[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20180731113152.4pes5xwhae2xydqg@gauss3.secunet.de>
Date: Tue, 31 Jul 2018 13:31:52 +0200
From: Steffen Klassert <steffen.klassert@...unet.com>
To: Tetsuo Handa <penguin-kernel@...ove.SAKURA.ne.jp>
CC: Herbert Xu <herbert@...dor.apana.org.au>,
"David S. Miller" <davem@...emloft.net>,
syzbot <syzbot+2dfb68e639f0621b19fb@...kaller.appspotmail.com>,
<ddstreet@...e.org>, <dvyukov@...gle.com>,
<linux-kernel@...r.kernel.org>, <netdev@...r.kernel.org>,
<syzkaller-bugs@...glegroups.com>
Subject: Re: unregister_netdevice: waiting for DEV to become free
On Tue, Jul 31, 2018 at 08:16:22PM +0900, Tetsuo Handa wrote:
> Steffen and Herbert,
>
> Do you have any question? I think I provided enough information for debugging.
It seems that I was not on Cc at the beginning of the threat,
so I had to search the web for some context.
I'm currently trying your reproducer (still with my .config) but I don't
hit the problem.
>
> This problem occurs because two dev_put() calls are missing (compared with not
> calling setsockopt(SOL_IPV6, IPV6_XFRM_POLICY)) because dst_release() is not
> called via fib6_info_destroy_rcu() when we called xfrm_compile_policy() from
> xfrm_user_policy() from setsockopt(SOL_IPV6, IPV6_XFRM_POLICY).
Thanks for the information! As said, I'm already working on it.
Powered by blists - more mailing lists