Hi! On Fri, Mar 15, 2019 at 04:25:01PM -0400, Vladimir Makarov wrote: > On 2019-03-15 2:30 p.m., Segher Boessenkool wrote: > >PR89721 shows LRA treating an unspec_volatile's result as invariant, > >which of course isn't correct. This patch fixes it. > Segher, thank you for fixing this. The patch is ok to commit.
Thanks, done. Is this okay for backports to 8 and 7 as well? After a while of course. > >Question. Is side_effects_p the correct check? Or do we want to > >allow PRE_INC etc. here? What about CALL? > > > No, we don't want INC/DEC. Inheritance helps to reuse some reloaded > values which are expensive for calculation. So if we have two INCs, we > will have only one, which is wrong. The same is for calls. And we could handle it, but that requires code that isn't there yet, and is it worth it anyway. Okay :-) Segher
