linux-mips-fnet
[Top] [All Lists]

Re: SGI

To: linux-mips@fnet.fr
Subject: Re: SGI
From: Systemkennung Linux <linux@mailhost.uni-koblenz.de>
Date: Tue, 16 Jul 1996 04:24:02 +0200 (MET DST)
In-reply-to: <199607152340.BAA00166@miston.any.net> from "Didier Frick" at Jul 16, 96 01:40:28 am
Bonjour Didier,

>     Ralf> For a long time noone of you DECstation people has posted.
>     Ralf> What's the status?  There are some people in switzerland
>     Ralf> that are working on R3000 stuff which will be very usefull
>     Ralf> for the DEC.
> 
> Well, we hope so :-)

Not to forget R3000 Indigos ...

> A little status report: our working version of the kernel is now moderately
> usable on our R3052-based (R3000 instruction set) board.

> I made major changes to the TLB code , which is now horrendously unefficient,
> but seems to work. We also changed various spots throughout the mips-specific
> parts to make them R3000 compatible and endianness-independent.

Indeed, as written the code wasn't the right thing to fit the R3000.

> The trouble is: our version is based on version 1.3.98 of the kernel. We
> would now like to integrate kernel updates up to 2.0.max and make our 
> R3000 adaptations available.

Luckily we're living in a time of code freeze, so this won't be that hard.

> I suggest that some assembly-language files (for example head.S) be splitted
> into R3000 and R4000 versions because they really become cluttered with
> conditional directives.

Yes, you're completly right.  My plan was to put all the code dealing with
exception handling from head.S into mips[1-4]/.  The parts of machine and
prozessor specific initialization that are in head.S should be rewritten
to C for maintainability and go into <arch/mips/<machine>/setup.c and
arch/mips/kernel/setup.c.  What will finally be left in head.S is more or
less nothing but initializing a few CPU register.

> Besides, we really need to have shared library support.

I'm still debugging.  GCC, binutils and the dynamic linker had bugs.  The
bugs in GCC are gone and since two days I finally have a set of binutils
where all the annoying bugs like abort when striping most files are
gone also; the dynamic linker is getting more reliable also.

> Does anybody have any hints on how to do all this in the most efficient
> way (if there's one :-) ?

The shared library stuff is the easy part to do for you:  nothing to do
because the minimal required support is already in the kernel and the
userland is completly o32 ABI/MIPS ISA I.

For the rest just send me a diff using options -urN.  I assume that most
of you patches can easily be integrated; we can then talk about how to
integrate the rest of the patches.  I'm not yet as swamped with patches
as Linus is, so this will probably pretty painless for you.

  Ralf

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