linux-kernel - Re: [PATCH] kstrtox: make kstrtobool_from

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]

Date:   Sun, 11 Feb 2018 13:27:58 -0800
From:   Kees Cook <keescook@...omium.org>
To:     Alexey Dobriyan <adobriyan@...il.com>
Cc:     Andrew Morton <akpm@...ux-foundation.org>,
        LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] kstrtox: make kstrtobool_from_user() very strict

On Sat, Feb 10, 2018 at 1:11 PM, Alexey Dobriyan <adobriyan@...il.com> wrote:
> Once upon a time module parameter parsing code accepted
> 0, 1, y, n, Y and N for boolean values. Gratituous but contained
> to module code and thus tolerable.

And kernel command line.

> Commit ef951599074ba4fad2d0efa0a977129b41e6d203
> ("lib: move strtobool() to kstrtobool()") promoted that ugly wart
> to kstrtobool() and, more importantly, kstrtobool_from_user().
>
> Later set of accepted values was expanded to "on" and "of".
> Now there are 6+8=14(!) valid strings for a boolean.
>
> This patch reduces set of accepted values to "0" and "1"
> (with optional newline) in spirit with other kstrto*() functions.
>
> I'm starting with kstrtobool_from_user() as it is explicitly designed
> to be used for interacting with userspace. Currently there are 9 users
> all debug code, so there is hope.
>
> Please send before 4.16 so no real users start to depend on verbose behaviour.
>
> Signed-off-by: Alexey Dobriyan <adobriyan@...il.com>

Why bother? What's wrong with generous input parsing? This just
creates a divergence where none is needed.

-Kees

-- 
Kees Cook
Pixel Security