DSM-G600, DNS-3xx and NSA-220 Hack Forum

indy75 wrote:

I may be wrong, but it seems this firmware lacks support for USB external HDD ? At least, my USB EXT2-formatted HD is not recognized since I tried Sala's firmware.
Just to be sure...

Something is screwed with hotplug scripts. I remember that original D-Link firmware did have same problem sometimes.
You have to mount USB drives manualy using fun_plug or telnet.

And for the record it is not "my firmware" or "sala firmware". I prefer to call it custom firmware. This firmware is here because of dns323.kood.org/dsmg600 community and not just because of me.
As matter of fact beattie's contributions to this firmware is many times greater than mine or anybody else's.

qn1234 wrote:

Oh, I got it from your download directory: people/sala/fw-build-script.tar.bz2 .  I renamed it to Sala_DSM-0.4.tar.bz2 since there is Beattie_DSM-0.4.tar.bz2.

It is pretty much the same. It was meant to implement busybox 1.0 final and downloading sources from original sites.

qn1234 wrote:

Did you have to do anything else besides extracting the archive and type make in order to build your fwimage?

--
Quang

There is wrong beginning in toolchain make file (wget line) and gdb still fails, so you need to remove mk file for gdb when you get the error.

My mistake : let's call it "custom firmware". And thanks to all who contribute !
About mounting USB drives by fun_plug, could anybody tell me what is the syntax ? I'll try it...
Unfortunately, I am not good enough in Unix to guess it by myself

Many thanks for providing the custom firmware. It's a huge improvement for me. The ability to use dtach is invaluable.

I got a bit excited when I saw the uboot stuff, but I'm assuming that there is still a blocking issue with the IDE and network controllers. However, would it be possible to do a boot-from-USB-only version, or am I being completely stupid?

beattie wrote:

Unless somebody can figure out how to enable the JTAG there is not going to be a new u-boot.

Understood.

What exactly is it that you'd like to do by booting from USB that you can't do with the mechinisim I developed.

The ability to load an experimental kernel from the USB without having to fall back on the serial interface would have been nice. From what I've read on the subject so far I'm guessing you can't do that with pivot_root. However, I'm quite happy with what's been provided so far, and I'd rather contribute in some way, such as trying to work out what chkbutton's up to, than waste people's time. I've already had to look at USB core driver assembler to work out why a very basic driver I'd written was crashing, though it helped having the source code available.

Last edited by StevieP (2007-02-07 20:52:10)

beattie wrote:

As thinks sit, no, but there is a kernel patch and set of tools called kexec/kexec-tools that will let you boot a new kernel from an existing kernel.  If you can get this working, you should be able to use my root diversion and run kexec instead of piviot_root.

i'm looking for this patch (and for ppc) at this time i found only 2.6 patch. Will 2.6 kernel fit in flash ?
And why D-link stay o, 2.4.21 pre1 and not in pure 2.4.21 ?

beattie wrote:

Here is one for an earlier 2.4 kernel version http://www.cwlinux.com/downloads/kexec/ … exec.patch I have not looked any further and I don't know if it is processor specific.

Only for Alpha and i386 architectures ....

I've removed iocharset=cp850 and now all works fine!
thanks!

sala wrote:

SystemR89 wrote:

Hi!  Good work!    I've immediately upgraded my DSMG600, but now I've encountered a problem: I can't no more mount my internal FAT32 partition    

This is the error: 
# mount
%root% on / type unknown (rw)
proc on /proc type proc (rw)
/image.cfs on /sys/crfs type cramfs (rw,loop=/dev/loop1)
/dev/sda2 on /mnt/HD_a2 type ext2 (rw) 
# mount /dev/sda3  mount: wrong fs type, bad option, bad superblock on /dev/sda3, or too many mounted file systems

Obviously I've replaced the default fstab with my fstab (that I've alredy used with "old" firmwire)

Works fine in here without any extra mount option.

Code:

# mount
%root% on / type unknown (rw)
proc on /proc type proc (rw)
/image.cfs on /sys/crfs type cramfs (rw,loop=/dev/loop1)
/dev/sda2 on /mnt/HD_a2 type ext3 (rw)
/dev/sdb1 on /mnt/HD_a2/gentoo type vfat (rw)
# df -m |grep sdb1
/dev/sdb1                19084         4     19080   0% /mnt/HD_a2/gentoo
# cat /proc/filesystems
nodev   rootfs
nodev   bdev
nodev   proc
nodev   sockfs
nodev   tmpfs
nodev   pipefs
        ext3
        ext2
        cramfs
nodev   ramfs
        minix
        msdos
        vfat
        iso9660
nodev   nfs
        ntfs
        udf
nodev   autofs
#

But I can tell that mounting through hotplug/usbount will give error. usbmount uses mount parameters such iocharset=cp850 and I think that kernel does not like it. I might try to set FAT_DEFAULT_IOCHARSET to kernel config.

Code:

# mount -t vfat -o umask=0,iocharset=cp850,shortname=winnt /dev/sdb1 /mnt/HD_a2/gentoo/
mount: wrong fs type, bad option, bad superblock on /dev/sdb1,
       or too many mounted file systems
#

Now there is this possibility that we have one kernel source and D-Link has the other (more patched/modified). So our firmware and D-Link's original firmware can act little differently.
To meet GPL needs D-Link must release new tarballs with new firmwares they release but this is what D-Link does not do.

Also I would very much like to see D-Link releasing Ralink rt2500 drivers sources. I have e-mailed couple of times to support but they just ignore the main point of my emails, tell me to contact my local support or call them instead of emailing.

If I only could have more time, English skills and better knowledges about GPL I would troll over http://gpl-violations.org/
I am sure that D-Link is known case to them.

SystemR89 wrote:

How I can use dtach? I've tried, but with no success

At the end of putty.log I have a example of dtach in action.

# dtach -n /tmp/test vi
# dtach -a /tmp/test
//cutout
[EOF - dtach terminating]

Also look up man pages @ http://linuxcommand.org/man_pages/dtach1.html

mikkel_robin wrote:

Hi all, i just got my DSN-G600 and would like to know, how close are you guys on releasing a firmware that can be used without the serial connection?

Current firmware do not need serial console at all, but there is always this thing, if firmware update process fails or something else goes wrong, then you should have serial port in place, which you can use to recover your DSM-G600.

But before you go updating firmware, you should think why you want to do that, because most thing (except domain sockets) is still doable thru fun_plug

I'm planning on a 0.5 release, the main change I have is to build the kernel from source, so I can add a printk to the serial port driver to see the events (button presses etc.) comming from the mcu.  I'd like to collect up any changes that others have done or would like where they fit in with the scheme of things.