[igt-dev] [PATCH i-g-t v2] igt/tests: Clear pipes properly in kms_atomic_transition

[Maarten Lankhorst]

Op 30-06-2020 om 15:24 schreef Stanislav Lisovskiy:
> There is an issue happening from time to time in kms_atomic_transition
> (bug #1918). We periodically get assertion that some two outputs
> attempt to use same pipe like this:
>
> "Failed assertion: output->pending_pipe != b->pending_pipe"
>
> After some investigation came to conclusion that this is happening
> because we are calling igt_output_set_pipe(output, PIPE_NONE) only
> for connected outputs, which is wrong.
> Periodically igt_display_refresh/igt_output_refresh call calls can
> update the output state to disconnected. However that doesn't clear
> the pipe being assigned(i.e output->pending_pipe).
> So this causes assertion to be triggered on next igt_display_refresh
> called during commit.
>
> Bugzilla: https://gitlab.freedesktop.org/drm/intel/-/issues/1918
>
> v2: - Do not use for_each_valid_output_on_pipe as it also iterates
>       only on connected outputs(Maarten)
>     - Also fix run_modeset_tests function
>
> Signed-off-by: Stanislav Lisovskiy <stanislav.lisovskiy at intel.com>
> ---
>  tests/kms_atomic_transition.c | 10 ++++++----
>  1 file changed, 6 insertions(+), 4 deletions(-)
>
> diff --git a/tests/kms_atomic_transition.c b/tests/kms_atomic_transition.c
> index 754a4969..656bd51e 100644
> --- a/tests/kms_atomic_transition.c
> +++ b/tests/kms_atomic_transition.c
> @@ -664,9 +664,10 @@ static unsigned set_combinations(igt_display_t *display, unsigned mask, struct i
>  	igt_output_t *output;
>  	enum pipe pipe;
>  	unsigned event_mask = 0;
> +	int i;
>  
> -	for_each_connected_output(display, output)
> -		igt_output_set_pipe(output, PIPE_NONE);
> +	for (i = 0; i < display->n_outputs; i++)
> +		igt_output_set_pipe(&display->outputs[i], PIPE_NONE);
>  
>  	for_each_pipe(display, pipe) {
>  		igt_plane_t *plane = igt_pipe_get_plane_type(&display->pipes[pipe],
> @@ -743,11 +744,12 @@ static void run_modeset_tests(igt_display_t *display, int howmany, bool nonblock
>  	igt_output_t *output;
>  	unsigned width = 0, height = 0;
>  
> +	for (i = 0; i < display->n_outputs; i++)
> +		igt_output_set_pipe(&display->outputs[i], PIPE_NONE);
> +
>  	for_each_connected_output(display, output) {
>  		drmModeModeInfo *mode = igt_output_get_mode(output);
>  
> -		igt_output_set_pipe(output, PIPE_NONE);
> -
>  		width = max(width, mode->hdisplay);
>  		height = max(height, mode->vdisplay);
>  	}

Much better!

Reviewed-by: Maarten Lankhorst <maarten.lankhorst at linux.intel.com>

Just worrying that we may lose some test coverage when outputs disappear, but oh well..