Re: [PATCH 0/8] PCI PM: Handling of PCI devices wake-up functionality (rev. 4)

[Jesse Barnes <jbarnes@virtuousgeek.org>]

On Sunday, July 06, 2008 6:30 pm Rafael J. Wysocki wrote:
> Hi Jesse,
>
> This is a new (4th) iteration of the series of patches intended to rework
> the power management of PCI devices so that the handling of their wake-up
> functionality is consistent with the other PM operations and so that the
> user space can manage the wake-up functionality of PCI devices using the
> /sys/devices/.../power/wakeup interface regardless of whether the wake-up
> function is based on the native PCI mechanism (PME#), or it is handled by
> the platform (eg. ACPI).
>
> Apart from the ACPI and PCI changes necessary for this purpose, the series
> includes some clean-ups that are not strictly required, but make the code
> more straightforward and (IMHO) easier to follow.
>
> The patchset is on top of the linux-next branch of the PCI tree and is
> regarded as 2.6.27 material.

I just tested on one of my 965 based platforms.  Suspend/resume works as well 
as I'd expect normally, but if I use wake-on-lan (using the wakeup code from 
this patchset) the machine wakes up but never seems to finish resuming; it 
stays off the network with video off and the console is unresponsive.  I'll 
try some more tests (including older kernels to see if this is actually 
regression or not), but it would be good if other people could give this 
patch set some attention too.

Thanks,
Jesse