tcp: tsq: remove one locked operation in tcp_wfree()

Instead of atomically clear TSQ_THROTTLED and atomically set TSQ_QUEUED
bits, use one cmpxchg() to perform a single locked operation.

Since the following patch will also set TCP_TSQ_DEFERRED here,
this cmpxchg() will make this addition free.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/ipv4/tcp_output.c

@@ -860,6 +860,7 @@ void tcp_wfree(struct sk_buff *skb)
 {
 	struct sock *sk = skb->sk;
 	struct tcp_sock *tp = tcp_sk(sk);
+	unsigned long flags, nval, oval;
 	int wmem;
 
 	/* Keep one reference on sk_wmem_alloc.
@@ -877,11 +878,17 @@ void tcp_wfree(struct sk_buff *skb)
 	if (wmem >= SKB_TRUESIZE(1) && this_cpu_ksoftirqd() == current)
 		goto out;
 
-	if (test_and_clear_bit(TSQ_THROTTLED, &tp->tsq_flags) &&
-	    !test_and_set_bit(TSQ_QUEUED, &tp->tsq_flags)) {
-		unsigned long flags;
+	for (oval = READ_ONCE(tp->tsq_flags);; oval = nval) {
 		struct tsq_tasklet *tsq;
 
+		if (!(oval & TSQF_THROTTLED) || (oval & TSQF_QUEUED))
+			goto out;
+
+		nval = (oval & ~TSQF_THROTTLED) | TSQF_QUEUED;
+		nval = cmpxchg(&tp->tsq_flags, oval, nval);
+		if (nval != oval)
+			continue;
+
 		/* queue this socket to tasklet queue */
 		local_irq_save(flags);
 		tsq = this_cpu_ptr(&tsq_tasklet);