DavidSpickett added a comment. > Yeah, this is a very aarch64 specific thing in the generic symbolizer, but > it's not an especially large function so I didn't feel too bad about it. If > we started needing to handle multi-instruction sequences like this in the > symbolizer (where we save state from previous instructions to determine a > result), then this definitely would not scale well.
Yeah I mainly mentioned it because of vague memories of some other function that grew a whole bunch of these. Fine for now. (And in general I like the change, we get this sort of request for the disassembly tools a lot but haven't really thought about it for lldb) > I've written it so that I only recognize the pattern with ADRP is immediately > followed by an ADD, that was maybe a choice I could have gone either way on. That bit makes sense to me. I suppose instruction scheduling might separate them but this probably hits 99% of them. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D107213/new/ https://reviews.llvm.org/D107213 _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
