blk-mq: mark flush request as IDLE in flush_end_io()

Mark flush request as IDLE in its .end_io(), aligning it with how normal
requests behave. The flush request stays in in-flight tags if we're not
using an IO scheduler, so we need to change its state into IDLE.
Otherwise, we will hang in blk_mq_tagset_wait_completed_request() during
error recovery because flush the request state is kept as COMPLETED.

Reported-by: Yi Zhang <yi.zhang@redhat.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Tested-by: Yi Zhang <yi.zhang@redhat.com>
Cc: Chao Leng <lengchao@huawei.com>
Cc: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Jens Axboe <axboe@kernel.dk>

diff --git a/block/blk-flush.c b/block/blk-flush.c
index 53abb5c73d99f90a90033855b60f2d27e3546104..e32958f0b687506c3815c1105ffa3acf62b7df6b 100644 (file)
--- a/block/blk-flush.c
+++ b/block/blk-flush.c
@@ -225,6 +225,7 @@ static void flush_end_io(struct request *flush_rq, blk_status_t error)
        /* release the tag's ownership to the req cloned from */
        spin_lock_irqsave(&fq->mq_flush_lock, flags);
 
+       WRITE_ONCE(flush_rq->state, MQ_RQ_IDLE);
        if (!refcount_dec_and_test(&flush_rq->ref)) {
                fq->rq_status = error;
                spin_unlock_irqrestore(&fq->mq_flush_lock, flags);