[Mesa-dev] [PATCH 5/7] i965: relax brw_texture_offset assert

[Kenneth Graunke]

On 10/08/2013 12:31 PM, Ian Romanick wrote:
> On 10/08/2013 02:34 AM, Chris Forbes wrote:
>> Some texturing ops are about to have nonconstant offset support; the
>> offset in the header in these cases should be zero.
>>
>> Signed-off-by: Chris Forbes <chrisf at ijw.co.nz>
>> ---
>>  src/mesa/drivers/dri/i965/brw_shader.cpp | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/src/mesa/drivers/dri/i965/brw_shader.cpp b/src/mesa/drivers/dri/i965/brw_shader.cpp
>> index 6b37f58..df7fa55 100644
>> --- a/src/mesa/drivers/dri/i965/brw_shader.cpp
>> +++ b/src/mesa/drivers/dri/i965/brw_shader.cpp
>> @@ -370,7 +370,7 @@ brw_math_function(enum opcode op)
>>  uint32_t
>>  brw_texture_offset(ir_constant *offset)
>>  {
>> -   assert(offset != NULL);
> 
> I think we should leave this assertion, but amend it to be
> 
>    /* If the driver does not support GL_ARB_gpu_shader5, the offset
>     * must be constant.
>     */
>    assert(offset != NULL || ctx->Extensions.ARB_gpu_shader5);

Once we actually enable ARB_gpu_shader5, this won't be useful on Gen7+.

I'm not sure if Chris is hoping to land his Sandybridge support for
ARB_texture_gather or not (it has mad workarounds, but he's already done
them...).  If so, then your suggestion seems reasonable.

I guess I have a mild preference for following Ian's suggestion, in case
we ever do add Sandybridge support.  But I'll leave the decision up to
you, Chris.

>> +   if (!offset) return 0;  /* nonconstant offset; caller will handle it. */
>>  
>>     signed char offsets[3];
>>     for (unsigned i = 0; i < offset->type->vector_elements; i++)
>>
> 
> _______________________________________________
> mesa-dev mailing list
> mesa-dev at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/mesa-dev
>