Did MS look to xen and others for virtualization-ware because they were aware of this? Although that would have been interesting if they had accepted it, then they would be under the GPL!

Nobody wanted to deal with this and still no one does

it's not a happy thing and pretty soon all hell will break

loose copywrite holders are going to have the final

say on this when it comes to Linus et al they aren't

so inflexible but you can bet if it turns out vmware is

breaking the rules they will be forced to see reason.

Look nobody wants anyone out of business for using

Linux nobody sane anyway but the community

expects their contributions will remain open source

otherwise they won't make them they also expect the

license under which their work is distributed be honored

if it isn't they will have their answer. Of course vmware

could pay them all off but that would be cheating.

Either VMware is making money selling other people's work or it's not. It's as simple as that. If they are using GPLed code, then they are in violation. It's not as if they are actually giving people Linux to run, most of their customers don't have the foggiest idea what's on the base. But if it is the Linux kernel, then they agreed to certain rules when they used it.

Brilliant logic there from Gordon. Obviously VMWare aren't doing anything wrong, because if they were they would have been caught and punished. They haven't been caught and punished, therefore they're all good.

Once again the Linux community takes aim at its own foot. Before anyone has a complete understanding both practical and legal of VMwares ESX Server Linux is mudding up its own image. Its bad enough that Linux complains about Microsoft standards ( when Linux really has no standards) now it is shouting unfair at one of Linux's strongest supporters.

This question has come up numerous times over the years now and everytime the answer is the same.

Deafening silence.

Enough is enough. It's time for VMware to answer the question. No more silence. VMware just had an IPO and made about a BILLION dollars and this is not how responsible companies act. If there was no issue here, they should have answered this question YEARS AGO, so we wouldn't be even having this debate. This is IRRESPONSIBLE.

As for Gordon Haff, he, once again, shows just how much an idiot he really is. He has zero credibility.

"Haff continues that legal issues aside, it would be at least manifestly unfair for such an issue to be realized at this late date when it's been sitting in more-or-less plain sight for years."

And it was manifestly unfair when Unisys sat quietly on the sidelines while the .GIF graphic format became a web standard, then turned around and claimed IP rights and licensing fees only once it became established as such. It was manifestly unfair when Fraunhofer did exactly that with the .MP3 audio format too. So, tough. If those are the rules these proprietary companies play by, they have no right to complain when the OS community pulls the same game. What's good for the goose is good for the gander, as the saying goes.

Long discussion here; some useful comments.

http://linux.slashdot.org/article.pl?sid=07/08/14/1618241

> Linux complains about Microsoft standards

Not sure what `Linux complains' means...

> ( when Linux really has no standards)

..anyway, trying to adhere to existing standards instead of calling any random crap `a new standard' really results in a very small number of `standards' created -- ideally only for the few things that are OS-specific by definition.

Unfortunately the premise that the linux kernel included in the service console in ESX bootstraps the vmkernel is incorrect - the vmkernel boots the system, and then brings up the service console.

As the base premise of this argument is incorrect - it follows that the argument is incorrect.

Move along ... nothing to see here ... as ever bloggers on the net making wild claims to get people along to their site ....

It seems to me that this supposed violation by VMware is in the same class as those by, say, nvidia distribution proprietary binary modules that link against the kernel. The controversy centres on the GPL/copyright notion of a 'derived work', and whether binary modules are 'derived' from the kernel because of their close symbiosis.

AFAIK, this interpretation of the GPL for kernel modules has yet to be tested in court. Opinions differ, but I can't see the linux world wanting to chuck out all the crucial existing proprietary binary modules in use. Apart from fedora, perhaps.

quote:

"Haff continues that legal issues aside, it would be at least manifestly unfair for such an issue to be realized at this late date when it's been sitting in more-or-less plain sight for years."

Yeah right, so they managed to *get away* with it for years, so now they go "come on we've always done it in (somewhat) full view, you can't sue us NOW!"

Laughable...

They may have a sensible policy of not allowing technical staff to be drawn into debates around legality and licensing and there is no need to reply to every blogger's speculation.

If there is something in the accusation then the FSF will presumably take it up in the courts. And if VMWare has been stupid enough to use the Linux kernel in a way that breaches the GPL then they will have to respect it.

If you're going to do something proprietary then don't touch GPL stuff. Go FreeBSD!

I think there are some misunderstandings flying about relating to the implications for VMWare. If ESX really qualifies as a derivative work of a GPL'd product, all that means is that VMWare ESX Server needs to be distributed under the GPL. The GPL does not forbid (in fact explicitly permits) "making money from other peoples' work" - the apparent problem with ESX server is that it is not covered by the GPL (copying, modification and redistribution being restricted is the real issue here).

So the fix is an apology from VMWare and the application of the GPL to the code from now on. This needn't take too big a chunk out of VMWare and I hope they realise that.

And this doesn't account for why MS avoided VMWare; Xen is GPL'd.

"Once again the Linux community takes aim at its own foot. "

So by preventing code theft by a company that is raking in millions the open source community is doing bad things?

Linux is published with a licence that prevents the code being abused, VMWare are abusing it and therefore they should stop doing so or comply with the licence.

Signed bands have released their song recordings in Garageband/Sony Acid for remix purposes before. While it's possible for me to just mixdown and record the output onto CD and sell it to everyone for profit the licence prevents me. How is this any different to what VMWare are doing?

VMWare could have used one of the BSDs which would allow them to withold the source code, but they chose Linux.

Let's face it, Linux isn't going to vanish anytime soon. There simply isn't another open source OS which has the vast levels of hardware support it has. Plus there are many embedded Linux experts out there now.

yes, and look at how popular gifs are these days?

"> Linux complains about Microsoft standards

Not sure what `Linux complains' means...

> ( when Linux really has no standards)

..anyway, trying to adhere to existing standards instead of calling any random crap `a new standard' really results in a very small number of `standards' created -- ideally only for the few things that are OS-specific by definition."

Linux follows broad standards, but it's all in the details. Why does software install in different locations when using different distrobutions? Why does it take ten years to get PDA syncing, music playback, and DVD burning working? Why do I have to install software to play videos on the web?

I use Linux everyday, and I use Windows everyday. Windows is a FAR better user experience. I'm not a programmer, as you may have noticed, but the reason that Mac OSX succeeded was because they focused on a polished product that was CONSISTENT and met the users needs. When a feature is added, it works, (bugs are bugs, so those aside), and if it isn't good enough they fix it. (Agreed that "Spaces" is a pretty funny take on multiple desktops)

I am sick of the Linux community's holier-than-though attitude. Until it works close to EVERY time for close to EVERY user, you got NOTHING!!

The first is that if the console uses linux, they have to provide the code for the modifications they did in the linux kernel to interface it with their software. This is what sony and ibm did with the cell. They put the cell spe support code into the free source code and only kept the non linux based drm code closed. Vmware should do the same. This would only allow other developers to write a new compatible console, not to replace the whole product.

Unless they used parts of the linux kernel in their virtualization system, which is a big error, since no company is allowed to steal the linux kernel and sell it as their own code. Knowing that even I borrow device driver code from linux from time to time, it wouldn't be suprising. The difference is that my projects are also gpl-ed and my company gives custormers the full source. (even non gpl-ed code, but only for review)

The fact is that it already happened that a company used linux on their own kit and sold it without telling the world what os they use and this is legally wrong, be it linux or priated windows.

It's understandable and sensible that a company would tell its developers and support staff to not engage in legal arguments on blogs.

However, this is has been an open question for some time now that goes back to the ESX 2.x days. If VMware is going to go public, make a BILLION dollars, and subject its shareholders to potential risk about their flagship product that brings in that BILLION dollars, that is indeed irresponsible and I'm more than a little surprised that no one from Wall Street is asking for a public response on this issue.

This could be considered a form of "cooking the books" if there's any merit behind it. The REASONABLE and RESPONSIBLE things to do is for VMware to respond so we can decide for ourselves. If there is indeed "nothing to see here," then GREAT! lets agree and move on. Not responding is the equivalent of Alberto Gonzales stating "he couldn't recall" a few dozen times in front of the Senate. No one's buying this.

As for the FSF, WTF? Why aren't you pressing this? What exactly do the FSF guys do other than go online and whine about OSS religion? This is a PERFECT EXAMPLE of what the FSF should be investigating.

We need to get to the bottom of this.

Ok, so after watching ESX boot *many* times, and dealing with the Service Console with some regularity, I agree that yes, the bare iron boots linux (a RHEL clone). ESX uses the cpci option on grub or lilo (depending on ESX vintage) to determine what hardware is 'owned or shared' by the service console vs. what's totally owned by the vmware kernel, which loads as a module and effectively takes ownership of the hardware that the service console cannot 'see'.

Should VMWare distribute source to first part? you bet, it's needed to comply with the GPL. Must VMWare release the source code for the vmkernel/hypervisor/etc. modules simply because linux launches it? I don't think so, but then again, I am unsure if it's a derivative or not. Binary modules are allowed under the GPL. I use more than one daily...

The service console in ESX 3 does not boot the vmkernel ( the hypervisor ). The vmkernel boots the system - and then brings up a virtual machine that contains linux that is called the service console. The service console is purely a place for people to put bits of software that do hardware monitoring, backup and so on.

If you actually read the article the author states the opposite - and is therefore 100% wrong - guaranteed - anyone who actualy has an ESX system knows this as they see it when ESX boots.

The vmkernel is not based on linux - if you dont believe me - download a trial copy from their website and try it for yourself

As some are taking issue with my quotes in the article, I wanted to expand and clarify a coupe of points.

1, First of all, based on what I know about how VMware works and the components that go into VMware, I have no reason to think that there's a GPL issue. Robin Turvey's comment is correct as far as I know. However, I'm not a lawyer and therefore didn't give a legal opinion in the article.

2. This is not a case of GPL'd code running under the covers in some router in a way that violates the license. Of course, under those circumstances the fact that "they've been doing it for years" doesn't change anything if such hidden use is subsequently discovered. But the structure of VMware ESX Server and the way it makes use of Linux (in the service console--roughly analogous to Dom0 in Xen) has never been a secret; it's discussed in any number of books--and even Wikipedia. As such, I would find it exceedingly strange if folks like the gpl-violations.org project would never (as far as I know) have raised as an issue such a high profile case--if there were anything to raise an issue about. I don't know for a fact who all has looked at VMware's use of Linux in the past, but it's really, really hard for me to believe that it hasn't been looked at and vetted by any number of knowledgeable people.

Thus, my comments. It's not that "getting away with it" provides amnesty. That's not my point. Rather, if there were a substantive issue, how could it not have been raised before now given how many far more obscure cases have been?

It has always been my understanding that the kernel was proprietary and only the console relied on Linux bits.

<quote>I am sick of the Linux community's holier-than-though attitude. Until it works close to EVERY time for close to EVERY user, you got NOTHING!!</quote>

Yeah Linux is nothing. Linux runs the World Wide Web and the Web is nothing too.

Yeah right!

No it doesn't. A large proportion of web servers run on Linux but that does not mean it 'runs' the www.

Linux is a great tool but it's just not as perfect as the evangelists seem to think it is.

As the gentlemen earlier mentions, the article mentions, the VMware advanced technical design guide shows, and the video attached the article shows, and as Zachary Amsden from VMware mentions:

vmkernel is started from the Linux kernel.

As does Robin, I invite the author to download a trial version of ESX. They'll find out that both sides of this debate are right, and she is wrong.

Nobody disputes that a Linux kernel is used to start vmkernel apart from Robin.

The question is, does the use of a Linux kernel module to load the vmkernel binary blob means that the binary blob must be licensed as GPL?

As mentioned in the article, Linus Torvalds personally considers such works as derivatives of Linux, unless they're probably not derivatives by having been ported from another platform.

Zachary Amsden has now replied to the article, and stated that he believes that since after the binary blob loads, it is no longer communicating with Linux, the fact that it requires Linux to start does not mean it is a derivative work.

This doesn't really make much sense - copyright infringement for a short time is still copyright infringement - but I invite you to read the full reply in the article comments are VentureCake and make up your own mind.

probably=provably

Ahem.