Skip to main content

Nvidia Wants to Remove Some GPL From Linux Kernel Code

The company's Robert Morell submitted a patch to remove the GPL license from the dma buffer interface in the Linux kernel so it can be used in Nvidia's driver. Not everybody is happy, especially Alan Cox, who has been involved in Linux development since 1991 and was most influential when he maintained the version 2.2 of the Linux kernel.

Cox unmistakably told Morell that he and "many others" are not happy with the idea of combining proprietary and GPL code. Cox declined to accept Morell's patch. The argument appears to be stuck at this time as words started flying and Cox told Morell that a patch acceptance will have to get the confirmation of all rightsholders of the code, some of apparently include Red Hat and Intel. The developer left Morell with the thought to get those confirmations via attorneys.

If it is up to Cox, it appears to be rather unlikely that Nvidia will get anywhere without making substantial concessions. However, he also noted that he will not fight tooth and nail either: "I also have better things to do with life than sue Nvidia and start an all out copyright and patent war in Linuxspace," he wrote. "It's simple enough. If Nvidia think their code is not derivative then why do they care about the _GPL being significant ?"

Morell, however, has not given up yet: "My intention is not to steal any code from the kernel or change any licenses," he posted on the discussion thread. "The goal here is to allow interoperation between drivers. […] I believe that the developers and maintainers of dma-buf have provided the needed signoff, both in person and in this thread.  If there are any objections from that group, I'm happy to discuss any changes necessary to get this merged."

Expect this discussion to last for some time and don't expect this discussion to improve the relationship between Nvidia and the Linux developer community.

Contact Us for News Tips, Corrections and Feedback

  • sten_gn
    Why do i have a cold feeling down my spine ...
    Reply
  • Dax corrin
    9418034 said:
    Why do i have a cold feeling down my spine ...

    Because it's fall and it's cold out?
    Reply
  • chewy1963
    WTF Nvidia? Just release your drivers via the GPL and you wouldn't have to worry about it! But noooo, you want to keep your drivers proprietary. In the words of the immortal Linus Torvalds: "F*&^ You, Nvidia!
    Reply
  • spasmolytic46
    Arg! The linux version of steam is so close. Don't f*ck this up for me nvidia.
    Reply
  • cookoy
    Don't allow a bad precedent no matter what its stated good intentions are. Just say no way.
    Reply
  • rantoc
    First Nvidia get stoned for not providing drivers and now when they work towards that end they get stoned, what a grateful bunch of peeps!
    Reply
  • alidan
    spasmolytic46Arg! The linux version of steam is so close. Don't f*ck this up for me nvidia.
    if the drivers go open, wouldnt you be able to code a work around to get cuda and physx up and going on an amd card?
    Reply
  • spasmolytic46
    9418079 said:
    if the drivers go open, wouldnt you be able to code a work around to get cuda and physx up and going on an amd card?

    I want open drivers (although I don't care about CUDA or Physx), but I was more worried about nvidia starting a pissing match that winds up with them abandoning Linux support after they get told off. This is why I'm worried nvidia will screw this up for me.
    Reply
  • division_9
    I understand why Nvidia wants to keep it's drivers closed. But I swear to god, this is about as awkward as a guy walking into the girls bathroom. It just ain't right.
    Reply
  • jhansonxi
    rantocFirst Nvidia get stoned for not providing drivers and now when they work towards that end they get stoned, what a grateful bunch of peeps!That's not the issue. The kernel developers have many problmes debugging involving closed-source modules and get often changes requests because of them. Since they don't have access to the code, they can't identify if the module or the kernel code is responsible for errors and can't tell if the modules have security holes.

    Nvidia sells hardware, not drivers, so releasing programming info for the GPUs doesn't cost them anything. Video development with Nvidia devices is slow because only Nvidia has the info. This also means that when older devices are no longer supported by Nvidia, they can't easily be supported by anyone else either. This is a significant problem with laptops because the GPUs are normally not changeable.
    Reply