linux-kernel - Re: [PATCH v5 00/11] vsprintf: Prevent silent crashes and consolidate error handling

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:   Fri, 27 Apr 2018 16:10:57 +0200
From:   Petr Mladek <pmladek@...e.com>
To:     Andy Shevchenko <andriy.shevchenko@...ux.intel.com>,
        Rasmus Villemoes <linux@...musvillemoes.dk>
Cc:     Linus Torvalds <torvalds@...ux-foundation.org>,
        "Tobin C . Harding" <me@...in.cc>, Joe Perches <joe@...ches.com>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Michal Hocko <mhocko@...e.cz>,
        Sergey Senozhatsky <sergey.senozhatsky@...il.com>,
        Steven Rostedt <rostedt@...dmis.org>,
        Sergey Senozhatsky <sergey.senozhatsky.work@...il.com>,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v5 00/11] vsprintf: Prevent silent crashes and
 consolidate error handling

On Wed 2018-04-25 13:12:40, Petr Mladek wrote:
> Crash in vsprintf() might be silent when it happens under logbuf_lock
> in vprintk_emit(). This patch set prevents most of the crashes by probing
> the address. The check is done only by %s and some %p* specifiers that need
> to dereference the address.
> 
> Only the first byte of the address is checked to keep it simple. It should
> be enough to catch most problems.
> 
> The check is explicitly done in each function that does the dereference.
> It helps to avoid the questionable strchr() of affected specifiers. This
> change motivated me to do some preparation patches that consolidated
> the error handling and cleaned the code a bit.
> 
> I did my best to address the feedback. Note that there is still the
> (efault) error message. But it is accompanied with WARN() when
> panic_on_warn is not enabled. I hope that it makes it more acceptable.
> 
> 
> Changes against v4:
> 
> 	+ rebased on top of
> git://git.kernel.org/pub/scm/linux/kernel/git/pmladek/printk.git for-4.18
> 	+ Added missing conts into ptr_to_ind() in a separate patch
> 	+ Renamed __string to valid_string()
> 	+ Avoid WARN() for invalid poimter specifiers
> 	+ Removed noinline_for_stack where it was not really useful
> 	+ WARN() when accessing invalid non-NULL address

Thanks a lot everyone for feedback. I'll incorporate it into v6. It
might take some time.

BTW: I also got report from 0day robot about that the size of vmlinux
increased by 545 bytes in i386-tinyconfig. I guess that it is mainly
because all the copies of

	if (!valid_pointer_access(&buf, end, bdev, spec))
		return buf;

got inlined. I guess that I would need to address it somehow
as well.

Best Regards,
Petr