From owner-freebsd-sparc64@FreeBSD.ORG Mon Jan 18 19:32:24 2010 Return-Path: Delivered-To: freebsd-sparc64@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 938E91065676 for ; Mon, 18 Jan 2010 19:32:24 +0000 (UTC) (envelope-from marius@alchemy.franken.de) Received: from alchemy.franken.de (alchemy.franken.de [194.94.249.214]) by mx1.freebsd.org (Postfix) with ESMTP id 01DBA8FC0C for ; Mon, 18 Jan 2010 19:32:23 +0000 (UTC) Received: from alchemy.franken.de (localhost [127.0.0.1]) by alchemy.franken.de (8.14.3/8.14.3/ALCHEMY.FRANKEN.DE) with ESMTP id o0IJWMIf003077; Mon, 18 Jan 2010 20:32:22 +0100 (CET) (envelope-from marius@alchemy.franken.de) Received: (from marius@localhost) by alchemy.franken.de (8.14.3/8.14.3/Submit) id o0IJWMN5003076; Mon, 18 Jan 2010 20:32:22 +0100 (CET) (envelope-from marius) Date: Mon, 18 Jan 2010 20:32:22 +0100 From: Marius Strobl To: Kaya Saman Message-ID: <20100118193222.GL18141@alchemy.franken.de> References: <4B549896.8090309@netscape.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <4B549896.8090309@netscape.net> User-Agent: Mutt/1.4.2.3i Cc: freebsd-sparc64@freebsd.org Subject: Re: Sun Fire V480 compatibility question? X-BeenThere: freebsd-sparc64@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Porting FreeBSD to the Sparc List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 18 Jan 2010 19:32:24 -0000 On Mon, Jan 18, 2010 at 07:21:26PM +0200, Kaya Saman wrote: > Hi guys and Marius, (sorry for the cc but you did bring up the topic so > I'm figuring that you probably know the most about this subject!) > > I am just in the process of figuring out which Sun bridge I will need > for the Fire v480 server as a quote from Marius Strobl suggests this: > > [quote] > > FreeBSD currently crashes on older models of V480 when attempting > to use an on-board NIC due to what appears to be a CPU bug which > needs to be worked around. It appears to work fine on later V480 > equipped with version 7 Schizo bridges (that's just an indicator > and unlikely related to the problem) though, but FreeBSD 7.3 will > be the first release (besides the current stable branches) to work > with their changed firmware behavior. Linux doesn't seem to do any > better in this regard: > http://marc.info/?l=linux-sparc&m=122220796209509&w=2 > http://lists.debian.org/debian-sparc/2009/12/msg00012.html > > Marius > > [quote] > > My supplier has given me an extract from the Sun handbook which doesn't > seem to show version 7: > > > A37 > > 501-5819 501-6733 > November 2001 > <=501-5819-12: SCHIZO 2.2 > >=501-5819-13: SCHIZO 2.5 August 2003 > reworked 501-5819-12 > SCHIZO 2.2 > only available from > Sun Services > 501-6780 501-6790 > December 2003 > FAB 270-5819-06 > SCHIZO 2.5 April 2004 > FAB 270-5819-07 > SCHIZO 2.5 > or ELE 1.1+ > > > A bit of Google'ing however got me to this: > > [quote] > Version /Schizo/ Rev /7/ 2.5 6 2.4 5 2.3 4 2.2. Symptoms. Should the > described issue occur, /Sun Fire/ 280R, /V480/, V880 and Netra 20 > systems may encounter a FATAL *... > *[/quote] > > From here: > http://sunsolve.sun.com/search/document.do?assetkey=1-66-200502-1 > > also this: > > [quote] > > * FreeBSD has been adopted to the changed firmware of newer Sun Fire > V480 (those equipped with version 7 Schizo bridges) and has been > reported to now run fine on these. The necessary change will be > part of 7.3-RELEASE. Unfortunately, using the on-board NICs in > older models of Sun Fire V480 (at least those equipped with version > 4 Schizo bridges) under FreeBSD still leads to the firmware issuing > a FATAL RESET due to what appears to be a CPU bug, which needs to > be worked around. > > > [/quote] > > Taken from the Sparc64 part of this page: > http://marc.info/?l=freebsd-hackers&m=126375029221240&w=2 > > What I can understand from all of this is that the version 7 is actually > a revision 7.... so any of the bus versions listed above in the table > will work as long as they are revision 7 or later? > > Is this the case, have I understood correctly?? > It's the other way around, version 4 and 7 refer to the contents of the "version#" property of these bridges in the Open Firmware device tree, which can be translated into 2.x revision numbers for exampled based on the table present in the Schizo Errata [1], i.e. version 7 translates to revision 2.5 and version 4 translates to revision 2.2. The sunsolve document you refer to actuall is in line with this information. I highly suspect this problem is related to the infamous BugID 4898531, just that Solaris can work around it to some extent or that FreeBSD happens to better a triggering it (I actually have a version of cas(4) which makes it less likely but still doesn't work around it completely), so V480 equipped with 501-6780 and 501-6790 centerplanes should be fine while those with 501-5819 and 501-6733 are prone to fail. Marius 1: http://www.sun.com/processors/manuals/External_Schizo_Errata.pdf