EDIT: This has been fixed, it all mostly chalked up to possibly a driver / kernel bug. Why things managed to work so well for so long, who can say? Anyway all you need to do is upgrade your Linux kernel to 6.5.x, and your Mesa driver to the latest Mesa git.

Hey all,

I’ve been playing Starfield for the last two weeks, mostly been having a great time with it! I’m using the latest kisak mesa drivers and ProtonGE, and it’s been a smooth and stable experience for the most part.

However…within the last two days, I’ve started having a major issue. Any time I land on a planet and exit my ship, the loading screen hangs, and I get kicked back to my OS’es login screen (Ubuntu Studio 23.04). The audio continues to play, not in a staggered way, but as if the game was still running. And yet, my desktop has no idea about the executable being active.

I dunno if a patch came out within the last day or so that fundamentally changed something, or maybe my hardware’s just finally started melting after weeks of playing. Performance is degraded out of nowhere.

What’s curious is that, every time I reload my save, the game runs fine from within my ship. But, if I try to travel to another planet, or exit out of my landing bay, everything I’ve just described happens all over again. This seems to happen regardless of whether the game is windowed, borderless, medium graphics settings, low graphics settings.

I decided to check whether my hardware was just straight-up melting. It’s not. My other Proton titles that are relatively heavy (Returnal, Resident Evil 4, The Callisto Protocol) all run smoothly at higher graphics settings.

Some specs info:

  • Operating System: Ubuntu Studio 23.04
  • KDE Plasma Version: 5.27.4
  • KDE Frameworks Version: 5.104.0
  • Qt Version: 5.15.8
  • Kernel Version: 6.2.0-1013-lowlatency (64-bit)
  • Graphics Platform: X11
  • Processors: 32 × AMD Ryzen 9 7950X 16-Core Processor
  • Memory: 62.5 GiB of RAM
  • Graphics Processor: AMD Radeon RX 6700 XT
  • Manufacturer: ASUS

Also, software:

  • ProtonGE: GE-Proton-8-15
  • OpenGL core profile version string: 4.6 (Core Profile) Mesa 23.1.8 - kisak-mesa PPA
  • OpenGL version string: 4.6 (Compatibility Profile) Mesa 23.1.8 - kisak-mesa PPA
  • OpenGL ES profile version string: OpenGL ES 3.2 Mesa 23.1.8 - kisak-mesa PPA-
  • azvasKvklenko@sh.itjust.worksEnglish
    5·
    1 year ago

    Going back to login screen is concerning at the very least, like it wasn’t your game crashing, but the Xserver or Plasma, which could potentially be a kernel bug (with amdgpu) or Mesa, or Xorg…

    Try to gather some logs as it happens. See if it’s kernel/amdgpu thing

    sudo dmesg

    See if it’s something in user space crashing (plasma session? Xorg?)

    journalctl -b # then press capital G to jump to the end of logs

    I’d try to:

    • trying the Plasma Wayland session to see if it’s not Xorg (should be straightforward given that you’re on AMD/RDNA2).
    • revert back to stock kernel
    • trying mainline kernel 6.5
    • revert to stock Mesa
  • conciselyverbose@kbin.social
    3·
    1 year ago

    Just a statement that this isn’t a global proton problem with the current build. I’m having no issue, also with GE (not positive exactly which build).

    Do you have any crash logs? People (more qualified than me) can sometimes use details from there to identify the root cause.

    • Sean Tilley@lemmy.mlOPEnglish
      2·
      1 year ago

      Unfortunately, I’m not sure as to how to find them, or enable them if they’re not already working. I agree that logs would definitely help here.

  • hallettj@beehaw.orgEnglish
    2·
    1 year ago

    My shot-in-the-dark guess is that you were unlucky, and hit a bug in the game. It might be something triggered by a particular detail in your game state which would explain why you didn’t have problems earlier. You could test that by starting a new character, and testing whether you see the same problem.

    If I were in your position and feeling motivated I would submit a bug report to Bethesda with as much detail as I could manage.

    • Sean Tilley@lemmy.mlOPEnglish
      2·
      1 year ago

      Yeah, it’s possible. I’m curious, though, as to why it affected my other saves retroactively? Like, I can run around within TMD / the old Red Devils facility just fine, but the game craps out when I set foot on Mars.

      • conciselyverbose@kbin.social
        4·
        1 year ago

        Have you tried backing up your saves and using the delete proton files option?

        This will delete your saves from your local device, but it effectively re-installs the game and could clean up corrupted files if that is the issue. I know there’s a lengthy shader compile process, so dumping your shader cache and allowing it to recompile could also be useful.

  • Sean Tilley@lemmy.mlOPEnglish
    2·
    1 year ago

    Hey, final update here - I managed to get the game working again! I think a patch came through somewhere that broke something - maybe it was ProtonGE, maybe it was Valve patching Proton Experimental, maybe it was Bethesda putting out something weird in a patch. Regardless, I got a tip that you can upgrade your Linux kernel to 6.5.x, and upgrade to the latest Mesa git drivers. The game is now running smoothly again without any issues on Wayland!

  • JohnWorks@sh.itjust.works
    2·
    1 year ago

    Have you tried a different proton version? I know on my steam deck it used to be where some games wouldn’t run well on certain proton versions so I had to manually set proton to run as experimental.

    • Sean Tilley@lemmy.mlOPEnglish
      2·
      1 year ago

      Yeah, I’ve tried several, of both the GE and Experimental varieties. The problem persists, sadly.

      It feels as though the game is suddenly only capable of loading interior spaces, despite the game running great as recently as two days ago.

  • Sean Tilley@lemmy.mlOPEnglish
    1·
    1 year ago

    Alright, bit of an update: after investigating my hardware, doing a fresh OS install, and trying just about everything that everybody’s suggested to me so far, I have logs. Definitely amdgpu related, but not 100% sure what is going on yet. This was on Wayland, searches for similar error messages suggest a possible driver bug, maybe issues with particular kernel versions? Still investigating.

    amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:2 pasid:32785, for process Starfield.exe pid 71830 thread vkd3d_queue pid 71960)
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x0000800131efe000 from client 0x1b (UTCL2)
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00201031
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          Faulty UTCL2 client ID: TCP (0x8)
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          MORE_FAULTS: 0x1
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          WALKER_ERROR: 0x0
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          PERMISSION_FAULTS: 0x3
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          MAPPING_ERROR: 0x0
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          RW: 0x0
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:2 pasid:32785, for process Starfield.exe pid 71830 thread vkd3d_queue pid 71960)
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x0000800131efe000 from client 0x1b (UTCL2)
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00201031
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          Faulty UTCL2 client ID: TCP (0x8)
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          MORE_FAULTS: 0x1
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          WALKER_ERROR: 0x0
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          PERMISSION_FAULTS: 0x3
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          MAPPING_ERROR: 0x0
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          RW: 0x0
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:2 pasid:32785, for process Starfield.exe pid 71830 thread vkd3d_queue pid 71960)
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x0000800131efe000 from client 0x1b (UTCL2)
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00201031
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          Faulty UTCL2 client ID: TCP (0x8)
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          MORE_FAULTS: 0x1
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          WALKER_ERROR: 0x0
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          PERMISSION_FAULTS: 0x3
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          MAPPING_ERROR: 0x0
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          RW: 0x0
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:2 pasid:32785, for process Starfield.exe pid 71830 thread vkd3d_queue pid 71960)
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x0000800131eff000 from client 0x1b (UTCL2)
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00000000
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          Faulty UTCL2 client ID: CB/DB (0x0)
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          MORE_FAULTS: 0x0
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          WALKER_ERROR: 0x0
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          PERMISSION_FAULTS: 0x0
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          MAPPING_ERROR: 0x0
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          RW: 0x0
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:2 pasid:32785, for process Starfield.exe pid 71830 thread vkd3d_queue pid 71960)
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x0000800131eff000 from client 0x1b (UTCL2)
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00000000
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          Faulty UTCL2 client ID: CB/DB (0x0)
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          MORE_FAULTS: 0x0
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          WALKER_ERROR: 0x0
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          PERMISSION_FAULTS: 0x0
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          MAPPING_ERROR: 0x0
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          RW: 0x0
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:2 pasid:32785, for process Starfield.exe pid 71830 thread vkd3d_queue pid 71960)
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x0000800131efe000 from client 0x1b (UTCL2)
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00000000
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          Faulty UTCL2 client ID: CB/DB (0x0)
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          MORE_FAULTS: 0x0
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          WALKER_ERROR: 0x0
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          PERMISSION_FAULTS: 0x0
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          MAPPING_ERROR: 0x0
Sep 24 02:29:50 Asmodeus kernel: amdgpu 0000:03:00.0: amdgpu:          RW: 0x0
Sep 24 02:30:00 Asmodeus kernel: [drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring gfx_0.0.0 timeout, but soft recovered