Merge branch 'ipv6-Separate-data-structures-for-FIB-and-data-path'

David Ahern says:

====================
net/ipv6: Separate data structures for FIB and data path

IPv6 uses the same data struct for both control plane (FIB entries) and
data path (dst entries). This struct has elements needed for both paths
adding memory overhead and complexity (taking a dst hold in most places
but an additional reference on rt6i_ref in a few). Furthermore, because
of the dst_alloc tie, all FIB entries are allocated with GFP_ATOMIC.

This patch set separates FIB entries from dst entries, better aligning
IPv6 code with IPv4, simplifying the reference counting and allowing
FIB entries added by userspace (not autoconf) to use GFP_KERNEL. It is
first step to a number of performance and scalability changes.

The end result of this patch set:
  - FIB entries (fib6_info):
        /* size: 208, cachelines: 4, members: 25 */
        /* sum members: 207, holes: 1, sum holes: 1 */

  - dst entries (rt6_info)
       /* size: 240, cachelines: 4, members: 11 */

Versus the the single rt6_info struct today for both paths:
      /* size: 320, cachelines: 5, members: 28 */

This amounts to a 35% reduction in memory use for FIB entries and a
25% reduction for dst entries.

With respect to locking FIB entries use RCU and a single atomic
counter with fib6_info_hold and fib6_info_release helpers to manage
the reference counting. dst entries use only the traditional dst
refcounts with dst_hold and dst_release.

FIB entries for host routes are referenced by inet6_ifaddr and
ifacaddr6. In both cases, additional holds are taken -- similar to
what is done for devices.

This set is the first of many changes to improve the scalability of the
IPv6 code. Follow on changes include:
- consolidating duplicate fib6_info references like IPv4 does with
  duplicate fib_info

- moving fib6_info into a slab cache to avoid allocation roundups to
  power of 2 (the 208 size becomes a 256 actual allocation)

- Allow FIB lookups without generating a dst (e.g., most rt6_lookup
  users just want to verify the egress device). Means moving dst
  allocation to the other side of fib6_rule_lookup which again aligns
  with IPv4 behavior

- using separate standalone nexthop objects which have performance
  benefits beyond fib_info consolidation

At this point I am not seeing any refcount leaks or underflows, no
oops or bug_ons, or warnings from kasan, so I think it is ready for
others to beat up on it finding errors in code paths I have missed.

v2 changes
- rebased to top of tree
- improved commit message on patch 7

v1 changes
- rebased to top of tree
- fix memory leak of metrics as noted by Ido
- MTU fixes based on pmtu tests (thanks Stefano Brivio for writing)

RFC v2 changes
- improved commit messages
- move common metrics code from dst.c to net/ipv4/metrics.c (comment
  from DaveM)
- address comments from Wei Wang and Martin KaFai Lau (let me know if
  I missed something)
- fixes detected by kernel test robots
  + added fib6_metric_set to change metric on a FIB entry which could
    be pointing to read-only dst_default_metrics
  + 0day testing found a problem with an intermediate patch; added
    dst_hold_safe on rt->from. Code is removed 3 patches later
- allow cacheinfo to handle NULL dst; means only expires is pushed to
  userspace
====================

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

5 files changed, 123 insertions, 83 deletions

diff --git a/include/net/if_inet6.h b/include/net/if_inet6.h
index d4088d1a688d..d6089b2e64fe 100644
--- a/include/net/if_inet6.h
+++ b/include/net/if_inet6.h
@@ -64,7 +64,7 @@ struct inet6_ifaddr {
 	struct delayed_work	dad_work;
 
 	struct inet6_dev	*idev;
-	struct rt6_info		*rt;
+	struct fib6_info	*rt;
 
 	struct hlist_node	addr_lst;
 	struct list_head	if_list;
@@ -144,7 +144,7 @@ struct ipv6_ac_socklist {
 struct ifacaddr6 {
 	struct in6_addr		aca_addr;
 	struct inet6_dev	*aca_idev;
-	struct rt6_info		*aca_rt;
+	struct fib6_info	*aca_rt;
 	struct ifacaddr6	*aca_next;
 	int			aca_users;
 	refcount_t		aca_refcnt;
diff --git a/include/net/ip.h b/include/net/ip.h
index ecffd843e7b8..dc4a2d6e58a5 100644
--- a/include/net/ip.h
+++ b/include/net/ip.h
@@ -396,6 +396,9 @@ static inline unsigned int ip_skb_dst_mtu(struct sock *sk,
 	return min(READ_ONCE(skb_dst(skb)->dev->mtu), IP_MAX_MTU);
 }
 
+int ip_metrics_convert(struct net *net, struct nlattr *fc_mx, int fc_mx_len,
+		       u32 *metrics);
+
 u32 ip_idents_reserve(u32 hash, int segs);
 void __ip_select_ident(struct net *net, struct iphdr *iph, int segs);
 
diff --git a/include/net/ip6_fib.h b/include/net/ip6_fib.h
index 5e86fd9dc857..a36116b92100 100644
--- a/include/net/ip6_fib.h
+++ b/include/net/ip6_fib.h
@@ -38,6 +38,7 @@
 #endif
 
 struct rt6_info;
+struct fib6_info;
 
 struct fib6_config {
 	u32		fc_table;
@@ -74,12 +75,12 @@ struct fib6_node {
 #ifdef CONFIG_IPV6_SUBTREES
 	struct fib6_node __rcu	*subtree;
 #endif
-	struct rt6_info __rcu	*leaf;
+	struct fib6_info __rcu	*leaf;
 
 	__u16			fn_bit;		/* bit key */
 	__u16			fn_flags;
 	int			fn_sernum;
-	struct rt6_info __rcu	*rr_ptr;
+	struct fib6_info __rcu	*rr_ptr;
 	struct rcu_head		rcu;
 };
 
@@ -94,11 +95,6 @@ struct fib6_gc_args {
 #define FIB6_SUBTREE(fn)	(rcu_dereference_protected((fn)->subtree, 1))
 #endif
 
-struct mx6_config {
-	const u32 *mx;
-	DECLARE_BITMAP(mx_valid, RTAX_MAX);
-};
-
 /*
  *	routing information
  *
@@ -127,56 +123,72 @@ struct rt6_exception {
 #define FIB6_EXCEPTION_BUCKET_SIZE (1 << FIB6_EXCEPTION_BUCKET_SIZE_SHIFT)
 #define FIB6_MAX_DEPTH 5
 
-struct rt6_info {
-	struct dst_entry		dst;
-	struct rt6_info __rcu		*rt6_next;
-	struct rt6_info			*from;
+struct fib6_nh {
+	struct in6_addr		nh_gw;
+	struct net_device	*nh_dev;
+	struct lwtunnel_state	*nh_lwtstate;
 
-	/*
-	 * Tail elements of dst_entry (__refcnt etc.)
-	 * and these elements (rarely used in hot path) are in
-	 * the same cache line.
-	 */
+	unsigned int		nh_flags;
+	atomic_t		nh_upper_bound;
+	int			nh_weight;
+};
+
+struct fib6_info {
 	struct fib6_table		*rt6i_table;
+	struct fib6_info __rcu		*rt6_next;
 	struct fib6_node __rcu		*rt6i_node;
 
-	struct in6_addr			rt6i_gateway;
-
 	/* Multipath routes:
-	 * siblings is a list of rt6_info that have the the same metric/weight,
+	 * siblings is a list of fib6_info that have the the same metric/weight,
 	 * destination, but not the same gateway. nsiblings is just a cache
 	 * to speed up lookup.
 	 */
 	struct list_head		rt6i_siblings;
 	unsigned int			rt6i_nsiblings;
-	atomic_t			rt6i_nh_upper_bound;
 
 	atomic_t			rt6i_ref;
+	struct inet6_dev		*rt6i_idev;
+	unsigned long			expires;
+	struct dst_metrics		*fib6_metrics;
+#define fib6_pmtu		fib6_metrics->metrics[RTAX_MTU-1]
 
-	unsigned int			rt6i_nh_flags;
-
-	/* These are in a separate cache line. */
-	struct rt6key			rt6i_dst ____cacheline_aligned_in_smp;
+	struct rt6key			rt6i_dst;
 	u32				rt6i_flags;
 	struct rt6key			rt6i_src;
 	struct rt6key			rt6i_prefsrc;
 
-	struct list_head		rt6i_uncached;
-	struct uncached_list		*rt6i_uncached_list;
-
-	struct inet6_dev		*rt6i_idev;
 	struct rt6_info * __percpu	*rt6i_pcpu;
 	struct rt6_exception_bucket __rcu *rt6i_exception_bucket;
 
 	u32				rt6i_metric;
-	u32				rt6i_pmtu;
-	/* more non-fragment space at head required */
-	int				rt6i_nh_weight;
-	unsigned short			rt6i_nfheader_len;
 	u8				rt6i_protocol;
+	u8				fib6_type;
 	u8				exception_bucket_flushed:1,
 					should_flush:1,
-					unused:6;
+					dst_nocount:1,
+					dst_nopolicy:1,
+					dst_host:1,
+					unused:3;
+
+	struct fib6_nh			fib6_nh;
+};
+
+struct rt6_info {
+	struct dst_entry		dst;
+	struct fib6_info		*from;
+
+	struct rt6key			rt6i_dst;
+	struct rt6key			rt6i_src;
+	struct in6_addr			rt6i_gateway;
+	struct inet6_dev		*rt6i_idev;
+	u32				rt6i_flags;
+	struct rt6key			rt6i_prefsrc;
+
+	struct list_head		rt6i_uncached;
+	struct uncached_list		*rt6i_uncached_list;
+
+	/* more non-fragment space at head required */
+	unsigned short			rt6i_nfheader_len;
 };
 
 #define for_each_fib6_node_rt_rcu(fn)					\
@@ -192,6 +204,26 @@ static inline struct inet6_dev *ip6_dst_idev(struct dst_entry *dst)
 	return ((struct rt6_info *)dst)->rt6i_idev;
 }
 
+static inline void fib6_clean_expires(struct fib6_info *f6i)
+{
+	f6i->rt6i_flags &= ~RTF_EXPIRES;
+	f6i->expires = 0;
+}
+
+static inline void fib6_set_expires(struct fib6_info *f6i,
+				    unsigned long expires)
+{
+	f6i->expires = expires;
+	f6i->rt6i_flags |= RTF_EXPIRES;
+}
+
+static inline bool fib6_check_expired(const struct fib6_info *f6i)
+{
+	if (f6i->rt6i_flags & RTF_EXPIRES)
+		return time_after(jiffies, f6i->expires);
+	return false;
+}
+
 static inline void rt6_clean_expires(struct rt6_info *rt)
 {
 	rt->rt6i_flags &= ~RTF_EXPIRES;
@@ -206,11 +238,9 @@ static inline void rt6_set_expires(struct rt6_info *rt, unsigned long expires)
 
 static inline void rt6_update_expires(struct rt6_info *rt0, int timeout)
 {
-	struct rt6_info *rt;
+	if (!(rt0->rt6i_flags & RTF_EXPIRES) && rt0->from)
+		rt0->dst.expires = rt0->from->expires;
 
-	for (rt = rt0; rt && !(rt->rt6i_flags & RTF_EXPIRES); rt = rt->from);
-	if (rt && rt != rt0)
-		rt0->dst.expires = rt->dst.expires;
 	dst_set_expires(&rt0->dst, timeout);
 	rt0->rt6i_flags |= RTF_EXPIRES;
 }
@@ -220,7 +250,7 @@ static inline void rt6_update_expires(struct rt6_info *rt0, int timeout)
  * Return true if we can get cookie safely
  * Return false if not
  */
-static inline bool rt6_get_cookie_safe(const struct rt6_info *rt,
+static inline bool rt6_get_cookie_safe(const struct fib6_info *rt,
 				       u32 *cookie)
 {
 	struct fib6_node *fn;
@@ -246,9 +276,7 @@ static inline u32 rt6_get_cookie(const struct rt6_info *rt)
 
 	if (rt->rt6i_flags & RTF_PCPU ||
 	    (unlikely(!list_empty(&rt->rt6i_uncached)) && rt->from))
-		rt = rt->from;
-
-	rt6_get_cookie_safe(rt, &cookie);
+		rt6_get_cookie_safe(rt->from, &cookie);
 
 	return cookie;
 }
@@ -262,20 +290,18 @@ static inline void ip6_rt_put(struct rt6_info *rt)
 	dst_release(&rt->dst);
 }
 
-void rt6_free_pcpu(struct rt6_info *non_pcpu_rt);
+struct fib6_info *fib6_info_alloc(gfp_t gfp_flags);
+void fib6_info_destroy(struct fib6_info *f6i);
 
-static inline void rt6_hold(struct rt6_info *rt)
+static inline void fib6_info_hold(struct fib6_info *f6i)
 {
-	atomic_inc(&rt->rt6i_ref);
+	atomic_inc(&f6i->rt6i_ref);
 }
 
-static inline void rt6_release(struct rt6_info *rt)
+static inline void fib6_info_release(struct fib6_info *f6i)
 {
-	if (atomic_dec_and_test(&rt->rt6i_ref)) {
-		rt6_free_pcpu(rt);
-		dst_dev_put(&rt->dst);
-		dst_release(&rt->dst);
-	}
+	if (f6i && atomic_dec_and_test(&f6i->rt6i_ref))
+		fib6_info_destroy(f6i);
 }
 
 enum fib6_walk_state {
@@ -291,7 +317,7 @@ enum fib6_walk_state {
 struct fib6_walker {
 	struct list_head lh;
 	struct fib6_node *root, *node;
-	struct rt6_info *leaf;
+	struct fib6_info *leaf;
 	enum fib6_walk_state state;
 	unsigned int skip;
 	unsigned int count;
@@ -355,7 +381,7 @@ typedef struct rt6_info *(*pol_lookup_t)(struct net *,
 
 struct fib6_entry_notifier_info {
 	struct fib_notifier_info info; /* must be first */
-	struct rt6_info *rt;
+	struct fib6_info *rt;
 };
 
 /*
@@ -377,15 +403,14 @@ struct fib6_node *fib6_locate(struct fib6_node *root,
 			      const struct in6_addr *saddr, int src_len,
 			      bool exact_match);
 
-void fib6_clean_all(struct net *net, int (*func)(struct rt6_info *, void *arg),
+void fib6_clean_all(struct net *net, int (*func)(struct fib6_info *, void *arg),
 		    void *arg);
 
-int fib6_add(struct fib6_node *root, struct rt6_info *rt,
-	     struct nl_info *info, struct mx6_config *mxc,
-	     struct netlink_ext_ack *extack);
-int fib6_del(struct rt6_info *rt, struct nl_info *info);
+int fib6_add(struct fib6_node *root, struct fib6_info *rt,
+	     struct nl_info *info, struct netlink_ext_ack *extack);
+int fib6_del(struct fib6_info *rt, struct nl_info *info);
 
-void inet6_rt_notify(int event, struct rt6_info *rt, struct nl_info *info,
+void inet6_rt_notify(int event, struct fib6_info *rt, struct nl_info *info,
 		     unsigned int flags);
 
 void fib6_run_gc(unsigned long expires, struct net *net, bool force);
@@ -408,8 +433,14 @@ void __net_exit fib6_notifier_exit(struct net *net);
 unsigned int fib6_tables_seq_read(struct net *net);
 int fib6_tables_dump(struct net *net, struct notifier_block *nb);
 
-void fib6_update_sernum(struct rt6_info *rt);
-void fib6_update_sernum_upto_root(struct net *net, struct rt6_info *rt);
+void fib6_update_sernum(struct net *net, struct fib6_info *rt);
+void fib6_update_sernum_upto_root(struct net *net, struct fib6_info *rt);
+
+void fib6_metric_set(struct fib6_info *f6i, int metric, u32 val);
+static inline bool fib6_metric_locked(struct fib6_info *f6i, int metric)
+{
+	return !!(f6i->fib6_metrics->metrics[RTAX_LOCK - 1] & (1 << metric));
+}
 
 #ifdef CONFIG_IPV6_MULTIPLE_TABLES
 int fib6_rules_init(void);
diff --git a/include/net/ip6_route.h b/include/net/ip6_route.h
index 08b132381984..d5fb1e4ae7ac 100644
--- a/include/net/ip6_route.h
+++ b/include/net/ip6_route.h
@@ -66,7 +66,7 @@ static inline bool rt6_need_strict(const struct in6_addr *daddr)
 		(IPV6_ADDR_MULTICAST | IPV6_ADDR_LINKLOCAL | IPV6_ADDR_LOOPBACK);
 }
 
-static inline bool rt6_qualify_for_ecmp(const struct rt6_info *rt)
+static inline bool rt6_qualify_for_ecmp(const struct fib6_info *rt)
 {
 	return (rt->rt6i_flags & (RTF_GATEWAY|RTF_ADDRCONF|RTF_DYNAMIC)) ==
 	       RTF_GATEWAY;
@@ -100,22 +100,21 @@ void ip6_route_cleanup(void);
 
 int ipv6_route_ioctl(struct net *net, unsigned int cmd, void __user *arg);
 
-int ip6_route_add(struct fib6_config *cfg, struct netlink_ext_ack *extack);
-int ip6_ins_rt(struct rt6_info *);
-int ip6_del_rt(struct rt6_info *);
+int ip6_route_add(struct fib6_config *cfg, gfp_t gfp_flags,
+		  struct netlink_ext_ack *extack);
+int ip6_ins_rt(struct net *net, struct fib6_info *rt);
+int ip6_del_rt(struct net *net, struct fib6_info *rt);
 
-void rt6_flush_exceptions(struct rt6_info *rt);
-int rt6_remove_exception_rt(struct rt6_info *rt);
-void rt6_age_exceptions(struct rt6_info *rt, struct fib6_gc_args *gc_args,
+void rt6_flush_exceptions(struct fib6_info *rt);
+void rt6_age_exceptions(struct fib6_info *rt, struct fib6_gc_args *gc_args,
 			unsigned long now);
 
-static inline int ip6_route_get_saddr(struct net *net, struct rt6_info *rt,
+static inline int ip6_route_get_saddr(struct net *net, struct fib6_info *rt,
 				      const struct in6_addr *daddr,
 				      unsigned int prefs,
 				      struct in6_addr *saddr)
 {
-	struct inet6_dev *idev =
-			rt ? ip6_dst_idev((struct dst_entry *)rt) : NULL;
+	struct inet6_dev *idev = rt ? rt->rt6i_idev : NULL;
 	int err = 0;
 
 	if (rt && rt->rt6i_prefsrc.plen)
@@ -137,8 +136,9 @@ struct dst_entry *icmp6_dst_alloc(struct net_device *dev, struct flowi6 *fl6);
 
 void fib6_force_start_gc(struct net *net);
 
-struct rt6_info *addrconf_dst_alloc(struct inet6_dev *idev,
-				    const struct in6_addr *addr, bool anycast);
+struct fib6_info *addrconf_dst_alloc(struct net *net, struct inet6_dev *idev,
+				     const struct in6_addr *addr, bool anycast,
+				     gfp_t gfp_flags);
 
 struct rt6_info *ip6_dst_alloc(struct net *net, struct net_device *dev,
 			       int flags);
@@ -147,9 +147,11 @@ struct rt6_info *ip6_dst_alloc(struct net *net, struct net_device *dev,
  *	support functions for ND
  *
  */
-struct rt6_info *rt6_get_dflt_router(const struct in6_addr *addr,
+struct fib6_info *rt6_get_dflt_router(struct net *net,
+				     const struct in6_addr *addr,
 				     struct net_device *dev);
-struct rt6_info *rt6_add_dflt_router(const struct in6_addr *gwaddr,
+struct fib6_info *rt6_add_dflt_router(struct net *net,
+				     const struct in6_addr *gwaddr,
 				     struct net_device *dev, unsigned int pref);
 
 void rt6_purge_dflt_routers(struct net *net);
@@ -174,14 +176,14 @@ struct rt6_rtnl_dump_arg {
 	struct net *net;
 };
 
-int rt6_dump_route(struct rt6_info *rt, void *p_arg);
+int rt6_dump_route(struct fib6_info *rt, void *p_arg);
 void rt6_mtu_change(struct net_device *dev, unsigned int mtu);
 void rt6_remove_prefsrc(struct inet6_ifaddr *ifp);
 void rt6_clean_tohost(struct net *net, struct in6_addr *gateway);
 void rt6_sync_up(struct net_device *dev, unsigned int nh_flags);
 void rt6_disable_ip(struct net_device *dev, unsigned long event);
 void rt6_sync_down_dev(struct net_device *dev, unsigned long event);
-void rt6_multipath_rebalance(struct rt6_info *rt);
+void rt6_multipath_rebalance(struct fib6_info *rt);
 
 void rt6_uncached_list_add(struct rt6_info *rt);
 void rt6_uncached_list_del(struct rt6_info *rt);
@@ -269,12 +271,15 @@ static inline struct in6_addr *rt6_nexthop(struct rt6_info *rt,
 		return daddr;
 }
 
-static inline bool rt6_duplicate_nexthop(struct rt6_info *a, struct rt6_info *b)
+static inline bool rt6_duplicate_nexthop(struct fib6_info *a, struct fib6_info *b)
 {
-	return a->dst.dev == b->dst.dev &&
+	return a->fib6_nh.nh_dev == b->fib6_nh.nh_dev &&
 	       a->rt6i_idev == b->rt6i_idev &&
-	       ipv6_addr_equal(&a->rt6i_gateway, &b->rt6i_gateway) &&
-	       !lwtunnel_cmp_encap(a->dst.lwtstate, b->dst.lwtstate);
+	       ipv6_addr_equal(&a->fib6_nh.nh_gw, &b->fib6_nh.nh_gw) &&
+	       !lwtunnel_cmp_encap(a->fib6_nh.nh_lwtstate, b->fib6_nh.nh_lwtstate);
 }
 
+struct neighbour *ip6_neigh_lookup(const struct in6_addr *gw,
+				   struct net_device *dev, struct sk_buff *skb,
+				   const void *daddr);
 #endif
diff --git a/include/net/netns/ipv6.h b/include/net/netns/ipv6.h
index c29f09cfc9d7..97b3a54579c8 100644
--- a/include/net/netns/ipv6.h
+++ b/include/net/netns/ipv6.h
@@ -60,7 +60,8 @@ struct netns_ipv6 {
 #endif
 	struct xt_table		*ip6table_nat;
 #endif
-	struct rt6_info         *ip6_null_entry;
+	struct fib6_info	*fib6_null_entry;
+	struct rt6_info		*ip6_null_entry;
 	struct rt6_statistics   *rt6_stats;
 	struct timer_list       ip6_fib_timer;
 	struct hlist_head       *fib_table_hash;