ipv6: remove hard coded limitation on ipv6_pinfo

IPv6 inet sockets are supposed to have a "struct ipv6_pinfo"
field at the end of their definition, so that inet6_sk_generic()
can derive from socket size the offset of the "struct ipv6_pinfo".

This is very fragile, and prevents adding bigger alignment
in sockets, because inet6_sk_generic() does not work
if the compiler adds padding after the ipv6_pinfo component.

We are currently working on a patch series to reorganize
TCP structures for better data locality and found issues
similar to the one fixed in commit f5d547676ca0
("tcp: fix tcp_inet6_sk() for 32bit kernels")

Alternative would be to force an alignment on "struct ipv6_pinfo",
greater or equal to __alignof__(any ipv6 sock) to ensure there is
no padding. This does not look great.

v2: fix typo in mptcp_proto_v6_init() (Paolo)

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Chao Wu <wwchao@google.com>
Cc: Wei Wang <weiwan@google.com>
Cc: Coco Li <lixiaoyan@google.com>
Cc: YiFei Zhu <zhuyifei@google.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

13 files changed, 16 insertions, 20 deletions

diff --git a/include/linux/ipv6.h b/include/linux/ipv6.h
index ed3d110c2eb5..0295b47c10a3 100644
--- a/include/linux/ipv6.h
+++ b/include/linux/ipv6.h
@@ -200,14 +200,7 @@ struct inet6_cork {
 	u8 tclass;
 };
 
-/**
- * struct ipv6_pinfo - ipv6 private area
- *
- * In the struct sock hierarchy (tcp6_sock, upd6_sock, etc)
- * this _must_ be the last member, so that inet6_sk_generic
- * is able to calculate its offset from the base struct sock
- * by using the struct proto->slab_obj_size member. -acme
- */
+/* struct ipv6_pinfo - ipv6 private area */
 struct ipv6_pinfo {
 	struct in6_addr 	saddr;
 	struct in6_pktinfo	sticky_pktinfo;
@@ -307,19 +300,19 @@ struct raw6_sock {
 	__u32			offset;		/* checksum offset  */
 	struct icmp6_filter	filter;
 	__u32			ip6mr_table;
-	/* ipv6_pinfo has to be the last member of raw6_sock, see inet6_sk_generic */
+
 	struct ipv6_pinfo	inet6;
 };
 
 struct udp6_sock {
 	struct udp_sock	  udp;
-	/* ipv6_pinfo has to be the last member of udp6_sock, see inet6_sk_generic */
+
 	struct ipv6_pinfo inet6;
 };
 
 struct tcp6_sock {
 	struct tcp_sock	  tcp;
-	/* ipv6_pinfo has to be the last member of tcp6_sock, see inet6_sk_generic */
+
 	struct ipv6_pinfo inet6;
 };
 
diff --git a/include/net/sock.h b/include/net/sock.h
index 2eb916d1ff64..7ae44bf866af 100644
--- a/include/net/sock.h
+++ b/include/net/sock.h
@@ -1339,6 +1339,7 @@ struct proto {
 
 	struct kmem_cache	*slab;
 	unsigned int		obj_size;
+	unsigned int		ipv6_pinfo_offset;
 	slab_flags_t		slab_flags;
 	unsigned int		useroffset;	/* Usercopy region offset */
 	unsigned int		usersize;	/* Usercopy region size */
diff --git a/net/dccp/ipv6.c b/net/dccp/ipv6.c
index 7249ef218178..e03b5331df6d 100644
--- a/net/dccp/ipv6.c
+++ b/net/dccp/ipv6.c
@@ -1056,6 +1056,7 @@ static struct proto dccp_v6_prot = {
 	.orphan_count	   = &dccp_orphan_count,
 	.max_header	   = MAX_DCCP_HEADER,
 	.obj_size	   = sizeof(struct dccp6_sock),
+	.ipv6_pinfo_offset = offsetof(struct dccp6_sock, inet6),
 	.slab_flags	   = SLAB_TYPESAFE_BY_RCU,
 	.rsk_prot	   = &dccp6_request_sock_ops,
 	.twsk_prot	   = &dccp6_timewait_sock_ops,
diff --git a/net/dccp/ipv6.h b/net/dccp/ipv6.h
index 7e4c2a3b322b..c5d14c48def1 100644
--- a/net/dccp/ipv6.h
+++ b/net/dccp/ipv6.h
@@ -13,10 +13,6 @@
 
 struct dccp6_sock {
 	struct dccp_sock  dccp;
-	/*
-	 * ipv6_pinfo has to be the last member of dccp6_sock,
-	 * see inet6_sk_generic.
-	 */
 	struct ipv6_pinfo inet6;
 };
 
diff --git a/net/ipv6/af_inet6.c b/net/ipv6/af_inet6.c
index 5d593ddc0347..9f9c4b838664 100644
--- a/net/ipv6/af_inet6.c
+++ b/net/ipv6/af_inet6.c
@@ -102,9 +102,9 @@ bool ipv6_mod_enabled(void)
 }
 EXPORT_SYMBOL_GPL(ipv6_mod_enabled);
 
-static __inline__ struct ipv6_pinfo *inet6_sk_generic(struct sock *sk)
+static struct ipv6_pinfo *inet6_sk_generic(struct sock *sk)
 {
-	const int offset = sk->sk_prot->obj_size - sizeof(struct ipv6_pinfo);
+	const int offset = sk->sk_prot->ipv6_pinfo_offset;
 
 	return (struct ipv6_pinfo *)(((u8 *)sk) + offset);
 }
diff --git a/net/ipv6/ping.c b/net/ipv6/ping.c
index f804c11e2146..2a0e8bc07398 100644
--- a/net/ipv6/ping.c
+++ b/net/ipv6/ping.c
@@ -215,6 +215,7 @@ struct proto pingv6_prot = {
 	.get_port =	ping_get_port,
 	.put_port =	ping_unhash,
 	.obj_size =	sizeof(struct raw6_sock),
+	.ipv6_pinfo_offset = offsetof(struct raw6_sock, inet6),
 };
 EXPORT_SYMBOL_GPL(pingv6_prot);
 
diff --git a/net/ipv6/raw.c b/net/ipv6/raw.c
index ac1cef094c5f..0fcf1b890807 100644
--- a/net/ipv6/raw.c
+++ b/net/ipv6/raw.c
@@ -1216,6 +1216,7 @@ struct proto rawv6_prot = {
 	.hash		   = raw_hash_sk,
 	.unhash		   = raw_unhash_sk,
 	.obj_size	   = sizeof(struct raw6_sock),
+	.ipv6_pinfo_offset = offsetof(struct raw6_sock, inet6),
 	.useroffset	   = offsetof(struct raw6_sock, filter),
 	.usersize	   = sizeof_field(struct raw6_sock, filter),
 	.h.raw_hash	   = &raw_v6_hashinfo,
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index 4714eb695913..1b4529e833a1 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -2175,6 +2175,7 @@ struct proto tcpv6_prot = {
 	.sysctl_rmem_offset	= offsetof(struct net, ipv4.sysctl_tcp_rmem),
 	.max_header		= MAX_TCP_HEADER,
 	.obj_size		= sizeof(struct tcp6_sock),
+	.ipv6_pinfo_offset = offsetof(struct tcp6_sock, inet6),
 	.slab_flags		= SLAB_TYPESAFE_BY_RCU,
 	.twsk_prot		= &tcp6_timewait_sock_ops,
 	.rsk_prot		= &tcp6_request_sock_ops,
diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c
index b7c972aa09a7..95c75d8f73d5 100644
--- a/net/ipv6/udp.c
+++ b/net/ipv6/udp.c
@@ -1798,6 +1798,7 @@ struct proto udpv6_prot = {
 	.sysctl_wmem_offset     = offsetof(struct net, ipv4.sysctl_udp_wmem_min),
 	.sysctl_rmem_offset     = offsetof(struct net, ipv4.sysctl_udp_rmem_min),
 	.obj_size		= sizeof(struct udp6_sock),
+	.ipv6_pinfo_offset = offsetof(struct udp6_sock, inet6),
 	.h.udp_table		= NULL,
 	.diag_destroy		= udp_abort,
 };
diff --git a/net/ipv6/udplite.c b/net/ipv6/udplite.c
index 8e010d07917a..267d491e9707 100644
--- a/net/ipv6/udplite.c
+++ b/net/ipv6/udplite.c
@@ -67,6 +67,7 @@ struct proto udplitev6_prot = {
 	.sysctl_wmem_offset = offsetof(struct net, ipv4.sysctl_udp_wmem_min),
 	.sysctl_rmem_offset = offsetof(struct net, ipv4.sysctl_udp_rmem_min),
 	.obj_size	   = sizeof(struct udp6_sock),
+	.ipv6_pinfo_offset = offsetof(struct udp6_sock, inet6),
 	.h.udp_table	   = &udplite_table,
 };
 
diff --git a/net/l2tp/l2tp_ip6.c b/net/l2tp/l2tp_ip6.c
index b1623f9c4f92..2eee95a00c05 100644
--- a/net/l2tp/l2tp_ip6.c
+++ b/net/l2tp/l2tp_ip6.c
@@ -36,9 +36,6 @@ struct l2tp_ip6_sock {
 	u32			conn_id;
 	u32			peer_conn_id;
 
-	/* ipv6_pinfo has to be the last member of l2tp_ip6_sock, see
-	 * inet6_sk_generic
-	 */
 	struct ipv6_pinfo	inet6;
 };
 
@@ -730,6 +727,7 @@ static struct proto l2tp_ip6_prot = {
 	.hash		   = l2tp_ip6_hash,
 	.unhash		   = l2tp_ip6_unhash,
 	.obj_size	   = sizeof(struct l2tp_ip6_sock),
+	.ipv6_pinfo_offset = offsetof(struct l2tp_ip6_sock, inet6),
 };
 
 static const struct proto_ops l2tp_ip6_ops = {
diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c
index 3613489eb6e3..22323b1fa184 100644
--- a/net/mptcp/protocol.c
+++ b/net/mptcp/protocol.c
@@ -3988,6 +3988,7 @@ int __init mptcp_proto_v6_init(void)
 	strcpy(mptcp_v6_prot.name, "MPTCPv6");
 	mptcp_v6_prot.slab = NULL;
 	mptcp_v6_prot.obj_size = sizeof(struct mptcp6_sock);
+	mptcp_v6_prot.ipv6_pinfo_offset = offsetof(struct mptcp6_sock, np);
 
 	err = proto_register(&mptcp_v6_prot, 1);
 	if (err)
diff --git a/net/sctp/socket.c b/net/sctp/socket.c
index 9388d98aebc0..6e3d28aa587c 100644
--- a/net/sctp/socket.c
+++ b/net/sctp/socket.c
@@ -9732,6 +9732,7 @@ struct proto sctpv6_prot = {
 	.unhash		= sctp_unhash,
 	.no_autobind	= true,
 	.obj_size	= sizeof(struct sctp6_sock),
+	.ipv6_pinfo_offset = offsetof(struct sctp6_sock, inet6),
 	.useroffset	= offsetof(struct sctp6_sock, sctp.subscribe),
 	.usersize	= offsetof(struct sctp6_sock, sctp.initmsg) -
 				offsetof(struct sctp6_sock, sctp.subscribe) +