io_uring: get rid of atomic FAA for cq_timeouts
authorPavel Begunkov <asml.silence@gmail.com>
Thu, 30 Jul 2020 15:43:50 +0000 (18:43 +0300)
committerJens Axboe <axboe@kernel.dk>
Thu, 30 Jul 2020 17:42:21 +0000 (11:42 -0600)
commit01cec8c18f5ad9c27eee9f21439072832181039e
tree5671fb2c3e87266395efd49695e7af7456ac3aa1
parent4693014340808e7f099e302c1dc40e9d79ff7667
io_uring: get rid of atomic FAA for cq_timeouts

If ->cq_timeouts modifications are done under ->completion_lock, we
don't really nee any fetch-and-add and other complex atomics. Replace it
with non-atomic FAA, that saves an implicit full memory barrier.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
fs/io_uring.c