linux-kernel - Re: linux.git: printk() problem

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: <1476354050.2164.7.camel@perches.com>
Date:   Thu, 13 Oct 2016 03:20:50 -0700
From:   Joe Perches <joe@...ches.com>
To:     Michal Hocko <mhocko@...nel.org>
Cc:     Tetsuo Handa <penguin-kernel@...ove.SAKURA.ne.jp>,
        torvalds@...ux-foundation.org, linux-kernel@...r.kernel.org,
        Colin Ian King <colin.king@...onical.com>
Subject: Re: linux.git: printk() problem

On Thu, 2016-10-13 at 12:04 +0200, Michal Hocko wrote:
> On Thu 13-10-16 02:29:46, Joe Perches wrote:
> > On Thu, 2016-10-13 at 08:26 +0200, Michal Hocko wrote:
> 
> [...]
> > > I think they are not critical and can be fix once somebody notices.
> > 
> >  
> > As do I, but Linus objected to applying a patch when Colin Ian King
> > noticed one.
> > 
> > I think the 250,000 or so uses with newlines are enough of a
> > precedence to keep using newlines everywhere.
> 
> 
> or simply fix missing KERN_CONTs and simply do not add any new missing \n
> 
> > Now we'll have to have patches adding hundreds to thousands of the
> > missing KERN_CONTs for continuation lines that weren't previously a
> > problem in logging output but are now.
> 
> 
> I would be really surprised if we really had that many continuation
> lines. They should be avoided as much as possible. Hundreds of thousands
> just sounds more than over exaggerated...

Hey Michal.

"Hundreds _to_ thousands" of instances.  Not "hundreds _of_ thousands".

> Not requiring \n at the end of strings just makes a lot of sense if we
> have a KERN_CONT with a well defined semantic.

True enough.  And I am not at all arguing against having a
well defined KERN_CONT semantic.

But using KERN_CONT alone is not enough information to be able to
perfectly reassemble message fragments post hoc given multiple
threads possibly interleaving KERN_CONT.

I do think the inconsistency of mixing styles with and without
newlines not particularly good.

cheers, Joe