netdev - Re: [PATCH net-next 0/2] Bonding: support new xfrm state offload functions

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <69a8ec12-ac65-46e3-9d0c-ba9a56dbe5bb@blackwall.org>
Date: Fri, 16 Aug 2024 12:58:30 +0300
From: Nikolay Aleksandrov <razor@...ckwall.org>
To: Hangbin Liu <liuhangbin@...il.com>
Cc: netdev@...r.kernel.org, Jay Vosburgh <j.vosburgh@...il.com>,
 "David S . Miller" <davem@...emloft.net>, Jakub Kicinski <kuba@...nel.org>,
 Paolo Abeni <pabeni@...hat.com>, Eric Dumazet <edumazet@...gle.com>,
 Tariq Toukan <tariqt@...dia.com>, Jianbo Liu <jianbol@...dia.com>,
 Sabrina Dubroca <sd@...asysnail.net>
Subject: Re: [PATCH net-next 0/2] Bonding: support new xfrm state offload
 functions

On 16/08/2024 12:04, Nikolay Aleksandrov wrote:
> On 16/08/2024 11:37, Nikolay Aleksandrov wrote:
>> On 16/08/2024 11:32, Hangbin Liu wrote:
>>> On Fri, Aug 16, 2024 at 09:06:12AM +0300, Nikolay Aleksandrov wrote:
>>>> On 16/08/2024 06:55, Hangbin Liu wrote:
>>>>> I planned to add the new XFRM state offload functions after Jianbo's
>>>>> patchset [1], but it seems that may take some time. Therefore, I am
>>>>> posting these two patches to net-next now, as our users are waiting for
>>>>> this functionality. If Jianbo's patch is applied first, I can update these
>>>>> patches accordingly.
>>>>>
>>>>> [1] https://lore.kernel.org/netdev/20240815142103.2253886-2-tariqt@nvidia.com
>>>>>
>>>>> Hangbin Liu (2):
>>>>>   bonding: Add ESN support to IPSec HW offload
>>>>>   bonding: support xfrm state update
>>>>>
>>>>>  drivers/net/bonding/bond_main.c | 76 +++++++++++++++++++++++++++++++++
>>>>>  1 file changed, 76 insertions(+)
>>>>>
>>>>
>>>> (not related to this set, but to bond xfrm)
>>>> By the way looking at bond's xfrm code, what prevents bond_ipsec_offload_ok()
>>>> from dereferencing a null ptr?
>>>> I mean it does:
>>>>         curr_active = rcu_dereference(bond->curr_active_slave);
>>>>         real_dev = curr_active->dev;
>>>>
>>>> If this is running only under RCU as the code suggests then
>>>> curr_active_slave can change to NULL in parallel. Should there be a
>>>> check for curr_active before deref or am I missing something?
>>>
>>> Yes, we can do like
>>> real_dev = curr_active ? curr_active->dev : NULL;
>>>
>>> Thanks
>>> Hangbin
>>
>> Right, let me try and trigger it and I'll send a patch. :)
>>
> 
> Just fyi I was able to trigger this null deref easily and one more
> thing - there is a second null deref after this one because real_dev is
> used directly :(
> 
> I'll send a patch to fix both in a bit.
> 
> 

TBH I don't know how bond xfrm code is running at all. There are *many*
bugs when I took a closer look. I'll try to prepare a patch-set to address
them all. This code is seriously broken...