[PATCH] D18895: [sanitizer] [SystemZ] Fix stack traces.

[Ulrich Weigand via llvm-commits]

uweigand added a comment.

OK, in that case I agree with option (2).  Implementing -mbackchain in itself doesn't hurt anyway, if only for compatibility with GCC.

[ Note that -fno-omit-frame-pointer doesn't suffice on s390x to enable unwinding; even though functions up the call stack will then use a frame pointer, and save it into their register save area, you still do not know without consulting DWARF CFI *where* that save area is (since it is at the top of the frame and you don't know the frame size). ]


Repository:
  rL LLVM

http://reviews.llvm.org/D18895