From owner-freebsd-smp@FreeBSD.ORG Wed Apr 9 16:54:57 2003 Return-Path: Delivered-To: freebsd-smp@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id D8EA437B401; Wed, 9 Apr 2003 16:54:57 -0700 (PDT) Received: from bluejay.mail.pas.earthlink.net (bluejay.mail.pas.earthlink.net [207.217.120.218]) by mx1.FreeBSD.org (Postfix) with ESMTP id 469B843FA3; Wed, 9 Apr 2003 16:54:57 -0700 (PDT) (envelope-from tlambert2@mindspring.com) Received: from pool0163.cvx22-bradley.dialup.earthlink.net ([209.179.198.163] helo=mindspring.com) by bluejay.mail.pas.earthlink.net with asmtp (SSLv3:RC4-MD5:128) (Exim 3.33 #1) id 193PP5-0000fs-00; Wed, 09 Apr 2003 16:54:56 -0700 Message-ID: <3E94B26D.9C13C13E@mindspring.com> Date: Wed, 09 Apr 2003 16:53:17 -0700 From: Terry Lambert X-Mailer: Mozilla 4.79 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 To: John Baldwin References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-ELNK-Trace: b1a02af9316fbb217a47c185c03b154d40683398e744b8a41cd0bc3545dab10b210f300b95c4ae99350badd9bab72f9c350badd9bab72f9c350badd9bab72f9c cc: Aleksandr Melentiev cc: freebsd-smp@freebsd.org Subject: Re: System freezes with SMP support enabled X-BeenThere: freebsd-smp@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: FreeBSD SMP implementation group List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 09 Apr 2003 23:54:58 -0000 John Baldwin wrote: > > Second, it seems to me that there's a possibility for a deadlock > > if an interrupt comes in on one CPU, and an ithread to handle it > > is scheduled to run on a different CPU. You may want to try using > > SCHED_4BSD to see if that changes anything. > > Huh? Where in the code do you see this happening exactly? All the > bits you should need to look at for this are in ithread_schedule() > and ithread_loop() in sys/kern/kern_intr.c. Off-list, I have been helping someone debug an ithread_exit() bug, after which they can submit the patches if they want. I saw that there might be an issue if affinity is applied to ithread scheduling, if the interrupt is not routed to the CPU with the affinity. THis was a side issue, having nothing to do with the problem I was trying to help debug at the time, but it tickled my suspic-o-meter. > Not only that, but 4.8 doesn't have ithreads so I doubt seriously > that this is causing the lockups on 4.x. Yeah, my bad. I was only considering the case of his complaint on 5.0, since he complained about both 5.0 and 4.8, and 5.0 last, so that's what was in my recall buffer. 8-) 8-). -- Terry