Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 10 Aug 1999 07:22:35 -0400 (EDT)
From:      Daniel Eischen <eischen@vigrid.com>
To:        hackers@FreeBSD.ORG, jb@cimlogic.com.au
Subject:   Re: FreeBSD on VMEbus
Message-ID:  <199908101122.HAA10329@pcnet1.pcnet.com>

next in thread | raw e-mail | index | archive | help
John Birrell wrote:
> I have an application which should use VxWorks like the rest of the
> project, but since this part has been subcontracted, Wind River wants
> A$35,000 for an <expletive deleted> OEM licence for the right to link to
> their libraries. Blood suckers. They only charge A$5990 for the tools!
> 
> I am being pressured to write a kernel driver for NT to run on the same
> machine that is doing some pretty heavy FFT calculations using LabView.
> I want to put a box in between, and if I can't have a realtime operating
> system, then I need something that I can embedded, like FreeBSD.
> 
> The question is: out of all the VMEbus PCs available, which one is the
> best choice based on drivers working out-of-the box? Ethernet is
> particularly important.

We've got a couple of General Micro Systems (www.gms4vme.com) SBC
PCs, one single CPU and one dual CPU Pentium.  They run Labview and
NT fine :(  I've booted FreeBSD on them a couple of years ago
successfully.  These SBCs (Mustang) were nice in that they had
on-board aic7880 SCSI controller and DEC 21040 10/100 Ethernet.
The newer models (see above web site) look rather nice also (dual
Intel 10/100 Ethernet and 53C875 SCSI).

A Tundra Universe II chip provides a PCI<->VMEbus interface.  I've
got the full documentation for the chip and it should be possible
to make it work under FreeBSD, especially with newbus.  You can
download the documents for the chip at:

  http://www.tundra.com/index_prod.cfm?tree_id=100362

This isn't going to work out of the box, but if you only want
to memory map VME devices, it shouldn't be that hard to teach
FreeBSD that much.  The bridge chip provides both master and
slave addresses (A16/A24/A32) that look to be easy to configure.
The chip also supports mapping of VMEbus interrupts to PCI bus
interrupts.

> For those who are interested, here is a link to the main project:
> <http://www.rlmsystems.com.au/projects.htm>; (JORN).
> 
> Please save me from a fate worse than death. I got this contact on the
> basis that it _wasn't_ a Microsoft development, so either I need to jump
> off the nearest tall building, walk away, or find an alternative...

Supporting VMEbus under FreeBSD is something that we'd like to
see also (especially the Tundra chip).

Believe it or not, we have had good luck running real-time simulations
under Solaris on Force Computers SPARC SBCs with VMEbus.  They provide
the VMEbus (nexus) driver for Solaris, and you can easily write Solaris
drivers for VMEbus devices.  mmap is also supported for simply mapping
memory to VMEbus devices.  We've done both without any problems and
have a fully populated 21-slot chassis.

Dan Eischen
eischen@vigrid.com


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199908101122.HAA10329>