[PATCH] D82828: [ELF] Don't resolve a relocation in .debug_line referencing an ICF folded symbol to the tombstone value

[Fangrui Song via Phabricator via llvm-commits]

MaskRay added a comment.

In D82828#2126155 <https://reviews.llvm.org/D82828#2126155>, @dblaikie wrote:

> In D82828#2126075 <https://reviews.llvm.org/D82828#2126075>, @MaskRay wrote:
>
> > @dblaikie Dave, are you ok with this change as well?
>
>
> Sure! Not the biggest fan, personally, of these layered special cases - but equally if Sony's been shipping with it/demonstrates practical usefulness, etc, I won't stand in the way & can appreciate the practical benefits.


For COMDAT, there is a prevailing (canonical) group. Usually the file:line is the same. For ICF, every function should be equal. If a function can't be set breakpoint just because the linker does not select it as the canonical one, the lost debugging experience looks genuine to me. Just committed https://sourceware.org/git/?p=binutils-gdb.git;a=commit;h=a8caed5d7faa639a1e6769eba551d15d8ddd9510 (woohoo! my first gdb patch) before it, the experience was even less pleasant because gdb would keep warning when you hit `continue`

  Warning:
  Cannot insert breakpoint 1.
  Cannot access memory at address 0x6


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D82828/new/

https://reviews.llvm.org/D82828