[PATCH] D67390: [LLDB][ELF] Load both, .symtab and .dynsym sections

Konrad Wilhelm Kleine via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Wed Oct 2 08:00:46 PDT 2019


kwk abandoned this revision.
kwk added a comment.

Here's the relevant transcript from #lldb at otfc for why this change is abandoned.

  [10/02/19 15:22:25] <jankratochvil> labath: Is it acceptable for you?  "BTW given how this unique-ness of symbols turns out to be non-trivial is it really needed?  Because for regular .symtab we can ignore .dynsym (as current LLDB does) and for .gnu_debugdata->.symtab we can just concatenate it with .dynsym and there will be no symbol duplicates anyway."
  [10/02/19 15:22:25] <jankratochvil> Otherwise it is either a big code complication or a big performance regression and it is not really needed for anything.
  [10/02/19 15:27:42] <labath> yeah, i suppose that would work
  [10/02/19 15:28:11] <labath> then you just need to make sure the merging does not kick in for non-gnu_debugdata sections
  [10/02/19 15:28:30] <labath> a simple check on the existing of this section might suffice
  [10/02/19 15:29:39] <labath> it seems to me that it would be useful to have information from both sources available, but it is true that nobody has really needed that so far
  [10/02/19 15:36:11] <jankratochvil> Thanks.
  [10/02/19 15:36:22] <jankratochvil> kkleine: ^^^ Let's call it a win-win. :-)
  [10/02/19 16:14:27] <kkleine> labath, jankratochvil I wonder if we the merging with .gnu_debugdata needs to be as "clever" as it is today and if it is enough to just add symbols without checking for uniqueness. That would simplify things even more.
  [10/02/19 16:16:32] <jankratochvil> I think just add it without checks.
  [10/02/19 16:16:53] <labath> works for me


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D67390/new/

https://reviews.llvm.org/D67390





More information about the llvm-commits mailing list