On Mon, Oct 17, 2016 at 4:23 PM, Ilia Mirkin <imir...@alum.mit.edu> wrote:
> On Mon, Oct 17, 2016 at 10:20 AM, Marek Olšák <mar...@gmail.com> wrote:
>> I'm adding back this:
>>
>> options->EmitNoMainReturn =
>> !screen->get_shader_param(screen, sh, PIPE_SHADER_CAP_SUBROUTINES);
>>
>> And:
>>
>> diff --git a/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
>> b/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
>> index 293654c..5f28d07 100644
>> --- a/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
>> +++ b/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
>> @@ -4176,7 +4176,9 @@ glsl_to_tgsi_visitor::visit(ir_texture *ir)
>> void
>> glsl_to_tgsi_visitor::visit(ir_return *ir)
>> {
>> - assert(0);
>> + assert(!ir->get_value());
>> +
>> + emit_asm(ir, TGSI_OPCODE_RET);
>> }
>>
>> void
>> @@ -4416,7 +4418,8 @@ glsl_to_tgsi_visitor::simplify_cmp(void)
>> inst->dst[1].reladdr || inst->dst[1].reladdr2 ||
>> tgsi_get_opcode_info(inst->op)->is_branch ||
>> inst->op == TGSI_OPCODE_CONT ||
>> - inst->op == TGSI_OPCODE_END) {
>> + inst->op == TGSI_OPCODE_END ||
>> + inst->op == TGSI_OPCODE_RET) {
>> break;
>> }
>>
>>
>> Hopefully that should restore the RET support.
>
> Thanks. I think that should be enough. I won't be able to test
> anything today, in all likelihood, but probably could tomorrow.
> Perhaps Samuel will have time to play with it.
No worries. I've tested it on softpipe.
Marek
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
