jasonmolenda accepted this revision. jasonmolenda added a comment. This revision is now accepted and ready to land.
Looks good to me. Yes, we need to change from a model of "the remote stub teaches lldb everything about registers" to "lldb knows all the architectural / ABI register number details, asks the remote stub what numbers it uses for them when connected". The need to go through RegisterContext::ConvertBetweenRegisterKinds / RegisterContext::GetRegisterInfoAtIndex to get any information about registers is unfortunate. I'm not super concerned about the size of the UnwindPlan or FuncUnwinders objects - we create them lazily as we unwind through functions in the process, so my guess is that even a long-running debug session will have on the order of thousands of these objects. The ABIs don't even bother to issue a single ArchDefault unwind plan and FunctionEntry unwind plan, handing out a new one for each FuncUnwinders object (ok that part is a little embarrassing, but again this isn't super high on the priority list to fix imo.) CHANGES SINCE LAST ACTION https://reviews.llvm.org/D61732/new/ https://reviews.llvm.org/D61732 _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits