Re: [PATCH V2] nvme: Add support for ACRE Command Interrupted status

[Keith Busch <kbusch@kernel.org>]

On Wed, Dec 04, 2019 at 02:26:06PM +0100, Hannes Reinecke wrote:
> > Why is the controller returning the same error so many times? Are we
> > not waiting the requested delay timed? If so, the controller told us
> > retrying should be successful.
> > 
> > It is possible we kick the requeue list early if one command error
> > has a valid CRD, but a subsequent retryable command does not. Is that
> > what's happening?
> > 
> No. The problem is that CRD needs to be observed _in all cases_ whenever
> we get a COMMAND INTERRUPTED status.
> Currently we're only observing CRD when multipathing is _not_ active,
> which is wrong.
> 
> And this is what the patch is trying to achieve.

Right, I totally got that part.

I just don't understand this follow-up from John that I cropped it out
of my original reply:

> The alternative is to stop incrementing nvme_req(req)->retries in nvme_retry_req() when CRD is set.
>
>  @@ -258,8 +258,8 @@ static void nvme_retry_req(struct request *req)
>        crd = (nvme_req(req)->status & NVME_SC_CRD) >> 11;
>        if (ns && crd)
>                delay = ns->ctrl->crdt[crd - 1] * 100;
>-
>-       nvme_req(req)->retries++;
>+       else
>+               nvme_req(req)->retries++;

_______________________________________________
linux-nvme mailing list
linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme