how can i update to kernal 2.6.33-X-Kirkwood

[realfenglin]

HI guys

     i have used  manually install a tar ball to install debian into my sheevaplug.(http://www.cyrius.com/debian/kirkwood/sheevaplug/unpack.html) i do not know why i cannot use Debian installer to install (http://www.cyrius.com/debian/kirkwood/sheevaplug/install.html). it is always fails at this end when some software are installed.

    right now, i am only using 2.6.30-x-kirkwood Kernel. but i found it does not include some driver i need. then i try to update the 2.6.30 kernal to 2.6.33.

     i used the following way:
wget http://sheeva.with-linux.com/sheeva/README-2.6.33
chmod 777 README-2.6.33
./README-2.6.33

then it automatically downloading some file and at the end tell me to do a new boot image.

then i do the following setting:

setenv bootcmd_mmc 'mmcinit; ext2load mmc 0:1 0x01100000 /uInitrd; ext2load mmc 0:1 0x00800000 /sheeva-2.6.33-uImage'

but when i boot the sheevaplug. there are some error happen:

FATAL: Could not load /lib/modules/2.6.33/modules.dep: No such file or directory

what should i do to solve this problem?

thank you

[realfenglin]

some information about my settings:

debian:~# uname -a
Linux debian 2.6.33 #2 PREEMPT Tue Mar 2 14:27:17 MST 2010 armv5tel GNU/Linux


debian:/boot# ls
config-2.6.30-2-kirkwood      System.map-2.6.30-2-kirkwood
config-2.6.32-5-kirkwood      System.map-2.6.32-5-kirkwood
initrd.img                    uImage
initrd.img-2.6.30-2-kirkwood  uInitrd
initrd.img-2.6.32-5-kirkwood  vmlinuz
initrd.img.old                vmlinuz-2.6.30-2-kirkwood
lost+found                    vmlinuz-2.6.32-5-kirkwood
sheeva-2.6.33-uImage          vmlinuz.old


bootcmd:
setenv bootcmd_mmc 'mmcinit; ext2load mmc 0:1 0x01100000 /uInitrd; ext2load mmc 0:1 0x00800000 /sheeva-2.6.33-uImage'


error happen:

Begin: Loading essential drivers ... FATAL: Could not load /lib/modules/2.6.33/m
odules.dep: No such file or directory
FATAL: Could not load /lib/modules/2.6.33/modules.dep: No such file or directory

FATAL: Could not load /lib/modules/2.6.33/modules.dep: No such file or directory

FATAL: Could not load /lib/modules/2.6.33/modules.dep: No such file or directory

done.
Begin: Running /scripts/init-premount ... FATAL: Could not load /lib/modules/2.6
.33/modules.dep: No such file or directory
done.
Begin: Mounting root file system ... Begin: Running /scripts/local-top ... Begin
: Loadingmodprobe: FATAL: Could not load /lib/modules/2.6.33/modules.dep: No suc
h file or directory

Failure: failed to load module linear.
modprobe: FATAL: Could not load /lib/modules/2.6.33/modules.dep: No such file or
 directory

Failure: failed to load module multipath.
modprobe: FATAL: Could not load /lib/modules/2.6.33/modules.dep: No such file or
 directory

Failure: failed to load module raid0.
modprobe: FATAL: Could not load /lib/modules/2.6.33/modules.dep: No such file or
 directory

Failure: failed to load module raid1.
modprobe: FATAL: Could not load /lib/modules/2.6.33/modules.dep: No such file or
 directory

Failure: failed to load module raid456.
modprobe: FATAL: Could not load /lib/modules/2.6.33/modules.dep: No such file or
 directory

Failure: failed to load module raid5.
modprobe: FATAL: Could not load /lib/modules/2.6.33/modules.dep: No such file or
 directory

Failure: failed to load module raid6.
modprobe: FATAL: Could not load /lib/modules/2.6.33/modules.dep: No such file or
 directory

Failure: failed to load module raid10.
done.
FATAL: Could not load /lib/modules/2.6.33/modules.dep: No such file or directory

done.
Begin: Running /scripts/local-premount ... done.
FATAL: Could not load /lib/modules/2.6.33/modules.dep: No such file or director

[big_ignoramus]

Those drivers (or modules) aren't part of the kernel; they're hosted on the root file system (although they are loaded into kernel space when they're loaded). It can't find them because the directory is based off the kernel version. If, for fun, you made a symbolic link to the old directory with the new name, the modules wouldn't load because they're built against the old kernel. In short you need to build a new root file system against the new kernel. I haven't gotten that far because I haven't settled on a distribution yet. However, this wiki page might help you: http://plugcomputer.org/plugwiki/index.php/Recreating_the_GuruPlug_RFS

[chaotix]

It looks like your boot arguments are not suitable for the installed kernel. There are basically two different methods to boot a system. One is to compile every driver you need for booting into the kernel directly. This is ok for fixed hardware like the sheevaplug or if you compile the kernel for you PC at home. Then there is the possibility to just compile basically everything else as a module and generate an initrd (Init RAM disk). After the kernel boots it starts the RAM disk which is loaded by the boot loader into RAM, which needs to contain all drivers necessary. This is a good way for distributions to support nearly every hardware out there.

So, why do I tell you this? Because it seems that your old kernel was using an initrd, but the sheeva.with-linux.com one does not need one, as all drivers necessary to boot should be compiled in. But your boot arguments still load the initrd from the old kernel. You need to drop that part from your boot arguments. Just remove 'ext2load mmc 0:1 0x01100000 /uInitrd;' from bootcmd_mmc and hopefully it should be fine.

Good luck,
Lars