[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Wed, 08 Aug 2012 12:14:42 -0700
From: John Stultz <johnstul@...ibm.com>
To: "Serge E. Hallyn" <serge@...lyn.com>
CC: Paul Moore <paul@...l-moore.com>,
lkml <linux-kernel@...r.kernel.org>,
James Morris <james.l.morris@...cle.com>,
selinux@...ho.nsa.gov, Eric Dumazet <edumazet@...gle.com>,
john.johansen@...onical.com
Subject: Re: NULL pointer dereference in selinux_ip_postroute_compat
On 08/07/2012 03:37 PM, John Stultz wrote:
> On 08/07/2012 03:17 PM, Serge E. Hallyn wrote:
>> Quoting Paul Moore (paul@...l-moore.com):
>>> On Tue, Aug 7, 2012 at 5:58 PM, John Stultz <john.stultz@...aro.org>
>>> wrote:
>>>> On 08/07/2012 02:50 PM, Paul Moore wrote:
>>>>> On Tue, Aug 7, 2012 at 2:12 PM, John Stultz <john.stultz@...aro.org>
>>>>> wrote:
>>>>>> Hi,
>>>>>> With my kvm environment using 3.6-rc1+, I'm seeing NULL
>>>>>> pointer
>>>>>> dereferences in selinux_ip_postroute_compat(). It looks like the
>>>>>> sksec
>>>>>> value
>>>>>> is null and we die in the following line:
>>>>>>
>>>>>> if (selinux_xfrm_postroute_last(sksec->sid, skb, &ad, proto))
>>>>>>
>>>>>> This triggers every time I shutdown the machine, but has also
>>>>>> triggered
>>>>>> randomly after a few hours.
[snip]
>> The problem seems to be that selinux_nf_ip_init() was called, which
>> registers the selinux_ipv4_ops (and ipv6). Those should not get
>> registered
>> if selinux ends up not being loaded (as in, if apparmor is loaded
>> first),
>> since as you've found here the selinux lsm hooks won't be called to set
>> call selinux_sk_alloc_security().
> This sounds about right:
> root@...tvm:~# dmesg | grep SELinux
> [ 0.004578] SELinux: Initializing.
> [ 0.005704] SELinux: Starting in permissive mode
> [ 2.235034] SELinux: Registering netfilter hooks
>
>> I assume what's happening is that
>> CONFIG_SECURITY_SELINUX_BOOTPARAM_VALUE was
>> set to 1, but selinux ended up being set to disabled after the
>> __initcall(selinux_nf_ip_init) ran? Weird.
> This looks right as well:
>
> # zcat config.gz | grep SELINUX
> CONFIG_SECURITY_SELINUX=y
> CONFIG_SECURITY_SELINUX_BOOTPARAM=y
> CONFIG_SECURITY_SELINUX_BOOTPARAM_VALUE=1
> CONFIG_SECURITY_SELINUX_DISABLE=y
> CONFIG_SECURITY_SELINUX_DEVELOP=y
> CONFIG_SECURITY_SELINUX_AVC_STATS=y
> CONFIG_SECURITY_SELINUX_CHECKREQPROT_VALUE=1
> # CONFIG_SECURITY_SELINUX_POLICYDB_VERSION_MAX is not set
> CONFIG_DEFAULT_SECURITY_SELINUX=y
>
>
> Since the problem isn't completely obvious, I'm starting a bisection
> to narrow this down some more.
So I bisected this down and it seems to be the following commit:
commit be9f4a44e7d41cee50ddb5f038fc2391cbbb4046
Author: Eric Dumazet <edumazet@...gle.com>
Date: Thu Jul 19 07:34:03 2012 +0000
ipv4: tcp: remove per net tcp_sock
It doesn't revert totally cleanly, but after fixing up the rejections
and booting with this patch removed on top of Linus' head the oops on
shutdown goes away.
thanks
-john
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists