> Date: Wed, 28 Feb 2024 16:16:09 +0300
> From: Vitaliy Makkoveev <m...@openbsd.org>
>
> On Wed, Feb 28, 2024 at 12:36:26PM +0100, Claudio Jeker wrote:
> > On Wed, Feb 28, 2024 at 12:26:43PM +0100, Marko Cupać wrote:
> > > Hi,
> > >
> > > thank you for looking into it, and for the advice.
> > >
> > > On Wed, 28 Feb 2024 10:13:06 +0000
> > > Stuart Henderson <s...@spacehopper.org> wrote:
> > >
> > > > Please try to re-type at least the most important bits from a
> > > > screenshot so readers can quickly see which subsystems are involved.
> > >
> > > Below is manual transcript of whole screenshot, hopefully no typos.
> > >
> > > If you have any advice on what should I do if it happens again in order
> > > to get as much info for debuggers as possible, please let me know.
> > >
> > > splassert: assertwaitok: want 0 have 4
> > > panic: kernel diagnostic assertion "p->p_wchan == NULL" failed: file
> > > "/usr/src/sys/kern/kern_sched.c", line 373
> > > Stopped at db_enter+0x14: popq %rbp
> > > TID PID UID PRFLAGS PFLAGS CPU COMMAND
> > > 199248 36172 577 0x10 0 1 openvpn
> > > 490874 47446 0 0x14000 0x200 2 wg_handshake
> > > 71544 9311 0 0x14000 0x200 3 softnet0
> > > db_enter() at db_enter+0x14
> > > panic(ffffffff820a4b9f) at panic+0xc3
> > > __assert(ffffffff82121fcb,ffffffff8209ae5f,175,ffffffff82092fbf) at
> > > assert+0x29
> > > sched_chooseproc() at sched_chooseproc+0x26d
> > > mi_switch() at mi_switch+0x17f
> > > sleep_finish(0,1) at sleep_finish+0x107
> > > rw_enter(ffff800008003cf0,2) at rw_enter+0x1ad
> > > noise_remote_ready(ffff800008003bf0) at noise_remote_ready+0x33
> > > wg_qstart(fff800000a622a8) at wg_qstart+0x18c
> > > ifq_serialize(ffff800000a622a8,ffff800000a62390) at ifq_serialize+0xfd
> > > hfsc_deferred(ffff800000a62000) at hfsc_deferred+0x68
> > > softclock_process_tick_timeout(ffff80000115e248,1) at
> > > softclock_process_tick_timeout+0xfb
> > > softclock(0) at softclock+0xb8
> > > softintr_dispatch(0) at softintr_dispatch+0xeb
> > > end trace frame: 0xffff800020dbc730, count:0
> > >
> >
> > WTF! wg(4) is just broken. How the hell should a sleeping rw_lock work
> > when called from inside a timeout aka softclock? This is interrupt context
> > code is not allowed to sleep there.
> >
>
> Not only wg(4). Depends on interface queue usage, ifq_start() schedules
> (*if_qstart)() or calls it, so all the interfaces with use rwlock(9) in
> (*if_qstart)() handler are in risk.
>
> What about to always schedule (*if_qstart)()?
Why would you want to introduce additional latence?
> Index: sys/net/hfsc.c
> ===================================================================
> RCS file: /cvs/src/sys/net/hfsc.c,v
> retrieving revision 1.49
> diff -u -p -r1.49 hfsc.c
> --- sys/net/hfsc.c 11 Apr 2023 00:45:09 -0000 1.49
> +++ sys/net/hfsc.c 28 Feb 2024 13:15:22 -0000
> @@ -953,8 +953,7 @@ hfsc_deferred(void *arg)
> if (!HFSC_ENABLED(ifq))
> return;
>
> - if (!ifq_empty(ifq))
> - ifq_start(ifq);
> + ifq_start_deferred(ifq);
>
> hif = ifq_q_enter(&ifp->if_snd, ifq_hfsc_ops);
> if (hif == NULL)
> Index: sys/net/ifq.c
> ===================================================================
> RCS file: /cvs/src/sys/net/ifq.c,v
> retrieving revision 1.53
> diff -u -p -r1.53 ifq.c
> --- sys/net/ifq.c 10 Nov 2023 15:51:24 -0000 1.53
> +++ sys/net/ifq.c 28 Feb 2024 13:15:22 -0000
> @@ -133,6 +133,12 @@ ifq_start(struct ifqueue *ifq)
> } else
> task_add(ifq->ifq_softnet, &ifq->ifq_bundle);
> }
> +void
> +ifq_start_deferred(struct ifqueue *ifq)
> +{
> + if (ifq_len(ifq))
> + task_add(ifq->ifq_softnet, &ifq->ifq_bundle);
> +}
>
> void
> ifq_start_task(void *p)
> Index: sys/net/ifq.h
> ===================================================================
> RCS file: /cvs/src/sys/net/ifq.h,v
> retrieving revision 1.41
> diff -u -p -r1.41 ifq.h
> --- sys/net/ifq.h 10 Nov 2023 15:51:24 -0000 1.41
> +++ sys/net/ifq.h 28 Feb 2024 13:15:22 -0000
> @@ -430,6 +430,7 @@ void ifq_destroy(struct ifqueue *);
> void ifq_add_data(struct ifqueue *, struct if_data *);
> int ifq_enqueue(struct ifqueue *, struct mbuf *);
> void ifq_start(struct ifqueue *);
> +void ifq_start_deferred(struct ifqueue *);
> struct mbuf *ifq_deq_begin(struct ifqueue *);
> void ifq_deq_commit(struct ifqueue *, struct mbuf *);
> void ifq_deq_rollback(struct ifqueue *, struct mbuf *);
>
>
