Re: [PATCH i-g-t v3 4/6] tests/kms_properties: keep primary plane enabled for each CRTC

On 8/11/26 08:23, Melissa Wen wrote:
> For AMD, a primary plane must be active to keep CRTC active. So stop
> reusing primary plane fb for overlays, give the latter its own fb.
> 
> Signed-off-by: Melissa Wen <[email protected]>
> 
> ---
> 
> v2:
> - detach different changes from a single commit (Chaitanya)

Were patches 4 & 5 a single patch? I think they belong to a single patch.

Ignore me if this was discussed before.

> ---
>   tests/kms_properties.c | 12 +++++++++++-
>   1 file changed, 11 insertions(+), 1 deletion(-)
> 
> diff --git a/tests/kms_properties.c b/tests/kms_properties.c
> index 292a04a2a..87614346a 100644
> --- a/tests/kms_properties.c
> +++ b/tests/kms_properties.c
> @@ -237,7 +237,7 @@ static void run_colorop_property_tests(igt_display_t *display,
>   				       igt_crtc_t *crtc, igt_output_t *output,
>   				       bool atomic)
>   {
> -	struct igt_fb fb;
> +	struct igt_fb fb, afb;
>   	igt_plane_t *plane;
>   	igt_colorop_t *colorop;
>   	int i;
> @@ -255,6 +255,16 @@ static void run_colorop_property_tests(igt_display_t *display,
>   			 igt_crtc_name(crtc), plane->index,
>   			 kmstest_plane_type_name(plane->type), output->name);
>   
> +		/* AMD requires primary plane enabled to keep CRTC enabled */
> +		if (plane->type != DRM_PLANE_TYPE_PRIMARY) {
> +			drmModeModeInfo *mode = igt_output_get_mode(output);
> +
> +			igt_create_pattern_fb(display->drm_fd, mode->hdisplay, mode->vdisplay,
> +					      DRM_FORMAT_XRGB8888, DRM_FORMAT_MOD_LINEAR, &afb);
> +
> +			igt_plane_set_fb(plane, &afb);
> +		}
> +
>   		/* iterate over all color pipelines on plane */
>   		for (i = 0; i < plane->num_color_pipelines; ++i) {
>   			/* iterate over all colorops in pipeline*/