Date: Sun, 14 Mar 1999 09:02:20 +0200 From: Mark Murray <mark@grondar.za> To: Greg Lehey <grog@lemis.com> Cc: hackers@FreeBSD.ORG Subject: Re: Randomness and vnodes Message-ID: <199903140702.JAA02264@greenpeace.grondar.za> In-Reply-To: Your message of " Sun, 14 Mar 1999 17:15:29 %2B1030." <19990314171529.Z429@lemis.com> References: <199903131704.TAA97969@greenpeace.grondar.za> <19990314112807.K429@lemis.com> <199903140618.IAA64577@greenpeace.grondar.za> <19990314171529.Z429@lemis.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Greg Lehey wrote: > > For interrupts, the interrupt number is used; for the keyboard, scancodes; > > the available environmental stuff in the namei cache is centred around > > vnodes, so I'm looking for dirt in them. Heck - I may just xor the whole > > thing into an int to get some junk if necessary. > > I don't think xoring would be very random, given the number of fields > with predictable content. That's my main concern with using vnodes in > the first place, but if you want to use them, you should at least > divide and get a remainder. Ah! There is a theorem (sez Schneier) that says that if you xor two bits and only one of them is truly random, then the result is truly random. By xor'ing the lot, I shall get any randomness that is anywhere in the structure. However, I may lose some, and I may waste time if only a subset of the struct is random enough. M -- Mark Murray Join the anti-SPAM movement: http://www.cauce.org To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199903140702.JAA02264>