From owner-freebsd-ppc@FreeBSD.ORG Sun Jul 27 19:24:30 2014 Return-Path: Delivered-To: freebsd-ppc@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [8.8.178.115]) (using TLSv1 with cipher ADH-AES256-SHA (256/256 bits)) (No client certificate requested) by hub.freebsd.org (Postfix) with ESMTPS id 17933423 for ; Sun, 27 Jul 2014 19:24:30 +0000 (UTC) Received: from asp.reflexion.net (outbound-245.asp.reflexion.net [69.84.129.245]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client did not present a certificate) by mx1.freebsd.org (Postfix) with ESMTPS id 94F472ABC for ; Sun, 27 Jul 2014 19:24:29 +0000 (UTC) Received: (qmail 2156 invoked from network); 27 Jul 2014 19:24:27 -0000 Received: from unknown (HELO mail-cs-04.app.dca.reflexion.local) (10.81.19.4) by 0 (rfx-qmail) with SMTP; 27 Jul 2014 19:24:27 -0000 Received: by mail-cs-04.app.dca.reflexion.local (Reflexion email security v7.30.6) with SMTP; Sun, 27 Jul 2014 15:24:27 -0400 (EDT) Received: (qmail 18170 invoked from network); 27 Jul 2014 19:24:26 -0000 Received: from unknown (HELO iron2.pdx.net) (69.64.224.71) by 0 (rfx-qmail) with (DHE-RSA-AES256-SHA encrypted) SMTP; 27 Jul 2014 19:24:26 -0000 X-No-Relay: not in my network X-No-Relay: not in my network X-No-Relay: not in my network Received: from [192.168.0.108] (ip98-160-185-241.lv.lv.cox.net [98.160.185.241]) by iron2.pdx.net (Postfix) with ESMTPSA id 2B07B1C4388; Sun, 27 Jul 2014 12:24:22 -0700 (PDT) Mime-Version: 1.0 (Mac OS X Mail 7.3 \(1878.6\)) Subject: Re: Running FreeBSD-ppc/ppc64 on emulators? [.iso based CD boot problems exist for PowerMac G5's too] From: Mark Millard In-Reply-To: <20140727094457.4827b851@zhabar.att.net> Date: Sun, 27 Jul 2014 12:24:23 -0700 Message-Id: <96A3A9F6-EB28-47CD-A3EF-903704DFF3FE@dsl-only.net> References: <9104D42E-5A30-49AF-B5B8-F2C97090802F@dsl-only.net> <0E217B0D-233B-4698-B482-75C3BFE66638@dsl-only.net> <53D49EDC.3050301@freebsd.org> <27E2583C-FA8E-4F6E-83D1-C950318133C2@dsl-only.net> <20140727094457.4827b851@zhabar.att.net> To: Justin Hibbits , freebsd-ppc@freebsd.org X-Mailer: Apple Mail (2.1878.6) Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: quoted-printable X-Content-Filtered-By: Mailman/MimeDel 2.1.18 X-BeenThere: freebsd-ppc@freebsd.org X-Mailman-Version: 2.1.18 Precedence: list List-Id: Porting FreeBSD to the PowerPC List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 27 Jul 2014 19:24:30 -0000 Hi Justin. The 3 G5 contexts that I've been using are: 8GB dual = single-core-processor, 12GB two dual-core processors (quad), 16 GB two = dual-core processors (quad). So "yes" for the amount of RAM being > 4GB = where I've seen the problem. I've not been running anything but the default pre-built kernels = --installed by some 10.0-STABLE build primarily. (Otherwise = 9.2-STABLE/early-9.3 or 10.0-RELEASE.) And I'm away from the PowerMac's = for weeks and so will not be trying any variations for some time. While I have done buildworld and buildkernel (without making any = changes) I've yet to try installing the result. Side note: As I remember one of the things that Mac OS X 10.5 does (so probably = Darwin too?) is to write some failure/traceback information into NVRAM = and if it later boots okay it notices the updated NVRAM log contents, = extracts the failure log information, and then presents the option to = see it and/or send it in. (Not that Apple cares about sending such in = any more.) It seems to have been their one public hook for getting = information from time frames where nothing at-the-time would display = about a crash that occurred while also not needing general writable = media to be in operation at the time: NVRAM is always present. But I'm = not sure just how early Mac OS X starts using this technique of storing = away such information for later extraction. =3D=3D=3D Mark Millard markmi@dsl-only.net On Jul 27, 2014, at 9:44 AM, Justin Hibbits = wrote: Hi Mark, Just a guess, but you probably have more than 4GB RAM in those machines, right? If so, then you hit the same problem I've encountered on my G5, which has 12GB RAM. Nathan and I both speculate that it's dropping into Open Firmware (we make extensive use of OFW), and then messing something up, taking a page fault or something. Off and on we're working on the problem, taking the solution that Linux and Darwin use, to cache all the OFW properties while in the loader, and hand this off to the kernel (in our case, we'd format it as FDT). If you have a debug kernel, you would see "KDB: Supported backend...." and then it would hang. - Justin On Sun, 27 Jul 2014 00:33:02 -0700 Mark Millard wrote: > You are welcome. >=20 > You may run into another G5 (PowerPC64) problem that I've not > reported because I've so little evidence and it has been very > intermittent. Also it is a normal boot issue, not an > installation-media specific issue. >=20 > Normally one goes through open firmware and either types boot or one > lets it autoboot (explicit return or timed out) and the screen > changes to black and the white text starts to show up, starting with > the copyright notice. >=20 > But some of the time on the G5's it instead hangs after going black > but before any part of the copyright text shows up. This has happened > on all 3 G5's, two being 7,11's (NVIDIA video) and one being a 7,2 > (Radeon video). The fans gradually wind up when this happens. > (Unlikely if only the video display was messed but it actually went > through the normal steps otherwise.) >=20 > I've had a wild mix: At times I've gone bunches of boots without the > problem. On rare occasion I've had to reboot up to a dozen times to > finally not have it hang. Normally once the problem happens I have to > retry at most a few times. >=20 > [I do not have access to a single-processor PowerMac G5 system. So I > can not compare/contrast behavior with those. All my PowerMac G5 > notes are multi-processor G5 notes.] >=20 >=20 >=20 >=20 > =3D=3D=3D > Mark Millard > markmi at dsl-only.net >=20 > On Jul 26, 2014, at 11:40 PM, Nathan Whitehorn > wrote: >=20 > Interesting. It seems loader is broken. I just tried with a > 10.0-STABLE image and can reproduce the problem -- 11 oddly is OK. > What really perplexes me is that the powerpc and powerpc64 loaders > are identical, since loader is built 32-bit in either case. Super > weird. I'll try to get it fixed. Thanks for the problem report and > reminder! -Nathan >=20 > On 07/26/14 23:33, Mark Millard wrote: >>=20 >> To make clear about the "/" in the line: >>=20 >> and stops while displaying "/" from its progress indicator. The >> fans then gradually spin up. >>=20 >> This "progress indicator" is the repeating sequence -, \, |, /. It >> always stopped with / as I remember. But I've no clue for if that >> is from the repeating sequence or is the leading / >> from /boot/kernel/kernel. At this point I'm limited to reporting >> just the visual result, not the internal stage it is tied to. >>=20 >> =3D=3D=3D >> Mark Millard >> markmi at dsl-only.net >>=20 >> On Jul 26, 2014, at 7:53 PM, Mark Millard >> wrote: >>=20 >> First I quote what I wrote about when the hangups occur on the G5's >> (combining quotes from the two messages so the overall notes are >> easier to understand): >>=20 >>> For a C-key style boot it hangs shortly after displaying >>>=20 >>> loading /boot/defaults/loader.conf >>>=20 >>> and stops while displaying "/" from its progress indicator. The >>> fans then gradually spin up. >>>=20 >>> In other words: It hangs just before it would normally display the >>> line that includes the text: /boot/kernel/kernel >>=20 >> The original messages were: >>=20 >> http://lists.freebsd.org/pipermail/freebsd-ppc/2014-June/007029.html >> and >> http://lists.freebsd.org/pipermail/freebsd-ppc/2014-June/007056.html >>=20 >> Even the /b prefix from /boot/kernel/kernel did not show up even >> though /boot/defaults/loader.conf did show up just fine. >>=20 >>=20 >>=20 >> As for trying 11-CURRENT: I've not tried 11-CURRENT for anything >> yet. And, unfortunately, I'm now away from the PowerMacs for weeks >> so I will not be doing new experiments for several weeks. I can try >> once I get back. >>=20 >> All the 9.2/9.3-STABLE's that I tried worked. And 10.0-RELELASE >> worked. Any 10.0-STABLE from before about Apr-27 I've never tried: >> Late April is about when I started the experiments with PowerMacs. >>=20 >> Also: I had no troubles with 10.0-STABLE for powerpc on PowerMacs. >> PowerPC64 on PowerMac G5's is where I had problems. (I do not have >> other kinds of PowerPC equipment available.) >>=20 >>=20 >>=20 >> =3D=3D=3D >> Mark Millard >> markmi at dsl-only.net >>=20 >> On Jul 26, 2014, at 2:23 PM, Mark Millard >> wrote: >>=20 >> This boot issue from .iso's (burned to CDs) is not specific to >> emulation (assuming it is the same problem in both places). >>=20 >> I'm afraid that it sounds like PowerMac G5's hang the same >> way/place whenever I've tried a fairly modern PowerPC64 >> 10.0-STABLE .iso burned to a CD. I was able to boot >> 10.0-RELEASE's .iso burned to a CD. I've not tried emulation. >>=20 >> Having an initial FreeBSD running I tend to use the MANIFEST >> and .txz files with "bsdinstall auto" to make new from-scratch (not >> buildworld/buildkernel) installations when I want such. >>=20 >> My notes on the G5 boot issue for the .iso's burned to CD's were >> reported in: (I'm not near the PowerMacs, unfortunately) >>=20 >> FreeBSD 10-STABLE powerpc64 r266807 CD PowerMac7, 2 (and 7, 11) >> boot hangs very early in the process >>=20 >> Has anyone booted a G5 PowerMac from a recent 10.0-STABLE .iso >> burned to CD? A multi-processor one? >>=20 >> =3D=3D=3D >> Mark Millard >> markmi at dsl-only.net >>=20 >>=20 >>=20 >=20 >=20 > _______________________________________________ > freebsd-ppc@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-ppc > To unsubscribe, send any mail to "freebsd-ppc-unsubscribe@freebsd.org"