Bug 206255 - kernel constantly changing thermal zones on yoga c940
Summary: kernel constantly changing thermal zones on yoga c940
Status: CLOSED INSUFFICIENT_DATA
Alias: None
Product: ACPI
Classification: Unclassified
Component: Power-Thermal (show other bugs)
Hardware: Intel Linux
: P1 normal
Assignee: Zhang Rui
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2020-01-19 05:06 UTC by Samuel Hsieh
Modified: 2021-07-01 08:54 UTC (History)
1 user (show)

See Also:
Kernel Version: 5.3.0-28-generic
Subsystem:
Regression: No
Bisected commit-id:


Attachments
sudo udevadm monitor (if started linux directly) (41.79 KB, text/plain)
2020-01-19 05:06 UTC, Samuel Hsieh
Details
grep . /sys/class/thermal/thermal*/* (33.44 KB, application/octet-stream)
2020-06-29 22:39 UTC, Samuel Hsieh
Details
grep . /sys/class/thermal/thermal*/* (13.01 KB, text/plain)
2020-11-28 01:45 UTC, Samuel Hsieh
Details
sudo udevadm monitor (43.18 KB, text/plain)
2020-11-28 01:50 UTC, Samuel Hsieh
Details
sci-before (98.12 KB, application/octet-stream)
2021-03-20 02:44 UTC, Samuel Hsieh
Details
before (98.12 KB, application/octet-stream)
2021-03-23 02:11 UTC, Samuel Hsieh
Details
after (98.12 KB, application/octet-stream)
2021-03-23 02:12 UTC, Samuel Hsieh
Details
udev (17.95 KB, text/plain)
2021-03-27 03:14 UTC, Samuel Hsieh
Details
acpidump (1.49 MB, text/plain)
2021-03-30 03:13 UTC, Samuel Hsieh
Details

Description Samuel Hsieh 2020-01-19 05:06:37 UTC
Created attachment 286879 [details]
sudo udevadm monitor (if started linux directly)

I discovered this while debugging for another issue. Are there any suggestions for next steps? By the way, this doesn't seem to happen if I boot into Windows first. Is it possible that Lenovo Ventage is enforcing some sort of CPU profile that Ubuntu wipes out during shut down?
Comment 1 Zhang Rui 2020-06-29 11:31:58 UTC
please attach the output of "grep . /sys/class/thermal/thermal*/*"
Comment 2 Zhang Rui 2020-06-29 11:32:25 UTC
what kernel are you using? does the problem still exist in latest upstream kernel?
Comment 3 Samuel Hsieh 2020-06-29 22:39:15 UTC
Created attachment 289953 [details]
grep . /sys/class/thermal/thermal*/*
Comment 4 Samuel Hsieh 2020-06-29 22:46:42 UTC
(In reply to Zhang Rui from comment #2)
> what kernel are you using? does the problem still exist in latest upstream
> kernel?

5.4.0-40-generic #44-Ubuntu SMP Tue Jun 23 00:01:04 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux

This is not the latest kernel,
Comment 5 Zhang Rui 2020-11-18 14:57:18 UTC
(In reply to Samuel Hsieh from comment #3)
> Created attachment 289953 [details]
> grep . /sys/class/thermal/thermal*/*

It is not plain/txt format, right? I can not read it.
can you confirm if the problem still exists in latest upstream kernel?
Comment 6 Zhang Rui 2020-11-18 14:57:36 UTC
(In reply to Samuel Hsieh from comment #3)
> Created attachment 289953 [details]
> grep . /sys/class/thermal/thermal*/*

It is not plain/txt format, right? I can not read it.
can you confirm if the problem still exists in latest upstream kernel?
Comment 7 Samuel Hsieh 2020-11-28 01:45:52 UTC
Created attachment 293853 [details]
grep . /sys/class/thermal/thermal*/*

I'm on 5.8.0-31-generic #33-Ubuntu now. Still see these change events.
Comment 8 Samuel Hsieh 2020-11-28 01:50:28 UTC
Created attachment 293855 [details]
sudo udevadm monitor

update both output after ubuntu 20.10 upgrade
Comment 9 Zhang Rui 2021-03-19 07:18:35 UTC
It seems that the thermal event keeps on firing
please attach the sci-before and sci-after file after running following command
grep . /sys/firmware/acpi/interrupts/* > sci-before; sleep 10; "grep . /sys/firmware/acpi/interrupts/*" both  > sci-after

Please make sure your test are done with latest vanilla kernel.
Comment 10 Samuel Hsieh 2021-03-20 02:44:48 UTC
Created attachment 295959 [details]
sci-before
Comment 11 Samuel Hsieh 2021-03-20 02:53:44 UTC
sci-after is empty
probably due to this error
bash: grep . /sys/firmware/acpi/interrupts/*: No such file or directory
Comment 12 Zhang Rui 2021-03-22 12:30:29 UTC
There must be something wrong,
please make sure you run the same command both before and after the sleep, we just need to redirect them to different files
Comment 13 Samuel Hsieh 2021-03-23 02:11:36 UTC
Created attachment 296005 [details]
before
Comment 14 Samuel Hsieh 2021-03-23 02:12:47 UTC
Created attachment 296007 [details]
after
Comment 15 Samuel Hsieh 2021-03-23 02:19:15 UTC
Oh, now I understand what you were trying to do, I removed the quotes and "both" from the command. This ran without issue.

grep . /sys/firmware/acpi/interrupts/* > sci-before; sleep 10; grep . /sys/firmware/acpi/interrupts/* > sci-after

Please let me know if I misunderstood your instructions.
Comment 16 Zhang Rui 2021-03-25 03:45:50 UTC
$ diff -pNur sci-before sci-after 
--- sci-before	2021-03-25 11:40:09.486037789 +0800
+++ sci-after	2021-03-25 11:40:07.223054124 +0800
@@ -114,7 +114,7 @@
-/sys/firmware/acpi/interrupts/gpe6E:   56720  EN     enabled      unmasked
+/sys/firmware/acpi/interrupts/gpe6E:   56792  EN     enabled      unmasked

72 SCI in 10 seconds, this is not consistent with the number of thermal change events you've seen in udevadmin.

Can you please first check if the problem still exists in the latest upstream kernel or not? because we have fixed a similar problem recently.
If not, please attach the acpidump output.
Comment 17 Samuel Hsieh 2021-03-27 03:14:11 UTC
Created attachment 296083 [details]
udev

still seems like a lot of change events
Comment 18 Zhang Rui 2021-03-29 05:31:27 UTC
what kernel version you're using?
and please attach the acpidump output.
Comment 19 Samuel Hsieh 2021-03-30 03:13:15 UTC
Created attachment 296143 [details]
acpidump

I'm running 5.8.0-49-generic.

Not sure how to compile the latest kernel though.
Comment 20 Zhang Rui 2021-06-03 03:21:41 UTC
please follow the guidance here
https://phoenixnap.com/kb/build-linux-kernel
Comment 21 Zhang Rui 2021-07-01 08:54:20 UTC
Bug closed as there is no response from the bug reporter.
Please feel free to open it if the problem still exists with latest upstream kernel.

Note You need to log in before you can comment on or make changes to this bug.