i915 keeps crashing with "pipe state doesn't match!" in verify_crtc_state

[vinibali]

I'm using a Docking station with two displays (1 HDMI + 1 DP) and one HDMI to the machine's port.

ProblemType: Bug
DistroRelease: Ubuntu 20.04
Package: xorg 1:7.7+19ubuntu14
ProcVersionSignature: Ubuntu 5.15.0-72.79~20.04.1-generic 5.15.98
Uname: Linux 5.15.0-72-generic x86_64
ApportVersion: 2.20.11-0ubuntu27.26
Architecture: amd64
BootLog: Error: [Errno 13] Permission denied: '/var/log/boot.log'
CasperMD5CheckResult: skip
Date: Thu May 18 15:11:56 2023
DistUpgraded: Fresh install
DistroCodename: focal
DistroVariant: ubuntu
ExtraDebuggingInterest: Yes
GraphicsCard:
 Intel Corporation Device [8086:9a49] (rev 01) (prog-if 00 [VGA controller])
   Subsystem: Hewlett-Packard Company Device [103c:8846]
InstallationDate: Installed on 2022-08-05 (286 days ago)
InstallationMedia: Ubuntu 20.04.4 LTS "Focal Fossa" - Release amd64 (20220223)
MachineType: HP HP EliteBook 850 G8 Notebook PC
ProcKernelCmdLine: BOOT_IMAGE=/vmlinuz-5.15.0-72-generic root=/dev/mapper/elx_vg-root ro mem_sleep_default=deep noaudit noquiet nosplash sysrq_always_enabled i915.enable_guc=1 intel_idle.max_cstate=1
SourcePackage: xorg
Symptom: display
Title: Xorg crash
UpgradeStatus: No upgrade log present (probably fresh install)
dmi.bios.date: 08/12/2021
dmi.bios.release: 6.2
dmi.bios.vendor: HP
dmi.bios.version: T76 Ver. 01.06.02
dmi.board.name: 8846
dmi.board.vendor: HP
dmi.board.version: KBC Version 30.33.00
dmi.chassis.type: 10
dmi.chassis.vendor: HP
dmi.ec.firmware.release: 48.51
dmi.modalias: dmi:bvnHP:bvrT76Ver.01.06.02:bd08/12/2021:br6.2:efr48.51:svnHP:pnHPEliteBook850G8NotebookPC:pvr:rvnHP:rn8846:rvrKBCVersion30.33.00:cvnHP:ct10:cvr:sku4T9G0EC#ABB:
dmi.product.family: 103C_5336AN HP EliteBook
dmi.product.name: HP EliteBook 850 G8 Notebook PC
dmi.product.sku: 4T9G0EC#ABB
dmi.sys.vendor: HP
modified.conffile..etc.default.apport:
 # set this to 0 to disable apport, or to 1 to enable it
 # you can temporarily override this with
 # sudo service apport start force_start=1
 enabled=0
mtime.conffile..etc.default.apport: 2022-08-05T12:33:01.637811
version.compiz: compiz N/A
version.libdrm2: libdrm2 2.4.107-8ubuntu1~20.04.2
version.libgl1-mesa-dri: libgl1-mesa-dri 21.2.6-0ubuntu0.1~20.04.2
version.libgl1-mesa-glx: libgl1-mesa-glx N/A
version.xserver-xorg-core: xserver-xorg-core 2:1.20.13-1ubuntu1~20.04.8
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 N/A
version.xserver-xorg-video-nouveau: xserver-xorg-video-nouveau 1:1.0.16-1

[vinibali (vinibali)]

All the time the displays got relocked again, right after I unlocked it with my password. Some parts of the desktop can be seen for a moment in the meanwhile.
Right after that, it shows the "This screen is locked" message with a padlock and turns into black and shows me the errors.
So for the first time, I can't unlock the display successfully, but for the 2nd time it works.
Apart from that it also happens, that I won't get back the login screen after the session has been locked. Super strange.

[Daniel van Vugt (vanvugt)]

Sounds like this isn't a crash (unless you find files in /var/crash). Can you provide a better summary of the problem than "Xorg crash"?

[vinibali (vinibali)]

That title was coming from the Apport report.
If you look at the dmesg logs, you can see the possible problems.
Do you think if I can help the debugging with something else?

[vinibali (vinibali)]

https://launchpadlibrarian.net/666927057/CurrentDmesg.txt

[Daniel van Vugt (vanvugt)]

This bug is probably fixed already in a newer release so please try:
https://ubuntu.com/download/desktop

Note you have the old Ubuntu 20.04 installed whereas the latest long term support release is 22.04. I know they sound similar.

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

This change was made by a bot.

[vinibali (vinibali)]

As this is a company machine, I need to stick to the current release.

[Daniel van Vugt (vanvugt)]

You're not on the current release. You're using an old release :)

[vinibali (vinibali)]

Okay, so I had to stick to an old release, now I'm on the 22.04 :)
In the past months there were many times when this bug has gone literally for weeks and than came back again without any reason.
I'll update this ticket if this happens again.

[Daniel van Vugt (vanvugt)]

Using 20.04 is certainly supported for some years yet (https://wiki.ubuntu.com/Releases) but:

 * Some people don't notice the difference in the version numbers and accidentally install the older version; and
 * Plenty of bugs get fixed without official tracking so trying the most recent long term release is certainly recommended.

I think you will enjoy 22.04 more too.

[vinibali (vinibali)]

I know it's still supported and I got your point. This the reason why companies love to stick to an LTS version anyways :)
I feel this makes those LTS versions much harder to maintain.

[Daniel van Vugt (vanvugt)]

If you are using Ubuntu 22.04 then please try the latest supported kernel:

  sudo apt install linux-generic-hwe-22.04

[vinibali (vinibali)]

Done, I'm testing this now!
I'll do the feedback, thanks.

[vinibali (vinibali)]

Hello Daniel!
I'm having errors with another docking station - which was working fine so far for months - it seems, the display manager is not handling the disconnect and reconnects.
https://bugs.launchpad.net/ubuntu/+source/xorg/+bug/2032905

[vinibali (vinibali)]

I reconnect the USB-C cable to the hub (Dell P2419HD), the display is identified, the desktop is being rearranged but there is no picture on that.
This was definitely working fine in this docking station before.

[vinibali (vinibali)]

[16580.286522] i915 0000:00:02.0: [drm] *ERROR* [ENCODER:325:DDI TC1/PHY TC1][DPRX] Failed to enable link training
[16587.708453] i915 0000:00:02.0: [drm] *ERROR* [ENCODER:325:DDI TC1/PHY TC1][DPRX] Failed to enable link training
[16589.071651] i915 0000:00:02.0: [drm] *ERROR* CPU pipe B FIFO underrun: transcoder,
[16978.104729] i915 0000:00:02.0: [drm] *ERROR* Failed to get ACT after 3000ms, last status: 00
[16981.138068] i915 0000:00:02.0: [drm] *ERROR* Failed to get ACT after 3000ms, last status: 00
[17179.530073] i915 0000:00:02.0: [drm] *ERROR* Failed to get ACT after 3000ms, last status: 01
[17182.640462] i915 0000:00:02.0: [drm] *ERROR* Failed to get ACT after 3000ms, last status: 01

[Launchpad Janitor (janitor)]

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

[vinibali (vinibali)]

Hello Daniel.
I've filled two new bugs, which possible related to this one.
https://bugs.launchpad.net/ubuntu/+source/xorg/+bug/2036160
https://bugs.launchpad.net/ubuntu/+source/xorg/+bug/2036159
I identified some new type of behaviors with this problem:
* this latest kernel is not always able to print the kernel messages to all three monitors, sometimes I need to restart the OS while booting to see the boot on the screen (noquiet text)
* if all three monitors are showing the boot, the login manager not shows the anything on the DP-2-2-2 output, which is connected with a DP-DP cable from the docking station.
** the native HDMI from the notebook and the native HDMI from the docking station works mostly all the time
* if all three monitors are showing the boot and the login manager can be seen on all theree monitors, the display manager will show the screen on all three if the $HOME/.config/xfce4/xfconf/xfce-perchannel-xml/displays.xml file contains only the displays for that setup
** if some fallback or other displays are in the config file, the DP-2-2-2 output remain black. however xrandr shows that it's configured and used. I can only get this work if I reconnect the DP cable

If works, that it's mostly flawless, but if the kernel boot or login manager or desktop cannot be seen - as I mentioned in the list - only the reconnect works.
I hope it helps you to understand this situation better.

[vinibali (vinibali)]

A new bug has been added to this story:
https://bugs.launchpad.net/ubuntu/+source/xorg/+bug/2038282
This case the original $HOME/.config/xfce4/xfconf/xfce-perchannel-xml/displays.xml remained in place, however the DP-2-2-2 didn't show any screen after I log in the login manager.