diff options
| author | Davidlohr Bueso <[email protected]> | 2015-01-30 01:14:24 -0800 |
|---|---|---|
| committer | Ingo Molnar <[email protected]> | 2015-02-18 16:57:10 +0100 |
| commit | 49e4b2bcf7b812e985e65b6c8a0255b1520a6e7e (patch) | |
| tree | 72b8d5e9e0ce330251a310e05b5633fc8f505371 | |
| parent | a21294644623ee41034db60e93aaebed4db0e57b (diff) | |
locking/rwsem: Document barrier need when waking tasks
The need for the smp_mb() in __rwsem_do_wake() should be
properly documented. Applies to both xadd and spinlock
variants.
Signed-off-by: Davidlohr Bueso <[email protected]>
Signed-off-by: Peter Zijlstra (Intel) <[email protected]>
Cc: Jason Low <[email protected]>
Cc: Linus Torvalds <[email protected]>
Cc: Michel Lespinasse <[email protected]>
Cc: Paul E. McKenney <[email protected]>
Cc: Tim Chen <[email protected]>
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Ingo Molnar <[email protected]>
| -rw-r--r-- | kernel/locking/rwsem-spinlock.c | 7 | ||||
| -rw-r--r-- | kernel/locking/rwsem-xadd.c | 7 |
2 files changed, 14 insertions, 0 deletions
diff --git a/kernel/locking/rwsem-spinlock.c b/kernel/locking/rwsem-spinlock.c index 2555ae15ec14..3a5048572065 100644 --- a/kernel/locking/rwsem-spinlock.c +++ b/kernel/locking/rwsem-spinlock.c @@ -85,6 +85,13 @@ __rwsem_do_wake(struct rw_semaphore *sem, int wakewrite) list_del(&waiter->list); tsk = waiter->task; + /* + * Make sure we do not wakeup the next reader before + * setting the nil condition to grant the next reader; + * otherwise we could miss the wakeup on the other + * side and end up sleeping again. See the pairing + * in rwsem_down_read_failed(). + */ smp_mb(); waiter->task = NULL; wake_up_process(tsk); diff --git a/kernel/locking/rwsem-xadd.c b/kernel/locking/rwsem-xadd.c index 2f7cc4076f50..82aba467564a 100644 --- a/kernel/locking/rwsem-xadd.c +++ b/kernel/locking/rwsem-xadd.c @@ -186,6 +186,13 @@ __rwsem_do_wake(struct rw_semaphore *sem, enum rwsem_wake_type wake_type) waiter = list_entry(next, struct rwsem_waiter, list); next = waiter->list.next; tsk = waiter->task; + /* + * Make sure we do not wakeup the next reader before + * setting the nil condition to grant the next reader; + * otherwise we could miss the wakeup on the other + * side and end up sleeping again. See the pairing + * in rwsem_down_read_failed(). + */ smp_mb(); waiter->task = NULL; wake_up_process(tsk); |