Bug 38082 - Fan sensor/controls not working in Thinkpad Edge E220s
Summary: Fan sensor/controls not working in Thinkpad Edge E220s
Status: RESOLVED WILL_NOT_FIX
Alias: None
Product: ACPI
Classification: Unclassified
Component: Power-Fan (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: acpi_power-fan
URL:
Keywords:
: 43048 (view as bug list)
Depends on:
Blocks:
 
Reported: 2011-06-23 14:54 UTC by Chow Loong Jin
Modified: 2024-09-10 21:04 UTC (History)
7 users (show)

See Also:
Kernel Version: 3.1
Subsystem:
Regression: No
Bisected commit-id:


Attachments
E555 DSDT (450.67 KB, text/x-csrc)
2024-09-05 12:41 UTC, Lorenzo M.
Details

Description Chow Loong Jin 2011-06-23 14:54:23 UTC
Eventhough I can hear the fan running, every userspace interface I have poked seems to say it's at 0RPM.

/proc/acpi/ibm:
status:         disabled
speed:          0
level:          0

/sys/devices/platform/thinkpad_hwmon/fan1_input:
0
Comment 1 Enrico Tagliavini 2012-07-27 16:14:30 UTC
This is true also for Thinkpad edge 530. They changed once again the EC config it seems. This bug is likely a dup of bug #43048 

If someone want a tester or would just mind to explain me the basis of this stuff so I can just write it myself, make a whistle.
Comment 2 Alan 2012-08-24 14:54:45 UTC
*** Bug 43048 has been marked as a duplicate of this bug. ***
Comment 3 Tommy Trussell 2015-02-27 03:17:18 UTC
Also affects Thinkpad E540.
Comment 4 Lorenzo M. 2024-09-01 20:45:06 UTC
I confirm the same bug on E555.

Fan speed is always reported as zero. Writing fan speed (with acpi on) is not effective, as the fan remains at minimum speed no matter the settings. This lead to CPU overheating and throttling.

At power up, fan runs at full speed briefly; as soon as kernel is loaded, the fan slows down to minimum speed and does not change speed any more.

With acpi=off, the fan is effective.
Comment 5 Lorenzo M. 2024-09-02 09:57:39 UTC
Please consider reassigning to acpi_power-fan@kernel-bugs.osdl.org
Comment 6 Artem S. Tashkinov 2024-09-05 12:15:03 UTC
I'm closing this bug report.

Fan speed reporting is more often than not proprietary and speed control has not been a thing for ages now (except for power profiles in modern laptops which are again proprietary).
Comment 7 Lorenzo M. 2024-09-05 12:41:09 UTC
Created attachment 306822 [details]
E555 DSDT
Comment 8 Lorenzo M. 2024-09-05 12:50:44 UTC
I sincerely object to closing this bug, for several reasons.

First, ist' not a matter of fan active control (which I recognise should not be used): the current kernel implementation breaks automatic fan management by EC, which otherwise works with acpi=off. Moreover, it causes fan monitoring not to work at all.

Second, Thinkpad ACPI driver is being actively maintained and is receiving patched for very related subjects, a this one from just a couple weeks ago introducing support on a very similar model:

https://patchwork.kernel.org/project/platform-driver-x86/patch/20240816141228.134529-1-kontakt@matthias-fetzer.de/

By the way this patch introduces support for FANG and FANW methods, which are also implemented in E555 DSDT (see attachment). So methods are recognised and used by the kernel, but the implementation fails on E555 (maybe because of model-specific arguments passed to the methods). If the patch is accepted in the mainline kernel, this would be the subject for a new bug report.

The patched module is loaded correctly and the related hierarchies in proc and sysfs are populated, but trying to monitor or control the fan is ineffective or causes errors:

    # cat /proc/acpi/ibm/fan
    cat: /proc/acpi/ibm/fan: No such device or address
    # cat /sys/class/hwmon/hwmon5/name
    thinkpad
    # echo 1 > /sys/class/hwmon/hwmon5/pwm1_enable
    # cat /sys/class/hwmon/hwmon5/pwm1
    109
    # echo 255 > /sys/class/hwmon/hwmon5/pwm1
    # cat /sys/class/hwmon/hwmon5/pwm1
    109
    # cat /sys/class/hwmon/hwmon5/fan1_input
    cat: /sys/class/hwmon/hwmon5/fan1_input: No such device or address
Comment 9 Artem S. Tashkinov 2024-09-10 21:04:14 UTC
Whether this bug report is open or not does not change a thing.

Either there are developers working on it or not.

And considering that this laptop model may have the whole ten people in world still using it, I guess developers' resources could be better spent on supporting modern hardware that is actually used by a lot of people.

I also very much doubt this model came with Linux (support). If it didn't, then expecting people to workaround the broken BIOS behavior is a tall order.

This laptop is also basically unusable for modern web.

I have a nine year old laptop that is twice as fast that I couldn't use any longer because of how slow it was at opening websites.

Again, the status of this bug report in inconsequential.

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