From a9322f6488b432ddc1e89be88242c827c633fb63 Mon Sep 17 00:00:00 2001 From: Stefan Assmann Date: Wed, 11 Jun 2008 16:35:14 +0200 Subject: x86, pci: introduce pci=noioapicquirk kernel cmdline option Introduce pci=noioapicquirk kernel cmdline option to disable all boot interrupt quirks Signed-off-by: Stefan Assmann Signed-off-by: Olaf Dabrunz Signed-off-by: Ingo Molnar --- arch/x86/pci/common.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/pci/common.c b/arch/x86/pci/common.c index 940185ecaeda..bc6a101ed7ec 100644 --- a/arch/x86/pci/common.c +++ b/arch/x86/pci/common.c @@ -22,6 +22,7 @@ unsigned int pci_probe = PCI_PROBE_BIOS | PCI_PROBE_CONF1 | PCI_PROBE_CONF2 | static int pci_bf_sort; int pci_routeirq; +int noioapicquirk; int pcibios_last_bus = -1; unsigned long pirq_table_addr; struct pci_bus *pci_root_bus; @@ -495,6 +496,9 @@ char * __devinit pcibios_setup(char *str) } else if (!strcmp(str, "skip_isa_align")) { pci_probe |= PCI_CAN_SKIP_ISA_ALIGN; return NULL; + } else if (!strcmp(str, "noioapicquirk")) { + noioapicquirk = 1; + return NULL; } return str; } -- cgit v1.2.3 From 9197979b518573999d52d9e85bce1680682ed85c Mon Sep 17 00:00:00 2001 From: Stefan Assmann Date: Wed, 11 Jun 2008 16:35:15 +0200 Subject: x86, pci: introduce pci=ioapicreroute kernel cmdline option Introduce pci=ioapicreroute kernel cmdline option to enable rerouting of boot interrupts to the primary io-apic. Signed-off-by: Stefan Assmann Signed-off-by: Olaf Dabrunz Signed-off-by: Ingo Molnar --- Documentation/kernel-parameters.txt | 4 ++++ arch/x86/pci/common.c | 5 +++++ include/asm-x86/io_apic.h | 4 ++++ include/asm-x86/pci.h | 1 + 4 files changed, 14 insertions(+) (limited to 'arch/x86') diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index 1aebe9dffbaa..df262b3c3d6e 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -1521,6 +1521,10 @@ and is between 256 and 4096 characters. It is defined in the file noioapicquirk [APIC] Disable all boot interrupt quirks. Safety option to keep boot IRQs enabled. This should never be necessary. + ioapicreroute [APIC] Enable rerouting of boot IRQs to the + primary IO-APIC for bridges that cannot disable + boot IRQs. This fixes a source of spurious IRQs + when the system masks IRQs. biosirq [X86-32] Use PCI BIOS calls to get the interrupt routing table. These calls are known to be buggy on several machines and they hang the machine diff --git a/arch/x86/pci/common.c b/arch/x86/pci/common.c index bc6a101ed7ec..0a9eaa736d94 100644 --- a/arch/x86/pci/common.c +++ b/arch/x86/pci/common.c @@ -23,6 +23,7 @@ unsigned int pci_probe = PCI_PROBE_BIOS | PCI_PROBE_CONF1 | PCI_PROBE_CONF2 | static int pci_bf_sort; int pci_routeirq; int noioapicquirk; +int noioapicreroute = 1; int pcibios_last_bus = -1; unsigned long pirq_table_addr; struct pci_bus *pci_root_bus; @@ -499,6 +500,10 @@ char * __devinit pcibios_setup(char *str) } else if (!strcmp(str, "noioapicquirk")) { noioapicquirk = 1; return NULL; + } else if (!strcmp(str, "ioapicreroute")) { + if (noioapicreroute != -1) + noioapicreroute = 0; + return NULL; } return str; } diff --git a/include/asm-x86/io_apic.h b/include/asm-x86/io_apic.h index 8ca0110819f4..a39670ae17df 100644 --- a/include/asm-x86/io_apic.h +++ b/include/asm-x86/io_apic.h @@ -160,12 +160,16 @@ extern int skip_ioapic_setup; /* 1 if "noapic" boot option passed */ extern int noioapicquirk; +/* -1 if "noapic" boot option passed */ +extern int noioapicreroute; + /* 1 if the timer IRQ uses the '8259A Virtual Wire' mode */ extern int timer_through_8259; static inline void disable_ioapic_setup(void) { noioapicquirk = 1; + noioapicreroute = -1; skip_ioapic_setup = 1; } diff --git a/include/asm-x86/pci.h b/include/asm-x86/pci.h index 30eec93a845e..52a29f7668ef 100644 --- a/include/asm-x86/pci.h +++ b/include/asm-x86/pci.h @@ -20,6 +20,7 @@ struct pci_sysdata { extern int pci_routeirq; extern int noioapicquirk; +extern int ioapicreroute; /* scan a bus after allocating a pci_sysdata for it */ extern struct pci_bus *pci_scan_bus_on_node(int busno, struct pci_ops *ops, -- cgit v1.2.3 From 41b9eb264c8407655db57b60b4457fe1b2ec9977 Mon Sep 17 00:00:00 2001 From: Stefan Assmann Date: Tue, 15 Jul 2008 13:48:55 +0200 Subject: x86, pci: introduce config option for pci reroute quirks (was: [PATCH 0/3] Boot IRQ quirks for Broadcom and AMD/ATI) This is against linux-2.6-tip, branch pci-ioapic-boot-irq-quirks. From: Stefan Assmann Subject: Introduce config option for pci reroute quirks The config option X86_REROUTE_FOR_BROKEN_BOOT_IRQS is introduced to enable (or disable) the redirection of the interrupt handler to the boot interrupt line by default. Depending on the existence of interrupt masking / threaded interrupt handling in the kernel (vanilla, rt, ...) and the maturity of the rerouting patch, users can enable or disable the redirection by default. This means that the reroute quirk can be applied to any kernel without changing it. Interrupt sharing could be increased if this option is enabled. However this option is vital for threaded interrupt handling, as done by the RT kernel. It should simplify the consolidation with the RT kernel. The option can be overridden by either pci=ioapicreroute or pci=noioapicreroute. Signed-off-by: Stefan Assmann Signed-off-by: Olaf Dabrunz Cc: Jesse Barnes Cc: Jon Masters Cc: Ihno Krumreich Cc: Sven Dietrich Cc: Daniel Gollub Cc: Felix Foerster Signed-off-by: Ingo Molnar --- Documentation/kernel-parameters.txt | 4 ++++ arch/x86/Kconfig | 24 ++++++++++++++++++++++++ arch/x86/pci/common.c | 8 ++++++++ drivers/pci/quirks.c | 2 +- include/asm-x86/pci.h | 2 +- 5 files changed, 38 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index f5662b7a34d1..62b6e8067a5b 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -1536,6 +1536,10 @@ and is between 256 and 4096 characters. It is defined in the file primary IO-APIC for bridges that cannot disable boot IRQs. This fixes a source of spurious IRQs when the system masks IRQs. + noioapicreroute [APIC] Disable workaround that uses the + boot IRQ equivalent of an IRQ that connects to + a chipset where boot IRQs cannot be disabled. + The opposite of ioapicreroute. biosirq [X86-32] Use PCI BIOS calls to get the interrupt routing table. These calls are known to be buggy on several machines and they hang the machine diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 96e0c2ebc388..09521332636b 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -665,6 +665,30 @@ config X86_VISWS_APIC def_bool y depends on X86_32 && X86_VISWS +config X86_REROUTE_FOR_BROKEN_BOOT_IRQS + bool "Reroute for broken boot IRQs" + default n + depends on X86_IO_APIC + help + This option enables a workaround that fixes a source of + spurious interrupts. This is recommended when threaded + interrupt handling is used on systems where the generation of + superfluous "boot interrupts" cannot be disabled. + + Some chipsets generate a legacy INTx "boot IRQ" when the IRQ + entry in the chipset's IO-APIC is masked (as, e.g. the RT + kernel does during interrupt handling). On chipsets where this + boot IRQ generation cannot be disabled, this workaround keeps + the original IRQ line masked so that only the equivalent "boot + IRQ" is delivered to the CPUs. The workaround also tells the + kernel to set up the IRQ handler on the boot IRQ line. In this + way only one interrupt is delivered to the kernel. Otherwise + the spurious second interrupt may cause the kernel to bring + down (vital) interrupt lines. + + Only affects "broken" chipsets. Interrupt sharing may be + increased on these systems. + config X86_MCE bool "Machine Check Exception" depends on !X86_VOYAGER diff --git a/arch/x86/pci/common.c b/arch/x86/pci/common.c index 1485a26ddcef..bb1a01f089e2 100644 --- a/arch/x86/pci/common.c +++ b/arch/x86/pci/common.c @@ -24,7 +24,11 @@ unsigned int pci_early_dump_regs; static int pci_bf_sort; int pci_routeirq; int noioapicquirk; +#ifdef CONFIG_X86_REROUTE_FOR_BROKEN_BOOT_IRQS +int noioapicreroute = 0; +#else int noioapicreroute = 1; +#endif int pcibios_last_bus = -1; unsigned long pirq_table_addr; struct pci_bus *pci_root_bus; @@ -528,6 +532,10 @@ char * __devinit pcibios_setup(char *str) if (noioapicreroute != -1) noioapicreroute = 0; return NULL; + } else if (!strcmp(str, "noioapicreroute")) { + if (noioapicreroute != -1) + noioapicreroute = 1; + return NULL; } return str; } diff --git a/drivers/pci/quirks.c b/drivers/pci/quirks.c index 0911b0c60b64..c880dd0bbfb5 100644 --- a/drivers/pci/quirks.c +++ b/drivers/pci/quirks.c @@ -1397,7 +1397,7 @@ DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_INTEL, 0x260b, quirk_intel_pcie_pm); */ static void quirk_reroute_to_boot_interrupts_intel(struct pci_dev *dev) { - if (noioapicquirk) + if (noioapicquirk || noioapicreroute) return; dev->irq_reroute_variant = INTEL_IRQ_REROUTE_VARIANT; diff --git a/include/asm-x86/pci.h b/include/asm-x86/pci.h index 52a29f7668ef..9584d6d5eb93 100644 --- a/include/asm-x86/pci.h +++ b/include/asm-x86/pci.h @@ -20,7 +20,7 @@ struct pci_sysdata { extern int pci_routeirq; extern int noioapicquirk; -extern int ioapicreroute; +extern int noioapicreroute; /* scan a bus after allocating a pci_sysdata for it */ extern struct pci_bus *pci_scan_bus_on_node(int busno, struct pci_ops *ops, -- cgit v1.2.3 From c10d38dda1774ed4540380333cabd229eff37094 Mon Sep 17 00:00:00 2001 From: Nick Piggin Date: Wed, 10 Sep 2008 13:37:17 +0200 Subject: x86: some lock annotations for user copy paths copy_to/from_user and all its variants (except the atomic ones) can take a page fault and perform non-trivial work like taking mmap_sem and entering the filesyste/pagecache. Unfortunately, this often escapes lockdep because a common pattern is to use it to read in some arguments just set up from userspace, or write data back to a hot buffer. In those cases, it will be unlikely for page reclaim to get a window in to cause copy_*_user to fault. With the new might_lock primitives, add some annotations to x86. I don't know if I caught all possible faulting points (it's a bit of a maze, and I didn't really look at 32-bit). But this is a starting point. Boots and runs OK so far. Signed-off-by: Nick Piggin Acked-by: Peter Zijlstra Signed-off-by: Ingo Molnar --- arch/x86/lib/usercopy_32.c | 7 ++++++- arch/x86/lib/usercopy_64.c | 4 ++++ include/asm-x86/uaccess.h | 14 ++++++++++++++ include/asm-x86/uaccess_32.h | 10 ++++++++-- include/asm-x86/uaccess_64.h | 12 ++++++++++++ 5 files changed, 44 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/lib/usercopy_32.c b/arch/x86/lib/usercopy_32.c index 24e60944971a..8eedde2a9cac 100644 --- a/arch/x86/lib/usercopy_32.c +++ b/arch/x86/lib/usercopy_32.c @@ -33,6 +33,8 @@ static inline int __movsl_is_ok(unsigned long a1, unsigned long a2, unsigned lon do { \ int __d0, __d1, __d2; \ might_sleep(); \ + if (current->mm) \ + might_lock_read(¤t->mm->mmap_sem); \ __asm__ __volatile__( \ " testl %1,%1\n" \ " jz 2f\n" \ @@ -120,6 +122,8 @@ EXPORT_SYMBOL(strncpy_from_user); do { \ int __d0; \ might_sleep(); \ + if (current->mm) \ + might_lock_read(¤t->mm->mmap_sem); \ __asm__ __volatile__( \ "0: rep; stosl\n" \ " movl %2,%0\n" \ @@ -148,7 +152,6 @@ do { \ unsigned long clear_user(void __user *to, unsigned long n) { - might_sleep(); if (access_ok(VERIFY_WRITE, to, n)) __do_clear_user(to, n); return n; @@ -191,6 +194,8 @@ long strnlen_user(const char __user *s, long n) unsigned long res, tmp; might_sleep(); + if (current->mm) + might_lock_read(¤t->mm->mmap_sem); __asm__ __volatile__( " testl %0, %0\n" diff --git a/arch/x86/lib/usercopy_64.c b/arch/x86/lib/usercopy_64.c index f4df6e7c718b..847d12945998 100644 --- a/arch/x86/lib/usercopy_64.c +++ b/arch/x86/lib/usercopy_64.c @@ -16,6 +16,8 @@ do { \ long __d0, __d1, __d2; \ might_sleep(); \ + if (current->mm) \ + might_lock_read(¤t->mm->mmap_sem); \ __asm__ __volatile__( \ " testq %1,%1\n" \ " jz 2f\n" \ @@ -65,6 +67,8 @@ unsigned long __clear_user(void __user *addr, unsigned long size) { long __d0; might_sleep(); + if (current->mm) + might_lock_read(¤t->mm->mmap_sem); /* no memory constraint because it doesn't change any memory gcc knows about */ asm volatile( diff --git a/include/asm-x86/uaccess.h b/include/asm-x86/uaccess.h index 5f702d1d5218..ad29752a1713 100644 --- a/include/asm-x86/uaccess.h +++ b/include/asm-x86/uaccess.h @@ -8,6 +8,8 @@ #include #include #include +#include +#include #include #include @@ -157,6 +159,9 @@ extern int __get_user_bad(void); int __ret_gu; \ unsigned long __val_gu; \ __chk_user_ptr(ptr); \ + might_sleep(); \ + if (current->mm) \ + might_lock_read(¤t->mm->mmap_sem); \ switch (sizeof(*(ptr))) { \ case 1: \ __get_user_x(1, __ret_gu, __val_gu, ptr); \ @@ -241,6 +246,9 @@ extern void __put_user_8(void); int __ret_pu; \ __typeof__(*(ptr)) __pu_val; \ __chk_user_ptr(ptr); \ + might_sleep(); \ + if (current->mm) \ + might_lock_read(¤t->mm->mmap_sem); \ __pu_val = x; \ switch (sizeof(*(ptr))) { \ case 1: \ @@ -265,6 +273,9 @@ extern void __put_user_8(void); #define __put_user_size(x, ptr, size, retval, errret) \ do { \ retval = 0; \ + might_sleep(); \ + if (current->mm) \ + might_lock_read(¤t->mm->mmap_sem); \ __chk_user_ptr(ptr); \ switch (size) { \ case 1: \ @@ -317,6 +328,9 @@ do { \ #define __get_user_size(x, ptr, size, retval, errret) \ do { \ retval = 0; \ + might_sleep(); \ + if (current->mm) \ + might_lock_read(¤t->mm->mmap_sem); \ __chk_user_ptr(ptr); \ switch (size) { \ case 1: \ diff --git a/include/asm-x86/uaccess_32.h b/include/asm-x86/uaccess_32.h index 6fdef39a0bcb..d725e2d703f7 100644 --- a/include/asm-x86/uaccess_32.h +++ b/include/asm-x86/uaccess_32.h @@ -82,8 +82,10 @@ __copy_to_user_inatomic(void __user *to, const void *from, unsigned long n) static __always_inline unsigned long __must_check __copy_to_user(void __user *to, const void *from, unsigned long n) { - might_sleep(); - return __copy_to_user_inatomic(to, from, n); + might_sleep(); + if (current->mm) + might_lock_read(¤t->mm->mmap_sem); + return __copy_to_user_inatomic(to, from, n); } static __always_inline unsigned long @@ -138,6 +140,8 @@ static __always_inline unsigned long __copy_from_user(void *to, const void __user *from, unsigned long n) { might_sleep(); + if (current->mm) + might_lock_read(¤t->mm->mmap_sem); if (__builtin_constant_p(n)) { unsigned long ret; @@ -160,6 +164,8 @@ static __always_inline unsigned long __copy_from_user_nocache(void *to, const void __user *from, unsigned long n) { might_sleep(); + if (current->mm) + might_lock_read(¤t->mm->mmap_sem); if (__builtin_constant_p(n)) { unsigned long ret; diff --git a/include/asm-x86/uaccess_64.h b/include/asm-x86/uaccess_64.h index 515d4dce96b5..40a7205fe576 100644 --- a/include/asm-x86/uaccess_64.h +++ b/include/asm-x86/uaccess_64.h @@ -28,6 +28,10 @@ static __always_inline __must_check int __copy_from_user(void *dst, const void __user *src, unsigned size) { int ret = 0; + + might_sleep(); + if (current->mm) + might_lock_read(¤t->mm->mmap_sem); if (!__builtin_constant_p(size)) return copy_user_generic(dst, (__force void *)src, size); switch (size) { @@ -70,6 +74,10 @@ static __always_inline __must_check int __copy_to_user(void __user *dst, const void *src, unsigned size) { int ret = 0; + + might_sleep(); + if (current->mm) + might_lock_read(¤t->mm->mmap_sem); if (!__builtin_constant_p(size)) return copy_user_generic((__force void *)dst, src, size); switch (size) { @@ -112,6 +120,10 @@ static __always_inline __must_check int __copy_in_user(void __user *dst, const void __user *src, unsigned size) { int ret = 0; + + might_sleep(); + if (current->mm) + might_lock_read(¤t->mm->mmap_sem); if (!__builtin_constant_p(size)) return copy_user_generic((__force void *)dst, (__force void *)src, size); -- cgit v1.2.3 From 3ee1afa308f2a38e5d1e2ad3752ad7abcf480da1 Mon Sep 17 00:00:00 2001 From: Nick Piggin Date: Wed, 10 Sep 2008 13:37:17 +0200 Subject: x86: some lock annotations for user copy paths, v2 - introduce might_fault() - handle the atomic user copy paths correctly [ mingo@elte.hu: move might_sleep() outside of in_atomic(). ] Signed-off-by: Nick Piggin Acked-by: Peter Zijlstra Signed-off-by: Ingo Molnar --- arch/x86/lib/usercopy_32.c | 12 +++--------- arch/x86/lib/usercopy_64.c | 8 ++------ include/asm-x86/uaccess.h | 18 ++++-------------- include/asm-x86/uaccess_32.h | 12 +++--------- include/asm-x86/uaccess_64.h | 12 +++--------- include/linux/kernel.h | 9 +++++++++ mm/memory.c | 15 +++++++++++++++ 7 files changed, 39 insertions(+), 47 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/lib/usercopy_32.c b/arch/x86/lib/usercopy_32.c index 8eedde2a9cac..7393152a252e 100644 --- a/arch/x86/lib/usercopy_32.c +++ b/arch/x86/lib/usercopy_32.c @@ -32,9 +32,7 @@ static inline int __movsl_is_ok(unsigned long a1, unsigned long a2, unsigned lon #define __do_strncpy_from_user(dst, src, count, res) \ do { \ int __d0, __d1, __d2; \ - might_sleep(); \ - if (current->mm) \ - might_lock_read(¤t->mm->mmap_sem); \ + might_fault(); \ __asm__ __volatile__( \ " testl %1,%1\n" \ " jz 2f\n" \ @@ -121,9 +119,7 @@ EXPORT_SYMBOL(strncpy_from_user); #define __do_clear_user(addr,size) \ do { \ int __d0; \ - might_sleep(); \ - if (current->mm) \ - might_lock_read(¤t->mm->mmap_sem); \ + might_fault(); \ __asm__ __volatile__( \ "0: rep; stosl\n" \ " movl %2,%0\n" \ @@ -193,9 +189,7 @@ long strnlen_user(const char __user *s, long n) unsigned long mask = -__addr_ok(s); unsigned long res, tmp; - might_sleep(); - if (current->mm) - might_lock_read(¤t->mm->mmap_sem); + might_fault(); __asm__ __volatile__( " testl %0, %0\n" diff --git a/arch/x86/lib/usercopy_64.c b/arch/x86/lib/usercopy_64.c index 847d12945998..64d6c84e6353 100644 --- a/arch/x86/lib/usercopy_64.c +++ b/arch/x86/lib/usercopy_64.c @@ -15,9 +15,7 @@ #define __do_strncpy_from_user(dst,src,count,res) \ do { \ long __d0, __d1, __d2; \ - might_sleep(); \ - if (current->mm) \ - might_lock_read(¤t->mm->mmap_sem); \ + might_fault(); \ __asm__ __volatile__( \ " testq %1,%1\n" \ " jz 2f\n" \ @@ -66,9 +64,7 @@ EXPORT_SYMBOL(strncpy_from_user); unsigned long __clear_user(void __user *addr, unsigned long size) { long __d0; - might_sleep(); - if (current->mm) - might_lock_read(¤t->mm->mmap_sem); + might_fault(); /* no memory constraint because it doesn't change any memory gcc knows about */ asm volatile( diff --git a/include/asm-x86/uaccess.h b/include/asm-x86/uaccess.h index ad29752a1713..39f8420c75d9 100644 --- a/include/asm-x86/uaccess.h +++ b/include/asm-x86/uaccess.h @@ -8,8 +8,6 @@ #include #include #include -#include -#include #include #include @@ -159,9 +157,7 @@ extern int __get_user_bad(void); int __ret_gu; \ unsigned long __val_gu; \ __chk_user_ptr(ptr); \ - might_sleep(); \ - if (current->mm) \ - might_lock_read(¤t->mm->mmap_sem); \ + might_fault(); \ switch (sizeof(*(ptr))) { \ case 1: \ __get_user_x(1, __ret_gu, __val_gu, ptr); \ @@ -246,9 +242,7 @@ extern void __put_user_8(void); int __ret_pu; \ __typeof__(*(ptr)) __pu_val; \ __chk_user_ptr(ptr); \ - might_sleep(); \ - if (current->mm) \ - might_lock_read(¤t->mm->mmap_sem); \ + might_fault(); \ __pu_val = x; \ switch (sizeof(*(ptr))) { \ case 1: \ @@ -273,9 +267,7 @@ extern void __put_user_8(void); #define __put_user_size(x, ptr, size, retval, errret) \ do { \ retval = 0; \ - might_sleep(); \ - if (current->mm) \ - might_lock_read(¤t->mm->mmap_sem); \ + might_fault(); \ __chk_user_ptr(ptr); \ switch (size) { \ case 1: \ @@ -328,9 +320,7 @@ do { \ #define __get_user_size(x, ptr, size, retval, errret) \ do { \ retval = 0; \ - might_sleep(); \ - if (current->mm) \ - might_lock_read(¤t->mm->mmap_sem); \ + might_fault(); \ __chk_user_ptr(ptr); \ switch (size) { \ case 1: \ diff --git a/include/asm-x86/uaccess_32.h b/include/asm-x86/uaccess_32.h index d725e2d703f7..d10e842ec3ee 100644 --- a/include/asm-x86/uaccess_32.h +++ b/include/asm-x86/uaccess_32.h @@ -82,9 +82,7 @@ __copy_to_user_inatomic(void __user *to, const void *from, unsigned long n) static __always_inline unsigned long __must_check __copy_to_user(void __user *to, const void *from, unsigned long n) { - might_sleep(); - if (current->mm) - might_lock_read(¤t->mm->mmap_sem); + might_fault(); return __copy_to_user_inatomic(to, from, n); } @@ -139,9 +137,7 @@ __copy_from_user_inatomic(void *to, const void __user *from, unsigned long n) static __always_inline unsigned long __copy_from_user(void *to, const void __user *from, unsigned long n) { - might_sleep(); - if (current->mm) - might_lock_read(¤t->mm->mmap_sem); + might_fault(); if (__builtin_constant_p(n)) { unsigned long ret; @@ -163,9 +159,7 @@ __copy_from_user(void *to, const void __user *from, unsigned long n) static __always_inline unsigned long __copy_from_user_nocache(void *to, const void __user *from, unsigned long n) { - might_sleep(); - if (current->mm) - might_lock_read(¤t->mm->mmap_sem); + might_fault(); if (__builtin_constant_p(n)) { unsigned long ret; diff --git a/include/asm-x86/uaccess_64.h b/include/asm-x86/uaccess_64.h index 40a7205fe576..13fd56fbc3ab 100644 --- a/include/asm-x86/uaccess_64.h +++ b/include/asm-x86/uaccess_64.h @@ -29,9 +29,7 @@ int __copy_from_user(void *dst, const void __user *src, unsigned size) { int ret = 0; - might_sleep(); - if (current->mm) - might_lock_read(¤t->mm->mmap_sem); + might_fault(); if (!__builtin_constant_p(size)) return copy_user_generic(dst, (__force void *)src, size); switch (size) { @@ -75,9 +73,7 @@ int __copy_to_user(void __user *dst, const void *src, unsigned size) { int ret = 0; - might_sleep(); - if (current->mm) - might_lock_read(¤t->mm->mmap_sem); + might_fault(); if (!__builtin_constant_p(size)) return copy_user_generic((__force void *)dst, src, size); switch (size) { @@ -121,9 +117,7 @@ int __copy_in_user(void __user *dst, const void __user *src, unsigned size) { int ret = 0; - might_sleep(); - if (current->mm) - might_lock_read(¤t->mm->mmap_sem); + might_fault(); if (!__builtin_constant_p(size)) return copy_user_generic((__force void *)dst, (__force void *)src, size); diff --git a/include/linux/kernel.h b/include/linux/kernel.h index 2651f805ba6d..e580ec095765 100644 --- a/include/linux/kernel.h +++ b/include/linux/kernel.h @@ -140,6 +140,15 @@ extern int _cond_resched(void); (__x < 0) ? -__x : __x; \ }) +#ifdef CONFIG_PROVE_LOCKING +void might_fault(void); +#else +static inline void might_fault(void) +{ + might_sleep(); +} +#endif + extern struct atomic_notifier_head panic_notifier_list; extern long (*panic_blink)(long time); NORET_TYPE void panic(const char * fmt, ...) diff --git a/mm/memory.c b/mm/memory.c index 1002f473f497..b8fdf4e5e65b 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -3016,3 +3016,18 @@ void print_vma_addr(char *prefix, unsigned long ip) } up_read(¤t->mm->mmap_sem); } + +#ifdef CONFIG_PROVE_LOCKING +void might_fault(void) +{ + might_sleep(); + /* + * it would be nicer only to annotate paths which are not under + * pagefault_disable, however that requires a larger audit and + * providing helpers like get_user_atomic. + */ + if (!in_atomic() && current->mm) + might_lock_read(¤t->mm->mmap_sem); +} +EXPORT_SYMBOL(might_fault); +#endif -- cgit v1.2.3 From 1d18ef489509314506328b9e464dd47c24c1d68f Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Thu, 11 Sep 2008 20:53:21 +0200 Subject: x86: some lock annotations for user copy paths, v3 - add annotation back to clear_user() - change probe_kernel_address() to _inatomic*() method Signed-off-by: Ingo Molnar --- arch/x86/lib/usercopy_32.c | 1 + include/asm-x86/uaccess.h | 2 -- include/linux/uaccess.h | 2 +- 3 files changed, 2 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/lib/usercopy_32.c b/arch/x86/lib/usercopy_32.c index 7393152a252e..fab5faba1d3e 100644 --- a/arch/x86/lib/usercopy_32.c +++ b/arch/x86/lib/usercopy_32.c @@ -148,6 +148,7 @@ do { \ unsigned long clear_user(void __user *to, unsigned long n) { + might_fault(); if (access_ok(VERIFY_WRITE, to, n)) __do_clear_user(to, n); return n; diff --git a/include/asm-x86/uaccess.h b/include/asm-x86/uaccess.h index 39f8420c75d9..dc8edb5c4659 100644 --- a/include/asm-x86/uaccess.h +++ b/include/asm-x86/uaccess.h @@ -267,7 +267,6 @@ extern void __put_user_8(void); #define __put_user_size(x, ptr, size, retval, errret) \ do { \ retval = 0; \ - might_fault(); \ __chk_user_ptr(ptr); \ switch (size) { \ case 1: \ @@ -320,7 +319,6 @@ do { \ #define __get_user_size(x, ptr, size, retval, errret) \ do { \ retval = 0; \ - might_fault(); \ __chk_user_ptr(ptr); \ switch (size) { \ case 1: \ diff --git a/include/linux/uaccess.h b/include/linux/uaccess.h index fec6decfb983..2062293e57e6 100644 --- a/include/linux/uaccess.h +++ b/include/linux/uaccess.h @@ -78,7 +78,7 @@ static inline unsigned long __copy_from_user_nocache(void *to, \ set_fs(KERNEL_DS); \ pagefault_disable(); \ - ret = __get_user(retval, (__force typeof(retval) __user *)(addr)); \ + ret = __copy_from_user_inatomic((__force typeof(retval) __user *)(addr), &(retval), sizeof(retval)); \ pagefault_enable(); \ set_fs(old_fs); \ ret; \ -- cgit v1.2.3 From b0f209898f1a177bd503d49215b8c6628797a81c Mon Sep 17 00:00:00 2001 From: Russ Anderson Date: Tue, 21 Oct 2008 14:09:51 -0500 Subject: x86, uv: use consistent names for region size and conherence id on x86 and ia64 Use consistent names for region size and conherence id on x86 and ia64. The SGI xp drivers are used on both ia64 and x86. Using the same names (sn_coherency_id, sn_region_size) simplies the driver code. Signed-off-by: Russ Anderson Signed-off-by: Ingo Molnar --- arch/x86/kernel/bios_uv.c | 8 ++++---- arch/x86/kernel/genx2apic_uv_x.c | 4 ++-- include/asm-x86/uv/bios.h | 6 +++--- 3 files changed, 9 insertions(+), 9 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/bios_uv.c b/arch/x86/kernel/bios_uv.c index f0dfe6f17e7e..7cefb7170e75 100644 --- a/arch/x86/kernel/bios_uv.c +++ b/arch/x86/kernel/bios_uv.c @@ -69,10 +69,10 @@ s64 uv_bios_call_reentrant(enum uv_bios_cmd which, u64 a1, u64 a2, u64 a3, long sn_partition_id; EXPORT_SYMBOL_GPL(sn_partition_id); -long uv_coherency_id; -EXPORT_SYMBOL_GPL(uv_coherency_id); -long uv_region_size; -EXPORT_SYMBOL_GPL(uv_region_size); +long sn_coherency_id; +EXPORT_SYMBOL_GPL(sn_coherency_id); +long sn_region_size; +EXPORT_SYMBOL_GPL(sn_region_size); int uv_type; diff --git a/arch/x86/kernel/genx2apic_uv_x.c b/arch/x86/kernel/genx2apic_uv_x.c index bfd532843df6..6cf35c8bd636 100644 --- a/arch/x86/kernel/genx2apic_uv_x.c +++ b/arch/x86/kernel/genx2apic_uv_x.c @@ -429,7 +429,7 @@ void __init uv_system_init(void) uv_bios_init(); uv_bios_get_sn_info(0, &uv_type, &sn_partition_id, - &uv_coherency_id, &uv_region_size); + &sn_coherency_id, &sn_region_size); uv_rtc_init(); for_each_present_cpu(cpu) { @@ -451,7 +451,7 @@ void __init uv_system_init(void) uv_cpu_hub_info(cpu)->gpa_mask = (1 << (m_val + n_val)) - 1; uv_cpu_hub_info(cpu)->gnode_upper = gnode_upper; uv_cpu_hub_info(cpu)->global_mmr_base = mmr_base; - uv_cpu_hub_info(cpu)->coherency_domain_number = uv_coherency_id; + uv_cpu_hub_info(cpu)->coherency_domain_number = sn_coherency_id; uv_node_to_blade[nid] = blade; uv_cpu_to_blade[cpu] = blade; max_pnode = max(pnode, max_pnode); diff --git a/include/asm-x86/uv/bios.h b/include/asm-x86/uv/bios.h index 215f1969c266..7b3d7022c639 100644 --- a/include/asm-x86/uv/bios.h +++ b/include/asm-x86/uv/bios.h @@ -85,9 +85,9 @@ extern void uv_bios_init(void); extern int uv_type; extern long sn_partition_id; -extern long uv_coherency_id; -extern long uv_region_size; -#define partition_coherence_id() (uv_coherency_id) +extern long sn_coherency_id; +extern long sn_region_size; +#define partition_coherence_id() (sn_coherency_id) extern struct kobject *sgi_uv_kobj; /* /sys/firmware/sgi_uv */ -- cgit v1.2.3 From 9e899816d126cc6f7d405c349f65363214fe7399 Mon Sep 17 00:00:00 2001 From: Nick Piggin Date: Wed, 22 Oct 2008 12:33:16 +0200 Subject: x86, mm: enable GBPAGES option by default DIRECT_GBPAGES was under DEBUG_KERNEL && EXPERIMENTAL and disabled by default. Turn it on by default and put it under EMBEDDED. Signed-off-by: Nick Piggin Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 9 +++++++++ arch/x86/Kconfig.debug | 12 ------------ 2 files changed, 9 insertions(+), 12 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 5b9b12321ad1..c00aefcb47d5 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -946,6 +946,15 @@ config X86_PAE config ARCH_PHYS_ADDR_T_64BIT def_bool X86_64 || X86_PAE +config DIRECT_GBPAGES + bool "Enable 1GB pages for kernel pagetables" if EMBEDDED + default y + depends on X86_64 + help + Allow the kernel linear mapping to use 1GB pages on CPUs that + support it. This can improve the kernel's performance a tiny bit by + reducing TLB pressure. If in doubt, say "Y". + # Common NUMA Features config NUMA bool "Numa Memory Allocation and Scheduler Support (EXPERIMENTAL)" diff --git a/arch/x86/Kconfig.debug b/arch/x86/Kconfig.debug index 2a3dfbd5e677..567fe543e09c 100644 --- a/arch/x86/Kconfig.debug +++ b/arch/x86/Kconfig.debug @@ -114,18 +114,6 @@ config DEBUG_RODATA data. This is recommended so that we can catch kernel bugs sooner. If in doubt, say "Y". -config DIRECT_GBPAGES - bool "Enable gbpages-mapped kernel pagetables" - depends on DEBUG_KERNEL && EXPERIMENTAL && X86_64 - help - Enable gigabyte pages support (if the CPU supports it). This can - improve the kernel's performance a tiny bit by reducing TLB - pressure. - - This is experimental code. - - If in doubt, say "N". - config DEBUG_RODATA_TEST bool "Testcase for the DEBUG_RODATA feature" depends on DEBUG_RODATA -- cgit v1.2.3 From b4b8f87bf4958cbad620654efc0882ac46c19846 Mon Sep 17 00:00:00 2001 From: Alexander van Heukelum Date: Wed, 22 Oct 2008 12:00:08 +0200 Subject: i386, dumpstack: move crash_kexec before bust_spinlocks(0) in oops_end crash_kexec should not be called with console_sem held. Move the call before bust_spinlocks(0) in oops_end to avoid the problem. Signed-off-by: Alexander van Heukelum Acked-by: "Neil Horman" Signed-off-by: Ingo Molnar --- arch/x86/kernel/dumpstack_32.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/dumpstack_32.c b/arch/x86/kernel/dumpstack_32.c index b3614752197b..5493d31be4e5 100644 --- a/arch/x86/kernel/dumpstack_32.c +++ b/arch/x86/kernel/dumpstack_32.c @@ -309,6 +309,9 @@ unsigned __kprobes long oops_begin(void) void __kprobes oops_end(unsigned long flags, struct pt_regs *regs, int signr) { + if (regs && kexec_should_crash(current)) + crash_kexec(regs); + bust_spinlocks(0); die_owner = -1; add_taint(TAINT_DIE); @@ -318,8 +321,6 @@ void __kprobes oops_end(unsigned long flags, struct pt_regs *regs, int signr) if (!regs) return; - if (kexec_should_crash(current)) - crash_kexec(regs); if (in_interrupt()) panic("Fatal exception in interrupt"); if (panic_on_oops) -- cgit v1.2.3 From 874d93d11823b2b861addac6a5dc31162e924ab2 Mon Sep 17 00:00:00 2001 From: Alexander van Heukelum Date: Wed, 22 Oct 2008 12:00:09 +0200 Subject: x86, dumpstack: let signr=0 signal no do_exit Change oops_end such that signr=0 signals that do_exit is not to be called. Currently, each use of __die is soon followed by a call to oops_end and 'regs' is set to NULL if oops_end is expected not to call do_exit. Change all such pairs to set signr=0 instead. On x86_64 oops_end is used 'bare' in die_nmi; use signr=0 instead of regs=NULL there, too. Signed-off-by: Alexander van Heukelum Acked-by: Neil Horman Signed-off-by: Ingo Molnar --- arch/x86/kernel/dumpstack_32.c | 7 ++++--- arch/x86/kernel/dumpstack_64.c | 9 +++++---- arch/x86/mm/fault.c | 11 +++++++---- 3 files changed, 16 insertions(+), 11 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/dumpstack_32.c b/arch/x86/kernel/dumpstack_32.c index 5493d31be4e5..7c22f99f0efb 100644 --- a/arch/x86/kernel/dumpstack_32.c +++ b/arch/x86/kernel/dumpstack_32.c @@ -318,7 +318,7 @@ void __kprobes oops_end(unsigned long flags, struct pt_regs *regs, int signr) __raw_spin_unlock(&die_lock); raw_local_irq_restore(flags); - if (!regs) + if (!signr) return; if (in_interrupt()) @@ -371,17 +371,18 @@ int __kprobes __die(const char *str, struct pt_regs *regs, long err) void die(const char *str, struct pt_regs *regs, long err) { unsigned long flags = oops_begin(); + int sig = SIGSEGV; if (die_nest_count < 3) { report_bug(regs->ip, regs); if (__die(str, regs, err)) - regs = NULL; + sig = 0; } else { printk(KERN_EMERG "Recursive die() failure, output suppressed\n"); } - oops_end(flags, regs, SIGSEGV); + oops_end(flags, regs, sig); } static DEFINE_SPINLOCK(nmi_print_lock); diff --git a/arch/x86/kernel/dumpstack_64.c b/arch/x86/kernel/dumpstack_64.c index 96a5db7da8a7..ffefea611ba3 100644 --- a/arch/x86/kernel/dumpstack_64.c +++ b/arch/x86/kernel/dumpstack_64.c @@ -465,7 +465,7 @@ void __kprobes oops_end(unsigned long flags, struct pt_regs *regs, int signr) /* Nest count reaches zero, release the lock. */ __raw_spin_unlock(&die_lock); raw_local_irq_restore(flags); - if (!regs) { + if (!signr) { oops_exit(); return; } @@ -509,13 +509,14 @@ int __kprobes __die(const char *str, struct pt_regs *regs, long err) void die(const char *str, struct pt_regs *regs, long err) { unsigned long flags = oops_begin(); + int sig = SIGSEGV; if (!user_mode(regs)) report_bug(regs->ip, regs); if (__die(str, regs, err)) - regs = NULL; - oops_end(flags, regs, SIGSEGV); + sig = 0; + oops_end(flags, regs, sig); } notrace __kprobes void @@ -539,7 +540,7 @@ die_nmi(char *str, struct pt_regs *regs, int do_panic) crash_kexec(regs); if (do_panic || panic_on_oops) panic("Non maskable interrupt"); - oops_end(flags, NULL, SIGBUS); + oops_end(flags, regs, 0); nmi_exit(); local_irq_enable(); do_exit(SIGBUS); diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c index 31e8730fa246..20ef272c412c 100644 --- a/arch/x86/mm/fault.c +++ b/arch/x86/mm/fault.c @@ -413,6 +413,7 @@ static noinline void pgtable_bad(unsigned long address, struct pt_regs *regs, unsigned long error_code) { unsigned long flags = oops_begin(); + int sig = SIGKILL; struct task_struct *tsk; printk(KERN_ALERT "%s: Corrupted page table at address %lx\n", @@ -423,8 +424,8 @@ static noinline void pgtable_bad(unsigned long address, struct pt_regs *regs, tsk->thread.trap_no = 14; tsk->thread.error_code = error_code; if (__die("Bad pagetable", regs, error_code)) - regs = NULL; - oops_end(flags, regs, SIGKILL); + sig = 0; + oops_end(flags, regs, sig); } #endif @@ -590,6 +591,7 @@ void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code) int fault; #ifdef CONFIG_X86_64 unsigned long flags; + int sig; #endif tsk = current; @@ -849,11 +851,12 @@ no_context: bust_spinlocks(0); do_exit(SIGKILL); #else + sig = SIGKILL; if (__die("Oops", regs, error_code)) - regs = NULL; + sig = 0; /* Executive summary in case the body of the oops scrolled away */ printk(KERN_EMERG "CR2: %016lx\n", address); - oops_end(flags, regs, SIGKILL); + oops_end(flags, regs, sig); #endif /* -- cgit v1.2.3 From 0ed7a498f416dcfa1cca478a559238a2a3396240 Mon Sep 17 00:00:00 2001 From: Alexander van Heukelum Date: Wed, 22 Oct 2008 12:00:10 +0200 Subject: x86_64, dumpstack: move kexec_crash from __die to oops_end oops_end is preceded by either a call to __die, or a conditional call to crash_kexec. Move the conditional call to crash_kexec from the end of __die to the start of oops_end and remove the superfluous call to crash_kexec in die_nmi. Signed-off-by: Alexander van Heukelum Acked-by: Neil Horman Signed-off-by: Ingo Molnar --- arch/x86/kernel/dumpstack_64.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/dumpstack_64.c b/arch/x86/kernel/dumpstack_64.c index ffefea611ba3..57ce11b895ce 100644 --- a/arch/x86/kernel/dumpstack_64.c +++ b/arch/x86/kernel/dumpstack_64.c @@ -458,6 +458,9 @@ unsigned __kprobes long oops_begin(void) void __kprobes oops_end(unsigned long flags, struct pt_regs *regs, int signr) { + if (regs && kexec_should_crash(current)) + crash_kexec(regs); + die_owner = -1; bust_spinlocks(0); die_nest_count--; @@ -501,8 +504,6 @@ int __kprobes __die(const char *str, struct pt_regs *regs, long err) printk(KERN_ALERT "RIP "); printk_address(regs->ip, 1); printk(" RSP <%016lx>\n", regs->sp); - if (kexec_should_crash(current)) - crash_kexec(regs); return 0; } @@ -536,11 +537,9 @@ die_nmi(char *str, struct pt_regs *regs, int do_panic) printk(" on CPU%d, ip %08lx, registers:\n", smp_processor_id(), regs->ip); show_registers(regs); - if (kexec_should_crash(current)) - crash_kexec(regs); + oops_end(flags, regs, 0); if (do_panic || panic_on_oops) panic("Non maskable interrupt"); - oops_end(flags, regs, 0); nmi_exit(); local_irq_enable(); do_exit(SIGBUS); -- cgit v1.2.3 From 10b14cb7eb7dd5bff8023f76a55c8ac20e586128 Mon Sep 17 00:00:00 2001 From: Alexander van Heukelum Date: Wed, 22 Oct 2008 12:00:11 +0200 Subject: x86, dumpstack: always call oops_exit from oops_end Always call oops_exit from oops_end, even if signr==0. Also, move add_taint(TAINT_DIE) from __die to oops_end on x86_64 and interchange two lines to make oops_end more similar to the i386-version. Signed-off-by: Alexander van Heukelum Acked-by: Neil Horman Signed-off-by: Ingo Molnar --- arch/x86/kernel/dumpstack_32.c | 2 +- arch/x86/kernel/dumpstack_64.c | 11 +++++------ 2 files changed, 6 insertions(+), 7 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/dumpstack_32.c b/arch/x86/kernel/dumpstack_32.c index 7c22f99f0efb..a29b88ffa346 100644 --- a/arch/x86/kernel/dumpstack_32.c +++ b/arch/x86/kernel/dumpstack_32.c @@ -318,6 +318,7 @@ void __kprobes oops_end(unsigned long flags, struct pt_regs *regs, int signr) __raw_spin_unlock(&die_lock); raw_local_irq_restore(flags); + oops_exit(); if (!signr) return; @@ -325,7 +326,6 @@ void __kprobes oops_end(unsigned long flags, struct pt_regs *regs, int signr) panic("Fatal exception in interrupt"); if (panic_on_oops) panic("Fatal exception"); - oops_exit(); do_exit(signr); } diff --git a/arch/x86/kernel/dumpstack_64.c b/arch/x86/kernel/dumpstack_64.c index 57ce11b895ce..dc6162bf7454 100644 --- a/arch/x86/kernel/dumpstack_64.c +++ b/arch/x86/kernel/dumpstack_64.c @@ -461,22 +461,22 @@ void __kprobes oops_end(unsigned long flags, struct pt_regs *regs, int signr) if (regs && kexec_should_crash(current)) crash_kexec(regs); - die_owner = -1; bust_spinlocks(0); + die_owner = -1; + add_taint(TAINT_DIE); die_nest_count--; if (!die_nest_count) /* Nest count reaches zero, release the lock. */ __raw_spin_unlock(&die_lock); raw_local_irq_restore(flags); - if (!signr) { - oops_exit(); + oops_exit(); + + if (!signr) return; - } if (in_interrupt()) panic("Fatal exception in interrupt"); if (panic_on_oops) panic("Fatal exception"); - oops_exit(); do_exit(signr); } @@ -499,7 +499,6 @@ int __kprobes __die(const char *str, struct pt_regs *regs, long err) return 1; show_registers(regs); - add_taint(TAINT_DIE); /* Executive summary in case the oops scrolled away */ printk(KERN_ALERT "RIP "); printk_address(regs->ip, 1); -- cgit v1.2.3 From e4955cfd2f5c81eb708f55769aa60173f207fd63 Mon Sep 17 00:00:00 2001 From: Alexander van Heukelum Date: Wed, 22 Oct 2008 12:00:12 +0200 Subject: i386, dumpstack: use x86_64's method to account die_nest_count oops_begin/oops_end should always be used in pairs. On x86_64 oops_begin increments die_nest_count, and oops_end decrements die_nest_count. Doing this makes oops_begin and oops_end equal to the x86_64 versions. Signed-off-by: Alexander van Heukelum Acked-by: Neil Horman Signed-off-by: Ingo Molnar --- arch/x86/kernel/dumpstack_32.c | 29 +++++++++++++++++------------ 1 file changed, 17 insertions(+), 12 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/dumpstack_32.c b/arch/x86/kernel/dumpstack_32.c index a29b88ffa346..7c7d691b32be 100644 --- a/arch/x86/kernel/dumpstack_32.c +++ b/arch/x86/kernel/dumpstack_32.c @@ -289,21 +289,24 @@ static unsigned int die_nest_count; unsigned __kprobes long oops_begin(void) { + int cpu; unsigned long flags; oops_enter(); - if (die_owner != raw_smp_processor_id()) { - console_verbose(); - raw_local_irq_save(flags); - __raw_spin_lock(&die_lock); - die_owner = smp_processor_id(); - die_nest_count = 0; - bust_spinlocks(1); - } else { - raw_local_irq_save(flags); + /* racy, but better than risking deadlock. */ + raw_local_irq_save(flags); + cpu = smp_processor_id(); + if (!__raw_spin_trylock(&die_lock)) { + if (cpu == die_owner) + /* nested oops. should stop eventually */; + else + __raw_spin_lock(&die_lock); } die_nest_count++; + die_owner = cpu; + console_verbose(); + bust_spinlocks(1); return flags; } @@ -315,13 +318,15 @@ void __kprobes oops_end(unsigned long flags, struct pt_regs *regs, int signr) bust_spinlocks(0); die_owner = -1; add_taint(TAINT_DIE); - __raw_spin_unlock(&die_lock); + die_nest_count--; + if (!die_nest_count) + /* Nest count reaches zero, release the lock. */ + __raw_spin_unlock(&die_lock); raw_local_irq_restore(flags); - oops_exit(); + if (!signr) return; - if (in_interrupt()) panic("Fatal exception in interrupt"); if (panic_on_oops) -- cgit v1.2.3 From e06ca430c3d0fddbd1c901ab3fb3e1f0bc8a786b Mon Sep 17 00:00:00 2001 From: Alexander van Heukelum Date: Wed, 22 Oct 2008 12:00:13 +0200 Subject: i386, dumpstack: use oops_begin/oops_end in die_nmi Use oops_begin and oops_end in die_nmi. Whitespace-only changes on x86_64, to make it equal to i386's version. Signed-off-by: Alexander van Heukelum Acked-by: Neil Horman Signed-off-by: Ingo Molnar --- arch/x86/kernel/dumpstack_32.c | 33 +++++++++++---------------------- arch/x86/kernel/dumpstack_64.c | 4 ++-- 2 files changed, 13 insertions(+), 24 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/dumpstack_32.c b/arch/x86/kernel/dumpstack_32.c index 7c7d691b32be..e91ae34f9684 100644 --- a/arch/x86/kernel/dumpstack_32.c +++ b/arch/x86/kernel/dumpstack_32.c @@ -390,40 +390,29 @@ void die(const char *str, struct pt_regs *regs, long err) oops_end(flags, regs, sig); } -static DEFINE_SPINLOCK(nmi_print_lock); - void notrace __kprobes die_nmi(char *str, struct pt_regs *regs, int do_panic) { + unsigned long flags; + if (notify_die(DIE_NMIWATCHDOG, str, regs, 0, 2, SIGINT) == NOTIFY_STOP) return; - spin_lock(&nmi_print_lock); /* - * We are in trouble anyway, lets at least try - * to get a message out: - */ - bust_spinlocks(1); + * We are in trouble anyway, lets at least try + * to get a message out. + */ + flags = oops_begin(); printk(KERN_EMERG "%s", str); printk(" on CPU%d, ip %08lx, registers:\n", smp_processor_id(), regs->ip); show_registers(regs); - if (do_panic) + oops_end(flags, regs, 0); + if (do_panic || panic_on_oops) panic("Non maskable interrupt"); - console_silent(); - spin_unlock(&nmi_print_lock); - - /* - * If we are in kernel we are probably nested up pretty bad - * and might aswell get out now while we still can: - */ - if (!user_mode_vm(regs)) { - current->thread.trap_no = 2; - crash_kexec(regs); - } - - bust_spinlocks(0); - do_exit(SIGSEGV); + nmi_exit(); + local_irq_enable(); + do_exit(SIGBUS); } static int __init oops_setup(char *s) diff --git a/arch/x86/kernel/dumpstack_64.c b/arch/x86/kernel/dumpstack_64.c index dc6162bf7454..831e1e159cb4 100644 --- a/arch/x86/kernel/dumpstack_64.c +++ b/arch/x86/kernel/dumpstack_64.c @@ -519,7 +519,7 @@ void die(const char *str, struct pt_regs *regs, long err) oops_end(flags, regs, sig); } -notrace __kprobes void +void notrace __kprobes die_nmi(char *str, struct pt_regs *regs, int do_panic) { unsigned long flags; @@ -527,11 +527,11 @@ die_nmi(char *str, struct pt_regs *regs, int do_panic) if (notify_die(DIE_NMIWATCHDOG, str, regs, 0, 2, SIGINT) == NOTIFY_STOP) return; - flags = oops_begin(); /* * We are in trouble anyway, lets at least try * to get a message out. */ + flags = oops_begin(); printk(KERN_EMERG "%s", str); printk(" on CPU%d, ip %08lx, registers:\n", smp_processor_id(), regs->ip); -- cgit v1.2.3 From 871d3779cba18b028e34d0d2f6cc6caae76a97b6 Mon Sep 17 00:00:00 2001 From: Alexander van Heukelum Date: Wed, 22 Oct 2008 12:00:14 +0200 Subject: i386, dumpstack: unify die() Make i386's die() equal to x86_64's version. Whitespace-only changes on x86_64, to make it equal to i386's version. (user_mode and user_mode_vm are equal on x86_64.) Signed-off-by: Alexander van Heukelum Acked-by: Neil Horman Signed-off-by: Ingo Molnar --- arch/x86/kernel/dumpstack_32.c | 10 +++------- arch/x86/kernel/dumpstack_64.c | 6 +++++- 2 files changed, 8 insertions(+), 8 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/dumpstack_32.c b/arch/x86/kernel/dumpstack_32.c index e91ae34f9684..f2046c5752d0 100644 --- a/arch/x86/kernel/dumpstack_32.c +++ b/arch/x86/kernel/dumpstack_32.c @@ -378,15 +378,11 @@ void die(const char *str, struct pt_regs *regs, long err) unsigned long flags = oops_begin(); int sig = SIGSEGV; - if (die_nest_count < 3) { + if (!user_mode_vm(regs)) report_bug(regs->ip, regs); - if (__die(str, regs, err)) - sig = 0; - } else { - printk(KERN_EMERG "Recursive die() failure, output suppressed\n"); - } - + if (__die(str, regs, err)) + sig = 0; oops_end(flags, regs, sig); } diff --git a/arch/x86/kernel/dumpstack_64.c b/arch/x86/kernel/dumpstack_64.c index 831e1e159cb4..28c67aae5562 100644 --- a/arch/x86/kernel/dumpstack_64.c +++ b/arch/x86/kernel/dumpstack_64.c @@ -506,12 +506,16 @@ int __kprobes __die(const char *str, struct pt_regs *regs, long err) return 0; } +/* + * This is gone through when something in the kernel has done something bad + * and is about to be terminated: + */ void die(const char *str, struct pt_regs *regs, long err) { unsigned long flags = oops_begin(); int sig = SIGSEGV; - if (!user_mode(regs)) + if (!user_mode_vm(regs)) report_bug(regs->ip, regs); if (__die(str, regs, err)) -- cgit v1.2.3 From 63fb70859f987f2b3b8028fa467fd63336315e9c Mon Sep 17 00:00:00 2001 From: Sitsofe Wheeler Date: Sat, 11 Oct 2008 20:27:53 +0100 Subject: x86: change OPTIMIZE_INLINING help to say enabling makes smaller kernels Impact: clarify Kconfig help text The OPTIMIZE_INLINING help currently says "The gcc 4.x series have a rewritten inlining algorithm and disabling this option will generate a smaller kernel there." This contradicts other parts of the help text and my own tests: 5463127 2008-10-11 19:51 vmlinux.no-opt 5456152 2008-10-11 19:56 vmlinux.opt Reword text to say that enabling OPTIMIZE_INLINING will lead to smaller kernels with gcc 4.x or later. Signed-off-by: Sitsofe Wheeler Signed-off-by: Ingo Molnar --- arch/x86/Kconfig.debug | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig.debug b/arch/x86/Kconfig.debug index 2a3dfbd5e677..2be1e6b8e18b 100644 --- a/arch/x86/Kconfig.debug +++ b/arch/x86/Kconfig.debug @@ -307,10 +307,10 @@ config OPTIMIZE_INLINING developers have marked 'inline'. Doing so takes away freedom from gcc to do what it thinks is best, which is desirable for the gcc 3.x series of compilers. The gcc 4.x series have a rewritten inlining algorithm and - disabling this option will generate a smaller kernel there. Hopefully - this algorithm is so good that allowing gcc4 to make the decision can - become the default in the future, until then this option is there to - test gcc for this. + enabling this option will generate a smaller kernel there. Hopefully + this algorithm is so good that allowing gcc 4.x and above to make the + decision will become the default in the future. Until then this option + is there to test gcc for this. If unsure, say N. -- cgit v1.2.3 From 8479d94e9f6a44b5050cbacf653272a561fbe0d0 Mon Sep 17 00:00:00 2001 From: Mikael Pettersson Date: Mon, 27 Oct 2008 09:30:57 +0100 Subject: x86, signals: remove duplicated register setup code in ia32 signal delivery Impact: cleanup, no functionality changed ia32_setup_rt_frame() has a duplicated code block labelled "Make -mregparm=3 work" for setting up the register parameters to the user-mode signal handler. This is harmless but ugly. Remove the redundant assignments. Signed-off-by: Mikael Pettersson Signed-off-by: Ingo Molnar --- arch/x86/ia32/ia32_signal.c | 5 ----- 1 file changed, 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/ia32/ia32_signal.c b/arch/x86/ia32/ia32_signal.c index 4bc02b23674b..e82ebd652263 100644 --- a/arch/x86/ia32/ia32_signal.c +++ b/arch/x86/ia32/ia32_signal.c @@ -572,11 +572,6 @@ int ia32_setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, regs->dx = (unsigned long) &frame->info; regs->cx = (unsigned long) &frame->uc; - /* Make -mregparm=3 work */ - regs->ax = sig; - regs->dx = (unsigned long) &frame->info; - regs->cx = (unsigned long) &frame->uc; - loadsegment(ds, __USER32_DS); loadsegment(es, __USER32_DS); -- cgit v1.2.3 From 7f1baa063e2582dd52d83bb31508e9e84468c666 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Fri, 24 Oct 2008 15:24:29 -0700 Subject: x86/uv: provide a System Activity Indicator driver Impact: start per CPU heartbeat LED timers on SGI UV systems The SGI UV system has no LEDS but uses one of the system controller regs to indicate the online internal state of the cpu. There is a heartbeat bit indicating that the cpu is responding to interrupts, and an idle bit indicating whether the cpu is idle when the heartbeat interrupt occurs. The current period is one second. When a cpu panics, an error code is written by BIOS to this same reg. This patchset provides the following: * x86_64: Add base functionality for writing to the specific SCIR's for each cpu. * heartbeat: Invert "heartbeat" bit to indicate the cpu is "interruptible". If the current thread is the idle thread, then indicate system is "idle". * if hotplug enabled, all bits are set (0xff) when the cpu is disabled. Signed-off-by: Mike Travis Signed-off-by: Ingo Molnar --- arch/x86/include/asm/uv/uv_hub.h | 63 +++++++++++++++++++++++- arch/x86/kernel/genx2apic_uv_x.c | 102 +++++++++++++++++++++++++++++++++++++++ 2 files changed, 164 insertions(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/uv/uv_hub.h b/arch/x86/include/asm/uv/uv_hub.h index c6ad93e315c8..400776dba9b5 100644 --- a/arch/x86/include/asm/uv/uv_hub.h +++ b/arch/x86/include/asm/uv/uv_hub.h @@ -112,6 +112,16 @@ */ #define UV_MAX_NASID_VALUE (UV_MAX_NUMALINK_NODES * 2) +struct uv_scir_s { + struct timer_list timer; + unsigned long offset; + unsigned long last; + unsigned long idle_on; + unsigned long idle_off; + unsigned char state; + unsigned char enabled; +}; + /* * The following defines attributes of the HUB chip. These attributes are * frequently referenced and are kept in the per-cpu data areas of each cpu. @@ -130,7 +140,9 @@ struct uv_hub_info_s { unsigned char blade_processor_id; unsigned char m_val; unsigned char n_val; + struct uv_scir_s scir; }; + DECLARE_PER_CPU(struct uv_hub_info_s, __uv_hub_info); #define uv_hub_info (&__get_cpu_var(__uv_hub_info)) #define uv_cpu_hub_info(cpu) (&per_cpu(__uv_hub_info, cpu)) @@ -162,6 +174,30 @@ DECLARE_PER_CPU(struct uv_hub_info_s, __uv_hub_info); #define UV_APIC_PNODE_SHIFT 6 +/* Local Bus from cpu's perspective */ +#define LOCAL_BUS_BASE 0x1c00000 +#define LOCAL_BUS_SIZE (4 * 1024 * 1024) + +/* + * System Controller Interface Reg + * + * Note there are NO leds on a UV system. This register is only + * used by the system controller to monitor system-wide operation. + * There are 64 regs per node. With Nahelem cpus (2 cores per node, + * 8 cpus per core, 2 threads per cpu) there are 32 cpu threads on + * a node. + * + * The window is located at top of ACPI MMR space + */ +#define SCIR_WINDOW_COUNT 64 +#define SCIR_LOCAL_MMR_BASE (LOCAL_BUS_BASE + \ + LOCAL_BUS_SIZE - \ + SCIR_WINDOW_COUNT) + +#define SCIR_CPU_HEARTBEAT 0x01 /* timer interrupt */ +#define SCIR_CPU_ACTIVITY 0x02 /* not idle */ +#define SCIR_CPU_HB_INTERVAL (HZ) /* once per second */ + /* * Macros for converting between kernel virtual addresses, socket local physical * addresses, and UV global physical addresses. @@ -276,6 +312,16 @@ static inline void uv_write_local_mmr(unsigned long offset, unsigned long val) *uv_local_mmr_address(offset) = val; } +static inline unsigned char uv_read_local_mmr8(unsigned long offset) +{ + return *((unsigned char *)uv_local_mmr_address(offset)); +} + +static inline void uv_write_local_mmr8(unsigned long offset, unsigned char val) +{ + *((unsigned char *)uv_local_mmr_address(offset)) = val; +} + /* * Structures and definitions for converting between cpu, node, pnode, and blade * numbers. @@ -350,5 +396,20 @@ static inline int uv_num_possible_blades(void) return uv_possible_blades; } -#endif /* _ASM_X86_UV_UV_HUB_H */ +/* Update SCIR state */ +static inline void uv_set_scir_bits(unsigned char value) +{ + if (uv_hub_info->scir.state != value) { + uv_hub_info->scir.state = value; + uv_write_local_mmr8(uv_hub_info->scir.offset, value); + } +} +static inline void uv_set_cpu_scir_bits(int cpu, unsigned char value) +{ + if (uv_cpu_hub_info(cpu)->scir.state != value) { + uv_cpu_hub_info(cpu)->scir.state = value; + uv_write_local_mmr8(uv_cpu_hub_info(cpu)->scir.offset, value); + } +} +#endif /* _ASM_X86_UV_UV_HUB_H */ diff --git a/arch/x86/kernel/genx2apic_uv_x.c b/arch/x86/kernel/genx2apic_uv_x.c index dc6b46961523..84367d84bb10 100644 --- a/arch/x86/kernel/genx2apic_uv_x.c +++ b/arch/x86/kernel/genx2apic_uv_x.c @@ -10,6 +10,7 @@ #include #include +#include #include #include #include @@ -18,6 +19,8 @@ #include #include #include +#include +#include #include #include #include @@ -356,6 +359,103 @@ static __init void uv_rtc_init(void) sn_rtc_cycles_per_second = ticks_per_sec; } +/* + * percpu heartbeat timer + */ +static void uv_heartbeat(unsigned long ignored) +{ + struct timer_list *timer = &uv_hub_info->scir.timer; + unsigned char bits = uv_hub_info->scir.state; + + /* flip heartbeat bit */ + bits ^= SCIR_CPU_HEARTBEAT; + + /* are we the idle thread? */ + if (current->pid == 0) + bits &= ~SCIR_CPU_ACTIVITY; + else + bits |= SCIR_CPU_ACTIVITY; + + /* update system controller interface reg */ + uv_set_scir_bits(bits); + + /* enable next timer period */ + mod_timer(timer, jiffies + SCIR_CPU_HB_INTERVAL); +} + +static void __cpuinit uv_heartbeat_enable(int cpu) +{ + if (!uv_cpu_hub_info(cpu)->scir.enabled) { + struct timer_list *timer = &uv_cpu_hub_info(cpu)->scir.timer; + + uv_set_cpu_scir_bits(cpu, SCIR_CPU_HEARTBEAT|SCIR_CPU_ACTIVITY); + setup_timer(timer, uv_heartbeat, cpu); + timer->expires = jiffies + SCIR_CPU_HB_INTERVAL; + add_timer_on(timer, cpu); + uv_cpu_hub_info(cpu)->scir.enabled = 1; + } + + /* check boot cpu */ + if (!uv_cpu_hub_info(0)->scir.enabled) + uv_heartbeat_enable(0); +} + +static void __cpuinit uv_heartbeat_disable(int cpu) +{ + if (uv_cpu_hub_info(cpu)->scir.enabled) { + uv_cpu_hub_info(cpu)->scir.enabled = 0; + del_timer(&uv_cpu_hub_info(cpu)->scir.timer); + } + uv_set_cpu_scir_bits(cpu, 0xff); +} + +#ifdef CONFIG_HOTPLUG_CPU +/* + * cpu hotplug notifier + */ +static __cpuinit int uv_scir_cpu_notify(struct notifier_block *self, + unsigned long action, void *hcpu) +{ + long cpu = (long)hcpu; + + switch (action) { + case CPU_ONLINE: + uv_heartbeat_enable(cpu); + break; + case CPU_DOWN_PREPARE: + uv_heartbeat_disable(cpu); + break; + default: + break; + } + return NOTIFY_OK; +} + +static __init void uv_scir_register_cpu_notifier(void) +{ + hotcpu_notifier(uv_scir_cpu_notify, 0); +} + +#else /* !CONFIG_HOTPLUG_CPU */ + +static __init void uv_scir_register_cpu_notifier(void) +{ +} + +static __init int uv_init_heartbeat(void) +{ + int cpu; + + if (is_uv_system()) + for_each_online_cpu(cpu) + uv_heartbeat_enable(cpu); + return 0; +} + +late_initcall(uv_init_heartbeat); + +#endif /* !CONFIG_HOTPLUG_CPU */ + /* * Called on each cpu to initialize the per_cpu UV data area. * ZZZ hotplug not supported yet @@ -452,6 +552,7 @@ void __init uv_system_init(void) uv_cpu_hub_info(cpu)->gnode_upper = gnode_upper; uv_cpu_hub_info(cpu)->global_mmr_base = mmr_base; uv_cpu_hub_info(cpu)->coherency_domain_number = sn_coherency_id; + uv_cpu_hub_info(cpu)->scir.offset = SCIR_LOCAL_MMR_BASE + lcpu; uv_node_to_blade[nid] = blade; uv_cpu_to_blade[cpu] = blade; max_pnode = max(pnode, max_pnode); @@ -468,4 +569,5 @@ void __init uv_system_init(void) map_mmioh_high(max_pnode); uv_cpu_init(); + uv_scir_register_cpu_notifier(); } -- cgit v1.2.3 From fd3fdf11d3c649769e02459c5f1b8081a15e9007 Mon Sep 17 00:00:00 2001 From: Pekka Paalanen Date: Fri, 24 Oct 2008 20:08:11 +0300 Subject: trace: add the MMIO-tracer to the tracer menu, cleanup Impact: cleanup We can remove MMIOTRACE_HOOKS and replace it with just MMIOTRACE. MMIOTRACE_HOOKS is a remnant from the time when I thought that something else could also use the kmmio facilities. Signed-off-by: Pekka Paalanen Acked-by: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/x86/Kconfig.debug | 4 ---- arch/x86/mm/Makefile | 3 +-- arch/x86/mm/fault.c | 2 +- 3 files changed, 2 insertions(+), 7 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig.debug b/arch/x86/Kconfig.debug index 2a3dfbd5e677..fa013f529b74 100644 --- a/arch/x86/Kconfig.debug +++ b/arch/x86/Kconfig.debug @@ -186,14 +186,10 @@ config IOMMU_LEAK Add a simple leak tracer to the IOMMU code. This is useful when you are debugging a buggy device driver that leaks IOMMU mappings. -config MMIOTRACE_HOOKS - bool - config MMIOTRACE bool "Memory mapped IO tracing" depends on DEBUG_KERNEL && PCI select TRACING - select MMIOTRACE_HOOKS help Mmiotrace traces Memory Mapped I/O access and is meant for debugging and reverse engineering. It is called from the ioremap diff --git a/arch/x86/mm/Makefile b/arch/x86/mm/Makefile index 59f89b434b45..0a21b7aab9dc 100644 --- a/arch/x86/mm/Makefile +++ b/arch/x86/mm/Makefile @@ -8,9 +8,8 @@ obj-$(CONFIG_X86_PTDUMP) += dump_pagetables.o obj-$(CONFIG_HIGHMEM) += highmem_32.o -obj-$(CONFIG_MMIOTRACE_HOOKS) += kmmio.o obj-$(CONFIG_MMIOTRACE) += mmiotrace.o -mmiotrace-y := pf_in.o mmio-mod.o +mmiotrace-y := kmmio.o pf_in.o mmio-mod.o obj-$(CONFIG_MMIOTRACE_TEST) += testmmiotrace.o obj-$(CONFIG_NUMA) += numa_$(BITS).o diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c index 31e8730fa246..4152d3c3b138 100644 --- a/arch/x86/mm/fault.c +++ b/arch/x86/mm/fault.c @@ -53,7 +53,7 @@ static inline int kmmio_fault(struct pt_regs *regs, unsigned long addr) { -#ifdef CONFIG_MMIOTRACE_HOOKS +#ifdef CONFIG_MMIOTRACE if (unlikely(is_kmmio_active())) if (kmmio_handler(regs, addr) == 1) return -1; -- cgit v1.2.3 From 709110bd5624094992579f5311541f2e2b7ce58a Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Thu, 23 Oct 2008 17:14:25 -0700 Subject: x86: signal: cosmetic unification of restore_sigcontext() Impact: cleanup Make restore_sigcontext() the same. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/kernel/signal_32.c | 22 ++++++++++++++++++++++ arch/x86/kernel/signal_64.c | 15 +++++++++++++++ 2 files changed, 37 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/signal_32.c b/arch/x86/kernel/signal_32.c index d6dd057d0f22..85a0d37cdae9 100644 --- a/arch/x86/kernel/signal_32.c +++ b/arch/x86/kernel/signal_32.c @@ -149,14 +149,36 @@ restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc, /* Always make any pending restarted system calls return -EINTR */ current_thread_info()->restart_block.fn = do_no_restart_syscall; +#ifdef CONFIG_X86_32 GET_SEG(gs); COPY_SEG(fs); COPY_SEG(es); COPY_SEG(ds); +#endif /* CONFIG_X86_32 */ + COPY(di); COPY(si); COPY(bp); COPY(sp); COPY(bx); COPY(dx); COPY(cx); COPY(ip); + +#ifdef CONFIG_X86_64 + COPY(r8); + COPY(r9); + COPY(r10); + COPY(r11); + COPY(r12); + COPY(r13); + COPY(r14); + COPY(r15); +#endif /* CONFIG_X86_64 */ + +#ifdef CONFIG_X86_32 COPY_SEG_STRICT(cs); COPY_SEG_STRICT(ss); +#else /* !CONFIG_X86_32 */ + /* Kernel saves and restores only the CS segment register on signals, + * which is the bare minimum needed to allow mixed 32/64-bit code. + * App's signal handler can save/restore other segments if needed. */ + COPY_SEG_STRICT(cs); +#endif /* CONFIG_X86_32 */ err |= __get_user(tmpflags, &sc->flags); regs->flags = (regs->flags & ~FIX_EFLAGS) | (tmpflags & FIX_EFLAGS); diff --git a/arch/x86/kernel/signal_64.c b/arch/x86/kernel/signal_64.c index a5c9627f4db9..9c469da7f9e8 100644 --- a/arch/x86/kernel/signal_64.c +++ b/arch/x86/kernel/signal_64.c @@ -76,8 +76,17 @@ restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc, /* Always make any pending restarted system calls return -EINTR */ current_thread_info()->restart_block.fn = do_no_restart_syscall; +#ifdef CONFIG_X86_32 + GET_SEG(gs); + COPY_SEG(fs); + COPY_SEG(es); + COPY_SEG(ds); +#endif /* CONFIG_X86_32 */ + COPY(di); COPY(si); COPY(bp); COPY(sp); COPY(bx); COPY(dx); COPY(cx); COPY(ip); + +#ifdef CONFIG_X86_64 COPY(r8); COPY(r9); COPY(r10); @@ -86,11 +95,17 @@ restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc, COPY(r13); COPY(r14); COPY(r15); +#endif /* CONFIG_X86_64 */ +#ifdef CONFIG_X86_32 + COPY_SEG_STRICT(cs); + COPY_SEG_STRICT(ss); +#else /* !CONFIG_X86_32 */ /* Kernel saves and restores only the CS segment register on signals, * which is the bare minimum needed to allow mixed 32/64-bit code. * App's signal handler can save/restore other segments if needed. */ COPY_SEG_STRICT(cs); +#endif /* CONFIG_X86_32 */ err |= __get_user(tmpflags, &sc->flags); regs->flags = (regs->flags & ~FIX_EFLAGS) | (tmpflags & FIX_EFLAGS); -- cgit v1.2.3 From fd4a2030a358b4818646031049d9631bd45b9915 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Thu, 23 Oct 2008 17:15:28 -0700 Subject: x86: signal_64.c: get_stack() doesn't need entire regs Impact: cleanup get_stack() uses sp only, entire regs is not needed. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/kernel/signal_64.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/signal_64.c b/arch/x86/kernel/signal_64.c index 9c469da7f9e8..3d0deb336745 100644 --- a/arch/x86/kernel/signal_64.c +++ b/arch/x86/kernel/signal_64.c @@ -199,12 +199,10 @@ setup_sigcontext(struct sigcontext __user *sc, struct pt_regs *regs, */ static void __user * -get_stack(struct k_sigaction *ka, struct pt_regs *regs, unsigned long size) +get_stack(struct k_sigaction *ka, unsigned long sp, unsigned long size) { - unsigned long sp; - /* Default to using normal stack - redzone*/ - sp = regs->sp - 128; + sp -= 128; /* This is the X/Open sanctioned signal stack switching. */ if (ka->sa.sa_flags & SA_ONSTACK) { @@ -224,14 +222,14 @@ static int __setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, struct task_struct *me = current; if (used_math()) { - fp = get_stack(ka, regs, sig_xstate_size); + fp = get_stack(ka, regs->sp, sig_xstate_size); frame = (void __user *)round_down( (unsigned long)fp - sizeof(struct rt_sigframe), 16) - 8; if (save_i387_xstate(fp) < 0) return -EFAULT; } else - frame = get_stack(ka, regs, sizeof(struct rt_sigframe)) - 8; + frame = get_stack(ka, regs->sp, sizeof(struct rt_sigframe)) - 8; if (!access_ok(VERIFY_WRITE, frame, sizeof(*frame))) return -EFAULT; -- cgit v1.2.3 From c63dfefd48d92b1db3400fe8de4886a519ac3949 Mon Sep 17 00:00:00 2001 From: Dan McGee Date: Thu, 23 Oct 2008 15:44:02 -0500 Subject: x86: remove dead IRQBALANCE code Impact: cleanup CONFIG_IRQBALANCE was removed in commit 8b8e8c1bf; this ifdef was still around. Signed-off-by: Dan McGee Signed-off-by: Ingo Molnar --- arch/x86/include/asm/irq.h | 4 ---- 1 file changed, 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/irq.h b/arch/x86/include/asm/irq.h index bae0eda95486..28e409fc73f3 100644 --- a/arch/x86/include/asm/irq.h +++ b/arch/x86/include/asm/irq.h @@ -31,10 +31,6 @@ static inline int irq_canonicalize(int irq) # endif #endif -#ifdef CONFIG_IRQBALANCE -extern int irqbalance_disable(char *str); -#endif - #ifdef CONFIG_HOTPLUG_CPU #include extern void fixup_irqs(cpumask_t map); -- cgit v1.2.3 From 04d2aac33eb54fd3084140f2db130530d71e97c6 Mon Sep 17 00:00:00 2001 From: Arjan van de Ven Date: Sun, 5 Oct 2008 11:08:10 -0700 Subject: x86: corruption-check: fix some style issues Impact: cleanup Before moving the code to it's own file, fix some style issues in the corruption check code. Signed-off-by: Arjan van de Ven Signed-off-by: Ingo Molnar --- arch/x86/kernel/setup.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index 0fa6790c1dd3..4f38e0305b07 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -662,7 +662,7 @@ static void __init setup_bios_corruption_check(void) corruption_check_size = round_up(corruption_check_size, PAGE_SIZE); - while(addr < corruption_check_size && num_scan_areas < MAX_SCAN_AREAS) { + while (addr < corruption_check_size && num_scan_areas < MAX_SCAN_AREAS) { u64 size; addr = find_e820_area_size(addr, &size, PAGE_SIZE); @@ -701,11 +701,11 @@ void check_for_bios_corruption(void) if (!memory_corruption_check) return; - for(i = 0; i < num_scan_areas; i++) { + for (i = 0; i < num_scan_areas; i++) { unsigned long *addr = __va(scan_areas[i].addr); unsigned long size = scan_areas[i].size; - for(; size; addr++, size -= sizeof(unsigned long)) { + for (; size; addr++, size -= sizeof(unsigned long)) { if (!*addr) continue; printk(KERN_ERR "Corrupted low memory at %p (%lx phys) = %08lx\n", @@ -721,7 +721,8 @@ void check_for_bios_corruption(void) static void periodic_check_for_corruption(unsigned long data) { check_for_bios_corruption(); - mod_timer(&periodic_check_timer, round_jiffies(jiffies + corruption_check_period*HZ)); + mod_timer(&periodic_check_timer, + round_jiffies(jiffies + corruption_check_period*HZ)); } void start_periodic_check_for_corruption(void) -- cgit v1.2.3 From 6784f7d0a5016a397d38be1134e63fc784c1ca8e Mon Sep 17 00:00:00 2001 From: Arjan van de Ven Date: Sun, 5 Oct 2008 11:33:42 -0700 Subject: x86: corruption check: move the corruption checks into their own file Impact: cleanup The corruption check code is rather sizable and it's likely to grow over time when we add checks for more types of corruptions (there's a few candidates in kerneloops.org that I want to add checks for)... so lets move it to its own file Signed-off-by: Arjan van de Ven Signed-off-by: Ingo Molnar --- arch/x86/include/asm/setup.h | 4 ++ arch/x86/kernel/Makefile | 1 + arch/x86/kernel/check.c | 158 +++++++++++++++++++++++++++++++++++++++++++ arch/x86/kernel/setup.c | 152 ----------------------------------------- 4 files changed, 163 insertions(+), 152 deletions(-) create mode 100644 arch/x86/kernel/check.c (limited to 'arch/x86') diff --git a/arch/x86/include/asm/setup.h b/arch/x86/include/asm/setup.h index f12d37237465..1ed8b2e80727 100644 --- a/arch/x86/include/asm/setup.h +++ b/arch/x86/include/asm/setup.h @@ -8,6 +8,10 @@ /* Interrupt control for vSMPowered x86_64 systems */ void vsmp_init(void); + +void setup_bios_corruption_check(void); + + #ifdef CONFIG_X86_VISWS extern void visws_early_detect(void); extern int is_visws_box(void); diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index d7e5a58ee22f..31fbcaf3df70 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -35,6 +35,7 @@ obj-y += bootflag.o e820.o obj-y += pci-dma.o quirks.o i8237.o topology.o kdebugfs.o obj-y += alternative.o i8253.o pci-nommu.o obj-y += tsc.o io_delay.o rtc.o +obj-y += check.o obj-$(CONFIG_X86_TRAMPOLINE) += trampoline.o obj-y += process.o diff --git a/arch/x86/kernel/check.c b/arch/x86/kernel/check.c new file mode 100644 index 000000000000..5056703e1b05 --- /dev/null +++ b/arch/x86/kernel/check.c @@ -0,0 +1,158 @@ +#include +#include + +#include +#include + +/* + * Some BIOSes seem to corrupt the low 64k of memory during events + * like suspend/resume and unplugging an HDMI cable. Reserve all + * remaining free memory in that area and fill it with a distinct + * pattern. + */ +#ifdef CONFIG_X86_CHECK_BIOS_CORRUPTION +#define MAX_SCAN_AREAS 8 + +static int __read_mostly memory_corruption_check = -1; + +static unsigned __read_mostly corruption_check_size = 64*1024; +static unsigned __read_mostly corruption_check_period = 60; /* seconds */ + +static struct e820entry scan_areas[MAX_SCAN_AREAS]; +static int num_scan_areas; + + +static int set_corruption_check(char *arg) +{ + char *end; + + memory_corruption_check = simple_strtol(arg, &end, 10); + + return (*end == 0) ? 0 : -EINVAL; +} +early_param("memory_corruption_check", set_corruption_check); + +static int set_corruption_check_period(char *arg) +{ + char *end; + + corruption_check_period = simple_strtoul(arg, &end, 10); + + return (*end == 0) ? 0 : -EINVAL; +} +early_param("memory_corruption_check_period", set_corruption_check_period); + +static int set_corruption_check_size(char *arg) +{ + char *end; + unsigned size; + + size = memparse(arg, &end); + + if (*end == '\0') + corruption_check_size = size; + + return (size == corruption_check_size) ? 0 : -EINVAL; +} +early_param("memory_corruption_check_size", set_corruption_check_size); + + +void __init setup_bios_corruption_check(void) +{ + u64 addr = PAGE_SIZE; /* assume first page is reserved anyway */ + + if (memory_corruption_check == -1) { + memory_corruption_check = +#ifdef CONFIG_X86_BOOTPARAM_MEMORY_CORRUPTION_CHECK + 1 +#else + 0 +#endif + ; + } + + if (corruption_check_size == 0) + memory_corruption_check = 0; + + if (!memory_corruption_check) + return; + + corruption_check_size = round_up(corruption_check_size, PAGE_SIZE); + + while (addr < corruption_check_size && num_scan_areas < MAX_SCAN_AREAS) { + u64 size; + addr = find_e820_area_size(addr, &size, PAGE_SIZE); + + if (addr == 0) + break; + + if ((addr + size) > corruption_check_size) + size = corruption_check_size - addr; + + if (size == 0) + break; + + e820_update_range(addr, size, E820_RAM, E820_RESERVED); + scan_areas[num_scan_areas].addr = addr; + scan_areas[num_scan_areas].size = size; + num_scan_areas++; + + /* Assume we've already mapped this early memory */ + memset(__va(addr), 0, size); + + addr += size; + } + + printk(KERN_INFO "Scanning %d areas for low memory corruption\n", + num_scan_areas); + update_e820(); +} + +static struct timer_list periodic_check_timer; + +void check_for_bios_corruption(void) +{ + int i; + int corruption = 0; + + if (!memory_corruption_check) + return; + + for (i = 0; i < num_scan_areas; i++) { + unsigned long *addr = __va(scan_areas[i].addr); + unsigned long size = scan_areas[i].size; + + for (; size; addr++, size -= sizeof(unsigned long)) { + if (!*addr) + continue; + printk(KERN_ERR "Corrupted low memory at %p (%lx phys) = %08lx\n", + addr, __pa(addr), *addr); + corruption = 1; + *addr = 0; + } + } + + WARN(corruption, KERN_ERR "Memory corruption detected in low memory\n"); +} + +static void periodic_check_for_corruption(unsigned long data) +{ + check_for_bios_corruption(); + mod_timer(&periodic_check_timer, + round_jiffies(jiffies + corruption_check_period*HZ)); +} + +void start_periodic_check_for_corruption(void) +{ + if (!memory_corruption_check || corruption_check_period == 0) + return; + + printk(KERN_INFO "Scanning for low memory corruption every %d seconds\n", + corruption_check_period); + + init_timer(&periodic_check_timer); + periodic_check_timer.function = &periodic_check_for_corruption; + periodic_check_for_corruption(0); +} +#endif + diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index 4f38e0305b07..af690aa593a9 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -587,158 +587,6 @@ static struct x86_quirks default_x86_quirks __initdata; struct x86_quirks *x86_quirks __initdata = &default_x86_quirks; -/* - * Some BIOSes seem to corrupt the low 64k of memory during events - * like suspend/resume and unplugging an HDMI cable. Reserve all - * remaining free memory in that area and fill it with a distinct - * pattern. - */ -#ifdef CONFIG_X86_CHECK_BIOS_CORRUPTION -#define MAX_SCAN_AREAS 8 - -static int __read_mostly memory_corruption_check = -1; - -static unsigned __read_mostly corruption_check_size = 64*1024; -static unsigned __read_mostly corruption_check_period = 60; /* seconds */ - -static struct e820entry scan_areas[MAX_SCAN_AREAS]; -static int num_scan_areas; - - -static int set_corruption_check(char *arg) -{ - char *end; - - memory_corruption_check = simple_strtol(arg, &end, 10); - - return (*end == 0) ? 0 : -EINVAL; -} -early_param("memory_corruption_check", set_corruption_check); - -static int set_corruption_check_period(char *arg) -{ - char *end; - - corruption_check_period = simple_strtoul(arg, &end, 10); - - return (*end == 0) ? 0 : -EINVAL; -} -early_param("memory_corruption_check_period", set_corruption_check_period); - -static int set_corruption_check_size(char *arg) -{ - char *end; - unsigned size; - - size = memparse(arg, &end); - - if (*end == '\0') - corruption_check_size = size; - - return (size == corruption_check_size) ? 0 : -EINVAL; -} -early_param("memory_corruption_check_size", set_corruption_check_size); - - -static void __init setup_bios_corruption_check(void) -{ - u64 addr = PAGE_SIZE; /* assume first page is reserved anyway */ - - if (memory_corruption_check == -1) { - memory_corruption_check = -#ifdef CONFIG_X86_BOOTPARAM_MEMORY_CORRUPTION_CHECK - 1 -#else - 0 -#endif - ; - } - - if (corruption_check_size == 0) - memory_corruption_check = 0; - - if (!memory_corruption_check) - return; - - corruption_check_size = round_up(corruption_check_size, PAGE_SIZE); - - while (addr < corruption_check_size && num_scan_areas < MAX_SCAN_AREAS) { - u64 size; - addr = find_e820_area_size(addr, &size, PAGE_SIZE); - - if (addr == 0) - break; - - if ((addr + size) > corruption_check_size) - size = corruption_check_size - addr; - - if (size == 0) - break; - - e820_update_range(addr, size, E820_RAM, E820_RESERVED); - scan_areas[num_scan_areas].addr = addr; - scan_areas[num_scan_areas].size = size; - num_scan_areas++; - - /* Assume we've already mapped this early memory */ - memset(__va(addr), 0, size); - - addr += size; - } - - printk(KERN_INFO "Scanning %d areas for low memory corruption\n", - num_scan_areas); - update_e820(); -} - -static struct timer_list periodic_check_timer; - -void check_for_bios_corruption(void) -{ - int i; - int corruption = 0; - - if (!memory_corruption_check) - return; - - for (i = 0; i < num_scan_areas; i++) { - unsigned long *addr = __va(scan_areas[i].addr); - unsigned long size = scan_areas[i].size; - - for (; size; addr++, size -= sizeof(unsigned long)) { - if (!*addr) - continue; - printk(KERN_ERR "Corrupted low memory at %p (%lx phys) = %08lx\n", - addr, __pa(addr), *addr); - corruption = 1; - *addr = 0; - } - } - - WARN(corruption, KERN_ERR "Memory corruption detected in low memory\n"); -} - -static void periodic_check_for_corruption(unsigned long data) -{ - check_for_bios_corruption(); - mod_timer(&periodic_check_timer, - round_jiffies(jiffies + corruption_check_period*HZ)); -} - -void start_periodic_check_for_corruption(void) -{ - if (!memory_corruption_check || corruption_check_period == 0) - return; - - printk(KERN_INFO "Scanning for low memory corruption every %d seconds\n", - corruption_check_period); - - init_timer(&periodic_check_timer); - periodic_check_timer.function = &periodic_check_for_corruption; - periodic_check_for_corruption(0); -} -#endif - static int __init dmi_low_memory_corruption(const struct dmi_system_id *d) { printk(KERN_NOTICE -- cgit v1.2.3 From 304e629bf4a3150a0bf6556fc45c52c5c082340f Mon Sep 17 00:00:00 2001 From: Arjan van de Ven Date: Sun, 5 Oct 2008 12:09:03 -0700 Subject: x86: corruption check: run the corruption checks from a work queue Impact: change the implementation of the debug feature the periodic corruption checks are better off run from a work queue; there's nothing time critical about them and this way the amount of interrupt-context work is reduced. Signed-off-by: Arjan van de Ven Signed-off-by: Ingo Molnar --- arch/x86/kernel/check.c | 27 +++++++++++++++++---------- arch/x86/mm/init_32.c | 2 -- arch/x86/mm/init_64.c | 2 -- 3 files changed, 17 insertions(+), 14 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/check.c b/arch/x86/kernel/check.c index 5056703e1b05..55eed1752b43 100644 --- a/arch/x86/kernel/check.c +++ b/arch/x86/kernel/check.c @@ -1,6 +1,7 @@ #include #include - +#include +#include #include #include @@ -108,13 +109,14 @@ void __init setup_bios_corruption_check(void) update_e820(); } -static struct timer_list periodic_check_timer; void check_for_bios_corruption(void) { int i; int corruption = 0; + printk("dot\n"); + if (!memory_corruption_check) return; @@ -135,24 +137,29 @@ void check_for_bios_corruption(void) WARN(corruption, KERN_ERR "Memory corruption detected in low memory\n"); } -static void periodic_check_for_corruption(unsigned long data) +static void check_corruption(struct work_struct *dummy); +static DECLARE_DELAYED_WORK(bios_check_work, check_corruption); + +static void check_corruption(struct work_struct *dummy) { check_for_bios_corruption(); - mod_timer(&periodic_check_timer, - round_jiffies(jiffies + corruption_check_period*HZ)); + schedule_delayed_work(&bios_check_work, + round_jiffies_relative(corruption_check_period*HZ)); } -void start_periodic_check_for_corruption(void) +static int start_periodic_check_for_corruption(void) { if (!memory_corruption_check || corruption_check_period == 0) - return; + return 0; printk(KERN_INFO "Scanning for low memory corruption every %d seconds\n", corruption_check_period); - init_timer(&periodic_check_timer); - periodic_check_timer.function = &periodic_check_for_corruption; - periodic_check_for_corruption(0); + /* First time we run the checks right away */ + schedule_delayed_work(&bios_check_work, 0); + return 0; } + +module_init(start_periodic_check_for_corruption); #endif diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c index 8396868e82c5..5e6377560ff1 100644 --- a/arch/x86/mm/init_32.c +++ b/arch/x86/mm/init_32.c @@ -970,8 +970,6 @@ void __init mem_init(void) int codesize, reservedpages, datasize, initsize; int tmp; - start_periodic_check_for_corruption(); - #ifdef CONFIG_FLATMEM BUG_ON(!mem_map); #endif diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c index b8e461d49412..d6ef1589b95a 100644 --- a/arch/x86/mm/init_64.c +++ b/arch/x86/mm/init_64.c @@ -879,8 +879,6 @@ void __init mem_init(void) { long codesize, reservedpages, datasize, initsize; - start_periodic_check_for_corruption(); - pci_iommu_alloc(); /* clear_bss() already clear the empty_zero_page */ -- cgit v1.2.3 From b43d196c4d3fe46d6dda7c987c47792612b80b1b Mon Sep 17 00:00:00 2001 From: Arjan van de Ven Date: Sun, 5 Oct 2008 12:21:32 -0700 Subject: x86: corruption-check: some post-move cleanups Impact: cleanup now that the code is moved and converted to a work queue, there's some minor cleanups that can be done. Signed-off-by: Arjan van de Ven Signed-off-by: Ingo Molnar --- arch/x86/kernel/Makefile | 3 ++- arch/x86/kernel/check.c | 12 ++++-------- 2 files changed, 6 insertions(+), 9 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index 31fbcaf3df70..f63a8034fb82 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -35,7 +35,6 @@ obj-y += bootflag.o e820.o obj-y += pci-dma.o quirks.o i8237.o topology.o kdebugfs.o obj-y += alternative.o i8253.o pci-nommu.o obj-y += tsc.o io_delay.o rtc.o -obj-y += check.o obj-$(CONFIG_X86_TRAMPOLINE) += trampoline.o obj-y += process.o @@ -105,6 +104,8 @@ microcode-$(CONFIG_MICROCODE_INTEL) += microcode_intel.o microcode-$(CONFIG_MICROCODE_AMD) += microcode_amd.o obj-$(CONFIG_MICROCODE) += microcode.o +obj-$(CONFIG_X86_CHECK_BIOS_CORRUPTION) += check.o + ### # 64 bit specific files ifeq ($(CONFIG_X86_64),y) diff --git a/arch/x86/kernel/check.c b/arch/x86/kernel/check.c index 55eed1752b43..2ac0ab71412a 100644 --- a/arch/x86/kernel/check.c +++ b/arch/x86/kernel/check.c @@ -11,7 +11,6 @@ * remaining free memory in that area and fill it with a distinct * pattern. */ -#ifdef CONFIG_X86_CHECK_BIOS_CORRUPTION #define MAX_SCAN_AREAS 8 static int __read_mostly memory_corruption_check = -1; @@ -23,7 +22,7 @@ static struct e820entry scan_areas[MAX_SCAN_AREAS]; static int num_scan_areas; -static int set_corruption_check(char *arg) +static __init int set_corruption_check(char *arg) { char *end; @@ -33,7 +32,7 @@ static int set_corruption_check(char *arg) } early_param("memory_corruption_check", set_corruption_check); -static int set_corruption_check_period(char *arg) +static __init int set_corruption_check_period(char *arg) { char *end; @@ -43,7 +42,7 @@ static int set_corruption_check_period(char *arg) } early_param("memory_corruption_check_period", set_corruption_check_period); -static int set_corruption_check_size(char *arg) +static __init int set_corruption_check_size(char *arg) { char *end; unsigned size; @@ -115,8 +114,6 @@ void check_for_bios_corruption(void) int i; int corruption = 0; - printk("dot\n"); - if (!memory_corruption_check) return; @@ -134,7 +131,7 @@ void check_for_bios_corruption(void) } } - WARN(corruption, KERN_ERR "Memory corruption detected in low memory\n"); + WARN_ONCE(corruption, KERN_ERR "Memory corruption detected in low memory\n"); } static void check_corruption(struct work_struct *dummy); @@ -161,5 +158,4 @@ static int start_periodic_check_for_corruption(void) } module_init(start_periodic_check_for_corruption); -#endif -- cgit v1.2.3 From 6f290b4e016d6c61511542cf6d9ebdef1965978e Mon Sep 17 00:00:00 2001 From: Aristeu Rozanski Date: Mon, 27 Oct 2008 12:42:34 -0400 Subject: x86, NMI watchdog: add support to enable and disable IOAPIC NMI Impact: change/improve the way /proc/sys/kernel/nmi_watchdog works This patch adds support to enable/disable IOAPIC NMI watchdog in runtime via procfs. Signed-off-by: Aristeu Rozanski Cc: "Maciej W. Rozycki" Signed-off-by: Ingo Molnar --- arch/x86/kernel/nmi.c | 25 +++++++++++++++++++++++++ 1 file changed, 25 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/nmi.c b/arch/x86/kernel/nmi.c index 2c97f07f1c2c..2c005fac6171 100644 --- a/arch/x86/kernel/nmi.c +++ b/arch/x86/kernel/nmi.c @@ -340,6 +340,8 @@ void stop_apic_nmi_watchdog(void *unused) return; if (nmi_watchdog == NMI_LOCAL_APIC) lapic_watchdog_stop(); + else + __acpi_nmi_disable(NULL); __get_cpu_var(wd_enabled) = 0; atomic_dec(&nmi_active); } @@ -465,6 +467,24 @@ nmi_watchdog_tick(struct pt_regs *regs, unsigned reason) #ifdef CONFIG_SYSCTL +static void enable_ioapic_nmi_watchdog_single(void *unused) +{ + __get_cpu_var(wd_enabled) = 1; + atomic_inc(&nmi_active); + __acpi_nmi_enable(NULL); +} + +static void enable_ioapic_nmi_watchdog(void) +{ + on_each_cpu(enable_ioapic_nmi_watchdog_single, NULL, 1); + touch_nmi_watchdog(); +} + +static void disable_ioapic_nmi_watchdog(void) +{ + on_each_cpu(stop_apic_nmi_watchdog, NULL, 1); +} + static int __init setup_unknown_nmi_panic(char *str) { unknown_nmi_panic = 1; @@ -507,6 +527,11 @@ int proc_nmi_enabled(struct ctl_table *table, int write, struct file *file, enable_lapic_nmi_watchdog(); else disable_lapic_nmi_watchdog(); + } else if (nmi_watchdog == NMI_IO_APIC) { + if (nmi_watchdog_enabled) + enable_ioapic_nmi_watchdog(); + else + disable_ioapic_nmi_watchdog(); } else { printk(KERN_WARNING "NMI watchdog doesn't know what hardware to touch\n"); -- cgit v1.2.3 From 7d5a78cd98c3a5eb83bd6a061c5ea6ef1e9b8fcb Mon Sep 17 00:00:00 2001 From: Aristeu Rozanski Date: Mon, 27 Oct 2008 12:42:35 -0400 Subject: x86, NMI watchdog: disable NMIs on LVT0 in case NMI watchdog is not working Impact: change NMI watchdog detection and disabling sequence Currently, if the NMI watchdog fails using IOAPIC method, it'll only disable interrupts on 8259 if the timer is passing thru it. This patch disables NMI delivery on LINT0 if the NMI watchdog initial test fails, just for safety. Signed-off-by: Aristeu Rozanski Cc: "Maciej W. Rozycki" Signed-off-by: Ingo Molnar --- arch/x86/kernel/nmi.c | 18 +++++++++++------- 1 file changed, 11 insertions(+), 7 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/nmi.c b/arch/x86/kernel/nmi.c index 2c005fac6171..13316cf57cdb 100644 --- a/arch/x86/kernel/nmi.c +++ b/arch/x86/kernel/nmi.c @@ -131,6 +131,11 @@ static void report_broken_nmi(int cpu, int *prev_nmi_count) atomic_dec(&nmi_active); } +static void __acpi_nmi_disable(void *__unused) +{ + apic_write(APIC_LVT0, APIC_DM_NMI | APIC_LVT_MASKED); +} + int __init check_nmi_watchdog(void) { unsigned int *prev_nmi_count; @@ -179,8 +184,12 @@ int __init check_nmi_watchdog(void) kfree(prev_nmi_count); return 0; error: - if (nmi_watchdog == NMI_IO_APIC && !timer_through_8259) - disable_8259A_irq(0); + if (nmi_watchdog == NMI_IO_APIC) { + if (!timer_through_8259) + disable_8259A_irq(0); + on_each_cpu(__acpi_nmi_disable, NULL, 1); + } + #ifdef CONFIG_X86_32 timer_ack = 0; #endif @@ -285,11 +294,6 @@ void acpi_nmi_enable(void) on_each_cpu(__acpi_nmi_enable, NULL, 1); } -static void __acpi_nmi_disable(void *__unused) -{ - apic_write(APIC_LVT0, APIC_DM_NMI | APIC_LVT_MASKED); -} - /* * Disable timer based NMIs on all CPUs: */ -- cgit v1.2.3 From 878719e831d9e076961aa15d4049a57a6668c67a Mon Sep 17 00:00:00 2001 From: Neil Horman Date: Thu, 23 Oct 2008 10:40:06 -0400 Subject: x86: unify appropriate bits from dumpstack_32 and dumpstack_64 Impact: cleanup As promised, now that dumpstack_32 and dumpstack_64 have so many bits in common, we should merge the in-sync bits into a common file, to prevent them from diverging again. This patch removes bits which are common between dumpstack_32.c and dumpstack_64.c and places them in a common dumpstack.c which is built for both 32 and 64 bit arches. Signed-off-by: Neil Horman Acked-by: Alexander van Heukelum Signed-off-by: Ingo Molnar Makefile | 2 arch/x86/kernel/Makefile | 2 arch/x86/kernel/Makefile | 2 arch/x86/kernel/Makefile | 2 arch/x86/kernel/Makefile | 2 arch/x86/kernel/Makefile | 2 arch/x86/kernel/dumpstack.c | 319 +++++++++++++++++++++++++++++++++++++++++ arch/x86/kernel/dumpstack.h | 39 +++++ arch/x86/kernel/dumpstack_32.c | 294 ------------------------------------- arch/x86/kernel/dumpstack_64.c | 285 ------------------------------------ 5 files changed, 363 insertions(+), 576 deletions(-) --- arch/x86/kernel/Makefile | 2 +- arch/x86/kernel/dumpstack.c | 319 +++++++++++++++++++++++++++++++++++++++++ arch/x86/kernel/dumpstack.h | 39 +++++ arch/x86/kernel/dumpstack_32.c | 294 +------------------------------------ arch/x86/kernel/dumpstack_64.c | 285 +----------------------------------- 5 files changed, 363 insertions(+), 576 deletions(-) create mode 100644 arch/x86/kernel/dumpstack.c create mode 100644 arch/x86/kernel/dumpstack.h (limited to 'arch/x86') diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index d7e5a58ee22f..db3216a9d2b9 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -24,7 +24,7 @@ CFLAGS_tsc.o := $(nostackp) obj-y := process_$(BITS).o signal_$(BITS).o entry_$(BITS).o obj-y += traps.o irq.o irq_$(BITS).o dumpstack_$(BITS).o -obj-y += time_$(BITS).o ioport.o ldt.o +obj-y += time_$(BITS).o ioport.o ldt.o dumpstack.o obj-y += setup.o i8259.o irqinit_$(BITS).o setup_percpu.o obj-$(CONFIG_X86_VISWS) += visws_quirks.o obj-$(CONFIG_X86_32) += probe_roms_32.o diff --git a/arch/x86/kernel/dumpstack.c b/arch/x86/kernel/dumpstack.c new file mode 100644 index 000000000000..5962176dfabb --- /dev/null +++ b/arch/x86/kernel/dumpstack.c @@ -0,0 +1,319 @@ +/* + * Copyright (C) 1991, 1992 Linus Torvalds + * Copyright (C) 2000, 2001, 2002 Andi Kleen, SuSE Labs + */ +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include + +#include "dumpstack.h" + +int panic_on_unrecovered_nmi; +unsigned int code_bytes = 64; +int kstack_depth_to_print = 3 * STACKSLOTS_PER_LINE; +static int die_counter; + +void printk_address(unsigned long address, int reliable) +{ + printk(" [<%p>] %s%pS\n", (void *) address, + reliable ? "" : "? ", (void *) address); +} + +/* + * x86-64 can have up to three kernel stacks: + * process stack + * interrupt stack + * severe exception (double fault, nmi, stack fault, debug, mce) hardware stack + */ + +static inline int valid_stack_ptr(struct thread_info *tinfo, + void *p, unsigned int size, void *end) +{ + void *t = tinfo; + if (end) { + if (p < end && p >= (end-THREAD_SIZE)) + return 1; + else + return 0; + } + return p > t && p < t + THREAD_SIZE - size; +} + +unsigned long +print_context_stack(struct thread_info *tinfo, + unsigned long *stack, unsigned long bp, + const struct stacktrace_ops *ops, void *data, + unsigned long *end) +{ + struct stack_frame *frame = (struct stack_frame *)bp; + + while (valid_stack_ptr(tinfo, stack, sizeof(*stack), end)) { + unsigned long addr; + + addr = *stack; + if (__kernel_text_address(addr)) { + if ((unsigned long) stack == bp + sizeof(long)) { + ops->address(data, addr, 1); + frame = frame->next_frame; + bp = (unsigned long) frame; + } else { + ops->address(data, addr, bp == 0); + } + } + stack++; + } + return bp; +} + + +static void +print_trace_warning_symbol(void *data, char *msg, unsigned long symbol) +{ + printk(data); + print_symbol(msg, symbol); + printk("\n"); +} + +static void print_trace_warning(void *data, char *msg) +{ + printk("%s%s\n", (char *)data, msg); +} + +static int print_trace_stack(void *data, char *name) +{ + printk("%s <%s> ", (char *)data, name); + return 0; +} + +/* + * Print one address/symbol entries per line. + */ +static void print_trace_address(void *data, unsigned long addr, int reliable) +{ + touch_nmi_watchdog(); + printk(data); + printk_address(addr, reliable); +} + +static const struct stacktrace_ops print_trace_ops = { + .warning = print_trace_warning, + .warning_symbol = print_trace_warning_symbol, + .stack = print_trace_stack, + .address = print_trace_address, +}; + +void +show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs, + unsigned long *stack, unsigned long bp, char *log_lvl) +{ + printk("%sCall Trace:\n", log_lvl); + dump_trace(task, regs, stack, bp, &print_trace_ops, log_lvl); +} + +void show_trace(struct task_struct *task, struct pt_regs *regs, + unsigned long *stack, unsigned long bp) +{ + show_trace_log_lvl(task, regs, stack, bp, ""); +} + +void show_stack(struct task_struct *task, unsigned long *sp) +{ + show_stack_log_lvl(task, NULL, sp, 0, ""); +} + +/* + * The architecture-independent dump_stack generator + */ +void dump_stack(void) +{ + unsigned long bp = 0; + unsigned long stack; + +#ifdef CONFIG_FRAME_POINTER + if (!bp) + get_bp(bp); +#endif + + printk("Pid: %d, comm: %.20s %s %s %.*s\n", + current->pid, current->comm, print_tainted(), + init_utsname()->release, + (int)strcspn(init_utsname()->version, " "), + init_utsname()->version); + show_trace(NULL, NULL, &stack, bp); +} +EXPORT_SYMBOL(dump_stack); + +static raw_spinlock_t die_lock = __RAW_SPIN_LOCK_UNLOCKED; +static int die_owner = -1; +static unsigned int die_nest_count; + +unsigned __kprobes long oops_begin(void) +{ + int cpu; + unsigned long flags; + + oops_enter(); + + /* racy, but better than risking deadlock. */ + raw_local_irq_save(flags); + cpu = smp_processor_id(); + if (!__raw_spin_trylock(&die_lock)) { + if (cpu == die_owner) + /* nested oops. should stop eventually */; + else + __raw_spin_lock(&die_lock); + } + die_nest_count++; + die_owner = cpu; + console_verbose(); + bust_spinlocks(1); + return flags; +} + +void __kprobes oops_end(unsigned long flags, struct pt_regs *regs, int signr) +{ + if (regs && kexec_should_crash(current)) + crash_kexec(regs); + + bust_spinlocks(0); + die_owner = -1; + add_taint(TAINT_DIE); + die_nest_count--; + if (!die_nest_count) + /* Nest count reaches zero, release the lock. */ + __raw_spin_unlock(&die_lock); + raw_local_irq_restore(flags); + oops_exit(); + + if (!signr) + return; + if (in_interrupt()) + panic("Fatal exception in interrupt"); + if (panic_on_oops) + panic("Fatal exception"); + do_exit(signr); +} + +int __kprobes __die(const char *str, struct pt_regs *regs, long err) +{ +#ifdef CONFIG_X86_32 + unsigned short ss; + unsigned long sp; +#endif + printk(KERN_EMERG "%s: %04lx [#%d] ", str, err & 0xffff, ++die_counter); +#ifdef CONFIG_PREEMPT + printk("PREEMPT "); +#endif +#ifdef CONFIG_SMP + printk("SMP "); +#endif +#ifdef CONFIG_DEBUG_PAGEALLOC + printk("DEBUG_PAGEALLOC"); +#endif + printk("\n"); + sysfs_printk_last_file(); + if (notify_die(DIE_OOPS, str, regs, err, + current->thread.trap_no, SIGSEGV) == NOTIFY_STOP) + return 1; + + show_registers(regs); +#ifdef CONFIG_X86_32 + sp = (unsigned long) (®s->sp); + savesegment(ss, ss); + if (user_mode(regs)) { + sp = regs->sp; + ss = regs->ss & 0xffff; + } + printk(KERN_EMERG "EIP: [<%08lx>] ", regs->ip); + print_symbol("%s", regs->ip); + printk(" SS:ESP %04x:%08lx\n", ss, sp); +#else + /* Executive summary in case the oops scrolled away */ + printk(KERN_ALERT "RIP "); + printk_address(regs->ip, 1); + printk(" RSP <%016lx>\n", regs->sp); +#endif + return 0; +} + +/* + * This is gone through when something in the kernel has done something bad + * and is about to be terminated: + */ +void die(const char *str, struct pt_regs *regs, long err) +{ + unsigned long flags = oops_begin(); + int sig = SIGSEGV; + + if (!user_mode_vm(regs)) + report_bug(regs->ip, regs); + + if (__die(str, regs, err)) + sig = 0; + oops_end(flags, regs, sig); +} + +void notrace __kprobes +die_nmi(char *str, struct pt_regs *regs, int do_panic) +{ + unsigned long flags; + + if (notify_die(DIE_NMIWATCHDOG, str, regs, 0, 2, SIGINT) == NOTIFY_STOP) + return; + + /* + * We are in trouble anyway, lets at least try + * to get a message out. + */ + flags = oops_begin(); + printk(KERN_EMERG "%s", str); + printk(" on CPU%d, ip %08lx, registers:\n", + smp_processor_id(), regs->ip); + show_registers(regs); + oops_end(flags, regs, 0); + if (do_panic || panic_on_oops) + panic("Non maskable interrupt"); + nmi_exit(); + local_irq_enable(); + do_exit(SIGBUS); +} + +static int __init oops_setup(char *s) +{ + if (!s) + return -EINVAL; + if (!strcmp(s, "panic")) + panic_on_oops = 1; + return 0; +} +early_param("oops", oops_setup); + +static int __init kstack_setup(char *s) +{ + if (!s) + return -EINVAL; + kstack_depth_to_print = simple_strtoul(s, NULL, 0); + return 0; +} +early_param("kstack", kstack_setup); + +static int __init code_bytes_setup(char *s) +{ + code_bytes = simple_strtoul(s, NULL, 0); + if (code_bytes > 8192) + code_bytes = 8192; + + return 1; +} +__setup("code_bytes=", code_bytes_setup); diff --git a/arch/x86/kernel/dumpstack.h b/arch/x86/kernel/dumpstack.h new file mode 100644 index 000000000000..3119a801c32b --- /dev/null +++ b/arch/x86/kernel/dumpstack.h @@ -0,0 +1,39 @@ +/* + * Copyright (C) 1991, 1992 Linus Torvalds + * Copyright (C) 2000, 2001, 2002 Andi Kleen, SuSE Labs + */ + +#ifndef DUMPSTACK_H +#define DUMPSTACK_H + +#ifdef CONFIG_X86_32 +#define STACKSLOTS_PER_LINE 8 +#define get_bp(bp) asm("movl %%ebp, %0" : "=r" (bp) :) +#else +#define STACKSLOTS_PER_LINE 4 +#define get_bp(bp) asm("movq %%rbp, %0" : "=r" (bp) :) +#endif + +extern unsigned long +print_context_stack(struct thread_info *tinfo, + unsigned long *stack, unsigned long bp, + const struct stacktrace_ops *ops, void *data, + unsigned long *end); + +extern void +show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs, + unsigned long *stack, unsigned long bp, char *log_lvl); + +extern void +show_stack_log_lvl(struct task_struct *task, struct pt_regs *regs, + unsigned long *sp, unsigned long bp, char *log_lvl); + +extern unsigned int code_bytes; +extern int kstack_depth_to_print; + +/* The form of the top of the frame on the stack */ +struct stack_frame { + struct stack_frame *next_frame; + unsigned long return_address; +}; +#endif diff --git a/arch/x86/kernel/dumpstack_32.c b/arch/x86/kernel/dumpstack_32.c index f2046c5752d0..7b031b106ec8 100644 --- a/arch/x86/kernel/dumpstack_32.c +++ b/arch/x86/kernel/dumpstack_32.c @@ -17,64 +17,7 @@ #include -#define STACKSLOTS_PER_LINE 8 -#define get_bp(bp) asm("movl %%ebp, %0" : "=r" (bp) :) - -int panic_on_unrecovered_nmi; -int kstack_depth_to_print = 3 * STACKSLOTS_PER_LINE; -static unsigned int code_bytes = 64; -static int die_counter; - -void printk_address(unsigned long address, int reliable) -{ - printk(" [<%p>] %s%pS\n", (void *) address, - reliable ? "" : "? ", (void *) address); -} - -static inline int valid_stack_ptr(struct thread_info *tinfo, - void *p, unsigned int size, void *end) -{ - void *t = tinfo; - if (end) { - if (p < end && p >= (end-THREAD_SIZE)) - return 1; - else - return 0; - } - return p > t && p < t + THREAD_SIZE - size; -} - -/* The form of the top of the frame on the stack */ -struct stack_frame { - struct stack_frame *next_frame; - unsigned long return_address; -}; - -static inline unsigned long -print_context_stack(struct thread_info *tinfo, - unsigned long *stack, unsigned long bp, - const struct stacktrace_ops *ops, void *data, - unsigned long *end) -{ - struct stack_frame *frame = (struct stack_frame *)bp; - - while (valid_stack_ptr(tinfo, stack, sizeof(*stack), end)) { - unsigned long addr; - - addr = *stack; - if (__kernel_text_address(addr)) { - if ((unsigned long) stack == bp + sizeof(long)) { - ops->address(data, addr, 1); - frame = frame->next_frame; - bp = (unsigned long) frame; - } else { - ops->address(data, addr, bp == 0); - } - } - stack++; - } - return bp; -} +#include "dumpstack.h" void dump_trace(struct task_struct *task, struct pt_regs *regs, unsigned long *stack, unsigned long bp, @@ -119,57 +62,7 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs, } EXPORT_SYMBOL(dump_trace); -static void -print_trace_warning_symbol(void *data, char *msg, unsigned long symbol) -{ - printk(data); - print_symbol(msg, symbol); - printk("\n"); -} - -static void print_trace_warning(void *data, char *msg) -{ - printk("%s%s\n", (char *)data, msg); -} - -static int print_trace_stack(void *data, char *name) -{ - printk("%s <%s> ", (char *)data, name); - return 0; -} - -/* - * Print one address/symbol entries per line. - */ -static void print_trace_address(void *data, unsigned long addr, int reliable) -{ - touch_nmi_watchdog(); - printk(data); - printk_address(addr, reliable); -} - -static const struct stacktrace_ops print_trace_ops = { - .warning = print_trace_warning, - .warning_symbol = print_trace_warning_symbol, - .stack = print_trace_stack, - .address = print_trace_address, -}; - -static void -show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs, - unsigned long *stack, unsigned long bp, char *log_lvl) -{ - printk("%sCall Trace:\n", log_lvl); - dump_trace(task, regs, stack, bp, &print_trace_ops, log_lvl); -} - -void show_trace(struct task_struct *task, struct pt_regs *regs, - unsigned long *stack, unsigned long bp) -{ - show_trace_log_lvl(task, regs, stack, bp, ""); -} - -static void +void show_stack_log_lvl(struct task_struct *task, struct pt_regs *regs, unsigned long *sp, unsigned long bp, char *log_lvl) { @@ -196,33 +89,6 @@ show_stack_log_lvl(struct task_struct *task, struct pt_regs *regs, show_trace_log_lvl(task, regs, sp, bp, log_lvl); } -void show_stack(struct task_struct *task, unsigned long *sp) -{ - show_stack_log_lvl(task, NULL, sp, 0, ""); -} - -/* - * The architecture-independent dump_stack generator - */ -void dump_stack(void) -{ - unsigned long bp = 0; - unsigned long stack; - -#ifdef CONFIG_FRAME_POINTER - if (!bp) - get_bp(bp); -#endif - - printk("Pid: %d, comm: %.20s %s %s %.*s\n", - current->pid, current->comm, print_tainted(), - init_utsname()->release, - (int)strcspn(init_utsname()->version, " "), - init_utsname()->version); - show_trace(NULL, NULL, &stack, bp); -} - -EXPORT_SYMBOL(dump_stack); void show_registers(struct pt_regs *regs) { @@ -283,159 +149,3 @@ int is_valid_bugaddr(unsigned long ip) return ud2 == 0x0b0f; } -static raw_spinlock_t die_lock = __RAW_SPIN_LOCK_UNLOCKED; -static int die_owner = -1; -static unsigned int die_nest_count; - -unsigned __kprobes long oops_begin(void) -{ - int cpu; - unsigned long flags; - - oops_enter(); - - /* racy, but better than risking deadlock. */ - raw_local_irq_save(flags); - cpu = smp_processor_id(); - if (!__raw_spin_trylock(&die_lock)) { - if (cpu == die_owner) - /* nested oops. should stop eventually */; - else - __raw_spin_lock(&die_lock); - } - die_nest_count++; - die_owner = cpu; - console_verbose(); - bust_spinlocks(1); - return flags; -} - -void __kprobes oops_end(unsigned long flags, struct pt_regs *regs, int signr) -{ - if (regs && kexec_should_crash(current)) - crash_kexec(regs); - - bust_spinlocks(0); - die_owner = -1; - add_taint(TAINT_DIE); - die_nest_count--; - if (!die_nest_count) - /* Nest count reaches zero, release the lock. */ - __raw_spin_unlock(&die_lock); - raw_local_irq_restore(flags); - oops_exit(); - - if (!signr) - return; - if (in_interrupt()) - panic("Fatal exception in interrupt"); - if (panic_on_oops) - panic("Fatal exception"); - do_exit(signr); -} - -int __kprobes __die(const char *str, struct pt_regs *regs, long err) -{ - unsigned short ss; - unsigned long sp; - - printk(KERN_EMERG "%s: %04lx [#%d] ", str, err & 0xffff, ++die_counter); -#ifdef CONFIG_PREEMPT - printk("PREEMPT "); -#endif -#ifdef CONFIG_SMP - printk("SMP "); -#endif -#ifdef CONFIG_DEBUG_PAGEALLOC - printk("DEBUG_PAGEALLOC"); -#endif - printk("\n"); - sysfs_printk_last_file(); - if (notify_die(DIE_OOPS, str, regs, err, - current->thread.trap_no, SIGSEGV) == NOTIFY_STOP) - return 1; - - show_registers(regs); - /* Executive summary in case the oops scrolled away */ - sp = (unsigned long) (®s->sp); - savesegment(ss, ss); - if (user_mode(regs)) { - sp = regs->sp; - ss = regs->ss & 0xffff; - } - printk(KERN_EMERG "EIP: [<%08lx>] ", regs->ip); - print_symbol("%s", regs->ip); - printk(" SS:ESP %04x:%08lx\n", ss, sp); - return 0; -} - -/* - * This is gone through when something in the kernel has done something bad - * and is about to be terminated: - */ -void die(const char *str, struct pt_regs *regs, long err) -{ - unsigned long flags = oops_begin(); - int sig = SIGSEGV; - - if (!user_mode_vm(regs)) - report_bug(regs->ip, regs); - - if (__die(str, regs, err)) - sig = 0; - oops_end(flags, regs, sig); -} - -void notrace __kprobes -die_nmi(char *str, struct pt_regs *regs, int do_panic) -{ - unsigned long flags; - - if (notify_die(DIE_NMIWATCHDOG, str, regs, 0, 2, SIGINT) == NOTIFY_STOP) - return; - - /* - * We are in trouble anyway, lets at least try - * to get a message out. - */ - flags = oops_begin(); - printk(KERN_EMERG "%s", str); - printk(" on CPU%d, ip %08lx, registers:\n", - smp_processor_id(), regs->ip); - show_registers(regs); - oops_end(flags, regs, 0); - if (do_panic || panic_on_oops) - panic("Non maskable interrupt"); - nmi_exit(); - local_irq_enable(); - do_exit(SIGBUS); -} - -static int __init oops_setup(char *s) -{ - if (!s) - return -EINVAL; - if (!strcmp(s, "panic")) - panic_on_oops = 1; - return 0; -} -early_param("oops", oops_setup); - -static int __init kstack_setup(char *s) -{ - if (!s) - return -EINVAL; - kstack_depth_to_print = simple_strtoul(s, NULL, 0); - return 0; -} -early_param("kstack", kstack_setup); - -static int __init code_bytes_setup(char *s) -{ - code_bytes = simple_strtoul(s, NULL, 0); - if (code_bytes > 8192) - code_bytes = 8192; - - return 1; -} -__setup("code_bytes=", code_bytes_setup); diff --git a/arch/x86/kernel/dumpstack_64.c b/arch/x86/kernel/dumpstack_64.c index 28c67aae5562..33ff10287a5d 100644 --- a/arch/x86/kernel/dumpstack_64.c +++ b/arch/x86/kernel/dumpstack_64.c @@ -17,19 +17,7 @@ #include -#define STACKSLOTS_PER_LINE 4 -#define get_bp(bp) asm("movq %%rbp, %0" : "=r" (bp) :) - -int panic_on_unrecovered_nmi; -int kstack_depth_to_print = 3 * STACKSLOTS_PER_LINE; -static unsigned int code_bytes = 64; -static int die_counter; - -void printk_address(unsigned long address, int reliable) -{ - printk(" [<%p>] %s%pS\n", (void *) address, - reliable ? "" : "? ", (void *) address); -} +#include "dumpstack.h" static unsigned long *in_exception_stack(unsigned cpu, unsigned long stack, unsigned *usedp, char **idp) @@ -113,51 +101,6 @@ static unsigned long *in_exception_stack(unsigned cpu, unsigned long stack, * severe exception (double fault, nmi, stack fault, debug, mce) hardware stack */ -static inline int valid_stack_ptr(struct thread_info *tinfo, - void *p, unsigned int size, void *end) -{ - void *t = tinfo; - if (end) { - if (p < end && p >= (end-THREAD_SIZE)) - return 1; - else - return 0; - } - return p > t && p < t + THREAD_SIZE - size; -} - -/* The form of the top of the frame on the stack */ -struct stack_frame { - struct stack_frame *next_frame; - unsigned long return_address; -}; - -static inline unsigned long -print_context_stack(struct thread_info *tinfo, - unsigned long *stack, unsigned long bp, - const struct stacktrace_ops *ops, void *data, - unsigned long *end) -{ - struct stack_frame *frame = (struct stack_frame *)bp; - - while (valid_stack_ptr(tinfo, stack, sizeof(*stack), end)) { - unsigned long addr; - - addr = *stack; - if (__kernel_text_address(addr)) { - if ((unsigned long) stack == bp + sizeof(long)) { - ops->address(data, addr, 1); - frame = frame->next_frame; - bp = (unsigned long) frame; - } else { - ops->address(data, addr, bp == 0); - } - } - stack++; - } - return bp; -} - void dump_trace(struct task_struct *task, struct pt_regs *regs, unsigned long *stack, unsigned long bp, const struct stacktrace_ops *ops, void *data) @@ -248,57 +191,7 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs, } EXPORT_SYMBOL(dump_trace); -static void -print_trace_warning_symbol(void *data, char *msg, unsigned long symbol) -{ - printk(data); - print_symbol(msg, symbol); - printk("\n"); -} - -static void print_trace_warning(void *data, char *msg) -{ - printk("%s%s\n", (char *)data, msg); -} - -static int print_trace_stack(void *data, char *name) -{ - printk("%s <%s> ", (char *)data, name); - return 0; -} - -/* - * Print one address/symbol entries per line. - */ -static void print_trace_address(void *data, unsigned long addr, int reliable) -{ - touch_nmi_watchdog(); - printk(data); - printk_address(addr, reliable); -} - -static const struct stacktrace_ops print_trace_ops = { - .warning = print_trace_warning, - .warning_symbol = print_trace_warning_symbol, - .stack = print_trace_stack, - .address = print_trace_address, -}; - -static void -show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs, - unsigned long *stack, unsigned long bp, char *log_lvl) -{ - printk("%sCall Trace:\n", log_lvl); - dump_trace(task, regs, stack, bp, &print_trace_ops, log_lvl); -} - -void show_trace(struct task_struct *task, struct pt_regs *regs, - unsigned long *stack, unsigned long bp) -{ - show_trace_log_lvl(task, regs, stack, bp, ""); -} - -static void +void show_stack_log_lvl(struct task_struct *task, struct pt_regs *regs, unsigned long *sp, unsigned long bp, char *log_lvl) { @@ -342,33 +235,6 @@ show_stack_log_lvl(struct task_struct *task, struct pt_regs *regs, show_trace_log_lvl(task, regs, sp, bp, log_lvl); } -void show_stack(struct task_struct *task, unsigned long *sp) -{ - show_stack_log_lvl(task, NULL, sp, 0, ""); -} - -/* - * The architecture-independent dump_stack generator - */ -void dump_stack(void) -{ - unsigned long bp = 0; - unsigned long stack; - -#ifdef CONFIG_FRAME_POINTER - if (!bp) - get_bp(bp); -#endif - - printk("Pid: %d, comm: %.20s %s %s %.*s\n", - current->pid, current->comm, print_tainted(), - init_utsname()->release, - (int)strcspn(init_utsname()->version, " "), - init_utsname()->version); - show_trace(NULL, NULL, &stack, bp); -} -EXPORT_SYMBOL(dump_stack); - void show_registers(struct pt_regs *regs) { int i; @@ -429,150 +295,3 @@ int is_valid_bugaddr(unsigned long ip) return ud2 == 0x0b0f; } -static raw_spinlock_t die_lock = __RAW_SPIN_LOCK_UNLOCKED; -static int die_owner = -1; -static unsigned int die_nest_count; - -unsigned __kprobes long oops_begin(void) -{ - int cpu; - unsigned long flags; - - oops_enter(); - - /* racy, but better than risking deadlock. */ - raw_local_irq_save(flags); - cpu = smp_processor_id(); - if (!__raw_spin_trylock(&die_lock)) { - if (cpu == die_owner) - /* nested oops. should stop eventually */; - else - __raw_spin_lock(&die_lock); - } - die_nest_count++; - die_owner = cpu; - console_verbose(); - bust_spinlocks(1); - return flags; -} - -void __kprobes oops_end(unsigned long flags, struct pt_regs *regs, int signr) -{ - if (regs && kexec_should_crash(current)) - crash_kexec(regs); - - bust_spinlocks(0); - die_owner = -1; - add_taint(TAINT_DIE); - die_nest_count--; - if (!die_nest_count) - /* Nest count reaches zero, release the lock. */ - __raw_spin_unlock(&die_lock); - raw_local_irq_restore(flags); - oops_exit(); - - if (!signr) - return; - if (in_interrupt()) - panic("Fatal exception in interrupt"); - if (panic_on_oops) - panic("Fatal exception"); - do_exit(signr); -} - -int __kprobes __die(const char *str, struct pt_regs *regs, long err) -{ - printk(KERN_EMERG "%s: %04lx [#%d] ", str, err & 0xffff, ++die_counter); -#ifdef CONFIG_PREEMPT - printk("PREEMPT "); -#endif -#ifdef CONFIG_SMP - printk("SMP "); -#endif -#ifdef CONFIG_DEBUG_PAGEALLOC - printk("DEBUG_PAGEALLOC"); -#endif - printk("\n"); - sysfs_printk_last_file(); - if (notify_die(DIE_OOPS, str, regs, err, - current->thread.trap_no, SIGSEGV) == NOTIFY_STOP) - return 1; - - show_registers(regs); - /* Executive summary in case the oops scrolled away */ - printk(KERN_ALERT "RIP "); - printk_address(regs->ip, 1); - printk(" RSP <%016lx>\n", regs->sp); - return 0; -} - -/* - * This is gone through when something in the kernel has done something bad - * and is about to be terminated: - */ -void die(const char *str, struct pt_regs *regs, long err) -{ - unsigned long flags = oops_begin(); - int sig = SIGSEGV; - - if (!user_mode_vm(regs)) - report_bug(regs->ip, regs); - - if (__die(str, regs, err)) - sig = 0; - oops_end(flags, regs, sig); -} - -void notrace __kprobes -die_nmi(char *str, struct pt_regs *regs, int do_panic) -{ - unsigned long flags; - - if (notify_die(DIE_NMIWATCHDOG, str, regs, 0, 2, SIGINT) == NOTIFY_STOP) - return; - - /* - * We are in trouble anyway, lets at least try - * to get a message out. - */ - flags = oops_begin(); - printk(KERN_EMERG "%s", str); - printk(" on CPU%d, ip %08lx, registers:\n", - smp_processor_id(), regs->ip); - show_registers(regs); - oops_end(flags, regs, 0); - if (do_panic || panic_on_oops) - panic("Non maskable interrupt"); - nmi_exit(); - local_irq_enable(); - do_exit(SIGBUS); -} - -static int __init oops_setup(char *s) -{ - if (!s) - return -EINVAL; - if (!strcmp(s, "panic")) - panic_on_oops = 1; - return 0; -} -early_param("oops", oops_setup); - -static int __init kstack_setup(char *s) -{ - if (!s) - return -EINVAL; - kstack_depth_to_print = simple_strtoul(s, NULL, 0); - return 0; -} -early_param("kstack", kstack_setup); - -static int __init code_bytes_setup(char *s) -{ - code_bytes = simple_strtoul(s, NULL, 0); - if (code_bytes > 8192) - code_bytes = 8192; - - return 1; -} -__setup("code_bytes=", code_bytes_setup); -- cgit v1.2.3 From 69a72a0e9337aad8c730e8e9942d5aa022bc4c5c Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Mon, 27 Oct 2008 07:51:20 -0700 Subject: x86/uv: update SCIR driver to use the idle_cpu() function Impact: cleanup Change UV heartbeat function to use idle_cpu to determine cpu's "idleness". Realign uv_hub definitions. Signed-of-by: Mike Travis Signed-off-by: Ingo Molnar --- arch/x86/include/asm/uv/uv_hub.h | 26 +++++++++++++------------- arch/x86/kernel/genx2apic_uv_x.c | 4 ++-- 2 files changed, 15 insertions(+), 15 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/uv/uv_hub.h b/arch/x86/include/asm/uv/uv_hub.h index 400776dba9b5..0ee12928e9ee 100644 --- a/arch/x86/include/asm/uv/uv_hub.h +++ b/arch/x86/include/asm/uv/uv_hub.h @@ -128,19 +128,19 @@ struct uv_scir_s { * They are kept together in a struct to minimize cache misses. */ struct uv_hub_info_s { - unsigned long global_mmr_base; - unsigned long gpa_mask; - unsigned long gnode_upper; - unsigned long lowmem_remap_top; - unsigned long lowmem_remap_base; - unsigned short pnode; - unsigned short pnode_mask; - unsigned short coherency_domain_number; - unsigned short numa_blade_id; - unsigned char blade_processor_id; - unsigned char m_val; - unsigned char n_val; - struct uv_scir_s scir; + unsigned long global_mmr_base; + unsigned long gpa_mask; + unsigned long gnode_upper; + unsigned long lowmem_remap_top; + unsigned long lowmem_remap_base; + unsigned short pnode; + unsigned short pnode_mask; + unsigned short coherency_domain_number; + unsigned short numa_blade_id; + unsigned char blade_processor_id; + unsigned char m_val; + unsigned char n_val; + struct uv_scir_s scir; }; DECLARE_PER_CPU(struct uv_hub_info_s, __uv_hub_info); diff --git a/arch/x86/kernel/genx2apic_uv_x.c b/arch/x86/kernel/genx2apic_uv_x.c index 84367d84bb10..85fb7dd48f67 100644 --- a/arch/x86/kernel/genx2apic_uv_x.c +++ b/arch/x86/kernel/genx2apic_uv_x.c @@ -370,8 +370,8 @@ static void uv_heartbeat(unsigned long ignored) /* flip heartbeat bit */ bits ^= SCIR_CPU_HEARTBEAT; - /* are we the idle thread? */ - if (current->pid == 0) + /* is this cpu idle? */ + if (idle_cpu(raw_smp_processor_id())) bits &= ~SCIR_CPU_ACTIVITY; else bits |= SCIR_CPU_ACTIVITY; -- cgit v1.2.3 From 30604bb410b53efa9c93ee8f03d7aa7494094faa Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Tue, 14 Oct 2008 18:59:18 -0700 Subject: x86: break up mtrr_cleanup() into several small functions. Ingo said mtrr_cleanup() is big and ugly. so break it up into more functions and make it more readable. Signed-off-by: Yinghai Lu Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/mtrr/main.c | 346 ++++++++++++++++++++-------------------- 1 file changed, 171 insertions(+), 175 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/mtrr/main.c b/arch/x86/kernel/cpu/mtrr/main.c index c78c04821ea1..1159e269e596 100644 --- a/arch/x86/kernel/cpu/mtrr/main.c +++ b/arch/x86/kernel/cpu/mtrr/main.c @@ -803,6 +803,7 @@ x86_get_mtrr_mem_range(struct res_range *range, int nr_range, } static struct res_range __initdata range[RANGE_NUM]; +static int __initdata nr_range; #ifdef CONFIG_MTRR_SANITIZER @@ -1206,39 +1207,43 @@ struct mtrr_cleanup_result { #define PSHIFT (PAGE_SHIFT - 10) static struct mtrr_cleanup_result __initdata result[NUM_RESULT]; -static struct res_range __initdata range_new[RANGE_NUM]; static unsigned long __initdata min_loss_pfn[RANGE_NUM]; -static int __init mtrr_cleanup(unsigned address_bits) +static void __init print_out_mtrr_range_state(void) { - unsigned long extra_remove_base, extra_remove_size; - unsigned long base, size, def, dummy; - mtrr_type type; - int nr_range, nr_range_new; - u64 chunk_size, gran_size; - unsigned long range_sums, range_sums_new; - int index_good; - int num_reg_good; int i; + char start_factor = 'K', size_factor = 'K'; + unsigned long start_base, size_base; + mtrr_type type; - /* extra one for all 0 */ - int num[MTRR_NUM_TYPES + 1]; + for (i = 0; i < num_var_ranges; i++) { - if (!is_cpu(INTEL) || enable_mtrr_cleanup < 1) - return 0; - rdmsr(MTRRdefType_MSR, def, dummy); - def &= 0xff; - if (def != MTRR_TYPE_UNCACHABLE) - return 0; + size_base = range_state[i].size_pfn << (PAGE_SHIFT - 10); + if (!size_base) + continue; - /* get it and store it aside */ - memset(range_state, 0, sizeof(range_state)); - for (i = 0; i < num_var_ranges; i++) { - mtrr_if->get(i, &base, &size, &type); - range_state[i].base_pfn = base; - range_state[i].size_pfn = size; - range_state[i].type = type; + size_base = to_size_factor(size_base, &size_factor), + start_base = range_state[i].base_pfn << (PAGE_SHIFT - 10); + start_base = to_size_factor(start_base, &start_factor), + type = range_state[i].type; + + printk(KERN_DEBUG "reg %d, base: %ld%cB, range: %ld%cB, type %s\n", + i, start_base, start_factor, + size_base, size_factor, + (type == MTRR_TYPE_UNCACHABLE) ? "UC" : + ((type == MTRR_TYPE_WRPROT) ? "WP" : + ((type == MTRR_TYPE_WRBACK) ? "WB" : "Other")) + ); } +} + +static int __init mtrr_need_cleanup(void) +{ + int i; + mtrr_type type; + unsigned long size; + /* extra one for all 0 */ + int num[MTRR_NUM_TYPES + 1]; /* check entries number */ memset(num, 0, sizeof(num)); @@ -1263,29 +1268,133 @@ static int __init mtrr_cleanup(unsigned address_bits) num_var_ranges - num[MTRR_NUM_TYPES]) return 0; - /* print original var MTRRs at first, for debugging: */ - printk(KERN_DEBUG "original variable MTRRs\n"); - for (i = 0; i < num_var_ranges; i++) { - char start_factor = 'K', size_factor = 'K'; - unsigned long start_base, size_base; + return 1; +} - size_base = range_state[i].size_pfn << (PAGE_SHIFT - 10); - if (!size_base) - continue; +static unsigned long __initdata range_sums; +static void __init mtrr_calc_range_state(u64 chunk_size, u64 gran_size, + unsigned long extra_remove_base, + unsigned long extra_remove_size, + int i) +{ + int num_reg; + static struct res_range range_new[RANGE_NUM]; + static int nr_range_new; + unsigned long range_sums_new; + + /* convert ranges to var ranges state */ + num_reg = x86_setup_var_mtrrs(range, nr_range, + chunk_size, gran_size); + + /* we got new setting in range_state, check it */ + memset(range_new, 0, sizeof(range_new)); + nr_range_new = x86_get_mtrr_mem_range(range_new, 0, + extra_remove_base, extra_remove_size); + range_sums_new = sum_ranges(range_new, nr_range_new); + + result[i].chunk_sizek = chunk_size >> 10; + result[i].gran_sizek = gran_size >> 10; + result[i].num_reg = num_reg; + if (range_sums < range_sums_new) { + result[i].lose_cover_sizek = + (range_sums_new - range_sums) << PSHIFT; + result[i].bad = 1; + } else + result[i].lose_cover_sizek = + (range_sums - range_sums_new) << PSHIFT; - size_base = to_size_factor(size_base, &size_factor), - start_base = range_state[i].base_pfn << (PAGE_SHIFT - 10); - start_base = to_size_factor(start_base, &start_factor), - type = range_state[i].type; + /* double check it */ + if (!result[i].bad && !result[i].lose_cover_sizek) { + if (nr_range_new != nr_range || + memcmp(range, range_new, sizeof(range))) + result[i].bad = 1; + } - printk(KERN_DEBUG "reg %d, base: %ld%cB, range: %ld%cB, type %s\n", - i, start_base, start_factor, - size_base, size_factor, - (type == MTRR_TYPE_UNCACHABLE) ? "UC" : - ((type == MTRR_TYPE_WRPROT) ? "WP" : - ((type == MTRR_TYPE_WRBACK) ? "WB" : "Other")) - ); + if (!result[i].bad && (range_sums - range_sums_new < + min_loss_pfn[num_reg])) { + min_loss_pfn[num_reg] = + range_sums - range_sums_new; } +} + +static void __init mtrr_print_out_one_result(int i) +{ + char gran_factor, chunk_factor, lose_factor; + unsigned long gran_base, chunk_base, lose_base; + + gran_base = to_size_factor(result[i].gran_sizek, &gran_factor), + chunk_base = to_size_factor(result[i].chunk_sizek, &chunk_factor), + lose_base = to_size_factor(result[i].lose_cover_sizek, &lose_factor), + printk(KERN_INFO "%sgran_size: %ld%c \tchunk_size: %ld%c \t", + result[i].bad ? "*BAD*" : " ", + gran_base, gran_factor, chunk_base, chunk_factor); + printk(KERN_CONT "num_reg: %d \tlose cover RAM: %s%ld%c\n", + result[i].num_reg, result[i].bad ? "-" : "", + lose_base, lose_factor); +} + +static int __init mtrr_search_optimal_index(void) +{ + int i; + int num_reg_good; + int index_good; + + if (nr_mtrr_spare_reg >= num_var_ranges) + nr_mtrr_spare_reg = num_var_ranges - 1; + num_reg_good = -1; + for (i = num_var_ranges - nr_mtrr_spare_reg; i > 0; i--) { + if (!min_loss_pfn[i]) + num_reg_good = i; + } + + index_good = -1; + if (num_reg_good != -1) { + for (i = 0; i < NUM_RESULT; i++) { + if (!result[i].bad && + result[i].num_reg == num_reg_good && + !result[i].lose_cover_sizek) { + index_good = i; + break; + } + } + } + + return index_good; +} + + +static int __init mtrr_cleanup(unsigned address_bits) +{ + unsigned long extra_remove_base, extra_remove_size; + unsigned long base, size, def, dummy; + mtrr_type type; + u64 chunk_size, gran_size; + int index_good; + int i; + + if (!is_cpu(INTEL) || enable_mtrr_cleanup < 1) + return 0; + rdmsr(MTRRdefType_MSR, def, dummy); + def &= 0xff; + if (def != MTRR_TYPE_UNCACHABLE) + return 0; + + /* get it and store it aside */ + memset(range_state, 0, sizeof(range_state)); + for (i = 0; i < num_var_ranges; i++) { + mtrr_if->get(i, &base, &size, &type); + range_state[i].base_pfn = base; + range_state[i].size_pfn = size; + range_state[i].type = type; + } + + /* check if we need handle it and can handle it */ + if (!mtrr_need_cleanup()) + return 0; + + /* print original var MTRRs at first, for debugging: */ + printk(KERN_DEBUG "original variable MTRRs\n"); + print_out_mtrr_range_state(); memset(range, 0, sizeof(range)); extra_remove_size = 0; @@ -1309,176 +1418,64 @@ static int __init mtrr_cleanup(unsigned address_bits) range_sums >> (20 - PAGE_SHIFT)); if (mtrr_chunk_size && mtrr_gran_size) { - int num_reg; - char gran_factor, chunk_factor, lose_factor; - unsigned long gran_base, chunk_base, lose_base; - - debug_print++; - /* convert ranges to var ranges state */ - num_reg = x86_setup_var_mtrrs(range, nr_range, mtrr_chunk_size, - mtrr_gran_size); + i = 0; + mtrr_calc_range_state(mtrr_chunk_size, mtrr_gran_size, + extra_remove_base, extra_remove_size, i); - /* we got new setting in range_state, check it */ - memset(range_new, 0, sizeof(range_new)); - nr_range_new = x86_get_mtrr_mem_range(range_new, 0, - extra_remove_base, - extra_remove_size); - range_sums_new = sum_ranges(range_new, nr_range_new); + mtrr_print_out_one_result(i); - i = 0; - result[i].chunk_sizek = mtrr_chunk_size >> 10; - result[i].gran_sizek = mtrr_gran_size >> 10; - result[i].num_reg = num_reg; - if (range_sums < range_sums_new) { - result[i].lose_cover_sizek = - (range_sums_new - range_sums) << PSHIFT; - result[i].bad = 1; - } else - result[i].lose_cover_sizek = - (range_sums - range_sums_new) << PSHIFT; - - gran_base = to_size_factor(result[i].gran_sizek, &gran_factor), - chunk_base = to_size_factor(result[i].chunk_sizek, &chunk_factor), - lose_base = to_size_factor(result[i].lose_cover_sizek, &lose_factor), - printk(KERN_INFO "%sgran_size: %ld%c \tchunk_size: %ld%c \t", - result[i].bad?"*BAD*":" ", - gran_base, gran_factor, chunk_base, chunk_factor); - printk(KERN_CONT "num_reg: %d \tlose cover RAM: %s%ld%c\n", - result[i].num_reg, result[i].bad?"-":"", - lose_base, lose_factor); if (!result[i].bad) { set_var_mtrr_all(address_bits); return 1; } printk(KERN_INFO "invalid mtrr_gran_size or mtrr_chunk_size, " "will find optimal one\n"); - debug_print--; - memset(result, 0, sizeof(result[0])); } i = 0; memset(min_loss_pfn, 0xff, sizeof(min_loss_pfn)); memset(result, 0, sizeof(result)); for (gran_size = (1ULL<<16); gran_size < (1ULL<<32); gran_size <<= 1) { - char gran_factor; - unsigned long gran_base; - - if (debug_print) - gran_base = to_size_factor(gran_size >> 10, &gran_factor); for (chunk_size = gran_size; chunk_size < (1ULL<<32); chunk_size <<= 1) { - int num_reg; - if (debug_print) { - char chunk_factor; - unsigned long chunk_base; - - chunk_base = to_size_factor(chunk_size>>10, &chunk_factor), - printk(KERN_INFO "\n"); - printk(KERN_INFO "gran_size: %ld%c chunk_size: %ld%c \n", - gran_base, gran_factor, chunk_base, chunk_factor); - } if (i >= NUM_RESULT) continue; - /* convert ranges to var ranges state */ - num_reg = x86_setup_var_mtrrs(range, nr_range, - chunk_size, gran_size); - - /* we got new setting in range_state, check it */ - memset(range_new, 0, sizeof(range_new)); - nr_range_new = x86_get_mtrr_mem_range(range_new, 0, - extra_remove_base, extra_remove_size); - range_sums_new = sum_ranges(range_new, nr_range_new); - - result[i].chunk_sizek = chunk_size >> 10; - result[i].gran_sizek = gran_size >> 10; - result[i].num_reg = num_reg; - if (range_sums < range_sums_new) { - result[i].lose_cover_sizek = - (range_sums_new - range_sums) << PSHIFT; - result[i].bad = 1; - } else - result[i].lose_cover_sizek = - (range_sums - range_sums_new) << PSHIFT; - - /* double check it */ - if (!result[i].bad && !result[i].lose_cover_sizek) { - if (nr_range_new != nr_range || - memcmp(range, range_new, sizeof(range))) - result[i].bad = 1; + mtrr_calc_range_state(chunk_size, gran_size, + extra_remove_base, extra_remove_size, i); + if (debug_print) { + mtrr_print_out_one_result(i); + printk(KERN_INFO "\n"); } - if (!result[i].bad && (range_sums - range_sums_new < - min_loss_pfn[num_reg])) { - min_loss_pfn[num_reg] = - range_sums - range_sums_new; - } i++; } } - /* print out all */ - for (i = 0; i < NUM_RESULT; i++) { - char gran_factor, chunk_factor, lose_factor; - unsigned long gran_base, chunk_base, lose_base; - - gran_base = to_size_factor(result[i].gran_sizek, &gran_factor), - chunk_base = to_size_factor(result[i].chunk_sizek, &chunk_factor), - lose_base = to_size_factor(result[i].lose_cover_sizek, &lose_factor), - printk(KERN_INFO "%sgran_size: %ld%c \tchunk_size: %ld%c \t", - result[i].bad?"*BAD*":" ", - gran_base, gran_factor, chunk_base, chunk_factor); - printk(KERN_CONT "num_reg: %d \tlose cover RAM: %s%ld%c\n", - result[i].num_reg, result[i].bad?"-":"", - lose_base, lose_factor); - } - /* try to find the optimal index */ - if (nr_mtrr_spare_reg >= num_var_ranges) - nr_mtrr_spare_reg = num_var_ranges - 1; - num_reg_good = -1; - for (i = num_var_ranges - nr_mtrr_spare_reg; i > 0; i--) { - if (!min_loss_pfn[i]) - num_reg_good = i; - } - - index_good = -1; - if (num_reg_good != -1) { - for (i = 0; i < NUM_RESULT; i++) { - if (!result[i].bad && - result[i].num_reg == num_reg_good && - !result[i].lose_cover_sizek) { - index_good = i; - break; - } - } - } + index_good = mtrr_search_optimal_index(); if (index_good != -1) { - char gran_factor, chunk_factor, lose_factor; - unsigned long gran_base, chunk_base, lose_base; - printk(KERN_INFO "Found optimal setting for mtrr clean up\n"); i = index_good; - gran_base = to_size_factor(result[i].gran_sizek, &gran_factor), - chunk_base = to_size_factor(result[i].chunk_sizek, &chunk_factor), - lose_base = to_size_factor(result[i].lose_cover_sizek, &lose_factor), - printk(KERN_INFO "gran_size: %ld%c \tchunk_size: %ld%c \t", - gran_base, gran_factor, chunk_base, chunk_factor); - printk(KERN_CONT "num_reg: %d \tlose RAM: %ld%c\n", - result[i].num_reg, lose_base, lose_factor); + mtrr_print_out_one_result(i); + /* convert ranges to var ranges state */ chunk_size = result[i].chunk_sizek; chunk_size <<= 10; gran_size = result[i].gran_sizek; gran_size <<= 10; - debug_print++; x86_setup_var_mtrrs(range, nr_range, chunk_size, gran_size); - debug_print--; set_var_mtrr_all(address_bits); + printk(KERN_DEBUG "New variable MTRRs\n"); + print_out_mtrr_range_state(); return 1; + } else { + /* print out all */ + for (i = 0; i < NUM_RESULT; i++) + mtrr_print_out_one_result(i); } printk(KERN_INFO "mtrr_cleanup: can not find optimal value\n"); @@ -1562,7 +1559,6 @@ int __init mtrr_trim_uncached_memory(unsigned long end_pfn) { unsigned long i, base, size, highest_pfn = 0, def, dummy; mtrr_type type; - int nr_range; u64 total_trim_size; /* extra one for all 0 */ -- cgit v1.2.3 From d4f1b10365d4f03dd802433e0014cf503e6e930c Mon Sep 17 00:00:00 2001 From: Jike Song Date: Fri, 17 Oct 2008 13:25:07 +0800 Subject: x86: clean up comments wrt. rd{msr|tsc|pmc} The rdmsr instruction(et al) for i386 and x86-64 are semantically same. The only difference is how gcc interpret constraint "A" for these targets. Signed-off-by: Jike Song Signed-off-by: Ingo Molnar --- arch/x86/include/asm/msr.h | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/msr.h b/arch/x86/include/asm/msr.h index 46be2fa7ac26..478a9245aae1 100644 --- a/arch/x86/include/asm/msr.h +++ b/arch/x86/include/asm/msr.h @@ -22,10 +22,10 @@ static inline unsigned long long native_read_tscp(unsigned int *aux) } /* - * i386 calling convention returns 64-bit value in edx:eax, while - * x86_64 returns at rax. Also, the "A" constraint does not really - * mean rdx:rax in x86_64, so we need specialized behaviour for each - * architecture + * both i386 and x86_64 returns 64-bit value in edx:eax, but gcc's "A" + * constraint has different meanings. For i386, "A" means exactly + * edx:eax, while for x86_64 it doesn't mean rdx:rax or edx:eax. Instead, + * it means rax *or* rdx. */ #ifdef CONFIG_X86_64 #define DECLARE_ARGS(val, low, high) unsigned low, high -- cgit v1.2.3 From ad38dab01323a01e825555fc46863b73cd0efdc7 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Mon, 27 Oct 2008 13:30:56 -0700 Subject: x86: use the new byteorder headers Impact: cleanup, no functionality changed Signed-off-by: Harvey Harrison Signed-off-by: Ingo Molnar --- arch/x86/include/asm/byteorder.h | 74 ++++++++++++++++------------------------ 1 file changed, 29 insertions(+), 45 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/byteorder.h b/arch/x86/include/asm/byteorder.h index e02ae2d89acf..f110ad417df3 100644 --- a/arch/x86/include/asm/byteorder.h +++ b/arch/x86/include/asm/byteorder.h @@ -4,26 +4,33 @@ #include #include -#ifdef __GNUC__ +#define __LITTLE_ENDIAN -#ifdef __i386__ - -static inline __attribute_const__ __u32 ___arch__swab32(__u32 x) +static inline __attribute_const__ __u32 __arch_swab32(__u32 val) { -#ifdef CONFIG_X86_BSWAP - asm("bswap %0" : "=r" (x) : "0" (x)); -#else +#ifdef __i386__ +# ifdef CONFIG_X86_BSWAP + asm("bswap %0" : "=r" (val) : "0" (val)); +# else asm("xchgb %b0,%h0\n\t" /* swap lower bytes */ "rorl $16,%0\n\t" /* swap words */ "xchgb %b0,%h0" /* swap higher bytes */ - : "=q" (x) - : "0" (x)); + : "=q" (val) + : "0" (val)); +# endif + +#else /* __i386__ */ + asm("bswapl %0" + : "=r" (val) + : "0" (val)); #endif - return x; + return val; } +#define __arch_swab32 __arch_swab32 -static inline __attribute_const__ __u64 ___arch__swab64(__u64 val) +static inline __attribute_const__ __u64 __arch_swab64(__u64 val) { +#ifdef __i386__ union { struct { __u32 a; @@ -32,50 +39,27 @@ static inline __attribute_const__ __u64 ___arch__swab64(__u64 val) __u64 u; } v; v.u = val; -#ifdef CONFIG_X86_BSWAP +# ifdef CONFIG_X86_BSWAP asm("bswapl %0 ; bswapl %1 ; xchgl %0,%1" : "=r" (v.s.a), "=r" (v.s.b) : "0" (v.s.a), "1" (v.s.b)); -#else - v.s.a = ___arch__swab32(v.s.a); - v.s.b = ___arch__swab32(v.s.b); +# else + v.s.a = __arch_swab32(v.s.a); + v.s.b = __arch_swab32(v.s.b); asm("xchgl %0,%1" : "=r" (v.s.a), "=r" (v.s.b) : "0" (v.s.a), "1" (v.s.b)); -#endif +# endif return v.u; -} - #else /* __i386__ */ - -static inline __attribute_const__ __u64 ___arch__swab64(__u64 x) -{ asm("bswapq %0" - : "=r" (x) - : "0" (x)); - return x; -} - -static inline __attribute_const__ __u32 ___arch__swab32(__u32 x) -{ - asm("bswapl %0" - : "=r" (x) - : "0" (x)); - return x; -} - + : "=r" (val) + : "0" (val)); + return val; #endif +} +#define __arch_swab64 __arch_swab64 -/* Do not define swab16. Gcc is smart enough to recognize "C" version and - convert it into rotation or exhange. */ - -#define __arch__swab64(x) ___arch__swab64(x) -#define __arch__swab32(x) ___arch__swab32(x) - -#define __BYTEORDER_HAS_U64__ - -#endif /* __GNUC__ */ - -#include +#include #endif /* _ASM_X86_BYTEORDER_H */ -- cgit v1.2.3 From 96bf84b71255b0ee4fcee91e9acd1b5e73030eaf Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Wed, 29 Oct 2008 18:44:08 -0700 Subject: x86: signal: cosmetic unification of signr_convert() Impact: cleanup Make signr_convert() same. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/kernel/signal_32.c | 2 ++ arch/x86/kernel/signal_64.c | 6 ++++++ 2 files changed, 8 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/signal_32.c b/arch/x86/kernel/signal_32.c index 85a0d37cdae9..abf0df700fd0 100644 --- a/arch/x86/kernel/signal_32.c +++ b/arch/x86/kernel/signal_32.c @@ -503,10 +503,12 @@ static int __setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, */ static int signr_convert(int sig) { +#ifdef CONFIG_X86_32 struct thread_info *info = current_thread_info(); if (info->exec_domain && info->exec_domain->signal_invmap && sig < 32) return info->exec_domain->signal_invmap[sig]; +#endif /* CONFIG_X86_32 */ return sig; } diff --git a/arch/x86/kernel/signal_64.c b/arch/x86/kernel/signal_64.c index 3d0deb336745..a4b46e6392b1 100644 --- a/arch/x86/kernel/signal_64.c +++ b/arch/x86/kernel/signal_64.c @@ -295,6 +295,12 @@ static int __setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, */ static int signr_convert(int sig) { +#ifdef CONFIG_X86_32 + struct thread_info *info = current_thread_info(); + + if (info->exec_domain && info->exec_domain->signal_invmap && sig < 32) + return info->exec_domain->signal_invmap[sig]; +#endif /* CONFIG_X86_32 */ return sig; } -- cgit v1.2.3 From cabf503588961d202a33b3fd872767e9f6abbef7 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Wed, 29 Oct 2008 18:46:07 -0700 Subject: x86: signal: cosmetic unification of macros for setup_rt_frame() Impact: cleanup Add #ifdef directive for unification. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/kernel/signal_32.c | 12 ++++++++++++ arch/x86/kernel/signal_64.c | 14 ++++++++++++-- 2 files changed, 24 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/signal_32.c b/arch/x86/kernel/signal_32.c index abf0df700fd0..6f3b9a9cc123 100644 --- a/arch/x86/kernel/signal_32.c +++ b/arch/x86/kernel/signal_32.c @@ -512,10 +512,22 @@ static int signr_convert(int sig) return sig; } +#ifdef CONFIG_X86_32 + #define is_ia32 1 #define ia32_setup_frame __setup_frame #define ia32_setup_rt_frame __setup_rt_frame +#else /* !CONFIG_X86_32 */ + +#ifdef CONFIG_IA32_EMULATION +#define is_ia32 test_thread_flag(TIF_IA32) +#else /* !CONFIG_IA32_EMULATION */ +#define is_ia32 0 +#endif /* CONFIG_IA32_EMULATION */ + +#endif /* CONFIG_X86_32 */ + static int setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, sigset_t *set, struct pt_regs *regs) diff --git a/arch/x86/kernel/signal_64.c b/arch/x86/kernel/signal_64.c index a4b46e6392b1..49df79e05111 100644 --- a/arch/x86/kernel/signal_64.c +++ b/arch/x86/kernel/signal_64.c @@ -304,11 +304,21 @@ static int signr_convert(int sig) return sig; } +#ifdef CONFIG_X86_32 + +#define is_ia32 1 +#define ia32_setup_frame __setup_frame +#define ia32_setup_rt_frame __setup_rt_frame + +#else /* !CONFIG_X86_32 */ + #ifdef CONFIG_IA32_EMULATION #define is_ia32 test_thread_flag(TIF_IA32) -#else +#else /* !CONFIG_IA32_EMULATION */ #define is_ia32 0 -#endif +#endif /* CONFIG_IA32_EMULATION */ + +#endif /* CONFIG_X86_32 */ static int setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, -- cgit v1.2.3 From 57917752f51bcead3bb6c83d74137fbe342504ec Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Wed, 29 Oct 2008 18:46:40 -0700 Subject: x86: signal: cosmetic unification of NR_restart_syscall Impact: cleanup Add #ifdef directive to unify NR_restart_syscall. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/kernel/signal_32.c | 6 ++++++ arch/x86/kernel/signal_64.c | 5 +++++ 2 files changed, 11 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/signal_32.c b/arch/x86/kernel/signal_32.c index 6f3b9a9cc123..a0efc1b3c4c9 100644 --- a/arch/x86/kernel/signal_32.c +++ b/arch/x86/kernel/signal_32.c @@ -628,7 +628,13 @@ handle_signal(unsigned long sig, siginfo_t *info, struct k_sigaction *ka, return 0; } +#ifdef CONFIG_X86_32 #define NR_restart_syscall __NR_restart_syscall +#else /* !CONFIG_X86_32 */ +#define NR_restart_syscall \ + test_thread_flag(TIF_IA32) ? __NR_ia32_restart_syscall : __NR_restart_syscall +#endif /* CONFIG_X86_32 */ + /* * Note that 'init' is a special process: it doesn't get signals it doesn't * want to handle. Thus you cannot kill init even with a SIGKILL even by diff --git a/arch/x86/kernel/signal_64.c b/arch/x86/kernel/signal_64.c index 49df79e05111..83990db82f74 100644 --- a/arch/x86/kernel/signal_64.c +++ b/arch/x86/kernel/signal_64.c @@ -420,8 +420,13 @@ handle_signal(unsigned long sig, siginfo_t *info, struct k_sigaction *ka, return 0; } +#ifdef CONFIG_X86_32 +#define NR_restart_syscall __NR_restart_syscall +#else /* !CONFIG_X86_32 */ #define NR_restart_syscall \ test_thread_flag(TIF_IA32) ? __NR_ia32_restart_syscall : __NR_restart_syscall +#endif /* CONFIG_X86_32 */ + /* * Note that 'init' is a special process: it doesn't get signals it doesn't * want to handle. Thus you cannot kill init even with a SIGKILL even by -- cgit v1.2.3 From 7a5276889cfa96619bf863c87581005f46139986 Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Thu, 30 Oct 2008 10:38:24 +0000 Subject: x86: simplify X86_MPPARSE config option Impact: cleanup Signed-off-by: Jan Beulich Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 12 ++---------- 1 file changed, 2 insertions(+), 10 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 350bee1d54dc..f843de13e242 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -235,21 +235,13 @@ config X86_FIND_SMP_CONFIG def_bool y depends on X86_MPPARSE || X86_VOYAGER -if ACPI config X86_MPPARSE - def_bool y - bool "Enable MPS table" + bool "Enable MPS table" if ACPI + default y depends on X86_LOCAL_APIC help For old smp systems that do not have proper acpi support. Newer systems (esp with 64bit cpus) with acpi support, MADT and DSDT will override it -endif - -if !ACPI -config X86_MPPARSE - def_bool y - depends on X86_LOCAL_APIC -endif choice prompt "Subarchitecture Type" -- cgit v1.2.3 From b062f841b569791d3054e975cd85f48562161565 Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Thu, 30 Oct 2008 19:16:46 +0300 Subject: x86: nmi - add sensible names to nmi_watchdog boot param Impact: introduce nmi_watchdog=lapic and nmi_watchdog=ioapic aliases Add sensible names as "lapic" and "ioapic" to nmi_watchdog boot parameter. Sometimes it is not that easy to recall what exactly nmi_watchdog=1 does mean so we allow the using of symbolic names here. Old numeric values remain valid. Signed-off-by: Cyrill Gorcunov Signed-off-by: Ingo Molnar --- arch/x86/kernel/nmi.c | 15 ++++++++++----- 1 file changed, 10 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/nmi.c b/arch/x86/kernel/nmi.c index 2c97f07f1c2c..c4869e4532a3 100644 --- a/arch/x86/kernel/nmi.c +++ b/arch/x86/kernel/nmi.c @@ -199,12 +199,17 @@ static int __init setup_nmi_watchdog(char *str) ++str; } - get_option(&str, &nmi); - - if (nmi >= NMI_INVALID) - return 0; + if (!strncmp(str, "lapic", 5)) + nmi_watchdog = NMI_LOCAL_APIC; + else if (!strncmp(str, "ioapic", 6)) + nmi_watchdog = NMI_IO_APIC; + else { + get_option(&str, &nmi); + if (nmi >= NMI_INVALID) + return 0; + nmi_watchdog = nmi; + } - nmi_watchdog = nmi; return 1; } __setup("nmi_watchdog=", setup_nmi_watchdog); -- cgit v1.2.3 From 1cbd8b3fdcf56a3c39a7596512095c9e33221fa1 Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Thu, 30 Oct 2008 10:45:36 +0000 Subject: x86: add two missing unwind annotations Impact: improve debuginfo Signed-off-by: Jan Beulich Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_64.S | 2 ++ 1 file changed, 2 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index b86f332c96a6..ddeeb1052583 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -255,6 +255,7 @@ ENTRY(ret_from_fork) call schedule_tail GET_THREAD_INFO(%rcx) testl $(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT),TI_flags(%rcx) + CFI_REMEMBER_STATE jnz rff_trace rff_action: RESTORE_REST @@ -264,6 +265,7 @@ rff_action: jnz int_ret_from_sys_call RESTORE_TOP_OF_STACK %rdi,ARGOFFSET jmp ret_from_sys_call + CFI_RESTORE_STATE rff_trace: movq %rsp,%rdi call syscall_trace_leave -- cgit v1.2.3 From 17666f02b118099028522dfc3df00a235700e216 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 30 Oct 2008 16:08:32 -0400 Subject: ftrace: nmi safe code modification Impact: fix crashes that can occur in NMI handlers, if their code is modified Modifying code is something that needs special care. On SMP boxes, if code that is being modified is also being executed on another CPU, that CPU will have undefined results. The dynamic ftrace uses kstop_machine to make the system act like a uniprocessor system. But this does not address NMIs, that can still run on other CPUs. One approach to handle this is to make all code that are used by NMIs not be traced. But NMIs can call notifiers that spread throughout the kernel and this will be very hard to maintain, and the chance of missing a function is very high. The approach that this patch takes is to have the NMIs modify the code if the modification is taking place. The way this works is that just writing to code executing on another CPU is not harmful if what is written is the same as what exists. Two buffers are used: an IP buffer and a "code" buffer. The steps that the patcher takes are: 1) Put in the instruction pointer into the IP buffer and the new code into the "code" buffer. 2) Set a flag that says we are modifying code 3) Wait for any running NMIs to finish. 4) Write the code 5) clear the flag. 6) Wait for any running NMIs to finish. If an NMI is executed, it will also write the pending code. Multiple writes are OK, because what is being written is the same. Then the patcher must wait for all running NMIs to finish before going to the next line that must be patched. This is basically the RCU approach to code modification. Thanks to Ingo Molnar for suggesting the idea, and to Arjan van de Ven for his guidence on what is safe and what is not. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/arm/include/asm/ftrace.h | 5 ++ arch/powerpc/include/asm/ftrace.h | 5 ++ arch/sh/include/asm/ftrace.h | 5 ++ arch/sparc/include/asm/ftrace.h | 5 ++ arch/x86/include/asm/ftrace.h | 15 ++++++ arch/x86/kernel/ftrace.c | 107 +++++++++++++++++++++++++++++++++++++- include/linux/hardirq.h | 15 +++++- 7 files changed, 154 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/arm/include/asm/ftrace.h b/arch/arm/include/asm/ftrace.h index 39c8bc1a006a..d4c24a7a9280 100644 --- a/arch/arm/include/asm/ftrace.h +++ b/arch/arm/include/asm/ftrace.h @@ -1,6 +1,11 @@ #ifndef _ASM_ARM_FTRACE #define _ASM_ARM_FTRACE +#ifndef __ASSEMBLY__ +#define ftrace_nmi_enter() do { } while (0) +#define ftrace_nmi_exit() do { } while (0) +#endif + #ifdef CONFIG_FUNCTION_TRACER #define MCOUNT_ADDR ((long)(mcount)) #define MCOUNT_INSN_SIZE 4 /* sizeof mcount call */ diff --git a/arch/powerpc/include/asm/ftrace.h b/arch/powerpc/include/asm/ftrace.h index b298f7a631e6..7652755dc000 100644 --- a/arch/powerpc/include/asm/ftrace.h +++ b/arch/powerpc/include/asm/ftrace.h @@ -1,6 +1,11 @@ #ifndef _ASM_POWERPC_FTRACE #define _ASM_POWERPC_FTRACE +#ifndef __ASSEMBLY__ +#define ftrace_nmi_enter() do { } while (0) +#define ftrace_nmi_exit() do { } while (0) +#endif + #ifdef CONFIG_FUNCTION_TRACER #define MCOUNT_ADDR ((long)(_mcount)) #define MCOUNT_INSN_SIZE 4 /* sizeof mcount call */ diff --git a/arch/sh/include/asm/ftrace.h b/arch/sh/include/asm/ftrace.h index 3aed362c9463..cdf2cb0b9ffe 100644 --- a/arch/sh/include/asm/ftrace.h +++ b/arch/sh/include/asm/ftrace.h @@ -1,6 +1,11 @@ #ifndef __ASM_SH_FTRACE_H #define __ASM_SH_FTRACE_H +#ifndef __ASSEMBLY__ +#define ftrace_nmi_enter() do { } while (0) +#define ftrace_nmi_exit() do { } while (0) +#endif + #ifndef __ASSEMBLY__ extern void mcount(void); #endif diff --git a/arch/sparc/include/asm/ftrace.h b/arch/sparc/include/asm/ftrace.h index d27716cd38c1..33a95feeb137 100644 --- a/arch/sparc/include/asm/ftrace.h +++ b/arch/sparc/include/asm/ftrace.h @@ -1,6 +1,11 @@ #ifndef _ASM_SPARC64_FTRACE #define _ASM_SPARC64_FTRACE +#ifndef __ASSEMBLY__ +#define ftrace_nmi_enter() do { } while (0) +#define ftrace_nmi_exit() do { } while (0) +#endif + #ifdef CONFIG_MCOUNT #define MCOUNT_ADDR ((long)(_mcount)) #define MCOUNT_INSN_SIZE 4 /* sizeof mcount call */ diff --git a/arch/x86/include/asm/ftrace.h b/arch/x86/include/asm/ftrace.h index 9e8bc29b8b17..f2ed6b704a75 100644 --- a/arch/x86/include/asm/ftrace.h +++ b/arch/x86/include/asm/ftrace.h @@ -17,6 +17,21 @@ static inline unsigned long ftrace_call_adjust(unsigned long addr) */ return addr - 1; } + +#ifdef CONFIG_DYNAMIC_FTRACE +extern void ftrace_nmi_enter(void); +extern void ftrace_nmi_exit(void); +#else +#define ftrace_nmi_enter() do { } while (0) +#define ftrace_nmi_exit() do { } while (0) +#endif +#endif + +#else /* CONFIG_FUNCTION_TRACER */ + +#ifndef __ASSEMBLY__ +#define ftrace_nmi_enter() do { } while (0) +#define ftrace_nmi_exit() do { } while (0) #endif #endif /* CONFIG_FUNCTION_TRACER */ diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 50ea0ac8c9bf..fe5f859130b5 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -56,6 +56,111 @@ unsigned char *ftrace_call_replace(unsigned long ip, unsigned long addr) return calc.code; } +/* + * Modifying code must take extra care. On an SMP machine, if + * the code being modified is also being executed on another CPU + * that CPU will have undefined results and possibly take a GPF. + * We use kstop_machine to stop other CPUS from exectuing code. + * But this does not stop NMIs from happening. We still need + * to protect against that. We separate out the modification of + * the code to take care of this. + * + * Two buffers are added: An IP buffer and a "code" buffer. + * + * 1) Put in the instruction pointer into the IP buffer + * and the new code into the "code" buffer. + * 2) Set a flag that says we are modifying code + * 3) Wait for any running NMIs to finish. + * 4) Write the code + * 5) clear the flag. + * 6) Wait for any running NMIs to finish. + * + * If an NMI is executed, the first thing it does is to call + * "ftrace_nmi_enter". This will check if the flag is set to write + * and if it is, it will write what is in the IP and "code" buffers. + * + * The trick is, it does not matter if everyone is writing the same + * content to the code location. Also, if a CPU is executing code + * it is OK to write to that code location if the contents being written + * are the same as what exists. + */ + +static atomic_t in_nmi; +static int mod_code_status; +static int mod_code_write; +static void *mod_code_ip; +static void *mod_code_newcode; + +static void ftrace_mod_code(void) +{ + /* + * Yes, more than one CPU process can be writing to mod_code_status. + * (and the code itself) + * But if one were to fail, then they all should, and if one were + * to succeed, then they all should. + */ + mod_code_status = probe_kernel_write(mod_code_ip, mod_code_newcode, + MCOUNT_INSN_SIZE); + +} + +void ftrace_nmi_enter(void) +{ + atomic_inc(&in_nmi); + /* Must have in_nmi seen before reading write flag */ + smp_mb(); + if (mod_code_write) + ftrace_mod_code(); +} + +void ftrace_nmi_exit(void) +{ + /* Finish all executions before clearing in_nmi */ + smp_wmb(); + atomic_dec(&in_nmi); +} + +static void wait_for_nmi(void) +{ + while (atomic_read(&in_nmi)) + cpu_relax(); +} + +static int +do_ftrace_mod_code(unsigned long ip, void *new_code) +{ + mod_code_ip = (void *)ip; + mod_code_newcode = new_code; + + /* The buffers need to be visible before we let NMIs write them */ + smp_wmb(); + + mod_code_write = 1; + + /* Make sure write bit is visible before we wait on NMIs */ + smp_mb(); + + wait_for_nmi(); + + /* Make sure all running NMIs have finished before we write the code */ + smp_mb(); + + ftrace_mod_code(); + + /* Make sure the write happens before clearing the bit */ + smp_wmb(); + + mod_code_write = 0; + + /* make sure NMIs see the cleared bit */ + smp_mb(); + + wait_for_nmi(); + + return mod_code_status; +} + + int ftrace_modify_code(unsigned long ip, unsigned char *old_code, unsigned char *new_code) @@ -81,7 +186,7 @@ ftrace_modify_code(unsigned long ip, unsigned char *old_code, return -EINVAL; /* replace the text with the new text */ - if (probe_kernel_write((void *)ip, new_code, MCOUNT_INSN_SIZE)) + if (do_ftrace_mod_code(ip, new_code)) return -EPERM; sync_core(); diff --git a/include/linux/hardirq.h b/include/linux/hardirq.h index 181006cc94a0..0087cb43becf 100644 --- a/include/linux/hardirq.h +++ b/include/linux/hardirq.h @@ -5,6 +5,7 @@ #include #include #include +#include #include /* @@ -161,7 +162,17 @@ extern void irq_enter(void); */ extern void irq_exit(void); -#define nmi_enter() do { lockdep_off(); __irq_enter(); } while (0) -#define nmi_exit() do { __irq_exit(); lockdep_on(); } while (0) +#define nmi_enter() \ + do { \ + ftrace_nmi_enter(); \ + lockdep_off(); \ + __irq_enter(); \ + } while (0) +#define nmi_exit() \ + do { \ + __irq_exit(); \ + lockdep_on(); \ + ftrace_nmi_exit(); \ + } while (0) #endif /* LINUX_HARDIRQ_H */ -- cgit v1.2.3 From b807c3d0f8e39ed7cbbbe6da162650e305e8de15 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 30 Oct 2008 16:08:33 -0400 Subject: ftrace: nmi update statistics Impact: add more debug info to /debugfs/tracing/dyn_ftrace_total_info This patch adds dynamic ftrace NMI update statistics to the /debugfs/tracing/dyn_ftrace_total_info stat file. Signed-off-by: Steven Rostedt Cc: Peter Zijlstra Cc: Andrew Morton Cc: Linus Torvalds Cc: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/x86/kernel/ftrace.c | 26 ++++++++++++++++++++++++-- kernel/trace/trace.c | 31 ++++++++++++++++++++++++------- 2 files changed, 48 insertions(+), 9 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index fe5f859130b5..6685b0fc1b44 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -91,6 +91,19 @@ static int mod_code_write; static void *mod_code_ip; static void *mod_code_newcode; +static int nmi_wait_count; +static atomic_t nmi_update_count; + +int ftrace_arch_read_dyn_info(char *buf, int size) +{ + int r; + + r = snprintf(buf, size, "%u %u", + nmi_wait_count, + atomic_read(&nmi_update_count)); + return r; +} + static void ftrace_mod_code(void) { /* @@ -109,8 +122,10 @@ void ftrace_nmi_enter(void) atomic_inc(&in_nmi); /* Must have in_nmi seen before reading write flag */ smp_mb(); - if (mod_code_write) + if (mod_code_write) { ftrace_mod_code(); + atomic_inc(&nmi_update_count); + } } void ftrace_nmi_exit(void) @@ -122,8 +137,15 @@ void ftrace_nmi_exit(void) static void wait_for_nmi(void) { - while (atomic_read(&in_nmi)) + int waited = 0; + + while (atomic_read(&in_nmi)) { + waited = 1; cpu_relax(); + } + + if (waited) + nmi_wait_count++; } static int diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index a610ca771558..bc36febc0771 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2815,22 +2815,39 @@ static struct file_operations tracing_mark_fops = { #ifdef CONFIG_DYNAMIC_FTRACE +#define DYN_INFO_BUF_SIZE 1023 +static char ftrace_dyn_info_buffer[DYN_INFO_BUF_SIZE+1]; +static DEFINE_MUTEX(dyn_info_mutex); + +int __weak ftrace_arch_read_dyn_info(char *buf, int size) +{ + return 0; +} + static ssize_t -tracing_read_long(struct file *filp, char __user *ubuf, +tracing_read_dyn_info(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) { unsigned long *p = filp->private_data; - char buf[64]; + char *buf = ftrace_dyn_info_buffer; int r; - r = sprintf(buf, "%ld\n", *p); + mutex_lock(&dyn_info_mutex); + r = sprintf(buf, "%ld ", *p); - return simple_read_from_buffer(ubuf, cnt, ppos, buf, r); + r += ftrace_arch_read_dyn_info(buf+r, DYN_INFO_BUF_SIZE-r); + buf[r++] = '\n'; + + r = simple_read_from_buffer(ubuf, cnt, ppos, buf, r); + + mutex_unlock(&dyn_info_mutex); + + return r; } -static struct file_operations tracing_read_long_fops = { +static struct file_operations tracing_dyn_info_fops = { .open = tracing_open_generic, - .read = tracing_read_long, + .read = tracing_read_dyn_info, }; #endif @@ -2939,7 +2956,7 @@ static __init int tracer_init_debugfs(void) #ifdef CONFIG_DYNAMIC_FTRACE entry = debugfs_create_file("dyn_ftrace_total_info", 0444, d_tracer, &ftrace_update_tot_cnt, - &tracing_read_long_fops); + &tracing_dyn_info_fops); if (!entry) pr_warning("Could not create debugfs " "'dyn_ftrace_total_info' entry\n"); -- cgit v1.2.3 From 31498a01496ffca3b542bae72b8ec499cd9302db Mon Sep 17 00:00:00 2001 From: Huang Ying Date: Fri, 31 Oct 2008 09:48:02 +0800 Subject: kexec/i386: remove PAGE_SIZE alignment from relocate_kernel Impact: save kernel .text by loosening kexec page alignment This patch removes PAGE_SIZE alignment from relocate_kernel(). Before kexec jump patches are merged, control page is mapped to relocate_kernel in kexec page tables, so relocate_kernel must be PAGE_SIZE aligned. Now, control page is mapped to identity mapped address, so relocate_kernel need not to be PAGE_SIZE aligned any more. This can reduce a few KB from kernel text segement. Signed-off-by: Huang Ying Signed-off-by: Ingo Molnar --- arch/x86/kernel/relocate_kernel_32.S | 1 - 1 file changed, 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/relocate_kernel_32.S b/arch/x86/kernel/relocate_kernel_32.S index 6f50664b2ba5..377da3f78e8c 100644 --- a/arch/x86/kernel/relocate_kernel_32.S +++ b/arch/x86/kernel/relocate_kernel_32.S @@ -39,7 +39,6 @@ #define CP_PA_BACKUP_PAGES_MAP DATA(0x1c) .text - .align PAGE_SIZE .globl relocate_kernel relocate_kernel: /* Save the CPU context, used for jumping back */ -- cgit v1.2.3 From 92be3d6bdf2cb34972ab50e12ad4da1076e690da Mon Sep 17 00:00:00 2001 From: Huang Ying Date: Fri, 31 Oct 2008 09:48:08 +0800 Subject: kexec/i386: allocate page table pages dynamically Impact: save .text size when kexec is built in but not loaded This patch adds an architecture specific struct kimage_arch into struct kimage. The pointers to page table pages used by kexec are added to struct kimage_arch. The page tables pages are dynamically allocated in machine_kexec_prepare instead of statically from BSS segment. This will save up to 20k memory when kexec image is not loaded. Signed-off-by: Huang Ying Signed-off-by: Ingo Molnar --- arch/x86/include/asm/kexec.h | 14 ++++++++ arch/x86/kernel/machine_kexec_32.c | 67 ++++++++++++++++++++++++++------------ include/linux/kexec.h | 4 +++ 3 files changed, 64 insertions(+), 21 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/kexec.h b/arch/x86/include/asm/kexec.h index a1f22771a15a..df9c41a9c6ae 100644 --- a/arch/x86/include/asm/kexec.h +++ b/arch/x86/include/asm/kexec.h @@ -170,6 +170,20 @@ relocate_kernel(unsigned long indirection_page, unsigned long start_address) ATTRIB_NORET; #endif +#ifdef CONFIG_X86_32 +#define ARCH_HAS_KIMAGE_ARCH + +struct kimage_arch { + pgd_t *pgd; +#ifdef CONFIG_X86_PAE + pmd_t *pmd0; + pmd_t *pmd1; +#endif + pte_t *pte0; + pte_t *pte1; +}; +#endif + #endif /* __ASSEMBLY__ */ #endif /* _ASM_X86_KEXEC_H */ diff --git a/arch/x86/kernel/machine_kexec_32.c b/arch/x86/kernel/machine_kexec_32.c index 7a385746509a..1100312847a5 100644 --- a/arch/x86/kernel/machine_kexec_32.c +++ b/arch/x86/kernel/machine_kexec_32.c @@ -13,6 +13,7 @@ #include #include #include +#include #include #include @@ -25,15 +26,6 @@ #include #include -#define PAGE_ALIGNED __attribute__ ((__aligned__(PAGE_SIZE))) -static u32 kexec_pgd[1024] PAGE_ALIGNED; -#ifdef CONFIG_X86_PAE -static u32 kexec_pmd0[1024] PAGE_ALIGNED; -static u32 kexec_pmd1[1024] PAGE_ALIGNED; -#endif -static u32 kexec_pte0[1024] PAGE_ALIGNED; -static u32 kexec_pte1[1024] PAGE_ALIGNED; - static void set_idt(void *newidt, __u16 limit) { struct desc_ptr curidt; @@ -76,6 +68,37 @@ static void load_segments(void) #undef __STR } +static void machine_kexec_free_page_tables(struct kimage *image) +{ + free_page((unsigned long)image->arch.pgd); +#ifdef CONFIG_X86_PAE + free_page((unsigned long)image->arch.pmd0); + free_page((unsigned long)image->arch.pmd1); +#endif + free_page((unsigned long)image->arch.pte0); + free_page((unsigned long)image->arch.pte1); +} + +static int machine_kexec_alloc_page_tables(struct kimage *image) +{ + image->arch.pgd = (pgd_t *)get_zeroed_page(GFP_KERNEL); +#ifdef CONFIG_X86_PAE + image->arch.pmd0 = (pmd_t *)get_zeroed_page(GFP_KERNEL); + image->arch.pmd1 = (pmd_t *)get_zeroed_page(GFP_KERNEL); +#endif + image->arch.pte0 = (pte_t *)get_zeroed_page(GFP_KERNEL); + image->arch.pte1 = (pte_t *)get_zeroed_page(GFP_KERNEL); + if (!image->arch.pgd || +#ifdef CONFIG_X86_PAE + !image->arch.pmd0 || !image->arch.pmd1 || +#endif + !image->arch.pte0 || !image->arch.pte1) { + machine_kexec_free_page_tables(image); + return -ENOMEM; + } + return 0; +} + /* * A architecture hook called to validate the * proposed image and prepare the control pages @@ -87,13 +110,14 @@ static void load_segments(void) * reboot code buffer to allow us to avoid allocations * later. * - * Make control page executable. + * - Make control page executable. + * - Allocate page tables */ int machine_kexec_prepare(struct kimage *image) { if (nx_enabled) set_pages_x(image->control_code_page, 1); - return 0; + return machine_kexec_alloc_page_tables(image); } /* @@ -104,6 +128,7 @@ void machine_kexec_cleanup(struct kimage *image) { if (nx_enabled) set_pages_nx(image->control_code_page, 1); + machine_kexec_free_page_tables(image); } /* @@ -150,18 +175,18 @@ void machine_kexec(struct kimage *image) relocate_kernel_ptr = control_page; page_list[PA_CONTROL_PAGE] = __pa(control_page); page_list[VA_CONTROL_PAGE] = (unsigned long)control_page; - page_list[PA_PGD] = __pa(kexec_pgd); - page_list[VA_PGD] = (unsigned long)kexec_pgd; + page_list[PA_PGD] = __pa(image->arch.pgd); + page_list[VA_PGD] = (unsigned long)image->arch.pgd; #ifdef CONFIG_X86_PAE - page_list[PA_PMD_0] = __pa(kexec_pmd0); - page_list[VA_PMD_0] = (unsigned long)kexec_pmd0; - page_list[PA_PMD_1] = __pa(kexec_pmd1); - page_list[VA_PMD_1] = (unsigned long)kexec_pmd1; + page_list[PA_PMD_0] = __pa(image->arch.pmd0); + page_list[VA_PMD_0] = (unsigned long)image->arch.pmd0; + page_list[PA_PMD_1] = __pa(image->arch.pmd1); + page_list[VA_PMD_1] = (unsigned long)image->arch.pmd1; #endif - page_list[PA_PTE_0] = __pa(kexec_pte0); - page_list[VA_PTE_0] = (unsigned long)kexec_pte0; - page_list[PA_PTE_1] = __pa(kexec_pte1); - page_list[VA_PTE_1] = (unsigned long)kexec_pte1; + page_list[PA_PTE_0] = __pa(image->arch.pte0); + page_list[VA_PTE_0] = (unsigned long)image->arch.pte0; + page_list[PA_PTE_1] = __pa(image->arch.pte1); + page_list[VA_PTE_1] = (unsigned long)image->arch.pte1; if (image->type == KEXEC_TYPE_DEFAULT) page_list[PA_SWAP_PAGE] = (page_to_pfn(image->swap_page) diff --git a/include/linux/kexec.h b/include/linux/kexec.h index 17f76fc05173..adc34f2c6eff 100644 --- a/include/linux/kexec.h +++ b/include/linux/kexec.h @@ -100,6 +100,10 @@ struct kimage { #define KEXEC_TYPE_DEFAULT 0 #define KEXEC_TYPE_CRASH 1 unsigned int preserve_context : 1; + +#ifdef ARCH_HAS_KIMAGE_ARCH + struct kimage_arch arch; +#endif }; -- cgit v1.2.3 From 9868ee63b896ee4d2ceb8c292e88d7f4e66caaf9 Mon Sep 17 00:00:00 2001 From: Huang Ying Date: Fri, 31 Oct 2008 09:48:15 +0800 Subject: kexec/i386: setup kexec page table in C Impact: change the kexec bootstrap code implementation from assembly to C This patch transforms the kexec page tables setup code from assembler code to C code in machine_kexec_prepare. This improves readability and reduces code line number. Signed-off-by: Huang Ying Signed-off-by: Ingo Molnar --- arch/x86/include/asm/kexec.h | 17 +----- arch/x86/kernel/machine_kexec_32.c | 59 ++++++++++++++---- arch/x86/kernel/relocate_kernel_32.S | 114 ----------------------------------- 3 files changed, 49 insertions(+), 141 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/kexec.h b/arch/x86/include/asm/kexec.h index df9c41a9c6ae..c61d8b2ab8b9 100644 --- a/arch/x86/include/asm/kexec.h +++ b/arch/x86/include/asm/kexec.h @@ -5,21 +5,8 @@ # define PA_CONTROL_PAGE 0 # define VA_CONTROL_PAGE 1 # define PA_PGD 2 -# define VA_PGD 3 -# define PA_PTE_0 4 -# define VA_PTE_0 5 -# define PA_PTE_1 6 -# define VA_PTE_1 7 -# define PA_SWAP_PAGE 8 -# ifdef CONFIG_X86_PAE -# define PA_PMD_0 9 -# define VA_PMD_0 10 -# define PA_PMD_1 11 -# define VA_PMD_1 12 -# define PAGES_NR 13 -# else -# define PAGES_NR 9 -# endif +# define PA_SWAP_PAGE 3 +# define PAGES_NR 4 #else # define PA_CONTROL_PAGE 0 # define VA_CONTROL_PAGE 1 diff --git a/arch/x86/kernel/machine_kexec_32.c b/arch/x86/kernel/machine_kexec_32.c index 1100312847a5..37f420018a41 100644 --- a/arch/x86/kernel/machine_kexec_32.c +++ b/arch/x86/kernel/machine_kexec_32.c @@ -99,6 +99,45 @@ static int machine_kexec_alloc_page_tables(struct kimage *image) return 0; } +static void machine_kexec_page_table_set_one( + pgd_t *pgd, pmd_t *pmd, pte_t *pte, + unsigned long vaddr, unsigned long paddr) +{ + pud_t *pud; + + pgd += pgd_index(vaddr); +#ifdef CONFIG_X86_PAE + if (!(pgd_val(*pgd) & _PAGE_PRESENT)) + set_pgd(pgd, __pgd(__pa(pmd) | _PAGE_PRESENT)); +#endif + pud = pud_offset(pgd, vaddr); + pmd = pmd_offset(pud, vaddr); + if (!(pmd_val(*pmd) & _PAGE_PRESENT)) + set_pmd(pmd, __pmd(__pa(pte) | _PAGE_TABLE)); + pte = pte_offset_kernel(pmd, vaddr); + set_pte(pte, pfn_pte(paddr >> PAGE_SHIFT, PAGE_KERNEL_EXEC)); +} + +static void machine_kexec_prepare_page_tables(struct kimage *image) +{ + void *control_page; + pmd_t *pmd = 0; + + control_page = page_address(image->control_code_page); +#ifdef CONFIG_X86_PAE + pmd = image->arch.pmd0; +#endif + machine_kexec_page_table_set_one( + image->arch.pgd, pmd, image->arch.pte0, + (unsigned long)control_page, __pa(control_page)); +#ifdef CONFIG_X86_PAE + pmd = image->arch.pmd1; +#endif + machine_kexec_page_table_set_one( + image->arch.pgd, pmd, image->arch.pte1, + __pa(control_page), __pa(control_page)); +} + /* * A architecture hook called to validate the * proposed image and prepare the control pages @@ -112,12 +151,19 @@ static int machine_kexec_alloc_page_tables(struct kimage *image) * * - Make control page executable. * - Allocate page tables + * - Setup page tables */ int machine_kexec_prepare(struct kimage *image) { + int error; + if (nx_enabled) set_pages_x(image->control_code_page, 1); - return machine_kexec_alloc_page_tables(image); + error = machine_kexec_alloc_page_tables(image); + if (error) + return error; + machine_kexec_prepare_page_tables(image); + return 0; } /* @@ -176,17 +222,6 @@ void machine_kexec(struct kimage *image) page_list[PA_CONTROL_PAGE] = __pa(control_page); page_list[VA_CONTROL_PAGE] = (unsigned long)control_page; page_list[PA_PGD] = __pa(image->arch.pgd); - page_list[VA_PGD] = (unsigned long)image->arch.pgd; -#ifdef CONFIG_X86_PAE - page_list[PA_PMD_0] = __pa(image->arch.pmd0); - page_list[VA_PMD_0] = (unsigned long)image->arch.pmd0; - page_list[PA_PMD_1] = __pa(image->arch.pmd1); - page_list[VA_PMD_1] = (unsigned long)image->arch.pmd1; -#endif - page_list[PA_PTE_0] = __pa(image->arch.pte0); - page_list[VA_PTE_0] = (unsigned long)image->arch.pte0; - page_list[PA_PTE_1] = __pa(image->arch.pte1); - page_list[VA_PTE_1] = (unsigned long)image->arch.pte1; if (image->type == KEXEC_TYPE_DEFAULT) page_list[PA_SWAP_PAGE] = (page_to_pfn(image->swap_page) diff --git a/arch/x86/kernel/relocate_kernel_32.S b/arch/x86/kernel/relocate_kernel_32.S index 377da3f78e8c..a160f3119725 100644 --- a/arch/x86/kernel/relocate_kernel_32.S +++ b/arch/x86/kernel/relocate_kernel_32.S @@ -10,15 +10,12 @@ #include #include #include -#include /* * Must be relocatable PIC code callable as a C function */ #define PTR(x) (x << 2) -#define PAGE_ATTR (_PAGE_PRESENT | _PAGE_RW | _PAGE_ACCESSED | _PAGE_DIRTY) -#define PAE_PGD_ATTR (_PAGE_PRESENT) /* control_page + KEXEC_CONTROL_CODE_MAX_SIZE * ~ control_page + PAGE_SIZE are used as data storage and stack for @@ -59,117 +56,6 @@ relocate_kernel: movl %cr4, %eax movl %eax, CR4(%edi) -#ifdef CONFIG_X86_PAE - /* map the control page at its virtual address */ - - movl PTR(VA_PGD)(%ebp), %edi - movl PTR(VA_CONTROL_PAGE)(%ebp), %eax - andl $0xc0000000, %eax - shrl $27, %eax - addl %edi, %eax - - movl PTR(PA_PMD_0)(%ebp), %edx - orl $PAE_PGD_ATTR, %edx - movl %edx, (%eax) - - movl PTR(VA_PMD_0)(%ebp), %edi - movl PTR(VA_CONTROL_PAGE)(%ebp), %eax - andl $0x3fe00000, %eax - shrl $18, %eax - addl %edi, %eax - - movl PTR(PA_PTE_0)(%ebp), %edx - orl $PAGE_ATTR, %edx - movl %edx, (%eax) - - movl PTR(VA_PTE_0)(%ebp), %edi - movl PTR(VA_CONTROL_PAGE)(%ebp), %eax - andl $0x001ff000, %eax - shrl $9, %eax - addl %edi, %eax - - movl PTR(PA_CONTROL_PAGE)(%ebp), %edx - orl $PAGE_ATTR, %edx - movl %edx, (%eax) - - /* identity map the control page at its physical address */ - - movl PTR(VA_PGD)(%ebp), %edi - movl PTR(PA_CONTROL_PAGE)(%ebp), %eax - andl $0xc0000000, %eax - shrl $27, %eax - addl %edi, %eax - - movl PTR(PA_PMD_1)(%ebp), %edx - orl $PAE_PGD_ATTR, %edx - movl %edx, (%eax) - - movl PTR(VA_PMD_1)(%ebp), %edi - movl PTR(PA_CONTROL_PAGE)(%ebp), %eax - andl $0x3fe00000, %eax - shrl $18, %eax - addl %edi, %eax - - movl PTR(PA_PTE_1)(%ebp), %edx - orl $PAGE_ATTR, %edx - movl %edx, (%eax) - - movl PTR(VA_PTE_1)(%ebp), %edi - movl PTR(PA_CONTROL_PAGE)(%ebp), %eax - andl $0x001ff000, %eax - shrl $9, %eax - addl %edi, %eax - - movl PTR(PA_CONTROL_PAGE)(%ebp), %edx - orl $PAGE_ATTR, %edx - movl %edx, (%eax) -#else - /* map the control page at its virtual address */ - - movl PTR(VA_PGD)(%ebp), %edi - movl PTR(VA_CONTROL_PAGE)(%ebp), %eax - andl $0xffc00000, %eax - shrl $20, %eax - addl %edi, %eax - - movl PTR(PA_PTE_0)(%ebp), %edx - orl $PAGE_ATTR, %edx - movl %edx, (%eax) - - movl PTR(VA_PTE_0)(%ebp), %edi - movl PTR(VA_CONTROL_PAGE)(%ebp), %eax - andl $0x003ff000, %eax - shrl $10, %eax - addl %edi, %eax - - movl PTR(PA_CONTROL_PAGE)(%ebp), %edx - orl $PAGE_ATTR, %edx - movl %edx, (%eax) - - /* identity map the control page at its physical address */ - - movl PTR(VA_PGD)(%ebp), %edi - movl PTR(PA_CONTROL_PAGE)(%ebp), %eax - andl $0xffc00000, %eax - shrl $20, %eax - addl %edi, %eax - - movl PTR(PA_PTE_1)(%ebp), %edx - orl $PAGE_ATTR, %edx - movl %edx, (%eax) - - movl PTR(VA_PTE_1)(%ebp), %edi - movl PTR(PA_CONTROL_PAGE)(%ebp), %eax - andl $0x003ff000, %eax - shrl $10, %eax - addl %edi, %eax - - movl PTR(PA_CONTROL_PAGE)(%ebp), %edx - orl $PAGE_ATTR, %edx - movl %edx, (%eax) -#endif - -relocate_new_kernel: /* read the arguments and say goodbye to the stack */ movl 20+4(%esp), %ebx /* page_list */ movl 20+8(%esp), %ebp /* list of pages */ -- cgit v1.2.3 From a26a2a27396c0a0877aa701f8f92d08ba550a6c9 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 31 Oct 2008 00:03:22 -0400 Subject: ftrace: nmi safe code clean ups Impact: cleanup This patch cleans up the NMI safe code for dynamic ftrace as suggested by Andrew Morton. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/arm/include/asm/ftrace.h | 4 ++-- arch/powerpc/include/asm/ftrace.h | 4 ++-- arch/sh/include/asm/ftrace.h | 4 ++-- arch/sparc/include/asm/ftrace.h | 4 ++-- arch/x86/include/asm/ftrace.h | 10 +++++----- arch/x86/kernel/ftrace.c | 16 ++++++++-------- include/linux/ftrace.h | 3 +++ kernel/trace/trace.c | 9 ++++----- 8 files changed, 28 insertions(+), 26 deletions(-) (limited to 'arch/x86') diff --git a/arch/arm/include/asm/ftrace.h b/arch/arm/include/asm/ftrace.h index d4c24a7a9280..3f3a1d1508ea 100644 --- a/arch/arm/include/asm/ftrace.h +++ b/arch/arm/include/asm/ftrace.h @@ -2,8 +2,8 @@ #define _ASM_ARM_FTRACE #ifndef __ASSEMBLY__ -#define ftrace_nmi_enter() do { } while (0) -#define ftrace_nmi_exit() do { } while (0) +static inline void ftrace_nmi_enter(void) { } +static inline void ftrace_nmi_exit(void) { } #endif #ifdef CONFIG_FUNCTION_TRACER diff --git a/arch/powerpc/include/asm/ftrace.h b/arch/powerpc/include/asm/ftrace.h index 7652755dc000..1cd72700fbc0 100644 --- a/arch/powerpc/include/asm/ftrace.h +++ b/arch/powerpc/include/asm/ftrace.h @@ -2,8 +2,8 @@ #define _ASM_POWERPC_FTRACE #ifndef __ASSEMBLY__ -#define ftrace_nmi_enter() do { } while (0) -#define ftrace_nmi_exit() do { } while (0) +static inline void ftrace_nmi_enter(void) { } +static inline void ftrace_nmi_exit(void) { } #endif #ifdef CONFIG_FUNCTION_TRACER diff --git a/arch/sh/include/asm/ftrace.h b/arch/sh/include/asm/ftrace.h index cdf2cb0b9ffe..31ada0370cb6 100644 --- a/arch/sh/include/asm/ftrace.h +++ b/arch/sh/include/asm/ftrace.h @@ -2,8 +2,8 @@ #define __ASM_SH_FTRACE_H #ifndef __ASSEMBLY__ -#define ftrace_nmi_enter() do { } while (0) -#define ftrace_nmi_exit() do { } while (0) +static inline void ftrace_nmi_enter(void) { } +static inline void ftrace_nmi_exit(void) { } #endif #ifndef __ASSEMBLY__ diff --git a/arch/sparc/include/asm/ftrace.h b/arch/sparc/include/asm/ftrace.h index 33a95feeb137..62055ac0496e 100644 --- a/arch/sparc/include/asm/ftrace.h +++ b/arch/sparc/include/asm/ftrace.h @@ -2,8 +2,8 @@ #define _ASM_SPARC64_FTRACE #ifndef __ASSEMBLY__ -#define ftrace_nmi_enter() do { } while (0) -#define ftrace_nmi_exit() do { } while (0) +static inline void ftrace_nmi_enter(void) { } +static inline void ftrace_nmi_exit(void) { } #endif #ifdef CONFIG_MCOUNT diff --git a/arch/x86/include/asm/ftrace.h b/arch/x86/include/asm/ftrace.h index f2ed6b704a75..a23468194b8c 100644 --- a/arch/x86/include/asm/ftrace.h +++ b/arch/x86/include/asm/ftrace.h @@ -22,16 +22,16 @@ static inline unsigned long ftrace_call_adjust(unsigned long addr) extern void ftrace_nmi_enter(void); extern void ftrace_nmi_exit(void); #else -#define ftrace_nmi_enter() do { } while (0) -#define ftrace_nmi_exit() do { } while (0) -#endif +static inline void ftrace_nmi_enter(void) { } +static inline void ftrace_nmi_exit(void) { } #endif +#endif /* __ASSEMBLY__ */ #else /* CONFIG_FUNCTION_TRACER */ #ifndef __ASSEMBLY__ -#define ftrace_nmi_enter() do { } while (0) -#define ftrace_nmi_exit() do { } while (0) +static inline void ftrace_nmi_enter(void) { } +static inline void ftrace_nmi_exit(void) { } #endif #endif /* CONFIG_FUNCTION_TRACER */ diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 6685b0fc1b44..69149337f2fe 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -67,7 +67,7 @@ unsigned char *ftrace_call_replace(unsigned long ip, unsigned long addr) * * Two buffers are added: An IP buffer and a "code" buffer. * - * 1) Put in the instruction pointer into the IP buffer + * 1) Put the instruction pointer into the IP buffer * and the new code into the "code" buffer. * 2) Set a flag that says we are modifying code * 3) Wait for any running NMIs to finish. @@ -85,14 +85,14 @@ unsigned char *ftrace_call_replace(unsigned long ip, unsigned long addr) * are the same as what exists. */ -static atomic_t in_nmi; -static int mod_code_status; -static int mod_code_write; -static void *mod_code_ip; -static void *mod_code_newcode; +static atomic_t in_nmi = ATOMIC_INIT(0); +static int mod_code_status; /* holds return value of text write */ +static int mod_code_write; /* set when NMI should do the write */ +static void *mod_code_ip; /* holds the IP to write to */ +static void *mod_code_newcode; /* holds the text to write to the IP */ -static int nmi_wait_count; -static atomic_t nmi_update_count; +static unsigned nmi_wait_count; +static atomic_t nmi_update_count = ATOMIC_INIT(0); int ftrace_arch_read_dyn_info(char *buf, int size) { diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 703eb53cfa2b..22240dfe912e 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -74,6 +74,9 @@ extern void ftrace_caller(void); extern void ftrace_call(void); extern void mcount_call(void); +/* May be defined in arch */ +extern int ftrace_arch_read_dyn_info(char *buf, int size); + /** * ftrace_modify_code - modify code segment * @ip: the address of the code segment diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index bc36febc0771..7f86067d760c 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2815,10 +2815,6 @@ static struct file_operations tracing_mark_fops = { #ifdef CONFIG_DYNAMIC_FTRACE -#define DYN_INFO_BUF_SIZE 1023 -static char ftrace_dyn_info_buffer[DYN_INFO_BUF_SIZE+1]; -static DEFINE_MUTEX(dyn_info_mutex); - int __weak ftrace_arch_read_dyn_info(char *buf, int size) { return 0; @@ -2828,14 +2824,17 @@ static ssize_t tracing_read_dyn_info(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) { + static char ftrace_dyn_info_buffer[1024]; + static DEFINE_MUTEX(dyn_info_mutex); unsigned long *p = filp->private_data; char *buf = ftrace_dyn_info_buffer; + int size = ARRAY_SIZE(ftrace_dyn_info_buffer); int r; mutex_lock(&dyn_info_mutex); r = sprintf(buf, "%ld ", *p); - r += ftrace_arch_read_dyn_info(buf+r, DYN_INFO_BUF_SIZE-r); + r += ftrace_arch_read_dyn_info(buf+r, (size-1)-r); buf[r++] = '\n'; r = simple_read_from_buffer(ubuf, cnt, ppos, buf, r); -- cgit v1.2.3 From a376f30a95a796cde81d6dffde0f5243c8bd8f92 Mon Sep 17 00:00:00 2001 From: Zhaolei Date: Fri, 31 Oct 2008 17:43:04 +0800 Subject: x86: avoid duplicate running of pud_offset and pmd_offset in one_md_table_init() Impact: simplify implementation, cleanup If !(pgd_val(*pgd) & _PAGE_PRESENT) in PAE mode, we need not get value of pmd_table again. Signed-off-by: Zhao Lei Signed-off-by: Ingo Molnar --- arch/x86/mm/init_32.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c index 8396868e82c5..7f8a2daa3fde 100644 --- a/arch/x86/mm/init_32.c +++ b/arch/x86/mm/init_32.c @@ -102,6 +102,8 @@ static pmd_t * __init one_md_table_init(pgd_t *pgd) set_pgd(pgd, __pgd(__pa(pmd_table) | _PAGE_PRESENT)); pud = pud_offset(pgd, 0); BUG_ON(pmd_table != pmd_offset(pud, 0)); + + return pmd_table; } #endif pud = pud_offset(pgd, 0); -- cgit v1.2.3 From b2bcc7b299f37037b4a78dc1538e5d6508ae8110 Mon Sep 17 00:00:00 2001 From: Alok Kataria Date: Fri, 31 Oct 2008 11:59:53 -0700 Subject: x86: add a synthetic TSC_RELIABLE feature bit Impact: None, bit reservation only Add a synthetic TSC_RELIABLE feature bit which will be used to mark TSC as reliable so that we could skip all the runtime checks for TSC stablity, which have false positives in virtual environment. Signed-off-by: Alok N Kataria Signed-off-by: Dan Hecht Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/cpufeature.h | 1 + 1 file changed, 1 insertion(+) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/cpufeature.h b/arch/x86/include/asm/cpufeature.h index cfdf8c2c5c31..e490a7932a0d 100644 --- a/arch/x86/include/asm/cpufeature.h +++ b/arch/x86/include/asm/cpufeature.h @@ -92,6 +92,7 @@ #define X86_FEATURE_NOPL (3*32+20) /* The NOPL (0F 1F) instructions */ #define X86_FEATURE_AMDC1E (3*32+21) /* AMD C1E detected */ #define X86_FEATURE_XTOPOLOGY (3*32+22) /* cpu topology enum extensions */ +#define X86_FEATURE_TSC_RELIABLE (3*32+23) /* TSC is known to be reliable */ /* Intel-defined CPU features, CPUID level 0x00000001 (ecx), word 4 */ #define X86_FEATURE_XMM3 (4*32+ 0) /* "pni" SSE-3 */ -- cgit v1.2.3 From 49ab56ac6e1b907b7dadb72a4012460359feaf0e Mon Sep 17 00:00:00 2001 From: Alok Kataria Date: Sat, 1 Nov 2008 18:34:37 -0700 Subject: x86: add X86_FEATURE_HYPERVISOR feature bit Impact: Number declaration only. Add X86_FEATURE_HYPERVISOR bit (CPUID level 1, ECX, bit 31). Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/cpufeature.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/cpufeature.h b/arch/x86/include/asm/cpufeature.h index e490a7932a0d..694d1f8f1bee 100644 --- a/arch/x86/include/asm/cpufeature.h +++ b/arch/x86/include/asm/cpufeature.h @@ -118,6 +118,7 @@ #define X86_FEATURE_XSAVE (4*32+26) /* XSAVE/XRSTOR/XSETBV/XGETBV */ #define X86_FEATURE_OSXSAVE (4*32+27) /* "" XSAVE enabled in the OS */ #define X86_FEATURE_AVX (4*32+28) /* Advanced Vector Extensions */ +#define X86_FEATURE_HYPERVISOR (4*32+31) /* Running on a hypervisor */ /* VIA/Cyrix/Centaur-defined CPU features, CPUID level 0xC0000001, word 5 */ #define X86_FEATURE_XSTORE (5*32+ 2) /* "rng" RNG present (xstore) */ @@ -238,6 +239,7 @@ extern const char * const x86_power_flags[32]; #define cpu_has_xmm4_2 boot_cpu_has(X86_FEATURE_XMM4_2) #define cpu_has_x2apic boot_cpu_has(X86_FEATURE_X2APIC) #define cpu_has_xsave boot_cpu_has(X86_FEATURE_XSAVE) +#define cpu_has_hypervisor boot_cpu_has(X86_FEATURE_HYPERVISOR) #if defined(CONFIG_X86_INVLPG) || defined(CONFIG_X86_64) # define cpu_has_invlpg 1 -- cgit v1.2.3 From 88b094fb8d4fe43b7025ea8d487059e8813e02cd Mon Sep 17 00:00:00 2001 From: Alok Kataria Date: Mon, 27 Oct 2008 10:41:46 -0700 Subject: x86: Hypervisor detection and get tsc_freq from hypervisor Impact: Changes timebase calibration on Vmware. v3->v2 : Abstract the hypervisor detection and feature (tsc_freq) request behind a hypervisor.c file v2->v1 : Add a x86_hyper_vendor field to the cpuinfo_x86 structure. This avoids multiple calls to the hypervisor detection function. This patch adds function to detect if we are running under VMware. The current way to check if we are on VMware is following, # check if "hypervisor present bit" is set, if so read the 0x40000000 cpuid leaf and check for "VMwareVMware" signature. # if the above fails, check the DMI vendors name for "VMware" string if we find one we query the VMware hypervisor port to check if we are under VMware. The DMI + "VMware hypervisor port check" is needed for older VMware products, which don't implement the hypervisor signature cpuid leaf. Also note that since we are checking for the DMI signature the hypervisor port should never be accessed on native hardware. This patch also adds a hypervisor_get_tsc_freq function, instead of calibrating the frequency which can be error prone in virtualized environment, we ask the hypervisor for it. We get the frequency from the hypervisor by accessing the hypervisor port if we are running on VMware. Other hypervisors too can add code to the generic routine to get frequency on their platform. Signed-off-by: Alok N Kataria Signed-off-by: Dan Hecht Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/hypervisor.h | 26 ++++++++++++ arch/x86/include/asm/processor.h | 4 ++ arch/x86/include/asm/vmware.h | 26 ++++++++++++ arch/x86/kernel/cpu/Makefile | 1 + arch/x86/kernel/cpu/common.c | 2 + arch/x86/kernel/cpu/hypervisor.c | 48 +++++++++++++++++++++ arch/x86/kernel/cpu/vmware.c | 88 +++++++++++++++++++++++++++++++++++++++ arch/x86/kernel/setup.c | 7 ++++ arch/x86/kernel/tsc.c | 9 +++- 9 files changed, 210 insertions(+), 1 deletion(-) create mode 100644 arch/x86/include/asm/hypervisor.h create mode 100644 arch/x86/include/asm/vmware.h create mode 100644 arch/x86/kernel/cpu/hypervisor.c create mode 100644 arch/x86/kernel/cpu/vmware.c (limited to 'arch/x86') diff --git a/arch/x86/include/asm/hypervisor.h b/arch/x86/include/asm/hypervisor.h new file mode 100644 index 000000000000..369f5c5d09a1 --- /dev/null +++ b/arch/x86/include/asm/hypervisor.h @@ -0,0 +1,26 @@ +/* + * Copyright (C) 2008, VMware, Inc. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or + * NON INFRINGEMENT. See the GNU General Public License for more + * details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. + * + */ +#ifndef ASM_X86__HYPERVISOR_H +#define ASM_X86__HYPERVISOR_H + +extern unsigned long get_hypervisor_tsc_freq(void); +extern void init_hypervisor(struct cpuinfo_x86 *c); + +#endif diff --git a/arch/x86/include/asm/processor.h b/arch/x86/include/asm/processor.h index 5ca01e383269..a570eafa4755 100644 --- a/arch/x86/include/asm/processor.h +++ b/arch/x86/include/asm/processor.h @@ -110,6 +110,7 @@ struct cpuinfo_x86 { /* Index into per_cpu list: */ u16 cpu_index; #endif + unsigned int x86_hyper_vendor; } __attribute__((__aligned__(SMP_CACHE_BYTES))); #define X86_VENDOR_INTEL 0 @@ -123,6 +124,9 @@ struct cpuinfo_x86 { #define X86_VENDOR_UNKNOWN 0xff +#define X86_HYPER_VENDOR_NONE 0 +#define X86_HYPER_VENDOR_VMWARE 1 + /* * capabilities of CPUs */ diff --git a/arch/x86/include/asm/vmware.h b/arch/x86/include/asm/vmware.h new file mode 100644 index 000000000000..02dfea5aebc4 --- /dev/null +++ b/arch/x86/include/asm/vmware.h @@ -0,0 +1,26 @@ +/* + * Copyright (C) 2008, VMware, Inc. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or + * NON INFRINGEMENT. See the GNU General Public License for more + * details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. + * + */ +#ifndef ASM_X86__VMWARE_H +#define ASM_X86__VMWARE_H + +extern unsigned long vmware_get_tsc_khz(void); +extern int vmware_platform(void); + +#endif diff --git a/arch/x86/kernel/cpu/Makefile b/arch/x86/kernel/cpu/Makefile index 82ec6075c057..a5c04e88777e 100644 --- a/arch/x86/kernel/cpu/Makefile +++ b/arch/x86/kernel/cpu/Makefile @@ -4,6 +4,7 @@ obj-y := intel_cacheinfo.o addon_cpuid_features.o obj-y += proc.o capflags.o powerflags.o common.o +obj-y += vmware.o hypervisor.o obj-$(CONFIG_X86_32) += bugs.o cmpxchg.o obj-$(CONFIG_X86_64) += bugs_64.o diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index b9c9ea0217a9..b88595c36254 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -36,6 +36,7 @@ #include #include #include +#include #include "cpu.h" @@ -703,6 +704,7 @@ static void __cpuinit identify_cpu(struct cpuinfo_x86 *c) detect_ht(c); #endif + init_hypervisor(c); /* * On SMP, boot_cpu_data holds the common feature set between * all CPUs; so make sure that we indicate which features are diff --git a/arch/x86/kernel/cpu/hypervisor.c b/arch/x86/kernel/cpu/hypervisor.c new file mode 100644 index 000000000000..7bd55064ffe9 --- /dev/null +++ b/arch/x86/kernel/cpu/hypervisor.c @@ -0,0 +1,48 @@ +/* + * Common hypervisor code + * + * Copyright (C) 2008, VMware, Inc. + * Author : Alok N Kataria + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or + * NON INFRINGEMENT. See the GNU General Public License for more + * details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. + * + */ + +#include +#include + +static inline void __cpuinit +detect_hypervisor_vendor(struct cpuinfo_x86 *c) +{ + if (vmware_platform()) { + c->x86_hyper_vendor = X86_HYPER_VENDOR_VMWARE; + } else { + c->x86_hyper_vendor = X86_HYPER_VENDOR_NONE; + } +} + +unsigned long get_hypervisor_tsc_freq(void) +{ + if (boot_cpu_data.x86_hyper_vendor == X86_HYPER_VENDOR_VMWARE) + return vmware_get_tsc_khz(); + return 0; +} + +void __cpuinit init_hypervisor(struct cpuinfo_x86 *c) +{ + detect_hypervisor_vendor(c); +} + diff --git a/arch/x86/kernel/cpu/vmware.c b/arch/x86/kernel/cpu/vmware.c new file mode 100644 index 000000000000..d5d1b75a4b77 --- /dev/null +++ b/arch/x86/kernel/cpu/vmware.c @@ -0,0 +1,88 @@ +/* + * VMware Detection code. + * + * Copyright (C) 2008, VMware, Inc. + * Author : Alok N Kataria + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or + * NON INFRINGEMENT. See the GNU General Public License for more + * details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. + * + */ + +#include +#include + +#define CPUID_VMWARE_INFO_LEAF 0x40000000 +#define VMWARE_HYPERVISOR_MAGIC 0x564D5868 +#define VMWARE_HYPERVISOR_PORT 0x5658 + +#define VMWARE_PORT_CMD_GETVERSION 10 +#define VMWARE_PORT_CMD_GETHZ 45 + +#define VMWARE_PORT(cmd, eax, ebx, ecx, edx) \ + __asm__("inl (%%dx)" : \ + "=a"(eax), "=c"(ecx), "=d"(edx), "=b"(ebx) : \ + "0"(VMWARE_HYPERVISOR_MAGIC), \ + "1"(VMWARE_PORT_CMD_##cmd), \ + "2"(VMWARE_HYPERVISOR_PORT), "3"(0) : \ + "memory"); + +static inline int __vmware_platform(void) +{ + uint32_t eax, ebx, ecx, edx; + VMWARE_PORT(GETVERSION, eax, ebx, ecx, edx); + return eax != (uint32_t)-1 && ebx == VMWARE_HYPERVISOR_MAGIC; +} + +static unsigned long __vmware_get_tsc_khz(void) +{ + uint64_t tsc_hz; + uint32_t eax, ebx, ecx, edx; + + VMWARE_PORT(GETHZ, eax, ebx, ecx, edx); + + if (eax == (uint32_t)-1) + return 0; + tsc_hz = eax | (((uint64_t)ebx) << 32); + do_div(tsc_hz, 1000); + BUG_ON(tsc_hz >> 32); + return tsc_hz; +} + +int vmware_platform(void) +{ + if (cpu_has_hypervisor) { + unsigned int eax, ebx, ecx, edx; + char hyper_vendor_id[13]; + + cpuid(CPUID_VMWARE_INFO_LEAF, &eax, &ebx, &ecx, &edx); + memcpy(hyper_vendor_id + 0, &ebx, 4); + memcpy(hyper_vendor_id + 4, &ecx, 4); + memcpy(hyper_vendor_id + 8, &edx, 4); + hyper_vendor_id[12] = '\0'; + if (!strcmp(hyper_vendor_id, "VMwareVMware")) + return 1; + } else if (dmi_available && dmi_name_in_vendors("VMware") && + __vmware_platform()) + return 1; + + return 0; +} + +unsigned long vmware_get_tsc_khz(void) +{ + BUG_ON(!vmware_platform()); + return __vmware_get_tsc_khz(); +} diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index 0fa6790c1dd3..f44dadfb32cf 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -98,6 +98,7 @@ #include #include +#include #include #include @@ -909,6 +910,12 @@ void __init setup_arch(char **cmdline_p) dmi_check_system(bad_bios_dmi_table); + /* + * VMware detection requires dmi to be available, so this + * needs to be done after dmi_scan_machine, for the BP. + */ + init_hypervisor(&boot_cpu_data); + #ifdef CONFIG_X86_32 probe_roms(); #endif diff --git a/arch/x86/kernel/tsc.c b/arch/x86/kernel/tsc.c index 62348e4fd8d1..6dbf0bcb44a8 100644 --- a/arch/x86/kernel/tsc.c +++ b/arch/x86/kernel/tsc.c @@ -15,6 +15,7 @@ #include #include #include +#include unsigned int cpu_khz; /* TSC clocks / usec, not used here */ EXPORT_SYMBOL(cpu_khz); @@ -352,9 +353,15 @@ unsigned long native_calibrate_tsc(void) { u64 tsc1, tsc2, delta, ref1, ref2; unsigned long tsc_pit_min = ULONG_MAX, tsc_ref_min = ULONG_MAX; - unsigned long flags, latch, ms, fast_calibrate; + unsigned long flags, latch, ms, fast_calibrate, tsc_khz; int hpet = is_hpet_enabled(), i, loopmin; + tsc_khz = get_hypervisor_tsc_freq(); + if (tsc_khz) { + printk(KERN_INFO "TSC: Frequency read from the hypervisor\n"); + return tsc_khz; + } + local_irq_save(flags); fast_calibrate = quick_pit_calibrate(); local_irq_restore(flags); -- cgit v1.2.3 From eca0cd028bdf0f6aaceb0d023e9c7501079a7dda Mon Sep 17 00:00:00 2001 From: Alok Kataria Date: Fri, 31 Oct 2008 12:01:58 -0700 Subject: x86: Add a synthetic TSC_RELIABLE feature bit. Impact: Changes timebase calibration on Vmware. Use the synthetic TSC_RELIABLE bit to workaround virtualization anomalies. Virtual TSCs can be kept nearly in sync, but because the virtual TSC offset is set by software, it's not perfect. So, the TSC synchronization test can fail. Even then the TSC can be used as a clocksource since the VMware platform exports a reliable TSC to the guest for timekeeping purposes. Use this bit to check if we need to skip the TSC sync checks. Along with this also set the CONSTANT_TSC bit when on VMware, since we still want to use TSC as clocksource on VM running over hardware which has unsynchronized TSC's (opteron's), since the hypervisor will take care of providing consistent TSC to the guest. Signed-off-by: Alok N Kataria Signed-off-by: Dan Hecht Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/vmware.h | 1 + arch/x86/kernel/cpu/hypervisor.c | 11 ++++++++++- arch/x86/kernel/cpu/vmware.c | 18 ++++++++++++++++++ arch/x86/kernel/tsc_sync.c | 8 +++++++- 4 files changed, 36 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/vmware.h b/arch/x86/include/asm/vmware.h index 02dfea5aebc4..c11b7e100d83 100644 --- a/arch/x86/include/asm/vmware.h +++ b/arch/x86/include/asm/vmware.h @@ -22,5 +22,6 @@ extern unsigned long vmware_get_tsc_khz(void); extern int vmware_platform(void); +extern void vmware_set_feature_bits(struct cpuinfo_x86 *c); #endif diff --git a/arch/x86/kernel/cpu/hypervisor.c b/arch/x86/kernel/cpu/hypervisor.c index 7bd55064ffe9..35ae2b75226d 100644 --- a/arch/x86/kernel/cpu/hypervisor.c +++ b/arch/x86/kernel/cpu/hypervisor.c @@ -41,8 +41,17 @@ unsigned long get_hypervisor_tsc_freq(void) return 0; } +static inline void __cpuinit +hypervisor_set_feature_bits(struct cpuinfo_x86 *c) +{ + if (boot_cpu_data.x86_hyper_vendor == X86_HYPER_VENDOR_VMWARE) { + vmware_set_feature_bits(c); + return; + } +} + void __cpuinit init_hypervisor(struct cpuinfo_x86 *c) { detect_hypervisor_vendor(c); + hypervisor_set_feature_bits(c); } - diff --git a/arch/x86/kernel/cpu/vmware.c b/arch/x86/kernel/cpu/vmware.c index d5d1b75a4b77..2ac4394fcb90 100644 --- a/arch/x86/kernel/cpu/vmware.c +++ b/arch/x86/kernel/cpu/vmware.c @@ -86,3 +86,21 @@ unsigned long vmware_get_tsc_khz(void) BUG_ON(!vmware_platform()); return __vmware_get_tsc_khz(); } + +/* + * VMware hypervisor takes care of exporting a reliable TSC to the guest. + * Still, due to timing difference when running on virtual cpus, the TSC can + * be marked as unstable in some cases. For example, the TSC sync check at + * bootup can fail due to a marginal offset between vcpus' TSCs (though the + * TSCs do not drift from each other). Also, the ACPI PM timer clocksource + * is not suitable as a watchdog when running on a hypervisor because the + * kernel may miss a wrap of the counter if the vcpu is descheduled for a + * long time. To skip these checks at runtime we set these capability bits, + * so that the kernel could just trust the hypervisor with providing a + * reliable virtual TSC that is suitable for timekeeping. + */ +void __cpuinit vmware_set_feature_bits(struct cpuinfo_x86 *c) +{ + set_cpu_cap(c, X86_FEATURE_CONSTANT_TSC); + set_cpu_cap(c, X86_FEATURE_TSC_RELIABLE); +} diff --git a/arch/x86/kernel/tsc_sync.c b/arch/x86/kernel/tsc_sync.c index 9ffb01c31c40..5977c40a138f 100644 --- a/arch/x86/kernel/tsc_sync.c +++ b/arch/x86/kernel/tsc_sync.c @@ -108,6 +108,12 @@ void __cpuinit check_tsc_sync_source(int cpu) if (unsynchronized_tsc()) return; + if (boot_cpu_has(X86_FEATURE_TSC_RELIABLE)) { + printk(KERN_INFO + "Skipping synchronization checks as TSC is reliable.\n"); + return; + } + printk(KERN_INFO "checking TSC synchronization [CPU#%d -> CPU#%d]:", smp_processor_id(), cpu); @@ -161,7 +167,7 @@ void __cpuinit check_tsc_sync_target(void) { int cpus = 2; - if (unsynchronized_tsc()) + if (unsynchronized_tsc() || boot_cpu_has(X86_FEATURE_TSC_RELIABLE)) return; /* -- cgit v1.2.3 From 395628ef4ea12ff0748099f145363b5e33c69acb Mon Sep 17 00:00:00 2001 From: Alok Kataria Date: Fri, 24 Oct 2008 17:22:01 -0700 Subject: x86: Skip verification by the watchdog for TSC clocksource. Impact: Changes timekeeping on Vmware (or with tsc=reliable). This is achieved by resetting the CLOCKSOURCE_MUST_VERIFY flag. We add a tsc=reliable commandline option to enable this. This enables legacy hardware without HPET, LAPIC, or ACPI timers to enter high-resolution timer mode. Along with that have extended this to be used in virtualization environement too. Now we also set this flag if the X86_FEATURE_TSC_RELIABLE bit is set. This is important since there is a wrap-around problem with the acpi_pm timer. The acpi_pm counter is just 24bits and this can overflow in ~4 seconds. With the NO_HZ kernels in virtualized environment, there can be situations when the guest is descheduled for longer duration, as a result we may miss the wrap of the acpi counter. When TSC is used as a clocksource and acpi_pm timer is being used as the watchdog clocksource this error in acpi_pm results in TSC being marked as unstable, and essentially results in time dropping in chunks of 4 seconds whenever this wrap is missed. Since the virtualized TSC is reliable on VMware, we should always use the TSCs clocksource on VMware, so we skip the verfication at runtime, by checking for the feature bit. Since we reset the flag for mgeode systems too, i have combined the mgeode case with the feature bit check. Signed-off-by: Jeff Hansen Signed-off-by: Alok N Kataria Signed-off-by: Dan Hecht Signed-off-by: H. Peter Anvin --- Documentation/kernel-parameters.txt | 7 +++++++ arch/x86/kernel/tsc.c | 33 +++++++++++++++++++++------------ 2 files changed, 28 insertions(+), 12 deletions(-) (limited to 'arch/x86') diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index 1bbcaa8982b6..dc6b06f67fca 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -2267,6 +2267,13 @@ and is between 256 and 4096 characters. It is defined in the file Format: ,,,,,,,, + tsc= Disable clocksource-must-verify flag for TSC. + Format: + [x86] reliable: mark tsc clocksource as reliable, this + disables clocksource verification at runtime. + Used to enable high-resolution timer mode on older + hardware, and in virtualized environment. + turbografx.map[2|3]= [HW,JOY] TurboGraFX parallel port interface Format: diff --git a/arch/x86/kernel/tsc.c b/arch/x86/kernel/tsc.c index 6dbf0bcb44a8..ee01cd96b5e1 100644 --- a/arch/x86/kernel/tsc.c +++ b/arch/x86/kernel/tsc.c @@ -32,6 +32,7 @@ static int tsc_unstable; erroneous rdtsc usage on !cpu_has_tsc processors */ static int tsc_disabled = -1; +static int tsc_clocksource_reliable; /* * Scheduler clock - returns current time in nanosec units. */ @@ -99,6 +100,15 @@ int __init notsc_setup(char *str) __setup("notsc", notsc_setup); +static int __init tsc_setup(char *str) +{ + if (!strcmp(str, "reliable")) + tsc_clocksource_reliable = 1; + return 1; +} + +__setup("tsc=", tsc_setup); + #define MAX_RETRIES 5 #define SMI_TRESHOLD 50000 @@ -738,24 +748,21 @@ static struct dmi_system_id __initdata bad_tsc_dmi_table[] = { {} }; -/* - * Geode_LX - the OLPC CPU has a possibly a very reliable TSC - */ +static void __init check_system_tsc_reliable(void) +{ #ifdef CONFIG_MGEODE_LX -/* RTSC counts during suspend */ + /* RTSC counts during suspend */ #define RTSC_SUSP 0x100 - -static void __init check_geode_tsc_reliable(void) -{ unsigned long res_low, res_high; rdmsr_safe(MSR_GEODE_BUSCONT_CONF0, &res_low, &res_high); + /* Geode_LX - the OLPC CPU has a possibly a very reliable TSC */ if (res_low & RTSC_SUSP) - clocksource_tsc.flags &= ~CLOCK_SOURCE_MUST_VERIFY; -} -#else -static inline void check_geode_tsc_reliable(void) { } + tsc_clocksource_reliable = 1; #endif + if (boot_cpu_has(X86_FEATURE_TSC_RELIABLE)) + tsc_clocksource_reliable = 1; +} /* * Make an educated guess if the TSC is trustworthy and synchronized @@ -790,6 +797,8 @@ static void __init init_tsc_clocksource(void) { clocksource_tsc.mult = clocksource_khz2mult(tsc_khz, clocksource_tsc.shift); + if (tsc_clocksource_reliable) + clocksource_tsc.flags &= ~CLOCK_SOURCE_MUST_VERIFY; /* lower the rating if we already know its unstable: */ if (check_tsc_unstable()) { clocksource_tsc.rating = 0; @@ -850,7 +859,7 @@ void __init tsc_init(void) if (unsynchronized_tsc()) mark_tsc_unstable("TSCs unsynchronized"); - check_geode_tsc_reliable(); + check_system_tsc_reliable(); init_tsc_clocksource(); } -- cgit v1.2.3 From 3555105333ae55414d0fe051557bd7dc590f5255 Mon Sep 17 00:00:00 2001 From: Gary Hade Date: Fri, 31 Oct 2008 10:52:03 -0700 Subject: x86: add memory hotremove config option Impact: enable CONFIG_MEMORY_HOTREMOVE feature on x86. (default-off) Memory hotremove functionality can currently be configured into the ia64, powerpc, and s390 kernels. This patch makes it possible to configure the memory hotremove functionality into the x86 kernel as well. Signed-off-by: Badari Pulavarty Signed-off-by: Gary Hade Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index c00aefcb47d5..25e711526116 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -1492,6 +1492,10 @@ config ARCH_ENABLE_MEMORY_HOTPLUG def_bool y depends on X86_64 || (X86_32 && HIGHMEM) +config ARCH_ENABLE_MEMORY_HOTREMOVE + def_bool y + depends on MEMORY_HOTPLUG + config HAVE_ARCH_EARLY_PFN_TO_NID def_bool X86_64 depends on NUMA -- cgit v1.2.3 From 7e5e26a3d8ac4bcadb380073dc9604c07a9a6198 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 31 Oct 2008 09:36:38 -0400 Subject: ftrace: fix hardirq header for non ftrace archs Impact: build fix for non-ftrace architectures Not all archs implement ftrace, and therefore do not have an asm/ftrace.h. This patch corrects the problem. The ftrace_nmi_enter/exit now must be defined for all archs that implement dynamic ftrace. Currently, only x86 does. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/arm/include/asm/ftrace.h | 5 ----- arch/powerpc/include/asm/ftrace.h | 5 ----- arch/sh/include/asm/ftrace.h | 5 ----- arch/sparc/include/asm/ftrace.h | 5 ----- arch/x86/include/asm/ftrace.h | 16 ---------------- include/linux/ftrace.h | 5 ++++- include/linux/hardirq.h | 2 +- 7 files changed, 5 insertions(+), 38 deletions(-) (limited to 'arch/x86') diff --git a/arch/arm/include/asm/ftrace.h b/arch/arm/include/asm/ftrace.h index 3f3a1d1508ea..39c8bc1a006a 100644 --- a/arch/arm/include/asm/ftrace.h +++ b/arch/arm/include/asm/ftrace.h @@ -1,11 +1,6 @@ #ifndef _ASM_ARM_FTRACE #define _ASM_ARM_FTRACE -#ifndef __ASSEMBLY__ -static inline void ftrace_nmi_enter(void) { } -static inline void ftrace_nmi_exit(void) { } -#endif - #ifdef CONFIG_FUNCTION_TRACER #define MCOUNT_ADDR ((long)(mcount)) #define MCOUNT_INSN_SIZE 4 /* sizeof mcount call */ diff --git a/arch/powerpc/include/asm/ftrace.h b/arch/powerpc/include/asm/ftrace.h index 1cd72700fbc0..b298f7a631e6 100644 --- a/arch/powerpc/include/asm/ftrace.h +++ b/arch/powerpc/include/asm/ftrace.h @@ -1,11 +1,6 @@ #ifndef _ASM_POWERPC_FTRACE #define _ASM_POWERPC_FTRACE -#ifndef __ASSEMBLY__ -static inline void ftrace_nmi_enter(void) { } -static inline void ftrace_nmi_exit(void) { } -#endif - #ifdef CONFIG_FUNCTION_TRACER #define MCOUNT_ADDR ((long)(_mcount)) #define MCOUNT_INSN_SIZE 4 /* sizeof mcount call */ diff --git a/arch/sh/include/asm/ftrace.h b/arch/sh/include/asm/ftrace.h index 31ada0370cb6..3aed362c9463 100644 --- a/arch/sh/include/asm/ftrace.h +++ b/arch/sh/include/asm/ftrace.h @@ -1,11 +1,6 @@ #ifndef __ASM_SH_FTRACE_H #define __ASM_SH_FTRACE_H -#ifndef __ASSEMBLY__ -static inline void ftrace_nmi_enter(void) { } -static inline void ftrace_nmi_exit(void) { } -#endif - #ifndef __ASSEMBLY__ extern void mcount(void); #endif diff --git a/arch/sparc/include/asm/ftrace.h b/arch/sparc/include/asm/ftrace.h index 62055ac0496e..d27716cd38c1 100644 --- a/arch/sparc/include/asm/ftrace.h +++ b/arch/sparc/include/asm/ftrace.h @@ -1,11 +1,6 @@ #ifndef _ASM_SPARC64_FTRACE #define _ASM_SPARC64_FTRACE -#ifndef __ASSEMBLY__ -static inline void ftrace_nmi_enter(void) { } -static inline void ftrace_nmi_exit(void) { } -#endif - #ifdef CONFIG_MCOUNT #define MCOUNT_ADDR ((long)(_mcount)) #define MCOUNT_INSN_SIZE 4 /* sizeof mcount call */ diff --git a/arch/x86/include/asm/ftrace.h b/arch/x86/include/asm/ftrace.h index a23468194b8c..f8173ed1c970 100644 --- a/arch/x86/include/asm/ftrace.h +++ b/arch/x86/include/asm/ftrace.h @@ -17,23 +17,7 @@ static inline unsigned long ftrace_call_adjust(unsigned long addr) */ return addr - 1; } - -#ifdef CONFIG_DYNAMIC_FTRACE -extern void ftrace_nmi_enter(void); -extern void ftrace_nmi_exit(void); -#else -static inline void ftrace_nmi_enter(void) { } -static inline void ftrace_nmi_exit(void) { } -#endif #endif /* __ASSEMBLY__ */ - -#else /* CONFIG_FUNCTION_TRACER */ - -#ifndef __ASSEMBLY__ -static inline void ftrace_nmi_enter(void) { } -static inline void ftrace_nmi_exit(void) { } -#endif - #endif /* CONFIG_FUNCTION_TRACER */ #endif /* _ASM_X86_FTRACE_H */ diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index e46a7b34037c..0ad1b48aea69 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -44,7 +44,6 @@ static inline void ftrace_kill(void) { } #endif /* CONFIG_FUNCTION_TRACER */ #ifdef CONFIG_DYNAMIC_FTRACE - enum { FTRACE_FL_FREE = (1 << 0), FTRACE_FL_FAILED = (1 << 1), @@ -105,6 +104,8 @@ extern void ftrace_release(void *start, unsigned long size); extern void ftrace_disable_daemon(void); extern void ftrace_enable_daemon(void); +extern void ftrace_nmi_enter(void); +extern void ftrace_nmi_exit(void); #else # define skip_trace(ip) ({ 0; }) @@ -113,6 +114,8 @@ extern void ftrace_enable_daemon(void); # define ftrace_disable_daemon() do { } while (0) # define ftrace_enable_daemon() do { } while (0) static inline void ftrace_release(void *start, unsigned long size) { } +static inline void ftrace_nmi_enter(void) { } +static inline void ftrace_nmi_exit(void) { } #endif /* CONFIG_DYNAMIC_FTRACE */ /* totally disable ftrace - can not re-enable after this */ diff --git a/include/linux/hardirq.h b/include/linux/hardirq.h index 0087cb43becf..ffc16ab5a878 100644 --- a/include/linux/hardirq.h +++ b/include/linux/hardirq.h @@ -4,8 +4,8 @@ #include #include #include +#include #include -#include #include /* -- cgit v1.2.3 From 6bdbfe99916398dbb28d83833cc04757110f2738 Mon Sep 17 00:00:00 2001 From: Alok Kataria Date: Mon, 3 Nov 2008 11:31:28 -0800 Subject: x86: VMware: Fix vmware_get_tsc code Impact: Fix possible failure to calibrate the TSC on Vmware near 4 GHz The current version of the code to get the tsc frequency from the VMware hypervisor, will be broken on processor with frequency (4G-1) HZ, because on such processors eax will have UINT_MAX and that would be legitimate. We instead check that EBX did change to decide if we were able to read the frequency from the hypervisor. Signed-off-by: Alok N Kataria Signed-off-by: H. Peter Anvin --- arch/x86/kernel/cpu/vmware.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/vmware.c b/arch/x86/kernel/cpu/vmware.c index 2ac4394fcb90..a0905ecfe7d2 100644 --- a/arch/x86/kernel/cpu/vmware.c +++ b/arch/x86/kernel/cpu/vmware.c @@ -36,7 +36,7 @@ "=a"(eax), "=c"(ecx), "=d"(edx), "=b"(ebx) : \ "0"(VMWARE_HYPERVISOR_MAGIC), \ "1"(VMWARE_PORT_CMD_##cmd), \ - "2"(VMWARE_HYPERVISOR_PORT), "3"(0) : \ + "2"(VMWARE_HYPERVISOR_PORT), "3"(UINT_MAX) : \ "memory"); static inline int __vmware_platform(void) @@ -53,7 +53,7 @@ static unsigned long __vmware_get_tsc_khz(void) VMWARE_PORT(GETHZ, eax, ebx, ecx, edx); - if (eax == (uint32_t)-1) + if (ebx == UINT_MAX) return 0; tsc_hz = eax | (((uint64_t)ebx) << 32); do_div(tsc_hz, 1000); -- cgit v1.2.3 From 124ffe1456d6efea5b32cc6d36e3fa434cdc84d9 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Mon, 3 Nov 2008 19:23:01 -0800 Subject: x86: signal_64: remove unused code in __setup_rt_frame() Impact: cleanup sizeof(*set) is always 8 on x86_64. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/kernel/signal_64.c | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/signal_64.c b/arch/x86/kernel/signal_64.c index 83990db82f74..cfbb60a5f9d2 100644 --- a/arch/x86/kernel/signal_64.c +++ b/arch/x86/kernel/signal_64.c @@ -251,11 +251,7 @@ static int __setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, err |= __put_user(me->sas_ss_size, &frame->uc.uc_stack.ss_size); err |= setup_sigcontext(&frame->uc.uc_mcontext, regs, set->sig[0], me); err |= __put_user(fp, &frame->uc.uc_mcontext.fpstate); - if (sizeof(*set) == 16) { - __put_user(set->sig[0], &frame->uc.uc_sigmask.sig[0]); - __put_user(set->sig[1], &frame->uc.uc_sigmask.sig[1]); - } else - err |= __copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set)); + err |= __copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set)); /* Set up to return from userspace. If provided, use a stub already in userspace. */ -- cgit v1.2.3 From 6cf87efbc7a3676e0ad7c9622ec6aec244a593bc Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Tue, 4 Nov 2008 10:42:23 +0100 Subject: x86 debug: mark early_printk.o as notrace Impact: do not do function-tracing in the early-printk code this is useful when earlyprintk=vga,keep is used to debug tracer plugins. Signed-off-by: Ingo Molnar --- arch/x86/kernel/Makefile | 1 + 1 file changed, 1 insertion(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index e489ff9cb3e2..943fe6026c64 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -12,6 +12,7 @@ CFLAGS_REMOVE_tsc.o = -pg CFLAGS_REMOVE_rtc.o = -pg CFLAGS_REMOVE_paravirt-spinlocks.o = -pg CFLAGS_REMOVE_ftrace.o = -pg +CFLAGS_REMOVE_early_printk.o = -pg endif # -- cgit v1.2.3 From fd8cd7e1919fc1c27fe2fdccd2a1cd32f791ef0f Mon Sep 17 00:00:00 2001 From: Alok Kataria Date: Mon, 3 Nov 2008 15:50:38 -0800 Subject: x86: vmware: look for DMI string in the product serial key Impact: Should permit VMware detection on older platforms where the vendor is changed. Could theoretically cause a regression if some weird serial number scheme contains the string "VMware" by pure chance. Seems unlikely, especially with the mixed case. In some user configured cases, VMware may choose not to put a VMware specific DMI string, but the product serial key is always there and is VMware specific. Add a interface to check the serial key, when checking for VMware in the DMI information. Signed-off-by: Alok N Kataria Signed-off-by: H. Peter Anvin --- arch/x86/kernel/cpu/vmware.c | 7 ++++++- drivers/firmware/dmi_scan.c | 11 +++++++++++ include/linux/dmi.h | 2 ++ 3 files changed, 19 insertions(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/vmware.c b/arch/x86/kernel/cpu/vmware.c index a0905ecfe7d2..c034bda842d9 100644 --- a/arch/x86/kernel/cpu/vmware.c +++ b/arch/x86/kernel/cpu/vmware.c @@ -61,6 +61,11 @@ static unsigned long __vmware_get_tsc_khz(void) return tsc_hz; } +/* + * While checking the dmi string infomation, just checking the product + * serial key should be enough, as this will always have a VMware + * specific string when running under VMware hypervisor. + */ int vmware_platform(void) { if (cpu_has_hypervisor) { @@ -74,7 +79,7 @@ int vmware_platform(void) hyper_vendor_id[12] = '\0'; if (!strcmp(hyper_vendor_id, "VMwareVMware")) return 1; - } else if (dmi_available && dmi_name_in_vendors("VMware") && + } else if (dmi_available && dmi_name_in_serial("VMware") && __vmware_platform()) return 1; diff --git a/drivers/firmware/dmi_scan.c b/drivers/firmware/dmi_scan.c index 3e526b6d00cb..d66d41282907 100644 --- a/drivers/firmware/dmi_scan.c +++ b/drivers/firmware/dmi_scan.c @@ -467,6 +467,17 @@ const char *dmi_get_system_info(int field) } EXPORT_SYMBOL(dmi_get_system_info); +/** + * dmi_name_in_serial - Check if string is in the DMI product serial + * information. + */ +int dmi_name_in_serial(const char *str) +{ + int f = DMI_PRODUCT_SERIAL; + if (dmi_ident[f] && strstr(dmi_ident[f], str)) + return 1; + return 0; +} /** * dmi_name_in_vendors - Check if string is anywhere in the DMI vendor information. diff --git a/include/linux/dmi.h b/include/linux/dmi.h index e5084eb5943a..2bfda178f274 100644 --- a/include/linux/dmi.h +++ b/include/linux/dmi.h @@ -44,6 +44,7 @@ extern const struct dmi_device * dmi_find_device(int type, const char *name, extern void dmi_scan_machine(void); extern int dmi_get_year(int field); extern int dmi_name_in_vendors(const char *str); +extern int dmi_name_in_serial(const char *str); extern int dmi_available; extern int dmi_walk(void (*decode)(const struct dmi_header *)); @@ -56,6 +57,7 @@ static inline const struct dmi_device * dmi_find_device(int type, const char *na static inline void dmi_scan_machine(void) { return; } static inline int dmi_get_year(int year) { return 0; } static inline int dmi_name_in_vendors(const char *s) { return 0; } +static inline int dmi_name_in_serial(const char *s) { return 0; } #define dmi_available 0 static inline int dmi_walk(void (*decode)(const struct dmi_header *)) { return -1; } -- cgit v1.2.3 From 838e8bb71dc0c892bf8f84abd3c709d8fe3a8d3c Mon Sep 17 00:00:00 2001 From: Uros Bizjak Date: Fri, 24 Oct 2008 16:53:33 +0200 Subject: x86: Implement change_bit with immediate operand as "lock xorb" Impact: Minor optimization. Implement change_bit with immediate bit count as "lock xorb". This is similar to "lock orb" and "lock andb" for set_bit and clear_bit functions. Signed-off-by: Uros Bizjak Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/bitops.h | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/bitops.h b/arch/x86/include/asm/bitops.h index 360010322711..9fa9dcdf344b 100644 --- a/arch/x86/include/asm/bitops.h +++ b/arch/x86/include/asm/bitops.h @@ -168,7 +168,15 @@ static inline void __change_bit(int nr, volatile unsigned long *addr) */ static inline void change_bit(int nr, volatile unsigned long *addr) { - asm volatile(LOCK_PREFIX "btc %1,%0" : ADDR : "Ir" (nr)); + if (IS_IMMEDIATE(nr)) { + asm volatile(LOCK_PREFIX "xorb %1,%0" + : CONST_MASK_ADDR(nr, addr) + : "iq" ((u8)CONST_MASK(nr))); + } else { + asm volatile(LOCK_PREFIX "btc %1,%0" + : BITOP_ADDR(addr) + : "Ir" (nr)); + } } /** -- cgit v1.2.3 From 64ccf2f9a70a06ba56cd8cedfa610b4e77181587 Mon Sep 17 00:00:00 2001 From: Russ Anderson Date: Wed, 5 Nov 2008 22:11:56 -0600 Subject: x86: uv: Add UV watchlist bios call Add UV bios calls to allocate and free watchlists. Signed-off-by: Russ Anderson Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/uv/bios.h | 17 ++++++++++++++++- arch/x86/kernel/bios_uv.c | 33 +++++++++++++++++++++++++++++++++ 2 files changed, 49 insertions(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/uv/bios.h b/arch/x86/include/asm/uv/bios.h index 51cadc645e6f..58105c5b0b4e 100644 --- a/arch/x86/include/asm/uv/bios.h +++ b/arch/x86/include/asm/uv/bios.h @@ -32,7 +32,9 @@ enum uv_bios_cmd { UV_BIOS_COMMON, UV_BIOS_GET_SN_INFO, - UV_BIOS_FREQ_BASE + UV_BIOS_FREQ_BASE, + UV_BIOS_WATCHLIST_ALLOC, + UV_BIOS_WATCHLIST_FREE }; /* @@ -71,6 +73,15 @@ union partition_info_u { }; }; +union uv_watchlist_u { + u64 val; + struct { + u64 blade : 16, + size : 32, + filler : 16; + }; +}; + /* * bios calls have 6 parameters */ @@ -80,9 +91,13 @@ extern s64 uv_bios_call_reentrant(enum uv_bios_cmd, u64, u64, u64, u64, u64); extern s64 uv_bios_get_sn_info(int, int *, long *, long *, long *); extern s64 uv_bios_freq_base(u64, u64 *); +extern int uv_bios_mq_watchlist_alloc(int, void *, unsigned int, + unsigned long *); +extern int uv_bios_mq_watchlist_free(int, int); extern void uv_bios_init(void); +extern unsigned long sn_rtc_cycles_per_second; extern int uv_type; extern long sn_partition_id; extern long sn_coherency_id; diff --git a/arch/x86/kernel/bios_uv.c b/arch/x86/kernel/bios_uv.c index 7cefb7170e75..4c02b2799216 100644 --- a/arch/x86/kernel/bios_uv.c +++ b/arch/x86/kernel/bios_uv.c @@ -100,6 +100,39 @@ s64 uv_bios_get_sn_info(int fc, int *uvtype, long *partid, long *coher, return ret; } +int +uv_bios_mq_watchlist_alloc(int blade, void *mq, unsigned int mq_size, + unsigned long *intr_mmr_offset) +{ + union uv_watchlist_u size_blade; + unsigned long addr; + u64 watchlist; + s64 ret; + + addr = (unsigned long)mq; + size_blade.size = mq_size; + size_blade.blade = blade; + + /* + * bios returns watchlist number or negative error number. + */ + ret = (int)uv_bios_call_irqsave(UV_BIOS_WATCHLIST_ALLOC, addr, + size_blade.val, (u64)intr_mmr_offset, + (u64)&watchlist, 0); + if (ret < BIOS_STATUS_SUCCESS) + return ret; + + return watchlist; +} +EXPORT_SYMBOL_GPL(uv_bios_mq_watchlist_alloc); + +int +uv_bios_mq_watchlist_free(int blade, int watchlist_num) +{ + return (int)uv_bios_call_irqsave(UV_BIOS_WATCHLIST_FREE, + blade, watchlist_num, 0, 0, 0); +} +EXPORT_SYMBOL_GPL(uv_bios_mq_watchlist_free); s64 uv_bios_freq_base(u64 clock_type, u64 *ticks_per_second) { -- cgit v1.2.3 From e8929c8a6acbecbd629b8e3f2d1a2546ec4ebdfc Mon Sep 17 00:00:00 2001 From: Russ Anderson Date: Wed, 5 Nov 2008 22:13:44 -0600 Subject: x86: uv: Add UV memory protection bios call Add UV bios call to change memory protections. Signed-off-by: Russ Anderson Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/uv/bios.h | 10 +++++++++- arch/x86/kernel/bios_uv.c | 8 ++++++++ 2 files changed, 17 insertions(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/uv/bios.h b/arch/x86/include/asm/uv/bios.h index 58105c5b0b4e..a301a56d4157 100644 --- a/arch/x86/include/asm/uv/bios.h +++ b/arch/x86/include/asm/uv/bios.h @@ -34,7 +34,8 @@ enum uv_bios_cmd { UV_BIOS_GET_SN_INFO, UV_BIOS_FREQ_BASE, UV_BIOS_WATCHLIST_ALLOC, - UV_BIOS_WATCHLIST_FREE + UV_BIOS_WATCHLIST_FREE, + UV_BIOS_MEMPROTECT }; /* @@ -82,6 +83,12 @@ union uv_watchlist_u { }; }; +enum uv_memprotect { + UV_MEMPROT_RESTRICT_ACCESS, + UV_MEMPROT_ALLOW_AMO, + UV_MEMPROT_ALLOW_RW +}; + /* * bios calls have 6 parameters */ @@ -94,6 +101,7 @@ extern s64 uv_bios_freq_base(u64, u64 *); extern int uv_bios_mq_watchlist_alloc(int, void *, unsigned int, unsigned long *); extern int uv_bios_mq_watchlist_free(int, int); +extern s64 uv_bios_change_memprotect(u64, u64, enum uv_memprotect); extern void uv_bios_init(void); diff --git a/arch/x86/kernel/bios_uv.c b/arch/x86/kernel/bios_uv.c index 4c02b2799216..7cf6fc3d1c10 100644 --- a/arch/x86/kernel/bios_uv.c +++ b/arch/x86/kernel/bios_uv.c @@ -134,6 +134,14 @@ uv_bios_mq_watchlist_free(int blade, int watchlist_num) } EXPORT_SYMBOL_GPL(uv_bios_mq_watchlist_free); +s64 +uv_bios_change_memprotect(u64 paddr, u64 len, enum uv_memprotect perms) +{ + return uv_bios_call_irqsave(UV_BIOS_MEMPROTECT, paddr, len, + perms, 0, 0); +} +EXPORT_SYMBOL_GPL(uv_bios_change_memprotect); + s64 uv_bios_freq_base(u64 clock_type, u64 *ticks_per_second) { return uv_bios_call(UV_BIOS_FREQ_BASE, clock_type, -- cgit v1.2.3 From 23c357003b3671cdfb17bc4d5383589e74b71511 Mon Sep 17 00:00:00 2001 From: Russ Anderson Date: Wed, 5 Nov 2008 22:15:13 -0600 Subject: x86: uv: Add UV reserved page bios call Add UV bios call to get the address of the reserved page. Signed-off-by: Russ Anderson Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/uv/bios.h | 5 ++++- arch/x86/kernel/bios_uv.c | 11 +++++++++++ 2 files changed, 15 insertions(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/uv/bios.h b/arch/x86/include/asm/uv/bios.h index a301a56d4157..da1c4e8e78fc 100644 --- a/arch/x86/include/asm/uv/bios.h +++ b/arch/x86/include/asm/uv/bios.h @@ -35,13 +35,15 @@ enum uv_bios_cmd { UV_BIOS_FREQ_BASE, UV_BIOS_WATCHLIST_ALLOC, UV_BIOS_WATCHLIST_FREE, - UV_BIOS_MEMPROTECT + UV_BIOS_MEMPROTECT, + UV_BIOS_GET_PARTITION_ADDR }; /* * Status values returned from a BIOS call. */ enum { + BIOS_STATUS_MORE_PASSES = 1, BIOS_STATUS_SUCCESS = 0, BIOS_STATUS_UNIMPLEMENTED = -ENOSYS, BIOS_STATUS_EINVAL = -EINVAL, @@ -102,6 +104,7 @@ extern int uv_bios_mq_watchlist_alloc(int, void *, unsigned int, unsigned long *); extern int uv_bios_mq_watchlist_free(int, int); extern s64 uv_bios_change_memprotect(u64, u64, enum uv_memprotect); +extern s64 uv_bios_reserved_page_pa(u64, u64 *, u64 *, u64 *); extern void uv_bios_init(void); diff --git a/arch/x86/kernel/bios_uv.c b/arch/x86/kernel/bios_uv.c index 7cf6fc3d1c10..d22d0f1bbea0 100644 --- a/arch/x86/kernel/bios_uv.c +++ b/arch/x86/kernel/bios_uv.c @@ -142,6 +142,17 @@ uv_bios_change_memprotect(u64 paddr, u64 len, enum uv_memprotect perms) } EXPORT_SYMBOL_GPL(uv_bios_change_memprotect); +s64 +uv_bios_reserved_page_pa(u64 buf, u64 *cookie, u64 *addr, u64 *len) +{ + s64 ret; + + ret = uv_bios_call_irqsave(UV_BIOS_GET_PARTITION_ADDR, (u64)cookie, + (u64)addr, buf, (u64)len, 0); + return ret; +} +EXPORT_SYMBOL_GPL(uv_bios_reserved_page_pa); + s64 uv_bios_freq_base(u64 clock_type, u64 *ticks_per_second) { return uv_bios_call(UV_BIOS_FREQ_BASE, clock_type, -- cgit v1.2.3 From 60a7ecf42661f2b22168751298592da6ee210c9e Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 5 Nov 2008 16:05:44 -0500 Subject: ftrace: add quick function trace stop Impact: quick start and stop of function tracer This patch adds a way to disable the function tracer quickly without the need to run kstop_machine. It adds a new variable called function_trace_stop which will stop the calls to functions from mcount when set. This is just an on/off switch and does not handle recursion like preempt_disable(). It's main purpose is to help other tracers/debuggers start and stop tracing fuctions without the need to call kstop_machine. The config option HAVE_FUNCTION_TRACE_MCOUNT_TEST is added for archs that implement the testing of the function_trace_stop in the mcount arch dependent code. Otherwise, the test is done in the C code. x86 is the only arch at the moment that supports this. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 1 + arch/x86/kernel/entry_32.S | 6 ++++++ arch/x86/kernel/entry_64.S | 5 +++++ include/linux/ftrace.h | 30 +++++++++++++++++++++++++++++ kernel/trace/Kconfig | 7 +++++++ kernel/trace/ftrace.c | 47 ++++++++++++++++++++++++++++++++++++---------- 6 files changed, 86 insertions(+), 10 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 6f20718d3156..d09e812c6223 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -29,6 +29,7 @@ config X86 select HAVE_FTRACE_MCOUNT_RECORD select HAVE_DYNAMIC_FTRACE select HAVE_FUNCTION_TRACER + select HAVE_FUNCTION_TRACE_MCOUNT_TEST select HAVE_KVM if ((X86_32 && !X86_VOYAGER && !X86_VISWS && !X86_NUMAQ) || X86_64) select HAVE_ARCH_KGDB if !X86_VOYAGER select HAVE_ARCH_TRACEHOOK diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S index 28b597ef9ca1..9134de814c97 100644 --- a/arch/x86/kernel/entry_32.S +++ b/arch/x86/kernel/entry_32.S @@ -1157,6 +1157,9 @@ ENTRY(mcount) END(mcount) ENTRY(ftrace_caller) + cmpl $0, function_trace_stop + jne ftrace_stub + pushl %eax pushl %ecx pushl %edx @@ -1180,6 +1183,9 @@ END(ftrace_caller) #else /* ! CONFIG_DYNAMIC_FTRACE */ ENTRY(mcount) + cmpl $0, function_trace_stop + jne ftrace_stub + cmpl $ftrace_stub, ftrace_trace_function jnz trace .globl ftrace_stub diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index b86f332c96a6..08aa6b10933c 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -68,6 +68,8 @@ ENTRY(mcount) END(mcount) ENTRY(ftrace_caller) + cmpl $0, function_trace_stop + jne ftrace_stub /* taken from glibc */ subq $0x38, %rsp @@ -103,6 +105,9 @@ END(ftrace_caller) #else /* ! CONFIG_DYNAMIC_FTRACE */ ENTRY(mcount) + cmpl $0, function_trace_stop + jne ftrace_stub + cmpq $ftrace_stub, ftrace_trace_function jnz trace .globl ftrace_stub diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 4642959e5bda..794ab907dbfe 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -23,6 +23,34 @@ struct ftrace_ops { struct ftrace_ops *next; }; +extern int function_trace_stop; + +/** + * ftrace_stop - stop function tracer. + * + * A quick way to stop the function tracer. Note this an on off switch, + * it is not something that is recursive like preempt_disable. + * This does not disable the calling of mcount, it only stops the + * calling of functions from mcount. + */ +static inline void ftrace_stop(void) +{ + function_trace_stop = 1; +} + +/** + * ftrace_start - start the function tracer. + * + * This function is the inverse of ftrace_stop. This does not enable + * the function tracing if the function tracer is disabled. This only + * sets the function tracer flag to continue calling the functions + * from mcount. + */ +static inline void ftrace_start(void) +{ + function_trace_stop = 0; +} + /* * The ftrace_ops must be a static and should also * be read_mostly. These functions do modify read_mostly variables @@ -41,6 +69,8 @@ extern void ftrace_stub(unsigned long a0, unsigned long a1); # define unregister_ftrace_function(ops) do { } while (0) # define clear_ftrace_function(ops) do { } while (0) static inline void ftrace_kill(void) { } +static inline void ftrace_stop(void) { } +static inline void ftrace_start(void) { } #endif /* CONFIG_FUNCTION_TRACER */ #ifdef CONFIG_DYNAMIC_FTRACE diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 33dbefd471e8..fc4febc3334a 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -9,6 +9,13 @@ config NOP_TRACER config HAVE_FUNCTION_TRACER bool +config HAVE_FUNCTION_TRACE_MCOUNT_TEST + bool + help + This gets selected when the arch tests the function_trace_stop + variable at the mcount call site. Otherwise, this variable + is tested by the called function. + config HAVE_DYNAMIC_FTRACE bool diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 4a39d24568c8..896c71f0f4c4 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -47,6 +47,9 @@ int ftrace_enabled __read_mostly; static int last_ftrace_enabled; +/* Quick disabling of function tracer. */ +int function_trace_stop; + /* * ftrace_disabled is set when an anomaly is discovered. * ftrace_disabled is much stronger than ftrace_enabled. @@ -63,6 +66,7 @@ static struct ftrace_ops ftrace_list_end __read_mostly = static struct ftrace_ops *ftrace_list __read_mostly = &ftrace_list_end; ftrace_func_t ftrace_trace_function __read_mostly = ftrace_stub; +ftrace_func_t __ftrace_trace_function __read_mostly = ftrace_stub; static void ftrace_list_func(unsigned long ip, unsigned long parent_ip) { @@ -88,8 +92,23 @@ static void ftrace_list_func(unsigned long ip, unsigned long parent_ip) void clear_ftrace_function(void) { ftrace_trace_function = ftrace_stub; + __ftrace_trace_function = ftrace_stub; } +#ifndef CONFIG_HAVE_FUNCTION_TRACE_MCOUNT_TEST +/* + * For those archs that do not test ftrace_trace_stop in their + * mcount call site, we need to do it from C. + */ +static void ftrace_test_stop_func(unsigned long ip, unsigned long parent_ip) +{ + if (function_trace_stop) + return; + + __ftrace_trace_function(ip, parent_ip); +} +#endif + static int __register_ftrace_function(struct ftrace_ops *ops) { /* should not be called from interrupt context */ @@ -110,10 +129,18 @@ static int __register_ftrace_function(struct ftrace_ops *ops) * For one func, simply call it directly. * For more than one func, call the chain. */ +#ifdef CONFIG_HAVE_FUNCTION_TRACE_MCOUNT_TEST if (ops->next == &ftrace_list_end) ftrace_trace_function = ops->func; else ftrace_trace_function = ftrace_list_func; +#else + if (ops->next == &ftrace_list_end) + __ftrace_trace_function = ops->func; + else + __ftrace_trace_function = ftrace_list_func; + ftrace_trace_function = ftrace_test_stop_func; +#endif } spin_unlock(&ftrace_lock); @@ -526,7 +553,7 @@ static void ftrace_run_update_code(int command) } static ftrace_func_t saved_ftrace_func; -static int ftrace_start; +static int ftrace_start_up; static DEFINE_MUTEX(ftrace_start_lock); static void ftrace_startup(void) @@ -537,8 +564,8 @@ static void ftrace_startup(void) return; mutex_lock(&ftrace_start_lock); - ftrace_start++; - if (ftrace_start == 1) + ftrace_start_up++; + if (ftrace_start_up == 1) command |= FTRACE_ENABLE_CALLS; if (saved_ftrace_func != ftrace_trace_function) { @@ -562,8 +589,8 @@ static void ftrace_shutdown(void) return; mutex_lock(&ftrace_start_lock); - ftrace_start--; - if (!ftrace_start) + ftrace_start_up--; + if (!ftrace_start_up) command |= FTRACE_DISABLE_CALLS; if (saved_ftrace_func != ftrace_trace_function) { @@ -589,8 +616,8 @@ static void ftrace_startup_sysctl(void) mutex_lock(&ftrace_start_lock); /* Force update next time */ saved_ftrace_func = NULL; - /* ftrace_start is true if we want ftrace running */ - if (ftrace_start) + /* ftrace_start_up is true if we want ftrace running */ + if (ftrace_start_up) command |= FTRACE_ENABLE_CALLS; ftrace_run_update_code(command); @@ -605,8 +632,8 @@ static void ftrace_shutdown_sysctl(void) return; mutex_lock(&ftrace_start_lock); - /* ftrace_start is true if ftrace is running */ - if (ftrace_start) + /* ftrace_start_up is true if ftrace is running */ + if (ftrace_start_up) command |= FTRACE_DISABLE_CALLS; ftrace_run_update_code(command); @@ -1186,7 +1213,7 @@ ftrace_regex_release(struct inode *inode, struct file *file, int enable) mutex_lock(&ftrace_sysctl_lock); mutex_lock(&ftrace_start_lock); - if (iter->filtered && ftrace_start && ftrace_enabled) + if (iter->filtered && ftrace_start_up && ftrace_enabled) ftrace_run_update_code(FTRACE_ENABLE_CALLS); mutex_unlock(&ftrace_start_lock); mutex_unlock(&ftrace_sysctl_lock); -- cgit v1.2.3 From 4b33669e817a01dd99ff91df330d504ccfb2e99c Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Wed, 5 Nov 2008 18:30:25 -0800 Subject: x86: signal_32: do save_i387_xstate() at get_sigframe() Impact: cleanup move calling save_i387_xstate() into get_sigframe() from setup_sigcontext() like 64bit. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/kernel/signal_32.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/signal_32.c b/arch/x86/kernel/signal_32.c index a0efc1b3c4c9..6a05c74b4084 100644 --- a/arch/x86/kernel/signal_32.c +++ b/arch/x86/kernel/signal_32.c @@ -303,11 +303,7 @@ setup_sigcontext(struct sigcontext __user *sc, void __user *fpstate, err |= __put_user(regs->sp, &sc->sp_at_signal); err |= __put_user(regs->ss, (unsigned int __user *)&sc->ss); - tmp = save_i387_xstate(fpstate); - if (tmp < 0) - err = 1; - else - err |= __put_user(tmp ? fpstate : NULL, &sc->fpstate); + err |= __put_user(fpstate, &sc->fpstate); /* non-iBCS2 extensions.. */ err |= __put_user(mask, &sc->oldmask); @@ -350,6 +346,8 @@ get_sigframe(struct k_sigaction *ka, struct pt_regs *regs, size_t frame_size, if (used_math()) { sp = sp - sig_xstate_size; *fpstate = (struct _fpstate *) sp; + if (save_i387_xstate(*fpstate) < 0) + return (void __user *)-1L; } sp -= frame_size; -- cgit v1.2.3 From 99ea1b93bf80a287dd70499b96d9c4d06f320ff2 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Wed, 5 Nov 2008 18:32:54 -0800 Subject: x86: ia32_signal: do save_i387_xstate_ia32 at get_sigframe() Impact: cleanup move calling save_i387_xstate_ia32() into get_sigframe() from setup_sigcontext(). Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/ia32/ia32_signal.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/ia32/ia32_signal.c b/arch/x86/ia32/ia32_signal.c index 4bc02b23674b..47ddc23f4f54 100644 --- a/arch/x86/ia32/ia32_signal.c +++ b/arch/x86/ia32/ia32_signal.c @@ -367,12 +367,7 @@ static int ia32_setup_sigcontext(struct sigcontext_ia32 __user *sc, err |= __put_user(regs->flags, &sc->flags); err |= __put_user(regs->sp, &sc->sp_at_signal); - tmp = save_i387_xstate_ia32(fpstate); - if (tmp < 0) - err = -EFAULT; - else - err |= __put_user(ptr_to_compat(tmp ? fpstate : NULL), - &sc->fpstate); + err |= __put_user(ptr_to_compat(fpstate), &sc->fpstate); /* non-iBCS2 extensions.. */ err |= __put_user(mask, &sc->oldmask); @@ -408,6 +403,8 @@ static void __user *get_sigframe(struct k_sigaction *ka, struct pt_regs *regs, if (used_math()) { sp = sp - sig_xstate_ia32_size; *fpstate = (struct _fpstate_ia32 *) sp; + if (save_i387_xstate_ia32(*fpstate) < 0) + return (void __user *) -1L; } sp -= frame_size; -- cgit v1.2.3 From ee7d523c124a186ce3a886868de9cd1d8bc991f3 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Wed, 5 Nov 2008 18:33:35 -0800 Subject: x86: signal_64: setup fpstate in setup_sigcontext() Impact: cleanup set fpstate field of signal context at setup_sigcontext(). Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/kernel/signal_64.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/signal_64.c b/arch/x86/kernel/signal_64.c index cfbb60a5f9d2..97d26fa62ac1 100644 --- a/arch/x86/kernel/signal_64.c +++ b/arch/x86/kernel/signal_64.c @@ -159,8 +159,9 @@ asmlinkage long sys_rt_sigreturn(struct pt_regs *regs) */ static inline int -setup_sigcontext(struct sigcontext __user *sc, struct pt_regs *regs, - unsigned long mask, struct task_struct *me) +setup_sigcontext(struct sigcontext __user *sc, void __user *fpstate, + struct pt_regs *regs, + unsigned long mask, struct task_struct *me) { int err = 0; @@ -188,6 +189,7 @@ setup_sigcontext(struct sigcontext __user *sc, struct pt_regs *regs, err |= __put_user(me->thread.error_code, &sc->err); err |= __put_user(regs->ip, &sc->ip); err |= __put_user(regs->flags, &sc->flags); + err |= __put_user(fpstate, &sc->fpstate); err |= __put_user(mask, &sc->oldmask); err |= __put_user(me->thread.cr2, &sc->cr2); @@ -249,8 +251,8 @@ static int __setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, err |= __put_user(sas_ss_flags(regs->sp), &frame->uc.uc_stack.ss_flags); err |= __put_user(me->sas_ss_size, &frame->uc.uc_stack.ss_size); - err |= setup_sigcontext(&frame->uc.uc_mcontext, regs, set->sig[0], me); - err |= __put_user(fp, &frame->uc.uc_mcontext.fpstate); + err |= setup_sigcontext(&frame->uc.uc_mcontext, fp, + regs, set->sig[0], me); err |= __copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set)); /* Set up to return from userspace. If provided, use a stub -- cgit v1.2.3 From 8735b7d0a2a6246faa406a8cdd1376bd0e689ba3 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Wed, 5 Nov 2008 18:34:35 -0800 Subject: x86: signal_64: make setup_sigcontext() similar Impact: cleanup remove passing task struct. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/kernel/signal_64.c | 12 +++++------- 1 file changed, 5 insertions(+), 7 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/signal_64.c b/arch/x86/kernel/signal_64.c index 97d26fa62ac1..3868c2a21793 100644 --- a/arch/x86/kernel/signal_64.c +++ b/arch/x86/kernel/signal_64.c @@ -160,8 +160,7 @@ asmlinkage long sys_rt_sigreturn(struct pt_regs *regs) static inline int setup_sigcontext(struct sigcontext __user *sc, void __user *fpstate, - struct pt_regs *regs, - unsigned long mask, struct task_struct *me) + struct pt_regs *regs, unsigned long mask) { int err = 0; @@ -185,13 +184,13 @@ setup_sigcontext(struct sigcontext __user *sc, void __user *fpstate, err |= __put_user(regs->r13, &sc->r13); err |= __put_user(regs->r14, &sc->r14); err |= __put_user(regs->r15, &sc->r15); - err |= __put_user(me->thread.trap_no, &sc->trapno); - err |= __put_user(me->thread.error_code, &sc->err); + err |= __put_user(current->thread.trap_no, &sc->trapno); + err |= __put_user(current->thread.error_code, &sc->err); err |= __put_user(regs->ip, &sc->ip); err |= __put_user(regs->flags, &sc->flags); err |= __put_user(fpstate, &sc->fpstate); err |= __put_user(mask, &sc->oldmask); - err |= __put_user(me->thread.cr2, &sc->cr2); + err |= __put_user(current->thread.cr2, &sc->cr2); return err; } @@ -251,8 +250,7 @@ static int __setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, err |= __put_user(sas_ss_flags(regs->sp), &frame->uc.uc_stack.ss_flags); err |= __put_user(me->sas_ss_size, &frame->uc.uc_stack.ss_size); - err |= setup_sigcontext(&frame->uc.uc_mcontext, fp, - regs, set->sig[0], me); + err |= setup_sigcontext(&frame->uc.uc_mcontext, fp, regs, set->sig[0]); err |= __copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set)); /* Set up to return from userspace. If provided, use a stub -- cgit v1.2.3 From fd51b2d7d5df932767b89e00d0871a38a2c53e74 Mon Sep 17 00:00:00 2001 From: KOSAKI Motohiro Date: Wed, 5 Nov 2008 02:27:19 +0900 Subject: x86: update CONFIG_NUMA description Impact: clarify/update CONFIG_NUMA text CONFIG_NUMA description talk about a bit old thing. So, following changes are better. o CONFIG_NUMA is no longer EXPERIMENTAL o Opteron is not the only processor of NUMA topology on x86_64 no longer, but also Intel Core7i has it. Signed-off-by: KOSAKI Motohiro Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 16 ++++++++++------ 1 file changed, 10 insertions(+), 6 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 350bee1d54dc..38ae04bf6514 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -951,22 +951,26 @@ config ARCH_PHYS_ADDR_T_64BIT # Common NUMA Features config NUMA - bool "Numa Memory Allocation and Scheduler Support (EXPERIMENTAL)" + bool "Numa Memory Allocation and Scheduler Support" depends on SMP depends on X86_64 || (X86_32 && HIGHMEM64G && (X86_NUMAQ || X86_BIGSMP || X86_SUMMIT && ACPI) && EXPERIMENTAL) default n if X86_PC default y if (X86_NUMAQ || X86_SUMMIT || X86_BIGSMP) help Enable NUMA (Non Uniform Memory Access) support. + The kernel will try to allocate memory used by a CPU on the local memory controller of the CPU and add some more NUMA awareness to the kernel. - For 32-bit this is currently highly experimental and should be only - used for kernel development. It might also cause boot failures. - For 64-bit this is recommended on all multiprocessor Opteron systems. - If the system is EM64T, you should say N unless your system is - EM64T NUMA. + For 64-bit this is recommended if the system is Intel Core 7i + (or later), AMD Opteron, or EM64T NUMA. + + For 32-bit this is only needed on (rare) 32-bit-only platforms + that support NUMA topologies, such as NUMAQ / Summit, or if you + boot a 32-bit kernel on a 64-bit NUMA platform. + + Otherwise, you should say N. comment "NUMA (Summit) requires SMP, 64GB highmem support, ACPI" depends on X86_32 && X86_SUMMIT && (!HIGHMEM64G || !ACPI) -- cgit v1.2.3 From a87d091434ed2a34d647979ab12084139ee1fe41 Mon Sep 17 00:00:00 2001 From: Ken Chen Date: Thu, 6 Nov 2008 11:10:49 -0800 Subject: x86, sched: enable wchan config menu item on 64-bit Enable the wchan config menu item for now on x86-64 arch? This will at least allow people to enable/disable frame pointers on scheduler functions. Signed-off-by: Ken Chen Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 6f20718d3156..488a4ecd0b59 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -367,7 +367,7 @@ config X86_RDC321X config SCHED_NO_NO_OMIT_FRAME_POINTER def_bool y prompt "Single-depth WCHAN output" - depends on X86_32 + depends on X86 help Calculate simpler /proc//wchan values. If this option is disabled then wchan values will recurse back to the -- cgit v1.2.3 From 15002fa9bf3a79ac9dcafba7ff308586936088b2 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Fri, 7 Nov 2008 19:25:36 -0800 Subject: x86: signal: cosmetic unification of setup_sigcontext() Impact: cleanup Make setup_sigcontext() same. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/kernel/signal_32.c | 32 ++++++++++++++++++++++++++++---- arch/x86/kernel/signal_64.c | 33 ++++++++++++++++++++++++++++----- 2 files changed, 56 insertions(+), 9 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/signal_32.c b/arch/x86/kernel/signal_32.c index 6a05c74b4084..27a5c8174322 100644 --- a/arch/x86/kernel/signal_32.c +++ b/arch/x86/kernel/signal_32.c @@ -279,14 +279,20 @@ static int setup_sigcontext(struct sigcontext __user *sc, void __user *fpstate, struct pt_regs *regs, unsigned long mask) { - int tmp, err = 0; + int err = 0; - err |= __put_user(regs->fs, (unsigned int __user *)&sc->fs); - savesegment(gs, tmp); - err |= __put_user(tmp, (unsigned int __user *)&sc->gs); +#ifdef CONFIG_X86_32 + { + unsigned int tmp; + savesegment(gs, tmp); + err |= __put_user(tmp, (unsigned int __user *)&sc->gs); + } + err |= __put_user(regs->fs, (unsigned int __user *)&sc->fs); err |= __put_user(regs->es, (unsigned int __user *)&sc->es); err |= __put_user(regs->ds, (unsigned int __user *)&sc->ds); +#endif /* CONFIG_X86_32 */ + err |= __put_user(regs->di, &sc->di); err |= __put_user(regs->si, &sc->si); err |= __put_user(regs->bp, &sc->bp); @@ -295,13 +301,31 @@ setup_sigcontext(struct sigcontext __user *sc, void __user *fpstate, err |= __put_user(regs->dx, &sc->dx); err |= __put_user(regs->cx, &sc->cx); err |= __put_user(regs->ax, &sc->ax); +#ifdef CONFIG_X86_64 + err |= __put_user(regs->r8, &sc->r8); + err |= __put_user(regs->r9, &sc->r9); + err |= __put_user(regs->r10, &sc->r10); + err |= __put_user(regs->r11, &sc->r11); + err |= __put_user(regs->r12, &sc->r12); + err |= __put_user(regs->r13, &sc->r13); + err |= __put_user(regs->r14, &sc->r14); + err |= __put_user(regs->r15, &sc->r15); +#endif /* CONFIG_X86_64 */ + err |= __put_user(current->thread.trap_no, &sc->trapno); err |= __put_user(current->thread.error_code, &sc->err); err |= __put_user(regs->ip, &sc->ip); +#ifdef CONFIG_X86_32 err |= __put_user(regs->cs, (unsigned int __user *)&sc->cs); err |= __put_user(regs->flags, &sc->flags); err |= __put_user(regs->sp, &sc->sp_at_signal); err |= __put_user(regs->ss, (unsigned int __user *)&sc->ss); +#else /* !CONFIG_X86_32 */ + err |= __put_user(regs->flags, &sc->flags); + err |= __put_user(regs->cs, &sc->cs); + err |= __put_user(0, &sc->gs); + err |= __put_user(0, &sc->fs); +#endif /* CONFIG_X86_32 */ err |= __put_user(fpstate, &sc->fpstate); diff --git a/arch/x86/kernel/signal_64.c b/arch/x86/kernel/signal_64.c index 3868c2a21793..d2307e41fbdb 100644 --- a/arch/x86/kernel/signal_64.c +++ b/arch/x86/kernel/signal_64.c @@ -157,16 +157,23 @@ asmlinkage long sys_rt_sigreturn(struct pt_regs *regs) /* * Set up a signal frame. */ - -static inline int +static int setup_sigcontext(struct sigcontext __user *sc, void __user *fpstate, struct pt_regs *regs, unsigned long mask) { int err = 0; - err |= __put_user(regs->cs, &sc->cs); - err |= __put_user(0, &sc->gs); - err |= __put_user(0, &sc->fs); +#ifdef CONFIG_X86_32 + { + unsigned int tmp; + + savesegment(gs, tmp); + err |= __put_user(tmp, (unsigned int __user *)&sc->gs); + } + err |= __put_user(regs->fs, (unsigned int __user *)&sc->fs); + err |= __put_user(regs->es, (unsigned int __user *)&sc->es); + err |= __put_user(regs->ds, (unsigned int __user *)&sc->ds); +#endif /* CONFIG_X86_32 */ err |= __put_user(regs->di, &sc->di); err |= __put_user(regs->si, &sc->si); @@ -176,6 +183,7 @@ setup_sigcontext(struct sigcontext __user *sc, void __user *fpstate, err |= __put_user(regs->dx, &sc->dx); err |= __put_user(regs->cx, &sc->cx); err |= __put_user(regs->ax, &sc->ax); +#ifdef CONFIG_X86_64 err |= __put_user(regs->r8, &sc->r8); err |= __put_user(regs->r9, &sc->r9); err |= __put_user(regs->r10, &sc->r10); @@ -184,11 +192,26 @@ setup_sigcontext(struct sigcontext __user *sc, void __user *fpstate, err |= __put_user(regs->r13, &sc->r13); err |= __put_user(regs->r14, &sc->r14); err |= __put_user(regs->r15, &sc->r15); +#endif /* CONFIG_X86_64 */ + err |= __put_user(current->thread.trap_no, &sc->trapno); err |= __put_user(current->thread.error_code, &sc->err); err |= __put_user(regs->ip, &sc->ip); +#ifdef CONFIG_X86_32 + err |= __put_user(regs->cs, (unsigned int __user *)&sc->cs); err |= __put_user(regs->flags, &sc->flags); + err |= __put_user(regs->sp, &sc->sp_at_signal); + err |= __put_user(regs->ss, (unsigned int __user *)&sc->ss); +#else /* !CONFIG_X86_32 */ + err |= __put_user(regs->flags, &sc->flags); + err |= __put_user(regs->cs, &sc->cs); + err |= __put_user(0, &sc->gs); + err |= __put_user(0, &sc->fs); +#endif /* CONFIG_X86_32 */ + err |= __put_user(fpstate, &sc->fpstate); + + /* non-iBCS2 extensions.. */ err |= __put_user(mask, &sc->oldmask); err |= __put_user(current->thread.cr2, &sc->cr2); -- cgit v1.2.3 From cb9e35dce94a1b9c59d46224e8a94377d673e204 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Sat, 8 Nov 2008 20:27:00 +0100 Subject: x86: clean up rdtsc_barrier() use Impact: cleanup Move rdtsc_barrier() use to vsyscall_64.c where it's relied on, and point out its role in the context of its use. Signed-off-by: Ingo Molnar --- arch/x86/include/asm/tsc.h | 6 +----- arch/x86/kernel/vsyscall_64.c | 9 +++++++++ 2 files changed, 10 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/tsc.h b/arch/x86/include/asm/tsc.h index 9cd83a8e40d5..700aeb8d2098 100644 --- a/arch/x86/include/asm/tsc.h +++ b/arch/x86/include/asm/tsc.h @@ -44,11 +44,7 @@ static __always_inline cycles_t vget_cycles(void) if (!cpu_has_tsc) return 0; #endif - rdtsc_barrier(); - cycles = (cycles_t)__native_read_tsc(); - rdtsc_barrier(); - - return cycles; + return (cycles_t)__native_read_tsc(); } extern void tsc_init(void); diff --git a/arch/x86/kernel/vsyscall_64.c b/arch/x86/kernel/vsyscall_64.c index 0b8b6690a86d..ebf2f12900f5 100644 --- a/arch/x86/kernel/vsyscall_64.c +++ b/arch/x86/kernel/vsyscall_64.c @@ -128,7 +128,16 @@ static __always_inline void do_vgettimeofday(struct timeval * tv) gettimeofday(tv,NULL); return; } + + /* + * Surround the RDTSC by barriers, to make sure it's not + * speculated to outside the seqlock critical section and + * does not cause time warps: + */ + rdtsc_barrier(); now = vread(); + rdtsc_barrier(); + base = __vsyscall_gtod_data.clock.cycle_last; mask = __vsyscall_gtod_data.clock.mask; mult = __vsyscall_gtod_data.clock.mult; -- cgit v1.2.3 From 4fcc50abdffb517cee36cec9cb22138d84fb62d0 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Sun, 9 Nov 2008 08:10:03 +0100 Subject: x86: clean up vget_cycles() Impact: remove unused variable I forgot to remove the now unused "cycles_t cycles" parameter from vget_cycles() - which triggers build warnings as tsc.h is included in a number of files. Remove it. Signed-off-by: Ingo Molnar --- arch/x86/include/asm/tsc.h | 2 -- 1 file changed, 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/tsc.h b/arch/x86/include/asm/tsc.h index 700aeb8d2098..38ae163cc91b 100644 --- a/arch/x86/include/asm/tsc.h +++ b/arch/x86/include/asm/tsc.h @@ -34,8 +34,6 @@ static inline cycles_t get_cycles(void) static __always_inline cycles_t vget_cycles(void) { - cycles_t cycles; - /* * We only do VDSOs on TSC capable CPUs, so this shouldnt * access boot_cpu_data (which is not VDSO-safe): -- cgit v1.2.3 From 19f47c634ea8c5a10ff7bb1a08c52fd0f49bc54c Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Sun, 9 Nov 2008 21:28:15 -0800 Subject: x86: x86_32 has its own irq_regs definition Impact: cleanup Arches that have their own irq_regs definition are expected to define ARCH_HAS_OWN_IRQ_REGS or else a generic (unused) set will also be defined in lib/irq_regs.c Sparse noticed the unused generic one had no prototype: lib/irq_regs.c:15:1: warning: symbol 'per_cpu____irq_regs' was not declared. Should it be static? Signed-off-by: Harvey Harrison Signed-off-by: Ingo Molnar --- arch/x86/include/asm/irq_regs_32.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/irq_regs_32.h b/arch/x86/include/asm/irq_regs_32.h index af2f02d27fc7..86afd7473457 100644 --- a/arch/x86/include/asm/irq_regs_32.h +++ b/arch/x86/include/asm/irq_regs_32.h @@ -9,6 +9,8 @@ #include +#define ARCH_HAS_OWN_IRQ_REGS + DECLARE_PER_CPU(struct pt_regs *, irq_regs); static inline struct pt_regs *get_irq_regs(void) -- cgit v1.2.3 From f4166c54bfe04f64603974058e44fbd7cfef0ccc Mon Sep 17 00:00:00 2001 From: Markus Metzger Date: Sun, 9 Nov 2008 14:29:21 +0100 Subject: x86, bts: DS and BTS initialization Impact: widen BTS/PEBS ptrace enablement to more CPU models Move BTS initialisation out of an #ifdef CONFIG_X86_64 guard. Assume core2 BTS and DS layout for future models of family 6 processors. Signed-off-by: Markus Metzger Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/intel.c | 3 +-- arch/x86/kernel/ds.c | 9 ++++----- arch/x86/kernel/ptrace.c | 9 ++++----- 3 files changed, 9 insertions(+), 12 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/intel.c b/arch/x86/kernel/cpu/intel.c index cce0b6118d55..816f27f289b1 100644 --- a/arch/x86/kernel/cpu/intel.c +++ b/arch/x86/kernel/cpu/intel.c @@ -307,12 +307,11 @@ static void __cpuinit init_intel(struct cpuinfo_x86 *c) set_cpu_cap(c, X86_FEATURE_P4); if (c->x86 == 6) set_cpu_cap(c, X86_FEATURE_P3); +#endif if (cpu_has_bts) ptrace_bts_init_intel(c); -#endif - detect_extended_topology(c); if (!cpu_has(c, X86_FEATURE_XTOPOLOGY)) { /* diff --git a/arch/x86/kernel/ds.c b/arch/x86/kernel/ds.c index 2b69994fd3a8..c570252905a1 100644 --- a/arch/x86/kernel/ds.c +++ b/arch/x86/kernel/ds.c @@ -821,17 +821,16 @@ void __cpuinit ds_init_intel(struct cpuinfo_x86 *c) switch (c->x86) { case 0x6: switch (c->x86_model) { + case 0 ... 0xC: + /* sorry, don't know about them */ + break; case 0xD: case 0xE: /* Pentium M */ ds_configure(&ds_cfg_var); break; - case 0xF: /* Core2 */ - case 0x1C: /* Atom */ + default: /* Core2, Atom, ... */ ds_configure(&ds_cfg_64); break; - default: - /* sorry, don't know about them */ - break; } break; case 0xF: diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c index 0a6d8c12e10d..06180dff5b2e 100644 --- a/arch/x86/kernel/ptrace.c +++ b/arch/x86/kernel/ptrace.c @@ -929,17 +929,16 @@ void __cpuinit ptrace_bts_init_intel(struct cpuinfo_x86 *c) switch (c->x86) { case 0x6: switch (c->x86_model) { + case 0 ... 0xC: + /* sorry, don't know about them */ + break; case 0xD: case 0xE: /* Pentium M */ bts_configure(&bts_cfg_pentium_m); break; - case 0xF: /* Core2 */ - case 0x1C: /* Atom */ + default: /* Core2, Atom, ... */ bts_configure(&bts_cfg_core2); break; - default: - /* sorry, don't know about them */ - break; } break; case 0xF: -- cgit v1.2.3 From 4e0304310f5180eee11b4edc72cf4cb78acdc634 Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Mon, 10 Nov 2008 09:16:40 +0100 Subject: x86: apic - calibrate_APIC_clock remove redundant irq-enable-disable Impact: cleanup lapic_timer_setup is self-protected with local_irq_save/restore no need to use them in caller and levt is the per-cpu variable so no concurrent access from another cpu. Signed-off-by: Cyrill Gorcunov Acked-by: "Maciej W. Rozycki" Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic.c | 4 ---- 1 file changed, 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/apic.c b/arch/x86/kernel/apic.c index 04a7f960bbc0..ce90dc184139 100644 --- a/arch/x86/kernel/apic.c +++ b/arch/x86/kernel/apic.c @@ -672,13 +672,9 @@ static int __init calibrate_APIC_clock(void) while (lapic_cal_loops <= LAPIC_CAL_LOOPS) cpu_relax(); - local_irq_disable(); - /* Stop the lapic timer */ lapic_timer_setup(CLOCK_EVT_MODE_SHUTDOWN, levt); - local_irq_enable(); - /* Jiffies delta */ deltaj = lapic_cal_j2 - lapic_cal_j1; apic_printk(APIC_VERBOSE, "... jiffies delta = %lu\n", deltaj); -- cgit v1.2.3 From ba21ebb6abac5c46e1d818d2ceda82420bd099ba Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Mon, 10 Nov 2008 09:16:41 +0100 Subject: x86: apic - use pr_ macros for logging Impact: cleanup It saves us some source lines and shift the code a bit righter. And a multiline comment style is fixed too :-) Signed-off-by: Cyrill Gorcunov Acked-by: "Maciej W. Rozycki" Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic.c | 116 ++++++++++++++++++++++--------------------------- 1 file changed, 53 insertions(+), 63 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/apic.c b/arch/x86/kernel/apic.c index ce90dc184139..70879c9e3936 100644 --- a/arch/x86/kernel/apic.c +++ b/arch/x86/kernel/apic.c @@ -559,13 +559,13 @@ static int __init calibrate_by_pmtimer(long deltapm, long *delta) } else { res = (((u64)deltapm) * mult) >> 22; do_div(res, 1000000); - printk(KERN_WARNING "APIC calibration not consistent " + pr_warning("APIC calibration not consistent " "with PM Timer: %ldms instead of 100ms\n", (long)res); /* Correct the lapic counter value */ res = (((u64)(*delta)) * pm_100ms); do_div(res, deltapm); - printk(KERN_INFO "APIC delta adjusted to PM-Timer: " + pr_info("APIC delta adjusted to PM-Timer: " "%lu (%ld)\n", (unsigned long)res, *delta); *delta = (long)res; } @@ -645,8 +645,7 @@ static int __init calibrate_APIC_clock(void) */ if (calibration_result < (1000000 / HZ)) { local_irq_enable(); - printk(KERN_WARNING - "APIC frequency too slow, disabling apic timer\n"); + pr_warning("APIC frequency too slow, disabling apic timer\n"); return -1; } @@ -688,8 +687,7 @@ static int __init calibrate_APIC_clock(void) local_irq_enable(); if (levt->features & CLOCK_EVT_FEAT_DUMMY) { - printk(KERN_WARNING - "APIC timer disabled due to verification failure.\n"); + pr_warning("APIC timer disabled due to verification failure.\n"); return -1; } @@ -710,7 +708,7 @@ void __init setup_boot_APIC_clock(void) * broadcast mechanism is used. On UP systems simply ignore it. */ if (disable_apic_timer) { - printk(KERN_INFO "Disabling APIC timer\n"); + pr_info("Disabling APIC timer\n"); /* No broadcast on UP ! */ if (num_possible_cpus() > 1) { lapic_clockevent.mult = 1; @@ -737,7 +735,7 @@ void __init setup_boot_APIC_clock(void) if (nmi_watchdog != NMI_IO_APIC) lapic_clockevent.features &= ~CLOCK_EVT_FEAT_DUMMY; else - printk(KERN_WARNING "APIC timer registered as dummy," + pr_warning("APIC timer registered as dummy," " due to nmi_watchdog=%d!\n", nmi_watchdog); /* Setup the lapic or request the broadcast */ @@ -769,8 +767,7 @@ static void local_apic_timer_interrupt(void) * spurious. */ if (!evt->event_handler) { - printk(KERN_WARNING - "Spurious LAPIC timer interrupt on cpu %d\n", cpu); + pr_warning("Spurious LAPIC timer interrupt on cpu %d\n", cpu); /* Switch it off */ lapic_timer_setup(CLOCK_EVT_MODE_SHUTDOWN, evt); return; @@ -1089,7 +1086,7 @@ static void __cpuinit lapic_setup_esr(void) unsigned int oldvalue, value, maxlvt; if (!lapic_is_integrated()) { - printk(KERN_INFO "No ESR for 82489DX.\n"); + pr_info("No ESR for 82489DX.\n"); return; } @@ -1100,7 +1097,7 @@ static void __cpuinit lapic_setup_esr(void) * ESR disabled - we can't do anything useful with the * errors anyway - mbligh */ - printk(KERN_INFO "Leaving ESR disabled.\n"); + pr_info("Leaving ESR disabled.\n"); return; } @@ -1294,7 +1291,7 @@ void check_x2apic(void) rdmsr(MSR_IA32_APICBASE, msr, msr2); if (msr & X2APIC_ENABLE) { - printk("x2apic enabled by BIOS, switching to x2apic ops\n"); + pr_info("x2apic enabled by BIOS, switching to x2apic ops\n"); x2apic_preenabled = x2apic = 1; apic_ops = &x2apic_ops; } @@ -1306,7 +1303,7 @@ void enable_x2apic(void) rdmsr(MSR_IA32_APICBASE, msr, msr2); if (!(msr & X2APIC_ENABLE)) { - printk("Enabling x2apic\n"); + pr_info("Enabling x2apic\n"); wrmsr(MSR_IA32_APICBASE, msr | X2APIC_ENABLE, 0); } } @@ -1321,9 +1318,8 @@ void enable_IR_x2apic(void) return; if (!x2apic_preenabled && disable_x2apic) { - printk(KERN_INFO - "Skipped enabling x2apic and Interrupt-remapping " - "because of nox2apic\n"); + pr_info("Skipped enabling x2apic and Interrupt-remapping " + "because of nox2apic\n"); return; } @@ -1331,22 +1327,19 @@ void enable_IR_x2apic(void) panic("Bios already enabled x2apic, can't enforce nox2apic"); if (!x2apic_preenabled && skip_ioapic_setup) { - printk(KERN_INFO - "Skipped enabling x2apic and Interrupt-remapping " - "because of skipping io-apic setup\n"); + pr_info("Skipped enabling x2apic and Interrupt-remapping " + "because of skipping io-apic setup\n"); return; } ret = dmar_table_init(); if (ret) { - printk(KERN_INFO - "dmar_table_init() failed with %d:\n", ret); + pr_info("dmar_table_init() failed with %d:\n", ret); if (x2apic_preenabled) panic("x2apic enabled by bios. But IR enabling failed"); else - printk(KERN_INFO - "Not enabling x2apic,Intr-remapping\n"); + pr_info("Not enabling x2apic,Intr-remapping\n"); return; } @@ -1355,7 +1348,7 @@ void enable_IR_x2apic(void) ret = save_mask_IO_APIC_setup(); if (ret) { - printk(KERN_INFO "Saving IO-APIC state failed: %d\n", ret); + pr_info("Saving IO-APIC state failed: %d\n", ret); goto end; } @@ -1390,14 +1383,11 @@ end: if (!ret) { if (!x2apic_preenabled) - printk(KERN_INFO - "Enabled x2apic and interrupt-remapping\n"); + pr_info("Enabled x2apic and interrupt-remapping\n"); else - printk(KERN_INFO - "Enabled Interrupt-remapping\n"); + pr_info("Enabled Interrupt-remapping\n"); } else - printk(KERN_ERR - "Failed to enable Interrupt-remapping and x2apic\n"); + pr_err("Failed to enable Interrupt-remapping and x2apic\n"); #else if (!cpu_has_x2apic) return; @@ -1406,8 +1396,8 @@ end: panic("x2apic enabled prior OS handover," " enable CONFIG_INTR_REMAP"); - printk(KERN_INFO "Enable CONFIG_INTR_REMAP for enabling intr-remapping " - " and x2apic\n"); + pr_info("Enable CONFIG_INTR_REMAP for enabling intr-remapping " + " and x2apic\n"); #endif return; @@ -1424,7 +1414,7 @@ end: static int __init detect_init_APIC(void) { if (!cpu_has_apic) { - printk(KERN_INFO "No local APIC present\n"); + pr_info("No local APIC present\n"); return -1; } @@ -1465,8 +1455,8 @@ static int __init detect_init_APIC(void) * "lapic" specified. */ if (!force_enable_local_apic) { - printk(KERN_INFO "Local APIC disabled by BIOS -- " - "you can enable it with \"lapic\"\n"); + pr_info("Local APIC disabled by BIOS -- " + "you can enable it with \"lapic\"\n"); return -1; } /* @@ -1476,8 +1466,7 @@ static int __init detect_init_APIC(void) */ rdmsr(MSR_IA32_APICBASE, l, h); if (!(l & MSR_IA32_APICBASE_ENABLE)) { - printk(KERN_INFO - "Local APIC disabled by BIOS -- reenabling.\n"); + pr_info("Local APIC disabled by BIOS -- reenabling.\n"); l &= ~MSR_IA32_APICBASE_BASE; l |= MSR_IA32_APICBASE_ENABLE | APIC_DEFAULT_PHYS_BASE; wrmsr(MSR_IA32_APICBASE, l, h); @@ -1490,7 +1479,7 @@ static int __init detect_init_APIC(void) */ features = cpuid_edx(1); if (!(features & (1 << X86_FEATURE_APIC))) { - printk(KERN_WARNING "Could not enable APIC!\n"); + pr_warning("Could not enable APIC!\n"); return -1; } set_cpu_cap(&boot_cpu_data, X86_FEATURE_APIC); @@ -1501,14 +1490,14 @@ static int __init detect_init_APIC(void) if (l & MSR_IA32_APICBASE_ENABLE) mp_lapic_addr = l & MSR_IA32_APICBASE_BASE; - printk(KERN_INFO "Found and enabled local APIC!\n"); + pr_info("Found and enabled local APIC!\n"); apic_pm_activate(); return 0; no_apic: - printk(KERN_INFO "No local APIC present or hardware disabled\n"); + pr_info("No local APIC present or hardware disabled\n"); return -1; } #endif @@ -1584,12 +1573,12 @@ int __init APIC_init_uniprocessor(void) { #ifdef CONFIG_X86_64 if (disable_apic) { - printk(KERN_INFO "Apic disabled\n"); + pr_info("Apic disabled\n"); return -1; } if (!cpu_has_apic) { disable_apic = 1; - printk(KERN_INFO "Apic disabled by BIOS\n"); + pr_info("Apic disabled by BIOS\n"); return -1; } #else @@ -1601,8 +1590,8 @@ int __init APIC_init_uniprocessor(void) */ if (!cpu_has_apic && APIC_INTEGRATED(apic_version[boot_cpu_physical_apicid])) { - printk(KERN_ERR "BIOS bug, local APIC 0x%x not detected!...\n", - boot_cpu_physical_apicid); + pr_err("BIOS bug, local APIC 0x%x not detected!...\n", + boot_cpu_physical_apicid); clear_cpu_cap(&boot_cpu_data, X86_FEATURE_APIC); return -1; } @@ -1695,8 +1684,8 @@ void smp_spurious_interrupt(struct pt_regs *regs) add_pda(irq_spurious_count, 1); #else /* see sw-dev-man vol 3, chapter 7.4.13.5 */ - printk(KERN_INFO "spurious APIC interrupt on CPU#%d, " - "should never happen.\n", smp_processor_id()); + pr_info("spurious APIC interrupt on CPU#%d, " + "should never happen.\n", smp_processor_id()); __get_cpu_var(irq_stat).irq_spurious_count++; #endif irq_exit(); @@ -1720,17 +1709,18 @@ void smp_error_interrupt(struct pt_regs *regs) ack_APIC_irq(); atomic_inc(&irq_err_count); - /* Here is what the APIC error bits mean: - 0: Send CS error - 1: Receive CS error - 2: Send accept error - 3: Receive accept error - 4: Reserved - 5: Send illegal vector - 6: Received illegal vector - 7: Illegal register address - */ - printk(KERN_DEBUG "APIC error on CPU%d: %02x(%02x)\n", + /* + * Here is what the APIC error bits mean: + * 0: Send CS error + * 1: Receive CS error + * 2: Send accept error + * 3: Receive accept error + * 4: Reserved + * 5: Send illegal vector + * 6: Received illegal vector + * 7: Illegal register address + */ + pr_debug("APIC error on CPU%d: %02x(%02x)\n", smp_processor_id(), v , v1); irq_exit(); } @@ -1834,15 +1824,15 @@ void __cpuinit generic_processor_info(int apicid, int version) * Validate version */ if (version == 0x0) { - printk(KERN_WARNING "BIOS bug, APIC version is 0 for CPU#%d! " - "fixing up to 0x10. (tell your hw vendor)\n", - version); + pr_warning("BIOS bug, APIC version is 0 for CPU#%d! " + "fixing up to 0x10. (tell your hw vendor)\n", + version); version = 0x10; } apic_version[apicid] = version; if (num_processors >= NR_CPUS) { - printk(KERN_WARNING "WARNING: NR_CPUS limit of %i reached." + pr_warning("WARNING: NR_CPUS limit of %i reached." " Processor ignored.\n", NR_CPUS); return; } @@ -2205,7 +2195,7 @@ static int __init apic_set_verbosity(char *arg) else if (strcmp("verbose", arg) == 0) apic_verbosity = APIC_VERBOSE; else { - printk(KERN_WARNING "APIC Verbosity level %s not recognised" + pr_warning("APIC Verbosity level %s not recognised" " use apic=verbose or apic=debug\n", arg); return -EINVAL; } -- cgit v1.2.3 From ae1e9130bfb9ad55eb97ec3fb17a122b7a118f98 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Tue, 11 Nov 2008 09:05:16 +0100 Subject: sched: rename SCHED_NO_NO_OMIT_FRAME_POINTER => SCHED_OMIT_FRAME_POINTER Impact: cleanup, change .config option name We had this ugly config name for a long time for hysteric raisons. Rename it to a saner name. We still cannot get rid of it completely, until /proc//stack usage replaces WCHAN usage for good. We'll be able to do that in the v2.6.29/v2.6.30 timeframe. Signed-off-by: Ingo Molnar --- arch/ia64/Kconfig | 2 +- arch/m32r/Kconfig | 2 +- arch/mips/Kconfig | 2 +- arch/powerpc/Kconfig | 2 +- arch/x86/Kconfig | 2 +- include/asm-m32r/system.h | 2 +- kernel/Makefile | 2 +- 7 files changed, 7 insertions(+), 7 deletions(-) (limited to 'arch/x86') diff --git a/arch/ia64/Kconfig b/arch/ia64/Kconfig index 27eec71429b0..59d12788b60c 100644 --- a/arch/ia64/Kconfig +++ b/arch/ia64/Kconfig @@ -99,7 +99,7 @@ config GENERIC_IOMAP bool default y -config SCHED_NO_NO_OMIT_FRAME_POINTER +config SCHED_OMIT_FRAME_POINTER bool default y diff --git a/arch/m32r/Kconfig b/arch/m32r/Kconfig index dbaed4a63815..29047d5c259a 100644 --- a/arch/m32r/Kconfig +++ b/arch/m32r/Kconfig @@ -273,7 +273,7 @@ config GENERIC_CALIBRATE_DELAY bool default y -config SCHED_NO_NO_OMIT_FRAME_POINTER +config SCHED_OMIT_FRAME_POINTER bool default y diff --git a/arch/mips/Kconfig b/arch/mips/Kconfig index f4af967a6b30..a5255e7c79e0 100644 --- a/arch/mips/Kconfig +++ b/arch/mips/Kconfig @@ -653,7 +653,7 @@ config GENERIC_CMOS_UPDATE bool default y -config SCHED_NO_NO_OMIT_FRAME_POINTER +config SCHED_OMIT_FRAME_POINTER bool default y diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig index 525c13a4de93..adb23ea1c1ef 100644 --- a/arch/powerpc/Kconfig +++ b/arch/powerpc/Kconfig @@ -141,7 +141,7 @@ config GENERIC_NVRAM bool default y if PPC32 -config SCHED_NO_NO_OMIT_FRAME_POINTER +config SCHED_OMIT_FRAME_POINTER bool default y diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 1d5550d19b66..74db682ec1ce 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -364,7 +364,7 @@ config X86_RDC321X as R-8610-(G). If you don't have one of these chips, you should say N here. -config SCHED_NO_NO_OMIT_FRAME_POINTER +config SCHED_OMIT_FRAME_POINTER def_bool y prompt "Single-depth WCHAN output" depends on X86 diff --git a/include/asm-m32r/system.h b/include/asm-m32r/system.h index 70a57c8c002b..c980f5ba8de7 100644 --- a/include/asm-m32r/system.h +++ b/include/asm-m32r/system.h @@ -23,7 +23,7 @@ */ #if defined(CONFIG_FRAME_POINTER) || \ - !defined(CONFIG_SCHED_NO_NO_OMIT_FRAME_POINTER) + !defined(CONFIG_SCHED_OMIT_FRAME_POINTER) #define M32R_PUSH_FP " push fp\n" #define M32R_POP_FP " pop fp\n" #else diff --git a/kernel/Makefile b/kernel/Makefile index e1af03972148..46e67a398495 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -91,7 +91,7 @@ obj-$(CONFIG_FUNCTION_TRACER) += trace/ obj-$(CONFIG_TRACING) += trace/ obj-$(CONFIG_SMP) += sched_cpupri.o -ifneq ($(CONFIG_SCHED_NO_NO_OMIT_FRAME_POINTER),y) +ifneq ($(CONFIG_SCHED_OMIT_FRAME_POINTER),y) # According to Alan Modra , the -fno-omit-frame-pointer is # needed for x86 only. Why this used to be enabled for all architectures is beyond # me. I suspect most platforms don't need this, but until we know that for sure -- cgit v1.2.3 From caf4b323b02a16c92fba449952ac6515ddc76d7a Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Tue, 11 Nov 2008 07:03:45 +0100 Subject: tracing, x86: add low level support for ftrace return tracing Impact: add infrastructure for function-return tracing Add low level support for ftrace return tracing. This plug-in stores return addresses on the thread_info structure of the current task. The index of the current return address is initialized when the task is the first one (init) and when a process forks (the child). It is not needed when a task does a sys_execve because after this syscall, it still needs to return on the kernel functions it called. Note that the code of return_to_handler has been suggested by Steven Rostedt as almost all of the ideas of improvements in this V3. For purpose of security, arch/x86/kernel/process_32.c is not traced because __switch_to() changes the current task during its execution. That could cause inconsistency in the stored return address of this function even if I didn't have any crash after testing with tracing on this function enabled. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 1 + arch/x86/include/asm/ftrace.h | 26 ++++++ arch/x86/include/asm/thread_info.h | 24 +++++ arch/x86/kernel/Makefile | 6 ++ arch/x86/kernel/entry_32.S | 33 +++++++ arch/x86/kernel/ftrace.c | 181 +++++++++++++++++++++++++++++++++++-- include/linux/ftrace.h | 20 ++++ include/linux/ftrace_irq.h | 2 +- include/linux/sched.h | 11 +++ kernel/Makefile | 4 + 10 files changed, 300 insertions(+), 8 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 27b8a3a39911..ca91e50bdb10 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -11,6 +11,7 @@ config 64BIT config X86_32 def_bool !64BIT + select HAVE_FUNCTION_RET_TRACER config X86_64 def_bool 64BIT diff --git a/arch/x86/include/asm/ftrace.h b/arch/x86/include/asm/ftrace.h index f8173ed1c970..9b6a1fa19e70 100644 --- a/arch/x86/include/asm/ftrace.h +++ b/arch/x86/include/asm/ftrace.h @@ -20,4 +20,30 @@ static inline unsigned long ftrace_call_adjust(unsigned long addr) #endif /* __ASSEMBLY__ */ #endif /* CONFIG_FUNCTION_TRACER */ +#ifdef CONFIG_FUNCTION_RET_TRACER +#define FTRACE_RET_STACK_SIZE 20 + +#ifndef __ASSEMBLY__ + +/* + * Stack of return addresses for functions + * of a thread. + * Used in struct thread_info + */ +struct ftrace_ret_stack { + unsigned long ret; + unsigned long func; + unsigned long long calltime; +}; + +/* + * Primary handler of a function return. + * It relays on ftrace_return_to_handler. + * Defined in entry32.S + */ +extern void return_to_handler(void); + +#endif /* __ASSEMBLY__ */ +#endif /* CONFIG_FUNCTION_RET_TRACER */ + #endif /* _ASM_X86_FTRACE_H */ diff --git a/arch/x86/include/asm/thread_info.h b/arch/x86/include/asm/thread_info.h index e44d379faad2..a71158369fd4 100644 --- a/arch/x86/include/asm/thread_info.h +++ b/arch/x86/include/asm/thread_info.h @@ -20,6 +20,7 @@ struct task_struct; struct exec_domain; #include +#include struct thread_info { struct task_struct *task; /* main task structure */ @@ -38,8 +39,30 @@ struct thread_info { */ __u8 supervisor_stack[0]; #endif + +#ifdef CONFIG_FUNCTION_RET_TRACER + /* Index of current stored adress in ret_stack */ + int curr_ret_stack; + /* Stack of return addresses for return function tracing */ + struct ftrace_ret_stack ret_stack[FTRACE_RET_STACK_SIZE]; +#endif }; +#ifdef CONFIG_FUNCTION_RET_TRACER +#define INIT_THREAD_INFO(tsk) \ +{ \ + .task = &tsk, \ + .exec_domain = &default_exec_domain, \ + .flags = 0, \ + .cpu = 0, \ + .preempt_count = 1, \ + .addr_limit = KERNEL_DS, \ + .restart_block = { \ + .fn = do_no_restart_syscall, \ + }, \ + .curr_ret_stack = -1,\ +} +#else #define INIT_THREAD_INFO(tsk) \ { \ .task = &tsk, \ @@ -52,6 +75,7 @@ struct thread_info { .fn = do_no_restart_syscall, \ }, \ } +#endif #define init_thread_info (init_thread_union.thread_info) #define init_stack (init_thread_union.stack) diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index e489ff9cb3e2..1d8ed95da846 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -14,6 +14,11 @@ CFLAGS_REMOVE_paravirt-spinlocks.o = -pg CFLAGS_REMOVE_ftrace.o = -pg endif +ifdef CONFIG_FUNCTION_RET_TRACER +# Don't trace __switch_to() but let it for function tracer +CFLAGS_REMOVE_process_32.o = -pg +endif + # # vsyscalls (which work on the user stack) should have # no stack-protector checks: @@ -65,6 +70,7 @@ obj-$(CONFIG_X86_LOCAL_APIC) += apic.o nmi.o obj-$(CONFIG_X86_IO_APIC) += io_apic.o obj-$(CONFIG_X86_REBOOTFIXUPS) += reboot_fixups_32.o obj-$(CONFIG_DYNAMIC_FTRACE) += ftrace.o +obj-$(CONFIG_FUNCTION_RET_TRACER) += ftrace.o obj-$(CONFIG_KEXEC) += machine_kexec_$(BITS).o obj-$(CONFIG_KEXEC) += relocate_kernel_$(BITS).o crash.o obj-$(CONFIG_CRASH_DUMP) += crash_dump_$(BITS).o diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S index 9134de814c97..9a0ac85946db 100644 --- a/arch/x86/kernel/entry_32.S +++ b/arch/x86/kernel/entry_32.S @@ -1188,6 +1188,10 @@ ENTRY(mcount) cmpl $ftrace_stub, ftrace_trace_function jnz trace +#ifdef CONFIG_FUNCTION_RET_TRACER + cmpl $ftrace_stub, ftrace_function_return + jnz trace_return +#endif .globl ftrace_stub ftrace_stub: ret @@ -1206,8 +1210,37 @@ trace: popl %edx popl %ecx popl %eax + jmp ftrace_stub +#ifdef CONFIG_FUNCTION_RET_TRACER +trace_return: + pushl %eax + pushl %ecx + pushl %edx + movl 0xc(%esp), %eax + pushl %eax + lea 0x4(%ebp), %eax + pushl %eax + call prepare_ftrace_return + addl $8, %esp + popl %edx + popl %ecx + popl %eax jmp ftrace_stub + +.globl return_to_handler +return_to_handler: + pushl $0 + pushl %eax + pushl %ecx + pushl %edx + call ftrace_return_to_handler + movl %eax, 0xc(%esp) + popl %edx + popl %ecx + popl %eax + ret +#endif /* CONFIG_FUNCTION_RET_TRACER */ END(mcount) #endif /* CONFIG_DYNAMIC_FTRACE */ #endif /* CONFIG_FUNCTION_TRACER */ diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 69149337f2fe..d68033bba223 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -18,10 +18,173 @@ #include #include +#include #include +#include -static unsigned char ftrace_nop[MCOUNT_INSN_SIZE]; + +#ifdef CONFIG_FUNCTION_RET_TRACER + +/* + * These functions are picked from those used on + * this page for dynamic ftrace. They have been + * simplified to ignore all traces in NMI context. + */ +static atomic_t in_nmi; + +void ftrace_nmi_enter(void) +{ + atomic_inc(&in_nmi); +} + +void ftrace_nmi_exit(void) +{ + atomic_dec(&in_nmi); +} + +/* + * Synchronize accesses to return adresses stack with + * interrupts. + */ +static raw_spinlock_t ret_stack_lock; + +/* Add a function return address to the trace stack on thread info.*/ +static int push_return_trace(unsigned long ret, unsigned long long time, + unsigned long func) +{ + int index; + struct thread_info *ti; + unsigned long flags; + int err = 0; + + raw_local_irq_save(flags); + __raw_spin_lock(&ret_stack_lock); + + ti = current_thread_info(); + /* The return trace stack is full */ + if (ti->curr_ret_stack == FTRACE_RET_STACK_SIZE - 1) { + err = -EBUSY; + goto out; + } + + index = ++ti->curr_ret_stack; + ti->ret_stack[index].ret = ret; + ti->ret_stack[index].func = func; + ti->ret_stack[index].calltime = time; + +out: + __raw_spin_unlock(&ret_stack_lock); + raw_local_irq_restore(flags); + return err; +} + +/* Retrieve a function return address to the trace stack on thread info.*/ +static void pop_return_trace(unsigned long *ret, unsigned long long *time, + unsigned long *func) +{ + struct thread_info *ti; + int index; + unsigned long flags; + + raw_local_irq_save(flags); + __raw_spin_lock(&ret_stack_lock); + + ti = current_thread_info(); + index = ti->curr_ret_stack; + *ret = ti->ret_stack[index].ret; + *func = ti->ret_stack[index].func; + *time = ti->ret_stack[index].calltime; + ti->curr_ret_stack--; + + __raw_spin_unlock(&ret_stack_lock); + raw_local_irq_restore(flags); +} + +/* + * Send the trace to the ring-buffer. + * @return the original return address. + */ +unsigned long ftrace_return_to_handler(void) +{ + struct ftrace_retfunc trace; + pop_return_trace(&trace.ret, &trace.calltime, &trace.func); + trace.rettime = cpu_clock(raw_smp_processor_id()); + ftrace_function_return(&trace); + + return trace.ret; +} + +/* + * Hook the return address and push it in the stack of return addrs + * in current thread info. + */ +asmlinkage +void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) +{ + unsigned long old; + unsigned long long calltime; + int faulted; + unsigned long return_hooker = (unsigned long) + &return_to_handler; + + /* Nmi's are currently unsupported */ + if (atomic_read(&in_nmi)) + return; + + /* + * Protect against fault, even if it shouldn't + * happen. This tool is too much intrusive to + * ignore such a protection. + */ + asm volatile( + "1: movl (%[parent_old]), %[old]\n" + "2: movl %[return_hooker], (%[parent_replaced])\n" + " movl $0, %[faulted]\n" + + ".section .fixup, \"ax\"\n" + "3: movl $1, %[faulted]\n" + ".previous\n" + + ".section __ex_table, \"a\"\n" + " .long 1b, 3b\n" + " .long 2b, 3b\n" + ".previous\n" + + : [parent_replaced] "=rm" (parent), [old] "=r" (old), + [faulted] "=r" (faulted) + : [parent_old] "0" (parent), [return_hooker] "r" (return_hooker) + : "memory" + ); + + if (WARN_ON(faulted)) { + unregister_ftrace_return(); + return; + } + + if (WARN_ON(!__kernel_text_address(old))) { + unregister_ftrace_return(); + *parent = old; + return; + } + + calltime = cpu_clock(raw_smp_processor_id()); + + if (push_return_trace(old, calltime, self_addr) == -EBUSY) + *parent = old; +} + +static int __init init_ftrace_function_return(void) +{ + ret_stack_lock = (raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED; + return 0; +} +device_initcall(init_ftrace_function_return); + + +#endif + +#ifdef CONFIG_DYNAMIC_FTRACE union ftrace_code_union { char code[MCOUNT_INSN_SIZE]; @@ -31,17 +194,11 @@ union ftrace_code_union { } __attribute__((packed)); }; - static int ftrace_calc_offset(long ip, long addr) { return (int)(addr - ip); } -unsigned char *ftrace_nop_replace(void) -{ - return ftrace_nop; -} - unsigned char *ftrace_call_replace(unsigned long ip, unsigned long addr) { static union ftrace_code_union calc; @@ -183,6 +340,15 @@ do_ftrace_mod_code(unsigned long ip, void *new_code) } + + +static unsigned char ftrace_nop[MCOUNT_INSN_SIZE]; + +unsigned char *ftrace_nop_replace(void) +{ + return ftrace_nop; +} + int ftrace_modify_code(unsigned long ip, unsigned char *old_code, unsigned char *new_code) @@ -292,3 +458,4 @@ int __init ftrace_dyn_arch_init(void *data) return 0; } +#endif diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 1f5608c11023..dcbbf72a88b1 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -267,6 +267,26 @@ ftrace_init_module(unsigned long *start, unsigned long *end) { } #endif +/* + * Structure that defines a return function trace. + */ +struct ftrace_retfunc { + unsigned long ret; /* Return address */ + unsigned long func; /* Current function */ + unsigned long long calltime; + unsigned long long rettime; +}; + +#ifdef CONFIG_FUNCTION_RET_TRACER +/* Type of a callback handler of tracing return function */ +typedef void (*trace_function_return_t)(struct ftrace_retfunc *); + +extern void register_ftrace_return(trace_function_return_t func); +/* The current handler in use */ +extern trace_function_return_t ftrace_function_return; +extern void unregister_ftrace_return(void); +#endif + /* * Structure which defines the trace of an initcall. * You don't have to fill the func field since it is diff --git a/include/linux/ftrace_irq.h b/include/linux/ftrace_irq.h index b1299d6729f2..0b4df55d7a74 100644 --- a/include/linux/ftrace_irq.h +++ b/include/linux/ftrace_irq.h @@ -2,7 +2,7 @@ #define _LINUX_FTRACE_IRQ_H -#ifdef CONFIG_DYNAMIC_FTRACE +#if defined(CONFIG_DYNAMIC_FTRACE) || defined(CONFIG_FUNCTION_RET_TRACER) extern void ftrace_nmi_enter(void); extern void ftrace_nmi_exit(void); #else diff --git a/include/linux/sched.h b/include/linux/sched.h index 295b7c756ca6..df77abe860c9 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -2005,6 +2005,17 @@ static inline void setup_thread_stack(struct task_struct *p, struct task_struct { *task_thread_info(p) = *task_thread_info(org); task_thread_info(p)->task = p; + +#ifdef CONFIG_FUNCTION_RET_TRACER + /* + * When fork() creates a child process, this function is called. + * But the child task may not inherit the return adresses traced + * by the return function tracer because it will directly execute + * in userspace and will not return to kernel functions its parent + * used. + */ + task_thread_info(p)->curr_ret_stack = -1; +#endif } static inline unsigned long *end_of_stack(struct task_struct *p) diff --git a/kernel/Makefile b/kernel/Makefile index 9a3ec66a9d84..af3be57acbbb 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -23,6 +23,10 @@ CFLAGS_REMOVE_cgroup-debug.o = -pg CFLAGS_REMOVE_sched_clock.o = -pg CFLAGS_REMOVE_sched.o = -mno-spe -pg endif +ifdef CONFIG_FUNCTION_RET_TRACER +CFLAGS_REMOVE_extable.o = -pg # For __kernel_text_address() +CFLAGS_REMOVE_module.o = -pg # For __module_text_address() +endif obj-$(CONFIG_FREEZER) += freezer.o obj-$(CONFIG_PROFILING) += profile.o -- cgit v1.2.3 From f1c4be5edad3756212cbbbeab39428fe90c27109 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Tue, 11 Nov 2008 10:22:36 +0100 Subject: tracing, x86: clean up FUNCTION_RET_TRACER Kconfig Impact: cleanup move FUNCTION_RET_TRACER to the X86 select section, where we have all the other options. Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index ca91e50bdb10..0de793cf2144 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -11,7 +11,6 @@ config 64BIT config X86_32 def_bool !64BIT - select HAVE_FUNCTION_RET_TRACER config X86_64 def_bool 64BIT @@ -30,6 +29,7 @@ config X86 select HAVE_FTRACE_MCOUNT_RECORD select HAVE_DYNAMIC_FTRACE select HAVE_FUNCTION_TRACER + select HAVE_FUNCTION_RET_TRACER if X86_32 select HAVE_FUNCTION_TRACE_MCOUNT_TEST select HAVE_KVM if ((X86_32 && !X86_VOYAGER && !X86_VISWS && !X86_NUMAQ) || X86_64) select HAVE_ARCH_KGDB if !X86_VOYAGER -- cgit v1.2.3 From 867f7fb3ebb831970847b179e7df5a9ab10da16d Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Tue, 11 Nov 2008 11:18:14 +0100 Subject: tracing, x86: function return tracer, fix assembly constraints fix: arch/x86/kernel/ftrace.c: Assembler messages: arch/x86/kernel/ftrace.c:140: Error: missing ')' arch/x86/kernel/ftrace.c:140: Error: junk `(%ebp))' after expression arch/x86/kernel/ftrace.c:141: Error: missing ')' arch/x86/kernel/ftrace.c:141: Error: junk `(%ebp))' after expression the [parent_replaced] is used in an =rm fashion, so that constraint is correct in isolation - but [parent_old] aliases register %0 and uses it in an addressing mode that is only valid with registers - so change the constraint from =rm to =r. This fixes the build failure. Signed-off-by: Ingo Molnar --- arch/x86/kernel/ftrace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index d68033bba223..9b2325a4d53c 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -151,7 +151,7 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) " .long 2b, 3b\n" ".previous\n" - : [parent_replaced] "=rm" (parent), [old] "=r" (old), + : [parent_replaced] "=r" (parent), [old] "=r" (old), [faulted] "=r" (faulted) : [parent_old] "0" (parent), [return_hooker] "r" (return_hooker) : "memory" -- cgit v1.2.3 From a3d732f93785da17e0137210deadb4616f5536fc Mon Sep 17 00:00:00 2001 From: Cliff Wickman Date: Mon, 10 Nov 2008 16:16:31 -0600 Subject: x86, UV: fix redundant creation of sgi_uv Impact: fix double entry creation in /proc There is a collision between two UV functions: both uv_ptc_init() and gru_proc_init() try to make /proc/sgi_uv So move it's creation to a single place: uv_system_init() Signed-off-by: Cliff Wickman Signed-off-by: Ingo Molnar --- arch/x86/kernel/genx2apic_uv_x.c | 2 ++ arch/x86/kernel/tlb_uv.c | 4 ---- drivers/misc/sgi-gru/gruprocfs.c | 1 - 3 files changed, 2 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/genx2apic_uv_x.c b/arch/x86/kernel/genx2apic_uv_x.c index 85fb7dd48f67..d7213a1cb784 100644 --- a/arch/x86/kernel/genx2apic_uv_x.c +++ b/arch/x86/kernel/genx2apic_uv_x.c @@ -20,6 +20,7 @@ #include #include #include +#include #include #include #include @@ -570,4 +571,5 @@ void __init uv_system_init(void) uv_cpu_init(); uv_scir_register_cpu_notifier(); + proc_mkdir("sgi_uv", NULL); } diff --git a/arch/x86/kernel/tlb_uv.c b/arch/x86/kernel/tlb_uv.c index 04431f34fd16..6a00e5faaa74 100644 --- a/arch/x86/kernel/tlb_uv.c +++ b/arch/x86/kernel/tlb_uv.c @@ -566,14 +566,10 @@ static int __init uv_ptc_init(void) if (!is_uv_system()) return 0; - if (!proc_mkdir("sgi_uv", NULL)) - return -EINVAL; - proc_uv_ptc = create_proc_entry(UV_PTC_BASENAME, 0444, NULL); if (!proc_uv_ptc) { printk(KERN_ERR "unable to create %s proc entry\n", UV_PTC_BASENAME); - remove_proc_entry("sgi_uv", NULL); return -EINVAL; } proc_uv_ptc->proc_fops = &proc_uv_ptc_operations; diff --git a/drivers/misc/sgi-gru/gruprocfs.c b/drivers/misc/sgi-gru/gruprocfs.c index 533923f83f1a..73b0ca061bb5 100644 --- a/drivers/misc/sgi-gru/gruprocfs.c +++ b/drivers/misc/sgi-gru/gruprocfs.c @@ -317,7 +317,6 @@ int gru_proc_init(void) { struct proc_entry *p; - proc_mkdir("sgi_uv", NULL); proc_gru = proc_mkdir("sgi_uv/gru", NULL); for (p = proc_files; p->name; p++) -- cgit v1.2.3 From 19b3e9671c5a219b8c34da2cc66e0ce7c3a501ae Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Tue, 11 Nov 2008 11:57:02 +0100 Subject: tracing: function return tracer, build fix fix: arch/x86/kernel/ftrace.c: In function 'ftrace_return_to_handler': arch/x86/kernel/ftrace.c:112: error: implicit declaration of function 'cpu_clock' cpu_clock() is implicitly included via a number of ways, but its real location is sched.h. (Build failure is triggerable if enough other kernel components are turned off.) Signed-off-by: Ingo Molnar --- arch/x86/kernel/ftrace.c | 1 + 1 file changed, 1 insertion(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 9b2325a4d53c..16a571dea2ef 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -14,6 +14,7 @@ #include #include #include +#include #include #include -- cgit v1.2.3 From d3ec5cae0921611ceae06464ef6291012dd9849f Mon Sep 17 00:00:00 2001 From: Ivan Vecera Date: Tue, 11 Nov 2008 14:33:44 +0100 Subject: x86: call machine_shutdown and stop all CPUs in native_machine_halt Impact: really halt all CPUs on halt Function machine_halt (resp. native_machine_halt) is empty for x86 architectures. When command 'halt -f' is invoked, the message "System halted." is displayed but this is not really true because all CPUs are still running. There are also similar inconsistencies for other arches (some uses power-off for halt or forever-loop with IRQs enabled/disabled). IMO there should be used the same approach for all architectures OR what does the message "System halted" really mean? This patch fixes it for x86. Signed-off-by: Ivan Vecera Signed-off-by: Ingo Molnar --- arch/x86/include/asm/apic.h | 1 + arch/x86/include/asm/system.h | 2 ++ arch/x86/kernel/process.c | 16 ++++++++++++++++ arch/x86/kernel/reboot.c | 5 +++++ arch/x86/kernel/smp.c | 13 ------------- 5 files changed, 24 insertions(+), 13 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/apic.h b/arch/x86/include/asm/apic.h index 3b1510b4fc57..25caa0738af5 100644 --- a/arch/x86/include/asm/apic.h +++ b/arch/x86/include/asm/apic.h @@ -193,6 +193,7 @@ extern u8 setup_APIC_eilvt_ibs(u8 vector, u8 msg_type, u8 mask); static inline void lapic_shutdown(void) { } #define local_apic_timer_c2_ok 1 static inline void init_apic_mappings(void) { } +static inline void disable_local_APIC(void) { } #endif /* !CONFIG_X86_LOCAL_APIC */ diff --git a/arch/x86/include/asm/system.h b/arch/x86/include/asm/system.h index 2ed3f0f44ff7..07c3e4048991 100644 --- a/arch/x86/include/asm/system.h +++ b/arch/x86/include/asm/system.h @@ -314,6 +314,8 @@ extern void free_init_pages(char *what, unsigned long begin, unsigned long end); void default_idle(void); +void stop_this_cpu(void *dummy); + /* * Force strict CPU ordering. * And yes, this is required on UP too when we're talking diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c index c622772744d8..a4da7c4b3129 100644 --- a/arch/x86/kernel/process.c +++ b/arch/x86/kernel/process.c @@ -8,6 +8,7 @@ #include #include #include +#include unsigned long idle_halt; EXPORT_SYMBOL(idle_halt); @@ -122,6 +123,21 @@ void default_idle(void) EXPORT_SYMBOL(default_idle); #endif +void stop_this_cpu(void *dummy) +{ + local_irq_disable(); + /* + * Remove this CPU: + */ + cpu_clear(smp_processor_id(), cpu_online_map); + disable_local_APIC(); + + for (;;) { + if (hlt_works(smp_processor_id())) + halt(); + } +} + static void do_nothing(void *unused) { } diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c index 724adfc63cb9..34f8d37ae3c5 100644 --- a/arch/x86/kernel/reboot.c +++ b/arch/x86/kernel/reboot.c @@ -461,6 +461,11 @@ static void native_machine_restart(char *__unused) static void native_machine_halt(void) { + /* stop other cpus and apics */ + machine_shutdown(); + + /* stop this cpu */ + stop_this_cpu(NULL); } static void native_machine_power_off(void) diff --git a/arch/x86/kernel/smp.c b/arch/x86/kernel/smp.c index 18f9b19f5f8f..3f92b134ab90 100644 --- a/arch/x86/kernel/smp.c +++ b/arch/x86/kernel/smp.c @@ -140,19 +140,6 @@ void native_send_call_func_ipi(cpumask_t mask) send_IPI_mask(mask, CALL_FUNCTION_VECTOR); } -static void stop_this_cpu(void *dummy) -{ - local_irq_disable(); - /* - * Remove this CPU: - */ - cpu_clear(smp_processor_id(), cpu_online_map); - disable_local_APIC(); - if (hlt_works(smp_processor_id())) - for (;;) halt(); - for (;;); -} - /* * this function calls the 'stop' function on all other CPUs in the system. */ -- cgit v1.2.3 From a98f8fd24fb24fcb9a359553e64dd6aac5cf4279 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Thu, 6 Nov 2008 01:13:39 +0100 Subject: x86: apic reset counter on shutdown Impact: avoid spurious lapic timer events on shutdown The apic timer might be close to firing when it is shutdown. We can not really disable the timer - we just mask the interrupt. That way we can get an extra interrupt when it is reenabled. Set the counter to max on shutdown to avoid this. Signed-off-by: Thomas Gleixner Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic.c | 1 + 1 file changed, 1 insertion(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/apic.c b/arch/x86/kernel/apic.c index 70879c9e3936..1d410ee4b064 100644 --- a/arch/x86/kernel/apic.c +++ b/arch/x86/kernel/apic.c @@ -441,6 +441,7 @@ static void lapic_timer_setup(enum clock_event_mode mode, v = apic_read(APIC_LVTT); v |= (APIC_LVT_MASKED | LOCAL_TIMER_VECTOR); apic_write(APIC_LVTT, v); + apic_write(APIC_TMICT, 0xffffffff); break; case CLOCK_EVT_MODE_RESUME: /* Nothing to do here */ -- cgit v1.2.3 From c280ea5e4c6ba0b38ed6b005150fe16a660e903b Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Sat, 8 Nov 2008 13:29:45 +0100 Subject: x86: fix documentation typo in arch/x86/Kconfig Impact: documentation update Chris Snook pointed out that it's Core i7, not Core 7i. Reported-by: Chris Snook Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 38ae04bf6514..bacac556b189 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -963,7 +963,7 @@ config NUMA local memory controller of the CPU and add some more NUMA awareness to the kernel. - For 64-bit this is recommended if the system is Intel Core 7i + For 64-bit this is recommended if the system is Intel Core i7 (or later), AMD Opteron, or EM64T NUMA. For 32-bit this is only needed on (rare) 32-bit-only platforms -- cgit v1.2.3 From 4687518c4cb7807fbeff21770e309080f9eb7f2f Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Tue, 11 Nov 2008 13:03:07 -0800 Subject: x86: 32 bit: interrupt stub consistency with 64 bit Don't generate interrupt stubs for interrupt vectors below FIRST_EXTERNAL_VECTOR, and make the table of interrupt vectors (interrupt[]) __initconst. Both of these changes both conserve memory and improve consistency with 64 bits. Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/hw_irq.h | 2 +- arch/x86/kernel/entry_32.S | 6 +++--- arch/x86/kernel/irqinit_32.c | 2 +- arch/x86/lguest/boot.c | 3 ++- 4 files changed, 7 insertions(+), 6 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/hw_irq.h b/arch/x86/include/asm/hw_irq.h index b97aecb0b61d..27d33f92afe2 100644 --- a/arch/x86/include/asm/hw_irq.h +++ b/arch/x86/include/asm/hw_irq.h @@ -110,7 +110,7 @@ extern asmlinkage void smp_invalidate_interrupt(struct pt_regs *); #endif #ifdef CONFIG_X86_32 -extern void (*const interrupt[NR_VECTORS])(void); +extern void (*__initconst interrupt[NR_VECTORS-FIRST_EXTERNAL_VECTOR])(void); #endif typedef int vector_irq_t[NR_VECTORS]; diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S index 28b597ef9ca1..4aea95652cff 100644 --- a/arch/x86/kernel/entry_32.S +++ b/arch/x86/kernel/entry_32.S @@ -622,16 +622,16 @@ END(syscall_badsys) * Build the entry stubs and pointer table with * some assembler magic. */ -.section .rodata,"a" +.section .init.rodata,"a" ENTRY(interrupt) .text ENTRY(irq_entries_start) RING0_INT_FRAME -vector=0 +vector=FIRST_EXTERNAL_VECTOR .rept NR_VECTORS ALIGN - .if vector + .if vector != FIRST_EXTERNAL_VECTOR CFI_ADJUST_CFA_OFFSET -4 .endif 1: pushl $~(vector) diff --git a/arch/x86/kernel/irqinit_32.c b/arch/x86/kernel/irqinit_32.c index 845aa9803e80..607db63044a5 100644 --- a/arch/x86/kernel/irqinit_32.c +++ b/arch/x86/kernel/irqinit_32.c @@ -129,7 +129,7 @@ void __init native_init_IRQ(void) for (i = FIRST_EXTERNAL_VECTOR; i < NR_VECTORS; i++) { /* SYSCALL_VECTOR was reserved in trap_init. */ if (i != SYSCALL_VECTOR) - set_intr_gate(i, interrupt[i]); + set_intr_gate(i, interrupt[i-FIRST_EXTERNAL_VECTOR]); } diff --git a/arch/x86/lguest/boot.c b/arch/x86/lguest/boot.c index a5d8e1ace1cf..50a779264bb1 100644 --- a/arch/x86/lguest/boot.c +++ b/arch/x86/lguest/boot.c @@ -590,7 +590,8 @@ static void __init lguest_init_IRQ(void) * a straightforward 1 to 1 mapping, so force that here. */ __get_cpu_var(vector_irq)[vector] = i; if (vector != SYSCALL_VECTOR) { - set_intr_gate(vector, interrupt[vector]); + set_intr_gate(vector, + interrupt[vector-FIRST_EXTERNAL_VECTOR]); set_irq_chip_and_handler_name(i, &lguest_irq_controller, handle_level_irq, "level"); -- cgit v1.2.3 From b7c6244f13d37592003b46e12500a90e9781ad9d Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Tue, 11 Nov 2008 13:24:58 -0800 Subject: x86: 32 bits: shrink and align IRQ stubs Shrink the IRQ stubs on 32 bits down to just over four bytes per (we fit seven into a 32-byte chunk.) This shrinks the total icache consumption of the IRQ stubs down to an even kilobyte, if all of them are in active use. The downside is that we end up with a double jump, which could have a negative effect on some pipelines. The double jump is always inside the same cacheline on any modern chips (the exception being 486/Elan/Geode which have only 16-byte cachelines, but are unlikely to have too many interrupt sources.) To get the most effect, cache-align the IRQ stubs. Signed-off-by: H. Peter Anvin --- arch/x86/kernel/entry_32.S | 34 ++++++++++++++++++++++------------ 1 file changed, 22 insertions(+), 12 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S index 4aea95652cff..dae81b9fd451 100644 --- a/arch/x86/kernel/entry_32.S +++ b/arch/x86/kernel/entry_32.S @@ -619,28 +619,37 @@ END(syscall_badsys) 27:; /* - * Build the entry stubs and pointer table with - * some assembler magic. + * Build the entry stubs and pointer table with some assembler magic. + * We pack 7 stubs into a single 32-byte chunk, which will fit in a + * single cache line on all modern x86 implementations. */ .section .init.rodata,"a" ENTRY(interrupt) .text - + .p2align 5 + .p2align CONFIG_X86_L1_CACHE_SHIFT ENTRY(irq_entries_start) RING0_INT_FRAME vector=FIRST_EXTERNAL_VECTOR -.rept NR_VECTORS - ALIGN - .if vector != FIRST_EXTERNAL_VECTOR +.rept (NR_VECTORS-FIRST_EXTERNAL_VECTOR+6)/7 + .balign 32 + .rept 7 + .if vector < NR_VECTORS + .if vector != FIRST_EXTERNAL_VECTOR CFI_ADJUST_CFA_OFFSET -4 - .endif -1: pushl $~(vector) + .endif +1: pushl $(~vector+0x80) /* Note: always in signed byte range */ CFI_ADJUST_CFA_OFFSET 4 - jmp common_interrupt - .previous + .if ((vector-FIRST_EXTERNAL_VECTOR)%7) != 6 + jmp 2f + .endif + .previous .long 1b - .text + .text vector=vector+1 + .endif + .endr +2: jmp common_interrupt .endr END(irq_entries_start) @@ -652,8 +661,9 @@ END(interrupt) * the CPU automatically disables interrupts when executing an IRQ vector, * so IRQ-flags tracing has to follow that: */ - ALIGN + .p2align CONFIG_X86_L1_CACHE_SHIFT common_interrupt: + addl $-0x80,(%esp) /* Adjust vector into the [-256,-1] range */ SAVE_ALL TRACE_IRQS_OFF movl %esp,%eax -- cgit v1.2.3 From 939b787130bf22887a09d8fd2641a094dcef8c22 Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Tue, 11 Nov 2008 13:51:52 -0800 Subject: x86: 64 bits: shrink and align IRQ stubs Move the IRQ stub generation to assembly to simplify it and for consistency with 32 bits. Doing it in a C file with asm() statements doesn't help clarity, and it prevents some optimizations. Shrink the IRQ stubs down to just over four bytes per (we fit seven into a 32-byte chunk.) This shrinks the total icache consumption of the IRQ stubs down to an even kilobyte, if all of them are in active use. The downside is that we end up with a double jump, which could have a negative effect on some pipelines. The double jump is always inside the same cacheline on any modern chips. To get the most effect, cache-align the IRQ stubs. This makes the 64-bit code match changes already done to the 32-bit code, and should open up irqinit*.c for unification. Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/hw_irq.h | 2 -- arch/x86/kernel/entry_64.S | 48 +++++++++++++++++++++++++++++-- arch/x86/kernel/irqinit_64.c | 66 ------------------------------------------- 3 files changed, 45 insertions(+), 71 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/hw_irq.h b/arch/x86/include/asm/hw_irq.h index 27d33f92afe2..8de644b6b959 100644 --- a/arch/x86/include/asm/hw_irq.h +++ b/arch/x86/include/asm/hw_irq.h @@ -109,9 +109,7 @@ extern asmlinkage void smp_invalidate_interrupt(struct pt_regs *); #endif #endif -#ifdef CONFIG_X86_32 extern void (*__initconst interrupt[NR_VECTORS-FIRST_EXTERNAL_VECTOR])(void); -#endif typedef int vector_irq_t[NR_VECTORS]; DECLARE_PER_CPU(vector_irq_t, vector_irq); diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index b86f332c96a6..9b2aeaac9a6b 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -627,6 +627,46 @@ END(stub_rt_sigreturn) vector already pushed) */ #define XCPT_FRAME _frame ORIG_RAX +/* + * Build the entry stubs and pointer table with some assembler magic. + * We pack 7 stubs into a single 32-byte chunk, which will fit in a + * single cache line on all modern x86 implementations. + */ + .section .init.rodata,"a" +ENTRY(interrupt) + .text + .p2align 5 + .p2align CONFIG_X86_L1_CACHE_SHIFT +ENTRY(irq_entries_start) + INTR_FRAME +vector=FIRST_EXTERNAL_VECTOR +.rept (NR_VECTORS-FIRST_EXTERNAL_VECTOR+6)/7 + .balign 32 + .rept 7 + .if vector < NR_VECTORS + .if vector != FIRST_EXTERNAL_VECTOR + CFI_ADJUST_CFA_OFFSET -8 + .endif +1: pushq $(~vector+0x80) /* Note: always in signed byte range */ + CFI_ADJUST_CFA_OFFSET 8 + .if ((vector-FIRST_EXTERNAL_VECTOR)%7) != 6 + jmp 2f + .endif + .previous + .quad 1b + .text +vector=vector+1 + .endif + .endr +2: jmp common_interrupt +.endr + CFI_ENDPROC +END(irq_entries_start) + +.previous +END(interrupt) +.previous + /* * Interrupt entry/exit. * @@ -635,11 +675,12 @@ END(stub_rt_sigreturn) * Entry runs with interrupts off. */ -/* 0(%rsp): interrupt number */ +/* 0(%rsp): ~(interrupt number)+0x80 */ .macro interrupt func + addq $-0x80,(%rsp) /* Adjust vector to [-256,-1] range */ cld SAVE_ARGS - leaq -ARGOFFSET(%rsp),%rdi # arg1 for handler + leaq -ARGOFFSET(%rsp),%rdi /* arg1 for handler */ pushq %rbp /* * Save rbp twice: One is for marking the stack frame, as usual, and the @@ -670,7 +711,8 @@ END(stub_rt_sigreturn) call \func .endm -ENTRY(common_interrupt) + .p2align CONFIG_X86_L1_CACHE_SHIFT +common_interrupt: XCPT_FRAME interrupt do_IRQ /* 0(%rsp): oldrsp-ARGOFFSET */ diff --git a/arch/x86/kernel/irqinit_64.c b/arch/x86/kernel/irqinit_64.c index ff0235391285..8670b3ce626e 100644 --- a/arch/x86/kernel/irqinit_64.c +++ b/arch/x86/kernel/irqinit_64.c @@ -23,41 +23,6 @@ #include #include -/* - * Common place to define all x86 IRQ vectors - * - * This builds up the IRQ handler stubs using some ugly macros in irq.h - * - * These macros create the low-level assembly IRQ routines that save - * register context and call do_IRQ(). do_IRQ() then does all the - * operations that are needed to keep the AT (or SMP IOAPIC) - * interrupt-controller happy. - */ - -#define IRQ_NAME2(nr) nr##_interrupt(void) -#define IRQ_NAME(nr) IRQ_NAME2(IRQ##nr) - -/* - * SMP has a few special interrupts for IPI messages - */ - -#define BUILD_IRQ(nr) \ - asmlinkage void IRQ_NAME(nr); \ - asm("\n.text\n.p2align\n" \ - "IRQ" #nr "_interrupt:\n\t" \ - "push $~(" #nr ") ; " \ - "jmp common_interrupt\n" \ - ".previous"); - -#define BI(x,y) \ - BUILD_IRQ(x##y) - -#define BUILD_16_IRQS(x) \ - BI(x,0) BI(x,1) BI(x,2) BI(x,3) \ - BI(x,4) BI(x,5) BI(x,6) BI(x,7) \ - BI(x,8) BI(x,9) BI(x,a) BI(x,b) \ - BI(x,c) BI(x,d) BI(x,e) BI(x,f) - /* * ISA PIC or low IO-APIC triggered (INTA-cycle or APIC) interrupts: * (these are usually mapped to vectors 0x30-0x3f) @@ -73,37 +38,6 @@ * * (these are usually mapped into the 0x30-0xff vector range) */ - BUILD_16_IRQS(0x2) BUILD_16_IRQS(0x3) -BUILD_16_IRQS(0x4) BUILD_16_IRQS(0x5) BUILD_16_IRQS(0x6) BUILD_16_IRQS(0x7) -BUILD_16_IRQS(0x8) BUILD_16_IRQS(0x9) BUILD_16_IRQS(0xa) BUILD_16_IRQS(0xb) -BUILD_16_IRQS(0xc) BUILD_16_IRQS(0xd) BUILD_16_IRQS(0xe) BUILD_16_IRQS(0xf) - -#undef BUILD_16_IRQS -#undef BI - - -#define IRQ(x,y) \ - IRQ##x##y##_interrupt - -#define IRQLIST_16(x) \ - IRQ(x,0), IRQ(x,1), IRQ(x,2), IRQ(x,3), \ - IRQ(x,4), IRQ(x,5), IRQ(x,6), IRQ(x,7), \ - IRQ(x,8), IRQ(x,9), IRQ(x,a), IRQ(x,b), \ - IRQ(x,c), IRQ(x,d), IRQ(x,e), IRQ(x,f) - -/* for the irq vectors */ -static void (*__initdata interrupt[NR_VECTORS - FIRST_EXTERNAL_VECTOR])(void) = { - IRQLIST_16(0x2), IRQLIST_16(0x3), - IRQLIST_16(0x4), IRQLIST_16(0x5), IRQLIST_16(0x6), IRQLIST_16(0x7), - IRQLIST_16(0x8), IRQLIST_16(0x9), IRQLIST_16(0xa), IRQLIST_16(0xb), - IRQLIST_16(0xc), IRQLIST_16(0xd), IRQLIST_16(0xe), IRQLIST_16(0xf) -}; - -#undef IRQ -#undef IRQLIST_16 - - - /* * IRQ2 is cascade interrupt to second interrupt controller -- cgit v1.2.3 From 14d7ca5c575853664d8fe4f225a77b8df1b7de7d Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Tue, 11 Nov 2008 16:19:48 -0800 Subject: x86: attempt reboot via port CF9 if we have standard PCI ports Impact: Changes reboot behavior. If port CF9 seems to be safe to touch, attempt it before trying the keyboard controller. Port CF9 is not available on all chipsets (a significant but decreasing number of modern chipsets don't implement it), but port CF9 itself should in general be safe to poke (no ill effects if unimplemented) on any system which has PCI Configuration Method #1 or #2, as it falls inside the PCI configuration port range in both cases. No chipset without PCI is known to have port CF9, either, although an explicit "pci=bios" would mean we miss this and therefore don't use port CF9. An explicit "reboot=pci" can be used to force the use of port CF9. Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/emergency-restart.h | 4 +++- arch/x86/kernel/reboot.c | 34 +++++++++++++++++++++++++------- arch/x86/pci/direct.c | 4 +++- arch/x86/pci/pci.h | 1 + 4 files changed, 34 insertions(+), 9 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/emergency-restart.h b/arch/x86/include/asm/emergency-restart.h index 94826cf87455..cc70c1c78ca4 100644 --- a/arch/x86/include/asm/emergency-restart.h +++ b/arch/x86/include/asm/emergency-restart.h @@ -8,7 +8,9 @@ enum reboot_type { BOOT_BIOS = 'b', #endif BOOT_ACPI = 'a', - BOOT_EFI = 'e' + BOOT_EFI = 'e', + BOOT_CF9 = 'p', + BOOT_CF9_COND = 'q', }; extern enum reboot_type reboot_type; diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c index 34f8d37ae3c5..ddc93891cdcd 100644 --- a/arch/x86/kernel/reboot.c +++ b/arch/x86/kernel/reboot.c @@ -29,14 +29,17 @@ EXPORT_SYMBOL(pm_power_off); static const struct desc_ptr no_idt = {}; static int reboot_mode; -enum reboot_type reboot_type = BOOT_KBD; +enum reboot_type reboot_type = BOOT_CF9_COND; int reboot_force; #if defined(CONFIG_X86_32) && defined(CONFIG_SMP) static int reboot_cpu = -1; #endif -/* reboot=b[ios] | s[mp] | t[riple] | k[bd] | e[fi] [, [w]arm | [c]old] +/* This is set by the PCI code if either type 1 or type 2 PCI is detected */ +bool port_cf9_safe = false; + +/* reboot=b[ios] | s[mp] | t[riple] | k[bd] | e[fi] [, [w]arm | [c]old] | p[ci] warm Don't set the cold reboot flag cold Set the cold reboot flag bios Reboot by jumping through the BIOS (only for X86_32) @@ -45,6 +48,7 @@ static int reboot_cpu = -1; kbd Use the keyboard controller. cold reset (default) acpi Use the RESET_REG in the FADT efi Use efi reset_system runtime service + pci Use the so-called "PCI reset register", CF9 force Avoid anything that could hang. */ static int __init reboot_setup(char *str) @@ -79,6 +83,7 @@ static int __init reboot_setup(char *str) case 'k': case 't': case 'e': + case 'p': reboot_type = *str; break; @@ -379,28 +384,43 @@ static void native_machine_emergency_restart(void) load_idt(&no_idt); __asm__ __volatile__("int3"); - reboot_type = BOOT_KBD; + reboot_type = BOOT_CF9_COND; break; #ifdef CONFIG_X86_32 case BOOT_BIOS: machine_real_restart(jump_to_bios, sizeof(jump_to_bios)); - reboot_type = BOOT_KBD; + reboot_type = BOOT_CF9_COND; break; #endif case BOOT_ACPI: acpi_reboot(); - reboot_type = BOOT_KBD; + reboot_type = BOOT_CF9_COND; break; - case BOOT_EFI: if (efi_enabled) - efi.reset_system(reboot_mode ? EFI_RESET_WARM : EFI_RESET_COLD, + efi.reset_system(reboot_mode ? + EFI_RESET_WARM : + EFI_RESET_COLD, EFI_SUCCESS, 0, NULL); + reboot_type = BOOT_CF9_COND; + break; + + case BOOT_CF9: + port_cf9_safe = true; + /* fall through */ + case BOOT_CF9_COND: + if (port_cf9_safe) { + u8 cf9 = inb(0xcf9) & ~6; + outb(cf9|2, 0xcf9); /* Request hard reset */ + udelay(50); + outb(cf9|6, 0xcf9); /* Actually do the reset */ + udelay(50); + } reboot_type = BOOT_KBD; break; } diff --git a/arch/x86/pci/direct.c b/arch/x86/pci/direct.c index 9915293500fb..9a5af6c8fbe9 100644 --- a/arch/x86/pci/direct.c +++ b/arch/x86/pci/direct.c @@ -173,7 +173,7 @@ static int pci_conf2_write(unsigned int seg, unsigned int bus, #undef PCI_CONF2_ADDRESS -static struct pci_raw_ops pci_direct_conf2 = { +struct pci_raw_ops pci_direct_conf2 = { .read = pci_conf2_read, .write = pci_conf2_write, }; @@ -289,6 +289,7 @@ int __init pci_direct_probe(void) if (pci_check_type1()) { raw_pci_ops = &pci_direct_conf1; + port_cf9_safe = true; return 1; } release_resource(region); @@ -305,6 +306,7 @@ int __init pci_direct_probe(void) if (pci_check_type2()) { raw_pci_ops = &pci_direct_conf2; + port_cf9_safe = true; return 2; } diff --git a/arch/x86/pci/pci.h b/arch/x86/pci/pci.h index 15b9cf6be729..1959018aac02 100644 --- a/arch/x86/pci/pci.h +++ b/arch/x86/pci/pci.h @@ -96,6 +96,7 @@ extern struct pci_raw_ops *raw_pci_ops; extern struct pci_raw_ops *raw_pci_ext_ops; extern struct pci_raw_ops pci_direct_conf1; +extern bool port_cf9_safe; /* arch_initcall level */ extern int pci_direct_probe(void); -- cgit v1.2.3 From 1f0d69a9fc815db82f15722bf05227190b1d714d Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 12 Nov 2008 00:14:39 -0500 Subject: tracing: profile likely and unlikely annotations Impact: new unlikely/likely profiler Andrew Morton recently suggested having an in-kernel way to profile likely and unlikely macros. This patch achieves that goal. When configured, every(*) likely and unlikely macro gets a counter attached to it. When the condition is hit, the hit and misses of that condition are recorded. These numbers can later be retrieved by: /debugfs/tracing/profile_likely - All likely markers /debugfs/tracing/profile_unlikely - All unlikely markers. # cat /debug/tracing/profile_unlikely | head correct incorrect % Function File Line ------- --------- - -------- ---- ---- 2167 0 0 do_arch_prctl process_64.c 832 0 0 0 do_arch_prctl process_64.c 804 2670 0 0 IS_ERR err.h 34 71230 5693 7 __switch_to process_64.c 673 76919 0 0 __switch_to process_64.c 639 43184 33743 43 __switch_to process_64.c 624 12740 64181 83 __switch_to process_64.c 594 12740 64174 83 __switch_to process_64.c 590 # cat /debug/tracing/profile_unlikely | \ awk '{ if ($3 > 25) print $0; }' |head -20 44963 35259 43 __switch_to process_64.c 624 12762 67454 84 __switch_to process_64.c 594 12762 67447 84 __switch_to process_64.c 590 1478 595 28 syscall_get_error syscall.h 51 0 2821 100 syscall_trace_leave ptrace.c 1567 0 1 100 native_smp_prepare_cpus smpboot.c 1237 86338 265881 75 calc_delta_fair sched_fair.c 408 210410 108540 34 calc_delta_mine sched.c 1267 0 54550 100 sched_info_queued sched_stats.h 222 51899 66435 56 pick_next_task_fair sched_fair.c 1422 6 10 62 yield_task_fair sched_fair.c 982 7325 2692 26 rt_policy sched.c 144 0 1270 100 pre_schedule_rt sched_rt.c 1261 1268 48073 97 pick_next_task_rt sched_rt.c 884 0 45181 100 sched_info_dequeued sched_stats.h 177 0 15 100 sched_move_task sched.c 8700 0 15 100 sched_move_task sched.c 8690 53167 33217 38 schedule sched.c 4457 0 80208 100 sched_info_switch sched_stats.h 270 30585 49631 61 context_switch sched.c 2619 # cat /debug/tracing/profile_likely | awk '{ if ($3 > 25) print $0; }' 39900 36577 47 pick_next_task sched.c 4397 20824 15233 42 switch_mm mmu_context_64.h 18 0 7 100 __cancel_work_timer workqueue.c 560 617 66484 99 clocksource_adjust timekeeping.c 456 0 346340 100 audit_syscall_exit auditsc.c 1570 38 347350 99 audit_get_context auditsc.c 732 0 345244 100 audit_syscall_entry auditsc.c 1541 38 1017 96 audit_free auditsc.c 1446 0 1090 100 audit_alloc auditsc.c 862 2618 1090 29 audit_alloc auditsc.c 858 0 6 100 move_masked_irq migration.c 9 1 198 99 probe_sched_wakeup trace_sched_switch.c 58 2 2 50 probe_wakeup trace_sched_wakeup.c 227 0 2 100 probe_wakeup_sched_switch trace_sched_wakeup.c 144 4514 2090 31 __grab_cache_page filemap.c 2149 12882 228786 94 mapping_unevictable pagemap.h 50 4 11 73 __flush_cpu_slab slub.c 1466 627757 330451 34 slab_free slub.c 1731 2959 61245 95 dentry_lru_del_init dcache.c 153 946 1217 56 load_elf_binary binfmt_elf.c 904 102 82 44 disk_put_part genhd.h 206 1 1 50 dst_gc_task dst.c 82 0 19 100 tcp_mss_split_point tcp_output.c 1126 As you can see by the above, there's a bit of work to do in rethinking the use of some unlikelys and likelys. Note: the unlikely case had 71 hits that were more than 25%. Note: After submitting my first version of this patch, Andrew Morton showed me a version written by Daniel Walker, where I picked up the following ideas from: 1) Using __builtin_constant_p to avoid profiling fixed values. 2) Using __FILE__ instead of instruction pointers. 3) Using the preprocessor to stop all profiling of likely annotations from vsyscall_64.c. Thanks to Andrew Morton, Arjan van de Ven, Theodore Tso and Ingo Molnar for their feed back on this patch. (*) Not ever unlikely is recorded, those that are used by vsyscalls (a few of them) had to have profiling disabled. Signed-off-by: Steven Rostedt Cc: Andrew Morton Cc: Frederic Weisbecker Cc: Theodore Tso Cc: Arjan van de Ven Cc: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/x86/kernel/vsyscall_64.c | 8 ++ include/asm-generic/vmlinux.lds.h | 14 +++- include/linux/compiler.h | 61 +++++++++++++- kernel/trace/Kconfig | 16 ++++ kernel/trace/Makefile | 1 + kernel/trace/trace_unlikely.c | 164 ++++++++++++++++++++++++++++++++++++++ 6 files changed, 261 insertions(+), 3 deletions(-) create mode 100644 kernel/trace/trace_unlikely.c (limited to 'arch/x86') diff --git a/arch/x86/kernel/vsyscall_64.c b/arch/x86/kernel/vsyscall_64.c index 0b8b6690a86d..2f90202e59b3 100644 --- a/arch/x86/kernel/vsyscall_64.c +++ b/arch/x86/kernel/vsyscall_64.c @@ -17,6 +17,14 @@ * want per guest time just set the kernel.vsyscall64 sysctl to 0. */ +/* Protect userspace from profiling */ +#ifdef CONFIG_TRACE_UNLIKELY_PROFILE +# undef likely +# undef unlikely +# define likely(x) likely_notrace(x) +# define unlikely(x) unlikely_notrace(x) +#endif + #include #include #include diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h index 80744606bad1..e10beb5335c9 100644 --- a/include/asm-generic/vmlinux.lds.h +++ b/include/asm-generic/vmlinux.lds.h @@ -45,6 +45,17 @@ #define MCOUNT_REC() #endif +#ifdef CONFIG_TRACE_UNLIKELY_PROFILE +#define LIKELY_PROFILE() VMLINUX_SYMBOL(__start_likely_profile) = .; \ + *(_ftrace_likely) \ + VMLINUX_SYMBOL(__stop_likely_profile) = .; \ + VMLINUX_SYMBOL(__start_unlikely_profile) = .; \ + *(_ftrace_unlikely) \ + VMLINUX_SYMBOL(__stop_unlikely_profile) = .; +#else +#define LIKELY_PROFILE() +#endif + /* .data section */ #define DATA_DATA \ *(.data) \ @@ -62,7 +73,8 @@ VMLINUX_SYMBOL(__stop___markers) = .; \ VMLINUX_SYMBOL(__start___tracepoints) = .; \ *(__tracepoints) \ - VMLINUX_SYMBOL(__stop___tracepoints) = .; + VMLINUX_SYMBOL(__stop___tracepoints) = .; \ + LIKELY_PROFILE() #define RO_DATA(align) \ . = ALIGN((align)); \ diff --git a/include/linux/compiler.h b/include/linux/compiler.h index 98115d9d04da..935e30cfaf3c 100644 --- a/include/linux/compiler.h +++ b/include/linux/compiler.h @@ -59,8 +59,65 @@ extern void __chk_io_ptr(const volatile void __iomem *); * specific implementations come from the above header files */ -#define likely(x) __builtin_expect(!!(x), 1) -#define unlikely(x) __builtin_expect(!!(x), 0) +#ifdef CONFIG_TRACE_UNLIKELY_PROFILE +struct ftrace_likely_data { + const char *func; + const char *file; + unsigned line; + unsigned long correct; + unsigned long incorrect; +}; +void ftrace_likely_update(struct ftrace_likely_data *f, int val, int expect); + +#define likely_notrace(x) __builtin_expect(!!(x), 1) +#define unlikely_notrace(x) __builtin_expect(!!(x), 0) + +#define likely_check(x) ({ \ + int ______r; \ + static struct ftrace_likely_data \ + __attribute__((__aligned__(4))) \ + __attribute__((section("_ftrace_likely"))) \ + ______f = { \ + .func = __func__, \ + .file = __FILE__, \ + .line = __LINE__, \ + }; \ + ______f.line = __LINE__; \ + ______r = likely_notrace(x); \ + ftrace_likely_update(&______f, ______r, 1); \ + ______r; \ + }) +#define unlikely_check(x) ({ \ + int ______r; \ + static struct ftrace_likely_data \ + __attribute__((__aligned__(4))) \ + __attribute__((section("_ftrace_unlikely"))) \ + ______f = { \ + .func = __func__, \ + .file = __FILE__, \ + .line = __LINE__, \ + }; \ + ______f.line = __LINE__; \ + ______r = unlikely_notrace(x); \ + ftrace_likely_update(&______f, ______r, 0); \ + ______r; \ + }) + +/* + * Using __builtin_constant_p(x) to ignore cases where the return + * value is always the same. This idea is taken from a similar patch + * written by Daniel Walker. + */ +# ifndef likely +# define likely(x) (__builtin_constant_p(x) ? !!(x) : likely_check(x)) +# endif +# ifndef unlikely +# define unlikely(x) (__builtin_constant_p(x) ? !!(x) : unlikely_check(x)) +# endif +#else +# define likely(x) __builtin_expect(!!(x), 1) +# define unlikely(x) __builtin_expect(!!(x), 0) +#endif /* Optimization barrier */ #ifndef barrier diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index d986216c8327..a604f24c755f 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -159,6 +159,22 @@ config BOOT_TRACER selected, because the self-tests are an initcall as well and that would invalidate the boot trace. ) +config TRACE_UNLIKELY_PROFILE + bool "Trace likely/unlikely profiler" + depends on DEBUG_KERNEL + select TRACING + help + This tracer profiles all the the likely and unlikely macros + in the kernel. It will display the results in: + + /debugfs/tracing/profile_likely + /debugfs/tracing/profile_unlikely + + Note: this will add a significant overhead, only turn this + on if you need to profile the system's use of these macros. + + Say N if unsure. + config STACK_TRACER bool "Trace max stack" depends on HAVE_FUNCTION_TRACER diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index 3e1f361bbc17..98e70ee27986 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -25,5 +25,6 @@ obj-$(CONFIG_STACK_TRACER) += trace_stack.o obj-$(CONFIG_MMIOTRACE) += trace_mmiotrace.o obj-$(CONFIG_BOOT_TRACER) += trace_boot.o obj-$(CONFIG_FUNCTION_RET_TRACER) += trace_functions_return.o +obj-$(CONFIG_TRACE_UNLIKELY_PROFILE) += trace_unlikely.o libftrace-y := ftrace.o diff --git a/kernel/trace/trace_unlikely.c b/kernel/trace/trace_unlikely.c new file mode 100644 index 000000000000..94932696069f --- /dev/null +++ b/kernel/trace/trace_unlikely.c @@ -0,0 +1,164 @@ +/* + * unlikely profiler + * + * Copyright (C) 2008 Steven Rostedt + */ +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include "trace.h" + +void ftrace_likely_update(struct ftrace_likely_data *f, int val, int expect) +{ + /* FIXME: Make this atomic! */ + if (val == expect) + f->correct++; + else + f->incorrect++; +} +EXPORT_SYMBOL(ftrace_likely_update); + +struct ftrace_pointer { + void *start; + void *stop; +}; + +static void * +t_next(struct seq_file *m, void *v, loff_t *pos) +{ + struct ftrace_pointer *f = m->private; + struct ftrace_likely_data *p = v; + + (*pos)++; + + if (v == (void *)1) + return f->start; + + ++p; + + if ((void *)p >= (void *)f->stop) + return NULL; + + return p; +} + +static void *t_start(struct seq_file *m, loff_t *pos) +{ + void *t = (void *)1; + loff_t l = 0; + + for (; t && l < *pos; t = t_next(m, t, &l)) + ; + + return t; +} + +static void t_stop(struct seq_file *m, void *p) +{ +} + +static int t_show(struct seq_file *m, void *v) +{ + struct ftrace_likely_data *p = v; + const char *f; + unsigned long percent; + + if (v == (void *)1) { + seq_printf(m, " correct incorrect %% " + " Function " + " File Line\n" + " ------- --------- - " + " -------- " + " ---- ----\n"); + return 0; + } + + /* Only print the file, not the path */ + f = p->file + strlen(p->file); + while (f >= p->file && *f != '/') + f--; + f++; + + if (p->correct) { + percent = p->incorrect * 100; + percent /= p->correct + p->incorrect; + } else + percent = p->incorrect ? 100 : 0; + + seq_printf(m, "%8lu %8lu %3lu ", p->correct, p->incorrect, percent); + seq_printf(m, "%-30.30s %-20.20s %d\n", p->func, f, p->line); + return 0; +} + +static struct seq_operations tracing_likely_seq_ops = { + .start = t_start, + .next = t_next, + .stop = t_stop, + .show = t_show, +}; + +static int tracing_likely_open(struct inode *inode, struct file *file) +{ + int ret; + + ret = seq_open(file, &tracing_likely_seq_ops); + if (!ret) { + struct seq_file *m = file->private_data; + m->private = (void *)inode->i_private; + } + + return ret; +} + +static struct file_operations tracing_likely_fops = { + .open = tracing_likely_open, + .read = seq_read, + .llseek = seq_lseek, +}; + +extern unsigned long __start_likely_profile[]; +extern unsigned long __stop_likely_profile[]; +extern unsigned long __start_unlikely_profile[]; +extern unsigned long __stop_unlikely_profile[]; + +static struct ftrace_pointer ftrace_likely_pos = { + .start = __start_likely_profile, + .stop = __stop_likely_profile, +}; + +static struct ftrace_pointer ftrace_unlikely_pos = { + .start = __start_unlikely_profile, + .stop = __stop_unlikely_profile, +}; + +static __init int ftrace_unlikely_init(void) +{ + struct dentry *d_tracer; + struct dentry *entry; + + d_tracer = tracing_init_dentry(); + + entry = debugfs_create_file("profile_likely", 0444, d_tracer, + &ftrace_likely_pos, + &tracing_likely_fops); + if (!entry) + pr_warning("Could not create debugfs 'profile_likely' entry\n"); + + entry = debugfs_create_file("profile_unlikely", 0444, d_tracer, + &ftrace_unlikely_pos, + &tracing_likely_fops); + if (!entry) + pr_warning("Could not create debugfs" + " 'profile_unlikely' entry\n"); + + return 0; +} + +device_initcall(ftrace_unlikely_init); -- cgit v1.2.3 From 4a61204856e8b28e9f5489a7875cb3a60afd1e43 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Tue, 11 Nov 2008 19:09:29 -0800 Subject: x86: signal_32: introduce retcode and rt_retcode Impact: cleanup Introduce retcode and rt_retcode to replace setting up frame->retcode. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/kernel/signal_32.c | 30 ++++++++++++++++++++++++------ 1 file changed, 24 insertions(+), 6 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/signal_32.c b/arch/x86/kernel/signal_32.c index 27a5c8174322..514171ac0d03 100644 --- a/arch/x86/kernel/signal_32.c +++ b/arch/x86/kernel/signal_32.c @@ -45,6 +45,28 @@ # define FIX_EFLAGS __FIX_EFLAGS #endif +static const struct { + u16 poplmovl; + u32 val; + u16 int80; +} __attribute__((packed)) retcode = { + 0xb858, /* popl %eax; movl $..., %eax */ + __NR_sigreturn, + 0x80cd, /* int $0x80 */ +}; + +static const struct { + u8 movl; + u32 val; + u16 int80; + u8 pad; +} __attribute__((packed)) rt_retcode = { + 0xb8, /* movl $..., %eax */ + __NR_rt_sigreturn, + 0x80cd, /* int $0x80 */ + 0 +}; + /* * Atomically swap in the new signal mask, and wait for a signal. */ @@ -427,9 +449,7 @@ __setup_frame(int sig, struct k_sigaction *ka, sigset_t *set, * reasons and because gdb uses it as a signature to notice * signal handler stack frames. */ - err |= __put_user(0xb858, (short __user *)(frame->retcode+0)); - err |= __put_user(__NR_sigreturn, (int __user *)(frame->retcode+2)); - err |= __put_user(0x80cd, (short __user *)(frame->retcode+6)); + err |= __put_user(*((u64 *)&retcode), (u64 *)frame->retcode); if (err) return -EFAULT; @@ -498,9 +518,7 @@ static int __setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, * reasons and because gdb uses it as a signature to notice * signal handler stack frames. */ - err |= __put_user(0xb8, (char __user *)(frame->retcode+0)); - err |= __put_user(__NR_rt_sigreturn, (int __user *)(frame->retcode+1)); - err |= __put_user(0x80cd, (short __user *)(frame->retcode+5)); + err |= __put_user(*((u64 *)&rt_retcode), (u64 *)frame->retcode); if (err) return -EFAULT; -- cgit v1.2.3 From 9cc3c49ed1b1dbf6e50de4055a6773bd162f24b7 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Tue, 11 Nov 2008 19:11:39 -0800 Subject: x86: ia32_signal: remove unnecessary padding Impact: reduce structure padding Remove unnecessary paddings, this saves 4 bytes. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/ia32/ia32_signal.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/ia32/ia32_signal.c b/arch/x86/ia32/ia32_signal.c index 47ddc23f4f54..e2d0bc779bf7 100644 --- a/arch/x86/ia32/ia32_signal.c +++ b/arch/x86/ia32/ia32_signal.c @@ -427,12 +427,10 @@ int ia32_setup_frame(int sig, struct k_sigaction *ka, u16 poplmovl; u32 val; u16 int80; - u16 pad; } __attribute__((packed)) code = { 0xb858, /* popl %eax ; movl $...,%eax */ __NR_ia32_sigreturn, 0x80cd, /* int $0x80 */ - 0, }; frame = get_sigframe(ka, regs, sizeof(*frame), &fpstate); @@ -508,8 +506,7 @@ int ia32_setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, u8 movl; u32 val; u16 int80; - u16 pad; - u8 pad2; + u8 pad; } __attribute__((packed)) code = { 0xb8, __NR_ia32_rt_sigreturn, -- cgit v1.2.3 From 2b7d0390a6d6d595f43ea3806639664afe5b9ebe Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Wed, 12 Nov 2008 13:17:38 +0100 Subject: tracing: branch tracer, fix vdso crash Impact: fix bootup crash the branch tracer missed arch/x86/vdso/vclock_gettime.c from disabling tracing, which caused such bootup crashes: [ 201.840097] init[1]: segfault at 7fffed3fe7c0 ip 00007fffed3fea2e sp 000077 also clean up the ugly ifdefs in arch/x86/kernel/vsyscall_64.c by creating DISABLE_UNLIKELY_PROFILE facility for code to turn off instrumentation on a per file basis. Signed-off-by: Ingo Molnar --- arch/x86/kernel/vsyscall_64.c | 9 ++------- arch/x86/vdso/vclock_gettime.c | 3 +++ include/linux/compiler.h | 6 +++++- 3 files changed, 10 insertions(+), 8 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/vsyscall_64.c b/arch/x86/kernel/vsyscall_64.c index 2f90202e59b3..ece02932ea57 100644 --- a/arch/x86/kernel/vsyscall_64.c +++ b/arch/x86/kernel/vsyscall_64.c @@ -17,13 +17,8 @@ * want per guest time just set the kernel.vsyscall64 sysctl to 0. */ -/* Protect userspace from profiling */ -#ifdef CONFIG_TRACE_UNLIKELY_PROFILE -# undef likely -# undef unlikely -# define likely(x) likely_notrace(x) -# define unlikely(x) unlikely_notrace(x) -#endif +/* Disable profiling for userspace code: */ +#define DISABLE_UNLIKELY_PROFILE #include #include diff --git a/arch/x86/vdso/vclock_gettime.c b/arch/x86/vdso/vclock_gettime.c index 1ef0f90813d6..6e667631e7dc 100644 --- a/arch/x86/vdso/vclock_gettime.c +++ b/arch/x86/vdso/vclock_gettime.c @@ -9,6 +9,9 @@ * Also alternative() doesn't work. */ +/* Disable profiling for userspace code: */ +#define DISABLE_UNLIKELY_PROFILE + #include #include #include diff --git a/include/linux/compiler.h b/include/linux/compiler.h index 935e30cfaf3c..63b7d9089d6e 100644 --- a/include/linux/compiler.h +++ b/include/linux/compiler.h @@ -59,7 +59,11 @@ extern void __chk_io_ptr(const volatile void __iomem *); * specific implementations come from the above header files */ -#ifdef CONFIG_TRACE_UNLIKELY_PROFILE +/* + * Note: DISABLE_UNLIKELY_PROFILE can be used by special lowlevel code + * to disable branch tracing on a per file basis. + */ +#if defined(CONFIG_TRACE_UNLIKELY_PROFILE) && !defined(DISABLE_UNLIKELY_PROFILE) struct ftrace_likely_data { const char *func; const char *file; -- cgit v1.2.3 From a7d41820f683c35b53af719210a51f6aa0f86a6a Mon Sep 17 00:00:00 2001 From: Eduardo Habkost Date: Wed, 12 Nov 2008 11:34:37 -0200 Subject: x86 kdump: extract kdump-specific code from crash_nmi_callback() Impact: cleanup The NMI CPU-halting code will be used on non-kdump cases, also (e.g. emergency_reboot when virtualization is enabled). Signed-off-by: Eduardo Habkost Signed-off-by: Ingo Molnar --- arch/x86/kernel/crash.c | 31 ++++++++++++++++++++----------- 1 file changed, 20 insertions(+), 11 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/crash.c b/arch/x86/kernel/crash.c index 268553817909..60475422a51a 100644 --- a/arch/x86/kernel/crash.c +++ b/arch/x86/kernel/crash.c @@ -35,19 +35,34 @@ static int crashing_cpu; #if defined(CONFIG_SMP) && defined(CONFIG_X86_LOCAL_APIC) static atomic_t waiting_for_crash_ipi; -static int crash_nmi_callback(struct notifier_block *self, - unsigned long val, void *data) +static void kdump_nmi_callback(int cpu, struct die_args *args) { struct pt_regs *regs; #ifdef CONFIG_X86_32 struct pt_regs fixed_regs; #endif + + regs = args->regs; + +#ifdef CONFIG_X86_32 + if (!user_mode_vm(regs)) { + crash_fixup_ss_esp(&fixed_regs, regs); + regs = &fixed_regs; + } +#endif + crash_save_cpu(regs, cpu); + + disable_local_APIC(); +} + +static int crash_nmi_callback(struct notifier_block *self, + unsigned long val, void *data) +{ int cpu; if (val != DIE_NMI_IPI) return NOTIFY_OK; - regs = ((struct die_args *)data)->regs; cpu = raw_smp_processor_id(); /* Don't do anything if this handler is invoked on crashing cpu. @@ -58,14 +73,8 @@ static int crash_nmi_callback(struct notifier_block *self, return NOTIFY_STOP; local_irq_disable(); -#ifdef CONFIG_X86_32 - if (!user_mode_vm(regs)) { - crash_fixup_ss_esp(&fixed_regs, regs); - regs = &fixed_regs; - } -#endif - crash_save_cpu(regs, cpu); - disable_local_APIC(); + kdump_nmi_callback(cpu, (struct die_args *)data); + atomic_dec(&waiting_for_crash_ipi); /* Assume hlt works */ halt(); -- cgit v1.2.3 From b2bbe71b829564fb65a6bc7e1e25e02d70cffce8 Mon Sep 17 00:00:00 2001 From: Eduardo Habkost Date: Wed, 12 Nov 2008 11:34:38 -0200 Subject: x86 kdump: move crashing_cpu assignment to nmi_shootdown_cpus() Impact: cleanup This variable will be moved to non-kdump-specific code. Signed-off-by: Eduardo Habkost Signed-off-by: Ingo Molnar --- arch/x86/kernel/crash.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/crash.c b/arch/x86/kernel/crash.c index 60475422a51a..ed2f0f9dc894 100644 --- a/arch/x86/kernel/crash.c +++ b/arch/x86/kernel/crash.c @@ -29,10 +29,11 @@ #include +#if defined(CONFIG_SMP) && defined(CONFIG_X86_LOCAL_APIC) + /* This keeps a track of which one is crashing cpu. */ static int crashing_cpu; -#if defined(CONFIG_SMP) && defined(CONFIG_X86_LOCAL_APIC) static atomic_t waiting_for_crash_ipi; static void kdump_nmi_callback(int cpu, struct die_args *args) @@ -100,6 +101,9 @@ static void nmi_shootdown_cpus(void) { unsigned long msecs; + /* Make a note of crashing cpu. Will be used in NMI callback.*/ + crashing_cpu = safe_smp_processor_id(); + atomic_set(&waiting_for_crash_ipi, num_online_cpus() - 1); /* Would it be better to replace the trap vector here? */ if (register_die_notifier(&crash_nmi_nb)) @@ -140,8 +144,6 @@ void native_machine_crash_shutdown(struct pt_regs *regs) /* The kernel is broken so disable interrupts */ local_irq_disable(); - /* Make a note of crashing cpu. Will be used in NMI callback.*/ - crashing_cpu = safe_smp_processor_id(); nmi_shootdown_cpus(); lapic_shutdown(); #if defined(CONFIG_X86_IO_APIC) -- cgit v1.2.3 From d1e7b91cfaa8fc5ed736dcfb8beb5134a2385228 Mon Sep 17 00:00:00 2001 From: Eduardo Habkost Date: Wed, 12 Nov 2008 11:34:39 -0200 Subject: x86 kdump: create kdump_nmi_shootdown_cpus() Impact: cleanup For the kdump-specific code that was living on nmi_shootdown_cpus(). Signed-off-by: Eduardo Habkost Signed-off-by: Ingo Molnar --- arch/x86/kernel/crash.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/crash.c b/arch/x86/kernel/crash.c index ed2f0f9dc894..75c468cc7e59 100644 --- a/arch/x86/kernel/crash.c +++ b/arch/x86/kernel/crash.c @@ -122,10 +122,17 @@ static void nmi_shootdown_cpus(void) } /* Leave the nmi callback set */ +} + +static void kdump_nmi_shootdown_cpus(void) +{ + nmi_shootdown_cpus(); + disable_local_APIC(); } + #else -static void nmi_shootdown_cpus(void) +static void kdump_nmi_shootdown_cpus(void) { /* There are no cpus to shootdown */ } @@ -144,7 +151,7 @@ void native_machine_crash_shutdown(struct pt_regs *regs) /* The kernel is broken so disable interrupts */ local_irq_disable(); - nmi_shootdown_cpus(); + kdump_nmi_shootdown_cpus(); lapic_shutdown(); #if defined(CONFIG_X86_IO_APIC) disable_IO_APIC(); -- cgit v1.2.3 From 8e294786316aca41c66b8b73ba1ee74a4ae7d452 Mon Sep 17 00:00:00 2001 From: Eduardo Habkost Date: Wed, 12 Nov 2008 11:34:40 -0200 Subject: x86 kdump: make kdump_nmi_callback() a function ptr on crash_nmi_callback() Impact: extend nmi_shootdown_cpus() with a callback The reboot code will use a different function on crash_nmi_callback(). Adding a function pointer parameter to nmi_shootdown_cpus() for that. Signed-off-by: Eduardo Habkost Signed-off-by: Ingo Molnar --- arch/x86/kernel/crash.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/crash.c b/arch/x86/kernel/crash.c index 75c468cc7e59..f23c2beeb37d 100644 --- a/arch/x86/kernel/crash.c +++ b/arch/x86/kernel/crash.c @@ -29,10 +29,13 @@ #include +typedef void (*nmi_shootdown_cb)(int, struct die_args*); + #if defined(CONFIG_SMP) && defined(CONFIG_X86_LOCAL_APIC) /* This keeps a track of which one is crashing cpu. */ static int crashing_cpu; +static nmi_shootdown_cb shootdown_callback; static atomic_t waiting_for_crash_ipi; @@ -74,7 +77,7 @@ static int crash_nmi_callback(struct notifier_block *self, return NOTIFY_STOP; local_irq_disable(); - kdump_nmi_callback(cpu, (struct die_args *)data); + shootdown_callback(cpu, (struct die_args *)data); atomic_dec(&waiting_for_crash_ipi); /* Assume hlt works */ @@ -97,13 +100,15 @@ static struct notifier_block crash_nmi_nb = { .notifier_call = crash_nmi_callback, }; -static void nmi_shootdown_cpus(void) +static void nmi_shootdown_cpus(nmi_shootdown_cb callback) { unsigned long msecs; /* Make a note of crashing cpu. Will be used in NMI callback.*/ crashing_cpu = safe_smp_processor_id(); + shootdown_callback = callback; + atomic_set(&waiting_for_crash_ipi, num_online_cpus() - 1); /* Would it be better to replace the trap vector here? */ if (register_die_notifier(&crash_nmi_nb)) @@ -126,7 +131,7 @@ static void nmi_shootdown_cpus(void) static void kdump_nmi_shootdown_cpus(void) { - nmi_shootdown_cpus(); + nmi_shootdown_cpus(kdump_nmi_callback); disable_local_APIC(); } -- cgit v1.2.3 From c370e5e089adfd5b1b863f3464cccae9ebf33cca Mon Sep 17 00:00:00 2001 From: Eduardo Habkost Date: Wed, 12 Nov 2008 11:34:41 -0200 Subject: x86 kdump: make nmi_shootdown_cpus() non-static Impact: make API available to the rest of x86 platform code Add prototype to asm/reboot.h. Signed-off-by: Eduardo Habkost Signed-off-by: Ingo Molnar --- arch/x86/include/asm/reboot.h | 5 +++++ arch/x86/kernel/crash.c | 3 +-- 2 files changed, 6 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/reboot.h b/arch/x86/include/asm/reboot.h index df7710354f85..562d4fd31ba8 100644 --- a/arch/x86/include/asm/reboot.h +++ b/arch/x86/include/asm/reboot.h @@ -1,6 +1,8 @@ #ifndef _ASM_X86_REBOOT_H #define _ASM_X86_REBOOT_H +#include + struct pt_regs; struct machine_ops { @@ -18,4 +20,7 @@ void native_machine_crash_shutdown(struct pt_regs *regs); void native_machine_shutdown(void); void machine_real_restart(const unsigned char *code, int length); +typedef void (*nmi_shootdown_cb)(int, struct die_args*); +void nmi_shootdown_cpus(nmi_shootdown_cb callback); + #endif /* _ASM_X86_REBOOT_H */ diff --git a/arch/x86/kernel/crash.c b/arch/x86/kernel/crash.c index f23c2beeb37d..fb298d1daac9 100644 --- a/arch/x86/kernel/crash.c +++ b/arch/x86/kernel/crash.c @@ -29,7 +29,6 @@ #include -typedef void (*nmi_shootdown_cb)(int, struct die_args*); #if defined(CONFIG_SMP) && defined(CONFIG_X86_LOCAL_APIC) @@ -100,7 +99,7 @@ static struct notifier_block crash_nmi_nb = { .notifier_call = crash_nmi_callback, }; -static void nmi_shootdown_cpus(nmi_shootdown_cb callback) +void nmi_shootdown_cpus(nmi_shootdown_cb callback) { unsigned long msecs; -- cgit v1.2.3 From 2ddded213895e41b9cfe1c084127e6c01632ac1a Mon Sep 17 00:00:00 2001 From: Eduardo Habkost Date: Wed, 12 Nov 2008 11:34:42 -0200 Subject: x86: move nmi_shootdown_cpus() to reboot.c Impact: make nmi_shootdown_cpus() available to the rest of the x86 platform Now nmi_shootdown_cpus() is ready to be used by non-kdump code also. Move it to reboot.c. Signed-off-by: Eduardo Habkost Signed-off-by: Ingo Molnar --- arch/x86/kernel/crash.c | 76 -------------------------------------------- arch/x86/kernel/reboot.c | 83 ++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 83 insertions(+), 76 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/crash.c b/arch/x86/kernel/crash.c index fb298d1daac9..d84a852e4cd7 100644 --- a/arch/x86/kernel/crash.c +++ b/arch/x86/kernel/crash.c @@ -32,12 +32,6 @@ #if defined(CONFIG_SMP) && defined(CONFIG_X86_LOCAL_APIC) -/* This keeps a track of which one is crashing cpu. */ -static int crashing_cpu; -static nmi_shootdown_cb shootdown_callback; - -static atomic_t waiting_for_crash_ipi; - static void kdump_nmi_callback(int cpu, struct die_args *args) { struct pt_regs *regs; @@ -58,76 +52,6 @@ static void kdump_nmi_callback(int cpu, struct die_args *args) disable_local_APIC(); } -static int crash_nmi_callback(struct notifier_block *self, - unsigned long val, void *data) -{ - int cpu; - - if (val != DIE_NMI_IPI) - return NOTIFY_OK; - - cpu = raw_smp_processor_id(); - - /* Don't do anything if this handler is invoked on crashing cpu. - * Otherwise, system will completely hang. Crashing cpu can get - * an NMI if system was initially booted with nmi_watchdog parameter. - */ - if (cpu == crashing_cpu) - return NOTIFY_STOP; - local_irq_disable(); - - shootdown_callback(cpu, (struct die_args *)data); - - atomic_dec(&waiting_for_crash_ipi); - /* Assume hlt works */ - halt(); - for (;;) - cpu_relax(); - - return 1; -} - -static void smp_send_nmi_allbutself(void) -{ - cpumask_t mask = cpu_online_map; - cpu_clear(safe_smp_processor_id(), mask); - if (!cpus_empty(mask)) - send_IPI_mask(mask, NMI_VECTOR); -} - -static struct notifier_block crash_nmi_nb = { - .notifier_call = crash_nmi_callback, -}; - -void nmi_shootdown_cpus(nmi_shootdown_cb callback) -{ - unsigned long msecs; - - /* Make a note of crashing cpu. Will be used in NMI callback.*/ - crashing_cpu = safe_smp_processor_id(); - - shootdown_callback = callback; - - atomic_set(&waiting_for_crash_ipi, num_online_cpus() - 1); - /* Would it be better to replace the trap vector here? */ - if (register_die_notifier(&crash_nmi_nb)) - return; /* return what? */ - /* Ensure the new callback function is set before sending - * out the NMI - */ - wmb(); - - smp_send_nmi_allbutself(); - - msecs = 1000; /* Wait at most a second for the other cpus to stop */ - while ((atomic_read(&waiting_for_crash_ipi) > 0) && msecs) { - mdelay(1); - msecs--; - } - - /* Leave the nmi callback set */ -} - static void kdump_nmi_shootdown_cpus(void) { nmi_shootdown_cpus(kdump_nmi_callback); diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c index 724adfc63cb9..364edeecc235 100644 --- a/arch/x86/kernel/reboot.c +++ b/arch/x86/kernel/reboot.c @@ -21,6 +21,9 @@ # include #endif +#include + + /* * Power off function, if any */ @@ -514,3 +517,83 @@ void machine_crash_shutdown(struct pt_regs *regs) machine_ops.crash_shutdown(regs); } #endif + + +#if defined(CONFIG_SMP) && defined(CONFIG_X86_LOCAL_APIC) + +/* This keeps a track of which one is crashing cpu. */ +static int crashing_cpu; +static nmi_shootdown_cb shootdown_callback; + +static atomic_t waiting_for_crash_ipi; + +static int crash_nmi_callback(struct notifier_block *self, + unsigned long val, void *data) +{ + int cpu; + + if (val != DIE_NMI_IPI) + return NOTIFY_OK; + + cpu = raw_smp_processor_id(); + + /* Don't do anything if this handler is invoked on crashing cpu. + * Otherwise, system will completely hang. Crashing cpu can get + * an NMI if system was initially booted with nmi_watchdog parameter. + */ + if (cpu == crashing_cpu) + return NOTIFY_STOP; + local_irq_disable(); + + shootdown_callback(cpu, (struct die_args *)data); + + atomic_dec(&waiting_for_crash_ipi); + /* Assume hlt works */ + halt(); + for (;;) + cpu_relax(); + + return 1; +} + +static void smp_send_nmi_allbutself(void) +{ + cpumask_t mask = cpu_online_map; + cpu_clear(safe_smp_processor_id(), mask); + if (!cpus_empty(mask)) + send_IPI_mask(mask, NMI_VECTOR); +} + +static struct notifier_block crash_nmi_nb = { + .notifier_call = crash_nmi_callback, +}; + +void nmi_shootdown_cpus(nmi_shootdown_cb callback) +{ + unsigned long msecs; + + /* Make a note of crashing cpu. Will be used in NMI callback.*/ + crashing_cpu = safe_smp_processor_id(); + + shootdown_callback = callback; + + atomic_set(&waiting_for_crash_ipi, num_online_cpus() - 1); + /* Would it be better to replace the trap vector here? */ + if (register_die_notifier(&crash_nmi_nb)) + return; /* return what? */ + /* Ensure the new callback function is set before sending + * out the NMI + */ + wmb(); + + smp_send_nmi_allbutself(); + + msecs = 1000; /* Wait at most a second for the other cpus to stop */ + while ((atomic_read(&waiting_for_crash_ipi) > 0) && msecs) { + mdelay(1); + msecs--; + } + + /* Leave the nmi callback set */ +} +#endif -- cgit v1.2.3 From bb8dd270e62217e2d2172094c6c352c4ddc0a127 Mon Sep 17 00:00:00 2001 From: Eduardo Habkost Date: Wed, 12 Nov 2008 11:34:43 -0200 Subject: x86: make nmi_shootdown_cpus() available on !SMP and !X86_LOCAL_APIC Impact: widen nmi_shootdown_cpus() availability The X86_LOCAL_APIC #ifdef was for kdump. For !SMP, the function simply does nothing. Signed-off-by: Eduardo Habkost Signed-off-by: Ingo Molnar --- arch/x86/kernel/reboot.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c index 364edeecc235..17a41e055565 100644 --- a/arch/x86/kernel/reboot.c +++ b/arch/x86/kernel/reboot.c @@ -519,7 +519,7 @@ void machine_crash_shutdown(struct pt_regs *regs) #endif -#if defined(CONFIG_SMP) && defined(CONFIG_X86_LOCAL_APIC) +#if defined(CONFIG_SMP) /* This keeps a track of which one is crashing cpu. */ static int crashing_cpu; @@ -568,6 +568,12 @@ static struct notifier_block crash_nmi_nb = { .notifier_call = crash_nmi_callback, }; +/* Halt all other CPUs, calling the specified function on each of them + * + * This function can be used to halt all other CPUs on crash + * or emergency reboot time. The function passed as parameter + * will be called inside a NMI handler on all CPUs. + */ void nmi_shootdown_cpus(nmi_shootdown_cb callback) { unsigned long msecs; @@ -596,4 +602,9 @@ void nmi_shootdown_cpus(nmi_shootdown_cb callback) /* Leave the nmi callback set */ } +#else /* !CONFIG_SMP */ +void nmi_shootdown_cpus(nmi_shootdown_cb callback) +{ + /* No other CPUs to shoot down */ +} #endif -- cgit v1.2.3 From c415b3dce30dfb41234e118662e8720f47343a4f Mon Sep 17 00:00:00 2001 From: Eduardo Habkost Date: Wed, 12 Nov 2008 11:34:44 -0200 Subject: x86: disable IRQs before doing anything on nmi_shootdown_cpus() Impact: make nmi_shootdown_cpus() callable from preemptible context We need to know on which CPU we are running on, and we don't want to be preempted while doing this. Signed-off-by: Eduardo Habkost Signed-off-by: Ingo Molnar --- arch/x86/kernel/reboot.c | 1 + 1 file changed, 1 insertion(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c index 17a41e055565..c3cd512484e5 100644 --- a/arch/x86/kernel/reboot.c +++ b/arch/x86/kernel/reboot.c @@ -577,6 +577,7 @@ static struct notifier_block crash_nmi_nb = { void nmi_shootdown_cpus(nmi_shootdown_cb callback) { unsigned long msecs; + local_irq_disable(); /* Make a note of crashing cpu. Will be used in NMI callback.*/ crashing_cpu = safe_smp_processor_id(); -- cgit v1.2.3 From 8665596ec05498525014436520b316ba174a068a Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Wed, 12 Nov 2008 10:27:35 -0800 Subject: x86: fix up the new IRQ code for older versions of gas Older versions of gas don't implement the C-style != operator, they instead want the Pascal-style <> operator. Change != to <> so we don't break compilation with those old versions of gas. Signed-off-by: H. Peter Anvin --- arch/x86/kernel/entry_32.S | 4 ++-- arch/x86/kernel/entry_64.S | 4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S index dae81b9fd451..bd02ec77edc4 100644 --- a/arch/x86/kernel/entry_32.S +++ b/arch/x86/kernel/entry_32.S @@ -635,12 +635,12 @@ vector=FIRST_EXTERNAL_VECTOR .balign 32 .rept 7 .if vector < NR_VECTORS - .if vector != FIRST_EXTERNAL_VECTOR + .if vector <> FIRST_EXTERNAL_VECTOR CFI_ADJUST_CFA_OFFSET -4 .endif 1: pushl $(~vector+0x80) /* Note: always in signed byte range */ CFI_ADJUST_CFA_OFFSET 4 - .if ((vector-FIRST_EXTERNAL_VECTOR)%7) != 6 + .if ((vector-FIRST_EXTERNAL_VECTOR)%7) <> 6 jmp 2f .endif .previous diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 9b2aeaac9a6b..2b42362a85b2 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -644,12 +644,12 @@ vector=FIRST_EXTERNAL_VECTOR .balign 32 .rept 7 .if vector < NR_VECTORS - .if vector != FIRST_EXTERNAL_VECTOR + .if vector <> FIRST_EXTERNAL_VECTOR CFI_ADJUST_CFA_OFFSET -8 .endif 1: pushq $(~vector+0x80) /* Note: always in signed byte range */ CFI_ADJUST_CFA_OFFSET 8 - .if ((vector-FIRST_EXTERNAL_VECTOR)%7) != 6 + .if ((vector-FIRST_EXTERNAL_VECTOR)%7) <> 6 jmp 2f .endif .previous -- cgit v1.2.3 From 8652cb4b0d87accbe78725fd2a13be2787059649 Mon Sep 17 00:00:00 2001 From: Prarit Bhargava Date: Wed, 12 Nov 2008 13:35:00 -0500 Subject: x86: warn of incorrect cpu_khz on AMD systems Impact: add debug check If none of the perfctrs are free when calculating cpu_khz we default to using ctr 3 (ie, we just choose 3). This may lead to an incorrect tsc freq value which can cause the system to be unstable. To aid in future debugging, WARN the user of a potential problem. Signed-off-by: Prarit Bhargava Signed-off-by: Ingo Molnar --- arch/x86/kernel/time_64.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/time_64.c b/arch/x86/kernel/time_64.c index cb19d650c216..418a095c5796 100644 --- a/arch/x86/kernel/time_64.c +++ b/arch/x86/kernel/time_64.c @@ -80,6 +80,8 @@ unsigned long __init calibrate_cpu(void) break; no_ctr_free = (i == 4); if (no_ctr_free) { + WARN(1, KERN_WARNING "Warning: AMD perfctrs busy ... " + "cpu_khz value may be incorrect.\n"); i = 3; rdmsrl(MSR_K7_EVNTSEL3, evntsel3); wrmsrl(MSR_K7_EVNTSEL3, 0); -- cgit v1.2.3 From 2ed84eeb8808cf3c9f039213ca137ffd7d753f0e Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 12 Nov 2008 15:24:24 -0500 Subject: trace: rename unlikely profiler to branch profiler Impact: name change of unlikely tracer and profiler Ingo Molnar suggested changing the config from UNLIKELY_PROFILE to BRANCH_PROFILING. I never did like the "unlikely" name so I went one step farther, and renamed all the unlikely configurations to a "BRANCH" variant. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/x86/kernel/vsyscall_64.c | 2 +- arch/x86/vdso/vclock_gettime.c | 2 +- include/asm-generic/vmlinux.lds.h | 2 +- include/linux/compiler.h | 19 ++++++++++--------- kernel/trace/Kconfig | 10 +++++----- kernel/trace/Makefile | 7 +++---- kernel/trace/trace.c | 2 +- kernel/trace/trace.h | 6 +++--- kernel/trace/trace_unlikely.c | 4 ++-- 9 files changed, 27 insertions(+), 27 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/vsyscall_64.c b/arch/x86/kernel/vsyscall_64.c index ece02932ea57..6f3d3d4cd973 100644 --- a/arch/x86/kernel/vsyscall_64.c +++ b/arch/x86/kernel/vsyscall_64.c @@ -18,7 +18,7 @@ */ /* Disable profiling for userspace code: */ -#define DISABLE_UNLIKELY_PROFILE +#define DISABLE_BRANCH_PROFILING #include #include diff --git a/arch/x86/vdso/vclock_gettime.c b/arch/x86/vdso/vclock_gettime.c index 6e667631e7dc..d9d35824c56f 100644 --- a/arch/x86/vdso/vclock_gettime.c +++ b/arch/x86/vdso/vclock_gettime.c @@ -10,7 +10,7 @@ */ /* Disable profiling for userspace code: */ -#define DISABLE_UNLIKELY_PROFILE +#define DISABLE_BRANCH_PROFILING #include #include diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h index e10beb5335c9..a5e4ed9baec8 100644 --- a/include/asm-generic/vmlinux.lds.h +++ b/include/asm-generic/vmlinux.lds.h @@ -45,7 +45,7 @@ #define MCOUNT_REC() #endif -#ifdef CONFIG_TRACE_UNLIKELY_PROFILE +#ifdef CONFIG_TRACE_BRANCH_PROFILING #define LIKELY_PROFILE() VMLINUX_SYMBOL(__start_likely_profile) = .; \ *(_ftrace_likely) \ VMLINUX_SYMBOL(__stop_likely_profile) = .; \ diff --git a/include/linux/compiler.h b/include/linux/compiler.h index 63b7d9089d6e..c7d804a7a4d6 100644 --- a/include/linux/compiler.h +++ b/include/linux/compiler.h @@ -59,26 +59,27 @@ extern void __chk_io_ptr(const volatile void __iomem *); * specific implementations come from the above header files */ -/* - * Note: DISABLE_UNLIKELY_PROFILE can be used by special lowlevel code - * to disable branch tracing on a per file basis. - */ -#if defined(CONFIG_TRACE_UNLIKELY_PROFILE) && !defined(DISABLE_UNLIKELY_PROFILE) -struct ftrace_likely_data { +struct ftrace_branch_data { const char *func; const char *file; unsigned line; unsigned long correct; unsigned long incorrect; }; -void ftrace_likely_update(struct ftrace_likely_data *f, int val, int expect); + +/* + * Note: DISABLE_BRANCH_PROFILING can be used by special lowlevel code + * to disable branch tracing on a per file basis. + */ +#if defined(CONFIG_TRACE_BRANCH_PROFILING) && !defined(DISABLE_BRANCH_PROFILING) +void ftrace_likely_update(struct ftrace_branch_data *f, int val, int expect); #define likely_notrace(x) __builtin_expect(!!(x), 1) #define unlikely_notrace(x) __builtin_expect(!!(x), 0) #define likely_check(x) ({ \ int ______r; \ - static struct ftrace_likely_data \ + static struct ftrace_branch_data \ __attribute__((__aligned__(4))) \ __attribute__((section("_ftrace_likely"))) \ ______f = { \ @@ -93,7 +94,7 @@ void ftrace_likely_update(struct ftrace_likely_data *f, int val, int expect); }) #define unlikely_check(x) ({ \ int ______r; \ - static struct ftrace_likely_data \ + static struct ftrace_branch_data \ __attribute__((__aligned__(4))) \ __attribute__((section("_ftrace_unlikely"))) \ ______f = { \ diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 8abcaf821beb..9c89526b6b7c 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -159,7 +159,7 @@ config BOOT_TRACER selected, because the self-tests are an initcall as well and that would invalidate the boot trace. ) -config TRACE_UNLIKELY_PROFILE +config TRACE_BRANCH_PROFILING bool "Trace likely/unlikely profiler" depends on DEBUG_KERNEL select TRACING @@ -175,7 +175,7 @@ config TRACE_UNLIKELY_PROFILE Say N if unsure. -config TRACING_UNLIKELY +config TRACING_BRANCHES bool help Selected by tracers that will trace the likely and unlikely @@ -183,10 +183,10 @@ config TRACING_UNLIKELY profiled. Profiling the tracing infrastructure can only happen when the likelys and unlikelys are not being traced. -config UNLIKELY_TRACER +config BRANCH_TRACER bool "Trace likely/unlikely instances" - depends on TRACE_UNLIKELY_PROFILE - select TRACING_UNLIKELY + depends on TRACE_BRANCH_PROFILING + select TRACING_BRANCHES help This traces the events of likely and unlikely condition calls in the kernel. The difference between this and the diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index c938d03516c0..0087df7ba44e 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -11,9 +11,8 @@ obj-y += trace_selftest_dynamic.o endif # If unlikely tracing is enabled, do not trace these files -ifdef CONFIG_TRACING_UNLIKELY -KBUILD_CFLAGS += '-Dlikely(x)=likely_notrace(x)' -KBUILD_CFLAGS += '-Dunlikely(x)=unlikely_notrace(x)' +ifdef CONFIG_TRACING_BRANCHES +KBUILD_CFLAGS += -DDISABLE_BRANCH_PROFILING endif obj-$(CONFIG_FUNCTION_TRACER) += libftrace.o @@ -31,6 +30,6 @@ obj-$(CONFIG_STACK_TRACER) += trace_stack.o obj-$(CONFIG_MMIOTRACE) += trace_mmiotrace.o obj-$(CONFIG_BOOT_TRACER) += trace_boot.o obj-$(CONFIG_FUNCTION_RET_TRACER) += trace_functions_return.o -obj-$(CONFIG_TRACE_UNLIKELY_PROFILE) += trace_unlikely.o +obj-$(CONFIG_TRACE_BRANCH_PROFILING) += trace_unlikely.o libftrace-y := ftrace.o diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index d842db14a59b..bad59d32a4a9 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -258,7 +258,7 @@ static const char *trace_options[] = { "sched-tree", "ftrace_printk", "ftrace_preempt", -#ifdef CONFIG_UNLIKELY_TRACER +#ifdef CONFIG_BRANCH_TRACER "unlikely", #endif NULL diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 9635aa2c4fc1..dccae6312941 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -468,7 +468,7 @@ enum trace_iterator_flags { TRACE_ITER_SCHED_TREE = 0x200, TRACE_ITER_PRINTK = 0x400, TRACE_ITER_PREEMPTONLY = 0x800, -#ifdef CONFIG_UNLIKELY_TRACER +#ifdef CONFIG_BRANCH_TRACER TRACE_ITER_UNLIKELY = 0x1000, #endif }; @@ -530,7 +530,7 @@ static inline void ftrace_preempt_enable(int resched) preempt_enable_notrace(); } -#ifdef CONFIG_UNLIKELY_TRACER +#ifdef CONFIG_BRANCH_TRACER extern int enable_unlikely_tracing(struct trace_array *tr); extern void disable_unlikely_tracing(void); static inline int trace_unlikely_enable(struct trace_array *tr) @@ -552,6 +552,6 @@ static inline int trace_unlikely_enable(struct trace_array *tr) static inline void trace_unlikely_disable(void) { } -#endif /* CONFIG_UNLIKELY_TRACER */ +#endif /* CONFIG_BRANCH_TRACER */ #endif /* _LINUX_KERNEL_TRACE_H */ diff --git a/kernel/trace/trace_unlikely.c b/kernel/trace/trace_unlikely.c index 7290e0e7b4e3..856eb3b7f694 100644 --- a/kernel/trace/trace_unlikely.c +++ b/kernel/trace/trace_unlikely.c @@ -15,7 +15,7 @@ #include #include "trace.h" -#ifdef CONFIG_UNLIKELY_TRACER +#ifdef CONFIG_BRANCH_TRACER static int unlikely_tracing_enabled __read_mostly; static DEFINE_MUTEX(unlikely_tracing_mutex); @@ -119,7 +119,7 @@ static inline void trace_likely_condition(struct ftrace_likely_data *f, int val, int expect) { } -#endif /* CONFIG_UNLIKELY_TRACER */ +#endif /* CONFIG_BRANCH_TRACER */ void ftrace_likely_update(struct ftrace_likely_data *f, int val, int expect) { -- cgit v1.2.3 From 62d59d17a5f98edb48b171742dfa531488802f07 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 12 Nov 2008 22:47:54 +0100 Subject: tracing/function-return-tracer: make the function return tracer lockless Impact: remove spinlocks and irq disabling in function return tracer. I've tried to figure out all of the race condition that could happen when the tracer pushes or pops a return address trace to/from the current thread_info. Theory: _ One thread can only execute on one cpu at a time. So this code doesn't need to be SMP-safe. Just drop the spinlock. _ The only race could happen between the current thread and an interrupt. If an interrupt is raised, it will increase the index of the return stack storage and then execute until the end of the tracing to finally free the index it used. We don't need to disable irqs. This is theorical. In practice, I've tested it with a two-core SMP and had no problem at all. Perhaps -tip testing could confirm it. Signed-off-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- arch/x86/kernel/ftrace.c | 43 +++++-------------------------------------- 1 file changed, 5 insertions(+), 38 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 16a571dea2ef..1db0e121a3e7 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -44,62 +44,37 @@ void ftrace_nmi_exit(void) atomic_dec(&in_nmi); } -/* - * Synchronize accesses to return adresses stack with - * interrupts. - */ -static raw_spinlock_t ret_stack_lock; - /* Add a function return address to the trace stack on thread info.*/ static int push_return_trace(unsigned long ret, unsigned long long time, unsigned long func) { int index; - struct thread_info *ti; - unsigned long flags; - int err = 0; - - raw_local_irq_save(flags); - __raw_spin_lock(&ret_stack_lock); + struct thread_info *ti = current_thread_info(); - ti = current_thread_info(); /* The return trace stack is full */ - if (ti->curr_ret_stack == FTRACE_RET_STACK_SIZE - 1) { - err = -EBUSY; - goto out; - } + if (ti->curr_ret_stack == FTRACE_RET_STACK_SIZE - 1) + return -EBUSY; index = ++ti->curr_ret_stack; ti->ret_stack[index].ret = ret; ti->ret_stack[index].func = func; ti->ret_stack[index].calltime = time; -out: - __raw_spin_unlock(&ret_stack_lock); - raw_local_irq_restore(flags); - return err; + return 0; } /* Retrieve a function return address to the trace stack on thread info.*/ static void pop_return_trace(unsigned long *ret, unsigned long long *time, unsigned long *func) { - struct thread_info *ti; int index; - unsigned long flags; - - raw_local_irq_save(flags); - __raw_spin_lock(&ret_stack_lock); - ti = current_thread_info(); + struct thread_info *ti = current_thread_info(); index = ti->curr_ret_stack; *ret = ti->ret_stack[index].ret; *func = ti->ret_stack[index].func; *time = ti->ret_stack[index].calltime; ti->curr_ret_stack--; - - __raw_spin_unlock(&ret_stack_lock); - raw_local_irq_restore(flags); } /* @@ -175,14 +150,6 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) *parent = old; } -static int __init init_ftrace_function_return(void) -{ - ret_stack_lock = (raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED; - return 0; -} -device_initcall(init_ftrace_function_return); - - #endif #ifdef CONFIG_DYNAMIC_FTRACE -- cgit v1.2.3 From 1dc1c6adf38bc5799d1594681645ced40ced4b6b Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 12 Nov 2008 22:49:23 +0100 Subject: tracing/function-return-tracer: call prepare_ftrace_return by registers Impact: Optimize a bit the function return tracer This patch changes the calling convention of prepare_ftrace_return to pass its arguments by register. This will optimize it a bit and prepare it to support dynamic tracing. Signed-off-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_32.S | 5 +---- arch/x86/kernel/ftrace.c | 1 - 2 files changed, 1 insertion(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S index 9a0ac85946db..f97621149839 100644 --- a/arch/x86/kernel/entry_32.S +++ b/arch/x86/kernel/entry_32.S @@ -1217,12 +1217,9 @@ trace_return: pushl %eax pushl %ecx pushl %edx - movl 0xc(%esp), %eax - pushl %eax + movl 0xc(%esp), %edx lea 0x4(%ebp), %eax - pushl %eax call prepare_ftrace_return - addl $8, %esp popl %edx popl %ecx popl %eax diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 1db0e121a3e7..fe832738e1e2 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -95,7 +95,6 @@ unsigned long ftrace_return_to_handler(void) * Hook the return address and push it in the stack of return addrs * in current thread info. */ -asmlinkage void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) { unsigned long old; -- cgit v1.2.3 From 350b4da71f8326b9319ada7b701f2bce2e1285b7 Mon Sep 17 00:00:00 2001 From: David Howells Date: Fri, 14 Nov 2008 10:38:40 +1100 Subject: CRED: Wrap task credential accesses in the x86 arch Wrap access to task credentials so that they can be separated more easily from the task_struct during the introduction of COW creds. Change most current->(|e|s|fs)[ug]id to current_(|e|s|fs)[ug]id(). Change some task->e?[ug]id to task_e?[ug]id(). In some places it makes more sense to use RCU directly rather than a convenient wrapper; these will be addressed by later patches. Signed-off-by: David Howells Reviewed-by: James Morris Acked-by: Serge Hallyn Cc: Thomas Gleixner Cc: Ingo Molnar Cc: H. Peter Anvin Signed-off-by: James Morris --- arch/x86/mm/fault.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c index 31e8730fa246..3a1b6ef4f05d 100644 --- a/arch/x86/mm/fault.c +++ b/arch/x86/mm/fault.c @@ -393,7 +393,7 @@ static void show_fault_oops(struct pt_regs *regs, unsigned long error_code, if (pte && pte_present(*pte) && !pte_exec(*pte)) printk(KERN_CRIT "kernel tried to execute " "NX-protected page - exploit attempt? " - "(uid: %d)\n", current->uid); + "(uid: %d)\n", current_uid()); } #endif -- cgit v1.2.3 From a6f76f23d297f70e2a6b3ec607f7aeeea9e37e8d Mon Sep 17 00:00:00 2001 From: David Howells Date: Fri, 14 Nov 2008 10:39:24 +1100 Subject: CRED: Make execve() take advantage of copy-on-write credentials Make execve() take advantage of copy-on-write credentials, allowing it to set up the credentials in advance, and then commit the whole lot after the point of no return. This patch and the preceding patches have been tested with the LTP SELinux testsuite. This patch makes several logical sets of alteration: (1) execve(). The credential bits from struct linux_binprm are, for the most part, replaced with a single credentials pointer (bprm->cred). This means that all the creds can be calculated in advance and then applied at the point of no return with no possibility of failure. I would like to replace bprm->cap_effective with: cap_isclear(bprm->cap_effective) but this seems impossible due to special behaviour for processes of pid 1 (they always retain their parent's capability masks where normally they'd be changed - see cap_bprm_set_creds()). The following sequence of events now happens: (a) At the start of do_execve, the current task's cred_exec_mutex is locked to prevent PTRACE_ATTACH from obsoleting the calculation of creds that we make. (a) prepare_exec_creds() is then called to make a copy of the current task's credentials and prepare it. This copy is then assigned to bprm->cred. This renders security_bprm_alloc() and security_bprm_free() unnecessary, and so they've been removed. (b) The determination of unsafe execution is now performed immediately after (a) rather than later on in the code. The result is stored in bprm->unsafe for future reference. (c) prepare_binprm() is called, possibly multiple times. (i) This applies the result of set[ug]id binaries to the new creds attached to bprm->cred. Personality bit clearance is recorded, but now deferred on the basis that the exec procedure may yet fail. (ii) This then calls the new security_bprm_set_creds(). This should calculate the new LSM and capability credentials into *bprm->cred. This folds together security_bprm_set() and parts of security_bprm_apply_creds() (these two have been removed). Anything that might fail must be done at this point. (iii) bprm->cred_prepared is set to 1. bprm->cred_prepared is 0 on the first pass of the security calculations, and 1 on all subsequent passes. This allows SELinux in (ii) to base its calculations only on the initial script and not on the interpreter. (d) flush_old_exec() is called to commit the task to execution. This performs the following steps with regard to credentials: (i) Clear pdeath_signal and set dumpable on certain circumstances that may not be covered by commit_creds(). (ii) Clear any bits in current->personality that were deferred from (c.i). (e) install_exec_creds() [compute_creds() as was] is called to install the new credentials. This performs the following steps with regard to credentials: (i) Calls security_bprm_committing_creds() to apply any security requirements, such as flushing unauthorised files in SELinux, that must be done before the credentials are changed. This is made up of bits of security_bprm_apply_creds() and security_bprm_post_apply_creds(), both of which have been removed. This function is not allowed to fail; anything that might fail must have been done in (c.ii). (ii) Calls commit_creds() to apply the new credentials in a single assignment (more or less). Possibly pdeath_signal and dumpable should be part of struct creds. (iii) Unlocks the task's cred_replace_mutex, thus allowing PTRACE_ATTACH to take place. (iv) Clears The bprm->cred pointer as the credentials it was holding are now immutable. (v) Calls security_bprm_committed_creds() to apply any security alterations that must be done after the creds have been changed. SELinux uses this to flush signals and signal handlers. (f) If an error occurs before (d.i), bprm_free() will call abort_creds() to destroy the proposed new credentials and will then unlock cred_replace_mutex. No changes to the credentials will have been made. (2) LSM interface. A number of functions have been changed, added or removed: (*) security_bprm_alloc(), ->bprm_alloc_security() (*) security_bprm_free(), ->bprm_free_security() Removed in favour of preparing new credentials and modifying those. (*) security_bprm_apply_creds(), ->bprm_apply_creds() (*) security_bprm_post_apply_creds(), ->bprm_post_apply_creds() Removed; split between security_bprm_set_creds(), security_bprm_committing_creds() and security_bprm_committed_creds(). (*) security_bprm_set(), ->bprm_set_security() Removed; folded into security_bprm_set_creds(). (*) security_bprm_set_creds(), ->bprm_set_creds() New. The new credentials in bprm->creds should be checked and set up as appropriate. bprm->cred_prepared is 0 on the first call, 1 on the second and subsequent calls. (*) security_bprm_committing_creds(), ->bprm_committing_creds() (*) security_bprm_committed_creds(), ->bprm_committed_creds() New. Apply the security effects of the new credentials. This includes closing unauthorised files in SELinux. This function may not fail. When the former is called, the creds haven't yet been applied to the process; when the latter is called, they have. The former may access bprm->cred, the latter may not. (3) SELinux. SELinux has a number of changes, in addition to those to support the LSM interface changes mentioned above: (a) The bprm_security_struct struct has been removed in favour of using the credentials-under-construction approach. (c) flush_unauthorized_files() now takes a cred pointer and passes it on to inode_has_perm(), file_has_perm() and dentry_open(). Signed-off-by: David Howells Acked-by: James Morris Acked-by: Serge Hallyn Signed-off-by: James Morris --- arch/x86/ia32/ia32_aout.c | 2 +- fs/binfmt_aout.c | 2 +- fs/binfmt_elf.c | 2 +- fs/binfmt_elf_fdpic.c | 2 +- fs/binfmt_flat.c | 2 +- fs/binfmt_som.c | 2 +- fs/compat.c | 42 +++--- fs/exec.c | 149 +++++++++++--------- fs/internal.h | 6 + include/linux/audit.h | 16 --- include/linux/binfmts.h | 16 ++- include/linux/cred.h | 3 +- include/linux/key.h | 2 - include/linux/security.h | 103 +++++--------- kernel/cred.c | 46 ++++++- security/capability.c | 19 +-- security/commoncap.c | 152 ++++++++++---------- security/keys/process_keys.c | 42 ------ security/root_plug.c | 13 +- security/security.c | 26 ++-- security/selinux/hooks.c | 283 ++++++++++++++++---------------------- security/selinux/include/objsec.h | 11 -- security/smack/smack_lsm.c | 3 +- 23 files changed, 429 insertions(+), 515 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/ia32/ia32_aout.c b/arch/x86/ia32/ia32_aout.c index 127ec3f07214..2a4d073d2cf1 100644 --- a/arch/x86/ia32/ia32_aout.c +++ b/arch/x86/ia32/ia32_aout.c @@ -327,7 +327,7 @@ static int load_aout_binary(struct linux_binprm *bprm, struct pt_regs *regs) current->mm->cached_hole_size = 0; current->mm->mmap = NULL; - compute_creds(bprm); + install_exec_creds(bprm); current->flags &= ~PF_FORKNOEXEC; if (N_MAGIC(ex) == OMAGIC) { diff --git a/fs/binfmt_aout.c b/fs/binfmt_aout.c index 204cfd1d7676..f1f3f4192a60 100644 --- a/fs/binfmt_aout.c +++ b/fs/binfmt_aout.c @@ -320,7 +320,7 @@ static int load_aout_binary(struct linux_binprm * bprm, struct pt_regs * regs) current->mm->free_area_cache = current->mm->mmap_base; current->mm->cached_hole_size = 0; - compute_creds(bprm); + install_exec_creds(bprm); current->flags &= ~PF_FORKNOEXEC; #ifdef __sparc__ if (N_MAGIC(ex) == NMAGIC) { diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c index 9142ff5dc8e6..f458c1217c5e 100644 --- a/fs/binfmt_elf.c +++ b/fs/binfmt_elf.c @@ -956,7 +956,7 @@ static int load_elf_binary(struct linux_binprm *bprm, struct pt_regs *regs) } #endif /* ARCH_HAS_SETUP_ADDITIONAL_PAGES */ - compute_creds(bprm); + install_exec_creds(bprm); current->flags &= ~PF_FORKNOEXEC; retval = create_elf_tables(bprm, &loc->elf_ex, load_addr, interp_load_addr); diff --git a/fs/binfmt_elf_fdpic.c b/fs/binfmt_elf_fdpic.c index 45dabd59936f..aa5b43205e37 100644 --- a/fs/binfmt_elf_fdpic.c +++ b/fs/binfmt_elf_fdpic.c @@ -404,7 +404,7 @@ static int load_elf_fdpic_binary(struct linux_binprm *bprm, current->mm->start_stack = current->mm->start_brk + stack_size; #endif - compute_creds(bprm); + install_exec_creds(bprm); current->flags &= ~PF_FORKNOEXEC; if (create_elf_fdpic_tables(bprm, current->mm, &exec_params, &interp_params) < 0) diff --git a/fs/binfmt_flat.c b/fs/binfmt_flat.c index ccb781a6a804..7bbd5c6b3725 100644 --- a/fs/binfmt_flat.c +++ b/fs/binfmt_flat.c @@ -880,7 +880,7 @@ static int load_flat_binary(struct linux_binprm * bprm, struct pt_regs * regs) (libinfo.lib_list[j].loaded)? libinfo.lib_list[j].start_data:UNLOADED_LIB; - compute_creds(bprm); + install_exec_creds(bprm); current->flags &= ~PF_FORKNOEXEC; set_binfmt(&flat_format); diff --git a/fs/binfmt_som.c b/fs/binfmt_som.c index 74e587a52796..08644a61616e 100644 --- a/fs/binfmt_som.c +++ b/fs/binfmt_som.c @@ -255,7 +255,7 @@ load_som_binary(struct linux_binprm * bprm, struct pt_regs * regs) kfree(hpuxhdr); set_binfmt(&som_format); - compute_creds(bprm); + install_exec_creds(bprm); setup_arg_pages(bprm, STACK_TOP, EXSTACK_DEFAULT); create_som_tables(bprm); diff --git a/fs/compat.c b/fs/compat.c index e5f49f538502..d1ece79b6411 100644 --- a/fs/compat.c +++ b/fs/compat.c @@ -1393,10 +1393,20 @@ int compat_do_execve(char * filename, if (!bprm) goto out_ret; + retval = mutex_lock_interruptible(¤t->cred_exec_mutex); + if (retval < 0) + goto out_free; + + retval = -ENOMEM; + bprm->cred = prepare_exec_creds(); + if (!bprm->cred) + goto out_unlock; + check_unsafe_exec(bprm); + file = open_exec(filename); retval = PTR_ERR(file); if (IS_ERR(file)) - goto out_kfree; + goto out_unlock; sched_exec(); @@ -1410,14 +1420,10 @@ int compat_do_execve(char * filename, bprm->argc = compat_count(argv, MAX_ARG_STRINGS); if ((retval = bprm->argc) < 0) - goto out_mm; + goto out; bprm->envc = compat_count(envp, MAX_ARG_STRINGS); if ((retval = bprm->envc) < 0) - goto out_mm; - - retval = security_bprm_alloc(bprm); - if (retval) goto out; retval = prepare_binprm(bprm); @@ -1438,19 +1444,16 @@ int compat_do_execve(char * filename, goto out; retval = search_binary_handler(bprm, regs); - if (retval >= 0) { - /* execve success */ - security_bprm_free(bprm); - acct_update_integrals(current); - free_bprm(bprm); - return retval; - } + if (retval < 0) + goto out; -out: - if (bprm->security) - security_bprm_free(bprm); + /* execve succeeded */ + mutex_unlock(¤t->cred_exec_mutex); + acct_update_integrals(current); + free_bprm(bprm); + return retval; -out_mm: +out: if (bprm->mm) mmput(bprm->mm); @@ -1460,7 +1463,10 @@ out_file: fput(bprm->file); } -out_kfree: +out_unlock: + mutex_unlock(¤t->cred_exec_mutex); + +out_free: free_bprm(bprm); out_ret: diff --git a/fs/exec.c b/fs/exec.c index 9bd3559ddece..32f13e299417 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -55,6 +55,7 @@ #include #include #include +#include "internal.h" #ifdef __alpha__ /* for /sbin/loader handling in search_binary_handler() */ @@ -1007,15 +1008,17 @@ int flush_old_exec(struct linux_binprm * bprm) */ current->mm->task_size = TASK_SIZE; - if (bprm->e_uid != current_euid() || - bprm->e_gid != current_egid()) { - set_dumpable(current->mm, suid_dumpable); + /* install the new credentials */ + if (bprm->cred->uid != current_euid() || + bprm->cred->gid != current_egid()) { current->pdeath_signal = 0; } else if (file_permission(bprm->file, MAY_READ) || - (bprm->interp_flags & BINPRM_FLAGS_ENFORCE_NONDUMP)) { + bprm->interp_flags & BINPRM_FLAGS_ENFORCE_NONDUMP) { set_dumpable(current->mm, suid_dumpable); } + current->personality &= ~bprm->per_clear; + /* An exec changes our domain. We are no longer part of the thread group */ @@ -1032,13 +1035,50 @@ out: EXPORT_SYMBOL(flush_old_exec); +/* + * install the new credentials for this executable + */ +void install_exec_creds(struct linux_binprm *bprm) +{ + security_bprm_committing_creds(bprm); + + commit_creds(bprm->cred); + bprm->cred = NULL; + + /* cred_exec_mutex must be held at least to this point to prevent + * ptrace_attach() from altering our determination of the task's + * credentials; any time after this it may be unlocked */ + + security_bprm_committed_creds(bprm); +} +EXPORT_SYMBOL(install_exec_creds); + +/* + * determine how safe it is to execute the proposed program + * - the caller must hold current->cred_exec_mutex to protect against + * PTRACE_ATTACH + */ +void check_unsafe_exec(struct linux_binprm *bprm) +{ + struct task_struct *p = current; + + bprm->unsafe = tracehook_unsafe_exec(p); + + if (atomic_read(&p->fs->count) > 1 || + atomic_read(&p->files->count) > 1 || + atomic_read(&p->sighand->count) > 1) + bprm->unsafe |= LSM_UNSAFE_SHARE; +} + /* * Fill the binprm structure from the inode. * Check permissions, then read the first 128 (BINPRM_BUF_SIZE) bytes + * + * This may be called multiple times for binary chains (scripts for example). */ int prepare_binprm(struct linux_binprm *bprm) { - int mode; + umode_t mode; struct inode * inode = bprm->file->f_path.dentry->d_inode; int retval; @@ -1046,14 +1086,15 @@ int prepare_binprm(struct linux_binprm *bprm) if (bprm->file->f_op == NULL) return -EACCES; - bprm->e_uid = current_euid(); - bprm->e_gid = current_egid(); + /* clear any previous set[ug]id data from a previous binary */ + bprm->cred->euid = current_euid(); + bprm->cred->egid = current_egid(); - if(!(bprm->file->f_path.mnt->mnt_flags & MNT_NOSUID)) { + if (!(bprm->file->f_path.mnt->mnt_flags & MNT_NOSUID)) { /* Set-uid? */ if (mode & S_ISUID) { - current->personality &= ~PER_CLEAR_ON_SETID; - bprm->e_uid = inode->i_uid; + bprm->per_clear |= PER_CLEAR_ON_SETID; + bprm->cred->euid = inode->i_uid; } /* Set-gid? */ @@ -1063,50 +1104,23 @@ int prepare_binprm(struct linux_binprm *bprm) * executable. */ if ((mode & (S_ISGID | S_IXGRP)) == (S_ISGID | S_IXGRP)) { - current->personality &= ~PER_CLEAR_ON_SETID; - bprm->e_gid = inode->i_gid; + bprm->per_clear |= PER_CLEAR_ON_SETID; + bprm->cred->egid = inode->i_gid; } } /* fill in binprm security blob */ - retval = security_bprm_set(bprm); + retval = security_bprm_set_creds(bprm); if (retval) return retval; + bprm->cred_prepared = 1; - memset(bprm->buf,0,BINPRM_BUF_SIZE); - return kernel_read(bprm->file,0,bprm->buf,BINPRM_BUF_SIZE); + memset(bprm->buf, 0, BINPRM_BUF_SIZE); + return kernel_read(bprm->file, 0, bprm->buf, BINPRM_BUF_SIZE); } EXPORT_SYMBOL(prepare_binprm); -static int unsafe_exec(struct task_struct *p) -{ - int unsafe = tracehook_unsafe_exec(p); - - if (atomic_read(&p->fs->count) > 1 || - atomic_read(&p->files->count) > 1 || - atomic_read(&p->sighand->count) > 1) - unsafe |= LSM_UNSAFE_SHARE; - - return unsafe; -} - -void compute_creds(struct linux_binprm *bprm) -{ - int unsafe; - - if (bprm->e_uid != current_uid()) - current->pdeath_signal = 0; - exec_keys(current); - - task_lock(current); - unsafe = unsafe_exec(current); - security_bprm_apply_creds(bprm, unsafe); - task_unlock(current); - security_bprm_post_apply_creds(bprm); -} -EXPORT_SYMBOL(compute_creds); - /* * Arguments are '\0' separated strings found at the location bprm->p * points to; chop off the first by relocating brpm->p to right after @@ -1259,6 +1273,8 @@ EXPORT_SYMBOL(search_binary_handler); void free_bprm(struct linux_binprm *bprm) { free_arg_pages(bprm); + if (bprm->cred) + abort_creds(bprm->cred); kfree(bprm); } @@ -1284,10 +1300,20 @@ int do_execve(char * filename, if (!bprm) goto out_files; + retval = mutex_lock_interruptible(¤t->cred_exec_mutex); + if (retval < 0) + goto out_free; + + retval = -ENOMEM; + bprm->cred = prepare_exec_creds(); + if (!bprm->cred) + goto out_unlock; + check_unsafe_exec(bprm); + file = open_exec(filename); retval = PTR_ERR(file); if (IS_ERR(file)) - goto out_kfree; + goto out_unlock; sched_exec(); @@ -1301,14 +1327,10 @@ int do_execve(char * filename, bprm->argc = count(argv, MAX_ARG_STRINGS); if ((retval = bprm->argc) < 0) - goto out_mm; + goto out; bprm->envc = count(envp, MAX_ARG_STRINGS); if ((retval = bprm->envc) < 0) - goto out_mm; - - retval = security_bprm_alloc(bprm); - if (retval) goto out; retval = prepare_binprm(bprm); @@ -1330,21 +1352,18 @@ int do_execve(char * filename, current->flags &= ~PF_KTHREAD; retval = search_binary_handler(bprm,regs); - if (retval >= 0) { - /* execve success */ - security_bprm_free(bprm); - acct_update_integrals(current); - free_bprm(bprm); - if (displaced) - put_files_struct(displaced); - return retval; - } + if (retval < 0) + goto out; -out: - if (bprm->security) - security_bprm_free(bprm); + /* execve succeeded */ + mutex_unlock(¤t->cred_exec_mutex); + acct_update_integrals(current); + free_bprm(bprm); + if (displaced) + put_files_struct(displaced); + return retval; -out_mm: +out: if (bprm->mm) mmput (bprm->mm); @@ -1353,7 +1372,11 @@ out_file: allow_write_access(bprm->file); fput(bprm->file); } -out_kfree: + +out_unlock: + mutex_unlock(¤t->cred_exec_mutex); + +out_free: free_bprm(bprm); out_files: diff --git a/fs/internal.h b/fs/internal.h index 80aa9a023372..53af885f1732 100644 --- a/fs/internal.h +++ b/fs/internal.h @@ -10,6 +10,7 @@ */ struct super_block; +struct linux_binprm; /* * block_dev.c @@ -39,6 +40,11 @@ static inline int sb_is_blkdev_sb(struct super_block *sb) */ extern void __init chrdev_init(void); +/* + * exec.c + */ +extern void check_unsafe_exec(struct linux_binprm *); + /* * namespace.c */ diff --git a/include/linux/audit.h b/include/linux/audit.h index 0b2fcb698a63..e8ce2c4c7ac7 100644 --- a/include/linux/audit.h +++ b/include/linux/audit.h @@ -508,22 +508,6 @@ static inline int audit_mq_getsetattr(mqd_t mqdes, struct mq_attr *mqstat) return 0; } -/* - * ieieeeeee, an audit function without a return code! - * - * This function might fail! I decided that it didn't matter. We are too late - * to fail the syscall and the information isn't REQUIRED for any purpose. It's - * just nice to have. We should be able to look at past audit logs to figure - * out this process's current cap set along with the fcaps from the PATH record - * and use that to come up with the final set. Yeah, its ugly, but all the info - * is still in the audit log. So I'm not going to bother mentioning we failed - * if we couldn't allocate memory. - * - * If someone changes their mind they could create the aux record earlier and - * then search here and use that earlier allocation. But I don't wanna. - * - * -Eric - */ static inline int audit_log_bprm_fcaps(struct linux_binprm *bprm, const struct cred *new, const struct cred *old) diff --git a/include/linux/binfmts.h b/include/linux/binfmts.h index 7394b5b349ff..6cbfbe297180 100644 --- a/include/linux/binfmts.h +++ b/include/linux/binfmts.h @@ -35,16 +35,20 @@ struct linux_binprm{ struct mm_struct *mm; unsigned long p; /* current top of mem */ unsigned int sh_bang:1, - misc_bang:1; + misc_bang:1, + cred_prepared:1,/* true if creds already prepared (multiple + * preps happen for interpreters) */ + cap_effective:1;/* true if has elevated effective capabilities, + * false if not; except for init which inherits + * its parent's caps anyway */ #ifdef __alpha__ unsigned int taso:1; #endif unsigned int recursion_depth; struct file * file; - int e_uid, e_gid; - kernel_cap_t cap_post_exec_permitted; - bool cap_effective; - void *security; + struct cred *cred; /* new credentials */ + int unsafe; /* how unsafe this exec is (mask of LSM_UNSAFE_*) */ + unsigned int per_clear; /* bits to clear in current->personality */ int argc, envc; char * filename; /* Name of binary as seen by procps */ char * interp; /* Name of the binary really executed. Most @@ -101,7 +105,7 @@ extern int setup_arg_pages(struct linux_binprm * bprm, int executable_stack); extern int bprm_mm_init(struct linux_binprm *bprm); extern int copy_strings_kernel(int argc,char ** argv,struct linux_binprm *bprm); -extern void compute_creds(struct linux_binprm *binprm); +extern void install_exec_creds(struct linux_binprm *bprm); extern int do_coredump(long signr, int exit_code, struct pt_regs * regs); extern int set_binfmt(struct linux_binfmt *new); extern void free_bprm(struct linux_binprm *); diff --git a/include/linux/cred.h b/include/linux/cred.h index eaf6fa695a04..8edb4d1d5427 100644 --- a/include/linux/cred.h +++ b/include/linux/cred.h @@ -84,8 +84,6 @@ struct thread_group_cred { struct key *process_keyring; /* keyring private to this process */ struct rcu_head rcu; /* RCU deletion hook */ }; - -extern void release_tgcred(struct cred *cred); #endif /* @@ -144,6 +142,7 @@ struct cred { extern void __put_cred(struct cred *); extern int copy_creds(struct task_struct *, unsigned long); extern struct cred *prepare_creds(void); +extern struct cred *prepare_exec_creds(void); extern struct cred *prepare_usermodehelper_creds(void); extern int commit_creds(struct cred *); extern void abort_creds(struct cred *); diff --git a/include/linux/key.h b/include/linux/key.h index 69ecf0934b02..21d32a142c00 100644 --- a/include/linux/key.h +++ b/include/linux/key.h @@ -278,7 +278,6 @@ extern ctl_table key_sysctls[]; * the userspace interface */ extern int install_thread_keyring_to_cred(struct cred *cred); -extern int exec_keys(struct task_struct *tsk); extern void key_fsuid_changed(struct task_struct *tsk); extern void key_fsgid_changed(struct task_struct *tsk); extern void key_init(void); @@ -294,7 +293,6 @@ extern void key_init(void); #define make_key_ref(k, p) NULL #define key_ref_to_ptr(k) NULL #define is_key_possessed(k) 0 -#define exec_keys(t) do { } while(0) #define key_fsuid_changed(t) do { } while(0) #define key_fsgid_changed(t) do { } while(0) #define key_init() do { } while(0) diff --git a/include/linux/security.h b/include/linux/security.h index 68be11251447..56a0eed65673 100644 --- a/include/linux/security.h +++ b/include/linux/security.h @@ -57,8 +57,7 @@ extern int cap_capset(struct cred *new, const struct cred *old, const kernel_cap_t *effective, const kernel_cap_t *inheritable, const kernel_cap_t *permitted); -extern int cap_bprm_set_security(struct linux_binprm *bprm); -extern int cap_bprm_apply_creds(struct linux_binprm *bprm, int unsafe); +extern int cap_bprm_set_creds(struct linux_binprm *bprm); extern int cap_bprm_secureexec(struct linux_binprm *bprm); extern int cap_inode_setxattr(struct dentry *dentry, const char *name, const void *value, size_t size, int flags); @@ -110,7 +109,7 @@ extern unsigned long mmap_min_addr; struct sched_param; struct request_sock; -/* bprm_apply_creds unsafe reasons */ +/* bprm->unsafe reasons */ #define LSM_UNSAFE_SHARE 1 #define LSM_UNSAFE_PTRACE 2 #define LSM_UNSAFE_PTRACE_CAP 4 @@ -154,36 +153,7 @@ static inline void security_free_mnt_opts(struct security_mnt_opts *opts) * * Security hooks for program execution operations. * - * @bprm_alloc_security: - * Allocate and attach a security structure to the @bprm->security field. - * The security field is initialized to NULL when the bprm structure is - * allocated. - * @bprm contains the linux_binprm structure to be modified. - * Return 0 if operation was successful. - * @bprm_free_security: - * @bprm contains the linux_binprm structure to be modified. - * Deallocate and clear the @bprm->security field. - * @bprm_apply_creds: - * Compute and set the security attributes of a process being transformed - * by an execve operation based on the old attributes (current->security) - * and the information saved in @bprm->security by the set_security hook. - * Since this function may return an error, in which case the process will - * be killed. However, it can leave the security attributes of the - * process unchanged if an access failure occurs at this point. - * bprm_apply_creds is called under task_lock. @unsafe indicates various - * reasons why it may be unsafe to change security state. - * @bprm contains the linux_binprm structure. - * @bprm_post_apply_creds: - * Runs after bprm_apply_creds with the task_lock dropped, so that - * functions which cannot be called safely under the task_lock can - * be used. This hook is a good place to perform state changes on - * the process such as closing open file descriptors to which access - * is no longer granted if the attributes were changed. - * Note that a security module might need to save state between - * bprm_apply_creds and bprm_post_apply_creds to store the decision - * on whether the process may proceed. - * @bprm contains the linux_binprm structure. - * @bprm_set_security: + * @bprm_set_creds: * Save security information in the bprm->security field, typically based * on information about the bprm->file, for later use by the apply_creds * hook. This hook may also optionally check permissions (e.g. for @@ -196,15 +166,30 @@ static inline void security_free_mnt_opts(struct security_mnt_opts *opts) * @bprm contains the linux_binprm structure. * Return 0 if the hook is successful and permission is granted. * @bprm_check_security: - * This hook mediates the point when a search for a binary handler will - * begin. It allows a check the @bprm->security value which is set in - * the preceding set_security call. The primary difference from - * set_security is that the argv list and envp list are reliably - * available in @bprm. This hook may be called multiple times - * during a single execve; and in each pass set_security is called - * first. + * This hook mediates the point when a search for a binary handler will + * begin. It allows a check the @bprm->security value which is set in the + * preceding set_creds call. The primary difference from set_creds is + * that the argv list and envp list are reliably available in @bprm. This + * hook may be called multiple times during a single execve; and in each + * pass set_creds is called first. * @bprm contains the linux_binprm structure. * Return 0 if the hook is successful and permission is granted. + * @bprm_committing_creds: + * Prepare to install the new security attributes of a process being + * transformed by an execve operation, based on the old credentials + * pointed to by @current->cred and the information set in @bprm->cred by + * the bprm_set_creds hook. @bprm points to the linux_binprm structure. + * This hook is a good place to perform state changes on the process such + * as closing open file descriptors to which access will no longer be + * granted when the attributes are changed. This is called immediately + * before commit_creds(). + * @bprm_committed_creds: + * Tidy up after the installation of the new security attributes of a + * process being transformed by an execve operation. The new credentials + * have, by this point, been set to @current->cred. @bprm points to the + * linux_binprm structure. This hook is a good place to perform state + * changes on the process such as clearing out non-inheritable signal + * state. This is called immediately after commit_creds(). * @bprm_secureexec: * Return a boolean value (0 or 1) indicating whether a "secure exec" * is required. The flag is passed in the auxiliary table @@ -1301,13 +1286,11 @@ struct security_operations { int (*settime) (struct timespec *ts, struct timezone *tz); int (*vm_enough_memory) (struct mm_struct *mm, long pages); - int (*bprm_alloc_security) (struct linux_binprm *bprm); - void (*bprm_free_security) (struct linux_binprm *bprm); - int (*bprm_apply_creds) (struct linux_binprm *bprm, int unsafe); - void (*bprm_post_apply_creds) (struct linux_binprm *bprm); - int (*bprm_set_security) (struct linux_binprm *bprm); + int (*bprm_set_creds) (struct linux_binprm *bprm); int (*bprm_check_security) (struct linux_binprm *bprm); int (*bprm_secureexec) (struct linux_binprm *bprm); + void (*bprm_committing_creds) (struct linux_binprm *bprm); + void (*bprm_committed_creds) (struct linux_binprm *bprm); int (*sb_alloc_security) (struct super_block *sb); void (*sb_free_security) (struct super_block *sb); @@ -1569,12 +1552,10 @@ int security_settime(struct timespec *ts, struct timezone *tz); int security_vm_enough_memory(long pages); int security_vm_enough_memory_mm(struct mm_struct *mm, long pages); int security_vm_enough_memory_kern(long pages); -int security_bprm_alloc(struct linux_binprm *bprm); -void security_bprm_free(struct linux_binprm *bprm); -int security_bprm_apply_creds(struct linux_binprm *bprm, int unsafe); -void security_bprm_post_apply_creds(struct linux_binprm *bprm); -int security_bprm_set(struct linux_binprm *bprm); +int security_bprm_set_creds(struct linux_binprm *bprm); int security_bprm_check(struct linux_binprm *bprm); +void security_bprm_committing_creds(struct linux_binprm *bprm); +void security_bprm_committed_creds(struct linux_binprm *bprm); int security_bprm_secureexec(struct linux_binprm *bprm); int security_sb_alloc(struct super_block *sb); void security_sb_free(struct super_block *sb); @@ -1812,32 +1793,22 @@ static inline int security_vm_enough_memory_mm(struct mm_struct *mm, long pages) return cap_vm_enough_memory(mm, pages); } -static inline int security_bprm_alloc(struct linux_binprm *bprm) -{ - return 0; -} - -static inline void security_bprm_free(struct linux_binprm *bprm) -{ } - -static inline int security_bprm_apply_creds(struct linux_binprm *bprm, int unsafe) +static inline int security_bprm_set_creds(struct linux_binprm *bprm) { - return cap_bprm_apply_creds(bprm, unsafe); + return cap_bprm_set_creds(bprm); } -static inline void security_bprm_post_apply_creds(struct linux_binprm *bprm) +static inline int security_bprm_check(struct linux_binprm *bprm) { - return; + return 0; } -static inline int security_bprm_set(struct linux_binprm *bprm) +static inline void security_bprm_committing_creds(struct linux_binprm *bprm) { - return cap_bprm_set_security(bprm); } -static inline int security_bprm_check(struct linux_binprm *bprm) +static inline void security_bprm_committed_creds(struct linux_binprm *bprm) { - return 0; } static inline int security_bprm_secureexec(struct linux_binprm *bprm) diff --git a/kernel/cred.c b/kernel/cred.c index cb6b5eda978d..e6fcdd67b2ec 100644 --- a/kernel/cred.c +++ b/kernel/cred.c @@ -68,7 +68,7 @@ static void release_tgcred_rcu(struct rcu_head *rcu) /* * Release a set of thread group credentials. */ -void release_tgcred(struct cred *cred) +static void release_tgcred(struct cred *cred) { #ifdef CONFIG_KEYS struct thread_group_cred *tgcred = cred->tgcred; @@ -163,6 +163,50 @@ error: } EXPORT_SYMBOL(prepare_creds); +/* + * Prepare credentials for current to perform an execve() + * - The caller must hold current->cred_exec_mutex + */ +struct cred *prepare_exec_creds(void) +{ + struct thread_group_cred *tgcred = NULL; + struct cred *new; + +#ifdef CONFIG_KEYS + tgcred = kmalloc(sizeof(*tgcred), GFP_KERNEL); + if (!tgcred) + return NULL; +#endif + + new = prepare_creds(); + if (!new) { + kfree(tgcred); + return new; + } + +#ifdef CONFIG_KEYS + /* newly exec'd tasks don't get a thread keyring */ + key_put(new->thread_keyring); + new->thread_keyring = NULL; + + /* create a new per-thread-group creds for all this set of threads to + * share */ + memcpy(tgcred, new->tgcred, sizeof(struct thread_group_cred)); + + atomic_set(&tgcred->usage, 1); + spin_lock_init(&tgcred->lock); + + /* inherit the session keyring; new process keyring */ + key_get(tgcred->session_keyring); + tgcred->process_keyring = NULL; + + release_tgcred(new); + new->tgcred = tgcred; +#endif + + return new; +} + /* * prepare new credentials for the usermode helper dispatcher */ diff --git a/security/capability.c b/security/capability.c index efeb6d9e0e6a..185804f99ad1 100644 --- a/security/capability.c +++ b/security/capability.c @@ -32,24 +32,19 @@ static int cap_quota_on(struct dentry *dentry) return 0; } -static int cap_bprm_alloc_security(struct linux_binprm *bprm) +static int cap_bprm_check_security (struct linux_binprm *bprm) { return 0; } -static void cap_bprm_free_security(struct linux_binprm *bprm) +static void cap_bprm_committing_creds(struct linux_binprm *bprm) { } -static void cap_bprm_post_apply_creds(struct linux_binprm *bprm) +static void cap_bprm_committed_creds(struct linux_binprm *bprm) { } -static int cap_bprm_check_security(struct linux_binprm *bprm) -{ - return 0; -} - static int cap_sb_alloc_security(struct super_block *sb) { return 0; @@ -827,11 +822,9 @@ void security_fixup_ops(struct security_operations *ops) set_to_cap_if_null(ops, syslog); set_to_cap_if_null(ops, settime); set_to_cap_if_null(ops, vm_enough_memory); - set_to_cap_if_null(ops, bprm_alloc_security); - set_to_cap_if_null(ops, bprm_free_security); - set_to_cap_if_null(ops, bprm_apply_creds); - set_to_cap_if_null(ops, bprm_post_apply_creds); - set_to_cap_if_null(ops, bprm_set_security); + set_to_cap_if_null(ops, bprm_set_creds); + set_to_cap_if_null(ops, bprm_committing_creds); + set_to_cap_if_null(ops, bprm_committed_creds); set_to_cap_if_null(ops, bprm_check_security); set_to_cap_if_null(ops, bprm_secureexec); set_to_cap_if_null(ops, sb_alloc_security); diff --git a/security/commoncap.c b/security/commoncap.c index b5419273f92d..51dfa11e8e56 100644 --- a/security/commoncap.c +++ b/security/commoncap.c @@ -167,7 +167,7 @@ int cap_capset(struct cred *new, static inline void bprm_clear_caps(struct linux_binprm *bprm) { - cap_clear(bprm->cap_post_exec_permitted); + cap_clear(bprm->cred->cap_permitted); bprm->cap_effective = false; } @@ -198,15 +198,15 @@ int cap_inode_killpriv(struct dentry *dentry) } static inline int bprm_caps_from_vfs_caps(struct cpu_vfs_cap_data *caps, - struct linux_binprm *bprm) + struct linux_binprm *bprm, + bool *effective) { + struct cred *new = bprm->cred; unsigned i; int ret = 0; if (caps->magic_etc & VFS_CAP_FLAGS_EFFECTIVE) - bprm->cap_effective = true; - else - bprm->cap_effective = false; + *effective = true; CAP_FOR_EACH_U32(i) { __u32 permitted = caps->permitted.cap[i]; @@ -215,16 +215,13 @@ static inline int bprm_caps_from_vfs_caps(struct cpu_vfs_cap_data *caps, /* * pP' = (X & fP) | (pI & fI) */ - bprm->cap_post_exec_permitted.cap[i] = - (current->cred->cap_bset.cap[i] & permitted) | - (current->cred->cap_inheritable.cap[i] & inheritable); + new->cap_permitted.cap[i] = + (new->cap_bset.cap[i] & permitted) | + (new->cap_inheritable.cap[i] & inheritable); - if (permitted & ~bprm->cap_post_exec_permitted.cap[i]) { - /* - * insufficient to execute correctly - */ + if (permitted & ~new->cap_permitted.cap[i]) + /* insufficient to execute correctly */ ret = -EPERM; - } } /* @@ -232,7 +229,7 @@ static inline int bprm_caps_from_vfs_caps(struct cpu_vfs_cap_data *caps, * do not have enough capabilities, we return an error if they are * missing some "forced" (aka file-permitted) capabilities. */ - return bprm->cap_effective ? ret : 0; + return *effective ? ret : 0; } int get_vfs_caps_from_disk(const struct dentry *dentry, struct cpu_vfs_cap_data *cpu_caps) @@ -250,10 +247,9 @@ int get_vfs_caps_from_disk(const struct dentry *dentry, struct cpu_vfs_cap_data size = inode->i_op->getxattr((struct dentry *)dentry, XATTR_NAME_CAPS, &caps, XATTR_CAPS_SZ); - if (size == -ENODATA || size == -EOPNOTSUPP) { + if (size == -ENODATA || size == -EOPNOTSUPP) /* no data, that's ok */ return -ENODATA; - } if (size < 0) return size; @@ -262,7 +258,7 @@ int get_vfs_caps_from_disk(const struct dentry *dentry, struct cpu_vfs_cap_data cpu_caps->magic_etc = magic_etc = le32_to_cpu(caps.magic_etc); - switch ((magic_etc & VFS_CAP_REVISION_MASK)) { + switch (magic_etc & VFS_CAP_REVISION_MASK) { case VFS_CAP_REVISION_1: if (size != XATTR_CAPS_SZ_1) return -EINVAL; @@ -283,11 +279,12 @@ int get_vfs_caps_from_disk(const struct dentry *dentry, struct cpu_vfs_cap_data cpu_caps->permitted.cap[i] = le32_to_cpu(caps.data[i].permitted); cpu_caps->inheritable.cap[i] = le32_to_cpu(caps.data[i].inheritable); } + return 0; } /* Locate any VFS capabilities: */ -static int get_file_caps(struct linux_binprm *bprm) +static int get_file_caps(struct linux_binprm *bprm, bool *effective) { struct dentry *dentry; int rc = 0; @@ -313,7 +310,10 @@ static int get_file_caps(struct linux_binprm *bprm) goto out; } - rc = bprm_caps_from_vfs_caps(&vcaps, bprm); + rc = bprm_caps_from_vfs_caps(&vcaps, bprm, effective); + if (rc == -EINVAL) + printk(KERN_NOTICE "%s: cap_from_disk returned %d for %s\n", + __func__, rc, bprm->filename); out: dput(dentry); @@ -334,18 +334,27 @@ int cap_inode_killpriv(struct dentry *dentry) return 0; } -static inline int get_file_caps(struct linux_binprm *bprm) +static inline int get_file_caps(struct linux_binprm *bprm, bool *effective) { bprm_clear_caps(bprm); return 0; } #endif -int cap_bprm_set_security (struct linux_binprm *bprm) +/* + * set up the new credentials for an exec'd task + */ +int cap_bprm_set_creds(struct linux_binprm *bprm) { + const struct cred *old = current_cred(); + struct cred *new = bprm->cred; + bool effective; int ret; - ret = get_file_caps(bprm); + effective = false; + ret = get_file_caps(bprm, &effective); + if (ret < 0) + return ret; if (!issecure(SECURE_NOROOT)) { /* @@ -353,63 +362,47 @@ int cap_bprm_set_security (struct linux_binprm *bprm) * executables under compatibility mode, we override the * capability sets for the file. * - * If only the real uid is 0, we do not set the effective - * bit. + * If only the real uid is 0, we do not set the effective bit. */ - if (bprm->e_uid == 0 || current_uid() == 0) { + if (new->euid == 0 || new->uid == 0) { /* pP' = (cap_bset & ~0) | (pI & ~0) */ - bprm->cap_post_exec_permitted = cap_combine( - current->cred->cap_bset, - current->cred->cap_inheritable); - bprm->cap_effective = (bprm->e_uid == 0); - ret = 0; + new->cap_permitted = cap_combine(old->cap_bset, + old->cap_inheritable); } + if (new->euid == 0) + effective = true; } - return ret; -} - -int cap_bprm_apply_creds (struct linux_binprm *bprm, int unsafe) -{ - const struct cred *old = current_cred(); - struct cred *new; - - new = prepare_creds(); - if (!new) - return -ENOMEM; - - if (bprm->e_uid != old->uid || bprm->e_gid != old->gid || - !cap_issubset(bprm->cap_post_exec_permitted, - old->cap_permitted)) { - set_dumpable(current->mm, suid_dumpable); - current->pdeath_signal = 0; - - if (unsafe & ~LSM_UNSAFE_PTRACE_CAP) { - if (!capable(CAP_SETUID)) { - bprm->e_uid = old->uid; - bprm->e_gid = old->gid; - } - if (cap_limit_ptraced_target()) { - bprm->cap_post_exec_permitted = cap_intersect( - bprm->cap_post_exec_permitted, - new->cap_permitted); - } + /* Don't let someone trace a set[ug]id/setpcap binary with the revised + * credentials unless they have the appropriate permit + */ + if ((new->euid != old->uid || + new->egid != old->gid || + !cap_issubset(new->cap_permitted, old->cap_permitted)) && + bprm->unsafe & ~LSM_UNSAFE_PTRACE_CAP) { + /* downgrade; they get no more than they had, and maybe less */ + if (!capable(CAP_SETUID)) { + new->euid = new->uid; + new->egid = new->gid; } + if (cap_limit_ptraced_target()) + new->cap_permitted = cap_intersect(new->cap_permitted, + old->cap_permitted); } - new->suid = new->euid = new->fsuid = bprm->e_uid; - new->sgid = new->egid = new->fsgid = bprm->e_gid; + new->suid = new->fsuid = new->euid; + new->sgid = new->fsgid = new->egid; - /* For init, we want to retain the capabilities set - * in the init_task struct. Thus we skip the usual - * capability rules */ + /* For init, we want to retain the capabilities set in the initial + * task. Thus we skip the usual capability rules + */ if (!is_global_init(current)) { - new->cap_permitted = bprm->cap_post_exec_permitted; - if (bprm->cap_effective) - new->cap_effective = bprm->cap_post_exec_permitted; + if (effective) + new->cap_effective = new->cap_permitted; else cap_clear(new->cap_effective); } + bprm->cap_effective = effective; /* * Audit candidate if current->cap_effective is set @@ -425,23 +418,31 @@ int cap_bprm_apply_creds (struct linux_binprm *bprm, int unsafe) */ if (!cap_isclear(new->cap_effective)) { if (!cap_issubset(CAP_FULL_SET, new->cap_effective) || - bprm->e_uid != 0 || new->uid != 0 || - issecure(SECURE_NOROOT)) - audit_log_bprm_fcaps(bprm, new, old); + new->euid != 0 || new->uid != 0 || + issecure(SECURE_NOROOT)) { + ret = audit_log_bprm_fcaps(bprm, new, old); + if (ret < 0) + return ret; + } } new->securebits &= ~issecure_mask(SECURE_KEEP_CAPS); - return commit_creds(new); + return 0; } -int cap_bprm_secureexec (struct linux_binprm *bprm) +/* + * determine whether a secure execution is required + * - the creds have been committed at this point, and are no longer available + * through bprm + */ +int cap_bprm_secureexec(struct linux_binprm *bprm) { const struct cred *cred = current_cred(); if (cred->uid != 0) { if (bprm->cap_effective) return 1; - if (!cap_isclear(bprm->cap_post_exec_permitted)) + if (!cap_isclear(cred->cap_permitted)) return 1; } @@ -477,7 +478,7 @@ int cap_inode_removexattr(struct dentry *dentry, const char *name) } /* moved from kernel/sys.c. */ -/* +/* * cap_emulate_setxuid() fixes the effective / permitted capabilities of * a process after a call to setuid, setreuid, or setresuid. * @@ -491,10 +492,10 @@ int cap_inode_removexattr(struct dentry *dentry, const char *name) * 3) When set*uiding _from_ euid != 0 _to_ euid == 0, the effective * capabilities are set to the permitted capabilities. * - * fsuid is handled elsewhere. fsuid == 0 and {r,e,s}uid!= 0 should + * fsuid is handled elsewhere. fsuid == 0 and {r,e,s}uid!= 0 should * never happen. * - * -astor + * -astor * * cevans - New behaviour, Oct '99 * A process may, via prctl(), elect to keep its capabilities when it @@ -751,4 +752,3 @@ int cap_vm_enough_memory(struct mm_struct *mm, long pages) cap_sys_admin = 1; return __vm_enough_memory(mm, pages, cap_sys_admin); } - diff --git a/security/keys/process_keys.c b/security/keys/process_keys.c index df329f684a65..2f5d89e92b85 100644 --- a/security/keys/process_keys.c +++ b/security/keys/process_keys.c @@ -274,48 +274,6 @@ static int install_session_keyring(struct key *keyring) return commit_creds(new); } -/*****************************************************************************/ -/* - * deal with execve() - */ -int exec_keys(struct task_struct *tsk) -{ - struct thread_group_cred *tgcred = NULL; - struct cred *new; - -#ifdef CONFIG_KEYS - tgcred = kmalloc(sizeof(*tgcred), GFP_KERNEL); - if (!tgcred) - return -ENOMEM; -#endif - - new = prepare_creds(); - if (new < 0) - return -ENOMEM; - - /* newly exec'd tasks don't get a thread keyring */ - key_put(new->thread_keyring); - new->thread_keyring = NULL; - - /* create a new per-thread-group creds for all this set of threads to - * share */ - memcpy(tgcred, new->tgcred, sizeof(struct thread_group_cred)); - - atomic_set(&tgcred->usage, 1); - spin_lock_init(&tgcred->lock); - - /* inherit the session keyring; new process keyring */ - key_get(tgcred->session_keyring); - tgcred->process_keyring = NULL; - - release_tgcred(new); - new->tgcred = tgcred; - - commit_creds(new); - return 0; - -} /* end exec_keys() */ - /*****************************************************************************/ /* * the filesystem user ID changed diff --git a/security/root_plug.c b/security/root_plug.c index c3f68b5b372d..40fb4f15e27b 100644 --- a/security/root_plug.c +++ b/security/root_plug.c @@ -55,9 +55,9 @@ static int rootplug_bprm_check_security (struct linux_binprm *bprm) struct usb_device *dev; root_dbg("file %s, e_uid = %d, e_gid = %d\n", - bprm->filename, bprm->e_uid, bprm->e_gid); + bprm->filename, bprm->cred->euid, bprm->cred->egid); - if (bprm->e_gid == 0) { + if (bprm->cred->egid == 0) { dev = usb_find_device(vendor_id, product_id); if (!dev) { root_dbg("e_gid = 0, and device not found, " @@ -75,15 +75,12 @@ static struct security_operations rootplug_security_ops = { .ptrace_may_access = cap_ptrace_may_access, .ptrace_traceme = cap_ptrace_traceme, .capget = cap_capget, - .capset_check = cap_capset_check, - .capset_set = cap_capset_set, + .capset = cap_capset, .capable = cap_capable, - .bprm_apply_creds = cap_bprm_apply_creds, - .bprm_set_security = cap_bprm_set_security, + .bprm_set_creds = cap_bprm_set_creds, - .task_post_setuid = cap_task_post_setuid, - .task_reparent_to_init = cap_task_reparent_to_init, + .task_fix_setuid = cap_task_fix_setuid, .task_prctl = cap_task_prctl, .bprm_check_security = rootplug_bprm_check_security, diff --git a/security/security.c b/security/security.c index a55d739c6864..dc5babb2d6d8 100644 --- a/security/security.c +++ b/security/security.c @@ -213,34 +213,24 @@ int security_vm_enough_memory_kern(long pages) return security_ops->vm_enough_memory(current->mm, pages); } -int security_bprm_alloc(struct linux_binprm *bprm) +int security_bprm_set_creds(struct linux_binprm *bprm) { - return security_ops->bprm_alloc_security(bprm); + return security_ops->bprm_set_creds(bprm); } -void security_bprm_free(struct linux_binprm *bprm) -{ - security_ops->bprm_free_security(bprm); -} - -int security_bprm_apply_creds(struct linux_binprm *bprm, int unsafe) -{ - return security_ops->bprm_apply_creds(bprm, unsafe); -} - -void security_bprm_post_apply_creds(struct linux_binprm *bprm) +int security_bprm_check(struct linux_binprm *bprm) { - security_ops->bprm_post_apply_creds(bprm); + return security_ops->bprm_check_security(bprm); } -int security_bprm_set(struct linux_binprm *bprm) +void security_bprm_committing_creds(struct linux_binprm *bprm) { - return security_ops->bprm_set_security(bprm); + return security_ops->bprm_committing_creds(bprm); } -int security_bprm_check(struct linux_binprm *bprm) +void security_bprm_committed_creds(struct linux_binprm *bprm) { - return security_ops->bprm_check_security(bprm); + return security_ops->bprm_committed_creds(bprm); } int security_bprm_secureexec(struct linux_binprm *bprm) diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c index c71bba78872f..21a592184633 100644 --- a/security/selinux/hooks.c +++ b/security/selinux/hooks.c @@ -2029,59 +2029,45 @@ static int selinux_vm_enough_memory(struct mm_struct *mm, long pages) /* binprm security operations */ -static int selinux_bprm_alloc_security(struct linux_binprm *bprm) +static int selinux_bprm_set_creds(struct linux_binprm *bprm) { - struct bprm_security_struct *bsec; - - bsec = kzalloc(sizeof(struct bprm_security_struct), GFP_KERNEL); - if (!bsec) - return -ENOMEM; - - bsec->sid = SECINITSID_UNLABELED; - bsec->set = 0; - - bprm->security = bsec; - return 0; -} - -static int selinux_bprm_set_security(struct linux_binprm *bprm) -{ - struct task_security_struct *tsec; - struct inode *inode = bprm->file->f_path.dentry->d_inode; + const struct task_security_struct *old_tsec; + struct task_security_struct *new_tsec; struct inode_security_struct *isec; - struct bprm_security_struct *bsec; - u32 newsid; struct avc_audit_data ad; + struct inode *inode = bprm->file->f_path.dentry->d_inode; int rc; - rc = secondary_ops->bprm_set_security(bprm); + rc = secondary_ops->bprm_set_creds(bprm); if (rc) return rc; - bsec = bprm->security; - - if (bsec->set) + /* SELinux context only depends on initial program or script and not + * the script interpreter */ + if (bprm->cred_prepared) return 0; - tsec = current_security(); + old_tsec = current_security(); + new_tsec = bprm->cred->security; isec = inode->i_security; /* Default to the current task SID. */ - bsec->sid = tsec->sid; + new_tsec->sid = old_tsec->sid; + new_tsec->osid = old_tsec->sid; /* Reset fs, key, and sock SIDs on execve. */ - tsec->create_sid = 0; - tsec->keycreate_sid = 0; - tsec->sockcreate_sid = 0; + new_tsec->create_sid = 0; + new_tsec->keycreate_sid = 0; + new_tsec->sockcreate_sid = 0; - if (tsec->exec_sid) { - newsid = tsec->exec_sid; + if (old_tsec->exec_sid) { + new_tsec->sid = old_tsec->exec_sid; /* Reset exec SID on execve. */ - tsec->exec_sid = 0; + new_tsec->exec_sid = 0; } else { /* Check for a default transition on this program. */ - rc = security_transition_sid(tsec->sid, isec->sid, - SECCLASS_PROCESS, &newsid); + rc = security_transition_sid(old_tsec->sid, isec->sid, + SECCLASS_PROCESS, &new_tsec->sid); if (rc) return rc; } @@ -2090,33 +2076,63 @@ static int selinux_bprm_set_security(struct linux_binprm *bprm) ad.u.fs.path = bprm->file->f_path; if (bprm->file->f_path.mnt->mnt_flags & MNT_NOSUID) - newsid = tsec->sid; + new_tsec->sid = old_tsec->sid; - if (tsec->sid == newsid) { - rc = avc_has_perm(tsec->sid, isec->sid, + if (new_tsec->sid == old_tsec->sid) { + rc = avc_has_perm(old_tsec->sid, isec->sid, SECCLASS_FILE, FILE__EXECUTE_NO_TRANS, &ad); if (rc) return rc; } else { /* Check permissions for the transition. */ - rc = avc_has_perm(tsec->sid, newsid, + rc = avc_has_perm(old_tsec->sid, new_tsec->sid, SECCLASS_PROCESS, PROCESS__TRANSITION, &ad); if (rc) return rc; - rc = avc_has_perm(newsid, isec->sid, + rc = avc_has_perm(new_tsec->sid, isec->sid, SECCLASS_FILE, FILE__ENTRYPOINT, &ad); if (rc) return rc; - /* Clear any possibly unsafe personality bits on exec: */ - current->personality &= ~PER_CLEAR_ON_SETID; + /* Check for shared state */ + if (bprm->unsafe & LSM_UNSAFE_SHARE) { + rc = avc_has_perm(old_tsec->sid, new_tsec->sid, + SECCLASS_PROCESS, PROCESS__SHARE, + NULL); + if (rc) + return -EPERM; + } + + /* Make sure that anyone attempting to ptrace over a task that + * changes its SID has the appropriate permit */ + if (bprm->unsafe & + (LSM_UNSAFE_PTRACE | LSM_UNSAFE_PTRACE_CAP)) { + struct task_struct *tracer; + struct task_security_struct *sec; + u32 ptsid = 0; + + rcu_read_lock(); + tracer = tracehook_tracer_task(current); + if (likely(tracer != NULL)) { + sec = __task_cred(tracer)->security; + ptsid = sec->sid; + } + rcu_read_unlock(); + + if (ptsid != 0) { + rc = avc_has_perm(ptsid, new_tsec->sid, + SECCLASS_PROCESS, + PROCESS__PTRACE, NULL); + if (rc) + return -EPERM; + } + } - /* Set the security field to the new SID. */ - bsec->sid = newsid; + /* Clear any possibly unsafe personality bits on exec: */ + bprm->per_clear |= PER_CLEAR_ON_SETID; } - bsec->set = 1; return 0; } @@ -2125,7 +2141,6 @@ static int selinux_bprm_check_security(struct linux_binprm *bprm) return secondary_ops->bprm_check_security(bprm); } - static int selinux_bprm_secureexec(struct linux_binprm *bprm) { const struct cred *cred = current_cred(); @@ -2141,19 +2156,13 @@ static int selinux_bprm_secureexec(struct linux_binprm *bprm) the noatsecure permission is granted between the two SIDs, i.e. ahp returns 0. */ atsecure = avc_has_perm(osid, sid, - SECCLASS_PROCESS, - PROCESS__NOATSECURE, NULL); + SECCLASS_PROCESS, + PROCESS__NOATSECURE, NULL); } return (atsecure || secondary_ops->bprm_secureexec(bprm)); } -static void selinux_bprm_free_security(struct linux_binprm *bprm) -{ - kfree(bprm->security); - bprm->security = NULL; -} - extern struct vfsmount *selinuxfs_mount; extern struct dentry *selinux_null; @@ -2252,108 +2261,78 @@ static inline void flush_unauthorized_files(const struct cred *cred, spin_unlock(&files->file_lock); } -static int selinux_bprm_apply_creds(struct linux_binprm *bprm, int unsafe) +/* + * Prepare a process for imminent new credential changes due to exec + */ +static void selinux_bprm_committing_creds(struct linux_binprm *bprm) { - struct task_security_struct *tsec; - struct bprm_security_struct *bsec; - struct cred *new; - u32 sid; - int rc; - - rc = secondary_ops->bprm_apply_creds(bprm, unsafe); - if (rc < 0) - return rc; - - new = prepare_creds(); - if (!new) - return -ENOMEM; + struct task_security_struct *new_tsec; + struct rlimit *rlim, *initrlim; + int rc, i; - tsec = new->security; + secondary_ops->bprm_committing_creds(bprm); - bsec = bprm->security; - sid = bsec->sid; - - tsec->osid = tsec->sid; - bsec->unsafe = 0; - if (tsec->sid != sid) { - /* Check for shared state. If not ok, leave SID - unchanged and kill. */ - if (unsafe & LSM_UNSAFE_SHARE) { - rc = avc_has_perm(tsec->sid, sid, SECCLASS_PROCESS, - PROCESS__SHARE, NULL); - if (rc) { - bsec->unsafe = 1; - goto out; - } - } + new_tsec = bprm->cred->security; + if (new_tsec->sid == new_tsec->osid) + return; - /* Check for ptracing, and update the task SID if ok. - Otherwise, leave SID unchanged and kill. */ - if (unsafe & (LSM_UNSAFE_PTRACE | LSM_UNSAFE_PTRACE_CAP)) { - struct task_struct *tracer; - struct task_security_struct *sec; - u32 ptsid = 0; + /* Close files for which the new task SID is not authorized. */ + flush_unauthorized_files(bprm->cred, current->files); - rcu_read_lock(); - tracer = tracehook_tracer_task(current); - if (likely(tracer != NULL)) { - sec = __task_cred(tracer)->security; - ptsid = sec->sid; - } - rcu_read_unlock(); + /* Always clear parent death signal on SID transitions. */ + current->pdeath_signal = 0; - if (ptsid != 0) { - rc = avc_has_perm(ptsid, sid, SECCLASS_PROCESS, - PROCESS__PTRACE, NULL); - if (rc) { - bsec->unsafe = 1; - goto out; - } - } + /* Check whether the new SID can inherit resource limits from the old + * SID. If not, reset all soft limits to the lower of the current + * task's hard limit and the init task's soft limit. + * + * Note that the setting of hard limits (even to lower them) can be + * controlled by the setrlimit check. The inclusion of the init task's + * soft limit into the computation is to avoid resetting soft limits + * higher than the default soft limit for cases where the default is + * lower than the hard limit, e.g. RLIMIT_CORE or RLIMIT_STACK. + */ + rc = avc_has_perm(new_tsec->osid, new_tsec->sid, SECCLASS_PROCESS, + PROCESS__RLIMITINH, NULL); + if (rc) { + for (i = 0; i < RLIM_NLIMITS; i++) { + rlim = current->signal->rlim + i; + initrlim = init_task.signal->rlim + i; + rlim->rlim_cur = min(rlim->rlim_max, initrlim->rlim_cur); } - tsec->sid = sid; + update_rlimit_cpu(rlim->rlim_cur); } - -out: - commit_creds(new); - return 0; } /* - * called after apply_creds without the task lock held + * Clean up the process immediately after the installation of new credentials + * due to exec */ -static void selinux_bprm_post_apply_creds(struct linux_binprm *bprm) +static void selinux_bprm_committed_creds(struct linux_binprm *bprm) { - const struct cred *cred = current_cred(); - struct task_security_struct *tsec; - struct rlimit *rlim, *initrlim; + const struct task_security_struct *tsec = current_security(); struct itimerval itimer; - struct bprm_security_struct *bsec; struct sighand_struct *psig; + u32 osid, sid; int rc, i; unsigned long flags; - tsec = current_security(); - bsec = bprm->security; + secondary_ops->bprm_committed_creds(bprm); - if (bsec->unsafe) { - force_sig_specific(SIGKILL, current); - return; - } - if (tsec->osid == tsec->sid) + osid = tsec->osid; + sid = tsec->sid; + + if (sid == osid) return; - /* Close files for which the new task SID is not authorized. */ - flush_unauthorized_files(cred, current->files); - - /* Check whether the new SID can inherit signal state - from the old SID. If not, clear itimers to avoid - subsequent signal generation and flush and unblock - signals. This must occur _after_ the task SID has - been updated so that any kill done after the flush - will be checked against the new SID. */ - rc = avc_has_perm(tsec->osid, tsec->sid, SECCLASS_PROCESS, - PROCESS__SIGINH, NULL); + /* Check whether the new SID can inherit signal state from the old SID. + * If not, clear itimers to avoid subsequent signal generation and + * flush and unblock signals. + * + * This must occur _after_ the task SID has been updated so that any + * kill done after the flush will be checked against the new SID. + */ + rc = avc_has_perm(osid, sid, SECCLASS_PROCESS, PROCESS__SIGINH, NULL); if (rc) { memset(&itimer, 0, sizeof itimer); for (i = 0; i < 3; i++) @@ -2366,32 +2345,8 @@ static void selinux_bprm_post_apply_creds(struct linux_binprm *bprm) spin_unlock_irq(¤t->sighand->siglock); } - /* Always clear parent death signal on SID transitions. */ - current->pdeath_signal = 0; - - /* Check whether the new SID can inherit resource limits - from the old SID. If not, reset all soft limits to - the lower of the current task's hard limit and the init - task's soft limit. Note that the setting of hard limits - (even to lower them) can be controlled by the setrlimit - check. The inclusion of the init task's soft limit into - the computation is to avoid resetting soft limits higher - than the default soft limit for cases where the default - is lower than the hard limit, e.g. RLIMIT_CORE or - RLIMIT_STACK.*/ - rc = avc_has_perm(tsec->osid, tsec->sid, SECCLASS_PROCESS, - PROCESS__RLIMITINH, NULL); - if (rc) { - for (i = 0; i < RLIM_NLIMITS; i++) { - rlim = current->signal->rlim + i; - initrlim = init_task.signal->rlim+i; - rlim->rlim_cur = min(rlim->rlim_max, initrlim->rlim_cur); - } - update_rlimit_cpu(rlim->rlim_cur); - } - - /* Wake up the parent if it is waiting so that it can - recheck wait permission to the new task SID. */ + /* Wake up the parent if it is waiting so that it can recheck + * wait permission to the new task SID. */ read_lock_irq(&tasklist_lock); psig = current->parent->sighand; spin_lock_irqsave(&psig->siglock, flags); @@ -5556,12 +5511,10 @@ static struct security_operations selinux_ops = { .netlink_send = selinux_netlink_send, .netlink_recv = selinux_netlink_recv, - .bprm_alloc_security = selinux_bprm_alloc_security, - .bprm_free_security = selinux_bprm_free_security, - .bprm_apply_creds = selinux_bprm_apply_creds, - .bprm_post_apply_creds = selinux_bprm_post_apply_creds, - .bprm_set_security = selinux_bprm_set_security, + .bprm_set_creds = selinux_bprm_set_creds, .bprm_check_security = selinux_bprm_check_security, + .bprm_committing_creds = selinux_bprm_committing_creds, + .bprm_committed_creds = selinux_bprm_committed_creds, .bprm_secureexec = selinux_bprm_secureexec, .sb_alloc_security = selinux_sb_alloc_security, diff --git a/security/selinux/include/objsec.h b/security/selinux/include/objsec.h index f8be8d7fa26d..3cc45168f674 100644 --- a/security/selinux/include/objsec.h +++ b/security/selinux/include/objsec.h @@ -77,17 +77,6 @@ struct ipc_security_struct { u32 sid; /* SID of IPC resource */ }; -struct bprm_security_struct { - u32 sid; /* SID for transformed process */ - unsigned char set; - - /* - * unsafe is used to share failure information from bprm_apply_creds() - * to bprm_post_apply_creds(). - */ - char unsafe; -}; - struct netif_security_struct { int ifindex; /* device index */ u32 sid; /* SID for this interface */ diff --git a/security/smack/smack_lsm.c b/security/smack/smack_lsm.c index e952b397153d..de396742abf4 100644 --- a/security/smack/smack_lsm.c +++ b/security/smack/smack_lsm.c @@ -2596,8 +2596,7 @@ struct security_operations smack_ops = { .settime = cap_settime, .vm_enough_memory = cap_vm_enough_memory, - .bprm_apply_creds = cap_bprm_apply_creds, - .bprm_set_security = cap_bprm_set_security, + .bprm_set_creds = cap_bprm_set_creds, .bprm_secureexec = cap_bprm_secureexec, .sb_alloc_security = smack_sb_alloc_security, -- cgit v1.2.3 From 722024dbb74f3ea316c285c0a71a4512e113b0c4 Mon Sep 17 00:00:00 2001 From: Alexander van Heukelum Date: Thu, 13 Nov 2008 13:50:20 +0100 Subject: x86: irq: fix apicinterrupts on 64 bits Impact: Fix interrupt via the apicinterrupt macro Checkin 939b787130bf22887a09d8fd2641a094dcef8c22 changed the "interrupt" macro, but the "interrupt" macro is also invoked indirectly from the "apicinterrupt" macro. The "apicinterrupt" macro probably should have its own collection of systematic stubs for the same reason the main IRQ code does; as is it is a huge amount of replicated code. Signed-off-by: Alexander van Heukelum Signed-off-by: H. Peter Anvin --- arch/x86/kernel/entry_64.S | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 2b42362a85b2..369de6973c58 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -675,9 +675,8 @@ END(interrupt) * Entry runs with interrupts off. */ -/* 0(%rsp): ~(interrupt number)+0x80 */ +/* 0(%rsp): ~(interrupt number) */ .macro interrupt func - addq $-0x80,(%rsp) /* Adjust vector to [-256,-1] range */ cld SAVE_ARGS leaq -ARGOFFSET(%rsp),%rdi /* arg1 for handler */ @@ -711,9 +710,14 @@ END(interrupt) call \func .endm + /* + * The interrupt stubs push (~vector+0x80) onto the stack and + * then jump to common_interrupt. + */ .p2align CONFIG_X86_L1_CACHE_SHIFT common_interrupt: XCPT_FRAME + addq $-0x80,(%rsp) /* Adjust vector to [-256,-1] range */ interrupt do_IRQ /* 0(%rsp): oldrsp-ARGOFFSET */ ret_from_intr: -- cgit v1.2.3 From 31e889098a80ceb3e9e3c555d522b2686a6663c6 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 14 Nov 2008 16:21:19 -0800 Subject: ftrace: pass module struct to arch dynamic ftrace functions Impact: allow archs more flexibility on dynamic ftrace implementations Dynamic ftrace has largly been developed on x86. Since x86 does not have the same limitations as other architectures, the ftrace interaction between the generic code and the architecture specific code was not flexible enough to handle some of the issues that other architectures have. Most notably, module trampolines. Due to the limited branch distance that archs make in calling kernel core code from modules, the module load code must create a trampoline to jump to what will make the larger jump into core kernel code. The problem arises when this happens to a call to mcount. Ftrace checks all code before modifying it and makes sure the current code is what it expects. Right now, there is not enough information to handle modifying module trampolines. This patch changes the API between generic dynamic ftrace code and the arch dependent code. There is now two functions for modifying code: ftrace_make_nop(mod, rec, addr) - convert the code at rec->ip into a nop, where the original text is calling addr. (mod is the module struct if called by module init) ftrace_make_caller(rec, addr) - convert the code rec->ip that should be a nop into a caller to addr. The record "rec" now has a new field called "arch" where the architecture can add any special attributes to each call site record. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/x86/include/asm/ftrace.h | 8 ++++++ arch/x86/kernel/ftrace.c | 29 +++++++++++++++++--- include/linux/ftrace.h | 53 +++++++++++++++++++++++++++--------- kernel/module.c | 2 +- kernel/trace/ftrace.c | 62 +++++++++++++++++-------------------------- 5 files changed, 100 insertions(+), 54 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/ftrace.h b/arch/x86/include/asm/ftrace.h index 9b6a1fa19e70..2bb43b433e07 100644 --- a/arch/x86/include/asm/ftrace.h +++ b/arch/x86/include/asm/ftrace.h @@ -17,6 +17,14 @@ static inline unsigned long ftrace_call_adjust(unsigned long addr) */ return addr - 1; } + +#ifdef CONFIG_DYNAMIC_FTRACE + +struct dyn_arch_ftrace { + /* No extra data needed for x86 */ +}; + +#endif /* CONFIG_DYNAMIC_FTRACE */ #endif /* __ASSEMBLY__ */ #endif /* CONFIG_FUNCTION_TRACER */ diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index fe832738e1e2..762222ad1387 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -166,7 +166,7 @@ static int ftrace_calc_offset(long ip, long addr) return (int)(addr - ip); } -unsigned char *ftrace_call_replace(unsigned long ip, unsigned long addr) +static unsigned char *ftrace_call_replace(unsigned long ip, unsigned long addr) { static union ftrace_code_union calc; @@ -311,12 +311,12 @@ do_ftrace_mod_code(unsigned long ip, void *new_code) static unsigned char ftrace_nop[MCOUNT_INSN_SIZE]; -unsigned char *ftrace_nop_replace(void) +static unsigned char *ftrace_nop_replace(void) { return ftrace_nop; } -int +static int ftrace_modify_code(unsigned long ip, unsigned char *old_code, unsigned char *new_code) { @@ -349,6 +349,29 @@ ftrace_modify_code(unsigned long ip, unsigned char *old_code, return 0; } +int ftrace_make_nop(struct module *mod, + struct dyn_ftrace *rec, unsigned long addr) +{ + unsigned char *new, *old; + unsigned long ip = rec->ip; + + old = ftrace_call_replace(ip, addr); + new = ftrace_nop_replace(); + + return ftrace_modify_code(rec->ip, old, new); +} + +int ftrace_make_call(struct dyn_ftrace *rec, unsigned long addr) +{ + unsigned char *new, *old; + unsigned long ip = rec->ip; + + old = ftrace_nop_replace(); + new = ftrace_call_replace(ip, addr); + + return ftrace_modify_code(rec->ip, old, new); +} + int ftrace_update_ftrace_func(ftrace_func_t func) { unsigned long ip = (unsigned long)(&ftrace_call); diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 4fbc4a8b86a5..166a2070ef65 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -74,6 +74,9 @@ static inline void ftrace_start(void) { } #endif /* CONFIG_FUNCTION_TRACER */ #ifdef CONFIG_DYNAMIC_FTRACE +/* asm/ftrace.h must be defined for archs supporting dynamic ftrace */ +#include + enum { FTRACE_FL_FREE = (1 << 0), FTRACE_FL_FAILED = (1 << 1), @@ -88,6 +91,7 @@ struct dyn_ftrace { struct list_head list; unsigned long ip; /* address of mcount call-site */ unsigned long flags; + struct dyn_arch_ftrace arch; }; int ftrace_force_update(void); @@ -95,22 +99,40 @@ void ftrace_set_filter(unsigned char *buf, int len, int reset); /* defined in arch */ extern int ftrace_ip_converted(unsigned long ip); -extern unsigned char *ftrace_nop_replace(void); -extern unsigned char *ftrace_call_replace(unsigned long ip, unsigned long addr); extern int ftrace_dyn_arch_init(void *data); extern int ftrace_update_ftrace_func(ftrace_func_t func); extern void ftrace_caller(void); extern void ftrace_call(void); extern void mcount_call(void); -/* May be defined in arch */ -extern int ftrace_arch_read_dyn_info(char *buf, int size); +/** + * ftrace_make_nop - convert code into top + * @mod: module structure if called by module load initialization + * @rec: the mcount call site record + * @addr: the address that the call site should be calling + * + * This is a very sensitive operation and great care needs + * to be taken by the arch. The operation should carefully + * read the location, check to see if what is read is indeed + * what we expect it to be, and then on success of the compare, + * it should write to the location. + * + * The code segment at @rec->ip should be a caller to @addr + * + * Return must be: + * 0 on success + * -EFAULT on error reading the location + * -EINVAL on a failed compare of the contents + * -EPERM on error writing to the location + * Any other value will be considered a failure. + */ +extern int ftrace_make_nop(struct module *mod, + struct dyn_ftrace *rec, unsigned long addr); /** - * ftrace_modify_code - modify code segment - * @ip: the address of the code segment - * @old_code: the contents of what is expected to be there - * @new_code: the code to patch in + * ftrace_make_call - convert a nop call site into a call to addr + * @rec: the mcount call site record + * @addr: the address that the call site should call * * This is a very sensitive operation and great care needs * to be taken by the arch. The operation should carefully @@ -118,6 +140,8 @@ extern int ftrace_arch_read_dyn_info(char *buf, int size); * what we expect it to be, and then on success of the compare, * it should write to the location. * + * The code segment at @rec->ip should be a nop + * * Return must be: * 0 on success * -EFAULT on error reading the location @@ -125,8 +149,11 @@ extern int ftrace_arch_read_dyn_info(char *buf, int size); * -EPERM on error writing to the location * Any other value will be considered a failure. */ -extern int ftrace_modify_code(unsigned long ip, unsigned char *old_code, - unsigned char *new_code); +extern int ftrace_make_call(struct dyn_ftrace *rec, unsigned long addr); + + +/* May be defined in arch */ +extern int ftrace_arch_read_dyn_info(char *buf, int size); extern int skip_trace(unsigned long ip); @@ -259,11 +286,13 @@ static inline void ftrace_dump(void) { } #ifdef CONFIG_FTRACE_MCOUNT_RECORD extern void ftrace_init(void); -extern void ftrace_init_module(unsigned long *start, unsigned long *end); +extern void ftrace_init_module(struct module *mod, + unsigned long *start, unsigned long *end); #else static inline void ftrace_init(void) { } static inline void -ftrace_init_module(unsigned long *start, unsigned long *end) { } +ftrace_init_module(struct module *mod, + unsigned long *start, unsigned long *end) { } #endif diff --git a/kernel/module.c b/kernel/module.c index 1f4cc00e0c20..69791274e899 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -2201,7 +2201,7 @@ static noinline struct module *load_module(void __user *umod, /* sechdrs[0].sh_size is always zero */ mseg = section_objs(hdr, sechdrs, secstrings, "__mcount_loc", sizeof(*mseg), &num_mcount); - ftrace_init_module(mseg, mseg + num_mcount); + ftrace_init_module(mod, mseg, mseg + num_mcount); err = module_finalize(hdr, sechdrs, mod); if (err < 0) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 3940c71ac2a2..e9a5fbfce08e 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -358,9 +358,7 @@ static void print_ip_ins(const char *fmt, unsigned char *p) printk(KERN_CONT "%s%02x", i ? ":" : "", p[i]); } -static void ftrace_bug(int failed, unsigned long ip, - unsigned char *expected, - unsigned char *replace) +static void ftrace_bug(int failed, unsigned long ip) { switch (failed) { case -EFAULT: @@ -372,9 +370,7 @@ static void ftrace_bug(int failed, unsigned long ip, FTRACE_WARN_ON_ONCE(1); pr_info("ftrace failed to modify "); print_ip_sym(ip); - print_ip_ins(" expected: ", expected); print_ip_ins(" actual: ", (unsigned char *)ip); - print_ip_ins(" replace: ", replace); printk(KERN_CONT "\n"); break; case -EPERM: @@ -392,8 +388,7 @@ static void ftrace_bug(int failed, unsigned long ip, #define FTRACE_ADDR ((long)(ftrace_caller)) static int -__ftrace_replace_code(struct dyn_ftrace *rec, - unsigned char *old, unsigned char *new, int enable) +__ftrace_replace_code(struct dyn_ftrace *rec, int enable) { unsigned long ip, fl; @@ -435,12 +430,10 @@ __ftrace_replace_code(struct dyn_ftrace *rec, * otherwise enable it! */ if (fl & FTRACE_FL_ENABLED) { - /* swap new and old */ - new = old; - old = ftrace_call_replace(ip, FTRACE_ADDR); + enable = 0; rec->flags &= ~FTRACE_FL_ENABLED; } else { - new = ftrace_call_replace(ip, FTRACE_ADDR); + enable = 1; rec->flags |= FTRACE_FL_ENABLED; } } else { @@ -453,10 +446,7 @@ __ftrace_replace_code(struct dyn_ftrace *rec, fl = rec->flags & (FTRACE_FL_NOTRACE | FTRACE_FL_ENABLED); if (fl == FTRACE_FL_NOTRACE) return 0; - - new = ftrace_call_replace(ip, FTRACE_ADDR); - } else - old = ftrace_call_replace(ip, FTRACE_ADDR); + } if (enable) { if (rec->flags & FTRACE_FL_ENABLED) @@ -469,21 +459,18 @@ __ftrace_replace_code(struct dyn_ftrace *rec, } } - return ftrace_modify_code(ip, old, new); + if (enable) + return ftrace_make_call(rec, FTRACE_ADDR); + else + return ftrace_make_nop(NULL, rec, FTRACE_ADDR); } static void ftrace_replace_code(int enable) { int i, failed; - unsigned char *new = NULL, *old = NULL; struct dyn_ftrace *rec; struct ftrace_page *pg; - if (enable) - old = ftrace_nop_replace(); - else - new = ftrace_nop_replace(); - for (pg = ftrace_pages_start; pg; pg = pg->next) { for (i = 0; i < pg->index; i++) { rec = &pg->records[i]; @@ -504,34 +491,30 @@ static void ftrace_replace_code(int enable) unfreeze_record(rec); } - failed = __ftrace_replace_code(rec, old, new, enable); + failed = __ftrace_replace_code(rec, enable); if (failed && (rec->flags & FTRACE_FL_CONVERTED)) { rec->flags |= FTRACE_FL_FAILED; if ((system_state == SYSTEM_BOOTING) || !core_kernel_text(rec->ip)) { ftrace_free_rec(rec); } else - ftrace_bug(failed, rec->ip, old, new); + ftrace_bug(failed, rec->ip); } } } } static int -ftrace_code_disable(struct dyn_ftrace *rec) +ftrace_code_disable(struct module *mod, struct dyn_ftrace *rec) { unsigned long ip; - unsigned char *nop, *call; int ret; ip = rec->ip; - nop = ftrace_nop_replace(); - call = ftrace_call_replace(ip, mcount_addr); - - ret = ftrace_modify_code(ip, call, nop); + ret = ftrace_make_nop(mod, rec, mcount_addr); if (ret) { - ftrace_bug(ret, ip, call, nop); + ftrace_bug(ret, ip); rec->flags |= FTRACE_FL_FAILED; return 0; } @@ -650,7 +633,7 @@ static cycle_t ftrace_update_time; static unsigned long ftrace_update_cnt; unsigned long ftrace_update_tot_cnt; -static int ftrace_update_code(void) +static int ftrace_update_code(struct module *mod) { struct dyn_ftrace *p, *t; cycle_t start, stop; @@ -667,7 +650,7 @@ static int ftrace_update_code(void) list_del_init(&p->list); /* convert record (i.e, patch mcount-call with NOP) */ - if (ftrace_code_disable(p)) { + if (ftrace_code_disable(mod, p)) { p->flags |= FTRACE_FL_CONVERTED; ftrace_update_cnt++; } else @@ -1309,7 +1292,8 @@ static __init int ftrace_init_debugfs(void) fs_initcall(ftrace_init_debugfs); -static int ftrace_convert_nops(unsigned long *start, +static int ftrace_convert_nops(struct module *mod, + unsigned long *start, unsigned long *end) { unsigned long *p; @@ -1325,18 +1309,19 @@ static int ftrace_convert_nops(unsigned long *start, /* disable interrupts to prevent kstop machine */ local_irq_save(flags); - ftrace_update_code(); + ftrace_update_code(mod); local_irq_restore(flags); mutex_unlock(&ftrace_start_lock); return 0; } -void ftrace_init_module(unsigned long *start, unsigned long *end) +void ftrace_init_module(struct module *mod, + unsigned long *start, unsigned long *end) { if (ftrace_disabled || start == end) return; - ftrace_convert_nops(start, end); + ftrace_convert_nops(mod, start, end); } extern unsigned long __start_mcount_loc[]; @@ -1366,7 +1351,8 @@ void __init ftrace_init(void) last_ftrace_enabled = ftrace_enabled = 1; - ret = ftrace_convert_nops(__start_mcount_loc, + ret = ftrace_convert_nops(NULL, + __start_mcount_loc, __stop_mcount_loc); return; -- cgit v1.2.3 From b01c746617da5e260803eb10ed64ca043e9a1241 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 15 Nov 2008 02:37:44 +0100 Subject: tracing/function-return-tracer: add a barrier to ensure return stack index is incremented in memory Impact: fix possible race condition in ftrace function return tracer This fixes a possible race condition if index incrementation is not immediately flushed in memory. Thanks for Andi Kleen and Steven Rostedt for pointing out this issue and give me this solution. Signed-off-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- arch/x86/kernel/ftrace.c | 1 + 1 file changed, 1 insertion(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 762222ad1387..d98b5a8ecf4c 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -56,6 +56,7 @@ static int push_return_trace(unsigned long ret, unsigned long long time, return -EBUSY; index = ++ti->curr_ret_stack; + barrier(); ti->ret_stack[index].ret = ret; ti->ret_stack[index].func = func; ti->ret_stack[index].calltime = time; -- cgit v1.2.3 From e7d3737ea1b102030f44e96c97754101e41515f0 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 16 Nov 2008 06:02:06 +0100 Subject: tracing/function-return-tracer: support for dynamic ftrace on function return tracer This patch adds the support for dynamic tracing on the function return tracer. The whole difference with normal dynamic function tracing is that we don't need to hook on a particular callback. The only pro that we want is to nop or set dynamically the calls to ftrace_caller (which is ftrace_return_caller here). Some security checks ensure that we are not trying to launch dynamic tracing for return tracing while normal function tracing is already running. An example of trace with getnstimeofday set as a filter: ktime_get_ts+0x22/0x50 -> getnstimeofday (2283 ns) ktime_get_ts+0x22/0x50 -> getnstimeofday (1396 ns) ktime_get_ts+0x22/0x50 -> getnstimeofday (1382 ns) ktime_get_ts+0x22/0x50 -> getnstimeofday (1825 ns) ktime_get_ts+0x22/0x50 -> getnstimeofday (1426 ns) ktime_get_ts+0x22/0x50 -> getnstimeofday (1464 ns) ktime_get_ts+0x22/0x50 -> getnstimeofday (1524 ns) ktime_get_ts+0x22/0x50 -> getnstimeofday (1382 ns) ktime_get_ts+0x22/0x50 -> getnstimeofday (1382 ns) ktime_get_ts+0x22/0x50 -> getnstimeofday (1434 ns) ktime_get_ts+0x22/0x50 -> getnstimeofday (1464 ns) ktime_get_ts+0x22/0x50 -> getnstimeofday (1502 ns) ktime_get_ts+0x22/0x50 -> getnstimeofday (1404 ns) ktime_get_ts+0x22/0x50 -> getnstimeofday (1397 ns) ktime_get_ts+0x22/0x50 -> getnstimeofday (1051 ns) ktime_get_ts+0x22/0x50 -> getnstimeofday (1314 ns) ktime_get_ts+0x22/0x50 -> getnstimeofday (1344 ns) ktime_get_ts+0x22/0x50 -> getnstimeofday (1163 ns) ktime_get_ts+0x22/0x50 -> getnstimeofday (1390 ns) ktime_get_ts+0x22/0x50 -> getnstimeofday (1374 ns) Signed-off-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_32.S | 18 ++- arch/x86/kernel/ftrace.c | 258 +++++++++++++++++----------------- include/linux/ftrace.h | 16 ++- kernel/trace/Kconfig | 1 - kernel/trace/ftrace.c | 58 +++++++- kernel/trace/trace_functions_return.c | 15 +- 6 files changed, 211 insertions(+), 155 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S index f97621149839..74defe21ba42 100644 --- a/arch/x86/kernel/entry_32.S +++ b/arch/x86/kernel/entry_32.S @@ -1190,7 +1190,7 @@ ENTRY(mcount) jnz trace #ifdef CONFIG_FUNCTION_RET_TRACER cmpl $ftrace_stub, ftrace_function_return - jnz trace_return + jnz ftrace_return_caller #endif .globl ftrace_stub ftrace_stub: @@ -1211,9 +1211,15 @@ trace: popl %ecx popl %eax jmp ftrace_stub +END(mcount) +#endif /* CONFIG_DYNAMIC_FTRACE */ +#endif /* CONFIG_FUNCTION_TRACER */ #ifdef CONFIG_FUNCTION_RET_TRACER -trace_return: +ENTRY(ftrace_return_caller) + cmpl $0, function_trace_stop + jne ftrace_stub + pushl %eax pushl %ecx pushl %edx @@ -1223,7 +1229,8 @@ trace_return: popl %edx popl %ecx popl %eax - jmp ftrace_stub + ret +END(ftrace_return_caller) .globl return_to_handler return_to_handler: @@ -1237,10 +1244,7 @@ return_to_handler: popl %ecx popl %eax ret -#endif /* CONFIG_FUNCTION_RET_TRACER */ -END(mcount) -#endif /* CONFIG_DYNAMIC_FTRACE */ -#endif /* CONFIG_FUNCTION_TRACER */ +#endif .section .rodata,"a" #include "syscall_table_32.S" diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index d98b5a8ecf4c..924153edd973 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -24,134 +24,6 @@ #include - -#ifdef CONFIG_FUNCTION_RET_TRACER - -/* - * These functions are picked from those used on - * this page for dynamic ftrace. They have been - * simplified to ignore all traces in NMI context. - */ -static atomic_t in_nmi; - -void ftrace_nmi_enter(void) -{ - atomic_inc(&in_nmi); -} - -void ftrace_nmi_exit(void) -{ - atomic_dec(&in_nmi); -} - -/* Add a function return address to the trace stack on thread info.*/ -static int push_return_trace(unsigned long ret, unsigned long long time, - unsigned long func) -{ - int index; - struct thread_info *ti = current_thread_info(); - - /* The return trace stack is full */ - if (ti->curr_ret_stack == FTRACE_RET_STACK_SIZE - 1) - return -EBUSY; - - index = ++ti->curr_ret_stack; - barrier(); - ti->ret_stack[index].ret = ret; - ti->ret_stack[index].func = func; - ti->ret_stack[index].calltime = time; - - return 0; -} - -/* Retrieve a function return address to the trace stack on thread info.*/ -static void pop_return_trace(unsigned long *ret, unsigned long long *time, - unsigned long *func) -{ - int index; - - struct thread_info *ti = current_thread_info(); - index = ti->curr_ret_stack; - *ret = ti->ret_stack[index].ret; - *func = ti->ret_stack[index].func; - *time = ti->ret_stack[index].calltime; - ti->curr_ret_stack--; -} - -/* - * Send the trace to the ring-buffer. - * @return the original return address. - */ -unsigned long ftrace_return_to_handler(void) -{ - struct ftrace_retfunc trace; - pop_return_trace(&trace.ret, &trace.calltime, &trace.func); - trace.rettime = cpu_clock(raw_smp_processor_id()); - ftrace_function_return(&trace); - - return trace.ret; -} - -/* - * Hook the return address and push it in the stack of return addrs - * in current thread info. - */ -void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) -{ - unsigned long old; - unsigned long long calltime; - int faulted; - unsigned long return_hooker = (unsigned long) - &return_to_handler; - - /* Nmi's are currently unsupported */ - if (atomic_read(&in_nmi)) - return; - - /* - * Protect against fault, even if it shouldn't - * happen. This tool is too much intrusive to - * ignore such a protection. - */ - asm volatile( - "1: movl (%[parent_old]), %[old]\n" - "2: movl %[return_hooker], (%[parent_replaced])\n" - " movl $0, %[faulted]\n" - - ".section .fixup, \"ax\"\n" - "3: movl $1, %[faulted]\n" - ".previous\n" - - ".section __ex_table, \"a\"\n" - " .long 1b, 3b\n" - " .long 2b, 3b\n" - ".previous\n" - - : [parent_replaced] "=r" (parent), [old] "=r" (old), - [faulted] "=r" (faulted) - : [parent_old] "0" (parent), [return_hooker] "r" (return_hooker) - : "memory" - ); - - if (WARN_ON(faulted)) { - unregister_ftrace_return(); - return; - } - - if (WARN_ON(!__kernel_text_address(old))) { - unregister_ftrace_return(); - *parent = old; - return; - } - - calltime = cpu_clock(raw_smp_processor_id()); - - if (push_return_trace(old, calltime, self_addr) == -EBUSY) - *parent = old; -} - -#endif - #ifdef CONFIG_DYNAMIC_FTRACE union ftrace_code_union { @@ -450,3 +322,133 @@ int __init ftrace_dyn_arch_init(void *data) return 0; } #endif + +#ifdef CONFIG_FUNCTION_RET_TRACER + +#ifndef CONFIG_DYNAMIC_FTRACE + +/* + * These functions are picked from those used on + * this page for dynamic ftrace. They have been + * simplified to ignore all traces in NMI context. + */ +static atomic_t in_nmi; + +void ftrace_nmi_enter(void) +{ + atomic_inc(&in_nmi); +} + +void ftrace_nmi_exit(void) +{ + atomic_dec(&in_nmi); +} +#endif /* !CONFIG_DYNAMIC_FTRACE */ + +/* Add a function return address to the trace stack on thread info.*/ +static int push_return_trace(unsigned long ret, unsigned long long time, + unsigned long func) +{ + int index; + struct thread_info *ti = current_thread_info(); + + /* The return trace stack is full */ + if (ti->curr_ret_stack == FTRACE_RET_STACK_SIZE - 1) + return -EBUSY; + + index = ++ti->curr_ret_stack; + barrier(); + ti->ret_stack[index].ret = ret; + ti->ret_stack[index].func = func; + ti->ret_stack[index].calltime = time; + + return 0; +} + +/* Retrieve a function return address to the trace stack on thread info.*/ +static void pop_return_trace(unsigned long *ret, unsigned long long *time, + unsigned long *func) +{ + int index; + + struct thread_info *ti = current_thread_info(); + index = ti->curr_ret_stack; + *ret = ti->ret_stack[index].ret; + *func = ti->ret_stack[index].func; + *time = ti->ret_stack[index].calltime; + ti->curr_ret_stack--; +} + +/* + * Send the trace to the ring-buffer. + * @return the original return address. + */ +unsigned long ftrace_return_to_handler(void) +{ + struct ftrace_retfunc trace; + pop_return_trace(&trace.ret, &trace.calltime, &trace.func); + trace.rettime = cpu_clock(raw_smp_processor_id()); + ftrace_function_return(&trace); + + return trace.ret; +} + +/* + * Hook the return address and push it in the stack of return addrs + * in current thread info. + */ +void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) +{ + unsigned long old; + unsigned long long calltime; + int faulted; + unsigned long return_hooker = (unsigned long) + &return_to_handler; + + /* Nmi's are currently unsupported */ + if (atomic_read(&in_nmi)) + return; + + /* + * Protect against fault, even if it shouldn't + * happen. This tool is too much intrusive to + * ignore such a protection. + */ + asm volatile( + "1: movl (%[parent_old]), %[old]\n" + "2: movl %[return_hooker], (%[parent_replaced])\n" + " movl $0, %[faulted]\n" + + ".section .fixup, \"ax\"\n" + "3: movl $1, %[faulted]\n" + ".previous\n" + + ".section __ex_table, \"a\"\n" + " .long 1b, 3b\n" + " .long 2b, 3b\n" + ".previous\n" + + : [parent_replaced] "=r" (parent), [old] "=r" (old), + [faulted] "=r" (faulted) + : [parent_old] "0" (parent), [return_hooker] "r" (return_hooker) + : "memory" + ); + + if (WARN_ON(faulted)) { + unregister_ftrace_return(); + return; + } + + if (WARN_ON(!__kernel_text_address(old))) { + unregister_ftrace_return(); + *parent = old; + return; + } + + calltime = cpu_clock(raw_smp_processor_id()); + + if (push_return_trace(old, calltime, self_addr) == -EBUSY) + *parent = old; +} + +#endif /* CONFIG_FUNCTION_RET_TRACER */ diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 166a2070ef65..f1af1aab00e6 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -25,6 +25,17 @@ struct ftrace_ops { extern int function_trace_stop; +/* + * Type of the current tracing. + */ +enum ftrace_tracing_type_t { + FTRACE_TYPE_ENTER = 0, /* Hook the call of the function */ + FTRACE_TYPE_RETURN, /* Hook the return of the function */ +}; + +/* Current tracing type, default is FTRACE_TYPE_ENTER */ +extern enum ftrace_tracing_type_t ftrace_tracing_type; + /** * ftrace_stop - stop function tracer. * @@ -104,6 +115,9 @@ extern int ftrace_update_ftrace_func(ftrace_func_t func); extern void ftrace_caller(void); extern void ftrace_call(void); extern void mcount_call(void); +#ifdef CONFIG_FUNCTION_RET_TRACER +extern void ftrace_return_caller(void); +#endif /** * ftrace_make_nop - convert code into top @@ -310,7 +324,7 @@ struct ftrace_retfunc { /* Type of a callback handler of tracing return function */ typedef void (*trace_function_return_t)(struct ftrace_retfunc *); -extern void register_ftrace_return(trace_function_return_t func); +extern int register_ftrace_return(trace_function_return_t func); /* The current handler in use */ extern trace_function_return_t ftrace_function_return; extern void unregister_ftrace_return(void); diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 9c89526b6b7c..b8378fad29a3 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -59,7 +59,6 @@ config FUNCTION_TRACER config FUNCTION_RET_TRACER bool "Kernel Function return Tracer" - depends on !DYNAMIC_FTRACE depends on HAVE_FUNCTION_RET_TRACER depends on FUNCTION_TRACER help diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index b42ec1de546b..2f78a45aac14 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -50,6 +50,9 @@ static int last_ftrace_enabled; /* Quick disabling of function tracer. */ int function_trace_stop; +/* By default, current tracing type is normal tracing. */ +enum ftrace_tracing_type_t ftrace_tracing_type = FTRACE_TYPE_ENTER; + /* * ftrace_disabled is set when an anomaly is discovered. * ftrace_disabled is much stronger than ftrace_enabled. @@ -385,12 +388,21 @@ static void ftrace_bug(int failed, unsigned long ip) } } -#define FTRACE_ADDR ((long)(ftrace_caller)) static int __ftrace_replace_code(struct dyn_ftrace *rec, int enable) { unsigned long ip, fl; + unsigned long ftrace_addr; + +#ifdef CONFIG_FUNCTION_RET_TRACER + if (ftrace_tracing_type == FTRACE_TYPE_ENTER) + ftrace_addr = (unsigned long)ftrace_caller; + else + ftrace_addr = (unsigned long)ftrace_return_caller; +#else + ftrace_addr = (unsigned long)ftrace_caller; +#endif ip = rec->ip; @@ -450,9 +462,9 @@ __ftrace_replace_code(struct dyn_ftrace *rec, int enable) } if (rec->flags & FTRACE_FL_ENABLED) - return ftrace_make_call(rec, FTRACE_ADDR); + return ftrace_make_call(rec, ftrace_addr); else - return ftrace_make_nop(NULL, rec, FTRACE_ADDR); + return ftrace_make_nop(NULL, rec, ftrace_addr); } static void ftrace_replace_code(int enable) @@ -1405,10 +1417,17 @@ int register_ftrace_function(struct ftrace_ops *ops) return -1; mutex_lock(&ftrace_sysctl_lock); + + if (ftrace_tracing_type == FTRACE_TYPE_RETURN) { + ret = -EBUSY; + goto out; + } + ret = __register_ftrace_function(ops); ftrace_startup(); - mutex_unlock(&ftrace_sysctl_lock); +out: + mutex_unlock(&ftrace_sysctl_lock); return ret; } @@ -1474,16 +1493,45 @@ ftrace_enable_sysctl(struct ctl_table *table, int write, } #ifdef CONFIG_FUNCTION_RET_TRACER + +/* The callback that hooks the return of a function */ trace_function_return_t ftrace_function_return = (trace_function_return_t)ftrace_stub; -void register_ftrace_return(trace_function_return_t func) + +int register_ftrace_return(trace_function_return_t func) { + int ret = 0; + + mutex_lock(&ftrace_sysctl_lock); + + /* + * Don't launch return tracing if normal function + * tracing is already running. + */ + if (ftrace_trace_function != ftrace_stub) { + ret = -EBUSY; + goto out; + } + + ftrace_tracing_type = FTRACE_TYPE_RETURN; ftrace_function_return = func; + ftrace_startup(); + +out: + mutex_unlock(&ftrace_sysctl_lock); + return ret; } void unregister_ftrace_return(void) { + mutex_lock(&ftrace_sysctl_lock); + ftrace_function_return = (trace_function_return_t)ftrace_stub; + ftrace_shutdown(); + /* Restore normal tracing type */ + ftrace_tracing_type = FTRACE_TYPE_ENTER; + + mutex_unlock(&ftrace_sysctl_lock); } #endif diff --git a/kernel/trace/trace_functions_return.c b/kernel/trace/trace_functions_return.c index 61185f756a13..a68564af022b 100644 --- a/kernel/trace/trace_functions_return.c +++ b/kernel/trace/trace_functions_return.c @@ -14,29 +14,18 @@ #include "trace.h" -static void start_return_trace(struct trace_array *tr) -{ - register_ftrace_return(&trace_function_return); -} - -static void stop_return_trace(struct trace_array *tr) -{ - unregister_ftrace_return(); -} - static int return_trace_init(struct trace_array *tr) { int cpu; for_each_online_cpu(cpu) tracing_reset(tr, cpu); - start_return_trace(tr); - return 0; + return register_ftrace_return(&trace_function_return); } static void return_trace_reset(struct trace_array *tr) { - stop_return_trace(tr); + unregister_ftrace_return(); } -- cgit v1.2.3 From 0bd7b79851d0f74b24a9ce87d088f2e7c718f668 Mon Sep 17 00:00:00 2001 From: Alexander van Heukelum Date: Sun, 16 Nov 2008 15:29:00 +0100 Subject: x86: entry_64.S: remove whitespace at end of lines Impact: cleanup All blame goes to: color white,red "[^[:graph:]]+$" in .nanorc ;). Signed-off-by: Alexander van Heukelum Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_64.S | 190 ++++++++++++++++++++++----------------------- 1 file changed, 95 insertions(+), 95 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index b86f332c96a6..54927784bab9 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -11,15 +11,15 @@ * * NOTE: This code handles signal-recognition, which happens every time * after an interrupt and after each system call. - * - * Normal syscalls and interrupts don't save a full stack frame, this is + * + * Normal syscalls and interrupts don't save a full stack frame, this is * only done for syscall tracing, signals or fork/exec et.al. - * - * A note on terminology: - * - top of stack: Architecture defined interrupt frame from SS to RIP - * at the top of the kernel process stack. + * + * A note on terminology: + * - top of stack: Architecture defined interrupt frame from SS to RIP + * at the top of the kernel process stack. * - partial stack frame: partially saved registers upto R11. - * - full stack frame: Like partial stack frame, but all register saved. + * - full stack frame: Like partial stack frame, but all register saved. * * Some macro usage: * - CFI macros are used to generate dwarf2 unwind information for better @@ -142,7 +142,7 @@ END(mcount) #ifndef CONFIG_PREEMPT #define retint_kernel retint_restore_args -#endif +#endif #ifdef CONFIG_PARAVIRT ENTRY(native_usergs_sysret64) @@ -161,14 +161,14 @@ ENTRY(native_usergs_sysret64) .endm /* - * C code is not supposed to know about undefined top of stack. Every time - * a C function with an pt_regs argument is called from the SYSCALL based + * C code is not supposed to know about undefined top of stack. Every time + * a C function with an pt_regs argument is called from the SYSCALL based * fast path FIXUP_TOP_OF_STACK is needed. * RESTORE_TOP_OF_STACK syncs the syscall state after any possible ptregs * manipulation. - */ - - /* %rsp:at FRAMEEND */ + */ + + /* %rsp:at FRAMEEND */ .macro FIXUP_TOP_OF_STACK tmp movq %gs:pda_oldrsp,\tmp movq \tmp,RSP(%rsp) @@ -244,8 +244,8 @@ ENTRY(native_usergs_sysret64) .endm /* * A newly forked process directly context switches into this. - */ -/* rdi: prev */ + */ +/* rdi: prev */ ENTRY(ret_from_fork) CFI_DEFAULT_STACK push kernel_eflags(%rip) @@ -256,7 +256,7 @@ ENTRY(ret_from_fork) GET_THREAD_INFO(%rcx) testl $(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT),TI_flags(%rcx) jnz rff_trace -rff_action: +rff_action: RESTORE_REST testl $3,CS-ARGOFFSET(%rsp) # from kernel_thread? je int_ret_from_sys_call @@ -267,7 +267,7 @@ rff_action: rff_trace: movq %rsp,%rdi call syscall_trace_leave - GET_THREAD_INFO(%rcx) + GET_THREAD_INFO(%rcx) jmp rff_action CFI_ENDPROC END(ret_from_fork) @@ -278,20 +278,20 @@ END(ret_from_fork) * SYSCALL does not save anything on the stack and does not change the * stack pointer. */ - + /* - * Register setup: + * Register setup: * rax system call number * rdi arg0 - * rcx return address for syscall/sysret, C arg3 + * rcx return address for syscall/sysret, C arg3 * rsi arg1 - * rdx arg2 + * rdx arg2 * r10 arg3 (--> moved to rcx for C) * r8 arg4 * r9 arg5 * r11 eflags for syscall/sysret, temporary for C - * r12-r15,rbp,rbx saved by C code, not touched. - * + * r12-r15,rbp,rbx saved by C code, not touched. + * * Interrupts are off on entry. * Only called from user space. * @@ -301,7 +301,7 @@ END(ret_from_fork) * When user can change the frames always force IRET. That is because * it deals with uncanonical addresses better. SYSRET has trouble * with them due to bugs in both AMD and Intel CPUs. - */ + */ ENTRY(system_call) CFI_STARTPROC simple @@ -317,7 +317,7 @@ ENTRY(system_call) */ ENTRY(system_call_after_swapgs) - movq %rsp,%gs:pda_oldrsp + movq %rsp,%gs:pda_oldrsp movq %gs:pda_kernelstack,%rsp /* * No need to follow this irqs off/on section - it's straight @@ -325,7 +325,7 @@ ENTRY(system_call_after_swapgs) */ ENABLE_INTERRUPTS(CLBR_NONE) SAVE_ARGS 8,1 - movq %rax,ORIG_RAX-ARGOFFSET(%rsp) + movq %rax,ORIG_RAX-ARGOFFSET(%rsp) movq %rcx,RIP-ARGOFFSET(%rsp) CFI_REL_OFFSET rip,RIP-ARGOFFSET GET_THREAD_INFO(%rcx) @@ -339,19 +339,19 @@ system_call_fastpath: movq %rax,RAX-ARGOFFSET(%rsp) /* * Syscall return path ending with SYSRET (fast path) - * Has incomplete stack frame and undefined top of stack. - */ + * Has incomplete stack frame and undefined top of stack. + */ ret_from_sys_call: movl $_TIF_ALLWORK_MASK,%edi /* edi: flagmask */ -sysret_check: +sysret_check: LOCKDEP_SYS_EXIT GET_THREAD_INFO(%rcx) DISABLE_INTERRUPTS(CLBR_NONE) TRACE_IRQS_OFF movl TI_flags(%rcx),%edx andl %edi,%edx - jnz sysret_careful + jnz sysret_careful CFI_REMEMBER_STATE /* * sysretq will re-enable interrupts: @@ -366,7 +366,7 @@ sysret_check: CFI_RESTORE_STATE /* Handle reschedules */ - /* edx: work, edi: workmask */ + /* edx: work, edi: workmask */ sysret_careful: bt $TIF_NEED_RESCHED,%edx jnc sysret_signal @@ -379,7 +379,7 @@ sysret_careful: CFI_ADJUST_CFA_OFFSET -8 jmp sysret_check - /* Handle a signal */ + /* Handle a signal */ sysret_signal: TRACE_IRQS_ON ENABLE_INTERRUPTS(CLBR_NONE) @@ -398,7 +398,7 @@ sysret_signal: DISABLE_INTERRUPTS(CLBR_NONE) TRACE_IRQS_OFF jmp int_with_check - + badsys: movq $-ENOSYS,RAX-ARGOFFSET(%rsp) jmp ret_from_sys_call @@ -437,7 +437,7 @@ sysret_audit: #endif /* CONFIG_AUDITSYSCALL */ /* Do syscall tracing */ -tracesys: +tracesys: #ifdef CONFIG_AUDITSYSCALL testl $(_TIF_WORK_SYSCALL_ENTRY & ~_TIF_SYSCALL_AUDIT),TI_flags(%rcx) jz auditsys @@ -460,8 +460,8 @@ tracesys: call *sys_call_table(,%rax,8) movq %rax,RAX-ARGOFFSET(%rsp) /* Use IRET because user could have changed frame */ - -/* + +/* * Syscall return path ending with IRET. * Has correct top of stack, but partial stack frame. */ @@ -505,18 +505,18 @@ int_very_careful: TRACE_IRQS_ON ENABLE_INTERRUPTS(CLBR_NONE) SAVE_REST - /* Check for syscall exit trace */ + /* Check for syscall exit trace */ testl $_TIF_WORK_SYSCALL_EXIT,%edx jz int_signal pushq %rdi CFI_ADJUST_CFA_OFFSET 8 - leaq 8(%rsp),%rdi # &ptregs -> arg1 + leaq 8(%rsp),%rdi # &ptregs -> arg1 call syscall_trace_leave popq %rdi CFI_ADJUST_CFA_OFFSET -8 andl $~(_TIF_WORK_SYSCALL_EXIT|_TIF_SYSCALL_EMU),%edi jmp int_restore_rest - + int_signal: testl $_TIF_DO_NOTIFY_MASK,%edx jz 1f @@ -531,11 +531,11 @@ int_restore_rest: jmp int_with_check CFI_ENDPROC END(system_call) - -/* + +/* * Certain special system calls that need to save a complete full stack frame. - */ - + */ + .macro PTREGSCALL label,func,arg .globl \label \label: @@ -572,7 +572,7 @@ ENTRY(ptregscall_common) ret CFI_ENDPROC END(ptregscall_common) - + ENTRY(stub_execve) CFI_STARTPROC popq %r11 @@ -588,11 +588,11 @@ ENTRY(stub_execve) jmp int_ret_from_sys_call CFI_ENDPROC END(stub_execve) - + /* * sigreturn is special because it needs to restore all registers on return. * This cannot be done with SYSRET, so use the IRET return path instead. - */ + */ ENTRY(stub_rt_sigreturn) CFI_STARTPROC addq $8, %rsp @@ -685,12 +685,12 @@ exit_intr: GET_THREAD_INFO(%rcx) testl $3,CS-ARGOFFSET(%rsp) je retint_kernel - + /* Interrupt came from user space */ /* * Has a correct top of stack, but a partial stack frame * %rcx: thread info. Interrupts off. - */ + */ retint_with_reschedule: movl $_TIF_WORK_MASK,%edi retint_check: @@ -763,20 +763,20 @@ retint_careful: pushq %rdi CFI_ADJUST_CFA_OFFSET 8 call schedule - popq %rdi + popq %rdi CFI_ADJUST_CFA_OFFSET -8 GET_THREAD_INFO(%rcx) DISABLE_INTERRUPTS(CLBR_NONE) TRACE_IRQS_OFF jmp retint_check - + retint_signal: testl $_TIF_DO_NOTIFY_MASK,%edx jz retint_swapgs TRACE_IRQS_ON ENABLE_INTERRUPTS(CLBR_NONE) SAVE_REST - movq $-1,ORIG_RAX(%rsp) + movq $-1,ORIG_RAX(%rsp) xorl %esi,%esi # oldset movq %rsp,%rdi # &pt_regs call do_notify_resume @@ -798,14 +798,14 @@ ENTRY(retint_kernel) jnc retint_restore_args call preempt_schedule_irq jmp exit_intr -#endif +#endif CFI_ENDPROC END(common_interrupt) - + /* * APIC interrupts. - */ + */ .macro apicinterrupt num,func INTR_FRAME pushq $~(\num) @@ -823,14 +823,14 @@ ENTRY(threshold_interrupt) apicinterrupt THRESHOLD_APIC_VECTOR,mce_threshold_interrupt END(threshold_interrupt) -#ifdef CONFIG_SMP +#ifdef CONFIG_SMP ENTRY(reschedule_interrupt) apicinterrupt RESCHEDULE_VECTOR,smp_reschedule_interrupt END(reschedule_interrupt) .macro INVALIDATE_ENTRY num ENTRY(invalidate_interrupt\num) - apicinterrupt INVALIDATE_TLB_VECTOR_START+\num,smp_invalidate_interrupt + apicinterrupt INVALIDATE_TLB_VECTOR_START+\num,smp_invalidate_interrupt END(invalidate_interrupt\num) .endm @@ -869,22 +869,22 @@ END(error_interrupt) ENTRY(spurious_interrupt) apicinterrupt SPURIOUS_APIC_VECTOR,smp_spurious_interrupt END(spurious_interrupt) - + /* * Exception entry points. - */ + */ .macro zeroentry sym INTR_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME - pushq $0 /* push error code/oldrax */ + pushq $0 /* push error code/oldrax */ CFI_ADJUST_CFA_OFFSET 8 - pushq %rax /* push real oldrax to the rdi slot */ + pushq %rax /* push real oldrax to the rdi slot */ CFI_ADJUST_CFA_OFFSET 8 CFI_REL_OFFSET rax,0 leaq \sym(%rip),%rax jmp error_entry CFI_ENDPROC - .endm + .endm .macro errorentry sym XCPT_FRAME @@ -998,13 +998,13 @@ paranoid_schedule\trace: /* * Exception entry point. This expects an error code/orig_rax on the stack - * and the exception handler in %rax. - */ + * and the exception handler in %rax. + */ KPROBE_ENTRY(error_entry) _frame RDI CFI_REL_OFFSET rax,0 /* rdi slot contains rax, oldrax contains error code */ - cld + cld subq $14*8,%rsp CFI_ADJUST_CFA_OFFSET (14*8) movq %rsi,13*8(%rsp) @@ -1015,7 +1015,7 @@ KPROBE_ENTRY(error_entry) CFI_REL_OFFSET rdx,RDX movq %rcx,11*8(%rsp) CFI_REL_OFFSET rcx,RCX - movq %rsi,10*8(%rsp) /* store rax */ + movq %rsi,10*8(%rsp) /* store rax */ CFI_REL_OFFSET rax,RAX movq %r8, 9*8(%rsp) CFI_REL_OFFSET r8,R8 @@ -1025,29 +1025,29 @@ KPROBE_ENTRY(error_entry) CFI_REL_OFFSET r10,R10 movq %r11,6*8(%rsp) CFI_REL_OFFSET r11,R11 - movq %rbx,5*8(%rsp) + movq %rbx,5*8(%rsp) CFI_REL_OFFSET rbx,RBX - movq %rbp,4*8(%rsp) + movq %rbp,4*8(%rsp) CFI_REL_OFFSET rbp,RBP - movq %r12,3*8(%rsp) + movq %r12,3*8(%rsp) CFI_REL_OFFSET r12,R12 - movq %r13,2*8(%rsp) + movq %r13,2*8(%rsp) CFI_REL_OFFSET r13,R13 - movq %r14,1*8(%rsp) + movq %r14,1*8(%rsp) CFI_REL_OFFSET r14,R14 - movq %r15,(%rsp) + movq %r15,(%rsp) CFI_REL_OFFSET r15,R15 - xorl %ebx,%ebx + xorl %ebx,%ebx testl $3,CS(%rsp) je error_kernelspace -error_swapgs: +error_swapgs: SWAPGS error_sti: TRACE_IRQS_OFF - movq %rdi,RDI(%rsp) + movq %rdi,RDI(%rsp) CFI_REL_OFFSET rdi,RDI movq %rsp,%rdi - movq ORIG_RAX(%rsp),%rsi /* get error code */ + movq ORIG_RAX(%rsp),%rsi /* get error code */ movq $-1,ORIG_RAX(%rsp) call *%rax /* ebx: no swapgs flag (1: don't need swapgs, 0: need it) */ @@ -1056,7 +1056,7 @@ error_exit: RESTORE_REST DISABLE_INTERRUPTS(CLBR_NONE) TRACE_IRQS_OFF - GET_THREAD_INFO(%rcx) + GET_THREAD_INFO(%rcx) testl %eax,%eax jne retint_kernel LOCKDEP_SYS_EXIT_IRQ @@ -1072,7 +1072,7 @@ error_kernelspace: /* There are two places in the kernel that can potentially fault with usergs. Handle them here. The exception handlers after iret run with kernel gs again, so don't set the user space flag. - B stepping K8s sometimes report an truncated RIP for IRET + B stepping K8s sometimes report an truncated RIP for IRET exceptions returning to compat mode. Check for these here too. */ leaq irq_return(%rip),%rcx cmpq %rcx,RIP(%rsp) @@ -1084,17 +1084,17 @@ error_kernelspace: je error_swapgs jmp error_sti KPROBE_END(error_entry) - + /* Reload gs selector with exception handling */ - /* edi: new selector */ + /* edi: new selector */ ENTRY(native_load_gs_index) CFI_STARTPROC pushf CFI_ADJUST_CFA_OFFSET 8 DISABLE_INTERRUPTS(CLBR_ANY | ~(CLBR_RDI)) SWAPGS -gs_change: - movl %edi,%gs +gs_change: + movl %edi,%gs 2: mfence /* workaround */ SWAPGS popf @@ -1102,20 +1102,20 @@ gs_change: ret CFI_ENDPROC ENDPROC(native_load_gs_index) - + .section __ex_table,"a" .align 8 .quad gs_change,bad_gs .previous .section .fixup,"ax" /* running with kernelgs */ -bad_gs: +bad_gs: SWAPGS /* switch back to user gs */ xorl %eax,%eax movl %eax,%gs jmp 2b - .previous - + .previous + /* * Create a kernel thread. * @@ -1138,7 +1138,7 @@ ENTRY(kernel_thread) xorl %r8d,%r8d xorl %r9d,%r9d - + # clone now call do_fork movq %rax,RAX(%rsp) @@ -1149,14 +1149,14 @@ ENTRY(kernel_thread) * so internally to the x86_64 port you can rely on kernel_thread() * not to reschedule the child before returning, this avoids the need * of hacks for example to fork off the per-CPU idle tasks. - * [Hopefully no generic code relies on the reschedule -AK] + * [Hopefully no generic code relies on the reschedule -AK] */ RESTORE_ALL UNFAKE_STACK_FRAME ret CFI_ENDPROC ENDPROC(kernel_thread) - + child_rip: pushq $0 # fake return address CFI_STARTPROC @@ -1191,10 +1191,10 @@ ENDPROC(child_rip) ENTRY(kernel_execve) CFI_STARTPROC FAKE_STACK_FRAME $0 - SAVE_ALL + SAVE_ALL movq %rsp,%rcx call sys_execve - movq %rax, RAX(%rsp) + movq %rax, RAX(%rsp) RESTORE_REST testq %rax,%rax je int_ret_from_sys_call @@ -1213,7 +1213,7 @@ ENTRY(coprocessor_error) END(coprocessor_error) ENTRY(simd_coprocessor_error) - zeroentry do_simd_coprocessor_error + zeroentry do_simd_coprocessor_error END(simd_coprocessor_error) ENTRY(device_not_available) @@ -1225,12 +1225,12 @@ KPROBE_ENTRY(debug) INTR_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME pushq $0 - CFI_ADJUST_CFA_OFFSET 8 + CFI_ADJUST_CFA_OFFSET 8 paranoidentry do_debug, DEBUG_STACK paranoidexit KPROBE_END(debug) - /* runs on exception stack */ + /* runs on exception stack */ KPROBE_ENTRY(nmi) INTR_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME @@ -1264,7 +1264,7 @@ ENTRY(bounds) END(bounds) ENTRY(invalid_op) - zeroentry do_invalid_op + zeroentry do_invalid_op END(invalid_op) ENTRY(coprocessor_segment_overrun) @@ -1319,7 +1319,7 @@ ENTRY(machine_check) INTR_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME pushq $0 - CFI_ADJUST_CFA_OFFSET 8 + CFI_ADJUST_CFA_OFFSET 8 paranoidentry do_machine_check jmp paranoid_exit1 CFI_ENDPROC -- cgit v1.2.3 From 569712b2b0970fa5b19673544d62ae661d04a220 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Sun, 16 Nov 2008 03:12:49 -0800 Subject: x86: fix wakeup_cpu with numaq/es7000, v2 Impact: fix secondary-CPU wakeup/init path with numaq and es7000 While looking at wakeup_secondary_cpu for WAKE_SECONDARY_VIA_NMI: |#ifdef WAKE_SECONDARY_VIA_NMI |/* | * Poke the other CPU in the eye via NMI to wake it up. Remember that the normal | * INIT, INIT, STARTUP sequence will reset the chip hard for us, and this | * won't ... remember to clear down the APIC, etc later. | */ |static int __devinit |wakeup_secondary_cpu(int logical_apicid, unsigned long start_eip) |{ | unsigned long send_status, accept_status = 0; | int maxlvt; |... | if (APIC_INTEGRATED(apic_version[phys_apicid])) { | maxlvt = lapic_get_maxlvt(); I noticed that there is no warning about undefined phys_apicid... because WAKE_SECONDARY_VIA_NMI and WAKE_SECONDARY_VIA_INIT can not be defined at the same time. So NUMAQ is using wrong wakeup_secondary_cpu. WAKE_SECONDARY_VIA_NMI, WAKE_SECONDARY_VIA_INIT and WAKE_SECONDARY_VIA_MIP are variants of a weird and fragile preprocessor-driven "HAL" mechanisms to specify the kind of secondary-CPU wakeup strategy a given x86 kernel will use. The vast majority of systems want to use INIT for secondary wakeup - NUMAQ uses an NMI, (old-style-) ES7000 uses 'MIP' (a firmware driven in-memory flag to let secondaries continue). So convert these mechanisms to x86_quirks and add a ->wakeup_secondary_cpu() method to specify the rare exception to the sane default. Extend genapic accordingly as well, for 32-bit. While looking further, I noticed that functions in wakecup.h for numaq and es7000 are different to the default in mach_wakecpu.h - but smpboot.c will only use default mach_wakecpu.h with smphook.h. So we need to add mach_wakecpu.h for mach_generic, to properly support numaq and es7000, and vectorize the following SMP init methods: int trampoline_phys_low; int trampoline_phys_high; void (*wait_for_init_deassert)(atomic_t *deassert); void (*smp_callin_clear_local_apic)(void); void (*store_NMI_vector)(unsigned short *high, unsigned short *low); void (*restore_NMI_vector)(unsigned short *high, unsigned short *low); void (*inquire_remote_apic)(int apicid); Signed-off-by: Yinghai Lu Signed-off-by: Ingo Molnar --- arch/x86/include/asm/bigsmp/apic.h | 2 -- arch/x86/include/asm/es7000/apic.h | 3 -- arch/x86/include/asm/es7000/wakecpu.h | 41 ++++++--------------- arch/x86/include/asm/genapic_32.h | 17 ++++++++- arch/x86/include/asm/mach-default/mach_wakecpu.h | 24 +++++-------- arch/x86/include/asm/mach-default/smpboot_hooks.h | 8 +++-- arch/x86/include/asm/mach-generic/mach_wakecpu.h | 12 +++++++ arch/x86/include/asm/numaq/wakecpu.h | 24 +++++++------ arch/x86/include/asm/setup.h | 2 ++ arch/x86/kernel/es7000_32.c | 44 ++++++++++++----------- arch/x86/kernel/numaq_32.c | 1 + arch/x86/kernel/smpboot.c | 24 ++++++++----- arch/x86/mach-generic/bigsmp.c | 1 + arch/x86/mach-generic/default.c | 1 + arch/x86/mach-generic/summit.c | 1 + 15 files changed, 110 insertions(+), 95 deletions(-) create mode 100644 arch/x86/include/asm/mach-generic/mach_wakecpu.h (limited to 'arch/x86') diff --git a/arch/x86/include/asm/bigsmp/apic.h b/arch/x86/include/asm/bigsmp/apic.h index 1d9543b9d358..ce547f24a1cd 100644 --- a/arch/x86/include/asm/bigsmp/apic.h +++ b/arch/x86/include/asm/bigsmp/apic.h @@ -24,8 +24,6 @@ static inline cpumask_t target_cpus(void) #define INT_DELIVERY_MODE (dest_Fixed) #define INT_DEST_MODE (0) /* phys delivery to target proc */ #define NO_BALANCE_IRQ (0) -#define WAKE_SECONDARY_VIA_INIT - static inline unsigned long check_apicid_used(physid_mask_t bitmap, int apicid) { diff --git a/arch/x86/include/asm/es7000/apic.h b/arch/x86/include/asm/es7000/apic.h index 380f0b4f17ed..9d8cf776c285 100644 --- a/arch/x86/include/asm/es7000/apic.h +++ b/arch/x86/include/asm/es7000/apic.h @@ -23,8 +23,6 @@ static inline cpumask_t target_cpus(void) #define INT_DELIVERY_MODE (dest_LowestPrio) #define INT_DEST_MODE (1) /* logical delivery broadcast to all procs */ #define NO_BALANCE_IRQ (1) -#undef WAKE_SECONDARY_VIA_INIT -#define WAKE_SECONDARY_VIA_MIP #else #define APIC_DFR_VALUE (APIC_DFR_FLAT) #define INT_DELIVERY_MODE (dest_Fixed) @@ -32,7 +30,6 @@ static inline cpumask_t target_cpus(void) #define NO_BALANCE_IRQ (0) #undef APIC_DEST_LOGICAL #define APIC_DEST_LOGICAL 0x0 -#define WAKE_SECONDARY_VIA_INIT #endif static inline unsigned long check_apicid_used(physid_mask_t bitmap, int apicid) diff --git a/arch/x86/include/asm/es7000/wakecpu.h b/arch/x86/include/asm/es7000/wakecpu.h index 398493461913..78f0daaee436 100644 --- a/arch/x86/include/asm/es7000/wakecpu.h +++ b/arch/x86/include/asm/es7000/wakecpu.h @@ -1,36 +1,12 @@ #ifndef __ASM_ES7000_WAKECPU_H #define __ASM_ES7000_WAKECPU_H -/* - * This file copes with machines that wakeup secondary CPUs by the - * INIT, INIT, STARTUP sequence. - */ - -#ifdef CONFIG_ES7000_CLUSTERED_APIC -#define WAKE_SECONDARY_VIA_MIP -#else -#define WAKE_SECONDARY_VIA_INIT -#endif - -#ifdef WAKE_SECONDARY_VIA_MIP -extern int es7000_start_cpu(int cpu, unsigned long eip); -static inline int -wakeup_secondary_cpu(int phys_apicid, unsigned long start_eip) -{ - int boot_error = 0; - boot_error = es7000_start_cpu(phys_apicid, start_eip); - return boot_error; -} -#endif - -#define TRAMPOLINE_LOW phys_to_virt(0x467) -#define TRAMPOLINE_HIGH phys_to_virt(0x469) - -#define boot_cpu_apicid boot_cpu_physical_apicid +#define TRAMPOLINE_PHYS_LOW 0x467 +#define TRAMPOLINE_PHYS_HIGH 0x469 static inline void wait_for_init_deassert(atomic_t *deassert) { -#ifdef WAKE_SECONDARY_VIA_INIT +#ifndef CONFIG_ES7000_CLUSTERED_APIC while (!atomic_read(deassert)) cpu_relax(); #endif @@ -50,9 +26,12 @@ static inline void restore_NMI_vector(unsigned short *high, unsigned short *low) { } -#define inquire_remote_apic(apicid) do { \ - if (apic_verbosity >= APIC_DEBUG) \ - __inquire_remote_apic(apicid); \ - } while (0) +extern void __inquire_remote_apic(int apicid); + +static inline void inquire_remote_apic(int apicid) +{ + if (apic_verbosity >= APIC_DEBUG) + __inquire_remote_apic(apicid); +} #endif /* __ASM_MACH_WAKECPU_H */ diff --git a/arch/x86/include/asm/genapic_32.h b/arch/x86/include/asm/genapic_32.h index 5cbd4fcc06fd..39bd8c1db3f5 100644 --- a/arch/x86/include/asm/genapic_32.h +++ b/arch/x86/include/asm/genapic_32.h @@ -2,6 +2,7 @@ #define _ASM_X86_GENAPIC_32_H #include +#include /* * Generic APIC driver interface. @@ -65,6 +66,13 @@ struct genapic { void (*send_IPI_allbutself)(int vector); void (*send_IPI_all)(int vector); #endif + int trampoline_phys_low; + int trampoline_phys_high; + void (*wait_for_init_deassert)(atomic_t *deassert); + void (*smp_callin_clear_local_apic)(void); + void (*store_NMI_vector)(unsigned short *high, unsigned short *low); + void (*restore_NMI_vector)(unsigned short *high, unsigned short *low); + void (*inquire_remote_apic)(int apicid); }; #define APICFUNC(x) .x = x, @@ -105,13 +113,20 @@ struct genapic { APICFUNC(get_apic_id) \ .apic_id_mask = APIC_ID_MASK, \ APICFUNC(cpu_mask_to_apicid) \ - APICFUNC(vector_allocation_domain) \ + APICFUNC(vector_allocation_domain) \ APICFUNC(acpi_madt_oem_check) \ IPIFUNC(send_IPI_mask) \ IPIFUNC(send_IPI_allbutself) \ IPIFUNC(send_IPI_all) \ APICFUNC(enable_apic_mode) \ APICFUNC(phys_pkg_id) \ + .trampoline_phys_low = TRAMPOLINE_PHYS_LOW, \ + .trampoline_phys_high = TRAMPOLINE_PHYS_HIGH, \ + APICFUNC(wait_for_init_deassert) \ + APICFUNC(smp_callin_clear_local_apic) \ + APICFUNC(store_NMI_vector) \ + APICFUNC(restore_NMI_vector) \ + APICFUNC(inquire_remote_apic) \ } extern struct genapic *genapic; diff --git a/arch/x86/include/asm/mach-default/mach_wakecpu.h b/arch/x86/include/asm/mach-default/mach_wakecpu.h index 9d80db91e992..ceb013660146 100644 --- a/arch/x86/include/asm/mach-default/mach_wakecpu.h +++ b/arch/x86/include/asm/mach-default/mach_wakecpu.h @@ -1,17 +1,8 @@ #ifndef _ASM_X86_MACH_DEFAULT_MACH_WAKECPU_H #define _ASM_X86_MACH_DEFAULT_MACH_WAKECPU_H -/* - * This file copes with machines that wakeup secondary CPUs by the - * INIT, INIT, STARTUP sequence. - */ - -#define WAKE_SECONDARY_VIA_INIT - -#define TRAMPOLINE_LOW phys_to_virt(0x467) -#define TRAMPOLINE_HIGH phys_to_virt(0x469) - -#define boot_cpu_apicid boot_cpu_physical_apicid +#define TRAMPOLINE_PHYS_LOW (0x467) +#define TRAMPOLINE_PHYS_HIGH (0x469) static inline void wait_for_init_deassert(atomic_t *deassert) { @@ -33,9 +24,12 @@ static inline void restore_NMI_vector(unsigned short *high, unsigned short *low) { } -#define inquire_remote_apic(apicid) do { \ - if (apic_verbosity >= APIC_DEBUG) \ - __inquire_remote_apic(apicid); \ - } while (0) +extern void __inquire_remote_apic(int apicid); + +static inline void inquire_remote_apic(int apicid) +{ + if (apic_verbosity >= APIC_DEBUG) + __inquire_remote_apic(apicid); +} #endif /* _ASM_X86_MACH_DEFAULT_MACH_WAKECPU_H */ diff --git a/arch/x86/include/asm/mach-default/smpboot_hooks.h b/arch/x86/include/asm/mach-default/smpboot_hooks.h index dbab36d64d48..23bf52103b89 100644 --- a/arch/x86/include/asm/mach-default/smpboot_hooks.h +++ b/arch/x86/include/asm/mach-default/smpboot_hooks.h @@ -13,9 +13,11 @@ static inline void smpboot_setup_warm_reset_vector(unsigned long start_eip) CMOS_WRITE(0xa, 0xf); local_flush_tlb(); pr_debug("1.\n"); - *((volatile unsigned short *) TRAMPOLINE_HIGH) = start_eip >> 4; + *((volatile unsigned short *)phys_to_virt(TRAMPOLINE_PHYS_HIGH)) = + start_eip >> 4; pr_debug("2.\n"); - *((volatile unsigned short *) TRAMPOLINE_LOW) = start_eip & 0xf; + *((volatile unsigned short *)phys_to_virt(TRAMPOLINE_PHYS_LOW)) = + start_eip & 0xf; pr_debug("3.\n"); } @@ -32,7 +34,7 @@ static inline void smpboot_restore_warm_reset_vector(void) */ CMOS_WRITE(0, 0xf); - *((volatile long *) phys_to_virt(0x467)) = 0; + *((volatile long *)phys_to_virt(TRAMPOLINE_PHYS_LOW)) = 0; } static inline void __init smpboot_setup_io_apic(void) diff --git a/arch/x86/include/asm/mach-generic/mach_wakecpu.h b/arch/x86/include/asm/mach-generic/mach_wakecpu.h new file mode 100644 index 000000000000..1ab16b168c8a --- /dev/null +++ b/arch/x86/include/asm/mach-generic/mach_wakecpu.h @@ -0,0 +1,12 @@ +#ifndef _ASM_X86_MACH_GENERIC_MACH_WAKECPU_H +#define _ASM_X86_MACH_GENERIC_MACH_WAKECPU_H + +#define TRAMPOLINE_PHYS_LOW (genapic->trampoline_phys_low) +#define TRAMPOLINE_PHYS_HIGH (genapic->trampoline_phys_high) +#define wait_for_init_deassert (genapic->wait_for_init_deassert) +#define smp_callin_clear_local_apic (genapic->smp_callin_clear_local_apic) +#define store_NMI_vector (genapic->store_NMI_vector) +#define restore_NMI_vector (genapic->restore_NMI_vector) +#define inquire_remote_apic (genapic->inquire_remote_apic) + +#endif /* _ASM_X86_MACH_GENERIC_MACH_APIC_H */ diff --git a/arch/x86/include/asm/numaq/wakecpu.h b/arch/x86/include/asm/numaq/wakecpu.h index c577bda5b1c5..6f499df8eddb 100644 --- a/arch/x86/include/asm/numaq/wakecpu.h +++ b/arch/x86/include/asm/numaq/wakecpu.h @@ -3,12 +3,8 @@ /* This file copes with machines that wakeup secondary CPUs by NMIs */ -#define WAKE_SECONDARY_VIA_NMI - -#define TRAMPOLINE_LOW phys_to_virt(0x8) -#define TRAMPOLINE_HIGH phys_to_virt(0xa) - -#define boot_cpu_apicid boot_cpu_logical_apicid +#define TRAMPOLINE_PHYS_LOW (0x8) +#define TRAMPOLINE_PHYS_HIGH (0xa) /* We don't do anything here because we use NMI's to boot instead */ static inline void wait_for_init_deassert(atomic_t *deassert) @@ -27,17 +23,23 @@ static inline void smp_callin_clear_local_apic(void) static inline void store_NMI_vector(unsigned short *high, unsigned short *low) { printk("Storing NMI vector\n"); - *high = *((volatile unsigned short *) TRAMPOLINE_HIGH); - *low = *((volatile unsigned short *) TRAMPOLINE_LOW); + *high = + *((volatile unsigned short *)phys_to_virt(TRAMPOLINE_PHYS_HIGH)); + *low = + *((volatile unsigned short *)phys_to_virt(TRAMPOLINE_PHYS_LOW)); } static inline void restore_NMI_vector(unsigned short *high, unsigned short *low) { printk("Restoring NMI vector\n"); - *((volatile unsigned short *) TRAMPOLINE_HIGH) = *high; - *((volatile unsigned short *) TRAMPOLINE_LOW) = *low; + *((volatile unsigned short *)phys_to_virt(TRAMPOLINE_PHYS_HIGH)) = + *high; + *((volatile unsigned short *)phys_to_virt(TRAMPOLINE_PHYS_LOW)) = + *low; } -#define inquire_remote_apic(apicid) {} +static inline void inquire_remote_apic(int apicid) +{ +} #endif /* __ASM_NUMAQ_WAKECPU_H */ diff --git a/arch/x86/include/asm/setup.h b/arch/x86/include/asm/setup.h index f12d37237465..40b2d3304911 100644 --- a/arch/x86/include/asm/setup.h +++ b/arch/x86/include/asm/setup.h @@ -16,6 +16,7 @@ static inline void visws_early_detect(void) { } static inline int is_visws_box(void) { return 0; } #endif +extern int wakeup_secondary_cpu_via_nmi(int apicid, unsigned long start_eip); /* * Any setup quirks to be performed? */ @@ -39,6 +40,7 @@ struct x86_quirks { void (*smp_read_mpc_oem)(struct mp_config_oemtable *oemtable, unsigned short oemsize); int (*setup_ioapic_ids)(void); + int (*wakeup_secondary_cpu)(int apicid, unsigned long start_eip); }; extern struct x86_quirks *x86_quirks; diff --git a/arch/x86/kernel/es7000_32.c b/arch/x86/kernel/es7000_32.c index f454c78fcef6..bed10dddf099 100644 --- a/arch/x86/kernel/es7000_32.c +++ b/arch/x86/kernel/es7000_32.c @@ -40,6 +40,7 @@ #include #include #include +#include /* * ES7000 chipsets @@ -161,6 +162,26 @@ es7000_rename_gsi(int ioapic, int gsi) return gsi; } +#ifdef CONFIG_ES7000_CLUSTERED_APIC +static int wakeup_secondary_cpu_via_mip(int cpu, unsigned long eip) +{ + unsigned long vect = 0, psaival = 0; + + if (psai == NULL) + return -1; + + vect = ((unsigned long)__pa(eip)/0x1000) << 16; + psaival = (0x1000000 | vect | cpu); + + while (*psai & 0x1000000) + ; + + *psai = psaival; + + return 0; +} +#endif + void __init setup_unisys(void) { @@ -176,6 +197,9 @@ setup_unisys(void) else es7000_plat = ES7000_CLASSIC; ioapic_renumber_irq = es7000_rename_gsi; +#ifdef CONFIG_ES7000_CLUSTERED_APIC + x86_quirks->wakeup_secondary_cpu = wakeup_secondary_cpu_via_mip; +#endif } /* @@ -324,26 +348,6 @@ es7000_mip_write(struct mip_reg *mip_reg) return status; } -int -es7000_start_cpu(int cpu, unsigned long eip) -{ - unsigned long vect = 0, psaival = 0; - - if (psai == NULL) - return -1; - - vect = ((unsigned long)__pa(eip)/0x1000) << 16; - psaival = (0x1000000 | vect | cpu); - - while (*psai & 0x1000000) - ; - - *psai = psaival; - - return 0; - -} - void __init es7000_sw_apic(void) { diff --git a/arch/x86/kernel/numaq_32.c b/arch/x86/kernel/numaq_32.c index 4caff39078e0..745891b7d0fb 100644 --- a/arch/x86/kernel/numaq_32.c +++ b/arch/x86/kernel/numaq_32.c @@ -250,6 +250,7 @@ static struct x86_quirks numaq_x86_quirks __initdata = { .mpc_oem_pci_bus = mpc_oem_pci_bus, .smp_read_mpc_oem = smp_read_mpc_oem, .setup_ioapic_ids = numaq_setup_ioapic_ids, + .wakeup_secondary_cpu = wakeup_secondary_cpu_via_nmi, }; void numaq_mps_oem_check(struct mp_config_table *mpc, char *oem, diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index 7b1093397319..498c1ef37fe0 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -62,6 +62,7 @@ #include #include #include +#include #include #include @@ -536,7 +537,7 @@ static void impress_friends(void) pr_debug("Before bogocount - setting activated=1.\n"); } -static inline void __inquire_remote_apic(int apicid) +void __inquire_remote_apic(int apicid) { unsigned i, regs[] = { APIC_ID >> 4, APIC_LVR >> 4, APIC_SPIV >> 4 }; char *names[] = { "ID", "VERSION", "SPIV" }; @@ -575,14 +576,13 @@ static inline void __inquire_remote_apic(int apicid) } } -#ifdef WAKE_SECONDARY_VIA_NMI /* * Poke the other CPU in the eye via NMI to wake it up. Remember that the normal * INIT, INIT, STARTUP sequence will reset the chip hard for us, and this * won't ... remember to clear down the APIC, etc later. */ -static int __devinit -wakeup_secondary_cpu(int logical_apicid, unsigned long start_eip) +int __devinit +wakeup_secondary_cpu_via_nmi(int logical_apicid, unsigned long start_eip) { unsigned long send_status, accept_status = 0; int maxlvt; @@ -599,7 +599,7 @@ wakeup_secondary_cpu(int logical_apicid, unsigned long start_eip) * Give the other CPU some time to accept the IPI. */ udelay(200); - if (APIC_INTEGRATED(apic_version[phys_apicid])) { + if (APIC_INTEGRATED(apic_version[boot_cpu_physical_apicid])) { maxlvt = lapic_get_maxlvt(); if (maxlvt > 3) /* Due to the Pentium erratum 3AP. */ apic_write(APIC_ESR, 0); @@ -614,11 +614,9 @@ wakeup_secondary_cpu(int logical_apicid, unsigned long start_eip) return (send_status | accept_status); } -#endif /* WAKE_SECONDARY_VIA_NMI */ -#ifdef WAKE_SECONDARY_VIA_INIT static int __devinit -wakeup_secondary_cpu(int phys_apicid, unsigned long start_eip) +wakeup_secondary_cpu_via_init(int phys_apicid, unsigned long start_eip) { unsigned long send_status, accept_status = 0; int maxlvt, num_starts, j; @@ -737,7 +735,15 @@ wakeup_secondary_cpu(int phys_apicid, unsigned long start_eip) return (send_status | accept_status); } -#endif /* WAKE_SECONDARY_VIA_INIT */ + +static int __devinit +wakeup_secondary_cpu(int apicid, unsigned long start_eip) +{ + if (x86_quirks->wakeup_secondary_cpu) + return x86_quirks->wakeup_secondary_cpu(apicid, start_eip); + + return wakeup_secondary_cpu_via_init(apicid, start_eip); +} struct create_idle { struct work_struct work; diff --git a/arch/x86/mach-generic/bigsmp.c b/arch/x86/mach-generic/bigsmp.c index 3c3b471ea496..3624a364b7f3 100644 --- a/arch/x86/mach-generic/bigsmp.c +++ b/arch/x86/mach-generic/bigsmp.c @@ -17,6 +17,7 @@ #include #include #include +#include static int dmi_bigsmp; /* can be set by dmi scanners */ diff --git a/arch/x86/mach-generic/default.c b/arch/x86/mach-generic/default.c index 9e835a11a13a..e63a4a76d8cd 100644 --- a/arch/x86/mach-generic/default.c +++ b/arch/x86/mach-generic/default.c @@ -16,6 +16,7 @@ #include #include #include +#include /* should be called last. */ static int probe_default(void) diff --git a/arch/x86/mach-generic/summit.c b/arch/x86/mach-generic/summit.c index 6272b5e69da6..2c6d234e0009 100644 --- a/arch/x86/mach-generic/summit.c +++ b/arch/x86/mach-generic/summit.c @@ -16,6 +16,7 @@ #include #include #include +#include static int probe_summit(void) { -- cgit v1.2.3 From 54ac14a8e982ae6c7ac71ee2b0d0173b974509e2 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Mon, 17 Nov 2008 15:19:53 -0800 Subject: x86: fix wakeup_cpu with numaq/es7000, v2, fix Impact: fix wakeup_secondary_cpu with hotplug We can not put that into x86_quirks, because that is __initdata. So try to move that to genapic, and add update_genapic in x86_quirks. later we even could use that stub to: 1. autodetect CONFIG_ES7000_CLUSTERED_APIC 2. more correct inquire_remote_apic with apic_verbosity setting. Signed-off-by: Yinghai Lu Signed-off-by: Ingo Molnar --- arch/x86/include/asm/genapic_32.h | 1 + arch/x86/include/asm/genapic_64.h | 2 ++ arch/x86/include/asm/mach-default/mach_apic.h | 2 ++ arch/x86/include/asm/mach-generic/mach_apic.h | 1 + arch/x86/include/asm/setup.h | 3 ++- arch/x86/kernel/es7000_32.c | 11 ++++++++++- arch/x86/kernel/genapic_64.c | 4 ++++ arch/x86/kernel/numaq_32.c | 11 +++++++++-- arch/x86/kernel/setup.c | 13 ++++++++++++- arch/x86/kernel/smpboot.c | 11 +---------- arch/x86/mach-generic/probe.c | 4 ++++ 11 files changed, 48 insertions(+), 15 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/genapic_32.h b/arch/x86/include/asm/genapic_32.h index 39bd8c1db3f5..455d6c27a98b 100644 --- a/arch/x86/include/asm/genapic_32.h +++ b/arch/x86/include/asm/genapic_32.h @@ -66,6 +66,7 @@ struct genapic { void (*send_IPI_allbutself)(int vector); void (*send_IPI_all)(int vector); #endif + int (*wakeup_cpu)(int apicid, unsigned long start_eip); int trampoline_phys_low; int trampoline_phys_high; void (*wait_for_init_deassert)(atomic_t *deassert); diff --git a/arch/x86/include/asm/genapic_64.h b/arch/x86/include/asm/genapic_64.h index 13c4e96199ea..2cae011668b7 100644 --- a/arch/x86/include/asm/genapic_64.h +++ b/arch/x86/include/asm/genapic_64.h @@ -32,6 +32,8 @@ struct genapic { unsigned int (*get_apic_id)(unsigned long x); unsigned long (*set_apic_id)(unsigned int id); unsigned long apic_id_mask; + /* wakeup_secondary_cpu */ + int (*wakeup_cpu)(int apicid, unsigned long start_eip); }; extern struct genapic *genapic; diff --git a/arch/x86/include/asm/mach-default/mach_apic.h b/arch/x86/include/asm/mach-default/mach_apic.h index ff3a6c236c00..6cb3a467e067 100644 --- a/arch/x86/include/asm/mach-default/mach_apic.h +++ b/arch/x86/include/asm/mach-default/mach_apic.h @@ -32,11 +32,13 @@ static inline cpumask_t target_cpus(void) #define vector_allocation_domain (genapic->vector_allocation_domain) #define read_apic_id() (GET_APIC_ID(apic_read(APIC_ID))) #define send_IPI_self (genapic->send_IPI_self) +#define wakeup_secondary_cpu (genapic->wakeup_cpu) extern void setup_apic_routing(void); #else #define INT_DELIVERY_MODE dest_LowestPrio #define INT_DEST_MODE 1 /* logical delivery broadcast to all procs */ #define TARGET_CPUS (target_cpus()) +#define wakeup_secondary_cpu wakeup_secondary_cpu_via_init /* * Set up the logical destination ID. * diff --git a/arch/x86/include/asm/mach-generic/mach_apic.h b/arch/x86/include/asm/mach-generic/mach_apic.h index 5180bd7478fb..e430f47df667 100644 --- a/arch/x86/include/asm/mach-generic/mach_apic.h +++ b/arch/x86/include/asm/mach-generic/mach_apic.h @@ -27,6 +27,7 @@ #define vector_allocation_domain (genapic->vector_allocation_domain) #define enable_apic_mode (genapic->enable_apic_mode) #define phys_pkg_id (genapic->phys_pkg_id) +#define wakeup_secondary_cpu (genapic->wakeup_cpu) extern void generic_bigsmp_probe(void); diff --git a/arch/x86/include/asm/setup.h b/arch/x86/include/asm/setup.h index 40b2d3304911..294daeb3a006 100644 --- a/arch/x86/include/asm/setup.h +++ b/arch/x86/include/asm/setup.h @@ -17,6 +17,7 @@ static inline int is_visws_box(void) { return 0; } #endif extern int wakeup_secondary_cpu_via_nmi(int apicid, unsigned long start_eip); +extern int wakeup_secondary_cpu_via_init(int apicid, unsigned long start_eip); /* * Any setup quirks to be performed? */ @@ -40,7 +41,7 @@ struct x86_quirks { void (*smp_read_mpc_oem)(struct mp_config_oemtable *oemtable, unsigned short oemsize); int (*setup_ioapic_ids)(void); - int (*wakeup_secondary_cpu)(int apicid, unsigned long start_eip); + int (*update_genapic)(void); }; extern struct x86_quirks *x86_quirks; diff --git a/arch/x86/kernel/es7000_32.c b/arch/x86/kernel/es7000_32.c index bed10dddf099..fb3bfe66fbe2 100644 --- a/arch/x86/kernel/es7000_32.c +++ b/arch/x86/kernel/es7000_32.c @@ -40,6 +40,7 @@ #include #include #include +#include #include /* @@ -180,6 +181,13 @@ static int wakeup_secondary_cpu_via_mip(int cpu, unsigned long eip) return 0; } + +static int __init es7000_update_genapic(void) +{ + genapic->wakeup_cpu = wakeup_secondary_cpu_via_mip; + + return 0; +} #endif void __init @@ -197,8 +205,9 @@ setup_unisys(void) else es7000_plat = ES7000_CLASSIC; ioapic_renumber_irq = es7000_rename_gsi; + #ifdef CONFIG_ES7000_CLUSTERED_APIC - x86_quirks->wakeup_secondary_cpu = wakeup_secondary_cpu_via_mip; + x86_quirks->update_genapic = es7000_update_genapic; #endif } diff --git a/arch/x86/kernel/genapic_64.c b/arch/x86/kernel/genapic_64.c index 6c9bfc9e1e95..2bced78b0b8e 100644 --- a/arch/x86/kernel/genapic_64.c +++ b/arch/x86/kernel/genapic_64.c @@ -21,6 +21,7 @@ #include #include #include +#include extern struct genapic apic_flat; extern struct genapic apic_physflat; @@ -53,6 +54,9 @@ void __init setup_apic_routing(void) genapic = &apic_physflat; printk(KERN_INFO "Setting APIC routing to %s\n", genapic->name); } + + if (x86_quirks->update_genapic) + x86_quirks->update_genapic(); } /* Same for both flat and physical. */ diff --git a/arch/x86/kernel/numaq_32.c b/arch/x86/kernel/numaq_32.c index 745891b7d0fb..0deea37a53cf 100644 --- a/arch/x86/kernel/numaq_32.c +++ b/arch/x86/kernel/numaq_32.c @@ -31,7 +31,7 @@ #include #include #include -#include +#include #include #include @@ -235,6 +235,13 @@ static int __init numaq_setup_ioapic_ids(void) return 1; } +static int __init numaq_update_genapic(void) +{ + genapic->wakeup_cpu = wakeup_secondary_cpu_via_nmi; + + return 0; +} + static struct x86_quirks numaq_x86_quirks __initdata = { .arch_pre_time_init = numaq_pre_time_init, .arch_time_init = NULL, @@ -250,7 +257,7 @@ static struct x86_quirks numaq_x86_quirks __initdata = { .mpc_oem_pci_bus = mpc_oem_pci_bus, .smp_read_mpc_oem = smp_read_mpc_oem, .setup_ioapic_ids = numaq_setup_ioapic_ids, - .wakeup_secondary_cpu = wakeup_secondary_cpu_via_nmi, + .update_genapic = numaq_update_genapic, }; void numaq_mps_oem_check(struct mp_config_table *mpc, char *oem, diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index 0fa6790c1dd3..c366e891e10b 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -583,7 +583,18 @@ static int __init setup_elfcorehdr(char *arg) early_param("elfcorehdr", setup_elfcorehdr); #endif -static struct x86_quirks default_x86_quirks __initdata; +static int __init default_update_genapic(void) +{ +#if defined(CONFIG_X86_GENERICARCH) || defined(CONFIG_X86_64) + genapic->wakeup_cpu = wakeup_secondary_cpu_via_nmi; +#endif + + return 0; +} + +static struct x86_quirks default_x86_quirks __initdata = { + .update_genapic = default_update_genapic, +}; struct x86_quirks *x86_quirks __initdata = &default_x86_quirks; diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index 498c1ef37fe0..0e9f446269f4 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -615,7 +615,7 @@ wakeup_secondary_cpu_via_nmi(int logical_apicid, unsigned long start_eip) return (send_status | accept_status); } -static int __devinit +int __devinit wakeup_secondary_cpu_via_init(int phys_apicid, unsigned long start_eip) { unsigned long send_status, accept_status = 0; @@ -736,15 +736,6 @@ wakeup_secondary_cpu_via_init(int phys_apicid, unsigned long start_eip) return (send_status | accept_status); } -static int __devinit -wakeup_secondary_cpu(int apicid, unsigned long start_eip) -{ - if (x86_quirks->wakeup_secondary_cpu) - return x86_quirks->wakeup_secondary_cpu(apicid, start_eip); - - return wakeup_secondary_cpu_via_init(apicid, start_eip); -} - struct create_idle { struct work_struct work; struct task_struct *idle; diff --git a/arch/x86/mach-generic/probe.c b/arch/x86/mach-generic/probe.c index 5a7e4619e1c4..90b134f3cd74 100644 --- a/arch/x86/mach-generic/probe.c +++ b/arch/x86/mach-generic/probe.c @@ -15,6 +15,7 @@ #include #include #include +#include extern struct genapic apic_numaq; extern struct genapic apic_summit; @@ -57,6 +58,9 @@ static int __init parse_apic(char *arg) } } + if (x86_quirks->update_genapic) + x86_quirks->update_genapic(); + /* Parsed again by __setup for debug/verbose */ return 0; } -- cgit v1.2.3 From 0231022cc32d5f2e7f3c06b75691dda0ad6aec33 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Mon, 17 Nov 2008 03:22:41 +0100 Subject: tracing/function-return-tracer: add the overrun field Impact: help to find the better depth of trace We decided to arbitrary define the depth of function return trace as "20". Perhaps this is not enough. To help finding an optimal depth, we measure now the overrun: the number of functions that have been missed for the current thread. By default this is not displayed, we have to do set a particular flag on the return tracer: echo overrun > /debug/tracing/trace_options And the overrun will be printed on the right. As the trace shows below, the current 20 depth is not enough. update_wall_time+0x37f/0x8c0 -> update_xtime_cache (345 ns) (Overruns: 2838) update_wall_time+0x384/0x8c0 -> clocksource_get_next (1141 ns) (Overruns: 2838) do_timer+0x23/0x100 -> update_wall_time (3882 ns) (Overruns: 2838) tick_do_update_jiffies64+0xbf/0x160 -> do_timer (5339 ns) (Overruns: 2838) tick_sched_timer+0x6a/0xf0 -> tick_do_update_jiffies64 (7209 ns) (Overruns: 2838) vgacon_set_cursor_size+0x98/0x120 -> native_io_delay (2613 ns) (Overruns: 274) vgacon_cursor+0x16e/0x1d0 -> vgacon_set_cursor_size (33151 ns) (Overruns: 274) set_cursor+0x5f/0x80 -> vgacon_cursor (36432 ns) (Overruns: 274) con_flush_chars+0x34/0x40 -> set_cursor (38790 ns) (Overruns: 274) release_console_sem+0x1ec/0x230 -> up (721 ns) (Overruns: 274) release_console_sem+0x225/0x230 -> wake_up_klogd (316 ns) (Overruns: 274) con_flush_chars+0x39/0x40 -> release_console_sem (2996 ns) (Overruns: 274) con_write+0x22/0x30 -> con_flush_chars (46067 ns) (Overruns: 274) n_tty_write+0x1cc/0x360 -> con_write (292670 ns) (Overruns: 274) smp_apic_timer_interrupt+0x2a/0x90 -> native_apic_mem_write (330 ns) (Overruns: 274) irq_enter+0x17/0x70 -> idle_cpu (413 ns) (Overruns: 274) smp_apic_timer_interrupt+0x2f/0x90 -> irq_enter (1525 ns) (Overruns: 274) ktime_get_ts+0x40/0x70 -> getnstimeofday (465 ns) (Overruns: 274) ktime_get_ts+0x60/0x70 -> set_normalized_timespec (436 ns) (Overruns: 274) ktime_get+0x16/0x30 -> ktime_get_ts (2501 ns) (Overruns: 274) hrtimer_interrupt+0x77/0x1a0 -> ktime_get (3439 ns) (Overruns: 274) Signed-off-by: Frederic Weisbecker Acked-by: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/x86/include/asm/thread_info.h | 7 +++++++ arch/x86/kernel/ftrace.c | 10 ++++++--- include/linux/ftrace.h | 2 ++ include/linux/sched.h | 1 + kernel/trace/trace.c | 1 + kernel/trace/trace.h | 1 + kernel/trace/trace_functions_return.c | 38 +++++++++++++++++++++++++++++------ 7 files changed, 51 insertions(+), 9 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/thread_info.h b/arch/x86/include/asm/thread_info.h index a71158369fd4..e90e81ef6ab9 100644 --- a/arch/x86/include/asm/thread_info.h +++ b/arch/x86/include/asm/thread_info.h @@ -21,6 +21,7 @@ struct task_struct; struct exec_domain; #include #include +#include struct thread_info { struct task_struct *task; /* main task structure */ @@ -45,6 +46,11 @@ struct thread_info { int curr_ret_stack; /* Stack of return addresses for return function tracing */ struct ftrace_ret_stack ret_stack[FTRACE_RET_STACK_SIZE]; + /* + * Number of functions that haven't been traced + * because of depth overrun. + */ + atomic_t trace_overrun; #endif }; @@ -61,6 +67,7 @@ struct thread_info { .fn = do_no_restart_syscall, \ }, \ .curr_ret_stack = -1,\ + .trace_overrun = ATOMIC_INIT(0) \ } #else #define INIT_THREAD_INFO(tsk) \ diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 924153edd973..356bb1eb6e9a 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -353,8 +353,10 @@ static int push_return_trace(unsigned long ret, unsigned long long time, struct thread_info *ti = current_thread_info(); /* The return trace stack is full */ - if (ti->curr_ret_stack == FTRACE_RET_STACK_SIZE - 1) + if (ti->curr_ret_stack == FTRACE_RET_STACK_SIZE - 1) { + atomic_inc(&ti->trace_overrun); return -EBUSY; + } index = ++ti->curr_ret_stack; barrier(); @@ -367,7 +369,7 @@ static int push_return_trace(unsigned long ret, unsigned long long time, /* Retrieve a function return address to the trace stack on thread info.*/ static void pop_return_trace(unsigned long *ret, unsigned long long *time, - unsigned long *func) + unsigned long *func, unsigned long *overrun) { int index; @@ -376,6 +378,7 @@ static void pop_return_trace(unsigned long *ret, unsigned long long *time, *ret = ti->ret_stack[index].ret; *func = ti->ret_stack[index].func; *time = ti->ret_stack[index].calltime; + *overrun = atomic_read(&ti->trace_overrun); ti->curr_ret_stack--; } @@ -386,7 +389,8 @@ static void pop_return_trace(unsigned long *ret, unsigned long long *time, unsigned long ftrace_return_to_handler(void) { struct ftrace_retfunc trace; - pop_return_trace(&trace.ret, &trace.calltime, &trace.func); + pop_return_trace(&trace.ret, &trace.calltime, &trace.func, + &trace.overrun); trace.rettime = cpu_clock(raw_smp_processor_id()); ftrace_function_return(&trace); diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index f1af1aab00e6..f7ba4ea5e128 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -318,6 +318,8 @@ struct ftrace_retfunc { unsigned long func; /* Current function */ unsigned long long calltime; unsigned long long rettime; + /* Number of functions that overran the depth limit for current task */ + unsigned long overrun; }; #ifdef CONFIG_FUNCTION_RET_TRACER diff --git a/include/linux/sched.h b/include/linux/sched.h index 61c8cc36028a..c8e0db464206 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -2016,6 +2016,7 @@ static inline void setup_thread_stack(struct task_struct *p, struct task_struct * used. */ task_thread_info(p)->curr_ret_stack = -1; + atomic_set(&task_thread_info(p)->trace_overrun, 0); #endif } diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 9531fddcfb8d..e97c29a6e7b0 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -853,6 +853,7 @@ static void __trace_function_return(struct trace_array *tr, entry->parent_ip = trace->ret; entry->rettime = trace->rettime; entry->calltime = trace->calltime; + entry->overrun = trace->overrun; ring_buffer_unlock_commit(global_trace.buffer, event, irq_flags); } #endif diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 9d22618bf99f..2cb12fd98f6b 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -60,6 +60,7 @@ struct ftrace_ret_entry { unsigned long parent_ip; unsigned long long calltime; unsigned long long rettime; + unsigned long overrun; }; extern struct tracer boot_tracer; diff --git a/kernel/trace/trace_functions_return.c b/kernel/trace/trace_functions_return.c index a68564af022b..e00d64509c9c 100644 --- a/kernel/trace/trace_functions_return.c +++ b/kernel/trace/trace_functions_return.c @@ -14,6 +14,19 @@ #include "trace.h" +#define TRACE_RETURN_PRINT_OVERRUN 0x1 +static struct tracer_opt trace_opts[] = { + /* Display overruns or not */ + { TRACER_OPT(overrun, TRACE_RETURN_PRINT_OVERRUN) }, + { } /* Empty entry */ +}; + +static struct tracer_flags tracer_flags = { + .val = 0, /* Don't display overruns by default */ + .opts = trace_opts +}; + + static int return_trace_init(struct trace_array *tr) { int cpu; @@ -42,26 +55,39 @@ print_return_function(struct trace_iterator *iter) ret = trace_seq_printf(s, "%pF -> ", (void *)field->parent_ip); if (!ret) return TRACE_TYPE_PARTIAL_LINE; + ret = seq_print_ip_sym(s, field->ip, trace_flags & TRACE_ITER_SYM_MASK); if (!ret) return TRACE_TYPE_PARTIAL_LINE; - ret = trace_seq_printf(s, " (%llu ns)\n", + + ret = trace_seq_printf(s, " (%llu ns)", field->rettime - field->calltime); if (!ret) return TRACE_TYPE_PARTIAL_LINE; - else - return TRACE_TYPE_HANDLED; + + if (tracer_flags.val & TRACE_RETURN_PRINT_OVERRUN) { + ret = trace_seq_printf(s, " (Overruns: %lu)", + field->overrun); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + } + + ret = trace_seq_printf(s, "\n"); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + return TRACE_TYPE_HANDLED; } return TRACE_TYPE_UNHANDLED; } -static struct tracer return_trace __read_mostly = -{ +static struct tracer return_trace __read_mostly = { .name = "return", .init = return_trace_init, .reset = return_trace_reset, - .print_line = print_return_function + .print_line = print_return_function, + .flags = &tracer_flags, }; static __init int init_return_trace(void) -- cgit v1.2.3 From a1afd01c175324656d0e8f1c82ea94b474953c04 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Tue, 18 Nov 2008 12:44:21 +0100 Subject: x86: default to SWIOTLB=y on x86_64 Impact: fixes korg bugzilla 11980 A kernel for a 64bit x86 system should always contain the swiotlb code in case it is booted on a machine without any hardware IOMMU supported by the kernel and more than 4GB of RAM. This patch changes Kconfig to always compile swiotlb into the kernel for x86_64. Signed-off-by: Joerg Roedel Cc: stable@kernel.org Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 93224b569187..669c6d588bde 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -566,7 +566,7 @@ config AMD_IOMMU # need this always selected by IOMMU for the VIA workaround config SWIOTLB - bool + def_bool y if X86_64 help Support for software bounce buffers used on x86-64 systems which don't have a hardware IOMMU (e.g. the current generation -- cgit v1.2.3 From b78a5b5260abf90d574911e7c7b8d35d5b48d6c0 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Mon, 17 Nov 2008 15:44:50 -0800 Subject: x86: ia32_signal: cleanup macro COPY Impact: cleanup No need to use temporary variable in this case. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/ia32/ia32_signal.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/ia32/ia32_signal.c b/arch/x86/ia32/ia32_signal.c index e2d0bc779bf7..610a17774ea2 100644 --- a/arch/x86/ia32/ia32_signal.c +++ b/arch/x86/ia32/ia32_signal.c @@ -197,10 +197,8 @@ struct rt_sigframe /* fp state follows here */ }; -#define COPY(x) { \ - unsigned int reg; \ - err |= __get_user(reg, &sc->x); \ - regs->x = reg; \ +#define COPY(x) { \ + err |= __get_user(regs->x, &sc->x); \ } #define RELOAD_SEG(seg,mask) \ -- cgit v1.2.3 From d71a68dca54756049e0eae62458a1705bf680d09 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Mon, 17 Nov 2008 15:47:06 -0800 Subject: x86: ia32_signal: introduce COPY_SEG_CPL3 Impact: cleanup Introduce COPY_SEG_CPL3 for ia32_restore_sigcontext(). Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/ia32/ia32_signal.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/ia32/ia32_signal.c b/arch/x86/ia32/ia32_signal.c index 610a17774ea2..fe44c314c9c0 100644 --- a/arch/x86/ia32/ia32_signal.c +++ b/arch/x86/ia32/ia32_signal.c @@ -201,6 +201,12 @@ struct rt_sigframe err |= __get_user(regs->x, &sc->x); \ } +#define COPY_SEG_CPL3(seg) { \ + unsigned short tmp; \ + err |= __get_user(tmp, &sc->seg); \ + regs->seg = tmp | 3; \ +} + #define RELOAD_SEG(seg,mask) \ { unsigned int cur; \ unsigned short pre; \ @@ -246,10 +252,8 @@ static int ia32_restore_sigcontext(struct pt_regs *regs, COPY(dx); COPY(cx); COPY(ip); /* Don't touch extended registers */ - err |= __get_user(regs->cs, &sc->cs); - regs->cs |= 3; - err |= __get_user(regs->ss, &sc->ss); - regs->ss |= 3; + COPY_SEG_CPL3(cs); + COPY_SEG_CPL3(ss); err |= __get_user(tmpflags, &sc->flags); regs->flags = (regs->flags & ~FIX_EFLAGS) | (tmpflags & FIX_EFLAGS); -- cgit v1.2.3 From 8c6e5ce0fd67c57ad5e19d1718e1250214e855db Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Mon, 17 Nov 2008 15:47:48 -0800 Subject: x86: ia32_signal: cleanup macro RELOAD_SEG Impact: cleanup Remove mask parameter because it's always 3. Cleanup coding styles. Signed-off-by: Hiroshi Shimamoto Reviewed-by: WANG Cong Signed-off-by: Ingo Molnar --- arch/x86/ia32/ia32_signal.c | 21 +++++++++++---------- 1 file changed, 11 insertions(+), 10 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/ia32/ia32_signal.c b/arch/x86/ia32/ia32_signal.c index fe44c314c9c0..2c56e6857d1a 100644 --- a/arch/x86/ia32/ia32_signal.c +++ b/arch/x86/ia32/ia32_signal.c @@ -207,13 +207,14 @@ struct rt_sigframe regs->seg = tmp | 3; \ } -#define RELOAD_SEG(seg,mask) \ - { unsigned int cur; \ - unsigned short pre; \ - err |= __get_user(pre, &sc->seg); \ - savesegment(seg, cur); \ - pre |= mask; \ - if (pre != cur) loadsegment(seg, pre); } +#define RELOAD_SEG(seg) { \ + unsigned int cur, pre; \ + err |= __get_user(pre, &sc->seg); \ + savesegment(seg, cur); \ + pre |= 3; \ + if (pre != cur) \ + loadsegment(seg, pre); \ +} static int ia32_restore_sigcontext(struct pt_regs *regs, struct sigcontext_ia32 __user *sc, @@ -244,9 +245,9 @@ static int ia32_restore_sigcontext(struct pt_regs *regs, if (gs != oldgs) load_gs_index(gs); - RELOAD_SEG(fs, 3); - RELOAD_SEG(ds, 3); - RELOAD_SEG(es, 3); + RELOAD_SEG(fs); + RELOAD_SEG(ds); + RELOAD_SEG(es); COPY(di); COPY(si); COPY(bp); COPY(sp); COPY(bx); COPY(dx); COPY(cx); COPY(ip); -- cgit v1.2.3 From 047ce93581ca122442ed3c13a62a645249a7db1d Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Mon, 17 Nov 2008 15:48:27 -0800 Subject: x86: ia32_signal: remove using temporary variable Impact: cleanup No need to use temporary variable. Also rename the variable same as arch/x86/kernel/signal_32.c. Signed-off-by: Hiroshi Shimamoto Reviewed-by: WANG Cong Signed-off-by: Ingo Molnar --- arch/x86/ia32/ia32_signal.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/ia32/ia32_signal.c b/arch/x86/ia32/ia32_signal.c index 2c56e6857d1a..e591e381611b 100644 --- a/arch/x86/ia32/ia32_signal.c +++ b/arch/x86/ia32/ia32_signal.c @@ -218,7 +218,7 @@ struct rt_sigframe static int ia32_restore_sigcontext(struct pt_regs *regs, struct sigcontext_ia32 __user *sc, - unsigned int *peax) + unsigned int *pax) { unsigned int tmpflags, gs, oldgs, err = 0; void __user *buf; @@ -265,9 +265,7 @@ static int ia32_restore_sigcontext(struct pt_regs *regs, buf = compat_ptr(tmp); err |= restore_i387_xstate_ia32(buf); - err |= __get_user(tmp, &sc->ax); - *peax = tmp; - + err |= __get_user(*pax, &sc->ax); return err; } -- cgit v1.2.3 From 64977609e316c86fad513d9bf0afff998581e59d Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Mon, 17 Nov 2008 15:49:14 -0800 Subject: x86: ia32_signal: change order of storing in setup_sigcontext() Impact: cleanup Change order of storing to match the sigcontext_ia32. And add casting to make this code same as arch/x86/kernel/signal_32.c. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/ia32/ia32_signal.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/ia32/ia32_signal.c b/arch/x86/ia32/ia32_signal.c index e591e381611b..1267977e7708 100644 --- a/arch/x86/ia32/ia32_signal.c +++ b/arch/x86/ia32/ia32_signal.c @@ -360,13 +360,13 @@ static int ia32_setup_sigcontext(struct sigcontext_ia32 __user *sc, err |= __put_user(regs->dx, &sc->dx); err |= __put_user(regs->cx, &sc->cx); err |= __put_user(regs->ax, &sc->ax); - err |= __put_user(regs->cs, &sc->cs); - err |= __put_user(regs->ss, &sc->ss); err |= __put_user(current->thread.trap_no, &sc->trapno); err |= __put_user(current->thread.error_code, &sc->err); err |= __put_user(regs->ip, &sc->ip); + err |= __put_user(regs->cs, (unsigned int __user *)&sc->cs); err |= __put_user(regs->flags, &sc->flags); err |= __put_user(regs->sp, &sc->sp_at_signal); + err |= __put_user(regs->ss, (unsigned int __user *)&sc->ss); err |= __put_user(ptr_to_compat(fpstate), &sc->fpstate); -- cgit v1.2.3 From f632ddcc0786149c0e4bef9b6b44c96a75c0d074 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Tue, 18 Nov 2008 17:32:26 +0100 Subject: x86: fix wakeup_cpu with numaq/es7000, v2, fix #2 Impact: fix boot crash fix default_update_genapic(). Signed-off-by: Yinghai Lu Signed-off-by: Ingo Molnar --- arch/x86/kernel/setup.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index c366e891e10b..31328909456e 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -585,8 +585,10 @@ early_param("elfcorehdr", setup_elfcorehdr); static int __init default_update_genapic(void) { -#if defined(CONFIG_X86_GENERICARCH) || defined(CONFIG_X86_64) - genapic->wakeup_cpu = wakeup_secondary_cpu_via_nmi; +#ifdef CONFIG_X86_SMP +# if defined(CONFIG_X86_GENERICARCH) || defined(CONFIG_X86_64) + genapic->wakeup_cpu = wakeup_secondary_cpu_via_init; +# endif #endif return 0; -- cgit v1.2.3 From b5fe363b7d89577fcfda9b6cf0efc32760bbccc6 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Tue, 18 Nov 2008 08:14:14 -0800 Subject: x86: use update_genapic to get rid of ES7000_CLUSTERED_APIC v2 Impact: clean up We can autodetect those system that need cluster apic, and update genapic accordingly. We can also remove wakeup.h for e7000, because it's default one is now the same as overall default mach_wakecpu.h Signed-off-by: Yinghai Lu Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 4 -- arch/x86/include/asm/es7000/apic.h | 76 ++++++++++++++++++++++++++------------ arch/x86/include/asm/genapic_32.h | 1 + arch/x86/kernel/es7000_32.c | 17 +++++++-- arch/x86/mach-generic/es7000.c | 14 ++++++- 5 files changed, 80 insertions(+), 32 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 93224b569187..7d0ab8942cfb 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -462,10 +462,6 @@ config X86_CYCLONE_TIMER def_bool y depends on X86_GENERICARCH -config ES7000_CLUSTERED_APIC - def_bool y - depends on SMP && X86_ES7000 && MPENTIUMIII - source "arch/x86/Kconfig.cpu" config HPET_TIMER diff --git a/arch/x86/include/asm/es7000/apic.h b/arch/x86/include/asm/es7000/apic.h index 9d8cf776c285..e24ef876915f 100644 --- a/arch/x86/include/asm/es7000/apic.h +++ b/arch/x86/include/asm/es7000/apic.h @@ -9,28 +9,27 @@ static inline int apic_id_registered(void) return (1); } -static inline cpumask_t target_cpus(void) +static inline cpumask_t target_cpus_cluster(void) { -#if defined CONFIG_ES7000_CLUSTERED_APIC return CPU_MASK_ALL; -#else +} + +static inline cpumask_t target_cpus(void) +{ return cpumask_of_cpu(smp_processor_id()); -#endif } -#if defined CONFIG_ES7000_CLUSTERED_APIC -#define APIC_DFR_VALUE (APIC_DFR_CLUSTER) -#define INT_DELIVERY_MODE (dest_LowestPrio) -#define INT_DEST_MODE (1) /* logical delivery broadcast to all procs */ -#define NO_BALANCE_IRQ (1) -#else +#define APIC_DFR_VALUE_CLUSTER (APIC_DFR_CLUSTER) +#define INT_DELIVERY_MODE_CLUSTER (dest_LowestPrio) +#define INT_DEST_MODE_CLUSTER (1) /* logical delivery broadcast to all procs */ +#define NO_BALANCE_IRQ_CLUSTER (1) + #define APIC_DFR_VALUE (APIC_DFR_FLAT) #define INT_DELIVERY_MODE (dest_Fixed) #define INT_DEST_MODE (0) /* phys delivery to target procs */ #define NO_BALANCE_IRQ (0) #undef APIC_DEST_LOGICAL #define APIC_DEST_LOGICAL 0x0 -#endif static inline unsigned long check_apicid_used(physid_mask_t bitmap, int apicid) { @@ -57,6 +56,16 @@ static inline unsigned long calculate_ldr(int cpu) * an APIC. See e.g. "AP-388 82489DX User's Manual" (Intel * document number 292116). So here it goes... */ +static inline void init_apic_ldr_cluster(void) +{ + unsigned long val; + int cpu = smp_processor_id(); + + apic_write(APIC_DFR, APIC_DFR_VALUE_CLUSTER); + val = calculate_ldr(cpu); + apic_write(APIC_LDR, val); +} + static inline void init_apic_ldr(void) { unsigned long val; @@ -67,10 +76,6 @@ static inline void init_apic_ldr(void) apic_write(APIC_LDR, val); } -#ifndef CONFIG_X86_GENERICARCH -extern void enable_apic_mode(void); -#endif - extern int apic_version [MAX_APICS]; static inline void setup_apic_routing(void) { @@ -141,7 +146,7 @@ static inline int check_phys_apicid_present(int cpu_physical_apicid) return (1); } -static inline unsigned int cpu_mask_to_apicid(cpumask_t cpumask) +static inline unsigned int cpu_mask_to_apicid_cluster(cpumask_t cpumask) { int num_bits_set; int cpus_found = 0; @@ -151,11 +156,7 @@ static inline unsigned int cpu_mask_to_apicid(cpumask_t cpumask) num_bits_set = cpus_weight(cpumask); /* Return id to all */ if (num_bits_set == NR_CPUS) -#if defined CONFIG_ES7000_CLUSTERED_APIC return 0xFF; -#else - return cpu_to_logical_apicid(0); -#endif /* * The cpus in the mask must all be on the apic cluster. If are not * on the same apicid cluster return default value of TARGET_CPUS. @@ -168,11 +169,40 @@ static inline unsigned int cpu_mask_to_apicid(cpumask_t cpumask) if (apicid_cluster(apicid) != apicid_cluster(new_apicid)){ printk ("%s: Not a valid mask!\n", __func__); -#if defined CONFIG_ES7000_CLUSTERED_APIC return 0xFF; -#else + } + apicid = new_apicid; + cpus_found++; + } + cpu++; + } + return apicid; +} + +static inline unsigned int cpu_mask_to_apicid(cpumask_t cpumask) +{ + int num_bits_set; + int cpus_found = 0; + int cpu; + int apicid; + + num_bits_set = cpus_weight(cpumask); + /* Return id to all */ + if (num_bits_set == NR_CPUS) + return cpu_to_logical_apicid(0); + /* + * The cpus in the mask must all be on the apic cluster. If are not + * on the same apicid cluster return default value of TARGET_CPUS. + */ + cpu = first_cpu(cpumask); + apicid = cpu_to_logical_apicid(cpu); + while (cpus_found < num_bits_set) { + if (cpu_isset(cpu, cpumask)) { + int new_apicid = cpu_to_logical_apicid(cpu); + if (apicid_cluster(apicid) != + apicid_cluster(new_apicid)){ + printk ("%s: Not a valid mask!\n", __func__); return cpu_to_logical_apicid(0); -#endif } apicid = new_apicid; cpus_found++; diff --git a/arch/x86/include/asm/genapic_32.h b/arch/x86/include/asm/genapic_32.h index 455d6c27a98b..0ac17d33a8c7 100644 --- a/arch/x86/include/asm/genapic_32.h +++ b/arch/x86/include/asm/genapic_32.h @@ -131,6 +131,7 @@ struct genapic { } extern struct genapic *genapic; +extern void es7000_update_genapic_to_cluster(void); enum uv_system_type {UV_NONE, UV_LEGACY_APIC, UV_X2APIC, UV_NON_UNIQUE_APIC}; #define get_uv_system_type() UV_NONE diff --git a/arch/x86/kernel/es7000_32.c b/arch/x86/kernel/es7000_32.c index fb3bfe66fbe2..71d7be624d46 100644 --- a/arch/x86/kernel/es7000_32.c +++ b/arch/x86/kernel/es7000_32.c @@ -38,6 +38,7 @@ #include #include #include +#include #include #include #include @@ -163,7 +164,6 @@ es7000_rename_gsi(int ioapic, int gsi) return gsi; } -#ifdef CONFIG_ES7000_CLUSTERED_APIC static int wakeup_secondary_cpu_via_mip(int cpu, unsigned long eip) { unsigned long vect = 0, psaival = 0; @@ -182,13 +182,24 @@ static int wakeup_secondary_cpu_via_mip(int cpu, unsigned long eip) return 0; } +static void noop_wait_for_deassert(atomic_t *deassert_not_used) +{ +} + static int __init es7000_update_genapic(void) { genapic->wakeup_cpu = wakeup_secondary_cpu_via_mip; + /* MPENTIUMIII */ + if (boot_cpu_data.x86 == 6 && + (boot_cpu_data.x86_model >= 7 || boot_cpu_data.x86_model <= 11)) { + es7000_update_genapic_to_cluster(); + genapic->wait_for_init_deassert = noop_wait_for_deassert; + genapic->wakeup_cpu = wakeup_secondary_cpu_via_mip; + } + return 0; } -#endif void __init setup_unisys(void) @@ -206,9 +217,7 @@ setup_unisys(void) es7000_plat = ES7000_CLASSIC; ioapic_renumber_irq = es7000_rename_gsi; -#ifdef CONFIG_ES7000_CLUSTERED_APIC x86_quirks->update_genapic = es7000_update_genapic; -#endif } /* diff --git a/arch/x86/mach-generic/es7000.c b/arch/x86/mach-generic/es7000.c index 28459cab3ddb..7b4e6d0d1690 100644 --- a/arch/x86/mach-generic/es7000.c +++ b/arch/x86/mach-generic/es7000.c @@ -16,7 +16,19 @@ #include #include #include -#include +#include + +void __init es7000_update_genapic_to_cluster(void) +{ + genapic->target_cpus = target_cpus_cluster; + genapic->int_delivery_mode = INT_DELIVERY_MODE_CLUSTER; + genapic->int_dest_mode = INT_DEST_MODE_CLUSTER; + genapic->no_balance_irq = NO_BALANCE_IRQ_CLUSTER; + + genapic->init_apic_ldr = init_apic_ldr_cluster; + + genapic->cpu_mask_to_apicid = cpu_mask_to_apicid_cluster; +} static int probe_es7000(void) { -- cgit v1.2.3 From 77be80e437fec44f8b7a620314b7d7b605b8d93b Mon Sep 17 00:00:00 2001 From: "Richard A. Holden III" Date: Wed, 19 Nov 2008 16:05:14 -0700 Subject: x86: fix arch/x86/kernel/genx2apic_uv_x.c build warning when !CONFIG_HOTPLUG_CPU Impact: cleanup, reduce size of the kernel image a bit Fix: arch/x86/kernel/genx2apic_uv_x.c:403: warning: 'uv_heartbeat_disable' defined but not used the function is only used when CONFIG_HOTPLUG_CPU is defined. Signed-off-by: Richard A. Holden III Signed-off-by: Ingo Molnar --- arch/x86/kernel/genx2apic_uv_x.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/genx2apic_uv_x.c b/arch/x86/kernel/genx2apic_uv_x.c index f02bbe5d0178..221299f4509f 100644 --- a/arch/x86/kernel/genx2apic_uv_x.c +++ b/arch/x86/kernel/genx2apic_uv_x.c @@ -400,6 +400,7 @@ static void __cpuinit uv_heartbeat_enable(int cpu) uv_heartbeat_enable(0); } +#ifdef CONFIG_HOTPLUG_CPU static void __cpuinit uv_heartbeat_disable(int cpu) { if (uv_cpu_hub_info(cpu)->scir.enabled) { @@ -409,7 +410,6 @@ static void __cpuinit uv_heartbeat_disable(int cpu) uv_set_cpu_scir_bits(cpu, 0xff); } -#ifdef CONFIG_HOTPLUG_CPU /* * cpu hotplug notifier */ -- cgit v1.2.3 From bb5574608a8375026510b4f983ffbb06ece33fe2 Mon Sep 17 00:00:00 2001 From: "Richard A. Holden III" Date: Wed, 19 Nov 2008 16:05:15 -0700 Subject: x86: fix arch/x86/kernel/setup.c build warning when !CONFIG_X86_RESERVE_LOW_64K Impact: cleanup Fix: arch/x86/kernel/setup.c:592: warning: 'dmi_low_memory_corruption' defined but not used this is only used if CONFIG_X86_RESERVE_LOW_64K is defined. Signed-off-by: Richard A. Holden III Signed-off-by: Ingo Molnar --- arch/x86/kernel/setup.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index e6c51433247d..13a5f592ac28 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -587,6 +587,7 @@ static struct x86_quirks default_x86_quirks __initdata; struct x86_quirks *x86_quirks __initdata = &default_x86_quirks; +#ifdef CONFIG_X86_RESERVE_LOW_64K static int __init dmi_low_memory_corruption(const struct dmi_system_id *d) { printk(KERN_NOTICE @@ -598,6 +599,7 @@ static int __init dmi_low_memory_corruption(const struct dmi_system_id *d) return 0; } +#endif /* List of systems that have known low memory corruption BIOS problems */ static struct dmi_system_id __initdata bad_bios_dmi_table[] = { -- cgit v1.2.3 From 87f7606591aea6a8a38ea4c8911b5eeeee2740b8 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Wed, 19 Nov 2008 20:50:53 -0800 Subject: x86: fix wakeup_cpu with numaq/es7000 v2 - call ->update_genapic() Impact: fix boot crash on 32-bit Hiroshi Shimamoto reported a boot failure on 32-bit x86. The setting of x86_quirks.wakeup_cpu is missing (when not passing in an explicit apic= boot parameter). Reported-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/mach-generic/probe.c | 12 +++++++++++- 1 file changed, 11 insertions(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/mach-generic/probe.c b/arch/x86/mach-generic/probe.c index 90b134f3cd74..c346d9d0226f 100644 --- a/arch/x86/mach-generic/probe.c +++ b/arch/x86/mach-generic/probe.c @@ -76,12 +76,15 @@ void __init generic_bigsmp_probe(void) * - we find more than 8 CPUs in acpi LAPIC listing with xAPIC support */ - if (!cmdline_apic && genapic == &apic_default) + if (!cmdline_apic && genapic == &apic_default) { if (apic_bigsmp.probe()) { genapic = &apic_bigsmp; + if (x86_quirks->update_genapic) + x86_quirks->update_genapic(); printk(KERN_INFO "Overriding APIC driver with %s\n", genapic->name); } + } #endif } @@ -98,6 +101,9 @@ void __init generic_apic_probe(void) /* Not visible without early console */ if (!apic_probe[i]) panic("Didn't find an APIC driver"); + + if (x86_quirks->update_genapic) + x86_quirks->update_genapic(); } printk(KERN_INFO "Using APIC driver %s\n", genapic->name); } @@ -112,6 +118,8 @@ int __init mps_oem_check(struct mp_config_table *mpc, char *oem, if (apic_probe[i]->mps_oem_check(mpc, oem, productid)) { if (!cmdline_apic) { genapic = apic_probe[i]; + if (x86_quirks->update_genapic) + x86_quirks->update_genapic(); printk(KERN_INFO "Switched to APIC driver `%s'.\n", genapic->name); } @@ -128,6 +136,8 @@ int __init acpi_madt_oem_check(char *oem_id, char *oem_table_id) if (apic_probe[i]->acpi_madt_oem_check(oem_id, oem_table_id)) { if (!cmdline_apic) { genapic = apic_probe[i]; + if (x86_quirks->update_genapic) + x86_quirks->update_genapic(); printk(KERN_INFO "Switched to APIC driver `%s'.\n", genapic->name); } -- cgit v1.2.3 From d99015b1abbad743aa049b439c1e1dede6d0fa49 Mon Sep 17 00:00:00 2001 From: Alexander van Heukelum Date: Wed, 19 Nov 2008 01:18:11 +0100 Subject: x86: move entry_64.S register saving out of the macros Here is a combined patch that moves "save_args" out-of-line for the interrupt macro and moves "error_entry" mostly out-of-line for the zeroentry and errorentry macros. The save_args function becomes really straightforward and easy to understand, with the possible exception of the stack switch code, which now needs to copy the return address of to the calling function. Normal interrupts arrive with ((~vector)-0x80) on the stack, which gets adjusted in common_interrupt: : (5) addq $0xffffffffffffff80,(%rsp) /* -> ~(vector) */ (4) sub $0x50,%rsp /* space for registers */ (5) callq ffffffff80211290 (5) callq ffffffff80214290 : ... An apic interrupt stub now look like this: : (5) pushq $0xffffffffffffff05 /* ~(vector) */ (4) sub $0x50,%rsp /* space for registers */ (5) callq ffffffff80211290 (5) callq ffffffff80212b8f (5) jmpq ffffffff80211f93 Similarly the exception handler register saving function becomes simpler, without the need of any parameter shuffling. The stub for an exception without errorcode looks like this: : (6) callq *0x1cad12(%rip) # ffffffff803dd448 (2) pushq $0xffffffffffffffff /* no syscall */ (4) sub $0x78,%rsp /* space for registers */ (5) callq ffffffff8030e3b0 (3) mov %rsp,%rdi /* pt_regs pointer */ (2) xor %esi,%esi /* no error code */ (5) callq ffffffff80213446 (5) jmpq ffffffff8030e460 And one for an exception with errorcode like this: : (6) callq *0x1cab92(%rip) # ffffffff803dd448 (4) sub $0x78,%rsp /* space for registers */ (5) callq ffffffff8030e3b0 (3) mov %rsp,%rdi /* pt_regs pointer */ (5) mov 0x78(%rsp),%rsi /* load error code */ (9) movq $0xffffffffffffffff,0x78(%rsp) /* no syscall */ (5) callq ffffffff80213209 (5) jmpq ffffffff8030e460 Unfortunately, this last type is more than 32 bytes. But the total space savings due to this patch is about 2500 bytes on an smp-configuration, and I think the code is clearer than it was before. The tested kernels were non-paravirt ones (i.e., without the indirect call at the top of the exception handlers). Anyhow, I tested this patch on top of a recent -tip. The machine was an 2x4-core Xeon at 2333MHz. Measured where the delays between (almost-)adjacent rdtsc instructions. The graphs show how much time is spent outside of the program as a function of the measured delay. The area under the graph represents the total time spent outside the program. Eight instances of the rdtsctest were started, each pinned to a single cpu. The histogams are added. For each kernel two measurements were done: one in mostly idle condition, the other while running "bonnie++ -f", bound to cpu 0. Each measurement took 40 minutes runtime. See the attached graphs for the results. The graphs overlap almost everywhere, but there are small differences. Signed-off-by: Alexander van Heukelum Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_64.S | 300 +++++++++++++++++++++++++-------------------- 1 file changed, 166 insertions(+), 134 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index dbf06a0ef3d5..5a12432ccdf9 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -242,6 +242,78 @@ ENTRY(native_usergs_sysret64) CFI_REL_OFFSET rsp,RSP /*CFI_REL_OFFSET ss,SS*/ .endm + +/* + * initial frame state for interrupts and exceptions + */ + .macro _frame ref + CFI_STARTPROC simple + CFI_SIGNAL_FRAME + CFI_DEF_CFA rsp,SS+8-\ref + /*CFI_REL_OFFSET ss,SS-\ref*/ + CFI_REL_OFFSET rsp,RSP-\ref + /*CFI_REL_OFFSET rflags,EFLAGS-\ref*/ + /*CFI_REL_OFFSET cs,CS-\ref*/ + CFI_REL_OFFSET rip,RIP-\ref + .endm + +/* + * initial frame state for interrupts (and exceptions without error code) + */ +#define INTR_FRAME _frame RIP +/* + * initial frame state for exceptions with error code (and interrupts + * with vector already pushed) + */ +#define XCPT_FRAME _frame ORIG_RAX + +/* save partial stack frame */ +ENTRY(save_args) + XCPT_FRAME + cld + movq %rdi, 8*8+16(%rsp) + CFI_REL_OFFSET rdi, 8*8+16 + movq %rsi, 7*8+16(%rsp) + CFI_REL_OFFSET rsi, 7*8+16 + movq %rdx, 6*8+16(%rsp) + CFI_REL_OFFSET rdx, 6*8+16 + movq %rcx, 5*8+16(%rsp) + CFI_REL_OFFSET rcx, 5*8+16 + movq %rax, 4*8+16(%rsp) + CFI_REL_OFFSET rax, 4*8+16 + movq %r8, 3*8+16(%rsp) + CFI_REL_OFFSET r8, 3*8+16 + movq %r9, 2*8+16(%rsp) + CFI_REL_OFFSET r9, 2*8+16 + movq %r10, 1*8+16(%rsp) + CFI_REL_OFFSET r10, 1*8+16 + movq %r11, 0*8+16(%rsp) + CFI_REL_OFFSET r11, 0*8+16 + leaq -ARGOFFSET+16(%rsp),%rdi /* arg1 for handler */ + movq %rbp, 8(%rsp) /* push %rbp */ + leaq 8(%rsp), %rbp /* mov %rsp, %ebp */ + testl $3, CS(%rdi) + je 1f + SWAPGS + /* + * irqcount is used to check if a CPU is already on an interrupt stack + * or not. While this is essentially redundant with preempt_count it is + * a little cheaper to use a separate counter in the PDA (short of + * moving irq_enter into assembly, which would be too much work) + */ +1: incl %gs:pda_irqcount + jne 2f + pop %rax /* move return address... */ + mov %gs:pda_irqstackptr,%rsp + push %rax /* ... to the new stack */ + /* + * We entered an interrupt context - irqs are off: + */ +2: TRACE_IRQS_OFF + ret + CFI_ENDPROC +END(save_args) + /* * A newly forked process directly context switches into this. */ @@ -607,26 +679,6 @@ ENTRY(stub_rt_sigreturn) CFI_ENDPROC END(stub_rt_sigreturn) -/* - * initial frame state for interrupts and exceptions - */ - .macro _frame ref - CFI_STARTPROC simple - CFI_SIGNAL_FRAME - CFI_DEF_CFA rsp,SS+8-\ref - /*CFI_REL_OFFSET ss,SS-\ref*/ - CFI_REL_OFFSET rsp,RSP-\ref - /*CFI_REL_OFFSET rflags,EFLAGS-\ref*/ - /*CFI_REL_OFFSET cs,CS-\ref*/ - CFI_REL_OFFSET rip,RIP-\ref - .endm - -/* initial frame state for interrupts (and exceptions without error code) */ -#define INTR_FRAME _frame RIP -/* initial frame state for exceptions with error code (and interrupts with - vector already pushed) */ -#define XCPT_FRAME _frame ORIG_RAX - /* * Build the entry stubs and pointer table with some assembler magic. * We pack 7 stubs into a single 32-byte chunk, which will fit in a @@ -667,46 +719,19 @@ END(irq_entries_start) END(interrupt) .previous -/* +/* * Interrupt entry/exit. * * Interrupt entry points save only callee clobbered registers in fast path. - * - * Entry runs with interrupts off. - */ + * + * Entry runs with interrupts off. + */ /* 0(%rsp): ~(interrupt number) */ .macro interrupt func - cld - SAVE_ARGS - leaq -ARGOFFSET(%rsp),%rdi /* arg1 for handler */ - pushq %rbp - /* - * Save rbp twice: One is for marking the stack frame, as usual, and the - * other, to fill pt_regs properly. This is because bx comes right - * before the last saved register in that structure, and not bp. If the - * base pointer were in the place bx is today, this would not be needed. - */ - movq %rbp, -8(%rsp) - CFI_ADJUST_CFA_OFFSET 8 - CFI_REL_OFFSET rbp, 0 - movq %rsp,%rbp - CFI_DEF_CFA_REGISTER rbp - testl $3,CS(%rdi) - je 1f - SWAPGS - /* irqcount is used to check if a CPU is already on an interrupt - stack or not. While this is essentially redundant with preempt_count - it is a little cheaper to use a separate counter in the PDA - (short of moving irq_enter into assembly, which would be too - much work) */ -1: incl %gs:pda_irqcount - cmoveq %gs:pda_irqstackptr,%rsp - push %rbp # backlink for old unwinder - /* - * We entered an interrupt context - irqs are off: - */ - TRACE_IRQS_OFF + subq $10*8, %rsp + CFI_ADJUST_CFA_OFFSET 10*8 + call save_args call \func .endm @@ -852,6 +877,8 @@ END(common_interrupt) /* * APIC interrupts. */ + .p2align 5 + .macro apicinterrupt num,func INTR_FRAME pushq $~(\num) @@ -922,24 +949,29 @@ END(spurious_interrupt) .macro zeroentry sym INTR_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME - pushq $0 /* push error code/oldrax */ + pushq $-1 /* ORIG_RAX: no syscall to restart */ CFI_ADJUST_CFA_OFFSET 8 - pushq %rax /* push real oldrax to the rdi slot */ - CFI_ADJUST_CFA_OFFSET 8 - CFI_REL_OFFSET rax,0 - leaq \sym(%rip),%rax - jmp error_entry + subq $15*8,%rsp + CFI_ADJUST_CFA_OFFSET 15*8 + call error_entry + movq %rsp,%rdi /* pt_regs pointer */ + xorl %esi,%esi /* no error code */ + call \sym + jmp error_exit /* %ebx: no swapgs flag */ CFI_ENDPROC .endm .macro errorentry sym XCPT_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME - pushq %rax - CFI_ADJUST_CFA_OFFSET 8 - CFI_REL_OFFSET rax,0 - leaq \sym(%rip),%rax - jmp error_entry + subq $15*8,%rsp + CFI_ADJUST_CFA_OFFSET 15*8 + call error_entry + movq %rsp,%rdi /* pt_regs pointer */ + movq ORIG_RAX(%rsp),%rsi /* get error code */ + movq $-1,ORIG_RAX(%rsp) /* no syscall to restart */ + call \sym + jmp error_exit /* %ebx: no swapgs flag */ CFI_ENDPROC .endm @@ -1043,93 +1075,93 @@ paranoid_schedule\trace: .endm /* - * Exception entry point. This expects an error code/orig_rax on the stack - * and the exception handler in %rax. + * Exception entry point. This expects an error code/orig_rax on the stack. + * returns in "no swapgs flag" in %ebx. */ KPROBE_ENTRY(error_entry) _frame RDI - CFI_REL_OFFSET rax,0 - /* rdi slot contains rax, oldrax contains error code */ + CFI_ADJUST_CFA_OFFSET 15*8 + /* oldrax contains error code */ cld - subq $14*8,%rsp - CFI_ADJUST_CFA_OFFSET (14*8) - movq %rsi,13*8(%rsp) - CFI_REL_OFFSET rsi,RSI - movq 14*8(%rsp),%rsi /* load rax from rdi slot */ - CFI_REGISTER rax,rsi - movq %rdx,12*8(%rsp) - CFI_REL_OFFSET rdx,RDX - movq %rcx,11*8(%rsp) - CFI_REL_OFFSET rcx,RCX - movq %rsi,10*8(%rsp) /* store rax */ - CFI_REL_OFFSET rax,RAX - movq %r8, 9*8(%rsp) - CFI_REL_OFFSET r8,R8 - movq %r9, 8*8(%rsp) - CFI_REL_OFFSET r9,R9 - movq %r10,7*8(%rsp) - CFI_REL_OFFSET r10,R10 - movq %r11,6*8(%rsp) - CFI_REL_OFFSET r11,R11 - movq %rbx,5*8(%rsp) - CFI_REL_OFFSET rbx,RBX - movq %rbp,4*8(%rsp) - CFI_REL_OFFSET rbp,RBP - movq %r12,3*8(%rsp) - CFI_REL_OFFSET r12,R12 - movq %r13,2*8(%rsp) - CFI_REL_OFFSET r13,R13 - movq %r14,1*8(%rsp) - CFI_REL_OFFSET r14,R14 - movq %r15,(%rsp) - CFI_REL_OFFSET r15,R15 + movq %rdi,14*8+8(%rsp) + CFI_REL_OFFSET rdi,RDI+8 + movq %rsi,13*8+8(%rsp) + CFI_REL_OFFSET rsi,RSI+8 + movq %rdx,12*8+8(%rsp) + CFI_REL_OFFSET rdx,RDX+8 + movq %rcx,11*8+8(%rsp) + CFI_REL_OFFSET rcx,RCX+8 + movq %rax,10*8+8(%rsp) + CFI_REL_OFFSET rax,RAX+8 + movq %r8, 9*8+8(%rsp) + CFI_REL_OFFSET r8,R8+8 + movq %r9, 8*8+8(%rsp) + CFI_REL_OFFSET r9,R9+8 + movq %r10,7*8+8(%rsp) + CFI_REL_OFFSET r10,R10+8 + movq %r11,6*8+8(%rsp) + CFI_REL_OFFSET r11,R11+8 + movq %rbx,5*8+8(%rsp) + CFI_REL_OFFSET rbx,RBX+8 + movq %rbp,4*8+8(%rsp) + CFI_REL_OFFSET rbp,RBP+8 + movq %r12,3*8+8(%rsp) + CFI_REL_OFFSET r12,R12+8 + movq %r13,2*8+8(%rsp) + CFI_REL_OFFSET r13,R13+8 + movq %r14,1*8+8(%rsp) + CFI_REL_OFFSET r14,R14+8 + movq %r15,0*8+8(%rsp) + CFI_REL_OFFSET r15,R15+8 xorl %ebx,%ebx - testl $3,CS(%rsp) - je error_kernelspace + testl $3,CS+8(%rsp) + je error_kernelspace error_swapgs: SWAPGS error_sti: TRACE_IRQS_OFF - movq %rdi,RDI(%rsp) - CFI_REL_OFFSET rdi,RDI - movq %rsp,%rdi - movq ORIG_RAX(%rsp),%rsi /* get error code */ - movq $-1,ORIG_RAX(%rsp) - call *%rax - /* ebx: no swapgs flag (1: don't need swapgs, 0: need it) */ -error_exit: + ret + CFI_ENDPROC + +/* + * There are two places in the kernel that can potentially fault with + * usergs. Handle them here. The exception handlers after iret run with + * kernel gs again, so don't set the user space flag. B stepping K8s + * sometimes report an truncated RIP for IRET exceptions returning to + * compat mode. Check for these here too. + */ +error_kernelspace: + incl %ebx + leaq irq_return(%rip),%rcx + cmpq %rcx,RIP+8(%rsp) + je error_swapgs + movl %ecx,%ecx /* zero extend */ + cmpq %rcx,RIP+8(%rsp) + je error_swapgs + cmpq $gs_change,RIP+8(%rsp) + je error_swapgs + jmp error_sti +KPROBE_END(error_entry) + + +/* ebx: no swapgs flag (1: don't need swapgs, 0: need it) */ +KPROBE_ENTRY(error_exit) + _frame R15 movl %ebx,%eax RESTORE_REST DISABLE_INTERRUPTS(CLBR_NONE) TRACE_IRQS_OFF GET_THREAD_INFO(%rcx) testl %eax,%eax - jne retint_kernel + jne retint_kernel LOCKDEP_SYS_EXIT_IRQ - movl TI_flags(%rcx),%edx - movl $_TIF_WORK_MASK,%edi - andl %edi,%edx - jnz retint_careful + movl TI_flags(%rcx),%edx + movl $_TIF_WORK_MASK,%edi + andl %edi,%edx + jnz retint_careful jmp retint_swapgs CFI_ENDPROC - -error_kernelspace: - incl %ebx - /* There are two places in the kernel that can potentially fault with - usergs. Handle them here. The exception handlers after - iret run with kernel gs again, so don't set the user space flag. - B stepping K8s sometimes report an truncated RIP for IRET - exceptions returning to compat mode. Check for these here too. */ - leaq irq_return(%rip),%rcx - cmpq %rcx,RIP(%rsp) - je error_swapgs - movl %ecx,%ecx /* zero extend */ - cmpq %rcx,RIP(%rsp) - je error_swapgs - cmpq $gs_change,RIP(%rsp) - je error_swapgs - jmp error_sti -KPROBE_END(error_entry) +KPROBE_END(error_exit) /* Reload gs selector with exception handling */ /* edi: new selector */ -- cgit v1.2.3 From dcd072e26055de600cecdc3f7a1e083ecd55c2e4 Mon Sep 17 00:00:00 2001 From: Alexander van Heukelum Date: Thu, 20 Nov 2008 14:40:11 +0100 Subject: x86: clean up after: move entry_64.S register saving out of the macros This add-on patch to x86: move entry_64.S register saving out of the macros visually cleans up the appearance of the code by introducing some basic helper macro's. It also adds some cfi annotations which were missing. Signed-off-by: Alexander van Heukelum Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_64.S | 220 +++++++++++++++++++++++---------------------- 1 file changed, 112 insertions(+), 108 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 5a12432ccdf9..7a04f696121d 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -60,6 +60,23 @@ #define __AUDIT_ARCH_LE 0x40000000 .code64 +/* + * Some macro's to hide the most frequently occuring CFI annotations. + */ + .macro CFI_PUSHQ reg + pushq \reg + CFI_ADJUST_CFA_OFFSET 8 + .endm + + .macro CFI_POPQ reg + popq \reg + CFI_ADJUST_CFA_OFFSET -8 + .endm + + .macro CFI_MOVQ reg offset=0 + movq %\reg, \offset(%rsp) + CFI_REL_OFFSET \reg, \offset + .endm #ifdef CONFIG_FUNCTION_TRACER #ifdef CONFIG_DYNAMIC_FTRACE @@ -213,84 +230,84 @@ ENTRY(native_usergs_sysret64) CFI_ADJUST_CFA_OFFSET -(6*8) .endm - .macro CFI_DEFAULT_STACK start=1 +/* + * initial frame state for interrupts (and exceptions without error code) + */ + .macro EMPTY_FRAME start=1 offset=0 .if \start - CFI_STARTPROC simple + CFI_STARTPROC simple CFI_SIGNAL_FRAME - CFI_DEF_CFA rsp,SS+8 + CFI_DEF_CFA rsp,8+\offset .else - CFI_DEF_CFA_OFFSET SS+8 + CFI_DEF_CFA_OFFSET 8+\offset .endif - CFI_REL_OFFSET r15,R15 - CFI_REL_OFFSET r14,R14 - CFI_REL_OFFSET r13,R13 - CFI_REL_OFFSET r12,R12 - CFI_REL_OFFSET rbp,RBP - CFI_REL_OFFSET rbx,RBX - CFI_REL_OFFSET r11,R11 - CFI_REL_OFFSET r10,R10 - CFI_REL_OFFSET r9,R9 - CFI_REL_OFFSET r8,R8 - CFI_REL_OFFSET rax,RAX - CFI_REL_OFFSET rcx,RCX - CFI_REL_OFFSET rdx,RDX - CFI_REL_OFFSET rsi,RSI - CFI_REL_OFFSET rdi,RDI - CFI_REL_OFFSET rip,RIP - /*CFI_REL_OFFSET cs,CS*/ - /*CFI_REL_OFFSET rflags,EFLAGS*/ - CFI_REL_OFFSET rsp,RSP - /*CFI_REL_OFFSET ss,SS*/ .endm /* - * initial frame state for interrupts and exceptions + * initial frame state for interrupts (and exceptions without error code) */ - .macro _frame ref - CFI_STARTPROC simple - CFI_SIGNAL_FRAME - CFI_DEF_CFA rsp,SS+8-\ref - /*CFI_REL_OFFSET ss,SS-\ref*/ - CFI_REL_OFFSET rsp,RSP-\ref - /*CFI_REL_OFFSET rflags,EFLAGS-\ref*/ - /*CFI_REL_OFFSET cs,CS-\ref*/ - CFI_REL_OFFSET rip,RIP-\ref + .macro INTR_FRAME start=1 offset=0 + EMPTY_FRAME \start, (SS+8-RIP)+\offset + /*CFI_REL_OFFSET ss, SS-RIP+\offset*/ + CFI_REL_OFFSET rsp, RSP-RIP+\offset + /*CFI_REL_OFFSET rflags, EFLAGS-RIP+\offset*/ + /*CFI_REL_OFFSET cs, CS-RIP+\offset*/ + CFI_REL_OFFSET rip, RIP-RIP+\offset .endm -/* - * initial frame state for interrupts (and exceptions without error code) - */ -#define INTR_FRAME _frame RIP /* * initial frame state for exceptions with error code (and interrupts * with vector already pushed) */ -#define XCPT_FRAME _frame ORIG_RAX + .macro XCPT_FRAME start=1 offset=0 + INTR_FRAME \start, (RIP-ORIG_RAX)+\offset + /*CFI_REL_OFFSET orig_rax, ORIG_RAX-ORIG_RAX*/ + .endm + +/* + * frame that enables calling into C. + */ + .macro PARTIAL_FRAME start=1 offset=0 + XCPT_FRAME \start, (ORIG_RAX-ARGOFFSET)+\offset + CFI_REL_OFFSET rdi, (RDI-ARGOFFSET)+\offset + CFI_REL_OFFSET rsi, (RSI-ARGOFFSET)+\offset + CFI_REL_OFFSET rdx, (RDX-ARGOFFSET)+\offset + CFI_REL_OFFSET rcx, (RCX-ARGOFFSET)+\offset + CFI_REL_OFFSET rax, (RAX-ARGOFFSET)+\offset + CFI_REL_OFFSET r8, (R8-ARGOFFSET)+\offset + CFI_REL_OFFSET r9, (R9-ARGOFFSET)+\offset + CFI_REL_OFFSET r10, (R10-ARGOFFSET)+\offset + CFI_REL_OFFSET r11, (R11-ARGOFFSET)+\offset + .endm + +/* + * frame that enables passing a complete pt_regs to a C function. + */ + .macro DEFAULT_FRAME start=1 offset=0 + PARTIAL_FRAME \start, (R11-R15)+\offset + CFI_REL_OFFSET rbx, RBX+\offset + CFI_REL_OFFSET rbp, RBP+\offset + CFI_REL_OFFSET r12, R12+\offset + CFI_REL_OFFSET r13, R13+\offset + CFI_REL_OFFSET r14, R14+\offset + CFI_REL_OFFSET r15, R15+\offset + .endm /* save partial stack frame */ ENTRY(save_args) XCPT_FRAME cld - movq %rdi, 8*8+16(%rsp) - CFI_REL_OFFSET rdi, 8*8+16 - movq %rsi, 7*8+16(%rsp) - CFI_REL_OFFSET rsi, 7*8+16 - movq %rdx, 6*8+16(%rsp) - CFI_REL_OFFSET rdx, 6*8+16 - movq %rcx, 5*8+16(%rsp) - CFI_REL_OFFSET rcx, 5*8+16 - movq %rax, 4*8+16(%rsp) - CFI_REL_OFFSET rax, 4*8+16 - movq %r8, 3*8+16(%rsp) - CFI_REL_OFFSET r8, 3*8+16 - movq %r9, 2*8+16(%rsp) - CFI_REL_OFFSET r9, 2*8+16 - movq %r10, 1*8+16(%rsp) - CFI_REL_OFFSET r10, 1*8+16 - movq %r11, 0*8+16(%rsp) - CFI_REL_OFFSET r11, 0*8+16 + CFI_MOVQ rdi, (RDI-ARGOFFSET)+16 + CFI_MOVQ rsi, (RSI-ARGOFFSET)+16 + CFI_MOVQ rdx, (RDX-ARGOFFSET)+16 + CFI_MOVQ rcx, (RCX-ARGOFFSET)+16 + CFI_MOVQ rax, (RAX-ARGOFFSET)+16 + CFI_MOVQ r8, (R8-ARGOFFSET)+16 + CFI_MOVQ r9, (R9-ARGOFFSET)+16 + CFI_MOVQ r10, (R10-ARGOFFSET)+16 + CFI_MOVQ r11, (R11-ARGOFFSET)+16 leaq -ARGOFFSET+16(%rsp),%rdi /* arg1 for handler */ - movq %rbp, 8(%rsp) /* push %rbp */ + CFI_MOVQ rbp, 8 /* push %rbp */ leaq 8(%rsp), %rbp /* mov %rsp, %ebp */ testl $3, CS(%rdi) je 1f @@ -303,9 +320,10 @@ ENTRY(save_args) */ 1: incl %gs:pda_irqcount jne 2f - pop %rax /* move return address... */ + CFI_POPQ %rax /* move return address... */ mov %gs:pda_irqstackptr,%rsp - push %rax /* ... to the new stack */ + EMPTY_FRAME 0 + CFI_PUSHQ %rax /* ... to the new stack */ /* * We entered an interrupt context - irqs are off: */ @@ -319,7 +337,7 @@ END(save_args) */ /* rdi: prev */ ENTRY(ret_from_fork) - CFI_DEFAULT_STACK + DEFAULT_FRAME push kernel_eflags(%rip) CFI_ADJUST_CFA_OFFSET 8 popf # reset kernel eflags @@ -732,6 +750,7 @@ END(interrupt) subq $10*8, %rsp CFI_ADJUST_CFA_OFFSET 10*8 call save_args + PARTIAL_FRAME 0 call \func .endm @@ -949,11 +968,11 @@ END(spurious_interrupt) .macro zeroentry sym INTR_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME - pushq $-1 /* ORIG_RAX: no syscall to restart */ - CFI_ADJUST_CFA_OFFSET 8 + CFI_PUSHQ $-1 /* ORIG_RAX: no syscall to restart */ subq $15*8,%rsp CFI_ADJUST_CFA_OFFSET 15*8 call error_entry + DEFAULT_FRAME 0 movq %rsp,%rdi /* pt_regs pointer */ xorl %esi,%esi /* no error code */ call \sym @@ -967,6 +986,7 @@ END(spurious_interrupt) subq $15*8,%rsp CFI_ADJUST_CFA_OFFSET 15*8 call error_entry + DEFAULT_FRAME 0 movq %rsp,%rdi /* pt_regs pointer */ movq ORIG_RAX(%rsp),%rsi /* get error code */ movq $-1,ORIG_RAX(%rsp) /* no syscall to restart */ @@ -1079,40 +1099,25 @@ paranoid_schedule\trace: * returns in "no swapgs flag" in %ebx. */ KPROBE_ENTRY(error_entry) - _frame RDI + XCPT_FRAME CFI_ADJUST_CFA_OFFSET 15*8 /* oldrax contains error code */ cld - movq %rdi,14*8+8(%rsp) - CFI_REL_OFFSET rdi,RDI+8 - movq %rsi,13*8+8(%rsp) - CFI_REL_OFFSET rsi,RSI+8 - movq %rdx,12*8+8(%rsp) - CFI_REL_OFFSET rdx,RDX+8 - movq %rcx,11*8+8(%rsp) - CFI_REL_OFFSET rcx,RCX+8 - movq %rax,10*8+8(%rsp) - CFI_REL_OFFSET rax,RAX+8 - movq %r8, 9*8+8(%rsp) - CFI_REL_OFFSET r8,R8+8 - movq %r9, 8*8+8(%rsp) - CFI_REL_OFFSET r9,R9+8 - movq %r10,7*8+8(%rsp) - CFI_REL_OFFSET r10,R10+8 - movq %r11,6*8+8(%rsp) - CFI_REL_OFFSET r11,R11+8 - movq %rbx,5*8+8(%rsp) - CFI_REL_OFFSET rbx,RBX+8 - movq %rbp,4*8+8(%rsp) - CFI_REL_OFFSET rbp,RBP+8 - movq %r12,3*8+8(%rsp) - CFI_REL_OFFSET r12,R12+8 - movq %r13,2*8+8(%rsp) - CFI_REL_OFFSET r13,R13+8 - movq %r14,1*8+8(%rsp) - CFI_REL_OFFSET r14,R14+8 - movq %r15,0*8+8(%rsp) - CFI_REL_OFFSET r15,R15+8 + CFI_MOVQ rdi, RDI+8 + CFI_MOVQ rsi, RSI+8 + CFI_MOVQ rdx, RDX+8 + CFI_MOVQ rcx, RCX+8 + CFI_MOVQ rax, RAX+8 + CFI_MOVQ r8, R8+8 + CFI_MOVQ r9, R9+8 + CFI_MOVQ r10, R10+8 + CFI_MOVQ r11, R11+8 + CFI_MOVQ rbx, RBX+8 + CFI_MOVQ rbp, RBP+8 + CFI_MOVQ r12, R12+8 + CFI_MOVQ r13, R13+8 + CFI_MOVQ r14, R14+8 + CFI_MOVQ r15, R15+8 xorl %ebx,%ebx testl $3,CS+8(%rsp) je error_kernelspace @@ -1146,7 +1151,7 @@ KPROBE_END(error_entry) /* ebx: no swapgs flag (1: don't need swapgs, 0: need it) */ KPROBE_ENTRY(error_exit) - _frame R15 + DEFAULT_FRAME movl %ebx,%eax RESTORE_REST DISABLE_INTERRUPTS(CLBR_NONE) @@ -1455,7 +1460,7 @@ ENTRY(xen_do_hypervisor_callback) # do_hypervisor_callback(struct *pt_regs) see the correct pointer to the pt_regs */ movq %rdi, %rsp # we don't return, adjust the stack frame CFI_ENDPROC - CFI_DEFAULT_STACK + DEFAULT_FRAME 11: incl %gs:pda_irqcount movq %rsp,%rbp CFI_DEF_CFA_REGISTER rbp @@ -1483,10 +1488,13 @@ END(do_hypervisor_callback) # with its current contents: any discrepancy means we in category 1. */ ENTRY(xen_failsafe_callback) - framesz = (RIP-0x30) /* workaround buggy gas */ - _frame framesz - CFI_REL_OFFSET rcx, 0 - CFI_REL_OFFSET r11, 8 + INTR_FRAME 1 (6*8) + /*CFI_REL_OFFSET gs,GS*/ + /*CFI_REL_OFFSET fs,FS*/ + /*CFI_REL_OFFSET es,ES*/ + /*CFI_REL_OFFSET ds,DS*/ + CFI_REL_OFFSET r11,8 + CFI_REL_OFFSET rcx,0 movw %ds,%cx cmpw %cx,0x10(%rsp) CFI_REMEMBER_STATE @@ -1507,12 +1515,9 @@ ENTRY(xen_failsafe_callback) CFI_RESTORE r11 addq $0x30,%rsp CFI_ADJUST_CFA_OFFSET -0x30 - pushq $0 - CFI_ADJUST_CFA_OFFSET 8 - pushq %r11 - CFI_ADJUST_CFA_OFFSET 8 - pushq %rcx - CFI_ADJUST_CFA_OFFSET 8 + CFI_PUSHQ $0 /* RIP */ + CFI_PUSHQ %r11 + CFI_PUSHQ %rcx jmp general_protection CFI_RESTORE_STATE 1: /* Segment mismatch => Category 1 (Bad segment). Retry the IRET. */ @@ -1522,8 +1527,7 @@ ENTRY(xen_failsafe_callback) CFI_RESTORE r11 addq $0x30,%rsp CFI_ADJUST_CFA_OFFSET -0x30 - pushq $0 - CFI_ADJUST_CFA_OFFSET 8 + CFI_PUSHQ $0 SAVE_ALL jmp error_exit CFI_ENDPROC -- cgit v1.2.3 From 3ddd972d970fdabbe6515aa2f95e0ef2c8df903d Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Thu, 20 Nov 2008 18:32:17 -0800 Subject: x86: signal: rename COPY_SEG_STRICT to COPY_SEG_CPL3 Impact: cleanup Rename macro COPY_SEG_STRICT to COPY_SEG_CPL3, as suggested by hpa. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/kernel/signal_32.c | 8 ++++---- arch/x86/kernel/signal_64.c | 8 ++++---- 2 files changed, 8 insertions(+), 8 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/signal_32.c b/arch/x86/kernel/signal_32.c index 514171ac0d03..c2aabeba27a5 100644 --- a/arch/x86/kernel/signal_32.c +++ b/arch/x86/kernel/signal_32.c @@ -145,7 +145,7 @@ asmlinkage int sys_sigaltstack(unsigned long bx) regs->seg = tmp; \ } -#define COPY_SEG_STRICT(seg) { \ +#define COPY_SEG_CPL3(seg) { \ unsigned short tmp; \ err |= __get_user(tmp, &sc->seg); \ regs->seg = tmp | 3; \ @@ -193,13 +193,13 @@ restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc, #endif /* CONFIG_X86_64 */ #ifdef CONFIG_X86_32 - COPY_SEG_STRICT(cs); - COPY_SEG_STRICT(ss); + COPY_SEG_CPL3(cs); + COPY_SEG_CPL3(ss); #else /* !CONFIG_X86_32 */ /* Kernel saves and restores only the CS segment register on signals, * which is the bare minimum needed to allow mixed 32/64-bit code. * App's signal handler can save/restore other segments if needed. */ - COPY_SEG_STRICT(cs); + COPY_SEG_CPL3(cs); #endif /* CONFIG_X86_32 */ err |= __get_user(tmpflags, &sc->flags); diff --git a/arch/x86/kernel/signal_64.c b/arch/x86/kernel/signal_64.c index d2307e41fbdb..3d54d366ccb2 100644 --- a/arch/x86/kernel/signal_64.c +++ b/arch/x86/kernel/signal_64.c @@ -56,7 +56,7 @@ sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss, err |= __get_user(regs->x, &sc->x); \ } -#define COPY_SEG_STRICT(seg) { \ +#define COPY_SEG_CPL3(seg) { \ unsigned short tmp; \ err |= __get_user(tmp, &sc->seg); \ regs->seg = tmp | 3; \ @@ -98,13 +98,13 @@ restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc, #endif /* CONFIG_X86_64 */ #ifdef CONFIG_X86_32 - COPY_SEG_STRICT(cs); - COPY_SEG_STRICT(ss); + COPY_SEG_CPL3(cs); + COPY_SEG_CPL3(ss); #else /* !CONFIG_X86_32 */ /* Kernel saves and restores only the CS segment register on signals, * which is the bare minimum needed to allow mixed 32/64-bit code. * App's signal handler can save/restore other segments if needed. */ - COPY_SEG_STRICT(cs); + COPY_SEG_CPL3(cs); #endif /* CONFIG_X86_32 */ err |= __get_user(tmpflags, &sc->flags); -- cgit v1.2.3 From e8a0e27662186f8856a0a6242e7a8386c9a64a53 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 21 Nov 2008 15:11:32 +0100 Subject: x86: clean up after: move entry_64.S register saving out of the macros, fix Impact: build fix The break builds with older binutils (2.16.1): arch/x86/kernel/entry_64.S: Assembler messages: arch/x86/kernel/entry_64.S:282: Error: too many positional arguments arch/x86/kernel/entry_64.S:283: Error: too many positional arguments arch/x86/kernel/entry_64.S:284: Error: too many positional arguments arch/x86/kernel/entry_64.S:285: Error: too many positional arguments arch/x86/kernel/entry_64.S:286: Error: too many positional arguments arch/x86/kernel/entry_64.S:287: Error: too many positional arguments arch/x86/kernel/entry_64.S:288: Error: too many positional arguments arch/x86/kernel/entry_64.S:289: Error: too many positional arguments arch/x86/kernel/entry_64.S:290: Error: too many positional arguments Took some time to figure out the detail that GAS chokes on: it's negative offsets. Rearrange the calculations to make sure we never go negative. Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_64.S | 54 +++++++++++++++++++++++----------------------- 1 file changed, 27 insertions(+), 27 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 7a04f696121d..4e3d83678f85 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -247,12 +247,12 @@ ENTRY(native_usergs_sysret64) * initial frame state for interrupts (and exceptions without error code) */ .macro INTR_FRAME start=1 offset=0 - EMPTY_FRAME \start, (SS+8-RIP)+\offset - /*CFI_REL_OFFSET ss, SS-RIP+\offset*/ - CFI_REL_OFFSET rsp, RSP-RIP+\offset - /*CFI_REL_OFFSET rflags, EFLAGS-RIP+\offset*/ - /*CFI_REL_OFFSET cs, CS-RIP+\offset*/ - CFI_REL_OFFSET rip, RIP-RIP+\offset + EMPTY_FRAME \start, SS+8+\offset-RIP + /*CFI_REL_OFFSET ss, SS+\offset-RIP*/ + CFI_REL_OFFSET rsp, RSP+\offset-RIP + /*CFI_REL_OFFSET rflags, EFLAGS+\offset-RIP*/ + /*CFI_REL_OFFSET cs, CS+\offset-RIP*/ + CFI_REL_OFFSET rip, RIP+\offset-RIP .endm /* @@ -260,7 +260,7 @@ ENTRY(native_usergs_sysret64) * with vector already pushed) */ .macro XCPT_FRAME start=1 offset=0 - INTR_FRAME \start, (RIP-ORIG_RAX)+\offset + INTR_FRAME \start, RIP+\offset-ORIG_RAX /*CFI_REL_OFFSET orig_rax, ORIG_RAX-ORIG_RAX*/ .endm @@ -268,23 +268,23 @@ ENTRY(native_usergs_sysret64) * frame that enables calling into C. */ .macro PARTIAL_FRAME start=1 offset=0 - XCPT_FRAME \start, (ORIG_RAX-ARGOFFSET)+\offset - CFI_REL_OFFSET rdi, (RDI-ARGOFFSET)+\offset - CFI_REL_OFFSET rsi, (RSI-ARGOFFSET)+\offset - CFI_REL_OFFSET rdx, (RDX-ARGOFFSET)+\offset - CFI_REL_OFFSET rcx, (RCX-ARGOFFSET)+\offset - CFI_REL_OFFSET rax, (RAX-ARGOFFSET)+\offset - CFI_REL_OFFSET r8, (R8-ARGOFFSET)+\offset - CFI_REL_OFFSET r9, (R9-ARGOFFSET)+\offset - CFI_REL_OFFSET r10, (R10-ARGOFFSET)+\offset - CFI_REL_OFFSET r11, (R11-ARGOFFSET)+\offset + XCPT_FRAME \start, ORIG_RAX+\offset-ARGOFFSET + CFI_REL_OFFSET rdi, RDI+\offset-ARGOFFSET + CFI_REL_OFFSET rsi, RSI+\offset-ARGOFFSET + CFI_REL_OFFSET rdx, RDX+\offset-ARGOFFSET + CFI_REL_OFFSET rcx, RCX+\offset-ARGOFFSET + CFI_REL_OFFSET rax, RAX+\offset-ARGOFFSET + CFI_REL_OFFSET r8, R8+\offset-ARGOFFSET + CFI_REL_OFFSET r9, R9+\offset-ARGOFFSET + CFI_REL_OFFSET r10, R10+\offset-ARGOFFSET + CFI_REL_OFFSET r11, R11+\offset-ARGOFFSET .endm /* * frame that enables passing a complete pt_regs to a C function. */ .macro DEFAULT_FRAME start=1 offset=0 - PARTIAL_FRAME \start, (R11-R15)+\offset + PARTIAL_FRAME \start, R11+\offset-R15 CFI_REL_OFFSET rbx, RBX+\offset CFI_REL_OFFSET rbp, RBP+\offset CFI_REL_OFFSET r12, R12+\offset @@ -297,15 +297,15 @@ ENTRY(native_usergs_sysret64) ENTRY(save_args) XCPT_FRAME cld - CFI_MOVQ rdi, (RDI-ARGOFFSET)+16 - CFI_MOVQ rsi, (RSI-ARGOFFSET)+16 - CFI_MOVQ rdx, (RDX-ARGOFFSET)+16 - CFI_MOVQ rcx, (RCX-ARGOFFSET)+16 - CFI_MOVQ rax, (RAX-ARGOFFSET)+16 - CFI_MOVQ r8, (R8-ARGOFFSET)+16 - CFI_MOVQ r9, (R9-ARGOFFSET)+16 - CFI_MOVQ r10, (R10-ARGOFFSET)+16 - CFI_MOVQ r11, (R11-ARGOFFSET)+16 + CFI_MOVQ rdi, RDI+16-ARGOFFSET + CFI_MOVQ rsi, RSI+16-ARGOFFSET + CFI_MOVQ rdx, RDX+16-ARGOFFSET + CFI_MOVQ rcx, RCX+16-ARGOFFSET + CFI_MOVQ rax, RAX+16-ARGOFFSET + CFI_MOVQ r8, R8+16-ARGOFFSET + CFI_MOVQ r9, R9+16-ARGOFFSET + CFI_MOVQ r10, R10+16-ARGOFFSET + CFI_MOVQ r11, R11+16-ARGOFFSET leaq -ARGOFFSET+16(%rsp),%rdi /* arg1 for handler */ CFI_MOVQ rbp, 8 /* push %rbp */ leaq 8(%rsp), %rbp /* mov %rsp, %ebp */ -- cgit v1.2.3 From 14ae22ba2b8bb3d53fb795f9b8074aa39ef7b6cd Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 21 Nov 2008 15:20:47 +0100 Subject: x86: entry_64.S: rename Impact: cleanup Rename: CFI_PUSHQ => pushq_cfi CFI_POPQ => popq_cfi CFI_MOVQ => movq_cfi To make it blend better into regular assembly code. Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_64.S | 71 +++++++++++++++++++++++----------------------- 1 file changed, 36 insertions(+), 35 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 4e3d83678f85..92c5e18340db 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -63,17 +63,17 @@ /* * Some macro's to hide the most frequently occuring CFI annotations. */ - .macro CFI_PUSHQ reg + .macro pushq_cfi reg pushq \reg CFI_ADJUST_CFA_OFFSET 8 .endm - .macro CFI_POPQ reg + .macro popq_cfi reg popq \reg CFI_ADJUST_CFA_OFFSET -8 .endm - .macro CFI_MOVQ reg offset=0 + .macro movq_cfi reg offset=0 movq %\reg, \offset(%rsp) CFI_REL_OFFSET \reg, \offset .endm @@ -297,17 +297,18 @@ ENTRY(native_usergs_sysret64) ENTRY(save_args) XCPT_FRAME cld - CFI_MOVQ rdi, RDI+16-ARGOFFSET - CFI_MOVQ rsi, RSI+16-ARGOFFSET - CFI_MOVQ rdx, RDX+16-ARGOFFSET - CFI_MOVQ rcx, RCX+16-ARGOFFSET - CFI_MOVQ rax, RAX+16-ARGOFFSET - CFI_MOVQ r8, R8+16-ARGOFFSET - CFI_MOVQ r9, R9+16-ARGOFFSET - CFI_MOVQ r10, R10+16-ARGOFFSET - CFI_MOVQ r11, R11+16-ARGOFFSET + movq_cfi rdi, RDI+16-ARGOFFSET + movq_cfi rsi, RSI+16-ARGOFFSET + movq_cfi rdx, RDX+16-ARGOFFSET + movq_cfi rcx, RCX+16-ARGOFFSET + movq_cfi rax, RAX+16-ARGOFFSET + movq_cfi r8, R8+16-ARGOFFSET + movq_cfi r9, R9+16-ARGOFFSET + movq_cfi r10, R10+16-ARGOFFSET + movq_cfi r11, R11+16-ARGOFFSET + leaq -ARGOFFSET+16(%rsp),%rdi /* arg1 for handler */ - CFI_MOVQ rbp, 8 /* push %rbp */ + movq_cfi rbp, 8 /* push %rbp */ leaq 8(%rsp), %rbp /* mov %rsp, %ebp */ testl $3, CS(%rdi) je 1f @@ -320,10 +321,10 @@ ENTRY(save_args) */ 1: incl %gs:pda_irqcount jne 2f - CFI_POPQ %rax /* move return address... */ + popq_cfi %rax /* move return address... */ mov %gs:pda_irqstackptr,%rsp EMPTY_FRAME 0 - CFI_PUSHQ %rax /* ... to the new stack */ + pushq_cfi %rax /* ... to the new stack */ /* * We entered an interrupt context - irqs are off: */ @@ -968,7 +969,7 @@ END(spurious_interrupt) .macro zeroentry sym INTR_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME - CFI_PUSHQ $-1 /* ORIG_RAX: no syscall to restart */ + pushq_cfi $-1 /* ORIG_RAX: no syscall to restart */ subq $15*8,%rsp CFI_ADJUST_CFA_OFFSET 15*8 call error_entry @@ -1103,21 +1104,21 @@ KPROBE_ENTRY(error_entry) CFI_ADJUST_CFA_OFFSET 15*8 /* oldrax contains error code */ cld - CFI_MOVQ rdi, RDI+8 - CFI_MOVQ rsi, RSI+8 - CFI_MOVQ rdx, RDX+8 - CFI_MOVQ rcx, RCX+8 - CFI_MOVQ rax, RAX+8 - CFI_MOVQ r8, R8+8 - CFI_MOVQ r9, R9+8 - CFI_MOVQ r10, R10+8 - CFI_MOVQ r11, R11+8 - CFI_MOVQ rbx, RBX+8 - CFI_MOVQ rbp, RBP+8 - CFI_MOVQ r12, R12+8 - CFI_MOVQ r13, R13+8 - CFI_MOVQ r14, R14+8 - CFI_MOVQ r15, R15+8 + movq_cfi rdi, RDI+8 + movq_cfi rsi, RSI+8 + movq_cfi rdx, RDX+8 + movq_cfi rcx, RCX+8 + movq_cfi rax, RAX+8 + movq_cfi r8, R8+8 + movq_cfi r9, R9+8 + movq_cfi r10, R10+8 + movq_cfi r11, R11+8 + movq_cfi rbx, RBX+8 + movq_cfi rbp, RBP+8 + movq_cfi r12, R12+8 + movq_cfi r13, R13+8 + movq_cfi r14, R14+8 + movq_cfi r15, R15+8 xorl %ebx,%ebx testl $3,CS+8(%rsp) je error_kernelspace @@ -1515,9 +1516,9 @@ ENTRY(xen_failsafe_callback) CFI_RESTORE r11 addq $0x30,%rsp CFI_ADJUST_CFA_OFFSET -0x30 - CFI_PUSHQ $0 /* RIP */ - CFI_PUSHQ %r11 - CFI_PUSHQ %rcx + pushq_cfi $0 /* RIP */ + pushq_cfi %r11 + pushq_cfi %rcx jmp general_protection CFI_RESTORE_STATE 1: /* Segment mismatch => Category 1 (Bad segment). Retry the IRET. */ @@ -1527,7 +1528,7 @@ ENTRY(xen_failsafe_callback) CFI_RESTORE r11 addq $0x30,%rsp CFI_ADJUST_CFA_OFFSET -0x30 - CFI_PUSHQ $0 + pushq_cfi $0 SAVE_ALL jmp error_exit CFI_ENDPROC -- cgit v1.2.3 From c002a1e6b6b6f07ae04e68987054bf1f2150ae48 Mon Sep 17 00:00:00 2001 From: Alexander van Heukelum Date: Fri, 21 Nov 2008 16:41:55 +0100 Subject: x86: introduce save_rest and restructure the PTREGSCALL macro in entry_64.S Impact: cleanup The save_rest function completes a partial stack frame for use by the PTREGSCALL macro. This also avoids the indirect call in PTREGSCALLs. This adds the macro movq_cfi_restore to hide the CFI_RESTORE annotation when restoring a register from the stack frame. Signed-off-by: Alexander van Heukelum Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_64.S | 91 +++++++++++++++++++++++++++------------------- 1 file changed, 53 insertions(+), 38 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 92c5e18340db..ef95c45b9269 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -78,6 +78,11 @@ CFI_REL_OFFSET \reg, \offset .endm + .macro movq_cfi_restore offset reg + movq \offset(%rsp), %\reg + CFI_RESTORE \reg + .endm + #ifdef CONFIG_FUNCTION_TRACER #ifdef CONFIG_DYNAMIC_FTRACE ENTRY(mcount) @@ -186,21 +191,21 @@ ENTRY(native_usergs_sysret64) */ /* %rsp:at FRAMEEND */ - .macro FIXUP_TOP_OF_STACK tmp - movq %gs:pda_oldrsp,\tmp - movq \tmp,RSP(%rsp) - movq $__USER_DS,SS(%rsp) - movq $__USER_CS,CS(%rsp) - movq $-1,RCX(%rsp) - movq R11(%rsp),\tmp /* get eflags */ - movq \tmp,EFLAGS(%rsp) + .macro FIXUP_TOP_OF_STACK tmp offset=0 + movq %gs:pda_oldrsp,\tmp + movq \tmp,RSP+\offset(%rsp) + movq $__USER_DS,SS+\offset(%rsp) + movq $__USER_CS,CS+\offset(%rsp) + movq $-1,RCX+\offset(%rsp) + movq R11+\offset(%rsp),\tmp /* get eflags */ + movq \tmp,EFLAGS+\offset(%rsp) .endm - .macro RESTORE_TOP_OF_STACK tmp,offset=0 - movq RSP-\offset(%rsp),\tmp - movq \tmp,%gs:pda_oldrsp - movq EFLAGS-\offset(%rsp),\tmp - movq \tmp,R11-\offset(%rsp) + .macro RESTORE_TOP_OF_STACK tmp offset=0 + movq RSP+\offset(%rsp),\tmp + movq \tmp,%gs:pda_oldrsp + movq EFLAGS+\offset(%rsp),\tmp + movq \tmp,R11+\offset(%rsp) .endm .macro FAKE_STACK_FRAME child_rip @@ -333,6 +338,21 @@ ENTRY(save_args) CFI_ENDPROC END(save_args) +ENTRY(save_rest) + PARTIAL_FRAME 1 REST_SKIP+8 + movq 5*8+16(%rsp), %r11 /* save return address */ + movq_cfi rbx, RBX+16 + movq_cfi rbp, RBP+16 + movq_cfi r12, R12+16 + movq_cfi r13, R13+16 + movq_cfi r14, R14+16 + movq_cfi r15, R15+16 + movq %r11, 8(%rsp) /* return address */ + FIXUP_TOP_OF_STACK %r11, 16 + ret + CFI_ENDPROC +END(save_rest) + /* * A newly forked process directly context switches into this. */ @@ -353,7 +373,7 @@ rff_action: je int_ret_from_sys_call testl $_TIF_IA32,TI_flags(%rcx) jnz int_ret_from_sys_call - RESTORE_TOP_OF_STACK %rdi,ARGOFFSET + RESTORE_TOP_OF_STACK %rdi, -ARGOFFSET jmp ret_from_sys_call rff_trace: movq %rsp,%rdi @@ -626,18 +646,20 @@ END(system_call) /* * Certain special system calls that need to save a complete full stack frame. */ - .macro PTREGSCALL label,func,arg - .globl \label -\label: - leaq \func(%rip),%rax - leaq -ARGOFFSET+8(%rsp),\arg /* 8 for return address */ - jmp ptregscall_common +ENTRY(\label) + PARTIAL_FRAME 1 8 /* offset 8: return address */ + subq $REST_SKIP, %rsp + CFI_ADJUST_CFA_OFFSET REST_SKIP + call save_rest + DEFAULT_FRAME 0 8 /* offset 8: return address */ + leaq 8(%rsp), \arg /* pt_regs pointer */ + call \func + jmp ptregscall_common + CFI_ENDPROC END(\label) .endm - CFI_STARTPROC - PTREGSCALL stub_clone, sys_clone, %r8 PTREGSCALL stub_fork, sys_fork, %rdi PTREGSCALL stub_vfork, sys_vfork, %rdi @@ -645,22 +667,15 @@ END(\label) PTREGSCALL stub_iopl, sys_iopl, %rsi ENTRY(ptregscall_common) - popq %r11 - CFI_ADJUST_CFA_OFFSET -8 - CFI_REGISTER rip, r11 - SAVE_REST - movq %r11, %r15 - CFI_REGISTER rip, r15 - FIXUP_TOP_OF_STACK %r11 - call *%rax - RESTORE_TOP_OF_STACK %r11 - movq %r15, %r11 - CFI_REGISTER rip, r11 - RESTORE_REST - pushq %r11 - CFI_ADJUST_CFA_OFFSET 8 - CFI_REL_OFFSET rip, 0 - ret + DEFAULT_FRAME 1 8 /* offset 8: return address */ + RESTORE_TOP_OF_STACK %r11, 8 + movq_cfi_restore R15+8, r15 + movq_cfi_restore R14+8, r14 + movq_cfi_restore R13+8, r13 + movq_cfi_restore R12+8, r12 + movq_cfi_restore RBP+8, rbp + movq_cfi_restore RBX+8, rbx + ret $REST_SKIP /* pop extended registers */ CFI_ENDPROC END(ptregscall_common) -- cgit v1.2.3 From e2f6bc25b98dbb10d809ee50262b43fcae67840a Mon Sep 17 00:00:00 2001 From: Alexander van Heukelum Date: Fri, 21 Nov 2008 16:43:18 +0100 Subject: x86: entry_64.S: factor out save_paranoid and paranoid_exit Impact: cleanup, shrink kernel image size Also expand the paranoid_exit0 macro into nmi_exit inside the nmi stub in the case of enabled irq-tracing. This gives a few hundred bytes code size reduction. Signed-off-by: Alexander van Heukelum Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_64.S | 151 ++++++++++++++++++++++++++++++--------------- 1 file changed, 102 insertions(+), 49 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index ef95c45b9269..fad777b11366 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -353,6 +353,36 @@ ENTRY(save_rest) CFI_ENDPROC END(save_rest) +/* save complete stack frame */ +ENTRY(save_paranoid) + XCPT_FRAME 1 RDI+8 + cld + movq_cfi rdi, RDI+8 + movq_cfi rsi, RSI+8 + movq_cfi rdx, RDX+8 + movq_cfi rcx, RCX+8 + movq_cfi rax, RAX+8 + movq_cfi r8, R8+8 + movq_cfi r9, R9+8 + movq_cfi r10, R10+8 + movq_cfi r11, R11+8 + movq_cfi rbx, RBX+8 + movq_cfi rbp, RBP+8 + movq_cfi r12, R12+8 + movq_cfi r13, R13+8 + movq_cfi r14, R14+8 + movq_cfi r15, R15+8 + movl $1,%ebx + movl $MSR_GS_BASE,%ecx + rdmsr + testl %edx,%edx + js 1f /* negative -> in kernel */ + SWAPGS + xorl %ebx,%ebx +1: ret + CFI_ENDPROC +END(save_paranoid) + /* * A newly forked process directly context switches into this. */ @@ -1012,24 +1042,15 @@ END(spurious_interrupt) .endm /* error code is on the stack already */ - /* handle NMI like exceptions that can happen everywhere */ - .macro paranoidentry sym, ist=0, irqtrace=1 - SAVE_ALL - cld - movl $1,%ebx - movl $MSR_GS_BASE,%ecx - rdmsr - testl %edx,%edx - js 1f - SWAPGS - xorl %ebx,%ebx -1: + .macro paranoidentry sym ist=0 + subq $15*8, %rsp + CFI_ADJUST_CFA_OFFSET 15*8 + call save_paranoid + DEFAULT_FRAME 0 .if \ist movq %gs:pda_data_offset, %rbp .endif - .if \irqtrace TRACE_IRQS_OFF - .endif movq %rsp,%rdi movq ORIG_RAX(%rsp),%rsi movq $-1,ORIG_RAX(%rsp) @@ -1041,9 +1062,7 @@ END(spurious_interrupt) addq $EXCEPTION_STKSZ, per_cpu__init_tss + TSS_ist + (\ist - 1) * 8(%rbp) .endif DISABLE_INTERRUPTS(CLBR_NONE) - .if \irqtrace TRACE_IRQS_OFF - .endif .endm /* @@ -1058,57 +1077,48 @@ END(spurious_interrupt) * is fundamentally NMI-unsafe. (we cannot change the soft and * hard flags at once, atomically) */ - .macro paranoidexit trace=1 + /* ebx: no swapgs flag */ -paranoid_exit\trace: +KPROBE_ENTRY(paranoid_exit) + INTR_FRAME testl %ebx,%ebx /* swapgs needed? */ - jnz paranoid_restore\trace + jnz paranoid_restore testl $3,CS(%rsp) - jnz paranoid_userspace\trace -paranoid_swapgs\trace: - .if \trace + jnz paranoid_userspace +paranoid_swapgs: TRACE_IRQS_IRETQ 0 - .endif SWAPGS_UNSAFE_STACK -paranoid_restore\trace: +paranoid_restore: RESTORE_ALL 8 jmp irq_return -paranoid_userspace\trace: +paranoid_userspace: GET_THREAD_INFO(%rcx) movl TI_flags(%rcx),%ebx andl $_TIF_WORK_MASK,%ebx - jz paranoid_swapgs\trace + jz paranoid_swapgs movq %rsp,%rdi /* &pt_regs */ call sync_regs movq %rax,%rsp /* switch stack for scheduling */ testl $_TIF_NEED_RESCHED,%ebx - jnz paranoid_schedule\trace + jnz paranoid_schedule movl %ebx,%edx /* arg3: thread flags */ - .if \trace TRACE_IRQS_ON - .endif ENABLE_INTERRUPTS(CLBR_NONE) xorl %esi,%esi /* arg2: oldset */ movq %rsp,%rdi /* arg1: &pt_regs */ call do_notify_resume DISABLE_INTERRUPTS(CLBR_NONE) - .if \trace TRACE_IRQS_OFF - .endif - jmp paranoid_userspace\trace -paranoid_schedule\trace: - .if \trace + jmp paranoid_userspace +paranoid_schedule: TRACE_IRQS_ON - .endif ENABLE_INTERRUPTS(CLBR_ANY) call schedule DISABLE_INTERRUPTS(CLBR_ANY) - .if \trace TRACE_IRQS_OFF - .endif - jmp paranoid_userspace\trace + jmp paranoid_userspace CFI_ENDPROC - .endm +END(paranoid_exit) /* * Exception entry point. This expects an error code/orig_rax on the stack. @@ -1326,20 +1336,63 @@ KPROBE_ENTRY(debug) pushq $0 CFI_ADJUST_CFA_OFFSET 8 paranoidentry do_debug, DEBUG_STACK - paranoidexit + jmp paranoid_exit + CFI_ENDPROC KPROBE_END(debug) /* runs on exception stack */ KPROBE_ENTRY(nmi) INTR_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME - pushq $-1 - CFI_ADJUST_CFA_OFFSET 8 - paranoidentry do_nmi, 0, 0 + pushq_cfi $-1 + subq $15*8, %rsp + CFI_ADJUST_CFA_OFFSET 15*8 + call save_paranoid + DEFAULT_FRAME 0 + /* paranoidentry do_nmi, 0; without TRACE_IRQS_OFF */ + movq %rsp,%rdi + movq ORIG_RAX(%rsp),%rsi + movq $-1,ORIG_RAX(%rsp) + call do_nmi + DISABLE_INTERRUPTS(CLBR_NONE) #ifdef CONFIG_TRACE_IRQFLAGS - paranoidexit 0 + /* paranoidexit; without TRACE_IRQS_OFF */ + /* ebx: no swapgs flag */ +nmi_exit: + testl %ebx,%ebx /* swapgs needed? */ + jnz nmi_restore + testl $3,CS(%rsp) + jnz nmi_userspace +nmi_swapgs: + SWAPGS_UNSAFE_STACK +nmi_restore: + RESTORE_ALL 8 + jmp irq_return +nmi_userspace: + GET_THREAD_INFO(%rcx) + movl TI_flags(%rcx),%ebx + andl $_TIF_WORK_MASK,%ebx + jz nmi_swapgs + movq %rsp,%rdi /* &pt_regs */ + call sync_regs + movq %rax,%rsp /* switch stack for scheduling */ + testl $_TIF_NEED_RESCHED,%ebx + jnz nmi_schedule + movl %ebx,%edx /* arg3: thread flags */ + ENABLE_INTERRUPTS(CLBR_NONE) + xorl %esi,%esi /* arg2: oldset */ + movq %rsp,%rdi /* arg1: &pt_regs */ + call do_notify_resume + DISABLE_INTERRUPTS(CLBR_NONE) + jmp nmi_userspace +nmi_schedule: + ENABLE_INTERRUPTS(CLBR_ANY) + call schedule + DISABLE_INTERRUPTS(CLBR_ANY) + jmp nmi_userspace + CFI_ENDPROC #else - jmp paranoid_exit1 + jmp paranoid_exit CFI_ENDPROC #endif KPROBE_END(nmi) @@ -1350,7 +1403,7 @@ KPROBE_ENTRY(int3) pushq $0 CFI_ADJUST_CFA_OFFSET 8 paranoidentry do_int3, DEBUG_STACK - jmp paranoid_exit1 + jmp paranoid_exit CFI_ENDPROC KPROBE_END(int3) @@ -1375,7 +1428,7 @@ ENTRY(double_fault) XCPT_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME paranoidentry do_double_fault - jmp paranoid_exit1 + jmp paranoid_exit CFI_ENDPROC END(double_fault) @@ -1392,7 +1445,7 @@ ENTRY(stack_segment) XCPT_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME paranoidentry do_stack_segment - jmp paranoid_exit1 + jmp paranoid_exit CFI_ENDPROC END(stack_segment) @@ -1420,7 +1473,7 @@ ENTRY(machine_check) pushq $0 CFI_ADJUST_CFA_OFFSET 8 paranoidentry do_machine_check - jmp paranoid_exit1 + jmp paranoid_exit CFI_ENDPROC END(machine_check) #endif -- cgit v1.2.3 From b8b1d08bf6fe7c09e6cb2294bc0e5e964b361241 Mon Sep 17 00:00:00 2001 From: Alexander van Heukelum Date: Fri, 21 Nov 2008 16:44:28 +0100 Subject: x86: entry_64.S: split out some macro's and move common code to paranoid_exit Impact: cleanup DISABLE_INTERRUPTS(CLBR_NONE)/TRACE_IRQS_OFF is now always executed just before paranoid_exit. Move it there. Split out paranoidzeroentry, paranoiderrorentry, and paranoidzeroentry_ist to get more readable macro's. Signed-off-by: Alexander van Heukelum Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_64.S | 102 ++++++++++++++++++++++----------------------- 1 file changed, 51 insertions(+), 51 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index fad777b11366..692c1da61905 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -1026,6 +1026,39 @@ END(spurious_interrupt) CFI_ENDPROC .endm + .macro paranoidzeroentry sym + INTR_FRAME + PARAVIRT_ADJUST_EXCEPTION_FRAME + pushq $-1 /* ORIG_RAX: no syscall to restart */ + CFI_ADJUST_CFA_OFFSET 8 + subq $15*8, %rsp + call save_paranoid + TRACE_IRQS_OFF + movq %rsp,%rdi /* pt_regs pointer */ + xorl %esi,%esi /* no error code */ + call \sym + jmp paranoid_exit /* %ebx: no swapgs flag */ + CFI_ENDPROC + .endm + + .macro paranoidzeroentry_ist sym ist + INTR_FRAME + PARAVIRT_ADJUST_EXCEPTION_FRAME + pushq $-1 /* ORIG_RAX: no syscall to restart */ + CFI_ADJUST_CFA_OFFSET 8 + subq $15*8, %rsp + call save_paranoid + TRACE_IRQS_OFF + movq %rsp,%rdi /* pt_regs pointer */ + xorl %esi,%esi /* no error code */ + movq %gs:pda_data_offset, %rbp + subq $EXCEPTION_STKSZ, per_cpu__init_tss + TSS_ist + (\ist - 1) * 8(%rbp) + call \sym + addq $EXCEPTION_STKSZ, per_cpu__init_tss + TSS_ist + (\ist - 1) * 8(%rbp) + jmp paranoid_exit /* %ebx: no swapgs flag */ + CFI_ENDPROC + .endm + .macro errorentry sym XCPT_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME @@ -1042,27 +1075,20 @@ END(spurious_interrupt) .endm /* error code is on the stack already */ - .macro paranoidentry sym ist=0 - subq $15*8, %rsp + .macro paranoiderrorentry sym + XCPT_FRAME + PARAVIRT_ADJUST_EXCEPTION_FRAME + subq $15*8,%rsp CFI_ADJUST_CFA_OFFSET 15*8 call save_paranoid DEFAULT_FRAME 0 - .if \ist - movq %gs:pda_data_offset, %rbp - .endif TRACE_IRQS_OFF - movq %rsp,%rdi - movq ORIG_RAX(%rsp),%rsi - movq $-1,ORIG_RAX(%rsp) - .if \ist - subq $EXCEPTION_STKSZ, per_cpu__init_tss + TSS_ist + (\ist - 1) * 8(%rbp) - .endif + movq %rsp,%rdi /* pt_regs pointer */ + movq ORIG_RAX(%rsp),%rsi /* get error code */ + movq $-1,ORIG_RAX(%rsp) /* no syscall to restart */ call \sym - .if \ist - addq $EXCEPTION_STKSZ, per_cpu__init_tss + TSS_ist + (\ist - 1) * 8(%rbp) - .endif - DISABLE_INTERRUPTS(CLBR_NONE) - TRACE_IRQS_OFF + jmp paranoid_exit /* %ebx: no swapgs flag */ + CFI_ENDPROC .endm /* @@ -1081,6 +1107,8 @@ END(spurious_interrupt) /* ebx: no swapgs flag */ KPROBE_ENTRY(paranoid_exit) INTR_FRAME + DISABLE_INTERRUPTS(CLBR_NONE) + TRACE_IRQS_OFF testl %ebx,%ebx /* swapgs needed? */ jnz paranoid_restore testl $3,CS(%rsp) @@ -1331,13 +1359,7 @@ END(device_not_available) /* runs on exception stack */ KPROBE_ENTRY(debug) - INTR_FRAME - PARAVIRT_ADJUST_EXCEPTION_FRAME - pushq $0 - CFI_ADJUST_CFA_OFFSET 8 - paranoidentry do_debug, DEBUG_STACK - jmp paranoid_exit - CFI_ENDPROC + paranoidzeroentry_ist do_debug, DEBUG_STACK KPROBE_END(debug) /* runs on exception stack */ @@ -1351,14 +1373,12 @@ KPROBE_ENTRY(nmi) DEFAULT_FRAME 0 /* paranoidentry do_nmi, 0; without TRACE_IRQS_OFF */ movq %rsp,%rdi - movq ORIG_RAX(%rsp),%rsi - movq $-1,ORIG_RAX(%rsp) + movq $-1,%rsi call do_nmi - DISABLE_INTERRUPTS(CLBR_NONE) #ifdef CONFIG_TRACE_IRQFLAGS /* paranoidexit; without TRACE_IRQS_OFF */ /* ebx: no swapgs flag */ -nmi_exit: + DISABLE_INTERRUPTS(CLBR_NONE) testl %ebx,%ebx /* swapgs needed? */ jnz nmi_restore testl $3,CS(%rsp) @@ -1398,13 +1418,7 @@ nmi_schedule: KPROBE_END(nmi) KPROBE_ENTRY(int3) - INTR_FRAME - PARAVIRT_ADJUST_EXCEPTION_FRAME - pushq $0 - CFI_ADJUST_CFA_OFFSET 8 - paranoidentry do_int3, DEBUG_STACK - jmp paranoid_exit - CFI_ENDPROC + paranoidzeroentry_ist do_int3, DEBUG_STACK KPROBE_END(int3) ENTRY(overflow) @@ -1425,11 +1439,7 @@ END(coprocessor_segment_overrun) /* runs on exception stack */ ENTRY(double_fault) - XCPT_FRAME - PARAVIRT_ADJUST_EXCEPTION_FRAME - paranoidentry do_double_fault - jmp paranoid_exit - CFI_ENDPROC + paranoiderrorentry do_double_fault END(double_fault) ENTRY(invalid_TSS) @@ -1442,11 +1452,7 @@ END(segment_not_present) /* runs on exception stack */ ENTRY(stack_segment) - XCPT_FRAME - PARAVIRT_ADJUST_EXCEPTION_FRAME - paranoidentry do_stack_segment - jmp paranoid_exit - CFI_ENDPROC + paranoiderrorentry do_stack_segment END(stack_segment) KPROBE_ENTRY(general_protection) @@ -1468,13 +1474,7 @@ END(spurious_interrupt_bug) #ifdef CONFIG_X86_MCE /* runs on exception stack */ ENTRY(machine_check) - INTR_FRAME - PARAVIRT_ADJUST_EXCEPTION_FRAME - pushq $0 - CFI_ADJUST_CFA_OFFSET 8 - paranoidentry do_machine_check - jmp paranoid_exit - CFI_ENDPROC + paranoidzeroentry do_machine_check END(machine_check) #endif -- cgit v1.2.3 From c81084114f6ff957bc6b5a0048350479c1c1f7b3 Mon Sep 17 00:00:00 2001 From: Alexander van Heukelum Date: Fri, 21 Nov 2008 22:59:52 +0100 Subject: x86: split out some macro's and move common code to paranoid_exit, fix Impact: fix bootup crash Even though it tested fine for me, there was still a bug in the first patch: I have overlooked a call to ptregscall_common. This patch fixes that, I think, but the code is never executed for me while running a debian install... (I tested this by putting an "1:jmp 1b" in there.) Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_64.S | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 692c1da61905..e5ddf573ded2 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -529,10 +529,13 @@ sysret_signal: jc sysret_audit #endif /* edx: work flags (arg3) */ - leaq do_notify_resume(%rip),%rax leaq -ARGOFFSET(%rsp),%rdi # &pt_regs -> arg1 xorl %esi,%esi # oldset -> arg2 - call ptregscall_common + SAVE_REST + FIXUP_TOP_OF_STACK %r11 + call do_notify_resume + RESTORE_TOP_OF_STACK %r11 + RESTORE_REST movl $_TIF_WORK_MASK,%edi /* Use IRET because user could have changed frame. This works because ptregscall_common has called FIXUP_TOP_OF_STACK. */ -- cgit v1.2.3 From 3889d0cea2b73049bdca062d9ff1e5d33468289c Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Sat, 22 Nov 2008 23:39:23 -0800 Subject: x86: revert default reboot method to REBOOT_KBD Impact: Reverts default reboot method. Checkin 14d7ca5c575853664d8fe4f225a77b8df1b7de7d changed the default reboot method to "pci", a.k.a. port CF9. Unfortunately this has been shown to cause lockups on at least two systems for which REBOOT_KBD worked, both Thinkpads with Intel chipsets. This reverts the default to REBOOT_KBD, while leaving the option to have "reboot=pci" specified explicitly or via a DMI match. Signed-off-by: H. Peter Anvin --- arch/x86/kernel/reboot.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c index ddc93891cdcd..790b09fbadcb 100644 --- a/arch/x86/kernel/reboot.c +++ b/arch/x86/kernel/reboot.c @@ -29,7 +29,7 @@ EXPORT_SYMBOL(pm_power_off); static const struct desc_ptr no_idt = {}; static int reboot_mode; -enum reboot_type reboot_type = BOOT_CF9_COND; +enum reboot_type reboot_type = BOOT_KBD; int reboot_force; #if defined(CONFIG_X86_32) && defined(CONFIG_SMP) -- cgit v1.2.3 From 3aeb95d5b7839708a8d8e11aa274ee4d0d4042cc Mon Sep 17 00:00:00 2001 From: jia zhang Date: Sun, 23 Nov 2008 09:51:41 +0800 Subject: x86_64: fix the check in stack_overflow_check Impact: make stack overflow debug check and printout narrower stack_overflow_check() should consider the stack usage of pt_regs, and thus it could warn us in advance. Additionally, it looks better for the warning time to start at INITIAL_JIFFIES. Assuming that rsp gets close to the check point before interrupt arrives: when interrupt really happens, thread_info will be partly overrode. Signed-off-by: jia zhang Signed-off-by: Ingo Molnar --- arch/x86/kernel/irq_64.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/irq_64.c b/arch/x86/kernel/irq_64.c index 60eb84eb77a0..b842fc82be15 100644 --- a/arch/x86/kernel/irq_64.c +++ b/arch/x86/kernel/irq_64.c @@ -29,11 +29,12 @@ static inline void stack_overflow_check(struct pt_regs *regs) { u64 curbase = (u64)task_stack_page(current); - static unsigned long warned = -60*HZ; + static unsigned long warned = INITIAL_JIFFIES - 60*HZ; if (regs->sp >= curbase && regs->sp <= curbase + THREAD_SIZE && - regs->sp < curbase + sizeof(struct thread_info) + 128 && - time_after(jiffies, warned + 60*HZ)) { + regs->sp < curbase + sizeof(struct thread_info) + + sizeof(struct pt_regs) + 128 && + time_after(jiffies, warned + 60*HZ)) { printk("do_IRQ: %s near stack overflow (cur:%Lx,sp:%lx)\n", current->comm, curbase, regs->sp); show_stack(NULL,NULL); -- cgit v1.2.3 From f377fa123d0ec621e8e361ecc3f2a8ee70e81a2e Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Sun, 23 Nov 2008 09:02:26 +0100 Subject: x86: clean up stack overflow debug check Impact: cleanup Simplify the irq-sampled stack overflow debug check: - eliminate an #idef - use WARN_ONCE() to emit a single warning (all bets are off after the first such warning anyway) Signed-off-by: Ingo Molnar --- arch/x86/kernel/irq_64.c | 25 ++++++++++--------------- 1 file changed, 10 insertions(+), 15 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/irq_64.c b/arch/x86/kernel/irq_64.c index b842fc82be15..1d3d0e71b044 100644 --- a/arch/x86/kernel/irq_64.c +++ b/arch/x86/kernel/irq_64.c @@ -18,7 +18,6 @@ #include #include -#ifdef CONFIG_DEBUG_STACKOVERFLOW /* * Probabilistic stack overflow check: * @@ -28,20 +27,18 @@ */ static inline void stack_overflow_check(struct pt_regs *regs) { +#ifdef CONFIG_DEBUG_STACKOVERFLOW u64 curbase = (u64)task_stack_page(current); - static unsigned long warned = INITIAL_JIFFIES - 60*HZ; - - if (regs->sp >= curbase && regs->sp <= curbase + THREAD_SIZE && - regs->sp < curbase + sizeof(struct thread_info) + - sizeof(struct pt_regs) + 128 && - time_after(jiffies, warned + 60*HZ)) { - printk("do_IRQ: %s near stack overflow (cur:%Lx,sp:%lx)\n", - current->comm, curbase, regs->sp); - show_stack(NULL,NULL); - warned = jiffies; - } -} + + WARN_ONCE(regs->sp >= curbase && + regs->sp <= curbase + THREAD_SIZE && + regs->sp < curbase + sizeof(struct thread_info) + + sizeof(struct pt_regs) + 128, + + "do_IRQ: %s near stack overflow (cur:%Lx,sp:%lx)\n", + current->comm, curbase, regs->sp); #endif +} /* * do_IRQ handles all normal device IRQ's (the special @@ -61,9 +58,7 @@ asmlinkage unsigned int do_IRQ(struct pt_regs *regs) irq_enter(); irq = __get_cpu_var(vector_irq)[vector]; -#ifdef CONFIG_DEBUG_STACKOVERFLOW stack_overflow_check(regs); -#endif desc = irq_to_desc(irq); if (likely(desc)) -- cgit v1.2.3 From f201ae2356c74bcae130b2177b3dca903ea98071 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 23 Nov 2008 06:22:56 +0100 Subject: tracing/function-return-tracer: store return stack into task_struct and allocate it dynamically Impact: use deeper function tracing depth safely Some tests showed that function return tracing needed a more deeper depth of function calls. But it could be unsafe to store these return addresses to the stack. So these arrays will now be allocated dynamically into task_struct of current only when the tracer is activated. Typical scheme when tracer is activated: - allocate a return stack for each task in global list. - fork: allocate the return stack for the newly created task - exit: free return stack of current - idle init: same as fork I chose a default depth of 50. I don't have overruns anymore. Signed-off-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- arch/x86/include/asm/ftrace.h | 1 - arch/x86/include/asm/thread_info.h | 29 ------------ arch/x86/kernel/ftrace.c | 29 ++++++------ include/linux/ftrace.h | 5 ++ include/linux/sched.h | 23 +++++---- kernel/exit.c | 5 +- kernel/fork.c | 4 ++ kernel/sched.c | 3 ++ kernel/trace/ftrace.c | 96 +++++++++++++++++++++++++++++++++++++- 9 files changed, 137 insertions(+), 58 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/ftrace.h b/arch/x86/include/asm/ftrace.h index 2bb43b433e07..754a3e082f94 100644 --- a/arch/x86/include/asm/ftrace.h +++ b/arch/x86/include/asm/ftrace.h @@ -29,7 +29,6 @@ struct dyn_arch_ftrace { #endif /* CONFIG_FUNCTION_TRACER */ #ifdef CONFIG_FUNCTION_RET_TRACER -#define FTRACE_RET_STACK_SIZE 20 #ifndef __ASSEMBLY__ diff --git a/arch/x86/include/asm/thread_info.h b/arch/x86/include/asm/thread_info.h index e90e81ef6ab9..0921b4018c11 100644 --- a/arch/x86/include/asm/thread_info.h +++ b/arch/x86/include/asm/thread_info.h @@ -40,36 +40,8 @@ struct thread_info { */ __u8 supervisor_stack[0]; #endif - -#ifdef CONFIG_FUNCTION_RET_TRACER - /* Index of current stored adress in ret_stack */ - int curr_ret_stack; - /* Stack of return addresses for return function tracing */ - struct ftrace_ret_stack ret_stack[FTRACE_RET_STACK_SIZE]; - /* - * Number of functions that haven't been traced - * because of depth overrun. - */ - atomic_t trace_overrun; -#endif }; -#ifdef CONFIG_FUNCTION_RET_TRACER -#define INIT_THREAD_INFO(tsk) \ -{ \ - .task = &tsk, \ - .exec_domain = &default_exec_domain, \ - .flags = 0, \ - .cpu = 0, \ - .preempt_count = 1, \ - .addr_limit = KERNEL_DS, \ - .restart_block = { \ - .fn = do_no_restart_syscall, \ - }, \ - .curr_ret_stack = -1,\ - .trace_overrun = ATOMIC_INIT(0) \ -} -#else #define INIT_THREAD_INFO(tsk) \ { \ .task = &tsk, \ @@ -82,7 +54,6 @@ struct thread_info { .fn = do_no_restart_syscall, \ }, \ } -#endif #define init_thread_info (init_thread_union.thread_info) #define init_stack (init_thread_union.stack) diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 356bb1eb6e9a..bb137f7297ed 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -350,19 +350,21 @@ static int push_return_trace(unsigned long ret, unsigned long long time, unsigned long func) { int index; - struct thread_info *ti = current_thread_info(); + + if (!current->ret_stack) + return -EBUSY; /* The return trace stack is full */ - if (ti->curr_ret_stack == FTRACE_RET_STACK_SIZE - 1) { - atomic_inc(&ti->trace_overrun); + if (current->curr_ret_stack == FTRACE_RETFUNC_DEPTH - 1) { + atomic_inc(¤t->trace_overrun); return -EBUSY; } - index = ++ti->curr_ret_stack; + index = ++current->curr_ret_stack; barrier(); - ti->ret_stack[index].ret = ret; - ti->ret_stack[index].func = func; - ti->ret_stack[index].calltime = time; + current->ret_stack[index].ret = ret; + current->ret_stack[index].func = func; + current->ret_stack[index].calltime = time; return 0; } @@ -373,13 +375,12 @@ static void pop_return_trace(unsigned long *ret, unsigned long long *time, { int index; - struct thread_info *ti = current_thread_info(); - index = ti->curr_ret_stack; - *ret = ti->ret_stack[index].ret; - *func = ti->ret_stack[index].func; - *time = ti->ret_stack[index].calltime; - *overrun = atomic_read(&ti->trace_overrun); - ti->curr_ret_stack--; + index = current->curr_ret_stack; + *ret = current->ret_stack[index].ret; + *func = current->ret_stack[index].func; + *time = current->ret_stack[index].calltime; + *overrun = atomic_read(¤t->trace_overrun); + current->curr_ret_stack--; } /* diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index f7ba4ea5e128..2ba259b2defa 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -323,6 +323,8 @@ struct ftrace_retfunc { }; #ifdef CONFIG_FUNCTION_RET_TRACER +#define FTRACE_RETFUNC_DEPTH 50 +#define FTRACE_RETSTACK_ALLOC_SIZE 32 /* Type of a callback handler of tracing return function */ typedef void (*trace_function_return_t)(struct ftrace_retfunc *); @@ -330,6 +332,9 @@ extern int register_ftrace_return(trace_function_return_t func); /* The current handler in use */ extern trace_function_return_t ftrace_function_return; extern void unregister_ftrace_return(void); + +extern void ftrace_retfunc_init_task(struct task_struct *t); +extern void ftrace_retfunc_exit_task(struct task_struct *t); #endif #endif /* _LINUX_FTRACE_H */ diff --git a/include/linux/sched.h b/include/linux/sched.h index c8e0db464206..bee1e93c95ad 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1352,6 +1352,17 @@ struct task_struct { unsigned long default_timer_slack_ns; struct list_head *scm_work_list; +#ifdef CONFIG_FUNCTION_RET_TRACER + /* Index of current stored adress in ret_stack */ + int curr_ret_stack; + /* Stack of return addresses for return function tracing */ + struct ftrace_ret_stack *ret_stack; + /* + * Number of functions that haven't been traced + * because of depth overrun. + */ + atomic_t trace_overrun; +#endif }; /* @@ -2006,18 +2017,6 @@ static inline void setup_thread_stack(struct task_struct *p, struct task_struct { *task_thread_info(p) = *task_thread_info(org); task_thread_info(p)->task = p; - -#ifdef CONFIG_FUNCTION_RET_TRACER - /* - * When fork() creates a child process, this function is called. - * But the child task may not inherit the return adresses traced - * by the return function tracer because it will directly execute - * in userspace and will not return to kernel functions its parent - * used. - */ - task_thread_info(p)->curr_ret_stack = -1; - atomic_set(&task_thread_info(p)->trace_overrun, 0); -#endif } static inline unsigned long *end_of_stack(struct task_struct *p) diff --git a/kernel/exit.c b/kernel/exit.c index 35c8ec2ba03a..b9d446329da1 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -47,6 +47,7 @@ #include #include #include +#include #include #include @@ -1127,7 +1128,9 @@ NORET_TYPE void do_exit(long code) preempt_disable(); /* causes final put_task_struct in finish_task_switch(). */ tsk->state = TASK_DEAD; - +#ifdef CONFIG_FUNCTION_RET_TRACER + ftrace_retfunc_exit_task(tsk); +#endif schedule(); BUG(); /* Avoid "noreturn function does return". */ diff --git a/kernel/fork.c b/kernel/fork.c index ac62f43ee430..d1eb30e69ccc 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -47,6 +47,7 @@ #include #include #include +#include #include #include #include @@ -1269,6 +1270,9 @@ static struct task_struct *copy_process(unsigned long clone_flags, total_forks++; spin_unlock(¤t->sighand->siglock); write_unlock_irq(&tasklist_lock); +#ifdef CONFIG_FUNCTION_RET_TRACER + ftrace_retfunc_init_task(p); +#endif proc_fork_connector(p); cgroup_post_fork(p); return p; diff --git a/kernel/sched.c b/kernel/sched.c index 4de56108c86f..fb17205950de 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -5901,6 +5901,9 @@ void __cpuinit init_idle(struct task_struct *idle, int cpu) * The idle tasks have their own, simple scheduling class: */ idle->sched_class = &idle_sched_class; +#ifdef CONFIG_FUNCTION_RET_TRACER + ftrace_retfunc_init_task(idle); +#endif } /* diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index f212da486689..90d99fb02ae4 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -1498,10 +1498,77 @@ ftrace_enable_sysctl(struct ctl_table *table, int write, #ifdef CONFIG_FUNCTION_RET_TRACER +static atomic_t ftrace_retfunc_active; + /* The callback that hooks the return of a function */ trace_function_return_t ftrace_function_return = (trace_function_return_t)ftrace_stub; + +/* Try to assign a return stack array on FTRACE_RETSTACK_ALLOC_SIZE tasks. */ +static int alloc_retstack_tasklist(struct ftrace_ret_stack **ret_stack_list) +{ + int i; + int ret = 0; + unsigned long flags; + int start = 0, end = FTRACE_RETSTACK_ALLOC_SIZE; + struct task_struct *g, *t; + + for (i = 0; i < FTRACE_RETSTACK_ALLOC_SIZE; i++) { + ret_stack_list[i] = kmalloc(FTRACE_RETFUNC_DEPTH + * sizeof(struct ftrace_ret_stack), + GFP_KERNEL); + if (!ret_stack_list[i]) { + start = 0; + end = i; + ret = -ENOMEM; + goto free; + } + } + + read_lock_irqsave(&tasklist_lock, flags); + do_each_thread(g, t) { + if (start == end) { + ret = -EAGAIN; + goto unlock; + } + + if (t->ret_stack == NULL) { + t->ret_stack = ret_stack_list[start++]; + t->curr_ret_stack = -1; + atomic_set(&t->trace_overrun, 0); + } + } while_each_thread(g, t); + +unlock: + read_unlock_irqrestore(&tasklist_lock, flags); +free: + for (i = start; i < end; i++) + kfree(ret_stack_list[i]); + return ret; +} + +/* Allocate a return stack for each task */ +static int start_return_tracing(void) +{ + struct ftrace_ret_stack **ret_stack_list; + int ret; + + ret_stack_list = kmalloc(FTRACE_RETSTACK_ALLOC_SIZE * + sizeof(struct ftrace_ret_stack *), + GFP_KERNEL); + + if (!ret_stack_list) + return -ENOMEM; + + do { + ret = alloc_retstack_tasklist(ret_stack_list); + } while (ret == -EAGAIN); + + kfree(ret_stack_list); + return ret; +} + int register_ftrace_return(trace_function_return_t func) { int ret = 0; @@ -1516,7 +1583,12 @@ int register_ftrace_return(trace_function_return_t func) ret = -EBUSY; goto out; } - + atomic_inc(&ftrace_retfunc_active); + ret = start_return_tracing(); + if (ret) { + atomic_dec(&ftrace_retfunc_active); + goto out; + } ftrace_tracing_type = FTRACE_TYPE_RETURN; ftrace_function_return = func; ftrace_startup(); @@ -1530,6 +1602,7 @@ void unregister_ftrace_return(void) { mutex_lock(&ftrace_sysctl_lock); + atomic_dec(&ftrace_retfunc_active); ftrace_function_return = (trace_function_return_t)ftrace_stub; ftrace_shutdown(); /* Restore normal tracing type */ @@ -1537,6 +1610,27 @@ void unregister_ftrace_return(void) mutex_unlock(&ftrace_sysctl_lock); } + +/* Allocate a return stack for newly created task */ +void ftrace_retfunc_init_task(struct task_struct *t) +{ + if (atomic_read(&ftrace_retfunc_active)) { + t->ret_stack = kmalloc(FTRACE_RETFUNC_DEPTH + * sizeof(struct ftrace_ret_stack), + GFP_KERNEL); + if (!t->ret_stack) + return; + t->curr_ret_stack = -1; + atomic_set(&t->trace_overrun, 0); + } else + t->ret_stack = NULL; +} + +void ftrace_retfunc_exit_task(struct task_struct *t) +{ + kfree(t->ret_stack); + t->ret_stack = NULL; +} #endif -- cgit v1.2.3 From 02b67518e2b1c490787dac7f35e1204e74fe21ba Mon Sep 17 00:00:00 2001 From: Török Edwin Date: Sat, 22 Nov 2008 13:28:47 +0200 Subject: tracing: add support for userspace stacktraces in tracing/iter_ctrl MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Impact: add new (default-off) tracing visualization feature Usage example: mount -t debugfs nodev /sys/kernel/debug cd /sys/kernel/debug/tracing echo userstacktrace >iter_ctrl echo sched_switch >current_tracer echo 1 >tracing_enabled .... run application ... echo 0 >tracing_enabled Then read one of 'trace','latency_trace','trace_pipe'. To get the best output you can compile your userspace programs with frame pointers (at least glibc + the app you are tracing). Signed-off-by: Török Edwin Signed-off-by: Ingo Molnar --- Documentation/ftrace.txt | 5 ++- arch/x86/kernel/stacktrace.c | 57 +++++++++++++++++++++++++++ include/linux/stacktrace.h | 8 ++++ kernel/trace/trace.c | 93 ++++++++++++++++++++++++++++++++++++++++++++ kernel/trace/trace.h | 9 +++++ 5 files changed, 171 insertions(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/Documentation/ftrace.txt b/Documentation/ftrace.txt index 753f4de4b175..79a80f79c062 100644 --- a/Documentation/ftrace.txt +++ b/Documentation/ftrace.txt @@ -324,7 +324,7 @@ output. To see what is available, simply cat the file: cat /debug/tracing/trace_options print-parent nosym-offset nosym-addr noverbose noraw nohex nobin \ - noblock nostacktrace nosched-tree + noblock nostacktrace nosched-tree nouserstacktrace To disable one of the options, echo in the option prepended with "no". @@ -378,6 +378,9 @@ Here are the available options: When a trace is recorded, so is the stack of functions. This allows for back traces of trace sites. + userstacktrace - This option changes the trace. + It records a stacktrace of the current userspace thread. + sched-tree - TBD (any users??) diff --git a/arch/x86/kernel/stacktrace.c b/arch/x86/kernel/stacktrace.c index a03e7f6d90c3..b15153060417 100644 --- a/arch/x86/kernel/stacktrace.c +++ b/arch/x86/kernel/stacktrace.c @@ -6,6 +6,7 @@ #include #include #include +#include #include static void save_stack_warning(void *data, char *msg) @@ -83,3 +84,59 @@ void save_stack_trace_tsk(struct task_struct *tsk, struct stack_trace *trace) trace->entries[trace->nr_entries++] = ULONG_MAX; } EXPORT_SYMBOL_GPL(save_stack_trace_tsk); + +/* Userspace stacktrace - based on kernel/trace/trace_sysprof.c */ + +struct stack_frame { + const void __user *next_fp; + unsigned long return_address; +}; + +static int copy_stack_frame(const void __user *fp, struct stack_frame *frame) +{ + int ret; + + if (!access_ok(VERIFY_READ, fp, sizeof(*frame))) + return 0; + + ret = 1; + pagefault_disable(); + if (__copy_from_user_inatomic(frame, fp, sizeof(*frame))) + ret = 0; + pagefault_enable(); + + return ret; +} + +void save_stack_trace_user(struct stack_trace *trace) +{ + /* + * Trace user stack if we are not a kernel thread + */ + if (current->mm) { + const struct pt_regs *regs = task_pt_regs(current); + const void __user *fp = (const void __user *)regs->bp; + + if (trace->nr_entries < trace->max_entries) + trace->entries[trace->nr_entries++] = regs->ip; + + while (trace->nr_entries < trace->max_entries) { + struct stack_frame frame; + frame.next_fp = NULL; + frame.return_address = 0; + if (!copy_stack_frame(fp, &frame)) + break; + if ((unsigned long)fp < regs->sp) + break; + if (frame.return_address) + trace->entries[trace->nr_entries++] = + frame.return_address; + if (fp == frame.next_fp) + break; + fp = frame.next_fp; + } + } + if (trace->nr_entries < trace->max_entries) + trace->entries[trace->nr_entries++] = ULONG_MAX; +} + diff --git a/include/linux/stacktrace.h b/include/linux/stacktrace.h index b106fd8e0d5c..68de51468f5d 100644 --- a/include/linux/stacktrace.h +++ b/include/linux/stacktrace.h @@ -15,9 +15,17 @@ extern void save_stack_trace_tsk(struct task_struct *tsk, struct stack_trace *trace); extern void print_stack_trace(struct stack_trace *trace, int spaces); + +#ifdef CONFIG_X86 +extern void save_stack_trace_user(struct stack_trace *trace); +#else +# define save_stack_trace_user(trace) do { } while (0) +#endif + #else # define save_stack_trace(trace) do { } while (0) # define save_stack_trace_tsk(tsk, trace) do { } while (0) +# define save_stack_trace_user(trace) do { } while (0) # define print_stack_trace(trace, spaces) do { } while (0) #endif diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 4ee6f0375222..ced8b4fa9f51 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -275,6 +275,7 @@ static const char *trace_options[] = { "ftrace_preempt", "branch", "annotate", + "userstacktrace", NULL }; @@ -918,6 +919,44 @@ void __trace_stack(struct trace_array *tr, ftrace_trace_stack(tr, data, flags, skip, preempt_count()); } +static void ftrace_trace_userstack(struct trace_array *tr, + struct trace_array_cpu *data, + unsigned long flags, int pc) +{ + struct userstack_entry *entry; + struct stack_trace trace; + struct ring_buffer_event *event; + unsigned long irq_flags; + + if (!(trace_flags & TRACE_ITER_USERSTACKTRACE)) + return; + + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), + &irq_flags); + if (!event) + return; + entry = ring_buffer_event_data(event); + tracing_generic_entry_update(&entry->ent, flags, pc); + entry->ent.type = TRACE_USER_STACK; + + memset(&entry->caller, 0, sizeof(entry->caller)); + + trace.nr_entries = 0; + trace.max_entries = FTRACE_STACK_ENTRIES; + trace.skip = 0; + trace.entries = entry->caller; + + save_stack_trace_user(&trace); + ring_buffer_unlock_commit(tr->buffer, event, irq_flags); +} + +void __trace_userstack(struct trace_array *tr, + struct trace_array_cpu *data, + unsigned long flags) +{ + ftrace_trace_userstack(tr, data, flags, preempt_count()); +} + static void ftrace_trace_special(void *__tr, void *__data, unsigned long arg1, unsigned long arg2, unsigned long arg3, @@ -941,6 +980,7 @@ ftrace_trace_special(void *__tr, void *__data, entry->arg3 = arg3; ring_buffer_unlock_commit(tr->buffer, event, irq_flags); ftrace_trace_stack(tr, data, irq_flags, 4, pc); + ftrace_trace_userstack(tr, data, irq_flags, pc); trace_wake_up(); } @@ -979,6 +1019,7 @@ tracing_sched_switch_trace(struct trace_array *tr, entry->next_cpu = task_cpu(next); ring_buffer_unlock_commit(tr->buffer, event, irq_flags); ftrace_trace_stack(tr, data, flags, 5, pc); + ftrace_trace_userstack(tr, data, flags, pc); } void @@ -1008,6 +1049,7 @@ tracing_sched_wakeup_trace(struct trace_array *tr, entry->next_cpu = task_cpu(wakee); ring_buffer_unlock_commit(tr->buffer, event, irq_flags); ftrace_trace_stack(tr, data, flags, 6, pc); + ftrace_trace_userstack(tr, data, flags, pc); trace_wake_up(); } @@ -1387,6 +1429,31 @@ seq_print_ip_sym(struct trace_seq *s, unsigned long ip, unsigned long sym_flags) return ret; } +static int +seq_print_userip_objs(const struct userstack_entry *entry, struct trace_seq *s, + unsigned long sym_flags) +{ + int ret = 1; + unsigned i; + + for (i = 0; i < FTRACE_STACK_ENTRIES; i++) { + unsigned long ip = entry->caller[i]; + + if (ip == ULONG_MAX || !ret) + break; + if (i) + ret = trace_seq_puts(s, " <- "); + if (!ip) { + ret = trace_seq_puts(s, "??"); + continue; + } + if (ret /*&& (sym_flags & TRACE_ITER_SYM_ADDR)*/) + ret = trace_seq_printf(s, " <" IP_FMT ">", ip); + } + + return ret; +} + static void print_lat_help_header(struct seq_file *m) { seq_puts(m, "# _------=> CPU# \n"); @@ -1702,6 +1769,16 @@ print_lat_fmt(struct trace_iterator *iter, unsigned int trace_idx, int cpu) field->line); break; } + case TRACE_USER_STACK: { + struct userstack_entry *field; + + trace_assign_type(field, entry); + + seq_print_userip_objs(field, s, sym_flags); + if (entry->flags & TRACE_FLAG_CONT) + trace_seq_print_cont(s, iter); + break; + } default: trace_seq_printf(s, "Unknown type %d\n", entry->type); } @@ -1853,6 +1930,19 @@ static enum print_line_t print_trace_fmt(struct trace_iterator *iter) field->line); break; } + case TRACE_USER_STACK: { + struct userstack_entry *field; + + trace_assign_type(field, entry); + + ret = seq_print_userip_objs(field, s, sym_flags); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + ret = trace_seq_putc(s, '\n'); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + break; + } } return TRACE_TYPE_HANDLED; } @@ -1912,6 +2002,7 @@ static enum print_line_t print_raw_fmt(struct trace_iterator *iter) break; } case TRACE_SPECIAL: + case TRACE_USER_STACK: case TRACE_STACK: { struct special_entry *field; @@ -2000,6 +2091,7 @@ static enum print_line_t print_hex_fmt(struct trace_iterator *iter) break; } case TRACE_SPECIAL: + case TRACE_USER_STACK: case TRACE_STACK: { struct special_entry *field; @@ -2054,6 +2146,7 @@ static enum print_line_t print_bin_fmt(struct trace_iterator *iter) break; } case TRACE_SPECIAL: + case TRACE_USER_STACK: case TRACE_STACK: { struct special_entry *field; diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 2cb12fd98f6b..17bb4c830b01 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -26,6 +26,7 @@ enum trace_type { TRACE_BOOT_CALL, TRACE_BOOT_RET, TRACE_FN_RET, + TRACE_USER_STACK, __TRACE_LAST_TYPE }; @@ -42,6 +43,7 @@ struct trace_entry { unsigned char flags; unsigned char preempt_count; int pid; + int tgid; }; /* @@ -99,6 +101,11 @@ struct stack_entry { unsigned long caller[FTRACE_STACK_ENTRIES]; }; +struct userstack_entry { + struct trace_entry ent; + unsigned long caller[FTRACE_STACK_ENTRIES]; +}; + /* * ftrace_printk entry: */ @@ -240,6 +247,7 @@ extern void __ftrace_bad_type(void); IF_ASSIGN(var, ent, struct ctx_switch_entry, 0); \ IF_ASSIGN(var, ent, struct trace_field_cont, TRACE_CONT); \ IF_ASSIGN(var, ent, struct stack_entry, TRACE_STACK); \ + IF_ASSIGN(var, ent, struct userstack_entry, TRACE_USER_STACK);\ IF_ASSIGN(var, ent, struct print_entry, TRACE_PRINT); \ IF_ASSIGN(var, ent, struct special_entry, 0); \ IF_ASSIGN(var, ent, struct trace_mmiotrace_rw, \ @@ -500,6 +508,7 @@ enum trace_iterator_flags { TRACE_ITER_PREEMPTONLY = 0x800, TRACE_ITER_BRANCH = 0x1000, TRACE_ITER_ANNOTATE = 0x2000, + TRACE_ITER_USERSTACKTRACE = 0x4000 }; /* -- cgit v1.2.3 From 5c9b3a0c7b8be3cdef3d7418f0a49127e7cdc998 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Fri, 21 Nov 2008 17:36:41 -0800 Subject: x86: signal: cosmetic unification of including headers Impact: cleanup Make the headers portion of signal_32.c and signal_64.c the same. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/kernel/signal_32.c | 27 +++++++++++++++------------ arch/x86/kernel/signal_64.c | 7 ++++++- 2 files changed, 21 insertions(+), 13 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/signal_32.c b/arch/x86/kernel/signal_32.c index c2aabeba27a5..0ff8d8750a7d 100644 --- a/arch/x86/kernel/signal_32.c +++ b/arch/x86/kernel/signal_32.c @@ -4,29 +4,32 @@ * 1997-11-28 Modified for POSIX.1b signals by Richard Henderson * 2000-06-20 Pentium III FXSR, SSE support by Gareth Hughes */ -#include -#include -#include -#include +#include +#include +#include #include -#include #include -#include -#include #include -#include #include +#include #include -#include -#include -#include +#include +#include +#include +#include #include #include -#include #include #include + +#ifdef CONFIG_X86_64 +#include +#include +#include +#endif /* CONFIG_X86_64 */ + #include #include diff --git a/arch/x86/kernel/signal_64.c b/arch/x86/kernel/signal_64.c index 3d54d366ccb2..c52244ac19fc 100644 --- a/arch/x86/kernel/signal_64.c +++ b/arch/x86/kernel/signal_64.c @@ -19,17 +19,22 @@ #include #include #include -#include #include #include #include #include +#include + +#ifdef CONFIG_X86_64 #include #include #include +#endif /* CONFIG_X86_64 */ + #include #include + #include "sigframe.h" #define _BLOCKABLE (~(sigmask(SIGKILL) | sigmask(SIGSTOP))) -- cgit v1.2.3 From 666ac7be049ec290625e65d5922ff59f7bdec527 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Fri, 21 Nov 2008 17:38:25 -0800 Subject: x86: signal: cosmetic unification of sys_sigaltstack() Impact: cleanup Add #ifdef directive for unification. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/kernel/signal_32.c | 9 +++++++++ arch/x86/kernel/signal_64.c | 15 +++++++++++++++ 2 files changed, 24 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/signal_32.c b/arch/x86/kernel/signal_32.c index 0ff8d8750a7d..d9909881ac66 100644 --- a/arch/x86/kernel/signal_32.c +++ b/arch/x86/kernel/signal_32.c @@ -125,6 +125,7 @@ sys_sigaction(int sig, const struct old_sigaction __user *act, return ret; } +#ifdef CONFIG_X86_32 asmlinkage int sys_sigaltstack(unsigned long bx) { /* @@ -137,6 +138,14 @@ asmlinkage int sys_sigaltstack(unsigned long bx) return do_sigaltstack(uss, uoss, regs->sp); } +#else /* !CONFIG_X86_32 */ +asmlinkage long +sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss, + struct pt_regs *regs) +{ + return do_sigaltstack(uss, uoss, regs->sp); +} +#endif /* CONFIG_X86_32 */ #define COPY(x) { \ err |= __get_user(regs->x, &sc->x); \ diff --git a/arch/x86/kernel/signal_64.c b/arch/x86/kernel/signal_64.c index c52244ac19fc..b6e4fe03a36b 100644 --- a/arch/x86/kernel/signal_64.c +++ b/arch/x86/kernel/signal_64.c @@ -50,12 +50,27 @@ # define FIX_EFLAGS __FIX_EFLAGS #endif +#ifdef CONFIG_X86_32 +asmlinkage int sys_sigaltstack(unsigned long bx) +{ + /* + * This is needed to make gcc realize it doesn't own the + * "struct pt_regs" + */ + struct pt_regs *regs = (struct pt_regs *)&bx; + const stack_t __user *uss = (const stack_t __user *)bx; + stack_t __user *uoss = (stack_t __user *)regs->cx; + + return do_sigaltstack(uss, uoss, regs->sp); +} +#else /* !CONFIG_X86_32 */ asmlinkage long sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss, struct pt_regs *regs) { return do_sigaltstack(uss, uoss, regs->sp); } +#endif /* CONFIG_X86_32 */ #define COPY(x) { \ err |= __get_user(regs->x, &sc->x); \ -- cgit v1.2.3 From 2456d738ef051f85170bf018faef63f83fa84eb5 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Fri, 21 Nov 2008 17:38:57 -0800 Subject: x86: signal: cosmetic unification of sys_rt_sigreturn() Impact: cleanup Add #ifdef directive for unification. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/kernel/signal_32.c | 7 +++++++ arch/x86/kernel/signal_64.c | 9 +++++++++ 2 files changed, 16 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/signal_32.c b/arch/x86/kernel/signal_32.c index d9909881ac66..f7dd6c44c042 100644 --- a/arch/x86/kernel/signal_32.c +++ b/arch/x86/kernel/signal_32.c @@ -299,12 +299,19 @@ badframe: return 0; } +#ifdef CONFIG_X86_32 asmlinkage int sys_rt_sigreturn(unsigned long __unused) { struct pt_regs *regs = (struct pt_regs *)&__unused; return do_rt_sigreturn(regs); } +#else /* !CONFIG_X86_32 */ +asmlinkage long sys_rt_sigreturn(struct pt_regs *regs) +{ + return do_rt_sigreturn(regs); +} +#endif /* CONFIG_X86_32 */ /* * Set up a signal frame. diff --git a/arch/x86/kernel/signal_64.c b/arch/x86/kernel/signal_64.c index b6e4fe03a36b..32718f5e4f61 100644 --- a/arch/x86/kernel/signal_64.c +++ b/arch/x86/kernel/signal_64.c @@ -169,10 +169,19 @@ badframe: return 0; } +#ifdef CONFIG_X86_32 +asmlinkage int sys_rt_sigreturn(unsigned long __unused) +{ + struct pt_regs *regs = (struct pt_regs *)&__unused; + + return do_rt_sigreturn(regs); +} +#else /* !CONFIG_X86_32 */ asmlinkage long sys_rt_sigreturn(struct pt_regs *regs) { return do_rt_sigreturn(regs); } +#endif /* CONFIG_X86_32 */ /* * Set up a signal frame. -- cgit v1.2.3 From c450d7805b2c5cac8846c5f490fddfd9030d2207 Mon Sep 17 00:00:00 2001 From: Hannes Eder Date: Fri, 21 Nov 2008 23:17:09 +0100 Subject: x86: vmware - fix sparse warnings Impact: fix sparse build warning Fix the following sparse warnings: arch/x86/kernel/cpu/vmware.c:69:5: warning: symbol 'vmware_platform' was not declared. Should it be static? arch/x86/kernel/cpu/vmware.c:89:15: warning: symbol 'vmware_get_tsc_khz' was not declared. Should it be static? arch/x86/kernel/cpu/vmware.c:107:16: warning: symbol 'vmware_set_feature_bits' was not declared. Should it be static? Signed-off-by: Hannes Eder Cc: "Alok N Kataria" Cc: "Dan Hecht" Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/vmware.c | 1 + 1 file changed, 1 insertion(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/vmware.c b/arch/x86/kernel/cpu/vmware.c index c034bda842d9..284c399e3234 100644 --- a/arch/x86/kernel/cpu/vmware.c +++ b/arch/x86/kernel/cpu/vmware.c @@ -23,6 +23,7 @@ #include #include +#include #define CPUID_VMWARE_INFO_LEAF 0x40000000 #define VMWARE_HYPERVISOR_MAGIC 0x564D5868 -- cgit v1.2.3 From 4e42ebd57b2e727b28bf5f6068e95cd19b0e807b Mon Sep 17 00:00:00 2001 From: Hannes Eder Date: Fri, 21 Nov 2008 22:56:17 +0100 Subject: x86: hypervisor - fix sparse warnings Impact: fix sparse build warning Fix the following sparse warnings: arch/x86/kernel/cpu/hypervisor.c:37:15: warning: symbol 'get_hypervisor_tsc_freq' was not declared. Should it be static? arch/x86/kernel/cpu/hypervisor.c:53:16: warning: symbol 'init_hypervisor' was not declared. Should it be static? Signed-off-by: Hannes Eder Cc: "Alok N Kataria" Cc: "Dan Hecht" Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/hypervisor.c | 1 + 1 file changed, 1 insertion(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/hypervisor.c b/arch/x86/kernel/cpu/hypervisor.c index 35ae2b75226d..fb5b86af0b01 100644 --- a/arch/x86/kernel/cpu/hypervisor.c +++ b/arch/x86/kernel/cpu/hypervisor.c @@ -23,6 +23,7 @@ #include #include +#include static inline void __cpuinit detect_hypervisor_vendor(struct cpuinfo_x86 *c) -- cgit v1.2.3 From 8d7c6a96164651dbbab449ef0b5c20ae1f76a3a1 Mon Sep 17 00:00:00 2001 From: Török Edwin Date: Sun, 23 Nov 2008 12:39:06 +0200 Subject: tracing/stack-tracer: fix style issues MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Impact: cleanup Signed-off-by: Török Edwin Signed-off-by: Ingo Molnar --- arch/x86/kernel/stacktrace.c | 51 +++++++++++++++++++++++++------------------- include/linux/stacktrace.h | 2 +- kernel/trace/trace.c | 7 +++--- 3 files changed, 33 insertions(+), 27 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/stacktrace.c b/arch/x86/kernel/stacktrace.c index b15153060417..10786af95545 100644 --- a/arch/x86/kernel/stacktrace.c +++ b/arch/x86/kernel/stacktrace.c @@ -89,7 +89,7 @@ EXPORT_SYMBOL_GPL(save_stack_trace_tsk); struct stack_frame { const void __user *next_fp; - unsigned long return_address; + unsigned long ret_addr; }; static int copy_stack_frame(const void __user *fp, struct stack_frame *frame) @@ -108,33 +108,40 @@ static int copy_stack_frame(const void __user *fp, struct stack_frame *frame) return ret; } +static inline void __save_stack_trace_user(struct stack_trace *trace) +{ + const struct pt_regs *regs = task_pt_regs(current); + const void __user *fp = (const void __user *)regs->bp; + + if (trace->nr_entries < trace->max_entries) + trace->entries[trace->nr_entries++] = regs->ip; + + while (trace->nr_entries < trace->max_entries) { + struct stack_frame frame; + + frame.next_fp = NULL; + frame.ret_addr = 0; + if (!copy_stack_frame(fp, &frame)) + break; + if ((unsigned long)fp < regs->sp) + break; + if (frame.ret_addr) { + trace->entries[trace->nr_entries++] = + frame.ret_addr; + } + if (fp == frame.next_fp) + break; + fp = frame.next_fp; + } +} + void save_stack_trace_user(struct stack_trace *trace) { /* * Trace user stack if we are not a kernel thread */ if (current->mm) { - const struct pt_regs *regs = task_pt_regs(current); - const void __user *fp = (const void __user *)regs->bp; - - if (trace->nr_entries < trace->max_entries) - trace->entries[trace->nr_entries++] = regs->ip; - - while (trace->nr_entries < trace->max_entries) { - struct stack_frame frame; - frame.next_fp = NULL; - frame.return_address = 0; - if (!copy_stack_frame(fp, &frame)) - break; - if ((unsigned long)fp < regs->sp) - break; - if (frame.return_address) - trace->entries[trace->nr_entries++] = - frame.return_address; - if (fp == frame.next_fp) - break; - fp = frame.next_fp; - } + __save_stack_trace_user(trace); } if (trace->nr_entries < trace->max_entries) trace->entries[trace->nr_entries++] = ULONG_MAX; diff --git a/include/linux/stacktrace.h b/include/linux/stacktrace.h index 68de51468f5d..fd42d6851109 100644 --- a/include/linux/stacktrace.h +++ b/include/linux/stacktrace.h @@ -25,7 +25,7 @@ extern void save_stack_trace_user(struct stack_trace *trace); #else # define save_stack_trace(trace) do { } while (0) # define save_stack_trace_tsk(tsk, trace) do { } while (0) -# define save_stack_trace_user(trace) do { } while (0) +# define save_stack_trace_user(trace) do { } while (0) # define print_stack_trace(trace, spaces) do { } while (0) #endif diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 62776b71b1c5..dedf35f36971 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -948,9 +948,9 @@ static void ftrace_trace_userstack(struct trace_array *tr, struct trace_array_cpu *data, unsigned long flags, int pc) { + struct ring_buffer_event *event; struct userstack_entry *entry; struct stack_trace trace; - struct ring_buffer_event *event; unsigned long irq_flags; if (!(trace_flags & TRACE_ITER_USERSTACKTRACE)) @@ -1471,8 +1471,7 @@ static inline int seq_print_user_ip(struct trace_seq *s, struct mm_struct *mm, if (file) { ret = trace_seq_path(s, &file->f_path); if (ret) - ret = trace_seq_printf(s, "[+0x%lx]", - ip - vmstart); + ret = trace_seq_printf(s, "[+0x%lx]", ip - vmstart); } if (ret && ((sym_flags & TRACE_ITER_SYM_ADDR) || !file)) ret = trace_seq_printf(s, " <" IP_FMT ">", ip); @@ -1485,7 +1484,7 @@ seq_print_userip_objs(const struct userstack_entry *entry, struct trace_seq *s, { struct mm_struct *mm = NULL; int ret = 1; - unsigned i; + unsigned int i; if (trace_flags & TRACE_ITER_SYM_USEROBJ) { struct task_struct *task; -- cgit v1.2.3 From 8d26487fd4ddda7a0237da418fb8669fb06ae557 Mon Sep 17 00:00:00 2001 From: Török Edwin Date: Sun, 23 Nov 2008 12:39:08 +0200 Subject: tracing/stack-tracer: introduce CONFIG_USER_STACKTRACE_SUPPORT MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Impact: cleanup User stack tracing is just implemented for x86, but it is not x86 specific. Introduce a generic config flag, that is currently enabled only for x86. When other arches implement it, they will have to SELECT USER_STACKTRACE_SUPPORT. Signed-off-by: Török Edwin Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 1 + include/linux/stacktrace.h | 2 +- kernel/trace/Kconfig | 3 +++ 3 files changed, 5 insertions(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 7a146baaa990..e49a4fd718fe 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -36,6 +36,7 @@ config X86 select HAVE_ARCH_TRACEHOOK select HAVE_GENERIC_DMA_COHERENT if X86_32 select HAVE_EFFICIENT_UNALIGNED_ACCESS + select USER_STACKTRACE_SUPPORT config ARCH_DEFCONFIG string diff --git a/include/linux/stacktrace.h b/include/linux/stacktrace.h index fd42d6851109..1a8cecc4f38c 100644 --- a/include/linux/stacktrace.h +++ b/include/linux/stacktrace.h @@ -16,7 +16,7 @@ extern void save_stack_trace_tsk(struct task_struct *tsk, extern void print_stack_trace(struct stack_trace *trace, int spaces); -#ifdef CONFIG_X86 +#ifdef CONFIG_USER_STACKTRACE_SUPPORT extern void save_stack_trace_user(struct stack_trace *trace); #else # define save_stack_trace_user(trace) do { } while (0) diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index b8378fad29a3..87fc34a1bb91 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -3,6 +3,9 @@ # select HAVE_FUNCTION_TRACER: # +config USER_STACKTRACE_SUPPORT + bool + config NOP_TRACER bool -- cgit v1.2.3 From 8a2503fa4a6fae8ee42140b339f37373fc6acaae Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Sun, 23 Nov 2008 14:53:43 +0300 Subject: x86: move dwarf2 related macro to dwarf2.h Impact: cleanup Move recently introduced dwarf2 macros to dwarf2.h file. It allow us to not duplicate them in assembly files. Active usage of _cfi macros don't make assembly files more obvious to understand but we already have a lot of macros there which requires to search the definitions of them *anyway*. But at least it make every cfi usage one line shorter. Also some code alignment is done. Signed-off-by: Cyrill Gorcunov Signed-off-by: Ingo Molnar --- arch/x86/include/asm/dwarf2.h | 97 +++++++++++++++++++++++++++++-------------- arch/x86/kernel/entry_64.S | 23 ---------- 2 files changed, 66 insertions(+), 54 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/dwarf2.h b/arch/x86/include/asm/dwarf2.h index 804b6e6be929..3afc5e87cfdd 100644 --- a/arch/x86/include/asm/dwarf2.h +++ b/arch/x86/include/asm/dwarf2.h @@ -6,56 +6,91 @@ #endif /* - Macros for dwarf2 CFI unwind table entries. - See "as.info" for details on these pseudo ops. Unfortunately - they are only supported in very new binutils, so define them - away for older version. + * Macros for dwarf2 CFI unwind table entries. + * See "as.info" for details on these pseudo ops. Unfortunately + * they are only supported in very new binutils, so define them + * away for older version. */ #ifdef CONFIG_AS_CFI -#define CFI_STARTPROC .cfi_startproc -#define CFI_ENDPROC .cfi_endproc -#define CFI_DEF_CFA .cfi_def_cfa -#define CFI_DEF_CFA_REGISTER .cfi_def_cfa_register -#define CFI_DEF_CFA_OFFSET .cfi_def_cfa_offset -#define CFI_ADJUST_CFA_OFFSET .cfi_adjust_cfa_offset -#define CFI_OFFSET .cfi_offset -#define CFI_REL_OFFSET .cfi_rel_offset -#define CFI_REGISTER .cfi_register -#define CFI_RESTORE .cfi_restore -#define CFI_REMEMBER_STATE .cfi_remember_state -#define CFI_RESTORE_STATE .cfi_restore_state -#define CFI_UNDEFINED .cfi_undefined +#define CFI_STARTPROC .cfi_startproc +#define CFI_ENDPROC .cfi_endproc +#define CFI_DEF_CFA .cfi_def_cfa +#define CFI_DEF_CFA_REGISTER .cfi_def_cfa_register +#define CFI_DEF_CFA_OFFSET .cfi_def_cfa_offset +#define CFI_ADJUST_CFA_OFFSET .cfi_adjust_cfa_offset +#define CFI_OFFSET .cfi_offset +#define CFI_REL_OFFSET .cfi_rel_offset +#define CFI_REGISTER .cfi_register +#define CFI_RESTORE .cfi_restore +#define CFI_REMEMBER_STATE .cfi_remember_state +#define CFI_RESTORE_STATE .cfi_restore_state +#define CFI_UNDEFINED .cfi_undefined #ifdef CONFIG_AS_CFI_SIGNAL_FRAME -#define CFI_SIGNAL_FRAME .cfi_signal_frame +#define CFI_SIGNAL_FRAME .cfi_signal_frame #else #define CFI_SIGNAL_FRAME #endif #else -/* Due to the structure of pre-exisiting code, don't use assembler line - comment character # to ignore the arguments. Instead, use a dummy macro. */ +/* + * Due to the structure of pre-exisiting code, don't use assembler line + * comment character # to ignore the arguments. Instead, use a dummy macro. + */ .macro cfi_ignore a=0, b=0, c=0, d=0 .endm -#define CFI_STARTPROC cfi_ignore -#define CFI_ENDPROC cfi_ignore -#define CFI_DEF_CFA cfi_ignore +#define CFI_STARTPROC cfi_ignore +#define CFI_ENDPROC cfi_ignore +#define CFI_DEF_CFA cfi_ignore #define CFI_DEF_CFA_REGISTER cfi_ignore #define CFI_DEF_CFA_OFFSET cfi_ignore #define CFI_ADJUST_CFA_OFFSET cfi_ignore -#define CFI_OFFSET cfi_ignore -#define CFI_REL_OFFSET cfi_ignore -#define CFI_REGISTER cfi_ignore -#define CFI_RESTORE cfi_ignore -#define CFI_REMEMBER_STATE cfi_ignore -#define CFI_RESTORE_STATE cfi_ignore -#define CFI_UNDEFINED cfi_ignore -#define CFI_SIGNAL_FRAME cfi_ignore +#define CFI_OFFSET cfi_ignore +#define CFI_REL_OFFSET cfi_ignore +#define CFI_REGISTER cfi_ignore +#define CFI_RESTORE cfi_ignore +#define CFI_REMEMBER_STATE cfi_ignore +#define CFI_RESTORE_STATE cfi_ignore +#define CFI_UNDEFINED cfi_ignore +#define CFI_SIGNAL_FRAME cfi_ignore #endif +/* + * An attempt to make CFI annotations more or less + * correct and shorter. It is implied that you know + * what you're doing if you use them. + */ +#ifdef __ASSEMBLY__ +#ifdef CONFIG_X86_64 + .macro pushq_cfi reg + pushq \reg + CFI_ADJUST_CFA_OFFSET 8 + .endm + + .macro popq_cfi reg + popq \reg + CFI_ADJUST_CFA_OFFSET -8 + .endm + + .macro movq_cfi reg offset=0 + movq %\reg, \offset(%rsp) + CFI_REL_OFFSET \reg, \offset + .endm + + .macro movq_cfi_restore offset reg + movq \offset(%rsp), %\reg + CFI_RESTORE \reg + .endm +#else /*!CONFIG_X86_64*/ + + /* 32bit defenitions are missed yet */ + +#endif /*!CONFIG_X86_64*/ +#endif /*__ASSEMBLY__*/ + #endif /* _ASM_X86_DWARF2_H */ diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index e5ddf573ded2..249eb604e71b 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -60,29 +60,6 @@ #define __AUDIT_ARCH_LE 0x40000000 .code64 -/* - * Some macro's to hide the most frequently occuring CFI annotations. - */ - .macro pushq_cfi reg - pushq \reg - CFI_ADJUST_CFA_OFFSET 8 - .endm - - .macro popq_cfi reg - popq \reg - CFI_ADJUST_CFA_OFFSET -8 - .endm - - .macro movq_cfi reg offset=0 - movq %\reg, \offset(%rsp) - CFI_REL_OFFSET \reg, \offset - .endm - - .macro movq_cfi_restore offset reg - movq \offset(%rsp), %\reg - CFI_RESTORE \reg - .endm - #ifdef CONFIG_FUNCTION_TRACER #ifdef CONFIG_DYNAMIC_FTRACE ENTRY(mcount) -- cgit v1.2.3 From 050dc6944b9ca2186f4729ab44e0da3743933941 Mon Sep 17 00:00:00 2001 From: Hannes Eder Date: Sun, 23 Nov 2008 13:35:48 +0100 Subject: x86: remove duplicate #define from 'cpufeature.h' Impact: cleanup Remove duplicate #define from 'cpufeature.h'. This also fixes the following sparse warning: arch/x86/kernel/cpu/capflags.c:54:3: warning: Initializer entry defined twice arch/x86/kernel/cpu/capflags.c:58:3: also defined here Signed-off-by: Hannes Eder Signed-off-by: Ingo Molnar --- arch/x86/include/asm/cpufeature.h | 1 - 1 file changed, 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/cpufeature.h b/arch/x86/include/asm/cpufeature.h index 694d1f8f1bee..5bce8ed02b44 100644 --- a/arch/x86/include/asm/cpufeature.h +++ b/arch/x86/include/asm/cpufeature.h @@ -80,7 +80,6 @@ #define X86_FEATURE_UP (3*32+ 9) /* smp kernel running on up */ #define X86_FEATURE_FXSAVE_LEAK (3*32+10) /* "" FXSAVE leaks FOP/FIP/FOP */ #define X86_FEATURE_ARCH_PERFMON (3*32+11) /* Intel Architectural PerfMon */ -#define X86_FEATURE_NOPL (3*32+20) /* The NOPL (0F 1F) instructions */ #define X86_FEATURE_PEBS (3*32+12) /* Precise-Event Based Sampling */ #define X86_FEATURE_BTS (3*32+13) /* Branch Trace Store */ #define X86_FEATURE_SYSCALL32 (3*32+14) /* "" syscall in ia32 userspace */ -- cgit v1.2.3 From 322648d1ba75280d62f114d47048beb0b35f5047 Mon Sep 17 00:00:00 2001 From: Alexander van Heukelum Date: Sun, 23 Nov 2008 10:08:28 +0100 Subject: x86: include ENTRY/END in entry handlers in entry_64.S Impact: cleanup of entry_64.S Except for the order and the place of the functions, this patch should not change the generated code. Signed-off-by: Alexander van Heukelum Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_64.S | 259 +++++++++++++++++++-------------------------- 1 file changed, 109 insertions(+), 150 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 249eb604e71b..1a856c0b21a8 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -922,76 +922,70 @@ END(common_interrupt) /* * APIC interrupts. */ - .p2align 5 - - .macro apicinterrupt num,func +.macro apicinterrupt num sym do_sym +ENTRY(\sym) INTR_FRAME pushq $~(\num) CFI_ADJUST_CFA_OFFSET 8 - interrupt \func + interrupt \do_sym jmp ret_from_intr CFI_ENDPROC - .endm +END(\sym) +.endm -ENTRY(thermal_interrupt) - apicinterrupt THERMAL_APIC_VECTOR,smp_thermal_interrupt -END(thermal_interrupt) +#ifdef CONFIG_SMP +apicinterrupt IRQ_MOVE_CLEANUP_VECTOR \ + irq_move_cleanup_interrupt smp_irq_move_cleanup_interrupt +#endif -ENTRY(threshold_interrupt) - apicinterrupt THRESHOLD_APIC_VECTOR,mce_threshold_interrupt -END(threshold_interrupt) +apicinterrupt 220 \ + uv_bau_message_intr1 uv_bau_message_interrupt +apicinterrupt LOCAL_TIMER_VECTOR \ + apic_timer_interrupt smp_apic_timer_interrupt #ifdef CONFIG_SMP -ENTRY(reschedule_interrupt) - apicinterrupt RESCHEDULE_VECTOR,smp_reschedule_interrupt -END(reschedule_interrupt) - - .macro INVALIDATE_ENTRY num -ENTRY(invalidate_interrupt\num) - apicinterrupt INVALIDATE_TLB_VECTOR_START+\num,smp_invalidate_interrupt -END(invalidate_interrupt\num) - .endm - - INVALIDATE_ENTRY 0 - INVALIDATE_ENTRY 1 - INVALIDATE_ENTRY 2 - INVALIDATE_ENTRY 3 - INVALIDATE_ENTRY 4 - INVALIDATE_ENTRY 5 - INVALIDATE_ENTRY 6 - INVALIDATE_ENTRY 7 - -ENTRY(call_function_interrupt) - apicinterrupt CALL_FUNCTION_VECTOR,smp_call_function_interrupt -END(call_function_interrupt) -ENTRY(call_function_single_interrupt) - apicinterrupt CALL_FUNCTION_SINGLE_VECTOR,smp_call_function_single_interrupt -END(call_function_single_interrupt) -ENTRY(irq_move_cleanup_interrupt) - apicinterrupt IRQ_MOVE_CLEANUP_VECTOR,smp_irq_move_cleanup_interrupt -END(irq_move_cleanup_interrupt) +apicinterrupt INVALIDATE_TLB_VECTOR_START+0 \ + invalidate_interrupt0 smp_invalidate_interrupt +apicinterrupt INVALIDATE_TLB_VECTOR_START+1 \ + invalidate_interrupt1 smp_invalidate_interrupt +apicinterrupt INVALIDATE_TLB_VECTOR_START+2 \ + invalidate_interrupt2 smp_invalidate_interrupt +apicinterrupt INVALIDATE_TLB_VECTOR_START+3 \ + invalidate_interrupt3 smp_invalidate_interrupt +apicinterrupt INVALIDATE_TLB_VECTOR_START+4 \ + invalidate_interrupt4 smp_invalidate_interrupt +apicinterrupt INVALIDATE_TLB_VECTOR_START+5 \ + invalidate_interrupt5 smp_invalidate_interrupt +apicinterrupt INVALIDATE_TLB_VECTOR_START+6 \ + invalidate_interrupt6 smp_invalidate_interrupt +apicinterrupt INVALIDATE_TLB_VECTOR_START+7 \ + invalidate_interrupt7 smp_invalidate_interrupt #endif -ENTRY(apic_timer_interrupt) - apicinterrupt LOCAL_TIMER_VECTOR,smp_apic_timer_interrupt -END(apic_timer_interrupt) - -ENTRY(uv_bau_message_intr1) - apicinterrupt 220,uv_bau_message_interrupt -END(uv_bau_message_intr1) +apicinterrupt THRESHOLD_APIC_VECTOR \ + threshold_interrupt mce_threshold_interrupt +apicinterrupt THERMAL_APIC_VECTOR \ + thermal_interrupt smp_thermal_interrupt -ENTRY(error_interrupt) - apicinterrupt ERROR_APIC_VECTOR,smp_error_interrupt -END(error_interrupt) +#ifdef CONFIG_SMP +apicinterrupt CALL_FUNCTION_SINGLE_VECTOR \ + call_function_single_interrupt smp_call_function_single_interrupt +apicinterrupt CALL_FUNCTION_VECTOR \ + call_function_interrupt smp_call_function_interrupt +apicinterrupt RESCHEDULE_VECTOR \ + reschedule_interrupt smp_reschedule_interrupt +#endif -ENTRY(spurious_interrupt) - apicinterrupt SPURIOUS_APIC_VECTOR,smp_spurious_interrupt -END(spurious_interrupt) +apicinterrupt ERROR_APIC_VECTOR \ + error_interrupt smp_error_interrupt +apicinterrupt SPURIOUS_APIC_VECTOR \ + spurious_interrupt smp_spurious_interrupt /* * Exception entry points. */ - .macro zeroentry sym +.macro zeroentry sym do_sym +ENTRY(\sym) INTR_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME pushq_cfi $-1 /* ORIG_RAX: no syscall to restart */ @@ -1001,12 +995,14 @@ END(spurious_interrupt) DEFAULT_FRAME 0 movq %rsp,%rdi /* pt_regs pointer */ xorl %esi,%esi /* no error code */ - call \sym + call \do_sym jmp error_exit /* %ebx: no swapgs flag */ CFI_ENDPROC - .endm +END(\sym) +.endm - .macro paranoidzeroentry sym +.macro paranoidzeroentry sym do_sym +KPROBE_ENTRY(\sym) INTR_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME pushq $-1 /* ORIG_RAX: no syscall to restart */ @@ -1016,12 +1012,14 @@ END(spurious_interrupt) TRACE_IRQS_OFF movq %rsp,%rdi /* pt_regs pointer */ xorl %esi,%esi /* no error code */ - call \sym + call \do_sym jmp paranoid_exit /* %ebx: no swapgs flag */ CFI_ENDPROC - .endm +KPROBE_END(\sym) +.endm - .macro paranoidzeroentry_ist sym ist +.macro paranoidzeroentry_ist sym do_sym ist +KPROBE_ENTRY(\sym) INTR_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME pushq $-1 /* ORIG_RAX: no syscall to restart */ @@ -1033,13 +1031,19 @@ END(spurious_interrupt) xorl %esi,%esi /* no error code */ movq %gs:pda_data_offset, %rbp subq $EXCEPTION_STKSZ, per_cpu__init_tss + TSS_ist + (\ist - 1) * 8(%rbp) - call \sym + call \do_sym addq $EXCEPTION_STKSZ, per_cpu__init_tss + TSS_ist + (\ist - 1) * 8(%rbp) jmp paranoid_exit /* %ebx: no swapgs flag */ CFI_ENDPROC - .endm +KPROBE_END(\sym) +.endm - .macro errorentry sym +.macro errorentry sym do_sym entry=0 +.if \entry +KPROBE_ENTRY(\sym) +.else +ENTRY(\sym) +.endif XCPT_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME subq $15*8,%rsp @@ -1049,13 +1053,23 @@ END(spurious_interrupt) movq %rsp,%rdi /* pt_regs pointer */ movq ORIG_RAX(%rsp),%rsi /* get error code */ movq $-1,ORIG_RAX(%rsp) /* no syscall to restart */ - call \sym + call \do_sym jmp error_exit /* %ebx: no swapgs flag */ CFI_ENDPROC - .endm +.if \entry +KPROBE_END(\sym) +.else +END(\sym) +.endif +.endm /* error code is on the stack already */ - .macro paranoiderrorentry sym +.macro paranoiderrorentry sym do_sym entry=1 +.if \entry +KPROBE_ENTRY(\sym) +.else +ENTRY(\sym) +.endif XCPT_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME subq $15*8,%rsp @@ -1066,10 +1080,37 @@ END(spurious_interrupt) movq %rsp,%rdi /* pt_regs pointer */ movq ORIG_RAX(%rsp),%rsi /* get error code */ movq $-1,ORIG_RAX(%rsp) /* no syscall to restart */ - call \sym + call \do_sym jmp paranoid_exit /* %ebx: no swapgs flag */ CFI_ENDPROC - .endm +.if \entry +KPROBE_END(\sym) +.else +END(\sym) +.endif +.endm + +zeroentry divide_error do_divide_error +paranoidzeroentry_ist debug do_debug DEBUG_STACK +paranoidzeroentry_ist int3 do_int3 DEBUG_STACK +zeroentry overflow do_overflow +zeroentry bounds do_bounds +zeroentry invalid_op do_invalid_op +zeroentry device_not_available do_device_not_available +paranoiderrorentry double_fault do_double_fault 0 +zeroentry coprocessor_segment_overrun do_coprocessor_segment_overrun +errorentry invalid_TSS do_invalid_TSS +errorentry segment_not_present do_segment_not_present +paranoiderrorentry stack_segment do_stack_segment +errorentry general_protection do_general_protection 1 +errorentry page_fault do_page_fault 1 +zeroentry spurious_interrupt_bug do_spurious_interrupt_bug +zeroentry coprocessor_error do_coprocessor_error +errorentry alignment_check do_alignment_check +#ifdef CONFIG_X86_MCE +paranoidzeroentry machine_check do_machine_check +#endif +zeroentry simd_coprocessor_error do_simd_coprocessor_error /* * "Paranoid" exit path from exception stack. @@ -1321,26 +1362,7 @@ ENTRY(kernel_execve) CFI_ENDPROC ENDPROC(kernel_execve) -KPROBE_ENTRY(page_fault) - errorentry do_page_fault -KPROBE_END(page_fault) -ENTRY(coprocessor_error) - zeroentry do_coprocessor_error -END(coprocessor_error) - -ENTRY(simd_coprocessor_error) - zeroentry do_simd_coprocessor_error -END(simd_coprocessor_error) - -ENTRY(device_not_available) - zeroentry do_device_not_available -END(device_not_available) - - /* runs on exception stack */ -KPROBE_ENTRY(debug) - paranoidzeroentry_ist do_debug, DEBUG_STACK -KPROBE_END(debug) /* runs on exception stack */ KPROBE_ENTRY(nmi) @@ -1397,67 +1419,6 @@ nmi_schedule: #endif KPROBE_END(nmi) -KPROBE_ENTRY(int3) - paranoidzeroentry_ist do_int3, DEBUG_STACK -KPROBE_END(int3) - -ENTRY(overflow) - zeroentry do_overflow -END(overflow) - -ENTRY(bounds) - zeroentry do_bounds -END(bounds) - -ENTRY(invalid_op) - zeroentry do_invalid_op -END(invalid_op) - -ENTRY(coprocessor_segment_overrun) - zeroentry do_coprocessor_segment_overrun -END(coprocessor_segment_overrun) - - /* runs on exception stack */ -ENTRY(double_fault) - paranoiderrorentry do_double_fault -END(double_fault) - -ENTRY(invalid_TSS) - errorentry do_invalid_TSS -END(invalid_TSS) - -ENTRY(segment_not_present) - errorentry do_segment_not_present -END(segment_not_present) - - /* runs on exception stack */ -ENTRY(stack_segment) - paranoiderrorentry do_stack_segment -END(stack_segment) - -KPROBE_ENTRY(general_protection) - errorentry do_general_protection -KPROBE_END(general_protection) - -ENTRY(alignment_check) - errorentry do_alignment_check -END(alignment_check) - -ENTRY(divide_error) - zeroentry do_divide_error -END(divide_error) - -ENTRY(spurious_interrupt_bug) - zeroentry do_spurious_interrupt_bug -END(spurious_interrupt_bug) - -#ifdef CONFIG_X86_MCE - /* runs on exception stack */ -ENTRY(machine_check) - paranoidzeroentry do_machine_check -END(machine_check) -#endif - /* Call softirq on interrupt stack. Interrupts are off. */ ENTRY(call_softirq) CFI_STARTPROC @@ -1486,9 +1447,7 @@ KPROBE_ENTRY(ignore_sysret) ENDPROC(ignore_sysret) #ifdef CONFIG_XEN -ENTRY(xen_hypervisor_callback) - zeroentry xen_do_hypervisor_callback -END(xen_hypervisor_callback) +zeroentry xen_hypervisor_callback xen_do_hypervisor_callback /* # A note on the "critical region" in our callback handler. -- cgit v1.2.3 From 6efdcfaf16cc4fc76651603e083cf3ec4bd1e6de Mon Sep 17 00:00:00 2001 From: Alexander van Heukelum Date: Sun, 23 Nov 2008 10:15:32 +0100 Subject: x86: KPROBE_ENTRY should be paired wth KPROBE_END Impact: move some code out of .kprobes.text KPROBE_ENTRY switches code generation to .kprobes.text, and KPROBE_END uses .popsection to get back to the previous section (.text, normally). Also replace ENDPROC by END, for consistency. Signed-off-by: Alexander van Heukelum Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_64.S | 16 +++++++--------- 1 file changed, 7 insertions(+), 9 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 1a856c0b21a8..f2d546e16354 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -1167,7 +1167,7 @@ paranoid_schedule: TRACE_IRQS_OFF jmp paranoid_userspace CFI_ENDPROC -END(paranoid_exit) +KPROBE_END(paranoid_exit) /* * Exception entry point. This expects an error code/orig_rax on the stack. @@ -1259,7 +1259,7 @@ gs_change: CFI_ADJUST_CFA_OFFSET -8 ret CFI_ENDPROC -ENDPROC(native_load_gs_index) +END(native_load_gs_index) .section __ex_table,"a" .align 8 @@ -1313,7 +1313,7 @@ ENTRY(kernel_thread) UNFAKE_STACK_FRAME ret CFI_ENDPROC -ENDPROC(kernel_thread) +END(kernel_thread) child_rip: pushq $0 # fake return address @@ -1329,7 +1329,7 @@ child_rip: mov %eax, %edi call do_exit CFI_ENDPROC -ENDPROC(child_rip) +END(child_rip) /* * execve(). This function needs to use IRET, not SYSRET, to set up all state properly. @@ -1360,9 +1360,7 @@ ENTRY(kernel_execve) UNFAKE_STACK_FRAME ret CFI_ENDPROC -ENDPROC(kernel_execve) - - +END(kernel_execve) /* runs on exception stack */ KPROBE_ENTRY(nmi) @@ -1437,14 +1435,14 @@ ENTRY(call_softirq) decl %gs:pda_irqcount ret CFI_ENDPROC -ENDPROC(call_softirq) +END(call_softirq) KPROBE_ENTRY(ignore_sysret) CFI_STARTPROC mov $-ENOSYS,%eax sysret CFI_ENDPROC -ENDPROC(ignore_sysret) +KPROBE_END(ignore_sysret) #ifdef CONFIG_XEN zeroentry xen_hypervisor_callback xen_do_hypervisor_callback -- cgit v1.2.3 From 3b6c52b5b634ae41d762cb174465272d69198160 Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Sun, 23 Nov 2008 20:21:39 +0300 Subject: x86: introduce ENTRY(KPROBE_ENTRY)_X86 assembly helpers to catch unbalanced declaration v3 Impact: make ENTRY()/END() macros more capable It's usefull to catch unbalanced or messed or mixed declarations of ENTRY and KPROBES. These macros would help a bit. For example the following code would compile without problems ENTRY_X86(mcount) retq END_X86(mcount) But if you forget and mess the following form ENTRY_X86(mcount) retq END(mcount) ENTRY_X86(ftrace_caller) The assembler will issue the following message: Error: ENTRY_X86/KPROBE_X86 unbalanced,missed,mixed Actually the checking is performed at every _X86 macro so maybe it's good idea to put ENTRY_KPROBE_FINAL_X86 at the end of .S file to be sure you didn't miss anything. Signed-off-by: Cyrill Gorcunov Cc: Alexander van Heukelum Signed-off-by: Ingo Molnar --- arch/x86/include/asm/linkage.h | 60 ++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 60 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/linkage.h b/arch/x86/include/asm/linkage.h index f61ee8f937e4..5d98d0b68ffc 100644 --- a/arch/x86/include/asm/linkage.h +++ b/arch/x86/include/asm/linkage.h @@ -57,5 +57,65 @@ #define __ALIGN_STR ".align 16,0x90" #endif +/* + * to check ENTRY_X86/END_X86 and + * KPROBE_ENTRY_X86/KPROBE_END_X86 + * unbalanced-missed-mixed appearance + */ +#define __set_entry_x86 .set ENTRY_X86_IN, 0 +#define __unset_entry_x86 .set ENTRY_X86_IN, 1 +#define __set_kprobe_x86 .set KPROBE_X86_IN, 0 +#define __unset_kprobe_x86 .set KPROBE_X86_IN, 1 + +#define __macro_err_x86 .error "ENTRY_X86/KPROBE_X86 unbalanced,missed,mixed" + +#define __check_entry_x86 \ + .ifdef ENTRY_X86_IN; \ + .ifeq ENTRY_X86_IN; \ + __macro_err_x86; \ + .abort; \ + .endif; \ + .endif + +#define __check_kprobe_x86 \ + .ifdef KPROBE_X86_IN; \ + .ifeq KPROBE_X86_IN; \ + __macro_err_x86; \ + .abort; \ + .endif; \ + .endif + +#define __check_entry_kprobe_x86 \ + __check_entry_x86; \ + __check_kprobe_x86 + +#define ENTRY_KPROBE_FINAL_X86 __check_entry_kprobe_x86 + +#define ENTRY_X86(name) \ + __check_entry_kprobe_x86; \ + __set_entry_x86; \ + .globl name; \ + __ALIGN; \ + name: + +#define END_X86(name) \ + __unset_entry_x86; \ + __check_entry_kprobe_x86; \ + .size name, .-name + +#define KPROBE_ENTRY_X86(name) \ + __check_entry_kprobe_x86; \ + __set_kprobe_x86; \ + .pushsection .kprobes.text, "ax"; \ + .globl name; \ + __ALIGN; \ + name: + +#define KPROBE_END_X86(name) \ + __unset_kprobe_x86; \ + __check_entry_kprobe_x86; \ + .size name, .-name; \ + .popsection + #endif /* _ASM_X86_LINKAGE_H */ -- cgit v1.2.3 From a1a00b58855ccdbedf556b4f5638d5208b454472 Mon Sep 17 00:00:00 2001 From: Hannes Eder Date: Sun, 23 Nov 2008 19:37:09 +0100 Subject: x86: boot - fix sparse warnings Impact: make global variables static Fix these sparse warnings: arch/x86/boot/video.c:233:3: warning: symbol 'saved' was not declared. Should it be static? arch/x86/boot/video-vga.c:37:13: warning: symbol 'video_vga' was not declared. Should it be static? Signed-off-by: Hannes Eder Signed-off-by: Ingo Molnar --- arch/x86/boot/video-vga.c | 4 ++-- arch/x86/boot/video.c | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/boot/video-vga.c b/arch/x86/boot/video-vga.c index b939cb476dec..5d4742ed4aa2 100644 --- a/arch/x86/boot/video-vga.c +++ b/arch/x86/boot/video-vga.c @@ -34,7 +34,7 @@ static struct mode_info cga_modes[] = { { VIDEO_80x25, 80, 25, 0 }, }; -__videocard video_vga; +static __videocard video_vga; /* Set basic 80x25 mode */ static u8 vga_set_basic_mode(void) @@ -259,7 +259,7 @@ static int vga_probe(void) return mode_count[adapter]; } -__videocard video_vga = { +static __videocard video_vga = { .card_name = "VGA", .probe = vga_probe, .set_mode = vga_set_mode, diff --git a/arch/x86/boot/video.c b/arch/x86/boot/video.c index 83598b23093a..3bef2c1febe9 100644 --- a/arch/x86/boot/video.c +++ b/arch/x86/boot/video.c @@ -226,7 +226,7 @@ static unsigned int mode_menu(void) #ifdef CONFIG_VIDEO_RETAIN /* Save screen content to the heap */ -struct saved_screen { +static struct saved_screen { int x, y; int curx, cury; u16 *data; -- cgit v1.2.3 From 5f5db591326779a80cfe490c5d6b6ce9fac08b31 Mon Sep 17 00:00:00 2001 From: jia zhang Date: Sun, 23 Nov 2008 22:47:10 +0800 Subject: x86, debug: remove the confusing entry in call trace Impact: improve backtrace quality avoid the confusion in call trace because of the lack of padding at the tail of function. When do_exit gets called, the return address behind call instruction is pushed into stack. If something get wrong in do_exit, for x86_64, the entry "kernel_execve +0x00/0xXX" rather than "child_rip +0xYY/0xZZ" is in the call trace. That looks confusing, so add a u2d to make the return address still part of the original call site. (This also catches any instances of us returning from that function somehow.) Signed-off-by: jia zhang Acked-by: Alexander van Heukelum Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_32.S | 1 + arch/x86/kernel/entry_64.S | 1 + 2 files changed, 2 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S index 28b597ef9ca1..f6402c4ba10d 100644 --- a/arch/x86/kernel/entry_32.S +++ b/arch/x86/kernel/entry_32.S @@ -1051,6 +1051,7 @@ ENTRY(kernel_thread_helper) push %eax CFI_ADJUST_CFA_OFFSET 4 call do_exit + ud2 # padding for call trace CFI_ENDPROC ENDPROC(kernel_thread_helper) diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index ddeeb1052583..4a16bf31c783 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -1172,6 +1172,7 @@ child_rip: # exit mov %eax, %edi call do_exit + ud2 # padding for call trace CFI_ENDPROC ENDPROC(child_rip) -- cgit v1.2.3 From 3b71e9e307b3406aa29960a7428247f8a48b810c Mon Sep 17 00:00:00 2001 From: Hannes Eder Date: Sun, 23 Nov 2008 20:19:33 +0100 Subject: x86: HPET: fix sparse warning Impact: make global variable static Fix this sparse warning: arch/x86/kernel/hpet.c:36:18: warning: symbol 'hpet_num_timers' was not declared. Should it be static? Signed-off-by: Hannes Eder Signed-off-by: Ingo Molnar --- arch/x86/kernel/hpet.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/hpet.c b/arch/x86/kernel/hpet.c index 067d8de913f6..15fcaacc1f84 100644 --- a/arch/x86/kernel/hpet.c +++ b/arch/x86/kernel/hpet.c @@ -33,7 +33,7 @@ * HPET address is set in acpi/boot.c, when an ACPI entry exists */ unsigned long hpet_address; -unsigned long hpet_num_timers; +static unsigned long hpet_num_timers; static void __iomem *hpet_virt_address; struct hpet_dev { -- cgit v1.2.3 From b47b92884212008b4bd044ba6b48b93c00b10ec6 Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Mon, 24 Nov 2008 00:50:09 -0800 Subject: x86: drop REBOOT_CF9_COND from reboot fallback chain Impact: Reverts sequence of reboot fallbacks Checkin 14d7ca5c575853664d8fe4f225a77b8df1b7de7d changed the default reboot method to "pci", a.k.a. port CF9. Unfortunately this has been shown to cause lockups on at least two systems for which REBOOT_KBD worked, both Thinkpads with Intel chipsets. Checkin 3889d0cea2b73049bdca062d9ff1e5d33468289c reverted the default, but did not revert the fallback chain. This checkin reverts the fallback chain; port CF9 is now only done by explicit "reboot=pci" or a future potential DMI key. Signed-off-by: H. Peter Anvin --- arch/x86/kernel/reboot.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c index 790b09fbadcb..bb387ab0eea8 100644 --- a/arch/x86/kernel/reboot.c +++ b/arch/x86/kernel/reboot.c @@ -384,20 +384,20 @@ static void native_machine_emergency_restart(void) load_idt(&no_idt); __asm__ __volatile__("int3"); - reboot_type = BOOT_CF9_COND; + reboot_type = BOOT_KBD; break; #ifdef CONFIG_X86_32 case BOOT_BIOS: machine_real_restart(jump_to_bios, sizeof(jump_to_bios)); - reboot_type = BOOT_CF9_COND; + reboot_type = BOOT_KBD; break; #endif case BOOT_ACPI: acpi_reboot(); - reboot_type = BOOT_CF9_COND; + reboot_type = BOOT_KBD; break; case BOOT_EFI: @@ -406,7 +406,7 @@ static void native_machine_emergency_restart(void) EFI_RESET_WARM : EFI_RESET_COLD, EFI_SUCCESS, 0, NULL); - reboot_type = BOOT_CF9_COND; + reboot_type = BOOT_KBD; break; case BOOT_CF9: -- cgit v1.2.3 From e45f2c07742d447597df001c878bc4a8aafcde37 Mon Sep 17 00:00:00 2001 From: "Denis V. Lunev" Date: Mon, 24 Nov 2008 11:28:36 +0300 Subject: x86: correct link to HPET timer specification Impact: update documentation / help text Original link is dead. Signed-off-by: Denis V. Lunev Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 2 +- drivers/char/hpet.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index ac22bb7719f7..19f0d97829ee 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -482,7 +482,7 @@ config HPET_TIMER The HPET provides a stable time base on SMP systems, unlike the TSC, but it is more expensive to access, as it is off-chip. You can find the HPET spec at - . + . You can safely choose Y here. However, HPET will only be activated if the platform and the BIOS support this feature. diff --git a/drivers/char/hpet.c b/drivers/char/hpet.c index 53fdc7ff3870..32b8bbf5003e 100644 --- a/drivers/char/hpet.c +++ b/drivers/char/hpet.c @@ -46,7 +46,7 @@ /* * The High Precision Event Timer driver. * This driver is closely modelled after the rtc.c driver. - * http://www.intel.com/hardwaredesign/hpetspec.htm + * http://www.intel.com/hardwaredesign/hpetspec_1.pdf */ #define HPET_USER_FREQ (64) #define HPET_DRIFT (500) -- cgit v1.2.3 From ad07e914e681f18ec0eaba60db17f497ee7e7e78 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Mon, 24 Nov 2008 11:33:12 +0100 Subject: x86 defconfig: increase CONFIG_LOG_BUF_SHIFT Impact: double the defconfig printk buffer Booting defconfigs produces more output than 128K so the output is truncated - double it to 256K. Signed-off-by: Ingo Molnar --- arch/x86/configs/i386_defconfig | 2 +- arch/x86/configs/x86_64_defconfig | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/configs/i386_defconfig b/arch/x86/configs/i386_defconfig index 13b8c86ae985..71fc39c70782 100644 --- a/arch/x86/configs/i386_defconfig +++ b/arch/x86/configs/i386_defconfig @@ -77,7 +77,7 @@ CONFIG_AUDIT=y CONFIG_AUDITSYSCALL=y CONFIG_AUDIT_TREE=y # CONFIG_IKCONFIG is not set -CONFIG_LOG_BUF_SHIFT=17 +CONFIG_LOG_BUF_SHIFT=18 CONFIG_CGROUPS=y # CONFIG_CGROUP_DEBUG is not set CONFIG_CGROUP_NS=y diff --git a/arch/x86/configs/x86_64_defconfig b/arch/x86/configs/x86_64_defconfig index f0a03d7a7d63..b38bbabc1706 100644 --- a/arch/x86/configs/x86_64_defconfig +++ b/arch/x86/configs/x86_64_defconfig @@ -77,7 +77,7 @@ CONFIG_AUDIT=y CONFIG_AUDITSYSCALL=y CONFIG_AUDIT_TREE=y # CONFIG_IKCONFIG is not set -CONFIG_LOG_BUF_SHIFT=17 +CONFIG_LOG_BUF_SHIFT=18 CONFIG_CGROUPS=y # CONFIG_CGROUP_DEBUG is not set CONFIG_CGROUP_NS=y -- cgit v1.2.3 From e951e4af2e399c46891004d4931333d2d8d520ab Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Tue, 25 Nov 2008 08:42:01 +0100 Subject: x86: fix unused variable warning in arch/x86/kernel/hpet.c MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Impact: fix build warning this warning: arch/x86/kernel/hpet.c:36: warning: ‘hpet_num_timers’ defined but not used Triggers because hpet_num_timers is unused in the !CONFIG_PCI_MSI case. Signed-off-by: Ingo Molnar --- arch/x86/kernel/hpet.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/hpet.c b/arch/x86/kernel/hpet.c index 15fcaacc1f84..3f0a3edf0a57 100644 --- a/arch/x86/kernel/hpet.c +++ b/arch/x86/kernel/hpet.c @@ -33,7 +33,9 @@ * HPET address is set in acpi/boot.c, when an ACPI entry exists */ unsigned long hpet_address; +#ifdef CONFIG_PCI_MSI static unsigned long hpet_num_timers; +#endif static void __iomem *hpet_virt_address; struct hpet_dev { -- cgit v1.2.3 From ca0002a179bfa532d009a9272d619732872c49bd Mon Sep 17 00:00:00 2001 From: Markus Metzger Date: Tue, 25 Nov 2008 09:01:25 +0100 Subject: x86, bts: base in-kernel ds interface on handles Impact: generalize the DS code to shared buffers Change the in-kernel ds.h interface to identify the tracer via a handle returned on ds_request_~(). Tracers used to be identified via their task_struct. The changes are required to allow DS to be shared between different tasks, which is needed for perfmon2 and for ftrace. For ptrace, the handle is stored in the traced task's task_struct. This should probably go into a (arch-specific) ptrace context some time. Signed-off-by: Markus Metzger Signed-off-by: Ingo Molnar --- arch/x86/include/asm/ds.h | 124 ++++----- arch/x86/kernel/ds.c | 679 +++++++++++++++++++++++----------------------- arch/x86/kernel/ptrace.c | 73 ++--- include/linux/sched.h | 9 + 4 files changed, 446 insertions(+), 439 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/ds.h b/arch/x86/include/asm/ds.h index a95008457ea4..0af997de5f01 100644 --- a/arch/x86/include/asm/ds.h +++ b/arch/x86/include/asm/ds.h @@ -26,11 +26,18 @@ #include #include +#include #ifdef CONFIG_X86_DS struct task_struct; +struct ds_tracer; +struct bts_tracer; +struct pebs_tracer; + +typedef void (*bts_ovfl_callback_t)(struct bts_tracer *); +typedef void (*pebs_ovfl_callback_t)(struct pebs_tracer *); /* * Request BTS or PEBS @@ -38,21 +45,29 @@ struct task_struct; * Due to alignement constraints, the actual buffer may be slightly * smaller than the requested or provided buffer. * - * Returns 0 on success; -Eerrno otherwise + * Returns a pointer to a tracer structure on success, or + * ERR_PTR(errcode) on failure. + * + * The interrupt threshold is independent from the overflow callback + * to allow users to use their own overflow interrupt handling mechanism. * * task: the task to request recording for; * NULL for per-cpu recording on the current cpu * base: the base pointer for the (non-pageable) buffer; * NULL if buffer allocation requested - * size: the size of the requested or provided buffer + * size: the size of the requested or provided buffer in bytes * ovfl: pointer to a function to be called on buffer overflow; * NULL if cyclic buffer requested + * th: the interrupt threshold in records from the end of the buffer; + * -1 if no interrupt threshold is requested. */ -typedef void (*ds_ovfl_callback_t)(struct task_struct *); -extern int ds_request_bts(struct task_struct *task, void *base, size_t size, - ds_ovfl_callback_t ovfl); -extern int ds_request_pebs(struct task_struct *task, void *base, size_t size, - ds_ovfl_callback_t ovfl); +extern struct bts_tracer *ds_request_bts(struct task_struct *task, + void *base, size_t size, + bts_ovfl_callback_t ovfl, size_t th); +extern struct pebs_tracer *ds_request_pebs(struct task_struct *task, + void *base, size_t size, + pebs_ovfl_callback_t ovfl, + size_t th); /* * Release BTS or PEBS resources @@ -61,37 +76,34 @@ extern int ds_request_pebs(struct task_struct *task, void *base, size_t size, * * Returns 0 on success; -Eerrno otherwise * - * task: the task to release resources for; - * NULL to release resources for the current cpu + * tracer: the tracer handle returned from ds_request_~() */ -extern int ds_release_bts(struct task_struct *task); -extern int ds_release_pebs(struct task_struct *task); +extern int ds_release_bts(struct bts_tracer *tracer); +extern int ds_release_pebs(struct pebs_tracer *tracer); /* - * Return the (array) index of the write pointer. + * Get the (array) index of the write pointer. * (assuming an array of BTS/PEBS records) * - * Returns -Eerrno on error + * Returns 0 on success; -Eerrno on error * - * task: the task to access; - * NULL to access the current cpu - * pos (out): if not NULL, will hold the result + * tracer: the tracer handle returned from ds_request_~() + * pos (out): will hold the result */ -extern int ds_get_bts_index(struct task_struct *task, size_t *pos); -extern int ds_get_pebs_index(struct task_struct *task, size_t *pos); +extern int ds_get_bts_index(struct bts_tracer *tracer, size_t *pos); +extern int ds_get_pebs_index(struct pebs_tracer *tracer, size_t *pos); /* - * Return the (array) index one record beyond the end of the array. + * Get the (array) index one record beyond the end of the array. * (assuming an array of BTS/PEBS records) * - * Returns -Eerrno on error + * Returns 0 on success; -Eerrno on error * - * task: the task to access; - * NULL to access the current cpu - * pos (out): if not NULL, will hold the result + * tracer: the tracer handle returned from ds_request_~() + * pos (out): will hold the result */ -extern int ds_get_bts_end(struct task_struct *task, size_t *pos); -extern int ds_get_pebs_end(struct task_struct *task, size_t *pos); +extern int ds_get_bts_end(struct bts_tracer *tracer, size_t *pos); +extern int ds_get_pebs_end(struct pebs_tracer *tracer, size_t *pos); /* * Provide a pointer to the BTS/PEBS record at parameter index. @@ -102,14 +114,13 @@ extern int ds_get_pebs_end(struct task_struct *task, size_t *pos); * * Returns the size of a single record on success; -Eerrno on error * - * task: the task to access; - * NULL to access the current cpu + * tracer: the tracer handle returned from ds_request_~() * index: the index of the requested record * record (out): pointer to the requested record */ -extern int ds_access_bts(struct task_struct *task, +extern int ds_access_bts(struct bts_tracer *tracer, size_t index, const void **record); -extern int ds_access_pebs(struct task_struct *task, +extern int ds_access_pebs(struct pebs_tracer *tracer, size_t index, const void **record); /* @@ -129,38 +140,24 @@ extern int ds_access_pebs(struct task_struct *task, * * Returns the number of bytes written or -Eerrno. * - * task: the task to access; - * NULL to access the current cpu + * tracer: the tracer handle returned from ds_request_~() * buffer: the buffer to write * size: the size of the buffer */ -extern int ds_write_bts(struct task_struct *task, +extern int ds_write_bts(struct bts_tracer *tracer, const void *buffer, size_t size); -extern int ds_write_pebs(struct task_struct *task, +extern int ds_write_pebs(struct pebs_tracer *tracer, const void *buffer, size_t size); -/* - * Same as ds_write_bts/pebs, but omit ownership checks. - * - * This is needed to have some other task than the owner of the - * BTS/PEBS buffer or the parameter task itself write into the - * respective buffer. - */ -extern int ds_unchecked_write_bts(struct task_struct *task, - const void *buffer, size_t size); -extern int ds_unchecked_write_pebs(struct task_struct *task, - const void *buffer, size_t size); - /* * Reset the write pointer of the BTS/PEBS buffer. * * Returns 0 on success; -Eerrno on error * - * task: the task to access; - * NULL to access the current cpu + * tracer: the tracer handle returned from ds_request_~() */ -extern int ds_reset_bts(struct task_struct *task); -extern int ds_reset_pebs(struct task_struct *task); +extern int ds_reset_bts(struct bts_tracer *tracer); +extern int ds_reset_pebs(struct pebs_tracer *tracer); /* * Clear the BTS/PEBS buffer and reset the write pointer. @@ -168,33 +165,30 @@ extern int ds_reset_pebs(struct task_struct *task); * * Returns 0 on success; -Eerrno on error * - * task: the task to access; - * NULL to access the current cpu + * tracer: the tracer handle returned from ds_request_~() */ -extern int ds_clear_bts(struct task_struct *task); -extern int ds_clear_pebs(struct task_struct *task); +extern int ds_clear_bts(struct bts_tracer *tracer); +extern int ds_clear_pebs(struct pebs_tracer *tracer); /* * Provide the PEBS counter reset value. * * Returns 0 on success; -Eerrno on error * - * task: the task to access; - * NULL to access the current cpu + * tracer: the tracer handle returned from ds_request_pebs() * value (out): the counter reset value */ -extern int ds_get_pebs_reset(struct task_struct *task, u64 *value); +extern int ds_get_pebs_reset(struct pebs_tracer *tracer, u64 *value); /* * Set the PEBS counter reset value. * * Returns 0 on success; -Eerrno on error * - * task: the task to access; - * NULL to access the current cpu + * tracer: the tracer handle returned from ds_request_pebs() * value: the new counter reset value */ -extern int ds_set_pebs_reset(struct task_struct *task, u64 value); +extern int ds_set_pebs_reset(struct pebs_tracer *tracer, u64 value); /* * Initialization @@ -207,17 +201,13 @@ extern void __cpuinit ds_init_intel(struct cpuinfo_x86 *); /* * The DS context - part of struct thread_struct. */ +#define MAX_SIZEOF_DS (12 * 8) + struct ds_context { /* pointer to the DS configuration; goes into MSR_IA32_DS_AREA */ - unsigned char *ds; + unsigned char ds[MAX_SIZEOF_DS]; /* the owner of the BTS and PEBS configuration, respectively */ - struct task_struct *owner[2]; - /* buffer overflow notification function for BTS and PEBS */ - ds_ovfl_callback_t callback[2]; - /* the original buffer address */ - void *buffer[2]; - /* the number of allocated pages for on-request allocated buffers */ - unsigned int pages[2]; + struct ds_tracer *owner[2]; /* use count */ unsigned long count; /* a pointer to the context location inside the thread_struct diff --git a/arch/x86/kernel/ds.c b/arch/x86/kernel/ds.c index d6938d9351cf..96768e9cce99 100644 --- a/arch/x86/kernel/ds.c +++ b/arch/x86/kernel/ds.c @@ -28,6 +28,7 @@ #include #include #include +#include /* @@ -44,6 +45,35 @@ struct ds_configuration { }; static struct ds_configuration ds_cfg; +/* + * A BTS or PEBS tracer. + * + * This holds the configuration of the tracer and serves as a handle + * to identify tracers. + */ +struct ds_tracer { + /* the DS context (partially) owned by this tracer */ + struct ds_context *context; + /* the buffer provided on ds_request() and its size in bytes */ + void *buffer; + size_t size; + /* the number of allocated pages for on-request allocated buffers */ + unsigned int pages; +}; + +struct bts_tracer { + /* the common DS part */ + struct ds_tracer ds; + /* buffer overflow notification function */ + bts_ovfl_callback_t ovfl; +}; + +struct pebs_tracer { + /* the common DS part */ + struct ds_tracer ds; + /* buffer overflow notification function */ + pebs_ovfl_callback_t ovfl; +}; /* * Debug Store (DS) save area configuration (see Intel64 and IA32 @@ -107,35 +137,15 @@ static inline void ds_set(unsigned char *base, enum ds_qualifier qual, (*(unsigned long *)base) = value; } +#define DS_ALIGNMENT (1 << 3) /* BTS and PEBS buffer alignment */ + /* * Locking is done only for allocating BTS or PEBS resources and for * guarding context and buffer memory allocation. - * - * Most functions require the current task to own the ds context part - * they are going to access. All the locking is done when validating - * access to the context. */ static spinlock_t ds_lock = __SPIN_LOCK_UNLOCKED(ds_lock); -/* - * Validate that the current task is allowed to access the BTS/PEBS - * buffer of the parameter task. - * - * Returns 0, if access is granted; -Eerrno, otherwise. - */ -static inline int ds_validate_access(struct ds_context *context, - enum ds_qualifier qual) -{ - if (!context) - return -EPERM; - - if (context->owner[qual] == current) - return 0; - - return -EPERM; -} - /* * We either support (system-wide) per-cpu or per-thread allocation. @@ -183,50 +193,12 @@ static inline int check_tracer(struct task_struct *task) * * Contexts are use-counted. They are allocated on first access and * deallocated when the last user puts the context. - * - * We distinguish between an allocating and a non-allocating get of a - * context: - * - the allocating get is used for requesting BTS/PEBS resources. It - * requires the caller to hold the global ds_lock. - * - the non-allocating get is used for all other cases. A - * non-existing context indicates an error. It acquires and releases - * the ds_lock itself for obtaining the context. - * - * A context and its DS configuration are allocated and deallocated - * together. A context always has a DS configuration of the - * appropriate size. */ static DEFINE_PER_CPU(struct ds_context *, system_context); #define this_system_context per_cpu(system_context, smp_processor_id()) -/* - * Returns the pointer to the parameter task's context or to the - * system-wide context, if task is NULL. - * - * Increases the use count of the returned context, if not NULL. - */ static inline struct ds_context *ds_get_context(struct task_struct *task) -{ - struct ds_context *context; - unsigned long irq; - - spin_lock_irqsave(&ds_lock, irq); - - context = (task ? task->thread.ds_ctx : this_system_context); - if (context) - context->count++; - - spin_unlock_irqrestore(&ds_lock, irq); - - return context; -} - -/* - * Same as ds_get_context, but allocates the context and it's DS - * structure, if necessary; returns NULL; if out of memory. - */ -static inline struct ds_context *ds_alloc_context(struct task_struct *task) { struct ds_context **p_context = (task ? &task->thread.ds_ctx : &this_system_context); @@ -238,16 +210,9 @@ static inline struct ds_context *ds_alloc_context(struct task_struct *task) if (!context) return NULL; - context->ds = kzalloc(ds_cfg.sizeof_ds, GFP_KERNEL); - if (!context->ds) { - kfree(context); - return NULL; - } - spin_lock_irqsave(&ds_lock, irq); if (*p_context) { - kfree(context->ds); kfree(context); context = *p_context; @@ -272,10 +237,6 @@ static inline struct ds_context *ds_alloc_context(struct task_struct *task) return context; } -/* - * Decreases the use count of the parameter context, if not NULL. - * Deallocates the context, if the use count reaches zero. - */ static inline void ds_put_context(struct ds_context *context) { unsigned long irq; @@ -296,13 +257,6 @@ static inline void ds_put_context(struct ds_context *context) if (!context->task || (context->task == current)) wrmsrl(MSR_IA32_DS_AREA, 0); - put_tracer(context->task); - - /* free any leftover buffers from tracers that did not - * deallocate them properly. */ - kfree(context->buffer[ds_bts]); - kfree(context->buffer[ds_pebs]); - kfree(context->ds); kfree(context); out: spin_unlock_irqrestore(&ds_lock, irq); @@ -312,21 +266,29 @@ static inline void ds_put_context(struct ds_context *context) /* * Handle a buffer overflow * - * task: the task whose buffers are overflowing; - * NULL for a buffer overflow on the current cpu * context: the ds context * qual: the buffer type */ -static void ds_overflow(struct task_struct *task, struct ds_context *context, - enum ds_qualifier qual) -{ - if (!context) - return; - - if (context->callback[qual]) - (*context->callback[qual])(task); - - /* todo: do some more overflow handling */ +static void ds_overflow(struct ds_context *context, enum ds_qualifier qual) +{ + switch (qual) { + case ds_bts: { + struct bts_tracer *tracer = + container_of(context->owner[qual], + struct bts_tracer, ds); + if (tracer->ovfl) + tracer->ovfl(tracer); + } + break; + case ds_pebs: { + struct pebs_tracer *tracer = + container_of(context->owner[qual], + struct pebs_tracer, ds); + if (tracer->ovfl) + tracer->ovfl(tracer); + } + break; + } } @@ -343,23 +305,25 @@ static void ds_overflow(struct task_struct *task, struct ds_context *context, static inline void *ds_allocate_buffer(size_t size, unsigned int *pages) { unsigned long rlim, vm, pgsz; - void *buffer; + void *buffer = NULL; pgsz = PAGE_ALIGN(size) >> PAGE_SHIFT; + down_write(¤t->mm->mmap_sem); + rlim = current->signal->rlim[RLIMIT_AS].rlim_cur >> PAGE_SHIFT; vm = current->mm->total_vm + pgsz; if (rlim < vm) - return NULL; + goto out; rlim = current->signal->rlim[RLIMIT_MEMLOCK].rlim_cur >> PAGE_SHIFT; vm = current->mm->locked_vm + pgsz; if (rlim < vm) - return NULL; + goto out; buffer = kzalloc(size, GFP_KERNEL); if (!buffer) - return NULL; + goto out; current->mm->total_vm += pgsz; current->mm->locked_vm += pgsz; @@ -367,290 +331,337 @@ static inline void *ds_allocate_buffer(size_t size, unsigned int *pages) if (pages) *pages = pgsz; + out: + up_write(¤t->mm->mmap_sem); return buffer; } -static int ds_request(struct task_struct *task, void *base, size_t size, - ds_ovfl_callback_t ovfl, enum ds_qualifier qual) +static void ds_install_ds_config(struct ds_context *context, + enum ds_qualifier qual, + void *base, size_t size, size_t ith) { - struct ds_context *context; unsigned long buffer, adj; - const unsigned long alignment = (1 << 3); + + /* adjust the buffer address and size to meet alignment + * constraints: + * - buffer is double-word aligned + * - size is multiple of record size + * + * We checked the size at the very beginning; we have enough + * space to do the adjustment. + */ + buffer = (unsigned long)base; + + adj = ALIGN(buffer, DS_ALIGNMENT) - buffer; + buffer += adj; + size -= adj; + + size /= ds_cfg.sizeof_rec[qual]; + size *= ds_cfg.sizeof_rec[qual]; + + ds_set(context->ds, qual, ds_buffer_base, buffer); + ds_set(context->ds, qual, ds_index, buffer); + ds_set(context->ds, qual, ds_absolute_maximum, buffer + size); + + /* The value for 'no threshold' is -1, which will set the + * threshold outside of the buffer, just like we want it. + */ + ds_set(context->ds, qual, + ds_interrupt_threshold, buffer + size - ith); +} + +static int ds_request(struct ds_tracer *tracer, enum ds_qualifier qual, + struct task_struct *task, + void *base, size_t size, size_t th) +{ + struct ds_context *context; unsigned long irq; - int error = 0; + int error; + error = -EOPNOTSUPP; if (!ds_cfg.sizeof_ds) - return -EOPNOTSUPP; + goto out; /* we require some space to do alignment adjustments below */ - if (size < (alignment + ds_cfg.sizeof_rec[qual])) - return -EINVAL; + error = -EINVAL; + if (size < (DS_ALIGNMENT + ds_cfg.sizeof_rec[qual])) + goto out; - /* buffer overflow notification is not yet implemented */ - if (ovfl) - return -EOPNOTSUPP; + if (th != (size_t)-1) { + th *= ds_cfg.sizeof_rec[qual]; + + error = -EINVAL; + if (size <= th) + goto out; + } + + error = -ENOMEM; + if (!base) { + base = ds_allocate_buffer(size, &tracer->pages); + if (!base) + goto out; + } + tracer->buffer = base; + tracer->size = size; - context = ds_alloc_context(task); + error = -ENOMEM; + context = ds_get_context(task); if (!context) - return -ENOMEM; + goto out; + tracer->context = context; + spin_lock_irqsave(&ds_lock, irq); error = -EPERM; if (!check_tracer(task)) goto out_unlock; - get_tracer(task); - error = -EALREADY; - if (context->owner[qual] == current) - goto out_put_tracer; error = -EPERM; - if (context->owner[qual] != NULL) + if (context->owner[qual]) goto out_put_tracer; - context->owner[qual] = current; + context->owner[qual] = tracer; spin_unlock_irqrestore(&ds_lock, irq); - error = -ENOMEM; - if (!base) { - base = ds_allocate_buffer(size, &context->pages[qual]); - if (!base) - goto out_release; - - context->buffer[qual] = base; - } - error = 0; + ds_install_ds_config(context, qual, base, size, th); - context->callback[qual] = ovfl; - - /* adjust the buffer address and size to meet alignment - * constraints: - * - buffer is double-word aligned - * - size is multiple of record size - * - * We checked the size at the very beginning; we have enough - * space to do the adjustment. - */ - buffer = (unsigned long)base; - - adj = ALIGN(buffer, alignment) - buffer; - buffer += adj; - size -= adj; - - size /= ds_cfg.sizeof_rec[qual]; - size *= ds_cfg.sizeof_rec[qual]; - - ds_set(context->ds, qual, ds_buffer_base, buffer); - ds_set(context->ds, qual, ds_index, buffer); - ds_set(context->ds, qual, ds_absolute_maximum, buffer + size); - - if (ovfl) { - /* todo: select a suitable interrupt threshold */ - } else - ds_set(context->ds, qual, - ds_interrupt_threshold, buffer + size + 1); - - /* we keep the context until ds_release */ - return error; - - out_release: - context->owner[qual] = NULL; - ds_put_context(context); - put_tracer(task); - return error; + return 0; out_put_tracer: - spin_unlock_irqrestore(&ds_lock, irq); - ds_put_context(context); put_tracer(task); - return error; - out_unlock: spin_unlock_irqrestore(&ds_lock, irq); ds_put_context(context); + tracer->context = NULL; + out: return error; } -int ds_request_bts(struct task_struct *task, void *base, size_t size, - ds_ovfl_callback_t ovfl) +struct bts_tracer *ds_request_bts(struct task_struct *task, + void *base, size_t size, + bts_ovfl_callback_t ovfl, size_t th) { - return ds_request(task, base, size, ovfl, ds_bts); -} + struct bts_tracer *tracer; + int error; -int ds_request_pebs(struct task_struct *task, void *base, size_t size, - ds_ovfl_callback_t ovfl) -{ - return ds_request(task, base, size, ovfl, ds_pebs); + /* buffer overflow notification is not yet implemented */ + error = -EOPNOTSUPP; + if (ovfl) + goto out; + + error = -ENOMEM; + tracer = kzalloc(sizeof(*tracer), GFP_KERNEL); + if (!tracer) + goto out; + tracer->ovfl = ovfl; + + error = ds_request(&tracer->ds, ds_bts, task, base, size, th); + if (error < 0) + goto out_tracer; + + return tracer; + + out_tracer: + (void)ds_release_bts(tracer); + out: + return ERR_PTR(error); } -static int ds_release(struct task_struct *task, enum ds_qualifier qual) +struct pebs_tracer *ds_request_pebs(struct task_struct *task, + void *base, size_t size, + pebs_ovfl_callback_t ovfl, size_t th) { - struct ds_context *context; + struct pebs_tracer *tracer; int error; - context = ds_get_context(task); - error = ds_validate_access(context, qual); - if (error < 0) + /* buffer overflow notification is not yet implemented */ + error = -EOPNOTSUPP; + if (ovfl) goto out; - kfree(context->buffer[qual]); - context->buffer[qual] = NULL; + error = -ENOMEM; + tracer = kzalloc(sizeof(*tracer), GFP_KERNEL); + if (!tracer) + goto out; + tracer->ovfl = ovfl; - current->mm->total_vm -= context->pages[qual]; - current->mm->locked_vm -= context->pages[qual]; - context->pages[qual] = 0; - context->owner[qual] = NULL; + error = ds_request(&tracer->ds, ds_pebs, task, base, size, th); + if (error < 0) + goto out_tracer; - /* - * we put the context twice: - * once for the ds_get_context - * once for the corresponding ds_request - */ - ds_put_context(context); + return tracer; + + out_tracer: + (void)ds_release_pebs(tracer); out: - ds_put_context(context); - return error; + return ERR_PTR(error); +} + +static void ds_release(struct ds_tracer *tracer, enum ds_qualifier qual) +{ + if (tracer->context) { + BUG_ON(tracer->context->owner[qual] != tracer); + tracer->context->owner[qual] = NULL; + + put_tracer(tracer->context->task); + ds_put_context(tracer->context); + } + + if (tracer->pages) { + kfree(tracer->buffer); + + down_write(¤t->mm->mmap_sem); + + current->mm->total_vm -= tracer->pages; + current->mm->locked_vm -= tracer->pages; + + up_write(¤t->mm->mmap_sem); + } } -int ds_release_bts(struct task_struct *task) +int ds_release_bts(struct bts_tracer *tracer) { - return ds_release(task, ds_bts); + if (!tracer) + return -EINVAL; + + ds_release(&tracer->ds, ds_bts); + kfree(tracer); + + return 0; } -int ds_release_pebs(struct task_struct *task) +int ds_release_pebs(struct pebs_tracer *tracer) { - return ds_release(task, ds_pebs); + if (!tracer) + return -EINVAL; + + ds_release(&tracer->ds, ds_pebs); + kfree(tracer); + + return 0; } -static int ds_get_index(struct task_struct *task, size_t *pos, - enum ds_qualifier qual) +static size_t ds_get_index(struct ds_context *context, enum ds_qualifier qual) { - struct ds_context *context; unsigned long base, index; - int error; - - context = ds_get_context(task); - error = ds_validate_access(context, qual); - if (error < 0) - goto out; base = ds_get(context->ds, qual, ds_buffer_base); index = ds_get(context->ds, qual, ds_index); - error = ((index - base) / ds_cfg.sizeof_rec[qual]); - if (pos) - *pos = error; - out: - ds_put_context(context); - return error; + return (index - base) / ds_cfg.sizeof_rec[qual]; } -int ds_get_bts_index(struct task_struct *task, size_t *pos) +int ds_get_bts_index(struct bts_tracer *tracer, size_t *pos) { - return ds_get_index(task, pos, ds_bts); + if (!tracer) + return -EINVAL; + + if (!pos) + return -EINVAL; + + *pos = ds_get_index(tracer->ds.context, ds_bts); + + return 0; } -int ds_get_pebs_index(struct task_struct *task, size_t *pos) +int ds_get_pebs_index(struct pebs_tracer *tracer, size_t *pos) { - return ds_get_index(task, pos, ds_pebs); + if (!tracer) + return -EINVAL; + + if (!pos) + return -EINVAL; + + *pos = ds_get_index(tracer->ds.context, ds_pebs); + + return 0; } -static int ds_get_end(struct task_struct *task, size_t *pos, - enum ds_qualifier qual) +static size_t ds_get_end(struct ds_context *context, enum ds_qualifier qual) { - struct ds_context *context; - unsigned long base, end; - int error; - - context = ds_get_context(task); - error = ds_validate_access(context, qual); - if (error < 0) - goto out; + unsigned long base, max; base = ds_get(context->ds, qual, ds_buffer_base); - end = ds_get(context->ds, qual, ds_absolute_maximum); + max = ds_get(context->ds, qual, ds_absolute_maximum); - error = ((end - base) / ds_cfg.sizeof_rec[qual]); - if (pos) - *pos = error; - out: - ds_put_context(context); - return error; + return (max - base) / ds_cfg.sizeof_rec[qual]; } -int ds_get_bts_end(struct task_struct *task, size_t *pos) +int ds_get_bts_end(struct bts_tracer *tracer, size_t *pos) { - return ds_get_end(task, pos, ds_bts); + if (!tracer) + return -EINVAL; + + if (!pos) + return -EINVAL; + + *pos = ds_get_end(tracer->ds.context, ds_bts); + + return 0; } -int ds_get_pebs_end(struct task_struct *task, size_t *pos) +int ds_get_pebs_end(struct pebs_tracer *tracer, size_t *pos) { - return ds_get_end(task, pos, ds_pebs); + if (!tracer) + return -EINVAL; + + if (!pos) + return -EINVAL; + + *pos = ds_get_end(tracer->ds.context, ds_pebs); + + return 0; } -static int ds_access(struct task_struct *task, size_t index, - const void **record, enum ds_qualifier qual) +static int ds_access(struct ds_context *context, enum ds_qualifier qual, + size_t index, const void **record) { - struct ds_context *context; unsigned long base, idx; - int error; if (!record) return -EINVAL; - context = ds_get_context(task); - error = ds_validate_access(context, qual); - if (error < 0) - goto out; - base = ds_get(context->ds, qual, ds_buffer_base); idx = base + (index * ds_cfg.sizeof_rec[qual]); - error = -EINVAL; if (idx > ds_get(context->ds, qual, ds_absolute_maximum)) - goto out; + return -EINVAL; *record = (const void *)idx; - error = ds_cfg.sizeof_rec[qual]; - out: - ds_put_context(context); - return error; + + return ds_cfg.sizeof_rec[qual]; } -int ds_access_bts(struct task_struct *task, size_t index, const void **record) +int ds_access_bts(struct bts_tracer *tracer, size_t index, + const void **record) { - return ds_access(task, index, record, ds_bts); + if (!tracer) + return -EINVAL; + + return ds_access(tracer->ds.context, ds_bts, index, record); } -int ds_access_pebs(struct task_struct *task, size_t index, const void **record) +int ds_access_pebs(struct pebs_tracer *tracer, size_t index, + const void **record) { - return ds_access(task, index, record, ds_pebs); + if (!tracer) + return -EINVAL; + + return ds_access(tracer->ds.context, ds_pebs, index, record); } -static int ds_write(struct task_struct *task, const void *record, size_t size, - enum ds_qualifier qual, int force) +static int ds_write(struct ds_context *context, enum ds_qualifier qual, + const void *record, size_t size) { - struct ds_context *context; - int error; + int bytes_written = 0; if (!record) return -EINVAL; - error = -EPERM; - context = ds_get_context(task); - if (!context) - goto out; - - if (!force) { - error = ds_validate_access(context, qual); - if (error < 0) - goto out; - } - - error = 0; while (size) { unsigned long base, index, end, write_end, int_th; unsigned long write_size, adj_write_size; @@ -678,14 +689,14 @@ static int ds_write(struct task_struct *task, const void *record, size_t size, write_end = end; if (write_end <= index) - goto out; + break; write_size = min((unsigned long) size, write_end - index); memcpy((void *)index, record, write_size); record = (const char *)record + write_size; - size -= write_size; - error += write_size; + size -= write_size; + bytes_written += write_size; adj_write_size = write_size / ds_cfg.sizeof_rec[qual]; adj_write_size *= ds_cfg.sizeof_rec[qual]; @@ -700,47 +711,32 @@ static int ds_write(struct task_struct *task, const void *record, size_t size, ds_set(context->ds, qual, ds_index, index); if (index >= int_th) - ds_overflow(task, context, qual); + ds_overflow(context, qual); } - out: - ds_put_context(context); - return error; + return bytes_written; } -int ds_write_bts(struct task_struct *task, const void *record, size_t size) +int ds_write_bts(struct bts_tracer *tracer, const void *record, size_t size) { - return ds_write(task, record, size, ds_bts, /* force = */ 0); -} + if (!tracer) + return -EINVAL; -int ds_write_pebs(struct task_struct *task, const void *record, size_t size) -{ - return ds_write(task, record, size, ds_pebs, /* force = */ 0); + return ds_write(tracer->ds.context, ds_bts, record, size); } -int ds_unchecked_write_bts(struct task_struct *task, - const void *record, size_t size) +int ds_write_pebs(struct pebs_tracer *tracer, const void *record, size_t size) { - return ds_write(task, record, size, ds_bts, /* force = */ 1); -} + if (!tracer) + return -EINVAL; -int ds_unchecked_write_pebs(struct task_struct *task, - const void *record, size_t size) -{ - return ds_write(task, record, size, ds_pebs, /* force = */ 1); + return ds_write(tracer->ds.context, ds_pebs, record, size); } -static int ds_reset_or_clear(struct task_struct *task, - enum ds_qualifier qual, int clear) +static void ds_reset_or_clear(struct ds_context *context, + enum ds_qualifier qual, int clear) { - struct ds_context *context; unsigned long base, end; - int error; - - context = ds_get_context(task); - error = ds_validate_access(context, qual); - if (error < 0) - goto out; base = ds_get(context->ds, qual, ds_buffer_base); end = ds_get(context->ds, qual, ds_absolute_maximum); @@ -749,70 +745,69 @@ static int ds_reset_or_clear(struct task_struct *task, memset((void *)base, 0, end - base); ds_set(context->ds, qual, ds_index, base); - - error = 0; - out: - ds_put_context(context); - return error; } -int ds_reset_bts(struct task_struct *task) +int ds_reset_bts(struct bts_tracer *tracer) { - return ds_reset_or_clear(task, ds_bts, /* clear = */ 0); + if (!tracer) + return -EINVAL; + + ds_reset_or_clear(tracer->ds.context, ds_bts, /* clear = */ 0); + + return 0; } -int ds_reset_pebs(struct task_struct *task) +int ds_reset_pebs(struct pebs_tracer *tracer) { - return ds_reset_or_clear(task, ds_pebs, /* clear = */ 0); + if (!tracer) + return -EINVAL; + + ds_reset_or_clear(tracer->ds.context, ds_pebs, /* clear = */ 0); + + return 0; } -int ds_clear_bts(struct task_struct *task) +int ds_clear_bts(struct bts_tracer *tracer) { - return ds_reset_or_clear(task, ds_bts, /* clear = */ 1); + if (!tracer) + return -EINVAL; + + ds_reset_or_clear(tracer->ds.context, ds_bts, /* clear = */ 1); + + return 0; } -int ds_clear_pebs(struct task_struct *task) +int ds_clear_pebs(struct pebs_tracer *tracer) { - return ds_reset_or_clear(task, ds_pebs, /* clear = */ 1); + if (!tracer) + return -EINVAL; + + ds_reset_or_clear(tracer->ds.context, ds_pebs, /* clear = */ 1); + + return 0; } -int ds_get_pebs_reset(struct task_struct *task, u64 *value) +int ds_get_pebs_reset(struct pebs_tracer *tracer, u64 *value) { - struct ds_context *context; - int error; + if (!tracer) + return -EINVAL; if (!value) return -EINVAL; - context = ds_get_context(task); - error = ds_validate_access(context, ds_pebs); - if (error < 0) - goto out; - - *value = *(u64 *)(context->ds + (ds_cfg.sizeof_field * 8)); + *value = *(u64 *)(tracer->ds.context->ds + (ds_cfg.sizeof_field * 8)); - error = 0; - out: - ds_put_context(context); - return error; + return 0; } -int ds_set_pebs_reset(struct task_struct *task, u64 value) +int ds_set_pebs_reset(struct pebs_tracer *tracer, u64 value) { - struct ds_context *context; - int error; - - context = ds_get_context(task); - error = ds_validate_access(context, ds_pebs); - if (error < 0) - goto out; + if (!tracer) + return -EINVAL; - *(u64 *)(context->ds + (ds_cfg.sizeof_field * 8)) = value; + *(u64 *)(tracer->ds.context->ds + (ds_cfg.sizeof_field * 8)) = value; - error = 0; - out: - ds_put_context(context); - return error; + return 0; } static const struct ds_configuration ds_cfg_var = { @@ -840,6 +835,10 @@ static inline void ds_configure(const struct ds_configuration *cfg) { ds_cfg = *cfg; + + printk(KERN_INFO "DS available\n"); + + BUG_ON(MAX_SIZEOF_DS < ds_cfg.sizeof_ds); } void __cpuinit ds_init_intel(struct cpuinfo_x86 *c) @@ -883,6 +882,8 @@ void ds_free(struct ds_context *context) * is dying. There should not be any user of that context left * to disturb us, anymore. */ unsigned long leftovers = context->count; - while (leftovers--) + while (leftovers--) { + put_tracer(context->task); ds_put_context(context); + } } diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c index 06180dff5b2e..76adf5b640ff 100644 --- a/arch/x86/kernel/ptrace.c +++ b/arch/x86/kernel/ptrace.c @@ -668,14 +668,14 @@ static int ptrace_bts_read_record(struct task_struct *child, size_t index, size_t bts_index, bts_end; int error; - error = ds_get_bts_end(child, &bts_end); + error = ds_get_bts_end(child->bts, &bts_end); if (error < 0) return error; if (bts_end <= index) return -EINVAL; - error = ds_get_bts_index(child, &bts_index); + error = ds_get_bts_index(child->bts, &bts_index); if (error < 0) return error; @@ -684,7 +684,7 @@ static int ptrace_bts_read_record(struct task_struct *child, size_t index, if (bts_end <= bts_index) bts_index -= bts_end; - error = ds_access_bts(child, bts_index, &bts_record); + error = ds_access_bts(child->bts, bts_index, &bts_record); if (error < 0) return error; @@ -705,14 +705,14 @@ static int ptrace_bts_drain(struct task_struct *child, size_t end, i; int error; - error = ds_get_bts_index(child, &end); + error = ds_get_bts_index(child->bts, &end); if (error < 0) return error; if (size < (end * sizeof(struct bts_struct))) return -EIO; - error = ds_access_bts(child, 0, (const void **)&raw); + error = ds_access_bts(child->bts, 0, (const void **)&raw); if (error < 0) return error; @@ -723,18 +723,13 @@ static int ptrace_bts_drain(struct task_struct *child, return -EFAULT; } - error = ds_clear_bts(child); + error = ds_clear_bts(child->bts); if (error < 0) return error; return end; } -static void ptrace_bts_ovfl(struct task_struct *child) -{ - send_sig(child->thread.bts_ovfl_signal, child, 0); -} - static int ptrace_bts_config(struct task_struct *child, long cfg_size, const struct ptrace_bts_config __user *ucfg) @@ -760,23 +755,29 @@ static int ptrace_bts_config(struct task_struct *child, goto errout; if (cfg.flags & PTRACE_BTS_O_ALLOC) { - ds_ovfl_callback_t ovfl = NULL; + bts_ovfl_callback_t ovfl = NULL; unsigned int sig = 0; - /* we ignore the error in case we were not tracing child */ - (void)ds_release_bts(child); - if (cfg.flags & PTRACE_BTS_O_SIGNAL) { if (!cfg.signal) goto errout; + error = -EOPNOTSUPP; + goto errout; + sig = cfg.signal; - ovfl = ptrace_bts_ovfl; } - error = ds_request_bts(child, /* base = */ NULL, cfg.size, ovfl); - if (error < 0) + if (child->bts) + (void)ds_release_bts(child->bts); + + child->bts = ds_request_bts(child, /* base = */ NULL, cfg.size, + ovfl, /* th = */ (size_t)-1); + if (IS_ERR(child->bts)) { + error = PTR_ERR(child->bts); + child->bts = NULL; goto errout; + } child->thread.bts_ovfl_signal = sig; } @@ -823,15 +824,15 @@ static int ptrace_bts_status(struct task_struct *child, if (cfg_size < sizeof(cfg)) return -EIO; - error = ds_get_bts_end(child, &end); + error = ds_get_bts_end(child->bts, &end); if (error < 0) return error; - error = ds_access_bts(child, /* index = */ 0, &base); + error = ds_access_bts(child->bts, /* index = */ 0, &base); if (error < 0) return error; - error = ds_access_bts(child, /* index = */ end, &max); + error = ds_access_bts(child->bts, /* index = */ end, &max); if (error < 0) return error; @@ -884,10 +885,7 @@ static int ptrace_bts_write_record(struct task_struct *child, return -EINVAL; } - /* The writing task will be the switched-to task on a context - * switch. It needs to write into the switched-from task's BTS - * buffer. */ - return ds_unchecked_write_bts(child, bts_record, bts_cfg.sizeof_bts); + return ds_write_bts(child->bts, bts_record, bts_cfg.sizeof_bts); } void ptrace_bts_take_timestamp(struct task_struct *tsk, @@ -972,13 +970,15 @@ void ptrace_disable(struct task_struct *child) clear_tsk_thread_flag(child, TIF_SYSCALL_EMU); #endif #ifdef CONFIG_X86_PTRACE_BTS - (void)ds_release_bts(child); + if (child->bts) { + (void)ds_release_bts(child->bts); - child->thread.debugctlmsr &= ~bts_cfg.debugctl_mask; - if (!child->thread.debugctlmsr) - clear_tsk_thread_flag(child, TIF_DEBUGCTLMSR); + child->thread.debugctlmsr &= ~bts_cfg.debugctl_mask; + if (!child->thread.debugctlmsr) + clear_tsk_thread_flag(child, TIF_DEBUGCTLMSR); - clear_tsk_thread_flag(child, TIF_BTS_TRACE_TS); + clear_tsk_thread_flag(child, TIF_BTS_TRACE_TS); + } #endif /* CONFIG_X86_PTRACE_BTS */ } @@ -1110,9 +1110,16 @@ long arch_ptrace(struct task_struct *child, long request, long addr, long data) (child, data, (struct ptrace_bts_config __user *)addr); break; - case PTRACE_BTS_SIZE: - ret = ds_get_bts_index(child, /* pos = */ NULL); + case PTRACE_BTS_SIZE: { + size_t size; + + ret = ds_get_bts_index(child->bts, &size); + if (ret == 0) { + BUG_ON(size != (int) size); + ret = (int) size; + } break; + } case PTRACE_BTS_GET: ret = ptrace_bts_read_record @@ -1120,7 +1127,7 @@ long arch_ptrace(struct task_struct *child, long request, long addr, long data) break; case PTRACE_BTS_CLEAR: - ret = ds_clear_bts(child); + ret = ds_clear_bts(child->bts); break; case PTRACE_BTS_DRAIN: diff --git a/include/linux/sched.h b/include/linux/sched.h index bee1e93c95ad..a9780eaa6737 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -96,6 +96,7 @@ struct exec_domain; struct futex_pi_state; struct robust_list_head; struct bio; +struct bts_tracer; /* * List of flags we want to share for kernel threads, @@ -1161,6 +1162,14 @@ struct task_struct { struct list_head ptraced; struct list_head ptrace_entry; +#ifdef CONFIG_X86_PTRACE_BTS + /* + * This is the tracer handle for the ptrace BTS extension. + * This field actually belongs to the ptracer task. + */ + struct bts_tracer *bts; +#endif /* CONFIG_X86_PTRACE_BTS */ + /* PID/PID hash table linkage. */ struct pid_link pids[PIDTYPE_MAX]; struct list_head thread_group; -- cgit v1.2.3 From 6abb11aecd888d1da6276399380b7355f127c006 Mon Sep 17 00:00:00 2001 From: Markus Metzger Date: Tue, 25 Nov 2008 09:05:27 +0100 Subject: x86, bts, ptrace: move BTS buffer allocation from ds.c into ptrace.c Impact: restructure DS memory allocation to be done by the usage site of DS Require pre-allocated buffers in ds.h. Move the BTS buffer allocation for ptrace into ptrace.c. The pointer to the allocated buffer is stored in the traced task's task_struct together with the handle returned by ds_request_bts(). Removes memory accounting code. Signed-off-by: Markus Metzger Signed-off-by: Ingo Molnar --- arch/x86/include/asm/ds.h | 12 +++---- arch/x86/kernel/ds.c | 92 ++++++++--------------------------------------- arch/x86/kernel/ptrace.c | 22 ++++++++++-- include/linux/sched.h | 4 +++ 4 files changed, 42 insertions(+), 88 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/ds.h b/arch/x86/include/asm/ds.h index 0af997de5f01..99b6c39774a4 100644 --- a/arch/x86/include/asm/ds.h +++ b/arch/x86/include/asm/ds.h @@ -7,13 +7,12 @@ * * It manages: * - per-thread and per-cpu allocation of BTS and PEBS - * - buffer memory allocation (optional) - * - buffer overflow handling + * - buffer overflow handling (to be done) * - buffer access * * It assumes: - * - get_task_struct on all parameter tasks - * - current is allowed to trace parameter tasks + * - get_task_struct on all traced tasks + * - current is allowed to trace tasks * * * Copyright (C) 2007-2008 Intel Corporation. @@ -54,8 +53,7 @@ typedef void (*pebs_ovfl_callback_t)(struct pebs_tracer *); * task: the task to request recording for; * NULL for per-cpu recording on the current cpu * base: the base pointer for the (non-pageable) buffer; - * NULL if buffer allocation requested - * size: the size of the requested or provided buffer in bytes + * size: the size of the provided buffer in bytes * ovfl: pointer to a function to be called on buffer overflow; * NULL if cyclic buffer requested * th: the interrupt threshold in records from the end of the buffer; @@ -72,8 +70,6 @@ extern struct pebs_tracer *ds_request_pebs(struct task_struct *task, /* * Release BTS or PEBS resources * - * Frees buffers allocated on ds_request. - * * Returns 0 on success; -Eerrno otherwise * * tracer: the tracer handle returned from ds_request_~() diff --git a/arch/x86/kernel/ds.c b/arch/x86/kernel/ds.c index 96768e9cce99..19a8c2c0389f 100644 --- a/arch/x86/kernel/ds.c +++ b/arch/x86/kernel/ds.c @@ -7,13 +7,12 @@ * * It manages: * - per-thread and per-cpu allocation of BTS and PEBS - * - buffer memory allocation (optional) - * - buffer overflow handling + * - buffer overflow handling (to be done) * - buffer access * * It assumes: - * - get_task_struct on all parameter tasks - * - current is allowed to trace parameter tasks + * - get_task_struct on all traced tasks + * - current is allowed to trace tasks * * * Copyright (C) 2007-2008 Intel Corporation. @@ -57,8 +56,6 @@ struct ds_tracer { /* the buffer provided on ds_request() and its size in bytes */ void *buffer; size_t size; - /* the number of allocated pages for on-request allocated buffers */ - unsigned int pages; }; struct bts_tracer { @@ -141,8 +138,7 @@ static inline void ds_set(unsigned char *base, enum ds_qualifier qual, /* - * Locking is done only for allocating BTS or PEBS resources and for - * guarding context and buffer memory allocation. + * Locking is done only for allocating BTS or PEBS resources. */ static spinlock_t ds_lock = __SPIN_LOCK_UNLOCKED(ds_lock); @@ -292,50 +288,6 @@ static void ds_overflow(struct ds_context *context, enum ds_qualifier qual) } -/* - * Allocate a non-pageable buffer of the parameter size. - * Checks the memory and the locked memory rlimit. - * - * Returns the buffer, if successful; - * NULL, if out of memory or rlimit exceeded. - * - * size: the requested buffer size in bytes - * pages (out): if not NULL, contains the number of pages reserved - */ -static inline void *ds_allocate_buffer(size_t size, unsigned int *pages) -{ - unsigned long rlim, vm, pgsz; - void *buffer = NULL; - - pgsz = PAGE_ALIGN(size) >> PAGE_SHIFT; - - down_write(¤t->mm->mmap_sem); - - rlim = current->signal->rlim[RLIMIT_AS].rlim_cur >> PAGE_SHIFT; - vm = current->mm->total_vm + pgsz; - if (rlim < vm) - goto out; - - rlim = current->signal->rlim[RLIMIT_MEMLOCK].rlim_cur >> PAGE_SHIFT; - vm = current->mm->locked_vm + pgsz; - if (rlim < vm) - goto out; - - buffer = kzalloc(size, GFP_KERNEL); - if (!buffer) - goto out; - - current->mm->total_vm += pgsz; - current->mm->locked_vm += pgsz; - - if (pages) - *pages = pgsz; - - out: - up_write(¤t->mm->mmap_sem); - return buffer; -} - static void ds_install_ds_config(struct ds_context *context, enum ds_qualifier qual, void *base, size_t size, size_t ith) @@ -382,6 +334,10 @@ static int ds_request(struct ds_tracer *tracer, enum ds_qualifier qual, if (!ds_cfg.sizeof_ds) goto out; + error = -EINVAL; + if (!base) + goto out; + /* we require some space to do alignment adjustments below */ error = -EINVAL; if (size < (DS_ALIGNMENT + ds_cfg.sizeof_rec[qual])) @@ -395,13 +351,6 @@ static int ds_request(struct ds_tracer *tracer, enum ds_qualifier qual, goto out; } - error = -ENOMEM; - if (!base) { - base = ds_allocate_buffer(size, &tracer->pages); - if (!base) - goto out; - } - tracer->buffer = base; tracer->size = size; @@ -466,7 +415,7 @@ struct bts_tracer *ds_request_bts(struct task_struct *task, return tracer; out_tracer: - (void)ds_release_bts(tracer); + kfree(tracer); out: return ERR_PTR(error); } @@ -496,31 +445,18 @@ struct pebs_tracer *ds_request_pebs(struct task_struct *task, return tracer; out_tracer: - (void)ds_release_pebs(tracer); + kfree(tracer); out: return ERR_PTR(error); } static void ds_release(struct ds_tracer *tracer, enum ds_qualifier qual) { - if (tracer->context) { - BUG_ON(tracer->context->owner[qual] != tracer); - tracer->context->owner[qual] = NULL; - - put_tracer(tracer->context->task); - ds_put_context(tracer->context); - } + BUG_ON(tracer->context->owner[qual] != tracer); + tracer->context->owner[qual] = NULL; - if (tracer->pages) { - kfree(tracer->buffer); - - down_write(¤t->mm->mmap_sem); - - current->mm->total_vm -= tracer->pages; - current->mm->locked_vm -= tracer->pages; - - up_write(¤t->mm->mmap_sem); - } + put_tracer(tracer->context->task); + ds_put_context(tracer->context); } int ds_release_bts(struct bts_tracer *tracer) diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c index 76adf5b640ff..2c8ec1ba75e6 100644 --- a/arch/x86/kernel/ptrace.c +++ b/arch/x86/kernel/ptrace.c @@ -758,6 +758,10 @@ static int ptrace_bts_config(struct task_struct *child, bts_ovfl_callback_t ovfl = NULL; unsigned int sig = 0; + error = -EINVAL; + if (cfg.size < (10 * bts_cfg.sizeof_bts)) + goto errout; + if (cfg.flags & PTRACE_BTS_O_SIGNAL) { if (!cfg.signal) goto errout; @@ -768,14 +772,26 @@ static int ptrace_bts_config(struct task_struct *child, sig = cfg.signal; } - if (child->bts) + if (child->bts) { (void)ds_release_bts(child->bts); + kfree(child->bts_buffer); + + child->bts = NULL; + child->bts_buffer = NULL; + } + + error = -ENOMEM; + child->bts_buffer = kzalloc(cfg.size, GFP_KERNEL); + if (!child->bts_buffer) + goto errout; - child->bts = ds_request_bts(child, /* base = */ NULL, cfg.size, + child->bts = ds_request_bts(child, child->bts_buffer, cfg.size, ovfl, /* th = */ (size_t)-1); if (IS_ERR(child->bts)) { error = PTR_ERR(child->bts); + kfree(child->bts_buffer); child->bts = NULL; + child->bts_buffer = NULL; goto errout; } @@ -972,6 +988,8 @@ void ptrace_disable(struct task_struct *child) #ifdef CONFIG_X86_PTRACE_BTS if (child->bts) { (void)ds_release_bts(child->bts); + kfree(child->bts_buffer); + child->bts_buffer = NULL; child->thread.debugctlmsr &= ~bts_cfg.debugctl_mask; if (!child->thread.debugctlmsr) diff --git a/include/linux/sched.h b/include/linux/sched.h index a9780eaa6737..d02a0ca70ee9 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1168,6 +1168,10 @@ struct task_struct { * This field actually belongs to the ptracer task. */ struct bts_tracer *bts; + /* + * The buffer to hold the BTS data. + */ + void *bts_buffer; #endif /* CONFIG_X86_PTRACE_BTS */ /* PID/PID hash table linkage. */ -- cgit v1.2.3 From 1e9b51c28312f7334394aa30be56ff52c2b65b7e Mon Sep 17 00:00:00 2001 From: Markus Metzger Date: Tue, 25 Nov 2008 09:24:15 +0100 Subject: x86, bts, ftrace: a BTS ftrace plug-in prototype Impact: add new ftrace plugin A prototype for a BTS ftrace plug-in. The tracer collects branch trace in a cyclic buffer for each cpu. The tracer is not configurable and the trace for each snapshot is appended when doing cat /debug/tracing/trace. This is a proof of concept that will be extended with future patches to become a (hopefully) useful tool. Signed-off-by: Markus Metzger Signed-off-by: Ingo Molnar --- arch/x86/Kconfig.cpu | 1 + kernel/trace/Kconfig | 11 ++ kernel/trace/Makefile | 1 + kernel/trace/trace.h | 12 +++ kernel/trace/trace_bts.c | 276 +++++++++++++++++++++++++++++++++++++++++++++++ 5 files changed, 301 insertions(+) create mode 100644 kernel/trace/trace_bts.c (limited to 'arch/x86') diff --git a/arch/x86/Kconfig.cpu b/arch/x86/Kconfig.cpu index b815664fe370..85a78575956c 100644 --- a/arch/x86/Kconfig.cpu +++ b/arch/x86/Kconfig.cpu @@ -515,6 +515,7 @@ config CPU_SUP_UMC_32 config X86_DS def_bool X86_PTRACE_BTS depends on X86_DEBUGCTLMSR + select HAVE_HW_BRANCH_TRACER config X86_PTRACE_BTS bool "Branch Trace Store" diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 9cbf7761f498..620feadff67a 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -28,6 +28,9 @@ config HAVE_DYNAMIC_FTRACE config HAVE_FTRACE_MCOUNT_RECORD bool +config HAVE_HW_BRANCH_TRACER + bool + config TRACER_MAX_TRACE bool @@ -233,6 +236,14 @@ config STACK_TRACER Say N if unsure. +config BTS_TRACER + depends on HAVE_HW_BRANCH_TRACER + bool "Trace branches" + select TRACING + help + This tracer records all branches on the system in a circular + buffer giving access to the last N branches for each cpu. + config DYNAMIC_FTRACE bool "enable/disable ftrace tracepoints dynamically" depends on FUNCTION_TRACER diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index 1a8c9259dc69..cef4bcb4e822 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -31,5 +31,6 @@ obj-$(CONFIG_MMIOTRACE) += trace_mmiotrace.o obj-$(CONFIG_BOOT_TRACER) += trace_boot.o obj-$(CONFIG_FUNCTION_RET_TRACER) += trace_functions_return.o obj-$(CONFIG_TRACE_BRANCH_PROFILING) += trace_branch.o +obj-$(CONFIG_BTS_TRACER) += trace_bts.o libftrace-y := ftrace.o diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 717f9f045c6f..3abd645e8af2 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -27,6 +27,7 @@ enum trace_type { TRACE_BOOT_RET, TRACE_FN_RET, TRACE_USER_STACK, + TRACE_BTS, __TRACE_LAST_TYPE }; @@ -153,6 +154,12 @@ struct trace_branch { char correct; }; +struct bts_entry { + struct trace_entry ent; + unsigned long from; + unsigned long to; +}; + /* * trace_flag_type is an enumeration that holds different * states when a trace occurs. These are: @@ -258,6 +265,7 @@ extern void __ftrace_bad_type(void); IF_ASSIGN(var, ent, struct trace_boot_ret, TRACE_BOOT_RET);\ IF_ASSIGN(var, ent, struct trace_branch, TRACE_BRANCH); \ IF_ASSIGN(var, ent, struct ftrace_ret_entry, TRACE_FN_RET);\ + IF_ASSIGN(var, ent, struct bts_entry, TRACE_BTS);\ __ftrace_bad_type(); \ } while (0) @@ -392,6 +400,10 @@ void trace_function(struct trace_array *tr, void trace_function_return(struct ftrace_retfunc *trace); +void trace_bts(struct trace_array *tr, + unsigned long from, + unsigned long to); + void tracing_start_cmdline_record(void); void tracing_stop_cmdline_record(void); void tracing_sched_switch_assign_trace(struct trace_array *tr); diff --git a/kernel/trace/trace_bts.c b/kernel/trace/trace_bts.c new file mode 100644 index 000000000000..23b76e4690ef --- /dev/null +++ b/kernel/trace/trace_bts.c @@ -0,0 +1,276 @@ +/* + * BTS tracer + * + * Copyright (C) 2008 Markus Metzger + * + */ + +#include +#include +#include +#include +#include + +#include + +#include "trace.h" + + +#define SIZEOF_BTS (1 << 13) + +static DEFINE_PER_CPU(struct bts_tracer *, tracer); +static DEFINE_PER_CPU(unsigned char[SIZEOF_BTS], buffer); + +#define this_tracer per_cpu(tracer, smp_processor_id()) +#define this_buffer per_cpu(buffer, smp_processor_id()) + + +/* + * Information to interpret a BTS record. + * This will go into an in-kernel BTS interface. + */ +static unsigned char sizeof_field; +static unsigned long debugctl_mask; + +#define sizeof_bts (3 * sizeof_field) + +static void bts_trace_cpuinit(struct cpuinfo_x86 *c) +{ + switch (c->x86) { + case 0x6: + switch (c->x86_model) { + case 0x0 ... 0xC: + break; + case 0xD: + case 0xE: /* Pentium M */ + sizeof_field = sizeof(long); + debugctl_mask = (1<<6)|(1<<7); + break; + default: + sizeof_field = 8; + debugctl_mask = (1<<6)|(1<<7); + break; + } + break; + case 0xF: + switch (c->x86_model) { + case 0x0: + case 0x1: + case 0x2: /* Netburst */ + sizeof_field = sizeof(long); + debugctl_mask = (1<<2)|(1<<3); + break; + default: + /* sorry, don't know about them */ + break; + } + break; + default: + /* sorry, don't know about them */ + break; + } +} + +static inline void bts_enable(void) +{ + unsigned long debugctl; + + rdmsrl(MSR_IA32_DEBUGCTLMSR, debugctl); + wrmsrl(MSR_IA32_DEBUGCTLMSR, debugctl | debugctl_mask); +} + +static inline void bts_disable(void) +{ + unsigned long debugctl; + + rdmsrl(MSR_IA32_DEBUGCTLMSR, debugctl); + wrmsrl(MSR_IA32_DEBUGCTLMSR, debugctl & ~debugctl_mask); +} + +static void bts_trace_reset(struct trace_array *tr) +{ + int cpu; + + tr->time_start = ftrace_now(tr->cpu); + + for_each_online_cpu(cpu) + tracing_reset(tr, cpu); +} + +static void bts_trace_start_cpu(void *arg) +{ + this_tracer = + ds_request_bts(/* task = */ NULL, this_buffer, SIZEOF_BTS, + /* ovfl = */ NULL, /* th = */ (size_t)-1); + if (IS_ERR(this_tracer)) { + this_tracer = NULL; + return; + } + + bts_enable(); +} + +static void bts_trace_start(struct trace_array *tr) +{ + int cpu; + + bts_trace_reset(tr); + + for_each_cpu_mask(cpu, cpu_possible_map) + smp_call_function_single(cpu, bts_trace_start_cpu, NULL, 1); +} + +static void bts_trace_stop_cpu(void *arg) +{ + if (this_tracer) { + bts_disable(); + + ds_release_bts(this_tracer); + this_tracer = NULL; + } +} + +static void bts_trace_stop(struct trace_array *tr) +{ + int cpu; + + for_each_cpu_mask(cpu, cpu_possible_map) + smp_call_function_single(cpu, bts_trace_stop_cpu, NULL, 1); +} + +static int bts_trace_init(struct trace_array *tr) +{ + bts_trace_cpuinit(&boot_cpu_data); + bts_trace_reset(tr); + bts_trace_start(tr); + + return 0; +} + +static void bts_trace_print_header(struct seq_file *m) +{ +#ifdef __i386__ + seq_puts(m, "# CPU# FROM TO FUNCTION\n"); + seq_puts(m, "# | | | |\n"); +#else + seq_puts(m, + "# CPU# FROM TO FUNCTION\n"); + seq_puts(m, + "# | | | |\n"); +#endif +} + +static enum print_line_t bts_trace_print_line(struct trace_iterator *iter) +{ + struct trace_entry *entry = iter->ent; + struct trace_seq *seq = &iter->seq; + struct bts_entry *it; + + trace_assign_type(it, entry); + + if (entry->type == TRACE_BTS) { + int ret; +#ifdef CONFIG_KALLSYMS + char function[KSYM_SYMBOL_LEN]; + sprint_symbol(function, it->from); +#else + char *function = ""; +#endif + + ret = trace_seq_printf(seq, "%4d 0x%lx -> 0x%lx [%s]\n", + entry->cpu, it->from, it->to, function); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE;; + return TRACE_TYPE_HANDLED; + } + return TRACE_TYPE_UNHANDLED; +} + +void trace_bts(struct trace_array *tr, unsigned long from, unsigned long to) +{ + struct ring_buffer_event *event; + struct bts_entry *entry; + unsigned long irq; + + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), &irq); + if (!event) + return; + entry = ring_buffer_event_data(event); + tracing_generic_entry_update(&entry->ent, 0, from); + entry->ent.type = TRACE_BTS; + entry->ent.cpu = smp_processor_id(); + entry->from = from; + entry->to = to; + ring_buffer_unlock_commit(tr->buffer, event, irq); +} + +static void trace_bts_at(struct trace_array *tr, size_t index) +{ + const void *raw = NULL; + unsigned long from, to; + int err; + + err = ds_access_bts(this_tracer, index, &raw); + if (err < 0) + return; + + from = *(const unsigned long *)raw; + to = *(const unsigned long *)((const char *)raw + sizeof_field); + + trace_bts(tr, from, to); +} + +static void trace_bts_cpu(void *arg) +{ + struct trace_array *tr = (struct trace_array *) arg; + size_t index = 0, end = 0, i; + int err; + + if (!this_tracer) + return; + + bts_disable(); + + err = ds_get_bts_index(this_tracer, &index); + if (err < 0) + goto out; + + err = ds_get_bts_end(this_tracer, &end); + if (err < 0) + goto out; + + for (i = index; i < end; i++) + trace_bts_at(tr, i); + + for (i = 0; i < index; i++) + trace_bts_at(tr, i); + +out: + bts_enable(); +} + +static void trace_bts_prepare(struct trace_iterator *iter) +{ + int cpu; + + for_each_cpu_mask(cpu, cpu_possible_map) + smp_call_function_single(cpu, trace_bts_cpu, iter->tr, 1); +} + +struct tracer bts_tracer __read_mostly = +{ + .name = "bts", + .init = bts_trace_init, + .reset = bts_trace_stop, + .print_header = bts_trace_print_header, + .print_line = bts_trace_print_line, + .start = bts_trace_start, + .stop = bts_trace_stop, + .open = trace_bts_prepare +}; + +__init static int init_bts_trace(void) +{ + return register_tracer(&bts_tracer); +} +device_initcall(init_bts_trace); -- cgit v1.2.3 From fb52607afcd0629776f1dc9e657647ceae81dd50 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Tue, 25 Nov 2008 21:07:04 +0100 Subject: tracing/function-return-tracer: change the name into function-graph-tracer Impact: cleanup This patch changes the name of the "return function tracer" into function-graph-tracer which is a more suitable name for a tracing which makes one able to retrieve the ordered call stack during the code flow. Signed-off-by: Frederic Weisbecker Acked-by: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 2 +- arch/x86/include/asm/ftrace.h | 4 +- arch/x86/kernel/Makefile | 4 +- arch/x86/kernel/entry_32.S | 12 ++--- arch/x86/kernel/ftrace.c | 12 ++--- include/linux/ftrace.h | 24 ++++----- include/linux/ftrace_irq.h | 2 +- include/linux/sched.h | 2 +- kernel/Makefile | 2 +- kernel/fork.c | 4 +- kernel/sched.c | 2 +- kernel/trace/Kconfig | 19 ++++--- kernel/trace/Makefile | 2 +- kernel/trace/ftrace.c | 26 +++++----- kernel/trace/trace.c | 18 +++---- kernel/trace/trace.h | 12 ++--- kernel/trace/trace_functions_graph.c | 98 ++++++++++++++++++++++++++++++++++++ 17 files changed, 173 insertions(+), 72 deletions(-) create mode 100644 kernel/trace/trace_functions_graph.c (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index e49a4fd718fe..0842b1127684 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -29,7 +29,7 @@ config X86 select HAVE_FTRACE_MCOUNT_RECORD select HAVE_DYNAMIC_FTRACE select HAVE_FUNCTION_TRACER - select HAVE_FUNCTION_RET_TRACER if X86_32 + select HAVE_FUNCTION_GRAPH_TRACER if X86_32 select HAVE_FUNCTION_TRACE_MCOUNT_TEST select HAVE_KVM if ((X86_32 && !X86_VOYAGER && !X86_VISWS && !X86_NUMAQ) || X86_64) select HAVE_ARCH_KGDB if !X86_VOYAGER diff --git a/arch/x86/include/asm/ftrace.h b/arch/x86/include/asm/ftrace.h index 754a3e082f94..7e61b4ceb9a4 100644 --- a/arch/x86/include/asm/ftrace.h +++ b/arch/x86/include/asm/ftrace.h @@ -28,7 +28,7 @@ struct dyn_arch_ftrace { #endif /* __ASSEMBLY__ */ #endif /* CONFIG_FUNCTION_TRACER */ -#ifdef CONFIG_FUNCTION_RET_TRACER +#ifdef CONFIG_FUNCTION_GRAPH_TRACER #ifndef __ASSEMBLY__ @@ -51,6 +51,6 @@ struct ftrace_ret_stack { extern void return_to_handler(void); #endif /* __ASSEMBLY__ */ -#endif /* CONFIG_FUNCTION_RET_TRACER */ +#endif /* CONFIG_FUNCTION_GRAPH_TRACER */ #endif /* _ASM_X86_FTRACE_H */ diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index af2bc36ca1c4..64939a0c3986 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -14,7 +14,7 @@ CFLAGS_REMOVE_paravirt-spinlocks.o = -pg CFLAGS_REMOVE_ftrace.o = -pg endif -ifdef CONFIG_FUNCTION_RET_TRACER +ifdef CONFIG_FUNCTION_GRAPH_TRACER # Don't trace __switch_to() but let it for function tracer CFLAGS_REMOVE_process_32.o = -pg endif @@ -70,7 +70,7 @@ obj-$(CONFIG_X86_LOCAL_APIC) += apic.o nmi.o obj-$(CONFIG_X86_IO_APIC) += io_apic.o obj-$(CONFIG_X86_REBOOTFIXUPS) += reboot_fixups_32.o obj-$(CONFIG_DYNAMIC_FTRACE) += ftrace.o -obj-$(CONFIG_FUNCTION_RET_TRACER) += ftrace.o +obj-$(CONFIG_FUNCTION_GRAPH_TRACER) += ftrace.o obj-$(CONFIG_KEXEC) += machine_kexec_$(BITS).o obj-$(CONFIG_KEXEC) += relocate_kernel_$(BITS).o crash.o obj-$(CONFIG_CRASH_DUMP) += crash_dump_$(BITS).o diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S index 74defe21ba42..2b1f0f081a6b 100644 --- a/arch/x86/kernel/entry_32.S +++ b/arch/x86/kernel/entry_32.S @@ -1188,9 +1188,9 @@ ENTRY(mcount) cmpl $ftrace_stub, ftrace_trace_function jnz trace -#ifdef CONFIG_FUNCTION_RET_TRACER - cmpl $ftrace_stub, ftrace_function_return - jnz ftrace_return_caller +#ifdef CONFIG_FUNCTION_GRAPH_TRACER + cmpl $ftrace_stub, ftrace_graph_function + jnz ftrace_graph_caller #endif .globl ftrace_stub ftrace_stub: @@ -1215,8 +1215,8 @@ END(mcount) #endif /* CONFIG_DYNAMIC_FTRACE */ #endif /* CONFIG_FUNCTION_TRACER */ -#ifdef CONFIG_FUNCTION_RET_TRACER -ENTRY(ftrace_return_caller) +#ifdef CONFIG_FUNCTION_GRAPH_TRACER +ENTRY(ftrace_graph_caller) cmpl $0, function_trace_stop jne ftrace_stub @@ -1230,7 +1230,7 @@ ENTRY(ftrace_return_caller) popl %ecx popl %eax ret -END(ftrace_return_caller) +END(ftrace_graph_caller) .globl return_to_handler return_to_handler: diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index bb137f7297ed..3595a4c14aba 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -323,7 +323,7 @@ int __init ftrace_dyn_arch_init(void *data) } #endif -#ifdef CONFIG_FUNCTION_RET_TRACER +#ifdef CONFIG_FUNCTION_GRAPH_TRACER #ifndef CONFIG_DYNAMIC_FTRACE @@ -389,11 +389,11 @@ static void pop_return_trace(unsigned long *ret, unsigned long long *time, */ unsigned long ftrace_return_to_handler(void) { - struct ftrace_retfunc trace; + struct ftrace_graph_ret trace; pop_return_trace(&trace.ret, &trace.calltime, &trace.func, &trace.overrun); trace.rettime = cpu_clock(raw_smp_processor_id()); - ftrace_function_return(&trace); + ftrace_graph_function(&trace); return trace.ret; } @@ -440,12 +440,12 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) ); if (WARN_ON(faulted)) { - unregister_ftrace_return(); + unregister_ftrace_graph(); return; } if (WARN_ON(!__kernel_text_address(old))) { - unregister_ftrace_return(); + unregister_ftrace_graph(); *parent = old; return; } @@ -456,4 +456,4 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) *parent = old; } -#endif /* CONFIG_FUNCTION_RET_TRACER */ +#endif /* CONFIG_FUNCTION_GRAPH_TRACER */ diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 7854d87b97b2..b4ac734ad8d6 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -115,8 +115,8 @@ extern int ftrace_update_ftrace_func(ftrace_func_t func); extern void ftrace_caller(void); extern void ftrace_call(void); extern void mcount_call(void); -#ifdef CONFIG_FUNCTION_RET_TRACER -extern void ftrace_return_caller(void); +#ifdef CONFIG_FUNCTION_GRAPH_TRACER +extern void ftrace_graph_caller(void); #endif /** @@ -315,7 +315,7 @@ ftrace_init_module(struct module *mod, /* * Structure that defines a return function trace. */ -struct ftrace_retfunc { +struct ftrace_graph_ret { unsigned long ret; /* Return address */ unsigned long func; /* Current function */ unsigned long long calltime; @@ -324,22 +324,22 @@ struct ftrace_retfunc { unsigned long overrun; }; -#ifdef CONFIG_FUNCTION_RET_TRACER +#ifdef CONFIG_FUNCTION_GRAPH_TRACER #define FTRACE_RETFUNC_DEPTH 50 #define FTRACE_RETSTACK_ALLOC_SIZE 32 /* Type of a callback handler of tracing return function */ -typedef void (*trace_function_return_t)(struct ftrace_retfunc *); +typedef void (*trace_function_graph_t)(struct ftrace_graph_ret *); -extern int register_ftrace_return(trace_function_return_t func); +extern int register_ftrace_graph(trace_function_graph_t func); /* The current handler in use */ -extern trace_function_return_t ftrace_function_return; -extern void unregister_ftrace_return(void); +extern trace_function_graph_t ftrace_graph_function; +extern void unregister_ftrace_graph(void); -extern void ftrace_retfunc_init_task(struct task_struct *t); -extern void ftrace_retfunc_exit_task(struct task_struct *t); +extern void ftrace_graph_init_task(struct task_struct *t); +extern void ftrace_graph_exit_task(struct task_struct *t); #else -static inline void ftrace_retfunc_init_task(struct task_struct *t) { } -static inline void ftrace_retfunc_exit_task(struct task_struct *t) { } +static inline void ftrace_graph_init_task(struct task_struct *t) { } +static inline void ftrace_graph_exit_task(struct task_struct *t) { } #endif #endif /* _LINUX_FTRACE_H */ diff --git a/include/linux/ftrace_irq.h b/include/linux/ftrace_irq.h index 0b4df55d7a74..366a054d0b05 100644 --- a/include/linux/ftrace_irq.h +++ b/include/linux/ftrace_irq.h @@ -2,7 +2,7 @@ #define _LINUX_FTRACE_IRQ_H -#if defined(CONFIG_DYNAMIC_FTRACE) || defined(CONFIG_FUNCTION_RET_TRACER) +#if defined(CONFIG_DYNAMIC_FTRACE) || defined(CONFIG_FUNCTION_GRAPH_TRACER) extern void ftrace_nmi_enter(void); extern void ftrace_nmi_exit(void); #else diff --git a/include/linux/sched.h b/include/linux/sched.h index d02a0ca70ee9..7ad48f2a2758 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1365,7 +1365,7 @@ struct task_struct { unsigned long default_timer_slack_ns; struct list_head *scm_work_list; -#ifdef CONFIG_FUNCTION_RET_TRACER +#ifdef CONFIG_FUNCTION_GRAPH_TRACER /* Index of current stored adress in ret_stack */ int curr_ret_stack; /* Stack of return addresses for return function tracing */ diff --git a/kernel/Makefile b/kernel/Makefile index 03a45e7e87b7..703cf3b7389c 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -21,7 +21,7 @@ CFLAGS_REMOVE_cgroup-debug.o = -pg CFLAGS_REMOVE_sched_clock.o = -pg CFLAGS_REMOVE_sched.o = -pg endif -ifdef CONFIG_FUNCTION_RET_TRACER +ifdef CONFIG_FUNCTION_GRAPH_TRACER CFLAGS_REMOVE_extable.o = -pg # For __kernel_text_address() CFLAGS_REMOVE_module.o = -pg # For __module_text_address() endif diff --git a/kernel/fork.c b/kernel/fork.c index d6e1a3205f62..5f82a999c032 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -140,7 +140,7 @@ void free_task(struct task_struct *tsk) prop_local_destroy_single(&tsk->dirties); free_thread_info(tsk->stack); rt_mutex_debug_task_free(tsk); - ftrace_retfunc_exit_task(tsk); + ftrace_graph_exit_task(tsk); free_task_struct(tsk); } EXPORT_SYMBOL(free_task); @@ -1271,7 +1271,7 @@ static struct task_struct *copy_process(unsigned long clone_flags, total_forks++; spin_unlock(¤t->sighand->siglock); write_unlock_irq(&tasklist_lock); - ftrace_retfunc_init_task(p); + ftrace_graph_init_task(p); proc_fork_connector(p); cgroup_post_fork(p); return p; diff --git a/kernel/sched.c b/kernel/sched.c index 388d9db044ab..52490bf6b884 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -5901,7 +5901,7 @@ void __cpuinit init_idle(struct task_struct *idle, int cpu) * The idle tasks have their own, simple scheduling class: */ idle->sched_class = &idle_sched_class; - ftrace_retfunc_init_task(idle); + ftrace_graph_init_task(idle); } /* diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 620feadff67a..eb9b901e0777 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -12,7 +12,7 @@ config NOP_TRACER config HAVE_FUNCTION_TRACER bool -config HAVE_FUNCTION_RET_TRACER +config HAVE_FUNCTION_GRAPH_TRACER bool config HAVE_FUNCTION_TRACE_MCOUNT_TEST @@ -63,15 +63,18 @@ config FUNCTION_TRACER (the bootup default), then the overhead of the instructions is very small and not measurable even in micro-benchmarks. -config FUNCTION_RET_TRACER - bool "Kernel Function return Tracer" - depends on HAVE_FUNCTION_RET_TRACER +config FUNCTION_GRAPH_TRACER + bool "Kernel Function Graph Tracer" + depends on HAVE_FUNCTION_GRAPH_TRACER depends on FUNCTION_TRACER help - Enable the kernel to trace a function at its return. - It's first purpose is to trace the duration of functions. - This is done by setting the current return address on the thread - info structure of the current task. + Enable the kernel to trace a function at both its return + and its entry. + It's first purpose is to trace the duration of functions and + draw a call graph for each thread with some informations like + the return value. + This is done by setting the current return address on the current + task structure into a stack of calls. config IRQSOFF_TRACER bool "Interrupts-off Latency Tracer" diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index cef4bcb4e822..08c5fe6ddc09 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -29,7 +29,7 @@ obj-$(CONFIG_NOP_TRACER) += trace_nop.o obj-$(CONFIG_STACK_TRACER) += trace_stack.o obj-$(CONFIG_MMIOTRACE) += trace_mmiotrace.o obj-$(CONFIG_BOOT_TRACER) += trace_boot.o -obj-$(CONFIG_FUNCTION_RET_TRACER) += trace_functions_return.o +obj-$(CONFIG_FUNCTION_GRAPH_TRACER) += trace_functions_graph.o obj-$(CONFIG_TRACE_BRANCH_PROFILING) += trace_branch.o obj-$(CONFIG_BTS_TRACER) += trace_bts.o diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 53042f118f23..9e19976af727 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -395,11 +395,11 @@ __ftrace_replace_code(struct dyn_ftrace *rec, int enable) unsigned long ip, fl; unsigned long ftrace_addr; -#ifdef CONFIG_FUNCTION_RET_TRACER +#ifdef CONFIG_FUNCTION_GRAPH_TRACER if (ftrace_tracing_type == FTRACE_TYPE_ENTER) ftrace_addr = (unsigned long)ftrace_caller; else - ftrace_addr = (unsigned long)ftrace_return_caller; + ftrace_addr = (unsigned long)ftrace_graph_caller; #else ftrace_addr = (unsigned long)ftrace_caller; #endif @@ -1496,13 +1496,13 @@ ftrace_enable_sysctl(struct ctl_table *table, int write, return ret; } -#ifdef CONFIG_FUNCTION_RET_TRACER +#ifdef CONFIG_FUNCTION_GRAPH_TRACER static atomic_t ftrace_retfunc_active; /* The callback that hooks the return of a function */ -trace_function_return_t ftrace_function_return = - (trace_function_return_t)ftrace_stub; +trace_function_graph_t ftrace_graph_function = + (trace_function_graph_t)ftrace_stub; /* Try to assign a return stack array on FTRACE_RETSTACK_ALLOC_SIZE tasks. */ @@ -1549,7 +1549,7 @@ free: } /* Allocate a return stack for each task */ -static int start_return_tracing(void) +static int start_graph_tracing(void) { struct ftrace_ret_stack **ret_stack_list; int ret; @@ -1569,7 +1569,7 @@ static int start_return_tracing(void) return ret; } -int register_ftrace_return(trace_function_return_t func) +int register_ftrace_graph(trace_function_graph_t func) { int ret = 0; @@ -1584,13 +1584,13 @@ int register_ftrace_return(trace_function_return_t func) goto out; } atomic_inc(&ftrace_retfunc_active); - ret = start_return_tracing(); + ret = start_graph_tracing(); if (ret) { atomic_dec(&ftrace_retfunc_active); goto out; } ftrace_tracing_type = FTRACE_TYPE_RETURN; - ftrace_function_return = func; + ftrace_graph_function = func; ftrace_startup(); out: @@ -1598,12 +1598,12 @@ out: return ret; } -void unregister_ftrace_return(void) +void unregister_ftrace_graph(void) { mutex_lock(&ftrace_sysctl_lock); atomic_dec(&ftrace_retfunc_active); - ftrace_function_return = (trace_function_return_t)ftrace_stub; + ftrace_graph_function = (trace_function_graph_t)ftrace_stub; ftrace_shutdown(); /* Restore normal tracing type */ ftrace_tracing_type = FTRACE_TYPE_ENTER; @@ -1612,7 +1612,7 @@ void unregister_ftrace_return(void) } /* Allocate a return stack for newly created task */ -void ftrace_retfunc_init_task(struct task_struct *t) +void ftrace_graph_init_task(struct task_struct *t) { if (atomic_read(&ftrace_retfunc_active)) { t->ret_stack = kmalloc(FTRACE_RETFUNC_DEPTH @@ -1626,7 +1626,7 @@ void ftrace_retfunc_init_task(struct task_struct *t) t->ret_stack = NULL; } -void ftrace_retfunc_exit_task(struct task_struct *t) +void ftrace_graph_exit_task(struct task_struct *t) { struct ftrace_ret_stack *ret_stack = t->ret_stack; diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 8df8fdd69c95..f21ab2c68fd4 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -878,15 +878,15 @@ trace_function(struct trace_array *tr, struct trace_array_cpu *data, ring_buffer_unlock_commit(tr->buffer, event, irq_flags); } -#ifdef CONFIG_FUNCTION_RET_TRACER -static void __trace_function_return(struct trace_array *tr, +#ifdef CONFIG_FUNCTION_GRAPH_TRACER +static void __trace_function_graph(struct trace_array *tr, struct trace_array_cpu *data, - struct ftrace_retfunc *trace, + struct ftrace_graph_ret *trace, unsigned long flags, int pc) { struct ring_buffer_event *event; - struct ftrace_ret_entry *entry; + struct ftrace_graph_entry *entry; unsigned long irq_flags; if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled)))) @@ -1177,8 +1177,8 @@ function_trace_call(unsigned long ip, unsigned long parent_ip) local_irq_restore(flags); } -#ifdef CONFIG_FUNCTION_RET_TRACER -void trace_function_return(struct ftrace_retfunc *trace) +#ifdef CONFIG_FUNCTION_GRAPH_TRACER +void trace_function_graph(struct ftrace_graph_ret *trace) { struct trace_array *tr = &global_trace; struct trace_array_cpu *data; @@ -1193,12 +1193,12 @@ void trace_function_return(struct ftrace_retfunc *trace) disabled = atomic_inc_return(&data->disabled); if (likely(disabled == 1)) { pc = preempt_count(); - __trace_function_return(tr, data, trace, flags, pc); + __trace_function_graph(tr, data, trace, flags, pc); } atomic_dec(&data->disabled); raw_local_irq_restore(flags); } -#endif /* CONFIG_FUNCTION_RET_TRACER */ +#endif /* CONFIG_FUNCTION_GRAPH_TRACER */ static struct ftrace_ops trace_ops __read_mostly = { @@ -2001,7 +2001,7 @@ static enum print_line_t print_trace_fmt(struct trace_iterator *iter) break; } case TRACE_FN_RET: { - return print_return_function(iter); + return print_graph_function(iter); break; } case TRACE_BRANCH: { diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 3abd645e8af2..72b5ef868765 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -57,7 +57,7 @@ struct ftrace_entry { }; /* Function return entry */ -struct ftrace_ret_entry { +struct ftrace_graph_entry { struct trace_entry ent; unsigned long ip; unsigned long parent_ip; @@ -264,7 +264,7 @@ extern void __ftrace_bad_type(void); IF_ASSIGN(var, ent, struct trace_boot_call, TRACE_BOOT_CALL);\ IF_ASSIGN(var, ent, struct trace_boot_ret, TRACE_BOOT_RET);\ IF_ASSIGN(var, ent, struct trace_branch, TRACE_BRANCH); \ - IF_ASSIGN(var, ent, struct ftrace_ret_entry, TRACE_FN_RET);\ + IF_ASSIGN(var, ent, struct ftrace_graph_entry, TRACE_FN_RET);\ IF_ASSIGN(var, ent, struct bts_entry, TRACE_BTS);\ __ftrace_bad_type(); \ } while (0) @@ -398,7 +398,7 @@ void trace_function(struct trace_array *tr, unsigned long parent_ip, unsigned long flags, int pc); void -trace_function_return(struct ftrace_retfunc *trace); +trace_function_graph(struct ftrace_graph_ret *trace); void trace_bts(struct trace_array *tr, unsigned long from, @@ -489,11 +489,11 @@ extern int trace_vprintk(unsigned long ip, const char *fmt, va_list args); extern unsigned long trace_flags; /* Standard output formatting function used for function return traces */ -#ifdef CONFIG_FUNCTION_RET_TRACER -extern enum print_line_t print_return_function(struct trace_iterator *iter); +#ifdef CONFIG_FUNCTION_GRAPH_TRACER +extern enum print_line_t print_graph_function(struct trace_iterator *iter); #else static inline enum print_line_t -print_return_function(struct trace_iterator *iter) +print_graph_function(struct trace_iterator *iter) { return TRACE_TYPE_UNHANDLED; } diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c new file mode 100644 index 000000000000..f5bad4624d2b --- /dev/null +++ b/kernel/trace/trace_functions_graph.c @@ -0,0 +1,98 @@ +/* + * + * Function graph tracer. + * Copyright (c) 2008 Frederic Weisbecker + * Mostly borrowed from function tracer which + * is Copyright (c) Steven Rostedt + * + */ +#include +#include +#include +#include + +#include "trace.h" + + +#define TRACE_GRAPH_PRINT_OVERRUN 0x1 +static struct tracer_opt trace_opts[] = { + /* Display overruns or not */ + { TRACER_OPT(overrun, TRACE_GRAPH_PRINT_OVERRUN) }, + { } /* Empty entry */ +}; + +static struct tracer_flags tracer_flags = { + .val = 0, /* Don't display overruns by default */ + .opts = trace_opts +}; + + +static int graph_trace_init(struct trace_array *tr) +{ + int cpu; + for_each_online_cpu(cpu) + tracing_reset(tr, cpu); + + return register_ftrace_graph(&trace_function_graph); +} + +static void graph_trace_reset(struct trace_array *tr) +{ + unregister_ftrace_graph(); +} + + +enum print_line_t +print_graph_function(struct trace_iterator *iter) +{ + struct trace_seq *s = &iter->seq; + struct trace_entry *entry = iter->ent; + struct ftrace_graph_entry *field; + int ret; + + if (entry->type == TRACE_FN_RET) { + trace_assign_type(field, entry); + ret = trace_seq_printf(s, "%pF -> ", (void *)field->parent_ip); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + ret = seq_print_ip_sym(s, field->ip, + trace_flags & TRACE_ITER_SYM_MASK); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + ret = trace_seq_printf(s, " (%llu ns)", + field->rettime - field->calltime); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + if (tracer_flags.val & TRACE_GRAPH_PRINT_OVERRUN) { + ret = trace_seq_printf(s, " (Overruns: %lu)", + field->overrun); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + } + + ret = trace_seq_printf(s, "\n"); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + return TRACE_TYPE_HANDLED; + } + return TRACE_TYPE_UNHANDLED; +} + +static struct tracer graph_trace __read_mostly = { + .name = "function-graph", + .init = graph_trace_init, + .reset = graph_trace_reset, + .print_line = print_graph_function, + .flags = &tracer_flags, +}; + +static __init int init_graph_trace(void) +{ + return register_tracer(&graph_trace); +} + +device_initcall(init_graph_trace); -- cgit v1.2.3 From 287b6e68ca7209caec40b2f44f837c580a413bae Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 26 Nov 2008 00:57:25 +0100 Subject: tracing/function-return-tracer: set a more human readable output Impact: feature This patch sets a C-like output for the function graph tracing. For this aim, we now call two handler for each function: one on the entry and one other on return. This way we can draw a well-ordered call stack. The pid of the previous trace is loosely stored to be compared against the one of the current trace to see if there were a context switch. Without this little feature, the call tree would seem broken at some locations. We could use the sched_tracer to capture these sched_events but this way of processing is much more simpler. 2 spaces have been chosen for indentation to fit the screen while deep calls. The time of execution in nanosecs is printed just after closed braces, it seems more easy this way to find the corresponding function. If the time was printed as a first column, it would be not so easy to find the corresponding function if it is called on a deep depth. I plan to output the return value but on 32 bits CPU, the return value can be 32 or 64, and its difficult to guess on which case we are. I don't know what would be the better solution on X86-32: only print eax (low-part) or even edx (high-part). Actually it's thee same problem when a function return a 8 bits value, the high part of eax could contain junk values... Here is an example of trace: sys_read() { fget_light() { } 526 vfs_read() { rw_verify_area() { security_file_permission() { cap_file_permission() { } 519 } 1564 } 2640 do_sync_read() { pipe_read() { __might_sleep() { } 511 pipe_wait() { prepare_to_wait() { } 760 deactivate_task() { dequeue_task() { dequeue_task_fair() { dequeue_entity() { update_curr() { update_min_vruntime() { } 504 } 1587 clear_buddies() { } 512 add_cfs_task_weight() { } 519 update_min_vruntime() { } 511 } 5602 dequeue_entity() { update_curr() { update_min_vruntime() { } 496 } 1631 clear_buddies() { } 496 update_min_vruntime() { } 527 } 4580 hrtick_update() { hrtick_start_fair() { } 488 } 1489 } 13700 } 14949 } 16016 msecs_to_jiffies() { } 496 put_prev_task_fair() { } 504 pick_next_task_fair() { } 489 pick_next_task_rt() { } 496 pick_next_task_fair() { } 489 pick_next_task_idle() { } 489 ------------8<---------- thread 4 ------------8<---------- finish_task_switch() { } 1203 do_softirq() { __do_softirq() { __local_bh_disable() { } 669 rcu_process_callbacks() { __rcu_process_callbacks() { cpu_quiet() { rcu_start_batch() { } 503 } 1647 } 3128 __rcu_process_callbacks() { } 542 } 5362 _local_bh_enable() { } 587 } 8880 } 9986 kthread_should_stop() { } 669 deactivate_task() { dequeue_task() { dequeue_task_fair() { dequeue_entity() { update_curr() { calc_delta_mine() { } 511 update_min_vruntime() { } 511 } 2813 Signed-off-by: Frederic Weisbecker Acked-by: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/x86/kernel/ftrace.c | 32 +++++++---- include/linux/ftrace.h | 25 ++++++-- kernel/trace/ftrace.c | 30 +++++----- kernel/trace/trace.c | 67 ++++++++++++++++++---- kernel/trace/trace.h | 28 +++++---- kernel/trace/trace_functions_graph.c | 104 ++++++++++++++++++++++++++-------- kernel/trace/trace_functions_return.c | 98 -------------------------------- 7 files changed, 208 insertions(+), 176 deletions(-) delete mode 100644 kernel/trace/trace_functions_return.c (limited to 'arch/x86') diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 3595a4c14aba..26b2d92d48b3 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -347,7 +347,7 @@ void ftrace_nmi_exit(void) /* Add a function return address to the trace stack on thread info.*/ static int push_return_trace(unsigned long ret, unsigned long long time, - unsigned long func) + unsigned long func, int *depth) { int index; @@ -365,21 +365,22 @@ static int push_return_trace(unsigned long ret, unsigned long long time, current->ret_stack[index].ret = ret; current->ret_stack[index].func = func; current->ret_stack[index].calltime = time; + *depth = index; return 0; } /* Retrieve a function return address to the trace stack on thread info.*/ -static void pop_return_trace(unsigned long *ret, unsigned long long *time, - unsigned long *func, unsigned long *overrun) +static void pop_return_trace(struct ftrace_graph_ret *trace, unsigned long *ret) { int index; index = current->curr_ret_stack; *ret = current->ret_stack[index].ret; - *func = current->ret_stack[index].func; - *time = current->ret_stack[index].calltime; - *overrun = atomic_read(¤t->trace_overrun); + trace->func = current->ret_stack[index].func; + trace->calltime = current->ret_stack[index].calltime; + trace->overrun = atomic_read(¤t->trace_overrun); + trace->depth = index; current->curr_ret_stack--; } @@ -390,12 +391,13 @@ static void pop_return_trace(unsigned long *ret, unsigned long long *time, unsigned long ftrace_return_to_handler(void) { struct ftrace_graph_ret trace; - pop_return_trace(&trace.ret, &trace.calltime, &trace.func, - &trace.overrun); + unsigned long ret; + + pop_return_trace(&trace, &ret); trace.rettime = cpu_clock(raw_smp_processor_id()); - ftrace_graph_function(&trace); + ftrace_graph_return(&trace); - return trace.ret; + return ret; } /* @@ -407,6 +409,7 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) unsigned long old; unsigned long long calltime; int faulted; + struct ftrace_graph_ent trace; unsigned long return_hooker = (unsigned long) &return_to_handler; @@ -452,8 +455,15 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) calltime = cpu_clock(raw_smp_processor_id()); - if (push_return_trace(old, calltime, self_addr) == -EBUSY) + if (push_return_trace(old, calltime, + self_addr, &trace.depth) == -EBUSY) { *parent = old; + return; + } + + trace.func = self_addr; + ftrace_graph_entry(&trace); + } #endif /* CONFIG_FUNCTION_GRAPH_TRACER */ diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index b4ac734ad8d6..fc2d54987198 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -312,27 +312,40 @@ ftrace_init_module(struct module *mod, #endif +/* + * Structure that defines an entry function trace. + */ +struct ftrace_graph_ent { + unsigned long func; /* Current function */ + int depth; +}; + /* * Structure that defines a return function trace. */ struct ftrace_graph_ret { - unsigned long ret; /* Return address */ unsigned long func; /* Current function */ unsigned long long calltime; unsigned long long rettime; /* Number of functions that overran the depth limit for current task */ unsigned long overrun; + int depth; }; #ifdef CONFIG_FUNCTION_GRAPH_TRACER #define FTRACE_RETFUNC_DEPTH 50 #define FTRACE_RETSTACK_ALLOC_SIZE 32 -/* Type of a callback handler of tracing return function */ -typedef void (*trace_function_graph_t)(struct ftrace_graph_ret *); +/* Type of the callback handlers for tracing function graph*/ +typedef void (*trace_func_graph_ret_t)(struct ftrace_graph_ret *); /* return */ +typedef void (*trace_func_graph_ent_t)(struct ftrace_graph_ent *); /* entry */ + +extern int register_ftrace_graph(trace_func_graph_ret_t retfunc, + trace_func_graph_ent_t entryfunc); + +/* The current handlers in use */ +extern trace_func_graph_ret_t ftrace_graph_return; +extern trace_func_graph_ent_t ftrace_graph_entry; -extern int register_ftrace_graph(trace_function_graph_t func); -/* The current handler in use */ -extern trace_function_graph_t ftrace_graph_function; extern void unregister_ftrace_graph(void); extern void ftrace_graph_init_task(struct task_struct *t); diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 9e19976af727..7e2d3b91692d 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -1498,12 +1498,13 @@ ftrace_enable_sysctl(struct ctl_table *table, int write, #ifdef CONFIG_FUNCTION_GRAPH_TRACER -static atomic_t ftrace_retfunc_active; - -/* The callback that hooks the return of a function */ -trace_function_graph_t ftrace_graph_function = - (trace_function_graph_t)ftrace_stub; +static atomic_t ftrace_graph_active; +/* The callbacks that hook a function */ +trace_func_graph_ret_t ftrace_graph_return = + (trace_func_graph_ret_t)ftrace_stub; +trace_func_graph_ent_t ftrace_graph_entry = + (trace_func_graph_ent_t)ftrace_stub; /* Try to assign a return stack array on FTRACE_RETSTACK_ALLOC_SIZE tasks. */ static int alloc_retstack_tasklist(struct ftrace_ret_stack **ret_stack_list) @@ -1569,7 +1570,8 @@ static int start_graph_tracing(void) return ret; } -int register_ftrace_graph(trace_function_graph_t func) +int register_ftrace_graph(trace_func_graph_ret_t retfunc, + trace_func_graph_ent_t entryfunc) { int ret = 0; @@ -1583,14 +1585,15 @@ int register_ftrace_graph(trace_function_graph_t func) ret = -EBUSY; goto out; } - atomic_inc(&ftrace_retfunc_active); + atomic_inc(&ftrace_graph_active); ret = start_graph_tracing(); if (ret) { - atomic_dec(&ftrace_retfunc_active); + atomic_dec(&ftrace_graph_active); goto out; } ftrace_tracing_type = FTRACE_TYPE_RETURN; - ftrace_graph_function = func; + ftrace_graph_return = retfunc; + ftrace_graph_entry = entryfunc; ftrace_startup(); out: @@ -1602,8 +1605,9 @@ void unregister_ftrace_graph(void) { mutex_lock(&ftrace_sysctl_lock); - atomic_dec(&ftrace_retfunc_active); - ftrace_graph_function = (trace_function_graph_t)ftrace_stub; + atomic_dec(&ftrace_graph_active); + ftrace_graph_return = (trace_func_graph_ret_t)ftrace_stub; + ftrace_graph_entry = (trace_func_graph_ent_t)ftrace_stub; ftrace_shutdown(); /* Restore normal tracing type */ ftrace_tracing_type = FTRACE_TYPE_ENTER; @@ -1614,7 +1618,7 @@ void unregister_ftrace_graph(void) /* Allocate a return stack for newly created task */ void ftrace_graph_init_task(struct task_struct *t) { - if (atomic_read(&ftrace_retfunc_active)) { + if (atomic_read(&ftrace_graph_active)) { t->ret_stack = kmalloc(FTRACE_RETFUNC_DEPTH * sizeof(struct ftrace_ret_stack), GFP_KERNEL); @@ -1638,5 +1642,3 @@ void ftrace_graph_exit_task(struct task_struct *t) } #endif - - diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index f21ab2c68fd4..9d5f7c94f251 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -879,14 +879,38 @@ trace_function(struct trace_array *tr, struct trace_array_cpu *data, } #ifdef CONFIG_FUNCTION_GRAPH_TRACER -static void __trace_function_graph(struct trace_array *tr, +static void __trace_graph_entry(struct trace_array *tr, + struct trace_array_cpu *data, + struct ftrace_graph_ent *trace, + unsigned long flags, + int pc) +{ + struct ring_buffer_event *event; + struct ftrace_graph_ent_entry *entry; + unsigned long irq_flags; + + if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled)))) + return; + + event = ring_buffer_lock_reserve(global_trace.buffer, sizeof(*entry), + &irq_flags); + if (!event) + return; + entry = ring_buffer_event_data(event); + tracing_generic_entry_update(&entry->ent, flags, pc); + entry->ent.type = TRACE_GRAPH_ENT; + entry->graph_ent = *trace; + ring_buffer_unlock_commit(global_trace.buffer, event, irq_flags); +} + +static void __trace_graph_return(struct trace_array *tr, struct trace_array_cpu *data, struct ftrace_graph_ret *trace, unsigned long flags, int pc) { struct ring_buffer_event *event; - struct ftrace_graph_entry *entry; + struct ftrace_graph_ret_entry *entry; unsigned long irq_flags; if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled)))) @@ -898,12 +922,8 @@ static void __trace_function_graph(struct trace_array *tr, return; entry = ring_buffer_event_data(event); tracing_generic_entry_update(&entry->ent, flags, pc); - entry->ent.type = TRACE_FN_RET; - entry->ip = trace->func; - entry->parent_ip = trace->ret; - entry->rettime = trace->rettime; - entry->calltime = trace->calltime; - entry->overrun = trace->overrun; + entry->ent.type = TRACE_GRAPH_RET; + entry->ret = *trace; ring_buffer_unlock_commit(global_trace.buffer, event, irq_flags); } #endif @@ -1178,7 +1198,7 @@ function_trace_call(unsigned long ip, unsigned long parent_ip) } #ifdef CONFIG_FUNCTION_GRAPH_TRACER -void trace_function_graph(struct ftrace_graph_ret *trace) +void trace_graph_entry(struct ftrace_graph_ent *trace) { struct trace_array *tr = &global_trace; struct trace_array_cpu *data; @@ -1193,7 +1213,28 @@ void trace_function_graph(struct ftrace_graph_ret *trace) disabled = atomic_inc_return(&data->disabled); if (likely(disabled == 1)) { pc = preempt_count(); - __trace_function_graph(tr, data, trace, flags, pc); + __trace_graph_entry(tr, data, trace, flags, pc); + } + atomic_dec(&data->disabled); + raw_local_irq_restore(flags); +} + +void trace_graph_return(struct ftrace_graph_ret *trace) +{ + struct trace_array *tr = &global_trace; + struct trace_array_cpu *data; + unsigned long flags; + long disabled; + int cpu; + int pc; + + raw_local_irq_save(flags); + cpu = raw_smp_processor_id(); + data = tr->data[cpu]; + disabled = atomic_inc_return(&data->disabled); + if (likely(disabled == 1)) { + pc = preempt_count(); + __trace_graph_return(tr, data, trace, flags, pc); } atomic_dec(&data->disabled); raw_local_irq_restore(flags); @@ -2000,9 +2041,11 @@ static enum print_line_t print_trace_fmt(struct trace_iterator *iter) trace_seq_print_cont(s, iter); break; } - case TRACE_FN_RET: { + case TRACE_GRAPH_RET: { + return print_graph_function(iter); + } + case TRACE_GRAPH_ENT: { return print_graph_function(iter); - break; } case TRACE_BRANCH: { struct trace_branch *field; diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 72b5ef868765..ffe1bb1eb620 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -25,7 +25,8 @@ enum trace_type { TRACE_BRANCH, TRACE_BOOT_CALL, TRACE_BOOT_RET, - TRACE_FN_RET, + TRACE_GRAPH_RET, + TRACE_GRAPH_ENT, TRACE_USER_STACK, TRACE_BTS, @@ -56,14 +57,16 @@ struct ftrace_entry { unsigned long parent_ip; }; +/* Function call entry */ +struct ftrace_graph_ent_entry { + struct trace_entry ent; + struct ftrace_graph_ent graph_ent; +}; + /* Function return entry */ -struct ftrace_graph_entry { - struct trace_entry ent; - unsigned long ip; - unsigned long parent_ip; - unsigned long long calltime; - unsigned long long rettime; - unsigned long overrun; +struct ftrace_graph_ret_entry { + struct trace_entry ent; + struct ftrace_graph_ret ret; }; extern struct tracer boot_tracer; @@ -264,7 +267,10 @@ extern void __ftrace_bad_type(void); IF_ASSIGN(var, ent, struct trace_boot_call, TRACE_BOOT_CALL);\ IF_ASSIGN(var, ent, struct trace_boot_ret, TRACE_BOOT_RET);\ IF_ASSIGN(var, ent, struct trace_branch, TRACE_BRANCH); \ - IF_ASSIGN(var, ent, struct ftrace_graph_entry, TRACE_FN_RET);\ + IF_ASSIGN(var, ent, struct ftrace_graph_ent_entry, \ + TRACE_GRAPH_ENT); \ + IF_ASSIGN(var, ent, struct ftrace_graph_ret_entry, \ + TRACE_GRAPH_RET); \ IF_ASSIGN(var, ent, struct bts_entry, TRACE_BTS);\ __ftrace_bad_type(); \ } while (0) @@ -397,9 +403,9 @@ void trace_function(struct trace_array *tr, unsigned long ip, unsigned long parent_ip, unsigned long flags, int pc); -void -trace_function_graph(struct ftrace_graph_ret *trace); +void trace_graph_return(struct ftrace_graph_ret *trace); +void trace_graph_entry(struct ftrace_graph_ent *trace); void trace_bts(struct trace_array *tr, unsigned long from, unsigned long to); diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index f5bad4624d2b..b6f0cc2a00cb 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -13,6 +13,7 @@ #include "trace.h" +#define TRACE_GRAPH_INDENT 2 #define TRACE_GRAPH_PRINT_OVERRUN 0x1 static struct tracer_opt trace_opts[] = { @@ -26,6 +27,8 @@ static struct tracer_flags tracer_flags = { .opts = trace_opts }; +/* pid on the last trace processed */ +static pid_t last_pid = -1; static int graph_trace_init(struct trace_array *tr) { @@ -33,7 +36,8 @@ static int graph_trace_init(struct trace_array *tr) for_each_online_cpu(cpu) tracing_reset(tr, cpu); - return register_ftrace_graph(&trace_function_graph); + return register_ftrace_graph(&trace_graph_return, + &trace_graph_entry); } static void graph_trace_reset(struct trace_array *tr) @@ -41,45 +45,97 @@ static void graph_trace_reset(struct trace_array *tr) unregister_ftrace_graph(); } +/* If the pid changed since the last trace, output this event */ +static int verif_pid(struct trace_seq *s, pid_t pid) +{ + if (last_pid != -1 && last_pid == pid) + return 1; -enum print_line_t -print_graph_function(struct trace_iterator *iter) + last_pid = pid; + return trace_seq_printf(s, "\n------------8<---------- thread %d" + " ------------8<----------\n\n", + pid); +} + +static enum print_line_t +print_graph_entry(struct ftrace_graph_ent *call, struct trace_seq *s, + struct trace_entry *ent) { - struct trace_seq *s = &iter->seq; - struct trace_entry *entry = iter->ent; - struct ftrace_graph_entry *field; + int i; int ret; - if (entry->type == TRACE_FN_RET) { - trace_assign_type(field, entry); - ret = trace_seq_printf(s, "%pF -> ", (void *)field->parent_ip); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; + if (!verif_pid(s, ent->pid)) + return TRACE_TYPE_PARTIAL_LINE; - ret = seq_print_ip_sym(s, field->ip, - trace_flags & TRACE_ITER_SYM_MASK); + for (i = 0; i < call->depth * TRACE_GRAPH_INDENT; i++) { + ret = trace_seq_printf(s, " "); if (!ret) return TRACE_TYPE_PARTIAL_LINE; + } + + ret = seq_print_ip_sym(s, call->func, 0); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + ret = trace_seq_printf(s, "() {\n"); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + return TRACE_TYPE_HANDLED; +} + +static enum print_line_t +print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s, + struct trace_entry *ent) +{ + int i; + int ret; + + if (!verif_pid(s, ent->pid)) + return TRACE_TYPE_PARTIAL_LINE; - ret = trace_seq_printf(s, " (%llu ns)", - field->rettime - field->calltime); + for (i = 0; i < trace->depth * TRACE_GRAPH_INDENT; i++) { + ret = trace_seq_printf(s, " "); if (!ret) return TRACE_TYPE_PARTIAL_LINE; + } + + ret = trace_seq_printf(s, "} "); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; - if (tracer_flags.val & TRACE_GRAPH_PRINT_OVERRUN) { - ret = trace_seq_printf(s, " (Overruns: %lu)", - field->overrun); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - } + ret = trace_seq_printf(s, "%llu\n", trace->rettime - trace->calltime); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; - ret = trace_seq_printf(s, "\n"); + if (tracer_flags.val & TRACE_GRAPH_PRINT_OVERRUN) { + ret = trace_seq_printf(s, " (Overruns: %lu)\n", + trace->overrun); if (!ret) return TRACE_TYPE_PARTIAL_LINE; + } + return TRACE_TYPE_HANDLED; +} + +enum print_line_t +print_graph_function(struct trace_iterator *iter) +{ + struct trace_seq *s = &iter->seq; + struct trace_entry *entry = iter->ent; - return TRACE_TYPE_HANDLED; + switch (entry->type) { + case TRACE_GRAPH_ENT: { + struct ftrace_graph_ent_entry *field; + trace_assign_type(field, entry); + return print_graph_entry(&field->graph_ent, s, entry); + } + case TRACE_GRAPH_RET: { + struct ftrace_graph_ret_entry *field; + trace_assign_type(field, entry); + return print_graph_return(&field->ret, s, entry); + } + default: + return TRACE_TYPE_UNHANDLED; } - return TRACE_TYPE_UNHANDLED; } static struct tracer graph_trace __read_mostly = { diff --git a/kernel/trace/trace_functions_return.c b/kernel/trace/trace_functions_return.c deleted file mode 100644 index e00d64509c9c..000000000000 --- a/kernel/trace/trace_functions_return.c +++ /dev/null @@ -1,98 +0,0 @@ -/* - * - * Function return tracer. - * Copyright (c) 2008 Frederic Weisbecker - * Mostly borrowed from function tracer which - * is Copyright (c) Steven Rostedt - * - */ -#include -#include -#include -#include - -#include "trace.h" - - -#define TRACE_RETURN_PRINT_OVERRUN 0x1 -static struct tracer_opt trace_opts[] = { - /* Display overruns or not */ - { TRACER_OPT(overrun, TRACE_RETURN_PRINT_OVERRUN) }, - { } /* Empty entry */ -}; - -static struct tracer_flags tracer_flags = { - .val = 0, /* Don't display overruns by default */ - .opts = trace_opts -}; - - -static int return_trace_init(struct trace_array *tr) -{ - int cpu; - for_each_online_cpu(cpu) - tracing_reset(tr, cpu); - - return register_ftrace_return(&trace_function_return); -} - -static void return_trace_reset(struct trace_array *tr) -{ - unregister_ftrace_return(); -} - - -enum print_line_t -print_return_function(struct trace_iterator *iter) -{ - struct trace_seq *s = &iter->seq; - struct trace_entry *entry = iter->ent; - struct ftrace_ret_entry *field; - int ret; - - if (entry->type == TRACE_FN_RET) { - trace_assign_type(field, entry); - ret = trace_seq_printf(s, "%pF -> ", (void *)field->parent_ip); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - - ret = seq_print_ip_sym(s, field->ip, - trace_flags & TRACE_ITER_SYM_MASK); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - - ret = trace_seq_printf(s, " (%llu ns)", - field->rettime - field->calltime); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - - if (tracer_flags.val & TRACE_RETURN_PRINT_OVERRUN) { - ret = trace_seq_printf(s, " (Overruns: %lu)", - field->overrun); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - } - - ret = trace_seq_printf(s, "\n"); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - - return TRACE_TYPE_HANDLED; - } - return TRACE_TYPE_UNHANDLED; -} - -static struct tracer return_trace __read_mostly = { - .name = "return", - .init = return_trace_init, - .reset = return_trace_reset, - .print_line = print_return_function, - .flags = &tracer_flags, -}; - -static __init int init_return_trace(void) -{ - return register_tracer(&return_trace); -} - -device_initcall(init_return_trace); -- cgit v1.2.3 From c2324b694fa8ffee382a124198c68754088e483c Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Wed, 26 Nov 2008 03:10:01 +0100 Subject: tracing: function graph tracer, fix fix return-tracer => graph-tracer namespace rename fallout. Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_32.S | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S index 2b1f0f081a6b..7def9fd5c1e6 100644 --- a/arch/x86/kernel/entry_32.S +++ b/arch/x86/kernel/entry_32.S @@ -1189,7 +1189,7 @@ ENTRY(mcount) cmpl $ftrace_stub, ftrace_trace_function jnz trace #ifdef CONFIG_FUNCTION_GRAPH_TRACER - cmpl $ftrace_stub, ftrace_graph_function + cmpl $ftrace_stub, ftrace_graph_return jnz ftrace_graph_caller #endif .globl ftrace_stub -- cgit v1.2.3 From 2601657d223d82053d4e1fe1063091401e6b860a Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Mon, 24 Nov 2008 18:21:37 -0800 Subject: x86: signal: move {setup|restore}_sigcontext() Impact: cleanup Move {setup|restore}_sigcontext() declaration onto head of file. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/kernel/signal_32.c | 271 ++++++++++++++++++++++---------------------- arch/x86/kernel/signal_64.c | 148 ++++++++++++------------ 2 files changed, 210 insertions(+), 209 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/signal_32.c b/arch/x86/kernel/signal_32.c index f7dd6c44c042..b3f30d2a2178 100644 --- a/arch/x86/kernel/signal_32.c +++ b/arch/x86/kernel/signal_32.c @@ -70,6 +70,142 @@ static const struct { 0 }; +#define COPY(x) { \ + err |= __get_user(regs->x, &sc->x); \ +} + +#define COPY_SEG(seg) { \ + unsigned short tmp; \ + err |= __get_user(tmp, &sc->seg); \ + regs->seg = tmp; \ +} + +#define COPY_SEG_CPL3(seg) { \ + unsigned short tmp; \ + err |= __get_user(tmp, &sc->seg); \ + regs->seg = tmp | 3; \ +} + +#define GET_SEG(seg) { \ + unsigned short tmp; \ + err |= __get_user(tmp, &sc->seg); \ + loadsegment(seg, tmp); \ +} + +static int +restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc, + unsigned long *pax) +{ + void __user *buf; + unsigned int tmpflags; + unsigned int err = 0; + + /* Always make any pending restarted system calls return -EINTR */ + current_thread_info()->restart_block.fn = do_no_restart_syscall; + +#ifdef CONFIG_X86_32 + GET_SEG(gs); + COPY_SEG(fs); + COPY_SEG(es); + COPY_SEG(ds); +#endif /* CONFIG_X86_32 */ + + COPY(di); COPY(si); COPY(bp); COPY(sp); COPY(bx); + COPY(dx); COPY(cx); COPY(ip); + +#ifdef CONFIG_X86_64 + COPY(r8); + COPY(r9); + COPY(r10); + COPY(r11); + COPY(r12); + COPY(r13); + COPY(r14); + COPY(r15); +#endif /* CONFIG_X86_64 */ + +#ifdef CONFIG_X86_32 + COPY_SEG_CPL3(cs); + COPY_SEG_CPL3(ss); +#else /* !CONFIG_X86_32 */ + /* Kernel saves and restores only the CS segment register on signals, + * which is the bare minimum needed to allow mixed 32/64-bit code. + * App's signal handler can save/restore other segments if needed. */ + COPY_SEG_CPL3(cs); +#endif /* CONFIG_X86_32 */ + + err |= __get_user(tmpflags, &sc->flags); + regs->flags = (regs->flags & ~FIX_EFLAGS) | (tmpflags & FIX_EFLAGS); + regs->orig_ax = -1; /* disable syscall checks */ + + err |= __get_user(buf, &sc->fpstate); + err |= restore_i387_xstate(buf); + + err |= __get_user(*pax, &sc->ax); + return err; +} + +static int +setup_sigcontext(struct sigcontext __user *sc, void __user *fpstate, + struct pt_regs *regs, unsigned long mask) +{ + int err = 0; + +#ifdef CONFIG_X86_32 + { + unsigned int tmp; + + savesegment(gs, tmp); + err |= __put_user(tmp, (unsigned int __user *)&sc->gs); + } + err |= __put_user(regs->fs, (unsigned int __user *)&sc->fs); + err |= __put_user(regs->es, (unsigned int __user *)&sc->es); + err |= __put_user(regs->ds, (unsigned int __user *)&sc->ds); +#endif /* CONFIG_X86_32 */ + + err |= __put_user(regs->di, &sc->di); + err |= __put_user(regs->si, &sc->si); + err |= __put_user(regs->bp, &sc->bp); + err |= __put_user(regs->sp, &sc->sp); + err |= __put_user(regs->bx, &sc->bx); + err |= __put_user(regs->dx, &sc->dx); + err |= __put_user(regs->cx, &sc->cx); + err |= __put_user(regs->ax, &sc->ax); +#ifdef CONFIG_X86_64 + err |= __put_user(regs->r8, &sc->r8); + err |= __put_user(regs->r9, &sc->r9); + err |= __put_user(regs->r10, &sc->r10); + err |= __put_user(regs->r11, &sc->r11); + err |= __put_user(regs->r12, &sc->r12); + err |= __put_user(regs->r13, &sc->r13); + err |= __put_user(regs->r14, &sc->r14); + err |= __put_user(regs->r15, &sc->r15); +#endif /* CONFIG_X86_64 */ + + err |= __put_user(current->thread.trap_no, &sc->trapno); + err |= __put_user(current->thread.error_code, &sc->err); + err |= __put_user(regs->ip, &sc->ip); +#ifdef CONFIG_X86_32 + err |= __put_user(regs->cs, (unsigned int __user *)&sc->cs); + err |= __put_user(regs->flags, &sc->flags); + err |= __put_user(regs->sp, &sc->sp_at_signal); + err |= __put_user(regs->ss, (unsigned int __user *)&sc->ss); +#else /* !CONFIG_X86_32 */ + err |= __put_user(regs->flags, &sc->flags); + err |= __put_user(regs->cs, &sc->cs); + err |= __put_user(0, &sc->gs); + err |= __put_user(0, &sc->fs); +#endif /* CONFIG_X86_32 */ + + err |= __put_user(fpstate, &sc->fpstate); + + /* non-iBCS2 extensions.. */ + err |= __put_user(mask, &sc->oldmask); + err |= __put_user(current->thread.cr2, &sc->cr2); + + return err; +} + /* * Atomically swap in the new signal mask, and wait for a signal. */ @@ -147,84 +283,9 @@ sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss, } #endif /* CONFIG_X86_32 */ -#define COPY(x) { \ - err |= __get_user(regs->x, &sc->x); \ -} - -#define COPY_SEG(seg) { \ - unsigned short tmp; \ - err |= __get_user(tmp, &sc->seg); \ - regs->seg = tmp; \ -} - -#define COPY_SEG_CPL3(seg) { \ - unsigned short tmp; \ - err |= __get_user(tmp, &sc->seg); \ - regs->seg = tmp | 3; \ -} - -#define GET_SEG(seg) { \ - unsigned short tmp; \ - err |= __get_user(tmp, &sc->seg); \ - loadsegment(seg, tmp); \ -} - /* * Do a signal return; undo the signal stack. */ -static int -restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc, - unsigned long *pax) -{ - void __user *buf; - unsigned int tmpflags; - unsigned int err = 0; - - /* Always make any pending restarted system calls return -EINTR */ - current_thread_info()->restart_block.fn = do_no_restart_syscall; - -#ifdef CONFIG_X86_32 - GET_SEG(gs); - COPY_SEG(fs); - COPY_SEG(es); - COPY_SEG(ds); -#endif /* CONFIG_X86_32 */ - - COPY(di); COPY(si); COPY(bp); COPY(sp); COPY(bx); - COPY(dx); COPY(cx); COPY(ip); - -#ifdef CONFIG_X86_64 - COPY(r8); - COPY(r9); - COPY(r10); - COPY(r11); - COPY(r12); - COPY(r13); - COPY(r14); - COPY(r15); -#endif /* CONFIG_X86_64 */ - -#ifdef CONFIG_X86_32 - COPY_SEG_CPL3(cs); - COPY_SEG_CPL3(ss); -#else /* !CONFIG_X86_32 */ - /* Kernel saves and restores only the CS segment register on signals, - * which is the bare minimum needed to allow mixed 32/64-bit code. - * App's signal handler can save/restore other segments if needed. */ - COPY_SEG_CPL3(cs); -#endif /* CONFIG_X86_32 */ - - err |= __get_user(tmpflags, &sc->flags); - regs->flags = (regs->flags & ~FIX_EFLAGS) | (tmpflags & FIX_EFLAGS); - regs->orig_ax = -1; /* disable syscall checks */ - - err |= __get_user(buf, &sc->fpstate); - err |= restore_i387_xstate(buf); - - err |= __get_user(*pax, &sc->ax); - return err; -} - asmlinkage unsigned long sys_sigreturn(unsigned long __unused) { struct sigframe __user *frame; @@ -316,66 +377,6 @@ asmlinkage long sys_rt_sigreturn(struct pt_regs *regs) /* * Set up a signal frame. */ -static int -setup_sigcontext(struct sigcontext __user *sc, void __user *fpstate, - struct pt_regs *regs, unsigned long mask) -{ - int err = 0; - -#ifdef CONFIG_X86_32 - { - unsigned int tmp; - - savesegment(gs, tmp); - err |= __put_user(tmp, (unsigned int __user *)&sc->gs); - } - err |= __put_user(regs->fs, (unsigned int __user *)&sc->fs); - err |= __put_user(regs->es, (unsigned int __user *)&sc->es); - err |= __put_user(regs->ds, (unsigned int __user *)&sc->ds); -#endif /* CONFIG_X86_32 */ - - err |= __put_user(regs->di, &sc->di); - err |= __put_user(regs->si, &sc->si); - err |= __put_user(regs->bp, &sc->bp); - err |= __put_user(regs->sp, &sc->sp); - err |= __put_user(regs->bx, &sc->bx); - err |= __put_user(regs->dx, &sc->dx); - err |= __put_user(regs->cx, &sc->cx); - err |= __put_user(regs->ax, &sc->ax); -#ifdef CONFIG_X86_64 - err |= __put_user(regs->r8, &sc->r8); - err |= __put_user(regs->r9, &sc->r9); - err |= __put_user(regs->r10, &sc->r10); - err |= __put_user(regs->r11, &sc->r11); - err |= __put_user(regs->r12, &sc->r12); - err |= __put_user(regs->r13, &sc->r13); - err |= __put_user(regs->r14, &sc->r14); - err |= __put_user(regs->r15, &sc->r15); -#endif /* CONFIG_X86_64 */ - - err |= __put_user(current->thread.trap_no, &sc->trapno); - err |= __put_user(current->thread.error_code, &sc->err); - err |= __put_user(regs->ip, &sc->ip); -#ifdef CONFIG_X86_32 - err |= __put_user(regs->cs, (unsigned int __user *)&sc->cs); - err |= __put_user(regs->flags, &sc->flags); - err |= __put_user(regs->sp, &sc->sp_at_signal); - err |= __put_user(regs->ss, (unsigned int __user *)&sc->ss); -#else /* !CONFIG_X86_32 */ - err |= __put_user(regs->flags, &sc->flags); - err |= __put_user(regs->cs, &sc->cs); - err |= __put_user(0, &sc->gs); - err |= __put_user(0, &sc->fs); -#endif /* CONFIG_X86_32 */ - - err |= __put_user(fpstate, &sc->fpstate); - - /* non-iBCS2 extensions.. */ - err |= __put_user(mask, &sc->oldmask); - err |= __put_user(current->thread.cr2, &sc->cr2); - - return err; -} /* * Determine which stack to use.. diff --git a/arch/x86/kernel/signal_64.c b/arch/x86/kernel/signal_64.c index 32718f5e4f61..771c8fcc8b0d 100644 --- a/arch/x86/kernel/signal_64.c +++ b/arch/x86/kernel/signal_64.c @@ -50,28 +50,6 @@ # define FIX_EFLAGS __FIX_EFLAGS #endif -#ifdef CONFIG_X86_32 -asmlinkage int sys_sigaltstack(unsigned long bx) -{ - /* - * This is needed to make gcc realize it doesn't own the - * "struct pt_regs" - */ - struct pt_regs *regs = (struct pt_regs *)&bx; - const stack_t __user *uss = (const stack_t __user *)bx; - stack_t __user *uoss = (stack_t __user *)regs->cx; - - return do_sigaltstack(uss, uoss, regs->sp); -} -#else /* !CONFIG_X86_32 */ -asmlinkage long -sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss, - struct pt_regs *regs) -{ - return do_sigaltstack(uss, uoss, regs->sp); -} -#endif /* CONFIG_X86_32 */ - #define COPY(x) { \ err |= __get_user(regs->x, &sc->x); \ } @@ -82,9 +60,6 @@ sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss, regs->seg = tmp | 3; \ } -/* - * Do a signal return; undo the signal stack. - */ static int restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc, unsigned long *pax) @@ -138,54 +113,6 @@ restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc, return err; } -static long do_rt_sigreturn(struct pt_regs *regs) -{ - struct rt_sigframe __user *frame; - unsigned long ax; - sigset_t set; - - frame = (struct rt_sigframe __user *)(regs->sp - sizeof(long)); - if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) - goto badframe; - if (__copy_from_user(&set, &frame->uc.uc_sigmask, sizeof(set))) - goto badframe; - - sigdelsetmask(&set, ~_BLOCKABLE); - spin_lock_irq(¤t->sighand->siglock); - current->blocked = set; - recalc_sigpending(); - spin_unlock_irq(¤t->sighand->siglock); - - if (restore_sigcontext(regs, &frame->uc.uc_mcontext, &ax)) - goto badframe; - - if (do_sigaltstack(&frame->uc.uc_stack, NULL, regs->sp) == -EFAULT) - goto badframe; - - return ax; - -badframe: - signal_fault(regs, frame, "rt_sigreturn"); - return 0; -} - -#ifdef CONFIG_X86_32 -asmlinkage int sys_rt_sigreturn(unsigned long __unused) -{ - struct pt_regs *regs = (struct pt_regs *)&__unused; - - return do_rt_sigreturn(regs); -} -#else /* !CONFIG_X86_32 */ -asmlinkage long sys_rt_sigreturn(struct pt_regs *regs) -{ - return do_rt_sigreturn(regs); -} -#endif /* CONFIG_X86_32 */ - -/* - * Set up a signal frame. - */ static int setup_sigcontext(struct sigcontext __user *sc, void __user *fpstate, struct pt_regs *regs, unsigned long mask) @@ -247,10 +174,83 @@ setup_sigcontext(struct sigcontext __user *sc, void __user *fpstate, return err; } +#ifdef CONFIG_X86_32 +asmlinkage int sys_sigaltstack(unsigned long bx) +{ + /* + * This is needed to make gcc realize it doesn't own the + * "struct pt_regs" + */ + struct pt_regs *regs = (struct pt_regs *)&bx; + const stack_t __user *uss = (const stack_t __user *)bx; + stack_t __user *uoss = (stack_t __user *)regs->cx; + + return do_sigaltstack(uss, uoss, regs->sp); +} +#else /* !CONFIG_X86_32 */ +asmlinkage long +sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss, + struct pt_regs *regs) +{ + return do_sigaltstack(uss, uoss, regs->sp); +} +#endif /* CONFIG_X86_32 */ + /* - * Determine which stack to use.. + * Do a signal return; undo the signal stack. + */ +static long do_rt_sigreturn(struct pt_regs *regs) +{ + struct rt_sigframe __user *frame; + unsigned long ax; + sigset_t set; + + frame = (struct rt_sigframe __user *)(regs->sp - sizeof(long)); + if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) + goto badframe; + if (__copy_from_user(&set, &frame->uc.uc_sigmask, sizeof(set))) + goto badframe; + + sigdelsetmask(&set, ~_BLOCKABLE); + spin_lock_irq(¤t->sighand->siglock); + current->blocked = set; + recalc_sigpending(); + spin_unlock_irq(¤t->sighand->siglock); + + if (restore_sigcontext(regs, &frame->uc.uc_mcontext, &ax)) + goto badframe; + + if (do_sigaltstack(&frame->uc.uc_stack, NULL, regs->sp) == -EFAULT) + goto badframe; + + return ax; + +badframe: + signal_fault(regs, frame, "rt_sigreturn"); + return 0; +} + +#ifdef CONFIG_X86_32 +asmlinkage int sys_rt_sigreturn(unsigned long __unused) +{ + struct pt_regs *regs = (struct pt_regs *)&__unused; + + return do_rt_sigreturn(regs); +} +#else /* !CONFIG_X86_32 */ +asmlinkage long sys_rt_sigreturn(struct pt_regs *regs) +{ + return do_rt_sigreturn(regs); +} +#endif /* CONFIG_X86_32 */ + +/* + * Set up a signal frame. */ +/* + * Determine which stack to use.. + */ static void __user * get_stack(struct k_sigaction *ka, unsigned long sp, unsigned long size) { -- cgit v1.2.3 From bfeb91a9435889ef4fe7bfbb4b673f625e69e790 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Mon, 24 Nov 2008 18:23:12 -0800 Subject: x86: signal: cosmetic unification of __setup_sigframe() and __setup_rt_sigframe() Impact: cleanup Add #ifdef directive to unify __setup_sigframe() and __setup_rt_sigframe(). Move them after {setup|restore}_sigcontext() declaration. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/kernel/signal_32.c | 469 ++++++++++++++++++++++++++------------------ arch/x86/kernel/signal_64.c | 309 ++++++++++++++++++++++++----- 2 files changed, 536 insertions(+), 242 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/signal_32.c b/arch/x86/kernel/signal_32.c index b3f30d2a2178..e9f71298e746 100644 --- a/arch/x86/kernel/signal_32.c +++ b/arch/x86/kernel/signal_32.c @@ -48,28 +48,6 @@ # define FIX_EFLAGS __FIX_EFLAGS #endif -static const struct { - u16 poplmovl; - u32 val; - u16 int80; -} __attribute__((packed)) retcode = { - 0xb858, /* popl %eax; movl $..., %eax */ - __NR_sigreturn, - 0x80cd, /* int $0x80 */ -}; - -static const struct { - u8 movl; - u32 val; - u16 int80; - u8 pad; -} __attribute__((packed)) rt_retcode = { - 0xb8, /* movl $..., %eax */ - __NR_rt_sigreturn, - 0x80cd, /* int $0x80 */ - 0 -}; - #define COPY(x) { \ err |= __get_user(regs->x, &sc->x); \ } @@ -207,176 +185,30 @@ setup_sigcontext(struct sigcontext __user *sc, void __user *fpstate, } /* - * Atomically swap in the new signal mask, and wait for a signal. + * Set up a signal frame. */ -asmlinkage int -sys_sigsuspend(int history0, int history1, old_sigset_t mask) -{ - mask &= _BLOCKABLE; - spin_lock_irq(¤t->sighand->siglock); - current->saved_sigmask = current->blocked; - siginitset(¤t->blocked, mask); - recalc_sigpending(); - spin_unlock_irq(¤t->sighand->siglock); - - current->state = TASK_INTERRUPTIBLE; - schedule(); - set_restore_sigmask(); - - return -ERESTARTNOHAND; -} - -asmlinkage int -sys_sigaction(int sig, const struct old_sigaction __user *act, - struct old_sigaction __user *oact) -{ - struct k_sigaction new_ka, old_ka; - int ret; - - if (act) { - old_sigset_t mask; - - if (!access_ok(VERIFY_READ, act, sizeof(*act)) || - __get_user(new_ka.sa.sa_handler, &act->sa_handler) || - __get_user(new_ka.sa.sa_restorer, &act->sa_restorer)) - return -EFAULT; - - __get_user(new_ka.sa.sa_flags, &act->sa_flags); - __get_user(mask, &act->sa_mask); - siginitset(&new_ka.sa.sa_mask, mask); - } - - ret = do_sigaction(sig, act ? &new_ka : NULL, oact ? &old_ka : NULL); - - if (!ret && oact) { - if (!access_ok(VERIFY_WRITE, oact, sizeof(*oact)) || - __put_user(old_ka.sa.sa_handler, &oact->sa_handler) || - __put_user(old_ka.sa.sa_restorer, &oact->sa_restorer)) - return -EFAULT; - - __put_user(old_ka.sa.sa_flags, &oact->sa_flags); - __put_user(old_ka.sa.sa_mask.sig[0], &oact->sa_mask); - } - - return ret; -} - #ifdef CONFIG_X86_32 -asmlinkage int sys_sigaltstack(unsigned long bx) -{ - /* - * This is needed to make gcc realize it doesn't own the - * "struct pt_regs" - */ - struct pt_regs *regs = (struct pt_regs *)&bx; - const stack_t __user *uss = (const stack_t __user *)bx; - stack_t __user *uoss = (stack_t __user *)regs->cx; - - return do_sigaltstack(uss, uoss, regs->sp); -} -#else /* !CONFIG_X86_32 */ -asmlinkage long -sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss, - struct pt_regs *regs) -{ - return do_sigaltstack(uss, uoss, regs->sp); -} -#endif /* CONFIG_X86_32 */ - -/* - * Do a signal return; undo the signal stack. - */ -asmlinkage unsigned long sys_sigreturn(unsigned long __unused) -{ - struct sigframe __user *frame; - struct pt_regs *regs; - unsigned long ax; - sigset_t set; - - regs = (struct pt_regs *) &__unused; - frame = (struct sigframe __user *)(regs->sp - 8); - - if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) - goto badframe; - if (__get_user(set.sig[0], &frame->sc.oldmask) || (_NSIG_WORDS > 1 - && __copy_from_user(&set.sig[1], &frame->extramask, - sizeof(frame->extramask)))) - goto badframe; - - sigdelsetmask(&set, ~_BLOCKABLE); - spin_lock_irq(¤t->sighand->siglock); - current->blocked = set; - recalc_sigpending(); - spin_unlock_irq(¤t->sighand->siglock); - - if (restore_sigcontext(regs, &frame->sc, &ax)) - goto badframe; - return ax; - -badframe: - if (show_unhandled_signals && printk_ratelimit()) { - printk("%s%s[%d] bad frame in sigreturn frame:" - "%p ip:%lx sp:%lx oeax:%lx", - task_pid_nr(current) > 1 ? KERN_INFO : KERN_EMERG, - current->comm, task_pid_nr(current), frame, regs->ip, - regs->sp, regs->orig_ax); - print_vma_addr(" in ", regs->ip); - printk(KERN_CONT "\n"); - } - - force_sig(SIGSEGV, current); - - return 0; -} - -static long do_rt_sigreturn(struct pt_regs *regs) -{ - struct rt_sigframe __user *frame; - unsigned long ax; - sigset_t set; - - frame = (struct rt_sigframe __user *)(regs->sp - sizeof(long)); - if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) - goto badframe; - if (__copy_from_user(&set, &frame->uc.uc_sigmask, sizeof(set))) - goto badframe; - - sigdelsetmask(&set, ~_BLOCKABLE); - spin_lock_irq(¤t->sighand->siglock); - current->blocked = set; - recalc_sigpending(); - spin_unlock_irq(¤t->sighand->siglock); - - if (restore_sigcontext(regs, &frame->uc.uc_mcontext, &ax)) - goto badframe; - - if (do_sigaltstack(&frame->uc.uc_stack, NULL, regs->sp) == -EFAULT) - goto badframe; - - return ax; - -badframe: - signal_fault(regs, frame, "rt_sigreturn"); - return 0; -} - -#ifdef CONFIG_X86_32 -asmlinkage int sys_rt_sigreturn(unsigned long __unused) -{ - struct pt_regs *regs = (struct pt_regs *)&__unused; - - return do_rt_sigreturn(regs); -} -#else /* !CONFIG_X86_32 */ -asmlinkage long sys_rt_sigreturn(struct pt_regs *regs) -{ - return do_rt_sigreturn(regs); -} -#endif /* CONFIG_X86_32 */ +static const struct { + u16 poplmovl; + u32 val; + u16 int80; +} __attribute__((packed)) retcode = { + 0xb858, /* popl %eax; movl $..., %eax */ + __NR_sigreturn, + 0x80cd, /* int $0x80 */ +}; -/* - * Set up a signal frame. - */ +static const struct { + u8 movl; + u32 val; + u16 int80; + u8 pad; +} __attribute__((packed)) rt_retcode = { + 0xb8, /* movl $..., %eax */ + __NR_rt_sigreturn, + 0x80cd, /* int $0x80 */ + 0 +}; /* * Determine which stack to use.. @@ -557,6 +389,265 @@ static int __setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, return 0; } +#else /* !CONFIG_X86_32 */ +/* + * Determine which stack to use.. + */ +static void __user * +get_stack(struct k_sigaction *ka, unsigned long sp, unsigned long size) +{ + /* Default to using normal stack - redzone*/ + sp -= 128; + + /* This is the X/Open sanctioned signal stack switching. */ + if (ka->sa.sa_flags & SA_ONSTACK) { + if (sas_ss_flags(sp) == 0) + sp = current->sas_ss_sp + current->sas_ss_size; + } + + return (void __user *)round_down(sp - size, 64); +} + +static int __setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, + sigset_t *set, struct pt_regs *regs) +{ + struct rt_sigframe __user *frame; + void __user *fp = NULL; + int err = 0; + struct task_struct *me = current; + + if (used_math()) { + fp = get_stack(ka, regs->sp, sig_xstate_size); + frame = (void __user *)round_down( + (unsigned long)fp - sizeof(struct rt_sigframe), 16) - 8; + + if (save_i387_xstate(fp) < 0) + return -EFAULT; + } else + frame = get_stack(ka, regs->sp, sizeof(struct rt_sigframe)) - 8; + + if (!access_ok(VERIFY_WRITE, frame, sizeof(*frame))) + return -EFAULT; + + if (ka->sa.sa_flags & SA_SIGINFO) { + if (copy_siginfo_to_user(&frame->info, info)) + return -EFAULT; + } + + /* Create the ucontext. */ + if (cpu_has_xsave) + err |= __put_user(UC_FP_XSTATE, &frame->uc.uc_flags); + else + err |= __put_user(0, &frame->uc.uc_flags); + err |= __put_user(0, &frame->uc.uc_link); + err |= __put_user(me->sas_ss_sp, &frame->uc.uc_stack.ss_sp); + err |= __put_user(sas_ss_flags(regs->sp), + &frame->uc.uc_stack.ss_flags); + err |= __put_user(me->sas_ss_size, &frame->uc.uc_stack.ss_size); + err |= setup_sigcontext(&frame->uc.uc_mcontext, fp, regs, set->sig[0]); + err |= __copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set)); + + /* Set up to return from userspace. If provided, use a stub + already in userspace. */ + /* x86-64 should always use SA_RESTORER. */ + if (ka->sa.sa_flags & SA_RESTORER) { + err |= __put_user(ka->sa.sa_restorer, &frame->pretcode); + } else { + /* could use a vstub here */ + return -EFAULT; + } + + if (err) + return -EFAULT; + + /* Set up registers for signal handler */ + regs->di = sig; + /* In case the signal handler was declared without prototypes */ + regs->ax = 0; + + /* This also works for non SA_SIGINFO handlers because they expect the + next argument after the signal number on the stack. */ + regs->si = (unsigned long)&frame->info; + regs->dx = (unsigned long)&frame->uc; + regs->ip = (unsigned long) ka->sa.sa_handler; + + regs->sp = (unsigned long)frame; + + /* Set up the CS register to run signal handlers in 64-bit mode, + even if the handler happens to be interrupting 32-bit code. */ + regs->cs = __USER_CS; + + return 0; +} +#endif /* CONFIG_X86_32 */ + +/* + * Atomically swap in the new signal mask, and wait for a signal. + */ +asmlinkage int +sys_sigsuspend(int history0, int history1, old_sigset_t mask) +{ + mask &= _BLOCKABLE; + spin_lock_irq(¤t->sighand->siglock); + current->saved_sigmask = current->blocked; + siginitset(¤t->blocked, mask); + recalc_sigpending(); + spin_unlock_irq(¤t->sighand->siglock); + + current->state = TASK_INTERRUPTIBLE; + schedule(); + set_restore_sigmask(); + + return -ERESTARTNOHAND; +} + +asmlinkage int +sys_sigaction(int sig, const struct old_sigaction __user *act, + struct old_sigaction __user *oact) +{ + struct k_sigaction new_ka, old_ka; + int ret; + + if (act) { + old_sigset_t mask; + + if (!access_ok(VERIFY_READ, act, sizeof(*act)) || + __get_user(new_ka.sa.sa_handler, &act->sa_handler) || + __get_user(new_ka.sa.sa_restorer, &act->sa_restorer)) + return -EFAULT; + + __get_user(new_ka.sa.sa_flags, &act->sa_flags); + __get_user(mask, &act->sa_mask); + siginitset(&new_ka.sa.sa_mask, mask); + } + + ret = do_sigaction(sig, act ? &new_ka : NULL, oact ? &old_ka : NULL); + + if (!ret && oact) { + if (!access_ok(VERIFY_WRITE, oact, sizeof(*oact)) || + __put_user(old_ka.sa.sa_handler, &oact->sa_handler) || + __put_user(old_ka.sa.sa_restorer, &oact->sa_restorer)) + return -EFAULT; + + __put_user(old_ka.sa.sa_flags, &oact->sa_flags); + __put_user(old_ka.sa.sa_mask.sig[0], &oact->sa_mask); + } + + return ret; +} + +#ifdef CONFIG_X86_32 +asmlinkage int sys_sigaltstack(unsigned long bx) +{ + /* + * This is needed to make gcc realize it doesn't own the + * "struct pt_regs" + */ + struct pt_regs *regs = (struct pt_regs *)&bx; + const stack_t __user *uss = (const stack_t __user *)bx; + stack_t __user *uoss = (stack_t __user *)regs->cx; + + return do_sigaltstack(uss, uoss, regs->sp); +} +#else /* !CONFIG_X86_32 */ +asmlinkage long +sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss, + struct pt_regs *regs) +{ + return do_sigaltstack(uss, uoss, regs->sp); +} +#endif /* CONFIG_X86_32 */ + +/* + * Do a signal return; undo the signal stack. + */ +asmlinkage unsigned long sys_sigreturn(unsigned long __unused) +{ + struct sigframe __user *frame; + struct pt_regs *regs; + unsigned long ax; + sigset_t set; + + regs = (struct pt_regs *) &__unused; + frame = (struct sigframe __user *)(regs->sp - 8); + + if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) + goto badframe; + if (__get_user(set.sig[0], &frame->sc.oldmask) || (_NSIG_WORDS > 1 + && __copy_from_user(&set.sig[1], &frame->extramask, + sizeof(frame->extramask)))) + goto badframe; + + sigdelsetmask(&set, ~_BLOCKABLE); + spin_lock_irq(¤t->sighand->siglock); + current->blocked = set; + recalc_sigpending(); + spin_unlock_irq(¤t->sighand->siglock); + + if (restore_sigcontext(regs, &frame->sc, &ax)) + goto badframe; + return ax; + +badframe: + if (show_unhandled_signals && printk_ratelimit()) { + printk("%s%s[%d] bad frame in sigreturn frame:" + "%p ip:%lx sp:%lx oeax:%lx", + task_pid_nr(current) > 1 ? KERN_INFO : KERN_EMERG, + current->comm, task_pid_nr(current), frame, regs->ip, + regs->sp, regs->orig_ax); + print_vma_addr(" in ", regs->ip); + printk(KERN_CONT "\n"); + } + + force_sig(SIGSEGV, current); + + return 0; +} + +static long do_rt_sigreturn(struct pt_regs *regs) +{ + struct rt_sigframe __user *frame; + unsigned long ax; + sigset_t set; + + frame = (struct rt_sigframe __user *)(regs->sp - sizeof(long)); + if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) + goto badframe; + if (__copy_from_user(&set, &frame->uc.uc_sigmask, sizeof(set))) + goto badframe; + + sigdelsetmask(&set, ~_BLOCKABLE); + spin_lock_irq(¤t->sighand->siglock); + current->blocked = set; + recalc_sigpending(); + spin_unlock_irq(¤t->sighand->siglock); + + if (restore_sigcontext(regs, &frame->uc.uc_mcontext, &ax)) + goto badframe; + + if (do_sigaltstack(&frame->uc.uc_stack, NULL, regs->sp) == -EFAULT) + goto badframe; + + return ax; + +badframe: + signal_fault(regs, frame, "rt_sigreturn"); + return 0; +} + +#ifdef CONFIG_X86_32 +asmlinkage int sys_rt_sigreturn(unsigned long __unused) +{ + struct pt_regs *regs = (struct pt_regs *)&__unused; + + return do_rt_sigreturn(regs); +} +#else /* !CONFIG_X86_32 */ +asmlinkage long sys_rt_sigreturn(struct pt_regs *regs) +{ + return do_rt_sigreturn(regs); +} +#endif /* CONFIG_X86_32 */ /* * OK, we're invoking a handler: diff --git a/arch/x86/kernel/signal_64.c b/arch/x86/kernel/signal_64.c index 771c8fcc8b0d..2da7e6e60807 100644 --- a/arch/x86/kernel/signal_64.c +++ b/arch/x86/kernel/signal_64.c @@ -174,80 +174,212 @@ setup_sigcontext(struct sigcontext __user *sc, void __user *fpstate, return err; } +/* + * Set up a signal frame. + */ #ifdef CONFIG_X86_32 -asmlinkage int sys_sigaltstack(unsigned long bx) +static const struct { + u16 poplmovl; + u32 val; + u16 int80; +} __attribute__((packed)) retcode = { + 0xb858, /* popl %eax; movl $..., %eax */ + __NR_sigreturn, + 0x80cd, /* int $0x80 */ +}; + +static const struct { + u8 movl; + u32 val; + u16 int80; + u8 pad; +} __attribute__((packed)) rt_retcode = { + 0xb8, /* movl $..., %eax */ + __NR_rt_sigreturn, + 0x80cd, /* int $0x80 */ + 0 +}; + +/* + * Determine which stack to use.. + */ +static inline void __user * +get_sigframe(struct k_sigaction *ka, struct pt_regs *regs, size_t frame_size, + void **fpstate) { + unsigned long sp; + + /* Default to using normal stack */ + sp = regs->sp; + /* - * This is needed to make gcc realize it doesn't own the - * "struct pt_regs" + * If we are on the alternate signal stack and would overflow it, don't. + * Return an always-bogus address instead so we will die with SIGSEGV. */ - struct pt_regs *regs = (struct pt_regs *)&bx; - const stack_t __user *uss = (const stack_t __user *)bx; - stack_t __user *uoss = (stack_t __user *)regs->cx; + if (on_sig_stack(sp) && !likely(on_sig_stack(sp - frame_size))) + return (void __user *) -1L; - return do_sigaltstack(uss, uoss, regs->sp); -} -#else /* !CONFIG_X86_32 */ -asmlinkage long -sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss, - struct pt_regs *regs) -{ - return do_sigaltstack(uss, uoss, regs->sp); + /* This is the X/Open sanctioned signal stack switching. */ + if (ka->sa.sa_flags & SA_ONSTACK) { + if (sas_ss_flags(sp) == 0) + sp = current->sas_ss_sp + current->sas_ss_size; + } else { + /* This is the legacy signal stack switching. */ + if ((regs->ss & 0xffff) != __USER_DS && + !(ka->sa.sa_flags & SA_RESTORER) && + ka->sa.sa_restorer) + sp = (unsigned long) ka->sa.sa_restorer; + } + + if (used_math()) { + sp = sp - sig_xstate_size; + *fpstate = (struct _fpstate *) sp; + if (save_i387_xstate(*fpstate) < 0) + return (void __user *)-1L; + } + + sp -= frame_size; + /* + * Align the stack pointer according to the i386 ABI, + * i.e. so that on function entry ((sp + 4) & 15) == 0. + */ + sp = ((sp + 4) & -16ul) - 4; + + return (void __user *) sp; } -#endif /* CONFIG_X86_32 */ -/* - * Do a signal return; undo the signal stack. - */ -static long do_rt_sigreturn(struct pt_regs *regs) +static int +__setup_frame(int sig, struct k_sigaction *ka, sigset_t *set, + struct pt_regs *regs) { - struct rt_sigframe __user *frame; - unsigned long ax; - sigset_t set; + struct sigframe __user *frame; + void __user *restorer; + int err = 0; + void __user *fpstate = NULL; - frame = (struct rt_sigframe __user *)(regs->sp - sizeof(long)); - if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) - goto badframe; - if (__copy_from_user(&set, &frame->uc.uc_sigmask, sizeof(set))) - goto badframe; + frame = get_sigframe(ka, regs, sizeof(*frame), &fpstate); - sigdelsetmask(&set, ~_BLOCKABLE); - spin_lock_irq(¤t->sighand->siglock); - current->blocked = set; - recalc_sigpending(); - spin_unlock_irq(¤t->sighand->siglock); + if (!access_ok(VERIFY_WRITE, frame, sizeof(*frame))) + return -EFAULT; - if (restore_sigcontext(regs, &frame->uc.uc_mcontext, &ax)) - goto badframe; + if (__put_user(sig, &frame->sig)) + return -EFAULT; - if (do_sigaltstack(&frame->uc.uc_stack, NULL, regs->sp) == -EFAULT) - goto badframe; + if (setup_sigcontext(&frame->sc, fpstate, regs, set->sig[0])) + return -EFAULT; - return ax; + if (_NSIG_WORDS > 1) { + if (__copy_to_user(&frame->extramask, &set->sig[1], + sizeof(frame->extramask))) + return -EFAULT; + } + + if (current->mm->context.vdso) + restorer = VDSO32_SYMBOL(current->mm->context.vdso, sigreturn); + else + restorer = &frame->retcode; + if (ka->sa.sa_flags & SA_RESTORER) + restorer = ka->sa.sa_restorer; + + /* Set up to return from userspace. */ + err |= __put_user(restorer, &frame->pretcode); + + /* + * This is popl %eax ; movl $__NR_sigreturn, %eax ; int $0x80 + * + * WE DO NOT USE IT ANY MORE! It's only left here for historical + * reasons and because gdb uses it as a signature to notice + * signal handler stack frames. + */ + err |= __put_user(*((u64 *)&retcode), (u64 *)frame->retcode); + + if (err) + return -EFAULT; + + /* Set up registers for signal handler */ + regs->sp = (unsigned long)frame; + regs->ip = (unsigned long)ka->sa.sa_handler; + regs->ax = (unsigned long)sig; + regs->dx = 0; + regs->cx = 0; + + regs->ds = __USER_DS; + regs->es = __USER_DS; + regs->ss = __USER_DS; + regs->cs = __USER_CS; -badframe: - signal_fault(regs, frame, "rt_sigreturn"); return 0; } -#ifdef CONFIG_X86_32 -asmlinkage int sys_rt_sigreturn(unsigned long __unused) +static int __setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, + sigset_t *set, struct pt_regs *regs) { - struct pt_regs *regs = (struct pt_regs *)&__unused; + struct rt_sigframe __user *frame; + void __user *restorer; + int err = 0; + void __user *fpstate = NULL; - return do_rt_sigreturn(regs); -} -#else /* !CONFIG_X86_32 */ -asmlinkage long sys_rt_sigreturn(struct pt_regs *regs) -{ - return do_rt_sigreturn(regs); -} -#endif /* CONFIG_X86_32 */ + frame = get_sigframe(ka, regs, sizeof(*frame), &fpstate); -/* - * Set up a signal frame. - */ + if (!access_ok(VERIFY_WRITE, frame, sizeof(*frame))) + return -EFAULT; + + err |= __put_user(sig, &frame->sig); + err |= __put_user(&frame->info, &frame->pinfo); + err |= __put_user(&frame->uc, &frame->puc); + err |= copy_siginfo_to_user(&frame->info, info); + if (err) + return -EFAULT; + + /* Create the ucontext. */ + if (cpu_has_xsave) + err |= __put_user(UC_FP_XSTATE, &frame->uc.uc_flags); + else + err |= __put_user(0, &frame->uc.uc_flags); + err |= __put_user(0, &frame->uc.uc_link); + err |= __put_user(current->sas_ss_sp, &frame->uc.uc_stack.ss_sp); + err |= __put_user(sas_ss_flags(regs->sp), + &frame->uc.uc_stack.ss_flags); + err |= __put_user(current->sas_ss_size, &frame->uc.uc_stack.ss_size); + err |= setup_sigcontext(&frame->uc.uc_mcontext, fpstate, + regs, set->sig[0]); + err |= __copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set)); + if (err) + return -EFAULT; + + /* Set up to return from userspace. */ + restorer = VDSO32_SYMBOL(current->mm->context.vdso, rt_sigreturn); + if (ka->sa.sa_flags & SA_RESTORER) + restorer = ka->sa.sa_restorer; + err |= __put_user(restorer, &frame->pretcode); + + /* + * This is movl $__NR_rt_sigreturn, %ax ; int $0x80 + * + * WE DO NOT USE IT ANY MORE! It's only left here for historical + * reasons and because gdb uses it as a signature to notice + * signal handler stack frames. + */ + err |= __put_user(*((u64 *)&rt_retcode), (u64 *)frame->retcode); + if (err) + return -EFAULT; + + /* Set up registers for signal handler */ + regs->sp = (unsigned long)frame; + regs->ip = (unsigned long)ka->sa.sa_handler; + regs->ax = (unsigned long)sig; + regs->dx = (unsigned long)&frame->info; + regs->cx = (unsigned long)&frame->uc; + + regs->ds = __USER_DS; + regs->es = __USER_DS; + regs->ss = __USER_DS; + regs->cs = __USER_CS; + + return 0; +} +#else /* !CONFIG_X86_32 */ /* * Determine which stack to use.. */ @@ -337,6 +469,77 @@ static int __setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, return 0; } +#endif /* CONFIG_X86_32 */ + +#ifdef CONFIG_X86_32 +asmlinkage int sys_sigaltstack(unsigned long bx) +{ + /* + * This is needed to make gcc realize it doesn't own the + * "struct pt_regs" + */ + struct pt_regs *regs = (struct pt_regs *)&bx; + const stack_t __user *uss = (const stack_t __user *)bx; + stack_t __user *uoss = (stack_t __user *)regs->cx; + + return do_sigaltstack(uss, uoss, regs->sp); +} +#else /* !CONFIG_X86_32 */ +asmlinkage long +sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss, + struct pt_regs *regs) +{ + return do_sigaltstack(uss, uoss, regs->sp); +} +#endif /* CONFIG_X86_32 */ + +/* + * Do a signal return; undo the signal stack. + */ +static long do_rt_sigreturn(struct pt_regs *regs) +{ + struct rt_sigframe __user *frame; + unsigned long ax; + sigset_t set; + + frame = (struct rt_sigframe __user *)(regs->sp - sizeof(long)); + if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) + goto badframe; + if (__copy_from_user(&set, &frame->uc.uc_sigmask, sizeof(set))) + goto badframe; + + sigdelsetmask(&set, ~_BLOCKABLE); + spin_lock_irq(¤t->sighand->siglock); + current->blocked = set; + recalc_sigpending(); + spin_unlock_irq(¤t->sighand->siglock); + + if (restore_sigcontext(regs, &frame->uc.uc_mcontext, &ax)) + goto badframe; + + if (do_sigaltstack(&frame->uc.uc_stack, NULL, regs->sp) == -EFAULT) + goto badframe; + + return ax; + +badframe: + signal_fault(regs, frame, "rt_sigreturn"); + return 0; +} + +#ifdef CONFIG_X86_32 +asmlinkage int sys_rt_sigreturn(unsigned long __unused) +{ + struct pt_regs *regs = (struct pt_regs *)&__unused; + + return do_rt_sigreturn(regs); +} +#else /* !CONFIG_X86_32 */ +asmlinkage long sys_rt_sigreturn(struct pt_regs *regs) +{ + return do_rt_sigreturn(regs); +} +#endif /* CONFIG_X86_32 */ /* * OK, we're invoking a handler -- cgit v1.2.3 From e5fa2d063cf2ca38eae5fb3469315db669d5c041 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Mon, 24 Nov 2008 18:24:11 -0800 Subject: x86: signal: unify signal_{32|64}.c, prepare Impact: cleanup Add #ifdef directive for 32-bit only code. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/kernel/signal_32.c | 6 +++ arch/x86/kernel/signal_64.c | 116 +++++++++++++++++++++++++++++++++++++++++++- 2 files changed, 121 insertions(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/signal_32.c b/arch/x86/kernel/signal_32.c index e9f71298e746..b1f4d34e0a38 100644 --- a/arch/x86/kernel/signal_32.c +++ b/arch/x86/kernel/signal_32.c @@ -1,8 +1,10 @@ /* * Copyright (C) 1991, 1992 Linus Torvalds + * Copyright (C) 2000, 2001, 2002 Andi Kleen SuSE Labs * * 1997-11-28 Modified for POSIX.1b signals by Richard Henderson * 2000-06-20 Pentium III FXSR, SSE support by Gareth Hughes + * 2000-2002 x86-64 support by Andi Kleen */ #include @@ -481,6 +483,7 @@ static int __setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, } #endif /* CONFIG_X86_32 */ +#ifdef CONFIG_X86_32 /* * Atomically swap in the new signal mask, and wait for a signal. */ @@ -535,6 +538,7 @@ sys_sigaction(int sig, const struct old_sigaction __user *act, return ret; } +#endif /* CONFIG_X86_32 */ #ifdef CONFIG_X86_32 asmlinkage int sys_sigaltstack(unsigned long bx) @@ -561,6 +565,7 @@ sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss, /* * Do a signal return; undo the signal stack. */ +#ifdef CONFIG_X86_32 asmlinkage unsigned long sys_sigreturn(unsigned long __unused) { struct sigframe __user *frame; @@ -603,6 +608,7 @@ badframe: return 0; } +#endif /* CONFIG_X86_32 */ static long do_rt_sigreturn(struct pt_regs *regs) { diff --git a/arch/x86/kernel/signal_64.c b/arch/x86/kernel/signal_64.c index 2da7e6e60807..b1f4d34e0a38 100644 --- a/arch/x86/kernel/signal_64.c +++ b/arch/x86/kernel/signal_64.c @@ -54,12 +54,24 @@ err |= __get_user(regs->x, &sc->x); \ } +#define COPY_SEG(seg) { \ + unsigned short tmp; \ + err |= __get_user(tmp, &sc->seg); \ + regs->seg = tmp; \ +} + #define COPY_SEG_CPL3(seg) { \ unsigned short tmp; \ err |= __get_user(tmp, &sc->seg); \ regs->seg = tmp | 3; \ } +#define GET_SEG(seg) { \ + unsigned short tmp; \ + err |= __get_user(tmp, &sc->seg); \ + loadsegment(seg, tmp); \ +} + static int restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc, unsigned long *pax) @@ -471,6 +483,63 @@ static int __setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, } #endif /* CONFIG_X86_32 */ +#ifdef CONFIG_X86_32 +/* + * Atomically swap in the new signal mask, and wait for a signal. + */ +asmlinkage int +sys_sigsuspend(int history0, int history1, old_sigset_t mask) +{ + mask &= _BLOCKABLE; + spin_lock_irq(¤t->sighand->siglock); + current->saved_sigmask = current->blocked; + siginitset(¤t->blocked, mask); + recalc_sigpending(); + spin_unlock_irq(¤t->sighand->siglock); + + current->state = TASK_INTERRUPTIBLE; + schedule(); + set_restore_sigmask(); + + return -ERESTARTNOHAND; +} + +asmlinkage int +sys_sigaction(int sig, const struct old_sigaction __user *act, + struct old_sigaction __user *oact) +{ + struct k_sigaction new_ka, old_ka; + int ret; + + if (act) { + old_sigset_t mask; + + if (!access_ok(VERIFY_READ, act, sizeof(*act)) || + __get_user(new_ka.sa.sa_handler, &act->sa_handler) || + __get_user(new_ka.sa.sa_restorer, &act->sa_restorer)) + return -EFAULT; + + __get_user(new_ka.sa.sa_flags, &act->sa_flags); + __get_user(mask, &act->sa_mask); + siginitset(&new_ka.sa.sa_mask, mask); + } + + ret = do_sigaction(sig, act ? &new_ka : NULL, oact ? &old_ka : NULL); + + if (!ret && oact) { + if (!access_ok(VERIFY_WRITE, oact, sizeof(*oact)) || + __put_user(old_ka.sa.sa_handler, &oact->sa_handler) || + __put_user(old_ka.sa.sa_restorer, &oact->sa_restorer)) + return -EFAULT; + + __put_user(old_ka.sa.sa_flags, &oact->sa_flags); + __put_user(old_ka.sa.sa_mask.sig[0], &oact->sa_mask); + } + + return ret; +} +#endif /* CONFIG_X86_32 */ + #ifdef CONFIG_X86_32 asmlinkage int sys_sigaltstack(unsigned long bx) { @@ -496,6 +565,51 @@ sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss, /* * Do a signal return; undo the signal stack. */ +#ifdef CONFIG_X86_32 +asmlinkage unsigned long sys_sigreturn(unsigned long __unused) +{ + struct sigframe __user *frame; + struct pt_regs *regs; + unsigned long ax; + sigset_t set; + + regs = (struct pt_regs *) &__unused; + frame = (struct sigframe __user *)(regs->sp - 8); + + if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) + goto badframe; + if (__get_user(set.sig[0], &frame->sc.oldmask) || (_NSIG_WORDS > 1 + && __copy_from_user(&set.sig[1], &frame->extramask, + sizeof(frame->extramask)))) + goto badframe; + + sigdelsetmask(&set, ~_BLOCKABLE); + spin_lock_irq(¤t->sighand->siglock); + current->blocked = set; + recalc_sigpending(); + spin_unlock_irq(¤t->sighand->siglock); + + if (restore_sigcontext(regs, &frame->sc, &ax)) + goto badframe; + return ax; + +badframe: + if (show_unhandled_signals && printk_ratelimit()) { + printk("%s%s[%d] bad frame in sigreturn frame:" + "%p ip:%lx sp:%lx oeax:%lx", + task_pid_nr(current) > 1 ? KERN_INFO : KERN_EMERG, + current->comm, task_pid_nr(current), frame, regs->ip, + regs->sp, regs->orig_ax); + print_vma_addr(" in ", regs->ip); + printk(KERN_CONT "\n"); + } + + force_sig(SIGSEGV, current); + + return 0; +} +#endif /* CONFIG_X86_32 */ + static long do_rt_sigreturn(struct pt_regs *regs) { struct rt_sigframe __user *frame; @@ -542,7 +656,7 @@ asmlinkage long sys_rt_sigreturn(struct pt_regs *regs) #endif /* CONFIG_X86_32 */ /* - * OK, we're invoking a handler + * OK, we're invoking a handler: */ static int signr_convert(int sig) { -- cgit v1.2.3 From 5ceb40da9bacc8b056805d72efb1a52502d56b6b Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Mon, 24 Nov 2008 18:24:11 -0800 Subject: x86: signal: unify signal_{32|64}.c Impact: cleanup Unify signal_{32|64}.c! Mechanic unification - the two files are the same. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/kernel/Makefile | 2 +- arch/x86/kernel/signal.c | 915 ++++++++++++++++++++++++++++++++++++++++++++ arch/x86/kernel/signal_32.c | 915 -------------------------------------------- arch/x86/kernel/signal_64.c | 915 -------------------------------------------- 4 files changed, 916 insertions(+), 1831 deletions(-) create mode 100644 arch/x86/kernel/signal.c delete mode 100644 arch/x86/kernel/signal_32.c delete mode 100644 arch/x86/kernel/signal_64.c (limited to 'arch/x86') diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index d7e5a58ee22f..ef28c210ebf8 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -22,7 +22,7 @@ CFLAGS_vsyscall_64.o := $(PROFILING) -g0 $(nostackp) CFLAGS_hpet.o := $(nostackp) CFLAGS_tsc.o := $(nostackp) -obj-y := process_$(BITS).o signal_$(BITS).o entry_$(BITS).o +obj-y := process_$(BITS).o signal.o entry_$(BITS).o obj-y += traps.o irq.o irq_$(BITS).o dumpstack_$(BITS).o obj-y += time_$(BITS).o ioport.o ldt.o obj-y += setup.o i8259.o irqinit_$(BITS).o setup_percpu.o diff --git a/arch/x86/kernel/signal.c b/arch/x86/kernel/signal.c new file mode 100644 index 000000000000..b1f4d34e0a38 --- /dev/null +++ b/arch/x86/kernel/signal.c @@ -0,0 +1,915 @@ +/* + * Copyright (C) 1991, 1992 Linus Torvalds + * Copyright (C) 2000, 2001, 2002 Andi Kleen SuSE Labs + * + * 1997-11-28 Modified for POSIX.1b signals by Richard Henderson + * 2000-06-20 Pentium III FXSR, SSE support by Gareth Hughes + * 2000-2002 x86-64 support by Andi Kleen + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include +#include + +#ifdef CONFIG_X86_64 +#include +#include +#include +#endif /* CONFIG_X86_64 */ + +#include +#include + +#include "sigframe.h" + +#define _BLOCKABLE (~(sigmask(SIGKILL) | sigmask(SIGSTOP))) + +#define __FIX_EFLAGS (X86_EFLAGS_AC | X86_EFLAGS_OF | \ + X86_EFLAGS_DF | X86_EFLAGS_TF | X86_EFLAGS_SF | \ + X86_EFLAGS_ZF | X86_EFLAGS_AF | X86_EFLAGS_PF | \ + X86_EFLAGS_CF) + +#ifdef CONFIG_X86_32 +# define FIX_EFLAGS (__FIX_EFLAGS | X86_EFLAGS_RF) +#else +# define FIX_EFLAGS __FIX_EFLAGS +#endif + +#define COPY(x) { \ + err |= __get_user(regs->x, &sc->x); \ +} + +#define COPY_SEG(seg) { \ + unsigned short tmp; \ + err |= __get_user(tmp, &sc->seg); \ + regs->seg = tmp; \ +} + +#define COPY_SEG_CPL3(seg) { \ + unsigned short tmp; \ + err |= __get_user(tmp, &sc->seg); \ + regs->seg = tmp | 3; \ +} + +#define GET_SEG(seg) { \ + unsigned short tmp; \ + err |= __get_user(tmp, &sc->seg); \ + loadsegment(seg, tmp); \ +} + +static int +restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc, + unsigned long *pax) +{ + void __user *buf; + unsigned int tmpflags; + unsigned int err = 0; + + /* Always make any pending restarted system calls return -EINTR */ + current_thread_info()->restart_block.fn = do_no_restart_syscall; + +#ifdef CONFIG_X86_32 + GET_SEG(gs); + COPY_SEG(fs); + COPY_SEG(es); + COPY_SEG(ds); +#endif /* CONFIG_X86_32 */ + + COPY(di); COPY(si); COPY(bp); COPY(sp); COPY(bx); + COPY(dx); COPY(cx); COPY(ip); + +#ifdef CONFIG_X86_64 + COPY(r8); + COPY(r9); + COPY(r10); + COPY(r11); + COPY(r12); + COPY(r13); + COPY(r14); + COPY(r15); +#endif /* CONFIG_X86_64 */ + +#ifdef CONFIG_X86_32 + COPY_SEG_CPL3(cs); + COPY_SEG_CPL3(ss); +#else /* !CONFIG_X86_32 */ + /* Kernel saves and restores only the CS segment register on signals, + * which is the bare minimum needed to allow mixed 32/64-bit code. + * App's signal handler can save/restore other segments if needed. */ + COPY_SEG_CPL3(cs); +#endif /* CONFIG_X86_32 */ + + err |= __get_user(tmpflags, &sc->flags); + regs->flags = (regs->flags & ~FIX_EFLAGS) | (tmpflags & FIX_EFLAGS); + regs->orig_ax = -1; /* disable syscall checks */ + + err |= __get_user(buf, &sc->fpstate); + err |= restore_i387_xstate(buf); + + err |= __get_user(*pax, &sc->ax); + return err; +} + +static int +setup_sigcontext(struct sigcontext __user *sc, void __user *fpstate, + struct pt_regs *regs, unsigned long mask) +{ + int err = 0; + +#ifdef CONFIG_X86_32 + { + unsigned int tmp; + + savesegment(gs, tmp); + err |= __put_user(tmp, (unsigned int __user *)&sc->gs); + } + err |= __put_user(regs->fs, (unsigned int __user *)&sc->fs); + err |= __put_user(regs->es, (unsigned int __user *)&sc->es); + err |= __put_user(regs->ds, (unsigned int __user *)&sc->ds); +#endif /* CONFIG_X86_32 */ + + err |= __put_user(regs->di, &sc->di); + err |= __put_user(regs->si, &sc->si); + err |= __put_user(regs->bp, &sc->bp); + err |= __put_user(regs->sp, &sc->sp); + err |= __put_user(regs->bx, &sc->bx); + err |= __put_user(regs->dx, &sc->dx); + err |= __put_user(regs->cx, &sc->cx); + err |= __put_user(regs->ax, &sc->ax); +#ifdef CONFIG_X86_64 + err |= __put_user(regs->r8, &sc->r8); + err |= __put_user(regs->r9, &sc->r9); + err |= __put_user(regs->r10, &sc->r10); + err |= __put_user(regs->r11, &sc->r11); + err |= __put_user(regs->r12, &sc->r12); + err |= __put_user(regs->r13, &sc->r13); + err |= __put_user(regs->r14, &sc->r14); + err |= __put_user(regs->r15, &sc->r15); +#endif /* CONFIG_X86_64 */ + + err |= __put_user(current->thread.trap_no, &sc->trapno); + err |= __put_user(current->thread.error_code, &sc->err); + err |= __put_user(regs->ip, &sc->ip); +#ifdef CONFIG_X86_32 + err |= __put_user(regs->cs, (unsigned int __user *)&sc->cs); + err |= __put_user(regs->flags, &sc->flags); + err |= __put_user(regs->sp, &sc->sp_at_signal); + err |= __put_user(regs->ss, (unsigned int __user *)&sc->ss); +#else /* !CONFIG_X86_32 */ + err |= __put_user(regs->flags, &sc->flags); + err |= __put_user(regs->cs, &sc->cs); + err |= __put_user(0, &sc->gs); + err |= __put_user(0, &sc->fs); +#endif /* CONFIG_X86_32 */ + + err |= __put_user(fpstate, &sc->fpstate); + + /* non-iBCS2 extensions.. */ + err |= __put_user(mask, &sc->oldmask); + err |= __put_user(current->thread.cr2, &sc->cr2); + + return err; +} + +/* + * Set up a signal frame. + */ +#ifdef CONFIG_X86_32 +static const struct { + u16 poplmovl; + u32 val; + u16 int80; +} __attribute__((packed)) retcode = { + 0xb858, /* popl %eax; movl $..., %eax */ + __NR_sigreturn, + 0x80cd, /* int $0x80 */ +}; + +static const struct { + u8 movl; + u32 val; + u16 int80; + u8 pad; +} __attribute__((packed)) rt_retcode = { + 0xb8, /* movl $..., %eax */ + __NR_rt_sigreturn, + 0x80cd, /* int $0x80 */ + 0 +}; + +/* + * Determine which stack to use.. + */ +static inline void __user * +get_sigframe(struct k_sigaction *ka, struct pt_regs *regs, size_t frame_size, + void **fpstate) +{ + unsigned long sp; + + /* Default to using normal stack */ + sp = regs->sp; + + /* + * If we are on the alternate signal stack and would overflow it, don't. + * Return an always-bogus address instead so we will die with SIGSEGV. + */ + if (on_sig_stack(sp) && !likely(on_sig_stack(sp - frame_size))) + return (void __user *) -1L; + + /* This is the X/Open sanctioned signal stack switching. */ + if (ka->sa.sa_flags & SA_ONSTACK) { + if (sas_ss_flags(sp) == 0) + sp = current->sas_ss_sp + current->sas_ss_size; + } else { + /* This is the legacy signal stack switching. */ + if ((regs->ss & 0xffff) != __USER_DS && + !(ka->sa.sa_flags & SA_RESTORER) && + ka->sa.sa_restorer) + sp = (unsigned long) ka->sa.sa_restorer; + } + + if (used_math()) { + sp = sp - sig_xstate_size; + *fpstate = (struct _fpstate *) sp; + if (save_i387_xstate(*fpstate) < 0) + return (void __user *)-1L; + } + + sp -= frame_size; + /* + * Align the stack pointer according to the i386 ABI, + * i.e. so that on function entry ((sp + 4) & 15) == 0. + */ + sp = ((sp + 4) & -16ul) - 4; + + return (void __user *) sp; +} + +static int +__setup_frame(int sig, struct k_sigaction *ka, sigset_t *set, + struct pt_regs *regs) +{ + struct sigframe __user *frame; + void __user *restorer; + int err = 0; + void __user *fpstate = NULL; + + frame = get_sigframe(ka, regs, sizeof(*frame), &fpstate); + + if (!access_ok(VERIFY_WRITE, frame, sizeof(*frame))) + return -EFAULT; + + if (__put_user(sig, &frame->sig)) + return -EFAULT; + + if (setup_sigcontext(&frame->sc, fpstate, regs, set->sig[0])) + return -EFAULT; + + if (_NSIG_WORDS > 1) { + if (__copy_to_user(&frame->extramask, &set->sig[1], + sizeof(frame->extramask))) + return -EFAULT; + } + + if (current->mm->context.vdso) + restorer = VDSO32_SYMBOL(current->mm->context.vdso, sigreturn); + else + restorer = &frame->retcode; + if (ka->sa.sa_flags & SA_RESTORER) + restorer = ka->sa.sa_restorer; + + /* Set up to return from userspace. */ + err |= __put_user(restorer, &frame->pretcode); + + /* + * This is popl %eax ; movl $__NR_sigreturn, %eax ; int $0x80 + * + * WE DO NOT USE IT ANY MORE! It's only left here for historical + * reasons and because gdb uses it as a signature to notice + * signal handler stack frames. + */ + err |= __put_user(*((u64 *)&retcode), (u64 *)frame->retcode); + + if (err) + return -EFAULT; + + /* Set up registers for signal handler */ + regs->sp = (unsigned long)frame; + regs->ip = (unsigned long)ka->sa.sa_handler; + regs->ax = (unsigned long)sig; + regs->dx = 0; + regs->cx = 0; + + regs->ds = __USER_DS; + regs->es = __USER_DS; + regs->ss = __USER_DS; + regs->cs = __USER_CS; + + return 0; +} + +static int __setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, + sigset_t *set, struct pt_regs *regs) +{ + struct rt_sigframe __user *frame; + void __user *restorer; + int err = 0; + void __user *fpstate = NULL; + + frame = get_sigframe(ka, regs, sizeof(*frame), &fpstate); + + if (!access_ok(VERIFY_WRITE, frame, sizeof(*frame))) + return -EFAULT; + + err |= __put_user(sig, &frame->sig); + err |= __put_user(&frame->info, &frame->pinfo); + err |= __put_user(&frame->uc, &frame->puc); + err |= copy_siginfo_to_user(&frame->info, info); + if (err) + return -EFAULT; + + /* Create the ucontext. */ + if (cpu_has_xsave) + err |= __put_user(UC_FP_XSTATE, &frame->uc.uc_flags); + else + err |= __put_user(0, &frame->uc.uc_flags); + err |= __put_user(0, &frame->uc.uc_link); + err |= __put_user(current->sas_ss_sp, &frame->uc.uc_stack.ss_sp); + err |= __put_user(sas_ss_flags(regs->sp), + &frame->uc.uc_stack.ss_flags); + err |= __put_user(current->sas_ss_size, &frame->uc.uc_stack.ss_size); + err |= setup_sigcontext(&frame->uc.uc_mcontext, fpstate, + regs, set->sig[0]); + err |= __copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set)); + if (err) + return -EFAULT; + + /* Set up to return from userspace. */ + restorer = VDSO32_SYMBOL(current->mm->context.vdso, rt_sigreturn); + if (ka->sa.sa_flags & SA_RESTORER) + restorer = ka->sa.sa_restorer; + err |= __put_user(restorer, &frame->pretcode); + + /* + * This is movl $__NR_rt_sigreturn, %ax ; int $0x80 + * + * WE DO NOT USE IT ANY MORE! It's only left here for historical + * reasons and because gdb uses it as a signature to notice + * signal handler stack frames. + */ + err |= __put_user(*((u64 *)&rt_retcode), (u64 *)frame->retcode); + + if (err) + return -EFAULT; + + /* Set up registers for signal handler */ + regs->sp = (unsigned long)frame; + regs->ip = (unsigned long)ka->sa.sa_handler; + regs->ax = (unsigned long)sig; + regs->dx = (unsigned long)&frame->info; + regs->cx = (unsigned long)&frame->uc; + + regs->ds = __USER_DS; + regs->es = __USER_DS; + regs->ss = __USER_DS; + regs->cs = __USER_CS; + + return 0; +} +#else /* !CONFIG_X86_32 */ +/* + * Determine which stack to use.. + */ +static void __user * +get_stack(struct k_sigaction *ka, unsigned long sp, unsigned long size) +{ + /* Default to using normal stack - redzone*/ + sp -= 128; + + /* This is the X/Open sanctioned signal stack switching. */ + if (ka->sa.sa_flags & SA_ONSTACK) { + if (sas_ss_flags(sp) == 0) + sp = current->sas_ss_sp + current->sas_ss_size; + } + + return (void __user *)round_down(sp - size, 64); +} + +static int __setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, + sigset_t *set, struct pt_regs *regs) +{ + struct rt_sigframe __user *frame; + void __user *fp = NULL; + int err = 0; + struct task_struct *me = current; + + if (used_math()) { + fp = get_stack(ka, regs->sp, sig_xstate_size); + frame = (void __user *)round_down( + (unsigned long)fp - sizeof(struct rt_sigframe), 16) - 8; + + if (save_i387_xstate(fp) < 0) + return -EFAULT; + } else + frame = get_stack(ka, regs->sp, sizeof(struct rt_sigframe)) - 8; + + if (!access_ok(VERIFY_WRITE, frame, sizeof(*frame))) + return -EFAULT; + + if (ka->sa.sa_flags & SA_SIGINFO) { + if (copy_siginfo_to_user(&frame->info, info)) + return -EFAULT; + } + + /* Create the ucontext. */ + if (cpu_has_xsave) + err |= __put_user(UC_FP_XSTATE, &frame->uc.uc_flags); + else + err |= __put_user(0, &frame->uc.uc_flags); + err |= __put_user(0, &frame->uc.uc_link); + err |= __put_user(me->sas_ss_sp, &frame->uc.uc_stack.ss_sp); + err |= __put_user(sas_ss_flags(regs->sp), + &frame->uc.uc_stack.ss_flags); + err |= __put_user(me->sas_ss_size, &frame->uc.uc_stack.ss_size); + err |= setup_sigcontext(&frame->uc.uc_mcontext, fp, regs, set->sig[0]); + err |= __copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set)); + + /* Set up to return from userspace. If provided, use a stub + already in userspace. */ + /* x86-64 should always use SA_RESTORER. */ + if (ka->sa.sa_flags & SA_RESTORER) { + err |= __put_user(ka->sa.sa_restorer, &frame->pretcode); + } else { + /* could use a vstub here */ + return -EFAULT; + } + + if (err) + return -EFAULT; + + /* Set up registers for signal handler */ + regs->di = sig; + /* In case the signal handler was declared without prototypes */ + regs->ax = 0; + + /* This also works for non SA_SIGINFO handlers because they expect the + next argument after the signal number on the stack. */ + regs->si = (unsigned long)&frame->info; + regs->dx = (unsigned long)&frame->uc; + regs->ip = (unsigned long) ka->sa.sa_handler; + + regs->sp = (unsigned long)frame; + + /* Set up the CS register to run signal handlers in 64-bit mode, + even if the handler happens to be interrupting 32-bit code. */ + regs->cs = __USER_CS; + + return 0; +} +#endif /* CONFIG_X86_32 */ + +#ifdef CONFIG_X86_32 +/* + * Atomically swap in the new signal mask, and wait for a signal. + */ +asmlinkage int +sys_sigsuspend(int history0, int history1, old_sigset_t mask) +{ + mask &= _BLOCKABLE; + spin_lock_irq(¤t->sighand->siglock); + current->saved_sigmask = current->blocked; + siginitset(¤t->blocked, mask); + recalc_sigpending(); + spin_unlock_irq(¤t->sighand->siglock); + + current->state = TASK_INTERRUPTIBLE; + schedule(); + set_restore_sigmask(); + + return -ERESTARTNOHAND; +} + +asmlinkage int +sys_sigaction(int sig, const struct old_sigaction __user *act, + struct old_sigaction __user *oact) +{ + struct k_sigaction new_ka, old_ka; + int ret; + + if (act) { + old_sigset_t mask; + + if (!access_ok(VERIFY_READ, act, sizeof(*act)) || + __get_user(new_ka.sa.sa_handler, &act->sa_handler) || + __get_user(new_ka.sa.sa_restorer, &act->sa_restorer)) + return -EFAULT; + + __get_user(new_ka.sa.sa_flags, &act->sa_flags); + __get_user(mask, &act->sa_mask); + siginitset(&new_ka.sa.sa_mask, mask); + } + + ret = do_sigaction(sig, act ? &new_ka : NULL, oact ? &old_ka : NULL); + + if (!ret && oact) { + if (!access_ok(VERIFY_WRITE, oact, sizeof(*oact)) || + __put_user(old_ka.sa.sa_handler, &oact->sa_handler) || + __put_user(old_ka.sa.sa_restorer, &oact->sa_restorer)) + return -EFAULT; + + __put_user(old_ka.sa.sa_flags, &oact->sa_flags); + __put_user(old_ka.sa.sa_mask.sig[0], &oact->sa_mask); + } + + return ret; +} +#endif /* CONFIG_X86_32 */ + +#ifdef CONFIG_X86_32 +asmlinkage int sys_sigaltstack(unsigned long bx) +{ + /* + * This is needed to make gcc realize it doesn't own the + * "struct pt_regs" + */ + struct pt_regs *regs = (struct pt_regs *)&bx; + const stack_t __user *uss = (const stack_t __user *)bx; + stack_t __user *uoss = (stack_t __user *)regs->cx; + + return do_sigaltstack(uss, uoss, regs->sp); +} +#else /* !CONFIG_X86_32 */ +asmlinkage long +sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss, + struct pt_regs *regs) +{ + return do_sigaltstack(uss, uoss, regs->sp); +} +#endif /* CONFIG_X86_32 */ + +/* + * Do a signal return; undo the signal stack. + */ +#ifdef CONFIG_X86_32 +asmlinkage unsigned long sys_sigreturn(unsigned long __unused) +{ + struct sigframe __user *frame; + struct pt_regs *regs; + unsigned long ax; + sigset_t set; + + regs = (struct pt_regs *) &__unused; + frame = (struct sigframe __user *)(regs->sp - 8); + + if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) + goto badframe; + if (__get_user(set.sig[0], &frame->sc.oldmask) || (_NSIG_WORDS > 1 + && __copy_from_user(&set.sig[1], &frame->extramask, + sizeof(frame->extramask)))) + goto badframe; + + sigdelsetmask(&set, ~_BLOCKABLE); + spin_lock_irq(¤t->sighand->siglock); + current->blocked = set; + recalc_sigpending(); + spin_unlock_irq(¤t->sighand->siglock); + + if (restore_sigcontext(regs, &frame->sc, &ax)) + goto badframe; + return ax; + +badframe: + if (show_unhandled_signals && printk_ratelimit()) { + printk("%s%s[%d] bad frame in sigreturn frame:" + "%p ip:%lx sp:%lx oeax:%lx", + task_pid_nr(current) > 1 ? KERN_INFO : KERN_EMERG, + current->comm, task_pid_nr(current), frame, regs->ip, + regs->sp, regs->orig_ax); + print_vma_addr(" in ", regs->ip); + printk(KERN_CONT "\n"); + } + + force_sig(SIGSEGV, current); + + return 0; +} +#endif /* CONFIG_X86_32 */ + +static long do_rt_sigreturn(struct pt_regs *regs) +{ + struct rt_sigframe __user *frame; + unsigned long ax; + sigset_t set; + + frame = (struct rt_sigframe __user *)(regs->sp - sizeof(long)); + if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) + goto badframe; + if (__copy_from_user(&set, &frame->uc.uc_sigmask, sizeof(set))) + goto badframe; + + sigdelsetmask(&set, ~_BLOCKABLE); + spin_lock_irq(¤t->sighand->siglock); + current->blocked = set; + recalc_sigpending(); + spin_unlock_irq(¤t->sighand->siglock); + + if (restore_sigcontext(regs, &frame->uc.uc_mcontext, &ax)) + goto badframe; + + if (do_sigaltstack(&frame->uc.uc_stack, NULL, regs->sp) == -EFAULT) + goto badframe; + + return ax; + +badframe: + signal_fault(regs, frame, "rt_sigreturn"); + return 0; +} + +#ifdef CONFIG_X86_32 +asmlinkage int sys_rt_sigreturn(unsigned long __unused) +{ + struct pt_regs *regs = (struct pt_regs *)&__unused; + + return do_rt_sigreturn(regs); +} +#else /* !CONFIG_X86_32 */ +asmlinkage long sys_rt_sigreturn(struct pt_regs *regs) +{ + return do_rt_sigreturn(regs); +} +#endif /* CONFIG_X86_32 */ + +/* + * OK, we're invoking a handler: + */ +static int signr_convert(int sig) +{ +#ifdef CONFIG_X86_32 + struct thread_info *info = current_thread_info(); + + if (info->exec_domain && info->exec_domain->signal_invmap && sig < 32) + return info->exec_domain->signal_invmap[sig]; +#endif /* CONFIG_X86_32 */ + return sig; +} + +#ifdef CONFIG_X86_32 + +#define is_ia32 1 +#define ia32_setup_frame __setup_frame +#define ia32_setup_rt_frame __setup_rt_frame + +#else /* !CONFIG_X86_32 */ + +#ifdef CONFIG_IA32_EMULATION +#define is_ia32 test_thread_flag(TIF_IA32) +#else /* !CONFIG_IA32_EMULATION */ +#define is_ia32 0 +#endif /* CONFIG_IA32_EMULATION */ + +#endif /* CONFIG_X86_32 */ + +static int +setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, + sigset_t *set, struct pt_regs *regs) +{ + int usig = signr_convert(sig); + int ret; + + /* Set up the stack frame */ + if (is_ia32) { + if (ka->sa.sa_flags & SA_SIGINFO) + ret = ia32_setup_rt_frame(usig, ka, info, set, regs); + else + ret = ia32_setup_frame(usig, ka, set, regs); + } else + ret = __setup_rt_frame(sig, ka, info, set, regs); + + if (ret) { + force_sigsegv(sig, current); + return -EFAULT; + } + + return ret; +} + +static int +handle_signal(unsigned long sig, siginfo_t *info, struct k_sigaction *ka, + sigset_t *oldset, struct pt_regs *regs) +{ + int ret; + + /* Are we from a system call? */ + if (syscall_get_nr(current, regs) >= 0) { + /* If so, check system call restarting.. */ + switch (syscall_get_error(current, regs)) { + case -ERESTART_RESTARTBLOCK: + case -ERESTARTNOHAND: + regs->ax = -EINTR; + break; + + case -ERESTARTSYS: + if (!(ka->sa.sa_flags & SA_RESTART)) { + regs->ax = -EINTR; + break; + } + /* fallthrough */ + case -ERESTARTNOINTR: + regs->ax = regs->orig_ax; + regs->ip -= 2; + break; + } + } + + /* + * If TF is set due to a debugger (TIF_FORCED_TF), clear the TF + * flag so that register information in the sigcontext is correct. + */ + if (unlikely(regs->flags & X86_EFLAGS_TF) && + likely(test_and_clear_thread_flag(TIF_FORCED_TF))) + regs->flags &= ~X86_EFLAGS_TF; + + ret = setup_rt_frame(sig, ka, info, oldset, regs); + + if (ret) + return ret; + +#ifdef CONFIG_X86_64 + /* + * This has nothing to do with segment registers, + * despite the name. This magic affects uaccess.h + * macros' behavior. Reset it to the normal setting. + */ + set_fs(USER_DS); +#endif + + /* + * Clear the direction flag as per the ABI for function entry. + */ + regs->flags &= ~X86_EFLAGS_DF; + + /* + * Clear TF when entering the signal handler, but + * notify any tracer that was single-stepping it. + * The tracer may want to single-step inside the + * handler too. + */ + regs->flags &= ~X86_EFLAGS_TF; + + spin_lock_irq(¤t->sighand->siglock); + sigorsets(¤t->blocked, ¤t->blocked, &ka->sa.sa_mask); + if (!(ka->sa.sa_flags & SA_NODEFER)) + sigaddset(¤t->blocked, sig); + recalc_sigpending(); + spin_unlock_irq(¤t->sighand->siglock); + + tracehook_signal_handler(sig, info, ka, regs, + test_thread_flag(TIF_SINGLESTEP)); + + return 0; +} + +#ifdef CONFIG_X86_32 +#define NR_restart_syscall __NR_restart_syscall +#else /* !CONFIG_X86_32 */ +#define NR_restart_syscall \ + test_thread_flag(TIF_IA32) ? __NR_ia32_restart_syscall : __NR_restart_syscall +#endif /* CONFIG_X86_32 */ + +/* + * Note that 'init' is a special process: it doesn't get signals it doesn't + * want to handle. Thus you cannot kill init even with a SIGKILL even by + * mistake. + */ +static void do_signal(struct pt_regs *regs) +{ + struct k_sigaction ka; + siginfo_t info; + int signr; + sigset_t *oldset; + + /* + * We want the common case to go fast, which is why we may in certain + * cases get here from kernel mode. Just return without doing anything + * if so. + * X86_32: vm86 regs switched out by assembly code before reaching + * here, so testing against kernel CS suffices. + */ + if (!user_mode(regs)) + return; + + if (current_thread_info()->status & TS_RESTORE_SIGMASK) + oldset = ¤t->saved_sigmask; + else + oldset = ¤t->blocked; + + signr = get_signal_to_deliver(&info, &ka, regs, NULL); + if (signr > 0) { + /* + * Re-enable any watchpoints before delivering the + * signal to user space. The processor register will + * have been cleared if the watchpoint triggered + * inside the kernel. + */ + if (current->thread.debugreg7) + set_debugreg(current->thread.debugreg7, 7); + + /* Whee! Actually deliver the signal. */ + if (handle_signal(signr, &info, &ka, oldset, regs) == 0) { + /* + * A signal was successfully delivered; the saved + * sigmask will have been stored in the signal frame, + * and will be restored by sigreturn, so we can simply + * clear the TS_RESTORE_SIGMASK flag. + */ + current_thread_info()->status &= ~TS_RESTORE_SIGMASK; + } + return; + } + + /* Did we come from a system call? */ + if (syscall_get_nr(current, regs) >= 0) { + /* Restart the system call - no handlers present */ + switch (syscall_get_error(current, regs)) { + case -ERESTARTNOHAND: + case -ERESTARTSYS: + case -ERESTARTNOINTR: + regs->ax = regs->orig_ax; + regs->ip -= 2; + break; + + case -ERESTART_RESTARTBLOCK: + regs->ax = NR_restart_syscall; + regs->ip -= 2; + break; + } + } + + /* + * If there's no signal to deliver, we just put the saved sigmask + * back. + */ + if (current_thread_info()->status & TS_RESTORE_SIGMASK) { + current_thread_info()->status &= ~TS_RESTORE_SIGMASK; + sigprocmask(SIG_SETMASK, ¤t->saved_sigmask, NULL); + } +} + +/* + * notification of userspace execution resumption + * - triggered by the TIF_WORK_MASK flags + */ +void +do_notify_resume(struct pt_regs *regs, void *unused, __u32 thread_info_flags) +{ +#if defined(CONFIG_X86_64) && defined(CONFIG_X86_MCE) + /* notify userspace of pending MCEs */ + if (thread_info_flags & _TIF_MCE_NOTIFY) + mce_notify_user(); +#endif /* CONFIG_X86_64 && CONFIG_X86_MCE */ + + /* deal with pending signal delivery */ + if (thread_info_flags & _TIF_SIGPENDING) + do_signal(regs); + + if (thread_info_flags & _TIF_NOTIFY_RESUME) { + clear_thread_flag(TIF_NOTIFY_RESUME); + tracehook_notify_resume(regs); + } + +#ifdef CONFIG_X86_32 + clear_thread_flag(TIF_IRET); +#endif /* CONFIG_X86_32 */ +} + +void signal_fault(struct pt_regs *regs, void __user *frame, char *where) +{ + struct task_struct *me = current; + + if (show_unhandled_signals && printk_ratelimit()) { + printk(KERN_INFO + "%s[%d] bad frame in %s frame:%p ip:%lx sp:%lx orax:%lx", + me->comm, me->pid, where, frame, + regs->ip, regs->sp, regs->orig_ax); + print_vma_addr(" in ", regs->ip); + printk(KERN_CONT "\n"); + } + + force_sig(SIGSEGV, me); +} diff --git a/arch/x86/kernel/signal_32.c b/arch/x86/kernel/signal_32.c deleted file mode 100644 index b1f4d34e0a38..000000000000 --- a/arch/x86/kernel/signal_32.c +++ /dev/null @@ -1,915 +0,0 @@ -/* - * Copyright (C) 1991, 1992 Linus Torvalds - * Copyright (C) 2000, 2001, 2002 Andi Kleen SuSE Labs - * - * 1997-11-28 Modified for POSIX.1b signals by Richard Henderson - * 2000-06-20 Pentium III FXSR, SSE support by Gareth Hughes - * 2000-2002 x86-64 support by Andi Kleen - */ - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include -#include -#include -#include - -#ifdef CONFIG_X86_64 -#include -#include -#include -#endif /* CONFIG_X86_64 */ - -#include -#include - -#include "sigframe.h" - -#define _BLOCKABLE (~(sigmask(SIGKILL) | sigmask(SIGSTOP))) - -#define __FIX_EFLAGS (X86_EFLAGS_AC | X86_EFLAGS_OF | \ - X86_EFLAGS_DF | X86_EFLAGS_TF | X86_EFLAGS_SF | \ - X86_EFLAGS_ZF | X86_EFLAGS_AF | X86_EFLAGS_PF | \ - X86_EFLAGS_CF) - -#ifdef CONFIG_X86_32 -# define FIX_EFLAGS (__FIX_EFLAGS | X86_EFLAGS_RF) -#else -# define FIX_EFLAGS __FIX_EFLAGS -#endif - -#define COPY(x) { \ - err |= __get_user(regs->x, &sc->x); \ -} - -#define COPY_SEG(seg) { \ - unsigned short tmp; \ - err |= __get_user(tmp, &sc->seg); \ - regs->seg = tmp; \ -} - -#define COPY_SEG_CPL3(seg) { \ - unsigned short tmp; \ - err |= __get_user(tmp, &sc->seg); \ - regs->seg = tmp | 3; \ -} - -#define GET_SEG(seg) { \ - unsigned short tmp; \ - err |= __get_user(tmp, &sc->seg); \ - loadsegment(seg, tmp); \ -} - -static int -restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc, - unsigned long *pax) -{ - void __user *buf; - unsigned int tmpflags; - unsigned int err = 0; - - /* Always make any pending restarted system calls return -EINTR */ - current_thread_info()->restart_block.fn = do_no_restart_syscall; - -#ifdef CONFIG_X86_32 - GET_SEG(gs); - COPY_SEG(fs); - COPY_SEG(es); - COPY_SEG(ds); -#endif /* CONFIG_X86_32 */ - - COPY(di); COPY(si); COPY(bp); COPY(sp); COPY(bx); - COPY(dx); COPY(cx); COPY(ip); - -#ifdef CONFIG_X86_64 - COPY(r8); - COPY(r9); - COPY(r10); - COPY(r11); - COPY(r12); - COPY(r13); - COPY(r14); - COPY(r15); -#endif /* CONFIG_X86_64 */ - -#ifdef CONFIG_X86_32 - COPY_SEG_CPL3(cs); - COPY_SEG_CPL3(ss); -#else /* !CONFIG_X86_32 */ - /* Kernel saves and restores only the CS segment register on signals, - * which is the bare minimum needed to allow mixed 32/64-bit code. - * App's signal handler can save/restore other segments if needed. */ - COPY_SEG_CPL3(cs); -#endif /* CONFIG_X86_32 */ - - err |= __get_user(tmpflags, &sc->flags); - regs->flags = (regs->flags & ~FIX_EFLAGS) | (tmpflags & FIX_EFLAGS); - regs->orig_ax = -1; /* disable syscall checks */ - - err |= __get_user(buf, &sc->fpstate); - err |= restore_i387_xstate(buf); - - err |= __get_user(*pax, &sc->ax); - return err; -} - -static int -setup_sigcontext(struct sigcontext __user *sc, void __user *fpstate, - struct pt_regs *regs, unsigned long mask) -{ - int err = 0; - -#ifdef CONFIG_X86_32 - { - unsigned int tmp; - - savesegment(gs, tmp); - err |= __put_user(tmp, (unsigned int __user *)&sc->gs); - } - err |= __put_user(regs->fs, (unsigned int __user *)&sc->fs); - err |= __put_user(regs->es, (unsigned int __user *)&sc->es); - err |= __put_user(regs->ds, (unsigned int __user *)&sc->ds); -#endif /* CONFIG_X86_32 */ - - err |= __put_user(regs->di, &sc->di); - err |= __put_user(regs->si, &sc->si); - err |= __put_user(regs->bp, &sc->bp); - err |= __put_user(regs->sp, &sc->sp); - err |= __put_user(regs->bx, &sc->bx); - err |= __put_user(regs->dx, &sc->dx); - err |= __put_user(regs->cx, &sc->cx); - err |= __put_user(regs->ax, &sc->ax); -#ifdef CONFIG_X86_64 - err |= __put_user(regs->r8, &sc->r8); - err |= __put_user(regs->r9, &sc->r9); - err |= __put_user(regs->r10, &sc->r10); - err |= __put_user(regs->r11, &sc->r11); - err |= __put_user(regs->r12, &sc->r12); - err |= __put_user(regs->r13, &sc->r13); - err |= __put_user(regs->r14, &sc->r14); - err |= __put_user(regs->r15, &sc->r15); -#endif /* CONFIG_X86_64 */ - - err |= __put_user(current->thread.trap_no, &sc->trapno); - err |= __put_user(current->thread.error_code, &sc->err); - err |= __put_user(regs->ip, &sc->ip); -#ifdef CONFIG_X86_32 - err |= __put_user(regs->cs, (unsigned int __user *)&sc->cs); - err |= __put_user(regs->flags, &sc->flags); - err |= __put_user(regs->sp, &sc->sp_at_signal); - err |= __put_user(regs->ss, (unsigned int __user *)&sc->ss); -#else /* !CONFIG_X86_32 */ - err |= __put_user(regs->flags, &sc->flags); - err |= __put_user(regs->cs, &sc->cs); - err |= __put_user(0, &sc->gs); - err |= __put_user(0, &sc->fs); -#endif /* CONFIG_X86_32 */ - - err |= __put_user(fpstate, &sc->fpstate); - - /* non-iBCS2 extensions.. */ - err |= __put_user(mask, &sc->oldmask); - err |= __put_user(current->thread.cr2, &sc->cr2); - - return err; -} - -/* - * Set up a signal frame. - */ -#ifdef CONFIG_X86_32 -static const struct { - u16 poplmovl; - u32 val; - u16 int80; -} __attribute__((packed)) retcode = { - 0xb858, /* popl %eax; movl $..., %eax */ - __NR_sigreturn, - 0x80cd, /* int $0x80 */ -}; - -static const struct { - u8 movl; - u32 val; - u16 int80; - u8 pad; -} __attribute__((packed)) rt_retcode = { - 0xb8, /* movl $..., %eax */ - __NR_rt_sigreturn, - 0x80cd, /* int $0x80 */ - 0 -}; - -/* - * Determine which stack to use.. - */ -static inline void __user * -get_sigframe(struct k_sigaction *ka, struct pt_regs *regs, size_t frame_size, - void **fpstate) -{ - unsigned long sp; - - /* Default to using normal stack */ - sp = regs->sp; - - /* - * If we are on the alternate signal stack and would overflow it, don't. - * Return an always-bogus address instead so we will die with SIGSEGV. - */ - if (on_sig_stack(sp) && !likely(on_sig_stack(sp - frame_size))) - return (void __user *) -1L; - - /* This is the X/Open sanctioned signal stack switching. */ - if (ka->sa.sa_flags & SA_ONSTACK) { - if (sas_ss_flags(sp) == 0) - sp = current->sas_ss_sp + current->sas_ss_size; - } else { - /* This is the legacy signal stack switching. */ - if ((regs->ss & 0xffff) != __USER_DS && - !(ka->sa.sa_flags & SA_RESTORER) && - ka->sa.sa_restorer) - sp = (unsigned long) ka->sa.sa_restorer; - } - - if (used_math()) { - sp = sp - sig_xstate_size; - *fpstate = (struct _fpstate *) sp; - if (save_i387_xstate(*fpstate) < 0) - return (void __user *)-1L; - } - - sp -= frame_size; - /* - * Align the stack pointer according to the i386 ABI, - * i.e. so that on function entry ((sp + 4) & 15) == 0. - */ - sp = ((sp + 4) & -16ul) - 4; - - return (void __user *) sp; -} - -static int -__setup_frame(int sig, struct k_sigaction *ka, sigset_t *set, - struct pt_regs *regs) -{ - struct sigframe __user *frame; - void __user *restorer; - int err = 0; - void __user *fpstate = NULL; - - frame = get_sigframe(ka, regs, sizeof(*frame), &fpstate); - - if (!access_ok(VERIFY_WRITE, frame, sizeof(*frame))) - return -EFAULT; - - if (__put_user(sig, &frame->sig)) - return -EFAULT; - - if (setup_sigcontext(&frame->sc, fpstate, regs, set->sig[0])) - return -EFAULT; - - if (_NSIG_WORDS > 1) { - if (__copy_to_user(&frame->extramask, &set->sig[1], - sizeof(frame->extramask))) - return -EFAULT; - } - - if (current->mm->context.vdso) - restorer = VDSO32_SYMBOL(current->mm->context.vdso, sigreturn); - else - restorer = &frame->retcode; - if (ka->sa.sa_flags & SA_RESTORER) - restorer = ka->sa.sa_restorer; - - /* Set up to return from userspace. */ - err |= __put_user(restorer, &frame->pretcode); - - /* - * This is popl %eax ; movl $__NR_sigreturn, %eax ; int $0x80 - * - * WE DO NOT USE IT ANY MORE! It's only left here for historical - * reasons and because gdb uses it as a signature to notice - * signal handler stack frames. - */ - err |= __put_user(*((u64 *)&retcode), (u64 *)frame->retcode); - - if (err) - return -EFAULT; - - /* Set up registers for signal handler */ - regs->sp = (unsigned long)frame; - regs->ip = (unsigned long)ka->sa.sa_handler; - regs->ax = (unsigned long)sig; - regs->dx = 0; - regs->cx = 0; - - regs->ds = __USER_DS; - regs->es = __USER_DS; - regs->ss = __USER_DS; - regs->cs = __USER_CS; - - return 0; -} - -static int __setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, - sigset_t *set, struct pt_regs *regs) -{ - struct rt_sigframe __user *frame; - void __user *restorer; - int err = 0; - void __user *fpstate = NULL; - - frame = get_sigframe(ka, regs, sizeof(*frame), &fpstate); - - if (!access_ok(VERIFY_WRITE, frame, sizeof(*frame))) - return -EFAULT; - - err |= __put_user(sig, &frame->sig); - err |= __put_user(&frame->info, &frame->pinfo); - err |= __put_user(&frame->uc, &frame->puc); - err |= copy_siginfo_to_user(&frame->info, info); - if (err) - return -EFAULT; - - /* Create the ucontext. */ - if (cpu_has_xsave) - err |= __put_user(UC_FP_XSTATE, &frame->uc.uc_flags); - else - err |= __put_user(0, &frame->uc.uc_flags); - err |= __put_user(0, &frame->uc.uc_link); - err |= __put_user(current->sas_ss_sp, &frame->uc.uc_stack.ss_sp); - err |= __put_user(sas_ss_flags(regs->sp), - &frame->uc.uc_stack.ss_flags); - err |= __put_user(current->sas_ss_size, &frame->uc.uc_stack.ss_size); - err |= setup_sigcontext(&frame->uc.uc_mcontext, fpstate, - regs, set->sig[0]); - err |= __copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set)); - if (err) - return -EFAULT; - - /* Set up to return from userspace. */ - restorer = VDSO32_SYMBOL(current->mm->context.vdso, rt_sigreturn); - if (ka->sa.sa_flags & SA_RESTORER) - restorer = ka->sa.sa_restorer; - err |= __put_user(restorer, &frame->pretcode); - - /* - * This is movl $__NR_rt_sigreturn, %ax ; int $0x80 - * - * WE DO NOT USE IT ANY MORE! It's only left here for historical - * reasons and because gdb uses it as a signature to notice - * signal handler stack frames. - */ - err |= __put_user(*((u64 *)&rt_retcode), (u64 *)frame->retcode); - - if (err) - return -EFAULT; - - /* Set up registers for signal handler */ - regs->sp = (unsigned long)frame; - regs->ip = (unsigned long)ka->sa.sa_handler; - regs->ax = (unsigned long)sig; - regs->dx = (unsigned long)&frame->info; - regs->cx = (unsigned long)&frame->uc; - - regs->ds = __USER_DS; - regs->es = __USER_DS; - regs->ss = __USER_DS; - regs->cs = __USER_CS; - - return 0; -} -#else /* !CONFIG_X86_32 */ -/* - * Determine which stack to use.. - */ -static void __user * -get_stack(struct k_sigaction *ka, unsigned long sp, unsigned long size) -{ - /* Default to using normal stack - redzone*/ - sp -= 128; - - /* This is the X/Open sanctioned signal stack switching. */ - if (ka->sa.sa_flags & SA_ONSTACK) { - if (sas_ss_flags(sp) == 0) - sp = current->sas_ss_sp + current->sas_ss_size; - } - - return (void __user *)round_down(sp - size, 64); -} - -static int __setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, - sigset_t *set, struct pt_regs *regs) -{ - struct rt_sigframe __user *frame; - void __user *fp = NULL; - int err = 0; - struct task_struct *me = current; - - if (used_math()) { - fp = get_stack(ka, regs->sp, sig_xstate_size); - frame = (void __user *)round_down( - (unsigned long)fp - sizeof(struct rt_sigframe), 16) - 8; - - if (save_i387_xstate(fp) < 0) - return -EFAULT; - } else - frame = get_stack(ka, regs->sp, sizeof(struct rt_sigframe)) - 8; - - if (!access_ok(VERIFY_WRITE, frame, sizeof(*frame))) - return -EFAULT; - - if (ka->sa.sa_flags & SA_SIGINFO) { - if (copy_siginfo_to_user(&frame->info, info)) - return -EFAULT; - } - - /* Create the ucontext. */ - if (cpu_has_xsave) - err |= __put_user(UC_FP_XSTATE, &frame->uc.uc_flags); - else - err |= __put_user(0, &frame->uc.uc_flags); - err |= __put_user(0, &frame->uc.uc_link); - err |= __put_user(me->sas_ss_sp, &frame->uc.uc_stack.ss_sp); - err |= __put_user(sas_ss_flags(regs->sp), - &frame->uc.uc_stack.ss_flags); - err |= __put_user(me->sas_ss_size, &frame->uc.uc_stack.ss_size); - err |= setup_sigcontext(&frame->uc.uc_mcontext, fp, regs, set->sig[0]); - err |= __copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set)); - - /* Set up to return from userspace. If provided, use a stub - already in userspace. */ - /* x86-64 should always use SA_RESTORER. */ - if (ka->sa.sa_flags & SA_RESTORER) { - err |= __put_user(ka->sa.sa_restorer, &frame->pretcode); - } else { - /* could use a vstub here */ - return -EFAULT; - } - - if (err) - return -EFAULT; - - /* Set up registers for signal handler */ - regs->di = sig; - /* In case the signal handler was declared without prototypes */ - regs->ax = 0; - - /* This also works for non SA_SIGINFO handlers because they expect the - next argument after the signal number on the stack. */ - regs->si = (unsigned long)&frame->info; - regs->dx = (unsigned long)&frame->uc; - regs->ip = (unsigned long) ka->sa.sa_handler; - - regs->sp = (unsigned long)frame; - - /* Set up the CS register to run signal handlers in 64-bit mode, - even if the handler happens to be interrupting 32-bit code. */ - regs->cs = __USER_CS; - - return 0; -} -#endif /* CONFIG_X86_32 */ - -#ifdef CONFIG_X86_32 -/* - * Atomically swap in the new signal mask, and wait for a signal. - */ -asmlinkage int -sys_sigsuspend(int history0, int history1, old_sigset_t mask) -{ - mask &= _BLOCKABLE; - spin_lock_irq(¤t->sighand->siglock); - current->saved_sigmask = current->blocked; - siginitset(¤t->blocked, mask); - recalc_sigpending(); - spin_unlock_irq(¤t->sighand->siglock); - - current->state = TASK_INTERRUPTIBLE; - schedule(); - set_restore_sigmask(); - - return -ERESTARTNOHAND; -} - -asmlinkage int -sys_sigaction(int sig, const struct old_sigaction __user *act, - struct old_sigaction __user *oact) -{ - struct k_sigaction new_ka, old_ka; - int ret; - - if (act) { - old_sigset_t mask; - - if (!access_ok(VERIFY_READ, act, sizeof(*act)) || - __get_user(new_ka.sa.sa_handler, &act->sa_handler) || - __get_user(new_ka.sa.sa_restorer, &act->sa_restorer)) - return -EFAULT; - - __get_user(new_ka.sa.sa_flags, &act->sa_flags); - __get_user(mask, &act->sa_mask); - siginitset(&new_ka.sa.sa_mask, mask); - } - - ret = do_sigaction(sig, act ? &new_ka : NULL, oact ? &old_ka : NULL); - - if (!ret && oact) { - if (!access_ok(VERIFY_WRITE, oact, sizeof(*oact)) || - __put_user(old_ka.sa.sa_handler, &oact->sa_handler) || - __put_user(old_ka.sa.sa_restorer, &oact->sa_restorer)) - return -EFAULT; - - __put_user(old_ka.sa.sa_flags, &oact->sa_flags); - __put_user(old_ka.sa.sa_mask.sig[0], &oact->sa_mask); - } - - return ret; -} -#endif /* CONFIG_X86_32 */ - -#ifdef CONFIG_X86_32 -asmlinkage int sys_sigaltstack(unsigned long bx) -{ - /* - * This is needed to make gcc realize it doesn't own the - * "struct pt_regs" - */ - struct pt_regs *regs = (struct pt_regs *)&bx; - const stack_t __user *uss = (const stack_t __user *)bx; - stack_t __user *uoss = (stack_t __user *)regs->cx; - - return do_sigaltstack(uss, uoss, regs->sp); -} -#else /* !CONFIG_X86_32 */ -asmlinkage long -sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss, - struct pt_regs *regs) -{ - return do_sigaltstack(uss, uoss, regs->sp); -} -#endif /* CONFIG_X86_32 */ - -/* - * Do a signal return; undo the signal stack. - */ -#ifdef CONFIG_X86_32 -asmlinkage unsigned long sys_sigreturn(unsigned long __unused) -{ - struct sigframe __user *frame; - struct pt_regs *regs; - unsigned long ax; - sigset_t set; - - regs = (struct pt_regs *) &__unused; - frame = (struct sigframe __user *)(regs->sp - 8); - - if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) - goto badframe; - if (__get_user(set.sig[0], &frame->sc.oldmask) || (_NSIG_WORDS > 1 - && __copy_from_user(&set.sig[1], &frame->extramask, - sizeof(frame->extramask)))) - goto badframe; - - sigdelsetmask(&set, ~_BLOCKABLE); - spin_lock_irq(¤t->sighand->siglock); - current->blocked = set; - recalc_sigpending(); - spin_unlock_irq(¤t->sighand->siglock); - - if (restore_sigcontext(regs, &frame->sc, &ax)) - goto badframe; - return ax; - -badframe: - if (show_unhandled_signals && printk_ratelimit()) { - printk("%s%s[%d] bad frame in sigreturn frame:" - "%p ip:%lx sp:%lx oeax:%lx", - task_pid_nr(current) > 1 ? KERN_INFO : KERN_EMERG, - current->comm, task_pid_nr(current), frame, regs->ip, - regs->sp, regs->orig_ax); - print_vma_addr(" in ", regs->ip); - printk(KERN_CONT "\n"); - } - - force_sig(SIGSEGV, current); - - return 0; -} -#endif /* CONFIG_X86_32 */ - -static long do_rt_sigreturn(struct pt_regs *regs) -{ - struct rt_sigframe __user *frame; - unsigned long ax; - sigset_t set; - - frame = (struct rt_sigframe __user *)(regs->sp - sizeof(long)); - if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) - goto badframe; - if (__copy_from_user(&set, &frame->uc.uc_sigmask, sizeof(set))) - goto badframe; - - sigdelsetmask(&set, ~_BLOCKABLE); - spin_lock_irq(¤t->sighand->siglock); - current->blocked = set; - recalc_sigpending(); - spin_unlock_irq(¤t->sighand->siglock); - - if (restore_sigcontext(regs, &frame->uc.uc_mcontext, &ax)) - goto badframe; - - if (do_sigaltstack(&frame->uc.uc_stack, NULL, regs->sp) == -EFAULT) - goto badframe; - - return ax; - -badframe: - signal_fault(regs, frame, "rt_sigreturn"); - return 0; -} - -#ifdef CONFIG_X86_32 -asmlinkage int sys_rt_sigreturn(unsigned long __unused) -{ - struct pt_regs *regs = (struct pt_regs *)&__unused; - - return do_rt_sigreturn(regs); -} -#else /* !CONFIG_X86_32 */ -asmlinkage long sys_rt_sigreturn(struct pt_regs *regs) -{ - return do_rt_sigreturn(regs); -} -#endif /* CONFIG_X86_32 */ - -/* - * OK, we're invoking a handler: - */ -static int signr_convert(int sig) -{ -#ifdef CONFIG_X86_32 - struct thread_info *info = current_thread_info(); - - if (info->exec_domain && info->exec_domain->signal_invmap && sig < 32) - return info->exec_domain->signal_invmap[sig]; -#endif /* CONFIG_X86_32 */ - return sig; -} - -#ifdef CONFIG_X86_32 - -#define is_ia32 1 -#define ia32_setup_frame __setup_frame -#define ia32_setup_rt_frame __setup_rt_frame - -#else /* !CONFIG_X86_32 */ - -#ifdef CONFIG_IA32_EMULATION -#define is_ia32 test_thread_flag(TIF_IA32) -#else /* !CONFIG_IA32_EMULATION */ -#define is_ia32 0 -#endif /* CONFIG_IA32_EMULATION */ - -#endif /* CONFIG_X86_32 */ - -static int -setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, - sigset_t *set, struct pt_regs *regs) -{ - int usig = signr_convert(sig); - int ret; - - /* Set up the stack frame */ - if (is_ia32) { - if (ka->sa.sa_flags & SA_SIGINFO) - ret = ia32_setup_rt_frame(usig, ka, info, set, regs); - else - ret = ia32_setup_frame(usig, ka, set, regs); - } else - ret = __setup_rt_frame(sig, ka, info, set, regs); - - if (ret) { - force_sigsegv(sig, current); - return -EFAULT; - } - - return ret; -} - -static int -handle_signal(unsigned long sig, siginfo_t *info, struct k_sigaction *ka, - sigset_t *oldset, struct pt_regs *regs) -{ - int ret; - - /* Are we from a system call? */ - if (syscall_get_nr(current, regs) >= 0) { - /* If so, check system call restarting.. */ - switch (syscall_get_error(current, regs)) { - case -ERESTART_RESTARTBLOCK: - case -ERESTARTNOHAND: - regs->ax = -EINTR; - break; - - case -ERESTARTSYS: - if (!(ka->sa.sa_flags & SA_RESTART)) { - regs->ax = -EINTR; - break; - } - /* fallthrough */ - case -ERESTARTNOINTR: - regs->ax = regs->orig_ax; - regs->ip -= 2; - break; - } - } - - /* - * If TF is set due to a debugger (TIF_FORCED_TF), clear the TF - * flag so that register information in the sigcontext is correct. - */ - if (unlikely(regs->flags & X86_EFLAGS_TF) && - likely(test_and_clear_thread_flag(TIF_FORCED_TF))) - regs->flags &= ~X86_EFLAGS_TF; - - ret = setup_rt_frame(sig, ka, info, oldset, regs); - - if (ret) - return ret; - -#ifdef CONFIG_X86_64 - /* - * This has nothing to do with segment registers, - * despite the name. This magic affects uaccess.h - * macros' behavior. Reset it to the normal setting. - */ - set_fs(USER_DS); -#endif - - /* - * Clear the direction flag as per the ABI for function entry. - */ - regs->flags &= ~X86_EFLAGS_DF; - - /* - * Clear TF when entering the signal handler, but - * notify any tracer that was single-stepping it. - * The tracer may want to single-step inside the - * handler too. - */ - regs->flags &= ~X86_EFLAGS_TF; - - spin_lock_irq(¤t->sighand->siglock); - sigorsets(¤t->blocked, ¤t->blocked, &ka->sa.sa_mask); - if (!(ka->sa.sa_flags & SA_NODEFER)) - sigaddset(¤t->blocked, sig); - recalc_sigpending(); - spin_unlock_irq(¤t->sighand->siglock); - - tracehook_signal_handler(sig, info, ka, regs, - test_thread_flag(TIF_SINGLESTEP)); - - return 0; -} - -#ifdef CONFIG_X86_32 -#define NR_restart_syscall __NR_restart_syscall -#else /* !CONFIG_X86_32 */ -#define NR_restart_syscall \ - test_thread_flag(TIF_IA32) ? __NR_ia32_restart_syscall : __NR_restart_syscall -#endif /* CONFIG_X86_32 */ - -/* - * Note that 'init' is a special process: it doesn't get signals it doesn't - * want to handle. Thus you cannot kill init even with a SIGKILL even by - * mistake. - */ -static void do_signal(struct pt_regs *regs) -{ - struct k_sigaction ka; - siginfo_t info; - int signr; - sigset_t *oldset; - - /* - * We want the common case to go fast, which is why we may in certain - * cases get here from kernel mode. Just return without doing anything - * if so. - * X86_32: vm86 regs switched out by assembly code before reaching - * here, so testing against kernel CS suffices. - */ - if (!user_mode(regs)) - return; - - if (current_thread_info()->status & TS_RESTORE_SIGMASK) - oldset = ¤t->saved_sigmask; - else - oldset = ¤t->blocked; - - signr = get_signal_to_deliver(&info, &ka, regs, NULL); - if (signr > 0) { - /* - * Re-enable any watchpoints before delivering the - * signal to user space. The processor register will - * have been cleared if the watchpoint triggered - * inside the kernel. - */ - if (current->thread.debugreg7) - set_debugreg(current->thread.debugreg7, 7); - - /* Whee! Actually deliver the signal. */ - if (handle_signal(signr, &info, &ka, oldset, regs) == 0) { - /* - * A signal was successfully delivered; the saved - * sigmask will have been stored in the signal frame, - * and will be restored by sigreturn, so we can simply - * clear the TS_RESTORE_SIGMASK flag. - */ - current_thread_info()->status &= ~TS_RESTORE_SIGMASK; - } - return; - } - - /* Did we come from a system call? */ - if (syscall_get_nr(current, regs) >= 0) { - /* Restart the system call - no handlers present */ - switch (syscall_get_error(current, regs)) { - case -ERESTARTNOHAND: - case -ERESTARTSYS: - case -ERESTARTNOINTR: - regs->ax = regs->orig_ax; - regs->ip -= 2; - break; - - case -ERESTART_RESTARTBLOCK: - regs->ax = NR_restart_syscall; - regs->ip -= 2; - break; - } - } - - /* - * If there's no signal to deliver, we just put the saved sigmask - * back. - */ - if (current_thread_info()->status & TS_RESTORE_SIGMASK) { - current_thread_info()->status &= ~TS_RESTORE_SIGMASK; - sigprocmask(SIG_SETMASK, ¤t->saved_sigmask, NULL); - } -} - -/* - * notification of userspace execution resumption - * - triggered by the TIF_WORK_MASK flags - */ -void -do_notify_resume(struct pt_regs *regs, void *unused, __u32 thread_info_flags) -{ -#if defined(CONFIG_X86_64) && defined(CONFIG_X86_MCE) - /* notify userspace of pending MCEs */ - if (thread_info_flags & _TIF_MCE_NOTIFY) - mce_notify_user(); -#endif /* CONFIG_X86_64 && CONFIG_X86_MCE */ - - /* deal with pending signal delivery */ - if (thread_info_flags & _TIF_SIGPENDING) - do_signal(regs); - - if (thread_info_flags & _TIF_NOTIFY_RESUME) { - clear_thread_flag(TIF_NOTIFY_RESUME); - tracehook_notify_resume(regs); - } - -#ifdef CONFIG_X86_32 - clear_thread_flag(TIF_IRET); -#endif /* CONFIG_X86_32 */ -} - -void signal_fault(struct pt_regs *regs, void __user *frame, char *where) -{ - struct task_struct *me = current; - - if (show_unhandled_signals && printk_ratelimit()) { - printk(KERN_INFO - "%s[%d] bad frame in %s frame:%p ip:%lx sp:%lx orax:%lx", - me->comm, me->pid, where, frame, - regs->ip, regs->sp, regs->orig_ax); - print_vma_addr(" in ", regs->ip); - printk(KERN_CONT "\n"); - } - - force_sig(SIGSEGV, me); -} diff --git a/arch/x86/kernel/signal_64.c b/arch/x86/kernel/signal_64.c deleted file mode 100644 index b1f4d34e0a38..000000000000 --- a/arch/x86/kernel/signal_64.c +++ /dev/null @@ -1,915 +0,0 @@ -/* - * Copyright (C) 1991, 1992 Linus Torvalds - * Copyright (C) 2000, 2001, 2002 Andi Kleen SuSE Labs - * - * 1997-11-28 Modified for POSIX.1b signals by Richard Henderson - * 2000-06-20 Pentium III FXSR, SSE support by Gareth Hughes - * 2000-2002 x86-64 support by Andi Kleen - */ - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include -#include -#include -#include - -#ifdef CONFIG_X86_64 -#include -#include -#include -#endif /* CONFIG_X86_64 */ - -#include -#include - -#include "sigframe.h" - -#define _BLOCKABLE (~(sigmask(SIGKILL) | sigmask(SIGSTOP))) - -#define __FIX_EFLAGS (X86_EFLAGS_AC | X86_EFLAGS_OF | \ - X86_EFLAGS_DF | X86_EFLAGS_TF | X86_EFLAGS_SF | \ - X86_EFLAGS_ZF | X86_EFLAGS_AF | X86_EFLAGS_PF | \ - X86_EFLAGS_CF) - -#ifdef CONFIG_X86_32 -# define FIX_EFLAGS (__FIX_EFLAGS | X86_EFLAGS_RF) -#else -# define FIX_EFLAGS __FIX_EFLAGS -#endif - -#define COPY(x) { \ - err |= __get_user(regs->x, &sc->x); \ -} - -#define COPY_SEG(seg) { \ - unsigned short tmp; \ - err |= __get_user(tmp, &sc->seg); \ - regs->seg = tmp; \ -} - -#define COPY_SEG_CPL3(seg) { \ - unsigned short tmp; \ - err |= __get_user(tmp, &sc->seg); \ - regs->seg = tmp | 3; \ -} - -#define GET_SEG(seg) { \ - unsigned short tmp; \ - err |= __get_user(tmp, &sc->seg); \ - loadsegment(seg, tmp); \ -} - -static int -restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc, - unsigned long *pax) -{ - void __user *buf; - unsigned int tmpflags; - unsigned int err = 0; - - /* Always make any pending restarted system calls return -EINTR */ - current_thread_info()->restart_block.fn = do_no_restart_syscall; - -#ifdef CONFIG_X86_32 - GET_SEG(gs); - COPY_SEG(fs); - COPY_SEG(es); - COPY_SEG(ds); -#endif /* CONFIG_X86_32 */ - - COPY(di); COPY(si); COPY(bp); COPY(sp); COPY(bx); - COPY(dx); COPY(cx); COPY(ip); - -#ifdef CONFIG_X86_64 - COPY(r8); - COPY(r9); - COPY(r10); - COPY(r11); - COPY(r12); - COPY(r13); - COPY(r14); - COPY(r15); -#endif /* CONFIG_X86_64 */ - -#ifdef CONFIG_X86_32 - COPY_SEG_CPL3(cs); - COPY_SEG_CPL3(ss); -#else /* !CONFIG_X86_32 */ - /* Kernel saves and restores only the CS segment register on signals, - * which is the bare minimum needed to allow mixed 32/64-bit code. - * App's signal handler can save/restore other segments if needed. */ - COPY_SEG_CPL3(cs); -#endif /* CONFIG_X86_32 */ - - err |= __get_user(tmpflags, &sc->flags); - regs->flags = (regs->flags & ~FIX_EFLAGS) | (tmpflags & FIX_EFLAGS); - regs->orig_ax = -1; /* disable syscall checks */ - - err |= __get_user(buf, &sc->fpstate); - err |= restore_i387_xstate(buf); - - err |= __get_user(*pax, &sc->ax); - return err; -} - -static int -setup_sigcontext(struct sigcontext __user *sc, void __user *fpstate, - struct pt_regs *regs, unsigned long mask) -{ - int err = 0; - -#ifdef CONFIG_X86_32 - { - unsigned int tmp; - - savesegment(gs, tmp); - err |= __put_user(tmp, (unsigned int __user *)&sc->gs); - } - err |= __put_user(regs->fs, (unsigned int __user *)&sc->fs); - err |= __put_user(regs->es, (unsigned int __user *)&sc->es); - err |= __put_user(regs->ds, (unsigned int __user *)&sc->ds); -#endif /* CONFIG_X86_32 */ - - err |= __put_user(regs->di, &sc->di); - err |= __put_user(regs->si, &sc->si); - err |= __put_user(regs->bp, &sc->bp); - err |= __put_user(regs->sp, &sc->sp); - err |= __put_user(regs->bx, &sc->bx); - err |= __put_user(regs->dx, &sc->dx); - err |= __put_user(regs->cx, &sc->cx); - err |= __put_user(regs->ax, &sc->ax); -#ifdef CONFIG_X86_64 - err |= __put_user(regs->r8, &sc->r8); - err |= __put_user(regs->r9, &sc->r9); - err |= __put_user(regs->r10, &sc->r10); - err |= __put_user(regs->r11, &sc->r11); - err |= __put_user(regs->r12, &sc->r12); - err |= __put_user(regs->r13, &sc->r13); - err |= __put_user(regs->r14, &sc->r14); - err |= __put_user(regs->r15, &sc->r15); -#endif /* CONFIG_X86_64 */ - - err |= __put_user(current->thread.trap_no, &sc->trapno); - err |= __put_user(current->thread.error_code, &sc->err); - err |= __put_user(regs->ip, &sc->ip); -#ifdef CONFIG_X86_32 - err |= __put_user(regs->cs, (unsigned int __user *)&sc->cs); - err |= __put_user(regs->flags, &sc->flags); - err |= __put_user(regs->sp, &sc->sp_at_signal); - err |= __put_user(regs->ss, (unsigned int __user *)&sc->ss); -#else /* !CONFIG_X86_32 */ - err |= __put_user(regs->flags, &sc->flags); - err |= __put_user(regs->cs, &sc->cs); - err |= __put_user(0, &sc->gs); - err |= __put_user(0, &sc->fs); -#endif /* CONFIG_X86_32 */ - - err |= __put_user(fpstate, &sc->fpstate); - - /* non-iBCS2 extensions.. */ - err |= __put_user(mask, &sc->oldmask); - err |= __put_user(current->thread.cr2, &sc->cr2); - - return err; -} - -/* - * Set up a signal frame. - */ -#ifdef CONFIG_X86_32 -static const struct { - u16 poplmovl; - u32 val; - u16 int80; -} __attribute__((packed)) retcode = { - 0xb858, /* popl %eax; movl $..., %eax */ - __NR_sigreturn, - 0x80cd, /* int $0x80 */ -}; - -static const struct { - u8 movl; - u32 val; - u16 int80; - u8 pad; -} __attribute__((packed)) rt_retcode = { - 0xb8, /* movl $..., %eax */ - __NR_rt_sigreturn, - 0x80cd, /* int $0x80 */ - 0 -}; - -/* - * Determine which stack to use.. - */ -static inline void __user * -get_sigframe(struct k_sigaction *ka, struct pt_regs *regs, size_t frame_size, - void **fpstate) -{ - unsigned long sp; - - /* Default to using normal stack */ - sp = regs->sp; - - /* - * If we are on the alternate signal stack and would overflow it, don't. - * Return an always-bogus address instead so we will die with SIGSEGV. - */ - if (on_sig_stack(sp) && !likely(on_sig_stack(sp - frame_size))) - return (void __user *) -1L; - - /* This is the X/Open sanctioned signal stack switching. */ - if (ka->sa.sa_flags & SA_ONSTACK) { - if (sas_ss_flags(sp) == 0) - sp = current->sas_ss_sp + current->sas_ss_size; - } else { - /* This is the legacy signal stack switching. */ - if ((regs->ss & 0xffff) != __USER_DS && - !(ka->sa.sa_flags & SA_RESTORER) && - ka->sa.sa_restorer) - sp = (unsigned long) ka->sa.sa_restorer; - } - - if (used_math()) { - sp = sp - sig_xstate_size; - *fpstate = (struct _fpstate *) sp; - if (save_i387_xstate(*fpstate) < 0) - return (void __user *)-1L; - } - - sp -= frame_size; - /* - * Align the stack pointer according to the i386 ABI, - * i.e. so that on function entry ((sp + 4) & 15) == 0. - */ - sp = ((sp + 4) & -16ul) - 4; - - return (void __user *) sp; -} - -static int -__setup_frame(int sig, struct k_sigaction *ka, sigset_t *set, - struct pt_regs *regs) -{ - struct sigframe __user *frame; - void __user *restorer; - int err = 0; - void __user *fpstate = NULL; - - frame = get_sigframe(ka, regs, sizeof(*frame), &fpstate); - - if (!access_ok(VERIFY_WRITE, frame, sizeof(*frame))) - return -EFAULT; - - if (__put_user(sig, &frame->sig)) - return -EFAULT; - - if (setup_sigcontext(&frame->sc, fpstate, regs, set->sig[0])) - return -EFAULT; - - if (_NSIG_WORDS > 1) { - if (__copy_to_user(&frame->extramask, &set->sig[1], - sizeof(frame->extramask))) - return -EFAULT; - } - - if (current->mm->context.vdso) - restorer = VDSO32_SYMBOL(current->mm->context.vdso, sigreturn); - else - restorer = &frame->retcode; - if (ka->sa.sa_flags & SA_RESTORER) - restorer = ka->sa.sa_restorer; - - /* Set up to return from userspace. */ - err |= __put_user(restorer, &frame->pretcode); - - /* - * This is popl %eax ; movl $__NR_sigreturn, %eax ; int $0x80 - * - * WE DO NOT USE IT ANY MORE! It's only left here for historical - * reasons and because gdb uses it as a signature to notice - * signal handler stack frames. - */ - err |= __put_user(*((u64 *)&retcode), (u64 *)frame->retcode); - - if (err) - return -EFAULT; - - /* Set up registers for signal handler */ - regs->sp = (unsigned long)frame; - regs->ip = (unsigned long)ka->sa.sa_handler; - regs->ax = (unsigned long)sig; - regs->dx = 0; - regs->cx = 0; - - regs->ds = __USER_DS; - regs->es = __USER_DS; - regs->ss = __USER_DS; - regs->cs = __USER_CS; - - return 0; -} - -static int __setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, - sigset_t *set, struct pt_regs *regs) -{ - struct rt_sigframe __user *frame; - void __user *restorer; - int err = 0; - void __user *fpstate = NULL; - - frame = get_sigframe(ka, regs, sizeof(*frame), &fpstate); - - if (!access_ok(VERIFY_WRITE, frame, sizeof(*frame))) - return -EFAULT; - - err |= __put_user(sig, &frame->sig); - err |= __put_user(&frame->info, &frame->pinfo); - err |= __put_user(&frame->uc, &frame->puc); - err |= copy_siginfo_to_user(&frame->info, info); - if (err) - return -EFAULT; - - /* Create the ucontext. */ - if (cpu_has_xsave) - err |= __put_user(UC_FP_XSTATE, &frame->uc.uc_flags); - else - err |= __put_user(0, &frame->uc.uc_flags); - err |= __put_user(0, &frame->uc.uc_link); - err |= __put_user(current->sas_ss_sp, &frame->uc.uc_stack.ss_sp); - err |= __put_user(sas_ss_flags(regs->sp), - &frame->uc.uc_stack.ss_flags); - err |= __put_user(current->sas_ss_size, &frame->uc.uc_stack.ss_size); - err |= setup_sigcontext(&frame->uc.uc_mcontext, fpstate, - regs, set->sig[0]); - err |= __copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set)); - if (err) - return -EFAULT; - - /* Set up to return from userspace. */ - restorer = VDSO32_SYMBOL(current->mm->context.vdso, rt_sigreturn); - if (ka->sa.sa_flags & SA_RESTORER) - restorer = ka->sa.sa_restorer; - err |= __put_user(restorer, &frame->pretcode); - - /* - * This is movl $__NR_rt_sigreturn, %ax ; int $0x80 - * - * WE DO NOT USE IT ANY MORE! It's only left here for historical - * reasons and because gdb uses it as a signature to notice - * signal handler stack frames. - */ - err |= __put_user(*((u64 *)&rt_retcode), (u64 *)frame->retcode); - - if (err) - return -EFAULT; - - /* Set up registers for signal handler */ - regs->sp = (unsigned long)frame; - regs->ip = (unsigned long)ka->sa.sa_handler; - regs->ax = (unsigned long)sig; - regs->dx = (unsigned long)&frame->info; - regs->cx = (unsigned long)&frame->uc; - - regs->ds = __USER_DS; - regs->es = __USER_DS; - regs->ss = __USER_DS; - regs->cs = __USER_CS; - - return 0; -} -#else /* !CONFIG_X86_32 */ -/* - * Determine which stack to use.. - */ -static void __user * -get_stack(struct k_sigaction *ka, unsigned long sp, unsigned long size) -{ - /* Default to using normal stack - redzone*/ - sp -= 128; - - /* This is the X/Open sanctioned signal stack switching. */ - if (ka->sa.sa_flags & SA_ONSTACK) { - if (sas_ss_flags(sp) == 0) - sp = current->sas_ss_sp + current->sas_ss_size; - } - - return (void __user *)round_down(sp - size, 64); -} - -static int __setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, - sigset_t *set, struct pt_regs *regs) -{ - struct rt_sigframe __user *frame; - void __user *fp = NULL; - int err = 0; - struct task_struct *me = current; - - if (used_math()) { - fp = get_stack(ka, regs->sp, sig_xstate_size); - frame = (void __user *)round_down( - (unsigned long)fp - sizeof(struct rt_sigframe), 16) - 8; - - if (save_i387_xstate(fp) < 0) - return -EFAULT; - } else - frame = get_stack(ka, regs->sp, sizeof(struct rt_sigframe)) - 8; - - if (!access_ok(VERIFY_WRITE, frame, sizeof(*frame))) - return -EFAULT; - - if (ka->sa.sa_flags & SA_SIGINFO) { - if (copy_siginfo_to_user(&frame->info, info)) - return -EFAULT; - } - - /* Create the ucontext. */ - if (cpu_has_xsave) - err |= __put_user(UC_FP_XSTATE, &frame->uc.uc_flags); - else - err |= __put_user(0, &frame->uc.uc_flags); - err |= __put_user(0, &frame->uc.uc_link); - err |= __put_user(me->sas_ss_sp, &frame->uc.uc_stack.ss_sp); - err |= __put_user(sas_ss_flags(regs->sp), - &frame->uc.uc_stack.ss_flags); - err |= __put_user(me->sas_ss_size, &frame->uc.uc_stack.ss_size); - err |= setup_sigcontext(&frame->uc.uc_mcontext, fp, regs, set->sig[0]); - err |= __copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set)); - - /* Set up to return from userspace. If provided, use a stub - already in userspace. */ - /* x86-64 should always use SA_RESTORER. */ - if (ka->sa.sa_flags & SA_RESTORER) { - err |= __put_user(ka->sa.sa_restorer, &frame->pretcode); - } else { - /* could use a vstub here */ - return -EFAULT; - } - - if (err) - return -EFAULT; - - /* Set up registers for signal handler */ - regs->di = sig; - /* In case the signal handler was declared without prototypes */ - regs->ax = 0; - - /* This also works for non SA_SIGINFO handlers because they expect the - next argument after the signal number on the stack. */ - regs->si = (unsigned long)&frame->info; - regs->dx = (unsigned long)&frame->uc; - regs->ip = (unsigned long) ka->sa.sa_handler; - - regs->sp = (unsigned long)frame; - - /* Set up the CS register to run signal handlers in 64-bit mode, - even if the handler happens to be interrupting 32-bit code. */ - regs->cs = __USER_CS; - - return 0; -} -#endif /* CONFIG_X86_32 */ - -#ifdef CONFIG_X86_32 -/* - * Atomically swap in the new signal mask, and wait for a signal. - */ -asmlinkage int -sys_sigsuspend(int history0, int history1, old_sigset_t mask) -{ - mask &= _BLOCKABLE; - spin_lock_irq(¤t->sighand->siglock); - current->saved_sigmask = current->blocked; - siginitset(¤t->blocked, mask); - recalc_sigpending(); - spin_unlock_irq(¤t->sighand->siglock); - - current->state = TASK_INTERRUPTIBLE; - schedule(); - set_restore_sigmask(); - - return -ERESTARTNOHAND; -} - -asmlinkage int -sys_sigaction(int sig, const struct old_sigaction __user *act, - struct old_sigaction __user *oact) -{ - struct k_sigaction new_ka, old_ka; - int ret; - - if (act) { - old_sigset_t mask; - - if (!access_ok(VERIFY_READ, act, sizeof(*act)) || - __get_user(new_ka.sa.sa_handler, &act->sa_handler) || - __get_user(new_ka.sa.sa_restorer, &act->sa_restorer)) - return -EFAULT; - - __get_user(new_ka.sa.sa_flags, &act->sa_flags); - __get_user(mask, &act->sa_mask); - siginitset(&new_ka.sa.sa_mask, mask); - } - - ret = do_sigaction(sig, act ? &new_ka : NULL, oact ? &old_ka : NULL); - - if (!ret && oact) { - if (!access_ok(VERIFY_WRITE, oact, sizeof(*oact)) || - __put_user(old_ka.sa.sa_handler, &oact->sa_handler) || - __put_user(old_ka.sa.sa_restorer, &oact->sa_restorer)) - return -EFAULT; - - __put_user(old_ka.sa.sa_flags, &oact->sa_flags); - __put_user(old_ka.sa.sa_mask.sig[0], &oact->sa_mask); - } - - return ret; -} -#endif /* CONFIG_X86_32 */ - -#ifdef CONFIG_X86_32 -asmlinkage int sys_sigaltstack(unsigned long bx) -{ - /* - * This is needed to make gcc realize it doesn't own the - * "struct pt_regs" - */ - struct pt_regs *regs = (struct pt_regs *)&bx; - const stack_t __user *uss = (const stack_t __user *)bx; - stack_t __user *uoss = (stack_t __user *)regs->cx; - - return do_sigaltstack(uss, uoss, regs->sp); -} -#else /* !CONFIG_X86_32 */ -asmlinkage long -sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss, - struct pt_regs *regs) -{ - return do_sigaltstack(uss, uoss, regs->sp); -} -#endif /* CONFIG_X86_32 */ - -/* - * Do a signal return; undo the signal stack. - */ -#ifdef CONFIG_X86_32 -asmlinkage unsigned long sys_sigreturn(unsigned long __unused) -{ - struct sigframe __user *frame; - struct pt_regs *regs; - unsigned long ax; - sigset_t set; - - regs = (struct pt_regs *) &__unused; - frame = (struct sigframe __user *)(regs->sp - 8); - - if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) - goto badframe; - if (__get_user(set.sig[0], &frame->sc.oldmask) || (_NSIG_WORDS > 1 - && __copy_from_user(&set.sig[1], &frame->extramask, - sizeof(frame->extramask)))) - goto badframe; - - sigdelsetmask(&set, ~_BLOCKABLE); - spin_lock_irq(¤t->sighand->siglock); - current->blocked = set; - recalc_sigpending(); - spin_unlock_irq(¤t->sighand->siglock); - - if (restore_sigcontext(regs, &frame->sc, &ax)) - goto badframe; - return ax; - -badframe: - if (show_unhandled_signals && printk_ratelimit()) { - printk("%s%s[%d] bad frame in sigreturn frame:" - "%p ip:%lx sp:%lx oeax:%lx", - task_pid_nr(current) > 1 ? KERN_INFO : KERN_EMERG, - current->comm, task_pid_nr(current), frame, regs->ip, - regs->sp, regs->orig_ax); - print_vma_addr(" in ", regs->ip); - printk(KERN_CONT "\n"); - } - - force_sig(SIGSEGV, current); - - return 0; -} -#endif /* CONFIG_X86_32 */ - -static long do_rt_sigreturn(struct pt_regs *regs) -{ - struct rt_sigframe __user *frame; - unsigned long ax; - sigset_t set; - - frame = (struct rt_sigframe __user *)(regs->sp - sizeof(long)); - if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) - goto badframe; - if (__copy_from_user(&set, &frame->uc.uc_sigmask, sizeof(set))) - goto badframe; - - sigdelsetmask(&set, ~_BLOCKABLE); - spin_lock_irq(¤t->sighand->siglock); - current->blocked = set; - recalc_sigpending(); - spin_unlock_irq(¤t->sighand->siglock); - - if (restore_sigcontext(regs, &frame->uc.uc_mcontext, &ax)) - goto badframe; - - if (do_sigaltstack(&frame->uc.uc_stack, NULL, regs->sp) == -EFAULT) - goto badframe; - - return ax; - -badframe: - signal_fault(regs, frame, "rt_sigreturn"); - return 0; -} - -#ifdef CONFIG_X86_32 -asmlinkage int sys_rt_sigreturn(unsigned long __unused) -{ - struct pt_regs *regs = (struct pt_regs *)&__unused; - - return do_rt_sigreturn(regs); -} -#else /* !CONFIG_X86_32 */ -asmlinkage long sys_rt_sigreturn(struct pt_regs *regs) -{ - return do_rt_sigreturn(regs); -} -#endif /* CONFIG_X86_32 */ - -/* - * OK, we're invoking a handler: - */ -static int signr_convert(int sig) -{ -#ifdef CONFIG_X86_32 - struct thread_info *info = current_thread_info(); - - if (info->exec_domain && info->exec_domain->signal_invmap && sig < 32) - return info->exec_domain->signal_invmap[sig]; -#endif /* CONFIG_X86_32 */ - return sig; -} - -#ifdef CONFIG_X86_32 - -#define is_ia32 1 -#define ia32_setup_frame __setup_frame -#define ia32_setup_rt_frame __setup_rt_frame - -#else /* !CONFIG_X86_32 */ - -#ifdef CONFIG_IA32_EMULATION -#define is_ia32 test_thread_flag(TIF_IA32) -#else /* !CONFIG_IA32_EMULATION */ -#define is_ia32 0 -#endif /* CONFIG_IA32_EMULATION */ - -#endif /* CONFIG_X86_32 */ - -static int -setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, - sigset_t *set, struct pt_regs *regs) -{ - int usig = signr_convert(sig); - int ret; - - /* Set up the stack frame */ - if (is_ia32) { - if (ka->sa.sa_flags & SA_SIGINFO) - ret = ia32_setup_rt_frame(usig, ka, info, set, regs); - else - ret = ia32_setup_frame(usig, ka, set, regs); - } else - ret = __setup_rt_frame(sig, ka, info, set, regs); - - if (ret) { - force_sigsegv(sig, current); - return -EFAULT; - } - - return ret; -} - -static int -handle_signal(unsigned long sig, siginfo_t *info, struct k_sigaction *ka, - sigset_t *oldset, struct pt_regs *regs) -{ - int ret; - - /* Are we from a system call? */ - if (syscall_get_nr(current, regs) >= 0) { - /* If so, check system call restarting.. */ - switch (syscall_get_error(current, regs)) { - case -ERESTART_RESTARTBLOCK: - case -ERESTARTNOHAND: - regs->ax = -EINTR; - break; - - case -ERESTARTSYS: - if (!(ka->sa.sa_flags & SA_RESTART)) { - regs->ax = -EINTR; - break; - } - /* fallthrough */ - case -ERESTARTNOINTR: - regs->ax = regs->orig_ax; - regs->ip -= 2; - break; - } - } - - /* - * If TF is set due to a debugger (TIF_FORCED_TF), clear the TF - * flag so that register information in the sigcontext is correct. - */ - if (unlikely(regs->flags & X86_EFLAGS_TF) && - likely(test_and_clear_thread_flag(TIF_FORCED_TF))) - regs->flags &= ~X86_EFLAGS_TF; - - ret = setup_rt_frame(sig, ka, info, oldset, regs); - - if (ret) - return ret; - -#ifdef CONFIG_X86_64 - /* - * This has nothing to do with segment registers, - * despite the name. This magic affects uaccess.h - * macros' behavior. Reset it to the normal setting. - */ - set_fs(USER_DS); -#endif - - /* - * Clear the direction flag as per the ABI for function entry. - */ - regs->flags &= ~X86_EFLAGS_DF; - - /* - * Clear TF when entering the signal handler, but - * notify any tracer that was single-stepping it. - * The tracer may want to single-step inside the - * handler too. - */ - regs->flags &= ~X86_EFLAGS_TF; - - spin_lock_irq(¤t->sighand->siglock); - sigorsets(¤t->blocked, ¤t->blocked, &ka->sa.sa_mask); - if (!(ka->sa.sa_flags & SA_NODEFER)) - sigaddset(¤t->blocked, sig); - recalc_sigpending(); - spin_unlock_irq(¤t->sighand->siglock); - - tracehook_signal_handler(sig, info, ka, regs, - test_thread_flag(TIF_SINGLESTEP)); - - return 0; -} - -#ifdef CONFIG_X86_32 -#define NR_restart_syscall __NR_restart_syscall -#else /* !CONFIG_X86_32 */ -#define NR_restart_syscall \ - test_thread_flag(TIF_IA32) ? __NR_ia32_restart_syscall : __NR_restart_syscall -#endif /* CONFIG_X86_32 */ - -/* - * Note that 'init' is a special process: it doesn't get signals it doesn't - * want to handle. Thus you cannot kill init even with a SIGKILL even by - * mistake. - */ -static void do_signal(struct pt_regs *regs) -{ - struct k_sigaction ka; - siginfo_t info; - int signr; - sigset_t *oldset; - - /* - * We want the common case to go fast, which is why we may in certain - * cases get here from kernel mode. Just return without doing anything - * if so. - * X86_32: vm86 regs switched out by assembly code before reaching - * here, so testing against kernel CS suffices. - */ - if (!user_mode(regs)) - return; - - if (current_thread_info()->status & TS_RESTORE_SIGMASK) - oldset = ¤t->saved_sigmask; - else - oldset = ¤t->blocked; - - signr = get_signal_to_deliver(&info, &ka, regs, NULL); - if (signr > 0) { - /* - * Re-enable any watchpoints before delivering the - * signal to user space. The processor register will - * have been cleared if the watchpoint triggered - * inside the kernel. - */ - if (current->thread.debugreg7) - set_debugreg(current->thread.debugreg7, 7); - - /* Whee! Actually deliver the signal. */ - if (handle_signal(signr, &info, &ka, oldset, regs) == 0) { - /* - * A signal was successfully delivered; the saved - * sigmask will have been stored in the signal frame, - * and will be restored by sigreturn, so we can simply - * clear the TS_RESTORE_SIGMASK flag. - */ - current_thread_info()->status &= ~TS_RESTORE_SIGMASK; - } - return; - } - - /* Did we come from a system call? */ - if (syscall_get_nr(current, regs) >= 0) { - /* Restart the system call - no handlers present */ - switch (syscall_get_error(current, regs)) { - case -ERESTARTNOHAND: - case -ERESTARTSYS: - case -ERESTARTNOINTR: - regs->ax = regs->orig_ax; - regs->ip -= 2; - break; - - case -ERESTART_RESTARTBLOCK: - regs->ax = NR_restart_syscall; - regs->ip -= 2; - break; - } - } - - /* - * If there's no signal to deliver, we just put the saved sigmask - * back. - */ - if (current_thread_info()->status & TS_RESTORE_SIGMASK) { - current_thread_info()->status &= ~TS_RESTORE_SIGMASK; - sigprocmask(SIG_SETMASK, ¤t->saved_sigmask, NULL); - } -} - -/* - * notification of userspace execution resumption - * - triggered by the TIF_WORK_MASK flags - */ -void -do_notify_resume(struct pt_regs *regs, void *unused, __u32 thread_info_flags) -{ -#if defined(CONFIG_X86_64) && defined(CONFIG_X86_MCE) - /* notify userspace of pending MCEs */ - if (thread_info_flags & _TIF_MCE_NOTIFY) - mce_notify_user(); -#endif /* CONFIG_X86_64 && CONFIG_X86_MCE */ - - /* deal with pending signal delivery */ - if (thread_info_flags & _TIF_SIGPENDING) - do_signal(regs); - - if (thread_info_flags & _TIF_NOTIFY_RESUME) { - clear_thread_flag(TIF_NOTIFY_RESUME); - tracehook_notify_resume(regs); - } - -#ifdef CONFIG_X86_32 - clear_thread_flag(TIF_IRET); -#endif /* CONFIG_X86_32 */ -} - -void signal_fault(struct pt_regs *regs, void __user *frame, char *where) -{ - struct task_struct *me = current; - - if (show_unhandled_signals && printk_ratelimit()) { - printk(KERN_INFO - "%s[%d] bad frame in %s frame:%p ip:%lx sp:%lx orax:%lx", - me->comm, me->pid, where, frame, - regs->ip, regs->sp, regs->orig_ax); - print_vma_addr(" in ", regs->ip); - printk(KERN_CONT "\n"); - } - - force_sig(SIGSEGV, me); -} -- cgit v1.2.3 From 5a45cfe1c64862e8cd3b0d79d7c4ba71c3118915 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 26 Nov 2008 00:16:24 -0500 Subject: ftrace: use code patching for ftrace graph tracer Impact: more efficient code for ftrace graph tracer This patch uses the dynamic patching, when available, to patch the function graph code into the kernel. This patch will ease the way for letting both function tracing and function graph tracing run together. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_32.S | 5 +++++ arch/x86/kernel/ftrace.c | 48 ++++++++++++++++++++++++++++++++++++++++++++-- include/linux/ftrace.h | 5 +++++ kernel/trace/ftrace.c | 35 ++++++++++++++++----------------- 4 files changed, 72 insertions(+), 21 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S index 7def9fd5c1e6..958af86186c4 100644 --- a/arch/x86/kernel/entry_32.S +++ b/arch/x86/kernel/entry_32.S @@ -1174,6 +1174,11 @@ ftrace_call: popl %edx popl %ecx popl %eax +#ifdef CONFIG_FUNCTION_GRAPH_TRACER +.globl ftrace_graph_call +ftrace_graph_call: + jmp ftrace_stub +#endif .globl ftrace_stub ftrace_stub: diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 26b2d92d48b3..7ef914e6a2f6 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -111,7 +111,6 @@ static void ftrace_mod_code(void) */ mod_code_status = probe_kernel_write(mod_code_ip, mod_code_newcode, MCOUNT_INSN_SIZE); - } void ftrace_nmi_enter(void) @@ -325,7 +324,51 @@ int __init ftrace_dyn_arch_init(void *data) #ifdef CONFIG_FUNCTION_GRAPH_TRACER -#ifndef CONFIG_DYNAMIC_FTRACE +#ifdef CONFIG_DYNAMIC_FTRACE +extern void ftrace_graph_call(void); + +static int ftrace_mod_jmp(unsigned long ip, + int old_offset, int new_offset) +{ + unsigned char code[MCOUNT_INSN_SIZE]; + + if (probe_kernel_read(code, (void *)ip, MCOUNT_INSN_SIZE)) + return -EFAULT; + + if (code[0] != 0xe9 || old_offset != *(int *)(&code[1])) + return -EINVAL; + + *(int *)(&code[1]) = new_offset; + + if (do_ftrace_mod_code(ip, &code)) + return -EPERM; + + return 0; +} + +int ftrace_enable_ftrace_graph_caller(void) +{ + unsigned long ip = (unsigned long)(&ftrace_graph_call); + int old_offset, new_offset; + + old_offset = (unsigned long)(&ftrace_stub) - (ip + MCOUNT_INSN_SIZE); + new_offset = (unsigned long)(&ftrace_graph_caller) - (ip + MCOUNT_INSN_SIZE); + + return ftrace_mod_jmp(ip, old_offset, new_offset); +} + +int ftrace_disable_ftrace_graph_caller(void) +{ + unsigned long ip = (unsigned long)(&ftrace_graph_call); + int old_offset, new_offset; + + old_offset = (unsigned long)(&ftrace_graph_caller) - (ip + MCOUNT_INSN_SIZE); + new_offset = (unsigned long)(&ftrace_stub) - (ip + MCOUNT_INSN_SIZE); + + return ftrace_mod_jmp(ip, old_offset, new_offset); +} + +#else /* CONFIG_DYNAMIC_FTRACE */ /* * These functions are picked from those used on @@ -343,6 +386,7 @@ void ftrace_nmi_exit(void) { atomic_dec(&in_nmi); } + #endif /* !CONFIG_DYNAMIC_FTRACE */ /* Add a function return address to the trace stack on thread info.*/ diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index fc2d54987198..f9792c0d73f6 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -117,6 +117,11 @@ extern void ftrace_call(void); extern void mcount_call(void); #ifdef CONFIG_FUNCTION_GRAPH_TRACER extern void ftrace_graph_caller(void); +extern int ftrace_enable_ftrace_graph_caller(void); +extern int ftrace_disable_ftrace_graph_caller(void); +#else +static inline int ftrace_enable_ftrace_graph_caller(void) { return 0; } +static inline int ftrace_disable_ftrace_graph_caller(void) { return 0; } #endif /** diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 00d98c65fad0..5f7c8642d58b 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -281,6 +281,8 @@ enum { FTRACE_UPDATE_TRACE_FUNC = (1 << 2), FTRACE_ENABLE_MCOUNT = (1 << 3), FTRACE_DISABLE_MCOUNT = (1 << 4), + FTRACE_START_FUNC_RET = (1 << 5), + FTRACE_STOP_FUNC_RET = (1 << 6), }; static int ftrace_filtered; @@ -465,14 +467,7 @@ __ftrace_replace_code(struct dyn_ftrace *rec, int enable) unsigned long ip, fl; unsigned long ftrace_addr; -#ifdef CONFIG_FUNCTION_GRAPH_TRACER - if (ftrace_tracing_type == FTRACE_TYPE_ENTER) - ftrace_addr = (unsigned long)ftrace_caller; - else - ftrace_addr = (unsigned long)ftrace_graph_caller; -#else ftrace_addr = (unsigned long)ftrace_caller; -#endif ip = rec->ip; @@ -605,6 +600,11 @@ static int __ftrace_modify_code(void *data) if (*command & FTRACE_UPDATE_TRACE_FUNC) ftrace_update_ftrace_func(ftrace_trace_function); + if (*command & FTRACE_START_FUNC_RET) + ftrace_enable_ftrace_graph_caller(); + else if (*command & FTRACE_STOP_FUNC_RET) + ftrace_disable_ftrace_graph_caller(); + return 0; } @@ -629,10 +629,8 @@ static void ftrace_startup_enable(int command) ftrace_run_update_code(command); } -static void ftrace_startup(void) +static void ftrace_startup(int command) { - int command = 0; - if (unlikely(ftrace_disabled)) return; @@ -645,10 +643,8 @@ static void ftrace_startup(void) mutex_unlock(&ftrace_start_lock); } -static void ftrace_shutdown(void) +static void ftrace_shutdown(int command) { - int command = 0; - if (unlikely(ftrace_disabled)) return; @@ -1453,8 +1449,9 @@ device_initcall(ftrace_nodyn_init); static inline int ftrace_init_dyn_debugfs(struct dentry *d_tracer) { return 0; } static inline void ftrace_startup_enable(int command) { } -# define ftrace_startup() do { } while (0) -# define ftrace_shutdown() do { } while (0) +/* Keep as macros so we do not need to define the commands */ +# define ftrace_startup(command) do { } while (0) +# define ftrace_shutdown(command) do { } while (0) # define ftrace_startup_sysctl() do { } while (0) # define ftrace_shutdown_sysctl() do { } while (0) #endif /* CONFIG_DYNAMIC_FTRACE */ @@ -1585,7 +1582,7 @@ int register_ftrace_function(struct ftrace_ops *ops) } ret = __register_ftrace_function(ops); - ftrace_startup(); + ftrace_startup(0); out: mutex_unlock(&ftrace_sysctl_lock); @@ -1604,7 +1601,7 @@ int unregister_ftrace_function(struct ftrace_ops *ops) mutex_lock(&ftrace_sysctl_lock); ret = __unregister_ftrace_function(ops); - ftrace_shutdown(); + ftrace_shutdown(0); mutex_unlock(&ftrace_sysctl_lock); return ret; @@ -1751,7 +1748,7 @@ int register_ftrace_graph(trace_func_graph_ret_t retfunc, ftrace_tracing_type = FTRACE_TYPE_RETURN; ftrace_graph_return = retfunc; ftrace_graph_entry = entryfunc; - ftrace_startup(); + ftrace_startup(FTRACE_START_FUNC_RET); out: mutex_unlock(&ftrace_sysctl_lock); @@ -1765,7 +1762,7 @@ void unregister_ftrace_graph(void) atomic_dec(&ftrace_graph_active); ftrace_graph_return = (trace_func_graph_ret_t)ftrace_stub; ftrace_graph_entry = (trace_func_graph_ent_t)ftrace_stub; - ftrace_shutdown(); + ftrace_shutdown(FTRACE_STOP_FUNC_RET); /* Restore normal tracing type */ ftrace_tracing_type = FTRACE_TYPE_ENTER; -- cgit v1.2.3 From f3f47a6768a29448866da4422b6f6bee485c947f Mon Sep 17 00:00:00 2001 From: Arjan van de Ven Date: Sun, 23 Nov 2008 16:49:58 -0800 Subject: tracing: add "power-tracer": C/P state tracer to help power optimization Impact: new "power-tracer" ftrace plugin This patch adds a C/P-state ftrace plugin that will generate detailed statistics about the C/P-states that are being used, so that we can look at detailed decisions that the C/P-state code is making, rather than the too high level "average" that we have today. An example way of using this is: mount -t debugfs none /sys/kernel/debug echo cstate > /sys/kernel/debug/tracing/current_tracer echo 1 > /sys/kernel/debug/tracing/tracing_enabled sleep 1 echo 0 > /sys/kernel/debug/tracing/tracing_enabled cat /sys/kernel/debug/tracing/trace | perl scripts/trace/cstate.pl > out.svg Signed-off-by: Arjan van de Ven Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c | 4 + arch/x86/kernel/process.c | 16 +++ include/linux/ftrace.h | 29 +++++ kernel/trace/Kconfig | 11 ++ kernel/trace/Makefile | 1 + kernel/trace/trace.h | 7 ++ kernel/trace/trace_power.c | 179 +++++++++++++++++++++++++++++ scripts/trace/power.pl | 108 +++++++++++++++++ 8 files changed, 355 insertions(+) create mode 100644 kernel/trace/trace_power.c create mode 100644 scripts/trace/power.pl (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c index 8e48c5d4467d..88ea02dcb622 100644 --- a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c +++ b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c @@ -33,6 +33,7 @@ #include #include #include +#include #include #include @@ -391,6 +392,7 @@ static int acpi_cpufreq_target(struct cpufreq_policy *policy, unsigned int next_perf_state = 0; /* Index into perf table */ unsigned int i; int result = 0; + struct power_trace it; dprintk("acpi_cpufreq_target %d (%d)\n", target_freq, policy->cpu); @@ -427,6 +429,8 @@ static int acpi_cpufreq_target(struct cpufreq_policy *policy, } } + trace_power_mark(&it, POWER_PSTATE, next_perf_state); + switch (data->cpu_feature) { case SYSTEM_INTEL_MSR_CAPABLE: cmd.type = SYSTEM_INTEL_MSR_CAPABLE; diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c index c622772744d8..c27af49a4ede 100644 --- a/arch/x86/kernel/process.c +++ b/arch/x86/kernel/process.c @@ -7,6 +7,7 @@ #include #include #include +#include #include unsigned long idle_halt; @@ -100,6 +101,9 @@ static inline int hlt_use_halt(void) void default_idle(void) { if (hlt_use_halt()) { + struct power_trace it; + + trace_power_start(&it, POWER_CSTATE, 1); current_thread_info()->status &= ~TS_POLLING; /* * TS_POLLING-cleared state must be visible before we @@ -112,6 +116,7 @@ void default_idle(void) else local_irq_enable(); current_thread_info()->status |= TS_POLLING; + trace_power_end(&it); } else { local_irq_enable(); /* loop is done by the caller */ @@ -154,24 +159,31 @@ EXPORT_SYMBOL_GPL(cpu_idle_wait); */ void mwait_idle_with_hints(unsigned long ax, unsigned long cx) { + struct power_trace it; + + trace_power_start(&it, POWER_CSTATE, (ax>>4)+1); if (!need_resched()) { __monitor((void *)¤t_thread_info()->flags, 0, 0); smp_mb(); if (!need_resched()) __mwait(ax, cx); } + trace_power_end(&it); } /* Default MONITOR/MWAIT with no hints, used for default C1 state */ static void mwait_idle(void) { + struct power_trace it; if (!need_resched()) { + trace_power_start(&it, POWER_CSTATE, 1); __monitor((void *)¤t_thread_info()->flags, 0, 0); smp_mb(); if (!need_resched()) __sti_mwait(0, 0); else local_irq_enable(); + trace_power_end(&it); } else local_irq_enable(); } @@ -183,9 +195,13 @@ static void mwait_idle(void) */ static void poll_idle(void) { + struct power_trace it; + + trace_power_start(&it, POWER_CSTATE, 0); local_irq_enable(); while (!need_resched()) cpu_relax(); + trace_power_end(&it); } /* diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 7854d87b97b2..0df288666201 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -311,6 +311,35 @@ ftrace_init_module(struct module *mod, unsigned long *start, unsigned long *end) { } #endif +enum { + POWER_NONE = 0, + POWER_CSTATE = 1, + POWER_PSTATE = 2, +}; + +struct power_trace { +#ifdef CONFIG_POWER_TRACER + ktime_t stamp; + ktime_t end; + int type; + int state; +#endif +}; + +#ifdef CONFIG_POWER_TRACER +extern void trace_power_start(struct power_trace *it, unsigned int type, + unsigned int state); +extern void trace_power_mark(struct power_trace *it, unsigned int type, + unsigned int state); +extern void trace_power_end(struct power_trace *it); +#else +static inline void trace_power_start(struct power_trace *it, unsigned int type, + unsigned int state) { } +static inline void trace_power_mark(struct power_trace *it, unsigned int type, + unsigned int state) { } +static inline void trace_power_end(struct power_trace *it) { } +#endif + /* * Structure that defines a return function trace. diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 620feadff67a..d151aab48ed6 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -217,6 +217,17 @@ config BRANCH_TRACER Say N if unsure. +config POWER_TRACER + bool "Trace power consumption behavior" + depends on DEBUG_KERNEL + depends on X86 + select TRACING + help + This tracer helps developers to analyze and optimize the kernels + power management decisions, specifically the C-state and P-state + behavior. + + config STACK_TRACER bool "Trace max stack" depends on HAVE_FUNCTION_TRACER diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index cef4bcb4e822..acaa06553eca 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -32,5 +32,6 @@ obj-$(CONFIG_BOOT_TRACER) += trace_boot.o obj-$(CONFIG_FUNCTION_RET_TRACER) += trace_functions_return.o obj-$(CONFIG_TRACE_BRANCH_PROFILING) += trace_branch.o obj-$(CONFIG_BTS_TRACER) += trace_bts.o +obj-$(CONFIG_POWER_TRACER) += trace_power.o libftrace-y := ftrace.o diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 3abd645e8af2..4c453778a6ab 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -28,6 +28,7 @@ enum trace_type { TRACE_FN_RET, TRACE_USER_STACK, TRACE_BTS, + TRACE_POWER, __TRACE_LAST_TYPE }; @@ -160,6 +161,11 @@ struct bts_entry { unsigned long to; }; +struct trace_power { + struct trace_entry ent; + struct power_trace state_data; +}; + /* * trace_flag_type is an enumeration that holds different * states when a trace occurs. These are: @@ -266,6 +272,7 @@ extern void __ftrace_bad_type(void); IF_ASSIGN(var, ent, struct trace_branch, TRACE_BRANCH); \ IF_ASSIGN(var, ent, struct ftrace_ret_entry, TRACE_FN_RET);\ IF_ASSIGN(var, ent, struct bts_entry, TRACE_BTS);\ + IF_ASSIGN(var, ent, struct trace_power, TRACE_POWER); \ __ftrace_bad_type(); \ } while (0) diff --git a/kernel/trace/trace_power.c b/kernel/trace/trace_power.c new file mode 100644 index 000000000000..a7172a352f62 --- /dev/null +++ b/kernel/trace/trace_power.c @@ -0,0 +1,179 @@ +/* + * ring buffer based C-state tracer + * + * Arjan van de Ven + * Copyright (C) 2008 Intel Corporation + * + * Much is borrowed from trace_boot.c which is + * Copyright (C) 2008 Frederic Weisbecker + * + */ + +#include +#include +#include +#include +#include + +#include "trace.h" + +static struct trace_array *power_trace; +static int __read_mostly trace_power_enabled; + + +static void start_power_trace(struct trace_array *tr) +{ + trace_power_enabled = 1; +} + +static void stop_power_trace(struct trace_array *tr) +{ + trace_power_enabled = 0; +} + + +static int power_trace_init(struct trace_array *tr) +{ + int cpu; + power_trace = tr; + + trace_power_enabled = 1; + + for_each_cpu_mask(cpu, cpu_possible_map) + tracing_reset(tr, cpu); + return 0; +} + +static enum print_line_t power_print_line(struct trace_iterator *iter) +{ + int ret = 0; + struct trace_entry *entry = iter->ent; + struct trace_power *field ; + struct power_trace *it; + struct trace_seq *s = &iter->seq; + struct timespec stamp; + struct timespec duration; + + trace_assign_type(field, entry); + it = &field->state_data; + stamp = ktime_to_timespec(it->stamp); + duration = ktime_to_timespec(ktime_sub(it->end, it->stamp)); + + if (entry->type == TRACE_POWER) { + if (it->type == POWER_CSTATE) + ret = trace_seq_printf(s, "[%5ld.%09ld] CSTATE: Going to C%i on cpu %i for %ld.%09ld\n", + stamp.tv_sec, + stamp.tv_nsec, + it->state, iter->cpu, + duration.tv_sec, + duration.tv_nsec); + if (it->type == POWER_PSTATE) + ret = trace_seq_printf(s, "[%5ld.%09ld] PSTATE: Going to P%i on cpu %i\n", + stamp.tv_sec, + stamp.tv_nsec, + it->state, iter->cpu); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + return TRACE_TYPE_HANDLED; + } + return TRACE_TYPE_UNHANDLED; +} + +static struct tracer power_tracer __read_mostly = +{ + .name = "power", + .init = power_trace_init, + .start = start_power_trace, + .stop = stop_power_trace, + .reset = stop_power_trace, + .print_line = power_print_line, +}; + +static int init_power_trace(void) +{ + return register_tracer(&power_tracer); +} +device_initcall(init_power_trace); + +void trace_power_start(struct power_trace *it, unsigned int type, + unsigned int level) +{ + if (!trace_power_enabled) + return; + + memset(it, 0, sizeof(struct power_trace)); + it->state = level; + it->type = type; + it->stamp = ktime_get(); +} +EXPORT_SYMBOL_GPL(trace_power_start); + + +void trace_power_end(struct power_trace *it) +{ + struct ring_buffer_event *event; + struct trace_power *entry; + struct trace_array_cpu *data; + unsigned long irq_flags; + struct trace_array *tr = power_trace; + + if (!trace_power_enabled) + return; + + preempt_disable(); + it->end = ktime_get(); + data = tr->data[smp_processor_id()]; + + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), + &irq_flags); + if (!event) + goto out; + entry = ring_buffer_event_data(event); + tracing_generic_entry_update(&entry->ent, 0, 0); + entry->ent.type = TRACE_POWER; + entry->state_data = *it; + ring_buffer_unlock_commit(tr->buffer, event, irq_flags); + + trace_wake_up(); + + out: + preempt_enable(); +} +EXPORT_SYMBOL_GPL(trace_power_end); + +void trace_power_mark(struct power_trace *it, unsigned int type, + unsigned int level) +{ + struct ring_buffer_event *event; + struct trace_power *entry; + struct trace_array_cpu *data; + unsigned long irq_flags; + struct trace_array *tr = power_trace; + + if (!trace_power_enabled) + return; + + memset(it, 0, sizeof(struct power_trace)); + it->state = level; + it->type = type; + it->stamp = ktime_get(); + preempt_disable(); + it->end = it->stamp; + data = tr->data[smp_processor_id()]; + + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), + &irq_flags); + if (!event) + goto out; + entry = ring_buffer_event_data(event); + tracing_generic_entry_update(&entry->ent, 0, 0); + entry->ent.type = TRACE_POWER; + entry->state_data = *it; + ring_buffer_unlock_commit(tr->buffer, event, irq_flags); + + trace_wake_up(); + + out: + preempt_enable(); +} +EXPORT_SYMBOL_GPL(trace_power_mark); diff --git a/scripts/trace/power.pl b/scripts/trace/power.pl new file mode 100644 index 000000000000..4f729b3501e0 --- /dev/null +++ b/scripts/trace/power.pl @@ -0,0 +1,108 @@ +#!/usr/bin/perl + +# Copyright 2008, Intel Corporation +# +# This file is part of the Linux kernel +# +# This program file is free software; you can redistribute it and/or modify it +# under the terms of the GNU General Public License as published by the +# Free Software Foundation; version 2 of the License. +# +# This program is distributed in the hope that it will be useful, but WITHOUT +# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or +# FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License +# for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program in a file named COPYING; if not, write to the +# Free Software Foundation, Inc., +# 51 Franklin Street, Fifth Floor, +# Boston, MA 02110-1301 USA +# +# Authors: +# Arjan van de Ven + + +# +# This script turns a cstate ftrace output into a SVG graphic that shows +# historic C-state information +# +# +# cat /sys/kernel/debug/tracing/trace | perl power.pl > out.svg +# + +my @styles; +my $base = 0; + +my @pstate_last; +my @pstate_level; + +$styles[0] = "fill:rgb(0,0,255);fill-opacity:0.5;stroke-width:1;stroke:rgb(0,0,0)"; +$styles[1] = "fill:rgb(0,255,0);fill-opacity:0.5;stroke-width:1;stroke:rgb(0,0,0)"; +$styles[2] = "fill:rgb(255,0,20);fill-opacity:0.5;stroke-width:1;stroke:rgb(0,0,0)"; +$styles[3] = "fill:rgb(255,255,20);fill-opacity:0.5;stroke-width:1;stroke:rgb(0,0,0)"; +$styles[4] = "fill:rgb(255,0,255);fill-opacity:0.5;stroke-width:1;stroke:rgb(0,0,0)"; +$styles[5] = "fill:rgb(0,255,255);fill-opacity:0.5;stroke-width:1;stroke:rgb(0,0,0)"; +$styles[6] = "fill:rgb(0,128,255);fill-opacity:0.5;stroke-width:1;stroke:rgb(0,0,0)"; +$styles[7] = "fill:rgb(0,255,128);fill-opacity:0.5;stroke-width:1;stroke:rgb(0,0,0)"; +$styles[8] = "fill:rgb(0,25,20);fill-opacity:0.5;stroke-width:1;stroke:rgb(0,0,0)"; + + +print " \n"; +print "\n"; + +my $scale = 30000.0; +while (<>) { + my $line = $_; + if ($line =~ /([0-9\.]+)\] CSTATE: Going to C([0-9]) on cpu ([0-9]+) for ([0-9\.]+)/) { + if ($base == 0) { + $base = $1; + } + my $time = $1 - $base; + $time = $time * $scale; + my $C = $2; + my $cpu = $3; + my $y = 400 * $cpu; + my $duration = $4 * $scale; + my $msec = int($4 * 100000)/100.0; + my $height = $C * 20; + $style = $styles[$C]; + + $y = $y + 140 - $height; + + $x2 = $time + 4; + $y2 = $y + 4; + + + print "\n"; + print "C$C $msec\n"; + } + if ($line =~ /([0-9\.]+)\] PSTATE: Going to P([0-9]) on cpu ([0-9]+)/) { + my $time = $1 - $base; + my $state = $2; + my $cpu = $3; + + if (defined($pstate_last[$cpu])) { + my $from = $pstate_last[$cpu]; + my $oldstate = $pstate_state[$cpu]; + my $duration = ($time-$from) * $scale; + + $from = $from * $scale; + my $to = $from + $duration; + my $height = 140 - ($oldstate * (140/8)); + + my $y = 400 * $cpu + 200 + $height; + my $y2 = $y+4; + my $style = $styles[8]; + + print "\n"; + print "P$oldstate (cpu $cpu)\n"; + }; + + $pstate_last[$cpu] = $time; + $pstate_state[$cpu] = $state; + } +} + + +print "\n"; -- cgit v1.2.3 From 4db646b1af8fdcf01d690d29eeae44cd937edb0d Mon Sep 17 00:00:00 2001 From: Hannes Eder Date: Sun, 23 Nov 2008 20:49:52 +0100 Subject: x86: microcode: fix sparse warnings Impact: make global variables and a function static Fix following sparse warnings: arch/x86/kernel/microcode_core.c:102:22: warning: symbol 'microcode_ops' was not declared. Should it be static? arch/x86/kernel/microcode_core.c:206:24: warning: symbol 'microcode_pdev' was not declared. Should it be static? arch/x86/kernel/microcode_core.c:322:6: warning: symbol 'microcode_update_cpu' was not declared. Should it be static? arch/x86/kernel/microcode_intel.c:468:22: warning: symbol 'microcode_intel_ops' was not declared. Should it be static? Signed-off-by: Hannes Eder Signed-off-by: Ingo Molnar --- arch/x86/kernel/microcode_core.c | 6 +++--- arch/x86/kernel/microcode_intel.c | 2 +- 2 files changed, 4 insertions(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/microcode_core.c b/arch/x86/kernel/microcode_core.c index 82fb2809ce32..5b711a534495 100644 --- a/arch/x86/kernel/microcode_core.c +++ b/arch/x86/kernel/microcode_core.c @@ -99,7 +99,7 @@ MODULE_LICENSE("GPL"); #define MICROCODE_VERSION "2.00" -struct microcode_ops *microcode_ops; +static struct microcode_ops *microcode_ops; /* no concurrent ->write()s are allowed on /dev/cpu/microcode */ static DEFINE_MUTEX(microcode_mutex); @@ -203,7 +203,7 @@ MODULE_ALIAS_MISCDEV(MICROCODE_MINOR); #endif /* fake device for request_firmware */ -struct platform_device *microcode_pdev; +static struct platform_device *microcode_pdev; static ssize_t reload_store(struct sys_device *dev, struct sysdev_attribute *attr, @@ -319,7 +319,7 @@ static int microcode_resume_cpu(int cpu) return 0; } -void microcode_update_cpu(int cpu) +static void microcode_update_cpu(int cpu) { struct ucode_cpu_info *uci = ucode_cpu_info + cpu; int err = 0; diff --git a/arch/x86/kernel/microcode_intel.c b/arch/x86/kernel/microcode_intel.c index 622dc4a21784..c34c820ee486 100644 --- a/arch/x86/kernel/microcode_intel.c +++ b/arch/x86/kernel/microcode_intel.c @@ -465,7 +465,7 @@ static void microcode_fini_cpu(int cpu) uci->mc = NULL; } -struct microcode_ops microcode_intel_ops = { +static struct microcode_ops microcode_intel_ops = { .request_microcode_user = request_microcode_user, .request_microcode_fw = request_microcode_fw, .collect_cpu_info = collect_cpu_info, -- cgit v1.2.3 From ddeb8f2149de280d54f0c8910cead42e6042b2cb Mon Sep 17 00:00:00 2001 From: Alexander van Heukelum Date: Mon, 24 Nov 2008 13:24:28 +0100 Subject: x86_64: get rid of the use of KPROBE_ENTRY / KPROBE_END Impact: clean up assembly macros and annotations - with some object impact entry_64.S is the only user of KPROBE_ENTRY / KPROBE_END on x86_64. This patch reorders entry_64.S and explicitly generates a separate section for functions that need the protection. The generated code before and after the patch is equal. Implicitly changing sections in assembly files makes it more difficult to follow why the assembler is doing certain things. For example, .p2align 5 KPROBE_ENTRY(...) was not doing what you would expect. Other section changes (__ex_table, .fixup, .init.rodata) are done explicitly already. Signed-off-by: Alexander van Heukelum Acked-by: Jan Beulich Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_64.S | 444 ++++++++++++++++++++++----------------------- 1 file changed, 220 insertions(+), 224 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index f2d546e16354..38fcd0517c31 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -1002,7 +1002,7 @@ END(\sym) .endm .macro paranoidzeroentry sym do_sym -KPROBE_ENTRY(\sym) +ENTRY(\sym) INTR_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME pushq $-1 /* ORIG_RAX: no syscall to restart */ @@ -1015,11 +1015,11 @@ KPROBE_ENTRY(\sym) call \do_sym jmp paranoid_exit /* %ebx: no swapgs flag */ CFI_ENDPROC -KPROBE_END(\sym) +END(\sym) .endm .macro paranoidzeroentry_ist sym do_sym ist -KPROBE_ENTRY(\sym) +ENTRY(\sym) INTR_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME pushq $-1 /* ORIG_RAX: no syscall to restart */ @@ -1035,15 +1035,11 @@ KPROBE_ENTRY(\sym) addq $EXCEPTION_STKSZ, per_cpu__init_tss + TSS_ist + (\ist - 1) * 8(%rbp) jmp paranoid_exit /* %ebx: no swapgs flag */ CFI_ENDPROC -KPROBE_END(\sym) +END(\sym) .endm -.macro errorentry sym do_sym entry=0 -.if \entry -KPROBE_ENTRY(\sym) -.else +.macro errorentry sym do_sym ENTRY(\sym) -.endif XCPT_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME subq $15*8,%rsp @@ -1056,20 +1052,12 @@ ENTRY(\sym) call \do_sym jmp error_exit /* %ebx: no swapgs flag */ CFI_ENDPROC -.if \entry -KPROBE_END(\sym) -.else END(\sym) -.endif .endm /* error code is on the stack already */ -.macro paranoiderrorentry sym do_sym entry=1 -.if \entry -KPROBE_ENTRY(\sym) -.else +.macro paranoiderrorentry sym do_sym ENTRY(\sym) -.endif XCPT_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME subq $15*8,%rsp @@ -1083,166 +1071,23 @@ ENTRY(\sym) call \do_sym jmp paranoid_exit /* %ebx: no swapgs flag */ CFI_ENDPROC -.if \entry -KPROBE_END(\sym) -.else END(\sym) -.endif .endm zeroentry divide_error do_divide_error -paranoidzeroentry_ist debug do_debug DEBUG_STACK -paranoidzeroentry_ist int3 do_int3 DEBUG_STACK zeroentry overflow do_overflow zeroentry bounds do_bounds zeroentry invalid_op do_invalid_op zeroentry device_not_available do_device_not_available -paranoiderrorentry double_fault do_double_fault 0 +paranoiderrorentry double_fault do_double_fault zeroentry coprocessor_segment_overrun do_coprocessor_segment_overrun errorentry invalid_TSS do_invalid_TSS errorentry segment_not_present do_segment_not_present -paranoiderrorentry stack_segment do_stack_segment -errorentry general_protection do_general_protection 1 -errorentry page_fault do_page_fault 1 zeroentry spurious_interrupt_bug do_spurious_interrupt_bug zeroentry coprocessor_error do_coprocessor_error errorentry alignment_check do_alignment_check -#ifdef CONFIG_X86_MCE -paranoidzeroentry machine_check do_machine_check -#endif zeroentry simd_coprocessor_error do_simd_coprocessor_error - /* - * "Paranoid" exit path from exception stack. - * Paranoid because this is used by NMIs and cannot take - * any kernel state for granted. - * We don't do kernel preemption checks here, because only - * NMI should be common and it does not enable IRQs and - * cannot get reschedule ticks. - * - * "trace" is 0 for the NMI handler only, because irq-tracing - * is fundamentally NMI-unsafe. (we cannot change the soft and - * hard flags at once, atomically) - */ - - /* ebx: no swapgs flag */ -KPROBE_ENTRY(paranoid_exit) - INTR_FRAME - DISABLE_INTERRUPTS(CLBR_NONE) - TRACE_IRQS_OFF - testl %ebx,%ebx /* swapgs needed? */ - jnz paranoid_restore - testl $3,CS(%rsp) - jnz paranoid_userspace -paranoid_swapgs: - TRACE_IRQS_IRETQ 0 - SWAPGS_UNSAFE_STACK -paranoid_restore: - RESTORE_ALL 8 - jmp irq_return -paranoid_userspace: - GET_THREAD_INFO(%rcx) - movl TI_flags(%rcx),%ebx - andl $_TIF_WORK_MASK,%ebx - jz paranoid_swapgs - movq %rsp,%rdi /* &pt_regs */ - call sync_regs - movq %rax,%rsp /* switch stack for scheduling */ - testl $_TIF_NEED_RESCHED,%ebx - jnz paranoid_schedule - movl %ebx,%edx /* arg3: thread flags */ - TRACE_IRQS_ON - ENABLE_INTERRUPTS(CLBR_NONE) - xorl %esi,%esi /* arg2: oldset */ - movq %rsp,%rdi /* arg1: &pt_regs */ - call do_notify_resume - DISABLE_INTERRUPTS(CLBR_NONE) - TRACE_IRQS_OFF - jmp paranoid_userspace -paranoid_schedule: - TRACE_IRQS_ON - ENABLE_INTERRUPTS(CLBR_ANY) - call schedule - DISABLE_INTERRUPTS(CLBR_ANY) - TRACE_IRQS_OFF - jmp paranoid_userspace - CFI_ENDPROC -KPROBE_END(paranoid_exit) - -/* - * Exception entry point. This expects an error code/orig_rax on the stack. - * returns in "no swapgs flag" in %ebx. - */ -KPROBE_ENTRY(error_entry) - XCPT_FRAME - CFI_ADJUST_CFA_OFFSET 15*8 - /* oldrax contains error code */ - cld - movq_cfi rdi, RDI+8 - movq_cfi rsi, RSI+8 - movq_cfi rdx, RDX+8 - movq_cfi rcx, RCX+8 - movq_cfi rax, RAX+8 - movq_cfi r8, R8+8 - movq_cfi r9, R9+8 - movq_cfi r10, R10+8 - movq_cfi r11, R11+8 - movq_cfi rbx, RBX+8 - movq_cfi rbp, RBP+8 - movq_cfi r12, R12+8 - movq_cfi r13, R13+8 - movq_cfi r14, R14+8 - movq_cfi r15, R15+8 - xorl %ebx,%ebx - testl $3,CS+8(%rsp) - je error_kernelspace -error_swapgs: - SWAPGS -error_sti: - TRACE_IRQS_OFF - ret - CFI_ENDPROC - -/* - * There are two places in the kernel that can potentially fault with - * usergs. Handle them here. The exception handlers after iret run with - * kernel gs again, so don't set the user space flag. B stepping K8s - * sometimes report an truncated RIP for IRET exceptions returning to - * compat mode. Check for these here too. - */ -error_kernelspace: - incl %ebx - leaq irq_return(%rip),%rcx - cmpq %rcx,RIP+8(%rsp) - je error_swapgs - movl %ecx,%ecx /* zero extend */ - cmpq %rcx,RIP+8(%rsp) - je error_swapgs - cmpq $gs_change,RIP+8(%rsp) - je error_swapgs - jmp error_sti -KPROBE_END(error_entry) - - -/* ebx: no swapgs flag (1: don't need swapgs, 0: need it) */ -KPROBE_ENTRY(error_exit) - DEFAULT_FRAME - movl %ebx,%eax - RESTORE_REST - DISABLE_INTERRUPTS(CLBR_NONE) - TRACE_IRQS_OFF - GET_THREAD_INFO(%rcx) - testl %eax,%eax - jne retint_kernel - LOCKDEP_SYS_EXIT_IRQ - movl TI_flags(%rcx),%edx - movl $_TIF_WORK_MASK,%edi - andl %edi,%edx - jnz retint_careful - jmp retint_swapgs - CFI_ENDPROC -KPROBE_END(error_exit) - /* Reload gs selector with exception handling */ /* edi: new selector */ ENTRY(native_load_gs_index) @@ -1362,61 +1207,6 @@ ENTRY(kernel_execve) CFI_ENDPROC END(kernel_execve) - /* runs on exception stack */ -KPROBE_ENTRY(nmi) - INTR_FRAME - PARAVIRT_ADJUST_EXCEPTION_FRAME - pushq_cfi $-1 - subq $15*8, %rsp - CFI_ADJUST_CFA_OFFSET 15*8 - call save_paranoid - DEFAULT_FRAME 0 - /* paranoidentry do_nmi, 0; without TRACE_IRQS_OFF */ - movq %rsp,%rdi - movq $-1,%rsi - call do_nmi -#ifdef CONFIG_TRACE_IRQFLAGS - /* paranoidexit; without TRACE_IRQS_OFF */ - /* ebx: no swapgs flag */ - DISABLE_INTERRUPTS(CLBR_NONE) - testl %ebx,%ebx /* swapgs needed? */ - jnz nmi_restore - testl $3,CS(%rsp) - jnz nmi_userspace -nmi_swapgs: - SWAPGS_UNSAFE_STACK -nmi_restore: - RESTORE_ALL 8 - jmp irq_return -nmi_userspace: - GET_THREAD_INFO(%rcx) - movl TI_flags(%rcx),%ebx - andl $_TIF_WORK_MASK,%ebx - jz nmi_swapgs - movq %rsp,%rdi /* &pt_regs */ - call sync_regs - movq %rax,%rsp /* switch stack for scheduling */ - testl $_TIF_NEED_RESCHED,%ebx - jnz nmi_schedule - movl %ebx,%edx /* arg3: thread flags */ - ENABLE_INTERRUPTS(CLBR_NONE) - xorl %esi,%esi /* arg2: oldset */ - movq %rsp,%rdi /* arg1: &pt_regs */ - call do_notify_resume - DISABLE_INTERRUPTS(CLBR_NONE) - jmp nmi_userspace -nmi_schedule: - ENABLE_INTERRUPTS(CLBR_ANY) - call schedule - DISABLE_INTERRUPTS(CLBR_ANY) - jmp nmi_userspace - CFI_ENDPROC -#else - jmp paranoid_exit - CFI_ENDPROC -#endif -KPROBE_END(nmi) - /* Call softirq on interrupt stack. Interrupts are off. */ ENTRY(call_softirq) CFI_STARTPROC @@ -1437,13 +1227,6 @@ ENTRY(call_softirq) CFI_ENDPROC END(call_softirq) -KPROBE_ENTRY(ignore_sysret) - CFI_STARTPROC - mov $-ENOSYS,%eax - sysret - CFI_ENDPROC -KPROBE_END(ignore_sysret) - #ifdef CONFIG_XEN zeroentry xen_hypervisor_callback xen_do_hypervisor_callback @@ -1540,3 +1323,216 @@ ENTRY(xen_failsafe_callback) END(xen_failsafe_callback) #endif /* CONFIG_XEN */ + +/* + * Some functions should be protected against kprobes + */ + .pushsection .kprobes.text, "ax" + +paranoidzeroentry_ist debug do_debug DEBUG_STACK +paranoidzeroentry_ist int3 do_int3 DEBUG_STACK +paranoiderrorentry stack_segment do_stack_segment +errorentry general_protection do_general_protection +errorentry page_fault do_page_fault +#ifdef CONFIG_X86_MCE +paranoidzeroentry machine_check do_machine_check +#endif + + /* + * "Paranoid" exit path from exception stack. + * Paranoid because this is used by NMIs and cannot take + * any kernel state for granted. + * We don't do kernel preemption checks here, because only + * NMI should be common and it does not enable IRQs and + * cannot get reschedule ticks. + * + * "trace" is 0 for the NMI handler only, because irq-tracing + * is fundamentally NMI-unsafe. (we cannot change the soft and + * hard flags at once, atomically) + */ + + /* ebx: no swapgs flag */ +ENTRY(paranoid_exit) + INTR_FRAME + DISABLE_INTERRUPTS(CLBR_NONE) + TRACE_IRQS_OFF + testl %ebx,%ebx /* swapgs needed? */ + jnz paranoid_restore + testl $3,CS(%rsp) + jnz paranoid_userspace +paranoid_swapgs: + TRACE_IRQS_IRETQ 0 + SWAPGS_UNSAFE_STACK +paranoid_restore: + RESTORE_ALL 8 + jmp irq_return +paranoid_userspace: + GET_THREAD_INFO(%rcx) + movl TI_flags(%rcx),%ebx + andl $_TIF_WORK_MASK,%ebx + jz paranoid_swapgs + movq %rsp,%rdi /* &pt_regs */ + call sync_regs + movq %rax,%rsp /* switch stack for scheduling */ + testl $_TIF_NEED_RESCHED,%ebx + jnz paranoid_schedule + movl %ebx,%edx /* arg3: thread flags */ + TRACE_IRQS_ON + ENABLE_INTERRUPTS(CLBR_NONE) + xorl %esi,%esi /* arg2: oldset */ + movq %rsp,%rdi /* arg1: &pt_regs */ + call do_notify_resume + DISABLE_INTERRUPTS(CLBR_NONE) + TRACE_IRQS_OFF + jmp paranoid_userspace +paranoid_schedule: + TRACE_IRQS_ON + ENABLE_INTERRUPTS(CLBR_ANY) + call schedule + DISABLE_INTERRUPTS(CLBR_ANY) + TRACE_IRQS_OFF + jmp paranoid_userspace + CFI_ENDPROC +END(paranoid_exit) + +/* + * Exception entry point. This expects an error code/orig_rax on the stack. + * returns in "no swapgs flag" in %ebx. + */ +ENTRY(error_entry) + XCPT_FRAME + CFI_ADJUST_CFA_OFFSET 15*8 + /* oldrax contains error code */ + cld + movq_cfi rdi, RDI+8 + movq_cfi rsi, RSI+8 + movq_cfi rdx, RDX+8 + movq_cfi rcx, RCX+8 + movq_cfi rax, RAX+8 + movq_cfi r8, R8+8 + movq_cfi r9, R9+8 + movq_cfi r10, R10+8 + movq_cfi r11, R11+8 + movq_cfi rbx, RBX+8 + movq_cfi rbp, RBP+8 + movq_cfi r12, R12+8 + movq_cfi r13, R13+8 + movq_cfi r14, R14+8 + movq_cfi r15, R15+8 + xorl %ebx,%ebx + testl $3,CS+8(%rsp) + je error_kernelspace +error_swapgs: + SWAPGS +error_sti: + TRACE_IRQS_OFF + ret + CFI_ENDPROC + +/* + * There are two places in the kernel that can potentially fault with + * usergs. Handle them here. The exception handlers after iret run with + * kernel gs again, so don't set the user space flag. B stepping K8s + * sometimes report an truncated RIP for IRET exceptions returning to + * compat mode. Check for these here too. + */ +error_kernelspace: + incl %ebx + leaq irq_return(%rip),%rcx + cmpq %rcx,RIP+8(%rsp) + je error_swapgs + movl %ecx,%ecx /* zero extend */ + cmpq %rcx,RIP+8(%rsp) + je error_swapgs + cmpq $gs_change,RIP+8(%rsp) + je error_swapgs + jmp error_sti +END(error_entry) + + +/* ebx: no swapgs flag (1: don't need swapgs, 0: need it) */ +ENTRY(error_exit) + DEFAULT_FRAME + movl %ebx,%eax + RESTORE_REST + DISABLE_INTERRUPTS(CLBR_NONE) + TRACE_IRQS_OFF + GET_THREAD_INFO(%rcx) + testl %eax,%eax + jne retint_kernel + LOCKDEP_SYS_EXIT_IRQ + movl TI_flags(%rcx),%edx + movl $_TIF_WORK_MASK,%edi + andl %edi,%edx + jnz retint_careful + jmp retint_swapgs + CFI_ENDPROC +END(error_exit) + + + /* runs on exception stack */ +ENTRY(nmi) + INTR_FRAME + PARAVIRT_ADJUST_EXCEPTION_FRAME + pushq_cfi $-1 + subq $15*8, %rsp + CFI_ADJUST_CFA_OFFSET 15*8 + call save_paranoid + DEFAULT_FRAME 0 + /* paranoidentry do_nmi, 0; without TRACE_IRQS_OFF */ + movq %rsp,%rdi + movq $-1,%rsi + call do_nmi +#ifdef CONFIG_TRACE_IRQFLAGS + /* paranoidexit; without TRACE_IRQS_OFF */ + /* ebx: no swapgs flag */ + DISABLE_INTERRUPTS(CLBR_NONE) + testl %ebx,%ebx /* swapgs needed? */ + jnz nmi_restore + testl $3,CS(%rsp) + jnz nmi_userspace +nmi_swapgs: + SWAPGS_UNSAFE_STACK +nmi_restore: + RESTORE_ALL 8 + jmp irq_return +nmi_userspace: + GET_THREAD_INFO(%rcx) + movl TI_flags(%rcx),%ebx + andl $_TIF_WORK_MASK,%ebx + jz nmi_swapgs + movq %rsp,%rdi /* &pt_regs */ + call sync_regs + movq %rax,%rsp /* switch stack for scheduling */ + testl $_TIF_NEED_RESCHED,%ebx + jnz nmi_schedule + movl %ebx,%edx /* arg3: thread flags */ + ENABLE_INTERRUPTS(CLBR_NONE) + xorl %esi,%esi /* arg2: oldset */ + movq %rsp,%rdi /* arg1: &pt_regs */ + call do_notify_resume + DISABLE_INTERRUPTS(CLBR_NONE) + jmp nmi_userspace +nmi_schedule: + ENABLE_INTERRUPTS(CLBR_ANY) + call schedule + DISABLE_INTERRUPTS(CLBR_ANY) + jmp nmi_userspace + CFI_ENDPROC +#else + jmp paranoid_exit + CFI_ENDPROC +#endif +END(nmi) + +ENTRY(ignore_sysret) + CFI_STARTPROC + mov $-ENOSYS,%eax + sysret + CFI_ENDPROC +END(ignore_sysret) + +/* + * End of kprobes section + */ + .popsection -- cgit v1.2.3 From d211af055d0c12dc3416c2886e6fbdc6eb74a381 Mon Sep 17 00:00:00 2001 From: Alexander van Heukelum Date: Mon, 24 Nov 2008 15:38:45 +0100 Subject: i386: get rid of the use of KPROBE_ENTRY / KPROBE_END entry_32.S is now the only user of KPROBE_ENTRY / KPROBE_END, treewide. This patch reorders entry_64.S and explicitly generates a separate section for functions that need the protection. The generated code before and after the patch is equal. The KPROBE_ENTRY and KPROBE_END macro's are removed too. Signed-off-by: Alexander van Heukelum Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_32.S | 438 +++++++++++++++++++++++---------------------- include/linux/linkage.h | 8 - 2 files changed, 224 insertions(+), 222 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S index bd02ec77edc4..6e96028d1a9c 100644 --- a/arch/x86/kernel/entry_32.S +++ b/arch/x86/kernel/entry_32.S @@ -688,65 +688,6 @@ ENDPROC(name) /* The include is where all of the SMP etc. interrupts come from */ #include "entry_arch.h" -KPROBE_ENTRY(page_fault) - RING0_EC_FRAME - pushl $do_page_fault - CFI_ADJUST_CFA_OFFSET 4 - ALIGN -error_code: - /* the function address is in %fs's slot on the stack */ - pushl %es - CFI_ADJUST_CFA_OFFSET 4 - /*CFI_REL_OFFSET es, 0*/ - pushl %ds - CFI_ADJUST_CFA_OFFSET 4 - /*CFI_REL_OFFSET ds, 0*/ - pushl %eax - CFI_ADJUST_CFA_OFFSET 4 - CFI_REL_OFFSET eax, 0 - pushl %ebp - CFI_ADJUST_CFA_OFFSET 4 - CFI_REL_OFFSET ebp, 0 - pushl %edi - CFI_ADJUST_CFA_OFFSET 4 - CFI_REL_OFFSET edi, 0 - pushl %esi - CFI_ADJUST_CFA_OFFSET 4 - CFI_REL_OFFSET esi, 0 - pushl %edx - CFI_ADJUST_CFA_OFFSET 4 - CFI_REL_OFFSET edx, 0 - pushl %ecx - CFI_ADJUST_CFA_OFFSET 4 - CFI_REL_OFFSET ecx, 0 - pushl %ebx - CFI_ADJUST_CFA_OFFSET 4 - CFI_REL_OFFSET ebx, 0 - cld - pushl %fs - CFI_ADJUST_CFA_OFFSET 4 - /*CFI_REL_OFFSET fs, 0*/ - movl $(__KERNEL_PERCPU), %ecx - movl %ecx, %fs - UNWIND_ESPFIX_STACK - popl %ecx - CFI_ADJUST_CFA_OFFSET -4 - /*CFI_REGISTER es, ecx*/ - movl PT_FS(%esp), %edi # get the function address - movl PT_ORIG_EAX(%esp), %edx # get the error code - movl $-1, PT_ORIG_EAX(%esp) # no syscall to restart - mov %ecx, PT_FS(%esp) - /*CFI_REL_OFFSET fs, ES*/ - movl $(__USER_DS), %ecx - movl %ecx, %ds - movl %ecx, %es - TRACE_IRQS_OFF - movl %esp,%eax # pt_regs pointer - call *%edi - jmp ret_from_exception - CFI_ENDPROC -KPROBE_END(page_fault) - ENTRY(coprocessor_error) RING0_INT_FRAME pushl $0 @@ -777,140 +718,6 @@ ENTRY(device_not_available) CFI_ENDPROC END(device_not_available) -/* - * Debug traps and NMI can happen at the one SYSENTER instruction - * that sets up the real kernel stack. Check here, since we can't - * allow the wrong stack to be used. - * - * "TSS_sysenter_sp0+12" is because the NMI/debug handler will have - * already pushed 3 words if it hits on the sysenter instruction: - * eflags, cs and eip. - * - * We just load the right stack, and push the three (known) values - * by hand onto the new stack - while updating the return eip past - * the instruction that would have done it for sysenter. - */ -#define FIX_STACK(offset, ok, label) \ - cmpw $__KERNEL_CS,4(%esp); \ - jne ok; \ -label: \ - movl TSS_sysenter_sp0+offset(%esp),%esp; \ - CFI_DEF_CFA esp, 0; \ - CFI_UNDEFINED eip; \ - pushfl; \ - CFI_ADJUST_CFA_OFFSET 4; \ - pushl $__KERNEL_CS; \ - CFI_ADJUST_CFA_OFFSET 4; \ - pushl $sysenter_past_esp; \ - CFI_ADJUST_CFA_OFFSET 4; \ - CFI_REL_OFFSET eip, 0 - -KPROBE_ENTRY(debug) - RING0_INT_FRAME - cmpl $ia32_sysenter_target,(%esp) - jne debug_stack_correct - FIX_STACK(12, debug_stack_correct, debug_esp_fix_insn) -debug_stack_correct: - pushl $-1 # mark this as an int - CFI_ADJUST_CFA_OFFSET 4 - SAVE_ALL - TRACE_IRQS_OFF - xorl %edx,%edx # error code 0 - movl %esp,%eax # pt_regs pointer - call do_debug - jmp ret_from_exception - CFI_ENDPROC -KPROBE_END(debug) - -/* - * NMI is doubly nasty. It can happen _while_ we're handling - * a debug fault, and the debug fault hasn't yet been able to - * clear up the stack. So we first check whether we got an - * NMI on the sysenter entry path, but after that we need to - * check whether we got an NMI on the debug path where the debug - * fault happened on the sysenter path. - */ -KPROBE_ENTRY(nmi) - RING0_INT_FRAME - pushl %eax - CFI_ADJUST_CFA_OFFSET 4 - movl %ss, %eax - cmpw $__ESPFIX_SS, %ax - popl %eax - CFI_ADJUST_CFA_OFFSET -4 - je nmi_espfix_stack - cmpl $ia32_sysenter_target,(%esp) - je nmi_stack_fixup - pushl %eax - CFI_ADJUST_CFA_OFFSET 4 - movl %esp,%eax - /* Do not access memory above the end of our stack page, - * it might not exist. - */ - andl $(THREAD_SIZE-1),%eax - cmpl $(THREAD_SIZE-20),%eax - popl %eax - CFI_ADJUST_CFA_OFFSET -4 - jae nmi_stack_correct - cmpl $ia32_sysenter_target,12(%esp) - je nmi_debug_stack_check -nmi_stack_correct: - /* We have a RING0_INT_FRAME here */ - pushl %eax - CFI_ADJUST_CFA_OFFSET 4 - SAVE_ALL - TRACE_IRQS_OFF - xorl %edx,%edx # zero error code - movl %esp,%eax # pt_regs pointer - call do_nmi - jmp restore_nocheck_notrace - CFI_ENDPROC - -nmi_stack_fixup: - RING0_INT_FRAME - FIX_STACK(12,nmi_stack_correct, 1) - jmp nmi_stack_correct - -nmi_debug_stack_check: - /* We have a RING0_INT_FRAME here */ - cmpw $__KERNEL_CS,16(%esp) - jne nmi_stack_correct - cmpl $debug,(%esp) - jb nmi_stack_correct - cmpl $debug_esp_fix_insn,(%esp) - ja nmi_stack_correct - FIX_STACK(24,nmi_stack_correct, 1) - jmp nmi_stack_correct - -nmi_espfix_stack: - /* We have a RING0_INT_FRAME here. - * - * create the pointer to lss back - */ - pushl %ss - CFI_ADJUST_CFA_OFFSET 4 - pushl %esp - CFI_ADJUST_CFA_OFFSET 4 - addw $4, (%esp) - /* copy the iret frame of 12 bytes */ - .rept 3 - pushl 16(%esp) - CFI_ADJUST_CFA_OFFSET 4 - .endr - pushl %eax - CFI_ADJUST_CFA_OFFSET 4 - SAVE_ALL - TRACE_IRQS_OFF - FIXUP_ESPFIX_STACK # %eax == %esp - xorl %edx,%edx # zero error code - call do_nmi - RESTORE_REGS - lss 12+4(%esp), %esp # back to espfix stack - CFI_ADJUST_CFA_OFFSET -24 - jmp irq_return - CFI_ENDPROC -KPROBE_END(nmi) - #ifdef CONFIG_PARAVIRT ENTRY(native_iret) iret @@ -926,19 +733,6 @@ ENTRY(native_irq_enable_sysexit) END(native_irq_enable_sysexit) #endif -KPROBE_ENTRY(int3) - RING0_INT_FRAME - pushl $-1 # mark this as an int - CFI_ADJUST_CFA_OFFSET 4 - SAVE_ALL - TRACE_IRQS_OFF - xorl %edx,%edx # zero error code - movl %esp,%eax # pt_regs pointer - call do_int3 - jmp ret_from_exception - CFI_ENDPROC -KPROBE_END(int3) - ENTRY(overflow) RING0_INT_FRAME pushl $0 @@ -1003,14 +797,6 @@ ENTRY(stack_segment) CFI_ENDPROC END(stack_segment) -KPROBE_ENTRY(general_protection) - RING0_EC_FRAME - pushl $do_general_protection - CFI_ADJUST_CFA_OFFSET 4 - jmp error_code - CFI_ENDPROC -KPROBE_END(general_protection) - ENTRY(alignment_check) RING0_EC_FRAME pushl $do_alignment_check @@ -1220,3 +1006,227 @@ END(mcount) #include "syscall_table_32.S" syscall_table_size=(.-sys_call_table) + +/* + * Some functions should be protected against kprobes + */ + .pushsection .kprobes.text, "ax" + +ENTRY(page_fault) + RING0_EC_FRAME + pushl $do_page_fault + CFI_ADJUST_CFA_OFFSET 4 + ALIGN +error_code: + /* the function address is in %fs's slot on the stack */ + pushl %es + CFI_ADJUST_CFA_OFFSET 4 + /*CFI_REL_OFFSET es, 0*/ + pushl %ds + CFI_ADJUST_CFA_OFFSET 4 + /*CFI_REL_OFFSET ds, 0*/ + pushl %eax + CFI_ADJUST_CFA_OFFSET 4 + CFI_REL_OFFSET eax, 0 + pushl %ebp + CFI_ADJUST_CFA_OFFSET 4 + CFI_REL_OFFSET ebp, 0 + pushl %edi + CFI_ADJUST_CFA_OFFSET 4 + CFI_REL_OFFSET edi, 0 + pushl %esi + CFI_ADJUST_CFA_OFFSET 4 + CFI_REL_OFFSET esi, 0 + pushl %edx + CFI_ADJUST_CFA_OFFSET 4 + CFI_REL_OFFSET edx, 0 + pushl %ecx + CFI_ADJUST_CFA_OFFSET 4 + CFI_REL_OFFSET ecx, 0 + pushl %ebx + CFI_ADJUST_CFA_OFFSET 4 + CFI_REL_OFFSET ebx, 0 + cld + pushl %fs + CFI_ADJUST_CFA_OFFSET 4 + /*CFI_REL_OFFSET fs, 0*/ + movl $(__KERNEL_PERCPU), %ecx + movl %ecx, %fs + UNWIND_ESPFIX_STACK + popl %ecx + CFI_ADJUST_CFA_OFFSET -4 + /*CFI_REGISTER es, ecx*/ + movl PT_FS(%esp), %edi # get the function address + movl PT_ORIG_EAX(%esp), %edx # get the error code + movl $-1, PT_ORIG_EAX(%esp) # no syscall to restart + mov %ecx, PT_FS(%esp) + /*CFI_REL_OFFSET fs, ES*/ + movl $(__USER_DS), %ecx + movl %ecx, %ds + movl %ecx, %es + TRACE_IRQS_OFF + movl %esp,%eax # pt_regs pointer + call *%edi + jmp ret_from_exception + CFI_ENDPROC +END(page_fault) + +/* + * Debug traps and NMI can happen at the one SYSENTER instruction + * that sets up the real kernel stack. Check here, since we can't + * allow the wrong stack to be used. + * + * "TSS_sysenter_sp0+12" is because the NMI/debug handler will have + * already pushed 3 words if it hits on the sysenter instruction: + * eflags, cs and eip. + * + * We just load the right stack, and push the three (known) values + * by hand onto the new stack - while updating the return eip past + * the instruction that would have done it for sysenter. + */ +#define FIX_STACK(offset, ok, label) \ + cmpw $__KERNEL_CS,4(%esp); \ + jne ok; \ +label: \ + movl TSS_sysenter_sp0+offset(%esp),%esp; \ + CFI_DEF_CFA esp, 0; \ + CFI_UNDEFINED eip; \ + pushfl; \ + CFI_ADJUST_CFA_OFFSET 4; \ + pushl $__KERNEL_CS; \ + CFI_ADJUST_CFA_OFFSET 4; \ + pushl $sysenter_past_esp; \ + CFI_ADJUST_CFA_OFFSET 4; \ + CFI_REL_OFFSET eip, 0 + +ENTRY(debug) + RING0_INT_FRAME + cmpl $ia32_sysenter_target,(%esp) + jne debug_stack_correct + FIX_STACK(12, debug_stack_correct, debug_esp_fix_insn) +debug_stack_correct: + pushl $-1 # mark this as an int + CFI_ADJUST_CFA_OFFSET 4 + SAVE_ALL + TRACE_IRQS_OFF + xorl %edx,%edx # error code 0 + movl %esp,%eax # pt_regs pointer + call do_debug + jmp ret_from_exception + CFI_ENDPROC +END(debug) + +/* + * NMI is doubly nasty. It can happen _while_ we're handling + * a debug fault, and the debug fault hasn't yet been able to + * clear up the stack. So we first check whether we got an + * NMI on the sysenter entry path, but after that we need to + * check whether we got an NMI on the debug path where the debug + * fault happened on the sysenter path. + */ +ENTRY(nmi) + RING0_INT_FRAME + pushl %eax + CFI_ADJUST_CFA_OFFSET 4 + movl %ss, %eax + cmpw $__ESPFIX_SS, %ax + popl %eax + CFI_ADJUST_CFA_OFFSET -4 + je nmi_espfix_stack + cmpl $ia32_sysenter_target,(%esp) + je nmi_stack_fixup + pushl %eax + CFI_ADJUST_CFA_OFFSET 4 + movl %esp,%eax + /* Do not access memory above the end of our stack page, + * it might not exist. + */ + andl $(THREAD_SIZE-1),%eax + cmpl $(THREAD_SIZE-20),%eax + popl %eax + CFI_ADJUST_CFA_OFFSET -4 + jae nmi_stack_correct + cmpl $ia32_sysenter_target,12(%esp) + je nmi_debug_stack_check +nmi_stack_correct: + /* We have a RING0_INT_FRAME here */ + pushl %eax + CFI_ADJUST_CFA_OFFSET 4 + SAVE_ALL + TRACE_IRQS_OFF + xorl %edx,%edx # zero error code + movl %esp,%eax # pt_regs pointer + call do_nmi + jmp restore_nocheck_notrace + CFI_ENDPROC + +nmi_stack_fixup: + RING0_INT_FRAME + FIX_STACK(12,nmi_stack_correct, 1) + jmp nmi_stack_correct + +nmi_debug_stack_check: + /* We have a RING0_INT_FRAME here */ + cmpw $__KERNEL_CS,16(%esp) + jne nmi_stack_correct + cmpl $debug,(%esp) + jb nmi_stack_correct + cmpl $debug_esp_fix_insn,(%esp) + ja nmi_stack_correct + FIX_STACK(24,nmi_stack_correct, 1) + jmp nmi_stack_correct + +nmi_espfix_stack: + /* We have a RING0_INT_FRAME here. + * + * create the pointer to lss back + */ + pushl %ss + CFI_ADJUST_CFA_OFFSET 4 + pushl %esp + CFI_ADJUST_CFA_OFFSET 4 + addw $4, (%esp) + /* copy the iret frame of 12 bytes */ + .rept 3 + pushl 16(%esp) + CFI_ADJUST_CFA_OFFSET 4 + .endr + pushl %eax + CFI_ADJUST_CFA_OFFSET 4 + SAVE_ALL + TRACE_IRQS_OFF + FIXUP_ESPFIX_STACK # %eax == %esp + xorl %edx,%edx # zero error code + call do_nmi + RESTORE_REGS + lss 12+4(%esp), %esp # back to espfix stack + CFI_ADJUST_CFA_OFFSET -24 + jmp irq_return + CFI_ENDPROC +END(nmi) + +ENTRY(int3) + RING0_INT_FRAME + pushl $-1 # mark this as an int + CFI_ADJUST_CFA_OFFSET 4 + SAVE_ALL + TRACE_IRQS_OFF + xorl %edx,%edx # zero error code + movl %esp,%eax # pt_regs pointer + call do_int3 + jmp ret_from_exception + CFI_ENDPROC +END(int3) + +ENTRY(general_protection) + RING0_EC_FRAME + pushl $do_general_protection + CFI_ADJUST_CFA_OFFSET 4 + jmp error_code + CFI_ENDPROC +END(general_protection) + +/* + * End of kprobes section + */ + .popsection diff --git a/include/linux/linkage.h b/include/linux/linkage.h index 9fd1f859021b..fee9e59649c1 100644 --- a/include/linux/linkage.h +++ b/include/linux/linkage.h @@ -64,14 +64,6 @@ name: #endif -#define KPROBE_ENTRY(name) \ - .pushsection .kprobes.text, "ax"; \ - ENTRY(name) - -#define KPROBE_END(name) \ - END(name); \ - .popsection - #ifndef END #define END(name) \ .size name, .-name -- cgit v1.2.3 From 33454539f386a2beb38269bea5fff82b3d56b0e9 Mon Sep 17 00:00:00 2001 From: "gorcunov@gmail.com" Date: Wed, 26 Nov 2008 22:17:02 +0300 Subject: x86: entry_64.S - use X86_EFLAGS_IF instead of hardcoded number Impact: cleanup Signed-off-by: Cyrill Gorcunov Acked-by: Alexander van Heukelum Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_64.S | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 38fcd0517c31..1c309d546518 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -194,7 +194,7 @@ ENTRY(native_usergs_sysret64) pushq %rax /* rsp */ CFI_ADJUST_CFA_OFFSET 8 CFI_REL_OFFSET rsp,0 - pushq $(1<<9) /* eflags - interrupts on */ + pushq $X86_EFLAGS_IF /* eflags - interrupts on */ CFI_ADJUST_CFA_OFFSET 8 /*CFI_REL_OFFSET rflags,0*/ pushq $__KERNEL_CS /* cs */ -- cgit v1.2.3 From c2c631e318091118587f3b766347d259c9265b8b Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Wed, 26 Nov 2008 22:17:00 +0300 Subject: x86: entry_64.S - use ENTRY to define child_rip child_rip is called not by its name but indirectly rather so make it global and aligned. Signed-off-by: Cyrill Gorcunov Acked-by: Alexander van Heukelum Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_64.S | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 1c309d546518..0a910a7f85f5 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -1160,7 +1160,7 @@ ENTRY(kernel_thread) CFI_ENDPROC END(kernel_thread) -child_rip: +ENTRY(child_rip) pushq $0 # fake return address CFI_STARTPROC /* -- cgit v1.2.3 From 1d9b16d1690fe5edb1c907fe4746681cf026cdf3 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Thu, 27 Nov 2008 18:39:15 +0100 Subject: x86: move GART specific stuff from iommu.h to gart.h Impact: cleanup Signed-off-by: Joerg Roedel Signed-off-by: Ingo Molnar --- arch/x86/include/asm/gart.h | 33 +++++++++++++++++++++++++++++++++ arch/x86/include/asm/iommu.h | 33 --------------------------------- arch/x86/kernel/amd_iommu.c | 1 + arch/x86/kernel/amd_iommu_init.c | 1 + arch/x86/kernel/early-quirks.c | 1 + arch/x86/kernel/pci-dma.c | 1 + arch/x86/kernel/setup.c | 1 + 7 files changed, 38 insertions(+), 33 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/gart.h b/arch/x86/include/asm/gart.h index 74252264433d..6cfdafa409d8 100644 --- a/arch/x86/include/asm/gart.h +++ b/arch/x86/include/asm/gart.h @@ -29,6 +29,39 @@ extern int fix_aperture; #define AMD64_GARTCACHECTL 0x9c #define AMD64_GARTEN (1<<0) +#ifdef CONFIG_GART_IOMMU +extern int gart_iommu_aperture; +extern int gart_iommu_aperture_allowed; +extern int gart_iommu_aperture_disabled; + +extern void early_gart_iommu_check(void); +extern void gart_iommu_init(void); +extern void gart_iommu_shutdown(void); +extern void __init gart_parse_options(char *); +extern void gart_iommu_hole_init(void); + +#else +#define gart_iommu_aperture 0 +#define gart_iommu_aperture_allowed 0 +#define gart_iommu_aperture_disabled 1 + +static inline void early_gart_iommu_check(void) +{ +} +static inline void gart_iommu_init(void) +{ +} +static inline void gart_iommu_shutdown(void) +{ +} +static inline void gart_parse_options(char *options) +{ +} +static inline void gart_iommu_hole_init(void) +{ +} +#endif + extern int agp_amd64_init(void); static inline void enable_gart_translation(struct pci_dev *dev, u64 addr) diff --git a/arch/x86/include/asm/iommu.h b/arch/x86/include/asm/iommu.h index 0b500c5b6446..295b13193f4d 100644 --- a/arch/x86/include/asm/iommu.h +++ b/arch/x86/include/asm/iommu.h @@ -12,37 +12,4 @@ extern unsigned long iommu_nr_pages(unsigned long addr, unsigned long len); /* 10 seconds */ #define DMAR_OPERATION_TIMEOUT ((cycles_t) tsc_khz*10*1000) -#ifdef CONFIG_GART_IOMMU -extern int gart_iommu_aperture; -extern int gart_iommu_aperture_allowed; -extern int gart_iommu_aperture_disabled; - -extern void early_gart_iommu_check(void); -extern void gart_iommu_init(void); -extern void gart_iommu_shutdown(void); -extern void __init gart_parse_options(char *); -extern void gart_iommu_hole_init(void); - -#else -#define gart_iommu_aperture 0 -#define gart_iommu_aperture_allowed 0 -#define gart_iommu_aperture_disabled 1 - -static inline void early_gart_iommu_check(void) -{ -} -static inline void gart_iommu_init(void) -{ -} -static inline void gart_iommu_shutdown(void) -{ -} -static inline void gart_parse_options(char *options) -{ -} -static inline void gart_iommu_hole_init(void) -{ -} -#endif - #endif /* _ASM_X86_IOMMU_H */ diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 331b318304eb..172e0dc4641e 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -24,6 +24,7 @@ #include #include #include +#include #include #include diff --git a/arch/x86/kernel/amd_iommu_init.c b/arch/x86/kernel/amd_iommu_init.c index 0cdcda35a05f..7685f0774a8f 100644 --- a/arch/x86/kernel/amd_iommu_init.c +++ b/arch/x86/kernel/amd_iommu_init.c @@ -28,6 +28,7 @@ #include #include #include +#include /* * definitions for the ACPI scanning code diff --git a/arch/x86/kernel/early-quirks.c b/arch/x86/kernel/early-quirks.c index 1b894b72c0f5..744aa7fc49d5 100644 --- a/arch/x86/kernel/early-quirks.c +++ b/arch/x86/kernel/early-quirks.c @@ -17,6 +17,7 @@ #include #include #include +#include static void __init fix_hypertransport_config(int num, int slot, int func) { diff --git a/arch/x86/kernel/pci-dma.c b/arch/x86/kernel/pci-dma.c index 192624820217..12eeb4bfcdeb 100644 --- a/arch/x86/kernel/pci-dma.c +++ b/arch/x86/kernel/pci-dma.c @@ -6,6 +6,7 @@ #include #include #include +#include #include #include diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index 0fa6790c1dd3..67d5979e654e 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -93,6 +93,7 @@ #include #include #include +#include #include #include -- cgit v1.2.3 From 5ae3a139cf4fc2349f1dfa1993a66c1dcc119468 Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Thu, 27 Nov 2008 00:02:10 +0300 Subject: x86: uv bau interrupt -- use proper interrupt number Signed-off-by: Cyrill Gorcunov Acked-by: Cliff Wickman Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_64.S | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 0a910a7f85f5..57d7f7a5ad2f 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -938,7 +938,7 @@ apicinterrupt IRQ_MOVE_CLEANUP_VECTOR \ irq_move_cleanup_interrupt smp_irq_move_cleanup_interrupt #endif -apicinterrupt 220 \ +apicinterrupt UV_BAU_MESSAGE \ uv_bau_message_intr1 uv_bau_message_interrupt apicinterrupt LOCAL_TIMER_VECTOR \ apic_timer_interrupt smp_apic_timer_interrupt -- cgit v1.2.3 From 9f1e87ea3ecb3c46c21f6a1a202ec82f99ed2473 Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Thu, 27 Nov 2008 21:10:08 +0300 Subject: x86: entry_64.S - trivial: space, comments fixup Impact: cleanup Signed-off-by: Cyrill Gorcunov Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_64.S | 94 +++++++++++++++++++++++----------------------- 1 file changed, 48 insertions(+), 46 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 57d7f7a5ad2f..08c0c9777a09 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -1020,7 +1020,7 @@ END(\sym) .macro paranoidzeroentry_ist sym do_sym ist ENTRY(\sym) - INTR_FRAME + INTR_FRAME PARAVIRT_ADJUST_EXCEPTION_FRAME pushq $-1 /* ORIG_RAX: no syscall to restart */ CFI_ADJUST_CFA_OFFSET 8 @@ -1088,36 +1088,36 @@ zeroentry coprocessor_error do_coprocessor_error errorentry alignment_check do_alignment_check zeroentry simd_coprocessor_error do_simd_coprocessor_error - /* Reload gs selector with exception handling */ - /* edi: new selector */ + /* Reload gs selector with exception handling */ + /* edi: new selector */ ENTRY(native_load_gs_index) CFI_STARTPROC pushf CFI_ADJUST_CFA_OFFSET 8 DISABLE_INTERRUPTS(CLBR_ANY | ~(CLBR_RDI)) - SWAPGS + SWAPGS gs_change: - movl %edi,%gs + movl %edi,%gs 2: mfence /* workaround */ SWAPGS - popf + popf CFI_ADJUST_CFA_OFFSET -8 - ret + ret CFI_ENDPROC END(native_load_gs_index) - .section __ex_table,"a" - .align 8 - .quad gs_change,bad_gs - .previous - .section .fixup,"ax" + .section __ex_table,"a" + .align 8 + .quad gs_change,bad_gs + .previous + .section .fixup,"ax" /* running with kernelgs */ bad_gs: SWAPGS /* switch back to user gs */ xorl %eax,%eax - movl %eax,%gs - jmp 2b - .previous + movl %eax,%gs + jmp 2b + .previous /* * Create a kernel thread. @@ -1152,7 +1152,7 @@ ENTRY(kernel_thread) * so internally to the x86_64 port you can rely on kernel_thread() * not to reschedule the child before returning, this avoids the need * of hacks for example to fork off the per-CPU idle tasks. - * [Hopefully no generic code relies on the reschedule -AK] + * [Hopefully no generic code relies on the reschedule -AK] */ RESTORE_ALL UNFAKE_STACK_FRAME @@ -1231,22 +1231,24 @@ END(call_softirq) zeroentry xen_hypervisor_callback xen_do_hypervisor_callback /* -# A note on the "critical region" in our callback handler. -# We want to avoid stacking callback handlers due to events occurring -# during handling of the last event. To do this, we keep events disabled -# until we've done all processing. HOWEVER, we must enable events before -# popping the stack frame (can't be done atomically) and so it would still -# be possible to get enough handler activations to overflow the stack. -# Although unlikely, bugs of that kind are hard to track down, so we'd -# like to avoid the possibility. -# So, on entry to the handler we detect whether we interrupted an -# existing activation in its critical region -- if so, we pop the current -# activation and restart the handler using the previous one. -*/ + * A note on the "critical region" in our callback handler. + * We want to avoid stacking callback handlers due to events occurring + * during handling of the last event. To do this, we keep events disabled + * until we've done all processing. HOWEVER, we must enable events before + * popping the stack frame (can't be done atomically) and so it would still + * be possible to get enough handler activations to overflow the stack. + * Although unlikely, bugs of that kind are hard to track down, so we'd + * like to avoid the possibility. + * So, on entry to the handler we detect whether we interrupted an + * existing activation in its critical region -- if so, we pop the current + * activation and restart the handler using the previous one. + */ ENTRY(xen_do_hypervisor_callback) # do_hypervisor_callback(struct *pt_regs) CFI_STARTPROC -/* Since we don't modify %rdi, evtchn_do_upall(struct *pt_regs) will - see the correct pointer to the pt_regs */ +/* + * Since we don't modify %rdi, evtchn_do_upall(struct *pt_regs) will + * see the correct pointer to the pt_regs + */ movq %rdi, %rsp # we don't return, adjust the stack frame CFI_ENDPROC DEFAULT_FRAME @@ -1264,18 +1266,18 @@ ENTRY(xen_do_hypervisor_callback) # do_hypervisor_callback(struct *pt_regs) END(do_hypervisor_callback) /* -# Hypervisor uses this for application faults while it executes. -# We get here for two reasons: -# 1. Fault while reloading DS, ES, FS or GS -# 2. Fault while executing IRET -# Category 1 we do not need to fix up as Xen has already reloaded all segment -# registers that could be reloaded and zeroed the others. -# Category 2 we fix up by killing the current process. We cannot use the -# normal Linux return path in this case because if we use the IRET hypercall -# to pop the stack frame we end up in an infinite loop of failsafe callbacks. -# We distinguish between categories by comparing each saved segment register -# with its current contents: any discrepancy means we in category 1. -*/ + * Hypervisor uses this for application faults while it executes. + * We get here for two reasons: + * 1. Fault while reloading DS, ES, FS or GS + * 2. Fault while executing IRET + * Category 1 we do not need to fix up as Xen has already reloaded all segment + * registers that could be reloaded and zeroed the others. + * Category 2 we fix up by killing the current process. We cannot use the + * normal Linux return path in this case because if we use the IRET hypercall + * to pop the stack frame we end up in an infinite loop of failsafe callbacks. + * We distinguish between categories by comparing each saved segment register + * with its current contents: any discrepancy means we in category 1. + */ ENTRY(xen_failsafe_callback) INTR_FRAME 1 (6*8) /*CFI_REL_OFFSET gs,GS*/ @@ -1339,8 +1341,8 @@ paranoidzeroentry machine_check do_machine_check #endif /* - * "Paranoid" exit path from exception stack. - * Paranoid because this is used by NMIs and cannot take + * "Paranoid" exit path from exception stack. + * Paranoid because this is used by NMIs and cannot take * any kernel state for granted. * We don't do kernel preemption checks here, because only * NMI should be common and it does not enable IRQs and @@ -1445,7 +1447,7 @@ error_kernelspace: cmpq %rcx,RIP+8(%rsp) je error_swapgs cmpq $gs_change,RIP+8(%rsp) - je error_swapgs + je error_swapgs jmp error_sti END(error_entry) @@ -1521,7 +1523,7 @@ nmi_schedule: CFI_ENDPROC #else jmp paranoid_exit - CFI_ENDPROC + CFI_ENDPROC #endif END(nmi) -- cgit v1.2.3 From 5b3eec0c80038c8739ccd465b897a35c0dff1cc4 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Thu, 27 Nov 2008 14:41:21 +0100 Subject: x86: ret_from_fork - get rid of jump back Impact: remove dead code If we take a closer look at the rff_trace/rff_action ret_from_fork code, we have to realize that it does all the wrong things: for example it checks the TIF flag - while later on jumping back to the ret-from-syscall path - duplicating the check needlessly. But checking for _TIF_SYSCALL_TRACE is completely unnecessary here because we clear that flag for every freshly forked task. So the whole "tracing" code here, for which there is a out of line jump optimization that makes it even harder to read, is in reality completely dead code ... Reported-by: Cyrill Gorcunov Signed-off-by: Ingo Molnar Tested-by: Cyrill Gorcunov --- arch/x86/kernel/entry_64.S | 31 ++++++++++++++++--------------- 1 file changed, 16 insertions(+), 15 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index e41734a537bd..3194636a4293 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -361,34 +361,35 @@ ENTRY(save_paranoid) END(save_paranoid) /* - * A newly forked process directly context switches into this. + * A newly forked process directly context switches into this address. + * + * rdi: prev task we switched from */ -/* rdi: prev */ ENTRY(ret_from_fork) DEFAULT_FRAME + push kernel_eflags(%rip) CFI_ADJUST_CFA_OFFSET 8 - popf # reset kernel eflags + popf # reset kernel eflags CFI_ADJUST_CFA_OFFSET -8 - call schedule_tail + + call schedule_tail # rdi: 'prev' task parameter + GET_THREAD_INFO(%rcx) - testl $(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT),TI_flags(%rcx) + CFI_REMEMBER_STATE - jnz rff_trace -rff_action: RESTORE_REST - testl $3,CS-ARGOFFSET(%rsp) # from kernel_thread? + + testl $3, CS-ARGOFFSET(%rsp) # from kernel_thread? je int_ret_from_sys_call - testl $_TIF_IA32,TI_flags(%rcx) + + testl $_TIF_IA32, TI_flags(%rcx) # 32-bit compat task needs IRET jnz int_ret_from_sys_call + RESTORE_TOP_OF_STACK %rdi, -ARGOFFSET - jmp ret_from_sys_call + jmp ret_from_sys_call # go to the SYSRET fastpath + CFI_RESTORE_STATE -rff_trace: - movq %rsp,%rdi - call syscall_trace_leave - GET_THREAD_INFO(%rcx) - jmp rff_action CFI_ENDPROC END(ret_from_fork) -- cgit v1.2.3 From 8caac56305cef98f9357b060a77939d17699937d Mon Sep 17 00:00:00 2001 From: Pavel Machek Date: Wed, 26 Nov 2008 17:15:27 +0100 Subject: aperture_64.c: clarify that too small aperture is valid reason for this code Impact: update comment Clarify that too small aperture is valid reason for this code. Signed-off-by: Pavel Machek Signed-off-by: Ingo Molnar --- arch/x86/kernel/aperture_64.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/aperture_64.c b/arch/x86/kernel/aperture_64.c index 9a32b37ee2ee..676debfc1702 100644 --- a/arch/x86/kernel/aperture_64.c +++ b/arch/x86/kernel/aperture_64.c @@ -1,8 +1,9 @@ /* * Firmware replacement code. * - * Work around broken BIOSes that don't set an aperture or only set the - * aperture in the AGP bridge. + * Work around broken BIOSes that don't set an aperture, only set the + * aperture in the AGP bridge, or set too small aperture. + * * If all fails map the aperture over some low memory. This is cheaper than * doing bounce buffering. The memory is lost. This is done at early boot * because only the bootmem allocator can allocate 32+MB. -- cgit v1.2.3 From 4385cecf1f5866fb33fc95e2ee26a44e9b6f6be2 Mon Sep 17 00:00:00 2001 From: Jiri Slaby Date: Sat, 29 Nov 2008 22:33:16 +0100 Subject: x86: intel_cacheinfo, minor show_type cleanup Impact: cleanup Signed-off-by: Jiri Slaby Cc: Jiri Slaby Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/intel_cacheinfo.c | 17 +++++++---------- 1 file changed, 7 insertions(+), 10 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/intel_cacheinfo.c b/arch/x86/kernel/cpu/intel_cacheinfo.c index 3f46afbb1cf1..68b5d8681cbb 100644 --- a/arch/x86/kernel/cpu/intel_cacheinfo.c +++ b/arch/x86/kernel/cpu/intel_cacheinfo.c @@ -644,20 +644,17 @@ static inline ssize_t show_shared_cpu_list(struct _cpuid4_info *leaf, char *buf) return show_shared_cpu_map_func(leaf, 1, buf); } -static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf) { - switch(this_leaf->eax.split.type) { - case CACHE_TYPE_DATA: +static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf) +{ + switch (this_leaf->eax.split.type) { + case CACHE_TYPE_DATA: return sprintf(buf, "Data\n"); - break; - case CACHE_TYPE_INST: + case CACHE_TYPE_INST: return sprintf(buf, "Instruction\n"); - break; - case CACHE_TYPE_UNIFIED: + case CACHE_TYPE_UNIFIED: return sprintf(buf, "Unified\n"); - break; - default: + default: return sprintf(buf, "Unknown\n"); - break; } } -- cgit v1.2.3 From 2c5643b1c5c7fbb13f340d4c58944d9642f41796 Mon Sep 17 00:00:00 2001 From: Hitoshi Mitake Date: Sun, 30 Nov 2008 17:16:04 +0900 Subject: x86: provide readq()/writeq() on 32-bit too Impact: add new API for drivers Add implementation of readq/writeq to x86_32, and add config value to the x86 architecture to determine existence of readq/writeq. Signed-off-by: Hitoshi Mitake Acked-by: Sam Ravnborg Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 2 ++ arch/x86/include/asm/io.h | 24 ++++++++++++++++++++++++ 2 files changed, 26 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index ac22bb7719f7..a7d50f5d118c 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -19,6 +19,8 @@ config X86_64 config X86 def_bool y select HAVE_AOUT if X86_32 + select HAVE_READQ + select HAVE_WRITEQ select HAVE_UNSTABLE_SCHED_CLOCK select HAVE_IDE select HAVE_OPROFILE diff --git a/arch/x86/include/asm/io.h b/arch/x86/include/asm/io.h index ac2abc88cd95..25946449df4f 100644 --- a/arch/x86/include/asm/io.h +++ b/arch/x86/include/asm/io.h @@ -4,6 +4,7 @@ #define ARCH_HAS_IOREMAP_WC #include +#include #define build_mmio_read(name, size, type, reg, barrier) \ static inline type name(const volatile void __iomem *addr) \ @@ -57,6 +58,29 @@ build_mmio_write(__writeq, "q", unsigned long, "r", ) /* Let people know we have them */ #define readq readq #define writeq writeq + +#else /* CONFIG_X86_32 from here */ + +static inline __u64 readq(const volatile void __iomem *addr) +{ + const volatile u32 __iomem *p = addr; + u32 l, h; + + l = readl(p); + h = readl(p + 1); + + return l + ((u64)h << 32); +} + +static inline void writeq(__u64 val, volatile void __iomem *addr) +{ + writel(val, addr); + writel(val >> 32, addr+4); +} + +#define readq readq +#define writeq writeq + #endif extern int iommu_bio_merge; -- cgit v1.2.3 From a0b1131e479e5af32eefac8bc54c9742e23d638e Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Sun, 30 Nov 2008 09:33:55 +0100 Subject: x86: provide readq()/writeq() on 32-bit too, cleanup Impact: cleanup Signed-off-by: Ingo Molnar --- arch/x86/include/asm/io.h | 17 +++++++---------- 1 file changed, 7 insertions(+), 10 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/io.h b/arch/x86/include/asm/io.h index 25946449df4f..3ccfaf610c89 100644 --- a/arch/x86/include/asm/io.h +++ b/arch/x86/include/asm/io.h @@ -55,21 +55,17 @@ build_mmio_write(__writeq, "q", unsigned long, "r", ) #define __raw_readq __readq #define __raw_writeq writeq -/* Let people know we have them */ -#define readq readq -#define writeq writeq - #else /* CONFIG_X86_32 from here */ static inline __u64 readq(const volatile void __iomem *addr) { const volatile u32 __iomem *p = addr; - u32 l, h; + u32 low, high; - l = readl(p); - h = readl(p + 1); + low = readl(p); + high = readl(p + 1); - return l + ((u64)h << 32); + return low + ((u64)high << 32); } static inline void writeq(__u64 val, volatile void __iomem *addr) @@ -78,11 +74,12 @@ static inline void writeq(__u64 val, volatile void __iomem *addr) writel(val >> 32, addr+4); } +#endif + +/* Let people know that we have them */ #define readq readq #define writeq writeq -#endif - extern int iommu_bio_merge; #ifdef CONFIG_X86_32 -- cgit v1.2.3 From 93093d099e5dd0c258fd530c12668e828c20df41 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Sun, 30 Nov 2008 10:20:20 +0100 Subject: x86: provide readq()/writeq() on 32-bit too, complete if HAVE_READQ/HAVE_WRITEQ are defined, the full range of readq/writeq APIs has to be provided to drivers: drivers/infiniband/hw/amso1100/c2.c: In function 'c2_tx_ring_alloc': drivers/infiniband/hw/amso1100/c2.c:133: error: implicit declaration of function '__raw_writeq' So provide them on 32-bit as well. Also, map all the APIs to the strongest ordering variant. It's way too easy to mess such details up in drivers and the difference between "memory" and "" constrained asm() constructs is in the noise range. Signed-off-by: Ingo Molnar --- arch/x86/include/asm/io.h | 18 +++++++++--------- 1 file changed, 9 insertions(+), 9 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/io.h b/arch/x86/include/asm/io.h index 3ccfaf610c89..33513b9a67f3 100644 --- a/arch/x86/include/asm/io.h +++ b/arch/x86/include/asm/io.h @@ -46,16 +46,11 @@ build_mmio_write(__writel, "l", unsigned int, "r", ) #define mmiowb() barrier() #ifdef CONFIG_X86_64 + build_mmio_read(readq, "q", unsigned long, "=r", :"memory") -build_mmio_read(__readq, "q", unsigned long, "=r", ) build_mmio_write(writeq, "q", unsigned long, "r", :"memory") -build_mmio_write(__writeq, "q", unsigned long, "r", ) - -#define readq_relaxed(a) __readq(a) -#define __raw_readq __readq -#define __raw_writeq writeq -#else /* CONFIG_X86_32 from here */ +#else static inline __u64 readq(const volatile void __iomem *addr) { @@ -76,9 +71,14 @@ static inline void writeq(__u64 val, volatile void __iomem *addr) #endif +#define readq_relaxed(a) readq(a) + +#define __raw_readq(a) readq(a) +#define __raw_writeq(val, addr) writeq(val, addr) + /* Let people know that we have them */ -#define readq readq -#define writeq writeq +#define readq readq +#define writeq writeq extern int iommu_bio_merge; -- cgit v1.2.3 From 50cec5c51c18301ff60262fdbe920f4a907c9d81 Mon Sep 17 00:00:00 2001 From: FUJITA Tomonori Date: Tue, 2 Dec 2008 02:17:15 +0900 Subject: x86: fix dma_mapping_error for 32bit x86, cleanup This removes ifdef CONFIG_X86_64 in dma_mapping_error(): 1) Xen people plan to use swiotlb on X86_32 for Dom0 support. swiotlb uses ops->mapping_error so X86_32 also needs to check ops->mapping_error. 2) Removing #ifdef hack is almost always a good thing. Signed-off-by: FUJITA Tomonori Signed-off-by: Ingo Molnar --- arch/x86/include/asm/dma-mapping.h | 2 -- 1 file changed, 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/dma-mapping.h b/arch/x86/include/asm/dma-mapping.h index 097794ff6b79..dc22c0733282 100644 --- a/arch/x86/include/asm/dma-mapping.h +++ b/arch/x86/include/asm/dma-mapping.h @@ -71,12 +71,10 @@ static inline struct dma_mapping_ops *get_dma_ops(struct device *dev) /* Make sure we keep the same behaviour */ static inline int dma_mapping_error(struct device *dev, dma_addr_t dma_addr) { -#ifdef CONFIG_X86_64 struct dma_mapping_ops *ops = get_dma_ops(dev); if (ops->mapping_error) return ops->mapping_error(dev, dma_addr); -#endif return (dma_addr == bad_dma_address); } -- cgit v1.2.3 From 48d68b20d00865035b8b65e69af343d0f53fac9d Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Tue, 2 Dec 2008 00:20:39 +0100 Subject: tracing/function-graph-tracer: support for x86-64 Impact: extend and enable the function graph tracer to 64-bit x86 This patch implements the support for function graph tracer under x86-64. Both static and dynamic tracing are supported. This causes some small CPP conditional asm on arch/x86/kernel/ftrace.c I wanted to use probe_kernel_read/write to make the return address saving/patching code more generic but it causes tracing recursion. That would be perhaps useful to implement a notrace version of these function for other archs ports. Note that arch/x86/process_64.c is not traced, as in X86-32. I first thought __switch_to() was responsible of crashes during tracing because I believed current task were changed inside but that's actually not the case (actually yes, but not the "current" pointer). So I will have to investigate to find the functions that harm here, to enable tracing of the other functions inside (but there is no issue at this time, while process_64.c stays out of -pg flags). A little possible race condition is fixed inside this patch too. When the tracer allocate a return stack dynamically, the current depth is not initialized before but after. An interrupt could occur at this time and, after seeing that the return stack is allocated, the tracer could try to trace it with a random uninitialized depth. It's a prevention, even if I hadn't problems with it. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Tim Bird Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 2 +- arch/x86/kernel/Makefile | 1 + arch/x86/kernel/entry_64.S | 74 ++++++++++++++++++++++++++++++++++++++++++++++ arch/x86/kernel/ftrace.c | 11 ++++++- kernel/trace/ftrace.c | 4 ++- 5 files changed, 89 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 0842b1127684..45c86fb94132 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -29,7 +29,7 @@ config X86 select HAVE_FTRACE_MCOUNT_RECORD select HAVE_DYNAMIC_FTRACE select HAVE_FUNCTION_TRACER - select HAVE_FUNCTION_GRAPH_TRACER if X86_32 + select HAVE_FUNCTION_GRAPH_TRACER select HAVE_FUNCTION_TRACE_MCOUNT_TEST select HAVE_KVM if ((X86_32 && !X86_VOYAGER && !X86_VISWS && !X86_NUMAQ) || X86_64) select HAVE_ARCH_KGDB if !X86_VOYAGER diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index 64939a0c3986..d274425fb076 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -17,6 +17,7 @@ endif ifdef CONFIG_FUNCTION_GRAPH_TRACER # Don't trace __switch_to() but let it for function tracer CFLAGS_REMOVE_process_32.o = -pg +CFLAGS_REMOVE_process_64.o = -pg endif # diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 08aa6b10933c..2aa0526ac30e 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -98,6 +98,12 @@ ftrace_call: movq (%rsp), %rax addq $0x38, %rsp +#ifdef CONFIG_FUNCTION_GRAPH_TRACER +.globl ftrace_graph_call +ftrace_graph_call: + jmp ftrace_stub +#endif + .globl ftrace_stub ftrace_stub: retq @@ -110,6 +116,12 @@ ENTRY(mcount) cmpq $ftrace_stub, ftrace_trace_function jnz trace + +#ifdef CONFIG_FUNCTION_GRAPH_TRACER + cmpq $ftrace_stub, ftrace_graph_return + jnz ftrace_graph_caller +#endif + .globl ftrace_stub ftrace_stub: retq @@ -145,6 +157,68 @@ END(mcount) #endif /* CONFIG_DYNAMIC_FTRACE */ #endif /* CONFIG_FUNCTION_TRACER */ +#ifdef CONFIG_FUNCTION_GRAPH_TRACER +ENTRY(ftrace_graph_caller) + cmpl $0, function_trace_stop + jne ftrace_stub + + subq $0x38, %rsp + movq %rax, (%rsp) + movq %rcx, 8(%rsp) + movq %rdx, 16(%rsp) + movq %rsi, 24(%rsp) + movq %rdi, 32(%rsp) + movq %r8, 40(%rsp) + movq %r9, 48(%rsp) + + leaq 8(%rbp), %rdi + movq 0x38(%rsp), %rsi + + call prepare_ftrace_return + + movq 48(%rsp), %r9 + movq 40(%rsp), %r8 + movq 32(%rsp), %rdi + movq 24(%rsp), %rsi + movq 16(%rsp), %rdx + movq 8(%rsp), %rcx + movq (%rsp), %rax + addq $0x38, %rsp + retq +END(ftrace_graph_caller) + + +.globl return_to_handler +return_to_handler: + subq $80, %rsp + + movq %rax, (%rsp) + movq %rcx, 8(%rsp) + movq %rdx, 16(%rsp) + movq %rsi, 24(%rsp) + movq %rdi, 32(%rsp) + movq %r8, 40(%rsp) + movq %r9, 48(%rsp) + movq %r10, 56(%rsp) + movq %r11, 64(%rsp) + + call ftrace_return_to_handler + + movq %rax, 72(%rsp) + movq 64(%rsp), %r11 + movq 56(%rsp), %r10 + movq 48(%rsp), %r9 + movq 40(%rsp), %r8 + movq 32(%rsp), %rdi + movq 24(%rsp), %rsi + movq 16(%rsp), %rdx + movq 8(%rsp), %rcx + movq (%rsp), %rax + addq $72, %rsp + retq +#endif + + #ifndef CONFIG_PREEMPT #define retint_kernel retint_restore_args #endif diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 7ef914e6a2f6..58832478b94e 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -467,8 +467,13 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) * ignore such a protection. */ asm volatile( +#ifdef CONFIG_X86_64 + "1: movq (%[parent_old]), %[old]\n" + "2: movq %[return_hooker], (%[parent_replaced])\n" +#else "1: movl (%[parent_old]), %[old]\n" "2: movl %[return_hooker], (%[parent_replaced])\n" +#endif " movl $0, %[faulted]\n" ".section .fixup, \"ax\"\n" @@ -476,8 +481,13 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) ".previous\n" ".section __ex_table, \"a\"\n" +#ifdef CONFIG_X86_64 + " .quad 1b, 3b\n" + " .quad 2b, 3b\n" +#else " .long 1b, 3b\n" " .long 2b, 3b\n" +#endif ".previous\n" : [parent_replaced] "=r" (parent), [old] "=r" (old), @@ -509,5 +519,4 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) ftrace_graph_entry(&trace); } - #endif /* CONFIG_FUNCTION_GRAPH_TRACER */ diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 08b536a2614e..f72499627525 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -1673,8 +1673,10 @@ static int alloc_retstack_tasklist(struct ftrace_ret_stack **ret_stack_list) } if (t->ret_stack == NULL) { - t->ret_stack = ret_stack_list[start++]; t->curr_ret_stack = -1; + /* Make sure IRQs see the -1 first: */ + barrier(); + t->ret_stack = ret_stack_list[start++]; atomic_set(&t->trace_overrun, 0); } } while_each_thread(g, t); -- cgit v1.2.3 From 8daa19051e1c7369c89ace7b18e74fe1f55dfa29 Mon Sep 17 00:00:00 2001 From: Niels de Vos Date: Mon, 1 Dec 2008 14:13:53 -0800 Subject: x86, apm: remove CONFIG_APM_REAL_MODE_POWER_OFF in favor of a kernel parameter Remove CONFIG_APM_REAL_MODE_POWER_OFF like CONFIG_APM_POWER_OFF which has been done for linux-2.2.14pre8 (http://lkml.org/lkml/1999/11/23/3). Re-introducing CONFIG_APM_POWER_OFF got nack-ed. Stephen didn't bother to remove CONFIG_APM_REAL_MODE_POWER_OFF, let's get rid of it now. Reference: http://lkml.org/lkml/2008/5/7/97 Signed-off-by: Andrew Morton Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 7 ------- arch/x86/kernel/apm_32.c | 4 ---- 2 files changed, 11 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 4cf0ab13d187..ebcad15ccf35 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -1629,13 +1629,6 @@ config APM_ALLOW_INTS many of the newer IBM Thinkpads. If you experience hangs when you suspend, try setting this to Y. Otherwise, say N. -config APM_REAL_MODE_POWER_OFF - bool "Use real mode APM BIOS call to power off" - help - Use real mode APM BIOS calls to switch off the computer. This is - a work-around for a number of buggy BIOSes. Switch this option on if - your computer crashes instead of powering off properly. - endif # APM source "arch/x86/kernel/cpu/cpufreq/Kconfig" diff --git a/arch/x86/kernel/apm_32.c b/arch/x86/kernel/apm_32.c index 5145a6e72bbb..3a26525a3f31 100644 --- a/arch/x86/kernel/apm_32.c +++ b/arch/x86/kernel/apm_32.c @@ -391,11 +391,7 @@ static int power_off; #else static int power_off = 1; #endif -#ifdef CONFIG_APM_REAL_MODE_POWER_OFF -static int realmode_power_off = 1; -#else static int realmode_power_off; -#endif #ifdef CONFIG_APM_ALLOW_INTS static int allow_ints = 1; #else -- cgit v1.2.3 From dcb7731a185efbf3d800618d874af99895df5afb Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Tue, 2 Dec 2008 20:16:03 +0100 Subject: x86: fix broken flushing in GART nofullflush path Impact: remove stale IOTLB entries In the non-default nofullflush case the GART is only flushed when next_bit wraps around. But it can happen that an unmap operation unmaps memory which is behind the current next_bit location. If these addresses are reused it may result in stale GART IO/TLB entries. Fix this by setting the GART next_bit always behind an unmapped location. Signed-off-by: Joerg Roedel Signed-off-by: Ingo Molnar --- arch/x86/kernel/pci-gart_64.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/pci-gart_64.c b/arch/x86/kernel/pci-gart_64.c index a42b02b4df68..ba7ad83e20a8 100644 --- a/arch/x86/kernel/pci-gart_64.c +++ b/arch/x86/kernel/pci-gart_64.c @@ -123,6 +123,8 @@ static void free_iommu(unsigned long offset, int size) spin_lock_irqsave(&iommu_bitmap_lock, flags); iommu_area_free(iommu_gart_bitmap, offset, size); + if (offset >= next_bit) + next_bit = offset + size; spin_unlock_irqrestore(&iommu_bitmap_lock, flags); } -- cgit v1.2.3 From 181de82ee3ffda1175f89d50c991dae31b79280c Mon Sep 17 00:00:00 2001 From: FUJITA Tomonori Date: Wed, 3 Dec 2008 14:53:04 +0900 Subject: x86: remove dead BIO_VMERGE_BOUNDARY definition Impact: cleanup, remove dead code The block layer dropped the virtual merge feature (b8b3e16cfe6435d961f6aaebcfd52a1ff2a988c5). BIO_VMERGE_BOUNDARY definition is meaningless now. Signed-off-by: FUJITA Tomonori Acked-by: Jens Axboe Signed-off-by: Ingo Molnar --- arch/x86/include/asm/io.h | 2 -- arch/x86/include/asm/io_64.h | 2 -- arch/x86/kernel/pci-dma.c | 6 ------ 3 files changed, 10 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/io.h b/arch/x86/include/asm/io.h index 33513b9a67f3..05cfed4485fa 100644 --- a/arch/x86/include/asm/io.h +++ b/arch/x86/include/asm/io.h @@ -80,8 +80,6 @@ static inline void writeq(__u64 val, volatile void __iomem *addr) #define readq readq #define writeq writeq -extern int iommu_bio_merge; - #ifdef CONFIG_X86_32 # include "io_32.h" #else diff --git a/arch/x86/include/asm/io_64.h b/arch/x86/include/asm/io_64.h index fea325a1122f..563c16270ba6 100644 --- a/arch/x86/include/asm/io_64.h +++ b/arch/x86/include/asm/io_64.h @@ -232,8 +232,6 @@ void memset_io(volatile void __iomem *a, int b, size_t c); #define flush_write_buffers() -#define BIO_VMERGE_BOUNDARY iommu_bio_merge - /* * Convert a virtual cached pointer to an uncached pointer */ diff --git a/arch/x86/kernel/pci-dma.c b/arch/x86/kernel/pci-dma.c index 12eeb4bfcdeb..da93c65f8f0b 100644 --- a/arch/x86/kernel/pci-dma.c +++ b/arch/x86/kernel/pci-dma.c @@ -31,11 +31,6 @@ int no_iommu __read_mostly; /* Set this to 1 if there is a HW IOMMU in the system */ int iommu_detected __read_mostly = 0; -/* This tells the BIO block layer to assume merging. Default to off - because we cannot guarantee merging later. */ -int iommu_bio_merge __read_mostly = 0; -EXPORT_SYMBOL(iommu_bio_merge); - dma_addr_t bad_dma_address __read_mostly = 0; EXPORT_SYMBOL(bad_dma_address); @@ -189,7 +184,6 @@ static __init int iommu_setup(char *p) } if (!strncmp(p, "biomerge", 8)) { - iommu_bio_merge = 4096; iommu_merge = 1; force_iommu = 1; } -- cgit v1.2.3 From 347fdd9dd4e5d3f3a4e415925c35bdff1d59c3a9 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 2 Dec 2008 15:34:08 -0500 Subject: ftrace: clean up function graph asm Impact: clean up There exists macros for x86 asm to handle x86_64 and i386. This patch updates function graph asm to use them. Signed-off-by: Steven Rostedt Acked-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- arch/x86/kernel/ftrace.c | 20 ++++---------------- 1 file changed, 4 insertions(+), 16 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 58832478b94e..1a5b8f8cb3cc 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -467,28 +467,16 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) * ignore such a protection. */ asm volatile( -#ifdef CONFIG_X86_64 - "1: movq (%[parent_old]), %[old]\n" - "2: movq %[return_hooker], (%[parent_replaced])\n" -#else - "1: movl (%[parent_old]), %[old]\n" - "2: movl %[return_hooker], (%[parent_replaced])\n" -#endif + "1: " _ASM_MOV " (%[parent_old]), %[old]\n" + "2: " _ASM_MOV " %[return_hooker], (%[parent_replaced])\n" " movl $0, %[faulted]\n" ".section .fixup, \"ax\"\n" "3: movl $1, %[faulted]\n" ".previous\n" - ".section __ex_table, \"a\"\n" -#ifdef CONFIG_X86_64 - " .quad 1b, 3b\n" - " .quad 2b, 3b\n" -#else - " .long 1b, 3b\n" - " .long 2b, 3b\n" -#endif - ".previous\n" + _ASM_EXTABLE(1b, 3b) + _ASM_EXTABLE(2b, 3b) : [parent_replaced] "=r" (parent), [old] "=r" (old), [faulted] "=r" (faulted) -- cgit v1.2.3 From bb4304c71c97bf727ec43cd2f195c2c237c27fd3 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 2 Dec 2008 15:34:09 -0500 Subject: ftrace: have function graph use mcount caller address Impact: consistency change for function graph This patch makes function graph record the mcount caller address the same way the function tracer does. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_32.S | 1 + arch/x86/kernel/entry_64.S | 1 + 2 files changed, 2 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S index 958af86186c4..826682abed1d 100644 --- a/arch/x86/kernel/entry_32.S +++ b/arch/x86/kernel/entry_32.S @@ -1230,6 +1230,7 @@ ENTRY(ftrace_graph_caller) pushl %edx movl 0xc(%esp), %edx lea 0x4(%ebp), %eax + subl $MCOUNT_INSN_SIZE, %edx call prepare_ftrace_return popl %edx popl %ecx diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 2aa0526ac30e..9060ba6497e2 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -173,6 +173,7 @@ ENTRY(ftrace_graph_caller) leaq 8(%rbp), %rdi movq 0x38(%rsp), %rsi + subq $MCOUNT_INSN_SIZE, %rsi call prepare_ftrace_return -- cgit v1.2.3 From 14a866c567e040ccf6240d68b083dd1dbbde63e6 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 2 Dec 2008 23:50:02 -0500 Subject: ftrace: add ftrace_graph_stop() Impact: new ftrace_graph_stop function While developing more features of function graph, I hit a bug that caused the WARN_ON to trigger in the prepare_ftrace_return function. Well, it was hard for me to find out that was happening because the bug would not print, it would just cause a hard lockup or reboot. The reason is that it is not safe to call printk from this function. Looking further, I also found that it calls unregister_ftrace_graph, which grabs a mutex and calls kstop machine. This would definitely lock the box up if it were to trigger. This patch adds a fast and safe ftrace_graph_stop() which will stop the function tracer. Then it is safe to call the WARN ON. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/x86/kernel/ftrace.c | 10 ++++++---- include/linux/ftrace.h | 2 ++ kernel/trace/ftrace.c | 5 +++++ 3 files changed, 13 insertions(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 1a5b8f8cb3cc..adba8e9a427c 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -484,14 +484,16 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) : "memory" ); - if (WARN_ON(faulted)) { - unregister_ftrace_graph(); + if (unlikely(faulted)) { + ftrace_graph_stop(); + WARN_ON(1); return; } - if (WARN_ON(!__kernel_text_address(old))) { - unregister_ftrace_graph(); + if (unlikely(!__kernel_text_address(old))) { + ftrace_graph_stop(); *parent = old; + WARN_ON(1); return; } diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index afba918c623c..58ca1c3a3f4d 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -376,6 +376,8 @@ typedef void (*trace_func_graph_ent_t)(struct ftrace_graph_ent *); /* entry */ extern int register_ftrace_graph(trace_func_graph_ret_t retfunc, trace_func_graph_ent_t entryfunc); +extern void ftrace_graph_stop(void); + /* The current handlers in use */ extern trace_func_graph_ret_t ftrace_graph_return; extern trace_func_graph_ent_t ftrace_graph_entry; diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 2e78628443e8..a44af05ae2d0 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -1769,5 +1769,10 @@ void ftrace_graph_exit_task(struct task_struct *t) kfree(ret_stack); } + +void ftrace_graph_stop(void) +{ + ftrace_stop(); +} #endif -- cgit v1.2.3 From 7ee991fbc6f947e9b04f29c9c6c1d057d0671a16 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 2 Dec 2008 23:50:04 -0500 Subject: ftrace: print real return in dumpstack for function graph Impact: better dumpstack output I noticed in my crash dumps and even in the stack tracer that a lot of functions listed in the stack trace are simply return_to_handler which is ftrace graphs way to insert its own call into the return of a function. But we lose out where the actually function was called from. This patch adds in hooks to the dumpstack mechanism that detects this and finds the real function to print. Both are printed to let the user know that a hook is still in place. This does give a funny side effect in the stack tracer output: Depth Size Location (80 entries) ----- ---- -------- 0) 4144 48 save_stack_trace+0x2f/0x4d 1) 4096 128 ftrace_call+0x5/0x2b 2) 3968 16 mempool_alloc_slab+0x16/0x18 3) 3952 384 return_to_handler+0x0/0x73 4) 3568 -240 stack_trace_call+0x11d/0x209 5) 3808 144 return_to_handler+0x0/0x73 6) 3664 -128 mempool_alloc+0x4d/0xfe 7) 3792 128 return_to_handler+0x0/0x73 8) 3664 -32 scsi_sg_alloc+0x48/0x4a [scsi_mod] As you can see, the real functions are now negative. This is due to them not being found inside the stack. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/x86/kernel/dumpstack.c | 34 +++++++++++++++++++++++++++++++++- arch/x86/kernel/dumpstack.h | 2 +- arch/x86/kernel/dumpstack_32.c | 5 ++++- arch/x86/kernel/dumpstack_64.c | 7 ++++--- 4 files changed, 42 insertions(+), 6 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/dumpstack.c b/arch/x86/kernel/dumpstack.c index 5962176dfabb..6b1f6f6f8661 100644 --- a/arch/x86/kernel/dumpstack.c +++ b/arch/x86/kernel/dumpstack.c @@ -30,6 +30,37 @@ void printk_address(unsigned long address, int reliable) reliable ? "" : "? ", (void *) address); } +#ifdef CONFIG_FUNCTION_GRAPH_TRACER +static void +print_ftrace_graph_addr(unsigned long addr, void *data, + const struct stacktrace_ops *ops, + struct thread_info *tinfo, int *graph) +{ + struct task_struct *task = tinfo->task; + unsigned long ret_addr; + int index = task->curr_ret_stack; + + if (addr != (unsigned long)return_to_handler) + return; + + if (!task->ret_stack || index < *graph) + return; + + index -= *graph; + ret_addr = task->ret_stack[index].ret; + + ops->address(data, ret_addr, 1); + + (*graph)++; +} +#else +static inline void +print_ftrace_graph_addr(unsigned long addr, void *data, + const struct stacktrace_ops *ops, + struct thread_info *tinfo, int *graph) +{ } +#endif + /* * x86-64 can have up to three kernel stacks: * process stack @@ -54,7 +85,7 @@ unsigned long print_context_stack(struct thread_info *tinfo, unsigned long *stack, unsigned long bp, const struct stacktrace_ops *ops, void *data, - unsigned long *end) + unsigned long *end, int *graph) { struct stack_frame *frame = (struct stack_frame *)bp; @@ -70,6 +101,7 @@ print_context_stack(struct thread_info *tinfo, } else { ops->address(data, addr, bp == 0); } + print_ftrace_graph_addr(addr, data, ops, tinfo, graph); } stack++; } diff --git a/arch/x86/kernel/dumpstack.h b/arch/x86/kernel/dumpstack.h index 3119a801c32b..da87590b8698 100644 --- a/arch/x86/kernel/dumpstack.h +++ b/arch/x86/kernel/dumpstack.h @@ -18,7 +18,7 @@ extern unsigned long print_context_stack(struct thread_info *tinfo, unsigned long *stack, unsigned long bp, const struct stacktrace_ops *ops, void *data, - unsigned long *end); + unsigned long *end, int *graph); extern void show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs, diff --git a/arch/x86/kernel/dumpstack_32.c b/arch/x86/kernel/dumpstack_32.c index 7b031b106ec8..d593cd1f58dc 100644 --- a/arch/x86/kernel/dumpstack_32.c +++ b/arch/x86/kernel/dumpstack_32.c @@ -23,6 +23,8 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs, unsigned long *stack, unsigned long bp, const struct stacktrace_ops *ops, void *data) { + int graph = 0; + if (!task) task = current; @@ -50,7 +52,8 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs, context = (struct thread_info *) ((unsigned long)stack & (~(THREAD_SIZE - 1))); - bp = print_context_stack(context, stack, bp, ops, data, NULL); + bp = print_context_stack(context, stack, bp, ops, + data, NULL, &graph); stack = (unsigned long *)context->previous_esp; if (!stack) diff --git a/arch/x86/kernel/dumpstack_64.c b/arch/x86/kernel/dumpstack_64.c index 33ff10287a5d..c302d0707048 100644 --- a/arch/x86/kernel/dumpstack_64.c +++ b/arch/x86/kernel/dumpstack_64.c @@ -109,6 +109,7 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs, unsigned long *irqstack_end = (unsigned long *)cpu_pda(cpu)->irqstackptr; unsigned used = 0; struct thread_info *tinfo; + int graph = 0; if (!task) task = current; @@ -149,7 +150,7 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs, break; bp = print_context_stack(tinfo, stack, bp, ops, - data, estack_end); + data, estack_end, &graph); ops->stack(data, ""); /* * We link to the next stack via the @@ -168,7 +169,7 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs, if (ops->stack(data, "IRQ") < 0) break; bp = print_context_stack(tinfo, stack, bp, - ops, data, irqstack_end); + ops, data, irqstack_end, &graph); /* * We link to the next stack (which would be * the process stack normally) the last @@ -186,7 +187,7 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs, /* * This handles the process stack: */ - bp = print_context_stack(tinfo, stack, bp, ops, data, NULL); + bp = print_context_stack(tinfo, stack, bp, ops, data, NULL, &graph); put_cpu(); } EXPORT_SYMBOL(dump_trace); -- cgit v1.2.3 From e49dc19c6a19ea112fcb94b7c62ec62cdd5c08aa Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 2 Dec 2008 23:50:05 -0500 Subject: ftrace: function graph return for function entry Impact: feature, let entry function decide to trace or not This patch lets the graph tracer entry function decide if the tracing should be done at the end as well. This requires all function graph entry functions return 1 if it should trace, or 0 if the return should not be traced. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/x86/kernel/entry_32.S | 3 +++ arch/x86/kernel/entry_64.S | 3 +++ arch/x86/kernel/ftrace.c | 7 ++++++- include/linux/ftrace.h | 2 +- kernel/trace/ftrace.c | 10 +++++++--- kernel/trace/trace.c | 4 +++- kernel/trace/trace.h | 2 +- 7 files changed, 24 insertions(+), 7 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S index 826682abed1d..43ceb3f454bf 100644 --- a/arch/x86/kernel/entry_32.S +++ b/arch/x86/kernel/entry_32.S @@ -1196,6 +1196,9 @@ ENTRY(mcount) #ifdef CONFIG_FUNCTION_GRAPH_TRACER cmpl $ftrace_stub, ftrace_graph_return jnz ftrace_graph_caller + + cmpl $ftrace_graph_entry_stub, ftrace_graph_entry + jnz ftrace_graph_caller #endif .globl ftrace_stub ftrace_stub: diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 9060ba6497e2..54e0bbdccb99 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -120,6 +120,9 @@ ENTRY(mcount) #ifdef CONFIG_FUNCTION_GRAPH_TRACER cmpq $ftrace_stub, ftrace_graph_return jnz ftrace_graph_caller + + cmpq $ftrace_graph_entry_stub, ftrace_graph_entry + jnz ftrace_graph_caller #endif .globl ftrace_stub diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index adba8e9a427c..d278ad2ebda2 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -425,6 +425,7 @@ static void pop_return_trace(struct ftrace_graph_ret *trace, unsigned long *ret) trace->calltime = current->ret_stack[index].calltime; trace->overrun = atomic_read(¤t->trace_overrun); trace->depth = index; + barrier(); current->curr_ret_stack--; } @@ -506,7 +507,11 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) } trace.func = self_addr; - ftrace_graph_entry(&trace); + /* Only trace if the calling function expects to */ + if (!ftrace_graph_entry(&trace)) { + current->curr_ret_stack--; + *parent = old; + } } #endif /* CONFIG_FUNCTION_GRAPH_TRACER */ diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 58ca1c3a3f4d..469ceb3e85ba 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -371,7 +371,7 @@ struct ftrace_graph_ret { #define FTRACE_RETSTACK_ALLOC_SIZE 32 /* Type of the callback handlers for tracing function graph*/ typedef void (*trace_func_graph_ret_t)(struct ftrace_graph_ret *); /* return */ -typedef void (*trace_func_graph_ent_t)(struct ftrace_graph_ent *); /* entry */ +typedef int (*trace_func_graph_ent_t)(struct ftrace_graph_ent *); /* entry */ extern int register_ftrace_graph(trace_func_graph_ret_t retfunc, trace_func_graph_ent_t entryfunc); diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index a44af05ae2d0..65b9e863056b 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -1636,11 +1636,15 @@ ftrace_enable_sysctl(struct ctl_table *table, int write, static atomic_t ftrace_graph_active; +int ftrace_graph_entry_stub(struct ftrace_graph_ent *trace) +{ + return 0; +} + /* The callbacks that hook a function */ trace_func_graph_ret_t ftrace_graph_return = (trace_func_graph_ret_t)ftrace_stub; -trace_func_graph_ent_t ftrace_graph_entry = - (trace_func_graph_ent_t)ftrace_stub; +trace_func_graph_ent_t ftrace_graph_entry = ftrace_graph_entry_stub; /* Try to assign a return stack array on FTRACE_RETSTACK_ALLOC_SIZE tasks. */ static int alloc_retstack_tasklist(struct ftrace_ret_stack **ret_stack_list) @@ -1738,7 +1742,7 @@ void unregister_ftrace_graph(void) atomic_dec(&ftrace_graph_active); ftrace_graph_return = (trace_func_graph_ret_t)ftrace_stub; - ftrace_graph_entry = (trace_func_graph_ent_t)ftrace_stub; + ftrace_graph_entry = ftrace_graph_entry_stub; ftrace_shutdown(FTRACE_STOP_FUNC_RET); mutex_unlock(&ftrace_sysctl_lock); diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 380de630ebce..8b6409a62b54 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1200,7 +1200,7 @@ function_trace_call(unsigned long ip, unsigned long parent_ip) } #ifdef CONFIG_FUNCTION_GRAPH_TRACER -void trace_graph_entry(struct ftrace_graph_ent *trace) +int trace_graph_entry(struct ftrace_graph_ent *trace) { struct trace_array *tr = &global_trace; struct trace_array_cpu *data; @@ -1219,6 +1219,8 @@ void trace_graph_entry(struct ftrace_graph_ent *trace) } atomic_dec(&data->disabled); local_irq_restore(flags); + + return 1; } void trace_graph_return(struct ftrace_graph_ret *trace) diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index f96f4e787ff3..0565ae9a2210 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -412,7 +412,7 @@ void trace_function(struct trace_array *tr, unsigned long flags, int pc); void trace_graph_return(struct ftrace_graph_ret *trace); -void trace_graph_entry(struct ftrace_graph_ent *trace); +int trace_graph_entry(struct ftrace_graph_ent *trace); void trace_bts(struct trace_array *tr, unsigned long from, unsigned long to); -- cgit v1.2.3 From 62679efe0a5f02987a621942afc5979a80a6ca5a Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 2 Dec 2008 23:50:06 -0500 Subject: ftrace: add checks on ret stack in function graph Import: robustness checks Add more checks in the function graph code to detect errors and perhaps print out better information if a bug happens. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/x86/kernel/ftrace.c | 17 +++++++++++++++++ 1 file changed, 17 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index d278ad2ebda2..f98c4076a170 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -420,6 +420,15 @@ static void pop_return_trace(struct ftrace_graph_ret *trace, unsigned long *ret) int index; index = current->curr_ret_stack; + + if (unlikely(index < 0)) { + ftrace_graph_stop(); + WARN_ON(1); + /* Might as well panic, otherwise we have no where to go */ + *ret = (unsigned long)panic; + return; + } + *ret = current->ret_stack[index].ret; trace->func = current->ret_stack[index].func; trace->calltime = current->ret_stack[index].calltime; @@ -427,6 +436,7 @@ static void pop_return_trace(struct ftrace_graph_ret *trace, unsigned long *ret) trace->depth = index; barrier(); current->curr_ret_stack--; + } /* @@ -442,6 +452,13 @@ unsigned long ftrace_return_to_handler(void) trace.rettime = cpu_clock(raw_smp_processor_id()); ftrace_graph_return(&trace); + if (unlikely(!ret)) { + ftrace_graph_stop(); + WARN_ON(1); + /* Might as well panic. What else to do? */ + ret = (unsigned long)panic; + } + return ret; } -- cgit v1.2.3 From affa219b60a11b3295637a97f5b1b8ef231490fc Mon Sep 17 00:00:00 2001 From: Joe Korty Date: Wed, 3 Dec 2008 18:58:19 -0500 Subject: x86: change thread_info's flag field back to 32 bits Impact: pack struct thread_info more tightly Change x86_64's thread_info 'flags' field back to __u32. This was changed to 'unsigned long' when the thread_info*.h for i386 and x86_64 were merged. Change it back. We can do this as only 27 bits of 'flags' are actually used. This change actually packs down thread_info by 64 bits: 32 bits are saved by the smaller flags, and 32 bits are saved by the following 'mm_segment_t field' becoming naturally 64-bit aligned. Signed-off-by: Joe Korty Signed-off-by: Ingo Molnar --- arch/x86/include/asm/thread_info.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/thread_info.h b/arch/x86/include/asm/thread_info.h index e44d379faad2..8dbc57390d25 100644 --- a/arch/x86/include/asm/thread_info.h +++ b/arch/x86/include/asm/thread_info.h @@ -24,7 +24,7 @@ struct exec_domain; struct thread_info { struct task_struct *task; /* main task structure */ struct exec_domain *exec_domain; /* execution domain */ - unsigned long flags; /* low level flags */ + __u32 flags; /* low level flags */ __u32 status; /* thread synchronous flags */ __u32 cpu; /* current CPU */ int preempt_count; /* 0 => preemptable, -- cgit v1.2.3 From 55c395b47042e12d5c25aa07f271f56ffe44f793 Mon Sep 17 00:00:00 2001 From: Michael Tokarev Date: Fri, 5 Dec 2008 14:42:20 +0300 Subject: x86: fix missing space in printk Just come across this when booting on an old hw.. Looks somewhat ugly, that single missing space ;) Signed-off-by: Michael Tokarev Signed-off-by: Ingo Molnar --- arch/x86/kernel/smpboot.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index 7b1093397319..1a3c3253f0ed 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -1086,8 +1086,10 @@ static int __init smp_sanity_check(unsigned max_cpus) #endif if (!physid_isset(hard_smp_processor_id(), phys_cpu_present_map)) { - printk(KERN_WARNING "weird, boot CPU (#%d) not listed" - "by the BIOS.\n", hard_smp_processor_id()); + printk(KERN_WARNING + "weird, boot CPU (#%d) not listed by the BIOS.\n", + hard_smp_processor_id()); + physid_set(hard_smp_processor_id(), phys_cpu_present_map); } -- cgit v1.2.3 From a0286c94f07636380082608196d41dd725a83229 Mon Sep 17 00:00:00 2001 From: Michael Tokarev Date: Fri, 5 Dec 2008 15:47:29 +0300 Subject: x86: fix missing space in printk, #2 Impact: clean up printk Signed-off-by: Michael Tokarev Signed-off-by: Ingo Molnar --- arch/x86/kernel/pci-dma.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/pci-dma.c b/arch/x86/kernel/pci-dma.c index 192624820217..dc572994703d 100644 --- a/arch/x86/kernel/pci-dma.c +++ b/arch/x86/kernel/pci-dma.c @@ -300,8 +300,8 @@ fs_initcall(pci_iommu_init); static __devinit void via_no_dac(struct pci_dev *dev) { if ((dev->class >> 8) == PCI_CLASS_BRIDGE_PCI && forbid_dac == 0) { - printk(KERN_INFO "PCI: VIA PCI bridge detected." - "Disabling DAC.\n"); + printk(KERN_INFO + "PCI: VIA PCI bridge detected. Disabling DAC.\n"); forbid_dac = 1; } } -- cgit v1.2.3 From 3e1e9002aa8b32bd4c95ac6c8fad376b7a8127fb Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Mon, 8 Dec 2008 00:50:22 +0100 Subject: x86: change static allocation of trampoline area Impact: fix trampoline sizing bug, save space While debugging a suspend-to-RAM related issue it occured to me that if the trampoline code had grown past 4 KB, we would have been allocating too little memory for it, since the 4 KB size of the trampoline is hardcoded into arch/x86/kernel/e820.c . Change that by making the kernel compute the trampoline size and allocate as much memory as necessary. Signed-off-by: Rafael J. Wysocki Signed-off-by: Ingo Molnar --- arch/x86/include/asm/trampoline.h | 7 +++++++ arch/x86/kernel/e820.c | 16 ---------------- arch/x86/kernel/head32.c | 3 +++ arch/x86/kernel/head64.c | 3 +++ arch/x86/kernel/trampoline.c | 19 +++++++++++++++++-- 5 files changed, 30 insertions(+), 18 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/trampoline.h b/arch/x86/include/asm/trampoline.h index fa0d79facdbc..780ba0ab94f9 100644 --- a/arch/x86/include/asm/trampoline.h +++ b/arch/x86/include/asm/trampoline.h @@ -3,6 +3,7 @@ #ifndef __ASSEMBLY__ +#ifdef CONFIG_X86_TRAMPOLINE /* * Trampoline 80x86 program as an array. */ @@ -13,8 +14,14 @@ extern unsigned char *trampoline_base; extern unsigned long init_rsp; extern unsigned long initial_code; +#define TRAMPOLINE_SIZE roundup(trampoline_end - trampoline_data, PAGE_SIZE) #define TRAMPOLINE_BASE 0x6000 + extern unsigned long setup_trampoline(void); +extern void __init reserve_trampoline_memory(void); +#else +static inline void reserve_trampoline_memory(void) {}; +#endif /* CONFIG_X86_TRAMPOLINE */ #endif /* __ASSEMBLY__ */ diff --git a/arch/x86/kernel/e820.c b/arch/x86/kernel/e820.c index 7aafeb5263ef..65a13943e098 100644 --- a/arch/x86/kernel/e820.c +++ b/arch/x86/kernel/e820.c @@ -677,22 +677,6 @@ struct early_res { }; static struct early_res early_res[MAX_EARLY_RES] __initdata = { { 0, PAGE_SIZE, "BIOS data page" }, /* BIOS data page */ -#if defined(CONFIG_X86_64) && defined(CONFIG_X86_TRAMPOLINE) - { TRAMPOLINE_BASE, TRAMPOLINE_BASE + 2 * PAGE_SIZE, "TRAMPOLINE" }, -#endif -#if defined(CONFIG_X86_32) && defined(CONFIG_SMP) - /* - * But first pinch a few for the stack/trampoline stuff - * FIXME: Don't need the extra page at 4K, but need to fix - * trampoline before removing it. (see the GDT stuff) - */ - { PAGE_SIZE, PAGE_SIZE + PAGE_SIZE, "EX TRAMPOLINE" }, - /* - * Has to be in very low memory so we can execute - * real-mode AP code. - */ - { TRAMPOLINE_BASE, TRAMPOLINE_BASE + PAGE_SIZE, "TRAMPOLINE" }, -#endif {} }; diff --git a/arch/x86/kernel/head32.c b/arch/x86/kernel/head32.c index fa1d25dd83e3..ac108d1fe182 100644 --- a/arch/x86/kernel/head32.c +++ b/arch/x86/kernel/head32.c @@ -12,9 +12,12 @@ #include #include #include +#include void __init i386_start_kernel(void) { + reserve_trampoline_memory(); + reserve_early(__pa_symbol(&_text), __pa_symbol(&_end), "TEXT DATA BSS"); #ifdef CONFIG_BLK_DEV_INITRD diff --git a/arch/x86/kernel/head64.c b/arch/x86/kernel/head64.c index d16084f90649..388e05a5fc17 100644 --- a/arch/x86/kernel/head64.c +++ b/arch/x86/kernel/head64.c @@ -24,6 +24,7 @@ #include #include #include +#include /* boot cpu pda */ static struct x8664_pda _boot_cpu_pda __read_mostly; @@ -120,6 +121,8 @@ void __init x86_64_start_reservations(char *real_mode_data) { copy_bootdata(__va(real_mode_data)); + reserve_trampoline_memory(); + reserve_early(__pa_symbol(&_text), __pa_symbol(&_end), "TEXT DATA BSS"); #ifdef CONFIG_BLK_DEV_INITRD diff --git a/arch/x86/kernel/trampoline.c b/arch/x86/kernel/trampoline.c index 1106fac6024d..808031a5ba19 100644 --- a/arch/x86/kernel/trampoline.c +++ b/arch/x86/kernel/trampoline.c @@ -1,10 +1,26 @@ #include #include +#include /* ready for x86_64 and x86 */ unsigned char *trampoline_base = __va(TRAMPOLINE_BASE); +void __init reserve_trampoline_memory(void) +{ +#ifdef CONFIG_X86_32 + /* + * But first pinch a few for the stack/trampoline stuff + * FIXME: Don't need the extra page at 4K, but need to fix + * trampoline before removing it. (see the GDT stuff) + */ + reserve_early(PAGE_SIZE, PAGE_SIZE + PAGE_SIZE, "EX TRAMPOLINE"); +#endif + /* Has to be in very low memory so we can execute real-mode AP code. */ + reserve_early(TRAMPOLINE_BASE, TRAMPOLINE_BASE + TRAMPOLINE_SIZE, + "TRAMPOLINE"); +} + /* * Currently trivial. Write the real->protected mode * bootstrap into the page concerned. The caller @@ -12,7 +28,6 @@ unsigned char *trampoline_base = __va(TRAMPOLINE_BASE); */ unsigned long setup_trampoline(void) { - memcpy(trampoline_base, trampoline_data, - trampoline_end - trampoline_data); + memcpy(trampoline_base, trampoline_data, TRAMPOLINE_SIZE); return virt_to_phys(trampoline_base); } -- cgit v1.2.3 From 0b8f1efad30bd58f89961b82dfe68b9edf8fd2ac Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Fri, 5 Dec 2008 18:58:31 -0800 Subject: sparse irq_desc[] array: core kernel and x86 changes Impact: new feature Problem on distro kernels: irq_desc[NR_IRQS] takes megabytes of RAM with NR_CPUS set to large values. The goal is to be able to scale up to much larger NR_IRQS value without impacting the (important) common case. To solve this, we generalize irq_desc[NR_IRQS] to an (optional) array of irq_desc pointers. When CONFIG_SPARSE_IRQ=y is used, we use kzalloc_node to get irq_desc, this also makes the IRQ descriptors NUMA-local (to the site that calls request_irq()). This gets rid of the irq_cfg[] static array on x86 as well: irq_cfg now uses desc->chip_data for x86 to store irq_cfg. Signed-off-by: Yinghai Lu Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 10 ++ arch/x86/include/asm/irq_vectors.h | 9 ++ arch/x86/kernel/io_apic.c | 301 +++++++++++++++++++++++-------------- arch/x86/kernel/irq.c | 3 + arch/x86/kernel/irq_32.c | 2 + arch/x86/kernel/irq_64.c | 2 + arch/x86/kernel/irqinit_32.c | 1 - arch/x86/kernel/irqinit_64.c | 1 - drivers/char/random.c | 22 +-- drivers/pci/intr_remapping.c | 76 +++++++++- drivers/xen/events.c | 12 +- fs/proc/stat.c | 17 ++- include/linux/interrupt.h | 2 + include/linux/irq.h | 54 ++++++- include/linux/irqnr.h | 14 +- include/linux/kernel_stat.h | 14 +- include/linux/random.h | 51 +++++++ init/main.c | 11 ++ kernel/irq/autoprobe.c | 15 ++ kernel/irq/chip.c | 3 +- kernel/irq/handle.c | 181 +++++++++++++++++++++- kernel/irq/proc.c | 6 +- kernel/irq/spurious.c | 5 + 23 files changed, 649 insertions(+), 163 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index ac22bb7719f7..48ac688de3cd 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -238,6 +238,16 @@ config X86_HAS_BOOT_CPU_ID def_bool y depends on X86_VOYAGER +config SPARSE_IRQ + bool "Support sparse irq numbering" + depends on PCI_MSI || HT_IRQ + default y + help + This enables support for sparse irq, esp for msi/msi-x. You may need + if you have lots of cards supports msi-x installed. + + If you don't know what to do here, say Y. + config X86_FIND_SMP_CONFIG def_bool y depends on X86_MPPARSE || X86_VOYAGER diff --git a/arch/x86/include/asm/irq_vectors.h b/arch/x86/include/asm/irq_vectors.h index 0005adb0f941..bb6b69a6b125 100644 --- a/arch/x86/include/asm/irq_vectors.h +++ b/arch/x86/include/asm/irq_vectors.h @@ -102,11 +102,20 @@ #define invalid_vm86_irq(irq) ((irq) < 3 || (irq) > 15) #if defined(CONFIG_X86_IO_APIC) && !defined(CONFIG_X86_VOYAGER) + +#ifndef CONFIG_SPARSE_IRQ # if NR_CPUS < MAX_IO_APICS # define NR_IRQS (NR_VECTORS + (32 * NR_CPUS)) # else # define NR_IRQS (NR_VECTORS + (32 * MAX_IO_APICS)) # endif +#else +# if (8 * NR_CPUS) > (32 * MAX_IO_APICS) +# define NR_IRQS (NR_VECTORS + (8 * NR_CPUS)) +# else +# define NR_IRQS (NR_VECTORS + (32 * MAX_IO_APICS)) +# endif +#endif #elif defined(CONFIG_X86_VOYAGER) diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index 9043251210fb..9de17f5c1125 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -108,8 +108,33 @@ static int __init parse_noapic(char *str) early_param("noapic", parse_noapic); struct irq_pin_list; + +/* + * This is performance-critical, we want to do it O(1) + * + * the indexing order of this array favors 1:1 mappings + * between pins and IRQs. + */ + +struct irq_pin_list { + int apic, pin; + struct irq_pin_list *next; +}; + +static struct irq_pin_list *get_one_free_irq_2_pin(int cpu) +{ + struct irq_pin_list *pin; + int node; + + node = cpu_to_node(cpu); + + pin = kzalloc_node(sizeof(*pin), GFP_ATOMIC, node); + printk(KERN_DEBUG " alloc irq_2_pin on cpu %d node %d\n", cpu, node); + + return pin; +} + struct irq_cfg { - unsigned int irq; struct irq_pin_list *irq_2_pin; cpumask_t domain; cpumask_t old_domain; @@ -119,83 +144,93 @@ struct irq_cfg { }; /* irq_cfg is indexed by the sum of all RTEs in all I/O APICs. */ +#ifdef CONFIG_SPARSE_IRQ +static struct irq_cfg irq_cfgx[] = { +#else static struct irq_cfg irq_cfgx[NR_IRQS] = { - [0] = { .irq = 0, .domain = CPU_MASK_ALL, .vector = IRQ0_VECTOR, }, - [1] = { .irq = 1, .domain = CPU_MASK_ALL, .vector = IRQ1_VECTOR, }, - [2] = { .irq = 2, .domain = CPU_MASK_ALL, .vector = IRQ2_VECTOR, }, - [3] = { .irq = 3, .domain = CPU_MASK_ALL, .vector = IRQ3_VECTOR, }, - [4] = { .irq = 4, .domain = CPU_MASK_ALL, .vector = IRQ4_VECTOR, }, - [5] = { .irq = 5, .domain = CPU_MASK_ALL, .vector = IRQ5_VECTOR, }, - [6] = { .irq = 6, .domain = CPU_MASK_ALL, .vector = IRQ6_VECTOR, }, - [7] = { .irq = 7, .domain = CPU_MASK_ALL, .vector = IRQ7_VECTOR, }, - [8] = { .irq = 8, .domain = CPU_MASK_ALL, .vector = IRQ8_VECTOR, }, - [9] = { .irq = 9, .domain = CPU_MASK_ALL, .vector = IRQ9_VECTOR, }, - [10] = { .irq = 10, .domain = CPU_MASK_ALL, .vector = IRQ10_VECTOR, }, - [11] = { .irq = 11, .domain = CPU_MASK_ALL, .vector = IRQ11_VECTOR, }, - [12] = { .irq = 12, .domain = CPU_MASK_ALL, .vector = IRQ12_VECTOR, }, - [13] = { .irq = 13, .domain = CPU_MASK_ALL, .vector = IRQ13_VECTOR, }, - [14] = { .irq = 14, .domain = CPU_MASK_ALL, .vector = IRQ14_VECTOR, }, - [15] = { .irq = 15, .domain = CPU_MASK_ALL, .vector = IRQ15_VECTOR, }, +#endif + [0] = { .domain = CPU_MASK_ALL, .vector = IRQ0_VECTOR, }, + [1] = { .domain = CPU_MASK_ALL, .vector = IRQ1_VECTOR, }, + [2] = { .domain = CPU_MASK_ALL, .vector = IRQ2_VECTOR, }, + [3] = { .domain = CPU_MASK_ALL, .vector = IRQ3_VECTOR, }, + [4] = { .domain = CPU_MASK_ALL, .vector = IRQ4_VECTOR, }, + [5] = { .domain = CPU_MASK_ALL, .vector = IRQ5_VECTOR, }, + [6] = { .domain = CPU_MASK_ALL, .vector = IRQ6_VECTOR, }, + [7] = { .domain = CPU_MASK_ALL, .vector = IRQ7_VECTOR, }, + [8] = { .domain = CPU_MASK_ALL, .vector = IRQ8_VECTOR, }, + [9] = { .domain = CPU_MASK_ALL, .vector = IRQ9_VECTOR, }, + [10] = { .domain = CPU_MASK_ALL, .vector = IRQ10_VECTOR, }, + [11] = { .domain = CPU_MASK_ALL, .vector = IRQ11_VECTOR, }, + [12] = { .domain = CPU_MASK_ALL, .vector = IRQ12_VECTOR, }, + [13] = { .domain = CPU_MASK_ALL, .vector = IRQ13_VECTOR, }, + [14] = { .domain = CPU_MASK_ALL, .vector = IRQ14_VECTOR, }, + [15] = { .domain = CPU_MASK_ALL, .vector = IRQ15_VECTOR, }, }; -#define for_each_irq_cfg(irq, cfg) \ - for (irq = 0, cfg = irq_cfgx; irq < nr_irqs; irq++, cfg++) - -static struct irq_cfg *irq_cfg(unsigned int irq) +void __init arch_early_irq_init(void) { - return irq < nr_irqs ? irq_cfgx + irq : NULL; -} + struct irq_cfg *cfg; + struct irq_desc *desc; + int count; + int i; -static struct irq_cfg *irq_cfg_alloc(unsigned int irq) -{ - return irq_cfg(irq); -} + cfg = irq_cfgx; + count = ARRAY_SIZE(irq_cfgx); -/* - * Rough estimation of how many shared IRQs there are, can be changed - * anytime. - */ -#define MAX_PLUS_SHARED_IRQS NR_IRQS -#define PIN_MAP_SIZE (MAX_PLUS_SHARED_IRQS + NR_IRQS) + for (i = 0; i < count; i++) { + desc = irq_to_desc(i); + desc->chip_data = &cfg[i]; + } +} -/* - * This is performance-critical, we want to do it O(1) - * - * the indexing order of this array favors 1:1 mappings - * between pins and IRQs. - */ +#ifdef CONFIG_SPARSE_IRQ +static struct irq_cfg *irq_cfg(unsigned int irq) +{ + struct irq_cfg *cfg = NULL; + struct irq_desc *desc; -struct irq_pin_list { - int apic, pin; - struct irq_pin_list *next; -}; + desc = irq_to_desc(irq); + if (desc) + cfg = desc->chip_data; -static struct irq_pin_list irq_2_pin_head[PIN_MAP_SIZE]; -static struct irq_pin_list *irq_2_pin_ptr; + return cfg; +} -static void __init irq_2_pin_init(void) +static struct irq_cfg *get_one_free_irq_cfg(int cpu) { - struct irq_pin_list *pin = irq_2_pin_head; - int i; + struct irq_cfg *cfg; + int node; + + node = cpu_to_node(cpu); - for (i = 1; i < PIN_MAP_SIZE; i++) - pin[i-1].next = &pin[i]; + cfg = kzalloc_node(sizeof(*cfg), GFP_ATOMIC, node); + printk(KERN_DEBUG " alloc irq_cfg on cpu %d node %d\n", cpu, node); - irq_2_pin_ptr = &pin[0]; + return cfg; } -static struct irq_pin_list *get_one_free_irq_2_pin(void) +void arch_init_chip_data(struct irq_desc *desc, int cpu) { - struct irq_pin_list *pin = irq_2_pin_ptr; + struct irq_cfg *cfg; - if (!pin) - panic("can not get more irq_2_pin\n"); + cfg = desc->chip_data; + if (!cfg) { + desc->chip_data = get_one_free_irq_cfg(cpu); + if (!desc->chip_data) { + printk(KERN_ERR "can not alloc irq_cfg\n"); + BUG_ON(1); + } + } +} - irq_2_pin_ptr = pin->next; - pin->next = NULL; - return pin; +#else +static struct irq_cfg *irq_cfg(unsigned int irq) +{ + return irq < nr_irqs ? irq_cfgx + irq : NULL; } +#endif + struct io_apic { unsigned int index; unsigned int unused[3]; @@ -397,16 +432,19 @@ static void set_ioapic_affinity_irq(unsigned int irq, cpumask_t mask) * shared ISA-space IRQs, so we have to support them. We are super * fast in the common case, and fast for shared ISA-space IRQs. */ -static void add_pin_to_irq(unsigned int irq, int apic, int pin) +static void add_pin_to_irq_cpu(unsigned int irq, int cpu, int apic, int pin) { - struct irq_cfg *cfg; struct irq_pin_list *entry; + struct irq_cfg *cfg = irq_cfg(irq); - /* first time to refer irq_cfg, so with new */ - cfg = irq_cfg_alloc(irq); entry = cfg->irq_2_pin; if (!entry) { - entry = get_one_free_irq_2_pin(); + entry = get_one_free_irq_2_pin(cpu); + if (!entry) { + printk(KERN_ERR "can not alloc irq_2_pin to add %d - %d\n", + apic, pin); + return; + } cfg->irq_2_pin = entry; entry->apic = apic; entry->pin = pin; @@ -421,7 +459,7 @@ static void add_pin_to_irq(unsigned int irq, int apic, int pin) entry = entry->next; } - entry->next = get_one_free_irq_2_pin(); + entry->next = get_one_free_irq_2_pin(cpu); entry = entry->next; entry->apic = apic; entry->pin = pin; @@ -430,7 +468,7 @@ static void add_pin_to_irq(unsigned int irq, int apic, int pin) /* * Reroute an IRQ to a different pin. */ -static void __init replace_pin_at_irq(unsigned int irq, +static void __init replace_pin_at_irq(unsigned int irq, int cpu, int oldapic, int oldpin, int newapic, int newpin) { @@ -451,7 +489,7 @@ static void __init replace_pin_at_irq(unsigned int irq, /* why? call replace before add? */ if (!replaced) - add_pin_to_irq(irq, newapic, newpin); + add_pin_to_irq_cpu(irq, cpu, newapic, newpin); } static inline void io_apic_modify_irq(unsigned int irq, @@ -1162,9 +1200,13 @@ void __setup_vector_irq(int cpu) /* This function must be called with vector_lock held */ int irq, vector; struct irq_cfg *cfg; + struct irq_desc *desc; /* Mark the inuse vectors */ - for_each_irq_cfg(irq, cfg) { + for_each_irq_desc(irq, desc) { + if (!desc) + continue; + cfg = desc->chip_data; if (!cpu_isset(cpu, cfg->domain)) continue; vector = cfg->vector; @@ -1356,6 +1398,8 @@ static void __init setup_IO_APIC_irqs(void) { int apic, pin, idx, irq; int notcon = 0; + struct irq_desc *desc; + int cpu = boot_cpu_id; apic_printk(APIC_VERBOSE, KERN_DEBUG "init IO_APIC IRQs\n"); @@ -1387,7 +1431,12 @@ static void __init setup_IO_APIC_irqs(void) if (multi_timer_check(apic, irq)) continue; #endif - add_pin_to_irq(irq, apic, pin); + desc = irq_to_desc_alloc_cpu(irq, cpu); + if (!desc) { + printk(KERN_INFO "can not get irq_desc for %d\n", irq); + continue; + } + add_pin_to_irq_cpu(irq, cpu, apic, pin); setup_IO_APIC_irq(apic, pin, irq, irq_trigger(idx), irq_polarity(idx)); @@ -1448,6 +1497,7 @@ __apicdebuginit(void) print_IO_APIC(void) union IO_APIC_reg_03 reg_03; unsigned long flags; struct irq_cfg *cfg; + struct irq_desc *desc; unsigned int irq; if (apic_verbosity == APIC_QUIET) @@ -1537,8 +1587,13 @@ __apicdebuginit(void) print_IO_APIC(void) } } printk(KERN_DEBUG "IRQ to pin mappings:\n"); - for_each_irq_cfg(irq, cfg) { - struct irq_pin_list *entry = cfg->irq_2_pin; + for_each_irq_desc(irq, desc) { + struct irq_pin_list *entry; + + if (!desc) + continue; + cfg = desc->chip_data; + entry = cfg->irq_2_pin; if (!entry) continue; printk(KERN_DEBUG "IRQ%d ", irq); @@ -2022,6 +2077,7 @@ static unsigned int startup_ioapic_irq(unsigned int irq) { int was_pending = 0; unsigned long flags; + struct irq_cfg *cfg; spin_lock_irqsave(&ioapic_lock, flags); if (irq < 16) { @@ -2029,6 +2085,7 @@ static unsigned int startup_ioapic_irq(unsigned int irq) if (i8259A_irq_pending(irq)) was_pending = 1; } + cfg = irq_cfg(irq); __unmask_IO_APIC_irq(irq); spin_unlock_irqrestore(&ioapic_lock, flags); @@ -2178,6 +2235,9 @@ static void ir_irq_migration(struct work_struct *work) struct irq_desc *desc; for_each_irq_desc(irq, desc) { + if (!desc) + continue; + if (desc->status & IRQ_MOVE_PENDING) { unsigned long flags; @@ -2229,6 +2289,9 @@ asmlinkage void smp_irq_move_cleanup_interrupt(void) struct irq_cfg *cfg; irq = __get_cpu_var(vector_irq)[vector]; + if (irq == -1) + continue; + desc = irq_to_desc(irq); if (!desc) continue; @@ -2430,8 +2493,12 @@ static inline void init_IO_APIC_traps(void) * Also, we've got to be careful not to trash gate * 0x80, because int 0x80 is hm, kind of importantish. ;) */ - for_each_irq_cfg(irq, cfg) { - if (IO_APIC_IRQ(irq) && !cfg->vector) { + for_each_irq_desc(irq, desc) { + if (!desc) + continue; + + cfg = desc->chip_data; + if (IO_APIC_IRQ(irq) && cfg && !cfg->vector) { /* * Hmm.. We don't have an entry for this, * so default to an old-fashioned 8259 @@ -2439,11 +2506,9 @@ static inline void init_IO_APIC_traps(void) */ if (irq < 16) make_8259A_irq(irq); - else { - desc = irq_to_desc(irq); + else /* Strange. Oh, well.. */ desc->chip = &no_irq_chip; - } } } } @@ -2654,7 +2719,7 @@ static inline void __init check_timer(void) * Ok, does IRQ0 through the IOAPIC work? */ if (no_pin1) { - add_pin_to_irq(0, apic1, pin1); + add_pin_to_irq_cpu(0, boot_cpu_id, apic1, pin1); setup_timer_IRQ0_pin(apic1, pin1, cfg->vector); } unmask_IO_APIC_irq(0); @@ -2683,7 +2748,7 @@ static inline void __init check_timer(void) /* * legacy devices should be connected to IO APIC #0 */ - replace_pin_at_irq(0, apic1, pin1, apic2, pin2); + replace_pin_at_irq(0, boot_cpu_id, apic1, pin1, apic2, pin2); setup_timer_IRQ0_pin(apic2, pin2, cfg->vector); unmask_IO_APIC_irq(0); enable_8259A_irq(0); @@ -2902,21 +2967,25 @@ unsigned int create_irq_nr(unsigned int irq_want) unsigned int irq; unsigned int new; unsigned long flags; - struct irq_cfg *cfg_new; - - irq_want = nr_irqs - 1; + struct irq_cfg *cfg_new = NULL; + int cpu = boot_cpu_id; + struct irq_desc *desc_new = NULL; irq = 0; spin_lock_irqsave(&vector_lock, flags); for (new = irq_want; new > 0; new--) { if (platform_legacy_irq(new)) continue; - cfg_new = irq_cfg(new); - if (cfg_new && cfg_new->vector != 0) + + desc_new = irq_to_desc_alloc_cpu(new, cpu); + if (!desc_new) { + printk(KERN_INFO "can not get irq_desc for %d\n", new); + continue; + } + cfg_new = desc_new->chip_data; + + if (cfg_new->vector != 0) continue; - /* check if need to create one */ - if (!cfg_new) - cfg_new = irq_cfg_alloc(new); if (__assign_irq_vector(new, TARGET_CPUS) == 0) irq = new; break; @@ -2925,6 +2994,9 @@ unsigned int create_irq_nr(unsigned int irq_want) if (irq > 0) { dynamic_irq_init(irq); + /* restore it, in case dynamic_irq_init clear it */ + if (desc_new) + desc_new->chip_data = cfg_new; } return irq; } @@ -2944,8 +3016,16 @@ int create_irq(void) void destroy_irq(unsigned int irq) { unsigned long flags; + struct irq_cfg *cfg; + struct irq_desc *desc; + /* store it, in case dynamic_irq_cleanup clear it */ + desc = irq_to_desc(irq); + cfg = desc->chip_data; dynamic_irq_cleanup(irq); + /* connect back irq_cfg */ + if (desc) + desc->chip_data = cfg; #ifdef CONFIG_INTR_REMAP free_irte(irq); @@ -3195,26 +3275,13 @@ static int setup_msi_irq(struct pci_dev *dev, struct msi_desc *desc, int irq) return 0; } -static unsigned int build_irq_for_pci_dev(struct pci_dev *dev) -{ - unsigned int irq; - - irq = dev->bus->number; - irq <<= 8; - irq |= dev->devfn; - irq <<= 12; - - return irq; -} - -int arch_setup_msi_irq(struct pci_dev *dev, struct msi_desc *desc) +int arch_setup_msi_irq(struct pci_dev *dev, struct msi_desc *msidesc) { unsigned int irq; int ret; unsigned int irq_want; - irq_want = build_irq_for_pci_dev(dev) + 0x100; - + irq_want = nr_irqs - 1; irq = create_irq_nr(irq_want); if (irq == 0) return -1; @@ -3228,7 +3295,7 @@ int arch_setup_msi_irq(struct pci_dev *dev, struct msi_desc *desc) goto error; no_ir: #endif - ret = setup_msi_irq(dev, desc, irq); + ret = setup_msi_irq(dev, msidesc, irq); if (ret < 0) { destroy_irq(irq); return ret; @@ -3246,7 +3313,7 @@ int arch_setup_msi_irqs(struct pci_dev *dev, int nvec, int type) { unsigned int irq; int ret, sub_handle; - struct msi_desc *desc; + struct msi_desc *msidesc; unsigned int irq_want; #ifdef CONFIG_INTR_REMAP @@ -3254,10 +3321,11 @@ int arch_setup_msi_irqs(struct pci_dev *dev, int nvec, int type) int index = 0; #endif - irq_want = build_irq_for_pci_dev(dev) + 0x100; + irq_want = nr_irqs - 1; sub_handle = 0; - list_for_each_entry(desc, &dev->msi_list, list) { - irq = create_irq_nr(irq_want--); + list_for_each_entry(msidesc, &dev->msi_list, list) { + irq = create_irq_nr(irq_want); + irq_want--; if (irq == 0) return -1; #ifdef CONFIG_INTR_REMAP @@ -3289,7 +3357,7 @@ int arch_setup_msi_irqs(struct pci_dev *dev, int nvec, int type) } no_ir: #endif - ret = setup_msi_irq(dev, desc, irq); + ret = setup_msi_irq(dev, msidesc, irq); if (ret < 0) goto error; sub_handle++; @@ -3707,17 +3775,29 @@ int __init io_apic_get_version(int ioapic) int io_apic_set_pci_routing (int ioapic, int pin, int irq, int triggering, int polarity) { + struct irq_desc *desc; + struct irq_cfg *cfg; + int cpu = boot_cpu_id; + if (!IO_APIC_IRQ(irq)) { apic_printk(APIC_QUIET,KERN_ERR "IOAPIC[%d]: Invalid reference to IRQ 0\n", ioapic); return -EINVAL; } + desc = irq_to_desc_alloc_cpu(irq, cpu); + if (!desc) { + printk(KERN_INFO "can not get irq_desc %d\n", irq); + return 0; + } + /* * IRQs < 16 are already in the irq_2_pin[] map */ - if (irq >= 16) - add_pin_to_irq(irq, ioapic, pin); + if (irq >= 16) { + cfg = desc->chip_data; + add_pin_to_irq_cpu(irq, cpu, ioapic, pin); + } setup_IO_APIC_irq(ioapic, pin, irq, triggering, polarity); @@ -3773,7 +3853,8 @@ void __init setup_ioapic_dest(void) * when you have too many devices, because at that time only boot * cpu is online. */ - cfg = irq_cfg(irq); + desc = irq_to_desc(irq); + cfg = desc->chip_data; if (!cfg->vector) { setup_IO_APIC_irq(ioapic, pin, irq, irq_trigger(irq_entry), @@ -3785,7 +3866,6 @@ void __init setup_ioapic_dest(void) /* * Honour affinities which have been set in early boot */ - desc = irq_to_desc(irq); if (desc->status & (IRQ_NO_BALANCING | IRQ_AFFINITY_SET)) mask = desc->affinity; @@ -3846,7 +3926,6 @@ void __init ioapic_init_mappings(void) struct resource *ioapic_res; int i; - irq_2_pin_init(); ioapic_res = ioapic_setup_resources(); for (i = 0; i < nr_ioapics; i++) { if (smp_found_config) { diff --git a/arch/x86/kernel/irq.c b/arch/x86/kernel/irq.c index d1d4dc52f649..3f1d9d18df67 100644 --- a/arch/x86/kernel/irq.c +++ b/arch/x86/kernel/irq.c @@ -118,6 +118,9 @@ int show_interrupts(struct seq_file *p, void *v) } desc = irq_to_desc(i); + if (!desc) + return 0; + spin_lock_irqsave(&desc->lock, flags); #ifndef CONFIG_SMP any_count = kstat_irqs(i); diff --git a/arch/x86/kernel/irq_32.c b/arch/x86/kernel/irq_32.c index a51382672de0..119fc9c8ff7f 100644 --- a/arch/x86/kernel/irq_32.c +++ b/arch/x86/kernel/irq_32.c @@ -242,6 +242,8 @@ void fixup_irqs(cpumask_t map) for_each_irq_desc(irq, desc) { cpumask_t mask; + if (!desc) + continue; if (irq == 2) continue; diff --git a/arch/x86/kernel/irq_64.c b/arch/x86/kernel/irq_64.c index 60eb84eb77a0..900009c70591 100644 --- a/arch/x86/kernel/irq_64.c +++ b/arch/x86/kernel/irq_64.c @@ -94,6 +94,8 @@ void fixup_irqs(cpumask_t map) int break_affinity = 0; int set_affinity = 1; + if (!desc) + continue; if (irq == 2) continue; diff --git a/arch/x86/kernel/irqinit_32.c b/arch/x86/kernel/irqinit_32.c index 845aa9803e80..5a5651b7f9e6 100644 --- a/arch/x86/kernel/irqinit_32.c +++ b/arch/x86/kernel/irqinit_32.c @@ -69,7 +69,6 @@ void __init init_ISA_irqs (void) * 16 old-style INTA-cycle interrupts: */ for (i = 0; i < 16; i++) { - /* first time call this irq_desc */ struct irq_desc *desc = irq_to_desc(i); desc->status = IRQ_DISABLED; diff --git a/arch/x86/kernel/irqinit_64.c b/arch/x86/kernel/irqinit_64.c index ff0235391285..cd9f42d028d9 100644 --- a/arch/x86/kernel/irqinit_64.c +++ b/arch/x86/kernel/irqinit_64.c @@ -143,7 +143,6 @@ void __init init_ISA_irqs(void) init_8259A(0); for (i = 0; i < 16; i++) { - /* first time call this irq_desc */ struct irq_desc *desc = irq_to_desc(i); desc->status = IRQ_DISABLED; diff --git a/drivers/char/random.c b/drivers/char/random.c index 675076f5fca8..d26891bfcd41 100644 --- a/drivers/char/random.c +++ b/drivers/char/random.c @@ -558,23 +558,9 @@ struct timer_rand_state { unsigned dont_count_entropy:1; }; -static struct timer_rand_state *irq_timer_state[NR_IRQS]; - -static struct timer_rand_state *get_timer_rand_state(unsigned int irq) -{ - if (irq >= nr_irqs) - return NULL; - - return irq_timer_state[irq]; -} - -static void set_timer_rand_state(unsigned int irq, struct timer_rand_state *state) -{ - if (irq >= nr_irqs) - return; - - irq_timer_state[irq] = state; -} +#ifndef CONFIG_SPARSE_IRQ +struct timer_rand_state *irq_timer_state[NR_IRQS]; +#endif static struct timer_rand_state input_timer_state; @@ -933,8 +919,10 @@ void rand_initialize_irq(int irq) { struct timer_rand_state *state; +#ifndef CONFIG_SPARSE_IRQ if (irq >= nr_irqs) return; +#endif state = get_timer_rand_state(irq); diff --git a/drivers/pci/intr_remapping.c b/drivers/pci/intr_remapping.c index 2de5a3238c94..c9958ec5e25e 100644 --- a/drivers/pci/intr_remapping.c +++ b/drivers/pci/intr_remapping.c @@ -19,17 +19,75 @@ struct irq_2_iommu { u8 irte_mask; }; -static struct irq_2_iommu irq_2_iommuX[NR_IRQS]; +#ifdef CONFIG_SPARSE_IRQ +static struct irq_2_iommu *get_one_free_irq_2_iommu(int cpu) +{ + struct irq_2_iommu *iommu; + int node; + + node = cpu_to_node(cpu); + + iommu = kzalloc_node(sizeof(*iommu), GFP_ATOMIC, node); + printk(KERN_DEBUG "alloc irq_2_iommu on cpu %d node %d\n", cpu, node); + + return iommu; +} static struct irq_2_iommu *irq_2_iommu(unsigned int irq) { - return (irq < nr_irqs) ? irq_2_iommuX + irq : NULL; + struct irq_desc *desc; + + desc = irq_to_desc(irq); + + if (WARN_ON_ONCE(!desc)) + return NULL; + + return desc->irq_2_iommu; +} + +static struct irq_2_iommu *irq_2_iommu_alloc_cpu(unsigned int irq, int cpu) +{ + struct irq_desc *desc; + struct irq_2_iommu *irq_iommu; + + /* + * alloc irq desc if not allocated already. + */ + desc = irq_to_desc_alloc_cpu(irq, cpu); + if (!desc) { + printk(KERN_INFO "can not get irq_desc for %d\n", irq); + return NULL; + } + + irq_iommu = desc->irq_2_iommu; + + if (!irq_iommu) + desc->irq_2_iommu = get_one_free_irq_2_iommu(cpu); + + return desc->irq_2_iommu; } +static struct irq_2_iommu *irq_2_iommu_alloc(unsigned int irq) +{ + return irq_2_iommu_alloc_cpu(irq, boot_cpu_id); +} + +#else /* !CONFIG_SPARSE_IRQ */ + +static struct irq_2_iommu irq_2_iommuX[NR_IRQS]; + +static struct irq_2_iommu *irq_2_iommu(unsigned int irq) +{ + if (irq < nr_irqs) + return &irq_2_iommuX[irq]; + + return NULL; +} static struct irq_2_iommu *irq_2_iommu_alloc(unsigned int irq) { return irq_2_iommu(irq); } +#endif static DEFINE_SPINLOCK(irq_2_ir_lock); @@ -86,9 +144,11 @@ int alloc_irte(struct intel_iommu *iommu, int irq, u16 count) if (!count) return -1; +#ifndef CONFIG_SPARSE_IRQ /* protect irq_2_iommu_alloc later */ if (irq >= nr_irqs) return -1; +#endif /* * start the IRTE search from index 0. @@ -130,6 +190,12 @@ int alloc_irte(struct intel_iommu *iommu, int irq, u16 count) table->base[i].present = 1; irq_iommu = irq_2_iommu_alloc(irq); + if (!irq_iommu) { + spin_unlock(&irq_2_ir_lock); + printk(KERN_ERR "can't allocate irq_2_iommu\n"); + return -1; + } + irq_iommu->iommu = iommu; irq_iommu->irte_index = index; irq_iommu->sub_handle = 0; @@ -177,6 +243,12 @@ int set_irte_irq(int irq, struct intel_iommu *iommu, u16 index, u16 subhandle) irq_iommu = irq_2_iommu_alloc(irq); + if (!irq_iommu) { + spin_unlock(&irq_2_ir_lock); + printk(KERN_ERR "can't allocate irq_2_iommu\n"); + return -1; + } + irq_iommu->iommu = iommu; irq_iommu->irte_index = index; irq_iommu->sub_handle = subhandle; diff --git a/drivers/xen/events.c b/drivers/xen/events.c index 1e3b934a4cf7..2924faa7f6c4 100644 --- a/drivers/xen/events.c +++ b/drivers/xen/events.c @@ -141,8 +141,12 @@ static void init_evtchn_cpu_bindings(void) int i; /* By default all event channels notify CPU#0. */ - for_each_irq_desc(i, desc) + for_each_irq_desc(i, desc) { + if (!desc) + continue; + desc->affinity = cpumask_of_cpu(0); + } #endif memset(cpu_evtchn, 0, sizeof(cpu_evtchn)); @@ -231,7 +235,7 @@ static int find_unbound_irq(void) int irq; /* Only allocate from dynirq range */ - for_each_irq_nr(irq) + for (irq = 0; irq < nr_irqs; irq++) if (irq_bindcount[irq] == 0) break; @@ -792,7 +796,7 @@ void xen_irq_resume(void) mask_evtchn(evtchn); /* No IRQ <-> event-channel mappings. */ - for_each_irq_nr(irq) + for (irq = 0; irq < nr_irqs; irq++) irq_info[irq].evtchn = 0; /* zap event-channel binding */ for (evtchn = 0; evtchn < NR_EVENT_CHANNELS; evtchn++) @@ -824,7 +828,7 @@ void __init xen_init_IRQ(void) mask_evtchn(i); /* Dynamic IRQ space is currently unbound. Zero the refcnts. */ - for_each_irq_nr(i) + for (i = 0; i < nr_irqs; i++) irq_bindcount[i] = 0; irq_ctx_init(smp_processor_id()); diff --git a/fs/proc/stat.c b/fs/proc/stat.c index 81904f07679d..a13431ab7c65 100644 --- a/fs/proc/stat.c +++ b/fs/proc/stat.c @@ -27,6 +27,7 @@ static int show_stat(struct seq_file *p, void *v) u64 sum = 0; struct timespec boottime; unsigned int per_irq_sum; + struct irq_desc *desc; user = nice = system = idle = iowait = irq = softirq = steal = cputime64_zero; @@ -44,10 +45,11 @@ static int show_stat(struct seq_file *p, void *v) softirq = cputime64_add(softirq, kstat_cpu(i).cpustat.softirq); steal = cputime64_add(steal, kstat_cpu(i).cpustat.steal); guest = cputime64_add(guest, kstat_cpu(i).cpustat.guest); - - for_each_irq_nr(j) + for_each_irq_desc(j, desc) { + if (!desc) + continue; sum += kstat_irqs_cpu(j, i); - + } sum += arch_irq_stat_cpu(i); } sum += arch_irq_stat(); @@ -90,11 +92,14 @@ static int show_stat(struct seq_file *p, void *v) seq_printf(p, "intr %llu", (unsigned long long)sum); /* sum again ? it could be updated? */ - for_each_irq_nr(j) { + for (j = 0; j < NR_IRQS; j++) { + desc = irq_to_desc(j); per_irq_sum = 0; - for_each_possible_cpu(i) - per_irq_sum += kstat_irqs_cpu(j, i); + if (desc) { + for_each_possible_cpu(i) + per_irq_sum += kstat_irqs_cpu(j, i); + } seq_printf(p, " %u", per_irq_sum); } diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h index f58a0cf8929a..79e915e7e8a5 100644 --- a/include/linux/interrupt.h +++ b/include/linux/interrupt.h @@ -18,6 +18,8 @@ #include #include +extern int nr_irqs; + /* * These correspond to the IORESOURCE_IRQ_* defines in * linux/ioport.h to select the interrupt line behaviour. When diff --git a/include/linux/irq.h b/include/linux/irq.h index 3dddfa703ebd..63b00439d4d2 100644 --- a/include/linux/irq.h +++ b/include/linux/irq.h @@ -129,6 +129,8 @@ struct irq_chip { const char *typename; }; +struct timer_rand_state; +struct irq_2_iommu; /** * struct irq_desc - interrupt descriptor * @irq: interrupt number for this descriptor @@ -154,6 +156,13 @@ struct irq_chip { */ struct irq_desc { unsigned int irq; +#ifdef CONFIG_SPARSE_IRQ + struct timer_rand_state *timer_rand_state; + unsigned int *kstat_irqs; +# ifdef CONFIG_INTR_REMAP + struct irq_2_iommu *irq_2_iommu; +# endif +#endif irq_flow_handler_t handle_irq; struct irq_chip *chip; struct msi_desc *msi_desc; @@ -181,14 +190,52 @@ struct irq_desc { const char *name; } ____cacheline_internodealigned_in_smp; +extern void early_irq_init(void); +extern void arch_early_irq_init(void); +extern void arch_init_chip_data(struct irq_desc *desc, int cpu); +extern void arch_init_copy_chip_data(struct irq_desc *old_desc, + struct irq_desc *desc, int cpu); +extern void arch_free_chip_data(struct irq_desc *old_desc, struct irq_desc *desc); + +#ifndef CONFIG_SPARSE_IRQ extern struct irq_desc irq_desc[NR_IRQS]; static inline struct irq_desc *irq_to_desc(unsigned int irq) { - return (irq < nr_irqs) ? irq_desc + irq : NULL; + return (irq < NR_IRQS) ? irq_desc + irq : NULL; +} +static inline struct irq_desc *irq_to_desc_alloc_cpu(unsigned int irq, int cpu) +{ + return irq_to_desc(irq); } +#ifdef CONFIG_GENERIC_HARDIRQS +# define for_each_irq_desc(irq, desc) \ + for (irq = 0, desc = irq_desc; irq < nr_irqs; irq++, desc++) +# define for_each_irq_desc_reverse(irq, desc) \ + for (irq = nr_irqs - 1, desc = irq_desc + (nr_irqs - 1); \ + irq >= 0; irq--, desc--) +#endif + +#else + +extern struct irq_desc *irq_to_desc(unsigned int irq); +extern struct irq_desc *irq_to_desc_alloc_cpu(unsigned int irq, int cpu); +extern struct irq_desc *move_irq_desc(struct irq_desc *old_desc, int cpu); + +# define for_each_irq_desc(irq, desc) \ + for (irq = 0, desc = irq_to_desc(irq); irq < nr_irqs; irq++, desc = irq_to_desc(irq)) +# define for_each_irq_desc_reverse(irq, desc) \ + for (irq = nr_irqs - 1, desc = irq_to_desc(irq); irq >= 0; irq--, desc = irq_to_desc(irq)) + +#define kstat_irqs_this_cpu(DESC) \ + ((DESC)->kstat_irqs[smp_processor_id()]) +#define kstat_incr_irqs_this_cpu(irqno, DESC) \ + ((DESC)->kstat_irqs[smp_processor_id()]++) + +#endif + /* * Migration helpers for obsolete names, they will go away: */ @@ -380,6 +427,11 @@ extern int set_irq_msi(unsigned int irq, struct msi_desc *entry); #define get_irq_data(irq) (irq_to_desc(irq)->handler_data) #define get_irq_msi(irq) (irq_to_desc(irq)->msi_desc) +#define get_irq_desc_chip(desc) ((desc)->chip) +#define get_irq_desc_chip_data(desc) ((desc)->chip_data) +#define get_irq_desc_data(desc) ((desc)->handler_data) +#define get_irq_desc_msi(desc) ((desc)->msi_desc) + #endif /* CONFIG_GENERIC_HARDIRQS */ #endif /* !CONFIG_S390 */ diff --git a/include/linux/irqnr.h b/include/linux/irqnr.h index 452c280c8115..7a299e989f8b 100644 --- a/include/linux/irqnr.h +++ b/include/linux/irqnr.h @@ -7,18 +7,10 @@ # define for_each_irq_desc(irq, desc) \ for (irq = 0; irq < nr_irqs; irq++) -#else -extern int nr_irqs; -# define for_each_irq_desc(irq, desc) \ - for (irq = 0, desc = irq_desc; irq < nr_irqs; irq++, desc++) - -# define for_each_irq_desc_reverse(irq, desc) \ - for (irq = nr_irqs - 1, desc = irq_desc + (nr_irqs - 1); \ - irq >= 0; irq--, desc--) +static inline early_sparse_irq_init(void) +{ +} #endif -#define for_each_irq_nr(irq) \ - for (irq = 0; irq < nr_irqs; irq++) - #endif diff --git a/include/linux/kernel_stat.h b/include/linux/kernel_stat.h index 4a145caeee07..4ee4b3d2316f 100644 --- a/include/linux/kernel_stat.h +++ b/include/linux/kernel_stat.h @@ -28,7 +28,9 @@ struct cpu_usage_stat { struct kernel_stat { struct cpu_usage_stat cpustat; - unsigned int irqs[NR_IRQS]; +#ifndef CONFIG_SPARSE_IRQ + unsigned int irqs[NR_IRQS]; +#endif }; DECLARE_PER_CPU(struct kernel_stat, kstat); @@ -39,6 +41,10 @@ DECLARE_PER_CPU(struct kernel_stat, kstat); extern unsigned long long nr_context_switches(void); +#ifndef CONFIG_SPARSE_IRQ +#define kstat_irqs_this_cpu(irq) \ + (kstat_this_cpu.irqs[irq]) + struct irq_desc; static inline void kstat_incr_irqs_this_cpu(unsigned int irq, @@ -46,11 +52,17 @@ static inline void kstat_incr_irqs_this_cpu(unsigned int irq, { kstat_this_cpu.irqs[irq]++; } +#endif + +#ifndef CONFIG_SPARSE_IRQ static inline unsigned int kstat_irqs_cpu(unsigned int irq, int cpu) { return kstat_cpu(cpu).irqs[irq]; } +#else +extern unsigned int kstat_irqs_cpu(unsigned int irq, int cpu); +#endif /* * Number of interrupts per specific IRQ source, since bootup diff --git a/include/linux/random.h b/include/linux/random.h index 36f125c0c603..ad9daa2374d5 100644 --- a/include/linux/random.h +++ b/include/linux/random.h @@ -44,6 +44,57 @@ struct rand_pool_info { extern void rand_initialize_irq(int irq); +struct timer_rand_state; +#ifndef CONFIG_SPARSE_IRQ + +extern struct timer_rand_state *irq_timer_state[]; + +extern int nr_irqs; +static inline struct timer_rand_state *get_timer_rand_state(unsigned int irq) +{ + if (irq >= nr_irqs) + return NULL; + + return irq_timer_state[irq]; +} + +static inline void set_timer_rand_state(unsigned int irq, struct timer_rand_state *state) +{ + if (irq >= nr_irqs) + return; + + irq_timer_state[irq] = state; +} + +#else + +#include +static inline struct timer_rand_state *get_timer_rand_state(unsigned int irq) +{ + struct irq_desc *desc; + + desc = irq_to_desc(irq); + + if (!desc) + return NULL; + + return desc->timer_rand_state; +} + +static inline void set_timer_rand_state(unsigned int irq, struct timer_rand_state *state) +{ + struct irq_desc *desc; + + desc = irq_to_desc(irq); + + if (!desc) + return; + + desc->timer_rand_state = state; +} +#endif + + extern void add_input_randomness(unsigned int type, unsigned int code, unsigned int value); extern void add_interrupt_randomness(int irq); diff --git a/init/main.c b/init/main.c index 7e117a231af1..c1f999a3cf31 100644 --- a/init/main.c +++ b/init/main.c @@ -539,6 +539,15 @@ void __init __weak thread_info_cache_init(void) { } +void __init __weak arch_early_irq_init(void) +{ +} + +void __init __weak early_irq_init(void) +{ + arch_early_irq_init(); +} + asmlinkage void __init start_kernel(void) { char * command_line; @@ -603,6 +612,8 @@ asmlinkage void __init start_kernel(void) sort_main_extable(); trap_init(); rcu_init(); + /* init some links before init_ISA_irqs() */ + early_irq_init(); init_IRQ(); pidhash_init(); init_timers(); diff --git a/kernel/irq/autoprobe.c b/kernel/irq/autoprobe.c index cc0f7321b8ce..650ce4102a63 100644 --- a/kernel/irq/autoprobe.c +++ b/kernel/irq/autoprobe.c @@ -40,6 +40,9 @@ unsigned long probe_irq_on(void) * flush such a longstanding irq before considering it as spurious. */ for_each_irq_desc_reverse(i, desc) { + if (!desc) + continue; + spin_lock_irq(&desc->lock); if (!desc->action && !(desc->status & IRQ_NOPROBE)) { /* @@ -68,6 +71,9 @@ unsigned long probe_irq_on(void) * happened in the previous stage, it may have masked itself) */ for_each_irq_desc_reverse(i, desc) { + if (!desc) + continue; + spin_lock_irq(&desc->lock); if (!desc->action && !(desc->status & IRQ_NOPROBE)) { desc->status |= IRQ_AUTODETECT | IRQ_WAITING; @@ -86,6 +92,9 @@ unsigned long probe_irq_on(void) * Now filter out any obviously spurious interrupts */ for_each_irq_desc(i, desc) { + if (!desc) + continue; + spin_lock_irq(&desc->lock); status = desc->status; @@ -124,6 +133,9 @@ unsigned int probe_irq_mask(unsigned long val) int i; for_each_irq_desc(i, desc) { + if (!desc) + continue; + spin_lock_irq(&desc->lock); status = desc->status; @@ -166,6 +178,9 @@ int probe_irq_off(unsigned long val) unsigned int status; for_each_irq_desc(i, desc) { + if (!desc) + continue; + spin_lock_irq(&desc->lock); status = desc->status; diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index 10b5092e9bfe..8e4fce4a1b1f 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -24,9 +24,10 @@ */ void dynamic_irq_init(unsigned int irq) { - struct irq_desc *desc = irq_to_desc(irq); + struct irq_desc *desc; unsigned long flags; + desc = irq_to_desc(irq); if (!desc) { WARN(1, KERN_ERR "Trying to initialize invalid IRQ%d\n", irq); return; diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index c815b42d0f5b..96ca203eb51b 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -15,9 +15,16 @@ #include #include #include +#include +#include #include "internals.h" +/* + * lockdep: we want to handle all irq_desc locks as a single lock-class: + */ +static struct lock_class_key irq_desc_lock_class; + /** * handle_bad_irq - handle spurious and unhandled irqs * @irq: the interrupt number @@ -49,6 +56,155 @@ void handle_bad_irq(unsigned int irq, struct irq_desc *desc) int nr_irqs = NR_IRQS; EXPORT_SYMBOL_GPL(nr_irqs); +void __init __attribute__((weak)) arch_early_irq_init(void) +{ +} + +#ifdef CONFIG_SPARSE_IRQ +static struct irq_desc irq_desc_init = { + .irq = -1, + .status = IRQ_DISABLED, + .chip = &no_irq_chip, + .handle_irq = handle_bad_irq, + .depth = 1, + .lock = __SPIN_LOCK_UNLOCKED(irq_desc_init.lock), +#ifdef CONFIG_SMP + .affinity = CPU_MASK_ALL +#endif +}; + +static void init_kstat_irqs(struct irq_desc *desc, int cpu, int nr) +{ + unsigned long bytes; + char *ptr; + int node; + + /* Compute how many bytes we need per irq and allocate them */ + bytes = nr * sizeof(unsigned int); + + node = cpu_to_node(cpu); + ptr = kzalloc_node(bytes, GFP_ATOMIC, node); + printk(KERN_DEBUG " alloc kstat_irqs on cpu %d node %d\n", cpu, node); + + if (ptr) + desc->kstat_irqs = (unsigned int *)ptr; +} + +void __attribute__((weak)) arch_init_chip_data(struct irq_desc *desc, int cpu) +{ +} + +static void init_one_irq_desc(int irq, struct irq_desc *desc, int cpu) +{ + memcpy(desc, &irq_desc_init, sizeof(struct irq_desc)); + desc->irq = irq; +#ifdef CONFIG_SMP + desc->cpu = cpu; +#endif + lockdep_set_class(&desc->lock, &irq_desc_lock_class); + init_kstat_irqs(desc, cpu, nr_cpu_ids); + if (!desc->kstat_irqs) { + printk(KERN_ERR "can not alloc kstat_irqs\n"); + BUG_ON(1); + } + arch_init_chip_data(desc, cpu); +} + +/* + * Protect the sparse_irqs: + */ +static DEFINE_SPINLOCK(sparse_irq_lock); + +struct irq_desc *irq_desc_ptrs[NR_IRQS] __read_mostly; + +static struct irq_desc irq_desc_legacy[16] __cacheline_aligned_in_smp = { + [0 ... 15] = { + .irq = -1, + .status = IRQ_DISABLED, + .chip = &no_irq_chip, + .handle_irq = handle_bad_irq, + .depth = 1, + .lock = __SPIN_LOCK_UNLOCKED(irq_desc_init.lock), +#ifdef CONFIG_SMP + .affinity = CPU_MASK_ALL +#endif + } +}; + +/* FIXME: use bootmem alloc ...*/ +static unsigned int kstat_irqs_legacy[16][NR_CPUS]; + +void __init early_irq_init(void) +{ + struct irq_desc *desc; + int legacy_count; + int i; + + desc = irq_desc_legacy; + legacy_count = ARRAY_SIZE(irq_desc_legacy); + + for (i = 0; i < legacy_count; i++) { + desc[i].irq = i; + desc[i].kstat_irqs = kstat_irqs_legacy[i]; + + irq_desc_ptrs[i] = desc + i; + } + + for (i = legacy_count; i < NR_IRQS; i++) + irq_desc_ptrs[i] = NULL; + + arch_early_irq_init(); +} + +struct irq_desc *irq_to_desc(unsigned int irq) +{ + return (irq < NR_IRQS) ? irq_desc_ptrs[irq] : NULL; +} + +struct irq_desc *irq_to_desc_alloc_cpu(unsigned int irq, int cpu) +{ + struct irq_desc *desc; + unsigned long flags; + int node; + + if (irq >= NR_IRQS) { + printk(KERN_WARNING "irq >= NR_IRQS in irq_to_desc_alloc: %d %d\n", + irq, NR_IRQS); + WARN_ON(1); + return NULL; + } + + desc = irq_desc_ptrs[irq]; + if (desc) + return desc; + + spin_lock_irqsave(&sparse_irq_lock, flags); + + /* We have to check it to avoid races with another CPU */ + desc = irq_desc_ptrs[irq]; + if (desc) + goto out_unlock; + + node = cpu_to_node(cpu); + desc = kzalloc_node(sizeof(*desc), GFP_ATOMIC, node); + printk(KERN_DEBUG " alloc irq_desc for %d on cpu %d node %d\n", + irq, cpu, node); + if (!desc) { + printk(KERN_ERR "can not alloc irq_desc\n"); + BUG_ON(1); + } + init_one_irq_desc(irq, desc, cpu); + + irq_desc_ptrs[irq] = desc; + +out_unlock: + spin_unlock_irqrestore(&sparse_irq_lock, flags); + + return desc; +} + +#else + struct irq_desc irq_desc[NR_IRQS] __cacheline_aligned_in_smp = { [0 ... NR_IRQS-1] = { .status = IRQ_DISABLED, @@ -62,6 +218,8 @@ struct irq_desc irq_desc[NR_IRQS] __cacheline_aligned_in_smp = { } }; +#endif + /* * What should we do if we get a hw irq event on an illegal vector? * Each architecture has to answer this themself. @@ -261,17 +419,28 @@ out: #ifdef CONFIG_TRACE_IRQFLAGS -/* - * lockdep: we want to handle all irq_desc locks as a single lock-class: - */ -static struct lock_class_key irq_desc_lock_class; - void early_init_irq_lock_class(void) { +#ifndef CONFIG_SPARSE_IRQ struct irq_desc *desc; int i; - for_each_irq_desc(i, desc) + for_each_irq_desc(i, desc) { + if (!desc) + continue; + lockdep_set_class(&desc->lock, &irq_desc_lock_class); + } +#endif +} +#endif + +#ifdef CONFIG_SPARSE_IRQ +unsigned int kstat_irqs_cpu(unsigned int irq, int cpu) +{ + struct irq_desc *desc = irq_to_desc(irq); + return desc->kstat_irqs[cpu]; } #endif +EXPORT_SYMBOL(kstat_irqs_cpu); + diff --git a/kernel/irq/proc.c b/kernel/irq/proc.c index d257e7d6a8a4..f6b3440f05bc 100644 --- a/kernel/irq/proc.c +++ b/kernel/irq/proc.c @@ -243,7 +243,11 @@ void init_irq_proc(void) /* * Create entries for all existing IRQs. */ - for_each_irq_desc(irq, desc) + for_each_irq_desc(irq, desc) { + if (!desc) + continue; + register_irq_proc(irq, desc); + } } diff --git a/kernel/irq/spurious.c b/kernel/irq/spurious.c index dd364c11e56e..3738107531fd 100644 --- a/kernel/irq/spurious.c +++ b/kernel/irq/spurious.c @@ -91,6 +91,9 @@ static int misrouted_irq(int irq) int i, ok = 0; for_each_irq_desc(i, desc) { + if (!desc) + continue; + if (!i) continue; @@ -112,6 +115,8 @@ static void poll_spurious_irqs(unsigned long dummy) for_each_irq_desc(i, desc) { unsigned int status; + if (!desc) + continue; if (!i) continue; -- cgit v1.2.3 From 99d093d12897562a253540a902bbf65ec16042ac Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Fri, 5 Dec 2008 18:58:32 -0800 Subject: x86: use NR_IRQS_LEGACY Impact: cleanup Introduce NR_IRQS_LEGACY instead of hard coded number. Signed-off-by: Yinghai Lu Signed-off-by: Ingo Molnar --- arch/x86/include/asm/irq_vectors.h | 2 ++ arch/x86/kernel/io_apic.c | 10 +++++----- arch/x86/kernel/irqinit_32.c | 2 +- arch/x86/kernel/irqinit_64.c | 2 +- kernel/irq/handle.c | 6 +++--- 5 files changed, 12 insertions(+), 10 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/irq_vectors.h b/arch/x86/include/asm/irq_vectors.h index bb6b69a6b125..f7ff65032b9d 100644 --- a/arch/x86/include/asm/irq_vectors.h +++ b/arch/x86/include/asm/irq_vectors.h @@ -101,6 +101,8 @@ #define LAST_VM86_IRQ 15 #define invalid_vm86_irq(irq) ((irq) < 3 || (irq) > 15) +#define NR_IRQS_LEGACY 16 + #if defined(CONFIG_X86_IO_APIC) && !defined(CONFIG_X86_VOYAGER) #ifndef CONFIG_SPARSE_IRQ diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index 9de17f5c1125..9870461ae933 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -847,7 +847,7 @@ EXPORT_SYMBOL(IO_APIC_get_PCI_irq_vector); */ static int EISA_ELCR(unsigned int irq) { - if (irq < 16) { + if (irq < NR_IRQS_LEGACY) { unsigned int port = 0x4d0 + (irq >> 3); return (inb(port) >> (irq & 7)) & 1; } @@ -1388,7 +1388,7 @@ static void setup_IO_APIC_irq(int apic, int pin, unsigned int irq, } ioapic_register_intr(irq, trigger); - if (irq < 16) + if (irq < NR_IRQS_LEGACY) disable_8259A_irq(irq); ioapic_write_entry(apic, pin, entry); @@ -2080,7 +2080,7 @@ static unsigned int startup_ioapic_irq(unsigned int irq) struct irq_cfg *cfg; spin_lock_irqsave(&ioapic_lock, flags); - if (irq < 16) { + if (irq < NR_IRQS_LEGACY) { disable_8259A_irq(irq); if (i8259A_irq_pending(irq)) was_pending = 1; @@ -2504,7 +2504,7 @@ static inline void init_IO_APIC_traps(void) * so default to an old-fashioned 8259 * interrupt if we can.. */ - if (irq < 16) + if (irq < NR_IRQS_LEGACY) make_8259A_irq(irq); else /* Strange. Oh, well.. */ @@ -3794,7 +3794,7 @@ int io_apic_set_pci_routing (int ioapic, int pin, int irq, int triggering, int p /* * IRQs < 16 are already in the irq_2_pin[] map */ - if (irq >= 16) { + if (irq >= NR_IRQS_LEGACY) { cfg = desc->chip_data; add_pin_to_irq_cpu(irq, cpu, ioapic, pin); } diff --git a/arch/x86/kernel/irqinit_32.c b/arch/x86/kernel/irqinit_32.c index 5a5651b7f9e6..6a92f47c52e7 100644 --- a/arch/x86/kernel/irqinit_32.c +++ b/arch/x86/kernel/irqinit_32.c @@ -68,7 +68,7 @@ void __init init_ISA_irqs (void) /* * 16 old-style INTA-cycle interrupts: */ - for (i = 0; i < 16; i++) { + for (i = 0; i < NR_IRQS_LEGACY; i++) { struct irq_desc *desc = irq_to_desc(i); desc->status = IRQ_DISABLED; diff --git a/arch/x86/kernel/irqinit_64.c b/arch/x86/kernel/irqinit_64.c index cd9f42d028d9..40c1e62ec785 100644 --- a/arch/x86/kernel/irqinit_64.c +++ b/arch/x86/kernel/irqinit_64.c @@ -142,7 +142,7 @@ void __init init_ISA_irqs(void) init_bsp_APIC(); init_8259A(0); - for (i = 0; i < 16; i++) { + for (i = 0; i < NR_IRQS_LEGACY; i++) { struct irq_desc *desc = irq_to_desc(i); desc->status = IRQ_DISABLED; diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index 96ca203eb51b..8aa09547f5ef 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -117,8 +117,8 @@ static DEFINE_SPINLOCK(sparse_irq_lock); struct irq_desc *irq_desc_ptrs[NR_IRQS] __read_mostly; -static struct irq_desc irq_desc_legacy[16] __cacheline_aligned_in_smp = { - [0 ... 15] = { +static struct irq_desc irq_desc_legacy[NR_IRQS_LEGACY] __cacheline_aligned_in_smp = { + [0 ... NR_IRQS_LEGACY-1] = { .irq = -1, .status = IRQ_DISABLED, .chip = &no_irq_chip, @@ -132,7 +132,7 @@ static struct irq_desc irq_desc_legacy[16] __cacheline_aligned_in_smp = { }; /* FIXME: use bootmem alloc ...*/ -static unsigned int kstat_irqs_legacy[16][NR_CPUS]; +static unsigned int kstat_irqs_legacy[NR_IRQS_LEGACY][NR_CPUS]; void __init early_irq_init(void) { -- cgit v1.2.3 From be5d5350a937cd8513b258739f1099420129e96f Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Fri, 5 Dec 2008 18:58:33 -0800 Subject: x86: MSI start irq numbering from nr_irqs_gsi Impact: sanitize MSI irq number ordering from top-down to bottom-up Increase new MSI IRQs starting from nr_irqs_gsi (which is somewhere below 256), instead of decreasing from NR_IRQS. (The latter method can result in confusingly high IRQ numbers - if NR_CPUS is set to a high value and NR_IRQS scales up to a high value.) Signed-off-by: Yinghai Lu Signed-off-by: Ingo Molnar --- arch/x86/include/asm/io_apic.h | 2 +- arch/x86/kernel/io_apic.c | 24 +++++++++++++++++------- arch/x86/kernel/setup.c | 2 +- 3 files changed, 19 insertions(+), 9 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/io_apic.h b/arch/x86/include/asm/io_apic.h index 6afd9933a7dd..b35e94c2d6df 100644 --- a/arch/x86/include/asm/io_apic.h +++ b/arch/x86/include/asm/io_apic.h @@ -188,7 +188,7 @@ extern void restore_IO_APIC_setup(void); extern void reinit_intr_remapped_IO_APIC(int); #endif -extern int probe_nr_irqs(void); +extern void probe_nr_irqs_gsi(void); #else /* !CONFIG_X86_IO_APIC */ #define io_apic_assign_pci_irqs 0 diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index 9870461ae933..0dcde74abd1d 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -2973,7 +2973,7 @@ unsigned int create_irq_nr(unsigned int irq_want) irq = 0; spin_lock_irqsave(&vector_lock, flags); - for (new = irq_want; new > 0; new--) { + for (new = irq_want; new < NR_IRQS; new++) { if (platform_legacy_irq(new)) continue; @@ -3001,11 +3001,14 @@ unsigned int create_irq_nr(unsigned int irq_want) return irq; } +static int nr_irqs_gsi = NR_IRQS_LEGACY; int create_irq(void) { + unsigned int irq_want; int irq; - irq = create_irq_nr(nr_irqs - 1); + irq_want = nr_irqs_gsi; + irq = create_irq_nr(irq_want); if (irq == 0) irq = -1; @@ -3281,7 +3284,7 @@ int arch_setup_msi_irq(struct pci_dev *dev, struct msi_desc *msidesc) int ret; unsigned int irq_want; - irq_want = nr_irqs - 1; + irq_want = nr_irqs_gsi; irq = create_irq_nr(irq_want); if (irq == 0) return -1; @@ -3321,11 +3324,11 @@ int arch_setup_msi_irqs(struct pci_dev *dev, int nvec, int type) int index = 0; #endif - irq_want = nr_irqs - 1; + irq_want = nr_irqs_gsi; sub_handle = 0; list_for_each_entry(msidesc, &dev->msi_list, list) { irq = create_irq_nr(irq_want); - irq_want--; + irq_want++; if (irq == 0) return -1; #ifdef CONFIG_INTR_REMAP @@ -3674,9 +3677,16 @@ int __init io_apic_get_redir_entries (int ioapic) return reg_01.bits.entries; } -int __init probe_nr_irqs(void) +void __init probe_nr_irqs_gsi(void) { - return NR_IRQS; + int idx; + int nr = 0; + + for (idx = 0; idx < nr_ioapics; idx++) + nr += io_apic_get_redir_entries(idx) + 1; + + if (nr > nr_irqs_gsi) + nr_irqs_gsi = nr; } /* -------------------------------------------------------------------------- diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index 9d5674f7b6cc..a3834f123206 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -1082,7 +1082,7 @@ void __init setup_arch(char **cmdline_p) ioapic_init_mappings(); /* need to wait for io_apic is mapped */ - nr_irqs = probe_nr_irqs(); + probe_nr_irqs_gsi(); kvm_guest_init(); -- cgit v1.2.3 From 3145e941fcfe2548fa2270afb1a05bab3a6bc418 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Fri, 5 Dec 2008 18:58:34 -0800 Subject: x86, MSI: pass irq_cfg and irq_desc Impact: simplify code Pass irq_desc and cfg around, instead of raw IRQ numbers - this way we dont have to look it up again and again. Signed-off-by: Yinghai Lu Signed-off-by: Ingo Molnar --- arch/x86/kernel/io_apic.c | 318 ++++++++++++++++++++++++++-------------------- drivers/pci/msi.c | 55 +++++--- include/linux/msi.h | 3 + 3 files changed, 222 insertions(+), 154 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index 0dcde74abd1d..a1a2e070f31a 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -231,6 +231,10 @@ static struct irq_cfg *irq_cfg(unsigned int irq) #endif +static inline void set_extra_move_desc(struct irq_desc *desc, cpumask_t mask) +{ +} + struct io_apic { unsigned int index; unsigned int unused[3]; @@ -272,11 +276,10 @@ static inline void io_apic_modify(unsigned int apic, unsigned int reg, unsigned writel(value, &io_apic->data); } -static bool io_apic_level_ack_pending(unsigned int irq) +static bool io_apic_level_ack_pending(struct irq_cfg *cfg) { struct irq_pin_list *entry; unsigned long flags; - struct irq_cfg *cfg = irq_cfg(irq); spin_lock_irqsave(&ioapic_lock, flags); entry = cfg->irq_2_pin; @@ -358,13 +361,12 @@ static void ioapic_mask_entry(int apic, int pin) } #ifdef CONFIG_SMP -static void __target_IO_APIC_irq(unsigned int irq, unsigned int dest, u8 vector) +static void __target_IO_APIC_irq(unsigned int irq, unsigned int dest, struct irq_cfg *cfg) { int apic, pin; - struct irq_cfg *cfg; struct irq_pin_list *entry; + u8 vector = cfg->vector; - cfg = irq_cfg(irq); entry = cfg->irq_2_pin; for (;;) { unsigned int reg; @@ -394,24 +396,27 @@ static void __target_IO_APIC_irq(unsigned int irq, unsigned int dest, u8 vector) } } -static int assign_irq_vector(int irq, cpumask_t mask); +static int assign_irq_vector(int irq, struct irq_cfg *cfg, cpumask_t mask); -static void set_ioapic_affinity_irq(unsigned int irq, cpumask_t mask) +static void set_ioapic_affinity_irq_desc(struct irq_desc *desc, cpumask_t mask) { struct irq_cfg *cfg; unsigned long flags; unsigned int dest; cpumask_t tmp; - struct irq_desc *desc; + unsigned int irq; cpus_and(tmp, mask, cpu_online_map); if (cpus_empty(tmp)) return; - cfg = irq_cfg(irq); - if (assign_irq_vector(irq, mask)) + irq = desc->irq; + cfg = desc->chip_data; + if (assign_irq_vector(irq, cfg, mask)) return; + set_extra_move_desc(desc, mask); + cpus_and(tmp, cfg->domain, mask); dest = cpu_mask_to_apicid(tmp); /* @@ -419,12 +424,20 @@ static void set_ioapic_affinity_irq(unsigned int irq, cpumask_t mask) */ dest = SET_APIC_LOGICAL_ID(dest); - desc = irq_to_desc(irq); spin_lock_irqsave(&ioapic_lock, flags); - __target_IO_APIC_irq(irq, dest, cfg->vector); + __target_IO_APIC_irq(irq, dest, cfg); desc->affinity = mask; spin_unlock_irqrestore(&ioapic_lock, flags); } + +static void set_ioapic_affinity_irq(unsigned int irq, cpumask_t mask) +{ + struct irq_desc *desc; + + desc = irq_to_desc(irq); + + set_ioapic_affinity_irq_desc(desc, mask); +} #endif /* CONFIG_SMP */ /* @@ -432,10 +445,9 @@ static void set_ioapic_affinity_irq(unsigned int irq, cpumask_t mask) * shared ISA-space IRQs, so we have to support them. We are super * fast in the common case, and fast for shared ISA-space IRQs. */ -static void add_pin_to_irq_cpu(unsigned int irq, int cpu, int apic, int pin) +static void add_pin_to_irq_cpu(struct irq_cfg *cfg, int cpu, int apic, int pin) { struct irq_pin_list *entry; - struct irq_cfg *cfg = irq_cfg(irq); entry = cfg->irq_2_pin; if (!entry) { @@ -468,11 +480,10 @@ static void add_pin_to_irq_cpu(unsigned int irq, int cpu, int apic, int pin) /* * Reroute an IRQ to a different pin. */ -static void __init replace_pin_at_irq(unsigned int irq, int cpu, +static void __init replace_pin_at_irq_cpu(struct irq_cfg *cfg, int cpu, int oldapic, int oldpin, int newapic, int newpin) { - struct irq_cfg *cfg = irq_cfg(irq); struct irq_pin_list *entry = cfg->irq_2_pin; int replaced = 0; @@ -489,18 +500,16 @@ static void __init replace_pin_at_irq(unsigned int irq, int cpu, /* why? call replace before add? */ if (!replaced) - add_pin_to_irq_cpu(irq, cpu, newapic, newpin); + add_pin_to_irq_cpu(cfg, cpu, newapic, newpin); } -static inline void io_apic_modify_irq(unsigned int irq, +static inline void io_apic_modify_irq(struct irq_cfg *cfg, int mask_and, int mask_or, void (*final)(struct irq_pin_list *entry)) { int pin; - struct irq_cfg *cfg; struct irq_pin_list *entry; - cfg = irq_cfg(irq); for (entry = cfg->irq_2_pin; entry != NULL; entry = entry->next) { unsigned int reg; pin = entry->pin; @@ -513,9 +522,9 @@ static inline void io_apic_modify_irq(unsigned int irq, } } -static void __unmask_IO_APIC_irq(unsigned int irq) +static void __unmask_IO_APIC_irq(struct irq_cfg *cfg) { - io_apic_modify_irq(irq, ~IO_APIC_REDIR_MASKED, 0, NULL); + io_apic_modify_irq(cfg, ~IO_APIC_REDIR_MASKED, 0, NULL); } #ifdef CONFIG_X86_64 @@ -530,47 +539,64 @@ void io_apic_sync(struct irq_pin_list *entry) readl(&io_apic->data); } -static void __mask_IO_APIC_irq(unsigned int irq) +static void __mask_IO_APIC_irq(struct irq_cfg *cfg) { - io_apic_modify_irq(irq, ~0, IO_APIC_REDIR_MASKED, &io_apic_sync); + io_apic_modify_irq(cfg, ~0, IO_APIC_REDIR_MASKED, &io_apic_sync); } #else /* CONFIG_X86_32 */ -static void __mask_IO_APIC_irq(unsigned int irq) +static void __mask_IO_APIC_irq(struct irq_cfg *cfg) { - io_apic_modify_irq(irq, ~0, IO_APIC_REDIR_MASKED, NULL); + io_apic_modify_irq(cfg, ~0, IO_APIC_REDIR_MASKED, NULL); } -static void __mask_and_edge_IO_APIC_irq(unsigned int irq) +static void __mask_and_edge_IO_APIC_irq(struct irq_cfg *cfg) { - io_apic_modify_irq(irq, ~IO_APIC_REDIR_LEVEL_TRIGGER, + io_apic_modify_irq(cfg, ~IO_APIC_REDIR_LEVEL_TRIGGER, IO_APIC_REDIR_MASKED, NULL); } -static void __unmask_and_level_IO_APIC_irq(unsigned int irq) +static void __unmask_and_level_IO_APIC_irq(struct irq_cfg *cfg) { - io_apic_modify_irq(irq, ~IO_APIC_REDIR_MASKED, + io_apic_modify_irq(cfg, ~IO_APIC_REDIR_MASKED, IO_APIC_REDIR_LEVEL_TRIGGER, NULL); } #endif /* CONFIG_X86_32 */ -static void mask_IO_APIC_irq (unsigned int irq) +static void mask_IO_APIC_irq_desc(struct irq_desc *desc) { + struct irq_cfg *cfg = desc->chip_data; unsigned long flags; + BUG_ON(!cfg); + spin_lock_irqsave(&ioapic_lock, flags); - __mask_IO_APIC_irq(irq); + __mask_IO_APIC_irq(cfg); spin_unlock_irqrestore(&ioapic_lock, flags); } -static void unmask_IO_APIC_irq (unsigned int irq) +static void unmask_IO_APIC_irq_desc(struct irq_desc *desc) { + struct irq_cfg *cfg = desc->chip_data; unsigned long flags; spin_lock_irqsave(&ioapic_lock, flags); - __unmask_IO_APIC_irq(irq); + __unmask_IO_APIC_irq(cfg); spin_unlock_irqrestore(&ioapic_lock, flags); } +static void mask_IO_APIC_irq(unsigned int irq) +{ + struct irq_desc *desc = irq_to_desc(irq); + + mask_IO_APIC_irq_desc(desc); +} +static void unmask_IO_APIC_irq(unsigned int irq) +{ + struct irq_desc *desc = irq_to_desc(irq); + + unmask_IO_APIC_irq_desc(desc); +} + static void clear_IO_APIC_pin(unsigned int apic, unsigned int pin) { struct IO_APIC_route_entry entry; @@ -1072,7 +1098,7 @@ void unlock_vector_lock(void) spin_unlock(&vector_lock); } -static int __assign_irq_vector(int irq, cpumask_t mask) +static int __assign_irq_vector(int irq, struct irq_cfg *cfg, cpumask_t mask) { /* * NOTE! The local APIC isn't very good at handling @@ -1088,16 +1114,13 @@ static int __assign_irq_vector(int irq, cpumask_t mask) static int current_vector = FIRST_DEVICE_VECTOR, current_offset = 0; unsigned int old_vector; int cpu; - struct irq_cfg *cfg; - cfg = irq_cfg(irq); + if ((cfg->move_in_progress) || cfg->move_cleanup_count) + return -EBUSY; /* Only try and allocate irqs on cpus that are present */ cpus_and(mask, mask, cpu_online_map); - if ((cfg->move_in_progress) || cfg->move_cleanup_count) - return -EBUSY; - old_vector = cfg->vector; if (old_vector) { cpumask_t tmp; @@ -1151,24 +1174,22 @@ next: return -ENOSPC; } -static int assign_irq_vector(int irq, cpumask_t mask) +static int assign_irq_vector(int irq, struct irq_cfg *cfg, cpumask_t mask) { int err; unsigned long flags; spin_lock_irqsave(&vector_lock, flags); - err = __assign_irq_vector(irq, mask); + err = __assign_irq_vector(irq, cfg, mask); spin_unlock_irqrestore(&vector_lock, flags); return err; } -static void __clear_irq_vector(int irq) +static void __clear_irq_vector(int irq, struct irq_cfg *cfg) { - struct irq_cfg *cfg; cpumask_t mask; int cpu, vector; - cfg = irq_cfg(irq); BUG_ON(!cfg->vector); vector = cfg->vector; @@ -1257,11 +1278,8 @@ static inline int IO_APIC_irq_trigger(int irq) } #endif -static void ioapic_register_intr(int irq, unsigned long trigger) +static void ioapic_register_intr(int irq, struct irq_desc *desc, unsigned long trigger) { - struct irq_desc *desc; - - desc = irq_to_desc(irq); if ((trigger == IOAPIC_AUTO && IO_APIC_irq_trigger(irq)) || trigger == IOAPIC_LEVEL) @@ -1353,7 +1371,7 @@ static int setup_ioapic_entry(int apic, int irq, return 0; } -static void setup_IO_APIC_irq(int apic, int pin, unsigned int irq, +static void setup_IO_APIC_irq(int apic, int pin, unsigned int irq, struct irq_desc *desc, int trigger, int polarity) { struct irq_cfg *cfg; @@ -1363,10 +1381,10 @@ static void setup_IO_APIC_irq(int apic, int pin, unsigned int irq, if (!IO_APIC_IRQ(irq)) return; - cfg = irq_cfg(irq); + cfg = desc->chip_data; mask = TARGET_CPUS; - if (assign_irq_vector(irq, mask)) + if (assign_irq_vector(irq, cfg, mask)) return; cpus_and(mask, cfg->domain, mask); @@ -1383,11 +1401,11 @@ static void setup_IO_APIC_irq(int apic, int pin, unsigned int irq, cfg->vector)) { printk("Failed to setup ioapic entry for ioapic %d, pin %d\n", mp_ioapics[apic].mp_apicid, pin); - __clear_irq_vector(irq); + __clear_irq_vector(irq, cfg); return; } - ioapic_register_intr(irq, trigger); + ioapic_register_intr(irq, desc, trigger); if (irq < NR_IRQS_LEGACY) disable_8259A_irq(irq); @@ -1399,6 +1417,7 @@ static void __init setup_IO_APIC_irqs(void) int apic, pin, idx, irq; int notcon = 0; struct irq_desc *desc; + struct irq_cfg *cfg; int cpu = boot_cpu_id; apic_printk(APIC_VERBOSE, KERN_DEBUG "init IO_APIC IRQs\n"); @@ -1436,9 +1455,10 @@ static void __init setup_IO_APIC_irqs(void) printk(KERN_INFO "can not get irq_desc for %d\n", irq); continue; } - add_pin_to_irq_cpu(irq, cpu, apic, pin); + cfg = desc->chip_data; + add_pin_to_irq_cpu(cfg, cpu, apic, pin); - setup_IO_APIC_irq(apic, pin, irq, + setup_IO_APIC_irq(apic, pin, irq, desc, irq_trigger(idx), irq_polarity(idx)); } } @@ -2086,7 +2106,7 @@ static unsigned int startup_ioapic_irq(unsigned int irq) was_pending = 1; } cfg = irq_cfg(irq); - __unmask_IO_APIC_irq(irq); + __unmask_IO_APIC_irq(cfg); spin_unlock_irqrestore(&ioapic_lock, flags); return was_pending; @@ -2149,35 +2169,37 @@ static DECLARE_DELAYED_WORK(ir_migration_work, ir_irq_migration); * as simple as edge triggered migration and we can do the irq migration * with a simple atomic update to IO-APIC RTE. */ -static void migrate_ioapic_irq(int irq, cpumask_t mask) +static void migrate_ioapic_irq_desc(struct irq_desc *desc, cpumask_t mask) { struct irq_cfg *cfg; - struct irq_desc *desc; cpumask_t tmp, cleanup_mask; struct irte irte; int modify_ioapic_rte; unsigned int dest; unsigned long flags; + unsigned int irq; cpus_and(tmp, mask, cpu_online_map); if (cpus_empty(tmp)) return; + irq = desc->irq; if (get_irte(irq, &irte)) return; - if (assign_irq_vector(irq, mask)) + cfg = desc->chip_data; + if (assign_irq_vector(irq, cfg, mask)) return; - cfg = irq_cfg(irq); + set_extra_move_desc(desc, mask); + cpus_and(tmp, cfg->domain, mask); dest = cpu_mask_to_apicid(tmp); - desc = irq_to_desc(irq); modify_ioapic_rte = desc->status & IRQ_LEVEL; if (modify_ioapic_rte) { spin_lock_irqsave(&ioapic_lock, flags); - __target_IO_APIC_irq(irq, dest, cfg->vector); + __target_IO_APIC_irq(irq, dest, cfg); spin_unlock_irqrestore(&ioapic_lock, flags); } @@ -2199,14 +2221,14 @@ static void migrate_ioapic_irq(int irq, cpumask_t mask) desc->affinity = mask; } -static int migrate_irq_remapped_level(int irq) +static int migrate_irq_remapped_level_desc(struct irq_desc *desc) { int ret = -1; - struct irq_desc *desc = irq_to_desc(irq); + struct irq_cfg *cfg = desc->chip_data; - mask_IO_APIC_irq(irq); + mask_IO_APIC_irq_desc(desc); - if (io_apic_level_ack_pending(irq)) { + if (io_apic_level_ack_pending(cfg)) { /* * Interrupt in progress. Migrating irq now will change the * vector information in the IO-APIC RTE and that will confuse @@ -2218,14 +2240,15 @@ static int migrate_irq_remapped_level(int irq) } /* everthing is clear. we have right of way */ - migrate_ioapic_irq(irq, desc->pending_mask); + migrate_ioapic_irq_desc(desc, desc->pending_mask); ret = 0; desc->status &= ~IRQ_MOVE_PENDING; cpus_clear(desc->pending_mask); unmask: - unmask_IO_APIC_irq(irq); + unmask_IO_APIC_irq_desc(desc); + return ret; } @@ -2258,18 +2281,22 @@ static void ir_irq_migration(struct work_struct *work) /* * Migrates the IRQ destination in the process context. */ -static void set_ir_ioapic_affinity_irq(unsigned int irq, cpumask_t mask) +static void set_ir_ioapic_affinity_irq_desc(struct irq_desc *desc, cpumask_t mask) { - struct irq_desc *desc = irq_to_desc(irq); - if (desc->status & IRQ_LEVEL) { desc->status |= IRQ_MOVE_PENDING; desc->pending_mask = mask; - migrate_irq_remapped_level(irq); + migrate_irq_remapped_level_desc(desc); return; } - migrate_ioapic_irq(irq, mask); + migrate_ioapic_irq_desc(desc, mask); +} +static void set_ir_ioapic_affinity_irq(unsigned int irq, cpumask_t mask) +{ + struct irq_desc *desc = irq_to_desc(irq); + + set_ir_ioapic_affinity_irq_desc(desc, mask); } #endif @@ -2313,9 +2340,10 @@ unlock: irq_exit(); } -static void irq_complete_move(unsigned int irq) +static void irq_complete_move(struct irq_desc **descp) { - struct irq_cfg *cfg = irq_cfg(irq); + struct irq_desc *desc = *descp; + struct irq_cfg *cfg = desc->chip_data; unsigned vector, me; if (likely(!cfg->move_in_progress)) @@ -2333,8 +2361,9 @@ static void irq_complete_move(unsigned int irq) } } #else -static inline void irq_complete_move(unsigned int irq) {} +static inline void irq_complete_move(struct irq_desc **descp) {} #endif + #ifdef CONFIG_INTR_REMAP static void ack_x2apic_level(unsigned int irq) { @@ -2345,11 +2374,14 @@ static void ack_x2apic_edge(unsigned int irq) { ack_x2APIC_irq(); } + #endif static void ack_apic_edge(unsigned int irq) { - irq_complete_move(irq); + struct irq_desc *desc = irq_to_desc(irq); + + irq_complete_move(&desc); move_native_irq(irq); ack_APIC_irq(); } @@ -2358,18 +2390,21 @@ atomic_t irq_mis_count; static void ack_apic_level(unsigned int irq) { + struct irq_desc *desc = irq_to_desc(irq); + #ifdef CONFIG_X86_32 unsigned long v; int i; #endif + struct irq_cfg *cfg; int do_unmask_irq = 0; - irq_complete_move(irq); + irq_complete_move(&desc); #ifdef CONFIG_GENERIC_PENDING_IRQ /* If we are moving the irq we need to mask it */ - if (unlikely(irq_to_desc(irq)->status & IRQ_MOVE_PENDING)) { + if (unlikely(desc->status & IRQ_MOVE_PENDING)) { do_unmask_irq = 1; - mask_IO_APIC_irq(irq); + mask_IO_APIC_irq_desc(desc); } #endif @@ -2393,7 +2428,8 @@ static void ack_apic_level(unsigned int irq) * operation to prevent an edge-triggered interrupt escaping meanwhile. * The idea is from Manfred Spraul. --macro */ - i = irq_cfg(irq)->vector; + cfg = desc->chip_data; + i = cfg->vector; v = apic_read(APIC_TMR + ((i & ~0x1f) >> 1)); #endif @@ -2432,17 +2468,18 @@ static void ack_apic_level(unsigned int irq) * accurate and is causing problems then it is a hardware bug * and you can go talk to the chipset vendor about it. */ - if (!io_apic_level_ack_pending(irq)) + cfg = desc->chip_data; + if (!io_apic_level_ack_pending(cfg)) move_masked_irq(irq); - unmask_IO_APIC_irq(irq); + unmask_IO_APIC_irq_desc(desc); } #ifdef CONFIG_X86_32 if (!(v & (1 << (i & 0x1f)))) { atomic_inc(&irq_mis_count); spin_lock(&ioapic_lock); - __mask_and_edge_IO_APIC_irq(irq); - __unmask_and_level_IO_APIC_irq(irq); + __mask_and_edge_IO_APIC_irq(cfg); + __unmask_and_level_IO_APIC_irq(cfg); spin_unlock(&ioapic_lock); } #endif @@ -2533,7 +2570,7 @@ static void unmask_lapic_irq(unsigned int irq) apic_write(APIC_LVT0, v & ~APIC_LVT_MASKED); } -static void ack_lapic_irq (unsigned int irq) +static void ack_lapic_irq(unsigned int irq) { ack_APIC_irq(); } @@ -2545,11 +2582,8 @@ static struct irq_chip lapic_chip __read_mostly = { .ack = ack_lapic_irq, }; -static void lapic_register_intr(int irq) +static void lapic_register_intr(int irq, struct irq_desc *desc) { - struct irq_desc *desc; - - desc = irq_to_desc(irq); desc->status &= ~IRQ_LEVEL; set_irq_chip_and_handler_name(irq, &lapic_chip, handle_edge_irq, "edge"); @@ -2653,7 +2687,9 @@ int timer_through_8259 __initdata; */ static inline void __init check_timer(void) { - struct irq_cfg *cfg = irq_cfg(0); + struct irq_desc *desc = irq_to_desc(0); + struct irq_cfg *cfg = desc->chip_data; + int cpu = boot_cpu_id; int apic1, pin1, apic2, pin2; unsigned long flags; unsigned int ver; @@ -2668,7 +2704,7 @@ static inline void __init check_timer(void) * get/set the timer IRQ vector: */ disable_8259A_irq(0); - assign_irq_vector(0, TARGET_CPUS); + assign_irq_vector(0, cfg, TARGET_CPUS); /* * As IRQ0 is to be enabled in the 8259A, the virtual @@ -2719,10 +2755,10 @@ static inline void __init check_timer(void) * Ok, does IRQ0 through the IOAPIC work? */ if (no_pin1) { - add_pin_to_irq_cpu(0, boot_cpu_id, apic1, pin1); + add_pin_to_irq_cpu(cfg, cpu, apic1, pin1); setup_timer_IRQ0_pin(apic1, pin1, cfg->vector); } - unmask_IO_APIC_irq(0); + unmask_IO_APIC_irq_desc(desc); if (timer_irq_works()) { if (nmi_watchdog == NMI_IO_APIC) { setup_nmi(); @@ -2748,9 +2784,9 @@ static inline void __init check_timer(void) /* * legacy devices should be connected to IO APIC #0 */ - replace_pin_at_irq(0, boot_cpu_id, apic1, pin1, apic2, pin2); + replace_pin_at_irq_cpu(cfg, cpu, apic1, pin1, apic2, pin2); setup_timer_IRQ0_pin(apic2, pin2, cfg->vector); - unmask_IO_APIC_irq(0); + unmask_IO_APIC_irq_desc(desc); enable_8259A_irq(0); if (timer_irq_works()) { apic_printk(APIC_QUIET, KERN_INFO "....... works.\n"); @@ -2782,7 +2818,7 @@ static inline void __init check_timer(void) apic_printk(APIC_QUIET, KERN_INFO "...trying to set up timer as Virtual Wire IRQ...\n"); - lapic_register_intr(0); + lapic_register_intr(0, desc); apic_write(APIC_LVT0, APIC_DM_FIXED | cfg->vector); /* Fixed mode */ enable_8259A_irq(0); @@ -2986,7 +3022,7 @@ unsigned int create_irq_nr(unsigned int irq_want) if (cfg_new->vector != 0) continue; - if (__assign_irq_vector(new, TARGET_CPUS) == 0) + if (__assign_irq_vector(new, cfg_new, TARGET_CPUS) == 0) irq = new; break; } @@ -3034,7 +3070,7 @@ void destroy_irq(unsigned int irq) free_irte(irq); #endif spin_lock_irqsave(&vector_lock, flags); - __clear_irq_vector(irq); + __clear_irq_vector(irq, cfg); spin_unlock_irqrestore(&vector_lock, flags); } @@ -3049,12 +3085,12 @@ static int msi_compose_msg(struct pci_dev *pdev, unsigned int irq, struct msi_ms unsigned dest; cpumask_t tmp; + cfg = irq_cfg(irq); tmp = TARGET_CPUS; - err = assign_irq_vector(irq, tmp); + err = assign_irq_vector(irq, cfg, tmp); if (err) return err; - cfg = irq_cfg(irq); cpus_and(tmp, cfg->domain, tmp); dest = cpu_mask_to_apicid(tmp); @@ -3112,35 +3148,35 @@ static int msi_compose_msg(struct pci_dev *pdev, unsigned int irq, struct msi_ms #ifdef CONFIG_SMP static void set_msi_irq_affinity(unsigned int irq, cpumask_t mask) { + struct irq_desc *desc = irq_to_desc(irq); struct irq_cfg *cfg; struct msi_msg msg; unsigned int dest; cpumask_t tmp; - struct irq_desc *desc; cpus_and(tmp, mask, cpu_online_map); if (cpus_empty(tmp)) return; - if (assign_irq_vector(irq, mask)) + cfg = desc->chip_data; + if (assign_irq_vector(irq, cfg, mask)) return; - cfg = irq_cfg(irq); + set_extra_move_desc(desc, mask); + cpus_and(tmp, cfg->domain, mask); dest = cpu_mask_to_apicid(tmp); - read_msi_msg(irq, &msg); + read_msi_msg_desc(desc, &msg); msg.data &= ~MSI_DATA_VECTOR_MASK; msg.data |= MSI_DATA_VECTOR(cfg->vector); msg.address_lo &= ~MSI_ADDR_DEST_ID_MASK; msg.address_lo |= MSI_ADDR_DEST_ID(dest); - write_msi_msg(irq, &msg); - desc = irq_to_desc(irq); + write_msi_msg_desc(desc, &msg); desc->affinity = mask; } - #ifdef CONFIG_INTR_REMAP /* * Migrate the MSI irq to another cpumask. This migration is @@ -3148,11 +3184,11 @@ static void set_msi_irq_affinity(unsigned int irq, cpumask_t mask) */ static void ir_set_msi_irq_affinity(unsigned int irq, cpumask_t mask) { + struct irq_desc *desc = irq_to_desc(irq); struct irq_cfg *cfg; unsigned int dest; cpumask_t tmp, cleanup_mask; struct irte irte; - struct irq_desc *desc; cpus_and(tmp, mask, cpu_online_map); if (cpus_empty(tmp)) @@ -3161,10 +3197,12 @@ static void ir_set_msi_irq_affinity(unsigned int irq, cpumask_t mask) if (get_irte(irq, &irte)) return; - if (assign_irq_vector(irq, mask)) + cfg = desc->chip_data; + if (assign_irq_vector(irq, cfg, mask)) return; - cfg = irq_cfg(irq); + set_extra_move_desc(desc, mask); + cpus_and(tmp, cfg->domain, mask); dest = cpu_mask_to_apicid(tmp); @@ -3188,9 +3226,9 @@ static void ir_set_msi_irq_affinity(unsigned int irq, cpumask_t mask) cfg->move_in_progress = 0; } - desc = irq_to_desc(irq); desc->affinity = mask; } + #endif #endif /* CONFIG_SMP */ @@ -3249,7 +3287,7 @@ static int msi_alloc_irte(struct pci_dev *dev, int irq, int nvec) } #endif -static int setup_msi_irq(struct pci_dev *dev, struct msi_desc *desc, int irq) +static int setup_msi_irq(struct pci_dev *dev, struct msi_desc *msidesc, int irq) { int ret; struct msi_msg msg; @@ -3258,7 +3296,7 @@ static int setup_msi_irq(struct pci_dev *dev, struct msi_desc *desc, int irq) if (ret < 0) return ret; - set_irq_msi(irq, desc); + set_irq_msi(irq, msidesc); write_msi_msg(irq, &msg); #ifdef CONFIG_INTR_REMAP @@ -3381,20 +3419,22 @@ void arch_teardown_msi_irq(unsigned int irq) #ifdef CONFIG_SMP static void dmar_msi_set_affinity(unsigned int irq, cpumask_t mask) { + struct irq_desc *desc = irq_to_desc(irq); struct irq_cfg *cfg; struct msi_msg msg; unsigned int dest; cpumask_t tmp; - struct irq_desc *desc; cpus_and(tmp, mask, cpu_online_map); if (cpus_empty(tmp)) return; - if (assign_irq_vector(irq, mask)) + cfg = desc->chip_data; + if (assign_irq_vector(irq, cfg, mask)) return; - cfg = irq_cfg(irq); + set_extra_move_desc(desc, mask); + cpus_and(tmp, cfg->domain, mask); dest = cpu_mask_to_apicid(tmp); @@ -3406,9 +3446,9 @@ static void dmar_msi_set_affinity(unsigned int irq, cpumask_t mask) msg.address_lo |= MSI_ADDR_DEST_ID(dest); dmar_msi_write(irq, &msg); - desc = irq_to_desc(irq); desc->affinity = mask; } + #endif /* CONFIG_SMP */ struct irq_chip dmar_msi_type = { @@ -3442,8 +3482,8 @@ int arch_setup_dmar_msi(unsigned int irq) #ifdef CONFIG_SMP static void hpet_msi_set_affinity(unsigned int irq, cpumask_t mask) { + struct irq_desc *desc = irq_to_desc(irq); struct irq_cfg *cfg; - struct irq_desc *desc; struct msi_msg msg; unsigned int dest; cpumask_t tmp; @@ -3452,10 +3492,12 @@ static void hpet_msi_set_affinity(unsigned int irq, cpumask_t mask) if (cpus_empty(tmp)) return; - if (assign_irq_vector(irq, mask)) + cfg = desc->chip_data; + if (assign_irq_vector(irq, cfg, mask)) return; - cfg = irq_cfg(irq); + set_extra_move_desc(desc, mask); + cpus_and(tmp, cfg->domain, mask); dest = cpu_mask_to_apicid(tmp); @@ -3467,9 +3509,9 @@ static void hpet_msi_set_affinity(unsigned int irq, cpumask_t mask) msg.address_lo |= MSI_ADDR_DEST_ID(dest); hpet_msi_write(irq, &msg); - desc = irq_to_desc(irq); desc->affinity = mask; } + #endif /* CONFIG_SMP */ struct irq_chip hpet_msi_type = { @@ -3524,26 +3566,28 @@ static void target_ht_irq(unsigned int irq, unsigned int dest, u8 vector) static void set_ht_irq_affinity(unsigned int irq, cpumask_t mask) { + struct irq_desc *desc = irq_to_desc(irq); struct irq_cfg *cfg; unsigned int dest; cpumask_t tmp; - struct irq_desc *desc; cpus_and(tmp, mask, cpu_online_map); if (cpus_empty(tmp)) return; - if (assign_irq_vector(irq, mask)) + cfg = desc->chip_data; + if (assign_irq_vector(irq, cfg, mask)) return; - cfg = irq_cfg(irq); + set_extra_move_desc(desc, mask); + cpus_and(tmp, cfg->domain, mask); dest = cpu_mask_to_apicid(tmp); target_ht_irq(irq, dest, cfg->vector); - desc = irq_to_desc(irq); desc->affinity = mask; } + #endif static struct irq_chip ht_irq_chip = { @@ -3563,13 +3607,13 @@ int arch_setup_ht_irq(unsigned int irq, struct pci_dev *dev) int err; cpumask_t tmp; + cfg = irq_cfg(irq); tmp = TARGET_CPUS; - err = assign_irq_vector(irq, tmp); + err = assign_irq_vector(irq, cfg, tmp); if (!err) { struct ht_irq_msg msg; unsigned dest; - cfg = irq_cfg(irq); cpus_and(tmp, cfg->domain, tmp); dest = cpu_mask_to_apicid(tmp); @@ -3615,7 +3659,9 @@ int arch_enable_uv_irq(char *irq_name, unsigned int irq, int cpu, int mmr_blade, unsigned long flags; int err; - err = assign_irq_vector(irq, *eligible_cpu); + cfg = irq_cfg(irq); + + err = assign_irq_vector(irq, cfg, *eligible_cpu); if (err != 0) return err; @@ -3624,8 +3670,6 @@ int arch_enable_uv_irq(char *irq_name, unsigned int irq, int cpu, int mmr_blade, irq_name); spin_unlock_irqrestore(&vector_lock, flags); - cfg = irq_cfg(irq); - mmr_value = 0; entry = (struct uv_IO_APIC_route_entry *)&mmr_value; BUG_ON(sizeof(struct uv_IO_APIC_route_entry) != sizeof(unsigned long)); @@ -3806,10 +3850,10 @@ int io_apic_set_pci_routing (int ioapic, int pin, int irq, int triggering, int p */ if (irq >= NR_IRQS_LEGACY) { cfg = desc->chip_data; - add_pin_to_irq_cpu(irq, cpu, ioapic, pin); + add_pin_to_irq_cpu(cfg, cpu, ioapic, pin); } - setup_IO_APIC_irq(ioapic, pin, irq, triggering, polarity); + setup_IO_APIC_irq(ioapic, pin, irq, desc, triggering, polarity); return 0; } @@ -3866,7 +3910,7 @@ void __init setup_ioapic_dest(void) desc = irq_to_desc(irq); cfg = desc->chip_data; if (!cfg->vector) { - setup_IO_APIC_irq(ioapic, pin, irq, + setup_IO_APIC_irq(ioapic, pin, irq, desc, irq_trigger(irq_entry), irq_polarity(irq_entry)); continue; @@ -3884,10 +3928,10 @@ void __init setup_ioapic_dest(void) #ifdef CONFIG_INTR_REMAP if (intr_remapping_enabled) - set_ir_ioapic_affinity_irq(irq, mask); + set_ir_ioapic_affinity_irq_desc(desc, mask); else #endif - set_ioapic_affinity_irq(irq, mask); + set_ioapic_affinity_irq_desc(desc, mask); } } diff --git a/drivers/pci/msi.c b/drivers/pci/msi.c index 74801f7df9c9..11a51f8ed3b3 100644 --- a/drivers/pci/msi.c +++ b/drivers/pci/msi.c @@ -103,11 +103,11 @@ static void msix_set_enable(struct pci_dev *dev, int enable) } } -static void msix_flush_writes(unsigned int irq) +static void msix_flush_writes(struct irq_desc *desc) { struct msi_desc *entry; - entry = get_irq_msi(irq); + entry = get_irq_desc_msi(desc); BUG_ON(!entry || !entry->dev); switch (entry->msi_attrib.type) { case PCI_CAP_ID_MSI: @@ -135,11 +135,11 @@ static void msix_flush_writes(unsigned int irq) * Returns 1 if it succeeded in masking the interrupt and 0 if the device * doesn't support MSI masking. */ -static int msi_set_mask_bits(unsigned int irq, u32 mask, u32 flag) +static int msi_set_mask_bits(struct irq_desc *desc, u32 mask, u32 flag) { struct msi_desc *entry; - entry = get_irq_msi(irq); + entry = get_irq_desc_msi(desc); BUG_ON(!entry || !entry->dev); switch (entry->msi_attrib.type) { case PCI_CAP_ID_MSI: @@ -172,9 +172,9 @@ static int msi_set_mask_bits(unsigned int irq, u32 mask, u32 flag) return 1; } -void read_msi_msg(unsigned int irq, struct msi_msg *msg) +void read_msi_msg_desc(struct irq_desc *desc, struct msi_msg *msg) { - struct msi_desc *entry = get_irq_msi(irq); + struct msi_desc *entry = get_irq_desc_msi(desc); switch(entry->msi_attrib.type) { case PCI_CAP_ID_MSI: { @@ -211,9 +211,16 @@ void read_msi_msg(unsigned int irq, struct msi_msg *msg) } } -void write_msi_msg(unsigned int irq, struct msi_msg *msg) +void read_msi_msg(unsigned int irq, struct msi_msg *msg) { - struct msi_desc *entry = get_irq_msi(irq); + struct irq_desc *desc = irq_to_desc(irq); + + read_msi_msg_desc(desc, msg); +} + +void write_msi_msg_desc(struct irq_desc *desc, struct msi_msg *msg) +{ + struct msi_desc *entry = get_irq_desc_msi(desc); switch (entry->msi_attrib.type) { case PCI_CAP_ID_MSI: { @@ -252,21 +259,31 @@ void write_msi_msg(unsigned int irq, struct msi_msg *msg) entry->msg = *msg; } +void write_msi_msg(unsigned int irq, struct msi_msg *msg) +{ + struct irq_desc *desc = irq_to_desc(irq); + + write_msi_msg_desc(desc, msg); +} + void mask_msi_irq(unsigned int irq) { - msi_set_mask_bits(irq, 1, 1); - msix_flush_writes(irq); + struct irq_desc *desc = irq_to_desc(irq); + + msi_set_mask_bits(desc, 1, 1); + msix_flush_writes(desc); } void unmask_msi_irq(unsigned int irq) { - msi_set_mask_bits(irq, 1, 0); - msix_flush_writes(irq); + struct irq_desc *desc = irq_to_desc(irq); + + msi_set_mask_bits(desc, 1, 0); + msix_flush_writes(desc); } static int msi_free_irqs(struct pci_dev* dev); - static struct msi_desc* alloc_msi_entry(void) { struct msi_desc *entry; @@ -303,9 +320,11 @@ static void __pci_restore_msi_state(struct pci_dev *dev) pci_intx_for_msi(dev, 0); msi_set_enable(dev, 0); write_msi_msg(dev->irq, &entry->msg); - if (entry->msi_attrib.maskbit) - msi_set_mask_bits(dev->irq, entry->msi_attrib.maskbits_mask, + if (entry->msi_attrib.maskbit) { + struct irq_desc *desc = irq_to_desc(dev->irq); + msi_set_mask_bits(desc, entry->msi_attrib.maskbits_mask, entry->msi_attrib.masked); + } pci_read_config_word(dev, pos + PCI_MSI_FLAGS, &control); control &= ~PCI_MSI_FLAGS_QSIZE; @@ -327,8 +346,9 @@ static void __pci_restore_msix_state(struct pci_dev *dev) msix_set_enable(dev, 0); list_for_each_entry(entry, &dev->msi_list, list) { + struct irq_desc *desc = irq_to_desc(entry->irq); write_msi_msg(entry->irq, &entry->msg); - msi_set_mask_bits(entry->irq, 1, entry->msi_attrib.masked); + msi_set_mask_bits(desc, 1, entry->msi_attrib.masked); } BUG_ON(list_empty(&dev->msi_list)); @@ -596,7 +616,8 @@ void pci_msi_shutdown(struct pci_dev* dev) /* Return the the pci reset with msi irqs unmasked */ if (entry->msi_attrib.maskbit) { u32 mask = entry->msi_attrib.maskbits_mask; - msi_set_mask_bits(dev->irq, mask, ~mask); + struct irq_desc *desc = irq_to_desc(dev->irq); + msi_set_mask_bits(desc, mask, ~mask); } if (!entry->dev || entry->msi_attrib.type != PCI_CAP_ID_MSI) return; diff --git a/include/linux/msi.h b/include/linux/msi.h index 8f2939227207..d2b8a1e8ca11 100644 --- a/include/linux/msi.h +++ b/include/linux/msi.h @@ -10,8 +10,11 @@ struct msi_msg { }; /* Helper functions */ +struct irq_desc; extern void mask_msi_irq(unsigned int irq); extern void unmask_msi_irq(unsigned int irq); +extern void read_msi_msg_desc(struct irq_desc *desc, struct msi_msg *msg); +extern void write_msi_msg_desc(struct irq_desc *desc, struct msi_msg *msg); extern void read_msi_msg(unsigned int irq, struct msi_msg *msg); extern void write_msi_msg(unsigned int irq, struct msi_msg *msg); -- cgit v1.2.3 From 69b88afa8d114a43a3c0431722b79e31d9920692 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Fri, 5 Dec 2008 22:45:50 -0800 Subject: x86: clean up get_smp_config() Impact: cleanup reorder exit path in __get_smp_config(). also move two print outs to acpi_process_madt Signed-off-by: Yinghai Lu Signed-off-by: Ingo Molnar --- arch/x86/kernel/acpi/boot.c | 11 +++++++++++ arch/x86/kernel/mpparse.c | 25 +++++++++++-------------- 2 files changed, 22 insertions(+), 14 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c index 4c51a2f8fd31..65d0b72777ea 100644 --- a/arch/x86/kernel/acpi/boot.c +++ b/arch/x86/kernel/acpi/boot.c @@ -1360,6 +1360,17 @@ static void __init acpi_process_madt(void) disable_acpi(); } } + + /* + * ACPI supports both logical (e.g. Hyper-Threading) and physical + * processors, where MPS only supports physical. + */ + if (acpi_lapic && acpi_ioapic) + printk(KERN_INFO "Using ACPI (MADT) for SMP configuration " + "information\n"); + else if (acpi_lapic) + printk(KERN_INFO "Using ACPI for processor (LAPIC) " + "configuration information\n"); #endif return; } diff --git a/arch/x86/kernel/mpparse.c b/arch/x86/kernel/mpparse.c index 0f4c1fd5a1f4..45e3b69808ba 100644 --- a/arch/x86/kernel/mpparse.c +++ b/arch/x86/kernel/mpparse.c @@ -586,26 +586,23 @@ static void __init __get_smp_config(unsigned int early) { struct intel_mp_floating *mpf = mpf_found; - if (x86_quirks->mach_get_smp_config) { - if (x86_quirks->mach_get_smp_config(early)) - return; - } + if (!mpf) + return; + if (acpi_lapic && early) return; + /* - * ACPI supports both logical (e.g. Hyper-Threading) and physical - * processors, where MPS only supports physical. + * MPS doesn't support hyperthreading, aka only have + * thread 0 apic id in MPS table */ - if (acpi_lapic && acpi_ioapic) { - printk(KERN_INFO "Using ACPI (MADT) for SMP configuration " - "information\n"); + if (acpi_lapic && acpi_ioapic) return; - } else if (acpi_lapic) - printk(KERN_INFO "Using ACPI for processor (LAPIC) " - "configuration information\n"); - if (!mpf) - return; + if (x86_quirks->mach_get_smp_config) { + if (x86_quirks->mach_get_smp_config(early)) + return; + } printk(KERN_INFO "Intel MultiProcessor Specification v1.%d\n", mpf->mpf_specification); -- cgit v1.2.3 From 8b96f0119818964e4944fd1c423bf6770027d3ac Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 6 Dec 2008 03:40:00 +0100 Subject: tracing/function-graph-tracer: introduce __notrace_funcgraph to filter special functions Impact: trace more functions When the function graph tracer is configured, three more files are not traced to prevent only four functions to be traced. And this impacts the normal function tracer too. arch/x86/kernel/process_64/32.c: I had crashes when I let this file traced. After some debugging, I saw that the "current" task point was changed inside__swtich_to(), ie: "write_pda(pcurrent, next_p);" inside process_64.c Since the tracer store the original return address of the function inside current, we had crashes. Only __switch_to() has to be excluded from tracing. kernel/module.c and kernel/extable.c: Because of a function used internally by the function graph tracer: __kernel_text_address() To let the other functions inside these files to be traced, this patch introduces the __notrace_funcgraph function prefix which is __notrace if function graph tracer is configured and nothing if not. Signed-off-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- arch/x86/kernel/Makefile | 6 ------ arch/x86/kernel/process_32.c | 4 +++- arch/x86/kernel/process_64.c | 4 +++- include/linux/ftrace.h | 11 +++++++++++ kernel/Makefile | 4 ---- kernel/extable.c | 5 +++-- kernel/module.c | 2 +- 7 files changed, 21 insertions(+), 15 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index a3049da61985..1cad9318d217 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -14,12 +14,6 @@ CFLAGS_REMOVE_paravirt-spinlocks.o = -pg CFLAGS_REMOVE_ftrace.o = -pg endif -ifdef CONFIG_FUNCTION_GRAPH_TRACER -# Don't trace __switch_to() but let it for function tracer -CFLAGS_REMOVE_process_32.o = -pg -CFLAGS_REMOVE_process_64.o = -pg -endif - # # vsyscalls (which work on the user stack) should have # no stack-protector checks: diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c index 0a1302fe6d45..24c2276aa453 100644 --- a/arch/x86/kernel/process_32.c +++ b/arch/x86/kernel/process_32.c @@ -38,6 +38,7 @@ #include #include #include +#include #include #include @@ -548,7 +549,8 @@ __switch_to_xtra(struct task_struct *prev_p, struct task_struct *next_p, * the task-switch, and shows up in ret_from_fork in entry.S, * for example. */ -struct task_struct * __switch_to(struct task_struct *prev_p, struct task_struct *next_p) +__notrace_funcgraph struct task_struct * +__switch_to(struct task_struct *prev_p, struct task_struct *next_p) { struct thread_struct *prev = &prev_p->thread, *next = &next_p->thread; diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c index c958120fb1b6..fbb321d53d34 100644 --- a/arch/x86/kernel/process_64.c +++ b/arch/x86/kernel/process_64.c @@ -39,6 +39,7 @@ #include #include #include +#include #include #include @@ -551,8 +552,9 @@ static inline void __switch_to_xtra(struct task_struct *prev_p, * - could test fs/gs bitsliced * * Kprobes not supported here. Set the probe on schedule instead. + * Function graph tracer not supported too. */ -struct task_struct * +__notrace_funcgraph struct task_struct * __switch_to(struct task_struct *prev_p, struct task_struct *next_p) { struct thread_struct *prev = &prev_p->thread; diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index b9b4d0a22d10..449fa8e9e34f 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -369,6 +369,14 @@ struct ftrace_graph_ret { }; #ifdef CONFIG_FUNCTION_GRAPH_TRACER + +/* + * Sometimes we don't want to trace a function with the function + * graph tracer but we want them to keep traced by the usual function + * tracer if the function graph tracer is not configured. + */ +#define __notrace_funcgraph notrace + #define FTRACE_RETFUNC_DEPTH 50 #define FTRACE_RETSTACK_ALLOC_SIZE 32 /* Type of the callback handlers for tracing function graph*/ @@ -394,6 +402,9 @@ static inline int task_curr_ret_stack(struct task_struct *t) return t->curr_ret_stack; } #else + +#define __notrace_funcgraph + static inline void ftrace_graph_init_task(struct task_struct *t) { } static inline void ftrace_graph_exit_task(struct task_struct *t) { } diff --git a/kernel/Makefile b/kernel/Makefile index 703cf3b7389c..19fad003b19d 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -21,10 +21,6 @@ CFLAGS_REMOVE_cgroup-debug.o = -pg CFLAGS_REMOVE_sched_clock.o = -pg CFLAGS_REMOVE_sched.o = -pg endif -ifdef CONFIG_FUNCTION_GRAPH_TRACER -CFLAGS_REMOVE_extable.o = -pg # For __kernel_text_address() -CFLAGS_REMOVE_module.o = -pg # For __module_text_address() -endif obj-$(CONFIG_FREEZER) += freezer.o obj-$(CONFIG_PROFILING) += profile.o diff --git a/kernel/extable.c b/kernel/extable.c index a26cb2e17023..feb0317cf09a 100644 --- a/kernel/extable.c +++ b/kernel/extable.c @@ -17,6 +17,7 @@ */ #include #include +#include #include #include @@ -40,7 +41,7 @@ const struct exception_table_entry *search_exception_tables(unsigned long addr) return e; } -int core_kernel_text(unsigned long addr) +__notrace_funcgraph int core_kernel_text(unsigned long addr) { if (addr >= (unsigned long)_stext && addr <= (unsigned long)_etext) @@ -53,7 +54,7 @@ int core_kernel_text(unsigned long addr) return 0; } -int __kernel_text_address(unsigned long addr) +__notrace_funcgraph int __kernel_text_address(unsigned long addr) { if (core_kernel_text(addr)) return 1; diff --git a/kernel/module.c b/kernel/module.c index 89bcf7c1327d..dd2a54155b54 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -2704,7 +2704,7 @@ int is_module_address(unsigned long addr) /* Is this a valid kernel address? */ -struct module *__module_text_address(unsigned long addr) +__notrace_funcgraph struct module *__module_text_address(unsigned long addr) { struct module *mod; -- cgit v1.2.3 From 380c4b1411ccd6885f92b2c8ceb08433a720f44e Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 6 Dec 2008 03:43:41 +0100 Subject: tracing/function-graph-tracer: append the tracing_graph_flag Impact: Provide a way to pause the function graph tracer As suggested by Steven Rostedt, the previous patch that prevented from spinlock function tracing shouldn't use the raw_spinlock to fix it. It's much better to follow lockdep with normal spinlock, so this patch adds a new flag for each task to make the function graph tracer able to be paused. We also can send an ftrace_printk whithout worrying of the irrelevant traced spinlock during insertion. Signed-off-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- arch/x86/kernel/ftrace.c | 5 ++++- include/linux/ftrace.h | 13 +++++++++++++ include/linux/sched.h | 2 ++ kernel/trace/ftrace.c | 2 ++ kernel/trace/trace.c | 18 +++++------------- 5 files changed, 26 insertions(+), 14 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index f98c4076a170..1b43086b097a 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -476,7 +476,10 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) &return_to_handler; /* Nmi's are currently unsupported */ - if (atomic_read(&in_nmi)) + if (unlikely(atomic_read(&in_nmi))) + return; + + if (unlikely(atomic_read(¤t->tracing_graph_pause))) return; /* diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 449fa8e9e34f..11cac81eed08 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -401,6 +401,16 @@ static inline int task_curr_ret_stack(struct task_struct *t) { return t->curr_ret_stack; } + +static inline void pause_graph_tracing(void) +{ + atomic_inc(¤t->tracing_graph_pause); +} + +static inline void unpause_graph_tracing(void) +{ + atomic_dec(¤t->tracing_graph_pause); +} #else #define __notrace_funcgraph @@ -412,6 +422,9 @@ static inline int task_curr_ret_stack(struct task_struct *tsk) { return -1; } + +static inline void pause_graph_tracing(void) { } +static inline void unpause_graph_tracing(void) { } #endif #ifdef CONFIG_TRACING diff --git a/include/linux/sched.h b/include/linux/sched.h index 4c152e0acc9e..4b81fc5f7731 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1379,6 +1379,8 @@ struct task_struct { * because of depth overrun. */ atomic_t trace_overrun; + /* Pause for the tracing */ + atomic_t tracing_graph_pause; #endif #ifdef CONFIG_TRACING /* state flags for use by tracers */ diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 2971fe48f55e..a12f80efceaa 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -1998,6 +1998,7 @@ static int alloc_retstack_tasklist(struct ftrace_ret_stack **ret_stack_list) /* Make sure IRQs see the -1 first: */ barrier(); t->ret_stack = ret_stack_list[start++]; + atomic_set(&t->tracing_graph_pause, 0); atomic_set(&t->trace_overrun, 0); } } while_each_thread(g, t); @@ -2077,6 +2078,7 @@ void ftrace_graph_init_task(struct task_struct *t) if (!t->ret_stack) return; t->curr_ret_stack = -1; + atomic_set(&t->tracing_graph_pause, 0); atomic_set(&t->trace_overrun, 0); } else t->ret_stack = NULL; diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 33549537f30f..0b8659bd5ad2 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -3590,14 +3590,7 @@ static __init int tracer_init_debugfs(void) int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args) { - /* - * Raw Spinlock because a normal spinlock would be traced here - * and append an irrelevant couple spin_lock_irqsave/ - * spin_unlock_irqrestore traced by ftrace around this - * TRACE_PRINTK trace. - */ - static raw_spinlock_t trace_buf_lock = - (raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED; + static DEFINE_SPINLOCK(trace_buf_lock); static char trace_buf[TRACE_BUF_SIZE]; struct ring_buffer_event *event; @@ -3618,8 +3611,8 @@ int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args) if (unlikely(atomic_read(&data->disabled))) goto out; - local_irq_save(flags); - __raw_spin_lock(&trace_buf_lock); + pause_graph_tracing(); + spin_lock_irqsave(&trace_buf_lock, irq_flags); len = vsnprintf(trace_buf, TRACE_BUF_SIZE, fmt, args); len = min(len, TRACE_BUF_SIZE-1); @@ -3640,9 +3633,8 @@ int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args) ring_buffer_unlock_commit(tr->buffer, event, irq_flags); out_unlock: - __raw_spin_unlock(&trace_buf_lock); - local_irq_restore(flags); - + spin_unlock_irqrestore(&trace_buf_lock, irq_flags); + unpause_graph_tracing(); out: preempt_enable_notrace(); -- cgit v1.2.3 From 4217458dafaa57d8e26a46f5d05ab8c53cf64191 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Fri, 5 Dec 2008 17:17:09 -0800 Subject: x86: signal: change type of paramter for sys_rt_sigreturn() Impact: cleanup on 32-bit Peter pointed this parameter can be changed. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/include/asm/syscalls.h | 2 +- arch/x86/kernel/signal.c | 6 ++---- 2 files changed, 3 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/syscalls.h b/arch/x86/include/asm/syscalls.h index 87803da44010..3a5252c4b8d6 100644 --- a/arch/x86/include/asm/syscalls.h +++ b/arch/x86/include/asm/syscalls.h @@ -33,7 +33,7 @@ asmlinkage int sys_sigaction(int, const struct old_sigaction __user *, struct old_sigaction __user *); asmlinkage int sys_sigaltstack(unsigned long); asmlinkage unsigned long sys_sigreturn(unsigned long); -asmlinkage int sys_rt_sigreturn(unsigned long); +asmlinkage int sys_rt_sigreturn(struct pt_regs); /* kernel/ioport.c */ asmlinkage long sys_iopl(unsigned long); diff --git a/arch/x86/kernel/signal.c b/arch/x86/kernel/signal.c index b1f4d34e0a38..b1cc6da64208 100644 --- a/arch/x86/kernel/signal.c +++ b/arch/x86/kernel/signal.c @@ -642,11 +642,9 @@ badframe: } #ifdef CONFIG_X86_32 -asmlinkage int sys_rt_sigreturn(unsigned long __unused) +asmlinkage int sys_rt_sigreturn(struct pt_regs regs) { - struct pt_regs *regs = (struct pt_regs *)&__unused; - - return do_rt_sigreturn(regs); + return do_rt_sigreturn(®s); } #else /* !CONFIG_X86_32 */ asmlinkage long sys_rt_sigreturn(struct pt_regs *regs) -- cgit v1.2.3 From 50dd94e017ec39f85c26b6c10ed9fb2d7a7d8042 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Mon, 8 Dec 2008 18:47:51 +0100 Subject: sparseirq: fix typo in !CONFIG_IO_APIC case Impact: build fix Signed-off-by: Ingo Molnar --- arch/x86/include/asm/io_apic.h | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/io_apic.h b/arch/x86/include/asm/io_apic.h index b35e94c2d6df..25d527ca1362 100644 --- a/arch/x86/include/asm/io_apic.h +++ b/arch/x86/include/asm/io_apic.h @@ -193,12 +193,9 @@ extern void probe_nr_irqs_gsi(void); #else /* !CONFIG_X86_IO_APIC */ #define io_apic_assign_pci_irqs 0 static const int timer_through_8259 = 0; -static inline void ioapic_init_mappings(void) { } +static inline void ioapic_init_mappings(void) { } -static inline int probe_nr_irqs(void) -{ - return NR_IRQS; -} +static inline void probe_nr_irqs_gsi(void) { } #endif #endif /* _ASM_X86_IO_APIC_H */ -- cgit v1.2.3 From 8a4830f8891be6b4e04809693a24771a4694e0b0 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Tue, 9 Dec 2008 11:56:20 -0800 Subject: sparseirq: fix !SMP && !PCI_MSI && !HT_IRQ build Ingo Molnar wrote: >>> drivers/pci/intr_remapping.c: In function 'irq_2_iommu_alloc': >>> drivers/pci/intr_remapping.c:72: error: 'boot_cpu_id' undeclared (first use in this function) >>> drivers/pci/intr_remapping.c:72: error: (Each undeclared identifier is reported only once >>> drivers/pci/intr_remapping.c:72: error: for each function it appears in.) sparseirq should only be used with SMP for now. --- arch/x86/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 48ac688de3cd..8943c13502c6 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -240,7 +240,7 @@ config X86_HAS_BOOT_CPU_ID config SPARSE_IRQ bool "Support sparse irq numbering" - depends on PCI_MSI || HT_IRQ + depends on (PCI_MSI || HT_IRQ) && SMP default y help This enables support for sparse irq, esp for msi/msi-x. You may need -- cgit v1.2.3 From fd13f6c85144bb2026c534a35be1d7cb7628a64a Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Sun, 19 Oct 2008 21:00:09 +0200 Subject: oprofile: comment cleanup This fixes the coding style of some comments. Signed-off-by: Robert Richter --- arch/x86/oprofile/op_model_amd.c | 31 ++++++++++++++++--------------- drivers/oprofile/cpu_buffer.c | 25 ++++++++++++++++--------- 2 files changed, 32 insertions(+), 24 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/oprofile/op_model_amd.c b/arch/x86/oprofile/op_model_amd.c index 509513760a6e..fb67e1999d85 100644 --- a/arch/x86/oprofile/op_model_amd.c +++ b/arch/x86/oprofile/op_model_amd.c @@ -65,8 +65,10 @@ static unsigned long reset_value[NUM_COUNTERS]; #define IBS_FETCH_BEGIN 3 #define IBS_OP_BEGIN 4 -/* The function interface needs to be fixed, something like add - data. Should then be added to linux/oprofile.h. */ +/* + * The function interface needs to be fixed, something like add + * data. Should then be added to linux/oprofile.h. + */ extern void oprofile_add_ibs_sample(struct pt_regs *const regs, unsigned int *const ibs_sample, int ibs_code); @@ -106,7 +108,7 @@ struct ibs_op_sample { /* * unitialize the APIC for the IBS interrupts if needed on AMD Family10h+ -*/ + */ static void clear_ibs_nmi(void); static int ibs_allowed; /* AMD Family10h and later */ @@ -223,7 +225,7 @@ op_amd_handle_ibs(struct pt_regs * const regs, (unsigned int *)&ibs_fetch, IBS_FETCH_BEGIN); - /*reenable the IRQ */ + /* reenable the IRQ */ rdmsr(MSR_AMD64_IBSFETCHCTL, low, high); high &= ~IBS_FETCH_HIGH_VALID_BIT; high |= IBS_FETCH_HIGH_ENABLE; @@ -331,8 +333,10 @@ static void op_amd_stop(struct op_msrs const * const msrs) unsigned int low, high; int i; - /* Subtle: stop on all counters to avoid race with - * setting our pm callback */ + /* + * Subtle: stop on all counters to avoid race with setting our + * pm callback + */ for (i = 0 ; i < NUM_COUNTERS ; ++i) { if (!reset_value[i]) continue; @@ -343,13 +347,15 @@ static void op_amd_stop(struct op_msrs const * const msrs) #ifdef CONFIG_OPROFILE_IBS if (ibs_allowed && ibs_config.fetch_enabled) { - low = 0; /* clear max count and enable */ + /* clear max count and enable */ + low = 0; high = 0; wrmsr(MSR_AMD64_IBSFETCHCTL, low, high); } if (ibs_allowed && ibs_config.op_enabled) { - low = 0; /* clear max count and enable */ + /* clear max count and enable */ + low = 0; high = 0; wrmsr(MSR_AMD64_IBSOPCTL, low, high); } @@ -443,10 +449,7 @@ static int pfm_amd64_setup_eilvt(void) return 0; } -/* - * initialize the APIC for the IBS interrupts - * if available (AMD Family10h rev B0 and later) - */ +/* initialize the APIC for the IBS interrupts if available */ static void setup_ibs(void) { ibs_allowed = boot_cpu_has(X86_FEATURE_IBS); @@ -463,9 +466,7 @@ static void setup_ibs(void) } -/* - * unitialize the APIC for the IBS interrupts if needed on AMD Family10h - * rev B0 and later */ +/* uninitialize the APIC for the IBS interrupts if needed */ static void clear_ibs_nmi(void) { if (ibs_allowed) diff --git a/drivers/oprofile/cpu_buffer.c b/drivers/oprofile/cpu_buffer.c index 01d38e78cde1..3958107723fb 100644 --- a/drivers/oprofile/cpu_buffer.c +++ b/drivers/oprofile/cpu_buffer.c @@ -127,9 +127,10 @@ void end_cpu_work(void) /* Resets the cpu buffer to a sane state. */ void cpu_buffer_reset(struct oprofile_cpu_buffer *cpu_buf) { - /* reset these to invalid values; the next sample - * collected will populate the buffer with proper - * values to initialize the buffer + /* + * reset these to invalid values; the next sample collected + * will populate the buffer with proper values to initialize + * the buffer */ cpu_buf->last_is_kernel = -1; cpu_buf->last_task = NULL; @@ -151,8 +152,10 @@ static void increment_head(struct oprofile_cpu_buffer *b) { unsigned long new_head = b->head_pos + 1; - /* Ensure anything written to the slot before we - * increment is visible */ + /* + * Ensure anything written to the slot before we increment is + * visible + */ wmb(); if (new_head < b->buffer_size) @@ -253,8 +256,10 @@ void oprofile_add_ext_sample(unsigned long pc, struct pt_regs * const regs, if (!oprofile_begin_trace(cpu_buf)) return; - /* if log_sample() fail we can't backtrace since we lost the source - * of this event */ + /* + * if log_sample() fail we can't backtrace since we lost the + * source of this event + */ if (log_sample(cpu_buf, pc, is_kernel, event)) oprofile_ops.backtrace(regs, backtrace_depth); oprofile_end_trace(cpu_buf); @@ -338,8 +343,10 @@ void oprofile_add_trace(unsigned long pc) return; } - /* broken frame can give an eip with the same value as an escape code, - * abort the trace if we get it */ + /* + * broken frame can give an eip with the same value as an + * escape code, abort the trace if we get it + */ if (pc == ESCAPE_CODE) { cpu_buf->tracing = 0; cpu_buf->backtrace_aborted++; -- cgit v1.2.3 From cdc1834d1aa2e5b574a25e66f82625b44cdd0d8f Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Fri, 26 Sep 2008 22:18:44 -0400 Subject: oprofile: whitspace changes only Signed-off-by: Robert Richter --- arch/x86/oprofile/op_model_amd.c | 4 ++-- drivers/oprofile/cpu_buffer.c | 4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/oprofile/op_model_amd.c b/arch/x86/oprofile/op_model_amd.c index fb67e1999d85..f71bd218b488 100644 --- a/arch/x86/oprofile/op_model_amd.c +++ b/arch/x86/oprofile/op_model_amd.c @@ -70,8 +70,8 @@ static unsigned long reset_value[NUM_COUNTERS]; * data. Should then be added to linux/oprofile.h. */ extern void -oprofile_add_ibs_sample(struct pt_regs *const regs, - unsigned int *const ibs_sample, int ibs_code); +oprofile_add_ibs_sample(struct pt_regs * const regs, + unsigned int * const ibs_sample, int ibs_code); struct ibs_fetch_sample { /* MSRC001_1031 IBS Fetch Linear Address Register */ diff --git a/drivers/oprofile/cpu_buffer.c b/drivers/oprofile/cpu_buffer.c index 3958107723fb..2c4d54187b90 100644 --- a/drivers/oprofile/cpu_buffer.c +++ b/drivers/oprofile/cpu_buffer.c @@ -277,8 +277,8 @@ void oprofile_add_sample(struct pt_regs * const regs, unsigned long event) #define MAX_IBS_SAMPLE_SIZE 14 -void oprofile_add_ibs_sample(struct pt_regs *const regs, - unsigned int *const ibs_sample, int ibs_code) +void oprofile_add_ibs_sample(struct pt_regs * const regs, + unsigned int * const ibs_sample, int ibs_code) { int is_kernel = !user_mode(regs); struct oprofile_cpu_buffer *cpu_buf = &__get_cpu_var(cpu_buffer); -- cgit v1.2.3 From 9fa6812dbab9207f7af52c3d0417f1f9eb89c386 Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Mon, 24 Nov 2008 14:21:03 +0100 Subject: x86/oprofile: reordering IBS code in op_model_amd.c This is part of the cpu buffer rework. Signed-off-by: Robert Richter --- arch/x86/oprofile/op_model_amd.c | 26 +++++++++++++------------- 1 file changed, 13 insertions(+), 13 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/oprofile/op_model_amd.c b/arch/x86/oprofile/op_model_amd.c index f71bd218b488..8ff657b3ff89 100644 --- a/arch/x86/oprofile/op_model_amd.c +++ b/arch/x86/oprofile/op_model_amd.c @@ -376,18 +376,7 @@ static void op_amd_shutdown(struct op_msrs const * const msrs) } } -#ifndef CONFIG_OPROFILE_IBS - -/* no IBS support */ - -static int op_amd_init(struct oprofile_operations *ops) -{ - return 0; -} - -static void op_amd_exit(void) {} - -#else +#ifdef CONFIG_OPROFILE_IBS static u8 ibs_eilvt_off; @@ -531,7 +520,18 @@ static void op_amd_exit(void) clear_ibs_nmi(); } -#endif +#else + +/* no IBS support */ + +static int op_amd_init(struct oprofile_operations *ops) +{ + return 0; +} + +static void op_amd_exit(void) {} + +#endif /* CONFIG_OPROFILE_IBS */ struct op_x86_model_spec const op_amd_spec = { .init = op_amd_init, -- cgit v1.2.3 From fe615cbf34fc6a1c53c359417da4696328a488ed Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Mon, 24 Nov 2008 14:58:03 +0100 Subject: x86/oprofile: cleanup IBS init/exit functions in op_model_amd.c Implementation of pairwise init/exit funcions for IBS and IBS NMI setup. There are also some function renames and the removal of forward function declarations. Signed-off-by: Robert Richter --- arch/x86/oprofile/op_model_amd.c | 32 +++++++++++++++++--------------- 1 file changed, 17 insertions(+), 15 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/oprofile/op_model_amd.c b/arch/x86/oprofile/op_model_amd.c index 8ff657b3ff89..98658f25f542 100644 --- a/arch/x86/oprofile/op_model_amd.c +++ b/arch/x86/oprofile/op_model_amd.c @@ -106,11 +106,6 @@ struct ibs_op_sample { unsigned int ibs_dc_phys_high; }; -/* - * unitialize the APIC for the IBS interrupts if needed on AMD Family10h+ - */ -static void clear_ibs_nmi(void); - static int ibs_allowed; /* AMD Family10h and later */ struct op_ibs_config { @@ -390,7 +385,7 @@ static inline void apic_clear_ibs_nmi_per_cpu(void *arg) setup_APIC_eilvt_ibs(0, APIC_EILVT_MSG_FIX, 1); } -static int pfm_amd64_setup_eilvt(void) +static int init_ibs_nmi(void) { #define IBSCTL_LVTOFFSETVAL (1 << 8) #define IBSCTL 0x1cc @@ -438,15 +433,22 @@ static int pfm_amd64_setup_eilvt(void) return 0; } +/* uninitialize the APIC for the IBS interrupts if needed */ +static void clear_ibs_nmi(void) +{ + if (ibs_allowed) + on_each_cpu(apic_clear_ibs_nmi_per_cpu, NULL, 1); +} + /* initialize the APIC for the IBS interrupts if available */ -static void setup_ibs(void) +static void ibs_init(void) { ibs_allowed = boot_cpu_has(X86_FEATURE_IBS); if (!ibs_allowed) return; - if (pfm_amd64_setup_eilvt()) { + if (init_ibs_nmi()) { ibs_allowed = 0; return; } @@ -454,12 +456,12 @@ static void setup_ibs(void) printk(KERN_INFO "oprofile: AMD IBS detected\n"); } - -/* uninitialize the APIC for the IBS interrupts if needed */ -static void clear_ibs_nmi(void) +static void ibs_exit(void) { - if (ibs_allowed) - on_each_cpu(apic_clear_ibs_nmi_per_cpu, NULL, 1); + if (!ibs_allowed) + return; + + clear_ibs_nmi(); } static int (*create_arch_files)(struct super_block *sb, struct dentry *root); @@ -509,7 +511,7 @@ static int setup_ibs_files(struct super_block *sb, struct dentry *root) static int op_amd_init(struct oprofile_operations *ops) { - setup_ibs(); + ibs_init(); create_arch_files = ops->create_files; ops->create_files = setup_ibs_files; return 0; @@ -517,7 +519,7 @@ static int op_amd_init(struct oprofile_operations *ops) static void op_amd_exit(void) { - clear_ibs_nmi(); + ibs_exit(); } #else -- cgit v1.2.3 From b0884e25fe361f2ca228808fb5fd1b74cb04e711 Mon Sep 17 00:00:00 2001 From: Markus Metzger Date: Thu, 11 Dec 2008 13:45:23 +0100 Subject: x86, bts: turn BUG_ON into WARN_ON_ONCE Impact: make the ds code more debuggable Turn BUG_ON's into WARN_ON_ONCE. Signed-off-by: Markus Metzger Signed-off-by: Ingo Molnar --- arch/x86/kernel/ds.c | 4 ++-- arch/x86/kernel/ptrace.c | 5 +++-- 2 files changed, 5 insertions(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/ds.c b/arch/x86/kernel/ds.c index 19a8c2c0389f..095306988667 100644 --- a/arch/x86/kernel/ds.c +++ b/arch/x86/kernel/ds.c @@ -452,7 +452,7 @@ struct pebs_tracer *ds_request_pebs(struct task_struct *task, static void ds_release(struct ds_tracer *tracer, enum ds_qualifier qual) { - BUG_ON(tracer->context->owner[qual] != tracer); + WARN_ON_ONCE(tracer->context->owner[qual] != tracer); tracer->context->owner[qual] = NULL; put_tracer(tracer->context->task); @@ -774,7 +774,7 @@ ds_configure(const struct ds_configuration *cfg) printk(KERN_INFO "DS available\n"); - BUG_ON(MAX_SIZEOF_DS < ds_cfg.sizeof_ds); + WARN_ON_ONCE(MAX_SIZEOF_DS < ds_cfg.sizeof_ds); } void __cpuinit ds_init_intel(struct cpuinfo_x86 *c) diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c index 2c8ec1ba75e6..b2998fe1166b 100644 --- a/arch/x86/kernel/ptrace.c +++ b/arch/x86/kernel/ptrace.c @@ -878,7 +878,8 @@ static int ptrace_bts_write_record(struct task_struct *child, { unsigned char bts_record[BTS_MAX_RECORD_SIZE]; - BUG_ON(BTS_MAX_RECORD_SIZE < bts_cfg.sizeof_bts); + if (BTS_MAX_RECORD_SIZE < bts_cfg.sizeof_bts) + return -EOVERFLOW; memset(bts_record, 0, bts_cfg.sizeof_bts); switch (in->qualifier) { @@ -1133,7 +1134,7 @@ long arch_ptrace(struct task_struct *child, long request, long addr, long data) ret = ds_get_bts_index(child->bts, &size); if (ret == 0) { - BUG_ON(size != (int) size); + WARN_ON_ONCE(size != (int) size); ret = (int) size; } break; -- cgit v1.2.3 From c2724775ce57c98b8af9694857b941dc61056516 Mon Sep 17 00:00:00 2001 From: Markus Metzger Date: Thu, 11 Dec 2008 13:49:59 +0100 Subject: x86, bts: provide in-kernel branch-trace interface Impact: cleanup Move the BTS bits from ptrace.c into ds.c. Signed-off-by: Markus Metzger Signed-off-by: Ingo Molnar --- arch/x86/include/asm/ds.h | 241 ++++++----- arch/x86/include/asm/processor.h | 13 + arch/x86/include/asm/ptrace.h | 36 -- arch/x86/include/asm/thread_info.h | 5 +- arch/x86/kernel/cpu/intel.c | 4 - arch/x86/kernel/ds.c | 857 +++++++++++++++++++++++-------------- arch/x86/kernel/process_32.c | 59 +-- arch/x86/kernel/process_64.c | 50 +-- arch/x86/kernel/ptrace.c | 416 +++++------------- include/linux/sched.h | 1 + 10 files changed, 811 insertions(+), 871 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/ds.h b/arch/x86/include/asm/ds.h index 99b6c39774a4..ee0ea3a96c11 100644 --- a/arch/x86/include/asm/ds.h +++ b/arch/x86/include/asm/ds.h @@ -6,13 +6,13 @@ * precise-event based sampling (PEBS). * * It manages: - * - per-thread and per-cpu allocation of BTS and PEBS + * - DS and BTS hardware configuration * - buffer overflow handling (to be done) * - buffer access * - * It assumes: - * - get_task_struct on all traced tasks - * - current is allowed to trace tasks + * It does not do: + * - security checking (is the caller allowed to trace the task) + * - buffer allocation (memory accounting) * * * Copyright (C) 2007-2008 Intel Corporation. @@ -31,6 +31,7 @@ #ifdef CONFIG_X86_DS struct task_struct; +struct ds_context; struct ds_tracer; struct bts_tracer; struct pebs_tracer; @@ -38,6 +39,38 @@ struct pebs_tracer; typedef void (*bts_ovfl_callback_t)(struct bts_tracer *); typedef void (*pebs_ovfl_callback_t)(struct pebs_tracer *); + +/* + * A list of features plus corresponding macros to talk about them in + * the ds_request function's flags parameter. + * + * We use the enum to index an array of corresponding control bits; + * we use the macro to index a flags bit-vector. + */ +enum ds_feature { + dsf_bts = 0, + dsf_bts_kernel, +#define BTS_KERNEL (1 << dsf_bts_kernel) + /* trace kernel-mode branches */ + + dsf_bts_user, +#define BTS_USER (1 << dsf_bts_user) + /* trace user-mode branches */ + + dsf_bts_overflow, + dsf_bts_max, + dsf_pebs = dsf_bts_max, + + dsf_pebs_max, + dsf_ctl_max = dsf_pebs_max, + dsf_bts_timestamps = dsf_ctl_max, +#define BTS_TIMESTAMPS (1 << dsf_bts_timestamps) + /* add timestamps into BTS trace */ + +#define BTS_USER_FLAGS (BTS_KERNEL | BTS_USER | BTS_TIMESTAMPS) +}; + + /* * Request BTS or PEBS * @@ -58,92 +91,135 @@ typedef void (*pebs_ovfl_callback_t)(struct pebs_tracer *); * NULL if cyclic buffer requested * th: the interrupt threshold in records from the end of the buffer; * -1 if no interrupt threshold is requested. + * flags: a bit-mask of the above flags */ extern struct bts_tracer *ds_request_bts(struct task_struct *task, void *base, size_t size, - bts_ovfl_callback_t ovfl, size_t th); + bts_ovfl_callback_t ovfl, + size_t th, unsigned int flags); extern struct pebs_tracer *ds_request_pebs(struct task_struct *task, void *base, size_t size, pebs_ovfl_callback_t ovfl, - size_t th); + size_t th, unsigned int flags); /* * Release BTS or PEBS resources - * - * Returns 0 on success; -Eerrno otherwise + * Suspend and resume BTS or PEBS tracing * * tracer: the tracer handle returned from ds_request_~() */ -extern int ds_release_bts(struct bts_tracer *tracer); -extern int ds_release_pebs(struct pebs_tracer *tracer); +extern void ds_release_bts(struct bts_tracer *tracer); +extern void ds_suspend_bts(struct bts_tracer *tracer); +extern void ds_resume_bts(struct bts_tracer *tracer); +extern void ds_release_pebs(struct pebs_tracer *tracer); +extern void ds_suspend_pebs(struct pebs_tracer *tracer); +extern void ds_resume_pebs(struct pebs_tracer *tracer); + /* - * Get the (array) index of the write pointer. - * (assuming an array of BTS/PEBS records) - * - * Returns 0 on success; -Eerrno on error + * The raw DS buffer state as it is used for BTS and PEBS recording. * - * tracer: the tracer handle returned from ds_request_~() - * pos (out): will hold the result + * This is the low-level, arch-dependent interface for working + * directly on the raw trace data. */ -extern int ds_get_bts_index(struct bts_tracer *tracer, size_t *pos); -extern int ds_get_pebs_index(struct pebs_tracer *tracer, size_t *pos); +struct ds_trace { + /* the number of bts/pebs records */ + size_t n; + /* the size of a bts/pebs record in bytes */ + size_t size; + /* pointers into the raw buffer: + - to the first entry */ + void *begin; + /* - one beyond the last entry */ + void *end; + /* - one beyond the newest entry */ + void *top; + /* - the interrupt threshold */ + void *ith; + /* flags given on ds_request() */ + unsigned int flags; +}; /* - * Get the (array) index one record beyond the end of the array. - * (assuming an array of BTS/PEBS records) - * - * Returns 0 on success; -Eerrno on error - * - * tracer: the tracer handle returned from ds_request_~() - * pos (out): will hold the result + * An arch-independent view on branch trace data. */ -extern int ds_get_bts_end(struct bts_tracer *tracer, size_t *pos); -extern int ds_get_pebs_end(struct pebs_tracer *tracer, size_t *pos); +enum bts_qualifier { + bts_invalid, +#define BTS_INVALID bts_invalid + + bts_branch, +#define BTS_BRANCH bts_branch + + bts_task_arrives, +#define BTS_TASK_ARRIVES bts_task_arrives + + bts_task_departs, +#define BTS_TASK_DEPARTS bts_task_departs + + bts_qual_bit_size = 4, + bts_qual_max = (1 << bts_qual_bit_size), +}; + +struct bts_struct { + __u64 qualifier; + union { + /* BTS_BRANCH */ + struct { + __u64 from; + __u64 to; + } lbr; + /* BTS_TASK_ARRIVES or BTS_TASK_DEPARTS */ + struct { + __u64 jiffies; + pid_t pid; + } timestamp; + } variant; +}; + /* - * Provide a pointer to the BTS/PEBS record at parameter index. - * (assuming an array of BTS/PEBS records) - * - * The pointer points directly into the buffer. The user is - * responsible for copying the record. - * - * Returns the size of a single record on success; -Eerrno on error + * The BTS state. * - * tracer: the tracer handle returned from ds_request_~() - * index: the index of the requested record - * record (out): pointer to the requested record + * This gives access to the raw DS state and adds functions to provide + * an arch-independent view of the BTS data. */ -extern int ds_access_bts(struct bts_tracer *tracer, - size_t index, const void **record); -extern int ds_access_pebs(struct pebs_tracer *tracer, - size_t index, const void **record); +struct bts_trace { + struct ds_trace ds; + + int (*read)(struct bts_tracer *tracer, const void *at, + struct bts_struct *out); + int (*write)(struct bts_tracer *tracer, const struct bts_struct *in); +}; + /* - * Write one or more BTS/PEBS records at the write pointer index and - * advance the write pointer. + * The PEBS state. * - * If size is not a multiple of the record size, trailing bytes are - * zeroed out. - * - * May result in one or more overflow notifications. - * - * If called during overflow handling, that is, with index >= - * interrupt threshold, the write will wrap around. + * This gives access to the raw DS state and the PEBS-specific counter + * reset value. + */ +struct pebs_trace { + struct ds_trace ds; + + /* the PEBS reset value */ + unsigned long long reset_value; +}; + + +/* + * Read the BTS or PEBS trace. * - * An overflow notification is given if and when the interrupt - * threshold is reached during or after the write. + * Returns a view on the trace collected for the parameter tracer. * - * Returns the number of bytes written or -Eerrno. + * The view remains valid as long as the traced task is not running or + * the tracer is suspended. + * Writes into the trace buffer are not reflected. * * tracer: the tracer handle returned from ds_request_~() - * buffer: the buffer to write - * size: the size of the buffer */ -extern int ds_write_bts(struct bts_tracer *tracer, - const void *buffer, size_t size); -extern int ds_write_pebs(struct pebs_tracer *tracer, - const void *buffer, size_t size); +extern const struct bts_trace *ds_read_bts(struct bts_tracer *tracer); +extern const struct pebs_trace *ds_read_pebs(struct pebs_tracer *tracer); + /* * Reset the write pointer of the BTS/PEBS buffer. @@ -155,27 +231,6 @@ extern int ds_write_pebs(struct pebs_tracer *tracer, extern int ds_reset_bts(struct bts_tracer *tracer); extern int ds_reset_pebs(struct pebs_tracer *tracer); -/* - * Clear the BTS/PEBS buffer and reset the write pointer. - * The entire buffer will be zeroed out. - * - * Returns 0 on success; -Eerrno on error - * - * tracer: the tracer handle returned from ds_request_~() - */ -extern int ds_clear_bts(struct bts_tracer *tracer); -extern int ds_clear_pebs(struct pebs_tracer *tracer); - -/* - * Provide the PEBS counter reset value. - * - * Returns 0 on success; -Eerrno on error - * - * tracer: the tracer handle returned from ds_request_pebs() - * value (out): the counter reset value - */ -extern int ds_get_pebs_reset(struct pebs_tracer *tracer, u64 *value); - /* * Set the PEBS counter reset value. * @@ -192,35 +247,17 @@ extern int ds_set_pebs_reset(struct pebs_tracer *tracer, u64 value); struct cpuinfo_x86; extern void __cpuinit ds_init_intel(struct cpuinfo_x86 *); - - /* - * The DS context - part of struct thread_struct. + * Context switch work */ -#define MAX_SIZEOF_DS (12 * 8) - -struct ds_context { - /* pointer to the DS configuration; goes into MSR_IA32_DS_AREA */ - unsigned char ds[MAX_SIZEOF_DS]; - /* the owner of the BTS and PEBS configuration, respectively */ - struct ds_tracer *owner[2]; - /* use count */ - unsigned long count; - /* a pointer to the context location inside the thread_struct - * or the per_cpu context array */ - struct ds_context **this; - /* a pointer to the task owning this context, or NULL, if the - * context is owned by a cpu */ - struct task_struct *task; -}; - -/* called by exit_thread() to free leftover contexts */ -extern void ds_free(struct ds_context *context); +extern void ds_switch_to(struct task_struct *prev, struct task_struct *next); #else /* CONFIG_X86_DS */ struct cpuinfo_x86; static inline void __cpuinit ds_init_intel(struct cpuinfo_x86 *ignored) {} +static inline void ds_switch_to(struct task_struct *prev, + struct task_struct *next) {} #endif /* CONFIG_X86_DS */ #endif /* _ASM_X86_DS_H */ diff --git a/arch/x86/include/asm/processor.h b/arch/x86/include/asm/processor.h index 5ca01e383269..aa5914f8e501 100644 --- a/arch/x86/include/asm/processor.h +++ b/arch/x86/include/asm/processor.h @@ -752,6 +752,19 @@ extern void switch_to_new_gdt(void); extern void cpu_init(void); extern void init_gdt(int cpu); +static inline unsigned long get_debugctlmsr(void) +{ + unsigned long debugctlmsr = 0; + +#ifndef CONFIG_X86_DEBUGCTLMSR + if (boot_cpu_data.x86 < 6) + return 0; +#endif + rdmsrl(MSR_IA32_DEBUGCTLMSR, debugctlmsr); + + return debugctlmsr; +} + static inline void update_debugctlmsr(unsigned long debugctlmsr) { #ifndef CONFIG_X86_DEBUGCTLMSR diff --git a/arch/x86/include/asm/ptrace.h b/arch/x86/include/asm/ptrace.h index eefb0594b058..fbf744215911 100644 --- a/arch/x86/include/asm/ptrace.h +++ b/arch/x86/include/asm/ptrace.h @@ -6,7 +6,6 @@ #include #ifdef __KERNEL__ -#include /* the DS BTS struct is used for ptrace too */ #include #endif @@ -128,34 +127,6 @@ struct pt_regs { #endif /* !__i386__ */ -#ifdef CONFIG_X86_PTRACE_BTS -/* a branch trace record entry - * - * In order to unify the interface between various processor versions, - * we use the below data structure for all processors. - */ -enum bts_qualifier { - BTS_INVALID = 0, - BTS_BRANCH, - BTS_TASK_ARRIVES, - BTS_TASK_DEPARTS -}; - -struct bts_struct { - __u64 qualifier; - union { - /* BTS_BRANCH */ - struct { - __u64 from_ip; - __u64 to_ip; - } lbr; - /* BTS_TASK_ARRIVES or - BTS_TASK_DEPARTS */ - __u64 jiffies; - } variant; -}; -#endif /* CONFIG_X86_PTRACE_BTS */ - #ifdef __KERNEL__ #include @@ -163,13 +134,6 @@ struct bts_struct { struct cpuinfo_x86; struct task_struct; -#ifdef CONFIG_X86_PTRACE_BTS -extern void __cpuinit ptrace_bts_init_intel(struct cpuinfo_x86 *); -extern void ptrace_bts_take_timestamp(struct task_struct *, enum bts_qualifier); -#else -#define ptrace_bts_init_intel(config) do {} while (0) -#endif /* CONFIG_X86_PTRACE_BTS */ - extern unsigned long profile_pc(struct pt_regs *regs); extern unsigned long diff --git a/arch/x86/include/asm/thread_info.h b/arch/x86/include/asm/thread_info.h index 0921b4018c11..bf8113d16a33 100644 --- a/arch/x86/include/asm/thread_info.h +++ b/arch/x86/include/asm/thread_info.h @@ -93,7 +93,6 @@ struct thread_info { #define TIF_FORCED_TF 24 /* true if TF in eflags artificially */ #define TIF_DEBUGCTLMSR 25 /* uses thread_struct.debugctlmsr */ #define TIF_DS_AREA_MSR 26 /* uses thread_struct.ds_area_msr */ -#define TIF_BTS_TRACE_TS 27 /* record scheduling event timestamps */ #define _TIF_SYSCALL_TRACE (1 << TIF_SYSCALL_TRACE) #define _TIF_NOTIFY_RESUME (1 << TIF_NOTIFY_RESUME) @@ -115,7 +114,6 @@ struct thread_info { #define _TIF_FORCED_TF (1 << TIF_FORCED_TF) #define _TIF_DEBUGCTLMSR (1 << TIF_DEBUGCTLMSR) #define _TIF_DS_AREA_MSR (1 << TIF_DS_AREA_MSR) -#define _TIF_BTS_TRACE_TS (1 << TIF_BTS_TRACE_TS) /* work to do in syscall_trace_enter() */ #define _TIF_WORK_SYSCALL_ENTRY \ @@ -141,8 +139,7 @@ struct thread_info { /* flags to check in __switch_to() */ #define _TIF_WORK_CTXSW \ - (_TIF_IO_BITMAP|_TIF_DEBUGCTLMSR|_TIF_DS_AREA_MSR|_TIF_BTS_TRACE_TS| \ - _TIF_NOTSC) + (_TIF_IO_BITMAP|_TIF_DEBUGCTLMSR|_TIF_DS_AREA_MSR|_TIF_NOTSC) #define _TIF_WORK_CTXSW_PREV _TIF_WORK_CTXSW #define _TIF_WORK_CTXSW_NEXT (_TIF_WORK_CTXSW|_TIF_DEBUG) diff --git a/arch/x86/kernel/cpu/intel.c b/arch/x86/kernel/cpu/intel.c index 816f27f289b1..cd413d9a0218 100644 --- a/arch/x86/kernel/cpu/intel.c +++ b/arch/x86/kernel/cpu/intel.c @@ -11,7 +11,6 @@ #include #include #include -#include #include #include @@ -309,9 +308,6 @@ static void __cpuinit init_intel(struct cpuinfo_x86 *c) set_cpu_cap(c, X86_FEATURE_P3); #endif - if (cpu_has_bts) - ptrace_bts_init_intel(c); - detect_extended_topology(c); if (!cpu_has(c, X86_FEATURE_XTOPOLOGY)) { /* diff --git a/arch/x86/kernel/ds.c b/arch/x86/kernel/ds.c index 095306988667..f0583005b75e 100644 --- a/arch/x86/kernel/ds.c +++ b/arch/x86/kernel/ds.c @@ -6,13 +6,13 @@ * precise-event based sampling (PEBS). * * It manages: - * - per-thread and per-cpu allocation of BTS and PEBS + * - DS and BTS hardware configuration * - buffer overflow handling (to be done) * - buffer access * - * It assumes: - * - get_task_struct on all traced tasks - * - current is allowed to trace tasks + * It does not do: + * - security checking (is the caller allowed to trace the task) + * - buffer allocation (memory accounting) * * * Copyright (C) 2007-2008 Intel Corporation. @@ -34,15 +34,30 @@ * The configuration for a particular DS hardware implementation. */ struct ds_configuration { - /* the size of the DS structure in bytes */ - unsigned char sizeof_ds; - /* the size of one pointer-typed field in the DS structure in bytes; - this covers the first 8 fields related to buffer management. */ + /* the name of the configuration */ + const char *name; + /* the size of one pointer-typed field in the DS structure and + in the BTS and PEBS buffers in bytes; + this covers the first 8 DS fields related to buffer management. */ unsigned char sizeof_field; /* the size of a BTS/PEBS record in bytes */ unsigned char sizeof_rec[2]; + /* a series of bit-masks to control various features indexed + * by enum ds_feature */ + unsigned long ctl[dsf_ctl_max]; }; -static struct ds_configuration ds_cfg; +static DEFINE_PER_CPU(struct ds_configuration, ds_cfg_array); + +#define ds_cfg per_cpu(ds_cfg_array, smp_processor_id()) + +#define MAX_SIZEOF_DS (12 * 8) /* maximal size of a DS configuration */ +#define MAX_SIZEOF_BTS (3 * 8) /* maximal size of a BTS record */ +#define DS_ALIGNMENT (1 << 3) /* BTS and PEBS buffer alignment */ + +#define BTS_CONTROL \ + (ds_cfg.ctl[dsf_bts] | ds_cfg.ctl[dsf_bts_kernel] | ds_cfg.ctl[dsf_bts_user] |\ + ds_cfg.ctl[dsf_bts_overflow]) + /* * A BTS or PEBS tracer. @@ -61,6 +76,8 @@ struct ds_tracer { struct bts_tracer { /* the common DS part */ struct ds_tracer ds; + /* the trace including the DS configuration */ + struct bts_trace trace; /* buffer overflow notification function */ bts_ovfl_callback_t ovfl; }; @@ -68,6 +85,8 @@ struct bts_tracer { struct pebs_tracer { /* the common DS part */ struct ds_tracer ds; + /* the trace including the DS configuration */ + struct pebs_trace trace; /* buffer overflow notification function */ pebs_ovfl_callback_t ovfl; }; @@ -134,13 +153,11 @@ static inline void ds_set(unsigned char *base, enum ds_qualifier qual, (*(unsigned long *)base) = value; } -#define DS_ALIGNMENT (1 << 3) /* BTS and PEBS buffer alignment */ - /* * Locking is done only for allocating BTS or PEBS resources. */ -static spinlock_t ds_lock = __SPIN_LOCK_UNLOCKED(ds_lock); +static DEFINE_SPINLOCK(ds_lock); /* @@ -156,27 +173,32 @@ static spinlock_t ds_lock = __SPIN_LOCK_UNLOCKED(ds_lock); * >0 number of per-thread tracers * <0 number of per-cpu tracers * - * The below functions to get and put tracers and to check the - * allocation type require the ds_lock to be held by the caller. - * * Tracers essentially gives the number of ds contexts for a certain * type of allocation. */ -static long tracers; +static atomic_t tracers = ATOMIC_INIT(0); static inline void get_tracer(struct task_struct *task) { - tracers += (task ? 1 : -1); + if (task) + atomic_inc(&tracers); + else + atomic_dec(&tracers); } static inline void put_tracer(struct task_struct *task) { - tracers -= (task ? 1 : -1); + if (task) + atomic_dec(&tracers); + else + atomic_inc(&tracers); } static inline int check_tracer(struct task_struct *task) { - return (task ? (tracers >= 0) : (tracers <= 0)); + return task ? + (atomic_read(&tracers) >= 0) : + (atomic_read(&tracers) <= 0); } @@ -190,14 +212,30 @@ static inline int check_tracer(struct task_struct *task) * Contexts are use-counted. They are allocated on first access and * deallocated when the last user puts the context. */ -static DEFINE_PER_CPU(struct ds_context *, system_context); +struct ds_context { + /* pointer to the DS configuration; goes into MSR_IA32_DS_AREA */ + unsigned char ds[MAX_SIZEOF_DS]; + /* the owner of the BTS and PEBS configuration, respectively */ + struct bts_tracer *bts_master; + struct pebs_tracer *pebs_master; + /* use count */ + unsigned long count; + /* a pointer to the context location inside the thread_struct + * or the per_cpu context array */ + struct ds_context **this; + /* a pointer to the task owning this context, or NULL, if the + * context is owned by a cpu */ + struct task_struct *task; +}; + +static DEFINE_PER_CPU(struct ds_context *, system_context_array); -#define this_system_context per_cpu(system_context, smp_processor_id()) +#define system_context per_cpu(system_context_array, smp_processor_id()) static inline struct ds_context *ds_get_context(struct task_struct *task) { struct ds_context **p_context = - (task ? &task->thread.ds_ctx : &this_system_context); + (task ? &task->thread.ds_ctx : &system_context); struct ds_context *context = *p_context; unsigned long irq; @@ -225,10 +263,22 @@ static inline struct ds_context *ds_get_context(struct task_struct *task) wrmsrl(MSR_IA32_DS_AREA, (unsigned long)context->ds); } + + context->count++; + + spin_unlock_irqrestore(&ds_lock, irq); + } else { + spin_lock_irqsave(&ds_lock, irq); + + context = *p_context; + if (context) + context->count++; + spin_unlock_irqrestore(&ds_lock, irq); - } - context->count++; + if (!context) + context = ds_get_context(task); + } return context; } @@ -242,8 +292,10 @@ static inline void ds_put_context(struct ds_context *context) spin_lock_irqsave(&ds_lock, irq); - if (--context->count) - goto out; + if (--context->count) { + spin_unlock_irqrestore(&ds_lock, irq); + return; + } *(context->this) = NULL; @@ -253,14 +305,14 @@ static inline void ds_put_context(struct ds_context *context) if (!context->task || (context->task == current)) wrmsrl(MSR_IA32_DS_AREA, 0); - kfree(context); - out: spin_unlock_irqrestore(&ds_lock, irq); + + kfree(context); } /* - * Handle a buffer overflow + * Call the tracer's callback on a buffer overflow. * * context: the ds context * qual: the buffer type @@ -268,30 +320,244 @@ static inline void ds_put_context(struct ds_context *context) static void ds_overflow(struct ds_context *context, enum ds_qualifier qual) { switch (qual) { - case ds_bts: { - struct bts_tracer *tracer = - container_of(context->owner[qual], - struct bts_tracer, ds); - if (tracer->ovfl) - tracer->ovfl(tracer); - } + case ds_bts: + if (context->bts_master && + context->bts_master->ovfl) + context->bts_master->ovfl(context->bts_master); + break; + case ds_pebs: + if (context->pebs_master && + context->pebs_master->ovfl) + context->pebs_master->ovfl(context->pebs_master); break; - case ds_pebs: { - struct pebs_tracer *tracer = - container_of(context->owner[qual], - struct pebs_tracer, ds); - if (tracer->ovfl) - tracer->ovfl(tracer); } +} + + +/* + * Write raw data into the BTS or PEBS buffer. + * + * The remainder of any partially written record is zeroed out. + * + * context: the DS context + * qual: the buffer type + * record: the data to write + * size: the size of the data + */ +static int ds_write(struct ds_context *context, enum ds_qualifier qual, + const void *record, size_t size) +{ + int bytes_written = 0; + + if (!record) + return -EINVAL; + + while (size) { + unsigned long base, index, end, write_end, int_th; + unsigned long write_size, adj_write_size; + + /* + * write as much as possible without producing an + * overflow interrupt. + * + * interrupt_threshold must either be + * - bigger than absolute_maximum or + * - point to a record between buffer_base and absolute_maximum + * + * index points to a valid record. + */ + base = ds_get(context->ds, qual, ds_buffer_base); + index = ds_get(context->ds, qual, ds_index); + end = ds_get(context->ds, qual, ds_absolute_maximum); + int_th = ds_get(context->ds, qual, ds_interrupt_threshold); + + write_end = min(end, int_th); + + /* if we are already beyond the interrupt threshold, + * we fill the entire buffer */ + if (write_end <= index) + write_end = end; + + if (write_end <= index) + break; + + write_size = min((unsigned long) size, write_end - index); + memcpy((void *)index, record, write_size); + + record = (const char *)record + write_size; + size -= write_size; + bytes_written += write_size; + + adj_write_size = write_size / ds_cfg.sizeof_rec[qual]; + adj_write_size *= ds_cfg.sizeof_rec[qual]; + + /* zero out trailing bytes */ + memset((char *)index + write_size, 0, + adj_write_size - write_size); + index += adj_write_size; + + if (index >= end) + index = base; + ds_set(context->ds, qual, ds_index, index); + + if (index >= int_th) + ds_overflow(context, qual); + } + + return bytes_written; +} + + +/* + * Branch Trace Store (BTS) uses the following format. Different + * architectures vary in the size of those fields. + * - source linear address + * - destination linear address + * - flags + * + * Later architectures use 64bit pointers throughout, whereas earlier + * architectures use 32bit pointers in 32bit mode. + * + * We compute the base address for the first 8 fields based on: + * - the field size stored in the DS configuration + * - the relative field position + * + * In order to store additional information in the BTS buffer, we use + * a special source address to indicate that the record requires + * special interpretation. + * + * Netburst indicated via a bit in the flags field whether the branch + * was predicted; this is ignored. + * + * We use two levels of abstraction: + * - the raw data level defined here + * - an arch-independent level defined in ds.h + */ + +enum bts_field { + bts_from, + bts_to, + bts_flags, + + bts_qual = bts_from, + bts_jiffies = bts_to, + bts_pid = bts_flags, + + bts_qual_mask = (bts_qual_max - 1), + bts_escape = ((unsigned long)-1 & ~bts_qual_mask) +}; + +static inline unsigned long bts_get(const char *base, enum bts_field field) +{ + base += (ds_cfg.sizeof_field * field); + return *(unsigned long *)base; +} + +static inline void bts_set(char *base, enum bts_field field, unsigned long val) +{ + base += (ds_cfg.sizeof_field * field);; + (*(unsigned long *)base) = val; +} + + +/* + * The raw BTS data is architecture dependent. + * + * For higher-level users, we give an arch-independent view. + * - ds.h defines struct bts_struct + * - bts_read translates one raw bts record into a bts_struct + * - bts_write translates one bts_struct into the raw format and + * writes it into the top of the parameter tracer's buffer. + * + * return: bytes read/written on success; -Eerrno, otherwise + */ +static int bts_read(struct bts_tracer *tracer, const void *at, + struct bts_struct *out) +{ + if (!tracer) + return -EINVAL; + + if (at < tracer->trace.ds.begin) + return -EINVAL; + + if (tracer->trace.ds.end < (at + tracer->trace.ds.size)) + return -EINVAL; + + memset(out, 0, sizeof(*out)); + if ((bts_get(at, bts_qual) & ~bts_qual_mask) == bts_escape) { + out->qualifier = (bts_get(at, bts_qual) & bts_qual_mask); + out->variant.timestamp.jiffies = bts_get(at, bts_jiffies); + out->variant.timestamp.pid = bts_get(at, bts_pid); + } else { + out->qualifier = bts_branch; + out->variant.lbr.from = bts_get(at, bts_from); + out->variant.lbr.to = bts_get(at, bts_to); + } + + return ds_cfg.sizeof_rec[ds_bts]; +} + +static int bts_write(struct bts_tracer *tracer, const struct bts_struct *in) +{ + unsigned char raw[MAX_SIZEOF_BTS]; + + if (!tracer) + return -EINVAL; + + if (MAX_SIZEOF_BTS < ds_cfg.sizeof_rec[ds_bts]) + return -EOVERFLOW; + + switch (in->qualifier) { + case bts_invalid: + bts_set(raw, bts_from, 0); + bts_set(raw, bts_to, 0); + bts_set(raw, bts_flags, 0); + break; + case bts_branch: + bts_set(raw, bts_from, in->variant.lbr.from); + bts_set(raw, bts_to, in->variant.lbr.to); + bts_set(raw, bts_flags, 0); + break; + case bts_task_arrives: + case bts_task_departs: + bts_set(raw, bts_qual, (bts_escape | in->qualifier)); + bts_set(raw, bts_jiffies, in->variant.timestamp.jiffies); + bts_set(raw, bts_pid, in->variant.timestamp.pid); break; + default: + return -EINVAL; } + + return ds_write(tracer->ds.context, ds_bts, raw, + ds_cfg.sizeof_rec[ds_bts]); } -static void ds_install_ds_config(struct ds_context *context, - enum ds_qualifier qual, - void *base, size_t size, size_t ith) +static void ds_write_config(struct ds_context *context, + struct ds_trace *cfg, enum ds_qualifier qual) +{ + unsigned char *ds = context->ds; + + ds_set(ds, qual, ds_buffer_base, (unsigned long)cfg->begin); + ds_set(ds, qual, ds_index, (unsigned long)cfg->top); + ds_set(ds, qual, ds_absolute_maximum, (unsigned long)cfg->end); + ds_set(ds, qual, ds_interrupt_threshold, (unsigned long)cfg->ith); +} + +static void ds_read_config(struct ds_context *context, + struct ds_trace *cfg, enum ds_qualifier qual) { + unsigned char *ds = context->ds; + + cfg->begin = (void *)ds_get(ds, qual, ds_buffer_base); + cfg->top = (void *)ds_get(ds, qual, ds_index); + cfg->end = (void *)ds_get(ds, qual, ds_absolute_maximum); + cfg->ith = (void *)ds_get(ds, qual, ds_interrupt_threshold); +} + +static void ds_init_ds_trace(struct ds_trace *trace, enum ds_qualifier qual, + void *base, size_t size, size_t ith, + unsigned int flags) { unsigned long buffer, adj; /* adjust the buffer address and size to meet alignment @@ -308,32 +574,30 @@ static void ds_install_ds_config(struct ds_context *context, buffer += adj; size -= adj; - size /= ds_cfg.sizeof_rec[qual]; - size *= ds_cfg.sizeof_rec[qual]; + trace->n = size / ds_cfg.sizeof_rec[qual]; + trace->size = ds_cfg.sizeof_rec[qual]; - ds_set(context->ds, qual, ds_buffer_base, buffer); - ds_set(context->ds, qual, ds_index, buffer); - ds_set(context->ds, qual, ds_absolute_maximum, buffer + size); + size = (trace->n * trace->size); + trace->begin = (void *)buffer; + trace->top = trace->begin; + trace->end = (void *)(buffer + size); /* The value for 'no threshold' is -1, which will set the * threshold outside of the buffer, just like we want it. */ - ds_set(context->ds, qual, - ds_interrupt_threshold, buffer + size - ith); + trace->ith = (void *)(buffer + size - ith); + + trace->flags = flags; } -static int ds_request(struct ds_tracer *tracer, enum ds_qualifier qual, - struct task_struct *task, - void *base, size_t size, size_t th) + +static int ds_request(struct ds_tracer *tracer, struct ds_trace *trace, + enum ds_qualifier qual, struct task_struct *task, + void *base, size_t size, size_t th, unsigned int flags) { struct ds_context *context; - unsigned long irq; int error; - error = -EOPNOTSUPP; - if (!ds_cfg.sizeof_ds) - goto out; - error = -EINVAL; if (!base) goto out; @@ -360,43 +624,26 @@ static int ds_request(struct ds_tracer *tracer, enum ds_qualifier qual, goto out; tracer->context = context; + ds_init_ds_trace(trace, qual, base, size, th, flags); - spin_lock_irqsave(&ds_lock, irq); - - error = -EPERM; - if (!check_tracer(task)) - goto out_unlock; - get_tracer(task); - - error = -EPERM; - if (context->owner[qual]) - goto out_put_tracer; - context->owner[qual] = tracer; - - spin_unlock_irqrestore(&ds_lock, irq); - - - ds_install_ds_config(context, qual, base, size, th); - - return 0; - - out_put_tracer: - put_tracer(task); - out_unlock: - spin_unlock_irqrestore(&ds_lock, irq); - ds_put_context(context); - tracer->context = NULL; + error = 0; out: return error; } struct bts_tracer *ds_request_bts(struct task_struct *task, void *base, size_t size, - bts_ovfl_callback_t ovfl, size_t th) + bts_ovfl_callback_t ovfl, size_t th, + unsigned int flags) { struct bts_tracer *tracer; + unsigned long irq; int error; + error = -EOPNOTSUPP; + if (!ds_cfg.ctl[dsf_bts]) + goto out; + /* buffer overflow notification is not yet implemented */ error = -EOPNOTSUPP; if (ovfl) @@ -408,12 +655,40 @@ struct bts_tracer *ds_request_bts(struct task_struct *task, goto out; tracer->ovfl = ovfl; - error = ds_request(&tracer->ds, ds_bts, task, base, size, th); + error = ds_request(&tracer->ds, &tracer->trace.ds, + ds_bts, task, base, size, th, flags); if (error < 0) goto out_tracer; + + spin_lock_irqsave(&ds_lock, irq); + + error = -EPERM; + if (!check_tracer(task)) + goto out_unlock; + get_tracer(task); + + error = -EPERM; + if (tracer->ds.context->bts_master) + goto out_put_tracer; + tracer->ds.context->bts_master = tracer; + + spin_unlock_irqrestore(&ds_lock, irq); + + + tracer->trace.read = bts_read; + tracer->trace.write = bts_write; + + ds_write_config(tracer->ds.context, &tracer->trace.ds, ds_bts); + ds_resume_bts(tracer); + return tracer; + out_put_tracer: + put_tracer(task); + out_unlock: + spin_unlock_irqrestore(&ds_lock, irq); + ds_put_context(tracer->ds.context); out_tracer: kfree(tracer); out: @@ -422,9 +697,11 @@ struct bts_tracer *ds_request_bts(struct task_struct *task, struct pebs_tracer *ds_request_pebs(struct task_struct *task, void *base, size_t size, - pebs_ovfl_callback_t ovfl, size_t th) + pebs_ovfl_callback_t ovfl, size_t th, + unsigned int flags) { struct pebs_tracer *tracer; + unsigned long irq; int error; /* buffer overflow notification is not yet implemented */ @@ -438,300 +715,171 @@ struct pebs_tracer *ds_request_pebs(struct task_struct *task, goto out; tracer->ovfl = ovfl; - error = ds_request(&tracer->ds, ds_pebs, task, base, size, th); + error = ds_request(&tracer->ds, &tracer->trace.ds, + ds_pebs, task, base, size, th, flags); if (error < 0) goto out_tracer; + spin_lock_irqsave(&ds_lock, irq); + + error = -EPERM; + if (!check_tracer(task)) + goto out_unlock; + get_tracer(task); + + error = -EPERM; + if (tracer->ds.context->pebs_master) + goto out_put_tracer; + tracer->ds.context->pebs_master = tracer; + + spin_unlock_irqrestore(&ds_lock, irq); + + ds_write_config(tracer->ds.context, &tracer->trace.ds, ds_bts); + ds_resume_pebs(tracer); + return tracer; + out_put_tracer: + put_tracer(task); + out_unlock: + spin_unlock_irqrestore(&ds_lock, irq); + ds_put_context(tracer->ds.context); out_tracer: kfree(tracer); out: return ERR_PTR(error); } -static void ds_release(struct ds_tracer *tracer, enum ds_qualifier qual) -{ - WARN_ON_ONCE(tracer->context->owner[qual] != tracer); - tracer->context->owner[qual] = NULL; - - put_tracer(tracer->context->task); - ds_put_context(tracer->context); -} - -int ds_release_bts(struct bts_tracer *tracer) +void ds_release_bts(struct bts_tracer *tracer) { if (!tracer) - return -EINVAL; + return; - ds_release(&tracer->ds, ds_bts); - kfree(tracer); + ds_suspend_bts(tracer); - return 0; -} + WARN_ON_ONCE(tracer->ds.context->bts_master != tracer); + tracer->ds.context->bts_master = NULL; -int ds_release_pebs(struct pebs_tracer *tracer) -{ - if (!tracer) - return -EINVAL; + put_tracer(tracer->ds.context->task); + ds_put_context(tracer->ds.context); - ds_release(&tracer->ds, ds_pebs); kfree(tracer); - - return 0; -} - -static size_t ds_get_index(struct ds_context *context, enum ds_qualifier qual) -{ - unsigned long base, index; - - base = ds_get(context->ds, qual, ds_buffer_base); - index = ds_get(context->ds, qual, ds_index); - - return (index - base) / ds_cfg.sizeof_rec[qual]; } -int ds_get_bts_index(struct bts_tracer *tracer, size_t *pos) +void ds_suspend_bts(struct bts_tracer *tracer) { - if (!tracer) - return -EINVAL; + struct task_struct *task; - if (!pos) - return -EINVAL; - - *pos = ds_get_index(tracer->ds.context, ds_bts); - - return 0; -} - -int ds_get_pebs_index(struct pebs_tracer *tracer, size_t *pos) -{ if (!tracer) - return -EINVAL; + return; - if (!pos) - return -EINVAL; + task = tracer->ds.context->task; - *pos = ds_get_index(tracer->ds.context, ds_pebs); + if (!task || (task == current)) + update_debugctlmsr(get_debugctlmsr() & ~BTS_CONTROL); - return 0; -} + if (task) { + task->thread.debugctlmsr &= ~BTS_CONTROL; -static size_t ds_get_end(struct ds_context *context, enum ds_qualifier qual) -{ - unsigned long base, max; - - base = ds_get(context->ds, qual, ds_buffer_base); - max = ds_get(context->ds, qual, ds_absolute_maximum); - - return (max - base) / ds_cfg.sizeof_rec[qual]; + if (!task->thread.debugctlmsr) + clear_tsk_thread_flag(task, TIF_DEBUGCTLMSR); + } } -int ds_get_bts_end(struct bts_tracer *tracer, size_t *pos) +void ds_resume_bts(struct bts_tracer *tracer) { - if (!tracer) - return -EINVAL; - - if (!pos) - return -EINVAL; - - *pos = ds_get_end(tracer->ds.context, ds_bts); - - return 0; -} + struct task_struct *task; + unsigned long control; -int ds_get_pebs_end(struct pebs_tracer *tracer, size_t *pos) -{ if (!tracer) - return -EINVAL; - - if (!pos) - return -EINVAL; - - *pos = ds_get_end(tracer->ds.context, ds_pebs); - - return 0; -} - -static int ds_access(struct ds_context *context, enum ds_qualifier qual, - size_t index, const void **record) -{ - unsigned long base, idx; - - if (!record) - return -EINVAL; - - base = ds_get(context->ds, qual, ds_buffer_base); - idx = base + (index * ds_cfg.sizeof_rec[qual]); - - if (idx > ds_get(context->ds, qual, ds_absolute_maximum)) - return -EINVAL; + return; - *record = (const void *)idx; + task = tracer->ds.context->task; - return ds_cfg.sizeof_rec[qual]; -} + control = ds_cfg.ctl[dsf_bts]; + if (!(tracer->trace.ds.flags & BTS_KERNEL)) + control |= ds_cfg.ctl[dsf_bts_kernel]; + if (!(tracer->trace.ds.flags & BTS_USER)) + control |= ds_cfg.ctl[dsf_bts_user]; -int ds_access_bts(struct bts_tracer *tracer, size_t index, - const void **record) -{ - if (!tracer) - return -EINVAL; + if (task) { + task->thread.debugctlmsr |= control; + set_tsk_thread_flag(task, TIF_DEBUGCTLMSR); + } - return ds_access(tracer->ds.context, ds_bts, index, record); + if (!task || (task == current)) + update_debugctlmsr(get_debugctlmsr() | control); } -int ds_access_pebs(struct pebs_tracer *tracer, size_t index, - const void **record) +void ds_release_pebs(struct pebs_tracer *tracer) { if (!tracer) - return -EINVAL; - - return ds_access(tracer->ds.context, ds_pebs, index, record); -} - -static int ds_write(struct ds_context *context, enum ds_qualifier qual, - const void *record, size_t size) -{ - int bytes_written = 0; - - if (!record) - return -EINVAL; - - while (size) { - unsigned long base, index, end, write_end, int_th; - unsigned long write_size, adj_write_size; - - /* - * write as much as possible without producing an - * overflow interrupt. - * - * interrupt_threshold must either be - * - bigger than absolute_maximum or - * - point to a record between buffer_base and absolute_maximum - * - * index points to a valid record. - */ - base = ds_get(context->ds, qual, ds_buffer_base); - index = ds_get(context->ds, qual, ds_index); - end = ds_get(context->ds, qual, ds_absolute_maximum); - int_th = ds_get(context->ds, qual, ds_interrupt_threshold); - - write_end = min(end, int_th); - - /* if we are already beyond the interrupt threshold, - * we fill the entire buffer */ - if (write_end <= index) - write_end = end; - - if (write_end <= index) - break; - - write_size = min((unsigned long) size, write_end - index); - memcpy((void *)index, record, write_size); - - record = (const char *)record + write_size; - size -= write_size; - bytes_written += write_size; - - adj_write_size = write_size / ds_cfg.sizeof_rec[qual]; - adj_write_size *= ds_cfg.sizeof_rec[qual]; - - /* zero out trailing bytes */ - memset((char *)index + write_size, 0, - adj_write_size - write_size); - index += adj_write_size; + return; - if (index >= end) - index = base; - ds_set(context->ds, qual, ds_index, index); + ds_suspend_pebs(tracer); - if (index >= int_th) - ds_overflow(context, qual); - } + WARN_ON_ONCE(tracer->ds.context->pebs_master != tracer); + tracer->ds.context->pebs_master = NULL; - return bytes_written; -} + put_tracer(tracer->ds.context->task); + ds_put_context(tracer->ds.context); -int ds_write_bts(struct bts_tracer *tracer, const void *record, size_t size) -{ - if (!tracer) - return -EINVAL; - - return ds_write(tracer->ds.context, ds_bts, record, size); + kfree(tracer); } -int ds_write_pebs(struct pebs_tracer *tracer, const void *record, size_t size) +void ds_suspend_pebs(struct pebs_tracer *tracer) { - if (!tracer) - return -EINVAL; - return ds_write(tracer->ds.context, ds_pebs, record, size); } -static void ds_reset_or_clear(struct ds_context *context, - enum ds_qualifier qual, int clear) +void ds_resume_pebs(struct pebs_tracer *tracer) { - unsigned long base, end; - - base = ds_get(context->ds, qual, ds_buffer_base); - end = ds_get(context->ds, qual, ds_absolute_maximum); - - if (clear) - memset((void *)base, 0, end - base); - ds_set(context->ds, qual, ds_index, base); } -int ds_reset_bts(struct bts_tracer *tracer) +const struct bts_trace *ds_read_bts(struct bts_tracer *tracer) { if (!tracer) - return -EINVAL; - - ds_reset_or_clear(tracer->ds.context, ds_bts, /* clear = */ 0); + return NULL; - return 0; + ds_read_config(tracer->ds.context, &tracer->trace.ds, ds_bts); + return &tracer->trace; } -int ds_reset_pebs(struct pebs_tracer *tracer) +const struct pebs_trace *ds_read_pebs(struct pebs_tracer *tracer) { if (!tracer) - return -EINVAL; + return NULL; - ds_reset_or_clear(tracer->ds.context, ds_pebs, /* clear = */ 0); + ds_read_config(tracer->ds.context, &tracer->trace.ds, ds_pebs); + tracer->trace.reset_value = + *(u64 *)(tracer->ds.context->ds + (ds_cfg.sizeof_field * 8)); - return 0; + return &tracer->trace; } -int ds_clear_bts(struct bts_tracer *tracer) +int ds_reset_bts(struct bts_tracer *tracer) { if (!tracer) return -EINVAL; - ds_reset_or_clear(tracer->ds.context, ds_bts, /* clear = */ 1); - - return 0; -} - -int ds_clear_pebs(struct pebs_tracer *tracer) -{ - if (!tracer) - return -EINVAL; + tracer->trace.ds.top = tracer->trace.ds.begin; - ds_reset_or_clear(tracer->ds.context, ds_pebs, /* clear = */ 1); + ds_set(tracer->ds.context->ds, ds_bts, ds_index, + (unsigned long)tracer->trace.ds.top); return 0; } -int ds_get_pebs_reset(struct pebs_tracer *tracer, u64 *value) +int ds_reset_pebs(struct pebs_tracer *tracer) { if (!tracer) return -EINVAL; - if (!value) - return -EINVAL; + tracer->trace.ds.top = tracer->trace.ds.begin; - *value = *(u64 *)(tracer->ds.context->ds + (ds_cfg.sizeof_field * 8)); + ds_set(tracer->ds.context->ds, ds_bts, ds_index, + (unsigned long)tracer->trace.ds.top); return 0; } @@ -746,35 +894,59 @@ int ds_set_pebs_reset(struct pebs_tracer *tracer, u64 value) return 0; } -static const struct ds_configuration ds_cfg_var = { - .sizeof_ds = sizeof(long) * 12, - .sizeof_field = sizeof(long), - .sizeof_rec[ds_bts] = sizeof(long) * 3, +static const struct ds_configuration ds_cfg_netburst = { + .name = "netburst", + .ctl[dsf_bts] = (1 << 2) | (1 << 3), + .ctl[dsf_bts_kernel] = (1 << 5), + .ctl[dsf_bts_user] = (1 << 6), + + .sizeof_field = sizeof(long), + .sizeof_rec[ds_bts] = sizeof(long) * 3, #ifdef __i386__ - .sizeof_rec[ds_pebs] = sizeof(long) * 10 + .sizeof_rec[ds_pebs] = sizeof(long) * 10, #else - .sizeof_rec[ds_pebs] = sizeof(long) * 18 + .sizeof_rec[ds_pebs] = sizeof(long) * 18, #endif }; -static const struct ds_configuration ds_cfg_64 = { - .sizeof_ds = 8 * 12, - .sizeof_field = 8, - .sizeof_rec[ds_bts] = 8 * 3, +static const struct ds_configuration ds_cfg_pentium_m = { + .name = "pentium m", + .ctl[dsf_bts] = (1 << 6) | (1 << 7), + + .sizeof_field = sizeof(long), + .sizeof_rec[ds_bts] = sizeof(long) * 3, #ifdef __i386__ - .sizeof_rec[ds_pebs] = 8 * 10 + .sizeof_rec[ds_pebs] = sizeof(long) * 10, #else - .sizeof_rec[ds_pebs] = 8 * 18 + .sizeof_rec[ds_pebs] = sizeof(long) * 18, #endif }; +static const struct ds_configuration ds_cfg_core2 = { + .name = "core 2", + .ctl[dsf_bts] = (1 << 6) | (1 << 7), + .ctl[dsf_bts_kernel] = (1 << 9), + .ctl[dsf_bts_user] = (1 << 10), + + .sizeof_field = 8, + .sizeof_rec[ds_bts] = 8 * 3, + .sizeof_rec[ds_pebs] = 8 * 18, +}; -static inline void +static void ds_configure(const struct ds_configuration *cfg) { + memset(&ds_cfg, 0, sizeof(ds_cfg)); ds_cfg = *cfg; - printk(KERN_INFO "DS available\n"); + printk(KERN_INFO "[ds] using %s configuration\n", ds_cfg.name); + + if (!cpu_has_bts) { + ds_cfg.ctl[dsf_bts] = 0; + printk(KERN_INFO "[ds] bts not available\n"); + } + if (!cpu_has_pebs) + printk(KERN_INFO "[ds] pebs not available\n"); - WARN_ON_ONCE(MAX_SIZEOF_DS < ds_cfg.sizeof_ds); + WARN_ON_ONCE(MAX_SIZEOF_DS < (12 * ds_cfg.sizeof_field)); } void __cpuinit ds_init_intel(struct cpuinfo_x86 *c) @@ -787,10 +959,10 @@ void __cpuinit ds_init_intel(struct cpuinfo_x86 *c) break; case 0xD: case 0xE: /* Pentium M */ - ds_configure(&ds_cfg_var); + ds_configure(&ds_cfg_pentium_m); break; default: /* Core2, Atom, ... */ - ds_configure(&ds_cfg_64); + ds_configure(&ds_cfg_core2); break; } break; @@ -799,7 +971,7 @@ void __cpuinit ds_init_intel(struct cpuinfo_x86 *c) case 0x0: case 0x1: case 0x2: /* Netburst */ - ds_configure(&ds_cfg_var); + ds_configure(&ds_cfg_netburst); break; default: /* sorry, don't know about them */ @@ -812,14 +984,41 @@ void __cpuinit ds_init_intel(struct cpuinfo_x86 *c) } } -void ds_free(struct ds_context *context) +/* + * Change the DS configuration from tracing prev to tracing next. + */ +void ds_switch_to(struct task_struct *prev, struct task_struct *next) { - /* This is called when the task owning the parameter context - * is dying. There should not be any user of that context left - * to disturb us, anymore. */ - unsigned long leftovers = context->count; - while (leftovers--) { - put_tracer(context->task); - ds_put_context(context); + struct ds_context *prev_ctx = prev->thread.ds_ctx; + struct ds_context *next_ctx = next->thread.ds_ctx; + + if (prev_ctx) { + update_debugctlmsr(0); + + if (prev_ctx->bts_master && + (prev_ctx->bts_master->trace.ds.flags & BTS_TIMESTAMPS)) { + struct bts_struct ts = { + .qualifier = bts_task_departs, + .variant.timestamp.jiffies = jiffies_64, + .variant.timestamp.pid = prev->pid + }; + bts_write(prev_ctx->bts_master, &ts); + } + } + + if (next_ctx) { + if (next_ctx->bts_master && + (next_ctx->bts_master->trace.ds.flags & BTS_TIMESTAMPS)) { + struct bts_struct ts = { + .qualifier = bts_task_arrives, + .variant.timestamp.jiffies = jiffies_64, + .variant.timestamp.pid = next->pid + }; + bts_write(next_ctx->bts_master, &ts); + } + + wrmsrl(MSR_IA32_DS_AREA, (unsigned long)next_ctx->ds); } + + update_debugctlmsr(next->thread.debugctlmsr); } diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c index 24c2276aa453..605eff9a8ac0 100644 --- a/arch/x86/kernel/process_32.c +++ b/arch/x86/kernel/process_32.c @@ -252,11 +252,14 @@ void exit_thread(void) put_cpu(); } #ifdef CONFIG_X86_DS - /* Free any DS contexts that have not been properly released. */ - if (unlikely(current->thread.ds_ctx)) { - /* we clear debugctl to make sure DS is not used. */ - update_debugctlmsr(0); - ds_free(current->thread.ds_ctx); + /* Free any BTS tracers that have not been properly released. */ + if (unlikely(current->bts)) { + ds_release_bts(current->bts); + current->bts = NULL; + + kfree(current->bts_buffer); + current->bts_buffer = NULL; + current->bts_size = 0; } #endif /* CONFIG_X86_DS */ } @@ -420,48 +423,19 @@ int set_tsc_mode(unsigned int val) return 0; } -#ifdef CONFIG_X86_DS -static int update_debugctl(struct thread_struct *prev, - struct thread_struct *next, unsigned long debugctl) -{ - unsigned long ds_prev = 0; - unsigned long ds_next = 0; - - if (prev->ds_ctx) - ds_prev = (unsigned long)prev->ds_ctx->ds; - if (next->ds_ctx) - ds_next = (unsigned long)next->ds_ctx->ds; - - if (ds_next != ds_prev) { - /* we clear debugctl to make sure DS - * is not in use when we change it */ - debugctl = 0; - update_debugctlmsr(0); - wrmsr(MSR_IA32_DS_AREA, ds_next, 0); - } - return debugctl; -} -#else -static int update_debugctl(struct thread_struct *prev, - struct thread_struct *next, unsigned long debugctl) -{ - return debugctl; -} -#endif /* CONFIG_X86_DS */ - static noinline void __switch_to_xtra(struct task_struct *prev_p, struct task_struct *next_p, struct tss_struct *tss) { struct thread_struct *prev, *next; - unsigned long debugctl; prev = &prev_p->thread; next = &next_p->thread; - debugctl = update_debugctl(prev, next, prev->debugctlmsr); - - if (next->debugctlmsr != debugctl) + if (test_tsk_thread_flag(next_p, TIF_DS_AREA_MSR) || + test_tsk_thread_flag(prev_p, TIF_DS_AREA_MSR)) + ds_switch_to(prev_p, next_p); + else if (next->debugctlmsr != prev->debugctlmsr) update_debugctlmsr(next->debugctlmsr); if (test_tsk_thread_flag(next_p, TIF_DEBUG)) { @@ -483,15 +457,6 @@ __switch_to_xtra(struct task_struct *prev_p, struct task_struct *next_p, hard_enable_TSC(); } -#ifdef CONFIG_X86_PTRACE_BTS - if (test_tsk_thread_flag(prev_p, TIF_BTS_TRACE_TS)) - ptrace_bts_take_timestamp(prev_p, BTS_TASK_DEPARTS); - - if (test_tsk_thread_flag(next_p, TIF_BTS_TRACE_TS)) - ptrace_bts_take_timestamp(next_p, BTS_TASK_ARRIVES); -#endif /* CONFIG_X86_PTRACE_BTS */ - - if (!test_tsk_thread_flag(next_p, TIF_IO_BITMAP)) { /* * Disable the bitmap via an invalid offset. We still cache diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c index fbb321d53d34..1cfd2a4bf853 100644 --- a/arch/x86/kernel/process_64.c +++ b/arch/x86/kernel/process_64.c @@ -237,11 +237,14 @@ void exit_thread(void) put_cpu(); } #ifdef CONFIG_X86_DS - /* Free any DS contexts that have not been properly released. */ - if (unlikely(t->ds_ctx)) { - /* we clear debugctl to make sure DS is not used. */ - update_debugctlmsr(0); - ds_free(t->ds_ctx); + /* Free any BTS tracers that have not been properly released. */ + if (unlikely(current->bts)) { + ds_release_bts(current->bts); + current->bts = NULL; + + kfree(current->bts_buffer); + current->bts_buffer = NULL; + current->bts_size = 0; } #endif /* CONFIG_X86_DS */ } @@ -471,35 +474,14 @@ static inline void __switch_to_xtra(struct task_struct *prev_p, struct tss_struct *tss) { struct thread_struct *prev, *next; - unsigned long debugctl; prev = &prev_p->thread, next = &next_p->thread; - debugctl = prev->debugctlmsr; - -#ifdef CONFIG_X86_DS - { - unsigned long ds_prev = 0, ds_next = 0; - - if (prev->ds_ctx) - ds_prev = (unsigned long)prev->ds_ctx->ds; - if (next->ds_ctx) - ds_next = (unsigned long)next->ds_ctx->ds; - - if (ds_next != ds_prev) { - /* - * We clear debugctl to make sure DS - * is not in use when we change it: - */ - debugctl = 0; - update_debugctlmsr(0); - wrmsrl(MSR_IA32_DS_AREA, ds_next); - } - } -#endif /* CONFIG_X86_DS */ - - if (next->debugctlmsr != debugctl) + if (test_tsk_thread_flag(next_p, TIF_DS_AREA_MSR) || + test_tsk_thread_flag(prev_p, TIF_DS_AREA_MSR)) + ds_switch_to(prev_p, next_p); + else if (next->debugctlmsr != prev->debugctlmsr) update_debugctlmsr(next->debugctlmsr); if (test_tsk_thread_flag(next_p, TIF_DEBUG)) { @@ -534,14 +516,6 @@ static inline void __switch_to_xtra(struct task_struct *prev_p, */ memset(tss->io_bitmap, 0xff, prev->io_bitmap_max); } - -#ifdef CONFIG_X86_PTRACE_BTS - if (test_tsk_thread_flag(prev_p, TIF_BTS_TRACE_TS)) - ptrace_bts_take_timestamp(prev_p, BTS_TASK_DEPARTS); - - if (test_tsk_thread_flag(next_p, TIF_BTS_TRACE_TS)) - ptrace_bts_take_timestamp(next_p, BTS_TASK_ARRIVES); -#endif /* CONFIG_X86_PTRACE_BTS */ } /* diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c index b2998fe1166b..45e9855da2d2 100644 --- a/arch/x86/kernel/ptrace.c +++ b/arch/x86/kernel/ptrace.c @@ -581,153 +581,73 @@ static int ioperm_get(struct task_struct *target, } #ifdef CONFIG_X86_PTRACE_BTS -/* - * The configuration for a particular BTS hardware implementation. - */ -struct bts_configuration { - /* the size of a BTS record in bytes; at most BTS_MAX_RECORD_SIZE */ - unsigned char sizeof_bts; - /* the size of a field in the BTS record in bytes */ - unsigned char sizeof_field; - /* a bitmask to enable/disable BTS in DEBUGCTL MSR */ - unsigned long debugctl_mask; -}; -static struct bts_configuration bts_cfg; - -#define BTS_MAX_RECORD_SIZE (8 * 3) - - -/* - * Branch Trace Store (BTS) uses the following format. Different - * architectures vary in the size of those fields. - * - source linear address - * - destination linear address - * - flags - * - * Later architectures use 64bit pointers throughout, whereas earlier - * architectures use 32bit pointers in 32bit mode. - * - * We compute the base address for the first 8 fields based on: - * - the field size stored in the DS configuration - * - the relative field position - * - * In order to store additional information in the BTS buffer, we use - * a special source address to indicate that the record requires - * special interpretation. - * - * Netburst indicated via a bit in the flags field whether the branch - * was predicted; this is ignored. - */ - -enum bts_field { - bts_from = 0, - bts_to, - bts_flags, - - bts_escape = (unsigned long)-1, - bts_qual = bts_to, - bts_jiffies = bts_flags -}; - -static inline unsigned long bts_get(const char *base, enum bts_field field) -{ - base += (bts_cfg.sizeof_field * field); - return *(unsigned long *)base; -} - -static inline void bts_set(char *base, enum bts_field field, unsigned long val) -{ - base += (bts_cfg.sizeof_field * field);; - (*(unsigned long *)base) = val; -} - -/* - * Translate a BTS record from the raw format into the bts_struct format - * - * out (out): bts_struct interpretation - * raw: raw BTS record - */ -static void ptrace_bts_translate_record(struct bts_struct *out, const void *raw) -{ - memset(out, 0, sizeof(*out)); - if (bts_get(raw, bts_from) == bts_escape) { - out->qualifier = bts_get(raw, bts_qual); - out->variant.jiffies = bts_get(raw, bts_jiffies); - } else { - out->qualifier = BTS_BRANCH; - out->variant.lbr.from_ip = bts_get(raw, bts_from); - out->variant.lbr.to_ip = bts_get(raw, bts_to); - } -} - static int ptrace_bts_read_record(struct task_struct *child, size_t index, struct bts_struct __user *out) { - struct bts_struct ret; - const void *bts_record; - size_t bts_index, bts_end; + const struct bts_trace *trace; + struct bts_struct bts; + const unsigned char *at; int error; - error = ds_get_bts_end(child->bts, &bts_end); - if (error < 0) - return error; - - if (bts_end <= index) - return -EINVAL; + trace = ds_read_bts(child->bts); + if (!trace) + return -EPERM; - error = ds_get_bts_index(child->bts, &bts_index); - if (error < 0) - return error; + at = trace->ds.top - ((index + 1) * trace->ds.size); + if ((void *)at < trace->ds.begin) + at += (trace->ds.n * trace->ds.size); - /* translate the ptrace bts index into the ds bts index */ - bts_index += bts_end - (index + 1); - if (bts_end <= bts_index) - bts_index -= bts_end; + if (!trace->read) + return -EOPNOTSUPP; - error = ds_access_bts(child->bts, bts_index, &bts_record); + error = trace->read(child->bts, at, &bts); if (error < 0) return error; - ptrace_bts_translate_record(&ret, bts_record); - - if (copy_to_user(out, &ret, sizeof(ret))) + if (copy_to_user(out, &bts, sizeof(bts))) return -EFAULT; - return sizeof(ret); + return sizeof(bts); } static int ptrace_bts_drain(struct task_struct *child, long size, struct bts_struct __user *out) { - struct bts_struct ret; - const unsigned char *raw; - size_t end, i; - int error; + const struct bts_trace *trace; + const unsigned char *at; + int error, drained = 0; - error = ds_get_bts_index(child->bts, &end); - if (error < 0) - return error; + trace = ds_read_bts(child->bts); + if (!trace) + return -EPERM; - if (size < (end * sizeof(struct bts_struct))) + if (!trace->read) + return -EOPNOTSUPP; + + if (size < (trace->ds.top - trace->ds.begin)) return -EIO; - error = ds_access_bts(child->bts, 0, (const void **)&raw); - if (error < 0) - return error; + for (at = trace->ds.begin; (void *)at < trace->ds.top; + out++, drained++, at += trace->ds.size) { + struct bts_struct bts; + int error; - for (i = 0; i < end; i++, out++, raw += bts_cfg.sizeof_bts) { - ptrace_bts_translate_record(&ret, raw); + error = trace->read(child->bts, at, &bts); + if (error < 0) + return error; - if (copy_to_user(out, &ret, sizeof(ret))) + if (copy_to_user(out, &bts, sizeof(bts))) return -EFAULT; } - error = ds_clear_bts(child->bts); + memset(trace->ds.begin, 0, trace->ds.n * trace->ds.size); + + error = ds_reset_bts(child->bts); if (error < 0) return error; - return end; + return drained; } static int ptrace_bts_config(struct task_struct *child, @@ -735,136 +655,89 @@ static int ptrace_bts_config(struct task_struct *child, const struct ptrace_bts_config __user *ucfg) { struct ptrace_bts_config cfg; - int error = 0; - - error = -EOPNOTSUPP; - if (!bts_cfg.sizeof_bts) - goto errout; + unsigned int flags = 0; - error = -EIO; if (cfg_size < sizeof(cfg)) - goto errout; + return -EIO; - error = -EFAULT; if (copy_from_user(&cfg, ucfg, sizeof(cfg))) - goto errout; - - error = -EINVAL; - if ((cfg.flags & PTRACE_BTS_O_SIGNAL) && - !(cfg.flags & PTRACE_BTS_O_ALLOC)) - goto errout; - - if (cfg.flags & PTRACE_BTS_O_ALLOC) { - bts_ovfl_callback_t ovfl = NULL; - unsigned int sig = 0; - - error = -EINVAL; - if (cfg.size < (10 * bts_cfg.sizeof_bts)) - goto errout; + return -EFAULT; - if (cfg.flags & PTRACE_BTS_O_SIGNAL) { - if (!cfg.signal) - goto errout; + if (child->bts) { + ds_release_bts(child->bts); + child->bts = NULL; + } - error = -EOPNOTSUPP; - goto errout; + if (cfg.flags & PTRACE_BTS_O_SIGNAL) { + if (!cfg.signal) + return -EINVAL; - sig = cfg.signal; - } + return -EOPNOTSUPP; - if (child->bts) { - (void)ds_release_bts(child->bts); - kfree(child->bts_buffer); + child->thread.bts_ovfl_signal = cfg.signal; + } - child->bts = NULL; - child->bts_buffer = NULL; - } + if ((cfg.flags & PTRACE_BTS_O_ALLOC) && + (cfg.size != child->bts_size)) { + kfree(child->bts_buffer); - error = -ENOMEM; + child->bts_size = cfg.size; child->bts_buffer = kzalloc(cfg.size, GFP_KERNEL); - if (!child->bts_buffer) - goto errout; - - child->bts = ds_request_bts(child, child->bts_buffer, cfg.size, - ovfl, /* th = */ (size_t)-1); - if (IS_ERR(child->bts)) { - error = PTR_ERR(child->bts); - kfree(child->bts_buffer); - child->bts = NULL; - child->bts_buffer = NULL; - goto errout; + if (!child->bts_buffer) { + child->bts_size = 0; + return -ENOMEM; } - - child->thread.bts_ovfl_signal = sig; } - error = -EINVAL; - if (!child->thread.ds_ctx && cfg.flags) - goto errout; - if (cfg.flags & PTRACE_BTS_O_TRACE) - child->thread.debugctlmsr |= bts_cfg.debugctl_mask; - else - child->thread.debugctlmsr &= ~bts_cfg.debugctl_mask; + flags |= BTS_USER; if (cfg.flags & PTRACE_BTS_O_SCHED) - set_tsk_thread_flag(child, TIF_BTS_TRACE_TS); - else - clear_tsk_thread_flag(child, TIF_BTS_TRACE_TS); + flags |= BTS_TIMESTAMPS; - error = sizeof(cfg); + child->bts = ds_request_bts(child, child->bts_buffer, child->bts_size, + /* ovfl = */ NULL, /* th = */ (size_t)-1, + flags); + if (IS_ERR(child->bts)) { + int error = PTR_ERR(child->bts); -out: - if (child->thread.debugctlmsr) - set_tsk_thread_flag(child, TIF_DEBUGCTLMSR); - else - clear_tsk_thread_flag(child, TIF_DEBUGCTLMSR); + kfree(child->bts_buffer); + child->bts = NULL; + child->bts_buffer = NULL; + child->bts_size = 0; - return error; + return error; + } -errout: - child->thread.debugctlmsr &= ~bts_cfg.debugctl_mask; - clear_tsk_thread_flag(child, TIF_BTS_TRACE_TS); - goto out; + return sizeof(cfg); } static int ptrace_bts_status(struct task_struct *child, long cfg_size, struct ptrace_bts_config __user *ucfg) { + const struct bts_trace *trace; struct ptrace_bts_config cfg; - size_t end; - const void *base, *max; - int error; if (cfg_size < sizeof(cfg)) return -EIO; - error = ds_get_bts_end(child->bts, &end); - if (error < 0) - return error; - - error = ds_access_bts(child->bts, /* index = */ 0, &base); - if (error < 0) - return error; - - error = ds_access_bts(child->bts, /* index = */ end, &max); - if (error < 0) - return error; + trace = ds_read_bts(child->bts); + if (!trace) + return -EPERM; memset(&cfg, 0, sizeof(cfg)); - cfg.size = (max - base); + cfg.size = trace->ds.end - trace->ds.begin; cfg.signal = child->thread.bts_ovfl_signal; cfg.bts_size = sizeof(struct bts_struct); if (cfg.signal) cfg.flags |= PTRACE_BTS_O_SIGNAL; - if (test_tsk_thread_flag(child, TIF_DEBUGCTLMSR) && - child->thread.debugctlmsr & bts_cfg.debugctl_mask) + if (trace->ds.flags & BTS_USER) cfg.flags |= PTRACE_BTS_O_TRACE; - if (test_tsk_thread_flag(child, TIF_BTS_TRACE_TS)) + if (trace->ds.flags & BTS_TIMESTAMPS) cfg.flags |= PTRACE_BTS_O_SCHED; if (copy_to_user(ucfg, &cfg, sizeof(cfg))) @@ -873,105 +746,28 @@ static int ptrace_bts_status(struct task_struct *child, return sizeof(cfg); } -static int ptrace_bts_write_record(struct task_struct *child, - const struct bts_struct *in) +static int ptrace_bts_clear(struct task_struct *child) { - unsigned char bts_record[BTS_MAX_RECORD_SIZE]; + const struct bts_trace *trace; - if (BTS_MAX_RECORD_SIZE < bts_cfg.sizeof_bts) - return -EOVERFLOW; + trace = ds_read_bts(child->bts); + if (!trace) + return -EPERM; - memset(bts_record, 0, bts_cfg.sizeof_bts); - switch (in->qualifier) { - case BTS_INVALID: - break; + memset(trace->ds.begin, 0, trace->ds.n * trace->ds.size); - case BTS_BRANCH: - bts_set(bts_record, bts_from, in->variant.lbr.from_ip); - bts_set(bts_record, bts_to, in->variant.lbr.to_ip); - break; - - case BTS_TASK_ARRIVES: - case BTS_TASK_DEPARTS: - bts_set(bts_record, bts_from, bts_escape); - bts_set(bts_record, bts_qual, in->qualifier); - bts_set(bts_record, bts_jiffies, in->variant.jiffies); - break; - - default: - return -EINVAL; - } - - return ds_write_bts(child->bts, bts_record, bts_cfg.sizeof_bts); + return ds_reset_bts(child->bts); } -void ptrace_bts_take_timestamp(struct task_struct *tsk, - enum bts_qualifier qualifier) +static int ptrace_bts_size(struct task_struct *child) { - struct bts_struct rec = { - .qualifier = qualifier, - .variant.jiffies = jiffies_64 - }; - - ptrace_bts_write_record(tsk, &rec); -} - -static const struct bts_configuration bts_cfg_netburst = { - .sizeof_bts = sizeof(long) * 3, - .sizeof_field = sizeof(long), - .debugctl_mask = (1<<2)|(1<<3)|(1<<5) -}; + const struct bts_trace *trace; -static const struct bts_configuration bts_cfg_pentium_m = { - .sizeof_bts = sizeof(long) * 3, - .sizeof_field = sizeof(long), - .debugctl_mask = (1<<6)|(1<<7) -}; + trace = ds_read_bts(child->bts); + if (!trace) + return -EPERM; -static const struct bts_configuration bts_cfg_core2 = { - .sizeof_bts = 8 * 3, - .sizeof_field = 8, - .debugctl_mask = (1<<6)|(1<<7)|(1<<9) -}; - -static inline void bts_configure(const struct bts_configuration *cfg) -{ - bts_cfg = *cfg; -} - -void __cpuinit ptrace_bts_init_intel(struct cpuinfo_x86 *c) -{ - switch (c->x86) { - case 0x6: - switch (c->x86_model) { - case 0 ... 0xC: - /* sorry, don't know about them */ - break; - case 0xD: - case 0xE: /* Pentium M */ - bts_configure(&bts_cfg_pentium_m); - break; - default: /* Core2, Atom, ... */ - bts_configure(&bts_cfg_core2); - break; - } - break; - case 0xF: - switch (c->x86_model) { - case 0x0: - case 0x1: - case 0x2: /* Netburst */ - bts_configure(&bts_cfg_netburst); - break; - default: - /* sorry, don't know about them */ - break; - } - break; - default: - /* sorry, don't know about them */ - break; - } + return (trace->ds.top - trace->ds.begin) / trace->ds.size; } #endif /* CONFIG_X86_PTRACE_BTS */ @@ -988,15 +784,12 @@ void ptrace_disable(struct task_struct *child) #endif #ifdef CONFIG_X86_PTRACE_BTS if (child->bts) { - (void)ds_release_bts(child->bts); + ds_release_bts(child->bts); + child->bts = NULL; + kfree(child->bts_buffer); child->bts_buffer = NULL; - - child->thread.debugctlmsr &= ~bts_cfg.debugctl_mask; - if (!child->thread.debugctlmsr) - clear_tsk_thread_flag(child, TIF_DEBUGCTLMSR); - - clear_tsk_thread_flag(child, TIF_BTS_TRACE_TS); + child->bts_size = 0; } #endif /* CONFIG_X86_PTRACE_BTS */ } @@ -1129,16 +922,9 @@ long arch_ptrace(struct task_struct *child, long request, long addr, long data) (child, data, (struct ptrace_bts_config __user *)addr); break; - case PTRACE_BTS_SIZE: { - size_t size; - - ret = ds_get_bts_index(child->bts, &size); - if (ret == 0) { - WARN_ON_ONCE(size != (int) size); - ret = (int) size; - } + case PTRACE_BTS_SIZE: + ret = ptrace_bts_size(child); break; - } case PTRACE_BTS_GET: ret = ptrace_bts_read_record @@ -1146,7 +932,7 @@ long arch_ptrace(struct task_struct *child, long request, long addr, long data) break; case PTRACE_BTS_CLEAR: - ret = ds_clear_bts(child->bts); + ret = ptrace_bts_clear(child); break; case PTRACE_BTS_DRAIN: @@ -1409,6 +1195,14 @@ long compat_arch_ptrace(struct task_struct *child, compat_long_t request, case PTRACE_GET_THREAD_AREA: case PTRACE_SET_THREAD_AREA: +#ifdef CONFIG_X86_PTRACE_BTS + case PTRACE_BTS_CONFIG: + case PTRACE_BTS_STATUS: + case PTRACE_BTS_SIZE: + case PTRACE_BTS_GET: + case PTRACE_BTS_CLEAR: + case PTRACE_BTS_DRAIN: +#endif /* CONFIG_X86_PTRACE_BTS */ return arch_ptrace(child, request, addr, data); default: diff --git a/include/linux/sched.h b/include/linux/sched.h index 4b81fc5f7731..dc5ea65dc716 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1176,6 +1176,7 @@ struct task_struct { * The buffer to hold the BTS data. */ void *bts_buffer; + size_t bts_size; #endif /* CONFIG_X86_PTRACE_BTS */ /* PID/PID hash table linkage. */ -- cgit v1.2.3 From ffc2238af8431d930d2c15f16feecf1fd6d75642 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 12 Dec 2008 08:21:19 +0100 Subject: x86, bts: fix build error Impact: build fix arch/x86/kernel/ds.c: In function 'ds_request': arch/x86/kernel/ds.c:236: sorry, unimplemented: inlining failed in call to 'ds_get_context': recursive inlining but the recursion here is scary ... Signed-off-by: Ingo Molnar --- arch/x86/kernel/ds.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/ds.c b/arch/x86/kernel/ds.c index f0583005b75e..dc1e7123ea4e 100644 --- a/arch/x86/kernel/ds.c +++ b/arch/x86/kernel/ds.c @@ -232,7 +232,7 @@ static DEFINE_PER_CPU(struct ds_context *, system_context_array); #define system_context per_cpu(system_context_array, smp_processor_id()) -static inline struct ds_context *ds_get_context(struct task_struct *task) +static struct ds_context *ds_get_context(struct task_struct *task) { struct ds_context **p_context = (task ? &task->thread.ds_ctx : &system_context); -- cgit v1.2.3 From 8808500f26a61757cb414da76b271bbd09d5958c Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 12 Dec 2008 09:20:12 +0100 Subject: x86: soften multi-BAR mapping sanity check warning message Impact: make debug warning less scary The ioremap() time multi-BAR map warning has been causing false positives: http://lkml.org/lkml/2008/12/10/432 http://lkml.org/lkml/2008/12/11/136 So make it less scary by making it once-per-boot, by making it KERN_INFO and by adding this text: "Info: mapping multiple BARs. Your kernel is fine." Signed-off-by: Ingo Molnar --- arch/x86/mm/ioremap.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/mm/ioremap.c b/arch/x86/mm/ioremap.c index d4c4307ff3e0..bd85d42819e1 100644 --- a/arch/x86/mm/ioremap.c +++ b/arch/x86/mm/ioremap.c @@ -223,7 +223,8 @@ static void __iomem *__ioremap_caller(resource_size_t phys_addr, * Check if the request spans more than any BAR in the iomem resource * tree. */ - WARN_ON(iomem_map_sanity_check(phys_addr, size)); + WARN_ONCE(iomem_map_sanity_check(phys_addr, size), + KERN_INFO "Info: mapping multiple BARs. Your kernel is fine."); /* * Don't allow anybody to remap normal RAM that we're using.. -- cgit v1.2.3 From 85072bd55219231b8ca5d9d3fa3492eb4fa6635f Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 12 Dec 2008 11:08:42 +0100 Subject: x86, debug: remove EBDA debug printk Remove leftover EBDA debug message. Reported-by: Andrew Morton Signed-off-by: Ingo Molnar --- arch/x86/kernel/head.c | 1 - 1 file changed, 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/head.c b/arch/x86/kernel/head.c index 1dcb0f13897e..3e66bd364a9d 100644 --- a/arch/x86/kernel/head.c +++ b/arch/x86/kernel/head.c @@ -35,7 +35,6 @@ void __init reserve_ebda_region(void) /* start of EBDA area */ ebda_addr = get_bios_ebda(); - printk(KERN_INFO "BIOS EBDA/lowmem at: %08x/%08x\n", ebda_addr, lowmem); /* Fixup: bios puts an EBDA in the top 64K segment */ /* of conventional memory, but does not adjust lowmem. */ -- cgit v1.2.3 From a0343e823184070f55364d8359f832dcb33c57c7 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Tue, 9 Dec 2008 23:53:16 +0100 Subject: tracing/function-graph-tracer: add a new .irqentry.text section Impact: let the function-graph-tracer be aware of the irq entrypoints Add a new .irqentry.text section to store the irq entrypoints functions inside the same section. This way, the tracer will be able to signal an interrupts triggering on output by recognizing these entrypoints. Also, make this section recordable for dynamic tracing. Signed-off-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- arch/x86/kernel/vmlinux_64.lds.S | 1 + include/asm-generic/vmlinux.lds.h | 10 ++++++++++ scripts/recordmcount.pl | 1 + 3 files changed, 12 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/vmlinux_64.lds.S b/arch/x86/kernel/vmlinux_64.lds.S index 46e05447405b..1a614c0e6bef 100644 --- a/arch/x86/kernel/vmlinux_64.lds.S +++ b/arch/x86/kernel/vmlinux_64.lds.S @@ -35,6 +35,7 @@ SECTIONS SCHED_TEXT LOCK_TEXT KPROBES_TEXT + IRQENTRY_TEXT *(.fixup) *(.gnu.warning) _etext = .; /* End of text section */ diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h index eba835a2c2cd..c61fab1dd2f8 100644 --- a/include/asm-generic/vmlinux.lds.h +++ b/include/asm-generic/vmlinux.lds.h @@ -288,6 +288,16 @@ *(.kprobes.text) \ VMLINUX_SYMBOL(__kprobes_text_end) = .; +#ifdef CONFIG_FUNCTION_GRAPH_TRACER +#define IRQENTRY_TEXT \ + ALIGN_FUNCTION(); \ + VMLINUX_SYMBOL(__irqentry_text_start) = .; \ + *(.irqentry.text) \ + VMLINUX_SYMBOL(__irqentry_text_end) = .; +#else +#define IRQENTRY_TEXT +#endif + /* Section used for early init (in .S files) */ #define HEAD_TEXT *(.head.text) diff --git a/scripts/recordmcount.pl b/scripts/recordmcount.pl index 0b1dc9f9bb06..fe831412bea9 100755 --- a/scripts/recordmcount.pl +++ b/scripts/recordmcount.pl @@ -114,6 +114,7 @@ my %text_sections = ( ".text" => 1, ".sched.text" => 1, ".spinlock.text" => 1, + ".irqentry.text" => 1, ); $objdump = "objdump" if ((length $objdump) == 0); -- cgit v1.2.3 From bcbc4f20b52c2c40c43a4d2337707dcdfe81bc3a Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Tue, 9 Dec 2008 23:54:20 +0100 Subject: tracing/function-graph-tracer: annotate do_IRQ and smp_apic_timer_interrupt Impact: move most important x86 irq entry-points to a separate subsection Annotate do_IRQ and smp_apic_timer_interrupt to put them into the .irqentry.text subsection. These function will so be recognized as hardirq entrypoints for the function-graph-tracer. We could also annotate other irq entries but the others are far less important but they can be added on request. Signed-off-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic.c | 3 ++- arch/x86/kernel/irq_64.c | 3 ++- include/linux/ftrace.h | 11 +++++++++++ 3 files changed, 15 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/apic.c b/arch/x86/kernel/apic.c index 16f94879b525..b946ac19753b 100644 --- a/arch/x86/kernel/apic.c +++ b/arch/x86/kernel/apic.c @@ -30,6 +30,7 @@ #include #include #include +#include #include #include @@ -800,7 +801,7 @@ static void local_apic_timer_interrupt(void) * [ if a single-CPU system runs an SMP kernel then we call the local * interrupt as well. Thus we cannot inline the local irq ... ] */ -void smp_apic_timer_interrupt(struct pt_regs *regs) +void __irq_entry smp_apic_timer_interrupt(struct pt_regs *regs) { struct pt_regs *old_regs = set_irq_regs(regs); diff --git a/arch/x86/kernel/irq_64.c b/arch/x86/kernel/irq_64.c index 60eb84eb77a0..11c65e811ffe 100644 --- a/arch/x86/kernel/irq_64.c +++ b/arch/x86/kernel/irq_64.c @@ -13,6 +13,7 @@ #include #include #include +#include #include #include #include @@ -47,7 +48,7 @@ static inline void stack_overflow_check(struct pt_regs *regs) * SMP cross-CPU interrupts have their own specific * handlers). */ -asmlinkage unsigned int do_IRQ(struct pt_regs *regs) +asmlinkage unsigned int __irq_entry do_IRQ(struct pt_regs *regs) { struct pt_regs *old_regs = set_irq_regs(regs); struct irq_desc *desc; diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 11cac81eed08..44020f31bd81 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -377,6 +377,16 @@ struct ftrace_graph_ret { */ #define __notrace_funcgraph notrace +/* + * We want to which function is an entrypoint of a hardirq. + * That will help us to put a signal on output. + */ +#define __irq_entry __attribute__((__section__(".irqentry.text"))) + +/* Limits of hardirq entrypoints */ +extern char __irqentry_text_start[]; +extern char __irqentry_text_end[]; + #define FTRACE_RETFUNC_DEPTH 50 #define FTRACE_RETSTACK_ALLOC_SIZE 32 /* Type of the callback handlers for tracing function graph*/ @@ -414,6 +424,7 @@ static inline void unpause_graph_tracing(void) #else #define __notrace_funcgraph +#define __irq_entry static inline void ftrace_graph_init_task(struct task_struct *t) { } static inline void ftrace_graph_exit_task(struct task_struct *t) { } -- cgit v1.2.3 From 16855f878d7127a8bb3925753463485f3071ad76 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Mon, 8 Dec 2008 19:18:38 -0800 Subject: x86: uaccess: return value of __{get|put}_user() can be int Impact: cleanup The type of return value of __{get|put}_user() can be int. There is no user to refer the return value of __{get|put}_user() as long. This reduces code size a bit on 64-bit. $ size vmlinux.* text data bss dec hex filename 4509265 479988 673588 5662841 566879 vmlinux.new 4511462 479988 673588 5665038 56710e vmlinux.old Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/include/asm/uaccess.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/uaccess.h b/arch/x86/include/asm/uaccess.h index 35c54921b2e4..580c3ee6c58c 100644 --- a/arch/x86/include/asm/uaccess.h +++ b/arch/x86/include/asm/uaccess.h @@ -350,14 +350,14 @@ do { \ #define __put_user_nocheck(x, ptr, size) \ ({ \ - long __pu_err; \ + int __pu_err; \ __put_user_size((x), (ptr), (size), __pu_err, -EFAULT); \ __pu_err; \ }) #define __get_user_nocheck(x, ptr, size) \ ({ \ - long __gu_err; \ + int __gu_err; \ unsigned long __gu_val; \ __get_user_size(__gu_val, (ptr), (size), __gu_err, -EFAULT); \ (x) = (__force __typeof__(*(ptr)))__gu_val; \ -- cgit v1.2.3 From 8f2466f45f75e3cbe3aa2b69d33fd9d6e343b9cc Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Mon, 8 Dec 2008 19:19:07 -0800 Subject: x86: kill #ifdef for exit_idle() Impact: cleanup Introduce helper inline function in arch/x86/include/asm/idle.h to remove #ifdefs around exit_idle(). Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/include/asm/idle.h | 5 +++++ arch/x86/kernel/apic.c | 6 ------ arch/x86/kernel/io_apic.c | 3 +-- 3 files changed, 6 insertions(+), 8 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/idle.h b/arch/x86/include/asm/idle.h index 44c89c3a23e9..38d87379e270 100644 --- a/arch/x86/include/asm/idle.h +++ b/arch/x86/include/asm/idle.h @@ -8,8 +8,13 @@ struct notifier_block; void idle_notifier_register(struct notifier_block *n); void idle_notifier_unregister(struct notifier_block *n); +#ifdef CONFIG_X86_64 void enter_idle(void); void exit_idle(void); +#else /* !CONFIG_X86_64 */ +static inline void enter_idle(void) { } +static inline void exit_idle(void) { } +#endif /* CONFIG_X86_64 */ void c1e_remove_cpu(int cpu); diff --git a/arch/x86/kernel/apic.c b/arch/x86/kernel/apic.c index 16f94879b525..0fd083713f62 100644 --- a/arch/x86/kernel/apic.c +++ b/arch/x86/kernel/apic.c @@ -814,9 +814,7 @@ void smp_apic_timer_interrupt(struct pt_regs *regs) * Besides, if we don't timer interrupts ignore the global * interrupt lock, which is the WrongThing (tm) to do. */ -#ifdef CONFIG_X86_64 exit_idle(); -#endif irq_enter(); local_apic_timer_interrupt(); irq_exit(); @@ -1682,9 +1680,7 @@ void smp_spurious_interrupt(struct pt_regs *regs) { u32 v; -#ifdef CONFIG_X86_64 exit_idle(); -#endif irq_enter(); /* * Check if this really is a spurious interrupt and ACK it @@ -1713,9 +1709,7 @@ void smp_error_interrupt(struct pt_regs *regs) { u32 v, v1; -#ifdef CONFIG_X86_64 exit_idle(); -#endif irq_enter(); /* First tickle the hardware, only then report what went on. -- REW */ v = apic_read(APIC_ESR); diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index 9043251210fb..679e7bbbbcd6 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -2216,10 +2216,9 @@ static void set_ir_ioapic_affinity_irq(unsigned int irq, cpumask_t mask) asmlinkage void smp_irq_move_cleanup_interrupt(void) { unsigned vector, me; + ack_APIC_irq(); -#ifdef CONFIG_X86_64 exit_idle(); -#endif irq_enter(); me = smp_processor_id(); -- cgit v1.2.3 From 915b0d0104b72fd36af088ba4b11b5690bc96a6c Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Mon, 8 Dec 2008 19:19:26 -0800 Subject: x86: hardirq: introduce inc_irq_stat() Impact: cleanup Introduce inc_irq_stat() macro and unify irq_stat accounting code. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/include/asm/hardirq_32.h | 2 ++ arch/x86/include/asm/hardirq_64.h | 2 ++ arch/x86/kernel/apic.c | 13 +++---------- arch/x86/kernel/smp.c | 18 +++--------------- arch/x86/kernel/traps.c | 6 +----- 5 files changed, 11 insertions(+), 30 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/hardirq_32.h b/arch/x86/include/asm/hardirq_32.h index 5ca135e72f2b..cf7954d1405f 100644 --- a/arch/x86/include/asm/hardirq_32.h +++ b/arch/x86/include/asm/hardirq_32.h @@ -22,6 +22,8 @@ DECLARE_PER_CPU(irq_cpustat_t, irq_stat); #define __ARCH_IRQ_STAT #define __IRQ_STAT(cpu, member) (per_cpu(irq_stat, cpu).member) +#define inc_irq_stat(member) (__get_cpu_var(irq_stat).member++) + void ack_bad_irq(unsigned int irq); #include diff --git a/arch/x86/include/asm/hardirq_64.h b/arch/x86/include/asm/hardirq_64.h index 1ba381fc51d3..b5a6b5d56704 100644 --- a/arch/x86/include/asm/hardirq_64.h +++ b/arch/x86/include/asm/hardirq_64.h @@ -11,6 +11,8 @@ #define __ARCH_IRQ_STAT 1 +#define inc_irq_stat(member) add_pda(member, 1) + #define local_softirq_pending() read_pda(__softirq_pending) #define __ARCH_SET_SOFTIRQ_PENDING 1 diff --git a/arch/x86/kernel/apic.c b/arch/x86/kernel/apic.c index 16f94879b525..1771dd746811 100644 --- a/arch/x86/kernel/apic.c +++ b/arch/x86/kernel/apic.c @@ -783,11 +783,7 @@ static void local_apic_timer_interrupt(void) /* * the NMI deadlock-detector uses this. */ -#ifdef CONFIG_X86_64 - add_pda(apic_timer_irqs, 1); -#else - per_cpu(irq_stat, cpu).apic_timer_irqs++; -#endif + inc_irq_stat(apic_timer_irqs); evt->event_handler(evt); } @@ -1695,14 +1691,11 @@ void smp_spurious_interrupt(struct pt_regs *regs) if (v & (1 << (SPURIOUS_APIC_VECTOR & 0x1f))) ack_APIC_irq(); -#ifdef CONFIG_X86_64 - add_pda(irq_spurious_count, 1); -#else + inc_irq_stat(irq_spurious_count); + /* see sw-dev-man vol 3, chapter 7.4.13.5 */ printk(KERN_INFO "spurious APIC interrupt on CPU#%d, " "should never happen.\n", smp_processor_id()); - __get_cpu_var(irq_stat).irq_spurious_count++; -#endif irq_exit(); } diff --git a/arch/x86/kernel/smp.c b/arch/x86/kernel/smp.c index 18f9b19f5f8f..d18537ce2c79 100644 --- a/arch/x86/kernel/smp.c +++ b/arch/x86/kernel/smp.c @@ -178,11 +178,7 @@ static void native_smp_send_stop(void) void smp_reschedule_interrupt(struct pt_regs *regs) { ack_APIC_irq(); -#ifdef CONFIG_X86_32 - __get_cpu_var(irq_stat).irq_resched_count++; -#else - add_pda(irq_resched_count, 1); -#endif + inc_irq_stat(irq_resched_count); } void smp_call_function_interrupt(struct pt_regs *regs) @@ -190,11 +186,7 @@ void smp_call_function_interrupt(struct pt_regs *regs) ack_APIC_irq(); irq_enter(); generic_smp_call_function_interrupt(); -#ifdef CONFIG_X86_32 - __get_cpu_var(irq_stat).irq_call_count++; -#else - add_pda(irq_call_count, 1); -#endif + inc_irq_stat(irq_call_count); irq_exit(); } @@ -203,11 +195,7 @@ void smp_call_function_single_interrupt(struct pt_regs *regs) ack_APIC_irq(); irq_enter(); generic_smp_call_function_single_interrupt(); -#ifdef CONFIG_X86_32 - __get_cpu_var(irq_stat).irq_call_count++; -#else - add_pda(irq_call_count, 1); -#endif + inc_irq_stat(irq_call_count); irq_exit(); } diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c index 04d242ab0161..d815293e6d94 100644 --- a/arch/x86/kernel/traps.c +++ b/arch/x86/kernel/traps.c @@ -481,11 +481,7 @@ do_nmi(struct pt_regs *regs, long error_code) { nmi_enter(); -#ifdef CONFIG_X86_32 - { int cpu; cpu = smp_processor_id(); ++nmi_count(cpu); } -#else - add_pda(__nmi_count, 1); -#endif + inc_irq_stat(__nmi_count); if (!ignore_nmis) default_do_nmi(regs); -- cgit v1.2.3 From 2bed8446819a7c5033aa1da138d9f230ae212edc Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 12 Dec 2008 12:13:36 +0100 Subject: tracing/function-graph-tracer: add a new .irqentry.text section, fix Impact: build fix 32-bit x86 needs this section too. Signed-off-by: Ingo Molnar --- arch/x86/kernel/vmlinux_32.lds.S | 1 + 1 file changed, 1 insertion(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/vmlinux_32.lds.S b/arch/x86/kernel/vmlinux_32.lds.S index a9b8560adbc2..82c67559dde7 100644 --- a/arch/x86/kernel/vmlinux_32.lds.S +++ b/arch/x86/kernel/vmlinux_32.lds.S @@ -44,6 +44,7 @@ SECTIONS SCHED_TEXT LOCK_TEXT KPROBES_TEXT + IRQENTRY_TEXT *(.fixup) *(.gnu.warning) _etext = .; /* End of text section */ -- cgit v1.2.3 From 9470565579f29486f4ed0ffa50774268b64994b0 Mon Sep 17 00:00:00 2001 From: Dave Jones Date: Mon, 1 Dec 2008 14:13:50 -0800 Subject: x86: remove init_mm export as planned for 2.6.26 Impact: remove deprecated export Signed-off-by: Dave Jones Signed-off-by: Andrew Morton Signed-off-by: Thomas Gleixner --- Documentation/feature-removal-schedule.txt | 12 ------------ arch/x86/kernel/init_task.c | 1 - 2 files changed, 13 deletions(-) (limited to 'arch/x86') diff --git a/Documentation/feature-removal-schedule.txt b/Documentation/feature-removal-schedule.txt index c28a2ac88f9d..1a8af7354e79 100644 --- a/Documentation/feature-removal-schedule.txt +++ b/Documentation/feature-removal-schedule.txt @@ -244,18 +244,6 @@ Who: Michael Buesch --------------------------- -What: init_mm export -When: 2.6.26 -Why: Not used in-tree. The current out-of-tree users used it to - work around problems in the CPA code which should be resolved - by now. One usecase was described to provide verification code - of the CPA operation. That's a good idea in general, but such - code / infrastructure should be in the kernel and not in some - out-of-tree driver. -Who: Thomas Gleixner - ----------------------------- - What: usedac i386 kernel parameter When: 2.6.27 Why: replaced by allowdac and no dac combination diff --git a/arch/x86/kernel/init_task.c b/arch/x86/kernel/init_task.c index a4f93b4120c1..d39918076bb4 100644 --- a/arch/x86/kernel/init_task.c +++ b/arch/x86/kernel/init_task.c @@ -14,7 +14,6 @@ static struct fs_struct init_fs = INIT_FS; static struct signal_struct init_signals = INIT_SIGNALS(init_signals); static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand); struct mm_struct init_mm = INIT_MM(init_mm); -EXPORT_UNUSED_SYMBOL(init_mm); /* will be removed in 2.6.26 */ /* * Initial thread structure. -- cgit v1.2.3 From fd28a5b58dddf5cb5df162ae5c8797a63171c31d Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Tue, 21 Oct 2008 14:05:00 +0200 Subject: x86: remove simnow earlyprintk support Impact: remove obsolete code The later versions of SimNow! actually all have serial console emulation, so the direct interface isn't needed anymore. So remove the undocumented simnow earlyprintk console. Signed-off-by: Andi Kleen Signed-off-by: Thomas Gleixner --- arch/x86/kernel/early_printk.c | 47 ------------------------------------------ 1 file changed, 47 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/early_printk.c b/arch/x86/kernel/early_printk.c index 34ad997d3834..23b138e31e9c 100644 --- a/arch/x86/kernel/early_printk.c +++ b/arch/x86/kernel/early_printk.c @@ -875,49 +875,6 @@ static struct console early_dbgp_console = { }; #endif -/* Console interface to a host file on AMD's SimNow! */ - -static int simnow_fd; - -enum { - MAGIC1 = 0xBACCD00A, - MAGIC2 = 0xCA110000, - XOPEN = 5, - XWRITE = 4, -}; - -static noinline long simnow(long cmd, long a, long b, long c) -{ - long ret; - - asm volatile("cpuid" : - "=a" (ret) : - "b" (a), "c" (b), "d" (c), "0" (MAGIC1), "D" (cmd + MAGIC2)); - return ret; -} - -static void __init simnow_init(char *str) -{ - char *fn = "klog"; - - if (*str == '=') - fn = ++str; - /* error ignored */ - simnow_fd = simnow(XOPEN, (unsigned long)fn, O_WRONLY|O_APPEND|O_CREAT, 0644); -} - -static void simnow_write(struct console *con, const char *s, unsigned n) -{ - simnow(XWRITE, simnow_fd, (unsigned long)s, n); -} - -static struct console simnow_console = { - .name = "simnow", - .write = simnow_write, - .flags = CON_PRINTBUFFER, - .index = -1, -}; - /* Direct interface for emergencies */ static struct console *early_console = &early_vga_console; static int __initdata early_console_initialized; @@ -960,10 +917,6 @@ static int __init setup_early_printk(char *buf) max_ypos = boot_params.screen_info.orig_video_lines; current_ypos = boot_params.screen_info.orig_y; early_console = &early_vga_console; - } else if (!strncmp(buf, "simnow", 6)) { - simnow_init(buf + 6); - early_console = &simnow_console; - keep_early = 1; #ifdef CONFIG_EARLY_PRINTK_DBGP } else if (!strncmp(buf, "dbgp", 4)) { if (early_dbgp_init(buf+4) < 0) -- cgit v1.2.3 From 98a79d6a50181ca1ecf7400eda01d5dc1bc0dbf0 Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Sat, 13 Dec 2008 21:19:41 +1030 Subject: cpumask: centralize cpu_online_map and cpu_possible_map Impact: cleanup Each SMP arch defines these themselves. Move them to a central location. Twists: 1) Some archs (m32, parisc, s390) set possible_map to all 1, so we add a CONFIG_INIT_ALL_POSSIBLE for this rather than break them. 2) mips and sparc32 '#define cpu_possible_map phys_cpu_present_map'. Those archs simply have phys_cpu_present_map replaced everywhere. 3) Alpha defined cpu_possible_map to cpu_present_map; this is tricky so I just manipulate them both in sync. 4) IA64, cris and m32r have gratuitous 'extern cpumask_t cpu_possible_map' declarations. Signed-off-by: Rusty Russell Reviewed-by: Grant Grundler Tested-by: Tony Luck Acked-by: Ingo Molnar Cc: Mike Travis Cc: ink@jurassic.park.msu.ru Cc: rmk@arm.linux.org.uk Cc: starvik@axis.com Cc: tony.luck@intel.com Cc: takata@linux-m32r.org Cc: ralf@linux-mips.org Cc: grundler@parisc-linux.org Cc: paulus@samba.org Cc: schwidefsky@de.ibm.com Cc: lethal@linux-sh.org Cc: wli@holomorphy.com Cc: davem@davemloft.net Cc: jdike@addtoit.com Cc: mingo@redhat.com --- arch/alpha/include/asm/smp.h | 1 - arch/alpha/kernel/process.c | 2 ++ arch/alpha/kernel/smp.c | 7 ++----- arch/arm/kernel/smp.c | 10 ---------- arch/cris/arch-v32/kernel/smp.c | 4 ---- arch/cris/include/asm/smp.h | 1 - arch/ia64/include/asm/smp.h | 1 - arch/ia64/kernel/smpboot.c | 6 ------ arch/m32r/Kconfig | 1 + arch/m32r/kernel/smpboot.c | 6 ------ arch/mips/include/asm/smp.h | 3 --- arch/mips/kernel/smp-cmp.c | 2 +- arch/mips/kernel/smp-mt.c | 2 +- arch/mips/kernel/smp.c | 7 +------ arch/mips/kernel/smtc.c | 6 +++--- arch/mips/pmc-sierra/yosemite/smp.c | 6 +++--- arch/mips/sgi-ip27/ip27-smp.c | 2 +- arch/mips/sibyte/bcm1480/smp.c | 8 ++++---- arch/mips/sibyte/sb1250/smp.c | 8 ++++---- arch/parisc/Kconfig | 1 + arch/parisc/kernel/smp.c | 15 --------------- arch/powerpc/kernel/smp.c | 4 ---- arch/s390/Kconfig | 1 + arch/s390/kernel/smp.c | 6 ------ arch/sh/kernel/smp.c | 6 ------ arch/sparc/include/asm/smp_32.h | 2 -- arch/sparc/kernel/smp.c | 6 ++---- arch/sparc/kernel/sparc_ksyms.c | 4 ---- arch/sparc64/kernel/smp.c | 4 ---- arch/um/kernel/smp.c | 7 ------- arch/x86/kernel/smpboot.c | 6 ------ arch/x86/mach-voyager/voyager_smp.c | 7 ------- include/asm-m32r/smp.h | 2 -- init/Kconfig | 9 +++++++++ kernel/cpu.c | 11 ++++++----- 35 files changed, 42 insertions(+), 132 deletions(-) (limited to 'arch/x86') diff --git a/arch/alpha/include/asm/smp.h b/arch/alpha/include/asm/smp.h index 544c69af8168..547e90951cec 100644 --- a/arch/alpha/include/asm/smp.h +++ b/arch/alpha/include/asm/smp.h @@ -45,7 +45,6 @@ extern struct cpuinfo_alpha cpu_data[NR_CPUS]; #define raw_smp_processor_id() (current_thread_info()->cpu) extern int smp_num_cpus; -#define cpu_possible_map cpu_present_map extern void arch_send_call_function_single_ipi(int cpu); extern void arch_send_call_function_ipi(cpumask_t mask); diff --git a/arch/alpha/kernel/process.c b/arch/alpha/kernel/process.c index 351407e07e71..f238370c907d 100644 --- a/arch/alpha/kernel/process.c +++ b/arch/alpha/kernel/process.c @@ -94,6 +94,7 @@ common_shutdown_1(void *generic_ptr) flags |= 0x00040000UL; /* "remain halted" */ *pflags = flags; cpu_clear(cpuid, cpu_present_map); + cpu_clear(cpuid, cpu_possible_map); halt(); } #endif @@ -120,6 +121,7 @@ common_shutdown_1(void *generic_ptr) #ifdef CONFIG_SMP /* Wait for the secondaries to halt. */ cpu_clear(boot_cpuid, cpu_present_map); + cpu_clear(boot_cpuid, cpu_possible_map); while (cpus_weight(cpu_present_map)) barrier(); #endif diff --git a/arch/alpha/kernel/smp.c b/arch/alpha/kernel/smp.c index cf7da10097bb..d953e510f68d 100644 --- a/arch/alpha/kernel/smp.c +++ b/arch/alpha/kernel/smp.c @@ -70,11 +70,6 @@ enum ipi_message_type { /* Set to a secondary's cpuid when it comes online. */ static int smp_secondary_alive __devinitdata = 0; -/* Which cpus ids came online. */ -cpumask_t cpu_online_map; - -EXPORT_SYMBOL(cpu_online_map); - int smp_num_probed; /* Internal processor count */ int smp_num_cpus = 1; /* Number that came online. */ EXPORT_SYMBOL(smp_num_cpus); @@ -440,6 +435,7 @@ setup_smp(void) ((char *)cpubase + i*hwrpb->processor_size); if ((cpu->flags & 0x1cc) == 0x1cc) { smp_num_probed++; + cpu_set(i, cpu_possible_map); cpu_set(i, cpu_present_map); cpu->pal_revision = boot_cpu_palrev; } @@ -473,6 +469,7 @@ smp_prepare_cpus(unsigned int max_cpus) /* Nothing to do on a UP box, or when told not to. */ if (smp_num_probed == 1 || max_cpus == 0) { + cpu_possible_map = cpumask_of_cpu(boot_cpuid); cpu_present_map = cpumask_of_cpu(boot_cpuid); printk(KERN_INFO "SMP mode deactivated.\n"); return; diff --git a/arch/arm/kernel/smp.c b/arch/arm/kernel/smp.c index e42a749a56dd..bd905c0a7365 100644 --- a/arch/arm/kernel/smp.c +++ b/arch/arm/kernel/smp.c @@ -33,16 +33,6 @@ #include #include -/* - * bitmask of present and online CPUs. - * The present bitmask indicates that the CPU is physically present. - * The online bitmask indicates that the CPU is up and running. - */ -cpumask_t cpu_possible_map; -EXPORT_SYMBOL(cpu_possible_map); -cpumask_t cpu_online_map; -EXPORT_SYMBOL(cpu_online_map); - /* * as from 2.5, kernels no longer have an init_tasks structure * so we need some other way of telling a new secondary core diff --git a/arch/cris/arch-v32/kernel/smp.c b/arch/cris/arch-v32/kernel/smp.c index 52e16c6436f9..9dac17334640 100644 --- a/arch/cris/arch-v32/kernel/smp.c +++ b/arch/cris/arch-v32/kernel/smp.c @@ -29,11 +29,7 @@ spinlock_t cris_atomic_locks[] = { [0 ... LOCK_COUNT - 1] = SPIN_LOCK_UNLOCKED}; /* CPU masks */ -cpumask_t cpu_online_map = CPU_MASK_NONE; -EXPORT_SYMBOL(cpu_online_map); cpumask_t phys_cpu_present_map = CPU_MASK_NONE; -cpumask_t cpu_possible_map; -EXPORT_SYMBOL(cpu_possible_map); EXPORT_SYMBOL(phys_cpu_present_map); /* Variables used during SMP boot */ diff --git a/arch/cris/include/asm/smp.h b/arch/cris/include/asm/smp.h index dba33aba3e95..c615a06dd757 100644 --- a/arch/cris/include/asm/smp.h +++ b/arch/cris/include/asm/smp.h @@ -4,7 +4,6 @@ #include extern cpumask_t phys_cpu_present_map; -extern cpumask_t cpu_possible_map; #define raw_smp_processor_id() (current_thread_info()->cpu) diff --git a/arch/ia64/include/asm/smp.h b/arch/ia64/include/asm/smp.h index 12d96e0cd513..21c402365d0e 100644 --- a/arch/ia64/include/asm/smp.h +++ b/arch/ia64/include/asm/smp.h @@ -57,7 +57,6 @@ extern struct smp_boot_data { extern char no_int_routing __devinitdata; -extern cpumask_t cpu_online_map; extern cpumask_t cpu_core_map[NR_CPUS]; DECLARE_PER_CPU(cpumask_t, cpu_sibling_map); extern int smp_num_siblings; diff --git a/arch/ia64/kernel/smpboot.c b/arch/ia64/kernel/smpboot.c index 1dcbb85fc4ee..4ede6e571c38 100644 --- a/arch/ia64/kernel/smpboot.c +++ b/arch/ia64/kernel/smpboot.c @@ -131,12 +131,6 @@ struct task_struct *task_for_booting_cpu; */ DEFINE_PER_CPU(int, cpu_state); -/* Bitmasks of currently online, and possible CPUs */ -cpumask_t cpu_online_map; -EXPORT_SYMBOL(cpu_online_map); -cpumask_t cpu_possible_map = CPU_MASK_NONE; -EXPORT_SYMBOL(cpu_possible_map); - cpumask_t cpu_core_map[NR_CPUS] __cacheline_aligned; EXPORT_SYMBOL(cpu_core_map); DEFINE_PER_CPU_SHARED_ALIGNED(cpumask_t, cpu_sibling_map); diff --git a/arch/m32r/Kconfig b/arch/m32r/Kconfig index dbaed4a63815..17a6dab09319 100644 --- a/arch/m32r/Kconfig +++ b/arch/m32r/Kconfig @@ -10,6 +10,7 @@ config M32R default y select HAVE_IDE select HAVE_OPROFILE + select INIT_ALL_POSSIBLE config SBUS bool diff --git a/arch/m32r/kernel/smpboot.c b/arch/m32r/kernel/smpboot.c index 39cb6da72dcb..0f06b3722e96 100644 --- a/arch/m32r/kernel/smpboot.c +++ b/arch/m32r/kernel/smpboot.c @@ -73,17 +73,11 @@ static unsigned int bsp_phys_id = -1; /* Bitmask of physically existing CPUs */ physid_mask_t phys_cpu_present_map; -/* Bitmask of currently online CPUs */ -cpumask_t cpu_online_map; -EXPORT_SYMBOL(cpu_online_map); - cpumask_t cpu_bootout_map; cpumask_t cpu_bootin_map; static cpumask_t cpu_callin_map; cpumask_t cpu_callout_map; EXPORT_SYMBOL(cpu_callout_map); -cpumask_t cpu_possible_map = CPU_MASK_ALL; -EXPORT_SYMBOL(cpu_possible_map); /* Per CPU bogomips and other parameters */ struct cpuinfo_m32r cpu_data[NR_CPUS] __cacheline_aligned; diff --git a/arch/mips/include/asm/smp.h b/arch/mips/include/asm/smp.h index 0ff5b523ea77..86557b5d1b3f 100644 --- a/arch/mips/include/asm/smp.h +++ b/arch/mips/include/asm/smp.h @@ -38,9 +38,6 @@ extern int __cpu_logical_map[NR_CPUS]; #define SMP_RESCHEDULE_YOURSELF 0x1 /* XXX braindead */ #define SMP_CALL_FUNCTION 0x2 -extern cpumask_t phys_cpu_present_map; -#define cpu_possible_map phys_cpu_present_map - extern void asmlinkage smp_bootstrap(void); /* diff --git a/arch/mips/kernel/smp-cmp.c b/arch/mips/kernel/smp-cmp.c index ca476c4f62a5..6789c1a12120 100644 --- a/arch/mips/kernel/smp-cmp.c +++ b/arch/mips/kernel/smp-cmp.c @@ -226,7 +226,7 @@ void __init cmp_smp_setup(void) for (i = 1; i < NR_CPUS; i++) { if (amon_cpu_avail(i)) { - cpu_set(i, phys_cpu_present_map); + cpu_set(i, cpu_possible_map); __cpu_number_map[i] = ++ncpu; __cpu_logical_map[ncpu] = i; } diff --git a/arch/mips/kernel/smp-mt.c b/arch/mips/kernel/smp-mt.c index 87a1816c1f45..6f7ee5ac46ee 100644 --- a/arch/mips/kernel/smp-mt.c +++ b/arch/mips/kernel/smp-mt.c @@ -70,7 +70,7 @@ static unsigned int __init smvp_vpe_init(unsigned int tc, unsigned int mvpconf0, write_vpe_c0_vpeconf0(tmp); /* Record this as available CPU */ - cpu_set(tc, phys_cpu_present_map); + cpu_set(tc, cpu_possible_map); __cpu_number_map[tc] = ++ncpu; __cpu_logical_map[ncpu] = tc; } diff --git a/arch/mips/kernel/smp.c b/arch/mips/kernel/smp.c index 8bf88faf5afd..3da94704f816 100644 --- a/arch/mips/kernel/smp.c +++ b/arch/mips/kernel/smp.c @@ -44,15 +44,10 @@ #include #endif /* CONFIG_MIPS_MT_SMTC */ -cpumask_t phys_cpu_present_map; /* Bitmask of available CPUs */ volatile cpumask_t cpu_callin_map; /* Bitmask of started secondaries */ -cpumask_t cpu_online_map; /* Bitmask of currently online CPUs */ int __cpu_number_map[NR_CPUS]; /* Map physical to logical */ int __cpu_logical_map[NR_CPUS]; /* Map logical to physical */ -EXPORT_SYMBOL(phys_cpu_present_map); -EXPORT_SYMBOL(cpu_online_map); - extern void cpu_idle(void); /* Number of TCs (or siblings in Intel speak) per CPU core */ @@ -195,7 +190,7 @@ void __init smp_prepare_cpus(unsigned int max_cpus) /* preload SMP state for boot cpu */ void __devinit smp_prepare_boot_cpu(void) { - cpu_set(0, phys_cpu_present_map); + cpu_set(0, cpu_possible_map); cpu_set(0, cpu_online_map); cpu_set(0, cpu_callin_map); } diff --git a/arch/mips/kernel/smtc.c b/arch/mips/kernel/smtc.c index 897fb2b4751c..b6cca01ff82b 100644 --- a/arch/mips/kernel/smtc.c +++ b/arch/mips/kernel/smtc.c @@ -290,7 +290,7 @@ static void smtc_configure_tlb(void) * possibly leave some TCs/VPEs as "slave" processors. * * Use c0_MVPConf0 to find out how many TCs are available, setting up - * phys_cpu_present_map and the logical/physical mappings. + * cpu_possible_map and the logical/physical mappings. */ int __init smtc_build_cpu_map(int start_cpu_slot) @@ -304,7 +304,7 @@ int __init smtc_build_cpu_map(int start_cpu_slot) */ ntcs = ((read_c0_mvpconf0() & MVPCONF0_PTC) >> MVPCONF0_PTC_SHIFT) + 1; for (i=start_cpu_slot; i> MVPCONF0_PTC_SHIFT) + 1)) { - cpu_clear(tc, phys_cpu_present_map); + cpu_clear(tc, cpu_possible_map); cpu_clear(tc, cpu_present_map); tc++; } diff --git a/arch/mips/pmc-sierra/yosemite/smp.c b/arch/mips/pmc-sierra/yosemite/smp.c index 3a7df647ca77..f78c29b68d77 100644 --- a/arch/mips/pmc-sierra/yosemite/smp.c +++ b/arch/mips/pmc-sierra/yosemite/smp.c @@ -141,7 +141,7 @@ static void __cpuinit yos_boot_secondary(int cpu, struct task_struct *idle) } /* - * Detect available CPUs, populate phys_cpu_present_map before smp_init + * Detect available CPUs, populate cpu_possible_map before smp_init * * We don't want to start the secondary CPU yet nor do we have a nice probing * feature in PMON so we just assume presence of the secondary core. @@ -150,10 +150,10 @@ static void __init yos_smp_setup(void) { int i; - cpus_clear(phys_cpu_present_map); + cpus_clear(cpu_possible_map); for (i = 0; i < 2; i++) { - cpu_set(i, phys_cpu_present_map); + cpu_set(i, cpu_possible_map); __cpu_number_map[i] = i; __cpu_logical_map[i] = i; } diff --git a/arch/mips/sgi-ip27/ip27-smp.c b/arch/mips/sgi-ip27/ip27-smp.c index ba5cdebeaf0d..5b47d6b65275 100644 --- a/arch/mips/sgi-ip27/ip27-smp.c +++ b/arch/mips/sgi-ip27/ip27-smp.c @@ -76,7 +76,7 @@ static int do_cpumask(cnodeid_t cnode, nasid_t nasid, int highest) /* Only let it join in if it's marked enabled */ if ((acpu->cpu_info.flags & KLINFO_ENABLE) && (tot_cpus_found != NR_CPUS)) { - cpu_set(cpuid, phys_cpu_present_map); + cpu_set(cpuid, cpu_possible_map); alloc_cpupda(cpuid, tot_cpus_found); cpus_found++; tot_cpus_found++; diff --git a/arch/mips/sibyte/bcm1480/smp.c b/arch/mips/sibyte/bcm1480/smp.c index bd9eeb43ed0e..dddfda8e8294 100644 --- a/arch/mips/sibyte/bcm1480/smp.c +++ b/arch/mips/sibyte/bcm1480/smp.c @@ -136,7 +136,7 @@ static void __cpuinit bcm1480_boot_secondary(int cpu, struct task_struct *idle) /* * Use CFE to find out how many CPUs are available, setting up - * phys_cpu_present_map and the logical/physical mappings. + * cpu_possible_map and the logical/physical mappings. * XXXKW will the boot CPU ever not be physical 0? * * Common setup before any secondaries are started @@ -145,14 +145,14 @@ static void __init bcm1480_smp_setup(void) { int i, num; - cpus_clear(phys_cpu_present_map); - cpu_set(0, phys_cpu_present_map); + cpus_clear(cpu_possible_map); + cpu_set(0, cpu_possible_map); __cpu_number_map[0] = 0; __cpu_logical_map[0] = 0; for (i = 1, num = 0; i < NR_CPUS; i++) { if (cfe_cpu_stop(i) == 0) { - cpu_set(i, phys_cpu_present_map); + cpu_set(i, cpu_possible_map); __cpu_number_map[i] = ++num; __cpu_logical_map[num] = i; } diff --git a/arch/mips/sibyte/sb1250/smp.c b/arch/mips/sibyte/sb1250/smp.c index 0734b933e969..5950a288a7da 100644 --- a/arch/mips/sibyte/sb1250/smp.c +++ b/arch/mips/sibyte/sb1250/smp.c @@ -124,7 +124,7 @@ static void __cpuinit sb1250_boot_secondary(int cpu, struct task_struct *idle) /* * Use CFE to find out how many CPUs are available, setting up - * phys_cpu_present_map and the logical/physical mappings. + * cpu_possible_map and the logical/physical mappings. * XXXKW will the boot CPU ever not be physical 0? * * Common setup before any secondaries are started @@ -133,14 +133,14 @@ static void __init sb1250_smp_setup(void) { int i, num; - cpus_clear(phys_cpu_present_map); - cpu_set(0, phys_cpu_present_map); + cpus_clear(cpu_possible_map); + cpu_set(0, cpu_possible_map); __cpu_number_map[0] = 0; __cpu_logical_map[0] = 0; for (i = 1, num = 0; i < NR_CPUS; i++) { if (cfe_cpu_stop(i) == 0) { - cpu_set(i, phys_cpu_present_map); + cpu_set(i, cpu_possible_map); __cpu_number_map[i] = ++num; __cpu_logical_map[num] = i; } diff --git a/arch/parisc/Kconfig b/arch/parisc/Kconfig index 644a70b1b04e..aacf11d33723 100644 --- a/arch/parisc/Kconfig +++ b/arch/parisc/Kconfig @@ -11,6 +11,7 @@ config PARISC select HAVE_OPROFILE select RTC_CLASS select RTC_DRV_PARISC + select INIT_ALL_POSSIBLE help The PA-RISC microprocessor is designed by Hewlett-Packard and used in many of their workstations & servers (HP9000 700 and 800 series, diff --git a/arch/parisc/kernel/smp.c b/arch/parisc/kernel/smp.c index d47f3975c9c6..80bc000523fa 100644 --- a/arch/parisc/kernel/smp.c +++ b/arch/parisc/kernel/smp.c @@ -67,21 +67,6 @@ static volatile int cpu_now_booting __read_mostly = 0; /* track which CPU is boo static int parisc_max_cpus __read_mostly = 1; -/* online cpus are ones that we've managed to bring up completely - * possible cpus are all valid cpu - * present cpus are all detected cpu - * - * On startup we bring up the "possible" cpus. Since we discover - * CPUs later, we add them as hotplug, so the possible cpu mask is - * empty in the beginning. - */ - -cpumask_t cpu_online_map __read_mostly = CPU_MASK_NONE; /* Bitmap of online CPUs */ -cpumask_t cpu_possible_map __read_mostly = CPU_MASK_ALL; /* Bitmap of Present CPUs */ - -EXPORT_SYMBOL(cpu_online_map); -EXPORT_SYMBOL(cpu_possible_map); - DEFINE_PER_CPU(spinlock_t, ipi_lock) = SPIN_LOCK_UNLOCKED; enum ipi_message_type { diff --git a/arch/powerpc/kernel/smp.c b/arch/powerpc/kernel/smp.c index ff9f7010097d..d1165566f064 100644 --- a/arch/powerpc/kernel/smp.c +++ b/arch/powerpc/kernel/smp.c @@ -60,13 +60,9 @@ int smp_hw_index[NR_CPUS]; struct thread_info *secondary_ti; -cpumask_t cpu_possible_map = CPU_MASK_NONE; -cpumask_t cpu_online_map = CPU_MASK_NONE; DEFINE_PER_CPU(cpumask_t, cpu_sibling_map) = CPU_MASK_NONE; DEFINE_PER_CPU(cpumask_t, cpu_core_map) = CPU_MASK_NONE; -EXPORT_SYMBOL(cpu_online_map); -EXPORT_SYMBOL(cpu_possible_map); EXPORT_PER_CPU_SYMBOL(cpu_sibling_map); EXPORT_PER_CPU_SYMBOL(cpu_core_map); diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig index 8116a3328a19..b4aa5869c7f9 100644 --- a/arch/s390/Kconfig +++ b/arch/s390/Kconfig @@ -75,6 +75,7 @@ config S390 select HAVE_KRETPROBES select HAVE_KVM if 64BIT select HAVE_ARCH_TRACEHOOK + select INIT_ALL_POSSIBLE source "init/Kconfig" diff --git a/arch/s390/kernel/smp.c b/arch/s390/kernel/smp.c index b5595688a477..f03914b8ed2f 100644 --- a/arch/s390/kernel/smp.c +++ b/arch/s390/kernel/smp.c @@ -52,12 +52,6 @@ struct _lowcore *lowcore_ptr[NR_CPUS]; EXPORT_SYMBOL(lowcore_ptr); -cpumask_t cpu_online_map = CPU_MASK_NONE; -EXPORT_SYMBOL(cpu_online_map); - -cpumask_t cpu_possible_map = CPU_MASK_ALL; -EXPORT_SYMBOL(cpu_possible_map); - static struct task_struct *current_set[NR_CPUS]; static u8 smp_cpu_type; diff --git a/arch/sh/kernel/smp.c b/arch/sh/kernel/smp.c index 3c5ad1660bbc..593937d0c495 100644 --- a/arch/sh/kernel/smp.c +++ b/arch/sh/kernel/smp.c @@ -31,12 +31,6 @@ int __cpu_number_map[NR_CPUS]; /* Map physical to logical */ int __cpu_logical_map[NR_CPUS]; /* Map logical to physical */ -cpumask_t cpu_possible_map; -EXPORT_SYMBOL(cpu_possible_map); - -cpumask_t cpu_online_map; -EXPORT_SYMBOL(cpu_online_map); - static inline void __init smp_store_cpu_info(unsigned int cpu) { struct sh_cpuinfo *c = cpu_data + cpu; diff --git a/arch/sparc/include/asm/smp_32.h b/arch/sparc/include/asm/smp_32.h index a8180e546a48..8408d9d2a662 100644 --- a/arch/sparc/include/asm/smp_32.h +++ b/arch/sparc/include/asm/smp_32.h @@ -29,8 +29,6 @@ */ extern unsigned char boot_cpu_id; -extern cpumask_t phys_cpu_present_map; -#define cpu_possible_map phys_cpu_present_map typedef void (*smpfunc_t)(unsigned long, unsigned long, unsigned long, unsigned long, unsigned long); diff --git a/arch/sparc/kernel/smp.c b/arch/sparc/kernel/smp.c index e396c1f17a92..1e5ac4e282e1 100644 --- a/arch/sparc/kernel/smp.c +++ b/arch/sparc/kernel/smp.c @@ -39,8 +39,6 @@ volatile unsigned long cpu_callin_map[NR_CPUS] __cpuinitdata = {0,}; unsigned char boot_cpu_id = 0; unsigned char boot_cpu_id4 = 0; /* boot_cpu_id << 2 */ -cpumask_t cpu_online_map = CPU_MASK_NONE; -cpumask_t phys_cpu_present_map = CPU_MASK_NONE; cpumask_t smp_commenced_mask = CPU_MASK_NONE; /* The only guaranteed locking primitive available on all Sparc @@ -334,7 +332,7 @@ void __init smp_setup_cpu_possible_map(void) instance = 0; while (!cpu_find_by_instance(instance, NULL, &mid)) { if (mid < NR_CPUS) { - cpu_set(mid, phys_cpu_present_map); + cpu_set(mid, cpu_possible_map); cpu_set(mid, cpu_present_map); } instance++; @@ -354,7 +352,7 @@ void __init smp_prepare_boot_cpu(void) current_thread_info()->cpu = cpuid; cpu_set(cpuid, cpu_online_map); - cpu_set(cpuid, phys_cpu_present_map); + cpu_set(cpuid, cpu_possible_map); } int __cpuinit __cpu_up(unsigned int cpu) diff --git a/arch/sparc/kernel/sparc_ksyms.c b/arch/sparc/kernel/sparc_ksyms.c index b0dfff848653..32d11a5fe3a8 100644 --- a/arch/sparc/kernel/sparc_ksyms.c +++ b/arch/sparc/kernel/sparc_ksyms.c @@ -113,10 +113,6 @@ EXPORT_PER_CPU_SYMBOL(__cpu_data); #ifdef CONFIG_SMP /* IRQ implementation. */ EXPORT_SYMBOL(synchronize_irq); - -/* CPU online map and active count. */ -EXPORT_SYMBOL(cpu_online_map); -EXPORT_SYMBOL(phys_cpu_present_map); #endif EXPORT_SYMBOL(__udelay); diff --git a/arch/sparc64/kernel/smp.c b/arch/sparc64/kernel/smp.c index f500b0618bb0..a97b8822c22c 100644 --- a/arch/sparc64/kernel/smp.c +++ b/arch/sparc64/kernel/smp.c @@ -49,14 +49,10 @@ int sparc64_multi_core __read_mostly; -cpumask_t cpu_possible_map __read_mostly = CPU_MASK_NONE; -cpumask_t cpu_online_map __read_mostly = CPU_MASK_NONE; DEFINE_PER_CPU(cpumask_t, cpu_sibling_map) = CPU_MASK_NONE; cpumask_t cpu_core_map[NR_CPUS] __read_mostly = { [0 ... NR_CPUS-1] = CPU_MASK_NONE }; -EXPORT_SYMBOL(cpu_possible_map); -EXPORT_SYMBOL(cpu_online_map); EXPORT_PER_CPU_SYMBOL(cpu_sibling_map); EXPORT_SYMBOL(cpu_core_map); diff --git a/arch/um/kernel/smp.c b/arch/um/kernel/smp.c index 045772142844..98351c78bc81 100644 --- a/arch/um/kernel/smp.c +++ b/arch/um/kernel/smp.c @@ -25,13 +25,6 @@ DEFINE_PER_CPU(struct mmu_gather, mmu_gathers); #include "irq_user.h" #include "os.h" -/* CPU online map, set by smp_boot_cpus */ -cpumask_t cpu_online_map = CPU_MASK_NONE; -cpumask_t cpu_possible_map = CPU_MASK_NONE; - -EXPORT_SYMBOL(cpu_online_map); -EXPORT_SYMBOL(cpu_possible_map); - /* Per CPU bogomips and other parameters * The only piece used here is the ipi pipe, which is set before SMP is * started and never changed. diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index 7b1093397319..468c2f9d47ae 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -101,14 +101,8 @@ EXPORT_SYMBOL(smp_num_siblings); /* Last level cache ID of each logical CPU */ DEFINE_PER_CPU(u16, cpu_llc_id) = BAD_APICID; -/* bitmap of online cpus */ -cpumask_t cpu_online_map __read_mostly; -EXPORT_SYMBOL(cpu_online_map); - cpumask_t cpu_callin_map; cpumask_t cpu_callout_map; -cpumask_t cpu_possible_map; -EXPORT_SYMBOL(cpu_possible_map); /* representing HT siblings of each logical CPU */ DEFINE_PER_CPU(cpumask_t, cpu_sibling_map); diff --git a/arch/x86/mach-voyager/voyager_smp.c b/arch/x86/mach-voyager/voyager_smp.c index 52145007bd7e..9c990185e9f2 100644 --- a/arch/x86/mach-voyager/voyager_smp.c +++ b/arch/x86/mach-voyager/voyager_smp.c @@ -63,11 +63,6 @@ static int voyager_extended_cpus = 1; /* Used for the invalidate map that's also checked in the spinlock */ static volatile unsigned long smp_invalidate_needed; -/* Bitmask of currently online CPUs - used by setup.c for - /proc/cpuinfo, visible externally but still physical */ -cpumask_t cpu_online_map = CPU_MASK_NONE; -EXPORT_SYMBOL(cpu_online_map); - /* Bitmask of CPUs present in the system - exported by i386_syms.c, used * by scheduler but indexed physically */ cpumask_t phys_cpu_present_map = CPU_MASK_NONE; @@ -218,8 +213,6 @@ static cpumask_t smp_commenced_mask = CPU_MASK_NONE; /* This is for the new dynamic CPU boot code */ cpumask_t cpu_callin_map = CPU_MASK_NONE; cpumask_t cpu_callout_map = CPU_MASK_NONE; -cpumask_t cpu_possible_map = CPU_MASK_NONE; -EXPORT_SYMBOL(cpu_possible_map); /* The per processor IRQ masks (these are usually kept in sync) */ static __u16 vic_irq_mask[NR_CPUS] __cacheline_aligned; diff --git a/include/asm-m32r/smp.h b/include/asm-m32r/smp.h index c5dd66916692..b96a6d2ffbc3 100644 --- a/include/asm-m32r/smp.h +++ b/include/asm-m32r/smp.h @@ -63,8 +63,6 @@ extern volatile int cpu_2_physid[NR_CPUS]; #define raw_smp_processor_id() (current_thread_info()->cpu) extern cpumask_t cpu_callout_map; -extern cpumask_t cpu_possible_map; -extern cpumask_t cpu_present_map; static __inline__ int hard_smp_processor_id(void) { diff --git a/init/Kconfig b/init/Kconfig index f763762d544a..7656623f5006 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -916,6 +916,15 @@ config KMOD endif # MODULES +config INIT_ALL_POSSIBLE + bool + help + Back when each arch used to define their own cpu_online_map and + cpu_possible_map, some of them chose to initialize cpu_possible_map + with all 1s, and others with all 0s. When they were centralised, + it was better to provide this option than to break all the archs + and have several arch maintainers persuing me down dark alleys. + config STOP_MACHINE bool default y diff --git a/kernel/cpu.c b/kernel/cpu.c index 8ea32e8d68b0..bae131a1211b 100644 --- a/kernel/cpu.c +++ b/kernel/cpu.c @@ -24,19 +24,20 @@ cpumask_t cpu_present_map __read_mostly; EXPORT_SYMBOL(cpu_present_map); -#ifndef CONFIG_SMP - /* * Represents all cpu's that are currently online. */ -cpumask_t cpu_online_map __read_mostly = CPU_MASK_ALL; +cpumask_t cpu_online_map __read_mostly; EXPORT_SYMBOL(cpu_online_map); +#ifdef CONFIG_INIT_ALL_POSSIBLE cpumask_t cpu_possible_map __read_mostly = CPU_MASK_ALL; +#else +cpumask_t cpu_possible_map __read_mostly; +#endif EXPORT_SYMBOL(cpu_possible_map); -#else /* CONFIG_SMP */ - +#ifdef CONFIG_SMP /* Serializes the updates to cpu_online_map, cpu_present_map */ static DEFINE_MUTEX(cpu_add_remove_lock); -- cgit v1.2.3 From 29c0177e6a4ac094302bed54a1d4bbb6b740a9ef Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Sat, 13 Dec 2008 21:20:25 +1030 Subject: cpumask: change cpumask_scnprintf, cpumask_parse_user, cpulist_parse, and cpulist_scnprintf to take pointers. Impact: change calling convention of existing cpumask APIs Most cpumask functions started with cpus_: these have been replaced by cpumask_ ones which take struct cpumask pointers as expected. These four functions don't have good replacement names; fortunately they're rarely used, so we just change them over. Signed-off-by: Rusty Russell Signed-off-by: Mike Travis Acked-by: Ingo Molnar Cc: paulus@samba.org Cc: mingo@redhat.com Cc: tony.luck@intel.com Cc: ralf@linux-mips.org Cc: Greg Kroah-Hartman Cc: cl@linux-foundation.org Cc: srostedt@redhat.com --- arch/ia64/kernel/topology.c | 2 +- arch/mips/kernel/smp-cmp.c | 4 +- arch/powerpc/platforms/pseries/xics.c | 2 +- arch/x86/kernel/cpu/intel_cacheinfo.c | 4 +- arch/x86/kernel/setup_percpu.c | 2 +- drivers/base/cpu.c | 2 +- drivers/base/node.c | 4 +- drivers/base/topology.c | 4 +- drivers/pci/pci-sysfs.c | 4 +- drivers/pci/probe.c | 4 +- include/linux/cpumask.h | 87 +++++++++++++++++++++++------------ kernel/cpuset.c | 4 +- kernel/irq/proc.c | 4 +- kernel/profile.c | 4 +- kernel/sched.c | 4 +- kernel/sched_stats.h | 2 +- kernel/taskstats.c | 2 +- kernel/trace/trace.c | 4 +- mm/slub.c | 2 +- 19 files changed, 86 insertions(+), 59 deletions(-) (limited to 'arch/x86') diff --git a/arch/ia64/kernel/topology.c b/arch/ia64/kernel/topology.c index c75b914f2d6b..a8d61a3e9a94 100644 --- a/arch/ia64/kernel/topology.c +++ b/arch/ia64/kernel/topology.c @@ -219,7 +219,7 @@ static ssize_t show_shared_cpu_map(struct cache_info *this_leaf, char *buf) cpumask_t shared_cpu_map; cpus_and(shared_cpu_map, this_leaf->shared_cpu_map, cpu_online_map); - len = cpumask_scnprintf(buf, NR_CPUS+1, shared_cpu_map); + len = cpumask_scnprintf(buf, NR_CPUS+1, &shared_cpu_map); len += sprintf(buf+len, "\n"); return len; } diff --git a/arch/mips/kernel/smp-cmp.c b/arch/mips/kernel/smp-cmp.c index 6789c1a12120..f27beca4b26d 100644 --- a/arch/mips/kernel/smp-cmp.c +++ b/arch/mips/kernel/smp-cmp.c @@ -51,10 +51,10 @@ static int __init allowcpus(char *str) int len; cpus_clear(cpu_allow_map); - if (cpulist_parse(str, cpu_allow_map) == 0) { + if (cpulist_parse(str, &cpu_allow_map) == 0) { cpu_set(0, cpu_allow_map); cpus_and(cpu_possible_map, cpu_possible_map, cpu_allow_map); - len = cpulist_scnprintf(buf, sizeof(buf)-1, cpu_possible_map); + len = cpulist_scnprintf(buf, sizeof(buf)-1, &cpu_possible_map); buf[len] = '\0'; pr_debug("Allowable CPUs: %s\n", buf); return 1; diff --git a/arch/powerpc/platforms/pseries/xics.c b/arch/powerpc/platforms/pseries/xics.c index e1904774a70f..64d24310ce7e 100644 --- a/arch/powerpc/platforms/pseries/xics.c +++ b/arch/powerpc/platforms/pseries/xics.c @@ -358,7 +358,7 @@ static void xics_set_affinity(unsigned int virq, cpumask_t cpumask) irq_server = get_irq_server(virq, 1); if (irq_server == -1) { char cpulist[128]; - cpumask_scnprintf(cpulist, sizeof(cpulist), cpumask); + cpumask_scnprintf(cpulist, sizeof(cpulist), &cpumask); printk(KERN_WARNING "%s: No online cpus in the mask %s for irq %d\n", __func__, cpulist, virq); diff --git a/arch/x86/kernel/cpu/intel_cacheinfo.c b/arch/x86/kernel/cpu/intel_cacheinfo.c index 3f46afbb1cf1..43ea612d3e9d 100644 --- a/arch/x86/kernel/cpu/intel_cacheinfo.c +++ b/arch/x86/kernel/cpu/intel_cacheinfo.c @@ -626,8 +626,8 @@ static ssize_t show_shared_cpu_map_func(struct _cpuid4_info *this_leaf, cpumask_t *mask = &this_leaf->shared_cpu_map; n = type? - cpulist_scnprintf(buf, len-2, *mask): - cpumask_scnprintf(buf, len-2, *mask); + cpulist_scnprintf(buf, len-2, mask) : + cpumask_scnprintf(buf, len-2, mask); buf[n++] = '\n'; buf[n] = '\0'; } diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c index ae0c0d3bb770..1c2084291f97 100644 --- a/arch/x86/kernel/setup_percpu.c +++ b/arch/x86/kernel/setup_percpu.c @@ -282,7 +282,7 @@ static void __cpuinit numa_set_cpumask(int cpu, int enable) else cpu_clear(cpu, *mask); - cpulist_scnprintf(buf, sizeof(buf), *mask); + cpulist_scnprintf(buf, sizeof(buf), mask); printk(KERN_DEBUG "%s cpu %d node %d: mask now %s\n", enable? "numa_add_cpu":"numa_remove_cpu", cpu, node, buf); } diff --git a/drivers/base/cpu.c b/drivers/base/cpu.c index 64f5d54f7edc..4259072f5bd0 100644 --- a/drivers/base/cpu.c +++ b/drivers/base/cpu.c @@ -109,7 +109,7 @@ static SYSDEV_ATTR(crash_notes, 0400, show_crash_notes, NULL); */ static ssize_t print_cpus_map(char *buf, cpumask_t *map) { - int n = cpulist_scnprintf(buf, PAGE_SIZE-2, *map); + int n = cpulist_scnprintf(buf, PAGE_SIZE-2, map); buf[n++] = '\n'; buf[n] = '\0'; diff --git a/drivers/base/node.c b/drivers/base/node.c index f5207090885a..91636cd8b6c9 100644 --- a/drivers/base/node.c +++ b/drivers/base/node.c @@ -30,8 +30,8 @@ static ssize_t node_read_cpumap(struct sys_device *dev, int type, char *buf) BUILD_BUG_ON((NR_CPUS/32 * 9) > (PAGE_SIZE-1)); len = type? - cpulist_scnprintf(buf, PAGE_SIZE-2, *mask): - cpumask_scnprintf(buf, PAGE_SIZE-2, *mask); + cpulist_scnprintf(buf, PAGE_SIZE-2, mask) : + cpumask_scnprintf(buf, PAGE_SIZE-2, mask); buf[len++] = '\n'; buf[len] = '\0'; return len; diff --git a/drivers/base/topology.c b/drivers/base/topology.c index 199cd97e32e6..a8bc1cbcfa7c 100644 --- a/drivers/base/topology.c +++ b/drivers/base/topology.c @@ -49,8 +49,8 @@ static ssize_t show_cpumap(int type, cpumask_t *mask, char *buf) if (len > 1) { n = type? - cpulist_scnprintf(buf, len-2, *mask): - cpumask_scnprintf(buf, len-2, *mask); + cpulist_scnprintf(buf, len-2, mask) : + cpumask_scnprintf(buf, len-2, mask); buf[n++] = '\n'; buf[n] = '\0'; } diff --git a/drivers/pci/pci-sysfs.c b/drivers/pci/pci-sysfs.c index 5d72866897a8..c88485860a0a 100644 --- a/drivers/pci/pci-sysfs.c +++ b/drivers/pci/pci-sysfs.c @@ -74,7 +74,7 @@ static ssize_t local_cpus_show(struct device *dev, int len; mask = pcibus_to_cpumask(to_pci_dev(dev)->bus); - len = cpumask_scnprintf(buf, PAGE_SIZE-2, mask); + len = cpumask_scnprintf(buf, PAGE_SIZE-2, &mask); buf[len++] = '\n'; buf[len] = '\0'; return len; @@ -88,7 +88,7 @@ static ssize_t local_cpulist_show(struct device *dev, int len; mask = pcibus_to_cpumask(to_pci_dev(dev)->bus); - len = cpulist_scnprintf(buf, PAGE_SIZE-2, mask); + len = cpulist_scnprintf(buf, PAGE_SIZE-2, &mask); buf[len++] = '\n'; buf[len] = '\0'; return len; diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index 003a9b3c293f..5b3f5937ecf5 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -55,8 +55,8 @@ static ssize_t pci_bus_show_cpuaffinity(struct device *dev, cpumask = pcibus_to_cpumask(to_pci_bus(dev)); ret = type? - cpulist_scnprintf(buf, PAGE_SIZE-2, cpumask): - cpumask_scnprintf(buf, PAGE_SIZE-2, cpumask); + cpulist_scnprintf(buf, PAGE_SIZE-2, &cpumask) : + cpumask_scnprintf(buf, PAGE_SIZE-2, &cpumask); buf[ret++] = '\n'; buf[ret] = '\0'; return ret; diff --git a/include/linux/cpumask.h b/include/linux/cpumask.h index 21e1dd43e52a..94a2ab88ae85 100644 --- a/include/linux/cpumask.h +++ b/include/linux/cpumask.h @@ -339,36 +339,6 @@ extern cpumask_t cpu_mask_all; #endif #define CPUMASK_PTR(v, m) cpumask_t *v = &(m->v) -#define cpumask_scnprintf(buf, len, src) \ - __cpumask_scnprintf((buf), (len), &(src), NR_CPUS) -static inline int __cpumask_scnprintf(char *buf, int len, - const cpumask_t *srcp, int nbits) -{ - return bitmap_scnprintf(buf, len, srcp->bits, nbits); -} - -#define cpumask_parse_user(ubuf, ulen, dst) \ - __cpumask_parse_user((ubuf), (ulen), &(dst), NR_CPUS) -static inline int __cpumask_parse_user(const char __user *buf, int len, - cpumask_t *dstp, int nbits) -{ - return bitmap_parse_user(buf, len, dstp->bits, nbits); -} - -#define cpulist_scnprintf(buf, len, src) \ - __cpulist_scnprintf((buf), (len), &(src), NR_CPUS) -static inline int __cpulist_scnprintf(char *buf, int len, - const cpumask_t *srcp, int nbits) -{ - return bitmap_scnlistprintf(buf, len, srcp->bits, nbits); -} - -#define cpulist_parse(buf, dst) __cpulist_parse((buf), &(dst), NR_CPUS) -static inline int __cpulist_parse(const char *buf, cpumask_t *dstp, int nbits) -{ - return bitmap_parselist(buf, dstp->bits, nbits); -} - #define cpu_remap(oldbit, old, new) \ __cpu_remap((oldbit), &(old), &(new), NR_CPUS) static inline int __cpu_remap(int oldbit, @@ -945,6 +915,63 @@ static inline void cpumask_copy(struct cpumask *dstp, */ #define cpumask_of(cpu) (get_cpu_mask(cpu)) +/** + * cpumask_scnprintf - print a cpumask into a string as comma-separated hex + * @buf: the buffer to sprintf into + * @len: the length of the buffer + * @srcp: the cpumask to print + * + * If len is zero, returns zero. Otherwise returns the length of the + * (nul-terminated) @buf string. + */ +static inline int cpumask_scnprintf(char *buf, int len, + const struct cpumask *srcp) +{ + return bitmap_scnprintf(buf, len, srcp->bits, nr_cpumask_bits); +} + +/** + * cpumask_parse_user - extract a cpumask from a user string + * @buf: the buffer to extract from + * @len: the length of the buffer + * @dstp: the cpumask to set. + * + * Returns -errno, or 0 for success. + */ +static inline int cpumask_parse_user(const char __user *buf, int len, + struct cpumask *dstp) +{ + return bitmap_parse_user(buf, len, dstp->bits, nr_cpumask_bits); +} + +/** + * cpulist_scnprintf - print a cpumask into a string as comma-separated list + * @buf: the buffer to sprintf into + * @len: the length of the buffer + * @srcp: the cpumask to print + * + * If len is zero, returns zero. Otherwise returns the length of the + * (nul-terminated) @buf string. + */ +static inline int cpulist_scnprintf(char *buf, int len, + const struct cpumask *srcp) +{ + return bitmap_scnlistprintf(buf, len, srcp->bits, nr_cpumask_bits); +} + +/** + * cpulist_parse_user - extract a cpumask from a user string of ranges + * @buf: the buffer to extract from + * @len: the length of the buffer + * @dstp: the cpumask to set. + * + * Returns -errno, or 0 for success. + */ +static inline int cpulist_parse(const char *buf, struct cpumask *dstp) +{ + return bitmap_parselist(buf, dstp->bits, nr_cpumask_bits); +} + /** * to_cpumask - convert an NR_CPUS bitmap to a struct cpumask * * @bitmap: the bitmap diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 96c0ba13b8cd..39c1a4c1c5a9 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -896,7 +896,7 @@ static int update_cpumask(struct cpuset *cs, const char *buf) if (!*buf) { cpus_clear(trialcs.cpus_allowed); } else { - retval = cpulist_parse(buf, trialcs.cpus_allowed); + retval = cpulist_parse(buf, &trialcs.cpus_allowed); if (retval < 0) return retval; @@ -1482,7 +1482,7 @@ static int cpuset_sprintf_cpulist(char *page, struct cpuset *cs) mask = cs->cpus_allowed; mutex_unlock(&callback_mutex); - return cpulist_scnprintf(page, PAGE_SIZE, mask); + return cpulist_scnprintf(page, PAGE_SIZE, &mask); } static int cpuset_sprintf_memlist(char *page, struct cpuset *cs) diff --git a/kernel/irq/proc.c b/kernel/irq/proc.c index d257e7d6a8a4..f293349d49d0 100644 --- a/kernel/irq/proc.c +++ b/kernel/irq/proc.c @@ -47,7 +47,7 @@ static ssize_t irq_affinity_proc_write(struct file *file, irq_balancing_disabled(irq)) return -EIO; - err = cpumask_parse_user(buffer, count, new_value); + err = cpumask_parse_user(buffer, count, &new_value); if (err) return err; @@ -95,7 +95,7 @@ static ssize_t default_affinity_write(struct file *file, cpumask_t new_value; int err; - err = cpumask_parse_user(buffer, count, new_value); + err = cpumask_parse_user(buffer, count, &new_value); if (err) return err; diff --git a/kernel/profile.c b/kernel/profile.c index dc41827fbfee..7d620dfdde59 100644 --- a/kernel/profile.c +++ b/kernel/profile.c @@ -442,7 +442,7 @@ void profile_tick(int type) static int prof_cpu_mask_read_proc(char *page, char **start, off_t off, int count, int *eof, void *data) { - int len = cpumask_scnprintf(page, count, *(cpumask_t *)data); + int len = cpumask_scnprintf(page, count, (cpumask_t *)data); if (count - len < 2) return -EINVAL; len += sprintf(page + len, "\n"); @@ -456,7 +456,7 @@ static int prof_cpu_mask_write_proc(struct file *file, unsigned long full_count = count, err; cpumask_t new_value; - err = cpumask_parse_user(buffer, count, new_value); + err = cpumask_parse_user(buffer, count, &new_value); if (err) return err; diff --git a/kernel/sched.c b/kernel/sched.c index e4bb1dd7b308..d2d16d1273b1 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -6666,7 +6666,7 @@ static int sched_domain_debug_one(struct sched_domain *sd, int cpu, int level, struct sched_group *group = sd->groups; char str[256]; - cpulist_scnprintf(str, sizeof(str), sd->span); + cpulist_scnprintf(str, sizeof(str), &sd->span); cpus_clear(*groupmask); printk(KERN_DEBUG "%*s domain %d: ", level, "", level); @@ -6720,7 +6720,7 @@ static int sched_domain_debug_one(struct sched_domain *sd, int cpu, int level, cpus_or(*groupmask, *groupmask, group->cpumask); - cpulist_scnprintf(str, sizeof(str), group->cpumask); + cpulist_scnprintf(str, sizeof(str), &group->cpumask); printk(KERN_CONT " %s", str); group = group->next; diff --git a/kernel/sched_stats.h b/kernel/sched_stats.h index 7dbf72a2b02c..6beff1e4eeae 100644 --- a/kernel/sched_stats.h +++ b/kernel/sched_stats.h @@ -42,7 +42,7 @@ static int show_schedstat(struct seq_file *seq, void *v) for_each_domain(cpu, sd) { enum cpu_idle_type itype; - cpumask_scnprintf(mask_str, mask_len, sd->span); + cpumask_scnprintf(mask_str, mask_len, &sd->span); seq_printf(seq, "domain%d %s", dcount++, mask_str); for (itype = CPU_IDLE; itype < CPU_MAX_IDLE_TYPES; itype++) { diff --git a/kernel/taskstats.c b/kernel/taskstats.c index bd6be76303cf..6d7dc4ec4aa5 100644 --- a/kernel/taskstats.c +++ b/kernel/taskstats.c @@ -352,7 +352,7 @@ static int parse(struct nlattr *na, cpumask_t *mask) if (!data) return -ENOMEM; nla_strlcpy(data, na, len); - ret = cpulist_parse(data, *mask); + ret = cpulist_parse(data, mask); kfree(data); return ret; } diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index d86e3252f300..d2e75479dc50 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2126,7 +2126,7 @@ tracing_cpumask_read(struct file *filp, char __user *ubuf, mutex_lock(&tracing_cpumask_update_lock); - len = cpumask_scnprintf(mask_str, count, tracing_cpumask); + len = cpumask_scnprintf(mask_str, count, &tracing_cpumask); if (count - len < 2) { count = -EINVAL; goto out_err; @@ -2147,7 +2147,7 @@ tracing_cpumask_write(struct file *filp, const char __user *ubuf, int err, cpu; mutex_lock(&tracing_cpumask_update_lock); - err = cpumask_parse_user(ubuf, count, tracing_cpumask_new); + err = cpumask_parse_user(ubuf, count, &tracing_cpumask_new); if (err) goto err_unlock; diff --git a/mm/slub.c b/mm/slub.c index a2cd47d89e0a..8e516e29f989 100644 --- a/mm/slub.c +++ b/mm/slub.c @@ -3626,7 +3626,7 @@ static int list_locations(struct kmem_cache *s, char *buf, len < PAGE_SIZE - 60) { len += sprintf(buf + len, " cpus="); len += cpulist_scnprintf(buf + len, PAGE_SIZE - len - 50, - l->cpus); + &l->cpus); } if (num_online_nodes() > 1 && !nodes_empty(l->nodes) && -- cgit v1.2.3 From 0de26520c7cabf36e1de090ea8092f011a6106ce Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Sat, 13 Dec 2008 21:20:26 +1030 Subject: cpumask: make irq_set_affinity() take a const struct cpumask Impact: change existing irq_chip API Not much point with gentle transition here: the struct irq_chip's setaffinity method signature needs to change. Fortunately, not widely used code, but hits a few architectures. Note: In irq_select_affinity() I save a temporary in by mangling irq_desc[irq].affinity directly. Ingo, does this break anything? (Folded in fix from KOSAKI Motohiro) Signed-off-by: Rusty Russell Signed-off-by: Mike Travis Reviewed-by: Grant Grundler Acked-by: Ingo Molnar Cc: ralf@linux-mips.org Cc: grundler@parisc-linux.org Cc: jeremy@xensource.com Cc: KOSAKI Motohiro --- arch/alpha/kernel/irq.c | 2 +- arch/alpha/kernel/sys_dp264.c | 8 ++-- arch/alpha/kernel/sys_titan.c | 4 +- arch/arm/common/gic.c | 4 +- arch/arm/kernel/irq.c | 2 +- arch/arm/oprofile/op_model_mpcore.c | 4 +- arch/cris/arch-v32/kernel/irq.c | 4 +- arch/ia64/hp/sim/hpsim_irq.c | 2 +- arch/ia64/kernel/iosapic.c | 12 +++--- arch/ia64/kernel/irq.c | 9 ++-- arch/ia64/kernel/msi_ia64.c | 12 +++--- arch/ia64/kernel/smpboot.c | 4 +- arch/ia64/sn/kernel/irq.c | 6 +-- arch/ia64/sn/kernel/msi_sn.c | 7 +-- arch/mips/include/asm/irq.h | 3 +- arch/mips/kernel/cevt-bcm1480.c | 2 +- arch/mips/kernel/cevt-sb1250.c | 2 +- arch/mips/kernel/irq-gic.c | 6 +-- arch/mips/mti-malta/malta-smtc.c | 6 +-- arch/mips/sibyte/bcm1480/irq.c | 8 ++-- arch/mips/sibyte/sb1250/irq.c | 8 ++-- arch/parisc/kernel/irq.c | 6 +-- arch/powerpc/kernel/irq.c | 2 +- arch/powerpc/platforms/pseries/xics.c | 6 +-- arch/powerpc/sysdev/mpic.c | 4 +- arch/powerpc/sysdev/mpic.h | 2 +- arch/sparc64/kernel/irq.c | 11 +++-- arch/sparc64/kernel/of_device.c | 2 +- arch/sparc64/kernel/pci_msi.c | 2 +- arch/x86/kernel/hpet.c | 4 +- arch/x86/kernel/io_apic.c | 81 +++++++++++++++++------------------ arch/x86/kernel/irq_32.c | 2 +- arch/x86/kernel/irq_64.c | 2 +- drivers/parisc/iosapic.c | 7 +-- drivers/xen/events.c | 6 +-- include/linux/interrupt.h | 4 +- include/linux/irq.h | 3 +- kernel/irq/chip.c | 2 +- kernel/irq/manage.c | 22 +++++----- kernel/irq/migration.c | 14 +++--- kernel/irq/proc.c | 29 ++++++++----- kernel/time/tick-common.c | 6 +-- 42 files changed, 171 insertions(+), 161 deletions(-) (limited to 'arch/x86') diff --git a/arch/alpha/kernel/irq.c b/arch/alpha/kernel/irq.c index c626a821cdcb..d0f1620007f7 100644 --- a/arch/alpha/kernel/irq.c +++ b/arch/alpha/kernel/irq.c @@ -55,7 +55,7 @@ int irq_select_affinity(unsigned int irq) last_cpu = cpu; irq_desc[irq].affinity = cpumask_of_cpu(cpu); - irq_desc[irq].chip->set_affinity(irq, cpumask_of_cpu(cpu)); + irq_desc[irq].chip->set_affinity(irq, cpumask_of(cpu)); return 0; } #endif /* CONFIG_SMP */ diff --git a/arch/alpha/kernel/sys_dp264.c b/arch/alpha/kernel/sys_dp264.c index c71b0fd7a61f..ab44c164d9d4 100644 --- a/arch/alpha/kernel/sys_dp264.c +++ b/arch/alpha/kernel/sys_dp264.c @@ -177,19 +177,19 @@ cpu_set_irq_affinity(unsigned int irq, cpumask_t affinity) } static void -dp264_set_affinity(unsigned int irq, cpumask_t affinity) +dp264_set_affinity(unsigned int irq, const struct cpumask *affinity) { spin_lock(&dp264_irq_lock); - cpu_set_irq_affinity(irq, affinity); + cpu_set_irq_affinity(irq, *affinity); tsunami_update_irq_hw(cached_irq_mask); spin_unlock(&dp264_irq_lock); } static void -clipper_set_affinity(unsigned int irq, cpumask_t affinity) +clipper_set_affinity(unsigned int irq, const struct cpumask *affinity) { spin_lock(&dp264_irq_lock); - cpu_set_irq_affinity(irq - 16, affinity); + cpu_set_irq_affinity(irq - 16, *affinity); tsunami_update_irq_hw(cached_irq_mask); spin_unlock(&dp264_irq_lock); } diff --git a/arch/alpha/kernel/sys_titan.c b/arch/alpha/kernel/sys_titan.c index 52c91ccc1648..27f840a4ad3d 100644 --- a/arch/alpha/kernel/sys_titan.c +++ b/arch/alpha/kernel/sys_titan.c @@ -158,10 +158,10 @@ titan_cpu_set_irq_affinity(unsigned int irq, cpumask_t affinity) } static void -titan_set_irq_affinity(unsigned int irq, cpumask_t affinity) +titan_set_irq_affinity(unsigned int irq, const struct cpumask *affinity) { spin_lock(&titan_irq_lock); - titan_cpu_set_irq_affinity(irq - 16, affinity); + titan_cpu_set_irq_affinity(irq - 16, *affinity); titan_update_irq_hw(titan_cached_irq_mask); spin_unlock(&titan_irq_lock); } diff --git a/arch/arm/common/gic.c b/arch/arm/common/gic.c index 7fc9860a97d7..c6884ba1d5ed 100644 --- a/arch/arm/common/gic.c +++ b/arch/arm/common/gic.c @@ -109,11 +109,11 @@ static void gic_unmask_irq(unsigned int irq) } #ifdef CONFIG_SMP -static void gic_set_cpu(unsigned int irq, cpumask_t mask_val) +static void gic_set_cpu(unsigned int irq, const struct cpumask *mask_val) { void __iomem *reg = gic_dist_base(irq) + GIC_DIST_TARGET + (gic_irq(irq) & ~3); unsigned int shift = (irq % 4) * 8; - unsigned int cpu = first_cpu(mask_val); + unsigned int cpu = cpumask_first(mask_val); u32 val; spin_lock(&irq_controller_lock); diff --git a/arch/arm/kernel/irq.c b/arch/arm/kernel/irq.c index 2f3eb795fa6e..7141cee1fab7 100644 --- a/arch/arm/kernel/irq.c +++ b/arch/arm/kernel/irq.c @@ -174,7 +174,7 @@ static void route_irq(struct irq_desc *desc, unsigned int irq, unsigned int cpu) pr_debug("IRQ%u: moving from cpu%u to cpu%u\n", irq, desc->cpu, cpu); spin_lock_irq(&desc->lock); - desc->chip->set_affinity(irq, cpumask_of_cpu(cpu)); + desc->chip->set_affinity(irq, cpumask_of(cpu)); spin_unlock_irq(&desc->lock); } diff --git a/arch/arm/oprofile/op_model_mpcore.c b/arch/arm/oprofile/op_model_mpcore.c index 4de366e8b4c5..6d6bd5899240 100644 --- a/arch/arm/oprofile/op_model_mpcore.c +++ b/arch/arm/oprofile/op_model_mpcore.c @@ -260,10 +260,10 @@ static void em_stop(void) static void em_route_irq(int irq, unsigned int cpu) { struct irq_desc *desc = irq_desc + irq; - cpumask_t mask = cpumask_of_cpu(cpu); + const struct cpumask *mask = cpumask_of(cpu); spin_lock_irq(&desc->lock); - desc->affinity = mask; + desc->affinity = *mask; desc->chip->set_affinity(irq, mask); spin_unlock_irq(&desc->lock); } diff --git a/arch/cris/arch-v32/kernel/irq.c b/arch/cris/arch-v32/kernel/irq.c index 173c141ac9ba..295131fee710 100644 --- a/arch/cris/arch-v32/kernel/irq.c +++ b/arch/cris/arch-v32/kernel/irq.c @@ -325,11 +325,11 @@ static void end_crisv32_irq(unsigned int irq) { } -void set_affinity_crisv32_irq(unsigned int irq, cpumask_t dest) +void set_affinity_crisv32_irq(unsigned int irq, const struct cpumask *dest) { unsigned long flags; spin_lock_irqsave(&irq_lock, flags); - irq_allocations[irq - FIRST_IRQ].mask = dest; + irq_allocations[irq - FIRST_IRQ].mask = *dest; spin_unlock_irqrestore(&irq_lock, flags); } diff --git a/arch/ia64/hp/sim/hpsim_irq.c b/arch/ia64/hp/sim/hpsim_irq.c index c2f58ff364e7..cc0a3182db3c 100644 --- a/arch/ia64/hp/sim/hpsim_irq.c +++ b/arch/ia64/hp/sim/hpsim_irq.c @@ -22,7 +22,7 @@ hpsim_irq_noop (unsigned int irq) } static void -hpsim_set_affinity_noop (unsigned int a, cpumask_t b) +hpsim_set_affinity_noop(unsigned int a, const struct cpumask *b) { } diff --git a/arch/ia64/kernel/iosapic.c b/arch/ia64/kernel/iosapic.c index 5c4674ae8aea..c8adecd5b416 100644 --- a/arch/ia64/kernel/iosapic.c +++ b/arch/ia64/kernel/iosapic.c @@ -330,25 +330,25 @@ unmask_irq (unsigned int irq) static void -iosapic_set_affinity (unsigned int irq, cpumask_t mask) +iosapic_set_affinity(unsigned int irq, const struct cpumask *mask) { #ifdef CONFIG_SMP u32 high32, low32; - int dest, rte_index; + int cpu, dest, rte_index; int redir = (irq & IA64_IRQ_REDIRECTED) ? 1 : 0; struct iosapic_rte_info *rte; struct iosapic *iosapic; irq &= (~IA64_IRQ_REDIRECTED); - cpus_and(mask, mask, cpu_online_map); - if (cpus_empty(mask)) + cpu = cpumask_first_and(cpu_online_mask, mask); + if (cpu >= nr_cpu_ids) return; - if (irq_prepare_move(irq, first_cpu(mask))) + if (irq_prepare_move(irq, cpu)) return; - dest = cpu_physical_id(first_cpu(mask)); + dest = cpu_physical_id(cpu); if (!iosapic_intr_info[irq].count) return; /* not an IOSAPIC interrupt */ diff --git a/arch/ia64/kernel/irq.c b/arch/ia64/kernel/irq.c index 7fd18f54c056..0b6db53fedcf 100644 --- a/arch/ia64/kernel/irq.c +++ b/arch/ia64/kernel/irq.c @@ -133,7 +133,6 @@ unsigned int vectors_in_migration[NR_IRQS]; */ static void migrate_irqs(void) { - cpumask_t mask; irq_desc_t *desc; int irq, new_cpu; @@ -152,15 +151,14 @@ static void migrate_irqs(void) if (desc->status == IRQ_PER_CPU) continue; - cpus_and(mask, irq_desc[irq].affinity, cpu_online_map); - if (any_online_cpu(mask) == NR_CPUS) { + if (cpumask_any_and(&irq_desc[irq].affinity, cpu_online_mask) + >= nr_cpu_ids) { /* * Save it for phase 2 processing */ vectors_in_migration[irq] = irq; new_cpu = any_online_cpu(cpu_online_map); - mask = cpumask_of_cpu(new_cpu); /* * Al three are essential, currently WARN_ON.. maybe panic? @@ -168,7 +166,8 @@ static void migrate_irqs(void) if (desc->chip && desc->chip->disable && desc->chip->enable && desc->chip->set_affinity) { desc->chip->disable(irq); - desc->chip->set_affinity(irq, mask); + desc->chip->set_affinity(irq, + cpumask_of(new_cpu)); desc->chip->enable(irq); } else { WARN_ON((!(desc->chip) || !(desc->chip->disable) || diff --git a/arch/ia64/kernel/msi_ia64.c b/arch/ia64/kernel/msi_ia64.c index 702a09c13238..890339339035 100644 --- a/arch/ia64/kernel/msi_ia64.c +++ b/arch/ia64/kernel/msi_ia64.c @@ -49,11 +49,12 @@ static struct irq_chip ia64_msi_chip; #ifdef CONFIG_SMP -static void ia64_set_msi_irq_affinity(unsigned int irq, cpumask_t cpu_mask) +static void ia64_set_msi_irq_affinity(unsigned int irq, + const cpumask_t *cpu_mask) { struct msi_msg msg; u32 addr, data; - int cpu = first_cpu(cpu_mask); + int cpu = first_cpu(*cpu_mask); if (!cpu_online(cpu)) return; @@ -166,12 +167,11 @@ void arch_teardown_msi_irq(unsigned int irq) #ifdef CONFIG_DMAR #ifdef CONFIG_SMP -static void dmar_msi_set_affinity(unsigned int irq, cpumask_t mask) +static void dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask) { struct irq_cfg *cfg = irq_cfg + irq; struct msi_msg msg; - int cpu = first_cpu(mask); - + int cpu = cpumask_first(mask); if (!cpu_online(cpu)) return; @@ -187,7 +187,7 @@ static void dmar_msi_set_affinity(unsigned int irq, cpumask_t mask) msg.address_lo |= MSI_ADDR_DESTID_CPU(cpu_physical_id(cpu)); dmar_msi_write(irq, &msg); - irq_desc[irq].affinity = mask; + irq_desc[irq].affinity = *mask; } #endif /* CONFIG_SMP */ diff --git a/arch/ia64/kernel/smpboot.c b/arch/ia64/kernel/smpboot.c index 4ede6e571c38..11463994a7d5 100644 --- a/arch/ia64/kernel/smpboot.c +++ b/arch/ia64/kernel/smpboot.c @@ -682,7 +682,7 @@ int migrate_platform_irqs(unsigned int cpu) { int new_cpei_cpu; irq_desc_t *desc = NULL; - cpumask_t mask; + const struct cpumask *mask; int retval = 0; /* @@ -695,7 +695,7 @@ int migrate_platform_irqs(unsigned int cpu) * Now re-target the CPEI to a different processor */ new_cpei_cpu = any_online_cpu(cpu_online_map); - mask = cpumask_of_cpu(new_cpei_cpu); + mask = cpumask_of(new_cpei_cpu); set_cpei_target_cpu(new_cpei_cpu); desc = irq_desc + ia64_cpe_irq; /* diff --git a/arch/ia64/sn/kernel/irq.c b/arch/ia64/sn/kernel/irq.c index 0c66dbdd1d72..66fd705e82c0 100644 --- a/arch/ia64/sn/kernel/irq.c +++ b/arch/ia64/sn/kernel/irq.c @@ -227,14 +227,14 @@ finish_up: return new_irq_info; } -static void sn_set_affinity_irq(unsigned int irq, cpumask_t mask) +static void sn_set_affinity_irq(unsigned int irq, const struct cpumask *mask) { struct sn_irq_info *sn_irq_info, *sn_irq_info_safe; nasid_t nasid; int slice; - nasid = cpuid_to_nasid(first_cpu(mask)); - slice = cpuid_to_slice(first_cpu(mask)); + nasid = cpuid_to_nasid(cpumask_first(mask)); + slice = cpuid_to_slice(cpumask_first(mask)); list_for_each_entry_safe(sn_irq_info, sn_irq_info_safe, sn_irq_lh[irq], list) diff --git a/arch/ia64/sn/kernel/msi_sn.c b/arch/ia64/sn/kernel/msi_sn.c index 83f190ffe350..ca553b0429ce 100644 --- a/arch/ia64/sn/kernel/msi_sn.c +++ b/arch/ia64/sn/kernel/msi_sn.c @@ -151,7 +151,8 @@ int sn_setup_msi_irq(struct pci_dev *pdev, struct msi_desc *entry) } #ifdef CONFIG_SMP -static void sn_set_msi_irq_affinity(unsigned int irq, cpumask_t cpu_mask) +static void sn_set_msi_irq_affinity(unsigned int irq, + const struct cpumask *cpu_mask) { struct msi_msg msg; int slice; @@ -164,7 +165,7 @@ static void sn_set_msi_irq_affinity(unsigned int irq, cpumask_t cpu_mask) struct sn_pcibus_provider *provider; unsigned int cpu; - cpu = first_cpu(cpu_mask); + cpu = cpumask_first(cpu_mask); sn_irq_info = sn_msi_info[irq].sn_irq_info; if (sn_irq_info == NULL || sn_irq_info->irq_int_bit >= 0) return; @@ -204,7 +205,7 @@ static void sn_set_msi_irq_affinity(unsigned int irq, cpumask_t cpu_mask) msg.address_lo = (u32)(bus_addr & 0x00000000ffffffff); write_msi_msg(irq, &msg); - irq_desc[irq].affinity = cpu_mask; + irq_desc[irq].affinity = *cpu_mask; } #endif /* CONFIG_SMP */ diff --git a/arch/mips/include/asm/irq.h b/arch/mips/include/asm/irq.h index a58f0eecc68f..abc62aa744ac 100644 --- a/arch/mips/include/asm/irq.h +++ b/arch/mips/include/asm/irq.h @@ -49,7 +49,8 @@ static inline void smtc_im_ack_irq(unsigned int irq) #ifdef CONFIG_MIPS_MT_SMTC_IRQAFF #include -extern void plat_set_irq_affinity(unsigned int irq, cpumask_t affinity); +extern void plat_set_irq_affinity(unsigned int irq, + const struct cpumask *affinity); extern void smtc_forward_irq(unsigned int irq); /* diff --git a/arch/mips/kernel/cevt-bcm1480.c b/arch/mips/kernel/cevt-bcm1480.c index 0a57f86945f1..d7e21bc8cd21 100644 --- a/arch/mips/kernel/cevt-bcm1480.c +++ b/arch/mips/kernel/cevt-bcm1480.c @@ -148,6 +148,6 @@ void __cpuinit sb1480_clockevent_init(void) action->name = name; action->dev_id = cd; - irq_set_affinity(irq, cpumask_of_cpu(cpu)); + irq_set_affinity(irq, cpumask_of(cpu)); setup_irq(irq, action); } diff --git a/arch/mips/kernel/cevt-sb1250.c b/arch/mips/kernel/cevt-sb1250.c index 63ac3ad462bc..0f188cd46e03 100644 --- a/arch/mips/kernel/cevt-sb1250.c +++ b/arch/mips/kernel/cevt-sb1250.c @@ -147,6 +147,6 @@ void __cpuinit sb1250_clockevent_init(void) action->name = name; action->dev_id = cd; - irq_set_affinity(irq, cpumask_of_cpu(cpu)); + irq_set_affinity(irq, cpumask_of(cpu)); setup_irq(irq, action); } diff --git a/arch/mips/kernel/irq-gic.c b/arch/mips/kernel/irq-gic.c index f0a4bb19e096..494a49a317e9 100644 --- a/arch/mips/kernel/irq-gic.c +++ b/arch/mips/kernel/irq-gic.c @@ -155,7 +155,7 @@ static void gic_unmask_irq(unsigned int irq) static DEFINE_SPINLOCK(gic_lock); -static void gic_set_affinity(unsigned int irq, cpumask_t cpumask) +static void gic_set_affinity(unsigned int irq, const struct cpumask *cpumask) { cpumask_t tmp = CPU_MASK_NONE; unsigned long flags; @@ -164,7 +164,7 @@ static void gic_set_affinity(unsigned int irq, cpumask_t cpumask) pr_debug(KERN_DEBUG "%s called\n", __func__); irq -= _irqbase; - cpus_and(tmp, cpumask, cpu_online_map); + cpumask_and(&tmp, cpumask, cpu_online_mask); if (cpus_empty(tmp)) return; @@ -187,7 +187,7 @@ static void gic_set_affinity(unsigned int irq, cpumask_t cpumask) set_bit(irq, pcpu_masks[first_cpu(tmp)].pcpu_mask); } - irq_desc[irq].affinity = cpumask; + irq_desc[irq].affinity = *cpumask; spin_unlock_irqrestore(&gic_lock, flags); } diff --git a/arch/mips/mti-malta/malta-smtc.c b/arch/mips/mti-malta/malta-smtc.c index f84a46a8ae6e..aabd7274507b 100644 --- a/arch/mips/mti-malta/malta-smtc.c +++ b/arch/mips/mti-malta/malta-smtc.c @@ -114,9 +114,9 @@ struct plat_smp_ops msmtc_smp_ops = { */ -void plat_set_irq_affinity(unsigned int irq, cpumask_t affinity) +void plat_set_irq_affinity(unsigned int irq, const struct cpumask *affinity) { - cpumask_t tmask = affinity; + cpumask_t tmask = *affinity; int cpu = 0; void smtc_set_irq_affinity(unsigned int irq, cpumask_t aff); @@ -139,7 +139,7 @@ void plat_set_irq_affinity(unsigned int irq, cpumask_t affinity) * be made to forward to an offline "CPU". */ - for_each_cpu_mask(cpu, affinity) { + for_each_cpu(cpu, affinity) { if ((cpu_data[cpu].vpe_id != 0) || !cpu_online(cpu)) cpu_clear(cpu, tmask); } diff --git a/arch/mips/sibyte/bcm1480/irq.c b/arch/mips/sibyte/bcm1480/irq.c index a35818ed4263..12b465d404df 100644 --- a/arch/mips/sibyte/bcm1480/irq.c +++ b/arch/mips/sibyte/bcm1480/irq.c @@ -50,7 +50,7 @@ static void enable_bcm1480_irq(unsigned int irq); static void disable_bcm1480_irq(unsigned int irq); static void ack_bcm1480_irq(unsigned int irq); #ifdef CONFIG_SMP -static void bcm1480_set_affinity(unsigned int irq, cpumask_t mask); +static void bcm1480_set_affinity(unsigned int irq, const struct cpumask *mask); #endif #ifdef CONFIG_PCI @@ -109,7 +109,7 @@ void bcm1480_unmask_irq(int cpu, int irq) } #ifdef CONFIG_SMP -static void bcm1480_set_affinity(unsigned int irq, cpumask_t mask) +static void bcm1480_set_affinity(unsigned int irq, const struct cpumask *mask) { int i = 0, old_cpu, cpu, int_on, k; u64 cur_ints; @@ -117,11 +117,11 @@ static void bcm1480_set_affinity(unsigned int irq, cpumask_t mask) unsigned long flags; unsigned int irq_dirty; - if (cpus_weight(mask) != 1) { + if (cpumask_weight(mask) != 1) { printk("attempted to set irq affinity for irq %d to multiple CPUs\n", irq); return; } - i = first_cpu(mask); + i = cpumask_first(mask); /* Convert logical CPU to physical CPU */ cpu = cpu_logical_map(i); diff --git a/arch/mips/sibyte/sb1250/irq.c b/arch/mips/sibyte/sb1250/irq.c index a5158483986e..808ac2959b8c 100644 --- a/arch/mips/sibyte/sb1250/irq.c +++ b/arch/mips/sibyte/sb1250/irq.c @@ -50,7 +50,7 @@ static void enable_sb1250_irq(unsigned int irq); static void disable_sb1250_irq(unsigned int irq); static void ack_sb1250_irq(unsigned int irq); #ifdef CONFIG_SMP -static void sb1250_set_affinity(unsigned int irq, cpumask_t mask); +static void sb1250_set_affinity(unsigned int irq, const struct cpumask *mask); #endif #ifdef CONFIG_SIBYTE_HAS_LDT @@ -103,16 +103,16 @@ void sb1250_unmask_irq(int cpu, int irq) } #ifdef CONFIG_SMP -static void sb1250_set_affinity(unsigned int irq, cpumask_t mask) +static void sb1250_set_affinity(unsigned int irq, const struct cpumask *mask) { int i = 0, old_cpu, cpu, int_on; u64 cur_ints; struct irq_desc *desc = irq_desc + irq; unsigned long flags; - i = first_cpu(mask); + i = cpumask_first(mask); - if (cpus_weight(mask) > 1) { + if (cpumask_weight(mask) > 1) { printk("attempted to set irq affinity for irq %d to multiple CPUs\n", irq); return; } diff --git a/arch/parisc/kernel/irq.c b/arch/parisc/kernel/irq.c index 23ef950df008..4cea935e2f99 100644 --- a/arch/parisc/kernel/irq.c +++ b/arch/parisc/kernel/irq.c @@ -131,12 +131,12 @@ int cpu_check_affinity(unsigned int irq, cpumask_t *dest) return 0; } -static void cpu_set_affinity_irq(unsigned int irq, cpumask_t dest) +static void cpu_set_affinity_irq(unsigned int irq, const struct cpumask *dest) { - if (cpu_check_affinity(irq, &dest)) + if (cpu_check_affinity(irq, dest)) return; - irq_desc[irq].affinity = dest; + irq_desc[irq].affinity = *dest; } #endif diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c index ac222d0ab12e..23b8b5e36f98 100644 --- a/arch/powerpc/kernel/irq.c +++ b/arch/powerpc/kernel/irq.c @@ -237,7 +237,7 @@ void fixup_irqs(cpumask_t map) mask = map; } if (irq_desc[irq].chip->set_affinity) - irq_desc[irq].chip->set_affinity(irq, mask); + irq_desc[irq].chip->set_affinity(irq, &mask); else if (irq_desc[irq].action && !(warned++)) printk("Cannot set affinity for irq %i\n", irq); } diff --git a/arch/powerpc/platforms/pseries/xics.c b/arch/powerpc/platforms/pseries/xics.c index 64d24310ce7e..424b335a71c8 100644 --- a/arch/powerpc/platforms/pseries/xics.c +++ b/arch/powerpc/platforms/pseries/xics.c @@ -332,7 +332,7 @@ static void xics_eoi_lpar(unsigned int virq) lpar_xirr_info_set((0xff << 24) | irq); } -static void xics_set_affinity(unsigned int virq, cpumask_t cpumask) +static void xics_set_affinity(unsigned int virq, const struct cpumask *cpumask) { unsigned int irq; int status; @@ -358,7 +358,7 @@ static void xics_set_affinity(unsigned int virq, cpumask_t cpumask) irq_server = get_irq_server(virq, 1); if (irq_server == -1) { char cpulist[128]; - cpumask_scnprintf(cpulist, sizeof(cpulist), &cpumask); + cpumask_scnprintf(cpulist, sizeof(cpulist), cpumask); printk(KERN_WARNING "%s: No online cpus in the mask %s for irq %d\n", __func__, cpulist, virq); @@ -845,7 +845,7 @@ void xics_migrate_irqs_away(void) /* Reset affinity to all cpus */ irq_desc[virq].affinity = CPU_MASK_ALL; - desc->chip->set_affinity(virq, CPU_MASK_ALL); + desc->chip->set_affinity(virq, cpu_all_mask); unlock: spin_unlock_irqrestore(&desc->lock, flags); } diff --git a/arch/powerpc/sysdev/mpic.c b/arch/powerpc/sysdev/mpic.c index 1890fb085cde..5d7f9f0c93c3 100644 --- a/arch/powerpc/sysdev/mpic.c +++ b/arch/powerpc/sysdev/mpic.c @@ -817,7 +817,7 @@ static void mpic_end_ipi(unsigned int irq) #endif /* CONFIG_SMP */ -void mpic_set_affinity(unsigned int irq, cpumask_t cpumask) +void mpic_set_affinity(unsigned int irq, const struct cpumask *cpumask) { struct mpic *mpic = mpic_from_irq(irq); unsigned int src = mpic_irq_to_hw(irq); @@ -829,7 +829,7 @@ void mpic_set_affinity(unsigned int irq, cpumask_t cpumask) } else { cpumask_t tmp; - cpus_and(tmp, cpumask, cpu_online_map); + cpumask_and(&tmp, cpumask, cpu_online_mask); mpic_irq_write(src, MPIC_INFO(IRQ_DESTINATION), mpic_physmask(cpus_addr(tmp)[0])); diff --git a/arch/powerpc/sysdev/mpic.h b/arch/powerpc/sysdev/mpic.h index 6209c62a426d..3cef2af10f42 100644 --- a/arch/powerpc/sysdev/mpic.h +++ b/arch/powerpc/sysdev/mpic.h @@ -36,6 +36,6 @@ static inline int mpic_pasemi_msi_init(struct mpic *mpic) extern int mpic_set_irq_type(unsigned int virq, unsigned int flow_type); extern void mpic_set_vector(unsigned int virq, unsigned int vector); -extern void mpic_set_affinity(unsigned int irq, cpumask_t cpumask); +extern void mpic_set_affinity(unsigned int irq, const struct cpumask *cpumask); #endif /* _POWERPC_SYSDEV_MPIC_H */ diff --git a/arch/sparc64/kernel/irq.c b/arch/sparc64/kernel/irq.c index 52fc836f464d..4aaf18e83c8c 100644 --- a/arch/sparc64/kernel/irq.c +++ b/arch/sparc64/kernel/irq.c @@ -312,7 +312,8 @@ static void sun4u_irq_enable(unsigned int virt_irq) } } -static void sun4u_set_affinity(unsigned int virt_irq, cpumask_t mask) +static void sun4u_set_affinity(unsigned int virt_irq, + const struct cpumask *mask) { sun4u_irq_enable(virt_irq); } @@ -362,7 +363,8 @@ static void sun4v_irq_enable(unsigned int virt_irq) ino, err); } -static void sun4v_set_affinity(unsigned int virt_irq, cpumask_t mask) +static void sun4v_set_affinity(unsigned int virt_irq, + const struct cpumask *mask) { unsigned int ino = virt_irq_table[virt_irq].dev_ino; unsigned long cpuid = irq_choose_cpu(virt_irq); @@ -429,7 +431,8 @@ static void sun4v_virq_enable(unsigned int virt_irq) dev_handle, dev_ino, err); } -static void sun4v_virt_set_affinity(unsigned int virt_irq, cpumask_t mask) +static void sun4v_virt_set_affinity(unsigned int virt_irq, + const struct cpumask *mask) { unsigned long cpuid, dev_handle, dev_ino; int err; @@ -788,7 +791,7 @@ void fixup_irqs(void) !(irq_desc[irq].status & IRQ_PER_CPU)) { if (irq_desc[irq].chip->set_affinity) irq_desc[irq].chip->set_affinity(irq, - irq_desc[irq].affinity); + &irq_desc[irq].affinity); } spin_unlock_irqrestore(&irq_desc[irq].lock, flags); } diff --git a/arch/sparc64/kernel/of_device.c b/arch/sparc64/kernel/of_device.c index 0f616ae3246c..df2efb7fc14c 100644 --- a/arch/sparc64/kernel/of_device.c +++ b/arch/sparc64/kernel/of_device.c @@ -780,7 +780,7 @@ out: if (nid != -1) { cpumask_t numa_mask = node_to_cpumask(nid); - irq_set_affinity(irq, numa_mask); + irq_set_affinity(irq, &numa_mask); } return irq; diff --git a/arch/sparc64/kernel/pci_msi.c b/arch/sparc64/kernel/pci_msi.c index 2e680f34f727..0d0cd815e83e 100644 --- a/arch/sparc64/kernel/pci_msi.c +++ b/arch/sparc64/kernel/pci_msi.c @@ -288,7 +288,7 @@ static int bringup_one_msi_queue(struct pci_pbm_info *pbm, if (nid != -1) { cpumask_t numa_mask = node_to_cpumask(nid); - irq_set_affinity(irq, numa_mask); + irq_set_affinity(irq, &numa_mask); } err = request_irq(irq, sparc64_msiq_interrupt, 0, "MSIQ", diff --git a/arch/x86/kernel/hpet.c b/arch/x86/kernel/hpet.c index 067d8de913f6..940f25851e1e 100644 --- a/arch/x86/kernel/hpet.c +++ b/arch/x86/kernel/hpet.c @@ -301,7 +301,7 @@ static void hpet_set_mode(enum clock_event_mode mode, struct hpet_dev *hdev = EVT_TO_HPET_DEV(evt); hpet_setup_msi_irq(hdev->irq); disable_irq(hdev->irq); - irq_set_affinity(hdev->irq, cpumask_of_cpu(hdev->cpu)); + irq_set_affinity(hdev->irq, cpumask_of(hdev->cpu)); enable_irq(hdev->irq); } break; @@ -449,7 +449,7 @@ static int hpet_setup_irq(struct hpet_dev *dev) return -1; disable_irq(dev->irq); - irq_set_affinity(dev->irq, cpumask_of_cpu(dev->cpu)); + irq_set_affinity(dev->irq, cpumask_of(dev->cpu)); enable_irq(dev->irq); printk(KERN_DEBUG "hpet: %s irq %d for MSI\n", diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index 9043251210fb..1184210e6d0c 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -361,7 +361,8 @@ static void __target_IO_APIC_irq(unsigned int irq, unsigned int dest, u8 vector) static int assign_irq_vector(int irq, cpumask_t mask); -static void set_ioapic_affinity_irq(unsigned int irq, cpumask_t mask) +static void set_ioapic_affinity_irq(unsigned int irq, + const struct cpumask *mask) { struct irq_cfg *cfg; unsigned long flags; @@ -369,15 +370,14 @@ static void set_ioapic_affinity_irq(unsigned int irq, cpumask_t mask) cpumask_t tmp; struct irq_desc *desc; - cpus_and(tmp, mask, cpu_online_map); - if (cpus_empty(tmp)) + if (!cpumask_intersects(mask, cpu_online_mask)) return; cfg = irq_cfg(irq); - if (assign_irq_vector(irq, mask)) + if (assign_irq_vector(irq, *mask)) return; - cpus_and(tmp, cfg->domain, mask); + cpumask_and(&tmp, &cfg->domain, mask); dest = cpu_mask_to_apicid(tmp); /* * Only the high 8 bits are valid. @@ -387,7 +387,7 @@ static void set_ioapic_affinity_irq(unsigned int irq, cpumask_t mask) desc = irq_to_desc(irq); spin_lock_irqsave(&ioapic_lock, flags); __target_IO_APIC_irq(irq, dest, cfg->vector); - desc->affinity = mask; + cpumask_copy(&desc->affinity, mask); spin_unlock_irqrestore(&ioapic_lock, flags); } #endif /* CONFIG_SMP */ @@ -2189,7 +2189,7 @@ static void ir_irq_migration(struct work_struct *work) continue; } - desc->chip->set_affinity(irq, desc->pending_mask); + desc->chip->set_affinity(irq, &desc->pending_mask); spin_unlock_irqrestore(&desc->lock, flags); } } @@ -2198,18 +2198,19 @@ static void ir_irq_migration(struct work_struct *work) /* * Migrates the IRQ destination in the process context. */ -static void set_ir_ioapic_affinity_irq(unsigned int irq, cpumask_t mask) +static void set_ir_ioapic_affinity_irq(unsigned int irq, + const struct cpumask *mask) { struct irq_desc *desc = irq_to_desc(irq); if (desc->status & IRQ_LEVEL) { desc->status |= IRQ_MOVE_PENDING; - desc->pending_mask = mask; + cpumask_copy(&desc->pending_mask, mask); migrate_irq_remapped_level(irq); return; } - migrate_ioapic_irq(irq, mask); + migrate_ioapic_irq(irq, *mask); } #endif @@ -3027,7 +3028,7 @@ static int msi_compose_msg(struct pci_dev *pdev, unsigned int irq, struct msi_ms } #ifdef CONFIG_SMP -static void set_msi_irq_affinity(unsigned int irq, cpumask_t mask) +static void set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) { struct irq_cfg *cfg; struct msi_msg msg; @@ -3035,15 +3036,14 @@ static void set_msi_irq_affinity(unsigned int irq, cpumask_t mask) cpumask_t tmp; struct irq_desc *desc; - cpus_and(tmp, mask, cpu_online_map); - if (cpus_empty(tmp)) + if (!cpumask_intersects(mask, cpu_online_mask)) return; - if (assign_irq_vector(irq, mask)) + if (assign_irq_vector(irq, *mask)) return; cfg = irq_cfg(irq); - cpus_and(tmp, cfg->domain, mask); + cpumask_and(&tmp, &cfg->domain, mask); dest = cpu_mask_to_apicid(tmp); read_msi_msg(irq, &msg); @@ -3055,7 +3055,7 @@ static void set_msi_irq_affinity(unsigned int irq, cpumask_t mask) write_msi_msg(irq, &msg); desc = irq_to_desc(irq); - desc->affinity = mask; + cpumask_copy(&desc->affinity, mask); } #ifdef CONFIG_INTR_REMAP @@ -3063,7 +3063,8 @@ static void set_msi_irq_affinity(unsigned int irq, cpumask_t mask) * Migrate the MSI irq to another cpumask. This migration is * done in the process context using interrupt-remapping hardware. */ -static void ir_set_msi_irq_affinity(unsigned int irq, cpumask_t mask) +static void ir_set_msi_irq_affinity(unsigned int irq, + const struct cpumask *mask) { struct irq_cfg *cfg; unsigned int dest; @@ -3071,18 +3072,17 @@ static void ir_set_msi_irq_affinity(unsigned int irq, cpumask_t mask) struct irte irte; struct irq_desc *desc; - cpus_and(tmp, mask, cpu_online_map); - if (cpus_empty(tmp)) + if (!cpumask_intersects(mask, cpu_online_mask)) return; if (get_irte(irq, &irte)) return; - if (assign_irq_vector(irq, mask)) + if (assign_irq_vector(irq, *mask)) return; cfg = irq_cfg(irq); - cpus_and(tmp, cfg->domain, mask); + cpumask_and(&tmp, &cfg->domain, mask); dest = cpu_mask_to_apicid(tmp); irte.vector = cfg->vector; @@ -3106,7 +3106,7 @@ static void ir_set_msi_irq_affinity(unsigned int irq, cpumask_t mask) } desc = irq_to_desc(irq); - desc->affinity = mask; + cpumask_copy(&desc->affinity, mask); } #endif #endif /* CONFIG_SMP */ @@ -3308,7 +3308,7 @@ void arch_teardown_msi_irq(unsigned int irq) #ifdef CONFIG_DMAR #ifdef CONFIG_SMP -static void dmar_msi_set_affinity(unsigned int irq, cpumask_t mask) +static void dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask) { struct irq_cfg *cfg; struct msi_msg msg; @@ -3316,15 +3316,14 @@ static void dmar_msi_set_affinity(unsigned int irq, cpumask_t mask) cpumask_t tmp; struct irq_desc *desc; - cpus_and(tmp, mask, cpu_online_map); - if (cpus_empty(tmp)) + if (!cpumask_intersects(mask, cpu_online_mask)) return; - if (assign_irq_vector(irq, mask)) + if (assign_irq_vector(irq, *mask)) return; cfg = irq_cfg(irq); - cpus_and(tmp, cfg->domain, mask); + cpumask_and(&tmp, &cfg->domain, mask); dest = cpu_mask_to_apicid(tmp); dmar_msi_read(irq, &msg); @@ -3336,7 +3335,7 @@ static void dmar_msi_set_affinity(unsigned int irq, cpumask_t mask) dmar_msi_write(irq, &msg); desc = irq_to_desc(irq); - desc->affinity = mask; + cpumask_copy(&desc->affinity, mask); } #endif /* CONFIG_SMP */ @@ -3369,7 +3368,7 @@ int arch_setup_dmar_msi(unsigned int irq) #ifdef CONFIG_HPET_TIMER #ifdef CONFIG_SMP -static void hpet_msi_set_affinity(unsigned int irq, cpumask_t mask) +static void hpet_msi_set_affinity(unsigned int irq, const struct cpumask *mask) { struct irq_cfg *cfg; struct irq_desc *desc; @@ -3377,15 +3376,14 @@ static void hpet_msi_set_affinity(unsigned int irq, cpumask_t mask) unsigned int dest; cpumask_t tmp; - cpus_and(tmp, mask, cpu_online_map); - if (cpus_empty(tmp)) + if (!cpumask_intersects(mask, cpu_online_mask)) return; - if (assign_irq_vector(irq, mask)) + if (assign_irq_vector(irq, *mask)) return; cfg = irq_cfg(irq); - cpus_and(tmp, cfg->domain, mask); + cpumask_and(&tmp, &cfg->domain, mask); dest = cpu_mask_to_apicid(tmp); hpet_msi_read(irq, &msg); @@ -3397,7 +3395,7 @@ static void hpet_msi_set_affinity(unsigned int irq, cpumask_t mask) hpet_msi_write(irq, &msg); desc = irq_to_desc(irq); - desc->affinity = mask; + cpumask_copy(&desc->affinity, mask); } #endif /* CONFIG_SMP */ @@ -3451,27 +3449,26 @@ static void target_ht_irq(unsigned int irq, unsigned int dest, u8 vector) write_ht_irq_msg(irq, &msg); } -static void set_ht_irq_affinity(unsigned int irq, cpumask_t mask) +static void set_ht_irq_affinity(unsigned int irq, const struct cpumask *mask) { struct irq_cfg *cfg; unsigned int dest; cpumask_t tmp; struct irq_desc *desc; - cpus_and(tmp, mask, cpu_online_map); - if (cpus_empty(tmp)) + if (!cpumask_intersects(mask, cpu_online_mask)) return; - if (assign_irq_vector(irq, mask)) + if (assign_irq_vector(irq, *mask)) return; cfg = irq_cfg(irq); - cpus_and(tmp, cfg->domain, mask); + cpumask_and(&tmp, &cfg->domain, mask); dest = cpu_mask_to_apicid(tmp); target_ht_irq(irq, dest, cfg->vector); desc = irq_to_desc(irq); - desc->affinity = mask; + cpumask_copy(&desc->affinity, mask); } #endif @@ -3794,10 +3791,10 @@ void __init setup_ioapic_dest(void) #ifdef CONFIG_INTR_REMAP if (intr_remapping_enabled) - set_ir_ioapic_affinity_irq(irq, mask); + set_ir_ioapic_affinity_irq(irq, &mask); else #endif - set_ioapic_affinity_irq(irq, mask); + set_ioapic_affinity_irq(irq, &mask); } } diff --git a/arch/x86/kernel/irq_32.c b/arch/x86/kernel/irq_32.c index a51382672de0..87870a49be4e 100644 --- a/arch/x86/kernel/irq_32.c +++ b/arch/x86/kernel/irq_32.c @@ -251,7 +251,7 @@ void fixup_irqs(cpumask_t map) mask = map; } if (desc->chip->set_affinity) - desc->chip->set_affinity(irq, mask); + desc->chip->set_affinity(irq, &mask); else if (desc->action && !(warned++)) printk("Cannot set affinity for irq %i\n", irq); } diff --git a/arch/x86/kernel/irq_64.c b/arch/x86/kernel/irq_64.c index 60eb84eb77a0..7d37f847544d 100644 --- a/arch/x86/kernel/irq_64.c +++ b/arch/x86/kernel/irq_64.c @@ -116,7 +116,7 @@ void fixup_irqs(cpumask_t map) desc->chip->mask(irq); if (desc->chip->set_affinity) - desc->chip->set_affinity(irq, mask); + desc->chip->set_affinity(irq, &mask); else if (!(warned++)) set_affinity = 0; diff --git a/drivers/parisc/iosapic.c b/drivers/parisc/iosapic.c index 7beffcab2745..9dedbbd218c3 100644 --- a/drivers/parisc/iosapic.c +++ b/drivers/parisc/iosapic.c @@ -704,16 +704,17 @@ static unsigned int iosapic_startup_irq(unsigned int irq) } #ifdef CONFIG_SMP -static void iosapic_set_affinity_irq(unsigned int irq, cpumask_t dest) +static void iosapic_set_affinity_irq(unsigned int irq, + const struct cpumask *dest) { struct vector_info *vi = iosapic_get_vector(irq); u32 d0, d1, dummy_d0; unsigned long flags; - if (cpu_check_affinity(irq, &dest)) + if (cpu_check_affinity(irq, dest)) return; - vi->txn_addr = txn_affinity_addr(irq, first_cpu(dest)); + vi->txn_addr = txn_affinity_addr(irq, cpumask_first(dest)); spin_lock_irqsave(&iosapic_lock, flags); /* d1 contains the destination CPU, so only want to set that diff --git a/drivers/xen/events.c b/drivers/xen/events.c index 1e3b934a4cf7..eba5ec5b020e 100644 --- a/drivers/xen/events.c +++ b/drivers/xen/events.c @@ -579,7 +579,7 @@ void rebind_evtchn_irq(int evtchn, int irq) spin_unlock(&irq_mapping_update_lock); /* new event channels are always bound to cpu 0 */ - irq_set_affinity(irq, cpumask_of_cpu(0)); + irq_set_affinity(irq, cpumask_of(0)); /* Unmask the event channel. */ enable_irq(irq); @@ -608,9 +608,9 @@ static void rebind_irq_to_cpu(unsigned irq, unsigned tcpu) } -static void set_affinity_irq(unsigned irq, cpumask_t dest) +static void set_affinity_irq(unsigned irq, const struct cpumask *dest) { - unsigned tcpu = first_cpu(dest); + unsigned tcpu = cpumask_first(dest); rebind_irq_to_cpu(irq, tcpu); } diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h index f58a0cf8929a..48e63934fabe 100644 --- a/include/linux/interrupt.h +++ b/include/linux/interrupt.h @@ -109,13 +109,13 @@ extern void enable_irq(unsigned int irq); extern cpumask_t irq_default_affinity; -extern int irq_set_affinity(unsigned int irq, cpumask_t cpumask); +extern int irq_set_affinity(unsigned int irq, const struct cpumask *cpumask); extern int irq_can_set_affinity(unsigned int irq); extern int irq_select_affinity(unsigned int irq); #else /* CONFIG_SMP */ -static inline int irq_set_affinity(unsigned int irq, cpumask_t cpumask) +static inline int irq_set_affinity(unsigned int irq, const struct cpumask *m) { return -EINVAL; } diff --git a/include/linux/irq.h b/include/linux/irq.h index 3dddfa703ebd..ab70fd604d3a 100644 --- a/include/linux/irq.h +++ b/include/linux/irq.h @@ -113,7 +113,8 @@ struct irq_chip { void (*eoi)(unsigned int irq); void (*end)(unsigned int irq); - void (*set_affinity)(unsigned int irq, cpumask_t dest); + void (*set_affinity)(unsigned int irq, + const struct cpumask *dest); int (*retrigger)(unsigned int irq); int (*set_type)(unsigned int irq, unsigned int flow_type); int (*set_wake)(unsigned int irq, unsigned int on); diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index 10b5092e9bfe..58d8e31daa49 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -45,7 +45,7 @@ void dynamic_irq_init(unsigned int irq) desc->irq_count = 0; desc->irqs_unhandled = 0; #ifdef CONFIG_SMP - cpus_setall(desc->affinity); + cpumask_setall(&desc->affinity); #endif spin_unlock_irqrestore(&desc->lock, flags); } diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index 801addda3c43..10ad2f87ed9a 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -79,7 +79,7 @@ int irq_can_set_affinity(unsigned int irq) * @cpumask: cpumask * */ -int irq_set_affinity(unsigned int irq, cpumask_t cpumask) +int irq_set_affinity(unsigned int irq, const struct cpumask *cpumask) { struct irq_desc *desc = irq_to_desc(irq); unsigned long flags; @@ -91,14 +91,14 @@ int irq_set_affinity(unsigned int irq, cpumask_t cpumask) #ifdef CONFIG_GENERIC_PENDING_IRQ if (desc->status & IRQ_MOVE_PCNTXT || desc->status & IRQ_DISABLED) { - desc->affinity = cpumask; + cpumask_copy(&desc->affinity, cpumask); desc->chip->set_affinity(irq, cpumask); } else { desc->status |= IRQ_MOVE_PENDING; - desc->pending_mask = cpumask; + cpumask_copy(&desc->pending_mask, cpumask); } #else - desc->affinity = cpumask; + cpumask_copy(&desc->affinity, cpumask); desc->chip->set_affinity(irq, cpumask); #endif desc->status |= IRQ_AFFINITY_SET; @@ -112,26 +112,24 @@ int irq_set_affinity(unsigned int irq, cpumask_t cpumask) */ int do_irq_select_affinity(unsigned int irq, struct irq_desc *desc) { - cpumask_t mask; - if (!irq_can_set_affinity(irq)) return 0; - cpus_and(mask, cpu_online_map, irq_default_affinity); - /* * Preserve an userspace affinity setup, but make sure that * one of the targets is online. */ if (desc->status & (IRQ_AFFINITY_SET | IRQ_NO_BALANCING)) { - if (cpus_intersects(desc->affinity, cpu_online_map)) - mask = desc->affinity; + if (cpumask_any_and(&desc->affinity, cpu_online_mask) + < nr_cpu_ids) + goto set_affinity; else desc->status &= ~IRQ_AFFINITY_SET; } - desc->affinity = mask; - desc->chip->set_affinity(irq, mask); + cpumask_and(&desc->affinity, cpu_online_mask, &irq_default_affinity); +set_affinity: + desc->chip->set_affinity(irq, &desc->affinity); return 0; } diff --git a/kernel/irq/migration.c b/kernel/irq/migration.c index 9db681d95814..bd72329e630c 100644 --- a/kernel/irq/migration.c +++ b/kernel/irq/migration.c @@ -4,7 +4,6 @@ void move_masked_irq(int irq) { struct irq_desc *desc = irq_to_desc(irq); - cpumask_t tmp; if (likely(!(desc->status & IRQ_MOVE_PENDING))) return; @@ -19,7 +18,7 @@ void move_masked_irq(int irq) desc->status &= ~IRQ_MOVE_PENDING; - if (unlikely(cpus_empty(desc->pending_mask))) + if (unlikely(cpumask_empty(&desc->pending_mask))) return; if (!desc->chip->set_affinity) @@ -27,8 +26,6 @@ void move_masked_irq(int irq) assert_spin_locked(&desc->lock); - cpus_and(tmp, desc->pending_mask, cpu_online_map); - /* * If there was a valid mask to work with, please * do the disable, re-program, enable sequence. @@ -41,10 +38,13 @@ void move_masked_irq(int irq) * For correct operation this depends on the caller * masking the irqs. */ - if (likely(!cpus_empty(tmp))) { - desc->chip->set_affinity(irq,tmp); + if (likely(cpumask_any_and(&desc->pending_mask, cpu_online_mask) + < nr_cpu_ids)) { + cpumask_and(&desc->affinity, + &desc->pending_mask, cpu_online_mask); + desc->chip->set_affinity(irq, &desc->affinity); } - cpus_clear(desc->pending_mask); + cpumask_clear(&desc->pending_mask); } void move_native_irq(int irq) diff --git a/kernel/irq/proc.c b/kernel/irq/proc.c index f293349d49d0..8e91c9762520 100644 --- a/kernel/irq/proc.c +++ b/kernel/irq/proc.c @@ -40,33 +40,42 @@ static ssize_t irq_affinity_proc_write(struct file *file, const char __user *buffer, size_t count, loff_t *pos) { unsigned int irq = (int)(long)PDE(file->f_path.dentry->d_inode)->data; - cpumask_t new_value; + cpumask_var_t new_value; int err; if (!irq_to_desc(irq)->chip->set_affinity || no_irq_affinity || irq_balancing_disabled(irq)) return -EIO; - err = cpumask_parse_user(buffer, count, &new_value); + if (!alloc_cpumask_var(&new_value, GFP_KERNEL)) + return -ENOMEM; + + err = cpumask_parse_user(buffer, count, new_value); if (err) - return err; + goto free_cpumask; - if (!is_affinity_mask_valid(new_value)) - return -EINVAL; + if (!is_affinity_mask_valid(*new_value)) { + err = -EINVAL; + goto free_cpumask; + } /* * Do not allow disabling IRQs completely - it's a too easy * way to make the system unusable accidentally :-) At least * one online CPU still has to be targeted. */ - if (!cpus_intersects(new_value, cpu_online_map)) + if (!cpumask_intersects(new_value, cpu_online_mask)) { /* Special case for empty set - allow the architecture code to set default SMP affinity. */ - return irq_select_affinity_usr(irq) ? -EINVAL : count; - - irq_set_affinity(irq, new_value); + err = irq_select_affinity_usr(irq) ? -EINVAL : count; + } else { + irq_set_affinity(irq, new_value); + err = count; + } - return count; +free_cpumask: + free_cpumask_var(new_value); + return err; } static int irq_affinity_proc_open(struct inode *inode, struct file *file) diff --git a/kernel/time/tick-common.c b/kernel/time/tick-common.c index df12434b43ca..ab65d217583f 100644 --- a/kernel/time/tick-common.c +++ b/kernel/time/tick-common.c @@ -136,7 +136,7 @@ void tick_setup_periodic(struct clock_event_device *dev, int broadcast) */ static void tick_setup_device(struct tick_device *td, struct clock_event_device *newdev, int cpu, - const cpumask_t *cpumask) + const struct cpumask *cpumask) { ktime_t next_event; void (*handler)(struct clock_event_device *) = NULL; @@ -171,8 +171,8 @@ static void tick_setup_device(struct tick_device *td, * When the device is not per cpu, pin the interrupt to the * current cpu: */ - if (!cpus_equal(newdev->cpumask, *cpumask)) - irq_set_affinity(newdev->irq, *cpumask); + if (!cpumask_equal(&newdev->cpumask, cpumask)) + irq_set_affinity(newdev->irq, cpumask); /* * When global broadcasting is active, check if the current -- cgit v1.2.3 From 320ab2b0b1e08e3805a3e1084a2f0eb1938d5d67 Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Sat, 13 Dec 2008 21:20:26 +1030 Subject: cpumask: convert struct clock_event_device to cpumask pointers. Impact: change calling convention of existing clock_event APIs struct clock_event_timer's cpumask field gets changed to take pointer, as does the ->broadcast function. Another single-patch change. For safety, we BUG_ON() in clockevents_register_device() if it's not set. Signed-off-by: Rusty Russell Cc: Ingo Molnar --- arch/arm/mach-at91/at91rm9200_time.c | 3 +-- arch/arm/mach-at91/at91sam926x_time.c | 2 +- arch/arm/mach-davinci/time.c | 2 +- arch/arm/mach-imx/time.c | 2 +- arch/arm/mach-ixp4xx/common.c | 2 +- arch/arm/mach-msm/timer.c | 2 +- arch/arm/mach-ns9xxx/time-ns9360.c | 2 +- arch/arm/mach-omap1/time.c | 2 +- arch/arm/mach-omap1/timer32k.c | 2 +- arch/arm/mach-omap2/timer-gp.c | 2 +- arch/arm/mach-pxa/time.c | 2 +- arch/arm/mach-realview/core.c | 2 +- arch/arm/mach-realview/localtimer.c | 4 ++-- arch/arm/mach-sa1100/time.c | 2 +- arch/arm/mach-versatile/core.c | 2 +- arch/arm/plat-mxc/time.c | 2 +- arch/arm/plat-orion/time.c | 2 +- arch/avr32/kernel/time.c | 2 +- arch/blackfin/kernel/time-ts.c | 2 +- arch/m68knommu/platform/coldfire/pit.c | 2 +- arch/mips/jazz/irq.c | 2 +- arch/mips/kernel/cevt-bcm1480.c | 2 +- arch/mips/kernel/cevt-ds1287.c | 2 +- arch/mips/kernel/cevt-gt641xx.c | 2 +- arch/mips/kernel/cevt-r4k.c | 2 +- arch/mips/kernel/cevt-sb1250.c | 2 +- arch/mips/kernel/cevt-smtc.c | 2 +- arch/mips/kernel/cevt-txx9.c | 2 +- arch/mips/kernel/i8253.c | 2 +- arch/mips/nxp/pnx8550/common/time.c | 1 + arch/mips/sgi-ip27/ip27-timer.c | 2 +- arch/mips/sni/time.c | 2 +- arch/powerpc/kernel/time.c | 2 +- arch/s390/kernel/time.c | 2 +- arch/sh/include/asm/smp.h | 2 +- arch/sh/kernel/smp.c | 4 ++-- arch/sh/kernel/timers/timer-broadcast.c | 2 +- arch/sh/kernel/timers/timer-tmu.c | 2 +- arch/sparc64/kernel/time.c | 2 +- arch/um/kernel/time.c | 2 +- arch/x86/kernel/apic.c | 8 ++++---- arch/x86/kernel/hpet.c | 4 ++-- arch/x86/kernel/i8253.c | 2 +- arch/x86/kernel/mfgpt_32.c | 2 +- arch/x86/kernel/vmiclock_32.c | 2 +- arch/x86/lguest/boot.c | 2 +- arch/x86/xen/time.c | 2 +- drivers/clocksource/tcb_clksrc.c | 2 +- include/linux/clockchips.h | 4 ++-- kernel/time/clockevents.c | 2 ++ kernel/time/tick-broadcast.c | 2 +- kernel/time/tick-common.c | 8 ++++---- 52 files changed, 63 insertions(+), 61 deletions(-) (limited to 'arch/x86') diff --git a/arch/arm/mach-at91/at91rm9200_time.c b/arch/arm/mach-at91/at91rm9200_time.c index a72e798a2a40..72f51d39202c 100644 --- a/arch/arm/mach-at91/at91rm9200_time.c +++ b/arch/arm/mach-at91/at91rm9200_time.c @@ -169,7 +169,6 @@ static struct clock_event_device clkevt = { .features = CLOCK_EVT_FEAT_PERIODIC | CLOCK_EVT_FEAT_ONESHOT, .shift = 32, .rating = 150, - .cpumask = CPU_MASK_CPU0, .set_next_event = clkevt32k_next_event, .set_mode = clkevt32k_mode, }; @@ -197,7 +196,7 @@ void __init at91rm9200_timer_init(void) clkevt.mult = div_sc(AT91_SLOW_CLOCK, NSEC_PER_SEC, clkevt.shift); clkevt.max_delta_ns = clockevent_delta2ns(AT91_ST_ALMV, &clkevt); clkevt.min_delta_ns = clockevent_delta2ns(2, &clkevt) + 1; - clkevt.cpumask = cpumask_of_cpu(0); + clkevt.cpumask = cpumask_of(0); clockevents_register_device(&clkevt); /* register clocksource */ diff --git a/arch/arm/mach-at91/at91sam926x_time.c b/arch/arm/mach-at91/at91sam926x_time.c index 122fd77ed580..b63e1d5f1bad 100644 --- a/arch/arm/mach-at91/at91sam926x_time.c +++ b/arch/arm/mach-at91/at91sam926x_time.c @@ -91,7 +91,6 @@ static struct clock_event_device pit_clkevt = { .features = CLOCK_EVT_FEAT_PERIODIC, .shift = 32, .rating = 100, - .cpumask = CPU_MASK_CPU0, .set_mode = pit_clkevt_mode, }; @@ -173,6 +172,7 @@ static void __init at91sam926x_pit_init(void) /* Set up and register clockevents */ pit_clkevt.mult = div_sc(pit_rate, NSEC_PER_SEC, pit_clkevt.shift); + pit_clkevt.cpumask = cpumask_of(0); clockevents_register_device(&pit_clkevt); } diff --git a/arch/arm/mach-davinci/time.c b/arch/arm/mach-davinci/time.c index 3b9a296b5c4b..f8bcd29d17a6 100644 --- a/arch/arm/mach-davinci/time.c +++ b/arch/arm/mach-davinci/time.c @@ -322,7 +322,7 @@ static void __init davinci_timer_init(void) clockevent_davinci.min_delta_ns = clockevent_delta2ns(1, &clockevent_davinci); - clockevent_davinci.cpumask = cpumask_of_cpu(0); + clockevent_davinci.cpumask = cpumask_of(0); clockevents_register_device(&clockevent_davinci); } diff --git a/arch/arm/mach-imx/time.c b/arch/arm/mach-imx/time.c index a11765f5f23b..aff0ebcfa847 100644 --- a/arch/arm/mach-imx/time.c +++ b/arch/arm/mach-imx/time.c @@ -184,7 +184,7 @@ static int __init imx_clockevent_init(unsigned long rate) clockevent_imx.min_delta_ns = clockevent_delta2ns(0xf, &clockevent_imx); - clockevent_imx.cpumask = cpumask_of_cpu(0); + clockevent_imx.cpumask = cpumask_of(0); clockevents_register_device(&clockevent_imx); diff --git a/arch/arm/mach-ixp4xx/common.c b/arch/arm/mach-ixp4xx/common.c index 7766f469456b..f4656d2ac8a8 100644 --- a/arch/arm/mach-ixp4xx/common.c +++ b/arch/arm/mach-ixp4xx/common.c @@ -487,7 +487,7 @@ static int __init ixp4xx_clockevent_init(void) clockevent_delta2ns(0xfffffffe, &clockevent_ixp4xx); clockevent_ixp4xx.min_delta_ns = clockevent_delta2ns(0xf, &clockevent_ixp4xx); - clockevent_ixp4xx.cpumask = cpumask_of_cpu(0); + clockevent_ixp4xx.cpumask = cpumask_of(0); clockevents_register_device(&clockevent_ixp4xx); return 0; diff --git a/arch/arm/mach-msm/timer.c b/arch/arm/mach-msm/timer.c index 345a14cb73c3..444d9c0f5ca6 100644 --- a/arch/arm/mach-msm/timer.c +++ b/arch/arm/mach-msm/timer.c @@ -182,7 +182,7 @@ static void __init msm_timer_init(void) clockevent_delta2ns(0xf0000000 >> clock->shift, ce); /* 4 gets rounded down to 3 */ ce->min_delta_ns = clockevent_delta2ns(4, ce); - ce->cpumask = cpumask_of_cpu(0); + ce->cpumask = cpumask_of(0); cs->mult = clocksource_hz2mult(clock->freq, cs->shift); res = clocksource_register(cs); diff --git a/arch/arm/mach-ns9xxx/time-ns9360.c b/arch/arm/mach-ns9xxx/time-ns9360.c index a63424d083d9..41df69721769 100644 --- a/arch/arm/mach-ns9xxx/time-ns9360.c +++ b/arch/arm/mach-ns9xxx/time-ns9360.c @@ -173,7 +173,7 @@ static void __init ns9360_timer_init(void) ns9360_clockevent_device.min_delta_ns = clockevent_delta2ns(1, &ns9360_clockevent_device); - ns9360_clockevent_device.cpumask = cpumask_of_cpu(0); + ns9360_clockevent_device.cpumask = cpumask_of(0); clockevents_register_device(&ns9360_clockevent_device); setup_irq(IRQ_NS9360_TIMER0 + TIMER_CLOCKEVENT, diff --git a/arch/arm/mach-omap1/time.c b/arch/arm/mach-omap1/time.c index 2cf7e32bd293..495a32c287b4 100644 --- a/arch/arm/mach-omap1/time.c +++ b/arch/arm/mach-omap1/time.c @@ -173,7 +173,7 @@ static __init void omap_init_mpu_timer(unsigned long rate) clockevent_mpu_timer1.min_delta_ns = clockevent_delta2ns(1, &clockevent_mpu_timer1); - clockevent_mpu_timer1.cpumask = cpumask_of_cpu(0); + clockevent_mpu_timer1.cpumask = cpumask_of(0); clockevents_register_device(&clockevent_mpu_timer1); } diff --git a/arch/arm/mach-omap1/timer32k.c b/arch/arm/mach-omap1/timer32k.c index 705367ece174..fd3f7396e162 100644 --- a/arch/arm/mach-omap1/timer32k.c +++ b/arch/arm/mach-omap1/timer32k.c @@ -187,7 +187,7 @@ static __init void omap_init_32k_timer(void) clockevent_32k_timer.min_delta_ns = clockevent_delta2ns(1, &clockevent_32k_timer); - clockevent_32k_timer.cpumask = cpumask_of_cpu(0); + clockevent_32k_timer.cpumask = cpumask_of(0); clockevents_register_device(&clockevent_32k_timer); } diff --git a/arch/arm/mach-omap2/timer-gp.c b/arch/arm/mach-omap2/timer-gp.c index 589393bedade..ae6036300f60 100644 --- a/arch/arm/mach-omap2/timer-gp.c +++ b/arch/arm/mach-omap2/timer-gp.c @@ -120,7 +120,7 @@ static void __init omap2_gp_clockevent_init(void) clockevent_gpt.min_delta_ns = clockevent_delta2ns(1, &clockevent_gpt); - clockevent_gpt.cpumask = cpumask_of_cpu(0); + clockevent_gpt.cpumask = cpumask_of(0); clockevents_register_device(&clockevent_gpt); } diff --git a/arch/arm/mach-pxa/time.c b/arch/arm/mach-pxa/time.c index f8a9a62959e5..bf3c9a4aad50 100644 --- a/arch/arm/mach-pxa/time.c +++ b/arch/arm/mach-pxa/time.c @@ -122,7 +122,6 @@ static struct clock_event_device ckevt_pxa_osmr0 = { .features = CLOCK_EVT_FEAT_ONESHOT, .shift = 32, .rating = 200, - .cpumask = CPU_MASK_CPU0, .set_next_event = pxa_osmr0_set_next_event, .set_mode = pxa_osmr0_set_mode, }; @@ -170,6 +169,7 @@ static void __init pxa_timer_init(void) clockevent_delta2ns(0x7fffffff, &ckevt_pxa_osmr0); ckevt_pxa_osmr0.min_delta_ns = clockevent_delta2ns(MIN_OSCR_DELTA * 2, &ckevt_pxa_osmr0) + 1; + ckevt_pxa_osmr0.cpumask = cpumask_of(0); cksrc_pxa_oscr0.mult = clocksource_hz2mult(clock_tick_rate, cksrc_pxa_oscr0.shift); diff --git a/arch/arm/mach-realview/core.c b/arch/arm/mach-realview/core.c index 2f04d54711e7..b07cb9b7adb1 100644 --- a/arch/arm/mach-realview/core.c +++ b/arch/arm/mach-realview/core.c @@ -511,7 +511,7 @@ static struct clock_event_device timer0_clockevent = { .set_mode = timer_set_mode, .set_next_event = timer_set_next_event, .rating = 300, - .cpumask = CPU_MASK_ALL, + .cpumask = cpu_all_mask, }; static void __init realview_clockevents_init(unsigned int timer_irq) diff --git a/arch/arm/mach-realview/localtimer.c b/arch/arm/mach-realview/localtimer.c index 44d178cd5733..504961ef343c 100644 --- a/arch/arm/mach-realview/localtimer.c +++ b/arch/arm/mach-realview/localtimer.c @@ -161,7 +161,7 @@ void __cpuinit local_timer_setup(unsigned int cpu) clk->set_mode = local_timer_set_mode; clk->set_next_event = local_timer_set_next_event; clk->irq = IRQ_LOCALTIMER; - clk->cpumask = cpumask_of_cpu(cpu); + clk->cpumask = cpumask_of(cpu); clk->shift = 20; clk->mult = div_sc(mpcore_timer_rate, NSEC_PER_SEC, clk->shift); clk->max_delta_ns = clockevent_delta2ns(0xffffffff, clk); @@ -199,7 +199,7 @@ void __cpuinit local_timer_setup(unsigned int cpu) clk->rating = 200; clk->set_mode = dummy_timer_set_mode; clk->broadcast = smp_timer_broadcast; - clk->cpumask = cpumask_of_cpu(cpu); + clk->cpumask = cpumask_of(cpu); clockevents_register_device(clk); } diff --git a/arch/arm/mach-sa1100/time.c b/arch/arm/mach-sa1100/time.c index 24c0a4bae850..1cac4ac0b4b8 100644 --- a/arch/arm/mach-sa1100/time.c +++ b/arch/arm/mach-sa1100/time.c @@ -73,7 +73,6 @@ static struct clock_event_device ckevt_sa1100_osmr0 = { .features = CLOCK_EVT_FEAT_ONESHOT, .shift = 32, .rating = 200, - .cpumask = CPU_MASK_CPU0, .set_next_event = sa1100_osmr0_set_next_event, .set_mode = sa1100_osmr0_set_mode, }; @@ -110,6 +109,7 @@ static void __init sa1100_timer_init(void) clockevent_delta2ns(0x7fffffff, &ckevt_sa1100_osmr0); ckevt_sa1100_osmr0.min_delta_ns = clockevent_delta2ns(MIN_OSCR_DELTA * 2, &ckevt_sa1100_osmr0) + 1; + ckevt_sa1100_osmr0.cpumask = cpumask_of(0); cksrc_sa1100_oscr.mult = clocksource_hz2mult(CLOCK_TICK_RATE, cksrc_sa1100_oscr.shift); diff --git a/arch/arm/mach-versatile/core.c b/arch/arm/mach-versatile/core.c index 565e0ba0d67e..a3f1933434e2 100644 --- a/arch/arm/mach-versatile/core.c +++ b/arch/arm/mach-versatile/core.c @@ -965,7 +965,7 @@ static void __init versatile_timer_init(void) timer0_clockevent.min_delta_ns = clockevent_delta2ns(0xf, &timer0_clockevent); - timer0_clockevent.cpumask = cpumask_of_cpu(0); + timer0_clockevent.cpumask = cpumask_of(0); clockevents_register_device(&timer0_clockevent); } diff --git a/arch/arm/plat-mxc/time.c b/arch/arm/plat-mxc/time.c index fd28f5194f71..758a1293bcfa 100644 --- a/arch/arm/plat-mxc/time.c +++ b/arch/arm/plat-mxc/time.c @@ -190,7 +190,7 @@ static int __init mxc_clockevent_init(void) clockevent_mxc.min_delta_ns = clockevent_delta2ns(0xff, &clockevent_mxc); - clockevent_mxc.cpumask = cpumask_of_cpu(0); + clockevent_mxc.cpumask = cpumask_of(0); clockevents_register_device(&clockevent_mxc); diff --git a/arch/arm/plat-orion/time.c b/arch/arm/plat-orion/time.c index 544d6b327f3a..6fa2923e6dca 100644 --- a/arch/arm/plat-orion/time.c +++ b/arch/arm/plat-orion/time.c @@ -149,7 +149,6 @@ static struct clock_event_device orion_clkevt = { .features = CLOCK_EVT_FEAT_ONESHOT | CLOCK_EVT_FEAT_PERIODIC, .shift = 32, .rating = 300, - .cpumask = CPU_MASK_CPU0, .set_next_event = orion_clkevt_next_event, .set_mode = orion_clkevt_mode, }; @@ -199,5 +198,6 @@ void __init orion_time_init(unsigned int irq, unsigned int tclk) orion_clkevt.mult = div_sc(tclk, NSEC_PER_SEC, orion_clkevt.shift); orion_clkevt.max_delta_ns = clockevent_delta2ns(0xfffffffe, &orion_clkevt); orion_clkevt.min_delta_ns = clockevent_delta2ns(1, &orion_clkevt); + orion_clkevt.cpumask = cpumask_of(0); clockevents_register_device(&orion_clkevt); } diff --git a/arch/avr32/kernel/time.c b/arch/avr32/kernel/time.c index 283481d74a5b..0ff46bf873b0 100644 --- a/arch/avr32/kernel/time.c +++ b/arch/avr32/kernel/time.c @@ -106,7 +106,6 @@ static struct clock_event_device comparator = { .features = CLOCK_EVT_FEAT_ONESHOT, .shift = 16, .rating = 50, - .cpumask = CPU_MASK_CPU0, .set_next_event = comparator_next_event, .set_mode = comparator_mode, }; @@ -134,6 +133,7 @@ void __init time_init(void) comparator.mult = div_sc(counter_hz, NSEC_PER_SEC, comparator.shift); comparator.max_delta_ns = clockevent_delta2ns((u32)~0, &comparator); comparator.min_delta_ns = clockevent_delta2ns(50, &comparator) + 1; + comparator.cpumask = cpumask_of(0); sysreg_write(COMPARE, 0); timer_irqaction.dev_id = &comparator; diff --git a/arch/blackfin/kernel/time-ts.c b/arch/blackfin/kernel/time-ts.c index e887efc86c29..0ed2badfd746 100644 --- a/arch/blackfin/kernel/time-ts.c +++ b/arch/blackfin/kernel/time-ts.c @@ -162,7 +162,6 @@ static struct clock_event_device clockevent_bfin = { .name = "bfin_core_timer", .features = CLOCK_EVT_FEAT_PERIODIC | CLOCK_EVT_FEAT_ONESHOT, .shift = 32, - .cpumask = CPU_MASK_CPU0, .set_next_event = bfin_timer_set_next_event, .set_mode = bfin_timer_set_mode, }; @@ -193,6 +192,7 @@ static int __init bfin_clockevent_init(void) clockevent_bfin.mult = div_sc(timer_clk, NSEC_PER_SEC, clockevent_bfin.shift); clockevent_bfin.max_delta_ns = clockevent_delta2ns(-1, &clockevent_bfin); clockevent_bfin.min_delta_ns = clockevent_delta2ns(100, &clockevent_bfin); + clockevent_bfin.cpumask = cpumask_of(0); clockevents_register_device(&clockevent_bfin); return 0; diff --git a/arch/m68knommu/platform/coldfire/pit.c b/arch/m68knommu/platform/coldfire/pit.c index c5b916700b22..2a12e7fa9748 100644 --- a/arch/m68knommu/platform/coldfire/pit.c +++ b/arch/m68knommu/platform/coldfire/pit.c @@ -156,7 +156,7 @@ void hw_timer_init(void) { u32 imr; - cf_pit_clockevent.cpumask = cpumask_of_cpu(smp_processor_id()); + cf_pit_clockevent.cpumask = cpumask_of(smp_processor_id()); cf_pit_clockevent.mult = div_sc(FREQ, NSEC_PER_SEC, 32); cf_pit_clockevent.max_delta_ns = clockevent_delta2ns(0xFFFF, &cf_pit_clockevent); diff --git a/arch/mips/jazz/irq.c b/arch/mips/jazz/irq.c index d7f8a782aae4..03965cb1b252 100644 --- a/arch/mips/jazz/irq.c +++ b/arch/mips/jazz/irq.c @@ -146,7 +146,7 @@ void __init plat_time_init(void) BUG_ON(HZ != 100); - cd->cpumask = cpumask_of_cpu(cpu); + cd->cpumask = cpumask_of(cpu); clockevents_register_device(cd); action->dev_id = cd; setup_irq(JAZZ_TIMER_IRQ, action); diff --git a/arch/mips/kernel/cevt-bcm1480.c b/arch/mips/kernel/cevt-bcm1480.c index d7e21bc8cd21..b820661678b0 100644 --- a/arch/mips/kernel/cevt-bcm1480.c +++ b/arch/mips/kernel/cevt-bcm1480.c @@ -126,7 +126,7 @@ void __cpuinit sb1480_clockevent_init(void) cd->min_delta_ns = clockevent_delta2ns(2, cd); cd->rating = 200; cd->irq = irq; - cd->cpumask = cpumask_of_cpu(cpu); + cd->cpumask = cpumask_of(cpu); cd->set_next_event = sibyte_next_event; cd->set_mode = sibyte_set_mode; clockevents_register_device(cd); diff --git a/arch/mips/kernel/cevt-ds1287.c b/arch/mips/kernel/cevt-ds1287.c index df4acb68bfb5..1ada45ea0700 100644 --- a/arch/mips/kernel/cevt-ds1287.c +++ b/arch/mips/kernel/cevt-ds1287.c @@ -88,7 +88,6 @@ static void ds1287_event_handler(struct clock_event_device *dev) static struct clock_event_device ds1287_clockevent = { .name = "ds1287", .features = CLOCK_EVT_FEAT_PERIODIC, - .cpumask = CPU_MASK_CPU0, .set_next_event = ds1287_set_next_event, .set_mode = ds1287_set_mode, .event_handler = ds1287_event_handler, @@ -122,6 +121,7 @@ int __init ds1287_clockevent_init(int irq) clockevent_set_clock(cd, 32768); cd->max_delta_ns = clockevent_delta2ns(0x7fffffff, cd); cd->min_delta_ns = clockevent_delta2ns(0x300, cd); + cd->cpumask = cpumask_of(0); clockevents_register_device(&ds1287_clockevent); diff --git a/arch/mips/kernel/cevt-gt641xx.c b/arch/mips/kernel/cevt-gt641xx.c index 6e2f58520afb..e9b787feedcb 100644 --- a/arch/mips/kernel/cevt-gt641xx.c +++ b/arch/mips/kernel/cevt-gt641xx.c @@ -96,7 +96,6 @@ static void gt641xx_timer0_event_handler(struct clock_event_device *dev) static struct clock_event_device gt641xx_timer0_clockevent = { .name = "gt641xx-timer0", .features = CLOCK_EVT_FEAT_PERIODIC | CLOCK_EVT_FEAT_ONESHOT, - .cpumask = CPU_MASK_CPU0, .irq = GT641XX_TIMER0_IRQ, .set_next_event = gt641xx_timer0_set_next_event, .set_mode = gt641xx_timer0_set_mode, @@ -132,6 +131,7 @@ static int __init gt641xx_timer0_clockevent_init(void) clockevent_set_clock(cd, gt641xx_base_clock); cd->max_delta_ns = clockevent_delta2ns(0x7fffffff, cd); cd->min_delta_ns = clockevent_delta2ns(0x300, cd); + cd->cpumask = cpumask_of(0); clockevents_register_device(>641xx_timer0_clockevent); diff --git a/arch/mips/kernel/cevt-r4k.c b/arch/mips/kernel/cevt-r4k.c index 4a4c59f2737a..e1ec83b68031 100644 --- a/arch/mips/kernel/cevt-r4k.c +++ b/arch/mips/kernel/cevt-r4k.c @@ -195,7 +195,7 @@ int __cpuinit mips_clockevent_init(void) cd->rating = 300; cd->irq = irq; - cd->cpumask = cpumask_of_cpu(cpu); + cd->cpumask = cpumask_of(cpu); cd->set_next_event = mips_next_event; cd->set_mode = mips_set_clock_mode; cd->event_handler = mips_event_handler; diff --git a/arch/mips/kernel/cevt-sb1250.c b/arch/mips/kernel/cevt-sb1250.c index 0f188cd46e03..a2eebaafda52 100644 --- a/arch/mips/kernel/cevt-sb1250.c +++ b/arch/mips/kernel/cevt-sb1250.c @@ -125,7 +125,7 @@ void __cpuinit sb1250_clockevent_init(void) cd->min_delta_ns = clockevent_delta2ns(2, cd); cd->rating = 200; cd->irq = irq; - cd->cpumask = cpumask_of_cpu(cpu); + cd->cpumask = cpumask_of(cpu); cd->set_next_event = sibyte_next_event; cd->set_mode = sibyte_set_mode; clockevents_register_device(cd); diff --git a/arch/mips/kernel/cevt-smtc.c b/arch/mips/kernel/cevt-smtc.c index 5162fe4b5952..6d45e24db5bf 100644 --- a/arch/mips/kernel/cevt-smtc.c +++ b/arch/mips/kernel/cevt-smtc.c @@ -292,7 +292,7 @@ int __cpuinit mips_clockevent_init(void) cd->rating = 300; cd->irq = irq; - cd->cpumask = cpumask_of_cpu(cpu); + cd->cpumask = cpumask_of(cpu); cd->set_next_event = mips_next_event; cd->set_mode = mips_set_clock_mode; cd->event_handler = mips_event_handler; diff --git a/arch/mips/kernel/cevt-txx9.c b/arch/mips/kernel/cevt-txx9.c index b5fc4eb412d2..eccf7d6096bd 100644 --- a/arch/mips/kernel/cevt-txx9.c +++ b/arch/mips/kernel/cevt-txx9.c @@ -112,7 +112,6 @@ static struct clock_event_device txx9tmr_clock_event_device = { .name = "TXx9", .features = CLOCK_EVT_FEAT_PERIODIC | CLOCK_EVT_FEAT_ONESHOT, .rating = 200, - .cpumask = CPU_MASK_CPU0, .set_mode = txx9tmr_set_mode, .set_next_event = txx9tmr_set_next_event, }; @@ -150,6 +149,7 @@ void __init txx9_clockevent_init(unsigned long baseaddr, int irq, clockevent_delta2ns(0xffffffff >> (32 - TXX9_TIMER_BITS), cd); cd->min_delta_ns = clockevent_delta2ns(0xf, cd); cd->irq = irq; + cd->cpumask = cpumask_of(0), clockevents_register_device(cd); setup_irq(irq, &txx9tmr_irq); printk(KERN_INFO "TXx9: clockevent device at 0x%lx, irq %d\n", diff --git a/arch/mips/kernel/i8253.c b/arch/mips/kernel/i8253.c index b6ac55162b9a..f4d187825f96 100644 --- a/arch/mips/kernel/i8253.c +++ b/arch/mips/kernel/i8253.c @@ -115,7 +115,7 @@ void __init setup_pit_timer(void) * Start pit with the boot cpu mask and make it global after the * IO_APIC has been initialized. */ - cd->cpumask = cpumask_of_cpu(cpu); + cd->cpumask = cpumask_of(cpu); clockevent_set_clock(cd, CLOCK_TICK_RATE); cd->max_delta_ns = clockevent_delta2ns(0x7FFF, cd); cd->min_delta_ns = clockevent_delta2ns(0xF, cd); diff --git a/arch/mips/nxp/pnx8550/common/time.c b/arch/mips/nxp/pnx8550/common/time.c index 62f495b57f93..cf293b279098 100644 --- a/arch/mips/nxp/pnx8550/common/time.c +++ b/arch/mips/nxp/pnx8550/common/time.c @@ -102,6 +102,7 @@ __init void plat_time_init(void) unsigned int p; unsigned int pow2p; + pnx8xxx_clockevent.cpumask = cpu_none_mask; clockevents_register_device(&pnx8xxx_clockevent); clocksource_register(&pnx_clocksource); diff --git a/arch/mips/sgi-ip27/ip27-timer.c b/arch/mips/sgi-ip27/ip27-timer.c index 1327c2746fb7..f024057a35f8 100644 --- a/arch/mips/sgi-ip27/ip27-timer.c +++ b/arch/mips/sgi-ip27/ip27-timer.c @@ -134,7 +134,7 @@ void __cpuinit hub_rt_clock_event_init(void) cd->min_delta_ns = clockevent_delta2ns(0x300, cd); cd->rating = 200; cd->irq = irq; - cd->cpumask = cpumask_of_cpu(cpu); + cd->cpumask = cpumask_of(cpu); cd->set_next_event = rt_next_event; cd->set_mode = rt_set_mode; clockevents_register_device(cd); diff --git a/arch/mips/sni/time.c b/arch/mips/sni/time.c index 796e3ce28720..69f5f88711cc 100644 --- a/arch/mips/sni/time.c +++ b/arch/mips/sni/time.c @@ -80,7 +80,7 @@ static void __init sni_a20r_timer_setup(void) struct irqaction *action = &a20r_irqaction; unsigned int cpu = smp_processor_id(); - cd->cpumask = cpumask_of_cpu(cpu); + cd->cpumask = cpumask_of(cpu); clockevents_register_device(cd); action->dev_id = cd; setup_irq(SNI_A20R_IRQ_TIMER, &a20r_irqaction); diff --git a/arch/powerpc/kernel/time.c b/arch/powerpc/kernel/time.c index e2ee66b5831d..6f39d35d6f55 100644 --- a/arch/powerpc/kernel/time.c +++ b/arch/powerpc/kernel/time.c @@ -869,7 +869,7 @@ static void register_decrementer_clockevent(int cpu) struct clock_event_device *dec = &per_cpu(decrementers, cpu).event; *dec = decrementer_clockevent; - dec->cpumask = cpumask_of_cpu(cpu); + dec->cpumask = cpumask_of(cpu); printk(KERN_DEBUG "clockevent: %s mult[%lx] shift[%d] cpu[%d]\n", dec->name, dec->mult, dec->shift, cpu); diff --git a/arch/s390/kernel/time.c b/arch/s390/kernel/time.c index eccefbbff887..f5bd141c8443 100644 --- a/arch/s390/kernel/time.c +++ b/arch/s390/kernel/time.c @@ -154,7 +154,7 @@ void init_cpu_timer(void) cd->min_delta_ns = 1; cd->max_delta_ns = LONG_MAX; cd->rating = 400; - cd->cpumask = cpumask_of_cpu(cpu); + cd->cpumask = cpumask_of(cpu); cd->set_next_event = s390_next_event; cd->set_mode = s390_set_mode; diff --git a/arch/sh/include/asm/smp.h b/arch/sh/include/asm/smp.h index 85b660c17eb0..c24e9c6a1736 100644 --- a/arch/sh/include/asm/smp.h +++ b/arch/sh/include/asm/smp.h @@ -31,7 +31,7 @@ enum { }; void smp_message_recv(unsigned int msg); -void smp_timer_broadcast(cpumask_t mask); +void smp_timer_broadcast(const struct cpumask *mask); void local_timer_interrupt(void); void local_timer_setup(unsigned int cpu); diff --git a/arch/sh/kernel/smp.c b/arch/sh/kernel/smp.c index 593937d0c495..8f4027412614 100644 --- a/arch/sh/kernel/smp.c +++ b/arch/sh/kernel/smp.c @@ -184,11 +184,11 @@ void arch_send_call_function_single_ipi(int cpu) plat_send_ipi(cpu, SMP_MSG_FUNCTION_SINGLE); } -void smp_timer_broadcast(cpumask_t mask) +void smp_timer_broadcast(const struct cpumask *mask) { int cpu; - for_each_cpu_mask(cpu, mask) + for_each_cpu(cpu, mask) plat_send_ipi(cpu, SMP_MSG_TIMER); } diff --git a/arch/sh/kernel/timers/timer-broadcast.c b/arch/sh/kernel/timers/timer-broadcast.c index c2317635230f..96e8eaea1e62 100644 --- a/arch/sh/kernel/timers/timer-broadcast.c +++ b/arch/sh/kernel/timers/timer-broadcast.c @@ -51,7 +51,7 @@ void __cpuinit local_timer_setup(unsigned int cpu) clk->mult = 1; clk->set_mode = dummy_timer_set_mode; clk->broadcast = smp_timer_broadcast; - clk->cpumask = cpumask_of_cpu(cpu); + clk->cpumask = cpumask_of(cpu); clockevents_register_device(clk); } diff --git a/arch/sh/kernel/timers/timer-tmu.c b/arch/sh/kernel/timers/timer-tmu.c index 3c61ddd4d43e..0db3f9510336 100644 --- a/arch/sh/kernel/timers/timer-tmu.c +++ b/arch/sh/kernel/timers/timer-tmu.c @@ -263,7 +263,7 @@ static int tmu_timer_init(void) tmu0_clockevent.min_delta_ns = clockevent_delta2ns(1, &tmu0_clockevent); - tmu0_clockevent.cpumask = cpumask_of_cpu(0); + tmu0_clockevent.cpumask = cpumask_of(0); clockevents_register_device(&tmu0_clockevent); diff --git a/arch/sparc64/kernel/time.c b/arch/sparc64/kernel/time.c index 141da3759091..9df8f095a8b1 100644 --- a/arch/sparc64/kernel/time.c +++ b/arch/sparc64/kernel/time.c @@ -763,7 +763,7 @@ void __devinit setup_sparc64_timer(void) sevt = &__get_cpu_var(sparc64_events); memcpy(sevt, &sparc64_clockevent, sizeof(*sevt)); - sevt->cpumask = cpumask_of_cpu(smp_processor_id()); + sevt->cpumask = cpumask_of(smp_processor_id()); clockevents_register_device(sevt); } diff --git a/arch/um/kernel/time.c b/arch/um/kernel/time.c index 47f04f4a3464..b13a87a3ec95 100644 --- a/arch/um/kernel/time.c +++ b/arch/um/kernel/time.c @@ -50,7 +50,7 @@ static int itimer_next_event(unsigned long delta, static struct clock_event_device itimer_clockevent = { .name = "itimer", .rating = 250, - .cpumask = CPU_MASK_ALL, + .cpumask = cpu_all_mask, .features = CLOCK_EVT_FEAT_PERIODIC | CLOCK_EVT_FEAT_ONESHOT, .set_mode = itimer_set_mode, .set_next_event = itimer_next_event, diff --git a/arch/x86/kernel/apic.c b/arch/x86/kernel/apic.c index 16f94879b525..b2cef49f3085 100644 --- a/arch/x86/kernel/apic.c +++ b/arch/x86/kernel/apic.c @@ -141,7 +141,7 @@ static int lapic_next_event(unsigned long delta, struct clock_event_device *evt); static void lapic_timer_setup(enum clock_event_mode mode, struct clock_event_device *evt); -static void lapic_timer_broadcast(cpumask_t mask); +static void lapic_timer_broadcast(const struct cpumask *mask); static void apic_pm_activate(void); /* @@ -453,10 +453,10 @@ static void lapic_timer_setup(enum clock_event_mode mode, /* * Local APIC timer broadcast function */ -static void lapic_timer_broadcast(cpumask_t mask) +static void lapic_timer_broadcast(const struct cpumask *mask) { #ifdef CONFIG_SMP - send_IPI_mask(mask, LOCAL_TIMER_VECTOR); + send_IPI_mask(*mask, LOCAL_TIMER_VECTOR); #endif } @@ -469,7 +469,7 @@ static void __cpuinit setup_APIC_timer(void) struct clock_event_device *levt = &__get_cpu_var(lapic_events); memcpy(levt, &lapic_clockevent, sizeof(*levt)); - levt->cpumask = cpumask_of_cpu(smp_processor_id()); + levt->cpumask = cpumask_of(smp_processor_id()); clockevents_register_device(levt); } diff --git a/arch/x86/kernel/hpet.c b/arch/x86/kernel/hpet.c index 940f25851e1e..e76d7e272974 100644 --- a/arch/x86/kernel/hpet.c +++ b/arch/x86/kernel/hpet.c @@ -246,7 +246,7 @@ static void hpet_legacy_clockevent_register(void) * Start hpet with the boot cpu mask and make it * global after the IO_APIC has been initialized. */ - hpet_clockevent.cpumask = cpumask_of_cpu(smp_processor_id()); + hpet_clockevent.cpumask = cpumask_of(smp_processor_id()); clockevents_register_device(&hpet_clockevent); global_clock_event = &hpet_clockevent; printk(KERN_DEBUG "hpet clockevent registered\n"); @@ -500,7 +500,7 @@ static void init_one_hpet_msi_clockevent(struct hpet_dev *hdev, int cpu) /* 5 usec minimum reprogramming delta. */ evt->min_delta_ns = 5000; - evt->cpumask = cpumask_of_cpu(hdev->cpu); + evt->cpumask = cpumask_of(hdev->cpu); clockevents_register_device(evt); } diff --git a/arch/x86/kernel/i8253.c b/arch/x86/kernel/i8253.c index c1b5e3ece1f2..10f92fb532f3 100644 --- a/arch/x86/kernel/i8253.c +++ b/arch/x86/kernel/i8253.c @@ -114,7 +114,7 @@ void __init setup_pit_timer(void) * Start pit with the boot cpu mask and make it global after the * IO_APIC has been initialized. */ - pit_clockevent.cpumask = cpumask_of_cpu(smp_processor_id()); + pit_clockevent.cpumask = cpumask_of(smp_processor_id()); pit_clockevent.mult = div_sc(CLOCK_TICK_RATE, NSEC_PER_SEC, pit_clockevent.shift); pit_clockevent.max_delta_ns = diff --git a/arch/x86/kernel/mfgpt_32.c b/arch/x86/kernel/mfgpt_32.c index 3b599518c322..c12314c9e86f 100644 --- a/arch/x86/kernel/mfgpt_32.c +++ b/arch/x86/kernel/mfgpt_32.c @@ -287,7 +287,7 @@ static struct clock_event_device mfgpt_clockevent = { .set_mode = mfgpt_set_mode, .set_next_event = mfgpt_next_event, .rating = 250, - .cpumask = CPU_MASK_ALL, + .cpumask = cpu_all_mask, .shift = 32 }; diff --git a/arch/x86/kernel/vmiclock_32.c b/arch/x86/kernel/vmiclock_32.c index 254ee07f8635..c4c1f9e09402 100644 --- a/arch/x86/kernel/vmiclock_32.c +++ b/arch/x86/kernel/vmiclock_32.c @@ -226,7 +226,7 @@ static void __devinit vmi_time_init_clockevent(void) /* Upper bound is clockevent's use of ulong for cycle deltas. */ evt->max_delta_ns = clockevent_delta2ns(ULONG_MAX, evt); evt->min_delta_ns = clockevent_delta2ns(1, evt); - evt->cpumask = cpumask_of_cpu(cpu); + evt->cpumask = cpumask_of(cpu); printk(KERN_WARNING "vmi: registering clock event %s. mult=%lu shift=%u\n", evt->name, evt->mult, evt->shift); diff --git a/arch/x86/lguest/boot.c b/arch/x86/lguest/boot.c index a5d8e1ace1cf..104c8220a383 100644 --- a/arch/x86/lguest/boot.c +++ b/arch/x86/lguest/boot.c @@ -737,7 +737,7 @@ static void lguest_time_init(void) /* We can't set cpumask in the initializer: damn C limitations! Set it * here and register our timer device. */ - lguest_clockevent.cpumask = cpumask_of_cpu(0); + lguest_clockevent.cpumask = cpumask_of(0); clockevents_register_device(&lguest_clockevent); /* Finally, we unblock the timer interrupt. */ diff --git a/arch/x86/xen/time.c b/arch/x86/xen/time.c index c9f7cda48ed7..65d75a6be0ba 100644 --- a/arch/x86/xen/time.c +++ b/arch/x86/xen/time.c @@ -437,7 +437,7 @@ void xen_setup_timer(int cpu) evt = &per_cpu(xen_clock_events, cpu); memcpy(evt, xen_clockevent, sizeof(*evt)); - evt->cpumask = cpumask_of_cpu(cpu); + evt->cpumask = cpumask_of(cpu); evt->irq = irq; setup_runstate_info(cpu); diff --git a/drivers/clocksource/tcb_clksrc.c b/drivers/clocksource/tcb_clksrc.c index f450588e5858..254f1064d973 100644 --- a/drivers/clocksource/tcb_clksrc.c +++ b/drivers/clocksource/tcb_clksrc.c @@ -154,7 +154,6 @@ static struct tc_clkevt_device clkevt = { .shift = 32, /* Should be lower than at91rm9200's system timer */ .rating = 125, - .cpumask = CPU_MASK_CPU0, .set_next_event = tc_next_event, .set_mode = tc_mode, }, @@ -195,6 +194,7 @@ static void __init setup_clkevents(struct atmel_tc *tc, int clk32k_divisor_idx) clkevt.clkevt.max_delta_ns = clockevent_delta2ns(0xffff, &clkevt.clkevt); clkevt.clkevt.min_delta_ns = clockevent_delta2ns(1, &clkevt.clkevt) + 1; + clkevt.clkevt.cpumask = cpumask_of(0); setup_irq(irq, &tc_irqaction); diff --git a/include/linux/clockchips.h b/include/linux/clockchips.h index ed3a5d473e52..cea153697ec7 100644 --- a/include/linux/clockchips.h +++ b/include/linux/clockchips.h @@ -82,13 +82,13 @@ struct clock_event_device { int shift; int rating; int irq; - cpumask_t cpumask; + const struct cpumask *cpumask; int (*set_next_event)(unsigned long evt, struct clock_event_device *); void (*set_mode)(enum clock_event_mode mode, struct clock_event_device *); void (*event_handler)(struct clock_event_device *); - void (*broadcast)(cpumask_t mask); + void (*broadcast)(const struct cpumask *mask); struct list_head list; enum clock_event_mode mode; ktime_t next_event; diff --git a/kernel/time/clockevents.c b/kernel/time/clockevents.c index f8d968063cea..ea2f48af83cf 100644 --- a/kernel/time/clockevents.c +++ b/kernel/time/clockevents.c @@ -166,6 +166,8 @@ static void clockevents_notify_released(void) void clockevents_register_device(struct clock_event_device *dev) { BUG_ON(dev->mode != CLOCK_EVT_MODE_UNUSED); + BUG_ON(!dev->cpumask); + /* * A nsec2cyc multiplicator of 0 is invalid and we'd crash * on it, so fix it up and emit a warning: diff --git a/kernel/time/tick-broadcast.c b/kernel/time/tick-broadcast.c index f98a1b7b16e9..9590af2327be 100644 --- a/kernel/time/tick-broadcast.c +++ b/kernel/time/tick-broadcast.c @@ -150,7 +150,7 @@ static void tick_do_broadcast(cpumask_t mask) */ cpu = first_cpu(mask); td = &per_cpu(tick_cpu_device, cpu); - td->evtdev->broadcast(mask); + td->evtdev->broadcast(&mask); } } diff --git a/kernel/time/tick-common.c b/kernel/time/tick-common.c index ab65d217583f..f8372be74122 100644 --- a/kernel/time/tick-common.c +++ b/kernel/time/tick-common.c @@ -171,7 +171,7 @@ static void tick_setup_device(struct tick_device *td, * When the device is not per cpu, pin the interrupt to the * current cpu: */ - if (!cpumask_equal(&newdev->cpumask, cpumask)) + if (!cpumask_equal(newdev->cpumask, cpumask)) irq_set_affinity(newdev->irq, cpumask); /* @@ -202,14 +202,14 @@ static int tick_check_new_device(struct clock_event_device *newdev) spin_lock_irqsave(&tick_device_lock, flags); cpu = smp_processor_id(); - if (!cpu_isset(cpu, newdev->cpumask)) + if (!cpumask_test_cpu(cpu, newdev->cpumask)) goto out_bc; td = &per_cpu(tick_cpu_device, cpu); curdev = td->evtdev; /* cpu local device ? */ - if (!cpus_equal(newdev->cpumask, cpumask_of_cpu(cpu))) { + if (!cpumask_equal(newdev->cpumask, cpumask_of(cpu))) { /* * If the cpu affinity of the device interrupt can not @@ -222,7 +222,7 @@ static int tick_check_new_device(struct clock_event_device *newdev) * If we have a cpu local device already, do not replace it * by a non cpu local device */ - if (curdev && cpus_equal(curdev->cpumask, cpumask_of_cpu(cpu))) + if (curdev && cpumask_equal(curdev->cpumask, cpumask_of(cpu))) goto out_bc; } -- cgit v1.2.3 From ae8d04e2ecbb233926860e9ce145eac19c7835dc Mon Sep 17 00:00:00 2001 From: Zachary Amsden Date: Sat, 13 Dec 2008 12:36:58 -0800 Subject: x86 Fix VMI crash on boot in 2.6.28-rc8 VMI initialiation can relocate the fixmap, causing early_ioremap to malfunction if it is initialized before the relocation. To fix this, VMI activation is split into two phases; the detection, which must happen before setting up ioremap, and the activation, which must happen after parsing early boot parameters. This fixes a crash on boot when VMI is enabled under VMware. Signed-off-by: Zachary Amsden Signed-off-by: Linus Torvalds --- arch/x86/include/asm/vmi.h | 8 +++++++- arch/x86/kernel/setup.c | 12 +++++------- arch/x86/kernel/smpboot.c | 2 -- arch/x86/kernel/vmi_32.c | 16 +++++++++++----- 4 files changed, 23 insertions(+), 15 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/vmi.h b/arch/x86/include/asm/vmi.h index b7c0dea119fe..61e08c0a2907 100644 --- a/arch/x86/include/asm/vmi.h +++ b/arch/x86/include/asm/vmi.h @@ -223,9 +223,15 @@ struct pci_header { } __attribute__((packed)); /* Function prototypes for bootstrapping */ +#ifdef CONFIG_VMI extern void vmi_init(void); +extern void vmi_activate(void); extern void vmi_bringup(void); -extern void vmi_apply_boot_page_allocations(void); +#else +static inline void vmi_init(void) {} +static inline void vmi_activate(void) {} +static inline void vmi_bringup(void) {} +#endif /* State needed to start an application processor in an SMP system. */ struct vmi_ap_state { diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index 9d5674f7b6cc..bdec76e55594 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -794,6 +794,9 @@ void __init setup_arch(char **cmdline_p) printk(KERN_INFO "Command line: %s\n", boot_command_line); #endif + /* VMI may relocate the fixmap; do this before touching ioremap area */ + vmi_init(); + early_cpu_init(); early_ioremap_init(); @@ -880,13 +883,8 @@ void __init setup_arch(char **cmdline_p) check_efer(); #endif -#if defined(CONFIG_VMI) && defined(CONFIG_X86_32) - /* - * Must be before kernel pagetables are setup - * or fixmap area is touched. - */ - vmi_init(); -#endif + /* Must be before kernel pagetables are setup */ + vmi_activate(); /* after early param, so could get panic from serial */ reserve_early_setup_data(); diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index 7b1093397319..f71f96fc9e62 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -294,9 +294,7 @@ static void __cpuinit start_secondary(void *unused) * fragile that we want to limit the things done here to the * most necessary things. */ -#ifdef CONFIG_VMI vmi_bringup(); -#endif cpu_init(); preempt_disable(); smp_callin(); diff --git a/arch/x86/kernel/vmi_32.c b/arch/x86/kernel/vmi_32.c index 8b6c393ab9fd..22fd6577156a 100644 --- a/arch/x86/kernel/vmi_32.c +++ b/arch/x86/kernel/vmi_32.c @@ -960,8 +960,6 @@ static inline int __init activate_vmi(void) void __init vmi_init(void) { - unsigned long flags; - if (!vmi_rom) probe_vmi_rom(); else @@ -973,13 +971,21 @@ void __init vmi_init(void) reserve_top_address(-vmi_rom->virtual_top); - local_irq_save(flags); - activate_vmi(); - #ifdef CONFIG_X86_IO_APIC /* This is virtual hardware; timer routing is wired correctly */ no_timer_check = 1; #endif +} + +void vmi_activate(void) +{ + unsigned long flags; + + if (!vmi_rom) + return; + + local_irq_save(flags); + activate_vmi(); local_irq_restore(flags & X86_EFLAGS_IF); } -- cgit v1.2.3 From 205516c12dbba003c26b42cfb41e598631300106 Mon Sep 17 00:00:00 2001 From: Ken Chen Date: Tue, 16 Dec 2008 00:32:21 -0800 Subject: x86: convert rdtscll() to use __native_read_tsc Impact: micro-optimization Is there any reason why x86 rdtscll have to use the out of line function instead of inline __native_read_tsc()? native_read_tsc and __native_read_tsc is essentially the same functions. Patch to let x86 rdtscll() to use the inline version of read_tsc. Signed-off-by: Ken Chen Signed-off-by: Ingo Molnar --- arch/x86/include/asm/msr.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/msr.h b/arch/x86/include/asm/msr.h index c2a812ebde89..42f639b991b4 100644 --- a/arch/x86/include/asm/msr.h +++ b/arch/x86/include/asm/msr.h @@ -181,10 +181,10 @@ static inline int rdmsrl_amd_safe(unsigned msr, unsigned long long *p) } #define rdtscl(low) \ - ((low) = (u32)native_read_tsc()) + ((low) = (u32)__native_read_tsc()) #define rdtscll(val) \ - ((val) = native_read_tsc()) + ((val) = __native_read_tsc()) #define rdpmc(counter, low, high) \ do { \ -- cgit v1.2.3 From cc1dc6d039ced64c2f8b8457bf1cccf4ecfc5942 Mon Sep 17 00:00:00 2001 From: Markus Metzger Date: Tue, 16 Dec 2008 15:51:03 +0100 Subject: x86, bts: remove recursion from get_context Impact: cleanup Optimistically allocate a DS context. It is extremely unlikely that one already existed. This simplifies the code a lot. Signed-off-by: Markus Metzger Signed-off-by: Ingo Molnar --- arch/x86/kernel/ds.c | 58 ++++++++++++++++++++++------------------------------ 1 file changed, 25 insertions(+), 33 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/ds.c b/arch/x86/kernel/ds.c index dc1e7123ea4e..0dc795951d73 100644 --- a/arch/x86/kernel/ds.c +++ b/arch/x86/kernel/ds.c @@ -232,53 +232,45 @@ static DEFINE_PER_CPU(struct ds_context *, system_context_array); #define system_context per_cpu(system_context_array, smp_processor_id()) -static struct ds_context *ds_get_context(struct task_struct *task) + +static inline struct ds_context *ds_get_context(struct task_struct *task) { struct ds_context **p_context = (task ? &task->thread.ds_ctx : &system_context); - struct ds_context *context = *p_context; + struct ds_context *context = NULL; + struct ds_context *new_context = NULL; unsigned long irq; - if (!context) { - context = kzalloc(sizeof(*context), GFP_KERNEL); - if (!context) - return NULL; - - spin_lock_irqsave(&ds_lock, irq); - - if (*p_context) { - kfree(context); + /* Chances are small that we already have a context. */ + new_context = kzalloc(sizeof(*new_context), GFP_KERNEL); + if (!new_context) + return NULL; - context = *p_context; - } else { - *p_context = context; + spin_lock_irqsave(&ds_lock, irq); - context->this = p_context; - context->task = task; + context = *p_context; + if (!context) { + context = new_context; - if (task) - set_tsk_thread_flag(task, TIF_DS_AREA_MSR); + context->this = p_context; + context->task = task; + context->count = 0; - if (!task || (task == current)) - wrmsrl(MSR_IA32_DS_AREA, - (unsigned long)context->ds); - } + if (task) + set_tsk_thread_flag(task, TIF_DS_AREA_MSR); - context->count++; + if (!task || (task == current)) + wrmsrl(MSR_IA32_DS_AREA, (unsigned long)context->ds); - spin_unlock_irqrestore(&ds_lock, irq); - } else { - spin_lock_irqsave(&ds_lock, irq); + *p_context = context; + } - context = *p_context; - if (context) - context->count++; + context->count++; - spin_unlock_irqrestore(&ds_lock, irq); + spin_unlock_irqrestore(&ds_lock, irq); - if (!context) - context = ds_get_context(task); - } + if (context != new_context) + kfree(new_context); return context; } -- cgit v1.2.3 From d072c25f531c6513994960401d2c7f059434c0d2 Mon Sep 17 00:00:00 2001 From: Markus Metzger Date: Tue, 16 Dec 2008 15:53:11 +0100 Subject: x86, bts: correctly report invalid bts records Impact: change the reporting of empty BTS records Correctly report a cleared BTS record as invalid. Used to be reported as branch from 0 to 0. Signed-off-by: Markus Metzger Signed-off-by: Ingo Molnar --- arch/x86/kernel/ds.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/ds.c b/arch/x86/kernel/ds.c index 0dc795951d73..98d271e60e08 100644 --- a/arch/x86/kernel/ds.c +++ b/arch/x86/kernel/ds.c @@ -484,6 +484,9 @@ static int bts_read(struct bts_tracer *tracer, const void *at, out->qualifier = bts_branch; out->variant.lbr.from = bts_get(at, bts_from); out->variant.lbr.to = bts_get(at, bts_to); + + if (!out->variant.lbr.from && !out->variant.lbr.to) + out->qualifier = bts_invalid; } return ds_cfg.sizeof_rec[ds_bts]; -- cgit v1.2.3 From 1796316a8b028a148be48ba5d4e7be493a39d173 Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Tue, 16 Dec 2008 11:35:24 +0000 Subject: x86: consolidate __swp_XXX() macros Impact: cleanup, code robustization The __swp_...() macros silently relied upon which bits are used for _PAGE_FILE and _PAGE_PROTNONE. After having changed _PAGE_PROTNONE in our Xen kernel to no longer overlap _PAGE_PAT, live locks and crashes were reported that could have been avoided if these macros properly used the symbolic constants. Since, as pointed out earlier, for Xen Dom0 support mainline likewise will need to eliminate the conflict between _PAGE_PAT and _PAGE_PROTNONE, this patch does all the necessary adjustments, plus it introduces a mechanism to check consistency between MAX_SWAPFILES_SHIFT and the actual encoding macros. This also fixes a latent bug in that x86-64 used a 6-bit mask in __swp_type(), and if MAX_SWAPFILES_SHIFT was increased beyond 5 in (the seemingly unrelated) linux/swap.h, this would have resulted in a collision with _PAGE_FILE. Non-PAE 32-bit code gets similarly adjusted for its pte_to_pgoff() and pgoff_to_pte() calculations. Signed-off-by: Jan Beulich Signed-off-by: Ingo Molnar --- arch/x86/include/asm/pgtable-2level.h | 50 ++++++++++++++++++++++++++++------- arch/x86/include/asm/pgtable-3level.h | 1 + arch/x86/include/asm/pgtable.h | 14 +++++----- arch/x86/include/asm/pgtable_64.h | 20 +++++++++++--- mm/swapfile.c | 9 +++++++ 5 files changed, 75 insertions(+), 19 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/pgtable-2level.h b/arch/x86/include/asm/pgtable-2level.h index b17edfd23628..e0d199fe1d83 100644 --- a/arch/x86/include/asm/pgtable-2level.h +++ b/arch/x86/include/asm/pgtable-2level.h @@ -56,23 +56,55 @@ static inline pte_t native_ptep_get_and_clear(pte_t *xp) #define pte_none(x) (!(x).pte_low) /* - * Bits 0, 6 and 7 are taken, split up the 29 bits of offset - * into this range: + * Bits _PAGE_BIT_PRESENT, _PAGE_BIT_FILE and _PAGE_BIT_PROTNONE are taken, + * split up the 29 bits of offset into this range: */ #define PTE_FILE_MAX_BITS 29 +#define PTE_FILE_SHIFT1 (_PAGE_BIT_PRESENT + 1) +#if _PAGE_BIT_FILE < _PAGE_BIT_PROTNONE +#define PTE_FILE_SHIFT2 (_PAGE_BIT_FILE + 1) +#define PTE_FILE_SHIFT3 (_PAGE_BIT_PROTNONE + 1) +#else +#define PTE_FILE_SHIFT2 (_PAGE_BIT_PROTNONE + 1) +#define PTE_FILE_SHIFT3 (_PAGE_BIT_FILE + 1) +#endif +#define PTE_FILE_BITS1 (PTE_FILE_SHIFT2 - PTE_FILE_SHIFT1 - 1) +#define PTE_FILE_BITS2 (PTE_FILE_SHIFT3 - PTE_FILE_SHIFT2 - 1) #define pte_to_pgoff(pte) \ - ((((pte).pte_low >> 1) & 0x1f) + (((pte).pte_low >> 8) << 5)) + ((((pte).pte_low >> PTE_FILE_SHIFT1) \ + & ((1U << PTE_FILE_BITS1) - 1)) \ + + ((((pte).pte_low >> PTE_FILE_SHIFT2) \ + & ((1U << PTE_FILE_BITS2) - 1)) << PTE_FILE_BITS1) \ + + (((pte).pte_low >> PTE_FILE_SHIFT3) \ + << (PTE_FILE_BITS1 + PTE_FILE_BITS2))) #define pgoff_to_pte(off) \ - ((pte_t) { .pte_low = (((off) & 0x1f) << 1) + \ - (((off) >> 5) << 8) + _PAGE_FILE }) + ((pte_t) { .pte_low = \ + (((off) & ((1U << PTE_FILE_BITS1) - 1)) << PTE_FILE_SHIFT1) \ + + ((((off) >> PTE_FILE_BITS1) & ((1U << PTE_FILE_BITS2) - 1)) \ + << PTE_FILE_SHIFT2) \ + + (((off) >> (PTE_FILE_BITS1 + PTE_FILE_BITS2)) \ + << PTE_FILE_SHIFT3) \ + + _PAGE_FILE }) /* Encode and de-code a swap entry */ -#define __swp_type(x) (((x).val >> 1) & 0x1f) -#define __swp_offset(x) ((x).val >> 8) -#define __swp_entry(type, offset) \ - ((swp_entry_t) { ((type) << 1) | ((offset) << 8) }) +#if _PAGE_BIT_FILE < _PAGE_BIT_PROTNONE +#define SWP_TYPE_BITS (_PAGE_BIT_FILE - _PAGE_BIT_PRESENT - 1) +#define SWP_OFFSET_SHIFT (_PAGE_BIT_PROTNONE + 1) +#else +#define SWP_TYPE_BITS (_PAGE_BIT_PROTNONE - _PAGE_BIT_PRESENT - 1) +#define SWP_OFFSET_SHIFT (_PAGE_BIT_FILE + 1) +#endif + +#define MAX_SWAPFILES_CHECK() BUILD_BUG_ON(MAX_SWAPFILES_SHIFT > SWP_TYPE_BITS) + +#define __swp_type(x) (((x).val >> (_PAGE_BIT_PRESENT + 1)) \ + & ((1U << SWP_TYPE_BITS) - 1)) +#define __swp_offset(x) ((x).val >> SWP_OFFSET_SHIFT) +#define __swp_entry(type, offset) ((swp_entry_t) { \ + ((type) << (_PAGE_BIT_PRESENT + 1)) \ + | ((offset) << SWP_OFFSET_SHIFT) }) #define __pte_to_swp_entry(pte) ((swp_entry_t) { (pte).pte_low }) #define __swp_entry_to_pte(x) ((pte_t) { .pte = (x).val }) diff --git a/arch/x86/include/asm/pgtable-3level.h b/arch/x86/include/asm/pgtable-3level.h index 52597aeadfff..447da43cddb3 100644 --- a/arch/x86/include/asm/pgtable-3level.h +++ b/arch/x86/include/asm/pgtable-3level.h @@ -166,6 +166,7 @@ static inline int pte_none(pte_t pte) #define PTE_FILE_MAX_BITS 32 /* Encode and de-code a swap entry */ +#define MAX_SWAPFILES_CHECK() BUILD_BUG_ON(MAX_SWAPFILES_SHIFT > 5) #define __swp_type(x) (((x).val) & 0x1f) #define __swp_offset(x) ((x).val >> 5) #define __swp_entry(type, offset) ((swp_entry_t){(type) | (offset) << 5}) diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h index c012f3b11671..b7c2ecdb7658 100644 --- a/arch/x86/include/asm/pgtable.h +++ b/arch/x86/include/asm/pgtable.h @@ -10,7 +10,6 @@ #define _PAGE_BIT_PCD 4 /* page cache disabled */ #define _PAGE_BIT_ACCESSED 5 /* was accessed (raised by CPU) */ #define _PAGE_BIT_DIRTY 6 /* was written to (raised by CPU) */ -#define _PAGE_BIT_FILE 6 #define _PAGE_BIT_PSE 7 /* 4 MB (or 2MB) page */ #define _PAGE_BIT_PAT 7 /* on 4KB pages */ #define _PAGE_BIT_GLOBAL 8 /* Global TLB entry PPro+ */ @@ -22,6 +21,12 @@ #define _PAGE_BIT_CPA_TEST _PAGE_BIT_UNUSED1 #define _PAGE_BIT_NX 63 /* No execute: only valid after cpuid check */ +/* If _PAGE_BIT_PRESENT is clear, we use these: */ +/* - if the user mapped it with PROT_NONE; pte_present gives true */ +#define _PAGE_BIT_PROTNONE _PAGE_BIT_GLOBAL +/* - set: nonlinear file mapping, saved PTE; unset:swap */ +#define _PAGE_BIT_FILE _PAGE_BIT_DIRTY + #define _PAGE_PRESENT (_AT(pteval_t, 1) << _PAGE_BIT_PRESENT) #define _PAGE_RW (_AT(pteval_t, 1) << _PAGE_BIT_RW) #define _PAGE_USER (_AT(pteval_t, 1) << _PAGE_BIT_USER) @@ -46,11 +51,8 @@ #define _PAGE_NX (_AT(pteval_t, 0)) #endif -/* If _PAGE_PRESENT is clear, we use these: */ -#define _PAGE_FILE _PAGE_DIRTY /* nonlinear file mapping, - * saved PTE; unset:swap */ -#define _PAGE_PROTNONE _PAGE_PSE /* if the user mapped it with PROT_NONE; - pte_present gives true */ +#define _PAGE_FILE (_AT(pteval_t, 1) << _PAGE_BIT_FILE) +#define _PAGE_PROTNONE (_AT(pteval_t, 1) << _PAGE_BIT_PROTNONE) #define _PAGE_TABLE (_PAGE_PRESENT | _PAGE_RW | _PAGE_USER | \ _PAGE_ACCESSED | _PAGE_DIRTY) diff --git a/arch/x86/include/asm/pgtable_64.h b/arch/x86/include/asm/pgtable_64.h index 545a0e042bb2..65b6be6677c7 100644 --- a/arch/x86/include/asm/pgtable_64.h +++ b/arch/x86/include/asm/pgtable_64.h @@ -250,10 +250,22 @@ static inline int pud_large(pud_t pte) extern int direct_gbpages; /* Encode and de-code a swap entry */ -#define __swp_type(x) (((x).val >> 1) & 0x3f) -#define __swp_offset(x) ((x).val >> 8) -#define __swp_entry(type, offset) ((swp_entry_t) { ((type) << 1) | \ - ((offset) << 8) }) +#if _PAGE_BIT_FILE < _PAGE_BIT_PROTNONE +#define SWP_TYPE_BITS (_PAGE_BIT_FILE - _PAGE_BIT_PRESENT - 1) +#define SWP_OFFSET_SHIFT (_PAGE_BIT_PROTNONE + 1) +#else +#define SWP_TYPE_BITS (_PAGE_BIT_PROTNONE - _PAGE_BIT_PRESENT - 1) +#define SWP_OFFSET_SHIFT (_PAGE_BIT_FILE + 1) +#endif + +#define MAX_SWAPFILES_CHECK() BUILD_BUG_ON(MAX_SWAPFILES_SHIFT > SWP_TYPE_BITS) + +#define __swp_type(x) (((x).val >> (_PAGE_BIT_PRESENT + 1)) \ + & ((1U << SWP_TYPE_BITS) - 1)) +#define __swp_offset(x) ((x).val >> SWP_OFFSET_SHIFT) +#define __swp_entry(type, offset) ((swp_entry_t) { \ + ((type) << (_PAGE_BIT_PRESENT + 1)) \ + | ((offset) << SWP_OFFSET_SHIFT) }) #define __pte_to_swp_entry(pte) ((swp_entry_t) { pte_val((pte)) }) #define __swp_entry_to_pte(x) ((pte_t) { .pte = (x).val }) diff --git a/mm/swapfile.c b/mm/swapfile.c index 90cb67a5417c..54a9f87e5162 100644 --- a/mm/swapfile.c +++ b/mm/swapfile.c @@ -1462,6 +1462,15 @@ static int __init procswaps_init(void) __initcall(procswaps_init); #endif /* CONFIG_PROC_FS */ +#ifdef MAX_SWAPFILES_CHECK +static int __init max_swapfiles_check(void) +{ + MAX_SWAPFILES_CHECK(); + return 0; +} +late_initcall(max_swapfiles_check); +#endif + /* * Written 01/25/92 by Simmule Turner, heavily changed by Linus. * -- cgit v1.2.3 From b93a531e315e97ef00367099e6b5f19651936e20 Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Tue, 16 Dec 2008 11:40:27 +0000 Subject: allow bug table entries to use relative pointers (and use it on x86-64) Impact: reduce bug table size This allows reducing the bug table size by half. Perhaps there are other 64-bit architectures that could also make use of this. Signed-off-by: Jan Beulich Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 4 ++++ arch/x86/include/asm/bug.h | 2 +- include/asm-generic/bug.h | 8 ++++++++ lib/bug.c | 19 +++++++++++++++++-- 4 files changed, 30 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index ac22bb7719f7..ab98cca84e1b 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -87,6 +87,10 @@ config GENERIC_IOMAP config GENERIC_BUG def_bool y depends on BUG + select GENERIC_BUG_RELATIVE_POINTERS if X86_64 + +config GENERIC_BUG_RELATIVE_POINTERS + bool config GENERIC_HWEIGHT def_bool y diff --git a/arch/x86/include/asm/bug.h b/arch/x86/include/asm/bug.h index 3def2065fcea..d9cf1cd156d2 100644 --- a/arch/x86/include/asm/bug.h +++ b/arch/x86/include/asm/bug.h @@ -9,7 +9,7 @@ #ifdef CONFIG_X86_32 # define __BUG_C0 "2:\t.long 1b, %c0\n" #else -# define __BUG_C0 "2:\t.quad 1b, %c0\n" +# define __BUG_C0 "2:\t.long 1b - 2b, %c0 - 2b\n" #endif #define BUG() \ diff --git a/include/asm-generic/bug.h b/include/asm-generic/bug.h index 12c07c1866b2..4c794d73fb84 100644 --- a/include/asm-generic/bug.h +++ b/include/asm-generic/bug.h @@ -8,9 +8,17 @@ #ifdef CONFIG_GENERIC_BUG #ifndef __ASSEMBLY__ struct bug_entry { +#ifndef CONFIG_GENERIC_BUG_RELATIVE_POINTERS unsigned long bug_addr; +#else + signed int bug_addr_disp; +#endif #ifdef CONFIG_DEBUG_BUGVERBOSE +#ifndef CONFIG_GENERIC_BUG_RELATIVE_POINTERS const char *file; +#else + signed int file_disp; +#endif unsigned short line; #endif unsigned short flags; diff --git a/lib/bug.c b/lib/bug.c index bfeafd60ee9f..300e41afbf97 100644 --- a/lib/bug.c +++ b/lib/bug.c @@ -5,6 +5,8 @@ CONFIG_BUG - emit BUG traps. Nothing happens without this. CONFIG_GENERIC_BUG - enable this code. + CONFIG_GENERIC_BUG_RELATIVE_POINTERS - use 32-bit pointers relative to + the containing struct bug_entry for bug_addr and file. CONFIG_DEBUG_BUGVERBOSE - emit full file+line information for each BUG CONFIG_BUG and CONFIG_DEBUG_BUGVERBOSE are potentially user-settable @@ -43,6 +45,15 @@ extern const struct bug_entry __start___bug_table[], __stop___bug_table[]; +static inline unsigned long bug_addr(const struct bug_entry *bug) +{ +#ifndef CONFIG_GENERIC_BUG_RELATIVE_POINTERS + return bug->bug_addr; +#else + return (unsigned long)bug + bug->bug_addr_disp; +#endif +} + #ifdef CONFIG_MODULES static LIST_HEAD(module_bug_list); @@ -55,7 +66,7 @@ static const struct bug_entry *module_find_bug(unsigned long bugaddr) unsigned i; for (i = 0; i < mod->num_bugs; ++i, ++bug) - if (bugaddr == bug->bug_addr) + if (bugaddr == bug_addr(bug)) return bug; } return NULL; @@ -108,7 +119,7 @@ const struct bug_entry *find_bug(unsigned long bugaddr) const struct bug_entry *bug; for (bug = __start___bug_table; bug < __stop___bug_table; ++bug) - if (bugaddr == bug->bug_addr) + if (bugaddr == bug_addr(bug)) return bug; return module_find_bug(bugaddr); @@ -133,7 +144,11 @@ enum bug_trap_type report_bug(unsigned long bugaddr, struct pt_regs *regs) if (bug) { #ifdef CONFIG_DEBUG_BUGVERBOSE +#ifndef CONFIG_GENERIC_BUG_RELATIVE_POINTERS file = bug->file; +#else + file = (const char *)bug + bug->file_disp; +#endif line = bug->line; #endif warning = (bug->flags & BUGFLAG_WARNING) != 0; -- cgit v1.2.3 From d6be89ad660c5d03edef91715093d447025df59b Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Tue, 16 Dec 2008 11:42:45 +0000 Subject: x86, 32-bit: simplify alloc_low_page() Impact: cleanup Neither of the callers really needs the physical address this function returns, so eliminate the pointless argument. Signed-off-by: Jan Beulich Signed-off-by: Ingo Molnar --- arch/x86/mm/init_32.c | 12 ++++-------- 1 file changed, 4 insertions(+), 8 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c index 3ffed259883e..333c9e79d46f 100644 --- a/arch/x86/mm/init_32.c +++ b/arch/x86/mm/init_32.c @@ -67,7 +67,7 @@ static unsigned long __meminitdata table_top; static int __initdata after_init_bootmem; -static __init void *alloc_low_page(unsigned long *phys) +static __init void *alloc_low_page(void) { unsigned long pfn = table_end++; void *adr; @@ -77,7 +77,6 @@ static __init void *alloc_low_page(unsigned long *phys) adr = __va(pfn * PAGE_SIZE); memset(adr, 0, PAGE_SIZE); - *phys = pfn * PAGE_SIZE; return adr; } @@ -92,12 +91,11 @@ static pmd_t * __init one_md_table_init(pgd_t *pgd) pmd_t *pmd_table; #ifdef CONFIG_X86_PAE - unsigned long phys; if (!(pgd_val(*pgd) & _PAGE_PRESENT)) { if (after_init_bootmem) pmd_table = (pmd_t *)alloc_bootmem_low_pages(PAGE_SIZE); else - pmd_table = (pmd_t *)alloc_low_page(&phys); + pmd_table = (pmd_t *)alloc_low_page(); paravirt_alloc_pmd(&init_mm, __pa(pmd_table) >> PAGE_SHIFT); set_pgd(pgd, __pgd(__pa(pmd_table) | _PAGE_PRESENT)); pud = pud_offset(pgd, 0); @@ -128,10 +126,8 @@ static pte_t * __init one_page_table_init(pmd_t *pmd) if (!page_table) page_table = (pte_t *)alloc_bootmem_low_pages(PAGE_SIZE); - } else { - unsigned long phys; - page_table = (pte_t *)alloc_low_page(&phys); - } + } else + page_table = (pte_t *)alloc_low_page(); paravirt_alloc_pte(&init_mm, __pa(page_table) >> PAGE_SHIFT); set_pmd(pmd, __pmd(__pa(page_table) | _PAGE_TABLE)); -- cgit v1.2.3 From beeb4195cbc80b7489631361b7ed38b7518af433 Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Tue, 16 Dec 2008 11:45:56 +0000 Subject: x86, 32-bit: add some compile time checks to mem_init() Some of the inconsistencies checked for at run time can be detected at build time already, so duplicate the checks done at run time to also be done at build time. Signed-off-by: Jan Beulich Signed-off-by: Ingo Molnar --- arch/x86/mm/init_32.c | 16 +++++++++++++++- 1 file changed, 15 insertions(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c index c483f4242079..d3a45d54547a 100644 --- a/arch/x86/mm/init_32.c +++ b/arch/x86/mm/init_32.c @@ -1040,11 +1040,25 @@ void __init mem_init(void) (unsigned long)&_text, (unsigned long)&_etext, ((unsigned long)&_etext - (unsigned long)&_text) >> 10); + /* + * Check boundaries twice: Some fundamental inconsistencies can + * be detected at build time already. + */ +#define __FIXADDR_TOP (-PAGE_SIZE) +#ifdef CONFIG_HIGHMEM + BUILD_BUG_ON(PKMAP_BASE + LAST_PKMAP*PAGE_SIZE > FIXADDR_START); + BUILD_BUG_ON(VMALLOC_END > PKMAP_BASE); +#endif +#define high_memory (-128UL << 20) + BUILD_BUG_ON(VMALLOC_START >= VMALLOC_END); +#undef high_memory +#undef __FIXADDR_TOP + #ifdef CONFIG_HIGHMEM BUG_ON(PKMAP_BASE + LAST_PKMAP*PAGE_SIZE > FIXADDR_START); BUG_ON(VMALLOC_END > PKMAP_BASE); #endif - BUG_ON(VMALLOC_START > VMALLOC_END); + BUG_ON(VMALLOC_START >= VMALLOC_END); BUG_ON((unsigned long)high_memory > VMALLOC_START); if (boot_cpu_data.wp_works_ok < 0) -- cgit v1.2.3 From cfc319833b5b359bf3bce99564dbac00af7925ac Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Tue, 16 Dec 2008 11:46:58 +0000 Subject: x86, 32-bit: improve lazy TLB handling code Impact: micro-optimize the 32-bit TLB flush code Use the faster x86_{read,write}_percpu() accessors here. Signed-off-by: Jan Beulich Signed-off-by: Ingo Molnar --- arch/x86/include/asm/mmu_context_32.h | 13 ++++++------- arch/x86/kernel/tlb_32.c | 11 +++++------ 2 files changed, 11 insertions(+), 13 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/mmu_context_32.h b/arch/x86/include/asm/mmu_context_32.h index 8e10015781fb..7e98ce1d2c0e 100644 --- a/arch/x86/include/asm/mmu_context_32.h +++ b/arch/x86/include/asm/mmu_context_32.h @@ -4,9 +4,8 @@ static inline void enter_lazy_tlb(struct mm_struct *mm, struct task_struct *tsk) { #ifdef CONFIG_SMP - unsigned cpu = smp_processor_id(); - if (per_cpu(cpu_tlbstate, cpu).state == TLBSTATE_OK) - per_cpu(cpu_tlbstate, cpu).state = TLBSTATE_LAZY; + if (x86_read_percpu(cpu_tlbstate.state) == TLBSTATE_OK) + x86_write_percpu(cpu_tlbstate.state, TLBSTATE_LAZY); #endif } @@ -20,8 +19,8 @@ static inline void switch_mm(struct mm_struct *prev, /* stop flush ipis for the previous mm */ cpu_clear(cpu, prev->cpu_vm_mask); #ifdef CONFIG_SMP - per_cpu(cpu_tlbstate, cpu).state = TLBSTATE_OK; - per_cpu(cpu_tlbstate, cpu).active_mm = next; + x86_write_percpu(cpu_tlbstate.state, TLBSTATE_OK); + x86_write_percpu(cpu_tlbstate.active_mm, next); #endif cpu_set(cpu, next->cpu_vm_mask); @@ -36,8 +35,8 @@ static inline void switch_mm(struct mm_struct *prev, } #ifdef CONFIG_SMP else { - per_cpu(cpu_tlbstate, cpu).state = TLBSTATE_OK; - BUG_ON(per_cpu(cpu_tlbstate, cpu).active_mm != next); + x86_write_percpu(cpu_tlbstate.state, TLBSTATE_OK); + BUG_ON(x86_read_percpu(cpu_tlbstate.active_mm) != next); if (!cpu_test_and_set(cpu, next->cpu_vm_mask)) { /* We were in lazy tlb mode and leave_mm disabled diff --git a/arch/x86/kernel/tlb_32.c b/arch/x86/kernel/tlb_32.c index f4049f3513b6..4290d918b58a 100644 --- a/arch/x86/kernel/tlb_32.c +++ b/arch/x86/kernel/tlb_32.c @@ -34,9 +34,8 @@ static DEFINE_SPINLOCK(tlbstate_lock); */ void leave_mm(int cpu) { - if (per_cpu(cpu_tlbstate, cpu).state == TLBSTATE_OK) - BUG(); - cpu_clear(cpu, per_cpu(cpu_tlbstate, cpu).active_mm->cpu_vm_mask); + BUG_ON(x86_read_percpu(cpu_tlbstate.state) == TLBSTATE_OK); + cpu_clear(cpu, x86_read_percpu(cpu_tlbstate.active_mm)->cpu_vm_mask); load_cr3(swapper_pg_dir); } EXPORT_SYMBOL_GPL(leave_mm); @@ -104,8 +103,8 @@ void smp_invalidate_interrupt(struct pt_regs *regs) * BUG(); */ - if (flush_mm == per_cpu(cpu_tlbstate, cpu).active_mm) { - if (per_cpu(cpu_tlbstate, cpu).state == TLBSTATE_OK) { + if (flush_mm == x86_read_percpu(cpu_tlbstate.active_mm)) { + if (x86_read_percpu(cpu_tlbstate.state) == TLBSTATE_OK) { if (flush_va == TLB_FLUSH_ALL) local_flush_tlb(); else @@ -238,7 +237,7 @@ static void do_flush_tlb_all(void *info) unsigned long cpu = smp_processor_id(); __flush_tlb_all(); - if (per_cpu(cpu_tlbstate, cpu).state == TLBSTATE_LAZY) + if (x86_read_percpu(cpu_tlbstate.state) == TLBSTATE_LAZY) leave_mm(cpu); } -- cgit v1.2.3 From 83fd5cc6481c6b7fa8b45f8a7e0aa7120213430b Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Tue, 16 Dec 2008 19:17:11 +0100 Subject: AMD IOMMU: allocate rlookup_table with __GFP_ZERO Impact: fix bug which can lead to panic in prealloc_protection_domains() Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu_init.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu_init.c b/arch/x86/kernel/amd_iommu_init.c index 30ae2701b3df..c90a15eba5c5 100644 --- a/arch/x86/kernel/amd_iommu_init.c +++ b/arch/x86/kernel/amd_iommu_init.c @@ -1074,7 +1074,8 @@ int __init amd_iommu_init(void) goto free; /* IOMMU rlookup table - find the IOMMU for a specific device */ - amd_iommu_rlookup_table = (void *)__get_free_pages(GFP_KERNEL, + amd_iommu_rlookup_table = (void *)__get_free_pages( + GFP_KERNEL | __GFP_ZERO, get_order(rlookup_table_size)); if (amd_iommu_rlookup_table == NULL) goto free; -- cgit v1.2.3 From b6fd6f26733e864fba2ea3eb1d716e23d2e66f3a Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Tue, 16 Dec 2008 19:23:36 +0100 Subject: x86, mm: limit MAXMEM on 64-bit on 64-bit x86 the physical memory limit is controlled by the sparsemem bits - which are 44 bits right now. But MAXMEM (the max pfn number e820 parsing will allow to enter our sizing routines) is set to 0x00003fffffffffff, i.e. 46 bits - that's too large because it overlaps into the vmalloc range. So couple MAXMEM to MAX_PHYSMEM_BITS, and add a comment that the maximum of MAX_PHYSMEM_BITS is 45 bits. Signed-off-by: Ingo Molnar --- arch/x86/include/asm/pgtable_64.h | 2 +- arch/x86/include/asm/sparsemem.h | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/pgtable_64.h b/arch/x86/include/asm/pgtable_64.h index 65b6be6677c7..c54ba69608bd 100644 --- a/arch/x86/include/asm/pgtable_64.h +++ b/arch/x86/include/asm/pgtable_64.h @@ -146,7 +146,7 @@ static inline void native_pgd_clear(pgd_t *pgd) #define PGDIR_MASK (~(PGDIR_SIZE - 1)) -#define MAXMEM _AC(0x00003fffffffffff, UL) +#define MAXMEM _AC(__AC(1, UL) << MAX_PHYSMEM_BITS, UL) #define VMALLOC_START _AC(0xffffc20000000000, UL) #define VMALLOC_END _AC(0xffffe1ffffffffff, UL) #define VMEMMAP_START _AC(0xffffe20000000000, UL) diff --git a/arch/x86/include/asm/sparsemem.h b/arch/x86/include/asm/sparsemem.h index be44f7dab395..e3cc3c063ec5 100644 --- a/arch/x86/include/asm/sparsemem.h +++ b/arch/x86/include/asm/sparsemem.h @@ -27,7 +27,7 @@ #else /* CONFIG_X86_32 */ # define SECTION_SIZE_BITS 27 /* matt - 128 is convenient right now */ # define MAX_PHYSADDR_BITS 44 -# define MAX_PHYSMEM_BITS 44 +# define MAX_PHYSMEM_BITS 44 /* Can be max 45 bits */ #endif #endif /* CONFIG_SPARSEMEM */ -- cgit v1.2.3 From 3c763fd77e66e55d029052da31df0abd9920cb1e Mon Sep 17 00:00:00 2001 From: Andreas Herrmann Date: Tue, 16 Dec 2008 19:07:47 +0100 Subject: x86: microcode_amd: fix wrong handling of equivalent CPU id Impact: fix bug resulting in non-loaded AMD microcode mc_header->processor_rev_id is a 2 byte value. Similar is true for equiv_cpu in an equiv_cpu_entry -- only 2 bytes are of interest. Signed-off-by: Andreas Herrmann Signed-off-by: Ingo Molnar --- arch/x86/kernel/microcode_amd.c | 23 ++++++----------------- 1 file changed, 6 insertions(+), 17 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/microcode_amd.c b/arch/x86/kernel/microcode_amd.c index 5f8e5d75a254..b5bc81470bcf 100644 --- a/arch/x86/kernel/microcode_amd.c +++ b/arch/x86/kernel/microcode_amd.c @@ -62,7 +62,7 @@ struct microcode_header_amd { unsigned int mc_patch_data_checksum; unsigned int nb_dev_id; unsigned int sb_dev_id; - unsigned char processor_rev_id[2]; + u16 processor_rev_id; unsigned char nb_rev_id; unsigned char sb_rev_id; unsigned char bios_api_rev; @@ -125,7 +125,7 @@ static int get_matching_microcode(int cpu, void *mc, int rev) while (equiv_cpu_table[i].installed_cpu != 0) { if (current_cpu_id == equiv_cpu_table[i].installed_cpu) { - equiv_cpu_id = equiv_cpu_table[i].equiv_cpu; + equiv_cpu_id = equiv_cpu_table[i].equiv_cpu & 0xffff; break; } i++; @@ -137,21 +137,10 @@ static int get_matching_microcode(int cpu, void *mc, int rev) return 0; } - if ((mc_header->processor_rev_id[0]) != (equiv_cpu_id & 0xff)) { - printk(KERN_ERR - "microcode: CPU%d patch does not match " - "(patch is %x, cpu extended is %x) \n", - cpu, mc_header->processor_rev_id[0], - (equiv_cpu_id & 0xff)); - return 0; - } - - if ((mc_header->processor_rev_id[1]) != ((equiv_cpu_id >> 16) & 0xff)) { - printk(KERN_ERR "microcode: CPU%d patch does not match " - "(patch is %x, cpu base id is %x) \n", - cpu, mc_header->processor_rev_id[1], - ((equiv_cpu_id >> 16) & 0xff)); - + if (mc_header->processor_rev_id != equiv_cpu_id) { + printk(KERN_ERR "microcode: CPU%d patch does not match " + "(processor_rev_id: %x, eqiv_cpu_id: %x)\n", + cpu, mc_header->processor_rev_id, equiv_cpu_id); return 0; } -- cgit v1.2.3 From 2a3282a77b02fb47576ffbdb4867c8c6eeb83ed5 Mon Sep 17 00:00:00 2001 From: Andreas Herrmann Date: Tue, 16 Dec 2008 19:08:53 +0100 Subject: x86: microcode_amd: fix typos and trailing whitespaces in log messages Impact: fix printk typos Signed-off-by: Andreas Herrmann Signed-off-by: Ingo Molnar --- arch/x86/kernel/microcode_amd.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/microcode_amd.c b/arch/x86/kernel/microcode_amd.c index b5bc81470bcf..83a9fa321d9b 100644 --- a/arch/x86/kernel/microcode_amd.c +++ b/arch/x86/kernel/microcode_amd.c @@ -10,7 +10,7 @@ * This driver allows to upgrade microcode on AMD * family 0x10 and 0x11 processors. * - * Licensed unter the terms of the GNU General Public + * Licensed under the terms of the GNU General Public * License version 2. See file COPYING for details. */ @@ -133,7 +133,7 @@ static int get_matching_microcode(int cpu, void *mc, int rev) if (!equiv_cpu_id) { printk(KERN_ERR "microcode: CPU%d cpu_id " - "not found in equivalent cpu table \n", cpu); + "not found in equivalent cpu table\n", cpu); return 0; } @@ -151,7 +151,7 @@ static int get_matching_microcode(int cpu, void *mc, int rev) NULL); if ((!nb_pci_dev) || (mc_header->nb_rev_id != nb_pci_dev->revision)) { - printk(KERN_ERR "microcode: CPU%d NB mismatch \n", cpu); + printk(KERN_ERR "microcode: CPU%d NB mismatch\n", cpu); pci_dev_put(nb_pci_dev); return 0; } @@ -165,7 +165,7 @@ static int get_matching_microcode(int cpu, void *mc, int rev) NULL); if ((!sb_pci_dev) || (mc_header->sb_rev_id != sb_pci_dev->revision)) { - printk(KERN_ERR "microcode: CPU%d SB mismatch \n", cpu); + printk(KERN_ERR "microcode: CPU%d SB mismatch\n", cpu); pci_dev_put(sb_pci_dev); return 0; } @@ -219,7 +219,7 @@ static void apply_microcode_amd(int cpu) } printk(KERN_INFO "microcode: CPU%d updated from revision " - "0x%x to 0x%x \n", + "0x%x to 0x%x\n", cpu_num, uci->cpu_sig.rev, mc_amd->hdr.patch_id); uci->cpu_sig.rev = rev; @@ -282,7 +282,7 @@ static int install_equiv_cpu_table(u8 *buf, if (buf_pos[1] != UCODE_EQUIV_CPU_TABLE_TYPE || !size) { printk(KERN_ERR "microcode: error! " - "Wrong microcode equivalnet cpu table\n"); + "Wrong microcode equivalent cpu table\n"); return 0; } -- cgit v1.2.3 From be957763b01905d33b53cdd25c8df110f94f499a Mon Sep 17 00:00:00 2001 From: Andreas Herrmann Date: Tue, 16 Dec 2008 19:11:23 +0100 Subject: x86: microcode_amd: fix checkpatch warnings/errors Impact: cleanup Signed-off-by: Andreas Herrmann Signed-off-by: Ingo Molnar --- arch/x86/kernel/microcode_amd.c | 25 ++++++++++++++----------- 1 file changed, 14 insertions(+), 11 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/microcode_amd.c b/arch/x86/kernel/microcode_amd.c index 83a9fa321d9b..a8a0ec600554 100644 --- a/arch/x86/kernel/microcode_amd.c +++ b/arch/x86/kernel/microcode_amd.c @@ -32,9 +32,9 @@ #include #include #include +#include #include -#include #include #include @@ -225,7 +225,7 @@ static void apply_microcode_amd(int cpu) uci->cpu_sig.rev = rev; } -static void * get_next_ucode(u8 *buf, unsigned int size, +static void *get_next_ucode(u8 *buf, unsigned int size, int (*get_ucode_data)(void *, const void *, size_t), unsigned int *mc_size) { @@ -256,7 +256,8 @@ static void * get_next_ucode(u8 *buf, unsigned int size, mc = vmalloc(UCODE_MAX_SIZE); if (mc) { memset(mc, 0, UCODE_MAX_SIZE); - if (get_ucode_data(mc, buf + UCODE_CONTAINER_SECTION_HDR, total_size)) { + if (get_ucode_data(mc, buf + UCODE_CONTAINER_SECTION_HDR, + total_size)) { vfree(mc); mc = NULL; } else @@ -332,7 +333,8 @@ static int generic_load_microcode(int cpu, void *data, size_t size, unsigned int uninitialized_var(mc_size); struct microcode_header_amd *mc_header; - mc = get_next_ucode(ucode_ptr, leftover, get_ucode_data, &mc_size); + mc = get_next_ucode(ucode_ptr, leftover, get_ucode_data, + &mc_size); if (!mc) break; @@ -342,7 +344,7 @@ static int generic_load_microcode(int cpu, void *data, size_t size, vfree(new_mc); new_rev = mc_header->patch_id; new_mc = mc; - } else + } else vfree(mc); ucode_ptr += mc_size; @@ -354,9 +356,9 @@ static int generic_load_microcode(int cpu, void *data, size_t size, if (uci->mc) vfree(uci->mc); uci->mc = new_mc; - pr_debug("microcode: CPU%d found a matching microcode update with" - " version 0x%x (current=0x%x)\n", - cpu, new_rev, uci->cpu_sig.rev); + pr_debug("microcode: CPU%d found a matching microcode " + "update with version 0x%x (current=0x%x)\n", + cpu, new_rev, uci->cpu_sig.rev); } else vfree(new_mc); } @@ -383,7 +385,8 @@ static int request_microcode_fw(int cpu, struct device *device) ret = request_firmware(&firmware, fw_name, device); if (ret) { - printk(KERN_ERR "microcode: ucode data file %s load failed\n", fw_name); + printk(KERN_ERR "microcode: ucode data file %s load failed\n", + fw_name); return ret; } @@ -397,8 +400,8 @@ static int request_microcode_fw(int cpu, struct device *device) static int request_microcode_user(int cpu, const void __user *buf, size_t size) { - printk(KERN_WARNING "microcode: AMD microcode update via /dev/cpu/microcode" - "is not supported\n"); + printk(KERN_WARNING "microcode: AMD microcode update via " + "/dev/cpu/microcode is not supported\n"); return -1; } -- cgit v1.2.3 From 8c135206c826095c852c16d94a0a74eeaf05c90d Mon Sep 17 00:00:00 2001 From: Andreas Herrmann Date: Tue, 16 Dec 2008 19:13:00 +0100 Subject: x86: microcode_amd: fix compile warning MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Impact: fix build warning CC arch/x86/kernel/microcode_amd.o arch/x86/kernel/microcode_amd.c: In function ‘request_microcode_fw’: arch/x86/kernel/microcode_amd.c:393: warning: passing argument 2 of ‘generic_load_microcode’ discards qualifiers from pointer target type (Respect "const" qualifier of firmware->data.) Signed-off-by: Andreas Herrmann Signed-off-by: Ingo Molnar --- arch/x86/kernel/microcode_amd.c | 20 +++++++++++--------- 1 file changed, 11 insertions(+), 9 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/microcode_amd.c b/arch/x86/kernel/microcode_amd.c index a8a0ec600554..89b386c901fd 100644 --- a/arch/x86/kernel/microcode_amd.c +++ b/arch/x86/kernel/microcode_amd.c @@ -225,8 +225,8 @@ static void apply_microcode_amd(int cpu) uci->cpu_sig.rev = rev; } -static void *get_next_ucode(u8 *buf, unsigned int size, - int (*get_ucode_data)(void *, const void *, size_t), +static void *get_next_ucode(const u8 *buf, unsigned int size, + int (*get_ucode_data)(void *, const u8 *, size_t), unsigned int *mc_size) { unsigned int total_size; @@ -268,8 +268,8 @@ static void *get_next_ucode(u8 *buf, unsigned int size, } -static int install_equiv_cpu_table(u8 *buf, - int (*get_ucode_data)(void *, const void *, size_t)) +static int install_equiv_cpu_table(const u8 *buf, + int (*get_ucode_data)(void *, const u8 *, size_t)) { #define UCODE_CONTAINER_HEADER_SIZE 12 u8 *container_hdr[UCODE_CONTAINER_HEADER_SIZE]; @@ -311,11 +311,13 @@ static void free_equiv_cpu_table(void) } } -static int generic_load_microcode(int cpu, void *data, size_t size, - int (*get_ucode_data)(void *, const void *, size_t)) +static int generic_load_microcode(int cpu, const u8 *data, size_t size, + int (*get_ucode_data)(void *, const u8 *, size_t)) { struct ucode_cpu_info *uci = ucode_cpu_info + cpu; - u8 *ucode_ptr = data, *new_mc = NULL, *mc; + const u8 *ucode_ptr = data; + void *new_mc = NULL; + void *mc; int new_rev = uci->cpu_sig.rev; unsigned int leftover; unsigned long offset; @@ -368,7 +370,7 @@ static int generic_load_microcode(int cpu, void *data, size_t size, return (int)leftover; } -static int get_ucode_fw(void *to, const void *from, size_t n) +static int get_ucode_fw(void *to, const u8 *from, size_t n) { memcpy(to, from, n); return 0; @@ -390,7 +392,7 @@ static int request_microcode_fw(int cpu, struct device *device) return ret; } - ret = generic_load_microcode(cpu, (void*)firmware->data, firmware->size, + ret = generic_load_microcode(cpu, firmware->data, firmware->size, &get_ucode_fw); release_firmware(firmware); -- cgit v1.2.3 From 0657d9ebff186dcdb17e582dcb909028775a7707 Mon Sep 17 00:00:00 2001 From: Andreas Herrmann Date: Tue, 16 Dec 2008 19:14:05 +0100 Subject: x86: microcode_amd: don't pass superfluous function pointer for get_ucode_data Impact: cleanup Signed-off-by: Andreas Herrmann Signed-off-by: Ingo Molnar --- arch/x86/kernel/microcode_amd.c | 29 ++++++++++++----------------- 1 file changed, 12 insertions(+), 17 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/microcode_amd.c b/arch/x86/kernel/microcode_amd.c index 89b386c901fd..c7f225c7e481 100644 --- a/arch/x86/kernel/microcode_amd.c +++ b/arch/x86/kernel/microcode_amd.c @@ -225,9 +225,14 @@ static void apply_microcode_amd(int cpu) uci->cpu_sig.rev = rev; } +static int get_ucode_data(void *to, const u8 *from, size_t n) +{ + memcpy(to, from, n); + return 0; +} + static void *get_next_ucode(const u8 *buf, unsigned int size, - int (*get_ucode_data)(void *, const u8 *, size_t), - unsigned int *mc_size) + unsigned int *mc_size) { unsigned int total_size; #define UCODE_CONTAINER_SECTION_HDR 8 @@ -268,8 +273,7 @@ static void *get_next_ucode(const u8 *buf, unsigned int size, } -static int install_equiv_cpu_table(const u8 *buf, - int (*get_ucode_data)(void *, const u8 *, size_t)) +static int install_equiv_cpu_table(const u8 *buf) { #define UCODE_CONTAINER_HEADER_SIZE 12 u8 *container_hdr[UCODE_CONTAINER_HEADER_SIZE]; @@ -311,8 +315,7 @@ static void free_equiv_cpu_table(void) } } -static int generic_load_microcode(int cpu, const u8 *data, size_t size, - int (*get_ucode_data)(void *, const u8 *, size_t)) +static int generic_load_microcode(int cpu, const u8 *data, size_t size) { struct ucode_cpu_info *uci = ucode_cpu_info + cpu; const u8 *ucode_ptr = data; @@ -322,7 +325,7 @@ static int generic_load_microcode(int cpu, const u8 *data, size_t size, unsigned int leftover; unsigned long offset; - offset = install_equiv_cpu_table(ucode_ptr, get_ucode_data); + offset = install_equiv_cpu_table(ucode_ptr); if (!offset) { printk(KERN_ERR "microcode: installing equivalent cpu table failed\n"); return -EINVAL; @@ -335,8 +338,7 @@ static int generic_load_microcode(int cpu, const u8 *data, size_t size, unsigned int uninitialized_var(mc_size); struct microcode_header_amd *mc_header; - mc = get_next_ucode(ucode_ptr, leftover, get_ucode_data, - &mc_size); + mc = get_next_ucode(ucode_ptr, leftover, &mc_size); if (!mc) break; @@ -370,12 +372,6 @@ static int generic_load_microcode(int cpu, const u8 *data, size_t size, return (int)leftover; } -static int get_ucode_fw(void *to, const u8 *from, size_t n) -{ - memcpy(to, from, n); - return 0; -} - static int request_microcode_fw(int cpu, struct device *device) { const char *fw_name = "amd-ucode/microcode_amd.bin"; @@ -392,8 +388,7 @@ static int request_microcode_fw(int cpu, struct device *device) return ret; } - ret = generic_load_microcode(cpu, firmware->data, firmware->size, - &get_ucode_fw); + ret = generic_load_microcode(cpu, firmware->data, firmware->size); release_firmware(firmware); -- cgit v1.2.3 From 29d0887ffd084cde9d6a1286cb82b71701a974dd Mon Sep 17 00:00:00 2001 From: Andreas Herrmann Date: Tue, 16 Dec 2008 19:16:34 +0100 Subject: x86: microcode_amd: replace inline asm by common rdmsr/wrmsr functions Impact: cleanup Signed-off-by: Andreas Herrmann Signed-off-by: Ingo Molnar --- arch/x86/include/asm/msr-index.h | 2 ++ arch/x86/kernel/microcode_amd.c | 23 +++++------------------ 2 files changed, 7 insertions(+), 18 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/msr-index.h b/arch/x86/include/asm/msr-index.h index e38859d577a1..cb58643947b9 100644 --- a/arch/x86/include/asm/msr-index.h +++ b/arch/x86/include/asm/msr-index.h @@ -85,7 +85,9 @@ /* AMD64 MSRs. Not complete. See the architecture manual for a more complete list. */ +#define MSR_AMD64_PATCH_LEVEL 0x0000008b #define MSR_AMD64_NB_CFG 0xc001001f +#define MSR_AMD64_PATCH_LOADER 0xc0010020 #define MSR_AMD64_IBSFETCHCTL 0xc0011030 #define MSR_AMD64_IBSFETCHLINAD 0xc0011031 #define MSR_AMD64_IBSFETCHPHYSAD 0xc0011032 diff --git a/arch/x86/kernel/microcode_amd.c b/arch/x86/kernel/microcode_amd.c index c7f225c7e481..2856955ddab1 100644 --- a/arch/x86/kernel/microcode_amd.c +++ b/arch/x86/kernel/microcode_amd.c @@ -93,6 +93,7 @@ static struct equiv_cpu_entry *equiv_cpu_table; static int collect_cpu_info_amd(int cpu, struct cpu_signature *csig) { struct cpuinfo_x86 *c = &cpu_data(cpu); + u32 dummy; memset(csig, 0, sizeof(*csig)); @@ -102,9 +103,7 @@ static int collect_cpu_info_amd(int cpu, struct cpu_signature *csig) return -1; } - asm volatile("movl %1, %%ecx; rdmsr" - : "=a" (csig->rev) - : "i" (0x0000008B) : "ecx"); + rdmsr(MSR_AMD64_PATCH_LEVEL, csig->rev, dummy); printk(KERN_INFO "microcode: collect_cpu_info_amd : patch_id=0x%x\n", csig->rev); @@ -181,12 +180,10 @@ static int get_matching_microcode(int cpu, void *mc, int rev) static void apply_microcode_amd(int cpu) { unsigned long flags; - unsigned int eax, edx; - unsigned int rev; + u32 rev, dummy; int cpu_num = raw_smp_processor_id(); struct ucode_cpu_info *uci = ucode_cpu_info + cpu_num; struct microcode_amd *mc_amd = uci->mc; - unsigned long addr; /* We should bind the task to the CPU */ BUG_ON(cpu_num != cpu); @@ -195,19 +192,9 @@ static void apply_microcode_amd(int cpu) return; spin_lock_irqsave(µcode_update_lock, flags); - - addr = (unsigned long)&mc_amd->hdr.data_code; - edx = (unsigned int)(((unsigned long)upper_32_bits(addr))); - eax = (unsigned int)(((unsigned long)lower_32_bits(addr))); - - asm volatile("movl %0, %%ecx; wrmsr" : - : "i" (0xc0010020), "a" (eax), "d" (edx) : "ecx"); - + wrmsrl(MSR_AMD64_PATCH_LOADER, &mc_amd->hdr.data_code); /* get patch id after patching */ - asm volatile("movl %1, %%ecx; rdmsr" - : "=a" (rev) - : "i" (0x0000008B) : "ecx"); - + rdmsr(MSR_AMD64_PATCH_LEVEL, rev, dummy); spin_unlock_irqrestore(µcode_update_lock, flags); /* check current patch id and patch's id for match */ -- cgit v1.2.3 From 6cc9b6d94b6fee23b0671970f67d297fa76b68b3 Mon Sep 17 00:00:00 2001 From: Andreas Herrmann Date: Tue, 16 Dec 2008 19:17:45 +0100 Subject: x86: microcode_amd: consolidate macro definitions Impact: cleanup Signed-off-by: Andreas Herrmann Signed-off-by: Ingo Molnar --- arch/x86/kernel/microcode_amd.c | 16 +++------------- 1 file changed, 3 insertions(+), 13 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/microcode_amd.c b/arch/x86/kernel/microcode_amd.c index 2856955ddab1..e68e723490a3 100644 --- a/arch/x86/kernel/microcode_amd.c +++ b/arch/x86/kernel/microcode_amd.c @@ -75,15 +75,9 @@ struct microcode_amd { unsigned int mpb[0]; }; -#define UCODE_MAX_SIZE (2048) -#define DEFAULT_UCODE_DATASIZE (896) -#define MC_HEADER_SIZE (sizeof(struct microcode_header_amd)) -#define DEFAULT_UCODE_TOTALSIZE (DEFAULT_UCODE_DATASIZE + MC_HEADER_SIZE) -#define DWSIZE (sizeof(u32)) -/* For now we support a fixed ucode total size only */ -#define get_totalsize(mc) \ - ((((struct microcode_amd *)mc)->hdr.mc_patch_data_len * 28) \ - + MC_HEADER_SIZE) +#define UCODE_MAX_SIZE 2048 +#define UCODE_CONTAINER_SECTION_HDR 8 +#define UCODE_CONTAINER_HEADER_SIZE 12 /* serialize access to the physical write */ static DEFINE_SPINLOCK(microcode_update_lock); @@ -222,7 +216,6 @@ static void *get_next_ucode(const u8 *buf, unsigned int size, unsigned int *mc_size) { unsigned int total_size; -#define UCODE_CONTAINER_SECTION_HDR 8 u8 section_hdr[UCODE_CONTAINER_SECTION_HDR]; void *mc; @@ -255,14 +248,12 @@ static void *get_next_ucode(const u8 *buf, unsigned int size, } else *mc_size = total_size + UCODE_CONTAINER_SECTION_HDR; } -#undef UCODE_CONTAINER_SECTION_HDR return mc; } static int install_equiv_cpu_table(const u8 *buf) { -#define UCODE_CONTAINER_HEADER_SIZE 12 u8 *container_hdr[UCODE_CONTAINER_HEADER_SIZE]; unsigned int *buf_pos = (unsigned int *)container_hdr; unsigned long size; @@ -291,7 +282,6 @@ static int install_equiv_cpu_table(const u8 *buf) } return size + UCODE_CONTAINER_HEADER_SIZE; /* add header length */ -#undef UCODE_CONTAINER_HEADER_SIZE } static void free_equiv_cpu_table(void) -- cgit v1.2.3 From 98415301ea2dd389539ab429bcfa9da07219eabc Mon Sep 17 00:00:00 2001 From: Andreas Herrmann Date: Tue, 16 Dec 2008 19:20:21 +0100 Subject: x86: microcode_amd: remove (wrong) chipset deivce ID checks Impact: remove dead/incorrect code Currently there is no chipset specific ucode. The checks are incorrect anyway (e.g. pci device IDs are 16 bit and not 8 bit). Thus I remove the stuff for the time being and will reintroduce it if it's foreseeable that it is really needed. Signed-off-by: Andreas Herrmann Signed-off-by: Ingo Molnar --- arch/x86/kernel/microcode_amd.c | 32 +++++--------------------------- 1 file changed, 5 insertions(+), 27 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/microcode_amd.c b/arch/x86/kernel/microcode_amd.c index e68e723490a3..2e8af6ef3da9 100644 --- a/arch/x86/kernel/microcode_amd.c +++ b/arch/x86/kernel/microcode_amd.c @@ -108,7 +108,6 @@ static int collect_cpu_info_amd(int cpu, struct cpu_signature *csig) static int get_matching_microcode(int cpu, void *mc, int rev) { struct microcode_header_amd *mc_header = mc; - struct pci_dev *nb_pci_dev, *sb_pci_dev; unsigned int current_cpu_id; unsigned int equiv_cpu_id = 0x00; unsigned int i = 0; @@ -137,32 +136,11 @@ static int get_matching_microcode(int cpu, void *mc, int rev) return 0; } - /* ucode may be northbridge specific */ - if (mc_header->nb_dev_id) { - nb_pci_dev = pci_get_device(PCI_VENDOR_ID_AMD, - (mc_header->nb_dev_id & 0xff), - NULL); - if ((!nb_pci_dev) || - (mc_header->nb_rev_id != nb_pci_dev->revision)) { - printk(KERN_ERR "microcode: CPU%d NB mismatch\n", cpu); - pci_dev_put(nb_pci_dev); - return 0; - } - pci_dev_put(nb_pci_dev); - } - - /* ucode may be southbridge specific */ - if (mc_header->sb_dev_id) { - sb_pci_dev = pci_get_device(PCI_VENDOR_ID_AMD, - (mc_header->sb_dev_id & 0xff), - NULL); - if ((!sb_pci_dev) || - (mc_header->sb_rev_id != sb_pci_dev->revision)) { - printk(KERN_ERR "microcode: CPU%d SB mismatch\n", cpu); - pci_dev_put(sb_pci_dev); - return 0; - } - pci_dev_put(sb_pci_dev); + /* ucode might be chipset specific -- currently we don't support this */ + if (mc_header->nb_dev_id || mc_header->sb_dev_id) { + printk(KERN_WARNING "microcode: CPU%d loading of chipset " + "specific code not yet supported\n", cpu); + return 0; } if (mc_header->patch_id <= rev) -- cgit v1.2.3 From 5549b94bc74c3e7edd44e0aeb7d9f773e82d2d20 Mon Sep 17 00:00:00 2001 From: Andreas Herrmann Date: Tue, 16 Dec 2008 19:21:30 +0100 Subject: x86: microcode_amd: use 'packed' attribute for structs Impact: cleanup Signed-off-by: Andreas Herrmann Signed-off-by: Ingo Molnar --- arch/x86/kernel/microcode_amd.c | 45 +++++++++++++++++++++-------------------- 1 file changed, 23 insertions(+), 22 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/microcode_amd.c b/arch/x86/kernel/microcode_amd.c index 2e8af6ef3da9..e1ce650f276b 100644 --- a/arch/x86/kernel/microcode_amd.c +++ b/arch/x86/kernel/microcode_amd.c @@ -47,28 +47,29 @@ MODULE_LICENSE("GPL v2"); #define UCODE_UCODE_TYPE 0x00000001 struct equiv_cpu_entry { - unsigned int installed_cpu; - unsigned int fixed_errata_mask; - unsigned int fixed_errata_compare; - unsigned int equiv_cpu; -}; + u32 installed_cpu; + u32 fixed_errata_mask; + u32 fixed_errata_compare; + u16 equiv_cpu; + u16 res; +} __attribute__((packed)); struct microcode_header_amd { - unsigned int data_code; - unsigned int patch_id; - unsigned char mc_patch_data_id[2]; - unsigned char mc_patch_data_len; - unsigned char init_flag; - unsigned int mc_patch_data_checksum; - unsigned int nb_dev_id; - unsigned int sb_dev_id; - u16 processor_rev_id; - unsigned char nb_rev_id; - unsigned char sb_rev_id; - unsigned char bios_api_rev; - unsigned char reserved1[3]; - unsigned int match_reg[8]; -}; + u32 data_code; + u32 patch_id; + u16 mc_patch_data_id; + u8 mc_patch_data_len; + u8 init_flag; + u32 mc_patch_data_checksum; + u32 nb_dev_id; + u32 sb_dev_id; + u16 processor_rev_id; + u8 nb_rev_id; + u8 sb_rev_id; + u8 bios_api_rev; + u8 reserved1[3]; + u32 match_reg[8]; +} __attribute__((packed)); struct microcode_amd { struct microcode_header_amd hdr; @@ -109,7 +110,7 @@ static int get_matching_microcode(int cpu, void *mc, int rev) { struct microcode_header_amd *mc_header = mc; unsigned int current_cpu_id; - unsigned int equiv_cpu_id = 0x00; + u16 equiv_cpu_id = 0; unsigned int i = 0; BUG_ON(equiv_cpu_table == NULL); @@ -117,7 +118,7 @@ static int get_matching_microcode(int cpu, void *mc, int rev) while (equiv_cpu_table[i].installed_cpu != 0) { if (current_cpu_id == equiv_cpu_table[i].installed_cpu) { - equiv_cpu_id = equiv_cpu_table[i].equiv_cpu & 0xffff; + equiv_cpu_id = equiv_cpu_table[i].equiv_cpu; break; } i++; -- cgit v1.2.3 From df23cab563912ba43f7e9bc8ac517e5a2ddc9cd2 Mon Sep 17 00:00:00 2001 From: Andreas Herrmann Date: Tue, 16 Dec 2008 19:22:36 +0100 Subject: x86: microcode_amd: modify log messages Impact: change microcode printk content Change log level and provide (at least I tried to;-) consistent, short, meaningful content. Signed-off-by: Andreas Herrmann Cc: Dmitry Adamushko Signed-off-by: Ingo Molnar --- arch/x86/kernel/microcode_amd.c | 58 ++++++++++++++++++----------------------- 1 file changed, 26 insertions(+), 32 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/microcode_amd.c b/arch/x86/kernel/microcode_amd.c index e1ce650f276b..24c256f4e50a 100644 --- a/arch/x86/kernel/microcode_amd.c +++ b/arch/x86/kernel/microcode_amd.c @@ -91,18 +91,13 @@ static int collect_cpu_info_amd(int cpu, struct cpu_signature *csig) u32 dummy; memset(csig, 0, sizeof(*csig)); - if (c->x86_vendor != X86_VENDOR_AMD || c->x86 < 0x10) { - printk(KERN_ERR "microcode: CPU%d not a capable AMD processor\n", - cpu); + printk(KERN_WARNING "microcode: CPU%d: AMD CPU family 0x%x not " + "supported\n", cpu, c->x86); return -1; } - rdmsr(MSR_AMD64_PATCH_LEVEL, csig->rev, dummy); - - printk(KERN_INFO "microcode: collect_cpu_info_amd : patch_id=0x%x\n", - csig->rev); - + printk(KERN_INFO "microcode: CPU%d: patch_level=0x%x\n", cpu, csig->rev); return 0; } @@ -125,21 +120,21 @@ static int get_matching_microcode(int cpu, void *mc, int rev) } if (!equiv_cpu_id) { - printk(KERN_ERR "microcode: CPU%d cpu_id " - "not found in equivalent cpu table\n", cpu); + printk(KERN_WARNING "microcode: CPU%d: cpu revision " + "not listed in equivalent cpu table\n", cpu); return 0; } if (mc_header->processor_rev_id != equiv_cpu_id) { - printk(KERN_ERR "microcode: CPU%d patch does not match " - "(processor_rev_id: %x, eqiv_cpu_id: %x)\n", + printk(KERN_ERR "microcode: CPU%d: patch mismatch " + "(processor_rev_id: %x, equiv_cpu_id: %x)\n", cpu, mc_header->processor_rev_id, equiv_cpu_id); return 0; } /* ucode might be chipset specific -- currently we don't support this */ if (mc_header->nb_dev_id || mc_header->sb_dev_id) { - printk(KERN_WARNING "microcode: CPU%d loading of chipset " + printk(KERN_ERR "microcode: CPU%d: loading of chipset " "specific code not yet supported\n", cpu); return 0; } @@ -172,15 +167,13 @@ static void apply_microcode_amd(int cpu) /* check current patch id and patch's id for match */ if (rev != mc_amd->hdr.patch_id) { - printk(KERN_ERR "microcode: CPU%d update from revision " - "0x%x to 0x%x failed\n", cpu_num, - mc_amd->hdr.patch_id, rev); + printk(KERN_ERR "microcode: CPU%d: update failed " + "(for patch_level=0x%x)\n", cpu, mc_amd->hdr.patch_id); return; } - printk(KERN_INFO "microcode: CPU%d updated from revision " - "0x%x to 0x%x\n", - cpu_num, uci->cpu_sig.rev, mc_amd->hdr.patch_id); + printk(KERN_INFO "microcode: CPU%d: updated (new patch_level=0x%x)\n", + cpu, rev); uci->cpu_sig.rev = rev; } @@ -202,18 +195,18 @@ static void *get_next_ucode(const u8 *buf, unsigned int size, return NULL; if (section_hdr[0] != UCODE_UCODE_TYPE) { - printk(KERN_ERR "microcode: error! " - "Wrong microcode payload type field\n"); + printk(KERN_ERR "microcode: error: invalid type field in " + "container file section header\n"); return NULL; } total_size = (unsigned long) (section_hdr[4] + (section_hdr[5] << 8)); - printk(KERN_INFO "microcode: size %u, total_size %u\n", - size, total_size); + printk(KERN_DEBUG "microcode: size %u, total_size %u\n", + size, total_size); if (total_size > size || total_size > UCODE_MAX_SIZE) { - printk(KERN_ERR "microcode: error! Bad data in microcode data file\n"); + printk(KERN_ERR "microcode: error: size mismatch\n"); return NULL; } @@ -243,14 +236,15 @@ static int install_equiv_cpu_table(const u8 *buf) size = buf_pos[2]; if (buf_pos[1] != UCODE_EQUIV_CPU_TABLE_TYPE || !size) { - printk(KERN_ERR "microcode: error! " - "Wrong microcode equivalent cpu table\n"); + printk(KERN_ERR "microcode: error: invalid type field in " + "container file section header\n"); return 0; } equiv_cpu_table = (struct equiv_cpu_entry *) vmalloc(size); if (!equiv_cpu_table) { - printk(KERN_ERR "microcode: error, can't allocate memory for equiv CPU table\n"); + printk(KERN_ERR "microcode: failed to allocate " + "equivalent CPU table\n"); return 0; } @@ -283,7 +277,8 @@ static int generic_load_microcode(int cpu, const u8 *data, size_t size) offset = install_equiv_cpu_table(ucode_ptr); if (!offset) { - printk(KERN_ERR "microcode: installing equivalent cpu table failed\n"); + printk(KERN_ERR "microcode: failed to create " + "equivalent cpu table\n"); return -EINVAL; } @@ -339,8 +334,7 @@ static int request_microcode_fw(int cpu, struct device *device) ret = request_firmware(&firmware, fw_name, device); if (ret) { - printk(KERN_ERR "microcode: ucode data file %s load failed\n", - fw_name); + printk(KERN_ERR "microcode: failed to load file %s\n", fw_name); return ret; } @@ -353,8 +347,8 @@ static int request_microcode_fw(int cpu, struct device *device) static int request_microcode_user(int cpu, const void __user *buf, size_t size) { - printk(KERN_WARNING "microcode: AMD microcode update via " - "/dev/cpu/microcode is not supported\n"); + printk(KERN_INFO "microcode: AMD microcode update via " + "/dev/cpu/microcode not supported\n"); return -1; } -- cgit v1.2.3 From bacbe9994541c70aa3abd1a013ac738e58d4bfb2 Mon Sep 17 00:00:00 2001 From: Janne Kulmala Date: Tue, 16 Dec 2008 13:39:57 +0200 Subject: x86: enable HPET on Fujitsu u9200 Impact: auto-enable HPET on Fujitsu u9200 HPET timer is listed in the ACPI table, but needs a quirk entry in order to work. Unfortunately, the quirk code runs after first HPET hpet_enable() which has already determined that the timer doesn't work (reads 0xFFFFFFFF). This patch allows hpet_enable() to be called again after running the quirk code. Signed-off-by: Janne Kulmala Signed-off-by: Ingo Molnar --- arch/x86/kernel/hpet.c | 2 +- arch/x86/kernel/quirks.c | 2 ++ 2 files changed, 3 insertions(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/hpet.c b/arch/x86/kernel/hpet.c index 067d8de913f6..84089dc8fd1d 100644 --- a/arch/x86/kernel/hpet.c +++ b/arch/x86/kernel/hpet.c @@ -811,7 +811,7 @@ int __init hpet_enable(void) out_nohpet: hpet_clear_mapping(); - boot_hpet_disable = 1; + hpet_address = 0; return 0; } diff --git a/arch/x86/kernel/quirks.c b/arch/x86/kernel/quirks.c index 67465ed89310..309949e9e1c1 100644 --- a/arch/x86/kernel/quirks.c +++ b/arch/x86/kernel/quirks.c @@ -168,6 +168,8 @@ DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_ICH7_31, ich_force_enable_hpet); DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_ICH8_1, ich_force_enable_hpet); +DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_ICH8_4, + ich_force_enable_hpet); DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_ICH9_7, ich_force_enable_hpet); -- cgit v1.2.3 From d4377974062122d6d9be0bbd8a910a0954714194 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Tue, 16 Dec 2008 20:59:24 +0100 Subject: x86: support always running TSC on Intel CPUs, add cpufeature definition Impact: add new synthetic-cpuid bit definition add X86_FEATURE_NONSTOP_TSC to the cpufeature bits - this is in preparation of Venki's always-running-TSC patch. Signed-off-by: Ingo Molnar --- arch/x86/include/asm/cpufeature.h | 1 + 1 file changed, 1 insertion(+) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/cpufeature.h b/arch/x86/include/asm/cpufeature.h index 5bce8ed02b44..ea408dcba513 100644 --- a/arch/x86/include/asm/cpufeature.h +++ b/arch/x86/include/asm/cpufeature.h @@ -92,6 +92,7 @@ #define X86_FEATURE_AMDC1E (3*32+21) /* AMD C1E detected */ #define X86_FEATURE_XTOPOLOGY (3*32+22) /* cpu topology enum extensions */ #define X86_FEATURE_TSC_RELIABLE (3*32+23) /* TSC is known to be reliable */ +#define X86_FEATURE_NONSTOP_TSC (3*32+24) /* TSC does not stop in C states */ /* Intel-defined CPU features, CPUID level 0x00000001 (ecx), word 4 */ #define X86_FEATURE_XMM3 (4*32+ 0) /* "pni" SSE-3 */ -- cgit v1.2.3 From 40fb17152c50a69dc304dd632131c2f41281ce44 Mon Sep 17 00:00:00 2001 From: Venki Pallipadi Date: Mon, 17 Nov 2008 16:11:37 -0800 Subject: x86: support always running TSC on Intel CPUs Impact: reward non-stop TSCs with good TSC-based clocksources, etc. Add support for CPUID_0x80000007_Bit8 on Intel CPUs as well. This bit means that the TSC is invariant with C/P/T states and always runs at constant frequency. With Intel CPUs, we have 3 classes * CPUs where TSC runs at constant rate and does not stop n C-states * CPUs where TSC runs at constant rate, but will stop in deep C-states * CPUs where TSC rate will vary based on P/T-states and TSC will stop in deep C-states. To cover these 3, one feature bit (CONSTANT_TSC) is not enough. So, add a second bit (NONSTOP_TSC). CONSTANT_TSC indicates that the TSC runs at constant frequency irrespective of P/T-states, and NONSTOP_TSC indicates that TSC does not stop in deep C-states. CPUID_0x8000000_Bit8 indicates both these feature bit can be set. We still have CONSTANT_TSC _set_ and NONSTOP_TSC _not_set_ on some older Intel CPUs, based on model checks. We can use TSC on such CPUs for time, as long as those CPUs do not support/enter deep C-states. Signed-off-by: Venkatesh Pallipadi Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/amd.c | 9 +++++++-- arch/x86/kernel/cpu/intel.c | 10 ++++++++++ arch/x86/kernel/process.c | 2 +- drivers/acpi/processor_idle.c | 6 +++--- 4 files changed, 21 insertions(+), 6 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c index 8f1e31db2ad5..7c878f6aa919 100644 --- a/arch/x86/kernel/cpu/amd.c +++ b/arch/x86/kernel/cpu/amd.c @@ -283,9 +283,14 @@ static void __cpuinit early_init_amd(struct cpuinfo_x86 *c) { early_init_amd_mc(c); - /* c->x86_power is 8000_0007 edx. Bit 8 is constant TSC */ - if (c->x86_power & (1<<8)) + /* + * c->x86_power is 8000_0007 edx. Bit 8 is TSC runs at constant rate + * with P/T states and does not stop in deep C-states + */ + if (c->x86_power & (1 << 8)) { set_cpu_cap(c, X86_FEATURE_CONSTANT_TSC); + set_cpu_cap(c, X86_FEATURE_NONSTOP_TSC); + } #ifdef CONFIG_X86_64 set_cpu_cap(c, X86_FEATURE_SYSCALL32); diff --git a/arch/x86/kernel/cpu/intel.c b/arch/x86/kernel/cpu/intel.c index cce0b6118d55..caec59437a22 100644 --- a/arch/x86/kernel/cpu/intel.c +++ b/arch/x86/kernel/cpu/intel.c @@ -41,6 +41,16 @@ static void __cpuinit early_init_intel(struct cpuinfo_x86 *c) if (c->x86 == 15 && c->x86_cache_alignment == 64) c->x86_cache_alignment = 128; #endif + + /* + * c->x86_power is 8000_0007 edx. Bit 8 is TSC runs at constant rate + * with P/T states and does not stop in deep C-states + */ + if (c->x86_power & (1 << 8)) { + set_cpu_cap(c, X86_FEATURE_CONSTANT_TSC); + set_cpu_cap(c, X86_FEATURE_NONSTOP_TSC); + } + } #ifdef CONFIG_X86_32 diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c index c622772744d8..18c70fedba32 100644 --- a/arch/x86/kernel/process.c +++ b/arch/x86/kernel/process.c @@ -270,7 +270,7 @@ static void c1e_idle(void) rdmsr(MSR_K8_INT_PENDING_MSG, lo, hi); if (lo & K8_INTP_C1E_ACTIVE_MASK) { c1e_detected = 1; - if (!boot_cpu_has(X86_FEATURE_CONSTANT_TSC)) + if (!boot_cpu_has(X86_FEATURE_NONSTOP_TSC)) mark_tsc_unstable("TSC halt in AMD C1E"); printk(KERN_INFO "System has AMD C1E enabled\n"); set_cpu_cap(&boot_cpu_data, X86_FEATURE_AMDC1E); diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c index 5f8d746a9b81..38aca048e951 100644 --- a/drivers/acpi/processor_idle.c +++ b/drivers/acpi/processor_idle.c @@ -374,15 +374,15 @@ static int tsc_halts_in_c(int state) { switch (boot_cpu_data.x86_vendor) { case X86_VENDOR_AMD: + case X86_VENDOR_INTEL: /* * AMD Fam10h TSC will tick in all * C/P/S0/S1 states when this bit is set. */ - if (boot_cpu_has(X86_FEATURE_CONSTANT_TSC)) + if (boot_cpu_has(X86_FEATURE_NONSTOP_TSC)) return 0; + /*FALL THROUGH*/ - case X86_VENDOR_INTEL: - /* Several cases known where TSC halts in C2 too */ default: return state > ACPI_STATE_C1; } -- cgit v1.2.3 From f63c2f248959366cd11bfa476f866737047cf663 Mon Sep 17 00:00:00 2001 From: Tej Date: Tue, 16 Dec 2008 11:56:06 -0800 Subject: xen: whitespace/checkpatch cleanup Impact: cleanup Signed-off-by: Tej Signed-off-by: Jeremy Fitzhardinge Signed-off-by: Ingo Molnar --- arch/x86/xen/enlighten.c | 16 +++++++++------- arch/x86/xen/mmu.c | 17 ++++++++++------- arch/x86/xen/multicalls.c | 2 +- arch/x86/xen/setup.c | 9 +++++---- 4 files changed, 25 insertions(+), 19 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c index 5e4686d70f62..86cd2f829683 100644 --- a/arch/x86/xen/enlighten.c +++ b/arch/x86/xen/enlighten.c @@ -793,7 +793,7 @@ static int xen_write_msr_safe(unsigned int msr, unsigned low, unsigned high) ret = 0; - switch(msr) { + switch (msr) { #ifdef CONFIG_X86_64 unsigned which; u64 base; @@ -1453,7 +1453,7 @@ static __init void xen_map_identity_early(pmd_t *pmd, unsigned long max_pfn) ident_pte = 0; pfn = 0; - for(pmdidx = 0; pmdidx < PTRS_PER_PMD && pfn < max_pfn; pmdidx++) { + for (pmdidx = 0; pmdidx < PTRS_PER_PMD && pfn < max_pfn; pmdidx++) { pte_t *pte_page; /* Reuse or allocate a page of ptes */ @@ -1471,7 +1471,7 @@ static __init void xen_map_identity_early(pmd_t *pmd, unsigned long max_pfn) } /* Install mappings */ - for(pteidx = 0; pteidx < PTRS_PER_PTE; pteidx++, pfn++) { + for (pteidx = 0; pteidx < PTRS_PER_PTE; pteidx++, pfn++) { pte_t pte; if (pfn > max_pfn_mapped) @@ -1485,7 +1485,7 @@ static __init void xen_map_identity_early(pmd_t *pmd, unsigned long max_pfn) } } - for(pteidx = 0; pteidx < ident_pte; pteidx += PTRS_PER_PTE) + for (pteidx = 0; pteidx < ident_pte; pteidx += PTRS_PER_PTE) set_page_prot(&level1_ident_pgt[pteidx], PAGE_KERNEL_RO); set_page_prot(pmd, PAGE_KERNEL_RO); @@ -1499,7 +1499,7 @@ static void convert_pfn_mfn(void *v) /* All levels are converted the same way, so just treat them as ptes. */ - for(i = 0; i < PTRS_PER_PTE; i++) + for (i = 0; i < PTRS_PER_PTE; i++) pte[i] = xen_make_pte(pte[i].pte); } @@ -1514,7 +1514,8 @@ static void convert_pfn_mfn(void *v) * of the physical mapping once some sort of allocator has been set * up. */ -static __init pgd_t *xen_setup_kernel_pagetable(pgd_t *pgd, unsigned long max_pfn) +static __init pgd_t *xen_setup_kernel_pagetable(pgd_t *pgd, + unsigned long max_pfn) { pud_t *l3; pmd_t *l2; @@ -1577,7 +1578,8 @@ static __init pgd_t *xen_setup_kernel_pagetable(pgd_t *pgd, unsigned long max_pf #else /* !CONFIG_X86_64 */ static pmd_t level2_kernel_pgt[PTRS_PER_PMD] __page_aligned_bss; -static __init pgd_t *xen_setup_kernel_pagetable(pgd_t *pgd, unsigned long max_pfn) +static __init pgd_t *xen_setup_kernel_pagetable(pgd_t *pgd, + unsigned long max_pfn) { pmd_t *kernel_pmd; diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c index 636ef4caa52d..773d68d3e912 100644 --- a/arch/x86/xen/mmu.c +++ b/arch/x86/xen/mmu.c @@ -154,13 +154,13 @@ void xen_setup_mfn_list_list(void) { unsigned pfn, idx; - for(pfn = 0; pfn < MAX_DOMAIN_PAGES; pfn += P2M_ENTRIES_PER_PAGE) { + for (pfn = 0; pfn < MAX_DOMAIN_PAGES; pfn += P2M_ENTRIES_PER_PAGE) { unsigned topidx = p2m_top_index(pfn); p2m_top_mfn[topidx] = virt_to_mfn(p2m_top[topidx]); } - for(idx = 0; idx < ARRAY_SIZE(p2m_top_mfn_list); idx++) { + for (idx = 0; idx < ARRAY_SIZE(p2m_top_mfn_list); idx++) { unsigned topidx = idx * P2M_ENTRIES_PER_PAGE; p2m_top_mfn_list[idx] = virt_to_mfn(&p2m_top_mfn[topidx]); } @@ -179,7 +179,7 @@ void __init xen_build_dynamic_phys_to_machine(void) unsigned long max_pfn = min(MAX_DOMAIN_PAGES, xen_start_info->nr_pages); unsigned pfn; - for(pfn = 0; pfn < max_pfn; pfn += P2M_ENTRIES_PER_PAGE) { + for (pfn = 0; pfn < max_pfn; pfn += P2M_ENTRIES_PER_PAGE) { unsigned topidx = p2m_top_index(pfn); p2m_top[topidx] = &mfn_list[pfn]; @@ -207,7 +207,7 @@ static void alloc_p2m(unsigned long **pp, unsigned long *mfnp) p = (void *)__get_free_page(GFP_KERNEL | __GFP_NOFAIL); BUG_ON(p == NULL); - for(i = 0; i < P2M_ENTRIES_PER_PAGE; i++) + for (i = 0; i < P2M_ENTRIES_PER_PAGE; i++) p[i] = INVALID_P2M_ENTRY; if (cmpxchg(pp, p2m_missing, p) != p2m_missing) @@ -407,7 +407,8 @@ out: preempt_enable(); } -pte_t xen_ptep_modify_prot_start(struct mm_struct *mm, unsigned long addr, pte_t *ptep) +pte_t xen_ptep_modify_prot_start(struct mm_struct *mm, + unsigned long addr, pte_t *ptep) { /* Just return the pte as-is. We preserve the bits on commit */ return *ptep; @@ -878,7 +879,8 @@ static void __xen_pgd_pin(struct mm_struct *mm, pgd_t *pgd) if (user_pgd) { xen_pin_page(mm, virt_to_page(user_pgd), PT_PGD); - xen_do_pin(MMUEXT_PIN_L4_TABLE, PFN_DOWN(__pa(user_pgd))); + xen_do_pin(MMUEXT_PIN_L4_TABLE, + PFN_DOWN(__pa(user_pgd))); } } #else /* CONFIG_X86_32 */ @@ -993,7 +995,8 @@ static void __xen_pgd_unpin(struct mm_struct *mm, pgd_t *pgd) pgd_t *user_pgd = xen_get_user_pgd(pgd); if (user_pgd) { - xen_do_pin(MMUEXT_UNPIN_TABLE, PFN_DOWN(__pa(user_pgd))); + xen_do_pin(MMUEXT_UNPIN_TABLE, + PFN_DOWN(__pa(user_pgd))); xen_unpin_page(mm, virt_to_page(user_pgd), PT_PGD); } } diff --git a/arch/x86/xen/multicalls.c b/arch/x86/xen/multicalls.c index 8ea8a0d0b0de..c738644b5435 100644 --- a/arch/x86/xen/multicalls.c +++ b/arch/x86/xen/multicalls.c @@ -154,7 +154,7 @@ void xen_mc_flush(void) ret, smp_processor_id()); dump_stack(); for (i = 0; i < b->mcidx; i++) { - printk(" call %2d/%d: op=%lu arg=[%lx] result=%ld\n", + printk(KERN_DEBUG " call %2d/%d: op=%lu arg=[%lx] result=%ld\n", i+1, b->mcidx, b->debug[i].op, b->debug[i].args[0], diff --git a/arch/x86/xen/setup.c b/arch/x86/xen/setup.c index d67901083888..15c6c68db6a2 100644 --- a/arch/x86/xen/setup.c +++ b/arch/x86/xen/setup.c @@ -28,6 +28,9 @@ /* These are code, but not functions. Defined in entry.S */ extern const char xen_hypervisor_callback[]; extern const char xen_failsafe_callback[]; +extern void xen_sysenter_target(void); +extern void xen_syscall_target(void); +extern void xen_syscall32_target(void); /** @@ -110,7 +113,6 @@ static __cpuinit int register_callback(unsigned type, const void *func) void __cpuinit xen_enable_sysenter(void) { - extern void xen_sysenter_target(void); int ret; unsigned sysenter_feature; @@ -132,8 +134,6 @@ void __cpuinit xen_enable_syscall(void) { #ifdef CONFIG_X86_64 int ret; - extern void xen_syscall_target(void); - extern void xen_syscall32_target(void); ret = register_callback(CALLBACKTYPE_syscall, xen_syscall_target); if (ret != 0) { @@ -160,7 +160,8 @@ void __init xen_arch_setup(void) HYPERVISOR_vm_assist(VMASST_CMD_enable, VMASST_TYPE_writable_pagetables); if (!xen_feature(XENFEAT_auto_translated_physmap)) - HYPERVISOR_vm_assist(VMASST_CMD_enable, VMASST_TYPE_pae_extended_cr3); + HYPERVISOR_vm_assist(VMASST_CMD_enable, + VMASST_TYPE_pae_extended_cr3); if (register_callback(CALLBACKTYPE_event, xen_hypervisor_callback) || register_callback(CALLBACKTYPE_failsafe, xen_failsafe_callback)) -- cgit v1.2.3 From aab02f0ae20b8fe0fe891e9f107c6e392256ca01 Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Date: Mon, 15 Dec 2008 22:23:54 +0530 Subject: x86: process_64.c declare __switch_to() and sys_arch_prctl before they get used Impact: cleanup In asm/system.h moved out __switch_to from CONFIG_X86_32 as it is common for both 32 and 64 bit. In asm/pctl.h defined sys_arch_prctl Signed-off-by: Jaswinder Singh Signed-off-by: Ingo Molnar --- arch/x86/include/asm/prctl.h | 3 +++ arch/x86/include/asm/system.h | 4 ++-- 2 files changed, 5 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/prctl.h b/arch/x86/include/asm/prctl.h index fe681147a4f7..a8894647dd9a 100644 --- a/arch/x86/include/asm/prctl.h +++ b/arch/x86/include/asm/prctl.h @@ -6,5 +6,8 @@ #define ARCH_GET_FS 0x1003 #define ARCH_GET_GS 0x1004 +#ifdef CONFIG_X86_64 +extern long sys_arch_prctl(int, unsigned long); +#endif /* CONFIG_X86_64 */ #endif /* _ASM_X86_PRCTL_H */ diff --git a/arch/x86/include/asm/system.h b/arch/x86/include/asm/system.h index 2ed3f0f44ff7..59555f48bf4c 100644 --- a/arch/x86/include/asm/system.h +++ b/arch/x86/include/asm/system.h @@ -17,12 +17,12 @@ # define AT_VECTOR_SIZE_ARCH 1 #endif -#ifdef CONFIG_X86_32 - struct task_struct; /* one of the stranger aspects of C forward declarations */ struct task_struct *__switch_to(struct task_struct *prev, struct task_struct *next); +#ifdef CONFIG_X86_32 + /* * Saving eflags is important. It switches not only IOPL between tasks, * it also protects other tasks from NT leaking through sysenter etc. -- cgit v1.2.3 From 7b5b50f1be9e07714cfaa620d102c8daf3cdd814 Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Date: Mon, 15 Dec 2008 22:24:48 +0530 Subject: x86: signal.c declare do_notify_resume before they get used Impact: cleanup In asm/signal.h moved out do_notify_resume from __i386__ as it is common for both 32 and 64 bit. Signed-off-by: Jaswinder Singh Signed-off-by: Ingo Molnar arch/x86/include/asm/signal.h | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) --- arch/x86/include/asm/signal.h | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/signal.h b/arch/x86/include/asm/signal.h index 96ac44f275da..7761a5d554bb 100644 --- a/arch/x86/include/asm/signal.h +++ b/arch/x86/include/asm/signal.h @@ -121,6 +121,10 @@ typedef unsigned long sigset_t; #ifndef __ASSEMBLY__ +# ifdef __KERNEL__ +extern void do_notify_resume(struct pt_regs *, void *, __u32); +# endif /* __KERNEL__ */ + #ifdef __i386__ # ifdef __KERNEL__ struct old_sigaction { @@ -141,8 +145,6 @@ struct k_sigaction { struct sigaction sa; }; -extern void do_notify_resume(struct pt_regs *, void *, __u32); - # else /* __KERNEL__ */ /* Here we must cater to libcs that poke about in kernel headers. */ -- cgit v1.2.3 From c0195b6da08c4ddd8c8ea830f6c3c40bc7f82071 Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Date: Mon, 15 Dec 2008 22:26:30 +0530 Subject: x86: ldt.c declare sys_modify_ldt before they get used Impact: cleanup In asm/syscalls.h moved out sys_modify_ldt from CONFIG_X86_32 as it is common for both 32 and 64 bit. Signed-off-by: Jaswinder Singh Signed-off-by: Ingo Molnar --- arch/x86/include/asm/syscalls.h | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/syscalls.h b/arch/x86/include/asm/syscalls.h index 87803da44010..75d4a6afc36f 100644 --- a/arch/x86/include/asm/syscalls.h +++ b/arch/x86/include/asm/syscalls.h @@ -19,6 +19,9 @@ /* kernel/ioport.c */ asmlinkage long sys_ioperm(unsigned long, unsigned long, int); +/* kernel/ldt.c */ +asmlinkage int sys_modify_ldt(int, void __user *, unsigned long); + /* X86_32 only */ #ifdef CONFIG_X86_32 /* kernel/process_32.c */ @@ -38,9 +41,6 @@ asmlinkage int sys_rt_sigreturn(unsigned long); /* kernel/ioport.c */ asmlinkage long sys_iopl(unsigned long); -/* kernel/ldt.c */ -asmlinkage int sys_modify_ldt(int, void __user *, unsigned long); - /* kernel/sys_i386_32.c */ asmlinkage long sys_mmap2(unsigned long, unsigned long, unsigned long, unsigned long, unsigned long, unsigned long); -- cgit v1.2.3 From a9b43c7d9890066709609df849959009645c1a19 Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Date: Mon, 15 Dec 2008 23:11:10 +0530 Subject: x86: setup.c find_and_reserve_crashkernel should be static Impact: cleanup, reduce kernel size a bit Signed-off-by: Jaswinder Singh Signed-off-by: Ingo Molnar --- arch/x86/kernel/setup.c | 1 + 1 file changed, 1 insertion(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index 9d5674f7b6cc..81f5d22747ae 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -448,6 +448,7 @@ static void __init reserve_early_setup_data(void) * @size: Size of the crashkernel memory to reserve. * Returns the base address on success, and -1ULL on failure. */ +static unsigned long long __init find_and_reserve_crashkernel(unsigned long long size) { const unsigned long long alignment = 16<<20; /* 16M */ -- cgit v1.2.3 From a79b7a2a758c39315344f0d86b5adb21d90d786e Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Tue, 16 Dec 2008 12:17:25 -0800 Subject: x86: remove unused iommu_nr_pages Impact: cleanup, remove dead code The last usage was removed by the patch set culminating in | commit e3c449f526cebb8d287241c7e82faafd9709668b | Author: Joerg Roedel | Date: Wed Oct 15 22:02:11 2008 -0700 | | x86, AMD IOMMU: convert driver to generic iommu_num_pages function Signed-off-by: Ian Campbell Signed-off-by: Jeremy Fitzhardinge Signed-off-by: Ingo Molnar --- arch/x86/include/asm/iommu.h | 2 -- arch/x86/kernel/pci-dma.c | 7 ------- 2 files changed, 9 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/iommu.h b/arch/x86/include/asm/iommu.h index 0b500c5b6446..35276ec5925b 100644 --- a/arch/x86/include/asm/iommu.h +++ b/arch/x86/include/asm/iommu.h @@ -7,8 +7,6 @@ extern struct dma_mapping_ops nommu_dma_ops; extern int force_iommu, no_iommu; extern int iommu_detected; -extern unsigned long iommu_nr_pages(unsigned long addr, unsigned long len); - /* 10 seconds */ #define DMAR_OPERATION_TIMEOUT ((cycles_t) tsc_khz*10*1000) diff --git a/arch/x86/kernel/pci-dma.c b/arch/x86/kernel/pci-dma.c index 192624820217..e150ad4f0ccc 100644 --- a/arch/x86/kernel/pci-dma.c +++ b/arch/x86/kernel/pci-dma.c @@ -125,13 +125,6 @@ void __init pci_iommu_alloc(void) pci_swiotlb_init(); } -unsigned long iommu_nr_pages(unsigned long addr, unsigned long len) -{ - unsigned long size = roundup((addr & ~PAGE_MASK) + len, PAGE_SIZE); - - return size >> PAGE_SHIFT; -} -EXPORT_SYMBOL(iommu_nr_pages); #endif void *dma_generic_alloc_coherent(struct device *dev, size_t size, -- cgit v1.2.3 From ecbf29cdb3990c83d90d0c4187c89fb2ce423367 Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Tue, 16 Dec 2008 12:37:07 -0800 Subject: xen: clean up asm/xen/hypervisor.h Impact: cleanup hypervisor.h had accumulated a lot of crud, including lots of spurious #includes. Clean it all up, and go around fixing up everything else accordingly. Signed-off-by: Jeremy Fitzhardinge Signed-off-by: Ingo Molnar --- arch/x86/include/asm/xen/hypercall.h | 6 ++++++ arch/x86/include/asm/xen/hypervisor.h | 39 +++++++---------------------------- arch/x86/include/asm/xen/page.h | 5 +++++ arch/x86/xen/enlighten.c | 1 + drivers/xen/balloon.c | 4 +++- drivers/xen/features.c | 6 +++++- drivers/xen/grant-table.c | 1 + include/xen/interface/event_channel.h | 2 ++ 8 files changed, 31 insertions(+), 33 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/xen/hypercall.h b/arch/x86/include/asm/xen/hypercall.h index 3f6000d95fe2..5e79ca694326 100644 --- a/arch/x86/include/asm/xen/hypercall.h +++ b/arch/x86/include/asm/xen/hypercall.h @@ -33,8 +33,14 @@ #ifndef _ASM_X86_XEN_HYPERCALL_H #define _ASM_X86_XEN_HYPERCALL_H +#include +#include #include #include +#include + +#include +#include #include #include diff --git a/arch/x86/include/asm/xen/hypervisor.h b/arch/x86/include/asm/xen/hypervisor.h index a38d25ac87d2..81fbd735aec4 100644 --- a/arch/x86/include/asm/xen/hypervisor.h +++ b/arch/x86/include/asm/xen/hypervisor.h @@ -33,39 +33,10 @@ #ifndef _ASM_X86_XEN_HYPERVISOR_H #define _ASM_X86_XEN_HYPERVISOR_H -#include -#include - -#include -#include - -#include -#include -#include -#if defined(__i386__) -# ifdef CONFIG_X86_PAE -# include -# else -# include -# endif -#endif -#include - /* arch/i386/kernel/setup.c */ extern struct shared_info *HYPERVISOR_shared_info; extern struct start_info *xen_start_info; -/* arch/i386/mach-xen/evtchn.c */ -/* Force a proper event-channel callback from Xen. */ -extern void force_evtchn_callback(void); - -/* Turn jiffies into Xen system time. */ -u64 jiffies_to_st(unsigned long jiffies); - - -#define MULTI_UVMFLAGS_INDEX 3 -#define MULTI_UVMDOMID_INDEX 4 - enum xen_domain_type { XEN_NATIVE, XEN_PV_DOMAIN, @@ -74,9 +45,15 @@ enum xen_domain_type { extern enum xen_domain_type xen_domain_type; +#ifdef CONFIG_XEN #define xen_domain() (xen_domain_type != XEN_NATIVE) -#define xen_pv_domain() (xen_domain_type == XEN_PV_DOMAIN) +#else +#define xen_domain() (0) +#endif + +#define xen_pv_domain() (xen_domain() && xen_domain_type == XEN_PV_DOMAIN) +#define xen_hvm_domain() (xen_domain() && xen_domain_type == XEN_HVM_DOMAIN) + #define xen_initial_domain() (xen_pv_domain() && xen_start_info->flags & SIF_INITDOMAIN) -#define xen_hvm_domain() (xen_domain_type == XEN_HVM_DOMAIN) #endif /* _ASM_X86_XEN_HYPERVISOR_H */ diff --git a/arch/x86/include/asm/xen/page.h b/arch/x86/include/asm/xen/page.h index bc628998a1b9..7ef617ef1df3 100644 --- a/arch/x86/include/asm/xen/page.h +++ b/arch/x86/include/asm/xen/page.h @@ -1,11 +1,16 @@ #ifndef _ASM_X86_XEN_PAGE_H #define _ASM_X86_XEN_PAGE_H +#include +#include +#include #include #include +#include #include +#include #include /* Xen machine address */ diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c index 86cd2f829683..bea215230b20 100644 --- a/arch/x86/xen/enlighten.c +++ b/arch/x86/xen/enlighten.c @@ -28,6 +28,7 @@ #include #include +#include #include #include #include diff --git a/drivers/xen/balloon.c b/drivers/xen/balloon.c index 526c191e84ea..8dc7109d61b7 100644 --- a/drivers/xen/balloon.c +++ b/drivers/xen/balloon.c @@ -44,13 +44,15 @@ #include #include -#include #include #include #include #include #include +#include +#include +#include #include #include #include diff --git a/drivers/xen/features.c b/drivers/xen/features.c index 0707714e40d6..99eda169c779 100644 --- a/drivers/xen/features.c +++ b/drivers/xen/features.c @@ -8,7 +8,11 @@ #include #include #include -#include + +#include + +#include +#include #include u8 xen_features[XENFEAT_NR_SUBMAPS * 32] __read_mostly; diff --git a/drivers/xen/grant-table.c b/drivers/xen/grant-table.c index 06592b9da83c..7d8f531fb8e8 100644 --- a/drivers/xen/grant-table.c +++ b/drivers/xen/grant-table.c @@ -40,6 +40,7 @@ #include #include #include +#include #include #include diff --git a/include/xen/interface/event_channel.h b/include/xen/interface/event_channel.h index 919b5bdcb2bd..2090881c3650 100644 --- a/include/xen/interface/event_channel.h +++ b/include/xen/interface/event_channel.h @@ -9,6 +9,8 @@ #ifndef __XEN_PUBLIC_EVENT_CHANNEL_H__ #define __XEN_PUBLIC_EVENT_CHANNEL_H__ +#include + typedef uint32_t evtchn_port_t; DEFINE_GUEST_HANDLE(evtchn_port_t); -- cgit v1.2.3 From 39c04b55240342d0742ac48538d3d8c71bfc0a94 Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Tue, 16 Dec 2008 12:32:23 -0800 Subject: x86: make sure we really have an hpet mapping before using it Impact: prepare the hpet code for Xen dom0 booting When booting in Xen dom0, the hpet isn't really accessible, so make sure the mapping is non-NULL before use. Signed-off-by: Jeremy Fitzhardinge Signed-off-by: Ingo Molnar --- arch/x86/kernel/hpet.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/hpet.c b/arch/x86/kernel/hpet.c index 84089dc8fd1d..a1f6ed5e1a05 100644 --- a/arch/x86/kernel/hpet.c +++ b/arch/x86/kernel/hpet.c @@ -834,10 +834,11 @@ static __init int hpet_late_init(void) hpet_address = force_hpet_address; hpet_enable(); - if (!hpet_virt_address) - return -ENODEV; } + if (!hpet_virt_address) + return -ENODEV; + hpet_reserve_platform_timers(hpet_readl(HPET_ID)); for_each_online_cpu(cpu) { -- cgit v1.2.3 From 8ae936690972dfcad73d0dde1095b9f32af5ee95 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Fri, 12 Dec 2008 15:52:26 -0800 Subject: x86: hardirq: use inc_irq_stat() in non-unified functions Impact: cleanup Replace incrementing irq stat with inc_irq_stat() in non-unified functions. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/mcheck/mce_amd_64.c | 2 +- arch/x86/kernel/cpu/mcheck/mce_intel_64.c | 2 +- arch/x86/kernel/time_32.c | 2 +- arch/x86/kernel/time_64.c | 2 +- arch/x86/kernel/tlb_32.c | 2 +- arch/x86/kernel/tlb_64.c | 2 +- 6 files changed, 6 insertions(+), 6 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd_64.c b/arch/x86/kernel/cpu/mcheck/mce_amd_64.c index 5eb390a4b2e9..748c8f9e7a05 100644 --- a/arch/x86/kernel/cpu/mcheck/mce_amd_64.c +++ b/arch/x86/kernel/cpu/mcheck/mce_amd_64.c @@ -237,7 +237,7 @@ asmlinkage void mce_threshold_interrupt(void) } } out: - add_pda(irq_threshold_count, 1); + inc_irq_stat(irq_threshold_count); irq_exit(); } diff --git a/arch/x86/kernel/cpu/mcheck/mce_intel_64.c b/arch/x86/kernel/cpu/mcheck/mce_intel_64.c index c17eaf5dd6dd..4b48f251fd39 100644 --- a/arch/x86/kernel/cpu/mcheck/mce_intel_64.c +++ b/arch/x86/kernel/cpu/mcheck/mce_intel_64.c @@ -26,7 +26,7 @@ asmlinkage void smp_thermal_interrupt(void) if (therm_throt_process(msr_val & 1)) mce_log_therm_throt_event(smp_processor_id(), msr_val); - add_pda(irq_thermal_count, 1); + inc_irq_stat(irq_thermal_count); irq_exit(); } diff --git a/arch/x86/kernel/time_32.c b/arch/x86/kernel/time_32.c index 77b400f06ea2..65309e4cb1c0 100644 --- a/arch/x86/kernel/time_32.c +++ b/arch/x86/kernel/time_32.c @@ -75,7 +75,7 @@ EXPORT_SYMBOL(profile_pc); irqreturn_t timer_interrupt(int irq, void *dev_id) { /* Keep nmi watchdog up to date */ - per_cpu(irq_stat, smp_processor_id()).irq0_irqs++; + inc_irq_stat(irq0_irqs); #ifdef CONFIG_X86_IO_APIC if (timer_ack) { diff --git a/arch/x86/kernel/time_64.c b/arch/x86/kernel/time_64.c index 418a095c5796..1749cacde8b9 100644 --- a/arch/x86/kernel/time_64.c +++ b/arch/x86/kernel/time_64.c @@ -51,7 +51,7 @@ EXPORT_SYMBOL(profile_pc); irqreturn_t timer_interrupt(int irq, void *dev_id) { - add_pda(irq0_irqs, 1); + inc_irq_stat(irq0_irqs); global_clock_event->event_handler(global_clock_event); diff --git a/arch/x86/kernel/tlb_32.c b/arch/x86/kernel/tlb_32.c index f4049f3513b6..f374f83fca42 100644 --- a/arch/x86/kernel/tlb_32.c +++ b/arch/x86/kernel/tlb_32.c @@ -119,7 +119,7 @@ void smp_invalidate_interrupt(struct pt_regs *regs) smp_mb__after_clear_bit(); out: put_cpu_no_resched(); - __get_cpu_var(irq_stat).irq_tlb_count++; + inc_irq_stat(irq_tlb_count); } void native_flush_tlb_others(const cpumask_t *cpumaskp, struct mm_struct *mm, diff --git a/arch/x86/kernel/tlb_64.c b/arch/x86/kernel/tlb_64.c index 8f919ca69494..29887d7081a9 100644 --- a/arch/x86/kernel/tlb_64.c +++ b/arch/x86/kernel/tlb_64.c @@ -154,7 +154,7 @@ asmlinkage void smp_invalidate_interrupt(struct pt_regs *regs) out: ack_APIC_irq(); cpu_clear(cpu, f->flush_cpumask); - add_pda(irq_tlb_count, 1); + inc_irq_stat(irq_tlb_count); } void native_flush_tlb_others(const cpumask_t *cpumaskp, struct mm_struct *mm, -- cgit v1.2.3 From fde9071167c4624281553b23232aa8b81e71c790 Mon Sep 17 00:00:00 2001 From: Zachary Amsden Date: Fri, 12 Dec 2008 11:26:35 -0800 Subject: x86: clean up dead code in vmi_32.c Impact: cleanup, remove dead debug code I ran across some old debugging code in vmi paravirt-ops code that was already dead, but still potentially useful. After reviewing recent changes to the way kernel page tables are allocated and initialized, and the lack of bugs caught by this debugging code, I've concluded it is now totally useless to have around, and it's already been #if 0'd for quite some time. There's no rush to get this in mainline, but it's also totally harmless, so I'll let the x86 maintainers decide where it should be tucked. I've been out of the mainstream dev loop for a couple months, so apologies if I haven't got any protocol changes in order. Remove mummified remains found in vmi_32.c Signed-off-by: Zachary Amsden Signed-off-by: Ingo Molnar --- arch/x86/kernel/vmi_32.c | 119 ----------------------------------------------- 1 file changed, 119 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/vmi_32.c b/arch/x86/kernel/vmi_32.c index 8b6c393ab9fd..8087e0cd877d 100644 --- a/arch/x86/kernel/vmi_32.c +++ b/arch/x86/kernel/vmi_32.c @@ -266,109 +266,6 @@ static void vmi_nop(void) { } -#ifdef CONFIG_DEBUG_PAGE_TYPE - -#ifdef CONFIG_X86_PAE -#define MAX_BOOT_PTS (2048+4+1) -#else -#define MAX_BOOT_PTS (1024+1) -#endif - -/* - * During boot, mem_map is not yet available in paging_init, so stash - * all the boot page allocations here. - */ -static struct { - u32 pfn; - int type; -} boot_page_allocations[MAX_BOOT_PTS]; -static int num_boot_page_allocations; -static int boot_allocations_applied; - -void vmi_apply_boot_page_allocations(void) -{ - int i; - BUG_ON(!mem_map); - for (i = 0; i < num_boot_page_allocations; i++) { - struct page *page = pfn_to_page(boot_page_allocations[i].pfn); - page->type = boot_page_allocations[i].type; - page->type = boot_page_allocations[i].type & - ~(VMI_PAGE_ZEROED | VMI_PAGE_CLONE); - } - boot_allocations_applied = 1; -} - -static void record_page_type(u32 pfn, int type) -{ - BUG_ON(num_boot_page_allocations >= MAX_BOOT_PTS); - boot_page_allocations[num_boot_page_allocations].pfn = pfn; - boot_page_allocations[num_boot_page_allocations].type = type; - num_boot_page_allocations++; -} - -static void check_zeroed_page(u32 pfn, int type, struct page *page) -{ - u32 *ptr; - int i; - int limit = PAGE_SIZE / sizeof(int); - - if (page_address(page)) - ptr = (u32 *)page_address(page); - else - ptr = (u32 *)__va(pfn << PAGE_SHIFT); - /* - * When cloning the root in non-PAE mode, only the userspace - * pdes need to be zeroed. - */ - if (type & VMI_PAGE_CLONE) - limit = KERNEL_PGD_BOUNDARY; - for (i = 0; i < limit; i++) - BUG_ON(ptr[i]); -} - -/* - * We stash the page type into struct page so we can verify the page - * types are used properly. - */ -static void vmi_set_page_type(u32 pfn, int type) -{ - /* PAE can have multiple roots per page - don't track */ - if (PTRS_PER_PMD > 1 && (type & VMI_PAGE_PDP)) - return; - - if (boot_allocations_applied) { - struct page *page = pfn_to_page(pfn); - if (type != VMI_PAGE_NORMAL) - BUG_ON(page->type); - else - BUG_ON(page->type == VMI_PAGE_NORMAL); - page->type = type & ~(VMI_PAGE_ZEROED | VMI_PAGE_CLONE); - if (type & VMI_PAGE_ZEROED) - check_zeroed_page(pfn, type, page); - } else { - record_page_type(pfn, type); - } -} - -static void vmi_check_page_type(u32 pfn, int type) -{ - /* PAE can have multiple roots per page - skip checks */ - if (PTRS_PER_PMD > 1 && (type & VMI_PAGE_PDP)) - return; - - type &= ~(VMI_PAGE_ZEROED | VMI_PAGE_CLONE); - if (boot_allocations_applied) { - struct page *page = pfn_to_page(pfn); - BUG_ON((page->type ^ type) & VMI_PAGE_PAE); - BUG_ON(type == VMI_PAGE_NORMAL && page->type); - BUG_ON((type & page->type) == 0); - } -} -#else -#define vmi_set_page_type(p,t) do { } while (0) -#define vmi_check_page_type(p,t) do { } while (0) -#endif - #ifdef CONFIG_HIGHPTE static void *vmi_kmap_atomic_pte(struct page *page, enum km_type type) { @@ -395,7 +292,6 @@ static void *vmi_kmap_atomic_pte(struct page *page, enum km_type type) static void vmi_allocate_pte(struct mm_struct *mm, unsigned long pfn) { - vmi_set_page_type(pfn, VMI_PAGE_L1); vmi_ops.allocate_page(pfn, VMI_PAGE_L1, 0, 0, 0); } @@ -406,27 +302,22 @@ static void vmi_allocate_pmd(struct mm_struct *mm, unsigned long pfn) * It is called only for swapper_pg_dir, which already has * data on it. */ - vmi_set_page_type(pfn, VMI_PAGE_L2); vmi_ops.allocate_page(pfn, VMI_PAGE_L2, 0, 0, 0); } static void vmi_allocate_pmd_clone(unsigned long pfn, unsigned long clonepfn, unsigned long start, unsigned long count) { - vmi_set_page_type(pfn, VMI_PAGE_L2 | VMI_PAGE_CLONE); - vmi_check_page_type(clonepfn, VMI_PAGE_L2); vmi_ops.allocate_page(pfn, VMI_PAGE_L2 | VMI_PAGE_CLONE, clonepfn, start, count); } static void vmi_release_pte(unsigned long pfn) { vmi_ops.release_page(pfn, VMI_PAGE_L1); - vmi_set_page_type(pfn, VMI_PAGE_NORMAL); } static void vmi_release_pmd(unsigned long pfn) { vmi_ops.release_page(pfn, VMI_PAGE_L2); - vmi_set_page_type(pfn, VMI_PAGE_NORMAL); } /* @@ -450,26 +341,22 @@ static void vmi_release_pmd(unsigned long pfn) static void vmi_update_pte(struct mm_struct *mm, unsigned long addr, pte_t *ptep) { - vmi_check_page_type(__pa(ptep) >> PAGE_SHIFT, VMI_PAGE_PTE); vmi_ops.update_pte(ptep, vmi_flags_addr(mm, addr, VMI_PAGE_PT, 0)); } static void vmi_update_pte_defer(struct mm_struct *mm, unsigned long addr, pte_t *ptep) { - vmi_check_page_type(__pa(ptep) >> PAGE_SHIFT, VMI_PAGE_PTE); vmi_ops.update_pte(ptep, vmi_flags_addr_defer(mm, addr, VMI_PAGE_PT, 0)); } static void vmi_set_pte(pte_t *ptep, pte_t pte) { /* XXX because of set_pmd_pte, this can be called on PT or PD layers */ - vmi_check_page_type(__pa(ptep) >> PAGE_SHIFT, VMI_PAGE_PTE | VMI_PAGE_PD); vmi_ops.set_pte(pte, ptep, VMI_PAGE_PT); } static void vmi_set_pte_at(struct mm_struct *mm, unsigned long addr, pte_t *ptep, pte_t pte) { - vmi_check_page_type(__pa(ptep) >> PAGE_SHIFT, VMI_PAGE_PTE); vmi_ops.set_pte(pte, ptep, vmi_flags_addr(mm, addr, VMI_PAGE_PT, 0)); } @@ -477,10 +364,8 @@ static void vmi_set_pmd(pmd_t *pmdp, pmd_t pmdval) { #ifdef CONFIG_X86_PAE const pte_t pte = { .pte = pmdval.pmd }; - vmi_check_page_type(__pa(pmdp) >> PAGE_SHIFT, VMI_PAGE_PMD); #else const pte_t pte = { pmdval.pud.pgd.pgd }; - vmi_check_page_type(__pa(pmdp) >> PAGE_SHIFT, VMI_PAGE_PGD); #endif vmi_ops.set_pte(pte, (pte_t *)pmdp, VMI_PAGE_PD); } @@ -502,7 +387,6 @@ static void vmi_set_pte_atomic(pte_t *ptep, pte_t pteval) static void vmi_set_pte_present(struct mm_struct *mm, unsigned long addr, pte_t *ptep, pte_t pte) { - vmi_check_page_type(__pa(ptep) >> PAGE_SHIFT, VMI_PAGE_PTE); vmi_ops.set_pte(pte, ptep, vmi_flags_addr_defer(mm, addr, VMI_PAGE_PT, 1)); } @@ -510,21 +394,18 @@ static void vmi_set_pud(pud_t *pudp, pud_t pudval) { /* Um, eww */ const pte_t pte = { .pte = pudval.pgd.pgd }; - vmi_check_page_type(__pa(pudp) >> PAGE_SHIFT, VMI_PAGE_PGD); vmi_ops.set_pte(pte, (pte_t *)pudp, VMI_PAGE_PDP); } static void vmi_pte_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep) { const pte_t pte = { .pte = 0 }; - vmi_check_page_type(__pa(ptep) >> PAGE_SHIFT, VMI_PAGE_PTE); vmi_ops.set_pte(pte, ptep, vmi_flags_addr(mm, addr, VMI_PAGE_PT, 0)); } static void vmi_pmd_clear(pmd_t *pmd) { const pte_t pte = { .pte = 0 }; - vmi_check_page_type(__pa(pmd) >> PAGE_SHIFT, VMI_PAGE_PMD); vmi_ops.set_pte(pte, (pte_t *)pmd, VMI_PAGE_PD); } #endif -- cgit v1.2.3 From 189f67c4408806563a1f061f5c8bf184a6658477 Mon Sep 17 00:00:00 2001 From: Jack Steiner Date: Fri, 12 Dec 2008 14:50:40 -0600 Subject: x86: UV fix for global physical addresses Impact: fix UV boot crash This fixes a UV bug related to generating global memory addresses on partitioned systems. Partition systems do not have physical memory at address 0. Instead, a chunk of high memory is remapped by the chipset so that it appears to be at address 0. This remapping is INVISIBLE to most of the OS. The only OS functions that need to be aware of the remaping are functions that directly interface to the chipset. The GRU is one example. Also, delete a couple of unused macros related to global memory addresses. Signed-off-by: Jack Steiner Signed-off-by: Ingo Molnar --- arch/x86/include/asm/uv/uv_hub.h | 16 ++-------------- arch/x86/kernel/genx2apic_uv_x.c | 3 +-- 2 files changed, 3 insertions(+), 16 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/uv/uv_hub.h b/arch/x86/include/asm/uv/uv_hub.h index 52aa943c634f..777327ef05c1 100644 --- a/arch/x86/include/asm/uv/uv_hub.h +++ b/arch/x86/include/asm/uv/uv_hub.h @@ -210,7 +210,7 @@ DECLARE_PER_CPU(struct uv_hub_info_s, __uv_hub_info); static inline unsigned long uv_soc_phys_ram_to_gpa(unsigned long paddr) { if (paddr < uv_hub_info->lowmem_remap_top) - paddr += uv_hub_info->lowmem_remap_base; + paddr |= uv_hub_info->lowmem_remap_base; return paddr | uv_hub_info->gnode_upper; } @@ -218,19 +218,7 @@ static inline unsigned long uv_soc_phys_ram_to_gpa(unsigned long paddr) /* socket virtual --> UV global physical address */ static inline unsigned long uv_gpa(void *v) { - return __pa(v) | uv_hub_info->gnode_upper; -} - -/* socket virtual --> UV global physical address */ -static inline void *uv_vgpa(void *v) -{ - return (void *)uv_gpa(v); -} - -/* UV global physical address --> socket virtual */ -static inline void *uv_va(unsigned long gpa) -{ - return __va(gpa & uv_hub_info->gpa_mask); + return uv_soc_phys_ram_to_gpa(__pa(v)); } /* pnode, offset --> socket virtual */ diff --git a/arch/x86/kernel/genx2apic_uv_x.c b/arch/x86/kernel/genx2apic_uv_x.c index 221299f4509f..dece17289731 100644 --- a/arch/x86/kernel/genx2apic_uv_x.c +++ b/arch/x86/kernel/genx2apic_uv_x.c @@ -540,8 +540,7 @@ void __init uv_system_init(void) uv_blade_info[blade].nr_possible_cpus++; uv_cpu_hub_info(cpu)->lowmem_remap_base = lowmem_redir_base; - uv_cpu_hub_info(cpu)->lowmem_remap_top = - lowmem_redir_base + lowmem_redir_size; + uv_cpu_hub_info(cpu)->lowmem_remap_top = lowmem_redir_size; uv_cpu_hub_info(cpu)->m_val = m_val; uv_cpu_hub_info(cpu)->n_val = m_val; uv_cpu_hub_info(cpu)->numa_blade_id = blade; -- cgit v1.2.3 From cf9b303e55da810255638c0b616b1a3f7eda9320 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Mon, 15 Dec 2008 23:33:10 +0100 Subject: x86: re-enable MCE on secondary CPUS after suspend/resume Impact: fix disabled MCE after resume Don't prevent multiple initialization of MCEs. Back from early prehistory mcheck_init() has a reentry check. Presumably that was needed in very old kernels to prevent it entering twice. But as Andreas points out this prevents CPU hotplug (and therefore resume) to correctly reinitialize MCEs when a AP boots again after being offlined. Just drop the check. Reported-by: Andreas Herrmann Signed-off-by: Andi Kleen Tested-by: Andreas Herrmann Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/mcheck/mce_64.c | 3 --- 1 file changed, 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/mcheck/mce_64.c b/arch/x86/kernel/cpu/mcheck/mce_64.c index 4b031a4ac856..1c838032fd37 100644 --- a/arch/x86/kernel/cpu/mcheck/mce_64.c +++ b/arch/x86/kernel/cpu/mcheck/mce_64.c @@ -510,12 +510,9 @@ static void __cpuinit mce_cpu_features(struct cpuinfo_x86 *c) */ void __cpuinit mcheck_init(struct cpuinfo_x86 *c) { - static cpumask_t mce_cpus = CPU_MASK_NONE; - mce_cpu_quirks(c); if (mce_dont_init || - cpu_test_and_set(smp_processor_id(), mce_cpus) || !mce_available(c)) return; -- cgit v1.2.3 From c8182f0016fb65a721c4fbe487909a2d56178135 Mon Sep 17 00:00:00 2001 From: Russ Anderson Date: Fri, 12 Dec 2008 11:07:00 -0600 Subject: sgi-xp: xpc needs to pass the physical address, not virtual Impact: fix crash xpc needs to pass the physical address, not virtual. Testing uncovered this problem. The virtual address happens to work most of the time due to the way bios was masking off the node bits. Passing the physical address makes it work all of the time. Signed-off-by: Russ Anderson Acked-by: Dean Nelson Signed-off-by: Ingo Molnar --- arch/x86/include/asm/uv/bios.h | 2 +- arch/x86/kernel/bios_uv.c | 4 +--- drivers/misc/sgi-xp/xpc_uv.c | 8 ++++---- 3 files changed, 6 insertions(+), 8 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/uv/bios.h b/arch/x86/include/asm/uv/bios.h index da1c4e8e78fc..7ed17ff502b9 100644 --- a/arch/x86/include/asm/uv/bios.h +++ b/arch/x86/include/asm/uv/bios.h @@ -100,7 +100,7 @@ extern s64 uv_bios_call_reentrant(enum uv_bios_cmd, u64, u64, u64, u64, u64); extern s64 uv_bios_get_sn_info(int, int *, long *, long *, long *); extern s64 uv_bios_freq_base(u64, u64 *); -extern int uv_bios_mq_watchlist_alloc(int, void *, unsigned int, +extern int uv_bios_mq_watchlist_alloc(int, unsigned long, unsigned int, unsigned long *); extern int uv_bios_mq_watchlist_free(int, int); extern s64 uv_bios_change_memprotect(u64, u64, enum uv_memprotect); diff --git a/arch/x86/kernel/bios_uv.c b/arch/x86/kernel/bios_uv.c index d22d0f1bbea0..2a0a2a3cac26 100644 --- a/arch/x86/kernel/bios_uv.c +++ b/arch/x86/kernel/bios_uv.c @@ -101,15 +101,13 @@ s64 uv_bios_get_sn_info(int fc, int *uvtype, long *partid, long *coher, } int -uv_bios_mq_watchlist_alloc(int blade, void *mq, unsigned int mq_size, +uv_bios_mq_watchlist_alloc(int blade, unsigned long addr, unsigned int mq_size, unsigned long *intr_mmr_offset) { union uv_watchlist_u size_blade; - unsigned long addr; u64 watchlist; s64 ret; - addr = (unsigned long)mq; size_blade.size = mq_size; size_blade.blade = blade; diff --git a/drivers/misc/sgi-xp/xpc_uv.c b/drivers/misc/sgi-xp/xpc_uv.c index 684b2dd17583..91a55b1b1037 100644 --- a/drivers/misc/sgi-xp/xpc_uv.c +++ b/drivers/misc/sgi-xp/xpc_uv.c @@ -119,16 +119,16 @@ xpc_gru_mq_watchlist_alloc_uv(struct xpc_gru_mq_uv *mq) int ret; #if defined CONFIG_X86_64 - ret = uv_bios_mq_watchlist_alloc(mq->mmr_blade, mq->address, mq->order, - &mq->mmr_offset); + ret = uv_bios_mq_watchlist_alloc(mq->mmr_blade, uv_gpa(mq->address), + mq->order, &mq->mmr_offset); if (ret < 0) { dev_err(xpc_part, "uv_bios_mq_watchlist_alloc() failed, " "ret=%d\n", ret); return ret; } #elif defined CONFIG_IA64_GENERIC || defined CONFIG_IA64_SGI_UV - ret = sn_mq_watchlist_alloc(mq->mmr_blade, mq->address, mq->order, - &mq->mmr_offset); + ret = sn_mq_watchlist_alloc(mq->mmr_blade, uv_gpa(mq->address), + mq->order, &mq->mmr_offset); if (ret < 0) { dev_err(xpc_part, "sn_mq_watchlist_alloc() failed, ret=%d\n", ret); -- cgit v1.2.3 From ae417bb487e3bb88dc862b83b4bf00d87ba67ec8 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Tue, 16 Dec 2008 14:02:16 -0800 Subject: x86: signal: use signal_fault() in sys_sigreturn() Impact: cleanup Call signal_fault() in error route of sys_sigreturn(). Change log level to KERN_EMERG if current is init. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/kernel/signal.c | 15 +++------------ 1 file changed, 3 insertions(+), 12 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/signal.c b/arch/x86/kernel/signal.c index b1cc6da64208..2725a294d734 100644 --- a/arch/x86/kernel/signal.c +++ b/arch/x86/kernel/signal.c @@ -594,17 +594,7 @@ asmlinkage unsigned long sys_sigreturn(unsigned long __unused) return ax; badframe: - if (show_unhandled_signals && printk_ratelimit()) { - printk("%s%s[%d] bad frame in sigreturn frame:" - "%p ip:%lx sp:%lx oeax:%lx", - task_pid_nr(current) > 1 ? KERN_INFO : KERN_EMERG, - current->comm, task_pid_nr(current), frame, regs->ip, - regs->sp, regs->orig_ax); - print_vma_addr(" in ", regs->ip); - printk(KERN_CONT "\n"); - } - - force_sig(SIGSEGV, current); + signal_fault(regs, frame, "sigreturn"); return 0; } @@ -901,8 +891,9 @@ void signal_fault(struct pt_regs *regs, void __user *frame, char *where) struct task_struct *me = current; if (show_unhandled_signals && printk_ratelimit()) { - printk(KERN_INFO + printk("%s" "%s[%d] bad frame in %s frame:%p ip:%lx sp:%lx orax:%lx", + task_pid_nr(current) > 1 ? KERN_INFO : KERN_EMERG, me->comm, me->pid, where, frame, regs->ip, regs->sp, regs->orig_ax); print_vma_addr(" in ", regs->ip); -- cgit v1.2.3 From d0b48ca189523b638d8674fa41e94d1950a17038 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Tue, 16 Dec 2008 14:03:36 -0800 Subject: x86: ia32_signal: use __put_user() instead of __copy_to_user() Impact: cleanup __put_user() can be used for constant size 8, like arch/x86/kernel/signal.c. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/ia32/ia32_signal.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/ia32/ia32_signal.c b/arch/x86/ia32/ia32_signal.c index 1267977e7708..e4f2a5045743 100644 --- a/arch/x86/ia32/ia32_signal.c +++ b/arch/x86/ia32/ia32_signal.c @@ -467,7 +467,7 @@ int ia32_setup_frame(int sig, struct k_sigaction *ka, * These are actually not used anymore, but left because some * gdb versions depend on them as a marker. */ - err |= __copy_to_user(frame->retcode, &code, 8); + err |= __put_user(*((u64 *)&code), (u64 *)frame->retcode); if (err) return -EFAULT; @@ -554,7 +554,7 @@ int ia32_setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, * Not actually used anymore, but left because some gdb * versions need it. */ - err |= __copy_to_user(frame->retcode, &code, 8); + err |= __put_user(*((u64 *)&code), (u64 *)frame->retcode); if (err) return -EFAULT; -- cgit v1.2.3 From 8bee3f0a662ad9c3d6bb705b0530a3b90f089c55 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Tue, 16 Dec 2008 14:04:43 -0800 Subject: x86: ia32_signal: use proper macro __USER32_DS Impact: cleanup Use __USER32_DS instead of __USER_DS in ia32_signal.c. No impact, because __USER32_DS is defined __USER_DS. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/ia32/ia32_signal.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/ia32/ia32_signal.c b/arch/x86/ia32/ia32_signal.c index e4f2a5045743..9c99c429a20d 100644 --- a/arch/x86/ia32/ia32_signal.c +++ b/arch/x86/ia32/ia32_signal.c @@ -396,7 +396,7 @@ static void __user *get_sigframe(struct k_sigaction *ka, struct pt_regs *regs, } /* This is the legacy signal stack switching. */ - else if ((regs->ss & 0xffff) != __USER_DS && + else if ((regs->ss & 0xffff) != __USER32_DS && !(ka->sa.sa_flags & SA_RESTORER) && ka->sa.sa_restorer) sp = (unsigned long) ka->sa.sa_restorer; -- cgit v1.2.3 From c8cae544bba6aee0f5cb0756dbab1a71d2c68737 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Tue, 16 Dec 2008 09:13:11 -0800 Subject: x86: fix build error with post-merge of tip/cpus4096 and rr-for-ingo/master. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Ingo Molnar wrote: > allyes64 build failure: > > arch/x86/kernel/io_apic.c: In function ‘set_ir_ioapic_affinity_irq_desc’: > arch/x86/kernel/io_apic.c:2295: error: incompatible type for argument 2 of > ‘migrate_ioapic_irq_desc’ > arch/x86/kernel/io_apic.c: In function ‘ir_set_msi_irq_affinity’: > arch/x86/kernel/io_apic.c:3205: error: incompatible type for argument 2 of > ‘set_extra_move_desc’ > make[1]: *** wait: No child processes. Stop. Here's a small patch to correct the build error with the post-merge tree. Built and boot-tested. I'll will reset the follow on patches in my brand new git tree to accommodate this change. Fix two references in io_apic.c that were incorrect. Signed-off-by: Mike Travis Signed-off-by: Ingo Molnar --- arch/x86/kernel/io_apic.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index d7f0993b8056..3d7d0d55253f 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -2292,7 +2292,7 @@ static void set_ir_ioapic_affinity_irq_desc(struct irq_desc *desc, return; } - migrate_ioapic_irq_desc(desc, mask); + migrate_ioapic_irq_desc(desc, *mask); } static void set_ir_ioapic_affinity_irq(unsigned int irq, const struct cpumask *mask) @@ -3203,7 +3203,7 @@ static void ir_set_msi_irq_affinity(unsigned int irq, if (assign_irq_vector(irq, cfg, *mask)) return; - set_extra_move_desc(desc, mask); + set_extra_move_desc(desc, *mask); cpumask_and(&tmp, &cfg->domain, mask); dest = cpu_mask_to_apicid(tmp); -- cgit v1.2.3 From 48a1b10aff588833b73994704c47bbd0deb73e9c Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Thu, 11 Dec 2008 00:15:01 -0800 Subject: x86, sparseirq: move irq_desc according to smp_affinity, v7 Impact: improve NUMA handling by migrating irq_desc on smp_affinity changes if CONFIG_NUMA_MIGRATE_IRQ_DESC is set: - make irq_desc to go with affinity aka irq_desc moving etc - call move_irq_desc in irq_complete_move() - legacy irq_desc is not moved, because they are allocated via static array for logical apic mode, need to add move_desc_in_progress_in_same_domain, otherwise it will not be moved ==> also could need two phases to get irq_desc moved. Signed-off-by: Yinghai Lu Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 9 +++ arch/x86/kernel/io_apic.c | 142 +++++++++++++++++++++++++++++++++++++++++++++- include/linux/irq.h | 10 ++++ kernel/irq/Makefile | 1 + kernel/irq/chip.c | 12 +++- kernel/irq/handle.c | 15 +++-- kernel/irq/internals.h | 5 ++ kernel/irq/numa_migrate.c | 127 +++++++++++++++++++++++++++++++++++++++++ 8 files changed, 313 insertions(+), 8 deletions(-) create mode 100644 kernel/irq/numa_migrate.c (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 8943c13502c6..29073532f94c 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -248,6 +248,15 @@ config SPARSE_IRQ If you don't know what to do here, say Y. +config NUMA_MIGRATE_IRQ_DESC + bool "Move irq desc when changing irq smp_affinity" + depends on SPARSE_IRQ && SMP + default n + help + This enables moving irq_desc to cpu/node that irq will use handled. + + If you don't know what to do here, say N. + config X86_FIND_SMP_CONFIG def_bool y depends on X86_MPPARSE || X86_VOYAGER diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index a1a2e070f31a..bfe1245b1a3e 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -141,6 +141,9 @@ struct irq_cfg { unsigned move_cleanup_count; u8 vector; u8 move_in_progress : 1; +#ifdef CONFIG_NUMA_MIGRATE_IRQ_DESC + u8 move_desc_pending : 1; +#endif }; /* irq_cfg is indexed by the sum of all RTEs in all I/O APICs. */ @@ -223,6 +226,121 @@ void arch_init_chip_data(struct irq_desc *desc, int cpu) } } +#ifdef CONFIG_NUMA_MIGRATE_IRQ_DESC + +static void +init_copy_irq_2_pin(struct irq_cfg *old_cfg, struct irq_cfg *cfg, int cpu) +{ + struct irq_pin_list *old_entry, *head, *tail, *entry; + + cfg->irq_2_pin = NULL; + old_entry = old_cfg->irq_2_pin; + if (!old_entry) + return; + + entry = get_one_free_irq_2_pin(cpu); + if (!entry) + return; + + entry->apic = old_entry->apic; + entry->pin = old_entry->pin; + head = entry; + tail = entry; + old_entry = old_entry->next; + while (old_entry) { + entry = get_one_free_irq_2_pin(cpu); + if (!entry) { + entry = head; + while (entry) { + head = entry->next; + kfree(entry); + entry = head; + } + /* still use the old one */ + return; + } + entry->apic = old_entry->apic; + entry->pin = old_entry->pin; + tail->next = entry; + tail = entry; + old_entry = old_entry->next; + } + + tail->next = NULL; + cfg->irq_2_pin = head; +} + +static void free_irq_2_pin(struct irq_cfg *old_cfg, struct irq_cfg *cfg) +{ + struct irq_pin_list *entry, *next; + + if (old_cfg->irq_2_pin == cfg->irq_2_pin) + return; + + entry = old_cfg->irq_2_pin; + + while (entry) { + next = entry->next; + kfree(entry); + entry = next; + } + old_cfg->irq_2_pin = NULL; +} + +void arch_init_copy_chip_data(struct irq_desc *old_desc, + struct irq_desc *desc, int cpu) +{ + struct irq_cfg *cfg; + struct irq_cfg *old_cfg; + + cfg = get_one_free_irq_cfg(cpu); + + if (!cfg) + return; + + desc->chip_data = cfg; + + old_cfg = old_desc->chip_data; + + memcpy(cfg, old_cfg, sizeof(struct irq_cfg)); + + init_copy_irq_2_pin(old_cfg, cfg, cpu); +} + +static void free_irq_cfg(struct irq_cfg *old_cfg) +{ + kfree(old_cfg); +} + +void arch_free_chip_data(struct irq_desc *old_desc, struct irq_desc *desc) +{ + struct irq_cfg *old_cfg, *cfg; + + old_cfg = old_desc->chip_data; + cfg = desc->chip_data; + + if (old_cfg == cfg) + return; + + if (old_cfg) { + free_irq_2_pin(old_cfg, cfg); + free_irq_cfg(old_cfg); + old_desc->chip_data = NULL; + } +} + +static void set_extra_move_desc(struct irq_desc *desc, cpumask_t mask) +{ + struct irq_cfg *cfg = desc->chip_data; + + if (!cfg->move_in_progress) { + /* it means that domain is not changed */ + if (!cpus_intersects(desc->affinity, mask)) + cfg->move_desc_pending = 1; + } +} +#endif + #else static struct irq_cfg *irq_cfg(unsigned int irq) { @@ -231,9 +349,11 @@ static struct irq_cfg *irq_cfg(unsigned int irq) #endif +#ifndef CONFIG_NUMA_MIGRATE_IRQ_DESC static inline void set_extra_move_desc(struct irq_desc *desc, cpumask_t mask) { } +#endif struct io_apic { unsigned int index; @@ -2346,14 +2466,34 @@ static void irq_complete_move(struct irq_desc **descp) struct irq_cfg *cfg = desc->chip_data; unsigned vector, me; - if (likely(!cfg->move_in_progress)) + if (likely(!cfg->move_in_progress)) { +#ifdef CONFIG_NUMA_MIGRATE_IRQ_DESC + if (likely(!cfg->move_desc_pending)) + return; + + /* domain is not change, but affinity is changed */ + me = smp_processor_id(); + if (cpu_isset(me, desc->affinity)) { + *descp = desc = move_irq_desc(desc, me); + /* get the new one */ + cfg = desc->chip_data; + cfg->move_desc_pending = 0; + } +#endif return; + } vector = ~get_irq_regs()->orig_ax; me = smp_processor_id(); if ((vector == cfg->vector) && cpu_isset(me, cfg->domain)) { cpumask_t cleanup_mask; +#ifdef CONFIG_NUMA_MIGRATE_IRQ_DESC + *descp = desc = move_irq_desc(desc, me); + /* get the new one */ + cfg = desc->chip_data; +#endif + cpus_and(cleanup_mask, cfg->old_domain, cpu_online_map); cfg->move_cleanup_count = cpus_weight(cleanup_mask); send_IPI_mask(cleanup_mask, IRQ_MOVE_CLEANUP_VECTOR); diff --git a/include/linux/irq.h b/include/linux/irq.h index b5749db3e5a1..36a015746788 100644 --- a/include/linux/irq.h +++ b/include/linux/irq.h @@ -227,6 +227,16 @@ extern struct irq_desc *move_irq_desc(struct irq_desc *old_desc, int cpu); #endif +static inline struct irq_desc * +irq_remap_to_desc(unsigned int irq, struct irq_desc *desc) +{ +#ifdef CONFIG_NUMA_MIGRATE_IRQ_DESC + return irq_to_desc(irq); +#else + return desc; +#endif +} + /* * Migration helpers for obsolete names, they will go away: */ diff --git a/kernel/irq/Makefile b/kernel/irq/Makefile index 681c52dbfe22..4dd5b1edac98 100644 --- a/kernel/irq/Makefile +++ b/kernel/irq/Makefile @@ -3,3 +3,4 @@ obj-y := handle.o manage.o spurious.o resend.o chip.o devres.o obj-$(CONFIG_GENERIC_IRQ_PROBE) += autoprobe.o obj-$(CONFIG_PROC_FS) += proc.o obj-$(CONFIG_GENERIC_PENDING_IRQ) += migration.o +obj-$(CONFIG_NUMA_MIGRATE_IRQ_DESC) += numa_migrate.o diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index 8e4fce4a1b1f..de210f4b7a92 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -353,6 +353,7 @@ handle_level_irq(unsigned int irq, struct irq_desc *desc) spin_lock(&desc->lock); mask_ack_irq(desc, irq); + desc = irq_remap_to_desc(irq, desc); if (unlikely(desc->status & IRQ_INPROGRESS)) goto out_unlock; @@ -430,6 +431,7 @@ handle_fasteoi_irq(unsigned int irq, struct irq_desc *desc) desc->status &= ~IRQ_INPROGRESS; out: desc->chip->eoi(irq); + desc = irq_remap_to_desc(irq, desc); spin_unlock(&desc->lock); } @@ -466,12 +468,14 @@ handle_edge_irq(unsigned int irq, struct irq_desc *desc) !desc->action)) { desc->status |= (IRQ_PENDING | IRQ_MASKED); mask_ack_irq(desc, irq); + desc = irq_remap_to_desc(irq, desc); goto out_unlock; } kstat_incr_irqs_this_cpu(irq, desc); /* Start handling the irq */ desc->chip->ack(irq); + desc = irq_remap_to_desc(irq, desc); /* Mark the IRQ currently in progress.*/ desc->status |= IRQ_INPROGRESS; @@ -532,8 +536,10 @@ handle_percpu_irq(unsigned int irq, struct irq_desc *desc) if (!noirqdebug) note_interrupt(irq, desc, action_ret); - if (desc->chip->eoi) + if (desc->chip->eoi) { desc->chip->eoi(irq); + desc = irq_remap_to_desc(irq, desc); + } } void @@ -568,8 +574,10 @@ __set_irq_handler(unsigned int irq, irq_flow_handler_t handle, int is_chained, /* Uninstall? */ if (handle == handle_bad_irq) { - if (desc->chip != &no_irq_chip) + if (desc->chip != &no_irq_chip) { mask_ack_irq(desc, irq); + desc = irq_remap_to_desc(irq, desc); + } desc->status |= IRQ_DISABLED; desc->depth = 1; } diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index 8aa09547f5ef..f1a23069c20a 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -23,7 +23,7 @@ /* * lockdep: we want to handle all irq_desc locks as a single lock-class: */ -static struct lock_class_key irq_desc_lock_class; +struct lock_class_key irq_desc_lock_class; /** * handle_bad_irq - handle spurious and unhandled irqs @@ -73,7 +73,7 @@ static struct irq_desc irq_desc_init = { #endif }; -static void init_kstat_irqs(struct irq_desc *desc, int cpu, int nr) +void init_kstat_irqs(struct irq_desc *desc, int cpu, int nr) { unsigned long bytes; char *ptr; @@ -113,7 +113,7 @@ static void init_one_irq_desc(int irq, struct irq_desc *desc, int cpu) /* * Protect the sparse_irqs: */ -static DEFINE_SPINLOCK(sparse_irq_lock); +DEFINE_SPINLOCK(sparse_irq_lock); struct irq_desc *irq_desc_ptrs[NR_IRQS] __read_mostly; @@ -337,8 +337,11 @@ unsigned int __do_IRQ(unsigned int irq) /* * No locking required for CPU-local interrupts: */ - if (desc->chip->ack) + if (desc->chip->ack) { desc->chip->ack(irq); + /* get new one */ + desc = irq_remap_to_desc(irq, desc); + } if (likely(!(desc->status & IRQ_DISABLED))) { action_ret = handle_IRQ_event(irq, desc->action); if (!noirqdebug) @@ -349,8 +352,10 @@ unsigned int __do_IRQ(unsigned int irq) } spin_lock(&desc->lock); - if (desc->chip->ack) + if (desc->chip->ack) { desc->chip->ack(irq); + desc = irq_remap_to_desc(irq, desc); + } /* * REPLAY is when Linux resends an IRQ that was dropped earlier * WAITING is used by probe to mark irqs that are being tested diff --git a/kernel/irq/internals.h b/kernel/irq/internals.h index 64c1c7253dae..e6d0a43cc125 100644 --- a/kernel/irq/internals.h +++ b/kernel/irq/internals.h @@ -13,6 +13,11 @@ extern void compat_irq_chip_set_default_handler(struct irq_desc *desc); extern int __irq_set_trigger(struct irq_desc *desc, unsigned int irq, unsigned long flags); +extern struct lock_class_key irq_desc_lock_class; +extern void init_kstat_irqs(struct irq_desc *desc, int cpu, int nr); +extern spinlock_t sparse_irq_lock; +extern struct irq_desc *irq_desc_ptrs[NR_IRQS]; + #ifdef CONFIG_PROC_FS extern void register_irq_proc(unsigned int irq, struct irq_desc *desc); extern void register_handler_proc(unsigned int irq, struct irqaction *action); diff --git a/kernel/irq/numa_migrate.c b/kernel/irq/numa_migrate.c new file mode 100644 index 000000000000..0178e2296990 --- /dev/null +++ b/kernel/irq/numa_migrate.c @@ -0,0 +1,127 @@ +/* + * linux/kernel/irq/handle.c + * + * Copyright (C) 1992, 1998-2006 Linus Torvalds, Ingo Molnar + * Copyright (C) 2005-2006, Thomas Gleixner, Russell King + * + * This file contains the core interrupt handling code. + * + * Detailed information is available in Documentation/DocBook/genericirq + * + */ + +#include +#include +#include +#include +#include + +#include "internals.h" + +static void init_copy_kstat_irqs(struct irq_desc *old_desc, + struct irq_desc *desc, + int cpu, int nr) +{ + unsigned long bytes; + + init_kstat_irqs(desc, cpu, nr); + + if (desc->kstat_irqs != old_desc->kstat_irqs) { + /* Compute how many bytes we need per irq and allocate them */ + bytes = nr * sizeof(unsigned int); + + memcpy(desc->kstat_irqs, old_desc->kstat_irqs, bytes); + } +} + +static void free_kstat_irqs(struct irq_desc *old_desc, struct irq_desc *desc) +{ + if (old_desc->kstat_irqs == desc->kstat_irqs) + return; + + kfree(old_desc->kstat_irqs); + old_desc->kstat_irqs = NULL; +} + +static void init_copy_one_irq_desc(int irq, struct irq_desc *old_desc, + struct irq_desc *desc, int cpu) +{ + memcpy(desc, old_desc, sizeof(struct irq_desc)); + desc->cpu = cpu; + lockdep_set_class(&desc->lock, &irq_desc_lock_class); + init_copy_kstat_irqs(old_desc, desc, cpu, nr_cpu_ids); + arch_init_copy_chip_data(old_desc, desc, cpu); +} + +static void free_one_irq_desc(struct irq_desc *old_desc, struct irq_desc *desc) +{ + free_kstat_irqs(old_desc, desc); + arch_free_chip_data(old_desc, desc); +} + +static struct irq_desc *__real_move_irq_desc(struct irq_desc *old_desc, + int cpu) +{ + struct irq_desc *desc; + unsigned int irq; + unsigned long flags; + int node; + + irq = old_desc->irq; + + spin_lock_irqsave(&sparse_irq_lock, flags); + + /* We have to check it to avoid races with another CPU */ + desc = irq_desc_ptrs[irq]; + + if (desc && old_desc != desc) + goto out_unlock; + + node = cpu_to_node(cpu); + desc = kzalloc_node(sizeof(*desc), GFP_ATOMIC, node); + printk(KERN_DEBUG " move irq_desc for %d to cpu %d node %d\n", + irq, cpu, node); + if (!desc) { + printk(KERN_ERR "can not get new irq_desc for moving\n"); + /* still use old one */ + desc = old_desc; + goto out_unlock; + } + init_copy_one_irq_desc(irq, old_desc, desc, cpu); + + irq_desc_ptrs[irq] = desc; + + /* free the old one */ + free_one_irq_desc(old_desc, desc); + kfree(old_desc); + +out_unlock: + spin_unlock_irqrestore(&sparse_irq_lock, flags); + + return desc; +} + +struct irq_desc *move_irq_desc(struct irq_desc *desc, int cpu) +{ + int old_cpu; + int node, old_node; + + /* those all static, do move them */ + if (desc->irq < NR_IRQS_LEGACY) + return desc; + + old_cpu = desc->cpu; + printk(KERN_DEBUG + "try to move irq_desc from cpu %d to %d\n", old_cpu, cpu); + if (old_cpu != cpu) { + node = cpu_to_node(cpu); + old_node = cpu_to_node(old_cpu); + if (old_node != node) + desc = __real_move_irq_desc(desc, cpu); + else + desc->cpu = cpu; + } + + return desc; +} + -- cgit v1.2.3 From 17483a1f34c970e6c2cb8c082d4441bfabbe88a9 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Fri, 12 Dec 2008 13:14:18 -0800 Subject: sparseirq: fix !SMP building, #2 Impact: build fix make intr_remapping.c to include smp.h, so could use boot_cpu_id there also remove old change that disabling sparseirq with !SMP Signed-off-by: Yinghai Lu Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 2 +- drivers/pci/intr_remapping.c | 1 + 2 files changed, 2 insertions(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 29073532f94c..60a008857a38 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -240,7 +240,7 @@ config X86_HAS_BOOT_CPU_ID config SPARSE_IRQ bool "Support sparse irq numbering" - depends on (PCI_MSI || HT_IRQ) && SMP + depends on PCI_MSI || HT_IRQ default y help This enables support for sparse irq, esp for msi/msi-x. You may need diff --git a/drivers/pci/intr_remapping.c b/drivers/pci/intr_remapping.c index c9958ec5e25e..f78371b22529 100644 --- a/drivers/pci/intr_remapping.c +++ b/drivers/pci/intr_remapping.c @@ -5,6 +5,7 @@ #include #include #include +#include #include #include "intr_remapping.h" -- cgit v1.2.3 From d680fe44775ed17a80035462d9898f5e77bfd7dd Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Sat, 13 Dec 2008 00:09:08 +0300 Subject: x86: entry_64 - introduce FTRACE_ frame macro v2 Impact: clean up Itroduce MCOUNT_SAVE/RESTORE_FRAME which allow us to save a number of lines on source level. Also fix a comment in ftrace.h. Signed-off-by: Ingo Molnar --- arch/x86/include/asm/ftrace.h | 29 +++++++++++++++++++++- arch/x86/kernel/entry_64.S | 57 ++++++------------------------------------- 2 files changed, 35 insertions(+), 51 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/ftrace.h b/arch/x86/include/asm/ftrace.h index 7e61b4ceb9a4..b55b4a7fbefd 100644 --- a/arch/x86/include/asm/ftrace.h +++ b/arch/x86/include/asm/ftrace.h @@ -1,6 +1,33 @@ #ifndef _ASM_X86_FTRACE_H #define _ASM_X86_FTRACE_H +#ifdef __ASSEMBLY__ + + .macro MCOUNT_SAVE_FRAME + /* taken from glibc */ + subq $0x38, %rsp + movq %rax, (%rsp) + movq %rcx, 8(%rsp) + movq %rdx, 16(%rsp) + movq %rsi, 24(%rsp) + movq %rdi, 32(%rsp) + movq %r8, 40(%rsp) + movq %r9, 48(%rsp) + .endm + + .macro MCOUNT_RESTORE_FRAME + movq 48(%rsp), %r9 + movq 40(%rsp), %r8 + movq 32(%rsp), %rdi + movq 24(%rsp), %rsi + movq 16(%rsp), %rdx + movq 8(%rsp), %rcx + movq (%rsp), %rax + addq $0x38, %rsp + .endm + +#endif + #ifdef CONFIG_FUNCTION_TRACER #define MCOUNT_ADDR ((long)(mcount)) #define MCOUNT_INSN_SIZE 5 /* sizeof mcount call */ @@ -46,7 +73,7 @@ struct ftrace_ret_stack { /* * Primary handler of a function return. * It relays on ftrace_return_to_handler. - * Defined in entry32.S + * Defined in entry_32/64.S */ extern void return_to_handler(void); diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 54e0bbdccb99..303dd84d2a98 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -71,15 +71,7 @@ ENTRY(ftrace_caller) cmpl $0, function_trace_stop jne ftrace_stub - /* taken from glibc */ - subq $0x38, %rsp - movq %rax, (%rsp) - movq %rcx, 8(%rsp) - movq %rdx, 16(%rsp) - movq %rsi, 24(%rsp) - movq %rdi, 32(%rsp) - movq %r8, 40(%rsp) - movq %r9, 48(%rsp) + MCOUNT_SAVE_FRAME movq 0x38(%rsp), %rdi movq 8(%rbp), %rsi @@ -89,14 +81,7 @@ ENTRY(ftrace_caller) ftrace_call: call ftrace_stub - movq 48(%rsp), %r9 - movq 40(%rsp), %r8 - movq 32(%rsp), %rdi - movq 24(%rsp), %rsi - movq 16(%rsp), %rdx - movq 8(%rsp), %rcx - movq (%rsp), %rax - addq $0x38, %rsp + MCOUNT_RESTORE_FRAME #ifdef CONFIG_FUNCTION_GRAPH_TRACER .globl ftrace_graph_call @@ -130,15 +115,7 @@ ftrace_stub: retq trace: - /* taken from glibc */ - subq $0x38, %rsp - movq %rax, (%rsp) - movq %rcx, 8(%rsp) - movq %rdx, 16(%rsp) - movq %rsi, 24(%rsp) - movq %rdi, 32(%rsp) - movq %r8, 40(%rsp) - movq %r9, 48(%rsp) + MCOUNT_SAVE_FRAME movq 0x38(%rsp), %rdi movq 8(%rbp), %rsi @@ -146,14 +123,7 @@ trace: call *ftrace_trace_function - movq 48(%rsp), %r9 - movq 40(%rsp), %r8 - movq 32(%rsp), %rdi - movq 24(%rsp), %rsi - movq 16(%rsp), %rdx - movq 8(%rsp), %rcx - movq (%rsp), %rax - addq $0x38, %rsp + MCOUNT_RESTORE_FRAME jmp ftrace_stub END(mcount) @@ -165,14 +135,7 @@ ENTRY(ftrace_graph_caller) cmpl $0, function_trace_stop jne ftrace_stub - subq $0x38, %rsp - movq %rax, (%rsp) - movq %rcx, 8(%rsp) - movq %rdx, 16(%rsp) - movq %rsi, 24(%rsp) - movq %rdi, 32(%rsp) - movq %r8, 40(%rsp) - movq %r9, 48(%rsp) + MCOUNT_SAVE_FRAME leaq 8(%rbp), %rdi movq 0x38(%rsp), %rsi @@ -180,14 +143,8 @@ ENTRY(ftrace_graph_caller) call prepare_ftrace_return - movq 48(%rsp), %r9 - movq 40(%rsp), %r8 - movq 32(%rsp), %rdi - movq 24(%rsp), %rsi - movq 16(%rsp), %rdx - movq 8(%rsp), %rcx - movq (%rsp), %rax - addq $0x38, %rsp + MCOUNT_RESTORE_FRAME + retq END(ftrace_graph_caller) -- cgit v1.2.3 From 36f5101a60de8f79c0d1ca06e50660bf5129e02c Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Tue, 16 Dec 2008 17:33:51 -0800 Subject: x86: enable MAXSMP Impact: activates new off-stack cpumask code on MAXSMP (non-default) x86 configs Set MAXSMP to enable CONFIG_CPUMASK_OFFSTACK which moves cpumask's off the stack (and in structs) when using cpumask_var_t. Signed-off-by: Mike Travis Signed-off-by: Rusty Russell Acked-by: Ingo Molnar --- arch/x86/Kconfig | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index d99eeb7915c6..1fd44352f27c 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -591,16 +591,17 @@ config IOMMU_HELPER config MAXSMP bool "Configure Maximum number of SMP Processors and NUMA Nodes" - depends on X86_64 && SMP && BROKEN + depends on X86_64 && SMP && DEBUG_KERNEL && EXPERIMENTAL + select CPUMASK_OFFSTACK default n help Configure maximum number of CPUS and NUMA Nodes for this architecture. If unsure, say N. config NR_CPUS - int "Maximum number of CPUs (2-512)" if !MAXSMP - range 2 512 depends on SMP + int "Maximum number of CPUs" if SMP && !MAXSMP + range 2 512 if SMP && !MAXSMP default "4096" if MAXSMP default "32" if X86_NUMAQ || X86_SUMMIT || X86_BIGSMP || X86_ES7000 default "8" -- cgit v1.2.3 From e7986739a76cde5079da08809d8bbc6878387ae0 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Tue, 16 Dec 2008 17:33:52 -0800 Subject: x86 smp: modify send_IPI_mask interface to accept cpumask_t pointers Impact: cleanup, change parameter passing * Change genapic interfaces to accept cpumask_t pointers where possible. * Modify external callers to use cpumask_t pointers in function calls. * Create new send_IPI_mask_allbutself which is the same as the send_IPI_mask functions but removes smp_processor_id() from list. This removes another common need for a temporary cpumask_t variable. * Functions that used a temp cpumask_t variable for: cpumask_t allbutme = cpu_online_map; cpu_clear(smp_processor_id(), allbutme); if (!cpus_empty(allbutme)) ... become: if (!cpus_equal(cpu_online_map, cpumask_of_cpu(cpu))) ... * Other minor code optimizations (like using cpus_clear instead of CPU_MASK_NONE, etc.) Applies to linux-2.6.tip/master. Signed-off-by: Mike Travis Signed-off-by: Rusty Russell Acked-by: Ingo Molnar --- arch/x86/include/asm/bigsmp/apic.h | 14 +-- arch/x86/include/asm/bigsmp/ipi.h | 9 +- arch/x86/include/asm/es7000/apic.h | 38 +++---- arch/x86/include/asm/es7000/ipi.h | 9 +- arch/x86/include/asm/genapic_32.h | 9 +- arch/x86/include/asm/genapic_64.h | 11 +- arch/x86/include/asm/ipi.h | 21 +++- arch/x86/include/asm/mach-default/mach_apic.h | 17 ++- arch/x86/include/asm/mach-default/mach_ipi.h | 18 ++-- arch/x86/include/asm/numaq/apic.h | 6 +- arch/x86/include/asm/numaq/ipi.h | 9 +- arch/x86/include/asm/smp.h | 6 +- arch/x86/include/asm/summit/apic.h | 12 +-- arch/x86/include/asm/summit/ipi.h | 9 +- arch/x86/kernel/apic.c | 6 +- arch/x86/kernel/crash.c | 5 +- arch/x86/kernel/genapic_flat_64.c | 76 +++++++++----- arch/x86/kernel/genx2apic_cluster.c | 60 +++++++---- arch/x86/kernel/genx2apic_phys.c | 55 +++++++--- arch/x86/kernel/genx2apic_uv_x.c | 43 +++++--- arch/x86/kernel/io_apic.c | 145 +++++++++++++------------- arch/x86/kernel/ipi.c | 26 +++-- arch/x86/kernel/smp.c | 8 +- arch/x86/kernel/tlb_32.c | 2 +- arch/x86/kernel/tlb_64.c | 2 +- arch/x86/mach-generic/bigsmp.c | 5 +- arch/x86/mach-generic/es7000.c | 5 +- arch/x86/mach-generic/numaq.c | 5 +- arch/x86/mach-generic/summit.c | 5 +- arch/x86/xen/smp.c | 17 ++- 30 files changed, 380 insertions(+), 273 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/bigsmp/apic.h b/arch/x86/include/asm/bigsmp/apic.h index ce547f24a1cd..dc6225ca48ad 100644 --- a/arch/x86/include/asm/bigsmp/apic.h +++ b/arch/x86/include/asm/bigsmp/apic.h @@ -9,12 +9,12 @@ static inline int apic_id_registered(void) return (1); } -static inline cpumask_t target_cpus(void) +static inline const cpumask_t *target_cpus(void) { #ifdef CONFIG_SMP - return cpu_online_map; + return &cpu_online_map; #else - return cpumask_of_cpu(0); + return &cpumask_of_cpu(0); #endif } @@ -79,7 +79,7 @@ static inline int apicid_to_node(int logical_apicid) static inline int cpu_present_to_apicid(int mps_cpu) { - if (mps_cpu < NR_CPUS) + if (mps_cpu < nr_cpu_ids) return (int) per_cpu(x86_bios_cpu_apicid, mps_cpu); return BAD_APICID; @@ -94,7 +94,7 @@ extern u8 cpu_2_logical_apicid[]; /* Mapping from cpu number to logical apicid */ static inline int cpu_to_logical_apicid(int cpu) { - if (cpu >= NR_CPUS) + if (cpu >= nr_cpu_ids) return BAD_APICID; return cpu_physical_id(cpu); } @@ -119,12 +119,12 @@ static inline int check_phys_apicid_present(int boot_cpu_physical_apicid) } /* As we are using single CPU as destination, pick only one CPU here */ -static inline unsigned int cpu_mask_to_apicid(cpumask_t cpumask) +static inline unsigned int cpu_mask_to_apicid(const cpumask_t *cpumask) { int cpu; int apicid; - cpu = first_cpu(cpumask); + cpu = first_cpu(*cpumask); apicid = cpu_to_logical_apicid(cpu); return apicid; } diff --git a/arch/x86/include/asm/bigsmp/ipi.h b/arch/x86/include/asm/bigsmp/ipi.h index 9404c535b7ec..63553e9f22b2 100644 --- a/arch/x86/include/asm/bigsmp/ipi.h +++ b/arch/x86/include/asm/bigsmp/ipi.h @@ -1,9 +1,10 @@ #ifndef __ASM_MACH_IPI_H #define __ASM_MACH_IPI_H -void send_IPI_mask_sequence(cpumask_t mask, int vector); +void send_IPI_mask_sequence(const cpumask_t *mask, int vector); +void send_IPI_mask_allbutself(const cpumask_t *mask, int vector); -static inline void send_IPI_mask(cpumask_t mask, int vector) +static inline void send_IPI_mask(const cpumask_t *mask, int vector) { send_IPI_mask_sequence(mask, vector); } @@ -14,12 +15,12 @@ static inline void send_IPI_allbutself(int vector) cpu_clear(smp_processor_id(), mask); if (!cpus_empty(mask)) - send_IPI_mask(mask, vector); + send_IPI_mask(&mask, vector); } static inline void send_IPI_all(int vector) { - send_IPI_mask(cpu_online_map, vector); + send_IPI_mask(&cpu_online_map, vector); } #endif /* __ASM_MACH_IPI_H */ diff --git a/arch/x86/include/asm/es7000/apic.h b/arch/x86/include/asm/es7000/apic.h index e24ef876915f..4cac0837bb40 100644 --- a/arch/x86/include/asm/es7000/apic.h +++ b/arch/x86/include/asm/es7000/apic.h @@ -9,14 +9,14 @@ static inline int apic_id_registered(void) return (1); } -static inline cpumask_t target_cpus_cluster(void) +static inline const cpumask_t *target_cpus_cluster(void) { - return CPU_MASK_ALL; + return &CPU_MASK_ALL; } -static inline cpumask_t target_cpus(void) +static inline const cpumask_t *target_cpus(void) { - return cpumask_of_cpu(smp_processor_id()); + return &cpumask_of_cpu(smp_processor_id()); } #define APIC_DFR_VALUE_CLUSTER (APIC_DFR_CLUSTER) @@ -80,9 +80,10 @@ extern int apic_version [MAX_APICS]; static inline void setup_apic_routing(void) { int apic = per_cpu(x86_bios_cpu_apicid, smp_processor_id()); - printk("Enabling APIC mode: %s. Using %d I/O APICs, target cpus %lx\n", + printk("Enabling APIC mode: %s. Using %d I/O APICs, target cpus %lx\n", (apic_version[apic] == 0x14) ? - "Physical Cluster" : "Logical Cluster", nr_ioapics, cpus_addr(target_cpus())[0]); + "Physical Cluster" : "Logical Cluster", + nr_ioapics, cpus_addr(*target_cpus())[0]); } static inline int multi_timer_check(int apic, int irq) @@ -100,7 +101,7 @@ static inline int cpu_present_to_apicid(int mps_cpu) { if (!mps_cpu) return boot_cpu_physical_apicid; - else if (mps_cpu < NR_CPUS) + else if (mps_cpu < nr_cpu_ids) return (int) per_cpu(x86_bios_cpu_apicid, mps_cpu); else return BAD_APICID; @@ -120,9 +121,9 @@ extern u8 cpu_2_logical_apicid[]; static inline int cpu_to_logical_apicid(int cpu) { #ifdef CONFIG_SMP - if (cpu >= NR_CPUS) - return BAD_APICID; - return (int)cpu_2_logical_apicid[cpu]; + if (cpu >= nr_cpu_ids) + return BAD_APICID; + return (int)cpu_2_logical_apicid[cpu]; #else return logical_smp_processor_id(); #endif @@ -146,14 +147,15 @@ static inline int check_phys_apicid_present(int cpu_physical_apicid) return (1); } -static inline unsigned int cpu_mask_to_apicid_cluster(cpumask_t cpumask) +static inline unsigned int +cpu_mask_to_apicid_cluster(const struct cpumask *cpumask) { int num_bits_set; int cpus_found = 0; int cpu; int apicid; - num_bits_set = cpus_weight(cpumask); + num_bits_set = cpumask_weight(cpumask); /* Return id to all */ if (num_bits_set == NR_CPUS) return 0xFF; @@ -161,10 +163,10 @@ static inline unsigned int cpu_mask_to_apicid_cluster(cpumask_t cpumask) * The cpus in the mask must all be on the apic cluster. If are not * on the same apicid cluster return default value of TARGET_CPUS. */ - cpu = first_cpu(cpumask); + cpu = cpumask_first(cpumask); apicid = cpu_to_logical_apicid(cpu); while (cpus_found < num_bits_set) { - if (cpu_isset(cpu, cpumask)) { + if (cpumask_test_cpu(cpu, cpumask)) { int new_apicid = cpu_to_logical_apicid(cpu); if (apicid_cluster(apicid) != apicid_cluster(new_apicid)){ @@ -179,14 +181,14 @@ static inline unsigned int cpu_mask_to_apicid_cluster(cpumask_t cpumask) return apicid; } -static inline unsigned int cpu_mask_to_apicid(cpumask_t cpumask) +static inline unsigned int cpu_mask_to_apicid(const cpumask_t *cpumask) { int num_bits_set; int cpus_found = 0; int cpu; int apicid; - num_bits_set = cpus_weight(cpumask); + num_bits_set = cpus_weight(*cpumask); /* Return id to all */ if (num_bits_set == NR_CPUS) return cpu_to_logical_apicid(0); @@ -194,10 +196,10 @@ static inline unsigned int cpu_mask_to_apicid(cpumask_t cpumask) * The cpus in the mask must all be on the apic cluster. If are not * on the same apicid cluster return default value of TARGET_CPUS. */ - cpu = first_cpu(cpumask); + cpu = first_cpu(*cpumask); apicid = cpu_to_logical_apicid(cpu); while (cpus_found < num_bits_set) { - if (cpu_isset(cpu, cpumask)) { + if (cpu_isset(cpu, *cpumask)) { int new_apicid = cpu_to_logical_apicid(cpu); if (apicid_cluster(apicid) != apicid_cluster(new_apicid)){ diff --git a/arch/x86/include/asm/es7000/ipi.h b/arch/x86/include/asm/es7000/ipi.h index 632a955fcc0a..1a8507265f91 100644 --- a/arch/x86/include/asm/es7000/ipi.h +++ b/arch/x86/include/asm/es7000/ipi.h @@ -1,9 +1,10 @@ #ifndef __ASM_ES7000_IPI_H #define __ASM_ES7000_IPI_H -void send_IPI_mask_sequence(cpumask_t mask, int vector); +void send_IPI_mask_sequence(const cpumask_t *mask, int vector); +void send_IPI_mask_allbutself(const cpumask_t *mask, int vector); -static inline void send_IPI_mask(cpumask_t mask, int vector) +static inline void send_IPI_mask(const cpumask_t *mask, int vector) { send_IPI_mask_sequence(mask, vector); } @@ -13,12 +14,12 @@ static inline void send_IPI_allbutself(int vector) cpumask_t mask = cpu_online_map; cpu_clear(smp_processor_id(), mask); if (!cpus_empty(mask)) - send_IPI_mask(mask, vector); + send_IPI_mask(&mask, vector); } static inline void send_IPI_all(int vector) { - send_IPI_mask(cpu_online_map, vector); + send_IPI_mask(&cpu_online_map, vector); } #endif /* __ASM_ES7000_IPI_H */ diff --git a/arch/x86/include/asm/genapic_32.h b/arch/x86/include/asm/genapic_32.h index 0ac17d33a8c7..b21ed21c574d 100644 --- a/arch/x86/include/asm/genapic_32.h +++ b/arch/x86/include/asm/genapic_32.h @@ -24,7 +24,7 @@ struct genapic { int (*probe)(void); int (*apic_id_registered)(void); - cpumask_t (*target_cpus)(void); + const cpumask_t *(*target_cpus)(void); int int_delivery_mode; int int_dest_mode; int ESR_DISABLE; @@ -57,12 +57,13 @@ struct genapic { unsigned (*get_apic_id)(unsigned long x); unsigned long apic_id_mask; - unsigned int (*cpu_mask_to_apicid)(cpumask_t cpumask); - cpumask_t (*vector_allocation_domain)(int cpu); + unsigned int (*cpu_mask_to_apicid)(const cpumask_t *cpumask); + void (*vector_allocation_domain)(int cpu, cpumask_t *retmask); #ifdef CONFIG_SMP /* ipi */ - void (*send_IPI_mask)(cpumask_t mask, int vector); + void (*send_IPI_mask)(const cpumask_t *mask, int vector); + void (*send_IPI_mask_allbutself)(const cpumask_t *mask, int vector); void (*send_IPI_allbutself)(int vector); void (*send_IPI_all)(int vector); #endif diff --git a/arch/x86/include/asm/genapic_64.h b/arch/x86/include/asm/genapic_64.h index 2cae011668b7..a020e7d35a40 100644 --- a/arch/x86/include/asm/genapic_64.h +++ b/arch/x86/include/asm/genapic_64.h @@ -1,6 +1,8 @@ #ifndef _ASM_X86_GENAPIC_64_H #define _ASM_X86_GENAPIC_64_H +#include + /* * Copyright 2004 James Cleverdon, IBM. * Subject to the GNU Public License, v.2 @@ -18,16 +20,17 @@ struct genapic { u32 int_delivery_mode; u32 int_dest_mode; int (*apic_id_registered)(void); - cpumask_t (*target_cpus)(void); - cpumask_t (*vector_allocation_domain)(int cpu); + const cpumask_t *(*target_cpus)(void); + void (*vector_allocation_domain)(int cpu, cpumask_t *retmask); void (*init_apic_ldr)(void); /* ipi */ - void (*send_IPI_mask)(cpumask_t mask, int vector); + void (*send_IPI_mask)(const cpumask_t *mask, int vector); + void (*send_IPI_mask_allbutself)(const cpumask_t *mask, int vector); void (*send_IPI_allbutself)(int vector); void (*send_IPI_all)(int vector); void (*send_IPI_self)(int vector); /* */ - unsigned int (*cpu_mask_to_apicid)(cpumask_t cpumask); + unsigned int (*cpu_mask_to_apicid)(const cpumask_t *cpumask); unsigned int (*phys_pkg_id)(int index_msb); unsigned int (*get_apic_id)(unsigned long x); unsigned long (*set_apic_id)(unsigned int id); diff --git a/arch/x86/include/asm/ipi.h b/arch/x86/include/asm/ipi.h index f89dffb28aa9..24b6e613edfa 100644 --- a/arch/x86/include/asm/ipi.h +++ b/arch/x86/include/asm/ipi.h @@ -117,7 +117,7 @@ static inline void __send_IPI_dest_field(unsigned int mask, int vector, native_apic_mem_write(APIC_ICR, cfg); } -static inline void send_IPI_mask_sequence(cpumask_t mask, int vector) +static inline void send_IPI_mask_sequence(const cpumask_t *mask, int vector) { unsigned long flags; unsigned long query_cpu; @@ -128,11 +128,28 @@ static inline void send_IPI_mask_sequence(cpumask_t mask, int vector) * - mbligh */ local_irq_save(flags); - for_each_cpu_mask_nr(query_cpu, mask) { + for_each_cpu_mask_nr(query_cpu, *mask) { __send_IPI_dest_field(per_cpu(x86_cpu_to_apicid, query_cpu), vector, APIC_DEST_PHYSICAL); } local_irq_restore(flags); } +static inline void send_IPI_mask_allbutself(const cpumask_t *mask, int vector) +{ + unsigned long flags; + unsigned int query_cpu; + unsigned int this_cpu = smp_processor_id(); + + /* See Hack comment above */ + + local_irq_save(flags); + for_each_cpu_mask_nr(query_cpu, *mask) + if (query_cpu != this_cpu) + __send_IPI_dest_field( + per_cpu(x86_cpu_to_apicid, query_cpu), + vector, APIC_DEST_PHYSICAL); + local_irq_restore(flags); +} + #endif /* _ASM_X86_IPI_H */ diff --git a/arch/x86/include/asm/mach-default/mach_apic.h b/arch/x86/include/asm/mach-default/mach_apic.h index 6cb3a467e067..c18896b0508c 100644 --- a/arch/x86/include/asm/mach-default/mach_apic.h +++ b/arch/x86/include/asm/mach-default/mach_apic.h @@ -8,12 +8,12 @@ #define APIC_DFR_VALUE (APIC_DFR_FLAT) -static inline cpumask_t target_cpus(void) +static inline const cpumask_t *target_cpus(void) { #ifdef CONFIG_SMP - return cpu_online_map; + return &cpu_online_map; #else - return cpumask_of_cpu(0); + return &cpumask_of_cpu(0); #endif } @@ -61,9 +61,9 @@ static inline int apic_id_registered(void) return physid_isset(read_apic_id(), phys_cpu_present_map); } -static inline unsigned int cpu_mask_to_apicid(cpumask_t cpumask) +static inline unsigned int cpu_mask_to_apicid(const cpumask_t *cpumask) { - return cpus_addr(cpumask)[0]; + return cpus_addr(*cpumask)[0]; } static inline u32 phys_pkg_id(u32 cpuid_apic, int index_msb) @@ -88,7 +88,7 @@ static inline int apicid_to_node(int logical_apicid) #endif } -static inline cpumask_t vector_allocation_domain(int cpu) +static inline void vector_allocation_domain(int cpu, cpumask_t *retmask) { /* Careful. Some cpus do not strictly honor the set of cpus * specified in the interrupt destination when using lowest @@ -98,8 +98,7 @@ static inline cpumask_t vector_allocation_domain(int cpu) * deliver interrupts to the wrong hyperthread when only one * hyperthread was specified in the interrupt desitination. */ - cpumask_t domain = { { [0] = APIC_ALL_CPUS, } }; - return domain; + *retmask = (cpumask_t) { { [0] = APIC_ALL_CPUS } }; } #endif @@ -131,7 +130,7 @@ static inline int cpu_to_logical_apicid(int cpu) static inline int cpu_present_to_apicid(int mps_cpu) { - if (mps_cpu < NR_CPUS && cpu_present(mps_cpu)) + if (mps_cpu < nr_cpu_ids && cpu_present(mps_cpu)) return (int)per_cpu(x86_bios_cpu_apicid, mps_cpu); else return BAD_APICID; diff --git a/arch/x86/include/asm/mach-default/mach_ipi.h b/arch/x86/include/asm/mach-default/mach_ipi.h index fabca01ebacf..9353ab854a10 100644 --- a/arch/x86/include/asm/mach-default/mach_ipi.h +++ b/arch/x86/include/asm/mach-default/mach_ipi.h @@ -4,7 +4,8 @@ /* Avoid include hell */ #define NMI_VECTOR 0x02 -void send_IPI_mask_bitmask(cpumask_t mask, int vector); +void send_IPI_mask_bitmask(const cpumask_t *mask, int vector); +void send_IPI_mask_allbutself(const cpumask_t *mask, int vector); void __send_IPI_shortcut(unsigned int shortcut, int vector); extern int no_broadcast; @@ -12,28 +13,27 @@ extern int no_broadcast; #ifdef CONFIG_X86_64 #include #define send_IPI_mask (genapic->send_IPI_mask) +#define send_IPI_mask_allbutself (genapic->send_IPI_mask_allbutself) #else -static inline void send_IPI_mask(cpumask_t mask, int vector) +static inline void send_IPI_mask(const cpumask_t *mask, int vector) { send_IPI_mask_bitmask(mask, vector); } +void send_IPI_mask_allbutself(const cpumask_t *mask, int vector); #endif static inline void __local_send_IPI_allbutself(int vector) { - if (no_broadcast || vector == NMI_VECTOR) { - cpumask_t mask = cpu_online_map; - - cpu_clear(smp_processor_id(), mask); - send_IPI_mask(mask, vector); - } else + if (no_broadcast || vector == NMI_VECTOR) + send_IPI_mask_allbutself(&cpu_online_map, vector); + else __send_IPI_shortcut(APIC_DEST_ALLBUT, vector); } static inline void __local_send_IPI_all(int vector) { if (no_broadcast || vector == NMI_VECTOR) - send_IPI_mask(cpu_online_map, vector); + send_IPI_mask(&cpu_online_map, vector); else __send_IPI_shortcut(APIC_DEST_ALLINC, vector); } diff --git a/arch/x86/include/asm/numaq/apic.h b/arch/x86/include/asm/numaq/apic.h index 0bf2a06b7a4e..1df7ebe738e5 100644 --- a/arch/x86/include/asm/numaq/apic.h +++ b/arch/x86/include/asm/numaq/apic.h @@ -7,9 +7,9 @@ #define APIC_DFR_VALUE (APIC_DFR_CLUSTER) -static inline cpumask_t target_cpus(void) +static inline const cpumask_t *target_cpus(void) { - return CPU_MASK_ALL; + return &CPU_MASK_ALL; } #define NO_BALANCE_IRQ (1) @@ -122,7 +122,7 @@ static inline void enable_apic_mode(void) * We use physical apicids here, not logical, so just return the default * physical broadcast to stop people from breaking us */ -static inline unsigned int cpu_mask_to_apicid(cpumask_t cpumask) +static inline unsigned int cpu_mask_to_apicid(const cpumask_t *cpumask) { return (int) 0xF; } diff --git a/arch/x86/include/asm/numaq/ipi.h b/arch/x86/include/asm/numaq/ipi.h index 935588d286cf..c734d7acc430 100644 --- a/arch/x86/include/asm/numaq/ipi.h +++ b/arch/x86/include/asm/numaq/ipi.h @@ -1,9 +1,10 @@ #ifndef __ASM_NUMAQ_IPI_H #define __ASM_NUMAQ_IPI_H -void send_IPI_mask_sequence(cpumask_t, int vector); +void send_IPI_mask_sequence(const cpumask_t *mask, int vector); +void send_IPI_mask_allbutself(const cpumask_t *mask, int vector); -static inline void send_IPI_mask(cpumask_t mask, int vector) +static inline void send_IPI_mask(const cpumask_t *mask, int vector) { send_IPI_mask_sequence(mask, vector); } @@ -14,12 +15,12 @@ static inline void send_IPI_allbutself(int vector) cpu_clear(smp_processor_id(), mask); if (!cpus_empty(mask)) - send_IPI_mask(mask, vector); + send_IPI_mask(&mask, vector); } static inline void send_IPI_all(int vector) { - send_IPI_mask(cpu_online_map, vector); + send_IPI_mask(&cpu_online_map, vector); } #endif /* __ASM_NUMAQ_IPI_H */ diff --git a/arch/x86/include/asm/smp.h b/arch/x86/include/asm/smp.h index d12811ce51d9..c4a9aa52df6e 100644 --- a/arch/x86/include/asm/smp.h +++ b/arch/x86/include/asm/smp.h @@ -60,7 +60,7 @@ struct smp_ops { void (*cpu_die)(unsigned int cpu); void (*play_dead)(void); - void (*send_call_func_ipi)(cpumask_t mask); + void (*send_call_func_ipi)(const cpumask_t *mask); void (*send_call_func_single_ipi)(int cpu); }; @@ -125,7 +125,7 @@ static inline void arch_send_call_function_single_ipi(int cpu) static inline void arch_send_call_function_ipi(cpumask_t mask) { - smp_ops.send_call_func_ipi(mask); + smp_ops.send_call_func_ipi(&mask); } void cpu_disable_common(void); @@ -138,7 +138,7 @@ void native_cpu_die(unsigned int cpu); void native_play_dead(void); void play_dead_common(void); -void native_send_call_func_ipi(cpumask_t mask); +void native_send_call_func_ipi(const cpumask_t *mask); void native_send_call_func_single_ipi(int cpu); extern void prefill_possible_map(void); diff --git a/arch/x86/include/asm/summit/apic.h b/arch/x86/include/asm/summit/apic.h index 9b3070f1c2ac..437dc83725ca 100644 --- a/arch/x86/include/asm/summit/apic.h +++ b/arch/x86/include/asm/summit/apic.h @@ -14,13 +14,13 @@ #define APIC_DFR_VALUE (APIC_DFR_CLUSTER) -static inline cpumask_t target_cpus(void) +static inline const cpumask_t *target_cpus(void) { /* CPU_MASK_ALL (0xff) has undefined behaviour with * dest_LowestPrio mode logical clustered apic interrupt routing * Just start on cpu 0. IRQ balancing will spread load */ - return cpumask_of_cpu(0); + return &cpumask_of_cpu(0); } #define INT_DELIVERY_MODE (dest_LowestPrio) @@ -137,14 +137,14 @@ static inline void enable_apic_mode(void) { } -static inline unsigned int cpu_mask_to_apicid(cpumask_t cpumask) +static inline unsigned int cpu_mask_to_apicid(const cpumask_t *cpumask) { int num_bits_set; int cpus_found = 0; int cpu; int apicid; - num_bits_set = cpus_weight(cpumask); + num_bits_set = cpus_weight(*cpumask); /* Return id to all */ if (num_bits_set == NR_CPUS) return (int) 0xFF; @@ -152,10 +152,10 @@ static inline unsigned int cpu_mask_to_apicid(cpumask_t cpumask) * The cpus in the mask must all be on the apic cluster. If are not * on the same apicid cluster return default value of TARGET_CPUS. */ - cpu = first_cpu(cpumask); + cpu = first_cpu(*cpumask); apicid = cpu_to_logical_apicid(cpu); while (cpus_found < num_bits_set) { - if (cpu_isset(cpu, cpumask)) { + if (cpu_isset(cpu, *cpumask)) { int new_apicid = cpu_to_logical_apicid(cpu); if (apicid_cluster(apicid) != apicid_cluster(new_apicid)){ diff --git a/arch/x86/include/asm/summit/ipi.h b/arch/x86/include/asm/summit/ipi.h index 53bd1e7bd7b4..a8a2c24f50cc 100644 --- a/arch/x86/include/asm/summit/ipi.h +++ b/arch/x86/include/asm/summit/ipi.h @@ -1,9 +1,10 @@ #ifndef __ASM_SUMMIT_IPI_H #define __ASM_SUMMIT_IPI_H -void send_IPI_mask_sequence(cpumask_t mask, int vector); +void send_IPI_mask_sequence(const cpumask_t *mask, int vector); +void send_IPI_mask_allbutself(const cpumask_t *mask, int vector); -static inline void send_IPI_mask(cpumask_t mask, int vector) +static inline void send_IPI_mask(const cpumask_t *mask, int vector) { send_IPI_mask_sequence(mask, vector); } @@ -14,12 +15,12 @@ static inline void send_IPI_allbutself(int vector) cpu_clear(smp_processor_id(), mask); if (!cpus_empty(mask)) - send_IPI_mask(mask, vector); + send_IPI_mask(&mask, vector); } static inline void send_IPI_all(int vector) { - send_IPI_mask(cpu_online_map, vector); + send_IPI_mask(&cpu_online_map, vector); } #endif /* __ASM_SUMMIT_IPI_H */ diff --git a/arch/x86/kernel/apic.c b/arch/x86/kernel/apic.c index b2cef49f3085..a375791c08ca 100644 --- a/arch/x86/kernel/apic.c +++ b/arch/x86/kernel/apic.c @@ -141,7 +141,7 @@ static int lapic_next_event(unsigned long delta, struct clock_event_device *evt); static void lapic_timer_setup(enum clock_event_mode mode, struct clock_event_device *evt); -static void lapic_timer_broadcast(const struct cpumask *mask); +static void lapic_timer_broadcast(const cpumask_t *mask); static void apic_pm_activate(void); /* @@ -453,10 +453,10 @@ static void lapic_timer_setup(enum clock_event_mode mode, /* * Local APIC timer broadcast function */ -static void lapic_timer_broadcast(const struct cpumask *mask) +static void lapic_timer_broadcast(const cpumask_t *mask) { #ifdef CONFIG_SMP - send_IPI_mask(*mask, LOCAL_TIMER_VECTOR); + send_IPI_mask(mask, LOCAL_TIMER_VECTOR); #endif } diff --git a/arch/x86/kernel/crash.c b/arch/x86/kernel/crash.c index 268553817909..81e01f7b1d12 100644 --- a/arch/x86/kernel/crash.c +++ b/arch/x86/kernel/crash.c @@ -77,10 +77,7 @@ static int crash_nmi_callback(struct notifier_block *self, static void smp_send_nmi_allbutself(void) { - cpumask_t mask = cpu_online_map; - cpu_clear(safe_smp_processor_id(), mask); - if (!cpus_empty(mask)) - send_IPI_mask(mask, NMI_VECTOR); + send_IPI_allbutself(NMI_VECTOR); } static struct notifier_block crash_nmi_nb = { diff --git a/arch/x86/kernel/genapic_flat_64.c b/arch/x86/kernel/genapic_flat_64.c index c0262791bda4..50eebd0328fe 100644 --- a/arch/x86/kernel/genapic_flat_64.c +++ b/arch/x86/kernel/genapic_flat_64.c @@ -30,12 +30,12 @@ static int flat_acpi_madt_oem_check(char *oem_id, char *oem_table_id) return 1; } -static cpumask_t flat_target_cpus(void) +static const cpumask_t *flat_target_cpus(void) { - return cpu_online_map; + return &cpu_online_map; } -static cpumask_t flat_vector_allocation_domain(int cpu) +static void flat_vector_allocation_domain(int cpu, cpumask_t *retmask) { /* Careful. Some cpus do not strictly honor the set of cpus * specified in the interrupt destination when using lowest @@ -45,8 +45,7 @@ static cpumask_t flat_vector_allocation_domain(int cpu) * deliver interrupts to the wrong hyperthread when only one * hyperthread was specified in the interrupt desitination. */ - cpumask_t domain = { { [0] = APIC_ALL_CPUS, } }; - return domain; + *retmask = (cpumask_t) { {[0] = APIC_ALL_CPUS, } }; } /* @@ -69,9 +68,8 @@ static void flat_init_apic_ldr(void) apic_write(APIC_LDR, val); } -static void flat_send_IPI_mask(cpumask_t cpumask, int vector) +static inline void _flat_send_IPI_mask(unsigned long mask, int vector) { - unsigned long mask = cpus_addr(cpumask)[0]; unsigned long flags; local_irq_save(flags); @@ -79,20 +77,40 @@ static void flat_send_IPI_mask(cpumask_t cpumask, int vector) local_irq_restore(flags); } +static void flat_send_IPI_mask(const cpumask_t *cpumask, int vector) +{ + unsigned long mask = cpus_addr(*cpumask)[0]; + + _flat_send_IPI_mask(mask, vector); +} + +static void flat_send_IPI_mask_allbutself(const cpumask_t *cpumask, int vector) +{ + unsigned long mask = cpus_addr(*cpumask)[0]; + int cpu = smp_processor_id(); + + if (cpu < BITS_PER_LONG) + clear_bit(cpu, &mask); + _flat_send_IPI_mask(mask, vector); +} + static void flat_send_IPI_allbutself(int vector) { + int cpu = smp_processor_id(); #ifdef CONFIG_HOTPLUG_CPU int hotplug = 1; #else int hotplug = 0; #endif if (hotplug || vector == NMI_VECTOR) { - cpumask_t allbutme = cpu_online_map; + if (!cpus_equal(cpu_online_map, cpumask_of_cpu(cpu))) { + unsigned long mask = cpus_addr(cpu_online_map)[0]; - cpu_clear(smp_processor_id(), allbutme); + if (cpu < BITS_PER_LONG) + clear_bit(cpu, &mask); - if (!cpus_empty(allbutme)) - flat_send_IPI_mask(allbutme, vector); + _flat_send_IPI_mask(mask, vector); + } } else if (num_online_cpus() > 1) { __send_IPI_shortcut(APIC_DEST_ALLBUT, vector,APIC_DEST_LOGICAL); } @@ -101,7 +119,7 @@ static void flat_send_IPI_allbutself(int vector) static void flat_send_IPI_all(int vector) { if (vector == NMI_VECTOR) - flat_send_IPI_mask(cpu_online_map, vector); + flat_send_IPI_mask(&cpu_online_map, vector); else __send_IPI_shortcut(APIC_DEST_ALLINC, vector, APIC_DEST_LOGICAL); } @@ -135,9 +153,9 @@ static int flat_apic_id_registered(void) return physid_isset(read_xapic_id(), phys_cpu_present_map); } -static unsigned int flat_cpu_mask_to_apicid(cpumask_t cpumask) +static unsigned int flat_cpu_mask_to_apicid(const cpumask_t *cpumask) { - return cpus_addr(cpumask)[0] & APIC_ALL_CPUS; + return cpus_addr(*cpumask)[0] & APIC_ALL_CPUS; } static unsigned int phys_pkg_id(int index_msb) @@ -157,6 +175,7 @@ struct genapic apic_flat = { .send_IPI_all = flat_send_IPI_all, .send_IPI_allbutself = flat_send_IPI_allbutself, .send_IPI_mask = flat_send_IPI_mask, + .send_IPI_mask_allbutself = flat_send_IPI_mask_allbutself, .send_IPI_self = apic_send_IPI_self, .cpu_mask_to_apicid = flat_cpu_mask_to_apicid, .phys_pkg_id = phys_pkg_id, @@ -188,35 +207,39 @@ static int physflat_acpi_madt_oem_check(char *oem_id, char *oem_table_id) return 0; } -static cpumask_t physflat_target_cpus(void) +static const cpumask_t *physflat_target_cpus(void) { - return cpu_online_map; + return &cpu_online_map; } -static cpumask_t physflat_vector_allocation_domain(int cpu) +static void physflat_vector_allocation_domain(int cpu, cpumask_t *retmask) { - return cpumask_of_cpu(cpu); + cpus_clear(*retmask); + cpu_set(cpu, *retmask); } -static void physflat_send_IPI_mask(cpumask_t cpumask, int vector) +static void physflat_send_IPI_mask(const cpumask_t *cpumask, int vector) { send_IPI_mask_sequence(cpumask, vector); } -static void physflat_send_IPI_allbutself(int vector) +static void physflat_send_IPI_mask_allbutself(const cpumask_t *cpumask, + int vector) { - cpumask_t allbutme = cpu_online_map; + send_IPI_mask_allbutself(cpumask, vector); +} - cpu_clear(smp_processor_id(), allbutme); - physflat_send_IPI_mask(allbutme, vector); +static void physflat_send_IPI_allbutself(int vector) +{ + send_IPI_mask_allbutself(&cpu_online_map, vector); } static void physflat_send_IPI_all(int vector) { - physflat_send_IPI_mask(cpu_online_map, vector); + physflat_send_IPI_mask(&cpu_online_map, vector); } -static unsigned int physflat_cpu_mask_to_apicid(cpumask_t cpumask) +static unsigned int physflat_cpu_mask_to_apicid(const cpumask_t *cpumask) { int cpu; @@ -224,7 +247,7 @@ static unsigned int physflat_cpu_mask_to_apicid(cpumask_t cpumask) * We're using fixed IRQ delivery, can only return one phys APIC ID. * May as well be the first. */ - cpu = first_cpu(cpumask); + cpu = first_cpu(*cpumask); if ((unsigned)cpu < nr_cpu_ids) return per_cpu(x86_cpu_to_apicid, cpu); else @@ -243,6 +266,7 @@ struct genapic apic_physflat = { .send_IPI_all = physflat_send_IPI_all, .send_IPI_allbutself = physflat_send_IPI_allbutself, .send_IPI_mask = physflat_send_IPI_mask, + .send_IPI_mask_allbutself = physflat_send_IPI_mask_allbutself, .send_IPI_self = apic_send_IPI_self, .cpu_mask_to_apicid = physflat_cpu_mask_to_apicid, .phys_pkg_id = phys_pkg_id, diff --git a/arch/x86/kernel/genx2apic_cluster.c b/arch/x86/kernel/genx2apic_cluster.c index f6a2c8eb48a6..f5fa9a91ad38 100644 --- a/arch/x86/kernel/genx2apic_cluster.c +++ b/arch/x86/kernel/genx2apic_cluster.c @@ -22,19 +22,18 @@ static int x2apic_acpi_madt_oem_check(char *oem_id, char *oem_table_id) /* Start with all IRQs pointing to boot CPU. IRQ balancing will shift them. */ -static cpumask_t x2apic_target_cpus(void) +static const cpumask_t *x2apic_target_cpus(void) { - return cpumask_of_cpu(0); + return &cpumask_of_cpu(0); } /* * for now each logical cpu is in its own vector allocation domain. */ -static cpumask_t x2apic_vector_allocation_domain(int cpu) +static void x2apic_vector_allocation_domain(int cpu, cpumask_t *retmask) { - cpumask_t domain = CPU_MASK_NONE; - cpu_set(cpu, domain); - return domain; + cpus_clear(*retmask); + cpu_set(cpu, *retmask); } static void __x2apic_send_IPI_dest(unsigned int apicid, int vector, @@ -56,32 +55,52 @@ static void __x2apic_send_IPI_dest(unsigned int apicid, int vector, * at once. We have 16 cpu's in a cluster. This will minimize IPI register * writes. */ -static void x2apic_send_IPI_mask(cpumask_t mask, int vector) +static void x2apic_send_IPI_mask(const cpumask_t *mask, int vector) { unsigned long flags; unsigned long query_cpu; local_irq_save(flags); - for_each_cpu_mask(query_cpu, mask) { - __x2apic_send_IPI_dest(per_cpu(x86_cpu_to_logical_apicid, query_cpu), - vector, APIC_DEST_LOGICAL); - } + for_each_cpu_mask_nr(query_cpu, *mask) + __x2apic_send_IPI_dest( + per_cpu(x86_cpu_to_logical_apicid, query_cpu), + vector, APIC_DEST_LOGICAL); local_irq_restore(flags); } -static void x2apic_send_IPI_allbutself(int vector) +static void x2apic_send_IPI_mask_allbutself(const cpumask_t *mask, int vector) { - cpumask_t mask = cpu_online_map; + unsigned long flags; + unsigned long query_cpu; + unsigned long this_cpu = smp_processor_id(); - cpu_clear(smp_processor_id(), mask); + local_irq_save(flags); + for_each_cpu_mask_nr(query_cpu, *mask) + if (query_cpu != this_cpu) + __x2apic_send_IPI_dest( + per_cpu(x86_cpu_to_logical_apicid, query_cpu), + vector, APIC_DEST_LOGICAL); + local_irq_restore(flags); +} - if (!cpus_empty(mask)) - x2apic_send_IPI_mask(mask, vector); +static void x2apic_send_IPI_allbutself(int vector) +{ + unsigned long flags; + unsigned long query_cpu; + unsigned long this_cpu = smp_processor_id(); + + local_irq_save(flags); + for_each_online_cpu(query_cpu) + if (query_cpu != this_cpu) + __x2apic_send_IPI_dest( + per_cpu(x86_cpu_to_logical_apicid, query_cpu), + vector, APIC_DEST_LOGICAL); + local_irq_restore(flags); } static void x2apic_send_IPI_all(int vector) { - x2apic_send_IPI_mask(cpu_online_map, vector); + x2apic_send_IPI_mask(&cpu_online_map, vector); } static int x2apic_apic_id_registered(void) @@ -89,7 +108,7 @@ static int x2apic_apic_id_registered(void) return 1; } -static unsigned int x2apic_cpu_mask_to_apicid(cpumask_t cpumask) +static unsigned int x2apic_cpu_mask_to_apicid(const cpumask_t *cpumask) { int cpu; @@ -97,8 +116,8 @@ static unsigned int x2apic_cpu_mask_to_apicid(cpumask_t cpumask) * We're using fixed IRQ delivery, can only return one phys APIC ID. * May as well be the first. */ - cpu = first_cpu(cpumask); - if ((unsigned)cpu < NR_CPUS) + cpu = first_cpu(*cpumask); + if ((unsigned)cpu < nr_cpu_ids) return per_cpu(x86_cpu_to_logical_apicid, cpu); else return BAD_APICID; @@ -150,6 +169,7 @@ struct genapic apic_x2apic_cluster = { .send_IPI_all = x2apic_send_IPI_all, .send_IPI_allbutself = x2apic_send_IPI_allbutself, .send_IPI_mask = x2apic_send_IPI_mask, + .send_IPI_mask_allbutself = x2apic_send_IPI_mask_allbutself, .send_IPI_self = x2apic_send_IPI_self, .cpu_mask_to_apicid = x2apic_cpu_mask_to_apicid, .phys_pkg_id = phys_pkg_id, diff --git a/arch/x86/kernel/genx2apic_phys.c b/arch/x86/kernel/genx2apic_phys.c index d042211768b7..41c27b2f3d01 100644 --- a/arch/x86/kernel/genx2apic_phys.c +++ b/arch/x86/kernel/genx2apic_phys.c @@ -29,16 +29,15 @@ static int x2apic_acpi_madt_oem_check(char *oem_id, char *oem_table_id) /* Start with all IRQs pointing to boot CPU. IRQ balancing will shift them. */ -static cpumask_t x2apic_target_cpus(void) +static const cpumask_t *x2apic_target_cpus(void) { - return cpumask_of_cpu(0); + return &cpumask_of_cpu(0); } -static cpumask_t x2apic_vector_allocation_domain(int cpu) +static void x2apic_vector_allocation_domain(int cpu, cpumask_t *retmask) { - cpumask_t domain = CPU_MASK_NONE; - cpu_set(cpu, domain); - return domain; + cpus_clear(*retmask); + cpu_set(cpu, *retmask); } static void __x2apic_send_IPI_dest(unsigned int apicid, int vector, @@ -54,32 +53,53 @@ static void __x2apic_send_IPI_dest(unsigned int apicid, int vector, x2apic_icr_write(cfg, apicid); } -static void x2apic_send_IPI_mask(cpumask_t mask, int vector) +static void x2apic_send_IPI_mask(const cpumask_t *mask, int vector) { unsigned long flags; unsigned long query_cpu; local_irq_save(flags); - for_each_cpu_mask(query_cpu, mask) { + for_each_cpu_mask_nr(query_cpu, *mask) { __x2apic_send_IPI_dest(per_cpu(x86_cpu_to_apicid, query_cpu), vector, APIC_DEST_PHYSICAL); } local_irq_restore(flags); } -static void x2apic_send_IPI_allbutself(int vector) +static void x2apic_send_IPI_mask_allbutself(const cpumask_t *mask, int vector) { - cpumask_t mask = cpu_online_map; + unsigned long flags; + unsigned long query_cpu; + unsigned long this_cpu = smp_processor_id(); + + local_irq_save(flags); + for_each_cpu_mask_nr(query_cpu, *mask) { + if (query_cpu != this_cpu) + __x2apic_send_IPI_dest( + per_cpu(x86_cpu_to_apicid, query_cpu), + vector, APIC_DEST_PHYSICAL); + } + local_irq_restore(flags); +} - cpu_clear(smp_processor_id(), mask); +static void x2apic_send_IPI_allbutself(int vector) +{ + unsigned long flags; + unsigned long query_cpu; + unsigned long this_cpu = smp_processor_id(); - if (!cpus_empty(mask)) - x2apic_send_IPI_mask(mask, vector); + local_irq_save(flags); + for_each_online_cpu(query_cpu) + if (query_cpu != this_cpu) + __x2apic_send_IPI_dest( + per_cpu(x86_cpu_to_apicid, query_cpu), + vector, APIC_DEST_PHYSICAL); + local_irq_restore(flags); } static void x2apic_send_IPI_all(int vector) { - x2apic_send_IPI_mask(cpu_online_map, vector); + x2apic_send_IPI_mask(&cpu_online_map, vector); } static int x2apic_apic_id_registered(void) @@ -87,7 +107,7 @@ static int x2apic_apic_id_registered(void) return 1; } -static unsigned int x2apic_cpu_mask_to_apicid(cpumask_t cpumask) +static unsigned int x2apic_cpu_mask_to_apicid(const cpumask_t *cpumask) { int cpu; @@ -95,8 +115,8 @@ static unsigned int x2apic_cpu_mask_to_apicid(cpumask_t cpumask) * We're using fixed IRQ delivery, can only return one phys APIC ID. * May as well be the first. */ - cpu = first_cpu(cpumask); - if ((unsigned)cpu < NR_CPUS) + cpu = first_cpu(*cpumask); + if ((unsigned)cpu < nr_cpu_ids) return per_cpu(x86_cpu_to_apicid, cpu); else return BAD_APICID; @@ -145,6 +165,7 @@ struct genapic apic_x2apic_phys = { .send_IPI_all = x2apic_send_IPI_all, .send_IPI_allbutself = x2apic_send_IPI_allbutself, .send_IPI_mask = x2apic_send_IPI_mask, + .send_IPI_mask_allbutself = x2apic_send_IPI_mask_allbutself, .send_IPI_self = x2apic_send_IPI_self, .cpu_mask_to_apicid = x2apic_cpu_mask_to_apicid, .phys_pkg_id = phys_pkg_id, diff --git a/arch/x86/kernel/genx2apic_uv_x.c b/arch/x86/kernel/genx2apic_uv_x.c index 2c7dbdb98278..010659415ae4 100644 --- a/arch/x86/kernel/genx2apic_uv_x.c +++ b/arch/x86/kernel/genx2apic_uv_x.c @@ -75,16 +75,15 @@ EXPORT_SYMBOL(sn_rtc_cycles_per_second); /* Start with all IRQs pointing to boot CPU. IRQ balancing will shift them. */ -static cpumask_t uv_target_cpus(void) +static const cpumask_t *uv_target_cpus(void) { - return cpumask_of_cpu(0); + return &cpumask_of_cpu(0); } -static cpumask_t uv_vector_allocation_domain(int cpu) +static void uv_vector_allocation_domain(int cpu, cpumask_t *retmask) { - cpumask_t domain = CPU_MASK_NONE; - cpu_set(cpu, domain); - return domain; + cpus_clear(*retmask); + cpu_set(cpu, *retmask); } int uv_wakeup_secondary(int phys_apicid, unsigned int start_rip) @@ -123,28 +122,37 @@ static void uv_send_IPI_one(int cpu, int vector) uv_write_global_mmr64(pnode, UVH_IPI_INT, val); } -static void uv_send_IPI_mask(cpumask_t mask, int vector) +static void uv_send_IPI_mask(const cpumask_t *mask, int vector) { unsigned int cpu; - for_each_possible_cpu(cpu) - if (cpu_isset(cpu, mask)) + for_each_cpu_mask_nr(cpu, *mask) + uv_send_IPI_one(cpu, vector); +} + +static void uv_send_IPI_mask_allbutself(const cpumask_t *mask, int vector) +{ + unsigned int cpu; + unsigned int this_cpu = smp_processor_id(); + + for_each_cpu_mask_nr(cpu, *mask) + if (cpu != this_cpu) uv_send_IPI_one(cpu, vector); } static void uv_send_IPI_allbutself(int vector) { - cpumask_t mask = cpu_online_map; - - cpu_clear(smp_processor_id(), mask); + unsigned int cpu; + unsigned int this_cpu = smp_processor_id(); - if (!cpus_empty(mask)) - uv_send_IPI_mask(mask, vector); + for_each_online_cpu(cpu) + if (cpu != this_cpu) + uv_send_IPI_one(cpu, vector); } static void uv_send_IPI_all(int vector) { - uv_send_IPI_mask(cpu_online_map, vector); + uv_send_IPI_mask(&cpu_online_map, vector); } static int uv_apic_id_registered(void) @@ -156,7 +164,7 @@ static void uv_init_apic_ldr(void) { } -static unsigned int uv_cpu_mask_to_apicid(cpumask_t cpumask) +static unsigned int uv_cpu_mask_to_apicid(const cpumask_t *cpumask) { int cpu; @@ -164,7 +172,7 @@ static unsigned int uv_cpu_mask_to_apicid(cpumask_t cpumask) * We're using fixed IRQ delivery, can only return one phys APIC ID. * May as well be the first. */ - cpu = first_cpu(cpumask); + cpu = first_cpu(*cpumask); if ((unsigned)cpu < nr_cpu_ids) return per_cpu(x86_cpu_to_apicid, cpu); else @@ -218,6 +226,7 @@ struct genapic apic_x2apic_uv_x = { .send_IPI_all = uv_send_IPI_all, .send_IPI_allbutself = uv_send_IPI_allbutself, .send_IPI_mask = uv_send_IPI_mask, + .send_IPI_mask_allbutself = uv_send_IPI_mask_allbutself, .send_IPI_self = uv_send_IPI_self, .cpu_mask_to_apicid = uv_cpu_mask_to_apicid, .phys_pkg_id = phys_pkg_id, diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index 3d7d0d55253f..7f23ce7f5518 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -231,7 +231,8 @@ static struct irq_cfg *irq_cfg(unsigned int irq) #endif -static inline void set_extra_move_desc(struct irq_desc *desc, cpumask_t mask) +static inline void +set_extra_move_desc(struct irq_desc *desc, const struct cpumask *mask) { } @@ -396,7 +397,8 @@ static void __target_IO_APIC_irq(unsigned int irq, unsigned int dest, struct irq } } -static int assign_irq_vector(int irq, struct irq_cfg *cfg, cpumask_t mask); +static int +assign_irq_vector(int irq, struct irq_cfg *cfg, const struct cpumask *mask); static void set_ioapic_affinity_irq_desc(struct irq_desc *desc, const struct cpumask *mask) @@ -412,13 +414,13 @@ static void set_ioapic_affinity_irq_desc(struct irq_desc *desc, irq = desc->irq; cfg = desc->chip_data; - if (assign_irq_vector(irq, cfg, *mask)) + if (assign_irq_vector(irq, cfg, mask)) return; - set_extra_move_desc(desc, *mask); + set_extra_move_desc(desc, mask); cpumask_and(&tmp, &cfg->domain, mask); - dest = cpu_mask_to_apicid(tmp); + dest = cpu_mask_to_apicid(&tmp); /* * Only the high 8 bits are valid. */ @@ -1099,7 +1101,8 @@ void unlock_vector_lock(void) spin_unlock(&vector_lock); } -static int __assign_irq_vector(int irq, struct irq_cfg *cfg, cpumask_t mask) +static int +__assign_irq_vector(int irq, struct irq_cfg *cfg, const struct cpumask *mask) { /* * NOTE! The local APIC isn't very good at handling @@ -1115,35 +1118,32 @@ static int __assign_irq_vector(int irq, struct irq_cfg *cfg, cpumask_t mask) static int current_vector = FIRST_DEVICE_VECTOR, current_offset = 0; unsigned int old_vector; int cpu; + cpumask_t tmp_mask; if ((cfg->move_in_progress) || cfg->move_cleanup_count) return -EBUSY; - /* Only try and allocate irqs on cpus that are present */ - cpus_and(mask, mask, cpu_online_map); - old_vector = cfg->vector; if (old_vector) { - cpumask_t tmp; - cpus_and(tmp, cfg->domain, mask); - if (!cpus_empty(tmp)) + cpus_and(tmp_mask, *mask, cpu_online_map); + cpus_and(tmp_mask, cfg->domain, tmp_mask); + if (!cpus_empty(tmp_mask)) return 0; } - for_each_cpu_mask_nr(cpu, mask) { - cpumask_t domain, new_mask; + /* Only try and allocate irqs on cpus that are present */ + for_each_cpu_and(cpu, mask, &cpu_online_map) { int new_cpu; int vector, offset; - domain = vector_allocation_domain(cpu); - cpus_and(new_mask, domain, cpu_online_map); + vector_allocation_domain(cpu, &tmp_mask); vector = current_vector; offset = current_offset; next: vector += 8; if (vector >= first_system_vector) { - /* If we run out of vectors on large boxen, must share them. */ + /* If out of vectors on large boxen, must share them. */ offset = (offset + 1) % 8; vector = FIRST_DEVICE_VECTOR + offset; } @@ -1156,7 +1156,7 @@ next: if (vector == SYSCALL_VECTOR) goto next; #endif - for_each_cpu_mask_nr(new_cpu, new_mask) + for_each_cpu_and(new_cpu, &tmp_mask, &cpu_online_map) if (per_cpu(vector_irq, new_cpu)[vector] != -1) goto next; /* Found one! */ @@ -1166,16 +1166,17 @@ next: cfg->move_in_progress = 1; cfg->old_domain = cfg->domain; } - for_each_cpu_mask_nr(new_cpu, new_mask) + for_each_cpu_and(new_cpu, &tmp_mask, &cpu_online_map) per_cpu(vector_irq, new_cpu)[vector] = irq; cfg->vector = vector; - cfg->domain = domain; + cfg->domain = tmp_mask; return 0; } return -ENOSPC; } -static int assign_irq_vector(int irq, struct irq_cfg *cfg, cpumask_t mask) +static int +assign_irq_vector(int irq, struct irq_cfg *cfg, const struct cpumask *mask) { int err; unsigned long flags; @@ -1384,8 +1385,8 @@ static void setup_IO_APIC_irq(int apic, int pin, unsigned int irq, struct irq_de cfg = desc->chip_data; - mask = TARGET_CPUS; - if (assign_irq_vector(irq, cfg, mask)) + mask = *TARGET_CPUS; + if (assign_irq_vector(irq, cfg, &mask)) return; cpus_and(mask, cfg->domain, mask); @@ -1398,7 +1399,7 @@ static void setup_IO_APIC_irq(int apic, int pin, unsigned int irq, struct irq_de if (setup_ioapic_entry(mp_ioapics[apic].mp_apicid, irq, &entry, - cpu_mask_to_apicid(mask), trigger, polarity, + cpu_mask_to_apicid(&mask), trigger, polarity, cfg->vector)) { printk("Failed to setup ioapic entry for ioapic %d, pin %d\n", mp_ioapics[apic].mp_apicid, pin); @@ -2121,7 +2122,7 @@ static int ioapic_retrigger_irq(unsigned int irq) unsigned long flags; spin_lock_irqsave(&vector_lock, flags); - send_IPI_mask(cpumask_of_cpu(first_cpu(cfg->domain)), cfg->vector); + send_IPI_mask(&cpumask_of_cpu(first_cpu(cfg->domain)), cfg->vector); spin_unlock_irqrestore(&vector_lock, flags); return 1; @@ -2170,18 +2171,19 @@ static DECLARE_DELAYED_WORK(ir_migration_work, ir_irq_migration); * as simple as edge triggered migration and we can do the irq migration * with a simple atomic update to IO-APIC RTE. */ -static void migrate_ioapic_irq_desc(struct irq_desc *desc, cpumask_t mask) +static void +migrate_ioapic_irq_desc(struct irq_desc *desc, const struct cpumask *mask) { struct irq_cfg *cfg; - cpumask_t tmp, cleanup_mask; + cpumask_t tmpmask; struct irte irte; int modify_ioapic_rte; unsigned int dest; unsigned long flags; unsigned int irq; - cpus_and(tmp, mask, cpu_online_map); - if (cpus_empty(tmp)) + cpus_and(tmpmask, *mask, cpu_online_map); + if (cpus_empty(tmpmask)) return; irq = desc->irq; @@ -2194,8 +2196,8 @@ static void migrate_ioapic_irq_desc(struct irq_desc *desc, cpumask_t mask) set_extra_move_desc(desc, mask); - cpus_and(tmp, cfg->domain, mask); - dest = cpu_mask_to_apicid(tmp); + cpus_and(tmpmask, cfg->domain, *mask); + dest = cpu_mask_to_apicid(&tmpmask); modify_ioapic_rte = desc->status & IRQ_LEVEL; if (modify_ioapic_rte) { @@ -2213,13 +2215,13 @@ static void migrate_ioapic_irq_desc(struct irq_desc *desc, cpumask_t mask) modify_irte(irq, &irte); if (cfg->move_in_progress) { - cpus_and(cleanup_mask, cfg->old_domain, cpu_online_map); - cfg->move_cleanup_count = cpus_weight(cleanup_mask); - send_IPI_mask(cleanup_mask, IRQ_MOVE_CLEANUP_VECTOR); + cpus_and(tmpmask, cfg->old_domain, cpu_online_map); + cfg->move_cleanup_count = cpus_weight(tmpmask); + send_IPI_mask(&tmpmask, IRQ_MOVE_CLEANUP_VECTOR); cfg->move_in_progress = 0; } - desc->affinity = mask; + desc->affinity = *mask; } static int migrate_irq_remapped_level_desc(struct irq_desc *desc) @@ -2241,7 +2243,7 @@ static int migrate_irq_remapped_level_desc(struct irq_desc *desc) } /* everthing is clear. we have right of way */ - migrate_ioapic_irq_desc(desc, desc->pending_mask); + migrate_ioapic_irq_desc(desc, &desc->pending_mask); ret = 0; desc->status &= ~IRQ_MOVE_PENDING; @@ -2292,7 +2294,7 @@ static void set_ir_ioapic_affinity_irq_desc(struct irq_desc *desc, return; } - migrate_ioapic_irq_desc(desc, *mask); + migrate_ioapic_irq_desc(desc, mask); } static void set_ir_ioapic_affinity_irq(unsigned int irq, const struct cpumask *mask) @@ -2359,7 +2361,7 @@ static void irq_complete_move(struct irq_desc **descp) cpus_and(cleanup_mask, cfg->old_domain, cpu_online_map); cfg->move_cleanup_count = cpus_weight(cleanup_mask); - send_IPI_mask(cleanup_mask, IRQ_MOVE_CLEANUP_VECTOR); + send_IPI_mask(&cleanup_mask, IRQ_MOVE_CLEANUP_VECTOR); cfg->move_in_progress = 0; } } @@ -3089,13 +3091,13 @@ static int msi_compose_msg(struct pci_dev *pdev, unsigned int irq, struct msi_ms cpumask_t tmp; cfg = irq_cfg(irq); - tmp = TARGET_CPUS; - err = assign_irq_vector(irq, cfg, tmp); + tmp = *TARGET_CPUS; + err = assign_irq_vector(irq, cfg, &tmp); if (err) return err; cpus_and(tmp, cfg->domain, tmp); - dest = cpu_mask_to_apicid(tmp); + dest = cpu_mask_to_apicid(&tmp); #ifdef CONFIG_INTR_REMAP if (irq_remapped(irq)) { @@ -3161,13 +3163,13 @@ static void set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) return; cfg = desc->chip_data; - if (assign_irq_vector(irq, cfg, *mask)) + if (assign_irq_vector(irq, cfg, mask)) return; - set_extra_move_desc(desc, *mask); + set_extra_move_desc(desc, mask); cpumask_and(&tmp, &cfg->domain, mask); - dest = cpu_mask_to_apicid(tmp); + dest = cpu_mask_to_apicid(&tmp); read_msi_msg_desc(desc, &msg); @@ -3184,8 +3186,8 @@ static void set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) * Migrate the MSI irq to another cpumask. This migration is * done in the process context using interrupt-remapping hardware. */ -static void ir_set_msi_irq_affinity(unsigned int irq, - const struct cpumask *mask) +static void +ir_set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) { struct irq_desc *desc = irq_to_desc(irq); struct irq_cfg *cfg; @@ -3200,13 +3202,13 @@ static void ir_set_msi_irq_affinity(unsigned int irq, return; cfg = desc->chip_data; - if (assign_irq_vector(irq, cfg, *mask)) + if (assign_irq_vector(irq, cfg, mask)) return; - set_extra_move_desc(desc, *mask); + set_extra_move_desc(desc, mask); cpumask_and(&tmp, &cfg->domain, mask); - dest = cpu_mask_to_apicid(tmp); + dest = cpu_mask_to_apicid(&tmp); irte.vector = cfg->vector; irte.dest_id = IRTE_DEST(dest); @@ -3224,7 +3226,7 @@ static void ir_set_msi_irq_affinity(unsigned int irq, if (cfg->move_in_progress) { cpus_and(cleanup_mask, cfg->old_domain, cpu_online_map); cfg->move_cleanup_count = cpus_weight(cleanup_mask); - send_IPI_mask(cleanup_mask, IRQ_MOVE_CLEANUP_VECTOR); + send_IPI_mask(&cleanup_mask, IRQ_MOVE_CLEANUP_VECTOR); cfg->move_in_progress = 0; } @@ -3419,7 +3421,7 @@ void arch_teardown_msi_irq(unsigned int irq) #ifdef CONFIG_DMAR #ifdef CONFIG_SMP -static void dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask) +static void dmar_msi_set_affinity(unsigned int irq, const cpumask_t *mask) { struct irq_desc *desc = irq_to_desc(irq); struct irq_cfg *cfg; @@ -3431,13 +3433,13 @@ static void dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask) return; cfg = desc->chip_data; - if (assign_irq_vector(irq, cfg, *mask)) + if (assign_irq_vector(irq, cfg, mask)) return; - set_extra_move_desc(desc, *mask); + set_extra_move_desc(desc, mask); cpumask_and(&tmp, &cfg->domain, mask); - dest = cpu_mask_to_apicid(tmp); + dest = cpu_mask_to_apicid(&tmp); dmar_msi_read(irq, &msg); @@ -3481,7 +3483,7 @@ int arch_setup_dmar_msi(unsigned int irq) #ifdef CONFIG_HPET_TIMER #ifdef CONFIG_SMP -static void hpet_msi_set_affinity(unsigned int irq, const struct cpumask *mask) +static void hpet_msi_set_affinity(unsigned int irq, const cpumask_t *mask) { struct irq_desc *desc = irq_to_desc(irq); struct irq_cfg *cfg; @@ -3493,13 +3495,13 @@ static void hpet_msi_set_affinity(unsigned int irq, const struct cpumask *mask) return; cfg = desc->chip_data; - if (assign_irq_vector(irq, cfg, *mask)) + if (assign_irq_vector(irq, cfg, mask)) return; - set_extra_move_desc(desc, *mask); + set_extra_move_desc(desc, mask); cpumask_and(&tmp, &cfg->domain, mask); - dest = cpu_mask_to_apicid(tmp); + dest = cpu_mask_to_apicid(&tmp); hpet_msi_read(irq, &msg); @@ -3564,7 +3566,7 @@ static void target_ht_irq(unsigned int irq, unsigned int dest, u8 vector) write_ht_irq_msg(irq, &msg); } -static void set_ht_irq_affinity(unsigned int irq, const struct cpumask *mask) +static void set_ht_irq_affinity(unsigned int irq, const cpumask_t *mask) { struct irq_desc *desc = irq_to_desc(irq); struct irq_cfg *cfg; @@ -3575,13 +3577,13 @@ static void set_ht_irq_affinity(unsigned int irq, const struct cpumask *mask) return; cfg = desc->chip_data; - if (assign_irq_vector(irq, cfg, *mask)) + if (assign_irq_vector(irq, cfg, mask)) return; - set_extra_move_desc(desc, *mask); + set_extra_move_desc(desc, mask); cpumask_and(&tmp, &cfg->domain, mask); - dest = cpu_mask_to_apicid(tmp); + dest = cpu_mask_to_apicid(&tmp); target_ht_irq(irq, dest, cfg->vector); cpumask_copy(&desc->affinity, mask); @@ -3607,14 +3609,13 @@ int arch_setup_ht_irq(unsigned int irq, struct pci_dev *dev) cpumask_t tmp; cfg = irq_cfg(irq); - tmp = TARGET_CPUS; - err = assign_irq_vector(irq, cfg, tmp); + err = assign_irq_vector(irq, cfg, TARGET_CPUS); if (!err) { struct ht_irq_msg msg; unsigned dest; cpus_and(tmp, cfg->domain, tmp); - dest = cpu_mask_to_apicid(tmp); + dest = cpu_mask_to_apicid(&tmp); msg.address_hi = HT_IRQ_HIGH_DEST_ID(dest); @@ -3650,7 +3651,7 @@ int arch_setup_ht_irq(unsigned int irq, struct pci_dev *dev) int arch_enable_uv_irq(char *irq_name, unsigned int irq, int cpu, int mmr_blade, unsigned long mmr_offset) { - const cpumask_t *eligible_cpu = get_cpu_mask(cpu); + const cpumask_t *eligible_cpu = &cpumask_of_cpu(cpu); struct irq_cfg *cfg; int mmr_pnode; unsigned long mmr_value; @@ -3660,7 +3661,7 @@ int arch_enable_uv_irq(char *irq_name, unsigned int irq, int cpu, int mmr_blade, cfg = irq_cfg(irq); - err = assign_irq_vector(irq, cfg, *eligible_cpu); + err = assign_irq_vector(irq, cfg, eligible_cpu); if (err != 0) return err; @@ -3679,7 +3680,7 @@ int arch_enable_uv_irq(char *irq_name, unsigned int irq, int cpu, int mmr_blade, entry->polarity = 0; entry->trigger = 0; entry->mask = 0; - entry->dest = cpu_mask_to_apicid(*eligible_cpu); + entry->dest = cpu_mask_to_apicid(eligible_cpu); mmr_pnode = uv_blade_to_pnode(mmr_blade); uv_write_global_mmr64(mmr_pnode, mmr_offset, mmr_value); @@ -3890,7 +3891,7 @@ void __init setup_ioapic_dest(void) int pin, ioapic, irq, irq_entry; struct irq_desc *desc; struct irq_cfg *cfg; - cpumask_t mask; + const cpumask_t *mask; if (skip_ioapic_setup == 1) return; @@ -3921,16 +3922,16 @@ void __init setup_ioapic_dest(void) */ if (desc->status & (IRQ_NO_BALANCING | IRQ_AFFINITY_SET)) - mask = desc->affinity; + mask = &desc->affinity; else mask = TARGET_CPUS; #ifdef CONFIG_INTR_REMAP if (intr_remapping_enabled) - set_ir_ioapic_affinity_irq_desc(desc, &mask); + set_ir_ioapic_affinity_irq_desc(desc, mask); else #endif - set_ioapic_affinity_irq_desc(desc, &mask); + set_ioapic_affinity_irq_desc(desc, mask); } } diff --git a/arch/x86/kernel/ipi.c b/arch/x86/kernel/ipi.c index f1c688e46f35..86aa50fc65a1 100644 --- a/arch/x86/kernel/ipi.c +++ b/arch/x86/kernel/ipi.c @@ -116,9 +116,9 @@ static inline void __send_IPI_dest_field(unsigned long mask, int vector) /* * This is only used on smaller machines. */ -void send_IPI_mask_bitmask(cpumask_t cpumask, int vector) +void send_IPI_mask_bitmask(const cpumask_t *cpumask, int vector) { - unsigned long mask = cpus_addr(cpumask)[0]; + unsigned long mask = cpus_addr(*cpumask)[0]; unsigned long flags; local_irq_save(flags); @@ -127,7 +127,7 @@ void send_IPI_mask_bitmask(cpumask_t cpumask, int vector) local_irq_restore(flags); } -void send_IPI_mask_sequence(cpumask_t mask, int vector) +void send_IPI_mask_sequence(const cpumask_t *mask, int vector) { unsigned long flags; unsigned int query_cpu; @@ -139,12 +139,24 @@ void send_IPI_mask_sequence(cpumask_t mask, int vector) */ local_irq_save(flags); - for_each_possible_cpu(query_cpu) { - if (cpu_isset(query_cpu, mask)) { + for_each_cpu_mask_nr(query_cpu, *mask) + __send_IPI_dest_field(cpu_to_logical_apicid(query_cpu), vector); + local_irq_restore(flags); +} + +void send_IPI_mask_allbutself(const cpumask_t *mask, int vector) +{ + unsigned long flags; + unsigned int query_cpu; + unsigned int this_cpu = smp_processor_id(); + + /* See Hack comment above */ + + local_irq_save(flags); + for_each_cpu_mask_nr(query_cpu, *mask) + if (query_cpu != this_cpu) __send_IPI_dest_field(cpu_to_logical_apicid(query_cpu), vector); - } - } local_irq_restore(flags); } diff --git a/arch/x86/kernel/smp.c b/arch/x86/kernel/smp.c index 3f92b134ab90..341df946f9a9 100644 --- a/arch/x86/kernel/smp.c +++ b/arch/x86/kernel/smp.c @@ -118,22 +118,22 @@ static void native_smp_send_reschedule(int cpu) WARN_ON(1); return; } - send_IPI_mask(cpumask_of_cpu(cpu), RESCHEDULE_VECTOR); + send_IPI_mask(&cpumask_of_cpu(cpu), RESCHEDULE_VECTOR); } void native_send_call_func_single_ipi(int cpu) { - send_IPI_mask(cpumask_of_cpu(cpu), CALL_FUNCTION_SINGLE_VECTOR); + send_IPI_mask(&cpumask_of_cpu(cpu), CALL_FUNCTION_SINGLE_VECTOR); } -void native_send_call_func_ipi(cpumask_t mask) +void native_send_call_func_ipi(const cpumask_t *mask) { cpumask_t allbutself; allbutself = cpu_online_map; cpu_clear(smp_processor_id(), allbutself); - if (cpus_equal(mask, allbutself) && + if (cpus_equal(*mask, allbutself) && cpus_equal(cpu_online_map, cpu_callout_map)) send_IPI_allbutself(CALL_FUNCTION_VECTOR); else diff --git a/arch/x86/kernel/tlb_32.c b/arch/x86/kernel/tlb_32.c index f4049f3513b6..174ea90d1cbd 100644 --- a/arch/x86/kernel/tlb_32.c +++ b/arch/x86/kernel/tlb_32.c @@ -164,7 +164,7 @@ void native_flush_tlb_others(const cpumask_t *cpumaskp, struct mm_struct *mm, * We have to send the IPI only to * CPUs affected. */ - send_IPI_mask(cpumask, INVALIDATE_TLB_VECTOR); + send_IPI_mask(&cpumask, INVALIDATE_TLB_VECTOR); while (!cpus_empty(flush_cpumask)) /* nothing. lockup detection does not belong here */ diff --git a/arch/x86/kernel/tlb_64.c b/arch/x86/kernel/tlb_64.c index 8f919ca69494..de6f1bda0c50 100644 --- a/arch/x86/kernel/tlb_64.c +++ b/arch/x86/kernel/tlb_64.c @@ -191,7 +191,7 @@ void native_flush_tlb_others(const cpumask_t *cpumaskp, struct mm_struct *mm, * We have to send the IPI only to * CPUs affected. */ - send_IPI_mask(cpumask, INVALIDATE_TLB_VECTOR_START + sender); + send_IPI_mask(&cpumask, INVALIDATE_TLB_VECTOR_START + sender); while (!cpus_empty(f->flush_cpumask)) cpu_relax(); diff --git a/arch/x86/mach-generic/bigsmp.c b/arch/x86/mach-generic/bigsmp.c index 3624a364b7f3..bc4c7840b2a8 100644 --- a/arch/x86/mach-generic/bigsmp.c +++ b/arch/x86/mach-generic/bigsmp.c @@ -42,9 +42,10 @@ static const struct dmi_system_id bigsmp_dmi_table[] = { { } }; -static cpumask_t vector_allocation_domain(int cpu) +static void vector_allocation_domain(int cpu, cpumask_t *retmask) { - return cpumask_of_cpu(cpu); + cpus_clear(*retmask); + cpu_set(cpu, *retmask); } static int probe_bigsmp(void) diff --git a/arch/x86/mach-generic/es7000.c b/arch/x86/mach-generic/es7000.c index 7b4e6d0d1690..4ba5ccaa1584 100644 --- a/arch/x86/mach-generic/es7000.c +++ b/arch/x86/mach-generic/es7000.c @@ -87,7 +87,7 @@ static int __init acpi_madt_oem_check(char *oem_id, char *oem_table_id) } #endif -static cpumask_t vector_allocation_domain(int cpu) +static void vector_allocation_domain(int cpu, cpumask_t *retmask) { /* Careful. Some cpus do not strictly honor the set of cpus * specified in the interrupt destination when using lowest @@ -97,8 +97,7 @@ static cpumask_t vector_allocation_domain(int cpu) * deliver interrupts to the wrong hyperthread when only one * hyperthread was specified in the interrupt desitination. */ - cpumask_t domain = { { [0] = APIC_ALL_CPUS, } }; - return domain; + *retmask = (cpumask_t){ { [0] = APIC_ALL_CPUS, } }; } struct genapic __initdata_refok apic_es7000 = APIC_INIT("es7000", probe_es7000); diff --git a/arch/x86/mach-generic/numaq.c b/arch/x86/mach-generic/numaq.c index 71a309b122e6..511d7941364f 100644 --- a/arch/x86/mach-generic/numaq.c +++ b/arch/x86/mach-generic/numaq.c @@ -38,7 +38,7 @@ static int acpi_madt_oem_check(char *oem_id, char *oem_table_id) return 0; } -static cpumask_t vector_allocation_domain(int cpu) +static void vector_allocation_domain(int cpu, cpumask_t *retmask) { /* Careful. Some cpus do not strictly honor the set of cpus * specified in the interrupt destination when using lowest @@ -48,8 +48,7 @@ static cpumask_t vector_allocation_domain(int cpu) * deliver interrupts to the wrong hyperthread when only one * hyperthread was specified in the interrupt desitination. */ - cpumask_t domain = { { [0] = APIC_ALL_CPUS, } }; - return domain; + *retmask = (cpumask_t){ { [0] = APIC_ALL_CPUS, } }; } struct genapic apic_numaq = APIC_INIT("NUMAQ", probe_numaq); diff --git a/arch/x86/mach-generic/summit.c b/arch/x86/mach-generic/summit.c index 2c6d234e0009..2821ffc188b5 100644 --- a/arch/x86/mach-generic/summit.c +++ b/arch/x86/mach-generic/summit.c @@ -24,7 +24,7 @@ static int probe_summit(void) return 0; } -static cpumask_t vector_allocation_domain(int cpu) +static void vector_allocation_domain(int cpu, cpumask_t *retmask) { /* Careful. Some cpus do not strictly honor the set of cpus * specified in the interrupt destination when using lowest @@ -34,8 +34,7 @@ static cpumask_t vector_allocation_domain(int cpu) * deliver interrupts to the wrong hyperthread when only one * hyperthread was specified in the interrupt desitination. */ - cpumask_t domain = { { [0] = APIC_ALL_CPUS, } }; - return domain; + *retmask = (cpumask_t){ { [0] = APIC_ALL_CPUS, } }; } struct genapic apic_summit = APIC_INIT("summit", probe_summit); diff --git a/arch/x86/xen/smp.c b/arch/x86/xen/smp.c index acd9b6705e02..2cce362c9874 100644 --- a/arch/x86/xen/smp.c +++ b/arch/x86/xen/smp.c @@ -158,7 +158,7 @@ static void __init xen_fill_possible_map(void) { int i, rc; - for (i = 0; i < NR_CPUS; i++) { + for (i = 0; i < nr_cpu_ids; i++) { rc = HYPERVISOR_vcpu_op(VCPUOP_is_up, i, NULL); if (rc >= 0) { num_processors++; @@ -196,7 +196,7 @@ static void __init xen_smp_prepare_cpus(unsigned int max_cpus) /* Restrict the possible_map according to max_cpus. */ while ((num_possible_cpus() > 1) && (num_possible_cpus() > max_cpus)) { - for (cpu = NR_CPUS - 1; !cpu_possible(cpu); cpu--) + for (cpu = nr_cpu_ids - 1; !cpu_possible(cpu); cpu--) continue; cpu_clear(cpu, cpu_possible_map); } @@ -408,24 +408,22 @@ static void xen_smp_send_reschedule(int cpu) xen_send_IPI_one(cpu, XEN_RESCHEDULE_VECTOR); } -static void xen_send_IPI_mask(cpumask_t mask, enum ipi_vector vector) +static void xen_send_IPI_mask(const cpumask_t *mask, enum ipi_vector vector) { unsigned cpu; - cpus_and(mask, mask, cpu_online_map); - - for_each_cpu_mask_nr(cpu, mask) + for_each_cpu_and(cpu, mask, &cpu_online_map) xen_send_IPI_one(cpu, vector); } -static void xen_smp_send_call_function_ipi(cpumask_t mask) +static void xen_smp_send_call_function_ipi(const cpumask_t *mask) { int cpu; xen_send_IPI_mask(mask, XEN_CALL_FUNCTION_VECTOR); /* Make sure other vcpus get a chance to run if they need to. */ - for_each_cpu_mask_nr(cpu, mask) { + for_each_cpu_mask_nr(cpu, *mask) { if (xen_vcpu_stolen(cpu)) { HYPERVISOR_sched_op(SCHEDOP_yield, 0); break; @@ -435,7 +433,8 @@ static void xen_smp_send_call_function_ipi(cpumask_t mask) static void xen_smp_send_call_function_single_ipi(int cpu) { - xen_send_IPI_mask(cpumask_of_cpu(cpu), XEN_CALL_FUNCTION_SINGLE_VECTOR); + xen_send_IPI_mask(&cpumask_of_cpu(cpu), + XEN_CALL_FUNCTION_SINGLE_VECTOR); } static irqreturn_t xen_call_function_interrupt(int irq, void *dev_id) -- cgit v1.2.3 From a1681965011916c2f1f0f1f87e70784f5d5d5be5 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Tue, 16 Dec 2008 17:33:53 -0800 Subject: x86: move and enhance debug printk for nr_cpu_ids etc. Impact: cleanup, better debugging This has proven useful in debugging, *before* we try to use for_each_possible_cpu(). It also now shows nr_cpumask_bits. Signed-off-by: Mike Travis Signed-off-by: Rusty Russell --- arch/x86/kernel/setup_percpu.c | 17 +++++++++++------ 1 file changed, 11 insertions(+), 6 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c index 1c2084291f97..0b63b08e7530 100644 --- a/arch/x86/kernel/setup_percpu.c +++ b/arch/x86/kernel/setup_percpu.c @@ -152,6 +152,11 @@ void __init setup_per_cpu_areas(void) old_size = PERCPU_ENOUGH_ROOM; align = max_t(unsigned long, PAGE_SIZE, align); size = roundup(old_size, align); + + printk(KERN_INFO + "NR_CPUS:%d nr_cpumask_bits:%d nr_cpu_ids:%d nr_node_ids:%d\n", + NR_CPUS, nr_cpumask_bits, nr_cpu_ids, nr_node_ids); + printk(KERN_INFO "PERCPU: Allocating %zd bytes of per cpu data\n", size); @@ -168,24 +173,24 @@ void __init setup_per_cpu_areas(void) "cpu %d has no node %d or node-local memory\n", cpu, node); if (ptr) - printk(KERN_DEBUG "per cpu data for cpu%d at %016lx\n", + printk(KERN_DEBUG + "per cpu data for cpu%d at %016lx\n", cpu, __pa(ptr)); } else { ptr = __alloc_bootmem_node(NODE_DATA(node), size, align, __pa(MAX_DMA_ADDRESS)); if (ptr) - printk(KERN_DEBUG "per cpu data for cpu%d on node%d at %016lx\n", - cpu, node, __pa(ptr)); + printk(KERN_DEBUG + "per cpu data for cpu%d on node%d " + "at %016lx\n", + cpu, node, __pa(ptr)); } #endif per_cpu_offset(cpu) = ptr - __per_cpu_start; memcpy(ptr, __per_cpu_start, __per_cpu_end - __per_cpu_start); } - printk(KERN_DEBUG "NR_CPUS: %d, nr_cpu_ids: %d, nr_node_ids %d\n", - NR_CPUS, nr_cpu_ids, nr_node_ids); - /* Setup percpu data maps */ setup_per_cpu_maps(); -- cgit v1.2.3 From 95d313cf1c1ecedc8bec5727b09bdacbf67dfc45 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Tue, 16 Dec 2008 17:33:54 -0800 Subject: x86: Add cpu_mask_to_apicid_and Impact: new API Add a helper function that takes two cpumask's, and's them and then returns the apicid of the result. This removes a need in io_apic.c that uses a temporary cpumask to hold (mask & cfg->domain). Signed-off-by: Mike Travis Signed-off-by: Rusty Russell --- arch/x86/include/asm/bigsmp/apic.h | 16 +++++++++ arch/x86/include/asm/es7000/apic.h | 47 +++++++++++++++++++++++++++ arch/x86/include/asm/genapic_32.h | 3 ++ arch/x86/include/asm/genapic_64.h | 2 ++ arch/x86/include/asm/mach-default/mach_apic.h | 10 ++++++ arch/x86/include/asm/mach-generic/mach_apic.h | 1 + arch/x86/include/asm/numaq/apic.h | 6 ++++ arch/x86/include/asm/summit/apic.h | 39 ++++++++++++++++++++++ arch/x86/kernel/genapic_flat_64.c | 26 +++++++++++++++ arch/x86/kernel/genx2apic_cluster.c | 16 +++++++++ arch/x86/kernel/genx2apic_phys.c | 16 +++++++++ arch/x86/kernel/genx2apic_uv_x.c | 16 +++++++++ 12 files changed, 198 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/bigsmp/apic.h b/arch/x86/include/asm/bigsmp/apic.h index dc6225ca48ad..99f9abacf6a2 100644 --- a/arch/x86/include/asm/bigsmp/apic.h +++ b/arch/x86/include/asm/bigsmp/apic.h @@ -129,6 +129,22 @@ static inline unsigned int cpu_mask_to_apicid(const cpumask_t *cpumask) return apicid; } +static inline unsigned int cpu_mask_to_apicid_and(const cpumask_t *cpumask, + const cpumask_t *andmask) +{ + int cpu; + + /* + * We're using fixed IRQ delivery, can only return one phys APIC ID. + * May as well be the first. + */ + while ((cpu = next_cpu(-1, *cpumask)) < nr_cpu_ids) + if (cpu_isset(cpu, *andmask)) + return cpu_to_logical_apicid(cpu); + + return BAD_APICID; +} + static inline u32 phys_pkg_id(u32 cpuid_apic, int index_msb) { return cpuid_apic >> index_msb; diff --git a/arch/x86/include/asm/es7000/apic.h b/arch/x86/include/asm/es7000/apic.h index 4cac0837bb40..c2bed772af88 100644 --- a/arch/x86/include/asm/es7000/apic.h +++ b/arch/x86/include/asm/es7000/apic.h @@ -214,6 +214,53 @@ static inline unsigned int cpu_mask_to_apicid(const cpumask_t *cpumask) return apicid; } +static inline unsigned int cpu_mask_to_apicid_and(const cpumask_t *cpumask, + const cpumask_t *andmask) +{ + int num_bits_set; + int num_bits_set2; + int cpus_found = 0; + int cpu; + int apicid = 0; + + num_bits_set = cpus_weight(*cpumask); + num_bits_set2 = cpus_weight(*andmask); + num_bits_set = min_t(int, num_bits_set, num_bits_set2); + /* Return id to all */ + if (num_bits_set >= nr_cpu_ids) +#if defined CONFIG_ES7000_CLUSTERED_APIC + return 0xFF; +#else + return cpu_to_logical_apicid(0); +#endif + /* + * The cpus in the mask must all be on the apic cluster. If are not + * on the same apicid cluster return default value of TARGET_CPUS. + */ + while ((cpu = next_cpu(-1, *cpumask)) < nr_cpu_ids) + if (cpu_isset(cpu, *andmask) + apicid = cpu_to_logical_apicid(cpu); + while (cpus_found < num_bits_set) { + if (cpu_isset(cpu, *cpumask) && cpu_isset(cpu, *andmask)) { + int new_apicid = cpu_to_logical_apicid(cpu); + if (apicid_cluster(apicid) != + apicid_cluster(new_apicid)) { + printk(KERN_WARNING + "%s: Not a valid mask!\n", __func__); +#if defined CONFIG_ES7000_CLUSTERED_APIC + return 0xFF; +#else + return cpu_to_logical_apicid(0); +#endif + } + apicid = new_apicid; + cpus_found++; + } + cpu++; + } + return apicid; +} + static inline u32 phys_pkg_id(u32 cpuid_apic, int index_msb) { return cpuid_apic >> index_msb; diff --git a/arch/x86/include/asm/genapic_32.h b/arch/x86/include/asm/genapic_32.h index b21ed21c574d..325298a82237 100644 --- a/arch/x86/include/asm/genapic_32.h +++ b/arch/x86/include/asm/genapic_32.h @@ -58,6 +58,8 @@ struct genapic { unsigned (*get_apic_id)(unsigned long x); unsigned long apic_id_mask; unsigned int (*cpu_mask_to_apicid)(const cpumask_t *cpumask); + unsigned int (*cpu_mask_to_apicid_and)(const cpumask_t *cpumask, + const cpumask_t *andmask); void (*vector_allocation_domain)(int cpu, cpumask_t *retmask); #ifdef CONFIG_SMP @@ -115,6 +117,7 @@ struct genapic { APICFUNC(get_apic_id) \ .apic_id_mask = APIC_ID_MASK, \ APICFUNC(cpu_mask_to_apicid) \ + APICFUNC(cpu_mask_to_apicid_and) \ APICFUNC(vector_allocation_domain) \ APICFUNC(acpi_madt_oem_check) \ IPIFUNC(send_IPI_mask) \ diff --git a/arch/x86/include/asm/genapic_64.h b/arch/x86/include/asm/genapic_64.h index a020e7d35a40..301c7f41125d 100644 --- a/arch/x86/include/asm/genapic_64.h +++ b/arch/x86/include/asm/genapic_64.h @@ -31,6 +31,8 @@ struct genapic { void (*send_IPI_self)(int vector); /* */ unsigned int (*cpu_mask_to_apicid)(const cpumask_t *cpumask); + unsigned int (*cpu_mask_to_apicid_and)(const cpumask_t *cpumask, + const cpumask_t *andmask); unsigned int (*phys_pkg_id)(int index_msb); unsigned int (*get_apic_id)(unsigned long x); unsigned long (*set_apic_id)(unsigned int id); diff --git a/arch/x86/include/asm/mach-default/mach_apic.h b/arch/x86/include/asm/mach-default/mach_apic.h index c18896b0508c..229b605d104e 100644 --- a/arch/x86/include/asm/mach-default/mach_apic.h +++ b/arch/x86/include/asm/mach-default/mach_apic.h @@ -28,6 +28,7 @@ static inline const cpumask_t *target_cpus(void) #define apic_id_registered (genapic->apic_id_registered) #define init_apic_ldr (genapic->init_apic_ldr) #define cpu_mask_to_apicid (genapic->cpu_mask_to_apicid) +#define cpu_mask_to_apicid_and (genapic->cpu_mask_to_apicid_and) #define phys_pkg_id (genapic->phys_pkg_id) #define vector_allocation_domain (genapic->vector_allocation_domain) #define read_apic_id() (GET_APIC_ID(apic_read(APIC_ID))) @@ -66,6 +67,15 @@ static inline unsigned int cpu_mask_to_apicid(const cpumask_t *cpumask) return cpus_addr(*cpumask)[0]; } +static inline unsigned int cpu_mask_to_apicid(const cpumask_t *cpumask, + const cpumask_t *andmask) +{ + unsigned long mask1 = cpus_addr(*cpumask)[0]; + unsigned long mask2 = cpus_addr(*andmask)[0]; + + return (unsigned int)(mask1 & mask2); +} + static inline u32 phys_pkg_id(u32 cpuid_apic, int index_msb) { return cpuid_apic >> index_msb; diff --git a/arch/x86/include/asm/mach-generic/mach_apic.h b/arch/x86/include/asm/mach-generic/mach_apic.h index e430f47df667..48553e958ad5 100644 --- a/arch/x86/include/asm/mach-generic/mach_apic.h +++ b/arch/x86/include/asm/mach-generic/mach_apic.h @@ -24,6 +24,7 @@ #define check_phys_apicid_present (genapic->check_phys_apicid_present) #define check_apicid_used (genapic->check_apicid_used) #define cpu_mask_to_apicid (genapic->cpu_mask_to_apicid) +#define cpu_mask_to_apicid_and (genapic->cpu_mask_to_apicid_and) #define vector_allocation_domain (genapic->vector_allocation_domain) #define enable_apic_mode (genapic->enable_apic_mode) #define phys_pkg_id (genapic->phys_pkg_id) diff --git a/arch/x86/include/asm/numaq/apic.h b/arch/x86/include/asm/numaq/apic.h index 1df7ebe738e5..abf668ced503 100644 --- a/arch/x86/include/asm/numaq/apic.h +++ b/arch/x86/include/asm/numaq/apic.h @@ -127,6 +127,12 @@ static inline unsigned int cpu_mask_to_apicid(const cpumask_t *cpumask) return (int) 0xF; } +static inline unsigned int cpu_mask_to_apicid_and(const cpumask_t *cpumask, + const cpumask_t *andmask) +{ + return (int) 0xF; +} + /* No NUMA-Q box has a HT CPU, but it can't hurt to use the default code. */ static inline u32 phys_pkg_id(u32 cpuid_apic, int index_msb) { diff --git a/arch/x86/include/asm/summit/apic.h b/arch/x86/include/asm/summit/apic.h index 437dc83725ca..cbcc2c7eb1d5 100644 --- a/arch/x86/include/asm/summit/apic.h +++ b/arch/x86/include/asm/summit/apic.h @@ -170,6 +170,45 @@ static inline unsigned int cpu_mask_to_apicid(const cpumask_t *cpumask) return apicid; } +static inline unsigned int cpu_mask_to_apicid_and(const cpumask_t *cpumask, + const cpumask_t *andmask) +{ + int num_bits_set; + int num_bits_set2; + int cpus_found = 0; + int cpu; + int apicid = 0; + + num_bits_set = cpus_weight(*cpumask); + num_bits_set2 = cpus_weight(*andmask); + num_bits_set = min_t(int, num_bits_set, num_bits_set2); + /* Return id to all */ + if (num_bits_set >= nr_cpu_ids) + return 0xFF; + /* + * The cpus in the mask must all be on the apic cluster. If are not + * on the same apicid cluster return default value of TARGET_CPUS. + */ + while ((cpu = next_cpu(-1, *cpumask)) < nr_cpu_ids) + if (cpu_isset(cpu, *andmask) + apicid = cpu_to_logical_apicid(cpu); + while (cpus_found < num_bits_set) { + if (cpu_isset(cpu, *cpumask) && cpu_isset(cpu, *andmask)) { + int new_apicid = cpu_to_logical_apicid(cpu); + if (apicid_cluster(apicid) != + apicid_cluster(new_apicid)) { + printk(KERN_WARNING + "%s: Not a valid mask!\n", __func__); + return 0xFF; + } + apicid = apicid | new_apicid; + cpus_found++; + } + cpu++; + } + return apicid; +} + /* cpuid returns the value latched in the HW at reset, not the APIC ID * register's value. For any box whose BIOS changes APIC IDs, like * clustered APIC systems, we must use hard_smp_processor_id. diff --git a/arch/x86/kernel/genapic_flat_64.c b/arch/x86/kernel/genapic_flat_64.c index 50eebd0328fe..1efecd206a74 100644 --- a/arch/x86/kernel/genapic_flat_64.c +++ b/arch/x86/kernel/genapic_flat_64.c @@ -158,6 +158,15 @@ static unsigned int flat_cpu_mask_to_apicid(const cpumask_t *cpumask) return cpus_addr(*cpumask)[0] & APIC_ALL_CPUS; } +static unsigned int flat_cpu_mask_to_apicid_and(const cpumask_t *cpumask, + const cpumask_t *andmask) +{ + unsigned long mask1 = cpus_addr(*cpumask)[0] & APIC_ALL_CPUS; + unsigned long mask2 = cpus_addr(*andmask)[0] & APIC_ALL_CPUS; + + return (int)(mask1 & mask2); +} + static unsigned int phys_pkg_id(int index_msb) { return hard_smp_processor_id() >> index_msb; @@ -178,6 +187,7 @@ struct genapic apic_flat = { .send_IPI_mask_allbutself = flat_send_IPI_mask_allbutself, .send_IPI_self = apic_send_IPI_self, .cpu_mask_to_apicid = flat_cpu_mask_to_apicid, + .cpu_mask_to_apicid_and = flat_cpu_mask_to_apicid_and, .phys_pkg_id = phys_pkg_id, .get_apic_id = get_apic_id, .set_apic_id = set_apic_id, @@ -254,6 +264,21 @@ static unsigned int physflat_cpu_mask_to_apicid(const cpumask_t *cpumask) return BAD_APICID; } +static unsigned int physflat_cpu_mask_to_apicid_and(const cpumask_t *cpumask, + const cpumask_t *andmask) +{ + int cpu; + + /* + * We're using fixed IRQ delivery, can only return one phys APIC ID. + * May as well be the first. + */ + while ((cpu = next_cpu(-1, *cpumask)) < nr_cpu_ids) + if (cpu_isset(cpu, *andmask)) + return per_cpu(x86_cpu_to_apicid, cpu); + return BAD_APICID; +} + struct genapic apic_physflat = { .name = "physical flat", .acpi_madt_oem_check = physflat_acpi_madt_oem_check, @@ -269,6 +294,7 @@ struct genapic apic_physflat = { .send_IPI_mask_allbutself = physflat_send_IPI_mask_allbutself, .send_IPI_self = apic_send_IPI_self, .cpu_mask_to_apicid = physflat_cpu_mask_to_apicid, + .cpu_mask_to_apicid_and = physflat_cpu_mask_to_apicid_and, .phys_pkg_id = phys_pkg_id, .get_apic_id = get_apic_id, .set_apic_id = set_apic_id, diff --git a/arch/x86/kernel/genx2apic_cluster.c b/arch/x86/kernel/genx2apic_cluster.c index f5fa9a91ad38..fd8047f4e455 100644 --- a/arch/x86/kernel/genx2apic_cluster.c +++ b/arch/x86/kernel/genx2apic_cluster.c @@ -123,6 +123,21 @@ static unsigned int x2apic_cpu_mask_to_apicid(const cpumask_t *cpumask) return BAD_APICID; } +static unsigned int x2apic_cpu_mask_to_apicid_and(const cpumask_t *cpumask, + const cpumask_t *andmask) +{ + int cpu; + + /* + * We're using fixed IRQ delivery, can only return one phys APIC ID. + * May as well be the first. + */ + while ((cpu = next_cpu(-1, *cpumask)) < nr_cpu_ids) + if (cpu_isset(cpu, *andmask)) + return per_cpu(x86_cpu_to_apicid, cpu); + return BAD_APICID; +} + static unsigned int get_apic_id(unsigned long x) { unsigned int id; @@ -172,6 +187,7 @@ struct genapic apic_x2apic_cluster = { .send_IPI_mask_allbutself = x2apic_send_IPI_mask_allbutself, .send_IPI_self = x2apic_send_IPI_self, .cpu_mask_to_apicid = x2apic_cpu_mask_to_apicid, + .cpu_mask_to_apicid_and = x2apic_cpu_mask_to_apicid_and, .phys_pkg_id = phys_pkg_id, .get_apic_id = get_apic_id, .set_apic_id = set_apic_id, diff --git a/arch/x86/kernel/genx2apic_phys.c b/arch/x86/kernel/genx2apic_phys.c index 41c27b2f3d01..d5578bb8f165 100644 --- a/arch/x86/kernel/genx2apic_phys.c +++ b/arch/x86/kernel/genx2apic_phys.c @@ -122,6 +122,21 @@ static unsigned int x2apic_cpu_mask_to_apicid(const cpumask_t *cpumask) return BAD_APICID; } +static unsigned int x2apic_cpu_mask_to_apicid_and(const cpumask_t *cpumask, + const cpumask_t *andmask) +{ + int cpu; + + /* + * We're using fixed IRQ delivery, can only return one phys APIC ID. + * May as well be the first. + */ + while ((cpu = next_cpu(-1, *cpumask)) < nr_cpu_ids) + if (cpu_isset(cpu, *andmask)) + return per_cpu(x86_cpu_to_apicid, cpu); + return BAD_APICID; +} + static unsigned int get_apic_id(unsigned long x) { unsigned int id; @@ -168,6 +183,7 @@ struct genapic apic_x2apic_phys = { .send_IPI_mask_allbutself = x2apic_send_IPI_mask_allbutself, .send_IPI_self = x2apic_send_IPI_self, .cpu_mask_to_apicid = x2apic_cpu_mask_to_apicid, + .cpu_mask_to_apicid_and = x2apic_cpu_mask_to_apicid_and, .phys_pkg_id = phys_pkg_id, .get_apic_id = get_apic_id, .set_apic_id = set_apic_id, diff --git a/arch/x86/kernel/genx2apic_uv_x.c b/arch/x86/kernel/genx2apic_uv_x.c index 010659415ae4..53bd2570272c 100644 --- a/arch/x86/kernel/genx2apic_uv_x.c +++ b/arch/x86/kernel/genx2apic_uv_x.c @@ -179,6 +179,21 @@ static unsigned int uv_cpu_mask_to_apicid(const cpumask_t *cpumask) return BAD_APICID; } +static unsigned int uv_cpu_mask_to_apicid_and(const cpumask_t *cpumask, + const cpumask_t *andmask) +{ + int cpu; + + /* + * We're using fixed IRQ delivery, can only return one phys APIC ID. + * May as well be the first. + */ + while ((cpu = next_cpu(-1, *cpumask)) < nr_cpu_ids) + if (cpu_isset(cpu, *andmask)) + return per_cpu(x86_cpu_to_apicid, cpu); + return BAD_APICID; +} + static unsigned int get_apic_id(unsigned long x) { unsigned int id; @@ -229,6 +244,7 @@ struct genapic apic_x2apic_uv_x = { .send_IPI_mask_allbutself = uv_send_IPI_mask_allbutself, .send_IPI_self = uv_send_IPI_self, .cpu_mask_to_apicid = uv_cpu_mask_to_apicid, + .cpu_mask_to_apicid_and = uv_cpu_mask_to_apicid_and, .phys_pkg_id = phys_pkg_id, .get_apic_id = get_apic_id, .set_apic_id = set_apic_id, -- cgit v1.2.3 From 6eeb7c5a99434596c5953a95baa17d2f085664e3 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Tue, 16 Dec 2008 17:33:55 -0800 Subject: x86: update add-cpu_mask_to_apicid_and to use struct cpumask* Impact: use updated APIs Various API updates for x86:add-cpu_mask_to_apicid_and (Note: separate because previous patch has been "backported" to 2.6.27.) Signed-off-by: Rusty Russell Signed-off-by: Mike Travis --- arch/x86/include/asm/bigsmp/apic.h | 10 +++++----- arch/x86/include/asm/es7000/apic.h | 19 ++++++++++--------- arch/x86/include/asm/genapic_32.h | 4 ++-- arch/x86/include/asm/genapic_64.h | 4 ++-- arch/x86/include/asm/mach-default/mach_apic.h | 8 ++++---- arch/x86/include/asm/numaq/apic.h | 4 ++-- arch/x86/include/asm/summit/apic.h | 18 +++++++++--------- arch/x86/kernel/genapic_flat_64.c | 21 +++++++++++---------- arch/x86/kernel/genx2apic_cluster.c | 10 +++++----- arch/x86/kernel/genx2apic_phys.c | 10 +++++----- arch/x86/kernel/genx2apic_uv_x.c | 10 +++++----- 11 files changed, 60 insertions(+), 58 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/bigsmp/apic.h b/arch/x86/include/asm/bigsmp/apic.h index 99f9abacf6a2..976399debb3f 100644 --- a/arch/x86/include/asm/bigsmp/apic.h +++ b/arch/x86/include/asm/bigsmp/apic.h @@ -129,8 +129,8 @@ static inline unsigned int cpu_mask_to_apicid(const cpumask_t *cpumask) return apicid; } -static inline unsigned int cpu_mask_to_apicid_and(const cpumask_t *cpumask, - const cpumask_t *andmask) +static inline unsigned int cpu_mask_to_apicid_and(const struct cpumask *cpumask, + const struct cpumask *andmask) { int cpu; @@ -138,9 +138,9 @@ static inline unsigned int cpu_mask_to_apicid_and(const cpumask_t *cpumask, * We're using fixed IRQ delivery, can only return one phys APIC ID. * May as well be the first. */ - while ((cpu = next_cpu(-1, *cpumask)) < nr_cpu_ids) - if (cpu_isset(cpu, *andmask)) - return cpu_to_logical_apicid(cpu); + cpu = cpumask_any_and(cpumask, andmask); + if (cpu < nr_cpu_ids) + return cpu_to_logical_apicid(cpu); return BAD_APICID; } diff --git a/arch/x86/include/asm/es7000/apic.h b/arch/x86/include/asm/es7000/apic.h index c2bed772af88..ba8423c5363f 100644 --- a/arch/x86/include/asm/es7000/apic.h +++ b/arch/x86/include/asm/es7000/apic.h @@ -214,8 +214,8 @@ static inline unsigned int cpu_mask_to_apicid(const cpumask_t *cpumask) return apicid; } -static inline unsigned int cpu_mask_to_apicid_and(const cpumask_t *cpumask, - const cpumask_t *andmask) +static inline unsigned int cpu_mask_to_apicid_and(const struct cpumask *cpumask, + const struct cpumask *andmask) { int num_bits_set; int num_bits_set2; @@ -223,9 +223,9 @@ static inline unsigned int cpu_mask_to_apicid_and(const cpumask_t *cpumask, int cpu; int apicid = 0; - num_bits_set = cpus_weight(*cpumask); - num_bits_set2 = cpus_weight(*andmask); - num_bits_set = min_t(int, num_bits_set, num_bits_set2); + num_bits_set = cpumask_weight(cpumask); + num_bits_set2 = cpumask_weight(andmask); + num_bits_set = min(num_bits_set, num_bits_set2); /* Return id to all */ if (num_bits_set >= nr_cpu_ids) #if defined CONFIG_ES7000_CLUSTERED_APIC @@ -237,11 +237,12 @@ static inline unsigned int cpu_mask_to_apicid_and(const cpumask_t *cpumask, * The cpus in the mask must all be on the apic cluster. If are not * on the same apicid cluster return default value of TARGET_CPUS. */ - while ((cpu = next_cpu(-1, *cpumask)) < nr_cpu_ids) - if (cpu_isset(cpu, *andmask) - apicid = cpu_to_logical_apicid(cpu); + cpu = cpumask_first_and(cpumask, andmask); + apicid = cpu_to_logical_apicid(cpu); + while (cpus_found < num_bits_set) { - if (cpu_isset(cpu, *cpumask) && cpu_isset(cpu, *andmask)) { + if (cpumask_test_cpu(cpu, cpumask) && + cpumask_test_cpu(cpu, andmask)) { int new_apicid = cpu_to_logical_apicid(cpu); if (apicid_cluster(apicid) != apicid_cluster(new_apicid)) { diff --git a/arch/x86/include/asm/genapic_32.h b/arch/x86/include/asm/genapic_32.h index 325298a82237..eed6e305291f 100644 --- a/arch/x86/include/asm/genapic_32.h +++ b/arch/x86/include/asm/genapic_32.h @@ -58,8 +58,8 @@ struct genapic { unsigned (*get_apic_id)(unsigned long x); unsigned long apic_id_mask; unsigned int (*cpu_mask_to_apicid)(const cpumask_t *cpumask); - unsigned int (*cpu_mask_to_apicid_and)(const cpumask_t *cpumask, - const cpumask_t *andmask); + unsigned int (*cpu_mask_to_apicid_and)(const struct cpumask *cpumask, + const struct cpumask *andmask); void (*vector_allocation_domain)(int cpu, cpumask_t *retmask); #ifdef CONFIG_SMP diff --git a/arch/x86/include/asm/genapic_64.h b/arch/x86/include/asm/genapic_64.h index 301c7f41125d..244b71729ecb 100644 --- a/arch/x86/include/asm/genapic_64.h +++ b/arch/x86/include/asm/genapic_64.h @@ -31,8 +31,8 @@ struct genapic { void (*send_IPI_self)(int vector); /* */ unsigned int (*cpu_mask_to_apicid)(const cpumask_t *cpumask); - unsigned int (*cpu_mask_to_apicid_and)(const cpumask_t *cpumask, - const cpumask_t *andmask); + unsigned int (*cpu_mask_to_apicid_and)(const struct cpumask *cpumask, + const struct cpumask *andmask); unsigned int (*phys_pkg_id)(int index_msb); unsigned int (*get_apic_id)(unsigned long x); unsigned long (*set_apic_id)(unsigned int id); diff --git a/arch/x86/include/asm/mach-default/mach_apic.h b/arch/x86/include/asm/mach-default/mach_apic.h index 229b605d104e..df8e024c43c5 100644 --- a/arch/x86/include/asm/mach-default/mach_apic.h +++ b/arch/x86/include/asm/mach-default/mach_apic.h @@ -67,11 +67,11 @@ static inline unsigned int cpu_mask_to_apicid(const cpumask_t *cpumask) return cpus_addr(*cpumask)[0]; } -static inline unsigned int cpu_mask_to_apicid(const cpumask_t *cpumask, - const cpumask_t *andmask) +static inline unsigned int cpu_mask_to_apicid_and(const struct cpumask *cpumask, + const struct cpumask *andmask) { - unsigned long mask1 = cpus_addr(*cpumask)[0]; - unsigned long mask2 = cpus_addr(*andmask)[0]; + unsigned long mask1 = cpumask_bits(cpumask)[0]; + unsigned long mask2 = cpumask_bits(andmask)[0]; return (unsigned int)(mask1 & mask2); } diff --git a/arch/x86/include/asm/numaq/apic.h b/arch/x86/include/asm/numaq/apic.h index abf668ced503..c80f00d29965 100644 --- a/arch/x86/include/asm/numaq/apic.h +++ b/arch/x86/include/asm/numaq/apic.h @@ -127,8 +127,8 @@ static inline unsigned int cpu_mask_to_apicid(const cpumask_t *cpumask) return (int) 0xF; } -static inline unsigned int cpu_mask_to_apicid_and(const cpumask_t *cpumask, - const cpumask_t *andmask) +static inline unsigned int cpu_mask_to_apicid_and(const struct cpumask *cpumask, + const struct cpumask *andmask) { return (int) 0xF; } diff --git a/arch/x86/include/asm/summit/apic.h b/arch/x86/include/asm/summit/apic.h index cbcc2c7eb1d5..651a93849341 100644 --- a/arch/x86/include/asm/summit/apic.h +++ b/arch/x86/include/asm/summit/apic.h @@ -170,8 +170,8 @@ static inline unsigned int cpu_mask_to_apicid(const cpumask_t *cpumask) return apicid; } -static inline unsigned int cpu_mask_to_apicid_and(const cpumask_t *cpumask, - const cpumask_t *andmask) +static inline unsigned int cpu_mask_to_apicid_and(const struct cpumask *cpumask, + const struct cpumask *andmask) { int num_bits_set; int num_bits_set2; @@ -179,9 +179,9 @@ static inline unsigned int cpu_mask_to_apicid_and(const cpumask_t *cpumask, int cpu; int apicid = 0; - num_bits_set = cpus_weight(*cpumask); - num_bits_set2 = cpus_weight(*andmask); - num_bits_set = min_t(int, num_bits_set, num_bits_set2); + num_bits_set = cpumask_weight(cpumask); + num_bits_set2 = cpumask_weight(andmask); + num_bits_set = min(num_bits_set, num_bits_set2); /* Return id to all */ if (num_bits_set >= nr_cpu_ids) return 0xFF; @@ -189,11 +189,11 @@ static inline unsigned int cpu_mask_to_apicid_and(const cpumask_t *cpumask, * The cpus in the mask must all be on the apic cluster. If are not * on the same apicid cluster return default value of TARGET_CPUS. */ - while ((cpu = next_cpu(-1, *cpumask)) < nr_cpu_ids) - if (cpu_isset(cpu, *andmask) - apicid = cpu_to_logical_apicid(cpu); + cpu = cpumask_first_and(cpumask, andmask); + apicid = cpu_to_logical_apicid(cpu); while (cpus_found < num_bits_set) { - if (cpu_isset(cpu, *cpumask) && cpu_isset(cpu, *andmask)) { + if (cpumask_test_cpu(cpu, cpumask) + && cpumask_test_cpu(cpu, andmask)) { int new_apicid = cpu_to_logical_apicid(cpu); if (apicid_cluster(apicid) != apicid_cluster(new_apicid)) { diff --git a/arch/x86/kernel/genapic_flat_64.c b/arch/x86/kernel/genapic_flat_64.c index 1efecd206a74..c772bb10b173 100644 --- a/arch/x86/kernel/genapic_flat_64.c +++ b/arch/x86/kernel/genapic_flat_64.c @@ -158,13 +158,13 @@ static unsigned int flat_cpu_mask_to_apicid(const cpumask_t *cpumask) return cpus_addr(*cpumask)[0] & APIC_ALL_CPUS; } -static unsigned int flat_cpu_mask_to_apicid_and(const cpumask_t *cpumask, - const cpumask_t *andmask) +static unsigned int flat_cpu_mask_to_apicid_and(const struct cpumask *cpumask, + const struct cpumask *andmask) { - unsigned long mask1 = cpus_addr(*cpumask)[0] & APIC_ALL_CPUS; - unsigned long mask2 = cpus_addr(*andmask)[0] & APIC_ALL_CPUS; + unsigned long mask1 = cpumask_bits(cpumask)[0] & APIC_ALL_CPUS; + unsigned long mask2 = cpumask_bits(andmask)[0] & APIC_ALL_CPUS; - return (int)(mask1 & mask2); + return mask1 & mask2; } static unsigned int phys_pkg_id(int index_msb) @@ -264,8 +264,9 @@ static unsigned int physflat_cpu_mask_to_apicid(const cpumask_t *cpumask) return BAD_APICID; } -static unsigned int physflat_cpu_mask_to_apicid_and(const cpumask_t *cpumask, - const cpumask_t *andmask) +static unsigned int +physflat_cpu_mask_to_apicid_and(const struct cpumask *cpumask, + const struct cpumask *andmask) { int cpu; @@ -273,9 +274,9 @@ static unsigned int physflat_cpu_mask_to_apicid_and(const cpumask_t *cpumask, * We're using fixed IRQ delivery, can only return one phys APIC ID. * May as well be the first. */ - while ((cpu = next_cpu(-1, *cpumask)) < nr_cpu_ids) - if (cpu_isset(cpu, *andmask)) - return per_cpu(x86_cpu_to_apicid, cpu); + cpu = cpumask_any_and(cpumask, andmask); + if (cpu < nr_cpu_ids) + return per_cpu(x86_cpu_to_apicid, cpu); return BAD_APICID; } diff --git a/arch/x86/kernel/genx2apic_cluster.c b/arch/x86/kernel/genx2apic_cluster.c index fd8047f4e455..e7d16f53b9cd 100644 --- a/arch/x86/kernel/genx2apic_cluster.c +++ b/arch/x86/kernel/genx2apic_cluster.c @@ -123,8 +123,8 @@ static unsigned int x2apic_cpu_mask_to_apicid(const cpumask_t *cpumask) return BAD_APICID; } -static unsigned int x2apic_cpu_mask_to_apicid_and(const cpumask_t *cpumask, - const cpumask_t *andmask) +static unsigned int x2apic_cpu_mask_to_apicid_and(const struct cpumask *cpumask, + const struct cpumask *andmask) { int cpu; @@ -132,9 +132,9 @@ static unsigned int x2apic_cpu_mask_to_apicid_and(const cpumask_t *cpumask, * We're using fixed IRQ delivery, can only return one phys APIC ID. * May as well be the first. */ - while ((cpu = next_cpu(-1, *cpumask)) < nr_cpu_ids) - if (cpu_isset(cpu, *andmask)) - return per_cpu(x86_cpu_to_apicid, cpu); + cpu = cpumask_any_and(cpumask, andmask); + if (cpu < nr_cpu_ids) + return per_cpu(x86_cpu_to_apicid, cpu); return BAD_APICID; } diff --git a/arch/x86/kernel/genx2apic_phys.c b/arch/x86/kernel/genx2apic_phys.c index d5578bb8f165..9d0386c7e798 100644 --- a/arch/x86/kernel/genx2apic_phys.c +++ b/arch/x86/kernel/genx2apic_phys.c @@ -122,8 +122,8 @@ static unsigned int x2apic_cpu_mask_to_apicid(const cpumask_t *cpumask) return BAD_APICID; } -static unsigned int x2apic_cpu_mask_to_apicid_and(const cpumask_t *cpumask, - const cpumask_t *andmask) +static unsigned int x2apic_cpu_mask_to_apicid_and(const struct cpumask *cpumask, + const struct cpumask *andmask) { int cpu; @@ -131,9 +131,9 @@ static unsigned int x2apic_cpu_mask_to_apicid_and(const cpumask_t *cpumask, * We're using fixed IRQ delivery, can only return one phys APIC ID. * May as well be the first. */ - while ((cpu = next_cpu(-1, *cpumask)) < nr_cpu_ids) - if (cpu_isset(cpu, *andmask)) - return per_cpu(x86_cpu_to_apicid, cpu); + cpu = cpumask_any_and(cpumask, andmask); + if (cpu < nr_cpu_ids) + return per_cpu(x86_cpu_to_apicid, cpu); return BAD_APICID; } diff --git a/arch/x86/kernel/genx2apic_uv_x.c b/arch/x86/kernel/genx2apic_uv_x.c index 53bd2570272c..22596ec94c82 100644 --- a/arch/x86/kernel/genx2apic_uv_x.c +++ b/arch/x86/kernel/genx2apic_uv_x.c @@ -179,8 +179,8 @@ static unsigned int uv_cpu_mask_to_apicid(const cpumask_t *cpumask) return BAD_APICID; } -static unsigned int uv_cpu_mask_to_apicid_and(const cpumask_t *cpumask, - const cpumask_t *andmask) +static unsigned int uv_cpu_mask_to_apicid_and(const struct cpumask *cpumask, + const struct cpumask *andmask) { int cpu; @@ -188,9 +188,9 @@ static unsigned int uv_cpu_mask_to_apicid_and(const cpumask_t *cpumask, * We're using fixed IRQ delivery, can only return one phys APIC ID. * May as well be the first. */ - while ((cpu = next_cpu(-1, *cpumask)) < nr_cpu_ids) - if (cpu_isset(cpu, *andmask)) - return per_cpu(x86_cpu_to_apicid, cpu); + cpu = cpumask_any_and(cpumask, andmask); + if (cpu < nr_cpu_ids) + return per_cpu(x86_cpu_to_apicid, cpu); return BAD_APICID; } -- cgit v1.2.3 From 22f65d31b25a320a5246592160bcb102d2791c45 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Tue, 16 Dec 2008 17:33:56 -0800 Subject: x86: Update io_apic.c to use new cpumask API Impact: cleanup, consolidate patches, use new API Consolidate the following into a single patch to adapt to new sparseirq code in arch/x86/kernel/io_apic.c, add allocation of cpumask_var_t's in domain and old_domain, and reduce further merge conflicts. Only one file (arch/x86/kernel/io_apic.c) is changed in all of these patches. 0006-x86-io_apic-change-irq_cfg-domain-old_domain-to.patch 0007-x86-io_apic-set_desc_affinity.patch 0008-x86-io_apic-send_cleanup_vector.patch 0009-x86-io_apic-eliminate-remaining-cpumask_ts-from-st.patch 0021-x86-final-cleanups-in-io_apic-to-use-new-cpumask-AP.patch Signed-off-by: Rusty Russell Signed-off-by: Mike Travis --- arch/x86/kernel/io_apic.c | 302 ++++++++++++++++++++++------------------------ 1 file changed, 145 insertions(+), 157 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index 7f23ce7f5518..60bb8b19f4cd 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -136,8 +136,8 @@ static struct irq_pin_list *get_one_free_irq_2_pin(int cpu) struct irq_cfg { struct irq_pin_list *irq_2_pin; - cpumask_t domain; - cpumask_t old_domain; + cpumask_var_t domain; + cpumask_var_t old_domain; unsigned move_cleanup_count; u8 vector; u8 move_in_progress : 1; @@ -149,22 +149,22 @@ static struct irq_cfg irq_cfgx[] = { #else static struct irq_cfg irq_cfgx[NR_IRQS] = { #endif - [0] = { .domain = CPU_MASK_ALL, .vector = IRQ0_VECTOR, }, - [1] = { .domain = CPU_MASK_ALL, .vector = IRQ1_VECTOR, }, - [2] = { .domain = CPU_MASK_ALL, .vector = IRQ2_VECTOR, }, - [3] = { .domain = CPU_MASK_ALL, .vector = IRQ3_VECTOR, }, - [4] = { .domain = CPU_MASK_ALL, .vector = IRQ4_VECTOR, }, - [5] = { .domain = CPU_MASK_ALL, .vector = IRQ5_VECTOR, }, - [6] = { .domain = CPU_MASK_ALL, .vector = IRQ6_VECTOR, }, - [7] = { .domain = CPU_MASK_ALL, .vector = IRQ7_VECTOR, }, - [8] = { .domain = CPU_MASK_ALL, .vector = IRQ8_VECTOR, }, - [9] = { .domain = CPU_MASK_ALL, .vector = IRQ9_VECTOR, }, - [10] = { .domain = CPU_MASK_ALL, .vector = IRQ10_VECTOR, }, - [11] = { .domain = CPU_MASK_ALL, .vector = IRQ11_VECTOR, }, - [12] = { .domain = CPU_MASK_ALL, .vector = IRQ12_VECTOR, }, - [13] = { .domain = CPU_MASK_ALL, .vector = IRQ13_VECTOR, }, - [14] = { .domain = CPU_MASK_ALL, .vector = IRQ14_VECTOR, }, - [15] = { .domain = CPU_MASK_ALL, .vector = IRQ15_VECTOR, }, + [0] = { .vector = IRQ0_VECTOR, }, + [1] = { .vector = IRQ1_VECTOR, }, + [2] = { .vector = IRQ2_VECTOR, }, + [3] = { .vector = IRQ3_VECTOR, }, + [4] = { .vector = IRQ4_VECTOR, }, + [5] = { .vector = IRQ5_VECTOR, }, + [6] = { .vector = IRQ6_VECTOR, }, + [7] = { .vector = IRQ7_VECTOR, }, + [8] = { .vector = IRQ8_VECTOR, }, + [9] = { .vector = IRQ9_VECTOR, }, + [10] = { .vector = IRQ10_VECTOR, }, + [11] = { .vector = IRQ11_VECTOR, }, + [12] = { .vector = IRQ12_VECTOR, }, + [13] = { .vector = IRQ13_VECTOR, }, + [14] = { .vector = IRQ14_VECTOR, }, + [15] = { .vector = IRQ15_VECTOR, }, }; void __init arch_early_irq_init(void) @@ -180,6 +180,10 @@ void __init arch_early_irq_init(void) for (i = 0; i < count; i++) { desc = irq_to_desc(i); desc->chip_data = &cfg[i]; + alloc_bootmem_cpumask_var(&cfg[i].domain); + alloc_bootmem_cpumask_var(&cfg[i].old_domain); + if (i < NR_IRQS_LEGACY) + cpumask_setall(cfg[i].domain); } } @@ -204,6 +208,20 @@ static struct irq_cfg *get_one_free_irq_cfg(int cpu) node = cpu_to_node(cpu); cfg = kzalloc_node(sizeof(*cfg), GFP_ATOMIC, node); + if (cfg) { + /* FIXME: needs alloc_cpumask_var_node() */ + if (!alloc_cpumask_var(&cfg->domain, GFP_ATOMIC)) { + kfree(cfg); + cfg = NULL; + } else if (!alloc_cpumask_var(&cfg->old_domain, GFP_ATOMIC)) { + free_cpumask_var(cfg->domain); + kfree(cfg); + cfg = NULL; + } else { + cpumask_clear(cfg->domain); + cpumask_clear(cfg->old_domain); + } + } printk(KERN_DEBUG " alloc irq_cfg on cpu %d node %d\n", cpu, node); return cfg; @@ -362,6 +380,26 @@ static void ioapic_mask_entry(int apic, int pin) } #ifdef CONFIG_SMP +static void send_cleanup_vector(struct irq_cfg *cfg) +{ + cpumask_var_t cleanup_mask; + + if (unlikely(!alloc_cpumask_var(&cleanup_mask, GFP_ATOMIC))) { + unsigned int i; + cfg->move_cleanup_count = 0; + for_each_cpu_and(i, cfg->old_domain, cpu_online_mask) + cfg->move_cleanup_count++; + for_each_cpu_and(i, cfg->old_domain, cpu_online_mask) + send_IPI_mask(cpumask_of(i), IRQ_MOVE_CLEANUP_VECTOR); + } else { + cpumask_and(cleanup_mask, cfg->old_domain, cpu_online_mask); + cfg->move_cleanup_count = cpumask_weight(cleanup_mask); + send_IPI_mask(cleanup_mask, IRQ_MOVE_CLEANUP_VECTOR); + free_cpumask_var(cleanup_mask); + } + cfg->move_in_progress = 0; +} + static void __target_IO_APIC_irq(unsigned int irq, unsigned int dest, struct irq_cfg *cfg) { int apic, pin; @@ -400,40 +438,52 @@ static void __target_IO_APIC_irq(unsigned int irq, unsigned int dest, struct irq static int assign_irq_vector(int irq, struct irq_cfg *cfg, const struct cpumask *mask); -static void set_ioapic_affinity_irq_desc(struct irq_desc *desc, - const struct cpumask *mask) +/* + * Either sets desc->affinity to a valid value, and returns cpu_mask_to_apicid + * of that, or returns BAD_APICID and leaves desc->affinity untouched. + */ +static unsigned int +set_desc_affinity(struct irq_desc *desc, const struct cpumask *mask) { struct irq_cfg *cfg; - unsigned long flags; - unsigned int dest; - cpumask_t tmp; unsigned int irq; if (!cpumask_intersects(mask, cpu_online_mask)) - return; + return BAD_APICID; irq = desc->irq; cfg = desc->chip_data; if (assign_irq_vector(irq, cfg, mask)) - return; + return BAD_APICID; + cpumask_and(&desc->affinity, cfg->domain, mask); set_extra_move_desc(desc, mask); + return cpu_mask_to_apicid_and(&desc->affinity, cpu_online_mask); +} - cpumask_and(&tmp, &cfg->domain, mask); - dest = cpu_mask_to_apicid(&tmp); - /* - * Only the high 8 bits are valid. - */ - dest = SET_APIC_LOGICAL_ID(dest); +static void +set_ioapic_affinity_irq_desc(struct irq_desc *desc, const struct cpumask *mask) +{ + struct irq_cfg *cfg; + unsigned long flags; + unsigned int dest; + unsigned int irq; + + irq = desc->irq; + cfg = desc->chip_data; spin_lock_irqsave(&ioapic_lock, flags); - __target_IO_APIC_irq(irq, dest, cfg); - cpumask_copy(&desc->affinity, mask); + dest = set_desc_affinity(desc, mask); + if (dest != BAD_APICID) { + /* Only the high 8 bits are valid. */ + dest = SET_APIC_LOGICAL_ID(dest); + __target_IO_APIC_irq(irq, dest, cfg); + } spin_unlock_irqrestore(&ioapic_lock, flags); } -static void set_ioapic_affinity_irq(unsigned int irq, - const struct cpumask *mask) +static void +set_ioapic_affinity_irq(unsigned int irq, const struct cpumask *mask) { struct irq_desc *desc; @@ -1117,26 +1167,32 @@ __assign_irq_vector(int irq, struct irq_cfg *cfg, const struct cpumask *mask) */ static int current_vector = FIRST_DEVICE_VECTOR, current_offset = 0; unsigned int old_vector; - int cpu; - cpumask_t tmp_mask; + int cpu, err; + cpumask_var_t tmp_mask; if ((cfg->move_in_progress) || cfg->move_cleanup_count) return -EBUSY; + if (!alloc_cpumask_var(&tmp_mask, GFP_ATOMIC)) + return -ENOMEM; + old_vector = cfg->vector; if (old_vector) { - cpus_and(tmp_mask, *mask, cpu_online_map); - cpus_and(tmp_mask, cfg->domain, tmp_mask); - if (!cpus_empty(tmp_mask)) + cpumask_and(tmp_mask, mask, cpu_online_mask); + cpumask_and(tmp_mask, cfg->domain, tmp_mask); + if (!cpumask_empty(tmp_mask)) { + free_cpumask_var(tmp_mask); return 0; + } } /* Only try and allocate irqs on cpus that are present */ - for_each_cpu_and(cpu, mask, &cpu_online_map) { + err = -ENOSPC; + for_each_cpu_and(cpu, mask, cpu_online_mask) { int new_cpu; int vector, offset; - vector_allocation_domain(cpu, &tmp_mask); + vector_allocation_domain(cpu, tmp_mask); vector = current_vector; offset = current_offset; @@ -1156,7 +1212,7 @@ next: if (vector == SYSCALL_VECTOR) goto next; #endif - for_each_cpu_and(new_cpu, &tmp_mask, &cpu_online_map) + for_each_cpu_and(new_cpu, tmp_mask, cpu_online_mask) if (per_cpu(vector_irq, new_cpu)[vector] != -1) goto next; /* Found one! */ @@ -1164,15 +1220,17 @@ next: current_offset = offset; if (old_vector) { cfg->move_in_progress = 1; - cfg->old_domain = cfg->domain; + cpumask_copy(cfg->old_domain, cfg->domain); } - for_each_cpu_and(new_cpu, &tmp_mask, &cpu_online_map) + for_each_cpu_and(new_cpu, tmp_mask, cpu_online_mask) per_cpu(vector_irq, new_cpu)[vector] = irq; cfg->vector = vector; - cfg->domain = tmp_mask; - return 0; + cpumask_copy(cfg->domain, tmp_mask); + err = 0; + break; } - return -ENOSPC; + free_cpumask_var(tmp_mask); + return err; } static int @@ -1189,23 +1247,20 @@ assign_irq_vector(int irq, struct irq_cfg *cfg, const struct cpumask *mask) static void __clear_irq_vector(int irq, struct irq_cfg *cfg) { - cpumask_t mask; int cpu, vector; BUG_ON(!cfg->vector); vector = cfg->vector; - cpus_and(mask, cfg->domain, cpu_online_map); - for_each_cpu_mask_nr(cpu, mask) + for_each_cpu_and(cpu, cfg->domain, cpu_online_mask) per_cpu(vector_irq, cpu)[vector] = -1; cfg->vector = 0; - cpus_clear(cfg->domain); + cpumask_clear(cfg->domain); if (likely(!cfg->move_in_progress)) return; - cpus_and(mask, cfg->old_domain, cpu_online_map); - for_each_cpu_mask_nr(cpu, mask) { + for_each_cpu_and(cpu, cfg->old_domain, cpu_online_mask) { for (vector = FIRST_EXTERNAL_VECTOR; vector < NR_VECTORS; vector++) { if (per_cpu(vector_irq, cpu)[vector] != irq) @@ -1230,7 +1285,7 @@ void __setup_vector_irq(int cpu) if (!desc) continue; cfg = desc->chip_data; - if (!cpu_isset(cpu, cfg->domain)) + if (!cpumask_test_cpu(cpu, cfg->domain)) continue; vector = cfg->vector; per_cpu(vector_irq, cpu)[vector] = irq; @@ -1242,7 +1297,7 @@ void __setup_vector_irq(int cpu) continue; cfg = irq_cfg(irq); - if (!cpu_isset(cpu, cfg->domain)) + if (!cpumask_test_cpu(cpu, cfg->domain)) per_cpu(vector_irq, cpu)[vector] = -1; } } @@ -1378,18 +1433,17 @@ static void setup_IO_APIC_irq(int apic, int pin, unsigned int irq, struct irq_de { struct irq_cfg *cfg; struct IO_APIC_route_entry entry; - cpumask_t mask; + unsigned int dest; if (!IO_APIC_IRQ(irq)) return; cfg = desc->chip_data; - mask = *TARGET_CPUS; - if (assign_irq_vector(irq, cfg, &mask)) + if (assign_irq_vector(irq, cfg, TARGET_CPUS)) return; - cpus_and(mask, cfg->domain, mask); + dest = cpu_mask_to_apicid_and(cfg->domain, TARGET_CPUS); apic_printk(APIC_VERBOSE,KERN_DEBUG "IOAPIC[%d]: Set routing entry (%d-%d -> 0x%x -> " @@ -1399,8 +1453,7 @@ static void setup_IO_APIC_irq(int apic, int pin, unsigned int irq, struct irq_de if (setup_ioapic_entry(mp_ioapics[apic].mp_apicid, irq, &entry, - cpu_mask_to_apicid(&mask), trigger, polarity, - cfg->vector)) { + dest, trigger, polarity, cfg->vector)) { printk("Failed to setup ioapic entry for ioapic %d, pin %d\n", mp_ioapics[apic].mp_apicid, pin); __clear_irq_vector(irq, cfg); @@ -2122,7 +2175,7 @@ static int ioapic_retrigger_irq(unsigned int irq) unsigned long flags; spin_lock_irqsave(&vector_lock, flags); - send_IPI_mask(&cpumask_of_cpu(first_cpu(cfg->domain)), cfg->vector); + send_IPI_mask(cpumask_of(cpumask_first(cfg->domain)), cfg->vector); spin_unlock_irqrestore(&vector_lock, flags); return 1; @@ -2175,15 +2228,13 @@ static void migrate_ioapic_irq_desc(struct irq_desc *desc, const struct cpumask *mask) { struct irq_cfg *cfg; - cpumask_t tmpmask; struct irte irte; int modify_ioapic_rte; unsigned int dest; unsigned long flags; unsigned int irq; - cpus_and(tmpmask, *mask, cpu_online_map); - if (cpus_empty(tmpmask)) + if (!cpumask_intersects(mask, cpu_online_mask)) return; irq = desc->irq; @@ -2196,8 +2247,7 @@ migrate_ioapic_irq_desc(struct irq_desc *desc, const struct cpumask *mask) set_extra_move_desc(desc, mask); - cpus_and(tmpmask, cfg->domain, *mask); - dest = cpu_mask_to_apicid(&tmpmask); + dest = cpu_mask_to_apicid_and(cfg->domain, mask); modify_ioapic_rte = desc->status & IRQ_LEVEL; if (modify_ioapic_rte) { @@ -2214,14 +2264,10 @@ migrate_ioapic_irq_desc(struct irq_desc *desc, const struct cpumask *mask) */ modify_irte(irq, &irte); - if (cfg->move_in_progress) { - cpus_and(tmpmask, cfg->old_domain, cpu_online_map); - cfg->move_cleanup_count = cpus_weight(tmpmask); - send_IPI_mask(&tmpmask, IRQ_MOVE_CLEANUP_VECTOR); - cfg->move_in_progress = 0; - } + if (cfg->move_in_progress) + send_cleanup_vector(cfg); - desc->affinity = *mask; + cpumask_copy(&desc->affinity, mask); } static int migrate_irq_remapped_level_desc(struct irq_desc *desc) @@ -2247,7 +2293,7 @@ static int migrate_irq_remapped_level_desc(struct irq_desc *desc) ret = 0; desc->status &= ~IRQ_MOVE_PENDING; - cpus_clear(desc->pending_mask); + cpumask_clear(&desc->pending_mask); unmask: unmask_IO_APIC_irq_desc(desc); @@ -2333,7 +2379,7 @@ asmlinkage void smp_irq_move_cleanup_interrupt(void) if (!cfg->move_cleanup_count) goto unlock; - if ((vector == cfg->vector) && cpu_isset(me, cfg->domain)) + if (vector == cfg->vector && cpumask_test_cpu(me, cfg->domain)) goto unlock; __get_cpu_var(vector_irq)[vector] = -1; @@ -2356,14 +2402,8 @@ static void irq_complete_move(struct irq_desc **descp) vector = ~get_irq_regs()->orig_ax; me = smp_processor_id(); - if ((vector == cfg->vector) && cpu_isset(me, cfg->domain)) { - cpumask_t cleanup_mask; - - cpus_and(cleanup_mask, cfg->old_domain, cpu_online_map); - cfg->move_cleanup_count = cpus_weight(cleanup_mask); - send_IPI_mask(&cleanup_mask, IRQ_MOVE_CLEANUP_VECTOR); - cfg->move_in_progress = 0; - } + if (vector == cfg->vector && cpumask_test_cpu(me, cfg->domain)) + send_cleanup_vector(cfg); } #else static inline void irq_complete_move(struct irq_desc **descp) {} @@ -3088,16 +3128,13 @@ static int msi_compose_msg(struct pci_dev *pdev, unsigned int irq, struct msi_ms struct irq_cfg *cfg; int err; unsigned dest; - cpumask_t tmp; cfg = irq_cfg(irq); - tmp = *TARGET_CPUS; - err = assign_irq_vector(irq, cfg, &tmp); + err = assign_irq_vector(irq, cfg, TARGET_CPUS); if (err) return err; - cpus_and(tmp, cfg->domain, tmp); - dest = cpu_mask_to_apicid(&tmp); + dest = cpu_mask_to_apicid_and(cfg->domain, TARGET_CPUS); #ifdef CONFIG_INTR_REMAP if (irq_remapped(irq)) { @@ -3157,19 +3194,12 @@ static void set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) struct irq_cfg *cfg; struct msi_msg msg; unsigned int dest; - cpumask_t tmp; - if (!cpumask_intersects(mask, cpu_online_mask)) + dest = set_desc_affinity(desc, mask); + if (dest == BAD_APICID) return; cfg = desc->chip_data; - if (assign_irq_vector(irq, cfg, mask)) - return; - - set_extra_move_desc(desc, mask); - - cpumask_and(&tmp, &cfg->domain, mask); - dest = cpu_mask_to_apicid(&tmp); read_msi_msg_desc(desc, &msg); @@ -3179,7 +3209,6 @@ static void set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) msg.address_lo |= MSI_ADDR_DEST_ID(dest); write_msi_msg_desc(desc, &msg); - cpumask_copy(&desc->affinity, mask); } #ifdef CONFIG_INTR_REMAP /* @@ -3192,24 +3221,15 @@ ir_set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) struct irq_desc *desc = irq_to_desc(irq); struct irq_cfg *cfg; unsigned int dest; - cpumask_t tmp, cleanup_mask; struct irte irte; - if (!cpumask_intersects(mask, cpu_online_mask)) - return; - if (get_irte(irq, &irte)) return; - cfg = desc->chip_data; - if (assign_irq_vector(irq, cfg, mask)) + dest = set_desc_affinity(desc, mask); + if (dest == BAD_APICID) return; - set_extra_move_desc(desc, mask); - - cpumask_and(&tmp, &cfg->domain, mask); - dest = cpu_mask_to_apicid(&tmp); - irte.vector = cfg->vector; irte.dest_id = IRTE_DEST(dest); @@ -3223,14 +3243,8 @@ ir_set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) * at the new destination. So, time to cleanup the previous * vector allocation. */ - if (cfg->move_in_progress) { - cpus_and(cleanup_mask, cfg->old_domain, cpu_online_map); - cfg->move_cleanup_count = cpus_weight(cleanup_mask); - send_IPI_mask(&cleanup_mask, IRQ_MOVE_CLEANUP_VECTOR); - cfg->move_in_progress = 0; - } - - cpumask_copy(&desc->affinity, mask); + if (cfg->move_in_progress) + send_cleanup_vector(cfg); } #endif @@ -3421,25 +3435,18 @@ void arch_teardown_msi_irq(unsigned int irq) #ifdef CONFIG_DMAR #ifdef CONFIG_SMP -static void dmar_msi_set_affinity(unsigned int irq, const cpumask_t *mask) +static void dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask) { struct irq_desc *desc = irq_to_desc(irq); struct irq_cfg *cfg; struct msi_msg msg; unsigned int dest; - cpumask_t tmp; - if (!cpumask_intersects(mask, cpu_online_mask)) + dest = set_desc_affinity(desc, mask); + if (dest == BAD_APICID) return; cfg = desc->chip_data; - if (assign_irq_vector(irq, cfg, mask)) - return; - - set_extra_move_desc(desc, mask); - - cpumask_and(&tmp, &cfg->domain, mask); - dest = cpu_mask_to_apicid(&tmp); dmar_msi_read(irq, &msg); @@ -3449,7 +3456,6 @@ static void dmar_msi_set_affinity(unsigned int irq, const cpumask_t *mask) msg.address_lo |= MSI_ADDR_DEST_ID(dest); dmar_msi_write(irq, &msg); - cpumask_copy(&desc->affinity, mask); } #endif /* CONFIG_SMP */ @@ -3483,25 +3489,18 @@ int arch_setup_dmar_msi(unsigned int irq) #ifdef CONFIG_HPET_TIMER #ifdef CONFIG_SMP -static void hpet_msi_set_affinity(unsigned int irq, const cpumask_t *mask) +static void hpet_msi_set_affinity(unsigned int irq, const struct cpumask *mask) { struct irq_desc *desc = irq_to_desc(irq); struct irq_cfg *cfg; struct msi_msg msg; unsigned int dest; - cpumask_t tmp; - if (!cpumask_intersects(mask, cpu_online_mask)) + dest = set_desc_affinity(desc, mask); + if (dest == BAD_APICID) return; cfg = desc->chip_data; - if (assign_irq_vector(irq, cfg, mask)) - return; - - set_extra_move_desc(desc, mask); - - cpumask_and(&tmp, &cfg->domain, mask); - dest = cpu_mask_to_apicid(&tmp); hpet_msi_read(irq, &msg); @@ -3511,7 +3510,6 @@ static void hpet_msi_set_affinity(unsigned int irq, const cpumask_t *mask) msg.address_lo |= MSI_ADDR_DEST_ID(dest); hpet_msi_write(irq, &msg); - cpumask_copy(&desc->affinity, mask); } #endif /* CONFIG_SMP */ @@ -3566,27 +3564,19 @@ static void target_ht_irq(unsigned int irq, unsigned int dest, u8 vector) write_ht_irq_msg(irq, &msg); } -static void set_ht_irq_affinity(unsigned int irq, const cpumask_t *mask) +static void set_ht_irq_affinity(unsigned int irq, const struct cpumask *mask) { struct irq_desc *desc = irq_to_desc(irq); struct irq_cfg *cfg; unsigned int dest; - cpumask_t tmp; - if (!cpumask_intersects(mask, cpu_online_mask)) + dest = set_desc_affinity(desc, mask); + if (dest == BAD_APICID) return; cfg = desc->chip_data; - if (assign_irq_vector(irq, cfg, mask)) - return; - - set_extra_move_desc(desc, mask); - - cpumask_and(&tmp, &cfg->domain, mask); - dest = cpu_mask_to_apicid(&tmp); target_ht_irq(irq, dest, cfg->vector); - cpumask_copy(&desc->affinity, mask); } #endif @@ -3606,7 +3596,6 @@ int arch_setup_ht_irq(unsigned int irq, struct pci_dev *dev) { struct irq_cfg *cfg; int err; - cpumask_t tmp; cfg = irq_cfg(irq); err = assign_irq_vector(irq, cfg, TARGET_CPUS); @@ -3614,8 +3603,7 @@ int arch_setup_ht_irq(unsigned int irq, struct pci_dev *dev) struct ht_irq_msg msg; unsigned dest; - cpus_and(tmp, cfg->domain, tmp); - dest = cpu_mask_to_apicid(&tmp); + dest = cpu_mask_to_apicid_and(cfg->domain, TARGET_CPUS); msg.address_hi = HT_IRQ_HIGH_DEST_ID(dest); @@ -3651,7 +3639,7 @@ int arch_setup_ht_irq(unsigned int irq, struct pci_dev *dev) int arch_enable_uv_irq(char *irq_name, unsigned int irq, int cpu, int mmr_blade, unsigned long mmr_offset) { - const cpumask_t *eligible_cpu = &cpumask_of_cpu(cpu); + const struct cpumask *eligible_cpu = cpumask_of(cpu); struct irq_cfg *cfg; int mmr_pnode; unsigned long mmr_value; @@ -3891,7 +3879,7 @@ void __init setup_ioapic_dest(void) int pin, ioapic, irq, irq_entry; struct irq_desc *desc; struct irq_cfg *cfg; - const cpumask_t *mask; + const struct cpumask *mask; if (skip_ioapic_setup == 1) return; -- cgit v1.2.3 From b78936e14ee47b6b2d628501a0eab5270db80132 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Tue, 16 Dec 2008 17:33:57 -0800 Subject: xen: convert to cpumask_var_t and new cpumask primitives. Simple change, and eventual space saving when NR_CPUS >> nr_cpu_ids. Signed-off-by: Rusty Russell Signed-off-by: Mike Travis Cc: Jeremy Fitzhardinge --- arch/x86/xen/smp.c | 9 ++++++--- arch/x86/xen/suspend.c | 3 ++- arch/x86/xen/xen-ops.h | 2 +- 3 files changed, 9 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/xen/smp.c b/arch/x86/xen/smp.c index 2cce362c9874..b3a95868839b 100644 --- a/arch/x86/xen/smp.c +++ b/arch/x86/xen/smp.c @@ -33,7 +33,7 @@ #include "xen-ops.h" #include "mmu.h" -cpumask_t xen_cpu_initialized_map; +cpumask_var_t xen_cpu_initialized_map; static DEFINE_PER_CPU(int, resched_irq); static DEFINE_PER_CPU(int, callfunc_irq); @@ -192,7 +192,10 @@ static void __init xen_smp_prepare_cpus(unsigned int max_cpus) if (xen_smp_intr_init(0)) BUG(); - xen_cpu_initialized_map = cpumask_of_cpu(0); + if (!alloc_cpumask_var(&xen_cpu_initialized_map, GFP_KERNEL)) + panic("could not allocate xen_cpu_initialized_map\n"); + + cpumask_copy(xen_cpu_initialized_map, cpumask_of(0)); /* Restrict the possible_map according to max_cpus. */ while ((num_possible_cpus() > 1) && (num_possible_cpus() > max_cpus)) { @@ -221,7 +224,7 @@ cpu_initialize_context(unsigned int cpu, struct task_struct *idle) struct vcpu_guest_context *ctxt; struct desc_struct *gdt; - if (cpu_test_and_set(cpu, xen_cpu_initialized_map)) + if (cpumask_test_and_set_cpu(cpu, xen_cpu_initialized_map)) return 0; ctxt = kzalloc(sizeof(*ctxt), GFP_KERNEL); diff --git a/arch/x86/xen/suspend.c b/arch/x86/xen/suspend.c index 2a234db5949b..212ffe012b76 100644 --- a/arch/x86/xen/suspend.c +++ b/arch/x86/xen/suspend.c @@ -35,7 +35,8 @@ void xen_post_suspend(int suspend_cancelled) pfn_to_mfn(xen_start_info->console.domU.mfn); } else { #ifdef CONFIG_SMP - xen_cpu_initialized_map = cpu_online_map; + BUG_ON(xen_cpu_initialized_map == NULL); + cpumask_copy(xen_cpu_initialized_map, cpu_online_mask); #endif xen_vcpu_restore(); } diff --git a/arch/x86/xen/xen-ops.h b/arch/x86/xen/xen-ops.h index 9e1afae8461f..c1f8faf0a2c5 100644 --- a/arch/x86/xen/xen-ops.h +++ b/arch/x86/xen/xen-ops.h @@ -58,7 +58,7 @@ void __init xen_init_spinlocks(void); __cpuinit void xen_init_lock_cpu(int cpu); void xen_uninit_lock_cpu(int cpu); -extern cpumask_t xen_cpu_initialized_map; +extern cpumask_var_t xen_cpu_initialized_map; #else static inline void xen_smp_init(void) {} #endif -- cgit v1.2.3 From d7b381bb7b1ad69ff008ea063d26e988b686c8de Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Tue, 16 Dec 2008 17:33:58 -0800 Subject: x86: fixup_irqs() doesnt need an argument. Impact: cleanup, remove on-stack cpumask. The "map" arg is always cpu_online_mask. Importantly, set_affinity always ands the argument with cpu_online_mask anyway, so we don't need to do it in fixup_irqs(), avoiding a temporary. Signed-off-by: Rusty Russell Signed-off-by: Mike Travis --- arch/x86/include/asm/irq.h | 2 +- arch/x86/kernel/irq_32.c | 13 +++++++------ arch/x86/kernel/irq_64.c | 15 ++++++++------- arch/x86/kernel/smpboot.c | 2 +- 4 files changed, 17 insertions(+), 15 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/irq.h b/arch/x86/include/asm/irq.h index bae0eda95486..8766d30fb746 100644 --- a/arch/x86/include/asm/irq.h +++ b/arch/x86/include/asm/irq.h @@ -37,7 +37,7 @@ extern int irqbalance_disable(char *str); #ifdef CONFIG_HOTPLUG_CPU #include -extern void fixup_irqs(cpumask_t map); +extern void fixup_irqs(void); #endif extern unsigned int do_IRQ(struct pt_regs *regs); diff --git a/arch/x86/kernel/irq_32.c b/arch/x86/kernel/irq_32.c index 9cf9cbbf7a02..9dc5588f336a 100644 --- a/arch/x86/kernel/irq_32.c +++ b/arch/x86/kernel/irq_32.c @@ -233,27 +233,28 @@ unsigned int do_IRQ(struct pt_regs *regs) #ifdef CONFIG_HOTPLUG_CPU #include -void fixup_irqs(cpumask_t map) +/* A cpu has been removed from cpu_online_mask. Reset irq affinities. */ +void fixup_irqs(void) { unsigned int irq; static int warned; struct irq_desc *desc; for_each_irq_desc(irq, desc) { - cpumask_t mask; + const struct cpumask *affinity; if (!desc) continue; if (irq == 2) continue; - cpus_and(mask, desc->affinity, map); - if (any_online_cpu(mask) == NR_CPUS) { + affinity = &desc->affinity; + if (cpumask_any_and(affinity, cpu_online_mask) >= nr_cpu_ids) { printk("Breaking affinity for irq %i\n", irq); - mask = map; + affinity = cpu_all_mask; } if (desc->chip->set_affinity) - desc->chip->set_affinity(irq, &mask); + desc->chip->set_affinity(irq, affinity); else if (desc->action && !(warned++)) printk("Cannot set affinity for irq %i\n", irq); } diff --git a/arch/x86/kernel/irq_64.c b/arch/x86/kernel/irq_64.c index 27f2307b0a34..fca2991443f5 100644 --- a/arch/x86/kernel/irq_64.c +++ b/arch/x86/kernel/irq_64.c @@ -83,16 +83,17 @@ asmlinkage unsigned int do_IRQ(struct pt_regs *regs) } #ifdef CONFIG_HOTPLUG_CPU -void fixup_irqs(cpumask_t map) +/* A cpu has been removed from cpu_online_mask. Reset irq affinities. */ +void fixup_irqs(void) { unsigned int irq; static int warned; struct irq_desc *desc; for_each_irq_desc(irq, desc) { - cpumask_t mask; int break_affinity = 0; int set_affinity = 1; + const struct cpumask *affinity; if (!desc) continue; @@ -102,23 +103,23 @@ void fixup_irqs(cpumask_t map) /* interrupt's are disabled at this point */ spin_lock(&desc->lock); + affinity = &desc->affinity; if (!irq_has_action(irq) || - cpus_equal(desc->affinity, map)) { + cpumask_equal(affinity, cpu_online_mask)) { spin_unlock(&desc->lock); continue; } - cpus_and(mask, desc->affinity, map); - if (cpus_empty(mask)) { + if (cpumask_any_and(affinity, cpu_online_mask) >= nr_cpu_ids) { break_affinity = 1; - mask = map; + affinity = cpu_all_mask; } if (desc->chip->mask) desc->chip->mask(irq); if (desc->chip->set_affinity) - desc->chip->set_affinity(irq, &mask); + desc->chip->set_affinity(irq, affinity); else if (!(warned++)) set_affinity = 0; diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index 9d58134e0231..8b6f675b363b 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -1346,7 +1346,7 @@ void cpu_disable_common(void) lock_vector_lock(); remove_cpu_from_maps(cpu); unlock_vector_lock(); - fixup_irqs(cpu_online_map); + fixup_irqs(); } int native_cpu_disable(void) -- cgit v1.2.3 From bcda016eddd7a8b374bb371473c821a91ff1d8cc Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Tue, 16 Dec 2008 17:33:59 -0800 Subject: x86: cosmetic changes apic-related files. This patch simply changes cpumask_t to struct cpumask and similar trivial modernizations. Signed-off-by: Rusty Russell Signed-off-by: Mike Travis --- arch/x86/include/asm/bigsmp/ipi.h | 14 +++----- arch/x86/include/asm/es7000/ipi.h | 13 +++---- arch/x86/include/asm/genapic_32.h | 11 +++--- arch/x86/include/asm/genapic_64.h | 11 +++--- arch/x86/include/asm/ipi.h | 10 +++--- arch/x86/include/asm/mach-default/mach_apic.h | 12 +++---- arch/x86/include/asm/mach-default/mach_ipi.h | 12 +++---- arch/x86/include/asm/numaq/ipi.h | 14 +++----- arch/x86/include/asm/smp.h | 4 +-- arch/x86/kernel/genapic_flat_64.c | 50 ++++++++++++++------------- arch/x86/kernel/genx2apic_cluster.c | 25 +++++++------- arch/x86/kernel/genx2apic_phys.c | 25 +++++++------- arch/x86/kernel/genx2apic_uv_x.c | 24 ++++++------- arch/x86/kernel/ipi.c | 14 ++++---- arch/x86/kernel/smp.c | 6 ++-- arch/x86/xen/smp.c | 11 +++--- 16 files changed, 127 insertions(+), 129 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/bigsmp/ipi.h b/arch/x86/include/asm/bigsmp/ipi.h index 63553e9f22b2..27fcd01b3ae6 100644 --- a/arch/x86/include/asm/bigsmp/ipi.h +++ b/arch/x86/include/asm/bigsmp/ipi.h @@ -1,26 +1,22 @@ #ifndef __ASM_MACH_IPI_H #define __ASM_MACH_IPI_H -void send_IPI_mask_sequence(const cpumask_t *mask, int vector); -void send_IPI_mask_allbutself(const cpumask_t *mask, int vector); +void send_IPI_mask_sequence(const struct cpumask *mask, int vector); +void send_IPI_mask_allbutself(const struct cpumask *mask, int vector); -static inline void send_IPI_mask(const cpumask_t *mask, int vector) +static inline void send_IPI_mask(const struct cpumask *mask, int vector) { send_IPI_mask_sequence(mask, vector); } static inline void send_IPI_allbutself(int vector) { - cpumask_t mask = cpu_online_map; - cpu_clear(smp_processor_id(), mask); - - if (!cpus_empty(mask)) - send_IPI_mask(&mask, vector); + send_IPI_mask_allbutself(cpu_online_mask, vector); } static inline void send_IPI_all(int vector) { - send_IPI_mask(&cpu_online_map, vector); + send_IPI_mask(cpu_online_mask, vector); } #endif /* __ASM_MACH_IPI_H */ diff --git a/arch/x86/include/asm/es7000/ipi.h b/arch/x86/include/asm/es7000/ipi.h index 1a8507265f91..7e8ed24d4b8a 100644 --- a/arch/x86/include/asm/es7000/ipi.h +++ b/arch/x86/include/asm/es7000/ipi.h @@ -1,25 +1,22 @@ #ifndef __ASM_ES7000_IPI_H #define __ASM_ES7000_IPI_H -void send_IPI_mask_sequence(const cpumask_t *mask, int vector); -void send_IPI_mask_allbutself(const cpumask_t *mask, int vector); +void send_IPI_mask_sequence(const struct cpumask *mask, int vector); +void send_IPI_mask_allbutself(const struct cpumask *mask, int vector); -static inline void send_IPI_mask(const cpumask_t *mask, int vector) +static inline void send_IPI_mask(const struct cpumask *mask, int vector) { send_IPI_mask_sequence(mask, vector); } static inline void send_IPI_allbutself(int vector) { - cpumask_t mask = cpu_online_map; - cpu_clear(smp_processor_id(), mask); - if (!cpus_empty(mask)) - send_IPI_mask(&mask, vector); + send_IPI_mask_allbutself(cpu_online_mask, vector); } static inline void send_IPI_all(int vector) { - send_IPI_mask(&cpu_online_map, vector); + send_IPI_mask(cpu_online_mask, vector); } #endif /* __ASM_ES7000_IPI_H */ diff --git a/arch/x86/include/asm/genapic_32.h b/arch/x86/include/asm/genapic_32.h index eed6e305291f..746f37a7963a 100644 --- a/arch/x86/include/asm/genapic_32.h +++ b/arch/x86/include/asm/genapic_32.h @@ -24,7 +24,7 @@ struct genapic { int (*probe)(void); int (*apic_id_registered)(void); - const cpumask_t *(*target_cpus)(void); + const struct cpumask *(*target_cpus)(void); int int_delivery_mode; int int_dest_mode; int ESR_DISABLE; @@ -57,15 +57,16 @@ struct genapic { unsigned (*get_apic_id)(unsigned long x); unsigned long apic_id_mask; - unsigned int (*cpu_mask_to_apicid)(const cpumask_t *cpumask); + unsigned int (*cpu_mask_to_apicid)(const struct cpumask *cpumask); unsigned int (*cpu_mask_to_apicid_and)(const struct cpumask *cpumask, const struct cpumask *andmask); - void (*vector_allocation_domain)(int cpu, cpumask_t *retmask); + void (*vector_allocation_domain)(int cpu, struct cpumask *retmask); #ifdef CONFIG_SMP /* ipi */ - void (*send_IPI_mask)(const cpumask_t *mask, int vector); - void (*send_IPI_mask_allbutself)(const cpumask_t *mask, int vector); + void (*send_IPI_mask)(const struct cpumask *mask, int vector); + void (*send_IPI_mask_allbutself)(const struct cpumask *mask, + int vector); void (*send_IPI_allbutself)(int vector); void (*send_IPI_all)(int vector); #endif diff --git a/arch/x86/include/asm/genapic_64.h b/arch/x86/include/asm/genapic_64.h index 244b71729ecb..adf32fb56aa6 100644 --- a/arch/x86/include/asm/genapic_64.h +++ b/arch/x86/include/asm/genapic_64.h @@ -20,17 +20,18 @@ struct genapic { u32 int_delivery_mode; u32 int_dest_mode; int (*apic_id_registered)(void); - const cpumask_t *(*target_cpus)(void); - void (*vector_allocation_domain)(int cpu, cpumask_t *retmask); + const struct cpumask *(*target_cpus)(void); + void (*vector_allocation_domain)(int cpu, struct cpumask *retmask); void (*init_apic_ldr)(void); /* ipi */ - void (*send_IPI_mask)(const cpumask_t *mask, int vector); - void (*send_IPI_mask_allbutself)(const cpumask_t *mask, int vector); + void (*send_IPI_mask)(const struct cpumask *mask, int vector); + void (*send_IPI_mask_allbutself)(const struct cpumask *mask, + int vector); void (*send_IPI_allbutself)(int vector); void (*send_IPI_all)(int vector); void (*send_IPI_self)(int vector); /* */ - unsigned int (*cpu_mask_to_apicid)(const cpumask_t *cpumask); + unsigned int (*cpu_mask_to_apicid)(const struct cpumask *cpumask); unsigned int (*cpu_mask_to_apicid_and)(const struct cpumask *cpumask, const struct cpumask *andmask); unsigned int (*phys_pkg_id)(int index_msb); diff --git a/arch/x86/include/asm/ipi.h b/arch/x86/include/asm/ipi.h index 24b6e613edfa..c745a306f7d3 100644 --- a/arch/x86/include/asm/ipi.h +++ b/arch/x86/include/asm/ipi.h @@ -117,7 +117,8 @@ static inline void __send_IPI_dest_field(unsigned int mask, int vector, native_apic_mem_write(APIC_ICR, cfg); } -static inline void send_IPI_mask_sequence(const cpumask_t *mask, int vector) +static inline void send_IPI_mask_sequence(const struct cpumask *mask, + int vector) { unsigned long flags; unsigned long query_cpu; @@ -128,14 +129,15 @@ static inline void send_IPI_mask_sequence(const cpumask_t *mask, int vector) * - mbligh */ local_irq_save(flags); - for_each_cpu_mask_nr(query_cpu, *mask) { + for_each_cpu(query_cpu, mask) { __send_IPI_dest_field(per_cpu(x86_cpu_to_apicid, query_cpu), vector, APIC_DEST_PHYSICAL); } local_irq_restore(flags); } -static inline void send_IPI_mask_allbutself(const cpumask_t *mask, int vector) +static inline void send_IPI_mask_allbutself(const struct cpumask *mask, + int vector) { unsigned long flags; unsigned int query_cpu; @@ -144,7 +146,7 @@ static inline void send_IPI_mask_allbutself(const cpumask_t *mask, int vector) /* See Hack comment above */ local_irq_save(flags); - for_each_cpu_mask_nr(query_cpu, *mask) + for_each_cpu(query_cpu, mask) if (query_cpu != this_cpu) __send_IPI_dest_field( per_cpu(x86_cpu_to_apicid, query_cpu), diff --git a/arch/x86/include/asm/mach-default/mach_apic.h b/arch/x86/include/asm/mach-default/mach_apic.h index df8e024c43c5..8863d978cb96 100644 --- a/arch/x86/include/asm/mach-default/mach_apic.h +++ b/arch/x86/include/asm/mach-default/mach_apic.h @@ -8,12 +8,12 @@ #define APIC_DFR_VALUE (APIC_DFR_FLAT) -static inline const cpumask_t *target_cpus(void) +static inline const struct cpumask *target_cpus(void) { #ifdef CONFIG_SMP - return &cpu_online_map; + return cpu_online_mask; #else - return &cpumask_of_cpu(0); + return cpumask_of(0); #endif } @@ -62,9 +62,9 @@ static inline int apic_id_registered(void) return physid_isset(read_apic_id(), phys_cpu_present_map); } -static inline unsigned int cpu_mask_to_apicid(const cpumask_t *cpumask) +static inline unsigned int cpu_mask_to_apicid(const struct cpumask *cpumask) { - return cpus_addr(*cpumask)[0]; + return cpumask_bits(cpumask)[0]; } static inline unsigned int cpu_mask_to_apicid_and(const struct cpumask *cpumask, @@ -98,7 +98,7 @@ static inline int apicid_to_node(int logical_apicid) #endif } -static inline void vector_allocation_domain(int cpu, cpumask_t *retmask) +static inline void vector_allocation_domain(int cpu, struct cpumask *retmask) { /* Careful. Some cpus do not strictly honor the set of cpus * specified in the interrupt destination when using lowest diff --git a/arch/x86/include/asm/mach-default/mach_ipi.h b/arch/x86/include/asm/mach-default/mach_ipi.h index 9353ab854a10..191312d155da 100644 --- a/arch/x86/include/asm/mach-default/mach_ipi.h +++ b/arch/x86/include/asm/mach-default/mach_ipi.h @@ -4,8 +4,8 @@ /* Avoid include hell */ #define NMI_VECTOR 0x02 -void send_IPI_mask_bitmask(const cpumask_t *mask, int vector); -void send_IPI_mask_allbutself(const cpumask_t *mask, int vector); +void send_IPI_mask_bitmask(const struct cpumask *mask, int vector); +void send_IPI_mask_allbutself(const struct cpumask *mask, int vector); void __send_IPI_shortcut(unsigned int shortcut, int vector); extern int no_broadcast; @@ -15,17 +15,17 @@ extern int no_broadcast; #define send_IPI_mask (genapic->send_IPI_mask) #define send_IPI_mask_allbutself (genapic->send_IPI_mask_allbutself) #else -static inline void send_IPI_mask(const cpumask_t *mask, int vector) +static inline void send_IPI_mask(const struct cpumask *mask, int vector) { send_IPI_mask_bitmask(mask, vector); } -void send_IPI_mask_allbutself(const cpumask_t *mask, int vector); +void send_IPI_mask_allbutself(const struct cpumask *mask, int vector); #endif static inline void __local_send_IPI_allbutself(int vector) { if (no_broadcast || vector == NMI_VECTOR) - send_IPI_mask_allbutself(&cpu_online_map, vector); + send_IPI_mask_allbutself(cpu_online_mask, vector); else __send_IPI_shortcut(APIC_DEST_ALLBUT, vector); } @@ -33,7 +33,7 @@ static inline void __local_send_IPI_allbutself(int vector) static inline void __local_send_IPI_all(int vector) { if (no_broadcast || vector == NMI_VECTOR) - send_IPI_mask(&cpu_online_map, vector); + send_IPI_mask(cpu_online_mask, vector); else __send_IPI_shortcut(APIC_DEST_ALLINC, vector); } diff --git a/arch/x86/include/asm/numaq/ipi.h b/arch/x86/include/asm/numaq/ipi.h index c734d7acc430..a8374c652778 100644 --- a/arch/x86/include/asm/numaq/ipi.h +++ b/arch/x86/include/asm/numaq/ipi.h @@ -1,26 +1,22 @@ #ifndef __ASM_NUMAQ_IPI_H #define __ASM_NUMAQ_IPI_H -void send_IPI_mask_sequence(const cpumask_t *mask, int vector); -void send_IPI_mask_allbutself(const cpumask_t *mask, int vector); +void send_IPI_mask_sequence(const struct cpumask *mask, int vector); +void send_IPI_mask_allbutself(const struct cpumask *mask, int vector); -static inline void send_IPI_mask(const cpumask_t *mask, int vector) +static inline void send_IPI_mask(const struct cpumask *mask, int vector) { send_IPI_mask_sequence(mask, vector); } static inline void send_IPI_allbutself(int vector) { - cpumask_t mask = cpu_online_map; - cpu_clear(smp_processor_id(), mask); - - if (!cpus_empty(mask)) - send_IPI_mask(&mask, vector); + send_IPI_mask_allbutself(cpu_online_mask, vector); } static inline void send_IPI_all(int vector) { - send_IPI_mask(&cpu_online_map, vector); + send_IPI_mask(cpu_online_mask, vector); } #endif /* __ASM_NUMAQ_IPI_H */ diff --git a/arch/x86/include/asm/smp.h b/arch/x86/include/asm/smp.h index c4a9aa52df6e..830b9fcb6427 100644 --- a/arch/x86/include/asm/smp.h +++ b/arch/x86/include/asm/smp.h @@ -60,7 +60,7 @@ struct smp_ops { void (*cpu_die)(unsigned int cpu); void (*play_dead)(void); - void (*send_call_func_ipi)(const cpumask_t *mask); + void (*send_call_func_ipi)(const struct cpumask *mask); void (*send_call_func_single_ipi)(int cpu); }; @@ -138,7 +138,7 @@ void native_cpu_die(unsigned int cpu); void native_play_dead(void); void play_dead_common(void); -void native_send_call_func_ipi(const cpumask_t *mask); +void native_send_call_func_ipi(const struct cpumask *mask); void native_send_call_func_single_ipi(int cpu); extern void prefill_possible_map(void); diff --git a/arch/x86/kernel/genapic_flat_64.c b/arch/x86/kernel/genapic_flat_64.c index c772bb10b173..7fa5f49c2dda 100644 --- a/arch/x86/kernel/genapic_flat_64.c +++ b/arch/x86/kernel/genapic_flat_64.c @@ -30,12 +30,12 @@ static int flat_acpi_madt_oem_check(char *oem_id, char *oem_table_id) return 1; } -static const cpumask_t *flat_target_cpus(void) +static const struct cpumask *flat_target_cpus(void) { - return &cpu_online_map; + return cpu_online_mask; } -static void flat_vector_allocation_domain(int cpu, cpumask_t *retmask) +static void flat_vector_allocation_domain(int cpu, struct cpumask *retmask) { /* Careful. Some cpus do not strictly honor the set of cpus * specified in the interrupt destination when using lowest @@ -45,7 +45,8 @@ static void flat_vector_allocation_domain(int cpu, cpumask_t *retmask) * deliver interrupts to the wrong hyperthread when only one * hyperthread was specified in the interrupt desitination. */ - *retmask = (cpumask_t) { {[0] = APIC_ALL_CPUS, } }; + cpumask_clear(retmask); + cpumask_bits(retmask)[0] = APIC_ALL_CPUS; } /* @@ -77,16 +78,17 @@ static inline void _flat_send_IPI_mask(unsigned long mask, int vector) local_irq_restore(flags); } -static void flat_send_IPI_mask(const cpumask_t *cpumask, int vector) +static void flat_send_IPI_mask(const struct cpumask *cpumask, int vector) { - unsigned long mask = cpus_addr(*cpumask)[0]; + unsigned long mask = cpumask_bits(cpumask)[0]; _flat_send_IPI_mask(mask, vector); } -static void flat_send_IPI_mask_allbutself(const cpumask_t *cpumask, int vector) +static void flat_send_IPI_mask_allbutself(const struct cpumask *cpumask, + int vector) { - unsigned long mask = cpus_addr(*cpumask)[0]; + unsigned long mask = cpumask_bits(cpumask)[0]; int cpu = smp_processor_id(); if (cpu < BITS_PER_LONG) @@ -103,8 +105,8 @@ static void flat_send_IPI_allbutself(int vector) int hotplug = 0; #endif if (hotplug || vector == NMI_VECTOR) { - if (!cpus_equal(cpu_online_map, cpumask_of_cpu(cpu))) { - unsigned long mask = cpus_addr(cpu_online_map)[0]; + if (!cpumask_equal(cpu_online_mask, cpumask_of(cpu))) { + unsigned long mask = cpumask_bits(cpu_online_mask)[0]; if (cpu < BITS_PER_LONG) clear_bit(cpu, &mask); @@ -119,7 +121,7 @@ static void flat_send_IPI_allbutself(int vector) static void flat_send_IPI_all(int vector) { if (vector == NMI_VECTOR) - flat_send_IPI_mask(&cpu_online_map, vector); + flat_send_IPI_mask(cpu_online_mask, vector); else __send_IPI_shortcut(APIC_DEST_ALLINC, vector, APIC_DEST_LOGICAL); } @@ -153,9 +155,9 @@ static int flat_apic_id_registered(void) return physid_isset(read_xapic_id(), phys_cpu_present_map); } -static unsigned int flat_cpu_mask_to_apicid(const cpumask_t *cpumask) +static unsigned int flat_cpu_mask_to_apicid(const struct cpumask *cpumask) { - return cpus_addr(*cpumask)[0] & APIC_ALL_CPUS; + return cpumask_bits(cpumask)[0] & APIC_ALL_CPUS; } static unsigned int flat_cpu_mask_to_apicid_and(const struct cpumask *cpumask, @@ -217,23 +219,23 @@ static int physflat_acpi_madt_oem_check(char *oem_id, char *oem_table_id) return 0; } -static const cpumask_t *physflat_target_cpus(void) +static const struct cpumask *physflat_target_cpus(void) { - return &cpu_online_map; + return cpu_online_mask; } -static void physflat_vector_allocation_domain(int cpu, cpumask_t *retmask) +static void physflat_vector_allocation_domain(int cpu, struct cpumask *retmask) { - cpus_clear(*retmask); - cpu_set(cpu, *retmask); + cpumask_clear(retmask); + cpumask_set_cpu(cpu, retmask); } -static void physflat_send_IPI_mask(const cpumask_t *cpumask, int vector) +static void physflat_send_IPI_mask(const struct cpumask *cpumask, int vector) { send_IPI_mask_sequence(cpumask, vector); } -static void physflat_send_IPI_mask_allbutself(const cpumask_t *cpumask, +static void physflat_send_IPI_mask_allbutself(const struct cpumask *cpumask, int vector) { send_IPI_mask_allbutself(cpumask, vector); @@ -241,15 +243,15 @@ static void physflat_send_IPI_mask_allbutself(const cpumask_t *cpumask, static void physflat_send_IPI_allbutself(int vector) { - send_IPI_mask_allbutself(&cpu_online_map, vector); + send_IPI_mask_allbutself(cpu_online_mask, vector); } static void physflat_send_IPI_all(int vector) { - physflat_send_IPI_mask(&cpu_online_map, vector); + physflat_send_IPI_mask(cpu_online_mask, vector); } -static unsigned int physflat_cpu_mask_to_apicid(const cpumask_t *cpumask) +static unsigned int physflat_cpu_mask_to_apicid(const struct cpumask *cpumask) { int cpu; @@ -257,7 +259,7 @@ static unsigned int physflat_cpu_mask_to_apicid(const cpumask_t *cpumask) * We're using fixed IRQ delivery, can only return one phys APIC ID. * May as well be the first. */ - cpu = first_cpu(*cpumask); + cpu = cpumask_first(cpumask); if ((unsigned)cpu < nr_cpu_ids) return per_cpu(x86_cpu_to_apicid, cpu); else diff --git a/arch/x86/kernel/genx2apic_cluster.c b/arch/x86/kernel/genx2apic_cluster.c index e7d16f53b9cd..4716a0c9f936 100644 --- a/arch/x86/kernel/genx2apic_cluster.c +++ b/arch/x86/kernel/genx2apic_cluster.c @@ -22,18 +22,18 @@ static int x2apic_acpi_madt_oem_check(char *oem_id, char *oem_table_id) /* Start with all IRQs pointing to boot CPU. IRQ balancing will shift them. */ -static const cpumask_t *x2apic_target_cpus(void) +static const struct cpumask *x2apic_target_cpus(void) { - return &cpumask_of_cpu(0); + return cpumask_of(0); } /* * for now each logical cpu is in its own vector allocation domain. */ -static void x2apic_vector_allocation_domain(int cpu, cpumask_t *retmask) +static void x2apic_vector_allocation_domain(int cpu, struct cpumask *retmask) { - cpus_clear(*retmask); - cpu_set(cpu, *retmask); + cpumask_clear(retmask); + cpumask_set_cpu(cpu, retmask); } static void __x2apic_send_IPI_dest(unsigned int apicid, int vector, @@ -55,27 +55,28 @@ static void __x2apic_send_IPI_dest(unsigned int apicid, int vector, * at once. We have 16 cpu's in a cluster. This will minimize IPI register * writes. */ -static void x2apic_send_IPI_mask(const cpumask_t *mask, int vector) +static void x2apic_send_IPI_mask(const struct cpumask *mask, int vector) { unsigned long flags; unsigned long query_cpu; local_irq_save(flags); - for_each_cpu_mask_nr(query_cpu, *mask) + for_each_cpu(query_cpu, mask) __x2apic_send_IPI_dest( per_cpu(x86_cpu_to_logical_apicid, query_cpu), vector, APIC_DEST_LOGICAL); local_irq_restore(flags); } -static void x2apic_send_IPI_mask_allbutself(const cpumask_t *mask, int vector) +static void x2apic_send_IPI_mask_allbutself(const struct cpumask *mask, + int vector) { unsigned long flags; unsigned long query_cpu; unsigned long this_cpu = smp_processor_id(); local_irq_save(flags); - for_each_cpu_mask_nr(query_cpu, *mask) + for_each_cpu(query_cpu, mask) if (query_cpu != this_cpu) __x2apic_send_IPI_dest( per_cpu(x86_cpu_to_logical_apicid, query_cpu), @@ -100,7 +101,7 @@ static void x2apic_send_IPI_allbutself(int vector) static void x2apic_send_IPI_all(int vector) { - x2apic_send_IPI_mask(&cpu_online_map, vector); + x2apic_send_IPI_mask(cpu_online_mask, vector); } static int x2apic_apic_id_registered(void) @@ -108,7 +109,7 @@ static int x2apic_apic_id_registered(void) return 1; } -static unsigned int x2apic_cpu_mask_to_apicid(const cpumask_t *cpumask) +static unsigned int x2apic_cpu_mask_to_apicid(const struct cpumask *cpumask) { int cpu; @@ -116,7 +117,7 @@ static unsigned int x2apic_cpu_mask_to_apicid(const cpumask_t *cpumask) * We're using fixed IRQ delivery, can only return one phys APIC ID. * May as well be the first. */ - cpu = first_cpu(*cpumask); + cpu = cpumask_first(cpumask); if ((unsigned)cpu < nr_cpu_ids) return per_cpu(x86_cpu_to_logical_apicid, cpu); else diff --git a/arch/x86/kernel/genx2apic_phys.c b/arch/x86/kernel/genx2apic_phys.c index 9d0386c7e798..b255507884f2 100644 --- a/arch/x86/kernel/genx2apic_phys.c +++ b/arch/x86/kernel/genx2apic_phys.c @@ -29,15 +29,15 @@ static int x2apic_acpi_madt_oem_check(char *oem_id, char *oem_table_id) /* Start with all IRQs pointing to boot CPU. IRQ balancing will shift them. */ -static const cpumask_t *x2apic_target_cpus(void) +static const struct cpumask *x2apic_target_cpus(void) { - return &cpumask_of_cpu(0); + return cpumask_of(0); } -static void x2apic_vector_allocation_domain(int cpu, cpumask_t *retmask) +static void x2apic_vector_allocation_domain(int cpu, struct cpumask *retmask) { - cpus_clear(*retmask); - cpu_set(cpu, *retmask); + cpumask_clear(retmask); + cpumask_set_cpu(cpu, retmask); } static void __x2apic_send_IPI_dest(unsigned int apicid, int vector, @@ -53,27 +53,28 @@ static void __x2apic_send_IPI_dest(unsigned int apicid, int vector, x2apic_icr_write(cfg, apicid); } -static void x2apic_send_IPI_mask(const cpumask_t *mask, int vector) +static void x2apic_send_IPI_mask(const struct cpumask *mask, int vector) { unsigned long flags; unsigned long query_cpu; local_irq_save(flags); - for_each_cpu_mask_nr(query_cpu, *mask) { + for_each_cpu(query_cpu, mask) { __x2apic_send_IPI_dest(per_cpu(x86_cpu_to_apicid, query_cpu), vector, APIC_DEST_PHYSICAL); } local_irq_restore(flags); } -static void x2apic_send_IPI_mask_allbutself(const cpumask_t *mask, int vector) +static void x2apic_send_IPI_mask_allbutself(const struct cpumask *mask, + int vector) { unsigned long flags; unsigned long query_cpu; unsigned long this_cpu = smp_processor_id(); local_irq_save(flags); - for_each_cpu_mask_nr(query_cpu, *mask) { + for_each_cpu(query_cpu, mask) { if (query_cpu != this_cpu) __x2apic_send_IPI_dest( per_cpu(x86_cpu_to_apicid, query_cpu), @@ -99,7 +100,7 @@ static void x2apic_send_IPI_allbutself(int vector) static void x2apic_send_IPI_all(int vector) { - x2apic_send_IPI_mask(&cpu_online_map, vector); + x2apic_send_IPI_mask(cpu_online_mask, vector); } static int x2apic_apic_id_registered(void) @@ -107,7 +108,7 @@ static int x2apic_apic_id_registered(void) return 1; } -static unsigned int x2apic_cpu_mask_to_apicid(const cpumask_t *cpumask) +static unsigned int x2apic_cpu_mask_to_apicid(const struct cpumask *cpumask) { int cpu; @@ -115,7 +116,7 @@ static unsigned int x2apic_cpu_mask_to_apicid(const cpumask_t *cpumask) * We're using fixed IRQ delivery, can only return one phys APIC ID. * May as well be the first. */ - cpu = first_cpu(*cpumask); + cpu = cpumask_first(cpumask); if ((unsigned)cpu < nr_cpu_ids) return per_cpu(x86_cpu_to_apicid, cpu); else diff --git a/arch/x86/kernel/genx2apic_uv_x.c b/arch/x86/kernel/genx2apic_uv_x.c index 22596ec94c82..3984682cd849 100644 --- a/arch/x86/kernel/genx2apic_uv_x.c +++ b/arch/x86/kernel/genx2apic_uv_x.c @@ -75,15 +75,15 @@ EXPORT_SYMBOL(sn_rtc_cycles_per_second); /* Start with all IRQs pointing to boot CPU. IRQ balancing will shift them. */ -static const cpumask_t *uv_target_cpus(void) +static const struct cpumask *uv_target_cpus(void) { - return &cpumask_of_cpu(0); + return cpumask_of(0); } -static void uv_vector_allocation_domain(int cpu, cpumask_t *retmask) +static void uv_vector_allocation_domain(int cpu, struct cpumask *retmask) { - cpus_clear(*retmask); - cpu_set(cpu, *retmask); + cpumask_clear(retmask); + cpumask_set_cpu(cpu, retmask); } int uv_wakeup_secondary(int phys_apicid, unsigned int start_rip) @@ -122,20 +122,20 @@ static void uv_send_IPI_one(int cpu, int vector) uv_write_global_mmr64(pnode, UVH_IPI_INT, val); } -static void uv_send_IPI_mask(const cpumask_t *mask, int vector) +static void uv_send_IPI_mask(const struct cpumask *mask, int vector) { unsigned int cpu; - for_each_cpu_mask_nr(cpu, *mask) + for_each_cpu(cpu, mask) uv_send_IPI_one(cpu, vector); } -static void uv_send_IPI_mask_allbutself(const cpumask_t *mask, int vector) +static void uv_send_IPI_mask_allbutself(const struct cpumask *mask, int vector) { unsigned int cpu; unsigned int this_cpu = smp_processor_id(); - for_each_cpu_mask_nr(cpu, *mask) + for_each_cpu(cpu, mask) if (cpu != this_cpu) uv_send_IPI_one(cpu, vector); } @@ -152,7 +152,7 @@ static void uv_send_IPI_allbutself(int vector) static void uv_send_IPI_all(int vector) { - uv_send_IPI_mask(&cpu_online_map, vector); + uv_send_IPI_mask(cpu_online_mask, vector); } static int uv_apic_id_registered(void) @@ -164,7 +164,7 @@ static void uv_init_apic_ldr(void) { } -static unsigned int uv_cpu_mask_to_apicid(const cpumask_t *cpumask) +static unsigned int uv_cpu_mask_to_apicid(const struct cpumask *cpumask) { int cpu; @@ -172,7 +172,7 @@ static unsigned int uv_cpu_mask_to_apicid(const cpumask_t *cpumask) * We're using fixed IRQ delivery, can only return one phys APIC ID. * May as well be the first. */ - cpu = first_cpu(*cpumask); + cpu = cpumask_first(cpumask); if ((unsigned)cpu < nr_cpu_ids) return per_cpu(x86_cpu_to_apicid, cpu); else diff --git a/arch/x86/kernel/ipi.c b/arch/x86/kernel/ipi.c index 86aa50fc65a1..285bbf8831fa 100644 --- a/arch/x86/kernel/ipi.c +++ b/arch/x86/kernel/ipi.c @@ -116,18 +116,18 @@ static inline void __send_IPI_dest_field(unsigned long mask, int vector) /* * This is only used on smaller machines. */ -void send_IPI_mask_bitmask(const cpumask_t *cpumask, int vector) +void send_IPI_mask_bitmask(const struct cpumask *cpumask, int vector) { - unsigned long mask = cpus_addr(*cpumask)[0]; + unsigned long mask = cpumask_bits(cpumask)[0]; unsigned long flags; local_irq_save(flags); - WARN_ON(mask & ~cpus_addr(cpu_online_map)[0]); + WARN_ON(mask & ~cpumask_bits(cpu_online_mask)[0]); __send_IPI_dest_field(mask, vector); local_irq_restore(flags); } -void send_IPI_mask_sequence(const cpumask_t *mask, int vector) +void send_IPI_mask_sequence(const struct cpumask *mask, int vector) { unsigned long flags; unsigned int query_cpu; @@ -139,12 +139,12 @@ void send_IPI_mask_sequence(const cpumask_t *mask, int vector) */ local_irq_save(flags); - for_each_cpu_mask_nr(query_cpu, *mask) + for_each_cpu(query_cpu, mask) __send_IPI_dest_field(cpu_to_logical_apicid(query_cpu), vector); local_irq_restore(flags); } -void send_IPI_mask_allbutself(const cpumask_t *mask, int vector) +void send_IPI_mask_allbutself(const struct cpumask *mask, int vector) { unsigned long flags; unsigned int query_cpu; @@ -153,7 +153,7 @@ void send_IPI_mask_allbutself(const cpumask_t *mask, int vector) /* See Hack comment above */ local_irq_save(flags); - for_each_cpu_mask_nr(query_cpu, *mask) + for_each_cpu(query_cpu, mask) if (query_cpu != this_cpu) __send_IPI_dest_field(cpu_to_logical_apicid(query_cpu), vector); diff --git a/arch/x86/kernel/smp.c b/arch/x86/kernel/smp.c index 341df946f9a9..49ed667b06f3 100644 --- a/arch/x86/kernel/smp.c +++ b/arch/x86/kernel/smp.c @@ -118,15 +118,15 @@ static void native_smp_send_reschedule(int cpu) WARN_ON(1); return; } - send_IPI_mask(&cpumask_of_cpu(cpu), RESCHEDULE_VECTOR); + send_IPI_mask(cpumask_of(cpu), RESCHEDULE_VECTOR); } void native_send_call_func_single_ipi(int cpu) { - send_IPI_mask(&cpumask_of_cpu(cpu), CALL_FUNCTION_SINGLE_VECTOR); + send_IPI_mask(cpumask_of(cpu), CALL_FUNCTION_SINGLE_VECTOR); } -void native_send_call_func_ipi(const cpumask_t *mask) +void native_send_call_func_ipi(const struct cpumask *mask) { cpumask_t allbutself; diff --git a/arch/x86/xen/smp.c b/arch/x86/xen/smp.c index b3a95868839b..c44e2069c7c7 100644 --- a/arch/x86/xen/smp.c +++ b/arch/x86/xen/smp.c @@ -411,22 +411,23 @@ static void xen_smp_send_reschedule(int cpu) xen_send_IPI_one(cpu, XEN_RESCHEDULE_VECTOR); } -static void xen_send_IPI_mask(const cpumask_t *mask, enum ipi_vector vector) +static void xen_send_IPI_mask(const struct cpumask *mask, + enum ipi_vector vector) { unsigned cpu; - for_each_cpu_and(cpu, mask, &cpu_online_map) + for_each_cpu_and(cpu, mask, cpu_online_mask) xen_send_IPI_one(cpu, vector); } -static void xen_smp_send_call_function_ipi(const cpumask_t *mask) +static void xen_smp_send_call_function_ipi(const struct cpumask *mask) { int cpu; xen_send_IPI_mask(mask, XEN_CALL_FUNCTION_VECTOR); /* Make sure other vcpus get a chance to run if they need to. */ - for_each_cpu_mask_nr(cpu, *mask) { + for_each_cpu(cpu, mask) { if (xen_vcpu_stolen(cpu)) { HYPERVISOR_sched_op(SCHEDOP_yield, 0); break; @@ -436,7 +437,7 @@ static void xen_smp_send_call_function_ipi(const cpumask_t *mask) static void xen_smp_send_call_function_single_ipi(int cpu) { - xen_send_IPI_mask(&cpumask_of_cpu(cpu), + xen_send_IPI_mask(cpumask_of(cpu), XEN_CALL_FUNCTION_SINGLE_VECTOR); } -- cgit v1.2.3 From 78637a97b7fe1df51f40a460448df0b93d511176 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Tue, 16 Dec 2008 17:34:00 -0800 Subject: x86: Set CONFIG_NR_CPUS even on UP Impact: cleanup Signed-off-by: Rusty Russell Signed-off-by: Mike Travis --- arch/x86/Kconfig | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 1fd44352f27c..4a3f5851ec64 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -599,12 +599,12 @@ config MAXSMP If unsure, say N. config NR_CPUS - depends on SMP int "Maximum number of CPUs" if SMP && !MAXSMP range 2 512 if SMP && !MAXSMP + default "1" if !SMP default "4096" if MAXSMP - default "32" if X86_NUMAQ || X86_SUMMIT || X86_BIGSMP || X86_ES7000 - default "8" + default "32" if SMP && (X86_NUMAQ || X86_SUMMIT || X86_BIGSMP || X86_ES7000) + default "8" if SMP help This allows you to specify the maximum number of CPUs which this kernel will support. The maximum supported value is 512 and the -- cgit v1.2.3 From 168ef543a43678146e06b3911e987ac021d575b8 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Tue, 16 Dec 2008 17:34:01 -0800 Subject: x86: prepare for cpumask iterators to only go to nr_cpu_ids Impact: cleanup, futureproof In fact, all cpumask ops will only be valid (in general) for bit numbers < nr_cpu_ids. So use that instead of NR_CPUS in various places. This is always safe: no cpu number can be >= nr_cpu_ids, and nr_cpu_ids is initialized to NR_CPUS at boot. Signed-off-by: Rusty Russell Signed-off-by: Mike Travis Acked-by: Ingo Molnar --- arch/x86/kernel/apic.c | 2 +- arch/x86/mach-voyager/voyager_smp.c | 2 +- arch/x86/mm/numa_64.c | 4 ++-- arch/x86/mm/srat_64.c | 2 +- 4 files changed, 5 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/apic.c b/arch/x86/kernel/apic.c index a375791c08ca..3b630ec24935 100644 --- a/arch/x86/kernel/apic.c +++ b/arch/x86/kernel/apic.c @@ -2106,7 +2106,7 @@ __cpuinit int apic_is_clustered_box(void) bios_cpu_apicid = early_per_cpu_ptr(x86_bios_cpu_apicid); bitmap_zero(clustermap, NUM_APIC_CLUSTERS); - for (i = 0; i < NR_CPUS; i++) { + for (i = 0; i < nr_cpu_ids; i++) { /* are we being called early in kernel startup? */ if (bios_cpu_apicid) { id = bios_cpu_apicid[i]; diff --git a/arch/x86/mach-voyager/voyager_smp.c b/arch/x86/mach-voyager/voyager_smp.c index 9c990185e9f2..a5bc05492b1e 100644 --- a/arch/x86/mach-voyager/voyager_smp.c +++ b/arch/x86/mach-voyager/voyager_smp.c @@ -672,7 +672,7 @@ void __init smp_boot_cpus(void) /* loop over all the extended VIC CPUs and boot them. The * Quad CPUs must be bootstrapped by their extended VIC cpu */ - for (i = 0; i < NR_CPUS; i++) { + for (i = 0; i < nr_cpu_ids; i++) { if (i == boot_cpu_id || !cpu_isset(i, phys_cpu_present_map)) continue; do_boot_cpu(i); diff --git a/arch/x86/mm/numa_64.c b/arch/x86/mm/numa_64.c index cebcbf152d46..71a14f89f89e 100644 --- a/arch/x86/mm/numa_64.c +++ b/arch/x86/mm/numa_64.c @@ -278,7 +278,7 @@ void __init numa_init_array(void) int rr, i; rr = first_node(node_online_map); - for (i = 0; i < NR_CPUS; i++) { + for (i = 0; i < nr_cpu_ids; i++) { if (early_cpu_to_node(i) != NUMA_NO_NODE) continue; numa_set_node(i, rr); @@ -549,7 +549,7 @@ void __init initmem_init(unsigned long start_pfn, unsigned long last_pfn) memnodemap[0] = 0; node_set_online(0); node_set(0, node_possible_map); - for (i = 0; i < NR_CPUS; i++) + for (i = 0; i < nr_cpu_ids; i++) numa_set_node(i, 0); e820_register_active_regions(0, start_pfn, last_pfn); setup_node_bootmem(0, start_pfn << PAGE_SHIFT, last_pfn << PAGE_SHIFT); diff --git a/arch/x86/mm/srat_64.c b/arch/x86/mm/srat_64.c index 51c0a2fc14fe..09737c8af074 100644 --- a/arch/x86/mm/srat_64.c +++ b/arch/x86/mm/srat_64.c @@ -382,7 +382,7 @@ int __init acpi_scan_nodes(unsigned long start, unsigned long end) if (!node_online(i)) setup_node_bootmem(i, nodes[i].start, nodes[i].end); - for (i = 0; i < NR_CPUS; i++) { + for (i = 0; i < nr_cpu_ids; i++) { int node = early_cpu_to_node(i); if (node == NUMA_NO_NODE) -- cgit v1.2.3 From 1de88cd4a33fcc2fcf70cbce01688723f728675d Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Tue, 16 Dec 2008 17:34:02 -0800 Subject: x86: Use cpumask accessors code for possible/present maps. Impact: use new API Use the accessors rather than frobbing bits directly. Most of this is in arch code I haven't even compiled, but is straightforward. Signed-off-by: Rusty Russell Signed-off-by: Mike Travis --- arch/x86/kernel/apic.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/apic.c b/arch/x86/kernel/apic.c index 3b630ec24935..edda4c00e3d2 100644 --- a/arch/x86/kernel/apic.c +++ b/arch/x86/kernel/apic.c @@ -1903,8 +1903,8 @@ void __cpuinit generic_processor_info(int apicid, int version) } #endif - cpu_set(cpu, cpu_possible_map); - cpu_set(cpu, cpu_present_map); + set_cpu_possible(cpu, true); + set_cpu_present(cpu, true); } #ifdef CONFIG_X86_64 -- cgit v1.2.3 From b2bb85549134c005e997e5a7ed303bda6a1ae738 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Tue, 16 Dec 2008 17:34:03 -0800 Subject: x86: Remove cpumask games in x86/kernel/cpu/intel_cacheinfo.c Impact: remove cpumask_t from stack. We should not try to save and restore cpus_allowed on current. We can't use work_on_cpu() here, since it's in the hotplug cpu path (if anyone else tries to get the hotplug lock from a workqueue we could deadlock against them). Fortunately, we can just use smp_call_function_single() since the function can run from an interrupt. Signed-off-by: Rusty Russell Signed-off-by: Mike Travis Cc: Ingo Molnar Cc: Oleg Nesterov --- arch/x86/kernel/cpu/intel_cacheinfo.c | 41 ++++++++++++++++------------------- 1 file changed, 19 insertions(+), 22 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/intel_cacheinfo.c b/arch/x86/kernel/cpu/intel_cacheinfo.c index 43ea612d3e9d..fb7f946cb65e 100644 --- a/arch/x86/kernel/cpu/intel_cacheinfo.c +++ b/arch/x86/kernel/cpu/intel_cacheinfo.c @@ -534,31 +534,16 @@ static void __cpuinit free_cache_attributes(unsigned int cpu) per_cpu(cpuid4_info, cpu) = NULL; } -static int __cpuinit detect_cache_attributes(unsigned int cpu) +static void get_cpu_leaves(void *_retval) { - struct _cpuid4_info *this_leaf; - unsigned long j; - int retval; - cpumask_t oldmask; - - if (num_cache_leaves == 0) - return -ENOENT; - - per_cpu(cpuid4_info, cpu) = kzalloc( - sizeof(struct _cpuid4_info) * num_cache_leaves, GFP_KERNEL); - if (per_cpu(cpuid4_info, cpu) == NULL) - return -ENOMEM; - - oldmask = current->cpus_allowed; - retval = set_cpus_allowed_ptr(current, &cpumask_of_cpu(cpu)); - if (retval) - goto out; + int j, *retval = _retval, cpu = smp_processor_id(); /* Do cpuid and store the results */ for (j = 0; j < num_cache_leaves; j++) { + struct _cpuid4_info *this_leaf; this_leaf = CPUID4_INFO_IDX(cpu, j); - retval = cpuid4_cache_lookup(j, this_leaf); - if (unlikely(retval < 0)) { + *retval = cpuid4_cache_lookup(j, this_leaf); + if (unlikely(*retval < 0)) { int i; for (i = 0; i < j; i++) @@ -567,9 +552,21 @@ static int __cpuinit detect_cache_attributes(unsigned int cpu) } cache_shared_cpu_map_setup(cpu, j); } - set_cpus_allowed_ptr(current, &oldmask); +} + +static int __cpuinit detect_cache_attributes(unsigned int cpu) +{ + int retval; + + if (num_cache_leaves == 0) + return -ENOENT; + + per_cpu(cpuid4_info, cpu) = kzalloc( + sizeof(struct _cpuid4_info) * num_cache_leaves, GFP_KERNEL); + if (per_cpu(cpuid4_info, cpu) == NULL) + return -ENOMEM; -out: + smp_call_function_single(cpu, get_cpu_leaves, &retval, true); if (retval) { kfree(per_cpu(cpuid4_info, cpu)); per_cpu(cpuid4_info, cpu) = NULL; -- cgit v1.2.3 From 4cd4601d592d07b26e4b7d2bb8fcd55bbfd6cf6e Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Tue, 16 Dec 2008 17:34:04 -0800 Subject: x86: use work_on_cpu in x86/kernel/cpu/mcheck/mce_amd_64.c Impact: Remove cpumask_t's from stack. Simple transition to work_on_cpu(), rather than cpumask games. Signed-off-by: Rusty Russell Signed-off-by: Mike Travis Cc: Ingo Molnar Cc: Robert Richter Cc: jacob.shin@amd.com --- arch/x86/kernel/cpu/mcheck/mce_amd_64.c | 108 ++++++++++++++++---------------- 1 file changed, 55 insertions(+), 53 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd_64.c b/arch/x86/kernel/cpu/mcheck/mce_amd_64.c index 5eb390a4b2e9..a1de80f368f1 100644 --- a/arch/x86/kernel/cpu/mcheck/mce_amd_64.c +++ b/arch/x86/kernel/cpu/mcheck/mce_amd_64.c @@ -83,34 +83,41 @@ static DEFINE_PER_CPU(unsigned char, bank_map); /* see which banks are on */ * CPU Initialization */ +struct thresh_restart { + struct threshold_block *b; + int reset; + u16 old_limit; +}; + /* must be called with correct cpu affinity */ -static void threshold_restart_bank(struct threshold_block *b, - int reset, u16 old_limit) +static long threshold_restart_bank(void *_tr) { + struct thresh_restart *tr = _tr; u32 mci_misc_hi, mci_misc_lo; - rdmsr(b->address, mci_misc_lo, mci_misc_hi); + rdmsr(tr->b->address, mci_misc_lo, mci_misc_hi); - if (b->threshold_limit < (mci_misc_hi & THRESHOLD_MAX)) - reset = 1; /* limit cannot be lower than err count */ + if (tr->b->threshold_limit < (mci_misc_hi & THRESHOLD_MAX)) + tr->reset = 1; /* limit cannot be lower than err count */ - if (reset) { /* reset err count and overflow bit */ + if (tr->reset) { /* reset err count and overflow bit */ mci_misc_hi = (mci_misc_hi & ~(MASK_ERR_COUNT_HI | MASK_OVERFLOW_HI)) | - (THRESHOLD_MAX - b->threshold_limit); - } else if (old_limit) { /* change limit w/o reset */ + (THRESHOLD_MAX - tr->b->threshold_limit); + } else if (tr->old_limit) { /* change limit w/o reset */ int new_count = (mci_misc_hi & THRESHOLD_MAX) + - (old_limit - b->threshold_limit); + (tr->old_limit - tr->b->threshold_limit); mci_misc_hi = (mci_misc_hi & ~MASK_ERR_COUNT_HI) | (new_count & THRESHOLD_MAX); } - b->interrupt_enable ? + tr->b->interrupt_enable ? (mci_misc_hi = (mci_misc_hi & ~MASK_INT_TYPE_HI) | INT_TYPE_APIC) : (mci_misc_hi &= ~MASK_INT_TYPE_HI); mci_misc_hi |= MASK_COUNT_EN_HI; - wrmsr(b->address, mci_misc_lo, mci_misc_hi); + wrmsr(tr->b->address, mci_misc_lo, mci_misc_hi); + return 0; } /* cpu init entry point, called from mce.c with preempt off */ @@ -120,6 +127,7 @@ void __cpuinit mce_amd_feature_init(struct cpuinfo_x86 *c) unsigned int cpu = smp_processor_id(); u8 lvt_off; u32 low = 0, high = 0, address = 0; + struct thresh_restart tr; for (bank = 0; bank < NR_BANKS; ++bank) { for (block = 0; block < NR_BLOCKS; ++block) { @@ -162,7 +170,10 @@ void __cpuinit mce_amd_feature_init(struct cpuinfo_x86 *c) wrmsr(address, low, high); threshold_defaults.address = address; - threshold_restart_bank(&threshold_defaults, 0, 0); + tr.b = &threshold_defaults; + tr.reset = 0; + tr.old_limit = 0; + threshold_restart_bank(&tr); } } } @@ -251,20 +262,6 @@ struct threshold_attr { ssize_t(*store) (struct threshold_block *, const char *, size_t count); }; -static void affinity_set(unsigned int cpu, cpumask_t *oldmask, - cpumask_t *newmask) -{ - *oldmask = current->cpus_allowed; - cpus_clear(*newmask); - cpu_set(cpu, *newmask); - set_cpus_allowed_ptr(current, newmask); -} - -static void affinity_restore(const cpumask_t *oldmask) -{ - set_cpus_allowed_ptr(current, oldmask); -} - #define SHOW_FIELDS(name) \ static ssize_t show_ ## name(struct threshold_block * b, char *buf) \ { \ @@ -277,15 +274,16 @@ static ssize_t store_interrupt_enable(struct threshold_block *b, const char *buf, size_t count) { char *end; - cpumask_t oldmask, newmask; + struct thresh_restart tr; unsigned long new = simple_strtoul(buf, &end, 0); if (end == buf) return -EINVAL; b->interrupt_enable = !!new; - affinity_set(b->cpu, &oldmask, &newmask); - threshold_restart_bank(b, 0, 0); - affinity_restore(&oldmask); + tr.b = b; + tr.reset = 0; + tr.old_limit = 0; + work_on_cpu(b->cpu, threshold_restart_bank, &tr); return end - buf; } @@ -294,8 +292,7 @@ static ssize_t store_threshold_limit(struct threshold_block *b, const char *buf, size_t count) { char *end; - cpumask_t oldmask, newmask; - u16 old; + struct thresh_restart tr; unsigned long new = simple_strtoul(buf, &end, 0); if (end == buf) return -EINVAL; @@ -303,34 +300,36 @@ static ssize_t store_threshold_limit(struct threshold_block *b, new = THRESHOLD_MAX; if (new < 1) new = 1; - old = b->threshold_limit; + tr.old_limit = b->threshold_limit; b->threshold_limit = new; + tr.b = b; + tr.reset = 0; - affinity_set(b->cpu, &oldmask, &newmask); - threshold_restart_bank(b, 0, old); - affinity_restore(&oldmask); + work_on_cpu(b->cpu, threshold_restart_bank, &tr); return end - buf; } -static ssize_t show_error_count(struct threshold_block *b, char *buf) +static long local_error_count(void *_b) { - u32 high, low; - cpumask_t oldmask, newmask; - affinity_set(b->cpu, &oldmask, &newmask); + struct threshold_block *b = _b; + u32 low, high; + rdmsr(b->address, low, high); - affinity_restore(&oldmask); - return sprintf(buf, "%x\n", - (high & 0xFFF) - (THRESHOLD_MAX - b->threshold_limit)); + return (high & 0xFFF) - (THRESHOLD_MAX - b->threshold_limit); +} + +static ssize_t show_error_count(struct threshold_block *b, char *buf) +{ + return sprintf(buf, "%lx\n", work_on_cpu(b->cpu, local_error_count, b)); } static ssize_t store_error_count(struct threshold_block *b, const char *buf, size_t count) { - cpumask_t oldmask, newmask; - affinity_set(b->cpu, &oldmask, &newmask); - threshold_restart_bank(b, 1, 0); - affinity_restore(&oldmask); + struct thresh_restart tr = { .b = b, .reset = 1, .old_limit = 0 }; + + work_on_cpu(b->cpu, threshold_restart_bank, &tr); return 1; } @@ -463,12 +462,19 @@ out_free: return err; } +static long local_allocate_threshold_blocks(void *_bank) +{ + unsigned int *bank = _bank; + + return allocate_threshold_blocks(smp_processor_id(), *bank, 0, + MSR_IA32_MC0_MISC + *bank * 4); +} + /* symlinks sibling shared banks to first core. first core owns dir/files. */ static __cpuinit int threshold_create_bank(unsigned int cpu, unsigned int bank) { int i, err = 0; struct threshold_bank *b = NULL; - cpumask_t oldmask, newmask; char name[32]; sprintf(name, "threshold_bank%i", bank); @@ -519,11 +525,7 @@ static __cpuinit int threshold_create_bank(unsigned int cpu, unsigned int bank) per_cpu(threshold_banks, cpu)[bank] = b; - affinity_set(cpu, &oldmask, &newmask); - err = allocate_threshold_blocks(cpu, bank, 0, - MSR_IA32_MC0_MISC + bank * 4); - affinity_restore(&oldmask); - + err = work_on_cpu(cpu, local_allocate_threshold_blocks, &bank); if (err) goto out_free; -- cgit v1.2.3 From e4d98207ea3f3d15eb664282df16d18c4ac86f80 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Tue, 16 Dec 2008 17:34:05 -0800 Subject: x86: xen: use smp_call_function_many() Impact: use new API, remove cpumask from stack. Change smp_call_function_mask() callers to smp_call_function_many(). This removes a cpumask from the stack, and falls back should allocating the cpumask var fail (only possible with CONFIG_CPUMASKS_OFFSTACK). Signed-off-by: Rusty Russell Signed-off-by: Mike Travis Cc: jeremy@xensource.com --- arch/x86/xen/mmu.c | 20 +++++++++++++++----- 1 file changed, 15 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c index 636ef4caa52d..e59e53b11e2b 100644 --- a/arch/x86/xen/mmu.c +++ b/arch/x86/xen/mmu.c @@ -1079,7 +1079,7 @@ static void drop_other_mm_ref(void *info) static void xen_drop_mm_ref(struct mm_struct *mm) { - cpumask_t mask; + cpumask_var_t mask; unsigned cpu; if (current->active_mm == mm) { @@ -1091,7 +1091,16 @@ static void xen_drop_mm_ref(struct mm_struct *mm) } /* Get the "official" set of cpus referring to our pagetable. */ - mask = mm->cpu_vm_mask; + if (!alloc_cpumask_var(&mask, GFP_ATOMIC)) { + for_each_online_cpu(cpu) { + if (!cpumask_test_cpu(cpu, &mm->cpu_vm_mask) + && per_cpu(xen_current_cr3, cpu) != __pa(mm->pgd)) + continue; + smp_call_function_single(cpu, drop_other_mm_ref, mm, 1); + } + return; + } + cpumask_copy(mask, &mm->cpu_vm_mask); /* It's possible that a vcpu may have a stale reference to our cr3, because its in lazy mode, and it hasn't yet flushed @@ -1100,11 +1109,12 @@ static void xen_drop_mm_ref(struct mm_struct *mm) if needed. */ for_each_online_cpu(cpu) { if (per_cpu(xen_current_cr3, cpu) == __pa(mm->pgd)) - cpu_set(cpu, mask); + cpumask_set_cpu(cpu, mask); } - if (!cpus_empty(mask)) - smp_call_function_mask(mask, drop_other_mm_ref, mm, 1); + if (!cpumask_empty(mask)) + smp_call_function_many(mask, drop_other_mm_ref, mm, 1); + free_cpumask_var(mask); } #else static void xen_drop_mm_ref(struct mm_struct *mm) -- cgit v1.2.3 From 83b19597f793fd5f91533bda0dc2eb3d21936798 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Tue, 16 Dec 2008 17:34:06 -0800 Subject: x86: Introduce topology_core_cpumask()/topology_thread_cpumask() Impact: new API The old topology_core_siblings() and topology_thread_siblings() return a cpumask_t; these new ones return a (const) struct cpumask *. Signed-off-by: Rusty Russell Signed-off-by: Mike Travis --- arch/x86/include/asm/topology.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/topology.h b/arch/x86/include/asm/topology.h index ff386ff50ed7..79e31e9dcdda 100644 --- a/arch/x86/include/asm/topology.h +++ b/arch/x86/include/asm/topology.h @@ -226,6 +226,8 @@ extern cpumask_t cpu_coregroup_map(int cpu); #define topology_core_id(cpu) (cpu_data(cpu).cpu_core_id) #define topology_core_siblings(cpu) (per_cpu(cpu_core_map, cpu)) #define topology_thread_siblings(cpu) (per_cpu(cpu_sibling_map, cpu)) +#define topology_core_cpumask(cpu) (&per_cpu(cpu_core_map, cpu)) +#define topology_thread_cpumask(cpu) (&per_cpu(cpu_sibling_map, cpu)) /* indicates that pointers to the topology cpumask_t maps are valid */ #define arch_provides_topology_pointers yes -- cgit v1.2.3 From d733e00d7c10cc68333fbb88108bb15bb044f61b Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Wed, 17 Dec 2008 13:35:51 +0100 Subject: x86: update io_apic.c to the new cpumask code Impact: build fix The sparseirq tree crossed with the cpumask changes, fix the fallout. Signed-off-by: Ingo Molnar --- arch/x86/kernel/io_apic.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index 6bd51ce3ce32..58938cc4b7d3 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -347,13 +347,14 @@ void arch_free_chip_data(struct irq_desc *old_desc, struct irq_desc *desc) } } -static void set_extra_move_desc(struct irq_desc *desc, cpumask_t mask) +static void +set_extra_move_desc(struct irq_desc *desc, const struct cpumask *mask) { struct irq_cfg *cfg = desc->chip_data; if (!cfg->move_in_progress) { /* it means that domain is not changed */ - if (!cpus_intersects(desc->affinity, mask)) + if (!cpumask_intersects(&desc->affinity, mask)) cfg->move_desc_pending = 1; } } -- cgit v1.2.3 From cf558d25e5c9f70fa0279c9b7b8b4aed7cae9bd4 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Wed, 17 Dec 2008 15:06:01 +0100 Subject: AMD IOMMU: set cmd buffer pointers to zero manually Impact: set cmd buffer head and tail pointers to zero in case nobody else did Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu_init.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu_init.c b/arch/x86/kernel/amd_iommu_init.c index c90a15eba5c5..c6cc22815d35 100644 --- a/arch/x86/kernel/amd_iommu_init.c +++ b/arch/x86/kernel/amd_iommu_init.c @@ -427,6 +427,10 @@ static u8 * __init alloc_command_buffer(struct amd_iommu *iommu) memcpy_toio(iommu->mmio_base + MMIO_CMD_BUF_OFFSET, &entry, sizeof(entry)); + /* set head and tail to zero manually */ + writel(0x00, iommu->mmio_base + MMIO_CMD_HEAD_OFFSET); + writel(0x00, iommu->mmio_base + MMIO_CMD_TAIL_OFFSET); + iommu_feature_enable(iommu, CONTROL_CMDBUF_EN); return cmd_buf; -- cgit v1.2.3 From 84df81759590ad16b0024cf46b3423cca76b2e07 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Wed, 17 Dec 2008 16:36:44 +0100 Subject: AMD IOMMU: panic if completion wait loop fails Impact: prevents data corruption after a failed completion wait loop Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index a7b6dec6fc3f..0a60d60ed036 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -235,8 +235,9 @@ static int iommu_completion_wait(struct amd_iommu *iommu) status &= ~MMIO_STATUS_COM_WAIT_INT_MASK; writel(status, iommu->mmio_base + MMIO_STATUS_OFFSET); - if (unlikely((i == EXIT_LOOP_COUNT) && printk_ratelimit())) - printk(KERN_WARNING "AMD IOMMU: Completion wait loop failed\n"); + if (unlikely(i == EXIT_LOOP_COUNT)) + panic("AMD IOMMU: Completion wait loop failed\n"); + out: spin_unlock_irqrestore(&iommu->lock, flags); -- cgit v1.2.3 From 8ce7996009bab7b2d23e7af7ad831fed7eb6faa1 Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Tue, 16 Dec 2008 12:17:35 -0800 Subject: x86: add swiotlb allocation functions Add x86-specific swiotlb allocation functions. These are purely default for the moment. Signed-off-by: Jeremy Fitzhardinge Signed-off-by: Ingo Molnar --- arch/x86/kernel/pci-swiotlb_64.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/pci-swiotlb_64.c b/arch/x86/kernel/pci-swiotlb_64.c index 3c539d111abb..f47a097a135b 100644 --- a/arch/x86/kernel/pci-swiotlb_64.c +++ b/arch/x86/kernel/pci-swiotlb_64.c @@ -3,6 +3,8 @@ #include #include #include +#include +#include #include #include @@ -11,6 +13,16 @@ int swiotlb __read_mostly; +void *swiotlb_alloc_boot(size_t size, unsigned long nslabs) +{ + return alloc_bootmem_low_pages(size); +} + +void *swiotlb_alloc(unsigned order, unsigned long nslabs) +{ + return (void *)__get_free_pages(GFP_DMA | __GFP_NOWARN, order); +} + static dma_addr_t swiotlb_map_single_phys(struct device *hwdev, phys_addr_t paddr, size_t size, int direction) -- cgit v1.2.3 From cfb80c9eae8c7ed8f2ee81090062d15ead51cbe8 Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Tue, 16 Dec 2008 12:17:36 -0800 Subject: x86: unify pci iommu setup and allow swiotlb to compile for 32 bit swiotlb on 32 bit will be used by Xen domain 0 support. Signed-off-by: Ian Campbell Signed-off-by: Jeremy Fitzhardinge Signed-off-by: Ingo Molnar --- arch/x86/include/asm/dma-mapping.h | 2 +- arch/x86/include/asm/pci.h | 2 ++ arch/x86/include/asm/pci_64.h | 1 - arch/x86/kernel/Makefile | 3 ++- arch/x86/kernel/pci-dma.c | 6 ++++-- arch/x86/kernel/pci-swiotlb_64.c | 2 ++ arch/x86/mm/init_32.c | 3 +++ 7 files changed, 14 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/dma-mapping.h b/arch/x86/include/asm/dma-mapping.h index 097794ff6b79..3b43a65894c4 100644 --- a/arch/x86/include/asm/dma-mapping.h +++ b/arch/x86/include/asm/dma-mapping.h @@ -65,7 +65,7 @@ static inline struct dma_mapping_ops *get_dma_ops(struct device *dev) return dma_ops; else return dev->archdata.dma_ops; -#endif /* _ASM_X86_DMA_MAPPING_H */ +#endif } /* Make sure we keep the same behaviour */ diff --git a/arch/x86/include/asm/pci.h b/arch/x86/include/asm/pci.h index 875b38edf193..50ac542c9382 100644 --- a/arch/x86/include/asm/pci.h +++ b/arch/x86/include/asm/pci.h @@ -82,6 +82,8 @@ static inline void pci_dma_burst_advice(struct pci_dev *pdev, static inline void early_quirks(void) { } #endif +extern void pci_iommu_alloc(void); + #endif /* __KERNEL__ */ #ifdef CONFIG_X86_32 diff --git a/arch/x86/include/asm/pci_64.h b/arch/x86/include/asm/pci_64.h index d02d936840a3..4da207982777 100644 --- a/arch/x86/include/asm/pci_64.h +++ b/arch/x86/include/asm/pci_64.h @@ -23,7 +23,6 @@ extern int (*pci_config_write)(int seg, int bus, int dev, int fn, int reg, int len, u32 value); extern void dma32_reserve_bootmem(void); -extern void pci_iommu_alloc(void); /* The PCI address space does equal the physical memory * address space. The networking and block device layers use diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index b62a7667828e..a9c656f2d661 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -105,6 +105,8 @@ microcode-$(CONFIG_MICROCODE_INTEL) += microcode_intel.o microcode-$(CONFIG_MICROCODE_AMD) += microcode_amd.o obj-$(CONFIG_MICROCODE) += microcode.o +obj-$(CONFIG_SWIOTLB) += pci-swiotlb_64.o # NB rename without _64 + ### # 64 bit specific files ifeq ($(CONFIG_X86_64),y) @@ -118,7 +120,6 @@ ifeq ($(CONFIG_X86_64),y) obj-$(CONFIG_GART_IOMMU) += pci-gart_64.o aperture_64.o obj-$(CONFIG_CALGARY_IOMMU) += pci-calgary_64.o tce_64.o obj-$(CONFIG_AMD_IOMMU) += amd_iommu_init.o amd_iommu.o - obj-$(CONFIG_SWIOTLB) += pci-swiotlb_64.o obj-$(CONFIG_PCI_MMCONFIG) += mmconf-fam10h_64.o endif diff --git a/arch/x86/kernel/pci-dma.c b/arch/x86/kernel/pci-dma.c index e150ad4f0ccc..00e07447a5bd 100644 --- a/arch/x86/kernel/pci-dma.c +++ b/arch/x86/kernel/pci-dma.c @@ -105,11 +105,15 @@ static void __init dma32_free_bootmem(void) dma32_bootmem_ptr = NULL; dma32_bootmem_size = 0; } +#endif void __init pci_iommu_alloc(void) { +#ifdef CONFIG_X86_64 /* free the range so iommu could get some range less than 4G */ dma32_free_bootmem(); +#endif + /* * The order of these functions is important for * fall-back/fail-over reasons @@ -125,8 +129,6 @@ void __init pci_iommu_alloc(void) pci_swiotlb_init(); } -#endif - void *dma_generic_alloc_coherent(struct device *dev, size_t size, dma_addr_t *dma_addr, gfp_t flag) { diff --git a/arch/x86/kernel/pci-swiotlb_64.c b/arch/x86/kernel/pci-swiotlb_64.c index f47a097a135b..a991afea6700 100644 --- a/arch/x86/kernel/pci-swiotlb_64.c +++ b/arch/x86/kernel/pci-swiotlb_64.c @@ -62,8 +62,10 @@ struct dma_mapping_ops swiotlb_dma_ops = { void __init pci_swiotlb_init(void) { /* don't initialize swiotlb if iommu=off (no_iommu=1) */ +#ifdef CONFIG_X86_64 if (!iommu_detected && !no_iommu && max_pfn > MAX_DMA32_PFN) swiotlb = 1; +#endif if (swiotlb_force) swiotlb = 1; if (swiotlb) { diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c index c483f4242079..2b4b14fc0c04 100644 --- a/arch/x86/mm/init_32.c +++ b/arch/x86/mm/init_32.c @@ -21,6 +21,7 @@ #include #include #include +#include #include #include #include @@ -971,6 +972,8 @@ void __init mem_init(void) start_periodic_check_for_corruption(); + pci_iommu_alloc(); + #ifdef CONFIG_FLATMEM BUG_ON(!mem_map); #endif -- cgit v1.2.3 From 1d32251e846ccbcf9d2da041dffd1199f94b2a3b Mon Sep 17 00:00:00 2001 From: Ian Campbell Date: Tue, 16 Dec 2008 12:17:37 -0800 Subject: x86/swiotlb: add default phys<->bus conversion Xen will override these later on. Signed-off-by: Ian Campbell Signed-off-by: Jeremy Fitzhardinge Signed-off-by: Ingo Molnar --- arch/x86/kernel/pci-swiotlb_64.c | 10 ++++++++++ 1 file changed, 10 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/pci-swiotlb_64.c b/arch/x86/kernel/pci-swiotlb_64.c index a991afea6700..93a8371f2c22 100644 --- a/arch/x86/kernel/pci-swiotlb_64.c +++ b/arch/x86/kernel/pci-swiotlb_64.c @@ -23,6 +23,16 @@ void *swiotlb_alloc(unsigned order, unsigned long nslabs) return (void *)__get_free_pages(GFP_DMA | __GFP_NOWARN, order); } +dma_addr_t swiotlb_phys_to_bus(phys_addr_t paddr) +{ + return paddr; +} + +phys_addr_t swiotlb_bus_to_phys(dma_addr_t baddr) +{ + return baddr; +} + static dma_addr_t swiotlb_map_single_phys(struct device *hwdev, phys_addr_t paddr, size_t size, int direction) -- cgit v1.2.3 From a08636690d06b2e36cfb4c2b3ee133a81c47e1e0 Mon Sep 17 00:00:00 2001 From: Ian Campbell Date: Tue, 16 Dec 2008 12:17:38 -0800 Subject: x86/swiotlb: add default swiotlb_arch_range_needs_mapping Xen will override these later on. Signed-off-by: Ian Campbell Signed-off-by: Jeremy Fitzhardinge Signed-off-by: Ingo Molnar --- arch/x86/kernel/pci-swiotlb_64.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/pci-swiotlb_64.c b/arch/x86/kernel/pci-swiotlb_64.c index 93a8371f2c22..242c3440687f 100644 --- a/arch/x86/kernel/pci-swiotlb_64.c +++ b/arch/x86/kernel/pci-swiotlb_64.c @@ -33,6 +33,11 @@ phys_addr_t swiotlb_bus_to_phys(dma_addr_t baddr) return baddr; } +int __weak swiotlb_arch_range_needs_mapping(void *ptr, size_t size) +{ + return 0; +} + static dma_addr_t swiotlb_map_single_phys(struct device *hwdev, phys_addr_t paddr, size_t size, int direction) -- cgit v1.2.3 From f5223763a664da16771211f9d293e18cb242b246 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Wed, 17 Dec 2008 18:47:17 -0800 Subject: x86: signal: move ia32 func declarations into arch/x86/kernel/signal.c Impact: cleanup Move declarations of ia32_setup_rt_frame() and ia32_setup_frame() into arch/x86/kernel/signal.c. This is for future use of sigframe.h. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/kernel/sigframe.h | 5 ----- arch/x86/kernel/signal.c | 5 +++++ 2 files changed, 5 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/sigframe.h b/arch/x86/kernel/sigframe.h index cc673aa55ce4..6dd7e2b70a4b 100644 --- a/arch/x86/kernel/sigframe.h +++ b/arch/x86/kernel/sigframe.h @@ -34,9 +34,4 @@ struct rt_sigframe { struct siginfo info; /* fp state follows here */ }; - -int ia32_setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, - sigset_t *set, struct pt_regs *regs); -int ia32_setup_frame(int sig, struct k_sigaction *ka, - sigset_t *set, struct pt_regs *regs); #endif diff --git a/arch/x86/kernel/signal.c b/arch/x86/kernel/signal.c index 2725a294d734..848c2d64a289 100644 --- a/arch/x86/kernel/signal.c +++ b/arch/x86/kernel/signal.c @@ -671,6 +671,11 @@ static int signr_convert(int sig) #define is_ia32 0 #endif /* CONFIG_IA32_EMULATION */ +int ia32_setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, + sigset_t *set, struct pt_regs *regs); +int ia32_setup_frame(int sig, struct k_sigaction *ka, + sigset_t *set, struct pt_regs *regs); + #endif /* CONFIG_X86_32 */ static int -- cgit v1.2.3 From a5c56eb36f999ae0ecac278e51fd1cf8feb16c2f Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Wed, 17 Dec 2008 18:49:55 -0800 Subject: x86: signal: rename sigframe and rt_sigframe on 32-bit Impact: cleanup, prepare to move sigframe.h On 32-bit, rename struct sigrame to struct sigframe_ia32, struct rt_sigframe to struct rt_sigframe_ia32 and several structures. And add helper macros to access the above data in arch/x86/kernel/signal.c. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/kernel/sigframe.h | 28 +++++++++++++++++----------- 1 file changed, 17 insertions(+), 11 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/sigframe.h b/arch/x86/kernel/sigframe.h index 6dd7e2b70a4b..6718ed04b05b 100644 --- a/arch/x86/kernel/sigframe.h +++ b/arch/x86/kernel/sigframe.h @@ -1,8 +1,14 @@ #ifdef CONFIG_X86_32 -struct sigframe { - char __user *pretcode; +#define sigframe_ia32 sigframe +#define rt_sigframe_ia32 rt_sigframe +#define sigcontext_ia32 sigcontext +#define _fpstate_ia32 _fpstate +#define ucontext_ia32 ucontext + +struct sigframe_ia32 { + u32 pretcode; int sig; - struct sigcontext sc; + struct sigcontext_ia32 sc; /* * fpstate is unused. fpstate is moved/allocated after * retcode[] below. This movement allows to have the FP state and the @@ -11,27 +17,27 @@ struct sigframe { * the offset of extramask[] in the sigframe and thus prevent any * legacy application accessing/modifying it. */ - struct _fpstate fpstate_unused; + struct _fpstate_ia32 fpstate_unused; unsigned long extramask[_NSIG_WORDS-1]; char retcode[8]; /* fp state follows here */ }; -struct rt_sigframe { - char __user *pretcode; +struct rt_sigframe_ia32 { + u32 pretcode; int sig; - struct siginfo __user *pinfo; - void __user *puc; + u32 pinfo; + u32 puc; struct siginfo info; - struct ucontext uc; + struct ucontext_ia32 uc; char retcode[8]; /* fp state follows here */ }; -#else +#else /* !CONFIG_X86_32 */ struct rt_sigframe { char __user *pretcode; struct ucontext uc; struct siginfo info; /* fp state follows here */ }; -#endif +#endif /* CONFIG_X86_32 */ -- cgit v1.2.3 From 41af86fad3c40646b9748279e3862781e937a5d2 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Wed, 17 Dec 2008 18:50:32 -0800 Subject: x86: signal: move sigframe.h to arch/x86/include/asm Impact: cleanup, move header file Move arch/x86/kernel/sigframe.h to arch/x86/include/asm/sigframe.h. It will be used in arch/x86/ia32/ia32_signal.c. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/include/asm/sigframe.h | 43 ++++++++++++++++++++++++++++++++++++++++ arch/x86/kernel/asm-offsets_32.c | 2 +- arch/x86/kernel/sigframe.h | 43 ---------------------------------------- arch/x86/kernel/signal.c | 2 +- 4 files changed, 45 insertions(+), 45 deletions(-) create mode 100644 arch/x86/include/asm/sigframe.h delete mode 100644 arch/x86/kernel/sigframe.h (limited to 'arch/x86') diff --git a/arch/x86/include/asm/sigframe.h b/arch/x86/include/asm/sigframe.h new file mode 100644 index 000000000000..6718ed04b05b --- /dev/null +++ b/arch/x86/include/asm/sigframe.h @@ -0,0 +1,43 @@ +#ifdef CONFIG_X86_32 +#define sigframe_ia32 sigframe +#define rt_sigframe_ia32 rt_sigframe +#define sigcontext_ia32 sigcontext +#define _fpstate_ia32 _fpstate +#define ucontext_ia32 ucontext + +struct sigframe_ia32 { + u32 pretcode; + int sig; + struct sigcontext_ia32 sc; + /* + * fpstate is unused. fpstate is moved/allocated after + * retcode[] below. This movement allows to have the FP state and the + * future state extensions (xsave) stay together. + * And at the same time retaining the unused fpstate, prevents changing + * the offset of extramask[] in the sigframe and thus prevent any + * legacy application accessing/modifying it. + */ + struct _fpstate_ia32 fpstate_unused; + unsigned long extramask[_NSIG_WORDS-1]; + char retcode[8]; + /* fp state follows here */ +}; + +struct rt_sigframe_ia32 { + u32 pretcode; + int sig; + u32 pinfo; + u32 puc; + struct siginfo info; + struct ucontext_ia32 uc; + char retcode[8]; + /* fp state follows here */ +}; +#else /* !CONFIG_X86_32 */ +struct rt_sigframe { + char __user *pretcode; + struct ucontext uc; + struct siginfo info; + /* fp state follows here */ +}; +#endif /* CONFIG_X86_32 */ diff --git a/arch/x86/kernel/asm-offsets_32.c b/arch/x86/kernel/asm-offsets_32.c index 6649d09ad88f..ee4df08feee6 100644 --- a/arch/x86/kernel/asm-offsets_32.c +++ b/arch/x86/kernel/asm-offsets_32.c @@ -11,7 +11,7 @@ #include #include #include -#include "sigframe.h" +#include #include #include #include diff --git a/arch/x86/kernel/sigframe.h b/arch/x86/kernel/sigframe.h deleted file mode 100644 index 6718ed04b05b..000000000000 --- a/arch/x86/kernel/sigframe.h +++ /dev/null @@ -1,43 +0,0 @@ -#ifdef CONFIG_X86_32 -#define sigframe_ia32 sigframe -#define rt_sigframe_ia32 rt_sigframe -#define sigcontext_ia32 sigcontext -#define _fpstate_ia32 _fpstate -#define ucontext_ia32 ucontext - -struct sigframe_ia32 { - u32 pretcode; - int sig; - struct sigcontext_ia32 sc; - /* - * fpstate is unused. fpstate is moved/allocated after - * retcode[] below. This movement allows to have the FP state and the - * future state extensions (xsave) stay together. - * And at the same time retaining the unused fpstate, prevents changing - * the offset of extramask[] in the sigframe and thus prevent any - * legacy application accessing/modifying it. - */ - struct _fpstate_ia32 fpstate_unused; - unsigned long extramask[_NSIG_WORDS-1]; - char retcode[8]; - /* fp state follows here */ -}; - -struct rt_sigframe_ia32 { - u32 pretcode; - int sig; - u32 pinfo; - u32 puc; - struct siginfo info; - struct ucontext_ia32 uc; - char retcode[8]; - /* fp state follows here */ -}; -#else /* !CONFIG_X86_32 */ -struct rt_sigframe { - char __user *pretcode; - struct ucontext uc; - struct siginfo info; - /* fp state follows here */ -}; -#endif /* CONFIG_X86_32 */ diff --git a/arch/x86/kernel/signal.c b/arch/x86/kernel/signal.c index 848c2d64a289..89bb7668041d 100644 --- a/arch/x86/kernel/signal.c +++ b/arch/x86/kernel/signal.c @@ -35,7 +35,7 @@ #include #include -#include "sigframe.h" +#include #define _BLOCKABLE (~(sigmask(SIGKILL) | sigmask(SIGSTOP))) -- cgit v1.2.3 From c85c2ff877c9305f801f7d5b9e6382cb05a03d45 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Wed, 17 Dec 2008 18:51:08 -0800 Subject: x86: signal: prepare to include from ia32_signal.c Impact: cleanup, prepare to use from ia32_signal.c Make struct sigframe_ia32 and rt_sigframe_ia32 visible to ia32_signal.c. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/include/asm/sigframe.h | 22 ++++++++++++++++++++-- 1 file changed, 20 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/sigframe.h b/arch/x86/include/asm/sigframe.h index 6718ed04b05b..491a0878c3aa 100644 --- a/arch/x86/include/asm/sigframe.h +++ b/arch/x86/include/asm/sigframe.h @@ -4,7 +4,15 @@ #define sigcontext_ia32 sigcontext #define _fpstate_ia32 _fpstate #define ucontext_ia32 ucontext +#else /* !CONFIG_X86_32 */ + +#ifdef CONFIG_IA32_EMULATION +#include +#endif /* CONFIG_IA32_EMULATION */ + +#endif /* CONFIG_X86_32 */ +#if defined(CONFIG_X86_32) || defined(CONFIG_IA32_EMULATION) struct sigframe_ia32 { u32 pretcode; int sig; @@ -18,7 +26,11 @@ struct sigframe_ia32 { * legacy application accessing/modifying it. */ struct _fpstate_ia32 fpstate_unused; +#ifdef CONFIG_IA32_EMULATION + unsigned int extramask[_COMPAT_NSIG_WORDS-1]; +#else /* !CONFIG_IA32_EMULATION */ unsigned long extramask[_NSIG_WORDS-1]; +#endif /* CONFIG_IA32_EMULATION */ char retcode[8]; /* fp state follows here */ }; @@ -28,16 +40,22 @@ struct rt_sigframe_ia32 { int sig; u32 pinfo; u32 puc; +#ifdef CONFIG_IA32_EMULATION + compat_siginfo_t info; +#else /* !CONFIG_IA32_EMULATION */ struct siginfo info; +#endif /* CONFIG_IA32_EMULATION */ struct ucontext_ia32 uc; char retcode[8]; /* fp state follows here */ }; -#else /* !CONFIG_X86_32 */ +#endif /* defined(CONFIG_X86_32) || defined(CONFIG_IA32_EMULATION) */ + +#ifdef CONFIG_X86_64 struct rt_sigframe { char __user *pretcode; struct ucontext uc; struct siginfo info; /* fp state follows here */ }; -#endif /* CONFIG_X86_32 */ +#endif /* CONFIG_X86_64 */ -- cgit v1.2.3 From 3b0d29ee1c73b6b90bfddd10f7b8e86632b6b694 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Wed, 17 Dec 2008 18:51:46 -0800 Subject: x86: ia32_signal: rename struct sigframe and rt_sigframe Impact: cleanup, prepare to include sigframe.h Rename struct sigframe to struct sigframe_ia32 and struct rt_sigframe to struct rt_sigframe_ia32. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/ia32/ia32_signal.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/ia32/ia32_signal.c b/arch/x86/ia32/ia32_signal.c index 9c99c429a20d..334a4aa2e75b 100644 --- a/arch/x86/ia32/ia32_signal.c +++ b/arch/x86/ia32/ia32_signal.c @@ -174,7 +174,7 @@ asmlinkage long sys32_sigaltstack(const stack_ia32_t __user *uss_ptr, * Do a signal return; undo the signal stack. */ -struct sigframe +struct sigframe_ia32 { u32 pretcode; int sig; @@ -185,7 +185,7 @@ struct sigframe /* fp state follows here */ }; -struct rt_sigframe +struct rt_sigframe_ia32 { u32 pretcode; int sig; @@ -271,7 +271,7 @@ static int ia32_restore_sigcontext(struct pt_regs *regs, asmlinkage long sys32_sigreturn(struct pt_regs *regs) { - struct sigframe __user *frame = (struct sigframe __user *)(regs->sp-8); + struct sigframe_ia32 __user *frame = (struct sigframe_ia32 __user *)(regs->sp-8); sigset_t set; unsigned int ax; @@ -301,12 +301,12 @@ badframe: asmlinkage long sys32_rt_sigreturn(struct pt_regs *regs) { - struct rt_sigframe __user *frame; + struct rt_sigframe_ia32 __user *frame; sigset_t set; unsigned int ax; struct pt_regs tregs; - frame = (struct rt_sigframe __user *)(regs->sp - 4); + frame = (struct rt_sigframe_ia32 __user *)(regs->sp - 4); if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) goto badframe; @@ -418,7 +418,7 @@ static void __user *get_sigframe(struct k_sigaction *ka, struct pt_regs *regs, int ia32_setup_frame(int sig, struct k_sigaction *ka, compat_sigset_t *set, struct pt_regs *regs) { - struct sigframe __user *frame; + struct sigframe_ia32 __user *frame; void __user *restorer; int err = 0; void __user *fpstate = NULL; @@ -497,7 +497,7 @@ int ia32_setup_frame(int sig, struct k_sigaction *ka, int ia32_setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, compat_sigset_t *set, struct pt_regs *regs) { - struct rt_sigframe __user *frame; + struct rt_sigframe_ia32 __user *frame; void __user *restorer; int err = 0; void __user *fpstate = NULL; -- cgit v1.2.3 From d98f9d84422c393103dc7569dc8444bac628f7ac Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Wed, 17 Dec 2008 18:52:45 -0800 Subject: x86: ia32_signal: use sigframe.h Impact: cleanup Use arch/x86/include/asm/sigframe.h instead of defining redundant structures. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/ia32/ia32_signal.c | 26 ++------------------------ 1 file changed, 2 insertions(+), 24 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/ia32/ia32_signal.c b/arch/x86/ia32/ia32_signal.c index 334a4aa2e75b..3b3878a63bc2 100644 --- a/arch/x86/ia32/ia32_signal.c +++ b/arch/x86/ia32/ia32_signal.c @@ -32,6 +32,8 @@ #include #include +#include + #define DEBUG_SIG 0 #define _BLOCKABLE (~(sigmask(SIGKILL) | sigmask(SIGSTOP))) @@ -173,30 +175,6 @@ asmlinkage long sys32_sigaltstack(const stack_ia32_t __user *uss_ptr, /* * Do a signal return; undo the signal stack. */ - -struct sigframe_ia32 -{ - u32 pretcode; - int sig; - struct sigcontext_ia32 sc; - struct _fpstate_ia32 fpstate_unused; /* look at kernel/sigframe.h */ - unsigned int extramask[_COMPAT_NSIG_WORDS-1]; - char retcode[8]; - /* fp state follows here */ -}; - -struct rt_sigframe_ia32 -{ - u32 pretcode; - int sig; - u32 pinfo; - u32 puc; - compat_siginfo_t info; - struct ucontext_ia32 uc; - char retcode[8]; - /* fp state follows here */ -}; - #define COPY(x) { \ err |= __get_user(regs->x, &sc->x); \ } -- cgit v1.2.3 From a775a38b1353161a6d7af86b667d6523c12c1a37 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Wed, 17 Dec 2008 15:21:39 -0800 Subject: x86: fix cpu_mask_to_apicid_and to include cpu_online_mask Impact: fix potential APIC crash In determining the destination apicid, there are usually three cpumasks that are considered: the incoming cpumask arg, cfg->domain and the cpu_online_mask. Since we are just introducing the cpu_mask_to_apicid_and function, make sure it includes the cpu_online_mask in it's evaluation. [Added with this patch.] There are two io_apic.c functions that did not previously use the cpu_online_mask: setup_IO_APIC_irq and msi_compose_msg. Both of these simply used cpu_mask_to_apicid(cfg->domain & TARGET_CPUS), and all but one arch (NUMAQ[*]) returns only online cpus in the TARGET_CPUS mask, so the behavior is identical for all cases. [*: NUMAQ bug?] Note that alloc_cpumask_var is only used for the 32-bit cases where it's highly likely that the cpumask set size will be small and therefore CPUMASK_OFFSTACK=n. But if that's not the case, failing the allocate will cause the same return value as the default. Signed-off-by: Mike Travis Signed-off-by: Ingo Molnar --- arch/x86/include/asm/bigsmp/apic.h | 4 ++- arch/x86/include/asm/es7000/apic.h | 40 ++++++++++++--------------- arch/x86/include/asm/mach-default/mach_apic.h | 3 +- arch/x86/include/asm/summit/apic.h | 30 +++++++++++--------- arch/x86/kernel/genapic_flat_64.c | 4 ++- arch/x86/kernel/genx2apic_cluster.c | 4 ++- arch/x86/kernel/genx2apic_phys.c | 4 ++- arch/x86/kernel/genx2apic_uv_x.c | 4 ++- 8 files changed, 52 insertions(+), 41 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/bigsmp/apic.h b/arch/x86/include/asm/bigsmp/apic.h index 976399debb3f..d8dd9f537911 100644 --- a/arch/x86/include/asm/bigsmp/apic.h +++ b/arch/x86/include/asm/bigsmp/apic.h @@ -138,7 +138,9 @@ static inline unsigned int cpu_mask_to_apicid_and(const struct cpumask *cpumask, * We're using fixed IRQ delivery, can only return one phys APIC ID. * May as well be the first. */ - cpu = cpumask_any_and(cpumask, andmask); + for_each_cpu_and(cpu, cpumask, andmask) + if (cpumask_test_cpu(cpu, cpu_online_mask)) + break; if (cpu < nr_cpu_ids) return cpu_to_logical_apicid(cpu); diff --git a/arch/x86/include/asm/es7000/apic.h b/arch/x86/include/asm/es7000/apic.h index ba8423c5363f..51ac1230294e 100644 --- a/arch/x86/include/asm/es7000/apic.h +++ b/arch/x86/include/asm/es7000/apic.h @@ -214,51 +214,47 @@ static inline unsigned int cpu_mask_to_apicid(const cpumask_t *cpumask) return apicid; } -static inline unsigned int cpu_mask_to_apicid_and(const struct cpumask *cpumask, + +static inline unsigned int cpu_mask_to_apicid_and(const struct cpumask *inmask, const struct cpumask *andmask) { int num_bits_set; - int num_bits_set2; int cpus_found = 0; int cpu; - int apicid = 0; + int apicid = cpu_to_logical_apicid(0); + cpumask_var_t cpumask; + + if (!alloc_cpumask_var(&cpumask, GFP_ATOMIC)) + return apicid; + + cpumask_and(cpumask, inmask, andmask); + cpumask_and(cpumask, cpumask, cpu_online_mask); num_bits_set = cpumask_weight(cpumask); - num_bits_set2 = cpumask_weight(andmask); - num_bits_set = min(num_bits_set, num_bits_set2); /* Return id to all */ - if (num_bits_set >= nr_cpu_ids) -#if defined CONFIG_ES7000_CLUSTERED_APIC - return 0xFF; -#else - return cpu_to_logical_apicid(0); -#endif + if (num_bits_set == NR_CPUS) + goto exit; /* * The cpus in the mask must all be on the apic cluster. If are not * on the same apicid cluster return default value of TARGET_CPUS. */ - cpu = cpumask_first_and(cpumask, andmask); + cpu = cpumask_first(cpumask); apicid = cpu_to_logical_apicid(cpu); - while (cpus_found < num_bits_set) { - if (cpumask_test_cpu(cpu, cpumask) && - cpumask_test_cpu(cpu, andmask)) { + if (cpumask_test_cpu(cpu, cpumask)) { int new_apicid = cpu_to_logical_apicid(cpu); if (apicid_cluster(apicid) != - apicid_cluster(new_apicid)) { - printk(KERN_WARNING - "%s: Not a valid mask!\n", __func__); -#if defined CONFIG_ES7000_CLUSTERED_APIC - return 0xFF; -#else + apicid_cluster(new_apicid)){ + printk ("%s: Not a valid mask!\n", __func__); return cpu_to_logical_apicid(0); -#endif } apicid = new_apicid; cpus_found++; } cpu++; } +exit: + free_cpumask_var(cpumask); return apicid; } diff --git a/arch/x86/include/asm/mach-default/mach_apic.h b/arch/x86/include/asm/mach-default/mach_apic.h index 8863d978cb96..cc09cbbee27e 100644 --- a/arch/x86/include/asm/mach-default/mach_apic.h +++ b/arch/x86/include/asm/mach-default/mach_apic.h @@ -72,8 +72,9 @@ static inline unsigned int cpu_mask_to_apicid_and(const struct cpumask *cpumask, { unsigned long mask1 = cpumask_bits(cpumask)[0]; unsigned long mask2 = cpumask_bits(andmask)[0]; + unsigned long mask3 = cpumask_bits(cpu_online_mask)[0]; - return (unsigned int)(mask1 & mask2); + return (unsigned int)(mask1 & mask2 & mask3); } static inline u32 phys_pkg_id(u32 cpuid_apic, int index_msb) diff --git a/arch/x86/include/asm/summit/apic.h b/arch/x86/include/asm/summit/apic.h index 651a93849341..99327d1be49f 100644 --- a/arch/x86/include/asm/summit/apic.h +++ b/arch/x86/include/asm/summit/apic.h @@ -170,35 +170,37 @@ static inline unsigned int cpu_mask_to_apicid(const cpumask_t *cpumask) return apicid; } -static inline unsigned int cpu_mask_to_apicid_and(const struct cpumask *cpumask, +static inline unsigned int cpu_mask_to_apicid_and(const struct cpumask *inmask, const struct cpumask *andmask) { int num_bits_set; - int num_bits_set2; int cpus_found = 0; int cpu; - int apicid = 0; + int apicid = 0xFF; + cpumask_var_t cpumask; + + if (!alloc_cpumask_var(&cpumask, GFP_ATOMIC)) + return (int) 0xFF; + + cpumask_and(cpumask, inmask, andmask); + cpumask_and(cpumask, cpumask, cpu_online_mask); num_bits_set = cpumask_weight(cpumask); - num_bits_set2 = cpumask_weight(andmask); - num_bits_set = min(num_bits_set, num_bits_set2); /* Return id to all */ - if (num_bits_set >= nr_cpu_ids) - return 0xFF; + if (num_bits_set == nr_cpu_ids) + goto exit; /* * The cpus in the mask must all be on the apic cluster. If are not * on the same apicid cluster return default value of TARGET_CPUS. */ - cpu = cpumask_first_and(cpumask, andmask); + cpu = cpumask_first(cpumask); apicid = cpu_to_logical_apicid(cpu); while (cpus_found < num_bits_set) { - if (cpumask_test_cpu(cpu, cpumask) - && cpumask_test_cpu(cpu, andmask)) { + if (cpumask_test_cpu(cpu, cpumask)) { int new_apicid = cpu_to_logical_apicid(cpu); if (apicid_cluster(apicid) != - apicid_cluster(new_apicid)) { - printk(KERN_WARNING - "%s: Not a valid mask!\n", __func__); + apicid_cluster(new_apicid)){ + printk ("%s: Not a valid mask!\n", __func__); return 0xFF; } apicid = apicid | new_apicid; @@ -206,6 +208,8 @@ static inline unsigned int cpu_mask_to_apicid_and(const struct cpumask *cpumask, } cpu++; } +exit: + free_cpumask_var(cpumask); return apicid; } diff --git a/arch/x86/kernel/genapic_flat_64.c b/arch/x86/kernel/genapic_flat_64.c index 7fa5f49c2dda..34185488e4fb 100644 --- a/arch/x86/kernel/genapic_flat_64.c +++ b/arch/x86/kernel/genapic_flat_64.c @@ -276,7 +276,9 @@ physflat_cpu_mask_to_apicid_and(const struct cpumask *cpumask, * We're using fixed IRQ delivery, can only return one phys APIC ID. * May as well be the first. */ - cpu = cpumask_any_and(cpumask, andmask); + for_each_cpu_and(cpu, cpumask, andmask) + if (cpumask_test_cpu(cpu, cpu_online_mask)) + break; if (cpu < nr_cpu_ids) return per_cpu(x86_cpu_to_apicid, cpu); return BAD_APICID; diff --git a/arch/x86/kernel/genx2apic_cluster.c b/arch/x86/kernel/genx2apic_cluster.c index 4716a0c9f936..d451c9b9fdff 100644 --- a/arch/x86/kernel/genx2apic_cluster.c +++ b/arch/x86/kernel/genx2apic_cluster.c @@ -133,7 +133,9 @@ static unsigned int x2apic_cpu_mask_to_apicid_and(const struct cpumask *cpumask, * We're using fixed IRQ delivery, can only return one phys APIC ID. * May as well be the first. */ - cpu = cpumask_any_and(cpumask, andmask); + for_each_cpu_and(cpu, cpumask, andmask) + if (cpumask_test_cpu(cpu, cpu_online_mask)) + break; if (cpu < nr_cpu_ids) return per_cpu(x86_cpu_to_apicid, cpu); return BAD_APICID; diff --git a/arch/x86/kernel/genx2apic_phys.c b/arch/x86/kernel/genx2apic_phys.c index b255507884f2..62895cf315ff 100644 --- a/arch/x86/kernel/genx2apic_phys.c +++ b/arch/x86/kernel/genx2apic_phys.c @@ -132,7 +132,9 @@ static unsigned int x2apic_cpu_mask_to_apicid_and(const struct cpumask *cpumask, * We're using fixed IRQ delivery, can only return one phys APIC ID. * May as well be the first. */ - cpu = cpumask_any_and(cpumask, andmask); + for_each_cpu_and(cpu, cpumask, andmask) + if (cpumask_test_cpu(cpu, cpu_online_mask)) + break; if (cpu < nr_cpu_ids) return per_cpu(x86_cpu_to_apicid, cpu); return BAD_APICID; diff --git a/arch/x86/kernel/genx2apic_uv_x.c b/arch/x86/kernel/genx2apic_uv_x.c index 3984682cd849..0e88be11227d 100644 --- a/arch/x86/kernel/genx2apic_uv_x.c +++ b/arch/x86/kernel/genx2apic_uv_x.c @@ -188,7 +188,9 @@ static unsigned int uv_cpu_mask_to_apicid_and(const struct cpumask *cpumask, * We're using fixed IRQ delivery, can only return one phys APIC ID. * May as well be the first. */ - cpu = cpumask_any_and(cpumask, andmask); + for_each_cpu_and(cpu, cpumask, andmask) + if (cpumask_test_cpu(cpu, cpu_online_mask)) + break; if (cpu < nr_cpu_ids) return per_cpu(x86_cpu_to_apicid, cpu); return BAD_APICID; -- cgit v1.2.3 From 3b11ce7f542e415c90267b4482d4611410b468e6 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Wed, 17 Dec 2008 15:21:39 -0800 Subject: x86: use possible_cpus=NUM to extend the possible cpus allowed Impact: add new boot parameter Use possible_cpus=NUM kernel parameter to extend the number of possible cpus. The ability to HOTPLUG ON cpus that are "possible" but not "present" is dealt with in a later patch. Signed-off-by: Mike Travis --- Documentation/cpu-hotplug.txt | 17 +++++++++-------- arch/x86/kernel/apic.c | 20 ++++++++++++-------- arch/x86/kernel/smpboot.c | 25 +++++++++++++++++++++---- 3 files changed, 42 insertions(+), 20 deletions(-) (limited to 'arch/x86') diff --git a/Documentation/cpu-hotplug.txt b/Documentation/cpu-hotplug.txt index 94bbc27ddd4f..9d620c153b04 100644 --- a/Documentation/cpu-hotplug.txt +++ b/Documentation/cpu-hotplug.txt @@ -50,16 +50,17 @@ additional_cpus=n (*) Use this to limit hotpluggable cpus. This option sets cpu_possible_map = cpu_present_map + additional_cpus (*) Option valid only for following architectures -- x86_64, ia64 +- ia64 -ia64 and x86_64 use the number of disabled local apics in ACPI tables MADT -to determine the number of potentially hot-pluggable cpus. The implementation -should only rely on this to count the # of cpus, but *MUST* not rely on the -apicid values in those tables for disabled apics. In the event BIOS doesn't -mark such hot-pluggable cpus as disabled entries, one could use this -parameter "additional_cpus=x" to represent those cpus in the cpu_possible_map. +ia64 uses the number of disabled local apics in ACPI tables MADT to +determine the number of potentially hot-pluggable cpus. The implementation +should only rely on this to count the # of cpus, but *MUST* not rely +on the apicid values in those tables for disabled apics. In the event +BIOS doesn't mark such hot-pluggable cpus as disabled entries, one could +use this parameter "additional_cpus=x" to represent those cpus in the +cpu_possible_map. -possible_cpus=n [s390 only] use this to set hotpluggable cpus. +possible_cpus=n [s390,x86_64] use this to set hotpluggable cpus. This option sets possible_cpus bits in cpu_possible_map. Thus keeping the numbers of bits set constant even if the machine gets rebooted. diff --git a/arch/x86/kernel/apic.c b/arch/x86/kernel/apic.c index 93cf2d13f335..f7a32a3beb2f 100644 --- a/arch/x86/kernel/apic.c +++ b/arch/x86/kernel/apic.c @@ -1819,28 +1819,32 @@ void disconnect_bsp_APIC(int virt_wire_setup) void __cpuinit generic_processor_info(int apicid, int version) { int cpu; - cpumask_t tmp_map; /* * Validate version */ if (version == 0x0) { pr_warning("BIOS bug, APIC version is 0 for CPU#%d! " - "fixing up to 0x10. (tell your hw vendor)\n", - version); + "fixing up to 0x10. (tell your hw vendor)\n", + version); version = 0x10; } apic_version[apicid] = version; - if (num_processors >= NR_CPUS) { - pr_warning("WARNING: NR_CPUS limit of %i reached." - " Processor ignored.\n", NR_CPUS); + if (num_processors >= nr_cpu_ids) { + int max = nr_cpu_ids; + int thiscpu = max + disabled_cpus; + + pr_warning( + "ACPI: NR_CPUS/possible_cpus limit of %i reached." + " Processor %d/0x%x ignored.\n", max, thiscpu, apicid); + + disabled_cpus++; return; } num_processors++; - cpus_complement(tmp_map, cpu_present_map); - cpu = first_cpu(tmp_map); + cpu = cpumask_next_zero(-1, cpu_present_mask); physid_set(apicid, phys_cpu_present_map); if (apicid == boot_cpu_physical_apicid) { diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index be9466788043..1a9941b11150 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -1252,6 +1252,15 @@ void __init native_smp_cpus_done(unsigned int max_cpus) check_nmi_watchdog(); } +static int __initdata setup_possible_cpus = -1; +static int __init _setup_possible_cpus(char *str) +{ + get_option(&str, &setup_possible_cpus); + return 0; +} +early_param("possible_cpus", _setup_possible_cpus); + + /* * cpu_possible_map should be static, it cannot change as cpu's * are onlined, or offlined. The reason is per-cpu data-structures @@ -1264,7 +1273,7 @@ void __init native_smp_cpus_done(unsigned int max_cpus) * * Three ways to find out the number of additional hotplug CPUs: * - If the BIOS specified disabled CPUs in ACPI/mptables use that. - * - The user can overwrite it with additional_cpus=NUM + * - The user can overwrite it with possible_cpus=NUM * - Otherwise don't reserve additional CPUs. * We do this because additional CPUs waste a lot of memory. * -AK @@ -1277,9 +1286,17 @@ __init void prefill_possible_map(void) if (!num_processors) num_processors = 1; - possible = num_processors + disabled_cpus; - if (possible > NR_CPUS) - possible = NR_CPUS; + if (setup_possible_cpus == -1) + possible = num_processors + disabled_cpus; + else + possible = setup_possible_cpus; + + if (possible > CONFIG_NR_CPUS) { + printk(KERN_WARNING + "%d Processors exceeds NR_CPUS limit of %d\n", + possible, CONFIG_NR_CPUS); + possible = CONFIG_NR_CPUS; + } printk(KERN_INFO "SMP: Allowing %d CPUs, %d hotplug CPUs\n", possible, max_t(int, possible - num_processors, 0)); -- cgit v1.2.3 From 55aab5f49e384a361668d112eefdb33e90779af9 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Wed, 17 Dec 2008 12:52:34 -0700 Subject: x86 gart: don't complain if no AMD GART found Impact: remove annoying bootup printk It's perfectly normal for no AMD GART to be present, e.g., if you have Intel CPUs. None of the other iommu_init() functions makes noise when it finds nothing. Signed-off-by: Bjorn Helgaas Acked-by: Joerg Roedel Signed-off-by: Ingo Molnar --- arch/x86/kernel/pci-gart_64.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/pci-gart_64.c b/arch/x86/kernel/pci-gart_64.c index ba7ad83e20a8..a35eaa379ff6 100644 --- a/arch/x86/kernel/pci-gart_64.c +++ b/arch/x86/kernel/pci-gart_64.c @@ -745,10 +745,8 @@ void __init gart_iommu_init(void) unsigned long scratch; long i; - if (cache_k8_northbridges() < 0 || num_k8_northbridges == 0) { - printk(KERN_INFO "PCI-GART: No AMD GART found.\n"); + if (cache_k8_northbridges() < 0 || num_k8_northbridges == 0) return; - } #ifndef CONFIG_AGP_AMD64 no_agp = 1; -- cgit v1.2.3 From 57a37505d19f4dfeee26f0fd7ea38ed6f1d10cbe Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Date: Wed, 17 Dec 2008 23:17:21 +0530 Subject: x86: time_64.c timer_interrupt() should be static Impact: cleanup, reduce kernel size a bit Signed-off-by: Jaswinder Singh Signed-off-by: Ingo Molnar --- arch/x86/kernel/time_64.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/time_64.c b/arch/x86/kernel/time_64.c index cb19d650c216..083a4a5bb00f 100644 --- a/arch/x86/kernel/time_64.c +++ b/arch/x86/kernel/time_64.c @@ -49,7 +49,7 @@ unsigned long profile_pc(struct pt_regs *regs) } EXPORT_SYMBOL(profile_pc); -irqreturn_t timer_interrupt(int irq, void *dev_id) +static irqreturn_t timer_interrupt(int irq, void *dev_id) { add_pda(irq0_irqs, 1); -- cgit v1.2.3 From 7c9c160c54fc545efc23881344593868e5f717bd Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Date: Wed, 17 Dec 2008 23:18:52 +0530 Subject: x86: tls.c declare sys_set_thread_area and sys_get_thread_area before they get used Impact: cleanup In asm/syscalls.h move out sys_set_thread_area() and sys_get_thread_area() as they are common for both 32 and 64 bit. Signed-off-by: Jaswinder Singh Signed-off-by: Ingo Molnar --- arch/x86/include/asm/syscalls.h | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/syscalls.h b/arch/x86/include/asm/syscalls.h index 75d4a6afc36f..c0b0bda754ee 100644 --- a/arch/x86/include/asm/syscalls.h +++ b/arch/x86/include/asm/syscalls.h @@ -22,6 +22,10 @@ asmlinkage long sys_ioperm(unsigned long, unsigned long, int); /* kernel/ldt.c */ asmlinkage int sys_modify_ldt(int, void __user *, unsigned long); +/* kernel/tls.c */ +asmlinkage int sys_set_thread_area(struct user_desc __user *); +asmlinkage int sys_get_thread_area(struct user_desc __user *); + /* X86_32 only */ #ifdef CONFIG_X86_32 /* kernel/process_32.c */ @@ -54,10 +58,6 @@ asmlinkage int sys_uname(struct old_utsname __user *); struct oldold_utsname; asmlinkage int sys_olduname(struct oldold_utsname __user *); -/* kernel/tls.c */ -asmlinkage int sys_set_thread_area(struct user_desc __user *); -asmlinkage int sys_get_thread_area(struct user_desc __user *); - /* kernel/vm86_32.c */ asmlinkage int sys_vm86old(struct pt_regs); asmlinkage int sys_vm86(struct pt_regs); -- cgit v1.2.3 From f269b07e862c395d6981ab2c05d6bc34b0249e90 Mon Sep 17 00:00:00 2001 From: Vegard Nossum Date: Thu, 18 Dec 2008 18:35:06 +0100 Subject: x86: revert CONFIG_RELOCATABLE=y defconfig change This commit: commit 5cb04df8d3f03e37a19f2502591a84156be71772 Author: Ingo Molnar Date: Sun May 4 19:49:04 2008 +0200 x86: defconfig updates changed CONFIG_RELOCATABLE from n to y, which may lead to a mismatch between the vmlinux debug information and the runtime location of the kernel, even when the bootloader does not relocate the kernel. Revert the specific change. Works for me with GRUB and qemu. Reference: http://lkml.org/lkml/2008/11/25/243 Signed-off-by: Vegard Nossum Signed-off-by: Ingo Molnar --- arch/x86/configs/i386_defconfig | 2 +- arch/x86/configs/x86_64_defconfig | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/configs/i386_defconfig b/arch/x86/configs/i386_defconfig index 71fc39c70782..b30a08ed8eb4 100644 --- a/arch/x86/configs/i386_defconfig +++ b/arch/x86/configs/i386_defconfig @@ -298,7 +298,7 @@ CONFIG_KEXEC=y CONFIG_CRASH_DUMP=y # CONFIG_KEXEC_JUMP is not set CONFIG_PHYSICAL_START=0x1000000 -CONFIG_RELOCATABLE=y +# CONFIG_RELOCATABLE is not set CONFIG_PHYSICAL_ALIGN=0x200000 CONFIG_HOTPLUG_CPU=y # CONFIG_COMPAT_VDSO is not set diff --git a/arch/x86/configs/x86_64_defconfig b/arch/x86/configs/x86_64_defconfig index b38bbabc1706..0e7dbc0a3e46 100644 --- a/arch/x86/configs/x86_64_defconfig +++ b/arch/x86/configs/x86_64_defconfig @@ -298,7 +298,7 @@ CONFIG_SCHED_HRTICK=y CONFIG_KEXEC=y CONFIG_CRASH_DUMP=y CONFIG_PHYSICAL_START=0x1000000 -CONFIG_RELOCATABLE=y +# CONFIG_RELOCATABLE is not set CONFIG_PHYSICAL_ALIGN=0x200000 CONFIG_HOTPLUG_CPU=y # CONFIG_COMPAT_VDSO is not set -- cgit v1.2.3 From 5c2628e8b4f670d0954053444289e2b018be957a Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Thu, 18 Dec 2008 09:18:35 -0800 Subject: x86: sigframe.h: add guard macro Impact: cleanup Add missing guard macro _ASM_X86_SIGFRAME_H. Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/include/asm/sigframe.h | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/sigframe.h b/arch/x86/include/asm/sigframe.h index 491a0878c3aa..3bd0f4276000 100644 --- a/arch/x86/include/asm/sigframe.h +++ b/arch/x86/include/asm/sigframe.h @@ -1,3 +1,6 @@ +#ifndef _ASM_X86_SIGFRAME_H +#define _ASM_X86_SIGFRAME_H + #ifdef CONFIG_X86_32 #define sigframe_ia32 sigframe #define rt_sigframe_ia32 rt_sigframe @@ -59,3 +62,5 @@ struct rt_sigframe { /* fp state follows here */ }; #endif /* CONFIG_X86_64 */ + +#endif /* _ASM_X86_SIGFRAME_H */ -- cgit v1.2.3 From f0bc2202e0373eb8e9b1ddbec930e2e681357db8 Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Date: Wed, 17 Dec 2008 23:20:05 +0530 Subject: x86: process.c declare c1e_remove_cpu before they get used Impact: cleanup, avoid sparse warning Included asm/idle.h for c1e_remove_cpu() declaration. Fixes this sparse warning: CHECK arch/x86/kernel/process.c arch/x86/kernel/process.c:284:6: warning: symbol 'c1e_remove_cpu' was not declared. Should it be static? Signed-off-by: Jaswinder Singh Signed-off-by: Ingo Molnar --- arch/x86/kernel/process.c | 1 + 1 file changed, 1 insertion(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c index c622772744d8..b06100f1d612 100644 --- a/arch/x86/kernel/process.c +++ b/arch/x86/kernel/process.c @@ -1,6 +1,7 @@ #include #include #include +#include #include #include #include -- cgit v1.2.3 From 5899329b19100c0b82dc78e9b21ed8b920c9ffb3 Mon Sep 17 00:00:00 2001 From: "venkatesh.pallipadi@intel.com" Date: Thu, 18 Dec 2008 11:41:30 -0800 Subject: x86: PAT: implement track/untrack of pfnmap regions for x86 - v3 Impact: New mm functionality. Hookup remap_pfn_range and vm_insert_pfn and corresponding copy and free routines with reserve and free tracking. reserve and free here only takes care of non RAM region mapping. For RAM region, driver should use set_memory_[uc|wc|wb] to set the cache type and then setup the mapping for user pte. We can bypass below reserve/free in that case. Signed-off-by: Venkatesh Pallipadi Signed-off-by: Suresh Siddha Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/pgtable.h | 10 ++ arch/x86/mm/pat.c | 236 +++++++++++++++++++++++++++++++++++++++++ 2 files changed, 246 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h index c012f3b11671..7dcd94c29044 100644 --- a/arch/x86/include/asm/pgtable.h +++ b/arch/x86/include/asm/pgtable.h @@ -219,6 +219,11 @@ static inline unsigned long pte_pfn(pte_t pte) return (pte_val(pte) & PTE_PFN_MASK) >> PAGE_SHIFT; } +static inline u64 pte_pa(pte_t pte) +{ + return pte_val(pte) & PTE_PFN_MASK; +} + #define pte_page(pte) pfn_to_page(pte_pfn(pte)) static inline int pmd_large(pmd_t pte) @@ -328,6 +333,11 @@ static inline pgprot_t pgprot_modify(pgprot_t oldprot, pgprot_t newprot) #define canon_pgprot(p) __pgprot(pgprot_val(p) & __supported_pte_mask) +/* Indicate that x86 has its own track and untrack pfn vma functions */ +#define track_pfn_vma_new track_pfn_vma_new +#define track_pfn_vma_copy track_pfn_vma_copy +#define untrack_pfn_vma untrack_pfn_vma + #ifndef __ASSEMBLY__ #define __HAVE_PHYS_MEM_ACCESS_PROT struct file; diff --git a/arch/x86/mm/pat.c b/arch/x86/mm/pat.c index eb1bf000d12e..1069ffecf77d 100644 --- a/arch/x86/mm/pat.c +++ b/arch/x86/mm/pat.c @@ -596,6 +596,242 @@ void unmap_devmem(unsigned long pfn, unsigned long size, pgprot_t vma_prot) free_memtype(addr, addr + size); } +/* + * Internal interface to reserve a range of physical memory with prot. + * Reserved non RAM regions only and after successful reserve_memtype, + * this func also keeps identity mapping (if any) in sync with this new prot. + */ +static int reserve_pfn_range(u64 paddr, unsigned long size, pgprot_t vma_prot) +{ + int is_ram = 0; + int id_sz, ret; + unsigned long flags; + unsigned long want_flags = (pgprot_val(vma_prot) & _PAGE_CACHE_MASK); + + is_ram = pagerange_is_ram(paddr, paddr + size); + + if (is_ram != 0) { + /* + * For mapping RAM pages, drivers need to call + * set_memory_[uc|wc|wb] directly, for reserve and free, before + * setting up the PTE. + */ + WARN_ON_ONCE(1); + return 0; + } + + ret = reserve_memtype(paddr, paddr + size, want_flags, &flags); + if (ret) + return ret; + + if (flags != want_flags) { + free_memtype(paddr, paddr + size); + printk(KERN_ERR + "%s:%d map pfn expected mapping type %s for %Lx-%Lx, got %s\n", + current->comm, current->pid, + cattr_name(want_flags), + (unsigned long long)paddr, + (unsigned long long)(paddr + size), + cattr_name(flags)); + return -EINVAL; + } + + /* Need to keep identity mapping in sync */ + if (paddr >= __pa(high_memory)) + return 0; + + id_sz = (__pa(high_memory) < paddr + size) ? + __pa(high_memory) - paddr : + size; + + if (ioremap_change_attr((unsigned long)__va(paddr), id_sz, flags) < 0) { + free_memtype(paddr, paddr + size); + printk(KERN_ERR + "%s:%d reserve_pfn_range ioremap_change_attr failed %s " + "for %Lx-%Lx\n", + current->comm, current->pid, + cattr_name(flags), + (unsigned long long)paddr, + (unsigned long long)(paddr + size)); + return -EINVAL; + } + return 0; +} + +/* + * Internal interface to free a range of physical memory. + * Frees non RAM regions only. + */ +static void free_pfn_range(u64 paddr, unsigned long size) +{ + int is_ram; + + is_ram = pagerange_is_ram(paddr, paddr + size); + if (is_ram == 0) + free_memtype(paddr, paddr + size); +} + +/* + * track_pfn_vma_copy is called when vma that is covering the pfnmap gets + * copied through copy_page_range(). + * + * If the vma has a linear pfn mapping for the entire range, we get the prot + * from pte and reserve the entire vma range with single reserve_pfn_range call. + * Otherwise, we reserve the entire vma range, my ging through the PTEs page + * by page to get physical address and protection. + */ +int track_pfn_vma_copy(struct vm_area_struct *vma) +{ + int retval = 0; + unsigned long i, j; + u64 paddr; + pgprot_t prot; + pte_t pte; + unsigned long vma_start = vma->vm_start; + unsigned long vma_end = vma->vm_end; + unsigned long vma_size = vma_end - vma_start; + + if (!pat_enabled) + return 0; + + if (is_linear_pfn_mapping(vma)) { + /* + * reserve the whole chunk starting from vm_pgoff, + * But, we have to get the protection from pte. + */ + if (follow_pfnmap_pte(vma, vma_start, &pte)) { + WARN_ON_ONCE(1); + return -1; + } + prot = pte_pgprot(pte); + paddr = (u64)vma->vm_pgoff << PAGE_SHIFT; + return reserve_pfn_range(paddr, vma_size, prot); + } + + /* reserve entire vma page by page, using pfn and prot from pte */ + for (i = 0; i < vma_size; i += PAGE_SIZE) { + if (follow_pfnmap_pte(vma, vma_start + i, &pte)) + continue; + + paddr = pte_pa(pte); + prot = pte_pgprot(pte); + retval = reserve_pfn_range(paddr, PAGE_SIZE, prot); + if (retval) + goto cleanup_ret; + } + return 0; + +cleanup_ret: + /* Reserve error: Cleanup partial reservation and return error */ + for (j = 0; j < i; j += PAGE_SIZE) { + if (follow_pfnmap_pte(vma, vma_start + j, &pte)) + continue; + + paddr = pte_pa(pte); + free_pfn_range(paddr, PAGE_SIZE); + } + + return retval; +} + +/* + * track_pfn_vma_new is called when a _new_ pfn mapping is being established + * for physical range indicated by pfn and size. + * + * prot is passed in as a parameter for the new mapping. If the vma has a + * linear pfn mapping for the entire range reserve the entire vma range with + * single reserve_pfn_range call. + * Otherwise, we look t the pfn and size and reserve only the specified range + * page by page. + * + * Note that this function can be called with caller trying to map only a + * subrange/page inside the vma. + */ +int track_pfn_vma_new(struct vm_area_struct *vma, pgprot_t prot, + unsigned long pfn, unsigned long size) +{ + int retval = 0; + unsigned long i, j; + u64 base_paddr; + u64 paddr; + unsigned long vma_start = vma->vm_start; + unsigned long vma_end = vma->vm_end; + unsigned long vma_size = vma_end - vma_start; + + if (!pat_enabled) + return 0; + + if (is_linear_pfn_mapping(vma)) { + /* reserve the whole chunk starting from vm_pgoff */ + paddr = (u64)vma->vm_pgoff << PAGE_SHIFT; + return reserve_pfn_range(paddr, vma_size, prot); + } + + /* reserve page by page using pfn and size */ + base_paddr = (u64)pfn << PAGE_SHIFT; + for (i = 0; i < size; i += PAGE_SIZE) { + paddr = base_paddr + i; + retval = reserve_pfn_range(paddr, PAGE_SIZE, prot); + if (retval) + goto cleanup_ret; + } + return 0; + +cleanup_ret: + /* Reserve error: Cleanup partial reservation and return error */ + for (j = 0; j < i; j += PAGE_SIZE) { + paddr = base_paddr + j; + free_pfn_range(paddr, PAGE_SIZE); + } + + return retval; +} + +/* + * untrack_pfn_vma is called while unmapping a pfnmap for a region. + * untrack can be called for a specific region indicated by pfn and size or + * can be for the entire vma (in which case size can be zero). + */ +void untrack_pfn_vma(struct vm_area_struct *vma, unsigned long pfn, + unsigned long size) +{ + unsigned long i; + u64 paddr; + unsigned long vma_start = vma->vm_start; + unsigned long vma_end = vma->vm_end; + unsigned long vma_size = vma_end - vma_start; + + if (!pat_enabled) + return; + + if (is_linear_pfn_mapping(vma)) { + /* free the whole chunk starting from vm_pgoff */ + paddr = (u64)vma->vm_pgoff << PAGE_SHIFT; + free_pfn_range(paddr, vma_size); + return; + } + + if (size != 0 && size != vma_size) { + /* free page by page, using pfn and size */ + paddr = (u64)pfn << PAGE_SHIFT; + for (i = 0; i < size; i += PAGE_SIZE) { + paddr = paddr + i; + free_pfn_range(paddr, PAGE_SIZE); + } + } else { + /* free entire vma, page by page, using the pfn from pte */ + for (i = 0; i < vma_size; i += PAGE_SIZE) { + pte_t pte; + + if (follow_pfnmap_pte(vma, vma_start + i, &pte)) + continue; + + paddr = pte_pa(pte); + free_pfn_range(paddr, PAGE_SIZE); + } + } +} + #if defined(CONFIG_DEBUG_FS) && defined(CONFIG_X86_PAT) /* get Nth element of the linked list */ -- cgit v1.2.3 From 8a7b12f70fb135a1b1d865687de3edcdc780f6d1 Mon Sep 17 00:00:00 2001 From: "venkatesh.pallipadi@intel.com" Date: Thu, 18 Dec 2008 11:41:31 -0800 Subject: x86: PAT: change pgprot_noncached to uc_minus instead of strong uc - v3 Impact: mm behavior change. Make pgprot_noncached uc_minus instead of strong UC. This will make pgprot_noncached to be in line with ioremap_nocache() and all the other APIs that map page uc_minus on uc request. Signed-off-by: Venkatesh Pallipadi Signed-off-by: Suresh Siddha Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/pgtable.h | 8 ++++++++ arch/x86/include/asm/pgtable_32.h | 9 --------- arch/x86/include/asm/pgtable_64.h | 6 ------ 3 files changed, 8 insertions(+), 15 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h index 7dcd94c29044..6968d4f6be3e 100644 --- a/arch/x86/include/asm/pgtable.h +++ b/arch/x86/include/asm/pgtable.h @@ -158,6 +158,14 @@ #define PGD_IDENT_ATTR 0x001 /* PRESENT (no other attributes) */ #endif +/* + * Macro to mark a page protection value as UC- + */ +#define pgprot_noncached(prot) \ + ((boot_cpu_data.x86 > 3) \ + ? (__pgprot(pgprot_val(prot) | _PAGE_CACHE_UC_MINUS)) \ + : (prot)) + #ifndef __ASSEMBLY__ /* diff --git a/arch/x86/include/asm/pgtable_32.h b/arch/x86/include/asm/pgtable_32.h index f9d5889b336b..72b020deb46b 100644 --- a/arch/x86/include/asm/pgtable_32.h +++ b/arch/x86/include/asm/pgtable_32.h @@ -100,15 +100,6 @@ extern unsigned long pg0[]; # include #endif -/* - * Macro to mark a page protection value as "uncacheable". - * On processors which do not support it, this is a no-op. - */ -#define pgprot_noncached(prot) \ - ((boot_cpu_data.x86 > 3) \ - ? (__pgprot(pgprot_val(prot) | _PAGE_PCD | _PAGE_PWT)) \ - : (prot)) - /* * Conversion functions: convert a page and protection to a page entry, * and a page entry and page directory to the page they refer to. diff --git a/arch/x86/include/asm/pgtable_64.h b/arch/x86/include/asm/pgtable_64.h index 545a0e042bb2..4798a4033e3a 100644 --- a/arch/x86/include/asm/pgtable_64.h +++ b/arch/x86/include/asm/pgtable_64.h @@ -176,12 +176,6 @@ static inline int pmd_bad(pmd_t pmd) #define pages_to_mb(x) ((x) >> (20 - PAGE_SHIFT)) /* FIXME: is this right? */ -/* - * Macro to mark a page protection value as "uncacheable". - */ -#define pgprot_noncached(prot) \ - (__pgprot(pgprot_val((prot)) | _PAGE_PCD | _PAGE_PWT)) - /* * Conversion functions: convert a page and protection to a page entry, * and a page entry and page directory to the page they refer to. -- cgit v1.2.3 From 2520bd3123c00272f818a176c92d03c7d0a113d6 Mon Sep 17 00:00:00 2001 From: "venkatesh.pallipadi@intel.com" Date: Thu, 18 Dec 2008 11:41:32 -0800 Subject: x86: PAT: add pgprot_writecombine() interface for drivers - v3 Impact: New mm functionality. Add pgprot_writecombine. pgprot_writecombine will be aliased to pgprot_noncached when not supported by the architecture. Signed-off-by: Venkatesh Pallipadi Signed-off-by: Suresh Siddha Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/pgtable.h | 3 +++ arch/x86/mm/pat.c | 8 ++++++++ include/asm-generic/pgtable.h | 4 ++++ 3 files changed, 15 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h index 6968d4f6be3e..579f8ceee948 100644 --- a/arch/x86/include/asm/pgtable.h +++ b/arch/x86/include/asm/pgtable.h @@ -168,6 +168,9 @@ #ifndef __ASSEMBLY__ +#define pgprot_writecombine pgprot_writecombine +extern pgprot_t pgprot_writecombine(pgprot_t prot); + /* * ZERO_PAGE is a global shared page that is always zero: used * for zero-mapped memory areas etc.. diff --git a/arch/x86/mm/pat.c b/arch/x86/mm/pat.c index 1069ffecf77d..d5254bae84f4 100644 --- a/arch/x86/mm/pat.c +++ b/arch/x86/mm/pat.c @@ -832,6 +832,14 @@ void untrack_pfn_vma(struct vm_area_struct *vma, unsigned long pfn, } } +pgprot_t pgprot_writecombine(pgprot_t prot) +{ + if (pat_enabled) + return __pgprot(pgprot_val(prot) | _PAGE_CACHE_WC); + else + return pgprot_noncached(prot); +} + #if defined(CONFIG_DEBUG_FS) && defined(CONFIG_X86_PAT) /* get Nth element of the linked list */ diff --git a/include/asm-generic/pgtable.h b/include/asm-generic/pgtable.h index ef87f889ef62..b84633801fb6 100644 --- a/include/asm-generic/pgtable.h +++ b/include/asm-generic/pgtable.h @@ -129,6 +129,10 @@ static inline void ptep_set_wrprotect(struct mm_struct *mm, unsigned long addres #define move_pte(pte, prot, old_addr, new_addr) (pte) #endif +#ifndef pgprot_writecombine +#define pgprot_writecombine pgprot_noncached +#endif + /* * When walking page tables, get the address of the next boundary, * or the end address of the range if that comes earlier. Although no -- cgit v1.2.3 From d1769d5475176124af04fa69848b022c98c4bc37 Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Date: Fri, 19 Dec 2008 00:03:56 +0530 Subject: x86: traps.c declare functions before they get used Impact: cleanup In asm/traps.h :- do_double_fault : added under X86_64 sync_regs : added under X86_64 math_error : moved out from X86_32 as it is common for both 32 and 64 bit math_emulate : moved from X86_32 as it is common for both 32 and 64 bit smp_thermal_interrupt : added under X86_64 mce_threshold_interrupt : added under X86_64 Signed-off-by: Jaswinder Singh Signed-off-by: Ingo Molnar --- arch/x86/include/asm/traps.h | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/traps.h b/arch/x86/include/asm/traps.h index 45dee286e45c..2ee0a3bceedf 100644 --- a/arch/x86/include/asm/traps.h +++ b/arch/x86/include/asm/traps.h @@ -46,6 +46,10 @@ dotraplinkage void do_coprocessor_segment_overrun(struct pt_regs *, long); dotraplinkage void do_invalid_TSS(struct pt_regs *, long); dotraplinkage void do_segment_not_present(struct pt_regs *, long); dotraplinkage void do_stack_segment(struct pt_regs *, long); +#ifdef CONFIG_X86_64 +dotraplinkage void do_double_fault(struct pt_regs *, long); +asmlinkage __kprobes struct pt_regs *sync_regs(struct pt_regs *); +#endif dotraplinkage void do_general_protection(struct pt_regs *, long); dotraplinkage void do_page_fault(struct pt_regs *, unsigned long); dotraplinkage void do_spurious_interrupt_bug(struct pt_regs *, long); @@ -72,10 +76,13 @@ static inline int get_si_code(unsigned long condition) extern int panic_on_unrecovered_nmi; extern int kstack_depth_to_print; -#ifdef CONFIG_X86_32 void math_error(void __user *); -unsigned long patch_espfix_desc(unsigned long, unsigned long); asmlinkage void math_emulate(long); +#ifdef CONFIG_X86_32 +unsigned long patch_espfix_desc(unsigned long, unsigned long); +#else +asmlinkage void smp_thermal_interrupt(void); +asmlinkage void mce_threshold_interrupt(void); #endif #endif /* _ASM_X86_TRAPS_H */ -- cgit v1.2.3 From b2fa739c06931d167b6d2aa7b514ab7f30d04dc0 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Thu, 18 Dec 2008 14:43:34 -0800 Subject: x86: sigframe.h: include headers for dependency Impact: cleanup Include following headers for dependency. asm/sigcontext.h asm/siginfo.h asm/ucontext.h Signed-off-by: Hiroshi Shimamoto Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/sigframe.h | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/sigframe.h b/arch/x86/include/asm/sigframe.h index 3bd0f4276000..4e0fe26d27d3 100644 --- a/arch/x86/include/asm/sigframe.h +++ b/arch/x86/include/asm/sigframe.h @@ -1,6 +1,10 @@ #ifndef _ASM_X86_SIGFRAME_H #define _ASM_X86_SIGFRAME_H +#include +#include +#include + #ifdef CONFIG_X86_32 #define sigframe_ia32 sigframe #define rt_sigframe_ia32 rt_sigframe -- cgit v1.2.3 From 8869a2e5d3a66d5b63b948052d60cd13ede8b735 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Thu, 18 Dec 2008 14:46:52 -0800 Subject: x86: asm-offset_64: use rt_sigframe_ia32 Impact: cleanup Use rt_sigframe_ia32 instead of rt_sigframe32. Signed-off-by: Hiroshi Shimamoto Signed-off-by: H. Peter Anvin --- arch/x86/kernel/asm-offsets_64.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/asm-offsets_64.c b/arch/x86/kernel/asm-offsets_64.c index 7fcf63d22f8b..1d41d3f1edbc 100644 --- a/arch/x86/kernel/asm-offsets_64.c +++ b/arch/x86/kernel/asm-offsets_64.c @@ -20,6 +20,8 @@ #include +#include + #define __NO_STUBS 1 #undef __SYSCALL #undef _ASM_X86_UNISTD_64_H @@ -87,7 +89,7 @@ int main(void) BLANK(); #undef ENTRY DEFINE(IA32_RT_SIGFRAME_sigcontext, - offsetof (struct rt_sigframe32, uc.uc_mcontext)); + offsetof (struct rt_sigframe_ia32, uc.uc_mcontext)); BLANK(); #endif DEFINE(pbe_address, offsetof(struct pbe, address)); -- cgit v1.2.3 From 9f221495997d180df51ce4d8296669445dd3e7b3 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Thu, 18 Dec 2008 14:47:37 -0800 Subject: x86: ia32.h: remove unused struct sigfram32 and rt_sigframe32 Impact: cleanup Remove struct sigfram32 and rt_sigframe32 because there is no user. Signed-off-by: Hiroshi Shimamoto Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/ia32.h | 18 ------------------ 1 file changed, 18 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/ia32.h b/arch/x86/include/asm/ia32.h index 97989c0e534c..50ca486fd88c 100644 --- a/arch/x86/include/asm/ia32.h +++ b/arch/x86/include/asm/ia32.h @@ -129,24 +129,6 @@ typedef struct compat_siginfo { } _sifields; } compat_siginfo_t; -struct sigframe32 { - u32 pretcode; - int sig; - struct sigcontext_ia32 sc; - struct _fpstate_ia32 fpstate; - unsigned int extramask[_COMPAT_NSIG_WORDS-1]; -}; - -struct rt_sigframe32 { - u32 pretcode; - int sig; - u32 pinfo; - u32 puc; - compat_siginfo_t info; - struct ucontext_ia32 uc; - struct _fpstate_ia32 fpstate; -}; - struct ustat32 { __u32 f_tfree; compat_ino_t f_tinode; -- cgit v1.2.3 From f34a10bd9f8cc95ebdc69a079db195636b2e22e0 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 19 Dec 2008 01:36:14 +0100 Subject: x86: fix warning in arch/x86/kernel/microcode_amd.c MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit this warning: arch/x86/kernel/microcode_amd.c: In function ‘apply_microcode_amd’: arch/x86/kernel/microcode_amd.c:163: warning: cast from pointer to integer of different size arch/x86/kernel/microcode_amd.c:163: warning: cast from pointer to integer of different size triggers because we want to pass the address to the microcode MSR, which is 64-bit even on 32-bit. Cast it explicitly to express this. Signed-off-by: Ingo Molnar --- arch/x86/kernel/microcode_amd.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/microcode_amd.c b/arch/x86/kernel/microcode_amd.c index 24c256f4e50a..c25fdb382292 100644 --- a/arch/x86/kernel/microcode_amd.c +++ b/arch/x86/kernel/microcode_amd.c @@ -160,7 +160,7 @@ static void apply_microcode_amd(int cpu) return; spin_lock_irqsave(µcode_update_lock, flags); - wrmsrl(MSR_AMD64_PATCH_LOADER, &mc_amd->hdr.data_code); + wrmsrl(MSR_AMD64_PATCH_LOADER, (u64)(long)&mc_amd->hdr.data_code); /* get patch id after patching */ rdmsr(MSR_AMD64_PATCH_LEVEL, rev, dummy); spin_unlock_irqrestore(µcode_update_lock, flags); @@ -372,3 +372,4 @@ struct microcode_ops * __init init_amd_microcode(void) { return µcode_amd_ops; } + -- cgit v1.2.3 From 345077cd98ff5532b2d1158013c3fec7b1ae85ec Mon Sep 17 00:00:00 2001 From: Suresh Siddha Date: Thu, 18 Dec 2008 18:09:21 -0800 Subject: x86: fix intel x86_64 llc_shared_map/cpu_llc_id anomolies Impact: fix wrong cache sharing detection on platforms supporting > 8 bit apicid's In the presence of extended topology eumeration leaf 0xb provided by cpuid, 32bit extended initial_apicid in cpuinfo_x86 struct will be updated by detect_extended_topology(). At this instance, we should also reinit the apicid (which could also potentially be extended to 32bit). With out this there will potentially be duplicate apicid's populated in the per cpu's cpuinfo_x86 struct, resulting in wrong cache sharing topology etc detected by init_intel_cacheinfo(). Reported-by: Dimitri Sivanich Signed-off-by: Suresh Siddha Acked-by: Dimitri Sivanich Signed-off-by: Ingo Molnar Cc: --- arch/x86/kernel/cpu/addon_cpuid_features.c | 8 ++++++++ arch/x86/kernel/cpu/intel.c | 8 +++++++- 2 files changed, 15 insertions(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/addon_cpuid_features.c b/arch/x86/kernel/cpu/addon_cpuid_features.c index ef8f831af823..2cf23634b6d9 100644 --- a/arch/x86/kernel/cpu/addon_cpuid_features.c +++ b/arch/x86/kernel/cpu/addon_cpuid_features.c @@ -120,9 +120,17 @@ void __cpuinit detect_extended_topology(struct cpuinfo_x86 *c) c->cpu_core_id = phys_pkg_id(c->initial_apicid, ht_mask_width) & core_select_mask; c->phys_proc_id = phys_pkg_id(c->initial_apicid, core_plus_mask_width); + /* + * Reinit the apicid, now that we have extended initial_apicid. + */ + c->apicid = phys_pkg_id(c->initial_apicid, 0); #else c->cpu_core_id = phys_pkg_id(ht_mask_width) & core_select_mask; c->phys_proc_id = phys_pkg_id(core_plus_mask_width); + /* + * Reinit the apicid, now that we have extended initial_apicid. + */ + c->apicid = phys_pkg_id(0); #endif c->x86_max_cores = (core_level_siblings / smp_num_siblings); diff --git a/arch/x86/kernel/cpu/intel.c b/arch/x86/kernel/cpu/intel.c index caec59437a22..b21c37c060a2 100644 --- a/arch/x86/kernel/cpu/intel.c +++ b/arch/x86/kernel/cpu/intel.c @@ -252,6 +252,13 @@ static void __cpuinit init_intel(struct cpuinfo_x86 *c) intel_workarounds(c); + /* + * Detect the extended topology information if available. This + * will reinitialise the initial_apicid which will be used + * in init_intel_cacheinfo() + */ + detect_extended_topology(c); + l2 = init_intel_cacheinfo(c); if (c->cpuid_level > 9) { unsigned eax = cpuid_eax(10); @@ -323,7 +330,6 @@ static void __cpuinit init_intel(struct cpuinfo_x86 *c) #endif - detect_extended_topology(c); if (!cpu_has(c, X86_FEATURE_XTOPOLOGY)) { /* * let's use the legacy cpuid vector 0x1 and 0x4 for topology -- cgit v1.2.3 From a7883dece6ef82097e6bdf19c1d0a20351e06056 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 19 Dec 2008 00:59:09 +0100 Subject: x86: fix warning in arch/x86/kernel/io_apic.c MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit this warning: arch/x86/kernel/io_apic.c: In function ‘ir_set_msi_irq_affinity’: arch/x86/kernel/io_apic.c:3373: warning: ‘cfg’ may be used uninitialized in this function triggers because the variable was truly uninitialized. We'd crash on entering this code. Signed-off-by: Ingo Molnar --- arch/x86/kernel/io_apic.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index 58938cc4b7d3..908c1d00a5c7 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -3360,7 +3360,7 @@ static void ir_set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) { struct irq_desc *desc = irq_to_desc(irq); - struct irq_cfg *cfg; + struct irq_cfg *cfg = desc->chip_data; unsigned int dest; struct irte irte; -- cgit v1.2.3 From 9bb482476c6c9d1ae033306440c51ceac93ea80c Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Tue, 16 Dec 2008 11:30:08 +0000 Subject: allow stripping of generated symbols under CONFIG_KALLSYMS_ALL Building upon parts of the module stripping patch, this patch introduces similar stripping for vmlinux when CONFIG_KALLSYMS_ALL=y. Using CONFIG_KALLSYMS_STRIP_GENERATED reduces the overhead of CONFIG_KALLSYMS_ALL from 245k/310k to 65k/80k for the (i386/x86-64) kernels I tested with. The patch also does away with the need to special case the kallsyms- internal symbols by making them available even in the first linking stage. While it is a generated file, the patch includes the changes to scripts/genksyms/keywords.c_shipped, as I'm unsure what the procedure here is. Signed-off-by: Jan Beulich Signed-off-by: Sam Ravnborg --- Makefile | 47 ++++++--- arch/x86/scripts/strip-symbols | 1 + init/Kconfig | 7 ++ kernel/kallsyms.c | 16 ++- scripts/genksyms/keywords.c_shipped | 189 ++++++++++++++++++------------------ scripts/genksyms/keywords.gperf | 2 + scripts/kallsyms.c | 21 ++-- 7 files changed, 155 insertions(+), 128 deletions(-) create mode 100644 arch/x86/scripts/strip-symbols (limited to 'arch/x86') diff --git a/Makefile b/Makefile index 5dd0ed3b12c6..b3d1c8f1f4ce 100644 --- a/Makefile +++ b/Makefile @@ -604,6 +604,9 @@ export INSTALL_PATH ?= /boot MODLIB = $(INSTALL_MOD_PATH)/lib/modules/$(KERNELRELEASE) export MODLIB +strip-symbols := $(srctree)/scripts/strip-symbols \ + $(wildcard $(srctree)/arch/$(ARCH)/scripts/strip-symbols) + # # INSTALL_MOD_STRIP, if defined, will cause modules to be stripped while # they get installed. If INSTALL_MOD_STRIP is '1', then the default @@ -611,8 +614,10 @@ export MODLIB # be used as the option(s) to the objcopy command. ifdef INSTALL_MOD_STRIP ifeq ($(INSTALL_MOD_STRIP),1) -mod_strip_cmd = $(OBJCOPY) --strip-debug --strip-symbols \ - $(srctree)/scripts/strip-symbols --wildcard +mod_strip_cmd = $(OBJCOPY) --strip-debug +ifeq ($(CONFIG_KALLSYMS_ALL),$(CONFIG_KALLSYMS_STRIP_GENERATED)) +mod_strip_cmd += --wildcard $(addprefix --strip-symbols ,$(strip-symbols)) +endif else mod_strip_cmd = $(OBJCOPY) $(INSTALL_MOD_STRIP) endif # INSTALL_MOD_STRIP=1 @@ -747,6 +752,7 @@ last_kallsyms := 2 endif kallsyms.o := .tmp_kallsyms$(last_kallsyms).o +kallsyms.h := $(wildcard include/config/kallsyms/*.h) $(wildcard include/config/kallsyms/*/*.h) define verify_kallsyms $(Q)$(if $($(quiet)cmd_sysmap), \ @@ -771,24 +777,41 @@ endef # Generate .S file with all kernel symbols quiet_cmd_kallsyms = KSYM $@ - cmd_kallsyms = $(NM) -n $< | $(KALLSYMS) \ - $(if $(CONFIG_KALLSYMS_ALL),--all-symbols) > $@ + cmd_kallsyms = { test $* -eq 0 || $(NM) -n $<; } \ + | $(KALLSYMS) $(if $(CONFIG_KALLSYMS_ALL),--all-symbols) >$@ -.tmp_kallsyms1.o .tmp_kallsyms2.o .tmp_kallsyms3.o: %.o: %.S scripts FORCE +quiet_cmd_kstrip = STRIP $@ + cmd_kstrip = $(OBJCOPY) --wildcard $(addprefix --strip$(if $(CONFIG_RELOCATABLE),-unneeded)-symbols ,$(filter %/scripts/strip-symbols,$^)) $< $@ + +$(foreach n,0 1 2 3,.tmp_kallsyms$(n).o): KBUILD_AFLAGS += -Wa,--strip-local-absolute +$(foreach n,0 1 2 3,.tmp_kallsyms$(n).o): %.o: %.S scripts FORCE $(call if_changed_dep,as_o_S) -.tmp_kallsyms%.S: .tmp_vmlinux% $(KALLSYMS) +ifeq ($(CONFIG_KALLSYMS_STRIP_GENERATED),y) +strip-ext := .stripped +endif + +.tmp_kallsyms%.S: .tmp_vmlinux%$(strip-ext) $(KALLSYMS) $(kallsyms.h) $(call cmd,kallsyms) +# make -jN seems to have problems with intermediate files, see bug #3330. +.SECONDARY: $(foreach n,1 2 3,.tmp_vmlinux$(n).stripped) +.tmp_vmlinux%.stripped: .tmp_vmlinux% $(strip-symbols) $(kallsyms.h) + $(call cmd,kstrip) + +ifneq ($(CONFIG_DEBUG_INFO),y) +.tmp_vmlinux%: LDFLAGS_vmlinux += -S +endif # .tmp_vmlinux1 must be complete except kallsyms, so update vmlinux version -.tmp_vmlinux1: $(vmlinux-lds) $(vmlinux-all) FORCE - $(call if_changed_rule,ksym_ld) +.tmp_vmlinux%: $(vmlinux-lds) $(vmlinux-all) FORCE + $(if $(filter 1,$*),$(call if_changed_rule,ksym_ld),$(call if_changed,vmlinux__)) -.tmp_vmlinux2: $(vmlinux-lds) $(vmlinux-all) .tmp_kallsyms1.o FORCE - $(call if_changed,vmlinux__) +.tmp_vmlinux0$(strip-ext): + $(Q)echo "placeholder" >$@ -.tmp_vmlinux3: $(vmlinux-lds) $(vmlinux-all) .tmp_kallsyms2.o FORCE - $(call if_changed,vmlinux__) +.tmp_vmlinux1: .tmp_kallsyms0.o +.tmp_vmlinux2: .tmp_kallsyms1.o +.tmp_vmlinux3: .tmp_kallsyms2.o # Needs to visit scripts/ before $(KALLSYMS) can be used. $(KALLSYMS): scripts ; diff --git a/arch/x86/scripts/strip-symbols b/arch/x86/scripts/strip-symbols new file mode 100644 index 000000000000..a2f1ccb827c7 --- /dev/null +++ b/arch/x86/scripts/strip-symbols @@ -0,0 +1 @@ +__cpu_vendor_dev_X86_VENDOR_* diff --git a/init/Kconfig b/init/Kconfig index f763762d544a..0f5af409fef1 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -588,6 +588,13 @@ config KALLSYMS_ALL Say N. +config KALLSYMS_STRIP_GENERATED + bool "Strip machine generated symbols from kallsyms" + depends on KALLSYMS_ALL + default y + help + Say N if you want kallsyms to retain even machine generated symbols. + config KALLSYMS_EXTRA_PASS bool "Do an extra kallsyms pass" depends on KALLSYMS diff --git a/kernel/kallsyms.c b/kernel/kallsyms.c index 7b8b0f21a5b1..e694afa0eb8c 100644 --- a/kernel/kallsyms.c +++ b/kernel/kallsyms.c @@ -30,20 +30,19 @@ #define all_var 0 #endif -/* These will be re-linked against their real values during the second link stage */ -extern const unsigned long kallsyms_addresses[] __attribute__((weak)); -extern const u8 kallsyms_names[] __attribute__((weak)); +extern const unsigned long kallsyms_addresses[]; +extern const u8 kallsyms_names[]; /* tell the compiler that the count isn't in the small data section if the arch * has one (eg: FRV) */ extern const unsigned long kallsyms_num_syms -__attribute__((weak, section(".rodata"))); + __attribute__((__section__(".rodata"))); -extern const u8 kallsyms_token_table[] __attribute__((weak)); -extern const u16 kallsyms_token_index[] __attribute__((weak)); +extern const u8 kallsyms_token_table[]; +extern const u16 kallsyms_token_index[]; -extern const unsigned long kallsyms_markers[] __attribute__((weak)); +extern const unsigned long kallsyms_markers[]; static inline int is_kernel_inittext(unsigned long addr) { @@ -168,9 +167,6 @@ static unsigned long get_symbol_pos(unsigned long addr, unsigned long symbol_start = 0, symbol_end = 0; unsigned long i, low, high, mid; - /* This kernel should never had been booted. */ - BUG_ON(!kallsyms_addresses); - /* do a binary search on the sorted kallsyms_addresses array */ low = 0; high = kallsyms_num_syms; diff --git a/scripts/genksyms/keywords.c_shipped b/scripts/genksyms/keywords.c_shipped index 971e0113ae7a..83484fe93ede 100644 --- a/scripts/genksyms/keywords.c_shipped +++ b/scripts/genksyms/keywords.c_shipped @@ -1,4 +1,4 @@ -/* ANSI-C code produced by gperf version 3.0.2 */ +/* ANSI-C code produced by gperf version 3.0.1 */ /* Command-line: gperf -L ANSI-C -a -C -E -g -H is_reserved_hash -k '1,3,$' -N is_reserved_word -p -t scripts/genksyms/keywords.gperf */ #if !((' ' == 32) && ('!' == 33) && ('"' == 34) && ('#' == 35) \ @@ -32,7 +32,7 @@ #line 3 "scripts/genksyms/keywords.gperf" struct resword { const char *name; int token; }; -/* maximum key range = 62, duplicates = 0 */ +/* maximum key range = 64, duplicates = 0 */ #ifdef __GNUC__ __inline @@ -46,32 +46,32 @@ is_reserved_hash (register const char *str, register unsigned int len) { static const unsigned char asso_values[] = { - 65, 65, 65, 65, 65, 65, 65, 65, 65, 65, - 65, 65, 65, 65, 65, 65, 65, 65, 65, 65, - 65, 65, 65, 65, 65, 65, 65, 65, 65, 65, - 65, 65, 65, 65, 65, 65, 65, 65, 65, 65, - 65, 65, 65, 65, 65, 65, 65, 65, 65, 65, - 65, 65, 65, 65, 65, 65, 65, 65, 65, 65, - 65, 65, 65, 65, 65, 65, 65, 65, 65, 5, - 65, 65, 65, 65, 65, 65, 35, 65, 65, 65, - 0, 65, 65, 65, 65, 65, 65, 65, 65, 65, - 65, 65, 65, 65, 65, 0, 65, 0, 65, 5, - 20, 15, 10, 30, 65, 15, 65, 65, 20, 0, - 10, 35, 20, 65, 10, 5, 0, 10, 5, 65, - 65, 65, 65, 65, 65, 65, 65, 65, 65, 65, - 65, 65, 65, 65, 65, 65, 65, 65, 65, 65, - 65, 65, 65, 65, 65, 65, 65, 65, 65, 65, - 65, 65, 65, 65, 65, 65, 65, 65, 65, 65, - 65, 65, 65, 65, 65, 65, 65, 65, 65, 65, - 65, 65, 65, 65, 65, 65, 65, 65, 65, 65, - 65, 65, 65, 65, 65, 65, 65, 65, 65, 65, - 65, 65, 65, 65, 65, 65, 65, 65, 65, 65, - 65, 65, 65, 65, 65, 65, 65, 65, 65, 65, - 65, 65, 65, 65, 65, 65, 65, 65, 65, 65, - 65, 65, 65, 65, 65, 65, 65, 65, 65, 65, - 65, 65, 65, 65, 65, 65, 65, 65, 65, 65, - 65, 65, 65, 65, 65, 65, 65, 65, 65, 65, - 65, 65, 65, 65, 65, 65 + 67, 67, 67, 67, 67, 67, 67, 67, 67, 67, + 67, 67, 67, 67, 67, 67, 67, 67, 67, 67, + 67, 67, 67, 67, 67, 67, 67, 67, 67, 67, + 67, 67, 67, 67, 67, 67, 67, 67, 67, 67, + 67, 67, 67, 67, 67, 67, 67, 67, 67, 67, + 67, 67, 67, 67, 67, 67, 67, 67, 67, 67, + 67, 67, 67, 67, 67, 67, 67, 67, 67, 0, + 67, 67, 67, 67, 67, 67, 15, 67, 67, 67, + 0, 67, 67, 67, 67, 67, 67, 67, 67, 67, + 67, 67, 67, 67, 67, 0, 67, 0, 67, 5, + 25, 20, 15, 30, 67, 15, 67, 67, 10, 0, + 10, 40, 20, 67, 10, 5, 0, 10, 15, 67, + 67, 67, 67, 67, 67, 67, 67, 67, 67, 67, + 67, 67, 67, 67, 67, 67, 67, 67, 67, 67, + 67, 67, 67, 67, 67, 67, 67, 67, 67, 67, + 67, 67, 67, 67, 67, 67, 67, 67, 67, 67, + 67, 67, 67, 67, 67, 67, 67, 67, 67, 67, + 67, 67, 67, 67, 67, 67, 67, 67, 67, 67, + 67, 67, 67, 67, 67, 67, 67, 67, 67, 67, + 67, 67, 67, 67, 67, 67, 67, 67, 67, 67, + 67, 67, 67, 67, 67, 67, 67, 67, 67, 67, + 67, 67, 67, 67, 67, 67, 67, 67, 67, 67, + 67, 67, 67, 67, 67, 67, 67, 67, 67, 67, + 67, 67, 67, 67, 67, 67, 67, 67, 67, 67, + 67, 67, 67, 67, 67, 67, 67, 67, 67, 67, + 67, 67, 67, 67, 67, 67 }; return len + asso_values[(unsigned char)str[2]] + asso_values[(unsigned char)str[0]] + asso_values[(unsigned char)str[len - 1]]; } @@ -84,116 +84,119 @@ is_reserved_word (register const char *str, register unsigned int len) { enum { - TOTAL_KEYWORDS = 43, + TOTAL_KEYWORDS = 45, MIN_WORD_LENGTH = 3, MAX_WORD_LENGTH = 24, MIN_HASH_VALUE = 3, - MAX_HASH_VALUE = 64 + MAX_HASH_VALUE = 66 }; static const struct resword wordlist[] = { {""}, {""}, {""}, -#line 26 "scripts/genksyms/keywords.gperf" +#line 28 "scripts/genksyms/keywords.gperf" {"asm", ASM_KEYW}, {""}, -#line 8 "scripts/genksyms/keywords.gperf" +#line 10 "scripts/genksyms/keywords.gperf" {"__asm", ASM_KEYW}, {""}, -#line 9 "scripts/genksyms/keywords.gperf" +#line 11 "scripts/genksyms/keywords.gperf" {"__asm__", ASM_KEYW}, {""}, {""}, -#line 52 "scripts/genksyms/keywords.gperf" +#line 54 "scripts/genksyms/keywords.gperf" {"__typeof__", TYPEOF_KEYW}, {""}, -#line 12 "scripts/genksyms/keywords.gperf" +#line 14 "scripts/genksyms/keywords.gperf" {"__const", CONST_KEYW}, -#line 11 "scripts/genksyms/keywords.gperf" - {"__attribute__", ATTRIBUTE_KEYW}, #line 13 "scripts/genksyms/keywords.gperf" + {"__attribute__", ATTRIBUTE_KEYW}, +#line 15 "scripts/genksyms/keywords.gperf" {"__const__", CONST_KEYW}, -#line 18 "scripts/genksyms/keywords.gperf" +#line 20 "scripts/genksyms/keywords.gperf" {"__signed__", SIGNED_KEYW}, -#line 44 "scripts/genksyms/keywords.gperf" +#line 46 "scripts/genksyms/keywords.gperf" {"static", STATIC_KEYW}, -#line 20 "scripts/genksyms/keywords.gperf" - {"__volatile__", VOLATILE_KEYW}, -#line 39 "scripts/genksyms/keywords.gperf" + {""}, +#line 41 "scripts/genksyms/keywords.gperf" {"int", INT_KEYW}, -#line 32 "scripts/genksyms/keywords.gperf" +#line 34 "scripts/genksyms/keywords.gperf" {"char", CHAR_KEYW}, -#line 33 "scripts/genksyms/keywords.gperf" +#line 35 "scripts/genksyms/keywords.gperf" {"const", CONST_KEYW}, -#line 45 "scripts/genksyms/keywords.gperf" +#line 47 "scripts/genksyms/keywords.gperf" {"struct", STRUCT_KEYW}, -#line 24 "scripts/genksyms/keywords.gperf" +#line 26 "scripts/genksyms/keywords.gperf" {"__restrict__", RESTRICT_KEYW}, -#line 25 "scripts/genksyms/keywords.gperf" +#line 27 "scripts/genksyms/keywords.gperf" {"restrict", RESTRICT_KEYW}, -#line 23 "scripts/genksyms/keywords.gperf" - {"_restrict", RESTRICT_KEYW}, -#line 16 "scripts/genksyms/keywords.gperf" +#line 7 "scripts/genksyms/keywords.gperf" + {"EXPORT_SYMBOL_GPL_FUTURE", EXPORT_SYMBOL_KEYW}, +#line 18 "scripts/genksyms/keywords.gperf" {"__inline__", INLINE_KEYW}, -#line 10 "scripts/genksyms/keywords.gperf" - {"__attribute", ATTRIBUTE_KEYW}, {""}, -#line 14 "scripts/genksyms/keywords.gperf" +#line 22 "scripts/genksyms/keywords.gperf" + {"__volatile__", VOLATILE_KEYW}, +#line 5 "scripts/genksyms/keywords.gperf" + {"EXPORT_SYMBOL", EXPORT_SYMBOL_KEYW}, +#line 25 "scripts/genksyms/keywords.gperf" + {"_restrict", RESTRICT_KEYW}, + {""}, +#line 12 "scripts/genksyms/keywords.gperf" + {"__attribute", ATTRIBUTE_KEYW}, +#line 6 "scripts/genksyms/keywords.gperf" + {"EXPORT_SYMBOL_GPL", EXPORT_SYMBOL_KEYW}, +#line 16 "scripts/genksyms/keywords.gperf" {"__extension__", EXTENSION_KEYW}, -#line 35 "scripts/genksyms/keywords.gperf" +#line 37 "scripts/genksyms/keywords.gperf" {"enum", ENUM_KEYW}, -#line 19 "scripts/genksyms/keywords.gperf" - {"__volatile", VOLATILE_KEYW}, -#line 36 "scripts/genksyms/keywords.gperf" +#line 8 "scripts/genksyms/keywords.gperf" + {"EXPORT_UNUSED_SYMBOL", EXPORT_SYMBOL_KEYW}, +#line 38 "scripts/genksyms/keywords.gperf" {"extern", EXTERN_KEYW}, {""}, -#line 17 "scripts/genksyms/keywords.gperf" +#line 19 "scripts/genksyms/keywords.gperf" {"__signed", SIGNED_KEYW}, -#line 7 "scripts/genksyms/keywords.gperf" - {"EXPORT_SYMBOL_GPL_FUTURE", EXPORT_SYMBOL_KEYW}, - {""}, -#line 51 "scripts/genksyms/keywords.gperf" +#line 9 "scripts/genksyms/keywords.gperf" + {"EXPORT_UNUSED_SYMBOL_GPL", EXPORT_SYMBOL_KEYW}, +#line 49 "scripts/genksyms/keywords.gperf" + {"union", UNION_KEYW}, +#line 53 "scripts/genksyms/keywords.gperf" {"typeof", TYPEOF_KEYW}, -#line 46 "scripts/genksyms/keywords.gperf" +#line 48 "scripts/genksyms/keywords.gperf" {"typedef", TYPEDEF_KEYW}, -#line 15 "scripts/genksyms/keywords.gperf" +#line 17 "scripts/genksyms/keywords.gperf" {"__inline", INLINE_KEYW}, -#line 31 "scripts/genksyms/keywords.gperf" +#line 33 "scripts/genksyms/keywords.gperf" {"auto", AUTO_KEYW}, -#line 47 "scripts/genksyms/keywords.gperf" - {"union", UNION_KEYW}, - {""}, {""}, -#line 48 "scripts/genksyms/keywords.gperf" - {"unsigned", UNSIGNED_KEYW}, -#line 49 "scripts/genksyms/keywords.gperf" - {"void", VOID_KEYW}, -#line 42 "scripts/genksyms/keywords.gperf" - {"short", SHORT_KEYW}, +#line 21 "scripts/genksyms/keywords.gperf" + {"__volatile", VOLATILE_KEYW}, {""}, {""}, #line 50 "scripts/genksyms/keywords.gperf" - {"volatile", VOLATILE_KEYW}, - {""}, -#line 37 "scripts/genksyms/keywords.gperf" - {"float", FLOAT_KEYW}, -#line 34 "scripts/genksyms/keywords.gperf" - {"double", DOUBLE_KEYW}, + {"unsigned", UNSIGNED_KEYW}, {""}, -#line 5 "scripts/genksyms/keywords.gperf" - {"EXPORT_SYMBOL", EXPORT_SYMBOL_KEYW}, - {""}, {""}, -#line 38 "scripts/genksyms/keywords.gperf" +#line 44 "scripts/genksyms/keywords.gperf" + {"short", SHORT_KEYW}, +#line 40 "scripts/genksyms/keywords.gperf" {"inline", INLINE_KEYW}, -#line 6 "scripts/genksyms/keywords.gperf" - {"EXPORT_SYMBOL_GPL", EXPORT_SYMBOL_KEYW}, -#line 41 "scripts/genksyms/keywords.gperf" - {"register", REGISTER_KEYW}, {""}, -#line 22 "scripts/genksyms/keywords.gperf" +#line 52 "scripts/genksyms/keywords.gperf" + {"volatile", VOLATILE_KEYW}, +#line 42 "scripts/genksyms/keywords.gperf" + {"long", LONG_KEYW}, +#line 24 "scripts/genksyms/keywords.gperf" {"_Bool", BOOL_KEYW}, -#line 43 "scripts/genksyms/keywords.gperf" - {"signed", SIGNED_KEYW}, {""}, {""}, -#line 40 "scripts/genksyms/keywords.gperf" - {"long", LONG_KEYW} +#line 43 "scripts/genksyms/keywords.gperf" + {"register", REGISTER_KEYW}, +#line 51 "scripts/genksyms/keywords.gperf" + {"void", VOID_KEYW}, +#line 39 "scripts/genksyms/keywords.gperf" + {"float", FLOAT_KEYW}, +#line 36 "scripts/genksyms/keywords.gperf" + {"double", DOUBLE_KEYW}, + {""}, {""}, {""}, {""}, +#line 45 "scripts/genksyms/keywords.gperf" + {"signed", SIGNED_KEYW} }; if (len <= MAX_WORD_LENGTH && len >= MIN_WORD_LENGTH) diff --git a/scripts/genksyms/keywords.gperf b/scripts/genksyms/keywords.gperf index 5ef3733225fb..8abe7ab8d88f 100644 --- a/scripts/genksyms/keywords.gperf +++ b/scripts/genksyms/keywords.gperf @@ -5,6 +5,8 @@ struct resword { const char *name; int token; } EXPORT_SYMBOL, EXPORT_SYMBOL_KEYW EXPORT_SYMBOL_GPL, EXPORT_SYMBOL_KEYW EXPORT_SYMBOL_GPL_FUTURE, EXPORT_SYMBOL_KEYW +EXPORT_UNUSED_SYMBOL, EXPORT_SYMBOL_KEYW +EXPORT_UNUSED_SYMBOL_GPL, EXPORT_SYMBOL_KEYW __asm, ASM_KEYW __asm__, ASM_KEYW __attribute, ATTRIBUTE_KEYW diff --git a/scripts/kallsyms.c b/scripts/kallsyms.c index ad2434b26970..92758120a767 100644 --- a/scripts/kallsyms.c +++ b/scripts/kallsyms.c @@ -130,18 +130,9 @@ static int read_symbol(FILE *in, struct sym_entry *s) static int symbol_valid(struct sym_entry *s) { /* Symbols which vary between passes. Passes 1 and 2 must have - * identical symbol lists. The kallsyms_* symbols below are only added - * after pass 1, they would be included in pass 2 when --all-symbols is - * specified so exclude them to get a stable symbol list. + * identical symbol lists. */ static char *special_symbols[] = { - "kallsyms_addresses", - "kallsyms_num_syms", - "kallsyms_names", - "kallsyms_markers", - "kallsyms_token_table", - "kallsyms_token_index", - /* Exclude linker generated symbols which vary between passes */ "_SDA_BASE_", /* ppc */ "_SDA2_BASE_", /* ppc */ @@ -173,7 +164,9 @@ static int symbol_valid(struct sym_entry *s) } /* Exclude symbols which vary between passes. */ - if (strstr((char *)s->sym + offset, "_compiled.")) + if (strstr((char *)s->sym + offset, "_compiled.") || + strncmp((char*)s->sym + offset, "__compound_literal.", 19) == 0 || + strncmp((char*)s->sym + offset, "__compound_literal$", 19) == 0) return 0; for (i = 0; special_symbols[i]; i++) @@ -550,8 +543,10 @@ int main(int argc, char **argv) usage(); read_map(stdin); - sort_symbols(); - optimize_token_table(); + if (table_cnt) { + sort_symbols(); + optimize_token_table(); + } write_src(); return 0; -- cgit v1.2.3 From b909895739427874c089bc0e03dc119f99cab2dd Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Fri, 19 Dec 2008 13:48:34 -0800 Subject: sparseirq: fix numa_migrate_irq_desc dependency and comments Impact: reduce kconfig variable scope and clean up Bartlomiej pointed out that the config dependencies and comments are not right. update it depend to NUMA, and fix some comments Reported-by: Bartlomiej Zolnierkiewicz Signed-off-by: Yinghai Lu Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 2 +- arch/x86/kernel/io_apic.c | 2 +- kernel/irq/numa_migrate.c | 11 +++-------- 3 files changed, 5 insertions(+), 10 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 60a008857a38..5c243826334a 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -250,7 +250,7 @@ config SPARSE_IRQ config NUMA_MIGRATE_IRQ_DESC bool "Move irq desc when changing irq smp_affinity" - depends on SPARSE_IRQ && SMP + depends on SPARSE_IRQ && NUMA default n help This enables moving irq_desc to cpu/node that irq will use handled. diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index bfe1245b1a3e..a74887b416cc 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -2471,7 +2471,7 @@ static void irq_complete_move(struct irq_desc **descp) if (likely(!cfg->move_desc_pending)) return; - /* domain is not change, but affinity is changed */ + /* domain has not changed, but affinity did */ me = smp_processor_id(); if (cpu_isset(me, desc->affinity)) { *descp = desc = move_irq_desc(desc, me); diff --git a/kernel/irq/numa_migrate.c b/kernel/irq/numa_migrate.c index 0178e2296990..089c3746358a 100644 --- a/kernel/irq/numa_migrate.c +++ b/kernel/irq/numa_migrate.c @@ -1,13 +1,8 @@ /* - * linux/kernel/irq/handle.c - * - * Copyright (C) 1992, 1998-2006 Linus Torvalds, Ingo Molnar - * Copyright (C) 2005-2006, Thomas Gleixner, Russell King - * - * This file contains the core interrupt handling code. - * - * Detailed information is available in Documentation/DocBook/genericirq + * NUMA irq-desc migration code * + * Migrate IRQ data structures (irq_desc, chip_data, etc.) over to + * the new "home node" of the IRQ. */ #include -- cgit v1.2.3 From 34945ede31071ac7d72270cc6c1893323f392b3f Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Date: Fri, 19 Dec 2008 22:33:52 +0530 Subject: x86: common.c boot_cpu_stack and boot_exception_stacks should be static Impact: cleanup, avoid sparse warnings, reduce kernel size a bit Fixes these sparse warnings: arch/x86/kernel/cpu/common.c:869:6: warning: symbol 'boot_cpu_stack' was not declared. Should it be static? arch/x86/kernel/cpu/common.c:910:6: warning: symbol 'boot_exception_stacks' was not declared. Should it be static? Signed-off-by: Jaswinder Singh Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/common.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index b9c9ea0217a9..aba49c782fd6 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -862,7 +862,7 @@ EXPORT_SYMBOL(_cpu_pda); struct desc_ptr idt_descr = { 256 * 16 - 1, (unsigned long) idt_table }; -char boot_cpu_stack[IRQSTACKSIZE] __page_aligned_bss; +static char boot_cpu_stack[IRQSTACKSIZE] __page_aligned_bss; void __cpuinit pda_init(int cpu) { @@ -903,8 +903,8 @@ void __cpuinit pda_init(int cpu) } } -char boot_exception_stacks[(N_EXCEPTION_STACKS - 1) * EXCEPTION_STKSZ + - DEBUG_STKSZ] __page_aligned_bss; +static char boot_exception_stacks[(N_EXCEPTION_STACKS - 1) * EXCEPTION_STKSZ + + DEBUG_STKSZ] __page_aligned_bss; extern asmlinkage void ignore_sysret(void); -- cgit v1.2.3 From 8403295e0fa460f6240e2d781e25dc29189f33c7 Mon Sep 17 00:00:00 2001 From: Hiroshi Shimamoto Date: Fri, 19 Dec 2008 14:25:50 -0800 Subject: x86: ia32_signal: remove unnecessary declaration Impact: cleanup No need to declare do_signal(). Signed-off-by: Hiroshi Shimamoto Signed-off-by: Ingo Molnar --- arch/x86/ia32/ia32_signal.c | 1 - 1 file changed, 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/ia32/ia32_signal.c b/arch/x86/ia32/ia32_signal.c index 3b3878a63bc2..09513f8a2896 100644 --- a/arch/x86/ia32/ia32_signal.c +++ b/arch/x86/ia32/ia32_signal.c @@ -43,7 +43,6 @@ X86_EFLAGS_ZF | X86_EFLAGS_AF | X86_EFLAGS_PF | \ X86_EFLAGS_CF) -asmlinkage int do_signal(struct pt_regs *regs, sigset_t *oldset); void signal_fault(struct pt_regs *regs, void __user *frame, char *where); int copy_siginfo_to_user32(compat_siginfo_t __user *to, siginfo_t *from) -- cgit v1.2.3 From 982d789ab76c8a11426852fec2fdf2f412e21c0c Mon Sep 17 00:00:00 2001 From: "venkatesh.pallipadi@intel.com" Date: Fri, 19 Dec 2008 13:47:28 -0800 Subject: x86: PAT: remove follow_pfnmap_pte in favor of follow_phys Impact: Cleanup - removes a new function in favor of a recently modified older one. Replace follow_pfnmap_pte in pat code with follow_phys. follow_phys lso returns protection eliminating the need of pte_pgprot call. Using follow_phys also eliminates the need for pte_pa. Signed-off-by: Venkatesh Pallipadi Signed-off-by: Suresh Siddha Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/pgtable.h | 5 ----- arch/x86/mm/pat.c | 30 +++++++++++------------------ include/linux/mm.h | 3 --- mm/memory.c | 43 ------------------------------------------ 4 files changed, 11 insertions(+), 70 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h index 579f8ceee948..2aa792bbd7e0 100644 --- a/arch/x86/include/asm/pgtable.h +++ b/arch/x86/include/asm/pgtable.h @@ -230,11 +230,6 @@ static inline unsigned long pte_pfn(pte_t pte) return (pte_val(pte) & PTE_PFN_MASK) >> PAGE_SHIFT; } -static inline u64 pte_pa(pte_t pte) -{ - return pte_val(pte) & PTE_PFN_MASK; -} - #define pte_page(pte) pfn_to_page(pte_pfn(pte)) static inline int pmd_large(pmd_t pte) diff --git a/arch/x86/mm/pat.c b/arch/x86/mm/pat.c index d5254bae84f4..541bcc944a5b 100644 --- a/arch/x86/mm/pat.c +++ b/arch/x86/mm/pat.c @@ -685,8 +685,7 @@ int track_pfn_vma_copy(struct vm_area_struct *vma) int retval = 0; unsigned long i, j; u64 paddr; - pgprot_t prot; - pte_t pte; + unsigned long prot; unsigned long vma_start = vma->vm_start; unsigned long vma_end = vma->vm_end; unsigned long vma_size = vma_end - vma_start; @@ -696,26 +695,22 @@ int track_pfn_vma_copy(struct vm_area_struct *vma) if (is_linear_pfn_mapping(vma)) { /* - * reserve the whole chunk starting from vm_pgoff, - * But, we have to get the protection from pte. + * reserve the whole chunk covered by vma. We need the + * starting address and protection from pte. */ - if (follow_pfnmap_pte(vma, vma_start, &pte)) { + if (follow_phys(vma, vma_start, 0, &prot, &paddr)) { WARN_ON_ONCE(1); - return -1; + return -EINVAL; } - prot = pte_pgprot(pte); - paddr = (u64)vma->vm_pgoff << PAGE_SHIFT; - return reserve_pfn_range(paddr, vma_size, prot); + return reserve_pfn_range(paddr, vma_size, __pgprot(prot)); } /* reserve entire vma page by page, using pfn and prot from pte */ for (i = 0; i < vma_size; i += PAGE_SIZE) { - if (follow_pfnmap_pte(vma, vma_start + i, &pte)) + if (follow_phys(vma, vma_start + i, 0, &prot, &paddr)) continue; - paddr = pte_pa(pte); - prot = pte_pgprot(pte); - retval = reserve_pfn_range(paddr, PAGE_SIZE, prot); + retval = reserve_pfn_range(paddr, PAGE_SIZE, __pgprot(prot)); if (retval) goto cleanup_ret; } @@ -724,10 +719,9 @@ int track_pfn_vma_copy(struct vm_area_struct *vma) cleanup_ret: /* Reserve error: Cleanup partial reservation and return error */ for (j = 0; j < i; j += PAGE_SIZE) { - if (follow_pfnmap_pte(vma, vma_start + j, &pte)) + if (follow_phys(vma, vma_start + j, 0, &prot, &paddr)) continue; - paddr = pte_pa(pte); free_pfn_range(paddr, PAGE_SIZE); } @@ -797,6 +791,7 @@ void untrack_pfn_vma(struct vm_area_struct *vma, unsigned long pfn, { unsigned long i; u64 paddr; + unsigned long prot; unsigned long vma_start = vma->vm_start; unsigned long vma_end = vma->vm_end; unsigned long vma_size = vma_end - vma_start; @@ -821,12 +816,9 @@ void untrack_pfn_vma(struct vm_area_struct *vma, unsigned long pfn, } else { /* free entire vma, page by page, using the pfn from pte */ for (i = 0; i < vma_size; i += PAGE_SIZE) { - pte_t pte; - - if (follow_pfnmap_pte(vma, vma_start + i, &pte)) + if (follow_phys(vma, vma_start + i, 0, &prot, &paddr)) continue; - paddr = pte_pa(pte); free_pfn_range(paddr, PAGE_SIZE); } } diff --git a/include/linux/mm.h b/include/linux/mm.h index 2f6e2f886d4b..36f9b3fa5e15 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -1239,9 +1239,6 @@ struct page *follow_page(struct vm_area_struct *, unsigned long address, #define FOLL_GET 0x04 /* do get_page on page */ #define FOLL_ANON 0x08 /* give ZERO_PAGE if no pgtable */ -int follow_pfnmap_pte(struct vm_area_struct *vma, - unsigned long address, pte_t *ret_ptep); - typedef int (*pte_fn_t)(pte_t *pte, pgtable_t token, unsigned long addr, void *data); extern int apply_to_page_range(struct mm_struct *mm, unsigned long address, diff --git a/mm/memory.c b/mm/memory.c index 79f28e35d4fc..6b29f39a5a3e 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -1168,49 +1168,6 @@ no_page_table: return page; } -int follow_pfnmap_pte(struct vm_area_struct *vma, unsigned long address, - pte_t *ret_ptep) -{ - pgd_t *pgd; - pud_t *pud; - pmd_t *pmd; - pte_t *ptep, pte; - spinlock_t *ptl; - struct page *page; - struct mm_struct *mm = vma->vm_mm; - - if (!is_pfn_mapping(vma)) - goto err; - - page = NULL; - pgd = pgd_offset(mm, address); - if (pgd_none(*pgd) || unlikely(pgd_bad(*pgd))) - goto err; - - pud = pud_offset(pgd, address); - if (pud_none(*pud) || unlikely(pud_bad(*pud))) - goto err; - - pmd = pmd_offset(pud, address); - if (pmd_none(*pmd) || unlikely(pmd_bad(*pmd))) - goto err; - - ptep = pte_offset_map_lock(mm, pmd, address, &ptl); - - pte = *ptep; - if (!pte_present(pte)) - goto err_unlock; - - *ret_ptep = pte; - pte_unmap_unlock(ptep, ptl); - return 0; - -err_unlock: - pte_unmap_unlock(ptep, ptl); -err: - return -EINVAL; -} - /* Can we do the FOLL_ANON optimization? */ static inline int use_zero_page(struct vm_area_struct *vma) { -- cgit v1.2.3 From 34801ba9bf0381fcf0e2b08179d2c07f2c6ede74 Mon Sep 17 00:00:00 2001 From: "venkatesh.pallipadi@intel.com" Date: Fri, 19 Dec 2008 13:47:29 -0800 Subject: x86: PAT: move track untrack pfnmap stubs to asm-generic Impact: Cleanup and branch hints only. Move the track and untrack pfn stub routines from memory.c to asm-generic. Also add unlikely to pfnmap related calls in fork and exit path. Signed-off-by: Venkatesh Pallipadi Signed-off-by: Suresh Siddha Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/pgtable.h | 6 ++---- include/asm-generic/pgtable.h | 46 ++++++++++++++++++++++++++++++++++++++++ include/linux/mm.h | 6 ------ mm/memory.c | 48 ++---------------------------------------- 4 files changed, 50 insertions(+), 56 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h index 2aa792bbd7e0..875192bf72cb 100644 --- a/arch/x86/include/asm/pgtable.h +++ b/arch/x86/include/asm/pgtable.h @@ -339,12 +339,10 @@ static inline pgprot_t pgprot_modify(pgprot_t oldprot, pgprot_t newprot) #define canon_pgprot(p) __pgprot(pgprot_val(p) & __supported_pte_mask) +#ifndef __ASSEMBLY__ /* Indicate that x86 has its own track and untrack pfn vma functions */ -#define track_pfn_vma_new track_pfn_vma_new -#define track_pfn_vma_copy track_pfn_vma_copy -#define untrack_pfn_vma untrack_pfn_vma +#define __HAVE_PFNMAP_TRACKING -#ifndef __ASSEMBLY__ #define __HAVE_PHYS_MEM_ACCESS_PROT struct file; pgprot_t phys_mem_access_prot(struct file *file, unsigned long pfn, diff --git a/include/asm-generic/pgtable.h b/include/asm-generic/pgtable.h index b84633801fb6..72ebe91005a8 100644 --- a/include/asm-generic/pgtable.h +++ b/include/asm-generic/pgtable.h @@ -293,6 +293,52 @@ static inline void ptep_modify_prot_commit(struct mm_struct *mm, #define arch_flush_lazy_cpu_mode() do {} while (0) #endif +#ifndef __HAVE_PFNMAP_TRACKING +/* + * Interface that can be used by architecture code to keep track of + * memory type of pfn mappings (remap_pfn_range, vm_insert_pfn) + * + * track_pfn_vma_new is called when a _new_ pfn mapping is being established + * for physical range indicated by pfn and size. + */ +static inline int track_pfn_vma_new(struct vm_area_struct *vma, pgprot_t prot, + unsigned long pfn, unsigned long size) +{ + return 0; +} + +/* + * Interface that can be used by architecture code to keep track of + * memory type of pfn mappings (remap_pfn_range, vm_insert_pfn) + * + * track_pfn_vma_copy is called when vma that is covering the pfnmap gets + * copied through copy_page_range(). + */ +static inline int track_pfn_vma_copy(struct vm_area_struct *vma) +{ + return 0; +} + +/* + * Interface that can be used by architecture code to keep track of + * memory type of pfn mappings (remap_pfn_range, vm_insert_pfn) + * + * untrack_pfn_vma is called while unmapping a pfnmap for a region. + * untrack can be called for a specific region indicated by pfn and size or + * can be for the entire vma (in which case size can be zero). + */ +static inline void untrack_pfn_vma(struct vm_area_struct *vma, + unsigned long pfn, unsigned long size) +{ +} +#else +extern int track_pfn_vma_new(struct vm_area_struct *vma, pgprot_t prot, + unsigned long pfn, unsigned long size); +extern int track_pfn_vma_copy(struct vm_area_struct *vma); +extern void untrack_pfn_vma(struct vm_area_struct *vma, unsigned long pfn, + unsigned long size); +#endif + #endif /* !__ASSEMBLY__ */ #endif /* _ASM_GENERIC_PGTABLE_H */ diff --git a/include/linux/mm.h b/include/linux/mm.h index 36f9b3fa5e15..d3ddd735e375 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -163,12 +163,6 @@ static inline int is_pfn_mapping(struct vm_area_struct *vma) return (vma->vm_flags & VM_PFNMAP); } -extern int track_pfn_vma_new(struct vm_area_struct *vma, pgprot_t prot, - unsigned long pfn, unsigned long size); -extern int track_pfn_vma_copy(struct vm_area_struct *vma); -extern void untrack_pfn_vma(struct vm_area_struct *vma, unsigned long pfn, - unsigned long size); - /* * vm_fault is filled by the the pagefault handler and passed to the vma's * ->fault function. The vma's ->fault is responsible for returning a bitmask diff --git a/mm/memory.c b/mm/memory.c index 6b29f39a5a3e..f01b7eed6e16 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -99,50 +99,6 @@ int randomize_va_space __read_mostly = 2; #endif -#ifndef track_pfn_vma_new -/* - * Interface that can be used by architecture code to keep track of - * memory type of pfn mappings (remap_pfn_range, vm_insert_pfn) - * - * track_pfn_vma_new is called when a _new_ pfn mapping is being established - * for physical range indicated by pfn and size. - */ -int track_pfn_vma_new(struct vm_area_struct *vma, pgprot_t prot, - unsigned long pfn, unsigned long size) -{ - return 0; -} -#endif - -#ifndef track_pfn_vma_copy -/* - * Interface that can be used by architecture code to keep track of - * memory type of pfn mappings (remap_pfn_range, vm_insert_pfn) - * - * track_pfn_vma_copy is called when vma that is covering the pfnmap gets - * copied through copy_page_range(). - */ -int track_pfn_vma_copy(struct vm_area_struct *vma) -{ - return 0; -} -#endif - -#ifndef untrack_pfn_vma -/* - * Interface that can be used by architecture code to keep track of - * memory type of pfn mappings (remap_pfn_range, vm_insert_pfn) - * - * untrack_pfn_vma is called while unmapping a pfnmap for a region. - * untrack can be called for a specific region indicated by pfn and size or - * can be for the entire vma (in which case size can be zero). - */ -void untrack_pfn_vma(struct vm_area_struct *vma, unsigned long pfn, - unsigned long size) -{ -} -#endif - static int __init disable_randmaps(char *s) { randomize_va_space = 0; @@ -713,7 +669,7 @@ int copy_page_range(struct mm_struct *dst_mm, struct mm_struct *src_mm, if (is_vm_hugetlb_page(vma)) return copy_hugetlb_page_range(dst_mm, src_mm, vma); - if (is_pfn_mapping(vma)) { + if (unlikely(is_pfn_mapping(vma))) { /* * We do not free on error cases below as remove_vma * gets called on error from higher level routine @@ -969,7 +925,7 @@ unsigned long unmap_vmas(struct mmu_gather **tlbp, if (vma->vm_flags & VM_ACCOUNT) *nr_accounted += (end - start) >> PAGE_SHIFT; - if (is_pfn_mapping(vma)) + if (unlikely(is_pfn_mapping(vma))) untrack_pfn_vma(vma, 0, 0); while (start != end) { -- cgit v1.2.3 From bf53de907dfdaac178c92d774aae7370d7b97d20 Mon Sep 17 00:00:00 2001 From: Markus Metzger Date: Fri, 19 Dec 2008 15:10:24 +0100 Subject: x86, bts: add fork and exit handling Impact: introduce new ptrace facility Add arch_ptrace_untrace() function that is called when the tracer detaches (either voluntarily or when the tracing task dies); ptrace_disable() is only called on a voluntary detach. Add ptrace_fork() and arch_ptrace_fork(). They are called when a traced task is forked. Clear DS and BTS related fields on fork. Release DS resources and reclaim memory in ptrace_untrace(). This releases resources already when the tracing task dies. We used to do that when the traced task dies. Signed-off-by: Markus Metzger Signed-off-by: Ingo Molnar --- arch/x86/include/asm/ds.h | 9 ++++++++ arch/x86/include/asm/ptrace.h | 7 ++++++ arch/x86/kernel/ds.c | 11 ++++++++++ arch/x86/kernel/process_32.c | 20 ++++++++--------- arch/x86/kernel/process_64.c | 20 ++++++++--------- arch/x86/kernel/ptrace.c | 50 ++++++++++++++++++++++++++++++++++--------- include/linux/ptrace.h | 22 +++++++++++++++++++ kernel/fork.c | 2 ++ kernel/ptrace.c | 12 +++++++++++ 9 files changed, 121 insertions(+), 32 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/ds.h b/arch/x86/include/asm/ds.h index ee0ea3a96c11..a8f672ba100c 100644 --- a/arch/x86/include/asm/ds.h +++ b/arch/x86/include/asm/ds.h @@ -252,12 +252,21 @@ extern void __cpuinit ds_init_intel(struct cpuinfo_x86 *); */ extern void ds_switch_to(struct task_struct *prev, struct task_struct *next); +/* + * Task clone/init and cleanup work + */ +extern void ds_copy_thread(struct task_struct *tsk, struct task_struct *father); +extern void ds_exit_thread(struct task_struct *tsk); + #else /* CONFIG_X86_DS */ struct cpuinfo_x86; static inline void __cpuinit ds_init_intel(struct cpuinfo_x86 *ignored) {} static inline void ds_switch_to(struct task_struct *prev, struct task_struct *next) {} +static inline void ds_copy_thread(struct task_struct *tsk, + struct task_struct *father) {} +static inline void ds_exit_thread(struct task_struct *tsk) {} #endif /* CONFIG_X86_DS */ #endif /* _ASM_X86_DS_H */ diff --git a/arch/x86/include/asm/ptrace.h b/arch/x86/include/asm/ptrace.h index fbf744215911..6d34d954c228 100644 --- a/arch/x86/include/asm/ptrace.h +++ b/arch/x86/include/asm/ptrace.h @@ -235,6 +235,13 @@ extern int do_get_thread_area(struct task_struct *p, int idx, extern int do_set_thread_area(struct task_struct *p, int idx, struct user_desc __user *info, int can_allocate); +extern void x86_ptrace_untrace(struct task_struct *); +extern void x86_ptrace_fork(struct task_struct *child, + unsigned long clone_flags); + +#define arch_ptrace_untrace(tsk) x86_ptrace_untrace(tsk) +#define arch_ptrace_fork(child, flags) x86_ptrace_fork(child, flags) + #endif /* __KERNEL__ */ #endif /* !__ASSEMBLY__ */ diff --git a/arch/x86/kernel/ds.c b/arch/x86/kernel/ds.c index 98d271e60e08..da91701a2348 100644 --- a/arch/x86/kernel/ds.c +++ b/arch/x86/kernel/ds.c @@ -1017,3 +1017,14 @@ void ds_switch_to(struct task_struct *prev, struct task_struct *next) update_debugctlmsr(next->thread.debugctlmsr); } + +void ds_copy_thread(struct task_struct *tsk, struct task_struct *father) +{ + clear_tsk_thread_flag(tsk, TIF_DS_AREA_MSR); + tsk->thread.ds_ctx = NULL; +} + +void ds_exit_thread(struct task_struct *tsk) +{ + WARN_ON(tsk->thread.ds_ctx); +} diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c index 605eff9a8ac0..3ba155d24884 100644 --- a/arch/x86/kernel/process_32.c +++ b/arch/x86/kernel/process_32.c @@ -60,6 +60,7 @@ #include #include #include +#include asmlinkage void ret_from_fork(void) __asm__("ret_from_fork"); @@ -251,17 +252,8 @@ void exit_thread(void) tss->x86_tss.io_bitmap_base = INVALID_IO_BITMAP_OFFSET; put_cpu(); } -#ifdef CONFIG_X86_DS - /* Free any BTS tracers that have not been properly released. */ - if (unlikely(current->bts)) { - ds_release_bts(current->bts); - current->bts = NULL; - - kfree(current->bts_buffer); - current->bts_buffer = NULL; - current->bts_size = 0; - } -#endif /* CONFIG_X86_DS */ + + ds_exit_thread(current); } void flush_thread(void) @@ -343,6 +335,12 @@ int copy_thread(int nr, unsigned long clone_flags, unsigned long sp, kfree(p->thread.io_bitmap_ptr); p->thread.io_bitmap_max = 0; } + + ds_copy_thread(p, current); + + clear_tsk_thread_flag(p, TIF_DEBUGCTLMSR); + p->thread.debugctlmsr = 0; + return err; } diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c index 1cfd2a4bf853..416fb9282f4f 100644 --- a/arch/x86/kernel/process_64.c +++ b/arch/x86/kernel/process_64.c @@ -53,6 +53,7 @@ #include #include #include +#include asmlinkage extern void ret_from_fork(void); @@ -236,17 +237,8 @@ void exit_thread(void) t->io_bitmap_max = 0; put_cpu(); } -#ifdef CONFIG_X86_DS - /* Free any BTS tracers that have not been properly released. */ - if (unlikely(current->bts)) { - ds_release_bts(current->bts); - current->bts = NULL; - - kfree(current->bts_buffer); - current->bts_buffer = NULL; - current->bts_size = 0; - } -#endif /* CONFIG_X86_DS */ + + ds_exit_thread(current); } void flush_thread(void) @@ -376,6 +368,12 @@ int copy_thread(int nr, unsigned long clone_flags, unsigned long sp, if (err) goto out; } + + ds_copy_thread(p, me); + + clear_tsk_thread_flag(p, TIF_DEBUGCTLMSR); + p->thread.debugctlmsr = 0; + err = 0; out: if (err && p->thread.io_bitmap_ptr) { diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c index 45e9855da2d2..6ad2bb607650 100644 --- a/arch/x86/kernel/ptrace.c +++ b/arch/x86/kernel/ptrace.c @@ -769,8 +769,47 @@ static int ptrace_bts_size(struct task_struct *child) return (trace->ds.top - trace->ds.begin) / trace->ds.size; } + +static void ptrace_bts_fork(struct task_struct *tsk) +{ + tsk->bts = NULL; + tsk->bts_buffer = NULL; + tsk->bts_size = 0; + tsk->thread.bts_ovfl_signal = 0; +} + +static void ptrace_bts_untrace(struct task_struct *child) +{ + if (unlikely(child->bts)) { + ds_release_bts(child->bts); + child->bts = NULL; + + kfree(child->bts_buffer); + child->bts_buffer = NULL; + child->bts_size = 0; + } +} + +static void ptrace_bts_detach(struct task_struct *child) +{ + ptrace_bts_untrace(child); +} +#else +static inline void ptrace_bts_fork(struct task_struct *tsk) {} +static inline void ptrace_bts_detach(struct task_struct *child) {} +static inline void ptrace_bts_untrace(struct task_struct *child) {} #endif /* CONFIG_X86_PTRACE_BTS */ +void x86_ptrace_fork(struct task_struct *child, unsigned long clone_flags) +{ + ptrace_bts_fork(child); +} + +void x86_ptrace_untrace(struct task_struct *child) +{ + ptrace_bts_untrace(child); +} + /* * Called by kernel/ptrace.c when detaching.. * @@ -782,16 +821,7 @@ void ptrace_disable(struct task_struct *child) #ifdef TIF_SYSCALL_EMU clear_tsk_thread_flag(child, TIF_SYSCALL_EMU); #endif -#ifdef CONFIG_X86_PTRACE_BTS - if (child->bts) { - ds_release_bts(child->bts); - child->bts = NULL; - - kfree(child->bts_buffer); - child->bts_buffer = NULL; - child->bts_size = 0; - } -#endif /* CONFIG_X86_PTRACE_BTS */ + ptrace_bts_detach(child); } #if defined CONFIG_X86_32 || defined CONFIG_IA32_EMULATION diff --git a/include/linux/ptrace.h b/include/linux/ptrace.h index 22641d5d45df..98b93ca4db06 100644 --- a/include/linux/ptrace.h +++ b/include/linux/ptrace.h @@ -94,6 +94,7 @@ extern void ptrace_notify(int exit_code); extern void __ptrace_link(struct task_struct *child, struct task_struct *new_parent); extern void __ptrace_unlink(struct task_struct *child); +extern void ptrace_fork(struct task_struct *task, unsigned long clone_flags); #define PTRACE_MODE_READ 1 #define PTRACE_MODE_ATTACH 2 /* Returns 0 on success, -errno on denial. */ @@ -313,6 +314,27 @@ static inline void user_enable_block_step(struct task_struct *task) #define arch_ptrace_stop(code, info) do { } while (0) #endif +#ifndef arch_ptrace_untrace +/* + * Do machine-specific work before untracing child. + * + * This is called for a normal detach as well as from ptrace_exit() + * when the tracing task dies. + * + * Called with write_lock(&tasklist_lock) held. + */ +#define arch_ptrace_untrace(task) do { } while (0) +#endif + +#ifndef arch_ptrace_fork +/* + * Do machine-specific work to initialize a new task. + * + * This is called from copy_process(). + */ +#define arch_ptrace_fork(child, clone_flags) do { } while (0) +#endif + extern int task_current_syscall(struct task_struct *target, long *callno, unsigned long args[6], unsigned int maxargs, unsigned long *sp, unsigned long *pc); diff --git a/kernel/fork.c b/kernel/fork.c index 7b93da72d4a2..65ce60adc8e8 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1096,6 +1096,8 @@ static struct task_struct *copy_process(unsigned long clone_flags, #ifdef CONFIG_DEBUG_MUTEXES p->blocked_on = NULL; /* not blocked yet */ #endif + if (unlikely(ptrace_reparented(current))) + ptrace_fork(p, clone_flags); /* Perform scheduler related setup. Assign this task to a CPU. */ sched_fork(p, clone_flags); diff --git a/kernel/ptrace.c b/kernel/ptrace.c index 4c8bcd7dd8e0..100a71cfdaba 100644 --- a/kernel/ptrace.c +++ b/kernel/ptrace.c @@ -25,6 +25,17 @@ #include #include + +/* + * Initialize a new task whose father had been ptraced. + * + * Called from copy_process(). + */ +void ptrace_fork(struct task_struct *child, unsigned long clone_flags) +{ + arch_ptrace_fork(child, clone_flags); +} + /* * ptrace a task: make the debugger its new parent and * move it to the ptrace list. @@ -72,6 +83,7 @@ void __ptrace_unlink(struct task_struct *child) child->parent = child->real_parent; list_del_init(&child->ptrace_entry); + arch_ptrace_untrace(child); if (task_is_traced(child)) ptrace_untrace(child); } -- cgit v1.2.3 From c5dee6177f4bd2095aab7d9be9f6ebdddd6deee9 Mon Sep 17 00:00:00 2001 From: Markus Metzger Date: Fri, 19 Dec 2008 15:17:02 +0100 Subject: x86, bts: memory accounting Impact: move the BTS buffer accounting to the mlock bucket Add alloc_locked_buffer() and free_locked_buffer() functions to mm/mlock.c to kalloc a buffer and account the locked memory to current. Account the memory for the BTS buffer to the tracer. Signed-off-by: Markus Metzger Signed-off-by: Ingo Molnar --- arch/x86/kernel/ptrace.c | 45 ++++++++++++++++++++++++++++++++++----------- include/linux/mm.h | 2 ++ mm/mlock.c | 45 +++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 81 insertions(+), 11 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c index 6ad2bb607650..0a5df5f82fb9 100644 --- a/arch/x86/kernel/ptrace.c +++ b/arch/x86/kernel/ptrace.c @@ -650,6 +650,24 @@ static int ptrace_bts_drain(struct task_struct *child, return drained; } +static int ptrace_bts_allocate_buffer(struct task_struct *child, size_t size) +{ + child->bts_buffer = alloc_locked_buffer(size); + if (!child->bts_buffer) + return -ENOMEM; + + child->bts_size = size; + + return 0; +} + +static void ptrace_bts_free_buffer(struct task_struct *child) +{ + free_locked_buffer(child->bts_buffer, child->bts_size); + child->bts_buffer = NULL; + child->bts_size = 0; +} + static int ptrace_bts_config(struct task_struct *child, long cfg_size, const struct ptrace_bts_config __user *ucfg) @@ -679,14 +697,13 @@ static int ptrace_bts_config(struct task_struct *child, if ((cfg.flags & PTRACE_BTS_O_ALLOC) && (cfg.size != child->bts_size)) { - kfree(child->bts_buffer); + int error; - child->bts_size = cfg.size; - child->bts_buffer = kzalloc(cfg.size, GFP_KERNEL); - if (!child->bts_buffer) { - child->bts_size = 0; - return -ENOMEM; - } + ptrace_bts_free_buffer(child); + + error = ptrace_bts_allocate_buffer(child, cfg.size); + if (error < 0) + return error; } if (cfg.flags & PTRACE_BTS_O_TRACE) @@ -701,10 +718,8 @@ static int ptrace_bts_config(struct task_struct *child, if (IS_ERR(child->bts)) { int error = PTR_ERR(child->bts); - kfree(child->bts_buffer); + ptrace_bts_free_buffer(child); child->bts = NULL; - child->bts_buffer = NULL; - child->bts_size = 0; return error; } @@ -784,6 +799,9 @@ static void ptrace_bts_untrace(struct task_struct *child) ds_release_bts(child->bts); child->bts = NULL; + /* We cannot update total_vm and locked_vm since + child's mm is already gone. But we can reclaim the + memory. */ kfree(child->bts_buffer); child->bts_buffer = NULL; child->bts_size = 0; @@ -792,7 +810,12 @@ static void ptrace_bts_untrace(struct task_struct *child) static void ptrace_bts_detach(struct task_struct *child) { - ptrace_bts_untrace(child); + if (unlikely(child->bts)) { + ds_release_bts(child->bts); + child->bts = NULL; + + ptrace_bts_free_buffer(child); + } } #else static inline void ptrace_bts_fork(struct task_struct *tsk) {} diff --git a/include/linux/mm.h b/include/linux/mm.h index ffee2f743418..9979d3fab6e7 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -1286,5 +1286,7 @@ int vmemmap_populate_basepages(struct page *start_page, int vmemmap_populate(struct page *start_page, unsigned long pages, int node); void vmemmap_populate_print_last(void); +extern void *alloc_locked_buffer(size_t size); +extern void free_locked_buffer(void *buffer, size_t size); #endif /* __KERNEL__ */ #endif /* _LINUX_MM_H */ diff --git a/mm/mlock.c b/mm/mlock.c index 1ada366570cb..3035a56e7616 100644 --- a/mm/mlock.c +++ b/mm/mlock.c @@ -667,3 +667,48 @@ void user_shm_unlock(size_t size, struct user_struct *user) spin_unlock(&shmlock_user_lock); free_uid(user); } + +void *alloc_locked_buffer(size_t size) +{ + unsigned long rlim, vm, pgsz; + void *buffer = NULL; + + pgsz = PAGE_ALIGN(size) >> PAGE_SHIFT; + + down_write(¤t->mm->mmap_sem); + + rlim = current->signal->rlim[RLIMIT_AS].rlim_cur >> PAGE_SHIFT; + vm = current->mm->total_vm + pgsz; + if (rlim < vm) + goto out; + + rlim = current->signal->rlim[RLIMIT_MEMLOCK].rlim_cur >> PAGE_SHIFT; + vm = current->mm->locked_vm + pgsz; + if (rlim < vm) + goto out; + + buffer = kzalloc(size, GFP_KERNEL); + if (!buffer) + goto out; + + current->mm->total_vm += pgsz; + current->mm->locked_vm += pgsz; + + out: + up_write(¤t->mm->mmap_sem); + return buffer; +} + +void free_locked_buffer(void *buffer, size_t size) +{ + unsigned long pgsz = PAGE_ALIGN(size) >> PAGE_SHIFT; + + down_write(¤t->mm->mmap_sem); + + current->mm->total_vm -= pgsz; + current->mm->locked_vm -= pgsz; + + up_write(¤t->mm->mmap_sem); + + kfree(buffer); +} -- cgit v1.2.3 From 280a9ca5d0663b185ddc4443052076c29652a328 Mon Sep 17 00:00:00 2001 From: Dmitry Adamushko Date: Sat, 20 Dec 2008 00:15:24 +0100 Subject: x86: fix resume (S2R) broken by Intel microcode module, on A110L Impact: fix deadlock This is in response to the following bug report: Bug-Entry : http://bugzilla.kernel.org/show_bug.cgi?id=12100 Subject : resume (S2R) broken by Intel microcode module, on A110L Submitter : Andreas Mohr Date : 2008-11-25 08:48 (19 days old) Handled-By : Dmitry Adamushko [ The deadlock scenario has been discovered by Andreas Mohr ] I think I might have a logical explanation why the system: (http://bugzilla.kernel.org/show_bug.cgi?id=12100) might hang upon resuming, OTOH it should have likely hanged each and every time. (1) possible deadlock in microcode_resume_cpu() if either 'if' section is taken; (2) now, I don't see it in spec. and can't experimentally verify it (newer ucodes don't seem to be available for my Core2duo)... but logically-wise, I'd think that when read upon resuming, the 'microcode revision' (MSR 0x8B) should be back to its original one (we need to reload ucode anyway so it doesn't seem logical if a cpu doesn't drop the version)... if so, the comparison with memcmp() for the full 'struct cpu_signature' is wrong... and that's how one of the aforementioned 'if' sections might have been triggered - leading to a deadlock. Obviously, in my tests I simulated loading/resuming with the ucode of the same version (just to see that the file is loaded/re-loaded upon resuming) so this issue has never popped up. I'd appreciate if someone with an appropriate system might give a try to the 2nd patch (titled "fix a comparison && deadlock..."). In any case, the deadlock situation is a must-have fix. Reported-by: Andreas Mohr Signed-off-by: Dmitry Adamushko Tested-by: Andreas Mohr Signed-off-by: Ingo Molnar Cc: Signed-off-by: Ingo Molnar --- arch/x86/kernel/microcode_core.c | 19 ++++++++++++++----- arch/x86/kernel/microcode_intel.c | 6 ++++++ 2 files changed, 20 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/microcode_core.c b/arch/x86/kernel/microcode_core.c index 82fb2809ce32..c4b5b24e0217 100644 --- a/arch/x86/kernel/microcode_core.c +++ b/arch/x86/kernel/microcode_core.c @@ -272,13 +272,18 @@ static struct attribute_group mc_attr_group = { .name = "microcode", }; -static void microcode_fini_cpu(int cpu) +static void __microcode_fini_cpu(int cpu) { struct ucode_cpu_info *uci = ucode_cpu_info + cpu; - mutex_lock(µcode_mutex); microcode_ops->microcode_fini_cpu(cpu); uci->valid = 0; +} + +static void microcode_fini_cpu(int cpu) +{ + mutex_lock(µcode_mutex); + __microcode_fini_cpu(cpu); mutex_unlock(µcode_mutex); } @@ -306,12 +311,16 @@ static int microcode_resume_cpu(int cpu) * to this cpu (a bit of paranoia): */ if (microcode_ops->collect_cpu_info(cpu, &nsig)) { - microcode_fini_cpu(cpu); + __microcode_fini_cpu(cpu); + printk(KERN_ERR "failed to collect_cpu_info for resuming cpu #%d\n", + cpu); return -1; } - if (memcmp(&nsig, &uci->cpu_sig, sizeof(nsig))) { - microcode_fini_cpu(cpu); + if ((nsig.sig != uci->cpu_sig.sig) || (nsig.pf != uci->cpu_sig.pf)) { + __microcode_fini_cpu(cpu); + printk(KERN_ERR "cached ucode doesn't match the resuming cpu #%d\n", + cpu); /* Should we look for a new ucode here? */ return 1; } diff --git a/arch/x86/kernel/microcode_intel.c b/arch/x86/kernel/microcode_intel.c index 622dc4a21784..a8e62792d171 100644 --- a/arch/x86/kernel/microcode_intel.c +++ b/arch/x86/kernel/microcode_intel.c @@ -155,6 +155,7 @@ static DEFINE_SPINLOCK(microcode_update_lock); static int collect_cpu_info(int cpu_num, struct cpu_signature *csig) { struct cpuinfo_x86 *c = &cpu_data(cpu_num); + unsigned long flags; unsigned int val[2]; memset(csig, 0, sizeof(*csig)); @@ -174,11 +175,16 @@ static int collect_cpu_info(int cpu_num, struct cpu_signature *csig) csig->pf = 1 << ((val[1] >> 18) & 7); } + /* serialize access to the physical write to MSR 0x79 */ + spin_lock_irqsave(µcode_update_lock, flags); + wrmsr(MSR_IA32_UCODE_REV, 0, 0); /* see notes above for revision 1.07. Apparent chip bug */ sync_core(); /* get the current revision from MSR 0x8B */ rdmsr(MSR_IA32_UCODE_REV, val[0], csig->rev); + spin_unlock_irqrestore(µcode_update_lock, flags); + pr_debug("microcode: collect_cpu_info : sig=0x%x, pf=0x%x, rev=0x%x\n", csig->sig, csig->pf, csig->rev); -- cgit v1.2.3 From adf77bac052bb5bf0722b2ce2af9fefc5b2d2a71 Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Mon, 22 Dec 2008 17:56:05 -0800 Subject: x86: prioritize the FPU traps for the error code In the case of multiple FPU errors, prioritize the error codes, instead of returning __SI_FAULT, which ends up pushing a 0 as the error code to userspace, a POSIX violation. For i386, we will simply return if there are no errors at all; for x86-64 this is probably a "can't happen" (and the code should be unified), but for this patch, return __SI_FAULT|SI_KERNEL if this ever happens. Signed-off-by: H. Peter Anvin --- arch/x86/kernel/traps.c | 34 +++++++++++++++------------------- 1 file changed, 15 insertions(+), 19 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c index 04d242ab0161..c320c29255c2 100644 --- a/arch/x86/kernel/traps.c +++ b/arch/x86/kernel/traps.c @@ -664,7 +664,7 @@ void math_error(void __user *ip) { struct task_struct *task; siginfo_t info; - unsigned short cwd, swd; + unsigned short cwd, swd, err; /* * Save the info for the exception handler and clear the error. @@ -675,7 +675,6 @@ void math_error(void __user *ip) task->thread.error_code = 0; info.si_signo = SIGFPE; info.si_errno = 0; - info.si_code = __SI_FAULT; info.si_addr = ip; /* * (~cwd & swd) will mask out exceptions that are not set to unmasked @@ -689,34 +688,31 @@ void math_error(void __user *ip) */ cwd = get_fpu_cwd(task); swd = get_fpu_swd(task); - switch (swd & ~cwd & 0x3f) { - case 0x000: /* No unmasked exception */ -#ifdef CONFIG_X86_32 + + err = swd & ~cwd & 0x3f; + +#if CONFIG_X86_32 + if (!err) return; #endif - default: /* Multiple exceptions */ - break; - case 0x001: /* Invalid Op */ + + if (err & 0x001) { /* Invalid op */ /* * swd & 0x240 == 0x040: Stack Underflow * swd & 0x240 == 0x240: Stack Overflow * User must clear the SF bit (0x40) if set */ info.si_code = FPE_FLTINV; - break; - case 0x002: /* Denormalize */ - case 0x010: /* Underflow */ - info.si_code = FPE_FLTUND; - break; - case 0x004: /* Zero Divide */ + } else if (err & 0x004) { /* Divide by Zero */ info.si_code = FPE_FLTDIV; - break; - case 0x008: /* Overflow */ + } else if (err & 0x008) { /* Overflow */ info.si_code = FPE_FLTOVF; - break; - case 0x020: /* Precision */ + } else if (err & 0x012) { /* Denormal, Underflow */ + info.si_code = FPE_FLTUND; + } else if (err & 0x020) { /* Precision */ info.si_code = FPE_FLTRES; - break; + } else { + info.si_code = __SI_FAULT|SI_KERNEL; /* WTF? */ } force_sig_info(SIGFPE, &info, task); } -- cgit v1.2.3 From b77b881f21b29aa7efa668fde69ee3dc0372ae3f Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Fri, 19 Dec 2008 15:23:44 -0800 Subject: x86: fix lguest used_vectors breakage, -v2 Impact: fix lguest, clean up 32-bit lguest used used_vectors to record vectors, but that model of allocating vectors changed and got broken, after we changed vector allocation to a per_cpu array. Try enable that for 64bit, and the array is used for all vectors that are not managed by vector_irq per_cpu array. Also kill system_vectors[], that is now a duplication of the used_vectors bitmap. [ merged in cpus4096 due to io_apic.c cpumask changes. ] [ -v2, fix build failure ] Signed-off-by: Yinghai Lu Signed-off-by: Ingo Molnar Signed-off-by: Ingo Molnar --- arch/x86/include/asm/desc.h | 10 ++++------ arch/x86/include/asm/irq.h | 1 + arch/x86/kernel/apic.c | 2 -- arch/x86/kernel/io_apic.c | 9 +++------ arch/x86/kernel/irqinit_32.c | 16 +++++++++++++++- arch/x86/kernel/irqinit_64.c | 13 +++++++++++++ arch/x86/kernel/traps.c | 12 +++++++----- drivers/lguest/interrupts_and_traps.c | 13 +++++++++---- 8 files changed, 52 insertions(+), 24 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/desc.h b/arch/x86/include/asm/desc.h index e6b82b17b072..dc27705f5443 100644 --- a/arch/x86/include/asm/desc.h +++ b/arch/x86/include/asm/desc.h @@ -320,16 +320,14 @@ static inline void set_intr_gate(unsigned int n, void *addr) _set_gate(n, GATE_INTERRUPT, addr, 0, 0, __KERNEL_CS); } -#define SYS_VECTOR_FREE 0 -#define SYS_VECTOR_ALLOCED 1 - extern int first_system_vector; -extern char system_vectors[]; +/* used_vectors is BITMAP for irq is not managed by percpu vector_irq */ +extern unsigned long used_vectors[]; static inline void alloc_system_vector(int vector) { - if (system_vectors[vector] == SYS_VECTOR_FREE) { - system_vectors[vector] = SYS_VECTOR_ALLOCED; + if (!test_bit(vector, used_vectors)) { + set_bit(vector, used_vectors); if (first_system_vector > vector) first_system_vector = vector; } else diff --git a/arch/x86/include/asm/irq.h b/arch/x86/include/asm/irq.h index 8766d30fb746..4bb732e45a85 100644 --- a/arch/x86/include/asm/irq.h +++ b/arch/x86/include/asm/irq.h @@ -46,5 +46,6 @@ extern void native_init_IRQ(void); /* Interrupt vector management */ extern DECLARE_BITMAP(used_vectors, NR_VECTORS); +extern int vector_used_by_percpu_irq(unsigned int vector); #endif /* _ASM_X86_IRQ_H */ diff --git a/arch/x86/kernel/apic.c b/arch/x86/kernel/apic.c index f7a32a3beb2f..b9019271af62 100644 --- a/arch/x86/kernel/apic.c +++ b/arch/x86/kernel/apic.c @@ -118,8 +118,6 @@ EXPORT_SYMBOL_GPL(local_apic_timer_c2_ok); int first_system_vector = 0xfe; -char system_vectors[NR_VECTORS] = { [0 ... NR_VECTORS-1] = SYS_VECTOR_FREE}; - /* * Debug level, exported for io_apic.c */ diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index 908c1d00a5c7..1cbf7c8d46e0 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -1326,13 +1326,10 @@ next: } if (unlikely(current_vector == vector)) continue; -#ifdef CONFIG_X86_64 - if (vector == IA32_SYSCALL_VECTOR) - goto next; -#else - if (vector == SYSCALL_VECTOR) + + if (test_bit(vector, used_vectors)) goto next; -#endif + for_each_cpu_and(new_cpu, tmp_mask, cpu_online_mask) if (per_cpu(vector_irq, new_cpu)[vector] != -1) goto next; diff --git a/arch/x86/kernel/irqinit_32.c b/arch/x86/kernel/irqinit_32.c index 6a92f47c52e7..61aa2a1004b5 100644 --- a/arch/x86/kernel/irqinit_32.c +++ b/arch/x86/kernel/irqinit_32.c @@ -110,6 +110,18 @@ DEFINE_PER_CPU(vector_irq_t, vector_irq) = { [IRQ15_VECTOR + 1 ... NR_VECTORS - 1] = -1 }; +int vector_used_by_percpu_irq(unsigned int vector) +{ + int cpu; + + for_each_online_cpu(cpu) { + if (per_cpu(vector_irq, cpu)[vector] != -1) + return 1; + } + + return 0; +} + /* Overridden in paravirt.c */ void init_IRQ(void) __attribute__((weak, alias("native_init_IRQ"))); @@ -146,10 +158,12 @@ void __init native_init_IRQ(void) alloc_intr_gate(CALL_FUNCTION_VECTOR, call_function_interrupt); /* IPI for single call function */ - set_intr_gate(CALL_FUNCTION_SINGLE_VECTOR, call_function_single_interrupt); + alloc_intr_gate(CALL_FUNCTION_SINGLE_VECTOR, + call_function_single_interrupt); /* Low priority IPI to cleanup after moving an irq */ set_intr_gate(IRQ_MOVE_CLEANUP_VECTOR, irq_move_cleanup_interrupt); + set_bit(IRQ_MOVE_CLEANUP_VECTOR, used_vectors); #endif #ifdef CONFIG_X86_LOCAL_APIC diff --git a/arch/x86/kernel/irqinit_64.c b/arch/x86/kernel/irqinit_64.c index 40c1e62ec785..1020919efe1c 100644 --- a/arch/x86/kernel/irqinit_64.c +++ b/arch/x86/kernel/irqinit_64.c @@ -135,6 +135,18 @@ DEFINE_PER_CPU(vector_irq_t, vector_irq) = { [IRQ15_VECTOR + 1 ... NR_VECTORS - 1] = -1 }; +int vector_used_by_percpu_irq(unsigned int vector) +{ + int cpu; + + for_each_online_cpu(cpu) { + if (per_cpu(vector_irq, cpu)[vector] != -1) + return 1; + } + + return 0; +} + void __init init_ISA_irqs(void) { int i; @@ -187,6 +199,7 @@ static void __init smp_intr_init(void) /* Low priority IPI to cleanup after moving an irq */ set_intr_gate(IRQ_MOVE_CLEANUP_VECTOR, irq_move_cleanup_interrupt); + set_bit(IRQ_MOVE_CLEANUP_VECTOR, used_vectors); #endif } diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c index 04d242ab0161..4a6dff39a470 100644 --- a/arch/x86/kernel/traps.c +++ b/arch/x86/kernel/traps.c @@ -72,9 +72,6 @@ #include "cpu/mcheck/mce.h" -DECLARE_BITMAP(used_vectors, NR_VECTORS); -EXPORT_SYMBOL_GPL(used_vectors); - asmlinkage int system_call(void); /* Do we ignore FPU interrupts ? */ @@ -89,6 +86,9 @@ gate_desc idt_table[256] __attribute__((__section__(".data.idt"))) = { { { { 0, 0 } } }, }; #endif +DECLARE_BITMAP(used_vectors, NR_VECTORS); +EXPORT_SYMBOL_GPL(used_vectors); + static int ignore_nmis; static inline void conditional_sti(struct pt_regs *regs) @@ -949,9 +949,7 @@ dotraplinkage void do_iret_error(struct pt_regs *regs, long error_code) void __init trap_init(void) { -#ifdef CONFIG_X86_32 int i; -#endif #ifdef CONFIG_EISA void __iomem *p = early_ioremap(0x0FFFD9, 4); @@ -1008,11 +1006,15 @@ void __init trap_init(void) } set_system_trap_gate(SYSCALL_VECTOR, &system_call); +#endif /* Reserve all the builtin and the syscall vector: */ for (i = 0; i < FIRST_EXTERNAL_VECTOR; i++) set_bit(i, used_vectors); +#ifdef CONFIG_X86_64 + set_bit(IA32_SYSCALL_VECTOR, used_vectors); +#else set_bit(SYSCALL_VECTOR, used_vectors); #endif /* diff --git a/drivers/lguest/interrupts_and_traps.c b/drivers/lguest/interrupts_and_traps.c index a1039068f95c..415fab0125ac 100644 --- a/drivers/lguest/interrupts_and_traps.c +++ b/drivers/lguest/interrupts_and_traps.c @@ -222,11 +222,16 @@ bool check_syscall_vector(struct lguest *lg) int init_interrupts(void) { /* If they want some strange system call vector, reserve it now */ - if (syscall_vector != SYSCALL_VECTOR - && test_and_set_bit(syscall_vector, used_vectors)) { - printk("lg: couldn't reserve syscall %u\n", syscall_vector); - return -EBUSY; + if (syscall_vector != SYSCALL_VECTOR) { + if (test_bit(syscall_vector, used_vectors) || + vector_used_by_percpu_irq(syscall_vector)) { + printk(KERN_ERR "lg: couldn't reserve syscall %u\n", + syscall_vector); + return -EBUSY; + } + set_bit(syscall_vector, used_vectors); } + return 0; } -- cgit v1.2.3 From 7d87d5365556b1c6e8c00abcc632c3ad1fdc58b8 Mon Sep 17 00:00:00 2001 From: Suresh Siddha Date: Mon, 22 Dec 2008 17:33:28 -0800 Subject: x86: use logical apicid in x2apic_cluster's x2apic_cpu_mask_to_apicid_and() These commits: commit 95d313cf1c1ecedc8bec5727b09bdacbf67dfc45 Author: Mike Travis Date: Tue Dec 16 17:33:54 2008 -0800 x86: Add cpu_mask_to_apicid_and and commit 6eeb7c5a99434596c5953a95baa17d2f085664e3 Author: Mike Travis Date: Tue Dec 16 17:33:55 2008 -0800 x86: update add-cpu_mask_to_apicid_and to use struct cpumask* broke interrupt delivery on x2apic platforms. As x2apic cluster mode uses logical delivery mode, we need to use logical apicid instead of physical apicid in x2apic_cpu_mask_to_apicid_and() Impact: fixes the broken interrupt delivery issue on generic x2apic platforms. Signed-off-by: Suresh Siddha Acked-by: Mike Travis Signed-off-by: H. Peter Anvin Signed-off-by: Ingo Molnar --- arch/x86/kernel/genx2apic_cluster.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/genx2apic_cluster.c b/arch/x86/kernel/genx2apic_cluster.c index d451c9b9fdff..6ce497cc372d 100644 --- a/arch/x86/kernel/genx2apic_cluster.c +++ b/arch/x86/kernel/genx2apic_cluster.c @@ -114,7 +114,7 @@ static unsigned int x2apic_cpu_mask_to_apicid(const struct cpumask *cpumask) int cpu; /* - * We're using fixed IRQ delivery, can only return one phys APIC ID. + * We're using fixed IRQ delivery, can only return one logical APIC ID. * May as well be the first. */ cpu = cpumask_first(cpumask); @@ -130,14 +130,14 @@ static unsigned int x2apic_cpu_mask_to_apicid_and(const struct cpumask *cpumask, int cpu; /* - * We're using fixed IRQ delivery, can only return one phys APIC ID. + * We're using fixed IRQ delivery, can only return one logical APIC ID. * May as well be the first. */ for_each_cpu_and(cpu, cpumask, andmask) if (cpumask_test_cpu(cpu, cpu_online_mask)) break; if (cpu < nr_cpu_ids) - return per_cpu(x86_cpu_to_apicid, cpu); + return per_cpu(x86_cpu_to_logical_apicid, cpu); return BAD_APICID; } -- cgit v1.2.3 From c3d80000e3a812fe5a200d6bde755fbd7fa65481 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Tue, 23 Dec 2008 15:15:17 +0100 Subject: x86: export vector_used_by_percpu_irq Impact: build fix lguest can be built as a module and makes use of this new symbol: ERROR: "vector_used_by_percpu_irq" [drivers/lguest/lg.ko] undefined! export it. Signed-off-by: Ingo Molnar --- arch/x86/kernel/irq.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/irq.c b/arch/x86/kernel/irq.c index 3f1d9d18df67..bce53e1352a0 100644 --- a/arch/x86/kernel/irq.c +++ b/arch/x86/kernel/irq.c @@ -9,6 +9,7 @@ #include #include #include +#include atomic_t irq_err_count; @@ -190,3 +191,5 @@ u64 arch_irq_stat(void) #endif return sum; } + +EXPORT_SYMBOL_GPL(vector_used_by_percpu_irq); -- cgit v1.2.3 From c1c15b65ec30275575dac9322aae607075769fbc Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Tue, 23 Dec 2008 10:10:40 -0800 Subject: x86: PAT: fix address types in track_pfn_vma_new() Impact: cleanup, fix warning This warning: arch/x86/mm/pat.c: In function track_pfn_vma_copy: arch/x86/mm/pat.c:701: warning: passing argument 5 of follow_phys from incompatible pointer type Triggers because physical addresses are resource_size_t, not u64. This really matters when calling an interface like follow_phys() which takes a pointer to a physical address -- although on x86, being littleendian, it would generally work anyway as long as the memory region wasn't completely uninitialized. Signed-off-by: H. Peter Anvin Signed-off-by: Ingo Molnar --- arch/x86/mm/pat.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/mm/pat.c b/arch/x86/mm/pat.c index 541bcc944a5b..85cbd3cd3723 100644 --- a/arch/x86/mm/pat.c +++ b/arch/x86/mm/pat.c @@ -684,7 +684,7 @@ int track_pfn_vma_copy(struct vm_area_struct *vma) { int retval = 0; unsigned long i, j; - u64 paddr; + resource_size_t paddr; unsigned long prot; unsigned long vma_start = vma->vm_start; unsigned long vma_end = vma->vm_end; @@ -746,8 +746,8 @@ int track_pfn_vma_new(struct vm_area_struct *vma, pgprot_t prot, { int retval = 0; unsigned long i, j; - u64 base_paddr; - u64 paddr; + resource_size_t base_paddr; + resource_size_t paddr; unsigned long vma_start = vma->vm_start; unsigned long vma_end = vma->vm_end; unsigned long vma_size = vma_end - vma_start; @@ -757,12 +757,12 @@ int track_pfn_vma_new(struct vm_area_struct *vma, pgprot_t prot, if (is_linear_pfn_mapping(vma)) { /* reserve the whole chunk starting from vm_pgoff */ - paddr = (u64)vma->vm_pgoff << PAGE_SHIFT; + paddr = (resource_size_t)vma->vm_pgoff << PAGE_SHIFT; return reserve_pfn_range(paddr, vma_size, prot); } /* reserve page by page using pfn and size */ - base_paddr = (u64)pfn << PAGE_SHIFT; + base_paddr = (resource_size_t)pfn << PAGE_SHIFT; for (i = 0; i < size; i += PAGE_SIZE) { paddr = base_paddr + i; retval = reserve_pfn_range(paddr, PAGE_SIZE, prot); @@ -790,7 +790,7 @@ void untrack_pfn_vma(struct vm_area_struct *vma, unsigned long pfn, unsigned long size) { unsigned long i; - u64 paddr; + resource_size_t paddr; unsigned long prot; unsigned long vma_start = vma->vm_start; unsigned long vma_end = vma->vm_end; @@ -801,14 +801,14 @@ void untrack_pfn_vma(struct vm_area_struct *vma, unsigned long pfn, if (is_linear_pfn_mapping(vma)) { /* free the whole chunk starting from vm_pgoff */ - paddr = (u64)vma->vm_pgoff << PAGE_SHIFT; + paddr = (resource_size_t)vma->vm_pgoff << PAGE_SHIFT; free_pfn_range(paddr, vma_size); return; } if (size != 0 && size != vma_size) { /* free page by page, using pfn and size */ - paddr = (u64)pfn << PAGE_SHIFT; + paddr = (resource_size_t)pfn << PAGE_SHIFT; for (i = 0; i < size; i += PAGE_SIZE) { paddr = paddr + i; free_pfn_range(paddr, PAGE_SIZE); -- cgit v1.2.3 From 40f15ad8aadff5ebb621b17a6f303ad2cd3f847d Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Wed, 24 Dec 2008 10:49:51 +0100 Subject: x86: disable X86_PTRACE_BTS there's a new ptrace arch level feature in .28: config X86_PTRACE_BTS bool "Branch Trace Store" it has broken fork() handling: the old DS area gets copied over into a new task without clearing it. Fixes exist but they came too late: c5dee61: x86, bts: memory accounting bf53de9: x86, bts: add fork and exit handling and are queued up for v2.6.29. This shows that the facility is still not tested well enough to release into a stable kernel - disable it for now and reactivate in .29. In .29 the hardware-branch-tracer will use the DS/BTS facilities too - hopefully resulting in better code. Signed-off-by: Ingo Molnar --- arch/x86/Kconfig.cpu | 1 + 1 file changed, 1 insertion(+) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig.cpu b/arch/x86/Kconfig.cpu index b815664fe370..8e99073b9e0f 100644 --- a/arch/x86/Kconfig.cpu +++ b/arch/x86/Kconfig.cpu @@ -520,6 +520,7 @@ config X86_PTRACE_BTS bool "Branch Trace Store" default y depends on X86_DEBUGCTLMSR + depends on BROKEN help This adds a ptrace interface to the hardware's branch trace store. -- cgit v1.2.3 From 67be403d897f818b1a5ecc201967b0ee6a0332f9 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Wed, 24 Dec 2008 21:08:37 +0100 Subject: Revert "x86: disable X86_PTRACE_BTS" This reverts commit 40f15ad8aadff5ebb621b17a6f303ad2cd3f847d. The CONFIG_X86_PTRACE_BTS bugs have been fixed via: c5dee61: x86, bts: memory accounting bf53de9: x86, bts: add fork and exit handling Signed-off-by: Ingo Molnar --- arch/x86/Kconfig.cpu | 1 - 1 file changed, 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig.cpu b/arch/x86/Kconfig.cpu index b54903efb39e..85a78575956c 100644 --- a/arch/x86/Kconfig.cpu +++ b/arch/x86/Kconfig.cpu @@ -521,7 +521,6 @@ config X86_PTRACE_BTS bool "Branch Trace Store" default y depends on X86_DEBUGCTLMSR - depends on BROKEN help This adds a ptrace interface to the hardware's branch trace store. -- cgit v1.2.3 From 1c06da81a5d042d5fba67c4c533b16ae62a174ab Mon Sep 17 00:00:00 2001 From: Kent Liu Date: Fri, 31 Oct 2008 16:52:58 +0800 Subject: crypto: crc32c-intel - Update copyright head The original copyright head for crc32c-intel.c is incorrect. Please merge the patch to update it. Signed-Off-By: Kent Liu Signed-off-by: Herbert Xu --- arch/x86/crypto/crc32c-intel.c | 20 ++++++++++++++------ 1 file changed, 14 insertions(+), 6 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/crypto/crc32c-intel.c b/arch/x86/crypto/crc32c-intel.c index 070afc5b6c94..a2c539cc52b7 100644 --- a/arch/x86/crypto/crc32c-intel.c +++ b/arch/x86/crypto/crc32c-intel.c @@ -6,13 +6,22 @@ * Intel(R) 64 and IA-32 Architectures Software Developer's Manual * Volume 2A: Instruction Set Reference, A-M * - * Copyright (c) 2008 Austin Zhang - * Copyright (c) 2008 Kent Liu + * Copyright (C) 2008 Intel Corporation + * Authors: Austin Zhang + * Kent Liu * * This program is free software; you can redistribute it and/or modify it - * under the terms of the GNU General Public License as published by the Free - * Software Foundation; either version 2 of the License, or (at your option) - * any later version. + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. * */ #include @@ -194,4 +203,3 @@ MODULE_LICENSE("GPL"); MODULE_ALIAS("crc32c"); MODULE_ALIAS("crc32c-intel"); - -- cgit v1.2.3 From b7e8bdadce6317eb13c13b9451d7114614aa1450 Mon Sep 17 00:00:00 2001 From: Herbert Xu Date: Thu, 6 Nov 2008 16:56:41 +0800 Subject: crypto: crc32c-intel - Switch to shash This patch changes crc32c-intel to the new shash interface. Signed-off-by: Herbert Xu --- arch/x86/crypto/crc32c-intel.c | 101 +++++++++++++++++++---------------------- 1 file changed, 47 insertions(+), 54 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/crypto/crc32c-intel.c b/arch/x86/crypto/crc32c-intel.c index a2c539cc52b7..b9d00261703c 100644 --- a/arch/x86/crypto/crc32c-intel.c +++ b/arch/x86/crypto/crc32c-intel.c @@ -84,99 +84,92 @@ static u32 __pure crc32c_intel_le_hw(u32 crc, unsigned char const *p, size_t len * If your algorithm starts with ~0, then XOR with ~0 before you set * the seed. */ -static int crc32c_intel_setkey(struct crypto_ahash *hash, const u8 *key, +static int crc32c_intel_setkey(struct crypto_shash *hash, const u8 *key, unsigned int keylen) { - u32 *mctx = crypto_ahash_ctx(hash); + u32 *mctx = crypto_shash_ctx(hash); if (keylen != sizeof(u32)) { - crypto_ahash_set_flags(hash, CRYPTO_TFM_RES_BAD_KEY_LEN); + crypto_shash_set_flags(hash, CRYPTO_TFM_RES_BAD_KEY_LEN); return -EINVAL; } *mctx = le32_to_cpup((__le32 *)key); return 0; } -static int crc32c_intel_init(struct ahash_request *req) +static int crc32c_intel_init(struct shash_desc *desc) { - u32 *mctx = crypto_ahash_ctx(crypto_ahash_reqtfm(req)); - u32 *crcp = ahash_request_ctx(req); + u32 *mctx = crypto_shash_ctx(desc->tfm); + u32 *crcp = shash_desc_ctx(desc); *crcp = *mctx; return 0; } -static int crc32c_intel_update(struct ahash_request *req) +static int crc32c_intel_update(struct shash_desc *desc, const u8 *data, + unsigned int len) { - struct crypto_hash_walk walk; - u32 *crcp = ahash_request_ctx(req); - u32 crc = *crcp; - int nbytes; + u32 *crcp = shash_desc_ctx(desc); - for (nbytes = crypto_hash_walk_first(req, &walk); nbytes; - nbytes = crypto_hash_walk_done(&walk, 0)) - crc = crc32c_intel_le_hw(crc, walk.data, nbytes); - - *crcp = crc; + *crcp = crc32c_intel_le_hw(*crcp, data, len); return 0; } -static int crc32c_intel_final(struct ahash_request *req) +static int __crc32c_intel_finup(u32 *crcp, const u8 *data, unsigned int len, + u8 *out) { - u32 *crcp = ahash_request_ctx(req); - - *(__le32 *)req->result = ~cpu_to_le32p(crcp); + *(__le32 *)out = ~cpu_to_le32(crc32c_intel_le_hw(*crcp, data, len)); return 0; } -static int crc32c_intel_digest(struct ahash_request *req) +static int crc32c_intel_finup(struct shash_desc *desc, const u8 *data, + unsigned int len, u8 *out) { - struct crypto_hash_walk walk; - u32 *mctx = crypto_ahash_ctx(crypto_ahash_reqtfm(req)); - u32 crc = *mctx; - int nbytes; + return __crc32c_intel_finup(shash_desc_ctx(desc), data, len, out); +} - for (nbytes = crypto_hash_walk_first(req, &walk); nbytes; - nbytes = crypto_hash_walk_done(&walk, 0)) - crc = crc32c_intel_le_hw(crc, walk.data, nbytes); +static int crc32c_intel_final(struct shash_desc *desc, u8 *out) +{ + u32 *crcp = shash_desc_ctx(desc); - *(__le32 *)req->result = ~cpu_to_le32(crc); + *(__le32 *)out = ~cpu_to_le32p(crcp); return 0; } +static int crc32c_intel_digest(struct shash_desc *desc, const u8 *data, + unsigned int len, u8 *out) +{ + return __crc32c_intel_finup(crypto_shash_ctx(desc->tfm), data, len, + out); +} + static int crc32c_intel_cra_init(struct crypto_tfm *tfm) { u32 *key = crypto_tfm_ctx(tfm); *key = ~0; - tfm->crt_ahash.reqsize = sizeof(u32); - return 0; } -static struct crypto_alg alg = { - .cra_name = "crc32c", - .cra_driver_name = "crc32c-intel", - .cra_priority = 200, - .cra_flags = CRYPTO_ALG_TYPE_AHASH, - .cra_blocksize = CHKSUM_BLOCK_SIZE, - .cra_alignmask = 3, - .cra_ctxsize = sizeof(u32), - .cra_module = THIS_MODULE, - .cra_list = LIST_HEAD_INIT(alg.cra_list), - .cra_init = crc32c_intel_cra_init, - .cra_type = &crypto_ahash_type, - .cra_u = { - .ahash = { - .digestsize = CHKSUM_DIGEST_SIZE, - .setkey = crc32c_intel_setkey, - .init = crc32c_intel_init, - .update = crc32c_intel_update, - .final = crc32c_intel_final, - .digest = crc32c_intel_digest, - } +static struct shash_alg alg = { + .setkey = crc32c_intel_setkey, + .init = crc32c_intel_init, + .update = crc32c_intel_update, + .final = crc32c_intel_final, + .finup = crc32c_intel_finup, + .digest = crc32c_intel_digest, + .descsize = sizeof(u32), + .digestsize = CHKSUM_DIGEST_SIZE, + .base = { + .cra_name = "crc32c", + .cra_driver_name = "crc32c-intel", + .cra_priority = 200, + .cra_blocksize = CHKSUM_BLOCK_SIZE, + .cra_ctxsize = sizeof(u32), + .cra_module = THIS_MODULE, + .cra_init = crc32c_intel_cra_init, } }; @@ -184,14 +177,14 @@ static struct crypto_alg alg = { static int __init crc32c_intel_mod_init(void) { if (cpu_has_xmm4_2) - return crypto_register_alg(&alg); + return crypto_register_shash(&alg); else return -ENODEV; } static void __exit crc32c_intel_mod_fini(void) { - crypto_unregister_alg(&alg); + crypto_unregister_shash(&alg); } module_init(crc32c_intel_mod_init); -- cgit v1.2.3 From 0ca59dd948a51c95d5a366d35f897bc5ef9df55d Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 24 Dec 2008 23:30:02 +0100 Subject: tracing/ftrace: don't trace on early stage of a secondary cpu boot, v3 Impact: fix a crash/hard-reboot on certain configs while enabling cpu runtime On some archs, the boot of a secondary cpu can have an early fragile state. On x86-64, the pda is not initialized on the first stage of a cpu boot but it is needed to get the cpu number and the current task pointer. This data is needed during tracing. As they were dereferenced at this stage, we got a crash while tracing a cpu being enabled at runtime. Some other archs like ia64 can have such kind of issue too. Changes on v2: We dropped the previous solution of a per-arch called function to guess the current state of a cpu. That could slow down the tracing. This patch removes the -pg flag on arch/x86/kernel/cpu/common.c where the low level cpu boot functions exist, on start_secondary() and a helper function used at this stage. Signed-off-by: Frederic Weisbecker Acked-by: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/x86/include/asm/msr.h | 3 ++- arch/x86/kernel/cpu/Makefile | 5 +++++ arch/x86/kernel/smpboot.c | 2 +- 3 files changed, 8 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/msr.h b/arch/x86/include/asm/msr.h index c2a812ebde89..b8a1799ea871 100644 --- a/arch/x86/include/asm/msr.h +++ b/arch/x86/include/asm/msr.h @@ -85,7 +85,8 @@ static inline void native_write_msr(unsigned int msr, asm volatile("wrmsr" : : "c" (msr), "a"(low), "d" (high) : "memory"); } -static inline int native_write_msr_safe(unsigned int msr, +/* Can be uninlined because referenced by paravirt */ +notrace static inline int native_write_msr_safe(unsigned int msr, unsigned low, unsigned high) { int err; diff --git a/arch/x86/kernel/cpu/Makefile b/arch/x86/kernel/cpu/Makefile index 82ec6075c057..4ae495a313f3 100644 --- a/arch/x86/kernel/cpu/Makefile +++ b/arch/x86/kernel/cpu/Makefile @@ -2,6 +2,11 @@ # Makefile for x86-compatible CPU details and quirks # +# Don't trace early stages of a secondary CPU boot +ifdef CONFIG_FUNCTION_TRACER +CFLAGS_REMOVE_common.o = -pg +endif + obj-y := intel_cacheinfo.o addon_cpuid_features.o obj-y += proc.o capflags.o powerflags.o common.o diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index f71f96fc9e62..f6174d229024 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -287,7 +287,7 @@ static int __cpuinitdata unsafe_smp; /* * Activate a secondary processor. */ -static void __cpuinit start_secondary(void *unused) +notrace static void __cpuinit start_secondary(void *unused) { /* * Don't put *anything* before cpu_init(), SMP booting is too -- cgit v1.2.3 From 1fcccb008be12ea823aaa392758e1e41fb82de9a Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Date: Tue, 23 Dec 2008 21:50:11 +0530 Subject: x86: traps.c replace #if CONFIG_X86_32 with #ifdef CONFIG_X86_32 Impact: cleanup, avoid warning on X86_64 Fixes this warning on X86_64: CC arch/x86/kernel/traps.o arch/x86/kernel/traps.c:695:5: warning: "CONFIG_X86_32" is not defined Signed-off-by: Jaswinder Singh Signed-off-by: Ingo Molnar --- arch/x86/kernel/traps.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c index c320c29255c2..f37cee75ab58 100644 --- a/arch/x86/kernel/traps.c +++ b/arch/x86/kernel/traps.c @@ -691,7 +691,7 @@ void math_error(void __user *ip) err = swd & ~cwd & 0x3f; -#if CONFIG_X86_32 +#ifdef CONFIG_X86_32 if (!err) return; #endif -- cgit v1.2.3 From 4e17fee24a39448f3a20e9cf98887b7665825848 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Thu, 25 Dec 2008 12:04:17 +0100 Subject: x86: turn CONFIG_SPARSE_IRQ off by default New feature - lets disable it by default first - can flip it around later. Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 1 - 1 file changed, 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 5c243826334a..d14a8806227d 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -241,7 +241,6 @@ config X86_HAS_BOOT_CPU_ID config SPARSE_IRQ bool "Support sparse irq numbering" depends on PCI_MSI || HT_IRQ - default y help This enables support for sparse irq, esp for msi/msi-x. You may need if you have lots of cards supports msi-x installed. -- cgit v1.2.3 From fc5243d98ac2575ad14a974b3c097e9ba874c03d Mon Sep 17 00:00:00 2001 From: Martin Schwidefsky Date: Thu, 25 Dec 2008 13:38:35 +0100 Subject: [S390] arch_setup_additional_pages arguments arch_setup_additional_pages currently gets two arguments, the binary format descripton and an indication if the process uses an executable stack or not. The second argument is not used by anybody, it could be removed without replacement. What actually does make sense is to pass an indication if the process uses the elf interpreter or not. The glibc code will not use anything from the vdso if the process does not use the dynamic linker, so for statically linked binaries the architecture backend can choose not to map the vdso. Acked-by: Ingo Molnar Signed-off-by: Martin Schwidefsky --- arch/powerpc/include/asm/elf.h | 2 +- arch/powerpc/kernel/vdso.c | 3 +-- arch/sh/include/asm/elf.h | 2 +- arch/sh/kernel/vsyscall/vsyscall.c | 3 +-- arch/x86/include/asm/elf.h | 2 +- arch/x86/vdso/vdso32-setup.c | 2 +- arch/x86/vdso/vma.c | 2 +- fs/binfmt_elf.c | 2 +- 8 files changed, 8 insertions(+), 10 deletions(-) (limited to 'arch/x86') diff --git a/arch/powerpc/include/asm/elf.h b/arch/powerpc/include/asm/elf.h index d812929390e4..cd46f023ec6d 100644 --- a/arch/powerpc/include/asm/elf.h +++ b/arch/powerpc/include/asm/elf.h @@ -267,7 +267,7 @@ extern int ucache_bsize; #define ARCH_HAS_SETUP_ADDITIONAL_PAGES struct linux_binprm; extern int arch_setup_additional_pages(struct linux_binprm *bprm, - int executable_stack); + int uses_interp); #define VDSO_AUX_ENT(a,b) NEW_AUX_ENT(a,b); #endif /* __KERNEL__ */ diff --git a/arch/powerpc/kernel/vdso.c b/arch/powerpc/kernel/vdso.c index 65639a43e644..f7ec7d0888fe 100644 --- a/arch/powerpc/kernel/vdso.c +++ b/arch/powerpc/kernel/vdso.c @@ -184,8 +184,7 @@ static void dump_vdso_pages(struct vm_area_struct * vma) * This is called from binfmt_elf, we create the special vma for the * vDSO and insert it into the mm struct tree */ -int arch_setup_additional_pages(struct linux_binprm *bprm, - int executable_stack) +int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp) { struct mm_struct *mm = current->mm; struct page **vdso_pagelist; diff --git a/arch/sh/include/asm/elf.h b/arch/sh/include/asm/elf.h index 9eb9036a1bdc..9381397ebeb8 100644 --- a/arch/sh/include/asm/elf.h +++ b/arch/sh/include/asm/elf.h @@ -204,7 +204,7 @@ do { \ #define ARCH_HAS_SETUP_ADDITIONAL_PAGES struct linux_binprm; extern int arch_setup_additional_pages(struct linux_binprm *bprm, - int executable_stack); + int uses_interp); extern unsigned int vdso_enabled; extern void __kernel_vsyscall; diff --git a/arch/sh/kernel/vsyscall/vsyscall.c b/arch/sh/kernel/vsyscall/vsyscall.c index 95f4de0800ec..3f7e415be86a 100644 --- a/arch/sh/kernel/vsyscall/vsyscall.c +++ b/arch/sh/kernel/vsyscall/vsyscall.c @@ -59,8 +59,7 @@ int __init vsyscall_init(void) } /* Setup a VMA at program startup for the vsyscall page */ -int arch_setup_additional_pages(struct linux_binprm *bprm, - int executable_stack) +int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp) { struct mm_struct *mm = current->mm; unsigned long addr; diff --git a/arch/x86/include/asm/elf.h b/arch/x86/include/asm/elf.h index 40ca1bea7916..f51a3ddde01a 100644 --- a/arch/x86/include/asm/elf.h +++ b/arch/x86/include/asm/elf.h @@ -325,7 +325,7 @@ struct linux_binprm; #define ARCH_HAS_SETUP_ADDITIONAL_PAGES 1 extern int arch_setup_additional_pages(struct linux_binprm *bprm, - int executable_stack); + int uses_interp); extern int syscall32_setup_pages(struct linux_binprm *, int exstack); #define compat_arch_setup_additional_pages syscall32_setup_pages diff --git a/arch/x86/vdso/vdso32-setup.c b/arch/x86/vdso/vdso32-setup.c index 513f330c5832..1241f118ab56 100644 --- a/arch/x86/vdso/vdso32-setup.c +++ b/arch/x86/vdso/vdso32-setup.c @@ -310,7 +310,7 @@ int __init sysenter_setup(void) } /* Setup a VMA at program startup for the vsyscall page */ -int arch_setup_additional_pages(struct linux_binprm *bprm, int exstack) +int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp) { struct mm_struct *mm = current->mm; unsigned long addr; diff --git a/arch/x86/vdso/vma.c b/arch/x86/vdso/vma.c index 257ba4a10abf..9c98cc6ba978 100644 --- a/arch/x86/vdso/vma.c +++ b/arch/x86/vdso/vma.c @@ -98,7 +98,7 @@ static unsigned long vdso_addr(unsigned long start, unsigned len) /* Setup a VMA at program startup for the vsyscall page. Not called for compat tasks */ -int arch_setup_additional_pages(struct linux_binprm *bprm, int exstack) +int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp) { struct mm_struct *mm = current->mm; unsigned long addr; diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c index 8fcfa398d350..95a76ff9e01b 100644 --- a/fs/binfmt_elf.c +++ b/fs/binfmt_elf.c @@ -949,7 +949,7 @@ static int load_elf_binary(struct linux_binprm *bprm, struct pt_regs *regs) set_binfmt(&elf_format); #ifdef ARCH_HAS_SETUP_ADDITIONAL_PAGES - retval = arch_setup_additional_pages(bprm, executable_stack); + retval = arch_setup_additional_pages(bprm, !!elf_interpreter); if (retval < 0) { send_sig(SIGKILL, current, 0); goto out; -- cgit v1.2.3 From 973656fe1afb4adf95d7b9ab75d4660cd3821ea1 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Thu, 25 Dec 2008 16:26:47 +0100 Subject: x86, sparseirq: clean up Kconfig entry Impact: improve help text Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index d14a8806227d..e4c038abb71c 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -242,10 +242,14 @@ config SPARSE_IRQ bool "Support sparse irq numbering" depends on PCI_MSI || HT_IRQ help - This enables support for sparse irq, esp for msi/msi-x. You may need - if you have lots of cards supports msi-x installed. + This enables support for sparse irqs. This is useful for distro + kernels that want to define a high CONFIG_NR_CPUS value but still + want to have low kernel memory footprint on smaller machines. - If you don't know what to do here, say Y. + ( Sparse IRQs can also be beneficial on NUMA boxes, as they spread + out the irq_desc[] array in a more NUMA-friendly way. ) + + If you don't know what to do here, say N. config NUMA_MIGRATE_IRQ_DESC bool "Move irq desc when changing irq smp_affinity" -- cgit v1.2.3 From a73ad3331fdbf4191cf99b83ea9ac7082b6757ba Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Thu, 25 Dec 2008 10:39:01 -0800 Subject: x86: unify the implementation of FPU traps On 32 bits, we may suffer IRQ 13, or supposedly we might have a buggy implementation which gives spurious trap 16. We did not check for this on 64 bits, but there is no reason we can't make the code the same in both cases. Furthermore, this is presumably rare, so do the spurious check last, instead of first. Signed-off-by: H. Peter Anvin --- arch/x86/kernel/traps.c | 11 ++++------- 1 file changed, 4 insertions(+), 7 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c index f37cee75ab58..f5a640ba04bc 100644 --- a/arch/x86/kernel/traps.c +++ b/arch/x86/kernel/traps.c @@ -689,12 +689,7 @@ void math_error(void __user *ip) cwd = get_fpu_cwd(task); swd = get_fpu_swd(task); - err = swd & ~cwd & 0x3f; - -#ifdef CONFIG_X86_32 - if (!err) - return; -#endif + err = swd & ~cwd; if (err & 0x001) { /* Invalid op */ /* @@ -712,7 +707,9 @@ void math_error(void __user *ip) } else if (err & 0x020) { /* Precision */ info.si_code = FPE_FLTRES; } else { - info.si_code = __SI_FAULT|SI_KERNEL; /* WTF? */ + /* If we're using IRQ 13, or supposedly even some trap 16 + implementations, it's possible we get a spurious trap... */ + return; /* Spurious trap, no error */ } force_sig_info(SIGFPE, &info, task); } -- cgit v1.2.3 From 71ab6b245fda6e7597a667a67cce0d26c3c7a14b Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Thu, 25 Dec 2008 17:18:43 +1030 Subject: x86: remove impossible test in mtrr/main.c Impact: cleanup enable_mtrr_cleanup is static, and is never set to anything but 0 or 1. Signed-off-by: Rusty Russell Signed-off-by: H. Peter Anvin --- arch/x86/kernel/cpu/mtrr/main.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/mtrr/main.c b/arch/x86/kernel/cpu/mtrr/main.c index c78c04821ea1..acd9ac54d47f 100644 --- a/arch/x86/kernel/cpu/mtrr/main.c +++ b/arch/x86/kernel/cpu/mtrr/main.c @@ -823,16 +823,14 @@ static int enable_mtrr_cleanup __initdata = static int __init disable_mtrr_cleanup_setup(char *str) { - if (enable_mtrr_cleanup != -1) - enable_mtrr_cleanup = 0; + enable_mtrr_cleanup = 0; return 0; } early_param("disable_mtrr_cleanup", disable_mtrr_cleanup_setup); static int __init enable_mtrr_cleanup_setup(char *str) { - if (enable_mtrr_cleanup != -1) - enable_mtrr_cleanup = 1; + enable_mtrr_cleanup = 1; return 0; } early_param("enable_mtrr_cleanup", enable_mtrr_cleanup_setup); -- cgit v1.2.3 From bd8b96dfc216eebc72950a6c40da8d3eca8667df Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 26 Dec 2008 09:20:22 +0100 Subject: x86: clean up comment style in arch/x86/kernel/traps.c Impact: cleanup Signed-off-by: Ingo Molnar --- arch/x86/kernel/traps.c | 26 +++++++++++++++++--------- 1 file changed, 17 insertions(+), 9 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c index f5a640ba04bc..dbfb80289928 100644 --- a/arch/x86/kernel/traps.c +++ b/arch/x86/kernel/traps.c @@ -292,8 +292,10 @@ dotraplinkage void do_double_fault(struct pt_regs *regs, long error_code) tsk->thread.error_code = error_code; tsk->thread.trap_no = 8; - /* This is always a kernel trap and never fixable (and thus must - never return). */ + /* + * This is always a kernel trap and never fixable (and thus must + * never return). + */ for (;;) die(str, regs, error_code); } @@ -524,9 +526,11 @@ dotraplinkage void __kprobes do_int3(struct pt_regs *regs, long error_code) } #ifdef CONFIG_X86_64 -/* Help handler running on IST stack to switch back to user stack - for scheduling or signal handling. The actual stack switch is done in - entry.S */ +/* + * Help handler running on IST stack to switch back to user stack + * for scheduling or signal handling. The actual stack switch is done in + * entry.S + */ asmlinkage __kprobes struct pt_regs *sync_regs(struct pt_regs *eregs) { struct pt_regs *regs = eregs; @@ -536,8 +540,10 @@ asmlinkage __kprobes struct pt_regs *sync_regs(struct pt_regs *eregs) /* Exception from user space */ else if (user_mode(eregs)) regs = task_pt_regs(current); - /* Exception from kernel and interrupts are enabled. Move to - kernel process stack. */ + /* + * Exception from kernel and interrupts are enabled. Move to + * kernel process stack. + */ else if (eregs->flags & X86_EFLAGS_IF) regs = (struct pt_regs *)(eregs->sp -= sizeof(struct pt_regs)); if (eregs != regs) @@ -707,8 +713,10 @@ void math_error(void __user *ip) } else if (err & 0x020) { /* Precision */ info.si_code = FPE_FLTRES; } else { - /* If we're using IRQ 13, or supposedly even some trap 16 - implementations, it's possible we get a spurious trap... */ + /* + * If we're using IRQ 13, or supposedly even some trap 16 + * implementations, it's possible we get a spurious trap... + */ return; /* Spurious trap, no error */ } force_sig_info(SIGFPE, &info, task); -- cgit v1.2.3 From 18eefedfe8ad33e8fc7614c13359e29a9fab4644 Mon Sep 17 00:00:00 2001 From: KOSAKI Motohiro Date: Fri, 26 Dec 2008 12:29:48 +0900 Subject: irq: simplify for_each_irq_desc() usage Impact: cleanup all for_each_irq_desc() usage point have !desc check. then its check can move into for_each_irq_desc() macro. Signed-off-by: KOSAKI Motohiro Acked-by: Yinghai Lu Signed-off-by: Ingo Molnar --- arch/x86/kernel/io_apic.c | 10 ---------- drivers/xen/events.c | 3 --- include/linux/irqnr.h | 8 ++++++-- kernel/irq/autoprobe.c | 15 --------------- kernel/irq/handle.c | 3 --- kernel/irq/spurious.c | 5 ----- 6 files changed, 6 insertions(+), 38 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index a74887b416cc..2fe543f58ac8 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -1345,8 +1345,6 @@ void __setup_vector_irq(int cpu) /* Mark the inuse vectors */ for_each_irq_desc(irq, desc) { - if (!desc) - continue; cfg = desc->chip_data; if (!cpu_isset(cpu, cfg->domain)) continue; @@ -1730,8 +1728,6 @@ __apicdebuginit(void) print_IO_APIC(void) for_each_irq_desc(irq, desc) { struct irq_pin_list *entry; - if (!desc) - continue; cfg = desc->chip_data; entry = cfg->irq_2_pin; if (!entry) @@ -2378,9 +2374,6 @@ static void ir_irq_migration(struct work_struct *work) struct irq_desc *desc; for_each_irq_desc(irq, desc) { - if (!desc) - continue; - if (desc->status & IRQ_MOVE_PENDING) { unsigned long flags; @@ -2671,9 +2664,6 @@ static inline void init_IO_APIC_traps(void) * 0x80, because int 0x80 is hm, kind of importantish. ;) */ for_each_irq_desc(irq, desc) { - if (!desc) - continue; - cfg = desc->chip_data; if (IO_APIC_IRQ(irq) && cfg && !cfg->vector) { /* diff --git a/drivers/xen/events.c b/drivers/xen/events.c index 46625cd38743..e26733a9df21 100644 --- a/drivers/xen/events.c +++ b/drivers/xen/events.c @@ -142,9 +142,6 @@ static void init_evtchn_cpu_bindings(void) /* By default all event channels notify CPU#0. */ for_each_irq_desc(i, desc) { - if (!desc) - continue; - desc->affinity = cpumask_of_cpu(0); } #endif diff --git a/include/linux/irqnr.h b/include/linux/irqnr.h index c4a59c7a478b..5504a5c97836 100644 --- a/include/linux/irqnr.h +++ b/include/linux/irqnr.h @@ -22,10 +22,14 @@ extern struct irq_desc *irq_to_desc(unsigned int irq); # define for_each_irq_desc(irq, desc) \ for (irq = 0, desc = irq_to_desc(irq); irq < nr_irqs; \ - irq++, desc = irq_to_desc(irq)) + irq++, desc = irq_to_desc(irq)) \ + if (desc) + + # define for_each_irq_desc_reverse(irq, desc) \ for (irq = nr_irqs - 1, desc = irq_to_desc(irq); irq >= 0; \ - irq--, desc = irq_to_desc(irq)) + irq--, desc = irq_to_desc(irq)) \ + if (desc) #endif /* CONFIG_GENERIC_HARDIRQS */ diff --git a/kernel/irq/autoprobe.c b/kernel/irq/autoprobe.c index 650ce4102a63..cc0f7321b8ce 100644 --- a/kernel/irq/autoprobe.c +++ b/kernel/irq/autoprobe.c @@ -40,9 +40,6 @@ unsigned long probe_irq_on(void) * flush such a longstanding irq before considering it as spurious. */ for_each_irq_desc_reverse(i, desc) { - if (!desc) - continue; - spin_lock_irq(&desc->lock); if (!desc->action && !(desc->status & IRQ_NOPROBE)) { /* @@ -71,9 +68,6 @@ unsigned long probe_irq_on(void) * happened in the previous stage, it may have masked itself) */ for_each_irq_desc_reverse(i, desc) { - if (!desc) - continue; - spin_lock_irq(&desc->lock); if (!desc->action && !(desc->status & IRQ_NOPROBE)) { desc->status |= IRQ_AUTODETECT | IRQ_WAITING; @@ -92,9 +86,6 @@ unsigned long probe_irq_on(void) * Now filter out any obviously spurious interrupts */ for_each_irq_desc(i, desc) { - if (!desc) - continue; - spin_lock_irq(&desc->lock); status = desc->status; @@ -133,9 +124,6 @@ unsigned int probe_irq_mask(unsigned long val) int i; for_each_irq_desc(i, desc) { - if (!desc) - continue; - spin_lock_irq(&desc->lock); status = desc->status; @@ -178,9 +166,6 @@ int probe_irq_off(unsigned long val) unsigned int status; for_each_irq_desc(i, desc) { - if (!desc) - continue; - spin_lock_irq(&desc->lock); status = desc->status; diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index 03479dfdebb8..7dbdfe524693 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -437,9 +437,6 @@ void early_init_irq_lock_class(void) int i; for_each_irq_desc(i, desc) { - if (!desc) - continue; - lockdep_set_class(&desc->lock, &irq_desc_lock_class); } } diff --git a/kernel/irq/spurious.c b/kernel/irq/spurious.c index 3738107531fd..dd364c11e56e 100644 --- a/kernel/irq/spurious.c +++ b/kernel/irq/spurious.c @@ -91,9 +91,6 @@ static int misrouted_irq(int irq) int i, ok = 0; for_each_irq_desc(i, desc) { - if (!desc) - continue; - if (!i) continue; @@ -115,8 +112,6 @@ static void poll_spurious_irqs(unsigned long dummy) for_each_irq_desc(i, desc) { unsigned int status; - if (!desc) - continue; if (!i) continue; -- cgit v1.2.3 From 393d68fb9929817cde7ab31c82d66fcb28ad35fc Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Fri, 26 Dec 2008 22:23:38 +1030 Subject: cpumask: x86: Introduce cpumask_of_{node,pcibus} to replace {node,pcibus}_to_cpumask Impact: New APIs The old node_to_cpumask/node_to_pcibus returned a cpumask_t: these return a pointer to a struct cpumask. Part of removing cpumasks from the stack. Also makes __pcibus_to_node take a const pointer. Signed-off-by: Rusty Russell Acked-by: Ingo Molnar --- arch/x86/include/asm/pci.h | 10 ++++++++-- arch/x86/include/asm/topology.h | 35 +++++++++++++++++++++++------------ arch/x86/kernel/setup_percpu.c | 8 ++++---- 3 files changed, 35 insertions(+), 18 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/pci.h b/arch/x86/include/asm/pci.h index 875b38edf193..52d80d3d94f3 100644 --- a/arch/x86/include/asm/pci.h +++ b/arch/x86/include/asm/pci.h @@ -98,9 +98,9 @@ static inline void early_quirks(void) { } #ifdef CONFIG_NUMA /* Returns the node based on pci bus */ -static inline int __pcibus_to_node(struct pci_bus *bus) +static inline int __pcibus_to_node(const struct pci_bus *bus) { - struct pci_sysdata *sd = bus->sysdata; + const struct pci_sysdata *sd = bus->sysdata; return sd->node; } @@ -109,6 +109,12 @@ static inline cpumask_t __pcibus_to_cpumask(struct pci_bus *bus) { return node_to_cpumask(__pcibus_to_node(bus)); } + +static inline const struct cpumask * +cpumask_of_pcibus(const struct pci_bus *bus) +{ + return cpumask_of_node(__pcibus_to_node(bus)); +} #endif #endif /* _ASM_X86_PCI_H */ diff --git a/arch/x86/include/asm/topology.h b/arch/x86/include/asm/topology.h index ff386ff50ed7..45da5dc50fc8 100644 --- a/arch/x86/include/asm/topology.h +++ b/arch/x86/include/asm/topology.h @@ -61,13 +61,19 @@ static inline int cpu_to_node(int cpu) * * Side note: this function creates the returned cpumask on the stack * so with a high NR_CPUS count, excessive stack space is used. The - * node_to_cpumask_ptr function should be used whenever possible. + * cpumask_of_node function should be used whenever possible. */ static inline cpumask_t node_to_cpumask(int node) { return node_to_cpumask_map[node]; } +/* Returns a bitmask of CPUs on Node 'node'. */ +static inline const struct cpumask *cpumask_of_node(int node) +{ + return &node_to_cpumask_map[node]; +} + #else /* CONFIG_X86_64 */ /* Mappings between node number and cpus on that node. */ @@ -82,7 +88,7 @@ DECLARE_EARLY_PER_CPU(int, x86_cpu_to_node_map); #ifdef CONFIG_DEBUG_PER_CPU_MAPS extern int cpu_to_node(int cpu); extern int early_cpu_to_node(int cpu); -extern const cpumask_t *_node_to_cpumask_ptr(int node); +extern const cpumask_t *cpumask_of_node(int node); extern cpumask_t node_to_cpumask(int node); #else /* !CONFIG_DEBUG_PER_CPU_MAPS */ @@ -103,7 +109,7 @@ static inline int early_cpu_to_node(int cpu) } /* Returns a pointer to the cpumask of CPUs on Node 'node'. */ -static inline const cpumask_t *_node_to_cpumask_ptr(int node) +static inline const cpumask_t *cpumask_of_node(int node) { return &node_to_cpumask_map[node]; } @@ -116,12 +122,15 @@ static inline cpumask_t node_to_cpumask(int node) #endif /* !CONFIG_DEBUG_PER_CPU_MAPS */ -/* Replace default node_to_cpumask_ptr with optimized version */ +/* + * Replace default node_to_cpumask_ptr with optimized version + * Deprecated: use "const struct cpumask *mask = cpumask_of_node(node)" + */ #define node_to_cpumask_ptr(v, node) \ - const cpumask_t *v = _node_to_cpumask_ptr(node) + const cpumask_t *v = cpumask_of_node(node) #define node_to_cpumask_ptr_next(v, node) \ - v = _node_to_cpumask_ptr(node) + v = cpumask_of_node(node) #endif /* CONFIG_X86_64 */ @@ -187,7 +196,7 @@ extern int __node_distance(int, int); #define cpu_to_node(cpu) 0 #define early_cpu_to_node(cpu) 0 -static inline const cpumask_t *_node_to_cpumask_ptr(int node) +static inline const cpumask_t *cpumask_of_node(int node) { return &cpu_online_map; } @@ -200,12 +209,15 @@ static inline int node_to_first_cpu(int node) return first_cpu(cpu_online_map); } -/* Replace default node_to_cpumask_ptr with optimized version */ +/* + * Replace default node_to_cpumask_ptr with optimized version + * Deprecated: use "const struct cpumask *mask = cpumask_of_node(node)" + */ #define node_to_cpumask_ptr(v, node) \ - const cpumask_t *v = _node_to_cpumask_ptr(node) + const cpumask_t *v = cpumask_of_node(node) #define node_to_cpumask_ptr_next(v, node) \ - v = _node_to_cpumask_ptr(node) + v = cpumask_of_node(node) #endif #include @@ -214,8 +226,7 @@ static inline int node_to_first_cpu(int node) /* Returns the number of the first CPU on Node 'node'. */ static inline int node_to_first_cpu(int node) { - node_to_cpumask_ptr(mask, node); - return first_cpu(*mask); + return cpumask_first(cpumask_of_node(node)); } #endif diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c index 1c2084291f97..8e8b1193add5 100644 --- a/arch/x86/kernel/setup_percpu.c +++ b/arch/x86/kernel/setup_percpu.c @@ -334,25 +334,25 @@ static const cpumask_t cpu_mask_none; /* * Returns a pointer to the bitmask of CPUs on Node 'node'. */ -const cpumask_t *_node_to_cpumask_ptr(int node) +const cpumask_t *cpumask_of_node(int node) { if (node_to_cpumask_map == NULL) { printk(KERN_WARNING - "_node_to_cpumask_ptr(%d): no node_to_cpumask_map!\n", + "cpumask_of_node(%d): no node_to_cpumask_map!\n", node); dump_stack(); return (const cpumask_t *)&cpu_online_map; } if (node >= nr_node_ids) { printk(KERN_WARNING - "_node_to_cpumask_ptr(%d): node > nr_node_ids(%d)\n", + "cpumask_of_node(%d): node > nr_node_ids(%d)\n", node, nr_node_ids); dump_stack(); return &cpu_mask_none; } return &node_to_cpumask_map[node]; } -EXPORT_SYMBOL(_node_to_cpumask_ptr); +EXPORT_SYMBOL(cpumask_of_node); /* * Returns a bitmask of CPUs on Node 'node'. -- cgit v1.2.3 From 030bb203e01db12e3f2866799f4f03a114d06349 Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Fri, 26 Dec 2008 22:23:41 +1030 Subject: cpumask: cpu_coregroup_mask(): x86 Impact: New API Like cpu_coregroup_map, but returns a (const) pointer. Signed-off-by: Rusty Russell Signed-off-by: Mike Travis Cc: Ingo Molnar --- arch/x86/include/asm/topology.h | 1 + arch/x86/kernel/smpboot.c | 11 ++++++++--- 2 files changed, 9 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/topology.h b/arch/x86/include/asm/topology.h index 45da5dc50fc8..168203c0c316 100644 --- a/arch/x86/include/asm/topology.h +++ b/arch/x86/include/asm/topology.h @@ -231,6 +231,7 @@ static inline int node_to_first_cpu(int node) #endif extern cpumask_t cpu_coregroup_map(int cpu); +extern const struct cpumask *cpu_coregroup_mask(int cpu); #ifdef ENABLE_TOPO_DEFINES #define topology_physical_package_id(cpu) (cpu_data(cpu).phys_proc_id) diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index 468c2f9d47ae..d5274b6b088e 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -497,7 +497,7 @@ void __cpuinit set_cpu_sibling_map(int cpu) } /* maps the cpu to the sched domain representing multi-core */ -cpumask_t cpu_coregroup_map(int cpu) +const struct cpumask *cpu_coregroup_mask(int cpu) { struct cpuinfo_x86 *c = &cpu_data(cpu); /* @@ -505,9 +505,14 @@ cpumask_t cpu_coregroup_map(int cpu) * And for power savings, we return cpu_core_map */ if (sched_mc_power_savings || sched_smt_power_savings) - return per_cpu(cpu_core_map, cpu); + return &per_cpu(cpu_core_map, cpu); else - return c->llc_shared_map; + return &c->llc_shared_map; +} + +cpumask_t cpu_coregroup_map(int cpu) +{ + return *cpu_coregroup_mask(cpu); } static void impress_friends(void) -- cgit v1.2.3 From 0b936bfdeb85784b7df132b2c64fb34ad9b11ffa Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Date: Tue, 23 Dec 2008 21:51:28 +0530 Subject: x86: reboot.c declare port_cf9_safe before they get used Impact: cleanup, avoid sparse warning Include "../pci/pci.h" for port_cf9_safe Fixes this sparse warning: arch/x86/kernel/reboot.c:43:6: warning: symbol 'port_cf9_safe' was not declared. Should it be static? Signed-off-by: Jaswinder Singh Signed-off-by: Ingo Molnar --- arch/x86/kernel/reboot.c | 1 + 1 file changed, 1 insertion(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c index 61f718df6eec..b165eb0884ed 100644 --- a/arch/x86/kernel/reboot.c +++ b/arch/x86/kernel/reboot.c @@ -22,6 +22,7 @@ #endif #include +#include "../pci/pci.h" /* -- cgit v1.2.3 From b6b301aa9fba57b114c3a00f5f43abf672bd4ecd Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Date: Tue, 23 Dec 2008 21:52:33 +0530 Subject: x86: apic.c x2apic_preenabled and disable_x2apic should be static Impact: cleanup, reduce kernel size a bit, avoid sparse warning Fixes sparse warning: arch/x86/kernel/apic.c:103:5: warning: symbol 'disable_x2apic' was not declared. Should it be static? Signed-off-by: Jaswinder Singh Signed-off-by: Ingo Molnar --- arch/x86/include/asm/apic.h | 2 +- arch/x86/kernel/apic.c | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/apic.h b/arch/x86/include/asm/apic.h index 25caa0738af5..e644bf6f90dc 100644 --- a/arch/x86/include/asm/apic.h +++ b/arch/x86/include/asm/apic.h @@ -93,7 +93,7 @@ static inline u32 native_apic_msr_read(u32 reg) } #ifndef CONFIG_X86_32 -extern int x2apic, x2apic_preenabled; +extern int x2apic; extern void check_x2apic(void); extern void enable_x2apic(void); extern void enable_IR_x2apic(void); diff --git a/arch/x86/kernel/apic.c b/arch/x86/kernel/apic.c index 7397911f8478..3a961bd9f619 100644 --- a/arch/x86/kernel/apic.c +++ b/arch/x86/kernel/apic.c @@ -97,8 +97,8 @@ __setup("apicpmtimer", setup_apicpmtimer); #ifdef HAVE_X2APIC int x2apic; /* x2apic enabled before OS handover */ -int x2apic_preenabled; -int disable_x2apic; +static int x2apic_preenabled; +static int disable_x2apic; static __init int setup_nox2apic(char *str) { disable_x2apic = 1; -- cgit v1.2.3 From 13a0c3c269b223f60abfac8a9811d77111a8b4ba Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Fri, 26 Dec 2008 02:05:47 -0800 Subject: sparseirq: work around compiler optimizing away __weak functions Impact: fix panic on null pointer with sparseirq Some GCC versions seem to inline the weak global function, when that function is empty. Work it around, by making the functions return a (dummy) integer. Signed-off-by: Yinghai Signed-off-by: Ingo Molnar --- arch/x86/kernel/io_apic.c | 8 ++++++-- include/linux/irq.h | 6 +++--- init/main.c | 7 ++++--- kernel/irq/handle.c | 7 ++++--- 4 files changed, 17 insertions(+), 11 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index 2fe543f58ac8..976039377846 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -170,7 +170,7 @@ static struct irq_cfg irq_cfgx[NR_IRQS] = { [15] = { .domain = CPU_MASK_ALL, .vector = IRQ15_VECTOR, }, }; -void __init arch_early_irq_init(void) +int __init arch_early_irq_init(void) { struct irq_cfg *cfg; struct irq_desc *desc; @@ -184,6 +184,8 @@ void __init arch_early_irq_init(void) desc = irq_to_desc(i); desc->chip_data = &cfg[i]; } + + return 0; } #ifdef CONFIG_SPARSE_IRQ @@ -212,7 +214,7 @@ static struct irq_cfg *get_one_free_irq_cfg(int cpu) return cfg; } -void arch_init_chip_data(struct irq_desc *desc, int cpu) +int arch_init_chip_data(struct irq_desc *desc, int cpu) { struct irq_cfg *cfg; @@ -224,6 +226,8 @@ void arch_init_chip_data(struct irq_desc *desc, int cpu) BUG_ON(1); } } + + return 0; } #ifdef CONFIG_NUMA_MIGRATE_IRQ_DESC diff --git a/include/linux/irq.h b/include/linux/irq.h index 69da275c0ebd..0e40af4bac40 100644 --- a/include/linux/irq.h +++ b/include/linux/irq.h @@ -193,9 +193,9 @@ struct irq_desc { const char *name; } ____cacheline_internodealigned_in_smp; -extern void early_irq_init(void); -extern void arch_early_irq_init(void); -extern void arch_init_chip_data(struct irq_desc *desc, int cpu); +extern int early_irq_init(void); +extern int arch_early_irq_init(void); +extern int arch_init_chip_data(struct irq_desc *desc, int cpu); extern void arch_init_copy_chip_data(struct irq_desc *old_desc, struct irq_desc *desc, int cpu); extern void arch_free_chip_data(struct irq_desc *old_desc, struct irq_desc *desc); diff --git a/init/main.c b/init/main.c index c1f999a3cf31..c314aa15370e 100644 --- a/init/main.c +++ b/init/main.c @@ -539,13 +539,14 @@ void __init __weak thread_info_cache_init(void) { } -void __init __weak arch_early_irq_init(void) +int __init __weak arch_early_irq_init(void) { + return 0; } -void __init __weak early_irq_init(void) +int __init __weak early_irq_init(void) { - arch_early_irq_init(); + return arch_early_irq_init(); } asmlinkage void __init start_kernel(void) diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index 893da67b7781..0bef3ecb7a0e 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -86,8 +86,9 @@ void init_kstat_irqs(struct irq_desc *desc, int cpu, int nr) desc->kstat_irqs = (unsigned int *)ptr; } -void __attribute__((weak)) arch_init_chip_data(struct irq_desc *desc, int cpu) +int __weak arch_init_chip_data(struct irq_desc *desc, int cpu) { + return 0; } static void init_one_irq_desc(int irq, struct irq_desc *desc, int cpu) @@ -132,7 +133,7 @@ static struct irq_desc irq_desc_legacy[NR_IRQS_LEGACY] __cacheline_aligned_in_sm /* FIXME: use bootmem alloc ...*/ static unsigned int kstat_irqs_legacy[NR_IRQS_LEGACY][NR_CPUS]; -void __init early_irq_init(void) +int __init early_irq_init(void) { struct irq_desc *desc; int legacy_count; @@ -151,7 +152,7 @@ void __init early_irq_init(void) for (i = legacy_count; i < NR_IRQS; i++) irq_desc_ptrs[i] = NULL; - arch_early_irq_init(); + return arch_early_irq_init(); } struct irq_desc *irq_to_desc(unsigned int irq) -- cgit v1.2.3 From 6092848a2a23b660150a38bc06f59d75838d70c8 Mon Sep 17 00:00:00 2001 From: Sergio Luis Date: Sun, 28 Dec 2008 04:12:26 -0300 Subject: x86: mark get_cpu_leaves() with __cpuinit annotation Impact: fix section mismatch warning Commit b2bb85549134c005e997e5a7ed303bda6a1ae738 ("x86: Remove cpumask games in x86/kernel/cpu/intel_cacheinfo.c") introduced get_cpu_leaves(), which references __cpuinit cpuid4_cache_lookup(). Mark get_cpu_leaves() with a __cpuinit annotation. Signed-off-by: Sergio Luis Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/intel_cacheinfo.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/intel_cacheinfo.c b/arch/x86/kernel/cpu/intel_cacheinfo.c index fb7f946cb65e..7bd00a565672 100644 --- a/arch/x86/kernel/cpu/intel_cacheinfo.c +++ b/arch/x86/kernel/cpu/intel_cacheinfo.c @@ -534,7 +534,7 @@ static void __cpuinit free_cache_attributes(unsigned int cpu) per_cpu(cpuid4_info, cpu) = NULL; } -static void get_cpu_leaves(void *_retval) +static void __cpuinit get_cpu_leaves(void *_retval) { int j, *retval = _retval, cpu = smp_processor_id(); -- cgit v1.2.3 From c805b7300ed20ec4f10ea385988d6d3fa935b26c Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Sat, 27 Dec 2008 17:10:18 +0300 Subject: x86: mach-default setup.c cleanups Impact: cleanup - Break long lines into shorter form. - Use pr_ macros instead of plain printk. Signed-off-by: Cyrill Gorcunov Signed-off-by: Ingo Molnar --- arch/x86/mach-default/setup.c | 15 +++++++-------- 1 file changed, 7 insertions(+), 8 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/mach-default/setup.c b/arch/x86/mach-default/setup.c index 37b9ae4d44c5..df167f265622 100644 --- a/arch/x86/mach-default/setup.c +++ b/arch/x86/mach-default/setup.c @@ -133,29 +133,28 @@ void __init time_init_hook(void) **/ void mca_nmi_hook(void) { - /* If I recall correctly, there's a whole bunch of other things that + /* + * If I recall correctly, there's a whole bunch of other things that * we can do to check for NMI problems, but that's all I know about * at the moment. */ - - printk("NMI generated from unknown source!\n"); + pr_warning("NMI generated from unknown source!\n"); } #endif static __init int no_ipi_broadcast(char *str) { get_option(&str, &no_broadcast); - printk ("Using %s mode\n", no_broadcast ? "No IPI Broadcast" : - "IPI Broadcast"); + pr_info("Using %s mode\n", + no_broadcast ? "No IPI Broadcast" : "IPI Broadcast"); return 1; } - __setup("no_ipi_broadcast=", no_ipi_broadcast); static int __init print_ipi_mode(void) { - printk ("Using IPI %s mode\n", no_broadcast ? "No-Shortcut" : - "Shortcut"); + pr_info("Using IPI %s mode\n", + no_broadcast ? "No-Shortcut" : "Shortcut"); return 0; } -- cgit v1.2.3 From 2f06de0671096e19350c9efe21cfdbc0891aab20 Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Sat, 27 Dec 2008 21:37:10 +0530 Subject: x86: introducing asm/sys_ia32.h Impact: cleanup, avoid 44 sparse warnings, new file asm/sys_ia32.h Fixes following sparse warnings: CHECK arch/x86/ia32/sys_ia32.c arch/x86/ia32/sys_ia32.c:53:17: warning: symbol 'sys32_truncate64' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:60:17: warning: symbol 'sys32_ftruncate64' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:98:17: warning: symbol 'sys32_stat64' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:109:17: warning: symbol 'sys32_lstat64' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:119:17: warning: symbol 'sys32_fstat64' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:128:17: warning: symbol 'sys32_fstatat' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:164:17: warning: symbol 'sys32_mmap' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:195:17: warning: symbol 'sys32_mprotect' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:201:17: warning: symbol 'sys32_pipe' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:215:17: warning: symbol 'sys32_rt_sigaction' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:291:17: warning: symbol 'sys32_sigaction' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:330:17: warning: symbol 'sys32_rt_sigprocmask' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:370:17: warning: symbol 'sys32_alarm' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:383:17: warning: symbol 'sys32_old_select' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:393:17: warning: symbol 'sys32_waitpid' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:401:17: warning: symbol 'sys32_sysfs' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:406:17: warning: symbol 'sys32_sched_rr_get_interval' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:421:17: warning: symbol 'sys32_rt_sigpending' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:445:17: warning: symbol 'sys32_rt_sigqueueinfo' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:472:17: warning: symbol 'sys32_sysctl' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:517:17: warning: symbol 'sys32_pread' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:524:17: warning: symbol 'sys32_pwrite' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:532:17: warning: symbol 'sys32_personality' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:545:17: warning: symbol 'sys32_sendfile' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:565:17: warning: symbol 'sys32_mmap2' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:589:17: warning: symbol 'sys32_olduname' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:626:6: warning: symbol 'sys32_uname' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:641:6: warning: symbol 'sys32_ustat' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:663:17: warning: symbol 'sys32_execve' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:678:17: warning: symbol 'sys32_clone' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:693:6: warning: symbol 'sys32_lseek' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:698:6: warning: symbol 'sys32_kill' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:703:6: warning: symbol 'sys32_fadvise64_64' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:712:6: warning: symbol 'sys32_vm86_warning' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:726:6: warning: symbol 'sys32_lookup_dcookie' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:732:20: warning: symbol 'sys32_readahead' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:738:17: warning: symbol 'sys32_sync_file_range' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:746:17: warning: symbol 'sys32_fadvise64' was not declared. Should it be static? arch/x86/ia32/sys_ia32.c:753:17: warning: symbol 'sys32_fallocate' was not declared. Should it be static? CHECK arch/x86/ia32/ia32_signal.c arch/x86/ia32/ia32_signal.c:126:17: warning: symbol 'sys32_sigsuspend' was not declared. Should it be static? arch/x86/ia32/ia32_signal.c:141:17: warning: symbol 'sys32_sigaltstack' was not declared. Should it be static? arch/x86/ia32/ia32_signal.c:249:17: warning: symbol 'sys32_sigreturn' was not declared. Should it be static? arch/x86/ia32/ia32_signal.c:279:17: warning: symbol 'sys32_rt_sigreturn' was not declared. Should it be static? CHECK arch/x86/ia32/ipc32.c arch/x86/ia32/ipc32.c:12:17: warning: symbol 'sys32_ipc' was not declared. Should it be static? Signed-off-by: Jaswinder Singh Rajput Signed-off-by: Ingo Molnar --- arch/x86/ia32/ia32_signal.c | 3 +- arch/x86/ia32/ipc32.c | 1 + arch/x86/ia32/sys_ia32.c | 2 +- arch/x86/include/asm/sys_ia32.h | 101 ++++++++++++++++++++++++++++++++++++++++ 4 files changed, 104 insertions(+), 3 deletions(-) create mode 100644 arch/x86/include/asm/sys_ia32.h (limited to 'arch/x86') diff --git a/arch/x86/ia32/ia32_signal.c b/arch/x86/ia32/ia32_signal.c index b195f85526e3..9dabd00e9805 100644 --- a/arch/x86/ia32/ia32_signal.c +++ b/arch/x86/ia32/ia32_signal.c @@ -24,15 +24,14 @@ #include #include #include -#include #include #include #include #include #include #include - #include +#include #define DEBUG_SIG 0 diff --git a/arch/x86/ia32/ipc32.c b/arch/x86/ia32/ipc32.c index d21991ce606c..29cdcd02ead3 100644 --- a/arch/x86/ia32/ipc32.c +++ b/arch/x86/ia32/ipc32.c @@ -8,6 +8,7 @@ #include #include #include +#include asmlinkage long sys32_ipc(u32 call, int first, int second, int third, compat_uptr_t ptr, u32 fifth) diff --git a/arch/x86/ia32/sys_ia32.c b/arch/x86/ia32/sys_ia32.c index 2e09dcd3c0a6..6c0d7f6231af 100644 --- a/arch/x86/ia32/sys_ia32.c +++ b/arch/x86/ia32/sys_ia32.c @@ -44,8 +44,8 @@ #include #include #include -#include #include +#include #define AA(__x) ((unsigned long)(__x)) diff --git a/arch/x86/include/asm/sys_ia32.h b/arch/x86/include/asm/sys_ia32.h new file mode 100644 index 000000000000..ffb08be2a530 --- /dev/null +++ b/arch/x86/include/asm/sys_ia32.h @@ -0,0 +1,101 @@ +/* + * sys_ia32.h - Linux ia32 syscall interfaces + * + * Copyright (c) 2008 Jaswinder Singh Rajput + * + * This file is released under the GPLv2. + * See the file COPYING for more details. + */ + +#ifndef _ASM_X86_SYS_IA32_H +#define _ASM_X86_SYS_IA32_H + +#include +#include +#include +#include +#include +#include + +/* ia32/sys_ia32.c */ +asmlinkage long sys32_truncate64(char __user *, unsigned long, unsigned long); +asmlinkage long sys32_ftruncate64(unsigned int, unsigned long, unsigned long); + +asmlinkage long sys32_stat64(char __user *, struct stat64 __user *); +asmlinkage long sys32_lstat64(char __user *, struct stat64 __user *); +asmlinkage long sys32_fstat64(unsigned int, struct stat64 __user *); +asmlinkage long sys32_fstatat(unsigned int, char __user *, + struct stat64 __user *, int); +struct mmap_arg_struct; +asmlinkage long sys32_mmap(struct mmap_arg_struct __user *); +asmlinkage long sys32_mprotect(unsigned long, size_t, unsigned long); + +asmlinkage long sys32_pipe(int __user *); +struct sigaction32; +struct old_sigaction32; +asmlinkage long sys32_rt_sigaction(int, struct sigaction32 __user *, + struct sigaction32 __user *, unsigned int); +asmlinkage long sys32_sigaction(int, struct old_sigaction32 __user *, + struct old_sigaction32 __user *); +asmlinkage long sys32_rt_sigprocmask(int, compat_sigset_t __user *, + compat_sigset_t __user *, unsigned int); +asmlinkage long sys32_alarm(unsigned int); + +struct sel_arg_struct; +asmlinkage long sys32_old_select(struct sel_arg_struct __user *); +asmlinkage long sys32_waitpid(compat_pid_t, unsigned int *, int); +asmlinkage long sys32_sysfs(int, u32, u32); + +asmlinkage long sys32_sched_rr_get_interval(compat_pid_t, + struct compat_timespec __user *); +asmlinkage long sys32_rt_sigpending(compat_sigset_t __user *, compat_size_t); +asmlinkage long sys32_rt_sigqueueinfo(int, int, compat_siginfo_t __user *); + +#ifdef CONFIG_SYSCTL_SYSCALL +struct sysctl_ia32; +asmlinkage long sys32_sysctl(struct sysctl_ia32 __user *); +#endif + +asmlinkage long sys32_pread(unsigned int, char __user *, u32, u32, u32); +asmlinkage long sys32_pwrite(unsigned int, char __user *, u32, u32, u32); + +asmlinkage long sys32_personality(unsigned long); +asmlinkage long sys32_sendfile(int, int, compat_off_t __user *, s32); + +asmlinkage long sys32_mmap2(unsigned long, unsigned long, unsigned long, + unsigned long, unsigned long, unsigned long); + +struct oldold_utsname; +struct old_utsname; +asmlinkage long sys32_olduname(struct oldold_utsname __user *); +long sys32_uname(struct old_utsname __user *); + +long sys32_ustat(unsigned, struct ustat32 __user *); + +asmlinkage long sys32_execve(char __user *, compat_uptr_t __user *, + compat_uptr_t __user *, struct pt_regs *); +asmlinkage long sys32_clone(unsigned int, unsigned int, struct pt_regs *); + +long sys32_lseek(unsigned int, int, unsigned int); +long sys32_kill(int, int); +long sys32_fadvise64_64(int, __u32, __u32, __u32, __u32, int); +long sys32_vm86_warning(void); +long sys32_lookup_dcookie(u32, u32, char __user *, size_t); + +asmlinkage ssize_t sys32_readahead(int, unsigned, unsigned, size_t); +asmlinkage long sys32_sync_file_range(int, unsigned, unsigned, + unsigned, unsigned, int); +asmlinkage long sys32_fadvise64(int, unsigned, unsigned, size_t, int); +asmlinkage long sys32_fallocate(int, int, unsigned, + unsigned, unsigned, unsigned); + +/* ia32/ia32_signal.c */ +asmlinkage long sys32_sigsuspend(int, int, old_sigset_t); +asmlinkage long sys32_sigaltstack(const stack_ia32_t __user *, + stack_ia32_t __user *, struct pt_regs *); +asmlinkage long sys32_sigreturn(struct pt_regs *); +asmlinkage long sys32_rt_sigreturn(struct pt_regs *); + +/* ia32/ipc32.c */ +asmlinkage long sys32_ipc(u32, int, int, int, compat_uptr_t, u32); +#endif /* _ASM_X86_SYS_IA32_H */ -- cgit v1.2.3 From a1ae299dfb6ef219b296b61d1f222732391973b5 Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Mon, 29 Dec 2008 20:32:52 +0530 Subject: x86: apic.c declare pic_mode before they get used Impact: cleanup, avoid sparse warning In asm/mpspec.h moved out pic_mode from CONFIG_X86_32 as it is common for both 32 and 64 bit. Fixes this sparse warning for x86_64: arch/x86/kernel/apic.c:128:5: warning: symbol 'pic_mode' was not declared. Should it be static? Signed-off-by: Jaswinder Singh Rajput Signed-off-by: Ingo Molnar --- arch/x86/include/asm/mpspec.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/mpspec.h b/arch/x86/include/asm/mpspec.h index 91885c28f66b..62d14ce3cd00 100644 --- a/arch/x86/include/asm/mpspec.h +++ b/arch/x86/include/asm/mpspec.h @@ -6,13 +6,13 @@ #include extern int apic_version[MAX_APICS]; +extern int pic_mode; #ifdef CONFIG_X86_32 #include extern unsigned int def_to_bigsmp; extern u8 apicid_2_node[]; -extern int pic_mode; #ifdef CONFIG_X86_NUMAQ extern int mp_bus_id_to_node[MAX_MP_BUSSES]; -- cgit v1.2.3 From 7f3e632f9d8d234819bcdef7a68fc8b84f7d3d3d Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Mon, 29 Dec 2008 20:34:35 +0530 Subject: x86: io_apic.c io_apic_sync should be static Impact: cleanup, reduce kernel size a bit, avoid sparse warning Fixes sparse warning: arch/x86/kernel/io_apic.c:709:6: warning: symbol 'io_apic_sync' was not declared. Should it be static? Signed-off-by: Jaswinder Singh Rajput Signed-off-by: Ingo Molnar --- arch/x86/kernel/io_apic.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index 679e7bbbbcd6..b8c8a8e99341 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -481,7 +481,7 @@ static void __unmask_IO_APIC_irq(unsigned int irq) } #ifdef CONFIG_X86_64 -void io_apic_sync(struct irq_pin_list *entry) +static void io_apic_sync(struct irq_pin_list *entry) { /* * Synchronize the IO-APIC and the CPU by doing -- cgit v1.2.3 From cbafbc826bf645f7fbbfbb2ff20138e5ccb4700e Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Mon, 29 Dec 2008 20:36:40 +0530 Subject: x86: efi.c declare add_efi_memmap before they get used Impact: cleanup, avoid sparse warning Fixes this sparse warning: arch/x86/kernel/efi.c:67:5: warning: symbol 'add_efi_memmap' was not declared. Should it be static? Signed-off-by: Jaswinder Singh Rajput Signed-off-by: Ingo Molnar --- arch/x86/include/asm/efi.h | 1 + 1 file changed, 1 insertion(+) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/efi.h b/arch/x86/include/asm/efi.h index a2e545c91c35..ca5ffb2856b6 100644 --- a/arch/x86/include/asm/efi.h +++ b/arch/x86/include/asm/efi.h @@ -90,6 +90,7 @@ extern void __iomem *efi_ioremap(unsigned long addr, unsigned long size); #endif /* CONFIG_X86_32 */ +extern int add_efi_memmap; extern void efi_reserve_early(void); extern void efi_call_phys_prelog(void); extern void efi_call_phys_epilog(void); -- cgit v1.2.3 From c854c91979e0717c619bc55e124d41d60d5eb3d6 Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Mon, 29 Dec 2008 20:38:09 +0530 Subject: x86_64: pci-gart_64.c iommu_fullflush should be static Impact: cleanup, reduce kernel size a bit, avoid sparse warning Fixes sparse warning: arch/x86/kernel/pci-gart_64.c:55:5: warning: symbol 'iommu_fullflush' was not declared. Should it be static? Signed-off-by: Jaswinder Singh Rajput Signed-off-by: Ingo Molnar --- arch/x86/kernel/pci-gart_64.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/pci-gart_64.c b/arch/x86/kernel/pci-gart_64.c index a35eaa379ff6..00c2bcd41463 100644 --- a/arch/x86/kernel/pci-gart_64.c +++ b/arch/x86/kernel/pci-gart_64.c @@ -52,7 +52,7 @@ static u32 *iommu_gatt_base; /* Remapping table */ * to trigger bugs with some popular PCI cards, in particular 3ware (but * has been also also seen with Qlogic at least). */ -int iommu_fullflush = 1; +static int iommu_fullflush = 1; /* Allocation bitmap for the remapping area: */ static DEFINE_SPINLOCK(iommu_bitmap_lock); -- cgit v1.2.3 From 824877111cd7f2b4fd2fe6947c5c5cbbb3ac5bd8 Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Sat, 27 Dec 2008 18:32:28 +0530 Subject: x86, pci: move arch/x86/pci/pci.h to arch/x86/include/asm/pci_x86.h Impact: cleanup Now that arch/x86/pci/pci.h is used in a number of other places as well, move the lowlevel x86 pci definitions into the architecture include files. (not to be confused with the existing arch/x86/include/asm/pci.h file, which provides public details about x86 PCI) Tested on: X86_32_UP, X86_32_SMP and X86_64_SMP Signed-off-by: Jaswinder Singh Rajput Acked-by: Jesse Barnes Signed-off-by: Ingo Molnar --- arch/x86/include/asm/pci_x86.h | 165 +++++++++++++++++++++++++++++++++++++ arch/x86/kernel/mmconf-fam10h_64.c | 3 +- arch/x86/kernel/reboot.c | 3 +- arch/x86/pci/acpi.c | 2 +- arch/x86/pci/amd_bus.c | 2 +- arch/x86/pci/common.c | 3 +- arch/x86/pci/direct.c | 2 +- arch/x86/pci/early.c | 2 +- arch/x86/pci/fixup.c | 3 +- arch/x86/pci/i386.c | 2 +- arch/x86/pci/init.c | 2 +- arch/x86/pci/irq.c | 3 +- arch/x86/pci/legacy.c | 2 +- arch/x86/pci/mmconfig-shared.c | 3 +- arch/x86/pci/mmconfig_32.c | 2 +- arch/x86/pci/mmconfig_64.c | 3 +- arch/x86/pci/numaq_32.c | 2 +- arch/x86/pci/olpc.c | 2 +- arch/x86/pci/pcbios.c | 5 +- arch/x86/pci/pci.h | 162 ------------------------------------ arch/x86/pci/visws.c | 3 +- drivers/pci/hotplug/cpqphp_core.c | 2 +- drivers/pci/hotplug/cpqphp_pci.c | 2 +- drivers/pci/hotplug/ibmphp_core.c | 2 +- 24 files changed, 188 insertions(+), 194 deletions(-) create mode 100644 arch/x86/include/asm/pci_x86.h delete mode 100644 arch/x86/pci/pci.h (limited to 'arch/x86') diff --git a/arch/x86/include/asm/pci_x86.h b/arch/x86/include/asm/pci_x86.h new file mode 100644 index 000000000000..e60fd3e14bdf --- /dev/null +++ b/arch/x86/include/asm/pci_x86.h @@ -0,0 +1,165 @@ +/* + * Low-Level PCI Access for i386 machines. + * + * (c) 1999 Martin Mares + */ + +#undef DEBUG + +#ifdef DEBUG +#define DBG(x...) printk(x) +#else +#define DBG(x...) +#endif + +#define PCI_PROBE_BIOS 0x0001 +#define PCI_PROBE_CONF1 0x0002 +#define PCI_PROBE_CONF2 0x0004 +#define PCI_PROBE_MMCONF 0x0008 +#define PCI_PROBE_MASK 0x000f +#define PCI_PROBE_NOEARLY 0x0010 + +#define PCI_NO_CHECKS 0x0400 +#define PCI_USE_PIRQ_MASK 0x0800 +#define PCI_ASSIGN_ROMS 0x1000 +#define PCI_BIOS_IRQ_SCAN 0x2000 +#define PCI_ASSIGN_ALL_BUSSES 0x4000 +#define PCI_CAN_SKIP_ISA_ALIGN 0x8000 +#define PCI_USE__CRS 0x10000 +#define PCI_CHECK_ENABLE_AMD_MMCONF 0x20000 +#define PCI_HAS_IO_ECS 0x40000 +#define PCI_NOASSIGN_ROMS 0x80000 + +extern unsigned int pci_probe; +extern unsigned long pirq_table_addr; + +enum pci_bf_sort_state { + pci_bf_sort_default, + pci_force_nobf, + pci_force_bf, + pci_dmi_bf, +}; + +/* pci-i386.c */ + +extern unsigned int pcibios_max_latency; + +void pcibios_resource_survey(void); + +/* pci-pc.c */ + +extern int pcibios_last_bus; +extern struct pci_bus *pci_root_bus; +extern struct pci_ops pci_root_ops; + +/* pci-irq.c */ + +struct irq_info { + u8 bus, devfn; /* Bus, device and function */ + struct { + u8 link; /* IRQ line ID, chipset dependent, + 0 = not routed */ + u16 bitmap; /* Available IRQs */ + } __attribute__((packed)) irq[4]; + u8 slot; /* Slot number, 0=onboard */ + u8 rfu; +} __attribute__((packed)); + +struct irq_routing_table { + u32 signature; /* PIRQ_SIGNATURE should be here */ + u16 version; /* PIRQ_VERSION */ + u16 size; /* Table size in bytes */ + u8 rtr_bus, rtr_devfn; /* Where the interrupt router lies */ + u16 exclusive_irqs; /* IRQs devoted exclusively to + PCI usage */ + u16 rtr_vendor, rtr_device; /* Vendor and device ID of + interrupt router */ + u32 miniport_data; /* Crap */ + u8 rfu[11]; + u8 checksum; /* Modulo 256 checksum must give 0 */ + struct irq_info slots[0]; +} __attribute__((packed)); + +extern unsigned int pcibios_irq_mask; + +extern int pcibios_scanned; +extern spinlock_t pci_config_lock; + +extern int (*pcibios_enable_irq)(struct pci_dev *dev); +extern void (*pcibios_disable_irq)(struct pci_dev *dev); + +struct pci_raw_ops { + int (*read)(unsigned int domain, unsigned int bus, unsigned int devfn, + int reg, int len, u32 *val); + int (*write)(unsigned int domain, unsigned int bus, unsigned int devfn, + int reg, int len, u32 val); +}; + +extern struct pci_raw_ops *raw_pci_ops; +extern struct pci_raw_ops *raw_pci_ext_ops; + +extern struct pci_raw_ops pci_direct_conf1; +extern bool port_cf9_safe; + +/* arch_initcall level */ +extern int pci_direct_probe(void); +extern void pci_direct_init(int type); +extern void pci_pcbios_init(void); +extern int pci_olpc_init(void); +extern void __init dmi_check_pciprobe(void); +extern void __init dmi_check_skip_isa_align(void); + +/* some common used subsys_initcalls */ +extern int __init pci_acpi_init(void); +extern int __init pcibios_irq_init(void); +extern int __init pci_visws_init(void); +extern int __init pci_numaq_init(void); +extern int __init pcibios_init(void); + +/* pci-mmconfig.c */ + +extern int __init pci_mmcfg_arch_init(void); +extern void __init pci_mmcfg_arch_free(void); + +/* + * AMD Fam10h CPUs are buggy, and cannot access MMIO config space + * on their northbrige except through the * %eax register. As such, you MUST + * NOT use normal IOMEM accesses, you need to only use the magic mmio-config + * accessor functions. + * In fact just use pci_config_*, nothing else please. + */ +static inline unsigned char mmio_config_readb(void __iomem *pos) +{ + u8 val; + asm volatile("movb (%1),%%al" : "=a" (val) : "r" (pos)); + return val; +} + +static inline unsigned short mmio_config_readw(void __iomem *pos) +{ + u16 val; + asm volatile("movw (%1),%%ax" : "=a" (val) : "r" (pos)); + return val; +} + +static inline unsigned int mmio_config_readl(void __iomem *pos) +{ + u32 val; + asm volatile("movl (%1),%%eax" : "=a" (val) : "r" (pos)); + return val; +} + +static inline void mmio_config_writeb(void __iomem *pos, u8 val) +{ + asm volatile("movb %%al,(%1)" : : "a" (val), "r" (pos) : "memory"); +} + +static inline void mmio_config_writew(void __iomem *pos, u16 val) +{ + asm volatile("movw %%ax,(%1)" : : "a" (val), "r" (pos) : "memory"); +} + +static inline void mmio_config_writel(void __iomem *pos, u32 val) +{ + asm volatile("movl %%eax,(%1)" : : "a" (val), "r" (pos) : "memory"); +} diff --git a/arch/x86/kernel/mmconf-fam10h_64.c b/arch/x86/kernel/mmconf-fam10h_64.c index efc2f361fe85..666e43df51f9 100644 --- a/arch/x86/kernel/mmconf-fam10h_64.c +++ b/arch/x86/kernel/mmconf-fam10h_64.c @@ -13,8 +13,7 @@ #include #include #include - -#include "../pci/pci.h" +#include struct pci_hostbridge_probe { u32 bus; diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c index b165eb0884ed..a90913cccfb7 100644 --- a/arch/x86/kernel/reboot.c +++ b/arch/x86/kernel/reboot.c @@ -12,6 +12,7 @@ #include #include #include +#include #ifdef CONFIG_X86_32 # include @@ -22,8 +23,6 @@ #endif #include -#include "../pci/pci.h" - /* * Power off function, if any diff --git a/arch/x86/pci/acpi.c b/arch/x86/pci/acpi.c index 1d88d2b39771..9e5752fe4d15 100644 --- a/arch/x86/pci/acpi.c +++ b/arch/x86/pci/acpi.c @@ -4,7 +4,7 @@ #include #include #include -#include "pci.h" +#include struct pci_root_info { char *name; diff --git a/arch/x86/pci/amd_bus.c b/arch/x86/pci/amd_bus.c index 22e057665e55..9bb09823b362 100644 --- a/arch/x86/pci/amd_bus.c +++ b/arch/x86/pci/amd_bus.c @@ -2,7 +2,7 @@ #include #include #include -#include "pci.h" +#include #ifdef CONFIG_X86_64 #include diff --git a/arch/x86/pci/common.c b/arch/x86/pci/common.c index bb1a01f089e2..62ddb73e09ed 100644 --- a/arch/x86/pci/common.c +++ b/arch/x86/pci/common.c @@ -14,8 +14,7 @@ #include #include #include - -#include "pci.h" +#include unsigned int pci_probe = PCI_PROBE_BIOS | PCI_PROBE_CONF1 | PCI_PROBE_CONF2 | PCI_PROBE_MMCONF; diff --git a/arch/x86/pci/direct.c b/arch/x86/pci/direct.c index 9a5af6c8fbe9..bd13c3e4c6db 100644 --- a/arch/x86/pci/direct.c +++ b/arch/x86/pci/direct.c @@ -5,7 +5,7 @@ #include #include #include -#include "pci.h" +#include /* * Functions for accessing PCI base (first 256 bytes) and extended diff --git a/arch/x86/pci/early.c b/arch/x86/pci/early.c index 86631ccbc25a..f6adf2c6d751 100644 --- a/arch/x86/pci/early.c +++ b/arch/x86/pci/early.c @@ -2,7 +2,7 @@ #include #include #include -#include "pci.h" +#include /* Direct PCI access. This is used for PCI accesses in early boot before the PCI subsystem works. */ diff --git a/arch/x86/pci/fixup.c b/arch/x86/pci/fixup.c index 2051dc96b8e9..7d388d5cf548 100644 --- a/arch/x86/pci/fixup.c +++ b/arch/x86/pci/fixup.c @@ -6,8 +6,7 @@ #include #include #include -#include "pci.h" - +#include static void __devinit pci_fixup_i450nx(struct pci_dev *d) { diff --git a/arch/x86/pci/i386.c b/arch/x86/pci/i386.c index 844df0cbbd3e..e51bf2cda4b0 100644 --- a/arch/x86/pci/i386.c +++ b/arch/x86/pci/i386.c @@ -34,8 +34,8 @@ #include #include +#include -#include "pci.h" static int skip_isa_ioresource_align(struct pci_dev *dev) { diff --git a/arch/x86/pci/init.c b/arch/x86/pci/init.c index d6c950f81858..bec3b048e72b 100644 --- a/arch/x86/pci/init.c +++ b/arch/x86/pci/init.c @@ -1,6 +1,6 @@ #include #include -#include "pci.h" +#include /* arch_initcall has too random ordering, so call the initializers in the right sequence from here. */ diff --git a/arch/x86/pci/irq.c b/arch/x86/pci/irq.c index bf69dbe08bff..373b9afe6d44 100644 --- a/arch/x86/pci/irq.c +++ b/arch/x86/pci/irq.c @@ -16,8 +16,7 @@ #include #include #include - -#include "pci.h" +#include #define PIRQ_SIGNATURE (('$' << 0) + ('P' << 8) + ('I' << 16) + ('R' << 24)) #define PIRQ_VERSION 0x0100 diff --git a/arch/x86/pci/legacy.c b/arch/x86/pci/legacy.c index b722dd481b39..f1065b129e9c 100644 --- a/arch/x86/pci/legacy.c +++ b/arch/x86/pci/legacy.c @@ -3,7 +3,7 @@ */ #include #include -#include "pci.h" +#include /* * Discover remaining PCI buses in case there are peer host bridges. diff --git a/arch/x86/pci/mmconfig-shared.c b/arch/x86/pci/mmconfig-shared.c index 654a2234f8f3..89bf9242c80a 100644 --- a/arch/x86/pci/mmconfig-shared.c +++ b/arch/x86/pci/mmconfig-shared.c @@ -15,8 +15,7 @@ #include #include #include - -#include "pci.h" +#include /* aperture is up to 256MB but BIOS may reserve less */ #define MMCONFIG_APER_MIN (2 * 1024*1024) diff --git a/arch/x86/pci/mmconfig_32.c b/arch/x86/pci/mmconfig_32.c index f3c761dce695..8b2d561046a3 100644 --- a/arch/x86/pci/mmconfig_32.c +++ b/arch/x86/pci/mmconfig_32.c @@ -13,7 +13,7 @@ #include #include #include -#include "pci.h" +#include /* Assume systems with more busses have correct MCFG */ #define mmcfg_virt_addr ((void __iomem *) fix_to_virt(FIX_PCIE_MCFG)) diff --git a/arch/x86/pci/mmconfig_64.c b/arch/x86/pci/mmconfig_64.c index a1994163c99d..30007ffc8e11 100644 --- a/arch/x86/pci/mmconfig_64.c +++ b/arch/x86/pci/mmconfig_64.c @@ -10,8 +10,7 @@ #include #include #include - -#include "pci.h" +#include /* Static virtual mapping of the MMCONFIG aperture */ struct mmcfg_virt { diff --git a/arch/x86/pci/numaq_32.c b/arch/x86/pci/numaq_32.c index 1177845d3186..2089354968a2 100644 --- a/arch/x86/pci/numaq_32.c +++ b/arch/x86/pci/numaq_32.c @@ -7,7 +7,7 @@ #include #include #include -#include "pci.h" +#include #define XQUAD_PORTIO_BASE 0xfe400000 #define XQUAD_PORTIO_QUAD 0x40000 /* 256k per quad. */ diff --git a/arch/x86/pci/olpc.c b/arch/x86/pci/olpc.c index e11e9e803d5f..b889d824f7c6 100644 --- a/arch/x86/pci/olpc.c +++ b/arch/x86/pci/olpc.c @@ -29,7 +29,7 @@ #include #include #include -#include "pci.h" +#include /* * In the tables below, the first two line (8 longwords) are the diff --git a/arch/x86/pci/pcbios.c b/arch/x86/pci/pcbios.c index 37472fc6f729..b82cae970dfd 100644 --- a/arch/x86/pci/pcbios.c +++ b/arch/x86/pci/pcbios.c @@ -6,9 +6,8 @@ #include #include #include -#include "pci.h" -#include "pci-functions.h" - +#include +#include /* BIOS32 signature: "_32_" */ #define BIOS32_SIGNATURE (('_' << 0) + ('3' << 8) + ('2' << 16) + ('_' << 24)) diff --git a/arch/x86/pci/pci.h b/arch/x86/pci/pci.h deleted file mode 100644 index 1959018aac02..000000000000 --- a/arch/x86/pci/pci.h +++ /dev/null @@ -1,162 +0,0 @@ -/* - * Low-Level PCI Access for i386 machines. - * - * (c) 1999 Martin Mares - */ - -#undef DEBUG - -#ifdef DEBUG -#define DBG(x...) printk(x) -#else -#define DBG(x...) -#endif - -#define PCI_PROBE_BIOS 0x0001 -#define PCI_PROBE_CONF1 0x0002 -#define PCI_PROBE_CONF2 0x0004 -#define PCI_PROBE_MMCONF 0x0008 -#define PCI_PROBE_MASK 0x000f -#define PCI_PROBE_NOEARLY 0x0010 - -#define PCI_NO_CHECKS 0x0400 -#define PCI_USE_PIRQ_MASK 0x0800 -#define PCI_ASSIGN_ROMS 0x1000 -#define PCI_BIOS_IRQ_SCAN 0x2000 -#define PCI_ASSIGN_ALL_BUSSES 0x4000 -#define PCI_CAN_SKIP_ISA_ALIGN 0x8000 -#define PCI_USE__CRS 0x10000 -#define PCI_CHECK_ENABLE_AMD_MMCONF 0x20000 -#define PCI_HAS_IO_ECS 0x40000 -#define PCI_NOASSIGN_ROMS 0x80000 - -extern unsigned int pci_probe; -extern unsigned long pirq_table_addr; - -enum pci_bf_sort_state { - pci_bf_sort_default, - pci_force_nobf, - pci_force_bf, - pci_dmi_bf, -}; - -/* pci-i386.c */ - -extern unsigned int pcibios_max_latency; - -void pcibios_resource_survey(void); - -/* pci-pc.c */ - -extern int pcibios_last_bus; -extern struct pci_bus *pci_root_bus; -extern struct pci_ops pci_root_ops; - -/* pci-irq.c */ - -struct irq_info { - u8 bus, devfn; /* Bus, device and function */ - struct { - u8 link; /* IRQ line ID, chipset dependent, 0=not routed */ - u16 bitmap; /* Available IRQs */ - } __attribute__((packed)) irq[4]; - u8 slot; /* Slot number, 0=onboard */ - u8 rfu; -} __attribute__((packed)); - -struct irq_routing_table { - u32 signature; /* PIRQ_SIGNATURE should be here */ - u16 version; /* PIRQ_VERSION */ - u16 size; /* Table size in bytes */ - u8 rtr_bus, rtr_devfn; /* Where the interrupt router lies */ - u16 exclusive_irqs; /* IRQs devoted exclusively to PCI usage */ - u16 rtr_vendor, rtr_device; /* Vendor and device ID of interrupt router */ - u32 miniport_data; /* Crap */ - u8 rfu[11]; - u8 checksum; /* Modulo 256 checksum must give zero */ - struct irq_info slots[0]; -} __attribute__((packed)); - -extern unsigned int pcibios_irq_mask; - -extern int pcibios_scanned; -extern spinlock_t pci_config_lock; - -extern int (*pcibios_enable_irq)(struct pci_dev *dev); -extern void (*pcibios_disable_irq)(struct pci_dev *dev); - -struct pci_raw_ops { - int (*read)(unsigned int domain, unsigned int bus, unsigned int devfn, - int reg, int len, u32 *val); - int (*write)(unsigned int domain, unsigned int bus, unsigned int devfn, - int reg, int len, u32 val); -}; - -extern struct pci_raw_ops *raw_pci_ops; -extern struct pci_raw_ops *raw_pci_ext_ops; - -extern struct pci_raw_ops pci_direct_conf1; -extern bool port_cf9_safe; - -/* arch_initcall level */ -extern int pci_direct_probe(void); -extern void pci_direct_init(int type); -extern void pci_pcbios_init(void); -extern int pci_olpc_init(void); -extern void __init dmi_check_pciprobe(void); -extern void __init dmi_check_skip_isa_align(void); - -/* some common used subsys_initcalls */ -extern int __init pci_acpi_init(void); -extern int __init pcibios_irq_init(void); -extern int __init pci_visws_init(void); -extern int __init pci_numaq_init(void); -extern int __init pcibios_init(void); - -/* pci-mmconfig.c */ - -extern int __init pci_mmcfg_arch_init(void); -extern void __init pci_mmcfg_arch_free(void); - -/* - * AMD Fam10h CPUs are buggy, and cannot access MMIO config space - * on their northbrige except through the * %eax register. As such, you MUST - * NOT use normal IOMEM accesses, you need to only use the magic mmio-config - * accessor functions. - * In fact just use pci_config_*, nothing else please. - */ -static inline unsigned char mmio_config_readb(void __iomem *pos) -{ - u8 val; - asm volatile("movb (%1),%%al" : "=a" (val) : "r" (pos)); - return val; -} - -static inline unsigned short mmio_config_readw(void __iomem *pos) -{ - u16 val; - asm volatile("movw (%1),%%ax" : "=a" (val) : "r" (pos)); - return val; -} - -static inline unsigned int mmio_config_readl(void __iomem *pos) -{ - u32 val; - asm volatile("movl (%1),%%eax" : "=a" (val) : "r" (pos)); - return val; -} - -static inline void mmio_config_writeb(void __iomem *pos, u8 val) -{ - asm volatile("movb %%al,(%1)" :: "a" (val), "r" (pos) : "memory"); -} - -static inline void mmio_config_writew(void __iomem *pos, u16 val) -{ - asm volatile("movw %%ax,(%1)" :: "a" (val), "r" (pos) : "memory"); -} - -static inline void mmio_config_writel(void __iomem *pos, u32 val) -{ - asm volatile("movl %%eax,(%1)" :: "a" (val), "r" (pos) : "memory"); -} diff --git a/arch/x86/pci/visws.c b/arch/x86/pci/visws.c index 42f4cb19faca..16d0c0eb0d19 100644 --- a/arch/x86/pci/visws.c +++ b/arch/x86/pci/visws.c @@ -9,11 +9,10 @@ #include #include +#include #include #include -#include "pci.h" - static int pci_visws_enable_irq(struct pci_dev *dev) { return 0; } static void pci_visws_disable_irq(struct pci_dev *dev) { } diff --git a/drivers/pci/hotplug/cpqphp_core.c b/drivers/pci/hotplug/cpqphp_core.c index 8514c3a1746a..c2e1bcbb28a7 100644 --- a/drivers/pci/hotplug/cpqphp_core.c +++ b/drivers/pci/hotplug/cpqphp_core.c @@ -45,7 +45,7 @@ #include "cpqphp.h" #include "cpqphp_nvram.h" -#include "../../../arch/x86/pci/pci.h" /* horrible hack showing how processor dependent we are... */ +#include /* Global variables */ diff --git a/drivers/pci/hotplug/cpqphp_pci.c b/drivers/pci/hotplug/cpqphp_pci.c index 09021930589f..df146be9d2e9 100644 --- a/drivers/pci/hotplug/cpqphp_pci.c +++ b/drivers/pci/hotplug/cpqphp_pci.c @@ -37,7 +37,7 @@ #include "../pci.h" #include "cpqphp.h" #include "cpqphp_nvram.h" -#include "../../../arch/x86/pci/pci.h" /* horrible hack showing how processor dependent we are... */ +#include u8 cpqhp_nic_irq; diff --git a/drivers/pci/hotplug/ibmphp_core.c b/drivers/pci/hotplug/ibmphp_core.c index 633e743442ac..dd18f857dfb0 100644 --- a/drivers/pci/hotplug/ibmphp_core.c +++ b/drivers/pci/hotplug/ibmphp_core.c @@ -35,7 +35,7 @@ #include #include #include "../pci.h" -#include "../../../arch/x86/pci/pci.h" /* for struct irq_routing_table */ +#include /* for struct irq_routing_table */ #include "ibmphp.h" #define attn_on(sl) ibmphp_hpc_writeslot (sl, HPC_SLOT_ATTNON) -- cgit v1.2.3 From 58a24566449892dda409b9ad92c2e56c76c5670c Mon Sep 17 00:00:00 2001 From: Matias Zabaljauregui Date: Mon, 29 Sep 2008 01:40:07 -0300 Subject: lguest: move the initial guest page table creation code to the host This patch moves the initial guest page table creation code to the host, so the launcher keeps working with PAE enabled configs. Signed-off-by: Matias Zabaljauregui Signed-off-by: Rusty Russell --- Documentation/lguest/lguest.c | 60 ++++------------------------------ arch/x86/lguest/i386_head.S | 15 --------- drivers/lguest/lg.h | 2 +- drivers/lguest/lguest_user.c | 13 +++----- drivers/lguest/page_tables.c | 72 +++++++++++++++++++++++++++++++++++++++-- include/linux/lguest_launcher.h | 2 +- 6 files changed, 83 insertions(+), 81 deletions(-) (limited to 'arch/x86') diff --git a/Documentation/lguest/lguest.c b/Documentation/lguest/lguest.c index aa2574ca94c7..f2dbbf3bdeab 100644 --- a/Documentation/lguest/lguest.c +++ b/Documentation/lguest/lguest.c @@ -481,51 +481,6 @@ static unsigned long load_initrd(const char *name, unsigned long mem) /* We return the initrd size. */ return len; } - -/* Once we know how much memory we have we can construct simple linear page - * tables which set virtual == physical which will get the Guest far enough - * into the boot to create its own. - * - * We lay them out of the way, just below the initrd (which is why we need to - * know its size here). */ -static unsigned long setup_pagetables(unsigned long mem, - unsigned long initrd_size) -{ - unsigned long *pgdir, *linear; - unsigned int mapped_pages, i, linear_pages; - unsigned int ptes_per_page = getpagesize()/sizeof(void *); - - mapped_pages = mem/getpagesize(); - - /* Each PTE page can map ptes_per_page pages: how many do we need? */ - linear_pages = (mapped_pages + ptes_per_page-1)/ptes_per_page; - - /* We put the toplevel page directory page at the top of memory. */ - pgdir = from_guest_phys(mem) - initrd_size - getpagesize(); - - /* Now we use the next linear_pages pages as pte pages */ - linear = (void *)pgdir - linear_pages*getpagesize(); - - /* Linear mapping is easy: put every page's address into the mapping in - * order. PAGE_PRESENT contains the flags Present, Writable and - * Executable. */ - for (i = 0; i < mapped_pages; i++) - linear[i] = ((i * getpagesize()) | PAGE_PRESENT); - - /* The top level points to the linear page table pages above. */ - for (i = 0; i < mapped_pages; i += ptes_per_page) { - pgdir[i/ptes_per_page] - = ((to_guest_phys(linear) + i*sizeof(void *)) - | PAGE_PRESENT); - } - - verbose("Linear mapping of %u pages in %u pte pages at %#lx\n", - mapped_pages, linear_pages, to_guest_phys(linear)); - - /* We return the top level (guest-physical) address: the kernel needs - * to know where it is. */ - return to_guest_phys(pgdir); -} /*:*/ /* Simple routine to roll all the commandline arguments together with spaces @@ -548,13 +503,13 @@ static void concat(char *dst, char *args[]) /*L:185 This is where we actually tell the kernel to initialize the Guest. We * saw the arguments it expects when we looked at initialize() in lguest_user.c: - * the base of Guest "physical" memory, the top physical page to allow, the - * top level pagetable and the entry point for the Guest. */ -static int tell_kernel(unsigned long pgdir, unsigned long start) + * the base of Guest "physical" memory, the top physical page to allow and the + * entry point for the Guest. */ +static int tell_kernel(unsigned long start) { unsigned long args[] = { LHREQ_INITIALIZE, (unsigned long)guest_base, - guest_limit / getpagesize(), pgdir, start }; + guest_limit / getpagesize(), start }; int fd; verbose("Guest: %p - %p (%#lx)\n", @@ -1941,7 +1896,7 @@ int main(int argc, char *argv[]) { /* Memory, top-level pagetable, code startpoint and size of the * (optional) initrd. */ - unsigned long mem = 0, pgdir, start, initrd_size = 0; + unsigned long mem = 0, start, initrd_size = 0; /* Two temporaries and the /dev/lguest file descriptor. */ int i, c, lguest_fd; /* The boot information for the Guest. */ @@ -2040,9 +1995,6 @@ int main(int argc, char *argv[]) boot->hdr.type_of_loader = 0xFF; } - /* Set up the initial linear pagetables, starting below the initrd. */ - pgdir = setup_pagetables(mem, initrd_size); - /* The Linux boot header contains an "E820" memory map: ours is a * simple, single region. */ boot->e820_entries = 1; @@ -2064,7 +2016,7 @@ int main(int argc, char *argv[]) /* We tell the kernel to initialize the Guest: this returns the open * /dev/lguest file descriptor. */ - lguest_fd = tell_kernel(pgdir, start); + lguest_fd = tell_kernel(start); /* We clone off a thread, which wakes the Launcher whenever one of the * input file descriptors needs attention. We call this the Waker, and diff --git a/arch/x86/lguest/i386_head.S b/arch/x86/lguest/i386_head.S index 5c7cef34c9e7..10b9bd35a8ff 100644 --- a/arch/x86/lguest/i386_head.S +++ b/arch/x86/lguest/i386_head.S @@ -30,21 +30,6 @@ ENTRY(lguest_entry) movl $lguest_data - __PAGE_OFFSET, %edx int $LGUEST_TRAP_ENTRY - /* The Host put the toplevel pagetable in lguest_data.pgdir. The movsl - * instruction uses %esi implicitly as the source for the copy we're - * about to do. */ - movl lguest_data - __PAGE_OFFSET + LGUEST_DATA_pgdir, %esi - - /* Copy first 32 entries of page directory to __PAGE_OFFSET entries. - * This means the first 128M of kernel memory will be mapped at - * PAGE_OFFSET where the kernel expects to run. This will get it far - * enough through boot to switch to its own pagetables. */ - movl $32, %ecx - movl %esi, %edi - addl $((__PAGE_OFFSET >> 22) * 4), %edi - rep - movsl - /* Set up the initial stack so we can run C code. */ movl $(init_thread_union+THREAD_SIZE),%esp diff --git a/drivers/lguest/lg.h b/drivers/lguest/lg.h index 5faefeaf6790..f2c641e0bdde 100644 --- a/drivers/lguest/lg.h +++ b/drivers/lguest/lg.h @@ -164,7 +164,7 @@ void copy_gdt(const struct lg_cpu *cpu, struct desc_struct *gdt); void copy_gdt_tls(const struct lg_cpu *cpu, struct desc_struct *gdt); /* page_tables.c: */ -int init_guest_pagetable(struct lguest *lg, unsigned long pgtable); +int init_guest_pagetable(struct lguest *lg); void free_guest_pagetable(struct lguest *lg); void guest_new_pagetable(struct lg_cpu *cpu, unsigned long pgtable); void guest_set_pmd(struct lguest *lg, unsigned long gpgdir, u32 i); diff --git a/drivers/lguest/lguest_user.c b/drivers/lguest/lguest_user.c index e73a000473cc..34bc017b8b3c 100644 --- a/drivers/lguest/lguest_user.c +++ b/drivers/lguest/lguest_user.c @@ -146,7 +146,7 @@ static int lg_cpu_start(struct lg_cpu *cpu, unsigned id, unsigned long start_ip) return 0; } -/*L:020 The initialization write supplies 4 pointer sized (32 or 64 bit) +/*L:020 The initialization write supplies 3 pointer sized (32 or 64 bit) * values (in addition to the LHREQ_INITIALIZE value). These are: * * base: The start of the Guest-physical memory inside the Launcher memory. @@ -155,9 +155,6 @@ static int lg_cpu_start(struct lg_cpu *cpu, unsigned id, unsigned long start_ip) * allowed to access. The Guest memory lives inside the Launcher, so it sets * this to ensure the Guest can only reach its own memory. * - * pgdir: The (Guest-physical) address of the top of the initial Guest - * pagetables (which are set up by the Launcher). - * * start: The first instruction to execute ("eip" in x86-speak). */ static int initialize(struct file *file, const unsigned long __user *input) @@ -166,7 +163,7 @@ static int initialize(struct file *file, const unsigned long __user *input) * Guest. */ struct lguest *lg; int err; - unsigned long args[4]; + unsigned long args[3]; /* We grab the Big Lguest lock, which protects against multiple * simultaneous initializations. */ @@ -192,14 +189,14 @@ static int initialize(struct file *file, const unsigned long __user *input) lg->mem_base = (void __user *)args[0]; lg->pfn_limit = args[1]; - /* This is the first cpu (cpu 0) and it will start booting at args[3] */ - err = lg_cpu_start(&lg->cpus[0], 0, args[3]); + /* This is the first cpu (cpu 0) and it will start booting at args[2] */ + err = lg_cpu_start(&lg->cpus[0], 0, args[2]); if (err) goto release_guest; /* Initialize the Guest's shadow page tables, using the toplevel * address the Launcher gave us. This allocates memory, so can fail. */ - err = init_guest_pagetable(lg, args[2]); + err = init_guest_pagetable(lg); if (err) goto free_regs; diff --git a/drivers/lguest/page_tables.c b/drivers/lguest/page_tables.c index 81d0c6053447..576a8318221c 100644 --- a/drivers/lguest/page_tables.c +++ b/drivers/lguest/page_tables.c @@ -14,6 +14,7 @@ #include #include #include +#include #include "lg.h" /*M:008 We hold reference to pages, which prevents them from being swapped. @@ -581,15 +582,82 @@ void guest_set_pmd(struct lguest *lg, unsigned long gpgdir, u32 idx) release_pgd(lg, lg->pgdirs[pgdir].pgdir + idx); } +/* Once we know how much memory we have we can construct simple identity + * (which set virtual == physical) and linear mappings + * which will get the Guest far enough into the boot to create its own. + * + * We lay them out of the way, just below the initrd (which is why we need to + * know its size here). */ +static unsigned long setup_pagetables(struct lguest *lg, + unsigned long mem, + unsigned long initrd_size) +{ + pgd_t __user *pgdir; + pte_t __user *linear; + unsigned int mapped_pages, i, linear_pages, phys_linear; + unsigned long mem_base = (unsigned long)lg->mem_base; + + /* We have mapped_pages frames to map, so we need + * linear_pages page tables to map them. */ + mapped_pages = mem / PAGE_SIZE; + linear_pages = (mapped_pages + PTRS_PER_PTE - 1) / PTRS_PER_PTE; + + /* We put the toplevel page directory page at the top of memory. */ + pgdir = (pgd_t *)(mem + mem_base - initrd_size - PAGE_SIZE); + + /* Now we use the next linear_pages pages as pte pages */ + linear = (void *)pgdir - linear_pages * PAGE_SIZE; + + /* Linear mapping is easy: put every page's address into the + * mapping in order. */ + for (i = 0; i < mapped_pages; i++) { + pte_t pte; + pte = pfn_pte(i, __pgprot(_PAGE_PRESENT|_PAGE_RW|_PAGE_USER)); + if (copy_to_user(&linear[i], &pte, sizeof(pte)) != 0) + return -EFAULT; + } + + /* The top level points to the linear page table pages above. + * We setup the identity and linear mappings here. */ + phys_linear = (unsigned long)linear - mem_base; + for (i = 0; i < mapped_pages; i += PTRS_PER_PTE) { + pgd_t pgd; + pgd = __pgd((phys_linear + i * sizeof(pte_t)) | + (_PAGE_PRESENT | _PAGE_RW | _PAGE_USER)); + + if (copy_to_user(&pgdir[i / PTRS_PER_PTE], &pgd, sizeof(pgd)) + || copy_to_user(&pgdir[pgd_index(PAGE_OFFSET) + + i / PTRS_PER_PTE], + &pgd, sizeof(pgd))) + return -EFAULT; + } + + /* We return the top level (guest-physical) address: remember where + * this is. */ + return (unsigned long)pgdir - mem_base; +} + /*H:500 (vii) Setting up the page tables initially. * * When a Guest is first created, the Launcher tells us where the toplevel of * its first page table is. We set some things up here: */ -int init_guest_pagetable(struct lguest *lg, unsigned long pgtable) +int init_guest_pagetable(struct lguest *lg) { + u64 mem; + u32 initrd_size; + struct boot_params __user *boot = (struct boot_params *)lg->mem_base; + + /* Get the Guest memory size and the ramdisk size from the boot header + * located at lg->mem_base (Guest address 0). */ + if (copy_from_user(&mem, &boot->e820_map[0].size, sizeof(mem)) + || get_user(initrd_size, &boot->hdr.ramdisk_size)) + return -EFAULT; + /* We start on the first shadow page table, and give it a blank PGD * page. */ - lg->pgdirs[0].gpgdir = pgtable; + lg->pgdirs[0].gpgdir = setup_pagetables(lg, mem, initrd_size); + if (IS_ERR_VALUE(lg->pgdirs[0].gpgdir)) + return lg->pgdirs[0].gpgdir; lg->pgdirs[0].pgdir = (pgd_t *)get_zeroed_page(GFP_KERNEL); if (!lg->pgdirs[0].pgdir) return -ENOMEM; diff --git a/include/linux/lguest_launcher.h b/include/linux/lguest_launcher.h index bd0eba760522..a53407a4165c 100644 --- a/include/linux/lguest_launcher.h +++ b/include/linux/lguest_launcher.h @@ -54,7 +54,7 @@ struct lguest_vqconfig { /* Write command first word is a request. */ enum lguest_req { - LHREQ_INITIALIZE, /* + base, pfnlimit, pgdir, start */ + LHREQ_INITIALIZE, /* + base, pfnlimit, start */ LHREQ_GETDMA, /* No longer used */ LHREQ_IRQ, /* + irq */ LHREQ_BREAK, /* + on/off flag (on blocks until someone does off) */ -- cgit v1.2.3 From 412a1be265b894a45cebbfc2b57eb7a593bf34b2 Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Mon, 29 Dec 2008 21:44:12 +0530 Subject: x86: amd_iommu_init.c: iommu_enable and iommu_enable_event_logging should be static Impact: cleanup, reduce kernel size a bit, avoid sparse warning Fixes sparse warning: arch/x86/kernel/amd_iommu_init.c:246:13: warning: symbol 'iommu_enable' was not declared. Should it be static? arch/x86/kernel/amd_iommu_init.c:259:13: warning: symbol 'iommu_enable_event_logging' was not declared. Should it be static? Signed-off-by: Jaswinder Singh Rajput Signed-off-by: H. Peter Anvin --- arch/x86/kernel/amd_iommu_init.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu_init.c b/arch/x86/kernel/amd_iommu_init.c index c625800c55ca..fb85e8d466cc 100644 --- a/arch/x86/kernel/amd_iommu_init.c +++ b/arch/x86/kernel/amd_iommu_init.c @@ -243,7 +243,7 @@ static void __init iommu_feature_disable(struct amd_iommu *iommu, u8 bit) } /* Function to enable the hardware */ -void __init iommu_enable(struct amd_iommu *iommu) +static void __init iommu_enable(struct amd_iommu *iommu) { printk(KERN_INFO "AMD IOMMU: Enabling IOMMU " "at %02x:%02x.%x cap 0x%hx\n", @@ -256,7 +256,7 @@ void __init iommu_enable(struct amd_iommu *iommu) } /* Function to enable IOMMU event logging and event interrupts */ -void __init iommu_enable_event_logging(struct amd_iommu *iommu) +static void __init iommu_enable_event_logging(struct amd_iommu *iommu) { iommu_feature_enable(iommu, CONTROL_EVT_LOG_EN); iommu_feature_enable(iommu, CONTROL_EVT_INT_EN); -- cgit v1.2.3 From 557f687c87ddb8adb094b2dad4e1c83c7717982d Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Mon, 29 Dec 2008 21:45:22 +0530 Subject: x86: amd_iommu.c: prealloc_protection_domains should be static Impact: cleanup, reduce kernel size a bit, avoid sparse warning Fixes sparse warning: arch/x86/kernel/amd_iommu.c:1299:6: warning: symbol 'prealloc_protection_domains' was not declared. Should it be static? Signed-off-by: Jaswinder Singh Rajput Signed-off-by: H. Peter Anvin --- arch/x86/kernel/amd_iommu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 2e2da717b350..658e29e0f49b 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -1296,7 +1296,7 @@ static int amd_iommu_dma_supported(struct device *dev, u64 mask) * we don't need to preallocate the protection domains anymore. * For now we have to. */ -void prealloc_protection_domains(void) +static void prealloc_protection_domains(void) { struct pci_dev *dev = NULL; struct dma_ops_domain *dma_dom; -- cgit v1.2.3 From 4d08d97f5262dab4482af5bc91b30af4ca02269e Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Mon, 29 Dec 2008 22:11:40 +0530 Subject: x86: genx2apic_phys.c: x2apic_send_IPI_self and init_x2apic_ldr should be static Impact: cleanup, reduce kernel size a bit, avoid sparse warnings Fixes sparse warnings: arch/x86/kernel/genx2apic_phys.c:164:6: warning: symbol 'x2apic_send_IPI_self' was not declared. Should it be static? arch/x86/kernel/genx2apic_phys.c:169:6: warning: symbol 'init_x2apic_ldr' was not declared. Should it be static? Signed-off-by: Jaswinder Singh Rajput Signed-off-by: H. Peter Anvin --- arch/x86/kernel/genx2apic_phys.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/genx2apic_phys.c b/arch/x86/kernel/genx2apic_phys.c index d042211768b7..a177c7880ab5 100644 --- a/arch/x86/kernel/genx2apic_phys.c +++ b/arch/x86/kernel/genx2apic_phys.c @@ -123,12 +123,12 @@ static unsigned int phys_pkg_id(int index_msb) return current_cpu_data.initial_apicid >> index_msb; } -void x2apic_send_IPI_self(int vector) +static void x2apic_send_IPI_self(int vector) { apic_write(APIC_SELF_IPI, vector); } -void init_x2apic_ldr(void) +static void init_x2apic_ldr(void) { return; } -- cgit v1.2.3 From c62e9d56ea90ef94f9708ce3f11860c20fa5e135 Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Mon, 29 Dec 2008 22:12:50 +0530 Subject: x86: bios_uv.c: uv_systab should be static Impact: cleanup, reduce kernel size a bit, avoid sparse warning Fixes sparse warning: arch/x86/kernel/bios_uv.c:28:18: warning: symbol 'uv_systab' was not declared. Should it be static? Signed-off-by: Jaswinder Singh Rajput Signed-off-by: H. Peter Anvin --- arch/x86/kernel/bios_uv.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/bios_uv.c b/arch/x86/kernel/bios_uv.c index 2a0a2a3cac26..f63882728d91 100644 --- a/arch/x86/kernel/bios_uv.c +++ b/arch/x86/kernel/bios_uv.c @@ -25,7 +25,7 @@ #include #include -struct uv_systab uv_systab; +static struct uv_systab uv_systab; s64 uv_bios_call(enum uv_bios_cmd which, u64 a1, u64 a2, u64 a3, u64 a4, u64 a5) { -- cgit v1.2.3 From ec8c842a524888fdcccece337d91798e3e8af880 Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Tue, 30 Dec 2008 22:46:36 +0530 Subject: x86: apic.c: xapic_icr_read and x2apic_icr_read should be static Impact: cleanup, reduce kernel size a bit, avoid sparse warning Fixes sparse warning: arch/x86/kernel/apic.c:270:5: warning: symbol 'x2apic_icr_read' was not declared. Should it be static? Signed-off-by: Jaswinder Singh Rajput Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/apic.h | 1 - arch/x86/kernel/apic.c | 4 ++-- 2 files changed, 2 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/apic.h b/arch/x86/include/asm/apic.h index e644bf6f90dc..ab1d51a8855e 100644 --- a/arch/x86/include/asm/apic.h +++ b/arch/x86/include/asm/apic.h @@ -54,7 +54,6 @@ extern int disable_apic; extern int is_vsmp_box(void); extern void xapic_wait_icr_idle(void); extern u32 safe_xapic_wait_icr_idle(void); -extern u64 xapic_icr_read(void); extern void xapic_icr_write(u32, u32); extern int setup_profiling_timer(unsigned int); diff --git a/arch/x86/kernel/apic.c b/arch/x86/kernel/apic.c index c67722f010bd..66198cbe464d 100644 --- a/arch/x86/kernel/apic.c +++ b/arch/x86/kernel/apic.c @@ -228,7 +228,7 @@ void xapic_icr_write(u32 low, u32 id) apic_write(APIC_ICR, low); } -u64 xapic_icr_read(void) +static u64 xapic_icr_read(void) { u32 icr1, icr2; @@ -268,7 +268,7 @@ void x2apic_icr_write(u32 low, u32 id) wrmsrl(APIC_BASE_MSR + (APIC_ICR >> 4), ((__u64) id) << 32 | low); } -u64 x2apic_icr_read(void) +static u64 x2apic_icr_read(void) { unsigned long val; -- cgit v1.2.3 From fa95826fe0ddbc2a55373134d8d1a21b49d13434 Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Tue, 30 Dec 2008 20:13:49 +0530 Subject: x86: uv_bau.h: fix dubious bitfield Impact: cleanup, avoid sparse warnings declare bitfield as unsigned to avoid dubious bitfield issue CHECK arch/x86/kernel/tlb_64.c arch/x86/include/asm/uv/uv_bau.h:136:22: warning: dubious bitfield without explicit `signed' or `unsigned' arch/x86/include/asm/uv/uv_bau.h:138:25: warning: dubious bitfield without explicit `signed' or `unsigned' arch/x86/include/asm/uv/uv_bau.h:140:15: warning: dubious bitfield without explicit `signed' or `unsigned' arch/x86/include/asm/uv/uv_bau.h:143:14: warning: dubious bitfield without explicit `signed' or `unsigned' arch/x86/include/asm/uv/uv_bau.h:146:14: warning: dubious bitfield without explicit `signed' or `unsigned' arch/x86/include/asm/uv/uv_bau.h:149:18: warning: dubious bitfield without explicit `signed' or `unsigned' arch/x86/include/asm/uv/uv_bau.h:151:18: warning: dubious bitfield without explicit `signed' or `unsigned' arch/x86/include/asm/uv/uv_bau.h:155:14: error: dubious one-bit signed bitfield arch/x86/include/asm/uv/uv_bau.h:159:18: error: dubious one-bit signed bitfield arch/x86/include/asm/uv/uv_bau.h:173:19: error: dubious one-bit signed bitfield arch/x86/include/asm/uv/uv_bau.h:181:16: error: dubious one-bit signed bitfield arch/x86/include/asm/uv/uv_bau.h:185:18: error: dubious one-bit signed bitfield arch/x86/include/asm/uv/uv_bau.h:188:16: error: dubious one-bit signed bitfield CHECK arch/x86/kernel/tlb_uv.c arch/x86/include/asm/uv/uv_bau.h:136:22: warning: dubious bitfield without explicit `signed' or `unsigned' arch/x86/include/asm/uv/uv_bau.h:138:25: warning: dubious bitfield without explicit `signed' or `unsigned' arch/x86/include/asm/uv/uv_bau.h:140:15: warning: dubious bitfield without explicit `signed' or `unsigned' arch/x86/include/asm/uv/uv_bau.h:143:14: warning: dubious bitfield without explicit `signed' or `unsigned' arch/x86/include/asm/uv/uv_bau.h:146:14: warning: dubious bitfield without explicit `signed' or `unsigned' arch/x86/include/asm/uv/uv_bau.h:149:18: warning: dubious bitfield without explicit `signed' or `unsigned' arch/x86/include/asm/uv/uv_bau.h:151:18: warning: dubious bitfield without explicit `signed' or `unsigned' arch/x86/include/asm/uv/uv_bau.h:155:14: error: dubious one-bit signed bitfield arch/x86/include/asm/uv/uv_bau.h:159:18: error: dubious one-bit signed bitfield arch/x86/include/asm/uv/uv_bau.h:173:19: error: dubious one-bit signed bitfield arch/x86/include/asm/uv/uv_bau.h:181:16: error: dubious one-bit signed bitfield arch/x86/include/asm/uv/uv_bau.h:185:18: error: dubious one-bit signed bitfield arch/x86/include/asm/uv/uv_bau.h:188:16: error: dubious one-bit signed bitfield Signed-off-by: Jaswinder Singh Rajput Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/uv/uv_bau.h | 46 ++++++++++++++++++++-------------------- 1 file changed, 23 insertions(+), 23 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/uv/uv_bau.h b/arch/x86/include/asm/uv/uv_bau.h index e2363253bbbf..50423c7b56b2 100644 --- a/arch/x86/include/asm/uv/uv_bau.h +++ b/arch/x86/include/asm/uv/uv_bau.h @@ -133,61 +133,61 @@ struct bau_msg_payload { * see table 4.2.3.0.1 in broacast_assist spec. */ struct bau_msg_header { - int dest_subnodeid:6; /* must be zero */ + unsigned int dest_subnodeid:6; /* must be zero */ /* bits 5:0 */ - int base_dest_nodeid:15; /* nasid>>1 (pnode) of first bit in node_map */ - /* bits 20:6 */ - int command:8; /* message type */ + unsigned int base_dest_nodeid:15; /* nasid>>1 (pnode) of */ + /* bits 20:6 */ /* first bit in node_map */ + unsigned int command:8; /* message type */ /* bits 28:21 */ /* 0x38: SN3net EndPoint Message */ - int rsvd_1:3; /* must be zero */ + unsigned int rsvd_1:3; /* must be zero */ /* bits 31:29 */ /* int will align on 32 bits */ - int rsvd_2:9; /* must be zero */ + unsigned int rsvd_2:9; /* must be zero */ /* bits 40:32 */ /* Suppl_A is 56-41 */ - int payload_2a:8; /* becomes byte 16 of msg */ + unsigned int payload_2a:8;/* becomes byte 16 of msg */ /* bits 48:41 */ /* not currently using */ - int payload_2b:8; /* becomes byte 17 of msg */ + unsigned int payload_2b:8;/* becomes byte 17 of msg */ /* bits 56:49 */ /* not currently using */ /* Address field (96:57) is never used as an address (these are address bits 42:3) */ - int rsvd_3:1; /* must be zero */ + unsigned int rsvd_3:1; /* must be zero */ /* bit 57 */ /* address bits 27:4 are payload */ /* these 24 bits become bytes 12-14 of msg */ - int replied_to:1; /* sent as 0 by the source to byte 12 */ + unsigned int replied_to:1;/* sent as 0 by the source to byte 12 */ /* bit 58 */ - int payload_1a:5; /* not currently used */ + unsigned int payload_1a:5;/* not currently used */ /* bits 63:59 */ - int payload_1b:8; /* not currently used */ + unsigned int payload_1b:8;/* not currently used */ /* bits 71:64 */ - int payload_1c:8; /* not currently used */ + unsigned int payload_1c:8;/* not currently used */ /* bits 79:72 */ - int payload_1d:2; /* not currently used */ + unsigned int payload_1d:2;/* not currently used */ /* bits 81:80 */ - int rsvd_4:7; /* must be zero */ + unsigned int rsvd_4:7; /* must be zero */ /* bits 88:82 */ - int sw_ack_flag:1; /* software acknowledge flag */ + unsigned int sw_ack_flag:1;/* software acknowledge flag */ /* bit 89 */ /* INTD trasactions at destination are to wait for software acknowledge */ - int rsvd_5:6; /* must be zero */ + unsigned int rsvd_5:6; /* must be zero */ /* bits 95:90 */ - int rsvd_6:5; /* must be zero */ + unsigned int rsvd_6:5; /* must be zero */ /* bits 100:96 */ - int int_both:1; /* if 1, interrupt both sockets on the blade */ + unsigned int int_both:1;/* if 1, interrupt both sockets on the blade */ /* bit 101*/ - int fairness:3; /* usually zero */ + unsigned int fairness:3;/* usually zero */ /* bits 104:102 */ - int multilevel:1; /* multi-level multicast format */ + unsigned int multilevel:1; /* multi-level multicast format */ /* bit 105 */ /* 0 for TLB: endpoint multi-unicast messages */ - int chaining:1; /* next descriptor is part of this activation*/ + unsigned int chaining:1;/* next descriptor is part of this activation*/ /* bit 106 */ - int rsvd_7:21; /* must be zero */ + unsigned int rsvd_7:21; /* must be zero */ /* bits 127:107 */ }; -- cgit v1.2.3 From 7820b75643a763abf595c99fab963000ffc8b5f0 Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Tue, 30 Dec 2008 22:05:55 +0530 Subject: x86: xsave.c: restore_user_xstate should be static Impact: cleanup, reduce kernel size a bit, avoid sparse warning Fixes sparse warning: arch/x86/kernel/xsave.c:162:5: warning: symbol 'restore_user_xstate' was not declared. Should it be static? Signed-off-by: Jaswinder Singh Rajput Signed-off-by: H. Peter Anvin --- arch/x86/kernel/xsave.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/xsave.c b/arch/x86/kernel/xsave.c index 15c3e6999182..2b54fe002e94 100644 --- a/arch/x86/kernel/xsave.c +++ b/arch/x86/kernel/xsave.c @@ -159,7 +159,7 @@ int save_i387_xstate(void __user *buf) * Restore the extended state if present. Otherwise, restore the FP/SSE * state. */ -int restore_user_xstate(void __user *buf) +static int restore_user_xstate(void __user *buf) { struct _fpx_sw_bytes fx_sw_user; u64 mask; -- cgit v1.2.3 From 79741dd35713ff4f6fd0eafd59fa94e8a4ba922d Mon Sep 17 00:00:00 2001 From: Martin Schwidefsky Date: Wed, 31 Dec 2008 15:11:38 +0100 Subject: [PATCH] idle cputime accounting The cpu time spent by the idle process actually doing something is currently accounted as idle time. This is plain wrong, the architectures that support VIRT_CPU_ACCOUNTING=y can do better: distinguish between the time spent doing nothing and the time spent by idle doing work. The first is accounted with account_idle_time and the second with account_system_time. The architectures that use the account_xxx_time interface directly and not the account_xxx_ticks interface now need to do the check for the idle process in their arch code. In particular to improve the system vs true idle time accounting the arch code needs to measure the true idle time instead of just testing for the idle process. To improve the tick based accounting as well we would need an architecture primitive that can tell us if the pt_regs of the interrupted context points to the magic instruction that halts the cpu. In addition idle time is no more added to the stime of the idle process. This field now contains the system time of the idle process as it should be. On systems without VIRT_CPU_ACCOUNTING this will always be zero as every tick that occurs while idle is running will be accounted as idle time. This patch contains the necessary common code changes to be able to distinguish idle system time and true idle time. The architectures with support for VIRT_CPU_ACCOUNTING need some changes to exploit this. Signed-off-by: Martin Schwidefsky --- arch/ia64/kernel/time.c | 10 ++++-- arch/powerpc/kernel/process.c | 1 + arch/powerpc/kernel/time.c | 13 +++++-- arch/s390/kernel/vtime.c | 20 ++++++++--- arch/x86/xen/time.c | 10 +++--- include/linux/kernel_stat.h | 7 +++- include/linux/sched.h | 1 - kernel/sched.c | 80 ++++++++++++++++++++++++++++++++++--------- kernel/time/tick-sched.c | 13 ++++--- kernel/timer.c | 13 ------- 10 files changed, 114 insertions(+), 54 deletions(-) (limited to 'arch/x86') diff --git a/arch/ia64/kernel/time.c b/arch/ia64/kernel/time.c index 4ee367817049..f0ebb342409d 100644 --- a/arch/ia64/kernel/time.c +++ b/arch/ia64/kernel/time.c @@ -93,7 +93,10 @@ void ia64_account_on_switch(struct task_struct *prev, struct task_struct *next) now = ia64_get_itc(); delta_stime = cycle_to_cputime(pi->ac_stime + (now - pi->ac_stamp)); - account_system_time(prev, 0, delta_stime, delta_stime); + if (idle_task(smp_processor_id()) != prev) + account_system_time(prev, 0, delta_stime, delta_stime); + else + account_idle_time(delta_stime); if (pi->ac_utime) { delta_utime = cycle_to_cputime(pi->ac_utime); @@ -120,7 +123,10 @@ void account_system_vtime(struct task_struct *tsk) now = ia64_get_itc(); delta_stime = cycle_to_cputime(ti->ac_stime + (now - ti->ac_stamp)); - account_system_time(tsk, 0, delta_stime, delta_stime); + if (irq_count() || idle_task(smp_processor_id()) != tsk) + account_system_time(tsk, 0, delta_stime, delta_stime); + else + account_idle_time(delta_stime); ti->ac_stime = 0; ti->ac_stamp = now; diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c index 51b201ddf9a1..fb7049c054c0 100644 --- a/arch/powerpc/kernel/process.c +++ b/arch/powerpc/kernel/process.c @@ -33,6 +33,7 @@ #include #include #include +#include #include #include diff --git a/arch/powerpc/kernel/time.c b/arch/powerpc/kernel/time.c index 92650ccad2e1..3be355c1cfa7 100644 --- a/arch/powerpc/kernel/time.c +++ b/arch/powerpc/kernel/time.c @@ -256,7 +256,10 @@ void account_system_vtime(struct task_struct *tsk) delta += sys_time; get_paca()->system_time = 0; } - account_system_time(tsk, 0, delta, deltascaled); + if (in_irq() || idle_task(smp_processor_id()) != tsk) + account_system_time(tsk, 0, delta, deltascaled); + else + account_idle_time(delta); per_cpu(cputime_last_delta, smp_processor_id()) = delta; per_cpu(cputime_scaled_last_delta, smp_processor_id()) = deltascaled; local_irq_restore(flags); @@ -335,8 +338,12 @@ void calculate_steal_time(void) tb = mftb(); purr = mfspr(SPRN_PURR); stolen = (tb - pme->tb) - (purr - pme->purr); - if (stolen > 0) - account_steal_time(current, stolen); + if (stolen > 0) { + if (idle_task(smp_processor_id()) != current) + account_steal_time(stolen); + else + account_idle_time(stolen); + } pme->tb = tb; pme->purr = purr; } diff --git a/arch/s390/kernel/vtime.c b/arch/s390/kernel/vtime.c index 07283aea2e56..4a4a34caec55 100644 --- a/arch/s390/kernel/vtime.c +++ b/arch/s390/kernel/vtime.c @@ -55,13 +55,19 @@ void account_process_tick(struct task_struct *tsk, int user_tick) cputime = S390_lowcore.system_timer >> 12; S390_lowcore.system_timer -= cputime << 12; S390_lowcore.steal_clock -= cputime << 12; - account_system_time(tsk, HARDIRQ_OFFSET, cputime, cputime); + if (idle_task(smp_processor_id()) != current) + account_system_time(tsk, HARDIRQ_OFFSET, cputime, cputime); + else + account_idle_time(cputime); cputime = S390_lowcore.steal_clock; if ((__s64) cputime > 0) { cputime >>= 12; S390_lowcore.steal_clock -= cputime << 12; - account_steal_time(tsk, cputime); + if (idle_task(smp_processor_id()) != current) + account_steal_time(cputime); + else + account_idle_time(cputime); } } @@ -87,7 +93,10 @@ void account_vtime(struct task_struct *tsk) cputime = S390_lowcore.system_timer >> 12; S390_lowcore.system_timer -= cputime << 12; S390_lowcore.steal_clock -= cputime << 12; - account_system_time(tsk, 0, cputime, cputime); + if (idle_task(smp_processor_id()) != current) + account_system_time(tsk, 0, cputime, cputime); + else + account_idle_time(cputime); } /* @@ -107,7 +116,10 @@ void account_system_vtime(struct task_struct *tsk) cputime = S390_lowcore.system_timer >> 12; S390_lowcore.system_timer -= cputime << 12; S390_lowcore.steal_clock -= cputime << 12; - account_system_time(tsk, 0, cputime, cputime); + if (in_irq() || idle_task(smp_processor_id()) != current) + account_system_time(tsk, 0, cputime, cputime); + else + account_idle_time(cputime); } EXPORT_SYMBOL_GPL(account_system_vtime); diff --git a/arch/x86/xen/time.c b/arch/x86/xen/time.c index c9f7cda48ed7..732e52dc991a 100644 --- a/arch/x86/xen/time.c +++ b/arch/x86/xen/time.c @@ -132,8 +132,7 @@ static void do_stolen_accounting(void) *snap = state; /* Add the appropriate number of ticks of stolen time, - including any left-overs from last time. Passing NULL to - account_steal_time accounts the time as stolen. */ + including any left-overs from last time. */ stolen = runnable + offline + __get_cpu_var(residual_stolen); if (stolen < 0) @@ -141,11 +140,10 @@ static void do_stolen_accounting(void) ticks = iter_div_u64_rem(stolen, NS_PER_TICK, &stolen); __get_cpu_var(residual_stolen) = stolen; - account_steal_time(NULL, ticks); + account_steal_ticks(ticks); /* Add the appropriate number of ticks of blocked time, - including any left-overs from last time. Passing idle to - account_steal_time accounts the time as idle/wait. */ + including any left-overs from last time. */ blocked += __get_cpu_var(residual_blocked); if (blocked < 0) @@ -153,7 +151,7 @@ static void do_stolen_accounting(void) ticks = iter_div_u64_rem(blocked, NS_PER_TICK, &blocked); __get_cpu_var(residual_blocked) = blocked; - account_steal_time(idle_task(smp_processor_id()), ticks); + account_idle_ticks(ticks); } /* diff --git a/include/linux/kernel_stat.h b/include/linux/kernel_stat.h index c78a459662a6..570d20413119 100644 --- a/include/linux/kernel_stat.h +++ b/include/linux/kernel_stat.h @@ -81,6 +81,11 @@ static inline unsigned int kstat_irqs(unsigned int irq) extern unsigned long long task_delta_exec(struct task_struct *); extern void account_user_time(struct task_struct *, cputime_t, cputime_t); extern void account_system_time(struct task_struct *, int, cputime_t, cputime_t); -extern void account_steal_time(struct task_struct *, cputime_t); +extern void account_steal_time(cputime_t); +extern void account_idle_time(cputime_t); + +extern void account_process_tick(struct task_struct *, int user); +extern void account_steal_ticks(unsigned long ticks); +extern void account_idle_ticks(unsigned long ticks); #endif /* _LINUX_KERNEL_STAT_H */ diff --git a/include/linux/sched.h b/include/linux/sched.h index 8395e715809d..b475d4db8053 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -284,7 +284,6 @@ long io_schedule_timeout(long timeout); extern void cpu_init (void); extern void trap_init(void); -extern void account_process_tick(struct task_struct *task, int user); extern void update_process_times(int user); extern void scheduler_tick(void); diff --git a/kernel/sched.c b/kernel/sched.c index 5b03679ff712..635eaffe1e4c 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4139,7 +4139,6 @@ void account_system_time(struct task_struct *p, int hardirq_offset, cputime_t cputime, cputime_t cputime_scaled) { struct cpu_usage_stat *cpustat = &kstat_this_cpu.cpustat; - struct rq *rq = this_rq(); cputime64_t tmp; if ((p->flags & PF_VCPU) && (irq_count() - hardirq_offset == 0)) { @@ -4158,37 +4157,84 @@ void account_system_time(struct task_struct *p, int hardirq_offset, cpustat->irq = cputime64_add(cpustat->irq, tmp); else if (softirq_count()) cpustat->softirq = cputime64_add(cpustat->softirq, tmp); - else if (p != rq->idle) - cpustat->system = cputime64_add(cpustat->system, tmp); - else if (atomic_read(&rq->nr_iowait) > 0) - cpustat->iowait = cputime64_add(cpustat->iowait, tmp); else - cpustat->idle = cputime64_add(cpustat->idle, tmp); + cpustat->system = cputime64_add(cpustat->system, tmp); + /* Account for system time used */ acct_update_integrals(p); } /* * Account for involuntary wait time. - * @p: the process from which the cpu time has been stolen * @steal: the cpu time spent in involuntary wait */ -void account_steal_time(struct task_struct *p, cputime_t steal) +void account_steal_time(cputime_t cputime) +{ + struct cpu_usage_stat *cpustat = &kstat_this_cpu.cpustat; + cputime64_t cputime64 = cputime_to_cputime64(cputime); + + cpustat->steal = cputime64_add(cpustat->steal, cputime64); +} + +/* + * Account for idle time. + * @cputime: the cpu time spent in idle wait + */ +void account_idle_time(cputime_t cputime) { struct cpu_usage_stat *cpustat = &kstat_this_cpu.cpustat; - cputime64_t tmp = cputime_to_cputime64(steal); + cputime64_t cputime64 = cputime_to_cputime64(cputime); struct rq *rq = this_rq(); - if (p == rq->idle) { - p->stime = cputime_add(p->stime, steal); - if (atomic_read(&rq->nr_iowait) > 0) - cpustat->iowait = cputime64_add(cpustat->iowait, tmp); - else - cpustat->idle = cputime64_add(cpustat->idle, tmp); - } else - cpustat->steal = cputime64_add(cpustat->steal, tmp); + if (atomic_read(&rq->nr_iowait) > 0) + cpustat->iowait = cputime64_add(cpustat->iowait, cputime64); + else + cpustat->idle = cputime64_add(cpustat->idle, cputime64); +} + +#ifndef CONFIG_VIRT_CPU_ACCOUNTING + +/* + * Account a single tick of cpu time. + * @p: the process that the cpu time gets accounted to + * @user_tick: indicates if the tick is a user or a system tick + */ +void account_process_tick(struct task_struct *p, int user_tick) +{ + cputime_t one_jiffy = jiffies_to_cputime(1); + cputime_t one_jiffy_scaled = cputime_to_scaled(one_jiffy); + struct rq *rq = this_rq(); + + if (user_tick) + account_user_time(p, one_jiffy, one_jiffy_scaled); + else if (p != rq->idle) + account_system_time(p, HARDIRQ_OFFSET, one_jiffy, + one_jiffy_scaled); + else + account_idle_time(one_jiffy); +} + +/* + * Account multiple ticks of steal time. + * @p: the process from which the cpu time has been stolen + * @ticks: number of stolen ticks + */ +void account_steal_ticks(unsigned long ticks) +{ + account_steal_time(jiffies_to_cputime(ticks)); +} + +/* + * Account multiple ticks of idle time. + * @ticks: number of stolen ticks + */ +void account_idle_ticks(unsigned long ticks) +{ + account_idle_time(jiffies_to_cputime(ticks)); } +#endif + /* * Use precise platform statistics if available: */ diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index 1f2fce2479fe..611fa4c0baab 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -419,8 +419,9 @@ void tick_nohz_restart_sched_tick(void) { int cpu = smp_processor_id(); struct tick_sched *ts = &per_cpu(tick_cpu_sched, cpu); +#ifndef CONFIG_VIRT_CPU_ACCOUNTING unsigned long ticks; - cputime_t cputime; +#endif ktime_t now; local_irq_disable(); @@ -442,6 +443,7 @@ void tick_nohz_restart_sched_tick(void) tick_do_update_jiffies64(now); cpu_clear(cpu, nohz_cpu_mask); +#ifndef CONFIG_VIRT_CPU_ACCOUNTING /* * We stopped the tick in idle. Update process times would miss the * time we slept as update_process_times does only a 1 tick @@ -451,12 +453,9 @@ void tick_nohz_restart_sched_tick(void) /* * We might be one off. Do not randomly account a huge number of ticks! */ - if (ticks && ticks < LONG_MAX) { - add_preempt_count(HARDIRQ_OFFSET); - cputime = jiffies_to_cputime(ticks); - account_system_time(current, HARDIRQ_OFFSET, cputime, cputime); - sub_preempt_count(HARDIRQ_OFFSET); - } + if (ticks && ticks < LONG_MAX) + account_idle_ticks(ticks); +#endif touch_softlockup_watchdog(); /* diff --git a/kernel/timer.c b/kernel/timer.c index b5efb528aa1d..dee3f641a7a7 100644 --- a/kernel/timer.c +++ b/kernel/timer.c @@ -1018,19 +1018,6 @@ unsigned long get_next_timer_interrupt(unsigned long now) } #endif -#ifndef CONFIG_VIRT_CPU_ACCOUNTING -void account_process_tick(struct task_struct *p, int user_tick) -{ - cputime_t one_jiffy = jiffies_to_cputime(1); - - if (user_tick) - account_user_time(p, one_jiffy, cputime_to_scaled(one_jiffy)); - else - account_system_time(p, HARDIRQ_OFFSET, one_jiffy, - cputime_to_scaled(one_jiffy)); -} -#endif - /* * Called from the timer interrupt handler to charge one tick to the current * process. user_tick is 1 if the tick is user time, 0 for system. -- cgit v1.2.3 From 2786b014ec893c301ea52ef9962e7cc60f89f9b3 Mon Sep 17 00:00:00 2001 From: Guillaume Thouvenin Date: Mon, 22 Sep 2008 16:08:06 +0200 Subject: KVM: x86 emulator: consolidate push reg This patch consolidate the emulation of push reg instruction. Signed-off-by: Guillaume Thouvenin Signed-off-by: Laurent Vivier Signed-off-by: Avi Kivity --- arch/x86/kvm/x86_emulate.c | 8 +------- 1 file changed, 1 insertion(+), 7 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/x86_emulate.c b/arch/x86/kvm/x86_emulate.c index ea051173b0da..a391e213fe61 100644 --- a/arch/x86/kvm/x86_emulate.c +++ b/arch/x86/kvm/x86_emulate.c @@ -1415,13 +1415,7 @@ special_insn: emulate_1op("dec", c->dst, ctxt->eflags); break; case 0x50 ... 0x57: /* push reg */ - c->dst.type = OP_MEM; - c->dst.bytes = c->op_bytes; - c->dst.val = c->src.val; - register_address_increment(c, &c->regs[VCPU_REGS_RSP], - -c->op_bytes); - c->dst.ptr = (void *) register_address( - c, ss_base(ctxt), c->regs[VCPU_REGS_RSP]); + emulate_push(ctxt); break; case 0x58 ... 0x5f: /* pop reg */ pop_instruction: -- cgit v1.2.3 From a26bf12afb608eb5a96192eaee35fc08ffbf85aa Mon Sep 17 00:00:00 2001 From: Jan Kiszka Date: Fri, 26 Sep 2008 09:30:45 +0200 Subject: KVM: VMX: include all IRQ window exits in statistics irq_window_exits only tracks IRQ window exits due to user space requests, nmi_window_exits include all exits. The latter makes more sense, so let's adjust irq_window_exits accounting. Signed-off-by: Jan Kiszka Signed-off-by: Avi Kivity --- arch/x86/kvm/vmx.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index a4018b01e1f9..ac3453799c17 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -2767,6 +2767,7 @@ static int handle_interrupt_window(struct kvm_vcpu *vcpu, vmcs_write32(CPU_BASED_VM_EXEC_CONTROL, cpu_based_vm_exec_control); KVMTRACE_0D(PEND_INTR, vcpu, handler); + ++vcpu->stat.irq_window_exits; /* * If the user space waits to inject interrupts, exit as soon as @@ -2775,7 +2776,6 @@ static int handle_interrupt_window(struct kvm_vcpu *vcpu, if (kvm_run->request_interrupt_window && !vcpu->arch.irq_summary) { kvm_run->exit_reason = KVM_EXIT_IRQ_WINDOW_OPEN; - ++vcpu->stat.irq_window_exits; return 0; } return 1; -- cgit v1.2.3 From e4a41889ece6c95f390a7fa3a94255ab62470968 Mon Sep 17 00:00:00 2001 From: Jan Kiszka Date: Fri, 26 Sep 2008 09:30:46 +0200 Subject: KVM: VMX: Use INTR_TYPE_NMI_INTR instead of magic value Signed-off-by: Jan Kiszka Signed-off-by: Avi Kivity --- arch/x86/kvm/vmx.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index ac3453799c17..81cf12b8d12a 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -2492,7 +2492,7 @@ static int handle_exception(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run) set_bit(irq / BITS_PER_LONG, &vcpu->arch.irq_summary); } - if ((intr_info & INTR_INFO_INTR_TYPE_MASK) == 0x200) /* nmi */ + if ((intr_info & INTR_INFO_INTR_TYPE_MASK) == INTR_TYPE_NMI_INTR) return 1; /* already handled by vmx_vcpu_run() */ if (is_no_device(intr_info)) { @@ -3337,7 +3337,7 @@ static void vmx_vcpu_run(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run) intr_info = vmcs_read32(VM_EXIT_INTR_INFO); /* We need to handle NMIs before interrupts are enabled */ - if ((intr_info & INTR_INFO_INTR_TYPE_MASK) == 0x200 && + if ((intr_info & INTR_INFO_INTR_TYPE_MASK) == INTR_TYPE_NMI_INTR && (intr_info & INTR_INFO_VALID_MASK)) { KVMTRACE_0D(NMI, vcpu, handler); asm("int $2"); -- cgit v1.2.3 From 60637aacfd95c368e1fbc2157275d1b621b5dcdd Mon Sep 17 00:00:00 2001 From: Jan Kiszka Date: Fri, 26 Sep 2008 09:30:47 +0200 Subject: KVM: VMX: Support for NMI task gates Properly set GUEST_INTR_STATE_NMI and reset nmi_injected when a task-switch vmexit happened due to a task gate being used for handling NMIs. Also avoid the false warning about valid vectoring info in kvm_handle_exit. Based on original patch by Gleb Natapov. Signed-off-by: Gleb Natapov Signed-off-by: Jan Kiszka Signed-off-by: Avi Kivity --- arch/x86/kvm/vmx.c | 18 +++++++++++++++--- 1 file changed, 15 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index 81cf12b8d12a..8d0fc68fd4ec 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -2832,6 +2832,7 @@ static int handle_apic_access(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run) static int handle_task_switch(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run) { + struct vcpu_vmx *vmx = to_vmx(vcpu); unsigned long exit_qualification; u16 tss_selector; int reason; @@ -2839,6 +2840,15 @@ static int handle_task_switch(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run) exit_qualification = vmcs_readl(EXIT_QUALIFICATION); reason = (u32)exit_qualification >> 30; + if (reason == TASK_SWITCH_GATE && vmx->vcpu.arch.nmi_injected && + (vmx->idt_vectoring_info & VECTORING_INFO_VALID_MASK) && + (vmx->idt_vectoring_info & VECTORING_INFO_TYPE_MASK) + == INTR_TYPE_NMI_INTR) { + vcpu->arch.nmi_injected = false; + if (cpu_has_virtual_nmis()) + vmcs_set_bits(GUEST_INTERRUPTIBILITY_INFO, + GUEST_INTR_STATE_NMI); + } tss_selector = exit_qualification; return kvm_task_switch(vcpu, tss_selector, reason); @@ -3012,9 +3022,11 @@ static int kvm_handle_exit(struct kvm_run *kvm_run, struct kvm_vcpu *vcpu) if ((vectoring_info & VECTORING_INFO_VALID_MASK) && (exit_reason != EXIT_REASON_EXCEPTION_NMI && - exit_reason != EXIT_REASON_EPT_VIOLATION)) - printk(KERN_WARNING "%s: unexpected, valid vectoring info and " - "exit reason is 0x%x\n", __func__, exit_reason); + exit_reason != EXIT_REASON_EPT_VIOLATION && + exit_reason != EXIT_REASON_TASK_SWITCH)) + printk(KERN_WARNING "%s: unexpected, valid vectoring info " + "(0x%x) and exit reason is 0x%x\n", + __func__, vectoring_info, exit_reason); if (exit_reason < kvm_vmx_max_exit_handlers && kvm_vmx_exit_handlers[exit_reason]) return kvm_vmx_exit_handlers[exit_reason](vcpu, kvm_run); -- cgit v1.2.3 From 448fa4a9c5dbc6941dd19ed09692c588d815bb06 Mon Sep 17 00:00:00 2001 From: Jan Kiszka Date: Fri, 26 Sep 2008 09:30:48 +0200 Subject: KVM: x86: Reset pending/inject NMI state on CPU reset CPU reset invalidates pending or already injected NMIs, therefore reset the related state variables. Based on original patch by Gleb Natapov. Signed-off-by: Gleb Natapov Signed-off-by: Jan Kiszka Signed-off-by: Avi Kivity --- arch/x86/kvm/x86.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index f1f8ff2f1fa2..1a71f6735593 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -3925,6 +3925,9 @@ void kvm_arch_vcpu_destroy(struct kvm_vcpu *vcpu) int kvm_arch_vcpu_reset(struct kvm_vcpu *vcpu) { + vcpu->arch.nmi_pending = false; + vcpu->arch.nmi_injected = false; + return kvm_x86_ops->vcpu_reset(vcpu); } -- cgit v1.2.3 From 33f089ca5a61f7aead26e8e1866dfc961dd88a9e Mon Sep 17 00:00:00 2001 From: Jan Kiszka Date: Fri, 26 Sep 2008 09:30:49 +0200 Subject: KVM: VMX: refactor/fix IRQ and NMI injectability determination There are currently two ways in VMX to check if an IRQ or NMI can be injected: - vmx_{nmi|irq}_enabled and - vcpu.arch.{nmi|interrupt}_window_open. Even worse, one test (at the end of vmx_vcpu_run) uses an inconsistent, likely incorrect logic. This patch consolidates and unifies the tests over {nmi|interrupt}_window_open as cache + vmx_update_window_states for updating the cache content. Signed-off-by: Jan Kiszka Signed-off-by: Avi Kivity --- arch/x86/include/asm/kvm_host.h | 1 + arch/x86/kvm/vmx.c | 46 +++++++++++++++++++---------------------- 2 files changed, 22 insertions(+), 25 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index 8346be87cfa1..bfbbdea869bf 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -327,6 +327,7 @@ struct kvm_vcpu_arch { bool nmi_pending; bool nmi_injected; + bool nmi_window_open; u64 mtrr[0x100]; }; diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index 8d0fc68fd4ec..f0866e1d20ee 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -2362,6 +2362,21 @@ static void vmx_inject_nmi(struct kvm_vcpu *vcpu) INTR_TYPE_NMI_INTR | INTR_INFO_VALID_MASK | NMI_VECTOR); } +static void vmx_update_window_states(struct kvm_vcpu *vcpu) +{ + u32 guest_intr = vmcs_read32(GUEST_INTERRUPTIBILITY_INFO); + + vcpu->arch.nmi_window_open = + !(guest_intr & (GUEST_INTR_STATE_STI | + GUEST_INTR_STATE_MOV_SS | + GUEST_INTR_STATE_NMI)); + + vcpu->arch.interrupt_window_open = + ((vmcs_readl(GUEST_RFLAGS) & X86_EFLAGS_IF) && + !(guest_intr & (GUEST_INTR_STATE_STI | + GUEST_INTR_STATE_MOV_SS))); +} + static void kvm_do_inject_irq(struct kvm_vcpu *vcpu) { int word_index = __ffs(vcpu->arch.irq_summary); @@ -2374,15 +2389,12 @@ static void kvm_do_inject_irq(struct kvm_vcpu *vcpu) kvm_queue_interrupt(vcpu, irq); } - static void do_interrupt_requests(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run) { u32 cpu_based_vm_exec_control; - vcpu->arch.interrupt_window_open = - ((vmcs_readl(GUEST_RFLAGS) & X86_EFLAGS_IF) && - (vmcs_read32(GUEST_INTERRUPTIBILITY_INFO) & 3) == 0); + vmx_update_window_states(vcpu); if (vcpu->arch.interrupt_window_open && vcpu->arch.irq_summary && !vcpu->arch.interrupt.pending) @@ -3075,22 +3087,6 @@ static void enable_nmi_window(struct kvm_vcpu *vcpu) vmcs_write32(CPU_BASED_VM_EXEC_CONTROL, cpu_based_vm_exec_control); } -static int vmx_nmi_enabled(struct kvm_vcpu *vcpu) -{ - u32 guest_intr = vmcs_read32(GUEST_INTERRUPTIBILITY_INFO); - return !(guest_intr & (GUEST_INTR_STATE_NMI | - GUEST_INTR_STATE_MOV_SS | - GUEST_INTR_STATE_STI)); -} - -static int vmx_irq_enabled(struct kvm_vcpu *vcpu) -{ - u32 guest_intr = vmcs_read32(GUEST_INTERRUPTIBILITY_INFO); - return (!(guest_intr & (GUEST_INTR_STATE_MOV_SS | - GUEST_INTR_STATE_STI)) && - (vmcs_readl(GUEST_RFLAGS) & X86_EFLAGS_IF)); -} - static void enable_intr_window(struct kvm_vcpu *vcpu) { if (vcpu->arch.nmi_pending) @@ -3159,11 +3155,13 @@ static void vmx_intr_assist(struct kvm_vcpu *vcpu) { update_tpr_threshold(vcpu); + vmx_update_window_states(vcpu); + if (cpu_has_virtual_nmis()) { if (vcpu->arch.nmi_pending && !vcpu->arch.nmi_injected) { if (vcpu->arch.interrupt.pending) { enable_nmi_window(vcpu); - } else if (vmx_nmi_enabled(vcpu)) { + } else if (vcpu->arch.nmi_window_open) { vcpu->arch.nmi_pending = false; vcpu->arch.nmi_injected = true; } else { @@ -3178,7 +3176,7 @@ static void vmx_intr_assist(struct kvm_vcpu *vcpu) } } if (!vcpu->arch.interrupt.pending && kvm_cpu_has_interrupt(vcpu)) { - if (vmx_irq_enabled(vcpu)) + if (vcpu->arch.interrupt_window_open) kvm_queue_interrupt(vcpu, kvm_cpu_get_interrupt(vcpu)); else enable_irq_window(vcpu); @@ -3339,9 +3337,7 @@ static void vmx_vcpu_run(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run) if (vmx->rmode.irq.pending) fixup_rmode_irq(vmx); - vcpu->arch.interrupt_window_open = - (vmcs_read32(GUEST_INTERRUPTIBILITY_INFO) & - (GUEST_INTR_STATE_STI | GUEST_INTR_STATE_MOV_SS)) == 0; + vmx_update_window_states(vcpu); asm("mov %0, %%ds; mov %0, %%es" : : "r"(__USER_DS)); vmx->launched = 1; -- cgit v1.2.3 From f460ee43e250b675376246b1c4c9fe9b9af4ab16 Mon Sep 17 00:00:00 2001 From: Jan Kiszka Date: Fri, 26 Sep 2008 09:30:50 +0200 Subject: KVM: VMX: refactor IRQ and NMI window enabling do_interrupt_requests and vmx_intr_assist go different way for achieving the same: enabling the nmi/irq window start notification. Unify their code over enable_{irq|nmi}_window, get rid of a redundant call to enable_intr_window instead of direct enable_nmi_window invocation and unroll enable_intr_window for both in-kernel and user space irq injection accordingly. Signed-off-by: Jan Kiszka Signed-off-by: Avi Kivity --- arch/x86/kvm/vmx.c | 78 ++++++++++++++++++++++-------------------------------- 1 file changed, 32 insertions(+), 46 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index f0866e1d20ee..440f56cd4bdd 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -2389,30 +2389,42 @@ static void kvm_do_inject_irq(struct kvm_vcpu *vcpu) kvm_queue_interrupt(vcpu, irq); } -static void do_interrupt_requests(struct kvm_vcpu *vcpu, - struct kvm_run *kvm_run) +static void enable_irq_window(struct kvm_vcpu *vcpu) { u32 cpu_based_vm_exec_control; - vmx_update_window_states(vcpu); + cpu_based_vm_exec_control = vmcs_read32(CPU_BASED_VM_EXEC_CONTROL); + cpu_based_vm_exec_control |= CPU_BASED_VIRTUAL_INTR_PENDING; + vmcs_write32(CPU_BASED_VM_EXEC_CONTROL, cpu_based_vm_exec_control); +} - if (vcpu->arch.interrupt_window_open && - vcpu->arch.irq_summary && !vcpu->arch.interrupt.pending) - kvm_do_inject_irq(vcpu); +static void enable_nmi_window(struct kvm_vcpu *vcpu) +{ + u32 cpu_based_vm_exec_control; - if (vcpu->arch.interrupt_window_open && vcpu->arch.interrupt.pending) - vmx_inject_irq(vcpu, vcpu->arch.interrupt.nr); + if (!cpu_has_virtual_nmis()) + return; cpu_based_vm_exec_control = vmcs_read32(CPU_BASED_VM_EXEC_CONTROL); + cpu_based_vm_exec_control |= CPU_BASED_VIRTUAL_NMI_PENDING; + vmcs_write32(CPU_BASED_VM_EXEC_CONTROL, cpu_based_vm_exec_control); +} + +static void do_interrupt_requests(struct kvm_vcpu *vcpu, + struct kvm_run *kvm_run) +{ + vmx_update_window_states(vcpu); + + if (vcpu->arch.interrupt_window_open) { + if (vcpu->arch.irq_summary && !vcpu->arch.interrupt.pending) + kvm_do_inject_irq(vcpu); + + if (vcpu->arch.interrupt.pending) + vmx_inject_irq(vcpu, vcpu->arch.interrupt.nr); + } if (!vcpu->arch.interrupt_window_open && (vcpu->arch.irq_summary || kvm_run->request_interrupt_window)) - /* - * Interrupts blocked. Wait for unblock. - */ - cpu_based_vm_exec_control |= CPU_BASED_VIRTUAL_INTR_PENDING; - else - cpu_based_vm_exec_control &= ~CPU_BASED_VIRTUAL_INTR_PENDING; - vmcs_write32(CPU_BASED_VM_EXEC_CONTROL, cpu_based_vm_exec_control); + enable_irq_window(vcpu); } static int vmx_set_tss_addr(struct kvm *kvm, unsigned int addr) @@ -3066,35 +3078,6 @@ static void update_tpr_threshold(struct kvm_vcpu *vcpu) vmcs_write32(TPR_THRESHOLD, (max_irr > tpr) ? tpr >> 4 : max_irr >> 4); } -static void enable_irq_window(struct kvm_vcpu *vcpu) -{ - u32 cpu_based_vm_exec_control; - - cpu_based_vm_exec_control = vmcs_read32(CPU_BASED_VM_EXEC_CONTROL); - cpu_based_vm_exec_control |= CPU_BASED_VIRTUAL_INTR_PENDING; - vmcs_write32(CPU_BASED_VM_EXEC_CONTROL, cpu_based_vm_exec_control); -} - -static void enable_nmi_window(struct kvm_vcpu *vcpu) -{ - u32 cpu_based_vm_exec_control; - - if (!cpu_has_virtual_nmis()) - return; - - cpu_based_vm_exec_control = vmcs_read32(CPU_BASED_VM_EXEC_CONTROL); - cpu_based_vm_exec_control |= CPU_BASED_VIRTUAL_NMI_PENDING; - vmcs_write32(CPU_BASED_VM_EXEC_CONTROL, cpu_based_vm_exec_control); -} - -static void enable_intr_window(struct kvm_vcpu *vcpu) -{ - if (vcpu->arch.nmi_pending) - enable_nmi_window(vcpu); - else if (kvm_cpu_has_interrupt(vcpu)) - enable_irq_window(vcpu); -} - static void vmx_complete_interrupts(struct vcpu_vmx *vmx) { u32 exit_intr_info; @@ -3165,13 +3148,16 @@ static void vmx_intr_assist(struct kvm_vcpu *vcpu) vcpu->arch.nmi_pending = false; vcpu->arch.nmi_injected = true; } else { - enable_intr_window(vcpu); + enable_nmi_window(vcpu); return; } } if (vcpu->arch.nmi_injected) { vmx_inject_nmi(vcpu); - enable_intr_window(vcpu); + if (vcpu->arch.nmi_pending) + enable_nmi_window(vcpu); + else if (kvm_cpu_has_interrupt(vcpu)) + enable_irq_window(vcpu); return; } } -- cgit v1.2.3 From 66a5a347c2690db4c0756524a8eb5a05e0437aa8 Mon Sep 17 00:00:00 2001 From: Jan Kiszka Date: Fri, 26 Sep 2008 09:30:51 +0200 Subject: KVM: VMX: fix real-mode NMI support Fix NMI injection in real-mode with the same pattern we perform IRQ injection. Signed-off-by: Jan Kiszka Signed-off-by: Avi Kivity --- arch/x86/kvm/vmx.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index 440f56cd4bdd..38d138566617 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -2358,6 +2358,19 @@ static void vmx_inject_irq(struct kvm_vcpu *vcpu, int irq) static void vmx_inject_nmi(struct kvm_vcpu *vcpu) { + struct vcpu_vmx *vmx = to_vmx(vcpu); + + if (vcpu->arch.rmode.active) { + vmx->rmode.irq.pending = true; + vmx->rmode.irq.vector = NMI_VECTOR; + vmx->rmode.irq.rip = kvm_rip_read(vcpu); + vmcs_write32(VM_ENTRY_INTR_INFO_FIELD, + NMI_VECTOR | INTR_TYPE_SOFT_INTR | + INTR_INFO_VALID_MASK); + vmcs_write32(VM_ENTRY_INSTRUCTION_LEN, 1); + kvm_rip_write(vcpu, vmx->rmode.irq.rip - 1); + return; + } vmcs_write32(VM_ENTRY_INTR_INFO_FIELD, INTR_TYPE_NMI_INTR | INTR_INFO_VALID_MASK | NMI_VECTOR); } -- cgit v1.2.3 From 23930f9521c9c4d4aa96cdb9d1e1703f3782bb94 Mon Sep 17 00:00:00 2001 From: Jan Kiszka Date: Fri, 26 Sep 2008 09:30:52 +0200 Subject: KVM: x86: Enable NMI Watchdog via in-kernel PIT source LINT0 of the LAPIC can be used to route PIT events as NMI watchdog ticks into the guest. This patch aligns the in-kernel irqchip emulation with the user space irqchip with already supports this feature. The trick is to route PIT interrupts to all LAPIC's LVT0 lines. Rebased and slightly polished patch originally posted by Sheng Yang. Signed-off-by: Jan Kiszka Signed-off-by: Sheng Yang Signed-off-by: Avi Kivity --- arch/x86/kvm/i8254.c | 15 +++++++++++++++ arch/x86/kvm/irq.h | 1 + arch/x86/kvm/lapic.c | 34 +++++++++++++++++++++++++++++----- 3 files changed, 45 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/i8254.c b/arch/x86/kvm/i8254.c index 59ebd37ad79e..580cc1d01c7d 100644 --- a/arch/x86/kvm/i8254.c +++ b/arch/x86/kvm/i8254.c @@ -603,10 +603,25 @@ void kvm_free_pit(struct kvm *kvm) static void __inject_pit_timer_intr(struct kvm *kvm) { + struct kvm_vcpu *vcpu; + int i; + mutex_lock(&kvm->lock); kvm_set_irq(kvm, kvm->arch.vpit->irq_source_id, 0, 1); kvm_set_irq(kvm, kvm->arch.vpit->irq_source_id, 0, 0); mutex_unlock(&kvm->lock); + + /* + * Provides NMI watchdog support in IOAPIC mode. + * The route is: PIT -> PIC -> LVT0 in NMI mode, + * timer IRQs will continue to flow through the IOAPIC. + */ + for (i = 0; i < KVM_MAX_VCPUS; ++i) { + vcpu = kvm->vcpus[i]; + if (!vcpu) + continue; + kvm_apic_local_deliver(vcpu, APIC_LVT0); + } } void kvm_inject_pit_timer_irqs(struct kvm_vcpu *vcpu) diff --git a/arch/x86/kvm/irq.h b/arch/x86/kvm/irq.h index f17c8f5bbf31..71e37a530cf7 100644 --- a/arch/x86/kvm/irq.h +++ b/arch/x86/kvm/irq.h @@ -87,6 +87,7 @@ void kvm_pic_reset(struct kvm_kpic_state *s); void kvm_timer_intr_post(struct kvm_vcpu *vcpu, int vec); void kvm_inject_pending_timer_irqs(struct kvm_vcpu *vcpu); void kvm_inject_apic_timer_irqs(struct kvm_vcpu *vcpu); +int kvm_apic_local_deliver(struct kvm_vcpu *vcpu, int lvt_type); void __kvm_migrate_apic_timer(struct kvm_vcpu *vcpu); void __kvm_migrate_pit_timer(struct kvm_vcpu *vcpu); void __kvm_migrate_timers(struct kvm_vcpu *vcpu); diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c index 0fc3cab48943..206cc11a1c97 100644 --- a/arch/x86/kvm/lapic.c +++ b/arch/x86/kvm/lapic.c @@ -380,6 +380,14 @@ static int __apic_accept_irq(struct kvm_lapic *apic, int delivery_mode, } break; + case APIC_DM_EXTINT: + /* + * Should only be called by kvm_apic_local_deliver() with LVT0, + * before NMI watchdog was enabled. Already handled by + * kvm_apic_accept_pic_intr(). + */ + break; + default: printk(KERN_ERR "TODO: unsupported delivery mode %x\n", delivery_mode); @@ -743,10 +751,13 @@ static void apic_mmio_write(struct kvm_io_device *this, apic_set_reg(apic, APIC_ICR2, val & 0xff000000); break; + case APIC_LVT0: + if (val == APIC_DM_NMI) + apic_debug("Receive NMI setting on APIC_LVT0 " + "for cpu %d\n", apic->vcpu->vcpu_id); case APIC_LVTT: case APIC_LVTTHMR: case APIC_LVTPC: - case APIC_LVT0: case APIC_LVT1: case APIC_LVTERR: /* TODO: Check vector */ @@ -961,12 +972,25 @@ int apic_has_pending_timer(struct kvm_vcpu *vcpu) return 0; } -static int __inject_apic_timer_irq(struct kvm_lapic *apic) +int kvm_apic_local_deliver(struct kvm_vcpu *vcpu, int lvt_type) { - int vector; + struct kvm_lapic *apic = vcpu->arch.apic; + int vector, mode, trig_mode; + u32 reg; + + if (apic && apic_enabled(apic)) { + reg = apic_get_reg(apic, lvt_type); + vector = reg & APIC_VECTOR_MASK; + mode = reg & APIC_MODE_MASK; + trig_mode = reg & APIC_LVT_LEVEL_TRIGGER; + return __apic_accept_irq(apic, mode, vector, 1, trig_mode); + } + return 0; +} - vector = apic_lvt_vector(apic, APIC_LVTT); - return __apic_accept_irq(apic, APIC_DM_FIXED, vector, 1, 0); +static inline int __inject_apic_timer_irq(struct kvm_lapic *apic) +{ + return kvm_apic_local_deliver(apic->vcpu, APIC_LVTT); } static enum hrtimer_restart apic_timer_fn(struct hrtimer *data) -- cgit v1.2.3 From 0496fbb973ccc9477082e859ed0faab5acb805ba Mon Sep 17 00:00:00 2001 From: Jan Kiszka Date: Fri, 26 Sep 2008 09:30:53 +0200 Subject: KVM: x86: VCPU with pending NMI is runnabled Ensure that a VCPU with pending NMIs is considered runnable. Signed-off-by: Jan Kiszka Signed-off-by: Avi Kivity --- arch/x86/kvm/x86.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 1a71f6735593..1fa9a6db633d 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -4130,7 +4130,8 @@ void kvm_arch_flush_shadow(struct kvm *kvm) int kvm_arch_vcpu_runnable(struct kvm_vcpu *vcpu) { return vcpu->arch.mp_state == KVM_MP_STATE_RUNNABLE - || vcpu->arch.mp_state == KVM_MP_STATE_SIPI_RECEIVED; + || vcpu->arch.mp_state == KVM_MP_STATE_SIPI_RECEIVED + || vcpu->arch.nmi_pending; } static void vcpu_kick_intr(void *info) -- cgit v1.2.3 From 26df99c6c5807115f06d4e1abae397b7f5f3e00c Mon Sep 17 00:00:00 2001 From: Jan Kiszka Date: Fri, 26 Sep 2008 09:30:54 +0200 Subject: KVM: Kick NMI receiving VCPU Kick the NMI receiving VCPU in case the triggering caller runs in a different context. Signed-off-by: Jan Kiszka Signed-off-by: Avi Kivity --- arch/x86/kvm/lapic.c | 1 + virt/kvm/ioapic.c | 1 + 2 files changed, 2 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c index 206cc11a1c97..304f9ddbdd51 100644 --- a/arch/x86/kvm/lapic.c +++ b/arch/x86/kvm/lapic.c @@ -354,6 +354,7 @@ static int __apic_accept_irq(struct kvm_lapic *apic, int delivery_mode, case APIC_DM_NMI: kvm_inject_nmi(vcpu); + kvm_vcpu_kick(vcpu); break; case APIC_DM_INIT: diff --git a/virt/kvm/ioapic.c b/virt/kvm/ioapic.c index 53772bb46320..c8f939c55075 100644 --- a/virt/kvm/ioapic.c +++ b/virt/kvm/ioapic.c @@ -150,6 +150,7 @@ static int ioapic_inj_irq(struct kvm_ioapic *ioapic, static void ioapic_inj_nmi(struct kvm_vcpu *vcpu) { kvm_inject_nmi(vcpu); + kvm_vcpu_kick(vcpu); } static u32 ioapic_get_delivery_bitmask(struct kvm_ioapic *ioapic, u8 dest, -- cgit v1.2.3 From c4abb7c9cde24b7351a47328ef866e6a2bbb1ad0 Mon Sep 17 00:00:00 2001 From: Jan Kiszka Date: Fri, 26 Sep 2008 09:30:55 +0200 Subject: KVM: x86: Support for user space injected NMIs Introduces the KVM_NMI IOCTL to the generic x86 part of KVM for injecting NMIs from user space and also extends the statistic report accordingly. Based on the original patch by Sheng Yang. Signed-off-by: Jan Kiszka Signed-off-by: Sheng Yang Signed-off-by: Avi Kivity --- arch/x86/include/asm/kvm_host.h | 2 ++ arch/x86/kvm/x86.c | 46 +++++++++++++++++++++++++++++++++++++++-- include/linux/kvm.h | 11 ++++++++-- 3 files changed, 55 insertions(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index bfbbdea869bf..a40fa8478920 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -398,6 +398,7 @@ struct kvm_vcpu_stat { u32 halt_exits; u32 halt_wakeup; u32 request_irq_exits; + u32 request_nmi_exits; u32 irq_exits; u32 host_state_reload; u32 efer_reload; @@ -406,6 +407,7 @@ struct kvm_vcpu_stat { u32 insn_emulation_fail; u32 hypercalls; u32 irq_injections; + u32 nmi_injections; }; struct descriptor_table { diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 1fa9a6db633d..07971451b947 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -86,6 +86,7 @@ struct kvm_stats_debugfs_item debugfs_entries[] = { { "halt_wakeup", VCPU_STAT(halt_wakeup) }, { "hypercalls", VCPU_STAT(hypercalls) }, { "request_irq", VCPU_STAT(request_irq_exits) }, + { "request_nmi", VCPU_STAT(request_nmi_exits) }, { "irq_exits", VCPU_STAT(irq_exits) }, { "host_state_reload", VCPU_STAT(host_state_reload) }, { "efer_reload", VCPU_STAT(efer_reload) }, @@ -93,6 +94,7 @@ struct kvm_stats_debugfs_item debugfs_entries[] = { { "insn_emulation", VCPU_STAT(insn_emulation) }, { "insn_emulation_fail", VCPU_STAT(insn_emulation_fail) }, { "irq_injections", VCPU_STAT(irq_injections) }, + { "nmi_injections", VCPU_STAT(nmi_injections) }, { "mmu_shadow_zapped", VM_STAT(mmu_shadow_zapped) }, { "mmu_pte_write", VM_STAT(mmu_pte_write) }, { "mmu_pte_updated", VM_STAT(mmu_pte_updated) }, @@ -1318,6 +1320,15 @@ static int kvm_vcpu_ioctl_interrupt(struct kvm_vcpu *vcpu, return 0; } +static int kvm_vcpu_ioctl_nmi(struct kvm_vcpu *vcpu) +{ + vcpu_load(vcpu); + kvm_inject_nmi(vcpu); + vcpu_put(vcpu); + + return 0; +} + static int vcpu_ioctl_tpr_access_reporting(struct kvm_vcpu *vcpu, struct kvm_tpr_access_ctl *tac) { @@ -1377,6 +1388,13 @@ long kvm_arch_vcpu_ioctl(struct file *filp, r = 0; break; } + case KVM_NMI: { + r = kvm_vcpu_ioctl_nmi(vcpu); + if (r) + goto out; + r = 0; + break; + } case KVM_SET_CPUID: { struct kvm_cpuid __user *cpuid_arg = argp; struct kvm_cpuid cpuid; @@ -2812,18 +2830,37 @@ static int dm_request_for_irq_injection(struct kvm_vcpu *vcpu, (kvm_x86_ops->get_rflags(vcpu) & X86_EFLAGS_IF)); } +/* + * Check if userspace requested a NMI window, and that the NMI window + * is open. + * + * No need to exit to userspace if we already have a NMI queued. + */ +static int dm_request_for_nmi_injection(struct kvm_vcpu *vcpu, + struct kvm_run *kvm_run) +{ + return (!vcpu->arch.nmi_pending && + kvm_run->request_nmi_window && + vcpu->arch.nmi_window_open); +} + static void post_kvm_run_save(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run) { kvm_run->if_flag = (kvm_x86_ops->get_rflags(vcpu) & X86_EFLAGS_IF) != 0; kvm_run->cr8 = kvm_get_cr8(vcpu); kvm_run->apic_base = kvm_get_apic_base(vcpu); - if (irqchip_in_kernel(vcpu->kvm)) + if (irqchip_in_kernel(vcpu->kvm)) { kvm_run->ready_for_interrupt_injection = 1; - else + kvm_run->ready_for_nmi_injection = 1; + } else { kvm_run->ready_for_interrupt_injection = (vcpu->arch.interrupt_window_open && vcpu->arch.irq_summary == 0); + kvm_run->ready_for_nmi_injection = + (vcpu->arch.nmi_window_open && + vcpu->arch.nmi_pending == 0); + } } static void vapic_enter(struct kvm_vcpu *vcpu) @@ -2999,6 +3036,11 @@ static int __vcpu_run(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run) } if (r > 0) { + if (dm_request_for_nmi_injection(vcpu, kvm_run)) { + r = -EINTR; + kvm_run->exit_reason = KVM_EXIT_NMI; + ++vcpu->stat.request_nmi_exits; + } if (dm_request_for_irq_injection(vcpu, kvm_run)) { r = -EINTR; kvm_run->exit_reason = KVM_EXIT_INTR; diff --git a/include/linux/kvm.h b/include/linux/kvm.h index f18b86fa8655..44fd7fa0af2b 100644 --- a/include/linux/kvm.h +++ b/include/linux/kvm.h @@ -83,18 +83,22 @@ struct kvm_irqchip { #define KVM_EXIT_S390_SIEIC 13 #define KVM_EXIT_S390_RESET 14 #define KVM_EXIT_DCR 15 +#define KVM_EXIT_NMI 16 +#define KVM_EXIT_NMI_WINDOW_OPEN 17 /* for KVM_RUN, returned by mmap(vcpu_fd, offset=0) */ struct kvm_run { /* in */ __u8 request_interrupt_window; - __u8 padding1[7]; + __u8 request_nmi_window; + __u8 padding1[6]; /* out */ __u32 exit_reason; __u8 ready_for_interrupt_injection; __u8 if_flag; - __u8 padding2[2]; + __u8 ready_for_nmi_injection; + __u8 padding2; /* in (pre_kvm_run), out (post_kvm_run) */ __u64 cr8; @@ -387,6 +391,7 @@ struct kvm_trace_rec { #define KVM_CAP_DEVICE_ASSIGNMENT 17 #endif #define KVM_CAP_IOMMU 18 +#define KVM_CAP_NMI 19 /* * ioctls for VM fds @@ -458,6 +463,8 @@ struct kvm_trace_rec { #define KVM_S390_INITIAL_RESET _IO(KVMIO, 0x97) #define KVM_GET_MP_STATE _IOR(KVMIO, 0x98, struct kvm_mp_state) #define KVM_SET_MP_STATE _IOW(KVMIO, 0x99, struct kvm_mp_state) +/* Available with KVM_CAP_NMI */ +#define KVM_NMI _IO(KVMIO, 0x9a) #define KVM_TRC_INJ_VIRQ (KVM_TRC_HANDLER + 0x02) #define KVM_TRC_REDELIVER_EVT (KVM_TRC_HANDLER + 0x03) -- cgit v1.2.3 From 487b391d6ea9b1d0e2e0440466fb3130e78c98d9 Mon Sep 17 00:00:00 2001 From: Jan Kiszka Date: Fri, 26 Sep 2008 09:30:56 +0200 Subject: KVM: VMX: Provide support for user space injected NMIs This patch adds the required bits to the VMX side for user space injected NMIs. As with the preexisting in-kernel irqchip support, the CPU must provide the "virtual NMI" feature for proper tracking of the NMI blocking state. Based on the original patch by Sheng Yang. Signed-off-by: Jan Kiszka Signed-off-by: Sheng Yang Signed-off-by: Avi Kivity --- arch/x86/kvm/vmx.c | 33 +++++++++++++++++++++++++++++++++ 1 file changed, 33 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index 38d138566617..f16a62c79267 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -2360,6 +2360,7 @@ static void vmx_inject_nmi(struct kvm_vcpu *vcpu) { struct vcpu_vmx *vmx = to_vmx(vcpu); + ++vcpu->stat.nmi_injections; if (vcpu->arch.rmode.active) { vmx->rmode.irq.pending = true; vmx->rmode.irq.vector = NMI_VECTOR; @@ -2428,6 +2429,30 @@ static void do_interrupt_requests(struct kvm_vcpu *vcpu, { vmx_update_window_states(vcpu); + if (cpu_has_virtual_nmis()) { + if (vcpu->arch.nmi_pending && !vcpu->arch.nmi_injected) { + if (vcpu->arch.nmi_window_open) { + vcpu->arch.nmi_pending = false; + vcpu->arch.nmi_injected = true; + } else { + enable_nmi_window(vcpu); + return; + } + } + if (vcpu->arch.nmi_injected) { + vmx_inject_nmi(vcpu); + if (vcpu->arch.nmi_pending + || kvm_run->request_nmi_window) + enable_nmi_window(vcpu); + else if (vcpu->arch.irq_summary + || kvm_run->request_interrupt_window) + enable_irq_window(vcpu); + return; + } + if (!vcpu->arch.nmi_window_open || kvm_run->request_nmi_window) + enable_nmi_window(vcpu); + } + if (vcpu->arch.interrupt_window_open) { if (vcpu->arch.irq_summary && !vcpu->arch.interrupt.pending) kvm_do_inject_irq(vcpu); @@ -2959,6 +2984,14 @@ static int handle_nmi_window(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run) vmcs_write32(CPU_BASED_VM_EXEC_CONTROL, cpu_based_vm_exec_control); ++vcpu->stat.nmi_window_exits; + /* + * If the user space waits to inject a NMI, exit as soon as possible + */ + if (kvm_run->request_nmi_window && !vcpu->arch.nmi_pending) { + kvm_run->exit_reason = KVM_EXIT_NMI_WINDOW_OPEN; + return 0; + } + return 1; } -- cgit v1.2.3 From 3b86cd9967242f3f3d775ee015fb814a349ed5e6 Mon Sep 17 00:00:00 2001 From: Jan Kiszka Date: Fri, 26 Sep 2008 09:30:57 +0200 Subject: KVM: VMX: work around lacking VNMI support Older VMX supporting CPUs do not provide the "Virtual NMI" feature for tracking the NMI-blocked state after injecting such events. For now KVM is unable to inject NMIs on those CPUs. Derived from Sheng Yang's suggestion to use the IRQ window notification for detecting the end of NMI handlers, this patch implements virtual NMI support without impact on the host's ability to receive real NMIs. The downside is that the given approach requires some heuristics that can cause NMI nesting in vary rare corner cases. The approach works as follows: - inject NMI and set a software-based NMI-blocked flag - arm the IRQ window start notification whenever an NMI window is requested - if the guest exits due to an opening IRQ window, clear the emulated NMI-blocked flag - if the guest net execution time with NMI-blocked but without an IRQ window exceeds 1 second, force NMI-blocked reset and inject anyway This approach covers most practical scenarios: - succeeding NMIs are seperated by at least one open IRQ window - the guest may spin with IRQs disabled (e.g. due to a bug), but leaving the NMI handler takes much less time than one second - the guest does not rely on strict ordering or timing of NMIs (would be problematic in virtualized environments anyway) Successfully tested with the 'nmi n' monitor command, the kgdbts testsuite on smp guests (additional patches required to add debug register support to kvm) + the kernel's nmi_watchdog=1, and a Siemens- specific board emulation (+ guest) that comes with its own NMI watchdog mechanism. Signed-off-by: Jan Kiszka Signed-off-by: Avi Kivity --- arch/x86/kvm/vmx.c | 174 +++++++++++++++++++++++++++++++++++------------------ 1 file changed, 115 insertions(+), 59 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index f16a62c79267..2180109d794c 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -90,6 +90,11 @@ struct vcpu_vmx { } rmode; int vpid; bool emulation_required; + + /* Support for vnmi-less CPUs */ + int soft_vnmi_blocked; + ktime_t entry_time; + s64 vnmi_blocked_time; }; static inline struct vcpu_vmx *to_vmx(struct kvm_vcpu *vcpu) @@ -2230,6 +2235,8 @@ static int vmx_vcpu_reset(struct kvm_vcpu *vcpu) vmx->vcpu.arch.rmode.active = 0; + vmx->soft_vnmi_blocked = 0; + vmx->vcpu.arch.regs[VCPU_REGS_RDX] = get_rdx_init_val(); kvm_set_cr8(&vmx->vcpu, 0); msr = 0xfee00000 | MSR_IA32_APICBASE_ENABLE; @@ -2335,6 +2342,29 @@ out: return ret; } +static void enable_irq_window(struct kvm_vcpu *vcpu) +{ + u32 cpu_based_vm_exec_control; + + cpu_based_vm_exec_control = vmcs_read32(CPU_BASED_VM_EXEC_CONTROL); + cpu_based_vm_exec_control |= CPU_BASED_VIRTUAL_INTR_PENDING; + vmcs_write32(CPU_BASED_VM_EXEC_CONTROL, cpu_based_vm_exec_control); +} + +static void enable_nmi_window(struct kvm_vcpu *vcpu) +{ + u32 cpu_based_vm_exec_control; + + if (!cpu_has_virtual_nmis()) { + enable_irq_window(vcpu); + return; + } + + cpu_based_vm_exec_control = vmcs_read32(CPU_BASED_VM_EXEC_CONTROL); + cpu_based_vm_exec_control |= CPU_BASED_VIRTUAL_NMI_PENDING; + vmcs_write32(CPU_BASED_VM_EXEC_CONTROL, cpu_based_vm_exec_control); +} + static void vmx_inject_irq(struct kvm_vcpu *vcpu, int irq) { struct vcpu_vmx *vmx = to_vmx(vcpu); @@ -2360,6 +2390,19 @@ static void vmx_inject_nmi(struct kvm_vcpu *vcpu) { struct vcpu_vmx *vmx = to_vmx(vcpu); + if (!cpu_has_virtual_nmis()) { + /* + * Tracking the NMI-blocked state in software is built upon + * finding the next open IRQ window. This, in turn, depends on + * well-behaving guests: They have to keep IRQs disabled at + * least as long as the NMI handler runs. Otherwise we may + * cause NMI nesting, maybe breaking the guest. But as this is + * highly unlikely, we can live with the residual risk. + */ + vmx->soft_vnmi_blocked = 1; + vmx->vnmi_blocked_time = 0; + } + ++vcpu->stat.nmi_injections; if (vcpu->arch.rmode.active) { vmx->rmode.irq.pending = true; @@ -2384,6 +2427,8 @@ static void vmx_update_window_states(struct kvm_vcpu *vcpu) !(guest_intr & (GUEST_INTR_STATE_STI | GUEST_INTR_STATE_MOV_SS | GUEST_INTR_STATE_NMI)); + if (!cpu_has_virtual_nmis() && to_vmx(vcpu)->soft_vnmi_blocked) + vcpu->arch.nmi_window_open = 0; vcpu->arch.interrupt_window_open = ((vmcs_readl(GUEST_RFLAGS) & X86_EFLAGS_IF) && @@ -2403,55 +2448,31 @@ static void kvm_do_inject_irq(struct kvm_vcpu *vcpu) kvm_queue_interrupt(vcpu, irq); } -static void enable_irq_window(struct kvm_vcpu *vcpu) -{ - u32 cpu_based_vm_exec_control; - - cpu_based_vm_exec_control = vmcs_read32(CPU_BASED_VM_EXEC_CONTROL); - cpu_based_vm_exec_control |= CPU_BASED_VIRTUAL_INTR_PENDING; - vmcs_write32(CPU_BASED_VM_EXEC_CONTROL, cpu_based_vm_exec_control); -} - -static void enable_nmi_window(struct kvm_vcpu *vcpu) -{ - u32 cpu_based_vm_exec_control; - - if (!cpu_has_virtual_nmis()) - return; - - cpu_based_vm_exec_control = vmcs_read32(CPU_BASED_VM_EXEC_CONTROL); - cpu_based_vm_exec_control |= CPU_BASED_VIRTUAL_NMI_PENDING; - vmcs_write32(CPU_BASED_VM_EXEC_CONTROL, cpu_based_vm_exec_control); -} - static void do_interrupt_requests(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run) { vmx_update_window_states(vcpu); - if (cpu_has_virtual_nmis()) { - if (vcpu->arch.nmi_pending && !vcpu->arch.nmi_injected) { - if (vcpu->arch.nmi_window_open) { - vcpu->arch.nmi_pending = false; - vcpu->arch.nmi_injected = true; - } else { - enable_nmi_window(vcpu); - return; - } - } - if (vcpu->arch.nmi_injected) { - vmx_inject_nmi(vcpu); - if (vcpu->arch.nmi_pending - || kvm_run->request_nmi_window) - enable_nmi_window(vcpu); - else if (vcpu->arch.irq_summary - || kvm_run->request_interrupt_window) - enable_irq_window(vcpu); + if (vcpu->arch.nmi_pending && !vcpu->arch.nmi_injected) { + if (vcpu->arch.nmi_window_open) { + vcpu->arch.nmi_pending = false; + vcpu->arch.nmi_injected = true; + } else { + enable_nmi_window(vcpu); return; } - if (!vcpu->arch.nmi_window_open || kvm_run->request_nmi_window) + } + if (vcpu->arch.nmi_injected) { + vmx_inject_nmi(vcpu); + if (vcpu->arch.nmi_pending || kvm_run->request_nmi_window) enable_nmi_window(vcpu); + else if (vcpu->arch.irq_summary + || kvm_run->request_interrupt_window) + enable_irq_window(vcpu); + return; } + if (!vcpu->arch.nmi_window_open || kvm_run->request_nmi_window) + enable_nmi_window(vcpu); if (vcpu->arch.interrupt_window_open) { if (vcpu->arch.irq_summary && !vcpu->arch.interrupt.pending) @@ -3097,6 +3118,37 @@ static int kvm_handle_exit(struct kvm_run *kvm_run, struct kvm_vcpu *vcpu) printk(KERN_WARNING "%s: unexpected, valid vectoring info " "(0x%x) and exit reason is 0x%x\n", __func__, vectoring_info, exit_reason); + + if (unlikely(!cpu_has_virtual_nmis() && vmx->soft_vnmi_blocked)) { + if (vcpu->arch.interrupt_window_open) { + vmx->soft_vnmi_blocked = 0; + vcpu->arch.nmi_window_open = 1; + } else if (vmx->vnmi_blocked_time > 1000000000LL && + (kvm_run->request_nmi_window || vcpu->arch.nmi_pending)) { + /* + * This CPU don't support us in finding the end of an + * NMI-blocked window if the guest runs with IRQs + * disabled. So we pull the trigger after 1 s of + * futile waiting, but inform the user about this. + */ + printk(KERN_WARNING "%s: Breaking out of NMI-blocked " + "state on VCPU %d after 1 s timeout\n", + __func__, vcpu->vcpu_id); + vmx->soft_vnmi_blocked = 0; + vmx->vcpu.arch.nmi_window_open = 1; + } + + /* + * If the user space waits to inject an NNI, exit ASAP + */ + if (vcpu->arch.nmi_window_open && kvm_run->request_nmi_window + && !vcpu->arch.nmi_pending) { + kvm_run->exit_reason = KVM_EXIT_NMI_WINDOW_OPEN; + ++vcpu->stat.nmi_window_exits; + return 0; + } + } + if (exit_reason < kvm_vmx_max_exit_handlers && kvm_vmx_exit_handlers[exit_reason]) return kvm_vmx_exit_handlers[exit_reason](vcpu, kvm_run); @@ -3146,7 +3198,9 @@ static void vmx_complete_interrupts(struct vcpu_vmx *vmx) if (unblock_nmi && vector != DF_VECTOR) vmcs_set_bits(GUEST_INTERRUPTIBILITY_INFO, GUEST_INTR_STATE_NMI); - } + } else if (unlikely(vmx->soft_vnmi_blocked)) + vmx->vnmi_blocked_time += + ktime_to_ns(ktime_sub(ktime_get(), vmx->entry_time)); idt_vectoring_info = vmx->idt_vectoring_info; idtv_info_valid = idt_vectoring_info & VECTORING_INFO_VALID_MASK; @@ -3186,27 +3240,25 @@ static void vmx_intr_assist(struct kvm_vcpu *vcpu) vmx_update_window_states(vcpu); - if (cpu_has_virtual_nmis()) { - if (vcpu->arch.nmi_pending && !vcpu->arch.nmi_injected) { - if (vcpu->arch.interrupt.pending) { - enable_nmi_window(vcpu); - } else if (vcpu->arch.nmi_window_open) { - vcpu->arch.nmi_pending = false; - vcpu->arch.nmi_injected = true; - } else { - enable_nmi_window(vcpu); - return; - } - } - if (vcpu->arch.nmi_injected) { - vmx_inject_nmi(vcpu); - if (vcpu->arch.nmi_pending) - enable_nmi_window(vcpu); - else if (kvm_cpu_has_interrupt(vcpu)) - enable_irq_window(vcpu); + if (vcpu->arch.nmi_pending && !vcpu->arch.nmi_injected) { + if (vcpu->arch.interrupt.pending) { + enable_nmi_window(vcpu); + } else if (vcpu->arch.nmi_window_open) { + vcpu->arch.nmi_pending = false; + vcpu->arch.nmi_injected = true; + } else { + enable_nmi_window(vcpu); return; } } + if (vcpu->arch.nmi_injected) { + vmx_inject_nmi(vcpu); + if (vcpu->arch.nmi_pending) + enable_nmi_window(vcpu); + else if (kvm_cpu_has_interrupt(vcpu)) + enable_irq_window(vcpu); + return; + } if (!vcpu->arch.interrupt.pending && kvm_cpu_has_interrupt(vcpu)) { if (vcpu->arch.interrupt_window_open) kvm_queue_interrupt(vcpu, kvm_cpu_get_interrupt(vcpu)); @@ -3255,6 +3307,10 @@ static void vmx_vcpu_run(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run) struct vcpu_vmx *vmx = to_vmx(vcpu); u32 intr_info; + /* Record the guest's net vcpu time for enforced NMI injections. */ + if (unlikely(!cpu_has_virtual_nmis() && vmx->soft_vnmi_blocked)) + vmx->entry_time = ktime_get(); + /* Handle invalid guest state instead of entering VMX */ if (vmx->emulation_required && emulate_invalid_guest_state) { handle_invalid_guest_state(vcpu, kvm_run); -- cgit v1.2.3 From 5f179287fa02723215eecf681d812b303c243973 Mon Sep 17 00:00:00 2001 From: Gleb Natapov Date: Tue, 7 Oct 2008 15:42:33 +0200 Subject: KVM: call kvm_arch_vcpu_reset() instead of the kvm_x86_ops callback Call kvm_arch_vcpu_reset() instead of directly using arch callback. The function does additional things. Signed-off-by: Gleb Natapov Signed-off-by: Avi Kivity --- arch/x86/kvm/x86.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 07971451b947..a2c4b5594555 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -3010,7 +3010,7 @@ static int __vcpu_run(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run) pr_debug("vcpu %d received sipi with vector # %x\n", vcpu->vcpu_id, vcpu->arch.sipi_vector); kvm_lapic_reset(vcpu); - r = kvm_x86_ops->vcpu_reset(vcpu); + r = kvm_arch_vcpu_reset(vcpu); if (r) return r; vcpu->arch.mp_state = KVM_MP_STATE_RUNNABLE; -- cgit v1.2.3 From b558bc0a25c82ef2a9d2683b0beb3e4b87cea20b Mon Sep 17 00:00:00 2001 From: Sheng Yang Date: Thu, 9 Oct 2008 16:01:52 +0800 Subject: x86: Rename mtrr_state struct and macro names Prepare for exporting them. Signed-off-by: Sheng Yang Signed-off-by: Avi Kivity --- arch/x86/kernel/cpu/mtrr/generic.c | 8 ++++---- arch/x86/kernel/cpu/mtrr/main.c | 4 ++-- arch/x86/kernel/cpu/mtrr/mtrr.h | 7 ++++--- 3 files changed, 10 insertions(+), 9 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/mtrr/generic.c b/arch/x86/kernel/cpu/mtrr/generic.c index 4e8d77f01eeb..90db91e15931 100644 --- a/arch/x86/kernel/cpu/mtrr/generic.c +++ b/arch/x86/kernel/cpu/mtrr/generic.c @@ -14,9 +14,9 @@ #include #include "mtrr.h" -struct mtrr_state { - struct mtrr_var_range var_ranges[MAX_VAR_RANGES]; - mtrr_type fixed_ranges[NUM_FIXED_RANGES]; +struct mtrr_state_type { + struct mtrr_var_range var_ranges[MTRR_MAX_VAR_RANGES]; + mtrr_type fixed_ranges[MTRR_NUM_FIXED_RANGES]; unsigned char enabled; unsigned char have_fixed; mtrr_type def_type; @@ -35,7 +35,7 @@ static struct fixed_range_block fixed_range_blocks[] = { }; static unsigned long smp_changes_mask; -static struct mtrr_state mtrr_state = {}; +static struct mtrr_state_type mtrr_state = {}; static int mtrr_state_set; u64 mtrr_tom2; diff --git a/arch/x86/kernel/cpu/mtrr/main.c b/arch/x86/kernel/cpu/mtrr/main.c index 1159e269e596..d6ec7ec30274 100644 --- a/arch/x86/kernel/cpu/mtrr/main.c +++ b/arch/x86/kernel/cpu/mtrr/main.c @@ -49,7 +49,7 @@ u32 num_var_ranges = 0; -unsigned int mtrr_usage_table[MAX_VAR_RANGES]; +unsigned int mtrr_usage_table[MTRR_MAX_VAR_RANGES]; static DEFINE_MUTEX(mtrr_mutex); u64 size_or_mask, size_and_mask; @@ -574,7 +574,7 @@ struct mtrr_value { unsigned long lsize; }; -static struct mtrr_value mtrr_state[MAX_VAR_RANGES]; +static struct mtrr_value mtrr_state[MTRR_MAX_VAR_RANGES]; static int mtrr_save(struct sys_device * sysdev, pm_message_t state) { diff --git a/arch/x86/kernel/cpu/mtrr/mtrr.h b/arch/x86/kernel/cpu/mtrr/mtrr.h index 2dc4ec656b23..988538202ddc 100644 --- a/arch/x86/kernel/cpu/mtrr/mtrr.h +++ b/arch/x86/kernel/cpu/mtrr/mtrr.h @@ -11,8 +11,9 @@ #define MTRRphysBase_MSR(reg) (0x200 + 2 * (reg)) #define MTRRphysMask_MSR(reg) (0x200 + 2 * (reg) + 1) -#define NUM_FIXED_RANGES 88 -#define MAX_VAR_RANGES 256 +#define MTRR_NUM_FIXED_RANGES 88 +#define MTRR_MAX_VAR_RANGES 256 + #define MTRRfix64K_00000_MSR 0x250 #define MTRRfix16K_80000_MSR 0x258 #define MTRRfix16K_A0000_MSR 0x259 @@ -33,7 +34,7 @@ an 8 bit field: */ typedef u8 mtrr_type; -extern unsigned int mtrr_usage_table[MAX_VAR_RANGES]; +extern unsigned int mtrr_usage_table[MTRR_MAX_VAR_RANGES]; struct mtrr_ops { u32 vendor; -- cgit v1.2.3 From 932d27a7913fc6b3c64c6e6082628b0a1561dec9 Mon Sep 17 00:00:00 2001 From: Sheng Yang Date: Thu, 9 Oct 2008 16:01:53 +0800 Subject: x86: Export some definition of MTRR For KVM can reuse the type define, and need them to support shadow MTRR. Signed-off-by: Sheng Yang Signed-off-by: Avi Kivity --- arch/x86/include/asm/mtrr.h | 25 +++++++++++++++++++++++++ arch/x86/kernel/cpu/mtrr/generic.c | 12 +++--------- arch/x86/kernel/cpu/mtrr/mtrr.h | 17 ----------------- 3 files changed, 28 insertions(+), 26 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/mtrr.h b/arch/x86/include/asm/mtrr.h index 7c1e4258b31e..cb988aab716d 100644 --- a/arch/x86/include/asm/mtrr.h +++ b/arch/x86/include/asm/mtrr.h @@ -57,6 +57,31 @@ struct mtrr_gentry { }; #endif /* !__i386__ */ +struct mtrr_var_range { + u32 base_lo; + u32 base_hi; + u32 mask_lo; + u32 mask_hi; +}; + +/* In the Intel processor's MTRR interface, the MTRR type is always held in + an 8 bit field: */ +typedef u8 mtrr_type; + +#define MTRR_NUM_FIXED_RANGES 88 +#define MTRR_MAX_VAR_RANGES 256 + +struct mtrr_state_type { + struct mtrr_var_range var_ranges[MTRR_MAX_VAR_RANGES]; + mtrr_type fixed_ranges[MTRR_NUM_FIXED_RANGES]; + unsigned char enabled; + unsigned char have_fixed; + mtrr_type def_type; +}; + +#define MTRRphysBase_MSR(reg) (0x200 + 2 * (reg)) +#define MTRRphysMask_MSR(reg) (0x200 + 2 * (reg) + 1) + /* These are the various ioctls */ #define MTRRIOC_ADD_ENTRY _IOW(MTRR_IOCTL_BASE, 0, struct mtrr_sentry) #define MTRRIOC_SET_ENTRY _IOW(MTRR_IOCTL_BASE, 1, struct mtrr_sentry) diff --git a/arch/x86/kernel/cpu/mtrr/generic.c b/arch/x86/kernel/cpu/mtrr/generic.c index 90db91e15931..b59ddcc88cd8 100644 --- a/arch/x86/kernel/cpu/mtrr/generic.c +++ b/arch/x86/kernel/cpu/mtrr/generic.c @@ -14,14 +14,6 @@ #include #include "mtrr.h" -struct mtrr_state_type { - struct mtrr_var_range var_ranges[MTRR_MAX_VAR_RANGES]; - mtrr_type fixed_ranges[MTRR_NUM_FIXED_RANGES]; - unsigned char enabled; - unsigned char have_fixed; - mtrr_type def_type; -}; - struct fixed_range_block { int base_msr; /* start address of an MTRR block */ int ranges; /* number of MTRRs in this block */ @@ -35,10 +27,12 @@ static struct fixed_range_block fixed_range_blocks[] = { }; static unsigned long smp_changes_mask; -static struct mtrr_state_type mtrr_state = {}; static int mtrr_state_set; u64 mtrr_tom2; +struct mtrr_state_type mtrr_state = {}; +EXPORT_SYMBOL_GPL(mtrr_state); + #undef MODULE_PARAM_PREFIX #define MODULE_PARAM_PREFIX "mtrr." diff --git a/arch/x86/kernel/cpu/mtrr/mtrr.h b/arch/x86/kernel/cpu/mtrr/mtrr.h index 988538202ddc..ffd60409cc6d 100644 --- a/arch/x86/kernel/cpu/mtrr/mtrr.h +++ b/arch/x86/kernel/cpu/mtrr/mtrr.h @@ -8,12 +8,6 @@ #define MTRRcap_MSR 0x0fe #define MTRRdefType_MSR 0x2ff -#define MTRRphysBase_MSR(reg) (0x200 + 2 * (reg)) -#define MTRRphysMask_MSR(reg) (0x200 + 2 * (reg) + 1) - -#define MTRR_NUM_FIXED_RANGES 88 -#define MTRR_MAX_VAR_RANGES 256 - #define MTRRfix64K_00000_MSR 0x250 #define MTRRfix16K_80000_MSR 0x258 #define MTRRfix16K_A0000_MSR 0x259 @@ -30,10 +24,6 @@ #define MTRR_CHANGE_MASK_VARIABLE 0x02 #define MTRR_CHANGE_MASK_DEFTYPE 0x04 -/* In the Intel processor's MTRR interface, the MTRR type is always held in - an 8 bit field: */ -typedef u8 mtrr_type; - extern unsigned int mtrr_usage_table[MTRR_MAX_VAR_RANGES]; struct mtrr_ops { @@ -71,13 +61,6 @@ struct set_mtrr_context { u32 ccr3; }; -struct mtrr_var_range { - u32 base_lo; - u32 base_hi; - u32 mask_lo; - u32 mask_hi; -}; - void set_mtrr_done(struct set_mtrr_context *ctxt); void set_mtrr_cache_disable(struct set_mtrr_context *ctxt); void set_mtrr_prepare_save(struct set_mtrr_context *ctxt); -- cgit v1.2.3 From 0bed3b568b68e5835ef5da888a372b9beabf7544 Mon Sep 17 00:00:00 2001 From: Sheng Yang Date: Thu, 9 Oct 2008 16:01:54 +0800 Subject: KVM: Improve MTRR structure As well as reset mmu context when set MTRR. Signed-off-by: Sheng Yang Signed-off-by: Avi Kivity --- arch/x86/include/asm/kvm_host.h | 5 +++- arch/x86/kvm/x86.c | 61 +++++++++++++++++++++++++++++++++++++++-- 2 files changed, 63 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index a40fa8478920..8082e87f628d 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -21,6 +21,7 @@ #include #include +#include #define KVM_MAX_VCPUS 16 #define KVM_MEMORY_SLOTS 32 @@ -86,6 +87,7 @@ #define KVM_MIN_FREE_MMU_PAGES 5 #define KVM_REFILL_PAGES 25 #define KVM_MAX_CPUID_ENTRIES 40 +#define KVM_NR_FIXED_MTRR_REGION 88 #define KVM_NR_VAR_MTRR 8 extern spinlock_t kvm_lock; @@ -329,7 +331,8 @@ struct kvm_vcpu_arch { bool nmi_injected; bool nmi_window_open; - u64 mtrr[0x100]; + struct mtrr_state_type mtrr_state; + u32 pat; }; struct kvm_mem_alias { diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index a2c4b5594555..f5b2334c6bda 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -39,6 +39,7 @@ #include #include #include +#include #define MAX_IO_MSRS 256 #define CR0_RESERVED_BITS \ @@ -650,10 +651,38 @@ static bool msr_mtrr_valid(unsigned msr) static int set_msr_mtrr(struct kvm_vcpu *vcpu, u32 msr, u64 data) { + u64 *p = (u64 *)&vcpu->arch.mtrr_state.fixed_ranges; + if (!msr_mtrr_valid(msr)) return 1; - vcpu->arch.mtrr[msr - 0x200] = data; + if (msr == MSR_MTRRdefType) { + vcpu->arch.mtrr_state.def_type = data; + vcpu->arch.mtrr_state.enabled = (data & 0xc00) >> 10; + } else if (msr == MSR_MTRRfix64K_00000) + p[0] = data; + else if (msr == MSR_MTRRfix16K_80000 || msr == MSR_MTRRfix16K_A0000) + p[1 + msr - MSR_MTRRfix16K_80000] = data; + else if (msr >= MSR_MTRRfix4K_C0000 && msr <= MSR_MTRRfix4K_F8000) + p[3 + msr - MSR_MTRRfix4K_C0000] = data; + else if (msr == MSR_IA32_CR_PAT) + vcpu->arch.pat = data; + else { /* Variable MTRRs */ + int idx, is_mtrr_mask; + u64 *pt; + + idx = (msr - 0x200) / 2; + is_mtrr_mask = msr - 0x200 - 2 * idx; + if (!is_mtrr_mask) + pt = + (u64 *)&vcpu->arch.mtrr_state.var_ranges[idx].base_lo; + else + pt = + (u64 *)&vcpu->arch.mtrr_state.var_ranges[idx].mask_lo; + *pt = data; + } + + kvm_mmu_reset_context(vcpu); return 0; } @@ -749,10 +778,37 @@ int kvm_get_msr(struct kvm_vcpu *vcpu, u32 msr_index, u64 *pdata) static int get_msr_mtrr(struct kvm_vcpu *vcpu, u32 msr, u64 *pdata) { + u64 *p = (u64 *)&vcpu->arch.mtrr_state.fixed_ranges; + if (!msr_mtrr_valid(msr)) return 1; - *pdata = vcpu->arch.mtrr[msr - 0x200]; + if (msr == MSR_MTRRdefType) + *pdata = vcpu->arch.mtrr_state.def_type + + (vcpu->arch.mtrr_state.enabled << 10); + else if (msr == MSR_MTRRfix64K_00000) + *pdata = p[0]; + else if (msr == MSR_MTRRfix16K_80000 || msr == MSR_MTRRfix16K_A0000) + *pdata = p[1 + msr - MSR_MTRRfix16K_80000]; + else if (msr >= MSR_MTRRfix4K_C0000 && msr <= MSR_MTRRfix4K_F8000) + *pdata = p[3 + msr - MSR_MTRRfix4K_C0000]; + else if (msr == MSR_IA32_CR_PAT) + *pdata = vcpu->arch.pat; + else { /* Variable MTRRs */ + int idx, is_mtrr_mask; + u64 *pt; + + idx = (msr - 0x200) / 2; + is_mtrr_mask = msr - 0x200 - 2 * idx; + if (!is_mtrr_mask) + pt = + (u64 *)&vcpu->arch.mtrr_state.var_ranges[idx].base_lo; + else + pt = + (u64 *)&vcpu->arch.mtrr_state.var_ranges[idx].mask_lo; + *pdata = *pt; + } + return 0; } @@ -3942,6 +3998,7 @@ int kvm_arch_vcpu_setup(struct kvm_vcpu *vcpu) /* We do fxsave: this must be aligned. */ BUG_ON((unsigned long)&vcpu->arch.host_fx_image & 0xF); + vcpu->arch.mtrr_state.have_fixed = 1; vcpu_load(vcpu); r = kvm_arch_vcpu_reset(vcpu); if (r == 0) -- cgit v1.2.3 From 468d472f3f65100d5fb88c8d45043c85b874c294 Mon Sep 17 00:00:00 2001 From: Sheng Yang Date: Thu, 9 Oct 2008 16:01:55 +0800 Subject: KVM: VMX: Add PAT support for EPT GUEST_PAT support is a new feature introduced by Intel Core i7 architecture. With this, cpu would save/load guest and host PAT automatically, for EPT memory type in guest depends on MSR_IA32_CR_PAT. Also add save/restore for MSR_IA32_CR_PAT. Signed-off-by: Sheng Yang Signed-off-by: Avi Kivity --- arch/x86/kvm/vmx.c | 29 ++++++++++++++++++++++++++--- arch/x86/kvm/vmx.h | 7 +++++++ arch/x86/kvm/x86.c | 2 +- 3 files changed, 34 insertions(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index 2180109d794c..b4c95a501cca 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -962,6 +962,13 @@ static int vmx_set_msr(struct kvm_vcpu *vcpu, u32 msr_index, u64 data) pr_unimpl(vcpu, "unimplemented perfctr wrmsr: 0x%x data 0x%llx\n", msr_index, data); break; + case MSR_IA32_CR_PAT: + if (vmcs_config.vmentry_ctrl & VM_ENTRY_LOAD_IA32_PAT) { + vmcs_write64(GUEST_IA32_PAT, data); + vcpu->arch.pat = data; + break; + } + /* Otherwise falls through to kvm_set_msr_common */ default: vmx_load_host_state(vmx); msr = find_msr_entry(vmx, msr_index); @@ -1181,12 +1188,13 @@ static __init int setup_vmcs_config(struct vmcs_config *vmcs_conf) #ifdef CONFIG_X86_64 min |= VM_EXIT_HOST_ADDR_SPACE_SIZE; #endif - opt = 0; + opt = VM_EXIT_SAVE_IA32_PAT | VM_EXIT_LOAD_IA32_PAT; if (adjust_vmx_controls(min, opt, MSR_IA32_VMX_EXIT_CTLS, &_vmexit_control) < 0) return -EIO; - min = opt = 0; + min = 0; + opt = VM_ENTRY_LOAD_IA32_PAT; if (adjust_vmx_controls(min, opt, MSR_IA32_VMX_ENTRY_CTLS, &_vmentry_control) < 0) return -EIO; @@ -2092,8 +2100,9 @@ static void vmx_disable_intercept_for_msr(struct page *msr_bitmap, u32 msr) */ static int vmx_vcpu_setup(struct vcpu_vmx *vmx) { - u32 host_sysenter_cs; + u32 host_sysenter_cs, msr_low, msr_high; u32 junk; + u64 host_pat; unsigned long a; struct descriptor_table dt; int i; @@ -2181,6 +2190,20 @@ static int vmx_vcpu_setup(struct vcpu_vmx *vmx) rdmsrl(MSR_IA32_SYSENTER_EIP, a); vmcs_writel(HOST_IA32_SYSENTER_EIP, a); /* 22.2.3 */ + if (vmcs_config.vmexit_ctrl & VM_EXIT_LOAD_IA32_PAT) { + rdmsr(MSR_IA32_CR_PAT, msr_low, msr_high); + host_pat = msr_low | ((u64) msr_high << 32); + vmcs_write64(HOST_IA32_PAT, host_pat); + } + if (vmcs_config.vmentry_ctrl & VM_ENTRY_LOAD_IA32_PAT) { + rdmsr(MSR_IA32_CR_PAT, msr_low, msr_high); + host_pat = msr_low | ((u64) msr_high << 32); + /* Write the default value follow host pat */ + vmcs_write64(GUEST_IA32_PAT, host_pat); + /* Keep arch.pat sync with GUEST_IA32_PAT */ + vmx->vcpu.arch.pat = host_pat; + } + for (i = 0; i < NR_VMX_MSR; ++i) { u32 index = vmx_msr_index[i]; u32 data_low, data_high; diff --git a/arch/x86/kvm/vmx.h b/arch/x86/kvm/vmx.h index ec5edc339da6..18598afe52eb 100644 --- a/arch/x86/kvm/vmx.h +++ b/arch/x86/kvm/vmx.h @@ -63,10 +63,13 @@ #define VM_EXIT_HOST_ADDR_SPACE_SIZE 0x00000200 #define VM_EXIT_ACK_INTR_ON_EXIT 0x00008000 +#define VM_EXIT_SAVE_IA32_PAT 0x00040000 +#define VM_EXIT_LOAD_IA32_PAT 0x00080000 #define VM_ENTRY_IA32E_MODE 0x00000200 #define VM_ENTRY_SMM 0x00000400 #define VM_ENTRY_DEACT_DUAL_MONITOR 0x00000800 +#define VM_ENTRY_LOAD_IA32_PAT 0x00004000 /* VMCS Encodings */ enum vmcs_field { @@ -112,6 +115,8 @@ enum vmcs_field { VMCS_LINK_POINTER_HIGH = 0x00002801, GUEST_IA32_DEBUGCTL = 0x00002802, GUEST_IA32_DEBUGCTL_HIGH = 0x00002803, + GUEST_IA32_PAT = 0x00002804, + GUEST_IA32_PAT_HIGH = 0x00002805, GUEST_PDPTR0 = 0x0000280a, GUEST_PDPTR0_HIGH = 0x0000280b, GUEST_PDPTR1 = 0x0000280c, @@ -120,6 +125,8 @@ enum vmcs_field { GUEST_PDPTR2_HIGH = 0x0000280f, GUEST_PDPTR3 = 0x00002810, GUEST_PDPTR3_HIGH = 0x00002811, + HOST_IA32_PAT = 0x00002c00, + HOST_IA32_PAT_HIGH = 0x00002c01, PIN_BASED_VM_EXEC_CONTROL = 0x00004000, CPU_BASED_VM_EXEC_CONTROL = 0x00004002, EXCEPTION_BITMAP = 0x00004004, diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index f5b2334c6bda..0edf75339f3a 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -452,7 +452,7 @@ static u32 msrs_to_save[] = { MSR_CSTAR, MSR_KERNEL_GS_BASE, MSR_SYSCALL_MASK, MSR_LSTAR, #endif MSR_IA32_TIME_STAMP_COUNTER, MSR_KVM_SYSTEM_TIME, MSR_KVM_WALL_CLOCK, - MSR_IA32_PERF_STATUS, + MSR_IA32_PERF_STATUS, MSR_IA32_CR_PAT }; static unsigned num_msrs_to_save; -- cgit v1.2.3 From 74be52e3e6285fc6e872a2a7baea544106f399ea Mon Sep 17 00:00:00 2001 From: Sheng Yang Date: Thu, 9 Oct 2008 16:01:56 +0800 Subject: KVM: Add local get_mtrr_type() to support MTRR For EPT memory type support. Signed-off-by: Sheng Yang Signed-off-by: Avi Kivity --- arch/x86/kvm/mmu.c | 104 +++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 104 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c index 410ddbc1aa2e..ac2304fd173e 100644 --- a/arch/x86/kvm/mmu.c +++ b/arch/x86/kvm/mmu.c @@ -1393,6 +1393,110 @@ struct page *gva_to_page(struct kvm_vcpu *vcpu, gva_t gva) return page; } +/* + * The function is based on mtrr_type_lookup() in + * arch/x86/kernel/cpu/mtrr/generic.c + */ +static int get_mtrr_type(struct mtrr_state_type *mtrr_state, + u64 start, u64 end) +{ + int i; + u64 base, mask; + u8 prev_match, curr_match; + int num_var_ranges = KVM_NR_VAR_MTRR; + + if (!mtrr_state->enabled) + return 0xFF; + + /* Make end inclusive end, instead of exclusive */ + end--; + + /* Look in fixed ranges. Just return the type as per start */ + if (mtrr_state->have_fixed && (start < 0x100000)) { + int idx; + + if (start < 0x80000) { + idx = 0; + idx += (start >> 16); + return mtrr_state->fixed_ranges[idx]; + } else if (start < 0xC0000) { + idx = 1 * 8; + idx += ((start - 0x80000) >> 14); + return mtrr_state->fixed_ranges[idx]; + } else if (start < 0x1000000) { + idx = 3 * 8; + idx += ((start - 0xC0000) >> 12); + return mtrr_state->fixed_ranges[idx]; + } + } + + /* + * Look in variable ranges + * Look of multiple ranges matching this address and pick type + * as per MTRR precedence + */ + if (!(mtrr_state->enabled & 2)) + return mtrr_state->def_type; + + prev_match = 0xFF; + for (i = 0; i < num_var_ranges; ++i) { + unsigned short start_state, end_state; + + if (!(mtrr_state->var_ranges[i].mask_lo & (1 << 11))) + continue; + + base = (((u64)mtrr_state->var_ranges[i].base_hi) << 32) + + (mtrr_state->var_ranges[i].base_lo & PAGE_MASK); + mask = (((u64)mtrr_state->var_ranges[i].mask_hi) << 32) + + (mtrr_state->var_ranges[i].mask_lo & PAGE_MASK); + + start_state = ((start & mask) == (base & mask)); + end_state = ((end & mask) == (base & mask)); + if (start_state != end_state) + return 0xFE; + + if ((start & mask) != (base & mask)) + continue; + + curr_match = mtrr_state->var_ranges[i].base_lo & 0xff; + if (prev_match == 0xFF) { + prev_match = curr_match; + continue; + } + + if (prev_match == MTRR_TYPE_UNCACHABLE || + curr_match == MTRR_TYPE_UNCACHABLE) + return MTRR_TYPE_UNCACHABLE; + + if ((prev_match == MTRR_TYPE_WRBACK && + curr_match == MTRR_TYPE_WRTHROUGH) || + (prev_match == MTRR_TYPE_WRTHROUGH && + curr_match == MTRR_TYPE_WRBACK)) { + prev_match = MTRR_TYPE_WRTHROUGH; + curr_match = MTRR_TYPE_WRTHROUGH; + } + + if (prev_match != curr_match) + return MTRR_TYPE_UNCACHABLE; + } + + if (prev_match != 0xFF) + return prev_match; + + return mtrr_state->def_type; +} + +static u8 get_memory_type(struct kvm_vcpu *vcpu, gfn_t gfn) +{ + u8 mtrr; + + mtrr = get_mtrr_type(&vcpu->arch.mtrr_state, gfn << PAGE_SHIFT, + (gfn << PAGE_SHIFT) + PAGE_SIZE); + if (mtrr == 0xfe || mtrr == 0xff) + mtrr = MTRR_TYPE_WRBACK; + return mtrr; +} + static int kvm_unsync_page(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp) { unsigned index; -- cgit v1.2.3 From 64d4d521757117aa5c1cfe79d3baa6cf57703f81 Mon Sep 17 00:00:00 2001 From: Sheng Yang Date: Thu, 9 Oct 2008 16:01:57 +0800 Subject: KVM: Enable MTRR for EPT The effective memory type of EPT is the mixture of MSR_IA32_CR_PAT and memory type field of EPT entry. Signed-off-by: Sheng Yang Signed-off-by: Avi Kivity --- arch/x86/include/asm/kvm_host.h | 3 ++- arch/x86/kvm/mmu.c | 11 ++++++++++- arch/x86/kvm/svm.c | 6 ++++++ arch/x86/kvm/vmx.c | 10 ++++++++-- arch/x86/kvm/x86.c | 2 +- 5 files changed, 27 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index 8082e87f628d..93040b5eed96 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -483,6 +483,7 @@ struct kvm_x86_ops { int (*set_tss_addr)(struct kvm *kvm, unsigned int addr); int (*get_tdp_level)(void); + int (*get_mt_mask_shift)(void); }; extern struct kvm_x86_ops *kvm_x86_ops; @@ -496,7 +497,7 @@ int kvm_mmu_setup(struct kvm_vcpu *vcpu); void kvm_mmu_set_nonpresent_ptes(u64 trap_pte, u64 notrap_pte); void kvm_mmu_set_base_ptes(u64 base_pte); void kvm_mmu_set_mask_ptes(u64 user_mask, u64 accessed_mask, - u64 dirty_mask, u64 nx_mask, u64 x_mask); + u64 dirty_mask, u64 nx_mask, u64 x_mask, u64 mt_mask); int kvm_mmu_reset_context(struct kvm_vcpu *vcpu); void kvm_mmu_slot_remove_write_access(struct kvm *kvm, int slot); diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c index ac2304fd173e..09d05f57bf66 100644 --- a/arch/x86/kvm/mmu.c +++ b/arch/x86/kvm/mmu.c @@ -168,6 +168,7 @@ static u64 __read_mostly shadow_x_mask; /* mutual exclusive with nx_mask */ static u64 __read_mostly shadow_user_mask; static u64 __read_mostly shadow_accessed_mask; static u64 __read_mostly shadow_dirty_mask; +static u64 __read_mostly shadow_mt_mask; void kvm_mmu_set_nonpresent_ptes(u64 trap_pte, u64 notrap_pte) { @@ -183,13 +184,14 @@ void kvm_mmu_set_base_ptes(u64 base_pte) EXPORT_SYMBOL_GPL(kvm_mmu_set_base_ptes); void kvm_mmu_set_mask_ptes(u64 user_mask, u64 accessed_mask, - u64 dirty_mask, u64 nx_mask, u64 x_mask) + u64 dirty_mask, u64 nx_mask, u64 x_mask, u64 mt_mask) { shadow_user_mask = user_mask; shadow_accessed_mask = accessed_mask; shadow_dirty_mask = dirty_mask; shadow_nx_mask = nx_mask; shadow_x_mask = x_mask; + shadow_mt_mask = mt_mask; } EXPORT_SYMBOL_GPL(kvm_mmu_set_mask_ptes); @@ -1546,6 +1548,8 @@ static int set_spte(struct kvm_vcpu *vcpu, u64 *shadow_pte, { u64 spte; int ret = 0; + u64 mt_mask = shadow_mt_mask; + /* * We don't set the accessed bit, since we sometimes want to see * whether the guest actually used the pte (in order to detect @@ -1564,6 +1568,11 @@ static int set_spte(struct kvm_vcpu *vcpu, u64 *shadow_pte, spte |= shadow_user_mask; if (largepage) spte |= PT_PAGE_SIZE_MASK; + if (mt_mask) { + mt_mask = get_memory_type(vcpu, gfn) << + kvm_x86_ops->get_mt_mask_shift(); + spte |= mt_mask; + } spte |= (u64)pfn << PAGE_SHIFT; diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c index 9c4ce657d963..05efc4ef75a6 100644 --- a/arch/x86/kvm/svm.c +++ b/arch/x86/kvm/svm.c @@ -1912,6 +1912,11 @@ static int get_npt_level(void) #endif } +static int svm_get_mt_mask_shift(void) +{ + return 0; +} + static struct kvm_x86_ops svm_x86_ops = { .cpu_has_kvm_support = has_svm, .disabled_by_bios = is_disabled, @@ -1967,6 +1972,7 @@ static struct kvm_x86_ops svm_x86_ops = { .set_tss_addr = svm_set_tss_addr, .get_tdp_level = get_npt_level, + .get_mt_mask_shift = svm_get_mt_mask_shift, }; static int __init svm_init(void) diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index b4c95a501cca..dae134fa09e7 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -3574,6 +3574,11 @@ static int get_ept_level(void) return VMX_EPT_DEFAULT_GAW + 1; } +static int vmx_get_mt_mask_shift(void) +{ + return VMX_EPT_MT_EPTE_SHIFT; +} + static struct kvm_x86_ops vmx_x86_ops = { .cpu_has_kvm_support = cpu_has_kvm_support, .disabled_by_bios = vmx_disabled_by_bios, @@ -3629,6 +3634,7 @@ static struct kvm_x86_ops vmx_x86_ops = { .set_tss_addr = vmx_set_tss_addr, .get_tdp_level = get_ept_level, + .get_mt_mask_shift = vmx_get_mt_mask_shift, }; static int __init vmx_init(void) @@ -3685,10 +3691,10 @@ static int __init vmx_init(void) bypass_guest_pf = 0; kvm_mmu_set_base_ptes(VMX_EPT_READABLE_MASK | VMX_EPT_WRITABLE_MASK | - VMX_EPT_DEFAULT_MT << VMX_EPT_MT_EPTE_SHIFT | VMX_EPT_IGMT_BIT); kvm_mmu_set_mask_ptes(0ull, 0ull, 0ull, 0ull, - VMX_EPT_EXECUTABLE_MASK); + VMX_EPT_EXECUTABLE_MASK, + VMX_EPT_DEFAULT_MT << VMX_EPT_MT_EPTE_SHIFT); kvm_enable_tdp(); } else kvm_disable_tdp(); diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 0edf75339f3a..f175b796c2a6 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -2615,7 +2615,7 @@ int kvm_arch_init(void *opaque) kvm_mmu_set_nonpresent_ptes(0ull, 0ull); kvm_mmu_set_base_ptes(PT_PRESENT_MASK); kvm_mmu_set_mask_ptes(PT_USER_MASK, PT_ACCESSED_MASK, - PT_DIRTY_MASK, PT64_NX_MASK, 0); + PT_DIRTY_MASK, PT64_NX_MASK, 0, 0); return 0; out: -- cgit v1.2.3 From d73fa29a9b75b2af7f69dae276d2c602a23b329b Mon Sep 17 00:00:00 2001 From: Sheng Yang Date: Tue, 14 Oct 2008 15:59:10 +0800 Subject: KVM: Clean up kvm_x86_emulate.h Remove one left improper comment of removed CR2. Signed-off-by: Sheng Yang Signed-off-by: Avi Kivity --- arch/x86/include/asm/kvm_x86_emulate.h | 10 +++------- 1 file changed, 3 insertions(+), 7 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/kvm_x86_emulate.h b/arch/x86/include/asm/kvm_x86_emulate.h index 25179a29f208..16a002655f31 100644 --- a/arch/x86/include/asm/kvm_x86_emulate.h +++ b/arch/x86/include/asm/kvm_x86_emulate.h @@ -146,22 +146,18 @@ struct x86_emulate_ctxt { /* Register state before/after emulation. */ struct kvm_vcpu *vcpu; - /* Linear faulting address (if emulating a page-faulting instruction) */ unsigned long eflags; - /* Emulated execution mode, represented by an X86EMUL_MODE value. */ int mode; - u32 cs_base; /* decode cache */ - struct decode_cache decode; }; /* Repeat String Operation Prefix */ -#define REPE_PREFIX 1 -#define REPNE_PREFIX 2 +#define REPE_PREFIX 1 +#define REPNE_PREFIX 2 /* Execution mode, passed to the emulator. */ #define X86EMUL_MODE_REAL 0 /* Real mode. */ @@ -170,7 +166,7 @@ struct x86_emulate_ctxt { #define X86EMUL_MODE_PROT64 8 /* 64-bit (long) mode. */ /* Host execution mode. */ -#if defined(__i386__) +#if defined(CONFIG_X86_32) #define X86EMUL_MODE_HOST X86EMUL_MODE_PROT32 #elif defined(CONFIG_X86_64) #define X86EMUL_MODE_HOST X86EMUL_MODE_PROT64 -- cgit v1.2.3 From 291f26bc0f89518ad7ee3207c09eb8a743ac8fcc Mon Sep 17 00:00:00 2001 From: Sheng Yang Date: Thu, 16 Oct 2008 17:30:57 +0800 Subject: KVM: MMU: Extend kvm_mmu_page->slot_bitmap size Otherwise set_bit() for private memory slot(above KVM_MEMORY_SLOTS) would corrupted memory in 32bit host. Signed-off-by: Sheng Yang Signed-off-by: Avi Kivity --- arch/x86/include/asm/kvm_host.h | 8 +++++--- arch/x86/kvm/mmu.c | 6 +++--- 2 files changed, 8 insertions(+), 6 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index 93040b5eed96..59c3ae10de6c 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -192,9 +192,11 @@ struct kvm_mmu_page { u64 *spt; /* hold the gfn of each spte inside spt */ gfn_t *gfns; - unsigned long slot_bitmap; /* One bit set per slot which has memory - * in this shadow page. - */ + /* + * One bit set per slot which has memory + * in this shadow page. + */ + DECLARE_BITMAP(slot_bitmap, KVM_MEMORY_SLOTS + KVM_PRIVATE_MEM_SLOTS); int multimapped; /* More than one parent_pte? */ int root_count; /* Currently serving as active root */ bool unsync; diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c index 09d05f57bf66..8687758b5295 100644 --- a/arch/x86/kvm/mmu.c +++ b/arch/x86/kvm/mmu.c @@ -789,7 +789,7 @@ static struct kvm_mmu_page *kvm_mmu_alloc_page(struct kvm_vcpu *vcpu, set_page_private(virt_to_page(sp->spt), (unsigned long)sp); list_add(&sp->link, &vcpu->kvm->arch.active_mmu_pages); ASSERT(is_empty_shadow_page(sp->spt)); - sp->slot_bitmap = 0; + bitmap_zero(sp->slot_bitmap, KVM_MEMORY_SLOTS + KVM_PRIVATE_MEM_SLOTS); sp->multimapped = 0; sp->parent_pte = parent_pte; --vcpu->kvm->arch.n_free_mmu_pages; @@ -1364,7 +1364,7 @@ static void page_header_update_slot(struct kvm *kvm, void *pte, gfn_t gfn) int slot = memslot_id(kvm, gfn_to_memslot(kvm, gfn)); struct kvm_mmu_page *sp = page_header(__pa(pte)); - __set_bit(slot, &sp->slot_bitmap); + __set_bit(slot, sp->slot_bitmap); } static void mmu_convert_notrap(struct kvm_mmu_page *sp) @@ -2564,7 +2564,7 @@ void kvm_mmu_slot_remove_write_access(struct kvm *kvm, int slot) int i; u64 *pt; - if (!test_bit(slot, &sp->slot_bitmap)) + if (!test_bit(slot, sp->slot_bitmap)) continue; pt = sp->spt; -- cgit v1.2.3 From 6fe639792c7b8e462baeaac39ecc33541fd5da6e Mon Sep 17 00:00:00 2001 From: Sheng Yang Date: Thu, 16 Oct 2008 17:30:58 +0800 Subject: KVM: VMX: Move private memory slot position PCI device assignment would map guest MMIO spaces as separate slot, so it is possible that the device has more than 2 MMIO spaces and overwrite current private memslot. The patch move private memory slot to the top of userspace visible memory slots. Signed-off-by: Sheng Yang Signed-off-by: Avi Kivity --- arch/x86/kvm/vmx.c | 2 +- arch/x86/kvm/vmx.h | 5 +++-- 2 files changed, 4 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index dae134fa09e7..7623eb7b68d5 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -2513,7 +2513,7 @@ static int vmx_set_tss_addr(struct kvm *kvm, unsigned int addr) { int ret; struct kvm_userspace_memory_region tss_mem = { - .slot = 8, + .slot = TSS_PRIVATE_MEMSLOT, .guest_phys_addr = addr, .memory_size = PAGE_SIZE * 3, .flags = 0, diff --git a/arch/x86/kvm/vmx.h b/arch/x86/kvm/vmx.h index 18598afe52eb..3db236c26fa4 100644 --- a/arch/x86/kvm/vmx.h +++ b/arch/x86/kvm/vmx.h @@ -338,8 +338,9 @@ enum vmcs_field { #define AR_RESERVD_MASK 0xfffe0f00 -#define APIC_ACCESS_PAGE_PRIVATE_MEMSLOT 9 -#define IDENTITY_PAGETABLE_PRIVATE_MEMSLOT 10 +#define TSS_PRIVATE_MEMSLOT (KVM_MEMORY_SLOTS + 0) +#define APIC_ACCESS_PAGE_PRIVATE_MEMSLOT (KVM_MEMORY_SLOTS + 1) +#define IDENTITY_PAGETABLE_PRIVATE_MEMSLOT (KVM_MEMORY_SLOTS + 2) #define VMX_NR_VPIDS (1 << 16) #define VMX_VPID_EXTENT_SINGLE_CONTEXT 1 -- cgit v1.2.3 From 291fd39bfc2089c2dae79cf2d7cfca81b14ca769 Mon Sep 17 00:00:00 2001 From: Guillaume Thouvenin Date: Mon, 20 Oct 2008 13:11:58 +0200 Subject: KVM: x86 emulator: Add decode entries for 0x04 and 0x05 opcodes (add acc, imm) Add decode entries for 0x04 and 0x05 (ADD) opcodes, execution is already implemented. Signed-off-by: Guillaume Thouvenin Signed-off-by: Avi Kivity --- arch/x86/kvm/x86_emulate.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/x86_emulate.c b/arch/x86/kvm/x86_emulate.c index a391e213fe61..57d7cc45be44 100644 --- a/arch/x86/kvm/x86_emulate.c +++ b/arch/x86/kvm/x86_emulate.c @@ -80,7 +80,7 @@ static u16 opcode_table[256] = { /* 0x00 - 0x07 */ ByteOp | DstMem | SrcReg | ModRM, DstMem | SrcReg | ModRM, ByteOp | DstReg | SrcMem | ModRM, DstReg | SrcMem | ModRM, - 0, 0, 0, 0, + ByteOp | DstAcc | SrcImm, DstAcc | SrcImm, 0, 0, /* 0x08 - 0x0F */ ByteOp | DstMem | SrcReg | ModRM, DstMem | SrcReg | ModRM, ByteOp | DstReg | SrcMem | ModRM, DstReg | SrcMem | ModRM, -- cgit v1.2.3 From 8fdb2351d51b040146f10a624387bbd102d851c0 Mon Sep 17 00:00:00 2001 From: Jan Kiszka Date: Mon, 20 Oct 2008 10:20:02 +0200 Subject: KVM: x86: Fix and refactor NMI watchdog emulation This patch refactors the NMI watchdog delivery patch, consolidating tests and providing a proper API for delivering watchdog events. An included micro-optimization is to check only for apic_hw_enabled in kvm_apic_local_deliver (the test for LVT mask is covering the soft-disabled case already). Signed-off-by: Jan Kiszka Acked-by: Sheng Yang Signed-off-by: Avi Kivity --- arch/x86/kvm/i8254.c | 15 +++++++++------ arch/x86/kvm/irq.h | 2 +- arch/x86/kvm/lapic.c | 20 ++++++++++---------- 3 files changed, 20 insertions(+), 17 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/i8254.c b/arch/x86/kvm/i8254.c index 580cc1d01c7d..b6fcf5a9e502 100644 --- a/arch/x86/kvm/i8254.c +++ b/arch/x86/kvm/i8254.c @@ -612,15 +612,18 @@ static void __inject_pit_timer_intr(struct kvm *kvm) mutex_unlock(&kvm->lock); /* - * Provides NMI watchdog support in IOAPIC mode. - * The route is: PIT -> PIC -> LVT0 in NMI mode, - * timer IRQs will continue to flow through the IOAPIC. + * Provides NMI watchdog support via Virtual Wire mode. + * The route is: PIT -> PIC -> LVT0 in NMI mode. + * + * Note: Our Virtual Wire implementation is simplified, only + * propagating PIT interrupts to all VCPUs when they have set + * LVT0 to NMI delivery. Other PIC interrupts are just sent to + * VCPU0, and only if its LVT0 is in EXTINT mode. */ for (i = 0; i < KVM_MAX_VCPUS; ++i) { vcpu = kvm->vcpus[i]; - if (!vcpu) - continue; - kvm_apic_local_deliver(vcpu, APIC_LVT0); + if (vcpu) + kvm_apic_nmi_wd_deliver(vcpu); } } diff --git a/arch/x86/kvm/irq.h b/arch/x86/kvm/irq.h index 71e37a530cf7..b9e9051650ea 100644 --- a/arch/x86/kvm/irq.h +++ b/arch/x86/kvm/irq.h @@ -87,7 +87,7 @@ void kvm_pic_reset(struct kvm_kpic_state *s); void kvm_timer_intr_post(struct kvm_vcpu *vcpu, int vec); void kvm_inject_pending_timer_irqs(struct kvm_vcpu *vcpu); void kvm_inject_apic_timer_irqs(struct kvm_vcpu *vcpu); -int kvm_apic_local_deliver(struct kvm_vcpu *vcpu, int lvt_type); +void kvm_apic_nmi_wd_deliver(struct kvm_vcpu *vcpu); void __kvm_migrate_apic_timer(struct kvm_vcpu *vcpu); void __kvm_migrate_pit_timer(struct kvm_vcpu *vcpu); void __kvm_migrate_timers(struct kvm_vcpu *vcpu); diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c index 304f9ddbdd51..0b0d413f0af5 100644 --- a/arch/x86/kvm/lapic.c +++ b/arch/x86/kvm/lapic.c @@ -973,14 +973,12 @@ int apic_has_pending_timer(struct kvm_vcpu *vcpu) return 0; } -int kvm_apic_local_deliver(struct kvm_vcpu *vcpu, int lvt_type) +static int kvm_apic_local_deliver(struct kvm_lapic *apic, int lvt_type) { - struct kvm_lapic *apic = vcpu->arch.apic; + u32 reg = apic_get_reg(apic, lvt_type); int vector, mode, trig_mode; - u32 reg; - if (apic && apic_enabled(apic)) { - reg = apic_get_reg(apic, lvt_type); + if (apic_hw_enabled(apic) && !(reg & APIC_LVT_MASKED)) { vector = reg & APIC_VECTOR_MASK; mode = reg & APIC_MODE_MASK; trig_mode = reg & APIC_LVT_LEVEL_TRIGGER; @@ -989,9 +987,12 @@ int kvm_apic_local_deliver(struct kvm_vcpu *vcpu, int lvt_type) return 0; } -static inline int __inject_apic_timer_irq(struct kvm_lapic *apic) +void kvm_apic_nmi_wd_deliver(struct kvm_vcpu *vcpu) { - return kvm_apic_local_deliver(apic->vcpu, APIC_LVTT); + struct kvm_lapic *apic = vcpu->arch.apic; + + if (apic) + kvm_apic_local_deliver(apic, APIC_LVT0); } static enum hrtimer_restart apic_timer_fn(struct hrtimer *data) @@ -1086,9 +1087,8 @@ void kvm_inject_apic_timer_irqs(struct kvm_vcpu *vcpu) { struct kvm_lapic *apic = vcpu->arch.apic; - if (apic && apic_lvt_enabled(apic, APIC_LVTT) && - atomic_read(&apic->timer.pending) > 0) { - if (__inject_apic_timer_irq(apic)) + if (apic && atomic_read(&apic->timer.pending) > 0) { + if (kvm_apic_local_deliver(apic, APIC_LVTT)) atomic_dec(&apic->timer.pending); } } -- cgit v1.2.3 From cc6e462cd54e64858ea25816df87d033229efe56 Mon Sep 17 00:00:00 2001 From: Jan Kiszka Date: Mon, 20 Oct 2008 10:20:03 +0200 Subject: KVM: x86: Optimize NMI watchdog delivery As suggested by Avi, this patch introduces a counter of VCPUs that have LVT0 set to NMI mode. Only if the counter > 0, we push the PIT ticks via all LAPIC LVT0 lines to enable NMI watchdog support. Signed-off-by: Jan Kiszka Acked-by: Sheng Yang Signed-off-by: Avi Kivity --- arch/x86/include/asm/kvm_host.h | 1 + arch/x86/kvm/i8254.c | 11 ++++++----- arch/x86/kvm/lapic.c | 23 ++++++++++++++++++++--- 3 files changed, 27 insertions(+), 8 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index 59c3ae10de6c..09e6c56572cb 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -361,6 +361,7 @@ struct kvm_arch{ struct kvm_ioapic *vioapic; struct kvm_pit *vpit; struct hlist_head irq_ack_notifier_list; + int vapics_in_nmi_mode; int round_robin_prev_vcpu; unsigned int tss_addr; diff --git a/arch/x86/kvm/i8254.c b/arch/x86/kvm/i8254.c index b6fcf5a9e502..e665d1c623ca 100644 --- a/arch/x86/kvm/i8254.c +++ b/arch/x86/kvm/i8254.c @@ -620,11 +620,12 @@ static void __inject_pit_timer_intr(struct kvm *kvm) * LVT0 to NMI delivery. Other PIC interrupts are just sent to * VCPU0, and only if its LVT0 is in EXTINT mode. */ - for (i = 0; i < KVM_MAX_VCPUS; ++i) { - vcpu = kvm->vcpus[i]; - if (vcpu) - kvm_apic_nmi_wd_deliver(vcpu); - } + if (kvm->arch.vapics_in_nmi_mode > 0) + for (i = 0; i < KVM_MAX_VCPUS; ++i) { + vcpu = kvm->vcpus[i]; + if (vcpu) + kvm_apic_nmi_wd_deliver(vcpu); + } } void kvm_inject_pit_timer_irqs(struct kvm_vcpu *vcpu) diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c index 0b0d413f0af5..afac68c0815c 100644 --- a/arch/x86/kvm/lapic.c +++ b/arch/x86/kvm/lapic.c @@ -130,6 +130,11 @@ static inline int apic_lvtt_period(struct kvm_lapic *apic) return apic_get_reg(apic, APIC_LVTT) & APIC_LVT_TIMER_PERIODIC; } +static inline int apic_lvt_nmi_mode(u32 lvt_val) +{ + return (lvt_val & (APIC_MODE_MASK | APIC_LVT_MASKED)) == APIC_DM_NMI; +} + static unsigned int apic_lvt_mask[APIC_LVT_NUM] = { LVT_MASK | APIC_LVT_TIMER_PERIODIC, /* LVTT */ LVT_MASK | APIC_MODE_MASK, /* LVTTHMR */ @@ -672,6 +677,20 @@ static void start_apic_timer(struct kvm_lapic *apic) apic->timer.period))); } +static void apic_manage_nmi_watchdog(struct kvm_lapic *apic, u32 lvt0_val) +{ + int nmi_wd_enabled = apic_lvt_nmi_mode(apic_get_reg(apic, APIC_LVT0)); + + if (apic_lvt_nmi_mode(lvt0_val)) { + if (!nmi_wd_enabled) { + apic_debug("Receive NMI setting on APIC_LVT0 " + "for cpu %d\n", apic->vcpu->vcpu_id); + apic->vcpu->kvm->arch.vapics_in_nmi_mode++; + } + } else if (nmi_wd_enabled) + apic->vcpu->kvm->arch.vapics_in_nmi_mode--; +} + static void apic_mmio_write(struct kvm_io_device *this, gpa_t address, int len, const void *data) { @@ -753,9 +772,7 @@ static void apic_mmio_write(struct kvm_io_device *this, break; case APIC_LVT0: - if (val == APIC_DM_NMI) - apic_debug("Receive NMI setting on APIC_LVT0 " - "for cpu %d\n", apic->vcpu->vcpu_id); + apic_manage_nmi_watchdog(apic, val); case APIC_LVTT: case APIC_LVTTHMR: case APIC_LVTPC: -- cgit v1.2.3 From b8222ad2e52fd2c0c4e5e1c53e65d131f911b767 Mon Sep 17 00:00:00 2001 From: Amit Shah Date: Wed, 22 Oct 2008 16:39:47 +0530 Subject: KVM: x86: Fix typo in function name get_segment_descritptor_dtable() contains an obvious type. Signed-off-by: Amit Shah Signed-off-by: Avi Kivity --- arch/x86/kvm/x86.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index f175b796c2a6..ceeac8897143 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -3373,9 +3373,9 @@ static void seg_desct_to_kvm_desct(struct desc_struct *seg_desc, u16 selector, kvm_desct->padding = 0; } -static void get_segment_descritptor_dtable(struct kvm_vcpu *vcpu, - u16 selector, - struct descriptor_table *dtable) +static void get_segment_descriptor_dtable(struct kvm_vcpu *vcpu, + u16 selector, + struct descriptor_table *dtable) { if (selector & 1 << 2) { struct kvm_segment kvm_seg; @@ -3400,7 +3400,7 @@ static int load_guest_segment_descriptor(struct kvm_vcpu *vcpu, u16 selector, struct descriptor_table dtable; u16 index = selector >> 3; - get_segment_descritptor_dtable(vcpu, selector, &dtable); + get_segment_descriptor_dtable(vcpu, selector, &dtable); if (dtable.limit < index * 8 + 7) { kvm_queue_exception_e(vcpu, GP_VECTOR, selector & 0xfffc); @@ -3419,7 +3419,7 @@ static int save_guest_segment_descriptor(struct kvm_vcpu *vcpu, u16 selector, struct descriptor_table dtable; u16 index = selector >> 3; - get_segment_descritptor_dtable(vcpu, selector, &dtable); + get_segment_descriptor_dtable(vcpu, selector, &dtable); if (dtable.limit < index * 8 + 7) return 1; -- cgit v1.2.3 From 25022acc3dd5f0b54071c7ba7c371860f2971b52 Mon Sep 17 00:00:00 2001 From: Amit Shah Date: Mon, 27 Oct 2008 09:04:17 +0000 Subject: KVM: SVM: Set the 'g' bit of the cs selector for cross-vendor migration The hardware does not set the 'g' bit of the cs selector and this breaks migration from amd hosts to intel hosts. Set this bit if the segment limit is beyond 1 MB. Signed-off-by: Amit Shah Signed-off-by: Avi Kivity --- arch/x86/kvm/svm.c | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c index 05efc4ef75a6..665008d97856 100644 --- a/arch/x86/kvm/svm.c +++ b/arch/x86/kvm/svm.c @@ -772,6 +772,15 @@ static void svm_get_segment(struct kvm_vcpu *vcpu, var->l = (s->attrib >> SVM_SELECTOR_L_SHIFT) & 1; var->db = (s->attrib >> SVM_SELECTOR_DB_SHIFT) & 1; var->g = (s->attrib >> SVM_SELECTOR_G_SHIFT) & 1; + + /* + * SVM always stores 0 for the 'G' bit in the CS selector in + * the VMCB on a VMEXIT. This hurts cross-vendor migration: + * Intel's VMENTRY has a check on the 'G' bit. + */ + if (seg == VCPU_SREG_CS) + var->g = s->limit > 0xfffff; + var->unusable = !var->present; } -- cgit v1.2.3 From c0d09828c870f90c6bc72070ada281568f89c63b Mon Sep 17 00:00:00 2001 From: Amit Shah Date: Mon, 27 Oct 2008 09:04:18 +0000 Subject: KVM: SVM: Set the 'busy' flag of the TR selector The busy flag of the TR selector is not set by the hardware. This breaks migration from amd hosts to intel hosts. Signed-off-by: Amit Shah Signed-off-by: Avi Kivity --- arch/x86/kvm/svm.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c index 665008d97856..743aebd7bfcc 100644 --- a/arch/x86/kvm/svm.c +++ b/arch/x86/kvm/svm.c @@ -781,6 +781,13 @@ static void svm_get_segment(struct kvm_vcpu *vcpu, if (seg == VCPU_SREG_CS) var->g = s->limit > 0xfffff; + /* + * Work around a bug where the busy flag in the tr selector + * isn't exposed + */ + if (seg == VCPU_SREG_TR) + var->type |= 0x2; + var->unusable = !var->present; } -- cgit v1.2.3 From e93f36bcfaa9e899c595e1c446c784a69021854a Mon Sep 17 00:00:00 2001 From: Guillaume Thouvenin Date: Tue, 28 Oct 2008 10:51:30 +0100 Subject: KVM: allow emulator to adjust rip for emulated pio instructions If we call the emulator we shouldn't call skip_emulated_instruction() in the first place, since the emulator already computes the next rip for us. Thus we move ->skip_emulated_instruction() out of kvm_emulate_pio() and into handle_io() (and the svm equivalent). We also replaced "return 0" by "break" in the "do_io:" case because now the shadow register state needs to be committed. Otherwise eip will never be updated. Signed-off-by: Guillaume Thouvenin Signed-off-by: Avi Kivity --- arch/x86/kvm/svm.c | 1 + arch/x86/kvm/vmx.c | 1 + arch/x86/kvm/x86.c | 2 -- arch/x86/kvm/x86_emulate.c | 2 +- 4 files changed, 3 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c index 743aebd7bfcc..f0ad4d4217e4 100644 --- a/arch/x86/kvm/svm.c +++ b/arch/x86/kvm/svm.c @@ -1115,6 +1115,7 @@ static int io_interception(struct vcpu_svm *svm, struct kvm_run *kvm_run) rep = (io_info & SVM_IOIO_REP_MASK) != 0; down = (svm->vmcb->save.rflags & X86_EFLAGS_DF) != 0; + skip_emulated_instruction(&svm->vcpu); return kvm_emulate_pio(&svm->vcpu, kvm_run, in, size, port); } diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index 7623eb7b68d5..816d23185fb8 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -2687,6 +2687,7 @@ static int handle_io(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run) rep = (exit_qualification & 32) != 0; port = exit_qualification >> 16; + skip_emulated_instruction(vcpu); return kvm_emulate_pio(vcpu, kvm_run, in, size, port); } diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index ceeac8897143..38f79b6aaf1e 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -2478,8 +2478,6 @@ int kvm_emulate_pio(struct kvm_vcpu *vcpu, struct kvm_run *run, int in, val = kvm_register_read(vcpu, VCPU_REGS_RAX); memcpy(vcpu->arch.pio_data, &val, 4); - kvm_x86_ops->skip_emulated_instruction(vcpu); - pio_dev = vcpu_find_pio_dev(vcpu, port, size, !in); if (pio_dev) { kernel_pio(pio_dev, vcpu, vcpu->arch.pio_data); diff --git a/arch/x86/kvm/x86_emulate.c b/arch/x86/kvm/x86_emulate.c index 57d7cc45be44..8f60ace13874 100644 --- a/arch/x86/kvm/x86_emulate.c +++ b/arch/x86/kvm/x86_emulate.c @@ -1772,7 +1772,7 @@ special_insn: c->eip = saved_eip; goto cannot_emulate; } - return 0; + break; case 0xf4: /* hlt */ ctxt->vcpu->arch.halt_request = 1; break; -- cgit v1.2.3 From 1d5a4d9b92028d9fe77da34037bd5a1ebfecc733 Mon Sep 17 00:00:00 2001 From: Guillaume Thouvenin Date: Wed, 29 Oct 2008 09:39:42 +0100 Subject: KVM: VMX: Handle mmio emulation when guest state is invalid If emulate_invalid_guest_state is enabled, the emulator is called when guest state is invalid. Until now, we reported an mmio failure when emulate_instruction() returned EMULATE_DO_MMIO. This patch adds the case where emulate_instruction() failed and an MMIO emulation is needed. Signed-off-by: Guillaume Thouvenin Signed-off-by: Avi Kivity --- arch/x86/kvm/vmx.c | 27 +++++++++++++++------------ 1 file changed, 15 insertions(+), 12 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index 816d23185fb8..427dbc14fae9 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -3052,16 +3052,12 @@ static void handle_invalid_guest_state(struct kvm_vcpu *vcpu, while (!guest_state_valid(vcpu)) { err = emulate_instruction(vcpu, kvm_run, 0, 0, 0); - switch (err) { - case EMULATE_DONE: - break; - case EMULATE_DO_MMIO: - kvm_report_emulation_failure(vcpu, "mmio"); - /* TODO: Handle MMIO */ - return; - default: - kvm_report_emulation_failure(vcpu, "emulation failure"); - return; + if (err == EMULATE_DO_MMIO) + break; + + if (err != EMULATE_DONE) { + kvm_report_emulation_failure(vcpu, "emulation failure"); + return; } if (signal_pending(current)) @@ -3073,8 +3069,10 @@ static void handle_invalid_guest_state(struct kvm_vcpu *vcpu, local_irq_disable(); preempt_disable(); - /* Guest state should be valid now, no more emulation should be needed */ - vmx->emulation_required = 0; + /* Guest state should be valid now except if we need to + * emulate an MMIO */ + if (guest_state_valid(vcpu)) + vmx->emulation_required = 0; } /* @@ -3121,6 +3119,11 @@ static int kvm_handle_exit(struct kvm_run *kvm_run, struct kvm_vcpu *vcpu) KVMTRACE_3D(VMEXIT, vcpu, exit_reason, (u32)kvm_rip_read(vcpu), (u32)((u64)kvm_rip_read(vcpu) >> 32), entryexit); + /* If we need to emulate an MMIO from handle_invalid_guest_state + * we just return 0 */ + if (vmx->emulation_required && emulate_invalid_guest_state) + return 0; + /* Access CR3 don't cause VMExit in paging mode, so we need * to sync with guest real CR3. */ if (vm_need_ept() && is_paging(vcpu)) { -- cgit v1.2.3 From 6eb55818c043b097c83828da8430fcb9a02fdb89 Mon Sep 17 00:00:00 2001 From: Sheng Yang Date: Fri, 31 Oct 2008 12:37:41 +0800 Subject: KVM: Enable Function Level Reset for assigned device Ideally, every assigned device should in a clear condition before and after assignment, so that the former state of device won't affect later work. Some devices provide a mechanism named Function Level Reset, which is defined in PCI/PCI-e document. We should execute it before and after device assignment. (But sadly, the feature is new, and most device on the market now don't support it. We are considering using D0/D3hot transmit to emulate it later, but not that elegant and reliable as FLR itself.) [Update: Reminded by Xiantao, execute FLR after we ensure that the device can be assigned to the guest.] Signed-off-by: Sheng Yang Signed-off-by: Avi Kivity --- arch/x86/kvm/x86.c | 2 +- virt/kvm/kvm_main.c | 5 +++++ 2 files changed, 6 insertions(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 38f79b6aaf1e..9a4a39cfe6ef 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -4148,8 +4148,8 @@ static void kvm_free_vcpus(struct kvm *kvm) void kvm_arch_destroy_vm(struct kvm *kvm) { - kvm_iommu_unmap_guest(kvm); kvm_free_all_assigned_devices(kvm); + kvm_iommu_unmap_guest(kvm); kvm_free_pit(kvm); kfree(kvm->arch.vpic); kfree(kvm->arch.vioapic); diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c index 4f43abe198e4..1838052f3c9e 100644 --- a/virt/kvm/kvm_main.c +++ b/virt/kvm/kvm_main.c @@ -152,6 +152,8 @@ static void kvm_free_assigned_device(struct kvm *kvm, */ kvm_put_kvm(kvm); + pci_reset_function(assigned_dev->dev); + pci_release_regions(assigned_dev->dev); pci_disable_device(assigned_dev->dev); pci_dev_put(assigned_dev->dev); @@ -283,6 +285,9 @@ static int kvm_vm_ioctl_assign_device(struct kvm *kvm, __func__); goto out_disable; } + + pci_reset_function(dev); + match->assigned_dev_id = assigned_dev->assigned_dev_id; match->host_busnr = assigned_dev->busnr; match->host_devfn = assigned_dev->devfn; -- cgit v1.2.3 From 2843099fee32a6020e1caa95c6026f28b5d43bff Mon Sep 17 00:00:00 2001 From: Izik Eidus Date: Fri, 3 Oct 2008 17:40:32 +0300 Subject: KVM: MMU: Fix aliased gfns treated as unaliased Some areas of kvm x86 mmu are using gfn offset inside a slot without unaliasing the gfn first. This patch makes sure that the gfn will be unaliased and add gfn_to_memslot_unaliased() to save the calculating of the gfn unaliasing in case we have it unaliased already. Signed-off-by: Izik Eidus Acked-by: Marcelo Tosatti Signed-off-by: Avi Kivity --- arch/x86/include/asm/kvm_host.h | 2 ++ arch/x86/kvm/mmu.c | 14 ++++++++++---- virt/kvm/kvm_main.c | 9 +++++---- 3 files changed, 17 insertions(+), 8 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index 09e6c56572cb..99e3cc149d21 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -617,6 +617,8 @@ void kvm_disable_tdp(void); int load_pdptrs(struct kvm_vcpu *vcpu, unsigned long cr3); int complete_pio(struct kvm_vcpu *vcpu); +struct kvm_memory_slot *gfn_to_memslot_unaliased(struct kvm *kvm, gfn_t gfn); + static inline struct kvm_mmu_page *page_header(hpa_t shadow_page) { struct page *page = pfn_to_page(shadow_page >> PAGE_SHIFT); diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c index 8687758b5295..8904e8ada978 100644 --- a/arch/x86/kvm/mmu.c +++ b/arch/x86/kvm/mmu.c @@ -386,7 +386,9 @@ static void account_shadowed(struct kvm *kvm, gfn_t gfn) { int *write_count; - write_count = slot_largepage_idx(gfn, gfn_to_memslot(kvm, gfn)); + gfn = unalias_gfn(kvm, gfn); + write_count = slot_largepage_idx(gfn, + gfn_to_memslot_unaliased(kvm, gfn)); *write_count += 1; } @@ -394,16 +396,20 @@ static void unaccount_shadowed(struct kvm *kvm, gfn_t gfn) { int *write_count; - write_count = slot_largepage_idx(gfn, gfn_to_memslot(kvm, gfn)); + gfn = unalias_gfn(kvm, gfn); + write_count = slot_largepage_idx(gfn, + gfn_to_memslot_unaliased(kvm, gfn)); *write_count -= 1; WARN_ON(*write_count < 0); } static int has_wrprotected_page(struct kvm *kvm, gfn_t gfn) { - struct kvm_memory_slot *slot = gfn_to_memslot(kvm, gfn); + struct kvm_memory_slot *slot; int *largepage_idx; + gfn = unalias_gfn(kvm, gfn); + slot = gfn_to_memslot_unaliased(kvm, gfn); if (slot) { largepage_idx = slot_largepage_idx(gfn, slot); return *largepage_idx; @@ -2973,8 +2979,8 @@ static void audit_write_protection(struct kvm_vcpu *vcpu) if (sp->role.metaphysical) continue; - slot = gfn_to_memslot(vcpu->kvm, sp->gfn); gfn = unalias_gfn(vcpu->kvm, sp->gfn); + slot = gfn_to_memslot_unaliased(vcpu->kvm, sp->gfn); rmapp = &slot->rmap[gfn - slot->base_gfn]; if (*rmapp) printk(KERN_ERR "%s: (%s) shadow page has writable" diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c index 1838052f3c9e..a65baa9039d5 100644 --- a/virt/kvm/kvm_main.c +++ b/virt/kvm/kvm_main.c @@ -923,7 +923,7 @@ int kvm_is_error_hva(unsigned long addr) } EXPORT_SYMBOL_GPL(kvm_is_error_hva); -static struct kvm_memory_slot *__gfn_to_memslot(struct kvm *kvm, gfn_t gfn) +struct kvm_memory_slot *gfn_to_memslot_unaliased(struct kvm *kvm, gfn_t gfn) { int i; @@ -936,11 +936,12 @@ static struct kvm_memory_slot *__gfn_to_memslot(struct kvm *kvm, gfn_t gfn) } return NULL; } +EXPORT_SYMBOL_GPL(gfn_to_memslot_unaliased); struct kvm_memory_slot *gfn_to_memslot(struct kvm *kvm, gfn_t gfn) { gfn = unalias_gfn(kvm, gfn); - return __gfn_to_memslot(kvm, gfn); + return gfn_to_memslot_unaliased(kvm, gfn); } int kvm_is_visible_gfn(struct kvm *kvm, gfn_t gfn) @@ -964,7 +965,7 @@ unsigned long gfn_to_hva(struct kvm *kvm, gfn_t gfn) struct kvm_memory_slot *slot; gfn = unalias_gfn(kvm, gfn); - slot = __gfn_to_memslot(kvm, gfn); + slot = gfn_to_memslot_unaliased(kvm, gfn); if (!slot) return bad_hva(); return (slot->userspace_addr + (gfn - slot->base_gfn) * PAGE_SIZE); @@ -1215,7 +1216,7 @@ void mark_page_dirty(struct kvm *kvm, gfn_t gfn) struct kvm_memory_slot *memslot; gfn = unalias_gfn(kvm, gfn); - memslot = __gfn_to_memslot(kvm, gfn); + memslot = gfn_to_memslot_unaliased(kvm, gfn); if (memslot && memslot->dirty_bitmap) { unsigned long rel_gfn = gfn - memslot->base_gfn; -- cgit v1.2.3 From 0853d2c1d849ef69884d2447d90d04007590b72b Mon Sep 17 00:00:00 2001 From: Nitin A Kamble Date: Wed, 5 Nov 2008 15:37:36 -0800 Subject: KVM: Fix cpuid leaf 0xb loop termination For cpuid leaf 0xb the bits 8-15 in ECX register define the end of counting leaf. The previous code was using bits 0-7 for this purpose, which is a bug. Signed-off-by: Nitin A Kamble Signed-off-by: Avi Kivity --- arch/x86/kvm/x86.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 9a4a39cfe6ef..2889a0f359ea 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -1276,7 +1276,7 @@ static void do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function, entry->flags |= KVM_CPUID_FLAG_SIGNIFCANT_INDEX; /* read more entries until level_type is zero */ for (i = 1; *nent < maxnent; ++i) { - level_type = entry[i - 1].ecx & 0xff; + level_type = entry[i - 1].ecx & 0xff00; if (!level_type) break; do_cpuid_1_ent(&entry[i], function, i); -- cgit v1.2.3 From 0fdf8e59faa5c60e9d77c8e14abe3a0f8bfcf586 Mon Sep 17 00:00:00 2001 From: Nitin A Kamble Date: Wed, 5 Nov 2008 15:56:21 -0800 Subject: KVM: Fix cpuid iteration on multiple leaves per eac The code to traverse the cpuid data array list for counting type of leaves is currently broken. This patches fixes the 2 things in it. 1. Set the 1st counting entry's flag KVM_CPUID_FLAG_STATE_READ_NEXT. Without it the code will never find a valid entry. 2. Also the stop condition in the for loop while looking for the next unflaged entry is broken. It needs to stop when it find one matching entry; and in the case of count of 1, it will be the same entry found in this iteration. Signed-Off-By: Nitin A Kamble Signed-off-by: Avi Kivity --- arch/x86/kvm/x86.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 2889a0f359ea..7a2aeba0bfbd 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -1246,6 +1246,7 @@ static void do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function, int t, times = entry->eax & 0xff; entry->flags |= KVM_CPUID_FLAG_STATEFUL_FUNC; + entry->flags |= KVM_CPUID_FLAG_STATE_READ_NEXT; for (t = 1; t < times && *nent < maxnent; ++t) { do_cpuid_1_ent(&entry[t], function, 0); entry[t].flags |= KVM_CPUID_FLAG_STATEFUL_FUNC; @@ -2801,7 +2802,7 @@ static int move_to_next_stateful_cpuid_entry(struct kvm_vcpu *vcpu, int i) e->flags &= ~KVM_CPUID_FLAG_STATE_READ_NEXT; /* when no next entry is found, the current entry[i] is reselected */ - for (j = i + 1; j == i; j = (j + 1) % nent) { + for (j = i + 1; ; j = (j + 1) % nent) { struct kvm_cpuid_entry2 *ej = &vcpu->arch.cpuid_entries[j]; if (ej->function == e->function) { ej->flags |= KVM_CPUID_FLAG_STATE_READ_NEXT; -- cgit v1.2.3 From 13673a90f1cf88296f726265cc7cf3ec76ecba30 Mon Sep 17 00:00:00 2001 From: Eduardo Habkost Date: Mon, 17 Nov 2008 19:03:13 -0200 Subject: KVM: VMX: move vmx.h to include/asm vmx.h will be used by core code that is independent of KVM, so I am moving it outside the arch/x86/kvm directory. Signed-off-by: Eduardo Habkost Signed-off-by: Avi Kivity --- arch/x86/include/asm/vmx.h | 367 +++++++++++++++++++++++++++++++++++++++++++++ arch/x86/kvm/mmu.c | 2 +- arch/x86/kvm/vmx.c | 2 +- arch/x86/kvm/vmx.h | 367 --------------------------------------------- 4 files changed, 369 insertions(+), 369 deletions(-) create mode 100644 arch/x86/include/asm/vmx.h delete mode 100644 arch/x86/kvm/vmx.h (limited to 'arch/x86') diff --git a/arch/x86/include/asm/vmx.h b/arch/x86/include/asm/vmx.h new file mode 100644 index 000000000000..3db236c26fa4 --- /dev/null +++ b/arch/x86/include/asm/vmx.h @@ -0,0 +1,367 @@ +#ifndef VMX_H +#define VMX_H + +/* + * vmx.h: VMX Architecture related definitions + * Copyright (c) 2004, Intel Corporation. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., 59 Temple + * Place - Suite 330, Boston, MA 02111-1307 USA. + * + * A few random additions are: + * Copyright (C) 2006 Qumranet + * Avi Kivity + * Yaniv Kamay + * + */ + +/* + * Definitions of Primary Processor-Based VM-Execution Controls. + */ +#define CPU_BASED_VIRTUAL_INTR_PENDING 0x00000004 +#define CPU_BASED_USE_TSC_OFFSETING 0x00000008 +#define CPU_BASED_HLT_EXITING 0x00000080 +#define CPU_BASED_INVLPG_EXITING 0x00000200 +#define CPU_BASED_MWAIT_EXITING 0x00000400 +#define CPU_BASED_RDPMC_EXITING 0x00000800 +#define CPU_BASED_RDTSC_EXITING 0x00001000 +#define CPU_BASED_CR3_LOAD_EXITING 0x00008000 +#define CPU_BASED_CR3_STORE_EXITING 0x00010000 +#define CPU_BASED_CR8_LOAD_EXITING 0x00080000 +#define CPU_BASED_CR8_STORE_EXITING 0x00100000 +#define CPU_BASED_TPR_SHADOW 0x00200000 +#define CPU_BASED_VIRTUAL_NMI_PENDING 0x00400000 +#define CPU_BASED_MOV_DR_EXITING 0x00800000 +#define CPU_BASED_UNCOND_IO_EXITING 0x01000000 +#define CPU_BASED_USE_IO_BITMAPS 0x02000000 +#define CPU_BASED_USE_MSR_BITMAPS 0x10000000 +#define CPU_BASED_MONITOR_EXITING 0x20000000 +#define CPU_BASED_PAUSE_EXITING 0x40000000 +#define CPU_BASED_ACTIVATE_SECONDARY_CONTROLS 0x80000000 +/* + * Definitions of Secondary Processor-Based VM-Execution Controls. + */ +#define SECONDARY_EXEC_VIRTUALIZE_APIC_ACCESSES 0x00000001 +#define SECONDARY_EXEC_ENABLE_EPT 0x00000002 +#define SECONDARY_EXEC_ENABLE_VPID 0x00000020 +#define SECONDARY_EXEC_WBINVD_EXITING 0x00000040 + + +#define PIN_BASED_EXT_INTR_MASK 0x00000001 +#define PIN_BASED_NMI_EXITING 0x00000008 +#define PIN_BASED_VIRTUAL_NMIS 0x00000020 + +#define VM_EXIT_HOST_ADDR_SPACE_SIZE 0x00000200 +#define VM_EXIT_ACK_INTR_ON_EXIT 0x00008000 +#define VM_EXIT_SAVE_IA32_PAT 0x00040000 +#define VM_EXIT_LOAD_IA32_PAT 0x00080000 + +#define VM_ENTRY_IA32E_MODE 0x00000200 +#define VM_ENTRY_SMM 0x00000400 +#define VM_ENTRY_DEACT_DUAL_MONITOR 0x00000800 +#define VM_ENTRY_LOAD_IA32_PAT 0x00004000 + +/* VMCS Encodings */ +enum vmcs_field { + VIRTUAL_PROCESSOR_ID = 0x00000000, + GUEST_ES_SELECTOR = 0x00000800, + GUEST_CS_SELECTOR = 0x00000802, + GUEST_SS_SELECTOR = 0x00000804, + GUEST_DS_SELECTOR = 0x00000806, + GUEST_FS_SELECTOR = 0x00000808, + GUEST_GS_SELECTOR = 0x0000080a, + GUEST_LDTR_SELECTOR = 0x0000080c, + GUEST_TR_SELECTOR = 0x0000080e, + HOST_ES_SELECTOR = 0x00000c00, + HOST_CS_SELECTOR = 0x00000c02, + HOST_SS_SELECTOR = 0x00000c04, + HOST_DS_SELECTOR = 0x00000c06, + HOST_FS_SELECTOR = 0x00000c08, + HOST_GS_SELECTOR = 0x00000c0a, + HOST_TR_SELECTOR = 0x00000c0c, + IO_BITMAP_A = 0x00002000, + IO_BITMAP_A_HIGH = 0x00002001, + IO_BITMAP_B = 0x00002002, + IO_BITMAP_B_HIGH = 0x00002003, + MSR_BITMAP = 0x00002004, + MSR_BITMAP_HIGH = 0x00002005, + VM_EXIT_MSR_STORE_ADDR = 0x00002006, + VM_EXIT_MSR_STORE_ADDR_HIGH = 0x00002007, + VM_EXIT_MSR_LOAD_ADDR = 0x00002008, + VM_EXIT_MSR_LOAD_ADDR_HIGH = 0x00002009, + VM_ENTRY_MSR_LOAD_ADDR = 0x0000200a, + VM_ENTRY_MSR_LOAD_ADDR_HIGH = 0x0000200b, + TSC_OFFSET = 0x00002010, + TSC_OFFSET_HIGH = 0x00002011, + VIRTUAL_APIC_PAGE_ADDR = 0x00002012, + VIRTUAL_APIC_PAGE_ADDR_HIGH = 0x00002013, + APIC_ACCESS_ADDR = 0x00002014, + APIC_ACCESS_ADDR_HIGH = 0x00002015, + EPT_POINTER = 0x0000201a, + EPT_POINTER_HIGH = 0x0000201b, + GUEST_PHYSICAL_ADDRESS = 0x00002400, + GUEST_PHYSICAL_ADDRESS_HIGH = 0x00002401, + VMCS_LINK_POINTER = 0x00002800, + VMCS_LINK_POINTER_HIGH = 0x00002801, + GUEST_IA32_DEBUGCTL = 0x00002802, + GUEST_IA32_DEBUGCTL_HIGH = 0x00002803, + GUEST_IA32_PAT = 0x00002804, + GUEST_IA32_PAT_HIGH = 0x00002805, + GUEST_PDPTR0 = 0x0000280a, + GUEST_PDPTR0_HIGH = 0x0000280b, + GUEST_PDPTR1 = 0x0000280c, + GUEST_PDPTR1_HIGH = 0x0000280d, + GUEST_PDPTR2 = 0x0000280e, + GUEST_PDPTR2_HIGH = 0x0000280f, + GUEST_PDPTR3 = 0x00002810, + GUEST_PDPTR3_HIGH = 0x00002811, + HOST_IA32_PAT = 0x00002c00, + HOST_IA32_PAT_HIGH = 0x00002c01, + PIN_BASED_VM_EXEC_CONTROL = 0x00004000, + CPU_BASED_VM_EXEC_CONTROL = 0x00004002, + EXCEPTION_BITMAP = 0x00004004, + PAGE_FAULT_ERROR_CODE_MASK = 0x00004006, + PAGE_FAULT_ERROR_CODE_MATCH = 0x00004008, + CR3_TARGET_COUNT = 0x0000400a, + VM_EXIT_CONTROLS = 0x0000400c, + VM_EXIT_MSR_STORE_COUNT = 0x0000400e, + VM_EXIT_MSR_LOAD_COUNT = 0x00004010, + VM_ENTRY_CONTROLS = 0x00004012, + VM_ENTRY_MSR_LOAD_COUNT = 0x00004014, + VM_ENTRY_INTR_INFO_FIELD = 0x00004016, + VM_ENTRY_EXCEPTION_ERROR_CODE = 0x00004018, + VM_ENTRY_INSTRUCTION_LEN = 0x0000401a, + TPR_THRESHOLD = 0x0000401c, + SECONDARY_VM_EXEC_CONTROL = 0x0000401e, + VM_INSTRUCTION_ERROR = 0x00004400, + VM_EXIT_REASON = 0x00004402, + VM_EXIT_INTR_INFO = 0x00004404, + VM_EXIT_INTR_ERROR_CODE = 0x00004406, + IDT_VECTORING_INFO_FIELD = 0x00004408, + IDT_VECTORING_ERROR_CODE = 0x0000440a, + VM_EXIT_INSTRUCTION_LEN = 0x0000440c, + VMX_INSTRUCTION_INFO = 0x0000440e, + GUEST_ES_LIMIT = 0x00004800, + GUEST_CS_LIMIT = 0x00004802, + GUEST_SS_LIMIT = 0x00004804, + GUEST_DS_LIMIT = 0x00004806, + GUEST_FS_LIMIT = 0x00004808, + GUEST_GS_LIMIT = 0x0000480a, + GUEST_LDTR_LIMIT = 0x0000480c, + GUEST_TR_LIMIT = 0x0000480e, + GUEST_GDTR_LIMIT = 0x00004810, + GUEST_IDTR_LIMIT = 0x00004812, + GUEST_ES_AR_BYTES = 0x00004814, + GUEST_CS_AR_BYTES = 0x00004816, + GUEST_SS_AR_BYTES = 0x00004818, + GUEST_DS_AR_BYTES = 0x0000481a, + GUEST_FS_AR_BYTES = 0x0000481c, + GUEST_GS_AR_BYTES = 0x0000481e, + GUEST_LDTR_AR_BYTES = 0x00004820, + GUEST_TR_AR_BYTES = 0x00004822, + GUEST_INTERRUPTIBILITY_INFO = 0x00004824, + GUEST_ACTIVITY_STATE = 0X00004826, + GUEST_SYSENTER_CS = 0x0000482A, + HOST_IA32_SYSENTER_CS = 0x00004c00, + CR0_GUEST_HOST_MASK = 0x00006000, + CR4_GUEST_HOST_MASK = 0x00006002, + CR0_READ_SHADOW = 0x00006004, + CR4_READ_SHADOW = 0x00006006, + CR3_TARGET_VALUE0 = 0x00006008, + CR3_TARGET_VALUE1 = 0x0000600a, + CR3_TARGET_VALUE2 = 0x0000600c, + CR3_TARGET_VALUE3 = 0x0000600e, + EXIT_QUALIFICATION = 0x00006400, + GUEST_LINEAR_ADDRESS = 0x0000640a, + GUEST_CR0 = 0x00006800, + GUEST_CR3 = 0x00006802, + GUEST_CR4 = 0x00006804, + GUEST_ES_BASE = 0x00006806, + GUEST_CS_BASE = 0x00006808, + GUEST_SS_BASE = 0x0000680a, + GUEST_DS_BASE = 0x0000680c, + GUEST_FS_BASE = 0x0000680e, + GUEST_GS_BASE = 0x00006810, + GUEST_LDTR_BASE = 0x00006812, + GUEST_TR_BASE = 0x00006814, + GUEST_GDTR_BASE = 0x00006816, + GUEST_IDTR_BASE = 0x00006818, + GUEST_DR7 = 0x0000681a, + GUEST_RSP = 0x0000681c, + GUEST_RIP = 0x0000681e, + GUEST_RFLAGS = 0x00006820, + GUEST_PENDING_DBG_EXCEPTIONS = 0x00006822, + GUEST_SYSENTER_ESP = 0x00006824, + GUEST_SYSENTER_EIP = 0x00006826, + HOST_CR0 = 0x00006c00, + HOST_CR3 = 0x00006c02, + HOST_CR4 = 0x00006c04, + HOST_FS_BASE = 0x00006c06, + HOST_GS_BASE = 0x00006c08, + HOST_TR_BASE = 0x00006c0a, + HOST_GDTR_BASE = 0x00006c0c, + HOST_IDTR_BASE = 0x00006c0e, + HOST_IA32_SYSENTER_ESP = 0x00006c10, + HOST_IA32_SYSENTER_EIP = 0x00006c12, + HOST_RSP = 0x00006c14, + HOST_RIP = 0x00006c16, +}; + +#define VMX_EXIT_REASONS_FAILED_VMENTRY 0x80000000 + +#define EXIT_REASON_EXCEPTION_NMI 0 +#define EXIT_REASON_EXTERNAL_INTERRUPT 1 +#define EXIT_REASON_TRIPLE_FAULT 2 + +#define EXIT_REASON_PENDING_INTERRUPT 7 +#define EXIT_REASON_NMI_WINDOW 8 +#define EXIT_REASON_TASK_SWITCH 9 +#define EXIT_REASON_CPUID 10 +#define EXIT_REASON_HLT 12 +#define EXIT_REASON_INVLPG 14 +#define EXIT_REASON_RDPMC 15 +#define EXIT_REASON_RDTSC 16 +#define EXIT_REASON_VMCALL 18 +#define EXIT_REASON_VMCLEAR 19 +#define EXIT_REASON_VMLAUNCH 20 +#define EXIT_REASON_VMPTRLD 21 +#define EXIT_REASON_VMPTRST 22 +#define EXIT_REASON_VMREAD 23 +#define EXIT_REASON_VMRESUME 24 +#define EXIT_REASON_VMWRITE 25 +#define EXIT_REASON_VMOFF 26 +#define EXIT_REASON_VMON 27 +#define EXIT_REASON_CR_ACCESS 28 +#define EXIT_REASON_DR_ACCESS 29 +#define EXIT_REASON_IO_INSTRUCTION 30 +#define EXIT_REASON_MSR_READ 31 +#define EXIT_REASON_MSR_WRITE 32 +#define EXIT_REASON_MWAIT_INSTRUCTION 36 +#define EXIT_REASON_TPR_BELOW_THRESHOLD 43 +#define EXIT_REASON_APIC_ACCESS 44 +#define EXIT_REASON_EPT_VIOLATION 48 +#define EXIT_REASON_EPT_MISCONFIG 49 +#define EXIT_REASON_WBINVD 54 + +/* + * Interruption-information format + */ +#define INTR_INFO_VECTOR_MASK 0xff /* 7:0 */ +#define INTR_INFO_INTR_TYPE_MASK 0x700 /* 10:8 */ +#define INTR_INFO_DELIVER_CODE_MASK 0x800 /* 11 */ +#define INTR_INFO_UNBLOCK_NMI 0x1000 /* 12 */ +#define INTR_INFO_VALID_MASK 0x80000000 /* 31 */ +#define INTR_INFO_RESVD_BITS_MASK 0x7ffff000 + +#define VECTORING_INFO_VECTOR_MASK INTR_INFO_VECTOR_MASK +#define VECTORING_INFO_TYPE_MASK INTR_INFO_INTR_TYPE_MASK +#define VECTORING_INFO_DELIVER_CODE_MASK INTR_INFO_DELIVER_CODE_MASK +#define VECTORING_INFO_VALID_MASK INTR_INFO_VALID_MASK + +#define INTR_TYPE_EXT_INTR (0 << 8) /* external interrupt */ +#define INTR_TYPE_NMI_INTR (2 << 8) /* NMI */ +#define INTR_TYPE_EXCEPTION (3 << 8) /* processor exception */ +#define INTR_TYPE_SOFT_INTR (4 << 8) /* software interrupt */ + +/* GUEST_INTERRUPTIBILITY_INFO flags. */ +#define GUEST_INTR_STATE_STI 0x00000001 +#define GUEST_INTR_STATE_MOV_SS 0x00000002 +#define GUEST_INTR_STATE_SMI 0x00000004 +#define GUEST_INTR_STATE_NMI 0x00000008 + +/* + * Exit Qualifications for MOV for Control Register Access + */ +#define CONTROL_REG_ACCESS_NUM 0x7 /* 2:0, number of control reg.*/ +#define CONTROL_REG_ACCESS_TYPE 0x30 /* 5:4, access type */ +#define CONTROL_REG_ACCESS_REG 0xf00 /* 10:8, general purpose reg. */ +#define LMSW_SOURCE_DATA_SHIFT 16 +#define LMSW_SOURCE_DATA (0xFFFF << LMSW_SOURCE_DATA_SHIFT) /* 16:31 lmsw source */ +#define REG_EAX (0 << 8) +#define REG_ECX (1 << 8) +#define REG_EDX (2 << 8) +#define REG_EBX (3 << 8) +#define REG_ESP (4 << 8) +#define REG_EBP (5 << 8) +#define REG_ESI (6 << 8) +#define REG_EDI (7 << 8) +#define REG_R8 (8 << 8) +#define REG_R9 (9 << 8) +#define REG_R10 (10 << 8) +#define REG_R11 (11 << 8) +#define REG_R12 (12 << 8) +#define REG_R13 (13 << 8) +#define REG_R14 (14 << 8) +#define REG_R15 (15 << 8) + +/* + * Exit Qualifications for MOV for Debug Register Access + */ +#define DEBUG_REG_ACCESS_NUM 0x7 /* 2:0, number of debug reg. */ +#define DEBUG_REG_ACCESS_TYPE 0x10 /* 4, direction of access */ +#define TYPE_MOV_TO_DR (0 << 4) +#define TYPE_MOV_FROM_DR (1 << 4) +#define DEBUG_REG_ACCESS_REG 0xf00 /* 11:8, general purpose reg. */ + + +/* segment AR */ +#define SEGMENT_AR_L_MASK (1 << 13) + +#define AR_TYPE_ACCESSES_MASK 1 +#define AR_TYPE_READABLE_MASK (1 << 1) +#define AR_TYPE_WRITEABLE_MASK (1 << 2) +#define AR_TYPE_CODE_MASK (1 << 3) +#define AR_TYPE_MASK 0x0f +#define AR_TYPE_BUSY_64_TSS 11 +#define AR_TYPE_BUSY_32_TSS 11 +#define AR_TYPE_BUSY_16_TSS 3 +#define AR_TYPE_LDT 2 + +#define AR_UNUSABLE_MASK (1 << 16) +#define AR_S_MASK (1 << 4) +#define AR_P_MASK (1 << 7) +#define AR_L_MASK (1 << 13) +#define AR_DB_MASK (1 << 14) +#define AR_G_MASK (1 << 15) +#define AR_DPL_SHIFT 5 +#define AR_DPL(ar) (((ar) >> AR_DPL_SHIFT) & 3) + +#define AR_RESERVD_MASK 0xfffe0f00 + +#define TSS_PRIVATE_MEMSLOT (KVM_MEMORY_SLOTS + 0) +#define APIC_ACCESS_PAGE_PRIVATE_MEMSLOT (KVM_MEMORY_SLOTS + 1) +#define IDENTITY_PAGETABLE_PRIVATE_MEMSLOT (KVM_MEMORY_SLOTS + 2) + +#define VMX_NR_VPIDS (1 << 16) +#define VMX_VPID_EXTENT_SINGLE_CONTEXT 1 +#define VMX_VPID_EXTENT_ALL_CONTEXT 2 + +#define VMX_EPT_EXTENT_INDIVIDUAL_ADDR 0 +#define VMX_EPT_EXTENT_CONTEXT 1 +#define VMX_EPT_EXTENT_GLOBAL 2 +#define VMX_EPT_EXTENT_INDIVIDUAL_BIT (1ull << 24) +#define VMX_EPT_EXTENT_CONTEXT_BIT (1ull << 25) +#define VMX_EPT_EXTENT_GLOBAL_BIT (1ull << 26) +#define VMX_EPT_DEFAULT_GAW 3 +#define VMX_EPT_MAX_GAW 0x4 +#define VMX_EPT_MT_EPTE_SHIFT 3 +#define VMX_EPT_GAW_EPTP_SHIFT 3 +#define VMX_EPT_DEFAULT_MT 0x6ull +#define VMX_EPT_READABLE_MASK 0x1ull +#define VMX_EPT_WRITABLE_MASK 0x2ull +#define VMX_EPT_EXECUTABLE_MASK 0x4ull +#define VMX_EPT_IGMT_BIT (1ull << 6) + +#define VMX_EPT_IDENTITY_PAGETABLE_ADDR 0xfffbc000ul + +#endif diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c index 8904e8ada978..fa3486d64078 100644 --- a/arch/x86/kvm/mmu.c +++ b/arch/x86/kvm/mmu.c @@ -17,7 +17,6 @@ * */ -#include "vmx.h" #include "mmu.h" #include @@ -33,6 +32,7 @@ #include #include #include +#include /* * When setting this variable to true it enables Two-Dimensional-Paging diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index 427dbc14fae9..ec71f6464cfa 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -16,7 +16,6 @@ */ #include "irq.h" -#include "vmx.h" #include "mmu.h" #include @@ -31,6 +30,7 @@ #include #include +#include #define __ex(x) __kvm_handle_fault_on_reboot(x) diff --git a/arch/x86/kvm/vmx.h b/arch/x86/kvm/vmx.h deleted file mode 100644 index 3db236c26fa4..000000000000 --- a/arch/x86/kvm/vmx.h +++ /dev/null @@ -1,367 +0,0 @@ -#ifndef VMX_H -#define VMX_H - -/* - * vmx.h: VMX Architecture related definitions - * Copyright (c) 2004, Intel Corporation. - * - * This program is free software; you can redistribute it and/or modify it - * under the terms and conditions of the GNU General Public License, - * version 2, as published by the Free Software Foundation. - * - * This program is distributed in the hope it will be useful, but WITHOUT - * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or - * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for - * more details. - * - * You should have received a copy of the GNU General Public License along with - * this program; if not, write to the Free Software Foundation, Inc., 59 Temple - * Place - Suite 330, Boston, MA 02111-1307 USA. - * - * A few random additions are: - * Copyright (C) 2006 Qumranet - * Avi Kivity - * Yaniv Kamay - * - */ - -/* - * Definitions of Primary Processor-Based VM-Execution Controls. - */ -#define CPU_BASED_VIRTUAL_INTR_PENDING 0x00000004 -#define CPU_BASED_USE_TSC_OFFSETING 0x00000008 -#define CPU_BASED_HLT_EXITING 0x00000080 -#define CPU_BASED_INVLPG_EXITING 0x00000200 -#define CPU_BASED_MWAIT_EXITING 0x00000400 -#define CPU_BASED_RDPMC_EXITING 0x00000800 -#define CPU_BASED_RDTSC_EXITING 0x00001000 -#define CPU_BASED_CR3_LOAD_EXITING 0x00008000 -#define CPU_BASED_CR3_STORE_EXITING 0x00010000 -#define CPU_BASED_CR8_LOAD_EXITING 0x00080000 -#define CPU_BASED_CR8_STORE_EXITING 0x00100000 -#define CPU_BASED_TPR_SHADOW 0x00200000 -#define CPU_BASED_VIRTUAL_NMI_PENDING 0x00400000 -#define CPU_BASED_MOV_DR_EXITING 0x00800000 -#define CPU_BASED_UNCOND_IO_EXITING 0x01000000 -#define CPU_BASED_USE_IO_BITMAPS 0x02000000 -#define CPU_BASED_USE_MSR_BITMAPS 0x10000000 -#define CPU_BASED_MONITOR_EXITING 0x20000000 -#define CPU_BASED_PAUSE_EXITING 0x40000000 -#define CPU_BASED_ACTIVATE_SECONDARY_CONTROLS 0x80000000 -/* - * Definitions of Secondary Processor-Based VM-Execution Controls. - */ -#define SECONDARY_EXEC_VIRTUALIZE_APIC_ACCESSES 0x00000001 -#define SECONDARY_EXEC_ENABLE_EPT 0x00000002 -#define SECONDARY_EXEC_ENABLE_VPID 0x00000020 -#define SECONDARY_EXEC_WBINVD_EXITING 0x00000040 - - -#define PIN_BASED_EXT_INTR_MASK 0x00000001 -#define PIN_BASED_NMI_EXITING 0x00000008 -#define PIN_BASED_VIRTUAL_NMIS 0x00000020 - -#define VM_EXIT_HOST_ADDR_SPACE_SIZE 0x00000200 -#define VM_EXIT_ACK_INTR_ON_EXIT 0x00008000 -#define VM_EXIT_SAVE_IA32_PAT 0x00040000 -#define VM_EXIT_LOAD_IA32_PAT 0x00080000 - -#define VM_ENTRY_IA32E_MODE 0x00000200 -#define VM_ENTRY_SMM 0x00000400 -#define VM_ENTRY_DEACT_DUAL_MONITOR 0x00000800 -#define VM_ENTRY_LOAD_IA32_PAT 0x00004000 - -/* VMCS Encodings */ -enum vmcs_field { - VIRTUAL_PROCESSOR_ID = 0x00000000, - GUEST_ES_SELECTOR = 0x00000800, - GUEST_CS_SELECTOR = 0x00000802, - GUEST_SS_SELECTOR = 0x00000804, - GUEST_DS_SELECTOR = 0x00000806, - GUEST_FS_SELECTOR = 0x00000808, - GUEST_GS_SELECTOR = 0x0000080a, - GUEST_LDTR_SELECTOR = 0x0000080c, - GUEST_TR_SELECTOR = 0x0000080e, - HOST_ES_SELECTOR = 0x00000c00, - HOST_CS_SELECTOR = 0x00000c02, - HOST_SS_SELECTOR = 0x00000c04, - HOST_DS_SELECTOR = 0x00000c06, - HOST_FS_SELECTOR = 0x00000c08, - HOST_GS_SELECTOR = 0x00000c0a, - HOST_TR_SELECTOR = 0x00000c0c, - IO_BITMAP_A = 0x00002000, - IO_BITMAP_A_HIGH = 0x00002001, - IO_BITMAP_B = 0x00002002, - IO_BITMAP_B_HIGH = 0x00002003, - MSR_BITMAP = 0x00002004, - MSR_BITMAP_HIGH = 0x00002005, - VM_EXIT_MSR_STORE_ADDR = 0x00002006, - VM_EXIT_MSR_STORE_ADDR_HIGH = 0x00002007, - VM_EXIT_MSR_LOAD_ADDR = 0x00002008, - VM_EXIT_MSR_LOAD_ADDR_HIGH = 0x00002009, - VM_ENTRY_MSR_LOAD_ADDR = 0x0000200a, - VM_ENTRY_MSR_LOAD_ADDR_HIGH = 0x0000200b, - TSC_OFFSET = 0x00002010, - TSC_OFFSET_HIGH = 0x00002011, - VIRTUAL_APIC_PAGE_ADDR = 0x00002012, - VIRTUAL_APIC_PAGE_ADDR_HIGH = 0x00002013, - APIC_ACCESS_ADDR = 0x00002014, - APIC_ACCESS_ADDR_HIGH = 0x00002015, - EPT_POINTER = 0x0000201a, - EPT_POINTER_HIGH = 0x0000201b, - GUEST_PHYSICAL_ADDRESS = 0x00002400, - GUEST_PHYSICAL_ADDRESS_HIGH = 0x00002401, - VMCS_LINK_POINTER = 0x00002800, - VMCS_LINK_POINTER_HIGH = 0x00002801, - GUEST_IA32_DEBUGCTL = 0x00002802, - GUEST_IA32_DEBUGCTL_HIGH = 0x00002803, - GUEST_IA32_PAT = 0x00002804, - GUEST_IA32_PAT_HIGH = 0x00002805, - GUEST_PDPTR0 = 0x0000280a, - GUEST_PDPTR0_HIGH = 0x0000280b, - GUEST_PDPTR1 = 0x0000280c, - GUEST_PDPTR1_HIGH = 0x0000280d, - GUEST_PDPTR2 = 0x0000280e, - GUEST_PDPTR2_HIGH = 0x0000280f, - GUEST_PDPTR3 = 0x00002810, - GUEST_PDPTR3_HIGH = 0x00002811, - HOST_IA32_PAT = 0x00002c00, - HOST_IA32_PAT_HIGH = 0x00002c01, - PIN_BASED_VM_EXEC_CONTROL = 0x00004000, - CPU_BASED_VM_EXEC_CONTROL = 0x00004002, - EXCEPTION_BITMAP = 0x00004004, - PAGE_FAULT_ERROR_CODE_MASK = 0x00004006, - PAGE_FAULT_ERROR_CODE_MATCH = 0x00004008, - CR3_TARGET_COUNT = 0x0000400a, - VM_EXIT_CONTROLS = 0x0000400c, - VM_EXIT_MSR_STORE_COUNT = 0x0000400e, - VM_EXIT_MSR_LOAD_COUNT = 0x00004010, - VM_ENTRY_CONTROLS = 0x00004012, - VM_ENTRY_MSR_LOAD_COUNT = 0x00004014, - VM_ENTRY_INTR_INFO_FIELD = 0x00004016, - VM_ENTRY_EXCEPTION_ERROR_CODE = 0x00004018, - VM_ENTRY_INSTRUCTION_LEN = 0x0000401a, - TPR_THRESHOLD = 0x0000401c, - SECONDARY_VM_EXEC_CONTROL = 0x0000401e, - VM_INSTRUCTION_ERROR = 0x00004400, - VM_EXIT_REASON = 0x00004402, - VM_EXIT_INTR_INFO = 0x00004404, - VM_EXIT_INTR_ERROR_CODE = 0x00004406, - IDT_VECTORING_INFO_FIELD = 0x00004408, - IDT_VECTORING_ERROR_CODE = 0x0000440a, - VM_EXIT_INSTRUCTION_LEN = 0x0000440c, - VMX_INSTRUCTION_INFO = 0x0000440e, - GUEST_ES_LIMIT = 0x00004800, - GUEST_CS_LIMIT = 0x00004802, - GUEST_SS_LIMIT = 0x00004804, - GUEST_DS_LIMIT = 0x00004806, - GUEST_FS_LIMIT = 0x00004808, - GUEST_GS_LIMIT = 0x0000480a, - GUEST_LDTR_LIMIT = 0x0000480c, - GUEST_TR_LIMIT = 0x0000480e, - GUEST_GDTR_LIMIT = 0x00004810, - GUEST_IDTR_LIMIT = 0x00004812, - GUEST_ES_AR_BYTES = 0x00004814, - GUEST_CS_AR_BYTES = 0x00004816, - GUEST_SS_AR_BYTES = 0x00004818, - GUEST_DS_AR_BYTES = 0x0000481a, - GUEST_FS_AR_BYTES = 0x0000481c, - GUEST_GS_AR_BYTES = 0x0000481e, - GUEST_LDTR_AR_BYTES = 0x00004820, - GUEST_TR_AR_BYTES = 0x00004822, - GUEST_INTERRUPTIBILITY_INFO = 0x00004824, - GUEST_ACTIVITY_STATE = 0X00004826, - GUEST_SYSENTER_CS = 0x0000482A, - HOST_IA32_SYSENTER_CS = 0x00004c00, - CR0_GUEST_HOST_MASK = 0x00006000, - CR4_GUEST_HOST_MASK = 0x00006002, - CR0_READ_SHADOW = 0x00006004, - CR4_READ_SHADOW = 0x00006006, - CR3_TARGET_VALUE0 = 0x00006008, - CR3_TARGET_VALUE1 = 0x0000600a, - CR3_TARGET_VALUE2 = 0x0000600c, - CR3_TARGET_VALUE3 = 0x0000600e, - EXIT_QUALIFICATION = 0x00006400, - GUEST_LINEAR_ADDRESS = 0x0000640a, - GUEST_CR0 = 0x00006800, - GUEST_CR3 = 0x00006802, - GUEST_CR4 = 0x00006804, - GUEST_ES_BASE = 0x00006806, - GUEST_CS_BASE = 0x00006808, - GUEST_SS_BASE = 0x0000680a, - GUEST_DS_BASE = 0x0000680c, - GUEST_FS_BASE = 0x0000680e, - GUEST_GS_BASE = 0x00006810, - GUEST_LDTR_BASE = 0x00006812, - GUEST_TR_BASE = 0x00006814, - GUEST_GDTR_BASE = 0x00006816, - GUEST_IDTR_BASE = 0x00006818, - GUEST_DR7 = 0x0000681a, - GUEST_RSP = 0x0000681c, - GUEST_RIP = 0x0000681e, - GUEST_RFLAGS = 0x00006820, - GUEST_PENDING_DBG_EXCEPTIONS = 0x00006822, - GUEST_SYSENTER_ESP = 0x00006824, - GUEST_SYSENTER_EIP = 0x00006826, - HOST_CR0 = 0x00006c00, - HOST_CR3 = 0x00006c02, - HOST_CR4 = 0x00006c04, - HOST_FS_BASE = 0x00006c06, - HOST_GS_BASE = 0x00006c08, - HOST_TR_BASE = 0x00006c0a, - HOST_GDTR_BASE = 0x00006c0c, - HOST_IDTR_BASE = 0x00006c0e, - HOST_IA32_SYSENTER_ESP = 0x00006c10, - HOST_IA32_SYSENTER_EIP = 0x00006c12, - HOST_RSP = 0x00006c14, - HOST_RIP = 0x00006c16, -}; - -#define VMX_EXIT_REASONS_FAILED_VMENTRY 0x80000000 - -#define EXIT_REASON_EXCEPTION_NMI 0 -#define EXIT_REASON_EXTERNAL_INTERRUPT 1 -#define EXIT_REASON_TRIPLE_FAULT 2 - -#define EXIT_REASON_PENDING_INTERRUPT 7 -#define EXIT_REASON_NMI_WINDOW 8 -#define EXIT_REASON_TASK_SWITCH 9 -#define EXIT_REASON_CPUID 10 -#define EXIT_REASON_HLT 12 -#define EXIT_REASON_INVLPG 14 -#define EXIT_REASON_RDPMC 15 -#define EXIT_REASON_RDTSC 16 -#define EXIT_REASON_VMCALL 18 -#define EXIT_REASON_VMCLEAR 19 -#define EXIT_REASON_VMLAUNCH 20 -#define EXIT_REASON_VMPTRLD 21 -#define EXIT_REASON_VMPTRST 22 -#define EXIT_REASON_VMREAD 23 -#define EXIT_REASON_VMRESUME 24 -#define EXIT_REASON_VMWRITE 25 -#define EXIT_REASON_VMOFF 26 -#define EXIT_REASON_VMON 27 -#define EXIT_REASON_CR_ACCESS 28 -#define EXIT_REASON_DR_ACCESS 29 -#define EXIT_REASON_IO_INSTRUCTION 30 -#define EXIT_REASON_MSR_READ 31 -#define EXIT_REASON_MSR_WRITE 32 -#define EXIT_REASON_MWAIT_INSTRUCTION 36 -#define EXIT_REASON_TPR_BELOW_THRESHOLD 43 -#define EXIT_REASON_APIC_ACCESS 44 -#define EXIT_REASON_EPT_VIOLATION 48 -#define EXIT_REASON_EPT_MISCONFIG 49 -#define EXIT_REASON_WBINVD 54 - -/* - * Interruption-information format - */ -#define INTR_INFO_VECTOR_MASK 0xff /* 7:0 */ -#define INTR_INFO_INTR_TYPE_MASK 0x700 /* 10:8 */ -#define INTR_INFO_DELIVER_CODE_MASK 0x800 /* 11 */ -#define INTR_INFO_UNBLOCK_NMI 0x1000 /* 12 */ -#define INTR_INFO_VALID_MASK 0x80000000 /* 31 */ -#define INTR_INFO_RESVD_BITS_MASK 0x7ffff000 - -#define VECTORING_INFO_VECTOR_MASK INTR_INFO_VECTOR_MASK -#define VECTORING_INFO_TYPE_MASK INTR_INFO_INTR_TYPE_MASK -#define VECTORING_INFO_DELIVER_CODE_MASK INTR_INFO_DELIVER_CODE_MASK -#define VECTORING_INFO_VALID_MASK INTR_INFO_VALID_MASK - -#define INTR_TYPE_EXT_INTR (0 << 8) /* external interrupt */ -#define INTR_TYPE_NMI_INTR (2 << 8) /* NMI */ -#define INTR_TYPE_EXCEPTION (3 << 8) /* processor exception */ -#define INTR_TYPE_SOFT_INTR (4 << 8) /* software interrupt */ - -/* GUEST_INTERRUPTIBILITY_INFO flags. */ -#define GUEST_INTR_STATE_STI 0x00000001 -#define GUEST_INTR_STATE_MOV_SS 0x00000002 -#define GUEST_INTR_STATE_SMI 0x00000004 -#define GUEST_INTR_STATE_NMI 0x00000008 - -/* - * Exit Qualifications for MOV for Control Register Access - */ -#define CONTROL_REG_ACCESS_NUM 0x7 /* 2:0, number of control reg.*/ -#define CONTROL_REG_ACCESS_TYPE 0x30 /* 5:4, access type */ -#define CONTROL_REG_ACCESS_REG 0xf00 /* 10:8, general purpose reg. */ -#define LMSW_SOURCE_DATA_SHIFT 16 -#define LMSW_SOURCE_DATA (0xFFFF << LMSW_SOURCE_DATA_SHIFT) /* 16:31 lmsw source */ -#define REG_EAX (0 << 8) -#define REG_ECX (1 << 8) -#define REG_EDX (2 << 8) -#define REG_EBX (3 << 8) -#define REG_ESP (4 << 8) -#define REG_EBP (5 << 8) -#define REG_ESI (6 << 8) -#define REG_EDI (7 << 8) -#define REG_R8 (8 << 8) -#define REG_R9 (9 << 8) -#define REG_R10 (10 << 8) -#define REG_R11 (11 << 8) -#define REG_R12 (12 << 8) -#define REG_R13 (13 << 8) -#define REG_R14 (14 << 8) -#define REG_R15 (15 << 8) - -/* - * Exit Qualifications for MOV for Debug Register Access - */ -#define DEBUG_REG_ACCESS_NUM 0x7 /* 2:0, number of debug reg. */ -#define DEBUG_REG_ACCESS_TYPE 0x10 /* 4, direction of access */ -#define TYPE_MOV_TO_DR (0 << 4) -#define TYPE_MOV_FROM_DR (1 << 4) -#define DEBUG_REG_ACCESS_REG 0xf00 /* 11:8, general purpose reg. */ - - -/* segment AR */ -#define SEGMENT_AR_L_MASK (1 << 13) - -#define AR_TYPE_ACCESSES_MASK 1 -#define AR_TYPE_READABLE_MASK (1 << 1) -#define AR_TYPE_WRITEABLE_MASK (1 << 2) -#define AR_TYPE_CODE_MASK (1 << 3) -#define AR_TYPE_MASK 0x0f -#define AR_TYPE_BUSY_64_TSS 11 -#define AR_TYPE_BUSY_32_TSS 11 -#define AR_TYPE_BUSY_16_TSS 3 -#define AR_TYPE_LDT 2 - -#define AR_UNUSABLE_MASK (1 << 16) -#define AR_S_MASK (1 << 4) -#define AR_P_MASK (1 << 7) -#define AR_L_MASK (1 << 13) -#define AR_DB_MASK (1 << 14) -#define AR_G_MASK (1 << 15) -#define AR_DPL_SHIFT 5 -#define AR_DPL(ar) (((ar) >> AR_DPL_SHIFT) & 3) - -#define AR_RESERVD_MASK 0xfffe0f00 - -#define TSS_PRIVATE_MEMSLOT (KVM_MEMORY_SLOTS + 0) -#define APIC_ACCESS_PAGE_PRIVATE_MEMSLOT (KVM_MEMORY_SLOTS + 1) -#define IDENTITY_PAGETABLE_PRIVATE_MEMSLOT (KVM_MEMORY_SLOTS + 2) - -#define VMX_NR_VPIDS (1 << 16) -#define VMX_VPID_EXTENT_SINGLE_CONTEXT 1 -#define VMX_VPID_EXTENT_ALL_CONTEXT 2 - -#define VMX_EPT_EXTENT_INDIVIDUAL_ADDR 0 -#define VMX_EPT_EXTENT_CONTEXT 1 -#define VMX_EPT_EXTENT_GLOBAL 2 -#define VMX_EPT_EXTENT_INDIVIDUAL_BIT (1ull << 24) -#define VMX_EPT_EXTENT_CONTEXT_BIT (1ull << 25) -#define VMX_EPT_EXTENT_GLOBAL_BIT (1ull << 26) -#define VMX_EPT_DEFAULT_GAW 3 -#define VMX_EPT_MAX_GAW 0x4 -#define VMX_EPT_MT_EPTE_SHIFT 3 -#define VMX_EPT_GAW_EPTP_SHIFT 3 -#define VMX_EPT_DEFAULT_MT 0x6ull -#define VMX_EPT_READABLE_MASK 0x1ull -#define VMX_EPT_WRITABLE_MASK 0x2ull -#define VMX_EPT_EXECUTABLE_MASK 0x4ull -#define VMX_EPT_IGMT_BIT (1ull << 6) - -#define VMX_EPT_IDENTITY_PAGETABLE_ADDR 0xfffbc000ul - -#endif -- cgit v1.2.3 From c2cedf7be2017e3264c93a4c0d75b1d96d0d7104 Mon Sep 17 00:00:00 2001 From: Eduardo Habkost Date: Mon, 17 Nov 2008 19:03:14 -0200 Subject: KVM: SVM: move svm.h to include/asm svm.h will be used by core code that is independent of KVM, so I am moving it outside the arch/x86/kvm directory. Signed-off-by: Eduardo Habkost Signed-off-by: Avi Kivity --- arch/x86/include/asm/svm.h | 328 +++++++++++++++++++++++++++++++++++++++++++++ arch/x86/kvm/kvm_svm.h | 2 +- arch/x86/kvm/svm.h | 328 --------------------------------------------- 3 files changed, 329 insertions(+), 329 deletions(-) create mode 100644 arch/x86/include/asm/svm.h delete mode 100644 arch/x86/kvm/svm.h (limited to 'arch/x86') diff --git a/arch/x86/include/asm/svm.h b/arch/x86/include/asm/svm.h new file mode 100644 index 000000000000..1b8afa78e869 --- /dev/null +++ b/arch/x86/include/asm/svm.h @@ -0,0 +1,328 @@ +#ifndef __SVM_H +#define __SVM_H + +enum { + INTERCEPT_INTR, + INTERCEPT_NMI, + INTERCEPT_SMI, + INTERCEPT_INIT, + INTERCEPT_VINTR, + INTERCEPT_SELECTIVE_CR0, + INTERCEPT_STORE_IDTR, + INTERCEPT_STORE_GDTR, + INTERCEPT_STORE_LDTR, + INTERCEPT_STORE_TR, + INTERCEPT_LOAD_IDTR, + INTERCEPT_LOAD_GDTR, + INTERCEPT_LOAD_LDTR, + INTERCEPT_LOAD_TR, + INTERCEPT_RDTSC, + INTERCEPT_RDPMC, + INTERCEPT_PUSHF, + INTERCEPT_POPF, + INTERCEPT_CPUID, + INTERCEPT_RSM, + INTERCEPT_IRET, + INTERCEPT_INTn, + INTERCEPT_INVD, + INTERCEPT_PAUSE, + INTERCEPT_HLT, + INTERCEPT_INVLPG, + INTERCEPT_INVLPGA, + INTERCEPT_IOIO_PROT, + INTERCEPT_MSR_PROT, + INTERCEPT_TASK_SWITCH, + INTERCEPT_FERR_FREEZE, + INTERCEPT_SHUTDOWN, + INTERCEPT_VMRUN, + INTERCEPT_VMMCALL, + INTERCEPT_VMLOAD, + INTERCEPT_VMSAVE, + INTERCEPT_STGI, + INTERCEPT_CLGI, + INTERCEPT_SKINIT, + INTERCEPT_RDTSCP, + INTERCEPT_ICEBP, + INTERCEPT_WBINVD, + INTERCEPT_MONITOR, + INTERCEPT_MWAIT, + INTERCEPT_MWAIT_COND, +}; + + +struct __attribute__ ((__packed__)) vmcb_control_area { + u16 intercept_cr_read; + u16 intercept_cr_write; + u16 intercept_dr_read; + u16 intercept_dr_write; + u32 intercept_exceptions; + u64 intercept; + u8 reserved_1[44]; + u64 iopm_base_pa; + u64 msrpm_base_pa; + u64 tsc_offset; + u32 asid; + u8 tlb_ctl; + u8 reserved_2[3]; + u32 int_ctl; + u32 int_vector; + u32 int_state; + u8 reserved_3[4]; + u32 exit_code; + u32 exit_code_hi; + u64 exit_info_1; + u64 exit_info_2; + u32 exit_int_info; + u32 exit_int_info_err; + u64 nested_ctl; + u8 reserved_4[16]; + u32 event_inj; + u32 event_inj_err; + u64 nested_cr3; + u64 lbr_ctl; + u8 reserved_5[832]; +}; + + +#define TLB_CONTROL_DO_NOTHING 0 +#define TLB_CONTROL_FLUSH_ALL_ASID 1 + +#define V_TPR_MASK 0x0f + +#define V_IRQ_SHIFT 8 +#define V_IRQ_MASK (1 << V_IRQ_SHIFT) + +#define V_INTR_PRIO_SHIFT 16 +#define V_INTR_PRIO_MASK (0x0f << V_INTR_PRIO_SHIFT) + +#define V_IGN_TPR_SHIFT 20 +#define V_IGN_TPR_MASK (1 << V_IGN_TPR_SHIFT) + +#define V_INTR_MASKING_SHIFT 24 +#define V_INTR_MASKING_MASK (1 << V_INTR_MASKING_SHIFT) + +#define SVM_INTERRUPT_SHADOW_MASK 1 + +#define SVM_IOIO_STR_SHIFT 2 +#define SVM_IOIO_REP_SHIFT 3 +#define SVM_IOIO_SIZE_SHIFT 4 +#define SVM_IOIO_ASIZE_SHIFT 7 + +#define SVM_IOIO_TYPE_MASK 1 +#define SVM_IOIO_STR_MASK (1 << SVM_IOIO_STR_SHIFT) +#define SVM_IOIO_REP_MASK (1 << SVM_IOIO_REP_SHIFT) +#define SVM_IOIO_SIZE_MASK (7 << SVM_IOIO_SIZE_SHIFT) +#define SVM_IOIO_ASIZE_MASK (7 << SVM_IOIO_ASIZE_SHIFT) + +struct __attribute__ ((__packed__)) vmcb_seg { + u16 selector; + u16 attrib; + u32 limit; + u64 base; +}; + +struct __attribute__ ((__packed__)) vmcb_save_area { + struct vmcb_seg es; + struct vmcb_seg cs; + struct vmcb_seg ss; + struct vmcb_seg ds; + struct vmcb_seg fs; + struct vmcb_seg gs; + struct vmcb_seg gdtr; + struct vmcb_seg ldtr; + struct vmcb_seg idtr; + struct vmcb_seg tr; + u8 reserved_1[43]; + u8 cpl; + u8 reserved_2[4]; + u64 efer; + u8 reserved_3[112]; + u64 cr4; + u64 cr3; + u64 cr0; + u64 dr7; + u64 dr6; + u64 rflags; + u64 rip; + u8 reserved_4[88]; + u64 rsp; + u8 reserved_5[24]; + u64 rax; + u64 star; + u64 lstar; + u64 cstar; + u64 sfmask; + u64 kernel_gs_base; + u64 sysenter_cs; + u64 sysenter_esp; + u64 sysenter_eip; + u64 cr2; + u8 reserved_6[32]; + u64 g_pat; + u64 dbgctl; + u64 br_from; + u64 br_to; + u64 last_excp_from; + u64 last_excp_to; +}; + +struct __attribute__ ((__packed__)) vmcb { + struct vmcb_control_area control; + struct vmcb_save_area save; +}; + +#define SVM_CPUID_FEATURE_SHIFT 2 +#define SVM_CPUID_FUNC 0x8000000a + +#define MSR_EFER_SVME_MASK (1ULL << 12) +#define MSR_VM_CR 0xc0010114 +#define MSR_VM_HSAVE_PA 0xc0010117ULL + +#define SVM_VM_CR_SVM_DISABLE 4 + +#define SVM_SELECTOR_S_SHIFT 4 +#define SVM_SELECTOR_DPL_SHIFT 5 +#define SVM_SELECTOR_P_SHIFT 7 +#define SVM_SELECTOR_AVL_SHIFT 8 +#define SVM_SELECTOR_L_SHIFT 9 +#define SVM_SELECTOR_DB_SHIFT 10 +#define SVM_SELECTOR_G_SHIFT 11 + +#define SVM_SELECTOR_TYPE_MASK (0xf) +#define SVM_SELECTOR_S_MASK (1 << SVM_SELECTOR_S_SHIFT) +#define SVM_SELECTOR_DPL_MASK (3 << SVM_SELECTOR_DPL_SHIFT) +#define SVM_SELECTOR_P_MASK (1 << SVM_SELECTOR_P_SHIFT) +#define SVM_SELECTOR_AVL_MASK (1 << SVM_SELECTOR_AVL_SHIFT) +#define SVM_SELECTOR_L_MASK (1 << SVM_SELECTOR_L_SHIFT) +#define SVM_SELECTOR_DB_MASK (1 << SVM_SELECTOR_DB_SHIFT) +#define SVM_SELECTOR_G_MASK (1 << SVM_SELECTOR_G_SHIFT) + +#define SVM_SELECTOR_WRITE_MASK (1 << 1) +#define SVM_SELECTOR_READ_MASK SVM_SELECTOR_WRITE_MASK +#define SVM_SELECTOR_CODE_MASK (1 << 3) + +#define INTERCEPT_CR0_MASK 1 +#define INTERCEPT_CR3_MASK (1 << 3) +#define INTERCEPT_CR4_MASK (1 << 4) +#define INTERCEPT_CR8_MASK (1 << 8) + +#define INTERCEPT_DR0_MASK 1 +#define INTERCEPT_DR1_MASK (1 << 1) +#define INTERCEPT_DR2_MASK (1 << 2) +#define INTERCEPT_DR3_MASK (1 << 3) +#define INTERCEPT_DR4_MASK (1 << 4) +#define INTERCEPT_DR5_MASK (1 << 5) +#define INTERCEPT_DR6_MASK (1 << 6) +#define INTERCEPT_DR7_MASK (1 << 7) + +#define SVM_EVTINJ_VEC_MASK 0xff + +#define SVM_EVTINJ_TYPE_SHIFT 8 +#define SVM_EVTINJ_TYPE_MASK (7 << SVM_EVTINJ_TYPE_SHIFT) + +#define SVM_EVTINJ_TYPE_INTR (0 << SVM_EVTINJ_TYPE_SHIFT) +#define SVM_EVTINJ_TYPE_NMI (2 << SVM_EVTINJ_TYPE_SHIFT) +#define SVM_EVTINJ_TYPE_EXEPT (3 << SVM_EVTINJ_TYPE_SHIFT) +#define SVM_EVTINJ_TYPE_SOFT (4 << SVM_EVTINJ_TYPE_SHIFT) + +#define SVM_EVTINJ_VALID (1 << 31) +#define SVM_EVTINJ_VALID_ERR (1 << 11) + +#define SVM_EXITINTINFO_VEC_MASK SVM_EVTINJ_VEC_MASK + +#define SVM_EXITINTINFO_TYPE_INTR SVM_EVTINJ_TYPE_INTR +#define SVM_EXITINTINFO_TYPE_NMI SVM_EVTINJ_TYPE_NMI +#define SVM_EXITINTINFO_TYPE_EXEPT SVM_EVTINJ_TYPE_EXEPT +#define SVM_EXITINTINFO_TYPE_SOFT SVM_EVTINJ_TYPE_SOFT + +#define SVM_EXITINTINFO_VALID SVM_EVTINJ_VALID +#define SVM_EXITINTINFO_VALID_ERR SVM_EVTINJ_VALID_ERR + +#define SVM_EXITINFOSHIFT_TS_REASON_IRET 36 +#define SVM_EXITINFOSHIFT_TS_REASON_JMP 38 + +#define SVM_EXIT_READ_CR0 0x000 +#define SVM_EXIT_READ_CR3 0x003 +#define SVM_EXIT_READ_CR4 0x004 +#define SVM_EXIT_READ_CR8 0x008 +#define SVM_EXIT_WRITE_CR0 0x010 +#define SVM_EXIT_WRITE_CR3 0x013 +#define SVM_EXIT_WRITE_CR4 0x014 +#define SVM_EXIT_WRITE_CR8 0x018 +#define SVM_EXIT_READ_DR0 0x020 +#define SVM_EXIT_READ_DR1 0x021 +#define SVM_EXIT_READ_DR2 0x022 +#define SVM_EXIT_READ_DR3 0x023 +#define SVM_EXIT_READ_DR4 0x024 +#define SVM_EXIT_READ_DR5 0x025 +#define SVM_EXIT_READ_DR6 0x026 +#define SVM_EXIT_READ_DR7 0x027 +#define SVM_EXIT_WRITE_DR0 0x030 +#define SVM_EXIT_WRITE_DR1 0x031 +#define SVM_EXIT_WRITE_DR2 0x032 +#define SVM_EXIT_WRITE_DR3 0x033 +#define SVM_EXIT_WRITE_DR4 0x034 +#define SVM_EXIT_WRITE_DR5 0x035 +#define SVM_EXIT_WRITE_DR6 0x036 +#define SVM_EXIT_WRITE_DR7 0x037 +#define SVM_EXIT_EXCP_BASE 0x040 +#define SVM_EXIT_INTR 0x060 +#define SVM_EXIT_NMI 0x061 +#define SVM_EXIT_SMI 0x062 +#define SVM_EXIT_INIT 0x063 +#define SVM_EXIT_VINTR 0x064 +#define SVM_EXIT_CR0_SEL_WRITE 0x065 +#define SVM_EXIT_IDTR_READ 0x066 +#define SVM_EXIT_GDTR_READ 0x067 +#define SVM_EXIT_LDTR_READ 0x068 +#define SVM_EXIT_TR_READ 0x069 +#define SVM_EXIT_IDTR_WRITE 0x06a +#define SVM_EXIT_GDTR_WRITE 0x06b +#define SVM_EXIT_LDTR_WRITE 0x06c +#define SVM_EXIT_TR_WRITE 0x06d +#define SVM_EXIT_RDTSC 0x06e +#define SVM_EXIT_RDPMC 0x06f +#define SVM_EXIT_PUSHF 0x070 +#define SVM_EXIT_POPF 0x071 +#define SVM_EXIT_CPUID 0x072 +#define SVM_EXIT_RSM 0x073 +#define SVM_EXIT_IRET 0x074 +#define SVM_EXIT_SWINT 0x075 +#define SVM_EXIT_INVD 0x076 +#define SVM_EXIT_PAUSE 0x077 +#define SVM_EXIT_HLT 0x078 +#define SVM_EXIT_INVLPG 0x079 +#define SVM_EXIT_INVLPGA 0x07a +#define SVM_EXIT_IOIO 0x07b +#define SVM_EXIT_MSR 0x07c +#define SVM_EXIT_TASK_SWITCH 0x07d +#define SVM_EXIT_FERR_FREEZE 0x07e +#define SVM_EXIT_SHUTDOWN 0x07f +#define SVM_EXIT_VMRUN 0x080 +#define SVM_EXIT_VMMCALL 0x081 +#define SVM_EXIT_VMLOAD 0x082 +#define SVM_EXIT_VMSAVE 0x083 +#define SVM_EXIT_STGI 0x084 +#define SVM_EXIT_CLGI 0x085 +#define SVM_EXIT_SKINIT 0x086 +#define SVM_EXIT_RDTSCP 0x087 +#define SVM_EXIT_ICEBP 0x088 +#define SVM_EXIT_WBINVD 0x089 +#define SVM_EXIT_MONITOR 0x08a +#define SVM_EXIT_MWAIT 0x08b +#define SVM_EXIT_MWAIT_COND 0x08c +#define SVM_EXIT_NPF 0x400 + +#define SVM_EXIT_ERR -1 + +#define SVM_CR0_SELECTIVE_MASK (1 << 3 | 1) /* TS and MP */ + +#define SVM_VMLOAD ".byte 0x0f, 0x01, 0xda" +#define SVM_VMRUN ".byte 0x0f, 0x01, 0xd8" +#define SVM_VMSAVE ".byte 0x0f, 0x01, 0xdb" +#define SVM_CLGI ".byte 0x0f, 0x01, 0xdd" +#define SVM_STGI ".byte 0x0f, 0x01, 0xdc" +#define SVM_INVLPGA ".byte 0x0f, 0x01, 0xdf" + +#endif + diff --git a/arch/x86/kvm/kvm_svm.h b/arch/x86/kvm/kvm_svm.h index 65ef0fc2c036..8e5ee99551f6 100644 --- a/arch/x86/kvm/kvm_svm.h +++ b/arch/x86/kvm/kvm_svm.h @@ -7,7 +7,7 @@ #include #include -#include "svm.h" +#include static const u32 host_save_user_msrs[] = { #ifdef CONFIG_X86_64 diff --git a/arch/x86/kvm/svm.h b/arch/x86/kvm/svm.h deleted file mode 100644 index 1b8afa78e869..000000000000 --- a/arch/x86/kvm/svm.h +++ /dev/null @@ -1,328 +0,0 @@ -#ifndef __SVM_H -#define __SVM_H - -enum { - INTERCEPT_INTR, - INTERCEPT_NMI, - INTERCEPT_SMI, - INTERCEPT_INIT, - INTERCEPT_VINTR, - INTERCEPT_SELECTIVE_CR0, - INTERCEPT_STORE_IDTR, - INTERCEPT_STORE_GDTR, - INTERCEPT_STORE_LDTR, - INTERCEPT_STORE_TR, - INTERCEPT_LOAD_IDTR, - INTERCEPT_LOAD_GDTR, - INTERCEPT_LOAD_LDTR, - INTERCEPT_LOAD_TR, - INTERCEPT_RDTSC, - INTERCEPT_RDPMC, - INTERCEPT_PUSHF, - INTERCEPT_POPF, - INTERCEPT_CPUID, - INTERCEPT_RSM, - INTERCEPT_IRET, - INTERCEPT_INTn, - INTERCEPT_INVD, - INTERCEPT_PAUSE, - INTERCEPT_HLT, - INTERCEPT_INVLPG, - INTERCEPT_INVLPGA, - INTERCEPT_IOIO_PROT, - INTERCEPT_MSR_PROT, - INTERCEPT_TASK_SWITCH, - INTERCEPT_FERR_FREEZE, - INTERCEPT_SHUTDOWN, - INTERCEPT_VMRUN, - INTERCEPT_VMMCALL, - INTERCEPT_VMLOAD, - INTERCEPT_VMSAVE, - INTERCEPT_STGI, - INTERCEPT_CLGI, - INTERCEPT_SKINIT, - INTERCEPT_RDTSCP, - INTERCEPT_ICEBP, - INTERCEPT_WBINVD, - INTERCEPT_MONITOR, - INTERCEPT_MWAIT, - INTERCEPT_MWAIT_COND, -}; - - -struct __attribute__ ((__packed__)) vmcb_control_area { - u16 intercept_cr_read; - u16 intercept_cr_write; - u16 intercept_dr_read; - u16 intercept_dr_write; - u32 intercept_exceptions; - u64 intercept; - u8 reserved_1[44]; - u64 iopm_base_pa; - u64 msrpm_base_pa; - u64 tsc_offset; - u32 asid; - u8 tlb_ctl; - u8 reserved_2[3]; - u32 int_ctl; - u32 int_vector; - u32 int_state; - u8 reserved_3[4]; - u32 exit_code; - u32 exit_code_hi; - u64 exit_info_1; - u64 exit_info_2; - u32 exit_int_info; - u32 exit_int_info_err; - u64 nested_ctl; - u8 reserved_4[16]; - u32 event_inj; - u32 event_inj_err; - u64 nested_cr3; - u64 lbr_ctl; - u8 reserved_5[832]; -}; - - -#define TLB_CONTROL_DO_NOTHING 0 -#define TLB_CONTROL_FLUSH_ALL_ASID 1 - -#define V_TPR_MASK 0x0f - -#define V_IRQ_SHIFT 8 -#define V_IRQ_MASK (1 << V_IRQ_SHIFT) - -#define V_INTR_PRIO_SHIFT 16 -#define V_INTR_PRIO_MASK (0x0f << V_INTR_PRIO_SHIFT) - -#define V_IGN_TPR_SHIFT 20 -#define V_IGN_TPR_MASK (1 << V_IGN_TPR_SHIFT) - -#define V_INTR_MASKING_SHIFT 24 -#define V_INTR_MASKING_MASK (1 << V_INTR_MASKING_SHIFT) - -#define SVM_INTERRUPT_SHADOW_MASK 1 - -#define SVM_IOIO_STR_SHIFT 2 -#define SVM_IOIO_REP_SHIFT 3 -#define SVM_IOIO_SIZE_SHIFT 4 -#define SVM_IOIO_ASIZE_SHIFT 7 - -#define SVM_IOIO_TYPE_MASK 1 -#define SVM_IOIO_STR_MASK (1 << SVM_IOIO_STR_SHIFT) -#define SVM_IOIO_REP_MASK (1 << SVM_IOIO_REP_SHIFT) -#define SVM_IOIO_SIZE_MASK (7 << SVM_IOIO_SIZE_SHIFT) -#define SVM_IOIO_ASIZE_MASK (7 << SVM_IOIO_ASIZE_SHIFT) - -struct __attribute__ ((__packed__)) vmcb_seg { - u16 selector; - u16 attrib; - u32 limit; - u64 base; -}; - -struct __attribute__ ((__packed__)) vmcb_save_area { - struct vmcb_seg es; - struct vmcb_seg cs; - struct vmcb_seg ss; - struct vmcb_seg ds; - struct vmcb_seg fs; - struct vmcb_seg gs; - struct vmcb_seg gdtr; - struct vmcb_seg ldtr; - struct vmcb_seg idtr; - struct vmcb_seg tr; - u8 reserved_1[43]; - u8 cpl; - u8 reserved_2[4]; - u64 efer; - u8 reserved_3[112]; - u64 cr4; - u64 cr3; - u64 cr0; - u64 dr7; - u64 dr6; - u64 rflags; - u64 rip; - u8 reserved_4[88]; - u64 rsp; - u8 reserved_5[24]; - u64 rax; - u64 star; - u64 lstar; - u64 cstar; - u64 sfmask; - u64 kernel_gs_base; - u64 sysenter_cs; - u64 sysenter_esp; - u64 sysenter_eip; - u64 cr2; - u8 reserved_6[32]; - u64 g_pat; - u64 dbgctl; - u64 br_from; - u64 br_to; - u64 last_excp_from; - u64 last_excp_to; -}; - -struct __attribute__ ((__packed__)) vmcb { - struct vmcb_control_area control; - struct vmcb_save_area save; -}; - -#define SVM_CPUID_FEATURE_SHIFT 2 -#define SVM_CPUID_FUNC 0x8000000a - -#define MSR_EFER_SVME_MASK (1ULL << 12) -#define MSR_VM_CR 0xc0010114 -#define MSR_VM_HSAVE_PA 0xc0010117ULL - -#define SVM_VM_CR_SVM_DISABLE 4 - -#define SVM_SELECTOR_S_SHIFT 4 -#define SVM_SELECTOR_DPL_SHIFT 5 -#define SVM_SELECTOR_P_SHIFT 7 -#define SVM_SELECTOR_AVL_SHIFT 8 -#define SVM_SELECTOR_L_SHIFT 9 -#define SVM_SELECTOR_DB_SHIFT 10 -#define SVM_SELECTOR_G_SHIFT 11 - -#define SVM_SELECTOR_TYPE_MASK (0xf) -#define SVM_SELECTOR_S_MASK (1 << SVM_SELECTOR_S_SHIFT) -#define SVM_SELECTOR_DPL_MASK (3 << SVM_SELECTOR_DPL_SHIFT) -#define SVM_SELECTOR_P_MASK (1 << SVM_SELECTOR_P_SHIFT) -#define SVM_SELECTOR_AVL_MASK (1 << SVM_SELECTOR_AVL_SHIFT) -#define SVM_SELECTOR_L_MASK (1 << SVM_SELECTOR_L_SHIFT) -#define SVM_SELECTOR_DB_MASK (1 << SVM_SELECTOR_DB_SHIFT) -#define SVM_SELECTOR_G_MASK (1 << SVM_SELECTOR_G_SHIFT) - -#define SVM_SELECTOR_WRITE_MASK (1 << 1) -#define SVM_SELECTOR_READ_MASK SVM_SELECTOR_WRITE_MASK -#define SVM_SELECTOR_CODE_MASK (1 << 3) - -#define INTERCEPT_CR0_MASK 1 -#define INTERCEPT_CR3_MASK (1 << 3) -#define INTERCEPT_CR4_MASK (1 << 4) -#define INTERCEPT_CR8_MASK (1 << 8) - -#define INTERCEPT_DR0_MASK 1 -#define INTERCEPT_DR1_MASK (1 << 1) -#define INTERCEPT_DR2_MASK (1 << 2) -#define INTERCEPT_DR3_MASK (1 << 3) -#define INTERCEPT_DR4_MASK (1 << 4) -#define INTERCEPT_DR5_MASK (1 << 5) -#define INTERCEPT_DR6_MASK (1 << 6) -#define INTERCEPT_DR7_MASK (1 << 7) - -#define SVM_EVTINJ_VEC_MASK 0xff - -#define SVM_EVTINJ_TYPE_SHIFT 8 -#define SVM_EVTINJ_TYPE_MASK (7 << SVM_EVTINJ_TYPE_SHIFT) - -#define SVM_EVTINJ_TYPE_INTR (0 << SVM_EVTINJ_TYPE_SHIFT) -#define SVM_EVTINJ_TYPE_NMI (2 << SVM_EVTINJ_TYPE_SHIFT) -#define SVM_EVTINJ_TYPE_EXEPT (3 << SVM_EVTINJ_TYPE_SHIFT) -#define SVM_EVTINJ_TYPE_SOFT (4 << SVM_EVTINJ_TYPE_SHIFT) - -#define SVM_EVTINJ_VALID (1 << 31) -#define SVM_EVTINJ_VALID_ERR (1 << 11) - -#define SVM_EXITINTINFO_VEC_MASK SVM_EVTINJ_VEC_MASK - -#define SVM_EXITINTINFO_TYPE_INTR SVM_EVTINJ_TYPE_INTR -#define SVM_EXITINTINFO_TYPE_NMI SVM_EVTINJ_TYPE_NMI -#define SVM_EXITINTINFO_TYPE_EXEPT SVM_EVTINJ_TYPE_EXEPT -#define SVM_EXITINTINFO_TYPE_SOFT SVM_EVTINJ_TYPE_SOFT - -#define SVM_EXITINTINFO_VALID SVM_EVTINJ_VALID -#define SVM_EXITINTINFO_VALID_ERR SVM_EVTINJ_VALID_ERR - -#define SVM_EXITINFOSHIFT_TS_REASON_IRET 36 -#define SVM_EXITINFOSHIFT_TS_REASON_JMP 38 - -#define SVM_EXIT_READ_CR0 0x000 -#define SVM_EXIT_READ_CR3 0x003 -#define SVM_EXIT_READ_CR4 0x004 -#define SVM_EXIT_READ_CR8 0x008 -#define SVM_EXIT_WRITE_CR0 0x010 -#define SVM_EXIT_WRITE_CR3 0x013 -#define SVM_EXIT_WRITE_CR4 0x014 -#define SVM_EXIT_WRITE_CR8 0x018 -#define SVM_EXIT_READ_DR0 0x020 -#define SVM_EXIT_READ_DR1 0x021 -#define SVM_EXIT_READ_DR2 0x022 -#define SVM_EXIT_READ_DR3 0x023 -#define SVM_EXIT_READ_DR4 0x024 -#define SVM_EXIT_READ_DR5 0x025 -#define SVM_EXIT_READ_DR6 0x026 -#define SVM_EXIT_READ_DR7 0x027 -#define SVM_EXIT_WRITE_DR0 0x030 -#define SVM_EXIT_WRITE_DR1 0x031 -#define SVM_EXIT_WRITE_DR2 0x032 -#define SVM_EXIT_WRITE_DR3 0x033 -#define SVM_EXIT_WRITE_DR4 0x034 -#define SVM_EXIT_WRITE_DR5 0x035 -#define SVM_EXIT_WRITE_DR6 0x036 -#define SVM_EXIT_WRITE_DR7 0x037 -#define SVM_EXIT_EXCP_BASE 0x040 -#define SVM_EXIT_INTR 0x060 -#define SVM_EXIT_NMI 0x061 -#define SVM_EXIT_SMI 0x062 -#define SVM_EXIT_INIT 0x063 -#define SVM_EXIT_VINTR 0x064 -#define SVM_EXIT_CR0_SEL_WRITE 0x065 -#define SVM_EXIT_IDTR_READ 0x066 -#define SVM_EXIT_GDTR_READ 0x067 -#define SVM_EXIT_LDTR_READ 0x068 -#define SVM_EXIT_TR_READ 0x069 -#define SVM_EXIT_IDTR_WRITE 0x06a -#define SVM_EXIT_GDTR_WRITE 0x06b -#define SVM_EXIT_LDTR_WRITE 0x06c -#define SVM_EXIT_TR_WRITE 0x06d -#define SVM_EXIT_RDTSC 0x06e -#define SVM_EXIT_RDPMC 0x06f -#define SVM_EXIT_PUSHF 0x070 -#define SVM_EXIT_POPF 0x071 -#define SVM_EXIT_CPUID 0x072 -#define SVM_EXIT_RSM 0x073 -#define SVM_EXIT_IRET 0x074 -#define SVM_EXIT_SWINT 0x075 -#define SVM_EXIT_INVD 0x076 -#define SVM_EXIT_PAUSE 0x077 -#define SVM_EXIT_HLT 0x078 -#define SVM_EXIT_INVLPG 0x079 -#define SVM_EXIT_INVLPGA 0x07a -#define SVM_EXIT_IOIO 0x07b -#define SVM_EXIT_MSR 0x07c -#define SVM_EXIT_TASK_SWITCH 0x07d -#define SVM_EXIT_FERR_FREEZE 0x07e -#define SVM_EXIT_SHUTDOWN 0x07f -#define SVM_EXIT_VMRUN 0x080 -#define SVM_EXIT_VMMCALL 0x081 -#define SVM_EXIT_VMLOAD 0x082 -#define SVM_EXIT_VMSAVE 0x083 -#define SVM_EXIT_STGI 0x084 -#define SVM_EXIT_CLGI 0x085 -#define SVM_EXIT_SKINIT 0x086 -#define SVM_EXIT_RDTSCP 0x087 -#define SVM_EXIT_ICEBP 0x088 -#define SVM_EXIT_WBINVD 0x089 -#define SVM_EXIT_MONITOR 0x08a -#define SVM_EXIT_MWAIT 0x08b -#define SVM_EXIT_MWAIT_COND 0x08c -#define SVM_EXIT_NPF 0x400 - -#define SVM_EXIT_ERR -1 - -#define SVM_CR0_SELECTIVE_MASK (1 << 3 | 1) /* TS and MP */ - -#define SVM_VMLOAD ".byte 0x0f, 0x01, 0xda" -#define SVM_VMRUN ".byte 0x0f, 0x01, 0xd8" -#define SVM_VMSAVE ".byte 0x0f, 0x01, 0xdb" -#define SVM_CLGI ".byte 0x0f, 0x01, 0xdd" -#define SVM_STGI ".byte 0x0f, 0x01, 0xdc" -#define SVM_INVLPGA ".byte 0x0f, 0x01, 0xdf" - -#endif - -- cgit v1.2.3 From eca70fc5671b226966dfb7ee9953d59199288566 Mon Sep 17 00:00:00 2001 From: Eduardo Habkost Date: Mon, 17 Nov 2008 19:03:15 -0200 Subject: KVM: VMX: move ASM_VMX_* definitions from asm/kvm_host.h to asm/vmx.h Those definitions will be used by code outside KVM, so move it outside of a KVM-specific source file. Those definitions are used only on kvm/vmx.c, that already includes asm/vmx.h, so they can be moved safely. Signed-off-by: Eduardo Habkost Signed-off-by: Avi Kivity --- arch/x86/include/asm/kvm_host.h | 12 ------------ arch/x86/include/asm/vmx.h | 15 +++++++++++++++ 2 files changed, 15 insertions(+), 12 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index 99e3cc149d21..f58f7ebdea81 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -714,18 +714,6 @@ static inline void kvm_inject_gp(struct kvm_vcpu *vcpu, u32 error_code) kvm_queue_exception_e(vcpu, GP_VECTOR, error_code); } -#define ASM_VMX_VMCLEAR_RAX ".byte 0x66, 0x0f, 0xc7, 0x30" -#define ASM_VMX_VMLAUNCH ".byte 0x0f, 0x01, 0xc2" -#define ASM_VMX_VMRESUME ".byte 0x0f, 0x01, 0xc3" -#define ASM_VMX_VMPTRLD_RAX ".byte 0x0f, 0xc7, 0x30" -#define ASM_VMX_VMREAD_RDX_RAX ".byte 0x0f, 0x78, 0xd0" -#define ASM_VMX_VMWRITE_RAX_RDX ".byte 0x0f, 0x79, 0xd0" -#define ASM_VMX_VMWRITE_RSP_RDX ".byte 0x0f, 0x79, 0xd4" -#define ASM_VMX_VMXOFF ".byte 0x0f, 0x01, 0xc4" -#define ASM_VMX_VMXON_RAX ".byte 0xf3, 0x0f, 0xc7, 0x30" -#define ASM_VMX_INVEPT ".byte 0x66, 0x0f, 0x38, 0x80, 0x08" -#define ASM_VMX_INVVPID ".byte 0x66, 0x0f, 0x38, 0x81, 0x08" - #define MSR_IA32_TIME_STAMP_COUNTER 0x010 #define TSS_IOPB_BASE_OFFSET 0x66 diff --git a/arch/x86/include/asm/vmx.h b/arch/x86/include/asm/vmx.h index 3db236c26fa4..d0238e6151d8 100644 --- a/arch/x86/include/asm/vmx.h +++ b/arch/x86/include/asm/vmx.h @@ -364,4 +364,19 @@ enum vmcs_field { #define VMX_EPT_IDENTITY_PAGETABLE_ADDR 0xfffbc000ul + +#define ASM_VMX_VMCLEAR_RAX ".byte 0x66, 0x0f, 0xc7, 0x30" +#define ASM_VMX_VMLAUNCH ".byte 0x0f, 0x01, 0xc2" +#define ASM_VMX_VMRESUME ".byte 0x0f, 0x01, 0xc3" +#define ASM_VMX_VMPTRLD_RAX ".byte 0x0f, 0xc7, 0x30" +#define ASM_VMX_VMREAD_RDX_RAX ".byte 0x0f, 0x78, 0xd0" +#define ASM_VMX_VMWRITE_RAX_RDX ".byte 0x0f, 0x79, 0xd0" +#define ASM_VMX_VMWRITE_RSP_RDX ".byte 0x0f, 0x79, 0xd4" +#define ASM_VMX_VMXOFF ".byte 0x0f, 0x01, 0xc4" +#define ASM_VMX_VMXON_RAX ".byte 0xf3, 0x0f, 0xc7, 0x30" +#define ASM_VMX_INVEPT ".byte 0x66, 0x0f, 0x38, 0x80, 0x08" +#define ASM_VMX_INVVPID ".byte 0x66, 0x0f, 0x38, 0x81, 0x08" + + + #endif -- cgit v1.2.3 From 6210e37b122583643da335c0389f74098713e5ca Mon Sep 17 00:00:00 2001 From: Eduardo Habkost Date: Mon, 17 Nov 2008 19:03:16 -0200 Subject: KVM: VMX: move cpu_has_kvm_support() to an inline on asm/virtext.h It will be used by core code on kdump and reboot, to disable vmx if needed. Signed-off-by: Eduardo Habkost Signed-off-by: Avi Kivity --- arch/x86/include/asm/virtext.h | 31 +++++++++++++++++++++++++++++++ arch/x86/kvm/vmx.c | 4 ++-- 2 files changed, 33 insertions(+), 2 deletions(-) create mode 100644 arch/x86/include/asm/virtext.h (limited to 'arch/x86') diff --git a/arch/x86/include/asm/virtext.h b/arch/x86/include/asm/virtext.h new file mode 100644 index 000000000000..298b6a06110d --- /dev/null +++ b/arch/x86/include/asm/virtext.h @@ -0,0 +1,31 @@ +/* CPU virtualization extensions handling + * + * This should carry the code for handling CPU virtualization extensions + * that needs to live in the kernel core. + * + * Author: Eduardo Habkost + * + * Copyright (C) 2008, Red Hat Inc. + * + * Contains code from KVM, Copyright (C) 2006 Qumranet, Inc. + * + * This work is licensed under the terms of the GNU GPL, version 2. See + * the COPYING file in the top-level directory. + */ +#ifndef _ASM_X86_VIRTEX_H +#define _ASM_X86_VIRTEX_H + +#include +#include + +/* + * VMX functions: + */ + +static inline int cpu_has_vmx(void) +{ + unsigned long ecx = cpuid_ecx(1); + return test_bit(5, &ecx); /* CPUID.1:ECX.VMX[bit 5] -> VT */ +} + +#endif /* _ASM_X86_VIRTEX_H */ diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index ec71f6464cfa..defaeeb3f75c 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -31,6 +31,7 @@ #include #include #include +#include #define __ex(x) __kvm_handle_fault_on_reboot(x) @@ -1044,8 +1045,7 @@ static int vmx_get_irq(struct kvm_vcpu *vcpu) static __init int cpu_has_kvm_support(void) { - unsigned long ecx = cpuid_ecx(1); - return test_bit(5, &ecx); /* CPUID.1:ECX.VMX[bit 5] -> VT */ + return cpu_has_vmx(); } static __init int vmx_disabled_by_bios(void) -- cgit v1.2.3 From 1e9931146c748420343aeefadb3bb17bd1c14a37 Mon Sep 17 00:00:00 2001 From: Eduardo Habkost Date: Mon, 17 Nov 2008 19:03:17 -0200 Subject: x86: asm/virtext.h: add cpu_vmxoff() inline function Unfortunately we can't use exactly the same code from vmx hardware_disable(), because the KVM function uses the __kvm_handle_fault_on_reboot() tricks. Signed-off-by: Eduardo Habkost Signed-off-by: Avi Kivity --- arch/x86/include/asm/virtext.h | 15 +++++++++++++++ 1 file changed, 15 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/virtext.h b/arch/x86/include/asm/virtext.h index 298b6a06110d..7dee5b59930e 100644 --- a/arch/x86/include/asm/virtext.h +++ b/arch/x86/include/asm/virtext.h @@ -18,6 +18,8 @@ #include #include +#include + /* * VMX functions: */ @@ -28,4 +30,17 @@ static inline int cpu_has_vmx(void) return test_bit(5, &ecx); /* CPUID.1:ECX.VMX[bit 5] -> VT */ } + +/** Disable VMX on the current CPU + * + * vmxoff causes a undefined-opcode exception if vmxon was not run + * on the CPU previously. Only call this function if you know VMX + * is enabled. + */ +static inline void cpu_vmxoff(void) +{ + asm volatile (ASM_VMX_VMXOFF : : : "cc"); + write_cr4(read_cr4() & ~X86_CR4_VMXE); +} + #endif /* _ASM_X86_VIRTEX_H */ -- cgit v1.2.3 From 710ff4a855d0f3bf74b5b4a20328e2858a8a2968 Mon Sep 17 00:00:00 2001 From: Eduardo Habkost Date: Mon, 17 Nov 2008 19:03:18 -0200 Subject: KVM: VMX: extract kvm_cpu_vmxoff() from hardware_disable() Along with some comments on why it is different from the core cpu_vmxoff() function. Signed-off-by: Eduardo Habkost Signed-off-by: Avi Kivity --- arch/x86/kvm/vmx.c | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index defaeeb3f75c..f5958a7823f4 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -1091,13 +1091,22 @@ static void vmclear_local_vcpus(void) __vcpu_clear(vmx); } -static void hardware_disable(void *garbage) + +/* Just like cpu_vmxoff(), but with the __kvm_handle_fault_on_reboot() + * tricks. + */ +static void kvm_cpu_vmxoff(void) { - vmclear_local_vcpus(); asm volatile (__ex(ASM_VMX_VMXOFF) : : : "cc"); write_cr4(read_cr4() & ~X86_CR4_VMXE); } +static void hardware_disable(void *garbage) +{ + vmclear_local_vcpus(); + kvm_cpu_vmxoff(); +} + static __init int adjust_vmx_controls(u32 ctl_min, u32 ctl_opt, u32 msr, u32 *result) { -- cgit v1.2.3 From 6aa07a0d77f6aafbe69e4e8609ffaf2b7ee1b591 Mon Sep 17 00:00:00 2001 From: Eduardo Habkost Date: Mon, 17 Nov 2008 19:03:19 -0200 Subject: x86: cpu_emergency_vmxoff() function Add cpu_emergency_vmxoff() and its friends: cpu_vmx_enabled() and __cpu_emergency_vmxoff(). Signed-off-by: Eduardo Habkost Signed-off-by: Avi Kivity --- arch/x86/include/asm/virtext.h | 23 +++++++++++++++++++++++ 1 file changed, 23 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/virtext.h b/arch/x86/include/asm/virtext.h index 7dee5b59930e..6bcf0acb4ef1 100644 --- a/arch/x86/include/asm/virtext.h +++ b/arch/x86/include/asm/virtext.h @@ -43,4 +43,27 @@ static inline void cpu_vmxoff(void) write_cr4(read_cr4() & ~X86_CR4_VMXE); } +static inline int cpu_vmx_enabled(void) +{ + return read_cr4() & X86_CR4_VMXE; +} + +/** Disable VMX if it is enabled on the current CPU + * + * You shouldn't call this if cpu_has_vmx() returns 0. + */ +static inline void __cpu_emergency_vmxoff(void) +{ + if (cpu_vmx_enabled()) + cpu_vmxoff(); +} + +/** Disable VMX if it is supported and enabled on the current CPU + */ +static inline void cpu_emergency_vmxoff(void) +{ + if (cpu_has_vmx()) + __cpu_emergency_vmxoff(); +} + #endif /* _ASM_X86_VIRTEX_H */ -- cgit v1.2.3 From 63d1142f8f69e39468bc6079ab2239e902828134 Mon Sep 17 00:00:00 2001 From: Eduardo Habkost Date: Mon, 17 Nov 2008 19:03:20 -0200 Subject: KVM: SVM: move has_svm() code to asm/virtext.h Use a trick to keep the printk()s on has_svm() working as before. gcc will take care of not generating code for the 'msg' stuff when the function is called with a NULL msg argument. Signed-off-by: Eduardo Habkost Signed-off-by: Avi Kivity --- arch/x86/include/asm/virtext.h | 41 +++++++++++++++++++++++++++++++++++++++++ arch/x86/kvm/svm.c | 19 +++++-------------- 2 files changed, 46 insertions(+), 14 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/virtext.h b/arch/x86/include/asm/virtext.h index 6bcf0acb4ef1..6f0d409c3682 100644 --- a/arch/x86/include/asm/virtext.h +++ b/arch/x86/include/asm/virtext.h @@ -19,6 +19,7 @@ #include #include +#include /* * VMX functions: @@ -66,4 +67,44 @@ static inline void cpu_emergency_vmxoff(void) __cpu_emergency_vmxoff(); } + + + +/* + * SVM functions: + */ + +/** Check if the CPU has SVM support + * + * You can use the 'msg' arg to get a message describing the problem, + * if the function returns zero. Simply pass NULL if you are not interested + * on the messages; gcc should take care of not generating code for + * the messages on this case. + */ +static inline int cpu_has_svm(const char **msg) +{ + uint32_t eax, ebx, ecx, edx; + + if (boot_cpu_data.x86_vendor != X86_VENDOR_AMD) { + if (msg) + *msg = "not amd"; + return 0; + } + + cpuid(0x80000000, &eax, &ebx, &ecx, &edx); + if (eax < SVM_CPUID_FUNC) { + if (msg) + *msg = "can't execute cpuid_8000000a"; + return 0; + } + + cpuid(0x80000001, &eax, &ebx, &ecx, &edx); + if (!(ecx & (1 << SVM_CPUID_FEATURE_SHIFT))) { + if (msg) + *msg = "svm not available"; + return 0; + } + return 1; +} + #endif /* _ASM_X86_VIRTEX_H */ diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c index f0ad4d4217e4..0667c6d13d30 100644 --- a/arch/x86/kvm/svm.c +++ b/arch/x86/kvm/svm.c @@ -28,6 +28,8 @@ #include +#include + #define __ex(x) __kvm_handle_fault_on_reboot(x) MODULE_AUTHOR("Qumranet"); @@ -245,24 +247,13 @@ static void skip_emulated_instruction(struct kvm_vcpu *vcpu) static int has_svm(void) { - uint32_t eax, ebx, ecx, edx; - - if (boot_cpu_data.x86_vendor != X86_VENDOR_AMD) { - printk(KERN_INFO "has_svm: not amd\n"); - return 0; - } + const char *msg; - cpuid(0x80000000, &eax, &ebx, &ecx, &edx); - if (eax < SVM_CPUID_FUNC) { - printk(KERN_INFO "has_svm: can't execute cpuid_8000000a\n"); + if (!cpu_has_svm(&msg)) { + printk(KERN_INFO "has_svn: %s\n", msg); return 0; } - cpuid(0x80000001, &eax, &ebx, &ecx, &edx); - if (!(ecx & (1 << SVM_CPUID_FEATURE_SHIFT))) { - printk(KERN_DEBUG "has_svm: svm not available\n"); - return 0; - } return 1; } -- cgit v1.2.3 From 2c8dceebb238680d5577500f8283397d41ca5590 Mon Sep 17 00:00:00 2001 From: Eduardo Habkost Date: Mon, 17 Nov 2008 19:03:21 -0200 Subject: KVM: SVM: move svm_hardware_disable() code to asm/virtext.h Create cpu_svm_disable() function. Signed-off-by: Eduardo Habkost Signed-off-by: Avi Kivity --- arch/x86/include/asm/virtext.h | 14 ++++++++++++++ arch/x86/kvm/svm.c | 6 +----- 2 files changed, 15 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/virtext.h b/arch/x86/include/asm/virtext.h index 6f0d409c3682..2cfe363729c3 100644 --- a/arch/x86/include/asm/virtext.h +++ b/arch/x86/include/asm/virtext.h @@ -107,4 +107,18 @@ static inline int cpu_has_svm(const char **msg) return 1; } + +/** Disable SVM on the current CPU + * + * You should call this only if cpu_has_svm() returned true. + */ +static inline void cpu_svm_disable(void) +{ + uint64_t efer; + + wrmsrl(MSR_VM_HSAVE_PA, 0); + rdmsrl(MSR_EFER, efer); + wrmsrl(MSR_EFER, efer & ~MSR_EFER_SVME_MASK); +} + #endif /* _ASM_X86_VIRTEX_H */ diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c index 0667c6d13d30..1452851ae258 100644 --- a/arch/x86/kvm/svm.c +++ b/arch/x86/kvm/svm.c @@ -259,11 +259,7 @@ static int has_svm(void) static void svm_hardware_disable(void *garbage) { - uint64_t efer; - - wrmsrl(MSR_VM_HSAVE_PA, 0); - rdmsrl(MSR_EFER, efer); - wrmsrl(MSR_EFER, efer & ~MSR_EFER_SVME_MASK); + cpu_svm_disable(); } static void svm_hardware_enable(void *garbage) -- cgit v1.2.3 From 0f3e9eeba0ea212bbea88790729d054b700ab91e Mon Sep 17 00:00:00 2001 From: Eduardo Habkost Date: Mon, 17 Nov 2008 19:03:22 -0200 Subject: x86: cpu_emergency_svm_disable() function This function can be used by the reboot or kdump code to forcibly disable SVM on the CPU. Signed-off-by: Eduardo Habkost Signed-off-by: Avi Kivity --- arch/x86/include/asm/virtext.h | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/virtext.h b/arch/x86/include/asm/virtext.h index 2cfe363729c3..593636275238 100644 --- a/arch/x86/include/asm/virtext.h +++ b/arch/x86/include/asm/virtext.h @@ -121,4 +121,12 @@ static inline void cpu_svm_disable(void) wrmsrl(MSR_EFER, efer & ~MSR_EFER_SVME_MASK); } +/** Makes sure SVM is disabled, if it is supported on the CPU + */ +static inline void cpu_emergency_svm_disable(void) +{ + if (cpu_has_svm(NULL)) + cpu_svm_disable(); +} + #endif /* _ASM_X86_VIRTEX_H */ -- cgit v1.2.3 From 2340b62f77c782c305e6ae7748675a638436d1ef Mon Sep 17 00:00:00 2001 From: Eduardo Habkost Date: Mon, 17 Nov 2008 19:03:23 -0200 Subject: kdump: forcibly disable VMX and SVM on machine_crash_shutdown() We need to disable virtualization extensions on all CPUs before booting the kdump kernel, otherwise the kdump kernel booting will fail, and rebooting after the kdump kernel did its task may also fail. We do it using cpu_emergency_vmxoff() and cpu_emergency_svm_disable(), that should always work, because those functions check if the CPUs support SVM or VMX before doing their tasks. Signed-off-by: Eduardo Habkost Signed-off-by: Avi Kivity --- arch/x86/kernel/crash.c | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/crash.c b/arch/x86/kernel/crash.c index d84a852e4cd7..c689d19e35ab 100644 --- a/arch/x86/kernel/crash.c +++ b/arch/x86/kernel/crash.c @@ -26,6 +26,7 @@ #include #include #include +#include #include @@ -49,6 +50,15 @@ static void kdump_nmi_callback(int cpu, struct die_args *args) #endif crash_save_cpu(regs, cpu); + /* Disable VMX or SVM if needed. + * + * We need to disable virtualization on all CPUs. + * Having VMX or SVM enabled on any CPU may break rebooting + * after the kdump kernel has finished its task. + */ + cpu_emergency_vmxoff(); + cpu_emergency_svm_disable(); + disable_local_APIC(); } @@ -80,6 +90,14 @@ void native_machine_crash_shutdown(struct pt_regs *regs) local_irq_disable(); kdump_nmi_shootdown_cpus(); + + /* Booting kdump kernel with VMX or SVM enabled won't work, + * because (among other limitations) we can't disable paging + * with the virt flags. + */ + cpu_emergency_vmxoff(); + cpu_emergency_svm_disable(); + lapic_shutdown(); #if defined(CONFIG_X86_IO_APIC) disable_IO_APIC(); -- cgit v1.2.3 From d176720d34c72f7a8474a12204add93e54fe3ef1 Mon Sep 17 00:00:00 2001 From: Eduardo Habkost Date: Mon, 17 Nov 2008 19:03:24 -0200 Subject: x86: disable VMX on all CPUs on reboot On emergency_restart, we may need to use an NMI to disable virtualization on all CPUs. We do that using nmi_shootdown_cpus() if VMX is enabled. Note: With this patch, we will run the NMI stuff only when the CPU where emergency_restart() was called has VMX enabled. This should work on most cases because KVM enables VMX on all CPUs, but we may miss the small window where KVM is doing that. Also, I don't know if all code using VMX out there always enable VMX on all CPUs like KVM does. We have two other alternatives for that: a) Have an API that all code that enables VMX on any CPU should use to tell the kernel core that it is going to enable VMX on the CPUs. b) Always call nmi_shootdown_cpus() if the CPU supports VMX. This is a bit intrusive and more risky, as it would run nmi_shootdown_cpus() on emergency_reboot() even on systems where virtualization is never enabled. Finding a proper point to hook the nmi_shootdown_cpus() call isn't trivial, as the non-emergency machine_restart() (that doesn't need the NMI tricks) uses machine_emergency_restart() directly. The solution to make this work without adding a new function or argument to machine_ops was setting a 'reboot_emergency' flag that tells if native_machine_emergency_restart() needs to do the virt cleanup or not. Signed-off-by: Eduardo Habkost Signed-off-by: Avi Kivity --- arch/x86/kernel/reboot.c | 62 ++++++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 60 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c index 61f718df6eec..72e0e4e712d6 100644 --- a/arch/x86/kernel/reboot.c +++ b/arch/x86/kernel/reboot.c @@ -12,6 +12,7 @@ #include #include #include +#include #ifdef CONFIG_X86_32 # include @@ -39,6 +40,12 @@ int reboot_force; static int reboot_cpu = -1; #endif +/* This is set if we need to go through the 'emergency' path. + * When machine_emergency_restart() is called, we may be on + * an inconsistent state and won't be able to do a clean cleanup + */ +static int reboot_emergency; + /* This is set by the PCI code if either type 1 or type 2 PCI is detected */ bool port_cf9_safe = false; @@ -368,6 +375,48 @@ static inline void kb_wait(void) } } +static void vmxoff_nmi(int cpu, struct die_args *args) +{ + cpu_emergency_vmxoff(); +} + +/* Use NMIs as IPIs to tell all CPUs to disable virtualization + */ +static void emergency_vmx_disable_all(void) +{ + /* Just make sure we won't change CPUs while doing this */ + local_irq_disable(); + + /* We need to disable VMX on all CPUs before rebooting, otherwise + * we risk hanging up the machine, because the CPU ignore INIT + * signals when VMX is enabled. + * + * We can't take any locks and we may be on an inconsistent + * state, so we use NMIs as IPIs to tell the other CPUs to disable + * VMX and halt. + * + * For safety, we will avoid running the nmi_shootdown_cpus() + * stuff unnecessarily, but we don't have a way to check + * if other CPUs have VMX enabled. So we will call it only if the + * CPU we are running on has VMX enabled. + * + * We will miss cases where VMX is not enabled on all CPUs. This + * shouldn't do much harm because KVM always enable VMX on all + * CPUs anyway. But we can miss it on the small window where KVM + * is still enabling VMX. + */ + if (cpu_has_vmx() && cpu_vmx_enabled()) { + /* Disable VMX on this CPU. + */ + cpu_vmxoff(); + + /* Halt and disable VMX on the other CPUs */ + nmi_shootdown_cpus(vmxoff_nmi); + + } +} + + void __attribute__((weak)) mach_reboot_fixups(void) { } @@ -376,6 +425,9 @@ static void native_machine_emergency_restart(void) { int i; + if (reboot_emergency) + emergency_vmx_disable_all(); + /* Tell the BIOS if we want cold or warm reboot */ *((unsigned short *)__va(0x472)) = reboot_mode; @@ -482,13 +534,19 @@ void native_machine_shutdown(void) #endif } +static void __machine_emergency_restart(int emergency) +{ + reboot_emergency = emergency; + machine_ops.emergency_restart(); +} + static void native_machine_restart(char *__unused) { printk("machine restart\n"); if (!reboot_force) machine_shutdown(); - machine_emergency_restart(); + __machine_emergency_restart(0); } static void native_machine_halt(void) @@ -532,7 +590,7 @@ void machine_shutdown(void) void machine_emergency_restart(void) { - machine_ops.emergency_restart(); + __machine_emergency_restart(1); } void machine_restart(char *cmd) -- cgit v1.2.3 From df203ec9a77a7236cb90456664d714423b98a977 Mon Sep 17 00:00:00 2001 From: Avi Kivity Date: Sun, 23 Nov 2008 18:08:57 +0200 Subject: KVM: VMX: Conditionally request interrupt window after injecting irq If we're injecting an interrupt, and another one is pending, request an interrupt window notification so we don't have excess latency on the second interrupt. This shouldn't happen in practice since an EOI will be issued, giving a second chance to request an interrupt window, but... Signed-off-by: Avi Kivity --- arch/x86/kvm/vmx.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index f5958a7823f4..7ea485543cf8 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -3304,6 +3304,8 @@ static void vmx_intr_assist(struct kvm_vcpu *vcpu) if (vcpu->arch.interrupt.pending) { vmx_inject_irq(vcpu, vcpu->arch.interrupt.nr); kvm_timer_intr_post(vcpu, vcpu->arch.interrupt.nr); + if (kvm_cpu_has_interrupt(vcpu)) + enable_irq_window(vcpu); } } -- cgit v1.2.3 From 423cd25a5ade17b8a5cc85e6f0a0f37028d2c4a2 Mon Sep 17 00:00:00 2001 From: Glauber Costa Date: Mon, 24 Nov 2008 15:45:23 -0200 Subject: x86: KVM guest: sign kvmclock as paravirt Currently, we only set the KVM paravirt signature in case of CONFIG_KVM_GUEST. However, it is possible to have it turned off, while CONFIG_KVM_CLOCK is turned on. This is also a paravirt case, and should be shown accordingly. Signed-off-by: Glauber Costa Signed-off-by: Avi Kivity --- arch/x86/kernel/kvmclock.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/kvmclock.c b/arch/x86/kernel/kvmclock.c index e169ae9b6a62..b38e801014e3 100644 --- a/arch/x86/kernel/kvmclock.c +++ b/arch/x86/kernel/kvmclock.c @@ -194,5 +194,7 @@ void __init kvmclock_init(void) #endif kvm_get_preset_lpj(); clocksource_register(&kvm_clock); + pv_info.paravirt_enabled = 1; + pv_info.name = "KVM"; } } -- cgit v1.2.3 From ecc5589f19a52e7e6501fe449047b19087ae11bb Mon Sep 17 00:00:00 2001 From: Marcelo Tosatti Date: Tue, 25 Nov 2008 15:58:07 +0100 Subject: KVM: MMU: optimize set_spte for page sync The write protect verification in set_spte is unnecessary for page sync. Its guaranteed that, if the unsync spte was writable, the target page does not have a write protected shadow (if it had, the spte would have been write protected under mmu_lock by rmap_write_protect before). Same reasoning applies to mark_page_dirty: the gfn has been marked as dirty via the pagefault path. The cost of hash table and memslot lookups are quite significant if the workload is pagetable write intensive resulting in increased mmu_lock contention. Signed-off-by: Marcelo Tosatti Signed-off-by: Avi Kivity --- arch/x86/kvm/mmu.c | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c index fa3486d64078..dd20b199a7c0 100644 --- a/arch/x86/kvm/mmu.c +++ b/arch/x86/kvm/mmu.c @@ -1593,6 +1593,15 @@ static int set_spte(struct kvm_vcpu *vcpu, u64 *shadow_pte, spte |= PT_WRITABLE_MASK; + /* + * Optimization: for pte sync, if spte was writable the hash + * lookup is unnecessary (and expensive). Write protection + * is responsibility of mmu_get_page / kvm_sync_page. + * Same reasoning can be applied to dirty page accounting. + */ + if (!can_unsync && is_writeble_pte(*shadow_pte)) + goto set_pte; + if (mmu_need_write_protect(vcpu, gfn, can_unsync)) { pgprintk("%s: found shadow page for %lx, marking ro\n", __func__, gfn); -- cgit v1.2.3 From dda96d8f1b3de692cce09969ce28fe22e58e5acf Mon Sep 17 00:00:00 2001 From: Avi Kivity Date: Wed, 26 Nov 2008 15:14:10 +0200 Subject: KVM: x86 emulator: reduce duplication in one operand emulation thunks Signed-off-by: Avi Kivity --- arch/x86/kvm/x86_emulate.c | 66 ++++++++++++++++------------------------------ 1 file changed, 23 insertions(+), 43 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/x86_emulate.c b/arch/x86/kvm/x86_emulate.c index 8f60ace13874..5f87d3e59195 100644 --- a/arch/x86/kvm/x86_emulate.c +++ b/arch/x86/kvm/x86_emulate.c @@ -359,6 +359,12 @@ static u16 group2_table[] = { "andl %"_msk",%"_LO32 _tmp"; " \ "orl %"_LO32 _tmp",%"_sav"; " +#ifdef CONFIG_X86_64 +#define ON64(x) x +#else +#define ON64(x) +#endif + /* Raw emulation: instruction has two explicit operands. */ #define __emulate_2op_nobyte(_op,_src,_dst,_eflags,_wx,_wy,_lx,_ly,_qx,_qy) \ do { \ @@ -425,42 +431,27 @@ static u16 group2_table[] = { __emulate_2op_nobyte(_op, _src, _dst, _eflags, \ "w", "r", _LO32, "r", "", "r") -/* Instruction has only one explicit operand (no source operand). */ -#define emulate_1op(_op, _dst, _eflags) \ +#define __emulate_1op(_op, _dst, _eflags, _suffix) \ do { \ unsigned long _tmp; \ \ + __asm__ __volatile__ ( \ + _PRE_EFLAGS("0", "3", "2") \ + _op _suffix " %1; " \ + _POST_EFLAGS("0", "3", "2") \ + : "=m" (_eflags), "+m" ((_dst).val), \ + "=&r" (_tmp) \ + : "i" (EFLAGS_MASK)); \ + } while (0) + +/* Instruction has only one explicit operand (no source operand). */ +#define emulate_1op(_op, _dst, _eflags) \ + do { \ switch ((_dst).bytes) { \ - case 1: \ - __asm__ __volatile__ ( \ - _PRE_EFLAGS("0", "3", "2") \ - _op"b %1; " \ - _POST_EFLAGS("0", "3", "2") \ - : "=m" (_eflags), "=m" ((_dst).val), \ - "=&r" (_tmp) \ - : "i" (EFLAGS_MASK)); \ - break; \ - case 2: \ - __asm__ __volatile__ ( \ - _PRE_EFLAGS("0", "3", "2") \ - _op"w %1; " \ - _POST_EFLAGS("0", "3", "2") \ - : "=m" (_eflags), "=m" ((_dst).val), \ - "=&r" (_tmp) \ - : "i" (EFLAGS_MASK)); \ - break; \ - case 4: \ - __asm__ __volatile__ ( \ - _PRE_EFLAGS("0", "3", "2") \ - _op"l %1; " \ - _POST_EFLAGS("0", "3", "2") \ - : "=m" (_eflags), "=m" ((_dst).val), \ - "=&r" (_tmp) \ - : "i" (EFLAGS_MASK)); \ - break; \ - case 8: \ - __emulate_1op_8byte(_op, _dst, _eflags); \ - break; \ + case 1: __emulate_1op(_op, _dst, _eflags, "b"); break; \ + case 2: __emulate_1op(_op, _dst, _eflags, "w"); break; \ + case 4: __emulate_1op(_op, _dst, _eflags, "l"); break; \ + case 8: ON64(__emulate_1op(_op, _dst, _eflags, "q")); break; \ } \ } while (0) @@ -476,19 +467,8 @@ static u16 group2_table[] = { : _qy ((_src).val), "i" (EFLAGS_MASK)); \ } while (0) -#define __emulate_1op_8byte(_op, _dst, _eflags) \ - do { \ - __asm__ __volatile__ ( \ - _PRE_EFLAGS("0", "3", "2") \ - _op"q %1; " \ - _POST_EFLAGS("0", "3", "2") \ - : "=m" (_eflags), "=m" ((_dst).val), "=&r" (_tmp) \ - : "i" (EFLAGS_MASK)); \ - } while (0) - #elif defined(__i386__) #define __emulate_2op_8byte(_op, _src, _dst, _eflags, _qx, _qy) -#define __emulate_1op_8byte(_op, _dst, _eflags) #endif /* __i386__ */ /* Fetch next part of the instruction being emulated. */ -- cgit v1.2.3 From 6b7ad61ffb9ca110add6f7fb36cc8a4dd89696a4 Mon Sep 17 00:00:00 2001 From: Avi Kivity Date: Wed, 26 Nov 2008 15:30:45 +0200 Subject: KVM: x86 emulator: consolidate emulation of two operand instructions No need to repeat the same assembly block over and over. Signed-off-by: Avi Kivity --- arch/x86/kvm/x86_emulate.c | 79 ++++++++++++++++------------------------------ 1 file changed, 28 insertions(+), 51 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/x86_emulate.c b/arch/x86/kvm/x86_emulate.c index 5f87d3e59195..a11af6f74d68 100644 --- a/arch/x86/kvm/x86_emulate.c +++ b/arch/x86/kvm/x86_emulate.c @@ -365,49 +365,42 @@ static u16 group2_table[] = { #define ON64(x) #endif +#define ____emulate_2op(_op, _src, _dst, _eflags, _x, _y, _suffix) \ + do { \ + __asm__ __volatile__ ( \ + _PRE_EFLAGS("0", "4", "2") \ + _op _suffix " %"_x"3,%1; " \ + _POST_EFLAGS("0", "4", "2") \ + : "=m" (_eflags), "=m" ((_dst).val), \ + "=&r" (_tmp) \ + : _y ((_src).val), "i" (EFLAGS_MASK)); \ + } while (0); + + /* Raw emulation: instruction has two explicit operands. */ #define __emulate_2op_nobyte(_op,_src,_dst,_eflags,_wx,_wy,_lx,_ly,_qx,_qy) \ - do { \ - unsigned long _tmp; \ - \ - switch ((_dst).bytes) { \ - case 2: \ - __asm__ __volatile__ ( \ - _PRE_EFLAGS("0", "4", "2") \ - _op"w %"_wx"3,%1; " \ - _POST_EFLAGS("0", "4", "2") \ - : "=m" (_eflags), "=m" ((_dst).val), \ - "=&r" (_tmp) \ - : _wy ((_src).val), "i" (EFLAGS_MASK)); \ - break; \ - case 4: \ - __asm__ __volatile__ ( \ - _PRE_EFLAGS("0", "4", "2") \ - _op"l %"_lx"3,%1; " \ - _POST_EFLAGS("0", "4", "2") \ - : "=m" (_eflags), "=m" ((_dst).val), \ - "=&r" (_tmp) \ - : _ly ((_src).val), "i" (EFLAGS_MASK)); \ - break; \ - case 8: \ - __emulate_2op_8byte(_op, _src, _dst, \ - _eflags, _qx, _qy); \ - break; \ - } \ + do { \ + unsigned long _tmp; \ + \ + switch ((_dst).bytes) { \ + case 2: \ + ____emulate_2op(_op,_src,_dst,_eflags,_wx,_wy,"w"); \ + break; \ + case 4: \ + ____emulate_2op(_op,_src,_dst,_eflags,_lx,_ly,"l"); \ + break; \ + case 8: \ + ON64(____emulate_2op(_op,_src,_dst,_eflags,_qx,_qy,"q")); \ + break; \ + } \ } while (0) #define __emulate_2op(_op,_src,_dst,_eflags,_bx,_by,_wx,_wy,_lx,_ly,_qx,_qy) \ do { \ - unsigned long __tmp; \ + unsigned long _tmp; \ switch ((_dst).bytes) { \ case 1: \ - __asm__ __volatile__ ( \ - _PRE_EFLAGS("0", "4", "2") \ - _op"b %"_bx"3,%1; " \ - _POST_EFLAGS("0", "4", "2") \ - : "=m" (_eflags), "=m" ((_dst).val), \ - "=&r" (__tmp) \ - : _by ((_src).val), "i" (EFLAGS_MASK)); \ + ____emulate_2op(_op,_src,_dst,_eflags,_bx,_by,"b"); \ break; \ default: \ __emulate_2op_nobyte(_op, _src, _dst, _eflags, \ @@ -455,22 +448,6 @@ static u16 group2_table[] = { } \ } while (0) -/* Emulate an instruction with quadword operands (x86/64 only). */ -#if defined(CONFIG_X86_64) -#define __emulate_2op_8byte(_op, _src, _dst, _eflags, _qx, _qy) \ - do { \ - __asm__ __volatile__ ( \ - _PRE_EFLAGS("0", "4", "2") \ - _op"q %"_qx"3,%1; " \ - _POST_EFLAGS("0", "4", "2") \ - : "=m" (_eflags), "=m" ((_dst).val), "=&r" (_tmp) \ - : _qy ((_src).val), "i" (EFLAGS_MASK)); \ - } while (0) - -#elif defined(__i386__) -#define __emulate_2op_8byte(_op, _src, _dst, _eflags, _qx, _qy) -#endif /* __i386__ */ - /* Fetch next part of the instruction being emulated. */ #define insn_fetch(_type, _size, _eip) \ ({ unsigned long _x; \ -- cgit v1.2.3 From faa5a3ae39483aefc46a78299c811194f953af27 Mon Sep 17 00:00:00 2001 From: Avi Kivity Date: Thu, 27 Nov 2008 17:36:41 +0200 Subject: KVM: x86 emulator: Extract 'pop' sequence into a function Switch 'pop r/m' instruction to use the new function. Signed-off-by: Avi Kivity --- arch/x86/kvm/x86_emulate.c | 21 +++++++++++++++++---- 1 file changed, 17 insertions(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/x86_emulate.c b/arch/x86/kvm/x86_emulate.c index a11af6f74d68..2555762f4b42 100644 --- a/arch/x86/kvm/x86_emulate.c +++ b/arch/x86/kvm/x86_emulate.c @@ -1057,20 +1057,33 @@ static inline void emulate_push(struct x86_emulate_ctxt *ctxt) c->regs[VCPU_REGS_RSP]); } -static inline int emulate_grp1a(struct x86_emulate_ctxt *ctxt, - struct x86_emulate_ops *ops) +static int emulate_pop(struct x86_emulate_ctxt *ctxt, + struct x86_emulate_ops *ops) { struct decode_cache *c = &ctxt->decode; int rc; rc = ops->read_std(register_address(c, ss_base(ctxt), c->regs[VCPU_REGS_RSP]), - &c->dst.val, c->dst.bytes, ctxt->vcpu); + &c->src.val, c->src.bytes, ctxt->vcpu); if (rc != 0) return rc; - register_address_increment(c, &c->regs[VCPU_REGS_RSP], c->dst.bytes); + register_address_increment(c, &c->regs[VCPU_REGS_RSP], c->src.bytes); + return rc; +} +static inline int emulate_grp1a(struct x86_emulate_ctxt *ctxt, + struct x86_emulate_ops *ops) +{ + struct decode_cache *c = &ctxt->decode; + int rc; + + c->src.bytes = c->dst.bytes; + rc = emulate_pop(ctxt, ops); + if (rc != 0) + return rc; + c->dst.val = c->src.val; return 0; } -- cgit v1.2.3 From 781d0edc5fc5cfe7491a0c5081734e62f6dc66ee Mon Sep 17 00:00:00 2001 From: Avi Kivity Date: Thu, 27 Nov 2008 18:00:28 +0200 Subject: KVM: x86 emulator: allow pop from mmio Signed-off-by: Avi Kivity --- arch/x86/kvm/x86_emulate.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/x86_emulate.c b/arch/x86/kvm/x86_emulate.c index 2555762f4b42..70242f5f0964 100644 --- a/arch/x86/kvm/x86_emulate.c +++ b/arch/x86/kvm/x86_emulate.c @@ -1063,9 +1063,9 @@ static int emulate_pop(struct x86_emulate_ctxt *ctxt, struct decode_cache *c = &ctxt->decode; int rc; - rc = ops->read_std(register_address(c, ss_base(ctxt), - c->regs[VCPU_REGS_RSP]), - &c->src.val, c->src.bytes, ctxt->vcpu); + rc = ops->read_emulated(register_address(c, ss_base(ctxt), + c->regs[VCPU_REGS_RSP]), + &c->src.val, c->src.bytes, ctxt->vcpu); if (rc != 0) return rc; -- cgit v1.2.3 From 8a09b6877f3100207b3572e7e12ea796493fe914 Mon Sep 17 00:00:00 2001 From: Avi Kivity Date: Thu, 27 Nov 2008 18:06:33 +0200 Subject: KVM: x86 emulator: switch 'pop reg' instruction to emulate_pop() Signed-off-by: Avi Kivity --- arch/x86/kvm/x86_emulate.c | 11 ++++------- 1 file changed, 4 insertions(+), 7 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/x86_emulate.c b/arch/x86/kvm/x86_emulate.c index 70242f5f0964..702de9869c19 100644 --- a/arch/x86/kvm/x86_emulate.c +++ b/arch/x86/kvm/x86_emulate.c @@ -1389,14 +1389,11 @@ special_insn: break; case 0x58 ... 0x5f: /* pop reg */ pop_instruction: - if ((rc = ops->read_std(register_address(c, ss_base(ctxt), - c->regs[VCPU_REGS_RSP]), c->dst.ptr, - c->op_bytes, ctxt->vcpu)) != 0) + c->src.bytes = c->op_bytes; + rc = emulate_pop(ctxt, ops); + if (rc != 0) goto done; - - register_address_increment(c, &c->regs[VCPU_REGS_RSP], - c->op_bytes); - c->dst.type = OP_NONE; /* Disable writeback. */ + c->dst.val = c->src.val; break; case 0x63: /* movsxd */ if (ctxt->mode != X86EMUL_MODE_PROT64) -- cgit v1.2.3 From cf5de4f886116871c2ae2eee53524edd741a68ae Mon Sep 17 00:00:00 2001 From: Avi Kivity Date: Fri, 28 Nov 2008 00:14:07 +0200 Subject: KVM: x86 emulator: fix ret emulation 'ret' did not set the operand type or size for the destination, so writeback ignored it. Signed-off-by: Avi Kivity --- arch/x86/kvm/x86_emulate.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kvm/x86_emulate.c b/arch/x86/kvm/x86_emulate.c index 702de9869c19..72ae86b1b131 100644 --- a/arch/x86/kvm/x86_emulate.c +++ b/arch/x86/kvm/x86_emulate.c @@ -1650,7 +1650,9 @@ special_insn: emulate_grp2(ctxt); break; case 0xc3: /* ret */ + c->dst.type = OP_REG; c->dst.ptr = &c->eip; + c->dst.bytes = c->op_bytes; goto pop_instruction; case 0xc6 ... 0xc7: /* mov (sole member of Grp11) */ mov: -- cgit v1.2.3 From 2b48cc75b21431037d6f902b9d583b1aff198490 Mon Sep 17 00:00:00 2001 From: Avi Kivity Date: Sat, 29 Nov 2008 20:36:13 +0200 Subject: KVM: x86 emulator: fix popf emulation Set operand type and size to get correct writeback behavior. Signed-off-by: Avi Kivity --- arch/x86/kvm/x86_emulate.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kvm/x86_emulate.c b/arch/x86/kvm/x86_emulate.c index 72ae86b1b131..e8c87ccfe310 100644 --- a/arch/x86/kvm/x86_emulate.c +++ b/arch/x86/kvm/x86_emulate.c @@ -1552,7 +1552,9 @@ special_insn: emulate_push(ctxt); break; case 0x9d: /* popf */ + c->dst.type = OP_REG; c->dst.ptr = (unsigned long *) &ctxt->eflags; + c->dst.bytes = c->op_bytes; goto pop_instruction; case 0xa0 ... 0xa1: /* mov */ c->dst.ptr = (unsigned long *)&c->regs[VCPU_REGS_RAX]; -- cgit v1.2.3 From f3fd92fbdb7663bd889c136842afc3851351ea8f Mon Sep 17 00:00:00 2001 From: Avi Kivity Date: Sat, 29 Nov 2008 20:38:12 +0200 Subject: KVM: Remove extraneous semicolon after do/while Notices by Guillaume Thouvenin. Signed-off-by: Avi Kivity --- arch/x86/kvm/x86_emulate.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/x86_emulate.c b/arch/x86/kvm/x86_emulate.c index e8c87ccfe310..69b330ba0ad0 100644 --- a/arch/x86/kvm/x86_emulate.c +++ b/arch/x86/kvm/x86_emulate.c @@ -374,7 +374,7 @@ static u16 group2_table[] = { : "=m" (_eflags), "=m" ((_dst).val), \ "=&r" (_tmp) \ : _y ((_src).val), "i" (EFLAGS_MASK)); \ - } while (0); + } while (0) /* Raw emulation: instruction has two explicit operands. */ -- cgit v1.2.3 From efff9e538f6bfa8ee2ca03f7e9a55d98df115186 Mon Sep 17 00:00:00 2001 From: Hannes Eder Date: Fri, 28 Nov 2008 17:02:06 +0100 Subject: KVM: VMX: fix sparse warning Impact: make global function static arch/x86/kvm/vmx.c:134:3: warning: symbol 'vmx_capability' was not declared. Should it be static? Signed-off-by: Hannes Eder Signed-off-by: Avi Kivity --- arch/x86/kvm/vmx.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index 7ea485543cf8..e446f232588e 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -128,7 +128,7 @@ static struct vmcs_config { u32 vmentry_ctrl; } vmcs_config; -struct vmx_capability { +static struct vmx_capability { u32 ept; u32 vpid; } vmx_capability; -- cgit v1.2.3 From 45ed60b371aeae6ed80f7e9d594a5e6412edc176 Mon Sep 17 00:00:00 2001 From: Guillaume Thouvenin Date: Thu, 4 Dec 2008 14:25:38 +0100 Subject: KVM: x86 emulator: Extend the opcode descriptor Extend the opcode descriptor to 32 bits. This is needed by the introduction of a new Src2 operand type. Signed-off-by: Guillaume Thouvenin Signed-off-by: Avi Kivity --- arch/x86/kvm/x86_emulate.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/x86_emulate.c b/arch/x86/kvm/x86_emulate.c index 69b330ba0ad0..7a07ca46c8ae 100644 --- a/arch/x86/kvm/x86_emulate.c +++ b/arch/x86/kvm/x86_emulate.c @@ -76,7 +76,7 @@ enum { Group1A, Group3_Byte, Group3, Group4, Group5, Group7, }; -static u16 opcode_table[256] = { +static u32 opcode_table[256] = { /* 0x00 - 0x07 */ ByteOp | DstMem | SrcReg | ModRM, DstMem | SrcReg | ModRM, ByteOp | DstReg | SrcMem | ModRM, DstReg | SrcMem | ModRM, @@ -195,7 +195,7 @@ static u16 opcode_table[256] = { ImplicitOps, ImplicitOps, Group | Group4, Group | Group5, }; -static u16 twobyte_table[256] = { +static u32 twobyte_table[256] = { /* 0x00 - 0x0F */ 0, Group | GroupDual | Group7, 0, 0, 0, 0, ImplicitOps, 0, ImplicitOps, ImplicitOps, 0, 0, 0, ImplicitOps | ModRM, 0, 0, @@ -253,7 +253,7 @@ static u16 twobyte_table[256] = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }; -static u16 group_table[] = { +static u32 group_table[] = { [Group1_80*8] = ByteOp | DstMem | SrcImm | ModRM, ByteOp | DstMem | SrcImm | ModRM, ByteOp | DstMem | SrcImm | ModRM, ByteOp | DstMem | SrcImm | ModRM, @@ -297,7 +297,7 @@ static u16 group_table[] = { SrcMem16 | ModRM | Mov, SrcMem | ModRM | ByteOp, }; -static u16 group2_table[] = { +static u32 group2_table[] = { [Group7*8] = SrcNone | ModRM, 0, 0, 0, SrcNone | ModRM | DstMem | Mov, 0, -- cgit v1.2.3 From 0dc8d10f7d848b63c8d32cf6fd31ba7def792ac9 Mon Sep 17 00:00:00 2001 From: Guillaume Thouvenin Date: Thu, 4 Dec 2008 14:26:42 +0100 Subject: KVM: x86 emulator: add Src2 decode set Instruction like shld has three operands, so we need to add a Src2 decode set. We start with Src2None, Src2CL, and Src2ImmByte, Src2One to support shld/shrd and we will expand it later. Signed-off-by: Guillaume Thouvenin Signed-off-by: Avi Kivity --- arch/x86/include/asm/kvm_x86_emulate.h | 1 + arch/x86/kvm/x86_emulate.c | 29 +++++++++++++++++++++++++++++ 2 files changed, 30 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/kvm_x86_emulate.h b/arch/x86/include/asm/kvm_x86_emulate.h index 16a002655f31..6a159732881a 100644 --- a/arch/x86/include/asm/kvm_x86_emulate.h +++ b/arch/x86/include/asm/kvm_x86_emulate.h @@ -123,6 +123,7 @@ struct decode_cache { u8 ad_bytes; u8 rex_prefix; struct operand src; + struct operand src2; struct operand dst; bool has_seg_override; u8 seg_override; diff --git a/arch/x86/kvm/x86_emulate.c b/arch/x86/kvm/x86_emulate.c index 7a07ca46c8ae..7f5cd62362c5 100644 --- a/arch/x86/kvm/x86_emulate.c +++ b/arch/x86/kvm/x86_emulate.c @@ -70,6 +70,12 @@ #define Group (1<<14) /* Bits 3:5 of modrm byte extend opcode */ #define GroupDual (1<<15) /* Alternate decoding of mod == 3 */ #define GroupMask 0xff /* Group number stored in bits 0:7 */ +/* Source 2 operand type */ +#define Src2None (0<<29) +#define Src2CL (1<<29) +#define Src2ImmByte (2<<29) +#define Src2One (3<<29) +#define Src2Mask (7<<29) enum { Group1_80, Group1_81, Group1_82, Group1_83, @@ -1000,6 +1006,29 @@ done_prefixes: break; } + /* + * Decode and fetch the second source operand: register, memory + * or immediate. + */ + switch (c->d & Src2Mask) { + case Src2None: + break; + case Src2CL: + c->src2.bytes = 1; + c->src2.val = c->regs[VCPU_REGS_RCX] & 0x8; + break; + case Src2ImmByte: + c->src2.type = OP_IMM; + c->src2.ptr = (unsigned long *)c->eip; + c->src2.bytes = 1; + c->src2.val = insn_fetch(u8, 1, c->eip); + break; + case Src2One: + c->src2.bytes = 1; + c->src2.val = 1; + break; + } + /* Decode and fetch the destination operand: register or memory. */ switch (c->d & DstMask) { case ImplicitOps: -- cgit v1.2.3 From bfcadf83ec5aafe600e73dd427d997db7bcc1d12 Mon Sep 17 00:00:00 2001 From: Guillaume Thouvenin Date: Thu, 4 Dec 2008 14:27:38 +0100 Subject: KVM: x86 emulator: add a new "implied 1" Src decode type Add SrcOne operand type when we need to decode an implied '1' like with regular shift instruction Signed-off-by: Guillaume Thouvenin Signed-off-by: Avi Kivity --- arch/x86/kvm/x86_emulate.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kvm/x86_emulate.c b/arch/x86/kvm/x86_emulate.c index 7f5cd62362c5..0c75306e7a07 100644 --- a/arch/x86/kvm/x86_emulate.c +++ b/arch/x86/kvm/x86_emulate.c @@ -58,6 +58,7 @@ #define SrcMem32 (4<<4) /* Memory operand (32-bit). */ #define SrcImm (5<<4) /* Immediate operand. */ #define SrcImmByte (6<<4) /* 8-bit sign-extended immediate operand. */ +#define SrcOne (7<<4) /* Implied '1' */ #define SrcMask (7<<4) /* Generic ModRM decode. */ #define ModRM (1<<7) @@ -1004,6 +1005,10 @@ done_prefixes: c->src.bytes = 1; c->src.val = insn_fetch(s8, 1, c->eip); break; + case SrcOne: + c->src.bytes = 1; + c->src.val = 1; + break; } /* -- cgit v1.2.3 From d175226a5f54817ba427368c6b739aefa7780fb2 Mon Sep 17 00:00:00 2001 From: Guillaume Thouvenin Date: Thu, 4 Dec 2008 14:29:00 +0100 Subject: KVM: x86 emulator: add the assembler code for three operands Add the assembler code for instruction with three operands and one operand is stored in ECX register Signed-off-by: Guillaume Thouvenin Signed-off-by: Avi Kivity --- arch/x86/kvm/x86_emulate.c | 39 +++++++++++++++++++++++++++++++++++++++ 1 file changed, 39 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kvm/x86_emulate.c b/arch/x86/kvm/x86_emulate.c index 0c75306e7a07..9ae6d5b3e962 100644 --- a/arch/x86/kvm/x86_emulate.c +++ b/arch/x86/kvm/x86_emulate.c @@ -431,6 +431,45 @@ static u32 group2_table[] = { __emulate_2op_nobyte(_op, _src, _dst, _eflags, \ "w", "r", _LO32, "r", "", "r") +/* Instruction has three operands and one operand is stored in ECX register */ +#define __emulate_2op_cl(_op, _cl, _src, _dst, _eflags, _suffix, _type) \ + do { \ + unsigned long _tmp; \ + _type _clv = (_cl).val; \ + _type _srcv = (_src).val; \ + _type _dstv = (_dst).val; \ + \ + __asm__ __volatile__ ( \ + _PRE_EFLAGS("0", "5", "2") \ + _op _suffix " %4,%1 \n" \ + _POST_EFLAGS("0", "5", "2") \ + : "=m" (_eflags), "+r" (_dstv), "=&r" (_tmp) \ + : "c" (_clv) , "r" (_srcv), "i" (EFLAGS_MASK) \ + ); \ + \ + (_cl).val = (unsigned long) _clv; \ + (_src).val = (unsigned long) _srcv; \ + (_dst).val = (unsigned long) _dstv; \ + } while (0) + +#define emulate_2op_cl(_op, _cl, _src, _dst, _eflags) \ + do { \ + switch ((_dst).bytes) { \ + case 2: \ + __emulate_2op_cl(_op, _cl, _src, _dst, _eflags, \ + "w", unsigned short); \ + break; \ + case 4: \ + __emulate_2op_cl(_op, _cl, _src, _dst, _eflags, \ + "l", unsigned int); \ + break; \ + case 8: \ + ON64(__emulate_2op_cl(_op, _cl, _src, _dst, _eflags, \ + "q", unsigned long)); \ + break; \ + } \ + } while (0) + #define __emulate_1op(_op, _dst, _eflags, _suffix) \ do { \ unsigned long _tmp; \ -- cgit v1.2.3 From 9bf8ea42fe22d7d1c48044148fa658cb9083d49c Mon Sep 17 00:00:00 2001 From: Guillaume Thouvenin Date: Thu, 4 Dec 2008 14:30:13 +0100 Subject: KVM: x86 emulator: add the emulation of shld and shrd instructions Add emulation of shld and shrd instructions Signed-off-by: Guillaume Thouvenin Signed-off-by: Avi Kivity --- arch/x86/kvm/x86_emulate.c | 17 +++++++++++++++-- 1 file changed, 15 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/x86_emulate.c b/arch/x86/kvm/x86_emulate.c index 9ae6d5b3e962..219dc3110bf1 100644 --- a/arch/x86/kvm/x86_emulate.c +++ b/arch/x86/kvm/x86_emulate.c @@ -237,9 +237,14 @@ static u32 twobyte_table[256] = { /* 0x90 - 0x9F */ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 0xA0 - 0xA7 */ - 0, 0, 0, DstMem | SrcReg | ModRM | BitOp, 0, 0, 0, 0, + 0, 0, 0, DstMem | SrcReg | ModRM | BitOp, + DstMem | SrcReg | Src2ImmByte | ModRM, + DstMem | SrcReg | Src2CL | ModRM, 0, 0, /* 0xA8 - 0xAF */ - 0, 0, 0, DstMem | SrcReg | ModRM | BitOp, 0, 0, ModRM, 0, + 0, 0, 0, DstMem | SrcReg | ModRM | BitOp, + DstMem | SrcReg | Src2ImmByte | ModRM, + DstMem | SrcReg | Src2CL | ModRM, + ModRM, 0, /* 0xB0 - 0xB7 */ ByteOp | DstMem | SrcReg | ModRM, DstMem | SrcReg | ModRM, 0, DstMem | SrcReg | ModRM | BitOp, @@ -2037,12 +2042,20 @@ twobyte_insn: c->src.val &= (c->dst.bytes << 3) - 1; emulate_2op_SrcV_nobyte("bt", c->src, c->dst, ctxt->eflags); break; + case 0xa4: /* shld imm8, r, r/m */ + case 0xa5: /* shld cl, r, r/m */ + emulate_2op_cl("shld", c->src2, c->src, c->dst, ctxt->eflags); + break; case 0xab: bts: /* bts */ /* only subword offset */ c->src.val &= (c->dst.bytes << 3) - 1; emulate_2op_SrcV_nobyte("bts", c->src, c->dst, ctxt->eflags); break; + case 0xac: /* shrd imm8, r, r/m */ + case 0xad: /* shrd cl, r, r/m */ + emulate_2op_cl("shrd", c->src2, c->src, c->dst, ctxt->eflags); + break; case 0xae: /* clflush */ break; case 0xb0 ... 0xb1: /* cmpxchg */ -- cgit v1.2.3 From fbce554e940a983d005e29849636d0ef54b3eb18 Mon Sep 17 00:00:00 2001 From: Amit Shah Date: Thu, 4 Dec 2008 11:11:40 +0000 Subject: KVM: x86 emulator: Fix handling of VMMCALL instruction The VMMCALL instruction doesn't get recognised and isn't processed by the emulator. This is seen on an Intel host that tries to execute the VMMCALL instruction after a guest live migrates from an AMD host. Signed-off-by: Amit Shah Signed-off-by: Avi Kivity --- arch/x86/kvm/x86_emulate.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/x86_emulate.c b/arch/x86/kvm/x86_emulate.c index 219dc3110bf1..d174db7a3370 100644 --- a/arch/x86/kvm/x86_emulate.c +++ b/arch/x86/kvm/x86_emulate.c @@ -311,7 +311,7 @@ static u32 group_table[] = { static u32 group2_table[] = { [Group7*8] = - SrcNone | ModRM, 0, 0, 0, + SrcNone | ModRM, 0, 0, SrcNone | ModRM, SrcNone | ModRM | DstMem | Mov, 0, SrcMem16 | ModRM | Mov, 0, }; -- cgit v1.2.3 From 60c8aec6e2c9923492dabbd6b67e34692bd26c20 Mon Sep 17 00:00:00 2001 From: Marcelo Tosatti Date: Mon, 1 Dec 2008 22:32:02 -0200 Subject: KVM: MMU: use page array in unsync walk Instead of invoking the handler directly collect pages into an array so the caller can work with it. Simplifies TLB flush collapsing. Signed-off-by: Marcelo Tosatti Signed-off-by: Avi Kivity --- arch/x86/include/asm/kvm_host.h | 2 +- arch/x86/kvm/mmu.c | 195 ++++++++++++++++++++++++++++------------ 2 files changed, 141 insertions(+), 56 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index f58f7ebdea81..93d0aed35880 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -200,7 +200,7 @@ struct kvm_mmu_page { int multimapped; /* More than one parent_pte? */ int root_count; /* Currently serving as active root */ bool unsync; - bool unsync_children; + unsigned int unsync_children; union { u64 *parent_pte; /* !multimapped */ struct hlist_head parent_ptes; /* multimapped, kvm_pte_chain */ diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c index dd20b199a7c0..7ce92f78f337 100644 --- a/arch/x86/kvm/mmu.c +++ b/arch/x86/kvm/mmu.c @@ -908,8 +908,9 @@ static void kvm_mmu_update_unsync_bitmap(u64 *spte) struct kvm_mmu_page *sp = page_header(__pa(spte)); index = spte - sp->spt; - __set_bit(index, sp->unsync_child_bitmap); - sp->unsync_children = 1; + if (!__test_and_set_bit(index, sp->unsync_child_bitmap)) + sp->unsync_children++; + WARN_ON(!sp->unsync_children); } static void kvm_mmu_update_parents_unsync(struct kvm_mmu_page *sp) @@ -936,7 +937,6 @@ static void kvm_mmu_update_parents_unsync(struct kvm_mmu_page *sp) static int unsync_walk_fn(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp) { - sp->unsync_children = 1; kvm_mmu_update_parents_unsync(sp); return 1; } @@ -967,18 +967,41 @@ static void nonpaging_invlpg(struct kvm_vcpu *vcpu, gva_t gva) { } +#define KVM_PAGE_ARRAY_NR 16 + +struct kvm_mmu_pages { + struct mmu_page_and_offset { + struct kvm_mmu_page *sp; + unsigned int idx; + } page[KVM_PAGE_ARRAY_NR]; + unsigned int nr; +}; + #define for_each_unsync_children(bitmap, idx) \ for (idx = find_first_bit(bitmap, 512); \ idx < 512; \ idx = find_next_bit(bitmap, 512, idx+1)) -static int mmu_unsync_walk(struct kvm_mmu_page *sp, - struct kvm_unsync_walk *walker) +int mmu_pages_add(struct kvm_mmu_pages *pvec, struct kvm_mmu_page *sp, + int idx) { - int i, ret; + int i; - if (!sp->unsync_children) - return 0; + if (sp->unsync) + for (i=0; i < pvec->nr; i++) + if (pvec->page[i].sp == sp) + return 0; + + pvec->page[pvec->nr].sp = sp; + pvec->page[pvec->nr].idx = idx; + pvec->nr++; + return (pvec->nr == KVM_PAGE_ARRAY_NR); +} + +static int __mmu_unsync_walk(struct kvm_mmu_page *sp, + struct kvm_mmu_pages *pvec) +{ + int i, ret, nr_unsync_leaf = 0; for_each_unsync_children(sp->unsync_child_bitmap, i) { u64 ent = sp->spt[i]; @@ -988,17 +1011,22 @@ static int mmu_unsync_walk(struct kvm_mmu_page *sp, child = page_header(ent & PT64_BASE_ADDR_MASK); if (child->unsync_children) { - ret = mmu_unsync_walk(child, walker); - if (ret) + if (mmu_pages_add(pvec, child, i)) + return -ENOSPC; + + ret = __mmu_unsync_walk(child, pvec); + if (!ret) + __clear_bit(i, sp->unsync_child_bitmap); + else if (ret > 0) + nr_unsync_leaf += ret; + else return ret; - __clear_bit(i, sp->unsync_child_bitmap); } if (child->unsync) { - ret = walker->entry(child, walker); - __clear_bit(i, sp->unsync_child_bitmap); - if (ret) - return ret; + nr_unsync_leaf++; + if (mmu_pages_add(pvec, child, i)) + return -ENOSPC; } } } @@ -1006,7 +1034,17 @@ static int mmu_unsync_walk(struct kvm_mmu_page *sp, if (find_first_bit(sp->unsync_child_bitmap, 512) == 512) sp->unsync_children = 0; - return 0; + return nr_unsync_leaf; +} + +static int mmu_unsync_walk(struct kvm_mmu_page *sp, + struct kvm_mmu_pages *pvec) +{ + if (!sp->unsync_children) + return 0; + + mmu_pages_add(pvec, sp, 0); + return __mmu_unsync_walk(sp, pvec); } static struct kvm_mmu_page *kvm_mmu_lookup_page(struct kvm *kvm, gfn_t gfn) @@ -1056,30 +1094,81 @@ static int kvm_sync_page(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp) return 0; } -struct sync_walker { - struct kvm_vcpu *vcpu; - struct kvm_unsync_walk walker; +struct mmu_page_path { + struct kvm_mmu_page *parent[PT64_ROOT_LEVEL-1]; + unsigned int idx[PT64_ROOT_LEVEL-1]; }; -static int mmu_sync_fn(struct kvm_mmu_page *sp, struct kvm_unsync_walk *walk) +#define for_each_sp(pvec, sp, parents, i) \ + for (i = mmu_pages_next(&pvec, &parents, -1), \ + sp = pvec.page[i].sp; \ + i < pvec.nr && ({ sp = pvec.page[i].sp; 1;}); \ + i = mmu_pages_next(&pvec, &parents, i)) + +int mmu_pages_next(struct kvm_mmu_pages *pvec, struct mmu_page_path *parents, + int i) { - struct sync_walker *sync_walk = container_of(walk, struct sync_walker, - walker); - struct kvm_vcpu *vcpu = sync_walk->vcpu; + int n; + + for (n = i+1; n < pvec->nr; n++) { + struct kvm_mmu_page *sp = pvec->page[n].sp; + + if (sp->role.level == PT_PAGE_TABLE_LEVEL) { + parents->idx[0] = pvec->page[n].idx; + return n; + } - kvm_sync_page(vcpu, sp); - return (need_resched() || spin_needbreak(&vcpu->kvm->mmu_lock)); + parents->parent[sp->role.level-2] = sp; + parents->idx[sp->role.level-1] = pvec->page[n].idx; + } + + return n; } -static void mmu_sync_children(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp) +void mmu_pages_clear_parents(struct mmu_page_path *parents) { - struct sync_walker walker = { - .walker = { .entry = mmu_sync_fn, }, - .vcpu = vcpu, - }; + struct kvm_mmu_page *sp; + unsigned int level = 0; + + do { + unsigned int idx = parents->idx[level]; + + sp = parents->parent[level]; + if (!sp) + return; + + --sp->unsync_children; + WARN_ON((int)sp->unsync_children < 0); + __clear_bit(idx, sp->unsync_child_bitmap); + level++; + } while (level < PT64_ROOT_LEVEL-1 && !sp->unsync_children); +} + +static void kvm_mmu_pages_init(struct kvm_mmu_page *parent, + struct mmu_page_path *parents, + struct kvm_mmu_pages *pvec) +{ + parents->parent[parent->role.level-1] = NULL; + pvec->nr = 0; +} - while (mmu_unsync_walk(sp, &walker.walker)) +static void mmu_sync_children(struct kvm_vcpu *vcpu, + struct kvm_mmu_page *parent) +{ + int i; + struct kvm_mmu_page *sp; + struct mmu_page_path parents; + struct kvm_mmu_pages pages; + + kvm_mmu_pages_init(parent, &parents, &pages); + while (mmu_unsync_walk(parent, &pages)) { + for_each_sp(pages, sp, parents, i) { + kvm_sync_page(vcpu, sp); + mmu_pages_clear_parents(&parents); + } cond_resched_lock(&vcpu->kvm->mmu_lock); + kvm_mmu_pages_init(parent, &parents, &pages); + } } static struct kvm_mmu_page *kvm_mmu_get_page(struct kvm_vcpu *vcpu, @@ -1245,33 +1334,29 @@ static void kvm_mmu_unlink_parents(struct kvm *kvm, struct kvm_mmu_page *sp) } } -struct zap_walker { - struct kvm_unsync_walk walker; - struct kvm *kvm; - int zapped; -}; - -static int mmu_zap_fn(struct kvm_mmu_page *sp, struct kvm_unsync_walk *walk) -{ - struct zap_walker *zap_walk = container_of(walk, struct zap_walker, - walker); - kvm_mmu_zap_page(zap_walk->kvm, sp); - zap_walk->zapped = 1; - return 0; -} - -static int mmu_zap_unsync_children(struct kvm *kvm, struct kvm_mmu_page *sp) +static int mmu_zap_unsync_children(struct kvm *kvm, + struct kvm_mmu_page *parent) { - struct zap_walker walker = { - .walker = { .entry = mmu_zap_fn, }, - .kvm = kvm, - .zapped = 0, - }; + int i, zapped = 0; + struct mmu_page_path parents; + struct kvm_mmu_pages pages; - if (sp->role.level == PT_PAGE_TABLE_LEVEL) + if (parent->role.level == PT_PAGE_TABLE_LEVEL) return 0; - mmu_unsync_walk(sp, &walker.walker); - return walker.zapped; + + kvm_mmu_pages_init(parent, &parents, &pages); + while (mmu_unsync_walk(parent, &pages)) { + struct kvm_mmu_page *sp; + + for_each_sp(pages, sp, parents, i) { + kvm_mmu_zap_page(kvm, sp); + mmu_pages_clear_parents(&parents); + } + zapped += pages.nr; + kvm_mmu_pages_init(parent, &parents, &pages); + } + + return zapped; } static int kvm_mmu_zap_page(struct kvm *kvm, struct kvm_mmu_page *sp) -- cgit v1.2.3 From b1a368218ad5b6e62380c8f206f16e6f18bf154c Mon Sep 17 00:00:00 2001 From: Marcelo Tosatti Date: Mon, 1 Dec 2008 22:32:03 -0200 Subject: KVM: MMU: collapse remote TLB flushes on root sync Collapse remote TLB flushes on root sync. kernbench is 2.7% faster on 4-way guest. Improvements have been seen with other loads such as AIM7. Signed-off-by: Marcelo Tosatti Signed-off-by: Avi Kivity --- arch/x86/kvm/mmu.c | 19 ++++++++++++++----- 1 file changed, 14 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c index 7ce92f78f337..58c35dead321 100644 --- a/arch/x86/kvm/mmu.c +++ b/arch/x86/kvm/mmu.c @@ -621,7 +621,7 @@ static u64 *rmap_next(struct kvm *kvm, unsigned long *rmapp, u64 *spte) return NULL; } -static void rmap_write_protect(struct kvm *kvm, u64 gfn) +static int rmap_write_protect(struct kvm *kvm, u64 gfn) { unsigned long *rmapp; u64 *spte; @@ -667,8 +667,7 @@ static void rmap_write_protect(struct kvm *kvm, u64 gfn) spte = rmap_next(kvm, rmapp, spte); } - if (write_protected) - kvm_flush_remote_tlbs(kvm); + return write_protected; } static int kvm_unmap_rmapp(struct kvm *kvm, unsigned long *rmapp) @@ -1083,7 +1082,8 @@ static int kvm_sync_page(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp) return 1; } - rmap_write_protect(vcpu->kvm, sp->gfn); + if (rmap_write_protect(vcpu->kvm, sp->gfn)) + kvm_flush_remote_tlbs(vcpu->kvm); kvm_unlink_unsync_page(vcpu->kvm, sp); if (vcpu->arch.mmu.sync_page(vcpu, sp)) { kvm_mmu_zap_page(vcpu->kvm, sp); @@ -1162,6 +1162,14 @@ static void mmu_sync_children(struct kvm_vcpu *vcpu, kvm_mmu_pages_init(parent, &parents, &pages); while (mmu_unsync_walk(parent, &pages)) { + int protected = 0; + + for_each_sp(pages, sp, parents, i) + protected |= rmap_write_protect(vcpu->kvm, sp->gfn); + + if (protected) + kvm_flush_remote_tlbs(vcpu->kvm); + for_each_sp(pages, sp, parents, i) { kvm_sync_page(vcpu, sp); mmu_pages_clear_parents(&parents); @@ -1226,7 +1234,8 @@ static struct kvm_mmu_page *kvm_mmu_get_page(struct kvm_vcpu *vcpu, sp->role = role; hlist_add_head(&sp->hash_link, bucket); if (!metaphysical) { - rmap_write_protect(vcpu->kvm, gfn); + if (rmap_write_protect(vcpu->kvm, gfn)) + kvm_flush_remote_tlbs(vcpu->kvm); account_shadowed(vcpu->kvm, gfn); } if (shadow_trap_nonpresent_pte != shadow_notrap_nonpresent_pte) -- cgit v1.2.3 From 6cffe8ca4a2adf1ac5003d9cad08fe4434d6eee0 Mon Sep 17 00:00:00 2001 From: Marcelo Tosatti Date: Mon, 1 Dec 2008 22:32:04 -0200 Subject: KVM: MMU: skip global pgtables on sync due to cr3 switch Skip syncing global pages on cr3 switch (but not on cr4/cr0). This is important for Linux 32-bit guests with PAE, where the kmap page is marked as global. Signed-off-by: Marcelo Tosatti Signed-off-by: Avi Kivity --- arch/x86/include/asm/kvm_host.h | 6 +++++ arch/x86/kvm/mmu.c | 53 ++++++++++++++++++++++++++++++++++++----- arch/x86/kvm/paging_tmpl.h | 10 ++++---- arch/x86/kvm/x86.c | 4 ++++ 4 files changed, 63 insertions(+), 10 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index 93d0aed35880..65b1ed295698 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -182,6 +182,8 @@ struct kvm_mmu_page { struct list_head link; struct hlist_node hash_link; + struct list_head oos_link; + /* * The following two entries are used to key the shadow page in the * hash table. @@ -200,6 +202,7 @@ struct kvm_mmu_page { int multimapped; /* More than one parent_pte? */ int root_count; /* Currently serving as active root */ bool unsync; + bool global; unsigned int unsync_children; union { u64 *parent_pte; /* !multimapped */ @@ -356,6 +359,7 @@ struct kvm_arch{ */ struct list_head active_mmu_pages; struct list_head assigned_dev_head; + struct list_head oos_global_pages; struct dmar_domain *intel_iommu_domain; struct kvm_pic *vpic; struct kvm_ioapic *vioapic; @@ -385,6 +389,7 @@ struct kvm_vm_stat { u32 mmu_recycled; u32 mmu_cache_miss; u32 mmu_unsync; + u32 mmu_unsync_global; u32 remote_tlb_flush; u32 lpages; }; @@ -603,6 +608,7 @@ void __kvm_mmu_free_some_pages(struct kvm_vcpu *vcpu); int kvm_mmu_load(struct kvm_vcpu *vcpu); void kvm_mmu_unload(struct kvm_vcpu *vcpu); void kvm_mmu_sync_roots(struct kvm_vcpu *vcpu); +void kvm_mmu_sync_global(struct kvm_vcpu *vcpu); int kvm_emulate_hypercall(struct kvm_vcpu *vcpu); diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c index 58c35dead321..cbac9e4b156f 100644 --- a/arch/x86/kvm/mmu.c +++ b/arch/x86/kvm/mmu.c @@ -793,9 +793,11 @@ static struct kvm_mmu_page *kvm_mmu_alloc_page(struct kvm_vcpu *vcpu, sp->gfns = mmu_memory_cache_alloc(&vcpu->arch.mmu_page_cache, PAGE_SIZE); set_page_private(virt_to_page(sp->spt), (unsigned long)sp); list_add(&sp->link, &vcpu->kvm->arch.active_mmu_pages); + INIT_LIST_HEAD(&sp->oos_link); ASSERT(is_empty_shadow_page(sp->spt)); bitmap_zero(sp->slot_bitmap, KVM_MEMORY_SLOTS + KVM_PRIVATE_MEM_SLOTS); sp->multimapped = 0; + sp->global = 1; sp->parent_pte = parent_pte; --vcpu->kvm->arch.n_free_mmu_pages; return sp; @@ -1066,10 +1068,18 @@ static struct kvm_mmu_page *kvm_mmu_lookup_page(struct kvm *kvm, gfn_t gfn) return NULL; } +static void kvm_unlink_unsync_global(struct kvm *kvm, struct kvm_mmu_page *sp) +{ + list_del(&sp->oos_link); + --kvm->stat.mmu_unsync_global; +} + static void kvm_unlink_unsync_page(struct kvm *kvm, struct kvm_mmu_page *sp) { WARN_ON(!sp->unsync); sp->unsync = 0; + if (sp->global) + kvm_unlink_unsync_global(kvm, sp); --kvm->stat.mmu_unsync; } @@ -1615,9 +1625,15 @@ static int kvm_unsync_page(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp) if (s->role.word != sp->role.word) return 1; } - kvm_mmu_mark_parents_unsync(vcpu, sp); ++vcpu->kvm->stat.mmu_unsync; sp->unsync = 1; + + if (sp->global) { + list_add(&sp->oos_link, &vcpu->kvm->arch.oos_global_pages); + ++vcpu->kvm->stat.mmu_unsync_global; + } else + kvm_mmu_mark_parents_unsync(vcpu, sp); + mmu_convert_notrap(sp); return 0; } @@ -1643,12 +1659,21 @@ static int mmu_need_write_protect(struct kvm_vcpu *vcpu, gfn_t gfn, static int set_spte(struct kvm_vcpu *vcpu, u64 *shadow_pte, unsigned pte_access, int user_fault, int write_fault, int dirty, int largepage, - gfn_t gfn, pfn_t pfn, bool speculative, + int global, gfn_t gfn, pfn_t pfn, bool speculative, bool can_unsync) { u64 spte; int ret = 0; u64 mt_mask = shadow_mt_mask; + struct kvm_mmu_page *sp = page_header(__pa(shadow_pte)); + + if (!global && sp->global) { + sp->global = 0; + if (sp->unsync) { + kvm_unlink_unsync_global(vcpu->kvm, sp); + kvm_mmu_mark_parents_unsync(vcpu, sp); + } + } /* * We don't set the accessed bit, since we sometimes want to see @@ -1717,8 +1742,8 @@ set_pte: static void mmu_set_spte(struct kvm_vcpu *vcpu, u64 *shadow_pte, unsigned pt_access, unsigned pte_access, int user_fault, int write_fault, int dirty, - int *ptwrite, int largepage, gfn_t gfn, - pfn_t pfn, bool speculative) + int *ptwrite, int largepage, int global, + gfn_t gfn, pfn_t pfn, bool speculative) { int was_rmapped = 0; int was_writeble = is_writeble_pte(*shadow_pte); @@ -1751,7 +1776,7 @@ static void mmu_set_spte(struct kvm_vcpu *vcpu, u64 *shadow_pte, } } if (set_spte(vcpu, shadow_pte, pte_access, user_fault, write_fault, - dirty, largepage, gfn, pfn, speculative, true)) { + dirty, largepage, global, gfn, pfn, speculative, true)) { if (write_fault) *ptwrite = 1; kvm_x86_ops->tlb_flush(vcpu); @@ -1808,7 +1833,7 @@ static int direct_map_entry(struct kvm_shadow_walk *_walk, || (walk->largepage && level == PT_DIRECTORY_LEVEL)) { mmu_set_spte(vcpu, sptep, ACC_ALL, ACC_ALL, 0, walk->write, 1, &walk->pt_write, - walk->largepage, gfn, walk->pfn, false); + walk->largepage, 0, gfn, walk->pfn, false); ++vcpu->stat.pf_fixed; return 1; } @@ -1995,6 +2020,15 @@ static void mmu_sync_roots(struct kvm_vcpu *vcpu) } } +static void mmu_sync_global(struct kvm_vcpu *vcpu) +{ + struct kvm *kvm = vcpu->kvm; + struct kvm_mmu_page *sp, *n; + + list_for_each_entry_safe(sp, n, &kvm->arch.oos_global_pages, oos_link) + kvm_sync_page(vcpu, sp); +} + void kvm_mmu_sync_roots(struct kvm_vcpu *vcpu) { spin_lock(&vcpu->kvm->mmu_lock); @@ -2002,6 +2036,13 @@ void kvm_mmu_sync_roots(struct kvm_vcpu *vcpu) spin_unlock(&vcpu->kvm->mmu_lock); } +void kvm_mmu_sync_global(struct kvm_vcpu *vcpu) +{ + spin_lock(&vcpu->kvm->mmu_lock); + mmu_sync_global(vcpu); + spin_unlock(&vcpu->kvm->mmu_lock); +} + static gpa_t nonpaging_gva_to_gpa(struct kvm_vcpu *vcpu, gva_t vaddr) { return vaddr; diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h index 84eee43bbe74..e644d81979b6 100644 --- a/arch/x86/kvm/paging_tmpl.h +++ b/arch/x86/kvm/paging_tmpl.h @@ -274,7 +274,8 @@ static void FNAME(update_pte)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *page, return; kvm_get_pfn(pfn); mmu_set_spte(vcpu, spte, page->role.access, pte_access, 0, 0, - gpte & PT_DIRTY_MASK, NULL, largepage, gpte_to_gfn(gpte), + gpte & PT_DIRTY_MASK, NULL, largepage, + gpte & PT_GLOBAL_MASK, gpte_to_gfn(gpte), pfn, true); } @@ -301,8 +302,9 @@ static int FNAME(shadow_walk_entry)(struct kvm_shadow_walk *_sw, mmu_set_spte(vcpu, sptep, access, gw->pte_access & access, sw->user_fault, sw->write_fault, gw->ptes[gw->level-1] & PT_DIRTY_MASK, - sw->ptwrite, sw->largepage, gw->gfn, sw->pfn, - false); + sw->ptwrite, sw->largepage, + gw->ptes[gw->level-1] & PT_GLOBAL_MASK, + gw->gfn, sw->pfn, false); sw->sptep = sptep; return 1; } @@ -580,7 +582,7 @@ static int FNAME(sync_page)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp) nr_present++; pte_access = sp->role.access & FNAME(gpte_access)(vcpu, gpte); set_spte(vcpu, &sp->spt[i], pte_access, 0, 0, - is_dirty_pte(gpte), 0, gfn, + is_dirty_pte(gpte), 0, gpte & PT_GLOBAL_MASK, gfn, spte_to_pfn(sp->spt[i]), true, false); } diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 7a2aeba0bfbd..774db00d2db6 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -104,6 +104,7 @@ struct kvm_stats_debugfs_item debugfs_entries[] = { { "mmu_recycled", VM_STAT(mmu_recycled) }, { "mmu_cache_miss", VM_STAT(mmu_cache_miss) }, { "mmu_unsync", VM_STAT(mmu_unsync) }, + { "mmu_unsync_global", VM_STAT(mmu_unsync_global) }, { "remote_tlb_flush", VM_STAT(remote_tlb_flush) }, { "largepages", VM_STAT(lpages) }, { NULL } @@ -315,6 +316,7 @@ void kvm_set_cr0(struct kvm_vcpu *vcpu, unsigned long cr0) kvm_x86_ops->set_cr0(vcpu, cr0); vcpu->arch.cr0 = cr0; + kvm_mmu_sync_global(vcpu); kvm_mmu_reset_context(vcpu); return; } @@ -358,6 +360,7 @@ void kvm_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4) } kvm_x86_ops->set_cr4(vcpu, cr4); vcpu->arch.cr4 = cr4; + kvm_mmu_sync_global(vcpu); kvm_mmu_reset_context(vcpu); } EXPORT_SYMBOL_GPL(kvm_set_cr4); @@ -4113,6 +4116,7 @@ struct kvm *kvm_arch_create_vm(void) return ERR_PTR(-ENOMEM); INIT_LIST_HEAD(&kvm->arch.active_mmu_pages); + INIT_LIST_HEAD(&kvm->arch.oos_global_pages); INIT_LIST_HEAD(&kvm->arch.assigned_dev_head); /* Reserve bit 0 of irq_sources_bitmap for userspace irq source */ -- cgit v1.2.3 From ad218f85e388e8ca816ff09d91c246cd014c53a8 Mon Sep 17 00:00:00 2001 From: Marcelo Tosatti Date: Mon, 1 Dec 2008 22:32:05 -0200 Subject: KVM: MMU: prepopulate the shadow on invlpg If the guest executes invlpg, peek into the pagetable and attempt to prepopulate the shadow entry. Also stop dirty fault updates from interfering with the fork detector. 2% improvement on RHEL3/AIM7. Signed-off-by: Marcelo Tosatti Signed-off-by: Avi Kivity --- arch/x86/include/asm/kvm_host.h | 3 ++- arch/x86/kvm/mmu.c | 25 +++++++++++++------------ arch/x86/kvm/paging_tmpl.h | 25 ++++++++++++++++++++++++- arch/x86/kvm/x86.c | 2 +- 4 files changed, 40 insertions(+), 15 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index 65b1ed295698..97215a458e5f 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -602,7 +602,8 @@ unsigned long segment_base(u16 selector); void kvm_mmu_flush_tlb(struct kvm_vcpu *vcpu); void kvm_mmu_pte_write(struct kvm_vcpu *vcpu, gpa_t gpa, - const u8 *new, int bytes); + const u8 *new, int bytes, + bool guest_initiated); int kvm_mmu_unprotect_page_virt(struct kvm_vcpu *vcpu, gva_t gva); void __kvm_mmu_free_some_pages(struct kvm_vcpu *vcpu); int kvm_mmu_load(struct kvm_vcpu *vcpu); diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c index cbac9e4b156f..863baf70506e 100644 --- a/arch/x86/kvm/mmu.c +++ b/arch/x86/kvm/mmu.c @@ -2441,7 +2441,8 @@ static void kvm_mmu_access_page(struct kvm_vcpu *vcpu, gfn_t gfn) } void kvm_mmu_pte_write(struct kvm_vcpu *vcpu, gpa_t gpa, - const u8 *new, int bytes) + const u8 *new, int bytes, + bool guest_initiated) { gfn_t gfn = gpa >> PAGE_SHIFT; struct kvm_mmu_page *sp; @@ -2467,15 +2468,17 @@ void kvm_mmu_pte_write(struct kvm_vcpu *vcpu, gpa_t gpa, kvm_mmu_free_some_pages(vcpu); ++vcpu->kvm->stat.mmu_pte_write; kvm_mmu_audit(vcpu, "pre pte write"); - if (gfn == vcpu->arch.last_pt_write_gfn - && !last_updated_pte_accessed(vcpu)) { - ++vcpu->arch.last_pt_write_count; - if (vcpu->arch.last_pt_write_count >= 3) - flooded = 1; - } else { - vcpu->arch.last_pt_write_gfn = gfn; - vcpu->arch.last_pt_write_count = 1; - vcpu->arch.last_pte_updated = NULL; + if (guest_initiated) { + if (gfn == vcpu->arch.last_pt_write_gfn + && !last_updated_pte_accessed(vcpu)) { + ++vcpu->arch.last_pt_write_count; + if (vcpu->arch.last_pt_write_count >= 3) + flooded = 1; + } else { + vcpu->arch.last_pt_write_gfn = gfn; + vcpu->arch.last_pt_write_count = 1; + vcpu->arch.last_pte_updated = NULL; + } } index = kvm_page_table_hashfn(gfn); bucket = &vcpu->kvm->arch.mmu_page_hash[index]; @@ -2615,9 +2618,7 @@ EXPORT_SYMBOL_GPL(kvm_mmu_page_fault); void kvm_mmu_invlpg(struct kvm_vcpu *vcpu, gva_t gva) { - spin_lock(&vcpu->kvm->mmu_lock); vcpu->arch.mmu.invlpg(vcpu, gva); - spin_unlock(&vcpu->kvm->mmu_lock); kvm_mmu_flush_tlb(vcpu); ++vcpu->stat.invlpg; } diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h index e644d81979b6..d20640154216 100644 --- a/arch/x86/kvm/paging_tmpl.h +++ b/arch/x86/kvm/paging_tmpl.h @@ -82,6 +82,7 @@ struct shadow_walker { int *ptwrite; pfn_t pfn; u64 *sptep; + gpa_t pte_gpa; }; static gfn_t gpte_to_gfn(pt_element_t gpte) @@ -222,7 +223,7 @@ walk: if (ret) goto walk; pte |= PT_DIRTY_MASK; - kvm_mmu_pte_write(vcpu, pte_gpa, (u8 *)&pte, sizeof(pte)); + kvm_mmu_pte_write(vcpu, pte_gpa, (u8 *)&pte, sizeof(pte), 0); walker->ptes[walker->level - 1] = pte; } @@ -468,8 +469,15 @@ static int FNAME(shadow_invlpg_entry)(struct kvm_shadow_walk *_sw, struct kvm_vcpu *vcpu, u64 addr, u64 *sptep, int level) { + struct shadow_walker *sw = + container_of(_sw, struct shadow_walker, walker); if (level == PT_PAGE_TABLE_LEVEL) { + struct kvm_mmu_page *sp = page_header(__pa(sptep)); + + sw->pte_gpa = (sp->gfn << PAGE_SHIFT); + sw->pte_gpa += (sptep - sp->spt) * sizeof(pt_element_t); + if (is_shadow_present_pte(*sptep)) rmap_remove(vcpu->kvm, sptep); set_shadow_pte(sptep, shadow_trap_nonpresent_pte); @@ -482,11 +490,26 @@ static int FNAME(shadow_invlpg_entry)(struct kvm_shadow_walk *_sw, static void FNAME(invlpg)(struct kvm_vcpu *vcpu, gva_t gva) { + pt_element_t gpte; struct shadow_walker walker = { .walker = { .entry = FNAME(shadow_invlpg_entry), }, + .pte_gpa = -1, }; + spin_lock(&vcpu->kvm->mmu_lock); walk_shadow(&walker.walker, vcpu, gva); + spin_unlock(&vcpu->kvm->mmu_lock); + if (walker.pte_gpa == -1) + return; + if (kvm_read_guest_atomic(vcpu->kvm, walker.pte_gpa, &gpte, + sizeof(pt_element_t))) + return; + if (is_present_pte(gpte) && (gpte & PT_ACCESSED_MASK)) { + if (mmu_topup_memory_caches(vcpu)) + return; + kvm_mmu_pte_write(vcpu, walker.pte_gpa, (const u8 *)&gpte, + sizeof(pt_element_t), 0); + } } static gpa_t FNAME(gva_to_gpa)(struct kvm_vcpu *vcpu, gva_t vaddr) diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 774db00d2db6..ba102879de33 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -2046,7 +2046,7 @@ int emulator_write_phys(struct kvm_vcpu *vcpu, gpa_t gpa, ret = kvm_write_guest(vcpu->kvm, gpa, val, bytes); if (ret < 0) return 0; - kvm_mmu_pte_write(vcpu, gpa, val, bytes); + kvm_mmu_pte_write(vcpu, gpa, val, bytes, 1); return 1; } -- cgit v1.2.3 From e93353c93a3ba4215633ce930784f40a4e94e3f9 Mon Sep 17 00:00:00 2001 From: Eduardo Habkost Date: Fri, 5 Dec 2008 18:36:45 -0200 Subject: x86: KVM guest: kvm_get_tsc_khz: return khz, not lpj kvm_get_tsc_khz() currently returns the previously-calculated preset_lpj value, but it is in loops-per-jiffy, not kHz. The current code works correctly only when HZ=1000. Signed-off-by: Eduardo Habkost Signed-off-by: Avi Kivity --- arch/x86/kernel/kvmclock.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/kvmclock.c b/arch/x86/kernel/kvmclock.c index b38e801014e3..652fce6d2cce 100644 --- a/arch/x86/kernel/kvmclock.c +++ b/arch/x86/kernel/kvmclock.c @@ -89,17 +89,17 @@ static cycle_t kvm_clock_read(void) */ static unsigned long kvm_get_tsc_khz(void) { - return preset_lpj; + struct pvclock_vcpu_time_info *src; + src = &per_cpu(hv_clock, 0); + return pvclock_tsc_khz(src); } static void kvm_get_preset_lpj(void) { - struct pvclock_vcpu_time_info *src; unsigned long khz; u64 lpj; - src = &per_cpu(hv_clock, 0); - khz = pvclock_tsc_khz(src); + khz = kvm_get_tsc_khz(); lpj = ((u64)khz * 1000); do_div(lpj, HZ); -- cgit v1.2.3 From ca9edaee1aea34ebd9adb48910aba0b3d64b1b22 Mon Sep 17 00:00:00 2001 From: Avi Kivity Date: Mon, 8 Dec 2008 18:29:29 +0200 Subject: KVM: Consolidate userspace memory capability reporting into common code Signed-off-by: Avi Kivity --- arch/ia64/kvm/kvm-ia64.c | 1 - arch/powerpc/kvm/powerpc.c | 3 --- arch/s390/kvm/kvm-s390.c | 2 -- arch/x86/kvm/x86.c | 1 - virt/kvm/kvm_main.c | 1 + 5 files changed, 1 insertion(+), 7 deletions(-) (limited to 'arch/x86') diff --git a/arch/ia64/kvm/kvm-ia64.c b/arch/ia64/kvm/kvm-ia64.c index b4d24e2cce40..d2eb9691d61d 100644 --- a/arch/ia64/kvm/kvm-ia64.c +++ b/arch/ia64/kvm/kvm-ia64.c @@ -180,7 +180,6 @@ int kvm_dev_ioctl_check_extension(long ext) switch (ext) { case KVM_CAP_IRQCHIP: - case KVM_CAP_USER_MEMORY: case KVM_CAP_MP_STATE: r = 1; diff --git a/arch/powerpc/kvm/powerpc.c b/arch/powerpc/kvm/powerpc.c index 1deda37cb771..2822c8ccfaaf 100644 --- a/arch/powerpc/kvm/powerpc.c +++ b/arch/powerpc/kvm/powerpc.c @@ -137,9 +137,6 @@ int kvm_dev_ioctl_check_extension(long ext) int r; switch (ext) { - case KVM_CAP_USER_MEMORY: - r = 1; - break; case KVM_CAP_COALESCED_MMIO: r = KVM_COALESCED_MMIO_PAGE_OFFSET; break; diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c index 76f05ddaef10..be8497186b96 100644 --- a/arch/s390/kvm/kvm-s390.c +++ b/arch/s390/kvm/kvm-s390.c @@ -113,8 +113,6 @@ long kvm_arch_dev_ioctl(struct file *filp, int kvm_dev_ioctl_check_extension(long ext) { switch (ext) { - case KVM_CAP_USER_MEMORY: - return 1; default: return 0; } diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index ba102879de33..10302d3bd415 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -964,7 +964,6 @@ int kvm_dev_ioctl_check_extension(long ext) case KVM_CAP_IRQCHIP: case KVM_CAP_HLT: case KVM_CAP_MMU_SHADOW_CACHE_CONTROL: - case KVM_CAP_USER_MEMORY: case KVM_CAP_SET_TSS_ADDR: case KVM_CAP_EXT_CPUID: case KVM_CAP_CLOCKSOURCE: diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c index e066eb125e55..eb70ca6c7145 100644 --- a/virt/kvm/kvm_main.c +++ b/virt/kvm/kvm_main.c @@ -1908,6 +1908,7 @@ static int kvm_dev_ioctl_create_vm(void) static long kvm_dev_ioctl_check_extension_generic(long arg) { switch (arg) { + case KVM_CAP_USER_MEMORY: case KVM_CAP_DESTROY_MEMORY_REGION_WORKS: return 1; default: -- cgit v1.2.3 From eb64f1e8cd5c3cae912db30a77d062367f7a11a6 Mon Sep 17 00:00:00 2001 From: Marcelo Tosatti Date: Tue, 9 Dec 2008 16:07:22 +0100 Subject: KVM: MMU: check for present pdptr shadow page in walk_shadow walk_shadow assumes the caller verified validity of the pdptr pointer in question, which is not the case for the invlpg handler. Fixes oops during Solaris 10 install. Signed-off-by: Marcelo Tosatti Signed-off-by: Avi Kivity --- arch/x86/kvm/mmu.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c index 863baf70506e..641c07844e6e 100644 --- a/arch/x86/kvm/mmu.c +++ b/arch/x86/kvm/mmu.c @@ -1269,6 +1269,8 @@ static int walk_shadow(struct kvm_shadow_walk *walker, if (level == PT32E_ROOT_LEVEL) { shadow_addr = vcpu->arch.mmu.pae_root[(addr >> 30) & 3]; shadow_addr &= PT64_BASE_ADDR_MASK; + if (!shadow_addr) + return 1; --level; } -- cgit v1.2.3 From 264ff01d55b456932cef03082448b41d2edeb6a1 Mon Sep 17 00:00:00 2001 From: Jan Kiszka Date: Mon, 24 Nov 2008 12:26:19 +0100 Subject: KVM: VMX: Fix pending NMI-vs.-IRQ race for user space irqchip As with the kernel irqchip, don't allow an NMI to stomp over an already injected IRQ; instead wait for the IRQ injection to be completed. Signed-off-by: Jan Kiszka Signed-off-by: Avi Kivity --- arch/x86/kvm/vmx.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index e446f232588e..487e1dcdce33 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -2486,7 +2486,9 @@ static void do_interrupt_requests(struct kvm_vcpu *vcpu, vmx_update_window_states(vcpu); if (vcpu->arch.nmi_pending && !vcpu->arch.nmi_injected) { - if (vcpu->arch.nmi_window_open) { + if (vcpu->arch.interrupt.pending) { + enable_nmi_window(vcpu); + } else if (vcpu->arch.nmi_window_open) { vcpu->arch.nmi_pending = false; vcpu->arch.nmi_injected = true; } else { -- cgit v1.2.3 From 4531220b71f0399e71cda0c4cf749e7281a7416a Mon Sep 17 00:00:00 2001 From: Jan Kiszka Date: Thu, 11 Dec 2008 16:54:54 +0100 Subject: KVM: x86: Rework user space NMI injection as KVM_CAP_USER_NMI There is no point in doing the ready_for_nmi_injection/ request_nmi_window dance with user space. First, we don't do this for in-kernel irqchip anyway, while the code path is the same as for user space irqchip mode. And second, there is nothing to loose if a pending NMI is overwritten by another one (in contrast to IRQs where we have to save the number). Actually, there is even the risk of raising spurious NMIs this way because the reason for the held-back NMI might already be handled while processing the first one. Therefore this patch creates a simplified user space NMI injection interface, exporting it under KVM_CAP_USER_NMI and dropping the old KVM_CAP_NMI capability. And this time we also take care to provide the interface only on archs supporting NMIs via KVM (right now only x86). Signed-off-by: Jan Kiszka Signed-off-by: Avi Kivity --- arch/x86/kvm/vmx.c | 24 ++---------------------- arch/x86/kvm/x86.c | 28 ++-------------------------- include/linux/kvm.h | 11 +++++------ 3 files changed, 9 insertions(+), 54 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index 487e1dcdce33..6259d7467648 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -2498,15 +2498,13 @@ static void do_interrupt_requests(struct kvm_vcpu *vcpu, } if (vcpu->arch.nmi_injected) { vmx_inject_nmi(vcpu); - if (vcpu->arch.nmi_pending || kvm_run->request_nmi_window) + if (vcpu->arch.nmi_pending) enable_nmi_window(vcpu); else if (vcpu->arch.irq_summary || kvm_run->request_interrupt_window) enable_irq_window(vcpu); return; } - if (!vcpu->arch.nmi_window_open || kvm_run->request_nmi_window) - enable_nmi_window(vcpu); if (vcpu->arch.interrupt_window_open) { if (vcpu->arch.irq_summary && !vcpu->arch.interrupt.pending) @@ -3040,14 +3038,6 @@ static int handle_nmi_window(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run) vmcs_write32(CPU_BASED_VM_EXEC_CONTROL, cpu_based_vm_exec_control); ++vcpu->stat.nmi_window_exits; - /* - * If the user space waits to inject a NMI, exit as soon as possible - */ - if (kvm_run->request_nmi_window && !vcpu->arch.nmi_pending) { - kvm_run->exit_reason = KVM_EXIT_NMI_WINDOW_OPEN; - return 0; - } - return 1; } @@ -3162,7 +3152,7 @@ static int kvm_handle_exit(struct kvm_run *kvm_run, struct kvm_vcpu *vcpu) vmx->soft_vnmi_blocked = 0; vcpu->arch.nmi_window_open = 1; } else if (vmx->vnmi_blocked_time > 1000000000LL && - (kvm_run->request_nmi_window || vcpu->arch.nmi_pending)) { + vcpu->arch.nmi_pending) { /* * This CPU don't support us in finding the end of an * NMI-blocked window if the guest runs with IRQs @@ -3175,16 +3165,6 @@ static int kvm_handle_exit(struct kvm_run *kvm_run, struct kvm_vcpu *vcpu) vmx->soft_vnmi_blocked = 0; vmx->vcpu.arch.nmi_window_open = 1; } - - /* - * If the user space waits to inject an NNI, exit ASAP - */ - if (vcpu->arch.nmi_window_open && kvm_run->request_nmi_window - && !vcpu->arch.nmi_pending) { - kvm_run->exit_reason = KVM_EXIT_NMI_WINDOW_OPEN; - ++vcpu->stat.nmi_window_exits; - return 0; - } } if (exit_reason < kvm_vmx_max_exit_handlers diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 10302d3bd415..0e6aa8141dcd 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -2887,37 +2887,18 @@ static int dm_request_for_irq_injection(struct kvm_vcpu *vcpu, (kvm_x86_ops->get_rflags(vcpu) & X86_EFLAGS_IF)); } -/* - * Check if userspace requested a NMI window, and that the NMI window - * is open. - * - * No need to exit to userspace if we already have a NMI queued. - */ -static int dm_request_for_nmi_injection(struct kvm_vcpu *vcpu, - struct kvm_run *kvm_run) -{ - return (!vcpu->arch.nmi_pending && - kvm_run->request_nmi_window && - vcpu->arch.nmi_window_open); -} - static void post_kvm_run_save(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run) { kvm_run->if_flag = (kvm_x86_ops->get_rflags(vcpu) & X86_EFLAGS_IF) != 0; kvm_run->cr8 = kvm_get_cr8(vcpu); kvm_run->apic_base = kvm_get_apic_base(vcpu); - if (irqchip_in_kernel(vcpu->kvm)) { + if (irqchip_in_kernel(vcpu->kvm)) kvm_run->ready_for_interrupt_injection = 1; - kvm_run->ready_for_nmi_injection = 1; - } else { + else kvm_run->ready_for_interrupt_injection = (vcpu->arch.interrupt_window_open && vcpu->arch.irq_summary == 0); - kvm_run->ready_for_nmi_injection = - (vcpu->arch.nmi_window_open && - vcpu->arch.nmi_pending == 0); - } } static void vapic_enter(struct kvm_vcpu *vcpu) @@ -3093,11 +3074,6 @@ static int __vcpu_run(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run) } if (r > 0) { - if (dm_request_for_nmi_injection(vcpu, kvm_run)) { - r = -EINTR; - kvm_run->exit_reason = KVM_EXIT_NMI; - ++vcpu->stat.request_nmi_exits; - } if (dm_request_for_irq_injection(vcpu, kvm_run)) { r = -EINTR; kvm_run->exit_reason = KVM_EXIT_INTR; diff --git a/include/linux/kvm.h b/include/linux/kvm.h index 48807767e726..35525ac63337 100644 --- a/include/linux/kvm.h +++ b/include/linux/kvm.h @@ -84,21 +84,18 @@ struct kvm_irqchip { #define KVM_EXIT_S390_RESET 14 #define KVM_EXIT_DCR 15 #define KVM_EXIT_NMI 16 -#define KVM_EXIT_NMI_WINDOW_OPEN 17 /* for KVM_RUN, returned by mmap(vcpu_fd, offset=0) */ struct kvm_run { /* in */ __u8 request_interrupt_window; - __u8 request_nmi_window; - __u8 padding1[6]; + __u8 padding1[7]; /* out */ __u32 exit_reason; __u8 ready_for_interrupt_injection; __u8 if_flag; - __u8 ready_for_nmi_injection; - __u8 padding2; + __u8 padding2[2]; /* in (pre_kvm_run), out (post_kvm_run) */ __u64 cr8; @@ -391,12 +388,14 @@ struct kvm_trace_rec { #define KVM_CAP_DEVICE_ASSIGNMENT 17 #endif #define KVM_CAP_IOMMU 18 -#define KVM_CAP_NMI 19 #if defined(CONFIG_X86) #define KVM_CAP_DEVICE_MSI 20 #endif /* Bug in KVM_SET_USER_MEMORY_REGION fixed: */ #define KVM_CAP_DESTROY_MEMORY_REGION_WORKS 21 +#if defined(CONFIG_X86) +#define KVM_CAP_USER_NMI 22 +#endif /* * ioctls for VM fds -- cgit v1.2.3 From 25e2343246fe135fce672f41abe61e9d2c38caac Mon Sep 17 00:00:00 2001 From: Avi Kivity Date: Sun, 21 Dec 2008 18:31:10 +0200 Subject: KVM: MMU: Don't treat a global pte as such if cr4.pge is cleared The pte.g bit is meaningless if global pages are disabled; deferring mmu page synchronization on these ptes will lead to the guest using stale shadow ptes. Fixes Vista x86 smp bootloader failure. Signed-off-by: Avi Kivity --- arch/x86/kvm/mmu.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c index 641c07844e6e..d50ebac6a07f 100644 --- a/arch/x86/kvm/mmu.c +++ b/arch/x86/kvm/mmu.c @@ -1669,6 +1669,8 @@ static int set_spte(struct kvm_vcpu *vcpu, u64 *shadow_pte, u64 mt_mask = shadow_mt_mask; struct kvm_mmu_page *sp = page_header(__pa(shadow_pte)); + if (!(vcpu->arch.cr4 & X86_CR4_PGE)) + global = 0; if (!global && sp->global) { sp->global = 0; if (sp->unsync) { -- cgit v1.2.3 From 3f353858c98dbe0240dac558a89870f4600f81bb Mon Sep 17 00:00:00 2001 From: Avi Kivity Date: Sun, 21 Dec 2008 22:48:32 +0200 Subject: KVM: Add locking to virtual i8259 interrupt controller While most accesses to the i8259 are with the kvm mutex taken, the call to kvm_pic_read_irq() is not. We can't easily take the kvm mutex there since the function is called with interrupts disabled. Fix by adding a spinlock to the virtual interrupt controller. Since we can't send an IPI under the spinlock (we also take the same spinlock in an irq disabled context), we defer the IPI until the spinlock is released. Similarly, we defer irq ack notifications until after spinlock release to avoid lock recursion. Signed-off-by: Avi Kivity --- arch/x86/kvm/i8259.c | 52 ++++++++++++++++++++++++++++++++++++++++++++++++---- arch/x86/kvm/irq.h | 5 +++++ 2 files changed, 53 insertions(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/i8259.c b/arch/x86/kvm/i8259.c index 17e41e165f1a..179dcb0103fd 100644 --- a/arch/x86/kvm/i8259.c +++ b/arch/x86/kvm/i8259.c @@ -26,10 +26,40 @@ * Port from Qemu. */ #include +#include #include "irq.h" #include +static void pic_lock(struct kvm_pic *s) +{ + spin_lock(&s->lock); +} + +static void pic_unlock(struct kvm_pic *s) +{ + struct kvm *kvm = s->kvm; + unsigned acks = s->pending_acks; + bool wakeup = s->wakeup_needed; + struct kvm_vcpu *vcpu; + + s->pending_acks = 0; + s->wakeup_needed = false; + + spin_unlock(&s->lock); + + while (acks) { + kvm_notify_acked_irq(kvm, __ffs(acks)); + acks &= acks - 1; + } + + if (wakeup) { + vcpu = s->kvm->vcpus[0]; + if (vcpu) + kvm_vcpu_kick(vcpu); + } +} + static void pic_clear_isr(struct kvm_kpic_state *s, int irq) { s->isr &= ~(1 << irq); @@ -136,17 +166,21 @@ static void pic_update_irq(struct kvm_pic *s) void kvm_pic_update_irq(struct kvm_pic *s) { + pic_lock(s); pic_update_irq(s); + pic_unlock(s); } void kvm_pic_set_irq(void *opaque, int irq, int level) { struct kvm_pic *s = opaque; + pic_lock(s); if (irq >= 0 && irq < PIC_NUM_PINS) { pic_set_irq1(&s->pics[irq >> 3], irq & 7, level); pic_update_irq(s); } + pic_unlock(s); } /* @@ -172,6 +206,7 @@ int kvm_pic_read_irq(struct kvm *kvm) int irq, irq2, intno; struct kvm_pic *s = pic_irqchip(kvm); + pic_lock(s); irq = pic_get_irq(&s->pics[0]); if (irq >= 0) { pic_intack(&s->pics[0], irq); @@ -196,6 +231,7 @@ int kvm_pic_read_irq(struct kvm *kvm) intno = s->pics[0].irq_base + irq; } pic_update_irq(s); + pic_unlock(s); kvm_notify_acked_irq(kvm, irq); return intno; @@ -203,7 +239,7 @@ int kvm_pic_read_irq(struct kvm *kvm) void kvm_pic_reset(struct kvm_kpic_state *s) { - int irq, irqbase; + int irq, irqbase, n; struct kvm *kvm = s->pics_state->irq_request_opaque; struct kvm_vcpu *vcpu0 = kvm->vcpus[0]; @@ -214,8 +250,10 @@ void kvm_pic_reset(struct kvm_kpic_state *s) for (irq = 0; irq < PIC_NUM_PINS/2; irq++) { if (vcpu0 && kvm_apic_accept_pic_intr(vcpu0)) - if (s->irr & (1 << irq) || s->isr & (1 << irq)) - kvm_notify_acked_irq(kvm, irq+irqbase); + if (s->irr & (1 << irq) || s->isr & (1 << irq)) { + n = irq + irqbase; + s->pics_state->pending_acks |= 1 << n; + } } s->last_irr = 0; s->irr = 0; @@ -406,6 +444,7 @@ static void picdev_write(struct kvm_io_device *this, printk(KERN_ERR "PIC: non byte write\n"); return; } + pic_lock(s); switch (addr) { case 0x20: case 0x21: @@ -418,6 +457,7 @@ static void picdev_write(struct kvm_io_device *this, elcr_ioport_write(&s->pics[addr & 1], addr, data); break; } + pic_unlock(s); } static void picdev_read(struct kvm_io_device *this, @@ -431,6 +471,7 @@ static void picdev_read(struct kvm_io_device *this, printk(KERN_ERR "PIC: non byte read\n"); return; } + pic_lock(s); switch (addr) { case 0x20: case 0x21: @@ -444,6 +485,7 @@ static void picdev_read(struct kvm_io_device *this, break; } *(unsigned char *)val = data; + pic_unlock(s); } /* @@ -459,7 +501,7 @@ static void pic_irq_request(void *opaque, int level) s->output = level; if (vcpu && level && (s->pics[0].isr_ack & (1 << irq))) { s->pics[0].isr_ack &= ~(1 << irq); - kvm_vcpu_kick(vcpu); + s->wakeup_needed = true; } } @@ -469,6 +511,8 @@ struct kvm_pic *kvm_create_pic(struct kvm *kvm) s = kzalloc(sizeof(struct kvm_pic), GFP_KERNEL); if (!s) return NULL; + spin_lock_init(&s->lock); + s->kvm = kvm; s->pics[0].elcr_mask = 0xf8; s->pics[1].elcr_mask = 0xde; s->irq_request = pic_irq_request; diff --git a/arch/x86/kvm/irq.h b/arch/x86/kvm/irq.h index b9e9051650ea..2bf32a03ceec 100644 --- a/arch/x86/kvm/irq.h +++ b/arch/x86/kvm/irq.h @@ -25,6 +25,7 @@ #include #include #include +#include #include "iodev.h" #include "ioapic.h" @@ -59,6 +60,10 @@ struct kvm_kpic_state { }; struct kvm_pic { + spinlock_t lock; + bool wakeup_needed; + unsigned pending_acks; + struct kvm *kvm; struct kvm_kpic_state pics[2]; /* 0 is master pic, 1 is slave pic */ irq_request_func *irq_request; void *irq_request_opaque; -- cgit v1.2.3 From 87917239204d67a316cb89751750f86c9ed3640b Mon Sep 17 00:00:00 2001 From: Marcelo Tosatti Date: Mon, 22 Dec 2008 18:49:30 -0200 Subject: KVM: MMU: handle large host sptes on invlpg/resync The invlpg and sync walkers lack knowledge of large host sptes, descending to non-existant pagetable level. Stop at directory level in such case. Fixes SMP Windows XP with hugepages. Signed-off-by: Marcelo Tosatti Signed-off-by: Avi Kivity --- arch/x86/kvm/mmu.c | 2 +- arch/x86/kvm/paging_tmpl.h | 9 +++++++-- 2 files changed, 8 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c index d50ebac6a07f..83f11c7474a1 100644 --- a/arch/x86/kvm/mmu.c +++ b/arch/x86/kvm/mmu.c @@ -1007,7 +1007,7 @@ static int __mmu_unsync_walk(struct kvm_mmu_page *sp, for_each_unsync_children(sp->unsync_child_bitmap, i) { u64 ent = sp->spt[i]; - if (is_shadow_present_pte(ent)) { + if (is_shadow_present_pte(ent) && !is_large_pte(ent)) { struct kvm_mmu_page *child; child = page_header(ent & PT64_BASE_ADDR_MASK); diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h index d20640154216..9fd78b6e17ad 100644 --- a/arch/x86/kvm/paging_tmpl.h +++ b/arch/x86/kvm/paging_tmpl.h @@ -472,14 +472,19 @@ static int FNAME(shadow_invlpg_entry)(struct kvm_shadow_walk *_sw, struct shadow_walker *sw = container_of(_sw, struct shadow_walker, walker); - if (level == PT_PAGE_TABLE_LEVEL) { + /* FIXME: properly handle invlpg on large guest pages */ + if (level == PT_PAGE_TABLE_LEVEL || + ((level == PT_DIRECTORY_LEVEL) && is_large_pte(*sptep))) { struct kvm_mmu_page *sp = page_header(__pa(sptep)); sw->pte_gpa = (sp->gfn << PAGE_SHIFT); sw->pte_gpa += (sptep - sp->spt) * sizeof(pt_element_t); - if (is_shadow_present_pte(*sptep)) + if (is_shadow_present_pte(*sptep)) { rmap_remove(vcpu->kvm, sptep); + if (is_large_pte(*sptep)) + --vcpu->kvm->stat.lpages; + } set_shadow_pte(sptep, shadow_trap_nonpresent_pte); return 1; } -- cgit v1.2.3 From 18d8fda7c3c9439be04d7ea2e82da2513b121acb Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 26 Dec 2008 00:35:37 -0500 Subject: take init_fs to saner place Signed-off-by: Al Viro --- arch/alpha/kernel/init_task.c | 1 - arch/arm/kernel/init_task.c | 1 - arch/avr32/kernel/init_task.c | 1 - arch/blackfin/kernel/init_task.c | 1 - arch/cris/kernel/process.c | 1 - arch/frv/kernel/init_task.c | 1 - arch/h8300/kernel/init_task.c | 1 - arch/ia64/kernel/init_task.c | 1 - arch/m32r/kernel/init_task.c | 1 - arch/m68k/kernel/process.c | 1 - arch/m68knommu/kernel/init_task.c | 1 - arch/mips/kernel/init_task.c | 1 - arch/mn10300/kernel/init_task.c | 1 - arch/parisc/kernel/init_task.c | 1 - arch/powerpc/kernel/init_task.c | 1 - arch/s390/kernel/init_task.c | 1 - arch/sh/kernel/init_task.c | 1 - arch/sparc/kernel/init_task.c | 1 - arch/um/kernel/init_task.c | 1 - arch/x86/kernel/init_task.c | 1 - arch/xtensa/kernel/init_task.c | 1 - fs/namei.c | 7 +++++++ include/linux/fs_struct.h | 6 ------ include/linux/init_task.h | 1 + 24 files changed, 8 insertions(+), 27 deletions(-) (limited to 'arch/x86') diff --git a/arch/alpha/kernel/init_task.c b/arch/alpha/kernel/init_task.c index 1f762189fa64..c2938e574a40 100644 --- a/arch/alpha/kernel/init_task.c +++ b/arch/alpha/kernel/init_task.c @@ -8,7 +8,6 @@ #include -static struct fs_struct init_fs = INIT_FS; static struct signal_struct init_signals = INIT_SIGNALS(init_signals); static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand); struct mm_struct init_mm = INIT_MM(init_mm); diff --git a/arch/arm/kernel/init_task.c b/arch/arm/kernel/init_task.c index 0bbf80625395..e859af349467 100644 --- a/arch/arm/kernel/init_task.c +++ b/arch/arm/kernel/init_task.c @@ -12,7 +12,6 @@ #include -static struct fs_struct init_fs = INIT_FS; static struct signal_struct init_signals = INIT_SIGNALS(init_signals); static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand); struct mm_struct init_mm = INIT_MM(init_mm); diff --git a/arch/avr32/kernel/init_task.c b/arch/avr32/kernel/init_task.c index 44058469c6ec..993d56ee3cf3 100644 --- a/arch/avr32/kernel/init_task.c +++ b/arch/avr32/kernel/init_task.c @@ -13,7 +13,6 @@ #include -static struct fs_struct init_fs = INIT_FS; static struct signal_struct init_signals = INIT_SIGNALS(init_signals); static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand); struct mm_struct init_mm = INIT_MM(init_mm); diff --git a/arch/blackfin/kernel/init_task.c b/arch/blackfin/kernel/init_task.c index 6bdba7b21109..2c228c020978 100644 --- a/arch/blackfin/kernel/init_task.c +++ b/arch/blackfin/kernel/init_task.c @@ -33,7 +33,6 @@ #include #include -static struct fs_struct init_fs = INIT_FS; static struct signal_struct init_signals = INIT_SIGNALS(init_signals); static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand); diff --git a/arch/cris/kernel/process.c b/arch/cris/kernel/process.c index 5933656db5a2..60816e876455 100644 --- a/arch/cris/kernel/process.c +++ b/arch/cris/kernel/process.c @@ -37,7 +37,6 @@ * setup. */ -static struct fs_struct init_fs = INIT_FS; static struct signal_struct init_signals = INIT_SIGNALS(init_signals); static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand); struct mm_struct init_mm = INIT_MM(init_mm); diff --git a/arch/frv/kernel/init_task.c b/arch/frv/kernel/init_task.c index e2198815b630..29429a8b7f6a 100644 --- a/arch/frv/kernel/init_task.c +++ b/arch/frv/kernel/init_task.c @@ -10,7 +10,6 @@ #include -static struct fs_struct init_fs = INIT_FS; static struct signal_struct init_signals = INIT_SIGNALS(init_signals); static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand); struct mm_struct init_mm = INIT_MM(init_mm); diff --git a/arch/h8300/kernel/init_task.c b/arch/h8300/kernel/init_task.c index 93a4899e46c2..cb5dc552da97 100644 --- a/arch/h8300/kernel/init_task.c +++ b/arch/h8300/kernel/init_task.c @@ -12,7 +12,6 @@ #include #include -static struct fs_struct init_fs = INIT_FS; static struct signal_struct init_signals = INIT_SIGNALS(init_signals); static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand); struct mm_struct init_mm = INIT_MM(init_mm); diff --git a/arch/ia64/kernel/init_task.c b/arch/ia64/kernel/init_task.c index 9d7e1c66faf4..5b0e830c6f33 100644 --- a/arch/ia64/kernel/init_task.c +++ b/arch/ia64/kernel/init_task.c @@ -17,7 +17,6 @@ #include #include -static struct fs_struct init_fs = INIT_FS; static struct signal_struct init_signals = INIT_SIGNALS(init_signals); static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand); struct mm_struct init_mm = INIT_MM(init_mm); diff --git a/arch/m32r/kernel/init_task.c b/arch/m32r/kernel/init_task.c index 0d658dbb6766..016885c6f260 100644 --- a/arch/m32r/kernel/init_task.c +++ b/arch/m32r/kernel/init_task.c @@ -11,7 +11,6 @@ #include #include -static struct fs_struct init_fs = INIT_FS; static struct signal_struct init_signals = INIT_SIGNALS(init_signals); static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand); struct mm_struct init_mm = INIT_MM(init_mm); diff --git a/arch/m68k/kernel/process.c b/arch/m68k/kernel/process.c index 3042c2bc8c58..632ce016014d 100644 --- a/arch/m68k/kernel/process.c +++ b/arch/m68k/kernel/process.c @@ -40,7 +40,6 @@ * alignment requirements and potentially different initial * setup. */ -static struct fs_struct init_fs = INIT_FS; static struct signal_struct init_signals = INIT_SIGNALS(init_signals); static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand); struct mm_struct init_mm = INIT_MM(init_mm); diff --git a/arch/m68knommu/kernel/init_task.c b/arch/m68knommu/kernel/init_task.c index 344c01aede08..fe282de1d596 100644 --- a/arch/m68knommu/kernel/init_task.c +++ b/arch/m68knommu/kernel/init_task.c @@ -12,7 +12,6 @@ #include #include -static struct fs_struct init_fs = INIT_FS; static struct signal_struct init_signals = INIT_SIGNALS(init_signals); static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand); struct mm_struct init_mm = INIT_MM(init_mm); diff --git a/arch/mips/kernel/init_task.c b/arch/mips/kernel/init_task.c index d72487ad7c15..149cd914526e 100644 --- a/arch/mips/kernel/init_task.c +++ b/arch/mips/kernel/init_task.c @@ -9,7 +9,6 @@ #include #include -static struct fs_struct init_fs = INIT_FS; static struct signal_struct init_signals = INIT_SIGNALS(init_signals); static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand); struct mm_struct init_mm = INIT_MM(init_mm); diff --git a/arch/mn10300/kernel/init_task.c b/arch/mn10300/kernel/init_task.c index af16f6e5c918..5ac3566f8c98 100644 --- a/arch/mn10300/kernel/init_task.c +++ b/arch/mn10300/kernel/init_task.c @@ -18,7 +18,6 @@ #include #include -static struct fs_struct init_fs = INIT_FS; static struct signal_struct init_signals = INIT_SIGNALS(init_signals); static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand); struct mm_struct init_mm = INIT_MM(init_mm); diff --git a/arch/parisc/kernel/init_task.c b/arch/parisc/kernel/init_task.c index f5941c086551..1e25a45d64c1 100644 --- a/arch/parisc/kernel/init_task.c +++ b/arch/parisc/kernel/init_task.c @@ -34,7 +34,6 @@ #include #include -static struct fs_struct init_fs = INIT_FS; static struct signal_struct init_signals = INIT_SIGNALS(init_signals); static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand); struct mm_struct init_mm = INIT_MM(init_mm); diff --git a/arch/powerpc/kernel/init_task.c b/arch/powerpc/kernel/init_task.c index 4c85b8d56478..688b329800bd 100644 --- a/arch/powerpc/kernel/init_task.c +++ b/arch/powerpc/kernel/init_task.c @@ -7,7 +7,6 @@ #include #include -static struct fs_struct init_fs = INIT_FS; static struct signal_struct init_signals = INIT_SIGNALS(init_signals); static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand); struct mm_struct init_mm = INIT_MM(init_mm); diff --git a/arch/s390/kernel/init_task.c b/arch/s390/kernel/init_task.c index e80716843619..7db95c0b8693 100644 --- a/arch/s390/kernel/init_task.c +++ b/arch/s390/kernel/init_task.c @@ -16,7 +16,6 @@ #include #include -static struct fs_struct init_fs = INIT_FS; static struct signal_struct init_signals = INIT_SIGNALS(init_signals); static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand); struct mm_struct init_mm = INIT_MM(init_mm); diff --git a/arch/sh/kernel/init_task.c b/arch/sh/kernel/init_task.c index b151a25cb14d..80c35ff71d56 100644 --- a/arch/sh/kernel/init_task.c +++ b/arch/sh/kernel/init_task.c @@ -7,7 +7,6 @@ #include #include -static struct fs_struct init_fs = INIT_FS; static struct signal_struct init_signals = INIT_SIGNALS(init_signals); static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand); struct pt_regs fake_swapper_regs; diff --git a/arch/sparc/kernel/init_task.c b/arch/sparc/kernel/init_task.c index 62126e4cec54..f28cb8278e98 100644 --- a/arch/sparc/kernel/init_task.c +++ b/arch/sparc/kernel/init_task.c @@ -8,7 +8,6 @@ #include #include -static struct fs_struct init_fs = INIT_FS; static struct signal_struct init_signals = INIT_SIGNALS(init_signals); static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand); struct mm_struct init_mm = INIT_MM(init_mm); diff --git a/arch/um/kernel/init_task.c b/arch/um/kernel/init_task.c index 910eda8fca18..806d381947bf 100644 --- a/arch/um/kernel/init_task.c +++ b/arch/um/kernel/init_task.c @@ -10,7 +10,6 @@ #include "linux/mqueue.h" #include "asm/uaccess.h" -static struct fs_struct init_fs = INIT_FS; struct mm_struct init_mm = INIT_MM(init_mm); static struct signal_struct init_signals = INIT_SIGNALS(init_signals); static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand); diff --git a/arch/x86/kernel/init_task.c b/arch/x86/kernel/init_task.c index d39918076bb4..df3bf269beab 100644 --- a/arch/x86/kernel/init_task.c +++ b/arch/x86/kernel/init_task.c @@ -10,7 +10,6 @@ #include #include -static struct fs_struct init_fs = INIT_FS; static struct signal_struct init_signals = INIT_SIGNALS(init_signals); static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand); struct mm_struct init_mm = INIT_MM(init_mm); diff --git a/arch/xtensa/kernel/init_task.c b/arch/xtensa/kernel/init_task.c index 3df469dbe814..e07f5c9fcd35 100644 --- a/arch/xtensa/kernel/init_task.c +++ b/arch/xtensa/kernel/init_task.c @@ -21,7 +21,6 @@ #include -static struct fs_struct init_fs = INIT_FS; static struct signal_struct init_signals = INIT_SIGNALS(init_signals); static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand); struct mm_struct init_mm = INIT_MM(init_mm); diff --git a/fs/namei.c b/fs/namei.c index 3f88e043d459..e203691b9d12 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -2893,3 +2893,10 @@ EXPORT_SYMBOL(vfs_symlink); EXPORT_SYMBOL(vfs_unlink); EXPORT_SYMBOL(dentry_unhash); EXPORT_SYMBOL(generic_readlink); + +/* to be mentioned only in INIT_TASK */ +struct fs_struct init_fs = { + .count = ATOMIC_INIT(1), + .lock = RW_LOCK_UNLOCKED, + .umask = 0022, +}; diff --git a/include/linux/fs_struct.h b/include/linux/fs_struct.h index 9e5a06e78d02..a97c053d3a9a 100644 --- a/include/linux/fs_struct.h +++ b/include/linux/fs_struct.h @@ -10,12 +10,6 @@ struct fs_struct { struct path root, pwd; }; -#define INIT_FS { \ - .count = ATOMIC_INIT(1), \ - .lock = RW_LOCK_UNLOCKED, \ - .umask = 0022, \ -} - extern struct kmem_cache *fs_cachep; extern void exit_fs(struct task_struct *); diff --git a/include/linux/init_task.h b/include/linux/init_task.h index 959f5522d10a..2f3c2d4ef73b 100644 --- a/include/linux/init_task.h +++ b/include/linux/init_task.h @@ -12,6 +12,7 @@ #include extern struct files_struct init_files; +extern struct fs_struct init_fs; #define INIT_KIOCTX(name, which_mm) \ { \ -- cgit v1.2.3 From c64d8996bd758cedc2ddc04b86ca66fa1d8599cf Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Fri, 2 Jan 2009 11:27:18 +0300 Subject: x86: early_printk - use sizeof instead of hardcoded number Impact: cleanup Signed-off-by: Cyrill Gorcunov Signed-off-by: Ingo Molnar --- arch/x86/kernel/early_printk.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/early_printk.c b/arch/x86/kernel/early_printk.c index 23b138e31e9c..504ad198e4ad 100644 --- a/arch/x86/kernel/early_printk.c +++ b/arch/x86/kernel/early_printk.c @@ -886,7 +886,7 @@ asmlinkage void early_printk(const char *fmt, ...) va_list ap; va_start(ap, fmt); - n = vscnprintf(buf, 512, fmt, ap); + n = vscnprintf(buf, sizeof(buf), fmt, ap); early_console->write(early_console, buf, n); va_end(ap); } -- cgit v1.2.3 From a9067d537615d534dcef06c0d819472e43a0d152 Mon Sep 17 00:00:00 2001 From: Ingo Brueckl Date: Fri, 2 Jan 2009 14:42:00 +0100 Subject: x86: convert permanent_kmaps_init() from macro to inline Impact: cleanup This compiler warning: arch/x86/mm/init_32.c:515: warning: unused variable 'pgd_base' triggers because permanent_kmaps_init() is a CPP macro in the !CONFIG_HIGHMEM case, that does not tell the compiler that the 'pgd_base' parameter is used. Convert permanent_kmaps_init() (and set_highmem_pages_init()) to C inline functions - which gives the parameter a proper type and which gets rid of the compiler warning as well. Signed-off-by: Ingo Brueckl Signed-off-by: Ingo Molnar --- arch/x86/mm/init_32.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c index 800e1d94c1b5..ad98b18f2b48 100644 --- a/arch/x86/mm/init_32.c +++ b/arch/x86/mm/init_32.c @@ -434,8 +434,12 @@ static void __init set_highmem_pages_init(void) #endif /* !CONFIG_NUMA */ #else -# define permanent_kmaps_init(pgd_base) do { } while (0) -# define set_highmem_pages_init() do { } while (0) +static inline void permanent_kmaps_init(pgd_t *pgd_base) +{ +} +static inline void set_highmem_pages_init(void) +{ +} #endif /* CONFIG_HIGHMEM */ void __init native_pagetable_setup_start(pgd_t *base) -- cgit v1.2.3 From 26799a63110dcbe81291ea53178f6b4810d07424 Mon Sep 17 00:00:00 2001 From: Ravikiran G Thirumalai Date: Wed, 31 Dec 2008 13:44:46 -0800 Subject: x86: fix incorrect __read_mostly on _boot_cpu_pda The pda rework (commit 3461b0af025251bbc6b3d56c821c6ac2de6f7209) to remove static boot cpu pdas introduced a performance bug. _boot_cpu_pda is the actual pda used by the boot cpu and is definitely not "__read_mostly" and ended up polluting the read mostly section with writes. This bug caused regression of about 8-10% on certain syscall intensive workloads. Signed-off-by: Ravikiran Thirumalai Acked-by: Mike Travis Cc: Signed-off-by: Ingo Molnar --- arch/x86/kernel/head64.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/head64.c b/arch/x86/kernel/head64.c index 388e05a5fc17..b9a4d8c4b935 100644 --- a/arch/x86/kernel/head64.c +++ b/arch/x86/kernel/head64.c @@ -27,7 +27,7 @@ #include /* boot cpu pda */ -static struct x8664_pda _boot_cpu_pda __read_mostly; +static struct x8664_pda _boot_cpu_pda; #ifdef CONFIG_SMP /* -- cgit v1.2.3 From 46814dded1b972a07b1609d81632eef3009fbb10 Mon Sep 17 00:00:00 2001 From: Cliff Wickman Date: Wed, 31 Dec 2008 13:20:50 -0600 Subject: x86, UV: remove erroneous BAU initialization Impact: fix crash on x86/UV UV is the SGI "UltraViolet" machine, which is x86_64 based. BAU is the "Broadcast Assist Unit", used for TLB shootdown in UV. This patch removes the allocation and initialization of an unused table. This table is left over from a development test mode. It is unused in the present code. And it was incorrectly initialized: 8 entries allocated but 17 initialized, causing slab corruption. This patch should go into 2.6.27 and 2.6.28 as well as the current tree. Diffed against 2.6.28 (linux-next, 12/30/08) Signed-off-by: Cliff Wickman Cc: Signed-off-by: Ingo Molnar --- arch/x86/kernel/tlb_uv.c | 9 --------- 1 file changed, 9 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/tlb_uv.c b/arch/x86/kernel/tlb_uv.c index 6a00e5faaa74..f885023167e0 100644 --- a/arch/x86/kernel/tlb_uv.c +++ b/arch/x86/kernel/tlb_uv.c @@ -582,7 +582,6 @@ static int __init uv_ptc_init(void) static struct bau_control * __init uv_table_bases_init(int blade, int node) { int i; - int *ip; struct bau_msg_status *msp; struct bau_control *bau_tabp; @@ -599,13 +598,6 @@ static struct bau_control * __init uv_table_bases_init(int blade, int node) bau_cpubits_clear(&msp->seen_by, (int) uv_blade_nr_possible_cpus(blade)); - bau_tabp->watching = - kmalloc_node(sizeof(int) * DEST_NUM_RESOURCES, GFP_KERNEL, node); - BUG_ON(!bau_tabp->watching); - - for (i = 0, ip = bau_tabp->watching; i < DEST_Q_SIZE; i++, ip++) - *ip = 0; - uv_bau_table_bases[blade] = bau_tabp; return bau_tabp; @@ -628,7 +620,6 @@ uv_table_bases_finish(int blade, int node, int cur_cpu, bcp->bau_msg_head = bau_tablesp->va_queue_first; bcp->va_queue_first = bau_tablesp->va_queue_first; bcp->va_queue_last = bau_tablesp->va_queue_last; - bcp->watching = bau_tablesp->watching; bcp->msg_statuses = bau_tablesp->msg_statuses; bcp->descriptor_base = adp; } -- cgit v1.2.3 From f634fa941188a91dbf1dab961fe7a4509852fd6e Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Wed, 31 Dec 2008 16:29:48 +0530 Subject: x86: cpuid.c fix style problems Impact: cleanup Fixes style problems: WARNING: Use #include instead of ERROR: "foo * bar" should be "foo *bar" ERROR: trailing whitespace WARNING: usage of NR_CPUS is often wrong - consider using cpu_possible(), num_possible_cpus(), for_each_possible_cpu(), etc total: 2 errors, 2 warnings Signed-off-by: Jaswinder Singh Rajput Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpuid.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpuid.c b/arch/x86/kernel/cpuid.c index 72cefd1e649b..85d28d53f5d3 100644 --- a/arch/x86/kernel/cpuid.c +++ b/arch/x86/kernel/cpuid.c @@ -39,10 +39,10 @@ #include #include #include +#include #include #include -#include #include static struct class *cpuid_class; @@ -82,7 +82,7 @@ static loff_t cpuid_seek(struct file *file, loff_t offset, int orig) } static ssize_t cpuid_read(struct file *file, char __user *buf, - size_t count, loff_t * ppos) + size_t count, loff_t *ppos) { char __user *tmp = buf; struct cpuid_regs cmd; @@ -117,7 +117,7 @@ static int cpuid_open(struct inode *inode, struct file *file) unsigned int cpu; struct cpuinfo_x86 *c; int ret = 0; - + lock_kernel(); cpu = iminor(file->f_path.dentry->d_inode); -- cgit v1.2.3 From 423a54058f746579aff1430877dbc82f17442b34 Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Wed, 31 Dec 2008 16:42:20 +0530 Subject: x86: ldt.c fix style problems Impact: cleanup Fixes style problems: WARNING: Use #include instead of ERROR: space required before the open parenthesis '(' total: 1 errors, 1 warnings Signed-off-by: Jaswinder Singh Rajput Signed-off-by: Ingo Molnar --- arch/x86/kernel/ldt.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/ldt.c b/arch/x86/kernel/ldt.c index eee32b43fee3..71f1d99a635d 100644 --- a/arch/x86/kernel/ldt.c +++ b/arch/x86/kernel/ldt.c @@ -12,8 +12,8 @@ #include #include #include +#include -#include #include #include #include @@ -93,7 +93,7 @@ static inline int copy_ldt(mm_context_t *new, mm_context_t *old) if (err < 0) return err; - for(i = 0; i < old->size; i++) + for (i = 0; i < old->size; i++) write_ldt_entry(new->ldt, i, old->ldt + i * LDT_ENTRY_SIZE); return 0; } -- cgit v1.2.3 From dceb4521c8ed24b9fe4230e0d385cf4770260383 Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Wed, 31 Dec 2008 17:35:02 +0530 Subject: x86: nmi.c fix style problems Impact: cleanup, fix style problems Fixes style problems: WARNING: Use #include instead of WARNING: Use #include instead of total: 0 errors, 2 warnings Signed-off-by: Jaswinder Singh Rajput Signed-off-by: Ingo Molnar --- arch/x86/kernel/nmi.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/nmi.c b/arch/x86/kernel/nmi.c index 8bd1bf9622a7..45a09ccdc214 100644 --- a/arch/x86/kernel/nmi.c +++ b/arch/x86/kernel/nmi.c @@ -26,11 +26,10 @@ #include #include #include +#include #include #include -#include -#include #include #include -- cgit v1.2.3 From 103ceffb9501531f6931df6aebc11a05189201f0 Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Fri, 2 Jan 2009 23:43:25 +0530 Subject: x86: mpparse.c fix style problems Impact: cleanup, fix style problems, more readable Fixes style problems: WARNING: Use #include instead of WARNING: Use #include instead of WARNING: suspect code indent for conditional statements (8, 17) WARNING: space prohibited between function name and open parenthesis '(' total: 0 errors, 5 warnings Signed-off-by: Jaswinder Singh Rajput Signed-off-by: Ingo Molnar --- arch/x86/kernel/mpparse.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/mpparse.c b/arch/x86/kernel/mpparse.c index 45e3b69808ba..c5c5b8df1dbc 100644 --- a/arch/x86/kernel/mpparse.c +++ b/arch/x86/kernel/mpparse.c @@ -16,14 +16,14 @@ #include #include #include +#include +#include -#include #include #include #include #include #include -#include #include #include #include @@ -95,8 +95,8 @@ static void __init MP_bus_info(struct mpc_config_bus *m) #endif if (strncmp(str, BUSTYPE_ISA, sizeof(BUSTYPE_ISA) - 1) == 0) { - set_bit(m->mpc_busid, mp_bus_not_pci); -#if defined(CONFIG_EISA) || defined (CONFIG_MCA) + set_bit(m->mpc_busid, mp_bus_not_pci); +#if defined(CONFIG_EISA) || defined(CONFIG_MCA) mp_bus_id_to_type[m->mpc_busid] = MP_BUS_ISA; #endif } else if (strncmp(str, BUSTYPE_PCI, sizeof(BUSTYPE_PCI) - 1) == 0) { @@ -104,7 +104,7 @@ static void __init MP_bus_info(struct mpc_config_bus *m) x86_quirks->mpc_oem_pci_bus(m); clear_bit(m->mpc_busid, mp_bus_not_pci); -#if defined(CONFIG_EISA) || defined (CONFIG_MCA) +#if defined(CONFIG_EISA) || defined(CONFIG_MCA) mp_bus_id_to_type[m->mpc_busid] = MP_BUS_PCI; } else if (strncmp(str, BUSTYPE_EISA, sizeof(BUSTYPE_EISA) - 1) == 0) { mp_bus_id_to_type[m->mpc_busid] = MP_BUS_EISA; -- cgit v1.2.3 From e8e32326279cba3d049b4325111f76618953195c Mon Sep 17 00:00:00 2001 From: Ingo Brueckl Date: Fri, 2 Jan 2009 14:42:00 +0100 Subject: Fix compiler warning in arch/x86/mm/init_32.c Signed-off-by: Ingo Brueckl Signed-off-by: Linus Torvalds --- arch/x86/mm/init_32.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c index 8655b5bb0963..f99a6c6c432e 100644 --- a/arch/x86/mm/init_32.c +++ b/arch/x86/mm/init_32.c @@ -435,8 +435,12 @@ static void __init set_highmem_pages_init(void) #endif /* !CONFIG_NUMA */ #else -# define permanent_kmaps_init(pgd_base) do { } while (0) -# define set_highmem_pages_init() do { } while (0) +static inline void permanent_kmaps_init(pgd_t *pgd_base) +{ +} +static inline void set_highmem_pages_init(void) +{ +} #endif /* CONFIG_HIGHMEM */ void __init native_pagetable_setup_start(pgd_t *base) -- cgit v1.2.3 From c4fa3864281c7d88b7262cbc6cbd5c90bb59860e Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Wed, 26 Nov 2008 15:51:19 +0100 Subject: KVM: rename vtd.c to iommu.c Impact: file renamed The code in the vtd.c file can be reused for other IOMMUs as well. So rename it to make it clear that it handle more than VT-d. Signed-off-by: Joerg Roedel --- arch/ia64/kvm/Makefile | 2 +- arch/x86/kvm/Makefile | 2 +- virt/kvm/iommu.c | 217 +++++++++++++++++++++++++++++++++++++++++++++++++ virt/kvm/vtd.c | 217 ------------------------------------------------- 4 files changed, 219 insertions(+), 219 deletions(-) create mode 100644 virt/kvm/iommu.c delete mode 100644 virt/kvm/vtd.c (limited to 'arch/x86') diff --git a/arch/ia64/kvm/Makefile b/arch/ia64/kvm/Makefile index 76464dc312e6..cb69dfcfb1a6 100644 --- a/arch/ia64/kvm/Makefile +++ b/arch/ia64/kvm/Makefile @@ -52,7 +52,7 @@ common-objs = $(addprefix ../../../virt/kvm/, kvm_main.o ioapic.o \ coalesced_mmio.o irq_comm.o) ifeq ($(CONFIG_DMAR),y) -common-objs += $(addprefix ../../../virt/kvm/, vtd.o) +common-objs += $(addprefix ../../../virt/kvm/, iommu.o) endif kvm-objs := $(common-objs) kvm-ia64.o kvm_fw.o diff --git a/arch/x86/kvm/Makefile b/arch/x86/kvm/Makefile index c02343594b4d..00f46c2d14bd 100644 --- a/arch/x86/kvm/Makefile +++ b/arch/x86/kvm/Makefile @@ -8,7 +8,7 @@ ifeq ($(CONFIG_KVM_TRACE),y) common-objs += $(addprefix ../../../virt/kvm/, kvm_trace.o) endif ifeq ($(CONFIG_DMAR),y) -common-objs += $(addprefix ../../../virt/kvm/, vtd.o) +common-objs += $(addprefix ../../../virt/kvm/, iommu.o) endif EXTRA_CFLAGS += -Ivirt/kvm -Iarch/x86/kvm diff --git a/virt/kvm/iommu.c b/virt/kvm/iommu.c new file mode 100644 index 000000000000..d46de9af838b --- /dev/null +++ b/virt/kvm/iommu.c @@ -0,0 +1,217 @@ +/* + * Copyright (c) 2006, Intel Corporation. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., 59 Temple + * Place - Suite 330, Boston, MA 02111-1307 USA. + * + * Copyright (C) 2006-2008 Intel Corporation + * Copyright IBM Corporation, 2008 + * Author: Allen M. Kay + * Author: Weidong Han + * Author: Ben-Ami Yassour + */ + +#include +#include +#include +#include +#include + +static int kvm_iommu_unmap_memslots(struct kvm *kvm); +static void kvm_iommu_put_pages(struct kvm *kvm, + gfn_t base_gfn, unsigned long npages); + +int kvm_iommu_map_pages(struct kvm *kvm, + gfn_t base_gfn, unsigned long npages) +{ + gfn_t gfn = base_gfn; + pfn_t pfn; + int i, r = 0; + struct dmar_domain *domain = kvm->arch.intel_iommu_domain; + + /* check if iommu exists and in use */ + if (!domain) + return 0; + + for (i = 0; i < npages; i++) { + /* check if already mapped */ + if (intel_iommu_iova_to_phys(domain, + gfn_to_gpa(gfn))) + continue; + + pfn = gfn_to_pfn(kvm, gfn); + r = intel_iommu_map_address(domain, + gfn_to_gpa(gfn), + pfn_to_hpa(pfn), + PAGE_SIZE, + DMA_PTE_READ | DMA_PTE_WRITE); + if (r) { + printk(KERN_ERR "kvm_iommu_map_address:" + "iommu failed to map pfn=%lx\n", pfn); + goto unmap_pages; + } + gfn++; + } + return 0; + +unmap_pages: + kvm_iommu_put_pages(kvm, base_gfn, i); + return r; +} + +static int kvm_iommu_map_memslots(struct kvm *kvm) +{ + int i, r; + + down_read(&kvm->slots_lock); + for (i = 0; i < kvm->nmemslots; i++) { + r = kvm_iommu_map_pages(kvm, kvm->memslots[i].base_gfn, + kvm->memslots[i].npages); + if (r) + break; + } + up_read(&kvm->slots_lock); + return r; +} + +int kvm_assign_device(struct kvm *kvm, + struct kvm_assigned_dev_kernel *assigned_dev) +{ + struct pci_dev *pdev = NULL; + struct dmar_domain *domain = kvm->arch.intel_iommu_domain; + int r; + + /* check if iommu exists and in use */ + if (!domain) + return 0; + + pdev = assigned_dev->dev; + if (pdev == NULL) + return -ENODEV; + + r = intel_iommu_attach_device(domain, pdev); + if (r) { + printk(KERN_ERR "assign device %x:%x.%x failed", + pdev->bus->number, + PCI_SLOT(pdev->devfn), + PCI_FUNC(pdev->devfn)); + return r; + } + + printk(KERN_DEBUG "assign device: host bdf = %x:%x:%x\n", + assigned_dev->host_busnr, + PCI_SLOT(assigned_dev->host_devfn), + PCI_FUNC(assigned_dev->host_devfn)); + + return 0; +} + +int kvm_deassign_device(struct kvm *kvm, + struct kvm_assigned_dev_kernel *assigned_dev) +{ + struct dmar_domain *domain = kvm->arch.intel_iommu_domain; + struct pci_dev *pdev = NULL; + + /* check if iommu exists and in use */ + if (!domain) + return 0; + + pdev = assigned_dev->dev; + if (pdev == NULL) + return -ENODEV; + + intel_iommu_detach_device(domain, pdev); + + printk(KERN_DEBUG "deassign device: host bdf = %x:%x:%x\n", + assigned_dev->host_busnr, + PCI_SLOT(assigned_dev->host_devfn), + PCI_FUNC(assigned_dev->host_devfn)); + + return 0; +} + +int kvm_iommu_map_guest(struct kvm *kvm) +{ + int r; + + if (!intel_iommu_found()) { + printk(KERN_ERR "%s: intel iommu not found\n", __func__); + return -ENODEV; + } + + kvm->arch.intel_iommu_domain = intel_iommu_alloc_domain(); + if (!kvm->arch.intel_iommu_domain) + return -ENOMEM; + + r = kvm_iommu_map_memslots(kvm); + if (r) + goto out_unmap; + + return 0; + +out_unmap: + kvm_iommu_unmap_memslots(kvm); + return r; +} + +static void kvm_iommu_put_pages(struct kvm *kvm, + gfn_t base_gfn, unsigned long npages) +{ + gfn_t gfn = base_gfn; + pfn_t pfn; + struct dmar_domain *domain = kvm->arch.intel_iommu_domain; + unsigned long i; + u64 phys; + + /* check if iommu exists and in use */ + if (!domain) + return; + + for (i = 0; i < npages; i++) { + phys = intel_iommu_iova_to_phys(domain, + gfn_to_gpa(gfn)); + pfn = phys >> PAGE_SHIFT; + kvm_release_pfn_clean(pfn); + gfn++; + } + + intel_iommu_unmap_address(domain, + gfn_to_gpa(base_gfn), + PAGE_SIZE * npages); +} + +static int kvm_iommu_unmap_memslots(struct kvm *kvm) +{ + int i; + down_read(&kvm->slots_lock); + for (i = 0; i < kvm->nmemslots; i++) { + kvm_iommu_put_pages(kvm, kvm->memslots[i].base_gfn, + kvm->memslots[i].npages); + } + up_read(&kvm->slots_lock); + + return 0; +} + +int kvm_iommu_unmap_guest(struct kvm *kvm) +{ + struct dmar_domain *domain = kvm->arch.intel_iommu_domain; + + /* check if iommu exists and in use */ + if (!domain) + return 0; + + kvm_iommu_unmap_memslots(kvm); + intel_iommu_free_domain(domain); + return 0; +} diff --git a/virt/kvm/vtd.c b/virt/kvm/vtd.c deleted file mode 100644 index d46de9af838b..000000000000 --- a/virt/kvm/vtd.c +++ /dev/null @@ -1,217 +0,0 @@ -/* - * Copyright (c) 2006, Intel Corporation. - * - * This program is free software; you can redistribute it and/or modify it - * under the terms and conditions of the GNU General Public License, - * version 2, as published by the Free Software Foundation. - * - * This program is distributed in the hope it will be useful, but WITHOUT - * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or - * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for - * more details. - * - * You should have received a copy of the GNU General Public License along with - * this program; if not, write to the Free Software Foundation, Inc., 59 Temple - * Place - Suite 330, Boston, MA 02111-1307 USA. - * - * Copyright (C) 2006-2008 Intel Corporation - * Copyright IBM Corporation, 2008 - * Author: Allen M. Kay - * Author: Weidong Han - * Author: Ben-Ami Yassour - */ - -#include -#include -#include -#include -#include - -static int kvm_iommu_unmap_memslots(struct kvm *kvm); -static void kvm_iommu_put_pages(struct kvm *kvm, - gfn_t base_gfn, unsigned long npages); - -int kvm_iommu_map_pages(struct kvm *kvm, - gfn_t base_gfn, unsigned long npages) -{ - gfn_t gfn = base_gfn; - pfn_t pfn; - int i, r = 0; - struct dmar_domain *domain = kvm->arch.intel_iommu_domain; - - /* check if iommu exists and in use */ - if (!domain) - return 0; - - for (i = 0; i < npages; i++) { - /* check if already mapped */ - if (intel_iommu_iova_to_phys(domain, - gfn_to_gpa(gfn))) - continue; - - pfn = gfn_to_pfn(kvm, gfn); - r = intel_iommu_map_address(domain, - gfn_to_gpa(gfn), - pfn_to_hpa(pfn), - PAGE_SIZE, - DMA_PTE_READ | DMA_PTE_WRITE); - if (r) { - printk(KERN_ERR "kvm_iommu_map_address:" - "iommu failed to map pfn=%lx\n", pfn); - goto unmap_pages; - } - gfn++; - } - return 0; - -unmap_pages: - kvm_iommu_put_pages(kvm, base_gfn, i); - return r; -} - -static int kvm_iommu_map_memslots(struct kvm *kvm) -{ - int i, r; - - down_read(&kvm->slots_lock); - for (i = 0; i < kvm->nmemslots; i++) { - r = kvm_iommu_map_pages(kvm, kvm->memslots[i].base_gfn, - kvm->memslots[i].npages); - if (r) - break; - } - up_read(&kvm->slots_lock); - return r; -} - -int kvm_assign_device(struct kvm *kvm, - struct kvm_assigned_dev_kernel *assigned_dev) -{ - struct pci_dev *pdev = NULL; - struct dmar_domain *domain = kvm->arch.intel_iommu_domain; - int r; - - /* check if iommu exists and in use */ - if (!domain) - return 0; - - pdev = assigned_dev->dev; - if (pdev == NULL) - return -ENODEV; - - r = intel_iommu_attach_device(domain, pdev); - if (r) { - printk(KERN_ERR "assign device %x:%x.%x failed", - pdev->bus->number, - PCI_SLOT(pdev->devfn), - PCI_FUNC(pdev->devfn)); - return r; - } - - printk(KERN_DEBUG "assign device: host bdf = %x:%x:%x\n", - assigned_dev->host_busnr, - PCI_SLOT(assigned_dev->host_devfn), - PCI_FUNC(assigned_dev->host_devfn)); - - return 0; -} - -int kvm_deassign_device(struct kvm *kvm, - struct kvm_assigned_dev_kernel *assigned_dev) -{ - struct dmar_domain *domain = kvm->arch.intel_iommu_domain; - struct pci_dev *pdev = NULL; - - /* check if iommu exists and in use */ - if (!domain) - return 0; - - pdev = assigned_dev->dev; - if (pdev == NULL) - return -ENODEV; - - intel_iommu_detach_device(domain, pdev); - - printk(KERN_DEBUG "deassign device: host bdf = %x:%x:%x\n", - assigned_dev->host_busnr, - PCI_SLOT(assigned_dev->host_devfn), - PCI_FUNC(assigned_dev->host_devfn)); - - return 0; -} - -int kvm_iommu_map_guest(struct kvm *kvm) -{ - int r; - - if (!intel_iommu_found()) { - printk(KERN_ERR "%s: intel iommu not found\n", __func__); - return -ENODEV; - } - - kvm->arch.intel_iommu_domain = intel_iommu_alloc_domain(); - if (!kvm->arch.intel_iommu_domain) - return -ENOMEM; - - r = kvm_iommu_map_memslots(kvm); - if (r) - goto out_unmap; - - return 0; - -out_unmap: - kvm_iommu_unmap_memslots(kvm); - return r; -} - -static void kvm_iommu_put_pages(struct kvm *kvm, - gfn_t base_gfn, unsigned long npages) -{ - gfn_t gfn = base_gfn; - pfn_t pfn; - struct dmar_domain *domain = kvm->arch.intel_iommu_domain; - unsigned long i; - u64 phys; - - /* check if iommu exists and in use */ - if (!domain) - return; - - for (i = 0; i < npages; i++) { - phys = intel_iommu_iova_to_phys(domain, - gfn_to_gpa(gfn)); - pfn = phys >> PAGE_SHIFT; - kvm_release_pfn_clean(pfn); - gfn++; - } - - intel_iommu_unmap_address(domain, - gfn_to_gpa(base_gfn), - PAGE_SIZE * npages); -} - -static int kvm_iommu_unmap_memslots(struct kvm *kvm) -{ - int i; - down_read(&kvm->slots_lock); - for (i = 0; i < kvm->nmemslots; i++) { - kvm_iommu_put_pages(kvm, kvm->memslots[i].base_gfn, - kvm->memslots[i].npages); - } - up_read(&kvm->slots_lock); - - return 0; -} - -int kvm_iommu_unmap_guest(struct kvm *kvm) -{ - struct dmar_domain *domain = kvm->arch.intel_iommu_domain; - - /* check if iommu exists and in use */ - if (!domain) - return 0; - - kvm_iommu_unmap_memslots(kvm); - intel_iommu_free_domain(domain); - return 0; -} -- cgit v1.2.3 From 1aaf118352b85bb359ce28070bcc478f659a7031 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Wed, 26 Nov 2008 17:25:13 +0100 Subject: select IOMMU_API when DMAR and/or AMD_IOMMU is selected These two IOMMUs can implement the current version of this API. So select the API if one or both of these IOMMU drivers is selected. Signed-off-by: Joerg Roedel --- arch/ia64/Kconfig | 3 +++ arch/x86/Kconfig | 3 +++ drivers/base/Makefile | 1 + 3 files changed, 7 insertions(+) (limited to 'arch/x86') diff --git a/arch/ia64/Kconfig b/arch/ia64/Kconfig index 7fa8f615ba6e..3d31636cbafb 100644 --- a/arch/ia64/Kconfig +++ b/arch/ia64/Kconfig @@ -687,3 +687,6 @@ config IRQ_PER_CPU config IOMMU_HELPER def_bool (IA64_HP_ZX1 || IA64_HP_ZX1_SWIOTLB || IA64_GENERIC || SWIOTLB) + +config IOMMU_API + def_bool (DMAR) diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 249d1e0824b5..4737435b00d4 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -599,6 +599,9 @@ config SWIOTLB config IOMMU_HELPER def_bool (CALGARY_IOMMU || GART_IOMMU || SWIOTLB || AMD_IOMMU) +config IOMMU_API + def_bool (AMD_IOMMU || DMAR) + config MAXSMP bool "Configure Maximum number of SMP Processors and NUMA Nodes" depends on X86_64 && SMP && DEBUG_KERNEL && EXPERIMENTAL diff --git a/drivers/base/Makefile b/drivers/base/Makefile index c66637392bbc..b5b8ba512b28 100644 --- a/drivers/base/Makefile +++ b/drivers/base/Makefile @@ -11,6 +11,7 @@ obj-$(CONFIG_FW_LOADER) += firmware_class.o obj-$(CONFIG_NUMA) += node.o obj-$(CONFIG_MEMORY_HOTPLUG_SPARSE) += memory.o obj-$(CONFIG_SMP) += topology.o +obj-$(CONFIG_IOMMU_API) += iommu.o ifeq ($(CONFIG_SYSFS),y) obj-$(CONFIG_MODULES) += module.o endif -- cgit v1.2.3 From 19de40a8472fa64693eab844911eec277d489f6c Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Wed, 3 Dec 2008 14:43:34 +0100 Subject: KVM: change KVM to use IOMMU API Signed-off-by: Joerg Roedel --- arch/ia64/include/asm/kvm_host.h | 2 +- arch/ia64/kvm/Makefile | 2 +- arch/ia64/kvm/kvm-ia64.c | 3 ++- arch/x86/include/asm/kvm_host.h | 2 +- arch/x86/kvm/Makefile | 2 +- arch/x86/kvm/x86.c | 3 ++- include/linux/kvm_host.h | 6 +++--- virt/kvm/iommu.c | 45 +++++++++++++++++++--------------------- virt/kvm/kvm_main.c | 2 +- 9 files changed, 33 insertions(+), 34 deletions(-) (limited to 'arch/x86') diff --git a/arch/ia64/include/asm/kvm_host.h b/arch/ia64/include/asm/kvm_host.h index 0560f3fae538..348663661659 100644 --- a/arch/ia64/include/asm/kvm_host.h +++ b/arch/ia64/include/asm/kvm_host.h @@ -467,7 +467,7 @@ struct kvm_arch { struct kvm_sal_data rdv_sal_data; struct list_head assigned_dev_head; - struct dmar_domain *intel_iommu_domain; + struct iommu_domain *iommu_domain; struct hlist_head irq_ack_notifier_list; unsigned long irq_sources_bitmap; diff --git a/arch/ia64/kvm/Makefile b/arch/ia64/kvm/Makefile index cb69dfcfb1a6..0bb99b732908 100644 --- a/arch/ia64/kvm/Makefile +++ b/arch/ia64/kvm/Makefile @@ -51,7 +51,7 @@ EXTRA_AFLAGS += -Ivirt/kvm -Iarch/ia64/kvm/ common-objs = $(addprefix ../../../virt/kvm/, kvm_main.o ioapic.o \ coalesced_mmio.o irq_comm.o) -ifeq ($(CONFIG_DMAR),y) +ifeq ($(CONFIG_IOMMU_API),y) common-objs += $(addprefix ../../../virt/kvm/, iommu.o) endif diff --git a/arch/ia64/kvm/kvm-ia64.c b/arch/ia64/kvm/kvm-ia64.c index 0f5ebd948437..4e586f6110aa 100644 --- a/arch/ia64/kvm/kvm-ia64.c +++ b/arch/ia64/kvm/kvm-ia64.c @@ -31,6 +31,7 @@ #include #include #include +#include #include #include @@ -188,7 +189,7 @@ int kvm_dev_ioctl_check_extension(long ext) r = KVM_COALESCED_MMIO_PAGE_OFFSET; break; case KVM_CAP_IOMMU: - r = intel_iommu_found(); + r = iommu_found(); break; default: r = 0; diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index 97215a458e5f..730843d1d2fb 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -360,7 +360,7 @@ struct kvm_arch{ struct list_head active_mmu_pages; struct list_head assigned_dev_head; struct list_head oos_global_pages; - struct dmar_domain *intel_iommu_domain; + struct iommu_domain *iommu_domain; struct kvm_pic *vpic; struct kvm_ioapic *vioapic; struct kvm_pit *vpit; diff --git a/arch/x86/kvm/Makefile b/arch/x86/kvm/Makefile index 00f46c2d14bd..d3ec292f00f2 100644 --- a/arch/x86/kvm/Makefile +++ b/arch/x86/kvm/Makefile @@ -7,7 +7,7 @@ common-objs = $(addprefix ../../../virt/kvm/, kvm_main.o ioapic.o \ ifeq ($(CONFIG_KVM_TRACE),y) common-objs += $(addprefix ../../../virt/kvm/, kvm_trace.o) endif -ifeq ($(CONFIG_DMAR),y) +ifeq ($(CONFIG_IOMMU_API),y) common-objs += $(addprefix ../../../virt/kvm/, iommu.o) endif diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 0e6aa8141dcd..cc17546a2406 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -34,6 +34,7 @@ #include #include #include +#include #include #include @@ -989,7 +990,7 @@ int kvm_dev_ioctl_check_extension(long ext) r = !tdp_enabled; break; case KVM_CAP_IOMMU: - r = intel_iommu_found(); + r = iommu_found(); break; default: r = 0; diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h index e62a4629e51c..ec49d0be7f52 100644 --- a/include/linux/kvm_host.h +++ b/include/linux/kvm_host.h @@ -328,7 +328,7 @@ void kvm_unregister_irq_ack_notifier(struct kvm_irq_ack_notifier *kian); int kvm_request_irq_source_id(struct kvm *kvm); void kvm_free_irq_source_id(struct kvm *kvm, int irq_source_id); -#ifdef CONFIG_DMAR +#ifdef CONFIG_IOMMU_API int kvm_iommu_map_pages(struct kvm *kvm, gfn_t base_gfn, unsigned long npages); int kvm_iommu_map_guest(struct kvm *kvm); @@ -337,7 +337,7 @@ int kvm_assign_device(struct kvm *kvm, struct kvm_assigned_dev_kernel *assigned_dev); int kvm_deassign_device(struct kvm *kvm, struct kvm_assigned_dev_kernel *assigned_dev); -#else /* CONFIG_DMAR */ +#else /* CONFIG_IOMMU_API */ static inline int kvm_iommu_map_pages(struct kvm *kvm, gfn_t base_gfn, unsigned long npages) @@ -366,7 +366,7 @@ static inline int kvm_deassign_device(struct kvm *kvm, { return 0; } -#endif /* CONFIG_DMAR */ +#endif /* CONFIG_IOMMU_API */ static inline void kvm_guest_enter(void) { diff --git a/virt/kvm/iommu.c b/virt/kvm/iommu.c index d46de9af838b..d0bebaa5bf0a 100644 --- a/virt/kvm/iommu.c +++ b/virt/kvm/iommu.c @@ -25,6 +25,7 @@ #include #include #include +#include #include static int kvm_iommu_unmap_memslots(struct kvm *kvm); @@ -37,7 +38,7 @@ int kvm_iommu_map_pages(struct kvm *kvm, gfn_t gfn = base_gfn; pfn_t pfn; int i, r = 0; - struct dmar_domain *domain = kvm->arch.intel_iommu_domain; + struct iommu_domain *domain = kvm->arch.iommu_domain; /* check if iommu exists and in use */ if (!domain) @@ -45,16 +46,15 @@ int kvm_iommu_map_pages(struct kvm *kvm, for (i = 0; i < npages; i++) { /* check if already mapped */ - if (intel_iommu_iova_to_phys(domain, - gfn_to_gpa(gfn))) + if (iommu_iova_to_phys(domain, gfn_to_gpa(gfn))) continue; pfn = gfn_to_pfn(kvm, gfn); - r = intel_iommu_map_address(domain, - gfn_to_gpa(gfn), - pfn_to_hpa(pfn), - PAGE_SIZE, - DMA_PTE_READ | DMA_PTE_WRITE); + r = iommu_map_range(domain, + gfn_to_gpa(gfn), + pfn_to_hpa(pfn), + PAGE_SIZE, + IOMMU_READ | IOMMU_WRITE); if (r) { printk(KERN_ERR "kvm_iommu_map_address:" "iommu failed to map pfn=%lx\n", pfn); @@ -88,7 +88,7 @@ int kvm_assign_device(struct kvm *kvm, struct kvm_assigned_dev_kernel *assigned_dev) { struct pci_dev *pdev = NULL; - struct dmar_domain *domain = kvm->arch.intel_iommu_domain; + struct iommu_domain *domain = kvm->arch.iommu_domain; int r; /* check if iommu exists and in use */ @@ -99,7 +99,7 @@ int kvm_assign_device(struct kvm *kvm, if (pdev == NULL) return -ENODEV; - r = intel_iommu_attach_device(domain, pdev); + r = iommu_attach_device(domain, &pdev->dev); if (r) { printk(KERN_ERR "assign device %x:%x.%x failed", pdev->bus->number, @@ -119,7 +119,7 @@ int kvm_assign_device(struct kvm *kvm, int kvm_deassign_device(struct kvm *kvm, struct kvm_assigned_dev_kernel *assigned_dev) { - struct dmar_domain *domain = kvm->arch.intel_iommu_domain; + struct iommu_domain *domain = kvm->arch.iommu_domain; struct pci_dev *pdev = NULL; /* check if iommu exists and in use */ @@ -130,7 +130,7 @@ int kvm_deassign_device(struct kvm *kvm, if (pdev == NULL) return -ENODEV; - intel_iommu_detach_device(domain, pdev); + iommu_detach_device(domain, &pdev->dev); printk(KERN_DEBUG "deassign device: host bdf = %x:%x:%x\n", assigned_dev->host_busnr, @@ -144,13 +144,13 @@ int kvm_iommu_map_guest(struct kvm *kvm) { int r; - if (!intel_iommu_found()) { - printk(KERN_ERR "%s: intel iommu not found\n", __func__); + if (!iommu_found()) { + printk(KERN_ERR "%s: iommu not found\n", __func__); return -ENODEV; } - kvm->arch.intel_iommu_domain = intel_iommu_alloc_domain(); - if (!kvm->arch.intel_iommu_domain) + kvm->arch.iommu_domain = iommu_domain_alloc(); + if (!kvm->arch.iommu_domain) return -ENOMEM; r = kvm_iommu_map_memslots(kvm); @@ -169,7 +169,7 @@ static void kvm_iommu_put_pages(struct kvm *kvm, { gfn_t gfn = base_gfn; pfn_t pfn; - struct dmar_domain *domain = kvm->arch.intel_iommu_domain; + struct iommu_domain *domain = kvm->arch.iommu_domain; unsigned long i; u64 phys; @@ -178,16 +178,13 @@ static void kvm_iommu_put_pages(struct kvm *kvm, return; for (i = 0; i < npages; i++) { - phys = intel_iommu_iova_to_phys(domain, - gfn_to_gpa(gfn)); + phys = iommu_iova_to_phys(domain, gfn_to_gpa(gfn)); pfn = phys >> PAGE_SHIFT; kvm_release_pfn_clean(pfn); gfn++; } - intel_iommu_unmap_address(domain, - gfn_to_gpa(base_gfn), - PAGE_SIZE * npages); + iommu_unmap_range(domain, gfn_to_gpa(base_gfn), PAGE_SIZE * npages); } static int kvm_iommu_unmap_memslots(struct kvm *kvm) @@ -205,13 +202,13 @@ static int kvm_iommu_unmap_memslots(struct kvm *kvm) int kvm_iommu_unmap_guest(struct kvm *kvm) { - struct dmar_domain *domain = kvm->arch.intel_iommu_domain; + struct iommu_domain *domain = kvm->arch.iommu_domain; /* check if iommu exists and in use */ if (!domain) return 0; kvm_iommu_unmap_memslots(kvm); - intel_iommu_free_domain(domain); + iommu_domain_free(domain); return 0; } diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c index 4ef0fb43d1f9..3a5a08298aab 100644 --- a/virt/kvm/kvm_main.c +++ b/virt/kvm/kvm_main.c @@ -504,7 +504,7 @@ static int kvm_vm_ioctl_assign_device(struct kvm *kvm, list_add(&match->list, &kvm->arch.assigned_dev_head); if (assigned_dev->flags & KVM_DEV_ASSIGN_ENABLE_IOMMU) { - if (!kvm->arch.intel_iommu_domain) { + if (!kvm->arch.iommu_domain) { r = kvm_iommu_map_guest(kvm); if (r) goto out_list_del; -- cgit v1.2.3 From 38e817febe2f12bd2fbf92a1df36f41946d0c223 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Tue, 2 Dec 2008 17:27:52 +0100 Subject: AMD IOMMU: rename iommu_map to iommu_map_page Impact: function rename The iommu_map function maps only one page. Make this clear in the function name. Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 2e2da717b350..b11c855af7b4 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -338,10 +338,10 @@ static void iommu_flush_tlb(struct amd_iommu *iommu, u16 domid) * supporting all features of AMD IOMMU page tables like level skipping * and full 64 bit address spaces. */ -static int iommu_map(struct protection_domain *dom, - unsigned long bus_addr, - unsigned long phys_addr, - int prot) +static int iommu_map_page(struct protection_domain *dom, + unsigned long bus_addr, + unsigned long phys_addr, + int prot) { u64 __pte, *pte, *page; @@ -440,7 +440,7 @@ static int dma_ops_unity_map(struct dma_ops_domain *dma_dom, for (addr = e->address_start; addr < e->address_end; addr += PAGE_SIZE) { - ret = iommu_map(&dma_dom->domain, addr, addr, e->prot); + ret = iommu_map_page(&dma_dom->domain, addr, addr, e->prot); if (ret) return ret; /* -- cgit v1.2.3 From 86db2e5d47bfa61a151d6ac83263f4bde4d52290 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Tue, 2 Dec 2008 18:20:21 +0100 Subject: AMD IOMMU: make dma_ops_free_pagetable generic Impact: change code to free pagetables from protection domains The dma_ops_free_pagetable function can only free pagetables from dma_ops domains. Change that to free pagetables of pure protection domains. Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index b11c855af7b4..8a0fd3d09973 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -587,12 +587,12 @@ static void dma_ops_reserve_addresses(struct dma_ops_domain *dom, iommu_area_reserve(dom->bitmap, start_page, pages); } -static void dma_ops_free_pagetable(struct dma_ops_domain *dma_dom) +static void free_pagetable(struct protection_domain *domain) { int i, j; u64 *p1, *p2, *p3; - p1 = dma_dom->domain.pt_root; + p1 = domain->pt_root; if (!p1) return; @@ -613,6 +613,8 @@ static void dma_ops_free_pagetable(struct dma_ops_domain *dma_dom) } free_page((unsigned long)p1); + + domain->pt_root = NULL; } /* @@ -624,7 +626,7 @@ static void dma_ops_domain_free(struct dma_ops_domain *dom) if (!dom) return; - dma_ops_free_pagetable(dom); + free_pagetable(&dom->domain); kfree(dom->pte_pages); -- cgit v1.2.3 From a2acfb75792511a35586db80a38b8e4701a97730 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Tue, 2 Dec 2008 18:28:53 +0100 Subject: AMD IOMMU: add domain id free function Impact: add code to release a domain id Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 8a0fd3d09973..0922d5fe633c 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -571,6 +571,18 @@ static u16 domain_id_alloc(void) return id; } +#ifdef CONFIG_IOMMU_API +static void domain_id_free(int id) +{ + unsigned long flags; + + write_lock_irqsave(&amd_iommu_devtable_lock, flags); + if (id > 0 && id < MAX_DOMAIN_ID) + __clear_bit(id, amd_iommu_pd_alloc_bitmap); + write_unlock_irqrestore(&amd_iommu_devtable_lock, flags); +} +#endif + /* * Used to reserve address ranges in the aperture (e.g. for exclusion * ranges. -- cgit v1.2.3 From 8d201968e15f56ae2837b0d0b64d3fff098857b0 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Tue, 2 Dec 2008 20:34:41 +0100 Subject: AMD IOMMU: refactor completion wait handling into separate functions Impact: split one function into three The separate functions are required synchronize commands across all hardware IOMMUs in the system. Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 68 ++++++++++++++++++++++++++++++--------------- 1 file changed, 45 insertions(+), 23 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 0922d5fe633c..2280ef86651f 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -195,6 +195,46 @@ static int iommu_queue_command(struct amd_iommu *iommu, struct iommu_cmd *cmd) return ret; } +/* + * This function waits until an IOMMU has completed a completion + * wait command + */ +static void __iommu_wait_for_completion(struct amd_iommu *iommu) +{ + int ready = 0; + unsigned status = 0; + unsigned long i = 0; + + while (!ready && (i < EXIT_LOOP_COUNT)) { + ++i; + /* wait for the bit to become one */ + status = readl(iommu->mmio_base + MMIO_STATUS_OFFSET); + ready = status & MMIO_STATUS_COM_WAIT_INT_MASK; + } + + /* set bit back to zero */ + status &= ~MMIO_STATUS_COM_WAIT_INT_MASK; + writel(status, iommu->mmio_base + MMIO_STATUS_OFFSET); + + if (unlikely(i == EXIT_LOOP_COUNT)) + panic("AMD IOMMU: Completion wait loop failed\n"); +} + +/* + * This function queues a completion wait command into the command + * buffer of an IOMMU + */ +static int __iommu_completion_wait(struct amd_iommu *iommu) +{ + struct iommu_cmd cmd; + + memset(&cmd, 0, sizeof(cmd)); + cmd.data[0] = CMD_COMPL_WAIT_INT_MASK; + CMD_SET_TYPE(&cmd, CMD_COMPL_WAIT); + + return __iommu_queue_command(iommu, &cmd); +} + /* * This function is called whenever we need to ensure that the IOMMU has * completed execution of all commands we sent. It sends a @@ -204,40 +244,22 @@ static int iommu_queue_command(struct amd_iommu *iommu, struct iommu_cmd *cmd) */ static int iommu_completion_wait(struct amd_iommu *iommu) { - int ret = 0, ready = 0; - unsigned status = 0; - struct iommu_cmd cmd; - unsigned long flags, i = 0; - - memset(&cmd, 0, sizeof(cmd)); - cmd.data[0] = CMD_COMPL_WAIT_INT_MASK; - CMD_SET_TYPE(&cmd, CMD_COMPL_WAIT); + int ret = 0; + unsigned long flags; spin_lock_irqsave(&iommu->lock, flags); if (!iommu->need_sync) goto out; - iommu->need_sync = 0; + ret = __iommu_completion_wait(iommu); - ret = __iommu_queue_command(iommu, &cmd); + iommu->need_sync = 0; if (ret) goto out; - while (!ready && (i < EXIT_LOOP_COUNT)) { - ++i; - /* wait for the bit to become one */ - status = readl(iommu->mmio_base + MMIO_STATUS_OFFSET); - ready = status & MMIO_STATUS_COM_WAIT_INT_MASK; - } - - /* set bit back to zero */ - status &= ~MMIO_STATUS_COM_WAIT_INT_MASK; - writel(status, iommu->mmio_base + MMIO_STATUS_OFFSET); - - if (unlikely(i == EXIT_LOOP_COUNT)) - panic("AMD IOMMU: Completion wait loop failed\n"); + __iommu_wait_for_completion(iommu); out: spin_unlock_irqrestore(&iommu->lock, flags); -- cgit v1.2.3 From 237b6f33291394c337ae84e2d3782d5605803af2 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Tue, 2 Dec 2008 20:54:37 +0100 Subject: AMD IOMMU: move invalidation command building to a separate function Impact: refactoring of iommu_queue_inv_iommu_pages Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 26 ++++++++++++++++---------- 1 file changed, 16 insertions(+), 10 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 2280ef86651f..fee16fbf2f33 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -286,6 +286,21 @@ static int iommu_queue_inv_dev_entry(struct amd_iommu *iommu, u16 devid) return ret; } +static void __iommu_build_inv_iommu_pages(struct iommu_cmd *cmd, u64 address, + u16 domid, int pde, int s) +{ + memset(cmd, 0, sizeof(*cmd)); + address &= PAGE_MASK; + CMD_SET_TYPE(cmd, CMD_INV_IOMMU_PAGES); + cmd->data[1] |= domid; + cmd->data[2] = lower_32_bits(address); + cmd->data[3] = upper_32_bits(address); + if (s) /* size bit - we flush more than one 4kb page */ + cmd->data[2] |= CMD_INV_IOMMU_PAGES_SIZE_MASK; + if (pde) /* PDE bit - we wan't flush everything not only the PTEs */ + cmd->data[2] |= CMD_INV_IOMMU_PAGES_PDE_MASK; +} + /* * Generic command send function for invalidaing TLB entries */ @@ -295,16 +310,7 @@ static int iommu_queue_inv_iommu_pages(struct amd_iommu *iommu, struct iommu_cmd cmd; int ret; - memset(&cmd, 0, sizeof(cmd)); - address &= PAGE_MASK; - CMD_SET_TYPE(&cmd, CMD_INV_IOMMU_PAGES); - cmd.data[1] |= domid; - cmd.data[2] = lower_32_bits(address); - cmd.data[3] = upper_32_bits(address); - if (s) /* size bit - we flush more than one 4kb page */ - cmd.data[2] |= CMD_INV_IOMMU_PAGES_SIZE_MASK; - if (pde) /* PDE bit - we wan't flush everything not only the PTEs */ - cmd.data[2] |= CMD_INV_IOMMU_PAGES_PDE_MASK; + __iommu_build_inv_iommu_pages(&cmd, address, domid, pde, s); ret = iommu_queue_command(iommu, &cmd); -- cgit v1.2.3 From 9e919012e33c481991e46aa4cb13d807cd47b798 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Wed, 10 Dec 2008 20:05:52 +0100 Subject: AMD IOMMU: don't remove protection domain from iommu_pd_list Impact: save unneeded logic to add and remove domains to the list The removal of a protection domain from the iommu_pd_list is not necessary. Another benefit is that we save complexity because we don't have to readd it later when the device no longer uses the domain. Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 1 - 1 file changed, 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index fee16fbf2f33..b7b3067630cf 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -844,7 +844,6 @@ static struct dma_ops_domain *find_protection_domain(u16 devid) list_for_each_entry(entry, &iommu_pd_list, list) { if (entry->target_dev == devid) { ret = entry; - list_del(&ret->list); break; } } -- cgit v1.2.3 From 43f4960983a309568a6c4375f081e63fb2ff24a3 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Tue, 2 Dec 2008 21:01:12 +0100 Subject: AMD IOMMU: add iommu_flush_domain function Impact: add a function to flush a domain id on every IOMMU Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index b7b3067630cf..2b6b8e050bd8 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -352,6 +352,30 @@ static void iommu_flush_tlb(struct amd_iommu *iommu, u16 domid) iommu_queue_inv_iommu_pages(iommu, address, domid, 0, 1); } +#ifdef CONFIG_IOMMU_API +/* + * This function is used to flush the IO/TLB for a given protection domain + * on every IOMMU in the system + */ +static void iommu_flush_domain(u16 domid) +{ + unsigned long flags; + struct amd_iommu *iommu; + struct iommu_cmd cmd; + + __iommu_build_inv_iommu_pages(&cmd, CMD_INV_IOMMU_ALL_PAGES_ADDRESS, + domid, 1, 1); + + list_for_each_entry(iommu, &amd_iommu_list, list) { + spin_lock_irqsave(&iommu->lock, flags); + __iommu_queue_command(iommu, &cmd); + __iommu_completion_wait(iommu); + __iommu_wait_for_completion(iommu); + spin_unlock_irqrestore(&iommu->lock, flags); + } +} +#endif + /**************************************************************************** * * The functions below are used the create the page table mappings for -- cgit v1.2.3 From 9fdb19d64c0247f23343b51fc85f438f8e7a2f3c Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Tue, 2 Dec 2008 17:46:25 +0100 Subject: AMD IOMMU: add protection domain flags Imapct: add a new struct member to 'struct protection_domain' When using protection domains for dma_ops and KVM its better to know for which subsystem it was allocated. Add a flags member to struct protection domain for that purpose. Signed-off-by: Joerg Roedel --- arch/x86/include/asm/amd_iommu_types.h | 14 +++++++++----- arch/x86/kernel/amd_iommu.c | 1 + 2 files changed, 10 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/amd_iommu_types.h b/arch/x86/include/asm/amd_iommu_types.h index ac302a2fa339..4862a5be899c 100644 --- a/arch/x86/include/asm/amd_iommu_types.h +++ b/arch/x86/include/asm/amd_iommu_types.h @@ -190,16 +190,20 @@ /* FIXME: move this macro to */ #define PCI_BUS(x) (((x) >> 8) & 0xff) +/* Protection domain flags */ +#define PD_DMA_OPS_MASK (1UL << 0) /* domain used for dma_ops */ + /* * This structure contains generic data for IOMMU protection domains * independent of their use. */ struct protection_domain { - spinlock_t lock; /* mostly used to lock the page table*/ - u16 id; /* the domain id written to the device table */ - int mode; /* paging mode (0-6 levels) */ - u64 *pt_root; /* page table root pointer */ - void *priv; /* private data */ + spinlock_t lock; /* mostly used to lock the page table*/ + u16 id; /* the domain id written to the device table */ + int mode; /* paging mode (0-6 levels) */ + u64 *pt_root; /* page table root pointer */ + unsigned long flags; /* flags to find out type of domain */ + void *priv; /* private data */ }; /* diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 2b6b8e050bd8..bb28e2cda711 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -729,6 +729,7 @@ static struct dma_ops_domain *dma_ops_domain_alloc(struct amd_iommu *iommu, goto free_dma_dom; dma_dom->domain.mode = PAGE_MODE_3_LEVEL; dma_dom->domain.pt_root = (void *)get_zeroed_page(GFP_KERNEL); + dma_dom->domain.flags = PD_DMA_OPS_MASK; dma_dom->domain.priv = dma_dom; if (!dma_dom->domain.pt_root) goto free_dma_dom; -- cgit v1.2.3 From 5b28df6f43ac9878f310ad0cb7f11ddb262a7ac6 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Tue, 2 Dec 2008 17:49:42 +0100 Subject: AMD IOMMU: add checks for dma_ops domain to dma_ops functions Impact: detect when a driver uses a device assigned otherwise Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 41 +++++++++++++++++++++++++++++++++++------ 1 file changed, 35 insertions(+), 6 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index bb28e2cda711..5c465c91150e 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -791,6 +791,15 @@ free_dma_dom: return NULL; } +/* + * little helper function to check whether a given protection domain is a + * dma_ops domain + */ +static bool dma_ops_domain(struct protection_domain *domain) +{ + return domain->flags & PD_DMA_OPS_MASK; +} + /* * Find out the protection domain structure for a given PCI device. This * will give us the pointer to the page table root for example. @@ -1096,6 +1105,9 @@ static dma_addr_t map_single(struct device *dev, phys_addr_t paddr, /* device not handled by any AMD IOMMU */ return (dma_addr_t)paddr; + if (!dma_ops_domain(domain)) + return bad_dma_address; + spin_lock_irqsave(&domain->lock, flags); addr = __map_single(dev, iommu, domain->priv, paddr, size, dir, false, dma_mask); @@ -1126,6 +1138,9 @@ static void unmap_single(struct device *dev, dma_addr_t dma_addr, /* device not handled by any AMD IOMMU */ return; + if (!dma_ops_domain(domain)) + return; + spin_lock_irqsave(&domain->lock, flags); __unmap_single(iommu, domain->priv, dma_addr, size, dir); @@ -1180,6 +1195,9 @@ static int map_sg(struct device *dev, struct scatterlist *sglist, if (!iommu || !domain) return map_sg_no_iommu(dev, sglist, nelems, dir); + if (!dma_ops_domain(domain)) + return 0; + spin_lock_irqsave(&domain->lock, flags); for_each_sg(sglist, s, nelems, i) { @@ -1233,6 +1251,9 @@ static void unmap_sg(struct device *dev, struct scatterlist *sglist, !get_device_resources(dev, &iommu, &domain, &devid)) return; + if (!dma_ops_domain(domain)) + return; + spin_lock_irqsave(&domain->lock, flags); for_each_sg(sglist, s, nelems, i) { @@ -1278,6 +1299,9 @@ static void *alloc_coherent(struct device *dev, size_t size, return virt_addr; } + if (!dma_ops_domain(domain)) + goto out_free; + if (!dma_mask) dma_mask = *dev->dma_mask; @@ -1286,18 +1310,20 @@ static void *alloc_coherent(struct device *dev, size_t size, *dma_addr = __map_single(dev, iommu, domain->priv, paddr, size, DMA_BIDIRECTIONAL, true, dma_mask); - if (*dma_addr == bad_dma_address) { - free_pages((unsigned long)virt_addr, get_order(size)); - virt_addr = NULL; - goto out; - } + if (*dma_addr == bad_dma_address) + goto out_free; iommu_completion_wait(iommu); -out: spin_unlock_irqrestore(&domain->lock, flags); return virt_addr; + +out_free: + + free_pages((unsigned long)virt_addr, get_order(size)); + + return NULL; } /* @@ -1319,6 +1345,9 @@ static void free_coherent(struct device *dev, size_t size, if (!iommu || !domain) goto free_mem; + if (!dma_ops_domain(domain)) + goto free_mem; + spin_lock_irqsave(&domain->lock, flags); __unmap_single(iommu, domain->priv, dma_addr, size, DMA_BIDIRECTIONAL); -- cgit v1.2.3 From 863c74ebd0152b21bc4b11c1447b5d1429287d37 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Tue, 2 Dec 2008 17:56:36 +0100 Subject: AMD IOMMU: add device reference counting for protection domains Impact: know how many devices are assigned to a domain Signed-off-by: Joerg Roedel --- arch/x86/include/asm/amd_iommu_types.h | 1 + arch/x86/kernel/amd_iommu.c | 3 ++- 2 files changed, 3 insertions(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/amd_iommu_types.h b/arch/x86/include/asm/amd_iommu_types.h index 4862a5be899c..1c769f4e6cdf 100644 --- a/arch/x86/include/asm/amd_iommu_types.h +++ b/arch/x86/include/asm/amd_iommu_types.h @@ -203,6 +203,7 @@ struct protection_domain { int mode; /* paging mode (0-6 levels) */ u64 *pt_root; /* page table root pointer */ unsigned long flags; /* flags to find out type of domain */ + unsigned dev_cnt; /* devices assigned to this domain */ void *priv; /* private data */ }; diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 5c465c91150e..8b45bc49d1c5 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -825,9 +825,10 @@ static void set_device_domain(struct amd_iommu *iommu, u16 devid) { unsigned long flags; - u64 pte_root = virt_to_phys(domain->pt_root); + domain->dev_cnt += 1; + pte_root |= (domain->mode & DEV_ENTRY_MODE_MASK) << DEV_ENTRY_MODE_SHIFT; pte_root |= IOMMU_PTE_IR | IOMMU_PTE_IW | IOMMU_PTE_P | IOMMU_PTE_TV; -- cgit v1.2.3 From f1179dc005ee2b0e55c3f74f3552c3e9ef852265 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Wed, 10 Dec 2008 14:39:51 +0100 Subject: AMD IOMMU: rename set_device_domain function Impact: rename set_device_domain() to attach_device() Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 8b45bc49d1c5..12e8b67e5881 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -820,9 +820,9 @@ static struct protection_domain *domain_for_device(u16 devid) * If a device is not yet associated with a domain, this function does * assigns it visible for the hardware */ -static void set_device_domain(struct amd_iommu *iommu, - struct protection_domain *domain, - u16 devid) +static void attach_device(struct amd_iommu *iommu, + struct protection_domain *domain, + u16 devid) { unsigned long flags; u64 pte_root = virt_to_phys(domain->pt_root); @@ -929,14 +929,14 @@ static int get_device_resources(struct device *dev, if (!dma_dom) dma_dom = (*iommu)->default_dom; *domain = &dma_dom->domain; - set_device_domain(*iommu, *domain, *bdf); + attach_device(*iommu, *domain, *bdf); printk(KERN_INFO "AMD IOMMU: Using protection domain %d for " "device ", (*domain)->id); print_devid(_bdf, 1); } if (domain_for_device(_bdf) == NULL) - set_device_domain(*iommu, *domain, _bdf); + attach_device(*iommu, *domain, _bdf); return 1; } -- cgit v1.2.3 From 355bf553edb7fe21ada51f62c849180bec6da877 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Mon, 8 Dec 2008 12:02:41 +0100 Subject: AMD IOMMU: add device detach helper functions Impact: add helper functions to detach a device from a domain Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 39 +++++++++++++++++++++++++++++++++++++++ 1 file changed, 39 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 12e8b67e5881..15456a3a18c8 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -844,6 +844,45 @@ static void attach_device(struct amd_iommu *iommu, iommu_queue_inv_dev_entry(iommu, devid); } +#ifdef CONFIG_IOMMU_API +/* + * Removes a device from a protection domain (unlocked) + */ +static void __detach_device(struct protection_domain *domain, u16 devid) +{ + + /* lock domain */ + spin_lock(&domain->lock); + + /* remove domain from the lookup table */ + amd_iommu_pd_table[devid] = NULL; + + /* remove entry from the device table seen by the hardware */ + amd_iommu_dev_table[devid].data[0] = IOMMU_PTE_P | IOMMU_PTE_TV; + amd_iommu_dev_table[devid].data[1] = 0; + amd_iommu_dev_table[devid].data[2] = 0; + + /* decrease reference counter */ + domain->dev_cnt -= 1; + + /* ready */ + spin_unlock(&domain->lock); +} + +/* + * Removes a device from a protection domain (with devtable_lock held) + */ +static void detach_device(struct protection_domain *domain, u16 devid) +{ + unsigned long flags; + + /* lock device table */ + write_lock_irqsave(&amd_iommu_devtable_lock, flags); + __detach_device(domain, devid); + write_unlock_irqrestore(&amd_iommu_devtable_lock, flags); +} +#endif + /***************************************************************************** * * The next functions belong to the dma_ops mapping/unmapping code. -- cgit v1.2.3 From e275a2a0fc9e2168b15f6c7814e30b7ad58b1c7c Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Wed, 10 Dec 2008 18:27:25 +0100 Subject: AMD IOMMU: add device notifier callback Impact: inform IOMMU about state change of a device in the driver core Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 62 +++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 60 insertions(+), 2 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 15456a3a18c8..140875b3f6ef 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -47,6 +47,8 @@ struct iommu_cmd { static int dma_ops_unity_map(struct dma_ops_domain *dma_dom, struct unity_map_entry *e); +static struct dma_ops_domain *find_protection_domain(u16 devid); + /* returns !0 if the IOMMU is caching non-present entries in its TLB */ static int iommu_has_npcache(struct amd_iommu *iommu) @@ -844,7 +846,6 @@ static void attach_device(struct amd_iommu *iommu, iommu_queue_inv_dev_entry(iommu, devid); } -#ifdef CONFIG_IOMMU_API /* * Removes a device from a protection domain (unlocked) */ @@ -881,7 +882,62 @@ static void detach_device(struct protection_domain *domain, u16 devid) __detach_device(domain, devid); write_unlock_irqrestore(&amd_iommu_devtable_lock, flags); } -#endif + +static int device_change_notifier(struct notifier_block *nb, + unsigned long action, void *data) +{ + struct device *dev = data; + struct pci_dev *pdev = to_pci_dev(dev); + u16 devid = calc_devid(pdev->bus->number, pdev->devfn); + struct protection_domain *domain; + struct dma_ops_domain *dma_domain; + struct amd_iommu *iommu; + + if (devid > amd_iommu_last_bdf) + goto out; + + devid = amd_iommu_alias_table[devid]; + + iommu = amd_iommu_rlookup_table[devid]; + if (iommu == NULL) + goto out; + + domain = domain_for_device(devid); + + if (domain && !dma_ops_domain(domain)) + WARN_ONCE(1, "AMD IOMMU WARNING: device %s already bound " + "to a non-dma-ops domain\n", dev_name(dev)); + + switch (action) { + case BUS_NOTIFY_BOUND_DRIVER: + if (domain) + goto out; + dma_domain = find_protection_domain(devid); + if (!dma_domain) + dma_domain = iommu->default_dom; + attach_device(iommu, &dma_domain->domain, devid); + printk(KERN_INFO "AMD IOMMU: Using protection domain %d for " + "device %s\n", dma_domain->domain.id, dev_name(dev)); + break; + case BUS_NOTIFY_UNBIND_DRIVER: + if (!domain) + goto out; + detach_device(domain, devid); + break; + default: + goto out; + } + + iommu_queue_inv_dev_entry(iommu, devid); + iommu_completion_wait(iommu); + +out: + return 0; +} + +struct notifier_block device_nb = { + .notifier_call = device_change_notifier, +}; /***************************************************************************** * @@ -1510,6 +1566,8 @@ int __init amd_iommu_init_dma_ops(void) /* Make the driver finally visible to the drivers */ dma_ops = &amd_iommu_dma_ops; + bus_register_notifier(&pci_bus_type, &device_nb); + return 0; free_domains: -- cgit v1.2.3 From 6d98cd8043c13438e4ca8a9464893f0224198a30 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Mon, 8 Dec 2008 12:05:55 +0100 Subject: AMD IOMMU: add domain cleanup helper function Impact: add a function to remove all devices from a domain Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 28 ++++++++++++++++++++++++++++ 1 file changed, 28 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 140875b3f6ef..5d3f085289eb 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -1579,3 +1579,31 @@ free_domains: return ret; } + +/***************************************************************************** + * + * The following functions belong to the exported interface of AMD IOMMU + * + * This interface allows access to lower level functions of the IOMMU + * like protection domain handling and assignement of devices to domains + * which is not possible with the dma_ops interface. + * + *****************************************************************************/ + +#ifdef CONFIG_IOMMU_API + +static void cleanup_domain(struct protection_domain *domain) +{ + unsigned long flags; + u16 devid; + + write_lock_irqsave(&amd_iommu_devtable_lock, flags); + + for (devid = 0; devid <= amd_iommu_last_bdf; ++devid) + if (amd_iommu_pd_table[devid] == domain) + __detach_device(domain, devid); + + write_unlock_irqrestore(&amd_iommu_devtable_lock, flags); +} + +#endif -- cgit v1.2.3 From c156e347d6d3c36b6843c3b168eda61b9a02c827 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Tue, 2 Dec 2008 18:13:27 +0100 Subject: AMD IOMMU: add domain init function for IOMMU API Impact: add a generic function for allocation protection domains Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 30 ++++++++++++++++++++++++++++++ 1 file changed, 30 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 5d3f085289eb..6c0bd49cee5f 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -22,6 +22,9 @@ #include #include #include +#ifdef CONFIG_IOMMU_API +#include +#endif #include #include #include @@ -1606,4 +1609,31 @@ static void cleanup_domain(struct protection_domain *domain) write_unlock_irqrestore(&amd_iommu_devtable_lock, flags); } +static int amd_iommu_domain_init(struct iommu_domain *dom) +{ + struct protection_domain *domain; + + domain = kzalloc(sizeof(*domain), GFP_KERNEL); + if (!domain) + return -ENOMEM; + + spin_lock_init(&domain->lock); + domain->mode = PAGE_MODE_3_LEVEL; + domain->id = domain_id_alloc(); + if (!domain->id) + goto out_free; + domain->pt_root = (void *)get_zeroed_page(GFP_KERNEL); + if (!domain->pt_root) + goto out_free; + + dom->priv = domain; + + return 0; + +out_free: + kfree(domain); + + return -ENOMEM; +} + #endif -- cgit v1.2.3 From 98383fc301c6546af0f3a8a1d3cb8bf218f7e940 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Tue, 2 Dec 2008 18:34:12 +0100 Subject: AMD IOMMU: add domain destroy function for IOMMU API Impact: add a generic function for releasing protection domains Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 21 +++++++++++++++++++++ 1 file changed, 21 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 6c0bd49cee5f..891d713d9c96 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -1636,4 +1636,25 @@ out_free: return -ENOMEM; } +static void amd_iommu_domain_destroy(struct iommu_domain *dom) +{ + struct protection_domain *domain = dom->priv; + + if (!domain) + return; + + if (domain->dev_cnt > 0) + cleanup_domain(domain); + + BUG_ON(domain->dev_cnt != 0); + + free_pagetable(domain); + + domain_id_free(domain->id); + + kfree(domain); + + dom->priv = NULL; +} + #endif -- cgit v1.2.3 From 684f2888847b896faafed87ce4733501d2cc283c Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Mon, 8 Dec 2008 12:07:44 +0100 Subject: AMD IOMMU: add device detach function for IOMMU API Impact: add a generic function to detach devices from protection domains Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 26 ++++++++++++++++++++++++++ 1 file changed, 26 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 891d713d9c96..ef9b309e8e09 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -1657,4 +1657,30 @@ static void amd_iommu_domain_destroy(struct iommu_domain *dom) dom->priv = NULL; } +static void amd_iommu_detach_device(struct iommu_domain *dom, + struct device *dev) +{ + struct protection_domain *domain = dom->priv; + struct amd_iommu *iommu; + struct pci_dev *pdev; + u16 devid; + + if (dev->bus != &pci_bus_type) + return; + + pdev = to_pci_dev(dev); + + devid = calc_devid(pdev->bus->number, pdev->devfn); + + if (devid > 0) + detach_device(domain, devid); + + iommu = amd_iommu_rlookup_table[devid]; + if (!iommu) + return; + + iommu_queue_inv_dev_entry(iommu, devid); + iommu_completion_wait(iommu); +} + #endif -- cgit v1.2.3 From 01106066a6900b518debe990ddaadf376d433bd6 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Tue, 2 Dec 2008 19:34:11 +0100 Subject: AMD IOMMU: add device attach function for IOMMU API Impact: add a generic function to attach devices to protection domains Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 35 +++++++++++++++++++++++++++++++++++ 1 file changed, 35 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index ef9b309e8e09..2f7c0b3a448b 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -1683,4 +1683,39 @@ static void amd_iommu_detach_device(struct iommu_domain *dom, iommu_completion_wait(iommu); } +static int amd_iommu_attach_device(struct iommu_domain *dom, + struct device *dev) +{ + struct protection_domain *domain = dom->priv; + struct protection_domain *old_domain; + struct amd_iommu *iommu; + struct pci_dev *pdev; + u16 devid; + + if (dev->bus != &pci_bus_type) + return -EINVAL; + + pdev = to_pci_dev(dev); + + devid = calc_devid(pdev->bus->number, pdev->devfn); + + if (devid >= amd_iommu_last_bdf || + devid != amd_iommu_alias_table[devid]) + return -EINVAL; + + iommu = amd_iommu_rlookup_table[devid]; + if (!iommu) + return -EINVAL; + + old_domain = domain_for_device(devid); + if (old_domain) + return -EBUSY; + + attach_device(iommu, domain, devid); + + iommu_completion_wait(iommu); + + return 0; +} + #endif -- cgit v1.2.3 From c6229ca649aa9b312d1f1de20af8d2603b14eead Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Tue, 2 Dec 2008 19:48:43 +0100 Subject: AMD IOMMU: add domain map function for IOMMU API Impact: add a generic function to map pages into protection domains Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 29 +++++++++++++++++++++++++++++ 1 file changed, 29 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 2f7c0b3a448b..1fcedbe39a99 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -1718,4 +1718,33 @@ static int amd_iommu_attach_device(struct iommu_domain *dom, return 0; } +static int amd_iommu_map_range(struct iommu_domain *dom, + unsigned long iova, phys_addr_t paddr, + size_t size, int iommu_prot) +{ + struct protection_domain *domain = dom->priv; + unsigned long i, npages = iommu_num_pages(paddr, size, PAGE_SIZE); + int prot = 0; + int ret; + + if (iommu_prot & IOMMU_READ) + prot |= IOMMU_PROT_IR; + if (iommu_prot & IOMMU_WRITE) + prot |= IOMMU_PROT_IW; + + iova &= PAGE_MASK; + paddr &= PAGE_MASK; + + for (i = 0; i < npages; ++i) { + ret = iommu_map_page(domain, iova, paddr, prot); + if (ret) + return ret; + + iova += PAGE_SIZE; + paddr += PAGE_SIZE; + } + + return 0; +} + #endif -- cgit v1.2.3 From eb74ff6cc0080c7f6270fdfffba65c4eff23d3ad Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Tue, 2 Dec 2008 19:59:10 +0100 Subject: AMD IOMMU: add domain unmap function for IOMMU API Impact: add a generic function to unmap pages into protection domains Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 41 +++++++++++++++++++++++++++++++++++++++++ 1 file changed, 41 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 1fcedbe39a99..d8a0abf423ba 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -445,6 +445,30 @@ static int iommu_map_page(struct protection_domain *dom, return 0; } +#ifdef CONFIG_IOMMU_API +static void iommu_unmap_page(struct protection_domain *dom, + unsigned long bus_addr) +{ + u64 *pte; + + pte = &dom->pt_root[IOMMU_PTE_L2_INDEX(bus_addr)]; + + if (!IOMMU_PTE_PRESENT(*pte)) + return; + + pte = IOMMU_PTE_PAGE(*pte); + pte = &pte[IOMMU_PTE_L1_INDEX(bus_addr)]; + + if (!IOMMU_PTE_PRESENT(*pte)) + return; + + pte = IOMMU_PTE_PAGE(*pte); + pte = &pte[IOMMU_PTE_L1_INDEX(bus_addr)]; + + *pte = 0; +} +#endif + /* * This function checks if a specific unity mapping entry is needed for * this specific IOMMU. @@ -1747,4 +1771,21 @@ static int amd_iommu_map_range(struct iommu_domain *dom, return 0; } +static void amd_iommu_unmap_range(struct iommu_domain *dom, + unsigned long iova, size_t size) +{ + + struct protection_domain *domain = dom->priv; + unsigned long i, npages = iommu_num_pages(iova, size, PAGE_SIZE); + + iova &= PAGE_MASK; + + for (i = 0; i < npages; ++i) { + iommu_unmap_page(domain, iova); + iova += PAGE_SIZE; + } + + iommu_flush_domain(domain->id); +} + #endif -- cgit v1.2.3 From 645c4c8d7289a718c9c828ec217f2b94e3b3e6ff Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Tue, 2 Dec 2008 20:05:50 +0100 Subject: AMD IOMMU: add domain address lookup function for IOMMU API Impact: add a generic function to lockup addresses in protection domains Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 31 +++++++++++++++++++++++++++++++ 1 file changed, 31 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index d8a0abf423ba..b599e80051fc 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -1788,4 +1788,35 @@ static void amd_iommu_unmap_range(struct iommu_domain *dom, iommu_flush_domain(domain->id); } +static phys_addr_t amd_iommu_iova_to_phys(struct iommu_domain *dom, + unsigned long iova) +{ + struct protection_domain *domain = dom->priv; + unsigned long offset = iova & ~PAGE_MASK; + phys_addr_t paddr; + u64 *pte; + + pte = &domain->pt_root[IOMMU_PTE_L2_INDEX(iova)]; + + if (!IOMMU_PTE_PRESENT(*pte)) + return 0; + + pte = IOMMU_PTE_PAGE(*pte); + pte = &pte[IOMMU_PTE_L1_INDEX(iova)]; + + if (!IOMMU_PTE_PRESENT(*pte)) + return 0; + + pte = IOMMU_PTE_PAGE(*pte); + pte = &pte[IOMMU_PTE_L0_INDEX(iova)]; + + if (!IOMMU_PTE_PRESENT(*pte)) + return 0; + + paddr = *pte & IOMMU_PAGE_MASK; + paddr |= offset; + + return paddr; +} + #endif -- cgit v1.2.3 From 26961efe0dab9ca73f8fc3b6137b814252e04972 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Wed, 3 Dec 2008 17:00:17 +0100 Subject: AMD IOMMU: register functions for the IOMMU API Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index b599e80051fc..d9b651c01186 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -41,6 +41,10 @@ static DEFINE_RWLOCK(amd_iommu_devtable_lock); static LIST_HEAD(iommu_pd_list); static DEFINE_SPINLOCK(iommu_pd_list_lock); +#ifdef CONFIG_IOMMU_API +static struct iommu_ops amd_iommu_ops; +#endif + /* * general struct to manage commands send to an IOMMU */ @@ -1593,6 +1597,10 @@ int __init amd_iommu_init_dma_ops(void) /* Make the driver finally visible to the drivers */ dma_ops = &amd_iommu_dma_ops; +#ifdef CONFIG_IOMMU_API + register_iommu(&amd_iommu_ops); +#endif + bus_register_notifier(&pci_bus_type, &device_nb); return 0; @@ -1819,4 +1827,14 @@ static phys_addr_t amd_iommu_iova_to_phys(struct iommu_domain *dom, return paddr; } +static struct iommu_ops amd_iommu_ops = { + .domain_init = amd_iommu_domain_init, + .domain_destroy = amd_iommu_domain_destroy, + .attach_dev = amd_iommu_attach_device, + .detach_dev = amd_iommu_detach_device, + .map = amd_iommu_map_range, + .unmap = amd_iommu_unmap_range, + .iova_to_phys = amd_iommu_iova_to_phys, +}; + #endif -- cgit v1.2.3 From e2dc14a2a6c9a83baaafc51f06b7e73cec2167be Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Wed, 10 Dec 2008 18:48:59 +0100 Subject: AMD IOMMU: add a domain flag for default domains Impact: adds a new protection domain flag Signed-off-by: Joerg Roedel --- arch/x86/include/asm/amd_iommu_types.h | 2 ++ arch/x86/kernel/amd_iommu.c | 1 + 2 files changed, 3 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/amd_iommu_types.h b/arch/x86/include/asm/amd_iommu_types.h index 1c769f4e6cdf..6adc7020ef97 100644 --- a/arch/x86/include/asm/amd_iommu_types.h +++ b/arch/x86/include/asm/amd_iommu_types.h @@ -192,6 +192,8 @@ /* Protection domain flags */ #define PD_DMA_OPS_MASK (1UL << 0) /* domain used for dma_ops */ +#define PD_DEFAULT_MASK (1UL << 1) /* domain is a default dma_ops + domain for an IOMMU */ /* * This structure contains generic data for IOMMU protection domains diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index d9b651c01186..f2956546423b 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -1574,6 +1574,7 @@ int __init amd_iommu_init_dma_ops(void) iommu->default_dom = dma_ops_domain_alloc(iommu, order); if (iommu->default_dom == NULL) return -ENOMEM; + iommu->default_dom->domain.flags |= PD_DEFAULT_MASK; ret = iommu_init_unity_mappings(iommu); if (ret) goto free_domains; -- cgit v1.2.3 From 1ac4cbbc5eb56de96d264d10f464ba5222815b1b Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Wed, 10 Dec 2008 19:33:26 +0100 Subject: AMD IOMMU: allocate a new protection for hotplugged devices Impact: also hotplug devices benefit from device isolation Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 17 +++++++++++++++++ 1 file changed, 17 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index f2956546423b..f2260609eadc 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -923,6 +923,8 @@ static int device_change_notifier(struct notifier_block *nb, struct protection_domain *domain; struct dma_ops_domain *dma_domain; struct amd_iommu *iommu; + int order = amd_iommu_aperture_order; + unsigned long flags; if (devid > amd_iommu_last_bdf) goto out; @@ -954,6 +956,21 @@ static int device_change_notifier(struct notifier_block *nb, if (!domain) goto out; detach_device(domain, devid); + break; + case BUS_NOTIFY_ADD_DEVICE: + /* allocate a protection domain if a device is added */ + dma_domain = find_protection_domain(devid); + if (dma_domain) + goto out; + dma_domain = dma_ops_domain_alloc(iommu, order); + if (!dma_domain) + goto out; + dma_domain->target_dev = devid; + + spin_lock_irqsave(&iommu_pd_list_lock, flags); + list_add_tail(&dma_domain->list, &iommu_pd_list); + spin_unlock_irqrestore(&iommu_pd_list_lock, flags); + break; default: goto out; -- cgit v1.2.3 From ab896722867602eb0e836717e8b857ad513800d8 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Wed, 10 Dec 2008 19:43:07 +0100 Subject: AMD IOMMU: use dev_name instead of self-build print_devid Impact: use generic dev_name instead of own function Signed-off-by: Joerg Roedel --- arch/x86/include/asm/amd_iommu_types.h | 12 ------------ arch/x86/kernel/amd_iommu.c | 3 +-- 2 files changed, 1 insertion(+), 14 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/amd_iommu_types.h b/arch/x86/include/asm/amd_iommu_types.h index 6adc7020ef97..ee8cfa00017d 100644 --- a/arch/x86/include/asm/amd_iommu_types.h +++ b/arch/x86/include/asm/amd_iommu_types.h @@ -389,18 +389,6 @@ extern int amd_iommu_isolate; */ extern bool amd_iommu_unmap_flush; -/* takes a PCI device id and prints it out in a readable form */ -static inline void print_devid(u16 devid, int nl) -{ - int bus = devid >> 8; - int dev = devid >> 3 & 0x1f; - int fn = devid & 0x07; - - printk("%02x:%02x.%x", bus, dev, fn); - if (nl) - printk("\n"); -} - /* takes bus and device/function and returns the device id * FIXME: should that be in generic PCI code? */ static inline u16 calc_devid(u8 bus, u8 devfn) diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index f2260609eadc..a53cf48d3be8 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -1074,8 +1074,7 @@ static int get_device_resources(struct device *dev, *domain = &dma_dom->domain; attach_device(*iommu, *domain, *bdf); printk(KERN_INFO "AMD IOMMU: Using protection domain %d for " - "device ", (*domain)->id); - print_devid(_bdf, 1); + "device %s\n", (*domain)->id, dev_name(dev)); } if (domain_for_device(_bdf) == NULL) -- cgit v1.2.3 From 0cfd7aa90be83a4d278810d231f9ef03f189b4f0 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Wed, 10 Dec 2008 19:58:00 +0100 Subject: AMD IOMMU: convert iommu->need_sync to bool Impact: use bool instead of int for iommu->need_sync Signed-off-by: Joerg Roedel --- arch/x86/include/asm/amd_iommu_types.h | 2 +- arch/x86/kernel/amd_iommu.c | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/amd_iommu_types.h b/arch/x86/include/asm/amd_iommu_types.h index ee8cfa00017d..c4b144ece1f8 100644 --- a/arch/x86/include/asm/amd_iommu_types.h +++ b/arch/x86/include/asm/amd_iommu_types.h @@ -302,7 +302,7 @@ struct amd_iommu { bool int_enabled; /* if one, we need to send a completion wait command */ - int need_sync; + bool need_sync; /* default dma_ops domain for that IOMMU */ struct dma_ops_domain *default_dom; diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index a53cf48d3be8..e410e97e5b0a 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -198,7 +198,7 @@ static int iommu_queue_command(struct amd_iommu *iommu, struct iommu_cmd *cmd) spin_lock_irqsave(&iommu->lock, flags); ret = __iommu_queue_command(iommu, cmd); if (!ret) - iommu->need_sync = 1; + iommu->need_sync = true; spin_unlock_irqrestore(&iommu->lock, flags); return ret; @@ -263,7 +263,7 @@ static int iommu_completion_wait(struct amd_iommu *iommu) ret = __iommu_completion_wait(iommu); - iommu->need_sync = 0; + iommu->need_sync = false; if (ret) goto out; -- cgit v1.2.3 From c226f853091577e665ebc02c064af4834d8d4f28 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Fri, 12 Dec 2008 13:53:54 +0100 Subject: AMD IOMMU: convert amd_iommu_isolate to bool Impact: cleanup Signed-off-by: Joerg Roedel --- arch/x86/include/asm/amd_iommu_types.h | 2 +- arch/x86/kernel/amd_iommu_init.c | 7 ++++--- 2 files changed, 5 insertions(+), 4 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/amd_iommu_types.h b/arch/x86/include/asm/amd_iommu_types.h index c4b144ece1f8..7abf9cf0c1fe 100644 --- a/arch/x86/include/asm/amd_iommu_types.h +++ b/arch/x86/include/asm/amd_iommu_types.h @@ -381,7 +381,7 @@ extern struct protection_domain **amd_iommu_pd_table; extern unsigned long *amd_iommu_pd_alloc_bitmap; /* will be 1 if device isolation is enabled */ -extern int amd_iommu_isolate; +extern bool amd_iommu_isolate; /* * If true, the addresses will be flushed on unmap time, not when diff --git a/arch/x86/kernel/amd_iommu_init.c b/arch/x86/kernel/amd_iommu_init.c index c625800c55ca..47e163b4431e 100644 --- a/arch/x86/kernel/amd_iommu_init.c +++ b/arch/x86/kernel/amd_iommu_init.c @@ -122,7 +122,8 @@ u16 amd_iommu_last_bdf; /* largest PCI device id we have LIST_HEAD(amd_iommu_unity_map); /* a list of required unity mappings we find in ACPI */ unsigned amd_iommu_aperture_order = 26; /* size of aperture in power of 2 */ -int amd_iommu_isolate = 1; /* if 1, device isolation is enabled */ +bool amd_iommu_isolate = true; /* if true, device isolation is + enabled */ bool amd_iommu_unmap_flush; /* if true, flush on every unmap */ LIST_HEAD(amd_iommu_list); /* list of all AMD IOMMUs in the @@ -1218,9 +1219,9 @@ static int __init parse_amd_iommu_options(char *str) { for (; *str; ++str) { if (strncmp(str, "isolate", 7) == 0) - amd_iommu_isolate = 1; + amd_iommu_isolate = true; if (strncmp(str, "share", 5) == 0) - amd_iommu_isolate = 0; + amd_iommu_isolate = false; if (strncmp(str, "fullflush", 9) == 0) amd_iommu_unmap_flush = true; } -- cgit v1.2.3 From edcb34da259c503a2ffd37e51a658672ba3bc7a2 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Wed, 10 Dec 2008 20:01:45 +0100 Subject: AMD IOMMU: use calc_devid in prealloc_protection_domains Impact: cleanup Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index e410e97e5b0a..3011ea7a3f82 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -1543,7 +1543,7 @@ void prealloc_protection_domains(void) u16 devid; while ((dev = pci_get_device(PCI_ANY_ID, PCI_ANY_ID, dev)) != NULL) { - devid = (dev->bus->number << 8) | dev->devfn; + devid = calc_devid(dev->bus->number, dev->devfn); if (devid > amd_iommu_last_bdf) continue; devid = amd_iommu_alias_table[devid]; -- cgit v1.2.3 From a4e267c88b4acfc87ff2ab0cc8e9509878e9aaba Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Wed, 10 Dec 2008 20:04:18 +0100 Subject: AMD IOMMU: use dev_name in iommu_enable function Impact: cleanup Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu_init.c | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu_init.c b/arch/x86/kernel/amd_iommu_init.c index 47e163b4431e..be81f6125c51 100644 --- a/arch/x86/kernel/amd_iommu_init.c +++ b/arch/x86/kernel/amd_iommu_init.c @@ -246,12 +246,8 @@ static void __init iommu_feature_disable(struct amd_iommu *iommu, u8 bit) /* Function to enable the hardware */ void __init iommu_enable(struct amd_iommu *iommu) { - printk(KERN_INFO "AMD IOMMU: Enabling IOMMU " - "at %02x:%02x.%x cap 0x%hx\n", - iommu->dev->bus->number, - PCI_SLOT(iommu->dev->devfn), - PCI_FUNC(iommu->dev->devfn), - iommu->cap_ptr); + printk(KERN_INFO "AMD IOMMU: Enabling IOMMU at %s cap 0x%hx\n", + dev_name(&iommu->dev->dev), iommu->cap_ptr); iommu_feature_enable(iommu, CONTROL_IOMMU_EN); } -- cgit v1.2.3 From 2e117604a4e8f3f9cee4aec3373b0382159e152a Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Thu, 11 Dec 2008 19:00:12 +0100 Subject: AMD IOMMU: add Kconfig entry for statistic collection code Impact: adds new Kconfig entry Signed-off-by: Joerg Roedel --- arch/x86/Kconfig | 10 ++++++++++ 1 file changed, 10 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 249d1e0824b5..f9998d276eb3 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -586,6 +586,16 @@ config AMD_IOMMU your BIOS for an option to enable it or if you have an IVRS ACPI table. +config AMD_IOMMU_STATS + bool "Export AMD IOMMU statistics to debugfs" + depends on AMD_IOMMU + select DEBUG_FS + help + This option enables code in the AMD IOMMU driver to collect various + statistics about whats happening in the driver and exports that + information to userspace via debugfs. + If unsure, say N. + # need this always selected by IOMMU for the VIA workaround config SWIOTLB def_bool y if X86_64 -- cgit v1.2.3 From a9dddbe0497ab0df7ee729e8d4cb0ee2dec3e4ba Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Fri, 12 Dec 2008 12:33:06 +0100 Subject: AMD IOMMU: add necessary header defines for stats counting Impact: add defines to make iommu stats collection configurable Signed-off-by: Joerg Roedel --- arch/x86/include/asm/amd_iommu_types.h | 26 ++++++++++++++++++++++++++ 1 file changed, 26 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/amd_iommu_types.h b/arch/x86/include/asm/amd_iommu_types.h index 7abf9cf0c1fe..1379c5fe86d0 100644 --- a/arch/x86/include/asm/amd_iommu_types.h +++ b/arch/x86/include/asm/amd_iommu_types.h @@ -396,4 +396,30 @@ static inline u16 calc_devid(u8 bus, u8 devfn) return (((u16)bus) << 8) | devfn; } +#ifdef CONFIG_AMD_IOMMU_STATS + +struct __iommu_counter { + char *name; + struct dentry *dent; + u64 value; +}; + +#define DECLARE_STATS_COUNTER(nm) \ + static struct __iommu_counter nm = { \ + .name = #nm, \ + } + +#define INC_STATS_COUNTER(name) name.value += 1 +#define ADD_STATS_COUNTER(name, x) name.value += (x) +#define SUB_STATS_COUNTER(name, x) name.value -= (x) + +#else /* CONFIG_AMD_IOMMU_STATS */ + +#define DECLARE_STATS_COUNTER(name) +#define INC_STATS_COUNTER(name) +#define ADD_STATS_COUNTER(name, x) +#define SUB_STATS_COUNTER(name, x) + +#endif /* CONFIG_AMD_IOMMU_STATS */ + #endif /* _ASM_X86_AMD_IOMMU_TYPES_H */ -- cgit v1.2.3 From 7f26508bbb76ce86aad1130ef6b7f1a4bb7de0c2 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Fri, 12 Dec 2008 13:50:21 +0100 Subject: AMD IOMMU: add init code for statistic collection Impact: create a new debugfs directory Signed-off-by: Joerg Roedel --- arch/x86/include/asm/amd_iommu_types.h | 2 ++ arch/x86/kernel/amd_iommu.c | 37 ++++++++++++++++++++++++++++++++++ 2 files changed, 39 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/amd_iommu_types.h b/arch/x86/include/asm/amd_iommu_types.h index 1379c5fe86d0..95c8cd9d22b5 100644 --- a/arch/x86/include/asm/amd_iommu_types.h +++ b/arch/x86/include/asm/amd_iommu_types.h @@ -420,6 +420,8 @@ struct __iommu_counter { #define ADD_STATS_COUNTER(name, x) #define SUB_STATS_COUNTER(name, x) +static inline void amd_iommu_stats_init(void) { } + #endif /* CONFIG_AMD_IOMMU_STATS */ #endif /* _ASM_X86_AMD_IOMMU_TYPES_H */ diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 3011ea7a3f82..f98f70626bc6 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -20,6 +20,7 @@ #include #include #include +#include #include #include #ifdef CONFIG_IOMMU_API @@ -57,6 +58,40 @@ static int dma_ops_unity_map(struct dma_ops_domain *dma_dom, static struct dma_ops_domain *find_protection_domain(u16 devid); +#ifdef CONFIG_AMD_IOMMU_STATS + +/* + * Initialization code for statistics collection + */ + +static struct dentry *stats_dir; +static struct dentry *de_isolate; +static struct dentry *de_fflush; + +static void amd_iommu_stats_add(struct __iommu_counter *cnt) +{ + if (stats_dir == NULL) + return; + + cnt->dent = debugfs_create_u64(cnt->name, 0444, stats_dir, + &cnt->value); +} + +static void amd_iommu_stats_init(void) +{ + stats_dir = debugfs_create_dir("amd-iommu", NULL); + if (stats_dir == NULL) + return; + + de_isolate = debugfs_create_bool("isolation", 0444, stats_dir, + (u32 *)&amd_iommu_isolate); + + de_fflush = debugfs_create_bool("fullflush", 0444, stats_dir, + (u32 *)&amd_iommu_unmap_flush); +} + +#endif + /* returns !0 if the IOMMU is caching non-present entries in its TLB */ static int iommu_has_npcache(struct amd_iommu *iommu) { @@ -1620,6 +1655,8 @@ int __init amd_iommu_init_dma_ops(void) bus_register_notifier(&pci_bus_type, &device_nb); + amd_iommu_stats_init(); + return 0; free_domains: -- cgit v1.2.3 From da49f6df726ecaaee87757e8b65a560679d32f99 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Fri, 12 Dec 2008 14:59:58 +0100 Subject: AMD IOMMU: add stats counter for completion wait events Impact: see number of completion wait events in debugfs Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index f98f70626bc6..b21435748e0d 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -64,6 +64,8 @@ static struct dma_ops_domain *find_protection_domain(u16 devid); * Initialization code for statistics collection */ +DECLARE_STATS_COUNTER(compl_wait); + static struct dentry *stats_dir; static struct dentry *de_isolate; static struct dentry *de_fflush; @@ -88,6 +90,8 @@ static void amd_iommu_stats_init(void) de_fflush = debugfs_create_bool("fullflush", 0444, stats_dir, (u32 *)&amd_iommu_unmap_flush); + + amd_iommu_stats_add(&compl_wait); } #endif @@ -249,6 +253,8 @@ static void __iommu_wait_for_completion(struct amd_iommu *iommu) unsigned status = 0; unsigned long i = 0; + INC_STATS_COUNTER(compl_wait); + while (!ready && (i < EXIT_LOOP_COUNT)) { ++i; /* wait for the bit to become one */ -- cgit v1.2.3 From 0f2a86f200bc97ae6cefc5d3ac879094b3fcde48 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Fri, 12 Dec 2008 15:05:16 +0100 Subject: AMD IOMMU: add stats counter for map_single requests Impact: see number of map_single requests in debugfs Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index b21435748e0d..ef377865eb76 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -65,6 +65,7 @@ static struct dma_ops_domain *find_protection_domain(u16 devid); */ DECLARE_STATS_COUNTER(compl_wait); +DECLARE_STATS_COUNTER(cnt_map_single); static struct dentry *stats_dir; static struct dentry *de_isolate; @@ -92,6 +93,7 @@ static void amd_iommu_stats_init(void) (u32 *)&amd_iommu_unmap_flush); amd_iommu_stats_add(&compl_wait); + amd_iommu_stats_add(&cnt_map_single); } #endif @@ -1278,6 +1280,8 @@ static dma_addr_t map_single(struct device *dev, phys_addr_t paddr, dma_addr_t addr; u64 dma_mask; + INC_STATS_COUNTER(cnt_map_single); + if (!check_device(dev)) return bad_dma_address; -- cgit v1.2.3 From 146a6917fc30616401a090f55cff2b855ee5b2ab Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Fri, 12 Dec 2008 15:07:12 +0100 Subject: AMD IOMMU: add stats counter for unmap_single requests Impact: see number of unmap_single requests in debugfs Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index ef377865eb76..71646c81421a 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -66,6 +66,7 @@ static struct dma_ops_domain *find_protection_domain(u16 devid); DECLARE_STATS_COUNTER(compl_wait); DECLARE_STATS_COUNTER(cnt_map_single); +DECLARE_STATS_COUNTER(cnt_unmap_single); static struct dentry *stats_dir; static struct dentry *de_isolate; @@ -94,6 +95,7 @@ static void amd_iommu_stats_init(void) amd_iommu_stats_add(&compl_wait); amd_iommu_stats_add(&cnt_map_single); + amd_iommu_stats_add(&cnt_unmap_single); } #endif @@ -1321,6 +1323,8 @@ static void unmap_single(struct device *dev, dma_addr_t dma_addr, struct protection_domain *domain; u16 devid; + INC_STATS_COUNTER(cnt_unmap_single); + if (!check_device(dev) || !get_device_resources(dev, &iommu, &domain, &devid)) /* device not handled by any AMD IOMMU */ -- cgit v1.2.3 From d03f067a9d0a1cc09529427af9a15e15d32ba1de Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Fri, 12 Dec 2008 15:09:48 +0100 Subject: AMD IOMMU: add stats counter for map_sg requests Impact: see number of map_sg requests in debugfs Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 71646c81421a..859ca741745b 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -67,6 +67,7 @@ static struct dma_ops_domain *find_protection_domain(u16 devid); DECLARE_STATS_COUNTER(compl_wait); DECLARE_STATS_COUNTER(cnt_map_single); DECLARE_STATS_COUNTER(cnt_unmap_single); +DECLARE_STATS_COUNTER(cnt_map_sg); static struct dentry *stats_dir; static struct dentry *de_isolate; @@ -96,6 +97,7 @@ static void amd_iommu_stats_init(void) amd_iommu_stats_add(&compl_wait); amd_iommu_stats_add(&cnt_map_single); amd_iommu_stats_add(&cnt_unmap_single); + amd_iommu_stats_add(&cnt_map_sg); } #endif @@ -1377,6 +1379,8 @@ static int map_sg(struct device *dev, struct scatterlist *sglist, int mapped_elems = 0; u64 dma_mask; + INC_STATS_COUNTER(cnt_map_sg); + if (!check_device(dev)) return 0; -- cgit v1.2.3 From 55877a6bcdf0843414eecc658550c6551f5b5e1d Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Fri, 12 Dec 2008 15:12:14 +0100 Subject: AMD IOMMU: add stats counter for unmap_sg requests Impact: see number of unmap_sg requests in debugfs Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 859ca741745b..49f2c8533e12 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -68,6 +68,7 @@ DECLARE_STATS_COUNTER(compl_wait); DECLARE_STATS_COUNTER(cnt_map_single); DECLARE_STATS_COUNTER(cnt_unmap_single); DECLARE_STATS_COUNTER(cnt_map_sg); +DECLARE_STATS_COUNTER(cnt_unmap_sg); static struct dentry *stats_dir; static struct dentry *de_isolate; @@ -98,6 +99,7 @@ static void amd_iommu_stats_init(void) amd_iommu_stats_add(&cnt_map_single); amd_iommu_stats_add(&cnt_unmap_single); amd_iommu_stats_add(&cnt_map_sg); + amd_iommu_stats_add(&cnt_unmap_sg); } #endif @@ -1443,6 +1445,8 @@ static void unmap_sg(struct device *dev, struct scatterlist *sglist, u16 devid; int i; + INC_STATS_COUNTER(cnt_unmap_sg); + if (!check_device(dev) || !get_device_resources(dev, &iommu, &domain, &devid)) return; -- cgit v1.2.3 From c8f0fb36bffa9e21d214a2910b825567d52bfc2c Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Fri, 12 Dec 2008 15:14:21 +0100 Subject: AMD IOMMU: add stats counter for alloc_coherent requests Impact: see number of alloc_coherent requests in debugfs Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 49f2c8533e12..ecc89f8857b6 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -69,6 +69,7 @@ DECLARE_STATS_COUNTER(cnt_map_single); DECLARE_STATS_COUNTER(cnt_unmap_single); DECLARE_STATS_COUNTER(cnt_map_sg); DECLARE_STATS_COUNTER(cnt_unmap_sg); +DECLARE_STATS_COUNTER(cnt_alloc_coherent); static struct dentry *stats_dir; static struct dentry *de_isolate; @@ -100,6 +101,7 @@ static void amd_iommu_stats_init(void) amd_iommu_stats_add(&cnt_unmap_single); amd_iommu_stats_add(&cnt_map_sg); amd_iommu_stats_add(&cnt_unmap_sg); + amd_iommu_stats_add(&cnt_alloc_coherent); } #endif @@ -1481,6 +1483,8 @@ static void *alloc_coherent(struct device *dev, size_t size, phys_addr_t paddr; u64 dma_mask = dev->coherent_dma_mask; + INC_STATS_COUNTER(cnt_alloc_coherent); + if (!check_device(dev)) return NULL; -- cgit v1.2.3 From 5d31ee7e08b7713596b999a42e67491bdf3665b3 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Fri, 12 Dec 2008 15:16:38 +0100 Subject: AMD IOMMU: add stats counter for free_coherent requests Impact: see number of free_coherent requests in debugfs Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index ecc89f8857b6..112412d733ab 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -70,6 +70,7 @@ DECLARE_STATS_COUNTER(cnt_unmap_single); DECLARE_STATS_COUNTER(cnt_map_sg); DECLARE_STATS_COUNTER(cnt_unmap_sg); DECLARE_STATS_COUNTER(cnt_alloc_coherent); +DECLARE_STATS_COUNTER(cnt_free_coherent); static struct dentry *stats_dir; static struct dentry *de_isolate; @@ -102,6 +103,7 @@ static void amd_iommu_stats_init(void) amd_iommu_stats_add(&cnt_map_sg); amd_iommu_stats_add(&cnt_unmap_sg); amd_iommu_stats_add(&cnt_alloc_coherent); + amd_iommu_stats_add(&cnt_free_coherent); } #endif @@ -1541,6 +1543,8 @@ static void free_coherent(struct device *dev, size_t size, struct protection_domain *domain; u16 devid; + INC_STATS_COUNTER(cnt_free_coherent); + if (!check_device(dev)) return; -- cgit v1.2.3 From c1858976f5ef05122bb671f678beaf7e1fe1dd74 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Fri, 12 Dec 2008 15:42:39 +0100 Subject: AMD IOMMU: add stats counter for cross-page request Impact: see number of requests for more than one page in debugfs Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 112412d733ab..f5455039b609 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -71,6 +71,7 @@ DECLARE_STATS_COUNTER(cnt_map_sg); DECLARE_STATS_COUNTER(cnt_unmap_sg); DECLARE_STATS_COUNTER(cnt_alloc_coherent); DECLARE_STATS_COUNTER(cnt_free_coherent); +DECLARE_STATS_COUNTER(cross_page); static struct dentry *stats_dir; static struct dentry *de_isolate; @@ -104,6 +105,7 @@ static void amd_iommu_stats_init(void) amd_iommu_stats_add(&cnt_unmap_sg); amd_iommu_stats_add(&cnt_alloc_coherent); amd_iommu_stats_add(&cnt_free_coherent); + amd_iommu_stats_add(&cross_page); } #endif @@ -1217,6 +1219,9 @@ static dma_addr_t __map_single(struct device *dev, pages = iommu_num_pages(paddr, size, PAGE_SIZE); paddr &= PAGE_MASK; + if (pages > 1) + INC_STATS_COUNTER(cross_page); + if (align) align_mask = (1UL << get_order(size)) - 1; -- cgit v1.2.3 From f57d98ae6979f7bcbf758023b4716f485385f903 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Fri, 12 Dec 2008 15:46:29 +0100 Subject: AMD IOMMU: add stats counter for single iommu domain tlb flushes Impact: see number of single iommu domain tlb flushes in debugfs Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index f5455039b609..e99022d3a394 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -72,6 +72,7 @@ DECLARE_STATS_COUNTER(cnt_unmap_sg); DECLARE_STATS_COUNTER(cnt_alloc_coherent); DECLARE_STATS_COUNTER(cnt_free_coherent); DECLARE_STATS_COUNTER(cross_page); +DECLARE_STATS_COUNTER(domain_flush_single); static struct dentry *stats_dir; static struct dentry *de_isolate; @@ -106,6 +107,7 @@ static void amd_iommu_stats_init(void) amd_iommu_stats_add(&cnt_alloc_coherent); amd_iommu_stats_add(&cnt_free_coherent); amd_iommu_stats_add(&cross_page); + amd_iommu_stats_add(&domain_flush_single); } #endif @@ -413,6 +415,8 @@ static void iommu_flush_tlb(struct amd_iommu *iommu, u16 domid) { u64 address = CMD_INV_IOMMU_ALL_PAGES_ADDRESS; + INC_STATS_COUNTER(domain_flush_single); + iommu_queue_inv_iommu_pages(iommu, address, domid, 0, 1); } -- cgit v1.2.3 From 18811f55d48e5f3ee70c4744c592f940022fa592 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Fri, 12 Dec 2008 15:48:28 +0100 Subject: AMD IOMMU: add stats counter for domain tlb flushes Impact: see number of domain tlb flushes in debugfs Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index e99022d3a394..a897c7246dca 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -73,6 +73,7 @@ DECLARE_STATS_COUNTER(cnt_alloc_coherent); DECLARE_STATS_COUNTER(cnt_free_coherent); DECLARE_STATS_COUNTER(cross_page); DECLARE_STATS_COUNTER(domain_flush_single); +DECLARE_STATS_COUNTER(domain_flush_all); static struct dentry *stats_dir; static struct dentry *de_isolate; @@ -108,6 +109,7 @@ static void amd_iommu_stats_init(void) amd_iommu_stats_add(&cnt_free_coherent); amd_iommu_stats_add(&cross_page); amd_iommu_stats_add(&domain_flush_single); + amd_iommu_stats_add(&domain_flush_all); } #endif @@ -431,6 +433,8 @@ static void iommu_flush_domain(u16 domid) struct amd_iommu *iommu; struct iommu_cmd cmd; + INC_STATS_COUNTER(domain_flush_all); + __iommu_build_inv_iommu_pages(&cmd, CMD_INV_IOMMU_ALL_PAGES_ADDRESS, domid, 1, 1); -- cgit v1.2.3 From 5774f7c5fef2526bfa58eab628fbe91dce5e07b1 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Fri, 12 Dec 2008 15:57:30 +0100 Subject: AMD IOMMU: add statistics about allocated io memory Impact: see amount of allocated io memory in debugfs Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index a897c7246dca..69f367b033a2 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -74,6 +74,7 @@ DECLARE_STATS_COUNTER(cnt_free_coherent); DECLARE_STATS_COUNTER(cross_page); DECLARE_STATS_COUNTER(domain_flush_single); DECLARE_STATS_COUNTER(domain_flush_all); +DECLARE_STATS_COUNTER(alloced_io_mem); static struct dentry *stats_dir; static struct dentry *de_isolate; @@ -110,6 +111,7 @@ static void amd_iommu_stats_init(void) amd_iommu_stats_add(&cross_page); amd_iommu_stats_add(&domain_flush_single); amd_iommu_stats_add(&domain_flush_all); + amd_iommu_stats_add(&alloced_io_mem); } #endif @@ -1246,6 +1248,8 @@ static dma_addr_t __map_single(struct device *dev, } address += offset; + ADD_STATS_COUNTER(alloced_io_mem, size); + if (unlikely(dma_dom->need_flush && !amd_iommu_unmap_flush)) { iommu_flush_tlb(iommu, dma_dom->domain.id); dma_dom->need_flush = false; @@ -1282,6 +1286,8 @@ static void __unmap_single(struct amd_iommu *iommu, start += PAGE_SIZE; } + SUB_STATS_COUNTER(alloced_io_mem, size); + dma_ops_free_addresses(dma_dom, dma_addr, pages); if (amd_iommu_unmap_flush || dma_dom->need_flush) { -- cgit v1.2.3 From 8ecaf8f19f0f0627d6ac6d69ed9472e7d307f35b Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Fri, 12 Dec 2008 16:13:04 +0100 Subject: AMD IOMMU: add statistics about total number of map requests Impact: see total number of map requests in debugfs Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 69f367b033a2..0c504b207bf9 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -75,6 +75,7 @@ DECLARE_STATS_COUNTER(cross_page); DECLARE_STATS_COUNTER(domain_flush_single); DECLARE_STATS_COUNTER(domain_flush_all); DECLARE_STATS_COUNTER(alloced_io_mem); +DECLARE_STATS_COUNTER(total_map_requests); static struct dentry *stats_dir; static struct dentry *de_isolate; @@ -112,6 +113,7 @@ static void amd_iommu_stats_init(void) amd_iommu_stats_add(&domain_flush_single); amd_iommu_stats_add(&domain_flush_all); amd_iommu_stats_add(&alloced_io_mem); + amd_iommu_stats_add(&total_map_requests); } #endif @@ -1229,6 +1231,8 @@ static dma_addr_t __map_single(struct device *dev, pages = iommu_num_pages(paddr, size, PAGE_SIZE); paddr &= PAGE_MASK; + INC_STATS_COUNTER(total_map_requests); + if (pages > 1) INC_STATS_COUNTER(cross_page); -- cgit v1.2.3 From 0e93dd883537e628b809a2120854cd591c8935f1 Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Mon, 29 Dec 2008 21:45:22 +0530 Subject: AMD IOMMU: prealloc_protection_domains should be static Impact: cleanup, reduce kernel size a bit, avoid sparse warning Fixes sparse warning: arch/x86/kernel/amd_iommu.c:1299:6: warning: symbol 'prealloc_protection_domains' was not declared. Should it be static? Signed-off-by: Jaswinder Singh Rajput Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 0c504b207bf9..881c68ffdf2e 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -1622,7 +1622,7 @@ static int amd_iommu_dma_supported(struct device *dev, u64 mask) * we don't need to preallocate the protection domains anymore. * For now we have to. */ -void prealloc_protection_domains(void) +static void prealloc_protection_domains(void) { struct pci_dev *dev = NULL; struct dma_ops_domain *dma_dom; -- cgit v1.2.3 From 065a6d68c71af2a3bdd080fa5aa353b76eede8f5 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Sat, 3 Jan 2009 14:16:35 +0100 Subject: AMD IOMMU: remove now unnecessary #ifdefs The #ifdef's are no longer necessary when the iommu-api and the amd iommu updates are merged together. Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu.c | 11 ----------- 1 file changed, 11 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c index 881c68ffdf2e..5113c080f0c4 100644 --- a/arch/x86/kernel/amd_iommu.c +++ b/arch/x86/kernel/amd_iommu.c @@ -426,7 +426,6 @@ static void iommu_flush_tlb(struct amd_iommu *iommu, u16 domid) iommu_queue_inv_iommu_pages(iommu, address, domid, 0, 1); } -#ifdef CONFIG_IOMMU_API /* * This function is used to flush the IO/TLB for a given protection domain * on every IOMMU in the system @@ -450,7 +449,6 @@ static void iommu_flush_domain(u16 domid) spin_unlock_irqrestore(&iommu->lock, flags); } } -#endif /**************************************************************************** * @@ -516,7 +514,6 @@ static int iommu_map_page(struct protection_domain *dom, return 0; } -#ifdef CONFIG_IOMMU_API static void iommu_unmap_page(struct protection_domain *dom, unsigned long bus_addr) { @@ -538,7 +535,6 @@ static void iommu_unmap_page(struct protection_domain *dom, *pte = 0; } -#endif /* * This function checks if a specific unity mapping entry is needed for @@ -723,7 +719,6 @@ static u16 domain_id_alloc(void) return id; } -#ifdef CONFIG_IOMMU_API static void domain_id_free(int id) { unsigned long flags; @@ -733,7 +728,6 @@ static void domain_id_free(int id) __clear_bit(id, amd_iommu_pd_alloc_bitmap); write_unlock_irqrestore(&amd_iommu_devtable_lock, flags); } -#endif /* * Used to reserve address ranges in the aperture (e.g. for exclusion @@ -1702,9 +1696,7 @@ int __init amd_iommu_init_dma_ops(void) /* Make the driver finally visible to the drivers */ dma_ops = &amd_iommu_dma_ops; -#ifdef CONFIG_IOMMU_API register_iommu(&amd_iommu_ops); -#endif bus_register_notifier(&pci_bus_type, &device_nb); @@ -1732,8 +1724,6 @@ free_domains: * *****************************************************************************/ -#ifdef CONFIG_IOMMU_API - static void cleanup_domain(struct protection_domain *domain) { unsigned long flags; @@ -1944,4 +1934,3 @@ static struct iommu_ops amd_iommu_ops = { .iova_to_phys = amd_iommu_iova_to_phys, }; -#endif -- cgit v1.2.3 From 730cf27246225d56ca1603b2f3c4fdbf882d4e51 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Wed, 31 Dec 2008 18:08:45 -0800 Subject: x86: enable cpus display of kernel_max and offlined cpus Impact: enables /sys/devices/system/cpu/{kernel_max,offline} user interface By setting total_cpus, the drivers/base/cpu.c will display the values of kernel_max (NR_CPUS-1) and the offlined cpu map. Signed-off-by: Mike Travis Acked-by: Rusty Russell Signed-off-by: Ingo Molnar --- arch/x86/kernel/smpboot.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'arch/x86') diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index 9e177a4077ee..f49c26bd7e2d 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -1298,6 +1298,8 @@ __init void prefill_possible_map(void) else possible = setup_possible_cpus; + total_cpus = max_t(int, possible, num_processors + disabled_cpus); + if (possible > CONFIG_NR_CPUS) { printk(KERN_WARNING "%d Processors exceeds NR_CPUS limit of %d\n", -- cgit v1.2.3 From 9628937d5b37169151c5f6bbd40919c6ac958a46 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Wed, 31 Dec 2008 18:08:46 -0800 Subject: x86: cleanup some remaining usages of NR_CPUS where s/b nr_cpu_ids Impact: Reduce future system panics due to cpumask operations using NR_CPUS Insure that code does not look at bits >= nr_cpu_ids as when cpumasks are allocated based on nr_cpu_ids, these extra bits will not be defined. Also some other minor updates: * change in to use cpu accessor function set_cpu_present() instead of directly accessing cpu_present_map w/cpu_clear() [arch/x86/kernel/reboot.c] * use cpumask_of() instead of &cpumask_of_cpu() [arch/x86/kernel/reboot.c] * optimize some cpu_mask_to_apicid_and functions. Signed-off-by: Mike Travis Acked-by: Rusty Russell Signed-off-by: Ingo Molnar --- arch/x86/include/asm/es7000/apic.h | 32 +++------------------------- arch/x86/include/asm/lguest.h | 2 +- arch/x86/include/asm/numaq/apic.h | 4 ++-- arch/x86/include/asm/summit/apic.h | 42 +++++++------------------------------ arch/x86/kernel/acpi/boot.c | 2 +- arch/x86/kernel/apic.c | 4 ++-- arch/x86/kernel/cpu/common.c | 2 +- arch/x86/kernel/cpuid.c | 2 +- arch/x86/kernel/msr.c | 2 +- arch/x86/kernel/reboot.c | 4 ++-- arch/x86/kernel/smpboot.c | 2 +- arch/x86/mach-voyager/voyager_smp.c | 7 +++---- 12 files changed, 26 insertions(+), 79 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/include/asm/es7000/apic.h b/arch/x86/include/asm/es7000/apic.h index 51ac1230294e..bc53d5ef1386 100644 --- a/arch/x86/include/asm/es7000/apic.h +++ b/arch/x86/include/asm/es7000/apic.h @@ -157,7 +157,7 @@ cpu_mask_to_apicid_cluster(const struct cpumask *cpumask) num_bits_set = cpumask_weight(cpumask); /* Return id to all */ - if (num_bits_set == NR_CPUS) + if (num_bits_set == nr_cpu_ids) return 0xFF; /* * The cpus in the mask must all be on the apic cluster. If are not @@ -190,7 +190,7 @@ static inline unsigned int cpu_mask_to_apicid(const cpumask_t *cpumask) num_bits_set = cpus_weight(*cpumask); /* Return id to all */ - if (num_bits_set == NR_CPUS) + if (num_bits_set == nr_cpu_ids) return cpu_to_logical_apicid(0); /* * The cpus in the mask must all be on the apic cluster. If are not @@ -218,9 +218,6 @@ static inline unsigned int cpu_mask_to_apicid(const cpumask_t *cpumask) static inline unsigned int cpu_mask_to_apicid_and(const struct cpumask *inmask, const struct cpumask *andmask) { - int num_bits_set; - int cpus_found = 0; - int cpu; int apicid = cpu_to_logical_apicid(0); cpumask_var_t cpumask; @@ -229,31 +226,8 @@ static inline unsigned int cpu_mask_to_apicid_and(const struct cpumask *inmask, cpumask_and(cpumask, inmask, andmask); cpumask_and(cpumask, cpumask, cpu_online_mask); + apicid = cpu_mask_to_apicid(cpumask); - num_bits_set = cpumask_weight(cpumask); - /* Return id to all */ - if (num_bits_set == NR_CPUS) - goto exit; - /* - * The cpus in the mask must all be on the apic cluster. If are not - * on the same apicid cluster return default value of TARGET_CPUS. - */ - cpu = cpumask_first(cpumask); - apicid = cpu_to_logical_apicid(cpu); - while (cpus_found < num_bits_set) { - if (cpumask_test_cpu(cpu, cpumask)) { - int new_apicid = cpu_to_logical_apicid(cpu); - if (apicid_cluster(apicid) != - apicid_cluster(new_apicid)){ - printk ("%s: Not a valid mask!\n", __func__); - return cpu_to_logical_apicid(0); - } - apicid = new_apicid; - cpus_found++; - } - cpu++; - } -exit: free_cpumask_var(cpumask); return apicid; } diff --git a/arch/x86/include/asm/lguest.h b/arch/x86/include/asm/lguest.h index d28a507cef39..1caf57628b9c 100644 --- a/arch/x86/include/asm/lguest.h +++ b/arch/x86/include/asm/lguest.h @@ -15,7 +15,7 @@ #define SHARED_SWITCHER_PAGES \ DIV_ROUND_UP(end_switcher_text - start_switcher_text, PAGE_SIZE) /* Pages for switcher itself, then two pages per cpu */ -#define TOTAL_SWITCHER_PAGES (SHARED_SWITCHER_PAGES + 2 * NR_CPUS) +#define TOTAL_SWITCHER_PAGES (SHARED_SWITCHER_PAGES + 2 * nr_cpu_ids) /* We map at -4M for ease of mapping into the guest (one PTE page). */ #define SWITCHER_ADDR 0xFFC00000 diff --git a/arch/x86/include/asm/numaq/apic.h b/arch/x86/include/asm/numaq/apic.h index c80f00d29965..bf37bc49bd8e 100644 --- a/arch/x86/include/asm/numaq/apic.h +++ b/arch/x86/include/asm/numaq/apic.h @@ -63,8 +63,8 @@ static inline physid_mask_t ioapic_phys_id_map(physid_mask_t phys_map) extern u8 cpu_2_logical_apicid[]; static inline int cpu_to_logical_apicid(int cpu) { - if (cpu >= NR_CPUS) - return BAD_APICID; + if (cpu >= nr_cpu_ids) + return BAD_APICID; return (int)cpu_2_logical_apicid[cpu]; } diff --git a/arch/x86/include/asm/summit/apic.h b/arch/x86/include/asm/summit/apic.h index 99327d1be49f..4bb5fb34f030 100644 --- a/arch/x86/include/asm/summit/apic.h +++ b/arch/x86/include/asm/summit/apic.h @@ -52,7 +52,7 @@ static inline void init_apic_ldr(void) int i; /* Create logical APIC IDs by counting CPUs already in cluster. */ - for (count = 0, i = NR_CPUS; --i >= 0; ) { + for (count = 0, i = nr_cpu_ids; --i >= 0; ) { lid = cpu_2_logical_apicid[i]; if (lid != BAD_APICID && apicid_cluster(lid) == my_cluster) ++count; @@ -97,8 +97,8 @@ static inline int apicid_to_node(int logical_apicid) static inline int cpu_to_logical_apicid(int cpu) { #ifdef CONFIG_SMP - if (cpu >= NR_CPUS) - return BAD_APICID; + if (cpu >= nr_cpu_ids) + return BAD_APICID; return (int)cpu_2_logical_apicid[cpu]; #else return logical_smp_processor_id(); @@ -107,7 +107,7 @@ static inline int cpu_to_logical_apicid(int cpu) static inline int cpu_present_to_apicid(int mps_cpu) { - if (mps_cpu < NR_CPUS) + if (mps_cpu < nr_cpu_ids) return (int)per_cpu(x86_bios_cpu_apicid, mps_cpu); else return BAD_APICID; @@ -146,7 +146,7 @@ static inline unsigned int cpu_mask_to_apicid(const cpumask_t *cpumask) num_bits_set = cpus_weight(*cpumask); /* Return id to all */ - if (num_bits_set == NR_CPUS) + if (num_bits_set >= nr_cpu_ids) return (int) 0xFF; /* * The cpus in the mask must all be on the apic cluster. If are not @@ -173,42 +173,16 @@ static inline unsigned int cpu_mask_to_apicid(const cpumask_t *cpumask) static inline unsigned int cpu_mask_to_apicid_and(const struct cpumask *inmask, const struct cpumask *andmask) { - int num_bits_set; - int cpus_found = 0; - int cpu; - int apicid = 0xFF; + int apicid = cpu_to_logical_apicid(0); cpumask_var_t cpumask; if (!alloc_cpumask_var(&cpumask, GFP_ATOMIC)) - return (int) 0xFF; + return apicid; cpumask_and(cpumask, inmask, andmask); cpumask_and(cpumask, cpumask, cpu_online_mask); + apicid = cpu_mask_to_apicid(cpumask); - num_bits_set = cpumask_weight(cpumask); - /* Return id to all */ - if (num_bits_set == nr_cpu_ids) - goto exit; - /* - * The cpus in the mask must all be on the apic cluster. If are not - * on the same apicid cluster return default value of TARGET_CPUS. - */ - cpu = cpumask_first(cpumask); - apicid = cpu_to_logical_apicid(cpu); - while (cpus_found < num_bits_set) { - if (cpumask_test_cpu(cpu, cpumask)) { - int new_apicid = cpu_to_logical_apicid(cpu); - if (apicid_cluster(apicid) != - apicid_cluster(new_apicid)){ - printk ("%s: Not a valid mask!\n", __func__); - return 0xFF; - } - apicid = apicid | new_apicid; - cpus_found++; - } - cpu++; - } -exit: free_cpumask_var(cpumask); return apicid; } diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c index 65d0b72777ea..fd24c55e4ae2 100644 --- a/arch/x86/kernel/acpi/boot.c +++ b/arch/x86/kernel/acpi/boot.c @@ -598,7 +598,7 @@ EXPORT_SYMBOL(acpi_map_lsapic); int acpi_unmap_lsapic(int cpu) { per_cpu(x86_cpu_to_apicid, cpu) = -1; - cpu_clear(cpu, cpu_present_map); + set_cpu_present(cpu, false); num_processors--; return (0); diff --git a/arch/x86/kernel/apic.c b/arch/x86/kernel/apic.c index 6b7f824db160..99589245fd8d 100644 --- a/arch/x86/kernel/apic.c +++ b/arch/x86/kernel/apic.c @@ -140,7 +140,7 @@ static int lapic_next_event(unsigned long delta, struct clock_event_device *evt); static void lapic_timer_setup(enum clock_event_mode mode, struct clock_event_device *evt); -static void lapic_timer_broadcast(const cpumask_t *mask); +static void lapic_timer_broadcast(const struct cpumask *mask); static void apic_pm_activate(void); /* @@ -453,7 +453,7 @@ static void lapic_timer_setup(enum clock_event_mode mode, /* * Local APIC timer broadcast function */ -static void lapic_timer_broadcast(const cpumask_t *mask) +static void lapic_timer_broadcast(const struct cpumask *mask) { #ifdef CONFIG_SMP send_IPI_mask(mask, LOCAL_TIMER_VECTOR); diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index 42e0853030cb..3f95a40f718a 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -355,7 +355,7 @@ void __cpuinit detect_ht(struct cpuinfo_x86 *c) printk(KERN_INFO "CPU: Hyper-Threading is disabled\n"); } else if (smp_num_siblings > 1) { - if (smp_num_siblings > NR_CPUS) { + if (smp_num_siblings > nr_cpu_ids) { printk(KERN_WARNING "CPU: Unsupported number of siblings %d", smp_num_siblings); smp_num_siblings = 1; diff --git a/arch/x86/kernel/cpuid.c b/arch/x86/kernel/cpuid.c index 72cefd1e649b..62a3c23bd703 100644 --- a/arch/x86/kernel/cpuid.c +++ b/arch/x86/kernel/cpuid.c @@ -121,7 +121,7 @@ static int cpuid_open(struct inode *inode, struct file *file) lock_kernel(); cpu = iminor(file->f_path.dentry->d_inode); - if (cpu >= NR_CPUS || !cpu_online(cpu)) { + if (cpu >= nr_cpu_ids || !cpu_online(cpu)) { ret = -ENXIO; /* No such CPU */ goto out; } diff --git a/arch/x86/kernel/msr.c b/arch/x86/kernel/msr.c index 82a7c7ed6d45..726266695b2c 100644 --- a/arch/x86/kernel/msr.c +++ b/arch/x86/kernel/msr.c @@ -136,7 +136,7 @@ static int msr_open(struct inode *inode, struct file *file) lock_kernel(); cpu = iminor(file->f_path.dentry->d_inode); - if (cpu >= NR_CPUS || !cpu_online(cpu)) { + if (cpu >= nr_cpu_ids || !cpu_online(cpu)) { ret = -ENXIO; /* No such CPU */ goto out; } diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c index ba7b9a0e6063..de4a9d643bee 100644 --- a/arch/x86/kernel/reboot.c +++ b/arch/x86/kernel/reboot.c @@ -449,7 +449,7 @@ void native_machine_shutdown(void) #ifdef CONFIG_X86_32 /* See if there has been given a command line override */ - if ((reboot_cpu != -1) && (reboot_cpu < NR_CPUS) && + if ((reboot_cpu != -1) && (reboot_cpu < nr_cpu_ids) && cpu_online(reboot_cpu)) reboot_cpu_id = reboot_cpu; #endif @@ -459,7 +459,7 @@ void native_machine_shutdown(void) reboot_cpu_id = smp_processor_id(); /* Make certain I only run on the appropriate processor */ - set_cpus_allowed_ptr(current, &cpumask_of_cpu(reboot_cpu_id)); + set_cpus_allowed_ptr(current, cpumask_of(reboot_cpu_id)); /* O.K Now that I'm on the appropriate processor, * stop all of the others. diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index f49c26bd7e2d..6bd4d9b73870 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -1154,7 +1154,7 @@ static void __init smp_cpu_index_default(void) for_each_possible_cpu(i) { c = &cpu_data(i); /* mark all to hotplug */ - c->cpu_index = NR_CPUS; + c->cpu_index = nr_cpu_ids; } } diff --git a/arch/x86/mach-voyager/voyager_smp.c b/arch/x86/mach-voyager/voyager_smp.c index a5bc05492b1e..9840b7ec749a 100644 --- a/arch/x86/mach-voyager/voyager_smp.c +++ b/arch/x86/mach-voyager/voyager_smp.c @@ -357,9 +357,8 @@ void __init find_smp_config(void) printk("VOYAGER SMP: Boot cpu is %d\n", boot_cpu_id); /* initialize the CPU structures (moved from smp_boot_cpus) */ - for (i = 0; i < NR_CPUS; i++) { + for (i = 0; i < nr_cpu_ids; i++) cpu_irq_affinity[i] = ~0; - } cpu_online_map = cpumask_of_cpu(boot_cpu_id); /* The boot CPU must be extended */ @@ -1227,7 +1226,7 @@ int setup_profiling_timer(unsigned int multiplier) * new values until the next timer interrupt in which they do process * accounting. */ - for (i = 0; i < NR_CPUS; ++i) + for (i = 0; i < nr_cpu_ids; ++i) per_cpu(prof_multiplier, i) = multiplier; return 0; @@ -1257,7 +1256,7 @@ void __init voyager_smp_intr_init(void) int i; /* initialize the per cpu irq mask to all disabled */ - for (i = 0; i < NR_CPUS; i++) + for (i = 0; i < nr_cpu_ids; i++) vic_irq_mask[i] = 0xFFFF; VIC_SET_GATE(VIC_CPI_LEVEL0, vic_cpi_interrupt); -- cgit v1.2.3 From ee943a82b697456f9d2ac46f1e6d230beedb4b6c Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Wed, 31 Dec 2008 18:08:47 -0800 Subject: x86: use cpumask_var_t in acpi/boot.c Impact: reduce stack size, use new API. Replace cpumask_t with cpumask_var_t. Signed-off-by: Rusty Russell Signed-off-by: Mike Travis Signed-off-by: Ingo Molnar --- arch/x86/kernel/acpi/boot.c | 29 ++++++++++++++++++++++------- 1 file changed, 22 insertions(+), 7 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c index fd24c55e4ae2..29dc0c89d4af 100644 --- a/arch/x86/kernel/acpi/boot.c +++ b/arch/x86/kernel/acpi/boot.c @@ -538,9 +538,10 @@ static int __cpuinit _acpi_map_lsapic(acpi_handle handle, int *pcpu) struct acpi_buffer buffer = { ACPI_ALLOCATE_BUFFER, NULL }; union acpi_object *obj; struct acpi_madt_local_apic *lapic; - cpumask_t tmp_map, new_map; + cpumask_var_t tmp_map, new_map; u8 physid; int cpu; + int retval = -ENOMEM; if (ACPI_FAILURE(acpi_evaluate_object(handle, "_MAT", NULL, &buffer))) return -EINVAL; @@ -569,23 +570,37 @@ static int __cpuinit _acpi_map_lsapic(acpi_handle handle, int *pcpu) buffer.length = ACPI_ALLOCATE_BUFFER; buffer.pointer = NULL; - tmp_map = cpu_present_map; + if (!alloc_cpumask_var(&tmp_map, GFP_KERNEL)) + goto out; + + if (!alloc_cpumask_var(&new_map, GFP_KERNEL)) + goto free_tmp_map; + + cpumask_copy(tmp_map, cpu_present_mask); acpi_register_lapic(physid, lapic->lapic_flags & ACPI_MADT_ENABLED); /* * If mp_register_lapic successfully generates a new logical cpu * number, then the following will get us exactly what was mapped */ - cpus_andnot(new_map, cpu_present_map, tmp_map); - if (cpus_empty(new_map)) { + cpumask_andnot(new_map, cpu_present_mask, tmp_map); + if (cpumask_empty(new_map)) { printk ("Unable to map lapic to logical cpu number\n"); - return -EINVAL; + retval = -EINVAL; + goto free_new_map; } - cpu = first_cpu(new_map); + cpu = cpumask_first(new_map); *pcpu = cpu; - return 0; + retval = 0; + +free_new_map: + free_cpumask_var(new_map); +free_tmp_map: + free_cpumask_var(tmp_map); +out: + return retval; } /* wrapper to silence section mismatch warning */ -- cgit v1.2.3 From 2fdf66b491ac706657946442789ec644cc317e1a Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Wed, 31 Dec 2008 18:08:47 -0800 Subject: cpumask: convert shared_cpu_map in acpi_processor* structs to cpumask_var_t Impact: Reduce memory usage, use new API. This is part of an effort to reduce structure sizes for machines configured with large NR_CPUS. cpumask_t gets replaced by cpumask_var_t, which is either struct cpumask[1] (small NR_CPUS) or struct cpumask * (large NR_CPUS). (Changes to powernow-k* by .) Signed-off-by: Rusty Russell Signed-off-by: Mike Travis Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c | 27 ++++++++-- arch/x86/kernel/cpu/cpufreq/powernow-k7.c | 9 ++++ arch/x86/kernel/cpu/cpufreq/powernow-k8.c | 24 +++++---- drivers/acpi/processor_core.c | 14 ++++-- drivers/acpi/processor_perflib.c | 28 ++++++----- drivers/acpi/processor_throttling.c | 80 +++++++++++++++++++----------- include/acpi/processor.h | 4 +- 7 files changed, 128 insertions(+), 58 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c index 88ea02dcb622..d0a001093b2d 100644 --- a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c +++ b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c @@ -517,6 +517,17 @@ acpi_cpufreq_guess_freq(struct acpi_cpufreq_data *data, unsigned int cpu) } } +static void free_acpi_perf_data(void) +{ + unsigned int i; + + /* Freeing a NULL pointer is OK, and alloc_percpu zeroes. */ + for_each_possible_cpu(i) + free_cpumask_var(per_cpu_ptr(acpi_perf_data, i) + ->shared_cpu_map); + free_percpu(acpi_perf_data); +} + /* * acpi_cpufreq_early_init - initialize ACPI P-States library * @@ -527,6 +538,7 @@ acpi_cpufreq_guess_freq(struct acpi_cpufreq_data *data, unsigned int cpu) */ static int __init acpi_cpufreq_early_init(void) { + unsigned int i; dprintk("acpi_cpufreq_early_init\n"); acpi_perf_data = alloc_percpu(struct acpi_processor_performance); @@ -534,6 +546,15 @@ static int __init acpi_cpufreq_early_init(void) dprintk("Memory allocation error for acpi_perf_data.\n"); return -ENOMEM; } + for_each_possible_cpu(i) { + if (!alloc_cpumask_var(&per_cpu_ptr(acpi_perf_data, i) + ->shared_cpu_map, GFP_KERNEL)) { + + /* Freeing a NULL pointer is OK: alloc_percpu zeroes. */ + free_acpi_perf_data(); + return -ENOMEM; + } + } /* Do initialization in ACPI core */ acpi_processor_preregister_performance(acpi_perf_data); @@ -604,9 +625,9 @@ static int acpi_cpufreq_cpu_init(struct cpufreq_policy *policy) */ if (policy->shared_type == CPUFREQ_SHARED_TYPE_ALL || policy->shared_type == CPUFREQ_SHARED_TYPE_ANY) { - policy->cpus = perf->shared_cpu_map; + cpumask_copy(&policy->cpus, perf->shared_cpu_map); } - policy->related_cpus = perf->shared_cpu_map; + cpumask_copy(&policy->related_cpus, perf->shared_cpu_map); #ifdef CONFIG_SMP dmi_check_system(sw_any_bug_dmi_table); @@ -795,7 +816,7 @@ static int __init acpi_cpufreq_init(void) ret = cpufreq_register_driver(&acpi_cpufreq_driver); if (ret) - free_percpu(acpi_perf_data); + free_acpi_perf_data(); return ret; } diff --git a/arch/x86/kernel/cpu/cpufreq/powernow-k7.c b/arch/x86/kernel/cpu/cpufreq/powernow-k7.c index 7c7d56b43136..1b446d79a8fd 100644 --- a/arch/x86/kernel/cpu/cpufreq/powernow-k7.c +++ b/arch/x86/kernel/cpu/cpufreq/powernow-k7.c @@ -310,6 +310,12 @@ static int powernow_acpi_init(void) goto err0; } + if (!alloc_cpumask_var(&acpi_processor_perf->shared_cpu_map, + GFP_KERNEL)) { + retval = -ENOMEM; + goto err05; + } + if (acpi_processor_register_performance(acpi_processor_perf, 0)) { retval = -EIO; goto err1; @@ -412,6 +418,8 @@ static int powernow_acpi_init(void) err2: acpi_processor_unregister_performance(acpi_processor_perf, 0); err1: + free_cpumask_var(acpi_processor_perf->shared_cpu_map); +err05: kfree(acpi_processor_perf); err0: printk(KERN_WARNING PFX "ACPI perflib can not be used in this platform\n"); @@ -652,6 +660,7 @@ static int powernow_cpu_exit (struct cpufreq_policy *policy) { #ifdef CONFIG_X86_POWERNOW_K7_ACPI if (acpi_processor_perf) { acpi_processor_unregister_performance(acpi_processor_perf, 0); + free_cpumask_var(acpi_processor_perf->shared_cpu_map); kfree(acpi_processor_perf); } #endif diff --git a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c index 7f05f44b97e9..c3c9adbaa26f 100644 --- a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c +++ b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c @@ -766,7 +766,7 @@ static void powernow_k8_acpi_pst_values(struct powernow_k8_data *data, unsigned static int powernow_k8_cpu_init_acpi(struct powernow_k8_data *data) { struct cpufreq_frequency_table *powernow_table; - int ret_val; + int ret_val = -ENODEV; if (acpi_processor_register_performance(&data->acpi_data, data->cpu)) { dprintk("register performance failed: bad ACPI data\n"); @@ -815,6 +815,13 @@ static int powernow_k8_cpu_init_acpi(struct powernow_k8_data *data) /* notify BIOS that we exist */ acpi_processor_notify_smm(THIS_MODULE); + if (!alloc_cpumask_var(&data->acpi_data.shared_cpu_map, GFP_KERNEL)) { + printk(KERN_ERR PFX + "unable to alloc powernow_k8_data cpumask\n"); + ret_val = -ENOMEM; + goto err_out_mem; + } + return 0; err_out_mem: @@ -826,7 +833,7 @@ err_out: /* data->acpi_data.state_count informs us at ->exit() whether ACPI was used */ data->acpi_data.state_count = 0; - return -ENODEV; + return ret_val; } static int fill_powernow_table_pstate(struct powernow_k8_data *data, struct cpufreq_frequency_table *powernow_table) @@ -929,6 +936,7 @@ static void powernow_k8_cpu_exit_acpi(struct powernow_k8_data *data) { if (data->acpi_data.state_count) acpi_processor_unregister_performance(&data->acpi_data, data->cpu); + free_cpumask_var(data->acpi_data.shared_cpu_map); } #else @@ -1134,7 +1142,8 @@ static int __cpuinit powernowk8_cpu_init(struct cpufreq_policy *pol) data->cpu = pol->cpu; data->currpstate = HW_PSTATE_INVALID; - if (powernow_k8_cpu_init_acpi(data)) { + rc = powernow_k8_cpu_init_acpi(data); + if (rc) { /* * Use the PSB BIOS structure. This is only availabe on * an UP version, and is deprecated by AMD. @@ -1152,20 +1161,17 @@ static int __cpuinit powernowk8_cpu_init(struct cpufreq_policy *pol) "ACPI maintainers and complain to your BIOS " "vendor.\n"); #endif - kfree(data); - return -ENODEV; + goto err_out; } if (pol->cpu != 0) { printk(KERN_ERR FW_BUG PFX "No ACPI _PSS objects for " "CPU other than CPU0. Complain to your BIOS " "vendor.\n"); - kfree(data); - return -ENODEV; + goto err_out; } rc = find_psb_table(data); if (rc) { - kfree(data); - return -ENODEV; + goto err_out; } } diff --git a/drivers/acpi/processor_core.c b/drivers/acpi/processor_core.c index 34948362f41d..0cc2fd31e376 100644 --- a/drivers/acpi/processor_core.c +++ b/drivers/acpi/processor_core.c @@ -826,6 +826,11 @@ static int acpi_processor_add(struct acpi_device *device) if (!pr) return -ENOMEM; + if (!alloc_cpumask_var(&pr->throttling.shared_cpu_map, GFP_KERNEL)) { + kfree(pr); + return -ENOMEM; + } + pr->handle = device->handle; strcpy(acpi_device_name(device), ACPI_PROCESSOR_DEVICE_NAME); strcpy(acpi_device_class(device), ACPI_PROCESSOR_CLASS); @@ -845,10 +850,8 @@ static int acpi_processor_remove(struct acpi_device *device, int type) pr = acpi_driver_data(device); - if (pr->id >= nr_cpu_ids) { - kfree(pr); - return 0; - } + if (pr->id >= nr_cpu_ids) + goto free; if (type == ACPI_BUS_REMOVAL_EJECT) { if (acpi_processor_handle_eject(pr)) @@ -873,6 +876,9 @@ static int acpi_processor_remove(struct acpi_device *device, int type) per_cpu(processors, pr->id) = NULL; per_cpu(processor_device_array, pr->id) = NULL; + +free: + free_cpumask_var(pr->throttling.shared_cpu_map); kfree(pr); return 0; diff --git a/drivers/acpi/processor_perflib.c b/drivers/acpi/processor_perflib.c index 0d7b772bef50..846e227592d4 100644 --- a/drivers/acpi/processor_perflib.c +++ b/drivers/acpi/processor_perflib.c @@ -588,12 +588,15 @@ int acpi_processor_preregister_performance( int count, count_target; int retval = 0; unsigned int i, j; - cpumask_t covered_cpus; + cpumask_var_t covered_cpus; struct acpi_processor *pr; struct acpi_psd_package *pdomain; struct acpi_processor *match_pr; struct acpi_psd_package *match_pdomain; + if (!alloc_cpumask_var(&covered_cpus, GFP_KERNEL)) + return -ENOMEM; + mutex_lock(&performance_mutex); retval = 0; @@ -617,7 +620,7 @@ int acpi_processor_preregister_performance( } pr->performance = percpu_ptr(performance, i); - cpu_set(i, pr->performance->shared_cpu_map); + cpumask_set_cpu(i, pr->performance->shared_cpu_map); if (acpi_processor_get_psd(pr)) { retval = -EINVAL; continue; @@ -650,18 +653,18 @@ int acpi_processor_preregister_performance( } } - cpus_clear(covered_cpus); + cpumask_clear(covered_cpus); for_each_possible_cpu(i) { pr = per_cpu(processors, i); if (!pr) continue; - if (cpu_isset(i, covered_cpus)) + if (cpumask_test_cpu(i, covered_cpus)) continue; pdomain = &(pr->performance->domain_info); - cpu_set(i, pr->performance->shared_cpu_map); - cpu_set(i, covered_cpus); + cpumask_set_cpu(i, pr->performance->shared_cpu_map); + cpumask_set_cpu(i, covered_cpus); if (pdomain->num_processors <= 1) continue; @@ -699,8 +702,8 @@ int acpi_processor_preregister_performance( goto err_ret; } - cpu_set(j, covered_cpus); - cpu_set(j, pr->performance->shared_cpu_map); + cpumask_set_cpu(j, covered_cpus); + cpumask_set_cpu(j, pr->performance->shared_cpu_map); count++; } @@ -718,8 +721,8 @@ int acpi_processor_preregister_performance( match_pr->performance->shared_type = pr->performance->shared_type; - match_pr->performance->shared_cpu_map = - pr->performance->shared_cpu_map; + cpumask_copy(match_pr->performance->shared_cpu_map, + pr->performance->shared_cpu_map); } } @@ -731,14 +734,15 @@ err_ret: /* Assume no coordination on any error parsing domain info */ if (retval) { - cpus_clear(pr->performance->shared_cpu_map); - cpu_set(i, pr->performance->shared_cpu_map); + cpumask_clear(pr->performance->shared_cpu_map); + cpumask_set_cpu(i, pr->performance->shared_cpu_map); pr->performance->shared_type = CPUFREQ_SHARED_TYPE_ALL; } pr->performance = NULL; /* Will be set for real in register */ } mutex_unlock(&performance_mutex); + free_cpumask_var(covered_cpus); return retval; } EXPORT_SYMBOL(acpi_processor_preregister_performance); diff --git a/drivers/acpi/processor_throttling.c b/drivers/acpi/processor_throttling.c index a0c38c94a8a0..d27838171f4a 100644 --- a/drivers/acpi/processor_throttling.c +++ b/drivers/acpi/processor_throttling.c @@ -61,11 +61,14 @@ static int acpi_processor_update_tsd_coord(void) int count, count_target; int retval = 0; unsigned int i, j; - cpumask_t covered_cpus; + cpumask_var_t covered_cpus; struct acpi_processor *pr, *match_pr; struct acpi_tsd_package *pdomain, *match_pdomain; struct acpi_processor_throttling *pthrottling, *match_pthrottling; + if (!alloc_cpumask_var(&covered_cpus, GFP_KERNEL)) + return -ENOMEM; + /* * Now that we have _TSD data from all CPUs, lets setup T-state * coordination between all CPUs. @@ -91,19 +94,19 @@ static int acpi_processor_update_tsd_coord(void) if (retval) goto err_ret; - cpus_clear(covered_cpus); + cpumask_clear(covered_cpus); for_each_possible_cpu(i) { pr = per_cpu(processors, i); if (!pr) continue; - if (cpu_isset(i, covered_cpus)) + if (cpumask_test_cpu(i, covered_cpus)) continue; pthrottling = &pr->throttling; pdomain = &(pthrottling->domain_info); - cpu_set(i, pthrottling->shared_cpu_map); - cpu_set(i, covered_cpus); + cpumask_set_cpu(i, pthrottling->shared_cpu_map); + cpumask_set_cpu(i, covered_cpus); /* * If the number of processor in the TSD domain is 1, it is * unnecessary to parse the coordination for this CPU. @@ -144,8 +147,8 @@ static int acpi_processor_update_tsd_coord(void) goto err_ret; } - cpu_set(j, covered_cpus); - cpu_set(j, pthrottling->shared_cpu_map); + cpumask_set_cpu(j, covered_cpus); + cpumask_set_cpu(j, pthrottling->shared_cpu_map); count++; } for_each_possible_cpu(j) { @@ -165,12 +168,14 @@ static int acpi_processor_update_tsd_coord(void) * If some CPUS have the same domain, they * will have the same shared_cpu_map. */ - match_pthrottling->shared_cpu_map = - pthrottling->shared_cpu_map; + cpumask_copy(match_pthrottling->shared_cpu_map, + pthrottling->shared_cpu_map); } } err_ret: + free_cpumask_var(covered_cpus); + for_each_possible_cpu(i) { pr = per_cpu(processors, i); if (!pr) @@ -182,8 +187,8 @@ err_ret: */ if (retval) { pthrottling = &(pr->throttling); - cpus_clear(pthrottling->shared_cpu_map); - cpu_set(i, pthrottling->shared_cpu_map); + cpumask_clear(pthrottling->shared_cpu_map); + cpumask_set_cpu(i, pthrottling->shared_cpu_map); pthrottling->shared_type = DOMAIN_COORD_TYPE_SW_ALL; } } @@ -567,7 +572,7 @@ static int acpi_processor_get_tsd(struct acpi_processor *pr) pthrottling = &pr->throttling; pthrottling->tsd_valid_flag = 1; pthrottling->shared_type = pdomain->coord_type; - cpu_set(pr->id, pthrottling->shared_cpu_map); + cpumask_set_cpu(pr->id, pthrottling->shared_cpu_map); /* * If the coordination type is not defined in ACPI spec, * the tsd_valid_flag will be clear and coordination type @@ -826,7 +831,7 @@ static int acpi_processor_get_throttling_ptc(struct acpi_processor *pr) static int acpi_processor_get_throttling(struct acpi_processor *pr) { - cpumask_t saved_mask; + cpumask_var_t saved_mask; int ret; if (!pr) @@ -834,14 +839,20 @@ static int acpi_processor_get_throttling(struct acpi_processor *pr) if (!pr->flags.throttling) return -ENODEV; + + if (!alloc_cpumask_var(&saved_mask, GFP_KERNEL)) + return -ENOMEM; + /* * Migrate task to the cpu pointed by pr. */ - saved_mask = current->cpus_allowed; - set_cpus_allowed_ptr(current, &cpumask_of_cpu(pr->id)); + cpumask_copy(saved_mask, ¤t->cpus_allowed); + /* FIXME: use work_on_cpu() */ + set_cpus_allowed_ptr(current, cpumask_of(pr->id)); ret = pr->throttling.acpi_processor_get_throttling(pr); /* restore the previous state */ - set_cpus_allowed_ptr(current, &saved_mask); + set_cpus_allowed_ptr(current, saved_mask); + free_cpumask_var(saved_mask); return ret; } @@ -986,13 +997,13 @@ static int acpi_processor_set_throttling_ptc(struct acpi_processor *pr, int acpi_processor_set_throttling(struct acpi_processor *pr, int state) { - cpumask_t saved_mask; + cpumask_var_t saved_mask; int ret = 0; unsigned int i; struct acpi_processor *match_pr; struct acpi_processor_throttling *p_throttling; struct throttling_tstate t_state; - cpumask_t online_throttling_cpus; + cpumask_var_t online_throttling_cpus; if (!pr) return -EINVAL; @@ -1003,17 +1014,25 @@ int acpi_processor_set_throttling(struct acpi_processor *pr, int state) if ((state < 0) || (state > (pr->throttling.state_count - 1))) return -EINVAL; - saved_mask = current->cpus_allowed; + if (!alloc_cpumask_var(&saved_mask, GFP_KERNEL)) + return -ENOMEM; + + if (!alloc_cpumask_var(&online_throttling_cpus, GFP_KERNEL)) { + free_cpumask_var(saved_mask); + return -ENOMEM; + } + + cpumask_copy(saved_mask, ¤t->cpus_allowed); t_state.target_state = state; p_throttling = &(pr->throttling); - cpus_and(online_throttling_cpus, cpu_online_map, - p_throttling->shared_cpu_map); + cpumask_and(online_throttling_cpus, cpu_online_mask, + p_throttling->shared_cpu_map); /* * The throttling notifier will be called for every * affected cpu in order to get one proper T-state. * The notifier event is THROTTLING_PRECHANGE. */ - for_each_cpu_mask_nr(i, online_throttling_cpus) { + for_each_cpu(i, online_throttling_cpus) { t_state.cpu = i; acpi_processor_throttling_notifier(THROTTLING_PRECHANGE, &t_state); @@ -1025,7 +1044,8 @@ int acpi_processor_set_throttling(struct acpi_processor *pr, int state) * it can be called only for the cpu pointed by pr. */ if (p_throttling->shared_type == DOMAIN_COORD_TYPE_SW_ANY) { - set_cpus_allowed_ptr(current, &cpumask_of_cpu(pr->id)); + /* FIXME: use work_on_cpu() */ + set_cpus_allowed_ptr(current, cpumask_of(pr->id)); ret = p_throttling->acpi_processor_set_throttling(pr, t_state.target_state); } else { @@ -1034,7 +1054,7 @@ int acpi_processor_set_throttling(struct acpi_processor *pr, int state) * it is necessary to set T-state for every affected * cpus. */ - for_each_cpu_mask_nr(i, online_throttling_cpus) { + for_each_cpu(i, online_throttling_cpus) { match_pr = per_cpu(processors, i); /* * If the pointer is invalid, we will report the @@ -1056,7 +1076,8 @@ int acpi_processor_set_throttling(struct acpi_processor *pr, int state) continue; } t_state.cpu = i; - set_cpus_allowed_ptr(current, &cpumask_of_cpu(i)); + /* FIXME: use work_on_cpu() */ + set_cpus_allowed_ptr(current, cpumask_of(i)); ret = match_pr->throttling. acpi_processor_set_throttling( match_pr, t_state.target_state); @@ -1068,13 +1089,16 @@ int acpi_processor_set_throttling(struct acpi_processor *pr, int state) * affected cpu to update the T-states. * The notifier event is THROTTLING_POSTCHANGE */ - for_each_cpu_mask_nr(i, online_throttling_cpus) { + for_each_cpu(i, online_throttling_cpus) { t_state.cpu = i; acpi_processor_throttling_notifier(THROTTLING_POSTCHANGE, &t_state); } /* restore the previous state */ - set_cpus_allowed_ptr(current, &saved_mask); + /* FIXME: use work_on_cpu() */ + set_cpus_allowed_ptr(current, saved_mask); + free_cpumask_var(online_throttling_cpus); + free_cpumask_var(saved_mask); return ret; } @@ -1120,7 +1144,7 @@ int acpi_processor_get_throttling_info(struct acpi_processor *pr) if (acpi_processor_get_tsd(pr)) { pthrottling = &pr->throttling; pthrottling->tsd_valid_flag = 0; - cpu_set(pr->id, pthrottling->shared_cpu_map); + cpumask_set_cpu(pr->id, pthrottling->shared_cpu_map); pthrottling->shared_type = DOMAIN_COORD_TYPE_SW_ALL; } diff --git a/include/acpi/processor.h b/include/acpi/processor.h index 3795590e152a..0574add2a1e3 100644 --- a/include/acpi/processor.h +++ b/include/acpi/processor.h @@ -127,7 +127,7 @@ struct acpi_processor_performance { unsigned int state_count; struct acpi_processor_px *states; struct acpi_psd_package domain_info; - cpumask_t shared_cpu_map; + cpumask_var_t shared_cpu_map; unsigned int shared_type; }; @@ -172,7 +172,7 @@ struct acpi_processor_throttling { unsigned int state_count; struct acpi_processor_tx_tss *states_tss; struct acpi_tsd_package domain_info; - cpumask_t shared_cpu_map; + cpumask_var_t shared_cpu_map; int (*acpi_processor_get_throttling) (struct acpi_processor * pr); int (*acpi_processor_set_throttling) (struct acpi_processor * pr, int state); -- cgit v1.2.3 From 80855f7361eb68205e6bc1981928629d9b02d5c9 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Wed, 31 Dec 2008 18:08:47 -0800 Subject: cpumask: use alloc_cpumask_var_node where appropriate Impact: Reduce inter-node memory traffic. Reduces inter-node memory traffic (offloading the global system bus) by allocating referenced struct cpumasks on the same node as the referring struct. Signed-off-by: Mike Travis Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c | 5 +++-- arch/x86/kernel/io_apic.c | 6 +++--- 2 files changed, 6 insertions(+), 5 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c index d0a001093b2d..28102ad1a363 100644 --- a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c +++ b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c @@ -547,8 +547,9 @@ static int __init acpi_cpufreq_early_init(void) return -ENOMEM; } for_each_possible_cpu(i) { - if (!alloc_cpumask_var(&per_cpu_ptr(acpi_perf_data, i) - ->shared_cpu_map, GFP_KERNEL)) { + if (!alloc_cpumask_var_node( + &per_cpu_ptr(acpi_perf_data, i)->shared_cpu_map, + GFP_KERNEL, cpu_to_node(i))) { /* Freeing a NULL pointer is OK: alloc_percpu zeroes. */ free_acpi_perf_data(); diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index 3e070bb961d7..a25c3f76b8ac 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -212,11 +212,11 @@ static struct irq_cfg *get_one_free_irq_cfg(int cpu) cfg = kzalloc_node(sizeof(*cfg), GFP_ATOMIC, node); if (cfg) { - /* FIXME: needs alloc_cpumask_var_node() */ - if (!alloc_cpumask_var(&cfg->domain, GFP_ATOMIC)) { + if (!alloc_cpumask_var_node(&cfg->domain, GFP_ATOMIC, node)) { kfree(cfg); cfg = NULL; - } else if (!alloc_cpumask_var(&cfg->old_domain, GFP_ATOMIC)) { + } else if (!alloc_cpumask_var_node(&cfg->old_domain, + GFP_ATOMIC, node)) { free_cpumask_var(cfg->domain); kfree(cfg); cfg = NULL; -- cgit v1.2.3 From ab14398abd195af91a744c320a52a1bce814dd1e Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Fri, 2 Jan 2009 21:51:32 +0300 Subject: x86: setup_per_cpu_areas() cleanup Impact: cleanup __alloc_bootmem and __alloc_bootmem_node do panic for us in case of fail so no need for additional checks here. Also lets use pr_*() macros for printing. Signed-off-by: Cyrill Gorcunov Signed-off-by: Ingo Molnar --- arch/x86/kernel/setup_percpu.c | 25 ++++++++----------------- 1 file changed, 8 insertions(+), 17 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c index 49f3f709ee1f..a4b619c33106 100644 --- a/arch/x86/kernel/setup_percpu.c +++ b/arch/x86/kernel/setup_percpu.c @@ -153,12 +153,10 @@ void __init setup_per_cpu_areas(void) align = max_t(unsigned long, PAGE_SIZE, align); size = roundup(old_size, align); - printk(KERN_INFO - "NR_CPUS:%d nr_cpumask_bits:%d nr_cpu_ids:%d nr_node_ids:%d\n", + pr_info("NR_CPUS:%d nr_cpumask_bits:%d nr_cpu_ids:%d nr_node_ids:%d\n", NR_CPUS, nr_cpumask_bits, nr_cpu_ids, nr_node_ids); - printk(KERN_INFO "PERCPU: Allocating %zd bytes of per cpu data\n", - size); + pr_info("PERCPU: Allocating %zd bytes of per cpu data\n", size); for_each_possible_cpu(cpu) { #ifndef CONFIG_NEED_MULTIPLE_NODES @@ -169,22 +167,15 @@ void __init setup_per_cpu_areas(void) if (!node_online(node) || !NODE_DATA(node)) { ptr = __alloc_bootmem(size, align, __pa(MAX_DMA_ADDRESS)); - printk(KERN_INFO - "cpu %d has no node %d or node-local memory\n", + pr_info("cpu %d has no node %d or node-local memory\n", cpu, node); - if (ptr) - printk(KERN_DEBUG - "per cpu data for cpu%d at %016lx\n", - cpu, __pa(ptr)); - } - else { + pr_debug("per cpu data for cpu%d at %016lx\n", + cpu, __pa(ptr)); + } else { ptr = __alloc_bootmem_node(NODE_DATA(node), size, align, __pa(MAX_DMA_ADDRESS)); - if (ptr) - printk(KERN_DEBUG - "per cpu data for cpu%d on node%d " - "at %016lx\n", - cpu, node, __pa(ptr)); + pr_debug("per cpu data for cpu%d on node%d at %016lx\n", + cpu, node, __pa(ptr)); } #endif per_cpu_offset(cpu) = ptr - __per_cpu_start; -- cgit v1.2.3