From owner-cvs-all Fri Sep 15 18: 9:14 2000 Delivered-To: cvs-all@freebsd.org Received: from winston.osd.bsdi.com (winston.osd.bsdi.com [204.216.27.229]) by hub.freebsd.org (Postfix) with ESMTP id 3FDD837B423; Fri, 15 Sep 2000 18:09:11 -0700 (PDT) Received: from winston.osd.bsdi.com (jkh@localhost [127.0.0.1]) by winston.osd.bsdi.com (8.11.0/8.9.3) with ESMTP id e8G197698699; Fri, 15 Sep 2000 18:09:07 -0700 (PDT) (envelope-from jkh@winston.osd.bsdi.com) To: John Baldwin Cc: cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: cvs commit: src/sys/kern kern_idle.c src/sys/vm vm_meter.c src/sys/i386/isa ithread.c src/sys/sys proc.h In-Reply-To: Message from John Baldwin of "Fri, 15 Sep 2000 15:00:24 PDT." <200009152200.PAA79187@freefall.freebsd.org> Date: Fri, 15 Sep 2000 18:09:07 -0700 Message-ID: <98695.969066547@winston.osd.bsdi.com> From: Jordan Hubbard Sender: owner-cvs-all@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG > - Add a new process flag P_NOLOAD that marks a process that should be > ignored during load average calcuations. > - Set this flag for the idle processes and the softinterrupt process. Yet another reason why the concept of a "load average" simply bites to an increasing degree with any halfway rational OS, something Unix hasn't always been exactly on the money with. I'm not proposing the elimination of the "load average", it being sort of like the Linux BogoMIP in being so enshrined in lore as to be exempt from having to actually make any kind of sense. I'm simply wondering if we shouldn't start keeping separate bits of statistical information around for all idle, interrupt and user process information, perhaps as sysct variables. Then we can hack tools like systat to display more meaningful and varied information rather than relying to any further extent on increasingly misunderstood and meaningless metrics. - Jordan To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message