diff options
Diffstat (limited to 'net/ipv4')
-rw-r--r-- | net/ipv4/af_inet.c | 4 | ||||
-rw-r--r-- | net/ipv4/esp4.c | 4 | ||||
-rw-r--r-- | net/ipv4/esp4_offload.c | 3 | ||||
-rw-r--r-- | net/ipv4/fib_frontend.c | 3 | ||||
-rw-r--r-- | net/ipv4/icmp.c | 5 | ||||
-rw-r--r-- | net/ipv4/igmp.c | 9 | ||||
-rw-r--r-- | net/ipv4/inet_connection_sock.c | 18 | ||||
-rw-r--r-- | net/ipv4/inet_hashtables.c | 12 | ||||
-rw-r--r-- | net/ipv4/ip_gre.c | 21 | ||||
-rw-r--r-- | net/ipv4/ip_output.c | 18 | ||||
-rw-r--r-- | net/ipv4/ip_sockglue.c | 12 | ||||
-rw-r--r-- | net/ipv4/ip_tunnel.c | 12 | ||||
-rw-r--r-- | net/ipv4/ip_vti.c | 4 | ||||
-rw-r--r-- | net/ipv4/metrics.c | 2 | ||||
-rw-r--r-- | net/ipv4/netfilter/nf_socket_ipv4.c | 6 | ||||
-rw-r--r-- | net/ipv4/netfilter/nft_dup_ipv4.c | 18 | ||||
-rw-r--r-- | net/ipv4/raw.c | 5 | ||||
-rw-r--r-- | net/ipv4/route.c | 6 | ||||
-rw-r--r-- | net/ipv4/tcp.c | 63 | ||||
-rw-r--r-- | net/ipv4/tcp_fastopen.c | 6 | ||||
-rw-r--r-- | net/ipv4/tcp_input.c | 38 | ||||
-rw-r--r-- | net/ipv4/tcp_metrics.c | 85 | ||||
-rw-r--r-- | net/ipv4/tcp_minisocks.c | 7 | ||||
-rw-r--r-- | net/ipv4/tcp_output.c | 42 | ||||
-rw-r--r-- | net/ipv4/tcp_recovery.c | 2 | ||||
-rw-r--r-- | net/ipv4/tcp_timer.c | 4 | ||||
-rw-r--r-- | net/ipv4/udp.c | 9 | ||||
-rw-r--r-- | net/ipv4/udp_tunnel.c | 1 | ||||
-rw-r--r-- | net/ipv4/udplite.c | 10 |
29 files changed, 288 insertions, 141 deletions
diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c index 4b26ae525d6d..7c902a1efbbf 100644 --- a/net/ipv4/af_inet.c +++ b/net/ipv4/af_inet.c @@ -578,6 +578,7 @@ static long inet_wait_for_connect(struct sock *sk, long timeo, int writebias) add_wait_queue(sk_sleep(sk), &wait); sk->sk_write_pending += writebias; + sk->sk_wait_pending++; /* Basic assumption: if someone sets sk->sk_err, he _must_ * change state of the socket from TCP_SYN_*. @@ -593,6 +594,7 @@ static long inet_wait_for_connect(struct sock *sk, long timeo, int writebias) } remove_wait_queue(sk_sleep(sk), &wait); sk->sk_write_pending -= writebias; + sk->sk_wait_pending--; return timeo; } @@ -865,7 +867,7 @@ int inet_shutdown(struct socket *sock, int how) EPOLLHUP, even on eg. unconnected UDP sockets -- RR */ /* fall through */ default: - sk->sk_shutdown |= how; + WRITE_ONCE(sk->sk_shutdown, sk->sk_shutdown | how); if (sk->sk_prot->shutdown) sk->sk_prot->shutdown(sk, how); break; diff --git a/net/ipv4/esp4.c b/net/ipv4/esp4.c index 203569500b91..24cd5c9c7839 100644 --- a/net/ipv4/esp4.c +++ b/net/ipv4/esp4.c @@ -565,7 +565,9 @@ static inline int esp_remove_trailer(struct sk_buff *skb) skb->csum = csum_block_sub(skb->csum, csumdiff, skb->len - trimlen); } - pskb_trim(skb, skb->len - trimlen); + ret = pskb_trim(skb, skb->len - trimlen); + if (unlikely(ret)) + return ret; ret = nexthdr[1]; diff --git a/net/ipv4/esp4_offload.c b/net/ipv4/esp4_offload.c index 58834a10c0be..93045373e44b 100644 --- a/net/ipv4/esp4_offload.c +++ b/net/ipv4/esp4_offload.c @@ -237,6 +237,9 @@ static int esp_xmit(struct xfrm_state *x, struct sk_buff *skb, netdev_features_ secpath_reset(skb); + if (skb_needs_linearize(skb, skb->dev->features) && + __skb_linearize(skb)) + return -ENOMEM; return 0; } diff --git a/net/ipv4/fib_frontend.c b/net/ipv4/fib_frontend.c index 1885a2fbad86..9aa48b4c4096 100644 --- a/net/ipv4/fib_frontend.c +++ b/net/ipv4/fib_frontend.c @@ -557,6 +557,9 @@ static int rtentry_to_fib_config(struct net *net, int cmd, struct rtentry *rt, cfg->fc_scope = RT_SCOPE_UNIVERSE; } + if (!cfg->fc_table) + cfg->fc_table = RT_TABLE_MAIN; + if (cmd == SIOCDELRT) return 0; diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c index aa179e6461e1..af0ddaa55e43 100644 --- a/net/ipv4/icmp.c +++ b/net/ipv4/icmp.c @@ -759,6 +759,11 @@ void __icmp_send(struct sk_buff *skb_in, int type, int code, __be32 info, room = 576; room -= sizeof(struct iphdr) + icmp_param.replyopts.opt.opt.optlen; room -= sizeof(struct icmphdr); + /* Guard against tiny mtu. We need to include at least one + * IP network header for this message to make any sense. + */ + if (room <= (int)sizeof(struct iphdr)) + goto ende; icmp_param.data_len = skb_in->len - icmp_param.offset; if (icmp_param.data_len > room) diff --git a/net/ipv4/igmp.c b/net/ipv4/igmp.c index a08acb54b6b0..5edf426fa414 100644 --- a/net/ipv4/igmp.c +++ b/net/ipv4/igmp.c @@ -221,8 +221,10 @@ static void igmp_start_timer(struct ip_mc_list *im, int max_delay) int tv = prandom_u32() % max_delay; im->tm_running = 1; - if (!mod_timer(&im->timer, jiffies+tv+2)) - refcount_inc(&im->refcnt); + if (refcount_inc_not_zero(&im->refcnt)) { + if (mod_timer(&im->timer, jiffies + tv + 2)) + ip_ma_put(im); + } } static void igmp_gq_start_timer(struct in_device *in_dev) @@ -357,8 +359,9 @@ static struct sk_buff *igmpv3_newpack(struct net_device *dev, unsigned int mtu) struct flowi4 fl4; int hlen = LL_RESERVED_SPACE(dev); int tlen = dev->needed_tailroom; - unsigned int size = mtu; + unsigned int size; + size = min(mtu, IP_MAX_MTU); while (1) { skb = alloc_skb(size + hlen + tlen, GFP_ATOMIC | __GFP_NOWARN); diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c index 0a69f92da71b..7392a744c677 100644 --- a/net/ipv4/inet_connection_sock.c +++ b/net/ipv4/inet_connection_sock.c @@ -826,6 +826,7 @@ struct sock *inet_csk_clone_lock(const struct sock *sk, if (newsk) { struct inet_connection_sock *newicsk = inet_csk(newsk); + newsk->sk_wait_pending = 0; inet_sk_set_state(newsk, TCP_SYN_RECV); newicsk->icsk_bind_hash = NULL; @@ -903,11 +904,25 @@ void inet_csk_prepare_forced_close(struct sock *sk) } EXPORT_SYMBOL(inet_csk_prepare_forced_close); +static int inet_ulp_can_listen(const struct sock *sk) +{ + const struct inet_connection_sock *icsk = inet_csk(sk); + + if (icsk->icsk_ulp_ops) + return -EINVAL; + + return 0; +} + int inet_csk_listen_start(struct sock *sk, int backlog) { struct inet_connection_sock *icsk = inet_csk(sk); struct inet_sock *inet = inet_sk(sk); - int err = -EADDRINUSE; + int err; + + err = inet_ulp_can_listen(sk); + if (unlikely(err)) + return err; reqsk_queue_alloc(&icsk->icsk_accept_queue); @@ -920,6 +935,7 @@ int inet_csk_listen_start(struct sock *sk, int backlog) * It is OK, because this socket enters to hash table only * after validation is complete. */ + err = -EADDRINUSE; inet_sk_state_store(sk, TCP_LISTEN); if (!sk->sk_prot->get_port(sk, inet->inet_num)) { inet->inet_sport = htons(inet->inet_num); diff --git a/net/ipv4/inet_hashtables.c b/net/ipv4/inet_hashtables.c index 3c58019f0718..c6d670cd872f 100644 --- a/net/ipv4/inet_hashtables.c +++ b/net/ipv4/inet_hashtables.c @@ -743,17 +743,7 @@ int __inet_hash_connect(struct inet_timewait_death_row *death_row, u32 index; if (port) { - head = &hinfo->bhash[inet_bhashfn(net, port, - hinfo->bhash_size)]; - tb = inet_csk(sk)->icsk_bind_hash; - spin_lock_bh(&head->lock); - if (sk_head(&tb->owners) == sk && !sk->sk_bind_node.next) { - inet_ehash_nolisten(sk, NULL, NULL); - spin_unlock_bh(&head->lock); - return 0; - } - spin_unlock(&head->lock); - /* No definite answer... Walk to established hash table */ + local_bh_disable(); ret = check_established(death_row, sk, port, NULL); local_bh_enable(); return ret; diff --git a/net/ipv4/ip_gre.c b/net/ipv4/ip_gre.c index c72432ce9bf5..38c8db78cda1 100644 --- a/net/ipv4/ip_gre.c +++ b/net/ipv4/ip_gre.c @@ -440,7 +440,7 @@ static void __gre_xmit(struct sk_buff *skb, struct net_device *dev, /* Push GRE header. */ gre_build_header(skb, tunnel->tun_hlen, flags, proto, tunnel->parms.o_key, - (flags & TUNNEL_SEQ) ? htonl(tunnel->o_seqno++) : 0); + (flags & TUNNEL_SEQ) ? htonl(atomic_fetch_inc(&tunnel->o_seqno)) : 0); ip_tunnel_xmit(skb, dev, tnl_params, tnl_params->protocol); } @@ -546,7 +546,7 @@ static void gre_fb_xmit(struct sk_buff *skb, struct net_device *dev, (TUNNEL_CSUM | TUNNEL_KEY | TUNNEL_SEQ); gre_build_header(skb, tunnel_hlen, flags, proto, tunnel_id_to_key32(tun_info->key.tun_id), - (flags & TUNNEL_SEQ) ? htonl(tunnel->o_seqno++) : 0); + (flags & TUNNEL_SEQ) ? htonl(atomic_fetch_inc(&tunnel->o_seqno)) : 0); df = key->tun_flags & TUNNEL_DONT_FRAGMENT ? htons(IP_DF) : 0; @@ -603,7 +603,7 @@ static void erspan_fb_xmit(struct sk_buff *skb, struct net_device *dev) truncate = true; } - nhoff = skb_network_header(skb) - skb_mac_header(skb); + nhoff = skb_network_offset(skb); if (skb->protocol == htons(ETH_P_IP) && (ntohs(ip_hdr(skb)->tot_len) > skb->len - nhoff)) truncate = true; @@ -612,7 +612,7 @@ static void erspan_fb_xmit(struct sk_buff *skb, struct net_device *dev) int thoff; if (skb_transport_header_was_set(skb)) - thoff = skb_transport_header(skb) - skb_mac_header(skb); + thoff = skb_transport_offset(skb); else thoff = nhoff + sizeof(struct ipv6hdr); if (ntohs(ipv6_hdr(skb)->payload_len) > skb->len - thoff) @@ -635,7 +635,7 @@ static void erspan_fb_xmit(struct sk_buff *skb, struct net_device *dev) } gre_build_header(skb, 8, TUNNEL_SEQ, - proto, 0, htonl(tunnel->o_seqno++)); + proto, 0, htonl(atomic_fetch_inc(&tunnel->o_seqno))); df = key->tun_flags & TUNNEL_DONT_FRAGMENT ? htons(IP_DF) : 0; @@ -683,15 +683,18 @@ static netdev_tx_t ipgre_xmit(struct sk_buff *skb, } if (dev->header_ops) { + int pull_len = tunnel->hlen + sizeof(struct iphdr); + if (skb_cow_head(skb, 0)) goto free_skb; tnl_params = (const struct iphdr *)skb->data; - /* Pull skb since ip_tunnel_xmit() needs skb->data pointing - * to gre header. - */ - skb_pull(skb, tunnel->hlen + sizeof(struct iphdr)); + if (!pskb_network_may_pull(skb, pull_len)) + goto free_skb; + + /* ip_tunnel_xmit() needs skb->data pointing to gre header. */ + skb_pull(skb, pull_len); skb_reset_mac_header(skb); if (skb->ip_summed == CHECKSUM_PARTIAL && diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c index 06a981676356..6936f703758b 100644 --- a/net/ipv4/ip_output.c +++ b/net/ipv4/ip_output.c @@ -221,7 +221,7 @@ static int ip_finish_output2(struct net *net, struct sock *sk, struct sk_buff *s if (lwtunnel_xmit_redirect(dst->lwtstate)) { int res = lwtunnel_xmit(skb); - if (res < 0 || res == LWTUNNEL_XMIT_DONE) + if (res != LWTUNNEL_XMIT_CONTINUE) return res; } @@ -1443,9 +1443,19 @@ struct sk_buff *__ip_make_skb(struct sock *sk, cork->dst = NULL; skb_dst_set(skb, &rt->dst); - if (iph->protocol == IPPROTO_ICMP) - icmp_out_count(net, ((struct icmphdr *) - skb_transport_header(skb))->type); + if (iph->protocol == IPPROTO_ICMP) { + u8 icmp_type; + + /* For such sockets, transhdrlen is zero when do ip_append_data(), + * so icmphdr does not in skb linear region and can not get icmp_type + * by icmp_hdr(skb)->type. + */ + if (sk->sk_type == SOCK_RAW && !inet_sk(sk)->hdrincl) + icmp_type = fl4->fl4_icmp_type; + else + icmp_type = icmp_hdr(skb)->type; + icmp_out_count(net, icmp_type); + } ip_cork_release(cork); out: diff --git a/net/ipv4/ip_sockglue.c b/net/ipv4/ip_sockglue.c index 82f341e84fae..fbf39077fc54 100644 --- a/net/ipv4/ip_sockglue.c +++ b/net/ipv4/ip_sockglue.c @@ -316,7 +316,14 @@ int ip_cmsg_send(struct sock *sk, struct msghdr *msg, struct ipcm_cookie *ipc, ipc->tos = val; ipc->priority = rt_tos2priority(ipc->tos); break; - + case IP_PROTOCOL: + if (cmsg->cmsg_len != CMSG_LEN(sizeof(int))) + return -EINVAL; + val = *(int *)CMSG_DATA(cmsg); + if (val < 1 || val > 255) + return -EINVAL; + ipc->protocol = val; + break; default: return -EINVAL; } @@ -1522,6 +1529,9 @@ static int do_ip_getsockopt(struct sock *sk, int level, int optname, case IP_MINTTL: val = inet->min_ttl; break; + case IP_PROTOCOL: + val = inet_sk(sk)->inet_num; + break; default: release_sock(sk); return -ENOPROTOOPT; diff --git a/net/ipv4/ip_tunnel.c b/net/ipv4/ip_tunnel.c index 30e93b4f831f..9c2381cf675d 100644 --- a/net/ipv4/ip_tunnel.c +++ b/net/ipv4/ip_tunnel.c @@ -609,10 +609,10 @@ void ip_md_tunnel_xmit(struct sk_buff *skb, struct net_device *dev, u8 proto) else if (skb->protocol == htons(ETH_P_IP)) df = inner_iph->frag_off & htons(IP_DF); headroom += LL_RESERVED_SPACE(rt->dst.dev) + rt->dst.header_len; - if (headroom > dev->needed_headroom) - dev->needed_headroom = headroom; + if (headroom > READ_ONCE(dev->needed_headroom)) + WRITE_ONCE(dev->needed_headroom, headroom); - if (skb_cow_head(skb, dev->needed_headroom)) { + if (skb_cow_head(skb, READ_ONCE(dev->needed_headroom))) { ip_rt_put(rt); goto tx_dropped; } @@ -777,10 +777,10 @@ void ip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev, max_headroom = LL_RESERVED_SPACE(rt->dst.dev) + sizeof(struct iphdr) + rt->dst.header_len + ip_encap_hlen(&tunnel->encap); - if (max_headroom > dev->needed_headroom) - dev->needed_headroom = max_headroom; + if (max_headroom > READ_ONCE(dev->needed_headroom)) + WRITE_ONCE(dev->needed_headroom, max_headroom); - if (skb_cow_head(skb, dev->needed_headroom)) { + if (skb_cow_head(skb, READ_ONCE(dev->needed_headroom))) { ip_rt_put(rt); dev->stats.tx_dropped++; kfree_skb(skb); diff --git a/net/ipv4/ip_vti.c b/net/ipv4/ip_vti.c index 15c71b08c2df..a3536dfe9b16 100644 --- a/net/ipv4/ip_vti.c +++ b/net/ipv4/ip_vti.c @@ -319,12 +319,12 @@ static netdev_tx_t vti_tunnel_xmit(struct sk_buff *skb, struct net_device *dev) switch (skb->protocol) { case htons(ETH_P_IP): - xfrm_decode_session(skb, &fl, AF_INET); memset(IPCB(skb), 0, sizeof(*IPCB(skb))); + xfrm_decode_session(skb, &fl, AF_INET); break; case htons(ETH_P_IPV6): - xfrm_decode_session(skb, &fl, AF_INET6); memset(IP6CB(skb), 0, sizeof(*IP6CB(skb))); + xfrm_decode_session(skb, &fl, AF_INET6); break; default: goto tx_err; diff --git a/net/ipv4/metrics.c b/net/ipv4/metrics.c index 04311f7067e2..9a6b01d85cd0 100644 --- a/net/ipv4/metrics.c +++ b/net/ipv4/metrics.c @@ -1,4 +1,5 @@ #include <linux/netlink.h> +#include <linux/nospec.h> #include <linux/rtnetlink.h> #include <linux/types.h> #include <net/ip.h> @@ -24,6 +25,7 @@ int ip_metrics_convert(struct net *net, struct nlattr *fc_mx, int fc_mx_len, if (type > RTAX_MAX) return -EINVAL; + type = array_index_nospec(type, RTAX_MAX + 1); if (type == RTAX_CC_ALGO) { char tmp[TCP_CA_NAME_MAX]; diff --git a/net/ipv4/netfilter/nf_socket_ipv4.c b/net/ipv4/netfilter/nf_socket_ipv4.c index 4824b1e183a1..bff2b85c5fd6 100644 --- a/net/ipv4/netfilter/nf_socket_ipv4.c +++ b/net/ipv4/netfilter/nf_socket_ipv4.c @@ -96,11 +96,11 @@ nf_socket_get_sock_v4(struct net *net, struct sk_buff *skb, const int doff, struct sock *nf_sk_lookup_slow_v4(struct net *net, const struct sk_buff *skb, const struct net_device *indev) { - __be32 uninitialized_var(daddr), uninitialized_var(saddr); - __be16 uninitialized_var(dport), uninitialized_var(sport); + __be32 daddr, saddr; + __be16 dport, sport; const struct iphdr *iph = ip_hdr(skb); struct sk_buff *data_skb = NULL; - u8 uninitialized_var(protocol); + u8 protocol; #if IS_ENABLED(CONFIG_NF_CONNTRACK) enum ip_conntrack_info ctinfo; struct nf_conn const *ct; diff --git a/net/ipv4/netfilter/nft_dup_ipv4.c b/net/ipv4/netfilter/nft_dup_ipv4.c index 0af3d8df70dd..157bca240edc 100644 --- a/net/ipv4/netfilter/nft_dup_ipv4.c +++ b/net/ipv4/netfilter/nft_dup_ipv4.c @@ -16,8 +16,8 @@ #include <net/netfilter/ipv4/nf_dup_ipv4.h> struct nft_dup_ipv4 { - enum nft_registers sreg_addr:8; - enum nft_registers sreg_dev:8; + u8 sreg_addr; + u8 sreg_dev; }; static void nft_dup_ipv4_eval(const struct nft_expr *expr, @@ -43,16 +43,16 @@ static int nft_dup_ipv4_init(const struct nft_ctx *ctx, if (tb[NFTA_DUP_SREG_ADDR] == NULL) return -EINVAL; - priv->sreg_addr = nft_parse_register(tb[NFTA_DUP_SREG_ADDR]); - err = nft_validate_register_load(priv->sreg_addr, sizeof(struct in_addr)); + err = nft_parse_register_load(tb[NFTA_DUP_SREG_ADDR], &priv->sreg_addr, + sizeof(struct in_addr)); if (err < 0) return err; - if (tb[NFTA_DUP_SREG_DEV] != NULL) { - priv->sreg_dev = nft_parse_register(tb[NFTA_DUP_SREG_DEV]); - return nft_validate_register_load(priv->sreg_dev, sizeof(int)); - } - return 0; + if (tb[NFTA_DUP_SREG_DEV]) + err = nft_parse_register_load(tb[NFTA_DUP_SREG_DEV], + &priv->sreg_dev, sizeof(int)); + + return err; } static int nft_dup_ipv4_dump(struct sk_buff *skb, const struct nft_expr *expr) diff --git a/net/ipv4/raw.c b/net/ipv4/raw.c index 654f586fc0d7..8ad120c07096 100644 --- a/net/ipv4/raw.c +++ b/net/ipv4/raw.c @@ -563,6 +563,9 @@ static int raw_sendmsg(struct sock *sk, struct msghdr *msg, size_t len) } ipcm_init_sk(&ipc, inet); + /* Keep backward compat */ + if (hdrincl) + ipc.protocol = IPPROTO_RAW; if (msg->msg_controllen) { err = ip_cmsg_send(sk, msg, &ipc, false); @@ -630,7 +633,7 @@ static int raw_sendmsg(struct sock *sk, struct msghdr *msg, size_t len) flowi4_init_output(&fl4, ipc.oif, sk->sk_mark, tos, RT_SCOPE_UNIVERSE, - hdrincl ? IPPROTO_RAW : sk->sk_protocol, + hdrincl ? ipc.protocol : sk->sk_protocol, inet_sk_flowi_flags(sk) | (hdrincl ? FLOWI_FLAG_KNOWN_NH : 0), daddr, saddr, 0, 0, sk->sk_uid); diff --git a/net/ipv4/route.c b/net/ipv4/route.c index 57e2316529d0..f4d41ceef946 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -791,7 +791,7 @@ static void __ip_do_redirect(struct rtable *rt, struct sk_buff *skb, struct flow goto reject_redirect; } - n = __ipv4_neigh_lookup(rt->dst.dev, new_gw); + n = __ipv4_neigh_lookup(rt->dst.dev, (__force u32)new_gw); if (!n) n = neigh_create(&arp_tbl, &new_gw, rt->dst.dev); if (!IS_ERR(n)) { @@ -1215,6 +1215,7 @@ static struct dst_entry *ipv4_dst_check(struct dst_entry *dst, u32 cookie) static void ipv4_send_dest_unreach(struct sk_buff *skb) { + struct net_device *dev; struct ip_options opt; int res; @@ -1232,7 +1233,8 @@ static void ipv4_send_dest_unreach(struct sk_buff *skb) opt.optlen = ip_hdr(skb)->ihl * 4 - sizeof(struct iphdr); rcu_read_lock(); - res = __ip_options_compile(dev_net(skb->dev), &opt, skb, NULL); + dev = skb->dev ? skb->dev : skb_rtable(skb)->dst.dev; + res = __ip_options_compile(dev_net(dev), &opt, skb, NULL); rcu_read_unlock(); if (res) diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index 68f89fe7f923..00648a478c6a 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -515,6 +515,7 @@ __poll_t tcp_poll(struct file *file, struct socket *sock, poll_table *wait) __poll_t mask; struct sock *sk = sock->sk; const struct tcp_sock *tp = tcp_sk(sk); + u8 shutdown; int state; sock_poll_wait(file, sock, wait); @@ -557,9 +558,10 @@ __poll_t tcp_poll(struct file *file, struct socket *sock, poll_table *wait) * NOTE. Check for TCP_CLOSE is added. The goal is to prevent * blocking on fresh not-connected or disconnected socket. --ANK */ - if (sk->sk_shutdown == SHUTDOWN_MASK || state == TCP_CLOSE) + shutdown = READ_ONCE(sk->sk_shutdown); + if (shutdown == SHUTDOWN_MASK || state == TCP_CLOSE) mask |= EPOLLHUP; - if (sk->sk_shutdown & RCV_SHUTDOWN) + if (shutdown & RCV_SHUTDOWN) mask |= EPOLLIN | EPOLLRDNORM | EPOLLRDHUP; /* Connected or passive Fast Open socket? */ @@ -575,8 +577,8 @@ __poll_t tcp_poll(struct file *file, struct socket *sock, poll_table *wait) if (tcp_stream_is_readable(tp, target, sk)) mask |= EPOLLIN | EPOLLRDNORM; - if (!(sk->sk_shutdown & SEND_SHUTDOWN)) { - if (sk_stream_is_writeable(sk)) { + if (!(shutdown & SEND_SHUTDOWN)) { + if (__sk_stream_is_writeable(sk, 1)) { mask |= EPOLLOUT | EPOLLWRNORM; } else { /* send SIGIO later */ sk_set_bit(SOCKWQ_ASYNC_NOSPACE, sk); @@ -588,7 +590,7 @@ __poll_t tcp_poll(struct file *file, struct socket *sock, poll_table *wait) * pairs with the input side. */ smp_mb__after_atomic(); - if (sk_stream_is_writeable(sk)) + if (__sk_stream_is_writeable(sk, 1)) mask |= EPOLLOUT | EPOLLWRNORM; } } else @@ -2338,14 +2340,13 @@ bool tcp_check_oom(struct sock *sk, int shift) return too_many_orphans || out_of_socket_memory; } -void tcp_close(struct sock *sk, long timeout) +void __tcp_close(struct sock *sk, long timeout) { struct sk_buff *skb; int data_was_unread = 0; int state; - lock_sock(sk); - sk->sk_shutdown = SHUTDOWN_MASK; + WRITE_ONCE(sk->sk_shutdown, SHUTDOWN_MASK); if (sk->sk_state == TCP_LISTEN) { tcp_set_state(sk, TCP_CLOSE); @@ -2505,6 +2506,12 @@ adjudge_to_death: out: bh_unlock_sock(sk); local_bh_enable(); +} + +void tcp_close(struct sock *sk, long timeout) +{ + lock_sock(sk); + __tcp_close(sk, timeout); release_sock(sk); sock_put(sk); } @@ -2561,6 +2568,12 @@ int tcp_disconnect(struct sock *sk, int flags) int old_state = sk->sk_state; u32 seq; + /* Deny disconnect if other threads are blocked in sk_wait_event() + * or inet_wait_for_connect(). + */ + if (sk->sk_wait_pending) + return -EBUSY; + if (old_state != TCP_CLOSE) tcp_set_state(sk, TCP_CLOSE); @@ -2593,7 +2606,7 @@ int tcp_disconnect(struct sock *sk, int flags) if (!(sk->sk_userlocks & SOCK_BINDADDR_LOCK)) inet_reset_saddr(sk); - sk->sk_shutdown = 0; + WRITE_ONCE(sk->sk_shutdown, 0); sock_reset_flag(sk, SOCK_DONE); tp->srtt_us = 0; tp->rcv_rtt_last_tsecr = 0; @@ -2987,18 +3000,18 @@ static int do_tcp_setsockopt(struct sock *sk, int level, case TCP_LINGER2: if (val < 0) - tp->linger2 = -1; - else if (val > net->ipv4.sysctl_tcp_fin_timeout / HZ) - tp->linger2 = 0; + WRITE_ONCE(tp->linger2, -1); + else if (val > TCP_FIN_TIMEOUT_MAX / HZ) + WRITE_ONCE(tp->linger2, TCP_FIN_TIMEOUT_MAX); else - tp->linger2 = val * HZ; + WRITE_ONCE(tp->linger2, val * HZ); break; case TCP_DEFER_ACCEPT: /* Translate value in seconds to number of retransmits */ - icsk->icsk_accept_queue.rskq_defer_accept = - secs_to_retrans(val, TCP_TIMEOUT_INIT / HZ, - TCP_RTO_MAX / HZ); + WRITE_ONCE(icsk->icsk_accept_queue.rskq_defer_accept, + secs_to_retrans(val, TCP_TIMEOUT_INIT / HZ, + TCP_RTO_MAX / HZ)); break; case TCP_WINDOW_CLAMP: @@ -3086,7 +3099,7 @@ static int do_tcp_setsockopt(struct sock *sk, int level, err = tcp_repair_set_window(tp, optval, optlen); break; case TCP_NOTSENT_LOWAT: - tp->notsent_lowat = val; + WRITE_ONCE(tp->notsent_lowat, val); sk->sk_write_space(sk); break; case TCP_INQ: @@ -3363,7 +3376,8 @@ static int do_tcp_getsockopt(struct sock *sk, int level, switch (optname) { case TCP_MAXSEG: val = tp->mss_cache; - if (!val && ((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_LISTEN))) + if (tp->rx_opt.user_mss && + ((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_LISTEN))) val = tp->rx_opt.user_mss; if (tp->repair) val = tp->rx_opt.mss_clamp; @@ -3387,13 +3401,14 @@ static int do_tcp_getsockopt(struct sock *sk, int level, val = icsk->icsk_syn_retries ? : net->ipv4.sysctl_tcp_syn_retries; break; case TCP_LINGER2: - val = tp->linger2; + val = READ_ONCE(tp->linger2); if (val >= 0) val = (val ? : READ_ONCE(net->ipv4.sysctl_tcp_fin_timeout)) / HZ; break; case TCP_DEFER_ACCEPT: - val = retrans_to_secs(icsk->icsk_accept_queue.rskq_defer_accept, - TCP_TIMEOUT_INIT / HZ, TCP_RTO_MAX / HZ); + val = READ_ONCE(icsk->icsk_accept_queue.rskq_defer_accept); + val = retrans_to_secs(val, TCP_TIMEOUT_INIT / HZ, + TCP_RTO_MAX / HZ); break; case TCP_WINDOW_CLAMP: val = tp->window_clamp; @@ -3539,7 +3554,7 @@ static int do_tcp_getsockopt(struct sock *sk, int level, break; case TCP_FASTOPEN: - val = icsk->icsk_accept_queue.fastopenq.max_qlen; + val = READ_ONCE(icsk->icsk_accept_queue.fastopenq.max_qlen); break; case TCP_FASTOPEN_CONNECT: @@ -3554,7 +3569,7 @@ static int do_tcp_getsockopt(struct sock *sk, int level, val = tcp_time_stamp_raw() + tp->tsoffset; break; case TCP_NOTSENT_LOWAT: - val = tp->notsent_lowat; + val = READ_ONCE(tp->notsent_lowat); break; case TCP_INQ: val = tp->recvmsg_inq; @@ -3802,7 +3817,7 @@ void tcp_done(struct sock *sk) if (req) reqsk_fastopen_remove(sk, req, false); - sk->sk_shutdown = SHUTDOWN_MASK; + WRITE_ONCE(sk->sk_shutdown, SHUTDOWN_MASK); if (!sock_flag(sk, SOCK_DEAD)) sk->sk_state_change(sk); diff --git a/net/ipv4/tcp_fastopen.c b/net/ipv4/tcp_fastopen.c index f726591de7c7..f7bb78b443fa 100644 --- a/net/ipv4/tcp_fastopen.c +++ b/net/ipv4/tcp_fastopen.c @@ -276,6 +276,7 @@ static struct sock *tcp_fastopen_create_child(struct sock *sk, static bool tcp_fastopen_queue_check(struct sock *sk) { struct fastopen_queue *fastopenq; + int max_qlen; /* Make sure the listener has enabled fastopen, and we don't * exceed the max # of pending TFO requests allowed before trying @@ -288,10 +289,11 @@ static bool tcp_fastopen_queue_check(struct sock *sk) * temporarily vs a server not supporting Fast Open at all. */ fastopenq = &inet_csk(sk)->icsk_accept_queue.fastopenq; - if (fastopenq->max_qlen == 0) + max_qlen = READ_ONCE(fastopenq->max_qlen); + if (max_qlen == 0) return false; - if (fastopenq->qlen >= fastopenq->max_qlen) { + if (fastopenq->qlen >= max_qlen) { struct request_sock *req1; spin_lock(&fastopenq->lock); req1 = fastopenq->rskq_rst_head; diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 11716780667c..407ad07dc598 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -172,6 +172,19 @@ static void tcp_measure_rcv_mss(struct sock *sk, const struct sk_buff *skb) if (unlikely(len > icsk->icsk_ack.rcv_mss + MAX_TCP_OPTION_SPACE)) tcp_gro_dev_warn(sk, skb, len); + /* If the skb has a len of exactly 1*MSS and has the PSH bit + * set then it is likely the end of an application write. So + * more data may not be arriving soon, and yet the data sender + * may be waiting for an ACK if cwnd-bound or using TX zero + * copy. So we set ICSK_ACK_PUSHED here so that + * tcp_cleanup_rbuf() will send an ACK immediately if the app + * reads all of the data and is not ping-pong. If len > MSS + * then this logic does not matter (and does not hurt) because + * tcp_cleanup_rbuf() will always ACK immediately if the app + * reads data and there is more than an MSS of unACKed data. + */ + if (TCP_SKB_CB(skb)->tcp_flags & TCPHDR_PSH) + icsk->icsk_ack.pending |= ICSK_ACK_PUSHED; } else { /* Otherwise, we make more careful check taking into account, * that SACKs block is variable. @@ -216,7 +229,7 @@ static void tcp_incr_quickack(struct sock *sk, unsigned int max_quickacks) icsk->icsk_ack.quick = quickacks; } -void tcp_enter_quickack_mode(struct sock *sk, unsigned int max_quickacks) +static void tcp_enter_quickack_mode(struct sock *sk, unsigned int max_quickacks) { struct inet_connection_sock *icsk = inet_csk(sk); @@ -224,7 +237,6 @@ void tcp_enter_quickack_mode(struct sock *sk, unsigned int max_quickacks) icsk->icsk_ack.pingpong = 0; icsk->icsk_ack.ato = TCP_ATO_MIN; } -EXPORT_SYMBOL(tcp_enter_quickack_mode); /* Send ACKs quickly, if "quick" count is not exhausted * and the session is not interactive. @@ -3429,8 +3441,11 @@ static int tcp_ack_update_window(struct sock *sk, const struct sk_buff *skb, u32 static bool __tcp_oow_rate_limited(struct net *net, int mib_idx, u32 *last_oow_ack_time) { - if (*last_oow_ack_time) { - s32 elapsed = (s32)(tcp_jiffies32 - *last_oow_ack_time); + /* Paired with the WRITE_ONCE() in this function. */ + u32 val = READ_ONCE(*last_oow_ack_time); + + if (val) { + s32 elapsed = (s32)(tcp_jiffies32 - val); if (0 <= elapsed && elapsed < READ_ONCE(net->ipv4.sysctl_tcp_invalid_ratelimit)) { @@ -3439,7 +3454,10 @@ static bool __tcp_oow_rate_limited(struct net *net, int mib_idx, } } - *last_oow_ack_time = tcp_jiffies32; + /* Paired with the prior READ_ONCE() and with itself, + * as we might be lockless. + */ + WRITE_ONCE(*last_oow_ack_time, tcp_jiffies32); return false; /* not rate-limited: go ahead, send dupack now! */ } @@ -3622,8 +3640,12 @@ static int tcp_ack(struct sock *sk, const struct sk_buff *skb, int flag) * then we can probably ignore it. */ if (before(ack, prior_snd_una)) { + u32 max_window; + + /* do not accept ACK for bytes we never sent. */ + max_window = min_t(u64, tp->max_window, tp->bytes_acked); /* RFC 5961 5.2 [Blind Data Injection Attack].[Mitigation] */ - if (before(ack, prior_snd_una - tp->max_window)) { + if (before(ack, prior_snd_una - max_window)) { if (!(flag & FLAG_NO_CHALLENGE_ACK)) tcp_send_challenge_ack(sk, skb); return -1; @@ -4130,7 +4152,7 @@ void tcp_fin(struct sock *sk) inet_csk_schedule_ack(sk); - sk->sk_shutdown |= RCV_SHUTDOWN; + WRITE_ONCE(sk->sk_shutdown, sk->sk_shutdown | RCV_SHUTDOWN); sock_set_flag(sk, SOCK_DONE); switch (sk->sk_state) { @@ -6209,7 +6231,7 @@ int tcp_rcv_state_process(struct sock *sk, struct sk_buff *skb) break; tcp_set_state(sk, TCP_FIN_WAIT2); - sk->sk_shutdown |= SEND_SHUTDOWN; + WRITE_ONCE(sk->sk_shutdown, sk->sk_shutdown | SEND_SHUTDOWN); sk_dst_confirm(sk); diff --git a/net/ipv4/tcp_metrics.c b/net/ipv4/tcp_metrics.c index 4960e2b6bd7f..60619b1f4acd 100644 --- a/net/ipv4/tcp_metrics.c +++ b/net/ipv4/tcp_metrics.c @@ -40,7 +40,7 @@ struct tcp_fastopen_metrics { struct tcp_metrics_block { struct tcp_metrics_block __rcu *tcpm_next; - possible_net_t tcpm_net; + struct net *tcpm_net; struct inetpeer_addr tcpm_saddr; struct inetpeer_addr tcpm_daddr; unsigned long tcpm_stamp; @@ -51,34 +51,38 @@ struct tcp_metrics_block { struct rcu_head rcu_head; }; -static inline struct net *tm_net(struct tcp_metrics_block *tm) +static inline struct net *tm_net(const struct tcp_metrics_block *tm) { - return read_pnet(&tm->tcpm_net); + /* Paired with the WRITE_ONCE() in tcpm_new() */ + return READ_ONCE(tm->tcpm_net); } static bool tcp_metric_locked(struct tcp_metrics_block *tm, enum tcp_metric_index idx) { - return tm->tcpm_lock & (1 << idx); + /* Paired with WRITE_ONCE() in tcpm_suck_dst() */ + return READ_ONCE(tm->tcpm_lock) & (1 << idx); } -static u32 tcp_metric_get(struct tcp_metrics_block *tm, +static u32 tcp_metric_get(const struct tcp_metrics_block *tm, enum tcp_metric_index idx) { - return tm->tcpm_vals[idx]; + /* Paired with WRITE_ONCE() in tcp_metric_set() */ + return READ_ONCE(tm->tcpm_vals[idx]); } static void tcp_metric_set(struct tcp_metrics_block *tm, enum tcp_metric_index idx, u32 val) { - tm->tcpm_vals[idx] = val; + /* Paired with READ_ONCE() in tcp_metric_get() */ + WRITE_ONCE(tm->tcpm_vals[idx], val); } static bool addr_same(const struct inetpeer_addr *a, const struct inetpeer_addr *b) { - return inetpeer_addr_cmp(a, b) == 0; + return (a->family == b->family) && !inetpeer_addr_cmp(a, b); } struct tcpm_hash_bucket { @@ -89,6 +93,7 @@ static struct tcpm_hash_bucket *tcp_metrics_hash __read_mostly; static unsigned int tcp_metrics_hash_log __read_mostly; static DEFINE_SPINLOCK(tcp_metrics_lock); +static DEFINE_SEQLOCK(fastopen_seqlock); static void tcpm_suck_dst(struct tcp_metrics_block *tm, const struct dst_entry *dst, @@ -97,7 +102,7 @@ static void tcpm_suck_dst(struct tcp_metrics_block *tm, u32 msval; u32 val; - tm->tcpm_stamp = jiffies; + WRITE_ONCE(tm->tcpm_stamp, jiffies); val = 0; if (dst_metric_locked(dst, RTAX_RTT)) @@ -110,30 +115,42 @@ static void tcpm_suck_dst(struct tcp_metrics_block *tm, val |= 1 << TCP_METRIC_CWND; if (dst_metric_locked(dst, RTAX_REORDERING)) val |= 1 << TCP_METRIC_REORDERING; - tm->tcpm_lock = val; + /* Paired with READ_ONCE() in tcp_metric_locked() */ + WRITE_ONCE(tm->tcpm_lock, val); msval = dst_metric_raw(dst, RTAX_RTT); - tm->tcpm_vals[TCP_METRIC_RTT] = msval * USEC_PER_MSEC; + tcp_metric_set(tm, TCP_METRIC_RTT, msval * USEC_PER_MSEC); msval = dst_metric_raw(dst, RTAX_RTTVAR); - tm->tcpm_vals[TCP_METRIC_RTTVAR] = msval * USEC_PER_MSEC; - tm->tcpm_vals[TCP_METRIC_SSTHRESH] = dst_metric_raw(dst, RTAX_SSTHRESH); - tm->tcpm_vals[TCP_METRIC_CWND] = dst_metric_raw(dst, RTAX_CWND); - tm->tcpm_vals[TCP_METRIC_REORDERING] = dst_metric_raw(dst, RTAX_REORDERING); + tcp_metric_set(tm, TCP_METRIC_RTTVAR, msval * USEC_PER_MSEC); + tcp_metric_set(tm, TCP_METRIC_SSTHRESH, + dst_metric_raw(dst, RTAX_SSTHRESH)); + tcp_metric_set(tm, TCP_METRIC_CWND, + dst_metric_raw(dst, RTAX_CWND)); + tcp_metric_set(tm, TCP_METRIC_REORDERING, + dst_metric_raw(dst, RTAX_REORDERING)); if (fastopen_clear) { + write_seqlock(&fastopen_seqlock); tm->tcpm_fastopen.mss = 0; tm->tcpm_fastopen.syn_loss = 0; tm->tcpm_fastopen.try_exp = 0; tm->tcpm_fastopen.cookie.exp = false; tm->tcpm_fastopen.cookie.len = 0; + write_sequnlock(&fastopen_seqlock); } } #define TCP_METRICS_TIMEOUT (60 * 60 * HZ) -static void tcpm_check_stamp(struct tcp_metrics_block *tm, struct dst_entry *dst) +static void tcpm_check_stamp(struct tcp_metrics_block *tm, + const struct dst_entry *dst) { - if (tm && unlikely(time_after(jiffies, tm->tcpm_stamp + TCP_METRICS_TIMEOUT))) + unsigned long limit; + + if (!tm) + return; + limit = READ_ONCE(tm->tcpm_stamp) + TCP_METRICS_TIMEOUT; + if (unlikely(time_after(jiffies, limit))) tcpm_suck_dst(tm, dst, false); } @@ -174,20 +191,23 @@ static struct tcp_metrics_block *tcpm_new(struct dst_entry *dst, oldest = deref_locked(tcp_metrics_hash[hash].chain); for (tm = deref_locked(oldest->tcpm_next); tm; tm = deref_locked(tm->tcpm_next)) { - if (time_before(tm->tcpm_stamp, oldest->tcpm_stamp)) + if (time_before(READ_ONCE(tm->tcpm_stamp), + READ_ONCE(oldest->tcpm_stamp))) oldest = tm; } tm = oldest; } else { - tm = kmalloc(sizeof(*tm), GFP_ATOMIC); + tm = kzalloc(sizeof(*tm), GFP_ATOMIC); if (!tm) goto out_unlock; } - write_pnet(&tm->tcpm_net, net); + /* Paired with the READ_ONCE() in tm_net() */ + WRITE_ONCE(tm->tcpm_net, net); + tm->tcpm_saddr = *saddr; tm->tcpm_daddr = *daddr; - tcpm_suck_dst(tm, dst, true); + tcpm_suck_dst(tm, dst, reclaim); if (likely(!reclaim)) { tm->tcpm_next = tcp_metrics_hash[hash].chain; @@ -431,7 +451,7 @@ void tcp_update_metrics(struct sock *sk) tp->reordering); } } - tm->tcpm_stamp = jiffies; + WRITE_ONCE(tm->tcpm_stamp, jiffies); out_unlock: rcu_read_unlock(); } @@ -446,11 +466,15 @@ void tcp_init_metrics(struct sock *sk) u32 val, crtt = 0; /* cached RTT scaled by 8 */ sk_dst_confirm(sk); + /* ssthresh may have been reduced unnecessarily during. + * 3WHS. Restore it back to its initial default. + */ + tp->snd_ssthresh = TCP_INFINITE_SSTHRESH; if (!dst) goto reset; rcu_read_lock(); - tm = tcp_get_metrics(sk, dst, true); + tm = tcp_get_metrics(sk, dst, false); if (!tm) { rcu_read_unlock(); goto reset; @@ -464,11 +488,6 @@ void tcp_init_metrics(struct sock *sk) tp->snd_ssthresh = val; if (tp->snd_ssthresh > tp->snd_cwnd_clamp) tp->snd_ssthresh = tp->snd_cwnd_clamp; - } else { - /* ssthresh may have been reduced unnecessarily during. - * 3WHS. Restore it back to its initial default. - */ - tp->snd_ssthresh = TCP_INFINITE_SSTHRESH; } val = tcp_metric_get(tm, TCP_METRIC_REORDERING); if (val && tp->reordering != val) @@ -544,8 +563,6 @@ bool tcp_peer_is_proven(struct request_sock *req, struct dst_entry *dst) return ret; } -static DEFINE_SEQLOCK(fastopen_seqlock); - void tcp_fastopen_cache_get(struct sock *sk, u16 *mss, struct tcp_fastopen_cookie *cookie) { @@ -652,7 +669,7 @@ static int tcp_metrics_fill_info(struct sk_buff *msg, } if (nla_put_msecs(msg, TCP_METRICS_ATTR_AGE, - jiffies - tm->tcpm_stamp, + jiffies - READ_ONCE(tm->tcpm_stamp), TCP_METRICS_ATTR_PAD) < 0) goto nla_put_failure; @@ -663,7 +680,7 @@ static int tcp_metrics_fill_info(struct sk_buff *msg, if (!nest) goto nla_put_failure; for (i = 0; i < TCP_METRIC_MAX_KERNEL + 1; i++) { - u32 val = tm->tcpm_vals[i]; + u32 val = tcp_metric_get(tm, i); if (!val) continue; @@ -895,7 +912,7 @@ static void tcp_metrics_flush_all(struct net *net) match = net ? net_eq(tm_net(tm), net) : !refcount_read(&tm_net(tm)->count); if (match) { - *pp = tm->tcpm_next; + rcu_assign_pointer(*pp, tm->tcpm_next); kfree_rcu(tm, rcu_head); } else { pp = &tm->tcpm_next; @@ -936,7 +953,7 @@ static int tcp_metrics_nl_cmd_del(struct sk_buff *skb, struct genl_info *info) if (addr_same(&tm->tcpm_daddr, &daddr) && (!src || addr_same(&tm->tcpm_saddr, &saddr)) && net_eq(tm_net(tm), net)) { - *pp = tm->tcpm_next; + rcu_assign_pointer(*pp, tm->tcpm_next); kfree_rcu(tm, rcu_head); found = true; } else { diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c index 0fc238d79b03..bae0199a943b 100644 --- a/net/ipv4/tcp_minisocks.c +++ b/net/ipv4/tcp_minisocks.c @@ -582,6 +582,9 @@ EXPORT_SYMBOL(tcp_create_openreq_child); * validation and inside tcp_v4_reqsk_send_ack(). Can we do better? * * We don't need to initialize tmp_opt.sack_ok as we don't use the results + * + * Note: If @fastopen is true, this can be called from process context. + * Otherwise, this is from BH context. */ struct sock *tcp_check_req(struct sock *sk, struct sk_buff *skb, @@ -734,7 +737,7 @@ struct sock *tcp_check_req(struct sock *sk, struct sk_buff *skb, &tcp_rsk(req)->last_oow_ack_time)) req->rsk_ops->send_ack(sk, skb, req); if (paws_reject) - __NET_INC_STATS(sock_net(sk), LINUX_MIB_PAWSESTABREJECTED); + NET_INC_STATS(sock_net(sk), LINUX_MIB_PAWSESTABREJECTED); return NULL; } @@ -753,7 +756,7 @@ struct sock *tcp_check_req(struct sock *sk, struct sk_buff *skb, * "fourth, check the SYN bit" */ if (flg & (TCP_FLAG_RST|TCP_FLAG_SYN)) { - __TCP_INC_STATS(sock_net(sk), TCP_MIB_ATTEMPTFAILS); + TCP_INC_STATS(sock_net(sk), TCP_MIB_ATTEMPTFAILS); goto embryonic_reset; } diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 8962864223b4..670804d4c169 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -164,8 +164,7 @@ static void tcp_event_data_sent(struct tcp_sock *tp, } /* Account for an ACK we sent. */ -static inline void tcp_event_ack_sent(struct sock *sk, unsigned int pkts, - u32 rcv_nxt) +static inline void tcp_event_ack_sent(struct sock *sk, u32 rcv_nxt) { struct tcp_sock *tp = tcp_sk(sk); @@ -179,7 +178,7 @@ static inline void tcp_event_ack_sent(struct sock *sk, unsigned int pkts, if (unlikely(rcv_nxt != tp->rcv_nxt)) return; /* Special ACK sent by DCTCP to reflect ECN */ - tcp_dec_quickack_mode(sk, pkts); + tcp_dec_quickack_mode(sk); inet_csk_clear_xmit_timer(sk, ICSK_TIME_DACK); } @@ -1091,7 +1090,7 @@ static int __tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, skb_set_hash_from_sk(skb, sk); refcount_add(skb->truesize, &sk->sk_wmem_alloc); - skb_set_dst_pending_confirm(skb, sk->sk_dst_pending_confirm); + skb_set_dst_pending_confirm(skb, READ_ONCE(sk->sk_dst_pending_confirm)); /* Build TCP header and checksum it. */ th = (struct tcphdr *)skb->data; @@ -1139,7 +1138,7 @@ static int __tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, icsk->icsk_af_ops->send_check(sk, skb); if (likely(tcb->tcp_flags & TCPHDR_ACK)) - tcp_event_ack_sent(sk, tcp_skb_pcount(skb), rcv_nxt); + tcp_event_ack_sent(sk, rcv_nxt); if (skb->len != tcp_header_size) { tcp_event_data_sent(tp, sk); @@ -2221,6 +2220,18 @@ static int tcp_mtu_probe(struct sock *sk) return -1; } +static bool tcp_rtx_queue_empty_or_single_skb(const struct sock *sk) +{ + const struct rb_node *node = sk->tcp_rtx_queue.rb_node; + + /* No skb in the rtx queue. */ + if (!node) + return true; + + /* Only one skb in rtx queue. */ + return !node->rb_left && !node->rb_right; +} + /* TCP Small Queues : * Control number of packets in qdisc/devices to two packets / or ~1 ms. * (These limits are doubled for retransmits) @@ -2243,12 +2254,12 @@ static bool tcp_small_queue_check(struct sock *sk, const struct sk_buff *skb, limit <<= factor; if (refcount_read(&sk->sk_wmem_alloc) > limit) { - /* Always send skb if rtx queue is empty. + /* Always send skb if rtx queue is empty or has one skb. * No need to wait for TX completion to call us back, * after softirq/tasklet schedule. * This helps when TX completions are delayed too much. */ - if (tcp_rtx_queue_empty(sk)) + if (tcp_rtx_queue_empty_or_single_skb(sk)) return false; set_bit(TSQ_THROTTLED, &sk->sk_tsq_flags); @@ -2450,7 +2461,7 @@ bool tcp_schedule_loss_probe(struct sock *sk, bool advancing_rto) { struct inet_connection_sock *icsk = inet_csk(sk); struct tcp_sock *tp = tcp_sk(sk); - u32 timeout, rto_delta_us; + u32 timeout, timeout_us, rto_delta_us; int early_retrans; /* Don't do any loss probe on a Fast Open connection before 3WHS @@ -2474,11 +2485,12 @@ bool tcp_schedule_loss_probe(struct sock *sk, bool advancing_rto) * sample is available then probe after TCP_TIMEOUT_INIT. */ if (tp->srtt_us) { - timeout = usecs_to_jiffies(tp->srtt_us >> 2); + timeout_us = tp->srtt_us >> 2; if (tp->packets_out == 1) - timeout += TCP_RTO_MIN; + timeout_us += tcp_rto_min_us(sk); else - timeout += TCP_TIMEOUT_MIN; + timeout_us += TCP_TIMEOUT_MIN_US; + timeout = usecs_to_jiffies(timeout_us); } else { timeout = TCP_TIMEOUT_INIT; } @@ -2878,7 +2890,13 @@ int __tcp_retransmit_skb(struct sock *sk, struct sk_buff *skb, int segs) if (skb_still_in_host_queue(sk, skb)) return -EBUSY; +start: if (before(TCP_SKB_CB(skb)->seq, tp->snd_una)) { + if (unlikely(TCP_SKB_CB(skb)->tcp_flags & TCPHDR_SYN)) { + TCP_SKB_CB(skb)->tcp_flags &= ~TCPHDR_SYN; + TCP_SKB_CB(skb)->seq++; + goto start; + } if (unlikely(before(TCP_SKB_CB(skb)->end_seq, tp->snd_una))) { WARN_ON_ONCE(1); return -EINVAL; @@ -3307,7 +3325,7 @@ struct sk_buff *tcp_make_synack(const struct sock *sk, struct dst_entry *dst, th->window = htons(min(req->rsk_rcv_wnd, 65535U)); tcp_options_write((__be32 *)(th + 1), NULL, &opts); th->doff = (tcp_header_size >> 2); - __TCP_INC_STATS(sock_net(sk), TCP_MIB_OUTSEGS); + TCP_INC_STATS(sock_net(sk), TCP_MIB_OUTSEGS); #ifdef CONFIG_TCP_MD5SIG /* Okay, we have all we need - do the md5 hash if needed */ diff --git a/net/ipv4/tcp_recovery.c b/net/ipv4/tcp_recovery.c index 61969bb9395c..844ff390f726 100644 --- a/net/ipv4/tcp_recovery.c +++ b/net/ipv4/tcp_recovery.c @@ -122,7 +122,7 @@ bool tcp_rack_mark_lost(struct sock *sk) tp->rack.advanced = 0; tcp_rack_detect_loss(sk, &timeout); if (timeout) { - timeout = usecs_to_jiffies(timeout) + TCP_TIMEOUT_MIN; + timeout = usecs_to_jiffies(timeout + TCP_TIMEOUT_MIN_US); inet_csk_reset_xmit_timer(sk, ICSK_TIME_REO_TIMEOUT, timeout, inet_csk(sk)->icsk_rto); } diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c index 84069db0423a..d8d28ba169b4 100644 --- a/net/ipv4/tcp_timer.c +++ b/net/ipv4/tcp_timer.c @@ -549,7 +549,9 @@ out_reset_timer: tcp_stream_is_thin(tp) && icsk->icsk_retransmits <= TCP_THIN_LINEAR_RETRIES) { icsk->icsk_backoff = 0; - icsk->icsk_rto = min(__tcp_set_rto(tp), TCP_RTO_MAX); + icsk->icsk_rto = clamp(__tcp_set_rto(tp), + tcp_rto_min(sk), + TCP_RTO_MAX); } else { /* Use normal (exponential) backoff */ icsk->icsk_rto = min(icsk->icsk_rto << 1, TCP_RTO_MAX); diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index dd3fa0ff84e7..a6048cc7fc35 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -1458,7 +1458,7 @@ drop: } EXPORT_SYMBOL_GPL(__udp_enqueue_schedule_skb); -void udp_destruct_sock(struct sock *sk) +void udp_destruct_common(struct sock *sk) { /* reclaim completely the forward allocated memory */ struct udp_sock *up = udp_sk(sk); @@ -1471,10 +1471,14 @@ void udp_destruct_sock(struct sock *sk) kfree_skb(skb); } udp_rmem_release(sk, total, 0, true); +} +EXPORT_SYMBOL_GPL(udp_destruct_common); +static void udp_destruct_sock(struct sock *sk) +{ + udp_destruct_common(sk); inet_sock_destruct(sk); } -EXPORT_SYMBOL_GPL(udp_destruct_sock); int udp_init_sock(struct sock *sk) { @@ -1482,7 +1486,6 @@ int udp_init_sock(struct sock *sk) sk->sk_destruct = udp_destruct_sock; return 0; } -EXPORT_SYMBOL_GPL(udp_init_sock); void skb_consume_udp(struct sock *sk, struct sk_buff *skb, int len) { diff --git a/net/ipv4/udp_tunnel.c b/net/ipv4/udp_tunnel.c index 6539ff15e9a3..d03d74388870 100644 --- a/net/ipv4/udp_tunnel.c +++ b/net/ipv4/udp_tunnel.c @@ -186,6 +186,7 @@ EXPORT_SYMBOL_GPL(udp_tunnel_xmit_skb); void udp_tunnel_sock_release(struct socket *sock) { rcu_assign_sk_user_data(sock->sk, NULL); + synchronize_rcu(); kernel_sock_shutdown(sock, SHUT_RDWR); sock_release(sock); } diff --git a/net/ipv4/udplite.c b/net/ipv4/udplite.c index 8545457752fb..27173549b000 100644 --- a/net/ipv4/udplite.c +++ b/net/ipv4/udplite.c @@ -20,6 +20,14 @@ struct udp_table udplite_table __read_mostly; EXPORT_SYMBOL(udplite_table); +/* Designate sk as UDP-Lite socket */ +static int udplite_sk_init(struct sock *sk) +{ + udp_init_sock(sk); + udp_sk(sk)->pcflag = UDPLITE_BIT; + return 0; +} + static int udplite_rcv(struct sk_buff *skb) { return __udp4_lib_rcv(skb, &udplite_table, IPPROTO_UDPLITE); @@ -56,6 +64,8 @@ struct proto udplite_prot = { .get_port = udp_v4_get_port, .memory_allocated = &udp_memory_allocated, .sysctl_mem = sysctl_udp_mem, + .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 udp_sock), .h.udp_table = &udplite_table, #ifdef CONFIG_COMPAT |