linux-kernel - Re: [PATCH] ipc/sem: Avoid indexing past end of sem

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: <CAGXu5jJt9yB-4tbcBOBwqFB5STP7AVYFt-THpjHsCJNL7jdvKA@mail.gmail.com>
Date:   Mon, 15 May 2017 10:40:34 -0700
From:   Kees Cook <keescook@...omium.org>
To:     Manfred Spraul <manfred@...orfullife.com>
Cc:     Andrew Morton <akpm@...ux-foundation.org>,
        Davidlohr Bueso <dave@...olabs.net>,
        Ingo Molnar <mingo@...nel.org>,
        Peter Zijlstra <peterz@...radead.org>,
        Fabian Frederick <fabf@...net.be>,
        LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] ipc/sem: Avoid indexing past end of sem_array

On Sun, May 14, 2017 at 6:54 AM, Manfred Spraul
<manfred@...orfullife.com> wrote:
> Hi Kees,
>
> On 05/09/2017 12:23 AM, Kees Cook wrote:
>>
>> This changes the struct + trailing data pattern to using a void * so that
>> the end of sem_array is found without possibly indexing past the end which
>> can upset some static analyzers. Mostly, this ends up avoiding a cast
>> between different non-void types, which the future randstruct GCC plugin
>> was warning about.
>
> Two question:
> - Would the attached patch work with the randstruct plugin as well?
>   If we touch the code, then I would propose that we remove sem_base
> entirely.

I'll double check with your series, but I think your change makes
sense regardless (since it makes it very clear that there are
allocated sems after the struct due to the [0] entry).

>
> - ipc/util.h contains
>
>> #define ipc_rcu_to_struct(p)  ((void *)(p+1))
>
> Does this trigger a warning with randstruct as well?
> If we have to touch it, then I would remove it by merging struct
> kern_ipc_perm and struct ipc_rcu.
>
> And, obviously:
> Do you see any issues with the attached patch?

I'll test your series with the randstruct series and see what falls out. :)

Thanks!

-Kees

-- 
Kees Cook
Pixel Security