[Mesa-dev] [PATCH] egl: wayland: bufferAge: only request a back buffer if there isn't one

[Eric Engestrom]

On Saturday, 2016-12-03 22:47:17 +0000, Lionel Landwerlin wrote:
> Seeing gtk+ application lockup when they query the buffer age of a surface.
> 
> Since we update the buffer age field only when creating buffers & swaping
> them on the client side, there shouldn't be any need for requesting a new
> back buffer if there is already one available.
> 
> Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin at intel.com>
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=84016
> ---
>  src/egl/drivers/dri2/platform_wayland.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/src/egl/drivers/dri2/platform_wayland.c b/src/egl/drivers/dri2/platform_wayland.c
> index 395f1e1..84d9ddd 100644
> --- a/src/egl/drivers/dri2/platform_wayland.c
> +++ b/src/egl/drivers/dri2/platform_wayland.c
> @@ -798,7 +798,7 @@ dri2_wl_query_buffer_age(_EGLDriver *drv,
>  {
>     struct dri2_egl_surface *dri2_surf = dri2_egl_surface(surface);
>  
> -   if (get_back_bo(dri2_surf) < 0) {
> +   if (dri2_surf->back == NULL && get_back_bo(dri2_surf) < 0) {

I'm not convinced this is the right fix: this change the behaviour in
two ways:
1. the backing dri_image might not be allocated, which we don't care
   about here because we only want to read dri2_surf->back->age.
2. wl_display_dispatch_queue_pending() not being called, ie. we might
   operate on stale data.

If I understand (2) correctly, it means after your patch, we might
return the buffer age of a buffer about to be released, instead of
processing the queue, letting it get released and getting a new one.
This might fix your lockup, but I think this results in potentially
invalid data being returned to the client.

Cheers,
  Eric