[PATCH] D43241: [ELF] - Do not crash with --emit-relocs and --icf=all together.
Rafael Avila de Espindola via llvm-commits
llvm-commits at lists.llvm.org
Fri Feb 16 10:51:36 PST 2018
I would suggest first moving the loop to the caller as Rui
requested. With that we have test that we don't crash and can go take a
look at SHF_LINK_ORDER.
Looking at your patch for PR36370 next.
George Rimar <grimar at accesssoftek.com> writes:
>>An interesting case that might clarify where this code should be is what
>>ICF should do about SHF_LINK_ORDER.
>>What happens if a section has SHF_LINK_ORDER and another otherwise
>>identical one doesn't?
>>Should ICF avoid merging the sections? Should it merge and keep the
>>SHF_LINK_ORDER linked section?
> Yep. Patch for improving support of SHF_LINK_ORDER sections was D38180.
> If you want, I can resurrect and rebase it too.
> (A bit busy with PR36370 atm, but should finish it soon, probably today.
> After that was going to reimplement this patch).
More information about the llvm-commits