Kernel Developers Work To Block NVIDIA "GPL Condom" Effort Around New NetGPU Code

[tuxayo]

Sorry man, I just don't feel open source should have a whole bunch of restrictions attached to it. How is that free? Maybe you should join the M$ team if you like forcing people to do stuff.

A way that hopefully helps to phrase the conflict: It's about letting the freedom to restrict the freedom of other users or not.

Complement: upstream gives freedom over some piece of code:
- option A: very simple, everything can be done with it. Including restricting the freedom of users of the derivative work (permissive)
- option B: one might not want to give freedom over some piece of code just to have someone else to use it without giving the same freedom to users. (copyleft (strong or weak is another nuance))

A critic of the libre/open source movement and ecosystem is that in many cases, the end user doesn't see the freedom while the application maker was able to benefit from a huge quantity of existing work that immensely helped them to make a great thing.

Cheers,

[Svartalf]

Depends how far back you go. http://utah-glx.sourceforge.net/ Nvidia was asked way back in 2003 to open up their complete graphics stack.

Heh. If you go far enough back, you'd know that the NV1 chipset had ALL of it's programming details publicly available. I'd have provided a driver for Utah-GLX for it had I been able to lay hands on it at the time. Nobody was really using their chip in anything other than Engineering boards at the time and it was about as easy to lay hands on as a BitBoys OY part when they might've been a thing. There were several promising parts that fizzled or wouldn't give me or any of the other Utah-GLX people the time of day.

NVidia survived, but they closed off the subsequent generations of their parts. I won't comment TOO much on my take about their being as closed as they can be...but at one point, they were very, very open. Only one in the pack- because ATI and Matrox were much more tight-fisted with their info. Their being closed wan't the magic sauce they're thinking it is and didn't give them some magic advantage...unless they're cheating in the guts of the drivers somehow.

As for their stopping support for their parts under Linux... Not likely. Their future is kind of tied to Tegra...and there's a LOT of their customers that could grind them into fine dust if they pulled that one...

[Kerisun]

Yes, but the point is that manufacturers ain't doing it. So far I haven't been able to find any Ryzen 4k laptops with a thunderbolt.

[wyatt8740]

I'd find it great if any kernel point update deliberately broke any out of tree parasites. Some people either should learn the hard way or get lost.

I'm late, but cdemu/vhba is a nice out-of-tree project.

[duby229]

Sorry man, I just don't feel open source should have a whole bunch of restrictions attached to it. How is that free? Maybe you should join the M$ team if you like forcing people to do stuff.

Yeah, no... The entire point of the GPL is to ensure that any modifications you make you must share upstream. But nVidia has tried to do over and over again is to use kernel code like a shim so that they can compile their modifications into their binaries without sharing them upstream.... That's not ok.... It's an obvious GPL violation....