Re: bash coredumping?

["Christopher C. Chimelis" <chris@debian.org>]

On 14 Mar 2000, John Goerzen wrote:

> strace sez:
> 
> 
> munmap(0x20000026000, 8192)             = 0
> osf_sigprocmask(0x1, 0, 0x11ffff750)    = 0
> osf_sigprocmask(0x3, 0, 0)              = 0
> osf_sigprocmask(0x1, 0, 0x11ffff750)    = 0
> osf_sigprocmask(0x1, 0, 0x11ffff750)    = 0
> osf_sigprocmask(0x1, 0, 0x11ffff750)    = 0
> osf_sigprocmask(0x1, 0, 0x11ffff750)    = 0
> osf_sigprocmask(0x1, 0, 0x11ffff6c0)    = 0
> osf_sigprocmask(0x1, 0, 0x11ffff6c0)    = 0
> exit(139)                               = ?

Oh, boy....

> gdb sez:
> 
> (gdb) bt
> #0  0x20000652710 in memcpy () from /lib/libc.so.6.1
> #1  0x12004d8a4 in unwind_protect_var ()
> #2  0x120028d18 in execute_command_internal ()
> #3  0x12002945c in execute_shell_function ()
> warning: Hit heuristic-fence-post without finding
> warning: enclosing function for address 0xffffffffffffffff
> 
> that's about it.

Yep, that's what I thought...bounds-checking problem.  bash should
probably be recompiled for the new glibc.  I'll NMU it, but offer it to
you for testing first.  I think memcpy() was modified on our target, fyi.

C