linux-mips-fnet
[Top] [All Lists]

Re: Deskstation rPC44 and Milo 0.27

To: linux-mips@fnet.fr
Subject: Re: Deskstation rPC44 and Milo 0.27
From: Ralf Baechle <ralf@Julia.DE>
Date: Mon, 2 Sep 1996 21:14:17 +0200 (MET DST)
In-reply-to: <199609021827.MAA21034@rover.village.org> from "Warner Losh" at Sep 2, 96 12:27:13 pm
> : Can you hack the code so that it prints exactly what it gets passed by
> : the ARC firmware?
> 
> Yes.  I was going to do that to see what is going on with the args,
> since I'm really wanting to start to use them...

The way that theyuse to pass arguments/environment variables to the
started program is a bit wired and I think there might well be a bug.
Hell, if M$ can make NT run, then we'll get Milo running on every
MIPS box too, won't we?  Just angry againabout this ARC shit ...

> Hmmm.  I know that the rPC44 is happy to report correct information
> (or nearly so) about the memory.  Is it just the Tyne (which is a
> known rogue in many respects), or are there others as well.

The data you've posted looks pretty sane, that's right.

                                                                I'd
> suspect the PICA also, since that was an early board, or maybe the
> Magnum, since it is just a stripped down pica (or more accurately a
> pica is a built up magnum).

If the PICA is a built up or a stripped down Magnum is arguable.  The
Magnum has 16bit/48KHz stereo sound which the PICA doesn't.  Hell, so
far noone has yet written a noise^H^H^H^H^Hsound driver for it ...
Then again the PICA is faster due to higher clock, l2 cache and a
R4400 CPU.

> Too bad I don't have hardware docs on the rPC, or I'd know how to ask
> the hardware directly.  There doesn't seem to be a R4030 controller
> chip like the PICA has on my motherboard to ask (or is that integrated
> into the R4400PC and I've just not stumbled across references to it
> yet).

Sounds improbable to me.  So far I know that Acer, NEC and Olivetti had
OEM licenses for the Mips Magnum design.  Again you might try to
disassemble NT drivers/HAL.  I happen to have ported the binutils for
NT targets just for this purpose :->  The port is however horrible; it
links only with the option -noinhibit-exec to ld but that's good enough
for objdump -d and that what I did it for.  Tell me if you need the
patches ...

> : Even better - apparently NT's equivalents to Linux's inb/inw/inl/outb etc.
> : functions are macros compiled into NT which means that every MIPS NT
> : machine is assumed to have the same 1:1 mapping of ports into memory -
> : just the base address is different.
> 
> I've seen something similar when disassembling code from NT.  I think
> that the pica port in NetBSD and the arc port in OpenBSD do the same
> thing.

Well, what I say isn't just a guess.  It's what the M$ HAL specs for
WNT 3.51 say.

  Ralf

-- 
A weird imagination is most useful to gain full advantage
of all the features - manpage of amd(8).

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