On Wed, 23 Nov 2011, Sebastian Andrzej Siewior wrote: > >that's the same thing as setting no_interrupt on all first 255 requests > >and only calling completion on the last transfer. The difference between > >this and the host API is that host SG support is always blocking, it > >will sleep on a wait_for_completion(). > > This is true and I need to change this detail in order to fix UAS > support for hcd which don't support sgs as suggested by Alan. The hcd sg > interface is blocking but it is no neccesarry a must argument. There is > no reason why one could not perform it async. That's true. But remember there's a difference between "async" and "in interrupt context". Given a long SG list, usb_sg_init() and usb_sg_wait() end up spending a large amount of time creating and submitting URBs, which is not a good thing to do while in interrupt context. Alan Stern -- 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
