[Mesa-dev] [PATCH v2 22/27] i965/miptree: Allocate mcs_buf for an image's CCS

[Chad Versace]

On Wed 12 Jul 2017, Jason Ekstrand wrote:
> On Wed, Jul 12, 2017 at 11:05 AM, Chad Versace <[1]chadversary at chromium.org>
> wrote:
> 
>     On Thu 29 Jun 2017, Jason Ekstrand wrote:


>     Why require that aux_offset > 0? Why reject images where the aux surface
>     precedes the primary surface? This rejection seems arbitrary.
> 
> 
> All throughout these patches image->aux_offset == 0 is used for "no aux".  As
> you pointed out earlier, that's a bit on the bogus side but also always true. 
> For scanout, the hardware requires the aux buffer to be placed after the main
> surface so it's not *that* bogus.

I suspected that would be the answer.

>     The else-if-if chain here is ugly.
> 
> 
> Yes it is.  But I did it because it made more logical sense to me.  "if we have
> a modifier, follow it for aux else follow aux_usage."  However, the only
> aux_usage that will ever show up here is CCS_E.
>  
> 
>     The inner if's condition should replace the
>     outer else-if's condition.
> 
> 
> Here's a better idea.  How about I write an intel_miptree_alloc_aux helper
> function and call that for aux_usage != NONE?  It can handle CCS, MCS, and HiZ
> so that's all together in one place.

Nice. A new function seems like the right way to do it.