> PCI has an interface to register driver. It works well for this case
> where the driver is already loaded and a device is inserted.

I glanced at <linux/pci.h> and found a "struct pci_driver", and
it wasn't something I was familiar with from the USB subsystem.

Seems likely that we should be using that framework for all
HC drivers that are resident ... perhaps just by providing a
pci_driver struct and calling pci_module_init().

Was there a reason we're not using that framework?  Like, maybe
it was unstable way when the USB HCDs first got going?  Seems
like USB could be using those existing hotplug facilities.

- Dave




---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to