[igt-dev] [PATCH] tests/kms_atomic_interruptible: fix a warning message on AMD GPU

Rodrigo Siqueira Rodrigo.Siqueira at amd.com
Thu Aug 26 12:32:27 UTC 2021


On 08/25, Tianci Yin wrote:
> From: "Tianci.Yin" <tianci.yin at amd.com>
> 
> [why]
> For subtest legacy-setmode and atomic-setmode, fb buffer is
> unbinded from plane by block_plane(), then igt_display_commit2()
> will trigger a dmesg warning message in drm_atomic_helper_commit(),
> since for block commit, drm will wait fences that represents
> finishing fb writing, when it find this fence has no accompanying
> fb, it gives out a warning.
> 
> [how]
> Just remove igt_plane_set_fb() from block_plane() to retain
> the fb binding with plane.
> Verified on AMD Radeon Pro W5500 and Intel UHD Graphics 630
> (inside i7-9700K), both pass.
> 
> Signed-off-by: Tianci.Yin <tianci.yin at amd.com>
> ---
>  tests/kms_atomic_interruptible.c | 1 -
>  1 file changed, 1 deletion(-)
> 
> diff --git a/tests/kms_atomic_interruptible.c b/tests/kms_atomic_interruptible.c
> index 0a127921..03980d6e 100644
> --- a/tests/kms_atomic_interruptible.c
> +++ b/tests/kms_atomic_interruptible.c
> @@ -49,7 +49,6 @@ static int block_plane(igt_display_t *display, igt_output_t *output, enum plane_
>  
>  		if (test_type == test_legacy_modeset || test_type == test_atomic_modeset) {
>  			igt_output_set_pipe(output, PIPE_NONE);
> -			igt_plane_set_fb(plane, NULL);
>  		}
>  		igt_plane_set_fence_fd(plane, sw_sync_timeline_create_fence(timeline, 1));
>  
> -- 
> 2.25.1
>

Hi,

This change lgtm. I tested it on vkms, amdgpu, and vc4 (test skip).

Reviewed-by: Rodrigo Siqueira <Rodrigo.Siqueira at amd.com>

Thanks
Siqueira

-- 
Rodrigo Siqueira
https://siqueira.tech


More information about the igt-dev mailing list