[PATCH] D59553: [WIP][LLD][ELF][DebugInfo] Use predefined value to mark debug address ranges pointing to deleted code.

David Blaikie via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Wed Jun 3 14:56:12 PDT 2020


dblaikie added a comment.

In D59553#2070586 <https://reviews.llvm.org/D59553#2070586>, @jhenderson wrote:

> In D59553#2067279 <https://reviews.llvm.org/D59553#2067279>, @dblaikie wrote:
>
> > (again, thanks always for your ongoing patience on this complicated/long-term issue)
> >
> > The flag name for disabling this seems a little confusing to me - since there's no option to really "disable" this sort of marking, but different choices in what value is used. So maybe something like "--dead-reloc-zero-addend" or something (don't go changing the patch on my bad suggestion here, but perhaps @MaskRay & others on this thread might have better ideas for names based on linker flag name history). I wouldn't worry about debating/changing this urgently - just a thought to ponder while we continue with the rest of the discussion.
>
>
> Perhaps if we're adding the option, the option value should be configurable, so something like `--dead-reloc-addend=<value>` which specifies the value instead of just assuming zero. You could even go further e.g. `--dead-reloc-addend=<section>=<value>` (your syntax may vary) to allow configuring it per section. Thus, the default would be something like `--dead-reloc-addend=.debug_ranges=0xfffffffffffffffe --dead-reloc-addend=.debug_info=0xffffffffffffffff` and so on. This would allow users to configure things according to the needs of their individual system.


Doable if necessary (will you likely have a need for this in PS4? that'd make it easier to justify adding the feature) - but I'd try to avoid adding such flexibility for as long as possible to avoid proliferation of even more variants of debug info consumers might then have to deal with.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D59553/new/

https://reviews.llvm.org/D59553





More information about the llvm-commits mailing list