Announcement

Collapse
No announcement yet.

NVIDIA 375.26 Driver Rolls Out With Several Fixes

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

  • NVIDIA 375.26 Driver Rolls Out With Several Fixes

    Phoronix: NVIDIA 375.26 Driver Rolls Out With Several Fixes

    In addition to NVIDIA updating their legacy Linux drivers today (xorg-server 1.19 support!), they have released a new build in their 375 driver series...

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    > With the NVIDIA 375 driver already being stable since 375.20

    It was tagged as stable, but it's the first update I do in a while that actually cause game crashs..

    Comment


    • #3
      Originally posted by by.peroux View Post
      > With the NVIDIA 375 driver already being stable since 375.20

      It was tagged as stable, but it's the first update I do in a while that actually cause game crashs..
      It had significantly more issues than that:

      stability issues, multi-minute stalls of X, kwin compositing crashing & settings for it (systemsettings5) not working,
      significantly less performance

      also see: https://devtalk.nvidia.com/default/t...-on-a-gtx970/1


      hopefully it's fixed with 375.26


      Michael thanks for re-enabling the edit function

      Comment


      • #4
        What about proper Optimus support? Sorry Nvidia, but f....k you!

        Michael, please update the post ASAP and advert people to STILL NOT USE this version. Thanks in advance.
        Last edited by timofonic; 14 December 2016, 12:48 PM.

        Comment


        • #5
          Originally posted by timofonic View Post
          What about proper Optimus support? Sorry Nvidia, but f....k you!

          Michael, please update the post ASAP and advert people to STILL NOT USE this version. Thanks in advance.
          People get so hung up on Optimus. I'm the first to value working features, but in Optiums' case, I don't see what the fuss is about. An idle desktop dGPU uses like 5-6W. So what do you gain from switching to the iGPU?
          At this point, Optimus should be taken behind the barn and "dealt with".

          Comment


          • #6
            Originally posted by timofonic View Post
            What about proper Optimus support? Sorry Nvidia, but f....k you!

            Michael, please update the post ASAP and advert people to STILL NOT USE this version. Thanks in advance.
            Actually, official Optimus support already here, as it is the purpose of libglvnd, which is supported by Fedora 25. See more details here :


            I agree that Open Source drivers would be preferable, but when they actually do something I think it is justice to say it loud.

            Comment


            • #7
              pfft, I can't even get SLI working at 4k@60hz......

              Comment


              • #8
                Originally posted by timofonic View Post
                Michael, please update the post ASAP and advert people to STILL NOT USE this version. Thanks in advance.
                lol... yeah I'm sure he'll get right on that.

                Comment


                • #9
                  Sorry Nvidia, but f....k you!
                  You're not one of the cool kids unless you post the picture of Linus and his middle finger.

                  Originally posted by johnc View Post
                  lol... yeah I'm sure he'll get right on that.
                  Maybe if he had posted the picture, Michael would have honored his request. Caps lock alone doesn't go so far these days...

                  Comment


                  • #10
                    Fixed an interaction problem between the Linux kernel's CONFIG_HARDENED_USERCOPY feature (new in Linux kernel 4.8) and nvidia-drm.ko with modeset=1 enabled. This problem caused CONFIG_HARDENED_USERCOPY to erroneously halt the system, issuing reports of the form:

                    usercopy: kernel memory overwrite attempt detected to ffff8803e8ec7ce0 (<process stack>) (8 bytes)
                    If this fixes what I think it'll fix, this may mean that tear-free will work on Fedora 25 and Xorg 1.19 on Optimus hardware. If so, that'll be awesome

                    Of course, that'll also rely on having drivers that allow KMS to be enabled. negativo17's drivers do not. RPM Fusion's packages apparently do.

                    Comment

                    Working...
                    X