[Intel-gfx] [PATCH] i-g-t: Adding display NV12 testcase

Tvrtko Ursulin tvrtko.ursulin at linux.intel.com
Mon May 11 05:12:20 PDT 2015


Hi,

On 05/09/2015 05:32 AM, Chandra Konduru wrote:
> From: chandra konduru <chandra.konduru at intel.com>
>
> This patch adds kms_nv12 test case. It covers testing NV12 in
> all supported linear/tile-X/tile-Y/tile-Yf tile formats in
> 0 and 180 orientations. For each tiling format, it tests
> various combinations of planes and scaling.
>
> v2:
> -Added 90/270 tests (me)
> -took out crc test as it isn't adding much value due to chroma upsampling (me)
>
> Signed-off-by: chandra konduru <chandra.konduru at intel.com>
> ---
>   tests/.gitignore       |   1 +
>   tests/Android.mk       |   1 +
>   tests/Makefile.sources |   1 +
>   tests/kms_nv12.c       | 619 +++++++++++++++++++++++++++++++++++++++++++++++++
>   4 files changed, 622 insertions(+)
>   create mode 100644 tests/kms_nv12.c

[snip]

> +igt_main
> +{
> +	data_t data = {};
> +	igt_skip_on_simulation();
> +
> +	igt_fixture {
> +		data.drm_fd = drm_open_any();
> +
> +		kmstest_set_vt_graphics_mode();
> +
> +		igt_display_init(&data.display, data.drm_fd);
> +	}
> +	devid = intel_get_drm_devid(data.drm_fd);
> +
> +	data.num_scalers = intel_gen(devid) >= 9 ? 2 : 0;
> +
> +	igt_assert(intel_gen(devid) >= 9);

Not reviewing as such, but notice --list-subtest is broken unless you 
move the above three lines into the fixture.

After that I ran:

> +	data.rotation = IGT_ROTATION_90;
> +	igt_subtest_f("nv12-plane-tile-y-rot-90") {
> +		data.tiled = LOCAL_I915_FORMAT_MOD_Y_TILED;
> +		test_nv12_plane_rotation_90_or_270(&data);
> +	}
> +	igt_subtest_f("nv12-plane-tile-yf-rot-90") {
> +		data.tiled = LOCAL_I915_FORMAT_MOD_Yf_TILED;
> +		test_nv12_plane_rotation_90_or_270(&data);
> +	}

And both succeeded! :)

Your + mine patch series rebased on top of today's nightly.

Regards,

Tvrtko


More information about the Intel-gfx mailing list