Announcement

Collapse
No announcement yet.

NVIDIA Linux Developers Don't Sound Too Happy About The ChromeOS Driver Approach

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • chithanh
    replied
    Originally posted by DDF420 View Post
    You would think a bloody display server protocol, would be built to work with Nvidia/AMD/Intel gpu's and any driver the USER wishes, and not forced to use a second rate opensource alternative.
    If you want to use proprietary drivers then you can go use Windows.

    NVidia needs Linux. Linux doesn't need NVidia, except for a few fringe use cases in HPC. And even there I suspect that people would rather stick with Linux than stick with NVidia if they had to choose.

    Leave a comment:


  • DDF420
    replied
    Originally posted by droidhacker View Post

    Frankly, if any complaints come from nvidia, IGNORE them.
    Yeah thats the way,just ignore their input,what a fantastic way to get driver support for this new protocol. Nvidia sticks up its middle finger in return, we support x, that will do you uncompromising bearded fools, anyone else should use Mir, they are not so our way or the highway. You would think a bloody display server protocol, would be built to work with Nvidia/AMD/Intel gpu's and any driver the USER wishes, and not forced to use a second rate opensource alternative.

    Leave a comment:


  • droidhacker
    replied
    Originally posted by Delgarde View Post

    From what I can see, NVidia are bringing up their experiences with ChomeOS as an example of the *wrong* way to do drivers. Seems quite relevant to the discussion...
    While possibly true, it is important to keep in mind that the interactions between google and nvidia over chromeos are probably *NOT* google's fault. Likely, google has a perfectly good system set up, and nvidia decided to be dicks about keeping their driver as obfuscated and non-transportable as possible, hence the need for a bunch of hacks to work around their crap binary drivers.

    I mean, we didn't hear about these problems with any OTHER hardware vendor wrt chromeos....

    Frankly, if any complaints come from nvidia, IGNORE them.

    Leave a comment:


  • SystemCrasher
    replied
    Originally posted by pal666 View Post
    well it doesn't looks like invidia developers' opinion is bringing something isn't known by upstream. in the end nvidia developers become more educated and do as they have been told
    Nvidia fails to get the idea: they are not part of process, therefore they have no word in what's going to be next thing. If they want something else, maybe they should stop trying to feed Linux users with proprietary blobs and actually cooperate around DRM/KMS and surrounding? Idea of bringing own proprietary driver implementing most of DRM/KMS on its own (because Linux devs also did very technical middle fingering by GPL_ONLY) is set for FAIL. Just does not fits the rest of processes and doomed to be a bumpy ride.

    Leave a comment:


  • pal666
    replied
    well it doesn't looks like invidia developers' opinion is bringing something isn't known by upstream. in the end nvidia developers become more educated and do as they have been told

    Leave a comment:


  • carewolf
    replied
    I wonder if the ChromeOS extensions are hardware implementation of the Chrome OpenGL extensions that Chrome uses between processes. Those are some weird crap.

    Leave a comment:


  • Jedipottsy
    replied
    Originally posted by Daktyl198 View Post
    Unless the ChromeOS drivers are either:
    - very similar to the current way Wayland drivers work via GBM, thus used as a counter-example to current Wayland drivers or
    - used as an example of the "right" way to do drivers (either to back-up their claims or to offer a 3rd option)
    there is 0 use in bringing up a 3rd driver stack other than as a distraction.
    Someone else brought up chromeos driver as a point of reference, to which NVIDIA replied how bad (in their opinion) the chromeos driver stack is

    Leave a comment:


  • Daktyl198
    replied
    Originally posted by Delgarde View Post

    From what I can see, NVidia are bringing up their experiences with ChomeOS as an example of the *wrong* way to do drivers. Seems quite relevant to the discussion...
    Unless the ChromeOS drivers are either:
    - very similar to the current way Wayland drivers work via GBM, thus used as a counter-example to current Wayland drivers or
    - used as an example of the "right" way to do drivers (either to back-up their claims or to offer a 3rd option)
    there is 0 use in bringing up a 3rd driver stack other than as a distraction.

    Leave a comment:


  • Delgarde
    replied
    Originally posted by Daktyl198 View Post
    Good for them. I guess? I don't remember anybody ever bringing up ChromeOS as an example of "the right way to do drivers" and there sure as hell isn't any ChromeOS specific extensions getting added to the Wayland spec so this discussion is pointless to the overall discussion. Are the NVidia guys just trying to find a bad guy worse than them in this situation so that when people look at their solution, all they see is "not as bad as..."?
    From what I can see, NVidia are bringing up their experiences with ChomeOS as an example of the *wrong* way to do drivers. Seems quite relevant to the discussion...

    Leave a comment:


  • slacka
    replied
    For a good overview of the Wayland stack explaining EGL vs GBM, check out this link.

    Leave a comment:

Working...
X