[amdgpu] Screen corruption, distorted geometry and misplaced triangles

[Patrick Doyle]

Laptop is a Lenovo ThinkPad P14s Gen 2 AMD.
Ryzen 7 PRO 5850U
with Radeon (RX Vega 8?) integrated graphics
16 GB RAM

Running Ubuntu 23.10 (GNOME) from a clean install performed October 9, 2023 from a daily-live/current .iso generated on October 4, 2023.

Wayland
Kernel 6.5.0-9-generic

* * * * *

Installed a pre-release build of Ubuntu 23.10 to my ThinkPad the other day, was going through setting up and testing the usual programs.

Installed Steam through apt from the 'mantic' repositories. Installed Proton 8.0 and Steam Linux Runtime 3.0 (Sniper) alongside two compatible titles.

Screen corruption (white and grey streaks) present in-game when GNOME UI elements appeared on-screen (e.g., volume, brightness, and keyboard backlight indicators) and omnipresent after closing either game.

Artifacts remain on screen until log-out or reboot. Artifacts were not present beforehand.

Artifacts only appeared in Wayland session; not X11/Xorg.

I previously had been running the same games on Ubuntu 22.04 LTS (GNOME, Wayland) and Kubuntu 22.04 LTS (Plasma, X11/Xorg) on this computer without issue (kernel 6.2).

* * * * *

Please see subsequent posts for video/images.

Happy to provide any other information as needed. Thanks!

* * * * *

ProblemType: Bug
DistroRelease: Ubuntu 23.10
Package: xorg 1:7.7+23ubuntu2
ProcVersionSignature: Ubuntu 6.5.0-9.9-generic 6.5.3
Uname: Linux 6.5.0-9-generic x86_64
ApportVersion: 2.27.0-0ubuntu5
Architecture: amd64
BootLog: Error: [Errno 13] Permission denied: '/var/log/boot.log'
CasperMD5CheckResult: unknown
CompositorRunning: None
CurrentDesktop: ubuntu:GNOME
Date: Wed Oct 11 01:17:43 2023
DistUpgraded: Fresh install
DistroCodename: mantic
DistroVariant: ubuntu
ExtraDebuggingInterest: Yes, if not too technical
GraphicsCard:
 Advanced Micro Devices, Inc. [AMD/ATI] Cezanne [Radeon Vega Series / Radeon Vega Mobile Series] [1002:1638] (rev d1) (prog-if 00 [VGA controller])
   Subsystem: Lenovo Cezanne [Radeon Vega Series / Radeon Vega Mobile Series] [17aa:509b]
MachineType: {report['dmi.sys.vendor']} {report['dmi.product.name']}
ProcEnviron:
 LANG=en_US.UTF-8
 PATH=(custom, no user)
 SHELL=/bin/bash
 TERM=xterm-256color
 XDG_RUNTIME_DIR=<set>
ProcKernelCmdLine: BOOT_IMAGE=/vmlinuz-6.5.0-9-generic root=/dev/mapper/ubuntu--vg-ubuntu--lv ro quiet splash vt.handoff=7
SourcePackage: xorg
Symptom: display
UpgradeStatus: No upgrade log present (probably fresh install)
dmi.bios.date: 05/15/2023
dmi.bios.release: 1.24
dmi.bios.vendor: LENOVO
dmi.bios.version: R1MET54W (1.24 )
dmi.board.asset.tag: Not Available
dmi.board.name: 21A00068US
dmi.board.vendor: LENOVO
dmi.board.version: SDK0T76530 WIN
dmi.chassis.asset.tag: No Asset Information
dmi.chassis.type: 10
dmi.chassis.vendor: LENOVO
dmi.chassis.version: None
dmi.ec.firmware.release: 1.24
dmi.modalias: dmi:bvnLENOVO:bvrR1MET54W(1.24):bd05/15/2023:br1.24:efr1.24:svnLENOVO:pn21A00068US:pvrThinkPadP14sGen2a:rvnLENOVO:rn21A00068US:rvrSDK0T76530WIN:cvnLENOVO:ct10:cvrNone:skuLENOVO_MT_21A0_BU_Think_FM_ThinkPadP14sGen2a:
dmi.product.family: ThinkPad P14s Gen 2a
dmi.product.name: 21A00068US
dmi.product.sku: LENOVO_MT_21A0_BU_Think_FM_ThinkPad P14s Gen 2a
dmi.product.version: ThinkPad P14s Gen 2a
dmi.sys.vendor: LENOVO
version.compiz: compiz N/A
version.libdrm2: libdrm2 2.4.115-1
version.libgl1-mesa-dri: libgl1-mesa-dri 23.2.1-1ubuntu3
version.libgl1-mesa-glx: libgl1-mesa-glx N/A
version.xserver-xorg-core: xserver-xorg-core 2:21.1.7-3ubuntu2
version.xserver-xorg-input-evdev: xserver-xorg-input-evdev N/A
version.xserver-xorg-video-ati: xserver-xorg-video-ati 1:19.1.0-3
version.xserver-xorg-video-intel: xserver-xorg-video-intel 2:2.99.917+git20210115-1
version.xserver-xorg-video-nouveau: xserver-xorg-video-nouveau 1:1.0.17-2build1

[Daniel van Vugt (vanvugt)]

Also reported in:
https://gitlab.gnome.org/GNOME/gnome-shell/-/issues/7062

[Ubuntu Kernel Bot (ubuntu-kernel-bot)]

This change was made by a bot.

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Antoine DAVID (polemoss)]

I'm experiencing the same issue as well.

I'm using an Asus TUF A15 with an RTX 4050.

Gnome crashes when using multiple screens with Firefox and the file explorer.

One tip is to debug 'the frozen screen' by changing the screen order in Settings > Display.

Best regards,

[Noctis Bennington (noctis13)]

I just realised it happens when I change from 60hz to 120hz. Doesn't mean in 120hz my monitor crash, just when I make a change in my display, sometimes the desktop makes that strange glitch.

[Noctis Bennington (noctis13)]

https://gitlab.gnome.org/GNOME/gnome-shell/-/issues/7062

Seems it's closed. However, I'm still seeing this in 23.10. I see it when I open Steam. Definitely it has something to do with XWayland.

[Noctis Bennington (noctis13)]

New info: I realised this is something related to my integrated graphic in my laptop (AMD Radeon graphic integrated). So, if I launch Firefox, Steam etc. with my dedicated graphic (AMD Radeon RX 5600m), works perfectly, no issues.

[Matthias Uschok (uschok)]

Happened to me also. Saw the attached error in dmesg.

[Noctis Bennington (noctis13)]

How much do we have to wait if this fix is already released but not in Ubuntu 23.10?

[Daniel van Vugt (vanvugt)]

There was no fix released in gnome-shell. That's just a quirk of Launchpad that any upstream bug that is closed is set to 'Fix Released'.

The upstream gnome-shell bug was closed because we don't believe it is gnome-shell or mutter's fault. That's why this bug is now assigned to the kernel and Mesa.

[Noctis Bennington (noctis13)]

Thanks for the clarification, Daniel.

Could makes sense if it's related to Mesa. I have the feeling that these strange artifacts, appear especially when I launch windows and apps from the integrated graphics, instead of the dedicated.

[Daniel van Vugt (vanvugt)]

Please try adding COGL_DEBUG=disable-batching to /etc/environment and then reboot.

[Noctis Bennington (noctis13)]

Tested with COGL_DEBUG parameter added and at the moment, I can't reproduce these artifacts. I will keep trying it today.

[Noctis Bennington (noctis13)]

I tested it again, and at the moment I reproduce this issue with a xwayland window (in this case, Discord app), but all gnome was stable, which is strange (but good). See screenshot attached.

[DooMMasteR (winrootkit-w)]

This issue never happened to me with my 3x FHD setup at work, but now I am working from home with 2x 4K and 1x FHD and I am getting constant corruptions.

[Daniel van Vugt (vanvugt)]

More people, please try comment #16.

Please also ensure you don't have any unsupported extensions installed. Wobbly Windows comes to mind, but you can remove them all by running:

  cd ~/.local/share/gnome-shell/
  rm -rf extensions

and then log in again.

[Daniel van Vugt (vanvugt)]

The corruption also feels like something is broken with mid command buffer preemption so please try this kernel parameter:

  amdgpu.mcbp=0

[Niklas Schmelzle (nsmlzl)]

The graphic glitch just happened on my system (Ubuntu 23.10, Lenovo ThinkPad T14 Gen 2a with AMD Ryzen™ 5 PRO 5650U with Radeon™ Graphics × 12) again. I have `COGL_DEBUG=disable-batching` set in /etc/environment. Furthermore all extensions are deactivated (however, not deleted). System was also rebooted.

Interestingly, this time it did not occur to an XWayland window. I configured chrome to run wayland directly (which decreased the frequency of the error occuring, at least it seemed like it).

Here dmesg:

[70061.823842] amdgpu 0000:07:00.0: amdgpu: [gfxhub0] no-retry page fault (src_id:0 ring:24 vmid:1 pasid:32771, for process chrome pid 5084 thread chrome:cs0 pid 5121)
[70061.823858] amdgpu 0000:07:00.0: amdgpu: in page starting at address 0x0000000c5de7b000 from IH client 0x1b (UTCL2)
[70061.823865] amdgpu 0000:07:00.0: amdgpu: VM_L2_PROTECTION_FAULT_STATUS:0x00100430
[70061.823869] amdgpu 0000:07:00.0: amdgpu: Faulty UTCL2 client ID: IA (0x2)
[70061.823873] amdgpu 0000:07:00.0: amdgpu: MORE_FAULTS: 0x0
[70061.823877] amdgpu 0000:07:00.0: amdgpu: WALKER_ERROR: 0x0
[70061.823880] amdgpu 0000:07:00.0: amdgpu: PERMISSION_FAULTS: 0x3
[70061.823883] amdgpu 0000:07:00.0: amdgpu: MAPPING_ERROR: 0x0
[70061.823886] amdgpu 0000:07:00.0: amdgpu: RW: 0x0

I might try to set the kernel parameter next. For clarification: I would add amdgpu.mcb=0 to GRUB_CMDLINE_LINUX_DEFAULT in /etc/default/grub and then run 'sudo update-grub' (or update-grub2?). Afterwards I can just reboot?

BTW, just to rule other influences out: I do hibernate my system from time to time. This should not have any influence, or?

[Daniel van Vugt (vanvugt)]

> I might try to set the kernel parameter next. For clarification: I would add amdgpu.mcb=0 to
> GRUB_CMDLINE_LINUX_DEFAULT in /etc/default/grub and then run 'sudo update-grub' (or update-grub2?).
> Afterwards I can just reboot?

Yes that sounds correct other than a typo - it should be "amdgpu.mcbp=0".

> BTW, just to rule other influences out: I do hibernate my system from time to time. This should not
> have any influence, or?

It might be a factor, if you can confirm it..?

[Alexander Michler (amichler)]

I've tried the kernel parameter and those artifacts are still happening. But just on my FHD screen; not on my other screens (1440p).