On Fri,  3 Mar 2017 23:05:23 +0000
Daniel Stone <[email protected]> wrote:

> Call drm_output_render unconditionally, doing an early exit if we're
> already rendering a client buffer on the primary plane, and asserting
> for damage on the way out.
> 
> Differential Revision: https://phabricator.freedesktop.org/D1494
> 
> Signed-off-by: Daniel Stone <[email protected]>
> ---
>  libweston/compositor-drm.c | 8 ++++++--
>  1 file changed, 6 insertions(+), 2 deletions(-)
> 
> diff --git a/libweston/compositor-drm.c b/libweston/compositor-drm.c
> index 5a54e29..423bd51 100644
> --- a/libweston/compositor-drm.c
> +++ b/libweston/compositor-drm.c
> @@ -710,6 +710,11 @@ drm_output_render(struct drm_output *output, 
> pixman_region32_t *damage)
>       struct weston_compositor *c = output->base.compositor;
>       struct drm_backend *b = to_drm_backend(c);
>  
> +     /* If we already have a client buffer promoted to scanout, then we don't
> +      * want to render. */
> +     if (output->fb_pending)
> +             return;
> +
>       if (b->use_pixman)
>               drm_output_render_pixman(output, damage);
>       else
> @@ -781,8 +786,7 @@ drm_output_repaint(struct weston_output *output_base,
>       if (output->disable_pending || output->destroy_pending)
>               return -1;
>  
> -     if (!output->fb_pending)
> -             drm_output_render(output, damage);
> +     drm_output_render(output, damage);
>       if (!output->fb_pending)
>               return -1;
>  

Hi,

instead of this, my first idea would be to take the following patch and
extend it even further to return the drm_fb out from
drm_output_render() so that drm_output_repaint() can set
output->fb_pending.

Anyway.

What's the talk about asserting for damage? I don't see it. If you can
just remove that sentence, then:
Reviewed-by: Pekka Paalanen <[email protected]>


Thanks,
pq

Attachment: pgpSxGCt4l_C0.pgp
Description: OpenPGP digital signature

_______________________________________________
wayland-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/wayland-devel

Reply via email to