int ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len)
{
struct inet_sock *inet = inet_sk(sk);
struct sockaddr_in *usin = (struct sockaddr_in *) uaddr;
struct rtable *rt;
__be32 saddr;
int oif;
int err;
if (addr_len < sizeof(*usin))
return -EINVAL;
if (usin->sin_family != AF_INET)
return -EAFNOSUPPORT;
sk_dst_reset(sk);
oif = sk->sk_bound_dev_if;
saddr = inet->inet_saddr;
if (ipv4_is_multicast(usin->sin_addr.s_addr)) {
if (!oif)
oif = inet->mc_index;
if (!saddr)
saddr = inet->mc_addr;
}
err = ip_route_connect(&rt, usin->sin_addr.s_addr, saddr,
RT_CONN_FLAGS(sk), oif,
sk->sk_protocol,
inet->inet_sport, usin->sin_port, sk, 1);
if (err) {
if (err == -ENETUNREACH)
IP_INC_STATS_BH(sock_net(sk), IPSTATS_MIB_OUTNOROUTES);
return err;
}
if ((rt->rt_flags & RTCF_BROADCAST) && !sock_flag(sk, SOCK_BROADCAST)) {
ip_rt_put(rt);
return -EACCES;
}
if (!inet->inet_saddr)
inet->inet_saddr = rt->rt_src; /* Update source address */
if (!inet->inet_rcv_saddr) {
inet->inet_rcv_saddr = rt->rt_src;
if (sk->sk_prot->rehash)
sk->sk_prot->rehash(sk);
}
inet->inet_daddr = rt->rt_dst;
inet->inet_dport = usin->sin_port;
sk->sk_state = TCP_ESTABLISHED;
inet->inet_id = jiffies;
sk_dst_set(sk, &rt->dst);
return(0);
}
static struct dst_entry *rxe_find_route(struct net_device *ndev,
struct rxe_qp *qp,
struct rxe_av *av)
{
struct dst_entry *dst = NULL;
if (qp_type(qp) == IB_QPT_RC)
dst = sk_dst_get(qp->sk->sk);
if (!dst || !dst_check(dst, qp->dst_cookie)) {
if (dst)
dst_release(dst);
if (av->network_type == RDMA_NETWORK_IPV4) {
struct in_addr *saddr;
struct in_addr *daddr;
saddr = &av->sgid_addr._sockaddr_in.sin_addr;
daddr = &av->dgid_addr._sockaddr_in.sin_addr;
dst = rxe_find_route4(ndev, saddr, daddr);
} else if (av->network_type == RDMA_NETWORK_IPV6) {
struct in6_addr *saddr6;
struct in6_addr *daddr6;
saddr6 = &av->sgid_addr._sockaddr_in6.sin6_addr;
daddr6 = &av->dgid_addr._sockaddr_in6.sin6_addr;
dst = rxe_find_route6(ndev, saddr6, daddr6);
#if IS_ENABLED(CONFIG_IPV6)
if (dst)
qp->dst_cookie =
rt6_get_cookie((struct rt6_info *)dst);
#endif
}
if (dst && (qp_type(qp) == IB_QPT_RC)) {
dst_hold(dst);
sk_dst_set(qp->sk->sk, dst);
}
}
return dst;
}
static int
mptp_sendmsg(struct kiocb *iocb, struct socket *sock,
struct msghdr *msg, size_t len)
{
int err;
uint16_t dport;
__be32 daddr;
__be32 saddr;
uint16_t sport;
struct sk_buff *skb;
struct sock *sk;
struct inet_sock *isk;
struct mptp_sock *ssk;
struct mptphdr *shdr;
int connected = 0;
int totlen;
struct rtable *rt = NULL;
int dests = 0;
int i;
struct sockaddr_mptp *mptp_addr = NULL;
int ret = 0;
if (unlikely(sock == NULL)) {
log_error("Sock is NULL\n");
err = -EINVAL;
goto out;
}
sk = sock->sk;
if (unlikely(sk == NULL)) {
log_error("Sock->sk is NULL\n");
err = -EINVAL;
goto out;
}
isk = inet_sk(sk);
ssk = mptp_sk(sk);
sport = ssk->src;
saddr = isk->inet_saddr;
if (sport == 0) {
sport = get_next_free_port();
if (unlikely(sport == 0)) {
log_error("No free ports\n");
err = -ENOMEM;
goto out;
}
}
if (msg->msg_name) {
mptp_addr = (struct sockaddr_mptp *)msg->msg_name;
if (unlikely
(msg->msg_namelen <
sizeof(*mptp_addr) +
mptp_addr->count * sizeof(struct mptp_dest)
|| mptp_addr->count <= 0)) {
log_error
("Invalid size for msg_name (size=%u, addr_count=%u)\n",
msg->msg_namelen, mptp_addr->count);
err = -EINVAL;
goto out;
}
dests = mptp_addr->count;
} else {
BUG();
if (unlikely(!ssk->dst || !isk->inet_daddr)) {
log_error("No destination port/address\n");
err = -EDESTADDRREQ;
goto out;
}
dport = ssk->dst;
daddr = isk->inet_daddr;
log_debug
("Got from socket destination port=%u and address=%u\n",
dport, daddr);
connected = 1;
}
if (msg->msg_iovlen < dests)
dests = msg->msg_iovlen;
for (i = 0; i < dests; i++) {
struct mptp_dest *dest = &mptp_addr->dests[i];
struct iovec *iov = &msg->msg_iov[i];
char *payload;
#if LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 39)
struct flowi fl = {};
#endif
dport = ntohs(dest->port);
if (unlikely(dport == 0 || dport >= MAX_MPTP_PORT)) {
log_error("Invalid value for destination port(%u)\n",
dport);
err = -EINVAL;
goto out;
}
daddr = dest->addr;
#if LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 39)
fl.u.ip4.saddr = saddr;
fl.u.ip4.daddr = daddr;
fl.flowi_proto = sk->sk_protocol;
fl.flowi_flags = inet_sk_flowi_flags(sk);
#endif
log_debug
("Received from user space destination port=%u and address=%u\n",
dport, daddr);
len = iov->iov_len;
totlen = len + sizeof(struct mptphdr) + sizeof(struct iphdr);
skb =
sock_alloc_send_skb(sk, totlen,
msg->msg_flags & MSG_DONTWAIT, &err);
if (unlikely(!skb)) {
log_error("sock_alloc_send_skb failed\n");
goto out;
}
log_debug("Allocated %u bytes for skb (payload size=%u)\n",
totlen, len);
skb_reset_network_header(skb);
skb_reserve(skb, sizeof(struct iphdr));
log_debug("Reseted network header\n");
skb_reset_transport_header(skb);
skb_put(skb, sizeof(struct mptphdr));
log_debug("Reseted transport header\n");
shdr = (struct mptphdr *)skb_transport_header(skb);
shdr->dst = htons(dport);
shdr->src = htons(sport);
shdr->len = htons(len + sizeof(struct mptphdr));
payload = skb_put(skb, len);
log_debug("payload=%p\n", payload);
err =
skb_copy_datagram_from_iovec(skb, sizeof(struct mptphdr),
iov, 0, len);
if (unlikely(err)) {
log_error("skb_copy_datagram_from_iovec failed\n");
goto out_free;
}
log_debug("Copied %u bytes into the skb\n", len);
if (connected)
rt = (struct rtable *)__sk_dst_check(sk, 0);
if (rt == NULL) {
log_debug("rt == NULL\n");
#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 39)
struct flowi fl = {.fl4_dst = daddr,
.proto = sk->sk_protocol,
.flags = inet_sk_flowi_flags(sk),
};
err =
ip_route_output_flow(sock_net(sk), &rt, &fl, sk, 0);
if (unlikely(err)) {
log_error("Route lookup failed\n");
goto out_free;
}
#else
rt = ip_route_output_flow(sock_net(sk), &fl.u.ip4, sk);
log_debug("rt = %p\n", rt);
if (IS_ERR(rt)) {
log_error("Route lookup failed\n");
goto out_free;
}
#endif
#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 36)
sk_dst_set(sk, dst_clone(&rt->u.dst));
#else
sk_dst_set(sk, dst_clone(&rt->dst));
#endif
}
log_debug("rt != NULL\n");
skb->local_df = 1;
#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 39)
err = ip_queue_xmit(skb);
#else
err = ip_queue_xmit(skb, &fl);
#endif
if (likely(!err)) {
log_debug("Sent %u bytes on wire\n", len);
ret += len;
dest->bytes = len;
} else {
log_error("ip_queue_xmit failed\n");
dest->bytes = -1;
}
}
return ret;
out_free:
kfree(skb);
out:
return err;
}
int ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len)
{
struct inet_sock *inet = inet_sk(sk);
struct sockaddr_in *usin = (struct sockaddr_in *) uaddr;
struct flowi4 *fl4;
struct rtable *rt;
__be32 saddr;
int oif;
int err;
if (addr_len < sizeof(*usin))
return -EINVAL;
if (usin->sin_family != AF_INET)
return -EAFNOSUPPORT;
sk_dst_reset(sk);
lock_sock(sk);
oif = sk->sk_bound_dev_if;
saddr = inet->inet_saddr;
if (ipv4_is_multicast(usin->sin_addr.s_addr)) {
if (!oif)
oif = inet->mc_index;
if (!saddr)
saddr = inet->mc_addr;
}
fl4 = &inet->cork.fl.u.ip4;
rt = ip_route_connect(fl4, usin->sin_addr.s_addr, saddr,
RT_CONN_FLAGS(sk), oif,
sk->sk_protocol,
inet->inet_sport, usin->sin_port, sk, true);
if (IS_ERR(rt)) {
err = PTR_ERR(rt);
if (err == -ENETUNREACH)
IP_INC_STATS_BH(sock_net(sk), IPSTATS_MIB_OUTNOROUTES);
goto out;
}
if ((rt->rt_flags & RTCF_BROADCAST) && !sock_flag(sk, SOCK_BROADCAST)) {
ip_rt_put(rt);
err = -EACCES;
goto out;
}
if (!inet->inet_saddr)
inet->inet_saddr = fl4->saddr;
if (!inet->inet_rcv_saddr) {
inet->inet_rcv_saddr = fl4->saddr;
if (sk->sk_prot->rehash)
sk->sk_prot->rehash(sk);
}
inet->inet_daddr = fl4->daddr;
inet->inet_dport = usin->sin_port;
sk->sk_state = TCP_ESTABLISHED;
inet->inet_id = jiffies;
sk_dst_set(sk, &rt->dst);
err = 0;
out:
release_sock(sk);
return err;
}