From: Chris Wilson <chris@chris-wilson.co.uk>
To: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>,
intel-gfx@lists.freedesktop.org
Subject: Re: [Intel-gfx] [PATCH v2] drm/i915/gt: Use virtual_engine during execlists_dequeue
Date: Mon, 18 May 2020 14:09:04 +0100 [thread overview]
Message-ID: <158980734430.17769.7168882813683816716@build.alporthouse.com> (raw)
In-Reply-To: <2bcbbf4a-42ad-ac61-89f0-a1fb25fb2c04@linux.intel.com>
Quoting Tvrtko Ursulin (2020-05-18 14:01:27)
>
> On 18/05/2020 13:33, Chris Wilson wrote:
> > +static struct virtual_engine *
> > +first_virtual_engine(struct intel_engine_cs *engine)
> > +{
> > + struct intel_engine_execlists *el = &engine->execlists;
> > + struct rb_node *rb = rb_first_cached(&el->virtual);
> > +
> > + while (rb) {
> > + struct virtual_engine *ve =
> > + rb_entry(rb, typeof(*ve), nodes[engine->id].rb);
> > + struct i915_request *rq = READ_ONCE(ve->request);
> > +
> > + if (!rq) { /* lazily cleanup after another engine handled rq */
> > + rb_erase_cached(rb, &el->virtual);
> > + RB_CLEAR_NODE(rb);
> > + rb = rb_first_cached(&el->virtual);
> > + continue;
> > + }
> > +
> > + if (!virtual_matches(ve, rq, engine)) {
> > + rb = rb_next(rb);
> > + continue;
> > + }
> > +
> > + return ve;
> > + }
> > +
> > + return NULL;
> > +}
> > - while (rb) { /* XXX virtual is always taking precedence */
> > - struct virtual_engine *ve =
> > - rb_entry(rb, typeof(*ve), nodes[engine->id].rb);
> > + while (ve) { /* XXX virtual is always taking precedence */
> > struct i915_request *rq;
> >
> > spin_lock(&ve->base.active.lock);
> >
> > rq = ve->request;
> > - if (unlikely(!rq)) { /* lost the race to a sibling */
> > - spin_unlock(&ve->base.active.lock);
> > - rb_erase_cached(rb, &execlists->virtual);
> > - RB_CLEAR_NODE(rb);
> > - rb = rb_first_cached(&execlists->virtual);
> > - continue;
> > - }
> > + if (unlikely(!rq)) /* lost the race to a sibling */
> > + goto unlock;
>
> Doesn't this now rely on a later patch to clear the node?
This is cleared by first_virtual_engine
> >
> > GEM_BUG_ON(rq != ve->request);
> > GEM_BUG_ON(rq->engine != &ve->base);
> > GEM_BUG_ON(rq->context != &ve->context);
> >
> > - if (rq_prio(rq) >= queue_prio(execlists)) {
> > - if (!virtual_matches(ve, rq, engine)) {
> > - spin_unlock(&ve->base.active.lock);
> > - rb = rb_next(rb);
> > - continue;
> > - }
> > + if (rq_prio(rq) < queue_prio(execlists)) {
> > + spin_unlock(&ve->base.active.lock);
> > + break;
> > + }
> >
> > - if (last && !can_merge_rq(last, rq)) {
> > - spin_unlock(&ve->base.active.lock);
> > - start_timeslice(engine, rq_prio(rq));
> > - return; /* leave this for another sibling */
> > - }
> > + GEM_BUG_ON(!virtual_matches(ve, rq, engine));
>
> This as well.
As first_virtual_engine skips non-matching nodes, it should only
unmatch during the unlocked phase since the lookup if it is claimed by
another engine. Then ve->request would be set to NULL and we the above
check fails.
So I think this patch stands by itself.
-Chris
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
next prev parent reply other threads:[~2020-05-18 13:09 UTC|newest]
Thread overview: 32+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-05-18 8:14 [Intel-gfx] [PATCH 1/8] drm/i915: Move saturated workload detection back to the context Chris Wilson
2020-05-18 8:14 ` [Intel-gfx] [PATCH 2/8] drm/i915/selftests: Add tests for timeslicing virtual engines Chris Wilson
2020-05-18 10:12 ` Tvrtko Ursulin
2020-05-18 10:21 ` Chris Wilson
2020-05-18 8:14 ` [Intel-gfx] [PATCH 3/8] drm/i915/gt: Reuse the tasklet priority for virtual as their siblings Chris Wilson
2020-05-18 10:13 ` Tvrtko Ursulin
2020-05-18 8:14 ` [Intel-gfx] [PATCH 4/8] drm/i915/gt: Kick virtual siblings on timeslice out Chris Wilson
2020-05-18 10:29 ` Tvrtko Ursulin
2020-05-18 8:14 ` [Intel-gfx] [PATCH 5/8] drm/i915/gt: Incorporate the virtual engine into timeslicing Chris Wilson
2020-05-18 10:36 ` Tvrtko Ursulin
2020-05-18 10:38 ` Chris Wilson
2020-05-18 8:14 ` [Intel-gfx] [PATCH 6/8] drm/i915/gt: Use virtual_engine during execlists_dequeue Chris Wilson
2020-05-18 10:51 ` Tvrtko Ursulin
2020-05-18 10:57 ` Chris Wilson
2020-05-18 12:33 ` [Intel-gfx] [PATCH v2] " Chris Wilson
2020-05-18 13:01 ` Tvrtko Ursulin
2020-05-18 13:09 ` Chris Wilson [this message]
2020-05-18 8:14 ` [Intel-gfx] [PATCH 7/8] drm/i915/gt: Decouple inflight virtual engines Chris Wilson
2020-05-18 12:53 ` Tvrtko Ursulin
2020-05-18 13:00 ` Chris Wilson
2020-05-18 14:55 ` Tvrtko Ursulin
2020-05-18 15:40 ` Chris Wilson
2020-05-18 15:48 ` Chris Wilson
2020-05-18 8:14 ` [Intel-gfx] [PATCH 8/8] drm/i915/gt: Resubmit the virtual engine on schedule-out Chris Wilson
2020-05-18 9:53 ` [Intel-gfx] [PATCH 1/8] drm/i915: Move saturated workload detection back to the context Tvrtko Ursulin
2020-05-18 10:11 ` Chris Wilson
2020-05-18 11:55 ` [Intel-gfx] ✗ Fi.CI.CHECKPATCH: warning for series starting with [1/8] " Patchwork
2020-05-18 11:56 ` [Intel-gfx] ✗ Fi.CI.SPARSE: " Patchwork
2020-05-18 12:17 ` [Intel-gfx] ✓ Fi.CI.BAT: success " Patchwork
2020-05-18 15:55 ` [Intel-gfx] ✗ Fi.CI.CHECKPATCH: warning for series starting with [1/8] drm/i915: Move saturated workload detection back to the context (rev2) Patchwork
2020-05-18 15:57 ` [Intel-gfx] ✗ Fi.CI.SPARSE: " Patchwork
2020-05-18 16:28 ` [Intel-gfx] ✗ Fi.CI.BAT: failure " 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=158980734430.17769.7168882813683816716@build.alporthouse.com \
--to=chris@chris-wilson.co.uk \
--cc=intel-gfx@lists.freedesktop.org \
--cc=tvrtko.ursulin@linux.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;
as well as URLs for NNTP newsgroup(s).