On Mon, Mar 4, 2013 at 11:53 AM, Roland Scheidegger <srol...@vmware.com> wrote:
> Am 02.03.2013 04:26, schrieb Dave Airlie:
>> On Sat, Mar 2, 2013 at 12:02 PM, Roland Scheidegger <srol...@vmware.com>
>> wrote:
>>> Am 02.03.2013 01:36, schrieb Brian Paul:
>>>> ---
>>>> src/mesa/state_tracker/st_glsl_to_tgsi.cpp | 3 +++
>>>> 1 files changed, 3 insertions(+), 0 deletions(-)
>>>>
>>>> diff --git a/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
>>>> b/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
>>>> index 8e3e3b8..c41b583 100644
>>>> --- a/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
>>>> +++ b/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
>>>> @@ -2746,6 +2746,9 @@ glsl_to_tgsi_visitor::visit(ir_texture *ir)
>>>> offset = this->result;
>>>> }
>>>> break;
>>>> + case ir_txf_ms:
>>>> + assert(!"Unexpected ir_txf_ms opcode");
>>>> + break;
>>>> }
>>>>
>>>> if (ir->projector) {
>>>>
>>>
>>> Series looks good to me. I guess we need a new opcode like
>>> (TGSI_OPCODE_TXF_MS?), unless we switch everything over and only use the
>>> new sample style opcodes which already have that (SAMPLE_I_MS) :-).
>>
>> I'm not really sure we do, from what I can see TXF is sufficent since
>> LOD is bogus with MS textures so you can reuse the slot. r600 hw at
>> least only has one LD instruction from what I can see.
> Hmm yes but it doesn't sound like a very clean solution, to reuse the
> opcode for something a bit different. Though granted since with the
> combined texture/sampler state you always know the target (unlike with
> the sample_i opcodes) it should be always possible to distinguish
> txf-with-lod from txf-with-sample so it might indeed be ok.
There is no need to look at the texture or sampler state. TGSI texture
instructions specify their own texture target (TGSI_TEXTURE_*). For
example:
TXF OUT[0], IN[0], 2D_MSAA
Marek
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
