Newsgroups: sci.crypt
Path: msuinfo!caen!sdd.hp.com!cs.utexas.edu!uunet!walter!qualcom.qualcomm.com!chicago.qualcomm.com!karn
From: karn@chicago.qualcomm.com (Phil Karn)
Subject: Re: FEAL?
Message-ID: <1992Jan14.065830.17336@qualcomm.com>
Sender: news@qualcomm.com
Nntp-Posting-Host: chicago.qualcomm.com
Organization: Qualcomm, Inc
References: <550@eouk9.abccam.abcl.co.uk>
Date: Tue, 14 Jan 1992 06:58:30 GMT
Lines: 13

In article <550@eouk9.abccam.abcl.co.uk> ahaley@eouk9.UUCP (Andrew Haley) writes:
>It's really easy to implement from the spec.  It's designed to be
>assembly coded and would be a real dog in C because it uses lots of 8
>bit rotates, which C cannot do in any reasonable way.

Sure it can. I just created a lookup table of 256 unsigned chars,
which I then initialize with ((n << 2) | (n >> 6)) for n from 0 to
255. Probably not as fast as using the byte-rotate operation in most
machines' instruction sets, but a lot better than executing the above
C code all the time.

Phil

