On Wed, 20 Jul 2016, Michal Kubecek wrote:

On Fri, Jul 15, 2016 at 03:04:48PM +0200, Thomas Voegtle wrote:

quoted

And that one?
Happens while trying to start a firewall script with iptables-restore.

[  180.071999] NMI watchdog: BUG: soft lockup - CPU#1 stuck for 22s!
[iptables-restor:2338]

...

quoted

[  180.072000] Call Trace:
[  180.072000]  [<ffffffff8111a923>] ? __kmalloc_node+0x26/0x2d
[  180.072000]  [<ffffffff81667495>] ? cpumask_next+0x1b/0x1d
[  180.072000]  [<ffffffff81668ec3>] ? xt_alloc_table_info+0xc2/0xdf
[  180.072000]  [<ffffffffc05be6c3>] do_ipt_set_ctl+0xfd/0x193 [ip_tables]
[  180.072000]  [<ffffffff81665181>] nf_sockopt+0x64/0x7e
[  180.072000]  [<ffffffff816651cc>] nf_setsockopt+0x1c/0x20
[  180.072000]  [<ffffffff81674dd1>] ip_setsockopt+0x6e/0x8e
[  180.072000]  [<ffffffff8168fe47>] raw_setsockopt+0x11/0x4c
[  180.072000]  [<ffffffff816296ab>] sock_common_setsockopt+0xf/0x11
[  180.072000]  [<ffffffff81628ab2>] SyS_setsockopt+0x85/0xa5
[  180.072000]  [<ffffffff816e7dee>] system_call_fastpath+0x12/0x71

This should be the problem addressed by

  http://article.gmane.org/gmane.linux.network/422215

Michal Kubecek

Hi,

I re-checked your patch today on top of a 4.1.29 and I'm very sorry
to say it actually does fix the problem despite what I wrote a few days 
ago.

I don't know why I saw the problem while I was sure I had that patch
applied. Maybe I mixed up the builds. Should have double checked it.

So, if that patch would appear in the next 4.1.y, that would be great.

talking about that patch:
  https://patchwork.ozlabs.org/patch/650120/
(as gmame is down atm)

thanks,

    Thomas