[Mesa-dev] [PATCH] intel: Fix intel_texsubimage_tiled_memcpy to skip GL_EXT_unpack_subimage case

[Ian Romanick]

On 10/05/2012 03:38 AM, Neil Roberts wrote:
> Rob Bradford <rob at robster.org.uk> writes:
>
>> +       packing->RowLength != width)
>
> I wonder if this is a bit overzealous and we should change it to accept
> packing->RowLength==0 as well. Setting the row length to zero is
> equivalent to setting it to the width of the data. It could be that
> Chrome leaves it at zero and in that case we'd have broken the
> optimisation.

That is correct.  Almost every app leaves GL_PACK_ROW_LENGTH as 0.  I 
think we should have a helper function that checks that the pack mode is 
a no-op.  It should get passed the gl_pixelstore_attrib, width, height, 
and dim.

It looks like the check also misses the GL_PACK_SWAP_BYTES == GL_TRUE case.

> Regards,
> - Neil
> _______________________________________________
> mesa-dev mailing list
> mesa-dev at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/mesa-dev