Re: [libunwind] walking the callstack on an x86 machine

[Arun Sharma]

Alex Shye wrote:
> Hi David,
>
> Thanks for the reply.  I don't know much about the DWARF format I as I 
> understand, it is included with compiling with debug information, correct?
>
> I guess I didn't state my question clearly enough, but what if you only 
> have stripped binaries without any debug information to work with?  Are 
> there any resources for stack walking in such a scenario?

Stripping binaries to remove unwind information means language run times
(such as C++ exception handing) will have trouble as well. It's possible
to strip binaries without removing unwind info.

Another issue with speculative stack walking is safety. If the unwinder
is in-process, how do you check if the return address is valid in a
non-racy way?

        -Arun