Igt-dev Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Ray Wu <ray.wu@amd.com>
To: <igt-dev@lists.freedesktop.org>
Cc: <alex.hung@amd.com>, <sunpeng.li@amd.com>,
	<chiahsuan.chung@amd.com>, <ray.wu@amd.com>
Subject: [PATCH i-g-t 2/2] tests/kms_cursor_legacy: disable AMD Panel Replay during test
Date: Tue, 2 Jun 2026 14:41:22 +0800	[thread overview]
Message-ID: <20260602064543.58586-3-ray.wu@amd.com> (raw)
In-Reply-To: <20260602064543.58586-1-ray.wu@amd.com>

[Why]
On AMD, cursor updates can occasionally be delayed when Panel
Replay is active, causing intermittent timeouts in the cursor
legacy subtests.

[How]
Call igt_amd_disallow_replay_on_all_edp() in the setup fixture to
disable Panel Replay entry on any eDP that currently has it
enabled, and restore it in the teardown fixture.

Signed-off-by: Ray Wu <ray.wu@amd.com>
---
 tests/kms_cursor_legacy.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/tests/kms_cursor_legacy.c b/tests/kms_cursor_legacy.c
index 4d597af79..62339855e 100644
--- a/tests/kms_cursor_legacy.c
+++ b/tests/kms_cursor_legacy.c
@@ -35,6 +35,7 @@
 
 #include "i915/gem.h"
 #include "igt.h"
+#include "igt_amd.h"
 #include "igt_psr.h"
 #include "igt_rand.h"
 #include "igt_stats.h"
@@ -1875,6 +1876,7 @@ int igt_main()
 	const int ncpus = sysconf(_SC_NPROCESSORS_ONLN);
 	igt_display_t display = { .drm_fd = -1 };
 	bool intel_psr2_restore = false;
+	bool amd_replay_restore = false;
 	int i;
 	const char *modes[flip_test_last+1] = {
 		"legacy",
@@ -1898,6 +1900,14 @@ int igt_main()
 		 */
 		intel_psr2_restore = i915_psr2_sel_fetch_to_psr1(display.drm_fd, NULL);
 
+		/*
+		 * On AMD, when Panel Replay is enabled the HW cannot guarantee
+		 * that all required cursor updates land within a single vblank,
+		 * so disable Panel Replay (if any eDP output has it on) for the
+		 * duration of this test and restore it on teardown.
+		 */
+		amd_replay_restore = igt_amd_disallow_replay_on_all_edp(display.drm_fd, &display, true);
+
 		igt_install_exit_handler(igt_drm_debug_mask_reset_exit_handler);
 		update_debug_mask_if_ci(debug_mask_if_ci);
 	}
@@ -2128,6 +2138,8 @@ int igt_main()
 	}
 
 	igt_fixture() {
+		if (amd_replay_restore)
+			igt_amd_disallow_replay_on_all_edp(display.drm_fd, &display, false);
 		if (intel_psr2_restore)
 			i915_psr2_sel_fetch_restore(display.drm_fd, NULL);
 		igt_display_fini(&display);
-- 
2.43.0


  parent reply	other threads:[~2026-06-02  6:46 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-06-02  6:41 [PATCH i-g-t 0/2] kms_cursor_legacy: avoid AMD Panel Replay interference Ray Wu
2026-06-02  6:41 ` [PATCH i-g-t 1/2] lib/igt_amd: add helper to disable Panel Replay on all eDPs Ray Wu
2026-06-04  2:30   ` Tom Chung
2026-06-02  6:41 ` Ray Wu [this message]
2026-06-04  2:31   ` [PATCH i-g-t 2/2] tests/kms_cursor_legacy: disable AMD Panel Replay during test Tom Chung
2026-06-02  7:47 ` ✓ Xe.CI.BAT: success for kms_cursor_legacy: avoid AMD Panel Replay interference Patchwork
2026-06-02  8:07 ` ✓ i915.CI.BAT: " Patchwork
2026-06-02 14:07 ` ✗ i915.CI.Full: failure " Patchwork
2026-06-02 14:44 ` ✗ Xe.CI.FULL: " 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=20260602064543.58586-3-ray.wu@amd.com \
    --to=ray.wu@amd.com \
    --cc=alex.hung@amd.com \
    --cc=chiahsuan.chung@amd.com \
    --cc=igt-dev@lists.freedesktop.org \
    --cc=sunpeng.li@amd.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