From alex@linux.ucla.edu Thu Dec 5 12:30:54 2002 Received: with ECARTIS (v1.0.0; list devfs); Thu, 05 Dec 2002 12:30:58 -0800 (PST) Received: from linux.ucla.edu (fire2.LINUX.UCLA.EDU [131.179.104.17]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gB5KUsuR017307 for ; Thu, 5 Dec 2002 12:30:54 -0800 Received: from alex (helo=localhost) by linux.ucla.edu with local-esmtp (Exim 3.35 #1 (Debian)) id 18K2h9-0001wc-00 for ; Thu, 05 Dec 2002 12:34:03 -0800 Date: Thu, 5 Dec 2002 12:34:03 -0800 (PST) From: alex balk To: devfs@oss.sgi.com Subject: strange log messages from devfsd Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-archive-position: 34 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: alex@linux.ucla.edu Precedence: bulk X-list: devfs Content-Length: 260 Lines: 18 hi, sorry if this has been asked before. in my /var/log/syslog: `devfsd[35]: action_compat: error unlinking "vcs*" No suck file or directory' where * ranges from 1 to 6. i have a feeling minor changes to devfsd.conf would fix it. ... but what ? -alex From racsw@frontiernet.net Thu Dec 5 18:10:08 2002 Received: with ECARTIS (v1.0.0; list devfs); Thu, 05 Dec 2002 18:10:10 -0800 (PST) Received: from relay03.roc.frontiernet.net (relay03.roc.frontiernet.net [66.133.131.36]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gB62A7uR008001 for ; Thu, 5 Dec 2002 18:10:08 -0800 Received: (qmail 17883 invoked from network); 6 Dec 2002 02:13:23 -0000 Received: from unknown (HELO localhost.localdomain) ([65.37.13.251]) (envelope-sender ) by relay03.roc.frontiernet.net (qmail-ldap-1.03) with SMTP for ; 6 Dec 2002 02:13:23 -0000 Content-Type: text/plain; charset="us-ascii" From: Robert To: Devfs Subject: Mandrake 8.2 deleting symlink on boot in /dev Date: Thu, 5 Dec 2002 21:13:22 -0500 User-Agent: KMail/1.4.3 MIME-Version: 1.0 Message-Id: <200212052113.22463.racsw@frontiernet.net> Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by oss.sgi.com id gB62A7uR008001 X-archive-position: 35 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: racsw@frontiernet.net Precedence: bulk X-list: devfs Content-Length: 489 Lines: 13 Hello, I created a symlink to a valid block device ( in /dev ) to run my ATAPI 256 Mb internal IDE Zip drive, which it does with no problems. The result is "hdd4@" which is valid for my system configuration. Everything works fine at this point. However, if I reboot, or shutdown and restart, devfsd (I believe this is the culprit), deletes the symlink, and installs an empty directory "/hdd4 in /dev in it's place. Can someone help me stop this from happening? Thanks, Robert From russell@coker.com.au Fri Dec 6 00:56:44 2002 Received: with ECARTIS (v1.0.0; list devfs); Fri, 06 Dec 2002 00:56:46 -0800 (PST) Received: from tsv.sws.net.au (tsv.sws.net.au [203.36.46.2]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gB68uhuR016095 for ; Fri, 6 Dec 2002 00:56:43 -0800 Received: from lyta.coker.com.au (localhost [127.0.0.1]) by tsv.sws.net.au (Postfix) with ESMTP id D273892668; Fri, 6 Dec 2002 19:59:57 +1100 (EST) Received: from lyta (localhost [127.0.0.1]) by lyta.coker.com.au (Postfix) with ESMTP id 8797018CF; Fri, 6 Dec 2002 09:59:43 +0100 (CET) Content-Type: text/plain; charset="iso-8859-1" From: Russell Coker Reply-To: Russell Coker To: Robert , Devfs Subject: Re: Mandrake 8.2 deleting symlink on boot in /dev Date: Fri, 6 Dec 2002 09:59:42 +0100 User-Agent: KMail/1.4.3 References: <200212052113.22463.racsw@frontiernet.net> In-Reply-To: <200212052113.22463.racsw@frontiernet.net> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: <200212060959.42654.russell@coker.com.au> X-archive-position: 36 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: russell@coker.com.au Precedence: bulk X-list: devfs Content-Length: 941 Lines: 19 On Fri, 6 Dec 2002 03:13, Robert wrote: > I created a symlink to a valid block device ( in /dev ) to run my ATAPI > 256 Mb internal IDE Zip drive, which it does with no problems. > The result is "hdd4@" which is valid for my system configuration. > Everything works fine at this point. > However, if I reboot, or shutdown and restart, devfsd (I believe this is > the culprit), deletes the symlink, and installs an empty directory "/hdd4 > in /dev in it's place. On a devfs system /dev is basically a RAM disk. Every time you reboot it is wiped clean. If you want something to be there after a reboot you have to change the devfsd configuration to cause it to be regenerated. -- http://www.coker.com.au/selinux/ My NSA Security Enhanced Linux packages http://www.coker.com.au/bonnie++/ Bonnie++ hard drive benchmark http://www.coker.com.au/postal/ Postal SMTP/POP benchmark http://www.coker.com.au/~russell/ My home page From kpfleming@cox.net Fri Dec 6 10:04:41 2002 Received: with ECARTIS (v1.0.0; list devfs); Fri, 06 Dec 2002 10:04:43 -0800 (PST) Received: from fed1mtao01.cox.net (fed1mtao01.cox.net [68.6.19.244]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gB6I4fuR002493 for ; Fri, 6 Dec 2002 10:04:41 -0800 Received: from jeeves.kpf.internal ([24.56.45.215]) by fed1mtao01.cox.net (InterMail vM.5.01.04.05 201-253-122-122-105-20011231) with ESMTP id <20021206180755.UNU2256.fed1mtao01.cox.net@jeeves.kpf.internal>; Fri, 6 Dec 2002 13:07:55 -0500 Received: from [192.168.130.55] (helo=cox.net) by jeeves.kpf.internal with esmtp (Exim 4.05) id 18KMtG-0000Ic-00; Fri, 06 Dec 2002 11:07:54 -0700 Message-ID: <3DF0E77A.2080906@cox.net> Date: Fri, 06 Dec 2002 11:07:54 -0700 From: "Kevin P. Fleming" User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.2b) Gecko/20021016 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Russell Coker CC: Robert , Devfs Subject: Re: Mandrake 8.2 deleting symlink on boot in /dev References: <200212052113.22463.racsw@frontiernet.net> <200212060959.42654.russell@coker.com.au> In-Reply-To: <200212052113.22463.racsw@frontiernet.net> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-archive-position: 37 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: kpfleming@cox.net Precedence: bulk X-list: devfs Content-Length: 1037 Lines: 23 Russell Coker wrote: > On Fri, 6 Dec 2002 03:13, Robert wrote: > > > I created a symlink to a valid block device ( in /dev ) to run my ATAPI > >256 Mb internal IDE Zip drive, which it does with no problems. > > The result is "hdd4@" which is valid for my system configuration. > > Everything works fine at this point. > > However, if I reboot, or shutdown and restart, devfsd (I believe this is > >the culprit), deletes the symlink, and installs an empty directory "/hdd4 > >in /dev in it's place. > > > On a devfs system /dev is basically a RAM disk. Every time you reboot it is > wiped clean. If you want something to be there after a reboot you have to > change the devfsd configuration to cause it to be regenerated. > And the reason he's having problems with his ZIP drive in the first place is the infamous removable media problem discussed previously on the list... Robert, if you go to http://members.cox.net/kpfleming/ide-floppy and apply those patches to your kernel, your ZIP drive will behave in the way you expect. From john.pellett@adic.com Tue Dec 10 12:44:00 2002 Received: with ECARTIS (v1.0.0; list devfs); Tue, 10 Dec 2002 12:44:02 -0800 (PST) Received: from penguin.adic.com (outgoingmail.adic.com [63.81.117.28]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gBAKi0uR015173 for ; Tue, 10 Dec 2002 12:44:00 -0800 Message-ID: <995FF289C9D69747A09E429926445954B788B9@penguin.adic.com> From: john.pellett@adic.com To: devfs@oss.sgi.com Subject: devfs and redhat 8.0 Date: Tue, 10 Dec 2002 12:46:07 -0800 MIME-Version: 1.0 Content-Type: text/plain; charset="ISO-8859-1" X-archive-position: 38 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: john.pellett@adic.com Precedence: bulk X-list: devfs Content-Length: 517 Lines: 19 Has anybody been able to get devfs to work on redhat 8.0? Are there any tricks that I cannot seem to find? I got it working on redhat 7.3 and the same steps, plus many more tries, to get it working don't work on redhat 8.0. Thanks in advance, John --------------------------------------------------------------- John Pellett Advanced Digital Information Corporation PHONE: (720) 249-5740 FAX: (303) 792-2465 Email: john.pellett@adic.com visit us at www.adic.com From proski@gnu.org Tue Dec 10 13:42:36 2002 Received: with ECARTIS (v1.0.0; list devfs); Tue, 10 Dec 2002 13:42:38 -0800 (PST) Received: from fencepost.gnu.org (fencepost.gnu.org [199.232.76.164]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gBALgZuR016532 for ; Tue, 10 Dec 2002 13:42:36 -0800 Received: from proski by fencepost.gnu.org with local (Exim 4.10) id 18LsCi-0000dP-00; Tue, 10 Dec 2002 16:46:12 -0500 Date: Tue, 10 Dec 2002 16:46:11 -0500 (EST) From: Pavel Roskin X-X-Sender: proski@marabou.research.att.com To: john.pellett@adic.com cc: devfs@oss.sgi.com Subject: Re: devfs and redhat 8.0 In-Reply-To: <995FF289C9D69747A09E429926445954B788B9@penguin.adic.com> Message-ID: References: <995FF289C9D69747A09E429926445954B788B9@penguin.adic.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-archive-position: 39 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: proski@gnu.org Precedence: bulk X-list: devfs Content-Length: 772 Lines: 30 Hello! > Has anybody been able to get devfs to work on redhat 8.0? Are there any > tricks that I cannot seem to find? I got it working on redhat 7.3 and > the same steps, plus many more tries, to get it working don't work on > redhat 8.0. It would be better if you explained what the problem is. I'm typing these words on a Red Hat 8.0 system with devfs. The steps are: 1) Install devfsd 2) Edit /etc/devfsd.conf to create the symlinks you need and set permissions, e.g. REGISTER ^input/mice$ CFUNCTION GLOBAL mksymlink $devname mouse REGISTER ^cdroms/cdrom.$ CFUNCTION GLOBAL mksymlink $devname cdrom REGISTER ^sound/.*$ PERMISSIONS root.root 0666 3) Comment out /dev/pts in /etc/fstab 4) Install kernel with devfs support 5) Reboot -- Regards, Pavel Roskin From jabiru_croc@yahoo.com Tue Dec 10 14:09:52 2002 Received: with ECARTIS (v1.0.0; list devfs); Tue, 10 Dec 2002 14:09:54 -0800 (PST) Received: from web40002.mail.yahoo.com (web40002.mail.yahoo.com [66.218.78.20]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gBAM9quR016729 for ; Tue, 10 Dec 2002 14:09:52 -0800 Message-ID: <20021210221324.51074.qmail@web40002.mail.yahoo.com> Received: from [24.50.169.109] by web40002.mail.yahoo.com via HTTP; Tue, 10 Dec 2002 14:13:24 PST Date: Tue, 10 Dec 2002 14:13:24 -0800 (PST) From: Brad Chapman Subject: Re: devfs and redhat 8.0 To: john.pellett@adic.com Cc: devfs@oss.sgi.com In-Reply-To: <995FF289C9D69747A09E429926445954B788B9@penguin.adic.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-archive-position: 40 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: jabiru_croc@yahoo.com Precedence: bulk X-list: devfs Content-Length: 721 Lines: 23 Mr. Pellett, --- john.pellett@adic.com wrote: > Has anybody been able to get devfs to work on redhat 8.0? > Are there any tricks that I cannot seem to find? I got > it working on redhat 7.3 and the same steps, plus many more > tries, to get it working don't work on redhat 8.0. It works OK on Red Hat 8.0. You just need to edit /etc/inittab, /etc/fstab, /etc/updfstab.conf, /etc/devfsd.conf (to add compat entries), and maybe a few other files as well. It's not difficult, but it will take a few reboots to iron out all the errors. > > Thanks in advance, > John Brad __________________________________________________ Do you Yahoo!? Yahoo! Mail Plus - Powerful. Affordable. Sign up now. http://mailplus.yahoo.com From racsw@frontiernet.net Tue Dec 10 17:29:42 2002 Received: with ECARTIS (v1.0.0; list devfs); Tue, 10 Dec 2002 17:29:46 -0800 (PST) Received: from relay01.roc.frontiernet.net (relay01.roc.frontiernet.net [66.133.131.34]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gBB1TfuR022439 for ; Tue, 10 Dec 2002 17:29:41 -0800 Received: (qmail 26103 invoked from network); 11 Dec 2002 01:33:18 -0000 Received: from unknown (HELO localhost.localdomain) ([65.37.13.251]) (envelope-sender ) by relay01.roc.frontiernet.net (qmail-ldap-1.03) with SMTP for ; 11 Dec 2002 01:33:18 -0000 Content-Type: text/plain; charset="us-ascii" From: Robert To: Devfs Subject: Host0, Host2 for IDE devices Date: Tue, 10 Dec 2002 20:33:17 -0500 User-Agent: KMail/1.4.3 MIME-Version: 1.0 Message-Id: <200212102033.17836.racsw@frontiernet.net> Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by oss.sgi.com id gBB1TfuR022439 X-archive-position: 41 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: racsw@frontiernet.net Precedence: bulk X-list: devfs Content-Length: 958 Lines: 23 Hello, I'm trying to understand the structure of host0, host1, host2, etc in regards to IDE devices. I did read Richards FAQ, and he did discuss this for scsi devices, but I didn't see the logic explained for the IDE devices or whatever determines the host# scheme. ( ...or it was in there, but I didn't understand it ) If someone could explain this, I sure would appreciate it. In regards to the /bus#/target# logic, maybe someone could confirm my understanding of this as well as I have listed below. /bus0/target0 = primary master /bus0/target1 = primary slave /bus1/target0 = secondary master /bus1/target1 = secondary slave. I'm trying to get this fellows internal zip drive working on his system, but because he has a different HD and Zip configuration off his IDE slots than I do, it's necessary I understand how to figure this out so I can get his block device (part4) in the right place for his system. Thanks, Robert From Karl.Merritts@adic.com Wed Dec 11 12:00:23 2002 Received: with ECARTIS (v1.0.0; list devfs); Wed, 11 Dec 2002 12:00:25 -0800 (PST) Received: from penguin.adic.com (outgoingmail.adic.com [63.81.117.28]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gBBK0MuR002460 for ; Wed, 11 Dec 2002 12:00:23 -0800 Message-ID: <995FF289C9D69747A09E4299264459540526DAC9@penguin.adic.com> From: Karl.Merritts@adic.com To: devfs@oss.sgi.com Subject: DevFS and Red Hat 8.0 Date: Wed, 11 Dec 2002 12:02:28 -0800 MIME-Version: 1.0 Content-Type: text/plain; charset="ISO-8859-1" X-archive-position: 42 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: Karl.Merritts@adic.com Precedence: bulk X-list: devfs Content-Length: 2538 Lines: 49 John P. and I work together and I am acutally the one trying to get DevFS to work on Red Hat 8.0. As he said we had no problem with getting it to run in Red Hat 7.3. Mainly I am just wondering what you guys do with your init scripts. On 7.3 we used the wrapper.sh from Dan Robbins' DevFS tutorial http://www-106.ibm.com/developerworks/library/l-fs6/wrapper.sh. Trying it using the exact same technique now yields the below error messages. ...... mkdir: cannot create directory /dev/ide/host0/bus0/target0/lun0/: 2 failed to create /dev/ide/host0/bus0/target0/lun0/part1 mkdir: cannot create directory /dev/ide/host0/bus0/target0/lun0/: 2 failed to create /dev/ide/host0/bus0/target0/lun0/part2 mkdir: cannot create directory /dev/ide/host0/bus0/target0/lun0/: 2 failed to create /dev/ide/host0/bus0/target0/lun0/part3 mkdir: cannot create directory /dev/ide/host0/bus0/target0/lun0/: 2 failed to create /dev/ide/host0/bus0/target0/lun0/part4 mkdir: cannot create directory /dev/ide/host0/bus0/target0/lun0/: 2 failed to create /dev/ide/host0/bus0/target0/lun0/part5 mkdir: cannot create directory /dev/ide/host0/bus0/target1/lun0/: 2 failed to create /dev/ide/host0/bus0/target0/lun0/disc mkdir: cannot create directory /dev/ide/host0/bus0/target1/lun0/: 2 failed to create /dev/ide/host0/bus0/target0/lun0/part1 Creating root device mkdir: cannot create directory /dev/ide/host0/bus0/target0/lun0/: 2 mkdir: cannot create directory /dev/ide/host0/bus0/target0/lun0/: 2 mkdir: cannot create directory /dev/ide/host0/bus0/target0/lun0/: 2 mkdir: cannot create directory /dev/ide/host0/bus0/target0/lun0/: 2 mkdir: cannot create directory /dev/ide/host0/bus0/target1/lun0/: 2 mkrootdev: label / not found Mounting root filesystem mount: error 2 mounting ext3 pivotroot: pivot_root(/sysroot,/sysroot/initrd) failed: 2 umount /initrd/proc failed: 2 Freeing unused kernel memory: 244k freed Kernel panic: No init found. Try passing init= option to kernel The Kernel I am using is a rebuild of the 2.4.18-18.8.0(I have also tried on the previous two 2.4.18-14 and 2.4.18-17.8.0) with no luck and I do edit the Makefile in /usr/src/linux(both linux and linux-2.4 point to the right kernel). We were also able to successfully get the 2.4.18-17.7 to work in 7.3 which leads me to believe it is more of a OS than kernel problem. All of the kernel options were also set correctly since the .config used is the same one as the one used in the 2.4.18-17.7 on 7.3 also on the same computer. Thank very much for any help you can offer. Karl From proski@gnu.org Wed Dec 11 12:24:20 2002 Received: with ECARTIS (v1.0.0; list devfs); Wed, 11 Dec 2002 12:24:22 -0800 (PST) Received: from fencepost.gnu.org (fencepost.gnu.org [199.232.76.164]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gBBKOJuR002624 for ; Wed, 11 Dec 2002 12:24:20 -0800 Received: from proski by fencepost.gnu.org with local (Exim 4.10) id 18MDSZ-00057J-00; Wed, 11 Dec 2002 15:27:59 -0500 Date: Wed, 11 Dec 2002 15:27:58 -0500 (EST) From: Pavel Roskin X-X-Sender: proski@marabou.research.att.com To: Karl.Merritts@adic.com cc: devfs@oss.sgi.com Subject: Re: DevFS and Red Hat 8.0 In-Reply-To: <995FF289C9D69747A09E4299264459540526DAC9@penguin.adic.com> Message-ID: References: <995FF289C9D69747A09E4299264459540526DAC9@penguin.adic.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-archive-position: 43 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: proski@gnu.org Precedence: bulk X-list: devfs Content-Length: 418 Lines: 13 Hello! Check my previous message in the list. I don't change any init scripts, neither do I change /etc/fstab. I don't use any scripts to save and restore entries in /dev - I fully rely on /etc/devfsd.conf to create symlinks and set permissions. On the other hand, I don't use initrd. I don't use kudzu either. But I don't remember anything similar to the problems you are describing. -- Regards, Pavel Roskin From chutz@gg3.net Fri Dec 13 10:59:14 2002 Received: with ECARTIS (v1.0.0; list devfs); Fri, 13 Dec 2002 10:59:15 -0800 (PST) Received: from mail.my.home (atfwmeoj@142.13.111.219.st.excite.co.jp [219.111.13.142]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gBDIxDuR006024 for ; Fri, 13 Dec 2002 10:59:13 -0800 Received: (qmail 11322 invoked from network); 13 Dec 2002 19:02:56 -0000 Received: from tank.my.home (HELO gg3.net) (10.0.0.2) by 0 with SMTP; 13 Dec 2002 19:02:56 -0000 Message-ID: <3DFA2EE0.3050901@gg3.net> Date: Sat, 14 Dec 2002 04:02:56 +0900 From: Georgi Georgiev User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.1) Gecko/20021018 Netscape6/6.1 X-Accept-Language: bg, en, ja MIME-Version: 1.0 To: devfs@oss.sgi.com Subject: Re: Host0, Host2 for IDE devices References: <200212102033.17836.racsw@frontiernet.net> X-Enigmail-Version: 0.65.2.0 X-Enigmail-Supports: pgp-inline, pgp-mime Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-archive-position: 44 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: chutz@gg3.net Precedence: bulk X-list: devfs Content-Length: 1653 Lines: 40 I read somewhere that host0 and host1 are assigned for on-board IDE interfaces, and IDE controllers on PCI slots receive host2 and up. I was a little surprised myself, when the IDE controller I bought recently became host2... Robert wrote: > Hello, > I'm trying to understand the structure of host0, host1, host2, etc in > regards to IDE devices. I did read Richards FAQ, and he did discuss this > for scsi devices, but I didn't see the logic explained for the IDE devices or > whatever determines the host# scheme. ( ...or it was in there, but I didn't > understand it ) If someone could explain this, I sure would appreciate it. > > In regards to the /bus#/target# logic, maybe someone could confirm my > understanding of this as well as I have listed below. > > /bus0/target0 = primary master > /bus0/target1 = primary slave > /bus1/target0 = secondary master > /bus1/target1 = secondary slave. > > I'm trying to get this fellows internal zip drive working on his system, but > because he has a different HD and Zip configuration off his IDE slots than I > do, it's necessary I understand how to figure this out so I can get his block > device (part4) in the right place for his system. > > Thanks, > Robert > > -- /^^^^^^^^^^^^^^^^^^^^^^^^^^^\/^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\ / Georgi Georgiev (-< / In 1914, the first crossword \ \ chutz@chubaka.net /\ .o)\ puzzle was printed in a / / +81(90)6266-1163 V_/_ |(/)/ newspaper. The creator received \ \ ^^^^^^^^^\ $4000 down ... and $3000 across. / \___________________________/\________________________________/ From james_mcmechan@hotmail.com Fri Dec 13 18:30:52 2002 Received: with ECARTIS (v1.0.0; list devfs); Fri, 13 Dec 2002 18:30:54 -0800 (PST) Received: from hotmail.com (oe15.law7.hotmail.com [216.33.236.119]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gBE2Up3v014007 for ; Fri, 13 Dec 2002 18:30:52 -0800 Received: from mail pickup service by hotmail.com with Microsoft SMTPSVC; Fri, 13 Dec 2002 18:34:37 -0800 X-Originating-IP: [67.28.104.242] From: "James McMechan" To: References: <200212102033.17836.racsw@frontiernet.net> <3DFA2EE0.3050901@gg3.net> Subject: Perhaps a silly question about late 2.5 kernels Date: Fri, 13 Dec 2002 18:29:49 -0800 MIME-Version: 1.0 Content-Type: multipart/signed; protocol="application/x-pkcs7-signature"; micalg=SHA1; boundary="----=_NextPart_000_002A_01C2A2D5.9F8DCF80" X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.50.4522.1200 X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4522.1200 Message-ID: X-OriginalArrivalTime: 14 Dec 2002 02:34:37.0606 (UTC) FILETIME=[589E1460:01C2A319] X-archive-position: 45 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: James_McMechan@hotmail.com Precedence: bulk X-list: devfs Content-Length: 6296 Lines: 111 This is a multi-part message in MIME format. ------=_NextPart_000_002A_01C2A2D5.9F8DCF80 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit I have been working on the user-mode-linux (ARCH=um) block driver ubd_kern.c & ubd_user.c Previously I was putting "ubd" into the registration so we had /dev/ubd/discX/... devices nicely created by devfs. Recently in 2.5 the add_disk() helper has appeared to be the correct way to register a block device and it also worked with a disk->disk_name of "ubd" still creating /dev/ubd/disc0/disc & co. But in 2.5.50/51 sysfs reportedly is also using the name and kills the kernel by trying to allocate two objects with the same name failing and then deleting both objects without removing the references to the first object. If add_disk is given a unique disk_name for each device I tend to end up with /dev/ubd0/disc0/disc /dev/ubd1/disc1/disc which is slightly awkward. So what I was trying was to use ubd/compat/ubdX and then devfs_mk_symlink to recreate the /dev/ubd/discX entries is there some better way to do this I feel I am approaching the problem from the wrong direction. Any suggestions on a better approach to the problem, is it a purely sysfs problem and add_disk() should take "ubd" for all the devices in one major number? I have been using devfs much longer than sysfs has existed and so I am asking here first since this is the part I understand the most of. Also what is the proper method of disposing of the symlink when the device goes away? I can save the handle and devfs_unregister but I am not sure that is the proper method. - err = ubd_new_disk(MAJOR_NR, dev->size, "ubd", n, &ubd_gendisk[n], ubd_dir_handle, &dev->real); + /* glack sysfs now insists on seperate names for each device not just each major number */ + /* which causes devfs to create /dev/ubd/Y/discN */ + sprintf(name1,"ubd/compat/ubd%d",n); + sprintf(name2,"compat/ubd%d/disc%d", n, n); + sprintf(name3,"disc%d", n); + err = ubd_new_disk(MAJOR_NR, dev->size, name1, n, &ubd_gendisk[n], ubd_dir_handle, &dev->real); + devfs_mk_symlink(ubd_dir_handle, name3, DEVFS_FL_DEFAULT, name2, NULL, NULL); ------=_NextPart_000_002A_01C2A2D5.9F8DCF80 Content-Type: application/x-pkcs7-signature; name="smime.p7s" Content-Transfer-Encoding: base64 Content-Disposition: attachment; filename="smime.p7s" MIAGCSqGSIb3DQEHAqCAMIACAQExCzAJBgUrDgMCGgUAMIAGCSqGSIb3DQEHAQAAoIII/zCCAo4w ggH3oAMCAQICAwjK5DANBgkqhkiG9w0BAQQFADCBkjELMAkGA1UEBhMCWkExFTATBgNVBAgTDFdl c3Rlcm4gQ2FwZTESMBAGA1UEBxMJQ2FwZSBUb3duMQ8wDQYDVQQKEwZUaGF3dGUxHTAbBgNVBAsT FENlcnRpZmljYXRlIFNlcnZpY2VzMSgwJgYDVQQDEx9QZXJzb25hbCBGcmVlbWFpbCBSU0EgMjAw MC44LjMwMB4XDTAyMTEyNzAwMzExOVoXDTAzMTEyNzAwMzExOVowTDEfMB0GA1UEAxMWVGhhd3Rl IEZyZWVtYWlsIE1lbWJlcjEpMCcGCSqGSIb3DQEJARYaSmFtZXNfTWNNZWNoYW5AaG90bWFpbC5j b20wgZ8wDQYJKoZIhvcNAQEBBQADgY0AMIGJAoGBAPXSzIHYbVKpiTH4jq0HHE8SH7ztIDMG/t8o WrnOFcEfYHroWd/MCO4GXAaS6gG1xJpyii1GBdmcvhRqGo2XbaMHfzgdVqD42I7CQjJGcE3je/Oo ur7/JR+ZhB6f9pcbI0P7DuLkjz+WSCCT9at4cIujIXLaudUXRZtDq3+Lvgm5AgMBAAGjNzA1MCUG A1UdEQQeMByBGkphbWVzX01jTWVjaGFuQGhvdG1haWwuY29tMAwGA1UdEwEB/wQCMAAwDQYJKoZI hvcNAQEEBQADgYEASGz4bx4IWEV5dU6fWrNvlNpmnlvhbNICyMCct4JzDu6rxSAHs2oq7NNI8wAo dK1tUUAdRt9F2J/mFw9tgzGyKMquyUBJp0nj3k1pIE/9+AHj7YOyCHqf5NVuEkWPwsysVRYWRe/l glykR0llVWXLYvry4VjzCmuWy2QTBrE6sDYwggMtMIIClqADAgECAgEAMA0GCSqGSIb3DQEBBAUA MIHRMQswCQYDVQQGEwJaQTEVMBMGA1UECBMMV2VzdGVybiBDYXBlMRIwEAYDVQQHEwlDYXBlIFRv d24xGjAYBgNVBAoTEVRoYXd0ZSBDb25zdWx0aW5nMSgwJgYDVQQLEx9DZXJ0aWZpY2F0aW9uIFNl cnZpY2VzIERpdmlzaW9uMSQwIgYDVQQDExtUaGF3dGUgUGVyc29uYWwgRnJlZW1haWwgQ0ExKzAp BgkqhkiG9w0BCQEWHHBlcnNvbmFsLWZyZWVtYWlsQHRoYXd0ZS5jb20wHhcNOTYwMTAxMDAwMDAw WhcNMjAxMjMxMjM1OTU5WjCB0TELMAkGA1UEBhMCWkExFTATBgNVBAgTDFdlc3Rlcm4gQ2FwZTES MBAGA1UEBxMJQ2FwZSBUb3duMRowGAYDVQQKExFUaGF3dGUgQ29uc3VsdGluZzEoMCYGA1UECxMf Q2VydGlmaWNhdGlvbiBTZXJ2aWNlcyBEaXZpc2lvbjEkMCIGA1UEAxMbVGhhd3RlIFBlcnNvbmFs IEZyZWVtYWlsIENBMSswKQYJKoZIhvcNAQkBFhxwZXJzb25hbC1mcmVlbWFpbEB0aGF3dGUuY29t MIGfMA0GCSqGSIb3DQEBAQUAA4GNADCBiQKBgQDUadfUsJRkW3HpR9gMUbbqcpGwhF59LQ2PexLf hSV1KHQ6QixjJ5+Ve0vvfhmHHYbqo925zpZkGsIUbkSsfOaP6E0PcR9AOKYAo4d49vmUhl6t6sBe duvZFKNdbnp8DKVLVX8GGSl/npom1Wq7OCQIapjHsdqjmJH9edvlWsQcuQIDAQABoxMwETAPBgNV HRMBAf8EBTADAQH/MA0GCSqGSIb3DQEBBAUAA4GBAMfskn5O+PWWpWdiKqTwTRFg0G+NYFhhrCa7 UjVcCM8w+6hKloofYkIjjBcP9LpknBesRynfnZhe0mxgcVyirNx54+duAEcftQ0o6AKd5Jr9E/Sm 2Xyx+NxfIyYJkYBz0BQb3kOpgyXy5pwvFcr+pquKB3WLDN1RhGvk+NHOd6KBMIIDODCCAqGgAwIB AgIQZkVyt8x09c9jdkWE0C6RATANBgkqhkiG9w0BAQQFADCB0TELMAkGA1UEBhMCWkExFTATBgNV BAgTDFdlc3Rlcm4gQ2FwZTESMBAGA1UEBxMJQ2FwZSBUb3duMRowGAYDVQQKExFUaGF3dGUgQ29u c3VsdGluZzEoMCYGA1UECxMfQ2VydGlmaWNhdGlvbiBTZXJ2aWNlcyBEaXZpc2lvbjEkMCIGA1UE AxMbVGhhd3RlIFBlcnNvbmFsIEZyZWVtYWlsIENBMSswKQYJKoZIhvcNAQkBFhxwZXJzb25hbC1m cmVlbWFpbEB0aGF3dGUuY29tMB4XDTAwMDgzMDAwMDAwMFoXDTA0MDgyNzIzNTk1OVowgZIxCzAJ BgNVBAYTAlpBMRUwEwYDVQQIEwxXZXN0ZXJuIENhcGUxEjAQBgNVBAcTCUNhcGUgVG93bjEPMA0G A1UEChMGVGhhd3RlMR0wGwYDVQQLExRDZXJ0aWZpY2F0ZSBTZXJ2aWNlczEoMCYGA1UEAxMfUGVy c29uYWwgRnJlZW1haWwgUlNBIDIwMDAuOC4zMDCBnzANBgkqhkiG9w0BAQEFAAOBjQAwgYkCgYEA 3jMypmPHCSVFPtJueCdngcXaiBmClw7jRCmKYzUqbXA8+tyu9+50bzC8M5B/+TRxoKNtmPHDT6Jl 2w36S/HW3WGl+YXNVZo1Gp2Sdagnrthy+boC9tewkd4c6avgGAOofENCUFGHgzzwObSbVIoTh/+z m51JZgAtCYnslGvpoWkCAwEAAaNOMEwwKQYDVR0RBCIwIKQeMBwxGjAYBgNVBAMTEVByaXZhdGVM YWJlbDEtMjk3MBIGA1UdEwEB/wQIMAYBAf8CAQAwCwYDVR0PBAQDAgEGMA0GCSqGSIb3DQEBBAUA A4GBADGxS0dd+QFx5fVTbF151j2YwCYTYoEipxL4IpXoG0m3J3sEObr85vIk65H6vewNKjj3UFWo bPcNrUwbvAP0teuiR59sogxYjTFCCRFssBpp0SsSskBdavl50OouJd2K5PzbDR+dAvNa28o89kTq JmmHf0iezqWf54TYyWJirQXGMYIB/jCCAfoCAQEwgZowgZIxCzAJBgNVBAYTAlpBMRUwEwYDVQQI EwxXZXN0ZXJuIENhcGUxEjAQBgNVBAcTCUNhcGUgVG93bjEPMA0GA1UEChMGVGhhd3RlMR0wGwYD VQQLExRDZXJ0aWZpY2F0ZSBTZXJ2aWNlczEoMCYGA1UEAxMfUGVyc29uYWwgRnJlZW1haWwgUlNB IDIwMDAuOC4zMAIDCMrkMAkGBSsOAwIaBQCggbowGAYJKoZIhvcNAQkDMQsGCSqGSIb3DQEHATAc BgkqhkiG9w0BCQUxDxcNMDIxMjE0MDIyOTQ5WjAjBgkqhkiG9w0BCQQxFgQUpi7TzRMzJy2DklxO NtPPdSzVhYwwWwYJKoZIhvcNAQkPMU4wTDAKBggqhkiG9w0DBzAOBggqhkiG9w0DAgICAIAwDQYI KoZIhvcNAwICAUAwBwYFKw4DAgcwDQYIKoZIhvcNAwICASgwBwYFKw4DAh0wDQYJKoZIhvcNAQEB BQAEgYAE/HHn7nNzVoZfR+x67I9Hn/THVhhyPq9HlnUUO2RhvsDJm5SKP6VI9FYHVSkhayPfdEvO gZhthBBZWJaj0o9Y0tTuiJDsaA0uWDqPepkuRhDZi2wZ/ZwSl77ozo7c6na6JorjcuVycJoeEmg7 XvGQXabGpqBq5l51lYmQM25CeAAAAAAAAA== ------=_NextPart_000_002A_01C2A2D5.9F8DCF80-- From chutz@chubaka.homeip.net Fri Dec 13 20:48:56 2002 Received: with ECARTIS (v1.0.0; list devfs); Fri, 13 Dec 2002 20:48:59 -0800 (PST) Received: from mail.my.home (142.13.111.219.st.excite.co.jp [219.111.13.142]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gBE4mt3v021616 for ; Fri, 13 Dec 2002 20:48:56 -0800 Received: (qmail 12214 invoked from network); 14 Dec 2002 04:52:41 -0000 Received: from tank.my.home (10.0.0.2) by 0 with QMQP; 14 Dec 2002 04:52:41 -0000 Date: Sat, 14 Dec 2002 13:52:41 +0900 From: Georgi Georgiev To: James McMechan Cc: devfs@oss.sgi.com Subject: Re: Perhaps a silly question about late 2.5 kernels Message-ID: <20021214045241.GA19717@tank.my.home> References: <200212102033.17836.racsw@frontiernet.net> <3DFA2EE0.3050901@gg3.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.4i X-archive-position: 46 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: chutz@gg3.net Precedence: bulk X-list: devfs Content-Length: 429 Lines: 9 Please, do not reply to a message in a mailing list if what you want is to start a new thread. -- /^^^^^^^^^^^^^^^^^^^^^^^^^^^\/^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\ / Georgi Georgiev (-< / Debug is human, de-fix divine. \ \ chutz@chubaka.net /\ .o)\ / / +81(90)6266-1163 V_/_ |(/)/ \ \___________________________/\__________________________________/ From fokkema@nat.vu.nl Tue Dec 17 04:02:28 2002 Received: with ECARTIS (v1.0.0; list devfs); Tue, 17 Dec 2002 04:02:34 -0800 (PST) Received: from nat.vu.nl (hardy.nat.vu.nl [130.37.36.3]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gBHC2Q3v018649 for ; Tue, 17 Dec 2002 04:02:27 -0800 Received: from selene.nat.vu.nl (selene [130.37.32.13]) by nat.vu.nl (8.10.2+Sun/8.10.2) with ESMTP id gBHC6SN03842; Tue, 17 Dec 2002 13:06:28 +0100 (MET) Received: from localhost (fokkema@localhost) by selene.nat.vu.nl (8.11.6+Sun/8.9.3) with ESMTP id gBHC6RW04241; Tue, 17 Dec 2002 13:06:27 +0100 (MET) X-Authentication-Warning: selene.nat.vu.nl: fokkema owned process doing -bs Date: Tue, 17 Dec 2002 13:06:26 +0100 (MET) From: David Fokkema To: devfs@oss.sgi.com cc: David Fokkema Subject: Boot problems Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-archive-position: 47 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: fokkema@nat.vu.nl Precedence: bulk X-list: devfs Content-Length: 1369 Lines: 35 Hi there! It's awfully quiet in here, I hope this mailing list is still alive... Anyway, I'm having some problems with devfs during boot. My system is working fine, that's not the problem. What is, however, is that I want to make a bootable CD and I was very glad with the existence of /dev/cdroms/cdrom0 and alike. However, my kernel (on a test bootfloppy) complains that it can't access that device. It does work if I use /dev/ide/host0/bus0/target1/lun0/cd and alike. But then, I need to know where my CD is located. Also, on my system, when I want LILO to append a root statement to the kernel command line, I can't use /dev/discs/disc0/part1 either. Again, I have to use /dev/ide/host0/bus0/target0/lun0/part1. In the latter case, I'll survive (it's still annoying, though) but in the former, I do want it to work. I guess I could use an initrd image which will use /dev/cdroms/cdrom0, but that is just extra trouble. I would say that it should work, or else should be added to the kernel. My setup: kernel 2.4.19 compiled with devfs set to mount at boot, cdrom driver and isofs included. LILO append line: append = "devfs=only root=/dev/cdroms/cdrom0" (which doesn't work) or, (doesn't work either) append = "devfs=mount root=/dev/cdroms/cdrom0" If someone knows how to help me, please let me know! Thanks in advance, David Fokkema (fokkema@nat.vu.nl) From Andrey.Borzenkov@siemens.com Tue Dec 17 04:43:10 2002 Received: with ECARTIS (v1.0.0; list devfs); Tue, 17 Dec 2002 04:43:11 -0800 (PST) Received: from thoth.sbs.de (thoth.sbs.de [192.35.17.2]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gBHCh83v020288 for ; Tue, 17 Dec 2002 04:43:09 -0800 Received: from mail1.siemens.de (mail1.siemens.de [139.23.33.14]) by thoth.sbs.de (8.11.6/8.11.6) with ESMTP id gBHCl5i19551; Tue, 17 Dec 2002 13:47:05 +0100 (MET) Received: from MOWD019A.mow.siemens.ru ([139.24.18.3]) by mail1.siemens.de (8.11.6/8.11.6) with ESMTP id gBHCl4c03926; Tue, 17 Dec 2002 13:47:04 +0100 (MET) Received: by MOWD019A with Internet Mail Service (5.5.2653.19) id ; Tue, 17 Dec 2002 16:00:07 +0300 Received: from mw2b210c (163.242.193.12 [163.242.193.12]) by MOWD019A.mow.siemens.ru with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2653.13) id ZCG3Z708; Tue, 17 Dec 2002 16:00:05 +0300 From: Borzenkov Andrey To: "'David Fokkema'" , devfs@oss.sgi.com Subject: RE: Boot problems Date: Tue, 17 Dec 2002 15:46:57 +0300 Message-ID: <6134254DE87BD411908B00A0C99B044F03A0B59E@MOWD019A> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook, Build 10.0.4510 Importance: Normal In-Reply-To: X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1106 X-archive-position: 48 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: Andrey.Borzenkov@siemens.com Precedence: bulk X-list: devfs Content-Length: 1958 Lines: 56 > > It's awfully quiet in here, I hope this mailing list is still alive... > it probably means that devfs does not have any problems :) > Anyway, I'm having some problems with devfs during boot. My system is > working fine, that's not the problem. What is, however, is that I want to > make a bootable CD and I was very glad with the existence of > /dev/cdroms/cdrom0 and alike. However, my kernel (on a test bootfloppy) > complains that it can't access that device. It does work if I use > /dev/ide/host0/bus0/target1/lun0/cd and alike. But then, I need to know > where my CD is located. Also, on my system, when I want LILO to append a > root statement to the kernel command line, I can't use > /dev/discs/disc0/part1 either. Again, I have to use > /dev/ide/host0/bus0/target0/lun0/part1. I can. Probably lilo problem: image=/boot/vmlinuz-2.4.20-2mdk label=test1 root=/dev/discs/disc0/part5 read-only optional vga=788 append=" devfs=mount apm=power_off,realmode_power_off" initrd=/boot/initrd-2.4.20-2mdk.img bor@cooker% rpm -qi lilo Name : lilo Relocations: (not relocateable) Version : 22.3.2 Vendor: MandrakeSoft Release : 5mdk ... In the latter case, I'll survive > (it's still annoying, though) but in the former, I do want it to work. I > guess I could use an initrd image which will use /dev/cdroms/cdrom0, but > that is just extra trouble. I would say that it should work, or else > should be added to the kernel. > > My setup: kernel 2.4.19 compiled with devfs set to mount at boot, cdrom > driver and isofs included. LILO append line: > > append = "devfs=only root=/dev/cdroms/cdrom0" (which doesn't work) > > or, (doesn't work either) > > append = "devfs=mount root=/dev/cdroms/cdrom0" > > If someone knows how to help me, please let me know! > Is cdrom module or statically compiled in kernel? -andrey From marton.kadar@mail.com Tue Dec 17 04:57:51 2002 Received: with ECARTIS (v1.0.0; list devfs); Tue, 17 Dec 2002 04:57:55 -0800 (PST) Received: from ws1-4.us4.outblaze.com (205-158-62-50.outblaze.com [205.158.62.50]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gBHCvo3v020947 for ; Tue, 17 Dec 2002 04:57:51 -0800 Received: (qmail 19202 invoked by uid 1001); 17 Dec 2002 13:01:51 -0000 Message-ID: <20021217130151.19201.qmail@mail.com> Content-Type: text/plain; charset="iso-8859-1" Content-Disposition: inline Content-Transfer-Encoding: 7bit MIME-Version: 1.0 X-Mailer: MIME-tools 5.41 (Entity 5.404) Received: from [145.236.107.5] by ws1-4.us4.outblaze.com with http for marton.kadar@mail.com; Tue, 17 Dec 2002 21:01:51 +0800 From: "Marton Kadar" To: fokkema@nat.vu.nl Cc: devfs@oss.sgi.com Date: Tue, 17 Dec 2002 21:01:51 +0800 Subject: Re: Boot problems X-Originating-Ip: 145.236.107.5 X-Originating-Server: ws1-4.us4.outblaze.com X-archive-position: 49 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: marton.kadar@mail.com Precedence: bulk X-list: devfs Content-Length: 3088 Lines: 84 Hi, To me the problem seems to be that you aren't distinguishing properly between the basic devfs namespace, and the filesystem through which it is exposed to the world. When specifying the root device while booting, you refer to the namespace, as you have no filesystems alive yet. (root=ide/host0/... works without /dev/ prepended!) Once you have a root device, you can begin thinking about mounting filesystems (beginning with the root fs!). Once you also mount the devfs fs among others, then you can start referring to the devices through it. And the case seems to be that the comfortable symlinks you intend to use are symlinks only, and as such they can be used only as file references. They in themselves are not part of the bare underlying namespace that you must use to specify your root device. Question: Are you trying to make a floppy-CD combination that boot only together??? Your message hints that the kernel is on the floppy and once it is loaded, it will mount the CD as root filesystem... Sounds like a weird setup to me. In the extreme case you can't directly boot from the CD drive, couldn't you rather put e.g. GRUB on the floppy to help you in booting and prepare the CD itself so that with the proper hardware it boots alone, kernel, etc all included? Marton ----- Original Message ----- From: David Fokkema Date: Tue, 17 Dec 2002 13:06:26 +0100 (MET) To: devfs@oss.sgi.com Subject: Boot problems > Hi there! > > It's awfully quiet in here, I hope this mailing list is still alive... > > Anyway, I'm having some problems with devfs during boot. My system is > working fine, that's not the problem. What is, however, is that I want to > make a bootable CD and I was very glad with the existence of > /dev/cdroms/cdrom0 and alike. However, my kernel (on a test bootfloppy) > complains that it can't access that device. It does work if I use > /dev/ide/host0/bus0/target1/lun0/cd and alike. But then, I need to know > where my CD is located. Also, on my system, when I want LILO to append a > root statement to the kernel command line, I can't use > /dev/discs/disc0/part1 either. Again, I have to use > /dev/ide/host0/bus0/target0/lun0/part1. In the latter case, I'll survive > (it's still annoying, though) but in the former, I do want it to work. I > guess I could use an initrd image which will use /dev/cdroms/cdrom0, but > that is just extra trouble. I would say that it should work, or else > should be added to the kernel. > > My setup: kernel 2.4.19 compiled with devfs set to mount at boot, cdrom > driver and isofs included. LILO append line: > > append = "devfs=only root=/dev/cdroms/cdrom0" (which doesn't work) > > or, (doesn't work either) > > append = "devfs=mount root=/dev/cdroms/cdrom0" > > If someone knows how to help me, please let me know! > > Thanks in advance, > > David Fokkema > (fokkema@nat.vu.nl) > > -- __________________________________________________________ Sign-up for your own FREE Personalized E-mail at Mail.com http://www.mail.com/?sr=signup Meet Singles http://corp.mail.com/lavalife From fokkema@nat.vu.nl Tue Dec 17 05:03:52 2002 Received: with ECARTIS (v1.0.0; list devfs); Tue, 17 Dec 2002 05:03:54 -0800 (PST) Received: from nat.vu.nl (hardy.nat.vu.nl [130.37.36.3]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gBHD3p3v021023 for ; Tue, 17 Dec 2002 05:03:52 -0800 Received: from selene.nat.vu.nl (selene [130.37.32.13]) by nat.vu.nl (8.10.2+Sun/8.10.2) with ESMTP id gBHD7rN09903; Tue, 17 Dec 2002 14:07:53 +0100 (MET) Received: from localhost (fokkema@localhost) by selene.nat.vu.nl (8.11.6+Sun/8.9.3) with ESMTP id gBHD7pA04314; Tue, 17 Dec 2002 14:07:52 +0100 (MET) X-Authentication-Warning: selene.nat.vu.nl: fokkema owned process doing -bs Date: Tue, 17 Dec 2002 14:07:51 +0100 (MET) From: David Fokkema To: Marton Kadar cc: devfs@oss.sgi.com Subject: Re: Boot problems In-Reply-To: <20021217130151.19201.qmail@mail.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-archive-position: 50 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: fokkema@nat.vu.nl Precedence: bulk X-list: devfs Content-Length: 1056 Lines: 27 Hi Marton, > And the case seems to be that the comfortable symlinks > you intend to use are symlinks only, and as such they > can be used only as file references. They in themselves > are not part of the bare underlying namespace that you > must use to specify your root device. That would be awful. Even without devfsd, the kernel itself knows which symlink to attach to which device. So, why not implement this at boot time? > Question: Are you trying to make a floppy-CD combination > that boot only together??? Your message hints that the > kernel is on the floppy and once it is loaded, it will > mount the CD as root filesystem... Sounds like a weird > setup to me. In the extreme case you can't directly boot > from the CD drive, couldn't you rather put e.g. GRUB on > the floppy to help you in booting and prepare the CD > itself so that with the proper hardware it boots alone, > kernel, etc all included? My idea is to create a nice boot floppy, read all 1440Ks of it with dd to boot.img, and use that one on an eltorito bootable CD. David From fokkema@nat.vu.nl Tue Dec 17 05:08:54 2002 Received: with ECARTIS (v1.0.0; list devfs); Tue, 17 Dec 2002 05:08:55 -0800 (PST) Received: from nat.vu.nl (hardy.nat.vu.nl [130.37.36.3]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gBHD8r3v021080 for ; Tue, 17 Dec 2002 05:08:53 -0800 Received: from selene.nat.vu.nl (selene [130.37.32.13]) by nat.vu.nl (8.10.2+Sun/8.10.2) with ESMTP id gBHDCtN10878; Tue, 17 Dec 2002 14:12:55 +0100 (MET) Received: from localhost (fokkema@localhost) by selene.nat.vu.nl (8.11.6+Sun/8.9.3) with ESMTP id gBHDCsB04321; Tue, 17 Dec 2002 14:12:54 +0100 (MET) X-Authentication-Warning: selene.nat.vu.nl: fokkema owned process doing -bs Date: Tue, 17 Dec 2002 14:12:53 +0100 (MET) From: David Fokkema To: Borzenkov Andrey cc: devfs@oss.sgi.com Subject: RE: Boot problems In-Reply-To: <6134254DE87BD411908B00A0C99B044F03A0B59E@MOWD019A> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-archive-position: 51 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: fokkema@nat.vu.nl Precedence: bulk X-list: devfs Content-Length: 2283 Lines: 66 > > It's awfully quiet in here, I hope this mailing list is still alive... > > > > it probably means that devfs does not have any problems :) Very nice! :-) > > > Anyway, I'm having some problems with devfs during boot. My system is > > working fine, that's not the problem. What is, however, is that I want to > > make a bootable CD and I was very glad with the existence of > > /dev/cdroms/cdrom0 and alike. However, my kernel (on a test bootfloppy) > > complains that it can't access that device. It does work if I use > > /dev/ide/host0/bus0/target1/lun0/cd and alike. But then, I need to know > > where my CD is located. Also, on my system, when I want LILO to append a > > root statement to the kernel command line, I can't use > > /dev/discs/disc0/part1 either. Again, I have to use > > /dev/ide/host0/bus0/target0/lun0/part1. > > I can. Probably lilo problem: > > image=/boot/vmlinuz-2.4.20-2mdk > label=test1 > root=/dev/discs/disc0/part5 > read-only > optional > vga=788 > append=" devfs=mount apm=power_off,realmode_power_off" > initrd=/boot/initrd-2.4.20-2mdk.img But wait. Aren't you using an initrd image? I hoped to do this without initrd, but just booting from root=/dev/cdroms/cdrom0. (Or root=cdroms/cdrom0 as Marton seems to suggest?) > > bor@cooker% rpm -qi lilo > Name : lilo Relocations: (not relocateable) > Version : 22.3.2 Vendor: MandrakeSoft > Release : 5mdk > ... > > > In the latter case, I'll survive > > (it's still annoying, though) but in the former, I do want it to work. I > > guess I could use an initrd image which will use /dev/cdroms/cdrom0, but > > that is just extra trouble. I would say that it should work, or else > > should be added to the kernel. > > > > My setup: kernel 2.4.19 compiled with devfs set to mount at boot, cdrom > > driver and isofs included. LILO append line: > > > > append = "devfs=only root=/dev/cdroms/cdrom0" (which doesn't work) > > > > or, (doesn't work either) > > > > append = "devfs=mount root=/dev/cdroms/cdrom0" > > > > If someone knows how to help me, please let me know! > > > > Is cdrom module or statically compiled in kernel? It is statically compiled in the kernel. David From marton.kadar@mail.com Tue Dec 17 05:16:42 2002 Received: with ECARTIS (v1.0.0; list devfs); Tue, 17 Dec 2002 05:16:44 -0800 (PST) Received: from ws1-4.us4.outblaze.com (205-158-62-50.outblaze.com [205.158.62.50]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gBHDGe3v021118 for ; Tue, 17 Dec 2002 05:16:41 -0800 Received: (qmail 41820 invoked by uid 1001); 17 Dec 2002 13:20:40 -0000 Message-ID: <20021217132040.41819.qmail@mail.com> Content-Type: text/plain; charset="iso-8859-1" Content-Disposition: inline Content-Transfer-Encoding: 7bit MIME-Version: 1.0 X-Mailer: MIME-tools 5.41 (Entity 5.404) Received: from [145.236.107.5] by ws1-4.us4.outblaze.com with http for marton.kadar@mail.com; Tue, 17 Dec 2002 21:20:40 +0800 From: "Marton Kadar" To: fokkema@nat.vu.nl, marton.kadar@mail.com Cc: devfs@oss.sgi.com Date: Tue, 17 Dec 2002 21:20:40 +0800 Subject: Re: Boot problems X-Originating-Ip: 145.236.107.5 X-Originating-Server: ws1-4.us4.outblaze.com X-archive-position: 52 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: marton.kadar@mail.com Precedence: bulk X-list: devfs Content-Length: 1945 Lines: 57 Hi David, Right, it would probably be possible to move the symlinks to the namespace layer, but I don't know how much work that would be. OK, I got your idea. Unfortunately I know nothing about how bootable CD-s work. BTW Andrey's LILO setup works only because LILO converts the file reference to block notation when installing itself (not at boot time). You however need the append= parameter, which is interpreted by the kernel at boot time. Marton ----- Original Message ----- From: David Fokkema Date: Tue, 17 Dec 2002 14:07:51 +0100 (MET) To: Marton Kadar Subject: Re: Boot problems > Hi Marton, > > > And the case seems to be that the comfortable symlinks > > you intend to use are symlinks only, and as such they > > can be used only as file references. They in themselves > > are not part of the bare underlying namespace that you > > must use to specify your root device. > > That would be awful. Even without devfsd, the kernel itself knows which > symlink to attach to which device. So, why not implement this at boot > time? > > > Question: Are you trying to make a floppy-CD combination > > that boot only together??? Your message hints that the > > kernel is on the floppy and once it is loaded, it will > > mount the CD as root filesystem... Sounds like a weird > > setup to me. In the extreme case you can't directly boot > > from the CD drive, couldn't you rather put e.g. GRUB on > > the floppy to help you in booting and prepare the CD > > itself so that with the proper hardware it boots alone, > > kernel, etc all included? > > My idea is to create a nice boot floppy, read all 1440Ks of it with dd to > boot.img, and use that one on an eltorito bootable CD. > > David > -- __________________________________________________________ Sign-up for your own FREE Personalized E-mail at Mail.com http://www.mail.com/?sr=signup Meet Singles http://corp.mail.com/lavalife From Andrey.Borzenkov@siemens.com Tue Dec 17 06:34:40 2002 Received: with ECARTIS (v1.0.0; list devfs); Tue, 17 Dec 2002 06:34:43 -0800 (PST) Received: from goliath.siemens.de (goliath.siemens.de [192.35.17.28]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gBHEYc3v026322 for ; Tue, 17 Dec 2002 06:34:40 -0800 Received: from mail1.siemens.de (mail1.siemens.de [139.23.33.14]) by goliath.siemens.de (8.11.6/8.11.6) with ESMTP id gBHEcaQ28915; Tue, 17 Dec 2002 15:38:36 +0100 (MET) Received: from MOWD019A.mow.siemens.ru ([139.24.18.3]) by mail1.siemens.de (8.11.6/8.11.6) with ESMTP id gBHEcZc15161; Tue, 17 Dec 2002 15:38:36 +0100 (MET) Received: by MOWD019A with Internet Mail Service (5.5.2653.19) id ; Tue, 17 Dec 2002 17:51:39 +0300 Received: from mw2b210c (163.242.193.12 [163.242.193.12]) by MOWD019A.mow.siemens.ru with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2653.13) id ZCG3Z9DZ; Tue, 17 Dec 2002 17:51:35 +0300 From: Borzenkov Andrey To: "'David Fokkema'" , "'Marton Kadar'" Cc: devfs@oss.sgi.com Subject: RE: Boot problems Date: Tue, 17 Dec 2002 17:38:26 +0300 Message-ID: <6134254DE87BD411908B00A0C99B044F03A0B59F@MOWD019A> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook, Build 10.0.4510 Importance: Normal In-Reply-To: X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1106 Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by oss.sgi.com id gBHEYc3v026322 X-archive-position: 53 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: Andrey.Borzenkov@siemens.com Precedence: bulk X-list: devfs Content-Length: 2228 Lines: 59 First, sorry, I was not reading carefully enough initially. > > Hi Marton, > > > And the case seems to be that the comfortable symlinks > > you intend to use are symlinks only, and as such they > > can be used only as file references. They in themselves > > are not part of the bare underlying namespace that you > > must use to specify your root device. > > That would be awful. Even without devfsd, the kernel itself knows which > symlink to attach to which device. So, why not implement this at boot > time? > Actually this seems to be implemented, at least in my (vendor-modified) kernel sources; the relevant part is in init/do_mounts.c:create_dev() handle = devfs_find_handle(NULL, dev ? NULL : devfs_name, MAJOR(dev), MINOR(dev), DEVFS_SPECIAL_BLK, 1); Note that we call it with traverse_symlinks = 1 so it _should_ find cdroms/cdrom0 that is symlinked to ../../ide/... Which means that either this part is different in vanilla kernel or this part is executed before cdrom module gets chance to create symlink. I am using 2.4.19-16mdk (Mandrake 9.0). ... hmm ... what happens is, first /dev/root is created for MAJOR/MINOR and then we try to create /dev/root from passed name, but this should fail because /dev/root already exists ... would you try following: in init/do_mounts.c:mount_root() add sys_unlink("/dev/root"); just between devfs_make_root(root_device_name); => sys_unlink("/dev/root"); <= add here create_dev("/dev/root", ROOT_DEV, root_device_name); Note that code is already messy enough, so I am not sure if it is the correct solution in long term. But if this works you may want to drop a note to lkml. [...] > > My idea is to create a nice boot floppy, read all 1440Ks of it with dd to > boot.img, and use that one on an eltorito bootable CD. > Use initrd. It is simple and general enough. Besides, note, that you have no idea if CD-ROM you are booting from will be actually cdrom0. So any code general enough should ask what device has to be mounted as CD-ROM. And, finally, why would you actually _need_ root on CD-ROM? Put any tools you want there, mount it in any convenient place (/tools or /usr) and let / be in RAM. -andrey From fokkema@nat.vu.nl Tue Dec 17 06:48:11 2002 Received: with ECARTIS (v1.0.0; list devfs); Tue, 17 Dec 2002 06:48:13 -0800 (PST) Received: from nat.vu.nl (hardy.nat.vu.nl [130.37.36.3]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gBHEm93v026415 for ; Tue, 17 Dec 2002 06:48:10 -0800 Received: from selene.nat.vu.nl (selene [130.37.32.13]) by nat.vu.nl (8.10.2+Sun/8.10.2) with ESMTP id gBHEqBN20165; Tue, 17 Dec 2002 15:52:11 +0100 (MET) Received: from localhost (fokkema@localhost) by selene.nat.vu.nl (8.11.6+Sun/8.9.3) with ESMTP id gBHEqA115742; Tue, 17 Dec 2002 15:52:10 +0100 (MET) X-Authentication-Warning: selene.nat.vu.nl: fokkema owned process doing -bs Date: Tue, 17 Dec 2002 15:52:10 +0100 (MET) From: David Fokkema To: Borzenkov Andrey cc: "'Marton Kadar'" , Subject: RE: Boot problems In-Reply-To: <6134254DE87BD411908B00A0C99B044F03A0B59F@MOWD019A> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-archive-position: 54 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: fokkema@nat.vu.nl Precedence: bulk X-list: devfs Content-Length: 2787 Lines: 61 > Actually this seems to be implemented, at least in my (vendor-modified) > kernel sources; the relevant part is in init/do_mounts.c:create_dev() > > handle = devfs_find_handle(NULL, dev ? NULL : devfs_name, > MAJOR(dev), MINOR(dev), DEVFS_SPECIAL_BLK, 1); > > Note that we call it with traverse_symlinks = 1 so it _should_ find > cdroms/cdrom0 that is symlinked to ../../ide/... > > Which means that either this part is different in vanilla kernel or this > part is executed before cdrom module gets chance to create symlink. I am > using 2.4.19-16mdk (Mandrake 9.0). > > ... hmm ... what happens is, first /dev/root is created for MAJOR/MINOR and > then we try to create /dev/root from passed name, but this should fail > because /dev/root already exists ... would you try following: in > init/do_mounts.c:mount_root() add sys_unlink("/dev/root"); just between > > devfs_make_root(root_device_name); > => sys_unlink("/dev/root"); <= add here > create_dev("/dev/root", ROOT_DEV, root_device_name); > > Note that code is already messy enough, so I am not sure if it is the > correct solution in long term. But if this works you may want to drop a note > to lkml. Hmmm. I'll try, but I have to take a look at the code to really understand this, I guess. Why would it succeed with ide/host0/etc.? > > My idea is to create a nice boot floppy, read all 1440Ks of it with dd to > > boot.img, and use that one on an eltorito bootable CD. > > > > Use initrd. It is simple and general enough. Besides, note, that you have no > idea if CD-ROM you are booting from will be actually cdrom0. So any code > general enough should ask what device has to be mounted as CD-ROM. And, Got me there. Although I'm not sure the bios, on booting, will try ALL CD-ROM drives, except just the first one. > finally, why would you actually _need_ root on CD-ROM? Put any tools you > want there, mount it in any convenient place (/tools or /usr) and let / be > in RAM. Well, by 'hardwiring' the floppy to ide/host0/etc. and a non-bootable CD-ROM in the drive, I managed to put a whole server system on CD-ROM, including the RPMS I created for the distribution, and some simple install scripts. /tmp, /var and /etc are recreated in tmpfs, including original directories and files of /var and /etc. So, basically, if I boot from floppy and CD, I have a fully working linux system (w/o X) which I can test on different servers with different hardware. And also, I can install them from the CD. I use devfs without symlinks, and so far it works, although on my home system, I need the compatibility links for scsi devices for cdrtools. I guess the device locations are wired into the binaries. I will try the patch and try initrc. Thanks for the help! David From Andrey.Borzenkov@siemens.com Tue Dec 17 07:06:28 2002 Received: with ECARTIS (v1.0.0; list devfs); Tue, 17 Dec 2002 07:06:30 -0800 (PST) Received: from david.siemens.de (david.siemens.de [192.35.17.14]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gBHF6Q3v026685 for ; Tue, 17 Dec 2002 07:06:27 -0800 Received: from mail1.siemens.de (mail1.siemens.de [139.23.33.14]) by david.siemens.de (8.11.6/8.11.6) with ESMTP id gBHFAR012535; Tue, 17 Dec 2002 16:10:27 +0100 (MET) Received: from MOWD019A.mow.siemens.ru ([139.24.18.3]) by mail1.siemens.de (8.11.6/8.11.6) with ESMTP id gBHFAQc19844; Tue, 17 Dec 2002 16:10:26 +0100 (MET) Received: by MOWD019A with Internet Mail Service (5.5.2653.19) id ; Tue, 17 Dec 2002 18:23:30 +0300 Received: from mw2b210c (163.242.193.12 [163.242.193.12]) by MOWD019A.mow.siemens.ru with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2653.13) id ZCG3Z9MT; Tue, 17 Dec 2002 18:23:24 +0300 From: Borzenkov Andrey To: "'David Fokkema'" Cc: "'Marton Kadar'" , devfs@oss.sgi.com Subject: RE: Boot problems Date: Tue, 17 Dec 2002 18:10:15 +0300 Message-ID: <6134254DE87BD411908B00A0C99B044F03A0B5A1@MOWD019A> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook, Build 10.0.4510 Importance: Normal In-Reply-To: X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1106 X-archive-position: 55 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: Andrey.Borzenkov@siemens.com Precedence: bulk X-list: devfs Content-Length: 465 Lines: 12 > Hmmm. I'll try, but I have to take a look at the code to really understand > this, I guess. Why would it succeed with ide/host0/etc.? > you are right, create_dev removes file first so it is not a problem. Which leaves us with original idea that symlink is not yet created at this point. Have you also compiled ide-cd in kernel? Sorry for stupid question. If you are truly interested you could also recompile with devfs debug and watch what happens :) -andrey From fokkema@nat.vu.nl Wed Dec 18 04:54:11 2002 Received: with ECARTIS (v1.0.0; list devfs); Wed, 18 Dec 2002 04:54:15 -0800 (PST) Received: from nat.vu.nl (hardy.nat.vu.nl [130.37.38.3]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gBICs93v025522 for ; Wed, 18 Dec 2002 04:54:11 -0800 Received: from selene.nat.vu.nl (selene [130.37.32.13]) by nat.vu.nl (8.10.2+Sun/8.10.2) with ESMTP id gBICwEN09239; Wed, 18 Dec 2002 13:58:14 +0100 (MET) Received: from localhost (fokkema@localhost) by selene.nat.vu.nl (8.11.6+Sun/8.9.3) with ESMTP id gBICwBT16605; Wed, 18 Dec 2002 13:58:11 +0100 (MET) X-Authentication-Warning: selene.nat.vu.nl: fokkema owned process doing -bs Date: Wed, 18 Dec 2002 13:58:11 +0100 (MET) From: David Fokkema To: Borzenkov Andrey cc: "'Marton Kadar'" , Subject: RE: Boot problems In-Reply-To: <6134254DE87BD411908B00A0C99B044F03A0B5A1@MOWD019A> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-archive-position: 56 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: fokkema@nat.vu.nl Precedence: bulk X-list: devfs Content-Length: 710 Lines: 20 > > Hmmm. I'll try, but I have to take a look at the code to really understand > > this, I guess. Why would it succeed with ide/host0/etc.? > > > > you are right, create_dev removes file first so it is not a problem. Which > leaves us with original idea that symlink is not yet created at this point. > Have you also compiled ide-cd in kernel? Sorry for stupid question. Yes, I have. Otherwise /dev/ide/host0/etc wouldn't have worked either. > > If you are truly interested you could also recompile with devfs debug and > watch what happens :) I will! Also, kernel 2.4.20 is available now, so I'll compile that one. Maybe this whole thing was considered a bug and fixed already. That would be nice. David From racsw@frontiernet.net Thu Dec 19 15:53:32 2002 Received: with ECARTIS (v1.0.0; list devfs); Thu, 19 Dec 2002 15:53:34 -0800 (PST) Received: from relay01.roc.frontiernet.net (relay01.roc.frontiernet.net [66.133.131.34]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gBJNrV3v001352 for ; Thu, 19 Dec 2002 15:53:32 -0800 Received: (qmail 5346 invoked from network); 19 Dec 2002 23:57:47 -0000 Received: from unknown (HELO localhost.localdomain) ([65.37.13.251]) (envelope-sender ) by relay01.roc.frontiernet.net (FrontierMTA 2.3.1) with SMTP for ; 19 Dec 2002 23:57:47 -0000 Content-Type: text/plain; charset="iso-8859-1" From: Robert To: Georgi Georgiev , Devfs Subject: Re: Host0, Host2 for IDE devices Date: Thu, 19 Dec 2002 18:57:46 -0500 User-Agent: KMail/1.4.3 References: <200212102033.17836.racsw@frontiernet.net> <3DFA2EE0.3050901@gg3.net> In-Reply-To: <3DFA2EE0.3050901@gg3.net> MIME-Version: 1.0 Message-Id: <200212191857.46337.racsw@frontiernet.net> Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by oss.sgi.com id gBJNrV3v001352 X-archive-position: 57 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: racsw@frontiernet.net Precedence: bulk X-list: devfs Content-Length: 1885 Lines: 47 Georgi, Thanks for replying. I apologize for getting back so late, but Xmas has been consuming my time. I really would like to understand this logic. I'm trying to help this guy put a "part4" block device in the right location on his system so his Zip drive works correctly and because he has both host0 and host2 sub-directories under /dev/ide, I'm not sure where to locate it. If you or anyone else finds a document explaining the "hostx" business, I sure would appreciate a link to it's location. Thanks again, Robert On Friday 13 December 2002 2:02 pm, you wrote: > I read somewhere that host0 and host1 are assigned for on-board IDE > interfaces, and IDE controllers on PCI slots receive host2 and up. > > I was a little surprised myself, when the IDE controller I bought recently > became host2... > > Robert wrote: > > Hello, > > I'm trying to understand the structure of host0, host1, host2, etc in > > regards to IDE devices. I did read Richards FAQ, and he did discuss > > this for scsi devices, but I didn't see the logic explained for the IDE > > devices or whatever determines the host# scheme. ( ...or it was in > > there, but I didn't understand it ) If someone could explain this, I > > sure would appreciate it. > > > > In regards to the /bus#/target# logic, maybe someone could confirm my > > understanding of this as well as I have listed below. > > > > /bus0/target0 = primary master > > /bus0/target1 = primary slave > > /bus1/target0 = secondary master > > /bus1/target1 = secondary slave. > > > > I'm trying to get this fellows internal zip drive working on his > > system, but because he has a different HD and Zip configuration off his > > IDE slots than I do, it's necessary I understand how to figure this out > > so I can get his block device (part4) in the right place for his system. > > > > Thanks, > > Robert From chutz@chubaka.homeip.net Thu Dec 19 17:07:19 2002 Received: with ECARTIS (v1.0.0; list devfs); Thu, 19 Dec 2002 17:07:22 -0800 (PST) Received: from mail.my.home (jnydnag@142.13.111.219.st.excite.co.jp [219.111.13.142]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gBK17F3v007232 for ; Thu, 19 Dec 2002 17:07:18 -0800 Received: (qmail 29508 invoked from network); 20 Dec 2002 01:11:29 -0000 Received: from tank.my.home (10.0.0.2) by 0 with QMQP; 20 Dec 2002 01:11:29 -0000 Date: Fri, 20 Dec 2002 10:11:28 +0900 From: Georgi Georgiev To: devfs@oss.sgi.com Cc: devfs@oss.sgi.com Subject: a fundamental question Message-ID: <20021220011128.GA1318@tank.my.home> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.4i X-archive-position: 58 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: chutz@gg3.net Precedence: bulk X-list: devfs Content-Length: 1131 Lines: 21 I see there is something that obviously I had not understood for a long time, so.... Aren't symlinks the responsibility of the devfsd daemon? I mean, that's the whole point in calling them *compatibility* symlinks isn't it? I mean... one can run his system with no problems even without running devfsd, they only have to mount /dev, but they would then have to say farewell to all the nice features devfsd provides - compatibility symlinks, on-demand module loading etc. etc. A system with mounted /dev and NOT having devfsd running, has no symlinks in /dev, does it? Short of stdin, stdout etc I guess. Is this something I am completely confused about? I figured a long time ago, that when I am the only one with a different opinion, it is usually so, because I am wrong. But, please, explain. -- /^^^^^^^^^^^^^^^^^^^^^^^^^^^\/^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\ / Georgi Georgiev (-< / That's what she said. \ \ chutz@chubaka.net /\ .o)\ / / +81(90)6266-1163 V_/_ |(/)/ \ \___________________________/\__________________________________/ From kpfleming@cox.net Thu Dec 19 18:25:32 2002 Received: with ECARTIS (v1.0.0; list devfs); Thu, 19 Dec 2002 18:25:35 -0800 (PST) Received: from fed1mtao03.cox.net (fed1mtao03.cox.net [68.6.19.242]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gBK2PW3v009077 for ; Thu, 19 Dec 2002 18:25:32 -0800 Received: from jeeves.kpf.internal ([24.56.45.215]) by fed1mtao03.cox.net (InterMail vM.5.01.04.05 201-253-122-122-105-20011231) with ESMTP id <20021220022943.XAMM4125.fed1mtao03.cox.net@jeeves.kpf.internal> for ; Thu, 19 Dec 2002 21:29:43 -0500 Received: from [192.168.172.107] (helo=cox.net) by jeeves.kpf.internal with esmtp (Exim 4.05) id 18PCv1-0000c1-00 for devfs@oss.sgi.com; Thu, 19 Dec 2002 19:29:43 -0700 Message-ID: <3E02809A.3010606@cox.net> Date: Thu, 19 Dec 2002 19:29:46 -0700 From: "Kevin P. Fleming" User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.2b) Gecko/20021016 X-Accept-Language: en-us, en MIME-Version: 1.0 CC: devfs@oss.sgi.com Subject: Re: a fundamental question References: <20021220011128.GA1318@tank.my.home> In-Reply-To: Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-archive-position: 59 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: kpfleming@cox.net Precedence: bulk X-list: devfs Content-Length: 1476 Lines: 36 Georgi Georgiev wrote: > I see there is something that obviously I had not understood for a > long time, > so.... > > Aren't symlinks the responsibility of the devfsd daemon? Some are, yes. Those designed to emulate older /dev entries are the responsbility of _some_ userspace application, usually devfsd. > A system with mounted /dev and NOT having devfsd running, has no > symlinks in /dev, does it? Short of stdin, stdout etc I guess. Not at all; my main server uses devfs and does not even have devfsd installed. All the normal devfs entries are there, include /dev/discs, /dev/cdroms, etc. > Is this something I am completely confused about? I figured a long > time ago, > that when I am the only one with a different opinion, it is usually > so, because > I am wrong. But, please, explain. Yes, you are mistaken :-) However, if this is in reference to the recent discussion about mounting root filesystems and not having "short" /dev paths available, this is where the confusion comes from. For some reason, the /dev _namespace_ inside the kernel is a distinct entity from the devfs _filesystem_ that normally gets mounted on /dev. All the /dev _namespace_ entries appear in the devfs filesystem, but also there are many symlinks to make things easier to use. I don't know if those symlinks are visible in the /dev namespace inside the kernel, but I think not since many people have trouble using them as arguments to the "root=" kernel parameter. From fokkema@nat.vu.nl Fri Dec 20 01:58:08 2002 Received: with ECARTIS (v1.0.0; list devfs); Fri, 20 Dec 2002 01:58:10 -0800 (PST) Received: from nat.vu.nl (hardy.nat.vu.nl [130.37.38.3]) by oss.sgi.com (8.12.5/8.12.5) with SMTP id gBK9w73v014807 for ; Fri, 20 Dec 2002 01:58:08 -0800 Received: from selene.nat.vu.nl (selene [130.37.32.13]) by nat.vu.nl (8.10.2+Sun/8.10.2) with ESMTP id gBKA2LN21473 for ; Fri, 20 Dec 2002 11:02:22 +0100 (MET) Received: from localhost (fokkema@localhost) by selene.nat.vu.nl (8.11.6+Sun/8.9.3) with ESMTP id gBKA2K502662 for ; Fri, 20 Dec 2002 11:02:21 +0100 (MET) X-Authentication-Warning: selene.nat.vu.nl: fokkema owned process doing -bs Date: Fri, 20 Dec 2002 11:02:20 +0100 (MET) From: David Fokkema To: devfs@oss.sgi.com Subject: Devfs symlinks, booting etc. Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-archive-position: 60 X-ecartis-version: Ecartis v1.0.0 Sender: devfs-bounce@oss.sgi.com Errors-to: devfs-bounce@oss.sgi.com X-original-sender: fokkema@nat.vu.nl Precedence: bulk X-list: devfs Content-Length: 938 Lines: 22 Hi there! I think I have something to add to the discussion about symlinks and booting problems. If you followed the discussion about the booting, you know that /dev/cdroms/cdrom0 etc. were not supported during boot. We were a bit confused about that, since they appear in /dev after boot (even without devfsd). So indeed, Kevin is right. _Some_ symlinks end up in /dev anyway, without a user space program like devfsd. It is the compatibility links that need devfsd. But why do people (like me) have problems booting from /dev/cdroms/cdrom0 or /dev/discs/disc0/part1? That is because they don't run kernel 2.4.20. I've compiled it yesterday and it works fine. Apparently, the missing of those links during boot was considered a bug and is now fixed. So, the only thing you need devfsd for is to create compatibility symlinks or to do module autoloading when /dev/floppy/0 is looked up, for instance. Thanks for all the help, David