[igt-dev] [PATCH i-g-t] lib/igt_kms: commit only the primary plane which is prepared

Ville Syrjälä ville.syrjala at linux.intel.com
Wed Mar 23 13:52:14 UTC 2022


On Tue, Mar 22, 2022 at 04:21:29PM -0700, Abhinav Kumar wrote:
> kms_atomic's test-only sub-test prepares only one primary plane
> but igt_plane_commit() tries to commit all the primary planes.
> 
> For drivers having only one primary plane per display still
> work fine but when there is more than one primary plane, since
> FB_ID is not assigned for the second one, the API
> igt_primary_plane_commit_legacy() ends up calling the CRTC
> disable path incorrectly and hence failing the test.
> 
> Since only one primary plane has been prepared, commit only
> that one by matching the index correctly.
> 
> Signed-off-by: Abhinav Kumar <quic_abhinavk at quicinc.com>
> ---
>  lib/igt_kms.c | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/lib/igt_kms.c b/lib/igt_kms.c
> index 34a2aa0..98a6e1b 100644
> --- a/lib/igt_kms.c
> +++ b/lib/igt_kms.c
> @@ -3189,6 +3189,8 @@ static int igt_plane_commit(igt_plane_t *plane,
>  			    enum igt_commit_style s,
>  			    bool fail_on_error)
>  {
> +	igt_plane_t *plane_primary = igt_pipe_get_plane_type(pipe, DRM_PLANE_TYPE_PRIMARY);
> +
>  	if (pipe->display->first_commit || (s == COMMIT_UNIVERSAL &&
>  	     igt_plane_is_prop_changed(plane, IGT_PLANE_ROTATION))) {
>  		int ret;
> @@ -3199,7 +3201,8 @@ static int igt_plane_commit(igt_plane_t *plane,
>  
>  	if (plane->type == DRM_PLANE_TYPE_CURSOR && s == COMMIT_LEGACY) {
>  		return igt_cursor_commit_legacy(plane, pipe, fail_on_error);
> -	} else if (plane->type == DRM_PLANE_TYPE_PRIMARY && s == COMMIT_LEGACY) {
> +	} else if (plane->type == DRM_PLANE_TYPE_PRIMARY && s == COMMIT_LEGACY &&
> +			(plane->index == plane_primary->index)) {

I think you should be able to simplify that to just
'plane == plane_primary && s == COMMIT_LEGACY'

>  		return igt_primary_plane_commit_legacy(plane, pipe,
>  						       fail_on_error);
>  	} else {
> -- 
> 2.7.4

-- 
Ville Syrjälä
Intel


More information about the igt-dev mailing list