I-Appliance BBS
The Official Source for Internet Appliance Upgrades and Mods
Amazon Honor System Click Here to Pay Learn More
BBS Main List | Sign In | Sign Up | Search | Help | Linux-Hacker.netReply to Thread | Printer |

Home / Linux / Midori Linux General Board
Midori Boot Problems (Installation Steps)
Midori Boot Problems (Installation Steps)

New MessageMidori Boot Problems (Installation Steps) (modified 0 times) Adrenolin
Profile | Email
Hello,
Let me apologize in advance for the length of this post. I felt it important to supply all details as I'm currently at a loss for ideas.

Background and Steps Followed:
I've successfully installed Midori Linux onto a V5 IO but it refuses to install correctly on 3 other V5 systems (all 4 units were new and unused). After flashing the BIOS chips with a BadFlash v5.40 image I installed a 6Gig Fujitsu laptop hard drive running MS-DOS 6.2 freshly installed. This disk included dolly and m4i-2.2pre1.img which I saved as midori.img. Both files had been downloaded directly to the hard drive and unmodified once saved to disk.

Booting the first V5 IO I saw the V5.40 BIOS, Primary IDE Master was the hard drive and the Secondary IDE Slave was the SanDisk. BIOS settings had been set to AUTO detect both and to boot from 'C Only'. DOS started fine and gave me a prompt. I then ran fdisk, switched to the SanDisk, removed the partitions, create a new DOS partition and rebooted. After this I ran: "dolly midori.img hd129: /c" which completed fine. At this point I shutdown the IO, removed the hard drive and cable and restarted it.

It booted and listed 'none' as Master IDE and the SanDisk as the Slave IDE. It continued to boot through displaying the Midori bootup, heard sound initialize and then X started up! WooHoo Its Alive!! Checked ctl-alt-F4 and ctl-alt-F1 which gave no errors. DHCP worked great and I was able to statically assign an IP and everything works perfect. Note here that I have reflashed the BIOS chip in this unit again after this install to the current V5.40a BadFlash BIOS version and it continues to run perfectly. I was able to stream MP3s over the network for a 24hour period without any problems.

The Problem:
OK so the first IO is up and running perfectly and was very smooth to install. The problem now is that I've followed the exact same proceedure for 3 other identical V5 IOs and they all end in failure. Dolly seems to copy the image fine and says it completed successfully. I then reboot the IO (removing the hard drive and cable as before) only to have the Midori boot proceedure screwup with errors ('/tmp/loop30: No such device' is the first) leaving me at a prompt. I have attempted a number of different 'solutions' found on this BBS without success. I have replaced the flashed V5.40 with a flashed V5.40a and with an original V5.40a chip from BadFlash, swapped out IDE cables and have run Dolly with the /c switch and without. I've also redownloaded the dolly program and the Midori image to a different directory and tried those. I've formatted the hard drive, reinstalled DOS and downloaded dolly and the Midori image again. I've run fdisk with the /mbr switch, checked the md5sums of the images, run scandisk on both the hard drive and SanDisk and still nothing is working! I dont know what to try next!

Here is the Midori Boot sequence with errors:

Starting init (return here with ctl-alt-F1)...
Fixing USB by tweaking PCI registers...
Mounting RW filesystems...
Building devpts...
Creating loopback interfaces...
Mounting usr partition...
/tmp/loop30: No such device
packcramfs: error opening loopback device
Thawing config partition to /tmp/config...
/tmp/loop30: No such device
packcramfs: error opening loopback device
/tmp/loop30: No such device
packcramfs: error opening loopback device
Setting timezone...
Starting logd (use ctl-alt-F4 to see it)...
Starting network...
.: Can't open /tmp/config/network
/sbin/portmap: error in loading shared libraries: libnsl.so.1: cannot open shared object file: No such file or directory
Loading hardware monitoring modules (run 'sensors' to see values)...
/sbin/init: /usr/bin/sensors: No such file or directory
spawming shell (return to root shell with ctl-alt-F1)...


BusyBox v0.60.2 () Built-in shell (ash)
Enter 'help' for a list of built-in commands.

#_

Again, this has worked on 1 system perfectly and failed on 3 others all of which are new/unused V5s with no other changes or modifications except for the BIOS. If anyone can offer any suggestions I'd appreciate it muchly!

Thanks,
Adrenolin
i-opener@irchottub.net

01-29-2003 01:23:11

New MessageRE:Midori Boot Problems (Installation Steps) (modified 0 times) bob94025
Profile
I'm at a loss. It is failing to mount /usr, and the rest of the errors follow from that. I saw a similar error on my IA-1 when I tried a kernel that didn't have midori's 'addpart' patch. However, you are using the standard Midori kernel, which should be fine.

It could be that the loopback devies aren't getting created. I assume that you can use the shell, right? Try poking around and see if *any* of the loopback devices have been created (look in /tmp). Also, try running /bin/uplo and see if it throws any errors.

It could also be that the usr partition is corrupt, and thus not mounting properly. This could happen if the sandisk on the problematic Iopeners is a slightly different size (not all "16MB" have exactly the same # of bytes) and/or it has some bad sectors. Can you run a disk verification utility from DOS? (I'm not a DOS guy, so I wouldn't even know what to run- 'scandisk' maybe?)

I assume that you configured the BIOS settings the same as for the successful Iopener, right?

One more thing- at some point, I added a PCI tweak that is supposed to fix some USB data corruption issues on the Iopener (see http://vischeck.homeip.net/m4i/pub/extras/usbFix.txt ). I suppose it is possible that this 'fix' is causing problems? To check this, you can build a new image from a build-kit, after editing /sbin/init to remove this tweak. (If you go this route, I'd suggest starting with the 2.2 build kit). Or, if you don't want to mess with building images, you can try one of my older images (2.0) just to see if that's the problem.

bob

01-29-2003 11:35:45

New MessageRE:Midori Boot Problems (Installation Steps) (modified 0 times) Adrenolin
Profile | Email
bob94025.. Thanks for your reply!

Resolved! I setup my hard drive to dual boot Win98 & Debian Linux. I deleted the partitions on the SanDisk from Win98 and attempted to dolly the m4i-2.2pre1.img image. Again it completed correctly so I removed the drive, rebooted and got the same errors! I also attempted to format the SanDisk as a FAT and FAT16 FS and running dolly. Same errors again.
Getting tired of this I booted into Linux and ran "cfdisk -z /dev/hdb" (cfdisk errors if you dont force it with the -z). At this point cfdisk didnt display any partitions at all so I created a new partition, saved it and rebooted. After rebooting linux again I simply typed "dd if=m4i-2.2pre1.img of=/dev/hdb". Shutting down the system, removing the hard drive and turning it back on it boots directly into Midori and is running great!

I have now completed this on 2 more units all successfully! Looks like for some reason Win98 was corrupting either the Sandisk geometry or the image during transfer.

Now I did notice that when I formatted the SanDisk from Win98 Windows Explorer only showed that the SanDisk had 15.3MB of available space. While in Linux cfdisk shows it had 16.3MB or 16.6MB (I forget which right now) of available space. I'm not entirely sure what Win98 was doing wrong but it failed on 3 of the 4 systems. Linux on the other hand worked flawlessly on them all.

Adrenolin

02-01-2003 21:58:30

New MessageRE:Midori Boot Problems (Installation Steps) (modified 0 times) Jeff102410
Profile | Email
Well...now that you found the problem...

I had a different set of problems that were occurring when trying to use a CF/IDE adapter. When the CF/IDE adapter was plugged in, I could not get either the built in SanDisk or the CF/IDE adapter to boot (V540 BIOS). Removing the CF/IDE and the IOpener would boot, and putting the CF/IDE in another desktop PC, the CF/IDE would boot.

The resolution was to use RanishPartitionMagic instead of FDisk. For some reason, FDisk was not setting the master boot record (or something). Anyway, RPM worked.

On another unit, the SanDisk was FDisk'ed and formatted with DOS. I could not get Midori to install...and after several attempts, it still would not work correctly. I want back, set the SanDisk with RPM, set the boot partition with RPM, formatted the SanDisk with DOS, and booted again to make sure it was working. Next, hooked up a second HD, Dolly'ed the Midori image, and everything worked...first time.

Hope this helps.

02-02-2003 22:54:26

New MessageRE:Midori Boot Problems (Installation Steps) (modified 0 times) Jeff102410
Profile | Email
Also...RPM has to be run from DOS. Running it from a Win98 command prompt does not seems to set the drive correctly. I usually just shutdown/exit to MSDOS.
02-02-2003 22:55:40

New MessageRE:Midori Boot Problems (Installation Steps) (modified 0 times) ravic
Profile
I am having the exact same issue. It keeps saying "no loopback device". Is there anyway to fix this via DOS.
02-27-2003 17:49:23

Reply to Thread | Printer |
All times are PSTPowered by UltraBoard v1.62



Copyright © 2000, Netmake Inc. All Rights Reserved.
See Terms and Conditions for more information.




i-opener opener laptop notebook computer help drivers dll free windows dos repair fix linux mac macintosh 2000 95 98 nt pc configure hardware software sound video netscape explorer network networking lan wan software cmos fat bios printer card mouse modem ide scsi cd rom controllers scanner tape hard drive cgi scripts source code mp3