On Tue, Sep 22, 2015 at 11:32:51AM +0200, Yvan Coene wrote:
> A stale NFS file handle on a VirtualBox host shared with a Debian guest
> under /usr/global results in `man` segfaulting at startup on the guest:
>
> ycoene@debian:~$ strace man man
>
> [...]
>
> lstat64("/usr", {st_mode=S_IFDIR|0755, st_size=4096, ...}) = 0
> lstat64("/usr/share", {st_mode=S_IFDIR|0755, st_size=12288, ...}) = 0
> lstat64("/usr/share/man", {st_mode=S_IFDIR|0755, st_size=4096, ...}) = 0
> lstat64("/usr", {st_mode=S_IFDIR|0755, st_size=4096, ...}) = 0
> lstat64("/usr/global", 0xbfb37db0) = -1 EPROTO (Protocol error)
> --- SIGSEGV {si_signo=SIGSEGV, si_code=SEGV_MAPERR, si_addr=0} ---
> +++ killed by SIGSEGV +++
> Segmentation fault
This is difficult for me to reproduce locally. Could you please run
"strace -o man.trace -s 1024 man -d man" and attach the resulting
man.trace file to this bug, to give me the best chance of working out
where this crash happens?
Thanks,
--
Colin Watson [[email protected]]
