On Mon, Jun 06, 2022 at 05:10:38PM +0100, Alberto Faria wrote:
> Thanks for the feedback, and apologies for the delayed response.
>
> On Mon, May 30, 2022 at 1:49 PM Stefan Hajnoczi <stefa...@redhat.com> wrote:
> > If you find it's safe to change to -EINVAL then that's consistent with
> > how file I/O syscalls work and I think it would be nice.
>
> Switching to -EINVAL on negative bytes sounds good to me, but perhaps
> it should be done as a separate series. For now, switching just
> bdrv_{pread,pwrite}() to -EIO will make them consistent with all of
> bdrv_{preadv,pwritev}() and bdrv_co_{pread,pwrite,preadv,pwritev}(),
> accomplishing the purpose of this series with less changes and
> auditing.
>
> I can work on a subsequent series that changes -EIO to -EINVAL on
> negative bytes for all the bdrv_...() and blk_...() functions.
>
> Would this make sense?
Yes, that's fine. My main concern is that callers have been audited when
errnos are changed. If you switch bdrv_{pread,pwrite}() to -EIO and have
audited callers, then I'm happy.
Consistent -EINVAL would be nice in the future, but I think it's lower
priority and it doesn't have to be done any time soon.
Stefan
signature.asc
Description: PGP signature
