[PATCH] USB: Usbcore: Don't try to delete unregistered interfaces

This patch handles a rarely-encountered failure mode in usbcore.  It's
legal for device_add to fail (although now it happens even more rarely
than before since failure to bind a driver is no longer fatal).  So when
we destroy the interfaces in a configuration, we shouldn't try to delete
ones which weren't successfully registered.  Also, failure to register an
interface shouldn't be fatal either -- I think; you may disagree about
this part of the patch.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

diff --git a/drivers/usb/core/message.c b/drivers/usb/core/message.c
index a428ef479bd70d27b1cf435e5247927357bdd545..88d1b376f67cb2ee03d15ebfce7654de8a2d0b32 100644 (file)
--- a/drivers/usb/core/message.c
+++ b/drivers/usb/core/message.c
@@ -985,8 +985,10 @@ void usb_disable_device(struct usb_device *dev, int skip_ep0)
                for (i = 0; i < dev->actconfig->desc.bNumInterfaces; i++) {
                        struct usb_interface    *interface;
 
-                       /* remove this interface */
+                       /* remove this interface if it has been registered */
                        interface = dev->actconfig->interface[i];
+                       if (!klist_node_attached(&interface->dev.knode_bus))
+                               continue;
                        dev_dbg (&dev->dev, "unregistering interface %s\n",
                                interface->dev.bus_id);
                        usb_remove_sysfs_intf_files(interface);
@@ -1439,7 +1441,7 @@ free_interfaces:
                }
        }
 
-       return ret;
+       return 0;
 }
 
 // synchronous request completion model