On Wed, 2017-11-22 at 13:11 +0800, Ming Lei wrote: > diff --git a/block/blk-mq.c b/block/blk-mq.c > index 11097477eeab..3d3797327491 100644 > --- a/block/blk-mq.c > +++ b/block/blk-mq.c > @@ -161,6 +161,8 @@ void blk_freeze_queue(struct request_queue *q) > * exported to drivers as the only user for unfreeze is blk_mq. > */ > blk_freeze_queue_start(q); > + if (!q->mq_ops) > + blk_drain_queue(q); > blk_mq_freeze_queue_wait(q); > }
Since q_usage_counter now tracks legacy requests, is there any reason why we still need __blk_drain_queue()? Have you considered to eliminate __blk_drain_queue() and to call blk_run_queue() from inside blk_freeze_queue() instead of calling blk_drain_queue()? I'm asking this because blk_mq_freeze_queue_wait() uses a more efficient mechanism (wait_event()) than __blk_drain_queue() (while (...) msleep(10)). Bart.
