Vanishing Features Of The 2.6 Kernel

chromatic writes "Jerry Cooperstein has written an excellent article summarizing the features removed from the upcoming 2.6 kernel. One controversial change may be tightening restrictions on binary-only modules." And Lovechild writes with some more 2.6 news: "I recently did an inteview with famous kernel hacker extraordinare and all round nice guy Robert M. Love for Tinyminds.org, about kernel 2.6 and what can be expected for desktop Linux users, when the new kernel series is released.

Comments

They aren't gone!

[jlharris_50010]

Nothing's vanished... just not included... now you too can learn to kernel hack!

Arrogance

[rossz]

This may very well break some rather expensive commercial Linux products, but that doesn't seem to bother most kernel developers. Reminding the purveyors of binary modules that they continue to operate at the pleasure of the Linux kernel developers and their open-source licenses is seen to be a necessary (even enjoyable) task.

What an arrogant attitude. The kernel developers need to be reminded that widespread acceptance of Linux might very well require the support of big commercial enterprises, not just hobbiests and open source enthusiasts. Few companies are going to make major investments in Linux if they have to deal with prima donna programmers who love fucking with them.

Re:Arrogance

[pclminion]

What an arrogant attitude. The kernel developers need to be reminded that widespread acceptance of Linux might very well require the support of big commercial enterprises, not just hobbiests and open source enthusiasts

If you read over the LKML mailing list archives, it will become quite clear to you that the kernel developers don't give a cold shit in hell about user requests or complaints. They do what they want, and fuck everyone else.

At least, that was the sense I got from the conversation..

Re:Arrogance

[silvaran]

How about you go branch your own Linux kernel tree and add code back in that better supports binary-only modules? It's their tree, they can do what they want. If Red Hat doesn't like it, they can maintain their own tree (which they do). If nVidia doesn't like it, they can do likewise.

After all the hard work of the kernel folks, you seem to expect them to be perfectly happy with having to support binary modules that they can't debug, and that fall in a grey area of the GPL.

How about you go create your own open-source project. Then I'll go create some hardware, and keep all the information proprietary. To tease you a bit, I'll keep all information about my hardware proprietary, so you have two choices:

1) Reverse engineer my hardware and write your own software.
2) Indicate to me that your software is valuable, and depends on my hardware, until you convince me to release sources so they'll work better with your software.

That's the position nVidia's in. They've got one of the most widely-used video cards on the market. And the only way to use that video card effectively under Linux is to use their binary-only drivers. Get a clue. Linux doesn't belong to commercial entities any more than it belongs to Linus. What you do with your kernel tree is your own business.

Re:Arrogance

[Subcarrier]

What an arrogant attitude.

Arrogance indeed. Users are so used to being customers that they don't know when they are receiving a gift.

It's weird how, after over ten years, many people still don't get Linux. Every single Linux developer does what he does does because he gets something out of it. Not because he want's you to get something out of it. That's the beauty of the GPL. It guarantees that everyone can do what they want with the code, as long as they give others the same courtesy.

If you want something, just do it. If you can't, find a way to motivate someone to do it for you. You have the same rights as everyone else, no more, no less.

Re:Arrogance

[finkployd]

Perhaps, but they sure produce a great product with these attitudes...

Finkployd

It's open source, damnit!

[europrobe]

If non-GPL companies feel they can require users to install binary-only modules, why not simply requiring them to apply a kernel patch to remove this new limitation first?

Or, better still, why not delivering the whole product with an installer doing all this for them? It's not going to break GPL, as long as they publish the source code for the patch itself, which should be trivial.

I'm all for GPL, but this is not going to make that big an impact.

Re:Binary modules

[IamTheRealMike]

Why does the kernel module interface change so often though? I mean, surely there are only so many ways to write a mouse driver

Derivative works.

[OoSync]

I think the position the kernel developers are in is determining exactly what is a derivative work. In any other GPL piece of code, if you muddle with the inner workings of the code, then your code is a derivative work and must also be GPL if you release it. Well, according to this article, most modules fiddle with the inner workings of the kernel in a very intimate way. The logic follows that such code is a derivative work (i.e. it has to use kernel code to do such things) and must be GPL, which binary-only modules violate. This isn't about doing away with binary-only modules, its about clarifying the Linux kernel's license and enforcing it. Binary-only modules can still run, they are just restricted in what interfaces they can use to communicate with the kernel instead of being on intimate terms with GPL modules. Sounds fair. Furthermore, its the *legal* thing to do.

As other posters have already specified, you can distribute your own kernel or patch that doesn't enforce the GPL license, but in doing so you may indeed be violating the GPL yourself. Remember, the GPL is like any other license, you must abide by it or lose the privelidge of using the software.

Also, this does not break userspace (i.e. proprietary and binary-only applications), unless such software is dependent on a binary-only module.

I, for one, am curious to see how those people who really want to distribute binary modules will react. I think many have a market in Linux systems and will continue to provide their code. However, they may be well-served to develop a GPL module that provides very consistent interfaces for binary-only modules. The kernel developers don't want to do this, but if developers of binary-only modules develop it and apply it, well that's their business.

Drivers please

[visualight]

When a new product comes out on the market, the box almost always includes a cd with a windows driver on it. That driver is written and supported by the manufacturer of the product.

I love linux but I want so bad to be able to buy the latest neates thing and have it just work right away in linux and not a year later. Right now I have a radeon 9000 pro and a wintv pvr 250 I'm struggling with.

Drivers are the ONLY issue I have with linux and because of that I'm trying to learn C but I have so far to go and in the meantime I have to just be patient and wait for the vendor to produce a driver or for some linux developer go buy the same card I have and make it work. That's all fine and dandy except for this one thing: Some of these kernel developers are PAID to develop the linux kernel. That's right, it's their JOB. Not to show any disrespect to those kind hearted souls who sacrifice their spare time to do a community service, but as an enduser who has paid far far more for linux distro's than I ever would have paid if I was using Windows I have a right to insist on support for ME the end user. I have purchased Redhat 7.1, Mandrake 8.2,9.0 and every Suse since 6.2. It's my opinion that the kernel developer employed by the distro's have an obligation to develop the linux kernel with me in mind. And I want drivers. I don't care if they're GPL or not. I want hardware vendors to write the driver for the hardware they sell and distribute that driver WITH the hardware. And I want that driver to just work. That means the linux kernel must allow it to work. Politics and personal philosophy regarding open source have nothing to do with the discussion. I'm a paying customer and I should be treated that way.

Re:Binary modules

[vinsci]

No, they aren't. Palladium will NEVER stop you from running unsigned code. Never. Ever.

So what if Palladium lets your GPL program run, it wouldn't matter at all. This is because your GPL program will not be able, or rather allowed to, access the Palladium infrastructure in order to get the needed cryptographic credentials to perform certain tasks. So yes, your program will run, but no, it won't be able to do anything useful for you, not for anything that requires access to Palladium. And that would be mission completed for Microsoft: GPL software rendered useless.

The easily fooled will probably be glad to hear that the full source code to Palladium will likely be made available. Again, the source code is as powerless as your GPL program, and is of no use to you - even with the source, you can not make modifications to it and run the modified version in place of the version endorsed by Microsoft.

Furthermore, Microsoft already has a license in place that exludes open source development and specifically GPL and LGPL open source, see this article by Bruce Perens

Yes, there are people who are LYING about Palladium. Yes, LYING.

If you are aware of any lie in Ross Anderson's TCPA/Palladium FAQ, please state what it is.

Re:Binary modules

[Daniel Phillips]

It's exactly the same as if Microsoft changed the Windows license to say you could not run any GPL programs under the Windows operating system.

No, it's not. There is no restriction whatsoever in running any kind of application you want under Linux. Did you ignore the distinction between kernel and user mode intentionally?

It's more like Microsoft changing the Windows license to disallow any GPL code in the Windows kernel. Oh wait...

Re:Binary modules

[__aaahtg7394]

If you have 10k desktops and your IT guy/gal/team
1. doesn't have an automated system to push updates, or
2. decides you need to upgrade to a new kernel for no apparent reason (security updates are almost universally backported), or is unable to do the kernel patching to fix things themselves, or
3. you have software that's somehow hacked into the kernel on desktops but you don't have the source to,

you're fucked. Start writing your resignation letter, you totally blew it for anyone you're responsible to. Hope you weren't there too long so you can put it on your resume as a hiatus.

Linux doesn't have to worry about backwards ABI compatibility because most vendors use the API instead...

Re:Binary modules

[vinsci]

That document is a load of scare mongering [...]

...but it is a document which the inventor of TCPA himself, Bill Arbaugh, seem to agree with. He opens his comment on Ross Anderson with this statement:

"We are all aware of the criticisms that the TCPA has received. Ross Anderson did a good job of explaining the problems in an abstract fashion, but I felt that there were some things left out (Privacy concerns)." (my emphasis).

You write:

And I don't care if he uses words like "could" or "may" or any other wishy-washy term.

The rest of your comment, "Reality Master 101", builds on ignoring what Anderson is actually saying and is thus just a straw man argument, with which you have apparently fooled yourself.