[Lldb-commits] [PATCH] D91238: Recognize __builtin_debugtrap on arm64, advance pc past it so users can continue easily

[Greg Clayton via Phabricator via lldb-commits]

clayborg added a comment.

Another possible way to do this would be to make a new stop reason in LLDB. debugserver currently can return a "reason" whose values are one of "trap", "breakpoint", "watchpoint", "exception", or "exec". What if we added "debug-trap"? Then back in LLDB we could add "eStopReasonDebugTrap" to "lldb::StopReason". Then the ABI plug-in or some arch specific plug-in could do something in response to this on stop, or on continue. x86 would know that is doesn't need to do anything because the PC is already past the trap instructions. But ARM64 it could advance the PC when the process goes to start running again so it skips the trap instruction. I would presonally like to see the program tell the truth when you hit a "brk #0xf000" and show the PC being on this instruction, and then continue from it correctly when you resume. Your solution is much easier though, so either way works.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D91238/new/

https://reviews.llvm.org/D91238