From: Matthew Brost <matthew.brost@intel.com>
To: John Harrison <john.c.harrison@intel.com>
Cc: <intel-xe@lists.freedesktop.org>
Subject: Re: [PATCH v5 06/10] drm/xe: Improve unexpected state error messages
Date: Tue, 11 Jun 2024 01:43:18 +0000 [thread overview]
Message-ID: <ZmertnRU4jSTqhBq@DUT025-TGLU.fm.intel.com> (raw)
In-Reply-To: <328a4647-642d-4d39-8d3b-dc69cb7ee4ce@intel.com>
On Mon, Jun 10, 2024 at 05:09:01PM -0700, John Harrison wrote:
> On 6/10/2024 07:18, Matthew Brost wrote:
> > Include G2H handler name when an unexpected error state messages.
> >
> > Signed-off-by: Matthew Brost <matthew.brost@intel.com>
> > ---
> > drivers/gpu/drm/xe/xe_guc_submit.c | 8 ++++----
> > 1 file changed, 4 insertions(+), 4 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/xe/xe_guc_submit.c b/drivers/gpu/drm/xe/xe_guc_submit.c
> > index 4464ba337d12..766ff8e48dde 100644
> > --- a/drivers/gpu/drm/xe/xe_guc_submit.c
> > +++ b/drivers/gpu/drm/xe/xe_guc_submit.c
> > @@ -1632,8 +1632,8 @@ int xe_guc_sched_done_handler(struct xe_guc *guc, u32 *msg, u32 len)
> > if (unlikely(!exec_queue_pending_enable(q) &&
> > !exec_queue_pending_disable(q))) {
> > - drm_err(&xe->drm, "Unexpected engine state 0x%04x",
> > - atomic_read(&q->guc->state));
> > + drm_err(&xe->drm, "SCHED_DONE: Unexpected engine state 0x%04x, guc_id=%d",
> > + atomic_read(&q->guc->state), q->guc->id);
> > return -EPROTO;
> > }
> My earlier point was that this G2H notification provides two words of data -
> the context id and the runnable state. The code is currently not looking at
I was unaware, msg[1] contained runnable state. Will include in message
plus add a patch to assert this in the handle_sched_done.
> that second word. It is only looking at the internal KMD state and assuming
> that the notification will match. Instead, the code should read out the
> state from msg[1] and pass that in to handle_sched_done(). Which should
> complain if the notification is for an enable but only a disable is pending
> and vice versa. And by extension, should not do the disable processing if
> the notification was actually for an enable, and vice versa.
>
> And as per Michal's comment, any update to a drm_err|warn|info|dbg|etc.
> should take the opportunity to convert it to an xe_gt_ equivalent.
>
Already done.
Matt
> John.
>
> > @@ -1671,8 +1671,8 @@ int xe_guc_deregister_done_handler(struct xe_guc *guc, u32 *msg, u32 len)
> > if (!exec_queue_destroyed(q) || exec_queue_pending_disable(q) ||
> > exec_queue_pending_enable(q) || exec_queue_enabled(q)) {
> > - drm_err(&xe->drm, "Unexpected engine state 0x%04x",
> > - atomic_read(&q->guc->state));
> > + drm_err(&xe->drm, "DEREGISTER_DONE: Unexpected engine state 0x%04x, guc_id=%d",
> > + atomic_read(&q->guc->state), q->guc->id);
> > return -EPROTO;
> > }
>
next prev parent reply other threads:[~2024-06-11 1:43 UTC|newest]
Thread overview: 37+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-06-10 14:18 [PATCH v5 00/10] Only timeout jobs if they run longer than timeout period Matthew Brost
2024-06-10 14:18 ` [PATCH v5 01/10] drm/xe: Add LRC ctx timestamp support functions Matthew Brost
2024-06-10 17:32 ` Cavitt, Jonathan
2024-06-10 14:18 ` [PATCH v5 02/10] drm/xe: Add MI_COPY_MEM_MEM GPU instruction definitions Matthew Brost
2024-06-10 16:39 ` Cavitt, Jonathan
2024-06-10 14:18 ` [PATCH v5 03/10] drm/xe: Emit ctx timestamp copy in ring ops Matthew Brost
2024-06-10 14:44 ` Cavitt, Jonathan
2024-06-10 14:18 ` [PATCH v5 04/10] drm/xe: Add ctx timestamp to LRC snapshot Matthew Brost
2024-06-10 16:40 ` Cavitt, Jonathan
2024-06-10 14:18 ` [PATCH v5 05/10] drm/xe: Add xe_gt_clock_interval_to_ms helper Matthew Brost
2024-06-10 16:34 ` Cavitt, Jonathan
2024-06-12 12:09 ` Ghimiray, Himal Prasad
2024-06-10 14:18 ` [PATCH v5 06/10] drm/xe: Improve unexpected state error messages Matthew Brost
2024-06-10 16:36 ` Cavitt, Jonathan
2024-06-10 16:45 ` Michal Wajdeczko
2024-06-10 17:09 ` Matthew Brost
2024-06-11 0:09 ` John Harrison
2024-06-11 1:43 ` Matthew Brost [this message]
2024-06-10 14:18 ` [PATCH v5 07/10] drm/xe: Add GuC state asserts to deregister_exec_queue Matthew Brost
2024-06-10 17:12 ` Cavitt, Jonathan
2024-06-10 14:18 ` [PATCH v5 08/10] drm/xe: Add pending disable assert to handle_sched_done Matthew Brost
2024-06-10 16:35 ` Cavitt, Jonathan
2024-06-10 14:18 ` [PATCH v5 09/10] drm/xe: Add killed, banned, or wedged as stick bit during GuC reset Matthew Brost
2024-06-10 16:35 ` Cavitt, Jonathan
2024-06-10 14:18 ` [PATCH v5 10/10] drm/xe: Sample ctx timestamp to determine if jobs have timed out Matthew Brost
2024-06-10 19:32 ` Cavitt, Jonathan
2024-06-10 20:12 ` Matthew Brost
2024-06-11 0:36 ` John Harrison
2024-06-11 1:35 ` Matthew Brost
2024-06-12 4:21 ` Zbigniew Kempczyński
2024-06-10 14:23 ` ✓ CI.Patch_applied: success for Only timeout jobs if they run longer than timeout period Patchwork
2024-06-10 14:23 ` ✗ CI.checkpatch: warning " Patchwork
2024-06-10 14:24 ` ✓ CI.KUnit: success " Patchwork
2024-06-10 14:36 ` ✓ CI.Build: " Patchwork
2024-06-10 14:38 ` ✗ CI.Hooks: failure " Patchwork
2024-06-10 14:39 ` ✓ CI.checksparse: success " Patchwork
2024-06-10 15:16 ` ✓ CI.BAT: " 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=ZmertnRU4jSTqhBq@DUT025-TGLU.fm.intel.com \
--to=matthew.brost@intel.com \
--cc=intel-xe@lists.freedesktop.org \
--cc=john.c.harrison@intel.com \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox