Bickering Continues About NVIDIA Using DMA-BUF

[medeiros]

So NVIDIA needs to stop development of driver to develop infrasctructure for the system that his developers could use to write their drivers...

[smitty3268]

if you link against the kernel in windows, is that derivative work?

Yes, it can be. But MS allows it with their proprietary license conditions. The GPL does not.

[smitty3268]

So NVIDIA needs to stop development of driver to develop infrasctructure for the system that his developers could use to write their drivers...

Either that, or stop trying to use GPL code others have written. The alternative is for them to come up with their own proprietary buffer sharing code, put it in their driver, and then convince the other linux driver writers to play nice with it. That's probably less work for them, although it might be a struggle to convince everyone to play nicely with nvidia's API.

[medeiros]

Either that, or stop trying to use GPL code others have written. The alternative is for them to come up with their own proprietary buffer sharing code, put it in their driver, and then convince the other linux driver writers to play nice with it. That's probably less work for them, although it might be a struggle to convince everyone to play nicely with nvidia's API.

So the best idea is just forget about optimus support on linux...

[vertexSymphony]

This is simple ... this code has a licensing, and as a loong time policy in the kernel development is to let the Author hold the rights of his own work (that's why re-licensing to GPLv3 was practically impossible when that debate came up), that means that the only person that is able to change/add the actual contract (license) is the copyright holder (the author).

So, if Alan wrote it ... and he doesn't want to re-license or add an exception, then i't simple: shut the fuck up, it's the copyright holder decision and need as much respect as the decision to allow exceptions.
If this situation didn't change (example: dma-buf being reimplemented by an author ok with creating a drop-in remplacement with the exception), is not because Alan being monarch (really? blaming on him? cheap shot), simply because everyone seems to agree with that.

So yeah, move along ... nothing to see here
Oh, and btw ... nVIDIA isn't developing linux drivers just for being "cool", they have a clear attitude showing it when they don't even release docs ... they do because they have paying customers, and as long they have paying customers, they'll provide drivers for the platforms.

No need to play the "I feel insignificant, I need their help and I have to be friendly with them so they continue to provide drivers" syndrome ... they'll stay as long as it pay the bills, and it does.
Linux will continue to evolve without having ballast thanks to those pricks ... move along

Regards.

[johnc]

This is simple ... this code has a licensing, and as a loong time policy in the kernel development is to let the Author hold the rights of his own work (that's why re-licensing to GPLv3 was practically impossible when that debate came up), that means that the only person that is able to change/add the actual contract (license) is the copyright holder (the author).

So, if Alan wrote it ... and he doesn't want to re-license or add an exception, then i't simple: shut the fuck up, it's the copyright holder decision and need as much respect as the decision to allow exceptions.

Alan wasn't the original author of the code. And in fact it was initially written EXPORT_SYMBOL, and was only later changed -- presumably without the consent of all involved.

Whether there's really a difference between EXPORT_SYMBOL and EXPORT_SYMBOL_GPL is pretty tenuous, and NVIDIA is probably already accessing EXPORT_SYMBOL code.

From what I understand, NVIDIA has a sign-off from most of the people involved... just Alan and one or two others are simply saying that NVIDIA should consult their lawyers [because they would be presumably violating the license]. Not that any of that means the patch will be accepted.


Now I see why Sun released all their stuff under the CDDL. One so they don't have to deal with GPL bullshit, and two just so they can stab a fork in the eye of these Stallman nuthuggers.

[kayosiii]

Guess next time you will not buy something that is made by a company who has such horrible support, I always buy intel for the simple fact that I know it will work... I have no complaints graphically...my dual core celeron Sandy Bridge can play xonotic at mostly high settings....

I wish that were the case. I currently have two options Linux + nVidia binaries or windows. I need OpenGL 4 support, I need speed comparable to what nVidia gets on windows or better, I need enough stability to be a production platform. If there was an opensource driver that gave me that I would quite happily ditch nVidia; currently this is not the case.

Fortunately optimus support isn't part of my work requirement so this particular battle doesn't affect me other to say in a few years time I may wish to upgrade my laptop.

I can see that it is good in general to encourage companies to write opensource drivers for Linux and I think GPL is part of the reason so many people contribute to Linux (rather than say BSD). I am personally feel better contributing code under those terms.

I wonder why there is need for both the export and export_gpl the situation does seem like there is a gentlemans agreement on the former.

[necro-lover]

my suggestion is: switch to the AGPLv3 for the kernel source and then start a real fight against enemy of the freedom like Nvidia.

[vertexSymphony]

my suggestion is: switch to the AGPLv3 for the kernel source and then start a real fight against enemy of the freedom like Nvidia.

Well, take a deep breath and find EVERY single person that made the actual state of the kernel ( the copyright holders of their code) ... so they can sign the consent for the re-licensing of their work :S
I can't imagine what would happen if some legal framework changes and somehow in the future Linux needs an upgraded license.

Having this "you keep the ownership of your work" has a lots of situational pros, but also a lot of cons that are kinda scary.

Regards.

[joe_gunner]

Well, take a deep breath and find EVERY single person that made the actual state of the kernel ( the copyright holders of their code) ... so they can sign the consent for the re-licensing of their work :S
I can't imagine what would happen if some legal framework changes and somehow in the future Linux needs an upgraded license.

Seriously why the fuck not just dual-license the related code? Certainly that won't involve permissions from more than a couple of programmers and save everyone some trouble, but for that Alan Cox would have to take his and Stallmans cox out of his own mouth.