bpf: Add more stats to HBM

Adds more stats to HBM, including average cwnd and rtt of all TCP
flows, percents of packets that are ecn ce marked and distribution
of return values.

Signed-off-by: Lawrence Brakmo <brakmo@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

samples/bpf/hbm.c

@@ -316,6 +316,14 @@ static int run_bpf_prog(char *prog, int cg_id)
 		double percent_pkts, percent_bytes;
 		char fname[100];
 		FILE *fout;
+		int k;
+		static const char *returnValNames[] = {
+			"DROP_PKT",
+			"ALLOW_PKT",
+			"DROP_PKT_CWR",
+			"ALLOW_PKT_CWR"
+		};
+#define RET_VAL_COUNT 4
 
 // Future support of ingress
 //		if (!outFlag)
@@ -350,6 +358,31 @@ static int run_bpf_prog(char *prog, int cg_id)
 			(qstats.bytes_total + 1);
 		fprintf(fout, "pkts_dropped_percent:%6.2f\n", percent_pkts);
 		fprintf(fout, "bytes_dropped_percent:%6.2f\n", percent_bytes);
+
+		// ECN CE markings
+		percent_pkts = (qstats.pkts_ecn_ce * 100.0) /
+			(qstats.pkts_total + 1);
+		fprintf(fout, "pkts_ecn_ce:%6.2f (%d)\n", percent_pkts,
+			(int)qstats.pkts_ecn_ce);
+
+		// Average cwnd
+		fprintf(fout, "avg cwnd:%d\n",
+			(int)(qstats.sum_cwnd / (qstats.sum_cwnd_cnt + 1)));
+		// Average rtt
+		fprintf(fout, "avg rtt:%d\n",
+			(int)(qstats.sum_rtt / (qstats.pkts_total + 1)));
+		// Average credit
+		fprintf(fout, "avg credit:%d\n",
+			(int)(qstats.sum_credit /
+			      (1500 * ((int)qstats.pkts_total) + 1)));
+
+		// Return values stats
+		for (k = 0; k < RET_VAL_COUNT; k++) {
+			percent_pkts = (qstats.returnValCount[k] * 100.0) /
+				(qstats.pkts_total + 1);
+			fprintf(fout, "%s:%6.2f (%d)\n", returnValNames[k],
+				percent_pkts, (int)qstats.returnValCount[k]);
+		}
 		fclose(fout);
 	}
 

samples/bpf/hbm.h

@@ -29,4 +29,10 @@ struct hbm_queue_stats {
 	unsigned long long bytes_total;
 	unsigned long long firstPacketTime;
 	unsigned long long lastPacketTime;
+	unsigned long long pkts_ecn_ce;
+	unsigned long long returnValCount[4];
+	unsigned long long sum_cwnd;
+	unsigned long long sum_rtt;
+	unsigned long long sum_cwnd_cnt;
+	long long sum_credit;
 };

samples/bpf/hbm_kern.h

@@ -65,17 +65,43 @@ struct bpf_map_def SEC("maps") queue_stats = {
 BPF_ANNOTATE_KV_PAIR(queue_stats, int, struct hbm_queue_stats);
 
 struct hbm_pkt_info {
+	int	cwnd;
+	int	rtt;
 	bool	is_ip;
 	bool	is_tcp;
 	short	ecn;
 };
 
+static int get_tcp_info(struct __sk_buff *skb, struct hbm_pkt_info *pkti)
+{
+	struct bpf_sock *sk;
+	struct bpf_tcp_sock *tp;
+
+	sk = skb->sk;
+	if (sk) {
+		sk = bpf_sk_fullsock(sk);
+		if (sk) {
+			if (sk->protocol == IPPROTO_TCP) {
+				tp = bpf_tcp_sock(sk);
+				if (tp) {
+					pkti->cwnd = tp->snd_cwnd;
+					pkti->rtt = tp->srtt_us >> 3;
+					return 0;
+				}
+			}
+		}
+	}
+	return 1;
+}
+
 static __always_inline void hbm_get_pkt_info(struct __sk_buff *skb,
 					     struct hbm_pkt_info *pkti)
 {
 	struct iphdr iph;
 	struct ipv6hdr *ip6h;
 
+	pkti->cwnd = 0;
+	pkti->rtt = 0;
 	bpf_skb_load_bytes(skb, 0, &iph, 12);
 	if (iph.version == 6) {
 		ip6h = (struct ipv6hdr *)&iph;
@@ -91,6 +117,8 @@ static __always_inline void hbm_get_pkt_info(struct __sk_buff *skb,
 		pkti->is_tcp = false;
 		pkti->ecn = 0;
 	}
+	if (pkti->is_tcp)
+		get_tcp_info(skb, pkti);
 }
 
 static __always_inline void hbm_init_vqueue(struct hbm_vqueue *qdp, int rate)
@@ -105,8 +133,14 @@ static __always_inline void hbm_update_stats(struct hbm_queue_stats *qsp,
 					     int len,
 					     unsigned long long curtime,
 					     bool congestion_flag,
-					     bool drop_flag)
+					     bool drop_flag,
+					     bool cwr_flag,
+					     bool ecn_ce_flag,
+					     struct hbm_pkt_info *pkti,
+					     int credit)
 {
+	int rv = ALLOW_PKT;
+
 	if (qsp != NULL) {
 		// Following is needed for work conserving
 		__sync_add_and_fetch(&(qsp->bytes_total), len);
@@ -116,7 +150,7 @@ static __always_inline void hbm_update_stats(struct hbm_queue_stats *qsp,
 				qsp->firstPacketTime = curtime;
 			qsp->lastPacketTime = curtime;
 			__sync_add_and_fetch(&(qsp->pkts_total), 1);
-			if (congestion_flag || drop_flag) {
+			if (congestion_flag) {
 				__sync_add_and_fetch(&(qsp->pkts_marked), 1);
 				__sync_add_and_fetch(&(qsp->bytes_marked), len);
 			}
@@ -125,6 +159,34 @@ static __always_inline void hbm_update_stats(struct hbm_queue_stats *qsp,
 				__sync_add_and_fetch(&(qsp->bytes_dropped),
 						     len);
 			}
+			if (ecn_ce_flag)
+				__sync_add_and_fetch(&(qsp->pkts_ecn_ce), 1);
+			if (pkti->cwnd) {
+				__sync_add_and_fetch(&(qsp->sum_cwnd),
+						     pkti->cwnd);
+				__sync_add_and_fetch(&(qsp->sum_cwnd_cnt), 1);
+			}
+			if (pkti->rtt)
+				__sync_add_and_fetch(&(qsp->sum_rtt),
+						     pkti->rtt);
+			__sync_add_and_fetch(&(qsp->sum_credit), credit);
+
+			if (drop_flag)
+				rv = DROP_PKT;
+			if (cwr_flag)
+				rv |= 2;
+			if (rv == DROP_PKT)
+				__sync_add_and_fetch(&(qsp->returnValCount[0]),
+						     1);
+			else if (rv == ALLOW_PKT)
+				__sync_add_and_fetch(&(qsp->returnValCount[1]),
+						     1);
+			else if (rv == 2)
+				__sync_add_and_fetch(&(qsp->returnValCount[2]),
+						     1);
+			else if (rv == 3)
+				__sync_add_and_fetch(&(qsp->returnValCount[3]),
+						     1);
 		}
 	}
 }

samples/bpf/hbm_out_kern.c

@@ -62,11 +62,12 @@ int _hbm_out_cg(struct __sk_buff *skb)
 	unsigned int queue_index = 0;
 	unsigned long long curtime;
 	int credit;
-	signed long long delta = 0, zero = 0;
+	signed long long delta = 0, new_credit;
 	int max_credit = MAX_CREDIT;
 	bool congestion_flag = false;
 	bool drop_flag = false;
 	bool cwr_flag = false;
+	bool ecn_ce_flag = false;
 	struct hbm_vqueue *qdp;
 	struct hbm_queue_stats *qsp = NULL;
 	int rv = ALLOW_PKT;
@@ -99,9 +100,11 @@ int _hbm_out_cg(struct __sk_buff *skb)
 	 */
 	if (delta > 0) {
 		qdp->lasttime = curtime;
-		credit += CREDIT_PER_NS(delta, qdp->rate);
+		new_credit = credit + CREDIT_PER_NS(delta, qdp->rate);
-		if (credit > MAX_CREDIT)
+		if (new_credit > MAX_CREDIT)
 			credit = MAX_CREDIT;
+		else
+			credit = new_credit;
 	}
 	credit -= len;
 	qdp->credit = credit;
@@ -139,7 +142,9 @@ int _hbm_out_cg(struct __sk_buff *skb)
 	}
 
 	if (congestion_flag) {
-		if (!bpf_skb_ecn_set_ce(skb)) {
+		if (bpf_skb_ecn_set_ce(skb)) {
+			ecn_ce_flag = true;
+		} else {
 			if (pkti.is_tcp) {
 				unsigned int rand = bpf_get_prandom_u32();
 
@@ -155,16 +160,17 @@ int _hbm_out_cg(struct __sk_buff *skb)
 		}
 	}
 
-	if (drop_flag)
-		rv = DROP_PKT;
 	if (qsp != NULL)
 		if (qsp->no_cn)
 			cwr_flag = false;
 
-	hbm_update_stats(qsp, len, curtime, congestion_flag, drop_flag);
+	hbm_update_stats(qsp, len, curtime, congestion_flag, drop_flag,
+			 cwr_flag, ecn_ce_flag, &pkti, credit);
 
-	if (rv == DROP_PKT)
+	if (drop_flag) {
 		__sync_add_and_fetch(&(qdp->credit), len);
+		rv = DROP_PKT;
+	}
 
 	if (cwr_flag)
 		rv |= 2;