[net-next PATCH 3/3] net: frag queue per hash bucket locking

[Jesper Dangaard Brouer <brouer@redhat.com>]

This patch implements per hash bucket locking for the frag queue
hash.  This removes two write locks, and the only remaining write
lock is for protecting hash rebuild.  This essentially reduce the
readers-writer lock to a rebuild lock.

Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
---

 include/net/inet_frag.h  |    9 ++++++-
 net/ipv4/inet_fragment.c |   60 ++++++++++++++++++++++++++++++++++++----------
 2 files changed, 55 insertions(+), 14 deletions(-)

diff --git a/include/net/inet_frag.h b/include/net/inet_frag.h
index 7cac9c5..c4f5183 100644
--- a/include/net/inet_frag.h
+++ b/include/net/inet_frag.h
@@ -50,10 +50,17 @@ struct inet_frag_queue {
  */
 #define INETFRAGS_MAXDEPTH		128
 
+struct inet_frag_bucket {
+	struct hlist_head	chain;
+	spinlock_t		chain_lock;
+	u16			chain_len;
+};
+
 struct inet_frags {
-	struct hlist_head	hash[INETFRAGS_HASHSZ];
+	struct inet_frag_bucket	hash[INETFRAGS_HASHSZ];
 	/* This rwlock is a global lock (seperate per IPv4, IPv6 and
 	 * netfilter). Important to keep this on a seperate cacheline.
+	 * Its primarily a rebuild protection rwlock.
 	 */
 	rwlock_t		lock ____cacheline_aligned_in_smp;
 	int			secret_interval;
diff --git a/net/ipv4/inet_fragment.c b/net/ipv4/inet_fragment.c
index 1206ca6..3471599 100644
--- a/net/ipv4/inet_fragment.c
+++ b/net/ipv4/inet_fragment.c
@@ -52,20 +52,27 @@ static void inet_frag_secret_rebuild(unsigned long dummy)
 	unsigned long now = jiffies;
 	int i;
 
+	/* Per bucket lock NOT needed here, due to write lock protection */
 	write_lock(&f->lock);
+
 	get_random_bytes(&f->rnd, sizeof(u32));
 	for (i = 0; i < INETFRAGS_HASHSZ; i++) {
+		struct inet_frag_bucket *hb;
 		struct inet_frag_queue *q;
 		struct hlist_node *n;
 
-		hlist_for_each_entry_safe(q, n, &f->hash[i], list) {
+		hb = &f->hash[i];
+		hlist_for_each_entry_safe(q, n, &hb->chain, list) {
 			unsigned int hval = f->hashfn(q);
 
 			if (hval != i) {
+				struct inet_frag_bucket *hb_dest;
+
 				hlist_del(&q->list);
 
 				/* Relink to new hash chain. */
-				hlist_add_head(&q->list, &f->hash[hval]);
+				hb_dest = &f->hash[hval];
+				hlist_add_head(&q->list, &hb_dest->chain);
 			}
 		}
 	}
@@ -78,9 +85,13 @@ void inet_frags_init(struct inet_frags *f)
 {
 	int i;
 
-	for (i = 0; i < INETFRAGS_HASHSZ; i++)
-		INIT_HLIST_HEAD(&f->hash[i]);
+	for (i = 0; i < INETFRAGS_HASHSZ; i++) {
+		struct inet_frag_bucket *hb = &f->hash[i];
 
+		spin_lock_init(&hb->chain_lock);
+		INIT_HLIST_HEAD(&hb->chain);
+		hb->chain_len = 0;
+	}
 	rwlock_init(&f->lock);
 
 	f->rnd = (u32) ((num_physpages ^ (num_physpages>>7)) ^
@@ -122,9 +133,19 @@ EXPORT_SYMBOL(inet_frags_exit_net);
 
 static inline void fq_unlink(struct inet_frag_queue *fq, struct inet_frags *f)
 {
-	write_lock(&f->lock);
+	struct inet_frag_bucket *hb;
+	unsigned int hash;
+
+	read_lock(&f->lock);
+	hash = f->hashfn(fq);
+	hb = &f->hash[hash];
+
+	spin_lock_bh(&hb->chain_lock);
 	hlist_del(&fq->list);
-	write_unlock(&f->lock);
+	hb->chain_len--;
+	spin_unlock_bh(&hb->chain_lock);
+
+	read_unlock(&f->lock);
 	inet_frag_lru_del(fq);
 }
 
@@ -226,27 +247,32 @@ static struct inet_frag_queue *inet_frag_intern(struct netns_frags *nf,
 		struct inet_frag_queue *qp_in, struct inet_frags *f,
 		void *arg)
 {
+	struct inet_frag_bucket *hb;
 	struct inet_frag_queue *qp;
 #ifdef CONFIG_SMP
 #endif
 	unsigned int hash;
 
-	write_lock(&f->lock);
+	read_lock(&f->lock); /* Protects against hash rebuild */
 	/*
 	 * While we stayed w/o the lock other CPU could update
 	 * the rnd seed, so we need to re-calculate the hash
 	 * chain. Fortunatelly the qp_in can be used to get one.
 	 */
 	hash = f->hashfn(qp_in);
+	hb = &f->hash[hash];
+	spin_lock_bh(&hb->chain_lock);
+
 #ifdef CONFIG_SMP
 	/* With SMP race we have to recheck hash table, because
 	 * such entry could be created on other cpu, while we
-	 * promoted read lock to write lock.
+	 * released the hash bucket lock.
 	 */
-	hlist_for_each_entry(qp, &f->hash[hash], list) {
+	hlist_for_each_entry(qp, &hb->chain, list) {
 		if (qp->net == nf && f->match(qp, arg)) {
 			atomic_inc(&qp->refcnt);
-			write_unlock(&f->lock);
+			spin_unlock_bh(&hb->chain_lock);
+			read_unlock(&f->lock);
 			qp_in->last_in |= INET_FRAG_COMPLETE;
 			inet_frag_put(qp_in, f);
 			return qp;
@@ -258,8 +284,10 @@ static struct inet_frag_queue *inet_frag_intern(struct netns_frags *nf,
 		atomic_inc(&qp->refcnt);
 
 	atomic_inc(&qp->refcnt);
-	hlist_add_head(&qp->list, &f->hash[hash]);
-	write_unlock(&f->lock);
+	hlist_add_head(&qp->list, &hb->chain);
+	hb->chain_len++;
+	spin_unlock_bh(&hb->chain_lock);
+	read_unlock(&f->lock);
 	inet_frag_lru_add(nf, qp);
 	return qp;
 }
@@ -300,17 +328,23 @@ struct inet_frag_queue *inet_frag_find(struct netns_frags *nf,
 		struct inet_frags *f, void *key, unsigned int hash)
 	__releases(&f->lock)
 {
+	struct inet_frag_bucket *hb;
 	struct inet_frag_queue *q;
 	int depth = 0;
 
-	hlist_for_each_entry(q, &f->hash[hash], list) {
+	hb = &f->hash[hash];
+
+	spin_lock_bh(&hb->chain_lock);
+	hlist_for_each_entry(q, &hb->chain, list) {
 		if (q->net == nf && f->match(q, key)) {
 			atomic_inc(&q->refcnt);
+			spin_unlock_bh(&hb->chain_lock);
 			read_unlock(&f->lock);
 			return q;
 		}
 		depth++;
 	}
+	spin_unlock_bh(&hb->chain_lock);
 	read_unlock(&f->lock);
 
 	if (depth <= INETFRAGS_MAXDEPTH)