*BSD News Article 46622


Return to BSD News archive

#! rnews 2671 sserve.cc.adfa.oz.au
Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!simtel!oleane!dish.news.pipex.net!pipex!edi.news.pipex.net!pipex!sunic!sunic.sunet.se!mn6.swip.net!seunet!news2.swip.net!enea.se!pefo
From: pefo@gordon.enea.se (Per Fogelstrom)
Newsgroups: comp.unix.bsd.netbsd.misc
Subject: Re: Port to Mips ?????????
Date: 10 Jul 1995 09:08:20 GMT
Organization: Enea Data AB
Lines: 55
Message-ID: <3tqqm4$9j9@gordon.enea.se>
References: <805160740.26161@imp.demon.co.uk>
X-Newsreader: TIN [version 1.2 PL2]

Phil Taylor (phil@zipmail.co.uk) wrote:
: Is there any possibility of a Mips port for NetBSD ?

: I have an Olivetti M700-10 which is a Mips 4000 processor based system and 
: would like to install a BSD based unix.

Well, this is the current status:

The NetBSD R4400 port have just started to 'work' on a Acer Labs PICA_61
system. The PICA_61 is a R4400 pc-style motherboard similar to the Olivetti
intended for Win NT.

For now this is what is working (or almost):

	o Booting from Dos floppy. NetBSD boots from the dos
	  floppy and then abandons the BIOS completely. In the
	  future it will boot from a small DOS partition on the
	  SCSI disk. But this is when DOS-FS works with NetBSD/PICA.

	o Mapping works, although there is something strange sometimes.
	  More testing to verify that cache aliasing is taken care of.
	  (Cache aliasing is from the viritual indexed and physical
	   tagging cache scheeme).

	o Console driver (pcons) works.

	o SCSI driver works (taken from pmax asc driver). DMA handling
	  needs to be improved. Still uses buffers instead of real
	  scatter/gather. Also needs to be more flexibel in respect
	  to dma resource handling (Channels, mapping etc).

	o RTC 'driver' works.

This means that with the old miniroot i have a can boot up to single
user mode and do ls, disklabel, newfs etc. My next step is to work out
the strange mapping problem and create a new root on another disk and
update the system commands to latest versions. After that the mapping
code need some cleanup, it's quite paranoid right now :), always
flushing caches etc. Perhaps page allocation needs some improvement
to to improve on cache alias flushing.

Progress of the future work is completely dependent on how much time
i can spend on this work. It has been down ti'l about a week ago when
i picked it up again to bring it to the current status.

As you might understand there are still plenty of work to do before
'everything' works :).

If anyone have suggestions, want to do some work etc
please feel free to mail me and we figure something out.

Iv'e also heard that someone else is working on a port for the R4000
but that's also all i know.

Per