Date: Fri, 15 Sep 2000 18:43:35 -0700 (PDT) From: John Baldwin <jhb@pike.osd.bsdi.com> To: Greg Lehey <grog@lemis.com> Cc: John Baldwin <jhb@FreeBSD.org>, 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 Message-ID: <200009160143.SAA78215@pike.osd.bsdi.com> In-Reply-To: <20000916110946.D71517@wantadilla.lemis.com> from Greg Lehey at "Sep 16, 2000 11:09:46 am"
next in thread | previous in thread | raw e-mail | index | archive | help
Greg Lehey wrote: > On Friday, 15 September 2000 at 15:00:24 -0700, John Baldwin wrote: > > jhb 2000/09/15 15:00:24 PDT > > > > Modified files: > > sys/kern kern_idle.c > > sys/vm vm_meter.c > > sys/i386/isa ithread.c > > sys/sys proc.h > > Log: > > - 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. > > Why for the soft interrupt processes? Unlike the idle process, that's > real processor work. It should count towards interrupt time. Because the softinterrupt process is always running when this is calculated, it's deadtime is never taken into account. Since the softinterrupt is idle more than it is running, it is a greater error to assume it is running all the time than to assume it is never running. We need as accurate a load average as we can get as otherwise the scheduler tends to thrash around a bit because we punish CPU intensive processes more than we should due to the inflated load. -- John Baldwin <jhb@bsdi.com> -- http://www.FreeBSD.org/~jhb/ "Power Users Use the Power to Serve!" - http://www.FreeBSD.org/ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200009160143.SAA78215>