Is there any tut? I would be vey interessed by testing that! :p

Which calculator?
For CX I wrote some kind of tutorial in this thread.
If you have an older (Click-/Touchpad) nspire, you have to clone, config and build a kernel yourself.

BTW: Tangrs, you double-posted :-P
And my name is not "vloginator" xD

I'll send you a new patch (again, last time mail didn't work..) later.

Oh my apologies on both counts - I was a little too excited

What are the differences between the Touchpad, the Clickpad and the CX?

On the one side, Clickpad and Touchpad calculators, which are pretty much similar, save for the keyboard and its handling.
On the other side, the CX, with more RAM, more Flash, color screen, a newer USB controller, different interrupt controllers, etc. The CM-C is, in principle, a watered-down CX (half the amount of RAM) with a chinese keyboard, but little else different, AFAWCT.
You can find more detail about the hardware differences on Hackspire (e.g. memory-mapped I/O page)

Are there also major differences between these and their CAS models?

Nope. The non-CAS Nspires can run the CAS OS just fine (well, with some persistence: on some calculators, it can be necessary to make dozens of attempts to yield a success, while it succeeds pretty quickly on others), as showed by the unreleased RunOS and the released OSLauncher.

Yeah, by default, the extremely crappy screen of the Clickpad & Touchpad is used for 4 bpp screen, though the driver supports 8 bpp (used by e.g. nDOOM, and some nSDL-based games). I don't remember whether Clickpad/Touchpad and CX have the same display driver.

I just saw the linux kernel option "Kernel Execute In-Place from ROM".
At the moment linuxloader2 copies the complete kernel image to ram.
Wouldn't it be possible to use that option (more free ram)?

Nope, it wouldn't be possible: XIP requires random access, but NAND Flash can only be read page by page. NOR Flash would do the job: it's precisely used for XIP in TI-Z80 & TI-68k calcs (and also for the boot1 in Nspire Clickpad / Touchpad calcs).
Even if it were possible, execution from Flash would be slower than execution from the main RAM, due to narrower (at the very least, not wider) data bus.

Yeah, looks like someone is having fun downvoting posts, no matter they contain technical content relevant to the discussion

Sorting out the nspire-related patch needles in the Linux kernel patch haystack is possible through directory filtering, arch/arm/mach-nspire being the most interesting one.

Congratz

Just a little question, maybe it will sound silly to those who know ^^ :

The log says :
"Processor       : ARM926EJ-S rev 4 (v5l)
BogoMIPS        : 59.39"

Looking that proc reference I found this : http://pastebin.com/dFSsVaBe, which reads a BogoMIPS of 103.68.
How can it be different if it'as actually the exact same processor ?

Congratz

Just a little question, maybe it will sound silly to those who know ^^ :

The log says :
"Processor       : ARM926EJ-S rev 4 (v5l)
BogoMIPS        : 59.39"

Looking that proc reference I found this : http://pastebin.com/dFSsVaBe, which reads a BogoMIPS of 103.68.
How can it be different if it'as actually the exact same processor ?

Different clock speeds. You can see it will increase if you overclock with Nover before booting the kernel.

Is there any tut? I would be vey interessed by testing that! :p

I'm planning to set up nightly builds sometime so people can test the latest and greatest (and to see if I've accidentally broken something).

In other news, LCD support on Touchpads is now working