2020-05-29 11:41:59

by Daniel Wagner

Subject: [PATCH] nvme-fc: Only call nvme_cleanup_cmd() for normal operations

Asynchronous event notifications do not have an request
associated. When fcp_io() fails we unconditionally call
nvme_cleanup_cmd() which leads to a crash.

Fixes: 16686f3a6c3c ("nvme: move common call to nvme_cleanup_cmd to core layer")
Cc: Max Gurtovoy <[email protected]>
Signed-off-by: Daniel Wagner <[email protected]>
---
drivers/nvme/host/fc.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/nvme/host/fc.c b/drivers/nvme/host/fc.c
index 7dfc4a2ecf1e..287a3e8ea317 100644
--- a/drivers/nvme/host/fc.c
+++ b/drivers/nvme/host/fc.c
@@ -2300,10 +2300,11 @@ nvme_fc_start_fcp_op(struct nvme_fc_ctrl *ctrl, struct nvme_fc_queue *queue,
opstate = atomic_xchg(&op->state, FCPOP_STATE_COMPLETE);
__nvme_fc_fcpop_chk_teardowns(ctrl, op, opstate);

- if (!(op->flags & FCOP_FLAGS_AEN))
+ if (!(op->flags & FCOP_FLAGS_AEN)) {
nvme_fc_unmap_data(ctrl, op->rq, op);
+ nvme_cleanup_cmd(op->rq);
+ }

- nvme_cleanup_cmd(op->rq);
nvme_fc_ctrl_put(ctrl);

if (ctrl->rport->remoteport.port_state == FC_OBJSTATE_ONLINE &&
--
2.26.2

2020-06-02 12:52:57

by Hannes Reinecke

Subject: Re: [PATCH] nvme-fc: Only call nvme_cleanup_cmd() for normal operations

On 5/29/20 1:37 PM, Daniel Wagner wrote:
> Asynchronous event notifications do not have an request
> associated. When fcp_io() fails we unconditionally call
> nvme_cleanup_cmd() which leads to a crash.
>
> Fixes: 16686f3a6c3c ("nvme: move common call to nvme_cleanup_cmd to core layer")
> Cc: Max Gurtovoy <[email protected]>
> Signed-off-by: Daniel Wagner <[email protected]>
> ---
> drivers/nvme/host/fc.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/nvme/host/fc.c b/drivers/nvme/host/fc.c
> index 7dfc4a2ecf1e..287a3e8ea317 100644
> --- a/drivers/nvme/host/fc.c
> +++ b/drivers/nvme/host/fc.c
> @@ -2300,10 +2300,11 @@ nvme_fc_start_fcp_op(struct nvme_fc_ctrl *ctrl, struct nvme_fc_queue *queue,
> opstate = atomic_xchg(&op->state, FCPOP_STATE_COMPLETE);
> __nvme_fc_fcpop_chk_teardowns(ctrl, op, opstate);
>
> - if (!(op->flags & FCOP_FLAGS_AEN))
> + if (!(op->flags & FCOP_FLAGS_AEN)) {
> nvme_fc_unmap_data(ctrl, op->rq, op);
> + nvme_cleanup_cmd(op->rq);
> + }
>
> - nvme_cleanup_cmd(op->rq);
> nvme_fc_ctrl_put(ctrl);
>
> if (ctrl->rport->remoteport.port_state == FC_OBJSTATE_ONLINE &&
>

Reviewed-by: Hannes Reinecke <[email protected]>

Cheers,

Hannes
--
Dr. Hannes Reinecke Teamlead Storage & Networking
[email protected] +49 911 74053 688
SUSE Software Solutions GmbH, Maxfeldstr. 5, 90409 Nürnberg
HRB 36809 (AG Nürnberg), Geschäftsführer: Felix Imendörffer

2020-06-02 16:32:48

by James Smart

Subject: Re: [PATCH] nvme-fc: Only call nvme_cleanup_cmd() for normal operations

On 5/29/2020 4:37 AM, Daniel Wagner wrote:
> Asynchronous event notifications do not have an request
> associated. When fcp_io() fails we unconditionally call
> nvme_cleanup_cmd() which leads to a crash.
>
> Fixes: 16686f3a6c3c ("nvme: move common call to nvme_cleanup_cmd to core layer")
> Cc: Max Gurtovoy <[email protected]>
> Signed-off-by: Daniel Wagner <[email protected]>
> ---
> drivers/nvme/host/fc.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/nvme/host/fc.c b/drivers/nvme/host/fc.c
> index 7dfc4a2ecf1e..287a3e8ea317 100644
> --- a/drivers/nvme/host/fc.c
> +++ b/drivers/nvme/host/fc.c
> @@ -2300,10 +2300,11 @@ nvme_fc_start_fcp_op(struct nvme_fc_ctrl *ctrl, struct nvme_fc_queue *queue,
> opstate = atomic_xchg(&op->state, FCPOP_STATE_COMPLETE);
> __nvme_fc_fcpop_chk_teardowns(ctrl, op, opstate);
>
> - if (!(op->flags & FCOP_FLAGS_AEN))
> + if (!(op->flags & FCOP_FLAGS_AEN)) {
> nvme_fc_unmap_data(ctrl, op->rq, op);
> + nvme_cleanup_cmd(op->rq);
> + }
>
> - nvme_cleanup_cmd(op->rq);
> nvme_fc_ctrl_put(ctrl);
>
> if (ctrl->rport->remoteport.port_state == FC_OBJSTATE_ONLINE &&

Reviewed-by:  James Smart  <[email protected]>

-- james

2020-06-03 18:59:57

by Himanshu Madhani

Subject: Re: [PATCH] nvme-fc: Only call nvme_cleanup_cmd() for normal operations

On 5/29/20 6:37 AM, Daniel Wagner wrote:
> Asynchronous event notifications do not have an request
> associated. When fcp_io() fails we unconditionally call
> nvme_cleanup_cmd() which leads to a crash.
>
> Fixes: 16686f3a6c3c ("nvme: move common call to nvme_cleanup_cmd to core layer")
> Cc: Max Gurtovoy <[email protected]>
> Signed-off-by: Daniel Wagner <[email protected]>
> ---
> drivers/nvme/host/fc.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/nvme/host/fc.c b/drivers/nvme/host/fc.c
> index 7dfc4a2ecf1e..287a3e8ea317 100644
> --- a/drivers/nvme/host/fc.c
> +++ b/drivers/nvme/host/fc.c
> @@ -2300,10 +2300,11 @@ nvme_fc_start_fcp_op(struct nvme_fc_ctrl *ctrl, struct nvme_fc_queue *queue,
> opstate = atomic_xchg(&op->state, FCPOP_STATE_COMPLETE);
> __nvme_fc_fcpop_chk_teardowns(ctrl, op, opstate);
>
> - if (!(op->flags & FCOP_FLAGS_AEN))
> + if (!(op->flags & FCOP_FLAGS_AEN)) {
> nvme_fc_unmap_data(ctrl, op->rq, op);
> + nvme_cleanup_cmd(op->rq);
> + }
>
> - nvme_cleanup_cmd(op->rq);
> nvme_fc_ctrl_put(ctrl);
>
> if (ctrl->rport->remoteport.port_state == FC_OBJSTATE_ONLINE &&

Reviewed-by: Himanshu Madhani <[email protected]>

2020-06-09 14:18:00

by Christoph Hellwig

Subject: Re: [PATCH] nvme-fc: Only call nvme_cleanup_cmd() for normal operations

Thanks, applied to nvme-5.8 with some edits to the commit message.