[PATCH 3.10 317/319] dm: mark request_queue dead before destroying the DM device

From: Willy Tarreau
Date: Sun Feb 05 2017 - 14:46:11 EST


From: Bart Van Assche <bart.vanassche@xxxxxxxxxxx>

commit 3b785fbcf81c3533772c52b717f77293099498d3 upstream.

This avoids that new requests are queued while __dm_destroy() is in
progress.

[js] use md->queue instead of non-present helper

Signed-off-by: Bart Van Assche <bart.vanassche@xxxxxxxxxxx>
Signed-off-by: Mike Snitzer <snitzer@xxxxxxxxxx>
Signed-off-by: Jiri Slaby <jslaby@xxxxxxx>
Signed-off-by: Willy Tarreau <w@xxxxxx>
---
drivers/md/dm.c | 5 +++++
1 file changed, 5 insertions(+)

diff --git a/drivers/md/dm.c b/drivers/md/dm.c
index f69fed8..a77ef6c 100644
--- a/drivers/md/dm.c
+++ b/drivers/md/dm.c
@@ -2323,6 +2323,7 @@ EXPORT_SYMBOL_GPL(dm_device_name);

static void __dm_destroy(struct mapped_device *md, bool wait)
{
+ struct request_queue *q = md->queue;
struct dm_table *map;

might_sleep();
@@ -2333,6 +2334,10 @@ static void __dm_destroy(struct mapped_device *md, bool wait)
set_bit(DMF_FREEING, &md->flags);
spin_unlock(&_minor_lock);

+ spin_lock_irq(q->queue_lock);
+ queue_flag_set(QUEUE_FLAG_DYING, q);
+ spin_unlock_irq(q->queue_lock);
+
/*
* Take suspend_lock so that presuspend and postsuspend methods
* do not race with internal suspend.
--
2.8.0.rc2.1.gbe9624a