[PATCH AUTOSEL 4.19 34/45] nvme-multipath: fix possible I/O hang when paths are updated

[Sasha Levin]

From: Anton Eidelman <anton@xxxxxxxxxxxxxxxxx>

[ Upstream commit 504db087aaccdb32af61539916409f7dca31ceb5 ]

nvme_state_set_live() making a path available triggers requeue_work
in order to resubmit requests that ended up on requeue_list when no
paths were available.

This requeue_work may race with concurrent nvme_ns_head_make_request()
that do not observe the live path yet.
Such concurrent requests may by made by either:
- New IO submission.
- Requeue_work triggered by nvme_failover_req() or another ana_work.

A race may cause requeue_work capture the state of requeue_list before
more requests get onto the list. These requests will stay on the list
forever unless requeue_work is triggered again.

In order to prevent such race, nvme_state_set_live() should
synchronize_srcu(&head->srcu) before triggering the requeue_work and
prevent nvme_ns_head_make_request referencing an old snapshot of the
path list.

Reviewed-by: Christoph Hellwig <hch@xxxxxx>
Signed-off-by: Anton Eidelman <anton@xxxxxxxxxxxxxxxxx>
Signed-off-by: Sagi Grimberg <sagi@xxxxxxxxxxx>
Signed-off-by: Jens Axboe <axboe@xxxxxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
---
 drivers/nvme/host/multipath.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/nvme/host/multipath.c b/drivers/nvme/host/multipath.c
index a11e210d173e4..9e24ef3bd0f29 100644
--- a/drivers/nvme/host/multipath.c
+++ b/drivers/nvme/host/multipath.c
@@ -293,6 +293,7 @@ static void nvme_mpath_set_live(struct nvme_ns *ns)
 				 "failed to create id group.\n");
 	}
 
+	synchronize_srcu(&ns->head->srcu);
 	kblockd_schedule_work(&ns->head->requeue_work);
 }
 
-- 
2.20.1