https://gcc.gnu.org/bugzilla/show_bug.cgi?id=57534
--- Comment #18 from Jeffrey A. Law <law at redhat dot com> --- A couple more notes. It could also well be the case that reassociating in a way that encourages lea could be good for x86, but bad for other targets. I also suspect this is closely related to other BZs in the database where we've regressed due to poor addressing mode selections. If you're able to make progress here I'll dig them out and we can cross-check your patch against those other bugs.
