Am Mittwoch, den 17.10.2018, 12:56 -0400 schrieb Ilia Mirkin:
> On Wed, Oct 17, 2018 at 12:39 PM Gert Wollny <gw.foss...@gmail.com>
> wrote:
> > 
> > From: Gert Wollny <gert.wol...@collabora.com>
> > 
> > With this patch the extension EXT_sRGB_write_control is enabled for
> > gallium drivers that support sRGB formats as render targets.
> > 
> > Tested (and pass) on r600(evergreen) and softpipe:
> > 
> >   dEQP-
> > GLES31.functional.fbo.srgb_write_control.framebuffer_srgb_enabled*
> > 
> > with "MESA_GLES_VERSION_OVERRIDE=3.2" (the tests needlessly check
> > for this)
> > 
> > Signed-off-by: Gert Wollny <gert.wol...@collabora.com>
> > ---
> >  src/mesa/state_tracker/st_manager.c | 17 +++++++++--------
> >  1 file changed, 9 insertions(+), 8 deletions(-)
> > 
> > diff --git a/src/mesa/state_tracker/st_manager.c
> > b/src/mesa/state_tracker/st_manager.c
> > index ceb48dd490..562b12a1ef 100644
> > --- a/src/mesa/state_tracker/st_manager.c
> > +++ b/src/mesa/state_tracker/st_manager.c
> > @@ -457,14 +457,12 @@ st_framebuffer_create(struct st_context *st,
> >      * format such that util_format_srgb(visual->color_format) can
> > be supported
> >      * by the pipe driver.  We still need to advertise the
> > capability here.
> >      *
> > -    * For GLES, however, sRGB framebuffer write is controlled only
> > by the
> > -    * capability of the framebuffer.  There is
> > GL_EXT_sRGB_write_control to
> > -    * give applications the control back, but sRGB write is still
> > enabled by
> > -    * default.  To avoid unexpected results, we should not
> > advertise the
> > -    * capability.  This could change when we add support for
> > -    * EGL_KHR_gl_colorspace.
> > +    * For GLES, however, sRGB framebuffer write is initially only
> > controlled
> > +    * by the capability of the framebuffer, but with
> > GL_EXT_sRGB_write_control
> > +    * control is given back to the applications. Similar to
> > desktop GL
> > +    * support for this extension depends EXT_framebuffer_sRGB.
> >      */
> > -   if (_mesa_is_desktop_gl(st->ctx)) {
> > +   {
> >        struct pipe_screen *screen = st->pipe->screen;
> >        const enum pipe_format srgb_format =
> >           util_format_srgb(stfbi->visual->color_format);
> > @@ -475,8 +473,11 @@ st_framebuffer_create(struct st_context *st,
> >                                        PIPE_TEXTURE_2D, stfbi-
> > >visual->samples,
> >                                        stfbi->visual->samples,
> >                                        (PIPE_BIND_DISPLAY_TARGET |
> > -                                       PIPE_BIND_RENDER_TARGET)))
> > +                                       PIPE_BIND_RENDER_TARGET)))
> > {
> >           mode.sRGBCapable = GL_TRUE;
> > +         /* Exposing this as extension is only needed on GLES */
> > +         st->ctx->Extensions.EXT_sRGB_write_control =
> > !_mesa_is_desktop_gl(st->ctx);
> 
> Having weird dependencies in extension enables creates a lot of
> confusion. I'd just flip it to true.
My resasoning here was that this is a GLES only extension, but I now
see that this is acctually done via the extension table. 

Thanks for all the pointers. 
Gert 

> 
> > +      }
> >     }
> > 
> >     _mesa_initialize_window_framebuffer(&stfb->Base, &mode);
> > --
> > 2.18.1
> > 
> > _______________________________________________
> > mesa-dev mailing list
> > mesa-dev@lists.freedesktop.org
> > https://lists.freedesktop.org/mailman/listinfo/mesa-dev
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev

Reply via email to