Merge: net: Backport data race annotations in the networking stack (part 1).

MR: https://gitlab.com/redhat/centos-stream/src/kernel/centos-stream-9/-/merge_requests/1722

Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=2149949
Upstream Status: linux.git
Conflicts: Few minor conflicts, see description in affected commits.

Properly mark concurent reads and writes with READ_ONCE() and
WRITE_ONCE() in various parts of the networking stack. This is a
backport of the following upstream patch series:
  
  * Patch set A: merge commit e97e68b ("Merge branch 'sk_bound_dev_if-annotations'")
  * Patch set B: merge commit 32b3ad1 ("Merge branch 'sysctl-data-races'")
  * Patch set C: merge commit 7d5424b ("Merge branch 'net-sysctl-races'")
  * Patch set D: merge commit 782d86f ("Merge branch 'net-sysctl-races-round2'")
  * Patch set E: merge commit c9f2110 ("Merge branch 'net-ipv4-sysctl-races-part-3'")

Patch 1 is a standalone READ_ONCE() annotation for sk->sk_bound_dev_if.
It's a prerequisite for correctly backporting patch set A.

Patches 2-9 are backports of patch set A. The following upstream
patches have been omitted since they're already in Centos Stream:
  
  * Upstream commit a20ea29 ("sctp: read sk->sk_bound_dev_if once                                                                                                                                                                                                                                                         
    in sctp_rcv()"), backported by Centos Stream commit 5d539b8.

  * Upstream commit 70f87de ("net_sched: em_meta: add READ_ONCE()                                                                                                                                                                                                                                                         
    in var_sk_bound_if()"), backported by Centos Stream commit
    866ca28.

Patch 10 was in the original upstream series of patch set B, but was
resubmitted independently as that series was reworked before being
applied. Therefore, it doesn't strictly belong to patch set B, but is
closely related to it and is thus backported here.

Patches 11-21 are backports of patch set B. The following upstream
patch has been omitted since it's already in Centos Stream:
  
  * Upstream commit 310731e ("net: Fix data-races around                                                                                                                                                                                                                                                                  
    sysctl_mem.", backported by Centos Stream commit a99b2cb.

Patches 22-36 are backports corresponding to patch set C.

Patches 37-51 are backports corresponding to patch set D.

Patches 52-66 are backports corresponding to patch set E.

Signed-off-by: Guillaume Nault <gnault@redhat.com>

Approved-by: Antoine Tenart <atenart@redhat.com>
Approved-by: Michal Schmidt <mschmidt@redhat.com>

Signed-off-by: Herton R. Krzesinski <herton@redhat.com>

Author: Herton R. Krzesinski

Documentation/networking/ip-sysctl.rst

@@ -1066,7 +1066,7 @@ cipso_cache_enable - BOOLEAN
 cipso_cache_bucket_size - INTEGER
 	The CIPSO label cache consists of a fixed size hash table with each
 	hash bucket containing a number of cache entries.  This variable limits
-	the number of entries in each hash bucket; the larger the value the
+	the number of entries in each hash bucket; the larger the value is, the
 	more CIPSO label mappings that can be cached.  When the number of
 	entries in a given hash bucket reaches this limit adding new entries
 	causes the oldest entry in the bucket to be removed to make room.
@@ -1160,7 +1160,7 @@ ip_autobind_reuse - BOOLEAN
 	option should only be set by experts.
 	Default: 0
 
-ip_dynaddr - BOOLEAN
+ip_dynaddr - INTEGER
 	If set non-zero, enables support for dynamic addresses.
 	If set to a non-zero value larger than 1, a kernel log
 	message will be printed when dynamic address rewriting

drivers/net/ethernet/chelsio/inline_crypto/chtls/chtls_cm.c

@@ -1391,7 +1391,7 @@ static void chtls_pass_accept_request(struct sock *sk,
 	th_ecn = tcph->ece && tcph->cwr;
 	if (th_ecn) {
 		ect = !INET_ECN_is_not_ect(ip_dsfield);
-		ecn_ok = sock_net(sk)->ipv4.sysctl_tcp_ecn;
+		ecn_ok = READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_ecn);
 		if ((!ect && ecn_ok) || tcp_ca_needs_ecn(sk))
 			inet_rsk(oreq)->ecn_ok = 1;
 	}

drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c

@@ -10224,13 +10224,14 @@ static int mlxsw_sp_dscp_init(struct mlxsw_sp *mlxsw_sp)
 static int __mlxsw_sp_router_init(struct mlxsw_sp *mlxsw_sp)
 {
 	struct net *net = mlxsw_sp_net(mlxsw_sp);
-	bool usp = net->ipv4.sysctl_ip_fwd_update_priority;
 	char rgcr_pl[MLXSW_REG_RGCR_LEN];
 	u64 max_rifs;
+	bool usp;
 
 	if (!MLXSW_CORE_RES_VALID(mlxsw_sp->core, MAX_RIFS))
 		return -EIO;
 	max_rifs = MLXSW_CORE_RES_GET(mlxsw_sp->core, MAX_RIFS);
+	usp = READ_ONCE(net->ipv4.sysctl_ip_fwd_update_priority);
 
 	mlxsw_reg_rgcr_pack(rgcr_pl, true, true);
 	mlxsw_reg_rgcr_max_router_interfaces_set(rgcr_pl, max_rifs);

include/net/inet6_hashtables.h

@@ -103,15 +103,25 @@ struct sock *inet6_lookup(struct net *net, struct inet_hashinfo *hashinfo,
 			  const int dif);
 
 int inet6_hash(struct sock *sk);
-#endif /* IS_ENABLED(CONFIG_IPV6) */
 
-#define INET6_MATCH(__sk, __net, __saddr, __daddr, __ports, __dif, __sdif) \
-	(((__sk)->sk_portpair == (__ports))			&&	\
-	 ((__sk)->sk_family == AF_INET6)			&&	\
-	 ipv6_addr_equal(&(__sk)->sk_v6_daddr, (__saddr))		&&	\
-	 ipv6_addr_equal(&(__sk)->sk_v6_rcv_saddr, (__daddr))	&&	\
-	 (((__sk)->sk_bound_dev_if == (__dif))	||			\
-	  ((__sk)->sk_bound_dev_if == (__sdif)))		&&	\
-	 net_eq(sock_net(__sk), (__net)))
+static inline bool inet6_match(struct net *net, const struct sock *sk,
+			       const struct in6_addr *saddr,
+			       const struct in6_addr *daddr,
+			       const __portpair ports,
+			       const int dif, const int sdif)
+{
+	int bound_dev_if;
+
+	if (!net_eq(sock_net(sk), net) ||
+	    sk->sk_family != AF_INET6 ||
+	    sk->sk_portpair != ports ||
+	    !ipv6_addr_equal(&sk->sk_v6_daddr, saddr) ||
+	    !ipv6_addr_equal(&sk->sk_v6_rcv_saddr, daddr))
+		return false;
+
+	bound_dev_if = READ_ONCE(sk->sk_bound_dev_if);
+	return bound_dev_if == dif || bound_dev_if == sdif;
+}
+#endif /* IS_ENABLED(CONFIG_IPV6) */
 
 #endif /* _INET6_HASHTABLES_H */

include/net/inet_hashtables.h

@@ -179,7 +179,7 @@ static inline bool inet_sk_bound_dev_eq(struct net *net, int bound_dev_if,
 					int dif, int sdif)
 {
 #if IS_ENABLED(CONFIG_NET_L3_MASTER_DEV)
-	return inet_bound_dev_eq(!!net->ipv4.sysctl_tcp_l3mdev_accept,
+	return inet_bound_dev_eq(!!READ_ONCE(net->ipv4.sysctl_tcp_l3mdev_accept),
 				 bound_dev_if, dif, sdif);
 #else
 	return inet_bound_dev_eq(true, bound_dev_if, dif, sdif);
@@ -255,7 +255,6 @@ static inline struct sock *inet_lookup_listener(struct net *net,
 	((__force __portpair)(((__u32)(__dport) << 16) | (__force __u32)(__be16)(__sport)))
 #endif
 
-#if (BITS_PER_LONG == 64)
 #ifdef __BIG_ENDIAN
 #define INET_ADDR_COOKIE(__name, __saddr, __daddr) \
 	const __addrpair __name = (__force __addrpair) ( \
@@ -267,24 +266,22 @@ static inline struct sock *inet_lookup_listener(struct net *net,
 				   (((__force __u64)(__be32)(__daddr)) << 32) | \
 				   ((__force __u64)(__be32)(__saddr)))
 #endif /* __BIG_ENDIAN */
-#define INET_MATCH(__sk, __net, __cookie, __saddr, __daddr, __ports, __dif, __sdif) \
-	(((__sk)->sk_portpair == (__ports))			&&	\
-	 ((__sk)->sk_addrpair == (__cookie))			&&	\
-	 (((__sk)->sk_bound_dev_if == (__dif))			||	\
-	  ((__sk)->sk_bound_dev_if == (__sdif)))		&&	\
-	 net_eq(sock_net(__sk), (__net)))
-#else /* 32-bit arch */
-#define INET_ADDR_COOKIE(__name, __saddr, __daddr) \
-	const int __name __deprecated __attribute__((unused))
-
-#define INET_MATCH(__sk, __net, __cookie, __saddr, __daddr, __ports, __dif, __sdif) \
-	(((__sk)->sk_portpair == (__ports))		&&		\
-	 ((__sk)->sk_daddr	== (__saddr))		&&		\
-	 ((__sk)->sk_rcv_saddr	== (__daddr))		&&		\
-	 (((__sk)->sk_bound_dev_if == (__dif))		||		\
-	  ((__sk)->sk_bound_dev_if == (__sdif)))	&&		\
-	 net_eq(sock_net(__sk), (__net)))
-#endif /* 64-bit arch */
+
+static inline bool inet_match(struct net *net, const struct sock *sk,
+			      const __addrpair cookie, const __portpair ports,
+			      int dif, int sdif)
+{
+	int bound_dev_if;
+
+	if (!net_eq(sock_net(sk), net) ||
+	    sk->sk_portpair != ports ||
+	    sk->sk_addrpair != cookie)
+	        return false;
+
+	/* Paired with WRITE_ONCE() from sock_bindtoindex_locked() */
+	bound_dev_if = READ_ONCE(sk->sk_bound_dev_if);
+	return bound_dev_if == dif || bound_dev_if == sdif;
+}
 
 /* Sockets in TCP_CLOSE state are _always_ taken out of the hash, so we need
  * not check it for lookups anymore, thanks Alexey. -DaveM

include/net/inet_sock.h

@@ -107,7 +107,8 @@ static inline struct inet_request_sock *inet_rsk(const struct request_sock *sk)
 
 static inline u32 inet_request_mark(const struct sock *sk, struct sk_buff *skb)
 {
-	if (!sk->sk_mark && sock_net(sk)->ipv4.sysctl_tcp_fwmark_accept)
+	if (!sk->sk_mark &&
+	    READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_fwmark_accept))
 		return skb->mark;
 
 	return sk->sk_mark;
@@ -116,22 +117,23 @@ static inline u32 inet_request_mark(const struct sock *sk, struct sk_buff *skb)
 static inline int inet_request_bound_dev_if(const struct sock *sk,
 					    struct sk_buff *skb)
 {
+	int bound_dev_if = READ_ONCE(sk->sk_bound_dev_if);
 #ifdef CONFIG_NET_L3_MASTER_DEV
 	struct net *net = sock_net(sk);
 
-	if (!sk->sk_bound_dev_if && net->ipv4.sysctl_tcp_l3mdev_accept)
+	if (!bound_dev_if && READ_ONCE(net->ipv4.sysctl_tcp_l3mdev_accept))
 		return l3mdev_master_ifindex_by_index(net, skb->skb_iif);
 #endif
 
-	return sk->sk_bound_dev_if;
+	return bound_dev_if;
 }
 
 static inline int inet_sk_bound_l3mdev(const struct sock *sk)
 {
 #ifdef CONFIG_NET_L3_MASTER_DEV
 	struct net *net = sock_net(sk);
 
-	if (!net->ipv4.sysctl_tcp_l3mdev_accept)
+	if (!READ_ONCE(net->ipv4.sysctl_tcp_l3mdev_accept))
 		return l3mdev_master_ifindex_by_index(net,
 						      sk->sk_bound_dev_if);
 #endif
@@ -373,7 +375,7 @@ static inline bool inet_get_convert_csum(struct sock *sk)
 static inline bool inet_can_nonlocal_bind(struct net *net,
 					  struct inet_sock *inet)
 {
-	return net->ipv4.sysctl_ip_nonlocal_bind ||
+	return READ_ONCE(net->ipv4.sysctl_ip_nonlocal_bind) ||
 		inet->freebind || inet->transparent;
 }
 

include/net/ip.h

@@ -92,7 +92,7 @@ static inline void ipcm_init_sk(struct ipcm_cookie *ipcm,
 
 	ipcm->sockc.mark = inet->sk.sk_mark;
 	ipcm->sockc.tsflags = inet->sk.sk_tsflags;
-	ipcm->oif = inet->sk.sk_bound_dev_if;
+	ipcm->oif = READ_ONCE(inet->sk.sk_bound_dev_if);
 	ipcm->addr = inet->inet_saddr;
 }
 
@@ -382,7 +382,7 @@ void ipfrag_init(void);
 void ip_static_sysctl_init(void);
 
 #define IP4_REPLY_MARK(net, mark) \
-	((net)->ipv4.sysctl_fwmark_reflect ? (mark) : 0)
+	(READ_ONCE((net)->ipv4.sysctl_fwmark_reflect) ? (mark) : 0)
 
 static inline bool ip_is_fragment(const struct iphdr *iph)
 {
@@ -443,7 +443,7 @@ static inline unsigned int ip_dst_mtu_maybe_forward(const struct dst_entry *dst,
 	struct net *net = dev_net(dst->dev);
 	unsigned int mtu;
 
-	if (net->ipv4.sysctl_ip_fwd_use_pmtu ||
+	if (READ_ONCE(net->ipv4.sysctl_ip_fwd_use_pmtu) ||
 	    ip_mtu_locked(dst) ||
 	    !forwarding)
 		return dst_mtu(dst);

include/net/raw.h

@@ -83,7 +83,7 @@ static inline bool raw_sk_bound_dev_eq(struct net *net, int bound_dev_if,
 				       int dif, int sdif)
 {
 #if IS_ENABLED(CONFIG_NET_L3_MASTER_DEV)
-	return inet_bound_dev_eq(!!net->ipv4.sysctl_raw_l3mdev_accept,
+	return inet_bound_dev_eq(READ_ONCE(net->ipv4.sysctl_raw_l3mdev_accept),
 				 bound_dev_if, dif, sdif);
 #else
 	return inet_bound_dev_eq(true, bound_dev_if, dif, sdif);

include/net/route.h

@@ -361,7 +361,7 @@ static inline int ip4_dst_hoplimit(const struct dst_entry *dst)
 	struct net *net = dev_net(dst->dev);
 
 	if (hoplimit == 0)
-		hoplimit = net->ipv4.sysctl_ip_default_ttl;
+		hoplimit = READ_ONCE(net->ipv4.sysctl_ip_default_ttl);
 	return hoplimit;
 }
 

include/net/sock.h

@@ -160,9 +160,6 @@ typedef __u64 __bitwise __addrpair;
  *	for struct sock and struct inet_timewait_sock.
  */
 struct sock_common {
-	/* skc_daddr and skc_rcv_saddr must be grouped on a 8 bytes aligned
-	 * address on 64bit arches : cf INET_MATCH()
-	 */
 	union {
 		__addrpair	skc_addrpair;
 		struct {
@@ -2851,13 +2848,14 @@ static inline void sk_pacing_shift_update(struct sock *sk, int val)
  */
 static inline bool sk_dev_equal_l3scope(struct sock *sk, int dif)
 {
+	int bound_dev_if = READ_ONCE(sk->sk_bound_dev_if);
 	int mdif;
 
-	if (!sk->sk_bound_dev_if || sk->sk_bound_dev_if == dif)
+	if (!bound_dev_if || bound_dev_if == dif)
 		return true;
 
 	mdif = l3mdev_master_ifindex_by_index(sock_net(sk), dif);
-	if (mdif && mdif == sk->sk_bound_dev_if)
+	if (mdif && mdif == bound_dev_if)
 		return true;
 
 	return false;