Merge branch 'perf/urgent' into perf/core

Merge reason: We are going to queue up a dependent patch.

Signed-off-by: Ingo Molnar <mingo@elte.hu>

3 files changed, 97 insertions, 19 deletions

diff --git a/net/ipv4/inetpeer.c b/net/ipv4/inetpeer.c
index bf4a9c4808e..d4d61b694fa 100644
--- a/net/ipv4/inetpeer.c
+++ b/net/ipv4/inetpeer.c
@@ -17,6 +17,7 @@
 #include <linux/kernel.h>
 #include <linux/mm.h>
 #include <linux/net.h>
+#include <linux/workqueue.h>
 #include <net/ip.h>
 #include <net/inetpeer.h>
 #include <net/secure_seq.h>
@@ -66,6 +67,11 @@
 
 static struct kmem_cache *peer_cachep __read_mostly;
 
+static LIST_HEAD(gc_list);
+static const int gc_delay = 60 * HZ;
+static struct delayed_work gc_work;
+static DEFINE_SPINLOCK(gc_lock);
+
 #define node_height(x) x->avl_height
 
 #define peer_avl_empty ((struct inet_peer *)&peer_fake_node)
@@ -102,6 +108,50 @@ int inet_peer_threshold __read_mostly = 65536 + 128;	/* start to throw entries m
 int inet_peer_minttl __read_mostly = 120 * HZ;	/* TTL under high load: 120 sec */
 int inet_peer_maxttl __read_mostly = 10 * 60 * HZ;	/* usual time to live: 10 min */
 
+static void inetpeer_gc_worker(struct work_struct *work)
+{
+	struct inet_peer *p, *n;
+	LIST_HEAD(list);
+
+	spin_lock_bh(&gc_lock);
+	list_replace_init(&gc_list, &list);
+	spin_unlock_bh(&gc_lock);
+
+	if (list_empty(&list))
+		return;
+
+	list_for_each_entry_safe(p, n, &list, gc_list) {
+
+		if(need_resched())
+			cond_resched();
+
+		if (p->avl_left != peer_avl_empty) {
+			list_add_tail(&p->avl_left->gc_list, &list);
+			p->avl_left = peer_avl_empty;
+		}
+
+		if (p->avl_right != peer_avl_empty) {
+			list_add_tail(&p->avl_right->gc_list, &list);
+			p->avl_right = peer_avl_empty;
+		}
+
+		n = list_entry(p->gc_list.next, struct inet_peer, gc_list);
+
+		if (!atomic_read(&p->refcnt)) {
+			list_del(&p->gc_list);
+			kmem_cache_free(peer_cachep, p);
+		}
+	}
+
+	if (list_empty(&list))
+		return;
+
+	spin_lock_bh(&gc_lock);
+	list_splice(&list, &gc_list);
+	spin_unlock_bh(&gc_lock);
+
+	schedule_delayed_work(&gc_work, gc_delay);
+}
 
 /* Called from ip_output.c:ip_init  */
 void __init inet_initpeers(void)
@@ -126,6 +176,7 @@ void __init inet_initpeers(void)
 			0, SLAB_HWCACHE_ALIGN | SLAB_PANIC,
 			NULL);
 
+	INIT_DELAYED_WORK_DEFERRABLE(&gc_work, inetpeer_gc_worker);
 }
 
 static int addr_compare(const struct inetpeer_addr *a,
@@ -447,9 +498,8 @@ relookup:
 		p->rate_last = 0;
 		p->pmtu_expires = 0;
 		p->pmtu_orig = 0;
-		p->redirect_genid = 0;
 		memset(&p->redirect_learned, 0, sizeof(p->redirect_learned));
-
+		INIT_LIST_HEAD(&p->gc_list);
 
 		/* Link the node. */
 		link_to_pool(p, base);
@@ -509,3 +559,30 @@ bool inet_peer_xrlim_allow(struct inet_peer *peer, int timeout)
 	return rc;
 }
 EXPORT_SYMBOL(inet_peer_xrlim_allow);
+
+void inetpeer_invalidate_tree(int family)
+{
+	struct inet_peer *old, *new, *prev;
+	struct inet_peer_base *base = family_to_base(family);
+
+	write_seqlock_bh(&base->lock);
+
+	old = base->root;
+	if (old == peer_avl_empty_rcu)
+		goto out;
+
+	new = peer_avl_empty_rcu;
+
+	prev = cmpxchg(&base->root, old, new);
+	if (prev == old) {
+		base->total = 0;
+		spin_lock(&gc_lock);
+		list_add_tail(&prev->gc_list, &gc_list);
+		spin_unlock(&gc_lock);
+		schedule_delayed_work(&gc_work, gc_delay);
+	}
+
+out:
+	write_sequnlock_bh(&base->lock);
+}
+EXPORT_SYMBOL(inetpeer_invalidate_tree);
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index bcacf54e541..01977479617 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -132,7 +132,6 @@ static int ip_rt_mtu_expires __read_mostly	= 10 * 60 * HZ;
 static int ip_rt_min_pmtu __read_mostly		= 512 + 20 + 20;
 static int ip_rt_min_advmss __read_mostly	= 256;
 static int rt_chain_length_max __read_mostly	= 20;
-static int redirect_genid;
 
 static struct delayed_work expires_work;
 static unsigned long expires_ljiffies;
@@ -937,7 +936,7 @@ static void rt_cache_invalidate(struct net *net)
 
 	get_random_bytes(&shuffle, sizeof(shuffle));
 	atomic_add(shuffle + 1U, &net->ipv4.rt_genid);
-	redirect_genid++;
+	inetpeer_invalidate_tree(AF_INET);
 }
 
 /*
@@ -1485,10 +1484,8 @@ void ip_rt_redirect(__be32 old_gw, __be32 daddr, __be32 new_gw,
 
 				peer = rt->peer;
 				if (peer) {
-					if (peer->redirect_learned.a4 != new_gw ||
-					    peer->redirect_genid != redirect_genid) {
+					if (peer->redirect_learned.a4 != new_gw) {
 						peer->redirect_learned.a4 = new_gw;
-						peer->redirect_genid = redirect_genid;
 						atomic_inc(&__rt_peer_genid);
 					}
 					check_peer_redir(&rt->dst, peer);
@@ -1793,8 +1790,6 @@ static void ipv4_validate_peer(struct rtable *rt)
 		if (peer) {
 			check_peer_pmtu(&rt->dst, peer);
 
-			if (peer->redirect_genid != redirect_genid)
-				peer->redirect_learned.a4 = 0;
 			if (peer->redirect_learned.a4 &&
 			    peer->redirect_learned.a4 != rt->rt_gateway)
 				check_peer_redir(&rt->dst, peer);
@@ -1958,8 +1953,7 @@ static void rt_init_metrics(struct rtable *rt, const struct flowi4 *fl4,
 		dst_init_metrics(&rt->dst, peer->metrics, false);
 
 		check_peer_pmtu(&rt->dst, peer);
-		if (peer->redirect_genid != redirect_genid)
-			peer->redirect_learned.a4 = 0;
+
 		if (peer->redirect_learned.a4 &&
 		    peer->redirect_learned.a4 != rt->rt_gateway) {
 			rt->rt_gateway = peer->redirect_learned.a4;
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 53c8ce4046b..b5e315f1364 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -1403,8 +1403,16 @@ static int tcp_shifted_skb(struct sock *sk, struct sk_buff *skb,
 
 	BUG_ON(!pcount);
 
-	/* Adjust hint for FACK. Non-FACK is handled in tcp_sacktag_one(). */
-	if (tcp_is_fack(tp) && (skb == tp->lost_skb_hint))
+	/* Adjust counters and hints for the newly sacked sequence
+	 * range but discard the return value since prev is already
+	 * marked. We must tag the range first because the seq
+	 * advancement below implicitly advances
+	 * tcp_highest_sack_seq() when skb is highest_sack.
+	 */
+	tcp_sacktag_one(sk, state, TCP_SKB_CB(skb)->sacked,
+			start_seq, end_seq, dup_sack, pcount);
+
+	if (skb == tp->lost_skb_hint)
 		tp->lost_cnt_hint += pcount;
 
 	TCP_SKB_CB(prev)->end_seq += shifted;
@@ -1430,12 +1438,6 @@ static int tcp_shifted_skb(struct sock *sk, struct sk_buff *skb,
 		skb_shinfo(skb)->gso_type = 0;
 	}
 
-	/* Adjust counters and hints for the newly sacked sequence range but
-	 * discard the return value since prev is already marked.
-	 */
-	tcp_sacktag_one(sk, state, TCP_SKB_CB(skb)->sacked,
-			start_seq, end_seq, dup_sack, pcount);
-
 	/* Difference in this won't matter, both ACKed by the same cumul. ACK */
 	TCP_SKB_CB(prev)->sacked |= (TCP_SKB_CB(skb)->sacked & TCPCB_EVER_RETRANS);
 
@@ -1583,6 +1585,10 @@ static struct sk_buff *tcp_shift_skb_data(struct sock *sk, struct sk_buff *skb,
 		}
 	}
 
+	/* tcp_sacktag_one() won't SACK-tag ranges below snd_una */
+	if (!after(TCP_SKB_CB(skb)->seq + len, tp->snd_una))
+		goto fallback;
+
 	if (!skb_shift(prev, skb, len))
 		goto fallback;
 	if (!tcp_shifted_skb(sk, skb, state, pcount, len, mss, dup_sack))
@@ -2567,6 +2573,7 @@ static void tcp_mark_head_lost(struct sock *sk, int packets, int mark_head)
 
 		if (cnt > packets) {
 			if ((tcp_is_sack(tp) && !tcp_is_fack(tp)) ||
+			    (TCP_SKB_CB(skb)->sacked & TCPCB_SACKED_ACKED) ||
 			    (oldcnt >= packets))
 				break;