On Sat, Oct 18, 2025 at 1:02 PM Andrew Pinski <[email protected]> wrote: > > On Sat, Oct 18, 2025 at 10:33 AM Josef Melcr <[email protected]> wrote: > > > > Hi, > > this patch series removes the requirement for LTO when utilizing > > callback edges pointing to callbacks of builtin functions. It also > > improves the testability of this feature, as the testsuite practically > > wasn't being run on them without -flto. > > > > Removing this requirement uncovers several bugs present in the original > > callback edge patch, as the edges are now being tested across the entire > > libgomp testsuite. These bugs cause around 60 regressions. They are > > addressed and fixed in the first 3 patches, with the final patch > > implementing the title feature. Each patch is only several lines long, > > but since they all address a different bug, I decided to keep them separate. > > > > I bootstrapped and regtested the patches all at once on x86-64_linux. > > They fix the regressions mentioned earlier and introduce no new regressions. > > > > OK for master? > > This will fully fix https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80161 I > think. > Currently with x86_64 specific testcase in comment #0 and removing the > const for scale (so it is not done in the front-end), only `-O2 -flto` > works. From my understanding this patch series is designed to remove > the requirement of `-flto`, if so then please add the testcase (if it > works) and add `PR ipa/80161` to the changelog entry on when it gets > fix.
Looks like https://gcc.gnu.org/bugzilla/show_bug.cgi?id=49575 too. Thanks, Andrew > > Also Good work on this, it has definitely been a long time coming and > very much welcome improvement to GCC. > > Thanks, > Andrew > > > > > Best regards, > > Josef
