[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <20120628201208.GA24838@openwall.com>
Date: Fri, 29 Jun 2012 00:12:08 +0400
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: WPA-PSK big-endian fix
magnum, Lukas -
The attached patch makes the WPA-PSK format (CPU) work on big-endian
(tested with macosx-ppc32-altivec-cross). Please apply to magnum-jumbo.
BTW, we need a second test vector, which should be for the keyver == 1
case. When our only test vector is for keyver != 1, the test passes
regardless of the endianness of keyver, even though I think the code was
not actually right for big-endian, and a swap of keyver was needed (added
with this patch).
The byte-swap of eapol_size was definitely needed - I was getting a "Bus
error" without it since a huge eapol_size was passed to HMAC().
In fact, it could be best to have two SHA-1 test vectors and one MD5, so
that the MD5 one is not used for benchmarks (only the first two salts
are used). This affects wpapsk_postprocess() only, though, so the
performance difference should be minimal. The main loop uses SHA-1
either way.
Alexander
View attachment "john-wpapsk-bigendian.diff" of type "text/plain" (629 bytes)
Powered by blists - more mailing lists
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
