On Friday, 20 of June 2008, Len Brown wrote: > > > I've recently discovered that pci_set_power_state() calls the platform > > callback intended for changing the state of the device after it has done > > that using the native PCI mechanism. In my opinion this is not correct, > > because, for example, if the device is being put into D0 from a low power > > state, it may require some platform-controlled power resources to be turned > > on before that (in theory). > > > > The following series of patches changes this function so that the platform > > callback is performed first and then the native mechanism is used. > > Additionally, however, it checks if the devices is power manageable by the > > platform and only in that case the platform callback is invoked. For this > > purpose, I needed to add an ACPI function allowing PCI to check the power > > manageability of given device. > > We have 3 cases > 1. device has PCI PM only > 2. device has ACPI PM only > 3. device has both PCI and ACPI PM > > For #1 and #2, order is moot. > For #3, how can we be sure that the new order is better than the old > order? Is this a theoretical patch, or is there a failure case? This was a theoretical patch, but please see the new series posted. The difference is that in the new series the order is only changed for transitions to D0. Thanks, Rafael -- To unsubscribe from this list: send the line "unsubscribe linux-acpi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
