I have not used PS2 Linux but I am curious how was it possible to limit the low level graphics API? Did they not let you run VU1 code? The retail games were just writing registers on GS from VU1 (and few smart ones were also streaming textures directly to GS). There was no API involved other than kicking DMA chains to feed VU1 (and that was just writing a pointer into a user register anyway).
You didn't have direct hardware access from Linux, the access was exposed via an abstraction library which was a bit higher level than what the devkit allowed for.
Then there was PS2GL which was built on top of that, also not quite like the GL version on the devkit. But both shared the same fate of not being used that much.
Hi,
This is not true - PS2 Linux provided full hardware access to the graphics hardware via the SPS2 library, which just mapped a bunch of hardware registers into your userland address space.
The IOP chip was restricted but that wasn't really a big deal for most purposes, as it was generally only used for I/O.
Personally I think the main reason for Sony's delusion with the community was that instead of trying to make games, many were using their PS2Linux kits to turn the PS2 into routers, cheap GPUs or cheap PCs.
The community was never as vibrant as the XNA one in terms of game coding.
Well, if you could program VU1 you had as much low level API as anybody else, that's my point.
As for the whole Linux support - I remember Kutaragi pushing the PS2 to become a home computer. Even its devkits (T10000s) were looking like PCs and had a "workstation" mode. PS3's "other OS" seems to be a vestige of this policy.
