1 files changed, 58 insertions, 74 deletions

diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index f5c336f8b0c8..bc7f419184aa 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -287,8 +287,8 @@ enum {
 	TCP_CMSG_TS = 2
 };
 
-struct percpu_counter tcp_orphan_count;
-EXPORT_SYMBOL_GPL(tcp_orphan_count);
+DEFINE_PER_CPU(unsigned int, tcp_orphan_count);
+EXPORT_PER_CPU_SYMBOL_GPL(tcp_orphan_count);
 
 long sysctl_tcp_mem[3] __read_mostly;
 EXPORT_SYMBOL(sysctl_tcp_mem);
@@ -325,11 +325,6 @@ struct tcp_splice_state {
 unsigned long tcp_memory_pressure __read_mostly;
 EXPORT_SYMBOL_GPL(tcp_memory_pressure);
 
-DEFINE_STATIC_KEY_FALSE(tcp_rx_skb_cache_key);
-EXPORT_SYMBOL(tcp_rx_skb_cache_key);
-
-DEFINE_STATIC_KEY_FALSE(tcp_tx_skb_cache_key);
-
 void tcp_enter_memory_pressure(struct sock *sk)
 {
 	unsigned long val;
@@ -644,7 +639,7 @@ int tcp_ioctl(struct sock *sk, int cmd, unsigned long arg)
 }
 EXPORT_SYMBOL(tcp_ioctl);
 
-static inline void tcp_mark_push(struct tcp_sock *tp, struct sk_buff *skb)
+void tcp_mark_push(struct tcp_sock *tp, struct sk_buff *skb)
 {
 	TCP_SKB_CB(skb)->tcp_flags |= TCPHDR_PSH;
 	tp->pushed_seq = tp->write_seq;
@@ -655,15 +650,13 @@ static inline bool forced_push(const struct tcp_sock *tp)
 	return after(tp->write_seq, tp->pushed_seq + (tp->max_window >> 1));
 }
 
-static void skb_entail(struct sock *sk, struct sk_buff *skb)
+void tcp_skb_entail(struct sock *sk, struct sk_buff *skb)
 {
 	struct tcp_sock *tp = tcp_sk(sk);
 	struct tcp_skb_cb *tcb = TCP_SKB_CB(skb);
 
-	skb->csum    = 0;
 	tcb->seq     = tcb->end_seq = tp->write_seq;
 	tcb->tcp_flags = TCPHDR_ACK;
-	tcb->sacked  = 0;
 	__skb_header_release(skb);
 	tcp_add_write_queue_tail(sk, skb);
 	sk_wmem_queued_add(sk, skb->truesize);
@@ -858,30 +851,15 @@ ssize_t tcp_splice_read(struct socket *sock, loff_t *ppos,
 }
 EXPORT_SYMBOL(tcp_splice_read);
 
-struct sk_buff *sk_stream_alloc_skb(struct sock *sk, int size, gfp_t gfp,
-				    bool force_schedule)
+struct sk_buff *tcp_stream_alloc_skb(struct sock *sk, int size, gfp_t gfp,
+				     bool force_schedule)
 {
 	struct sk_buff *skb;
 
-	if (likely(!size)) {
-		skb = sk->sk_tx_skb_cache;
-		if (skb) {
-			skb->truesize = SKB_TRUESIZE(skb_end_offset(skb));
-			sk->sk_tx_skb_cache = NULL;
-			pskb_trim(skb, 0);
-			INIT_LIST_HEAD(&skb->tcp_tsorted_anchor);
-			skb_shinfo(skb)->tx_flags = 0;
-			memset(TCP_SKB_CB(skb), 0, sizeof(struct tcp_skb_cb));
-			return skb;
-		}
-	}
-	/* The TCP header must be at least 32-bit aligned.  */
-	size = ALIGN(size, 4);
-
 	if (unlikely(tcp_under_memory_pressure(sk)))
 		sk_mem_reclaim_partial(sk);
 
-	skb = alloc_skb_fclone(size + sk->sk_prot->max_header, gfp);
+	skb = alloc_skb_fclone(size + MAX_TCP_HEADER, gfp);
 	if (likely(skb)) {
 		bool mem_scheduled;
 
@@ -892,12 +870,8 @@ struct sk_buff *sk_stream_alloc_skb(struct sock *sk, int size, gfp_t gfp,
 			mem_scheduled = sk_wmem_schedule(sk, skb->truesize);
 		}
 		if (likely(mem_scheduled)) {
-			skb_reserve(skb, sk->sk_prot->max_header);
-			/*
-			 * Make sure that we have exactly size bytes
-			 * available to the caller, no more, no less.
-			 */
-			skb->reserved_tailroom = skb->end - skb->tail - size;
+			skb_reserve(skb, MAX_TCP_HEADER);
+			skb->ip_summed = CHECKSUM_PARTIAL;
 			INIT_LIST_HEAD(&skb->tcp_tsorted_anchor);
 			return skb;
 		}
@@ -950,18 +924,20 @@ int tcp_send_mss(struct sock *sk, int *size_goal, int flags)
  * importantly be able to generate EPOLLOUT for Edge Trigger epoll()
  * users.
  */
-void tcp_remove_empty_skb(struct sock *sk, struct sk_buff *skb)
+void tcp_remove_empty_skb(struct sock *sk)
 {
-	if (skb && !skb->len) {
+	struct sk_buff *skb = tcp_write_queue_tail(sk);
+
+	if (skb && TCP_SKB_CB(skb)->seq == TCP_SKB_CB(skb)->end_seq) {
 		tcp_unlink_write_queue(skb, sk);
 		if (tcp_write_queue_empty(sk))
 			tcp_chrono_stop(sk, TCP_CHRONO_BUSY);
-		sk_wmem_free_skb(sk, skb);
+		tcp_wmem_free_skb(sk, skb);
 	}
 }
 
-struct sk_buff *tcp_build_frag(struct sock *sk, int size_goal, int flags,
-			       struct page *page, int offset, size_t *size)
+static struct sk_buff *tcp_build_frag(struct sock *sk, int size_goal, int flags,
+				      struct page *page, int offset, size_t *size)
 {
 	struct sk_buff *skb = tcp_write_queue_tail(sk);
 	struct tcp_sock *tp = tcp_sk(sk);
@@ -974,15 +950,15 @@ new_segment:
 		if (!sk_stream_memory_free(sk))
 			return NULL;
 
-		skb = sk_stream_alloc_skb(sk, 0, sk->sk_allocation,
-					  tcp_rtx_and_write_queues_empty(sk));
+		skb = tcp_stream_alloc_skb(sk, 0, sk->sk_allocation,
+					   tcp_rtx_and_write_queues_empty(sk));
 		if (!skb)
 			return NULL;
 
 #ifdef CONFIG_TLS_DEVICE
 		skb->decrypted = !!(flags & MSG_SENDPAGE_DECRYPTED);
 #endif
-		skb_entail(sk, skb);
+		tcp_skb_entail(sk, skb);
 		copy = size_goal;
 	}
 
@@ -1013,7 +989,6 @@ new_segment:
 	skb->truesize += copy;
 	sk_wmem_queued_add(sk, copy);
 	sk_mem_charge(sk, copy);
-	skb->ip_summed = CHECKSUM_PARTIAL;
 	WRITE_ONCE(tp->write_seq, tp->write_seq + copy);
 	TCP_SKB_CB(skb)->end_seq += copy;
 	tcp_skb_pcount_set(skb, 0);
@@ -1104,7 +1079,7 @@ out:
 	return copied;
 
 do_error:
-	tcp_remove_empty_skb(sk, tcp_write_queue_tail(sk));
+	tcp_remove_empty_skb(sk);
 	if (copied)
 		goto out;
 out_err:
@@ -1303,15 +1278,14 @@ new_segment:
 					goto restart;
 			}
 			first_skb = tcp_rtx_and_write_queues_empty(sk);
-			skb = sk_stream_alloc_skb(sk, 0, sk->sk_allocation,
-						  first_skb);
+			skb = tcp_stream_alloc_skb(sk, 0, sk->sk_allocation,
+						   first_skb);
 			if (!skb)
 				goto wait_for_space;
 
 			process_backlog++;
-			skb->ip_summed = CHECKSUM_PARTIAL;
 
-			skb_entail(sk, skb);
+			tcp_skb_entail(sk, skb);
 			copy = size_goal;
 
 			/* All packets are restored as if they have
@@ -1326,14 +1300,7 @@ new_segment:
 		if (copy > msg_data_left(msg))
 			copy = msg_data_left(msg);
 
-		/* Where to copy to? */
-		if (skb_availroom(skb) > 0 && !zc) {
-			/* We have some space in skb head. Superb! */
-			copy = min_t(int, copy, skb_availroom(skb));
-			err = skb_add_data_nocache(sk, skb, &msg->msg_iter, copy);
-			if (err)
-				goto do_fault;
-		} else if (!zc) {
+		if (!zc) {
 			bool merge = true;
 			int i = skb_shinfo(skb)->nr_frags;
 			struct page_frag *pfrag = sk_page_frag(sk);
@@ -1432,9 +1399,7 @@ out_nopush:
 	return copied + copied_syn;
 
 do_error:
-	skb = tcp_write_queue_tail(sk);
-do_fault:
-	tcp_remove_empty_skb(sk, skb);
+	tcp_remove_empty_skb(sk);
 
 	if (copied + copied_syn)
 		goto out;
@@ -2687,11 +2652,36 @@ void tcp_shutdown(struct sock *sk, int how)
 }
 EXPORT_SYMBOL(tcp_shutdown);
 
+int tcp_orphan_count_sum(void)
+{
+	int i, total = 0;
+
+	for_each_possible_cpu(i)
+		total += per_cpu(tcp_orphan_count, i);
+
+	return max(total, 0);
+}
+
+static int tcp_orphan_cache;
+static struct timer_list tcp_orphan_timer;
+#define TCP_ORPHAN_TIMER_PERIOD msecs_to_jiffies(100)
+
+static void tcp_orphan_update(struct timer_list *unused)
+{
+	WRITE_ONCE(tcp_orphan_cache, tcp_orphan_count_sum());
+	mod_timer(&tcp_orphan_timer, jiffies + TCP_ORPHAN_TIMER_PERIOD);
+}
+
+static bool tcp_too_many_orphans(int shift)
+{
+	return READ_ONCE(tcp_orphan_cache) << shift > sysctl_tcp_max_orphans;
+}
+
 bool tcp_check_oom(struct sock *sk, int shift)
 {
 	bool too_many_orphans, out_of_socket_memory;
 
-	too_many_orphans = tcp_too_many_orphans(sk, shift);
+	too_many_orphans = tcp_too_many_orphans(shift);
 	out_of_socket_memory = tcp_out_of_memory(sk);
 
 	if (too_many_orphans)
@@ -2800,7 +2790,7 @@ adjudge_to_death:
 	/* remove backlog if any, without releasing ownership. */
 	__release_sock(sk);
 
-	percpu_counter_inc(sk->sk_prot->orphan_count);
+	this_cpu_inc(tcp_orphan_count);
 
 	/* Have we already been destroyed by a softirq or backlog? */
 	if (state != TCP_CLOSE && sk->sk_state == TCP_CLOSE)
@@ -2903,7 +2893,7 @@ static void tcp_rtx_queue_purge(struct sock *sk)
 		 * list_del(&skb->tcp_tsorted_anchor)
 		 */
 		tcp_rtx_queue_unlink(skb, sk);
-		sk_wmem_free_skb(sk, skb);
+		tcp_wmem_free_skb(sk, skb);
 	}
 }
 
@@ -2914,14 +2904,9 @@ void tcp_write_queue_purge(struct sock *sk)
 	tcp_chrono_stop(sk, TCP_CHRONO_BUSY);
 	while ((skb = __skb_dequeue(&sk->sk_write_queue)) != NULL) {
 		tcp_skb_tsorted_anchor_cleanup(skb);
-		sk_wmem_free_skb(sk, skb);
+		tcp_wmem_free_skb(sk, skb);
 	}
 	tcp_rtx_queue_purge(sk);
-	skb = sk->sk_tx_skb_cache;
-	if (skb) {
-		__kfree_skb(skb);
-		sk->sk_tx_skb_cache = NULL;
-	}
 	INIT_LIST_HEAD(&tcp_sk(sk)->tsorted_sent_queue);
 	sk_mem_reclaim(sk);
 	tcp_clear_all_retrans_hints(tcp_sk(sk));
@@ -2958,10 +2943,6 @@ int tcp_disconnect(struct sock *sk, int flags)
 
 	tcp_clear_xmit_timers(sk);
 	__skb_queue_purge(&sk->sk_receive_queue);
-	if (sk->sk_rx_skb_cache) {
-		__kfree_skb(sk->sk_rx_skb_cache);
-		sk->sk_rx_skb_cache = NULL;
-	}
 	WRITE_ONCE(tp->copied_seq, tp->rcv_nxt);
 	tp->urg_data = 0;
 	tcp_write_queue_purge(sk);
@@ -4502,7 +4483,10 @@ void __init tcp_init(void)
 		     sizeof_field(struct sk_buff, cb));
 
 	percpu_counter_init(&tcp_sockets_allocated, 0, GFP_KERNEL);
-	percpu_counter_init(&tcp_orphan_count, 0, GFP_KERNEL);
+
+	timer_setup(&tcp_orphan_timer, tcp_orphan_update, TIMER_DEFERRABLE);
+	mod_timer(&tcp_orphan_timer, jiffies + TCP_ORPHAN_TIMER_PERIOD);
+
 	inet_hashinfo_init(&tcp_hashinfo);
 	inet_hashinfo2_init(&tcp_hashinfo, "tcp_listen_portaddr_hash",
 			    thash_entries, 21,  /* one slot per 2 MB*/