On Fri, Mar 3, 2017 at 12:30 PM, Jakub Jelinek <ja...@redhat.com> wrote:
>> Regarding the implementation, I'd rather see another flag for >> i386-builtin.def entries, where we can mark if matching memory operand >> is allowed. The expander can then look at the flag, and do some >> special magic. > > I think maintaining such a flag would be a nightmare, I think more insns > will need it in the future and it will be hard to remember. Maybe it is > better > not to do anything about those (i.e. apply my patch as is), after all -O0 > is not optimized code. Yes, I agree. Uros.
