From owner-freebsd-smp Mon Dec 2 16:50:24 1996 Return-Path: owner-smp Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id QAA25006 for smp-outgoing; Mon, 2 Dec 1996 16:50:24 -0800 (PST) Received: from spinner.DIALix.COM (root@spinner.DIALix.COM [192.203.228.67]) by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id QAA24998 for ; Mon, 2 Dec 1996 16:50:19 -0800 (PST) Received: from spinner.DIALix.COM (peter@localhost.DIALix.oz.au [127.0.0.1]) by spinner.DIALix.COM (8.8.3/8.8.3) with ESMTP id IAA02845; Tue, 3 Dec 1996 08:49:46 +0800 (WST) Message-Id: <199612030049.IAA02845@spinner.DIALix.COM> To: "J.M. Chuang" cc: erich@lodgenet.com (Eric L. Hernes), smp@freebsd.org Subject: Re: Status report... In-reply-to: Your message of "Mon, 02 Dec 1996 18:58:22 -0400." <199612022258.SAA10867@bluenose.na.tuns.ca> Date: Tue, 03 Dec 1996 08:49:46 +0800 From: Peter Wemm Sender: owner-smp@freebsd.org X-Loop: FreeBSD.org Precedence: bulk "J.M. Chuang" wrote: > > I just tested a kernel from this morning (12/1). It is *much* better > > than the one I tried last week. Last week I got lots of sig11's, but > > I haven't seen one yet! But I am getting a few silo overflows > > from the mouse, dmesg shows: > > > I got the same thing. No sig11's. But when I run multiple processes, > it seems that all the processes are taken care of by the same cpu > and switched back and forth. On the screen of `top': [..] I know about some problems here, I bailed out last night before I fell asleep with a jumbo set of diffs to back out some more stuff again, given my .. umm... "deadly" ... track record of problems with commits while 99% asleep at the keyboard. If you were to time a 'make -j 8' or whatever, you'll currently notice that it is no faster than a 'make'.... Parallel scheduling is rather broken as a result of automatically starting the AP cpu's. I don't know why, but turning it off (again!) fixes it. > Besides, if the system is rebooted, I always got the message like: > > I'm on cpu #1, I need to be on cpu #0! Yes, according to the MPSPEC, the only cpu that is allowed to reboot or halt the system is the primary CPU. The reboot code has to try and shift itself onto cpu#0 if it starts elsewhere. > Jim Cheers, -Peter