Hi there, I'm hoping someone can help me - I'm having difficulty installing a new drive into my mk 2 empeg.

It currently has an original 18Gb disk and an aftermarket 20Gb disk, and I'm trying to replace them both with a single 80Gb disk.

Unfortunately, the builder ram image seems to be unable to format the new drive - I have tried it on the first and second IDE connector, with each of the master/slave/CS settings (it only boots with CS), etc. The disk is a Fujitsu MHT2080AT.

Here is a log of the unit's output booting with the new drive, and the builder image (captured via hyperterminal):

empeg-car bootstrap v1.00 20000601 ([email protected])
If there is anyone present who wants to upgrade the flash, let them
speak now,
or forever hold their peace...it seems not. Let

e000 v1.04
Copying kernel...
Calling linux kernel...
Uncompressing Linux..................................... done, booting
the kerne
l.
Linux version 2.2.14-rmk5-np17-empeg49 ([email protected])
(gcc versi
on 2.95.3 20010315 (release)) #130 Mon Oct 22 18:52:07 BST 2001
Processor: Intel StrongARM-1100 revision 11
NetWinder Floating Point Emulator V0.94.1 (c) 1998 Corel Computer Corp.
empeg-car player (hardware revision 7, serial number 80000279)
Command line: mem=12m
Calibrating delay loop... 207.67 B
Memory: 10968k/12M available (964k code, 20k reserved, 332k data, 4k
init)s.cpp 18
IP Protocols: ICMP, UDP, TCPd=5
Dentry hash table entries: 2048 (order 2, 16k)hash 16384 bhash
16384)econds
Buffer cache hash table entries: 16384 (order 4, 64k)
IrDA (tm) Protocols for Linux-2.2 (Dag Br
Page cache hash table entries: 4096 (order 2, 16k)
Stressi
Linux-IrDA: IrCOMM protocol ( re
POSIX conformance testing by UNIFIXs.cpp 189 ( 14): Poll result=0
Linux NET4.0 for Linux 2.2
Based upon Swansea University Computer Society NET3.039
ircomm_tty: virtual tty driver for IrCOMM ( re
NET4: Linux TCP/IP 1.0 for NET4.0
IP Protocols: ICMP, UDP, TCP
hdstress.cp
TCP: Hash tables configured (ehash 16384 bhash 16384)
Starting kswapd v 1
IrDA (tm) Protocols for Linux-2.2 (Dag Brattli)serial driver version
4.27
empeg audi
ttyS00 at 0xf8010000 (irq = 15) is a SA1100 UART
hda: FUJI
ttyS01 at 0xf8050000 (irq = 17) is a SA1100 UARTsed.
ttyS02 at 0xf8030000 (irq = 16) is a SA1100 UART
empeg usb initialised, PDIUSBD12 id 1012
Signature is ffffffff 'ÿÿÿÿ'
hda: FUJITSU MHT208
empeg display initialised.
empeg dsp audio initialisedalised 0089/88c1 (save to d
empeg dsp mixer initialised
empeg dsp initialisedflash driver
hdb: FUJITSU MHT2080AT, ATA DISK drive 320 bl
RAM disk driver initialized: 16 RAM disks of 4096K size
hdb: FU
empeg single channel IDE drive
Probing primary interface...warn
hdb: FUJITS
hda: FUJITSU MHT2080AT, ATA DISK driveed
hda: FUJITSU MHT2080AT, ATA DISK driverupt, status=0xff, count=1unted
root (
hda: FUJITSU MHT2080AT, ATA DISK drive
hda: FUJITSU MHT2080AT, 76319MB w/8192kB Cache, CHS=9729/255/63nmount
old root ...
empeg-flash driver initialized) at 0x4008000 IRQ:7 INTF:TP M
smc chip id/revision 0x33491
smc9194.c:v0.12 03/06/96 by Erik Stahlman ([email protected])ounting first
music partition

SMC9194: SMC91C94(r:9) at 0x4008000 IRQ:7 INTF:TP MEM:6144b MAC
00:02:d7:10:01:1
hdb: hdb1 < hdb5 hdb6 > hdb2 hdb3 hdb4
7
Partition check:
VFS: Can't
hda: hda1 < hda5 hda6 > hda2 hda3 hda4ISK: ext2 filesystem found at
block 0
RAMDISK: ext2 filesystem found at block 0

empeg-pump v0.03 (19980601)
Press Ctrl-A to enter pump...(ext2 filesystem) readonly.
change_root: old root has d_count=1
to unmount old root ... okay
Freeing unused kernel memory: 4k initVFS: Can't find an ext2 filesystem
on dev i
de0(3,4).
mount: wrong fs type, bad option, bad superblock on /dev/hda4,
or too many mounted file systems
warning: can't open /etc/mtab: No such file or directory
umount: /drive0: not mounted
Making first drive...
mke2fs 1.14, 9-Jan-1999 for EXT2 FS 0.5b, 95/08/09
ext2fs_check_if_mount: No such file or directory while determining
whether /dev/
hda4 is mounted.
no room for private writable mapping
error: -12
/dev/hda4: Memory allocation failed while setting up superblock
hdstress.cpp 189 ( 14): Poll result=0 on fd=5
hdstress.cpp 189 ( 14): Poll result=0 on fd=5
Stressing /dev/hda: Spinning down.
hdstress.cpp 189 ( 14): Poll result=0 on fd=5
Stressing /dev/hda: Spinning up.

Thanks for any information,
Matt

You want to read this and this.

Ok, thanks.

Any idea why the builder image doesn't work directly? Is 80Gb too large for <something>?

Fair enough:)

Another problem (sorry to be montonous): I have put the current developer image back onto the player, and I'm now trying to quit the player and drop to the shell, as per the FAQ entry - however, the player does not give any indication that it is accepting commands from hyperterminal. In fact, the player app may have exited - is there some way I can access the shell from this position?

The boot log now ends with:

Press 'q' to go into development mode. You Have Zero Seconds to Compl
Starting player
player.cpp : 385:empeg-car 2.00 2003/04/01
! tags.cpp : 61:Failed to open tags (0xc0041002)
Prolux 4 empeg car - 2.1434 Mar 26 2003
Vcb: 0x4054d000

Thanks again.

Control-C ???

Had you succeeded in building the drive at this point?

No, the builder image wasn't able to format the drive, so it's still unformatted.

Okay... So you loaded developer on an unbuilt drive? And it worked? I must be missing something - what is your current status / configuration?

Well, I initially loaded the builder image, but it was unable to format the disk. I was then directed to the manual partitioning-and-formatting notes by Roger - to do this I assumed I need to get a shell on the player. So, I loaded the developer image back onto the player's flash. This image boots, but it is unable to play anything, presumably because it can't find any files to play.
The player image seems fine, as does the visuals application - the only problem is that I don't seem to be able to get the player to quit to let me use the shell.

Can you hit control-c right just before it starts looking for the player application? Don't wait for the prompt.

No, hitting Ctrl-C continually doesn't do anything. Either the window of opportunity is too small, or the player isn't receiving any input at all.

A cheap solution.

You are upgrading from two smaller drives. Make sure you load the 2.0 kernel. Install the original master drive in the player and the new 80GB as the slave. Boot and drop to shell. From here, you will be able to build the new drive using the fully installed kernel from your old master. Just be careful not to destroy the data on the old drive (make sure that all your commands point to the secondary drive, hdc...)

Hmmm, that's scary.
If I boot from the old master drive, and then drop to the shell, can I disconnect the valid drive before I issue any commands (which might trash the first drive if I get something wrong)?
Or, does the shell need the valid drive to be available for loading commands?

I think the shell needs the valid drive for the commands to run. Basically, you would need to build the filesystem and all, so that is probably necessary. Either that or do you have access to a Mark2A? There should be enough memory there to finish the build.

Ok, I'll give that a shot - I don't have a Mk2a.

Thanks for the advice, I'll let you know how I get on.

No. Don't try to disconnect the drives when it's running. It needs to load stuff off the disk and it also uses the disk as a swap partition. This is not to mention you could blow something important if you start doing this.

Agreed. Not to mention that you could do very nasty things that way. Leave the hard drive in place. I looked through the instructions, and they are all for doing exactly this - booting from a good master and building a slave drive. Just follow them carefully.

Well, I have made some progress.

First of all, the fact that I couldn't quit out of the player was something to do with hyperterminal not flushing changes; I tried many changes to HT settings, none of which worked, and then eventually after a restart of HT, communication worked again.

Now, I have followed Roger's partitioning+formatting instructions, with the new drive as a slave drive. This worked well, and the player was able to mount and access the new drive correctly. Now, I want to move the new drive to be the master drive (and eventually the only drive), but the player won't boot with the new drive as the master.
It reports:
Kernel panic: VFS: Unable to mount root fs on 03:05

So, my next question is: can the empeg have an 80Gb disk as the master disk?

Yes. I am running dual 80s.

Now, you partitioned and formatted the drive. Next, remove the original drive, install your 80 as master and reload the 2.0 kernel. This will (or should) pump the partition to populate all of the supporting files.

Edit: Hmmm... Looking through the instruction pages, I saw nothing about making a root filesystem. Roger - is that missing from the instructions?

Gah! Too much solaris. I keep expecting the root filesystem on slice 0... Duh... Sorry about that.

Yes, that's what it was. I did read that last line, but I thought to myself, I already have a player image on there...
I guess I just didn't realise what 'pumping partition' meant.

Anyway, all's well that ends well. Thanks for your help, everybody.

Okay, I'm going to use all of the above as an example of the current desperate need to get those builder images updated.

PLEASE SOMEONE AT EMPEG PLEASE MAKE NEW BUILDER IMAGES THAT CAN HANDLE BIGGER DRIVES? PPPPPPLEASE?

Had a think about it and we shouldn't need to have the empeg guys do this. The steps required are already on Roger's page so making a new builder image is possible.

I don't have any 80GB drives to test it on however. Only a spare 6GB drive.

Any volunteers to test it if I make one?

I would be more than willing to exhaustively test it with 100GB drives...

Yeah. I'd love to hit that option accidentally.

Don't think it is a good idea to integrate it into the kernel. It's only ever going to be called once per drive and the rest of the time it'll be taking up memory which could be put to better use.

Worked for me!

You should do it this way:

- install your 80G as Master on the cable at position number 1 (closest connector to the motherboard)

- flash the player with the Mk2 Builder image, and let it run

- flash the player with the developer image, then follow the instructions in Roger's article to format the music partition and build the player directories

- connect with emplode and allow the FSCK to run and get a built database together.

You should be flying after this.