1 files changed, 309 insertions, 357 deletions

diff --git a/net/ipv4/ip_sockglue.c b/net/ipv4/ip_sockglue.c
index 8e97d8d4cc9d..6d9c5c20b1c4 100644
--- a/net/ipv4/ip_sockglue.c
+++ b/net/ipv4/ip_sockglue.c
@@ -47,8 +47,6 @@
 #include <linux/errqueue.h>
 #include <linux/uaccess.h>
 
-#include <linux/bpfilter.h>
-
 /*
  *	SOL_IP control messages.
  */
@@ -130,20 +128,20 @@ static void ip_cmsg_recv_checksum(struct msghdr *msg, struct sk_buff *skb,
 
 static void ip_cmsg_recv_security(struct msghdr *msg, struct sk_buff *skb)
 {
-	char *secdata;
-	u32 seclen, secid;
+	struct lsm_context ctx;
+	u32 secid;
 	int err;
 
 	err = security_socket_getpeersec_dgram(NULL, skb, &secid);
 	if (err)
 		return;
 
-	err = security_secid_to_secctx(secid, &secdata, &seclen);
-	if (err)
+	err = security_secid_to_secctx(secid, &ctx);
+	if (err < 0)
 		return;
 
-	put_cmsg(msg, SOL_IP, SCM_SECURITY, seclen, secdata);
-	security_release_secctx(secdata, seclen);
+	put_cmsg(msg, SOL_IP, SCM_SECURITY, ctx.len, ctx.context);
+	security_release_secctx(&ctx);
 }
 
 static void ip_cmsg_recv_dstaddr(struct msghdr *msg, struct sk_buff *skb)
@@ -171,8 +169,10 @@ static void ip_cmsg_recv_dstaddr(struct msghdr *msg, struct sk_buff *skb)
 void ip_cmsg_recv_offset(struct msghdr *msg, struct sock *sk,
 			 struct sk_buff *skb, int tlen, int offset)
 {
-	struct inet_sock *inet = inet_sk(sk);
-	unsigned int flags = inet->cmsg_flags;
+	unsigned long flags = inet_cmsg_flags(inet_sk(sk));
+
+	if (!flags)
+		return;
 
 	/* Ordered by supposed usage frequency */
 	if (flags & IP_CMSG_PKTINFO) {
@@ -315,7 +315,7 @@ int ip_cmsg_send(struct sock *sk, struct msghdr *msg, struct ipcm_cookie *ipc,
 			if (val < 0 || val > 255)
 				return -EINVAL;
 			ipc->tos = val;
-			ipc->priority = rt_tos2priority(ipc->tos);
+			ipc->sockc.priority = rt_tos2priority(ipc->tos);
 			break;
 		case IP_PROTOCOL:
 			if (cmsg->cmsg_len != CMSG_LEN(sizeof(int)))
@@ -431,7 +431,7 @@ void ip_icmp_error(struct sock *sk, struct sk_buff *skb, int err,
 	serr->port = port;
 
 	if (skb_pull(skb, payload - skb->data)) {
-		if (inet_sk(sk)->recverr_rfc4884)
+		if (inet_test_bit(RECVERR_RFC4884, sk))
 			ipv4_icmp_error_rfc4884(skb, &serr->ee.ee_rfc4884);
 
 		skb_reset_transport_header(skb);
@@ -444,12 +444,11 @@ EXPORT_SYMBOL_GPL(ip_icmp_error);
 
 void ip_local_error(struct sock *sk, int err, __be32 daddr, __be16 port, u32 info)
 {
-	struct inet_sock *inet = inet_sk(sk);
 	struct sock_exterr_skb *serr;
 	struct iphdr *iph;
 	struct sk_buff *skb;
 
-	if (!inet->recverr)
+	if (!inet_test_bit(RECVERR, sk))
 		return;
 
 	skb = alloc_skb(sizeof(struct iphdr), GFP_ATOMIC);
@@ -510,7 +509,7 @@ static bool ipv4_datagram_support_cmsg(const struct sock *sk,
 	 * or without payload (SOF_TIMESTAMPING_OPT_TSONLY).
 	 */
 	info = PKTINFO_SKB_CB(skb);
-	if (!(sk->sk_tsflags & SOF_TIMESTAMPING_OPT_CMSG) ||
+	if (!(READ_ONCE(sk->sk_tsflags) & SOF_TIMESTAMPING_OPT_CMSG) ||
 	    !info->ipi_ifindex)
 		return false;
 
@@ -568,7 +567,7 @@ int ip_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len)
 	if (ipv4_datagram_support_cmsg(sk, skb, serr->ee.ee_origin)) {
 		sin->sin_family = AF_INET;
 		sin->sin_addr.s_addr = ip_hdr(skb)->saddr;
-		if (inet_sk(sk)->cmsg_flags)
+		if (inet_cmsg_flags(inet_sk(sk)))
 			ip_cmsg_recv(msg, skb);
 	}
 
@@ -586,38 +585,36 @@ out:
 
 void __ip_sock_set_tos(struct sock *sk, int val)
 {
+	u8 old_tos = inet_sk(sk)->tos;
+
 	if (sk->sk_type == SOCK_STREAM) {
 		val &= ~INET_ECN_MASK;
-		val |= inet_sk(sk)->tos & INET_ECN_MASK;
+		val |= old_tos & INET_ECN_MASK;
 	}
-	if (inet_sk(sk)->tos != val) {
-		inet_sk(sk)->tos = val;
-		sk->sk_priority = rt_tos2priority(val);
+	if (old_tos != val) {
+		WRITE_ONCE(inet_sk(sk)->tos, val);
+		WRITE_ONCE(sk->sk_priority, rt_tos2priority(val));
 		sk_dst_reset(sk);
 	}
 }
 
 void ip_sock_set_tos(struct sock *sk, int val)
 {
-	lock_sock(sk);
+	sockopt_lock_sock(sk);
 	__ip_sock_set_tos(sk, val);
-	release_sock(sk);
+	sockopt_release_sock(sk);
 }
 EXPORT_SYMBOL(ip_sock_set_tos);
 
 void ip_sock_set_freebind(struct sock *sk)
 {
-	lock_sock(sk);
-	inet_sk(sk)->freebind = true;
-	release_sock(sk);
+	inet_set_bit(FREEBIND, sk);
 }
 EXPORT_SYMBOL(ip_sock_set_freebind);
 
 void ip_sock_set_recverr(struct sock *sk)
 {
-	lock_sock(sk);
-	inet_sk(sk)->recverr = true;
-	release_sock(sk);
+	inet_set_bit(RECVERR, sk);
 }
 EXPORT_SYMBOL(ip_sock_set_recverr);
 
@@ -625,18 +622,14 @@ int ip_sock_set_mtu_discover(struct sock *sk, int val)
 {
 	if (val < IP_PMTUDISC_DONT || val > IP_PMTUDISC_OMIT)
 		return -EINVAL;
-	lock_sock(sk);
-	inet_sk(sk)->pmtudisc = val;
-	release_sock(sk);
+	WRITE_ONCE(inet_sk(sk)->pmtudisc, val);
 	return 0;
 }
 EXPORT_SYMBOL(ip_sock_set_mtu_discover);
 
 void ip_sock_set_pktinfo(struct sock *sk)
 {
-	lock_sock(sk);
-	inet_sk(sk)->cmsg_flags |= IP_CMSG_PKTINFO;
-	release_sock(sk);
+	inet_set_bit(PKTINFO, sk);
 }
 EXPORT_SYMBOL(ip_sock_set_pktinfo);
 
@@ -780,7 +773,7 @@ static int ip_set_mcast_msfilter(struct sock *sk, sockptr_t optval, int optlen)
 
 	if (optlen < GROUP_FILTER_SIZE(0))
 		return -EINVAL;
-	if (optlen > READ_ONCE(sysctl_optmem_max))
+	if (optlen > READ_ONCE(sock_net(sk)->core.sysctl_optmem_max))
 		return -ENOBUFS;
 
 	gsf = memdup_sockptr(optval, optlen);
@@ -816,7 +809,7 @@ static int compat_ip_set_mcast_msfilter(struct sock *sk, sockptr_t optval,
 
 	if (optlen < size0)
 		return -EINVAL;
-	if (optlen > READ_ONCE(sysctl_optmem_max) - 4)
+	if (optlen > READ_ONCE(sock_net(sk)->core.sysctl_optmem_max) - 4)
 		return -ENOBUFS;
 
 	p = kmalloc(optlen + 4, GFP_KERNEL);
@@ -901,7 +894,7 @@ int do_ip_setsockopt(struct sock *sk, int level, int optname,
 {
 	struct inet_sock *inet = inet_sk(sk);
 	struct net *net = sock_net(sk);
-	int val = 0, err;
+	int val = 0, err, retv;
 	bool needs_rtnl = setsockopt_needs_rtnl(optname);
 
 	switch (optname) {
@@ -945,11 +938,140 @@ int do_ip_setsockopt(struct sock *sk, int level, int optname,
 
 	/* If optlen==0, it is equivalent to val == 0 */
 
-	if (optname == IP_ROUTER_ALERT)
-		return ip_ra_control(sk, val ? 1 : 0, NULL);
+	if (optname == IP_ROUTER_ALERT) {
+		retv = ip_ra_control(sk, val ? 1 : 0, NULL);
+		if (retv == 0)
+			inet_assign_bit(RTALERT, sk, val);
+		return retv;
+	}
 	if (ip_mroute_opt(optname))
 		return ip_mroute_setsockopt(sk, optname, optval, optlen);
 
+	/* Handle options that can be set without locking the socket. */
+	switch (optname) {
+	case IP_PKTINFO:
+		inet_assign_bit(PKTINFO, sk, val);
+		return 0;
+	case IP_RECVTTL:
+		inet_assign_bit(TTL, sk, val);
+		return 0;
+	case IP_RECVTOS:
+		inet_assign_bit(TOS, sk, val);
+		return 0;
+	case IP_RECVOPTS:
+		inet_assign_bit(RECVOPTS, sk, val);
+		return 0;
+	case IP_RETOPTS:
+		inet_assign_bit(RETOPTS, sk, val);
+		return 0;
+	case IP_PASSSEC:
+		inet_assign_bit(PASSSEC, sk, val);
+		return 0;
+	case IP_RECVORIGDSTADDR:
+		inet_assign_bit(ORIGDSTADDR, sk, val);
+		return 0;
+	case IP_RECVFRAGSIZE:
+		if (sk->sk_type != SOCK_RAW && sk->sk_type != SOCK_DGRAM)
+			return -EINVAL;
+		inet_assign_bit(RECVFRAGSIZE, sk, val);
+		return 0;
+	case IP_RECVERR:
+		inet_assign_bit(RECVERR, sk, val);
+		if (!val)
+			skb_errqueue_purge(&sk->sk_error_queue);
+		return 0;
+	case IP_RECVERR_RFC4884:
+		if (val < 0 || val > 1)
+			return -EINVAL;
+		inet_assign_bit(RECVERR_RFC4884, sk, val);
+		return 0;
+	case IP_FREEBIND:
+		if (optlen < 1)
+			return -EINVAL;
+		inet_assign_bit(FREEBIND, sk, val);
+		return 0;
+	case IP_HDRINCL:
+		if (sk->sk_type != SOCK_RAW)
+			return -ENOPROTOOPT;
+		inet_assign_bit(HDRINCL, sk, val);
+		return 0;
+	case IP_MULTICAST_LOOP:
+		if (optlen < 1)
+			return -EINVAL;
+		inet_assign_bit(MC_LOOP, sk, val);
+		return 0;
+	case IP_MULTICAST_ALL:
+		if (optlen < 1)
+			return -EINVAL;
+		if (val != 0 && val != 1)
+			return -EINVAL;
+		inet_assign_bit(MC_ALL, sk, val);
+		return 0;
+	case IP_TRANSPARENT:
+		if (!!val && !sockopt_ns_capable(sock_net(sk)->user_ns, CAP_NET_RAW) &&
+		    !sockopt_ns_capable(sock_net(sk)->user_ns, CAP_NET_ADMIN))
+			return -EPERM;
+		if (optlen < 1)
+			return -EINVAL;
+		inet_assign_bit(TRANSPARENT, sk, val);
+		return 0;
+	case IP_NODEFRAG:
+		if (sk->sk_type != SOCK_RAW)
+			return -ENOPROTOOPT;
+		inet_assign_bit(NODEFRAG, sk, val);
+		return 0;
+	case IP_BIND_ADDRESS_NO_PORT:
+		inet_assign_bit(BIND_ADDRESS_NO_PORT, sk, val);
+		return 0;
+	case IP_TTL:
+		if (optlen < 1)
+			return -EINVAL;
+		if (val != -1 && (val < 1 || val > 255))
+			return -EINVAL;
+		WRITE_ONCE(inet->uc_ttl, val);
+		return 0;
+	case IP_MINTTL:
+		if (optlen < 1)
+			return -EINVAL;
+		if (val < 0 || val > 255)
+			return -EINVAL;
+
+		if (val)
+			static_branch_enable(&ip4_min_ttl);
+
+		WRITE_ONCE(inet->min_ttl, val);
+		return 0;
+	case IP_MULTICAST_TTL:
+		if (sk->sk_type == SOCK_STREAM)
+			return -EINVAL;
+		if (optlen < 1)
+			return -EINVAL;
+		if (val == -1)
+			val = 1;
+		if (val < 0 || val > 255)
+			return -EINVAL;
+		WRITE_ONCE(inet->mc_ttl, val);
+		return 0;
+	case IP_MTU_DISCOVER:
+		return ip_sock_set_mtu_discover(sk, val);
+	case IP_TOS:	/* This sets both TOS and Precedence */
+		ip_sock_set_tos(sk, val);
+		return 0;
+	case IP_LOCAL_PORT_RANGE:
+	{
+		u16 lo = val;
+		u16 hi = val >> 16;
+
+		if (optlen != sizeof(u32))
+			return -EINVAL;
+		if (lo != 0 && hi != 0 && lo > hi)
+			return -EINVAL;
+
+		WRITE_ONCE(inet->local_port_range, val);
+		return 0;
+	}
+	}
+
 	err = 0;
 	if (needs_rtnl)
 		rtnl_lock();
@@ -967,7 +1089,7 @@ int do_ip_setsockopt(struct sock *sk, int level, int optname,
 			break;
 		old = rcu_dereference_protected(inet->inet_opt,
 						lockdep_sock_is_held(sk));
-		if (inet->is_icsk) {
+		if (inet_test_bit(IS_ICSK, sk)) {
 			struct inet_connection_sock *icsk = inet_csk(sk);
 #if IS_ENABLED(CONFIG_IPV6)
 			if (sk->sk_family == PF_INET ||
@@ -989,127 +1111,19 @@ int do_ip_setsockopt(struct sock *sk, int level, int optname,
 			kfree_rcu(old, rcu);
 		break;
 	}
-	case IP_PKTINFO:
-		if (val)
-			inet->cmsg_flags |= IP_CMSG_PKTINFO;
-		else
-			inet->cmsg_flags &= ~IP_CMSG_PKTINFO;
-		break;
-	case IP_RECVTTL:
-		if (val)
-			inet->cmsg_flags |=  IP_CMSG_TTL;
-		else
-			inet->cmsg_flags &= ~IP_CMSG_TTL;
-		break;
-	case IP_RECVTOS:
-		if (val)
-			inet->cmsg_flags |=  IP_CMSG_TOS;
-		else
-			inet->cmsg_flags &= ~IP_CMSG_TOS;
-		break;
-	case IP_RECVOPTS:
-		if (val)
-			inet->cmsg_flags |=  IP_CMSG_RECVOPTS;
-		else
-			inet->cmsg_flags &= ~IP_CMSG_RECVOPTS;
-		break;
-	case IP_RETOPTS:
-		if (val)
-			inet->cmsg_flags |= IP_CMSG_RETOPTS;
-		else
-			inet->cmsg_flags &= ~IP_CMSG_RETOPTS;
-		break;
-	case IP_PASSSEC:
-		if (val)
-			inet->cmsg_flags |= IP_CMSG_PASSSEC;
-		else
-			inet->cmsg_flags &= ~IP_CMSG_PASSSEC;
-		break;
-	case IP_RECVORIGDSTADDR:
-		if (val)
-			inet->cmsg_flags |= IP_CMSG_ORIGDSTADDR;
-		else
-			inet->cmsg_flags &= ~IP_CMSG_ORIGDSTADDR;
-		break;
 	case IP_CHECKSUM:
 		if (val) {
-			if (!(inet->cmsg_flags & IP_CMSG_CHECKSUM)) {
+			if (!(inet_test_bit(CHECKSUM, sk))) {
 				inet_inc_convert_csum(sk);
-				inet->cmsg_flags |= IP_CMSG_CHECKSUM;
+				inet_set_bit(CHECKSUM, sk);
 			}
 		} else {
-			if (inet->cmsg_flags & IP_CMSG_CHECKSUM) {
+			if (inet_test_bit(CHECKSUM, sk)) {
 				inet_dec_convert_csum(sk);
-				inet->cmsg_flags &= ~IP_CMSG_CHECKSUM;
+				inet_clear_bit(CHECKSUM, sk);
 			}
 		}
 		break;
-	case IP_RECVFRAGSIZE:
-		if (sk->sk_type != SOCK_RAW && sk->sk_type != SOCK_DGRAM)
-			goto e_inval;
-		if (val)
-			inet->cmsg_flags |= IP_CMSG_RECVFRAGSIZE;
-		else
-			inet->cmsg_flags &= ~IP_CMSG_RECVFRAGSIZE;
-		break;
-	case IP_TOS:	/* This sets both TOS and Precedence */
-		__ip_sock_set_tos(sk, val);
-		break;
-	case IP_TTL:
-		if (optlen < 1)
-			goto e_inval;
-		if (val != -1 && (val < 1 || val > 255))
-			goto e_inval;
-		inet->uc_ttl = val;
-		break;
-	case IP_HDRINCL:
-		if (sk->sk_type != SOCK_RAW) {
-			err = -ENOPROTOOPT;
-			break;
-		}
-		inet->hdrincl = val ? 1 : 0;
-		break;
-	case IP_NODEFRAG:
-		if (sk->sk_type != SOCK_RAW) {
-			err = -ENOPROTOOPT;
-			break;
-		}
-		inet->nodefrag = val ? 1 : 0;
-		break;
-	case IP_BIND_ADDRESS_NO_PORT:
-		inet->bind_address_no_port = val ? 1 : 0;
-		break;
-	case IP_MTU_DISCOVER:
-		if (val < IP_PMTUDISC_DONT || val > IP_PMTUDISC_OMIT)
-			goto e_inval;
-		inet->pmtudisc = val;
-		break;
-	case IP_RECVERR:
-		inet->recverr = !!val;
-		if (!val)
-			skb_queue_purge(&sk->sk_error_queue);
-		break;
-	case IP_RECVERR_RFC4884:
-		if (val < 0 || val > 1)
-			goto e_inval;
-		inet->recverr_rfc4884 = !!val;
-		break;
-	case IP_MULTICAST_TTL:
-		if (sk->sk_type == SOCK_STREAM)
-			goto e_inval;
-		if (optlen < 1)
-			goto e_inval;
-		if (val == -1)
-			val = 1;
-		if (val < 0 || val > 255)
-			goto e_inval;
-		inet->mc_ttl = val;
-		break;
-	case IP_MULTICAST_LOOP:
-		if (optlen < 1)
-			goto e_inval;
-		inet->mc_loop = !!val;
-		break;
 	case IP_UNICAST_IF:
 	{
 		struct net_device *dev = NULL;
@@ -1121,7 +1135,7 @@ int do_ip_setsockopt(struct sock *sk, int level, int optname,
 
 		ifindex = (__force int)ntohl((__force __be32)val);
 		if (ifindex == 0) {
-			inet->uc_index = 0;
+			WRITE_ONCE(inet->uc_index, 0);
 			err = 0;
 			break;
 		}
@@ -1138,7 +1152,7 @@ int do_ip_setsockopt(struct sock *sk, int level, int optname,
 		if (sk->sk_bound_dev_if && midx != sk->sk_bound_dev_if)
 			break;
 
-		inet->uc_index = ifindex;
+		WRITE_ONCE(inet->uc_index, ifindex);
 		err = 0;
 		break;
 	}
@@ -1176,8 +1190,8 @@ int do_ip_setsockopt(struct sock *sk, int level, int optname,
 
 		if (!mreq.imr_ifindex) {
 			if (mreq.imr_address.s_addr == htonl(INADDR_ANY)) {
-				inet->mc_index = 0;
-				inet->mc_addr  = 0;
+				WRITE_ONCE(inet->mc_index, 0);
+				WRITE_ONCE(inet->mc_addr, 0);
 				err = 0;
 				break;
 			}
@@ -1202,8 +1216,8 @@ int do_ip_setsockopt(struct sock *sk, int level, int optname,
 		    midx != sk->sk_bound_dev_if)
 			break;
 
-		inet->mc_index = mreq.imr_ifindex;
-		inet->mc_addr  = mreq.imr_address.s_addr;
+		WRITE_ONCE(inet->mc_index, mreq.imr_ifindex);
+		WRITE_ONCE(inet->mc_addr, mreq.imr_address.s_addr);
 		err = 0;
 		break;
 	}
@@ -1214,7 +1228,7 @@ int do_ip_setsockopt(struct sock *sk, int level, int optname,
 		struct ip_mreqn mreq;
 
 		err = -EPROTO;
-		if (inet_sk(sk)->is_icsk)
+		if (inet_test_bit(IS_ICSK, sk))
 			break;
 
 		if (optlen < sizeof(struct ip_mreq))
@@ -1242,7 +1256,7 @@ int do_ip_setsockopt(struct sock *sk, int level, int optname,
 
 		if (optlen < IP_MSFILTER_SIZE(0))
 			goto e_inval;
-		if (optlen > READ_ONCE(sysctl_optmem_max)) {
+		if (optlen > READ_ONCE(net->core.sysctl_optmem_max)) {
 			err = -ENOBUFS;
 			break;
 		}
@@ -1325,20 +1339,6 @@ int do_ip_setsockopt(struct sock *sk, int level, int optname,
 		else
 			err = ip_set_mcast_msfilter(sk, optval, optlen);
 		break;
-	case IP_MULTICAST_ALL:
-		if (optlen < 1)
-			goto e_inval;
-		if (val != 0 && val != 1)
-			goto e_inval;
-		inet->mc_all = val;
-		break;
-
-	case IP_FREEBIND:
-		if (optlen < 1)
-			goto e_inval;
-		inet->freebind = !!val;
-		break;
-
 	case IP_IPSEC_POLICY:
 	case IP_XFRM_POLICY:
 		err = -EPERM;
@@ -1347,46 +1347,6 @@ int do_ip_setsockopt(struct sock *sk, int level, int optname,
 		err = xfrm_user_policy(sk, optname, optval, optlen);
 		break;
 
-	case IP_TRANSPARENT:
-		if (!!val && !sockopt_ns_capable(sock_net(sk)->user_ns, CAP_NET_RAW) &&
-		    !sockopt_ns_capable(sock_net(sk)->user_ns, CAP_NET_ADMIN)) {
-			err = -EPERM;
-			break;
-		}
-		if (optlen < 1)
-			goto e_inval;
-		inet->transparent = !!val;
-		break;
-
-	case IP_MINTTL:
-		if (optlen < 1)
-			goto e_inval;
-		if (val < 0 || val > 255)
-			goto e_inval;
-
-		if (val)
-			static_branch_enable(&ip4_min_ttl);
-
-		/* tcp_v4_err() and tcp_v4_rcv() might read min_ttl
-		 * while we are changint it.
-		 */
-		WRITE_ONCE(inet->min_ttl, val);
-		break;
-
-	case IP_LOCAL_PORT_RANGE:
-	{
-		const __u16 lo = val;
-		const __u16 hi = val >> 16;
-
-		if (optlen != sizeof(__u32))
-			goto e_inval;
-		if (lo != 0 && hi != 0 && lo > hi)
-			goto e_inval;
-
-		inet->local_port_range.lo = lo;
-		inet->local_port_range.hi = hi;
-		break;
-	}
 	default:
 		err = -ENOPROTOOPT;
 		break;
@@ -1407,15 +1367,16 @@ e_inval:
  * ipv4_pktinfo_prepare - transfer some info from rtable to skb
  * @sk: socket
  * @skb: buffer
+ * @drop_dst: if true, drops skb dst
  *
  * To support IP_CMSG_PKTINFO option, we store rt_iif and specific
  * destination in skb->cb[] before dst drop.
  * This way, receiver doesn't make cache line misses to read rtable.
  */
-void ipv4_pktinfo_prepare(const struct sock *sk, struct sk_buff *skb)
+void ipv4_pktinfo_prepare(const struct sock *sk, struct sk_buff *skb, bool drop_dst)
 {
 	struct in_pktinfo *pktinfo = PKTINFO_SKB_CB(skb);
-	bool prepare = (inet_sk(sk)->cmsg_flags & IP_CMSG_PKTINFO) ||
+	bool prepare = inet_test_bit(PKTINFO, sk) ||
 		       ipv6_sk_rxinfo(sk);
 
 	if (prepare && skb_rtable(skb)) {
@@ -1441,7 +1402,8 @@ void ipv4_pktinfo_prepare(const struct sock *sk, struct sk_buff *skb)
 		pktinfo->ipi_ifindex = 0;
 		pktinfo->ipi_spec_dst.s_addr = 0;
 	}
-	skb_dst_drop(skb);
+	if (drop_dst)
+		skb_dst_drop(skb);
 }
 
 int ip_setsockopt(struct sock *sk, int level, int optname, sockptr_t optval,
@@ -1453,11 +1415,6 @@ int ip_setsockopt(struct sock *sk, int level, int optname, sockptr_t optval,
 		return -ENOPROTOOPT;
 
 	err = do_ip_setsockopt(sk, level, optname, optval, optlen);
-#if IS_ENABLED(CONFIG_BPFILTER_UMH)
-	if (optname >= BPFILTER_IPT_SO_SET_REPLACE &&
-	    optname < BPFILTER_IPT_SET_MAX)
-		err = bpfilter_ip_set_sockopt(sk, optname, optval, optlen);
-#endif
 #ifdef CONFIG_NETFILTER
 	/* we need to exclude all possible ENOPROTOOPTs except default case */
 	if (err == -ENOPROTOOPT && optname != IP_HDRINCL &&
@@ -1566,25 +1523,96 @@ int do_ip_getsockopt(struct sock *sk, int level, int optname,
 	if (len < 0)
 		return -EINVAL;
 
-	if (needs_rtnl)
-		rtnl_lock();
-	sockopt_lock_sock(sk);
-
+	/* Handle options that can be read without locking the socket. */
 	switch (optname) {
+	case IP_PKTINFO:
+		val = inet_test_bit(PKTINFO, sk);
+		goto copyval;
+	case IP_RECVTTL:
+		val = inet_test_bit(TTL, sk);
+		goto copyval;
+	case IP_RECVTOS:
+		val = inet_test_bit(TOS, sk);
+		goto copyval;
+	case IP_RECVOPTS:
+		val = inet_test_bit(RECVOPTS, sk);
+		goto copyval;
+	case IP_RETOPTS:
+		val = inet_test_bit(RETOPTS, sk);
+		goto copyval;
+	case IP_PASSSEC:
+		val = inet_test_bit(PASSSEC, sk);
+		goto copyval;
+	case IP_RECVORIGDSTADDR:
+		val = inet_test_bit(ORIGDSTADDR, sk);
+		goto copyval;
+	case IP_CHECKSUM:
+		val = inet_test_bit(CHECKSUM, sk);
+		goto copyval;
+	case IP_RECVFRAGSIZE:
+		val = inet_test_bit(RECVFRAGSIZE, sk);
+		goto copyval;
+	case IP_RECVERR:
+		val = inet_test_bit(RECVERR, sk);
+		goto copyval;
+	case IP_RECVERR_RFC4884:
+		val = inet_test_bit(RECVERR_RFC4884, sk);
+		goto copyval;
+	case IP_FREEBIND:
+		val = inet_test_bit(FREEBIND, sk);
+		goto copyval;
+	case IP_HDRINCL:
+		val = inet_test_bit(HDRINCL, sk);
+		goto copyval;
+	case IP_MULTICAST_LOOP:
+		val = inet_test_bit(MC_LOOP, sk);
+		goto copyval;
+	case IP_MULTICAST_ALL:
+		val = inet_test_bit(MC_ALL, sk);
+		goto copyval;
+	case IP_TRANSPARENT:
+		val = inet_test_bit(TRANSPARENT, sk);
+		goto copyval;
+	case IP_NODEFRAG:
+		val = inet_test_bit(NODEFRAG, sk);
+		goto copyval;
+	case IP_BIND_ADDRESS_NO_PORT:
+		val = inet_test_bit(BIND_ADDRESS_NO_PORT, sk);
+		goto copyval;
+	case IP_ROUTER_ALERT:
+		val = inet_test_bit(RTALERT, sk);
+		goto copyval;
+	case IP_TTL:
+		val = READ_ONCE(inet->uc_ttl);
+		if (val < 0)
+			val = READ_ONCE(sock_net(sk)->ipv4.sysctl_ip_default_ttl);
+		goto copyval;
+	case IP_MINTTL:
+		val = READ_ONCE(inet->min_ttl);
+		goto copyval;
+	case IP_MULTICAST_TTL:
+		val = READ_ONCE(inet->mc_ttl);
+		goto copyval;
+	case IP_MTU_DISCOVER:
+		val = READ_ONCE(inet->pmtudisc);
+		goto copyval;
+	case IP_TOS:
+		val = READ_ONCE(inet->tos);
+		goto copyval;
 	case IP_OPTIONS:
 	{
 		unsigned char optbuf[sizeof(struct ip_options)+40];
 		struct ip_options *opt = (struct ip_options *)optbuf;
 		struct ip_options_rcu *inet_opt;
 
-		inet_opt = rcu_dereference_protected(inet->inet_opt,
-						     lockdep_sock_is_held(sk));
+		rcu_read_lock();
+		inet_opt = rcu_dereference(inet->inet_opt);
 		opt->optlen = 0;
 		if (inet_opt)
 			memcpy(optbuf, &inet_opt->opt,
 			       sizeof(struct ip_options) +
 			       inet_opt->opt.optlen);
-		sockopt_release_sock(sk);
+		rcu_read_unlock();
 
 		if (opt->optlen == 0) {
 			len = 0;
@@ -1600,56 +1628,6 @@ int do_ip_getsockopt(struct sock *sk, int level, int optname,
 			return -EFAULT;
 		return 0;
 	}
-	case IP_PKTINFO:
-		val = (inet->cmsg_flags & IP_CMSG_PKTINFO) != 0;
-		break;
-	case IP_RECVTTL:
-		val = (inet->cmsg_flags & IP_CMSG_TTL) != 0;
-		break;
-	case IP_RECVTOS:
-		val = (inet->cmsg_flags & IP_CMSG_TOS) != 0;
-		break;
-	case IP_RECVOPTS:
-		val = (inet->cmsg_flags & IP_CMSG_RECVOPTS) != 0;
-		break;
-	case IP_RETOPTS:
-		val = (inet->cmsg_flags & IP_CMSG_RETOPTS) != 0;
-		break;
-	case IP_PASSSEC:
-		val = (inet->cmsg_flags & IP_CMSG_PASSSEC) != 0;
-		break;
-	case IP_RECVORIGDSTADDR:
-		val = (inet->cmsg_flags & IP_CMSG_ORIGDSTADDR) != 0;
-		break;
-	case IP_CHECKSUM:
-		val = (inet->cmsg_flags & IP_CMSG_CHECKSUM) != 0;
-		break;
-	case IP_RECVFRAGSIZE:
-		val = (inet->cmsg_flags & IP_CMSG_RECVFRAGSIZE) != 0;
-		break;
-	case IP_TOS:
-		val = inet->tos;
-		break;
-	case IP_TTL:
-	{
-		struct net *net = sock_net(sk);
-		val = (inet->uc_ttl == -1 ?
-		       READ_ONCE(net->ipv4.sysctl_ip_default_ttl) :
-		       inet->uc_ttl);
-		break;
-	}
-	case IP_HDRINCL:
-		val = inet->hdrincl;
-		break;
-	case IP_NODEFRAG:
-		val = inet->nodefrag;
-		break;
-	case IP_BIND_ADDRESS_NO_PORT:
-		val = inet->bind_address_no_port;
-		break;
-	case IP_MTU_DISCOVER:
-		val = inet->pmtudisc;
-		break;
 	case IP_MTU:
 	{
 		struct dst_entry *dst;
@@ -1659,33 +1637,55 @@ int do_ip_getsockopt(struct sock *sk, int level, int optname,
 			val = dst_mtu(dst);
 			dst_release(dst);
 		}
-		if (!val) {
-			sockopt_release_sock(sk);
+		if (!val)
 			return -ENOTCONN;
+		goto copyval;
+	}
+	case IP_PKTOPTIONS:
+	{
+		struct msghdr msg;
+
+		if (sk->sk_type != SOCK_STREAM)
+			return -ENOPROTOOPT;
+
+		if (optval.is_kernel) {
+			msg.msg_control_is_user = false;
+			msg.msg_control = optval.kernel;
+		} else {
+			msg.msg_control_is_user = true;
+			msg.msg_control_user = optval.user;
 		}
-		break;
+		msg.msg_controllen = len;
+		msg.msg_flags = in_compat_syscall() ? MSG_CMSG_COMPAT : 0;
+
+		if (inet_test_bit(PKTINFO, sk)) {
+			struct in_pktinfo info;
+
+			info.ipi_addr.s_addr = READ_ONCE(inet->inet_rcv_saddr);
+			info.ipi_spec_dst.s_addr = READ_ONCE(inet->inet_rcv_saddr);
+			info.ipi_ifindex = READ_ONCE(inet->mc_index);
+			put_cmsg(&msg, SOL_IP, IP_PKTINFO, sizeof(info), &info);
+		}
+		if (inet_test_bit(TTL, sk)) {
+			int hlim = READ_ONCE(inet->mc_ttl);
+
+			put_cmsg(&msg, SOL_IP, IP_TTL, sizeof(hlim), &hlim);
+		}
+		if (inet_test_bit(TOS, sk)) {
+			int tos = READ_ONCE(inet->rcv_tos);
+			put_cmsg(&msg, SOL_IP, IP_TOS, sizeof(tos), &tos);
+		}
+		len -= msg.msg_controllen;
+		return copy_to_sockptr(optlen, &len, sizeof(int));
 	}
-	case IP_RECVERR:
-		val = inet->recverr;
-		break;
-	case IP_RECVERR_RFC4884:
-		val = inet->recverr_rfc4884;
-		break;
-	case IP_MULTICAST_TTL:
-		val = inet->mc_ttl;
-		break;
-	case IP_MULTICAST_LOOP:
-		val = inet->mc_loop;
-		break;
 	case IP_UNICAST_IF:
-		val = (__force int)htonl((__u32) inet->uc_index);
-		break;
+		val = (__force int)htonl((__u32) READ_ONCE(inet->uc_index));
+		goto copyval;
 	case IP_MULTICAST_IF:
 	{
 		struct in_addr addr;
 		len = min_t(unsigned int, len, sizeof(struct in_addr));
-		addr.s_addr = inet->mc_addr;
-		sockopt_release_sock(sk);
+		addr.s_addr = READ_ONCE(inet->mc_addr);
 
 		if (copy_to_sockptr(optlen, &len, sizeof(int)))
 			return -EFAULT;
@@ -1693,6 +1693,16 @@ int do_ip_getsockopt(struct sock *sk, int level, int optname,
 			return -EFAULT;
 		return 0;
 	}
+	case IP_LOCAL_PORT_RANGE:
+		val = READ_ONCE(inet->local_port_range);
+		goto copyval;
+	}
+
+	if (needs_rtnl)
+		rtnl_lock();
+	sockopt_lock_sock(sk);
+
+	switch (optname) {
 	case IP_MSFILTER:
 	{
 		struct ip_msfilter msf;
@@ -1715,59 +1725,6 @@ int do_ip_getsockopt(struct sock *sk, int level, int optname,
 		else
 			err = ip_get_mcast_msfilter(sk, optval, optlen, len);
 		goto out;
-	case IP_MULTICAST_ALL:
-		val = inet->mc_all;
-		break;
-	case IP_PKTOPTIONS:
-	{
-		struct msghdr msg;
-
-		sockopt_release_sock(sk);
-
-		if (sk->sk_type != SOCK_STREAM)
-			return -ENOPROTOOPT;
-
-		if (optval.is_kernel) {
-			msg.msg_control_is_user = false;
-			msg.msg_control = optval.kernel;
-		} else {
-			msg.msg_control_is_user = true;
-			msg.msg_control_user = optval.user;
-		}
-		msg.msg_controllen = len;
-		msg.msg_flags = in_compat_syscall() ? MSG_CMSG_COMPAT : 0;
-
-		if (inet->cmsg_flags & IP_CMSG_PKTINFO) {
-			struct in_pktinfo info;
-
-			info.ipi_addr.s_addr = inet->inet_rcv_saddr;
-			info.ipi_spec_dst.s_addr = inet->inet_rcv_saddr;
-			info.ipi_ifindex = inet->mc_index;
-			put_cmsg(&msg, SOL_IP, IP_PKTINFO, sizeof(info), &info);
-		}
-		if (inet->cmsg_flags & IP_CMSG_TTL) {
-			int hlim = inet->mc_ttl;
-			put_cmsg(&msg, SOL_IP, IP_TTL, sizeof(hlim), &hlim);
-		}
-		if (inet->cmsg_flags & IP_CMSG_TOS) {
-			int tos = inet->rcv_tos;
-			put_cmsg(&msg, SOL_IP, IP_TOS, sizeof(tos), &tos);
-		}
-		len -= msg.msg_controllen;
-		return copy_to_sockptr(optlen, &len, sizeof(int));
-	}
-	case IP_FREEBIND:
-		val = inet->freebind;
-		break;
-	case IP_TRANSPARENT:
-		val = inet->transparent;
-		break;
-	case IP_MINTTL:
-		val = inet->min_ttl;
-		break;
-	case IP_LOCAL_PORT_RANGE:
-		val = inet->local_port_range.hi << 16 | inet->local_port_range.lo;
-		break;
 	case IP_PROTOCOL:
 		val = inet_sk(sk)->inet_num;
 		break;
@@ -1776,7 +1733,7 @@ int do_ip_getsockopt(struct sock *sk, int level, int optname,
 		return -ENOPROTOOPT;
 	}
 	sockopt_release_sock(sk);
-
+copyval:
 	if (len < sizeof(int) && len > 0 && val >= 0 && val <= 255) {
 		unsigned char ucval = (unsigned char)val;
 		len = 1;
@@ -1808,11 +1765,6 @@ int ip_getsockopt(struct sock *sk, int level,
 	err = do_ip_getsockopt(sk, level, optname,
 			       USER_SOCKPTR(optval), USER_SOCKPTR(optlen));
 
-#if IS_ENABLED(CONFIG_BPFILTER_UMH)
-	if (optname >= BPFILTER_IPT_SO_GET_INFO &&
-	    optname < BPFILTER_IPT_GET_MAX)
-		err = bpfilter_ip_get_sockopt(sk, optname, optval, optlen);
-#endif
 #ifdef CONFIG_NETFILTER
 	/* we need to exclude all possible ENOPROTOOPTs except default case */
 	if (err == -ENOPROTOOPT && optname != IP_PKTOPTIONS &&