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

Dear all,

I have been tinkering with the problem of USB devices with all 777-permissions which seems to be the main obstacle against FFP on a stick here. The results look promising enough to me that I dare opening a new thread here.

1. It seems that the problem is due to mounting with wrong umask value (0). I am pretty sure this happens in '/usr/sbin/usbmount' which tries mounting devices in order as VFAT, AUTO, EXT3, EXT2; the former two with umask=0. (BTW, I have never seen any piece of code that looked worse than this one!) AFAIK there does not seem to be any real solution for adapting this behavior (cf. the two threads by sinisterguy and magnusheino).

2. Playing ignorant, I tried the simple path and put the following into '/ffp/etc/fun_plug.local' (don't forget to chmod +x that file):

## Test mounting USB-stick
# status encountered
mount > /mnt/HD/HD_a2/MountTest.pre
ls -l /mnt/USB/HD_c1/Test >> /mnt/HD/HD_a2/MountTest.pre
# un-mount device
umount /dev/sdc1
# re-mount it with proper umask and identical other mount options
mount -o umask=0022,codepage=850,iocharset=iso8859-1 /dev/sdc1 /mnt/USB/HD_c1
mount > /mnt/HD/HD_a2/MountTest.post
ls -l /mnt/USB/HD_c1/Test >> /mnt/HD/HD_a2/MountTest.post

The stick is formated with VFAT and the file 'Test' has been created beforehand.

Resulting 'MountTest.pre':

/dev/sdc1 on /mnt/USB/HD_c1 type vfat (rw,umask=0,iocharset=utf8,shortname=winnt)
-rwxrwxrwx    1 root     root            0 Oct 22 01:57 /mnt/USB/HD_c1/Test

Resulting 'MountTest.post':

/dev/sdc1 on /mnt/USB/HD_c1 type vfat (rw,fmask=0022,dmask=0022,codepage=850,iocharset=iso8859-1)
-rwxr-xr-x    1 root     root            0 Oct 22 01:57 /mnt/USB/HD_c1/Test

3. AFAICT the door looks wide open for installing FFP on a stick.
a) Create some setup script in 'HD_a2/.bootstrap/setup.sh' containing the above re-mounting procedure
b) Hand over the control to FFP on the USB device
c) Follow the standard (?) procedure for FFP on a stick.
The fact that the device path will be hard-coded should not be too big of an obstacle unless one keeps changing USB devices on the DNS-320 which seems to not very likely to me.

Unfortunately I have no prior experience with FFP on a stick and am fighting an approaching dead-line. I have not tested with an EXT-formated stick. So I hope someone can step in and share their expertise on this topic. Am I overlooking something obvious?

Cheers!
Stefan Müller, Switzerland