netdev - Re: [PATCH 2/2] Revert "xfrm: Fix stack-out-of-bounds read in xfrm_state

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 for Android: free password hash cracker in your pocket

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

Message-Id: <20711E45-B316-42D6-B83D-4EB7B823C3C2@gmail.com>
Date:   Sun, 24 Dec 2017 07:10:30 +1300
From:   robsonde@...il.com
To:     Steffen Klassert <steffen.klassert@...unet.com>
Cc:     David Miller <davem@...emloft.net>, herbert@...dor.apana.org.au,
        netdev@...r.kernel.org
Subject: Re: [PATCH 2/2] Revert "xfrm: Fix stack-out-of-bounds read in xfrm_state_find."



> On 24/12/2017, at 5:09 AM, Steffen Klassert <steffen.klassert@...unet.com> wrote:
> 
>> On Sat, Dec 23, 2017 at 10:56:12AM -0500, David Miller wrote:
>> From: Steffen Klassert <steffen.klassert@...unet.com>
>> Date: Sat, 23 Dec 2017 10:22:17 +0100
>> 
>>>> On Thu, Nov 16, 2017 at 11:00:40AM +0100, Steffen Klassert wrote:
>>>> This reverts commit c9f3f813d462c72dbe412cee6a5cbacf13c4ad5e.
>>>> 
>>>> This commit breaks transport mode when the policy template
>>>> has widlcard addresses configured, so revert it.
>>>> 
>>>> Signed-off-by: Steffen Klassert <steffen.klassert@...unet.com>
>>> 
>>> David, can you please queue this one up for v4.14-stable?
>>> Commit ID is 94802151894d482e82c324edf2c658f8e6b96508
>>> 
>>> v4.14 is unusable for some people without this revert.
>> 
>> Yes, but it adds back the stack out-of-bounds bug.
>> 
>> If I queue up the revert, I would also need to queue up whatever
>> follow-on you used to fix the out-of-bounds bug properly.  Which
>> commit is that?
> 
> This is commit ddc47e4404b58f03e98345398fb12d38fe291512
> ("xfrm: Fix stack-out-of-bounds read on socket policy lookup.")
> 
> It is included in the pull request for the net tree that
> I sent yesterday. The patch looks save, but not so sure
> if it should go directly to stable. These bugs reported by
> the syzbot are usually quite subtile and I already broke
> something when I tried to fix the original stack out-of-bounds
> bug. So maybe we should wait until the v4.15 release before
> backporting...
> 

At this time I cant build an IPSec VPN on a 4.14 kernel. 4.14 is LTS and so has been used as kernel choice for a prod system. This is a production fault.

My only choice is going back to a 4.13 or waiting for 4.15 or a custom build, which I am not keen on.

Unless the memory bug effects security and can be exploited, then I ask that the revert be back ported to 4.14

Thanks