[PATCH bpf-next v4 06/16] inet6: Run SK_LOOKUP BPF program on socket lookup

[Jakub Sitnicki <jakub@cloudflare.com>]

Following ipv4 stack changes, run a BPF program attached to netns before
looking up a listening socket. Program can return a listening socket to use
as result of socket lookup, fail the lookup, or take no action.

Suggested-by: Marek Majkowski <marek@cloudflare.com>
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
---

Notes:
    v4:
    - Adapt to changes in BPF prog return codes.
    - Invert return value from bpf_sk_lookup_run_v6 to true on skip reuseport.
    
    v3:
    - Use a static_key to minimize the hook overhead when not used. (Alexei)
    - Don't copy struct in6_addr when populating BPF prog context. (Martin)
    - Adapt for running an array of attached programs. (Alexei)
    - Adapt for optionally skipping reuseport selection. (Martin)

 include/linux/filter.h      | 44 +++++++++++++++++++++++++++++++++++++
 net/ipv6/inet6_hashtables.c | 35 +++++++++++++++++++++++++++++
 2 files changed, 79 insertions(+)

diff --git a/include/linux/filter.h b/include/linux/filter.h
index b9ad0fdabca5..900b71af5580 100644
--- a/include/linux/filter.h
+++ b/include/linux/filter.h
@@ -1397,4 +1397,48 @@ static inline bool bpf_sk_lookup_run_v4(struct net *net, int protocol,
 	return no_reuseport;
 }
 
+#if IS_ENABLED(CONFIG_IPV6)
+static inline bool bpf_sk_lookup_run_v6(struct net *net, int protocol,
+					const struct in6_addr *saddr,
+					const __be16 sport,
+					const struct in6_addr *daddr,
+					const u16 dport,
+					struct sock **psk)
+{
+	struct bpf_prog_array *run_array;
+	struct sock *selected_sk = NULL;
+	bool no_reuseport = false;
+
+	rcu_read_lock();
+	run_array = rcu_dereference(net->bpf.run_array[NETNS_BPF_SK_LOOKUP]);
+	if (run_array) {
+		struct bpf_sk_lookup_kern ctx = {
+			.family		= AF_INET6,
+			.protocol	= protocol,
+			.v6.saddr	= saddr,
+			.v6.daddr	= daddr,
+			.sport		= sport,
+			.dport		= dport,
+		};
+		u32 act;
+
+		act = BPF_PROG_SK_LOOKUP_RUN_ARRAY(run_array, ctx, BPF_PROG_RUN);
+		if (act == SK_PASS) {
+			selected_sk = ctx.selected_sk;
+			no_reuseport = ctx.no_reuseport;
+			goto unlock;
+		}
+
+		selected_sk = ERR_PTR(-ECONNREFUSED);
+		WARN_ONCE(act != SK_DROP,
+			  "Illegal BPF SK_LOOKUP return value %u, expect packet loss!\n",
+			  act);
+	}
+unlock:
+	rcu_read_unlock();
+	*psk = selected_sk;
+	return no_reuseport;
+}
+#endif /* IS_ENABLED(CONFIG_IPV6) */
+
 #endif /* __LINUX_FILTER_H__ */
diff --git a/net/ipv6/inet6_hashtables.c b/net/ipv6/inet6_hashtables.c
index 03942eef8ab6..2d3add9e6116 100644
--- a/net/ipv6/inet6_hashtables.c
+++ b/net/ipv6/inet6_hashtables.c
@@ -21,6 +21,8 @@
 #include <net/ip.h>
 #include <net/sock_reuseport.h>
 
+extern struct inet_hashinfo tcp_hashinfo;
+
 u32 inet6_ehashfn(const struct net *net,
 		  const struct in6_addr *laddr, const u16 lport,
 		  const struct in6_addr *faddr, const __be16 fport)
@@ -159,6 +161,31 @@ static struct sock *inet6_lhash2_lookup(struct net *net,
 	return result;
 }
 
+static inline struct sock *inet6_lookup_run_bpf(struct net *net,
+						struct inet_hashinfo *hashinfo,
+						struct sk_buff *skb, int doff,
+						const struct in6_addr *saddr,
+						const __be16 sport,
+						const struct in6_addr *daddr,
+						const u16 hnum)
+{
+	struct sock *sk, *reuse_sk;
+	bool no_reuseport;
+
+	if (hashinfo != &tcp_hashinfo)
+		return NULL; /* only TCP is supported */
+
+	no_reuseport = bpf_sk_lookup_run_v6(net, IPPROTO_TCP,
+					    saddr, sport, daddr, hnum, &sk);
+	if (no_reuseport || IS_ERR_OR_NULL(sk))
+		return sk;
+
+	reuse_sk = lookup_reuseport(net, sk, skb, doff, saddr, sport, daddr, hnum);
+	if (reuse_sk)
+		sk = reuse_sk;
+	return sk;
+}
+
 struct sock *inet6_lookup_listener(struct net *net,
 		struct inet_hashinfo *hashinfo,
 		struct sk_buff *skb, int doff,
@@ -170,6 +197,14 @@ struct sock *inet6_lookup_listener(struct net *net,
 	struct sock *result = NULL;
 	unsigned int hash2;
 
+	/* Lookup redirect from BPF */
+	if (static_branch_unlikely(&bpf_sk_lookup_enabled)) {
+		result = inet6_lookup_run_bpf(net, hashinfo, skb, doff,
+					      saddr, sport, daddr, hnum);
+		if (result)
+			goto done;
+	}
+
 	hash2 = ipv6_portaddr_hash(net, daddr, hnum);
 	ilb2 = inet_lhash2_bucket(hashinfo, hash2);
 
-- 
2.25.4