Re: NVMe PCI driver ignores SQHD from completion entries

[Vaibhav Nagarnaik <vnagarnaik@google.com>]

[-- Attachment #1.1: Type: text/plain, Size: 1509 bytes --]

On Sat, Oct 5, 2019 at 7:27 AM Keith Busch <kbusch@kernel.org> wrote:
>
> On Fri, Oct 04, 2019 at 11:27:30AM -0700, Vaibhav Nagarnaik wrote:
> > According to NVMe spec:
> > A Submission Queue entry has been consumed by the controller when a
> > Completion Queue entry is posted that indicates that the Submission
> > Queue Head Pointer has moved past the slot in which that Submission
> > Queue entry was placed.
> >
> > Which means, the driver needs to verify SQ Head Pointer value reported
> > in the completion entries before considering a particular SQ entry
> > reusable. Otherwise it's undefined behavior.
>
> The spec allows the controller to process and complete commands out of
> order, but the controller must fetch those commands in order. It's in the
> "Theory of Operation" section 1.4.

Thanks for clarifying. That makes sense.

It would be nice to have the spec clarify other sections that makes
the host check the SQ Head Pointer value.

> Checking SQ head is required only if the host might submit more commands
> than there are entries. The Linux nvme driver allocates enough tags
> for the depth of the queue, leaving one entry empty, so having a tag
> available means the next sq entry must be available.

But the driver does overwrite submission queue entries under process
(which are already fetched by the controller). Are there guarantees
from controllers out there that once fetched, the SQ entries will not
be fetched again for any reason? The spec doesn't prohibit that.



Vaibhav

[-- Attachment #1.2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4851 bytes --]

[-- Attachment #2: Type: text/plain, Size: 158 bytes --]

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