On Tue, 2022-08-09 at 08:52 +0200, Christoph Hellwig wrote:
> >
> > An alternative to changing all the callers would be we could make
> > scsi_noretry_cmd
> > detect when it's an internal passthrough command and just retry
> > these types of
> > errors. For SG IO type of passthough we still want to fail right
> > away.
>
> Yes, I think one single place to do retries for setup path command
> is much better than growing ad-hoc logic.
>
> I just made a similar comment to similar nvme patch from SuSE a few
> days
> ago..
Are you suggesting to re-invent REQ_OP_SCSI_{IN,OUT} to distinguish
SG_IO from kernel-internal passthrough?
Martin
