* [PATCH v2] nvme-fc: clean up error messages
@ 2019-06-24 22:12 Ewan D. Milne
2019-06-25 16:37 ` James Smart
0 siblings, 1 reply; 5+ messages in thread
From: Ewan D. Milne @ 2019-06-24 22:12 UTC (permalink / raw)
Some of the error messages are inconsistent, and one of them is wrong
(i.e. "queue_size 128 > ctrl maxcmd 32, reducing to queue_size").
Make them more clear and distinguishable for log analysis.
Signed-off-by: Ewan D. Milne <emilne at redhat.com>
---
drivers/nvme/host/fc.c | 16 +++++++++-------
1 file changed, 9 insertions(+), 7 deletions(-)
diff --git a/drivers/nvme/host/fc.c b/drivers/nvme/host/fc.c
index 9b497d785ed7..ab162c57ddc2 100644
--- a/drivers/nvme/host/fc.c
+++ b/drivers/nvme/host/fc.c
@@ -1259,7 +1259,7 @@ nvme_fc_connect_admin_queue(struct nvme_fc_ctrl *ctrl,
if (fcret) {
ret = -EBADF;
dev_err(ctrl->dev,
- "q %d connect failed: %s\n",
+ "queue %d connect admin queue failed: %s\n",
queue->qnum, validation_errors[fcret]);
} else {
ctrl->association_id =
@@ -1358,7 +1358,7 @@ nvme_fc_connect_queue(struct nvme_fc_ctrl *ctrl, struct nvme_fc_queue *queue,
if (fcret) {
ret = -EBADF;
dev_err(ctrl->dev,
- "q %d connect failed: %s\n",
+ "queue %d connect queue failed: %s\n",
queue->qnum, validation_errors[fcret]);
} else {
queue->connection_id =
@@ -1371,7 +1371,7 @@ nvme_fc_connect_queue(struct nvme_fc_ctrl *ctrl, struct nvme_fc_queue *queue,
out_no_memory:
if (ret)
dev_err(ctrl->dev,
- "queue %d connect command failed (%d).\n",
+ "queue %d connect queue failed (%d).\n",
queue->qnum, ret);
return ret;
}
@@ -2678,16 +2678,18 @@ nvme_fc_create_association(struct nvme_fc_ctrl *ctrl)
/* warn if maxcmd is lower than queue_size */
dev_warn(ctrl->ctrl.device,
"queue_size %zu > ctrl maxcmd %u, reducing "
- "to queue_size\n",
- opts->queue_size, ctrl->ctrl.maxcmd);
+ "queue_size to %u\n",
+ opts->queue_size, ctrl->ctrl.maxcmd, ctrl->ctrl.maxcmd);
opts->queue_size = ctrl->ctrl.maxcmd;
}
if (opts->queue_size > ctrl->ctrl.sqsize + 1) {
/* warn if sqsize is lower than queue_size */
dev_warn(ctrl->ctrl.device,
- "queue_size %zu > ctrl sqsize %u, clamping down\n",
- opts->queue_size, ctrl->ctrl.sqsize + 1);
+ "queue_size %zu > ctrl sqsize %u, reducing "
+ "queue_size to %u\n",
+ opts->queue_size, ctrl->ctrl.sqsize + 1,
+ ctrl->ctrl.sqsize + 1);
opts->queue_size = ctrl->ctrl.sqsize + 1;
}
--
2.18.1
^ permalink raw reply related [flat|nested] 5+ messages in thread
* [PATCH v2] nvme-fc: clean up error messages
2019-06-24 22:12 [PATCH v2] nvme-fc: clean up error messages Ewan D. Milne
@ 2019-06-25 16:37 ` James Smart
2019-06-25 19:05 ` Ewan D. Milne
0 siblings, 1 reply; 5+ messages in thread
From: James Smart @ 2019-06-25 16:37 UTC (permalink / raw)
On 6/24/2019 3:12 PM, Ewan D. Milne wrote:
> Some of the error messages are inconsistent, and one of them is wrong
> (i.e. "queue_size 128 > ctrl maxcmd 32, reducing to queue_size").
> Make them more clear and distinguishable for log analysis.
>
> Signed-off-by: Ewan D. Milne <emilne at redhat.com>
> ---
> drivers/nvme/host/fc.c | 16 +++++++++-------
> 1 file changed, 9 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/nvme/host/fc.c b/drivers/nvme/host/fc.c
> index 9b497d785ed7..ab162c57ddc2 100644
> --- a/drivers/nvme/host/fc.c
> +++ b/drivers/nvme/host/fc.c
> @@ -1259,7 +1259,7 @@ nvme_fc_connect_admin_queue(struct nvme_fc_ctrl *ctrl,
> if (fcret) {
> ret = -EBADF;
> dev_err(ctrl->dev,
> - "q %d connect failed: %s\n",
> + "queue %d connect admin queue failed: %s\n",
> queue->qnum, validation_errors[fcret]);
> } else {
> ctrl->association_id =
Since we're changing them...
how about "queue %d Create Association LS failed: %s"
> @@ -1358,7 +1358,7 @@ nvme_fc_connect_queue(struct nvme_fc_ctrl *ctrl, struct nvme_fc_queue *queue,
> if (fcret) {
> ret = -EBADF;
> dev_err(ctrl->dev,
> - "q %d connect failed: %s\n",
> + "queue %d connect queue failed: %s\n",
how about "queue %d Create Connection LS failed: %s"
> queue->qnum, validation_errors[fcret]);
> } else {
> queue->connection_id =
> @@ -1371,7 +1371,7 @@ nvme_fc_connect_queue(struct nvme_fc_ctrl *ctrl, struct nvme_fc_queue *queue,
> out_no_memory:
> if (ret)
> dev_err(ctrl->dev,
> - "queue %d connect command failed (%d).\n",
> + "queue %d connect queue failed (%d).\n",
> queue->qnum, ret);
> return ret;
> }
> @@ -2678,16 +2678,18 @@ nvme_fc_create_association(struct nvme_fc_ctrl *ctrl)
> /* warn if maxcmd is lower than queue_size */
> dev_warn(ctrl->ctrl.device,
> "queue_size %zu > ctrl maxcmd %u, reducing "
> - "to queue_size\n",
> - opts->queue_size, ctrl->ctrl.maxcmd);
> + "queue_size to %u\n",
> + opts->queue_size, ctrl->ctrl.maxcmd, ctrl->ctrl.maxcmd);
> opts->queue_size = ctrl->ctrl.maxcmd;
> }
>
> if (opts->queue_size > ctrl->ctrl.sqsize + 1) {
> /* warn if sqsize is lower than queue_size */
> dev_warn(ctrl->ctrl.device,
> - "queue_size %zu > ctrl sqsize %u, clamping down\n",
> - opts->queue_size, ctrl->ctrl.sqsize + 1);
> + "queue_size %zu > ctrl sqsize %u, reducing "
> + "queue_size to %u\n",
> + opts->queue_size, ctrl->ctrl.sqsize + 1,
> + ctrl->ctrl.sqsize + 1);
> opts->queue_size = ctrl->ctrl.sqsize + 1;
> }
>
Given these last two came from a template in the rdma transport - you
may want to consider using the same wording in all the transports.
-- james
^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH v2] nvme-fc: clean up error messages
2019-06-25 16:37 ` James Smart
@ 2019-06-25 19:05 ` Ewan D. Milne
2019-06-25 19:55 ` James Smart
0 siblings, 1 reply; 5+ messages in thread
From: Ewan D. Milne @ 2019-06-25 19:05 UTC (permalink / raw)
On Tue, 2019-06-25@09:37 -0700, James Smart wrote:
>
> > @@ -1259,7 +1259,7 @@ nvme_fc_connect_admin_queue(struct nvme_fc_ctrl *ctrl,
> > if (fcret) {
> > ret = -EBADF;
> > dev_err(ctrl->dev,
> > - "q %d connect failed: %s\n",
> > + "queue %d connect admin queue failed: %s\n",
> > queue->qnum, validation_errors[fcret]);
> > } else {
> > ctrl->association_id =
>
> Since we're changing them...
>
> how about "queue %d Create Association LS failed: %s"
>
> > @@ -1358,7 +1358,7 @@ nvme_fc_connect_queue(struct nvme_fc_ctrl *ctrl, struct nvme_fc_queue *queue,
> > if (fcret) {
> > ret = -EBADF;
> > dev_err(ctrl->dev,
> > - "q %d connect failed: %s\n",
> > + "queue %d connect queue failed: %s\n",
>
> how about "queue %d Create Connection LS failed: %s"
Sure.
>
> > @@ -2678,16 +2678,18 @@ nvme_fc_create_association(struct nvme_fc_ctrl *ctrl)
> > /* warn if maxcmd is lower than queue_size */
> > dev_warn(ctrl->ctrl.device,
> > "queue_size %zu > ctrl maxcmd %u, reducing "
> > - "to queue_size\n",
> > - opts->queue_size, ctrl->ctrl.maxcmd);
> > + "queue_size to %u\n",
> > + opts->queue_size, ctrl->ctrl.maxcmd, ctrl->ctrl.maxcmd);
> > opts->queue_size = ctrl->ctrl.maxcmd;
> > }
> >
> > if (opts->queue_size > ctrl->ctrl.sqsize + 1) {
> > /* warn if sqsize is lower than queue_size */
> > dev_warn(ctrl->ctrl.device,
> > - "queue_size %zu > ctrl sqsize %u, clamping down\n",
> > - opts->queue_size, ctrl->ctrl.sqsize + 1);
> > + "queue_size %zu > ctrl sqsize %u, reducing "
> > + "queue_size to %u\n",
> > + opts->queue_size, ctrl->ctrl.sqsize + 1,
> > + ctrl->ctrl.sqsize + 1);
> > opts->queue_size = ctrl->ctrl.sqsize + 1;
> > }
> >
>
> Given these last two came from a template in the rdma transport - you
> may want to consider using the same wording in all the transports.
Hmm. The RDMA and TCP transports do not appear to actually change
opts->queue_size if > maxcmd, here, despite the warning message, but
they use sqsize when allocating the tagset.
And they both reduce the sqsize if > maxcmd, not if queue_size > sqsize
as in FC, (which does not change sqsize).
commit 5e77d61cbc7e766778037127dab69e6410a8fc48
Author: Sagi Grimberg <sagi at grimberg.me>
Date: Tue Jun 19 15:34:13 2018 +0300
nvme-rdma: don't override opts->queue_size
So should FC be using the same logic?
-Ewan
>
> -- james
>
>
^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH v2] nvme-fc: clean up error messages
2019-06-25 19:05 ` Ewan D. Milne
@ 2019-06-25 19:55 ` James Smart
2019-06-25 21:18 ` Sagi Grimberg
0 siblings, 1 reply; 5+ messages in thread
From: James Smart @ 2019-06-25 19:55 UTC (permalink / raw)
On 6/25/2019 12:05 PM, Ewan D. Milne wrote:
> commit 5e77d61cbc7e766778037127dab69e6410a8fc48
> Author: Sagi Grimberg<sagi at grimberg.me>
> Date: Tue Jun 19 15:34:13 2018 +0300
>
> nvme-rdma: don't override opts->queue_size
>
> So should FC be using the same logic?
not sure - I'll have to look why Sagi reverted it.
-- james
^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH v2] nvme-fc: clean up error messages
2019-06-25 19:55 ` James Smart
@ 2019-06-25 21:18 ` Sagi Grimberg
0 siblings, 0 replies; 5+ messages in thread
From: Sagi Grimberg @ 2019-06-25 21:18 UTC (permalink / raw)
>> commit 5e77d61cbc7e766778037127dab69e6410a8fc48
>> Author: Sagi Grimberg<sagi at grimberg.me>
>> Date:?? Tue Jun 19 15:34:13 2018 +0300
>>
>> ???? nvme-rdma: don't override opts->queue_size
>>
>> So should FC be using the same logic?
>
> not sure - I'll have to look why Sagi reverted it.
I reverted it because its a user parameter which we should
try to avoid modifying.
And maxcmd or other limits can theoretically change over resets.
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2019-06-25 21:18 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-06-24 22:12 [PATCH v2] nvme-fc: clean up error messages Ewan D. Milne
2019-06-25 16:37 ` James Smart
2019-06-25 19:05 ` Ewan D. Milne
2019-06-25 19:55 ` James Smart
2019-06-25 21:18 ` Sagi Grimberg
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.