On Wed, 2019-06-26 at 14:21 +0200, Thomas Stüfe wrote: > > One curious question though about the original bug, was this really a > measurable performance issue? I would have thought that libjsig is > not invoked often enough for this to make a difference..
It's a matter of missed optimizations. We've got tooling in RHEL 8 which checks for native libraries which aren't being optimized. This was one of the flagged libs throughout all of OpenJDK. While it probably isn't needed for performance reasons, it's probly also just an oversight that opt flags didn't get passed properly for this library. It shouldn't do any harm. Thanks, Severin