[PATCH v4 2/2] drm/xe: Upgrade job timeout message to warn

[Matthew Brost]

On Fri, Oct 25, 2024 at 04:24:41PM -0700, John Harrison wrote:
> On 10/25/2024 14:43, Matthew Brost wrote:
> > Ensure job timeout message shows up on distro configs for user bug
> > reports. Change job timeout message from notice to warn.
> I don't think you are allowed to do this. Job timeouts occur when the user
> submits bad batch buffers. They are expected error handling for untrusted
> user programs. As such, they should not generate warnings or errors. Same as
> passing dodgy data into an IOCTL should not generate a warning but just
> return EINVAL or similar.
> 

Yea ok, will drop this. I think the warning in the first patch is still
ok though as that indicates some sort of scheduling problem.

Matt

> John.
> 
> > 
> > Fixes: dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs")
> > Cc: stable at vger.kernel.org
> > Cc: Paulo Zanoni <paulo.r.zanoni at intel.com>
> > Signed-off-by: Matthew Brost <matthew.brost at intel.com>
> > ---
> >   drivers/gpu/drm/xe/xe_guc_submit.c | 6 +++---
> >   1 file changed, 3 insertions(+), 3 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/xe/xe_guc_submit.c b/drivers/gpu/drm/xe/xe_guc_submit.c
> > index 7afcc243037c..067cf2bfac1e 100644
> > --- a/drivers/gpu/drm/xe/xe_guc_submit.c
> > +++ b/drivers/gpu/drm/xe/xe_guc_submit.c
> > @@ -1154,9 +1154,9 @@ guc_exec_queue_timedout_job(struct drm_sched_job *drm_job)
> >   		process_name = q->vm->xef->process_name;
> >   		pid = q->vm->xef->pid;
> >   	}
> > -	xe_gt_notice(guc_to_gt(guc), "Timedout job: seqno=%u, lrc_seqno=%u, guc_id=%d, flags=0x%lx in %s [%d]",
> > -		     xe_sched_job_seqno(job), xe_sched_job_lrc_seqno(job),
> > -		     q->guc->id, q->flags, process_name, pid);
> > +	xe_gt_warn(guc_to_gt(guc), "Timedout job: seqno=%u, lrc_seqno=%u, guc_id=%d, flags=0x%lx in %s [%d]",
> > +		   xe_sched_job_seqno(job), xe_sched_job_lrc_seqno(job),
> > +		   q->guc->id, q->flags, process_name, pid);
> >   	trace_xe_sched_job_timedout(job);
>