From 61ecdb801ef2cd28e32442383106d7837d76deac Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Tue, 26 Oct 2010 14:21:47 -0700 Subject: mm: strictly nested kmap_atomic() Ensure kmap_atomic() usage is strictly nested Signed-off-by: Peter Zijlstra Reviewed-by: Rik van Riel Acked-by: Chris Metcalf Cc: David Howells Cc: Hugh Dickins Cc: Ingo Molnar Cc: Thomas Gleixner Cc: "H. Peter Anvin" Cc: Steven Rostedt Cc: Russell King Cc: Ralf Baechle Cc: David Miller Cc: Paul Mackerras Cc: Benjamin Herrenschmidt Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/highmem.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'include/linux/highmem.h') diff --git a/include/linux/highmem.h b/include/linux/highmem.h index e3060ef85b6d..283cd47bb34c 100644 --- a/include/linux/highmem.h +++ b/include/linux/highmem.h @@ -201,8 +201,8 @@ static inline void copy_user_highpage(struct page *to, struct page *from, vfrom = kmap_atomic(from, KM_USER0); vto = kmap_atomic(to, KM_USER1); copy_user_page(vto, vfrom, vaddr, to); - kunmap_atomic(vfrom, KM_USER0); kunmap_atomic(vto, KM_USER1); + kunmap_atomic(vfrom, KM_USER0); } #endif @@ -214,8 +214,8 @@ static inline void copy_highpage(struct page *to, struct page *from) vfrom = kmap_atomic(from, KM_USER0); vto = kmap_atomic(to, KM_USER1); copy_page(vto, vfrom); - kunmap_atomic(vfrom, KM_USER0); kunmap_atomic(vto, KM_USER1); + kunmap_atomic(vfrom, KM_USER0); } #endif /* _LINUX_HIGHMEM_H */ -- cgit v1.2.2 From 3e4d3af501cccdc8a8cca41bdbe57d54ad7e7e73 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Tue, 26 Oct 2010 14:21:51 -0700 Subject: mm: stack based kmap_atomic() Keep the current interface but ignore the KM_type and use a stack based approach. The advantage is that we get rid of crappy code like: #define __KM_PTE \ (in_nmi() ? KM_NMI_PTE : \ in_irq() ? KM_IRQ_PTE : \ KM_PTE0) and in general can stop worrying about what context we're in and what kmap slots might be appropriate for that. The downside is that FRV kmap_atomic() gets more expensive. For now we use a CPP trick suggested by Andrew: #define kmap_atomic(page, args...) __kmap_atomic(page) to avoid having to touch all kmap_atomic() users in a single patch. [ not compiled on: - mn10300: the arch doesn't actually build with highmem to begin with ] [akpm@linux-foundation.org: coding-style fixes] [akpm@linux-foundation.org: fix up drivers/gpu/drm/i915/intel_overlay.c] Acked-by: Rik van Riel Signed-off-by: Peter Zijlstra Acked-by: Chris Metcalf Cc: David Howells Cc: Hugh Dickins Cc: Ingo Molnar Cc: Thomas Gleixner Cc: "H. Peter Anvin" Cc: Steven Rostedt Cc: Russell King Cc: Ralf Baechle Cc: David Miller Cc: Paul Mackerras Cc: Benjamin Herrenschmidt Cc: Dave Airlie Cc: Li Zefan Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/highmem.h | 61 +++++++++++++++++++++++++++++++------------------ 1 file changed, 39 insertions(+), 22 deletions(-) (limited to 'include/linux/highmem.h') diff --git a/include/linux/highmem.h b/include/linux/highmem.h index 283cd47bb34c..8a85ec109a3a 100644 --- a/include/linux/highmem.h +++ b/include/linux/highmem.h @@ -28,18 +28,6 @@ static inline void invalidate_kernel_vmap_range(void *vaddr, int size) #include -#ifdef CONFIG_DEBUG_HIGHMEM - -void debug_kmap_atomic(enum km_type type); - -#else - -static inline void debug_kmap_atomic(enum km_type type) -{ -} - -#endif - #ifdef CONFIG_HIGHMEM #include @@ -49,6 +37,27 @@ extern unsigned long totalhigh_pages; void kmap_flush_unused(void); +DECLARE_PER_CPU(int, __kmap_atomic_idx); + +static inline int kmap_atomic_idx_push(void) +{ + int idx = __get_cpu_var(__kmap_atomic_idx)++; +#ifdef CONFIG_DEBUG_HIGHMEM + WARN_ON_ONCE(in_irq() && !irqs_disabled()); + BUG_ON(idx > KM_TYPE_NR); +#endif + return idx; +} + +static inline int kmap_atomic_idx_pop(void) +{ + int idx = --__get_cpu_var(__kmap_atomic_idx); +#ifdef CONFIG_DEBUG_HIGHMEM + BUG_ON(idx < 0); +#endif + return idx; +} + #else /* CONFIG_HIGHMEM */ static inline unsigned int nr_free_highpages(void) { return 0; } @@ -66,19 +75,19 @@ static inline void kunmap(struct page *page) { } -static inline void *kmap_atomic(struct page *page, enum km_type idx) +static inline void *__kmap_atomic(struct page *page) { pagefault_disable(); return page_address(page); } -#define kmap_atomic_prot(page, idx, prot) kmap_atomic(page, idx) +#define kmap_atomic_prot(page, prot) __kmap_atomic(page) -static inline void kunmap_atomic_notypecheck(void *addr, enum km_type idx) +static inline void __kunmap_atomic(void *addr) { pagefault_enable(); } -#define kmap_atomic_pfn(pfn, idx) kmap_atomic(pfn_to_page(pfn), (idx)) +#define kmap_atomic_pfn(pfn) kmap_atomic(pfn_to_page(pfn)) #define kmap_atomic_to_page(ptr) virt_to_page(ptr) #define kmap_flush_unused() do {} while(0) @@ -86,12 +95,20 @@ static inline void kunmap_atomic_notypecheck(void *addr, enum km_type idx) #endif /* CONFIG_HIGHMEM */ -/* Prevent people trying to call kunmap_atomic() as if it were kunmap() */ -/* kunmap_atomic() should get the return value of kmap_atomic, not the page. */ -#define kunmap_atomic(addr, idx) do { \ - BUILD_BUG_ON(__same_type((addr), struct page *)); \ - kunmap_atomic_notypecheck((addr), (idx)); \ - } while (0) +/* + * Make both: kmap_atomic(page, idx) and kmap_atomic(page) work. + */ +#define kmap_atomic(page, args...) __kmap_atomic(page) + +/* + * Prevent people trying to call kunmap_atomic() as if it were kunmap() + * kunmap_atomic() should get the return value of kmap_atomic, not the page. + */ +#define kunmap_atomic(addr, args...) \ +do { \ + BUILD_BUG_ON(__same_type((addr), struct page *)); \ + __kunmap_atomic(addr); \ +} while (0) /* when CONFIG_HIGHMEM is not set these will be plain clear/copy_page */ #ifndef clear_user_highpage -- cgit v1.2.2 From a8e23a291852cd7c4fb5ca696dbb93912185ad10 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 27 Oct 2010 15:32:57 -0700 Subject: mm,x86: fix kmap_atomic_push vs ioremap_32.c It appears i386 uses kmap_atomic infrastructure regardless of CONFIG_HIGHMEM which results in a compile error when highmem is disabled. Cure this by providing the needed few bits for both CONFIG_HIGHMEM and CONFIG_X86_32. Signed-off-by: Peter Zijlstra Reported-by: Chris Wilson Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/highmem.h | 46 +++++++++++++++++++++++++--------------------- 1 file changed, 25 insertions(+), 21 deletions(-) (limited to 'include/linux/highmem.h') diff --git a/include/linux/highmem.h b/include/linux/highmem.h index 8a85ec109a3a..102f76be90da 100644 --- a/include/linux/highmem.h +++ b/include/linux/highmem.h @@ -37,27 +37,6 @@ extern unsigned long totalhigh_pages; void kmap_flush_unused(void); -DECLARE_PER_CPU(int, __kmap_atomic_idx); - -static inline int kmap_atomic_idx_push(void) -{ - int idx = __get_cpu_var(__kmap_atomic_idx)++; -#ifdef CONFIG_DEBUG_HIGHMEM - WARN_ON_ONCE(in_irq() && !irqs_disabled()); - BUG_ON(idx > KM_TYPE_NR); -#endif - return idx; -} - -static inline int kmap_atomic_idx_pop(void) -{ - int idx = --__get_cpu_var(__kmap_atomic_idx); -#ifdef CONFIG_DEBUG_HIGHMEM - BUG_ON(idx < 0); -#endif - return idx; -} - #else /* CONFIG_HIGHMEM */ static inline unsigned int nr_free_highpages(void) { return 0; } @@ -95,6 +74,31 @@ static inline void __kunmap_atomic(void *addr) #endif /* CONFIG_HIGHMEM */ +#if defined(CONFIG_HIGHMEM) || defined(CONFIG_X86_32) + +DECLARE_PER_CPU(int, __kmap_atomic_idx); + +static inline int kmap_atomic_idx_push(void) +{ + int idx = __get_cpu_var(__kmap_atomic_idx)++; +#ifdef CONFIG_DEBUG_HIGHMEM + WARN_ON_ONCE(in_irq() && !irqs_disabled()); + BUG_ON(idx > KM_TYPE_NR); +#endif + return idx; +} + +static inline int kmap_atomic_idx_pop(void) +{ + int idx = --__get_cpu_var(__kmap_atomic_idx); +#ifdef CONFIG_DEBUG_HIGHMEM + BUG_ON(idx < 0); +#endif + return idx; +} + +#endif + /* * Make both: kmap_atomic(page, idx) and kmap_atomic(page) work. */ -- cgit v1.2.2 From 20273941f2129aa5a432796d98a276ed73d60782 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 27 Oct 2010 15:32:58 -0700 Subject: mm: fix race in kunmap_atomic() Christoph reported a nice splat which illustrated a race in the new stack based kmap_atomic implementation. The problem is that we pop our stack slot before we're completely done resetting its state -- in particular clearing the PTE (sometimes that's CONFIG_DEBUG_HIGHMEM). If an interrupt happens before we actually clear the PTE used for the last slot, that interrupt can reuse the slot in a dirty state, which triggers a BUG in kmap_atomic(). Fix this by introducing kmap_atomic_idx() which reports the current slot index without actually releasing it and use that to find the PTE and delay the _pop() until after we're completely done. Signed-off-by: Peter Zijlstra Reported-by: Christoph Hellwig Acked-by: Rik van Riel Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/highmem.h | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'include/linux/highmem.h') diff --git a/include/linux/highmem.h b/include/linux/highmem.h index 102f76be90da..e9138198e823 100644 --- a/include/linux/highmem.h +++ b/include/linux/highmem.h @@ -88,6 +88,11 @@ static inline int kmap_atomic_idx_push(void) return idx; } +static inline int kmap_atomic_idx(void) +{ + return __get_cpu_var(__kmap_atomic_idx) - 1; +} + static inline int kmap_atomic_idx_pop(void) { int idx = --__get_cpu_var(__kmap_atomic_idx); -- cgit v1.2.2 From 43b3a0c732776746f53d8ed2ba659583fc1692aa Mon Sep 17 00:00:00 2001 From: Catalin Marinas Date: Thu, 11 Nov 2010 14:05:10 -0800 Subject: include/linux/highmem.h needs hardirq.h Commit 3e4d3af501cc ("mm: stack based kmap_atomic()") introduced the kmap_atomic_idx_push() function which warns on in_irq() with CONFIG_DEBUG_HIGHMEM enabled. This patch includes linux/hardirq.h for the in_irq definition. Signed-off-by: Catalin Marinas Acked-by: Peter Zijlstra Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/highmem.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include/linux/highmem.h') diff --git a/include/linux/highmem.h b/include/linux/highmem.h index e9138198e823..b676c585574e 100644 --- a/include/linux/highmem.h +++ b/include/linux/highmem.h @@ -5,6 +5,7 @@ #include #include #include +#include #include -- cgit v1.2.2 From cfb824349556904b319464139be5c75fce983b0d Mon Sep 17 00:00:00 2001 From: Christoph Lameter Date: Mon, 6 Dec 2010 11:40:03 -0600 Subject: highmem: Use this_cpu_xx_return() operations Use this_cpu operations to optimize access primitives for highmem. The main effect is the avoidance of address calculations through the use of a segment prefix. V3->V4 - kmap_atomic_idx: Do not return a value. - Use __this_cpu_dec without HIGHMEM_DEBUG Cc: Peter Zijlstra Cc: Catalin Marinas Acked-by: H. Peter Anvin Signed-off-by: Christoph Lameter Signed-off-by: Tejun Heo --- include/linux/highmem.h | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) (limited to 'include/linux/highmem.h') diff --git a/include/linux/highmem.h b/include/linux/highmem.h index b676c585574e..3a93f73a8acc 100644 --- a/include/linux/highmem.h +++ b/include/linux/highmem.h @@ -81,7 +81,8 @@ DECLARE_PER_CPU(int, __kmap_atomic_idx); static inline int kmap_atomic_idx_push(void) { - int idx = __get_cpu_var(__kmap_atomic_idx)++; + int idx = __this_cpu_inc_return(__kmap_atomic_idx) - 1; + #ifdef CONFIG_DEBUG_HIGHMEM WARN_ON_ONCE(in_irq() && !irqs_disabled()); BUG_ON(idx > KM_TYPE_NR); @@ -91,16 +92,18 @@ static inline int kmap_atomic_idx_push(void) static inline int kmap_atomic_idx(void) { - return __get_cpu_var(__kmap_atomic_idx) - 1; + return __this_cpu_read(__kmap_atomic_idx) - 1; } -static inline int kmap_atomic_idx_pop(void) +static inline void kmap_atomic_idx_pop(void) { - int idx = --__get_cpu_var(__kmap_atomic_idx); #ifdef CONFIG_DEBUG_HIGHMEM + int idx = __this_cpu_dec_return(__kmap_atomic_idx); + BUG_ON(idx < 0); +#else + __this_cpu_dec(__kmap_atomic_idx); #endif - return idx; } #endif -- cgit v1.2.2