From: "Thasleem, Mohammed" <mohammed.thasleem@intel.com>
To: Jeevan B <jeevan.b@intel.com>, <igt-dev@lists.freedesktop.org>
Cc: <dibin.moolakadan.subrahmanian@intel.com>
Subject: Re: [PATCH i-g-t 4/4] tests/intel/kms_pm_dc: Add new test for dc3co framedrop validation
Date: Thu, 9 Apr 2026 16:33:15 +0530 [thread overview]
Message-ID: <51c24cf3-99e5-4a83-999b-47ce6ca0393d@intel.com> (raw)
In-Reply-To: <20260304043805.572087-5-jeevan.b@intel.com>
[-- Attachment #1: Type: text/plain, Size: 5259 bytes --]
On 04-03-2026 10:08 am, Jeevan B wrote:
> Add a new subtest to validate that no frame drops occur during
> DC3CO entry, ensuring that no frame drops are detected and DC3CO
> is successfully triggered during the test.
>
> Signed-off-by: Jeevan B<jeevan.b@intel.com>
> ---
> tests/intel/kms_pm_dc.c | 83 +++++++++++++++++++++++++++++++++++++++++
> 1 file changed, 83 insertions(+)
>
> diff --git a/tests/intel/kms_pm_dc.c b/tests/intel/kms_pm_dc.c
> index 10450474a..b0e35dca1 100644
> --- a/tests/intel/kms_pm_dc.c
> +++ b/tests/intel/kms_pm_dc.c
> @@ -51,6 +51,10 @@
> * Description: Make sure that system enters DC3CO when PSR2 is active and system
> * is in SLEEP state
> *
> + * SUBTEST: dc3co-framedrop-check
> + * Description: Verify that DC3CO entry does not cause frame drops and successfully
> + * enters the power state
> + *
> * SUBTEST: dc5-dpms
> * Description: Validate display engine entry to DC5 state while all connectors's
> * DPMS property set to OFF
> @@ -315,6 +319,50 @@ static void check_dc3co_with_videoplayback_like_load(data_t *data)
> check_dc_counter(data, IGT_INTEL_CHECK_DC3CO, dc3co_prev_cnt);
> }
>
> +static void check_framedrop(data_t *data)
> +{
> + igt_plane_t *primary;
> + uint32_t dc3co_prev_cnt, dc3co_cnt;
> + int delay, frame_count, max_count = 100, ret;
-->set frame_count with 0.
> + bool dc3co_flag = false;
> + drmVBlank wait;
> +
> + primary = igt_output_get_plane_type(data->output,
> + DRM_PLANE_TYPE_PRIMARY);
> + igt_plane_set_fb(primary, NULL);
> + dc3co_prev_cnt = igt_read_dc_counter(data->debugfs_fd,
> + IGT_INTEL_CHECK_DC3CO);
> +
> + /* Calculate delay to generate idle frame in usec*/
> + delay = 1.5 * ((1000 * 1000) / data->mode->vrefresh);
-->better to have : (int) (1.5 * ((<rest code>)); so proepr conversion
to integer usec
> +
> + for (int i = 0; i < max_count; i++) {
> + if (i % 2 == 0)
> + igt_plane_set_fb(primary, &data->fb_rgb);
> + else
> + igt_plane_set_fb(primary, &data->fb_rgr);
> +
> + igt_display_commit(&data->display);
> + frame_count++;
> +
> + memset(&wait, 0, sizeof(wait));
> + wait.request.type = DRM_VBLANK_RELATIVE;
> + wait.request.sequence = 1;
> +
> + ret = drmWaitVBlank(data->drm_fd, &wait);
> + igt_assert_eq(ret, 0);
> + dc3co_cnt = igt_read_dc_counter(data->debugfs_fd,
> + IGT_INTEL_CHECK_DC3CO);
> + if (dc3co_cnt > dc3co_prev_cnt)
> + dc3co_flag = true;
> +
> + usleep(delay);
> + }
> +
> + igt_assert_f(dc3co_flag, "DC3CO entry failed.\n");
> + igt_assert_f(frame_count == max_count, "Framedrop seen during vpb scenario.\n");
--> Here we are just counting a loop itiration, as per my understanding
the real frame drop should be vblank seq numers.
I think we have to do like:
before start loop read prev_vblank = wait.reply.sequence and
inside loop current_vblank = wait.reply.sequence;
then vblank_diff = current_vblank - prev_vblank, so the gap
greter than 1 indicates frame drop.
on everyloop if the gap is greter than previous, we can
increment frame_drop + = vblank_diff-1,
and updated the prev_vblank with current_vblank in side the
loop till the loop repeats.
which will give vblank interval w.r.t actual expected, Please
check it...
--> if it is intention then better to updated message some thing like
-->"Frame processing" instead of "Frame drop" ??
-->also print number of frame count and max count along with the message.
> +}
> +
> static void setup_dc3co(data_t *data, enum psr_mode mode)
> {
> data->op_psr_mode = mode;
> @@ -333,6 +381,16 @@ static void test_dc3co_vpb_simulation(data_t *data, enum psr_mode mode)
> cleanup_dc3co_fbs(data);
> }
>
> +static void test_framedrop_dc3co(data_t *data, enum psr_mode mode)
> +{
> + igt_require_dc_counter(data->debugfs_fd, IGT_INTEL_CHECK_DC3CO);
> + setup_output(data);
> + setup_dc3co(data, mode);
> + setup_videoplayback(data);
> + check_framedrop(data);
> + cleanup_dc3co_fbs(data);
> +}
> +
> static void test_dc5_retention_flops(data_t *data, int dc_flag)
> {
> uint32_t dc_counter_before_psr;
> @@ -660,6 +718,31 @@ int igt_main()
> }
> }
>
> + igt_describe("Verify that DC3CO entry does not cause frame drops "
> + "and successfully enters the power state");
> + igt_subtest_with_dynamic("dc3co-framedrop-check") {
> + int modes[] = {PSR_MODE_2, PR_MODE};
> + const char *append_subtest_name[2] = {
> + "psr2-",
> + "pr-",
-->here also same comments, provided at patch 2
> + };
> +
> + for (int i = 0; i < ARRAY_SIZE(modes); i++) {
> + igt_dynamic_f("%s-dc3co-framedrop", append_subtest_name[i]) {
> + igt_require(psr_sink_support(data.drm_fd, data.debugfs_fd,
> + modes[i], NULL));
> +
> + if (modes[i] == PSR_MODE_2)
> + igt_require(IS_TIGERLAKE(data.devid) ||
> + intel_display_ver(data.devid) >= 35);
> + else if (modes[i] == PR_MODE)
> + igt_require(intel_display_ver(data.devid) >= 35);
> +
> + test_framedrop_dc3co(&data, modes[i]);
> + }
> + }
> + }
> +
> igt_describe("This test validates display engine entry to DC5 state "
> "while PSR is active");
> igt_subtest("dc5-psr") {
[-- Attachment #2: Type: text/html, Size: 6637 bytes --]
next prev parent reply other threads:[~2026-04-09 11:03 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-03-04 4:38 [PATCH i-g-t 0/4] Enable and Add new tests for DC3CO Jeevan B
2026-03-04 4:38 ` [PATCH i-g-t 1/4] tests/intel/kms_pm_dc: Replace require with proper assertion Jeevan B
2026-04-07 9:29 ` Thasleem, Mohammed
2026-03-04 4:38 ` [PATCH i-g-t 2/4] tests/intel/kms_pm_dc: Enable DC3CO test for PSR2/PR modes Jeevan B
2026-04-09 6:06 ` Thasleem, Mohammed
2026-03-04 4:38 ` [PATCH i-g-t 3/4] tests/kms_vrr: Add new test for DC3CO validation with LOBF Jeevan B
2026-04-09 8:12 ` Thasleem, Mohammed
2026-03-04 4:38 ` [PATCH i-g-t 4/4] tests/intel/kms_pm_dc: Add new test for dc3co framedrop validation Jeevan B
2026-04-09 11:03 ` Thasleem, Mohammed [this message]
2026-03-04 7:30 ` ✓ Xe.CI.BAT: success for Enable and Add new tests for DC3CO (rev2) Patchwork
2026-03-04 7:30 ` ✓ i915.CI.BAT: " Patchwork
2026-03-05 5:30 ` ✗ Xe.CI.FULL: failure " Patchwork
2026-03-05 9:52 ` ✗ i915.CI.Full: " Patchwork
-- strict thread matches above, loose matches on Subject: below --
2026-03-03 9:39 [PATCH i-g-t 0/4] Enable and Add new tests for DC3CO Jeevan B
2026-03-03 9:39 ` [PATCH i-g-t 4/4] tests/intel/kms_pm_dc: Add new test for dc3co framedrop validation Jeevan B
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=51c24cf3-99e5-4a83-999b-47ce6ca0393d@intel.com \
--to=mohammed.thasleem@intel.com \
--cc=dibin.moolakadan.subrahmanian@intel.com \
--cc=igt-dev@lists.freedesktop.org \
--cc=jeevan.b@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