[prefix=PATCH weston] gl-renderer.c: Pass visual ID for choosing egl configs for pbuffer

Madhurkiran Harikrishnan MADHURKI at xilinx.com
Tue Jul 31 01:02:36 UTC 2018


Hi Daniel,

Thanks for the comments. For a not having configless_context,  can you explain why should I check gr->egl_config, because the attributes passed while getting this gr->egl_config does not have EGL_PBUFFER_BIT set for EGL_SURFACE_TYPE, which means config returned from the driver would not have this bit set. In that case, shouldn't pbuffer_config just suffice.

Thanks
Madhurkiran

-----Original Message-----
From: Daniel Stone [mailto:daniel at fooishbar.org] 
Sent: Sunday, July 22, 2018 3:30 AM
To: Madhurkiran Harikrishnan <MADHURKI at xilinx.com>
Cc: wayland <wayland-devel at lists.freedesktop.org>; nofooter <nofooter at xilinx.com>; Madhurkiran Harikrishnan <MADHURKI at xilinx.com>
Subject: Re: [prefix=PATCH weston] gl-renderer.c: Pass visual ID for choosing egl configs for pbuffer

Hi Madhurkiran,

On Wed, 21 Mar 2018 at 22:45, Madhurkiran Harikrishnan <madhurkiran.harikrishnan at xilinx.com> wrote:
> +       if (pbuffer_config != gr->egl_config &&
> +           !gr->has_configless_context) {
> +               weston_log("attempted to use a different EGL config for an "
> +                          "output but EGL_KHR_no_config_context or "
> +                          "EGL_MESA_configless_context is not supported\n");
> +               return -1;
> +       }

If we must ensure that the config is exactly identical (not having configless_context), then we should just use gr->egl_config directly, but ensure that it has EGL_WINDOW_BIT and EGL_PBUFFER_BIT set when it is selected for the main output surface. Could you please rework the patch so it does this? The rest seems fine to me though.

Cheers,
Daniel


More information about the wayland-devel mailing list