Lenovo Thinkpad P14s Gen3 amd Ryzen 7 PRO 6850U system freeze more than once a day

[__JEAN_FRANCOIS__]

Hello,

Having this computer for about 6 month now.
Tried Ubuntu 20.04, 23.04, now I'm back with 22.04 and many kernel versions.

I can't found a stable configuration. Very often, once a day in average, the system hangs and I have to hard reboot it.
I have plenty of kernel logs but most of the time they are not relevant because the last message may be hours old.

I added a question at lenovo forum, the logs can be found there but if they help I can upload them here too.
Url of the post on lenovo site: https://forums.lenovo.com/t5/Ubuntu/System-freeze-with-Ubuntu-22-04-to-23-04-on-P14s-gen3-AMD-version/m-p/5227346?page=4#6089001

I have some technical knowledge, so I can do some tests if that can help solving this issue.

Note: currently attached dmeg contains a OoPs, I'm no sure but I feel like this message is an early sign of future freeze. Hence the bug report.

Many thanks

ProblemType: Bug
DistroRelease: Ubuntu 22.04
Package: linux-image-5.15.0-82-generic 5.15.0-82.91
ProcVersionSignature: Ubuntu 5.15.0-82.91-generic 5.15.111
Uname: Linux 5.15.0-82-generic x86_64
ApportVersion: 2.20.11-0ubuntu82.5
Architecture: amd64
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC1: jf 2338 F.... pulseaudio
 /dev/snd/pcmC1D0p: jf 2338 F...m pulseaudio
 /dev/snd/controlC0: jf 2338 F.... pulseaudio
 /dev/snd/controlC2: jf 2338 F.... pulseaudio
CRDA: N/A
CasperMD5CheckResult: unknown
CurrentDesktop: ubuntu:GNOME
Date: Fri Sep 1 10:34:36 2023
DistributionChannelDescriptor:
 # This is the distribution channel descriptor for the OEM CDs
 # For more information see http://wiki.ubuntu.com/DistributionChannelDescriptor
 canonical-oem-sutton-focal-amd64-20220803-89+sutton-focal-amd64+X02
InstallationDate: Installed on 2023-06-02 (90 days ago)
InstallationMedia: Ubuntu 20.04 "Focal" - Build amd64 LIVE Binary 20220803-13:42
MachineType: LENOVO 21J5CTO1WW
ProcEnviron:
 TERM=alacritty
 PATH=(custom, no user)
 XDG_RUNTIME_DIR=<set>
 LANG=fr_FR.UTF-8
 SHELL=/bin/bash
ProcFB: 0 amdgpudrmfb
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-5.15.0-82-generic root=UUID=ad3cccf2-3072-4b29-84f0-ed7b41dd4de6 ro quiet splash vt.handoff=7
RebootRequiredPkgs: Error: path contained symlinks.
RelatedPackageVersions:
 linux-restricted-modules-5.15.0-82-generic N/A
 linux-backports-modules-5.15.0-82-generic N/A
 linux-firmware 20220329.git681281e4-0ubuntu3.18
SourcePackage: linux
UpgradeStatus: Upgraded to jammy on 2023-08-12 (19 days ago)
dmi.bios.date: 03/21/2023
dmi.bios.release: 1.35
dmi.bios.vendor: LENOVO
dmi.bios.version: R23ET65W (1.35 )
dmi.board.asset.tag: Not Available
dmi.board.name: 21J5CTO1WW
dmi.board.vendor: LENOVO
dmi.board.version: ThinkPad
dmi.chassis.asset.tag: No Asset Information
dmi.chassis.type: 10
dmi.chassis.vendor: LENOVO
dmi.chassis.version: None
dmi.ec.firmware.release: 1.28
dmi.modalias: dmi:bvnLENOVO:bvrR23ET65W(1.35):bd03/21/2023:br1.35:efr1.28:svnLENOVO:pn21J5CTO1WW:pvrThinkPadP14sGen3:rvnLENOVO:rn21J5CTO1WW:rvrThinkPad:cvnLENOVO:ct10:cvrNone:skuLENOVO_MT_21J5_BU_Think_FM_ThinkPadP14sGen3:
dmi.product.family: ThinkPad P14s Gen 3
dmi.product.name: 21J5CTO1WW
dmi.product.sku: LENOVO_MT_21J5_BU_Think_FM_ThinkPad P14s Gen 3
dmi.product.version: ThinkPad P14s Gen 3
dmi.sys.vendor: LENOVO

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

This change was made by a bot.

[Martin Randau (cmmrandau)]

Hello! I have had the same laptop for some time and have been battling with the same issues. According to my experience, the solution is to find a distro with a >6.3 kernel and set amd_pstate=active as boot parameters. Alternatively, if on an earlier kernel, you can try amdgpu.dcdebugmask=0x10 as boot parameter. I'm currently running Ubuntu 23.10 and it's very stable save for the occasional gnome shell crash which does not freeze or require restart. So far no gpu-related freezes. Only suspend and dim screen don't work in Wayland, but do in xorg, where, on the other hand, the touch pad three-finger swipe doesn't work.
BW, Martin

[__JEAN_FRANCOIS__ (jean-francois--)]

Hello Martin

Many thanks for your reply and the hope your message is bringing.

This morning I just decided to give another try to recent kernel versions, funny you suggested this ^^
Previously, I used the tool `mainline`, but I'd not much success (I experienced more crashes).
Today, I decided to go for https://github.com/zabbly/ and I'm now running a kernel 6.4.13 with ubuntu 22.04. I was not aware of the parameter `amd_pstate=active` so it's not enabled. I'll try this later. I want to give a chance to the 6.4 and change param one at a time.

With earlier kernels I tried `amdgpu.dcdebugmask=0x10` but it was not bullet proof for me... So since a moment I removed this flag.

Some questions, on your side did you experienced only system freezes or some other "strange bugs". Here I'm also having: Soft lockup, kernel NULL point dereference, kernel unable to handle page fault...
I've found those messages in the lots of log files I have now. From a user perspective I'd say I've noticed them because some firefox tab or some vscode windows were crashing.

Regards,
JF

[Martin Randau (cmmrandau)]

Hello JF
You can read about amd_pstate here: https://docs.kernel.org/admin-guide/pm/amd-pstate.html
and about the T14s gen 3 - which also has GPU-related crashes - here: https://wiki.archlinux.org/title/Lenovo_ThinkPad_T14s_(AMD)_Gen_3#GPU

I have not been looking at dmesg after each crash but can only conclude that with kernel 6.3 and above and amd_pstate=active, I'm not experiencing any problems. I've read somewhere that amd_pstate will be on by default in kernels >6.5 but until then we have to activate this feature by ourselves.

Good luck!

edit: as a side note I'm experiencing better battery management and CPU performance with amd_pstate=active. On idle it consume as little as 3.7 W and on performance mode the CPU is boosted to 4.2 GHz. I'm not certain all features are enabled without this boot parameter.

BW
Martin

[__JEAN_FRANCOIS__ (jean-francois--)]

Hello Martin,

Many thanks for the detailed information.
I'm gonna configure amd_pstate=active for the next reboot (I'll wait for few days if I faced no crash to see if this kernel version is intrinsically more stable).

Nice to learn that battery will last longer, it was a bit low to me.

Regards,
JF