Neat -- I was about to mention the dead_control_flow_eliminate interaction, but you beat me to it.
On Tue, Dec 3, 2013 at 10:43 AM, Matt Turner <matts...@gmail.com> wrote: > And since this change makes ENDIF start blocks, I'll squash this > change in as well to make sure dead_control_flow_eliminate continues > working: > > diff --git a/src/mesa/drivers/dri/i965/brw_dead_control_flow.cpp > b/src/mesa/drivers/dri/i965/brw_dead_control_flow.cpp > index ad8ed82..63a3e5b 100644 > --- a/src/mesa/drivers/dri/i965/brw_dead_control_flow.cpp > +++ b/src/mesa/drivers/dri/i965/brw_dead_control_flow.cpp > @@ -45,10 +45,10 @@ dead_control_flow_eliminate(backend_visitor *v) > bblock_t *block = cfg.blocks[b]; > bool found = false; > > - /* ENDIF instructions, by definition, can only be found at the ends of > + /* ENDIF instructions, by definition, can only be found at the start of > * basic blocks. > */ > - backend_instruction *endif_inst = block->end; > + backend_instruction *endif_inst = block->start; > if (endif_inst->opcode != BRW_OPCODE_ENDIF) > continue; > _______________________________________________ > mesa-dev mailing list > mesa-dev@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/mesa-dev _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
