Fan control stops working on resume for Dell Precision

[Jason Gunthorpe]

This system is a Dell Precision Tower 3420 with the latest BIOS 2.13.1, kernel version is 5.3.0-26-generic (HWE, bionic)

Freshly booting the system and doing something CPU intensive, like compiling a kernel, cause the CPU fan to ramp up as expected and the CPU temperatures remain acceptable.

After suspending (ie Alt-clicking the power button in Unity) and then resuming, the CPU fan runs at a permanent low frequency and does not ramp up. Doing something like a kernel build causes the CPU to overheat and logs messages to dmesg.

Investigating I see that the dell_smm_hwmon driver has a 'temp2' input that reflects the CPU temperature. Upon power up the value works properly, after resume it returns -ENODATA. The other sensor for CPU temperature (thermal_zone3/temp, x86_pkg_temp) still works correctly.

Blacklisting the dell_smm_hwmon driver has no impact

Dell Precisions are listed as compatible with Ubuntu, and this system was purchased with Ubuntu pre-installed from Dell, so it is surpsing something so basic isn't working?

ProblemType: Bug
DistroRelease: Ubuntu 18.04
Package: linux-image-5.3.0-26-generic 5.3.0-26.28~18.04.1
ProcVersionSignature: Ubuntu 5.3.0-26.28~18.04.1-generic 5.3.13
Uname: Linux 5.3.0-26-generic x86_64
ApportVersion: 2.20.9-0ubuntu7.9
Architecture: amd64
CurrentDesktop: ubuntu:GNOME
Date: Mon Jan 27 15:48:39 2020
InstallationDate: Installed on 2018-08-04 (540 days ago)
InstallationMedia: Ubuntu 18.04.1 LTS "Bionic Beaver" - Release amd64 (20180725)
SourcePackage: linux-signed-hwe
UpgradeStatus: No upgrade log present (probably fresh install)

[Jason Gunthorpe (jgunthorpe)]

This still happens on a fresh Focal install using linux-image-5.4.0-26-generic version 5.4.0-26.30