No this isn't fixed, and it won't be fixed in the Karmic release timeframe.  I anticipate mainlined LZMA support for Squashfs 4.0 will be available in kernel 2.6.33.

The underlying reason for this "bug" is because you got your squashfs-tools and squashfs kernel modules mixed up.  Squashfs has never officially supported LZMA compression, and the official Squashfs tools and kernel module releases available from http://squashfs.org has no support for LZMA.

What complicates the situation is there are unofficial (but approved by myself) third party LZMA patches available from http://squashfs-lzma.org for Squashfs 3.x.

What appears to have happened is Debian adopted the LZMA patches for their Squashfs 3.3 packages, which was inherited by Ubuntu (which is why comment #3 reports mksquashfs 3.3 switched to LZMA by default).  Debian will have patched both the squashfs-tools and the squashfs kernel code, however, while Ubuntu inherited the squashfs-tools changes, it didn't inherit the squashfs kernel code changes.  This is because the Ubuntu kernel is produced independently, and the kernel code carries the original unmodified Squashfs code from http://squashfs.org.  This has produced a mis-match which has existed up to and including Jaunty.

Having just tested Jaunty mksquashfs, even though it is clearly the LZMA patched mksquashfs (with the -lzma/-nolzma options), the default has changed at sometime to building gzipped squashfs filesystems which are supported by the Jaunty kernel.  So the mis-match isn't as apparent as it was earlier (although the -lzma option will still produce unmountable filesystems).

Squashfs 4.0 which I mainlined in kernel 2.6.29 (and so it is the version used by Karmic) complicates the situation yet further.  Squashfs 4.0 doesn't support LZMA, but there are no third-party LZMA patches available from http://squashfs-lzma.org.  The site states "There is no longer any interest in developing separate patches against squashfs 4.0" as Squashfs has been mainlined.  I can surmise the reason for this is first because I substantially rewrote the Squashfs 4.0 kernel code for mainlining, and so their LZMA patches won't work without a substantial rewrite, and second because the LZMA patches should preferably be mainlined themselves, and their LZMA patches aren't suitable for mainlining.

The upshot of this in the short term for Karmic is that your squashfs-tools/squashfs-kernel mismatch has gone away, because both the squashfs 4.0 tools and the kernel code does not support LZMA.  However, the downside is you're not going to have LZMA support in Karmic.

In the longer term people are asking me (thanks guys) where LZMA support for Squashfs 4.0 has gone.  Because Squashfs has been mainlined, and because there has also been a (poor) implementation of LZMA mainlined, my original reasons for not supporting LZMA in Squashfs no longer applies.  I'm therefore currently working on getting Squashfs LZMA support mainlined.  Because the kernel LZMA implementation requires substantial modifications to work efficiently with Squashfs, I don't anticipate the patches entering mainline before 2.6.33, although out of tree patches may be available before then.