In article <20020930180844.GA55908@ego.yagosys.com>,
Artem Belevich <art@riverstonenet.com> wrote:
>Hi,
>
>I keep getting gernel panics in 1.6 kernel whenever there is a process
>that does a lot of stat calls. That, unfortunately, involves
>/etc/daily script and nightly backups. The crash always happens in the
>same place (see kern/17107 and the stack trace below).
>
>This was happening with 1.6_BETA1 and it's still in 1.6-RELEASE.
>
>Frank van der Linden suggested that it may be a "race condition with
>reclaiming vnodes and unmounting a filesystem; this vnode seems to
>have its mountpoint zeroed out too soon".
>
>I would appreciate any hints/clues/suggestions/whatnot that would help
>me to find solution of workaround for the panic.

Yes, frank appears to be right; try changing:

	if ((nmp->nm_flag & NFSMNT_NQNFS) && np->n_timer.cqe_next != 0) {
to:

	if (nmp && (nmp->nm_flag & NFSMNT_NQNFS) && np->n_timer.cqe_next != 0) {

in nfs_node.c.

christos