All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Dixit, Ashutosh" <ashutosh.dixit@intel.com>
To: Daniel Charles <daniel.charles@intel.com>
Cc: <igt-dev@lists.freedesktop.org>
Subject: Re: [PATCH] tests/intel/xe_exec_reset: expect error or complete on CAT_ERROR
Date: Mon, 22 Sep 2025 08:35:54 -0700	[thread overview]
Message-ID: <87348esdnp.wl-ashutosh.dixit@intel.com> (raw)
In-Reply-To: <de1dbb13-b9ba-42e1-a9c4-a9ff57006dfc@intel.com>

On Fri, 19 Sep 2025 17:07:28 -0700, Daniel Charles wrote:
>
> On 9/19/2025 4:59 PM, Dixit, Ashutosh wrote:
> > On Fri, 19 Sep 2025 15:28:36 -0700, Daniel Charles wrote:
> >> when running cm job, it could be that after the fence is checked, a
> >> CAT_ERROR will throw an error in the same way as GT_RESET does.
> >>
> >> Signed-off-by: Daniel Charles <daniel.charles@intel.com>
> >> ---
> >>   tests/intel/xe_exec_reset.c | 2 +-
> >>   1 file changed, 1 insertion(+), 1 deletion(-)
> >>
> >> diff --git a/tests/intel/xe_exec_reset.c b/tests/intel/xe_exec_reset.c
> >> index 7ae53c679..73c5c7e20 100644
> >> --- a/tests/intel/xe_exec_reset.c
> >> +++ b/tests/intel/xe_exec_reset.c
> >> @@ -433,7 +433,7 @@ test_compute_mode(int fd, struct drm_xe_engine_class_instance *eci,
> >>
> >>		err = __xe_wait_ufence(fd, &data[i].exec_sync, USER_FENCE_VALUE,
> >>				       exec_queues[i % n_exec_queues], &timeout);
> >> -		if (flags & GT_RESET)
> >> +		if (flags & GT_RESET || flags & CAT_ERROR)
> > Hi Daniel,
> >
> > 1. Is there a gitlab bug about this? If yes, could you please add a
> >     'Closes:' tag to this patch (see git log)
> not that I'm aware of.

There should have been a previous failure which this patch is fixing,
correct? So we would expect there to be a previous bug. Anyway.

> >
> > 2. Would you have a reference to the kernel code that a CAT error returns
> >     -EIO?
>
> when there's an engine reset a job that was sent to execution and didn't
> finish before the cat error, the driver will return -EIO as it does with a
> full reset.

OK, from xe_guc_exec_queue_memory_cat_error_handler(), what happens is the
exec_queue is reset on CAT error, which in turn would return -EIO.

So this is:

Reviewed-by: Ashutosh Dixit <ashutosh.dixit@intel.com>


>
> > Thanks.
> > --
> > Ashutosh
> >
> >>			/* exec races with reset: may return -EIO or complete */
> >>			igt_assert(err == -EIO || !err);
> >>		else
> >> --
> >> 2.43.0
> >>

  reply	other threads:[~2025-09-22 15:35 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-09-19 22:28 [PATCH] tests/intel/xe_exec_reset: expect error or complete on CAT_ERROR Daniel Charles
2025-09-19 23:34 ` ✓ Xe.CI.BAT: success for " Patchwork
2025-09-19 23:39 ` ✓ i915.CI.BAT: " Patchwork
2025-09-19 23:59 ` [PATCH] " Dixit, Ashutosh
2025-09-20  0:07   ` Daniel Charles
2025-09-22 15:35     ` Dixit, Ashutosh [this message]
2025-09-20  6:51 ` ✗ Xe.CI.Full: failure for " Patchwork
2025-09-20 19:00 ` ✗ i915.CI.Full: " Patchwork

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=87348esdnp.wl-ashutosh.dixit@intel.com \
    --to=ashutosh.dixit@intel.com \
    --cc=daniel.charles@intel.com \
    --cc=igt-dev@lists.freedesktop.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.