namespace's request queue is frozen and quiesced during error recovering, writeback IO is blocked in bio_queue_enter(), so fsync_bdev() <- del_gendisk() can't move on, and causes IO hang. Removal could be from sysfs, hard unplug or error handling. Fix this kind of issue by marking controller as DEAD if removal breaks error recovery. This ways is reasonable too, because controller can't be recovered any more after being removed. Cc: stable@xxxxxxxxxxxxxxx Reported-by: Chunguang Xu <brookxu.cn@xxxxxxxxx> Closes: https://lore.kernel.org/linux-nvme/cover.1685350577.git.chunguang.xu@xxxxxxxxxx/ Reported-by: Yi Zhang <yi.zhang@xxxxxxxxxx> Signed-off-by: Ming Lei <ming.lei@xxxxxxxxxx> --- drivers/nvme/host/core.c | 4 +++- drivers/nvme/host/nvme.h | 1 + 2 files changed, 4 insertions(+), 1 deletion(-) diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c index fdfcf2781c85..b4cebc01cc00 100644 --- a/drivers/nvme/host/core.c +++ b/drivers/nvme/host/core.c @@ -567,6 +567,7 @@ bool nvme_change_ctrl_state(struct nvme_ctrl *ctrl, } if (changed) { + ctrl->old_state = ctrl->state; ctrl->state = new_state; wake_up_all(&ctrl->state_wq); } @@ -4055,7 +4056,8 @@ void nvme_remove_namespaces(struct nvme_ctrl *ctrl) * 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 || + ctrl->old_state != NVME_CTRL_LIVE) { nvme_mark_namespaces_dead(ctrl); nvme_unquiesce_io_queues(ctrl); } diff --git a/drivers/nvme/host/nvme.h b/drivers/nvme/host/nvme.h index 9a98c14c552a..ce67856d4d4f 100644 --- a/drivers/nvme/host/nvme.h +++ b/drivers/nvme/host/nvme.h @@ -254,6 +254,7 @@ struct nvme_ctrl { bool comp_seen; bool identified; enum nvme_ctrl_state state; + enum nvme_ctrl_state old_state; spinlock_t lock; struct mutex scan_lock; const struct nvme_ctrl_ops *ops; -- 2.40.1
