https://bugs.kde.org/show_bug.cgi?id=338252
Philippe Waroquiers <philippe.waroqui...@skynet.be> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |RESOLVED Resolution|--- |FIXED --- Comment #21 from Philippe Waroquiers <philippe.waroqui...@skynet.be> --- (In reply to Дилян Палаузов from comment #20) Thanks for looking at this further. > With the change below m_libcsetjmp.c links on amd64 correctly with LTO. Is > the change equivalent (apart from cheating gcc restoring rax in in > VG_MINIMAL_LONGJMP)? Is volatile after __asm__ necessary? I have tested the patch on amd64, and the regression tests are working. So, what you propose might be a way to go. However, I am not at ease to change the asm code as this is tricky, and must be done and tested for all platforms. So, I have pushed the patch which implements --enable-lto=yes|no as ab773096df7aaaf46e8883af5ed4690f4d4499af. If deemed desirable, another bug/wish could be opened to support a better configuration mechanism (i.e. use CFLAGS, instead of having predefined logic to setup CFLAGS and similar in configure). -- You are receiving this mail because: You are watching all bug changes.