[Mesa-dev] [PATCH 1/8] mesa: fix error handling for glBegin

[Yuanhan Liu]

On Tue, Sep 20, 2011 at 08:51:26AM -0600, Brian Paul wrote:
> On 09/20/2011 08:36 AM, Brian Paul wrote:
> >On 09/19/2011 08:39 PM, Yuanhan Liu wrote:
> >>On Mon, Sep 19, 2011 at 08:09:09AM -0600, Brian Paul wrote:
> >>>On 09/19/2011 01:02 AM, Yuanhan Liu wrote:
> >>>>According to opengl spec, trigger GL_INVALID_ENUM error if mode is
> >>>>not
> >>>>one of those valid primitive mode while calling glBegin.
> >>>>
> >>>>Signed-off-by: Yuanhan Liu<yuanhan.liu at linux.intel.com>
> >>>>---
> >>>>src/mesa/vbo/vbo_exec_api.c | 24 ++++++++++++++++++++++++
> >>>>1 files changed, 24 insertions(+), 0 deletions(-)
> >>>>
> >>>>diff --git a/src/mesa/vbo/vbo_exec_api.c
> >>>>b/src/mesa/vbo/vbo_exec_api.c
> >>>>index cad7c46..72f9661 100644
> >>>>--- a/src/mesa/vbo/vbo_exec_api.c
> >>>>+++ b/src/mesa/vbo/vbo_exec_api.c
> >>>>@@ -552,6 +552,25 @@ static void GLAPIENTRY vbo_exec_EvalPoint2(
> >>>>GLint i, GLint j )
> >>>>#endif /* FEATURE_evaluators */
> >>>>
> >>>>
> >>>>+static GLboolean _mesa_valid_prim_mode(GLenum mode)
> >>>>+{
> >>>>+ switch(mode) {
> >>>>+ case GL_POINTS:
> >>>>+ case GL_LINES:
> >>>>+ case GL_LINE_LOOP:
> >>>>+ case GL_LINE_STRIP:
> >>>>+ case GL_TRIANGLES:
> >>>>+ case GL_TRIANGLE_STRIP:
> >>>>+ case GL_TRIANGLE_FAN:
> >>>>+ case GL_QUADS:
> >>>>+ case GL_QUAD_STRIP:
> >>>>+ case GL_POLYGON:
> >>>>+ return GL_TRUE;
> >>>>+ default:
> >>>>+ return GL_FALSE;
> >>>>+ }
> >>>>+}
> >>>>+
> >>>>/**
> >>>>* Called via glBegin.
> >>>>*/
> >>>>@@ -570,6 +589,11 @@ static void GLAPIENTRY vbo_exec_Begin( GLenum
> >>>>mode )
> >>>>return;
> >>>>}
> >>>>
> >>>>+ if (!_mesa_valid_prim_mode(mode)) {
> >>>>+ _mesa_error(ctx, GL_INVALID_ENUM, "glBegin");
> >>>>+ return;
> >>>>+ }
> >>>>+
> >>>>if (!_mesa_valid_to_render(ctx, "glBegin")) {
> >>>>return;
> >>>>}
> >>>
> >>>I think we're actually validating the prim mode elsewhere, but it
> >>>looks like we're not generating GL_INVALID_ENUM as we should. Could
> >>>you investigate this a bit further?
> >>
> >>From my investigation, nope. I didn't find it does check both at the
> >>glBegin call
> >>time and the draw execution time.
> >
> >Hmmm, you're right. I couldn't find the validation code either (except
> >on the display list path).
> >
> >I'll apply your original patch soon.
> 
> Scratch that.  I've got a new patch that also properly checks the
> mode for glDrawArrays, glDrawElements, etc.

Yeah, I should check those too.  Thanks.