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: <df46bdfa-d149-4823-d6b8-e350275cd8f0@i-love.sakura.ne.jp>
Date:   Sat, 25 May 2019 09:14:15 +0900
From:   Tetsuo Handa <penguin-kernel@...ove.sakura.ne.jp>
To:     Linus Torvalds <torvalds@...ux-foundation.org>
Cc:     Andrew Morton <akpm@...ux-foundation.org>,
        Linux List Kernel Mailing <linux-kernel@...r.kernel.org>,
        Dmitry Vyukov <dvyukov@...gle.com>,
        Petr Mladek <pmladek@...e.com>,
        Sergey Senozhatsky <sergey.senozhatsky.work@...il.com>,
        Joe Perches <joe@...ches.com>
Subject: Re: [PATCH] printk: Monitor change of console loglevel.

On 2019/05/25 2:17, Linus Torvalds wrote:
> A config option or two that help syzbot doesn't sound like a bad idea to me.

Thanks for suggestion. I think that #ifdef'ing

  static bool suppress_message_printing(int level)
  {
  	return (level >= console_loglevel && !ignore_loglevel);
  }

is simpler. If the cause of unexpected change of console loglevel
turns out to be syz_execute_func(), we will want a config option
which controls suppress_message_printing() for syzbot. That option
would also be used for guarding printk("WARNING:" ...) users.

Well, syzbot does not want to use ignore_loglevel kernel command
line option because that option would generate too much output...

  https://lkml.kernel.org/r/CACT4Y+ay7nuT-7y2JARozV1s0VisuLdN6VT+w9OsEDs1PeBRoA@mail.gmail.com

On 2019/05/25 2:55, Linus Torvalds wrote:
> On Fri, May 24, 2019 at 10:41 AM Joe Perches <joe@...ches.com> wrote:
> >
> > That could also help eliminate unnecessary pr_<foo> output
> > from object code.
> 
> Indeed. The small-config people might like it (if they haven't already
> given up..)

Do you mean doing e.g.

  #define pr_debug(fmt, ...) no_printk(KERN_DEBUG pr_fmt(fmt), ##__VA_ARGS__)

depending on the minimal console loglevel kernel config option? Then, OK.
But callers using e.g. printk(KERN_DEBUG ...) and printk(KERN_SOH "%u" ...)
will remain unfiltered...