Hi Jerome, On Thu, Aug 23, 2012 at 9:39 PM, Jérôme Duval <korli@xxxxxxxxxxxxxxxx>wrote: > 2012/8/23 yongcong-github.master <community@xxxxxxxxxxxx>: > > @@ -563,6 +554,8 @@ acpi_cpuidle_uninit_driver(void *driverCookie) > > > > module_dependency module_dependencies[] = { > > { B_DEVICE_MANAGER_MODULE_NAME, (module_info **)&sDeviceManager > }, > > + { B_ACPI_MODULE_NAME, (module_info **)&sAcpi}, > > + { B_CPUIDLE_MODULE_NAME, (module_info **)&gIdle }, > > {} > > }; > > > > This change probably implies that x86_cpuidle is only loaded when the > ACPI module is loaded and enabled. > Yes. > I think that's acceptable even if x86_cpuidle is supposed to work with > ACPI disabled for newer Intel processors, right? > In theory, the intel drivers can work even if ACPI is disabled, but then the x86_cpuidle module loading can't be triggered during boot because currently it's loaded by acpi bus manager. If haiku can support trigger the module loading in userspace (insmod under Linux, for example) or we find other elegant way to automatically load x86_cpuidle, then we do support newer intel platforms even if acpi is disabled. Thanks very much, Yongcong