From owner-freebsd-current Thu Sep 11 12:31:25 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id MAA12151 for current-outgoing; Thu, 11 Sep 1997 12:31:25 -0700 (PDT) Received: from smtp.algonet.se (angel.algonet.se [194.213.74.112]) by hub.freebsd.org (8.8.7/8.8.7) with SMTP id MAA12145 for ; Thu, 11 Sep 1997 12:31:18 -0700 (PDT) Received: (qmail 16679 invoked from network); 11 Sep 1997 19:31:14 -0000 Received: from kairos.algonet.se (HELO kairos) (mal@194.213.74.18) by angel.algonet.se with SMTP; 11 Sep 1997 19:31:14 -0000 Received: (mal@localhost) by kairos (SMI-8.6/8.6.12) id VAA01882; Thu, 11 Sep 1997 21:31:14 +0200 Date: Thu, 11 Sep 1997 21:31:14 +0200 Message-Id: <199709111931.VAA01882@kairos> From: Mats Lofkvist To: current@FreeBSD.ORG Subject: Re: lousy disk perf. under cpu load (was IDE vs SCSI) Sender: owner-freebsd-current@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk The interrupt sharing of the ncr and ahc controllers indeed seemed to be the problem. With the ncr driver removed from the kernel, the ahc worked ok and now I'm running with them both but on separate interrupts and the bogus behaviour is gone. Convincing the Providence motherboard to use different irqs wasn't easy, I finally tricked it by telling it irq 10 (which it used to use for fxp0, ncr0 and ahc0) was needed for an isa card which made it use irq 9 for fxp0 and ncr0 and irq14 for ahc0. Conclusion: 2.2.2R doesn't like to have an ncr and an ahc on the same interrupt. Thanks for the help I got from all of you! _ Mats Lofkvist mal@algonet.se