Can't use SYSFS for "Proprietry" driver modules !!!.
Quoted 5 lines
So, the fact that someone else is doing something illegal, makes it acceptable for you to do the same thing? Please, talk to a lawyer about
this issue if you have _any_ questions.
How is what they are doing illegal? How it is even "bad"? They obviously can't give up their IP. Them providing binary modules wrapped in GPL glue (so anyone can fix most kernel incompatabilities) is a good thing for Linux. Many people and businesses would not be using Linux if they did not do that.
Aaron
Quoted 7 lines · Aaron Gyes
Quoted 4 lines
So, the fact that someone else is doing something illegal, makes it acceptable for you to do the same thing? Please, talk to a lawyer about
this issue if you have _any_ questions.How is what they are doing illegal? How it is even "bad"? They obviously can't give up their IP.
Many kernel developers feel that when people use their GPL'ed Linux code with the binary drivers they are giving up their IP to Nvidia. That argument is no less valid than Nvidia's.
Only a lawyer can answer the first question, and AFAIK there is not much precedent in the area so it's especially important to get legal advice. One would think Nvidia and ATI got the green light from their lawyers before releasing the binary drivers, so who knows.
Lee
Quoted 5 lines · Aaron Gyes
How is what they are doing illegal? How it is even "bad"? They obviously can't give up their IP. Them providing binary modules wrapped in GPL glue (so anyone can fix most kernel incompatabilities) is a good thing for Linux. Many people and businesses would not be using Linux if they did not do that.
I think that at the moment the general consensus is that it is ok to use the Linux kernel APIs (but not the EXPORT_SYMBOL_GPL ones) from binary modules _if_ _and_ _only_ _if_ the driver was originally written elsewhere and ported to the Linux kernel. Otherwise it's a derivative work, and must therefore be GPLed. Yes it's kinda draconian, but it's generally been for the betterment of the Open Source community.
BTW, to all you "But my drivers must be proprietary!" nerds out there, take a look at 3ware, Adaptec, etc. They have _great_ hardware and yet they release all of their drivers under the GPL. They get free updates to new kernel APIs too!
Cheers,
Kyle Moffett
-----BEGIN GEEK CODE BLOCK-----
Version: 3.12
GCM/CS/IT/U d- s++: a18 C++++>$ UB/L/X/*++++(+)>$ P+++(++++)>$ L++++(+++) E W++(+) N+++(++) o? K? w--- O? M++ V? PS+() PE+(-) Y+ PGP+++ t+(+++) 5 X R? tv-(--) b++++(++) DI+ D+ G e->++++$ h!*()>++$ r !y?(-) ------END GEEK CODE BLOCK------
Quoted 4 lines
BTW, to all you "But my drivers must be proprietary!" nerds out there, take a look at 3ware, Adaptec, etc. They have _great_ hardware and yet they release all of their drivers under the GPL. They get free updates to new kernel APIs too!
Well, it boils down to the full sourcecode. NVidia does only half. Other good examples besides 3ware are: VMware kernel modules and SUNWut/SRSS3 Linux Kernel modules.
Looks like there's only the GPU industry left that thinks somebody could "mis"use the kmod to make them (:one company) inferior on the market.
Jan Engelhardt
Signature · 1 line
No TOFU for me, please.
Quoted 11 lines · Jan Engelhardt
Quoted 4 lines
BTW, to all you "But my drivers must be proprietary!" nerds out there, take a look at 3ware, Adaptec, etc. They have _great_ hardware and yet they release all of their drivers under the GPL. They get free updates to new kernel APIs too!
Well, it boils down to the full sourcecode. NVidia does only half. Other good examples besides 3ware are: VMware kernel modules and SUNWut/SRSS3 Linux Kernel modules.
Looks like there's only the GPU industry left that thinks somebody could "mis"use the kmod to make them (:one company) inferior on the market.
Probably because it's true. What if everyone could notice that they are cheating such as rendering one inferior frame every other frame or things like this to increase performance ? Afterall, opensource also prevents you from cheating, or at least lets others fix your "bugs".
Willy
A message that is not in the archive (sent off-list or not yet received): [email protected]
Quoted 22 lines · Arjan van de Ven
Quoted 5 lines
I think that at the moment the general consensus is that it is ok to use the Linux kernel APIs (but not the
EXPORT_SYMBOL_GPLones) from binary modules _if_ _and_ _only_ _if_ the driver was originally written elsewhere
and ported to the Linux kernel.I disagree there. Only a few copyright holders of the kernel suggested that "previously written elsewhere" would be an exception. I haven't, Alan Cox has been very vocal about that he hasn't. I've not seen my employer say that it would make that exception either.
And it's a gray area.
Is it ok if you have 5000 lines from another OS and 0 specific/modified for linux (the technical impossibility of this aside)
Is 4990/10 still good?
is 4900/100 still good ?
is 4500/500 still good ?
is 4000/1000 still good ?
is 2500/2500 still good ?
is 2000/3000 still good ?
is 500/4500 still good ?
Quoted 7 lines
if anyone thinks this is a loophole their lawyers better have an answer for this...
(and note that I'm not claiming that those 4500 lines are a derived work when used elsewhere. But I do consider it a derived work if it's in a binary form where it does include linux specific code, and even code from the linux kernel via say inlines).
<flame type="Binary Driver Hatred">
NOTE: I *strongly* discourage binary drivers. They're crap and frustrate poor PowerPC users like me. Since this is purely a theoretical discussion, and I want to discourage binary crud, this email is Copyrighted:
This email is Copyright (C) 2005 Kyle Moffett.
The remainder of this email is available under the GNU General Public License, version 2. See http://www.gnu.org/licenses/gpl.txt for details. THE BELOW MAY NOT BE USED IN A BINARY DRIVER, SO DON'T EVEN THINK ABOUT IT! </flame>
Ok, so what if the _driver_ provides an API like this:
int start_driver(void);
int stop_driver(void);
void register_alloc(void *(*alloc)(unsigned long)); void register_free(void (*free)(void *));
[... more register functions here, generic functionality ...]
And a BSD licensed bit of glue based on that interface that connects the driver to a dozen different OSen by wrapping or using their interfaces directly.
Do you think _that's_ legal? As far as I can see, even that level is iffy, but it's a murky issue, and I doubt it will be decided one way or another until people test it in various courts.
Cheers,
Kyle Moffett
-----BEGIN GEEK CODE BLOCK-----
Version: 3.12
GCM/CS/IT/U d- s++: a18 C++++>$ UB/L/X/*++++(+)>$ P+++(++++)>$ L++++(+++) E W++(+) N+++(++) o? K? w--- O? M++ V? PS+() PE+(-) Y+ PGP+++ t+(+++) 5 X R? tv-(--) b++++(++) DI+ D+ G e->++++$ h!*()>++$ r !y?(-) ------END GEEK CODE BLOCK------
Quoted 10 lines
This email is Copyright (C) 2005 Kyle Moffett.
The remainder of this email is available under the GNU General Public License, version 2. See http://www.gnu.org/licenses/gpl.txt for details.
THE BELOW MAY NOT BE USED IN A BINARY DRIVER, SO DON'T EVEN THINK ABOUT IT!
</flame>
Quoted 8 lines
Ok, so what if the _driver_ provides an API like this:
int
start_driver(void);
intstop_driver(void);
voidregister_alloc(void *(*alloc)(unsigned long)); voidregister_free(void (*free)(void *));
[... more register functions here, generic functionality ...]
#GPL this e-mail my first C program,..
int main(int argc, char **argv)
{
}
damn every C program is a derived work.. it just means you need to get a better lawyer... more than likely American courts will be involved.. I suggest the Chewbacca defence[1] will work if you can pay enough money....
Dave.
Quoted 9 lines · Dave Airlie
#GPL this e-mail my first C program,..
intmain(int argc, char **argv)
{
}damn every C program is a derived work.. it just means you need to get a better lawyer... more than likely American courts will be involved.. I suggest the Chewbacca defence[1] will work if you can pay enough money....
Well, independently developed code is just that, independent. I don't care if somebody does something similar (aside from the fact that it pisses me off WRT binary drivers), I just don't want anybody to think my email is a clear way to dodge around the GPL, because it isn't :-D.
Cheers,
Kyle Moffett
-----BEGIN GEEK CODE BLOCK-----
Version: 3.12
GCM/CS/IT/U d- s++: a18 C++++>$ UB/L/X/*++++(+)>$ P+++(++++)>$ L++++(+++) E W++(+) N+++(++) o? K? w--- O? M++ V? PS+() PE+(-) Y+ PGP+++ t+(+++) 5 X R? tv-(--) b++++(++) DI+ D+ G e->++++$ h!*()>++$ r !y?(-) ------END GEEK CODE BLOCK------
Quoted 3 lines · * Kyle Moffett ([email protected]) wrote:
<flame type="Binary Driver Hatred">
NOTE: I *strongly* discourage binary drivers. They're crap and frustrate poor PowerPC users like me.
I mostly agree - there is one case where I think they *might* be acceptable; (and I think the original poster *may* fall into this category).
If you are making a very specialist piece of equipment; not the type of thing you can go and plug into any old PC; but say an entire box with some obscure piece of hardware in that no one would want to buy as a seperate add on. I just don't see the need to force someone to make drivers for this type of thing public.
Of course the poster could just go and use one of the BSDs which is probably his safest bet.
Dave
-----Open up your eyes, open up your mind, open up your code -------
/ Dr. David Alan Gilbert | Running GNU/Linux on Alpha,68K| Happy \ \ gro.gilbert @ treblig.org | MIPS,x86,ARM,SPARC,PPC & HPPA | In Hex / \ _________________________|_____ http://www.treblig.org |_______/
Quoted 10 lines · Dr. David Alan Gilbert
Quoted 3 lines · * Kyle Moffett ([email protected]) wrote:
<flame type="Binary Driver Hatred">
NOTE: I *strongly* discourage binary drivers. They're crap and frustrate poor PowerPC users like me.I mostly agree - there is one case where I think they *might* be acceptable; (and I think the original poster *may* fall into this category).
Quoted 7 lines
If you are making a very specialist piece of equipment; not the type of thing you can go and plug into any old PC; but say an entire box with some obscure piece of hardware in that no one would want to buy as a seperate add on. I just don't see the need to force someone to make drivers for this type of thing public.
...
And then the user want to upgrade the 2.0 kernel that shipped with this box although the company that made the hardware went bankrupt some years ago.
If the user has the source of the driver, he can port the driver or hire someone to port the driver (this "obscure piece of hardware" might also be an expensive piece of hardware).
Or if the driver is in the kernel sources, it might have even been ported.
Quoted 1 line
Dave
cu
Adrian
Signature · 4 lines
"Is there not promise of rain?" Ling Tan asked suddenly out
of the darkness. There had been need of rain for many days.
"Only a promise," Lao Er said.
Pearl S. Buck - Dragon SeedQuoted 7 lines
Quoted 2 lines
If you are making a very specialist piece of equipment; not
...If the user has the source of the driver, he can port the driver or hire someone to port the driver (this "obscure piece of hardware" might also be an expensive piece of hardware).
I am happy that nvidia (to name one) provides at least the source to its glue... here's a real world example:
http://groups-beta.google.com/group/linux.kernel/browse_frm/thread/52002e4f6d454e44/216311fff8f7c91b?rnum=1
could have been worse, but also could have been better, though.
Quoted 7 lines
And then the user want to upgrade the 2.0 kernel that shipped with this box although the company that made the hardware went bankrupt some years ago.
If the user has the source of the driver, he can port the driver or hire someone to port the driver (this "obscure piece of hardware" might also be an expensive piece of hardware).
So what? Sure, GPL'd drivers are easier for an end-user in that case. What does that have to do with law? What about what's better for the company that made the device? Should NVIDIA be forced to give up their secrets to all their competitors because some over zealous developers say so? Should the end-users of the current drivers be forced to lose out on features such as sysfs and udev compatability?
I love Linux, and a I love that free software has become mildly successful, but the zealots are hurting both.
Quoted 2 lines · Aaron Gyes
So what? Sure, GPL'd drivers are easier for an end-user in that case. What does that have to do with law?
Well, under most interpretations of the GPL, you are *NOT* allowed to even _link_ non-GPL code with GPL code. (Basically, by distributing such a linked binary, you are certifying that you have permission to GPL the entire source-code and are doing so.
Quoted 1 line
What about what's better for the company that made the device?
Who says that free maintenance and bugfixes *isn't* better for said company?
Quoted 2 lines
Should NVIDIA be forced to give up their secrets to all their competitors because some over zealous developers say so?
We don't care about their secrets, we just want to be able to interface with their hardware. Really, we don't care how the hardware does what it does internally, we just care how to tell it to do that. It's the difference between telling an artist to paint a big picture and watching every thought he makes while he does the painting with a brain scanner.
Quoted 2 lines
Should the end-users of the current drivers be forced to lose out on features such as sysfs and udev compatability?
Should the end-users even *have* features such as sysfs and udev? If the *open-source* developers hadn't *opened* their *source*, then that code wouldn't exist. One condition they made when they gave that code for free was that *only* people who also gave their code for free could use it.
Quoted 2 lines
I love Linux, and a I love that free software has become mildly successful, but the zealots are hurting both.
On the contrary, the zealots are what protect us from the even worse proprietary software zealots. You may not agree with them, but if there were only one kind of zealot then the world would be much worse off.
Cheers,
Kyle Moffett
-----BEGIN GEEK CODE BLOCK-----
Version: 3.12
GCM/CS/IT/U d- s++: a18 C++++>$ UB/L/X/*++++(+)>$ P+++(++++)>$ L++++(+++) E W++(+) N+++(++) o? K? w--- O? M++ V? PS+() PE+(-) Y+ PGP+++ t+(+++) 5 X R? tv-(--) b++++(++) DI+ D+ G e->++++$ h!*()>++$ r !y?(-) ------END GEEK CODE BLOCK------
Quoted 8 lines · Kyle Moffett
Quoted 2 lines · Aaron Gyes
So what? Sure, GPL'd drivers are easier for an end-user in that case. What does that have to do with law?
Well, under most interpretations of the GPL, you are *NOT* allowed to even _link_ non-GPL code with GPL code. (Basically, by distributing such a linked binary, you are certifying that you have permission to GPL the entire source-code and are doing so.
Wrong. You are free to do whatever you like in the privacy of your home, but not distribute the result. So you could very well distribute both pieces, one under GPL, the other not, and leave the linking to the end user.
Sure, /creating/ the piece to be linked with the GPLed code might make it GPL also, but that is another story.
Quoted 4 lines
Quoted 1 line
What about what's better for the company that made the device?
Who says that free maintenance and bugfixes *isn't* better for said company?
Not me. But it is not the only consideration involved.
Quoted 2 lines
Quoted 2 lines
Should NVIDIA be forced to give up their secrets to all their competitors because some over zealous developers say so?
nVidia doesn't want to tell, that is their decision to make. If they don't tell, Linux users don't get to use nVidia cards with OSS drivers. Both sides loose something, nVidia thinks (right or wrong) that what they loose this way is less than what they'd loose by opening up.
Quoted 6 lines
We don't care about their secrets, we just want to be able to interface with their hardware. Really, we don't care how the hardware does what it does internally, we just care how to tell it to do that. It's the difference between telling an artist to paint a big picture and watching every thought he makes while he does the painting with a brain scanner.
That the "Linux kernel community" (whoever that might be) doesn't care doesn't mean others wouldn't consider mining the Linux drivers for any data on their competition.
Quoted 8 lines
Quoted 2 lines
Should the end-users of the current drivers be forced to lose out on features such as sysfs and udev compatability?
Should the end-users even *have* features such as sysfs and udev? If the *open-source* developers hadn't *opened* their *source*, then that code wouldn't exist. One condition they made when they gave that code for free was that *only* people who also gave their code for free could use it.
Grey area... it could be argued that this is /public/ interfase to the kernel, and as such shouldn't be closed off.
Quoted 7 lines
Quoted 2 lines
I love Linux, and a I love that free software has become mildly successful, but the zealots are hurting both.
On the contrary, the zealots are what protect us from the even worse proprietary software zealots. You may not agree with them, but if there were only one kind of zealot then the world would be much worse off.
As long as the opposing zealots keep each other in check... ;-)
Signature · 4 lines
Dr. Horst H. von Brand User #22616 counter.li.org Departamento de Informatica Fono: +56 32 654431 Universidad Tecnica Federico Santa Maria +56 32 654239 Casilla 110-V, Valparaiso, Chile Fax: +56 32 797513
Quoted 5 lines
Quoted 3 lines
Quoted 3 lines
Should NVIDIA be forced to give up their secrets to all their competitors because some over zealous developers say so?
nVidia doesn't want to tell, that is their decision to make.
Well, they /could/, as to prove they are not cheating...(if they really don't)
Quoted 7 lines · Horst von Brand
Wrong. You are free to do whatever you like in the privacy of your home, but not distribute the result. So you could very well distribute both pieces, one under GPL, the other not, and leave the linking to the end user.
Sure, /creating/ the piece to be linked with the GPLed code might make it GPL also, but that is another story.
Actually this is an easy one. If you are the creator of the code, you can license it anyway you want. So you can make it both GPL and allow it to link with your code. Heck, put it under LGPL since GPL is allowed to link to that.
Anyway, I don't think that the GPL is that powerful to affect things not linked directly with it. Just like the MS license can't make you do certain things that were stated in the license, the GPL can't take too much control over what you do. If something in the license is reasonable, than it is easy to enforce (like taking the code from GPL source and using it in a binary) but if it starts to stretch (like controlling the code you write and how you can use it) then that will have to be fought in court, and will probably lose.
-- Stev
Quoted 24 lines · Steven Rostedt
Quoted 7 lines · Horst von Brand
Wrong. You are free to do whatever you like in the privacy of your home, but not distribute the result. So you could very well distribute both pieces, one under GPL, the other not, and leave the linking to the end user.
Sure, /creating/ the piece to be linked with the GPLed code might make it GPL also, but that is another story.
Actually this is an easy one. If you are the creator of the code, you can license it anyway you want. So you can make it both GPL and allow it to link with your code. Heck, put it under LGPL since GPL is allowed to link to that.
Anyway, I don't think that the GPL is that powerful to affect things not linked directly with it. Just like the MS license can't make you do certain things that were stated in the license, the GPL can't take too much control over what you do. If something in the license is reasonable, than it is easy to enforce (like taking the code from GPL source and using it in a binary) but if it starts to stretch (like controlling the code you write and how you can use it) then that will have to be fought in court, and will probably lose.
Quoted 1 line
-- Stev
Shims are used everywhere to interface with strange, incompatible, or otherwise difficult to interface operating systems. If Linux is too difficult or incompatible, use a shim.
In its simplest form, a shim might be a separate object containing only the GPL license, linked with the difficult-to-interface code. Don't jump on me yet. Read on.
In its most complicated form, the shim-file might contain some pointers, statically initialized to GPL-only symbols, and additional public symbols with which that your code interfaces.
In any event, you can do anything you want in the privacy of your own bedroom, in most all countries except, perhaps, Cuba. The problems come about when you try to sell a product, a portion of which was designed by somebody who wanted his works to be forever "free". The solution to this is to simply add the code that the "hold-out" prevented you from using, to your code. Certainly, given the complete source-code, you could come up with a perfect emulation that doesn't copy a single line. In fact, given the time, you could emulate all of Linux because you have the source-code available.
To me, the fact that you want to interface with symbols that the writer wanted to restrict, means that you want the writer to do your work and not get paid. Perhaps you are just too lazy to do your own thing and expect that something that already exists should surely be freely available so you can make money from it. Not so. I certainly am not allowed to pick my neighbor's flowers (GPL symbols) simply because they are where I can see them.
Of course there is the other side of the coin. I've seen persons who had nothing at all to do with the writing of some code, decide on their own, that its symbols are now GPL-only symbols. Funny how they seem to know the intention of some writer who wrote some code, left his mark, then traveled on.
This "GPL" junk will continue forever. There is no way to fight it. It's become a religion. If you want to write drivers for Linux, they really must contain the GPL License and you really need to make the source-code available. If your "trade secrets" are really that "secret", they won't be in a month or two, notwithstanding what the lawyers say.
Following the flock in this GPL issue insulates you from many future changes in the kernel. Major portions of the module code has already been rewritten to erect a solid barrier, marking what's in the kernel and what's without. What used to be done outside the kernel, the only reasonable place to do it, has now been moved inside the kernel for no other reason but isolation.
So tell your senior staff that you need to include the GPL license with your code. If you write good code, the chances of anybody outside your company actually reading it is near zero. If your "trade secrets" are so obvious that a look at the code will reveal them, you really need to get another job, your company will disappear in a month or two.
Cheers,
Dick Johnson
Penguin : Linux version 2.6.11 on an i686 machine (5537.79 BogoMips).
Notice : All mail here is now cached for review by Dictator Bush.
98.36% of all statistics are fiction.
Quoted 10 lines · linux-os
On Mon, 28 Mar 2005, Steven Rostedt wrote:
Following the flock in this GPL issue insulates you from many future changes in the kernel. Major portions of the module code has already been rewritten to erect a solid barrier, marking what's in the kernel and what's without. What used to be done outside the kernel, the only reasonable place to do it, has now been moved inside the kernel for no other reason but isolation.
Quoted 7 lines
So tell your senior staff that you need to include the GPL license with your code. If you write good code, the chances of anybody outside your company actually reading it is near zero. If your "trade secrets" are so obvious that a look at the code will reveal them, you really need to get another job, your company will disappear in a month or two.
I first started on this thread because of NVidia. Since I have many machines that use nvidia drivers and I suffer the consequences of this, but I'm a kernel programmer and can get around this too.
But now you hit on something that I'm fighting with. I may be part of the GPL religion, but I'm not a true believer. I like the concept. All code that I write on my own time is usually LGPL (otherwise it is just public use). I like to share and share alike. But the problem I have is that I don't have senior management. I'm a free lance programmer. I have companies (my customers) pay me to code. I can refuse to code if I don't agree with them, but then they get someone else and I go hungry. I strongly recommend to them that it is in their interest to release the code I write under GPL, but the managers don't see it. I may not be that strong of a spokesman, but they don't like to listen to me, the just tell me, do what we pay you to do. Like I said, I'm not a strong believer, so I won't risk not being able to feed my family for the FSF cause. So I just go on and code, and let their lawyers figure out what to do.
The customers I work for are actually more interested in selling their hardware than the software. But when they spend a lot of money to code for their hardware, they find it hard to understand that it is best to give it up as GPL code, especially when the workings of the hardware are explicit in the code. I usually have to also make changes to the kernel to handle the situation (which all goes under the GPL of course), so the modules I write are never expected to be used by the vanilla kernel, or by anyone elses kernel for that matter. The kernel is made to run on special hardware, and then have some special extensions put on that are in the form of modules. I'm still in the process of fighting to get these under GPL, but I'm not an employee, I'm a vendor. And the management sees me as such. It's very easy for them to pull the plug on me and find another approach to go.
These companies are rather large, and sell things for a niche market, that usually don't care about fighting for the GPL. This is not NVidia selling video cards to consumers. This is large companies selling larger systems to other large companies, and my part is just a small one. So, I don't think they'll disappear simply because they don't put everything under the GPL. They are smart enough to keep the extensions under GPL and allow me to send fixes if I find a bug with the code back to the maintainer. So the maintainer still benefits from this in the form of testing and updates.
The one way I do try to fight for the GPL is always make the imbedded code more efficient than the modules. So, to keep the code proprietary always has a impact on performance. This isn't hard to do, since obviously the code that is imbedded would be more efficient than code that needs to be called indirectly through hooks. Nothing has been decided yet, but if the benchmarks hold out, it all may be under GPL in the end anyway.
-- Steve
Quoted 4 lines
What used to be done outside the kernel, the only reasonable place to do it, has now been moved inside the kernel for no other reason but isolation.
I would not complain as much if nvidia was "more userspace" so that bug reports could be more valid than they are currently, when they are tainted.
Jan Engelhardt
Signature · 1 line
No TOFU for me, please.
A message that is not in the archive (sent off-list or not yet received): [email protected]
Quoted 11 lines · Arjan van de Ven
Quoted 3 lines
Anyway, I don't think that the GPL is that powerful to affect things not linked directly with it
the problem with kernel modules is.. that you actually create quite a few lines of code directly from the kernel (via the headers). Also.. derived work is both broader and smaller than "directly linked".
Like if you write an extra chapter to a harry potter novel... even if it's not in the same bundle, it's still a derived work.
If you don't use any of the names of the characters, is it still a derived work?
Having a GPL wrapper may be legal to do. You won't find out until you are actually taken to court. I don't see why people are very upset with doing this, since those that do must work very hard in keeping things compliant. And those that write the GPL code, can keep things hard for them, which could just be by ignoring them. Anyone who complains about a crash that has the nvidia module loaded will not get any help, except from those that also have the nvidia module.
Writing code that needs wrappers is not derived work, if that code can also have wrappers for BSD, QNX and perhaps Windows. It may be an added functionality, that needs some operating system, but if it is a separate functionality, than it should really work for any operating system, and thus it is not a derived work. I don't find nvidia modules a derived work from linux. Since they are used for other operating systems. And I believe you'll have a hard time convincing any court that nvidia is a derived work.
God! I must be in the middle. With management, I'm always fighting to go GPL, and here on LKML, I'm arguing for proprietary modules. I must be going psycho!
-- Steve
Quoted 2 lines
Writing code that needs wrappers is not derived work, if that code can also have wrappers for BSD, QNX and perhaps Windows.
Just because it works with another O.S. doesn't mean it is not derived from Linux code. Good grief. Try your lawyer, or at least a Google search for something like "copyright derived work", for a better idea of what "derived" means.
Signature · 3 lines
I won't rest till it's the best ...
Programmer, Linux Scalability
Paul Jackson <[email protected]> 1.650.933.1373, 1.925.600.0401Quoted 8 lines · Paul Jackson
Quoted 2 lines
Writing code that needs wrappers is not derived work, if that code can also have wrappers for BSD, QNX and perhaps Windows.
Just because it works with another
O.S. doesn't mean it is not derived from Linux code. Good grief. Try your lawyer, or at least a Google search for something like "copyright derived work", for a better idea of what "derived" means.
So you are saying that a stand alone section of code, that needs wrappers to work with Linux is a derived work of Linux? If there's some functionality, that you make, and it just happens to need some kind of operating system to work. Does that make it a derived work of any operating system?
OK, I took your advise and found this from googling:
http://www.pbwt.com/Attorney/files/ravicher_1.pdf
It's a good read (I recommend you to read it), and it goes back to one of my first points, and that is the interpretation of a derived work is basically up to the judge that is handling the case. There's nothing set in stone here. It covers mainly US copyright law, and that's what I'm mainly concerned with.
Unless you misunderstood me, and thought that I was talking about taking some part of Linux and making it work under another OS, I still stand by my statement.
-- Steve
Quoted 5 lines · Steven Rostedt
So you are saying that a stand alone section of code, that needs wrappers to work with Linux is a derived work of Linux? If there's some functionality, that you make, and it just happens to need some kind of operating system to work. Does that make it a derived work of any operating system?
It depends on how special and different the wrappers for Linux are from the wrappers for other operating systems. Like, for example, the sysfs stuff is so radically different from the APIs that other operating systems provide that anything using it is most likely copied from other in-kernel sysfs code, and is therefore derived from the Linux kernel.
Quoted 3 lines
OK, I took your advise and found this from googling:
Mmm, good reference, thanks!
Quoted 3 lines
Unless you misunderstood me, and thought that I was talking about taking some part of Linux and making it work under another OS, I still stand by my statement.
I think it really depends on the APIs implemented. Anything based on the sysfs code, even if only using the APIs, will probably be found to be a derivative work (NOTE: IANAL) because the sysfs API is so very different from everything else. Other interfaces like PCI management, memory management, etc, may not be so protectable, because they are standard across many systems. If Linux got a new and unique memory hotplug API, however, that might be a very different story. Similar things could be said about integration between drivers and the new Unified Driver Model, which appears to be quite original.
Cheers,
Kyle Moffett
-----BEGIN GEEK CODE BLOCK-----
Version: 3.12
GCM/CS/IT/U d- s++: a18 C++++>$ UB/L/X/*++++(+)>$ P+++(++++)>$ L++++(+++) E W++(+) N+++(++) o? K? w--- O? M++ V? PS+() PE+(-) Y+ PGP+++ t+(+++) 5 X R? tv-(--) b++++(++) DI+ D+ G e->++++$ h!*()>++$ r !y?(-) ------END GEEK CODE BLOCK------
Quoted 7 lines · Kyle Moffett
Quoted 5 lines · Steven Rostedt
So you are saying that a stand alone section of code, that needs wrappers to work with Linux is a derived work of Linux? If there's some functionality, that you make, and it just happens to need some kind of operating system to work. Does that make it a derived work of any operating system?
Quoted 7 lines
It depends on how special and different the wrappers for Linux are from the wrappers for other operating systems. Like, for example, the sysfs stuff is so radically different from the APIs that other operating systems provide that anything using it is most likely copied from other in-kernel sysfs code, and is therefore derived from the Linux kernel.
If your stand alone code has it's own API and your GPL wrapper handles the sysfs interface, then this might get around it.
Quoted 6 lines
Quoted 3 lines
OK, I took your advise and found this from googling:
Mmm, good reference, thanks!
You're welcome!
Quoted 4 lines
Quoted 3 lines
Unless you misunderstood me, and thought that I was talking about taking some part of Linux and making it work under another OS, I still stand by my statement.
Quoted 11 lines
I think it really depends on the APIs implemented. Anything based on the sysfs code, even if only using the APIs, will probably be found to be a derivative work (NOTE: IANAL) because the sysfs API is so very different from everything else. Other interfaces like PCI management, memory management, etc, may not be so protectable, because they are standard across many systems. If Linux got a new and unique memory hotplug API, however, that might be a very different story. Similar things could be said about integration between drivers and the new Unified Driver Model, which appears to be quite original.
Yes, but as the article states, ideas are not protected under copyright law. So an unique idea to handle hotplug then it may still not be covered. This is all very ambiguous, and is too confusing. I'll leave it up to the lawers!
-- Steve
Quoted 44 lines · Steven Rostedt
Quoted 7 lines · Kyle Moffett
Quoted 5 lines · Steven Rostedt
So you are saying that a stand alone section of code, that needs wrappers to work with Linux is a derived work of Linux? If there's some functionality, that you make, and it just happens to need some kind of operating system to work. Does that make it a derived work of any operating system?
Quoted 7 lines
It depends on how special and different the wrappers for Linux are from the wrappers for other operating systems. Like, for example, the sysfs stuff is so radically different from the APIs that other operating systems provide that anything using it is most likely copied from other in-kernel sysfs code, and is therefore derived from the Linux kernel.
If your stand alone code has it's own API and your GPL wrapper handles the sysfs interface, then this might get around it.
Quoted 6 lines
Quoted 3 lines
OK, I took your advise and found this from googling:
Mmm, good reference, thanks!
You're welcome!
Quoted 4 lines
Quoted 3 lines
Unless you misunderstood me, and thought that I was talking about taking some part of Linux and making it work under another OS, I still stand by my statement.
Quoted 11 lines
I think it really depends on the APIs implemented. Anything based on the sysfs code, even if only using the APIs, will probably be found to be a derivative work (NOTE: IANAL) because the sysfs API is so very different from everything else. Other interfaces like PCI management, memory management, etc, may not be so protectable, because they are standard across many systems. If Linux got a new and unique memory hotplug API, however, that might be a very different story. Similar things could be said about integration between drivers and the new Unified Driver Model, which appears to be quite original.
Quoted 6 lines
Yes, but as the article states, ideas are not protected under copyright law. So an unique idea to handle hotplug then it may still not be covered. This is all very ambiguous, and is too confusing. I'll leave it up to the lawers!
-- Steve
In the United States there is something called "restraint of trade". Suppose there was a long-time facility or API that got replaced with one that was highly restrictive. To use the new facility, one would have to buy a license or kiss somebody or something that was not previously required. If an action was brought against the person(s) who replaced the old facility with the new one, it is likely that the plaintiff would prevail.
If there is documented proof that those symbols were previously available and then they were changed to something more restrictive, I think one would prevail if a complaint were brought in court.
If course, you need to convince the person(s) who changed them that the action was unconscionable and therefore force them to change them back without making money for the lawyers by suing them. And, yes, somebody who modifies software in that manner can be sued. They could also be charged with criminal behavior (malicious mischief) in the State of Massachusetts or charged under federal law with restraint of trade. Modifying an existing policy to further an individual's ambitions can be fraught with consequences.
Cheers,
Dick Johnson
Penguin : Linux version 2.6.11 on an i686 machine (5537.79 BogoMips).
Notice : All mail here is now cached for review by Dictator Bush.
98.36% of all statistics are fiction.
Quoted 12 lines · linux-os
In the United States there is something called "restraint of trade". Suppose there was a long-time facility or API that got replaced with one that was highly restrictive. To use the new facility, one would have to buy a license or kiss somebody or something that was not previously required. If an action was brought against the person(s) who replaced the old facility with the new one, it is likely that the plaintiff would prevail.
If there is documented proof that those symbols were previously available and then they were changed to something more restrictive, I think one would prevail if a complaint were brought in court.
Quoted 10 lines
If course, you need to convince the person(s) who changed them that the action was unconscionable and therefore force them to change them back without making money for the lawyers by suing them. And, yes, somebody who modifies software in that manner can be sued. They could also be charged with criminal behavior (malicious mischief) in the State of Massachusetts or charged under federal law with restraint of trade. Modifying an existing policy to further an individual's ambitions can be fraught with consequences.
What the hell is this world coming to? Can we please just respect the
intent of the GPL and dispense with all this legalese crap; at least here on LKML? People who want a free ride and don't care about returning anything to the community that created Linux have many other forums in which to vent.
Regards,
Sean
Quoted 3 lines · linux-os
If there is documented proof that those symbols were previously available and then they were changed to something more restrictive, I think one would prevail if a complaint were brought in court.
They're still available. Just download an older version of Linux.
Signature · 3 lines
"Debugging is twice as hard as writing the code in the first place. Therefore, if you write the code as cleverly as possible, you are, by definition, not smart enough to debug it." - Brian W. Kernighan
Quoted 8 lines · Rik van Riel
Quoted 3 lines · linux-os
If there is documented proof that those symbols were previously available and then they were changed to something more restrictive, I think one would prevail if a complaint were brought in court.
They're still available. Just download an older version of Linux.
Yes. And this would show that whomever did that already violated the intent of the GPL by adding restrictions to use. NotGood(tm).
Cheers,
Dick Johnson
Penguin : Linux version 2.6.11 on an i686 machine (5537.79 BogoMips).
Notice : All mail here is now cached for review by Dictator Bush.
98.36% of all statistics are fiction.
Quoted 2 lines · linux-os
Yes. And this would show that whomever did that already violated the intent of the GPL by adding restrictions to use. NotGood(tm).
Dick,
You are so full of shit. There are no additonal restrictions, just the
restrictions of the GPL; period. Adding _GPL to the symbol does not
place any additional restriction on the people who are already bound by the GPL. You were much easier to endure when you were just pretending to have invented RLE.
Sean
Quoted 11 lines · Sean
Quoted 2 lines · linux-os
Yes. And this would show that whomever did that already violated the intent of the GPL by adding restrictions to use. NotGood(tm).
Dick,
You are so full of shit. There are no additonal restrictions, just the
restrictions of the GPL; period. Adding_GPLto the symbol does not
place any additional restriction on the people who are already bound by the GPL.
Sure it does. Before the GPL-only stuff the only problem one would have with a proprietary module, i.e., one that didn't contain the GPL "license" notice, was that the kernel would be marked "tainted". Everything would still work.
With the ADDITIONAL RESTRICTION added, the module won't even work because an ARTIFICIAL CONSTRAINT was added to prevent its use unless a GPL "license" notice existed.
Quoted 3 lines
You were much easier to endure when you were just pretending to have invented RLE.
No pretense, and the original implementation, used under CP/M for *.LBQ files, given to the world through my PROGRAM EXCHANGE BBS in 1980, long before there was any GPL, was a much better implementation than what became known as "RLL" when the disc- drive people got ahold of it. For those, not in their '60s, *.LBQ was the early precursor to Phil Katz's *.ZIP files. He wasn't able to take the heat of all the people who kept putting him down. He died an early death. I'm going to outlast all you nay-sayers because when you write facts, you don't have to remember what you previously wrote so the liars can't hurt you.
You see, with the original idea of free software, it was free. You were expected to leave any copyright notice alone. This free software idea was picked up by UC Berkeley once the Internet took hold.
The rest is history about a neat idea that became corrupted, first by the likes of Stallman, then by others when it became a religion.
Quoted 2 lines
Sean
Cheers,
Dick Johnson
Penguin : Linux version 2.6.11 on an i686 machine (5537.79 BogoMips).
Notice : All mail here is now cached for review by Dictator Bush.
98.36% of all statistics are fiction.
Quoted 10 lines · linux-os
Sure it does. Before the GPL-only stuff the only problem one would have with a proprietary module, i.e., one that didn't contain the GPL "license" notice, was that the kernel would be marked "tainted". Everything would still work.
With the ADDITIONAL RESTRICTION added, the module won't even work because an ARTIFICIAL CONSTRAINT was added to prevent its use unless a GPL "license" notice existed.
There are absolutely no additional restrictions for anyone that is in full compliance with the spirit and intent of the GPL. Full Stop.
Runtime restrictions do not fall under the GPL anyway, otherwise it would be illegal to impose _any_ security restrictions on a GPL system.
You are just DeadWrong(Tm) on this issue.
Sean
Quoted 4 lines · linux-os
Sure it does. Before the GPL-only stuff the only problem one would have with a proprietary module, i.e., one that didn't contain the GPL "license" notice, was that the kernel would be marked "tainted". Everything would still work.
Wait, you realize that the EXPORT_SYMBOL_GPL vs. EXPORT_SYMBOL bit doesn't mean a damn thing with respect to the license, right? The source code in entirety is licensed under the GPL. From my point of view, the _only_ difference between the uses of the two macros is that the "EXPORT_SYMBOL" macro states that the developer who last touched that code either:
o Has no plans to sue anyone who infringes on their code
o Thinks that the interface is so common and so generic that a lawsuit would be stupid.
Such a labeling is _only_ a bit of advice and a technical warning, nothing else. It can be removed or added without any issue, as long as you obey the GPL license. If somebody disagrees and sues you over it, that's a problem between you and your lawyers.
Quoted 3 lines
With the ADDITIONAL RESTRICTION added, the module won't even work because an ARTIFICIAL CONSTRAINT was added to prevent its use unless a GPL "license" notice existed.
Umm, according technically to many interpretations of the GPL, you can't link anyways unless you're GPL, so such labels are just advice, and have no major legal bearing.
Quoted 1 line
[... excess flames snipped ...]
Please cut the worthless rhetoric and irrelevant statements and either discuss useful information or stop posting.
Cheers,
Kyle Moffett
-----BEGIN GEEK CODE BLOCK-----
Version: 3.12
GCM/CS/IT/U d- s++: a18 C++++>$ UB/L/X/*++++(+)>$ P+++(++++)>$ L++++(+++) E W++(+) N+++(++) o? K? w--- O? M++ V? PS+() PE+(-) Y+ PGP+++ t+(+++) 5 X R? tv-(--) b++++(++) DI+ D+ G e->++++$ h!*()>++$ r !y?(-) ------END GEEK CODE BLOCK------
On Tue, Mar 29, 2005 at 07:15:01AM -0500, linux-os wrote [snip]
Quoted 9 lines
In the United States there is something called "restraint of trade". Suppose there was a long-time facility or API that got replaced with one that was highly restrictive. To use the new facility, one would have to buy a license or kiss somebody or something that was not previously required. If an action was brought against the person(s) who replaced the old facility with the new one, it is likely that the plaintiff would prevail.
[snip]
The key phrase here is 'restraint of /trade/'.
Noone has been selling these drivers, and nonone's been dealing in the right to use or write them. Without being a lawyer (US or otherwise), I'd suggest that it's highly unlikely any such legislation would apply.
E.g.: suppose there are 2 snack bars within 100 yards of a school; one is out of sight, across an intersection and down a side street, and one is clearly visible across an empty lot. For years the lot has been unfenced and, human nature being what it is, kids just walk across the open lot. The owner of the lot then decides to put up a high fence around it with a combination lock on the gate (now he's raising chinchillas, or peaches; he won't say) so all the kids start going to the other snackbar, except for a few that he trusts with the combination. It seems to me you're suggesting that the snackbar owner who's lost out would have an action for restraint of trade; I can't see it myself.
John.
[snip]
Quoted 15 lines
Cheers,
Dick Johnson
Penguin : Linux version 2.6.11 on an i686 machine (5537.79 BogoMips).
Notice : All mail here is now cached for review by Dictator Bush.
98.36% of all statistics are fiction.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/------------------------------
Signature · 20 lines
Voice: +61 8 8202 9040 Email: [email protected] Oasis Systems Pty Ltd 288 Glen Osmond Road Fullarton, South Australia 5063 Ph: + 61 8 82029000 Fax: +61 8 82029001 CAUTION: This email and any attachments may contain information that is confidential and subject to copyright. If you are not the intended recipient, you must not read, use, disseminate, distribute or copy this email or any attachments. If you have received this email in error, please notify the sender immediately by reply email and erase this email and any attachments. DISCLAIMER: OASIS Systems uses virus-scanning technology but accepts no responsibility for loss or damage arising from the use of the information transmitted by this email including damage from virus.
Quoted 11 lines · John Pearson
E.g.: suppose there are 2 snack bars within 100 yards of a school; one is out of sight, across an intersection and down a side street, and one is clearly visible across an empty lot. For years the lot has been unfenced and, human nature being what it is, kids just walk across the open lot. The owner of the lot then decides to put up a high fence around it with a combination lock on the gate (now he's raising chinchillas, or peaches; he won't say) so all the kids start going to the other snackbar, except for a few that he trusts with the combination. It seems to me you're suggesting that the snackbar owner who's lost out would have an action for restraint of trade; I can't see it myself.
Well in Austria there is a law: if you walk through an area that is not public and do this for a very long time years and suddenly the owner stops you from doing this, you could sue him for stopping you doing a usual thing.
But real life issues and software laws are more than two kind of shoes. They are two kind of universes.
Right and Code changes always happens, some approve i some not. And there will be always people not like it.
Fact is the kernel is a GPL thing so logically the coders want to keep it GPL,
lg, clemens
Kyle Moffett <[email protected]> said:
[...]
Quoted 10 lines
I think it really depends on the APIs implemented. Anything based on the sysfs code, even if only using the APIs, will probably be found to be a derivative work (NOTE: IANAL) because the sysfs API is so very different from everything else. Other interfaces like PCI management, memory management, etc, may not be so protectable, because they are standard across many systems. If Linux got a new and unique memory hotplug API, however, that might be a very different story. Similar things could be said about integration between drivers and the new Unified Driver Model, which appears to be quite original.
Sorry, but an /interfase/ is there to do exactly that. It can be placed under copyright protection as code, but /using/ it just can't be considered a derived work. It makes no sense that if I get a description (docu, example code, whatever) and learn from that how it is used, and then go and write my own, that my code it should be a derived work of what is at the other side of the interfase.
If this was true, Linux would be a ripoff from Unix (same system calls), and any program ever written for Unix would belong to Novell (or SCOXE, or UCB, depending on whom you believe in the current mess). Note that this is quite similar to the gargabe SCOXE tried claiming against Linux/IBM, and had to take back.
Signature · 4 lines
Dr. Horst H. von Brand User #22616 counter.li.org Departamento de Informatica Fono: +56 32 654431 Universidad Tecnica Federico Santa Maria +56 32 654239 Casilla 110-V, Valparaiso, Chile Fax: +56 32 797513
Quoted 6 lines · Horst von Brand
Sorry, but an /interfase/ is there to do exactly that. It can be placed under copyright protection as code, but /using/ it just can't be considered a derived work. It makes no sense that if I get a description (docu, example code, whatever) and learn from that how it is used, and then go and write my own, that my code it should be a derived work of what is at the other side of the interfase.
IANAL, but I've seen two extremes in the interpretation of "derived work" w.r.t the GPL:
1- If there isn't GPL code in your sources, it's not a derived
work. That interpretation would make the GPL and the LGPL pretty much equivalent.
2- If you depend on GPL code in any way to make your code useful,
then it's a derived work.
The truth is, as always, somewhere in the middle, and intentions, local law and jurisprudence matter a lot when a judge states on a case like that, whichever your country is. That's why the "check with a lawyer in your jurisdiction" is the only appropriate answer.
The FSF stance btw is a little less harsh than 2. It's pretty much, from what I understand it, "if you depend on gpl code, and there is no non-gpl alternative which means you really, really depend on gpl code no matter what, then it's a derived work.". Makes a lot of sense too. Whether it holds water is a matter of the court. It _is_ the intent behind the GPL though, they wrote the GPL and said so numerous times, so it will have its importance if someone puts that part of the GPL to the test.
You'll notice that I never talk of interfaces in all that. Interfaces is a technical matter, and maybe an intent matter[1]. Derived work status is a _legal_ matter which puts in play dependency and intent. Not a technical matter at all.
OG.
[1] Compare for intent differences:
- "Linux has a sysfs interface, check the sources for how it works"
- "Here is the sysfs interface documentation for use in the GPL-ed Linux kernel"
- "Here is the sysfs interface documentation, an implementation-neutral interface for <whatever>, which happens to be implemented in, among others, the Linux kernel".
Quoted 3 lines · Olivier Galibert
It _is_ the intent behind the GPL though, they wrote the GPL and said so numerous times, so it will have its importance if someone puts that part of the GPL to the test.
Note that for judges in many countries, intent matters.
Intent might not overrule the language in the license, but it is often used to clarify things and to guide the way in which the license is interpreted.
Signature · 3 lines
"Debugging is twice as hard as writing the code in the first place. Therefore, if you write the code as cleverly as possible, you are, by definition, not smart enough to debug it." - Brian W. Kernighan
Quoted 2 lines
So you are saying that a stand alone section of code, that needs wrappers to work with Linux is a derived work of Linux?
Not what I said.
Signature · 3 lines
I won't rest till it's the best ...
Programmer, Linux Scalability
Paul Jackson <[email protected]> 1.650.933.1373, 1.925.600.0401Quoted 5 lines · Paul Jackson
Quoted 2 lines
So you are saying that a stand alone section of code, that needs wrappers to work with Linux is a derived work of Linux?
Not what I said.
Good, so you most likely misunderstood me.
OK, I've had enough of being devil's advocate. I'm usually fighting to get things under the GPL, or at least the LGPL, and I'm getting worn out fighting for the other side.
Have a good one, I have work to do ;-)
-- Steve
Quoted 1 line · Steven Rostedt
On Sun, 2005-03-27 at 21:54 -0400, Horst von Brand wrote:
[...]
Quoted 5 lines
Quoted 5 lines
Wrong. You are free to do whatever you like in the privacy of your home, but not distribute the result. So you could very well distribute both pieces, one under GPL, the other not, and leave the linking to the end user.
Quoted 7 lines
Quoted 2 lines
Sure, /creating/ the piece to be linked with the GPLed code might make it GPL also, but that is another story.
Actually this is an easy one. If you are the creator of the code, you can license it anyway you want. So you can make it both GPL and allow it to link with your code. Heck, put it under LGPL since GPL is allowed to link to that.
Right, but that is not the point. If you need to look closely at the GPL part, or swipe stuff from other GPLed code that uses the same interface, the result might have to be GPL.
Quoted 8 lines
Anyway, I don't think that the GPL is that powerful to affect things not linked directly with it. Just like the MS license can't make you do certain things that were stated in the license, the GPL can't take too much control over what you do. If something in the license is reasonable, than it is easy to enforce (like taking the code from GPL source and using it in a binary) but if it starts to stretch (like controlling the code you write and how you can use it) then that will have to be fought in court, and will probably lose.
The problem is that noone is sure what exactly qualifies as "derived work", and what can be used/copied without concern for infringement (Can you use declarations of stuff in a public .h? What if the .h is "local use" in the code? Can you write your own declarations matching what you find in random code, and use them? What about inlined functions? Lists of e.g. IOCTL codes or system calls?). Does it matter that you are writing your stuff for compatibility with what is out there, or to add functionality? Does it matter if you are just using "standard" interfases to the GPLed stuff (whatever that means in each case)? How much copying makes it derived? 1 line, 5%, what?
Signature · 4 lines
Dr. Horst H. von Brand User #22616 counter.li.org Departamento de Informatica Fono: +56 32 654431 Universidad Tecnica Federico Santa Maria +56 32 654239 Casilla 110-V, Valparaiso, Chile Fax: +56 32 797513
Quoted 10 lines · Aaron Gyes
Quoted 7 lines
And then the user want to upgrade the 2.0 kernel that shipped with this box although the company that made the hardware went bankrupt some years ago.
If the user has the source of the driver, he can port the driver or hire someone to port the driver (this "obscure piece of hardware" might also be an expensive piece of hardware).
So what? Sure, GPL'd drivers are easier for an end-user in that case. What does that have to do with law? What about what's better for the
I wasn't talking about legal issues.
I was answering Dave's "very specialist piece of equipment" opinion.
And my point was that even in this case it's better for the user to have the source.
Quoted 7 lines
company that made the device? Should NVIDIA be forced to give up their secrets to all their competitors because some over zealous developers say so? Should the end-users of the current drivers be forced to lose out on features such as sysfs and udev compatability?
I love Linux, and a I love that free software has become mildly successful, but the zealots are hurting both.
There are many bug reports to linux-kernel that are undebuggable because they involve the nvidia binary-only module.
I do personally know several people who use the nvidia binary-only modules and have as a result experienced stability problems on their computer. Linux has an IMHO justified reputation for being stable. Users experiencing system stability problems due to binary-only modules might wrongfully attribute them to Linux harming the reputation of Linux.
cu
Adrian
Signature · 4 lines
"Is there not promise of rain?" Ling Tan asked suddenly out
of the darkness. There had been need of rain for many days.
"Only a promise," Lao Er said.
Pearl S. Buck - Dragon SeedQuoted 3 lines · Aaron Gyes
company that made the device? Should NVIDIA be forced to give up their secrets to all their competitors because some over zealous developers say so? Should the end-users of the current drivers be forced to lose
Is not just about zealotism, propietary drivers have to deal with locking, multiple architectures and so on, and often they're not very good at that.
Note that this is not a problem only with linux, Windows is suffering the same every day. Here: http://blogs.msdn.com/oldnewthing/archive/2004/03/05/84469.aspx
is a good example of how *scary* can be. This is what happens in a OS where a certification process is needed. I don't want to know what are they doing in linux, where such certification doesn't exist.....
Quoted 8 lines · Aaron Gyes
Quoted 7 lines
And then the user want to upgrade the 2.0 kernel that shipped with this box although the company that made the hardware went bankrupt some years ago.
If the user has the source of the driver, he can port the driver or hire someone to port the driver (this "obscure piece of hardware" might also be an expensive piece of hardware).
Quoted 6 lines
So what? Sure, GPL'd drivers are easier for an end-user in that case. What does that have to do with law? What about what's better for the company that made the device? Should NVIDIA be forced to give up their secrets to all their competitors because some over zealous developers say so? Should the end-users of the current drivers be forced to lose out on features such as sysfs and udev compatability?
It's not zealotry, it's called being compliant with the license of the kernel. It's as simple as that.
If you ignore the license, you will suffer the consequences of it, just like if you ignore the license of any closed source chunk of software. Would you expect the owner of that software to turn a blind eye toward violators?
greg k-h
Quoted 4 lines · Aaron Gyes
Quoted 4 lines
And then the user want to upgrade the 2.0 kernel that shipped with this box although the company that made the hardware went bankrupt some years ago.
Quoted 8 lines
Quoted 3 lines
If the user has the source of the driver, he can port the driver or hire someone to port the driver (this "obscure piece of hardware" might also be an expensive piece of hardware).
So what? Sure, GPL'd drivers are easier for an end-user in that case. What does that have to do with law? What about what's better for the company that made the device? Should NVIDIA be forced to give up their secrets to all their competitors because some over zealous developers
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Quoted 1 line
say so? Should the end-users of the current drivers be forced to lose
^^^^^^^
Quoted 1 line
out on features such as sysfs and udev compatability?
Because otherwise they are violating someone else's copyright?
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]
In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
Quoted 8 lines · Kyle Moffett
Quoted 6 lines · Aaron Gyes
How is what they are doing illegal? How it is even "bad"? They obviously
can't give up their IP. Them providing binary modules wrapped in GPL glue (so anyone can fix most kernel incompatabilities) is a good thing for Linux. Many people and businesses would not be using Linux if they did not do that.
Quoted 6 lines
I think that at the moment the general consensus is that it is ok to use the Linux kernel APIs (but not the
EXPORT_SYMBOL_GPLones) from binary modules _if_ _and_ _only_ _if_ the driver was originally written elsewhere and ported to the Linux kernel. Otherwise it's a derivative work, and must therefore be GPLed. Yes it's kinda draconian, but it's generally been for the betterment of the Open Source community.
No, that is not the general consensus at all. Please search the archives and the web for summaries of this discussion topic the last time it came up.
greg k-h
Quoted 5 lines · Greg KH
No, that is not the general consensus at all. Please search the archives and the web for summaries of this discussion topic the last time it came up.
greg k-h
Hi. I've searched the archives about this stuff. It looks like you attempted to change the EXPORT_SYMBOL's to EXPORT_SYMBOL_GPL for sysfs stuff back in February, and the issue in general has come up many times. A few people have made the point that Linus has said that changing EXPORT_SYMBOL to EXPORT_SYMBOL_GPL is not okay, that they are supposed to start off as EXPORT_SYMBOL_GPL or somesuch. And it seems last time you tried it, something made you change it back.
Can you explain this, please?
Quoted 10 lines · Aaron Gyes
Quoted 5 lines · Greg KH
No, that is not the general consensus at all. Please search the archives and the web for summaries of this discussion topic the last time it came up.
greg k-h
Hi. I've searched the archives about this stuff. It looks like you attempted to change the EXPORT_SYMBOL's to
EXPORT_SYMBOL_GPLfor sysfs stuff back in February, and the issue in general has come up many times.
I only tried to change the class_simple code at that time. This was because the entire other driver core code was converted by the author of that code at the same time. The reason people complained about the class_simple code was because of vmware and nvidia.
Since then, vmware has stated that they don't even like sysfs and udev, and dropped support for it entirely. Also, nvidia has told me that they do not like udev either, so they don't care about the sysfs code too.
Quoted 4 lines
A few people have made the point that Linus has said that changing
EXPORT_SYMBOLtoEXPORT_SYMBOL_GPLis not okay, that they are supposed to start off asEXPORT_SYMBOL_GPLor somesuch. And it seems last time you tried it, something made you change it back.
Yes, I was asked nicely to do so by some people I respect. Since then, I sought out the only closed source users of this code that I could find, and consulted them for what to do. As mentioned above, neither company has a problem with me doing this.
Also, the code has undergone a rewrite, fixing many issues, and changing the way things work to tie more closely into the main driver core code. As such, the class_simple code is now just gone, there is no such need for it. And as such, the new code contains the _GPL markings, as I do not think that _anyone_ can try to claim that their code would not be a derived work of Linux who wants to use it (as no other OS has such a driver model interface.)
Quoted 1 line
Can you explain this, please?
I hope the above explanation is acceptable. If you have further questions, please do not hesitate to ask. And I would personally like to thank you for your civil tone. My current inbox reflects the rants of people without such civility at this time.
thanks,
greg k-h
Quoted 7 lines · Greg KH
Also, the code has undergone a rewrite, fixing many issues, and changing the way things work to tie more closely into the main driver core code. As such, the
class_simplecode is now just gone, there is no such need for it. And as such, the new code contains the_GPLmarkings, as I do not think that _anyone_ can try to claim that their code would not be a derived work of Linux who wants to use it (as no other OS has such a driver model interface.)
That does not really make sense, as the driver model code could be used for ndiswrapper, for example. That would not make the Windows net drivers derived code of the Linux kernel. ndiswrapper, yes it would be. Binary driver blobs, no.
ndiswrapper is a perfect example, in fact. It is GPL, and implements an _interface_ to binary code that is not GPL.
It seems to me that any author has the right to create a public interface into the kernel.
If that interface is well-defined and public, implementing it from the other side in binary code does not create a derived work. It is especially obvious if there are multiple interface implementations. Hard to argue that the same binary that links unchanged into Windows, BSD and Linux is Linux-derived.
Signature · 1 line
Zan Lynx <[email protected]>
Quoted 9 lines · Zan Lynx
Quoted 7 lines · Greg KH
Also, the code has undergone a rewrite, fixing many issues, and changing the way things work to tie more closely into the main driver core code. As such, the
class_simplecode is now just gone, there is no such need for it. And as such, the new code contains the_GPLmarkings, as I do not think that _anyone_ can try to claim that their code would not be a derived work of Linux who wants to use it (as no other OS has such a driver model interface.)
Quoted 7 lines
That does not really make sense, as the driver model code could be used for ndiswrapper, for example. That would not make the Windows net drivers derived code of the Linux kernel. ndiswrapper, yes it would be. Binary driver blobs, no.
ndiswrapper is a perfect example, in fact. It is GPL, and implements an _interface_ to binary code that is not GPL.
And do your lawyers deem ndiswrapper as something that is legal under the GPL? The ones I have talked to definitely do not feel that way.
Again, why are we, non-lawyers arguing about this. If you work for a company that deals with Linux kernel issues, and you have any questions about the legality of _anything_, get a legal opinion. Don't rely on lkml for this.
greg k-h
Quoted 5 lines · Zan Lynx
That does not really make sense, as the driver model code could be used for ndiswrapper, for example. That would not make the Windows net drivers derived code of the Linux kernel. ndiswrapper, yes it would be. Binary driver blobs, no.
The Windows net drivers are not (we believe) compiled using GPL'd header files and their included macros, asm code, etc..
Probably all Linux binary drivers *are* compiled using GPL'd header files, and thus are themselves subject to the GPL.
Cheers
Quoted 2 lines · Mark
Probably all Linux binary drivers *are* compiled using GPL'd header files, and thus are themselves subject to the GPL.
I doubt that there is a consensus that simply compiling something with a GPL header necessarily and always subjects it to the GPL. See your lawyer.
Signature · 3 lines
I won't rest till it's the best ...
Programmer, Linux Scalability
Paul Jackson <[email protected]> 1.650.933.1373, 1.925.600.0401Quoted 6 lines · Paul Jackson
Quoted 2 lines · Mark
Probably all Linux binary drivers *are* compiled using GPL'd header files, and thus are themselves subject to the GPL.
I doubt that there is a consensus that simply compiling something with a GPL header necessarily and always subjects it to the GPL. See your lawyer.
For a header as in interface maybe not. For headers containing significant code in inline functions the binary drivers is definitly a derived work.
Quoted 4 lines · Greg KH
I hope the above explanation is acceptable. If you have further questions, please do not hesitate to ask. And I would personally like to thank you for your civil tone. My current inbox reflects the rants of people without such civility at this time.
Beyond acceptable, thank you. I'll admit I started off (other posts in this thread) somewhat upset ("That jerk broke the udev support in the drivers I use!") But I've been somewhat swayed by what I've read. And I guess if a license say something, it doesn't really matter how inconvienient it is for me. (Though "derived work" is still pretty vague and I eagerly await a lawsuite that will disambiguate it.
In other news: How do I get udev to create a static node?
Aaron Gyes
Quoted 2 lines · Aaron Gyes
In other news: How do I get udev to create a static node?
What do you mean by "static"? Something that persists over a reboot? Or after the device is removed?
If reboot, mount your /dev on a disk-backed filesystem, not a ramfs or tmpfs.
If after removed, that's not what udev is set up to do, sorry.
thanks,
greg k-h
Quoted 1 line · Greg KH
If after removed, that's not what udev is set up to do, sorry.
There's no way to either a) Hack udev.conf to always create a node with a certain major and minor or b) A way to make sysfs trick udev?
I'll kind of need to do this for nvidia and any other modules affected by this change, or else switch back to the inferior devfs.
Aaron Gyes
Quoted 5 lines · Aaron Gyes
Quoted 1 line · Greg KH
If after removed, that's not what udev is set up to do, sorry.
There's no way to either a) Hack udev.conf to always create a node with a certain major and minor
No.
Quoted 1 line
or b) A way to make sysfs trick udev?
From userspace? No.
Quoted 2 lines
I'll kind of need to do this for nvidia and any other modules affected by this change, or else switch back to the inferior devfs.
I know debian provides a way for the udev startup script to create any static device node that you want to have be created. I suggest you look at that as an example to do what you wish.
Also, vmware already supports using udev, yet it does not use sysfs. It does so by creating the device nodes in its startup script. There is no reason why you can't do the same thing in the nvidia driver.
thanks,
greg k-h
Quoted 2 lines · Greg KH
What do you mean by "static"? Something that persists over a reboot? Or after the device is removed?
Forgot to clarify. Create a node for something that's not in sysfs, with udev.
Quoted 9 lines · Aaron Gyes
Quoted 2 lines · Greg KH
What do you mean by "static"? Something that persists over a reboot? Or after the device is removed?
Forgot to clarify. Create a node for something that's not in sysfs, with udev.
At least in Fedora, /etc/udev/makedevices.d or /etc/udev/devices.
--
Brian Gerst