From owner-freebsd-ppc@FreeBSD.ORG Mon May 22 06:59:30 2006 Return-Path: X-Original-To: freebsd-ppc@freebsd.org Delivered-To: freebsd-ppc@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 16B3816A48C for ; Mon, 22 May 2006 06:59:30 +0000 (UTC) (envelope-from grehan@freebsd.org) Received: from dommail.onthenet.com.au (dommail.OntheNet.com.au [203.13.70.57]) by mx1.FreeBSD.org (Postfix) with ESMTP id 3018143D77 for ; Mon, 22 May 2006 06:59:28 +0000 (GMT) (envelope-from grehan@freebsd.org) Received: from [192.168.0.14] (dsl-63-249-90-35.cruzio.com [63.249.90.35]) by dommail.onthenet.com.au (MOS 3.5.7-GR) with ESMTP id BVM09734 (AUTH peterg@ptree32.com.au); Mon, 22 May 2006 16:59:22 +1000 (EST) Message-ID: <44716148.6060801@freebsd.org> Date: Sun, 21 May 2006 23:59:20 -0700 From: Peter Grehan User-Agent: Thunderbird 1.5 (Macintosh/20051201) MIME-Version: 1.0 To: Craig St Jean References: In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: freebsd-ppc@freebsd.org Subject: Re: I may be able to help with getting it on a G5 X-BeenThere: freebsd-ppc@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list Reply-To: grehan@freebsd.org List-Id: Porting FreeBSD to the PowerPC List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 22 May 2006 06:59:33 -0000 Hi Craig, > Not sure if anyone has tried working on getting FreeBSD on a G5 > machine, but I have some free time and would be happy to help test. I > have a quad 2.5GHz G5 PowerMac with plenty of memory and hard drive space. Great ! I'm still struggling to resurrect my PPC dev machines and environment, but when I do I'd like to get G5 support going. I'll take any and all offers of help :) > If someone gives me some direction on what may need to be > done, I would be happy to try to help make FreeBSD run on a G5 system - It's going to be tough. My plan is to get it running in 32-bit mode, since full 64-bit support requires a massive amount of work, essentially a new port. What's there: - the loader What's not there: - pmap support. The G5 uses 64-bit addresses in it's PTE format, so it's just different enough than the G3/G4 to be annoying. The code is already there to abstract out the differences, but a lot of implementation work has to be done. - the low-level assembler trap code needs to switch back to 32-bits. This isn't too difficult, and the plan was to conditionally compile the .S files similar to how ELF 32- and 64-bit support is done. - no BAT registers, so the kernel needs to have virtual mappings created for it. This complicates pmap startup, and various other parts of the the machine-dependent code that assumes a 1:1 static mapping for all memory (e.g. UMA_MD_SMALL_ALLOC, the syscons frame-buffer code, page zeroing) - cachelines are 128 bytes on the G5 as opposed to 32 bits on G3/G4. This is almost abstracted out but some extra work is required. - probably a bunch of other things I've forgotten. > Unless everyone feels its not worth the time... Definitely worth the time. later, Peter.