linux-kernel - Re: [PATCH 05/17] Blackfin: convert to generic checksum code

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:	Thu, 18 Jun 2009 21:19:41 -0400
From:	Mike Frysinger <vapier.adi@...il.com>
To:	Arnd Bergmann <arnd@...db.de>
Cc:	linux-kernel@...r.kernel.org,
	uclinux-dist-devel@...ckfin.uclinux.org
Subject: Re: [PATCH 05/17] Blackfin: convert to generic checksum code

On Mon, Jun 15, 2009 at 07:04, Arnd Bergmann wrote:
> On Sunday 14 June 2009, Mike Frysinger wrote:
>> The Blackfin port only implemented an optimized version of the
>> csum_tcpudp_nofold function, so convert everything else to the new
>> generic code.
>>
>> Signed-off-by: Mike Frysinger <vapier@...too.org>
>
> Have you tested this one well? I was as careful as possible with the
> version I added, but it was basically only tested on microblaze, which
> has a different endianess from blackfin. Some areas of the code may be
> sensitive to this.

i did some tests and it looks like do_csum() is broken :(

here's the input:
do_csum({ 0xff 0xfb 0x01 }, 3)

and here's the output:
Blackfin: 0xfc00
generic: 0xfcff

if i run the two funcs on my x86, i see similar behavior.  the
attached csum-test.c contains the csum code from lib/checksum.c and
arch/blackfin/lib/checksum.c and shows the problem.
-mike

View attachment "csum-test.c" of type "text/x-csrc" (1891 bytes)