Bug 1185008

Summary: CPU tainted in dmesg since 5.11
Product: [openSUSE] openSUSE Tumbleweed Reporter: Vitaly Bakulev <real86bitals>
Component: KernelAssignee: openSUSE Kernel Bugs <kernel-bugs>
Status: NEW --- QA Contact: E-mail List <qa-bugs>
Severity: Major    
Priority: P5 - None CC: jslaby, real86bitals, tiwai
Version: Current   
Target Milestone: ---   
Hardware: x86-64   
OS: openSUSE Tumbleweed   
Whiteboard:
Found By: --- Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---
Attachments: dmesg output
full dmesg output 5.12.3-1
Soft lockup during shutdown 1
dmesg July 22

Description Vitaly Bakulev 2021-04-20 08:00:21 UTC
Created attachment 848532 [details]
dmesg output

Hello,

Tumbleweed 20210417
Kernel 5.11.12-1.1 (any 5.11.* in fact)

Every time I boot with a 5.11 kernel, I experience a lot of problems. Touchpad is lagging, Wi-Fi not connecting and so on. But everything is fine on 5.10.16-1.3.

Dmesg is full of this:

[506.529437] CPU: 4 PID: 7993 Comm: kworker/4:0 Tainted: G S      W         5.11.12-1-default #1 openSUSE Tumbleweed

More in the attached file.
Comment 1 Takashi Iwai 2021-04-20 13:23:07 UTC
The taint flag is likely a red herring.  You might have done some MSR write from user-space?  That would taint the kernel ('S' flag).
Comment 2 Vitaly Bakulev 2021-04-20 17:44:39 UTC
You mean like described in this issue: https://github.com/kitsunyan/intel-undervolt/issues/55 ?

This seems to be my case as I do use that project, but that never caused severe perfomance issues (making PC unusable) I am experiencing on 5.11.
Comment 3 Jiri Slaby 2021-04-21 05:07:33 UTC
Could you attach fill dmesg? As the excerpt you provided is very past the original issue.
Comment 4 Jiri Slaby 2021-04-21 05:07:49 UTC
(In reply to Jiri Slaby from comment #3)
> Could you attach fill dmesg? As the excerpt you provided is very past the

"full" I mean.
Comment 5 Vitaly Bakulev 2021-05-08 13:04:10 UTC
Sorry for the lack of response, it's been a crunchy time.
Seems like everything is fine on 5.12 for a couple of days though.
Comment 6 Vitaly Bakulev 2021-05-17 11:35:54 UTC
And now with the 5.12.3-1-default kernel it's really bad again. Touchpad is unusable, the whole system is stuttering every few seconds.
Attaching the full dmesg output.
Comment 7 Vitaly Bakulev 2021-05-17 11:36:19 UTC
Created attachment 849388 [details]
full dmesg output 5.12.3-1
Comment 8 Takashi Iwai 2021-05-18 12:51:27 UTC
Could you check with the kernel package in OBS Kernel:stable repo?
  http://download.opensuse.org/repositories/Kernel:/stable/standard/
Comment 9 Vitaly Bakulev 2021-05-18 16:22:19 UTC
So far way smoother on startup and in general. Mouse cursor still lags occasionally, but overall a lot better than mainline kernel.
I will add more feedback after a couple days of running and sleeping.
Comment 10 Vitaly Bakulev 2021-05-19 08:34:18 UTC
Experienced a very strange behaviour on 5.12.4-4.g2d71c6c-default just now.
Booted up, logged into Gnome, successfully launched terminal, launched my standard set of work software (Brave, Telegram, Discord, QOwnNotes, Evolution, 1Password, VSCode, VM Manager) and nothing of it started. Waited for about 2 minutes, still none of GUI apps I launched have started. Terminal commands didn't run either, everything just sort of freezed on the inside. But Gnome Shell seemed to be ok, I was able to type and click stuff. No idea what happened, and no logs because of dead terminal.
Rebooted again and everything is fine.
Comment 11 Vitaly Bakulev 2021-05-22 11:51:46 UTC
Created attachment 849582 [details]
Soft lockup during shutdown 1
Comment 12 Vitaly Bakulev 2021-05-22 11:57:41 UTC
I now get errors like in the photo I just attached every shutdown. It takes ages to poweroff. After boot, it feels like I randomly get either a borderline usable system or a complete mess. It can also change at any moment (sometime Wi-Fi just stops working. It is shown as active, but nothing from LAN or internet loads, it doesn't react to me trying to switch Wi-Fi off. Not to mention touchpad, it might lag a bit or the cursor can just freeze at one position and not react to either touches or clicks for several minutes.
Comment 13 Vitaly Bakulev 2021-07-16 14:17:18 UTC
Still occurs on 5.13.1-4.g72aabc2-default
If there are any other logs needed - please tell me where and how to find them.
Comment 14 Takashi Iwai 2021-07-16 14:20:48 UTC
Please give the log from the latest state.
Comment 15 Takashi Iwai 2021-07-16 14:23:04 UTC
Also check whether you get the same problem even after disabling intel-undervolt, too.  It's the only suspicious piece, so far.
Comment 16 Vitaly Bakulev 2021-07-22 15:41:49 UTC
Created attachment 851177 [details]
dmesg July 22

Disabling intel_undervolt made no difference.
Comment 17 Takashi Iwai 2021-07-26 13:52:52 UTC
Which problem are you referring to...?  I see no error there.