[Bug go/64999] s390x libgo test failure in TestMemoryProfiler

[vogt at linux dot vnet.ibm.com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64999

--- Comment #10 from Dominik Vogt <vogt at linux dot vnet.ibm.com> ---
As far as I understand, the code in libbacktrace was originally only intended
for handling exceptions, not for generating stack traces?  For the former, the
code is fine.  But given a function's return address, identifying the call
instruction is non-trivial.  Gdb has complex heuristics to get that right in
case of inlined functions, tail calls or the hand written trampolines in
libffi.  Glibc probably has the same complexity for backtrace_symbols() call. 
With libgo's backtrace functionality we need to duplicate the same logic in
libbacktrace, no?