Error: Unable To Determine Maximum Fan Speed For Card0! about amdgpu-pro-fans HOT 3 OPEN

FYI, you need to be using amdgpu kernel driver, what this script is meant for, for this to work.

@ivanaska Your card Vesuvius (appears to be re-branded Hawaii card) is a Sea Island GCN 1.1 second generation card which amdgpu, last I checked in 2017, had experimental support for. For this card, the default driver as of Linux kernel =<4.15 is still radeon. If you like to experiment with the amdgpu driver. Make sure your kernel is built with CONFIG_DRM_AMDGPU_CIK=y enabled and add amdgpu.cik_support=1 to the kernel command line via grub and blacklist radeon.

@miriampl Your card on the other hand, Cayman PRO, is a Northern Island TeraScale 3 (VLIW4) card which amdgpu (amdgpu supports only GCN cards) does not support.

Since it's more than a year since the last commit, alternatively may I suggest you guys give https://github.com/marazmista/radeon-profile a shot, which supports both radeon and amdgpu drivers, although I am not sure how functional fan control is for radeon driver.

from amdgpu-pro-fans.

I have the same issue
OS: Mint 18.3 sylvia
Kernel: x86_64 Linux 4.13.0-32-generic
CPU: Intel Core i7 CPU 860 @ 2.801GHz
product: Cayman PRO [Radeon HD 6950]
vendor: Advanced Micro Devices, Inc. [AMD/ATI]

head: cannot open '/sys/class/drm/card0/device/hwmon/hwmon0/pwm1_max' for reading: No such file or directory ./amdgpu-pro-fans.sh: linia 58: [: -gt: oczekiwano operatora jednoargumentowego Error: Unable To Determine Maximum Fan Speed For Card0!

from amdgpu-pro-fans.

Okay, I've just installed the latest BIOS. Version 5102 2019/03/06 which might have helped a little. There are still 3 cards installed (out of 8) and I've managed rocm-smi to see them all. Currently, I'm using the upstream kernel in Ubuntu 18.10, because the system failed to boot with 18.04.

One strange thing is that amdgpu is not loaded upon boot. I have to manually load the driver using sudo modprobe amdgpu to get things under /sys/class/drm/ populated. After that's done, rocm-smi reports all 3 cards. rocminfo and /opt/rocm/opencl/bin/x86_64/clinfo however fail with

malloc(): memory corruption
Aborted

Just running the console gives me error messages every minute or so, similar to that in this ROCK-Kernel-Driver issue.

Bad DLLP
aer_layer=Physical lLayer, aer_agent=Receiver ID

Full journalctl of the last boot is here.
last_boot.txt

Shall I post some screens of the BIOS settings available for PCI devices?

from amdgpu-pro-fans.