On 02/26/2016 01:01 PM, Segher Boessenkool wrote: >> How do you imagine the rs6000 change will regress codegen? > > Combine of sequences with double-length adds.
What sort of test case are you imagining here? The trivial tests I've looked at have all been optimal (before and after). > It failed more recently though, that's what r215429 is about. Oh, before > that it did test just the code, didn't use register_operand. *facepalm* Ah yes, exactly. I didn't think to annotate ca_operand for more recent changes. ;-) r~
