tcp: accecn: AccECN option send control

Instead of sending the option in every ACK, limit sending to
those ACKs where the option is necessary:
- Handshake
- "Change-triggered ACK" + the ACK following it. The
  2nd ACK is necessary to unambiguously indicate which
  of the ECN byte counters in increasing. The first
  ACK has two counters increasing due to the ecnfield
  edge.
- ACKs with CE to allow CEP delta validations to take
  advantage of the option.
- Force option to be sent every at least once per 2^22
  bytes. The check is done using the bit edges of the
  byte counters (avoids need for extra variables).
- AccECN option beacon to send a few times per RTT even if
  nothing in the ECN state requires that. The default is 3
  times per RTT, and its period can be set via
  sysctl_tcp_ecn_option_beacon.

Below are the pahole outcomes before and after this patch,
in which the group size of tcp_sock_write_tx is increased
from 89 to 97 due to the new u64 accecn_opt_tstamp member:

[BEFORE THIS PATCH]
struct tcp_sock {
    [...]
    u64                        tcp_wstamp_ns;        /*  2488     8 */
    struct list_head           tsorted_sent_queue;   /*  2496    16 */

    [...]
    __cacheline_group_end__tcp_sock_write_tx[0];     /*  2521     0 */
    __cacheline_group_begin__tcp_sock_write_txrx[0]; /*  2521     0 */
    u8                         nonagle:4;            /*  2521: 0  1 */
    u8                         rate_app_limited:1;   /*  2521: 4  1 */
    /* XXX 3 bits hole, try to pack */

    /* Force alignment to the next boundary: */
    u8                         :0;
    u8                         received_ce_pending:4;/*  2522: 0  1 */
    u8                         unused2:4;            /*  2522: 4  1 */
    u8                         accecn_minlen:2;      /*  2523: 0  1 */
    u8                         est_ecnfield:2;       /*  2523: 2  1 */
    u8                         unused3:4;            /*  2523: 4  1 */

    [...]
    __cacheline_group_end__tcp_sock_write_txrx[0];   /*  2628     0 */

    [...]
    /* size: 3200, cachelines: 50, members: 171 */
}

[AFTER THIS PATCH]
struct tcp_sock {
    [...]
    u64                        tcp_wstamp_ns;        /*  2488     8 */
    u64                        accecn_opt_tstamp;    /*  2596     8 */
    struct list_head           tsorted_sent_queue;   /*  2504    16 */

    [...]
    __cacheline_group_end__tcp_sock_write_tx[0];     /*  2529     0 */
    __cacheline_group_begin__tcp_sock_write_txrx[0]; /*  2529     0 */
    u8                         nonagle:4;            /*  2529: 0  1 */
    u8                         rate_app_limited:1;   /*  2529: 4  1 */
    /* XXX 3 bits hole, try to pack */

    /* Force alignment to the next boundary: */
    u8                         :0;
    u8                         received_ce_pending:4;/*  2530: 0  1 */
    u8                         unused2:4;            /*  2530: 4  1 */
    u8                         accecn_minlen:2;      /*  2531: 0  1 */
    u8                         est_ecnfield:2;       /*  2531: 2  1 */
    u8                         accecn_opt_demand:2;  /*  2531: 4  1 */
    u8                         prev_ecnfield:2;      /*  2531: 6  1 */

    [...]
    __cacheline_group_end__tcp_sock_write_txrx[0];   /*  2636     0 */

    [...]
    /* size: 3200, cachelines: 50, members: 173 */
}

Signed-off-by: Chia-Yu Chang <chia-yu.chang@nokia-bell-labs.com>
Co-developed-by: Ilpo Järvinen <ij@kernel.org>
Signed-off-by: Ilpo Järvinen <ij@kernel.org>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/20250916082434.100722-8-chia-yu.chang@nokia-bell-labs.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

4 files changed, 59 insertions, 1 deletions

diff --git a/include/linux/tcp.h b/include/linux/tcp.h
index 73557656cb2d..f637b659b35a 100644
--- a/include/linux/tcp.h
+++ b/include/linux/tcp.h
@@ -275,6 +275,7 @@ struct tcp_sock {
 	u32	mdev_us;	/* medium deviation			*/
 	u32	rtt_seq;	/* sequence number to update rttvar	*/
 	u64	tcp_wstamp_ns;	/* departure time for next sent data packet */
+	u64	accecn_opt_tstamp;	/* Last AccECN option sent timestamp */
 	struct list_head tsorted_sent_queue; /* time-sorted sent but un-SACKed skbs */
 	struct sk_buff *highest_sack;   /* skb just after the highest
 					 * skb with SACKed bit set
@@ -296,7 +297,8 @@ struct tcp_sock {
 		unused2:4;
 	u8	accecn_minlen:2,/* Minimum length of AccECN option sent */
 		est_ecnfield:2,/* ECN field for AccECN delivered estimates */
-		unused3:4;
+		accecn_opt_demand:2,/* Demand AccECN option for n next ACKs */
+		prev_ecnfield:2; /* ECN bits from the previous segment */
 	__be32	pred_flags;
 	u64	tcp_clock_cache; /* cache last tcp_clock_ns() (see tcp_mstamp_refresh()) */
 	u64	tcp_mstamp;	/* most recent packet received/sent */
diff --git a/include/net/netns/ipv4.h b/include/net/netns/ipv4.h
index acbb7dd497e1..34eb3aecb3f2 100644
--- a/include/net/netns/ipv4.h
+++ b/include/net/netns/ipv4.h
@@ -149,6 +149,7 @@ struct netns_ipv4 {
 
 	u8 sysctl_tcp_ecn;
 	u8 sysctl_tcp_ecn_option;
+	u8 sysctl_tcp_ecn_option_beacon;
 	u8 sysctl_tcp_ecn_fallback;
 
 	u8 sysctl_ip_default_ttl;
diff --git a/include/net/tcp.h b/include/net/tcp.h
index 6be29129465e..78dd7b8a4145 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -100,6 +100,9 @@ void tcp_time_wait(struct sock *sk, int state, int timeo);
 /* Maximal number of window scale according to RFC1323 */
 #define TCP_MAX_WSCALE		14U
 
+/* Default sending frequency of accurate ECN option per RTT */
+#define TCP_ACCECN_OPTION_BEACON	3
+
 /* urg_data states */
 #define TCP_URG_VALID	0x0100
 #define TCP_URG_NOTYET	0x0200
diff --git a/include/net/tcp_ecn.h b/include/net/tcp_ecn.h
index 08c7f4757e4e..133fb6b79500 100644
--- a/include/net/tcp_ecn.h
+++ b/include/net/tcp_ecn.h
@@ -176,6 +176,17 @@ static inline void tcp_accecn_third_ack(struct sock *sk,
 	}
 }
 
+/* Demand the minimum # to send AccECN optnio */
+static inline void tcp_accecn_opt_demand_min(struct sock *sk,
+					     u8 opt_demand_min)
+{
+	struct tcp_sock *tp = tcp_sk(sk);
+	u8 opt_demand;
+
+	opt_demand = max_t(u8, opt_demand_min, tp->accecn_opt_demand);
+	tp->accecn_opt_demand = opt_demand;
+}
+
 /* Maps IP ECN field ECT/CE code point to AccECN option field number, given
  * we are sending fields with Accurate ECN Order 1: ECT(1), CE, ECT(0).
  */
@@ -256,6 +267,7 @@ static inline void tcp_ecn_received_counters(struct sock *sk,
 	u8 ecnfield = TCP_SKB_CB(skb)->ip_dsfield & INET_ECN_MASK;
 	u8 is_ce = INET_ECN_is_ce(ecnfield);
 	struct tcp_sock *tp = tcp_sk(sk);
+	bool ecn_edge;
 
 	if (!INET_ECN_is_not_ect(ecnfield)) {
 		u32 pcount = is_ce * max_t(u16, 1, skb_shinfo(skb)->gso_segs);
@@ -274,9 +286,34 @@ static inline void tcp_ecn_received_counters(struct sock *sk,
 
 		if (len > 0) {
 			u8 minlen = tcp_ecnfield_to_accecn_optfield(ecnfield);
+			u32 oldbytes = tp->received_ecn_bytes[ecnfield - 1];
+			u32 bytes_mask = GENMASK_U32(31, 22);
+
 			tp->received_ecn_bytes[ecnfield - 1] += len;
 			tp->accecn_minlen = max_t(u8, tp->accecn_minlen,
 						  minlen);
+
+			/* Send AccECN option at least once per 2^22-byte
+			 * increase in any ECN byte counter.
+			 */
+			if ((tp->received_ecn_bytes[ecnfield - 1] ^ oldbytes) &
+			    bytes_mask) {
+				tcp_accecn_opt_demand_min(sk, 1);
+			}
+		}
+	}
+
+	ecn_edge = tp->prev_ecnfield != ecnfield;
+	if (ecn_edge || is_ce) {
+		tp->prev_ecnfield = ecnfield;
+		/* Demand Accurate ECN change-triggered ACKs. Two ACK are
+		 * demanded to indicate unambiguously the ecnfield value
+		 * in the latter ACK.
+		 */
+		if (tcp_ecn_mode_accecn(tp)) {
+			if (ecn_edge)
+				inet_csk(sk)->icsk_ack.pending |= ICSK_ACK_NOW;
+			tp->accecn_opt_demand = 2;
 		}
 	}
 }
@@ -349,6 +386,7 @@ static inline void tcp_accecn_init_counters(struct tcp_sock *tp)
 	__tcp_accecn_init_bytes_counters(tp->received_ecn_bytes);
 	__tcp_accecn_init_bytes_counters(tp->delivered_ecn_bytes);
 	tp->accecn_minlen = 0;
+	tp->accecn_opt_demand = 0;
 	tp->est_ecnfield = 0;
 }
 
@@ -431,6 +469,7 @@ static inline void tcp_ecn_rcv_synack(struct sock *sk, const struct tcphdr *th,
 	default:
 		tcp_ecn_mode_set(tp, TCP_ECN_MODE_ACCECN);
 		tp->syn_ect_rcv = ip_dsfield & INET_ECN_MASK;
+		tp->accecn_opt_demand = 2;
 		if (INET_ECN_is_ce(ip_dsfield) &&
 		    tcp_accecn_validate_syn_feedback(sk, ace,
 						     tp->syn_ect_snt)) {
@@ -451,6 +490,7 @@ static inline void tcp_ecn_rcv_syn(struct tcp_sock *tp, const struct tcphdr *th,
 		} else {
 			tp->syn_ect_rcv = TCP_SKB_CB(skb)->ip_dsfield &
 					  INET_ECN_MASK;
+			tp->prev_ecnfield = tp->syn_ect_rcv;
 			tcp_ecn_mode_set(tp, TCP_ECN_MODE_ACCECN);
 		}
 	}
@@ -542,4 +582,16 @@ tcp_ecn_make_synack(const struct request_sock *req, struct tcphdr *th)
 		th->ece = 1;
 }
 
+static inline bool tcp_accecn_option_beacon_check(const struct sock *sk)
+{
+	u32 ecn_beacon = READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_ecn_option_beacon);
+	const struct tcp_sock *tp = tcp_sk(sk);
+
+	if (!ecn_beacon)
+		return false;
+
+	return tcp_stamp_us_delta(tp->tcp_mstamp, tp->accecn_opt_tstamp) * ecn_beacon >=
+	       (tp->srtt_us >> 3);
+}
+
 #endif /* _LINUX_TCP_ECN_H */