Amit Prakash Ambasta

Is this really a bug w/ LLVM? Compiling w/ LLVM succeeds ```# /usr/lib/llvm/17/bin/x86_64-pc-linux-gnu-clang-17 -std=gnu99 -U_FORTIFY_SOURCE -D_FORTIFY_SOURCE=2 -fstack-protector-strong -fstack-clash-protection -march=native -pipe -fno-stack-protector -fPIC -ffunction-sections -fdata-sections -fno-math-errno -pthread -pipe -O3 -fomit-frame-pointer -funwind-tables...

@rui314 Some additional guidance please, looking at https://github.com/llvm/llvm-project/blob/631e2911ea298bc12564df8acd16bba89790426a/lld/test/ELF/relocation-none.test#L38 ``` ## Both REL and RELA are supported. .rel.llvm.call-graph-profile uses REL even ## if the prevailing format is RELA. ``` This is...

Yes, but that would be unlikely. Even the kernel has patches to ignore modpost checks on call-graph-profile https://lore.kernel.org/lkml/CAK7LNARVi1sfBjv5a5OoQWPEeM-6bFuwPJE+i32NC=wdum-AKw@mail.gmail.com/t/#mcd82d7d511dcb4fe8939e21127fafda08f4f732e And here's the rationale behind it https://reviews.llvm.org/D104080

I genuinely do not understand any of this enough to comment. This issue was my first attempt at even trying to go through a compiler/linker source, so I'm not sure...