Re: Comparison between .28 and .29-rc3 for STR cycle

[Frans Pop <elendil@planet.nl>]

On Tuesday 03 February 2009, Frans Pop wrote:
> # Of the next 4 lines only the 3rd line also appears during normal
> boot.
>  +pci 0000:00:02.0: PME# disabled 
>  +pci 0000:00:02.1: PME# disabled
>  +pci 0000:00:03.0: PME# disabled
>  +pci 0000:00:03.2: PME# disabled

The 3rd device is also the only one that has a "PME# supported" line 
during normal boot.

Problem seems to be in drivers/pci/pci.c in pci_enable_wake(), which has:
        if (!enable || pci_pme_capable(dev, state)) {
                pci_pme_active(dev, enable);
                pme_done = true;
        }

So it's always called if enable is false, regardless of whether the device 
is PME# capable. Should a test for dev->pme_support be added somewhere?

It seems harmless, just inconsistent.