Commit ed00f643 authored by Michel Lespinasse's avatar Michel Lespinasse Committed by Linus Torvalds

rwsem: more agressive lock stealing in rwsem_down_write_failed

Some small code simplifications can be achieved by doing more agressive
lock stealing:

- When rwsem_down_write_failed() notices that there are no active locks
  (and thus no thread to wake us if we decided to sleep), it used to wake
  the first queued process. However, stealing the lock is also sufficient
  to deal with this case, so we don't need this check anymore.

- In try_get_writer_sem(), we can steal the lock even when the first waiter
  is a reader. This is correct because the code path that wakes readers is
  protected by the wait_lock. As to the performance effects of this change,
  they are expected to be minimal: readers are still granted the lock
  (rather than having to acquire it themselves) when they reach the front
  of the wait queue, so we have essentially the same behavior as in
  rwsem-spinlock.
Signed-off-by: default avatarMichel Lespinasse <walken@google.com>
Reviewed-by: default avatarRik van Riel <riel@redhat.com>
Reviewed-by: default avatarPeter Hurley <peter@hurleysoftware.com>
Acked-by: default avatarDavidlohr Bueso <davidlohr.bueso@hp.com>
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
parent 023fe4f7
...@@ -143,20 +143,12 @@ __rwsem_do_wake(struct rw_semaphore *sem, int wake_type) ...@@ -143,20 +143,12 @@ __rwsem_do_wake(struct rw_semaphore *sem, int wake_type)
} }
/* Try to get write sem, caller holds sem->wait_lock: */ /* Try to get write sem, caller holds sem->wait_lock: */
static int try_get_writer_sem(struct rw_semaphore *sem, static int try_get_writer_sem(struct rw_semaphore *sem)
struct rwsem_waiter *waiter)
{ {
struct rwsem_waiter *fwaiter;
long oldcount, adjustment; long oldcount, adjustment;
/* only steal when first waiter is writing */
fwaiter = list_entry(sem->wait_list.next, struct rwsem_waiter, list);
if (fwaiter->type != RWSEM_WAITING_FOR_WRITE)
return 0;
adjustment = RWSEM_ACTIVE_WRITE_BIAS; adjustment = RWSEM_ACTIVE_WRITE_BIAS;
/* Only one waiter in the queue: */ if (list_is_singular(&sem->wait_list))
if (fwaiter == waiter && waiter->list.next == &sem->wait_list)
adjustment -= RWSEM_WAITING_BIAS; adjustment -= RWSEM_WAITING_BIAS;
try_again_write: try_again_write:
...@@ -233,23 +225,18 @@ struct rw_semaphore __sched *rwsem_down_write_failed(struct rw_semaphore *sem) ...@@ -233,23 +225,18 @@ struct rw_semaphore __sched *rwsem_down_write_failed(struct rw_semaphore *sem)
/* we're now waiting on the lock, but no longer actively locking */ /* we're now waiting on the lock, but no longer actively locking */
count = rwsem_atomic_update(adjustment, sem); count = rwsem_atomic_update(adjustment, sem);
/* If there are no active locks, wake the front queued process(es) up. /* If there were already threads queued before us and there are no
* * active writers, the lock must be read owned; so we try to wake
* Alternatively, if we're called from a failed down_write(), there * any read locks that were queued ahead of us. */
* were already threads queued before us and there are no active if (count > RWSEM_WAITING_BIAS &&
* writers, the lock must be read owned; so we try to wake any read adjustment == -RWSEM_ACTIVE_WRITE_BIAS)
* locks that were queued ahead of us. */
if (count == RWSEM_WAITING_BIAS)
sem = __rwsem_do_wake(sem, RWSEM_WAKE_NO_ACTIVE);
else if (count > RWSEM_WAITING_BIAS &&
adjustment == -RWSEM_ACTIVE_WRITE_BIAS)
sem = __rwsem_do_wake(sem, RWSEM_WAKE_READ_OWNED); sem = __rwsem_do_wake(sem, RWSEM_WAKE_READ_OWNED);
/* wait until we successfully acquire the lock */ /* wait until we successfully acquire the lock */
while (true) { while (true) {
set_task_state(tsk, TASK_UNINTERRUPTIBLE); set_task_state(tsk, TASK_UNINTERRUPTIBLE);
if (try_get_writer_sem(sem, &waiter)) if (try_get_writer_sem(sem))
break; break;
raw_spin_unlock_irq(&sem->wait_lock); raw_spin_unlock_irq(&sem->wait_lock);
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment