On Tue, Aug 27, 2019 at 2:14 PM Richard Biener wrote:
>
>
> With forcing STV for all chains I run into the same issue as I
> fixed for the analysis phase which is looping over all defs
> of a pseudo rather just those interesting. The following fixes
> this by recording insn/pseudo pairs in mark_d
With forcing STV for all chains I run into the same issue as I
fixed for the analysis phase which is looping over all defs
of a pseudo rather just those interesting. The following fixes
this by recording insn/pseudo pairs in mark_dual_mode_def;
well, not actually pairs but tracking insns in addi
