jasonmolenda wrote: OK finished writing the test case. It starts at main() and instruction steps through the test program, checking that it can backtrace all the way to main() (including all the frames in the middle) at each instruction.
I also added tests for the non-ABI compliant fetching of x0 mid-stack and x20 being unfetchable mid-stack, with the hand-written eh_frame instructions in the assembly file. It has nothing to do with the frameless function that faults that I'm trying to fix here, but it was interesting that it works as I wanted so I'll add a test on it. I have the test marked as skipUnlessDarwin but I'm not sure there's much to keep it from working on Linux. Maybe that I had to prefix the symbol names with _ on dawin? Might be easiest to update the .c file to call either to_be_interrupted or _to_be_interrupted depending on the OS if that's it. https://github.com/llvm/llvm-project/pull/138805 _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
