linux-kernel - Re: Revert needed: udev spewing warnons on common systems in 3.0

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: <alpine.DEB.2.00.1108011916480.15596@chino.kir.corp.google.com>
Date:	Mon, 1 Aug 2011 19:22:08 -0700 (PDT)
From:	David Rientjes <rientjes@...gle.com>
To:	Linus Torvalds <torvalds@...ux-foundation.org>
cc:	Andrew Morton <akpm@...ux-foundation.org>,
	Andi Kleen <andi@...stfloor.org>, linux-kernel@...r.kernel.org
Subject: Re: Revert needed: udev spewing warnons on common systems in 3.0

On Mon, 1 Aug 2011, Linus Torvalds wrote:

> > aw, what crap.
> >
> > "/proc/%d/oom_adj is deprecated, please use /proc/%d/oom_score_adj instead."
> >
> > Once per boot.  That's not "scary".
> 
> It is when there is that totally pointless stack trace, and the kernel
> bug reporting tool picks it up and tells the user that there was a
> kernel oops.
> 
> Which it really does.
> 
> So if it was a "printk_once()" it would be fine. The WARN_ONCE() is
> just annoying.
> 

It was a printk_once() when the tunable was scheduled to be removed two 
years later.  Now that one year has passed since the oom killer rewrite 
has been merged, it was changed to a WARN_ONCE() simply to attract more 
attention to the issue.  The plan was to remove the tunable completely a 
year from now as specified in Documentation/feature-removal-schedule.txt.

The intent here was to be more helpful than annoying or causing a problem 
with a log parser that doesn't understand the different between a warning 
and a panic or oops.

We could change it to a more obvious multi-line notification that an 
application is using a deprecated interface that wouldn't parse like the 
WARN_ONCE() output if you'd like.  The single printk_once() obviously 
wasn't sufficient over the course of a year to get enough applications to 
change although things like udev, kde, and opensshd did.  chromium 
actually noticed the new WARN_ON() behavior and is in the process of 
fixing it because of that.