linux-mips
[Top] [All Lists]

Re: Indigo R3000 PROM calls /

To: Michael Engel <engel@math.uni-siegen.de>
Subject: Re: Indigo R3000 PROM calls /
From: Ralf Baechle <ralf@oss.sgi.com>
Date: Mon, 17 Apr 2000 14:11:40 -0700
Cc: linux@cthulhu.engr.sgi.com
In-reply-to: <200004162112.XAA02036@jordan.numerik.math.uni-siegen.de>; from engel@math.uni-siegen.de on Sun, Apr 16, 2000 at 11:12:02PM +0200
References: <200004162112.XAA02036@jordan.numerik.math.uni-siegen.de>
Sender: owner-linuxmips@oss.sgi.com
On Sun, Apr 16, 2000 at 11:12:02PM +0200, Michael Engel wrote:

> I had some time over the weekend and started to hack on Linux for my
> good old R3000 Indigo (oh yeah, I should better try to write drivers
> for the PMAG-F and the FDDI adapter in my DECstations but Indigo 
> hacking seemed to be more fun ;-)). I can load the kernel from sash 
> and it actually starts at kernel_entry (whow) and - no wonder - crashes 
> somewhere in init_arch (because I didn't change anything there ...).
> 
> Now, it would of course be nice to have some kind of debugging output
> early on. Does anyone know if the R3k Indigo has the same ARCS console 
> semantics as the Indy ? I.e. there is a PROMBLOCK struct at address 
> 0xa0001000 (as defined in include/asm-mips/sgiarcs.h) which points to 
> romvec which I can then use to dereference PROM functions ? Or is it 
> something completely different ? 

Check out include/asm-mips/mipsprom.h, it should be fairly close to what
you're looking for.  SGI used two different PROM styles for their machines,
those with 64-bit CPUs are ARCS, the older machines have another
firmware implementation.  I do have some tested code for this platform
but it's back at home while I'm here at SGI ...

> Of course, if someone unexpectedly finds some Indigo R3k hardware docs 
> somewhere, I'd appreciate it ;).

A few power series systems are the oldest stuff that I've seen over here
at SGI and even are mostly used as hightech door stops ...

  Ralf

<Prev in Thread] Current Thread [Next in Thread>