From 3d94ae0c70a71a9824479366775e2c7679a57d94 Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Wed, 28 Sep 2011 11:49:28 -0700 Subject: x86/cmpxchg: add a locked add() helper Mostly to remove some conditional code in spinlock.h. Signed-off-by: Jeremy Fitzhardinge --- arch/x86/include/asm/cmpxchg.h | 42 ++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 42 insertions(+) (limited to 'arch/x86/include/asm/cmpxchg.h') diff --git a/arch/x86/include/asm/cmpxchg.h b/arch/x86/include/asm/cmpxchg.h index 5d3acdf5a7a6..49eade13161c 100644 --- a/arch/x86/include/asm/cmpxchg.h +++ b/arch/x86/include/asm/cmpxchg.h @@ -14,6 +14,8 @@ extern void __cmpxchg_wrong_size(void) __compiletime_error("Bad argument size for cmpxchg"); extern void __xadd_wrong_size(void) __compiletime_error("Bad argument size for xadd"); +extern void __add_wrong_size(void) + __compiletime_error("Bad argument size for add"); /* * Constants for operation sizes. On 32-bit, the 64-bit size it set to @@ -207,4 +209,44 @@ extern void __xadd_wrong_size(void) #define xadd_sync(ptr, inc) __xadd((ptr), (inc), "lock; ") #define xadd_local(ptr, inc) __xadd((ptr), (inc), "") +#define __add(ptr, inc, lock) \ + ({ \ + __typeof__ (*(ptr)) __ret = (inc); \ + switch (sizeof(*(ptr))) { \ + case __X86_CASE_B: \ + asm volatile (lock "addb %b1, %0\n" \ + : "+m" (*(ptr)) : "ri" (inc) \ + : "memory", "cc"); \ + break; \ + case __X86_CASE_W: \ + asm volatile (lock "addw %w1, %0\n" \ + : "+m" (*(ptr)) : "ri" (inc) \ + : "memory", "cc"); \ + break; \ + case __X86_CASE_L: \ + asm volatile (lock "addl %1, %0\n" \ + : "+m" (*(ptr)) : "ri" (inc) \ + : "memory", "cc"); \ + break; \ + case __X86_CASE_Q: \ + asm volatile (lock "addq %1, %0\n" \ + : "+m" (*(ptr)) : "ri" (inc) \ + : "memory", "cc"); \ + break; \ + default: \ + __add_wrong_size(); \ + } \ + __ret; \ + }) + +/* + * add_*() adds "inc" to "*ptr" + * + * __add() takes a lock prefix + * add_smp() is locked when multiple CPUs are online + * add_sync() is always locked + */ +#define add_smp(ptr, inc) __add((ptr), (inc), LOCK_PREFIX) +#define add_sync(ptr, inc) __add((ptr), (inc), "lock; ") + #endif /* ASM_X86_CMPXCHG_H */ -- cgit v1.2.3 From 31a8394e069e47dc47f4c29e4213aa943342f19f Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Fri, 30 Sep 2011 12:14:10 -0700 Subject: x86: consolidate xchg and xadd macros They both have a basic "put new value in location, return old value" pattern, so they can use the same macro easily. Signed-off-by: Jeremy Fitzhardinge --- arch/x86/include/asm/cmpxchg.h | 114 +++++++++++++---------------------------- 1 file changed, 36 insertions(+), 78 deletions(-) (limited to 'arch/x86/include/asm/cmpxchg.h') diff --git a/arch/x86/include/asm/cmpxchg.h b/arch/x86/include/asm/cmpxchg.h index 49eade13161c..5488e10b9dba 100644 --- a/arch/x86/include/asm/cmpxchg.h +++ b/arch/x86/include/asm/cmpxchg.h @@ -33,60 +33,47 @@ extern void __add_wrong_size(void) #define __X86_CASE_Q -1 /* sizeof will never return -1 */ #endif +/* + * An exchange-type operation, which takes a value and a pointer, and + * returns a the old value. + */ +#define __xchg_op(ptr, arg, op, lock) \ + ({ \ + __typeof__ (*(ptr)) __ret = (arg); \ + switch (sizeof(*(ptr))) { \ + case __X86_CASE_B: \ + asm volatile (lock #op "b %b0, %1\n" \ + : "+r" (__ret), "+m" (*(ptr)) \ + : : "memory", "cc"); \ + break; \ + case __X86_CASE_W: \ + asm volatile (lock #op "w %w0, %1\n" \ + : "+r" (__ret), "+m" (*(ptr)) \ + : : "memory", "cc"); \ + break; \ + case __X86_CASE_L: \ + asm volatile (lock #op "l %0, %1\n" \ + : "+r" (__ret), "+m" (*(ptr)) \ + : : "memory", "cc"); \ + break; \ + case __X86_CASE_Q: \ + asm volatile (lock #op "q %q0, %1\n" \ + : "+r" (__ret), "+m" (*(ptr)) \ + : : "memory", "cc"); \ + break; \ + default: \ + __ ## op ## _wrong_size(); \ + } \ + __ret; \ + }) + /* * Note: no "lock" prefix even on SMP: xchg always implies lock anyway. * Since this is generally used to protect other memory information, we * use "asm volatile" and "memory" clobbers to prevent gcc from moving * information around. */ -#define __xchg(x, ptr, size) \ -({ \ - __typeof(*(ptr)) __x = (x); \ - switch (size) { \ - case __X86_CASE_B: \ - { \ - volatile u8 *__ptr = (volatile u8 *)(ptr); \ - asm volatile("xchgb %0,%1" \ - : "=q" (__x), "+m" (*__ptr) \ - : "0" (__x) \ - : "memory"); \ - break; \ - } \ - case __X86_CASE_W: \ - { \ - volatile u16 *__ptr = (volatile u16 *)(ptr); \ - asm volatile("xchgw %0,%1" \ - : "=r" (__x), "+m" (*__ptr) \ - : "0" (__x) \ - : "memory"); \ - break; \ - } \ - case __X86_CASE_L: \ - { \ - volatile u32 *__ptr = (volatile u32 *)(ptr); \ - asm volatile("xchgl %0,%1" \ - : "=r" (__x), "+m" (*__ptr) \ - : "0" (__x) \ - : "memory"); \ - break; \ - } \ - case __X86_CASE_Q: \ - { \ - volatile u64 *__ptr = (volatile u64 *)(ptr); \ - asm volatile("xchgq %0,%1" \ - : "=r" (__x), "+m" (*__ptr) \ - : "0" (__x) \ - : "memory"); \ - break; \ - } \ - default: \ - __xchg_wrong_size(); \ - } \ - __x; \ -}) - -#define xchg(ptr, v) \ - __xchg((v), (ptr), sizeof(*ptr)) +#define xchg(ptr, v) __xchg_op((ptr), (v), xchg, "") /* * Atomic compare and exchange. Compare OLD with MEM, if identical, @@ -167,36 +154,6 @@ extern void __add_wrong_size(void) __cmpxchg_local((ptr), (old), (new), sizeof(*ptr)) #endif -#define __xadd(ptr, inc, lock) \ - ({ \ - __typeof__ (*(ptr)) __ret = (inc); \ - switch (sizeof(*(ptr))) { \ - case __X86_CASE_B: \ - asm volatile (lock "xaddb %b0, %1\n" \ - : "+r" (__ret), "+m" (*(ptr)) \ - : : "memory", "cc"); \ - break; \ - case __X86_CASE_W: \ - asm volatile (lock "xaddw %w0, %1\n" \ - : "+r" (__ret), "+m" (*(ptr)) \ - : : "memory", "cc"); \ - break; \ - case __X86_CASE_L: \ - asm volatile (lock "xaddl %0, %1\n" \ - : "+r" (__ret), "+m" (*(ptr)) \ - : : "memory", "cc"); \ - break; \ - case __X86_CASE_Q: \ - asm volatile (lock "xaddq %q0, %1\n" \ - : "+r" (__ret), "+m" (*(ptr)) \ - : : "memory", "cc"); \ - break; \ - default: \ - __xadd_wrong_size(); \ - } \ - __ret; \ - }) - /* * xadd() adds "inc" to "*ptr" and atomically returns the previous * value of "*ptr". @@ -205,6 +162,7 @@ extern void __add_wrong_size(void) * xadd_sync() is always locked * xadd_local() is never locked */ +#define __xadd(ptr, inc, lock) __xchg_op((ptr), (inc), xadd, lock) #define xadd(ptr, inc) __xadd((ptr), (inc), LOCK_PREFIX) #define xadd_sync(ptr, inc) __xadd((ptr), (inc), "lock; ") #define xadd_local(ptr, inc) __xadd((ptr), (inc), "") -- cgit v1.2.3 From cdcd629869fabcd38ebd24a03b0a05ec1cbcafb0 Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Mon, 2 Jan 2012 17:02:18 +0000 Subject: x86: Fix and improve cmpxchg_double{,_local}() Just like the per-CPU ones they had several problems/shortcomings: Only the first memory operand was mentioned in the asm() operands, and the 2x64-bit version didn't have a memory clobber while the 2x32-bit one did. The former allowed the compiler to not recognize the need to re-load the data in case it had it cached in some register, while the latter was overly destructive. The types of the local copies of the old and new values were incorrect (the types of the pointed-to variables should be used here, to make sure the respective old/new variable types are compatible). The __dummy/__junk variables were pointless, given that local copies of the inputs already existed (and can hence be used for discarded outputs). The 32-bit variant of cmpxchg_double_local() referenced cmpxchg16b_local(). At once also: - change the return value type to what it really is: 'bool' - unify 32- and 64-bit variants - abstract out the common part of the 'normal' and 'local' variants Signed-off-by: Jan Beulich Cc: Christoph Lameter Cc: Linus Torvalds Cc: Andrew Morton Link: http://lkml.kernel.org/r/4F01F12A020000780006A19B@nat28.tlf.novell.com Signed-off-by: Ingo Molnar --- arch/x86/include/asm/cmpxchg.h | 23 ++++++++++++++++++++ arch/x86/include/asm/cmpxchg_32.h | 46 --------------------------------------- arch/x86/include/asm/cmpxchg_64.h | 43 ------------------------------------ mm/slub.c | 4 ++-- 4 files changed, 25 insertions(+), 91 deletions(-) (limited to 'arch/x86/include/asm/cmpxchg.h') diff --git a/arch/x86/include/asm/cmpxchg.h b/arch/x86/include/asm/cmpxchg.h index 5488e10b9dba..0c9fa2745f13 100644 --- a/arch/x86/include/asm/cmpxchg.h +++ b/arch/x86/include/asm/cmpxchg.h @@ -207,4 +207,27 @@ extern void __add_wrong_size(void) #define add_smp(ptr, inc) __add((ptr), (inc), LOCK_PREFIX) #define add_sync(ptr, inc) __add((ptr), (inc), "lock; ") +#define __cmpxchg_double(pfx, p1, p2, o1, o2, n1, n2) \ +({ \ + bool __ret; \ + __typeof__(*(p1)) __old1 = (o1), __new1 = (n1); \ + __typeof__(*(p2)) __old2 = (o2), __new2 = (n2); \ + BUILD_BUG_ON(sizeof(*(p1)) != sizeof(long)); \ + BUILD_BUG_ON(sizeof(*(p2)) != sizeof(long)); \ + VM_BUG_ON((unsigned long)(p1) % (2 * sizeof(long))); \ + VM_BUG_ON((unsigned long)((p1) + 1) != (unsigned long)(p2)); \ + asm volatile(pfx "cmpxchg%c4b %2; sete %0" \ + : "=a" (__ret), "+d" (__old2), \ + "+m" (*(p1)), "+m" (*(p2)) \ + : "i" (2 * sizeof(long)), "a" (__old1), \ + "b" (__new1), "c" (__new2)); \ + __ret; \ +}) + +#define cmpxchg_double(p1, p2, o1, o2, n1, n2) \ + __cmpxchg_double(LOCK_PREFIX, p1, p2, o1, o2, n1, n2) + +#define cmpxchg_double_local(p1, p2, o1, o2, n1, n2) \ + __cmpxchg_double(, p1, p2, o1, o2, n1, n2) + #endif /* ASM_X86_CMPXCHG_H */ diff --git a/arch/x86/include/asm/cmpxchg_32.h b/arch/x86/include/asm/cmpxchg_32.h index fbebb07dd80b..53f4b219336b 100644 --- a/arch/x86/include/asm/cmpxchg_32.h +++ b/arch/x86/include/asm/cmpxchg_32.h @@ -166,52 +166,6 @@ static inline unsigned long cmpxchg_386(volatile void *ptr, unsigned long old, #endif -#define cmpxchg8b(ptr, o1, o2, n1, n2) \ -({ \ - char __ret; \ - __typeof__(o2) __dummy; \ - __typeof__(*(ptr)) __old1 = (o1); \ - __typeof__(o2) __old2 = (o2); \ - __typeof__(*(ptr)) __new1 = (n1); \ - __typeof__(o2) __new2 = (n2); \ - asm volatile(LOCK_PREFIX "cmpxchg8b %2; setz %1" \ - : "=d"(__dummy), "=a" (__ret), "+m" (*ptr)\ - : "a" (__old1), "d"(__old2), \ - "b" (__new1), "c" (__new2) \ - : "memory"); \ - __ret; }) - - -#define cmpxchg8b_local(ptr, o1, o2, n1, n2) \ -({ \ - char __ret; \ - __typeof__(o2) __dummy; \ - __typeof__(*(ptr)) __old1 = (o1); \ - __typeof__(o2) __old2 = (o2); \ - __typeof__(*(ptr)) __new1 = (n1); \ - __typeof__(o2) __new2 = (n2); \ - asm volatile("cmpxchg8b %2; setz %1" \ - : "=d"(__dummy), "=a"(__ret), "+m" (*ptr)\ - : "a" (__old), "d"(__old2), \ - "b" (__new1), "c" (__new2), \ - : "memory"); \ - __ret; }) - - -#define cmpxchg_double(ptr, o1, o2, n1, n2) \ -({ \ - BUILD_BUG_ON(sizeof(*(ptr)) != 4); \ - VM_BUG_ON((unsigned long)(ptr) % 8); \ - cmpxchg8b((ptr), (o1), (o2), (n1), (n2)); \ -}) - -#define cmpxchg_double_local(ptr, o1, o2, n1, n2) \ -({ \ - BUILD_BUG_ON(sizeof(*(ptr)) != 4); \ - VM_BUG_ON((unsigned long)(ptr) % 8); \ - cmpxchg16b_local((ptr), (o1), (o2), (n1), (n2)); \ -}) - #define system_has_cmpxchg_double() cpu_has_cx8 #endif /* _ASM_X86_CMPXCHG_32_H */ diff --git a/arch/x86/include/asm/cmpxchg_64.h b/arch/x86/include/asm/cmpxchg_64.h index 285da02c38fa..614be87f1a9b 100644 --- a/arch/x86/include/asm/cmpxchg_64.h +++ b/arch/x86/include/asm/cmpxchg_64.h @@ -20,49 +20,6 @@ static inline void set_64bit(volatile u64 *ptr, u64 val) cmpxchg_local((ptr), (o), (n)); \ }) -#define cmpxchg16b(ptr, o1, o2, n1, n2) \ -({ \ - char __ret; \ - __typeof__(o2) __junk; \ - __typeof__(*(ptr)) __old1 = (o1); \ - __typeof__(o2) __old2 = (o2); \ - __typeof__(*(ptr)) __new1 = (n1); \ - __typeof__(o2) __new2 = (n2); \ - asm volatile(LOCK_PREFIX "cmpxchg16b %2;setz %1" \ - : "=d"(__junk), "=a"(__ret), "+m" (*ptr) \ - : "b"(__new1), "c"(__new2), \ - "a"(__old1), "d"(__old2)); \ - __ret; }) - - -#define cmpxchg16b_local(ptr, o1, o2, n1, n2) \ -({ \ - char __ret; \ - __typeof__(o2) __junk; \ - __typeof__(*(ptr)) __old1 = (o1); \ - __typeof__(o2) __old2 = (o2); \ - __typeof__(*(ptr)) __new1 = (n1); \ - __typeof__(o2) __new2 = (n2); \ - asm volatile("cmpxchg16b %2;setz %1" \ - : "=d"(__junk), "=a"(__ret), "+m" (*ptr) \ - : "b"(__new1), "c"(__new2), \ - "a"(__old1), "d"(__old2)); \ - __ret; }) - -#define cmpxchg_double(ptr, o1, o2, n1, n2) \ -({ \ - BUILD_BUG_ON(sizeof(*(ptr)) != 8); \ - VM_BUG_ON((unsigned long)(ptr) % 16); \ - cmpxchg16b((ptr), (o1), (o2), (n1), (n2)); \ -}) - -#define cmpxchg_double_local(ptr, o1, o2, n1, n2) \ -({ \ - BUILD_BUG_ON(sizeof(*(ptr)) != 8); \ - VM_BUG_ON((unsigned long)(ptr) % 16); \ - cmpxchg16b_local((ptr), (o1), (o2), (n1), (n2)); \ -}) - #define system_has_cmpxchg_double() cpu_has_cx16 #endif /* _ASM_X86_CMPXCHG_64_H */ diff --git a/mm/slub.c b/mm/slub.c index ed3334d9b6da..09ccee8fb58e 100644 --- a/mm/slub.c +++ b/mm/slub.c @@ -368,7 +368,7 @@ static inline bool __cmpxchg_double_slab(struct kmem_cache *s, struct page *page VM_BUG_ON(!irqs_disabled()); #ifdef CONFIG_CMPXCHG_DOUBLE if (s->flags & __CMPXCHG_DOUBLE) { - if (cmpxchg_double(&page->freelist, + if (cmpxchg_double(&page->freelist, &page->counters, freelist_old, counters_old, freelist_new, counters_new)) return 1; @@ -402,7 +402,7 @@ static inline bool cmpxchg_double_slab(struct kmem_cache *s, struct page *page, { #ifdef CONFIG_CMPXCHG_DOUBLE if (s->flags & __CMPXCHG_DOUBLE) { - if (cmpxchg_double(&page->freelist, + if (cmpxchg_double(&page->freelist, &page->counters, freelist_old, counters_old, freelist_new, counters_new)) return 1; -- cgit v1.2.3 From fc395b9291925b1880e0afc61274fe2f6ddc1269 Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Thu, 26 Jan 2012 15:47:37 +0000 Subject: x86: Properly parenthesize cmpxchg() macro arguments Quite oddly, all of the arguments passed through from the top level macros to the second level which didn't need parentheses had them, while the only expression (involving a parameter) needing them didn't. Very recently I got bitten by the lack thereof when using something like "array + index" for the first operand, with "array" being an array more narrow than int. Signed-off-by: Jan Beulich Cc: Linus Torvalds Link: http://lkml.kernel.org/r/4F2183A9020000780006F3E6@nat28.tlf.novell.com Signed-off-by: Ingo Molnar --- arch/x86/include/asm/cmpxchg.h | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'arch/x86/include/asm/cmpxchg.h') diff --git a/arch/x86/include/asm/cmpxchg.h b/arch/x86/include/asm/cmpxchg.h index 0c9fa2745f13..b3b733262909 100644 --- a/arch/x86/include/asm/cmpxchg.h +++ b/arch/x86/include/asm/cmpxchg.h @@ -145,13 +145,13 @@ extern void __add_wrong_size(void) #ifdef __HAVE_ARCH_CMPXCHG #define cmpxchg(ptr, old, new) \ - __cmpxchg((ptr), (old), (new), sizeof(*ptr)) + __cmpxchg(ptr, old, new, sizeof(*(ptr))) #define sync_cmpxchg(ptr, old, new) \ - __sync_cmpxchg((ptr), (old), (new), sizeof(*ptr)) + __sync_cmpxchg(ptr, old, new, sizeof(*(ptr))) #define cmpxchg_local(ptr, old, new) \ - __cmpxchg_local((ptr), (old), (new), sizeof(*ptr)) + __cmpxchg_local(ptr, old, new, sizeof(*(ptr))) #endif /* -- cgit v1.2.3