Merge branch 'net/rds/4.3-v3' of git://git.kernel.org/pub/scm/linux/k…

…ernel/git/ssantosh/linux

Santosh Shilimkar says:

====================
RDS: connection scalability and performance improvements

[v4]
Re-sending the same patches from v3 again since my repost of
patch 05/14 from v3 was whitespace damaged.

[v3]
Updated patch "[PATCH v2 05/14] RDS: defer the over_batch work to
send worker" as per David Miller's comment [4] to avoid the magic
value usage. Patch now makes use of already available but unused
send_batch_count module parameter. Rest of the patches are same as
earlier version v2 [3]

[v2]:
Dropped "[PATCH 05/15] RDS: increase size of hash-table to 8K" from
earlier version [1]. I plan to address the hash table scalability using
re-sizable hash tables as suggested by David Laight and David Miller [2]

This series addresses RDS connection bottlenecks on massive workloads and
improve the RDMA performance almost by 3X. RDS TCP also gets a small gain
of about 12%.

RDS is being used in massive systems with high scalability where several
hundred thousand end points and tens of thousands of local processes
are operating in tens of thousand sockets. Being RC(reliable connection),
socket bind and release happens very often and any inefficiencies in
bind hash look ups hurts the overall system performance. RDS bin hash-table
uses global spin-lock which is the biggest bottleneck. To make matter worst,
it uses rcu inside global lock for hash buckets.
This is being addressed by simply using per bucket rw lock which makes the
locking simple and very efficient. The hash table size is still an issue and
I plan to address it by using re-sizable hash tables as suggested on the list.

For RDS RDMA improvement, the completion handling is revamped so that we
can do batch completions. Both send and receive completion handlers are
split logically to achieve the same. RDS 8K messages being one of the
key usecase, mr pool is adapted to have the 8K mrs along with default 1M
mrs. And while doing this, few fixes and couple of bottlenecks seen with
rds_sendmsg() are addressed.

Series applies against 4.3-rc1 as well net-next. Its tested on Oracle
hardware with IB fabric for both bcopy as well as RDMA mode. RDS TCP is
tested with iXGB NIC. Like last time, iWARP transport is untested with
these changes. The patchset is also available at below git repo:

git://git.kernel.org/pub/scm/linux/kernel/git/ssantosh/linux.git net/rds/4.3-v3

As a side note, the IB HCA driver I used for testing misses at least 3
important patches in upstream to see the full blown IB performance and
am hoping to get that in mainline with help of them.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

net/rds/af_rds.c

@@ -72,13 +72,7 @@ static int rds_release(struct socket *sock)
 	rds_clear_recv_queue(rs);
 	rds_cong_remove_socket(rs);
 
-	/*
-	 * the binding lookup hash uses rcu, we need to
-	 * make sure we synchronize_rcu before we free our
-	 * entry
-	 */
 	rds_remove_bound(rs);
-	synchronize_rcu();
 
 	rds_send_drop_to(rs, NULL);
 	rds_rdma_drop_keys(rs);
@@ -588,6 +582,8 @@ static int rds_init(void)
 {
 	int ret;
 
+	rds_bind_lock_init();
+
 	ret = rds_conn_init();
 	if (ret)
 		goto out;

net/rds/bind.c

@@ -38,48 +38,50 @@
 #include <linux/ratelimit.h>
 #include "rds.h"
 
+struct bind_bucket {
+	rwlock_t                lock;
+	struct hlist_head	head;
+};
+
 #define BIND_HASH_SIZE 1024
-static struct hlist_head bind_hash_table[BIND_HASH_SIZE];
-static DEFINE_SPINLOCK(rds_bind_lock);
+static struct bind_bucket bind_hash_table[BIND_HASH_SIZE];
 
-static struct hlist_head *hash_to_bucket(__be32 addr, __be16 port)
+static struct bind_bucket *hash_to_bucket(__be32 addr, __be16 port)
 {
 	return bind_hash_table + (jhash_2words((u32)addr, (u32)port, 0) &
 				  (BIND_HASH_SIZE - 1));
 }
 
-static struct rds_sock *rds_bind_lookup(__be32 addr, __be16 port,
+/* must hold either read or write lock (write lock for insert != NULL) */
+static struct rds_sock *rds_bind_lookup(struct bind_bucket *bucket,
+					__be32 addr, __be16 port,
 					struct rds_sock *insert)
 {
 	struct rds_sock *rs;
-	struct hlist_head *head = hash_to_bucket(addr, port);
+	struct hlist_head *head = &bucket->head;
 	u64 cmp;
 	u64 needle = ((u64)be32_to_cpu(addr) << 32) | be16_to_cpu(port);
 
-	rcu_read_lock();
-	hlist_for_each_entry_rcu(rs, head, rs_bound_node) {
+	hlist_for_each_entry(rs, head, rs_bound_node) {
 		cmp = ((u64)be32_to_cpu(rs->rs_bound_addr) << 32) |
 		      be16_to_cpu(rs->rs_bound_port);
 
 		if (cmp == needle) {
-			rcu_read_unlock();
+			rds_sock_addref(rs);
 			return rs;
 		}
 	}
-	rcu_read_unlock();
 
 	if (insert) {
 		/*
 		 * make sure our addr and port are set before
-		 * we are added to the list, other people
-		 * in rcu will find us as soon as the
-		 * hlist_add_head_rcu is done
+		 * we are added to the list.
 		 */
 		insert->rs_bound_addr = addr;
 		insert->rs_bound_port = port;
 		rds_sock_addref(insert);
 
-		hlist_add_head_rcu(&insert->rs_bound_node, head);
+		hlist_add_head(&insert->rs_bound_node, head);
 	}
 	return NULL;
 }
@@ -93,16 +95,21 @@ static struct rds_sock *rds_bind_lookup(__be32 addr, __be16 port,
 struct rds_sock *rds_find_bound(__be32 addr, __be16 port)
 {
 	struct rds_sock *rs;
+	unsigned long flags;
+	struct bind_bucket *bucket = hash_to_bucket(addr, port);
 
-	rs = rds_bind_lookup(addr, port, NULL);
+	read_lock_irqsave(&bucket->lock, flags);
+	rs = rds_bind_lookup(bucket, addr, port, NULL);
+	read_unlock_irqrestore(&bucket->lock, flags);
 
-	if (rs && !sock_flag(rds_rs_to_sk(rs), SOCK_DEAD))
-		rds_sock_addref(rs);
-	else
+	if (rs && sock_flag(rds_rs_to_sk(rs), SOCK_DEAD)) {
+		rds_sock_put(rs);
 		rs = NULL;
+	}
 
 	rdsdebug("returning rs %p for %pI4:%u\n", rs, &addr,
 		ntohs(port));
+
 	return rs;
 }
 
@@ -112,6 +119,7 @@ static int rds_add_bound(struct rds_sock *rs, __be32 addr, __be16 *port)
 	unsigned long flags;
 	int ret = -EADDRINUSE;
 	u16 rover, last;
+	struct bind_bucket *bucket;
 
 	if (*port != 0) {
 		rover = be16_to_cpu(*port);
@@ -121,42 +129,48 @@ static int rds_add_bound(struct rds_sock *rs, __be32 addr, __be16 *port)
 		last = rover - 1;
 	}
 
-	spin_lock_irqsave(&rds_bind_lock, flags);
-
 	do {
+		struct rds_sock *rrs;
 		if (rover == 0)
 			rover++;
-		if (!rds_bind_lookup(addr, cpu_to_be16(rover), rs)) {
+
+		bucket = hash_to_bucket(addr, cpu_to_be16(rover));
+		write_lock_irqsave(&bucket->lock, flags);
+		rrs = rds_bind_lookup(bucket, addr, cpu_to_be16(rover), rs);
+		write_unlock_irqrestore(&bucket->lock, flags);
+		if (!rrs) {
 			*port = rs->rs_bound_port;
 			ret = 0;
 			rdsdebug("rs %p binding to %pI4:%d\n",
 			  rs, &addr, (int)ntohs(*port));
 			break;
+		} else {
+			rds_sock_put(rrs);
 		}
 	} while (rover++ != last);
 
-	spin_unlock_irqrestore(&rds_bind_lock, flags);
-
 	return ret;
 }
 
 void rds_remove_bound(struct rds_sock *rs)
 {
 	unsigned long flags;
+	struct bind_bucket *bucket =
+		hash_to_bucket(rs->rs_bound_addr, rs->rs_bound_port);
 
-	spin_lock_irqsave(&rds_bind_lock, flags);
+	write_lock_irqsave(&bucket->lock, flags);
 
 	if (rs->rs_bound_addr) {
 		rdsdebug("rs %p unbinding from %pI4:%d\n",
 		  rs, &rs->rs_bound_addr,
 		  ntohs(rs->rs_bound_port));
 
-		hlist_del_init_rcu(&rs->rs_bound_node);
+		hlist_del_init(&rs->rs_bound_node);
 		rds_sock_put(rs);
 		rs->rs_bound_addr = 0;
 	}
 
-	spin_unlock_irqrestore(&rds_bind_lock, flags);
+	write_unlock_irqrestore(&bucket->lock, flags);
 }
 
 int rds_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len)
@@ -200,9 +214,13 @@ int rds_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len)
 
 out:
 	release_sock(sk);
-
-	/* we might have called rds_remove_bound on error */
-	if (ret)
-		synchronize_rcu();
 	return ret;
 }
+
+void rds_bind_lock_init(void)
+{
+	int i;
+
+	for (i = 0; i < BIND_HASH_SIZE; i++)
+		rwlock_init(&bind_hash_table[i].lock);
+}

net/rds/ib.c

@@ -43,14 +43,14 @@
 #include "rds.h"
 #include "ib.h"
 
-static unsigned int fmr_pool_size = RDS_FMR_POOL_SIZE;
-unsigned int fmr_message_size = RDS_FMR_SIZE + 1; /* +1 allows for unaligned MRs */
+unsigned int rds_ib_fmr_1m_pool_size = RDS_FMR_1M_POOL_SIZE;
+unsigned int rds_ib_fmr_8k_pool_size = RDS_FMR_8K_POOL_SIZE;
 unsigned int rds_ib_retry_count = RDS_IB_DEFAULT_RETRY_COUNT;
 
-module_param(fmr_pool_size, int, 0444);
-MODULE_PARM_DESC(fmr_pool_size, " Max number of fmr per HCA");
-module_param(fmr_message_size, int, 0444);
-MODULE_PARM_DESC(fmr_message_size, " Max size of a RDMA transfer");
+module_param(rds_ib_fmr_1m_pool_size, int, 0444);
+MODULE_PARM_DESC(rds_ib_fmr_1m_pool_size, " Max number of 1M fmr per HCA");
+module_param(rds_ib_fmr_8k_pool_size, int, 0444);
+MODULE_PARM_DESC(rds_ib_fmr_8k_pool_size, " Max number of 8K fmr per HCA");
 module_param(rds_ib_retry_count, int, 0444);
 MODULE_PARM_DESC(rds_ib_retry_count, " Number of hw retries before reporting an error");
 
@@ -97,8 +97,10 @@ static void rds_ib_dev_free(struct work_struct *work)
 	struct rds_ib_device *rds_ibdev = container_of(work,
 					struct rds_ib_device, free_work);
 
-	if (rds_ibdev->mr_pool)
-		rds_ib_destroy_mr_pool(rds_ibdev->mr_pool);
+	if (rds_ibdev->mr_8k_pool)
+		rds_ib_destroy_mr_pool(rds_ibdev->mr_8k_pool);
+	if (rds_ibdev->mr_1m_pool)
+		rds_ib_destroy_mr_pool(rds_ibdev->mr_1m_pool);
 	if (rds_ibdev->pd)
 		ib_dealloc_pd(rds_ibdev->pd);
 
@@ -148,9 +150,13 @@ static void rds_ib_add_one(struct ib_device *device)
 	rds_ibdev->max_sge = min(dev_attr->max_sge, RDS_IB_MAX_SGE);
 
 	rds_ibdev->fmr_max_remaps = dev_attr->max_map_per_fmr?: 32;
-	rds_ibdev->max_fmrs = dev_attr->max_fmr ?
-			min_t(unsigned int, dev_attr->max_fmr, fmr_pool_size) :
-			fmr_pool_size;
+	rds_ibdev->max_1m_fmrs = dev_attr->max_mr ?
+		min_t(unsigned int, (dev_attr->max_mr / 2),
+		      rds_ib_fmr_1m_pool_size) : rds_ib_fmr_1m_pool_size;
+
+	rds_ibdev->max_8k_fmrs = dev_attr->max_mr ?
+		min_t(unsigned int, ((dev_attr->max_mr / 2) * RDS_MR_8K_SCALE),
+		      rds_ib_fmr_8k_pool_size) : rds_ib_fmr_8k_pool_size;
 
 	rds_ibdev->max_initiator_depth = dev_attr->max_qp_init_rd_atom;
 	rds_ibdev->max_responder_resources = dev_attr->max_qp_rd_atom;
@@ -162,12 +168,25 @@ static void rds_ib_add_one(struct ib_device *device)
 		goto put_dev;
 	}
 
-	rds_ibdev->mr_pool = rds_ib_create_mr_pool(rds_ibdev);
-	if (IS_ERR(rds_ibdev->mr_pool)) {
-		rds_ibdev->mr_pool = NULL;
+	rds_ibdev->mr_1m_pool =
+		rds_ib_create_mr_pool(rds_ibdev, RDS_IB_MR_1M_POOL);
+	if (IS_ERR(rds_ibdev->mr_1m_pool)) {
+		rds_ibdev->mr_1m_pool = NULL;
 		goto put_dev;
 	}
 
+	rds_ibdev->mr_8k_pool =
+		rds_ib_create_mr_pool(rds_ibdev, RDS_IB_MR_8K_POOL);
+	if (IS_ERR(rds_ibdev->mr_8k_pool)) {
+		rds_ibdev->mr_8k_pool = NULL;
+		goto put_dev;
+	}
+
+	rdsdebug("RDS/IB: max_mr = %d, max_wrs = %d, max_sge = %d, fmr_max_remaps = %d, max_1m_fmrs = %d, max_8k_fmrs = %d\n",
+		 dev_attr->max_fmr, rds_ibdev->max_wrs, rds_ibdev->max_sge,
+		 rds_ibdev->fmr_max_remaps, rds_ibdev->max_1m_fmrs,
+		 rds_ibdev->max_8k_fmrs);
+
 	INIT_LIST_HEAD(&rds_ibdev->ipaddr_list);
 	INIT_LIST_HEAD(&rds_ibdev->conn_list);