Re: FPGA "asm", yeah that probably was the wrong word. If you will permit the analogy for me ASM is learning an set of opcodes and their uses, FPGA coding (and to a lesser extent the other programmable chips) is making your own opcodes.
While I have not got much in the way of FPGA dev experience I find the best way (aside from a good bit of knowledge of low level electronics:
http://ocw.mit.edu/OcwWeb/Electrical-Engin...ience/index.htm and
http://webster.cs.ucr.edu/AsmTools/HLA/ I know it "normal" ASM but it does have some stuff if you needed links) is somewhere between thought exercise (this is problem (which will probably become apparent from the "holes" in the disassembly you have) and how would I pull it off) and emulation of components (with a side of brute force).
If I had to guess they probably use it as a kind of hardware interface (USB port, LEDs etc: most other makers use something a bit more suitable like a CPLD) and had maybe intended it as a kind of co-processor (probably more graphics card than math coprocessor although they could have kicked MP3 support to it).
Actel if I recall correctly (it has been about a year and a half since I laid hands on it) also have a reasonable (as in fairly featured) dev kit/cpmiler if you can lay your hands on it, assuming they used it you can even go some way to decompiling the code straight up.
Also on the subject of DLDI did the DSX ever get a good one (unaligned read, write, good speed etc)? I remember reporting on one ages ago but never followed up.
As for the updater: emulation of the PC/windows is the key here for me (I did a few rudimentary handshake protocols a few months back) as that is probably the best way to get full dumps of the code. You may have to go a bit back into the past as a lot of things have full hardware access now or an interpreter that can feed stuff straight into it.
