linux-kernel - Re: [BUG] Threaded printk breaks early debugging

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]

Date:   Tue, 14 Jun 2022 10:37:13 +0200
From:   Petr Mladek <pmladek@...e.com>
To:     David Laight <David.Laight@...lab.com>
Cc:     Sergey Senozhatsky <senozhatsky@...omium.org>,
        John Ogness <john.ogness@...utronix.de>,
        Peter Geis <pgwipeout@...il.com>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        "open list:ARM/Rockchip SoC..." <linux-rockchip@...ts.infradead.org>
Subject: Re: [BUG] Threaded printk breaks early debugging

On Mon 2022-06-13 16:11:19, David Laight wrote:
> From: Petr Mladek
> > Sent: 13 June 2022 11:14
> ...
> > Another interesting alternative is the Peter Zijlstra's mode
> > where all messages are printed to the console "immediately".
> > They are serialized only by the CPU-reentrant lock.
> > 
> > This mode is not good for production system. But it might
> > be good for debugging. The good thing is that the behavior
> > is well defined.
> 
> ISTM that all messages should be output (even if this means
> that cpu spin waiting for a serial console) until userspace
> has a chance to set an option to change the behaviour.

There are systems that do not boot because of softlockups
caused by slow consoles.

I am pretty sure that we will add the option to disable
printk kthreads sooner or later. But we prefer to improve
the code so that it works out of box. If we add the option
already know then people will disable kthreads and
do not report problematic scenarios.

> Oh, and can someone stop the distros hiding the console output.
> I want to see the 'oops' traceback when I break the kernel!

Different users have different needs. AFAIK, distros hide
the messages because they scare users and overload support.
It is probably more effective to enable them only for debugging
when users notice some malfunction.

I do not know the details. I guess that there were years of
evolution. The default in SUSE was probably set by some
project managers or people taking care of boot splash,
installation, ...

Best Regards,
Petr