On Saturday, June 16, 2012, huang ying wrote:
> On Sat, Jun 16, 2012 at 5:21 AM, Rafael J. Wysocki <rjw@xxxxxxx> wrote:
> [snip]
> >
> > > @@ -731,9 +791,13 @@ int pci_set_power_state(struct pci_dev *
> > > {
> > > int error;
> > >
> > > - /* bound the state we're entering */
> > > - if (state > PCI_D3hot)
> > > + /* D3cold during system suspend/hibernate is not supported */
> > > + if (!dev->runtime_d3cold && state > PCI_D3hot)
> >
> > I don't think this is the right place for the dev->runtime_d3cold check,
> > I've already told you that and haven't heard anything from you causing me
> > to change my mind.
> >
> > If you don't want D3cold to be used during system sleep transitions,
> > please make sure that D3cold is not passed to pci_set_power_state() in
> > those cases.
> >
>
> OK. Checked pci_set_power_state usage with following command line:
>
> $ grep 'pci_set_power_state' -r . | grep -v PCI_D3hot | grep -v PCI_D0 |
> grep -v pci_choose_state
> $ grep 'pci_set_power_state' -r . | grep -v PCI_D3hot | grep -v PCI_D0 |
> grep -v pci_choose_state | grep -i d3cold
>
> Found that, almost all pci_set_power_state usage is:
>
> 1) set to PCI_Dx (not PCI_D3cold)
> 2) set to the state that choosed by pci_choose_state()
> 3) driver use pci_prepare_to_sleep(), which will call pci_set_power_state()
>
> The only two exceptions are as follow:
>
> ./drivers/misc/cb710/core.c: pci_set_power_state(pdev,
> PCI_D3cold);
> ./arch/x86/pci/mrst.c: pci_set_power_state(dev, PCI_D3cold);
Well, no one is supposed to pass PCI_D3cold to pci_set_power_state() before
the $subject patch, so they should be fixed anyway.
> And pci_choose_state() will not return PCI_D3cold for system suspend.
>
> So what do you think about the following solution?
>
> a) Change pci_prepare_to_sleep(), so that it will not put device into
> PCI_D3cold.
> b) Change PCI_D3cold to PCI_D3hot in the following two places,
>
> ./drivers/misc/cb710/core.c: pci_set_power_state(pdev,
> PCI_D3cold);
> ./arch/x86/pci/mrst.c: pci_set_power_state(dev, PCI_D3cold);
Yes, that will do.
Thanks,
Rafael
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
