From 53770f0ec5fd417429775ba006bc4abe14002335 Mon Sep 17 00:00:00 2001
From: Mikulas Patocka <mpatocka@redhat.com>
Date: Mon, 17 Feb 2020 07:43:03 -0500
Subject: [PATCH] dm integrity: fix a deadlock due to offloading to an
 incorrect workqueue
Git-commit: 53770f0ec5fd417429775ba006bc4abe14002335
Patch-mainline: v5.6-rc5
References: git fixes (block drivers)

If we need to perform synchronous I/O in dm_integrity_map_continue(),
we must make sure that we are not in the map function - in order to
avoid the deadlock due to bio queuing in generic_make_request. To
avoid the deadlock, we offload the request to metadata_wq.

However, metadata_wq also processes metadata updates for write requests.
If there are too many requests that get offloaded to metadata_wq at the
beginning of dm_integrity_map_continue, the workqueue metadata_wq
becomes clogged and the system is incapable of processing any metadata
updates.

This causes a deadlock because all the requests that need to do metadata
updates wait for metadata_wq to proceed and metadata_wq waits inside
wait_and_add_new_range until some existing request releases its range
lock (which doesn't happen because the range lock is released after
metadata update).

In order to fix the deadlock, we create a new workqueue offload_wq and
offload requests to it - so that processing of offload_wq is independent
from processing of metadata_wq.

(Coly Li: rebase for Linux v4.12 based kernel, remove the changes to
 bitmap_block_work() because it is not backported yet)

Fixes: 7eada909bfd7 ("dm: add integrity target")
Cc: stable@vger.kernel.org # v4.12+
Reported-by: Heinz Mauelshagen <heinzm@redhat.com>
Tested-by: Heinz Mauelshagen <heinzm@redhat.com>
Signed-off-by: Heinz Mauelshagen <heinzm@redhat.com>
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Coly Li <colyli@suse.de>

---
 drivers/md/dm-integrity.c |   16 ++++++++++++++--
 1 file changed, 14 insertions(+), 2 deletions(-)

--- a/drivers/md/dm-integrity.c
+++ b/drivers/md/dm-integrity.c
@@ -197,6 +197,7 @@ struct dm_integrity_c {
 	struct list_head wait_list;
 	wait_queue_head_t endio_wait;
 	struct workqueue_struct *wait_wq;
+	struct workqueue_struct *offload_wq;
 
 	unsigned char commit_seq;
 	commit_id_t commit_ids[N_COMMIT_IDS];
@@ -1236,7 +1237,8 @@ static void dec_in_flight(struct dm_inte
 			dio->range.logical_sector += dio->range.n_sectors;
 			bio_advance(bio, dio->range.n_sectors << SECTOR_SHIFT);
 			INIT_WORK(&dio->work, integrity_bio_wait);
-			queue_work(ic->wait_wq, &dio->work);
+			queue_work(ic->offload_wq, &dio->work);
+			queue_work(ic->offload_wq, &dio->work);
 			return;
 		}
 		do_endio_flush(ic, dio);
@@ -1656,7 +1658,7 @@ static void dm_integrity_map_continue(st
 
 	if (need_sync_io && from_map) {
 		INIT_WORK(&dio->work, integrity_bio_wait);
-		queue_work(ic->metadata_wq, &dio->work);
+		queue_work(ic->offload_wq, &dio->work);
 		return;
 	}
 
@@ -3303,6 +3305,14 @@ static int dm_integrity_ctr(struct dm_ta
 		goto bad;
 	}
 
+	ic->offload_wq = alloc_workqueue("dm-integrity-offload", WQ_MEM_RECLAIM,
+					  METADATA_WORKQUEUE_MAX_ACTIVE);
+	if (!ic->offload_wq) {
+		ti->error = "Cannot allocate workqueue";
+		r = -ENOMEM;
+		goto bad;
+	}
+
 	ic->commit_wq = alloc_workqueue("dm-integrity-commit", WQ_MEM_RECLAIM, 1);
 	if (!ic->commit_wq) {
 		ti->error = "Cannot allocate workqueue";
@@ -3538,6 +3548,8 @@ static void dm_integrity_dtr(struct dm_t
 		destroy_workqueue(ic->metadata_wq);
 	if (ic->wait_wq)
 		destroy_workqueue(ic->wait_wq);
+	if (ic->offload_wq)
+		destroy_workqueue(ic->offload_wq);
 	if (ic->commit_wq)
 		destroy_workqueue(ic->commit_wq);
 	if (ic->writer_wq)