linux-kernel - Re: [PATCH] printk: Monitor change of console loglevel.

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: <68486fcf-325f-fbd3-adb4-14666d477917@i-love.sakura.ne.jp>
Date:   Fri, 24 May 2019 19:35:28 +0900
From:   Tetsuo Handa <penguin-kernel@...ove.sakura.ne.jp>
To:     Dmitry Vyukov <dvyukov@...gle.com>
Cc:     Andrew Morton <akpm@...ux-foundation.org>,
        Sergey Senozhatsky <sergey.senozhatsky.work@...il.com>,
        LKML <linux-kernel@...r.kernel.org>,
        Petr Mladek <pmladek@...e.com>,
        Steven Rostedt <rostedt@...dmis.org>,
        Linus Torvalds <torvalds@...ux-foundation.org>
Subject: Re: [PATCH] printk: Monitor change of console loglevel.

On 2019/05/24 16:55, Dmitry Vyukov wrote:
> On Thu, May 23, 2019 at 11:57 AM Tetsuo Handa
> <penguin-kernel@...ove.sakura.ne.jp> wrote:
>>
>> Well, the culprit of this problem might be syz_execute_func().
>>
>>   https://twitter.com/ed_maste/status/1131165065485398016
>>
>> Then, blacklisting specific syscalls/arguments might not work.
>> We will need to guard specific paths on the kernel side using
>> some kernel config option...
> 
> Yes, that's a nasty issue. We could stop running random code, or
> setuid into nobody, but then we will lose lots of test coverage...
> 

I think that guarding specific paths on the kernel side is better.
TOMOYO already added CONFIG_SECURITY_TOMOYO_INSECURE_BUILTIN_SETTING for
avoid emitting WARNING: string and getting more test coverage. There are
other codes emitting WARNING: string that confuses syzbot. If we guard
critical paths like reboot/poweroff request that will destroy the target
VM instance, we can get more test coverage while reducing pointless reports.