block: avoid ordered task state change for polled IO
authorJens Axboe <axboe@kernel.dk>
Fri, 16 Nov 2018 15:37:34 +0000 (08:37 -0700)
committerJens Axboe <axboe@kernel.dk>
Mon, 19 Nov 2018 15:34:49 +0000 (08:34 -0700)
commit849a370016a5489c49253338507ee6cc4a08df4b
tree3e8794bb4312b0fa73e3092849b8b6e163437a2d
parenta4668d9ba4be1ca9f4a39798ba3419fdfef0750d
block: avoid ordered task state change for polled IO

For the core poll helper, the task state setting don't need to imply any
atomics, as it's the current task itself that is being modified and
we're not going to sleep.

For IRQ driven, the wakeup path have the necessary barriers to not need
us using the heavy handed version of the task state setting.

Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
block/blk-mq.c
fs/block_dev.c
fs/iomap.c
mm/page_io.c