On Mon, 26 Mar 2018, Yavuz, Tuba wrote: > I agree with Alan that the spinlock must be dropped before calling > usb_ep_queue. An example can be found in the ep0_queue function of > the f_mass_storage driver.
Thanks for the note of support, but you must be looking at a different version of the f_mass_storage code than I am -- in my copy, ep0_queue does not acquire or release any locks. I forgot to mention in my earlier email... Felipe said that the problem case involved calling the completion handler for a failed submission. That definitely is not the right approach; if a submission fails then usb_ep_queue() should return an error code and the completion routine should not be called. In fact, there is a general principal here which should be documented but doesn't seem to be (as far as I can tell). Namely, a request's completion routine will be called if and only if usb_ep_queue() returns 0. That would make a good addition to the kerneldoc for usb_ep_queue(). Alan Stern -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
