Error recovery can be interrupted by controller removal, then the
controller is left as quiesced, and IO hang can be caused.
Fix the issue by unquiescing controller unconditionally when removing
namespaces.
This way is reasonable and safe given forward progress can be made
when removing namespaces.
Reviewed-by: Keith Busch <kbusch@kernel.org>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reported-by: Chunguang Xu <brookxu.cn@gmail.com>
Closes: https://lore.kernel.org/linux-nvme/cover.1685350577.git.chunguang.xu@shopee.com/
Cc: stable@vger.kernel.org
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
*/
nvme_mpath_clear_ctrl_paths(ctrl);
*/
nvme_mpath_clear_ctrl_paths(ctrl);
+ /*
+ * Unquiesce io queues so any pending IO won't hang, especially
+ * those submitted from scan work
+ */
+ nvme_unquiesce_io_queues(ctrl);
+
/* prevent racing with ns scanning */
flush_work(&ctrl->scan_work);
/* prevent racing with ns scanning */
flush_work(&ctrl->scan_work);
* removing the namespaces' disks; fail all the queues now to avoid
* potentially having to clean up the failed sync later.
*/
* removing the namespaces' disks; fail all the queues now to avoid
* potentially having to clean up the failed sync later.
*/
- if (ctrl->state == NVME_CTRL_DEAD) {
+ if (ctrl->state == NVME_CTRL_DEAD)
nvme_mark_namespaces_dead(ctrl);
nvme_mark_namespaces_dead(ctrl);
- nvme_unquiesce_io_queues(ctrl);
- }
/* this is a no-op when called from the controller reset handler */
nvme_change_ctrl_state(ctrl, NVME_CTRL_DELETING_NOIO);
/* this is a no-op when called from the controller reset handler */
nvme_change_ctrl_state(ctrl, NVME_CTRL_DELETING_NOIO);