[virglrenderer-devel] [PATCH] vrend_clear: clear and restore rasterizer discard and depth clamp

[Gert Wollny]

Am Freitag, den 11.05.2018, 14:33 -0700 schrieb Gurchetan Singh:
> This regresses the following dEQP tests when run on a Nvidia GL host:
> 
> dEQP-GLES3.functional.fragment_ops.interaction.basic_shader.2
> dEQP-GLES3.functional.fragment_ops.interaction.basic_shader.10
> dEQP-GLES3.functional.fragment_ops.interaction.basic_shader.11
> dEQP-GLES3.functional.fragment_ops.interaction.basic_shader.14
> dEQP-GLES3.functional.fragment_ops.interaction.basic_shader.16
> dEQP-GLES3.functional.fragment_ops.interaction.basic_shader.25
> dEQP-GLES3.functional.fragment_ops.interaction.basic_shader.26
> dEQP-GLES3.functional.fragment_ops.interaction.basic_shader.39
> dEQP-GLES3.functional.fragment_ops.interaction.basic_shader.40
> dEQP-GLES3.functional.fragment_ops.interaction.basic_shader.41
> dEQP-GLES3.functional.fragment_ops.interaction.basic_shader.56
> dEQP-GLES3.functional.fragment_ops.interaction.basic_shader.61
> dEQP-GLES3.functional.fragment_ops.interaction.basic_shader.68
> dEQP-GLES3.functional.fragment_ops.interaction.basic_shader.70
> dEQP-GLES3.functional.fragment_ops.interaction.basic_shader.73
> dEQP-GLES3.functional.fragment_ops.interaction.basic_shader.90
> dEQP-GLES3.functional.fragment_ops.interaction.basic_shader.96

Sorry, I probably didn't recheck on R600 after testing with my first
patch (the one on the guser side). 

> Commenting out the depth clamp changes seems to fix them, though...
Unfortunaltely it is needed for some of the rasterizer_ddiscard tests. 

Given that with the guest side patch also these tests pass, I guess
that there is some other state that needs to be cleared. I'll
investigate which state that might be. 

Best, 
Gert





> 
> On Tue, May 8, 2018 at 11:35 AM Gurchetan Singh
> <gurchetansingh at chromium.org>
> wrote:
> 
> > Reviewed-by: Gurchetan Singh <gurchetansingh at chromium.org>
> > On Tue, May 8, 2018 at 3:47 AM, Gert Wollny <gert.wollny at collabora.
> > com>
> 
> wrote:
> > > vrend_clear is passed in from gallium clear which, unlike
> > > glClear, must
> > > not obey rasterizer discard and depth_clmap. However, the guest
> > > doesn't
> > > foreard the latest update of the rendering states to the host
> > > before
> 
> issuing
> > > "clear". Therefore, the host clear command must clear these
> > > states
> 
> before
> > > calling glClear in turn. After calling glClear these states are
> > > again
> 
> brought
> > > into sync with what the host assumes.
> > > 
> > > Fixes CTS:
> > >    dEQP-GLES3.functional.rasterizer_discard.fbo.*
> > > 
> > > Signed-off-by: Gert Wollny <gert.wollny at collabora.com>
> > > ---
> > >  src/vrend_renderer.c | 17 ++++++++++++++++-
> > >  1 file changed, 16 insertions(+), 1 deletion(-)
> > > 
> > > diff --git a/src/vrend_renderer.c b/src/vrend_renderer.c
> > > index ac9723c..34fa32b 100644
> > > --- a/src/vrend_renderer.c
> > > +++ b/src/vrend_renderer.c
> > > @@ -2486,6 +2486,8 @@ void vrend_clear(struct vrend_context *ctx,
> > >        } else {
> > >           glClearDepth(depth);
> > >        }
> > > +      if (!ctx->sub->hw_rs_state.depth_clip)
> > > +          glDisable(GL_DEPTH_CLAMP);
> > >     }
> > > 
> > >     if (buffers & PIPE_CLEAR_STENCIL) {
> > > @@ -2493,6 +2495,9 @@ void vrend_clear(struct vrend_context *ctx,
> > >        glClearStencil(stencil);
> > >     }
> > > 
> > > +   if (ctx->sub->hw_rs_state.rasterizer_discard)
> > > +       glDisable(GL_RASTERIZER_DISCARD);
> > > +
> > >     if (buffers & PIPE_CLEAR_COLOR) {
> > >        uint32_t mask = 0;
> > >        int i;
> > > @@ -2526,9 +2531,19 @@ void vrend_clear(struct vrend_context
> > > *ctx,
> > >     if (bits)
> > >        glClear(bits);
> > > 
> > > -   if (buffers & PIPE_CLEAR_DEPTH)
> > > +   /* Is it really necessary to restore the old states? The only
> 
> reason we
> > > +    * get here is because the guest cleared all those states but
> 
> gallium
> > > +    * didn't forward them before calling the clear command
> > > +    */
> > > +   if (ctx->sub->hw_rs_state.rasterizer_discard)
> > > +       glEnable(GL_RASTERIZER_DISCARD);
> > > +
> > > +   if (buffers & PIPE_CLEAR_DEPTH) {
> > >        if (!ctx->sub->dsa_state.depth.writemask)
> > >           glDepthMask(GL_FALSE);
> > > +      if (!ctx->sub->hw_rs_state.depth_clip)
> > > +          glEnable(GL_DEPTH_CLAMP);
> > > +   }
> > > 
> > >     /* Restore previous stencil buffer write masks for both front
> > > and
> 
> back faces */
> > >     if (buffers & PIPE_CLEAR_STENCIL) {
> > > --
> > > 2.16.1
> > > 
> > > _______________________________________________
> > > virglrenderer-devel mailing list
> > > virglrenderer-devel at lists.freedesktop.org
> > > https://lists.freedesktop.org/mailman/listinfo/virglrenderer-deve
> > > l
> 
> _______________________________________________
> virglrenderer-devel mailing list
> virglrenderer-devel at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/virglrenderer-devel