[5/5] rq-qos: use a mb for got_token
diff mbox series

Message ID 20190716201929.79142-6-josef@toxicpanda.com
State New
Headers show
  • rq-qos memory barrier shenanigans
Related show

Commit Message

Josef Bacik July 16, 2019, 8:19 p.m. UTC
Oleg noticed that our checking of data.got_token is unsafe in the
cleanup case, and should really use a memory barrier.  Use a wmb on the
write side, and a rmb() on the read side.  We don't need one in the main
loop since we're saved by set_current_state().

Signed-off-by: Josef Bacik <josef@toxicpanda.com>
 block/blk-rq-qos.c | 3 +++
 1 file changed, 3 insertions(+)

diff mbox series

diff --git a/block/blk-rq-qos.c b/block/blk-rq-qos.c
index c450b8952eae..3954c0dc1443 100644
--- a/block/blk-rq-qos.c
+++ b/block/blk-rq-qos.c
@@ -202,6 +202,7 @@  static int rq_qos_wake_function(struct wait_queue_entry *curr,
 		return -1;
 	data->got_token = true;
+	smp_wmb();
 	return 1;
@@ -246,6 +247,7 @@  void rq_qos_wait(struct rq_wait *rqw, void *private_data,
 	prepare_to_wait_exclusive(&rqw->wait, &data.wq, TASK_UNINTERRUPTIBLE);
 	has_sleeper = !wq_has_single_sleeper(&rqw->wait);
 	do {
+		/* The memory barrier in set_task_state saves us here. */
 		if (data.got_token)
 		if (!has_sleeper && acquire_inflight_cb(rqw, private_data)) {
@@ -256,6 +258,7 @@  void rq_qos_wait(struct rq_wait *rqw, void *private_data,
 			 * which means we now have two. Put our local token
 			 * and wake anyone else potentially waiting for one.
+			smp_rmb();
 			if (data.got_token)
 				cleanup_cb(rqw, private_data);