Re: [PATCH] tests/kms: Skip amdgpu's writeback connectors on panning tests

[Harry Wentland <harry.wentland@amd.com>]

On 2024-03-20 22:06, Alex Hung wrote:
> With commit ff1d9a601 writeback connector will be enabled. The
> AMDGPU writeback connector is to be skipped.
> 

Why?

What happens if we try this test on a writeback connector?

Would we expect the same problem to happen on other drivers with writeback connector?

Harry

> Signed-off-by: Alex Hung <alex.hung@amd.com>
> ---
>  tests/kms_plane.c | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/tests/kms_plane.c b/tests/kms_plane.c
> index 406aecc04..d76b3a8f5 100644
> --- a/tests/kms_plane.c
> +++ b/tests/kms_plane.c
> @@ -482,6 +482,12 @@ test_plane_panning(data_t *data, enum pipe pipe)
>  		drmModeModeInfo *m = &output->config.connector->modes[j__];
>  		uint32_t fb_size = m->hdisplay * m->vdisplay * 4;
>  
> +		if (output->config.connector->connector_type == DRM_MODE_CONNECTOR_WRITEBACK &&
> +		   is_amdgpu_device(data->drm_fd)) {
> +			igt_debug("Skipping Writeback connector\n");
> +			continue;
> +		}
> +
>  		/* test allocates 2 double-dim fbs, add one more, to be safe */
>  		if (mem_size && 3 * 4 * fb_size > mem_size) {
>  			igt_debug("Skipping mode %s due to low memory\n", m->name);