On Thu, 22 Jan 2004, Henrik Nordstrom wrote: > I wonder if maybe valgrind can be extended to protect return and frame > pointer addresses on the stack. If it can then that would work quite well > for this type of error exacly where it happens. But valgrind should be > able to see this on the return if I only manage to trigger the bug under > valgrind..
did a test case for valgrind and it unfortunately can not provide any additional information on this type of bug without the above mentioned extensions as for all it knows the stack is active and it seems to rely on the stack for the backtraces like all other tools.. maybe time to build a mudflap instrumented libc.. Regards Henrik
