From owner-freebsd-stable@FreeBSD.ORG Wed Oct 1 11:24:51 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 3C6921065688; Wed, 1 Oct 2008 11:24:51 +0000 (UTC) (envelope-from mirya@zoc.com.ua) Received: from zoc.com.ua (zoc.com.ua [82.144.212.13]) by mx1.freebsd.org (Postfix) with ESMTP id B48B98FC23; Wed, 1 Oct 2008 11:24:49 +0000 (UTC) (envelope-from mirya@zoc.com.ua) Received: from [192.168.0.22] (port=50531) by zoc.com.ua with esmtp (Exim 4.62) (envelope-from ) id 1Kkzoq-0001b8-H5; Wed, 01 Oct 2008 14:24:36 +0300 From: Kyryll A Mirnenko aka Mirya Organization: zoc To: Roland Smith Date: Wed, 1 Oct 2008 14:24:01 +0300 User-Agent: KMail/1.9.4 References: <200809301454.47473.mirya@zoc.com.ua> <20080930173202.GA16426@slackbox.xs4all.nl> In-Reply-To: <20080930173202.GA16426@slackbox.xs4all.nl> MIME-Version: 1.0 Content-Type: text/plain; charset="koi8-u" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200810011424.02903.mirya@zoc.com.ua> X-CHECK-DB: NO Cc: Pawel Jakub Dawidek , freebsd-stable@freebsd.org Subject: Re: GELI partition mount on boot fails after 7.0 -> 7.1-PRERELEASE upgrade 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, 01 Oct 2008 11:24:51 -0000 On Tuesday 30 September 2008 20:32, Roland Smith wrote: > My GELI encrypted home partition works fine on amd64 7.1-PRERELEASE > (updated september 25th). I've been tracking stable since 7.0-RELEASE > without problems. > > My custom kernel includes GEOM_ELI, GEOM_LABEL, GEOM_MIRROR and > GEOM_PART_GPT and uses SCHED_ULE. Filesystem options are FFS, > SOFTUPDATES, UFS_ACL and UFS_DIRHASH. The ADAPTIVE_GIANT and VFS_AIO > options are also part of the kernel. > First, I get to the following: if you have GEOM_PART_BSD in the kernel alone, attaching GELI at the boot time works as expected. If you add GEOM_PART_MBR (so both GEOM_PART_BSD and GEOM_PART_MBR are in), you face the problem i've described. Second, i've tried to get kern.geom.confxml sysctl as Pawel suggested, but with no lack. First, the whole XML dump doesn't feet the console buffer, so i can't later extract it from dmesg; i've tried to dump it to some file, but due to the fact everything is mounted -ro at the point /etc/rc.d/geli is executed, i placed "mount -u -rw /" in the beginning of it. While that made a trick and i got the dump (see below), the GELI partition attached successfully (while instantly failed with "Cannot access ad0s1f (error=1)" without remounting / -rw), so I guess remounting / read-write changed something and such dump will be of no use: ACD acd0 1 r0w0e0 acd0 8796093020160 2048 MD ELI JOURNAL VOL_FFS VFS ffs.ad0s1a 4 r1w1e1 MBR msdosfs/WD Passport 4 r0w0e0 r0w0e0 msdosfs/WD Passports4 10924544 512 3 10924544 21337 714049363456 1394627663 73 da0s1 3 r0w0e0 r0w0e0 da0s1s4 10924544 512 3 10924544 21337 714049363456 1394627663 73 da0 2 r0w0e0 r0w0e0 da0s1 120031478784 512 0 120031478784 234436482 32256 63 12 ad0 2 r1w1e3 r1w1e2 ad0s1 40007729664 512 0 40007729664 78140097 32256 63 165 MBREXT BDE PART da0 2 MBR 4 63 234441647 63 255 r0w0e0 r0w0e0 da0s1 120031478784 512 1 !12 32256 120031478784 12 ad0s1 3 BSD 8 0 78140096 63 16 r1w1e2 r0w0e0 ad0s1f 5368709120 512 6 freebsd-ufs 1048576000 5368709120 7 r0w0e0 ad0s1e 734003200 512 5 freebsd-ufs 314572800 734003200 7 r0w0e0 ad0s1d 314572800 512 4 freebsd-ufs 0 314572800 7 r0w0e0 ad0s1b 402653184 512 2 freebsd-swap 6417285120 402653184 1 r1w1e1 ad0s1a 33187791360 512 1 freebsd-ufs 6819938304 33187791360 7 DISK cd0 1 r0w0e0 cd0 0 2048 0 0 da0 1 r0w0e0 da0 120034123776 512 255 63 ad0 1 r1w1e3 ad0 40007761920 512 16 63 LABEL da0s1 3 r0w0e0 r0w0e0 msdosfs/WD Passport 120031478784 512 0 120031478784 234436482 0 0 SWAP DEV msdosfs/WD Passports4 5 r0w0e0 da0s1s4 4 r0w0e0 msdosfs/WD Passport 4 r0w0e0 cd0 2 r0w0e0 da0s1 3 r0w0e0 da0s1 3 r0w0e0 da0 2 r0w0e0 ad0s1df 5 r0w0e0 ad0s1de 5 r0w0e0 ad0s1dd 5 r0w0e0 ad0s1dc 5 r0w0e0 ad0s1db 5 r0w0e0 ad0s1da 5 r0w0e0 ad0s1f 4 r0w0e0 ad0s1e 4 r0w0e0 ad0s1d 4 r0w0e0 ad0s1b 4 r0w0e0 ad0s1a 4 r0w0e0 acd0 2 r0w0e0 ad0s1 3 r0w0e0 ad0 2 r0w0e0 BSD ad0s1d 4 512 32256 32256 r0w0e0 r0w0e0 ad0s1df 5368709120 512 5 5368709120 10485760 1048576000 2048000 7 r0w0e0 ad0s1de 734003200 512 4 734003200 1433600 314572800 614400 7 r0w0e0 ad0s1dd 314572800 512 3 314572800 614400 0 0 7 r0w0e0 ad0s1dc 40007729664 512 2 40007729664 78140097 0 0 0 r0w0e0 ad0s1db 402653184 512 1 402653184 786432 6417285120 12533760 1 r0w0e0 ad0s1da 33187791360 512 0 33187791360 64819905 6819938304 13320192 7 Hope the above will help solving it, though as far as it's specific to my weird kernel configuration and the bad options combination is known it's low priority. Also, if someone can offer a simple way to get kern.geom.confxml at the time the problem occurs, i can experiment more. -- Regards, Mirya ICQ #313898202