The Pull That Finally Fixes ASPM Power Regression

[phoronix]

Phoronix: The Pull That Finally Fixes ASPM Power Regression

Queued up in the PCI sub-system pull for the Linux 3.3 kernel is the ASPM re-work that provides a proper fix to the well known Linux kernel power regression that was talked about for much of 2011...

http://www.phoronix.com/vr.php?view=MTA0MTM

[Sidicas]

Technically, I think it's not really "fixed" since hardware makers are still putting out garbage BIOSs.. Rather it's just band-aided over, the same way Microsoft did it.

[Viper_Scull]

Where can i find the patch for kernel 3.2?

[Ragas]

Technically, I think it's not really "fixed" since hardware makers are still putting out garbage BIOSs.. Rather it's just band-aided over, the same way Microsoft did it.

Let's say it's fixed on the kernel side. If a vendor puts out faulty BIOSes .... can you blame Linux?

[DaemonFC]

Technically, I think it's not really "fixed" since hardware makers are still putting out garbage BIOSs.. Rather it's just band-aided over, the same way Microsoft did it.

It's actually the norm for Linux to work around Microsoft partner standards violations in BIOSes.

[ua=42]

Is a standard a standard if no one follows it?
Yes, it is ugly, but doing what Microsoft did is the only way to make Linux work on portable devices.

[DanL]

Is a standard a standard if no one follows it?

There is a specification and the most common way of implementing it (using Microsoft's BIOS tools) seems to be half-assed. That doesn't change the specification or make the minority who implemented it correctly "wrong" in some way. The good news is that the issue was resolved, and I'm sure the patch will be backported to a lot of long-term/stable kernels.
More good news is that the Phoronix Test Suite found it and generated traffic to this site.

[marcelosoaressouza]

3.2.5-stable review with ASPM Fixes: https://lkml.org/lkml/2012/2/3/409