https://sourceware.org/bugzilla/show_bug.cgi?id=27491
--- Comment #5 from Fangrui Song <i at maskray dot me> --- (In reply to H.J. Lu from comment #4) > (In reply to Fangrui Song from comment #3) > > > > I don't know what happened to the 2010-01 patch. It likely did not let > > __start_ reference retain magic sections in other translation unit (in > > metadata section usage __start_/__stop_ references are always in a separate > > runtime file), so "__start_ retaining all magic sections" was not effective > > before 2015-10. Why do we need new start/stop symbols? The orphan > > __start_/__stop_ usage worked before 2015-10 and will work with -z > > start-stop-gc. > > You got it wrong. --gc-section never worked correctly before all these > fixed applied. > > (I think of the possibility enabling -z start-stop-gc by default because the > > rule was essentially invalid before 2015-10.) > > > > LLVM requires a different __start_/__stop_ semantic, which isn't wrong > by itself. But it shouldn't hijack the current one. On Mach-O, ld64 define section$start$__DATA$__data and section$end$__DATA$__data which are similar to the ELF encapsulation symbols __start_/__stop_. ld64's behavior has always been similar to ld.lld -z start-stop-gc. Wow, really unfortunate that gold broke the nice GC semantics in 2010-01 and GNU ld in 2015-10. It'd be really nice to fix the issue as this made x86-64 -z start-stop-gc not usable for -pie / -shared. All of Clang PGO/SanitizerCoverage/source based Coverage documentation have noted that --gc-sections may not work well <--- hope they can stop hinting the users this way one day. -- You are receiving this mail because: You are on the CC list for the bug.
