On 12/13/2010 12:08 AM, Pavan Kondeti wrote:
> On 12/13/2010 1:32 PM, Artem Leonenko wrote:
>>>> @@ -2170,7 +2169,9 @@ static int ep_queue(struct usb_ep *ep, struct
>>>>>> usb_request *req,
>>>>>> mReq->req.actual = 0;
>>>>>> list_add_tail(&mReq->queue, &mEp->qh[mEp->dir].queue);
>>>>>>
>>>>>> - retval = _hardware_enqueue(mEp, mReq);
>>>>>> + if (list_is_singular(&mEp->qh[mEp->dir].queue))
>>>>>> + retval = _hardware_enqueue(mEp, mReq);
>>>>>> +
>>>>>> if (retval == -EALREADY || retval == -EBUSY) {
>>>>>> dbg_event(_usb_addr(mEp), "QUEUE", retval);
>>>>>> retval = 0;
>>>>
>>>> Given that _hardware_enqueue is called only when endpoint is in
>>>> un-primed state, may be we should pass the value returned by
>>>> _hardware_enqueue back to usb_ep_queue() caller.
>>>>
>> How can we do that if caller puts more than one request to the queue? If
>> there are already requests in the queue then ep_queue returns and
>> _hardware_enqueue will be called later in isr_tr_complete_low().
>>
>>
> Earlier we are calling _hardware_enqueue every time ep_queue is called.
> So we were converting -EBUSY to success. But now we will call
> _hardware_enqueue only when we know that ep is unprimed and expect
> _hardware_enqueue return success. Then why converting -EBUSY to success?
Good point. I can't think of any scenario with primed endpoint and empty
queue. That's definitely an error.
--
Cheers,
Artem
--
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
