Can't unsquashfs Albis STB-7710

[Blagus]

Hi, I have a problem with extracting recovery kernel's rootfs on Albis STB-7710.
It's 640kB SquashFS image, without uImage header.
All available unsquashfs tools fail, fw-mod-kit gives me "no supported fs found", and unsquashfs gives "FATAL ERROR aborting: failed to read fragment table" error.

U-Boot env variable defines the image as squashfs, and "squashfs: version 3.3 (2007/10/31) Phillip Lougher" is started in kernel.

Here's the file, I hope someone can help me.
http://www.mediafire.com/?561vm65uzl1v58w

[Jeremy Collake]

The vendor should be required to distribute the source. In the source tarball, you will likely find a modified version of squashfs-tools. It might require a little work to make sure unsquashfs is operational with the same mods, or maybe everything is ready to go.

[Blagus]

Unfortunately, there's no squashfs tools in sources.
http://opensource.albistechnologies.com/STB/STB-7710/7710.ZAP.01.XX.0378.05.htm

All I found that could be of some use is squashfs in kernel source, but I have no idea is there anything useful.

One more note: gzip-based unsquashfs variants are giving me fragment errors, while LZMA-based ones crash and throw some debug data.

[Jeremy Collake]

Someone will have to look into it more closely. I won't lie, I have far too much of a backlog to possibly look into it myself.

[Blagus]

Well, I'd appreciate some help from someone about what should I look for in sources.

[BenYeeHua]

Well, I'd appreciate some help from someone about what should I look for in sources.
I'm having a project in mind with this device, which should be done by the end of April if I wanted to apply for the contest this year.

I am sorry that I can't help you.
Because I don't has any knowledge for this, and maybe making more trouble.

[Jeremy Collake]

Well, I'd appreciate some help from someone about what should I look for in sources.

If the squashfs-tools package isn't there, you'll need to possibly do some reverse engineering of the filesystem image. You might also check any other source tarballs from the vendor, as they are probably using the same squashfs variant. It's also possible if you keep trying variants, you will find one that works.

Best to walk the filesystem image, compare and contrast the headers, signatures, and data types to the stock variant. If all looks ok, then you go after the compressed data, compare and contrast to known compression algorithms and/or random try to decompress them with known algorithms. It's almost certain they used a common algorithm.

Sometimes obfuscation layers are added, such as a simple XOR to the post-compressed data, or entire filesystem image. Check for the squashfs kernel patch that allows for the filesystem to even be read... actually, probably should be first step, see if this exists. If it does, you can have your answer.