Slashdot Mirror
Linus Denounces NDISWrapper, Denies It GPL Status
eldavojohn writes "On message boards, Linus Torvalds was explaining why NDISWrapper is not eligible to be released under the GPL even though the project claims to be. Linus remarked, "Ndiswrapper itself is *not* compatible with the GPL. Trying to claim that ndiswrapper somehow itself is GPL'd even though it then loads modules that aren't is stupid and pointless. Clearly it just re-exports those GPLONLY functions to code that is *not* GPL'd." This all sprung up with someone restricted NDISWrapper's access to GPL-only symbols thereby breaking the utility. Linus merely replied that "If it loads non-GPL modules, it shouldn't be able to use GPLONLY symbols." As you may know, NDISWrapper implements Windows kernel API and then loads Windows binaries for a number of devices and runs them natively to avoid the cost and complication of emulation."
-- Linus, in this post
Thank you. I am an open source advocate, but the driver for my network card is a half-assed approach that doesn't connect to any access points, or do much else that can be called "useful." ndiswrapper is a bandage that can be used until the kernel team and third party module developers can produce something usable. Trying to get rid of it will only restrict Linux adoption.
Palm trees and 8
Summary is missing a HUGE portion of what actually happened. The discussion continued. After the discussion, Linus applied a patch to ALLOW access to GPL_ONLY symbols (for those who care, it's git commit 9b37ccfc637be27d9a652fcedc35e6e782c3aa78).
Look at the second entry from the top in the changelog:
http://kernel.org/pub/linux/kernel/v2.6/testing/ChangeLog-2.6.25-rc4
The battle is over, the discussion is at end and Linus has already signed off a change to restore Ndiswrapper functionality.
The summary doesn't imply that he's trying to get ndiswrapper out. This is how rumours start.
"Thanks for all the money you paid to us. We've used it to buy off ISO among other things" -Microsoft
The stance isn't as crazy as the context-free summary makes it out to be. Linus isn't talking about the license for the ndiswrapper code. He's talking about access to kernel functions which have been marked as "GPLONLY". These are functions which are intentionally not exported to non-GPL code. Linus is saying that allowing ndiswrapper to use them is equivalent to allowing calls from non-GPL windows binary drivers. Which is true.
The debate then is whether or not this should be considered a problem. The contributors who added many of the GPLONLY functions may have different opinions on the topic. Linus hints that the contributors for the USB functions would prefer a strict interpretation and deny ndiswrapper access to the GPLONLY kernel-level functions, because there is a perfectly good user-space API. But everyone involved agrees that ndiswrapper is will never live in user-space, because there's no programmer who would do it and it's a crazy idea anyway. Anyway you slice it, it's clear that ndiswrapper will get fixed one way or another, and nobody is accusing the ndiswrapper project of misusing the GPL.
In summmary, it's a tempest in a teapot: someone accidentally broke ndiswrapper, kernel API discussion ensues, Slashdot posts inflammatory summary, life goes on.
The ndiswrapper developers can release their code under any license they like, including the GPL; Linus has nothing to say about that. Furthermore, as long as Linux is under the GPL, Linus has no say over what I link into my kernel. If I want to link code under non-GPL compatible licenses into my kernel, that's my good right, under the GPL.
Linus possibly has a say over whether distributors can simultaneously distribute the Linux kernel and ndiswrapper as pre-packaged binaries. But even there, I don't see a problem: ndiswrapper itself is under the GPL and complies with the GPL. The fact that it allows end users to link code under non-GPL compliant licenses into the kernel doesn't change that.
While I think it would be nice if we didn't have to use ndiswrapper, and while one can argue either way about the desirability of its existence, now that it exists, Linus needs to honor the letter of the GPL and not try to redefine the terms after the fact. If he wants to, he can always relicense his code under different licenses in the future.
Userspace programs don't link against the kernel. Additionally, from http://www.kernel.org/pub/linux/kernel/COPYING:
Dewey, what part of this looks like authorities should be involved?
HUH??
No, the wording of the license and its interpretation by legally qualified people determines whether or not something is GPL compatible, not the whims and say-so of a person, be it Linus, RMS, or whomsoever.
Agree whole heartedly. If it weren't for NDIS, I wouldn't be typing this now. Pick another more deserving target, Linus.
Some mornings it's hardly worth chewing through the restraints to get out of bed.
Nobody's trying to get rid of it, read the numerous other posts correcting that assumption. This is just about the kernel losing GPLONLY status if you load ndiswrapper, which is important for debugging purposes and other things.
"When I first heard Daydream Nation it quite frankly scared the living shit out of me." -- Matthew Stearns
The nVidia driver is also not considered GPLONLY. Your kernel is considered 'tainted' if you use it. You will get no help or support from the kernel people if you have a kernel problem when your kernel is tainted.
Linus wants ndiswrapper to be in the same class. And he's right to. Maybe it's GPL, but it's whole purpose is to load stuff that isn't right into the kernel.
Need a Python, C++, Unix, Linux develop
How do I stop Mozilla Firefox from prompting me to install a plugin? (for Windows)
Open about:config, and set plugin.default_plugin_disabled to false. Then delete the file named npnul32.dll from your Mozilla Firefox plugins folder. You may have to enable showing hidden files to do this.
How do I stop Mozilla Firefox from prompting me to install a plugin? (for Linux)
1. Open about:config, and set plugin.default_plugin_disabled to false.
2. Delete the libnullplugin.so from your Mozilla Firefox plugins directory. You may have to do this as root if you do not have write access to your Mozilla Firefox installation from your user account.
"Oh drat these computers, they're so naughty and so complex. I could pinch them."
Marvin the Martian
NDIS wrapper has never been a great idea. It puts you at the mercy of Microsoft bugs and malice all for the benefit of a $30 network card. The kind of card that needs NDIS wrapper is usually worst of class and should be shunned. It's brain dead much like a winmodem and the "firmware" game is intentional. The card maker wants to be Windows only so don't buy it. Sooner or later hardware vendors will have to come around.
That's a little hard on just about any laptop with an AMD processor. Intel boards have Centrino, which works under Linux without trouble. AMD-based laptops...not so much.
"You can either have software quality or you can have pointer arithmetic, but you cannot have both at the same time."
it's quite important to understand the issue, i think. :) ) - nobody is _breaking_ anything.
:)
it has been repeated in the thread several times already, but i'll try again (from my, outsider viewpoint, but i'd hope somewhat educated one
short interpretation by me :
kernel has a variable which denotes that it is gpl only - that is, the core and all loaded modules are gpled.
this shows to people trying to debug things that they can debug everything and there are no binary modules that break shit in unexplainable ways.
now, if a binary module is loaded, kernel notes in the variable that it is no more gpl only and breakages can be extremely hard to debug and impossible to fix. i guess you'll agree we don't want the kernel devs to waste time on such cases.
now, ndiswrapper itself poses as gpl, thus it does not taint the kernel, but it then loads modules inside itself...
so you get a tainted kernel that does not identify as one.
and that is the only behaviour which is going to change.
if i have misunderstood things miserably, correct me, thanks
Rich
Trolltech, however, can grant exceptions to certain licenses, although I don't know which clause permits them to do so.
The fact that they own the copyright on the Qt code and can therefore license it however the hell they want permits them to do so.
General Relativity: Space-time tells matter where to go; Matter tells space-time what shape to be.
- MbM
Did you actually bother to read the entire thread? Linus is not anti-NDISwrapper. He is at worst ambivalent toward it. He is however unwilling to violate the will of the developer's who released their work under the GPL if they in fact feel that NDISwrapper is not legitimately a GPL module.
There is legitimate cause to believe that NDISwrapper cannot itself be licensed under the GPL if it links against non-GPL code. Since the GPLONLY flag defines symbols that are only exported to modules licensed under the GPL, this caused a problem. Linus was requiring that the owner of those symbols agree to NDISwrapper using them ( and preferably having them not defined as GPLONLY for consistency ). As the principal kernel god, he was right in flagging this problem.
Of course he and many others in the Linux community would prefer that linux native drivers existed for these devices, but anyone who has spent any time reading his comments would agree that Linus is a pragmatist ( he is an engineer ), not an idealist. He does have an obligation to enforce the license that all of the kernel developers are releasing their work under.
By the way, the end of the discussion seems to be Linus agreeing to roll back the change that broke the NDISwrapper. The hope is that if the change had been made intentionally to break NDISWrapper, then the submitter will resubmit the change and they discuss the reasons.
Read what I read again. I didn't say he was denying access to the driver. I said he was using his right to allow binary linking but denying the developers of the NDISWrapper code that same right. Sorry, poor choice of wording on my part.
This code is not giving binary-only drivers access to GPLONLY symbols. It is strictly providing an emulation layer that happens to require some of those symbols in order to work correctly. Those are two completely different things. About the only valid reason to complain would be if the NDISWrapper code didn't set the tainted flag. If it doesn't, that's a one line fix.
Check out my sci-fi/humor trilogy at PatriotsBooks.
Excerpted from Linus mail of 29 Feb:
What's confusing to slashdotters about this whole shebang is that there are two separate issues going on.
First, there is a technical/legal issue relating ndiswrapper's access to the Linux kernel (specifically, access to symbols marked GPLONLY). On this matter Linus is doing his job, which is to enforce existing policy for GPLONLY stuff. Workarounds had been discussed, including the possibility that the people who actually wrote the code (USB stuff mostly) agree to remove the GPLONLY restriction that *they* imposed. Linus is not opposed to the workarounds, but he won't brook discussion about bending enforcement of GPLONLY.
Secondly, Linus' expressed personal opinion about ndiswrapper (whose only purpose is to load Windows code) is complete indifference. He simply doesn't agree that because users depend heavily on ndiswrapper, he should go out of his way to bend the GPLONLY policy or make other special efforts. And he's not alone in the kernel community. Which freaks out the users who are afraid they won't be able to keep using their wireless cards and whatnot.
So people see these two issues fused together and think that Linus is killing off ndiswrapper by personal fiat.
NDISWrapper deals with Window's binary blob drivers.
That, under anyone's definition, means nothing GPLed can touch them.
NDISWrapper tried calling its self GPL while exposing all of Linux's GPLed interfaces to the binary blobs.
A very straight forward violation.
Personally I never liked NDISWrapper.
I use Linux to get away from Windows. I dont want their drivers running on my system.
Many people use it even when there are superior native drivers.
Its been portrayed as a quick fix so if your hardware doesnt work out of the box, just use NDISWrapper.
I was looking in to this recently and found lots of info at:
http://www.fsf.org/resources/hw/net/wireless/cards.html
No I am afraid that is not correct. The situation is a lot more complex as of Vista SP1:
http://thebackroomtech.wordpress.com/2007/11/05/howto-disabling-driver-signing-in-windows-vista-64-bit/
None of the workarounds you describe will work with x64 editions of Vista SP1, or for "boot drivers" in x86. I do plan to continue develop of the driver, but I doubt I will ever be able to get it signed using anything but the test certificate from MS. Still, will find out during testing...
Wow, it's an open source fundamentalist! Let's ban closed-open source marriages since they aren't pure and the most Holy of Programmers has spoken out against it.
Would you kindly mod me +1 insightful?
Since "you" is ambiguous referring to an AC, I will post my $.02 worth of searching.
FCC Rules on FOSS and Software-Defined Radio
Cognitive Radio Technologies and Software Defined Radios
As far as I can gather the main problem is that part of the licensing requirements is that "security measures" that need to be in place to prevent use of the device outside the specifications for which it is licensed.
With the boundary between driver software on the computer vs. firmware on the device shifting ever more away from the device, it becomes harder to implement these security measures.
The Hacker's Guide To The Kernel: Don't panic()!
I don't know how you got marked all the way up to 5, but what you say has never been tested in court - which is the only thing that matters. It is NOT repeat NOT clear whether linking constitutes creation of a derivative work. Providing the source and headers may well prove to constitute a published specification, and the right to interoperate is generally protected by US law (there's even an exception to the DMCA restrictions on reverse engineering specifically for the purpose of interoperability.) So really, you are quite wrong about the GP advocating an end to software copyrights in their comment. No such thing was stated or implied.
The question of whether any restrictions on linking are actually legal is a very real one.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
http://linux.slashdot.org/comments.pl?sid=476560&cid=22656000
by F452 (97091) Alter Relationship on 14:11 Wednesday 05 March 2008
Cheers to F452 for this information.
Yet Socrates himself is particularly missed.
A lovely little thinker but a bugger when he's pissed.