[PATCH 2/3] nvme-core: delete the dependency on blk status

[Christoph Hellwig]

On Wed, Aug 12, 2020 at 04:28:30PM -0400, Mike Snitzer wrote:
> On Wed, Aug 12 2020 at 11:10am -0400,
> Christoph Hellwig <hch at lst.de> wrote:
> 
> > On Wed, Aug 12, 2020 at 04:18:44PM +0800, Chao Leng wrote:
> > > nvme should not depend on blk status, just need check nvme status.
> > > Just need do translating nvme status to blk status for returning error.
> > 
> > While this doesn't look wrong it also doesn't save us a single
> > instruction and actually adds more lines of code.  Do you have a good
> > reason for this change?
> 
> It certainly saves nvme_error_status(nvme_req(req)->status if
> nvme_req_needs_retry().

Yeah, but the retry path isn't exactly the fast path, is it?

Not that I really care too much about this one..