ThinkPad P72 - Very serious issues with external monitors on Thunderbolt dock

[Alex Zero]

I'm running Ubuntu 20.04 LTS on a ThinkPad P72 with a Quadro P5200. I recently purchased the ThinkPad Workstation Dock Gen2 and have been having issues with external displays on it.

There are a multitude of issues with this. Sometimes you can plug in displays and both will show up fine after running xrandr --auto. Sometimes only one will appear and the other will stay black.

There is a chance for any of these things to happen after making a display configuration change:
- Nothing (displays work fine)
- One display stays black
- All displays including laptop internal display freeze, which is recoverable only by physically disconnecting the displays from the dock
- Displays flicker on and off rapidly before every one goes dark including the laptop's internal display, requiring a hard reset.

I have tried:
- Upgrading to a newer kernel version (5.6.7) via mainline
- Upgrading all system firmware, including dock, to latest via Lenovo Vantage in Windows dual boot
- Downgrading the dock firmware from 3166 to 3164 - despite suggestions by people in this thread in the Lenovo community https://forums.lenovo.com/t5/Ubuntu/Thinkpad-P1-Gen2-amp-Thunderbolt-3-Workstation-Dock-amp-Linux-Ubuntu-Monitor-s-black/m-p/4575709?page=3 - this did not work
- Downgrading the NVIDIA driver version to both 435 and 390 - neither worked

Downgrading to an older kernel (4.19) appears to work, but it still has some issues with locking up while repositioning displays using xrandr.

ProblemType: Bug
DistroRelease: Ubuntu 20.04
Package: xorg 1:7.7+19ubuntu14
ProcVersionSignature: Ubuntu 5.4.0-28.32-generic 5.4.30
Uname: Linux 5.4.0-28-generic x86_64
NonfreeKernelModules: zfs zunicode zavl icp zcommon znvpair nvidia_modeset nvidia
.proc.driver.nvidia.gpus.0000.01.00.0: Error: [Errno 21] Is a directory: '/proc/driver/nvidia/gpus/0000:01:00.0'
.proc.driver.nvidia.registry: Binary: ""
.proc.driver.nvidia.suspend: suspend hibernate resume
.proc.driver.nvidia.suspend_depth: default modeset uvm
.proc.driver.nvidia.version:
 NVRM version: NVIDIA UNIX x86_64 Kernel Module 440.64 Fri Feb 21 01:17:26 UTC 2020
 GCC version: gcc version 9.3.0 (Ubuntu 9.3.0-10ubuntu2)
ApportVersion: 2.20.11-0ubuntu27
Architecture: amd64
CasperMD5CheckResult: skip
CompizPlugins: No value set for `/apps/compiz-1/general/screen0/options/active_plugins'
CompositorRunning: None
Date: Sun May 3 00:37:27 2020
DistUpgraded: 2020-03-04 19:48:20,592 DEBUG Running PostInstallScript: './xorg_fix_proprietary.py'
DistroCodename: focal
DistroVariant: ubuntu
DkmsStatus:
 acpi-call, 1.1.0, 5.4.0-28-generic, x86_64: installed
 nvidia, 440.64, 5.4.0-28-generic, x86_64: installed
 v4l2loopback, 0.12.3, 5.4.0-28-generic, x86_64: installed
 wireguard, 1.0.20200413, 5.4.0-28-generic, x86_64: installed (WARNING! Diff between built and installed module!)
ExtraDebuggingInterest: Yes, including running git bisection searches
GpuHangFrequency: This is the first time
GraphicsCard:
 Intel Corporation Device [8086:3e94] (prog-if 00 [VGA controller])
   Subsystem: Lenovo Device [17aa:2269]
 NVIDIA Corporation GP104GLM [Quadro P5200 Mobile] [10de:1bb5] (rev a1) (prog-if 00 [VGA controller])
   Subsystem: Lenovo GP104GLM [Quadro P5200 Mobile] [17aa:2269]
InstallationDate: Installed on 2019-10-19 (196 days ago)
InstallationMedia: Ubuntu 19.10 "Eoan Ermine" - Release amd64 (20191017)
MachineType: LENOVO 20MCS08200
ProcEnviron:
 LANGUAGE=en_GB:en
 TERM=xterm-256color
 PATH=(custom, no user)
 LANG=en_GB.UTF-8
 SHELL=/bin/bash
ProcKernelCmdLine: BOOT_IMAGE=/vmlinuz-5.4.0-28-generic root=/dev/mapper/vgubuntu-root ro quiet splash intel_iommu=on default_hugepagesz=1G hugepagesz=1G hugepages=32 transparent_hugepage=never vt.handoff=7
SourcePackage: xorg
Symptom: display
Title: Xorg freeze
UpgradeStatus: Upgraded to focal on 2020-03-04 (59 days ago)
dmi.bios.date: 01/15/2020
dmi.bios.vendor: LENOVO
dmi.bios.version: N2CET50W (1.33 )
dmi.board.asset.tag: Not Available
dmi.board.name: 20MCS08200
dmi.board.vendor: LENOVO
dmi.board.version: SDK0Q40104 WIN
dmi.chassis.asset.tag: 00040
dmi.chassis.type: 10
dmi.chassis.vendor: LENOVO
dmi.chassis.version: None
dmi.modalias: dmi:bvnLENOVO:bvrN2CET50W(1.33):bd01/15/2020:svnLENOVO:pn20MCS08200:pvrThinkPadP72:rvnLENOVO:rn20MCS08200:rvrSDK0Q40104WIN:cvnLENOVO:ct10:cvrNone:
dmi.product.family: ThinkPad P72
dmi.product.name: 20MCS08200
dmi.product.sku: LENOVO_MT_20MC_BU_Think_FM_ThinkPad P72
dmi.product.version: ThinkPad P72
dmi.sys.vendor: LENOVO
nvidia-settings:
 ERROR: Unable to load info from any available system

 ERROR: Unable to load info from any available system
version.compiz: compiz N/A
version.libdrm2: libdrm2 2.4.101-2
version.libgl1-mesa-dri: libgl1-mesa-dri 20.0.4-2ubuntu1
version.libgl1-mesa-glx: libgl1-mesa-glx 20.0.4-2ubuntu1
version.nvidia-graphics-drivers: nvidia-graphics-drivers-* N/A
version.xserver-xorg-core: xserver-xorg-core 2:1.20.8-2ubuntu2
version.xserver-xorg-input-evdev: xserver-xorg-input-evdev N/A
version.xserver-xorg-video-ati: xserver-xorg-video-ati 1:19.1.0-1
version.xserver-xorg-video-intel: xserver-xorg-video-intel 2:2.99.917+git20200226-1
version.xserver-xorg-video-nouveau: xserver-xorg-video-nouveau 1:1.0.16-1

[Daniel van Vugt (vanvugt)]

Note:

1. It's likely all of your display ports are connected to the Intel GPU so we can probably ignore Nvidia here. To be sure please try disabling the discrete GPU in your BIOS and verify the bug still occurs.

2. If older kernels like 4.19 work then this should be assigned to the kernel.

[Alex Zero (citadelcore)]

Hi:

For (1): The NVIDIA GPU cannot be disabled in the BIOS, only switched to the "hybrid" mode. The outputs are shown as attached to the GPU with these commands:
$ cat '/sys/devices/pci0000:00/0000:00:01.0/0000:01:00.0/vendor'
0x10de
$ ls '/sys/devices/pci0000:00/0000:00:01.0/0000:01:00.0/drm/card0'
card0-DP-1 card0-DP-2 card0-DP-3 card0-eDP-1 card0-HDMI-A-1 dev device power subsystem uevent

As you can see, the vendor of this PCI device is NVIDIA and the outputs are attached directly to it.

For (2), I have tried with 4.19 on Debian 10 and it produced the same issue.

[Launchpad Janitor (janitor)]

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

[Alon Diamant (advance512)]

Just a remark. I had the same issue, a Lenovo P1 Gen 2 (Quadro T2000), a Thunderbolt 3 Workstation Dock Gen 2 (firmware sub-version 64) and nVidia drivers 455 over Ubuntu 20.04 with kernel 5.4.85-050485-generic.

More over, I could not log-in at all unless I had the Dock connected, for some strange reason. After entering the password, with the laptop undocked, Ubuntu would freeze - the session would not start. (The mouse still moved, but the UI was no longer responsive).
In Windows, everything worked fine.
Kernel 5.4.58 had the same issues. Oddly, in kernel 5.4.56 I could log in, but still had all the issues you mentioned above.

Changing the drivers to Nouveau meant I could log-in, but only one external monitor worked (I have 2).

All of this with Kernel DMA Disabled and running at "No Security" to make everything simpler.
What I ended up doing was:

Changed to nVidia drivers 418 (I lose the smart power management, so battery time is terrible).
Thunberbolt 3 Security at "User Authorization" (for now).
Now, everything works (so far). Suspending works, resuming works, shutting off works (all of these would sometimes freeze).

It's pretty sad that this is the state of Lenovo's state of the art laptops and docks, on Ubuntu which they are supposedly certified for.

We are now sadly considering moving our entire shop to Macs.