On Mon, Apr 27, 2015 at 5:25 PM, Jason Ekstrand <ja...@jlekstrand.net> wrote:
> On Mon, Apr 27, 2015 at 1:35 PM, Connor Abbott <cwabbo...@gmail.com> wrote:
>> On Fri, Apr 24, 2015 at 7:32 PM, Jason Ekstrand <ja...@jlekstrand.net> wrote:
>>> +struct nir_if;
>>> +
>>> typedef struct nir_src {
>>> union {
>>> + nir_instr *parent_instr;
>>> + struct nir_if *parent_if;
>>> + };
>>
>> There's something I'm not quite understanding about this... how are we
>> supposed to know which of parent_instr and parent_if are valid? If I
>> walk over all the sources for a given SSA def or register, how am I
>> supposed to know if it's part of an if-condition or an instruction? I
>> would think that you would need a boolean here or have parent_instr
>> and parent_if not be in a union.
>
> We do the same thing we did with the sets before. We have separate
> uses and if_uses sets. If it's in if_uses, you use the if. If it's
> in uses, it's an instr. We could put something in the source but that
> seems like it'll make for even more mess.
> --Jason
D'oh... right. That makes sense.
Connor
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
