Merge branch 'ipv6-auto-flow-labels'

Tom Herbert says:

====================
ipv6: Turn on auto IPv6 flow labels by default

BSD (MacOS) has already turned on flow labels by default and this does
not seem to be causing any problems in the Internet. Let's go ahead
and turn them on by default. We'll continue to monitor for any devices
start choking on them.

Flow labels are important since they are the desired solution for
network devices to perform ECMP and RSS (RFC6437 and RFC6438).
Traditionally, devices perform a 5-tuple hash on packets that
includes port numbers. For the most part, these devices can only
compute 5-tuple hashes for TCP and UDP. This severely limits our ability
to get good network load balancing for other protocols (IPIP, GRE,ESP,
etc.), and hence we are limited in using other protocols. Unfortunately,
this method is accepted as the de facto standard to the extent that
there are several proposals to encapsulate protocols in UDP _just_ for
the purposes for getting ECMP to work. With hosts generating flow labels
and devices taking them as input into ECMP (several already do), we can
start to fix this fundamental problem.

This patch set:
 - Changes IPV6_FLOWINFO sockopt to be opt-out of flow labels for
   connections rather than opt-in
 - Disable flow label state ranges sysctl by default
 - Enable auto flow labels sysctl by default

v2:
  - Added functions to create an skb->hash based on flowi4 and flowi6.
    These are called in output path when creating a packet
  - Call skb_get_hash_flowi6 in ip6_make_flowlabel
  - Implement the auto_flowlabels sysctl as a mode for auto flowlabels.
    There are four modes which correspond to flow labels being enabled
    and whether socket option can be used to opt in or opt out of
    using them
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Documentation/networking/ip-sysctl.txt

@@ -1215,14 +1215,20 @@ flowlabel_consistency - BOOLEAN
 	FALSE: disabled
 	Default: TRUE
 
-auto_flowlabels - BOOLEAN
-	Automatically generate flow labels based based on a flow hash
-	of the packet. This allows intermediate devices, such as routers,
-	to idenfify packet flows for mechanisms like Equal Cost Multipath
+auto_flowlabels - INTEGER
+	Automatically generate flow labels based on a flow hash of the
+	packet. This allows intermediate devices, such as routers, to
+	identify packet flows for mechanisms like Equal Cost Multipath
 	Routing (see RFC 6438).
-	TRUE: enabled
-	FALSE: disabled
-	Default: false
+	0: automatic flow labels are completely disabled
+	1: automatic flow labels are enabled by default, they can be
+	   disabled on a per socket basis using the IPV6_AUTOFLOWLABEL
+	   socket option
+	2: automatic flow labels are allowed, they may be enabled on a
+	   per socket basis using the IPV6_AUTOFLOWLABEL socket option
+	3: automatic flow labels are enabled and enforced, they cannot
+	   be disabled by the socket option
+	Default: 1
 
 flowlabel_state_ranges - BOOLEAN
 	Split the flow label number space into two ranges. 0-0x7FFFF is

include/linux/skbuff.h

@@ -37,6 +37,7 @@
 #include <net/flow_dissector.h>
 #include <linux/splice.h>
 #include <linux/in6.h>
+#include <net/flow.h>
 
 /* A. Checksumming of received packets by device.
  *
@@ -945,6 +946,26 @@ static inline __u32 skb_get_hash(struct sk_buff *skb)
 	return skb->hash;
 }
 
+__u32 __skb_get_hash_flowi6(struct sk_buff *skb, struct flowi6 *fl6);
+
+static inline __u32 skb_get_hash_flowi6(struct sk_buff *skb, struct flowi6 *fl6)
+{
+	if (!skb->l4_hash && !skb->sw_hash)
+		__skb_get_hash_flowi6(skb, fl6);
+
+	return skb->hash;
+}
+
+__u32 __skb_get_hash_flowi4(struct sk_buff *skb, struct flowi4 *fl);
+
+static inline __u32 skb_get_hash_flowi4(struct sk_buff *skb, struct flowi4 *fl4)
+{
+	if (!skb->l4_hash && !skb->sw_hash)
+		__skb_get_hash_flowi4(skb, fl4);
+
+	return skb->hash;
+}
+
 __u32 skb_get_hash_perturb(const struct sk_buff *skb, u32 perturb);
 
 static inline __u32 skb_get_hash_raw(const struct sk_buff *skb)

include/net/ipv6.h

@@ -707,35 +707,69 @@ static inline void iph_to_flow_copy_v6addrs(struct flow_keys *flow,
 }
 
 #if IS_ENABLED(CONFIG_IPV6)
+
+/* Sysctl settings for net ipv6.auto_flowlabels */
+#define IP6_AUTO_FLOW_LABEL_OFF		0
+#define IP6_AUTO_FLOW_LABEL_OPTOUT	1
+#define IP6_AUTO_FLOW_LABEL_OPTIN	2
+#define IP6_AUTO_FLOW_LABEL_FORCED	3
+
+#define IP6_AUTO_FLOW_LABEL_MAX		IP6_AUTO_FLOW_LABEL_FORCED
+
+#define IP6_DEFAULT_AUTO_FLOW_LABELS	IP6_AUTO_FLOW_LABEL_OPTOUT
+
 static inline __be32 ip6_make_flowlabel(struct net *net, struct sk_buff *skb,
-					__be32 flowlabel, bool autolabel)
+					__be32 flowlabel, bool autolabel,
+					struct flowi6 *fl6)
 {
-	if (!flowlabel && (autolabel || net->ipv6.sysctl.auto_flowlabels)) {
 	u32 hash;
 
-		hash = skb_get_hash(skb);
+	if (flowlabel ||
+	    net->ipv6.sysctl.auto_flowlabels == IP6_AUTO_FLOW_LABEL_OFF ||
+	    (!autolabel &&
+	     net->ipv6.sysctl.auto_flowlabels != IP6_AUTO_FLOW_LABEL_FORCED))
+		return flowlabel;
+
+	hash = skb_get_hash_flowi6(skb, fl6);
 
 	/* Since this is being sent on the wire obfuscate hash a bit
 	 * to minimize possbility that any useful information to an
 	 * attacker is leaked. Only lower 20 bits are relevant.
 	 */
-		hash ^= hash >> 12;
+	rol32(hash, 16);
 
 	flowlabel = (__force __be32)hash & IPV6_FLOWLABEL_MASK;
 
 	if (net->ipv6.sysctl.flowlabel_state_ranges)
 		flowlabel |= IPV6_FLOWLABEL_STATELESS_FLAG;
-	}
 
 	return flowlabel;
 }
+
+static inline int ip6_default_np_autolabel(struct net *net)
+{
+	switch (net->ipv6.sysctl.auto_flowlabels) {
+	case IP6_AUTO_FLOW_LABEL_OFF:
+	case IP6_AUTO_FLOW_LABEL_OPTIN:
+	default:
+		return 0;
+	case IP6_AUTO_FLOW_LABEL_OPTOUT:
+	case IP6_AUTO_FLOW_LABEL_FORCED:
+		return 1;
+	}
+}
 #else
 static inline void ip6_set_txhash(struct sock *sk) { }
 static inline __be32 ip6_make_flowlabel(struct net *net, struct sk_buff *skb,
-					__be32 flowlabel, bool autolabel)
+					__be32 flowlabel, bool autolabel,
+					struct flowi6 *fl6)
 {
 	return flowlabel;
 }
+static inline int ip6_default_np_autolabel(struct net *net)
+{
+	return 0;
+}
 #endif
 
 

net/core/flow_dissector.c

@@ -590,6 +590,15 @@ void make_flow_keys_digest(struct flow_keys_digest *digest,
 }
 EXPORT_SYMBOL(make_flow_keys_digest);
 
+static inline void __skb_set_sw_hash(struct sk_buff *skb, u32 hash,
+				     struct flow_keys *keys)
+{
+	if (keys->ports.ports)
+		skb->l4_hash = 1;
+	skb->sw_hash = 1;
+	skb->hash = hash;
+}
+
 /**
  * __skb_get_hash: calculate a flow hash
  * @skb: sk_buff to calculate flow hash from
@@ -609,10 +618,8 @@ void __skb_get_hash(struct sk_buff *skb)
 	hash = ___skb_get_hash(skb, &keys, hashrnd);
 	if (!hash)
 		return;
-	if (keys.ports.ports)
-		skb->l4_hash = 1;
-	skb->sw_hash = 1;
-	skb->hash = hash;
+
+	__skb_set_sw_hash(skb, hash, &keys);
 }
 EXPORT_SYMBOL(__skb_get_hash);
 
@@ -624,6 +631,49 @@ __u32 skb_get_hash_perturb(const struct sk_buff *skb, u32 perturb)
 }
 EXPORT_SYMBOL(skb_get_hash_perturb);
 
+__u32 __skb_get_hash_flowi6(struct sk_buff *skb, struct flowi6 *fl6)
+{
+	struct flow_keys keys;
+
+	memset(&keys, 0, sizeof(keys));
+
+	memcpy(&keys.addrs.v6addrs.src, &fl6->saddr,
+	       sizeof(keys.addrs.v6addrs.src));
+	memcpy(&keys.addrs.v6addrs.dst, &fl6->daddr,
+	       sizeof(keys.addrs.v6addrs.dst));
+	keys.control.addr_type = FLOW_DISSECTOR_KEY_IPV6_ADDRS;
+	keys.ports.src = fl6->fl6_sport;
+	keys.ports.dst = fl6->fl6_dport;
+	keys.keyid.keyid = fl6->fl6_gre_key;
+	keys.tags.flow_label = (__force u32)fl6->flowlabel;
+	keys.basic.ip_proto = fl6->flowi6_proto;
+
+	__skb_set_sw_hash(skb, flow_hash_from_keys(&keys), &keys);
+
+	return skb->hash;
+}
+EXPORT_SYMBOL(__skb_get_hash_flowi6);
+
+__u32 __skb_get_hash_flowi4(struct sk_buff *skb, struct flowi4 *fl4)
+{
+	struct flow_keys keys;
+
+	memset(&keys, 0, sizeof(keys));
+
+	keys.addrs.v4addrs.src = fl4->saddr;
+	keys.addrs.v4addrs.dst = fl4->daddr;
+	keys.control.addr_type = FLOW_DISSECTOR_KEY_IPV4_ADDRS;
+	keys.ports.src = fl4->fl4_sport;
+	keys.ports.dst = fl4->fl4_dport;
+	keys.keyid.keyid = fl4->fl4_gre_key;
+	keys.basic.ip_proto = fl4->flowi4_proto;
+
+	__skb_set_sw_hash(skb, flow_hash_from_keys(&keys), &keys);
+
+	return skb->hash;
+}
+EXPORT_SYMBOL(__skb_get_hash_flowi4);
+
 u32 __skb_get_poff(const struct sk_buff *skb, void *data,
 		   const struct flow_keys *keys, int hlen)
 {

net/ipv6/af_inet6.c

@@ -197,6 +197,7 @@ static int inet6_create(struct net *net, struct socket *sock, int protocol,
 	np->mcast_hops	= IPV6_DEFAULT_MCASTHOPS;
 	np->mc_loop	= 1;
 	np->pmtudisc	= IPV6_PMTUDISC_WANT;
+	np->autoflowlabel = ip6_default_np_autolabel(sock_net(sk));
 	sk->sk_ipv6only	= net->ipv6.sysctl.bindv6only;
 
 	/* Init the ipv4 part of the socket since we can have sockets
@@ -767,10 +768,10 @@ static int __net_init inet6_net_init(struct net *net)
 	net->ipv6.sysctl.bindv6only = 0;
 	net->ipv6.sysctl.icmpv6_time = 1*HZ;
 	net->ipv6.sysctl.flowlabel_consistency = 1;
-	net->ipv6.sysctl.auto_flowlabels = 0;
+	net->ipv6.sysctl.auto_flowlabels = IP6_DEFAULT_AUTO_FLOW_LABELS;
 	net->ipv6.sysctl.idgen_retries = 3;
 	net->ipv6.sysctl.idgen_delay = 1 * HZ;
-	net->ipv6.sysctl.flowlabel_state_ranges = 1;
+	net->ipv6.sysctl.flowlabel_state_ranges = 0;
 	atomic_set(&net->ipv6.fib6_sernum, 1);
 
 	err = ipv6_init_mibs(net);

net/ipv6/ip6_gre.c

@@ -728,7 +728,7 @@ static netdev_tx_t ip6gre_xmit2(struct sk_buff *skb,
 	 */
 	ipv6h = ipv6_hdr(skb);
 	ip6_flow_hdr(ipv6h, INET_ECN_encapsulate(0, dsfield),
-		     ip6_make_flowlabel(net, skb, fl6->flowlabel, false));
+		     ip6_make_flowlabel(net, skb, fl6->flowlabel, true, fl6));
 	ipv6h->hop_limit = tunnel->parms.hop_limit;
 	ipv6h->nexthdr = proto;
 	ipv6h->saddr = fl6->saddr;
@@ -1182,7 +1182,8 @@ static int ip6gre_header(struct sk_buff *skb, struct net_device *dev,
 
 	ip6_flow_hdr(ipv6h, 0,
 		     ip6_make_flowlabel(dev_net(dev), skb,
-					t->fl.u.ip6.flowlabel, false));
+					t->fl.u.ip6.flowlabel, true,
+					&t->fl.u.ip6));
 	ipv6h->hop_limit = t->parms.hop_limit;
 	ipv6h->nexthdr = NEXTHDR_GRE;
 	ipv6h->saddr = t->parms.laddr;

net/ipv6/ip6_output.c

@@ -207,7 +207,7 @@ int ip6_xmit(struct sock *sk, struct sk_buff *skb, struct flowi6 *fl6,
 		hlimit = ip6_dst_hoplimit(dst);
 
 	ip6_flow_hdr(hdr, tclass, ip6_make_flowlabel(net, skb, fl6->flowlabel,
-						     np->autoflowlabel));
+						     np->autoflowlabel, fl6));
 
 	hdr->payload_len = htons(seg_len);
 	hdr->nexthdr = proto;
@@ -1649,7 +1649,7 @@ struct sk_buff *__ip6_make_skb(struct sock *sk,
 
 	ip6_flow_hdr(hdr, v6_cork->tclass,
 		     ip6_make_flowlabel(net, skb, fl6->flowlabel,
-					np->autoflowlabel));
+					np->autoflowlabel, fl6));
 	hdr->hop_limit = v6_cork->hop_limit;
 	hdr->nexthdr = proto;
 	hdr->saddr = fl6->saddr;

net/ipv6/ip6_tunnel.c

@@ -1095,7 +1095,7 @@ static int ip6_tnl_xmit2(struct sk_buff *skb,
 	skb_reset_network_header(skb);
 	ipv6h = ipv6_hdr(skb);
 	ip6_flow_hdr(ipv6h, INET_ECN_encapsulate(0, dsfield),
-		     ip6_make_flowlabel(net, skb, fl6->flowlabel, false));
+		     ip6_make_flowlabel(net, skb, fl6->flowlabel, true, fl6));
 	ipv6h->hop_limit = t->parms.hop_limit;
 	ipv6h->nexthdr = proto;
 	ipv6h->saddr = fl6->saddr;

net/ipv6/sysctl_net_ipv6.c

@@ -17,6 +17,9 @@
 #include <net/inet_frag.h>
 
 static int one = 1;
+static int auto_flowlabels_min;
+static int auto_flowlabels_max = IP6_AUTO_FLOW_LABEL_MAX;
+
 
 static struct ctl_table ipv6_table_template[] = {
 	{
@@ -45,7 +48,9 @@ static struct ctl_table ipv6_table_template[] = {
 		.data		= &init_net.ipv6.sysctl.auto_flowlabels,
 		.maxlen		= sizeof(int),
 		.mode		= 0644,
-		.proc_handler	= proc_dointvec
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= &auto_flowlabels_min,
+		.extra2		= &auto_flowlabels_max
 	},
 	{
 		.procname	= "fwmark_reflect",