From owner-freebsd-bugs Sat Jun 24 07:20:49 1995 Return-Path: bugs-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id HAA09301 for bugs-outgoing; Sat, 24 Jun 1995 07:20:49 -0700 Received: (from dyson@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id HAA09292 ; Sat, 24 Jun 1995 07:20:47 -0700 From: John Dyson Message-Id: <199506241420.HAA09292@freefall.cdrom.com> Subject: Re: more on the paging thing To: dillon@best.com (Matt Dillon) Date: Sat, 24 Jun 1995 07:20:47 -0700 (PDT) Cc: bugs@freebsd.org, rdy@best.com In-Reply-To: <199506241155.EAA13277@shell1.best.com> from "Matt Dillon" at Jun 24, 95 04:55:27 am X-Mailer: ELM [version 2.4 PL24] Content-Type: text Content-Length: 998 Sender: bugs-owner@freebsd.org Precedence: bulk > > * run through ALL the pages in the active queue once every 5 seconds > testing pmap_is_referenced() and adjusting act_count up or down. > If the active count hits 0, insert the page at the head of the queue > If the page was referenced, move it to the tail. > 5 Seconds might be too often, and some of our iterations did do something like that. The major concern would be the scanning overhead on large systems, and the drain on CPU associated with it. > This would work quite well... you can optimize it by keeping a pointer > to the last page with a 0 act_count occuring before any page with > a non-0 act_count, and use that as the insertion point for new pages > that are decremented to 0, inserting them into the queue at that > point rather then at the head of the queue. > I'll experiment with it next time that I can get a chance. Your observations are quite astute -- and DG & I will look into them!!! John dyson@root.com