linux-kernel - Re: kstrtojunk

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] [day] [month] [year] [list]

Message-ID: <20160704212629.GA20890@p183.telecom.by>
Date:	Tue, 5 Jul 2016 00:26:30 +0300
From:	Alexey Dobriyan <adobriyan@...il.com>
To:	Kees Cook <keescook@...omium.org>
Cc:	Andrew Morton <akpm@...ux-foundation.org>,
	Linux Kernel <linux-kernel@...r.kernel.org>
Subject: Re: kstrtojunk

On Fri, Jul 01, 2016 at 09:57:19AM -0400, Kees Cook wrote:
> On Fri, Jul 1, 2016 at 5:53 AM, Alexey Dobriyan <adobriyan@...il.com> wrote:
> > Could you please stop converting to kstrtobool()?
> >
> > commit a81a5a17d44b26521fb1199f8ccf27f4af337a67
> > Author: Kees Cook <keescook@...omium.org>
> > Date:   Thu Mar 17 14:22:57 2016 -0700
> >
> >     lib: add "on"/"off" support to kstrtobool
> >
> > Especially after doing patches like this?
> >
> > How on earth this was accepted?
> >
> > Now the kernel is supposed to know about every pair of words
> > with the opposite meaning and accept them.
> >
> > If kstrtobool() is ever going to be added it should accept
> > only '0' and '1' characters because kernel is not there
> > to second guess (same logic applies to whitespace trimming
> > for proc/sysfs files).
> >
> > Another point is that in C/C++ any value other than 0
> > is true in for bool but kstrtobool() doesn't accept, say '2'
> > as true. This is why it wasn't added in the first place.
> >
> > It is amazing to see how people think that every 2 common
> > lines of code should be generalized and pushed into lib/.
> 
> There were plenty of common use-cases of the 0/1/no/yes/off/on usage.

The mistake was to allow y/n/Y/N/yes/no/on/off in the first place.
Supporting anything but 0/1 by core code legitimizes those usages.
But since y/n/Y/N/... can't be removed, it is a mistake to move them
to core code, let it rot.

> I was specifically asked to use a common function, and this seemed
> like the cleanest approach. kstrtobool already took y/n, and hasn't
> ever handled "2" to mean "true", and is clearly documented. There's no
> reason it couldn't be adjusted to do that, though.

Other things:
* "!s" check, other functions don't do that,
* doesn't handle \n,
* can't spell "off" right,
* no tests,