Am Monday 12 January 2009 09:57:41 schrieb Denis Joseph Barrow: > Oliver if you haven't given a detailed description of the patch you need to do so before the patch can be accepted > especially a patch of this size. The original driver has a race that left a window between submitting a network packet and incrementing the usage counters preventing autosuspension. In that window hard_start_xmit() races with suspend() called for autosuspension. The problem arises from having to increase the usage counters from a context that cannot be slept in. A driver has to be able to deal with suspend being called in that case and needing to return -EBUSY or to delay transmission of the packet until resume() is called. This needs to called with approriate (spin)locking. This patch implements such handling in suspend() and unifies the spinlocks from the network and the serial parts of the driver. Thus all code taking a spinlock had to be touched. Regards Oliver -- 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
