[Intel-gfx] [PATCH 01/12] tests/kms_atomic_transition: use select + read instead of blocking read

[Daniel Vetter]

On Mon, Nov 14, 2016 at 06:59:14PM +0900, Gustavo Padovan wrote:
> From: Gustavo Padovan <gustavo.padovan at collabora.co.uk>
> 
> If the event never arrives we can timeout with select and end the test.
> 
> Signed-off-by: Gustavo Padovan <gustavo.padovan at collabora.co.uk>
> ---
>  tests/kms_atomic_transition.c | 8 ++++++++
>  1 file changed, 8 insertions(+)
> 
> diff --git a/tests/kms_atomic_transition.c b/tests/kms_atomic_transition.c
> index 1977993..e693c88 100644
> --- a/tests/kms_atomic_transition.c
> +++ b/tests/kms_atomic_transition.c
> @@ -296,6 +296,14 @@ static void commit_display(igt_display_t *display, unsigned event_mask, bool non
>  		struct drm_event *e = (void *)buf;
>  		struct drm_event_vblank *vblank = (void *)buf;
>  		uint32_t crtc_id, pipe = I915_MAX_PIPES;
> +		struct timeval timeout = { .tv_sec = 3, .tv_usec = 0 };
> +		fd_set fds;
> +
> +		FD_ZERO(&fds);
> +		FD_SET(0, &fds);
> +		FD_SET(display->drm_fd, &fds);
> +		ret = select(display->drm_fd + 1, &fds, NULL, NULL, &timeout);
> +		igt_assert(ret > 0);

Hm, we have igt_timeout which sends a signal and kills the test if the
timeout expires. And drm event reading should be interruptible. That might
be an even simpler way to implement this.
-Daniel

>  
>  		ret = read(display->drm_fd, buf, sizeof(buf));
>  		if (ret < 0 && (errno == EINTR || errno == EAGAIN))
> -- 
> 2.5.5
> 
> _______________________________________________
> dri-devel mailing list
> dri-devel at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/dri-devel

-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch