Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/jesse/openvswitch

Jesse Gross says:

====================
A number of significant new features and optimizations for net-next/3.12.
Highlights are:
 * "Megaflows", an optimization that allows userspace to specify which
   flow fields were used to compute the results of the flow lookup.
   This allows for a major reduction in flow setups (the major
   performance bottleneck in Open vSwitch) without reducing flexibility.
 * Converting netlink dump operations to use RCU, allowing for
   additional parallelism in userspace.
 * Matching and modifying SCTP protocol fields.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Documentation/networking/openvswitch.txt

@@ -91,6 +91,46 @@ Often we ellipsize arguments not important to the discussion, e.g.:
     in_port(1), eth(...), eth_type(0x0800), ipv4(...), tcp(...)
 
 
+Wildcarded flow key format
+--------------------------
+
+A wildcarded flow is described with two sequences of Netlink attributes
+passed over the Netlink socket. A flow key, exactly as described above, and an
+optional corresponding flow mask.
+
+A wildcarded flow can represent a group of exact match flows. Each '1' bit
+in the mask specifies a exact match with the corresponding bit in the flow key.
+A '0' bit specifies a don't care bit, which will match either a '1' or '0' bit
+of a incoming packet. Using wildcarded flow can improve the flow set up rate
+by reduce the number of new flows need to be processed by the user space program.
+
+Support for the mask Netlink attribute is optional for both the kernel and user
+space program. The kernel can ignore the mask attribute, installing an exact
+match flow, or reduce the number of don't care bits in the kernel to less than
+what was specified by the user space program. In this case, variations in bits
+that the kernel does not implement will simply result in additional flow setups.
+The kernel module will also work with user space programs that neither support
+nor supply flow mask attributes.
+
+Since the kernel may ignore or modify wildcard bits, it can be difficult for
+the userspace program to know exactly what matches are installed. There are
+two possible approaches: reactively install flows as they miss the kernel
+flow table (and therefore not attempt to determine wildcard changes at all)
+or use the kernel's response messages to determine the installed wildcards.
+
+When interacting with userspace, the kernel should maintain the match portion
+of the key exactly as originally installed. This will provides a handle to
+identify the flow for all future operations. However, when reporting the
+mask of an installed flow, the mask should include any restrictions imposed
+by the kernel.
+
+The behavior when using overlapping wildcarded flows is undefined. It is the
+responsibility of the user space program to ensure that any incoming packet
+can match at most one flow, wildcarded or not. The current implementation
+performs best-effort detection of overlapping wildcarded flows and may reject
+some but not all of them. However, this behavior may change in future versions.
+
+
 Basic rule for evolving flow keys
 ---------------------------------
 

include/net/ipv6.h

@@ -41,6 +41,7 @@
 #define NEXTHDR_ICMP		58	/* ICMP for IPv6. */
 #define NEXTHDR_NONE		59	/* No next header */
 #define NEXTHDR_DEST		60	/* Destination options header. */
+#define NEXTHDR_SCTP		132	/* SCTP message. */
 #define NEXTHDR_MOBILITY	135	/* Mobility header. */
 
 #define NEXTHDR_MAX		255

include/uapi/linux/openvswitch.h

 
 /*
- * Copyright (c) 2007-2011 Nicira Networks.
+ * Copyright (c) 2007-2013 Nicira, Inc.
  *
  * This program is free software; you can redistribute it and/or
  * modify it under the terms of version 2 of the GNU General Public
@@ -259,6 +259,7 @@ enum ovs_key_attr {
 	OVS_KEY_ATTR_ND,        /* struct ovs_key_nd */
 	OVS_KEY_ATTR_SKB_MARK,  /* u32 skb mark */
 	OVS_KEY_ATTR_TUNNEL,    /* Nested set of ovs_tunnel attributes */
+	OVS_KEY_ATTR_SCTP,      /* struct ovs_key_sctp */
 
 #ifdef __KERNEL__
 	OVS_KEY_ATTR_IPV4_TUNNEL,  /* struct ovs_key_ipv4_tunnel */
@@ -333,6 +334,11 @@ struct ovs_key_udp {
 	__be16 udp_dst;
 };
 
+struct ovs_key_sctp {
+	__be16 sctp_src;
+	__be16 sctp_dst;
+};
+
 struct ovs_key_icmp {
 	__u8 icmp_type;
 	__u8 icmp_code;
@@ -379,6 +385,12 @@ struct ovs_key_nd {
  * @OVS_FLOW_ATTR_CLEAR: If present in a %OVS_FLOW_CMD_SET request, clears the
  * last-used time, accumulated TCP flags, and statistics for this flow.
  * Otherwise ignored in requests.  Never present in notifications.
+ * @OVS_FLOW_ATTR_MASK: Nested %OVS_KEY_ATTR_* attributes specifying the
+ * mask bits for wildcarded flow match. Mask bit value '1' specifies exact
+ * match with corresponding flow key bit, while mask bit value '0' specifies
+ * a wildcarded match. Omitting attribute is treated as wildcarding all
+ * corresponding fields. Optional for all requests. If not present,
+ * all flow key bits are exact match bits.
  *
  * These attributes follow the &struct ovs_header within the Generic Netlink
  * payload for %OVS_FLOW_* commands.
@@ -391,6 +403,7 @@ enum ovs_flow_attr {
 	OVS_FLOW_ATTR_TCP_FLAGS, /* 8-bit OR'd TCP flags. */
 	OVS_FLOW_ATTR_USED,      /* u64 msecs last used in monotonic time. */
 	OVS_FLOW_ATTR_CLEAR,     /* Flag to clear stats, tcp_flags, used. */
+	OVS_FLOW_ATTR_MASK,      /* Sequence of OVS_KEY_ATTR_* attributes. */
 	__OVS_FLOW_ATTR_MAX
 };
 

net/openvswitch/Kconfig

@@ -4,6 +4,7 @@
 
 config OPENVSWITCH
 	tristate "Open vSwitch"
+	select LIBCRC32C
 	---help---
 	  Open vSwitch is a multilayer Ethernet switch targeted at virtualized
 	  environments.  In addition to supporting a variety of features

net/openvswitch/Makefile

@@ -10,10 +10,13 @@ openvswitch-y := \
 	dp_notify.o \
 	flow.o \
 	vport.o \
-	vport-gre.o \
 	vport-internal_dev.o \
 	vport-netdev.o
 
 ifneq ($(CONFIG_OPENVSWITCH_VXLAN),)
 openvswitch-y += vport-vxlan.o
 endif
+
+ifneq ($(CONFIG_OPENVSWITCH_GRE),)
+openvswitch-y += vport-gre.o
+endif

net/openvswitch/actions.c

 /*
- * Copyright (c) 2007-2012 Nicira, Inc.
+ * Copyright (c) 2007-2013 Nicira, Inc.
  *
  * This program is free software; you can redistribute it and/or
  * modify it under the terms of version 2 of the GNU General Public
@@ -22,6 +22,7 @@
 #include <linux/in.h>
 #include <linux/ip.h>
 #include <linux/openvswitch.h>
+#include <linux/sctp.h>
 #include <linux/tcp.h>
 #include <linux/udp.h>
 #include <linux/in6.h>
@@ -31,6 +32,7 @@
 #include <net/ipv6.h>
 #include <net/checksum.h>
 #include <net/dsfield.h>
+#include <net/sctp/checksum.h>
 
 #include "datapath.h"
 #include "vport.h"
@@ -352,6 +354,39 @@ static int set_tcp(struct sk_buff *skb, const struct ovs_key_tcp *tcp_port_key)
 	return 0;
 }
 
+static int set_sctp(struct sk_buff *skb,
+		     const struct ovs_key_sctp *sctp_port_key)
+{
+	struct sctphdr *sh;
+	int err;
+	unsigned int sctphoff = skb_transport_offset(skb);
+
+	err = make_writable(skb, sctphoff + sizeof(struct sctphdr));
+	if (unlikely(err))
+		return err;
+
+	sh = sctp_hdr(skb);
+	if (sctp_port_key->sctp_src != sh->source ||
+	    sctp_port_key->sctp_dst != sh->dest) {
+		__le32 old_correct_csum, new_csum, old_csum;
+
+		old_csum = sh->checksum;
+		old_correct_csum = sctp_compute_cksum(skb, sctphoff);
+
+		sh->source = sctp_port_key->sctp_src;
+		sh->dest = sctp_port_key->sctp_dst;
+
+		new_csum = sctp_compute_cksum(skb, sctphoff);
+
+		/* Carry any checksum errors through. */
+		sh->checksum = old_csum ^ old_correct_csum ^ new_csum;
+
+		skb->rxhash = 0;
+	}
+
+	return 0;
+}
+
 static int do_output(struct datapath *dp, struct sk_buff *skb, int out_port)
 {
 	struct vport *vport;
@@ -376,8 +411,10 @@ static int output_userspace(struct datapath *dp, struct sk_buff *skb,
 	const struct nlattr *a;
 	int rem;
 
+	BUG_ON(!OVS_CB(skb)->pkt_key);
+
 	upcall.cmd = OVS_PACKET_CMD_ACTION;
-	upcall.key = &OVS_CB(skb)->flow->key;
+	upcall.key = OVS_CB(skb)->pkt_key;
 	upcall.userdata = NULL;
 	upcall.portid = 0;
 
@@ -459,6 +496,10 @@ static int execute_set_action(struct sk_buff *skb,
 	case OVS_KEY_ATTR_UDP:
 		err = set_udp(skb, nla_data(nested_attr));
 		break;
+
+	case OVS_KEY_ATTR_SCTP:
+		err = set_sctp(skb, nla_data(nested_attr));
+		break;
 	}
 
 	return err;

net/openvswitch/datapath.h

@@ -88,11 +88,13 @@ struct datapath {
 /**
  * struct ovs_skb_cb - OVS data in skb CB
  * @flow: The flow associated with this packet.  May be %NULL if no flow.
+ * @pkt_key: The flow information extracted from the packet.  Must be nonnull.
  * @tun_key: Key for the tunnel that encapsulated this packet. NULL if the
  * packet is not being tunneled.
  */
 struct ovs_skb_cb {
 	struct sw_flow		*flow;
+	struct sw_flow_key	*pkt_key;
 	struct ovs_key_ipv4_tunnel  *tun_key;
 };
 #define OVS_CB(skb) ((struct ovs_skb_cb *)(skb)->cb)
@@ -183,4 +185,8 @@ struct sk_buff *ovs_vport_cmd_build_info(struct vport *, u32 pid, u32 seq,
 
 int ovs_execute_actions(struct datapath *dp, struct sk_buff *skb);
 void ovs_dp_notify_wq(struct work_struct *work);
+
+#define OVS_NLERR(fmt, ...) \
+	pr_info_once("netlink: " fmt, ##__VA_ARGS__)
+
 #endif /* datapath.h */

net/openvswitch/flow.h

 /*
- * Copyright (c) 2007-2011 Nicira, Inc.
+ * Copyright (c) 2007-2013 Nicira, Inc.
  *
  * This program is free software; you can redistribute it and/or
  * modify it under the terms of version 2 of the GNU General Public
@@ -33,6 +33,8 @@
 #include <net/inet_ecn.h>
 
 struct sk_buff;
+struct sw_flow_mask;
+struct flow_table;
 
 struct sw_flow_actions {
 	struct rcu_head rcu;
@@ -97,8 +99,8 @@ struct sw_flow_key {
 			} addr;
 			union {
 				struct {
-					__be16 src;		/* TCP/UDP source port. */
-					__be16 dst;		/* TCP/UDP destination port. */
+					__be16 src;		/* TCP/UDP/SCTP source port. */
+					__be16 dst;		/* TCP/UDP/SCTP destination port. */
 				} tp;
 				struct {
 					u8 sha[ETH_ALEN];	/* ARP source hardware address. */
@@ -113,8 +115,8 @@ struct sw_flow_key {
 			} addr;
 			__be32 label;			/* IPv6 flow label. */
 			struct {
-				__be16 src;		/* TCP/UDP source port. */
-				__be16 dst;		/* TCP/UDP destination port. */
+				__be16 src;		/* TCP/UDP/SCTP source port. */
+				__be16 dst;		/* TCP/UDP/SCTP destination port. */
 			} tp;
 			struct {
 				struct in6_addr target;	/* ND target address. */
@@ -123,7 +125,7 @@ struct sw_flow_key {
 			} nd;
 		} ipv6;
 	};
-};
+} __aligned(__alignof__(long));
 
 struct sw_flow {
 	struct rcu_head rcu;
@@ -131,6 +133,8 @@ struct sw_flow {
 	u32 hash;
 
 	struct sw_flow_key key;
+	struct sw_flow_key unmasked_key;
+	struct sw_flow_mask *mask;
 	struct sw_flow_actions __rcu *sf_acts;
 
 	spinlock_t lock;	/* Lock for values below. */
@@ -140,6 +144,20 @@ struct sw_flow {
 	u8 tcp_flags;		/* Union of seen TCP flags. */
 };
 
+struct sw_flow_key_range {
+	size_t start;
+	size_t end;
+};
+
+struct sw_flow_match {
+	struct sw_flow_key *key;
+	struct sw_flow_key_range range;
+	struct sw_flow_mask *mask;
+};
+
+void ovs_match_init(struct sw_flow_match *match,
+		struct sw_flow_key *key, struct sw_flow_mask *mask);
+
 struct arp_eth_header {
 	__be16      ar_hrd;	/* format of hardware address   */
 	__be16      ar_pro;	/* format of protocol address   */
@@ -159,21 +177,21 @@ void ovs_flow_exit(void);
 
 struct sw_flow *ovs_flow_alloc(void);
 void ovs_flow_deferred_free(struct sw_flow *);
-void ovs_flow_free(struct sw_flow *flow);
+void ovs_flow_free(struct sw_flow *, bool deferred);
 
 struct sw_flow_actions *ovs_flow_actions_alloc(int actions_len);
 void ovs_flow_deferred_free_acts(struct sw_flow_actions *);
 
-int ovs_flow_extract(struct sk_buff *, u16 in_port, struct sw_flow_key *,
-		     int *key_lenp);
+int ovs_flow_extract(struct sk_buff *, u16 in_port, struct sw_flow_key *);
 void ovs_flow_used(struct sw_flow *, struct sk_buff *);
 u64 ovs_flow_used_time(unsigned long flow_jiffies);
-
-int ovs_flow_to_nlattrs(const struct sw_flow_key *, struct sk_buff *);
-int ovs_flow_from_nlattrs(struct sw_flow_key *swkey, int *key_lenp,
+int ovs_flow_to_nlattrs(const struct sw_flow_key *,
+		const struct sw_flow_key *, struct sk_buff *);
+int ovs_match_from_nlattrs(struct sw_flow_match *match,
+		      const struct nlattr *,
 		      const struct nlattr *);
-int ovs_flow_metadata_from_nlattrs(struct sw_flow *flow, int key_len,
-				  const struct nlattr *attr);
+int ovs_flow_metadata_from_nlattrs(struct sw_flow *flow,
+		const struct nlattr *attr);
 
 #define MAX_ACTIONS_BUFSIZE    (32 * 1024)
 #define TBL_MIN_BUCKETS		1024
@@ -182,6 +200,7 @@ struct flow_table {
 	struct flex_array *buckets;
 	unsigned int count, n_buckets;
 	struct rcu_head rcu;
+	struct list_head *mask_list;
 	int node_ver;
 	u32 hash_seed;
 	bool keep_flows;
@@ -197,22 +216,44 @@ static inline int ovs_flow_tbl_need_to_expand(struct flow_table *table)
 	return (table->count > table->n_buckets);
 }
 
-struct sw_flow *ovs_flow_tbl_lookup(struct flow_table *table,
-				    struct sw_flow_key *key, int len);
-void ovs_flow_tbl_destroy(struct flow_table *table);
-void ovs_flow_tbl_deferred_destroy(struct flow_table *table);
+struct sw_flow *ovs_flow_lookup(struct flow_table *,
+				const struct sw_flow_key *);
+struct sw_flow *ovs_flow_lookup_unmasked_key(struct flow_table *table,
+				    struct sw_flow_match *match);
+
+void ovs_flow_tbl_destroy(struct flow_table *table, bool deferred);
 struct flow_table *ovs_flow_tbl_alloc(int new_size);
 struct flow_table *ovs_flow_tbl_expand(struct flow_table *table);
 struct flow_table *ovs_flow_tbl_rehash(struct flow_table *table);
-void ovs_flow_tbl_insert(struct flow_table *table, struct sw_flow *flow,
-			 struct sw_flow_key *key, int key_len);
-void ovs_flow_tbl_remove(struct flow_table *table, struct sw_flow *flow);
 
-struct sw_flow *ovs_flow_tbl_next(struct flow_table *table, u32 *bucket, u32 *idx);
+void ovs_flow_insert(struct flow_table *table, struct sw_flow *flow);
+void ovs_flow_remove(struct flow_table *table, struct sw_flow *flow);
+
+struct sw_flow *ovs_flow_dump_next(struct flow_table *table, u32 *bucket, u32 *idx);
 extern const int ovs_key_lens[OVS_KEY_ATTR_MAX + 1];
 int ovs_ipv4_tun_from_nlattr(const struct nlattr *attr,
-			 struct ovs_key_ipv4_tunnel *tun_key);
+			     struct sw_flow_match *match, bool is_mask);
 int ovs_ipv4_tun_to_nlattr(struct sk_buff *skb,
-			const struct ovs_key_ipv4_tunnel *tun_key);
+			   const struct ovs_key_ipv4_tunnel *tun_key,
+			   const struct ovs_key_ipv4_tunnel *output);
+
+bool ovs_flow_cmp_unmasked_key(const struct sw_flow *flow,
+		const struct sw_flow_key *key, int key_end);
+
+struct sw_flow_mask {
+	int ref_count;
+	struct rcu_head rcu;
+	struct list_head list;
+	struct sw_flow_key_range range;
+	struct sw_flow_key key;
+};
 
+struct sw_flow_mask *ovs_sw_flow_mask_alloc(void);
+void ovs_sw_flow_mask_add_ref(struct sw_flow_mask *);
+void ovs_sw_flow_mask_del_ref(struct sw_flow_mask *, bool deferred);
+void ovs_sw_flow_mask_insert(struct flow_table *, struct sw_flow_mask *);
+struct sw_flow_mask *ovs_sw_flow_mask_find(const struct flow_table *,
+		const struct sw_flow_mask *);
+void ovs_flow_key_mask(struct sw_flow_key *dst, const struct sw_flow_key *src,
+		       const struct sw_flow_mask *mask);
 #endif /* flow.h */

net/openvswitch/vport-gre.c

@@ -16,7 +16,6 @@
  * 02110-1301, USA
  */
 
-#ifdef CONFIG_OPENVSWITCH_GRE
 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 
 #include <linux/if.h>
@@ -271,5 +270,3 @@ const struct vport_ops ovs_gre_vport_ops = {
 	.get_name	= gre_get_name,
 	.send		= gre_tnl_send,
 };
-
-#endif /* OPENVSWITCH_GRE */

net/openvswitch/vport-netdev.c

@@ -25,6 +25,7 @@
 #include <linux/llc.h>
 #include <linux/rtnetlink.h>
 #include <linux/skbuff.h>
+#include <linux/openvswitch.h>
 
 #include <net/llc.h>
 
@@ -74,6 +75,15 @@ static rx_handler_result_t netdev_frame_hook(struct sk_buff **pskb)
 	return RX_HANDLER_CONSUMED;
 }
 
+static struct net_device *get_dpdev(struct datapath *dp)
+{
+	struct vport *local;
+
+	local = ovs_vport_ovsl(dp, OVSP_LOCAL);
+	BUG_ON(!local);
+	return netdev_vport_priv(local)->dev;
+}
+
 static struct vport *netdev_create(const struct vport_parms *parms)
 {
 	struct vport *vport;
@@ -103,10 +113,15 @@ static struct vport *netdev_create(const struct vport_parms *parms)
 	}
 
 	rtnl_lock();
+	err = netdev_master_upper_dev_link(netdev_vport->dev,
+					   get_dpdev(vport->dp));
+	if (err)
+		goto error_unlock;
+
 	err = netdev_rx_handler_register(netdev_vport->dev, netdev_frame_hook,
 					 vport);
 	if (err)
-		goto error_unlock;
+		goto error_master_upper_dev_unlink;
 
 	dev_set_promiscuity(netdev_vport->dev, 1);
 	netdev_vport->dev->priv_flags |= IFF_OVS_DATAPATH;
@@ -114,6 +129,8 @@ static struct vport *netdev_create(const struct vport_parms *parms)
 
 	return vport;
 
+error_master_upper_dev_unlink:
+	netdev_upper_dev_unlink(netdev_vport->dev, get_dpdev(vport->dp));
 error_unlock:
 	rtnl_unlock();
 error_put:
@@ -140,6 +157,7 @@ static void netdev_destroy(struct vport *vport)
 	rtnl_lock();
 	netdev_vport->dev->priv_flags &= ~IFF_OVS_DATAPATH;
 	netdev_rx_handler_unregister(netdev_vport->dev);
+	netdev_upper_dev_unlink(netdev_vport->dev, get_dpdev(vport->dp));
 	dev_set_promiscuity(netdev_vport->dev, -1);
 	rtnl_unlock();
 

net/openvswitch/vport.c

@@ -203,7 +203,7 @@ struct vport *ovs_vport_add(const struct vport_parms *parms)
  *	ovs_vport_set_options - modify existing vport device (for kernel callers)
  *
  * @vport: vport to modify.
- * @port: New configuration.
+ * @options: New configuration.
  *
  * Modifies an existing device with the specified configuration (which is
  * dependent on device type).  ovs_mutex must be held.
@@ -328,6 +328,7 @@ int ovs_vport_get_options(const struct vport *vport, struct sk_buff *skb)
  *
  * @vport: vport that received the packet
  * @skb: skb that was received
+ * @tun_key: tunnel (if any) that carried packet
  *
  * Must be called with rcu_read_lock.  The packet cannot be shared and
  * skb->data should point to the Ethernet header.