On Thursday, 11 of December 2008, Alan Stern wrote:
> On Wed, 10 Dec 2008, Alan Stern wrote:
>
> > On Wed, 10 Dec 2008, Rafael J. Wysocki wrote:
> >
> > > > > Well, it looks like you forgot to actually call pci_save_power_state(). :-)
> > > >
> > > > I didn't forget it -- in my tree there is no such function! :-)
> > >
> > > That should have been pci_save_state(). [Note to self: don't send email when
> > > you're too tired.]
> >
> > Yes, adding pci_save_state() fixed the problem. Here is the new patch;
> > it seems to work. Anything else need to be changed?
>
> I spoke a little too soon. It appears that there is a small bug in
> pci_enable_wake():
>
> int pci_enable_wake(struct pci_dev *dev, pci_power_t state, int enable)
> {
> int error = 0;
> bool pme_done = false;
>
> if (!device_may_wakeup(&dev->dev))
> return -EINVAL;
>
> This is wrong; the test should be:
>
> if (enable && !device_may_wakeup(&dev->dev))
Hm, that makes sense.
Will you prepare a patch?
Rafael
_______________________________________________
linux-pm mailing list
linux-pm@xxxxxxxxxxxxxxxxxxxxxxxxxx
https://lists.linux-foundation.org/mailman/listinfo/linux-pm
