All of lore.kernel.org
 help / color / mirror / Atom feed
From: Chris Wilson <chris@chris-wilson.co.uk>
To: intel-gfx@lists.freedesktop.org
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Subject: [Intel-gfx] [PATCH 01/56] drm/i915/gt: Restore ce->signal flush before releasing virtual engine
Date: Tue, 29 Dec 2020 12:00:50 +0000	[thread overview]
Message-ID: <20201229120145.26045-1-chris@chris-wilson.co.uk> (raw)

Before we mark the virtual engine as no longer inflight, flush any
ongoing signaling that may be using the ce->signal_link along the
previous breadcrumbs. On switch to a new physical engine, that link will
be inserted into the new set of breadcrumbs, causing confusion to an
ongoing iterator.

This patch undoes a last minute mistake introduced into commit
bab0557c8dca ("drm/i915/gt: Remove virtual breadcrumb before transfer"),
whereby instead of unconditionally applying the flush, it was only
applied if the request itself was going to be reused.

Fixes: bab0557c8dca ("drm/i915/gt: Remove virtual breadcrumb before transfer")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
 drivers/gpu/drm/i915/gt/intel_execlists_submission.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
index 1fae6c6f3868..b79365b5159a 100644
--- a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
+++ b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
@@ -578,8 +578,6 @@ resubmit_virtual_request(struct i915_request *rq, struct virtual_engine *ve)
 		 * ce->signal_link.
 		 */
 		i915_request_cancel_breadcrumb(rq);
-		while (atomic_read(&engine->breadcrumbs->signaler_active))
-			cpu_relax();
 	}
 
 	spin_lock_irq(&engine->active.lock);
@@ -595,6 +593,7 @@ static void kick_siblings(struct i915_request *rq, struct intel_context *ce)
 {
 	struct virtual_engine *ve = container_of(ce, typeof(*ve), context);
 	struct intel_engine_cs *engine = rq->engine;
+	bool signals = !list_empty(&ce->signals);
 
 	/*
 	 * This engine is now too busy to run this virtual request, so
@@ -608,6 +607,10 @@ static void kick_siblings(struct i915_request *rq, struct intel_context *ce)
 
 	if (READ_ONCE(ve->request))
 		tasklet_hi_schedule(&ve->base.execlists.tasklet);
+
+	/* Flush concurrent signal_irq_work before we reuse the link */
+	while (signals && atomic_read(&engine->breadcrumbs->signaler_active))
+		cpu_relax();
 }
 
 static inline void __execlists_schedule_out(struct i915_request *rq)
-- 
2.20.1

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

             reply	other threads:[~2020-12-29 12:03 UTC|newest]

Thread overview: 65+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-12-29 12:00 Chris Wilson [this message]
2020-12-29 12:00 ` [Intel-gfx] [PATCH 02/56] drm/i915/gt: Only retire on the last breadcrumb if the last request Chris Wilson
2020-12-29 12:00 ` [Intel-gfx] [PATCH 03/56] drm/i915/gt: Cancel submitted requests upon context reset Chris Wilson
2020-12-30 21:07   ` Mika Kuoppala
2020-12-29 12:00 ` [Intel-gfx] [PATCH 04/56] drm/i915/gt: Pull context closure check from request submit to schedule-in Chris Wilson
2020-12-29 12:00 ` [Intel-gfx] [PATCH 05/56] drm/i915/gem: Peek at the inflight context Chris Wilson
2020-12-29 12:00 ` [Intel-gfx] [PATCH 06/56] drm/i915: Mark up protected uses of 'i915_request_completed' Chris Wilson
2020-12-29 12:00 ` [Intel-gfx] [PATCH 07/56] drm/i915: Drop i915_request.lock serialisation around await_start Chris Wilson
2020-12-29 12:00 ` [Intel-gfx] [PATCH 08/56] drm/i915: Drop i915_request.lock requirement for intel_rps_boost() Chris Wilson
2020-12-29 12:00 ` [Intel-gfx] [PATCH 09/56] drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source Chris Wilson
2020-12-29 12:00 ` [Intel-gfx] [PATCH 10/56] drm/i915/gem: Reduce ctx->engines_mutex for get_engines() Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 11/56] drm/i915: Reduce test_and_set_bit to set_bit in i915_request_submit() Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 12/56] drm/i915/gt: Drop atomic for engine->fw_active tracking Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 13/56] drm/i915/gt: Extract busy-stats for ring-scheduler Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 14/56] drm/i915/gt: Convert stats.active to plain unsigned int Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 15/56] drm/i915/gt: Do not suspend bonded requests if one hangs Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 16/56] drm/i915/gt: Remove timeslice suppression Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 17/56] drm/i915/gt: Skip over completed active execlists, again Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 18/56] drm/i915: Strip out internal priorities Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 19/56] drm/i915: Remove I915_USER_PRIORITY_SHIFT Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 20/56] drm/i915: Replace engine->schedule() with a known request operation Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 21/56] drm/i915: Teach the i915_dependency to use a double-lock Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 22/56] drm/i915: Restructure priority inheritance Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 23/56] drm/i915/selftests: Measure set-priority duration Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 24/56] drm/i915/selftests: Exercise priority inheritance around an engine loop Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 25/56] drm/i915: Improve DFS for priority inheritance Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 26/56] drm/i915: Extract request submission from execlists Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 27/56] drm/i915: Extract request rewinding " Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 28/56] drm/i915: Extract request suspension from the execlists backend Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 29/56] drm/i915: Extract the ability to defer and rerun a request later Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 30/56] drm/i915: Fix the iterative dfs for defering requests Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 31/56] drm/i915: Move common active lists from engine to i915_scheduler Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 32/56] drm/i915: Move scheduler queue Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 33/56] drm/i915: Move tasklet from execlists to sched Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 34/56] drm/i915: Replace priolist rbtree with a skiplist Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 35/56] drm/i915: Wrap cmpxchg64 with try_cmpxchg64() helper Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 36/56] drm/i915: Fair low-latency scheduling Chris Wilson
2021-01-07 16:05   ` Matthew Brost
2021-01-07 16:45     ` Chris Wilson
2021-01-07 17:10       ` Matthew Brost
2020-12-29 12:01 ` [Intel-gfx] [PATCH 37/56] drm/i915/gt: Specify a deadline for the heartbeat Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 38/56] drm/i915: Extend the priority boosting for the display with a deadline Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 39/56] drm/i915/gt: Support virtual engine queues Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 40/56] drm/i915: Move saturated workload detection back to the context Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 41/56] drm/i915: Bump default timeslicing quantum to 5ms Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 42/56] drm/i915/gt: Wrap intel_timeline.has_initial_breadcrumb Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 43/56] drm/i915/gt: Track timeline GGTT offset separately from subpage offset Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 44/56] drm/i915/gt: Add timeline "mode" Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 45/56] drm/i915/gt: Use indices for writing into relative timelines Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 46/56] drm/i915/selftests: Exercise relative timeline modes Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 47/56] drm/i915/gt: Use ppHWSP for unshared non-semaphore related timelines Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 48/56] Restore "drm/i915: drop engine_pin/unpin_breadcrumbs_irq" Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 49/56] drm/i915/gt: Couple tasklet scheduling for all CS interrupts Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 50/56] drm/i915/gt: Support creation of 'internal' rings Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 51/56] drm/i915/gt: Use client timeline address for seqno writes Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 52/56] drm/i915/gt: Infrastructure for ring scheduling Chris Wilson
2020-12-29 18:34   ` kernel test robot
2020-12-29 18:34     ` kernel test robot
2020-12-29 12:01 ` [Intel-gfx] [PATCH 53/56] drm/i915/gt: Enable busy-stats for ring-scheduler Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 54/56] drm/i915/gt: Implement ring scheduler for gen6/7 Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 55/56] drm/i915/gt: Enable ring scheduling " Chris Wilson
2020-12-29 12:01 ` [Intel-gfx] [PATCH 56/56] drm/i915/gt: Limit C-states while waiting for requests Chris Wilson
2020-12-29 12:20 ` [Intel-gfx] ✗ Fi.CI.CHECKPATCH: warning for series starting with [01/56] drm/i915/gt: Restore ce->signal flush before releasing virtual engine Patchwork
2020-12-29 12:22 ` [Intel-gfx] ✗ Fi.CI.SPARSE: " Patchwork
2020-12-29 12:49 ` [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=20201229120145.26045-1-chris@chris-wilson.co.uk \
    --to=chris@chris-wilson.co.uk \
    --cc=intel-gfx@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.