Thanks again! Actually I asked about the size of the whole blob you mentioned to get some idea how hard would it be to disassemble it fully. What do you think?
It's a bootcode.bin which runs a simple shell via a serial terminal which allows you to do things like up and download data, examine memory, etc. No ARM code involved anywhere! It doesn't initialise the SDRAM yet (although I think I know how).
Right now it's built using a hacky port of the Amsterdam Compiler Kit, which generates lousy code. I've tried porting both gcc and LLVM and ran into a brick wall --- they're both vile. I'm hopefully going to have a go with libfirm soon, which looks more tractable. However, Volker Barthelmann's non-open-source vbcc has experimental (and apparently very good) VC4 support: http://www.compilers.de/vbcc.html
Where did you/they get the information on the VC4 instruction set from? According to some sources about reverse-engineering VC4 which I could recover if I wanted to, finding out the instruction set that the blob uses was a serious problem at least in the past.
I believe it was figured out from a combination of analysis of the blob, examining the patents (which contain quite a lot of information), trial-and-error, and careful examination of the fragments of VC4 source released by Broadcom (their big source release a year or so back contained quite a lot). I know that at least one person wrote a program which would run arbitrary instructions on the Pi and analyse the state of the registers afterwards, which is a neat trick.
The VC4 is a really nice processor, BTW. Dual core, lots of registers, efficient instruction packing, 64x64x8bit vector unit, integrated single-precision FPU using the integer registers (no double-precision, alas), 1kB on-chip lookup table... but, bizarrely, no adc or sbc instructions, so 64-bit arithmetic is hard. Very weird.
