linux-kernel - Re: [PATCH next 4/8] test_hexdump: Check for buffer overrun of sample output buffer

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: <20250312193724.294c84b2@pumpkin>
Date: Wed, 12 Mar 2025 19:37:24 +0000
From: David Laight <david.laight.linux@...il.com>
To: Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
Cc: linux-kernel@...r.kernel.org, Andrew Morton <akpm@...ux-foundation.org>,
 Arnd Bergmann <arnd@...db.de>, Linus Torvalds
 <torvalds@...ux-foundation.org>, Christophe Leroy
 <christophe.leroy@....fr>, Rasmus Villemoes <linux@...musvillemoes.dk>,
 nnac123@...ux.ibm.com, horms@...nel.org
Subject: Re: [PATCH next 4/8] test_hexdump: Check for buffer overrun of
 sample output buffer

On Mon, 10 Mar 2025 11:01:29 +0200
Andy Shevchenko <andriy.shevchenko@...ux.intel.com> wrote:

> On Sat, Mar 08, 2025 at 09:34:48AM +0000, David Laight wrote:
> > While the output generated by test_hexdump_prepare_test() shouldn't
> > be longer than the size of the buffer passed, for safety verify that
> > the buffer is long enough.
> > If too short fill the buffer with an error message - output on
> > test failure.  
> 
> Isn't the function should behave snprintf() alike?
> I think this patch is simply wrong because it's based on a wrong assumption.
> 

The normal hex_dump_to_buffer() is snprintf() like.

But, as the tests are written, test_hexdump_prepare_test() is always
passed a fixed size buffer that should be big enough.
The test control code then truncates the output to the required length
and fills the rest of the buffer with '#' before the compare.

The 'testlen' (output buffer length) parameter is actually ignored.
So if someone adds a test that would request output longer than the
buffer (actually requires the 'rowsize' limit be relaxed) then the
test code overruns the on-stack buffer.
While the control code could be rewritten that would be more changes.
So it seemed best just to force the test to fail and make it obvious why.

	David