On Fri, Dec 23, 2016 at 11:41:00AM +0100, Martin Pieuchot wrote:
> On 23/12/16(Fri) 06:08, Visa Hankala wrote:
> > NET_LOCK() should raise IPL before acquiring the lock, and NET_UNLOCK()
> > should restore the level after releasing the lock. Otherwise, lock
> > recursion can occur, most likely right after the splx(). An example:
> >
> > nd6_slowtimo <- NET_LOCK()
> > timeout_run
> > softclock
> > softintr_dispatch
> > dosoftint
> > interrupt
> > k_intr
> > if_netisr <- NET_LOCK()
> > taskq_thread
> >
> > OK?
>
> This should never happen. Simply because the NET_LOCK() MUST NOT be
> taken in (soft) interrupt context.
>
> The real problem is that nd6_slowtimo() is set twice, once with
> timeout_set_proc(9) and once with timeout_set(9). Diff below fixes
> that.
>
> ok?
OK bluhm@
>
> Index: netinet6/nd6.c
> ===================================================================
> RCS file: /cvs/src/sys/netinet6/nd6.c,v
> retrieving revision 1.200
> diff -u -p -r1.200 nd6.c
> --- netinet6/nd6.c 22 Dec 2016 13:39:32 -0000 1.200
> +++ netinet6/nd6.c 23 Dec 2016 10:37:33 -0000
> @@ -1479,7 +1479,6 @@ nd6_slowtimo(void *ignored_arg)
>
> NET_LOCK(s);
>
> - timeout_set(&nd6_slowtimo_ch, nd6_slowtimo, NULL);
> timeout_add_sec(&nd6_slowtimo_ch, ND6_SLOWTIMER_INTERVAL);
>
> TAILQ_FOREACH(ifp, &ifnet, if_list) {
