From owner-freebsd-questions@FreeBSD.ORG Wed Mar 14 23:38:36 2007 Return-Path: X-Original-To: questions@freebsd.org Delivered-To: freebsd-questions@FreeBSD.ORG Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 9B06D16A404 for ; Wed, 14 Mar 2007 23:38:36 +0000 (UTC) (envelope-from Robert.Eckardt@Robert-Eckardt.de) Received: from mo-p00-ob.rzone.de (mo-p00-ob.rzone.de [81.169.146.162]) by mx1.freebsd.org (Postfix) with ESMTP id 3D11813C46C for ; Wed, 14 Mar 2007 23:38:35 +0000 (UTC) (envelope-from Robert.Eckardt@Robert-Eckardt.de) Received: from postit (frnk-590c4daf.pool.einsundeins.de [89.12.77.175]) by post.webmailer.de (mrclete mo12) (RZmta 5.1) with ESMTP id 901d3bj2EMNJ64 for ; Thu, 15 Mar 2007 00:27:27 +0100 (MET) Received: from www.home.roberte.eu (localhost.home.roberte.eu [127.0.0.1]) by postit (8.13.8/8.13.8) with ESMTP id l2ENRTWF019205 for ; Thu, 15 Mar 2007 00:27:29 +0100 (CET) (envelope-from Robert.Eckardt@Robert-Eckardt.de) From: "Robert Eckardt" To: questions@freebsd.org Date: Thu, 15 Mar 2007 00:27:29 +0100 Message-Id: <20070314215639.M99480@Robert-Eckardt.de> X-Mailer: OpenWebMail 2.52 20060502 X-OriginatingIP: 192.168.183.50 (roberte) MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 X-RZG-AUTH: kN+8IwndSMLtCuBFVkRfThJXBOX2xi0RUSsjpsWeoPRmTmyCUDrmirx0jJobKkIgjKZV X-RZG-CLASS-ID: mo00 Cc: Subject: Did I take the wrong bus with FreeBSD 6 to VMware? X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 14 Mar 2007 23:38:36 -0000 Hi, for some time I'm trying to get FreeBSD 6 running on my server as a host for VMware and several other functions. I'm using a 1.7GHz Pentium M 735 on an AOpen i855GMEm-LFS mobo w/ USB, VGA, 2xGbit/s, 2xPATA channels etc. on board. I used to run FBSD-5.2.1 with vmware3 on an Epox mobo w/ a 2GHz Celeron without problems. After changing HW (mobo, CPU, HDD) and OS (FBSD6.0) I found the system to "freeze" upon accessing an USB device when vmware was running. So my first investigations led to its driver, but in some cases heavy disk I/O was sufficient to cause a freeze. Since the situation got worse with FreeBSD 6.2 I started to work on it more systematically and found the following (actually I was on the verge to switch to Linux CentOS 4.4 or OpenSUSE 10.2 with VMware Server running nicely, but the HD and network performance were disappointing): 1)**ACPI off, "Assign USB IRQ" disabled in BIOS, vmware3 started: vmware3 runs fine, but no USB devices. 2)**ACPI off, "Assign USB IRQ" enabled in BIOS, vmware3 started: system "freezes" with network connections breaking, endless messages ad2: WARNING: - SETFEATURES SET TRANSFER MODE taskqueue timeout - completing request directly ad2: WARNING: - SETFEATURES SET TRANSFER MODE taskqueue timeout - completing request directly ad2: WARNING: - SETFEATURES ENABLE RCACHE taskqueue timeout - completing request directly ad2: WARNING: - SETFEATURES ENABLE WCACHE taskqueue timeout - completing request directly ad2: WARNING: - SET_MULTI taskq..... ad2: FAILURE [or TIMEOUT] - WRITE:DMA timed out [or retrying] LBA=.... g_vs_done():ad2s1e[WRITE(offset=...., length=....)]error = 5 typing reboot will finally reboot the system after several hours, nothing in the logs though. 3)**ACPI off, "Assign USB IRQ" enabled in BIOS, additional PCI-VGA card installed, using either PCI-VGA *or* on-board VGA, vmware3 started: vmware3 runs fine, also when accessing the USB device. 4)**ACPI on, "Assign USB IRQ" enabled in BIOS, additional PCI-VGA card installed, using on-board VGA, vmware3 started: system "freezes" with messages above. So, what's the relation between the scenarios? Where can I tweak the system to get it stable? Since I spend already several man-days on getting VMware running on my machine, I would like to help further debugging by making additional tests, but I don't know where to start. I can live without ACPI (for the time being) -- the old system consumes 125W while the Pentium M machine stays at 42W with ACPI taking about another 8W in idle-state. For me it seems essential why enabling/disabling USB in the BIOS or adding an additional PCI-VGA card stabilizes the system and why the unstable system behaves the same way like with enabling ACPI. I put some boot_verbose-logs on http://www.robert-eckardt.de/ghost/ Regards, Robert -- Dr. Robert Eckardt --- Robert.Eckardt@Robert-Eckardt.de