Re: Linux kernel

From: Johan Klockars <>
Date: Thu, 5 Aug 1999 01:07:32 +0200 (MET DST)

> How much does the RAGE-driver differ from the Nova-driver? I just wonder

The sources have always been 'unified', but some things (mostly in the
headers) use #ifdef's to select the correct code.

The main differences have to do with byte swapping. The NOVA is actually
better off there for now, since we have to do a lot of swapping during
mono-expand blits on the RageII. It's supposed to be possible to change
the byte order for these things, but we haven't got that to work yet (not
that we've spent much time on it). Driver optimization will come once
everything is working as it should (the mach64 driver code is still almost
100% C and Lattice does some decidedly funny things here and there).

> if it's just a matter of initializing the card properly, or if the RAGE's

We've never had any real NOVA initialization. fVDI relied on everything
already being set up properly.

For fVDI to set up the NOVA itself, we'd probably have to borrow code from
the X drivers. Much of the setup code could of course be the same for both
cards, but memory/fifo timings and PLLs are very different (and next to
impossible to figure out even with the docs (for the strangest parts we
actually use numbers dumped from a PC ;-)).

> Mach64-API is different from my old Graphics Xpression.

No, it shouldn't differ. There are a couple of new things on the RageII,
but I don't think we use any of them yet.

  Chalmers University   | Why are these |  e-mail:
     of Technology      |  .signatures  |  
                        | so hard to do |  WWW/ftp:
   Gothenburg, Sweden   |     well?     |            (MGIFv5, QLem, BAD MOOD)
Received on to. aug. 05 1999 - 20:59:45 CEST

This archive was generated by hypermail 2.3.0 : ti. nov. 03 2015 - 20:07:54 CET