From: Matthew Brost <matthew.brost@intel.com>
To: intel-xe@lists.freedesktop.org
Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>,
Carlos Santa <carlos.santa@intel.com>
Subject: [PATCH] drm/xe: Do not preempt fence signaling CS instructions
Date: Wed, 14 Jan 2026 16:45:46 -0800 [thread overview]
Message-ID: <20260115004546.58060-1-matthew.brost@intel.com> (raw)
If a batch buffer is complete, it makes little sense to preempt the
fence signaling instructions in the ring, as the largest portion of the
work (the batch buffer) is already done and fence signaling consists of
only a few instructions. If these instructions are preempted, the GuC
would need to perform a context switch just to signal the fence, which
is costly and delays fence signaling. Avoid this scenario by disabling
preemption immediately after the BB start instruction and re-enabling it
after executing the fence signaling instructions.
Fixes: dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs")
Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: Carlos Santa <carlos.santa@intel.com>
Signed-off-by: Matthew Brost <matthew.brost@intel.com>
---
drivers/gpu/drm/xe/xe_ring_ops.c | 9 +++++++++
1 file changed, 9 insertions(+)
diff --git a/drivers/gpu/drm/xe/xe_ring_ops.c b/drivers/gpu/drm/xe/xe_ring_ops.c
index a1fd99f2d539..cd645ee400b9 100644
--- a/drivers/gpu/drm/xe/xe_ring_ops.c
+++ b/drivers/gpu/drm/xe/xe_ring_ops.c
@@ -282,6 +282,9 @@ static void __emit_job_gen12_simple(struct xe_sched_job *job, struct xe_lrc *lrc
i = emit_bb_start(batch_addr, ppgtt_flag, dw, i);
+ /* Don't preempt fence signaling */
+ dw[i++] = MI_ARB_ON_OFF | MI_ARB_DISABLE;
+
if (job->user_fence.used) {
i = emit_flush_dw(dw, i);
i = emit_store_imm_ppgtt_posted(job->user_fence.addr,
@@ -347,6 +350,9 @@ static void __emit_job_gen12_video(struct xe_sched_job *job, struct xe_lrc *lrc,
i = emit_bb_start(batch_addr, ppgtt_flag, dw, i);
+ /* Don't preempt fence signaling */
+ dw[i++] = MI_ARB_ON_OFF | MI_ARB_DISABLE;
+
if (job->user_fence.used) {
i = emit_flush_dw(dw, i);
i = emit_store_imm_ppgtt_posted(job->user_fence.addr,
@@ -399,6 +405,9 @@ static void __emit_job_gen12_render_compute(struct xe_sched_job *job,
i = emit_bb_start(batch_addr, ppgtt_flag, dw, i);
+ /* Don't preempt fence signaling */
+ dw[i++] = MI_ARB_ON_OFF | MI_ARB_DISABLE;
+
i = emit_render_cache_flush(job, dw, i);
if (job->user_fence.used)
--
2.34.1
next reply other threads:[~2026-01-15 0:45 UTC|newest]
Thread overview: 25+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-01-15 0:45 Matthew Brost [this message]
2026-01-15 0:52 ` ✓ CI.KUnit: success for drm/xe: Do not preempt fence signaling CS instructions Patchwork
2026-01-15 1:35 ` ✓ Xe.CI.BAT: " Patchwork
2026-01-15 6:13 ` ✗ Xe.CI.Full: failure " Patchwork
2026-01-16 9:45 ` [PATCH] " Zbigniew Kempczyński
2026-01-16 10:12 ` Francois Dugast
2026-01-16 16:43 ` Daniele Ceraolo Spurio
2026-01-16 19:51 ` Summers, Stuart
2026-01-16 20:44 ` Matthew Brost
2026-01-16 21:07 ` Summers, Stuart
2026-01-16 21:19 ` Matthew Brost
2026-01-16 21:05 ` Matthew Brost
2026-01-19 12:01 ` Zbigniew Kempczyński
2026-01-20 21:10 ` Daniele Ceraolo Spurio
2026-01-20 21:26 ` Matthew Brost
2026-01-20 21:27 ` Matthew Brost
2026-01-22 9:22 ` Zbigniew Kempczyński
2026-01-22 17:47 ` Daniele Ceraolo Spurio
2026-01-27 20:14 ` Matthew Brost
2026-01-20 21:11 ` Matthew Brost
2026-01-22 8:44 ` Zbigniew Kempczyński
2026-01-27 7:20 ` Zbigniew Kempczyński
2026-01-27 20:15 ` Matthew Brost
2026-02-26 17:35 ` Matthew Brost
2026-02-26 17:49 ` Daniele Ceraolo Spurio
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=20260115004546.58060-1-matthew.brost@intel.com \
--to=matthew.brost@intel.com \
--cc=carlos.santa@intel.com \
--cc=daniele.ceraolospurio@intel.com \
--cc=intel-xe@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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox