[PATCH i-g-t v3 1/1] tests/kms_async_flips: subtests to validate async flips on atomic path
Karthik B S
karthik.b.s at intel.com
Thu Nov 7 05:33:27 UTC 2024
On 10/15/2024 8:22 PM, Santhosh Reddy Guddati wrote:
> subtest to validate async flips with atomic commit
>
> v2: refactor repeated code to function (Karthik)
> explicitly set atomic_path to false in legacy tests (Karthik)
>
> v3: fix test_crc and test_invalid subtests
>
> Signed-off-by: Santhosh Reddy Guddati<santhosh.reddy.guddati at intel.com>
> ---
> tests/kms_async_flips.c | 170 ++++++++++++++++++++++++++++++++++------
> 1 file changed, 144 insertions(+), 26 deletions(-)
>
> diff --git a/tests/kms_async_flips.c b/tests/kms_async_flips.c
> index 5dec71291..b5961028a 100644
> --- a/tests/kms_async_flips.c
> +++ b/tests/kms_async_flips.c
> @@ -67,6 +67,26 @@
> * SUBTEST: invalid-async-flip
> * Description: Negative case to verify if changes in fb are rejected from kernel as expected
> *
> + * SUBTEST: alternate-sync-async-flip-atomic
> + * Description: Verify the async flip functionality and the fps during async flips using atomic path
> + *
> + * SUBTEST: async-flip-with-page-flip-events-atomic
> + * Description: Wait for page flip events in between successive asynchronous flips using atomic path
> + *
> + * SUBTEST: test-time-stamp-atomic
> + * Description: Verify that the async flip timestamp does not coincide with either previous
> + * or next vblank when async flip is done using atomic path
> + *
> + * SUBTEST: test-cursor-atomic
> + * Description: Verify that the DRM_IOCTL_MODE_CURSOR passes after async flip with atomic commit
> + *
> + * SUBTEST: invalid-async-flip-atomic
> + * Description: Negative case to verify if changes in fb are rejected from kernel
> + * as expected when async flip is done using atomic path
> + *
> + * SUBTEST: crc-atomic
> + * Description: Use CRC to verify async flip scans out the correct framebuffer with atomic commit
> + *
> * SUBTEST: async-flip-suspend-resume
> * Description: Verify the async flip functionality with suspend and resume cycle
> */
> @@ -105,6 +125,7 @@ typedef struct {
> bool suspend_resume;
> bool allow_fail;
> struct buf_ops *bops;
> + bool atomic_path;
> } data_t;
>
> static void flip_handler(int fd_, unsigned int sequence, unsigned int tv_sec,
> @@ -259,6 +280,21 @@ static bool async_flip_needs_extra_frame(data_t *data)
> return intel_display_ver(devid) >= 9 || IS_BROADWELL(devid);
> }
>
> +static int perform_flip(data_t *data, int frame, int flags)
> +{
> + int ret;
> +
> + if (!data->atomic_path) {
> + ret = drmModePageFlip(data->drm_fd, data->crtc_id,
> + data->bufs[frame % NUM_FBS].fb_id, flags, data);
> + } else {
> + igt_plane_set_fb(data->plane, &data->bufs[frame % NUM_FBS]);
Hi Santhosh,
> + ret = igt_display_try_commit_atomic(&data->display, flags, data);
Can we use commit here instead of try commit so that the assert check
later can also be avoided?
> + }
> +
> + return ret;
> +}
> +
> static void test_async_flip(data_t *data)
> {
> int ret, frame;
> @@ -276,9 +312,7 @@ static void test_async_flip(data_t *data)
> if (data->alternate_sync_async) {
> flags &= ~DRM_MODE_PAGE_FLIP_ASYNC;
>
> - ret = drmModePageFlip(data->drm_fd, data->crtc_id,
> - data->bufs[frame % NUM_FBS].fb_id,
> - flags, data);
> + ret = perform_flip(data, frame, flags);
>
> igt_assert_eq(ret, 0);
>
> @@ -287,19 +321,16 @@ static void test_async_flip(data_t *data)
> flags |= DRM_MODE_PAGE_FLIP_ASYNC;
>
> if (async_flip_needs_extra_frame(data)) {
> - ret = drmModePageFlip(data->drm_fd, data->crtc_id,
> - data->bufs[frame % NUM_FBS].fb_id,
> - flags, data);
>
> + ret = perform_flip(data, frame, flags);
> igt_assert_eq(ret, 0);
>
> wait_flip_event(data);
> }
> }
>
> - ret = drmModePageFlip(data->drm_fd, data->crtc_id,
> - data->bufs[frame % NUM_FBS].fb_id,
> - flags, data);
> + ret = perform_flip(data, frame, flags);
> +
> if (frame == 1 && data->allow_fail)
> igt_skip_on(ret == -EINVAL);
> else
> @@ -428,9 +459,7 @@ static void test_cursor(data_t *data)
>
> do_ioctl(data->drm_fd, DRM_IOCTL_MODE_CURSOR, &cur);
>
> - ret = drmModePageFlip(data->drm_fd, data->crtc_id,
> - data->bufs[0].fb_id,
> - flags, data);
> + ret = perform_flip(data, 0, flags);
>
> igt_assert_eq(ret, 0);
>
> @@ -450,6 +479,7 @@ static void test_invalid(data_t *data)
> int ret, width, height;
> struct igt_fb fb[2];
> drmModeModeInfo *mode;
> + int flags;
>
> igt_display_commit2(&data->display, data->display.is_atomic ? COMMIT_ATOMIC : COMMIT_LEGACY);
>
> @@ -457,6 +487,8 @@ static void test_invalid(data_t *data)
> width = mode->hdisplay;
> height = mode->vdisplay;
>
> + flags = DRM_MODE_PAGE_FLIP_ASYNC | DRM_MODE_PAGE_FLIP_EVENT;
> +
> igt_create_fb(data->drm_fd, width, height, DRM_FORMAT_XRGB8888,
> I915_FORMAT_MOD_X_TILED, &fb[0]);
> igt_create_fb(data->drm_fd, width, height, DRM_FORMAT_XRGB8888,
> @@ -465,16 +497,27 @@ static void test_invalid(data_t *data)
> igt_plane_set_fb(data->plane, &fb[0]);
> igt_display_commit2(&data->display, data->display.is_atomic ? COMMIT_ATOMIC : COMMIT_LEGACY);
>
> - /* first async flip is expected to allow modifier changes */
> - ret = drmModePageFlip(data->drm_fd, data->crtc_id, fb[1].fb_id,
> - DRM_MODE_PAGE_FLIP_ASYNC | DRM_MODE_PAGE_FLIP_EVENT, data);
> - igt_assert_eq(ret, 0);
> - wait_flip_event(data);
> + if (!data->atomic_path) {
> + /* first async flip is expected to allow modifier changes */
> + ret = drmModePageFlip(data->drm_fd, data->crtc_id, fb[1].fb_id, flags, data);
> + igt_assert_eq(ret, 0);
>
> - /* subsequent async flips should reject modifier changes */
> - ret = drmModePageFlip(data->drm_fd, data->crtc_id, fb[0].fb_id,
> - DRM_MODE_PAGE_FLIP_ASYNC | DRM_MODE_PAGE_FLIP_EVENT, data);
> - igt_assert(ret == -EINVAL);
> + wait_flip_event(data);
> +
> + /* subsequent async flips should reject modifier changes */
> + ret = drmModePageFlip(data->drm_fd, data->crtc_id, fb[0].fb_id, flags, data);
> + igt_assert(ret == -EINVAL);
> + } else {
> + igt_plane_set_fb(data->plane, &fb[1]);
> + ret = igt_display_try_commit_atomic(&data->display, flags, data);
> + igt_assert_eq(ret, 0);
> +
> + wait_flip_event(data);
> +
> + igt_plane_set_fb(data->plane, &fb[0]);
> + ret = igt_display_try_commit_atomic(&data->display, flags, data);
> + igt_assert(ret == -EINVAL);
> + }
Please document the reason as to why we are not using perform_flip
function in the invalid subtest case.
Regards,
Karthik.B.S
>
> /* TODO: Add verification for changes in stride, pixel format */
>
> @@ -588,6 +631,8 @@ static void test_crc(data_t *data)
> int ret, width, height;
> drmModeModeInfoPtr mode;
>
> + igt_display_commit2(&data->display, data->display.is_atomic ? COMMIT_ATOMIC : COMMIT_LEGACY);
> +
> /* make things faster by using a smallish mode */
> mode = &data->output->config.connector->modes[0];
> width = mode->hdisplay;
> @@ -620,8 +665,9 @@ static void test_crc(data_t *data)
> paint_fb(data, &data->bufs[frame], 1, height, 0xff0000ff);
>
> data->flip_pending = true;
> - ret = drmModePageFlip(data->drm_fd, data->crtc_id, data->bufs[frame].fb_id,
> - DRM_MODE_PAGE_FLIP_ASYNC | DRM_MODE_PAGE_FLIP_EVENT, data);
> +
> + ret = perform_flip(data, frame, DRM_MODE_PAGE_FLIP_ASYNC |
> + DRM_MODE_PAGE_FLIP_EVENT);
> igt_assert_eq(ret, 0);
>
> wait_events_crc(data);
> @@ -708,6 +754,18 @@ igt_main
> igt_describe("Wait for page flip events in between successive asynchronous flips");
> igt_subtest_with_dynamic("async-flip-with-page-flip-events") {
> data.alternate_sync_async = false;
> + data.atomic_path = false;
> + if (is_intel_device(data.drm_fd))
> + run_test_with_modifiers(&data, test_async_flip);
> + else
> + run_test(&data, test_async_flip);
> + }
> +
> + igt_describe("Wait for page flip events in between successive "
> + "asynchronous flips using atomic path");
> + igt_subtest_with_dynamic("async-flip-with-page-flip-events-atomic") {
> + data.alternate_sync_async = false;
> + data.atomic_path = true;
> if (is_intel_device(data.drm_fd))
> run_test_with_modifiers(&data, test_async_flip);
> else
> @@ -717,12 +775,30 @@ igt_main
> igt_describe("Alternate between sync and async flips");
> igt_subtest_with_dynamic("alternate-sync-async-flip") {
> data.alternate_sync_async = true;
> + data.atomic_path = false;
> + run_test(&data, test_async_flip);
> + }
> +
> + igt_describe("Alternate between sync and async flips using atomic path");
> + igt_subtest_with_dynamic("alternate-sync-async-flip-atomic") {
> + data.alternate_sync_async = true;
> + data.atomic_path = true;
> run_test(&data, test_async_flip);
> }
>
> - igt_describe("Verify that the async flip timestamp does not coincide with either previous or next vblank");
> - igt_subtest_with_dynamic("test-time-stamp")
> + igt_describe("Verify that the async flip timestamp does not "
> + "coincide with either previous or next vblank");
> + igt_subtest_with_dynamic("test-time-stamp") {
> + data.atomic_path = false;
> run_test(&data, test_timestamp);
> + }
> +
> + igt_describe("Verify that the async flip timestamp does not coincide "
> + "with either previous or next vblank with atomic path");
> + igt_subtest_with_dynamic("test-time-stamp-atomic") {
> + data.atomic_path = true;
> + run_test(&data, test_timestamp);
> + }
> }
>
> igt_describe("Verify that the DRM_IOCTL_MODE_CURSOR passes after async flip");
> @@ -733,9 +809,25 @@ igt_main
> * supported in cursor plane.
> */
> igt_skip_on_f(i915_psr2_selective_fetch_check(data.drm_fd, NULL),
> - "PSR2 sel fetch causes cursor to be added to primary plane " \
> + "PSR2 sel fetch causes cursor to be added to primary plane "
> "pages flips and async flip is not supported in cursor\n");
>
> + data.atomic_path = false;
> + run_test(&data, test_cursor);
> + }
> +
> + igt_describe("Verify that the DRM_IOCTL_MODE_CURSOR passes after "
> + "async flip with atomic commit");
> + igt_subtest_with_dynamic("test-cursor-atomic") {
> + /*
> + * Intel's PSR2 selective fetch adds other planes to state when
> + * necessary, causing the async flip to fail because async flip is not
> + * supported in cursor plane.
> + */
> + igt_skip_on_f(i915_psr2_selective_fetch_check(data.drm_fd, NULL),
> + "PSR2 sel fetch causes cursor to be added to primary plane "
> + "pages flips and async flip is not supported in cursor\n");
> + data.atomic_path = true;
> run_test(&data, test_cursor);
> }
>
> @@ -748,6 +840,21 @@ igt_main
> igt_require(igt_display_has_format_mod(&data.display, DRM_FORMAT_XRGB8888,
> I915_FORMAT_MOD_Y_TILED));
>
> + data.atomic_path = false;
> + run_test(&data, test_invalid);
> + }
> +
> + igt_describe("Negative case to verify if changes in fb are rejected "
> + "from kernel as expected when async flip is done using atomic path");
> + igt_subtest_with_dynamic("invalid-async-flip-atomic") {
> + data.atomic_path = true;
> + /* TODO: support more vendors */
> + igt_require(is_intel_device(data.drm_fd));
> + igt_require(igt_display_has_format_mod(&data.display, DRM_FORMAT_XRGB8888,
> + I915_FORMAT_MOD_X_TILED));
> + igt_require(igt_display_has_format_mod(&data.display, DRM_FORMAT_XRGB8888,
> + I915_FORMAT_MOD_Y_TILED));
> +
> run_test(&data, test_invalid);
> }
>
> @@ -756,6 +863,17 @@ igt_main
> /* Devices without CRC can't run this test */
> igt_require_pipe_crc(data.drm_fd);
>
> + data.atomic_path = false;
> + run_test(&data, test_crc);
> + }
> +
> + igt_describe("Use CRC to verify async flip scans out the correct framebuffer "
> + "with atomic commit");
> + igt_subtest_with_dynamic("crc-atomic") {
> + /* Devices without CRC can't run this test */
> + igt_require_pipe_crc(data.drm_fd);
> +
> + data.atomic_path = true;
> run_test(&data, test_crc);
> }
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/igt-dev/attachments/20241107/30b5c72f/attachment-0001.htm>
More information about the igt-dev
mailing list