[Mesa-dev] [PATCH 2/2] meta: Handle bitmaps with alpha test enabled.

[Brian Paul]

On 11/03/2010 02:59 PM, Francisco Jerez wrote:
> Brian Paul<brianp at vmware.com>  writes:
>
>> On 11/03/2010 09:12 AM, Francisco Jerez wrote:
>>> Brian Paul<brianp at vmware.com>   writes:
>>>
>>>> On 11/02/2010 12:25 PM, Francisco Jerez wrote:
>>>>>
>>>>> We have several options.
>>>>>
>>>>> - Fall back to swrast in the floating point renderbuffer case. In other
>>>>>      cases the conversion to fixed point guarantees that EQUAL will behave
>>>>>      as expected. Simple and good enough ATM because all the users of
>>>>>      _mesa_meta_Bitmap don't support floating point color buffers (no-one
>>>>>      does). Some of them might at some point though.
>>>>>
>>>>> - Choose a background alpha value different to the current raster color
>>>>>      that is known to be represented exactly all the way down the pipeline
>>>>>      (e.g. 0.0/1.0), then use NOTEQUAL to cull them out.
>>>>>
>>>>> - Stick to GREATER/LESS for maximum robustness/paranoia.
>>>>>
>>>>> I'm fine with any of them. What do you suggest?
>>>>
>>>> Either of the later two.  Thanks.
>>>>
>>> OK, the following patch implements the second one.
>>>
>>
>> Looks good.  Thanks.
>>
>> Reviewed-by: Brian Paul<brianp at vmware.com>
>>
>>
>> Another thing: it looks like we don't really need to specify vertex
>> colors when drawing the bitmap quad.  Want to remove/fix that?
>>
> I don't expect immediate mode color specification to be much faster than
> the current solution, there'd be yet another layer in between and the
> constant color would still be copied multiple times on its way down.
>
> We could fix up the color array struct by hand and make it point
> directly to the current raster color (kind of a layering violation), or
> use flat-shaded quads (more state we'd have to track), but, AFAICT, this
> isn't so relevant performance-wise, right now the bottleneck is in the
> repeated texture re-uploads and DrawArrays calls.

I was thinking of using flat shading, but I guess it wouldn't buy us much.

To really speed up the meta bitmap code, a bitmap cache like in the 
gallium state tracker would be good.

-Brian