OpenBSD (Still) Seeks UltraSparc III Docs From Sun

An anonymous reader writes "There is a very interesting article on kerneltrap regarding OpenBSD's lingering battle with Sun over UltraSparc III documentation (that's right ... it still hasn't been resolved). Jeremy Andrews relates his efforts to get a position from Sun on the matter. In summary, he was completely stonewalled ... and that is exactly what makes the article so noteworthy."

Comments

Why ever would they give up the IP ?

[ThundaGaiden]

Sun has always seemed to be in the bussiness of
sharing , but oftens seems to do a stab in the
back

As an example Staroffice was a awesome piece of software, they release the source and everything ,
then all of a sudden you have to pay for it... and there's a open source solution that has to catch up to the new release that Sun just made.

I must admit I like Sun's approach more than MS's make it free , make it opensource , then make em pay...

MS just makes you pay :P

Re:Why ever would they give up the IP ?

[sql*kitten]

Sun has always seemed to be in the bussiness of sharing , but oftens seems to do a stab in the back

No, Sun are in the business of making money. All the cool stuff they do has to be paid for one way or the other. A fab is possibly the most expensive artifact on the planet, short of an aircraft carrier. Sun are perfectly willing to provide documentation to anyone willing to meet them halfway and sign an NDA, which is fair enough, since they need to protect their opportunity to earn a return on their investment. (Note that I said opportunity, not right).

This is not a flame or a troll, but the OpenBSD people's position is "we want you to respect our terms, but we aren't willing to respect yours". Well, you can't have your cake and eat it - and no amount of ranting will change that.

Re:Why ever would they give up the IP ?

[flynn_nrg]

"Sun doesn't actually fab anything. They only do the chip design, and have them fab out-of-house."

Absolutely. Given the small volume of UltraSparc cpus produced every year (very small compared to the x86 world), it would be extremely expensive for them to have an own fab. During my work at Sun I saw USIIi cpu boards made by companies like IBM (those had the well known ecache bug) and Sony.

Re:Why ever would they give up the IP ?

[TPS Report]

Sun are perfectly willing to provide documentation to anyone willing to meet them halfway and sign an NDA, which is fair enough, since they need to protect their opportunity to earn a return on their investment.

If they are so willing, as you say, why has an offer not been made? The fact remains: he has not been offered the documentation for something that is supposedly "open" and availble (whether under an NDA or not).

This is not a flame or a troll, but the OpenBSD people's position is "we want you to respect our terms, but we aren't willing to respect yours".

No, you have that wrong. The OpenBSD position is "What ARE the terms?" -- they haven't been offered anything at all.

I think they have been ignoring his requests because they realize Theo is not going to be happy with an NDA, and is going to call the bluff. On one hand, Sun says it's an open architecture, but on the other hand - they're telling people they need to sign an NDA to get the data. I believe Sun would have offered him the NDA a long time ago if they thought he would sign it -- just like they've done with other (Linux) developers - but they probably realize it's not going to work like that with him. I don't think Theo is as concerned about the actual documentation; I think his driving motivation here is to get Sun to "practice what they preach."

Theo is of the opinion that if you're going to say something is a certain way, it better be. Sun will need to either ignore him indefinately, thus avoiding the whole NDA paradox -- or they will change the policy and truly offer the documention in an open manner, as they claim.

What I find interesting is the comment near the end of the article that basically says Theo thinks they have everyone sign an NDA to avoid public discussions of architecture bugs in the US3. Am I the only one that remembers Sun having customers sign NDA's so they could get their faulty UltraSparc chips replaced with bad cache? I can't find a link to past discussions about this, but...

Re:Why ever would they give up the IP ?

[dagnabit]

Low volume isn't the only reason they don't fab themselves... outsourcing that process allows Sun to seek best-of-breed fab technology without having to invest huge $$ in fab infrastructure and then have to "get their $$ worth" out of it. Which makes the stockholders happy (happy as can be, anyway...)

And TI (their main fab partner) gets the symbiotic benefit of having a reason to accelerate and chase the latest and greatest fab stuff to then use on their _own_ products (mainly DSP chips).

They want to do a linux distro..

[Jondor]

The whole story leaves me with the idea that sun hopes to get most of the expensive development done by the OS community as they are preparing their own linux distro. As such nothing wrong with that..

For OpenBSD they couldn't care less other than to keep them waiting and to keep possible competition at a distance.

Just my first, uninformed, impression..

Re:They want to do a linux distro..

[JimDabell]

For OpenBSD they couldn't care less other than to keep them waiting and to keep possible competition at a distance.

If Sun want to create a distro, fine. It won't be to profit from the distro though - it will merely be a "value-add" to the hardware. As such, competition is hardly something they would be scared of - the more operating systems running on their hardware, the better.

I Don't Understand This

[tealover]

Wasn't Sun in part started by Bill Joy of the *original* BSD fame, of which OpenBSD is an off-shoot? You'd think that Bill would show the OpenBSD guys some love.

Yes, I know Bill Joy creating BSD was a long time ago but there is no need for such duplicity and passive hostility. Let's get it together Sun !

Re:I Don't Understand This

[arvindn]

That's the problem with these big companies. They're made of disparate groups each with its own world view. So statements like "company foo is pro/anti open source" become meaningless. There is a lack of central vision and co-ordination. (Hint: what was one of the reasons Microsoft became what it is?) Look at IBM. Invests heavily in Linux, but OTOH is extrememly protective of its IP. Look at HPQ. (Remember the Perens anti-DMCA demonstration circus?) Look at SCO. They're all the same, vacillating (no pun intended). Except for exclusively OSS companies like RH, you can never tell.

Re:I Don't Understand This

[sql*kitten]

That's the problem with these big companies. They're made of disparate groups each with its own world view.

Their own P&L you mean. Each group is responsible for reporting profits to head office, and if those profits aren't there, then heads will roll.

That means that if one business unit has to pursue a policy that doesn't help another in order to protect its own profits, it will. This happened at IBM: the PC hardware division wasn't willing to risk it own sales by by preloading OS/2 just because it would have made things easier for the OS/2 division.

Which do you suppose makes more profit (and hence has more influence at head office) at Sun: the Open Source advocacy group, or the UltraSPARC engineering group? (Hint: the workstation and server group probably "buy" the CPUs internally from the SPARC group).

The docs ARE available...

[Some Bitch]

...but there's an NDA associated with getting them. It looks to me like De Raadt doesn't want to agree to the terms and threw his teddy out of the cot when Sun told him 'tough luck then'.

Re:The docs ARE available...

[iMMersE]

Because the Berkley'd code can be used in other commercial software with just a little copyright message, whilst GPL'd code can be used in other commercial software, but then requires the source code of that software to be released.

The key of the article

[fruey]

This is very interesting, because it really hits on the blurry line between "open" and controlled (closed), and also between the way that Linux developers signed something to not disclose information on the hardware itself, although their source code to access this hardware is available.

Sun boasts their UltraSparc III as an "open" architecture, yet seem to recognize that there is insufficient information freely available for the open source community to support it with operating systems. I have been told that the required documentation does exist, however, with a Sun part number of 805-0408-05-P. An early version of this manual was allegedly made available to Linux developers once a Confidential Disclosure Agreement was signed (Sun's version of a Non-Disclosure Agreement), however no such offer has been made to the OpenBSD team, an offer that if made is likely counter to the project's goals.

Clearly then OpenBSD developers are sticking to their guns, their question is really how an "Open" architecture cannot be disclosed without some contractual agreement.

I begin to suspect that the other comment (against Linux kernel devs) about this may be key:-

There's always people who suggest it is possible, but the pain is so high, it's just not worth it. Especially when the Linux kernel's interface with hardware is detailed about as well as the Linux manual pages. Especially when Linux is famous for stuff like: writereg(0x4, 0xff01);

Now, if I were a conspiracy theorist, I might say that the precise reason that comments are sparse in these sections of CPU code for the Linux kernel is due to some clause in the Sun disclosure agreement.

In any case

• The Sun Ultra Sparc III cannot be open if you cannot access it publically, it is rather available under terms and conditions
• The Linux project clearly had some other motivation to write the kernel code for this architecture, perhaps even encouraged by Sun (think Cobalt Linux?)
• The OpenBSD project is somehow staking out that they are "purer" for adopting this stance, which is all very noble, but means ultimately that OpenBSD is unlikely to support this architecture (apart from OpenBSD zealots, most of us will just run Linux instead, I guess).

Still, I'd like to see as much openness from vendors as possible. They have to realise that the people who support closed source business models are going to be driven out by cheap commodity hardware which is now powerful enough to do amazing things (think clusters of cheap hardware on AMD/Intel/Sparc architectures all talking together via some OpenSource kernel and clustering project, think Google). Their days are numbered, sure they'll still have a place, but their creaming off profits from their current installed base will start seeing serious competition from value added service providers with no ties to specific hardware, and that is great for the consumer. Do not forget, it used to be IBM, Sperry (then Unisys) and Burroughs that did all hardware and software support. Now, as an independent consultant, you can get out there and do amazing stuff with commodity hardware and your own tailored solutions pulling from a wealth of great free server software solutions.

We are already in a phase (as are companies in European telecoms like France Telecom and British Telecom) where holding on to what you have for as long as you can, before competition really breaks you, is the only business plan they seem to have.

Score one for innovation and open projects, every time.

Re:The key of the article

[sql*kitten]

Clearly then OpenBSD developers are sticking to their guns, their question is really how an "Open" architecture cannot be disclosed without some contractual agreement.

You've misunderstood the word "open". SPARC is open in that anyone can download the specification and implement it - you can set up a rival SPARC-based hardware company, fab your own SPARCs and compete with Sun, if you want to, and they will have no legal means to stop you.

However, what is not open is Sun's own implementation of that SPARC specification. That's because they spent a lot of their own money on it, creating the best implementation they could. If you want your own SPARC, you will have to implement it on your own.

Think about it this way: if you want to have your own recording of a Mozart symphony, you can buy one that an orchestra has recorded, or you can perform it and record it yourself. What you can't do is take a pre-recorded copy and try to pass it off as your own - even tho' the score is in the public domain.

Re:The key of the article

[sczimme]

"The Sun Ultra Sparc III cannot be open if you cannot access it publically, it is rather available under terms and conditions". (original emphasis)

GPL = (terms and conditions), too.

[relax. think about it.]

Politics

[lockne]

Article: "An early version of this manual was allegedly made available to Linux developers once a Confidential Disclosure Agreement was signed (Sun's version of a Non-Disclosure Agreement), however no such offer has been made to the OpenBSD team, an offer that if made is likely counter to the project's goals."

So what they're essentially saying is that they want Sun to give them the documentation without the OpenBSD developers having to sign an NDA, because doing so wouldn't be in line with the OpenBSD goals?

Sun is free to refuse. And the OpenBSD folks are free to reread their own goals and start taking them seriously. For example these two:

- Be as politics-free as possible; solutions should be decided on the basis of technical merit.
- Do not let serious problems sit unsolved.

Now go sign that NDA! :-)

Sun And Documentation...

[Bowie J. Poag]

Sun, with ANY kind of documentation, is going to be a royal pain in the ass. Here, i'll give you a personal example.

One day, I picked up a SparcStation 1 at a surplus auction. Cool, I thought, I'll learn SPARC architecture, a bit about disaster recovery with Sun hardware, Solaris, you name it. So, I hacked the hell out of it, and learned everything I could without documentation. When it came time to look at a manual. I called Sun.

"Hi... I was wondering if you could send me the owners manual for a SparcStation 1."

"Sorry. Thats handled by SunStore."

"Whats SunStore?"

"They handle all our documentation."

So, I call SunStore, and ask the same question.

"Hi.. I was wondering if I could order a user's manual for a Sun SparcStation 1. I know the machine is like 10 years old, but do you still have the manuals?"

"Yes, we do."

"Great, i'd like to order one, then. Is Visa ok?"

"Uhh.. Well, we can't sell it to you."

"What do you mean?"

"Well, we cant sell you just one."

"Huh?"

"You need to order in lots of 500."

"You mean in order to buy a SparcStation 1 manual, I need to buy 500?!"

"Yes."

"Uhhh.... Ooooh-kaaay.. How much is a lot of 500?"

"$39.95"

"Oh, okay..I guess thats fine.. I dunno what i'm gonna do with 499 Sparc manuals tho. I guess you can keep them, and just send me one. Thats all I need."

"39.95 is the unit price, sir. You're looking at a total of.... $19,975."

"No way!"

"Yes sir. Will this be on a Visa or Mastercard?"

*click*

Re:Sun And Documentation...

[LeftHanded]

Great story! However, you can get all kinds of user manuals, hardware manuals, software manuals, etc from http://docs.sun.com You can read them on-line, download Adobe Acrobat versions, and purchase the documents as well. It's a lot easier than it used to be... HMM, no Sparc 1 manual, but the Sparc Classic and similar is at http://docs.sun.com/db/doc/801-2176-13:

Re:Sun And Documentation...

[Andrew Francis]

Sun, with ANY kind of documentation, is going to be a royal pain in the ass. Here, i'll give you a personal example. One day, I picked up a SparcStation 1 at a surplus auction.

My experience was quite different. I was trying to get some Sun Xterminals (rather old; they were basically Sparcstation 2's without hard drives) booting and serving up displays from a Red Hat machine, instead of the aging Sparcserver we were about to retire.

We had support contracts with Sun for several machines, but not the Xterminals or the Sparcserver they booted from. I put in a request with Sun (via a web form) anyway. Within a few hours, someone at the local Sun office was on the phone to me. The next morning, I had a single copy of the manuals on my desk, via courier.

PS - there's a heap of stuff on docs.sun.com

Sun may be in the fetal position as a company..

[gorjusborg]

Sun seems to be a company with an identity crisis. Are they a hardware company that dabbles in software? Are they a software company who dabbles in hardware?

Either way, they are looking at a bleak future considering the proliferation of Linux, and the availability of cheap, relatively high performance x686 hardware.

If Sun is a software company, they are probably not comfortable with the fact that Solaris, recently the 'standard' OS for low-end scientific/technical computing (at least in the semiconductor industry) is being passed over in favor of the cheaper (faster) alternative Linux.

If Sun is a hardware company, they should be worried. The semiconductor company I work for, which previously used Sun machines exclusively, has found that a dual XEON running Linux outperforms the new SunBlades it owns. The blades cost about 4-5 times more than the XEON systems, and have about half the speed for our applications.

I am not surprized if Sun is starting to feel that it has to protect itself.

Re:Sun may be in the fetal position as a company..

[joab_son_of_zeruiah]

Yes. And, (I speculate) that the internal struggle is between the (at least) "certain" hardware past and the (for sure) uncertain java future.

I think what we are seeing here is a "strategic" stonewalling of the hardware side to try to maintain its "margins" - which are often a consequence of some kind of proprietary software. OpenBSD kind of kills that idea.

I speculate that internally Sun has advanced a kind of "let's have the best security platform in the industry" motion for its hardware. Release Ultrasparc III specs? That's counter strategic (aka - shoot yourself in the foot) especially if its OpenBSD.

So you be McNeally: how are you going to explain to stock analysts your great plan for resurrecting your company when you give away your proprietary advantage?

I'm just a bitter shareholder.

Sun should do the right thing...

[Noryungi]

The i386-based machines are seriously threatening Sun.

Linux, OpenBSD and other open-source OSs (NetBSD, FreeBSD) have already proven they can replace Solaris in most cases.

Admit it, Sun: your best bet for survival against the Microsoft Juggernaut is not just to pay lip service to Open Source. It is to be truly open. Otherwise, platforms such as the UltraSparc will be abandoned in favor of cheaper and more competitive architectures...

Release these docs, Sun. Prove to the Open Source community, and not just to Linux kernel hackers, that you are serious about supporting alternative Operating Systems.

Fujitsu SPARC64GP

[southpolesammy]

If Sun won't cooperate with the UltraSPARC III, then why not engage Fujitsu to develop kernel code for their SPARC64GP processor line? It's at least as good, if not superior, to the UltraSPARC III, and it sends another message to Sun to say, watch out -- you're not the only fish in the pond...

Re:Fujitsu SPARC64GP

[Tet]

Running OpenBSD on anything but a single blade 1000/2000/ the upcoming jalapeno-workstations or a small server is - IMHO - completely nuts.

Given the lack of SMP support, then yes, OpenBSD doesn't make much sense on larger Sun boxen. However, running OpenBSD on a V100 or V120 makes perfect sense, particularly for firewalls. We currently use Compaq DL320/360s for this, but Intel hardware suck, and Sun hardware is a joy to work with remotely (particularly the Netra T1 and V100/V120 series, which can be remotely power cycled via the serial port). Does anyone know if the V100/V120s are supported by OpenBSD? They're not on the list of approved hardware, but they do have an UltraSPARC IIi, rather than the UltraSPARC III that is causing so many problems. I might have a go at installing it when we get our next batch in...

More pressing issues for Sun

[NDPTAL85]

I think Sun has a few more important concerns right now than helping a rather obscure version of BSD run on their proprietary hardware.

I think they may be trying to regain profitability right now and OpenBSD compatability just isn't going to help in that regard.

Re:More pressing issues for Sun

[Arandir]

I think Sun has a few more important concerns right now than helping a rather obscure version of BSD run on their proprietary hardware.

What I find funny about this whole article is the underlying hypocrisy. I'm not pointing at you necessarily, but I am pointing at a bunch of two faced Linux advocates posting here.

If this story was about GNU not getting the docs for use with Hurd, there would be a major hue and cry. If it was about Linus and Alan having to sign a proprietary contract before getting Sparc docs, Sun hardware the world over would be burning in bonfires. But it's OpenBSD, so they don't care. Linux people are telling Theo to "suck it up" and sign an NDA. Hey wake up you nimwits! Non-Disclosure Agreements are the very antithesis of Free Software. You cannot be an honest Free Software advocate while arguing that OpenBSD needs to sign an NDA. You cannot say that freedom is for Linux users but not for anyone else. You cannot say that it's wrong to dominate and subjugate Linux users, but that it's okay to shackle OpenBSD users under onerous restrictions.

Re:More pressing issues for Sun

[guacamole]

But it's OpenBSD, so they don't care. Linux people are telling Theo to "suck it up" and sign an NDA.

Acutally, the Linux/SPARC developers did suck it up and sign the NDA to get the docs.

What exactly do they need...

[georgewilliamherbert]

...beyond what is found in, for example;

[warning 5+mb download]

http://www.sun.com/processors/manuals/usIII_um.p df
?

If that's not good enough, fine, what areas need
more info? What exactly are they looking for?