Re: [2.6.21.1] soft lockup when removing netconsole module

[Jarek Poplawski <jarkao2@o2.pl>]

On Tue, May 29, 2007 at 12:56:28AM -0700, Andrew Morton wrote:
> On Sat, 26 May 2007 17:40:12 +0200 Folkert van Heusden <folkert@vanheusden.com> wrote:
> 
> > When trying to remove the netconsole module, I got the following kernel
> > output after a while (couple of minutes iirc):
> > 
> > [525720.117293] BUG: soft lockup detected on CPU#1!
> > [525720.117353]  [<c1004d53>] show_trace_log_lvl+0x1a/0x30
> > [525720.117439]  [<c1004d7b>] show_trace+0x12/0x14
> > [525720.117526]  [<c1004e75>] dump_stack+0x16/0x18
> > [525720.117613]  [<c104dd5b>] softlockup_tick+0xa6/0xc2
> > [525720.117694]  [<c1026855>] run_local_timers+0x12/0x14
> > [525720.117738]  [<c1026669>] update_process_times+0x72/0xa1
> > [525720.117744]  [<c1038673>] tick_sched_timer+0x53/0xb6
> > [525720.117748]  [<c1033d62>] hrtimer_interrupt+0x189/0x1e3
> > [525720.117753]  [<c100e9e2>] local_apic_timer_interrupt+0x55/0x5b
> > [525720.117761]  [<c100ea12>] smp_apic_timer_interrupt+0x2a/0x39
> > [525720.117766]  [<c1004a3f>] apic_timer_interrupt+0x33/0x38
> > [525720.117770]  [<c120f4b1>] mutex_lock+0x8/0xa
> > [525720.117775]  [<c102d2f0>] flush_workqueue+0x2f/0x8f
> > [525720.117780]  [<c102d7a0>] cancel_rearming_delayed_workqueue+0x29/0x2b
> > [525720.117785]  [<c102d7b1>] cancel_rearming_delayed_work+0xf/0x11
> > [525720.117790]  [<c11be143>] netpoll_cleanup+0x75/0xa5
> > [525720.117794]  [<f893712d>] cleanup_netconsole+0x17/0x1a [netconsole]
> > [525720.117804]  [<c1041f11>] sys_delete_module+0x12f/0x14f
> > [525720.117809]  [<c1003f74>] syscall_call+0x7/0xb
> > [525720.117812]  =======================
> > 
> > Also the rmmod hangs and would not exit even with kill -9. It also
> > sucks up 100% cpu.
> 
> Jason recently posted a mystery patch without telling us what problem it
> fixed.
> 

To be fair the problem should be known:

http://marc.info/?l=linux-kernel&m=117700287817801&w=2

List:       linux-kernel
Subject:    Re: [PATCH -mm] workqueue: debug possible endless loop in cancel_rearming_delayed_work
From:       Chuck Ebbert <cebbert () redhat ! com>
Date:       2007-04-19 17:07:11
Message-ID: 4627A1BF.8080406 () redhat ! com

> Okay, an easy test for it: insmod netconsole ; rmmod netconsole
> 
> In 2.6.20.x it loops forever and cancel_rearming_delayed_work()
> is part of the trace...

I hoped the discussion about cancel_rearming_delayed_work would
reach more people (there was also a patch proposal to add a warning
to the usage comment). But it seem it was not enough...

Of course such a problem should preferably be fixed by somebody who
knows the code (alas I don't know netconsole), to be sure all needed
cancels are still done after this change. I hope Jason's patch is
right but I'm a little surprised I can't see netdev in cc (I'll try
to fix this).

Cheers,
Jarek P.

PS: I'm very sorry for such late response (holidays).

> It looks like you just found it: cancel_rearming_delayed_work() will hang
> if the work isn't actually pending.  Please test this:
> 
> 
> From: Jason Wessel <jason.wessel@windriver.com>
> 
> Do not call cancel_rearming_delayed_work() if there is no
> pending work.
> 
> Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
> ---
> 
>  net/core/netpoll.c |    6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)
> 
> diff -puN net/core/netpoll.c~a net/core/netpoll.c
> --- a/net/core/netpoll.c~a
> +++ a/net/core/netpoll.c
> @@ -784,8 +784,10 @@ void netpoll_cleanup(struct netpoll *np)
>  			if (atomic_dec_and_test(&npinfo->refcnt)) {
>  				skb_queue_purge(&npinfo->arp_tx);
>  				skb_queue_purge(&npinfo->txq);
> -				cancel_rearming_delayed_work(&npinfo->tx_work);
> -				flush_scheduled_work();
> +				if (delayed_work_pending(&npinfo->tx_work)) {
> +					cancel_rearming_delayed_work(&npinfo->tx_work);
> +					flush_scheduled_work();
> +				}
>  
>  				kfree(npinfo);
>  			}
> _
>