[Mesa-dev] [PATCH 1/1] main: validate the type correctly against the OpenGL ES spec.

[Brian Paul]

On 09/11/2012 10:21 AM, Oliver McFadden wrote:
> On Tue, Sep 11, 2012 at 07:54:36AM -0600, Brian Paul wrote:
>> On 09/11/2012 06:09 AM, Oliver McFadden wrote:
>>> Previously Mesa would validate UNSIGNED_INT successfully in violation of
>>> the OpenGL(R) ES 1.0 and 2.0 Specification.
>>>
>>> http://www.khronos.org/registry/gles/specs/1.1/es_full_spec_1.1.12.pdf
>>> http://www.khronos.org/registry/gles/specs/2.0/es_full_spec_2.0.25.pdf
>>> http://www.khronos.org/registry/gles/specs/3.0/es_spec_3.0.0.pdf
>>>
>>> Signed-off-by: Oliver McFadden<oliver.mcfadden at linux.intel.com>
>>> ---
>>>    src/mesa/main/api_validate.c |   22 ++++++++++++++++++++++
>>>    1 files changed, 22 insertions(+), 0 deletions(-)
>>>
>>> diff --git a/src/mesa/main/api_validate.c b/src/mesa/main/api_validate.c
>>> index d0d2ca4..698394c 100644
>>> --- a/src/mesa/main/api_validate.c
>>> +++ b/src/mesa/main/api_validate.c
>>> @@ -296,10 +296,32 @@ _mesa_valid_prim_mode(struct gl_context *ctx, GLenum mode, const char *name)
>>>    static bool
>>>    valid_elements_type(struct gl_context *ctx, GLenum type, const char *name)
>>>    {
>>> +  /*
>>> +   * Page 22 of the OpenGL(R) ES Version 1.0 Specification (PDF)
>>> +   * and
>>> +   * Page 21 of the OpenGL(R) ES Version 2.0 Specification (PDF) says:
>>> +   *
>>> +   * `type' must be one of UNSIGNED_BYTE or UNSIGNED_SHORT, indicating that the
>>> +   * values in indices are indices of GL type ubyte or ushort, respectively.
>>> +   *
>>> +   * Page 29 of the OpenGL(R) ES Version 3.0 Specification (PDF) says:
>>> +   *
>>> +   * `type' must be one of UNSIGNED_BYTE, UNSIGNED_SHORT, or UNSIGNED_INT,
>>> +   * indicating that the index values are of GL type ubyte, ushort, or uint
>>> +   * respectively.
>>> +   *
>>> +   */
>>>       switch (type) {
>>>       case GL_UNSIGNED_BYTE:
>>>       case GL_UNSIGNED_SHORT:
>>> +      return true;
>>>       case GL_UNSIGNED_INT:
>>> +      if (ctx->API == API_OPENGLES ||
>>> +          (ctx->API == API_OPENGLES2&&   ctx->Version<   30)) {
>>> +         _mesa_error(ctx, GL_INVALID_ENUM, "%s(type = %s)", name,
>>> +                     _mesa_lookup_enum_by_nr(type));
>>> +         return false;
>>> +      }
>>>          return true;
>>>
>>>       default:
>>
>> I suspect Ian's covered this in his ES overhaul work, but it looks
>> good to me.
>
> You're giving the OK to push to master? I don't want to conflict with
> existing work...

Ian hasn't posted all his ES-cleanup patches yet (though maybe they're 
in a public tree of his) so I don't know if there's a conflict.

If you can hold off for a bit, maybe Ian can review and comment.

-Brian