Re: [PATCH v2 1/6] PCI/ASPM: Add locked helper for enabling link state

From: Bjorn Helgaas
Date: Thu Dec 14 2023 - 12:28:54 EST


On Wed, Dec 13, 2023 at 03:39:24PM -0800, David E. Box wrote:
> On Wed, 2023-12-13 at 14:45 -0600, Bjorn Helgaas wrote:
> ...

> > I'd be shocked if Windows treated the BIOS config as a "do not exceed
> > this" situation, so my secret hope is that some of these "broken"
> > devices are really caused by defects in the Linux ASPM support or the
> > driver, and that we can fix them if we find out about them.
> >
> > But I have no details about any of these alleged broken devices, so
> > it's hard to make progress on them.  
>
> I don't have a sense of the scope either. But I could see BIOS not
> enabling features that would provide no added power savings benefit.
> We use ASPM to manage package power. There are Intel devices that
> certainly don't require L1SS for the SoC to achieve the deepest
> power savings. L1 alone is fine for them. I don't know what the test
> coverage is for unenabled features. I've sent these questions to
> our BIOS folks.

Once upon a time there was a push to make it so firmware only had to
enumerate boot and console devices and it could skip enumeration and
configuration of other devices. But I don't think we've made much
progress on that, at least for x86, possibly because Linux depends so
much on BIOS resource assignment. IMO that's a Linux deficiency.

Bjorn