From a741e67969577163a4cfc78d7fd2753219087ef1 Mon Sep 17 00:00:00 2001 From: Benjamin Herrenschmidt Date: Tue, 10 Apr 2007 17:09:37 +1000 Subject: [POWERPC] Make tlb flush batch use lazy MMU mode The current tlb flush code on powerpc 64 bits has a subtle race since we lost the page table lock due to the possible faulting in of new PTEs after a previous one has been removed but before the corresponding hash entry has been evicted, which can leads to all sort of fatal problems. This patch reworks the batch code completely. It doesn't use the mmu_gather stuff anymore. Instead, we use the lazy mmu hooks that were added by the paravirt code. They have the nice property that the enter/leave lazy mmu mode pair is always fully contained by the PTE lock for a given range of PTEs. Thus we can guarantee that all batches are flushed on a given CPU before it drops that lock. We also generalize batching for any PTE update that require a flush. Batching is now enabled on a CPU by arch_enter_lazy_mmu_mode() and disabled by arch_leave_lazy_mmu_mode(). The code epects that this is always contained within a PTE lock section so no preemption can happen and no PTE insertion in that range from another CPU. When batching is enabled on a CPU, every PTE updates that need a hash flush will use the batch for that flush. Signed-off-by: Benjamin Herrenschmidt Signed-off-by: Paul Mackerras --- include/asm-powerpc/tlbflush.h | 39 +++++++++++++++++++++++++-------------- 1 file changed, 25 insertions(+), 14 deletions(-) (limited to 'include/asm-powerpc/tlbflush.h') diff --git a/include/asm-powerpc/tlbflush.h b/include/asm-powerpc/tlbflush.h index 93c7d0c7230f..0bc5a5e506be 100644 --- a/include/asm-powerpc/tlbflush.h +++ b/include/asm-powerpc/tlbflush.h @@ -28,25 +28,41 @@ struct mm_struct; #define PPC64_TLB_BATCH_NR 192 struct ppc64_tlb_batch { - unsigned long index; - struct mm_struct *mm; - real_pte_t pte[PPC64_TLB_BATCH_NR]; - unsigned long vaddr[PPC64_TLB_BATCH_NR]; - unsigned int psize; + int active; + unsigned long index; + struct mm_struct *mm; + real_pte_t pte[PPC64_TLB_BATCH_NR]; + unsigned long vaddr[PPC64_TLB_BATCH_NR]; + unsigned int psize; }; DECLARE_PER_CPU(struct ppc64_tlb_batch, ppc64_tlb_batch); extern void __flush_tlb_pending(struct ppc64_tlb_batch *batch); -static inline void flush_tlb_pending(void) +extern void hpte_need_flush(struct mm_struct *mm, unsigned long addr, + pte_t *ptep, unsigned long pte, int huge); + +#define __HAVE_ARCH_ENTER_LAZY_MMU_MODE + +static inline void arch_enter_lazy_mmu_mode(void) +{ + struct ppc64_tlb_batch *batch = &__get_cpu_var(ppc64_tlb_batch); + + batch->active = 1; +} + +static inline void arch_leave_lazy_mmu_mode(void) { - struct ppc64_tlb_batch *batch = &get_cpu_var(ppc64_tlb_batch); + struct ppc64_tlb_batch *batch = &__get_cpu_var(ppc64_tlb_batch); if (batch->index) __flush_tlb_pending(batch); - put_cpu_var(ppc64_tlb_batch); + batch->active = 0; } +#define arch_flush_lazy_mmu_mode() do {} while (0) + + extern void flush_hash_page(unsigned long va, real_pte_t pte, int psize, int local); extern void flush_hash_range(unsigned long number, int local); @@ -88,15 +104,12 @@ extern void update_mmu_cache(struct vm_area_struct *, unsigned long, pte_t); static inline void flush_tlb_mm(struct mm_struct *mm) { - flush_tlb_pending(); } static inline void flush_tlb_page(struct vm_area_struct *vma, unsigned long vmaddr) { -#ifdef CONFIG_PPC64 - flush_tlb_pending(); -#else +#ifndef CONFIG_PPC64 _tlbie(vmaddr); #endif } @@ -112,13 +125,11 @@ static inline void flush_tlb_page_nohash(struct vm_area_struct *vma, static inline void flush_tlb_range(struct vm_area_struct *vma, unsigned long start, unsigned long end) { - flush_tlb_pending(); } static inline void flush_tlb_kernel_range(unsigned long start, unsigned long end) { - flush_tlb_pending(); } #else /* 6xx, 7xx, 7xxx cpus */ -- cgit v1.2.1 From 621023072524fc0155ed16490255e1ea3aa11585 Mon Sep 17 00:00:00 2001 From: David Gibson Date: Tue, 24 Apr 2007 13:09:12 +1000 Subject: [POWERPC] Cleanup and fix breakage in tlbflush.h BenH's commit a741e67969577163a4cfc78d7fd2753219087ef1 in powerpc.git, although (AFAICT) only intended to affect ppc64, also has side-effects which break 44x. I think 40x, 8xx and Freescale Book E are also affected, though I haven't tested them. The problem lies in unconditionally removing flush_tlb_pending() from the versions of flush_tlb_mm(), flush_tlb_range() and flush_tlb_kernel_range() used on ppc64 - which are also used the embedded platforms mentioned above. The patch below cleans up the convoluted #ifdef logic in tlbflush.h, in the process restoring the necessary flushes for the software TLB platforms. There are three sets of definitions for the flushing hooks: the software TLB versions (revised to avoid using names which appear to related to TLB batching), the 32-bit hash based versions (external functions) amd the 64-bit hash based versions (which implement batching). It also moves the declaration of update_mmu_cache() to always be in tlbflush.h (previously it was in tlbflush.h except for PPC64, where it was in pgtable.h). Booted on Ebony (440GP) and compiled for 64-bit and 32-bit multiplatform. Signed-off-by: David Gibson Acked-by: Benjamin Herrenschmidt Signed-off-by: Paul Mackerras --- include/asm-powerpc/tlbflush.h | 135 ++++++++++++++++++++++++----------------- 1 file changed, 80 insertions(+), 55 deletions(-) (limited to 'include/asm-powerpc/tlbflush.h') diff --git a/include/asm-powerpc/tlbflush.h b/include/asm-powerpc/tlbflush.h index 0bc5a5e506be..86e6266a028b 100644 --- a/include/asm-powerpc/tlbflush.h +++ b/include/asm-powerpc/tlbflush.h @@ -17,10 +17,73 @@ */ #ifdef __KERNEL__ - struct mm_struct; +struct vm_area_struct; + +#if defined(CONFIG_4xx) || defined(CONFIG_8xx) || defined(CONFIG_FSL_BOOKE) +/* + * TLB flushing for software loaded TLB chips + * + * TODO: (CONFIG_FSL_BOOKE) determine if flush_tlb_range & + * flush_tlb_kernel_range are best implemented as tlbia vs + * specific tlbie's + */ + +extern void _tlbie(unsigned long address); + +#if defined(CONFIG_40x) || defined(CONFIG_8xx) +#define _tlbia() asm volatile ("tlbia; sync" : : : "memory") +#else /* CONFIG_44x || CONFIG_FSL_BOOKE */ +extern void _tlbia(void); +#endif -#ifdef CONFIG_PPC64 +static inline void flush_tlb_mm(struct mm_struct *mm) +{ + _tlbia(); +} + +static inline void flush_tlb_page(struct vm_area_struct *vma, + unsigned long vmaddr) +{ + _tlbie(vmaddr); +} + +static inline void flush_tlb_page_nohash(struct vm_area_struct *vma, + unsigned long vmaddr) +{ + _tlbie(vmaddr); +} + +static inline void flush_tlb_range(struct vm_area_struct *vma, + unsigned long start, unsigned long end) +{ + _tlbia(); +} + +static inline void flush_tlb_kernel_range(unsigned long start, + unsigned long end) +{ + _tlbia(); +} + +#elif defined(CONFIG_PPC32) +/* + * TLB flushing for "classic" hash-MMMU 32-bit CPUs, 6xx, 7xx, 7xxx + */ +extern void _tlbie(unsigned long address); +extern void _tlbia(void); + +extern void flush_tlb_mm(struct mm_struct *mm); +extern void flush_tlb_page(struct vm_area_struct *vma, unsigned long vmaddr); +extern void flush_tlb_page_nohash(struct vm_area_struct *vma, unsigned long addr); +extern void flush_tlb_range(struct vm_area_struct *vma, unsigned long start, + unsigned long end); +extern void flush_tlb_kernel_range(unsigned long start, unsigned long end); + +#else +/* + * TLB flushing for 64-bit has-MMU CPUs + */ #include #include @@ -67,89 +130,51 @@ extern void flush_hash_page(unsigned long va, real_pte_t pte, int psize, int local); extern void flush_hash_range(unsigned long number, int local); -#else /* CONFIG_PPC64 */ - -#include - -extern void _tlbie(unsigned long address); -extern void _tlbia(void); - -/* - * TODO: (CONFIG_FSL_BOOKE) determine if flush_tlb_range & - * flush_tlb_kernel_range are best implemented as tlbia vs - * specific tlbie's - */ - -#if (defined(CONFIG_4xx) && !defined(CONFIG_44x)) || defined(CONFIG_8xx) -#define flush_tlb_pending() asm volatile ("tlbia; sync" : : : "memory") -#elif defined(CONFIG_4xx) || defined(CONFIG_FSL_BOOKE) -#define flush_tlb_pending() _tlbia() -#endif - -/* - * This gets called at the end of handling a page fault, when - * the kernel has put a new PTE into the page table for the process. - * We use it to ensure coherency between the i-cache and d-cache - * for the page which has just been mapped in. - * On machines which use an MMU hash table, we use this to put a - * corresponding HPTE into the hash table ahead of time, instead of - * waiting for the inevitable extra hash-table miss exception. - */ -extern void update_mmu_cache(struct vm_area_struct *, unsigned long, pte_t); - -#endif /* CONFIG_PPC64 */ - -#if defined(CONFIG_PPC64) || defined(CONFIG_4xx) || \ - defined(CONFIG_FSL_BOOKE) || defined(CONFIG_8xx) static inline void flush_tlb_mm(struct mm_struct *mm) { } static inline void flush_tlb_page(struct vm_area_struct *vma, - unsigned long vmaddr) + unsigned long vmaddr) { -#ifndef CONFIG_PPC64 - _tlbie(vmaddr); -#endif } static inline void flush_tlb_page_nohash(struct vm_area_struct *vma, unsigned long vmaddr) { -#ifndef CONFIG_PPC64 - _tlbie(vmaddr); -#endif } static inline void flush_tlb_range(struct vm_area_struct *vma, - unsigned long start, unsigned long end) + unsigned long start, unsigned long end) { } static inline void flush_tlb_kernel_range(unsigned long start, - unsigned long end) + unsigned long end) { } -#else /* 6xx, 7xx, 7xxx cpus */ - -extern void flush_tlb_mm(struct mm_struct *mm); -extern void flush_tlb_page(struct vm_area_struct *vma, unsigned long vmaddr); -extern void flush_tlb_page_nohash(struct vm_area_struct *vma, unsigned long addr); -extern void flush_tlb_range(struct vm_area_struct *vma, unsigned long start, - unsigned long end); -extern void flush_tlb_kernel_range(unsigned long start, unsigned long end); - #endif +/* + * This gets called at the end of handling a page fault, when + * the kernel has put a new PTE into the page table for the process. + * We use it to ensure coherency between the i-cache and d-cache + * for the page which has just been mapped in. + * On machines which use an MMU hash table, we use this to put a + * corresponding HPTE into the hash table ahead of time, instead of + * waiting for the inevitable extra hash-table miss exception. + */ +extern void update_mmu_cache(struct vm_area_struct *, unsigned long, pte_t); + /* * This is called in munmap when we have freed up some page-table * pages. We don't need to do anything here, there's nothing special * about our page-table pages. -- paulus */ static inline void flush_tlb_pgtables(struct mm_struct *mm, - unsigned long start, unsigned long end) + unsigned long start, unsigned long end) { } -- cgit v1.2.1