From 308a878210cde6ab19df9f392c24db53ad6f56bf Mon Sep 17 00:00:00 2001 From: Russ Anderson Date: Tue, 18 Apr 2006 11:26:34 -0500 Subject: [IA64] Remove unused variable in sn_sal.h cnodeid was being set but not used. The dead code was left over from a previous version that grabbed a per node lock. Signed-off-by: Russ Anderson (rja@sgi.com) Signed-off-by: Tony Luck --- include/asm-ia64/sn/sn_sal.h | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'include') diff --git a/include/asm-ia64/sn/sn_sal.h b/include/asm-ia64/sn/sn_sal.h index bf4cc867a698..7ddce80e1918 100644 --- a/include/asm-ia64/sn/sn_sal.h +++ b/include/asm-ia64/sn/sn_sal.h @@ -8,7 +8,7 @@ * License. See the file "COPYING" in the main directory of this archive * for more details. * - * Copyright (c) 2000-2005 Silicon Graphics, Inc. All rights reserved. + * Copyright (c) 2000-2006 Silicon Graphics, Inc. All rights reserved. */ @@ -705,10 +705,8 @@ static inline int sn_change_memprotect(u64 paddr, u64 len, u64 perms, u64 *nasid_array) { struct ia64_sal_retval ret_stuff; - int cnodeid; unsigned long irq_flags; - cnodeid = nasid_to_cnodeid(get_node_number(paddr)); local_irq_save(irq_flags); ia64_sal_oemcall_nolock(&ret_stuff, SN_SAL_MEMPROTECT, paddr, len, (u64)nasid_array, perms, 0, 0, 0); -- cgit v1.2.3 From 0d9adec525b87d8ab7e64efeabffb5b3f293056e Mon Sep 17 00:00:00 2001 From: Jack Steiner Date: Tue, 18 Apr 2006 15:00:45 -0500 Subject: [IA64] - Fix MAX_PXM_DOMAINS for systems with > 256 nodes Correctly size the PXM-related arrays for systems that have more than 256 nodes. Signed-off-by: Jack Steiner Signed-off-by: Tony Luck --- include/asm-ia64/acpi.h | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'include') diff --git a/include/asm-ia64/acpi.h b/include/asm-ia64/acpi.h index d734585a23cf..09a5dd0e44a8 100644 --- a/include/asm-ia64/acpi.h +++ b/include/asm-ia64/acpi.h @@ -110,9 +110,8 @@ extern void prefill_possible_map(void); extern int additional_cpus; #ifdef CONFIG_ACPI_NUMA -/* Proximity bitmap length; _PXM is at most 255 (8 bit)*/ -#ifdef CONFIG_IA64_NR_NODES -#define MAX_PXM_DOMAINS CONFIG_IA64_NR_NODES +#if MAX_NUMNODES > 256 +#define MAX_PXM_DOMAINS MAX_NUMNODES #else #define MAX_PXM_DOMAINS (256) #endif -- cgit v1.2.3 From 86db2f4239e2556cd37b853c2307aa9d43041458 Mon Sep 17 00:00:00 2001 From: Russ Anderson Date: Thu, 20 Apr 2006 17:05:43 -0700 Subject: [IA64-SGI] SN SAL call to inject memory errors The SGI Altix SAL provides an interface for modifying the ECC on memory to create memory errors. The SAL call can be used to inject memory errors for testing MCA recovery code. Signed-off-by: Russ Anderson (rja@sgi.com) Signed-off-by: Tony Luck --- include/asm-ia64/sn/sn_sal.h | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'include') diff --git a/include/asm-ia64/sn/sn_sal.h b/include/asm-ia64/sn/sn_sal.h index 7ddce80e1918..51aca022cf39 100644 --- a/include/asm-ia64/sn/sn_sal.h +++ b/include/asm-ia64/sn/sn_sal.h @@ -85,6 +85,7 @@ #define SN_SAL_GET_PROM_FEATURE_SET 0x02000065 #define SN_SAL_SET_OS_FEATURE_SET 0x02000066 +#define SN_SAL_INJECT_ERROR 0x02000067 /* * Service-specific constants @@ -1138,4 +1139,16 @@ ia64_sn_set_os_feature(int feature) return rv.status; } +static inline int +sn_inject_error(u64 paddr, u64 *data, u64 *ecc) +{ + struct ia64_sal_retval ret_stuff; + unsigned long irq_flags; + + local_irq_save(irq_flags); + ia64_sal_oemcall_nolock(&ret_stuff, SN_SAL_INJECT_ERROR, paddr, (u64)data, + (u64)ecc, 0, 0, 0, 0); + local_irq_restore(irq_flags); + return ret_stuff.status; +} #endif /* _ASM_IA64_SN_SN_SAL_H */ -- cgit v1.2.3 From a72391e42f0a13116995045b3d492d660f96697d Mon Sep 17 00:00:00 2001 From: Satoru Takeuchi Date: Thu, 20 Apr 2006 18:49:48 +0900 Subject: [IA64] eliminate compile time warnings This patch removes following compile time warnings: drivers/pci/pci-sysfs.c: In function `pci_read_legacy_io': drivers/pci/pci-sysfs.c:257: warning: implicit declaration of function `ia64_pci_legacy_read' drivers/pci/pci-sysfs.c: In function `pci_write_legacy_io': drivers/pci/pci-sysfs.c:280: warning: implicit declaration of function `ia64_pci_legacy_write' It also fixes wrong definition of ia64_pci_legacy_write (type of `bus' is not `pci_dev', but `pci_bus'). Signed-Off-By: Satoru Takeuchi Signed-off-by: Tony Luck --- arch/ia64/pci/pci.c | 3 +-- include/asm-ia64/machvec.h | 2 ++ 2 files changed, 3 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/arch/ia64/pci/pci.c b/arch/ia64/pci/pci.c index 9ba32b2d96d0..ab829a22f8a4 100644 --- a/arch/ia64/pci/pci.c +++ b/arch/ia64/pci/pci.c @@ -31,7 +31,6 @@ #include #include - /* * Low-level SAL-based PCI configuration access functions. Note that SAL * calls are already serialized (via sal_lock), so we don't need another @@ -707,7 +706,7 @@ int ia64_pci_legacy_read(struct pci_bus *bus, u16 port, u32 *val, u8 size) * * Simply writes @size bytes of @val to @port. */ -int ia64_pci_legacy_write(struct pci_dev *bus, u16 port, u32 val, u8 size) +int ia64_pci_legacy_write(struct pci_bus *bus, u16 port, u32 val, u8 size) { int ret = size; diff --git a/include/asm-ia64/machvec.h b/include/asm-ia64/machvec.h index c3e4ed8a3e17..a9c995a86c21 100644 --- a/include/asm-ia64/machvec.h +++ b/include/asm-ia64/machvec.h @@ -347,9 +347,11 @@ extern ia64_mv_dma_supported swiotlb_dma_supported; #endif #ifndef platform_pci_legacy_read # define platform_pci_legacy_read ia64_pci_legacy_read +extern int ia64_pci_legacy_read(struct pci_bus *bus, u16 port, u32 *val, u8 size); #endif #ifndef platform_pci_legacy_write # define platform_pci_legacy_write ia64_pci_legacy_write +extern int ia64_pci_legacy_write(struct pci_bus *bus, u16 port, u32 val, u8 size); #endif #ifndef platform_inb # define platform_inb __ia64_inb -- cgit v1.2.3 From e5ecc192dfc5e0b325dd8c99ce4c755714c9acbf Mon Sep 17 00:00:00 2001 From: Christoph Lameter Date: Thu, 13 Apr 2006 18:23:53 -0700 Subject: [IA64] Setup an IA64 specific reclaim distance RECLAIM_DISTANCE is checked on bootup against the SLIT table distances. Zone reclaim is important for system that have higher latencies but not for systems that have multiple nodes on one motherboard and therefore low latencies. We found that on motherboard latencies are typically 1 to 1.4 of local memory access speed whereas multinode systems which benefit from zone reclaim have usually more than 1.5 times the latency of a local access. Set the reclaim distance for IA64 to 1.5 times. Signed-off-by: Christoph Lameter Signed-off-by: Tony Luck --- include/asm-ia64/topology.h | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'include') diff --git a/include/asm-ia64/topology.h b/include/asm-ia64/topology.h index 3ee19dfa46df..616b5ed2aa72 100644 --- a/include/asm-ia64/topology.h +++ b/include/asm-ia64/topology.h @@ -22,6 +22,11 @@ /* Nodes w/o CPUs are preferred for memory allocations, see build_zonelists */ #define PENALTY_FOR_NODE_WITH_CPUS 255 +/* + * Distance above which we begin to use zone reclaim + */ +#define RECLAIM_DISTANCE 15 + /* * Returns the number of the node containing CPU 'cpu' */ -- cgit v1.2.3 From 2fd83038160531245099c3c5b3511fa4b80765eb Mon Sep 17 00:00:00 2001 From: Helge Deller Date: Thu, 20 Apr 2006 20:40:23 +0000 Subject: [PARISC] Further work for multiple page sizes More work towards supporing multiple page sizes on 64-bit. Convert some assumptions that 64bit uses 3 level page tables into testing PT_NLEVELS. Also some BUG() to BUG_ON() conversions and some cleanups to assembler. Signed-off-by: Helge Deller Signed-off-by: Kyle McMartin --- arch/parisc/Kconfig | 31 ++++++++++++++++++++ arch/parisc/kernel/asm-offsets.c | 3 ++ arch/parisc/kernel/entry.S | 36 ++++++++++++++--------- arch/parisc/kernel/head.S | 15 +++++----- arch/parisc/kernel/init_task.c | 10 +++---- arch/parisc/kernel/pacache.S | 25 ++++++++-------- arch/parisc/kernel/syscall.S | 10 +++---- arch/parisc/kernel/vmlinux.lds.S | 54 ++++++++++++++++++++-------------- arch/parisc/mm/init.c | 28 +++++++++--------- include/asm-parisc/page.h | 25 +++++++++++++--- include/asm-parisc/pgtable.h | 63 ++++++++++++++++++++++++++++------------ 11 files changed, 198 insertions(+), 102 deletions(-) (limited to 'include') diff --git a/arch/parisc/Kconfig b/arch/parisc/Kconfig index 19f911c5dd58..910fb3afc0b5 100644 --- a/arch/parisc/Kconfig +++ b/arch/parisc/Kconfig @@ -138,6 +138,37 @@ config 64BIT enable this option otherwise. The 64bit kernel is significantly bigger and slower than the 32bit one. +choice + prompt "Kernel page size" + default PARISC_PAGE_SIZE_4KB if !64BIT + default PARISC_PAGE_SIZE_4KB if 64BIT +# default PARISC_PAGE_SIZE_16KB if 64BIT + +config PARISC_PAGE_SIZE_4KB + bool "4KB" + help + This lets you select the page size of the kernel. For best + performance, a page size of 16KB is recommended. For best + compatibility with 32bit applications, a page size of 4KB should be + selected (the vast majority of 32bit binaries work perfectly fine + with a larger page size). + + 4KB For best 32bit compatibility + 16KB For best performance + 64KB For best performance, might give more overhead. + + If you don't know what to do, choose 4KB. + +config PARISC_PAGE_SIZE_16KB + bool "16KB (EXPERIMENTAL)" + depends on PA8X00 && EXPERIMENTAL + +config PARISC_PAGE_SIZE_64KB + bool "64KB (EXPERIMENTAL)" + depends on PA8X00 && EXPERIMENTAL + +endchoice + config SMP bool "Symmetric multi-processing support" ---help--- diff --git a/arch/parisc/kernel/asm-offsets.c b/arch/parisc/kernel/asm-offsets.c index e23c4e1e3a25..c11a5bc7c067 100644 --- a/arch/parisc/kernel/asm-offsets.c +++ b/arch/parisc/kernel/asm-offsets.c @@ -288,8 +288,11 @@ int main(void) DEFINE(ASM_PGD_ENTRY_SIZE, PGD_ENTRY_SIZE); DEFINE(ASM_PMD_ENTRY_SIZE, PMD_ENTRY_SIZE); DEFINE(ASM_PTE_ENTRY_SIZE, PTE_ENTRY_SIZE); + DEFINE(ASM_PFN_PTE_SHIFT, PFN_PTE_SHIFT); DEFINE(ASM_PT_INITIAL, PT_INITIAL); DEFINE(ASM_PAGE_SIZE, PAGE_SIZE); + DEFINE(ASM_PAGE_SIZE_DIV64, PAGE_SIZE/64); + DEFINE(ASM_PAGE_SIZE_DIV128, PAGE_SIZE/128); BLANK(); DEFINE(EXCDATA_IP, offsetof(struct exception_data, fault_ip)); DEFINE(EXCDATA_SPACE, offsetof(struct exception_data, fault_space)); diff --git a/arch/parisc/kernel/entry.S b/arch/parisc/kernel/entry.S index 7c95d7663c29..d9e53cf0372b 100644 --- a/arch/parisc/kernel/entry.S +++ b/arch/parisc/kernel/entry.S @@ -502,18 +502,20 @@ * all ILP32 processes and all the kernel for machines with * under 4GB of memory) */ .macro L3_ptep pgd,pte,index,va,fault +#if PT_NLEVELS == 3 /* we might have a 2-Level scheme, e.g. with 16kb page size */ extrd,u \va,63-ASM_PGDIR_SHIFT,ASM_BITS_PER_PGD,\index copy %r0,\pte - extrd,u,*= \va,31,32,%r0 + extrd,u,*= \va,63-ASM_PGDIR_SHIFT,64-ASM_PGDIR_SHIFT,%r0 ldw,s \index(\pgd),\pgd - extrd,u,*= \va,31,32,%r0 + extrd,u,*= \va,63-ASM_PGDIR_SHIFT,64-ASM_PGDIR_SHIFT,%r0 bb,>=,n \pgd,_PxD_PRESENT_BIT,\fault - extrd,u,*= \va,31,32,%r0 + extrd,u,*= \va,63-ASM_PGDIR_SHIFT,64-ASM_PGDIR_SHIFT,%r0 shld \pgd,PxD_VALUE_SHIFT,\index - extrd,u,*= \va,31,32,%r0 + extrd,u,*= \va,63-ASM_PGDIR_SHIFT,64-ASM_PGDIR_SHIFT,%r0 copy \index,\pgd - extrd,u,*<> \va,31,32,%r0 + extrd,u,*<> \va,63-ASM_PGDIR_SHIFT,64-ASM_PGDIR_SHIFT,%r0 ldo ASM_PGD_PMD_OFFSET(\pgd),\pgd +#endif L2_ptep \pgd,\pte,\index,\va,\fault .endm @@ -563,10 +565,18 @@ extrd,u,*= \pte,_PAGE_GATEWAY_BIT+32,1,%r0 depd %r0,11,2,\prot /* If Gateway, Set PL2 to 0 */ - /* Get rid of prot bits and convert to page addr for iitlbt and idtlbt */ + /* Enforce uncacheable pages. + * This should ONLY be use for MMIO on PA 2.0 machines. + * Memory/DMA is cache coherent on all PA2.0 machines we support + * (that means T-class is NOT supported) and the memory controllers + * on most of those machines only handles cache transactions. + */ + extrd,u,*= \pte,_PAGE_NO_CACHE_BIT+32,1,%r0 + depi 1,12,1,\prot - depd %r0,63,PAGE_SHIFT,\pte - extrd,s \pte,(63-PAGE_SHIFT)+(63-58),64-PAGE_SHIFT,\pte + /* Drop prot bits and convert to page addr for iitlbt and idtlbt */ + extrd,u \pte,(63-ASM_PFN_PTE_SHIFT)+(63-58),64-PAGE_SHIFT,\pte + depdi _PAGE_SIZE_ENCODING_DEFAULT,63,63-58,\pte .endm /* Identical macro to make_insert_tlb above, except it @@ -584,9 +594,8 @@ /* Get rid of prot bits and convert to page addr for iitlba */ - depi 0,31,PAGE_SHIFT,\pte + depi _PAGE_SIZE_ENCODING_DEFAULT,31,ASM_PFN_PTE_SHIFT,\pte extru \pte,24,25,\pte - .endm /* This is for ILP32 PA2.0 only. The TLB insertion needs @@ -1201,10 +1210,9 @@ intr_save: */ /* adjust isr/ior. */ - - extrd,u %r16,63,7,%r1 /* get high bits from isr for ior */ - depd %r1,31,7,%r17 /* deposit them into ior */ - depdi 0,63,7,%r16 /* clear them from isr */ + extrd,u %r16,63,SPACEID_SHIFT,%r1 /* get high bits from isr for ior */ + depd %r1,31,SPACEID_SHIFT,%r17 /* deposit them into ior */ + depdi 0,63,SPACEID_SHIFT,%r16 /* clear them from isr */ #endif STREG %r16, PT_ISR(%r29) STREG %r17, PT_IOR(%r29) diff --git a/arch/parisc/kernel/head.S b/arch/parisc/kernel/head.S index 0b47afc20690..3e79e62f7b0b 100644 --- a/arch/parisc/kernel/head.S +++ b/arch/parisc/kernel/head.S @@ -76,16 +76,16 @@ $bss_loop: mtctl %r4,%cr24 /* Initialize kernel root pointer */ mtctl %r4,%cr25 /* Initialize user root pointer */ -#ifdef CONFIG_64BIT +#if PT_NLEVELS == 3 /* Set pmd in pgd */ load32 PA(pmd0),%r5 shrd %r5,PxD_VALUE_SHIFT,%r3 - ldo (PxD_FLAG_PRESENT+PxD_FLAG_VALID)(%r3),%r3 + ldo (PxD_FLAG_PRESENT+PxD_FLAG_VALID)(%r3),%r3 stw %r3,ASM_PGD_ENTRY*ASM_PGD_ENTRY_SIZE(%r4) ldo ASM_PMD_ENTRY*ASM_PMD_ENTRY_SIZE(%r5),%r4 #else /* 2-level page table, so pmd == pgd */ - ldo ASM_PGD_ENTRY*ASM_PGD_ENTRY_SIZE(%r4),%r4 + ldo ASM_PGD_ENTRY*ASM_PGD_ENTRY_SIZE(%r4),%r4 #endif /* Fill in pmd with enough pte directories */ @@ -99,7 +99,7 @@ $bss_loop: stw %r3,0(%r4) ldo (ASM_PAGE_SIZE >> PxD_VALUE_SHIFT)(%r3),%r3 addib,> -1,%r1,1b -#ifdef CONFIG_64BIT +#if PT_NLEVELS == 3 ldo ASM_PMD_ENTRY_SIZE(%r4),%r4 #else ldo ASM_PGD_ENTRY_SIZE(%r4),%r4 @@ -107,13 +107,14 @@ $bss_loop: /* Now initialize the PTEs themselves */ - ldo _PAGE_KERNEL(%r0),%r3 /* Hardwired 0 phys addr start */ + ldo 0+_PAGE_KERNEL(%r0),%r3 /* Hardwired 0 phys addr start */ + ldi (1<<(KERNEL_INITIAL_ORDER-PAGE_SHIFT)),%r11 /* PFN count */ load32 PA(pg0),%r1 $pgt_fill_loop: STREGM %r3,ASM_PTE_ENTRY_SIZE(%r1) - ldo ASM_PAGE_SIZE(%r3),%r3 - bb,>= %r3,31-KERNEL_INITIAL_ORDER,$pgt_fill_loop + ldo (1< -1,%r11,$pgt_fill_loop nop /* Load the return address...er...crash 'n burn */ diff --git a/arch/parisc/kernel/init_task.c b/arch/parisc/kernel/init_task.c index 7e898fd64415..8384bf9cecd2 100644 --- a/arch/parisc/kernel/init_task.c +++ b/arch/parisc/kernel/init_task.c @@ -53,17 +53,17 @@ union thread_union init_thread_union __attribute__((aligned(128))) __attribute__((__section__(".data.init_task"))) = { INIT_THREAD_INFO(init_task) }; -#ifdef __LP64__ +#if PT_NLEVELS == 3 /* NOTE: This layout exactly conforms to the hybrid L2/L3 page table layout * with the first pmd adjacent to the pgd and below it. gcc doesn't actually * guarantee that global objects will be laid out in memory in the same order * as the order of declaration, so put these in different sections and use * the linker script to order them. */ -pmd_t pmd0[PTRS_PER_PMD] __attribute__ ((aligned(PAGE_SIZE))) __attribute__ ((__section__ (".data.vm0.pmd"))) = { {0}, }; - +pmd_t pmd0[PTRS_PER_PMD] __attribute__ ((__section__ (".data.vm0.pmd"), aligned(PAGE_SIZE))); #endif -pgd_t swapper_pg_dir[PTRS_PER_PGD] __attribute__ ((aligned(PAGE_SIZE))) __attribute__ ((__section__ (".data.vm0.pgd"))) = { {0}, }; -pte_t pg0[PT_INITIAL * PTRS_PER_PTE] __attribute__ ((aligned(PAGE_SIZE))) __attribute__ ((__section__ (".data.vm0.pte"))) = { {0}, }; + +pgd_t swapper_pg_dir[PTRS_PER_PGD] __attribute__ ((__section__ (".data.vm0.pgd"), aligned(PAGE_SIZE))); +pte_t pg0[PT_INITIAL * PTRS_PER_PTE] __attribute__ ((__section__ (".data.vm0.pte"), aligned(PAGE_SIZE))); /* * Initial task structure. diff --git a/arch/parisc/kernel/pacache.S b/arch/parisc/kernel/pacache.S index 7a4f07e8d3c3..f600556414d1 100644 --- a/arch/parisc/kernel/pacache.S +++ b/arch/parisc/kernel/pacache.S @@ -65,7 +65,7 @@ flush_tlb_all_local: */ /* pcxt_ssm_bug - relied upon translation! PA 2.0 Arch. F-4 and F-5 */ - rsm PSW_SM_I, %r19 /* save I-bit state */ + rsm PSW_SM_I, %r19 /* save I-bit state */ load32 PA(1f), %r1 nop nop @@ -84,8 +84,7 @@ flush_tlb_all_local: rfi nop -1: ldil L%PA(cache_info), %r1 - ldo R%PA(cache_info)(%r1), %r1 +1: load32 PA(cache_info), %r1 /* Flush Instruction Tlb */ @@ -212,8 +211,7 @@ flush_instruction_cache_local: .entry mtsp %r0, %sr1 - ldil L%cache_info, %r1 - ldo R%cache_info(%r1), %r1 + load32 cache_info, %r1 /* Flush Instruction Cache */ @@ -254,8 +252,7 @@ flush_data_cache_local: .entry mtsp %r0, %sr1 - ldil L%cache_info, %r1 - ldo R%cache_info(%r1), %r1 + load32 cache_info, %r1 /* Flush Data Cache */ @@ -303,7 +300,8 @@ copy_user_page_asm: */ ldd 0(%r25), %r19 - ldi 32, %r1 /* PAGE_SIZE/128 == 32 */ + ldi ASM_PAGE_SIZE_DIV128, %r1 + ldw 64(%r25), %r0 /* prefetch 1 cacheline ahead */ ldw 128(%r25), %r0 /* prefetch 2 */ @@ -368,7 +366,7 @@ copy_user_page_asm: * use ldd/std on a 32 bit kernel. */ ldw 0(%r25), %r19 - ldi 64, %r1 /* PAGE_SIZE/64 == 64 */ + ldi ASM_PAGE_SIZE_DIV64, %r1 1: ldw 4(%r25), %r20 @@ -461,6 +459,7 @@ copy_user_page_asm: sub %r25, %r1, %r23 /* move physical addr into non shadowed reg */ ldil L%(TMPALIAS_MAP_START), %r28 + /* FIXME for different page sizes != 4k */ #ifdef CONFIG_64BIT extrd,u %r26,56,32, %r26 /* convert phys addr to tlb insert format */ extrd,u %r23,56,32, %r23 /* convert phys addr to tlb insert format */ @@ -551,6 +550,7 @@ __clear_user_page_asm: #ifdef CONFIG_64BIT #if (TMPALIAS_MAP_START >= 0x80000000) depdi 0, 31,32, %r28 /* clear any sign extension */ + /* FIXME: page size dependend */ #endif extrd,u %r26, 56,32, %r26 /* convert phys addr to tlb insert format */ depd %r25, 63,22, %r28 /* Form aliased virtual address 'to' */ @@ -566,10 +566,10 @@ __clear_user_page_asm: pdtlb 0(%r28) #ifdef CONFIG_64BIT - ldi 32, %r1 /* PAGE_SIZE/128 == 32 */ + ldi ASM_PAGE_SIZE_DIV128, %r1 /* PREFETCH (Write) has not (yet) been proven to help here */ -/* #define PREFETCHW_OP ldd 256(%0), %r0 */ + /* #define PREFETCHW_OP ldd 256(%0), %r0 */ 1: std %r0, 0(%r28) std %r0, 8(%r28) @@ -591,8 +591,7 @@ __clear_user_page_asm: ldo 128(%r28), %r28 #else /* ! CONFIG_64BIT */ - - ldi 64, %r1 /* PAGE_SIZE/64 == 64 */ + ldi ASM_PAGE_SIZE_DIV64, %r1 1: stw %r0, 0(%r28) diff --git a/arch/parisc/kernel/syscall.S b/arch/parisc/kernel/syscall.S index af88afef41bd..479d9a017cd1 100644 --- a/arch/parisc/kernel/syscall.S +++ b/arch/parisc/kernel/syscall.S @@ -55,7 +55,7 @@ * pointers. */ - .align 4096 + .align ASM_PAGE_SIZE linux_gateway_page: /* ADDRESS 0x00 to 0xb0 = 176 bytes / 4 bytes per insn = 44 insns */ @@ -632,7 +632,7 @@ cas_action: end_compare_and_swap: /* Make sure nothing else is placed on this page */ - .align 4096 + .align ASM_PAGE_SIZE .export end_linux_gateway_page end_linux_gateway_page: @@ -652,7 +652,7 @@ end_linux_gateway_page: .section .rodata,"a" - .align 4096 + .align ASM_PAGE_SIZE /* Light-weight-syscall table */ /* Start of lws table. */ .export lws_table @@ -662,14 +662,14 @@ lws_table: LWS_ENTRY(compare_and_swap64) /* 1 - ELF64 Atomic compare and swap */ /* End of lws table */ - .align 4096 + .align ASM_PAGE_SIZE .export sys_call_table .Lsys_call_table: sys_call_table: #include "syscall_table.S" #ifdef CONFIG_64BIT - .align 4096 + .align ASM_PAGE_SIZE .export sys_call_table64 .Lsys_call_table64: sys_call_table64: diff --git a/arch/parisc/kernel/vmlinux.lds.S b/arch/parisc/kernel/vmlinux.lds.S index 6d6436a6b624..94dcc03a28ed 100644 --- a/arch/parisc/kernel/vmlinux.lds.S +++ b/arch/parisc/kernel/vmlinux.lds.S @@ -6,6 +6,7 @@ * Copyright (C) 2000 Michael Ang * Copyright (C) 2002 Randolph Chung * Copyright (C) 2003 James Bottomley + * Copyright (C) 2006 Helge Deller * * * This program is free software; you can redistribute it and/or modify @@ -27,6 +28,7 @@ /* needed for the processor specific cache alignment size */ #include #include +#include /* ld script to make hppa Linux kernel */ #ifndef CONFIG_64BIT @@ -68,7 +70,7 @@ SECTIONS RODATA /* writeable */ - . = ALIGN(4096); /* Make sure this is page aligned so + . = ALIGN(ASM_PAGE_SIZE); /* Make sure this is page aligned so that we can properly leave these as writable */ data_start = .; @@ -81,23 +83,17 @@ SECTIONS __start___unwind = .; /* unwind info */ .PARISC.unwind : { *(.PARISC.unwind) } __stop___unwind = .; - + + /* rarely changed data like cpu maps */ + . = ALIGN(16); + .data.read_mostly : { *(.data.read_mostly) } + + . = ALIGN(L1_CACHE_BYTES); .data : { /* Data */ *(.data) - *(.data.vm0.pmd) - *(.data.vm0.pgd) - *(.data.vm0.pte) CONSTRUCTORS } - . = ALIGN(4096); - /* nosave data is really only used for software suspend...it's here - * just in case we ever implement it */ - __nosave_begin = .; - .data_nosave : { *(.data.nosave) } - . = ALIGN(4096); - __nosave_end = .; - . = ALIGN(L1_CACHE_BYTES); .data.cacheline_aligned : { *(.data.cacheline_aligned) } @@ -105,12 +101,29 @@ SECTIONS . = ALIGN(16); .data.lock_aligned : { *(.data.lock_aligned) } - /* rarely changed data like cpu maps */ - . = ALIGN(16); - .data.read_mostly : { *(.data.read_mostly) } + . = ALIGN(ASM_PAGE_SIZE); + /* nosave data is really only used for software suspend...it's here + * just in case we ever implement it */ + __nosave_begin = .; + .data_nosave : { *(.data.nosave) } + . = ALIGN(ASM_PAGE_SIZE); + __nosave_end = .; _edata = .; /* End of data section */ + __bss_start = .; /* BSS */ + /* page table entries need to be PAGE_SIZE aligned */ + . = ALIGN(ASM_PAGE_SIZE); + .data.vmpages : { + *(.data.vm0.pmd) + *(.data.vm0.pgd) + *(.data.vm0.pte) + } + .bss : { *(.bss) *(COMMON) } + __bss_stop = .; + + + /* assembler code expects init_task to be 16k aligned */ . = ALIGN(16384); /* init_task */ .data.init_task : { *(.data.init_task) } @@ -126,6 +139,7 @@ SECTIONS .dlt : { *(.dlt) } #endif + /* reserve space for interrupt stack by aligning __init* to 16k */ . = ALIGN(16384); __init_begin = .; .init.text : { @@ -166,7 +180,7 @@ SECTIONS from .altinstructions and .eh_frame */ .exit.text : { *(.exit.text) } .exit.data : { *(.exit.data) } - . = ALIGN(4096); + . = ALIGN(ASM_PAGE_SIZE); __initramfs_start = .; .init.ramfs : { *(.init.ramfs) } __initramfs_end = .; @@ -174,14 +188,10 @@ SECTIONS __per_cpu_start = .; .data.percpu : { *(.data.percpu) } __per_cpu_end = .; - . = ALIGN(4096); + . = ALIGN(ASM_PAGE_SIZE); __init_end = .; /* freed after init ends here */ - __bss_start = .; /* BSS */ - .bss : { *(.bss) *(COMMON) } - __bss_stop = .; - _end = . ; /* Sections to be discarded */ diff --git a/arch/parisc/mm/init.c b/arch/parisc/mm/init.c index 3796be67cd53..631712562656 100644 --- a/arch/parisc/mm/init.c +++ b/arch/parisc/mm/init.c @@ -6,6 +6,7 @@ * changed by Philipp Rumpf * Copyright 1999 Philipp Rumpf (prumpf@tux.org) * Copyright 2004 Randolph Chung (tausq@debian.org) + * Copyright 2006 Helge Deller (deller@gmx.de) * */ @@ -371,8 +372,8 @@ static void __init setup_bootmem(void) void free_initmem(void) { - unsigned long addr; - + unsigned long addr, init_begin, init_end; + printk(KERN_INFO "Freeing unused kernel memory: "); #ifdef CONFIG_DEBUG_KERNEL @@ -395,8 +396,11 @@ void free_initmem(void) local_irq_enable(); #endif - addr = (unsigned long)(&__init_begin); - for (; addr < (unsigned long)(&__init_end); addr += PAGE_SIZE) { + /* align __init_begin and __init_end to page size, + ignoring linker script where we might have tried to save RAM */ + init_begin = PAGE_ALIGN((unsigned long)(&__init_begin)); + init_end = PAGE_ALIGN((unsigned long)(&__init_end)); + for (addr = init_begin; addr < init_end; addr += PAGE_SIZE) { ClearPageReserved(virt_to_page(addr)); init_page_count(virt_to_page(addr)); free_page(addr); @@ -407,7 +411,7 @@ void free_initmem(void) /* set up a new led state on systems shipped LED State panel */ pdc_chassis_send_status(PDC_CHASSIS_DIRECT_BCOMPLETE); - printk("%luk freed\n", (unsigned long)(&__init_end - &__init_begin) >> 10); + printk("%luk freed\n", (init_end - init_begin) >> 10); } @@ -639,11 +643,13 @@ static void __init map_pages(unsigned long start_vaddr, unsigned long start_padd * Map the fault vector writable so we can * write the HPMC checksum. */ +#if defined(CONFIG_PARISC_PAGE_SIZE_4KB) if (address >= ro_start && address < ro_end && address != fv_addr && address != gw_addr) pte = __mk_pte(address, PAGE_KERNEL_RO); else +#endif pte = __mk_pte(address, pgprot); if (address >= end_paddr) @@ -874,8 +880,7 @@ unsigned long alloc_sid(void) flush_tlb_all(); /* flush_tlb_all() calls recycle_sids() */ spin_lock(&sid_lock); } - if (free_space_ids == 0) - BUG(); + BUG_ON(free_space_ids == 0); } free_space_ids--; @@ -899,8 +904,7 @@ void free_sid(unsigned long spaceid) spin_lock(&sid_lock); - if (*dirty_space_offset & (1L << index)) - BUG(); /* attempt to free space id twice */ + BUG_ON(*dirty_space_offset & (1L << index)); /* attempt to free space id twice */ *dirty_space_offset |= (1L << index); dirty_space_ids++; @@ -975,7 +979,7 @@ static void recycle_sids(void) static unsigned long recycle_ndirty; static unsigned long recycle_dirty_array[SID_ARRAY_SIZE]; -static unsigned int recycle_inuse = 0; +static unsigned int recycle_inuse; void flush_tlb_all(void) { @@ -984,9 +988,7 @@ void flush_tlb_all(void) do_recycle = 0; spin_lock(&sid_lock); if (dirty_space_ids > RECYCLE_THRESHOLD) { - if (recycle_inuse) { - BUG(); /* FIXME: Use a semaphore/wait queue here */ - } + BUG_ON(recycle_inuse); /* FIXME: Use a semaphore/wait queue here */ get_dirty_sids(&recycle_ndirty,recycle_dirty_array); recycle_inuse++; do_recycle++; diff --git a/include/asm-parisc/page.h b/include/asm-parisc/page.h index 45e02aa5bf4b..c0dd461fb8f1 100644 --- a/include/asm-parisc/page.h +++ b/include/asm-parisc/page.h @@ -1,13 +1,30 @@ #ifndef _PARISC_PAGE_H #define _PARISC_PAGE_H -/* PAGE_SHIFT determines the page size */ -#define PAGE_SHIFT 12 -#define PAGE_SIZE (1UL << PAGE_SHIFT) -#define PAGE_MASK (~(PAGE_SIZE-1)) +#if !defined(__KERNEL__) +/* this is for userspace applications (4k page size) */ +# define PAGE_SHIFT 12 /* 4k */ +# define PAGE_SIZE (1UL << PAGE_SHIFT) +# define PAGE_MASK (~(PAGE_SIZE-1)) +#endif + #ifdef __KERNEL__ #include + +#if defined(CONFIG_PARISC_PAGE_SIZE_4KB) +# define PAGE_SHIFT 12 /* 4k */ +#elif defined(CONFIG_PARISC_PAGE_SIZE_16KB) +# define PAGE_SHIFT 14 /* 16k */ +#elif defined(CONFIG_PARISC_PAGE_SIZE_64KB) +# define PAGE_SHIFT 16 /* 64k */ +#else +# error "unknown default kernel page size" +#endif +#define PAGE_SIZE (1UL << PAGE_SHIFT) +#define PAGE_MASK (~(PAGE_SIZE-1)) + + #ifndef __ASSEMBLY__ #include diff --git a/include/asm-parisc/pgtable.h b/include/asm-parisc/pgtable.h index 4e34c6b44059..aec089eb8b85 100644 --- a/include/asm-parisc/pgtable.h +++ b/include/asm-parisc/pgtable.h @@ -59,16 +59,15 @@ #define ISTACK_SIZE 32768 /* Interrupt Stack Size */ #define ISTACK_ORDER 3 -/* This is the size of the initially mapped kernel memory (i.e. currently - * 0 to 1<<23 == 8MB */ +/* This is the size of the initially mapped kernel memory */ #ifdef CONFIG_64BIT -#define KERNEL_INITIAL_ORDER 24 +#define KERNEL_INITIAL_ORDER 24 /* 0 to 1<<24 = 16MB */ #else -#define KERNEL_INITIAL_ORDER 23 +#define KERNEL_INITIAL_ORDER 23 /* 0 to 1<<23 = 8MB */ #endif #define KERNEL_INITIAL_SIZE (1 << KERNEL_INITIAL_ORDER) -#ifdef CONFIG_64BIT +#if defined(CONFIG_64BIT) && defined(CONFIG_PARISC_PAGE_SIZE_4KB) #define PT_NLEVELS 3 #define PGD_ORDER 1 /* Number of pages per pgd */ #define PMD_ORDER 1 /* Number of pages per pmd */ @@ -111,11 +110,15 @@ #define MAX_ADDRBITS (PGDIR_SHIFT + BITS_PER_PGD) #define MAX_ADDRESS (1UL << MAX_ADDRBITS) -#define SPACEID_SHIFT (MAX_ADDRBITS - 32) +#define SPACEID_SHIFT (MAX_ADDRBITS - 32) /* This calculates the number of initial pages we need for the initial * page tables */ -#define PT_INITIAL (1 << (KERNEL_INITIAL_ORDER - PMD_SHIFT)) +#if (KERNEL_INITIAL_ORDER) >= (PMD_SHIFT) +# define PT_INITIAL (1 << (KERNEL_INITIAL_ORDER - PMD_SHIFT)) +#else +# define PT_INITIAL (1) /* all initial PTEs fit into one page */ +#endif /* * pgd entries used up by user/kernel: @@ -160,6 +163,10 @@ extern void *vmalloc_start; * to zero */ #define PTE_SHIFT xlate_pabit(_PAGE_USER_BIT) +/* PFN_PTE_SHIFT defines the shift of a PTE value to access the PFN field */ +#define PFN_PTE_SHIFT 12 + + /* this is how many bits may be used by the file functions */ #define PTE_FILE_MAX_BITS (BITS_PER_LONG - PTE_SHIFT) @@ -188,7 +195,8 @@ extern void *vmalloc_start; /* The pgd/pmd contains a ptr (in phys addr space); since all pgds/pmds * are page-aligned, we don't care about the PAGE_OFFSET bits, except * for a few meta-information bits, so we shift the address to be - * able to effectively address 40-bits of physical address space. */ + * able to effectively address 40/42/44-bits of physical address space + * depending on 4k/16k/64k PAGE_SIZE */ #define _PxD_PRESENT_BIT 31 #define _PxD_ATTACHED_BIT 30 #define _PxD_VALID_BIT 29 @@ -198,7 +206,7 @@ extern void *vmalloc_start; #define PxD_FLAG_VALID (1 << xlate_pabit(_PxD_VALID_BIT)) #define PxD_FLAG_MASK (0xf) #define PxD_FLAG_SHIFT (4) -#define PxD_VALUE_SHIFT (8) +#define PxD_VALUE_SHIFT (8) /* (PAGE_SHIFT-PxD_FLAG_SHIFT) */ #ifndef __ASSEMBLY__ @@ -246,6 +254,7 @@ extern void *vmalloc_start; #define __S110 PAGE_RWX #define __S111 PAGE_RWX + extern pgd_t swapper_pg_dir[]; /* declared in init_task.c */ /* initial page tables for 0-8MB for kernel */ @@ -272,7 +281,7 @@ extern unsigned long *empty_zero_page; #define pgd_flag(x) (pgd_val(x) & PxD_FLAG_MASK) #define pgd_address(x) ((unsigned long)(pgd_val(x) &~ PxD_FLAG_MASK) << PxD_VALUE_SHIFT) -#ifdef CONFIG_64BIT +#if PT_NLEVELS == 3 /* The first entry of the permanent pmd is not there if it contains * the gateway marker */ #define pmd_none(x) (!pmd_val(x) || pmd_flag(x) == PxD_FLAG_ATTACHED) @@ -282,7 +291,7 @@ extern unsigned long *empty_zero_page; #define pmd_bad(x) (!(pmd_flag(x) & PxD_FLAG_VALID)) #define pmd_present(x) (pmd_flag(x) & PxD_FLAG_PRESENT) static inline void pmd_clear(pmd_t *pmd) { -#ifdef CONFIG_64BIT +#if PT_NLEVELS == 3 if (pmd_flag(*pmd) & PxD_FLAG_ATTACHED) /* This is the entry pointing to the permanent pmd * attached to the pgd; cannot clear it */ @@ -303,7 +312,7 @@ static inline void pmd_clear(pmd_t *pmd) { #define pgd_bad(x) (!(pgd_flag(x) & PxD_FLAG_VALID)) #define pgd_present(x) (pgd_flag(x) & PxD_FLAG_PRESENT) static inline void pgd_clear(pgd_t *pgd) { -#ifdef CONFIG_64BIT +#if PT_NLEVELS == 3 if(pgd_flag(*pgd) & PxD_FLAG_ATTACHED) /* This is the permanent pmd attached to the pgd; cannot * free it */ @@ -351,7 +360,7 @@ extern inline pte_t pte_mkwrite(pte_t pte) { pte_val(pte) |= _PAGE_WRITE; return ({ \ pte_t __pte; \ \ - pte_val(__pte) = ((addr)+pgprot_val(pgprot)); \ + pte_val(__pte) = ((((addr)>>PAGE_SHIFT)<> PAGE_SHIFT) +#define pte_pfn(x) (pte_val(x) >> PFN_PTE_SHIFT) #define pte_page(pte) (pfn_to_page(pte_pfn(pte))) @@ -499,6 +504,26 @@ static inline void ptep_set_wrprotect(struct mm_struct *mm, unsigned long addr, #endif /* !__ASSEMBLY__ */ + +/* TLB page size encoding - see table 3-1 in parisc20.pdf */ +#define _PAGE_SIZE_ENCODING_4K 0 +#define _PAGE_SIZE_ENCODING_16K 1 +#define _PAGE_SIZE_ENCODING_64K 2 +#define _PAGE_SIZE_ENCODING_256K 3 +#define _PAGE_SIZE_ENCODING_1M 4 +#define _PAGE_SIZE_ENCODING_4M 5 +#define _PAGE_SIZE_ENCODING_16M 6 +#define _PAGE_SIZE_ENCODING_64M 7 + +#if defined(CONFIG_PARISC_PAGE_SIZE_4KB) +# define _PAGE_SIZE_ENCODING_DEFAULT _PAGE_SIZE_ENCODING_4K +#elif defined(CONFIG_PARISC_PAGE_SIZE_16KB) +# define _PAGE_SIZE_ENCODING_DEFAULT _PAGE_SIZE_ENCODING_16K +#elif defined(CONFIG_PARISC_PAGE_SIZE_64KB) +# define _PAGE_SIZE_ENCODING_DEFAULT _PAGE_SIZE_ENCODING_64K +#endif + + #define io_remap_pfn_range(vma, vaddr, pfn, size, prot) \ remap_pfn_range(vma, vaddr, pfn, size, prot) -- cgit v1.2.3 From 6ca773cf8b9dc19989c9b44635292b1ba80f9112 Mon Sep 17 00:00:00 2001 From: Kyle McMartin Date: Thu, 20 Apr 2006 04:44:07 +0000 Subject: [PARISC] Add new entries to the syscall table Most are easy, but sync_file_range needed special handling when entering through the 32-bit syscall table. Signed-off-by: Kyle McMartin --- arch/parisc/kernel/sys_parisc.c | 8 ++++++++ arch/parisc/kernel/syscall_table.S | 8 +++++++- include/asm-parisc/unistd.h | 8 +++++++- 3 files changed, 22 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/arch/parisc/kernel/sys_parisc.c b/arch/parisc/kernel/sys_parisc.c index d15a1d53e101..8b5df98e2b31 100644 --- a/arch/parisc/kernel/sys_parisc.c +++ b/arch/parisc/kernel/sys_parisc.c @@ -231,6 +231,14 @@ asmlinkage long parisc_fadvise64_64(int fd, (loff_t)high_len << 32 | low_len, advice); } +asmlinkage long parisc_sync_file_range(int fd, + u32 hi_off, u32 lo_off, u32 hi_nbytes, u32 lo_nbytes, + unsigned int flags) +{ + return sys_sync_file_range(fd, (loff_t)hi_off << 32 | lo_off, + (loff_t)hi_nbytes << 32 | lo_nbytes, flags); +} + asmlinkage unsigned long sys_alloc_hugepages(int key, unsigned long addr, unsigned long len, int prot, int flag) { return -ENOMEM; diff --git a/arch/parisc/kernel/syscall_table.S b/arch/parisc/kernel/syscall_table.S index bbeeb614cfab..e27b432f90a8 100644 --- a/arch/parisc/kernel/syscall_table.S +++ b/arch/parisc/kernel/syscall_table.S @@ -13,7 +13,7 @@ * Copyright (C) 2001 Helge Deller * Copyright (C) 2000-2001 Thomas Bogendoerfer * Copyright (C) 2002 Randolph Chung - * + * Copyright (C) 2005-2006 Kyle McMartin * * 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 @@ -393,5 +393,11 @@ ENTRY_SAME(readlinkat) /* 285 */ ENTRY_SAME(fchmodat) ENTRY_SAME(faccessat) + ENTRY_SAME(unshare) + ENTRY_COMP(set_robust_list) + ENTRY_COMP(get_robust_list) /* 290 */ + ENTRY_SAME(splice) + ENTRY_OURS(sync_file_range) + ENTRY_SAME(tee) /* Nothing yet */ diff --git a/include/asm-parisc/unistd.h b/include/asm-parisc/unistd.h index c56fccbf34ad..0e1a30be2e30 100644 --- a/include/asm-parisc/unistd.h +++ b/include/asm-parisc/unistd.h @@ -780,8 +780,14 @@ #define __NR_readlinkat (__NR_Linux + 285) #define __NR_fchmodat (__NR_Linux + 286) #define __NR_faccessat (__NR_Linux + 287) +#define __NR_unshare (__NR_Linux + 288) +#define __NR_set_robust_list (__NR_Linux + 289) +#define __NR_get_robust_list (__NR_Linux + 290) +#define __NR_splice (__NR_Linux + 291) +#define __NR_sync_file_range (__NR_Linux + 292) +#define __NR_tee (__NR_Linux + 293) -#define __NR_Linux_syscalls 288 +#define __NR_Linux_syscalls 294 #define HPUX_GATEWAY_ADDR 0xC0000004 #define LINUX_GATEWAY_ADDR 0x100 -- cgit v1.2.3 From 1b52d7c2210b9a64c5cba6aded478c8217a8853c Mon Sep 17 00:00:00 2001 From: Kyle McMartin Date: Thu, 20 Apr 2006 21:16:32 +0000 Subject: [PARISC] Make ioremap default to _nocache Since it is way more work to change most drivers to comply with parisc, take the easy way out and make ioremap _NO_CACHE by default. This is in line with what powerpc does. Signed-off-by: Kyle McMartin --- include/asm-parisc/io.h | 17 +++++------------ 1 file changed, 5 insertions(+), 12 deletions(-) (limited to 'include') diff --git a/include/asm-parisc/io.h b/include/asm-parisc/io.h index 29da31194b91..244f6b8883f4 100644 --- a/include/asm-parisc/io.h +++ b/include/asm-parisc/io.h @@ -126,24 +126,17 @@ static inline void gsc_writeq(unsigned long long val, unsigned long addr) extern void __iomem * __ioremap(unsigned long offset, unsigned long size, unsigned long flags); -extern inline void __iomem * ioremap(unsigned long offset, unsigned long size) -{ - return __ioremap(offset, size, 0); -} - -/* - * This one maps high address device memory and turns off caching for that area. - * it's useful if some control registers are in such an area and write combining - * or read caching is not desirable: +/* Most machines react poorly to I/O-space being cacheable... Instead let's + * define ioremap() in terms of ioremap_nocache(). */ -extern inline void * ioremap_nocache(unsigned long offset, unsigned long size) +extern inline void __iomem * ioremap(unsigned long offset, unsigned long size) { - return __ioremap(offset, size, _PAGE_NO_CACHE /* _PAGE_PCD */); + return __ioremap(offset, size, _PAGE_NO_CACHE); } +#define ioremap_nocache(off, sz) ioremap((off), (sz)) extern void iounmap(void __iomem *addr); - static inline unsigned char __raw_readb(const volatile void __iomem *addr) { return (*(volatile unsigned char __force *) (addr)); -- cgit v1.2.3 From 818667f7c40dd0bd14029b5ac1d0f5282e12310e Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Thu, 20 Apr 2006 20:02:03 +0200 Subject: [PATCH] softmac: fix SIOCSIWAP There are some bugs in the current implementation of the SIOCSIWAP wext, for example that when you do it twice and it fails, it may still try another access point for some reason. This patch fixes this by introducing a new flag that tells the association code that the bssid that is in use was fixed by the user and shouldn't be deviated from. Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- include/net/ieee80211softmac.h | 5 ++++- net/ieee80211/softmac/ieee80211softmac_assoc.c | 20 ++++++++++++++---- net/ieee80211/softmac/ieee80211softmac_module.c | 2 ++ net/ieee80211/softmac/ieee80211softmac_wx.c | 27 ++++++++++++++++--------- 4 files changed, 39 insertions(+), 15 deletions(-) (limited to 'include') diff --git a/include/net/ieee80211softmac.h b/include/net/ieee80211softmac.h index 6b3693f05ca0..b1ebfbae397f 100644 --- a/include/net/ieee80211softmac.h +++ b/include/net/ieee80211softmac.h @@ -96,10 +96,13 @@ struct ieee80211softmac_assoc_info { * * bssvalid is true if we found a matching network * and saved it's BSSID into the bssid above. + * + * bssfixed is used for SIOCSIWAP. */ u8 static_essid:1, associating:1, - bssvalid:1; + bssvalid:1, + bssfixed:1; /* Scan retries remaining */ int scan_retry; diff --git a/net/ieee80211/softmac/ieee80211softmac_assoc.c b/net/ieee80211/softmac/ieee80211softmac_assoc.c index 4498023841dc..fb79ce7d6439 100644 --- a/net/ieee80211/softmac/ieee80211softmac_assoc.c +++ b/net/ieee80211/softmac/ieee80211softmac_assoc.c @@ -144,6 +144,12 @@ network_matches_request(struct ieee80211softmac_device *mac, struct ieee80211_ne if (!we_support_all_basic_rates(mac, net->rates_ex, net->rates_ex_len)) return 0; + /* assume that users know what they're doing ... + * (note we don't let them select a net we're incompatible with) */ + if (mac->associnfo.bssfixed) { + return !memcmp(mac->associnfo.bssid, net->bssid, ETH_ALEN); + } + /* if 'ANY' network requested, take any that doesn't have privacy enabled */ if (mac->associnfo.req_essid.len == 0 && !(net->capability & WLAN_CAPABILITY_PRIVACY)) @@ -176,7 +182,7 @@ ieee80211softmac_assoc_work(void *d) ieee80211softmac_disassoc(mac, WLAN_REASON_DISASSOC_STA_HAS_LEFT); /* try to find the requested network in our list, if we found one already */ - if (mac->associnfo.bssvalid) + if (mac->associnfo.bssvalid || mac->associnfo.bssfixed) found = ieee80211softmac_get_network_by_bssid(mac, mac->associnfo.bssid); /* Search the ieee80211 networks for this network if we didn't find it by bssid, @@ -241,19 +247,25 @@ ieee80211softmac_assoc_work(void *d) if (ieee80211softmac_start_scan(mac)) dprintk(KERN_INFO PFX "Associate: failed to initiate scan. Is device up?\n"); return; - } - else { + } else { spin_lock_irqsave(&mac->lock, flags); mac->associnfo.associating = 0; mac->associated = 0; spin_unlock_irqrestore(&mac->lock, flags); dprintk(KERN_INFO PFX "Unable to find matching network after scan!\n"); + /* reset the retry counter for the next user request since we + * break out and don't reschedule ourselves after this point. */ + mac->associnfo.scan_retry = IEEE80211SOFTMAC_ASSOC_SCAN_RETRY_LIMIT; ieee80211softmac_call_events(mac, IEEE80211SOFTMAC_EVENT_ASSOCIATE_NET_NOT_FOUND, NULL); return; } } - + + /* reset the retry counter for the next user request since we + * now found a net and will try to associate to it, but not + * schedule this function again. */ + mac->associnfo.scan_retry = IEEE80211SOFTMAC_ASSOC_SCAN_RETRY_LIMIT; mac->associnfo.bssvalid = 1; memcpy(mac->associnfo.bssid, found->bssid, ETH_ALEN); /* copy the ESSID for displaying it */ diff --git a/net/ieee80211/softmac/ieee80211softmac_module.c b/net/ieee80211/softmac/ieee80211softmac_module.c index 60f06a31f0d1..be83bdc1644a 100644 --- a/net/ieee80211/softmac/ieee80211softmac_module.c +++ b/net/ieee80211/softmac/ieee80211softmac_module.c @@ -45,6 +45,8 @@ struct net_device *alloc_ieee80211softmac(int sizeof_priv) softmac->ieee->handle_disassoc = ieee80211softmac_handle_disassoc; softmac->scaninfo = NULL; + softmac->associnfo.scan_retry = IEEE80211SOFTMAC_ASSOC_SCAN_RETRY_LIMIT; + /* TODO: initialise all the other callbacks in the ieee struct * (once they're written) */ diff --git a/net/ieee80211/softmac/ieee80211softmac_wx.c b/net/ieee80211/softmac/ieee80211softmac_wx.c index 00f0d4f71897..27edb2b5581a 100644 --- a/net/ieee80211/softmac/ieee80211softmac_wx.c +++ b/net/ieee80211/softmac/ieee80211softmac_wx.c @@ -27,7 +27,8 @@ #include "ieee80211softmac_priv.h" #include - +/* for is_broadcast_ether_addr and is_zero_ether_addr */ +#include int ieee80211softmac_wx_trigger_scan(struct net_device *net_dev, @@ -83,7 +84,6 @@ ieee80211softmac_wx_set_essid(struct net_device *net_dev, sm->associnfo.static_essid = 1; } } - sm->associnfo.scan_retry = IEEE80211SOFTMAC_ASSOC_SCAN_RETRY_LIMIT; /* set our requested ESSID length. * If applicable, we have already copied the data in */ @@ -310,8 +310,6 @@ ieee80211softmac_wx_set_wap(struct net_device *net_dev, char *extra) { struct ieee80211softmac_device *mac = ieee80211_priv(net_dev); - static const unsigned char any[] = {0xff, 0xff, 0xff, 0xff, 0xff, 0xff}; - static const unsigned char off[] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00}; unsigned long flags; /* sanity check */ @@ -320,10 +318,17 @@ ieee80211softmac_wx_set_wap(struct net_device *net_dev, } spin_lock_irqsave(&mac->lock, flags); - if (!memcmp(any, data->ap_addr.sa_data, ETH_ALEN) || - !memcmp(off, data->ap_addr.sa_data, ETH_ALEN)) { - schedule_work(&mac->associnfo.work); - goto out; + if (is_broadcast_ether_addr(data->ap_addr.sa_data)) { + /* the bssid we have is not to be fixed any longer, + * and we should reassociate to the best AP. */ + mac->associnfo.bssfixed = 0; + /* force reassociation */ + mac->associnfo.bssvalid = 0; + if (mac->associated) + schedule_work(&mac->associnfo.work); + } else if (is_zero_ether_addr(data->ap_addr.sa_data)) { + /* the bssid we have is no longer fixed */ + mac->associnfo.bssfixed = 0; } else { if (!memcmp(mac->associnfo.bssid, data->ap_addr.sa_data, ETH_ALEN)) { if (mac->associnfo.associating || mac->associated) { @@ -333,12 +338,14 @@ ieee80211softmac_wx_set_wap(struct net_device *net_dev, } else { /* copy new value in data->ap_addr.sa_data to bssid */ memcpy(mac->associnfo.bssid, data->ap_addr.sa_data, ETH_ALEN); - } + } + /* tell the other code that this bssid should be used no matter what */ + mac->associnfo.bssfixed = 1; /* queue associate if new bssid or (old one again and not associated) */ schedule_work(&mac->associnfo.work); } -out: + out: spin_unlock_irqrestore(&mac->lock, flags); return 0; } -- cgit v1.2.3 From d8fe3f19203b1f5070358aaa292d33295258b448 Mon Sep 17 00:00:00 2001 From: OGAWA Hirofumi Date: Mon, 24 Apr 2006 13:48:51 -0700 Subject: [SPARC]: __NR_sys removal __NR_sys_sync_file_range part was lost somewhere... [glibc is already checking __NR_sync_file_range] Signed-off-by: OGAWA Hirofumi Signed-off-by: David S. Miller --- include/asm-sparc/unistd.h | 2 +- include/asm-sparc64/unistd.h | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/include/asm-sparc/unistd.h b/include/asm-sparc/unistd.h index 45feff893b8e..32a48f623e2b 100644 --- a/include/asm-sparc/unistd.h +++ b/include/asm-sparc/unistd.h @@ -271,7 +271,7 @@ #define __NR_getsid 252 #define __NR_fdatasync 253 #define __NR_nfsservctl 254 -#define __NR_sys_sync_file_range 255 +#define __NR_sync_file_range 255 #define __NR_clock_settime 256 #define __NR_clock_gettime 257 #define __NR_clock_getres 258 diff --git a/include/asm-sparc64/unistd.h b/include/asm-sparc64/unistd.h index 597f6923a46e..ca80e8aca128 100644 --- a/include/asm-sparc64/unistd.h +++ b/include/asm-sparc64/unistd.h @@ -273,7 +273,7 @@ #define __NR_getsid 252 #define __NR_fdatasync 253 #define __NR_nfsservctl 254 -#define __NR_sys_sync_file_range 255 +#define __NR_sync_file_range 255 #define __NR_clock_settime 256 #define __NR_clock_gettime 257 #define __NR_clock_getres 258 -- cgit v1.2.3 From 55fe5866366ae42f259f27ae5962eb267d9ce172 Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Mon, 24 Apr 2006 17:16:28 -0700 Subject: [NETFILTER]: Fix compat_xt_counters alignment for non-x86 Some (?) non-x86 architectures require 8byte alignment for u_int64_t even when compiled for 32bit, using u_int32_t in compat_xt_counters breaks on these architectures, use u_int64_t for everything but x86. Reported by Andreas Schwab . Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- include/linux/netfilter/x_tables.h | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'include') diff --git a/include/linux/netfilter/x_tables.h b/include/linux/netfilter/x_tables.h index f6bdef82a322..38701454e197 100644 --- a/include/linux/netfilter/x_tables.h +++ b/include/linux/netfilter/x_tables.h @@ -361,7 +361,11 @@ struct compat_xt_entry_target struct compat_xt_counters { +#if defined(CONFIG_X86_64) || defined(CONFIG_IA64) u_int32_t cnt[4]; +#else + u_int64_t cnt[2]; +#endif }; struct compat_xt_counters_info -- cgit v1.2.3 From 912d35f86781e64d73be1ef358f703c08905ac37 Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Wed, 26 Apr 2006 10:59:21 +0200 Subject: [PATCH] Add support for the sys_vmsplice syscall sys_splice() moves data to/from pipes with a file input/output. sys_vmsplice() moves data to a pipe, with the input being a user address range instead. This uses an approach suggested by Linus, where we can hold partial ranges inside the pages[] map. Hopefully this will be useful for network receive support as well. Signed-off-by: Jens Axboe --- arch/ia64/kernel/entry.S | 1 + arch/powerpc/kernel/systbl.S | 1 + arch/powerpc/platforms/cell/spu_callbacks.c | 1 + fs/splice.c | 292 ++++++++++++++++++++++++---- include/asm-i386/unistd.h | 3 +- include/asm-ia64/unistd.h | 3 +- include/asm-powerpc/unistd.h | 3 +- include/asm-x86_64/unistd.h | 4 +- include/linux/syscalls.h | 3 + 9 files changed, 268 insertions(+), 43 deletions(-) (limited to 'include') diff --git a/arch/ia64/kernel/entry.S b/arch/ia64/kernel/entry.S index e30798811216..bcb80ca5cf40 100644 --- a/arch/ia64/kernel/entry.S +++ b/arch/ia64/kernel/entry.S @@ -1610,5 +1610,6 @@ sys_call_table: data8 sys_get_robust_list data8 sys_sync_file_range // 1300 data8 sys_tee + data8 sys_vmsplice .org sys_call_table + 8*NR_syscalls // guard against failures to increase NR_syscalls diff --git a/arch/powerpc/kernel/systbl.S b/arch/powerpc/kernel/systbl.S index 8d1522690501..0b98eea73c5e 100644 --- a/arch/powerpc/kernel/systbl.S +++ b/arch/powerpc/kernel/systbl.S @@ -324,6 +324,7 @@ COMPAT_SYS(ppoll) SYSCALL(unshare) SYSCALL(splice) SYSCALL(tee) +SYSCALL(vmsplice) /* * please add new calls to arch/powerpc/platforms/cell/spu_callbacks.c diff --git a/arch/powerpc/platforms/cell/spu_callbacks.c b/arch/powerpc/platforms/cell/spu_callbacks.c index deb3afb94484..b283380a2a18 100644 --- a/arch/powerpc/platforms/cell/spu_callbacks.c +++ b/arch/powerpc/platforms/cell/spu_callbacks.c @@ -318,6 +318,7 @@ void *spu_syscall_table[] = { [__NR_unshare] sys_unshare, [__NR_splice] sys_splice, [__NR_tee] sys_tee, + [__NR_vmsplice] sys_vmsplice, }; long spu_sys_callback(struct spu_syscall_block *s) diff --git a/fs/splice.c b/fs/splice.c index 8c6030c762e2..0b2c1f060cae 100644 --- a/fs/splice.c +++ b/fs/splice.c @@ -27,6 +27,7 @@ #include #include #include +#include /* * Passed to the actors @@ -38,6 +39,22 @@ struct splice_desc { loff_t pos; /* file position */ }; +struct partial_page { + unsigned int offset; + unsigned int len; +}; + +/* + * Passed to move_to_pipe + */ +struct splice_pipe_desc { + struct page **pages; /* page map */ + struct partial_page *partial; /* pages[] may not be contig */ + int nr_pages; /* number of pages in map */ + unsigned int flags; /* splice flags */ + struct pipe_buf_operations *ops;/* ops associated with output pipe */ +}; + /* * Attempt to steal a page from a pipe buffer. This should perhaps go into * a vm helper function, it's already simplified quite a bit by the @@ -128,6 +145,19 @@ static void page_cache_pipe_buf_unmap(struct pipe_inode_info *info, kunmap(buf->page); } +static void *user_page_pipe_buf_map(struct file *file, + struct pipe_inode_info *pipe, + struct pipe_buffer *buf) +{ + return kmap(buf->page); +} + +static void user_page_pipe_buf_unmap(struct pipe_inode_info *pipe, + struct pipe_buffer *buf) +{ + kunmap(buf->page); +} + static void page_cache_pipe_buf_get(struct pipe_inode_info *info, struct pipe_buffer *buf) { @@ -143,19 +173,33 @@ static struct pipe_buf_operations page_cache_pipe_buf_ops = { .get = page_cache_pipe_buf_get, }; +static int user_page_pipe_buf_steal(struct pipe_inode_info *pipe, + struct pipe_buffer *buf) +{ + return 1; +} + +static struct pipe_buf_operations user_page_pipe_buf_ops = { + .can_merge = 0, + .map = user_page_pipe_buf_map, + .unmap = user_page_pipe_buf_unmap, + .release = page_cache_pipe_buf_release, + .steal = user_page_pipe_buf_steal, + .get = page_cache_pipe_buf_get, +}; + /* * Pipe output worker. This sets up our pipe format with the page cache * pipe buffer operations. Otherwise very similar to the regular pipe_writev(). */ -static ssize_t move_to_pipe(struct pipe_inode_info *pipe, struct page **pages, - int nr_pages, unsigned long len, - unsigned int offset, unsigned int flags) +static ssize_t move_to_pipe(struct pipe_inode_info *pipe, + struct splice_pipe_desc *spd) { - int ret, do_wakeup, i; + int ret, do_wakeup, page_nr; ret = 0; do_wakeup = 0; - i = 0; + page_nr = 0; if (pipe->inode) mutex_lock(&pipe->inode->i_mutex); @@ -171,27 +215,19 @@ static ssize_t move_to_pipe(struct pipe_inode_info *pipe, struct page **pages, if (pipe->nrbufs < PIPE_BUFFERS) { int newbuf = (pipe->curbuf + pipe->nrbufs) & (PIPE_BUFFERS - 1); struct pipe_buffer *buf = pipe->bufs + newbuf; - struct page *page = pages[i++]; - unsigned long this_len; - this_len = PAGE_CACHE_SIZE - offset; - if (this_len > len) - this_len = len; - - buf->page = page; - buf->offset = offset; - buf->len = this_len; - buf->ops = &page_cache_pipe_buf_ops; + buf->page = spd->pages[page_nr]; + buf->offset = spd->partial[page_nr].offset; + buf->len = spd->partial[page_nr].len; + buf->ops = spd->ops; pipe->nrbufs++; + page_nr++; + ret += buf->len; + if (pipe->inode) do_wakeup = 1; - ret += this_len; - len -= this_len; - offset = 0; - if (!--nr_pages) - break; - if (!len) + if (!--spd->nr_pages) break; if (pipe->nrbufs < PIPE_BUFFERS) continue; @@ -199,7 +235,7 @@ static ssize_t move_to_pipe(struct pipe_inode_info *pipe, struct page **pages, break; } - if (flags & SPLICE_F_NONBLOCK) { + if (spd->flags & SPLICE_F_NONBLOCK) { if (!ret) ret = -EAGAIN; break; @@ -234,8 +270,8 @@ static ssize_t move_to_pipe(struct pipe_inode_info *pipe, struct page **pages, kill_fasync(&pipe->fasync_readers, SIGIO, POLL_IN); } - while (i < nr_pages) - page_cache_release(pages[i++]); + while (page_nr < spd->nr_pages) + page_cache_release(spd->pages[page_nr++]); return ret; } @@ -246,17 +282,24 @@ __generic_file_splice_read(struct file *in, loff_t *ppos, unsigned int flags) { struct address_space *mapping = in->f_mapping; - unsigned int loff, offset, nr_pages; + unsigned int loff, nr_pages; struct page *pages[PIPE_BUFFERS]; + struct partial_page partial[PIPE_BUFFERS]; struct page *page; pgoff_t index, end_index; loff_t isize; - size_t bytes; - int i, error; + size_t total_len; + int error; + struct splice_pipe_desc spd = { + .pages = pages, + .partial = partial, + .flags = flags, + .ops = &page_cache_pipe_buf_ops, + }; index = *ppos >> PAGE_CACHE_SHIFT; - loff = offset = *ppos & ~PAGE_CACHE_MASK; - nr_pages = (len + offset + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT; + loff = *ppos & ~PAGE_CACHE_MASK; + nr_pages = (len + loff + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT; if (nr_pages > PIPE_BUFFERS) nr_pages = PIPE_BUFFERS; @@ -266,15 +309,15 @@ __generic_file_splice_read(struct file *in, loff_t *ppos, * read-ahead if this is a non-zero offset (we are likely doing small * chunk splice and the page is already there) for a single page. */ - if (!offset || nr_pages > 1) - do_page_cache_readahead(mapping, in, index, nr_pages); + if (!loff || spd.nr_pages > 1) + do_page_cache_readahead(mapping, in, index, spd.nr_pages); /* * Now fill in the holes: */ error = 0; - bytes = 0; - for (i = 0; i < nr_pages; i++, index++) { + total_len = 0; + for (spd.nr_pages = 0; spd.nr_pages < nr_pages; spd.nr_pages++, index++) { unsigned int this_len; if (!len) @@ -367,26 +410,29 @@ readpage: */ if (end_index == index) { loff = PAGE_CACHE_SIZE - (isize & ~PAGE_CACHE_MASK); - if (bytes + loff > isize) { + if (total_len + loff > isize) { page_cache_release(page); break; } /* * force quit after adding this page */ - nr_pages = i; + nr_pages = spd.nr_pages; this_len = min(this_len, loff); + loff = 0; } } fill_it: - pages[i] = page; - bytes += this_len; + pages[spd.nr_pages] = page; + partial[spd.nr_pages].offset = loff; + partial[spd.nr_pages].len = this_len; len -= this_len; + total_len += this_len; loff = 0; } - if (i) - return move_to_pipe(pipe, pages, i, bytes, offset, flags); + if (spd.nr_pages) + return move_to_pipe(pipe, &spd); return error; } @@ -1018,6 +1064,174 @@ static long do_splice(struct file *in, loff_t __user *off_in, return -EINVAL; } +/* + * Map an iov into an array of pages and offset/length tupples. With the + * partial_page structure, we can map several non-contiguous ranges into + * our ones pages[] map instead of splitting that operation into pieces. + * Could easily be exported as a generic helper for other users, in which + * case one would probably want to add a 'max_nr_pages' parameter as well. + */ +static int get_iovec_page_array(const struct iovec __user *iov, + unsigned int nr_vecs, struct page **pages, + struct partial_page *partial) +{ + int buffers = 0, error = 0; + + /* + * It's ok to take the mmap_sem for reading, even + * across a "get_user()". + */ + down_read(¤t->mm->mmap_sem); + + while (nr_vecs) { + unsigned long off, npages; + void __user *base; + size_t len; + int i; + + /* + * Get user address base and length for this iovec. + */ + error = get_user(base, &iov->iov_base); + if (unlikely(error)) + break; + error = get_user(len, &iov->iov_len); + if (unlikely(error)) + break; + + /* + * Sanity check this iovec. 0 read succeeds. + */ + if (unlikely(!len)) + break; + error = -EFAULT; + if (unlikely(!base)) + break; + + /* + * Get this base offset and number of pages, then map + * in the user pages. + */ + off = (unsigned long) base & ~PAGE_MASK; + npages = (off + len + PAGE_SIZE - 1) >> PAGE_SHIFT; + if (npages > PIPE_BUFFERS - buffers) + npages = PIPE_BUFFERS - buffers; + + error = get_user_pages(current, current->mm, + (unsigned long) base, npages, 0, 0, + &pages[buffers], NULL); + + if (unlikely(error <= 0)) + break; + + /* + * Fill this contiguous range into the partial page map. + */ + for (i = 0; i < error; i++) { + const int plen = min_t(size_t, len, PAGE_SIZE) - off; + + partial[buffers].offset = off; + partial[buffers].len = plen; + + off = 0; + len -= plen; + buffers++; + } + + /* + * We didn't complete this iov, stop here since it probably + * means we have to move some of this into a pipe to + * be able to continue. + */ + if (len) + break; + + /* + * Don't continue if we mapped fewer pages than we asked for, + * or if we mapped the max number of pages that we have + * room for. + */ + if (error < npages || buffers == PIPE_BUFFERS) + break; + + nr_vecs--; + iov++; + } + + up_read(¤t->mm->mmap_sem); + + if (buffers) + return buffers; + + return error; +} + +/* + * vmsplice splices a user address range into a pipe. It can be thought of + * as splice-from-memory, where the regular splice is splice-from-file (or + * to file). In both cases the output is a pipe, naturally. + * + * Note that vmsplice only supports splicing _from_ user memory to a pipe, + * not the other way around. Splicing from user memory is a simple operation + * that can be supported without any funky alignment restrictions or nasty + * vm tricks. We simply map in the user memory and fill them into a pipe. + * The reverse isn't quite as easy, though. There are two possible solutions + * for that: + * + * - memcpy() the data internally, at which point we might as well just + * do a regular read() on the buffer anyway. + * - Lots of nasty vm tricks, that are neither fast nor flexible (it + * has restriction limitations on both ends of the pipe). + * + * Alas, it isn't here. + * + */ +static long do_vmsplice(struct file *file, const struct iovec __user *iov, + unsigned long nr_segs, unsigned int flags) +{ + struct pipe_inode_info *pipe = file->f_dentry->d_inode->i_pipe; + struct page *pages[PIPE_BUFFERS]; + struct partial_page partial[PIPE_BUFFERS]; + struct splice_pipe_desc spd = { + .pages = pages, + .partial = partial, + .flags = flags, + .ops = &user_page_pipe_buf_ops, + }; + + if (unlikely(!pipe)) + return -EBADF; + if (unlikely(nr_segs > UIO_MAXIOV)) + return -EINVAL; + else if (unlikely(!nr_segs)) + return 0; + + spd.nr_pages = get_iovec_page_array(iov, nr_segs, pages, partial); + if (spd.nr_pages <= 0) + return spd.nr_pages; + + return move_to_pipe(pipe, &spd); +} + +asmlinkage long sys_vmsplice(int fd, const struct iovec __user *iov, + unsigned long nr_segs, unsigned int flags) +{ + struct file *file; + long error; + int fput; + + error = -EBADF; + file = fget_light(fd, &fput); + if (file) { + if (file->f_mode & FMODE_WRITE) + error = do_vmsplice(file, iov, nr_segs, flags); + + fput_light(file, fput); + } + + return error; +} + asmlinkage long sys_splice(int fd_in, loff_t __user *off_in, int fd_out, loff_t __user *off_out, size_t len, unsigned int flags) diff --git a/include/asm-i386/unistd.h b/include/asm-i386/unistd.h index d81d6cfc1bb4..eb4b152c82fc 100644 --- a/include/asm-i386/unistd.h +++ b/include/asm-i386/unistd.h @@ -321,8 +321,9 @@ #define __NR_splice 313 #define __NR_sync_file_range 314 #define __NR_tee 315 +#define __NR_vmsplice 316 -#define NR_syscalls 316 +#define NR_syscalls 317 /* * user-visible error numbers are in the range -1 - -128: see diff --git a/include/asm-ia64/unistd.h b/include/asm-ia64/unistd.h index a40ebec6aeeb..7107763168bf 100644 --- a/include/asm-ia64/unistd.h +++ b/include/asm-ia64/unistd.h @@ -290,12 +290,13 @@ #define __NR_get_robust_list 1299 #define __NR_sync_file_range 1300 #define __NR_tee 1301 +#define __NR_vmsplice 1302 #ifdef __KERNEL__ #include -#define NR_syscalls 278 /* length of syscall table */ +#define NR_syscalls 279 /* length of syscall table */ #define __ARCH_WANT_SYS_RT_SIGACTION diff --git a/include/asm-powerpc/unistd.h b/include/asm-powerpc/unistd.h index c612f1a62772..34325e292596 100644 --- a/include/asm-powerpc/unistd.h +++ b/include/asm-powerpc/unistd.h @@ -303,8 +303,9 @@ #define __NR_unshare 282 #define __NR_splice 283 #define __NR_tee 284 +#define __NR_vmsplice 285 -#define __NR_syscalls 285 +#define __NR_syscalls 286 #ifdef __KERNEL__ #define __NR__exit __NR_exit diff --git a/include/asm-x86_64/unistd.h b/include/asm-x86_64/unistd.h index 98c36eae567c..feb77cb8c044 100644 --- a/include/asm-x86_64/unistd.h +++ b/include/asm-x86_64/unistd.h @@ -615,8 +615,10 @@ __SYSCALL(__NR_splice, sys_splice) __SYSCALL(__NR_tee, sys_tee) #define __NR_sync_file_range 277 __SYSCALL(__NR_sync_file_range, sys_sync_file_range) +#define __NR_vmsplice 278 +__SYSCALL(__NR_vmsplice, sys_vmsplice) -#define __NR_syscall_max __NR_sync_file_range +#define __NR_syscall_max __NR_vmsplice #ifndef __NO_STUBS diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h index d3ebc0e68b2b..3996960fc565 100644 --- a/include/linux/syscalls.h +++ b/include/linux/syscalls.h @@ -574,6 +574,9 @@ asmlinkage long sys_splice(int fd_in, loff_t __user *off_in, int fd_out, loff_t __user *off_out, size_t len, unsigned int flags); +asmlinkage long sys_vmsplice(int fd, const struct iovec __user *iov, + unsigned long nr_segs, unsigned int flags); + asmlinkage long sys_tee(int fdin, int fdout, size_t len, unsigned int flags); asmlinkage long sys_sync_file_range(int fd, loff_t offset, loff_t nbytes, -- cgit v1.2.3 From 734cbc363b159caee158d5a83408c72d98bcacf0 Mon Sep 17 00:00:00 2001 From: Stephen Hemminger Date: Tue, 25 Apr 2006 10:58:50 -0700 Subject: [PATCH] sky2: reschedule if irq still pending This is a workaround for the case edge-triggered irq's. Several users seem to have broken configurations sharing edge-triggered irq's. To avoid losing IRQ's, reshedule if more work arrives. The changes to netdevice.h are to extract the part that puts device back in list into separate inline. Signed-off-by: Stephen Hemminger Signed-off-by: Jeff Garzik --- drivers/net/sky2.c | 20 ++++++++++++++++---- include/linux/netdevice.h | 18 ++++++++++-------- 2 files changed, 26 insertions(+), 12 deletions(-) (limited to 'include') diff --git a/drivers/net/sky2.c b/drivers/net/sky2.c index 67b0eab16589..618fde8622ca 100644 --- a/drivers/net/sky2.c +++ b/drivers/net/sky2.c @@ -2093,6 +2093,7 @@ static int sky2_poll(struct net_device *dev0, int *budget) int work_done = 0; u32 status = sky2_read32(hw, B0_Y2_SP_EISR); + restart_poll: if (unlikely(status & ~Y2_IS_STAT_BMU)) { if (status & Y2_IS_HW_ERR) sky2_hw_intr(hw); @@ -2123,7 +2124,7 @@ static int sky2_poll(struct net_device *dev0, int *budget) } if (status & Y2_IS_STAT_BMU) { - work_done = sky2_status_intr(hw, work_limit); + work_done += sky2_status_intr(hw, work_limit - work_done); *budget -= work_done; dev0->quota -= work_done; @@ -2133,9 +2134,22 @@ static int sky2_poll(struct net_device *dev0, int *budget) sky2_write32(hw, STAT_CTRL, SC_STAT_CLR_IRQ); } - netif_rx_complete(dev0); + local_irq_disable(); + __netif_rx_complete(dev0); status = sky2_read32(hw, B0_Y2_SP_LISR); + + if (unlikely(status)) { + /* More work pending, try and keep going */ + if (__netif_rx_schedule_prep(dev0)) { + __netif_rx_reschedule(dev0, work_done); + status = sky2_read32(hw, B0_Y2_SP_EISR); + local_irq_enable(); + goto restart_poll; + } + } + + local_irq_enable(); return 0; } @@ -2153,8 +2167,6 @@ static irqreturn_t sky2_intr(int irq, void *dev_id, struct pt_regs *regs) prefetch(&hw->st_le[hw->st_idx]); if (likely(__netif_rx_schedule_prep(dev0))) __netif_rx_schedule(dev0); - else - printk(KERN_DEBUG PFX "irq race detected\n"); return IRQ_HANDLED; } diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 40ccf8cc4239..01db7b88a2b1 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -829,19 +829,21 @@ static inline void netif_rx_schedule(struct net_device *dev) __netif_rx_schedule(dev); } -/* Try to reschedule poll. Called by dev->poll() after netif_rx_complete(). - * Do not inline this? - */ + +static inline void __netif_rx_reschedule(struct net_device *dev, int undo) +{ + dev->quota += undo; + list_add_tail(&dev->poll_list, &__get_cpu_var(softnet_data).poll_list); + __raise_softirq_irqoff(NET_RX_SOFTIRQ); +} + +/* Try to reschedule poll. Called by dev->poll() after netif_rx_complete(). */ static inline int netif_rx_reschedule(struct net_device *dev, int undo) { if (netif_rx_schedule_prep(dev)) { unsigned long flags; - - dev->quota += undo; - local_irq_save(flags); - list_add_tail(&dev->poll_list, &__get_cpu_var(softnet_data).poll_list); - __raise_softirq_irqoff(NET_RX_SOFTIRQ); + __netif_rx_reschedule(dev, undo); local_irq_restore(flags); return 1; } -- cgit v1.2.3 From 00522fb41a2a9bf0f98a007c0e2b516a3873148c Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Wed, 26 Apr 2006 14:39:29 +0200 Subject: [PATCH] splice: rearrange moving to/from pipe helpers We need these for people writing their own ->splice_read/write hooks. Signed-off-by: Jens Axboe --- fs/splice.c | 35 +++++++++++------------------------ include/linux/pipe_fs_i.h | 17 +++++++++++++++++ 2 files changed, 28 insertions(+), 24 deletions(-) (limited to 'include') diff --git a/fs/splice.c b/fs/splice.c index 0b2c1f060cae..447ebc0a37f3 100644 --- a/fs/splice.c +++ b/fs/splice.c @@ -29,23 +29,13 @@ #include #include -/* - * Passed to the actors - */ -struct splice_desc { - unsigned int len, total_len; /* current and remaining length */ - unsigned int flags; /* splice flags */ - struct file *file; /* file to read/write */ - loff_t pos; /* file position */ -}; - struct partial_page { unsigned int offset; unsigned int len; }; /* - * Passed to move_to_pipe + * Passed to splice_to_pipe */ struct splice_pipe_desc { struct page **pages; /* page map */ @@ -192,8 +182,8 @@ static struct pipe_buf_operations user_page_pipe_buf_ops = { * Pipe output worker. This sets up our pipe format with the page cache * pipe buffer operations. Otherwise very similar to the regular pipe_writev(). */ -static ssize_t move_to_pipe(struct pipe_inode_info *pipe, - struct splice_pipe_desc *spd) +static ssize_t splice_to_pipe(struct pipe_inode_info *pipe, + struct splice_pipe_desc *spd) { int ret, do_wakeup, page_nr; @@ -432,7 +422,7 @@ fill_it: } if (spd.nr_pages) - return move_to_pipe(pipe, &spd); + return splice_to_pipe(pipe, &spd); return error; } @@ -666,17 +656,14 @@ out_nomem: return ret; } -typedef int (splice_actor)(struct pipe_inode_info *, struct pipe_buffer *, - struct splice_desc *); - /* * Pipe input worker. Most of this logic works like a regular pipe, the * key here is the 'actor' worker passed in that actually moves the data * to the wanted destination. See pipe_to_file/pipe_to_sendpage above. */ -static ssize_t move_from_pipe(struct pipe_inode_info *pipe, struct file *out, - loff_t *ppos, size_t len, unsigned int flags, - splice_actor *actor) +ssize_t splice_from_pipe(struct pipe_inode_info *pipe, struct file *out, + loff_t *ppos, size_t len, unsigned int flags, + splice_actor *actor) { int ret, do_wakeup, err; struct splice_desc sd; @@ -795,7 +782,7 @@ generic_file_splice_write(struct pipe_inode_info *pipe, struct file *out, struct address_space *mapping = out->f_mapping; ssize_t ret; - ret = move_from_pipe(pipe, out, ppos, len, flags, pipe_to_file); + ret = splice_from_pipe(pipe, out, ppos, len, flags, pipe_to_file); if (ret > 0) { struct inode *inode = mapping->host; @@ -837,7 +824,7 @@ EXPORT_SYMBOL(generic_file_splice_write); ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe, struct file *out, loff_t *ppos, size_t len, unsigned int flags) { - return move_from_pipe(pipe, out, ppos, len, flags, pipe_to_sendpage); + return splice_from_pipe(pipe, out, ppos, len, flags, pipe_to_sendpage); } EXPORT_SYMBOL(generic_splice_sendpage); @@ -924,7 +911,7 @@ long do_splice_direct(struct file *in, loff_t *ppos, struct file *out, /* * We don't have an immediate reader, but we'll read the stuff - * out of the pipe right after the move_to_pipe(). So set + * out of the pipe right after the splice_to_pipe(). So set * PIPE_READERS appropriately. */ pipe->readers = 1; @@ -1210,7 +1197,7 @@ static long do_vmsplice(struct file *file, const struct iovec __user *iov, if (spd.nr_pages <= 0) return spd.nr_pages; - return move_to_pipe(pipe, &spd); + return splice_to_pipe(pipe, &spd); } asmlinkage long sys_vmsplice(int fd, const struct iovec __user *iov, diff --git a/include/linux/pipe_fs_i.h b/include/linux/pipe_fs_i.h index ef7f33c0be19..0008d4bd4059 100644 --- a/include/linux/pipe_fs_i.h +++ b/include/linux/pipe_fs_i.h @@ -61,4 +61,21 @@ void __free_pipe_info(struct pipe_inode_info *); /* from/to, of course */ #define SPLICE_F_MORE (0x04) /* expect more data */ +/* + * Passed to the actors + */ +struct splice_desc { + unsigned int len, total_len; /* current and remaining length */ + unsigned int flags; /* splice flags */ + struct file *file; /* file to read/write */ + loff_t pos; /* file position */ +}; + +typedef int (splice_actor)(struct pipe_inode_info *, struct pipe_buffer *, + struct splice_desc *); + +extern ssize_t splice_from_pipe(struct pipe_inode_info *, struct file *, + loff_t *, size_t, unsigned int, + splice_actor *); + #endif -- cgit v1.2.3 From ebf43500ef148a380bd132743c3fc530111ac620 Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Thu, 27 Apr 2006 08:46:01 +0200 Subject: [PATCH] Add find_get_pages_contig(): contiguous variant of find_get_pages() find_get_pages_contig() will break out if we hit a hole in the page cache. From Andrew Morton, small modifications and documentation by me. Signed-off-by: Jens Axboe --- include/linux/pagemap.h | 2 ++ mm/filemap.c | 32 ++++++++++++++++++++++++++++++++ 2 files changed, 34 insertions(+) (limited to 'include') diff --git a/include/linux/pagemap.h b/include/linux/pagemap.h index 9539efd4f7e6..7a1af574dedf 100644 --- a/include/linux/pagemap.h +++ b/include/linux/pagemap.h @@ -78,6 +78,8 @@ extern struct page * find_or_create_page(struct address_space *mapping, unsigned long index, gfp_t gfp_mask); unsigned find_get_pages(struct address_space *mapping, pgoff_t start, unsigned int nr_pages, struct page **pages); +unsigned find_get_pages_contig(struct address_space *mapping, pgoff_t start, + unsigned int nr_pages, struct page **pages); unsigned find_get_pages_tag(struct address_space *mapping, pgoff_t *index, int tag, unsigned int nr_pages, struct page **pages); diff --git a/mm/filemap.c b/mm/filemap.c index 3ef20739e725..fd57442186cb 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -697,6 +697,38 @@ unsigned find_get_pages(struct address_space *mapping, pgoff_t start, return ret; } +/** + * find_get_pages_contig - gang contiguous pagecache lookup + * @mapping: The address_space to search + * @index: The starting page index + * @nr_pages: The maximum number of pages + * @pages: Where the resulting pages are placed + * + * find_get_pages_contig() works exactly like find_get_pages(), except + * that the returned number of pages are guaranteed to be contiguous. + * + * find_get_pages_contig() returns the number of pages which were found. + */ +unsigned find_get_pages_contig(struct address_space *mapping, pgoff_t index, + unsigned int nr_pages, struct page **pages) +{ + unsigned int i; + unsigned int ret; + + read_lock_irq(&mapping->tree_lock); + ret = radix_tree_gang_lookup(&mapping->page_tree, + (void **)pages, index, nr_pages); + for (i = 0; i < ret; i++) { + if (pages[i]->mapping == NULL || pages[i]->index != index) + break; + + page_cache_get(pages[i]); + index++; + } + read_unlock_irq(&mapping->tree_lock); + return i; +} + /* * Like find_get_pages, except we only return pages which are tagged with * `tag'. We update *index to index the next page for the traversal. -- cgit v1.2.3 From bc818247203a7bfc40296a3f5b760de84fb8e0d1 Mon Sep 17 00:00:00 2001 From: Atsushi Nemoto Date: Mon, 17 Apr 2006 21:19:12 +0900 Subject: [MIPS] Fix bitops for MIPS32/MIPS64 CPUs. With recent rewrite for generic bitops, fls() for 32bit kernel with MIPS64_CPU is broken. Also, ffs(), fls() should be defined the same way as the libc and compiler built-in routines (returns int instead of unsigned long). Signed-off-by: Atsushi Nemoto Signed-off-by: Ralf Baechle --- include/asm-mips/bitops.h | 56 ++++++++++++++++++++--------------------------- 1 file changed, 24 insertions(+), 32 deletions(-) (limited to 'include') diff --git a/include/asm-mips/bitops.h b/include/asm-mips/bitops.h index a1728f8c0705..d2f444537e4b 100644 --- a/include/asm-mips/bitops.h +++ b/include/asm-mips/bitops.h @@ -467,64 +467,56 @@ static inline unsigned long __ffs(unsigned long word) } /* - * ffs - find first bit set. + * fls - find last bit set. * @word: The word to search * - * Returns 1..SZLONG - * Returns 0 if no bit exists + * This is defined the same way as ffs. + * Note fls(0) = 0, fls(1) = 1, fls(0x80000000) = 32. */ - -static inline unsigned long ffs(unsigned long word) +static inline int fls(int word) { - if (!word) - return 0; + __asm__ ("clz %0, %1" : "=r" (word) : "r" (word)); - return __ffs(word) + 1; + return 32 - word; } -/* - * ffz - find first zero in word. - * @word: The word to search - * - * Undefined if no zero exists, so code should check against ~0UL first. - */ -static inline unsigned long ffz(unsigned long word) +#if defined(CONFIG_64BIT) && defined(CONFIG_CPU_MIPS64) +static inline int fls64(__u64 word) { - return __ffs (~word); + __asm__ ("dclz %0, %1" : "=r" (word) : "r" (word)); + + return 64 - word; } +#else +#include +#endif /* - * fls - find last bit set. + * ffs - find first bit set. * @word: The word to search * - * Returns 1..SZLONG - * Returns 0 if no bit exists + * This is defined the same way as + * the libc and compiler builtin ffs routines, therefore + * differs in spirit from the above ffz (man ffs). */ -static inline unsigned long fls(unsigned long word) +static inline int ffs(int word) { -#ifdef CONFIG_CPU_MIPS32 - __asm__ ("clz %0, %1" : "=r" (word) : "r" (word)); - - return 32 - word; -#endif - -#ifdef CONFIG_CPU_MIPS64 - __asm__ ("dclz %0, %1" : "=r" (word) : "r" (word)); + if (!word) + return 0; - return 64 - word; -#endif + return fls(word & -word); } #else #include #include -#include #include +#include #endif /*defined(CONFIG_CPU_MIPS32) || defined(CONFIG_CPU_MIPS64) */ -#include +#include #include #ifdef __KERNEL__ -- cgit v1.2.3 From 7a8341969fe0df4a1fffa141435e742456270ffd Mon Sep 17 00:00:00 2001 From: Chris Dearman Date: Sat, 15 Apr 2006 00:31:16 +0100 Subject: [MIPS] 24K LV: Add core card id. Signed-off-by: Ralf Baechle --- arch/mips/mips-boards/generic/init.c | 1 + arch/mips/mips-boards/generic/pci.c | 1 + arch/mips/mips-boards/malta/malta_int.c | 3 +++ include/asm-mips/mips-boards/generic.h | 1 + 4 files changed, 6 insertions(+) (limited to 'include') diff --git a/arch/mips/mips-boards/generic/init.c b/arch/mips/mips-boards/generic/init.c index 17dfe6a8cab9..df4e94735604 100644 --- a/arch/mips/mips-boards/generic/init.c +++ b/arch/mips/mips-boards/generic/init.c @@ -337,6 +337,7 @@ void __init prom_init(void) case MIPS_REVISION_CORID_CORE_MSC: case MIPS_REVISION_CORID_CORE_FPGA2: case MIPS_REVISION_CORID_CORE_FPGA3: + case MIPS_REVISION_CORID_CORE_24K: case MIPS_REVISION_CORID_CORE_EMUL_MSC: _pcictrl_msc = (unsigned long)ioremap(MIPS_MSC01_PCI_REG_BASE, 0x2000); diff --git a/arch/mips/mips-boards/generic/pci.c b/arch/mips/mips-boards/generic/pci.c index 1f6f9df74ab2..9337f6c8873a 100644 --- a/arch/mips/mips-boards/generic/pci.c +++ b/arch/mips/mips-boards/generic/pci.c @@ -198,6 +198,7 @@ void __init mips_pcibios_init(void) case MIPS_REVISION_CORID_CORE_MSC: case MIPS_REVISION_CORID_CORE_FPGA2: case MIPS_REVISION_CORID_CORE_FPGA3: + case MIPS_REVISION_CORID_CORE_24K: case MIPS_REVISION_CORID_CORE_EMUL_MSC: /* Set up resource ranges from the controller's registers. */ MSC_READ(MSC01_PCI_SC2PMBASL, start); diff --git a/arch/mips/mips-boards/malta/malta_int.c b/arch/mips/mips-boards/malta/malta_int.c index 64db07d4dbe5..7cc0ba4f553a 100644 --- a/arch/mips/mips-boards/malta/malta_int.c +++ b/arch/mips/mips-boards/malta/malta_int.c @@ -57,6 +57,7 @@ static inline int mips_pcibios_iack(void) case MIPS_REVISION_CORID_CORE_MSC: case MIPS_REVISION_CORID_CORE_FPGA2: case MIPS_REVISION_CORID_CORE_FPGA3: + case MIPS_REVISION_CORID_CORE_24K: case MIPS_REVISION_CORID_CORE_EMUL_MSC: MSC_READ(MSC01_PCI_IACK, irq); irq &= 0xff; @@ -143,6 +144,7 @@ void corehi_irqdispatch(struct pt_regs *regs) case MIPS_REVISION_CORID_CORE_MSC: case MIPS_REVISION_CORID_CORE_FPGA2: case MIPS_REVISION_CORID_CORE_FPGA3: + case MIPS_REVISION_CORID_CORE_24K: case MIPS_REVISION_CORID_CORE_EMUL_MSC: ll_msc_irq(regs); break; @@ -309,6 +311,7 @@ void __init arch_init_irq(void) case MIPS_REVISION_CORID_CORE_MSC: case MIPS_REVISION_CORID_CORE_FPGA2: case MIPS_REVISION_CORID_CORE_FPGA3: + case MIPS_REVISION_CORID_CORE_24K: case MIPS_REVISION_CORID_CORE_EMUL_MSC: if (cpu_has_veic) init_msc_irqs (MSC01E_INT_BASE, msc_eicirqmap, msc_nr_eicirqs); diff --git a/include/asm-mips/mips-boards/generic.h b/include/asm-mips/mips-boards/generic.h index 25b6ffc26623..fa8b913cc3e0 100644 --- a/include/asm-mips/mips-boards/generic.h +++ b/include/asm-mips/mips-boards/generic.h @@ -67,6 +67,7 @@ #define MIPS_REVISION_CORID_CORE_FPGA2 7 #define MIPS_REVISION_CORID_CORE_FPGAR2 8 #define MIPS_REVISION_CORID_CORE_FPGA3 9 +#define MIPS_REVISION_CORID_CORE_24K 10 /**** Artificial corid defines ****/ /* -- cgit v1.2.3 From 6e5882cfa24e1456702e463f6920fc0ca3c3d2b8 Mon Sep 17 00:00:00 2001 From: Zachary Amsden Date: Thu, 27 Apr 2006 11:32:29 -0700 Subject: [PATCH] x86/PAE: Fix pte_clear for the >4GB RAM case Proposed fix for ptep_get_and_clear_full PAE bug. Pte_clear had the same bug, so use the same fix for both. Turns out pmd_clear had it as well, but pgds are not affected. The problem is rather intricate. Page table entries in PAE mode are 64-bits wide, but the only atomic 8-byte write operation available in 32-bit mode is cmpxchg8b, which is expensive (at least on P4), and thus avoided. But it can happen that the processor may prefetch entries into the TLB in the middle of an operation which clears a page table entry. So one must always clear the P-bit in the low word of the page table entry first when clearing it. Since the sequence *ptep = __pte(0) leaves the order of the write dependent on the compiler, it must be coded explicitly as a clear of the low word followed by a clear of the high word. Further, there must be a write memory barrier here to enforce proper ordering by the compiler (and, in the future, by the processor as well). On > 4GB memory machines, the implementation of pte_clear for PAE was clearly deficient, as it could leave virtual mappings of physical memory above 4GB aliased to memory below 4GB in the TLB. The implementation of ptep_get_and_clear_full has a similar bug, although not nearly as likely to occur, since the mappings being cleared are in the process of being destroyed, and should never be dereferenced again. But, as luck would have it, it is possible to trigger bugs even without ever dereferencing these bogus TLB mappings, even if the clear is followed fairly soon after with a TLB flush or invalidation. The problem is that memory above 4GB may now be aliased into the first 4GB of memory, and in fact, may hit a region of memory with non-memory semantics. These regions include AGP and PCI space. As such, these memory regions are not cached by the processor. This introduces the bug. The processor can speculate memory operations, including memory writes, as long as they are committed with the proper ordering. Speculating a memory write to a linear address that has a bogus TLB mapping is possible. Normally, the speculation is harmless. But for cached memory, it does leave the falsely speculated cacheline unmodified, but in a dirty state. This cache line will be eventually written back. If this cacheline happens to intersect a region of memory that is not protected by the cache coherency protocol, it can corrupt data in I/O memory, which is generally a very bad thing to do, and can cause total system failure or just plain undefined behavior. These bugs are extremely unlikely, but the severity is of such magnitude, and the fix so simple that I think fixing them immediately is justified. Also, they are nearly impossible to debug. Signed-off-by: Zachary Amsden Signed-off-by: Linus Torvalds --- include/asm-i386/pgtable-2level.h | 3 +++ include/asm-i386/pgtable-3level.h | 20 ++++++++++++++++++++ include/asm-i386/pgtable.h | 4 +--- 3 files changed, 24 insertions(+), 3 deletions(-) (limited to 'include') diff --git a/include/asm-i386/pgtable-2level.h b/include/asm-i386/pgtable-2level.h index 27bde973abc7..2756d4b04c27 100644 --- a/include/asm-i386/pgtable-2level.h +++ b/include/asm-i386/pgtable-2level.h @@ -18,6 +18,9 @@ #define set_pte_atomic(pteptr, pteval) set_pte(pteptr,pteval) #define set_pmd(pmdptr, pmdval) (*(pmdptr) = (pmdval)) +#define pte_clear(mm,addr,xp) do { set_pte_at(mm, addr, xp, __pte(0)); } while (0) +#define pmd_clear(xp) do { set_pmd(xp, __pmd(0)); } while (0) + #define ptep_get_and_clear(mm,addr,xp) __pte(xchg(&(xp)->pte_low, 0)) #define pte_same(a, b) ((a).pte_low == (b).pte_low) #define pte_page(x) pfn_to_page(pte_pfn(x)) diff --git a/include/asm-i386/pgtable-3level.h b/include/asm-i386/pgtable-3level.h index 36a5aa63cbbf..dccb1b3337ad 100644 --- a/include/asm-i386/pgtable-3level.h +++ b/include/asm-i386/pgtable-3level.h @@ -85,6 +85,26 @@ static inline void pud_clear (pud_t * pud) { } #define pmd_offset(pud, address) ((pmd_t *) pud_page(*(pud)) + \ pmd_index(address)) +/* + * For PTEs and PDEs, we must clear the P-bit first when clearing a page table + * entry, so clear the bottom half first and enforce ordering with a compiler + * barrier. + */ +static inline void pte_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep) +{ + ptep->pte_low = 0; + smp_wmb(); + ptep->pte_high = 0; +} + +static inline void pmd_clear(pmd_t *pmd) +{ + u32 *tmp = (u32 *)pmd; + *tmp = 0; + smp_wmb(); + *(tmp + 1) = 0; +} + static inline pte_t ptep_get_and_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep) { pte_t res; diff --git a/include/asm-i386/pgtable.h b/include/asm-i386/pgtable.h index ee056c41a9fb..672c3f76b9df 100644 --- a/include/asm-i386/pgtable.h +++ b/include/asm-i386/pgtable.h @@ -204,12 +204,10 @@ extern unsigned long long __PAGE_KERNEL, __PAGE_KERNEL_EXEC; extern unsigned long pg0[]; #define pte_present(x) ((x).pte_low & (_PAGE_PRESENT | _PAGE_PROTNONE)) -#define pte_clear(mm,addr,xp) do { set_pte_at(mm, addr, xp, __pte(0)); } while (0) /* To avoid harmful races, pmd_none(x) should check only the lower when PAE */ #define pmd_none(x) (!(unsigned long)pmd_val(x)) #define pmd_present(x) (pmd_val(x) & _PAGE_PRESENT) -#define pmd_clear(xp) do { set_pmd(xp, __pmd(0)); } while (0) #define pmd_bad(x) ((pmd_val(x) & (~PAGE_MASK & ~_PAGE_USER)) != _KERNPG_TABLE) @@ -268,7 +266,7 @@ static inline pte_t ptep_get_and_clear_full(struct mm_struct *mm, unsigned long pte_t pte; if (full) { pte = *ptep; - *ptep = __pte(0); + pte_clear(mm, addr, ptep); } else { pte = ptep_get_and_clear(mm, addr, ptep); } -- cgit v1.2.3 From 4d17ffda331ba6030bb8c233c73d6a87954d8ea7 Mon Sep 17 00:00:00 2001 From: Kay Sievers Date: Tue, 25 Apr 2006 15:37:26 +0200 Subject: [PATCH] Kobject: fix build error This fixes a build error for various odd combinations of CONFIG_HOTPLUG and CONFIG_NET. Signed-off-by: Kay Sievers Cc: Nigel Cunningham Cc: Andrew Morton Signed-off-by: Greg Kroah-Hartman --- include/linux/kobject.h | 2 +- lib/kobject_uevent.c | 8 +++++++- 2 files changed, 8 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/include/linux/kobject.h b/include/linux/kobject.h index dcd0623be892..e38bb357282a 100644 --- a/include/linux/kobject.h +++ b/include/linux/kobject.h @@ -259,7 +259,7 @@ struct subsys_attribute { extern int subsys_create_file(struct subsystem * , struct subsys_attribute *); extern void subsys_remove_file(struct subsystem * , struct subsys_attribute *); -#if defined(CONFIG_HOTPLUG) && defined(CONFIG_NET) +#if defined(CONFIG_HOTPLUG) void kobject_uevent(struct kobject *kobj, enum kobject_action action); int add_uevent_var(char **envp, int num_envp, int *cur_index, diff --git a/lib/kobject_uevent.c b/lib/kobject_uevent.c index 982226daf939..7f20e7b857cb 100644 --- a/lib/kobject_uevent.c +++ b/lib/kobject_uevent.c @@ -25,11 +25,13 @@ #define BUFFER_SIZE 2048 /* buffer for the variables */ #define NUM_ENVP 32 /* number of env pointers */ -#if defined(CONFIG_HOTPLUG) && defined(CONFIG_NET) +#if defined(CONFIG_HOTPLUG) u64 uevent_seqnum; char uevent_helper[UEVENT_HELPER_PATH_LEN] = "/sbin/hotplug"; static DEFINE_SPINLOCK(sequence_lock); +#if defined(CONFIG_NET) static struct sock *uevent_sock; +#endif static char *action_to_string(enum kobject_action action) { @@ -155,6 +157,7 @@ void kobject_uevent(struct kobject *kobj, enum kobject_action action) spin_unlock(&sequence_lock); sprintf(seq_buff, "SEQNUM=%llu", (unsigned long long)seq); +#if defined(CONFIG_NET) /* send netlink message */ if (uevent_sock) { struct sk_buff *skb; @@ -179,6 +182,7 @@ void kobject_uevent(struct kobject *kobj, enum kobject_action action) netlink_broadcast(uevent_sock, skb, 0, 1, GFP_KERNEL); } } +#endif /* call uevent_helper, usually only enabled during early boot */ if (uevent_helper[0]) { @@ -249,6 +253,7 @@ int add_uevent_var(char **envp, int num_envp, int *cur_index, } EXPORT_SYMBOL_GPL(add_uevent_var); +#if defined(CONFIG_NET) static int __init kobject_uevent_init(void) { uevent_sock = netlink_kernel_create(NETLINK_KOBJECT_UEVENT, 1, NULL, @@ -264,5 +269,6 @@ static int __init kobject_uevent_init(void) } postcore_initcall(kobject_uevent_init); +#endif #endif /* CONFIG_HOTPLUG */ -- cgit v1.2.3 From bde11d794206ae8d72defd0e8a481181200f7dc4 Mon Sep 17 00:00:00 2001 From: Jean Delvare Date: Tue, 18 Apr 2006 21:30:22 -0700 Subject: [PATCH] Fix OCFS2 warning when DEBUG_FS is not enabled Fix the following warning which happens when OCFS2_FS is enabled but DEBUG_FS isn't: fs/ocfs2/dlmglue.c: In function `ocfs2_dlm_init_debug': fs/ocfs2/dlmglue.c:2036: warning: passing arg 5 of `debugfs_create_file' discards qualifiers from pointer target type Signed-off-by: Jean Delvare Cc: Arjan van de Ven Cc: Joel Becker Acked-by: Mark Fasheh Signed-off-by: Andrew Morton Signed-off-by: Greg Kroah-Hartman --- include/linux/debugfs.h | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'include') diff --git a/include/linux/debugfs.h b/include/linux/debugfs.h index 176e2d371577..047567d34ca7 100644 --- a/include/linux/debugfs.h +++ b/include/linux/debugfs.h @@ -58,9 +58,8 @@ struct dentry *debugfs_create_blob(const char *name, mode_t mode, */ static inline struct dentry *debugfs_create_file(const char *name, mode_t mode, - struct dentry *parent, - void *data, - struct file_operations *fops) + struct dentry *parent, void *data, + const struct file_operations *fops) { return ERR_PTR(-ENODEV); } -- cgit v1.2.3 From 5b3ef14e3e9d745a512d65fcb4ef9be541226d80 Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Sat, 22 Apr 2006 12:14:44 +0200 Subject: [PATCH] Kobject: possible cleanups This patch contains the following possible cleanups: - #if 0 the following unused global function: - subsys_remove_file() - remove the following unused EXPORT_SYMBOL's: - kset_find_obj - subsystem_init - remove the following unused EXPORT_SYMBOL_GPL: - kobject_add_dir Signed-off-by: Adrian Bunk Signed-off-by: Greg Kroah-Hartman --- include/linux/kobject.h | 1 - lib/kobject.c | 7 ++----- 2 files changed, 2 insertions(+), 6 deletions(-) (limited to 'include') diff --git a/include/linux/kobject.h b/include/linux/kobject.h index e38bb357282a..c187c53cecd0 100644 --- a/include/linux/kobject.h +++ b/include/linux/kobject.h @@ -257,7 +257,6 @@ struct subsys_attribute { }; extern int subsys_create_file(struct subsystem * , struct subsys_attribute *); -extern void subsys_remove_file(struct subsystem * , struct subsys_attribute *); #if defined(CONFIG_HOTPLUG) void kobject_uevent(struct kobject *kobj, enum kobject_action action); diff --git a/lib/kobject.c b/lib/kobject.c index 01d957513940..b46350c27837 100644 --- a/lib/kobject.c +++ b/lib/kobject.c @@ -422,7 +422,6 @@ struct kobject *kobject_add_dir(struct kobject *parent, const char *name) return k; } -EXPORT_SYMBOL_GPL(kobject_add_dir); /** * kset_init - initialize a kset for use @@ -569,7 +568,7 @@ int subsys_create_file(struct subsystem * s, struct subsys_attribute * a) * @s: subsystem. * @a: attribute desciptor. */ - +#if 0 void subsys_remove_file(struct subsystem * s, struct subsys_attribute * a) { if (subsys_get(s)) { @@ -577,6 +576,7 @@ void subsys_remove_file(struct subsystem * s, struct subsys_attribute * a) subsys_put(s); } } +#endif /* 0 */ EXPORT_SYMBOL(kobject_init); EXPORT_SYMBOL(kobject_register); @@ -588,10 +588,7 @@ EXPORT_SYMBOL(kobject_del); EXPORT_SYMBOL(kset_register); EXPORT_SYMBOL(kset_unregister); -EXPORT_SYMBOL(kset_find_obj); -EXPORT_SYMBOL(subsystem_init); EXPORT_SYMBOL(subsystem_register); EXPORT_SYMBOL(subsystem_unregister); EXPORT_SYMBOL(subsys_create_file); -EXPORT_SYMBOL(subsys_remove_file); -- cgit v1.2.3 From f0fe253c4719faf76d40f581cdc0e8aef77273bb Mon Sep 17 00:00:00 2001 From: Jack Steiner Date: Sat, 22 Apr 2006 09:36:07 -0500 Subject: [IA64-SGI] - Fix discover of nearest cpu node to IO node Fix a bug that causes discovery of the nearest node/cpu to a TIO (IO node) to fail. Signed-off-by: Jack Steiner Signed-off-by: Tony Luck --- arch/ia64/sn/kernel/sn2/sn_hwperf.c | 4 ++-- include/asm-ia64/sn/sn2/sn_hwperf.h | 6 +++++- 2 files changed, 7 insertions(+), 3 deletions(-) (limited to 'include') diff --git a/arch/ia64/sn/kernel/sn2/sn_hwperf.c b/arch/ia64/sn/kernel/sn2/sn_hwperf.c index d917afa30b27..7ec65bc0ccf6 100644 --- a/arch/ia64/sn/kernel/sn2/sn_hwperf.c +++ b/arch/ia64/sn/kernel/sn2/sn_hwperf.c @@ -284,6 +284,8 @@ static int sn_hwperf_get_nearest_node_objdata(struct sn_hwperf_object_info *objb /* find nearest node with cpus and nearest memory */ for (router=NULL, j=0; j < op->ports; j++) { dest = sn_hwperf_findobj_id(objbuf, nobj, ptdata[j].conn_id); + if (dest && SN_HWPERF_IS_ROUTER(dest)) + router = dest; if (!dest || SN_HWPERF_FOREIGN(dest) || !SN_HWPERF_IS_NODE(dest) || SN_HWPERF_IS_IONODE(dest)) { continue; @@ -299,8 +301,6 @@ static int sn_hwperf_get_nearest_node_objdata(struct sn_hwperf_object_info *objb *near_mem_node = c; found_mem++; } - if (SN_HWPERF_IS_ROUTER(dest)) - router = dest; } if (router && (!found_cpu || !found_mem)) { diff --git a/include/asm-ia64/sn/sn2/sn_hwperf.h b/include/asm-ia64/sn/sn2/sn_hwperf.h index 291ef3d69da2..e61ebac38cdd 100644 --- a/include/asm-ia64/sn/sn2/sn_hwperf.h +++ b/include/asm-ia64/sn/sn2/sn_hwperf.h @@ -45,8 +45,12 @@ struct sn_hwperf_object_info { #define SN_HWPERF_IS_NODE(x) ((x) && strstr((x)->name, "SHub")) #define SN_HWPERF_IS_NODE_SHUB2(x) ((x) && strstr((x)->name, "SHub 2.")) #define SN_HWPERF_IS_IONODE(x) ((x) && strstr((x)->name, "TIO")) -#define SN_HWPERF_IS_ROUTER(x) ((x) && strstr((x)->name, "Router")) #define SN_HWPERF_IS_NL3ROUTER(x) ((x) && strstr((x)->name, "NL3Router")) +#define SN_HWPERF_IS_NL4ROUTER(x) ((x) && strstr((x)->name, "NL4Router")) +#define SN_HWPERF_IS_OLDROUTER(x) ((x) && strstr((x)->name, "Router")) +#define SN_HWPERF_IS_ROUTER(x) (SN_HWPERF_IS_NL3ROUTER(x) || \ + SN_HWPERF_IS_NL4ROUTER(x) || \ + SN_HWPERF_IS_OLDROUTER(x)) #define SN_HWPERF_FOREIGN(x) ((x) && !(x)->sn_hwp_this_part && !(x)->sn_hwp_is_shared) #define SN_HWPERF_SAME_OBJTYPE(x,y) ((SN_HWPERF_IS_NODE(x) && SN_HWPERF_IS_NODE(y)) ||\ (SN_HWPERF_IS_IONODE(x) && SN_HWPERF_IS_IONODE(y)) ||\ -- cgit v1.2.3 From 1df57c0c21c92a6d4fcfe5304c84151ed9beb7a2 Mon Sep 17 00:00:00 2001 From: Cliff Wickman Date: Tue, 25 Apr 2006 10:47:48 -0500 Subject: [IA64] enable dumps to capture second page of kernel stack In SLES10 (2.6.16) crash dumping (in my experience, LKCD) is unable to capture the second page of the 2-page task/stack allocation. This is particularly troublesome for dump analysis, as the stack traceback cannot be done. (A similar convention is probably needed throughout the kernel to make kernel multi-page allocations detectable for dumping) Multi-page kernel allocations are represented by the single page structure associated with the first page of the allocation. The page structures associated with the other pages are unintialized. If the dumper is selecting only kernel pages it has no way to identify any but the first page of the allocation. The fix is to make the task/stack allocation a compound page. Signed-off-by: Cliff Wickman Signed-off-by: Tony Luck --- include/asm-ia64/thread_info.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include') diff --git a/include/asm-ia64/thread_info.h b/include/asm-ia64/thread_info.h index 56394a2c7055..e5392c4d30c6 100644 --- a/include/asm-ia64/thread_info.h +++ b/include/asm-ia64/thread_info.h @@ -67,7 +67,7 @@ struct thread_info { #define end_of_stack(p) (unsigned long *)((void *)(p) + IA64_RBS_OFFSET) #define __HAVE_ARCH_TASK_STRUCT_ALLOCATOR -#define alloc_task_struct() ((task_t *)__get_free_pages(GFP_KERNEL, KERNEL_STACK_SIZE_ORDER)) +#define alloc_task_struct() ((task_t *)__get_free_pages(GFP_KERNEL | __GFP_COMP, KERNEL_STACK_SIZE_ORDER)) #define free_task_struct(tsk) free_pages((unsigned long) (tsk), KERNEL_STACK_SIZE_ORDER) #endif /* !__ASSEMBLY */ -- cgit v1.2.3 From f10a04c034c7285a1b15dfa4a83d3e56578e34e8 Mon Sep 17 00:00:00 2001 From: David Gibson Date: Fri, 28 Apr 2006 15:02:51 +1000 Subject: [PATCH] powerpc: Fix pagetable bloat for hugepages At present, ARCH=powerpc kernels can waste considerable space in pagetables when making large hugepage mappings. Hugepage PTEs go in PMD pages, but each PMD page maps 256M and so contains only 16 hugepage PTEs (128 bytes of data), but takes up a 1024 byte allocation. With CONFIG_PPC_64K_PAGES enabled (64k base page size), the situation is worse. Now hugepage PTEs are at the PTE page level (also mapping 256M), so we store 16 hugepage PTEs in a 64k allocation. The PowerPC MMU already means that any 256M region is either all hugepage, or all normal pages. Thus, with some care, we can use a different allocation for the hugepage PTE tables and only allocate the 128 bytes necessary. Signed-off-by: Paul Mackerras --- arch/powerpc/mm/hugetlbpage.c | 295 ++++++++++++++++++++++++++++++++++++------ arch/powerpc/mm/init_64.c | 7 + include/asm-powerpc/page_64.h | 1 + include/asm-powerpc/pgalloc.h | 2 + 4 files changed, 269 insertions(+), 36 deletions(-) (limited to 'include') diff --git a/arch/powerpc/mm/hugetlbpage.c b/arch/powerpc/mm/hugetlbpage.c index 7370f9f33e29..266b8b2ceac9 100644 --- a/arch/powerpc/mm/hugetlbpage.c +++ b/arch/powerpc/mm/hugetlbpage.c @@ -30,13 +30,66 @@ #define NUM_LOW_AREAS (0x100000000UL >> SID_SHIFT) #define NUM_HIGH_AREAS (PGTABLE_RANGE >> HTLB_AREA_SHIFT) +#ifdef CONFIG_PPC_64K_PAGES +#define HUGEPTE_INDEX_SIZE (PMD_SHIFT-HPAGE_SHIFT) +#else +#define HUGEPTE_INDEX_SIZE (PUD_SHIFT-HPAGE_SHIFT) +#endif +#define PTRS_PER_HUGEPTE (1 << HUGEPTE_INDEX_SIZE) +#define HUGEPTE_TABLE_SIZE (sizeof(pte_t) << HUGEPTE_INDEX_SIZE) + +#define HUGEPD_SHIFT (HPAGE_SHIFT + HUGEPTE_INDEX_SIZE) +#define HUGEPD_SIZE (1UL << HUGEPD_SHIFT) +#define HUGEPD_MASK (~(HUGEPD_SIZE-1)) + +#define huge_pgtable_cache (pgtable_cache[HUGEPTE_CACHE_NUM]) + +/* Flag to mark huge PD pointers. This means pmd_bad() and pud_bad() + * will choke on pointers to hugepte tables, which is handy for + * catching screwups early. */ +#define HUGEPD_OK 0x1 + +typedef struct { unsigned long pd; } hugepd_t; + +#define hugepd_none(hpd) ((hpd).pd == 0) + +static inline pte_t *hugepd_page(hugepd_t hpd) +{ + BUG_ON(!(hpd.pd & HUGEPD_OK)); + return (pte_t *)(hpd.pd & ~HUGEPD_OK); +} + +static inline pte_t *hugepte_offset(hugepd_t *hpdp, unsigned long addr) +{ + unsigned long idx = ((addr >> HPAGE_SHIFT) & (PTRS_PER_HUGEPTE-1)); + pte_t *dir = hugepd_page(*hpdp); + + return dir + idx; +} + +static int __hugepte_alloc(struct mm_struct *mm, hugepd_t *hpdp, + unsigned long address) +{ + pte_t *new = kmem_cache_alloc(huge_pgtable_cache, + GFP_KERNEL|__GFP_REPEAT); + + if (! new) + return -ENOMEM; + + spin_lock(&mm->page_table_lock); + if (!hugepd_none(*hpdp)) + kmem_cache_free(huge_pgtable_cache, new); + else + hpdp->pd = (unsigned long)new | HUGEPD_OK; + spin_unlock(&mm->page_table_lock); + return 0; +} + /* Modelled after find_linux_pte() */ pte_t *huge_pte_offset(struct mm_struct *mm, unsigned long addr) { pgd_t *pg; pud_t *pu; - pmd_t *pm; - pte_t *pt; BUG_ON(! in_hugepage_area(mm->context, addr)); @@ -46,26 +99,14 @@ pte_t *huge_pte_offset(struct mm_struct *mm, unsigned long addr) if (!pgd_none(*pg)) { pu = pud_offset(pg, addr); if (!pud_none(*pu)) { - pm = pmd_offset(pu, addr); #ifdef CONFIG_PPC_64K_PAGES - /* Currently, we use the normal PTE offset within full - * size PTE pages, thus our huge PTEs are scattered in - * the PTE page and we do waste some. We may change - * that in the future, but the current mecanism keeps - * things much simpler - */ - if (!pmd_none(*pm)) { - /* Note: pte_offset_* are all equivalent on - * ppc64 as we don't have HIGHMEM - */ - pt = pte_offset_kernel(pm, addr); - return pt; - } -#else /* CONFIG_PPC_64K_PAGES */ - /* On 4k pages, we put huge PTEs in the PMD page */ - pt = (pte_t *)pm; - return pt; -#endif /* CONFIG_PPC_64K_PAGES */ + pmd_t *pm; + pm = pmd_offset(pu, addr); + if (!pmd_none(*pm)) + return hugepte_offset((hugepd_t *)pm, addr); +#else + return hugepte_offset((hugepd_t *)pu, addr); +#endif } } @@ -76,8 +117,7 @@ pte_t *huge_pte_alloc(struct mm_struct *mm, unsigned long addr) { pgd_t *pg; pud_t *pu; - pmd_t *pm; - pte_t *pt; + hugepd_t *hpdp = NULL; BUG_ON(! in_hugepage_area(mm->context, addr)); @@ -87,23 +127,182 @@ pte_t *huge_pte_alloc(struct mm_struct *mm, unsigned long addr) pu = pud_alloc(mm, pg, addr); if (pu) { +#ifdef CONFIG_PPC_64K_PAGES + pmd_t *pm; pm = pmd_alloc(mm, pu, addr); - if (pm) { + if (pm) + hpdp = (hugepd_t *)pm; +#else + hpdp = (hugepd_t *)pu; +#endif + } + + if (! hpdp) + return NULL; + + if (hugepd_none(*hpdp) && __hugepte_alloc(mm, hpdp, addr)) + return NULL; + + return hugepte_offset(hpdp, addr); +} + +static void free_hugepte_range(struct mmu_gather *tlb, hugepd_t *hpdp) +{ + pte_t *hugepte = hugepd_page(*hpdp); + + hpdp->pd = 0; + tlb->need_flush = 1; + pgtable_free_tlb(tlb, pgtable_free_cache(hugepte, HUGEPTE_CACHE_NUM, + HUGEPTE_TABLE_SIZE-1)); +} + #ifdef CONFIG_PPC_64K_PAGES - /* See comment in huge_pte_offset. Note that if we ever - * want to put the page size in the PMD, we would have - * to open code our own pte_alloc* function in order - * to populate and set the size atomically - */ - pt = pte_alloc_map(mm, pm, addr); -#else /* CONFIG_PPC_64K_PAGES */ - pt = (pte_t *)pm; -#endif /* CONFIG_PPC_64K_PAGES */ - return pt; - } +static void hugetlb_free_pmd_range(struct mmu_gather *tlb, pud_t *pud, + unsigned long addr, unsigned long end, + unsigned long floor, unsigned long ceiling) +{ + pmd_t *pmd; + unsigned long next; + unsigned long start; + + start = addr; + pmd = pmd_offset(pud, addr); + do { + next = pmd_addr_end(addr, end); + if (pmd_none(*pmd)) + continue; + free_hugepte_range(tlb, (hugepd_t *)pmd); + } while (pmd++, addr = next, addr != end); + + start &= PUD_MASK; + if (start < floor) + return; + if (ceiling) { + ceiling &= PUD_MASK; + if (!ceiling) + return; } + if (end - 1 > ceiling - 1) + return; - return NULL; + pmd = pmd_offset(pud, start); + pud_clear(pud); + pmd_free_tlb(tlb, pmd); +} +#endif + +static void hugetlb_free_pud_range(struct mmu_gather *tlb, pgd_t *pgd, + unsigned long addr, unsigned long end, + unsigned long floor, unsigned long ceiling) +{ + pud_t *pud; + unsigned long next; + unsigned long start; + + start = addr; + pud = pud_offset(pgd, addr); + do { + next = pud_addr_end(addr, end); +#ifdef CONFIG_PPC_64K_PAGES + if (pud_none_or_clear_bad(pud)) + continue; + hugetlb_free_pmd_range(tlb, pud, addr, next, floor, ceiling); +#else + if (pud_none(*pud)) + continue; + free_hugepte_range(tlb, (hugepd_t *)pud); +#endif + } while (pud++, addr = next, addr != end); + + start &= PGDIR_MASK; + if (start < floor) + return; + if (ceiling) { + ceiling &= PGDIR_MASK; + if (!ceiling) + return; + } + if (end - 1 > ceiling - 1) + return; + + pud = pud_offset(pgd, start); + pgd_clear(pgd); + pud_free_tlb(tlb, pud); +} + +/* + * This function frees user-level page tables of a process. + * + * Must be called with pagetable lock held. + */ +void hugetlb_free_pgd_range(struct mmu_gather **tlb, + unsigned long addr, unsigned long end, + unsigned long floor, unsigned long ceiling) +{ + pgd_t *pgd; + unsigned long next; + unsigned long start; + + /* + * Comments below take from the normal free_pgd_range(). They + * apply here too. The tests against HUGEPD_MASK below are + * essential, because we *don't* test for this at the bottom + * level. Without them we'll attempt to free a hugepte table + * when we unmap just part of it, even if there are other + * active mappings using it. + * + * The next few lines have given us lots of grief... + * + * Why are we testing HUGEPD* at this top level? Because + * often there will be no work to do at all, and we'd prefer + * not to go all the way down to the bottom just to discover + * that. + * + * Why all these "- 1"s? Because 0 represents both the bottom + * of the address space and the top of it (using -1 for the + * top wouldn't help much: the masks would do the wrong thing). + * The rule is that addr 0 and floor 0 refer to the bottom of + * the address space, but end 0 and ceiling 0 refer to the top + * Comparisons need to use "end - 1" and "ceiling - 1" (though + * that end 0 case should be mythical). + * + * Wherever addr is brought up or ceiling brought down, we + * must be careful to reject "the opposite 0" before it + * confuses the subsequent tests. But what about where end is + * brought down by HUGEPD_SIZE below? no, end can't go down to + * 0 there. + * + * Whereas we round start (addr) and ceiling down, by different + * masks at different levels, in order to test whether a table + * now has no other vmas using it, so can be freed, we don't + * bother to round floor or end up - the tests don't need that. + */ + + addr &= HUGEPD_MASK; + if (addr < floor) { + addr += HUGEPD_SIZE; + if (!addr) + return; + } + if (ceiling) { + ceiling &= HUGEPD_MASK; + if (!ceiling) + return; + } + if (end - 1 > ceiling - 1) + end -= HUGEPD_SIZE; + if (addr > end - 1) + return; + + start = addr; + pgd = pgd_offset((*tlb)->mm, addr); + do { + BUG_ON(! in_hugepage_area((*tlb)->mm->context, addr)); + next = pgd_addr_end(addr, end); + if (pgd_none_or_clear_bad(pgd)) + continue; + hugetlb_free_pud_range(*tlb, pgd, addr, next, floor, ceiling); + } while (pgd++, addr = next, addr != end); } void set_huge_pte_at(struct mm_struct *mm, unsigned long addr, @@ -841,3 +1040,27 @@ repeat: out: return err; } + +static void zero_ctor(void *addr, kmem_cache_t *cache, unsigned long flags) +{ + memset(addr, 0, kmem_cache_size(cache)); +} + +static int __init hugetlbpage_init(void) +{ + if (!cpu_has_feature(CPU_FTR_16M_PAGE)) + return -ENODEV; + + huge_pgtable_cache = kmem_cache_create("hugepte_cache", + HUGEPTE_TABLE_SIZE, + HUGEPTE_TABLE_SIZE, + SLAB_HWCACHE_ALIGN | + SLAB_MUST_HWCACHE_ALIGN, + zero_ctor, NULL); + if (! huge_pgtable_cache) + panic("hugetlbpage_init(): could not create hugepte cache\n"); + + return 0; +} + +module_init(hugetlbpage_init); diff --git a/arch/powerpc/mm/init_64.c b/arch/powerpc/mm/init_64.c index babebd15bdc4..9e30f968c184 100644 --- a/arch/powerpc/mm/init_64.c +++ b/arch/powerpc/mm/init_64.c @@ -162,7 +162,14 @@ static const char *pgtable_cache_name[ARRAY_SIZE(pgtable_cache_size)] = { }; #endif /* CONFIG_PPC_64K_PAGES */ +#ifdef CONFIG_HUGETLB_PAGE +/* Hugepages need one extra cache, initialized in hugetlbpage.c. We + * can't put into the tables above, because HPAGE_SHIFT is not compile + * time constant. */ +kmem_cache_t *pgtable_cache[ARRAY_SIZE(pgtable_cache_size)+1]; +#else kmem_cache_t *pgtable_cache[ARRAY_SIZE(pgtable_cache_size)]; +#endif void pgtable_cache_init(void) { diff --git a/include/asm-powerpc/page_64.h b/include/asm-powerpc/page_64.h index 3fb061bab9ec..eab779c21995 100644 --- a/include/asm-powerpc/page_64.h +++ b/include/asm-powerpc/page_64.h @@ -101,6 +101,7 @@ extern unsigned int HPAGE_SHIFT; - (1U << GET_HTLB_AREA(addr))) & 0xffff) #define ARCH_HAS_HUGEPAGE_ONLY_RANGE +#define ARCH_HAS_HUGETLB_FREE_PGD_RANGE #define ARCH_HAS_PREPARE_HUGEPAGE_RANGE #define ARCH_HAS_SETCLEAR_HUGE_PTE diff --git a/include/asm-powerpc/pgalloc.h b/include/asm-powerpc/pgalloc.h index a00ee002cd11..9f0917c68659 100644 --- a/include/asm-powerpc/pgalloc.h +++ b/include/asm-powerpc/pgalloc.h @@ -17,11 +17,13 @@ extern kmem_cache_t *pgtable_cache[]; #define PTE_CACHE_NUM 0 #define PMD_CACHE_NUM 1 #define PGD_CACHE_NUM 2 +#define HUGEPTE_CACHE_NUM 3 #else #define PTE_CACHE_NUM 0 #define PMD_CACHE_NUM 1 #define PUD_CACHE_NUM 1 #define PGD_CACHE_NUM 0 +#define HUGEPTE_CACHE_NUM 2 #endif /* -- cgit v1.2.3 From 1269277a5e7c6d7ae1852e648a8bcdb78035e9fa Mon Sep 17 00:00:00 2001 From: David Woodhouse Date: Mon, 24 Apr 2006 23:22:17 +0100 Subject: [PATCH] powerpc: Use check_legacy_ioport() on ppc32 too. Some people report that we die on some Macs when we are expecting to catch machine checks after poking at some random I/O address. I'd seen it happen on my dual G4 with serial ports until we fixed those to use OF, but now other users are reporting it with i8042. This expands the use of check_legacy_ioport() to avoid that situation even on 32-bit kernels. Signed-off-by: David Woodhouse Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/setup-common.c | 8 ++++++++ arch/powerpc/kernel/setup_64.c | 8 -------- drivers/block/floppy.c | 2 +- drivers/input/serio/i8042-io.h | 4 ++-- include/asm-powerpc/io.h | 6 +++--- 5 files changed, 14 insertions(+), 14 deletions(-) (limited to 'include') diff --git a/arch/powerpc/kernel/setup-common.c b/arch/powerpc/kernel/setup-common.c index 1d93e73a7003..684ab1d49c65 100644 --- a/arch/powerpc/kernel/setup-common.c +++ b/arch/powerpc/kernel/setup-common.c @@ -516,3 +516,11 @@ void probe_machine(void) printk(KERN_INFO "Using %s machine description\n", ppc_md.name); } + +int check_legacy_ioport(unsigned long base_port) +{ + if (ppc_md.check_legacy_ioport == NULL) + return 0; + return ppc_md.check_legacy_ioport(base_port); +} +EXPORT_SYMBOL(check_legacy_ioport); diff --git a/arch/powerpc/kernel/setup_64.c b/arch/powerpc/kernel/setup_64.c index 13e91c4d70a8..4467c49903b6 100644 --- a/arch/powerpc/kernel/setup_64.c +++ b/arch/powerpc/kernel/setup_64.c @@ -594,14 +594,6 @@ void ppc64_terminate_msg(unsigned int src, const char *msg) printk("[terminate]%04x %s\n", src, msg); } -int check_legacy_ioport(unsigned long base_port) -{ - if (ppc_md.check_legacy_ioport == NULL) - return 0; - return ppc_md.check_legacy_ioport(base_port); -} -EXPORT_SYMBOL(check_legacy_ioport); - void cpu_die(void) { if (ppc_md.cpu_die) diff --git a/drivers/block/floppy.c b/drivers/block/floppy.c index bedb689b051f..dff1e67b1dd4 100644 --- a/drivers/block/floppy.c +++ b/drivers/block/floppy.c @@ -4301,7 +4301,7 @@ static int __init floppy_init(void) } use_virtual_dma = can_use_virtual_dma & 1; -#if defined(CONFIG_PPC64) +#if defined(CONFIG_PPC_MERGE) if (check_legacy_ioport(FDC1)) { del_timer(&fd_timeout); err = -ENODEV; diff --git a/drivers/input/serio/i8042-io.h b/drivers/input/serio/i8042-io.h index 9a9221644250..cc21914fbc72 100644 --- a/drivers/input/serio/i8042-io.h +++ b/drivers/input/serio/i8042-io.h @@ -67,14 +67,14 @@ static inline int i8042_platform_init(void) * On some platforms touching the i8042 data register region can do really * bad things. Because of this the region is always reserved on such boxes. */ -#if !defined(__sh__) && !defined(__alpha__) && !defined(__mips__) && !defined(CONFIG_PPC64) +#if !defined(__sh__) && !defined(__alpha__) && !defined(__mips__) && !defined(CONFIG_PPC_MERGE) if (!request_region(I8042_DATA_REG, 16, "i8042")) return -EBUSY; #endif i8042_reset = 1; -#if defined(CONFIG_PPC64) +#if defined(CONFIG_PPC_MERGE) if (check_legacy_ioport(I8042_DATA_REG)) return -EBUSY; if (!request_region(I8042_DATA_REG, 16, "i8042")) diff --git a/include/asm-powerpc/io.h b/include/asm-powerpc/io.h index 68efbea379c9..f1c2469b8844 100644 --- a/include/asm-powerpc/io.h +++ b/include/asm-powerpc/io.h @@ -9,6 +9,9 @@ * 2 of the License, or (at your option) any later version. */ +/* Check of existence of legacy devices */ +extern int check_legacy_ioport(unsigned long base_port); + #ifndef CONFIG_PPC64 #include #else @@ -437,9 +440,6 @@ out: #define dma_cache_wback(_start,_size) do { } while (0) #define dma_cache_wback_inv(_start,_size) do { } while (0) -/* Check of existence of legacy devices */ -extern int check_legacy_ioport(unsigned long base_port); - /* * Convert a physical pointer to a virtual kernel pointer for /dev/mem -- cgit v1.2.3 From 30aacebed0f0619f23ce84df7c59ad033ca08d77 Mon Sep 17 00:00:00 2001 From: Eugene Surovegin Date: Tue, 25 Apr 2006 01:22:44 -0700 Subject: [PATCH] ppc32: add 440GX erratum 440_43 workaround This patch adds workaround for PPC 440GX erratum 440_43. According to this erratum spurious MachineChecks (caused by L1 cache parity) can happen during DataTLB miss processing. We disable L1 cache parity checking for 440GX rev.C and rev.F Signed-off-by: Eugene Surovegin Signed-off-by: Paul Mackerras --- arch/ppc/platforms/4xx/ocotea.c | 2 +- arch/ppc/syslib/ibm440gx_common.c | 13 ++++++++++++- arch/ppc/syslib/ibm440gx_common.h | 4 ++++ include/asm-ppc/reg_booke.h | 1 + 4 files changed, 18 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/arch/ppc/platforms/4xx/ocotea.c b/arch/ppc/platforms/4xx/ocotea.c index f841972f1fa9..554776d4b8ac 100644 --- a/arch/ppc/platforms/4xx/ocotea.c +++ b/arch/ppc/platforms/4xx/ocotea.c @@ -331,7 +331,7 @@ static void __init ocotea_init(void) void __init platform_init(unsigned long r3, unsigned long r4, unsigned long r5, unsigned long r6, unsigned long r7) { - ibm44x_platform_init(r3, r4, r5, r6, r7); + ibm440gx_platform_init(r3, r4, r5, r6, r7); ppc_md.setup_arch = ocotea_setup_arch; ppc_md.show_cpuinfo = ocotea_show_cpuinfo; diff --git a/arch/ppc/syslib/ibm440gx_common.c b/arch/ppc/syslib/ibm440gx_common.c index a7dd55f1c63e..f6cc16888527 100644 --- a/arch/ppc/syslib/ibm440gx_common.c +++ b/arch/ppc/syslib/ibm440gx_common.c @@ -2,7 +2,7 @@ * PPC440GX system library * * Eugene Surovegin or - * Copyright (c) 2003, 2004 Zultys Technologies + * Copyright (c) 2003 - 2006 Zultys Technologies * * 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 @@ -282,3 +282,14 @@ int ibm440gx_show_cpuinfo(struct seq_file *m){ return 0; } +void __init ibm440gx_platform_init(unsigned long r3, unsigned long r4, + unsigned long r5, unsigned long r6, + unsigned long r7) +{ + /* Erratum 440_43 workaround, disable L1 cache parity checking */ + if (!strcmp(cur_cpu_spec->cpu_name, "440GX Rev. C") || + !strcmp(cur_cpu_spec->cpu_name, "440GX Rev. F")) + mtspr(SPRN_CCR1, mfspr(SPRN_CCR1) | CCR1_DPC); + + ibm44x_platform_init(r3, r4, r5, r6, r7); +} diff --git a/arch/ppc/syslib/ibm440gx_common.h b/arch/ppc/syslib/ibm440gx_common.h index a2ab9fab8e34..a03ec6022e8f 100644 --- a/arch/ppc/syslib/ibm440gx_common.h +++ b/arch/ppc/syslib/ibm440gx_common.h @@ -29,6 +29,10 @@ void ibm440gx_get_clocks(struct ibm44x_clocks*, unsigned int sys_clk, unsigned int ser_clk) __init; +/* common 440GX platform init */ +void ibm440gx_platform_init(unsigned long r3, unsigned long r4, unsigned long r5, + unsigned long r6, unsigned long r7) __init; + /* Enable L2 cache */ void ibm440gx_l2c_enable(void) __init; diff --git a/include/asm-ppc/reg_booke.h b/include/asm-ppc/reg_booke.h index 00ad9c754c78..4944c0fb8bea 100644 --- a/include/asm-ppc/reg_booke.h +++ b/include/asm-ppc/reg_booke.h @@ -237,6 +237,7 @@ do { \ #endif /* Bit definitions for CCR1. */ +#define CCR1_DPC 0x00000100 /* Disable L1 I-Cache/D-Cache parity checking */ #define CCR1_TCS 0x00000080 /* Timer Clock Select */ /* Bit definitions for the MCSR. */ -- cgit v1.2.3 From 2833c28aa0d0326780acfa61149a2a02dcb2c9b4 Mon Sep 17 00:00:00 2001 From: Andreas Schwab Date: Thu, 27 Apr 2006 15:46:42 +0200 Subject: [PATCH] powerpc: Wire up *at syscalls Wire up *at syscalls. This patch has been tested on ppc64 (using glibc's testsuite, both 32bit and 64bit), and compile-tested for ppc32 (I have currently no ppc32 system available, but I expect no problems). Signed-off-by: Andreas Schwab Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/systbl.S | 13 +++++++++++++ arch/powerpc/platforms/cell/spu_callbacks.c | 13 +++++++++++++ fs/stat.c | 2 +- include/asm-powerpc/unistd.h | 20 +++++++++++++++++++- 4 files changed, 46 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/arch/powerpc/kernel/systbl.S b/arch/powerpc/kernel/systbl.S index 0b98eea73c5e..cf56a1d499ff 100644 --- a/arch/powerpc/kernel/systbl.S +++ b/arch/powerpc/kernel/systbl.S @@ -325,6 +325,19 @@ SYSCALL(unshare) SYSCALL(splice) SYSCALL(tee) SYSCALL(vmsplice) +COMPAT_SYS(openat) +SYSCALL(mkdirat) +SYSCALL(mknodat) +SYSCALL(fchownat) +COMPAT_SYS(futimesat) +SYSX(sys_newfstatat, sys_fstatat64, sys_fstatat64) +SYSCALL(unlinkat) +SYSCALL(renameat) +SYSCALL(linkat) +SYSCALL(symlinkat) +SYSCALL(readlinkat) +SYSCALL(fchmodat) +SYSCALL(faccessat) /* * please add new calls to arch/powerpc/platforms/cell/spu_callbacks.c diff --git a/arch/powerpc/platforms/cell/spu_callbacks.c b/arch/powerpc/platforms/cell/spu_callbacks.c index b283380a2a18..95b36430aa0f 100644 --- a/arch/powerpc/platforms/cell/spu_callbacks.c +++ b/arch/powerpc/platforms/cell/spu_callbacks.c @@ -319,6 +319,19 @@ void *spu_syscall_table[] = { [__NR_splice] sys_splice, [__NR_tee] sys_tee, [__NR_vmsplice] sys_vmsplice, + [__NR_openat] sys_openat, + [__NR_mkdirat] sys_mkdirat, + [__NR_mknodat] sys_mknodat, + [__NR_fchownat] sys_fchownat, + [__NR_futimesat] sys_futimesat, + [__NR_newfstatat] sys_newfstatat, + [__NR_unlinkat] sys_unlinkat, + [__NR_renameat] sys_renameat, + [__NR_linkat] sys_linkat, + [__NR_symlinkat] sys_symlinkat, + [__NR_readlinkat] sys_readlinkat, + [__NR_fchmodat] sys_fchmodat, + [__NR_faccessat] sys_faccessat, }; long spu_sys_callback(struct spu_syscall_block *s) diff --git a/fs/stat.c b/fs/stat.c index 9948cc1685a4..0f282face322 100644 --- a/fs/stat.c +++ b/fs/stat.c @@ -261,7 +261,7 @@ asmlinkage long sys_newlstat(char __user *filename, struct stat __user *statbuf) return error; } -#ifndef __ARCH_WANT_STAT64 +#if !defined(__ARCH_WANT_STAT64) || defined(__ARCH_WANT_SYS_NEWFSTATAT) asmlinkage long sys_newfstatat(int dfd, char __user *filename, struct stat __user *statbuf, int flag) { diff --git a/include/asm-powerpc/unistd.h b/include/asm-powerpc/unistd.h index 34325e292596..908acb44cb8a 100644 --- a/include/asm-powerpc/unistd.h +++ b/include/asm-powerpc/unistd.h @@ -304,8 +304,25 @@ #define __NR_splice 283 #define __NR_tee 284 #define __NR_vmsplice 285 +#define __NR_openat 286 +#define __NR_mkdirat 287 +#define __NR_mknodat 288 +#define __NR_fchownat 289 +#define __NR_futimesat 290 +#ifdef __powerpc64__ +#define __NR_newfstatat 291 +#else +#define __NR_fstatat64 291 +#endif +#define __NR_unlinkat 292 +#define __NR_renameat 293 +#define __NR_linkat 294 +#define __NR_symlinkat 295 +#define __NR_readlinkat 296 +#define __NR_fchmodat 297 +#define __NR_faccessat 298 -#define __NR_syscalls 286 +#define __NR_syscalls 299 #ifdef __KERNEL__ #define __NR__exit __NR_exit @@ -458,6 +475,7 @@ type name(type1 arg1, type2 arg2, type3 arg3, type4 arg4, type5 arg5, type6 arg6 #ifdef CONFIG_PPC64 #define __ARCH_WANT_COMPAT_SYS_TIME #define __ARCH_WANT_COMPAT_SYS_RT_SIGSUSPEND +#define __ARCH_WANT_SYS_NEWFSTATAT #endif /* -- cgit v1.2.3 From a73c87bfe98f4d54c919e920a6efb0a116115722 Mon Sep 17 00:00:00 2001 From: Vitaly Bordug Date: Tue, 25 Apr 2006 20:26:33 +0400 Subject: [PATCH] ppc32: odd fixes and improvements in ppc_sys This consists of offsets fix in ..._devices.c, and update of ppc_sys_fixup_mem_resource() function to prevent subsequent fixups Signed-off-by: Vitaly Bordug Signed-off-by: Paul Mackerras --- arch/ppc/syslib/mpc8xx_devices.c | 25 +++++++++++++++++++------ arch/ppc/syslib/ppc_sys.c | 4 +++- arch/ppc/syslib/pq2_sys.c | 8 ++++---- include/asm-ppc/ppc_sys.h | 2 ++ 4 files changed, 28 insertions(+), 11 deletions(-) (limited to 'include') diff --git a/arch/ppc/syslib/mpc8xx_devices.c b/arch/ppc/syslib/mpc8xx_devices.c index bd41ed83beb3..6f536383866e 100644 --- a/arch/ppc/syslib/mpc8xx_devices.c +++ b/arch/ppc/syslib/mpc8xx_devices.c @@ -170,12 +170,18 @@ struct platform_device ppc_sys_platform_devices[] = { [MPC8xx_CPM_SMC1] = { .name = "fsl-cpm-smc", .id = 1, - .num_resources = 2, + .num_resources = 3, .resource = (struct resource[]) { { .name = "regs", - .start = 0xa82, - .end = 0xa91, + .start = 0xa80, + .end = 0xa8f, + .flags = IORESOURCE_MEM, + }, + { + .name = "pram", + .start = 0x3e80, + .end = 0x3ebf, .flags = IORESOURCE_MEM, }, { @@ -189,14 +195,21 @@ struct platform_device ppc_sys_platform_devices[] = { [MPC8xx_CPM_SMC2] = { .name = "fsl-cpm-smc", .id = 2, - .num_resources = 2, + .num_resources = 3, .resource = (struct resource[]) { { .name = "regs", - .start = 0xa92, - .end = 0xaa1, + .start = 0xa90, + .end = 0xa9f, .flags = IORESOURCE_MEM, }, + { + .name = "pram", + .start = 0x3f80, + .end = 0x3fbf, + .flags = IORESOURCE_MEM, + + }, { .name = "interrupt", .start = MPC8xx_INT_SMC2, diff --git a/arch/ppc/syslib/ppc_sys.c b/arch/ppc/syslib/ppc_sys.c index 7662c4e6e7d6..2d48018b71d9 100644 --- a/arch/ppc/syslib/ppc_sys.c +++ b/arch/ppc/syslib/ppc_sys.c @@ -109,9 +109,11 @@ ppc_sys_fixup_mem_resource(struct platform_device *pdev, phys_addr_t paddr) int i; for (i = 0; i < pdev->num_resources; i++) { struct resource *r = &pdev->resource[i]; - if ((r->flags & IORESOURCE_MEM) == IORESOURCE_MEM) { + if (((r->flags & IORESOURCE_MEM) == IORESOURCE_MEM) && + ((r->flags & PPC_SYS_IORESOURCE_FIXUPPED) != PPC_SYS_IORESOURCE_FIXUPPED)) { r->start += paddr; r->end += paddr; + r->flags |= PPC_SYS_IORESOURCE_FIXUPPED; } } } diff --git a/arch/ppc/syslib/pq2_sys.c b/arch/ppc/syslib/pq2_sys.c index 75e64f1c144d..433b0fa203e1 100644 --- a/arch/ppc/syslib/pq2_sys.c +++ b/arch/ppc/syslib/pq2_sys.c @@ -113,13 +113,13 @@ struct ppc_sys_spec ppc_sys_specs[] = { .ppc_sys_name = "8248", .mask = 0x0000ff00, .value = 0x00000c00, - .num_devices = 11, + .num_devices = 12, .device_list = (enum ppc_sys_devices[]) { MPC82xx_CPM_FCC1, MPC82xx_CPM_FCC2, MPC82xx_CPM_SCC1, - MPC82xx_CPM_SCC2, MPC82xx_CPM_SCC3, MPC82xx_CPM_SMC1, - MPC82xx_CPM_SMC2, MPC82xx_CPM_SPI, MPC82xx_CPM_I2C, - MPC82xx_CPM_USB, MPC82xx_SEC1, + MPC82xx_CPM_SCC2, MPC82xx_CPM_SCC3, MPC82xx_CPM_SCC4, + MPC82xx_CPM_SMC1, MPC82xx_CPM_SMC2, MPC82xx_CPM_SPI, + MPC82xx_CPM_I2C, MPC82xx_CPM_USB, MPC82xx_SEC1, }, }, { diff --git a/include/asm-ppc/ppc_sys.h b/include/asm-ppc/ppc_sys.h index 4b94f7059ebe..40f197af6508 100644 --- a/include/asm-ppc/ppc_sys.h +++ b/include/asm-ppc/ppc_sys.h @@ -39,6 +39,8 @@ #error "need definition of ppc_sys_devices" #endif +#define PPC_SYS_IORESOURCE_FIXUPPED 0x00000001 + struct ppc_sys_spec { /* PPC sys is matched via (ID & mask) == value, id could be * PVR, SVR, IMMR, * etc. */ -- cgit v1.2.3 From 03054d51a70e8c273df5d9bc31fea6c843eaa1c3 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Sat, 29 Apr 2006 09:51:06 +1000 Subject: [PATCH] powerpc: Add cputable entry for POWER6 Add a cputable entry for the POWER6 processor. The SIHV and SIPR bits in the mmcra have moved in POWER6, so disable support for that until oprofile is fixed. Also tell firmware that we know about POWER6. Signed-off-by: Anton Blanchard Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/cputable.c | 16 ++++++++++++++++ arch/powerpc/kernel/prom_init.c | 1 + include/asm-powerpc/cputable.h | 14 ++++++++++---- 3 files changed, 27 insertions(+), 4 deletions(-) (limited to 'include') diff --git a/arch/powerpc/kernel/cputable.c b/arch/powerpc/kernel/cputable.c index 39e348a3ade2..3f7182db9ed5 100644 --- a/arch/powerpc/kernel/cputable.c +++ b/arch/powerpc/kernel/cputable.c @@ -57,6 +57,8 @@ extern void __setup_cpu_ppc970(unsigned long offset, struct cpu_spec* spec); PPC_FEATURE_SMT | PPC_FEATURE_ICACHE_SNOOP) #define COMMON_USER_POWER5_PLUS (COMMON_USER_PPC64 | PPC_FEATURE_POWER5_PLUS|\ PPC_FEATURE_SMT | PPC_FEATURE_ICACHE_SNOOP) +#define COMMON_USER_POWER6 (COMMON_USER_PPC64 | PPC_FEATURE_ARCH_2_05 |\ + PPC_FEATURE_SMT | PPC_FEATURE_ICACHE_SNOOP) #define COMMON_USER_BOOKE (PPC_FEATURE_32 | PPC_FEATURE_HAS_MMU | \ PPC_FEATURE_BOOKE) @@ -263,6 +265,20 @@ struct cpu_spec cpu_specs[] = { .oprofile_type = PPC_OPROFILE_POWER4, .platform = "power5+", }, + { /* Power6 */ + .pvr_mask = 0xffff0000, + .pvr_value = 0x003e0000, + .cpu_name = "POWER6", + .cpu_features = CPU_FTRS_POWER6, + .cpu_user_features = COMMON_USER_POWER6, + .icache_bsize = 128, + .dcache_bsize = 128, + .num_pmcs = 6, + .cpu_setup = __setup_cpu_power4, + .oprofile_cpu_type = "ppc64/power6", + .oprofile_type = PPC_OPROFILE_POWER4, + .platform = "power6", + }, { /* Cell Broadband Engine */ .pvr_mask = 0xffff0000, .pvr_value = 0x00700000, diff --git a/arch/powerpc/kernel/prom_init.c b/arch/powerpc/kernel/prom_init.c index 00ad7f981a0d..078fb5533541 100644 --- a/arch/powerpc/kernel/prom_init.c +++ b/arch/powerpc/kernel/prom_init.c @@ -686,6 +686,7 @@ static void __init early_cmdline_parse(void) */ static unsigned char ibm_architecture_vec[] = { W(0xfffe0000), W(0x003a0000), /* POWER5/POWER5+ */ + W(0xffff0000), W(0x003e0000), /* POWER6 */ W(0xfffffffe), W(0x0f000001), /* all 2.04-compliant and earlier */ 5 - 1, /* 5 option vectors */ diff --git a/include/asm-powerpc/cputable.h b/include/asm-powerpc/cputable.h index 4321483cce51..9fcf0162d859 100644 --- a/include/asm-powerpc/cputable.h +++ b/include/asm-powerpc/cputable.h @@ -22,6 +22,7 @@ #define PPC_FEATURE_BOOKE 0x00008000 #define PPC_FEATURE_SMT 0x00004000 #define PPC_FEATURE_ICACHE_SNOOP 0x00002000 +#define PPC_FEATURE_ARCH_2_05 0x00001000 #ifdef __KERNEL__ #ifndef __ASSEMBLY__ @@ -320,6 +321,11 @@ extern void do_cpu_ftr_fixups(unsigned long offset); CPU_FTR_MMCRA | CPU_FTR_SMT | \ CPU_FTR_COHERENT_ICACHE | CPU_FTR_LOCKLESS_TLBIE | \ CPU_FTR_MMCRA_SIHV | CPU_FTR_PURR) +#define CPU_FTRS_POWER6 (CPU_FTR_SPLIT_ID_CACHE | CPU_FTR_USE_TB | \ + CPU_FTR_HPTE_TABLE | CPU_FTR_PPCAS_ARCH_V2 | \ + CPU_FTR_MMCRA | CPU_FTR_SMT | \ + CPU_FTR_COHERENT_ICACHE | CPU_FTR_LOCKLESS_TLBIE | \ + CPU_FTR_PURR | CPU_FTR_CI_LARGE_PAGE) #define CPU_FTRS_CELL (CPU_FTR_SPLIT_ID_CACHE | CPU_FTR_USE_TB | \ CPU_FTR_HPTE_TABLE | CPU_FTR_PPCAS_ARCH_V2 | \ CPU_FTR_ALTIVEC_COMP | CPU_FTR_MMCRA | CPU_FTR_SMT | \ @@ -331,8 +337,8 @@ extern void do_cpu_ftr_fixups(unsigned long offset); #ifdef __powerpc64__ #define CPU_FTRS_POSSIBLE \ (CPU_FTRS_POWER3 | CPU_FTRS_RS64 | CPU_FTRS_POWER4 | \ - CPU_FTRS_PPC970 | CPU_FTRS_POWER5 | CPU_FTRS_CELL | \ - CPU_FTR_CI_LARGE_PAGE) + CPU_FTRS_PPC970 | CPU_FTRS_POWER5 | CPU_FTRS_POWER6 | \ + CPU_FTRS_CELL | CPU_FTR_CI_LARGE_PAGE) #else enum { CPU_FTRS_POSSIBLE = @@ -376,8 +382,8 @@ enum { #ifdef __powerpc64__ #define CPU_FTRS_ALWAYS \ (CPU_FTRS_POWER3 & CPU_FTRS_RS64 & CPU_FTRS_POWER4 & \ - CPU_FTRS_PPC970 & CPU_FTRS_POWER5 & CPU_FTRS_CELL & \ - CPU_FTRS_POSSIBLE) + CPU_FTRS_PPC970 & CPU_FTRS_POWER5 & CPU_FTRS_POWER6 & \ + CPU_FTRS_CELL & CPU_FTRS_POSSIBLE) #else enum { CPU_FTRS_ALWAYS = -- cgit v1.2.3