From owner-freebsd-stable@FreeBSD.ORG Wed Jul 23 19:57:26 2008 Return-Path: Delivered-To: freebsd-stable@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id DA9301065674; Wed, 23 Jul 2008 19:57:26 +0000 (UTC) (envelope-from ianjhart@ntlworld.com) Received: from mtaout03-winn.ispmail.ntl.com (mtaout03-winn.ispmail.ntl.com [81.103.221.49]) by mx1.freebsd.org (Postfix) with ESMTP id 4AC2E8FC19; Wed, 23 Jul 2008 19:57:25 +0000 (UTC) (envelope-from ianjhart@ntlworld.com) Received: from aamtaout01-winn.ispmail.ntl.com ([81.103.221.35]) by mtaout03-winn.ispmail.ntl.com with ESMTP id <20080723200355.VABX16629.mtaout03-winn.ispmail.ntl.com@aamtaout01-winn.ispmail.ntl.com>; Wed, 23 Jul 2008 21:03:55 +0100 Received: from cpc2-cove3-0-0-cust311.sol2.cable.ntl.com ([86.20.33.56]) by aamtaout01-winn.ispmail.ntl.com with ESMTP id <20080723200403.ELFJ16854.aamtaout01-winn.ispmail.ntl.com@cpc2-cove3-0-0-cust311.sol2.cable.ntl.com>; Wed, 23 Jul 2008 21:04:03 +0100 X-Virus-Scanned: amavisd-new at cpc2-cove3-0-0-cust311.sol2.cable.ntl.com Received: from gamma.private.lan (gamma.private.lan [192.168.0.12]) by cpc2-cove3-0-0-cust311.sol2.cable.ntl.com (8.14.2/8.14.2) with ESMTP id m6NJvFIE094689; Wed, 23 Jul 2008 20:57:15 +0100 (BST) (envelope-from ianjhart@ntlworld.com) From: ian j hart To: Jeremy Chadwick Date: Wed, 23 Jul 2008 20:57:15 +0100 User-Agent: KMail/1.9.7 References: <200807221727.52718.ianjhart@ntlworld.com> <200807221847.34935.ianjhart@ntlworld.com> <20080723001835.GA33136@eos.sc1.parodius.com> In-Reply-To: <20080723001835.GA33136@eos.sc1.parodius.com> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200807232057.15057.ianjhart@ntlworld.com> X-Spam-Status: No, score=-1.4 required=5.0 tests=ALL_TRUSTED autolearn=failed version=3.2.5 X-Spam-Checker-Version: SpamAssassin 3.2.5 (2008-06-10) on cpc2-cove3-0-0-cust311.sol2.cable.ntl.com Cc: freebsd-stable@freebsd.org Subject: Re: unable to boot 7.0-RELEASE cdrom on supermicro 5015b-mt X-BeenThere: freebsd-stable@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Production branch of FreeBSD source code List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 23 Jul 2008 19:57:27 -0000 On Wednesday 23 July 2008 01:18:35 Jeremy Chadwick wrote: > On Tue, Jul 22, 2008 at 06:47:34PM +0100, ian j hart wrote: > > On Tuesday 22 July 2008 17:37:24 Jeremy Chadwick wrote: > > > On Tue, Jul 22, 2008 at 05:27:52PM +0100, ian j hart wrote: > > > > Same hardware as my other thread. > > > > http://www.supermicro.com/products/system/1U/5015/SYS-5015B-MT.cfm > > > > > > > > [using 2Gb RAM and SATA in legacy mode] > > > > > > > > I'd like to focus only on making the CDROM boot complete. > > > > > > > > Summary: hangs just after the CPUs are launched. > > > > > > > > 6.2-RELEASE works okay, no AHCI support > > > > 6.3-RELEASE works okay > > > > 7.0-RELEASE hangs > > > > 7.0-STABLE-200806-SNAPSHOT hangs > > > > 8.0-CURRENT-200806-SNAPSHOT hangs > > > > > > > > I thought I could do a binary search using the current snapshot > > > > boot-only CDs but they only go back to March. Are there any older > > > > ones available? > > > > > > Have you tried disabling ACPI to see if it makes any sort of > > > difference? > > > > Yes, but I'm happy to re-try. > > > > Which method is "best"? Or is it 1 + 2 or 3? > > > > 1) BIOS > > 2) Beastie menu option > > 3) loader prompt set hint > > Item #2 is the easiest. You should really be able to leave the BIOS > settings at their defaults (Factory Defaults) and have this system work > on FreeBSD. > > Items #2 and #3 are the same. The loader menu option for disabling ACPI > simply sets the hint. > > > > Also, AHCI should work just fine on those systems -- I know because I > > > have fairly extensive experience with Supermicro hardware, although > > > what you're using is newer than what I presently have. I don't know > > > why you're setting Compatible/Legacy mode on your controller (you > > > mention doing this in your other thread as well). > > > > Because I don't know what's wrong yet and AHCI support is newer than SATA > > support and this is a newish board? [At least 6.2 doesn't seem to support > > it and it has an AHCI legacy option!] > > > > I'd be happy to swap this over. Slight problem; the drives get > > renumbered, so I'd rather not swap back and forth. > > You *absolutely* should have AHCI enabled. There's a lot of reasons > why, too. I highly recommend avoiding the "SATA Compatible" mode. > > AHCI should work fine on FreeBSD 6.3 as well as 7.0 -- I know, because > we have many Supermicro boards running those versions which do have AHCI > enabled. Please use it, and stick with it. > > Here's added proof that AHCI works fine on 6.3: > > $ dmesg -a | grep -i ahci > atapci1: port > 0x30e8-0x30ef,0x30dc-0x30df,0x30e0-0x30e7,0x30d8-0x30db,0x30b0-0x30bf mem > 0xe0000400-0xe00007ff irq 19 at device 31.2 on pci0 atapci1: AHCI Version > 01.10 controller with 4 ports detected > $ uname -r -s > FreeBSD 6.3-PRERELEASE > > The adX device renumbering is expected. There are workarounds for this, > but I recommend you simply enable AHCI. Do not keep toggling it on/off. > > > > Below is what we use on our systems; factory defaults, then make the > > > following changes. (The G-LAN1 OPROM option you can do whatever you > > > want with -- it's specific to our environment). > > > > > > * Main > > > * Date > > > --> Set to GMT, not local time!!! > > > * Serial ATA > > > --> SATA Controller Mode --> Enhanced > > > --> SATA AHCI --> Enabled > > > > > > * Advanced > > > * Boot Features > > > --> Quiet Mode --> Disabled > > > --> Enable Multimedia Timer --> Enabled > > > * PCI Configuration > > > --> Onboard G-LAN1 OPROM --> Disabled > > > --> Large Disk Access Mode --> Other > > > * Advanced Processor Options > > > --> Intel(R) Virtualization Technology --> Enabled > > > --> C1 Enhanced Mode --> Enabled > > > > I've got as close as I can to this. > > > > This board also has an AHCI legacy option [disabled] which hides ports 5 > > and 6. I also disabled quickboot and POST errors. I assume multimedia > > timer is the same as HPET. Doesn't seem to be any disk translation > > option. I took the fans off 'flat out'. > > Okay, I've had a chance to review the board manual that comes with the > X7SBi. You should set the following: > > Serial ATA: Enabled > Native Mode Operation: Serial ATA > SATA AHCI: Enabled > SATA AHCI Legacy: Disabled > > The name "SATA AHCI Legacy" a horrible name for what it does. The ICH9 > itself has support for 6 SATA ports, but (if I remember correctly, based > on reading some Intel design documents) there are extra registers you > have to tweak to get those ports to work, and the OS has to be fully > aware of how to do that. The BIOS option simply disables SATA ports 5 > and 6 altogether; the underlying OS never sees them. I'd recommend > keeping that setting Disabled (the default) unless you have disks on > those ports (I don't see how, since it's a 4-disk system!). > > I don't think this option is what's causing you problems, though. > > "Multimedia Timer" is indeed HPET. Looks like they changed the name to > be more reflective of what it actually is. > > The "Large Disk Access" mode does appear to be missing from that BIOS, > probably for a good reason. I can enable/disable it on our boards with > no repercussions (the options are "DOS" and "Other", which is why I > choose "Other"). I'm not entirely sure what it does. > > As for your problem... > > If the CDROM is the problem (which would be odd, since the disc does > boot and load the kernel successfully), can you try going into the BIOS > and setting IDE Channel 0 Master (which I think is the CDROM -- I could > be wrong here) and set "Transfer Mode" to PIO1 and "Ultra DMA Mode" to > Disabled? > > I have a feeling the problem isn't related to the CDROM, but I'm not > entirely sure how to debug it. I tried each BIOS setting in turn. No joy. I tried a 40 wire cable. No joy I tried an old CDROM (Mitsumi FX4830T circa Feb 2001). No joy Had a feeling that the drive always came up as UDMA33. Tried setting hw.ata.atapi_dma=0 Wahay! The drive throws errors but it boots. Did an install, and that hangs too without this setting in /boot/loader.conf. Frankly I'm kicking myself for not trying this earlier. That's ~16 hours I'll never get back. OTOH this has been the default for a long time now. Thank you Jeremy, I wouldn't have got there without you. > > There are other users using the X7SBi with success: > > http://groups.google.com/group/mailing.freebsd.current/browse_thread/thread >/d0a2d20f8965361a http://www.webhostingtalk.com/showthread.php?t=666686 > > Also, can you make sure your BIOS revision is 1.1a, just to rule out any > BIOS-related issues? -- ian j hart