From: Michael S. <st...@in...> - 2004-01-22 12:38:05
|
Hi Jimmi! I'm Cc:ing this to the mailing list and I would suggest that=20= you also join it (=20 http://lists.sourceforge.net/lists/listinfo/gc-linux-devel ), okay? On 20.01.2004, at 22:08, jl...@fr... wrote: > I've just seen the new GC-Linux porting project, after looking from=20 > time to time > on the Xbox Linux website ... These projects are both hi-tech and fun,=20= > it's great ! :-) > I'm dropping these lines to you, because i think there are already=20 > avaiable code > that may help you on this project. > > The Game Cube main CPU is a custom chip (ASIC) made for Nintendo,=20 > including the > PowerPC 405 CPU and various peripherals. Besides the PPC 405 GP from=20= > IBM, > another chip includes this processor core : the Virtex II Pro from=20 > Xilinx, which > is a FPGA (you know, a programmable logic device). And there is=20 > already a Linux > port on this chip :-) > > There are even two : Mind Linux Solutions (www.mind.be) did it for=20 > standard > Linux kernel, and Montavista Linux (www.mvista.com) did it for the=20 > real-time > version of the Kernel, plus a complete distribution. Both these ports=20= > are GPL, > so you could get PPC405 code without any legal headache, and grab some=20= > boot > code, drivers for Ethernet, RTC, or X11 port, with minimal changes=20 > required. Hmmm... The documentation on the internet that I found, said "PowerPC=20 750CXe" ( http://www.anandtech.com/showdoc.html?i=3D1566&p=3D2 ), and=20 that's PPC7xx, not PPC4xx... (what's the difference anyway?) Well, Linux exists for both PPC7xx and PPC4xx, i.e. all code that is=20 CPU specific is in the standard kernel tree. And I think neither of=20 Ethernet, RTC and video of the GameCube are compatible with these=20 devices, as they are all implemented outside the CPU. :-( > Are there already shcematics, or at least block-diagrams of the GC=20 > arhitecture, > available on the web ? It may be possible to map VHDL models of the=20 > peripherals > in Virtex II Pro FPGA, to get a developpment system for GC without the > limitations of the GC (memory size, CD drive ...). Finding some JTAG > connector/test points on the board could also give access to the debug=20= > port of > the PPC405, allowing easy trace and debug, esp for the boot process. That's the way a hardware person thinks. :-) As I software person, I=20 think quite differently - I would just use the real thing as use some=20 debug code (that I know that works) to find out whether my code is=20 still alive at a certain point. It is my impression that building=20 custom development hardware might be too much work, as so much is known=20= about the GC already. > I'm sorry i can not help you much for the moment (i'm already busy=20 > trying to > hack some Linux on Xilinx chips) and i don't even have a CG (despite=20= > it's > getting rather cheap here, around 100=80). May be i could help on some=20= > french > translation ? Certainly. Anyone interested in doing translations, please join the=20 "docs" mailing list (=20 http://lists.sourceforge.net/lists/listinfo/gc-linux-docs ) and=20 announce it there. But since you seem to be a very experienced hardware=20= person, I am sure you can help in other ways - a bit? > Good luck, and have fun with you Tux-customized consoles ! > Jimmy Michael |