On Sunday, July 18, 2010 07:52:28 pm ykzhao wrote: > On Sat, 2010-07-17 at 00:01 +0800, Bjorn Helgaas wrote: > > On Thursday, July 15, 2010 08:34:35 pm ykzhao wrote: > > > On Fri, 2010-07-16 at 00:27 +0800, Bjorn Helgaas wrote: > > > > The ACPI IPMI opregion code deals with an ACPI device, and the > > > > obvious place where you have that device is the driver "add" > > > > function, i.e., ipmi_pnp_probe(). That's the point where the > > > > PNP core will help you out by matching device IDs and handling > > > > hotplug for you. ... > > > Yes. The hotplug scenario should be considered. This will be handled > > > under two places: > > > a. the IPMI pnp detection: > > > b. install opregion handle for the hotplug IPI0001 device > > > > > > But in fact the acpi_pnp device is enumerated only once at the boot > > > time. > > > > Drivers cannot rely on anything like "the acpi_pnp device is > > enumerated only once at boot-time." Device enumeration happens > > in the ACPI core and PNP core, not in drivers, and drivers can't > > assume anything about when it happens. > > Yes. But in fact the ACPI pnp device is enumerated only once at the > boot time. NO! You can NOT assume the ACPI PNP device is enumerated once at boot time. That may be the way the current implementation behaves, but the driver interface does not specify that, and it is likely to change. Bjorn -- 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