[Devel] [PATCH hci-8.1 v12 08/12] dm-qcow2: fail the L2 table allocation request on L1 writeback error

[Alexander Ivanov]

A request that allocates an L2 table waits for the L1 writeback on the
page wait list. If the writeback fails, the L1 entry is reverted and the
request is dispatched again, so it allocates another L2 table and dirties
L1 once more. With a persistent write error the request never completes.

Queue the request on the wbd dependent list instead, like requests
waiting for an L2 update. It is dispatched again after a successful
writeback and ends with the error after a failed one.

https://virtuozzo.atlassian.net/browse/VSTOR-148171
Feature: dm-qcow2: block device over QCOW2 files driver
Fixes: 222b6f1291c1c ("dm-qcow2: Introduce driver to create block devices over QCOW2 files")

Signed-off-by: Alexander Ivanov <alexander.ivanov at virtuozzo.com>
---
 drivers/md/dm-qcow2-map.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/md/dm-qcow2-map.c b/drivers/md/dm-qcow2-map.c
index 995f611ae5bb..7f61ff67646c 100644
--- a/drivers/md/dm-qcow2-map.c
+++ b/drivers/md/dm-qcow2-map.c
@@ -2616,7 +2616,7 @@ static int prepare_l1l2_allocation(struct qcow2 *qcow2, struct qio *qio,
 		 * 2)For COW from backing file this is must.
 		 */
 		spin_lock_irq(&qcow2->md_pages_lock);
-		list_add_tail(&qio->link, &map->l1.md->wait_list);
+		list_add_tail(&qio->link, &map->l1.md->wbd->dependent_list);
 		spin_unlock_irq(&qcow2->md_pages_lock);
 		return 0;
 	}
-- 
2.53.0