From ba1f14fbe70965ae0fb1655a5275a62723f65b77 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 28 Nov 2013 14:26:41 +0100 Subject: sched: Remove PREEMPT_NEED_RESCHED from generic code While hunting a preemption issue with Alexander, Ben noticed that the currently generic PREEMPT_NEED_RESCHED stuff is horribly broken for load-store architectures. We currently rely on the IPI to fold TIF_NEED_RESCHED into PREEMPT_NEED_RESCHED, but when this IPI lands while we already have a load for the preempt-count but before the store, the store will erase the PREEMPT_NEED_RESCHED change. The current preempt-count only works on load-store archs because interrupts are assumed to be completely balanced wrt their preempt_count fiddling; the previous preempt_count load will match the preempt_count state after the interrupt and therefore nothing gets lost. This patch removes the PREEMPT_NEED_RESCHED usage from generic code and pushes it into x86 arch code; the generic code goes back to relying on TIF_NEED_RESCHED. Boot tested on x86_64 and compile tested on ppc64. Reported-by: Benjamin Herrenschmidt Reported-and-Tested-by: Alexander Graf Signed-off-by: Peter Zijlstra Cc: Linus Torvalds Link: http://lkml.kernel.org/r/20131128132641.GP10022@twins.programming.kicks-ass.net Signed-off-by: Ingo Molnar --- include/linux/sched.h | 2 -- 1 file changed, 2 deletions(-) (limited to 'include/linux') diff --git a/include/linux/sched.h b/include/linux/sched.h index 768b037dfacb..96d674ba3876 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -440,8 +440,6 @@ struct task_cputime { .sum_exec_runtime = 0, \ } -#define PREEMPT_ENABLED (PREEMPT_NEED_RESCHED) - #ifdef CONFIG_PREEMPT_COUNT #define PREEMPT_DISABLED (1 + PREEMPT_ENABLED) #else -- cgit v1.2.3 From be5e610c0fd6ef772cafb9e0bd4128134804aef3 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 18 Nov 2013 18:27:06 +0100 Subject: math64: Add mul_u64_u32_shr() Introduce mul_u64_u32_shr() as proposed by Andy a while back; it allows using 64x64->128 muls on 64bit archs and recent GCC which defines __SIZEOF_INT128__ and __int128. (This new method will be used by the scheduler.) Signed-off-by: Peter Zijlstra Cc: fweisbec@gmail.com Cc: Andy Lutomirski Cc: Linus Torvalds Link: http://lkml.kernel.org/n/tip-hxjoeuzmrcaumR0uZwjpe2pv@git.kernel.org Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 1 + include/linux/math64.h | 30 ++++++++++++++++++++++++++++++ init/Kconfig | 6 ++++++ 3 files changed, 37 insertions(+) (limited to 'include/linux') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index e903c71f7e69..0952ecd60eca 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -26,6 +26,7 @@ config X86 select HAVE_AOUT if X86_32 select HAVE_UNSTABLE_SCHED_CLOCK select ARCH_SUPPORTS_NUMA_BALANCING + select ARCH_SUPPORTS_INT128 if X86_64 select ARCH_WANTS_PROT_NUMA_PROT_NONE select HAVE_IDE select HAVE_OPROFILE diff --git a/include/linux/math64.h b/include/linux/math64.h index 69ed5f5e9f6e..c45c089bfdac 100644 --- a/include/linux/math64.h +++ b/include/linux/math64.h @@ -133,4 +133,34 @@ __iter_div_u64_rem(u64 dividend, u32 divisor, u64 *remainder) return ret; } +#if defined(CONFIG_ARCH_SUPPORTS_INT128) && defined(__SIZEOF_INT128__) + +#ifndef mul_u64_u32_shr +static inline u64 mul_u64_u32_shr(u64 a, u32 mul, unsigned int shift) +{ + return (u64)(((unsigned __int128)a * mul) >> shift); +} +#endif /* mul_u64_u32_shr */ + +#else + +#ifndef mul_u64_u32_shr +static inline u64 mul_u64_u32_shr(u64 a, u32 mul, unsigned int shift) +{ + u32 ah, al; + u64 ret; + + al = a; + ah = a >> 32; + + ret = ((u64)al * mul) >> shift; + if (ah) + ret += ((u64)ah * mul) << (32 - shift); + + return ret; +} +#endif /* mul_u64_u32_shr */ + +#endif + #endif /* _LINUX_MATH64_H */ diff --git a/init/Kconfig b/init/Kconfig index 79383d3aa5dc..4e5d96ab2034 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -809,6 +809,12 @@ config GENERIC_SCHED_CLOCK config ARCH_SUPPORTS_NUMA_BALANCING bool +# +# For architectures that know their GCC __int128 support is sound +# +config ARCH_SUPPORTS_INT128 + bool + # For architectures that (ab)use NUMA to represent different memory regions # all cpu-local but of different latencies, such as SuperH. # -- cgit v1.2.3 From 9dbdb155532395ba000c5d5d187658b0e17e529f Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 18 Nov 2013 18:27:06 +0100 Subject: sched/fair: Rework sched_fair time accounting Christian suffers from a bad BIOS that wrecks his i5's TSC sync. This results in him occasionally seeing time going backwards - which crashes the scheduler ... Most of our time accounting can actually handle that except the most common one; the tick time update of sched_fair. There is a further problem with that code; previously we assumed that because we get a tick every TICK_NSEC our time delta could never exceed 32bits and math was simpler. However, ever since Frederic managed to get NO_HZ_FULL merged; this is no longer the case since now a task can run for a long time indeed without getting a tick. It only takes about ~4.2 seconds to overflow our u32 in nanoseconds. This means we not only need to better deal with time going backwards; but also means we need to be able to deal with large deltas. This patch reworks the entire code and uses mul_u64_u32_shr() as proposed by Andy a long while ago. We express our virtual time scale factor in a u32 multiplier and shift right and the 32bit mul_u64_u32_shr() implementation reduces to a single 32x32->64 multiply if the time delta is still short (common case). For 64bit a 64x64->128 multiply can be used if ARCH_SUPPORTS_INT128. Reported-and-Tested-by: Christian Engelmayer Signed-off-by: Peter Zijlstra Cc: fweisbec@gmail.com Cc: Paul Turner Cc: Stanislaw Gruszka Cc: Andy Lutomirski Cc: Linus Torvalds Cc: Andrew Morton Link: http://lkml.kernel.org/r/20131118172706.GI3866@twins.programming.kicks-ass.net Signed-off-by: Ingo Molnar --- include/linux/sched.h | 3 +- kernel/sched/fair.c | 144 ++++++++++++++++++++++---------------------------- 2 files changed, 66 insertions(+), 81 deletions(-) (limited to 'include/linux') diff --git a/include/linux/sched.h b/include/linux/sched.h index 96d674ba3876..53f97eb8dbc7 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -930,7 +930,8 @@ struct pipe_inode_info; struct uts_namespace; struct load_weight { - unsigned long weight, inv_weight; + unsigned long weight; + u32 inv_weight; }; struct sched_avg { diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index fd773ade1a31..9030da7bcb15 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -178,59 +178,61 @@ void sched_init_granularity(void) update_sysctl(); } -#if BITS_PER_LONG == 32 -# define WMULT_CONST (~0UL) -#else -# define WMULT_CONST (1UL << 32) -#endif - +#define WMULT_CONST (~0U) #define WMULT_SHIFT 32 -/* - * Shift right and round: - */ -#define SRR(x, y) (((x) + (1UL << ((y) - 1))) >> (y)) +static void __update_inv_weight(struct load_weight *lw) +{ + unsigned long w; + + if (likely(lw->inv_weight)) + return; + + w = scale_load_down(lw->weight); + + if (BITS_PER_LONG > 32 && unlikely(w >= WMULT_CONST)) + lw->inv_weight = 1; + else if (unlikely(!w)) + lw->inv_weight = WMULT_CONST; + else + lw->inv_weight = WMULT_CONST / w; +} /* - * delta *= weight / lw + * delta_exec * weight / lw.weight + * OR + * (delta_exec * (weight * lw->inv_weight)) >> WMULT_SHIFT + * + * Either weight := NICE_0_LOAD and lw \e prio_to_wmult[], in which case + * we're guaranteed shift stays positive because inv_weight is guaranteed to + * fit 32 bits, and NICE_0_LOAD gives another 10 bits; therefore shift >= 22. + * + * Or, weight =< lw.weight (because lw.weight is the runqueue weight), thus + * weight/lw.weight <= 1, and therefore our shift will also be positive. */ -static unsigned long -calc_delta_mine(unsigned long delta_exec, unsigned long weight, - struct load_weight *lw) +static u64 __calc_delta(u64 delta_exec, unsigned long weight, struct load_weight *lw) { - u64 tmp; + u64 fact = scale_load_down(weight); + int shift = WMULT_SHIFT; - /* - * weight can be less than 2^SCHED_LOAD_RESOLUTION for task group sched - * entities since MIN_SHARES = 2. Treat weight as 1 if less than - * 2^SCHED_LOAD_RESOLUTION. - */ - if (likely(weight > (1UL << SCHED_LOAD_RESOLUTION))) - tmp = (u64)delta_exec * scale_load_down(weight); - else - tmp = (u64)delta_exec; + __update_inv_weight(lw); - if (!lw->inv_weight) { - unsigned long w = scale_load_down(lw->weight); - - if (BITS_PER_LONG > 32 && unlikely(w >= WMULT_CONST)) - lw->inv_weight = 1; - else if (unlikely(!w)) - lw->inv_weight = WMULT_CONST; - else - lw->inv_weight = WMULT_CONST / w; + if (unlikely(fact >> 32)) { + while (fact >> 32) { + fact >>= 1; + shift--; + } } - /* - * Check whether we'd overflow the 64-bit multiplication: - */ - if (unlikely(tmp > WMULT_CONST)) - tmp = SRR(SRR(tmp, WMULT_SHIFT/2) * lw->inv_weight, - WMULT_SHIFT/2); - else - tmp = SRR(tmp * lw->inv_weight, WMULT_SHIFT); + /* hint to use a 32x32->64 mul */ + fact = (u64)(u32)fact * lw->inv_weight; + + while (fact >> 32) { + fact >>= 1; + shift--; + } - return (unsigned long)min(tmp, (u64)(unsigned long)LONG_MAX); + return mul_u64_u32_shr(delta_exec, fact, shift); } @@ -443,7 +445,7 @@ find_matching_se(struct sched_entity **se, struct sched_entity **pse) #endif /* CONFIG_FAIR_GROUP_SCHED */ static __always_inline -void account_cfs_rq_runtime(struct cfs_rq *cfs_rq, unsigned long delta_exec); +void account_cfs_rq_runtime(struct cfs_rq *cfs_rq, u64 delta_exec); /************************************************************** * Scheduling class tree data structure manipulation methods: @@ -612,11 +614,10 @@ int sched_proc_update_handler(struct ctl_table *table, int write, /* * delta /= w */ -static inline unsigned long -calc_delta_fair(unsigned long delta, struct sched_entity *se) +static inline u64 calc_delta_fair(u64 delta, struct sched_entity *se) { if (unlikely(se->load.weight != NICE_0_LOAD)) - delta = calc_delta_mine(delta, NICE_0_LOAD, &se->load); + delta = __calc_delta(delta, NICE_0_LOAD, &se->load); return delta; } @@ -665,7 +666,7 @@ static u64 sched_slice(struct cfs_rq *cfs_rq, struct sched_entity *se) update_load_add(&lw, se->load.weight); load = &lw; } - slice = calc_delta_mine(slice, se->load.weight, load); + slice = __calc_delta(slice, se->load.weight, load); } return slice; } @@ -703,47 +704,32 @@ void init_task_runnable_average(struct task_struct *p) #endif /* - * Update the current task's runtime statistics. Skip current tasks that - * are not in our scheduling class. + * Update the current task's runtime statistics. */ -static inline void -__update_curr(struct cfs_rq *cfs_rq, struct sched_entity *curr, - unsigned long delta_exec) -{ - unsigned long delta_exec_weighted; - - schedstat_set(curr->statistics.exec_max, - max((u64)delta_exec, curr->statistics.exec_max)); - - curr->sum_exec_runtime += delta_exec; - schedstat_add(cfs_rq, exec_clock, delta_exec); - delta_exec_weighted = calc_delta_fair(delta_exec, curr); - - curr->vruntime += delta_exec_weighted; - update_min_vruntime(cfs_rq); -} - static void update_curr(struct cfs_rq *cfs_rq) { struct sched_entity *curr = cfs_rq->curr; u64 now = rq_clock_task(rq_of(cfs_rq)); - unsigned long delta_exec; + u64 delta_exec; if (unlikely(!curr)) return; - /* - * Get the amount of time the current task was running - * since the last time we changed load (this cannot - * overflow on 32 bits): - */ - delta_exec = (unsigned long)(now - curr->exec_start); - if (!delta_exec) + delta_exec = now - curr->exec_start; + if (unlikely((s64)delta_exec <= 0)) return; - __update_curr(cfs_rq, curr, delta_exec); curr->exec_start = now; + schedstat_set(curr->statistics.exec_max, + max(delta_exec, curr->statistics.exec_max)); + + curr->sum_exec_runtime += delta_exec; + schedstat_add(cfs_rq, exec_clock, delta_exec); + + curr->vruntime += calc_delta_fair(delta_exec, curr); + update_min_vruntime(cfs_rq); + if (entity_is_task(curr)) { struct task_struct *curtask = task_of(curr); @@ -3015,8 +3001,7 @@ static void expire_cfs_rq_runtime(struct cfs_rq *cfs_rq) } } -static void __account_cfs_rq_runtime(struct cfs_rq *cfs_rq, - unsigned long delta_exec) +static void __account_cfs_rq_runtime(struct cfs_rq *cfs_rq, u64 delta_exec) { /* dock delta_exec before expiring quota (as it could span periods) */ cfs_rq->runtime_remaining -= delta_exec; @@ -3034,7 +3019,7 @@ static void __account_cfs_rq_runtime(struct cfs_rq *cfs_rq, } static __always_inline -void account_cfs_rq_runtime(struct cfs_rq *cfs_rq, unsigned long delta_exec) +void account_cfs_rq_runtime(struct cfs_rq *cfs_rq, u64 delta_exec) { if (!cfs_bandwidth_used() || !cfs_rq->runtime_enabled) return; @@ -3574,8 +3559,7 @@ static inline u64 cfs_rq_clock_task(struct cfs_rq *cfs_rq) return rq_clock_task(rq_of(cfs_rq)); } -static void account_cfs_rq_runtime(struct cfs_rq *cfs_rq, - unsigned long delta_exec) {} +static void account_cfs_rq_runtime(struct cfs_rq *cfs_rq, u64 delta_exec) {} static void check_cfs_rq_runtime(struct cfs_rq *cfs_rq) {} static void check_enqueue_throttle(struct cfs_rq *cfs_rq) {} static __always_inline void return_cfs_rq_runtime(struct cfs_rq *cfs_rq) {} -- cgit v1.2.3 From c97102ba96324da330078ad8619ba4dfe840dbe3 Mon Sep 17 00:00:00 2001 From: Vivek Goyal Date: Wed, 18 Dec 2013 17:08:31 -0800 Subject: kexec: migrate to reboot cpu Commit 1b3a5d02ee07 ("reboot: move arch/x86 reboot= handling to generic kernel") moved reboot= handling to generic code. In the process it also removed the code in native_machine_shutdown() which are moving reboot process to reboot_cpu/cpu0. I guess that thought must have been that all reboot paths are calling migrate_to_reboot_cpu(), so we don't need this special handling. But kexec reboot path (kernel_kexec()) is not calling migrate_to_reboot_cpu() so above change broke kexec. Now reboot can happen on non-boot cpu and when INIT is sent in second kerneo to bring up BP, it brings down the machine. So start calling migrate_to_reboot_cpu() in kexec reboot path to avoid this problem. Bisected by WANG Chao. Reported-by: Matthew Whitehead Reported-by: Dave Young Signed-off-by: Vivek Goyal Tested-by: Baoquan He Tested-by: WANG Chao Acked-by: H. Peter Anvin Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/reboot.h | 1 + kernel/kexec.c | 1 + kernel/reboot.c | 2 +- 3 files changed, 3 insertions(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/reboot.h b/include/linux/reboot.h index 8e00f9f6f963..9e7db9e73cc1 100644 --- a/include/linux/reboot.h +++ b/include/linux/reboot.h @@ -43,6 +43,7 @@ extern int unregister_reboot_notifier(struct notifier_block *); * Architecture-specific implementations of sys_reboot commands. */ +extern void migrate_to_reboot_cpu(void); extern void machine_restart(char *cmd); extern void machine_halt(void); extern void machine_power_off(void); diff --git a/kernel/kexec.c b/kernel/kexec.c index d0d8fca54065..9c970167e402 100644 --- a/kernel/kexec.c +++ b/kernel/kexec.c @@ -1680,6 +1680,7 @@ int kernel_kexec(void) { kexec_in_progress = true; kernel_restart_prepare(NULL); + migrate_to_reboot_cpu(); printk(KERN_EMERG "Starting new kernel\n"); machine_shutdown(); } diff --git a/kernel/reboot.c b/kernel/reboot.c index f813b3474646..662c83fc16b7 100644 --- a/kernel/reboot.c +++ b/kernel/reboot.c @@ -104,7 +104,7 @@ int unregister_reboot_notifier(struct notifier_block *nb) } EXPORT_SYMBOL(unregister_reboot_notifier); -static void migrate_to_reboot_cpu(void) +void migrate_to_reboot_cpu(void) { /* The boot cpu is always logical cpu 0 */ int cpu = reboot_cpu; -- cgit v1.2.3 From de466bd628e8d663fdf3f791bc8db318ee85c714 Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Wed, 18 Dec 2013 17:08:42 -0800 Subject: mm: numa: avoid unnecessary disruption of NUMA hinting during migration do_huge_pmd_numa_page() handles the case where there is parallel THP migration. However, by the time it is checked the NUMA hinting information has already been disrupted. This patch adds an earlier check with some helpers. Signed-off-by: Mel Gorman Reviewed-by: Rik van Riel Cc: Alex Thorlton Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/migrate.h | 9 +++++++++ mm/huge_memory.c | 22 ++++++++++++++++------ mm/migrate.c | 12 ++++++++++++ 3 files changed, 37 insertions(+), 6 deletions(-) (limited to 'include/linux') diff --git a/include/linux/migrate.h b/include/linux/migrate.h index f5096b58b20d..b7717d74da7f 100644 --- a/include/linux/migrate.h +++ b/include/linux/migrate.h @@ -90,10 +90,19 @@ static inline int migrate_huge_page_move_mapping(struct address_space *mapping, #endif /* CONFIG_MIGRATION */ #ifdef CONFIG_NUMA_BALANCING +extern bool pmd_trans_migrating(pmd_t pmd); +extern void wait_migrate_huge_page(struct anon_vma *anon_vma, pmd_t *pmd); extern int migrate_misplaced_page(struct page *page, struct vm_area_struct *vma, int node); extern bool migrate_ratelimited(int node); #else +static inline bool pmd_trans_migrating(pmd_t pmd) +{ + return false; +} +static inline void wait_migrate_huge_page(struct anon_vma *anon_vma, pmd_t *pmd) +{ +} static inline int migrate_misplaced_page(struct page *page, struct vm_area_struct *vma, int node) { diff --git a/mm/huge_memory.c b/mm/huge_memory.c index 70e7429fd8ea..7de1bf85f683 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -882,6 +882,10 @@ int copy_huge_pmd(struct mm_struct *dst_mm, struct mm_struct *src_mm, ret = 0; goto out_unlock; } + + /* mmap_sem prevents this happening but warn if that changes */ + WARN_ON(pmd_trans_migrating(pmd)); + if (unlikely(pmd_trans_splitting(pmd))) { /* split huge page running from under us */ spin_unlock(src_ptl); @@ -1299,6 +1303,17 @@ int do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_struct *vma, if (unlikely(!pmd_same(pmd, *pmdp))) goto out_unlock; + /* + * If there are potential migrations, wait for completion and retry + * without disrupting NUMA hinting information. Do not relock and + * check_same as the page may no longer be mapped. + */ + if (unlikely(pmd_trans_migrating(*pmdp))) { + spin_unlock(ptl); + wait_migrate_huge_page(vma->anon_vma, pmdp); + goto out; + } + page = pmd_page(pmd); BUG_ON(is_huge_zero_page(page)); page_nid = page_to_nid(page); @@ -1329,12 +1344,7 @@ int do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_struct *vma, goto clear_pmdnuma; } - /* - * If there are potential migrations, wait for completion and retry. We - * do not relock and check_same as the page may no longer be mapped. - * Furtermore, even if the page is currently misplaced, there is no - * guarantee it is still misplaced after the migration completes. - */ + /* Migration could have started since the pmd_trans_migrating check */ if (!page_locked) { spin_unlock(ptl); wait_on_page_locked(page); diff --git a/mm/migrate.c b/mm/migrate.c index a987525810ae..cfb419085261 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -1655,6 +1655,18 @@ int numamigrate_isolate_page(pg_data_t *pgdat, struct page *page) return 1; } +bool pmd_trans_migrating(pmd_t pmd) +{ + struct page *page = pmd_page(pmd); + return PageLocked(page); +} + +void wait_migrate_huge_page(struct anon_vma *anon_vma, pmd_t *pmd) +{ + struct page *page = pmd_page(*pmd); + wait_on_page_locked(page); +} + /* * Attempt to migrate a misplaced page to the specified destination * node. Caller is expected to have an elevated reference count on -- cgit v1.2.3 From 20841405940e7be0617612d521e206e4b6b325db Mon Sep 17 00:00:00 2001 From: Rik van Riel Date: Wed, 18 Dec 2013 17:08:44 -0800 Subject: mm: fix TLB flush race between migration, and change_protection_range There are a few subtle races, between change_protection_range (used by mprotect and change_prot_numa) on one side, and NUMA page migration and compaction on the other side. The basic race is that there is a time window between when the PTE gets made non-present (PROT_NONE or NUMA), and the TLB is flushed. During that time, a CPU may continue writing to the page. This is fine most of the time, however compaction or the NUMA migration code may come in, and migrate the page away. When that happens, the CPU may continue writing, through the cached translation, to what is no longer the current memory location of the process. This only affects x86, which has a somewhat optimistic pte_accessible. All other architectures appear to be safe, and will either always flush, or flush whenever there is a valid mapping, even with no permissions (SPARC). The basic race looks like this: CPU A CPU B CPU C load TLB entry make entry PTE/PMD_NUMA fault on entry read/write old page start migrating page change PTE/PMD to new page read/write old page [*] flush TLB reload TLB from new entry read/write new page lose data [*] the old page may belong to a new user at this point! The obvious fix is to flush remote TLB entries, by making sure that pte_accessible aware of the fact that PROT_NONE and PROT_NUMA memory may still be accessible if there is a TLB flush pending for the mm. This should fix both NUMA migration and compaction. [mgorman@suse.de: fix build] Signed-off-by: Rik van Riel Signed-off-by: Mel Gorman Cc: Alex Thorlton Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/sparc/include/asm/pgtable_64.h | 4 ++-- arch/x86/include/asm/pgtable.h | 11 ++++++++-- include/asm-generic/pgtable.h | 2 +- include/linux/mm_types.h | 44 +++++++++++++++++++++++++++++++++++++ kernel/fork.c | 1 + mm/huge_memory.c | 7 ++++++ mm/mprotect.c | 2 ++ mm/pgtable-generic.c | 5 +++-- 8 files changed, 69 insertions(+), 7 deletions(-) (limited to 'include/linux') diff --git a/arch/sparc/include/asm/pgtable_64.h b/arch/sparc/include/asm/pgtable_64.h index 8358dc144959..0f9e94537eee 100644 --- a/arch/sparc/include/asm/pgtable_64.h +++ b/arch/sparc/include/asm/pgtable_64.h @@ -619,7 +619,7 @@ static inline unsigned long pte_present(pte_t pte) } #define pte_accessible pte_accessible -static inline unsigned long pte_accessible(pte_t a) +static inline unsigned long pte_accessible(struct mm_struct *mm, pte_t a) { return pte_val(a) & _PAGE_VALID; } @@ -847,7 +847,7 @@ static inline void __set_pte_at(struct mm_struct *mm, unsigned long addr, * SUN4V NOTE: _PAGE_VALID is the same value in both the SUN4U * and SUN4V pte layout, so this inline test is fine. */ - if (likely(mm != &init_mm) && pte_accessible(orig)) + if (likely(mm != &init_mm) && pte_accessible(mm, orig)) tlb_batch_add(mm, addr, ptep, orig, fullmm); } diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h index 3d1999458709..bbc8b12fa443 100644 --- a/arch/x86/include/asm/pgtable.h +++ b/arch/x86/include/asm/pgtable.h @@ -452,9 +452,16 @@ static inline int pte_present(pte_t a) } #define pte_accessible pte_accessible -static inline int pte_accessible(pte_t a) +static inline bool pte_accessible(struct mm_struct *mm, pte_t a) { - return pte_flags(a) & _PAGE_PRESENT; + if (pte_flags(a) & _PAGE_PRESENT) + return true; + + if ((pte_flags(a) & (_PAGE_PROTNONE | _PAGE_NUMA)) && + mm_tlb_flush_pending(mm)) + return true; + + return false; } static inline int pte_hidden(pte_t pte) diff --git a/include/asm-generic/pgtable.h b/include/asm-generic/pgtable.h index f330d28e4d0e..b12079afbd5f 100644 --- a/include/asm-generic/pgtable.h +++ b/include/asm-generic/pgtable.h @@ -217,7 +217,7 @@ static inline int pmd_same(pmd_t pmd_a, pmd_t pmd_b) #endif #ifndef pte_accessible -# define pte_accessible(pte) ((void)(pte),1) +# define pte_accessible(mm, pte) ((void)(pte), 1) #endif #ifndef flush_tlb_fix_spurious_fault diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index bd299418a934..e5c49c30460f 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -442,6 +442,14 @@ struct mm_struct { /* numa_scan_seq prevents two threads setting pte_numa */ int numa_scan_seq; +#endif +#if defined(CONFIG_NUMA_BALANCING) || defined(CONFIG_COMPACTION) + /* + * An operation with batched TLB flushing is going on. Anything that + * can move process memory needs to flush the TLB when moving a + * PROT_NONE or PROT_NUMA mapped page. + */ + bool tlb_flush_pending; #endif struct uprobes_state uprobes_state; }; @@ -459,4 +467,40 @@ static inline cpumask_t *mm_cpumask(struct mm_struct *mm) return mm->cpu_vm_mask_var; } +#if defined(CONFIG_NUMA_BALANCING) || defined(CONFIG_COMPACTION) +/* + * Memory barriers to keep this state in sync are graciously provided by + * the page table locks, outside of which no page table modifications happen. + * The barriers below prevent the compiler from re-ordering the instructions + * around the memory barriers that are already present in the code. + */ +static inline bool mm_tlb_flush_pending(struct mm_struct *mm) +{ + barrier(); + return mm->tlb_flush_pending; +} +static inline void set_tlb_flush_pending(struct mm_struct *mm) +{ + mm->tlb_flush_pending = true; + barrier(); +} +/* Clearing is done after a TLB flush, which also provides a barrier. */ +static inline void clear_tlb_flush_pending(struct mm_struct *mm) +{ + barrier(); + mm->tlb_flush_pending = false; +} +#else +static inline bool mm_tlb_flush_pending(struct mm_struct *mm) +{ + return false; +} +static inline void set_tlb_flush_pending(struct mm_struct *mm) +{ +} +static inline void clear_tlb_flush_pending(struct mm_struct *mm) +{ +} +#endif + #endif /* _LINUX_MM_TYPES_H */ diff --git a/kernel/fork.c b/kernel/fork.c index 728d5be9548c..5721f0e3f2da 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -537,6 +537,7 @@ static struct mm_struct *mm_init(struct mm_struct *mm, struct task_struct *p) spin_lock_init(&mm->page_table_lock); mm_init_aio(mm); mm_init_owner(mm, p); + clear_tlb_flush_pending(mm); if (likely(!mm_alloc_pgd(mm))) { mm->def_flags = 0; diff --git a/mm/huge_memory.c b/mm/huge_memory.c index 7de1bf85f683..3d2783e10596 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -1376,6 +1376,13 @@ int do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_struct *vma, goto clear_pmdnuma; } + /* + * The page_table_lock above provides a memory barrier + * with change_protection_range. + */ + if (mm_tlb_flush_pending(mm)) + flush_tlb_range(vma, haddr, haddr + HPAGE_PMD_SIZE); + /* * Migrate the THP to the requested node, returns with page unlocked * and pmd_numa cleared. diff --git a/mm/mprotect.c b/mm/mprotect.c index f8421722acb9..bb53a6591aea 100644 --- a/mm/mprotect.c +++ b/mm/mprotect.c @@ -188,6 +188,7 @@ static unsigned long change_protection_range(struct vm_area_struct *vma, BUG_ON(addr >= end); pgd = pgd_offset(mm, addr); flush_cache_range(vma, addr, end); + set_tlb_flush_pending(mm); do { next = pgd_addr_end(addr, end); if (pgd_none_or_clear_bad(pgd)) @@ -199,6 +200,7 @@ static unsigned long change_protection_range(struct vm_area_struct *vma, /* Only flush the TLB if we actually modified any entries: */ if (pages) flush_tlb_range(vma, start, end); + clear_tlb_flush_pending(mm); return pages; } diff --git a/mm/pgtable-generic.c b/mm/pgtable-generic.c index e84cad27a801..a8b919925934 100644 --- a/mm/pgtable-generic.c +++ b/mm/pgtable-generic.c @@ -110,9 +110,10 @@ int pmdp_clear_flush_young(struct vm_area_struct *vma, pte_t ptep_clear_flush(struct vm_area_struct *vma, unsigned long address, pte_t *ptep) { + struct mm_struct *mm = (vma)->vm_mm; pte_t pte; - pte = ptep_get_and_clear((vma)->vm_mm, address, ptep); - if (pte_accessible(pte)) + pte = ptep_get_and_clear(mm, address, ptep); + if (pte_accessible(mm, pte)) flush_tlb_page(vma, address); return pte; } -- cgit v1.2.3 From af2c1401e6f9177483be4fad876d0073669df9df Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Wed, 18 Dec 2013 17:08:45 -0800 Subject: mm: numa: guarantee that tlb_flush_pending updates are visible before page table updates According to documentation on barriers, stores issued before a LOCK can complete after the lock implying that it's possible tlb_flush_pending can be visible after a page table update. As per revised documentation, this patch adds a smp_mb__before_spinlock to guarantee the correct ordering. Signed-off-by: Mel Gorman Acked-by: Paul E. McKenney Reviewed-by: Rik van Riel Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/mm_types.h | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index e5c49c30460f..ad0616f2fe2c 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -482,7 +482,12 @@ static inline bool mm_tlb_flush_pending(struct mm_struct *mm) static inline void set_tlb_flush_pending(struct mm_struct *mm) { mm->tlb_flush_pending = true; - barrier(); + + /* + * Guarantee that the tlb_flush_pending store does not leak into the + * critical section updating the page tables + */ + smp_mb__before_spinlock(); } /* Clearing is done after a TLB flush, which also provides a barrier. */ static inline void clear_tlb_flush_pending(struct mm_struct *mm) -- cgit v1.2.3 From 597d795a2a786d22dd872332428e2b9439ede639 Mon Sep 17 00:00:00 2001 From: "Kirill A. Shutemov" Date: Fri, 20 Dec 2013 13:35:58 +0200 Subject: mm: do not allocate page->ptl dynamically, if spinlock_t fits to long In struct page we have enough space to fit long-size page->ptl there, but we use dynamically-allocated page->ptl if size(spinlock_t) is larger than sizeof(int). It hurts 64-bit architectures with CONFIG_GENERIC_LOCKBREAK, where sizeof(spinlock_t) == 8, but it easily fits into struct page. Signed-off-by: Kirill A. Shutemov Acked-by: Hugh Dickins Signed-off-by: Linus Torvalds --- include/linux/lockref.h | 2 +- include/linux/mm.h | 6 +++--- include/linux/mm_types.h | 3 ++- kernel/bounds.c | 2 +- mm/memory.c | 2 +- 5 files changed, 8 insertions(+), 7 deletions(-) (limited to 'include/linux') diff --git a/include/linux/lockref.h b/include/linux/lockref.h index c8929c3832db..4bfde0e99ed5 100644 --- a/include/linux/lockref.h +++ b/include/linux/lockref.h @@ -19,7 +19,7 @@ #define USE_CMPXCHG_LOCKREF \ (IS_ENABLED(CONFIG_ARCH_USE_CMPXCHG_LOCKREF) && \ - IS_ENABLED(CONFIG_SMP) && !BLOATED_SPINLOCKS) + IS_ENABLED(CONFIG_SMP) && SPINLOCK_SIZE <= 4) struct lockref { union { diff --git a/include/linux/mm.h b/include/linux/mm.h index 1cedd000cf29..35527173cf50 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -1317,7 +1317,7 @@ static inline pmd_t *pmd_alloc(struct mm_struct *mm, pud_t *pud, unsigned long a #endif /* CONFIG_MMU && !__ARCH_HAS_4LEVEL_HACK */ #if USE_SPLIT_PTE_PTLOCKS -#if BLOATED_SPINLOCKS +#if ALLOC_SPLIT_PTLOCKS extern bool ptlock_alloc(struct page *page); extern void ptlock_free(struct page *page); @@ -1325,7 +1325,7 @@ static inline spinlock_t *ptlock_ptr(struct page *page) { return page->ptl; } -#else /* BLOATED_SPINLOCKS */ +#else /* ALLOC_SPLIT_PTLOCKS */ static inline bool ptlock_alloc(struct page *page) { return true; @@ -1339,7 +1339,7 @@ static inline spinlock_t *ptlock_ptr(struct page *page) { return &page->ptl; } -#endif /* BLOATED_SPINLOCKS */ +#endif /* ALLOC_SPLIT_PTLOCKS */ static inline spinlock_t *pte_lockptr(struct mm_struct *mm, pmd_t *pmd) { diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index ad0616f2fe2c..290901a8c1de 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -26,6 +26,7 @@ struct address_space; #define USE_SPLIT_PTE_PTLOCKS (NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS) #define USE_SPLIT_PMD_PTLOCKS (USE_SPLIT_PTE_PTLOCKS && \ IS_ENABLED(CONFIG_ARCH_ENABLE_SPLIT_PMD_PTLOCK)) +#define ALLOC_SPLIT_PTLOCKS (SPINLOCK_SIZE > BITS_PER_LONG/8) /* * Each physical page in the system has a struct page associated with @@ -155,7 +156,7 @@ struct page { * system if PG_buddy is set. */ #if USE_SPLIT_PTE_PTLOCKS -#if BLOATED_SPINLOCKS +#if ALLOC_SPLIT_PTLOCKS spinlock_t *ptl; #else spinlock_t ptl; diff --git a/kernel/bounds.c b/kernel/bounds.c index 5253204afdca..9fd4246b04b8 100644 --- a/kernel/bounds.c +++ b/kernel/bounds.c @@ -22,6 +22,6 @@ void foo(void) #ifdef CONFIG_SMP DEFINE(NR_CPUS_BITS, ilog2(CONFIG_NR_CPUS)); #endif - DEFINE(BLOATED_SPINLOCKS, sizeof(spinlock_t) > sizeof(int)); + DEFINE(SPINLOCK_SIZE, sizeof(spinlock_t)); /* End of constants */ } diff --git a/mm/memory.c b/mm/memory.c index 5d9025f3b3e1..b6e211b779d0 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -4271,7 +4271,7 @@ void copy_user_huge_page(struct page *dst, struct page *src, } #endif /* CONFIG_TRANSPARENT_HUGEPAGE || CONFIG_HUGETLBFS */ -#if USE_SPLIT_PTE_PTLOCKS && BLOATED_SPINLOCKS +#if ALLOC_SPLIT_PTLOCKS bool ptlock_alloc(struct page *page) { spinlock_t *ptl; -- cgit v1.2.3 From df36ac1bc2a166eef90785d584e4cfed6f52bd32 Mon Sep 17 00:00:00 2001 From: "Luck, Tony" Date: Wed, 18 Dec 2013 15:17:10 -0800 Subject: pstore: Don't allow high traffic options on fragile devices Some pstore backing devices use on board flash as persistent storage. These have limited numbers of write cycles so it is a poor idea to use them from high frequency operations. Signed-off-by: Tony Luck Signed-off-by: Linus Torvalds --- drivers/acpi/apei/erst.c | 1 + drivers/firmware/efi/efi-pstore.c | 1 + fs/pstore/platform.c | 7 +++++-- include/linux/pstore.h | 3 +++ 4 files changed, 10 insertions(+), 2 deletions(-) (limited to 'include/linux') diff --git a/drivers/acpi/apei/erst.c b/drivers/acpi/apei/erst.c index 26311f23c824..cb1d557fc22c 100644 --- a/drivers/acpi/apei/erst.c +++ b/drivers/acpi/apei/erst.c @@ -942,6 +942,7 @@ static int erst_clearer(enum pstore_type_id type, u64 id, int count, static struct pstore_info erst_info = { .owner = THIS_MODULE, .name = "erst", + .flags = PSTORE_FLAGS_FRAGILE, .open = erst_open_pstore, .close = erst_close_pstore, .read = erst_reader, diff --git a/drivers/firmware/efi/efi-pstore.c b/drivers/firmware/efi/efi-pstore.c index 743fd426f21b..4b9dc836dcf9 100644 --- a/drivers/firmware/efi/efi-pstore.c +++ b/drivers/firmware/efi/efi-pstore.c @@ -356,6 +356,7 @@ static int efi_pstore_erase(enum pstore_type_id type, u64 id, int count, static struct pstore_info efi_pstore_info = { .owner = THIS_MODULE, .name = "efi", + .flags = PSTORE_FLAGS_FRAGILE, .open = efi_pstore_open, .close = efi_pstore_close, .read = efi_pstore_read, diff --git a/fs/pstore/platform.c b/fs/pstore/platform.c index b8e93a40a5d3..78c3c2097787 100644 --- a/fs/pstore/platform.c +++ b/fs/pstore/platform.c @@ -443,8 +443,11 @@ int pstore_register(struct pstore_info *psi) pstore_get_records(0); kmsg_dump_register(&pstore_dumper); - pstore_register_console(); - pstore_register_ftrace(); + + if ((psi->flags & PSTORE_FLAGS_FRAGILE) == 0) { + pstore_register_console(); + pstore_register_ftrace(); + } if (pstore_update_ms >= 0) { pstore_timer.expires = jiffies + diff --git a/include/linux/pstore.h b/include/linux/pstore.h index abd437d0a8a7..ece0c6bbfcc5 100644 --- a/include/linux/pstore.h +++ b/include/linux/pstore.h @@ -51,6 +51,7 @@ struct pstore_info { char *buf; size_t bufsize; struct mutex read_mutex; /* serialize open/read/close */ + int flags; int (*open)(struct pstore_info *psi); int (*close)(struct pstore_info *psi); ssize_t (*read)(u64 *id, enum pstore_type_id *type, @@ -70,6 +71,8 @@ struct pstore_info { void *data; }; +#define PSTORE_FLAGS_FRAGILE 1 + #ifdef CONFIG_PSTORE extern int pstore_register(struct pstore_info *); extern bool pstore_cannot_block_path(enum kmsg_dump_reason reason); -- cgit v1.2.3 From 8e321fefb0e60bae4e2a28d20fc4fa30758d27c6 Mon Sep 17 00:00:00 2001 From: Benjamin LaHaise Date: Sat, 21 Dec 2013 17:56:08 -0500 Subject: aio/migratepages: make aio migrate pages sane The arbitrary restriction on page counts offered by the core migrate_page_move_mapping() code results in rather suspicious looking fiddling with page reference counts in the aio_migratepage() operation. To fix this, make migrate_page_move_mapping() take an extra_count parameter that allows aio to tell the code about its own reference count on the page being migrated. While cleaning up aio_migratepage(), make it validate that the old page being passed in is actually what aio_migratepage() expects to prevent misbehaviour in the case of races. Signed-off-by: Benjamin LaHaise --- fs/aio.c | 52 +++++++++++++++++++++++++++++++++++++++++-------- include/linux/migrate.h | 3 ++- mm/migrate.c | 13 +++++++------ 3 files changed, 53 insertions(+), 15 deletions(-) (limited to 'include/linux') diff --git a/fs/aio.c b/fs/aio.c index fd1c0baf15bb..efa708b29054 100644 --- a/fs/aio.c +++ b/fs/aio.c @@ -244,9 +244,14 @@ static void aio_free_ring(struct kioctx *ctx) int i; for (i = 0; i < ctx->nr_pages; i++) { + struct page *page; pr_debug("pid(%d) [%d] page->count=%d\n", current->pid, i, page_count(ctx->ring_pages[i])); - put_page(ctx->ring_pages[i]); + page = ctx->ring_pages[i]; + if (!page) + continue; + ctx->ring_pages[i] = NULL; + put_page(page); } put_aio_ring_file(ctx); @@ -280,18 +285,38 @@ static int aio_migratepage(struct address_space *mapping, struct page *new, unsigned long flags; int rc; + rc = 0; + + /* Make sure the old page hasn't already been changed */ + spin_lock(&mapping->private_lock); + ctx = mapping->private_data; + if (ctx) { + pgoff_t idx; + spin_lock_irqsave(&ctx->completion_lock, flags); + idx = old->index; + if (idx < (pgoff_t)ctx->nr_pages) { + if (ctx->ring_pages[idx] != old) + rc = -EAGAIN; + } else + rc = -EINVAL; + spin_unlock_irqrestore(&ctx->completion_lock, flags); + } else + rc = -EINVAL; + spin_unlock(&mapping->private_lock); + + if (rc != 0) + return rc; + /* Writeback must be complete */ BUG_ON(PageWriteback(old)); - put_page(old); + get_page(new); - rc = migrate_page_move_mapping(mapping, new, old, NULL, mode); + rc = migrate_page_move_mapping(mapping, new, old, NULL, mode, 1); if (rc != MIGRATEPAGE_SUCCESS) { - get_page(old); + put_page(new); return rc; } - get_page(new); - /* We can potentially race against kioctx teardown here. Use the * address_space's private data lock to protect the mapping's * private_data. @@ -303,13 +328,24 @@ static int aio_migratepage(struct address_space *mapping, struct page *new, spin_lock_irqsave(&ctx->completion_lock, flags); migrate_page_copy(new, old); idx = old->index; - if (idx < (pgoff_t)ctx->nr_pages) - ctx->ring_pages[idx] = new; + if (idx < (pgoff_t)ctx->nr_pages) { + /* And only do the move if things haven't changed */ + if (ctx->ring_pages[idx] == old) + ctx->ring_pages[idx] = new; + else + rc = -EAGAIN; + } else + rc = -EINVAL; spin_unlock_irqrestore(&ctx->completion_lock, flags); } else rc = -EBUSY; spin_unlock(&mapping->private_lock); + if (rc == MIGRATEPAGE_SUCCESS) + put_page(old); + else + put_page(new); + return rc; } #endif diff --git a/include/linux/migrate.h b/include/linux/migrate.h index b7717d74da7f..f015c059e159 100644 --- a/include/linux/migrate.h +++ b/include/linux/migrate.h @@ -55,7 +55,8 @@ extern int migrate_huge_page_move_mapping(struct address_space *mapping, struct page *newpage, struct page *page); extern int migrate_page_move_mapping(struct address_space *mapping, struct page *newpage, struct page *page, - struct buffer_head *head, enum migrate_mode mode); + struct buffer_head *head, enum migrate_mode mode, + int extra_count); #else static inline void putback_lru_pages(struct list_head *l) {} diff --git a/mm/migrate.c b/mm/migrate.c index e9b710201335..9194375b2307 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -317,14 +317,15 @@ static inline bool buffer_migrate_lock_buffers(struct buffer_head *head, */ int migrate_page_move_mapping(struct address_space *mapping, struct page *newpage, struct page *page, - struct buffer_head *head, enum migrate_mode mode) + struct buffer_head *head, enum migrate_mode mode, + int extra_count) { - int expected_count = 0; + int expected_count = 1 + extra_count; void **pslot; if (!mapping) { /* Anonymous page without mapping */ - if (page_count(page) != 1) + if (page_count(page) != expected_count) return -EAGAIN; return MIGRATEPAGE_SUCCESS; } @@ -334,7 +335,7 @@ int migrate_page_move_mapping(struct address_space *mapping, pslot = radix_tree_lookup_slot(&mapping->page_tree, page_index(page)); - expected_count = 2 + page_has_private(page); + expected_count += 1 + page_has_private(page); if (page_count(page) != expected_count || radix_tree_deref_slot_protected(pslot, &mapping->tree_lock) != page) { spin_unlock_irq(&mapping->tree_lock); @@ -584,7 +585,7 @@ int migrate_page(struct address_space *mapping, BUG_ON(PageWriteback(page)); /* Writeback must be complete */ - rc = migrate_page_move_mapping(mapping, newpage, page, NULL, mode); + rc = migrate_page_move_mapping(mapping, newpage, page, NULL, mode, 0); if (rc != MIGRATEPAGE_SUCCESS) return rc; @@ -611,7 +612,7 @@ int buffer_migrate_page(struct address_space *mapping, head = page_buffers(page); - rc = migrate_page_move_mapping(mapping, newpage, page, head, mode); + rc = migrate_page_move_mapping(mapping, newpage, page, head, mode, 0); if (rc != MIGRATEPAGE_SUCCESS) return rc; -- cgit v1.2.3