OpenBSD Journal

Improved Sparc64 Support

Contributed by dwc on from the come-on-baby-light-my-sunfire dept.

Mark Kettenis recently wrote to the misc@ and sparc@ lists with the following good news about UltraSPARC-III, and asks everyone to see what else is at least partially working…

OpenBSD 4.0 brought support for UltraSPARC-III processors.
Unfortunately that support was not complete and we had to disable the
L1 data cache on the cpus.  Over the last few months we made
significant improvements to the code that made it possible to fully
enable the UltraSPARC-III on-chip caches, which makes machines with
these cpus at least twice as fast.  On top of that, thanks to donated
hardware, we were able to add cas(4), a driver for Sun's PCI
GigaSwift/Cassini ethernet adapters, which is found onboard on many
UltraSPARC-III based machines. like the Sun Fire 280R, V480 and V880.
So now would be an excellent time to test OpenBSD snapshots on your
favourite Sun Hardware.

We're also very interested in people trying OpenBSD on machines that
are still listed as unsupported on OpenBSD/sparc64 web page:

    http://www.openbsd.org/sparc64.html

Although these machines are listed as unsupported, there actually is a
chance that OpenBSD will run on them.  Reading through the
UltraSPARC-IV processor manual, there is no reason why our kernel
would not run on that CPU.  So OpenBSD might run just fine on the Sun
Fire V490 and V890.  The same goes for the new PCI-express base
machines like Ultra 25/45 and Sun Fire V125/V215/V245/V445.  And if
nobody ever tries running OpenBSD on these machines, we'll never know.
So if you have access to any of these machines, please try booting
OpenBSD on it and send us a dmesg.

Thanks,

Mark

(Comments are closed)

Comments

  1. By Brian (68.184.126.151) on

    Wee.. finally a use for our retired E6800 at work? I'll give it a shot and let you know. ;-) Depressing to see what a million+ a few years ago turns out to now.

    Comments

    1. By Anonymous Coward (208.48.231.12) on

      > Wee.. finally a use for our retired E6800 at work? I'll give it a shot and let you know. ;-) Depressing to see what a million+ a few years ago turns out to now.

      I don't think they have smp support for sparc64 yet... very unfortunately.

      Comments

      1. By Anonymous Coward (74.109.131.13) on

        > > Wee.. finally a use for our retired E6800 at work? I'll give it a shot and let you know. ;-) Depressing to see what a million+ a few years ago turns out to now.
        >
        > I don't think they have smp support for sparc64 yet... very unfortunately.

        Which will possibly never happen if people don't test things on their own machines, hence the use :)

      2. By Anonymous Coward (70.179.123.124) on

        > > Wee.. finally a use for our retired E6800 at work? I'll give it a shot and let you know. ;-) Depressing to see what a million+ a few years ago turns out to now.
        >
        > I don't think they have smp support for sparc64 yet... very unfortunately.
        Well, just like the god of premature ejaculation, it's coming soon
        (with apologies to Mel Brooks)

  2. By Terrell Prude' Jr. (70.169.167.212) tprude@cmosnetworks.com on

    I know that this doesn't pertain to the UltraSPARC III specifically, but I do enjoy the fact that UltraSPARC support is there. Ultra 5 pizza boxes make absolutely terrific firewalls and spamd hosts. It's good to see the UltraSPARC III now better supported...but I've always wondered what idiot at Sun made the decision to keep UltraSPARC III under wraps anyway, given the existence of AMD's Opteron....

    I wonder...is the Niagara UltraSPARC supported? I'd think it'd be pretty easy (for a kernel hacker) since that chip design is wide open. But of course you'd need SMP/SPARC to take full advantage of it.

    And yep, I bought my CD set. Thank you, OpenBSD team, for a solid platform that continues to make old--and rock-solid--boxes useful. May Puffy in all his incarnations--Pufferix, Ponderosa Puff, and my favorite, Harisenbon--continue to live long and prosper.

    --TP

    Comments

    1. By Joachim Schipper (Joachim) on

      > I wonder...is the Niagara UltraSPARC supported? I'd think it'd be pretty easy (for a kernel hacker) since that chip design is wide open. But of course you'd need SMP/SPARC to take full advantage of it.

      I don't use SPARCs (lack of cheap spares lying around, mostly), but I seem to recall someone influential (possibly Theo) pointing out that getting the chip to work may be easy, but that documentation for about all the other parts of hardware in such machines is lacking or nonexistent, and as such it's not as easy as it appears to be.

      Joachim

      Comments

      1. By Terrell Prude', Jr. (70.169.167.212) on

        > > I wonder...is the Niagara UltraSPARC supported? I'd think it'd be pretty easy (for a kernel hacker) since that chip design is wide open. But of course you'd need SMP/SPARC to take full advantage of it.
        >
        > I don't use SPARCs (lack of cheap spares lying around, mostly), but I seem to recall someone influential (possibly Theo) pointing out that getting the chip to work may be easy, but that documentation for about all the other parts of hardware in such machines is lacking or nonexistent, and as such it's not as easy as it appears to be.
        >
        > Joachim

        Yep, that, too, has been a traditional Sun problem, and it's why our Sun Blade 1000 didn't work with OpenBSD for years. I ran Aurora Linux on it for a while and then gave up. It's been collecting dust ever since. But now it may be time to blow off the dust on that guy and try again.

        spitz@cmosnetworks.com

Latest Articles

Credits

Copyright © - Daniel Hartmeier. All rights reserved. Articles and comments are copyright their respective authors, submission implies license to publish on this web site. Contents of the archive prior to as well as images and HTML templates were copied from the fabulous original deadly.org with Jose's and Jim's kind permission. This journal runs as CGI with httpd(8) on OpenBSD, the source code is BSD licensed. undeadly \Un*dead"ly\, a. Not subject to death; immortal. [Obs.]