Arch Linux

Lockheed

Member

Registered: 2010-03-16

Posts: 1,521

Going around impossibility of enabling ASPM

Adding pcie_aspm=force to kernel line doesn't work (any longer?) on recent kernels. This was the only known to me way of enabling it  (please, do tell, if you know another one).

The result is the unavoidable

$ dmesg | grep ASPM
[    0.000000] PCIe ASPM is forcibly enabled
[    0.094612] ACPI FADT declares the system doesn't support PCIe ASPM, so disable it

I can also confirm it with

$ cat /sys/module/pcie_aspm/parameters/policy
default [performance] powersave

which means it's disabled.

However, I can change this value with

$ echo default > /sys/module/pcie_aspm/parameters/policy
$ cat /sys/module/pcie_aspm/parameters/policy
[default] performance powersave 

The question is, does it actually enable ASPM, or does it just change the text value, leaving ASPM still inactive.

ANOKNUSA

Member

Registered: 2010-10-22

Posts: 2,141

Re: Going around impossibility of enabling ASPM

This is a blind guess, but I'd say you're running an Intel CPU, in which case using "pcie_aspm=force" for power management is waaaaaaaaaaayyyyy outdated.  The intel_pstate driver + cpupower and cpupower.service are intended to allow the performance governor to handle all power management more efficiently, including (I think) with PCIE devices.  But then, I could be totally wrong about that (and your situation) as well.

---

Linux != Windows  ::  Ask Questions the Smart Way  ::  Configs

mich41

Member

Registered: 2012-06-22

Posts: 796

Re: Going around impossibility of enabling ASPM

I'd try "powersave".

Lockheed

Member

Registered: 2010-03-16

Posts: 1,521

Re: Going around impossibility of enabling ASPM

@ANOKNUSA,
Good guess, it is Intel Core 2 Duo.
However, I've been unable to find any info regarding PCIE functions of cpupower or intel_pstate. I already have TLP to manage my governor and frequencies.

@mich41,
I have moved away from powersave in favour of TLP