Screen freeze when plug 4K DP monitor on Dell WD19TB docking

changed title from Screen freeze when plug 4K DP monitor through Dell WD19TB docking to Screen freeze when plug 4K DP monitor on Dell WD19TB docking

changed the description

added Community platform: ADL_P labels

added Partner label

@kaichuan.hsieh, can you please try out disabling psr and see if you can still reproduce observed flickering or the freeze? This can be done by adding i915.enable_psr=0 into kernel boot parameters.

Yes, disable PSR fixes the flickering issue. And the freeze is somehow hard to reproduce, when it happens, you can reproduce it many times in a row, but now it maybe in another state, so I can't tell if this also fixes the freeze issue.

BTW, I found drm-intel-next 05-28 fixes the flickering issue https://kernel.ubuntu.com/~kernel-ppa/mainline/drm-intel-next/2022-05-28/

Ok, flickering is fixed by my patches here:

https://patchwork.freedesktop.org/series/103659/

I was just a bit confused, because I thought drm-tip 20220526 is supposed to have those already?

With more recent drm-tip containing my patches: Are you still able to reproduce the freeze?

Also not necessarily related to this issue at all but make sure all FW's are loaded successfully. I see DMC has been loaded but GUC is not: May 27 06:12:29 ubuntu kernel: i915 0000:00:02.0: Direct firmware load for i915/adlp_guc_70.1.1.bin failed with error -2 May 27 06:12:29 ubuntu kernel: i915 0000:00:02.0: [drm] GuC firmware i915/adlp_guc_70.1.1.bin: fetch failed with error -2 May 27 06:12:29 ubuntu kernel: i915 0000:00:02.0: [drm] GuC firmware(s) can be downloaded from https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tree/i915

Also provide all logs indicated at: https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs

Has WD19TB latest FW updates? Are you able to try with WD22TB4?

System architecture: Linux u-XPS-9320 5.18.0-051800rc2drmintelnext20220528-generic #202205280203 SMP PREEMPT_DYNAMIC Sat May 28 02:07:30 UTC 2022 x86_64 x86_64 x86_64 GNU/Linux
Kernel version: 5.18.0-051800rc2drmintelnext20220528-generic
Linux distribution: Ubuntu Jammy
Display connector: eDP(3840x2400 60Hz) + Dock(DP:3440x2440 144Hz & HDMI:3440x1440 100Hz)
Docking station: Dell WD22TB4(update all the firmware today by fwupdmgr update)

Plug in the dock at 23s, un-plug the dock at around 48s When the dock is plugged, the screen freeze, keyboard and touchpad have no function, can't type in and can't move. After the docker is un-plugged, keyboard and touchpad work again. When the screen is freeze, I can ssh into the machine and found the gnome-shell occupy CPU 100%. Attached the perf record/report screenshot with debug symbols.

dmesg_5.18.0-051800rc2drmintelnext20220528-generic.log

About the freeze issue:

I see the following errors in the log, is the BIOS up-to-date?:

May 27 06:12:29 ubuntu kernel: UBSAN: array-index-out-of-bounds in /home/kernel/COD/linux/drivers/acpi/acpica/dswexec.c:401:12
May 27 06:12:29 ubuntu kernel: index -1 is out of range for type 'acpi_operand_object *[9]'
May 27 06:12:29 ubuntu kernel: CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.18.0-051800drmtip20220526-generic #202205260203
May 27 06:12:29 ubuntu kernel: Hardware name: Dell Inc. XPS 9320/, BIOS 1.3.0 04/19/2022
...
May 27 06:12:31 ubuntu kernel: ACPI BIOS Error (bug): Could not resolve symbol [\_TZ.ETMD], AE_NOT_FOUND (20211217/psargs-330)
May 27 06:12:31 ubuntu kernel: 
May 27 06:12:31 ubuntu kernel: No Local Variables are initialized for Method [_OSC]
May 27 06:12:31 ubuntu kernel: 
May 27 06:12:31 ubuntu kernel: Initialized Arguments for Method [_OSC]:  (4 arguments defined for method invocation)
May 27 06:12:31 ubuntu kernel:   Arg0:   0000000083f34cc5 <Obj>           Buffer(16) 5D A8 3B B2 B7 C8 42 35
May 27 06:12:31 ubuntu kernel:   Arg1:   000000005ec0c6a9 <Obj>           Integer 0000000000000001
May 27 06:12:31 ubuntu kernel:   Arg2:   00000000dd5c12bc <Obj>           Integer 0000000000000002
May 27 06:12:31 ubuntu kernel:   Arg3:   000000001d54b6f1 <Obj>           Buffer(8) 00 00 00 00 05 00 00 00
May 27 06:12:31 ubuntu kernel: 
May 27 06:12:31 ubuntu kernel: ACPI Error: Aborting method \_SB.IETM._OSC due to previous error (AE_NOT_FOUND) (20211217/psparse-529)

@acelan , was the attached dmesg.log from the hang case? I can see there one connect and disconnect sequence at 06:13:40 , without any obvious issues. Is the machine still accessible?

Can you reproduce the issue plugging the hub without any display connected to it? If not, could you try if the problem reproduces with i915.enable_dp_mst=0 ?

Yes, that's the latest BIOS we got
Yes, the log is from the hang case. The machine is currently on my hand. Kaichuan Hsieh will get another machine to reproduce this issue today.
Looks like i915.enable_dp_mst=0 fixes the issue. Boot up with this parameter 3 times, can't reproduce the issue, and encounter the issue the fourth time reboot without this parameter.

Tried i915.enable_dp_mst=0 on my Precision 5750 with Ubuntu 22.04 + 5.18 drm-intel-next 05-28, issue behavior changes to: 1, built in 4k panel still freezes 2, external 4K monitor changes to flicking instead of freeze

Fail rate: 100%Dmesg_Precision_5750_5.18_drmtip.txt

@acelan , could you provide the no MST dmesg log as well (unless you notice the issue is still reproducible with that)? Also with MST enabled at the freeze state are you able to stop the GUI (systemctl stop gdm, or killall Xorg) note if you get back the framebuffer console, and provide the dmesg log afterwards (marking somehow the timestamp in dmesg when the freeze happened, and also when you got back the FB console).

@sunnycrown , not sure if the two issues have the same root cause, you have an Nvidia card, while @acelan doesn't. Could you provide a drm.debug=0x10e log after the freeze and - if the machine is still accessible - after an attempt to stop the GUI and recover an FB console (marking the timestamps in dmesg for the freeze and recovery)?

dmesg_disable_dp_mst.log

plug in dock at 36s
unplug the dock at 56s

dmesg_kill_gnome-shell.log

plug in dock at 33s
totally stuck at 52s
kill gnome-shell at 88s and got blank screen
switch to vt3 at 121s and login, no further dmesg when login

Dock WD19 TBT Plug: 44s Dock WD19 TBT Un-Plug: 60s @ideak btw, what's the side effect with parameter "i915.enable_dp_mst=0"? Thanks. dmesg_Precision_5750_dpmst_0_drmdebug.txt

In my Precision 5750, on top of Ubuntu 22.04 image, after installed NV graphic driver via "sudo apt install nvidia-driver-510", without dp_mst disable, it works with 4K 60hz external monitor.

mentioned in issue #4343 (closed)

@acelan , after plugin two outputs on the hub are detected and modeset without any errors in the log:

[   38.971533] i915 0000:00:02.0: [drm:intel_dump_pipe_config [i915]] [CRTC:131:pipe B] enable: yes [modeset]
[   38.971892] i915 0000:00:02.0: [drm:intel_dump_pipe_config [i915]] adjusted mode: "3440x1440": 85 458500 3440 3488 3520 3600 1440 1443 1453 1499 0x48 0x9
[   38.972452] i915 0000:00:02.0: [drm:intel_dump_pipe_config [i915]] [CRTC:182:pipe C] enable: yes [modeset]
[   38.972799] i915 0000:00:02.0: [drm:intel_dump_pipe_config [i915]] adjusted mode: "3440x1440": 100 543500 3440 3488 3520 3600 1440 1443 1447 1510 0x48 0x9

Not sure what are the above two output streams if you have only one monitor plugged to the hub.

Is the eDP panel also frozen and can you check the state of outputs in the frozen state from an ssh console:

export DISPLAY=:0
xrandr

and try to turn off all outputs and then modeset them one-by-one with various supported modes:

xrandr --output <output> --off

for each of the outputs you get from xrandr's output state list above, and then

xrandr --output <output> --mode <mode>

where mode is one of the mode listed for each output in the list above.

Yes, there are 2 outputs connected to the dock, but the 2 outputs are to the same monitor with different type of connector(DP and HDMI). This is much easier to reproduce this issue.

I can see 2 monitors are connected xrandr.log
Those outputs are

u@u-XPS-9320:~$ xrandr --output XWAYLAND0 --off
X Error of failed request:  BadMatch (invalid parameter attributes)
  Major opcode of failed request:  139 (RANDR)
  Minor opcode of failed request:  7 (RRSetScreenSize)
  Serial number of failed request:  36
  Current serial number in output stream:  37
u@u-XPS-9320:~$ xrandr --output XWAYLAND1 --off
u@u-XPS-9320:~$ xrandr --output XWAYLAND2 --off
X Error of failed request:  BadMatch (invalid parameter attributes)
  Major opcode of failed request:  139 (RANDR)
  Minor opcode of failed request:  7 (RRSetScreenSize)
  Serial number of failed request:  34
  Current serial number in output stream:  36
u@u-XPS-9320:~$ xrandr --output XWAYLAND1 --mode 3440x1440

This is more like a gnome shell issue I'm afraid of. When the screen freeze, meta + l makes the screen dim(other hotkeys do not work) and then touch the touchpad leads me to gdm to login. Touchpad and keyboard work in gdm. And then login to desktop, keyboard and touchpad stop working again, and the screen looks like freeze.

@acelan , yes scanout seems to be okay then, so looks more like a user space issue.

You could also try unplugging one of the monitors before stopping the GUI, to see if you get back a working FB console (the modeset with 2 monitors fails then for an unrelated BW issue).

@sunnycrown , disabling MST means that only one output stream will be supported for each TBT DP tunnel, so probably not all connectors on the hub can be used at the same time.

@acelan Any update on this, as Imre mentioned it is more likely an userspace issue.

https://bugs.launchpad.net/oem-priority/+bug/1976204

Closing this issue.

closed

Screen freeze when plug 4K DP monitor on Dell WD19TB docking

Child items 0

Activity

Admin message

Screen freeze when plug 4K DP monitor on Dell WD19TB docking

Activity