Date: Tue, 16 Sep 1997 01:49:24 +0200 From: j@uriah.heep.sax.de (J Wunsch) To: hackers@FreeBSD.ORG Cc: djardine@hotmail.com (Douglas Jardine) Subject: Re: LRU implementation Message-ID: <19970916014924.RB17838@uriah.heep.sax.de> In-Reply-To: <19970915125503.19344.qmail@hotmail.com>; from Douglas Jardine on Sep 15, 1997 05:55:03 -0800 References: <19970915125503.19344.qmail@hotmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
As Douglas Jardine wrote: > I had a question on the implementation of Global-LRU in FreeBSD: > What exact implementation does it employ - CLOCK, 2-handed CLOCK > or K-bit LRU? > > The 4.3BSD book says that 4.3BSD uses 2-handed CLOCK but the 4.4BSD > book is silent on this topic. Did FreeBSD diverge from 4.4BSD in > this aspect? Does 4.4BSD use a 2-handed clock too? My opinion on this is not very authoritative, but if i remember the 4.3BSD daemon book correctly, the two-handed clock (*) was just a hack that was required on the Vax. The Vax MMU doesn't have a `page referenced' bit, so it had to be simulated by first invalidating the page (1st clock hand), and then look which pages weren't reclaimed (2nd hand). The i386 architecture does have a referenced bit. (*) I've unfortunately only got the (terrible) German translation of the 4.3BSD book. They indeed translated this into ``zweihändige Uhr''. :-O -- cheers, J"org joerg_wunsch@uriah.heep.sax.de -- http://www.sax.de/~joerg/ -- NIC: JW11-RIPE Never trust an operating system you don't have sources for. ;-)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19970916014924.RB17838>