diff options
Diffstat (limited to 'net/ipv4')
-rw-r--r-- | net/ipv4/ah4.c | 2 | ||||
-rw-r--r-- | net/ipv4/esp4.c | 2 | ||||
-rw-r--r-- | net/ipv4/fib_semantics.c | 4 | ||||
-rw-r--r-- | net/ipv4/inet_fragment.c | 2 | ||||
-rw-r--r-- | net/ipv4/inet_timewait_sock.c | 3 | ||||
-rw-r--r-- | net/ipv4/ip_gre.c | 58 | ||||
-rw-r--r-- | net/ipv4/netfilter/arp_tables.c | 22 | ||||
-rw-r--r-- | net/ipv4/netfilter/ip_tables.c | 23 | ||||
-rw-r--r-- | net/ipv4/netfilter/nf_conntrack_l3proto_ipv4.c | 2 | ||||
-rw-r--r-- | net/ipv4/netfilter/nf_conntrack_proto_icmp.c | 28 | ||||
-rw-r--r-- | net/ipv4/proc.c | 1 | ||||
-rw-r--r-- | net/ipv4/route.c | 16 | ||||
-rw-r--r-- | net/ipv4/sysctl_net_ipv4.c | 51 | ||||
-rw-r--r-- | net/ipv4/tcp.c | 24 | ||||
-rw-r--r-- | net/ipv4/tcp_cong.c | 76 | ||||
-rw-r--r-- | net/ipv4/tcp_input.c | 236 | ||||
-rw-r--r-- | net/ipv4/tcp_ipv4.c | 22 | ||||
-rw-r--r-- | net/ipv4/tcp_metrics.c | 4 | ||||
-rw-r--r-- | net/ipv4/tcp_minisocks.c | 6 | ||||
-rw-r--r-- | net/ipv4/tcp_offload.c | 12 | ||||
-rw-r--r-- | net/ipv4/tcp_output.c | 38 | ||||
-rw-r--r-- | net/ipv4/udp.c | 4 |
22 files changed, 329 insertions, 307 deletions
diff --git a/net/ipv4/ah4.c b/net/ipv4/ah4.c index 37db44f60718..4dd95cdd8070 100644 --- a/net/ipv4/ah4.c +++ b/net/ipv4/ah4.c @@ -240,7 +240,7 @@ static int ah_output(struct xfrm_state *x, struct sk_buff *skb) if (err == -EINPROGRESS) goto out; - if (err == -EBUSY) + if (err == -ENOSPC) err = NET_XMIT_DROP; goto out_free; } diff --git a/net/ipv4/esp4.c b/net/ipv4/esp4.c index b00e4a43b4dc..d57aa64fa7c7 100644 --- a/net/ipv4/esp4.c +++ b/net/ipv4/esp4.c @@ -432,7 +432,7 @@ int esp_output_tail(struct xfrm_state *x, struct sk_buff *skb, struct esp_info * case -EINPROGRESS: goto error; - case -EBUSY: + case -ENOSPC: err = NET_XMIT_DROP; break; diff --git a/net/ipv4/fib_semantics.c b/net/ipv4/fib_semantics.c index 589caaa90613..f04d944f8abe 100644 --- a/net/ipv4/fib_semantics.c +++ b/net/ipv4/fib_semantics.c @@ -710,7 +710,7 @@ bool fib_metrics_match(struct fib_config *cfg, struct fib_info *fi) bool ecn_ca = false; nla_strlcpy(tmp, nla, sizeof(tmp)); - val = tcp_ca_get_key_by_name(tmp, &ecn_ca); + val = tcp_ca_get_key_by_name(fi->fib_net, tmp, &ecn_ca); } else { val = nla_get_u32(nla); } @@ -1030,7 +1030,7 @@ fib_convert_metrics(struct fib_info *fi, const struct fib_config *cfg) char tmp[TCP_CA_NAME_MAX]; nla_strlcpy(tmp, nla, sizeof(tmp)); - val = tcp_ca_get_key_by_name(tmp, &ecn_ca); + val = tcp_ca_get_key_by_name(fi->fib_net, tmp, &ecn_ca); if (val == TCP_CA_UNSPEC) return -EINVAL; } else { diff --git a/net/ipv4/inet_fragment.c b/net/ipv4/inet_fragment.c index 7f3ef5c287a1..26a3d0315728 100644 --- a/net/ipv4/inet_fragment.c +++ b/net/ipv4/inet_fragment.c @@ -164,7 +164,7 @@ static void inet_frag_worker(struct work_struct *work) local_bh_disable(); - for (i = ACCESS_ONCE(f->next_bucket); budget; --budget) { + for (i = READ_ONCE(f->next_bucket); budget; --budget) { evicted += inet_evict_bucket(f, &f->hash[i]); i = (i + 1) & (INETFRAGS_HASHSZ - 1); if (evicted > INETFRAGS_EVICT_MAX) diff --git a/net/ipv4/inet_timewait_sock.c b/net/ipv4/inet_timewait_sock.c index a4bab81f1462..c690cd0d9b3f 100644 --- a/net/ipv4/inet_timewait_sock.c +++ b/net/ipv4/inet_timewait_sock.c @@ -9,7 +9,6 @@ */ #include <linux/kernel.h> -#include <linux/kmemcheck.h> #include <linux/slab.h> #include <linux/module.h> #include <net/inet_hashtables.h> @@ -167,8 +166,6 @@ struct inet_timewait_sock *inet_twsk_alloc(const struct sock *sk, if (tw) { const struct inet_sock *inet = inet_sk(sk); - kmemcheck_annotate_bitfield(tw, flags); - tw->tw_dr = dr; /* Give us an identity. */ tw->tw_daddr = inet->inet_daddr; diff --git a/net/ipv4/ip_gre.c b/net/ipv4/ip_gre.c index c105a315b1a3..bb6239169b1a 100644 --- a/net/ipv4/ip_gre.c +++ b/net/ipv4/ip_gre.c @@ -773,20 +773,46 @@ free_skb: return NETDEV_TX_OK; } +static void ipgre_link_update(struct net_device *dev, bool set_mtu) +{ + struct ip_tunnel *tunnel = netdev_priv(dev); + int len; + + len = tunnel->tun_hlen; + tunnel->tun_hlen = gre_calc_hlen(tunnel->parms.o_flags); + len = tunnel->tun_hlen - len; + tunnel->hlen = tunnel->hlen + len; + + dev->needed_headroom = dev->needed_headroom + len; + if (set_mtu) + dev->mtu = max_t(int, dev->mtu - len, 68); + + if (!(tunnel->parms.o_flags & TUNNEL_SEQ)) { + if (!(tunnel->parms.o_flags & TUNNEL_CSUM) || + tunnel->encap.type == TUNNEL_ENCAP_NONE) { + dev->features |= NETIF_F_GSO_SOFTWARE; + dev->hw_features |= NETIF_F_GSO_SOFTWARE; + } + dev->features |= NETIF_F_LLTX; + } +} + static int ipgre_tunnel_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd) { - int err; struct ip_tunnel_parm p; + int err; if (copy_from_user(&p, ifr->ifr_ifru.ifru_data, sizeof(p))) return -EFAULT; + if (cmd == SIOCADDTUNNEL || cmd == SIOCCHGTUNNEL) { if (p.iph.version != 4 || p.iph.protocol != IPPROTO_GRE || - p.iph.ihl != 5 || (p.iph.frag_off&htons(~IP_DF)) || - ((p.i_flags|p.o_flags)&(GRE_VERSION|GRE_ROUTING))) + p.iph.ihl != 5 || (p.iph.frag_off & htons(~IP_DF)) || + ((p.i_flags | p.o_flags) & (GRE_VERSION | GRE_ROUTING))) return -EINVAL; } + p.i_flags = gre_flags_to_tnl_flags(p.i_flags); p.o_flags = gre_flags_to_tnl_flags(p.o_flags); @@ -794,11 +820,22 @@ static int ipgre_tunnel_ioctl(struct net_device *dev, if (err) return err; + if (cmd == SIOCCHGTUNNEL) { + struct ip_tunnel *t = netdev_priv(dev); + + t->parms.i_flags = p.i_flags; + t->parms.o_flags = p.o_flags; + + if (strcmp(dev->rtnl_link_ops->kind, "erspan")) + ipgre_link_update(dev, true); + } + p.i_flags = gre_tnl_flags_to_gre_flags(p.i_flags); p.o_flags = gre_tnl_flags_to_gre_flags(p.o_flags); if (copy_to_user(ifr->ifr_ifru.ifru_data, &p, sizeof(p))) return -EFAULT; + return 0; } @@ -1307,9 +1344,9 @@ static int ipgre_changelink(struct net_device *dev, struct nlattr *tb[], struct netlink_ext_ack *extack) { struct ip_tunnel *t = netdev_priv(dev); - struct ip_tunnel_parm p; struct ip_tunnel_encap ipencap; __u32 fwmark = t->fwmark; + struct ip_tunnel_parm p; int err; if (ipgre_netlink_encap_parms(data, &ipencap)) { @@ -1322,7 +1359,18 @@ static int ipgre_changelink(struct net_device *dev, struct nlattr *tb[], err = ipgre_netlink_parms(dev, data, tb, &p, &fwmark); if (err < 0) return err; - return ip_tunnel_changelink(dev, tb, &p, fwmark); + + err = ip_tunnel_changelink(dev, tb, &p, fwmark); + if (err < 0) + return err; + + t->parms.i_flags = p.i_flags; + t->parms.o_flags = p.o_flags; + + if (strcmp(dev->rtnl_link_ops->kind, "erspan")) + ipgre_link_update(dev, !tb[IFLA_MTU]); + + return 0; } static size_t ipgre_get_size(const struct net_device *dev) diff --git a/net/ipv4/netfilter/arp_tables.c b/net/ipv4/netfilter/arp_tables.c index 9e2770fd00be..f88221aebc9d 100644 --- a/net/ipv4/netfilter/arp_tables.c +++ b/net/ipv4/netfilter/arp_tables.c @@ -634,6 +634,25 @@ static void get_counters(const struct xt_table_info *t, } } +static void get_old_counters(const struct xt_table_info *t, + struct xt_counters counters[]) +{ + struct arpt_entry *iter; + unsigned int cpu, i; + + for_each_possible_cpu(cpu) { + i = 0; + xt_entry_foreach(iter, t->entries, t->size) { + struct xt_counters *tmp; + + tmp = xt_get_per_cpu_counter(&iter->counters, cpu); + ADD_COUNTER(counters[i], tmp->bcnt, tmp->pcnt); + ++i; + } + cond_resched(); + } +} + static struct xt_counters *alloc_counters(const struct xt_table *table) { unsigned int countersize; @@ -910,8 +929,7 @@ static int __do_replace(struct net *net, const char *name, (newinfo->number <= oldinfo->initial_entries)) module_put(t->me); - /* Get the old counters, and synchronize with replace */ - get_counters(oldinfo, counters); + get_old_counters(oldinfo, counters); /* Decrease module usage counts and free resource */ loc_cpu_old_entry = oldinfo->entries; diff --git a/net/ipv4/netfilter/ip_tables.c b/net/ipv4/netfilter/ip_tables.c index 39286e543ee6..4cbe5e80f3bf 100644 --- a/net/ipv4/netfilter/ip_tables.c +++ b/net/ipv4/netfilter/ip_tables.c @@ -781,6 +781,26 @@ get_counters(const struct xt_table_info *t, } } +static void get_old_counters(const struct xt_table_info *t, + struct xt_counters counters[]) +{ + struct ipt_entry *iter; + unsigned int cpu, i; + + for_each_possible_cpu(cpu) { + i = 0; + xt_entry_foreach(iter, t->entries, t->size) { + const struct xt_counters *tmp; + + tmp = xt_get_per_cpu_counter(&iter->counters, cpu); + ADD_COUNTER(counters[i], tmp->bcnt, tmp->pcnt); + ++i; /* macro does multi eval of i */ + } + + cond_resched(); + } +} + static struct xt_counters *alloc_counters(const struct xt_table *table) { unsigned int countersize; @@ -1070,8 +1090,7 @@ __do_replace(struct net *net, const char *name, unsigned int valid_hooks, (newinfo->number <= oldinfo->initial_entries)) module_put(t->me); - /* Get the old counters, and synchronize with replace */ - get_counters(oldinfo, counters); + get_old_counters(oldinfo, counters); /* Decrease module usage counts and free resource */ xt_entry_foreach(iter, oldinfo->entries, oldinfo->size) diff --git a/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4.c b/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4.c index fe374da4bc13..89af9d88ca21 100644 --- a/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4.c +++ b/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4.c @@ -344,7 +344,7 @@ static void ipv4_hooks_unregister(struct net *net) mutex_unlock(®ister_ipv4_hooks); } -struct nf_conntrack_l3proto nf_conntrack_l3proto_ipv4 __read_mostly = { +const struct nf_conntrack_l3proto nf_conntrack_l3proto_ipv4 = { .l3proto = PF_INET, .pkt_to_tuple = ipv4_pkt_to_tuple, .invert_tuple = ipv4_invert_tuple, diff --git a/net/ipv4/netfilter/nf_conntrack_proto_icmp.c b/net/ipv4/netfilter/nf_conntrack_proto_icmp.c index a046c298413a..1849fedd9b81 100644 --- a/net/ipv4/netfilter/nf_conntrack_proto_icmp.c +++ b/net/ipv4/netfilter/nf_conntrack_proto_icmp.c @@ -81,7 +81,6 @@ static int icmp_packet(struct nf_conn *ct, const struct sk_buff *skb, unsigned int dataoff, enum ip_conntrack_info ctinfo, - u_int8_t pf, unsigned int *timeout) { /* Do not immediately delete the connection after the first @@ -165,6 +164,12 @@ icmp_error_message(struct net *net, struct nf_conn *tmpl, struct sk_buff *skb, return NF_ACCEPT; } +static void icmp_error_log(const struct sk_buff *skb, struct net *net, + u8 pf, const char *msg) +{ + nf_l4proto_log_invalid(skb, net, pf, IPPROTO_ICMP, "%s", msg); +} + /* Small and modified version of icmp_rcv */ static int icmp_error(struct net *net, struct nf_conn *tmpl, @@ -177,18 +182,14 @@ icmp_error(struct net *net, struct nf_conn *tmpl, /* Not enough header? */ icmph = skb_header_pointer(skb, ip_hdrlen(skb), sizeof(_ih), &_ih); if (icmph == NULL) { - if (LOG_INVALID(net, IPPROTO_ICMP)) - nf_log_packet(net, PF_INET, 0, skb, NULL, NULL, - NULL, "nf_ct_icmp: short packet "); + icmp_error_log(skb, net, pf, "short packet"); return -NF_ACCEPT; } /* See ip_conntrack_proto_tcp.c */ if (net->ct.sysctl_checksum && hooknum == NF_INET_PRE_ROUTING && nf_ip_checksum(skb, hooknum, dataoff, 0)) { - if (LOG_INVALID(net, IPPROTO_ICMP)) - nf_log_packet(net, PF_INET, 0, skb, NULL, NULL, NULL, - "nf_ct_icmp: bad HW ICMP checksum "); + icmp_error_log(skb, net, pf, "bad hw icmp checksum"); return -NF_ACCEPT; } @@ -199,9 +200,7 @@ icmp_error(struct net *net, struct nf_conn *tmpl, * discarded. */ if (icmph->type > NR_ICMP_TYPES) { - if (LOG_INVALID(net, IPPROTO_ICMP)) - nf_log_packet(net, PF_INET, 0, skb, NULL, NULL, NULL, - "nf_ct_icmp: invalid ICMP type "); + icmp_error_log(skb, net, pf, "invalid icmp type"); return -NF_ACCEPT; } @@ -259,9 +258,14 @@ static int icmp_nlattr_to_tuple(struct nlattr *tb[], return 0; } -static int icmp_nlattr_tuple_size(void) +static unsigned int icmp_nlattr_tuple_size(void) { - return nla_policy_len(icmp_nla_policy, CTA_PROTO_MAX + 1); + static unsigned int size __read_mostly; + + if (!size) + size = nla_policy_len(icmp_nla_policy, CTA_PROTO_MAX + 1); + + return size; } #endif diff --git a/net/ipv4/proc.c b/net/ipv4/proc.c index 127153f1ed8a..9f37c4727861 100644 --- a/net/ipv4/proc.c +++ b/net/ipv4/proc.c @@ -212,7 +212,6 @@ static const struct snmp_mib snmp4_net_list[] = { SNMP_MIB_ITEM("TCPRenoRecovery", LINUX_MIB_TCPRENORECOVERY), SNMP_MIB_ITEM("TCPSackRecovery", LINUX_MIB_TCPSACKRECOVERY), SNMP_MIB_ITEM("TCPSACKReneging", LINUX_MIB_TCPSACKRENEGING), - SNMP_MIB_ITEM("TCPFACKReorder", LINUX_MIB_TCPFACKREORDER), SNMP_MIB_ITEM("TCPSACKReorder", LINUX_MIB_TCPSACKREORDER), SNMP_MIB_ITEM("TCPRenoReorder", LINUX_MIB_TCPRENOREORDER), SNMP_MIB_ITEM("TCPTSReorder", LINUX_MIB_TCPTSREORDER), diff --git a/net/ipv4/route.c b/net/ipv4/route.c index bc40bd411196..43b69af242e1 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -495,7 +495,7 @@ u32 ip_idents_reserve(u32 hash, int segs) { u32 *p_tstamp = ip_tstamps + hash % IP_IDENTS_SZ; atomic_t *p_id = ip_idents + hash % IP_IDENTS_SZ; - u32 old = ACCESS_ONCE(*p_tstamp); + u32 old = READ_ONCE(*p_tstamp); u32 now = (u32)jiffies; u32 new, delta = 0; @@ -651,9 +651,12 @@ static void update_or_create_fnhe(struct fib_nh *nh, __be32 daddr, __be32 gw, struct fnhe_hash_bucket *hash; struct fib_nh_exception *fnhe; struct rtable *rt; + u32 genid, hval; unsigned int i; int depth; - u32 hval = fnhe_hashfun(daddr); + + genid = fnhe_genid(dev_net(nh->nh_dev)); + hval = fnhe_hashfun(daddr); spin_lock_bh(&fnhe_lock); @@ -676,12 +679,13 @@ static void update_or_create_fnhe(struct fib_nh *nh, __be32 daddr, __be32 gw, } if (fnhe) { + if (fnhe->fnhe_genid != genid) + fnhe->fnhe_genid = genid; if (gw) fnhe->fnhe_gw = gw; - if (pmtu) { + if (pmtu) fnhe->fnhe_pmtu = pmtu; - fnhe->fnhe_expires = max(1UL, expires); - } + fnhe->fnhe_expires = max(1UL, expires); /* Update all cached dsts too */ rt = rcu_dereference(fnhe->fnhe_rth_input); if (rt) @@ -700,7 +704,7 @@ static void update_or_create_fnhe(struct fib_nh *nh, __be32 daddr, __be32 gw, fnhe->fnhe_next = hash->chain; rcu_assign_pointer(hash->chain, fnhe); } - fnhe->fnhe_genid = fnhe_genid(dev_net(nh->nh_dev)); + fnhe->fnhe_genid = genid; fnhe->fnhe_daddr = daddr; fnhe->fnhe_gw = gw; fnhe->fnhe_pmtu = pmtu; diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c index a82b44038308..93e172118a94 100644 --- a/net/ipv4/sysctl_net_ipv4.c +++ b/net/ipv4/sysctl_net_ipv4.c @@ -201,6 +201,8 @@ static int ipv4_ping_group_range(struct ctl_table *table, int write, static int proc_tcp_congestion_control(struct ctl_table *ctl, int write, void __user *buffer, size_t *lenp, loff_t *ppos) { + struct net *net = container_of(ctl->data, struct net, + ipv4.tcp_congestion_control); char val[TCP_CA_NAME_MAX]; struct ctl_table tbl = { .data = val, @@ -208,11 +210,11 @@ static int proc_tcp_congestion_control(struct ctl_table *ctl, int write, }; int ret; - tcp_get_default_congestion_control(val); + tcp_get_default_congestion_control(net, val); ret = proc_dostring(&tbl, write, buffer, lenp, ppos); if (write && ret == 0) - ret = tcp_set_default_congestion_control(val); + ret = tcp_set_default_congestion_control(net, val); return ret; } @@ -441,34 +443,12 @@ static struct ctl_table ipv4_table[] = { .proc_handler = proc_doulongvec_minmax, }, { - .procname = "tcp_wmem", - .data = &sysctl_tcp_wmem, - .maxlen = sizeof(sysctl_tcp_wmem), - .mode = 0644, - .proc_handler = proc_dointvec_minmax, - .extra1 = &one, - }, - { - .procname = "tcp_rmem", - .data = &sysctl_tcp_rmem, - .maxlen = sizeof(sysctl_tcp_rmem), - .mode = 0644, - .proc_handler = proc_dointvec_minmax, - .extra1 = &one, - }, - { .procname = "tcp_low_latency", .data = &sysctl_tcp_low_latency, .maxlen = sizeof(int), .mode = 0644, .proc_handler = proc_dointvec }, - { - .procname = "tcp_congestion_control", - .mode = 0644, - .maxlen = TCP_CA_NAME_MAX, - .proc_handler = proc_tcp_congestion_control, - }, #ifdef CONFIG_NETLABEL { .procname = "cipso_cache_enable", @@ -780,6 +760,13 @@ static struct ctl_table ipv4_net_table[] = { }, #endif { + .procname = "tcp_congestion_control", + .data = &init_net.ipv4.tcp_congestion_control, + .mode = 0644, + .maxlen = TCP_CA_NAME_MAX, + .proc_handler = proc_tcp_congestion_control, + }, + { .procname = "tcp_keepalive_time", .data = &init_net.ipv4.sysctl_tcp_keepalive_time, .maxlen = sizeof(int), @@ -1164,6 +1151,22 @@ static struct ctl_table ipv4_net_table[] = { .extra1 = &zero, .extra2 = &thousand, }, + { + .procname = "tcp_wmem", + .data = &init_net.ipv4.sysctl_tcp_wmem, + .maxlen = sizeof(init_net.ipv4.sysctl_tcp_wmem), + .mode = 0644, + .proc_handler = proc_dointvec_minmax, + .extra1 = &one, + }, + { + .procname = "tcp_rmem", + .data = &init_net.ipv4.sysctl_tcp_rmem, + .maxlen = sizeof(init_net.ipv4.sysctl_tcp_rmem), + .mode = 0644, + .proc_handler = proc_dointvec_minmax, + .extra1 = &one, + }, { } }; diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index c4cb19ed4628..bf97317e6c97 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -289,12 +289,7 @@ struct percpu_counter tcp_orphan_count; EXPORT_SYMBOL_GPL(tcp_orphan_count); long sysctl_tcp_mem[3] __read_mostly; -int sysctl_tcp_wmem[3] __read_mostly; -int sysctl_tcp_rmem[3] __read_mostly; - EXPORT_SYMBOL(sysctl_tcp_mem); -EXPORT_SYMBOL(sysctl_tcp_rmem); -EXPORT_SYMBOL(sysctl_tcp_wmem); atomic_long_t tcp_memory_allocated; /* Current allocated memory. */ EXPORT_SYMBOL(tcp_memory_allocated); @@ -456,8 +451,8 @@ void tcp_init_sock(struct sock *sk) icsk->icsk_sync_mss = tcp_sync_mss; - sk->sk_sndbuf = sysctl_tcp_wmem[1]; - sk->sk_rcvbuf = sysctl_tcp_rmem[1]; + sk->sk_sndbuf = sock_net(sk)->ipv4.sysctl_tcp_wmem[1]; + sk->sk_rcvbuf = sock_net(sk)->ipv4.sysctl_tcp_rmem[1]; sk_sockets_allocated_inc(sk); } @@ -2514,8 +2509,6 @@ static int tcp_repair_options_est(struct sock *sk, return -EINVAL; tp->rx_opt.sack_ok |= TCP_SACK_SEEN; - if (sock_net(sk)->ipv4.sysctl_tcp_fack) - tcp_enable_fack(tp); break; case TCPOPT_TIMESTAMP: if (opt.opt_val != 0) @@ -2984,7 +2977,6 @@ void tcp_get_info(struct sock *sk, struct tcp_info *info) info->tcpi_lost = tp->lost_out; info->tcpi_retrans = tp->retrans_out; - info->tcpi_fackets = tp->fackets_out; now = tcp_jiffies32; info->tcpi_last_data_sent = jiffies_to_msecs(now - tp->lsndtime); @@ -3636,13 +3628,13 @@ void __init tcp_init(void) max_wshare = min(4UL*1024*1024, limit); max_rshare = min(6UL*1024*1024, limit); - sysctl_tcp_wmem[0] = SK_MEM_QUANTUM; - sysctl_tcp_wmem[1] = 16*1024; - sysctl_tcp_wmem[2] = max(64*1024, max_wshare); + init_net.ipv4.sysctl_tcp_wmem[0] = SK_MEM_QUANTUM; + init_net.ipv4.sysctl_tcp_wmem[1] = 16*1024; + init_net.ipv4.sysctl_tcp_wmem[2] = max(64*1024, max_wshare); - sysctl_tcp_rmem[0] = SK_MEM_QUANTUM; - sysctl_tcp_rmem[1] = 87380; - sysctl_tcp_rmem[2] = max(87380, max_rshare); + init_net.ipv4.sysctl_tcp_rmem[0] = SK_MEM_QUANTUM; + init_net.ipv4.sysctl_tcp_rmem[1] = 87380; + init_net.ipv4.sysctl_tcp_rmem[2] = max(87380, max_rshare); pr_info("Hash tables configured (established %u bind %u)\n", tcp_hashinfo.ehash_mask + 1, tcp_hashinfo.bhash_size); diff --git a/net/ipv4/tcp_cong.c b/net/ipv4/tcp_cong.c index 2f26124fd160..bc6c02f16243 100644 --- a/net/ipv4/tcp_cong.c +++ b/net/ipv4/tcp_cong.c @@ -33,9 +33,11 @@ static struct tcp_congestion_ops *tcp_ca_find(const char *name) } /* Must be called with rcu lock held */ -static const struct tcp_congestion_ops *__tcp_ca_find_autoload(const char *name) +static struct tcp_congestion_ops *tcp_ca_find_autoload(struct net *net, + const char *name) { - const struct tcp_congestion_ops *ca = tcp_ca_find(name); + struct tcp_congestion_ops *ca = tcp_ca_find(name); + #ifdef CONFIG_MODULES if (!ca && capable(CAP_NET_ADMIN)) { rcu_read_unlock(); @@ -115,7 +117,7 @@ void tcp_unregister_congestion_control(struct tcp_congestion_ops *ca) } EXPORT_SYMBOL_GPL(tcp_unregister_congestion_control); -u32 tcp_ca_get_key_by_name(const char *name, bool *ecn_ca) +u32 tcp_ca_get_key_by_name(struct net *net, const char *name, bool *ecn_ca) { const struct tcp_congestion_ops *ca; u32 key = TCP_CA_UNSPEC; @@ -123,7 +125,7 @@ u32 tcp_ca_get_key_by_name(const char *name, bool *ecn_ca) might_sleep(); rcu_read_lock(); - ca = __tcp_ca_find_autoload(name); + ca = tcp_ca_find_autoload(net, name); if (ca) { key = ca->key; *ecn_ca = ca->flags & TCP_CONG_NEEDS_ECN; @@ -153,23 +155,18 @@ EXPORT_SYMBOL_GPL(tcp_ca_get_name_by_key); /* Assign choice of congestion control. */ void tcp_assign_congestion_control(struct sock *sk) { + struct net *net = sock_net(sk); struct inet_connection_sock *icsk = inet_csk(sk); - struct tcp_congestion_ops *ca; + const struct tcp_congestion_ops *ca; rcu_read_lock(); - list_for_each_entry_rcu(ca, &tcp_cong_list, list) { - if (likely(try_module_get(ca->owner))) { - icsk->icsk_ca_ops = ca; - goto out; - } - /* Fallback to next available. The last really - * guaranteed fallback is Reno from this list. - */ - } -out: + ca = rcu_dereference(net->ipv4.tcp_congestion_control); + if (unlikely(!try_module_get(ca->owner))) + ca = &tcp_reno; + icsk->icsk_ca_ops = ca; rcu_read_unlock(); - memset(icsk->icsk_ca_priv, 0, sizeof(icsk->icsk_ca_priv)); + memset(icsk->icsk_ca_priv, 0, sizeof(icsk->icsk_ca_priv)); if (ca->flags & TCP_CONG_NEEDS_ECN) INET_ECN_xmit(sk); else @@ -214,29 +211,27 @@ void tcp_cleanup_congestion_control(struct sock *sk) } /* Used by sysctl to change default congestion control */ -int tcp_set_default_congestion_control(const char *name) +int tcp_set_default_congestion_control(struct net *net, const char *name) { struct tcp_congestion_ops *ca; - int ret = -ENOENT; - - spin_lock(&tcp_cong_list_lock); - ca = tcp_ca_find(name); -#ifdef CONFIG_MODULES - if (!ca && capable(CAP_NET_ADMIN)) { - spin_unlock(&tcp_cong_list_lock); + const struct tcp_congestion_ops *prev; + int ret; - request_module("tcp_%s", name); - spin_lock(&tcp_cong_list_lock); - ca = tcp_ca_find(name); - } -#endif + rcu_read_lock(); + ca = tcp_ca_find_autoload(net, name); + if (!ca) { + ret = -ENOENT; + } else if (!try_module_get(ca->owner)) { + ret = -EBUSY; + } else { + prev = xchg(&net->ipv4.tcp_congestion_control, ca); + if (prev) + module_put(prev->owner); - if (ca) { - ca->flags |= TCP_CONG_NON_RESTRICTED; /* default is always allowed */ - list_move(&ca->list, &tcp_cong_list); + ca->flags |= TCP_CONG_NON_RESTRICTED; ret = 0; } - spin_unlock(&tcp_cong_list_lock); + rcu_read_unlock(); return ret; } @@ -244,7 +239,8 @@ int tcp_set_default_congestion_control(const char *name) /* Set default value from kernel configuration at bootup */ static int __init tcp_congestion_default(void) { - return tcp_set_default_congestion_control(CONFIG_DEFAULT_TCP_CONG); + return tcp_set_default_congestion_control(&init_net, + CONFIG_DEFAULT_TCP_CONG); } late_initcall(tcp_congestion_default); @@ -264,14 +260,12 @@ void tcp_get_available_congestion_control(char *buf, size_t maxlen) } /* Get current default congestion control */ -void tcp_get_default_congestion_control(char *name) +void tcp_get_default_congestion_control(struct net *net, char *name) { - struct tcp_congestion_ops *ca; - /* We will always have reno... */ - BUG_ON(list_empty(&tcp_cong_list)); + const struct tcp_congestion_ops *ca; rcu_read_lock(); - ca = list_entry(tcp_cong_list.next, struct tcp_congestion_ops, list); + ca = rcu_dereference(net->ipv4.tcp_congestion_control); strncpy(name, ca->name, TCP_CA_NAME_MAX); rcu_read_unlock(); } @@ -351,12 +345,14 @@ int tcp_set_congestion_control(struct sock *sk, const char *name, bool load, boo if (!load) ca = tcp_ca_find(name); else - ca = __tcp_ca_find_autoload(name); + ca = tcp_ca_find_autoload(sock_net(sk), name); + /* No change asking for existing value */ if (ca == icsk->icsk_ca_ops) { icsk->icsk_ca_setsockopt = 1; goto out; } + if (!ca) { err = -ENOENT; } else if (!load) { diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 0ada8bfc2ebd..734cfc8ff76e 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -100,7 +100,7 @@ int sysctl_tcp_max_orphans __read_mostly = NR_FILE; #define FLAG_ACKED (FLAG_DATA_ACKED|FLAG_SYN_ACKED) #define FLAG_NOT_DUP (FLAG_DATA|FLAG_WIN_UPDATE|FLAG_ACKED) -#define FLAG_CA_ALERT (FLAG_DATA_SACKED|FLAG_ECE) +#define FLAG_CA_ALERT (FLAG_DATA_SACKED|FLAG_ECE|FLAG_DSACKING_ACK) #define FLAG_FORWARD_PROGRESS (FLAG_ACKED|FLAG_DATA_SACKED) #define TCP_REMNANT (TCP_FLAG_FIN|TCP_FLAG_URG|TCP_FLAG_SYN|TCP_FLAG_PSH) @@ -320,7 +320,7 @@ static void tcp_sndbuf_expand(struct sock *sk) sndmem *= nr_segs * per_mss; if (sk->sk_sndbuf < sndmem) - sk->sk_sndbuf = min(sndmem, sysctl_tcp_wmem[2]); + sk->sk_sndbuf = min(sndmem, sock_net(sk)->ipv4.sysctl_tcp_wmem[2]); } /* 2. Tuning advertised window (window_clamp, rcv_ssthresh) @@ -354,7 +354,7 @@ static int __tcp_grow_window(const struct sock *sk, const struct sk_buff *skb) struct tcp_sock *tp = tcp_sk(sk); /* Optimize this! */ int truesize = tcp_win_from_space(sk, skb->truesize) >> 1; - int window = tcp_win_from_space(sk, sysctl_tcp_rmem[2]) >> 1; + int window = tcp_win_from_space(sk, sock_net(sk)->ipv4.sysctl_tcp_rmem[2]) >> 1; while (tp->rcv_ssthresh <= window) { if (truesize <= skb->len) @@ -409,7 +409,7 @@ static void tcp_fixup_rcvbuf(struct sock *sk) rcvmem <<= 2; if (sk->sk_rcvbuf < rcvmem) - sk->sk_rcvbuf = min(rcvmem, sysctl_tcp_rmem[2]); + sk->sk_rcvbuf = min(rcvmem, sock_net(sk)->ipv4.sysctl_tcp_rmem[2]); } /* 4. Try to fixup all. It is made immediately after connection enters @@ -457,15 +457,16 @@ static void tcp_clamp_window(struct sock *sk) { struct tcp_sock *tp = tcp_sk(sk); struct inet_connection_sock *icsk = inet_csk(sk); + struct net *net = sock_net(sk); icsk->icsk_ack.quick = 0; - if (sk->sk_rcvbuf < sysctl_tcp_rmem[2] && + if (sk->sk_rcvbuf < net->ipv4.sysctl_tcp_rmem[2] && !(sk->sk_userlocks & SOCK_RCVBUF_LOCK) && !tcp_under_memory_pressure(sk) && sk_memory_allocated(sk) < sk_prot_mem_limits(sk, 0)) { sk->sk_rcvbuf = min(atomic_read(&sk->sk_rmem_alloc), - sysctl_tcp_rmem[2]); + net->ipv4.sysctl_tcp_rmem[2]); } if (atomic_read(&sk->sk_rmem_alloc) > sk->sk_rcvbuf) tp->rcv_ssthresh = min(tp->window_clamp, 2U * tp->advmss); @@ -623,7 +624,8 @@ void tcp_rcv_space_adjust(struct sock *sk) while (tcp_win_from_space(sk, rcvmem) < tp->advmss) rcvmem += 128; - rcvbuf = min(rcvwin / tp->advmss * rcvmem, sysctl_tcp_rmem[2]); + rcvbuf = min(rcvwin / tp->advmss * rcvmem, + sock_net(sk)->ipv4.sysctl_tcp_rmem[2]); if (rcvbuf > sk->sk_rcvbuf) { sk->sk_rcvbuf = rcvbuf; @@ -793,12 +795,12 @@ static void tcp_update_pacing_rate(struct sock *sk) if (likely(tp->srtt_us)) do_div(rate, tp->srtt_us); - /* ACCESS_ONCE() is needed because sch_fq fetches sk_pacing_rate + /* WRITE_ONCE() is needed because sch_fq fetches sk_pacing_rate * without any lock. We want to make sure compiler wont store * intermediate values in this location. */ - ACCESS_ONCE(sk->sk_pacing_rate) = min_t(u64, rate, - sk->sk_max_pacing_rate); + WRITE_ONCE(sk->sk_pacing_rate, min_t(u64, rate, + sk->sk_max_pacing_rate)); } /* Calculate rto without backoff. This is the second half of Van Jacobson's @@ -840,18 +842,6 @@ __u32 tcp_init_cwnd(const struct tcp_sock *tp, const struct dst_entry *dst) return min_t(__u32, cwnd, tp->snd_cwnd_clamp); } -/* - * Packet counting of FACK is based on in-order assumptions, therefore TCP - * disables it when reordering is detected - */ -void tcp_disable_fack(struct tcp_sock *tp) -{ - /* RFC3517 uses different metric in lost marker => reset on change */ - if (tcp_is_fack(tp)) - tp->lost_skb_hint = NULL; - tp->rx_opt.sack_ok &= ~TCP_FACK_ENABLED; -} - /* Take a notice that peer is sending D-SACKs */ static void tcp_dsack_seen(struct tcp_sock *tp) { @@ -859,42 +849,39 @@ static void tcp_dsack_seen(struct tcp_sock *tp) tp->rack.dsack_seen = 1; } -static void tcp_update_reordering(struct sock *sk, const int metric, - const int ts) +/* It's reordering when higher sequence was delivered (i.e. sacked) before + * some lower never-retransmitted sequence ("low_seq"). The maximum reordering + * distance is approximated in full-mss packet distance ("reordering"). + */ +static void tcp_check_sack_reordering(struct sock *sk, const u32 low_seq, + const int ts) { struct tcp_sock *tp = tcp_sk(sk); - int mib_idx; + const u32 mss = tp->mss_cache; + u32 fack, metric; - if (WARN_ON_ONCE(metric < 0)) + fack = tcp_highest_sack_seq(tp); + if (!before(low_seq, fack)) return; - if (metric > tp->reordering) { - tp->reordering = min(sock_net(sk)->ipv4.sysctl_tcp_max_reordering, metric); - + metric = fack - low_seq; + if ((metric > tp->reordering * mss) && mss) { #if FASTRETRANS_DEBUG > 1 pr_debug("Disorder%d %d %u f%u s%u rr%d\n", tp->rx_opt.sack_ok, inet_csk(sk)->icsk_ca_state, tp->reordering, - tp->fackets_out, + 0, tp->sacked_out, tp->undo_marker ? tp->undo_retrans : 0); #endif - tcp_disable_fack(tp); + tp->reordering = min_t(u32, (metric + mss - 1) / mss, + sock_net(sk)->ipv4.sysctl_tcp_max_reordering); } tp->rack.reord = 1; - /* This exciting event is worth to be remembered. 8) */ - if (ts) - mib_idx = LINUX_MIB_TCPTSREORDER; - else if (tcp_is_reno(tp)) - mib_idx = LINUX_MIB_TCPRENOREORDER; - else if (tcp_is_fack(tp)) - mib_idx = LINUX_MIB_TCPFACKREORDER; - else - mib_idx = LINUX_MIB_TCPSACKREORDER; - - NET_INC_STATS(sock_net(sk), mib_idx); + NET_INC_STATS(sock_net(sk), + ts ? LINUX_MIB_TCPTSREORDER : LINUX_MIB_TCPSACKREORDER); } /* This must be called before lost_out is incremented */ @@ -968,7 +955,6 @@ void tcp_skb_mark_lost_uncond_verify(struct tcp_sock *tp, struct sk_buff *skb) * 3. Loss detection event of two flavors: * A. Scoreboard estimator decided the packet is lost. * A'. Reno "three dupacks" marks head of queue lost. - * A''. Its FACK modification, head until snd.fack is lost. * B. SACK arrives sacking SND.NXT at the moment, when the * segment was retransmitted. * 4. D-SACK added new rule: D-SACK changes any tag to S. @@ -1111,8 +1097,7 @@ static bool tcp_check_dsack(struct sock *sk, const struct sk_buff *ack_skb, } struct tcp_sacktag_state { - int reord; - int fack_count; + u32 reord; /* Timestamps for earliest and latest never-retransmitted segment * that was SACKed. RTO needs the earliest RTT to stay conservative, * but congestion control should still get an accurate delay signal. @@ -1188,15 +1173,15 @@ static u8 tcp_sacktag_one(struct sock *sk, u64 xmit_time) { struct tcp_sock *tp = tcp_sk(sk); - int fack_count = state->fack_count; /* Account D-SACK for retransmitted packet. */ if (dup_sack && (sacked & TCPCB_RETRANS)) { if (tp->undo_marker && tp->undo_retrans > 0 && after(end_seq, tp->undo_marker)) tp->undo_retrans--; - if (sacked & TCPCB_SACKED_ACKED) - state->reord = min(fack_count, state->reord); + if ((sacked & TCPCB_SACKED_ACKED) && + before(start_seq, state->reord)) + state->reord = start_seq; } /* Nothing to do; acked frame is about to be dropped (was ACKed). */ @@ -1222,9 +1207,10 @@ static u8 tcp_sacktag_one(struct sock *sk, * which was in hole. It is reordering. */ if (before(start_seq, - tcp_highest_sack_seq(tp))) - state->reord = min(fack_count, - state->reord); + tcp_highest_sack_seq(tp)) && + before(start_seq, state->reord)) + state->reord = start_seq; + if (!after(end_seq, tp->high_seq)) state->flag |= FLAG_ORIG_SACK_ACKED; if (state->first_sackt == 0) @@ -1243,15 +1229,10 @@ static u8 tcp_sacktag_one(struct sock *sk, tp->sacked_out += pcount; tp->delivered += pcount; /* Out-of-order packets delivered */ - fack_count += pcount; - /* Lost marker hint past SACKed? Tweak RFC3517 cnt */ - if (!tcp_is_fack(tp) && tp->lost_skb_hint && + if (tp->lost_skb_hint && before(start_seq, TCP_SKB_CB(tp->lost_skb_hint)->seq)) tp->lost_cnt_hint += pcount; - - if (fack_count > tp->fackets_out) - tp->fackets_out = fack_count; } /* D-SACK. We can detect redundant retransmission in S|R and plain R @@ -1498,7 +1479,6 @@ static struct sk_buff *tcp_shift_skb_data(struct sock *sk, struct sk_buff *skb, } out: - state->fack_count += pcount; return prev; noop: @@ -1577,8 +1557,6 @@ static struct sk_buff *tcp_sacktag_walk(struct sk_buff *skb, struct sock *sk, tcp_highest_sack_seq(tp))) tcp_advance_highest_sack(sk, skb); } - - state->fack_count += tcp_skb_pcount(skb); } return skb; } @@ -1589,7 +1567,6 @@ static struct sk_buff *tcp_sacktag_bsearch(struct sock *sk, { struct rb_node *parent, **p = &sk->tcp_rtx_queue.rb_node; struct sk_buff *skb; - int unack_bytes; while (*p) { parent = *p; @@ -1602,12 +1579,6 @@ static struct sk_buff *tcp_sacktag_bsearch(struct sock *sk, p = &parent->rb_right; continue; } - - state->fack_count = 0; - unack_bytes = TCP_SKB_CB(skb)->seq - tcp_sk(sk)->snd_una; - if (state->mss_now && unack_bytes > 0) - state->fack_count = unack_bytes / state->mss_now; - return skb; } return NULL; @@ -1665,13 +1636,10 @@ tcp_sacktag_write_queue(struct sock *sk, const struct sk_buff *ack_skb, int first_sack_index; state->flag = 0; - state->reord = tp->packets_out; + state->reord = tp->snd_nxt; - if (!tp->sacked_out) { - if (WARN_ON(tp->fackets_out)) - tp->fackets_out = 0; + if (!tp->sacked_out) tcp_highest_sack_reset(sk); - } found_dup_sack = tcp_check_dsack(sk, ack_skb, sp_wire, num_sacks, prior_snd_una); @@ -1743,7 +1711,6 @@ tcp_sacktag_write_queue(struct sock *sk, const struct sk_buff *ack_skb, } state->mss_now = tcp_current_mss(sk); - state->fack_count = 0; skb = NULL; i = 0; @@ -1801,7 +1768,6 @@ tcp_sacktag_write_queue(struct sock *sk, const struct sk_buff *ack_skb, skb = tcp_highest_sack(sk); if (!skb) break; - state->fack_count = tp->fackets_out; cache++; goto walk; } @@ -1816,7 +1782,6 @@ tcp_sacktag_write_queue(struct sock *sk, const struct sk_buff *ack_skb, skb = tcp_highest_sack(sk); if (!skb) break; - state->fack_count = tp->fackets_out; } skb = tcp_sacktag_skip(skb, sk, state, start_seq); @@ -1836,9 +1801,8 @@ advance_sp: for (j = 0; j < used_sacks; j++) tp->recv_sack_cache[i++] = sp[j]; - if ((state->reord < tp->fackets_out) && - ((inet_csk(sk)->icsk_ca_state != TCP_CA_Loss) || tp->undo_marker)) - tcp_update_reordering(sk, tp->fackets_out - state->reord, 0); + if (inet_csk(sk)->icsk_ca_state != TCP_CA_Loss || tp->undo_marker) + tcp_check_sack_reordering(sk, state->reord, 0); tcp_verify_left_out(tp); out: @@ -1876,8 +1840,13 @@ static bool tcp_limit_reno_sacked(struct tcp_sock *tp) static void tcp_check_reno_reordering(struct sock *sk, const int addend) { struct tcp_sock *tp = tcp_sk(sk); - if (tcp_limit_reno_sacked(tp)) - tcp_update_reordering(sk, tp->packets_out + addend, 0); + + if (!tcp_limit_reno_sacked(tp)) + return; + + tp->reordering = min_t(u32, tp->packets_out + addend, + sock_net(sk)->ipv4.sysctl_tcp_max_reordering); + NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPRENOREORDER); } /* Emulate SACKs for SACKless connection: account for a new dupack. */ @@ -1923,7 +1892,6 @@ void tcp_clear_retrans(struct tcp_sock *tp) tp->lost_out = 0; tp->undo_marker = 0; tp->undo_retrans = -1; - tp->fackets_out = 0; tp->sacked_out = 0; } @@ -1973,7 +1941,6 @@ void tcp_enter_loss(struct sock *sk) if (is_reneg) { NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPSACKRENEGING); tp->sacked_out = 0; - tp->fackets_out = 0; } tcp_clear_all_retrans_hints(tp); @@ -2040,19 +2007,10 @@ static bool tcp_check_sack_reneging(struct sock *sk, int flag) return false; } -static inline int tcp_fackets_out(const struct tcp_sock *tp) -{ - return tcp_is_reno(tp) ? tp->sacked_out + 1 : tp->fackets_out; -} - /* Heurestics to calculate number of duplicate ACKs. There's no dupACKs * counter when SACK is enabled (without SACK, sacked_out is used for * that purpose). * - * Instead, with FACK TCP uses fackets_out that includes both SACKed - * segments up to the highest received SACK block so far and holes in - * between them. - * * With reordering, holes may still be in flight, so RFC3517 recovery * uses pure sacked_out (total number of SACKed segments) even though * it violates the RFC that uses duplicate ACKs, often these are equal @@ -2062,10 +2020,10 @@ static inline int tcp_fackets_out(const struct tcp_sock *tp) */ static inline int tcp_dupack_heuristics(const struct tcp_sock *tp) { - return tcp_is_fack(tp) ? tp->fackets_out : tp->sacked_out + 1; + return tp->sacked_out + 1; } -/* Linux NewReno/SACK/FACK/ECN state machine. +/* Linux NewReno/SACK/ECN state machine. * -------------------------------------- * * "Open" Normal state, no dubious events, fast path. @@ -2130,16 +2088,6 @@ static inline int tcp_dupack_heuristics(const struct tcp_sock *tp) * dynamically measured and adjusted. This is implemented in * tcp_rack_mark_lost. * - * FACK (Disabled by default. Subsumbed by RACK): - * It is the simplest heuristics. As soon as we decided - * that something is lost, we decide that _all_ not SACKed - * packets until the most forward SACK are lost. I.e. - * lost_out = fackets_out - sacked_out and left_out = fackets_out. - * It is absolutely correct estimate, if network does not reorder - * packets. And it loses any connection to reality when reordering - * takes place. We use FACK by default until reordering - * is suspected on the path to this destination. - * * If the receiver does not support SACK: * * NewReno (RFC6582): in Recovery we assume that one segment @@ -2188,7 +2136,7 @@ static bool tcp_time_to_recover(struct sock *sk, int flag) } /* Detect loss in event "A" above by marking head of queue up as lost. - * For FACK or non-SACK(Reno) senders, the first "packets" number of segments + * For non-SACK(Reno) senders, the first "packets" number of segments * are considered lost. For RFC3517 SACK, a segment is considered lost if it * has at least tp->reordering SACKed seqments above it; "packets" refers to * the maximum SACKed segments to pass before reaching this limit. @@ -2224,12 +2172,12 @@ static void tcp_mark_head_lost(struct sock *sk, int packets, int mark_head) break; oldcnt = cnt; - if (tcp_is_fack(tp) || tcp_is_reno(tp) || + if (tcp_is_reno(tp) || (TCP_SKB_CB(skb)->sacked & TCPCB_SACKED_ACKED)) cnt += tcp_skb_pcount(skb); if (cnt > packets) { - if ((tcp_is_sack(tp) && !tcp_is_fack(tp)) || + if (tcp_is_sack(tp) || (TCP_SKB_CB(skb)->sacked & TCPCB_SACKED_ACKED) || (oldcnt >= packets)) break; @@ -2260,11 +2208,6 @@ static void tcp_update_scoreboard(struct sock *sk, int fast_rexmit) if (tcp_is_reno(tp)) { tcp_mark_head_lost(sk, 1, 1); - } else if (tcp_is_fack(tp)) { - int lost = tp->fackets_out - tp->reordering; - if (lost <= 0) - lost = 1; - tcp_mark_head_lost(sk, lost, 0); } else { int sacked_upto = tp->sacked_out - tp->reordering; if (sacked_upto >= 0) @@ -2611,7 +2554,6 @@ void tcp_simple_retransmit(struct sock *sk) struct tcp_sock *tp = tcp_sk(sk); struct sk_buff *skb; unsigned int mss = tcp_current_mss(sk); - u32 prior_lost = tp->lost_out; skb_rbtree_walk(skb, &sk->tcp_rtx_queue) { if (tcp_skb_seglen(skb) > mss && @@ -2626,7 +2568,7 @@ void tcp_simple_retransmit(struct sock *sk) tcp_clear_retrans_hints_partial(tp); - if (prior_lost == tp->lost_out) + if (!tp->lost_out) return; if (tcp_is_reno(tp)) @@ -2734,15 +2676,15 @@ static void tcp_process_loss(struct sock *sk, int flag, bool is_dupack, } /* Undo during fast recovery after partial ACK. */ -static bool tcp_try_undo_partial(struct sock *sk, const int acked) +static bool tcp_try_undo_partial(struct sock *sk, u32 prior_snd_una) { struct tcp_sock *tp = tcp_sk(sk); if (tp->undo_marker && tcp_packet_delayed(tp)) { /* Plain luck! Hole if filled with delayed - * packet, rather than with a retransmit. + * packet, rather than with a retransmit. Check reordering. */ - tcp_update_reordering(sk, tcp_fackets_out(tp) + acked, 1); + tcp_check_sack_reordering(sk, prior_snd_una, 1); /* We are getting evidence that the reordering degree is higher * than we realized. If there are no retransmits out then we @@ -2778,6 +2720,14 @@ static void tcp_rack_identify_loss(struct sock *sk, int *ack_flag) } } +static bool tcp_force_fast_retransmit(struct sock *sk) +{ + struct tcp_sock *tp = tcp_sk(sk); + + return after(tcp_highest_sack_seq(tp), + tp->snd_una + tp->reordering * tp->mss_cache); +} + /* Process an event, which can update packets-in-flight not trivially. * Main goal of this function is to calculate new estimate for left_out, * taking into account both packets sitting in receiver's buffer and @@ -2790,19 +2740,17 @@ static void tcp_rack_identify_loss(struct sock *sk, int *ack_flag) * It does _not_ decide what to send, it is made in function * tcp_xmit_retransmit_queue(). */ -static void tcp_fastretrans_alert(struct sock *sk, const int acked, +static void tcp_fastretrans_alert(struct sock *sk, const u32 prior_snd_una, bool is_dupack, int *ack_flag, int *rexmit) { struct inet_connection_sock *icsk = inet_csk(sk); struct tcp_sock *tp = tcp_sk(sk); int fast_rexmit = 0, flag = *ack_flag; bool do_lost = is_dupack || ((flag & FLAG_DATA_SACKED) && - (tcp_fackets_out(tp) > tp->reordering)); + tcp_force_fast_retransmit(sk)); if (!tp->packets_out && tp->sacked_out) tp->sacked_out = 0; - if (!tp->sacked_out && tp->fackets_out) - tp->fackets_out = 0; /* Now state machine starts. * A. ECE, hence prohibit cwnd undoing, the reduction is required. */ @@ -2849,11 +2797,11 @@ static void tcp_fastretrans_alert(struct sock *sk, const int acked, if (tcp_is_reno(tp) && is_dupack) tcp_add_reno_sack(sk); } else { - if (tcp_try_undo_partial(sk, acked)) + if (tcp_try_undo_partial(sk, prior_snd_una)) return; /* Partial ACK arrived. Force fast retransmit. */ do_lost = tcp_is_reno(tp) || - tcp_fackets_out(tp) > tp->reordering; + tcp_force_fast_retransmit(sk); } if (tcp_try_undo_dsack(sk)) { tcp_try_keep_open(sk); @@ -3016,7 +2964,7 @@ void tcp_rearm_rto(struct sock *sk) /* Try to schedule a loss probe; if that doesn't work, then schedule an RTO. */ static void tcp_set_xmit_timer(struct sock *sk) { - if (!tcp_schedule_loss_probe(sk)) + if (!tcp_schedule_loss_probe(sk, true)) tcp_rearm_rto(sk); } @@ -3063,15 +3011,15 @@ static void tcp_ack_tstamp(struct sock *sk, struct sk_buff *skb, * is before the ack sequence we can discard it as it's confirmed to have * arrived at the other end. */ -static int tcp_clean_rtx_queue(struct sock *sk, int prior_fackets, - u32 prior_snd_una, int *acked, +static int tcp_clean_rtx_queue(struct sock *sk, u32 prior_fack, + u32 prior_snd_una, struct tcp_sacktag_state *sack) { const struct inet_connection_sock *icsk = inet_csk(sk); u64 first_ackt, last_ackt; struct tcp_sock *tp = tcp_sk(sk); u32 prior_sacked = tp->sacked_out; - u32 reord = tp->packets_out; + u32 reord = tp->snd_nxt; /* lowest acked un-retx un-sacked seq */ struct sk_buff *skb, *next; bool fully_acked = true; long sack_rtt_us = -1L; @@ -3086,6 +3034,7 @@ static int tcp_clean_rtx_queue(struct sock *sk, int prior_fackets, for (skb = skb_rb_first(&sk->tcp_rtx_queue); skb; skb = next) { struct tcp_skb_cb *scb = TCP_SKB_CB(skb); + const u32 start_seq = scb->seq; u8 sacked = scb->sacked; u32 acked_pcount; @@ -3116,7 +3065,8 @@ static int tcp_clean_rtx_queue(struct sock *sk, int prior_fackets, first_ackt = last_ackt; last_in_flight = TCP_SKB_CB(skb)->tx.in_flight; - reord = min(pkts_acked, reord); + if (before(start_seq, reord)) + reord = start_seq; if (!after(scb->end_seq, tp->high_seq)) flag |= FLAG_ORIG_SACK_ACKED; } @@ -3194,16 +3144,12 @@ static int tcp_clean_rtx_queue(struct sock *sk, int prior_fackets, int delta; /* Non-retransmitted hole got filled? That's reordering */ - if (reord < prior_fackets && reord <= tp->fackets_out) - tcp_update_reordering(sk, tp->fackets_out - reord, 0); + if (before(reord, prior_fack)) + tcp_check_sack_reordering(sk, reord, 0); - delta = tcp_is_fack(tp) ? pkts_acked : - prior_sacked - tp->sacked_out; + delta = prior_sacked - tp->sacked_out; tp->lost_cnt_hint -= min(tp->lost_cnt_hint, delta); } - - tp->fackets_out -= min(pkts_acked, tp->fackets_out); - } else if (skb && rtt_update && sack_rtt_us >= 0 && sack_rtt_us > tcp_stamp_us_delta(tp->tcp_mstamp, skb->skb_mstamp)) { /* Do not re-arm RTO if the sack RTT is measured from data sent @@ -3244,7 +3190,6 @@ static int tcp_clean_rtx_queue(struct sock *sk, int prior_fackets, } } #endif - *acked = pkts_acked; return flag; } @@ -3553,12 +3498,11 @@ static int tcp_ack(struct sock *sk, const struct sk_buff *skb, int flag) u32 ack_seq = TCP_SKB_CB(skb)->seq; u32 ack = TCP_SKB_CB(skb)->ack_seq; bool is_dupack = false; - u32 prior_fackets; int prior_packets = tp->packets_out; u32 delivered = tp->delivered; u32 lost = tp->lost; - int acked = 0; /* Number of packets newly acked */ int rexmit = REXMIT_NONE; /* Flag to (re)transmit to recover losses */ + u32 prior_fack; sack_state.first_sackt = 0; sack_state.rate = &rs; @@ -3590,7 +3534,7 @@ static int tcp_ack(struct sock *sk, const struct sk_buff *skb, int flag) icsk->icsk_retransmits = 0; } - prior_fackets = tp->fackets_out; + prior_fack = tcp_is_sack(tp) ? tcp_highest_sack_seq(tp) : tp->snd_una; rs.prior_in_flight = tcp_packets_in_flight(tp); /* ts_recent update must be made after we are sure that the packet @@ -3646,8 +3590,7 @@ static int tcp_ack(struct sock *sk, const struct sk_buff *skb, int flag) goto no_queue; /* See if we can take anything off of the retransmit queue. */ - flag |= tcp_clean_rtx_queue(sk, prior_fackets, prior_snd_una, &acked, - &sack_state); + flag |= tcp_clean_rtx_queue(sk, prior_fack, prior_snd_una, &sack_state); tcp_rack_update_reo_wnd(sk, &rs); @@ -3659,7 +3602,8 @@ static int tcp_ack(struct sock *sk, const struct sk_buff *skb, int flag) if (tcp_ack_is_dubious(sk, flag)) { is_dupack = !(flag & (FLAG_SND_UNA_ADVANCED | FLAG_NOT_DUP)); - tcp_fastretrans_alert(sk, acked, is_dupack, &flag, &rexmit); + tcp_fastretrans_alert(sk, prior_snd_una, is_dupack, &flag, + &rexmit); } if ((flag & FLAG_FORWARD_PROGRESS) || !(flag & FLAG_NOT_DUP)) @@ -3675,7 +3619,8 @@ static int tcp_ack(struct sock *sk, const struct sk_buff *skb, int flag) no_queue: /* If data was DSACKed, see if we can undo a cwnd reduction. */ if (flag & FLAG_DSACKING_ACK) - tcp_fastretrans_alert(sk, acked, is_dupack, &flag, &rexmit); + tcp_fastretrans_alert(sk, prior_snd_una, is_dupack, &flag, + &rexmit); /* If this ack opens up a zero window, clear backoff. It was * being used to time the probes, and is probably far higher than * it needs to be for normal retransmission. @@ -3697,7 +3642,8 @@ old_ack: if (TCP_SKB_CB(skb)->sacked) { flag |= tcp_sacktag_write_queue(sk, skb, prior_snd_una, &sack_state); - tcp_fastretrans_alert(sk, acked, is_dupack, &flag, &rexmit); + tcp_fastretrans_alert(sk, prior_snd_una, is_dupack, &flag, + &rexmit); tcp_xmit_recovery(sk, rexmit); } @@ -5706,9 +5652,6 @@ static int tcp_rcv_synsent_state_process(struct sock *sk, struct sk_buff *skb, tp->tcp_header_len = sizeof(struct tcphdr); } - if (tcp_is_sack(tp) && sock_net(sk)->ipv4.sysctl_tcp_fack) - tcp_enable_fack(tp); - tcp_sync_mss(sk, icsk->icsk_pmtu_cookie); tcp_initialize_rcv_mss(sk); @@ -6187,7 +6130,6 @@ struct request_sock *inet_reqsk_alloc(const struct request_sock_ops *ops, if (req) { struct inet_request_sock *ireq = inet_rsk(req); - kmemcheck_annotate_bitfield(ireq, flags); ireq->ireq_opt = NULL; #if IS_ENABLED(CONFIG_IPV6) ireq->pktopts = NULL; diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index 0162c577bb9c..c6bc0c4d19c6 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -2409,8 +2409,8 @@ struct proto tcp_prot = { .memory_allocated = &tcp_memory_allocated, .memory_pressure = &tcp_memory_pressure, .sysctl_mem = sysctl_tcp_mem, - .sysctl_wmem = sysctl_tcp_wmem, - .sysctl_rmem = sysctl_tcp_rmem, + .sysctl_wmem_offset = offsetof(struct net, ipv4.sysctl_tcp_wmem), + .sysctl_rmem_offset = offsetof(struct net, ipv4.sysctl_tcp_rmem), .max_header = MAX_TCP_HEADER, .obj_size = sizeof(struct tcp_sock), .slab_flags = SLAB_TYPESAFE_BY_RCU, @@ -2430,6 +2430,8 @@ static void __net_exit tcp_sk_exit(struct net *net) { int cpu; + module_put(net->ipv4.tcp_congestion_control->owner); + for_each_possible_cpu(cpu) inet_ctl_sock_destroy(*per_cpu_ptr(net->ipv4.tcp_sk, cpu)); free_percpu(net->ipv4.tcp_sk); @@ -2509,12 +2511,26 @@ static int __net_init tcp_sk_init(struct net *net) net->ipv4.sysctl_tcp_invalid_ratelimit = HZ/2; net->ipv4.sysctl_tcp_pacing_ss_ratio = 200; net->ipv4.sysctl_tcp_pacing_ca_ratio = 120; - + if (net != &init_net) { + memcpy(net->ipv4.sysctl_tcp_rmem, + init_net.ipv4.sysctl_tcp_rmem, + sizeof(init_net.ipv4.sysctl_tcp_rmem)); + memcpy(net->ipv4.sysctl_tcp_wmem, + init_net.ipv4.sysctl_tcp_wmem, + sizeof(init_net.ipv4.sysctl_tcp_wmem)); + } net->ipv4.sysctl_tcp_fastopen = TFO_CLIENT_ENABLE; spin_lock_init(&net->ipv4.tcp_fastopen_ctx_lock); net->ipv4.sysctl_tcp_fastopen_blackhole_timeout = 60 * 60; atomic_set(&net->ipv4.tfo_active_disable_times, 0); + /* Reno is always built in */ + if (!net_eq(net, &init_net) && + try_module_get(init_net.ipv4.tcp_congestion_control->owner)) + net->ipv4.tcp_congestion_control = init_net.ipv4.tcp_congestion_control; + else + net->ipv4.tcp_congestion_control = &tcp_reno; + return 0; fail: tcp_sk_exit(net); diff --git a/net/ipv4/tcp_metrics.c b/net/ipv4/tcp_metrics.c index 9d5ddebfd831..7097f92d16e5 100644 --- a/net/ipv4/tcp_metrics.c +++ b/net/ipv4/tcp_metrics.c @@ -470,10 +470,8 @@ void tcp_init_metrics(struct sock *sk) tp->snd_ssthresh = TCP_INFINITE_SSTHRESH; } val = tcp_metric_get(tm, TCP_METRIC_REORDERING); - if (val && tp->reordering != val) { - tcp_disable_fack(tp); + if (val && tp->reordering != val) tp->reordering = val; - } crtt = tcp_metric_get(tm, TCP_METRIC_RTT); rcu_read_unlock(); diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c index 4bb86580decd..e36eff0403f4 100644 --- a/net/ipv4/tcp_minisocks.c +++ b/net/ipv4/tcp_minisocks.c @@ -475,7 +475,6 @@ struct sock *tcp_create_openreq_child(const struct sock *sk, newtp->packets_out = 0; newtp->retrans_out = 0; newtp->sacked_out = 0; - newtp->fackets_out = 0; newtp->snd_ssthresh = TCP_INFINITE_SSTHRESH; newtp->tlp_high_seq = 0; newtp->lsndtime = tcp_jiffies32; @@ -509,10 +508,7 @@ struct sock *tcp_create_openreq_child(const struct sock *sk, keepalive_time_when(newtp)); newtp->rx_opt.tstamp_ok = ireq->tstamp_ok; - if ((newtp->rx_opt.sack_ok = ireq->sack_ok) != 0) { - if (sock_net(sk)->ipv4.sysctl_tcp_fack) - tcp_enable_fack(newtp); - } + newtp->rx_opt.sack_ok = ireq->sack_ok; newtp->window_clamp = req->rsk_window_clamp; newtp->rcv_ssthresh = req->rsk_rcv_wnd; newtp->rcv_wnd = req->rsk_rcv_wnd; diff --git a/net/ipv4/tcp_offload.c b/net/ipv4/tcp_offload.c index 11f69bbf9307..b6a2aa1dcf56 100644 --- a/net/ipv4/tcp_offload.c +++ b/net/ipv4/tcp_offload.c @@ -149,11 +149,19 @@ struct sk_buff *tcp_gso_segment(struct sk_buff *skb, * is freed by GSO engine */ if (copy_destructor) { + int delta; + swap(gso_skb->sk, skb->sk); swap(gso_skb->destructor, skb->destructor); sum_truesize += skb->truesize; - refcount_add(sum_truesize - gso_skb->truesize, - &skb->sk->sk_wmem_alloc); + delta = sum_truesize - gso_skb->truesize; + /* In some pathological cases, delta can be negative. + * We need to either use refcount_add() or refcount_sub_and_test() + */ + if (likely(delta >= 0)) + refcount_add(delta, &skb->sk->sk_wmem_alloc); + else + WARN_ON_ONCE(refcount_sub_and_test(-delta, &skb->sk->sk_wmem_alloc)); } delta = htonl(oldlen + (skb_tail_pointer(skb) - diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index a9d917e4dad5..a4d214c7b506 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -220,7 +220,7 @@ void tcp_select_initial_window(const struct sock *sk, int __space, __u32 mss, (*rcv_wscale) = 0; if (wscale_ok) { /* Set window scaling on max possible window */ - space = max_t(u32, space, sysctl_tcp_rmem[2]); + space = max_t(u32, space, sock_net(sk)->ipv4.sysctl_tcp_rmem[2]); space = max_t(u32, space, sysctl_rmem_max); space = min_t(u32, space, *window_clamp); while (space > U16_MAX && (*rcv_wscale) < TCP_MAX_WSCALE) { @@ -1218,21 +1218,6 @@ static void tcp_set_skb_tso_segs(struct sk_buff *skb, unsigned int mss_now) } } -/* When a modification to fackets out becomes necessary, we need to check - * skb is counted to fackets_out or not. - */ -static void tcp_adjust_fackets_out(struct sock *sk, const struct sk_buff *skb, - int decr) -{ - struct tcp_sock *tp = tcp_sk(sk); - - if (!tp->sacked_out || tcp_is_reno(tp)) - return; - - if (after(tcp_highest_sack_seq(tp), TCP_SKB_CB(skb)->seq)) - tp->fackets_out -= decr; -} - /* Pcount in the middle of the write queue got changed, we need to do various * tweaks to fix counters */ @@ -1253,11 +1238,9 @@ static void tcp_adjust_pcount(struct sock *sk, const struct sk_buff *skb, int de if (tcp_is_reno(tp) && decr > 0) tp->sacked_out -= min_t(u32, tp->sacked_out, decr); - tcp_adjust_fackets_out(sk, skb, decr); - if (tp->lost_skb_hint && before(TCP_SKB_CB(skb)->seq, TCP_SKB_CB(tp->lost_skb_hint)->seq) && - (tcp_is_fack(tp) || (TCP_SKB_CB(skb)->sacked & TCPCB_SACKED_ACKED))) + (TCP_SKB_CB(skb)->sacked & TCPCB_SACKED_ACKED)) tp->lost_cnt_hint -= decr; tcp_verify_left_out(tp); @@ -1737,7 +1720,7 @@ u32 tcp_tso_autosize(const struct sock *sk, unsigned int mss_now, { u32 bytes, segs; - bytes = min(sk->sk_pacing_rate >> 10, + bytes = min(sk->sk_pacing_rate >> sk->sk_pacing_shift, sk->sk_gso_max_size - 1 - MAX_TCP_HEADER); /* Goal is to send at least one packet per ms, @@ -1978,7 +1961,7 @@ static bool tcp_tso_should_defer(struct sock *sk, struct sk_buff *skb, if ((skb != tcp_write_queue_tail(sk)) && (limit >= skb->len)) goto send_now; - win_divisor = ACCESS_ONCE(sock_net(sk)->ipv4.sysctl_tcp_tso_win_divisor); + win_divisor = READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_tso_win_divisor); if (win_divisor) { u32 chunk = min(tp->snd_wnd, tp->snd_cwnd * tp->mss_cache); @@ -2215,7 +2198,7 @@ static bool tcp_small_queue_check(struct sock *sk, const struct sk_buff *skb, { unsigned int limit; - limit = max(2 * skb->truesize, sk->sk_pacing_rate >> 10); + limit = max(2 * skb->truesize, sk->sk_pacing_rate >> sk->sk_pacing_shift); limit = min_t(u32, limit, sock_net(sk)->ipv4.sysctl_tcp_limit_output_bytes); limit <<= factor; @@ -2408,7 +2391,7 @@ repair: /* Send one loss probe per tail loss episode. */ if (push_one != 2) - tcp_schedule_loss_probe(sk); + tcp_schedule_loss_probe(sk, false); is_cwnd_limited |= (tcp_packets_in_flight(tp) >= tp->snd_cwnd); tcp_cwnd_validate(sk, is_cwnd_limited); return false; @@ -2416,7 +2399,7 @@ repair: return !tp->packets_out && !tcp_write_queue_empty(sk); } -bool tcp_schedule_loss_probe(struct sock *sk) +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); @@ -2457,7 +2440,9 @@ bool tcp_schedule_loss_probe(struct sock *sk) } /* If the RTO formula yields an earlier time, then use that time. */ - rto_delta_us = tcp_rto_delta_us(sk); /* How far in future is RTO? */ + rto_delta_us = advancing_rto ? + jiffies_to_usecs(inet_csk(sk)->icsk_rto) : + tcp_rto_delta_us(sk); /* How far in future is RTO? */ if (rto_delta_us > 0) timeout = min_t(u32, timeout, usecs_to_jiffies(rto_delta_us)); @@ -2961,9 +2946,6 @@ int tcp_retransmit_skb(struct sock *sk, struct sk_buff *skb, int segs) * retransmitted data is acknowledged. It tries to continue * resending the rest of the retransmit queue, until either * we've sent it all or the congestion window limit is reached. - * If doing SACK, the first ACK which comes back for a timeout - * based retransmit packet might feed us FACK information again. - * If so, we use it to avoid unnecessarily retransmissions. */ void tcp_xmit_retransmit_queue(struct sock *sk) { diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index a6699af05539..e4ff25c947c5 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -1852,7 +1852,7 @@ static int udp_queue_rcv_skb(struct sock *sk, struct sk_buff *skb) */ /* if we're overly short, let UDP handle it */ - encap_rcv = ACCESS_ONCE(up->encap_rcv); + encap_rcv = READ_ONCE(up->encap_rcv); if (encap_rcv) { int ret; @@ -2297,7 +2297,7 @@ void udp_destroy_sock(struct sock *sk) unlock_sock_fast(sk, slow); if (static_key_false(&udp_encap_needed) && up->encap_type) { void (*encap_destroy)(struct sock *sk); - encap_destroy = ACCESS_ONCE(up->encap_destroy); + encap_destroy = READ_ONCE(up->encap_destroy); if (encap_destroy) encap_destroy(sk); } |