On Mon, 25 Apr 2005, Roman Kagan wrote: > > But it is needed for the klist version. I didn't send a patch for that > > version because I don't have the -mm kernel source handy. The analogous > > change to the code should work, however. > > OK then a matching change to usb_driver_claim_interface would be needed > too.
I think usb_driver_claim_interface is correct as it stands. It was a mistake to leave out from usb_driver_release_interface originally the line setting iface->condition to USB_INTERFACE_UNBINDING. > > Probably this is something you've missed. Let's say a driver X bound to > > device D is rmmod'ed at the same time that driver Y is insmod'ed. > > [...] > > device_bind_driver calls klist_add_tail for D. But D's > > knode_driver is still on X's klist_devices! > > Yes the scenario was on the surface, I just thought that this last step, > where knode is added to a list before being released from another, is > excluded by the klist core. No, it isn't excluded. The klist core just overwrites the pointers. > At any rate the driver can check for > klist_node_attached(&dev->knode_bus) before claiming the device, and > leave it alone if it's already attached to another driver. You mean ->knode_driver, not ->knode_bus. The _driver_ shouldn't check. The check should be part of the driver-model core in __driver_attach, at the place where it checks whether the device is already bound to a driver. In fact, this check can subsume that one, since the klist_node will always be attached whenever the device is bound. Alan Stern ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ [email protected] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
