From f7d71f2052555ae57b47322f2c2f6c29ff2438ae Mon Sep 17 00:00:00 2001 From: Waiman Long Date: Fri, 19 Jun 2015 11:50:00 -0400 Subject: locking/qrwlock: Rename functions to queued_*() To sync up with the naming convention used in qspinlock, all the qrwlock functions were renamed to started with "queued" instead of "queue". Signed-off-by: Waiman Long Signed-off-by: Peter Zijlstra (Intel) Cc: Arnd Bergmann Cc: Douglas Hatch Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Scott J Norton Cc: Thomas Gleixner Cc: Will Deacon Link: http://lkml.kernel.org/r/1434729002-57724-2-git-send-email-Waiman.Long@hp.com Signed-off-by: Ingo Molnar --- include/asm-generic/qrwlock.h | 58 +++++++++++++++++++++---------------------- 1 file changed, 29 insertions(+), 29 deletions(-) (limited to 'include/asm-generic') diff --git a/include/asm-generic/qrwlock.h b/include/asm-generic/qrwlock.h index 6383d54bf983..55e3ee1d2415 100644 --- a/include/asm-generic/qrwlock.h +++ b/include/asm-generic/qrwlock.h @@ -36,33 +36,33 @@ /* * External function declarations */ -extern void queue_read_lock_slowpath(struct qrwlock *lock); -extern void queue_write_lock_slowpath(struct qrwlock *lock); +extern void queued_read_lock_slowpath(struct qrwlock *lock); +extern void queued_write_lock_slowpath(struct qrwlock *lock); /** - * queue_read_can_lock- would read_trylock() succeed? + * queued_read_can_lock- would read_trylock() succeed? * @lock: Pointer to queue rwlock structure */ -static inline int queue_read_can_lock(struct qrwlock *lock) +static inline int queued_read_can_lock(struct qrwlock *lock) { return !(atomic_read(&lock->cnts) & _QW_WMASK); } /** - * queue_write_can_lock- would write_trylock() succeed? + * queued_write_can_lock- would write_trylock() succeed? * @lock: Pointer to queue rwlock structure */ -static inline int queue_write_can_lock(struct qrwlock *lock) +static inline int queued_write_can_lock(struct qrwlock *lock) { return !atomic_read(&lock->cnts); } /** - * queue_read_trylock - try to acquire read lock of a queue rwlock + * queued_read_trylock - try to acquire read lock of a queue rwlock * @lock : Pointer to queue rwlock structure * Return: 1 if lock acquired, 0 if failed */ -static inline int queue_read_trylock(struct qrwlock *lock) +static inline int queued_read_trylock(struct qrwlock *lock) { u32 cnts; @@ -77,11 +77,11 @@ static inline int queue_read_trylock(struct qrwlock *lock) } /** - * queue_write_trylock - try to acquire write lock of a queue rwlock + * queued_write_trylock - try to acquire write lock of a queue rwlock * @lock : Pointer to queue rwlock structure * Return: 1 if lock acquired, 0 if failed */ -static inline int queue_write_trylock(struct qrwlock *lock) +static inline int queued_write_trylock(struct qrwlock *lock) { u32 cnts; @@ -93,10 +93,10 @@ static inline int queue_write_trylock(struct qrwlock *lock) cnts, cnts | _QW_LOCKED) == cnts); } /** - * queue_read_lock - acquire read lock of a queue rwlock + * queued_read_lock - acquire read lock of a queue rwlock * @lock: Pointer to queue rwlock structure */ -static inline void queue_read_lock(struct qrwlock *lock) +static inline void queued_read_lock(struct qrwlock *lock) { u32 cnts; @@ -105,27 +105,27 @@ static inline void queue_read_lock(struct qrwlock *lock) return; /* The slowpath will decrement the reader count, if necessary. */ - queue_read_lock_slowpath(lock); + queued_read_lock_slowpath(lock); } /** - * queue_write_lock - acquire write lock of a queue rwlock + * queued_write_lock - acquire write lock of a queue rwlock * @lock : Pointer to queue rwlock structure */ -static inline void queue_write_lock(struct qrwlock *lock) +static inline void queued_write_lock(struct qrwlock *lock) { /* Optimize for the unfair lock case where the fair flag is 0. */ if (atomic_cmpxchg(&lock->cnts, 0, _QW_LOCKED) == 0) return; - queue_write_lock_slowpath(lock); + queued_write_lock_slowpath(lock); } /** - * queue_read_unlock - release read lock of a queue rwlock + * queued_read_unlock - release read lock of a queue rwlock * @lock : Pointer to queue rwlock structure */ -static inline void queue_read_unlock(struct qrwlock *lock) +static inline void queued_read_unlock(struct qrwlock *lock) { /* * Atomically decrement the reader count @@ -134,12 +134,12 @@ static inline void queue_read_unlock(struct qrwlock *lock) atomic_sub(_QR_BIAS, &lock->cnts); } -#ifndef queue_write_unlock +#ifndef queued_write_unlock /** - * queue_write_unlock - release write lock of a queue rwlock + * queued_write_unlock - release write lock of a queue rwlock * @lock : Pointer to queue rwlock structure */ -static inline void queue_write_unlock(struct qrwlock *lock) +static inline void queued_write_unlock(struct qrwlock *lock) { /* * If the writer field is atomic, it can be cleared directly. @@ -154,13 +154,13 @@ static inline void queue_write_unlock(struct qrwlock *lock) * Remapping rwlock architecture specific functions to the corresponding * queue rwlock functions. */ -#define arch_read_can_lock(l) queue_read_can_lock(l) -#define arch_write_can_lock(l) queue_write_can_lock(l) -#define arch_read_lock(l) queue_read_lock(l) -#define arch_write_lock(l) queue_write_lock(l) -#define arch_read_trylock(l) queue_read_trylock(l) -#define arch_write_trylock(l) queue_write_trylock(l) -#define arch_read_unlock(l) queue_read_unlock(l) -#define arch_write_unlock(l) queue_write_unlock(l) +#define arch_read_can_lock(l) queued_read_can_lock(l) +#define arch_write_can_lock(l) queued_write_can_lock(l) +#define arch_read_lock(l) queued_read_lock(l) +#define arch_write_lock(l) queued_write_lock(l) +#define arch_read_trylock(l) queued_read_trylock(l) +#define arch_write_trylock(l) queued_write_trylock(l) +#define arch_read_unlock(l) queued_read_unlock(l) +#define arch_write_unlock(l) queued_write_unlock(l) #endif /* __ASM_GENERIC_QRWLOCK_H */ -- cgit From 0e06e5be70d392aa842c1455ec2d0baf62aeed48 Mon Sep 17 00:00:00 2001 From: Waiman Long Date: Fri, 19 Jun 2015 11:50:01 -0400 Subject: locking/qrwlock: Better optimization for interrupt context readers The qrwlock is fair in the process context, but becoming unfair when in the interrupt context to support use cases like the tasklist_lock. The current code isn't that well-documented on what happens when in the interrupt context. The rspin_until_writer_unlock() will only spin if the writer has gotten the lock. If the writer is still in the waiting state, the increment in the reader count will cause the writer to remain in the waiting state and the new interrupt context reader will get the lock and return immediately. The current code, however, does an additional read of the lock value which is not necessary as the information has already been there in the fast path. This may sometime cause an additional cacheline transfer when the lock is highly contended. This patch passes the lock value information gotten in the fast path to the slow path to eliminate the additional read. It also documents the action for the interrupt context readers more clearly. Signed-off-by: Waiman Long Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Will Deacon Cc: Arnd Bergmann Cc: Douglas Hatch Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Scott J Norton Cc: Thomas Gleixner Link: http://lkml.kernel.org/r/1434729002-57724-3-git-send-email-Waiman.Long@hp.com Signed-off-by: Ingo Molnar --- include/asm-generic/qrwlock.h | 4 ++-- kernel/locking/qrwlock.c | 13 +++++++------ 2 files changed, 9 insertions(+), 8 deletions(-) (limited to 'include/asm-generic') diff --git a/include/asm-generic/qrwlock.h b/include/asm-generic/qrwlock.h index 55e3ee1d2415..deb9e8b0eb9e 100644 --- a/include/asm-generic/qrwlock.h +++ b/include/asm-generic/qrwlock.h @@ -36,7 +36,7 @@ /* * External function declarations */ -extern void queued_read_lock_slowpath(struct qrwlock *lock); +extern void queued_read_lock_slowpath(struct qrwlock *lock, u32 cnts); extern void queued_write_lock_slowpath(struct qrwlock *lock); /** @@ -105,7 +105,7 @@ static inline void queued_read_lock(struct qrwlock *lock) return; /* The slowpath will decrement the reader count, if necessary. */ - queued_read_lock_slowpath(lock); + queued_read_lock_slowpath(lock, cnts); } /** diff --git a/kernel/locking/qrwlock.c b/kernel/locking/qrwlock.c index 49057d413b6e..d9c36c5f5711 100644 --- a/kernel/locking/qrwlock.c +++ b/kernel/locking/qrwlock.c @@ -62,20 +62,21 @@ rspin_until_writer_unlock(struct qrwlock *lock, u32 cnts) /** * queued_read_lock_slowpath - acquire read lock of a queue rwlock * @lock: Pointer to queue rwlock structure + * @cnts: Current qrwlock lock value */ -void queued_read_lock_slowpath(struct qrwlock *lock) +void queued_read_lock_slowpath(struct qrwlock *lock, u32 cnts) { - u32 cnts; - /* * Readers come here when they cannot get the lock without waiting */ if (unlikely(in_interrupt())) { /* - * Readers in interrupt context will spin until the lock is - * available without waiting in the queue. + * Readers in interrupt context will get the lock immediately + * if the writer is just waiting (not holding the lock yet). + * The rspin_until_writer_unlock() function returns immediately + * in this case. Otherwise, they will spin until the lock + * is available without waiting in the queue. */ - cnts = smp_load_acquire((u32 *)&lock->cnts); rspin_until_writer_unlock(lock, cnts); return; } -- cgit From 56d1defe0bbddaa97d6e74b51490904130fd4f1d Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 15 Jul 2015 15:47:25 +0200 Subject: atomic: Prepare generic atomic implementation for logic ops Clean up the #ifdef guards a bit to prepare for architectures to supply their own logic ops. Signed-off-by: Peter Zijlstra (Intel) Signed-off-by: Thomas Gleixner --- include/asm-generic/atomic.h | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-) (limited to 'include/asm-generic') diff --git a/include/asm-generic/atomic.h b/include/asm-generic/atomic.h index 1973ad2b13f4..92947e0a532a 100644 --- a/include/asm-generic/atomic.h +++ b/include/asm-generic/atomic.h @@ -98,14 +98,22 @@ ATOMIC_OP_RETURN(add, +) ATOMIC_OP_RETURN(sub, -) #endif -#ifndef atomic_clear_mask +#ifndef atomic_and ATOMIC_OP(and, &) +#endif + +#ifndef atomic_clear_mask #define atomic_clear_mask(i, v) atomic_and(~(i), (v)) #endif -#ifndef atomic_set_mask +#ifndef atomic_or +#ifndef CONFIG_ARCH_HAS_ATOMIC_OR #define CONFIG_ARCH_HAS_ATOMIC_OR +#endif ATOMIC_OP(or, |) +#endif + +#ifndef atomic_set_mask #define atomic_set_mask(i, v) atomic_or((i), (v)) #endif -- cgit From e6942b7de2dfe44ebde9bae57dadece5abca9de8 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 23 Apr 2014 19:32:50 +0200 Subject: atomic: Provide atomic_{or,xor,and} Implement atomic logic ops -- atomic_{or,xor,and}. These will replace the atomic_{set,clear}_mask functions that are available on some archs. Signed-off-by: Peter Zijlstra (Intel) Signed-off-by: Thomas Gleixner --- arch/alpha/include/asm/atomic.h | 1 - arch/arc/include/asm/atomic.h | 1 - arch/arm/include/asm/atomic.h | 1 - arch/arm64/include/asm/atomic.h | 1 - arch/avr32/include/asm/atomic.h | 2 -- arch/blackfin/include/asm/atomic.h | 2 -- arch/frv/include/asm/atomic.h | 2 -- arch/h8300/include/asm/atomic.h | 2 -- arch/hexagon/include/asm/atomic.h | 2 -- arch/ia64/include/asm/atomic.h | 2 -- arch/m32r/include/asm/atomic.h | 2 -- arch/m68k/include/asm/atomic.h | 2 -- arch/metag/include/asm/atomic_lnkget.h | 2 -- arch/mips/include/asm/atomic.h | 2 -- arch/mn10300/include/asm/atomic.h | 2 -- arch/parisc/include/asm/atomic.h | 2 -- arch/powerpc/include/asm/atomic.h | 2 -- arch/s390/include/asm/atomic.h | 2 -- arch/sh/include/asm/atomic-grb.h | 2 -- arch/sparc/include/asm/atomic_32.h | 2 -- arch/sparc/include/asm/atomic_64.h | 2 -- arch/tile/include/asm/atomic_32.h | 2 -- arch/tile/include/asm/atomic_64.h | 2 -- arch/x86/include/asm/atomic.h | 2 -- arch/xtensa/include/asm/atomic.h | 2 -- include/asm-generic/atomic.h | 21 ++++++++++++--------- include/asm-generic/atomic64.h | 4 ++++ include/linux/atomic.h | 13 ------------- lib/atomic64.c | 3 +++ 29 files changed, 19 insertions(+), 68 deletions(-) (limited to 'include/asm-generic') diff --git a/arch/alpha/include/asm/atomic.h b/arch/alpha/include/asm/atomic.h index 0eff853398d2..e8c956098424 100644 --- a/arch/alpha/include/asm/atomic.h +++ b/arch/alpha/include/asm/atomic.h @@ -110,7 +110,6 @@ static __inline__ long atomic64_##op##_return(long i, atomic64_t * v) \ ATOMIC_OPS(add) ATOMIC_OPS(sub) -#define CONFIG_ARCH_HAS_ATOMIC_OR #define atomic_andnot atomic_andnot #define atomic64_andnot atomic64_andnot diff --git a/arch/arc/include/asm/atomic.h b/arch/arc/include/asm/atomic.h index e90b701fc6a8..2a847821dee1 100644 --- a/arch/arc/include/asm/atomic.h +++ b/arch/arc/include/asm/atomic.h @@ -144,7 +144,6 @@ static inline int atomic_##op##_return(int i, atomic_t *v) \ ATOMIC_OPS(add, +=, add) ATOMIC_OPS(sub, -=, sub) -#define CONFIG_ARCH_HAS_ATOMIC_OR #define atomic_andnot atomic_andnot ATOMIC_OP(and, &=, and) diff --git a/arch/arm/include/asm/atomic.h b/arch/arm/include/asm/atomic.h index ff214bac9cb4..82b75a7cb762 100644 --- a/arch/arm/include/asm/atomic.h +++ b/arch/arm/include/asm/atomic.h @@ -194,7 +194,6 @@ static inline int __atomic_add_unless(atomic_t *v, int a, int u) ATOMIC_OPS(add, +=, add) ATOMIC_OPS(sub, -=, sub) -#define CONFIG_ARCH_HAS_ATOMIC_OR #define atomic_andnot atomic_andnot ATOMIC_OP(and, &=, and) diff --git a/arch/arm64/include/asm/atomic.h b/arch/arm64/include/asm/atomic.h index 2876173397b2..866a71fca9a3 100644 --- a/arch/arm64/include/asm/atomic.h +++ b/arch/arm64/include/asm/atomic.h @@ -85,7 +85,6 @@ static inline int atomic_##op##_return(int i, atomic_t *v) \ ATOMIC_OPS(add, add) ATOMIC_OPS(sub, sub) -#define CONFIG_ARCH_HAS_ATOMIC_OR #define atomic_andnot atomic_andnot ATOMIC_OP(and, and) diff --git a/arch/avr32/include/asm/atomic.h b/arch/avr32/include/asm/atomic.h index 115d3005e4bc..97c9bdf83409 100644 --- a/arch/avr32/include/asm/atomic.h +++ b/arch/avr32/include/asm/atomic.h @@ -51,8 +51,6 @@ static inline void atomic_##op(int i, atomic_t *v) \ (void)__atomic_##op##_return(i, v); \ } -#define CONFIG_ARCH_HAS_ATOMIC_OR - ATOMIC_OP(and, and) ATOMIC_OP(or, or) ATOMIC_OP(xor, eor) diff --git a/arch/blackfin/include/asm/atomic.h b/arch/blackfin/include/asm/atomic.h index eafa55b81a7b..2d6a7a3823c3 100644 --- a/arch/blackfin/include/asm/atomic.h +++ b/arch/blackfin/include/asm/atomic.h @@ -28,8 +28,6 @@ asmlinkage int __raw_atomic_test_asm(const volatile int *ptr, int value); #define atomic_add_return(i, v) __raw_atomic_add_asm(&(v)->counter, i) #define atomic_sub_return(i, v) __raw_atomic_add_asm(&(v)->counter, -(i)) -#define CONFIG_ARCH_HAS_ATOMIC_OR - #define atomic_or(i, v) (void)__raw_atomic_or_asm(&(v)->counter, i) #define atomic_and(i, v) (void)__raw_atomic_and_asm(&(v)->counter, i) #define atomic_xor(i, v) (void)__raw_atomic_xor_asm(&(v)->counter, i) diff --git a/arch/frv/include/asm/atomic.h b/arch/frv/include/asm/atomic.h index 74d22454d7c6..fc48bea26b40 100644 --- a/arch/frv/include/asm/atomic.h +++ b/arch/frv/include/asm/atomic.h @@ -192,8 +192,6 @@ static inline void atomic64_##op(long long i, atomic64_t *v) \ (void)__atomic64_fetch_##op(i, &v->counter); \ } -#define CONFIG_ARCH_HAS_ATOMIC_OR - ATOMIC_OP(or) ATOMIC_OP(and) ATOMIC_OP(xor) diff --git a/arch/h8300/include/asm/atomic.h b/arch/h8300/include/asm/atomic.h index f181f820be33..c4d061f09c44 100644 --- a/arch/h8300/include/asm/atomic.h +++ b/arch/h8300/include/asm/atomic.h @@ -41,8 +41,6 @@ static inline void atomic_##op(int i, atomic_t *v) \ ATOMIC_OP_RETURN(add, +=) ATOMIC_OP_RETURN(sub, -=) -#define CONFIG_ARCH_HAS_ATOMIC_OR - ATOMIC_OP(and, &=) ATOMIC_OP(or, |=) ATOMIC_OP(xor, ^=) diff --git a/arch/hexagon/include/asm/atomic.h b/arch/hexagon/include/asm/atomic.h index 4efe2c7c0dd8..811d61f6422d 100644 --- a/arch/hexagon/include/asm/atomic.h +++ b/arch/hexagon/include/asm/atomic.h @@ -132,8 +132,6 @@ static inline int atomic_##op##_return(int i, atomic_t *v) \ ATOMIC_OPS(add) ATOMIC_OPS(sub) -#define CONFIG_ARCH_HAS_ATOMIC_OR - ATOMIC_OP(and) ATOMIC_OP(or) ATOMIC_OP(xor) diff --git a/arch/ia64/include/asm/atomic.h b/arch/ia64/include/asm/atomic.h index 0809ef5d6b9a..be4beeb77d57 100644 --- a/arch/ia64/include/asm/atomic.h +++ b/arch/ia64/include/asm/atomic.h @@ -69,8 +69,6 @@ ATOMIC_OP(sub, -) : ia64_atomic_sub(__ia64_asr_i, v); \ }) -#define CONFIG_ARCH_HAS_ATOMIC_OR - ATOMIC_OP(and, &) ATOMIC_OP(or, |) ATOMIC_OP(xor, ^) diff --git a/arch/m32r/include/asm/atomic.h b/arch/m32r/include/asm/atomic.h index 7245463c1e98..b2a13fbd5be0 100644 --- a/arch/m32r/include/asm/atomic.h +++ b/arch/m32r/include/asm/atomic.h @@ -94,8 +94,6 @@ static __inline__ int atomic_##op##_return(int i, atomic_t *v) \ ATOMIC_OPS(add) ATOMIC_OPS(sub) -#define CONFIG_ARCH_HAS_ATOMIC_OR - ATOMIC_OP(and) ATOMIC_OP(or) ATOMIC_OP(xor) diff --git a/arch/m68k/include/asm/atomic.h b/arch/m68k/include/asm/atomic.h index c30e43ea49a3..93ebd96aa494 100644 --- a/arch/m68k/include/asm/atomic.h +++ b/arch/m68k/include/asm/atomic.h @@ -77,8 +77,6 @@ static inline int atomic_##op##_return(int i, atomic_t * v) \ ATOMIC_OPS(add, +=, add) ATOMIC_OPS(sub, -=, sub) -#define CONFIG_ARCH_HAS_ATOMIC_OR - ATOMIC_OP(and, &=, and) ATOMIC_OP(or, |=, or) ATOMIC_OP(xor, ^=, eor) diff --git a/arch/metag/include/asm/atomic_lnkget.h b/arch/metag/include/asm/atomic_lnkget.h index 930c12cb8d37..0642606de901 100644 --- a/arch/metag/include/asm/atomic_lnkget.h +++ b/arch/metag/include/asm/atomic_lnkget.h @@ -74,8 +74,6 @@ static inline int atomic_##op##_return(int i, atomic_t *v) \ ATOMIC_OPS(add) ATOMIC_OPS(sub) -#define CONFIG_ARCH_HAS_ATOMIC_OR - ATOMIC_OP(and) ATOMIC_OP(or) ATOMIC_OP(xor) diff --git a/arch/mips/include/asm/atomic.h b/arch/mips/include/asm/atomic.h index 0430ba6ab762..4c42fd9af777 100644 --- a/arch/mips/include/asm/atomic.h +++ b/arch/mips/include/asm/atomic.h @@ -137,8 +137,6 @@ static __inline__ int atomic_##op##_return(int i, atomic_t * v) \ ATOMIC_OPS(add, +=, addu) ATOMIC_OPS(sub, -=, subu) -#define CONFIG_ARCH_HAS_ATOMIC_OR - ATOMIC_OP(and, &=, and) ATOMIC_OP(or, |=, or) ATOMIC_OP(xor, ^=, xor) diff --git a/arch/mn10300/include/asm/atomic.h b/arch/mn10300/include/asm/atomic.h index 03eea8158cf9..f5a63f0bda46 100644 --- a/arch/mn10300/include/asm/atomic.h +++ b/arch/mn10300/include/asm/atomic.h @@ -89,8 +89,6 @@ static inline int atomic_##op##_return(int i, atomic_t *v) \ ATOMIC_OPS(add) ATOMIC_OPS(sub) -#define CONFIG_ARCH_HAS_ATOMIC_OR - ATOMIC_OP(and) ATOMIC_OP(or) ATOMIC_OP(xor) diff --git a/arch/parisc/include/asm/atomic.h b/arch/parisc/include/asm/atomic.h index be2c50ddebd6..2536965d00ea 100644 --- a/arch/parisc/include/asm/atomic.h +++ b/arch/parisc/include/asm/atomic.h @@ -126,8 +126,6 @@ static __inline__ int atomic_##op##_return(int i, atomic_t *v) \ ATOMIC_OPS(add, +=) ATOMIC_OPS(sub, -=) -#define CONFIG_ARCH_HAS_ATOMIC_OR - ATOMIC_OP(and, &=) ATOMIC_OP(or, |=) ATOMIC_OP(xor, ^=) diff --git a/arch/powerpc/include/asm/atomic.h b/arch/powerpc/include/asm/atomic.h index 6ca89e2aca15..55f106ed12bf 100644 --- a/arch/powerpc/include/asm/atomic.h +++ b/arch/powerpc/include/asm/atomic.h @@ -67,8 +67,6 @@ static __inline__ int atomic_##op##_return(int a, atomic_t *v) \ ATOMIC_OPS(add, add) ATOMIC_OPS(sub, subf) -#define CONFIG_ARCH_HAS_ATOMIC_OR - ATOMIC_OP(and, and) ATOMIC_OP(or, or) ATOMIC_OP(xor, xor) diff --git a/arch/s390/include/asm/atomic.h b/arch/s390/include/asm/atomic.h index b3859d8e001f..d761aeff72da 100644 --- a/arch/s390/include/asm/atomic.h +++ b/arch/s390/include/asm/atomic.h @@ -282,8 +282,6 @@ static inline void atomic64_##op(long i, atomic64_t *v) \ __ATOMIC64_LOOP(v, i, __ATOMIC64_##OP, __ATOMIC64_NO_BARRIER); \ } -#define CONFIG_ARCH_HAS_ATOMIC_OR - ATOMIC64_OP(and, AND) ATOMIC64_OP(or, OR) ATOMIC64_OP(xor, XOR) diff --git a/arch/sh/include/asm/atomic-grb.h b/arch/sh/include/asm/atomic-grb.h index 4b03830d48c7..b94df40e5f2d 100644 --- a/arch/sh/include/asm/atomic-grb.h +++ b/arch/sh/include/asm/atomic-grb.h @@ -48,8 +48,6 @@ static inline int atomic_##op##_return(int i, atomic_t *v) \ ATOMIC_OPS(add) ATOMIC_OPS(sub) -#define CONFIG_ARCH_HAS_ATOMIC_OR - ATOMIC_OP(and) ATOMIC_OP(or) ATOMIC_OP(xor) diff --git a/arch/sparc/include/asm/atomic_32.h b/arch/sparc/include/asm/atomic_32.h index e19d8880b146..7dcbebbcaec6 100644 --- a/arch/sparc/include/asm/atomic_32.h +++ b/arch/sparc/include/asm/atomic_32.h @@ -17,8 +17,6 @@ #include #include -#define CONFIG_ARCH_HAS_ATOMIC_OR - #define ATOMIC_INIT(i) { (i) } int atomic_add_return(int, atomic_t *); diff --git a/arch/sparc/include/asm/atomic_64.h b/arch/sparc/include/asm/atomic_64.h index d6af27c93450..917084ace49d 100644 --- a/arch/sparc/include/asm/atomic_64.h +++ b/arch/sparc/include/asm/atomic_64.h @@ -33,8 +33,6 @@ long atomic64_##op##_return(long, atomic64_t *); ATOMIC_OPS(add) ATOMIC_OPS(sub) -#define CONFIG_ARCH_HAS_ATOMIC_OR - ATOMIC_OP(and) ATOMIC_OP(or) ATOMIC_OP(xor) diff --git a/arch/tile/include/asm/atomic_32.h b/arch/tile/include/asm/atomic_32.h index 94237922f0dd..d320ce253d86 100644 --- a/arch/tile/include/asm/atomic_32.h +++ b/arch/tile/include/asm/atomic_32.h @@ -41,8 +41,6 @@ static inline void atomic_##op(int i, atomic_t *v) \ _atomic_##op((unsigned long *)&v->counter, i); \ } -#define CONFIG_ARCH_HAS_ATOMIC_OR - ATOMIC_OP(and) ATOMIC_OP(or) ATOMIC_OP(xor) diff --git a/arch/tile/include/asm/atomic_64.h b/arch/tile/include/asm/atomic_64.h index d07d9fc6e2a1..096a56d6ead4 100644 --- a/arch/tile/include/asm/atomic_64.h +++ b/arch/tile/include/asm/atomic_64.h @@ -58,8 +58,6 @@ static inline int __atomic_add_unless(atomic_t *v, int a, int u) return oldval; } -#define CONFIG_ARCH_HAS_ATOMIC_OR - static inline void atomic_and(int i, atomic_t *v) { __insn_fetchand4((void *)&v->counter, i); diff --git a/arch/x86/include/asm/atomic.h b/arch/x86/include/asm/atomic.h index f3a3ec040694..b3493023efda 100644 --- a/arch/x86/include/asm/atomic.h +++ b/arch/x86/include/asm/atomic.h @@ -191,8 +191,6 @@ static inline void atomic_##op(int i, atomic_t *v) \ : "memory"); \ } -#define CONFIG_ARCH_HAS_ATOMIC_OR - ATOMIC_OP(and) ATOMIC_OP(or) ATOMIC_OP(xor) diff --git a/arch/xtensa/include/asm/atomic.h b/arch/xtensa/include/asm/atomic.h index 4dd2450300a6..31371f43c23b 100644 --- a/arch/xtensa/include/asm/atomic.h +++ b/arch/xtensa/include/asm/atomic.h @@ -145,8 +145,6 @@ static inline int atomic_##op##_return(int i, atomic_t * v) \ ATOMIC_OPS(add) ATOMIC_OPS(sub) -#define CONFIG_ARCH_HAS_ATOMIC_OR - ATOMIC_OP(and) ATOMIC_OP(or) ATOMIC_OP(xor) diff --git a/include/asm-generic/atomic.h b/include/asm-generic/atomic.h index 92947e0a532a..a41b0b8f7404 100644 --- a/include/asm-generic/atomic.h +++ b/include/asm-generic/atomic.h @@ -102,24 +102,27 @@ ATOMIC_OP_RETURN(sub, -) ATOMIC_OP(and, &) #endif -#ifndef atomic_clear_mask -#define atomic_clear_mask(i, v) atomic_and(~(i), (v)) -#endif - #ifndef atomic_or -#ifndef CONFIG_ARCH_HAS_ATOMIC_OR -#define CONFIG_ARCH_HAS_ATOMIC_OR -#endif ATOMIC_OP(or, |) #endif -#ifndef atomic_set_mask -#define atomic_set_mask(i, v) atomic_or((i), (v)) +#ifndef atomic_xor +ATOMIC_OP(xor, ^) #endif #undef ATOMIC_OP_RETURN #undef ATOMIC_OP +static inline __deprecated void atomic_clear_mask(unsigned int mask, atomic_t *v) +{ + atomic_and(~mask, v); +} + +static inline __deprecated void atomic_set_mask(unsigned int mask, atomic_t *v) +{ + atomic_or(mask, v); +} + /* * Atomic operations that C can't guarantee us. Useful for * resource counting etc.. diff --git a/include/asm-generic/atomic64.h b/include/asm-generic/atomic64.h index 30ad9c86cebb..d48e78ccad3d 100644 --- a/include/asm-generic/atomic64.h +++ b/include/asm-generic/atomic64.h @@ -32,6 +32,10 @@ extern long long atomic64_##op##_return(long long a, atomic64_t *v); ATOMIC64_OPS(add) ATOMIC64_OPS(sub) +ATOMIC64_OP(and) +ATOMIC64_OP(or) +ATOMIC64_OP(xor) + #undef ATOMIC64_OPS #undef ATOMIC64_OP_RETURN #undef ATOMIC64_OP diff --git a/include/linux/atomic.h b/include/linux/atomic.h index 5b08a8540ecf..7d6279012a1f 100644 --- a/include/linux/atomic.h +++ b/include/linux/atomic.h @@ -111,19 +111,6 @@ static inline int atomic_dec_if_positive(atomic_t *v) } #endif -#ifndef CONFIG_ARCH_HAS_ATOMIC_OR -static inline void atomic_or(int i, atomic_t *v) -{ - int old; - int new; - - do { - old = atomic_read(v); - new = old | i; - } while (atomic_cmpxchg(v, old, new) != old); -} -#endif /* #ifndef CONFIG_ARCH_HAS_ATOMIC_OR */ - #include #ifdef CONFIG_GENERIC_ATOMIC64 #include diff --git a/lib/atomic64.c b/lib/atomic64.c index 1298c05ef528..2886ebac6567 100644 --- a/lib/atomic64.c +++ b/lib/atomic64.c @@ -102,6 +102,9 @@ EXPORT_SYMBOL(atomic64_##op##_return); ATOMIC64_OPS(add, +=) ATOMIC64_OPS(sub, -=) +ATOMIC64_OP(and, &=) +ATOMIC64_OP(or, |=) +ATOMIC64_OP(xor, ^=) #undef ATOMIC64_OPS #undef ATOMIC64_OP_RETURN -- cgit From de9e432cb5de1bf2952919dc0b22e4bec0ed8d53 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 24 Apr 2015 01:12:32 +0200 Subject: atomic: Collapse all atomic_{set,clear}_mask definitions Move the now generic definitions of atomic_{set,clear}_mask() into linux/atomic.h to avoid endless and pointless repetition. Also, provide an atomic_andnot() wrapper for those few archs that can implement that. Signed-off-by: Peter Zijlstra (Intel) Signed-off-by: Thomas Gleixner --- arch/arc/include/asm/atomic.h | 10 ---------- arch/blackfin/include/asm/atomic.h | 10 ---------- arch/frv/include/asm/atomic.h | 10 ---------- arch/h8300/include/asm/atomic.h | 10 ---------- arch/m32r/include/asm/atomic.h | 11 ----------- arch/m68k/include/asm/atomic.h | 10 ---------- arch/metag/include/asm/atomic_lnkget.h | 10 ---------- arch/metag/include/asm/atomic_lock1.h | 10 ---------- arch/mn10300/include/asm/atomic.h | 24 ------------------------ arch/powerpc/kernel/misc_32.S | 19 ------------------- arch/s390/include/asm/atomic.h | 10 ---------- arch/sh/include/asm/atomic.h | 10 ---------- arch/x86/include/asm/atomic.h | 10 ---------- arch/xtensa/include/asm/atomic.h | 10 ---------- include/asm-generic/atomic.h | 10 ---------- include/linux/atomic.h | 25 +++++++++++++++++++++++++ 16 files changed, 25 insertions(+), 174 deletions(-) (limited to 'include/asm-generic') diff --git a/arch/arc/include/asm/atomic.h b/arch/arc/include/asm/atomic.h index 2a847821dee1..d8a85e706fba 100644 --- a/arch/arc/include/asm/atomic.h +++ b/arch/arc/include/asm/atomic.h @@ -155,16 +155,6 @@ ATOMIC_OP(xor, ^=, xor) #undef ATOMIC_OP_RETURN #undef ATOMIC_OP -static inline __deprecated void atomic_clear_mask(unsigned int mask, atomic_t *v) -{ - atomic_and(~mask, v); -} - -static inline __deprecated void atomic_set_mask(unsigned int mask, atomic_t *v) -{ - atomic_or(mask, v); -} - /** * __atomic_add_unless - add unless the number is a given value * @v: pointer of type atomic_t diff --git a/arch/blackfin/include/asm/atomic.h b/arch/blackfin/include/asm/atomic.h index 2d6a7a3823c3..1c1c42330c99 100644 --- a/arch/blackfin/include/asm/atomic.h +++ b/arch/blackfin/include/asm/atomic.h @@ -32,16 +32,6 @@ asmlinkage int __raw_atomic_test_asm(const volatile int *ptr, int value); #define atomic_and(i, v) (void)__raw_atomic_and_asm(&(v)->counter, i) #define atomic_xor(i, v) (void)__raw_atomic_xor_asm(&(v)->counter, i) -static inline __deprecated void atomic_clear_mask(unsigned int mask, atomic_t *v) -{ - atomic_and(~mask, v); -} - -static inline __deprecated void atomic_set_mask(unsigned int mask, atomic_t *v) -{ - atomic_or(mask, v); -} - #endif #include diff --git a/arch/frv/include/asm/atomic.h b/arch/frv/include/asm/atomic.h index fc48bea26b40..0da689def4cc 100644 --- a/arch/frv/include/asm/atomic.h +++ b/arch/frv/include/asm/atomic.h @@ -198,14 +198,4 @@ ATOMIC_OP(xor) #undef ATOMIC_OP -static inline __deprecated void atomic_clear_mask(unsigned int mask, atomic_t *v) -{ - atomic_and(~mask, v); -} - -static inline __deprecated void atomic_set_mask(unsigned int mask, atomic_t *v) -{ - atomic_or(mask, v); -} - #endif /* _ASM_ATOMIC_H */ diff --git a/arch/h8300/include/asm/atomic.h b/arch/h8300/include/asm/atomic.h index c4d061f09c44..702ee539f87d 100644 --- a/arch/h8300/include/asm/atomic.h +++ b/arch/h8300/include/asm/atomic.h @@ -89,14 +89,4 @@ static inline int __atomic_add_unless(atomic_t *v, int a, int u) return ret; } -static inline __deprecated void atomic_clear_mask(unsigned int mask, atomic_t *v) -{ - atomic_and(~mask, v); -} - -static inline __deprecated void atomic_set_mask(unsigned int mask, atomic_t *v) -{ - atomic_or(mask, v); -} - #endif /* __ARCH_H8300_ATOMIC __ */ diff --git a/arch/m32r/include/asm/atomic.h b/arch/m32r/include/asm/atomic.h index b2a13fbd5be0..025e2a170493 100644 --- a/arch/m32r/include/asm/atomic.h +++ b/arch/m32r/include/asm/atomic.h @@ -243,15 +243,4 @@ static __inline__ int __atomic_add_unless(atomic_t *v, int a, int u) return c; } - -static __inline__ __deprecated void atomic_clear_mask(unsigned int mask, atomic_t *v) -{ - atomic_and(~mask, v); -} - -static __inline__ __deprecated void atomic_set_mask(unsigned int mask, atomic_t *v) -{ - atomic_or(mask, v); -} - #endif /* _ASM_M32R_ATOMIC_H */ diff --git a/arch/m68k/include/asm/atomic.h b/arch/m68k/include/asm/atomic.h index 93ebd96aa494..039fac120cc0 100644 --- a/arch/m68k/include/asm/atomic.h +++ b/arch/m68k/include/asm/atomic.h @@ -174,16 +174,6 @@ static inline int atomic_add_negative(int i, atomic_t *v) return c != 0; } -static inline __deprecated void atomic_clear_mask(unsigned int mask, atomic_t *v) -{ - atomic_and(~mask, v); -} - -static inline __deprecated void atomic_set_mask(unsigned int mask, atomic_t *v) -{ - atomic_or(mask, v); -} - static __inline__ int __atomic_add_unless(atomic_t *v, int a, int u) { int c, old; diff --git a/arch/metag/include/asm/atomic_lnkget.h b/arch/metag/include/asm/atomic_lnkget.h index 0642606de901..21c4c268b86c 100644 --- a/arch/metag/include/asm/atomic_lnkget.h +++ b/arch/metag/include/asm/atomic_lnkget.h @@ -82,16 +82,6 @@ ATOMIC_OP(xor) #undef ATOMIC_OP_RETURN #undef ATOMIC_OP -static inline __deprecated void atomic_clear_mask(unsigned int mask, atomic_t *v) -{ - atomic_and(~mask, v); -} - -static inline __deprecated void atomic_set_mask(unsigned int mask, atomic_t *v) -{ - atomic_or(mask, v); -} - static inline int atomic_cmpxchg(atomic_t *v, int old, int new) { int result, temp; diff --git a/arch/metag/include/asm/atomic_lock1.h b/arch/metag/include/asm/atomic_lock1.h index 7d88725a85da..f8efe380fe8b 100644 --- a/arch/metag/include/asm/atomic_lock1.h +++ b/arch/metag/include/asm/atomic_lock1.h @@ -76,16 +76,6 @@ ATOMIC_OP(xor, ^=) #undef ATOMIC_OP_RETURN #undef ATOMIC_OP -static inline __deprecated void atomic_clear_mask(unsigned int mask, atomic_t *v) -{ - atomic_and(~mask, v); -} - -static inline __deprecated void atomic_set_mask(unsigned int mask, atomic_t *v) -{ - atomic_or(mask, v); -} - static inline int atomic_cmpxchg(atomic_t *v, int old, int new) { int ret; diff --git a/arch/mn10300/include/asm/atomic.h b/arch/mn10300/include/asm/atomic.h index f5a63f0bda46..375e59140c9c 100644 --- a/arch/mn10300/include/asm/atomic.h +++ b/arch/mn10300/include/asm/atomic.h @@ -131,30 +131,6 @@ static inline void atomic_dec(atomic_t *v) #define atomic_xchg(ptr, v) (xchg(&(ptr)->counter, (v))) #define atomic_cmpxchg(v, old, new) (cmpxchg(&((v)->counter), (old), (new))) -/** - * atomic_clear_mask - Atomically clear bits in memory - * @mask: Mask of the bits to be cleared - * @v: pointer to word in memory - * - * Atomically clears the bits set in mask from the memory word specified. - */ -static inline __deprecated void atomic_clear_mask(unsigned int mask, atomic_t *v) -{ - atomic_and(~mask, v); -} - -/** - * atomic_set_mask - Atomically set bits in memory - * @mask: Mask of the bits to be set - * @v: pointer to word in memory - * - * Atomically sets the bits set in mask from the memory word specified. - */ -static inline __deprecated void atomic_set_mask(unsigned int mask, atomic_t *v) -{ - atomic_or(mask, v); -} - #endif /* __KERNEL__ */ #endif /* CONFIG_SMP */ #endif /* _ASM_ATOMIC_H */ diff --git a/arch/powerpc/kernel/misc_32.S b/arch/powerpc/kernel/misc_32.S index 7c6bb4b17b49..ed3ab509faca 100644 --- a/arch/powerpc/kernel/misc_32.S +++ b/arch/powerpc/kernel/misc_32.S @@ -595,25 +595,6 @@ _GLOBAL(copy_page) li r11,4 b 2b -/* - * void atomic_clear_mask(atomic_t mask, atomic_t *addr) - * void atomic_set_mask(atomic_t mask, atomic_t *addr); - */ -_GLOBAL(atomic_clear_mask) -10: lwarx r5,0,r4 - andc r5,r5,r3 - PPC405_ERR77(0,r4) - stwcx. r5,0,r4 - bne- 10b - blr -_GLOBAL(atomic_set_mask) -10: lwarx r5,0,r4 - or r5,r5,r3 - PPC405_ERR77(0,r4) - stwcx. r5,0,r4 - bne- 10b - blr - /* * Extended precision shifts. * diff --git a/arch/s390/include/asm/atomic.h b/arch/s390/include/asm/atomic.h index d761aeff72da..117fa5c921c1 100644 --- a/arch/s390/include/asm/atomic.h +++ b/arch/s390/include/asm/atomic.h @@ -132,16 +132,6 @@ ATOMIC_OP(xor, XOR) #undef ATOMIC_OP -static inline __deprecated void atomic_clear_mask(unsigned int mask, atomic_t *v) -{ - atomic_and(~mask, v); -} - -static inline __deprecated void atomic_set_mask(unsigned int mask, atomic_t *v) -{ - atomic_or(mask, v); -} - #define atomic_xchg(v, new) (xchg(&((v)->counter), new)) static inline int atomic_cmpxchg(atomic_t *v, int old, int new) diff --git a/arch/sh/include/asm/atomic.h b/arch/sh/include/asm/atomic.h index cee0245257e1..05b9f74ce2d5 100644 --- a/arch/sh/include/asm/atomic.h +++ b/arch/sh/include/asm/atomic.h @@ -25,16 +25,6 @@ #include #endif -static inline __deprecated void atomic_clear_mask(unsigned int mask, atomic_t *v) -{ - atomic_and(~mask, v); -} - -static inline __deprecated void atomic_set_mask(unsigned int mask, atomic_t *v) -{ - atomic_or(mask, v); -} - #define atomic_add_negative(a, v) (atomic_add_return((a), (v)) < 0) #define atomic_dec_return(v) atomic_sub_return(1, (v)) #define atomic_inc_return(v) atomic_add_return(1, (v)) diff --git a/arch/x86/include/asm/atomic.h b/arch/x86/include/asm/atomic.h index b3493023efda..fb52aa644aab 100644 --- a/arch/x86/include/asm/atomic.h +++ b/arch/x86/include/asm/atomic.h @@ -234,16 +234,6 @@ static __always_inline short int atomic_inc_short(short int *v) return *v; } -static inline __deprecated void atomic_clear_mask(unsigned int mask, atomic_t *v) -{ - atomic_and(~mask, v); -} - -static inline __deprecated void atomic_set_mask(unsigned int mask, atomic_t *v) -{ - atomic_or(mask, v); -} - #ifdef CONFIG_X86_32 # include #else diff --git a/arch/xtensa/include/asm/atomic.h b/arch/xtensa/include/asm/atomic.h index 31371f43c23b..e0be67936990 100644 --- a/arch/xtensa/include/asm/atomic.h +++ b/arch/xtensa/include/asm/atomic.h @@ -153,16 +153,6 @@ ATOMIC_OP(xor) #undef ATOMIC_OP_RETURN #undef ATOMIC_OP -static inline __deprecated void atomic_set_mask(unsigned int mask, atomic_t *v) -{ - atomic_or(mask, v); -} - -static inline __deprecated void atomic_clear_mask(unsigned int mask, atomic_t *v) -{ - atomic_and(~mask, v); -} - /** * atomic_sub_and_test - subtract value from variable and test result * @i: integer value to subtract diff --git a/include/asm-generic/atomic.h b/include/asm-generic/atomic.h index a41b0b8f7404..d4d7e337fdcb 100644 --- a/include/asm-generic/atomic.h +++ b/include/asm-generic/atomic.h @@ -113,16 +113,6 @@ ATOMIC_OP(xor, ^) #undef ATOMIC_OP_RETURN #undef ATOMIC_OP -static inline __deprecated void atomic_clear_mask(unsigned int mask, atomic_t *v) -{ - atomic_and(~mask, v); -} - -static inline __deprecated void atomic_set_mask(unsigned int mask, atomic_t *v) -{ - atomic_or(mask, v); -} - /* * Atomic operations that C can't guarantee us. Useful for * resource counting etc.. diff --git a/include/linux/atomic.h b/include/linux/atomic.h index 7d6279012a1f..8b98b423388f 100644 --- a/include/linux/atomic.h +++ b/include/linux/atomic.h @@ -28,6 +28,23 @@ static inline int atomic_add_unless(atomic_t *v, int a, int u) #define atomic_inc_not_zero(v) atomic_add_unless((v), 1, 0) #endif +#ifndef atomic_andnot +static inline void atomic_andnot(int i, atomic_t *v) +{ + atomic_and(~i, v); +} +#endif + +static inline __deprecated void atomic_clear_mask(unsigned int mask, atomic_t *v) +{ + atomic_andnot(mask, v); +} + +static inline __deprecated void atomic_set_mask(unsigned int mask, atomic_t *v) +{ + atomic_or(mask, v); +} + /** * atomic_inc_not_zero_hint - increment if not null * @v: pointer of type atomic_t @@ -115,4 +132,12 @@ static inline int atomic_dec_if_positive(atomic_t *v) #ifdef CONFIG_GENERIC_ATOMIC64 #include #endif + +#ifndef atomic64_andnot +static inline void atomic64_andnot(long long i, atomic64_t *v) +{ + atomic64_and(~i, v); +} +#endif + #endif /* _LINUX_ATOMIC_H */ -- cgit From 76695af20c015206cffb84b15912be6797d0cca2 Mon Sep 17 00:00:00 2001 From: Andrey Konovalov Date: Sun, 2 Aug 2015 17:11:04 +0200 Subject: locking, arch: use WRITE_ONCE()/READ_ONCE() in smp_store_release()/smp_load_acquire() Replace ACCESS_ONCE() macro in smp_store_release() and smp_load_acquire() with WRITE_ONCE() and READ_ONCE() on x86, arm, arm64, ia64, metag, mips, powerpc, s390, sparc and asm-generic since ACCESS_ONCE() does not work reliably on non-scalar types. WRITE_ONCE() and READ_ONCE() were introduced in the following commits: 230fa253df63 ("kernel: Provide READ_ONCE and ASSIGN_ONCE") 43239cbe79fc ("kernel: Change ASSIGN_ONCE(val, x) to WRITE_ONCE(x, val)") Signed-off-by: Andrey Konovalov Signed-off-by: Peter Zijlstra (Intel) Acked-by: Davidlohr Bueso Acked-by: Michael Ellerman (powerpc) Acked-by: Ralf Baechle Cc: Alexander Duyck Cc: Andre Przywara Cc: Arnd Bergmann Cc: Benjamin Herrenschmidt Cc: Borislav Petkov Cc: Catalin Marinas Cc: Christian Borntraeger Cc: David S. Miller Cc: Davidlohr Bueso Cc: Dmitry Vyukov Cc: Fenghua Yu Cc: H. Peter Anvin Cc: Heiko Carstens Cc: James Hogan Cc: Linus Torvalds Cc: Martin Schwidefsky Cc: Paul E. McKenney Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Russell King Cc: Thomas Gleixner Cc: Tony Luck Cc: Will Deacon Cc: linux-arch@vger.kernel.org Link: http://lkml.kernel.org/r/1438528264-714-1-git-send-email-andreyknvl@google.com Signed-off-by: Ingo Molnar --- arch/arm/include/asm/barrier.h | 4 ++-- arch/arm64/include/asm/barrier.h | 4 ++-- arch/ia64/include/asm/barrier.h | 4 ++-- arch/metag/include/asm/barrier.h | 4 ++-- arch/mips/include/asm/barrier.h | 4 ++-- arch/powerpc/include/asm/barrier.h | 4 ++-- arch/s390/include/asm/barrier.h | 4 ++-- arch/sparc/include/asm/barrier_64.h | 4 ++-- arch/x86/include/asm/barrier.h | 8 ++++---- include/asm-generic/barrier.h | 4 ++-- 10 files changed, 22 insertions(+), 22 deletions(-) (limited to 'include/asm-generic') diff --git a/arch/arm/include/asm/barrier.h b/arch/arm/include/asm/barrier.h index 6c2327e1c732..70393574e0fa 100644 --- a/arch/arm/include/asm/barrier.h +++ b/arch/arm/include/asm/barrier.h @@ -67,12 +67,12 @@ do { \ compiletime_assert_atomic_type(*p); \ smp_mb(); \ - ACCESS_ONCE(*p) = (v); \ + WRITE_ONCE(*p, v); \ } while (0) #define smp_load_acquire(p) \ ({ \ - typeof(*p) ___p1 = ACCESS_ONCE(*p); \ + typeof(*p) ___p1 = READ_ONCE(*p); \ compiletime_assert_atomic_type(*p); \ smp_mb(); \ ___p1; \ diff --git a/arch/arm64/include/asm/barrier.h b/arch/arm64/include/asm/barrier.h index 0fa47c4275cb..ef93b20bc964 100644 --- a/arch/arm64/include/asm/barrier.h +++ b/arch/arm64/include/asm/barrier.h @@ -44,12 +44,12 @@ do { \ compiletime_assert_atomic_type(*p); \ barrier(); \ - ACCESS_ONCE(*p) = (v); \ + WRITE_ONCE(*p, v); \ } while (0) #define smp_load_acquire(p) \ ({ \ - typeof(*p) ___p1 = ACCESS_ONCE(*p); \ + typeof(*p) ___p1 = READ_ONCE(*p); \ compiletime_assert_atomic_type(*p); \ barrier(); \ ___p1; \ diff --git a/arch/ia64/include/asm/barrier.h b/arch/ia64/include/asm/barrier.h index 843ba435e43b..df896a1c41d3 100644 --- a/arch/ia64/include/asm/barrier.h +++ b/arch/ia64/include/asm/barrier.h @@ -66,12 +66,12 @@ do { \ compiletime_assert_atomic_type(*p); \ barrier(); \ - ACCESS_ONCE(*p) = (v); \ + WRITE_ONCE(*p, v); \ } while (0) #define smp_load_acquire(p) \ ({ \ - typeof(*p) ___p1 = ACCESS_ONCE(*p); \ + typeof(*p) ___p1 = READ_ONCE(*p); \ compiletime_assert_atomic_type(*p); \ barrier(); \ ___p1; \ diff --git a/arch/metag/include/asm/barrier.h b/arch/metag/include/asm/barrier.h index 5a696e507930..172b7e5efc53 100644 --- a/arch/metag/include/asm/barrier.h +++ b/arch/metag/include/asm/barrier.h @@ -90,12 +90,12 @@ static inline void fence(void) do { \ compiletime_assert_atomic_type(*p); \ smp_mb(); \ - ACCESS_ONCE(*p) = (v); \ + WRITE_ONCE(*p, v); \ } while (0) #define smp_load_acquire(p) \ ({ \ - typeof(*p) ___p1 = ACCESS_ONCE(*p); \ + typeof(*p) ___p1 = READ_ONCE(*p); \ compiletime_assert_atomic_type(*p); \ smp_mb(); \ ___p1; \ diff --git a/arch/mips/include/asm/barrier.h b/arch/mips/include/asm/barrier.h index 7ecba84656d4..752e0b86c171 100644 --- a/arch/mips/include/asm/barrier.h +++ b/arch/mips/include/asm/barrier.h @@ -133,12 +133,12 @@ do { \ compiletime_assert_atomic_type(*p); \ smp_mb(); \ - ACCESS_ONCE(*p) = (v); \ + WRITE_ONCE(*p, v); \ } while (0) #define smp_load_acquire(p) \ ({ \ - typeof(*p) ___p1 = ACCESS_ONCE(*p); \ + typeof(*p) ___p1 = READ_ONCE(*p); \ compiletime_assert_atomic_type(*p); \ smp_mb(); \ ___p1; \ diff --git a/arch/powerpc/include/asm/barrier.h b/arch/powerpc/include/asm/barrier.h index 51ccc7232042..0eca6efc0631 100644 --- a/arch/powerpc/include/asm/barrier.h +++ b/arch/powerpc/include/asm/barrier.h @@ -76,12 +76,12 @@ do { \ compiletime_assert_atomic_type(*p); \ smp_lwsync(); \ - ACCESS_ONCE(*p) = (v); \ + WRITE_ONCE(*p, v); \ } while (0) #define smp_load_acquire(p) \ ({ \ - typeof(*p) ___p1 = ACCESS_ONCE(*p); \ + typeof(*p) ___p1 = READ_ONCE(*p); \ compiletime_assert_atomic_type(*p); \ smp_lwsync(); \ ___p1; \ diff --git a/arch/s390/include/asm/barrier.h b/arch/s390/include/asm/barrier.h index e6f8615a11eb..d48fe0162331 100644 --- a/arch/s390/include/asm/barrier.h +++ b/arch/s390/include/asm/barrier.h @@ -42,12 +42,12 @@ do { \ compiletime_assert_atomic_type(*p); \ barrier(); \ - ACCESS_ONCE(*p) = (v); \ + WRITE_ONCE(*p, v); \ } while (0) #define smp_load_acquire(p) \ ({ \ - typeof(*p) ___p1 = ACCESS_ONCE(*p); \ + typeof(*p) ___p1 = READ_ONCE(*p); \ compiletime_assert_atomic_type(*p); \ barrier(); \ ___p1; \ diff --git a/arch/sparc/include/asm/barrier_64.h b/arch/sparc/include/asm/barrier_64.h index 809941e33e12..14a928601657 100644 --- a/arch/sparc/include/asm/barrier_64.h +++ b/arch/sparc/include/asm/barrier_64.h @@ -60,12 +60,12 @@ do { __asm__ __volatile__("ba,pt %%xcc, 1f\n\t" \ do { \ compiletime_assert_atomic_type(*p); \ barrier(); \ - ACCESS_ONCE(*p) = (v); \ + WRITE_ONCE(*p, v); \ } while (0) #define smp_load_acquire(p) \ ({ \ - typeof(*p) ___p1 = ACCESS_ONCE(*p); \ + typeof(*p) ___p1 = READ_ONCE(*p); \ compiletime_assert_atomic_type(*p); \ barrier(); \ ___p1; \ diff --git a/arch/x86/include/asm/barrier.h b/arch/x86/include/asm/barrier.h index e51a8f803f55..d2bcfbed11c3 100644 --- a/arch/x86/include/asm/barrier.h +++ b/arch/x86/include/asm/barrier.h @@ -57,12 +57,12 @@ do { \ compiletime_assert_atomic_type(*p); \ smp_mb(); \ - ACCESS_ONCE(*p) = (v); \ + WRITE_ONCE(*p, v); \ } while (0) #define smp_load_acquire(p) \ ({ \ - typeof(*p) ___p1 = ACCESS_ONCE(*p); \ + typeof(*p) ___p1 = READ_ONCE(*p); \ compiletime_assert_atomic_type(*p); \ smp_mb(); \ ___p1; \ @@ -74,12 +74,12 @@ do { \ do { \ compiletime_assert_atomic_type(*p); \ barrier(); \ - ACCESS_ONCE(*p) = (v); \ + WRITE_ONCE(*p, v); \ } while (0) #define smp_load_acquire(p) \ ({ \ - typeof(*p) ___p1 = ACCESS_ONCE(*p); \ + typeof(*p) ___p1 = READ_ONCE(*p); \ compiletime_assert_atomic_type(*p); \ barrier(); \ ___p1; \ diff --git a/include/asm-generic/barrier.h b/include/asm-generic/barrier.h index 55e3abc2d027..b42afada1280 100644 --- a/include/asm-generic/barrier.h +++ b/include/asm-generic/barrier.h @@ -108,12 +108,12 @@ do { \ compiletime_assert_atomic_type(*p); \ smp_mb(); \ - ACCESS_ONCE(*p) = (v); \ + WRITE_ONCE(*p, v); \ } while (0) #define smp_load_acquire(p) \ ({ \ - typeof(*p) ___p1 = ACCESS_ONCE(*p); \ + typeof(*p) ___p1 = READ_ONCE(*p); \ compiletime_assert_atomic_type(*p); \ smp_mb(); \ ___p1; \ -- cgit From 586b610e43a5ad5096640312fefa6ce931738c7d Mon Sep 17 00:00:00 2001 From: Will Deacon Date: Thu, 6 Aug 2015 17:54:38 +0100 Subject: locking, asm-generic: Rework atomic-long.h to avoid bulk code duplication We can use some (admittedly ugly) macros to generate the 32-bit and 64-bit based atomic_long implementations from the same code. Signed-off-by: Will Deacon Signed-off-by: Peter Zijlstra (Intel) Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: Waiman.Long@hp.com Cc: paulmck@linux.vnet.ibm.com Link: http://lkml.kernel.org/r/1438880084-18856-3-git-send-email-will.deacon@arm.com Signed-off-by: Ingo Molnar --- include/asm-generic/atomic-long.h | 189 ++++++++------------------------------ 1 file changed, 40 insertions(+), 149 deletions(-) (limited to 'include/asm-generic') diff --git a/include/asm-generic/atomic-long.h b/include/asm-generic/atomic-long.h index b7babf0206b8..beaea541adfb 100644 --- a/include/asm-generic/atomic-long.h +++ b/include/asm-generic/atomic-long.h @@ -23,236 +23,127 @@ typedef atomic64_t atomic_long_t; #define ATOMIC_LONG_INIT(i) ATOMIC64_INIT(i) +#define ATOMIC_LONG_PFX(x) atomic64 ## x -static inline long atomic_long_read(atomic_long_t *l) -{ - atomic64_t *v = (atomic64_t *)l; - - return (long)atomic64_read(v); -} - -static inline void atomic_long_set(atomic_long_t *l, long i) -{ - atomic64_t *v = (atomic64_t *)l; - - atomic64_set(v, i); -} - -static inline void atomic_long_inc(atomic_long_t *l) -{ - atomic64_t *v = (atomic64_t *)l; - - atomic64_inc(v); -} - -static inline void atomic_long_dec(atomic_long_t *l) -{ - atomic64_t *v = (atomic64_t *)l; - - atomic64_dec(v); -} - -static inline void atomic_long_add(long i, atomic_long_t *l) -{ - atomic64_t *v = (atomic64_t *)l; - - atomic64_add(i, v); -} - -static inline void atomic_long_sub(long i, atomic_long_t *l) -{ - atomic64_t *v = (atomic64_t *)l; - - atomic64_sub(i, v); -} - -static inline int atomic_long_sub_and_test(long i, atomic_long_t *l) -{ - atomic64_t *v = (atomic64_t *)l; - - return atomic64_sub_and_test(i, v); -} - -static inline int atomic_long_dec_and_test(atomic_long_t *l) -{ - atomic64_t *v = (atomic64_t *)l; - - return atomic64_dec_and_test(v); -} - -static inline int atomic_long_inc_and_test(atomic_long_t *l) -{ - atomic64_t *v = (atomic64_t *)l; - - return atomic64_inc_and_test(v); -} - -static inline int atomic_long_add_negative(long i, atomic_long_t *l) -{ - atomic64_t *v = (atomic64_t *)l; - - return atomic64_add_negative(i, v); -} - -static inline long atomic_long_add_return(long i, atomic_long_t *l) -{ - atomic64_t *v = (atomic64_t *)l; - - return (long)atomic64_add_return(i, v); -} - -static inline long atomic_long_sub_return(long i, atomic_long_t *l) -{ - atomic64_t *v = (atomic64_t *)l; - - return (long)atomic64_sub_return(i, v); -} - -static inline long atomic_long_inc_return(atomic_long_t *l) -{ - atomic64_t *v = (atomic64_t *)l; - - return (long)atomic64_inc_return(v); -} - -static inline long atomic_long_dec_return(atomic_long_t *l) -{ - atomic64_t *v = (atomic64_t *)l; - - return (long)atomic64_dec_return(v); -} - -static inline long atomic_long_add_unless(atomic_long_t *l, long a, long u) -{ - atomic64_t *v = (atomic64_t *)l; - - return (long)atomic64_add_unless(v, a, u); -} - -#define atomic_long_inc_not_zero(l) atomic64_inc_not_zero((atomic64_t *)(l)) - -#define atomic_long_cmpxchg(l, old, new) \ - (atomic64_cmpxchg((atomic64_t *)(l), (old), (new))) -#define atomic_long_xchg(v, new) \ - (atomic64_xchg((atomic64_t *)(v), (new))) - -#else /* BITS_PER_LONG == 64 */ +#else typedef atomic_t atomic_long_t; #define ATOMIC_LONG_INIT(i) ATOMIC_INIT(i) +#define ATOMIC_LONG_PFX(x) atomic ## x + +#endif + static inline long atomic_long_read(atomic_long_t *l) { - atomic_t *v = (atomic_t *)l; + ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l; - return (long)atomic_read(v); + return (long)ATOMIC_LONG_PFX(_read)(v); } static inline void atomic_long_set(atomic_long_t *l, long i) { - atomic_t *v = (atomic_t *)l; + ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l; - atomic_set(v, i); + ATOMIC_LONG_PFX(_set)(v, i); } static inline void atomic_long_inc(atomic_long_t *l) { - atomic_t *v = (atomic_t *)l; + ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l; - atomic_inc(v); + ATOMIC_LONG_PFX(_inc)(v); } static inline void atomic_long_dec(atomic_long_t *l) { - atomic_t *v = (atomic_t *)l; + ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l; - atomic_dec(v); + ATOMIC_LONG_PFX(_dec)(v); } static inline void atomic_long_add(long i, atomic_long_t *l) { - atomic_t *v = (atomic_t *)l; + ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l; - atomic_add(i, v); + ATOMIC_LONG_PFX(_add)(i, v); } static inline void atomic_long_sub(long i, atomic_long_t *l) { - atomic_t *v = (atomic_t *)l; + ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l; - atomic_sub(i, v); + ATOMIC_LONG_PFX(_sub)(i, v); } static inline int atomic_long_sub_and_test(long i, atomic_long_t *l) { - atomic_t *v = (atomic_t *)l; + ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l; - return atomic_sub_and_test(i, v); + return ATOMIC_LONG_PFX(_sub_and_test)(i, v); } static inline int atomic_long_dec_and_test(atomic_long_t *l) { - atomic_t *v = (atomic_t *)l; + ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l; - return atomic_dec_and_test(v); + return ATOMIC_LONG_PFX(_dec_and_test)(v); } static inline int atomic_long_inc_and_test(atomic_long_t *l) { - atomic_t *v = (atomic_t *)l; + ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l; - return atomic_inc_and_test(v); + return ATOMIC_LONG_PFX(_inc_and_test)(v); } static inline int atomic_long_add_negative(long i, atomic_long_t *l) { - atomic_t *v = (atomic_t *)l; + ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l; - return atomic_add_negative(i, v); + return ATOMIC_LONG_PFX(_add_negative)(i, v); } static inline long atomic_long_add_return(long i, atomic_long_t *l) { - atomic_t *v = (atomic_t *)l; + ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l; - return (long)atomic_add_return(i, v); + return (long)ATOMIC_LONG_PFX(_add_return)(i, v); } static inline long atomic_long_sub_return(long i, atomic_long_t *l) { - atomic_t *v = (atomic_t *)l; + ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l; - return (long)atomic_sub_return(i, v); + return (long)ATOMIC_LONG_PFX(_sub_return)(i, v); } static inline long atomic_long_inc_return(atomic_long_t *l) { - atomic_t *v = (atomic_t *)l; + ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l; - return (long)atomic_inc_return(v); + return (long)ATOMIC_LONG_PFX(_inc_return)(v); } static inline long atomic_long_dec_return(atomic_long_t *l) { - atomic_t *v = (atomic_t *)l; + ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l; - return (long)atomic_dec_return(v); + return (long)ATOMIC_LONG_PFX(_dec_return)(v); } static inline long atomic_long_add_unless(atomic_long_t *l, long a, long u) { - atomic_t *v = (atomic_t *)l; + ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l; - return (long)atomic_add_unless(v, a, u); + return (long)ATOMIC_LONG_PFX(_add_unless)(v, a, u); } -#define atomic_long_inc_not_zero(l) atomic_inc_not_zero((atomic_t *)(l)) - +#define atomic_long_inc_not_zero(l) \ + ATOMIC_LONG_PFX(_inc_not_zero)((ATOMIC_LONG_PFX(_t) *)(l)) #define atomic_long_cmpxchg(l, old, new) \ - (atomic_cmpxchg((atomic_t *)(l), (old), (new))) + (ATOMIC_LONG_PFX(_cmpxchg)((ATOMIC_LONG_PFX(_t) *)(l), (old), (new))) #define atomic_long_xchg(v, new) \ - (atomic_xchg((atomic_t *)(v), (new))) - -#endif /* BITS_PER_LONG == 64 */ + (ATOMIC_LONG_PFX(_xchg)((ATOMIC_LONG_PFX(_t) *)(v), (new))) #endif /* _ASM_GENERIC_ATOMIC_LONG_H */ -- cgit From 6d79ef2d30ee5af7315535d1e7bf6fce0008f815 Mon Sep 17 00:00:00 2001 From: Will Deacon Date: Thu, 6 Aug 2015 17:54:39 +0100 Subject: locking, asm-generic: Add _{relaxed|acquire|release}() variants for 'atomic_long_t' This patch adds 'atomic_long_t' wrappers for the new relaxed atomic operations. Signed-off-by: Will Deacon Signed-off-by: Peter Zijlstra (Intel) Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: Waiman.Long@hp.com Cc: paulmck@linux.vnet.ibm.com Link: http://lkml.kernel.org/r/1438880084-18856-4-git-send-email-will.deacon@arm.com Signed-off-by: Ingo Molnar --- include/asm-generic/atomic-long.h | 86 +++++++++++++++++++++++++++------------ 1 file changed, 59 insertions(+), 27 deletions(-) (limited to 'include/asm-generic') diff --git a/include/asm-generic/atomic-long.h b/include/asm-generic/atomic-long.h index beaea541adfb..a94cbebbc33d 100644 --- a/include/asm-generic/atomic-long.h +++ b/include/asm-generic/atomic-long.h @@ -34,19 +34,69 @@ typedef atomic_t atomic_long_t; #endif -static inline long atomic_long_read(atomic_long_t *l) -{ - ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l; - - return (long)ATOMIC_LONG_PFX(_read)(v); +#define ATOMIC_LONG_READ_OP(mo) \ +static inline long atomic_long_read##mo(atomic_long_t *l) \ +{ \ + ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l; \ + \ + return (long)ATOMIC_LONG_PFX(_read##mo)(v); \ } +ATOMIC_LONG_READ_OP() +ATOMIC_LONG_READ_OP(_acquire) -static inline void atomic_long_set(atomic_long_t *l, long i) -{ - ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l; +#undef ATOMIC_LONG_READ_OP - ATOMIC_LONG_PFX(_set)(v, i); +#define ATOMIC_LONG_SET_OP(mo) \ +static inline void atomic_long_set##mo(atomic_long_t *l, long i) \ +{ \ + ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l; \ + \ + ATOMIC_LONG_PFX(_set##mo)(v, i); \ +} +ATOMIC_LONG_SET_OP() +ATOMIC_LONG_SET_OP(_release) + +#undef ATOMIC_LONG_SET_OP + +#define ATOMIC_LONG_ADD_SUB_OP(op, mo) \ +static inline long \ +atomic_long_##op##_return##mo(long i, atomic_long_t *l) \ +{ \ + ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l; \ + \ + return (long)ATOMIC_LONG_PFX(_##op##_return##mo)(i, v); \ } +ATOMIC_LONG_ADD_SUB_OP(add,) +ATOMIC_LONG_ADD_SUB_OP(add, _relaxed) +ATOMIC_LONG_ADD_SUB_OP(add, _acquire) +ATOMIC_LONG_ADD_SUB_OP(add, _release) +ATOMIC_LONG_ADD_SUB_OP(sub,) +ATOMIC_LONG_ADD_SUB_OP(sub, _relaxed) +ATOMIC_LONG_ADD_SUB_OP(sub, _acquire) +ATOMIC_LONG_ADD_SUB_OP(sub, _release) + +#undef ATOMIC_LONG_ADD_SUB_OP + +#define atomic_long_cmpxchg_relaxed(l, old, new) \ + (ATOMIC_LONG_PFX(_cmpxchg_relaxed)((ATOMIC_LONG_PFX(_t) *)(l), \ + (old), (new))) +#define atomic_long_cmpxchg_acquire(l, old, new) \ + (ATOMIC_LONG_PFX(_cmpxchg_acquire)((ATOMIC_LONG_PFX(_t) *)(l), \ + (old), (new))) +#define atomic_long_cmpxchg_release(l, old, new) \ + (ATOMIC_LONG_PFX(_cmpxchg_release)((ATOMIC_LONG_PFX(_t) *)(l), \ + (old), (new))) +#define atomic_long_cmpxchg(l, old, new) \ + (ATOMIC_LONG_PFX(_cmpxchg)((ATOMIC_LONG_PFX(_t) *)(l), (old), (new))) + +#define atomic_long_xchg_relaxed(v, new) \ + (ATOMIC_LONG_PFX(_xchg_relaxed)((ATOMIC_LONG_PFX(_t) *)(v), (new))) +#define atomic_long_xchg_acquire(v, new) \ + (ATOMIC_LONG_PFX(_xchg_acquire)((ATOMIC_LONG_PFX(_t) *)(v), (new))) +#define atomic_long_xchg_release(v, new) \ + (ATOMIC_LONG_PFX(_xchg_release)((ATOMIC_LONG_PFX(_t) *)(v), (new))) +#define atomic_long_xchg(v, new) \ + (ATOMIC_LONG_PFX(_xchg)((ATOMIC_LONG_PFX(_t) *)(v), (new))) static inline void atomic_long_inc(atomic_long_t *l) { @@ -104,20 +154,6 @@ static inline int atomic_long_add_negative(long i, atomic_long_t *l) return ATOMIC_LONG_PFX(_add_negative)(i, v); } -static inline long atomic_long_add_return(long i, atomic_long_t *l) -{ - ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l; - - return (long)ATOMIC_LONG_PFX(_add_return)(i, v); -} - -static inline long atomic_long_sub_return(long i, atomic_long_t *l) -{ - ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l; - - return (long)ATOMIC_LONG_PFX(_sub_return)(i, v); -} - static inline long atomic_long_inc_return(atomic_long_t *l) { ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l; @@ -141,9 +177,5 @@ static inline long atomic_long_add_unless(atomic_long_t *l, long a, long u) #define atomic_long_inc_not_zero(l) \ ATOMIC_LONG_PFX(_inc_not_zero)((ATOMIC_LONG_PFX(_t) *)(l)) -#define atomic_long_cmpxchg(l, old, new) \ - (ATOMIC_LONG_PFX(_cmpxchg)((ATOMIC_LONG_PFX(_t) *)(l), (old), (new))) -#define atomic_long_xchg(v, new) \ - (ATOMIC_LONG_PFX(_xchg)((ATOMIC_LONG_PFX(_t) *)(v), (new))) #endif /* _ASM_GENERIC_ATOMIC_LONG_H */ -- cgit From 2b2a85a4d3534b8884fcfa5bb52837f0e1c672bc Mon Sep 17 00:00:00 2001 From: Will Deacon Date: Thu, 6 Aug 2015 17:54:41 +0100 Subject: locking/qrwlock: Implement queue_write_unlock() using smp_store_release() Since the following commit: 536fa402221f ("compiler: Allow 1- and 2-byte smp_load_acquire() and smp_store_release()") smp_store_release() supports byte accesses, so use that in writer unlock and remove the conditional macro override. Signed-off-by: Will Deacon Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Waiman Long Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: paulmck@linux.vnet.ibm.com Link: http://lkml.kernel.org/r/1438880084-18856-6-git-send-email-will.deacon@arm.com Signed-off-by: Ingo Molnar --- arch/x86/include/asm/qrwlock.h | 10 ---------- include/asm-generic/qrwlock.h | 9 +-------- 2 files changed, 1 insertion(+), 18 deletions(-) (limited to 'include/asm-generic') diff --git a/arch/x86/include/asm/qrwlock.h b/arch/x86/include/asm/qrwlock.h index a8810bf135ab..c537cbb038a7 100644 --- a/arch/x86/include/asm/qrwlock.h +++ b/arch/x86/include/asm/qrwlock.h @@ -2,16 +2,6 @@ #define _ASM_X86_QRWLOCK_H #include - -#ifndef CONFIG_X86_PPRO_FENCE -#define queued_write_unlock queued_write_unlock -static inline void queued_write_unlock(struct qrwlock *lock) -{ - barrier(); - ACCESS_ONCE(*(u8 *)&lock->cnts) = 0; -} -#endif - #include #endif /* _ASM_X86_QRWLOCK_H */ diff --git a/include/asm-generic/qrwlock.h b/include/asm-generic/qrwlock.h index deb9e8b0eb9e..eb673dde8879 100644 --- a/include/asm-generic/qrwlock.h +++ b/include/asm-generic/qrwlock.h @@ -134,21 +134,14 @@ static inline void queued_read_unlock(struct qrwlock *lock) atomic_sub(_QR_BIAS, &lock->cnts); } -#ifndef queued_write_unlock /** * queued_write_unlock - release write lock of a queue rwlock * @lock : Pointer to queue rwlock structure */ static inline void queued_write_unlock(struct qrwlock *lock) { - /* - * If the writer field is atomic, it can be cleared directly. - * Otherwise, an atomic subtraction will be used to clear it. - */ - smp_mb__before_atomic(); - atomic_sub(_QW_LOCKED, &lock->cnts); + smp_store_release((u8 *)&lock->cnts, 0); } -#endif /* * Remapping rwlock architecture specific functions to the corresponding -- cgit From 77e430e3e45662b696dc49aa53ea0f7ac63f2574 Mon Sep 17 00:00:00 2001 From: Will Deacon Date: Thu, 6 Aug 2015 17:54:42 +0100 Subject: locking/qrwlock: Make use of _{acquire|release|relaxed}() atomics The qrwlock implementation is slightly heavy in its use of memory barriers, mainly through the use of _cmpxchg() and _return() atomics, which imply full barrier semantics. This patch modifies the qrwlock code to use the more relaxed atomic routines so that we can reduce the unnecessary barrier overhead on weakly-ordered architectures. Signed-off-by: Will Deacon Signed-off-by: Peter Zijlstra (Intel) Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: Waiman.Long@hp.com Cc: paulmck@linux.vnet.ibm.com Link: http://lkml.kernel.org/r/1438880084-18856-7-git-send-email-will.deacon@arm.com Signed-off-by: Ingo Molnar --- include/asm-generic/qrwlock.h | 13 ++++++------- kernel/locking/qrwlock.c | 24 ++++++++++++------------ 2 files changed, 18 insertions(+), 19 deletions(-) (limited to 'include/asm-generic') diff --git a/include/asm-generic/qrwlock.h b/include/asm-generic/qrwlock.h index eb673dde8879..54a8e65e18b6 100644 --- a/include/asm-generic/qrwlock.h +++ b/include/asm-generic/qrwlock.h @@ -68,7 +68,7 @@ static inline int queued_read_trylock(struct qrwlock *lock) cnts = atomic_read(&lock->cnts); if (likely(!(cnts & _QW_WMASK))) { - cnts = (u32)atomic_add_return(_QR_BIAS, &lock->cnts); + cnts = (u32)atomic_add_return_acquire(_QR_BIAS, &lock->cnts); if (likely(!(cnts & _QW_WMASK))) return 1; atomic_sub(_QR_BIAS, &lock->cnts); @@ -89,8 +89,8 @@ static inline int queued_write_trylock(struct qrwlock *lock) if (unlikely(cnts)) return 0; - return likely(atomic_cmpxchg(&lock->cnts, - cnts, cnts | _QW_LOCKED) == cnts); + return likely(atomic_cmpxchg_acquire(&lock->cnts, + cnts, cnts | _QW_LOCKED) == cnts); } /** * queued_read_lock - acquire read lock of a queue rwlock @@ -100,7 +100,7 @@ static inline void queued_read_lock(struct qrwlock *lock) { u32 cnts; - cnts = atomic_add_return(_QR_BIAS, &lock->cnts); + cnts = atomic_add_return_acquire(_QR_BIAS, &lock->cnts); if (likely(!(cnts & _QW_WMASK))) return; @@ -115,7 +115,7 @@ static inline void queued_read_lock(struct qrwlock *lock) static inline void queued_write_lock(struct qrwlock *lock) { /* Optimize for the unfair lock case where the fair flag is 0. */ - if (atomic_cmpxchg(&lock->cnts, 0, _QW_LOCKED) == 0) + if (atomic_cmpxchg_acquire(&lock->cnts, 0, _QW_LOCKED) == 0) return; queued_write_lock_slowpath(lock); @@ -130,8 +130,7 @@ static inline void queued_read_unlock(struct qrwlock *lock) /* * Atomically decrement the reader count */ - smp_mb__before_atomic(); - atomic_sub(_QR_BIAS, &lock->cnts); + (void)atomic_sub_return_release(_QR_BIAS, &lock->cnts); } /** diff --git a/kernel/locking/qrwlock.c b/kernel/locking/qrwlock.c index 6a7a3b8d5ac9..f17a3e3b3550 100644 --- a/kernel/locking/qrwlock.c +++ b/kernel/locking/qrwlock.c @@ -55,7 +55,7 @@ rspin_until_writer_unlock(struct qrwlock *lock, u32 cnts) { while ((cnts & _QW_WMASK) == _QW_LOCKED) { cpu_relax_lowlatency(); - cnts = smp_load_acquire((u32 *)&lock->cnts); + cnts = atomic_read_acquire(&lock->cnts); } } @@ -74,8 +74,9 @@ void queued_read_lock_slowpath(struct qrwlock *lock, u32 cnts) * Readers in interrupt context will get the lock immediately * if the writer is just waiting (not holding the lock yet). * The rspin_until_writer_unlock() function returns immediately - * in this case. Otherwise, they will spin until the lock - * is available without waiting in the queue. + * in this case. Otherwise, they will spin (with ACQUIRE + * semantics) until the lock is available without waiting in + * the queue. */ rspin_until_writer_unlock(lock, cnts); return; @@ -88,12 +89,11 @@ void queued_read_lock_slowpath(struct qrwlock *lock, u32 cnts) arch_spin_lock(&lock->lock); /* - * At the head of the wait queue now, increment the reader count - * and wait until the writer, if it has the lock, has gone away. - * At ths stage, it is not possible for a writer to remain in the - * waiting state (_QW_WAITING). So there won't be any deadlock. + * The ACQUIRE semantics of the following spinning code ensure + * that accesses can't leak upwards out of our subsequent critical + * section in the case that the lock is currently held for write. */ - cnts = atomic_add_return(_QR_BIAS, &lock->cnts) - _QR_BIAS; + cnts = atomic_add_return_acquire(_QR_BIAS, &lock->cnts) - _QR_BIAS; rspin_until_writer_unlock(lock, cnts); /* @@ -116,7 +116,7 @@ void queued_write_lock_slowpath(struct qrwlock *lock) /* Try to acquire the lock directly if no reader is present */ if (!atomic_read(&lock->cnts) && - (atomic_cmpxchg(&lock->cnts, 0, _QW_LOCKED) == 0)) + (atomic_cmpxchg_acquire(&lock->cnts, 0, _QW_LOCKED) == 0)) goto unlock; /* @@ -127,7 +127,7 @@ void queued_write_lock_slowpath(struct qrwlock *lock) struct __qrwlock *l = (struct __qrwlock *)lock; if (!READ_ONCE(l->wmode) && - (cmpxchg(&l->wmode, 0, _QW_WAITING) == 0)) + (cmpxchg_relaxed(&l->wmode, 0, _QW_WAITING) == 0)) break; cpu_relax_lowlatency(); @@ -137,8 +137,8 @@ void queued_write_lock_slowpath(struct qrwlock *lock) for (;;) { cnts = atomic_read(&lock->cnts); if ((cnts == _QW_WAITING) && - (atomic_cmpxchg(&lock->cnts, _QW_WAITING, - _QW_LOCKED) == _QW_WAITING)) + (atomic_cmpxchg_acquire(&lock->cnts, _QW_WAITING, + _QW_LOCKED) == _QW_WAITING)) break; cpu_relax_lowlatency(); -- cgit