blk-mq, percpu_ref: implement a kludge for SCSI blk-mq stall during p…

…robe

blk-mq uses percpu_ref for its usage counter which tracks the number
of in-flight commands and used to synchronously drain the queue on
freeze.  percpu_ref shutdown takes measureable wallclock time as it
involves a sched RCU grace period.  This means that draining a blk-mq
takes measureable wallclock time.  One would think that this shouldn't
matter as queue shutdown should be a rare event which takes place
asynchronously w.r.t. userland.

Unfortunately, SCSI probing involves synchronously setting up and then
tearing down a lot of request_queues back-to-back for non-existent
LUNs.  This means that SCSI probing may take more than ten seconds
when scsi-mq is used.

This will be properly fixed by implementing a mechanism to keep
q->mq_usage_counter in atomic mode till genhd registration; however,
that involves rather big updates to percpu_ref which is difficult to
apply late in the devel cycle (v3.17-rc6 at the moment).  As a
stop-gap measure till the proper fix can be implemented in the next
cycle, this patch introduces __percpu_ref_kill_expedited() and makes
blk_mq_freeze_queue() use it.  This is heavy-handed but should work
for testing the experimental SCSI blk-mq implementation.

Signed-off-by: Tejun Heo <[email protected]>
Reported-by: Christoph Hellwig <[email protected]>
Link: http://lkml.kernel.org/g/[email protected]
Fixes: add703f ("blk-mq: use percpu_ref for mq usage count")
Cc: Kent Overstreet <[email protected]>
Cc: Jens Axboe <[email protected]>
Tested-by: Christoph Hellwig <[email protected]>
Signed-off-by: Jens Axboe <[email protected]>

block/blk-mq.c

@@ -119,7 +119,16 @@ void blk_mq_freeze_queue(struct request_queue *q)
 	spin_unlock_irq(q->queue_lock);
 
 	if (freeze) {
-		percpu_ref_kill(&q->mq_usage_counter);
+		/*
+		 * XXX: Temporary kludge to work around SCSI blk-mq stall.
+		 * SCSI synchronously creates and destroys many queues
+		 * back-to-back during probe leading to lengthy stalls.
+		 * This will be fixed by keeping ->mq_usage_counter in
+		 * atomic mode until genhd registration, but, for now,
+		 * let's work around using expedited synchronization.
+		 */
+		__percpu_ref_kill_expedited(&q->mq_usage_counter);
+
 		blk_mq_run_queues(q, false);
 	}
 	wait_event(q->mq_freeze_wq, percpu_ref_is_zero(&q->mq_usage_counter));

include/linux/percpu-refcount.h

@@ -71,6 +71,7 @@ void percpu_ref_reinit(struct percpu_ref *ref);
 void percpu_ref_exit(struct percpu_ref *ref);
 void percpu_ref_kill_and_confirm(struct percpu_ref *ref,
 				 percpu_ref_func_t *confirm_kill);
+void __percpu_ref_kill_expedited(struct percpu_ref *ref);
 
 /**
  * percpu_ref_kill - drop the initial ref

lib/percpu-refcount.c

@@ -184,3 +184,19 @@ void percpu_ref_kill_and_confirm(struct percpu_ref *ref,
 	call_rcu_sched(&ref->rcu, percpu_ref_kill_rcu);
 }
 EXPORT_SYMBOL_GPL(percpu_ref_kill_and_confirm);
+
+/*
+ * XXX: Temporary kludge to work around SCSI blk-mq stall.  Used only by
+ * block/blk-mq.c::blk_mq_freeze_queue().  Will be removed during v3.18
+ * devel cycle.  Do not use anywhere else.
+ */
+void __percpu_ref_kill_expedited(struct percpu_ref *ref)
+{
+	WARN_ONCE(ref->pcpu_count_ptr & PCPU_REF_DEAD,
+		  "percpu_ref_kill() called more than once on %pf!",
+		  ref->release);
+
+	ref->pcpu_count_ptr |= PCPU_REF_DEAD;
+	synchronize_sched_expedited();
+	percpu_ref_kill_rcu(&ref->rcu);
+}