Slashdot is powered by your submissions, so send in your scoop

 


Forgot your password?
Close
typodupeerror
×
BSD Operating Systems

OpenBSD Requests UltraSPARC III Documentation 79

Posted by michael from the please-kind-sir dept.
An anonymous submitter writes "OpenBSD wants to run on all hardware. They've asked Sun for documentation on the UltraSPARC III processors over and over, but been stonewalled. Theo recently asked users to talk to Sun about this issue. A fairly complete thread archive can be found here. The real kicker is that Sun has released this documentation through an NDA to Linux developers..."
This discussion has been archived. No new comments can be posted.

OpenBSD Requests UltraSPARC III Documentation More

OpenBSD Requests UltraSPARC III Documentation

Comments Filter:

  • all hardware? (Score:3, Funny)

    by tps12 ( 105590 ) on Tuesday November 26, 2002 @03:01PM (#4761461) Homepage Journal
    When will they port it to my bandsaw?

  • Theo's diplomacy (Score:5, Insightful)

    by Outland Traveller ( 12138 ) on Tuesday November 26, 2002 @03:29PM (#4761748)
    Theo de Raadt writes [theaimsgroup.com]:

    > PS No, I don't work for Sun, and I'm not in bed with them. But
    > working for a LARGE company has taught me many things about
    > Bureaucracy, and two of those are: 1) Assume a lack of action before
    > an action (i.e., things tend *not* to happen in a bureaucracy), and 2)
    > if you can, pointing to a thing is almost always better than asking
    > for an unknown.

    No, you misunderstand. We've tried so hard; that is no excuse.
    Perhaps this will teach them to be less opaque.

    I think there are some times when Theo's style is dead on, like with the ipf filter. However, in this case it may not be the most constructive way to effect a change.

  • slashdotting the phone number... (Score:3, Interesting)

    by dk379 ( 548403 ) <dk+slashdot@fa[ ]org ['rm.' in gap]> on Tuesday November 26, 2002 @03:32PM (#4761782) Homepage
    I have just read Theo's pledge to users and looks like there is a person's name listed who appears to be a sole decision maker on this issue.

    I wonder how many phone calls it would take for him to get it ;-)

  • Comment removed (Score:5, Interesting)

    by account_deleted ( 4530225 ) on Tuesday November 26, 2002 @03:51PM (#4761941)
    Comment removed based on user account deletion
    • Ok, I've look but I cant find it

      What is "per-page X bit protection"

      thanks

      • Re:Theo seems very Pro-SPARC (Score:4, Informative)

        by Anonymous Coward on Tuesday November 26, 2002 @05:11PM (#4762719)
        R=read, W=write and X=execute. So, "per-page X bit protection" means the OS can specify for each memory page whether the bytes on that page can be executed as machine code. IA-32 has execute permissions for entire segments only, not for individual pages.
        • The main reason for this issue is C or C++.

          This sort of feature won't help as much for many other popular languages and scenarios.
          • Kernel is written in C, most apps too. This bit is very helpful.

            Don't try to troll about irrevelance of C.
            • Yah, but only the experts who can always allocate enough memory, always remember to deallocate at the right time, etc, should be writing in C or similar 'minefield' languages (e.g. Forth[1]). The smart obsessive compulsive paranoid sort I suppose.

              Better for the rest to create their apps using programs created by those experts.

              Most of us crap programmers don't need to deal with all the layers and it's blindingly obvious from Bugtraq we can't.

              Clear separation of roles and responsibilities. The different programmers do what they are capable of or better at.

              So you have one bunch doing stuff in perl/python/etc, not worrying about buffer overflows, malloc, pointers, but just worrying about the high level correctness of their applications (SQL injection, logic, who can see/update which bank account). While you have another bunch closer to the hardware, concentrating on making sure perl/python/etc work correctly. There's already plenty to worry about in the different domains.

              Seems more scalable and manageable. Sure performance suffers. But hey if the experts have time or are cheaper than faster hardware, we'll get them to rewrite the critical bits.

              Link.

              [1] AFAIK Forth is just as buffer overflow prone as C. Forth may be worse too since data==program which was how I crashed a Forth webserver recently. First time I noticed a Forth webserver, type stuff, press enter and poof. Wonder how things are for Lisp esp when they do the data==program stuff.

  • Hmmm (Score:2)

    by Ashran ( 107876 )
    What about this thing [sun.com]?
    What else would you need to port a kernel to it? It descibes about everything you want to know about the UltraSparc
    • Did you read the mailing list [theaimsgroup.com]? Obviously not.

    • Re:Hmmm (Score:2, Informative)

      by dolmant_php ( 461584 )
      If you would have taken the time to read the messages from Theo in the thread, he says very clearly that they don't have enough documentation to write a kernel or an OS. Also, the type of documentation they need is nothing you can find on Sun's website.

  • Jawn! You tell them Theo ;) (Score:3, Insightful)

    by Carl ( 12719 ) on Tuesday November 26, 2002 @04:10PM (#4762100) Homepage
    http://marc.theaimsgroup.com/?l=openbsd-misc&m=103 830199306037&w=2

    > I'm sorry you don't yet believe that we are striving to be a traditional
    > company that works well with the Open Source community. Most of our
    > efforts to date have been in the software arena, and I think some of
    > what you ran into was the trailing edge of Open Source awareness (in the
    > hardware business).

    The other contributions from Sun are entirely irrelevant.

    I don't care about Jinu, Jxta, Jboring, Jawn, or any of that
    stuff.

    I care about running on ultrasparc III.
  • What does Sun have to with BeOS? [slashdot.org]
  • I can certainly understand why Theo would want Sun to show support to the BSD world as they have been for Linux, but....as a practical matter, couldn't the OpenBSD kernel developers look at the Solaris 8 source (which Sun was allowing download of) and the Linux code to get a pretty good idea of how to do UltraSparc-III specific kernel space optimizations?

    As an aside, it'll be awhile before I get to run OpenBSD on an Ultra-III, my 70MHz Sparc 5 is doing just fine serving my domain with OpenBSD 3.1....you could almost here it breath a sigh of relief when the slow, load-heavy Solaris 2.6 was removed
    • Please provide link to free Solaris source download. ;-)
      • Alas, Sun is not doing that anymore, but over 2,000 people took advantage of the ability to download the source code to Solaris 8 about 18 months ago (required Sun's compiler to actually COMPILE the thing, of course).

        http://www.itworld.com/Comp/2377/IDG010628solaris/

        Unless....you're asking *me* to put it in a public place...sorry, had to give my written word I wouldn't do that to get access to it.

        And yes, I did send e-mail to "Denasse" at Sun in the linked discussion thread, mentioning my over $700K procurement of Sun server and workstation gear at places I've worked in the past 11 years, and my desire to see Sun give some support to the Open Source BSD's as they have with Linux for the UltraSparc III processor line.
    • Um... wouldn't looking at the code for something that isn't released under the BSD license taint any developer who plans to release code under the BSDL? I mean, once they look at the code, they wouldn't be allowed to include it in OpenBSD. Unless, of course, they had one team of people reading the code and writing documentation for it, and another team taking that documentation and coding on their own. If I recall correctly, that's how Compaq had to do it to reproduce IBM's BIOS.

      I can see it being against a lot of rules to just look at the source to other projects and then writing your own. Especially if you plan to use a different license.
      • haha! You're saying they can't look at source of similar projects to learn and then write very own implentation of something. I have news for you, the BSD people often look at Linux device drivers (GPL or similar non-BSD) and write THEIR VERY OWN IMPLEMENTATION of device driversfor same device for NetBSD, OpenBSD, FreeBSD, etc.etc. And other things too! Should those people should have thrown up their hands and cried "I'm TAINTED.....TAAAAAINTEEEEEED! "...kind of like lepers in Old Testament yelling "UNCLEAN, UNCLEAN!" ??? Get Real!
        • First of all, wow. You need to calm down a little bit. Second, it seems rather immoral to me that they would be allowed to do that. I wouldn't do that.
          • Matt Dillon, of FreeBSD kernel/VM: "There is constant borrowing going on between the BSDs and even between BSD and Linux, especially in regards to driver code." http://www.osnews.com/printer.php?news_id=153

            Stealing or copying code or even algorithms (without proper credit) would be wrong, but it seems these developers don't think it is wrong to look to learn, like say learn additional new SCSI command a newer model controller might support.

            Would it be wrong to look at Sun's code to see how atomicity on SMP UltraSparcIII system is done? I wouldn't think so, as long as one didn't cut & paste their code

            The educational value of Open Source of whatever licensing should be open to all, to use for whatever constructive purpose.

            P.S. Exuberqnce of last post due to 2nd day of Thanksgiving celebration, was feeling fine. And it may happen again sometime!
    • openbsd outperforming solaris on a sparc machine? last time i tried openbsd (2.8) that was exactly the opposite of what happened, it was far less performant than solaris, and crashed several times.
  • OK, I'm karma capped, lets some good ol' flaming start...

    Theo de Raadt: (calls up Sun) Hello, I demand some documentation.
    Sun Guy: Who the f*** are you?
    TdR: I'm Theo de Raadt.
    SG: Which Theo de Raadt?
    TdR: The one that is incredibly smart and productive and gets real pissy when I don't get my way; the one that forked OpenBSD because the NetBSD folks didn't like how pissy I got and drove users away.
    SG: Oh that one. What documentation do you demand because you somehow infer a right to having?
    TdR: On the UltraSparc III processor.
    SG: Oh, the one that you spent no R & D money on, that you spent no manufacturing money on, but you feel you have an absolute right to have it and if you don't get it you get pissy?
    TdR: Yeah, thats the one.
    SG: OK, here is our link.
    TdR: This isn't enough. I want more.
    SG: What other documentation are you demanding?
    TdR: I don't know. It is your job to figure out what documentation I don't have and to get it to me when I demand it.
    SG: If you don't even know what to ask for, how are you demanding more?
    TdR: Those other guys get more.
    SG: Which guys?
    TdR: The Linux guys.
    SG: You mean the ones that we kind of work with because we have an Intel distro and we should really appease the guys that kind of put it together?
    TdR: Yeah, I want what they have. I deserve it.
    SG: Why?
    TdR: Because I want it to make a server.
    SG: Using what OS?
    TdR: A free one, that will put no money in your pocket for OS licenses, no money for support, that will most likely not sell any Sun software because it usually runs as a fairly stripped down firewall box, and won't even sell any of your real expensive hardware where you make the real money from since we don't support SMP. Since you lost a lot of money when the dot-com bubble burst, and your stock is now close to historic lows and have had a couple rounds of layoffs, you must be real enthused about doing some work which probably won't get your company any money at all?
    SG: Ahh, so you demand we get some internal engineers for you who luckily will be really eager to stop their real work fending off fierce competition from IBM Windows HP and Linux, gather all our UltraSparc-III stuff for you, run it through our lawyers who luckily enough will drop all work involving our lawsuits about Microsoft and Java (and possible shareholder and wrongful termination lawsuits) sanitize it for you because from your reputation for getting pissy over things (witness ipf) you won't take kindly to an NDA and rush it to you on your schedule not ours.
    TdR: If you don't, I'll get pissy. Yes, and make sure you get that NDA stuff out. We're opensource, and we don't like NDAs, and since we're always right your NDAs should go away because we say so.

    I know why Theo would want this, but I can't see the Sun guys dropping everything and making this their number one priority. Though childish, if I was a Sun person, I'd release this stuff first to FreeBSD and NetBSD, knowing it would eventually trickle down to OpenBSD, just to piss off Theo.
  • I just talked to Ahmad Zandi from Sun and he told me that they expect Version 2.0 of the documentation to be released 2 weeks from today (The 18th of December). He said he was "waiting for some final signatures" and unless "someone drops dead" it should be available by then. We'll see if this holds. He also claimed that no private documentation has been released to Linux developers under a NDA and that any documentation which is released will be made available publically.

Slashdot Top Deals

How many QA engineers does it take to screw in a lightbulb? 3: 1 to screw it in and 2 to say "I told you so" when it doesn't work.

Close