Hey. Starting relatively recently, when I wake from sleep, some windows (mostly the file explorer Gnome Files) are blurry as in the screenshot. Moving the window or passing the cursor over the blurry area often helps, but sometimes I need to close the window and re-open it.

Here are the sections of my various .nix files that look relevant. I’m worried I have a bad video card!

    nvidia.modesetting.enable = true;
    nvidia.powerManagement.enable = false;
    nvidia.open = false;
    nvidia.nvidiaSettings = true;

Below is what looks like the relevant part of my restore from suspend from dmesg. Let me know if I can provide anything else helpful!

[21230.467601] nouveau 0000:01:00.0: gr: TRAP ch 11 [01ff748000 gst-plugin-scan[21807]]
[21230.467613] nouveau 0000:01:00.0: gr: GPC0/PROP trap: 00000020 [RT_HEIGHT_OVERRUN] x = 0, y = 8, format = 37, storage type = 0
[21230.467622] nouveau 0000:01:00.0: gr: GPC1/PROP trap: 00000020 [RT_HEIGHT_OVERRUN] x = 8, y = 8, format = 37, storage type = 0
[21230.467629] nouveau 0000:01:00.0: gr: GPC2/PROP trap: 00000020 [RT_HEIGHT_OVERRUN] x = 16, y = 8, format = 37, storage type = 0
[21230.486521] nouveau 0000:01:00.0: gr: TRAP ch 11 [01ff748000 gst-plugin-scan[21807]]
[21230.486536] nouveau 0000:01:00.0: gr: GPC0/PROP trap: 00000020 [RT_HEIGHT_OVERRUN] x = 0, y = 8, format = 37, storage type = 0
[21230.486548] nouveau 0000:01:00.0: gr: GPC1/PROP trap: 00000020 [RT_HEIGHT_OVERRUN] x = 8, y = 8, format = 37, storage type = 0
[21230.486556] nouveau 0000:01:00.0: gr: GPC2/PROP trap: 00000020 [RT_HEIGHT_OVERRUN] x = 16, y = 8, format = 37, storage type = 0
[21230.505718] nouveau 0000:01:00.0: gr: TRAP ch 11 [01ff748000 gst-plugin-scan[21807]]
[21230.505730] nouveau 0000:01:00.0: gr: GPC0/PROP trap: 00000020 [RT_HEIGHT_OVERRUN] x = 0, y = 8, format = 37, storage type = 0
[21230.505742] nouveau 0000:01:00.0: gr: GPC1/PROP trap: 00000020 [RT_HEIGHT_OVERRUN] x = 8, y = 8, format = 37, storage type = 0
[21230.505752] nouveau 0000:01:00.0: gr: GPC2/PROP trap: 00000020 [RT_HEIGHT_OVERRUN] x = 16, y = 8, format = 37, storage type = 0
[21234.024015] logitech-djreceiver 0003:046D:C537.0003: Unexpected input report number 128
[21234.029073] nouveau 0000:01:00.0: gr: TRAP ch 12 [01ff709000 .kgx-wrapped[21863]]
[21234.029088] nouveau 0000:01:00.0: gr: GPC1/TPC2/MP trap: global 00000000 [] warp 3d0016 []
[21234.036492] nouveau 0000:01:00.0: gr: TRAP ch 12 [01ff709000 .kgx-wrapped[21863]]
[21234.036506] nouveau 0000:01:00.0: gr: GPC2/TPC3/MP trap: global 00000000 [] warp 3e0016 []
[21234.104020] logitech-djreceiver 0003:046D:C537.0003: Unexpected input report number 128
[21237.439351] nouveau 0000:01:00.0: gr: TRAP ch 12 [01ff709000 .kgx-wrapped[21863]]
[21237.439364] nouveau 0000:01:00.0: gr: GPC2/TPC1/MP trap: global 00000000 [] warp 3c0016 []
[21243.107011] nouveau 0000:01:00.0: gr: DATA_ERROR 00000004 [INVALID_VALUE] ch 2 [01ffb9c000 vivaldi-bin[2820]] subc 0 class c197 mthd 0804 data 343ef401
[21243.107029] nouveau 0000:01:00.0: gr: DATA_ERROR 00000004 [INVALID_VALUE] ch 2 [01ffb9c000 vivaldi-bin[2820]] subc 0 class c197 mthd 0804 data 343ef001
[21243.121826] pcieport 0000:00:1c.7: AER: Correctable error message received from 0000:05:00.0
[21243.121838] rtl8821ae 0000:05:00.0: PCIe Bus Error: severity=Correctable, type=Physical Layer, (Receiver ID)
[21243.121840] rtl8821ae 0000:05:00.0:   device [10ec:8812] error status/mask=00000001/00006000
[21243.121842] rtl8821ae 0000:05:00.0:    [ 0] RxErr                  (First)

    • Björn@swg-empire.de
      link
      fedilink
      arrow-up
      0
      ·
      2 days ago

      I think I had a similar error on my laptop (MX150). If I recall correctly it would unscramble a bit while passing with the mouse over it.

      I’ve since ditched the nvidia card and only use the also included Intel one.

  • Deebster@infosec.pub
    link
    fedilink
    arrow-up
    0
    ·
    2 days ago

    I think I’ve seen other nvidia users have this issue, and it turned out to be a problem with the driver not preserving font bitmaps. Does the corruption go away or change if you zoom in/out (i.e. change the font size)?

  • petrichornetrainfall@piefed.social
    link
    fedilink
    English
    arrow-up
    0
    ·
    2 days ago

    I wouldn’t jump to your gpu going bad. There’s been quite a few issues in the past with nvidia + sleep + power management.

    I would recommend trying to comment out those values you shared, the generally recommendation is to use to the defaults when it comes to nvidia, unless youre positive you need to override them for some reason you feel very confident about.

    A lot of them aren’t named clearly and dont always do what you expect, at least from what ive read and from my personal experience.

    It also completes things a bit because gnome and kde like to automatically decide what gpu to use for what, its a little opaque, and sometimes causes issues.

    • petrichornetrainfall@piefed.social
      link
      fedilink
      English
      arrow-up
      0
      ·
      2 days ago

      Also are you intentionally using “novaveu” instead of their “open” version? If so, not “judgement” or anything, just wanting to confirm. Also what gen nvidia gpu? Alot of the settings have different defaults depending on whether your over or under a certain gen.

      • Clay_pidgin@sh.itjust.worksOP
        link
        fedilink
        English
        arrow-up
        0
        ·
        edit-2
        2 days ago

        The Nixos Wiki is often annoyingly confusing and/or out of date. I built my config based on the wiki but options are introduced in different sections and it isn’t clear which ones are supposed to be used together. Maybe I’ll bother the wiki maintainers on Matrix.

        I’m not “intentionally” using either driver version, so I’m happy to switch!

        https://github.com/NixOS/nixos-hardware/blob/master/common/gpu/nvidia/pascal/default.nix

        Only lists two option lines. Trying that now.

          # The open source driver does not support Pascal GPUs.
          hardware.nvidia.open = false;
          hardware.nvidia.package = lib.mkDefault config.boot.kernelPackages.nvidiaPackages.legacy_580;
        

        Well this is confusing. I added it set to true because sometimes the system doesn’t respond to input when it’s suspended and I need to reboot. Wonder if it’s contributing. I’ll leave it on for now, and turn it off once I’ve tried the two liner above for a few days.

        Graphical corruption and system crashes on suspend/resume powerManagement.enable = true can sometimes fix this, but is itself unstable and is known to cause suspend issues. hardware.nvidia.powerManagement.enable = true can also sometimes fix this issue; it is false by default.

        I’m running an Nvidia GTX 1070 (10th gen, Pascal) and I hadn’t remembered my system being that old! I don’t remember when I bought it, and I guess it’s too late to do anything about it now. Though maybe I can still get a previous-gen AMD card at a not too-outrageous price and have a big upgrade.

        From lspci:

        [nix-shell:~]$ sudo lshw -c video
        [sudo] password for clay: 
          *-display                 
               product: nouveaudrmfb
               physical id: 0
               bus info: pci@0000:01:00.0
               logical name: /dev/fb0
               version: a1
               width: 64 bits
               clock: 33MHz
               capabilities: pm msi pciexpress bus_master cap_list rom fb
               configuration: depth=32 driver=nouveau latency=0 resolution=3840,1600
               resources: irq:138 memory:f6000000-f6ffffff memory:e0000000-efffffff memory:f0000000-f1ffffff ioport:e000(size=128) memory:c0000-dffff
        
        • petrichornetrainfall@piefed.social
          link
          fedilink
          English
          arrow-up
          0
          ·
          2 days ago

          You are 100% right about thr wiki, and it doesnt help that there’s an unofficial wiki that pops up in search results too. And the wiki it especially out of date and confusing for nvidia. Its a known problem, and a bunch of “high level” contributers complain about it in discorse threads all the time.

          You linked “nixos-hardware” repo, so you’ve obviously done your homework. If youre not importing that as a flake, and importing the “common-gpu-nvidia” (or whatever the module is actually exposed as, thats just the pathing) modules, and instead grabbing the values manually, then you also need to grab the ones from here: https://github.com/NixOS/nixos-hardware/blob/master/common/gpu/nvidia/default.nix

          Your gpu is too old to be includes in the cuttoff that nvidia did when starting to open source their drivers, but i wouldn’t say its too old to be usable or anything.

          The wiki is out of date when they say power Management is experimental and unstable, I recommend setting that to true.

          I think the main problem is using nouveau drivers, and setting the option provided in the link above should switch that to their closed source drivers. If you make that change, definitely do a rebuild boot instead of switch, and then reboot into the new Gen.

          Its so annoying thag theres open (nouveau), open (nvidia - but only newer gpus), and closed (nvidia). There’s been a lot of talk about making major changes to the nixos nvidia configuration settings, but there’s a lot of conflicting ideas about how to improve them so it hasn’t made much progress.

          And apologizes for any weird formatting, especially around option names, on my phone.

          • Clay_pidgin@sh.itjust.worksOP
            link
            fedilink
            English
            arrow-up
            0
            ·
            2 days ago

            I appreciate the common tip; I’ll add that.

            My rebuild switch reboot seems to be working fine, so I’ll see what happens…

            Thanks for the time to spent trying to help me!

            • petrichornetrainfall@piefed.social
              link
              fedilink
              English
              arrow-up
              0
              ·
              1 day ago

              Im glad to hear that! Whenever I have changes for nvidia, I always need to “rebuild boot” and reboot, but that might be because I use “open = true” or because how I have home-manager setup.

              No problem! I’ve been in a similar situation if you couldn’t tell lol, I hope the changes work out!

  • just_another_person@lemmy.world
    link
    fedilink
    arrow-up
    0
    ·
    3 days ago

    Did you update the kernel or the Nvidia driver package recently? This is a bug that comes and goes for both AMD and Nvidia cards. Your GPU is fine.

    • Clay_pidgin@sh.itjust.worksOP
      link
      fedilink
      English
      arrow-up
      0
      ·
      2 days ago

      Thank you for the reassurance. I update packages automatically, but I don’t see either of those having changed recently. I’ll keep an eye out for upgrades touching them.