[Mesa-dev] [PATCH] mesa: error out in indirect draw when vertex bindings mismatch
Fredrik Höglund
fredrik at kde.org
Tue Nov 17 15:31:59 PST 2015
On Tuesday 17 November 2015, Tapani Pälli wrote:
>
> On 11/16/2015 08:55 AM, Tapani Pälli wrote:
> >
> >
> > On 11/13/2015 07:18 PM, Fredrik Höglund wrote:
> >> On Friday 13 November 2015, Tapani Pälli wrote:
> >>> Patch adds additional mask for tracking which vertex buffer bindings
> >>> are set. This array can be directly compared to which vertex arrays
> >>> are enabled and should match when drawing.
> >>>
> >>> Fixes following CTS tests:
> >>>
> >>> ES31-CTS.draw_indirect.negative-noVBO-arrays
> >>> ES31-CTS.draw_indirect.negative-noVBO-elements
> >>>
> >>> Signed-off-by: Tapani Pälli <tapani.palli at intel.com>
> >>> ---
> >>> src/mesa/main/api_validate.c | 13 +++++++++++++
> >>> src/mesa/main/mtypes.h | 3 +++
> >>> src/mesa/main/varray.c | 5 +++++
> >>> 3 files changed, 21 insertions(+)
> >>>
> >>> diff --git a/src/mesa/main/api_validate.c b/src/mesa/main/api_validate.c
> >>> index a490189..e82e89a 100644
> >>> --- a/src/mesa/main/api_validate.c
> >>> +++ b/src/mesa/main/api_validate.c
> >>> @@ -710,6 +710,19 @@ valid_draw_indirect(struct gl_context *ctx,
> >>> return GL_FALSE;
> >>> }
> >>>
> >>> + /* From OpenGL ES 3.1 spec. section 10.5:
> >>> + * "An INVALID_OPERATION error is generated if zero is bound to
> >>> + * VERTEX_ARRAY_BINDING, DRAW_INDIRECT_BUFFER or to any enabled
> >>> + * vertex array."
> >>> + *
> >>> + * Here we check that vertex buffer bindings match with enabled
> >>> + * vertex arrays.
> >>> + */
> >>> + if (ctx->Array.VAO->_Enabled != ctx->Array.VAO->VertexBindingMask) {
> >>
> >> This test only works when the enabled vertex arrays are associated with
> >> their default vertex buffer binding points.
> >
> > Could you open up this more, is there some existing test or app that
> > would do this? Would be great for testing purposes, all the indirect
> > draw rendering CTS tests pass with this change.
>
> Sorry, the question does not make sense. What I meant is that do you
> know some app that would fail this test to help debugging/fixing the issue?
No, but the following example should trigger the problem:
/* Enable arrays 0 and 1 */
glEnableVertexAttribArray(0);
glEnableVertexAttribArray(1);
/* Make both arrays use VBO binding point #0 */
glVertexAttribBinding(0, 0);
glVertexAttribBinding(1, 0);
/* Bind a buffer object to VBO binding point #0 */
glBindVertexBuffer(0, ...);
/* Bind a draw indirect buffer */
glBindBuffer(GL_DRAW_INDIRECT_BUFFER, ...);
/* This call will now generate an INVALID_OPERATION error since
* no buffer is bound to VBO binding point #1, even though none
* of the enabled arrays use it.
*/
glDrawArraysIndirect(...);
> >
> >>> + _mesa_error(ctx, GL_INVALID_OPERATION, "%s(No VBO bound)", name);
> >>> + return GL_FALSE;
> >>> + }
> >>> +
> >>> if (!_mesa_valid_prim_mode(ctx, mode, name))
> >>> return GL_FALSE;
> >>>
> >>> diff --git a/src/mesa/main/mtypes.h b/src/mesa/main/mtypes.h
> >>> index 4efdf1e..6c6187f 100644
> >>> --- a/src/mesa/main/mtypes.h
> >>> +++ b/src/mesa/main/mtypes.h
> >>> @@ -1419,6 +1419,9 @@ struct gl_vertex_array_object
> >>> /** Vertex buffer bindings */
> >>> struct gl_vertex_buffer_binding VertexBinding[VERT_ATTRIB_MAX];
> >>>
> >>> + /** Mask indicating which binding points are set. */
> >>> + GLbitfield64 VertexBindingMask;
> >>> +
> >>> /** Mask of VERT_BIT_* values indicating which arrays are
> >>> enabled */
> >>> GLbitfield64 _Enabled;
> >>>
> >>> diff --git a/src/mesa/main/varray.c b/src/mesa/main/varray.c
> >>> index 887d0c0..0a94c5a 100644
> >>> --- a/src/mesa/main/varray.c
> >>> +++ b/src/mesa/main/varray.c
> >>> @@ -174,6 +174,11 @@ bind_vertex_buffer(struct gl_context *ctx,
> >>> binding->Offset = offset;
> >>> binding->Stride = stride;
> >>>
> >>> + if (vbo == ctx->Shared->NullBufferObj)
> >>> + vao->VertexBindingMask &= ~VERT_BIT(index);
> >>> + else
> >>> + vao->VertexBindingMask |= VERT_BIT(index);
> >>> +
> >>> vao->NewArrays |= binding->_BoundArrays;
> >>> }
> >>> }
> >>>
> >>
> > _______________________________________________
> > mesa-dev mailing list
> > mesa-dev at lists.freedesktop.org
> > http://lists.freedesktop.org/mailman/listinfo/mesa-dev
>
More information about the mesa-dev
mailing list