rwsem: more agressive lock stealing in rwsem_down_write_failed
authorMichel Lespinasse <walken@google.com>
Tue, 7 May 2013 13:45:54 +0000 (06:45 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Tue, 7 May 2013 14:20:16 +0000 (07:20 -0700)
commited00f64346631dff035adfb9b0240daaa8b46c4e
tree61afae642c8a3333587bf73a8dc19849416e243d
parent023fe4f712028d25b42d31984abae1f3d3f0e3e2
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: Michel Lespinasse <walken@google.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Reviewed-by: Peter Hurley <peter@hurleysoftware.com>
Acked-by: Davidlohr Bueso <davidlohr.bueso@hp.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
lib/rwsem.c