Re: [ACPI] Re: [PATCH 4/4]An ACPI callback for pci_set_power_state

[Li Shaohua <shaohua.li@intel.com>]

On Wed, 2005-01-05 at 16:49, Pavel Machek wrote:
> Hi!
> 
> > This is an ACPI callback for pci_set_power_state. Besides setting PCI
> > config space, changing device's power state sometimes requires to power
> > on device's power source and to invoke other firmware methods.
> 
> 
> > diff -puN drivers/pci/pci.h~acpi-pci-set-power-state-callback drivers/pci/pci.h
> > --- 2.5/drivers/pci/pci.h~acpi-pci-set-power-state-callback	2005-01-05 09:58:06.469923128 +0800
> > +++ 2.5-root/drivers/pci/pci.h	2005-01-05 09:58:06.473922520 +0800
> > @@ -13,6 +13,7 @@ extern int pci_bus_alloc_resource(struct
> >  				  void *alignf_data);
> >  /* Firmware callbacks */
> >  extern int (*platform_pci_get_suspend_state)(struct device *dev, u32 state);
> > +extern int (*platform_pci_set_power_state)(struct pci_dev *dev, int state);
> 
> What kind of state is passed here? Why is it u32 in one case and int
> in the second one?
This is the types of current pci routines (pci_set_power_state and pci_device_suspend). 
Both are PCI device states. I agree it's confused.

> I'm about to introduce separate types for pci power states and system
> power states; could you at least add comments which states are which?
> Also few lines of documentation would be very usefull...
I planed post the patches after your patches enter before, but it seems
have a long delay. What's your plan? I'm ok to merge the patches after
yours.

Thanks,
Shaohua