nvme: prevent warning triggered by nvme_stop_keep_alive
PATCH v2 1/1 nvme: prevent warning triggered by nvme_stop_keep_alive
Delayed keep alive work is queued on system workqueue and may be cancelled via nvme_stop_keep_alive from nvme_reset_wq, nvme_fc_wq or nvme_wq.
Check_flush_dependency detects mismatched attributes between the work-queue context used to cancel the keep alive work and system-wq. Specifically system-wq does not have the WQ_MEM_RECLAIM flag, whereas the contexts used to cancel keep alive work have WQ_MEM_RECLAIM flag.
Example warning:
workqueue: WQ_MEM_RECLAIM nvme-reset-wq:nvme_fc_reset_ctrl_work [nvme_fc]
is flushing !WQ_MEM_RECLAIM events:nvme_keep_alive_work [nvme_core]
To avoid the flags mismatch, delayed keep alive work is queued on nvme_wq.
However this creates a secondary concern where work and a request to cancel that work may be in the same work queue - namely err_work in the rdma and tcp transports, which will want to flush/cancel the keep alive work which will now be on nvme_wq.
After reviewing the transports, it looks like err_work can be moved to nvme_reset_wq. In fact that aligns them better with transition into RESETTING and performing related reset work in nvme_reset_wq.
Change nvme-rdma and nvme-tcp to perform err_work in nvme_reset_wq.
---
V2: combined the 2 patches into 1 to prevent regressions if one missed.
--- drivers/nvme/host/core.c | 10 +++++----- drivers/nvme/host/rdma.c | 2 +- drivers/nvme/host/tcp.c | 2 +- 3 files changed, 7 insertions(+), 7 deletions(-)
drivers/nvme/host/core.c +5 -5
@@ -66,8 +66,8 @@ MODULE_PARM_DESC(streams, "turn on support for Streams write directives");* nvme_reset_wq - hosts nvme reset works * nvme_delete_wq - hosts nvme delete works * -* nvme_wq will host works such are scan, aen handling, fw activation, -* keep-alive error recovery, periodic reconnects etc. nvme_reset_wq +* nvme_wq will host works such as scan, aen handling, fw activation, +* keep-alive, periodic reconnects etc. nvme_reset_wq * runs reset works which also flush works hosted on nvme_wq for * serialization purposes. nvme_delete_wq host controller deletion * works which flush reset works for serialization.
@@ -974,7 +974,7 @@ static void nvme_keep_alive_end_io(struct request *rq, blk_status_t status) startka = true;spin_unlock_irqrestore (&ctrl->lock, flags);if (startka) -schedule_delayed_work (&ctrl->ka_work, ctrl->kato *HZ ); +queue_delayed_work (nvme_wq, &ctrl->ka_work, ctrl->kato *HZ ); }static int nvme_keep_alive (struct nvme_ctrl *ctrl)
@@ -1004,7 +1004,7 @@ static void nvme_keep_alive_work(struct work_struct *work)dev_dbg (ctrl->device,"reschedule traffic based keep-alive timer\n" ); ctrl->comp_seen = false; -schedule_delayed_work (&ctrl->ka_work, ctrl->kato *HZ ); +queue_delayed_work (nvme_wq, &ctrl->ka_work, ctrl->kato *HZ );return ; }
@@ -1021,7 +1021,7 @@ static void nvme_start_keep_alive(struct nvme_ctrl *ctrl)if (unlikely (ctrl->kato ==0 ))return ; -schedule_delayed_work (&ctrl->ka_work, ctrl->kato *HZ ); +queue_delayed_work (nvme_wq, &ctrl->ka_work, ctrl->kato *HZ ); }void nvme_stop_keep_alive (struct nvme_ctrl *ctrl)
drivers/nvme/host/rdma.c +1 -1
@@ -1088,7 +1088,7 @@ static void nvme_rdma_error_recovery(struct nvme_rdma_ctrl *ctrl)if (!nvme_change_ctrl_state (&ctrl->ctrl,NVME_CTRL_RESETTING ))return ; -queue_work (nvme_wq, &ctrl->err_work); +queue_work (nvme_reset_wq, &ctrl->err_work); } static void nvme_rdma_wr_error(struct ib_cq *cq, struct ib_wc *wc,
drivers/nvme/host/tcp.c +1 -1
@@ -422,7 +422,7 @@ static void nvme_tcp_error_recovery(struct nvme_ctrl *ctrl)if (!nvme_change_ctrl_state (ctrl,NVME_CTRL_RESETTING ))return ; -queue_work (nvme_wq, &to_tcp_ctrl (ctrl)->err_work); +queue_work (nvme_reset_wq, &to_tcp_ctrl (ctrl)->err_work); } static int nvme_tcp_process_nvme_cqe(struct nvme_tcp_queue *queue,
Signature · 7 lines
2.13.7 _______________________________________________ linux-nvme mailing list [email protected] http://lists.infradead.org/mailman/listinfo/linux-nvme
Keith Busch
maintainerQuoted 24 lines · James Smart
Delayed keep alive work is queued on system workqueue and may be cancelled via
nvme_stop_keep_alivefromnvme_reset_wq,nvme_fc_wqornvme_wq.
Check_flush_dependencydetects mismatched attributes between the work-queue context used to cancel the keep alive work and system-wq. Specifically system-wq does not have theWQ_MEM_RECLAIMflag, whereas the contexts used to cancel keep alive work haveWQ_MEM_RECLAIMflag.Example warning:
workqueue:
WQ_MEM_RECLAIMnvme-reset-wq:nvme_fc_reset_ctrl_work [nvme_fc]
is flushing !WQ_MEM_RECLAIM events:nvme_keep_alive_work [nvme_core]To avoid the flags mismatch, delayed keep alive work is queued on
nvme_wq.However this creates a secondary concern where work and a request to cancel that work may be in the same work queue - namely
err_workin the rdma and tcp transports, which will want to flush/cancel the keep alive work which will now be onnvme_wq.After reviewing the transports, it looks like
err_workcan be moved tonvme_reset_wq. In fact that aligns them better with transition into RESETTING and performing related reset work innvme_reset_wq.
Quoted 6 lines
Change nvme-rdma and nvme-tcp to perform
err_workinnvme_reset_wq.Signed-off-by Nigel Kirkland <[email protected]>Signed-off-by James Smart <[email protected]>Reviewed-by Sagi Grimberg <[email protected]>Reviewed-by Christoph Hellwig <[email protected]>
Thanks James, applied to nvme-5.6-rc.
_______________________________________________
linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme