On Tue, 28 Sep 2010, Ming Lei wrote:
> The following patch can fix the oops, if no objections from other
> guys, I will submit a formal version.
>
> The oops is caused by usb_enable_interface called from
> usb_unbind_interface, usb_enable_interface can make
> udev->ep_out[epnum] point to usb_host_endpoint instance
> which will be freed from usb_destroy_configuration later, so when
> the device is removed from usb bus, oops will happen since
> usb_disable_device is called again from usb_disconnect to touch
> this freed usb_host_endpoint instance.
Your diagnosis looks right, but your solution is wrong. The problem is
that the endpoints need to be disabled reliably when a configuration is
changed, even though unbinding an interface can cause its endpoints to
be re-enabled. This really has nothing to do with authorization.
The patch below should work.
Alan Stern
Index: usb-2.6/drivers/usb/core/message.c
===================================================================
--- usb-2.6.orig/drivers/usb/core/message.c
+++ usb-2.6/drivers/usb/core/message.c
@@ -1218,6 +1218,8 @@ void usb_enable_interface(struct usb_dev
struct usb_host_interface *alt = intf->cur_altsetting;
int i;
+ if (intf->unregistering)
+ return;
for (i = 0; i < alt->desc.bNumEndpoints; ++i)
usb_enable_endpoint(dev, &alt->endpoint[i], reset_eps);
}
@@ -1274,6 +1276,8 @@ int usb_set_interface(struct usb_device
interface);
return -EINVAL;
}
+ if (iface->unregistering)
+ return -ESHUTDOWN;
alt = usb_altnum_to_altsetting(iface, alternate);
if (!alt) {
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
