[Mesa-dev] [PATCH] glx: Fix image size computation for EXT_texture_integer
Ian Romanick
idr at freedesktop.org
Fri Dec 12 13:25:03 PST 2014
Looking at src/glx/glxextensions.c (the known_glx_extensions array
specifically), I don't think we support GL_EXT_texture_integer on the
client side for indirect rendering. So... do we actually need this patch?
On 12/12/2014 10:34 AM, Adam Jackson wrote:
> Without this, this extension basically can't work in indirect contexts,
> TexImage2D will compute the image size as 0 and we'll send no image data
> to the server. Old servers would do undefined things upon receipt of
> such a teximage; new servers with the fix for CVE-2014-8098 will correctly
> reject the request without this patch.
>
> This is the client-side equivalent of:
>
> http://cgit.freedesktop.org/xorg/xserver/commit/?id=13d36923e0ddb077f4854e354c3d5c80590b5d9d
>
> Signed-off-by: Adam Jackson <ajax at redhat.com>
> ---
> src/glx/compsize.c | 10 ++++++++++
> 1 file changed, 10 insertions(+)
>
> diff --git a/src/glx/compsize.c b/src/glx/compsize.c
> index 99c7763..8055919 100644
> --- a/src/glx/compsize.c
> +++ b/src/glx/compsize.c
> @@ -65,6 +65,8 @@ __glElementsPerGroup(GLenum format, GLenum type)
> switch (format) {
> case GL_RGB:
> case GL_BGR:
> + case GL_RGB_INTEGER_EXT:
> + case GL_BGR_INTEGER_EXT:
> return 3;
> case GL_RG:
> case GL_422_EXT:
> @@ -74,10 +76,13 @@ __glElementsPerGroup(GLenum format, GLenum type)
> case GL_DEPTH_STENCIL_NV:
> case GL_YCBCR_422_APPLE:
> case GL_LUMINANCE_ALPHA:
> + case GL_LUMINANCE_ALPHA_INTEGER_EXT:
> return 2;
> case GL_RGBA:
> case GL_BGRA:
> case GL_ABGR_EXT:
> + case GL_RGBA_INTEGER_EXT:
> + case GL_BGRA_INTEGER_EXT:
> return 4;
> case GL_COLOR_INDEX:
> case GL_STENCIL_INDEX:
> @@ -88,6 +93,11 @@ __glElementsPerGroup(GLenum format, GLenum type)
> case GL_ALPHA:
> case GL_LUMINANCE:
> case GL_INTENSITY:
> + case GL_RED_INTEGER_EXT:
> + case GL_GREEN_INTEGER_EXT:
> + case GL_BLUE_INTEGER_EXT:
> + case GL_ALPHA_INTEGER_EXT:
> + case GL_LUMINANCE_INTEGER_EXT:
> return 1;
> default:
> return 0;
>
More information about the mesa-dev
mailing list