On 23 July 2012 09:15, Eric Anholt <span dir="ltr"><<a href="mailto:eric@anholt.net" target="_blank">eric@anholt.net</a>></span> wrote:<br><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">

<div><div>Chris Wilson <<a href="mailto:chris@chris-wilson.co.uk" target="_blank">chris@chris-wilson.co.uk</a>> writes:<br>
<br>
> On Sat, 21 Jul 2012 17:36:40 -0700, Chad Versace <<a href="mailto:chad.versace@linux.intel.com" target="_blank">chad.versace@linux.intel.com</a>> wrote:<br>
>> After commit "intel: Convert to using private depth/stencil buffers", we<br>
>> request from DRI2GetBuffersWithFormat only the front left and back left<br>
>> buffers. We no longer request depth and stencil buffers.<br>
>><br>
>> Assert that in intelAllocateBuffer and remove the related dead code.<br>
><br>
>> diff --git a/src/mesa/drivers/dri/intel/intel_screen.c b/src/mesa/drivers/dri/intel/intel_screen.c<br>
>> index 81953ce..2d46d67 100644<br>
>> --- a/src/mesa/drivers/dri/intel/intel_screen.c<br>
>> +++ b/src/mesa/drivers/dri/intel/intel_screen.c<br>
>> @@ -972,54 +972,6 @@ struct intel_buffer {<br>
>>     struct intel_region *region;<br>
><br>
>> -   ok = intel_get_dri_buffer_tiling(intelScreen, attachment, &tiling);<br>
><br>
>> +   /* The front and back buffers are color buffers, which are X tiled. */<br>
>>     intelBuffer->region = intel_region_alloc(intelScreen,<br>
>> -                                            tiling,<br>
>> -                                            region_cpp,<br>
>> -                                            region_width,<br>
>> -                                            region_height,<br>
>> +                                            I915_TILING_X,<br>
>> +                                            format / 8,<br>
>> +                                            width,<br>
>> +                                            height,<br>
>>                                              true);<br>
><br>
> I see no reason to bake in the assumption that the buffers are X-tiled.<br>
> Presumably Y-tiling will still be advantageous in many circumstances,<br>
> basically any surface which is not a candidate for pageflipping?<br>
<br>
</div></div>If there was some reason, it probably won't be based on just the<br>
attachment type, so this seems like a good change.<br></blockquote><div><br>Agreed.  Besides, the old code had a baked-in assumption of X-tiling for front buffers too, so there is no change.  This patch is:<br><br>Reviewed-by: Paul Berry <<a href="mailto:stereotype441@gmail.com" target="_blank">stereotype441@gmail.com</a>><br>

</div></div>