[PATCH 6/6] USB: gadget: mv_udc: fix corner case for missiong dTD

[Neil Zhang]

When the missing dTD issue is triggerred, queue_dtd may prime the new
request instead of the missing dTD.
We can just add the request to the queue end and jump out if there are
more than one request in the queue already.

Signed-off-by: Neil Zhang <zhangwm@xxxxxxxxxxx>
---
 drivers/usb/gadget/mv_udc_core.c |   12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)

diff --git a/drivers/usb/gadget/mv_udc_core.c b/drivers/usb/gadget/mv_udc_core.c
index 8df8606..918b603 100644
--- a/drivers/usb/gadget/mv_udc_core.c
+++ b/drivers/usb/gadget/mv_udc_core.c
@@ -295,13 +295,23 @@ static int queue_dtd(struct mv_ep *ep, struct mv_req *req)
 
 	/* check if the pipe is empty */
 	if (!(list_empty(&ep->queue))) {
-		struct mv_req *lastreq;
+		struct mv_req *firstreq, *lastreq;
+		firstreq = list_entry(ep->queue.next, struct mv_req, queue);
 		lastreq = list_entry(ep->queue.prev, struct mv_req, queue);
 		lastreq->tail->dtd_next =
 			req->head->td_dma & EP_QUEUE_HEAD_NEXT_POINTER_MASK;
 
 		wmb();
 
+		/*
+		 * If there are more than one reqs in the queue,
+		 * then it's safe to just add it to the list end.
+		 * It should be able to handle the missing DTD issue.
+		 * And suppose it can improve the performance too.
+		 */
+		if (firstreq != lastreq)
+			goto done;
+
 		if (readl(&udc->op_regs->epprime) & bit_pos)
 			goto done;
 
-- 
1.7.9.5

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/