block: mark blk_put_queue as potentially blocking

We can't just say that the last reference release may block, as any
reference dropped could be the last one.  So move the might_sleep() from
blk_free_queue to blk_put_queue and update the documentation.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/20221114042637.1009333-6-hch@lst.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>

Author: Christoph Hellwig

block/blk-core.c

@@ -260,8 +260,6 @@ static void blk_free_queue_rcu(struct rcu_head *rcu_head)
 
 static void blk_free_queue(struct request_queue *q)
 {
-	might_sleep();
-
 	percpu_ref_exit(&q->q_usage_counter);
 
 	if (q->poll_stat)
@@ -285,11 +283,11 @@ static void blk_free_queue(struct request_queue *q)
  * Decrements the refcount of the request_queue and free it when the refcount
  * reaches 0.
  *
- * Context: Any context, but the last reference must not be dropped from
- *          atomic context.
+ * Context: Can sleep.
  */
 void blk_put_queue(struct request_queue *q)
 {
+	might_sleep();
 	if (refcount_dec_and_test(&q->refs))
 		blk_free_queue(q);
 }