Buffalo NAS-Central Forums

Welcome to the Linkstation Wiki community
It is currently Sat Jul 21, 2018 4:27 pm

All times are UTC+01:00




Post new topic  Reply to topic  [ 43 posts ]  Go to page Previous 1 2 3 Next
Author Message
PostPosted: Fri Aug 14, 2009 4:52 am 
Offline
Newbie

Joined: Thu Dec 06, 2007 9:58 am
Posts: 22
Code:
printenv
nfsboot=bootp;run nfsargs;bootm
baudrate=57600
autoload=no
stdin=nc
stdout=nc
stderr=nc
netmask=255.255.255.0
nc=setenv stdin nc;setenv stdout nc;setenv stderr nc
ser=setenv stdin serial;setenv stdout serial;setenv stderr serial
ldaddr=800000
hdpart=0:1
hdfile=boot/uImage
hdload=echo Loading ${hdpart}:${hdfile};ext2load ide ${hdpart} ${ldaddr} ${hdfil
e};ext2load ide ${hdpart} 7f0000 boot/kuroboxHG.dtb
boothd=setenv bootargs root=/dev/sda1 netconsole=6666@192.168.11.150/,@192.168.1
1.149/ rtc-rs5c372.probe=0,0x32;bootm ${ldaddr} - 7f0000
hdboot=run hdload boothd
flboot=setenv bootargs root=/dev/hda1;bootm ffc00000
emboot=setenv bootargs root=/dev/ram0;bootm ffc00000
nfsargs=setenv bootargs root=/dev/nfs rw nfsroot=${serverip}:${rootpath} ip=${ip
addr}:${serverip}:${gatewayip}:${netmask}:${hostname}::off
bootcmd1=run hdboot;run flboot
bootcmd2=run flboot
bootcmd3=run emboot
writeng=protect off fff70000 fff7ffff;era fff70000 fff7ffff;mw.l 800000 4e474e47
 1;cp.b 800000 fff70000 4
writeok=protect off fff70000 fff7ffff;era fff70000 fff7ffff;mw.l 800000 4f4b4f4b
 1;cp.b 800000 fff70000 4
ubpart=0:1
ubfile=share/u-boot/u-boot-hg.flash.bin
ubload=echo Loading ${ubpart}:${ubfile};ext2load ide ${ubpart} ${ldaddr} ${ubfil
e}
ubsaddr=fff00000
ubeaddr=fff2ffff
ubflash=protect off ${ubsaddr} ${ubeaddr};era ${ubsaddr} ${ubeaddr};cp.b ${ldadd
r} ${ubsaddr} ${filesize};cmp.b ${ldaddr} ${ubsaddr} ${filesize}
upgrade=run ubload ubflash
ethact=RTL8169#0
bootargs=root=/dev/hda1
bootretry=60
bootdelay=10
hdpart3=0:3
hdload3=echo Loading ${hdpart3}:${hdfile};ext2load ide ${hdpart3} ${ldaddr} ${hd
file}
boothd3=setenv bootargs root=/dev/hda3;bootm ${ldaddr}
hdboot3=run hdload3 boothd3
bootcmd4=run hdboot3;run hdboot;run flboot
netretry=no
serverip=192.168.2.9
ncip=192.168.2.9
ipaddr=192.168.2.12
bootcmd=run bootcmd1

Environment size: 1817/65532 bytes
=>


When I put variant-uboot . variant-legacy . variant-ug in /boot and remove variant-latest
Run webinstall again, My LS comes back.
It can boot in 2.6.25.20, but after upgrade to squeeze it hangs again.


Top
   
PostPosted: Tue Nov 10, 2009 9:52 pm 
Offline
Newbie

Joined: Tue Nov 10, 2009 4:58 am
Posts: 10
Location: San Jose, CA
A few days ago I attempted to use the webinstaller to upgrade my KuroBox HG. It chose 2.6.30.1 and all appeared well until the reboot where it launched the old 2.4.17_mvl21 kernel.

The environment (no doubt responsible for the confusion) is one that I compiled almost 3 years ago based on 2.6.18.2. This seems to have convinced the webinstaller that I had one of the variants that used a native 2.6 kernel and it installed a version which seems to have dependencies on GLIBC 2.3 (the insmod command and its associates in particular).

I was able to force a return to my 2.6.18.2 kernel using an insmod.old that I found.

The only variant active was variant-headers, which I need to support the various added items.

I also note that the kurobox.dtb link in /boot points to kuroboxHD.dtb rather than the kuroboxHG.dtb that I would have expected.

I also note that the md5sum tool has been made dependent on GLIBC 2.3 (without an .old version). I am now rebuilding that utility so I can user the webinstaller again.

So, my question is, should I attempt an upgrade from the 2.4.17 kernel to avoid confusing the webinstaller or is there some variant or manual setting that might help guide it to the correct configuration even while booted from 2.6.18.2 (which would be much more convenient since the Kuro serves as my firewall)?


Top
   
PostPosted: Wed Nov 11, 2009 4:51 pm 
Offline
Site Admin
User avatar

Joined: Sun Jul 17, 2005 4:34 pm
Posts: 5332
Short answer: variant-legacy (nothing else), plus "echo 32768 > /proc/fs/nfsd/max_block_size"

I'm very unhappy with the current situation. While 2.6.31.6 runs fine here on uboot on both my HG and LS1, I'm unable to test no-uboot installations, so currently they don't work. Would you like to test the non-uboot kernel for me?

Thanks for the hint with the .dtb link, please correct it manually for now!

BTW the utilities were taken from Debian etch.


Top
   
PostPosted: Thu Nov 12, 2009 11:57 am 
Offline
Newbie

Joined: Tue Nov 10, 2009 4:58 am
Posts: 10
Location: San Jose, CA
Thanks for the pointers.

I spent a part of the day trying to build the 2.3.2 GLIBC that appears in the GPL source of this sites download section, with mixed success.

I would be interested in testing a new kernel, although I would admit an interest in installing UBoot. I have been tripped up in that direction by uncertainty about what is current and dead links in the wiki to the kurobox.com web site.

What would be involved in testing the kernel beyond installing and hoping?

Thanks again for your help.


Top
   
PostPosted: Thu Nov 12, 2009 9:21 pm 
Offline
Newbie

Joined: Tue Nov 10, 2009 4:58 am
Posts: 10
Location: San Jose, CA
Andre,
I did a dry run with the variant-legacy and got what appears to be a plausible /tmp/kernel-inst directory including a config file for kernel 2.6.25.20. Prior to the dry run, I deleted the directory, so hopefully it represents a clean start.

I am concerned, however that the insmod.ppc (using ldd) still reports
Code:
/tmp/kernel-inst/boot/insmod.ppc: /lib/libc.so.6: version `GLIBC_2.3' not found (required by /tmp/kernel-inst/boot/insmod.ppc)


On further investigation, I found that running the command:
Code:
find . -perm /111 -type f | xargs ldd

from '/tmp/kernel-inst' showed that the following executables had failed dependencies on GLIBC_2.3:
./usr/src/linux/scripts/basic/fixdep
./usr/src/linux/scripts/basic/docproc
./usr/src/linux/scripts/kconfig/conf (also other issues requiring GLIBC_2.3.4 and a missing libncursesw.so.5)
./usr/src/linux/scripts/kconfig/mconf (also other issues requiring GLIBC_2.3.4 and a missing libncursesw.so.5)
./usr/src/linux/scripts/conmakehash
./usr/src/linux/scripts/kallsyms
./usr/src/linux/scripts/mod/modpost
./usr/bin/diff
./usr/bin/md5sum
./sbin/genksyms
./sbin/depmod.modutils
./sbin/insmod.modutils
./sbin/modinfo.modutils
./boot/insmod.ppc
./bin/lsmod

Is there some webinstaller configuration that might still be compatible with earlier glibc's?

Alternatively, is there some pre-compiled version of a later glibc that I might install? Might the glibc from Debian Etch work?

I can probably replace most of these post-install with the earlier versions already installed. Would that be likely to work?

Thanks for your efforts.

John


Top
   
PostPosted: Sun Nov 15, 2009 1:23 pm 
Offline
Site Admin
User avatar

Joined: Sun Jul 17, 2005 4:34 pm
Posts: 5332
What's your OS again?

insmod & Co. were compiled against etch's libc6 (current version: 2.3.6.ds1-13etch9). The most important component is /boot/insmod.ppc. Without it, the new kernel will never be loaded. Copy your recent insmod to /boot/insmod.ppc, keeping it executable!


Top
   
PostPosted: Sun Nov 15, 2009 8:28 pm 
Offline
Newbie

Joined: Tue Nov 10, 2009 4:58 am
Posts: 10
Location: San Jose, CA
I am running a 2.6.18.2 kernel compiled from source and installed on top of the original software load. The kernel active after a failed 2.6 boot is the original 2.4.17_mvl21.

Over the years, I have recompiled and upgraded many components of the software load to support the primary firewall/router function that the KuroBox HG serves.

One exception to that is the GLIBC, which I have become aware is difficult both to get a clean compile and then very tricky to install because the tools involved in the install tend to depend upon the currently installed version.

I note that when module-init-tools is built, it produces a component 'insmod.static' which, as the name implies, is statically linked. This, of course, avoids any dependency on particular GLIBC version and would seem to be a preferable choice to the dynamically linked version being used for /boot/insmod.ppc.

I am currently putting together an archive that will restore usable versions of all the tools that have unresolved dependencies with GLIBC using statically linked versions that should be stable against most library changes that might occur when switching kernels and upgrading other software. The archive will be expanded into the root after running the webinstall.

I would be happy to contribute that archive, but it might not be helpful to all users and its GPL provenance (and the related source requirement) would be difficult to meet since its contents are from several projects. Perhaps a pre-webinstall script to create the archive from the already installed software would be more useful?

If the archive-restoration works properly and allows me to boot and use your kernel, I will let you know.

Thanks for your attention. I know you have little ongoing interest in the kernels for the old PPC Kuros.


Top
   
PostPosted: Mon Nov 16, 2009 4:44 pm 
Offline
Site Admin
User avatar

Joined: Sun Jul 17, 2005 4:34 pm
Posts: 5332
A static insmod would be just perfect :) Test it (this will lead to an unclean unmounting of the FS):

/path/to/insmod.static -f /boot/loaderHG.o kernel=/boot/vmlinux.bin

I'm still a happy PPC user myself. Unfortunately, I had to upgrade my own LS1 to uboot, so I've got two uboot and no stock boxes left to test. As I still compile the latest kernels for myself on a regular basis, I could distribute them as well -- if I managed to get the configuration right. My latest attempt was a failure as you know, and I depend on you guys. It's hard to the right kind of people to test such things, and I'm very grateful for any help.


Top
   
PostPosted: Mon Nov 16, 2009 8:25 pm 
Offline
Newbie

Joined: Tue Nov 10, 2009 4:58 am
Posts: 10
Location: San Jose, CA
It turns out to be somewhat more difficult than it appeared.

Static is the right way for insmod and similar tools. It is not sufficient however.

Insmod has apparently gone through at least two major revisions. The current 2.6 one is compatible back to some version of the kernel about ~2.5.48. Under 2.4 kernels, the module loading interface to the kernel was different that the modutils package was used.

To address compatibility, the modutils tools were renamed (on some distributions) to insmod.modutils, depmod.modutils, etc. and the module-init-tools were modified to detect the kernel version and run the older module if necessary.

It appears that this code tries both insmod.old and insmod.modutils in an attempt to find a version compatible with the current kernel.

You are probably aware of all this, since you included insmod.modutils in your WebInstall.

It fails, however, if any of the tried variations are not loadable (back to the GLIBC 2.3 issue).

I rebuilt all the components and fiddled the setup in /boot (the loaderHG.o is for a different 2.4 kernel than I have), using instead the loader.o that I have loading my current 2.6 kernel.

I also created a simple script that archives the current versions of all the items in the current WebInstall that are GLIBC_2.3 dependent. This is of little use if you run the WebInstall first, but it makes it easy to restore afterward if you know you might have a problem (as I do now.)

Since the selection of the box FLAVOR is made based on the kernel version, it may be that the KO loader initially selected might have worked.

In any case, I got the script setup so that it appeared it might work. The boot appeared to go normally according to the panel lights, but the device is inaccessible via the net (it was unresponsive on any of the addresses known to me). It responds to holding the front panel button with what seems to be an orderly shutdown. When it restarts, it eventually launches my 2.6.18.2 kernel (as it should when it finds a try_new_kernel.running file.)

I have moved the kuro box into my office from the closet, set the network switch to mirror the port where the kuro is connected, have WireShark running to watch it and have the serial port connected so I can see the early console output.

I hope to know more about the current state soon.

Thanks for the encouragement.


Top
   
PostPosted: Tue Nov 17, 2009 6:31 am 
Offline
Site Admin
User avatar

Joined: Sun Jul 17, 2005 4:34 pm
Posts: 5332
loader??.o is chosen dynamically according to /etc/default/boot_new.sh_local*. Please bear with me, it's been a while I looked at this code!

What you could always do is keep a local libc6 for compatibility -- a whole lot of files. A cleaner way would be a second Linux+distribution on /dev/hda3 with an adjusted fstab. loader.o can parameters to theoretically accomplish that. One big advantage is you could address /dev/hda3 as /dev/sda3 in this fstab, and use recent kernels without the hassle.


Top
   
PostPosted: Tue Nov 17, 2009 12:07 pm 
Offline
Newbie

Joined: Tue Nov 10, 2009 4:58 am
Posts: 10
Location: San Jose, CA
andre wrote:
loader??.o is chosen dynamically according to /etc/default/boot_new.sh_local*. Please bear with me, it's been a while I looked at this code!

What seems to happen in the boot_new is that a 'FLAVOR' is determined from the current uname value. In my case, that leads to a flavor of "KO" (expected "HG"). That value is then appended to both the loader name and a kernel name.

In the end, it may not matter that the suffix is not the expected "HG" since the same kernel is used in either case and (untested) the loaderKO is probably the right one if the only reason it is used is to match the kernel (not the hardware characteristics of the box).

Because of the way the kernel name test code is ordered, the "KO" variant wins even if the kernel name for the HG is fixed (at least for my case) in the /boot_new.sh_local. Again, it probably is not significant, on further reflection.


Top
   
PostPosted: Tue Nov 17, 2009 12:22 pm 
Offline
Newbie

Joined: Tue Nov 10, 2009 4:58 am
Posts: 10
Location: San Jose, CA
The serial console output from booting the 2.6.25.20 (legacy) kernel is below. It seems to be working up to the point of console handover. After that, the process continues until the 5 minute timer expires and the red light begins blinking.

It is not clear whether it would have ever finished or not. As it happens, the dns server load and a some other things take quite a long time (or can under some circumstances), so if there were much additional delay, it might not get to the ppc_uartd in time.

If you will, please look over the output for anything wrong that I might have missed.

Do you know if there is any way to delay this console handover (or attach to the 'real' one) so that I can see what else is going on?

Thanks.

Code:
******* Product Information *******
----------------------------------
Product Name: KURO-BOX/HG(IESHIGE)
          VER: 1.00
         Date: 2004/10/19 17:18:54
----------------------------------
Firmware check:done.

>>root=/dev/hda1
Now Loading...done.
Now Booting
Memory BAT mapping: BAT2=128Mb, BAT3=0Mb, residual: 0Mb
Linux version 2.4.17_mvl21 (root@toda_dev.melcoinc.co.jp) (gcc version 2.95.3 20010315 (release/MontaVista)) #24 2004G/ 107n 19F| 2PMKF| 17:17:03 JST
KURO-BOX (C) 2004 KUROUTO-SHIKOU.
On node 0 totalpages: 32768
zone(0): 32768 pages.
zone(1): 0 pages.
zone(2): 0 pages.
Kernel command line: root=/dev/hda1
OpenPIC Version 1.2 (1 CPUs and 139 IRQ sources) at 80040000
decrementer frequency = 32.446624 MHz
Calibrating delay loop... 172.85 BogoMIPS
Memory: 124760k available (1244k kernel code, 556k data, 188k init, 0k highmem)
Dentry-cache hash table entries: 16384 (order: 5, 131072 bytes)
Inode-cache hash table entries: 8192 (order: 4, 65536 bytes)
Mount-cache hash table entries: 2048 (order: 2, 16384 bytes)
Buffer-cache hash table entries: 8192 (order: 3, 32768 bytes)
Page-cache hash table entries: 32768 (order: 5, 131072 bytes)
POSIX conformance testing by UNIFIX
PCI: Probing PCI hardware
Linux NET4.0 for Linux 2.4
Based upon Swansea University Computer Society NET3.039
Initializing RT netlink socket
Starting kswapd
Disabling the Out Of Memory Killer
Journalled Block Device driver loaded
pty: 256 Unix98 ptys configured
MELCO INC. RTC driver ver 1.00
Serial driver version 5.05c (2001-07-08) with MANY_PORTS SHARE_IRQ SERIAL_PCI enabled
ttyS00 at 0x80004600 (irq = 138) is a 16550A
ttyS01 at 0x80004500 (irq = 137) is a 16550A
block: 128 slots per queue, batch=32
RAMDISK driver initialized: 16 RAM disks of 10000K size 1024 blocksize
Uniform Multi-Platform E-IDE driver Revision: 6.31
ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
CMD680: IDE controller on PCI bus 00 dev 60
CMD680: chipset revision 2
CMD680: 100% native mode on irq 17
    ide0: BM-DMA at 0xbffed0-0xbffed7, BIOS settings: hda:pio, hdb:pio
    ide1: BM-DMA at 0xbffed8-0xbffedf, BIOS settings: hdc:pio, hdd:pio
hda: ST3400632A, ATA DISK drive
ide0 at 0xbffef8-0xbffeff,0xbffef6 on irq 17
hda: 781422768 sectors (400088 MB) w/16384KiB Cache, CHS=48641/255/63, UDMA(100)
Partition check:
 hda: hda1 hda2 hda3
FLASHDISK:Initialized [FUJITSU MBM29PL32TM]
eth0: RTL8169s/8110s Gigabit Ethernet driver supports Jambo Frame 1.8n4b <2004-3-16> at 0xc9000f00, 00:0d:0b:69:26:de, IRQ 16
eth0: Auto-negotiation Enabled.
eth0: 1000Mbps Full-duplex operation.
SCSI subsystem driver Revision: 1.00
request_module[scsi_hostadapter]: Root fs not mounted
usb.c: registered new driver usbdevfs
usb.c: registered new driver hub
hcd.c: ehci-hcd @ 00:0e.2, PCI device 1033:00e0 (NEC Corporation)
hcd.c: irq 19, pci mem c9002f00
usb.c: new USB bus registered, assigned bus number 1
hcd/ehci-hcd.c: USB 2.0 support enabled, EHCI rev 1. 0
hub.c: USB hub found
hub.c: 5 ports detected
usb-ohci.c: USB OHCI at membase 0xc9004000, IRQ 19
usb-ohci.c: usb-00:0e.0, NEC Corporation USB
usb.c: new USB bus registered, assigned bus number 2
hub.c: USB hub found
hub.c: 3 ports detected
usb-ohci.c: USB OHCI at membase 0xc9006000, IRQ 19
usb-ohci.c: usb-00:0e.1, NEC Corporation USB (#2)
usb.c: new USB bus registered, assigned bus number 3
hub.c: USB hub found
hub.c: 2 ports detected
usb.c: registered new driver usblp
printer.c: v0.11: USB Printer Device Class driver
Initializing USB Mass Storage driver...
usb.c: registered new driver usb-storage
USB Mass Storage support registered.
NET4: Linux TCP/IP 1.0 for NET4.0
IP Protocols: ICMP, UDP, TCP, IGMP
IP: routing cache hash table of 1024 buckets, 8Kbytes
TCP: Hash tables configured (established 8192 bind 8192)
NET4: Unix domain sockets 1.0/SMP for Linux NET4.0.
NET4: AppleTalk 0.18a for Linux NET4.0
RAMDISK: Compressed image found at block 0
Freeing initrd memory: 2137k freed
fff70000:4f4b4f4b
EXT2-fs warning: checktime reached, running e2fsck is recommended
VFS: Mounted root (ext2 filesystem).
fff70000:4f4b4f4b
kjournald starting.  Commit interval 5 seconds
EXT3-fs: mounted filesystem with ordered data mode.
VFS: Mounted root (ext3 filesystem) readonly.
change_root: old root has d_count=2
Trying to unmount old root ... okay
Freeing unused kernel memory: 188k init
INIT: version 2.78 booting
EXT3-fs warning: maximal mount count reached, running e2fsck is recommended
EXT3 FS 2.4-0.9.17, 10 Jan 2002 on ide0(3,1), internal journal
Warning: loading /boot/loaderHG.loader module loaded
o will taint thekernel=/boot/vmlinux.bin
 kernel: forced cmdline=root=/dev/hda1
load
  See http://www.tux.org/lkml/#export-tainted for information about tainted modules
KERNELBASE=c0000000 (r6=8937ac r7=8937ba) r3=0 (r4=0 r5=0)
---cmdline=root=/dev/hda1
Linux version 2.6.25.20-svn365-dirty493 (root@ls) (gcc version 4.1.2 20061115 (prerelease) (Debian 4.1.1-21)) #27 Tue Nov 11 11:03:15 CET 2008
Early serial console at MMIO 0x80004600 (options '57600n8')
console [uart0] enabled
BUFFALO Network Attached Storage Series
(C) 2002-2005 BUFFALO INC.
Zone PFN ranges:
  DMA             0 ->    32768
  Normal      32768 ->    32768
Movable zone start PFN for each node
early_node_map[1] active PFN ranges
    0:        0 ->    32768
Built 1 zonelists in Zone order, mobility grouping on.  Total pages: 32512
Kernel command line: root=/dev/hda1
OpenPIC Version 1.2 (1 CPUs and 11 IRQ sources) at 80040000
PID hash table entries: 512 (order: 9, 2048 bytes)
Bus clock freq: 131072000
decrementer frequency = 32.521743 MHz
Console: colour dummy device 80x25
console handover: boot [uart0] -> real [tty0]


Top
   
PostPosted: Tue Nov 17, 2009 6:05 pm 
Offline
Site Admin
User avatar

Joined: Sun Jul 17, 2005 4:34 pm
Posts: 5332
loaderKO.o is the one for you, the fact you specified 'HG' elsewhere mislead me.

avr_evtd might be set to the wrong device; set either /dev/ttyS0 or /dev/ttyS1 in /etc/default/avr_evtd manually.

Please try and load variant-latest the way you loaded variant-legacy. I put some effort in compiling it for /dev/hda and am curious what came out of it!


Top
   
PostPosted: Tue Nov 17, 2009 6:59 pm 
Offline
Newbie

Joined: Tue Nov 10, 2009 4:58 am
Posts: 10
Location: San Jose, CA
andre wrote:
loaderKO.o is the one for you, the fact you specified 'HG' elsewhere mislead me.


The device is an HG (as the product header indicates), I assumed that the flavor code needed to reflect that. Instead it seems to be just an expression of the ROM kernel version.

andre wrote:
avr_evtd might be set to the wrong device; set either /dev/ttyS0 or /dev/ttyS1 in /etc/default/avr_evtd manually.


If the device name might change under a different kernel, I probably broke that. I was trying to figure out which code was switching the console output rate to 9600 baud after entering Run level 2 (on my 2.6.18 kernel). I set it to the value the scan produced under my current 2.6.18 kernel, which might somehow be different than under yours (that is probably why the scan mechanism exists). I will go back to scan mode.

andre wrote:
Please try and load variant-latest the way you loaded variant-legacy. I put some effort in compiling it for /dev/hda and am curious what came out of it!


I will do that.

I will also conditionalize the loading of 'named' on the /boot/try_new_kernel.running to eliminate a delay that may allow the deadman timer to trigger.

Thanks, again.


Top
   
PostPosted: Wed Nov 18, 2009 5:58 am 
Offline
Newbie

Joined: Tue Nov 10, 2009 4:58 am
Posts: 10
Location: San Jose, CA
I followed your advice to use the loaderKO.o and was reminded of a change I did not discuss. I have been using the loader.o that successfully loads my 2.6.18 kernel to attempt to load your kernels. I did so because of the following issues:
The loaderKO.o is compiled for a similar but not identical kernel (perhaps of no consequence). It also does not operate correctly, producing the error:
Code:
ompiled for kernel version 2.4.17_kuro-box
        while this kernel is version 2.4.17_mvl21
Warning: loading /boot/loaderKO.o will taint the kernel: forced load
  See http://www.tux.org/lkml/#export-tainted for information about tainted modules
/boot/loaderKO.o: init_module: Cannot allocate memory

You may recognize this immediately, but I have speculation if you do not. First, however, I was not successful in locating the source code for the various loader??.o modules you provide. So, I can only speculate.
I note from other sources (http://buffalo.nas-central.org/wiki/Kurobox_support_in_stock_2.6_kernels) that 2.6.20 and subsequent kernels require a different loader convention to pass the .dtb file that you provide with your kernels. Without that data the kernel will not run correctly.
I assume that you have added the necessary functionality to the loaders you provide.
Later in that article (at http://buffalo.nas-central.org/wiki/Kurobox_support_in_stock_2.6_kernels#Figure_out_your_uloader.o) the author relates his experience with the two uloader variants. The nominal HG loader assumes a 2.4.20 kernel (not true in his or my case) while the 2.4.17 uloader assumes a load address appropriate for the Kuro but not the HG.
The author resolved his problem by adding a parameter on the command line to the loader which specified the load address explicitly and used the 2.4.17 version. He thus changed the location from the 0x03f00000 found in the 2.4.17 code to the value 0x07f00000 found in the 2.4.20 version.
I used the strings utility to examine your loaders, but did not find a similar option.

So, some questions:
Do these loaders pass the .dtb file to the kernel as seems to be required?
If so, is the address at which they are loaded in the loaderKO.o module appropriate for the KuroBox HG hardware that I have? Is there a command line argument that can be passed to correct it if not?

I must say that UBoot approach is looking pretty attractive right now. :)

The whole boot attempt is shown below:
Code:
******* Product Information *******
----------------------------------
Product Name: KURO-BOX/HG(IESHIGE)
          VER: 1.00
         Date: 2004/10/19 17:18:54
----------------------------------
Firmware check:done.

>>root=/dev/hda1
Now Loading...done.
Now Booting
Memory BAT mapping: BAT2=128Mb, BAT3=0Mb, residual: 0Mb
Linux version 2.4.17_mvl21 (root@toda_dev.melcoinc.co.jp) (gcc version 2.95.3 20010315 (release/MontaVista)) #24 2004G/ 107n 19F| 2PMKF| 17:17:03 JST
KURO-BOX (C) 2004 KUROUTO-SHIKOU.
On node 0 totalpages: 32768
zone(0): 32768 pages.
zone(1): 0 pages.
zone(2): 0 pages.
Kernel command line: root=/dev/hda1
OpenPIC Version 1.2 (1 CPUs and 139 IRQ sources) at 80040000
decrementer frequency = 32.462339 MHz
Calibrating delay loop... 172.85 BogoMIPS
Memory: 124760k available (1244k kernel code, 556k data, 188k init, 0k highmem)
Dentry-cache hash table entries: 16384 (order: 5, 131072 bytes)
Inode-cache hash table entries: 8192 (order: 4, 65536 bytes)
Mount-cache hash table entries: 2048 (order: 2, 16384 bytes)
Buffer-cache hash table entries: 8192 (order: 3, 32768 bytes)
Page-cache hash table entries: 32768 (order: 5, 131072 bytes)
POSIX conformance testing by UNIFIX
PCI: Probing PCI hardware
Linux NET4.0 for Linux 2.4
Based upon Swansea University Computer Society NET3.039
Initializing RT netlink socket
Starting kswapd
Disabling the Out Of Memory Killer
Journalled Block Device driver loaded
pty: 256 Unix98 ptys configured
MELCO INC. RTC driver ver 1.00
Serial driver version 5.05c (2001-07-08) with MANY_PORTS SHARE_IRQ SERIAL_PCI enabled
ttyS00 at 0x80004600 (irq = 138) is a 16550A
ttyS01 at 0x80004500 (irq = 137) is a 16550A
block: 128 slots per queue, batch=32
RAMDISK driver initialized: 16 RAM disks of 10000K size 1024 blocksize
Uniform Multi-Platform E-IDE driver Revision: 6.31
ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
CMD680: IDE controller on PCI bus 00 dev 60
CMD680: chipset revision 2
CMD680: 100% native mode on irq 17
    ide0: BM-DMA at 0xbffed0-0xbffed7, BIOS settings: hda:pio, hdb:pio
    ide1: BM-DMA at 0xbffed8-0xbffedf, BIOS settings: hdc:pio, hdd:pio
hda: ST3400632A, ATA DISK drive
ide0 at 0xbffef8-0xbffeff,0xbffef6 on irq 17
hda: 781422768 sectors (400088 MB) w/16384KiB Cache, CHS=48641/255/63, UDMA(100)
Partition check:
 hda: hda1 hda2 hda3
FLASHDISK:Initialized [FUJITSU MBM29PL32TM]
eth0: RTL8169s/8110s Gigabit Ethernet driver supports Jambo Frame 1.8n4b <2004-3-16> at 0xc9000f00, 00:0d:0b:69:26:de, IRQ 16
eth0: Auto-negotiation Enabled.
eth0: 1000Mbps Full-duplex operation.
SCSI subsystem driver Revision: 1.00
request_module[scsi_hostadapter]: Root fs not mounted
usb.c: registered new driver usbdevfs
usb.c: registered new driver hub
hcd.c: ehci-hcd @ 00:0e.2, PCI device 1033:00e0 (NEC Corporation)
hcd.c: irq 19, pci mem c9002f00
usb.c: new USB bus registered, assigned bus number 1
hcd/ehci-hcd.c: USB 2.0 support enabled, EHCI rev 1. 0
hub.c: USB hub found
hub.c: 5 ports detected
usb-ohci.c: USB OHCI at membase 0xc9004000, IRQ 19
usb-ohci.c: usb-00:0e.0, NEC Corporation USB
usb.c: new USB bus registered, assigned bus number 2
hub.c: USB hub found
hub.c: 3 ports detected
usb-ohci.c: USB OHCI at membase 0xc9006000, IRQ 19
usb-ohci.c: usb-00:0e.1, NEC Corporation USB (#2)
usb.c: new USB bus registered, assigned bus number 3
hub.c: USB hub found
hub.c: 2 ports detected
usb.c: registered new driver usblp
printer.c: v0.11: USB Printer Device Class driver
Initializing USB Mass Storage driver...
usb.c: registered new driver usb-storage
USB Mass Storage support registered.
NET4: Linux TCP/IP 1.0 for NET4.0
IP Protocols: ICMP, UDP, TCP, IGMP
IP: routing cache hash table of 1024 buckets, 8Kbytes
TCP: Hash tables configured (established 8192 bind 8192)
NET4: Unix domain sockets 1.0/SMP for Linux NET4.0.
NET4: AppleTalk 0.18a for Linux NET4.0
RAMDISK: Compressed image found at block 0
Freeing initrd memory: 2137k freed
fff70000:4f4b4f4b
EXT2-fs warning: checktime reached, running e2fsck is recommended
VFS: Mounted root (ext2 filesystem).
fff70000:4f4b4f4b
kjournald starting.  Commit interval 5 seconds
EXT3-fs: mounted filesystem with ordered data mode.
VFS: Mounted root (ext3 filesystem) readonly.
change_root: old root has d_count=2
Trying to unmount old root ... okay
Freeing unused kernel memory: 188k init
INIT: version 2.78 booting
EXT3-fs warning: maximal mount count reached, running e2fsck is recommended
EXT3 FS 2.4-0.9.17, 10 Jan 2002 on ide0(3,1), internal journal
Warning: kernel-module version mloader module loaded
ismatch
        /boot/kernel=/boot/vmlinux.bin
loaderKO.o was ccmdline=root=/dev/hda1
ompiled for kernel version 2.4.17_kuro-box
        while this kernel is version 2.4.17_mvl21
Warning: loading /boot/loaderKO.o will taint the kernel: forced load
  See http://www.tux.org/lkml/#export-tainted for information about tainted modules
/boot/loaderKO.o: init_module: Cannot allocate memory
Hint: insmod errors can be caused by incorrect module parameters, including invalid IO or IRQ parameters.
      You may find more information in syslog or the output from dmesg
/etc/rc.d/rcS.d/S00boot_new.sh: line 72: /boot/boot.log: Read-only file system
Warning: kernel-module version mloader module loaded
ismatch
        /boot/kernel=/boot/vmlinux.bin
loaderKO.o was ccmdline=root=/dev/hda1
ompiled for kernel version 2.4.17_kuro-box
        while this kernel is version 2.4.17_mvl21
Warning: loading /boot/loaderKO.o will taint the kernel: forced load
  See http://www.tux.org/lkml/#export-tainted for information about tainted modules
/boot/loaderKO.o: init_module: Cannot allocate memory
Hint: insmod errors can be caused by incorrect module parameters, including invalid IO or IRQ parameters.
      You may find more information in syslog or the output from dmesg
/etc/rc.d/rcS.d/S00boot_new.sh: line 74: /boot/boot.log: Read-only file system
loading kernel 2.6
loader module loaded
kernel=/boot/vmlinux-2.6.18.2_kurobox_hg
cmdline=root=/dev/hda1
[    0.000000]
[    0.000000]
[    0.000000] KERNELBASE=c0000000 (r6=8a07ac r7=8a07ba) r3=0 (r4=0 r5=0)
[    0.000000] Linux version 2.6.18.2-kuroboxHG (root@kuro.naksuji.com) (gcc version 3.4.4) #4 Sun Dec 17 04:59:37 PST 2006
[    0.000000] gen550_init
[    0.000000] Early serial console at MMIO 0x80004600 (options '57600n8')
[    0.000000] BUFFALO Network Attached Storage Series
[    0.000000] (C) 2002-2005 BUFFALO INC.
arch: exit
[    0.000000] Built 1 zonelists.  Total pages: 32768
... (boots 2.6.18.2 successfully)


Top
   
Display posts from previous:  Sort by  
Post new topic  Reply to topic  [ 43 posts ]  Go to page Previous 1 2 3 Next

All times are UTC+01:00


Who is online

Users browsing this forum: No registered users and 3 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Powered by phpBB® Forum Software © phpBB Limited