git: b1258b76435a - main - tcp: add conservative d.cep accounting algorithm

The branch main has been updated by rscheff:

URL: https://cgit.FreeBSD.org/src/commit/?id=b1258b76435ac370ddd0f814a351779ddb267f6f

commit b1258b76435ac370ddd0f814a351779ddb267f6f
Author:     Richard Scheffenegger <rscheff@FreeBSD.org>
AuthorDate: 2022-11-06 10:59:55 +0000
Commit:     Richard Scheffenegger <rscheff@FreeBSD.org>
CommitDate: 2022-11-06 11:05:22 +0000

    tcp: add conservative d.cep accounting algorithm
    
    Accurate ECN asks to conservatively estimate, when the
    ACE counter may have wrapped due to a single ACK covering a larger
    number of segments. This is described in Annex A.2 of the
    accurate-ecn draft.
    
    Event:                  IETF 115 Hackathon
    Reviewed By:            tuexen, #transport
    Sponsored by:           NetApp, Inc.
    Differential Revision:  https://reviews.freebsd.org/D37281
---
 sys/netinet/tcp_ecn.c         | 17 ++++++++++-------
 sys/netinet/tcp_ecn.h         |  2 +-
 sys/netinet/tcp_input.c       |  4 +++-
 sys/netinet/tcp_stacks/rack.c | 10 +++++++---
 sys/netinet/tcp_var.h         |  6 ++++++
 5 files changed, 27 insertions(+), 12 deletions(-)

diff --git a/sys/netinet/tcp_ecn.c b/sys/netinet/tcp_ecn.c
index 1d693944ac40..40791172e55f 100644
--- a/sys/netinet/tcp_ecn.c
+++ b/sys/netinet/tcp_ecn.c
@@ -271,9 +271,9 @@ tcp_ecn_input_parallel_syn(struct tcpcb *tp, uint16_t thflags, int iptos)
  * TCP ECN processing.
  */
 int
-tcp_ecn_input_segment(struct tcpcb *tp, uint16_t thflags, int iptos)
+tcp_ecn_input_segment(struct tcpcb *tp, uint16_t thflags, int tlen, int pkts, int iptos)
 {
-	int delta_ace = 0;
+	int delta_cep = 0;
 
 	if (tp->t_flags2 & (TF2_ECN_PERMIT | TF2_ACE_PERMIT)) {
 		switch (iptos & IPTOS_ECN_MASK) {
@@ -292,9 +292,12 @@ tcp_ecn_input_segment(struct tcpcb *tp, uint16_t thflags, int iptos)
 			if ((iptos & IPTOS_ECN_MASK) == IPTOS_ECN_CE)
 				tp->t_rcep += 1;
 			if (tp->t_flags2 & TF2_ECN_PERMIT) {
-				delta_ace = (tcp_ecn_get_ace(thflags) + 8 -
-					    (tp->t_scep & 0x07)) & 0x07;
-				tp->t_scep += delta_ace;
+				delta_cep = (tcp_ecn_get_ace(thflags) + 8 -
+					    (tp->t_scep & 7)) & 7;
+				if (delta_cep < pkts)
+					delta_cep = pkts -
+					    ((pkts - delta_cep) & 7);
+				tp->t_scep += delta_cep;
 			} else {
 				/*
 				 * process the final ACK of the 3WHS
@@ -326,7 +329,7 @@ tcp_ecn_input_segment(struct tcpcb *tp, uint16_t thflags, int iptos)
 		} else {
 			/* RFC3168 ECN handling */
 			if ((thflags & (TH_SYN | TH_ECE)) == TH_ECE) {
-				delta_ace = 1;
+				delta_cep = 1;
 				tp->t_scep++;
 			}
 			if (thflags & TH_CWR) {
@@ -341,7 +344,7 @@ tcp_ecn_input_segment(struct tcpcb *tp, uint16_t thflags, int iptos)
 		cc_ecnpkt_handler_flags(tp, thflags, iptos);
 	}
 
-	return delta_ace;
+	return delta_cep;
 }
 
 /*
diff --git a/sys/netinet/tcp_ecn.h b/sys/netinet/tcp_ecn.h
index deade12b75d1..3cc7715b9562 100644
--- a/sys/netinet/tcp_ecn.h
+++ b/sys/netinet/tcp_ecn.h
@@ -43,7 +43,7 @@
 
 void	 tcp_ecn_input_syn_sent(struct tcpcb *, uint16_t, int);
 void	 tcp_ecn_input_parallel_syn(struct tcpcb *, uint16_t, int);
-int	 tcp_ecn_input_segment(struct tcpcb *, uint16_t, int);
+int	 tcp_ecn_input_segment(struct tcpcb *, uint16_t, int, int, int);
 uint16_t tcp_ecn_output_syn_sent(struct tcpcb *);
 int	 tcp_ecn_output_established(struct tcpcb *, uint16_t *, int, bool);
 void	 tcp_ecn_syncache_socket(struct tcpcb *, struct syncache *);
diff --git a/sys/netinet/tcp_input.c b/sys/netinet/tcp_input.c
index 84574aaa00ae..a1787a0f93db 100644
--- a/sys/netinet/tcp_input.c
+++ b/sys/netinet/tcp_input.c
@@ -1627,7 +1627,9 @@ tcp_do_segment(struct mbuf *m, struct tcphdr *th, struct socket *so,
 	/*
 	 * TCP ECN processing.
 	 */
-	if (tcp_ecn_input_segment(tp, thflags, iptos))
+	if (tcp_ecn_input_segment(tp, thflags, tlen,
+	    tcp_packets_this_ack(tp, th->th_ack),
+	    iptos))
 		cc_cong_signal(tp, th, CC_ECN);
 
 	/*
diff --git a/sys/netinet/tcp_stacks/rack.c b/sys/netinet/tcp_stacks/rack.c
index fdac23d0c5cc..e99f104bfa29 100644
--- a/sys/netinet/tcp_stacks/rack.c
+++ b/sys/netinet/tcp_stacks/rack.c
@@ -13528,8 +13528,10 @@ rack_do_compressed_ack_processing(struct tcpcb *tp, struct socket *so, struct mb
 			rack_cc_after_idle(rack, tp);
 		}
 		tp->t_rcvtime = ticks;
-		/* Now what about ECN? */
-		if (tcp_ecn_input_segment(tp, ae->flags, ae->codepoint))
+		/* Now what about ECN of a chain of pure ACKs? */
+		if (tcp_ecn_input_segment(tp, ae->flags, 0,
+			tcp_packets_this_ack(tp, ae->ack),
+			ae->codepoint))
 			rack_cong_signal(tp, CC_ECN, ae->ack, __LINE__);
 #ifdef TCP_ACCOUNTING
 		/* Count for the specific type of ack in */
@@ -14320,7 +14322,9 @@ rack_do_segment_nounlock(struct mbuf *m, struct tcphdr *th, struct socket *so,
 	 * TCP ECN processing. XXXJTL: If we ever use ECN, we need to move
 	 * this to occur after we've validated the segment.
 	 */
-	if (tcp_ecn_input_segment(tp, thflags, iptos))
+	if (tcp_ecn_input_segment(tp, thflags, tlen,
+	    tcp_packets_this_ack(tp, th->th_ack),
+	    iptos))
 		rack_cong_signal(tp, CC_ECN, th->th_ack, __LINE__);
 
 	/*
diff --git a/sys/netinet/tcp_var.h b/sys/netinet/tcp_var.h
index d115a18d66d5..01deeaad58cf 100644
--- a/sys/netinet/tcp_var.h
+++ b/sys/netinet/tcp_var.h
@@ -551,6 +551,12 @@ tcp_unlock_or_drop(struct tcpcb *tp, int tcp_output_retval)
 #endif
 
 #define	BYTES_THIS_ACK(tp, th)	(th->th_ack - tp->snd_una)
+static int inline
+tcp_packets_this_ack(struct tcpcb *tp, tcp_seq ack)
+{
+	return ((ack - tp->snd_una) / tp->t_maxseg +
+		((((ack - tp->snd_una) % tp->t_maxseg) != 0) ? 1 : 0));
+}
 
 /*
  * Flags for the t_oobflags field.