xsk: Rollback reservation at NETDEV_TX_BUSY

Rollback the reservation in the completion ring when we get a
NETDEV_TX_BUSY. When this error is received from the driver, we are
supposed to let the user application retry the transmit again. And in
order to do this, we need to roll back the failed send so it can be
retried. Unfortunately, we did not cancel the reservation we had made
in the completion ring. By not doing this, we actually make the
completion ring one entry smaller per NETDEV_TX_BUSY error we get, and
after enough of these errors the completion ring will be of size zero
and transmit will stop working.

Fix this by cancelling the reservation when we get a NETDEV_TX_BUSY
error.

Fixes: 642e450b ("xsk: Do not discard packet when NETDEV_TX_BUSY")
Reported-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
Signed-off-by: Magnus Karlsson <magnus.karlsson@intel.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Björn Töpel <bjorn.topel@intel.com>
Link: https://lore.kernel.org/bpf/20201218134525.13119-3-magnus.karlsson@gmail.com

net/xdp/xsk.c

@@ -487,6 +487,9 @@ static int xsk_generic_xmit(struct sock *sk)
 		if  (err == NETDEV_TX_BUSY) {
 			/* Tell user-space to retry the send */
 			skb->destructor = sock_wfree;
+			spin_lock_irqsave(&xs->pool->cq_lock, flags);
+			xskq_prod_cancel(xs->pool->cq);
+			spin_unlock_irqrestore(&xs->pool->cq_lock, flags);
 			/* Free skb without triggering the perf drop trace */
 			consume_skb(skb);
 			err = -EAGAIN;

net/xdp/xsk_queue.h

@@ -334,6 +334,11 @@ static inline bool xskq_prod_is_full(struct xsk_queue *q)
 	return xskq_prod_nb_free(q, 1) ? false : true;
 }
 
+static inline void xskq_prod_cancel(struct xsk_queue *q)
+{
+	q->cached_prod--;
+}
+
 static inline int xskq_prod_reserve(struct xsk_queue *q)
 {
 	if (xskq_prod_is_full(q))