diff options
Diffstat (limited to 'net/ipv4/tcp_output.c')
-rw-r--r-- | net/ipv4/tcp_output.c | 182 |
1 files changed, 135 insertions, 47 deletions
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 4858e19..4e985de 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -151,7 +151,7 @@ void tcp_cwnd_restart(struct sock *sk, s32 delta) while ((delta -= inet_csk(sk)->icsk_rto) > 0 && cwnd > restart_cwnd) cwnd >>= 1; tp->snd_cwnd = max(cwnd, restart_cwnd); - tp->snd_cwnd_stamp = tcp_time_stamp; + tp->snd_cwnd_stamp = tcp_jiffies32; tp->snd_cwnd_used = 0; } @@ -160,7 +160,7 @@ static void tcp_event_data_sent(struct tcp_sock *tp, struct sock *sk) { struct inet_connection_sock *icsk = inet_csk(sk); - const u32 now = tcp_time_stamp; + const u32 now = tcp_jiffies32; if (tcp_packets_in_flight(tp) == 0) tcp_ca_event(sk, CA_EVENT_TX_START); @@ -316,7 +316,8 @@ static void tcp_ecn_send_synack(struct sock *sk, struct sk_buff *skb) TCP_SKB_CB(skb)->tcp_flags &= ~TCPHDR_CWR; if (!(tp->ecn_flags & TCP_ECN_OK)) TCP_SKB_CB(skb)->tcp_flags &= ~TCPHDR_ECE; - else if (tcp_ca_needs_ecn(sk)) + else if (tcp_ca_needs_ecn(sk) || + tcp_bpf_ca_needs_ecn(sk)) INET_ECN_xmit(sk); } @@ -324,8 +325,9 @@ static void tcp_ecn_send_synack(struct sock *sk, struct sk_buff *skb) static void tcp_ecn_send_syn(struct sock *sk, struct sk_buff *skb) { struct tcp_sock *tp = tcp_sk(sk); + bool bpf_needs_ecn = tcp_bpf_ca_needs_ecn(sk); bool use_ecn = sock_net(sk)->ipv4.sysctl_tcp_ecn == 1 || - tcp_ca_needs_ecn(sk); + tcp_ca_needs_ecn(sk) || bpf_needs_ecn; if (!use_ecn) { const struct dst_entry *dst = __sk_dst_get(sk); @@ -339,7 +341,7 @@ static void tcp_ecn_send_syn(struct sock *sk, struct sk_buff *skb) if (use_ecn) { TCP_SKB_CB(skb)->tcp_flags |= TCPHDR_ECE | TCPHDR_CWR; tp->ecn_flags = TCP_ECN_OK; - if (tcp_ca_needs_ecn(sk)) + if (tcp_ca_needs_ecn(sk) || bpf_needs_ecn) INET_ECN_xmit(sk); } } @@ -569,18 +571,18 @@ static unsigned int tcp_syn_options(struct sock *sk, struct sk_buff *skb, opts->mss = tcp_advertise_mss(sk); remaining -= TCPOLEN_MSS_ALIGNED; - if (likely(sysctl_tcp_timestamps && !*md5)) { + if (likely(sock_net(sk)->ipv4.sysctl_tcp_timestamps && !*md5)) { opts->options |= OPTION_TS; opts->tsval = tcp_skb_timestamp(skb) + tp->tsoffset; opts->tsecr = tp->rx_opt.ts_recent; remaining -= TCPOLEN_TSTAMP_ALIGNED; } - if (likely(sysctl_tcp_window_scaling)) { + if (likely(sock_net(sk)->ipv4.sysctl_tcp_window_scaling)) { opts->ws = tp->rx_opt.rcv_wscale; opts->options |= OPTION_WSCALE; remaining -= TCPOLEN_WSCALE_ALIGNED; } - if (likely(sysctl_tcp_sack)) { + if (likely(sock_net(sk)->ipv4.sysctl_tcp_sack)) { opts->options |= OPTION_SACK_ADVERTISE; if (unlikely(!(OPTION_TS & opts->options))) remaining -= TCPOLEN_SACKPERM_ALIGNED; @@ -861,12 +863,11 @@ void tcp_wfree(struct sk_buff *skb) struct sock *sk = skb->sk; struct tcp_sock *tp = tcp_sk(sk); unsigned long flags, nval, oval; - int wmem; /* Keep one reference on sk_wmem_alloc. * Will be released by sk_free() from here or tcp_tasklet_func() */ - wmem = atomic_sub_return(skb->truesize - 1, &sk->sk_wmem_alloc); + WARN_ON(refcount_sub_and_test(skb->truesize - 1, &sk->sk_wmem_alloc)); /* If this softirq is serviced by ksoftirqd, we are likely under stress. * Wait until our queues (qdisc + devices) are drained. @@ -875,7 +876,7 @@ void tcp_wfree(struct sk_buff *skb) * - chance for incoming ACK (processed by another cpu maybe) * to migrate this flow (skb->ooo_okay will be eventually set) */ - if (wmem >= SKB_TRUESIZE(1) && this_cpu_ksoftirqd() == current) + if (refcount_read(&sk->sk_wmem_alloc) >= SKB_TRUESIZE(1) && this_cpu_ksoftirqd() == current) goto out; for (oval = READ_ONCE(sk->sk_tsq_flags);; oval = nval) { @@ -904,6 +905,72 @@ out: sk_free(sk); } +/* Note: Called under hard irq. + * We can not call TCP stack right away. + */ +enum hrtimer_restart tcp_pace_kick(struct hrtimer *timer) +{ + struct tcp_sock *tp = container_of(timer, struct tcp_sock, pacing_timer); + struct sock *sk = (struct sock *)tp; + unsigned long nval, oval; + + for (oval = READ_ONCE(sk->sk_tsq_flags);; oval = nval) { + struct tsq_tasklet *tsq; + bool empty; + + if (oval & TSQF_QUEUED) + break; + + nval = (oval & ~TSQF_THROTTLED) | TSQF_QUEUED | TCPF_TSQ_DEFERRED; + nval = cmpxchg(&sk->sk_tsq_flags, oval, nval); + if (nval != oval) + continue; + + if (!refcount_inc_not_zero(&sk->sk_wmem_alloc)) + break; + /* queue this socket to tasklet queue */ + tsq = this_cpu_ptr(&tsq_tasklet); + empty = list_empty(&tsq->head); + list_add(&tp->tsq_node, &tsq->head); + if (empty) + tasklet_schedule(&tsq->tasklet); + break; + } + return HRTIMER_NORESTART; +} + +/* BBR congestion control needs pacing. + * Same remark for SO_MAX_PACING_RATE. + * sch_fq packet scheduler is efficiently handling pacing, + * but is not always installed/used. + * Return true if TCP stack should pace packets itself. + */ +static bool tcp_needs_internal_pacing(const struct sock *sk) +{ + return smp_load_acquire(&sk->sk_pacing_status) == SK_PACING_NEEDED; +} + +static void tcp_internal_pacing(struct sock *sk, const struct sk_buff *skb) +{ + u64 len_ns; + u32 rate; + + if (!tcp_needs_internal_pacing(sk)) + return; + rate = sk->sk_pacing_rate; + if (!rate || rate == ~0U) + return; + + /* Should account for header sizes as sch_fq does, + * but lets make things simple. + */ + len_ns = (u64)skb->len * NSEC_PER_SEC; + do_div(len_ns, rate); + hrtimer_start(&tcp_sk(sk)->pacing_timer, + ktime_add_ns(ktime_get(), len_ns), + HRTIMER_MODE_ABS_PINNED); +} + /* This routine actually transmits TCP packets queued in by * tcp_do_sendmsg(). This is used by both the initial * transmission and possible later retransmissions. @@ -931,8 +998,8 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, int clone_it, BUG_ON(!skb || !tcp_skb_pcount(skb)); tp = tcp_sk(sk); + skb->skb_mstamp = tp->tcp_mstamp; if (clone_it) { - skb_mstamp_get(&skb->skb_mstamp); TCP_SKB_CB(skb)->tx.in_flight = TCP_SKB_CB(skb)->end_seq - tp->snd_una; tcp_rate_skb_sent(sk, skb); @@ -979,7 +1046,7 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, int clone_it, skb->sk = sk; skb->destructor = skb_is_tcp_pure_ack(skb) ? __sock_wfree : tcp_wfree; skb_set_hash_from_sk(skb, sk); - atomic_add(skb->truesize, &sk->sk_wmem_alloc); + refcount_add(skb->truesize, &sk->sk_wmem_alloc); skb_set_dst_pending_confirm(skb, sk->sk_dst_pending_confirm); @@ -1034,6 +1101,7 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, int clone_it, if (skb->len != tcp_header_size) { tcp_event_data_sent(tp, sk); tp->data_segs_out += tcp_skb_pcount(skb); + tcp_internal_pacing(sk, skb); } if (after(tcb->end_seq, tp->snd_nxt) || tcb->seq == tcb->end_seq) @@ -1261,9 +1329,8 @@ int tcp_fragment(struct sock *sk, struct sk_buff *skb, u32 len, return 0; } -/* This is similar to __pskb_pull_head() (it will go to core/skbuff.c - * eventually). The difference is that pulled data not copied, but - * immediately discarded. +/* This is similar to __pskb_pull_tail(). The difference is that pulled + * data is not copied, but immediately discarded. */ static int __pskb_trim_head(struct sk_buff *skb, int len) { @@ -1298,7 +1365,6 @@ static int __pskb_trim_head(struct sk_buff *skb, int len) } shinfo->nr_frags = k; - skb_reset_tail_pointer(skb); skb->data_len -= len; skb->len = skb->data_len; return len; @@ -1408,7 +1474,7 @@ void tcp_mtup_init(struct sock *sk) icsk->icsk_mtup.search_low = tcp_mss_to_mtu(sk, net->ipv4.sysctl_tcp_base_mss); icsk->icsk_mtup.probe_size = 0; if (icsk->icsk_mtup.enabled) - icsk->icsk_mtup.probe_timestamp = tcp_time_stamp; + icsk->icsk_mtup.probe_timestamp = tcp_jiffies32; } EXPORT_SYMBOL(tcp_mtup_init); @@ -1509,7 +1575,7 @@ static void tcp_cwnd_application_limited(struct sock *sk) } tp->snd_cwnd_used = 0; } - tp->snd_cwnd_stamp = tcp_time_stamp; + tp->snd_cwnd_stamp = tcp_jiffies32; } static void tcp_cwnd_validate(struct sock *sk, bool is_cwnd_limited) @@ -1530,14 +1596,14 @@ static void tcp_cwnd_validate(struct sock *sk, bool is_cwnd_limited) if (tcp_is_cwnd_limited(sk)) { /* Network is feed fully. */ tp->snd_cwnd_used = 0; - tp->snd_cwnd_stamp = tcp_time_stamp; + tp->snd_cwnd_stamp = tcp_jiffies32; } else { /* Network starves. */ if (tp->packets_out > tp->snd_cwnd_used) tp->snd_cwnd_used = tp->packets_out; if (sysctl_tcp_slow_start_after_idle && - (s32)(tcp_time_stamp - tp->snd_cwnd_stamp) >= inet_csk(sk)->icsk_rto && + (s32)(tcp_jiffies32 - tp->snd_cwnd_stamp) >= inet_csk(sk)->icsk_rto && !ca_ops->cong_control) tcp_cwnd_application_limited(sk); @@ -1839,7 +1905,6 @@ static bool tcp_tso_should_defer(struct sock *sk, struct sk_buff *skb, const struct inet_connection_sock *icsk = inet_csk(sk); u32 age, send_win, cong_win, limit, in_flight; struct tcp_sock *tp = tcp_sk(sk); - struct skb_mstamp now; struct sk_buff *head; int win_divisor; @@ -1852,7 +1917,7 @@ static bool tcp_tso_should_defer(struct sock *sk, struct sk_buff *skb, /* Avoid bursty behavior by allowing defer * only if the last write was recent. */ - if ((s32)(tcp_time_stamp - tp->lsndtime) > 0) + if ((s32)(tcp_jiffies32 - tp->lsndtime) > 0) goto send_now; in_flight = tcp_packets_in_flight(tp); @@ -1895,8 +1960,8 @@ static bool tcp_tso_should_defer(struct sock *sk, struct sk_buff *skb, } head = tcp_write_queue_head(sk); - skb_mstamp_get(&now); - age = skb_mstamp_us_delta(&now, &head->skb_mstamp); + + age = tcp_stamp_us_delta(tp->tcp_mstamp, head->skb_mstamp); /* If next ACK is likely to come too late (half srtt), do not defer */ if (age < (tp->srtt_us >> 4)) goto send_now; @@ -1921,7 +1986,7 @@ static inline void tcp_mtu_check_reprobe(struct sock *sk) s32 delta; interval = net->ipv4.sysctl_tcp_probe_interval; - delta = tcp_time_stamp - icsk->icsk_mtup.probe_timestamp; + delta = tcp_jiffies32 - icsk->icsk_mtup.probe_timestamp; if (unlikely(delta >= interval * HZ)) { int mss = tcp_current_mss(sk); @@ -1933,7 +1998,7 @@ static inline void tcp_mtu_check_reprobe(struct sock *sk) icsk->icsk_mtup.search_low = tcp_mss_to_mtu(sk, mss); /* Update probe time stamp */ - icsk->icsk_mtup.probe_timestamp = tcp_time_stamp; + icsk->icsk_mtup.probe_timestamp = tcp_jiffies32; } } @@ -2086,6 +2151,12 @@ static int tcp_mtu_probe(struct sock *sk) return -1; } +static bool tcp_pacing_check(const struct sock *sk) +{ + return tcp_needs_internal_pacing(sk) && + hrtimer_active(&tcp_sk(sk)->pacing_timer); +} + /* TCP Small Queues : * Control number of packets in qdisc/devices to two packets / or ~1 ms. * (These limits are doubled for retransmits) @@ -2106,7 +2177,7 @@ static bool tcp_small_queue_check(struct sock *sk, const struct sk_buff *skb, limit = min_t(u32, limit, sysctl_tcp_limit_output_bytes); limit <<= factor; - if (atomic_read(&sk->sk_wmem_alloc) > limit) { + if (refcount_read(&sk->sk_wmem_alloc) > limit) { /* Always send the 1st or 2nd skb in write queue. * No need to wait for TX completion to call us back, * after softirq/tasklet schedule. @@ -2122,7 +2193,7 @@ static bool tcp_small_queue_check(struct sock *sk, const struct sk_buff *skb, * test again the condition. */ smp_mb__after_atomic(); - if (atomic_read(&sk->sk_wmem_alloc) > limit) + if (refcount_read(&sk->sk_wmem_alloc) > limit) return true; } return false; @@ -2130,7 +2201,7 @@ static bool tcp_small_queue_check(struct sock *sk, const struct sk_buff *skb, static void tcp_chrono_set(struct tcp_sock *tp, const enum tcp_chrono new) { - const u32 now = tcp_time_stamp; + const u32 now = tcp_jiffies32; if (tp->chrono_type > TCP_CHRONO_UNSPEC) tp->chrono_stat[tp->chrono_type - 1] += now - tp->chrono_start; @@ -2207,15 +2278,19 @@ static bool tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle, } max_segs = tcp_tso_segs(sk, mss_now); + tcp_mstamp_refresh(tp); while ((skb = tcp_send_head(sk))) { unsigned int limit; + if (tcp_pacing_check(sk)) + break; + tso_segs = tcp_init_tso_segs(skb, mss_now); BUG_ON(!tso_segs); if (unlikely(tp->repair) && tp->repair_queue == TCP_SEND_QUEUE) { /* "skb_mstamp" is used as a start point for the retransmit timer */ - skb_mstamp_get(&skb->skb_mstamp); + skb->skb_mstamp = tp->tcp_mstamp; goto repair; /* Skip network transmission */ } @@ -2342,10 +2417,10 @@ bool tcp_schedule_loss_probe(struct sock *sk) timeout = max_t(u32, timeout, msecs_to_jiffies(10)); /* If RTO is shorter, just schedule TLP in its place. */ - tlp_time_stamp = tcp_time_stamp + timeout; + tlp_time_stamp = tcp_jiffies32 + timeout; rto_time_stamp = (u32)inet_csk(sk)->icsk_timeout; if ((s32)(tlp_time_stamp - rto_time_stamp) > 0) { - s32 delta = rto_time_stamp - tcp_time_stamp; + s32 delta = rto_time_stamp - tcp_jiffies32; if (delta > 0) timeout = delta; } @@ -2738,7 +2813,7 @@ int __tcp_retransmit_skb(struct sock *sk, struct sk_buff *skb, int segs) /* Do not sent more than we queued. 1/4 is reserved for possible * copying overhead: fragmentation, tunneling, mangling etc. */ - if (atomic_read(&sk->sk_wmem_alloc) > + if (refcount_read(&sk->sk_wmem_alloc) > min_t(u32, sk->sk_wmem_queued + (sk->sk_wmem_queued >> 2), sk->sk_sndbuf)) return -EAGAIN; @@ -2803,7 +2878,7 @@ int __tcp_retransmit_skb(struct sock *sk, struct sk_buff *skb, int segs) skb_headroom(skb) >= 0xFFFF)) { struct sk_buff *nskb; - skb_mstamp_get(&skb->skb_mstamp); + skb->skb_mstamp = tp->tcp_mstamp; nskb = __pskb_copy(skb, MAX_TCP_HEADER, GFP_ATOMIC); err = nskb ? tcp_transmit_skb(sk, nskb, 0, GFP_ATOMIC) : -ENOBUFS; @@ -2878,6 +2953,10 @@ void tcp_xmit_retransmit_queue(struct sock *sk) if (skb == tcp_send_head(sk)) break; + + if (tcp_pacing_check(sk)) + break; + /* we could do better than to assign each time */ if (!hole) tp->retransmit_skb_hint = skb; @@ -3015,7 +3094,7 @@ void tcp_send_active_reset(struct sock *sk, gfp_t priority) skb_reserve(skb, MAX_TCP_HEADER); tcp_init_nondata_skb(skb, tcp_acceptable_seq(sk), TCPHDR_ACK | TCPHDR_RST); - skb_mstamp_get(&skb->skb_mstamp); + tcp_mstamp_refresh(tcp_sk(sk)); /* Send it off. */ if (tcp_transmit_skb(sk, skb, 0, priority)) NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPABORTFAILED); @@ -3111,10 +3190,10 @@ struct sk_buff *tcp_make_synack(const struct sock *sk, struct dst_entry *dst, memset(&opts, 0, sizeof(opts)); #ifdef CONFIG_SYN_COOKIES if (unlikely(req->cookie_ts)) - skb->skb_mstamp.stamp_jiffies = cookie_init_timestamp(req); + skb->skb_mstamp = cookie_init_timestamp(req); else #endif - skb_mstamp_get(&skb->skb_mstamp); + skb->skb_mstamp = tcp_clock_us(); #ifdef CONFIG_TCP_MD5SIG rcu_read_lock(); @@ -3134,6 +3213,7 @@ struct sk_buff *tcp_make_synack(const struct sock *sk, struct dst_entry *dst, tcp_ecn_make_synack(req, th); th->source = htons(ireq->ir_num); th->dest = ireq->ir_rmt_port; + skb->mark = ireq->ir_mark; /* Setting of flags are superfluous here for callers (and ECE is * not even correctly set) */ @@ -3189,12 +3269,14 @@ static void tcp_connect_init(struct sock *sk) const struct dst_entry *dst = __sk_dst_get(sk); struct tcp_sock *tp = tcp_sk(sk); __u8 rcv_wscale; + u32 rcv_wnd; /* We'll fix this up when we get a response from the other end. * See tcp_input.c:tcp_rcv_state_process case TCP_SYN_SENT. */ - tp->tcp_header_len = sizeof(struct tcphdr) + - (sysctl_tcp_timestamps ? TCPOLEN_TSTAMP_ALIGNED : 0); + tp->tcp_header_len = sizeof(struct tcphdr); + if (sock_net(sk)->ipv4.sysctl_tcp_timestamps) + tp->tcp_header_len += TCPOLEN_TSTAMP_ALIGNED; #ifdef CONFIG_TCP_MD5SIG if (tp->af_specific->md5_lookup(sk, sk)) @@ -3221,13 +3303,17 @@ static void tcp_connect_init(struct sock *sk) (tp->window_clamp > tcp_full_space(sk) || tp->window_clamp == 0)) tp->window_clamp = tcp_full_space(sk); + rcv_wnd = tcp_rwnd_init_bpf(sk); + if (rcv_wnd == 0) + rcv_wnd = dst_metric(dst, RTAX_INITRWND); + tcp_select_initial_window(tcp_full_space(sk), tp->advmss - (tp->rx_opt.ts_recent_stamp ? tp->tcp_header_len - sizeof(struct tcphdr) : 0), &tp->rcv_wnd, &tp->window_clamp, - sysctl_tcp_window_scaling, + sock_net(sk)->ipv4.sysctl_tcp_window_scaling, &rcv_wscale, - dst_metric(dst, RTAX_INITRWND)); + rcv_wnd); tp->rx_opt.rcv_wscale = rcv_wscale; tp->rcv_ssthresh = tp->rcv_wnd; @@ -3244,11 +3330,11 @@ static void tcp_connect_init(struct sock *sk) if (likely(!tp->repair)) tp->rcv_nxt = 0; else - tp->rcv_tstamp = tcp_time_stamp; + tp->rcv_tstamp = tcp_jiffies32; tp->rcv_wup = tp->rcv_nxt; tp->copied_seq = tp->rcv_nxt; - inet_csk(sk)->icsk_rto = TCP_TIMEOUT_INIT; + inet_csk(sk)->icsk_rto = tcp_timeout_init(sk); inet_csk(sk)->icsk_retransmits = 0; tcp_clear_retrans(tp); } @@ -3361,6 +3447,7 @@ int tcp_connect(struct sock *sk) struct sk_buff *buff; int err; + tcp_call_bpf(sk, BPF_SOCK_OPS_TCP_CONNECT_CB); tcp_connect_init(sk); if (unlikely(tp->repair)) { @@ -3373,7 +3460,8 @@ int tcp_connect(struct sock *sk) return -ENOBUFS; tcp_init_nondata_skb(buff, tp->write_seq++, TCPHDR_SYN); - tp->retrans_stamp = tcp_time_stamp; + tcp_mstamp_refresh(tp); + tp->retrans_stamp = tcp_time_stamp(tp); tcp_connect_queue_skb(sk, buff); tcp_ecn_send_syn(sk, buff); @@ -3492,7 +3580,6 @@ void tcp_send_ack(struct sock *sk) skb_set_tcp_pure_ack(buff); /* Send it off, this clears delayed acks for us. */ - skb_mstamp_get(&buff->skb_mstamp); tcp_transmit_skb(sk, buff, 0, (__force gfp_t)0); } EXPORT_SYMBOL_GPL(tcp_send_ack); @@ -3526,15 +3613,16 @@ static int tcp_xmit_probe_skb(struct sock *sk, int urgent, int mib) * send it. */ tcp_init_nondata_skb(skb, tp->snd_una - !urgent, TCPHDR_ACK); - skb_mstamp_get(&skb->skb_mstamp); NET_INC_STATS(sock_net(sk), mib); return tcp_transmit_skb(sk, skb, 0, (__force gfp_t)0); } +/* Called from setsockopt( ... TCP_REPAIR ) */ void tcp_send_window_probe(struct sock *sk) { if (sk->sk_state == TCP_ESTABLISHED) { tcp_sk(sk)->snd_wl1 = tcp_sk(sk)->rcv_nxt - 1; + tcp_mstamp_refresh(tcp_sk(sk)); tcp_xmit_probe_skb(sk, 0, LINUX_MIB_TCPWINPROBE); } } |