[tip:locking/core] locking/qspinlock: Use atomic_sub_return_release() in queued_spin_unlock()
From: tip-bot for Pan Xinhui
Date: Wed Jun 08 2016 - 10:27:52 EST
Commit-ID: ca50e426f96c905e7d14a9c7a6bd4e0330516047
Gitweb: http://git.kernel.org/tip/ca50e426f96c905e7d14a9c7a6bd4e0330516047
Author: Pan Xinhui <xinhui.pan@xxxxxxxxxxxxxxxxxx>
AuthorDate: Fri, 3 Jun 2016 16:38:14 +0800
Committer: Ingo Molnar <mingo@xxxxxxxxxx>
CommitDate: Wed, 8 Jun 2016 15:17:01 +0200
locking/qspinlock: Use atomic_sub_return_release() in queued_spin_unlock()
The existing version uses a heavy barrier while only release semantics
is required. So use atomic_sub_return_release() instead.
Suggested-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx>
Signed-off-by: Pan Xinhui <xinhui.pan@xxxxxxxxxxxxxxxxxx>
Signed-off-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx>
Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
Cc: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
Cc: Paul E. McKenney <paulmck@xxxxxxxxxxxxxxxxxx>
Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Cc: arnd@xxxxxxxx
Cc: waiman.long@xxxxxx
Link: http://lkml.kernel.org/r/1464943094-3129-1-git-send-email-xinhui.pan@xxxxxxxxxxxxxxxxxx
Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
---
include/asm-generic/qspinlock.h | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/include/asm-generic/qspinlock.h b/include/asm-generic/qspinlock.h
index 05f05f1..9f0681b 100644
--- a/include/asm-generic/qspinlock.h
+++ b/include/asm-generic/qspinlock.h
@@ -111,10 +111,9 @@ static __always_inline void queued_spin_lock(struct qspinlock *lock)
static __always_inline void queued_spin_unlock(struct qspinlock *lock)
{
/*
- * smp_mb__before_atomic() in order to guarantee release semantics
+ * unlock() needs release semantics:
*/
- smp_mb__before_atomic();
- atomic_sub(_Q_LOCKED_VAL, &lock->val);
+ (void)atomic_sub_return_release(_Q_LOCKED_VAL, &lock->val);
}
#endif