--- zzzz-none-000/linux-3.10.107/net/core/gen_stats.c	2017-06-27 09:49:32.000000000 +0000
+++ scorpion-7490-727/linux-3.10.107/net/core/gen_stats.c	2021-02-04 17:41:59.000000000 +0000
@@ -100,6 +100,43 @@
 }
 EXPORT_SYMBOL(gnet_stats_start_copy);
 
+static void
+__gnet_stats_copy_basic_cpu(struct gnet_stats_basic_packed *bstats,
+			    struct gnet_stats_basic_cpu __percpu *cpu)
+{
+	int i;
+
+	for_each_possible_cpu(i) {
+		struct gnet_stats_basic_cpu *bcpu = per_cpu_ptr(cpu, i);
+		unsigned int start;
+		u64 bytes;
+		u32 packets;
+
+		do {
+			start = u64_stats_fetch_begin_irq(&bcpu->syncp);
+			bytes = bcpu->bstats.bytes;
+			packets = bcpu->bstats.packets;
+		} while (u64_stats_fetch_retry_irq(&bcpu->syncp, start));
+
+		bstats->bytes += bytes;
+		bstats->packets += packets;
+	}
+}
+
+void
+__gnet_stats_copy_basic(struct gnet_stats_basic_packed *bstats,
+			struct gnet_stats_basic_cpu __percpu *cpu,
+			struct gnet_stats_basic_packed *b)
+{
+	if (cpu) {
+		__gnet_stats_copy_basic_cpu(bstats, cpu);
+	} else {
+		bstats->bytes = b->bytes;
+		bstats->packets = b->packets;
+	}
+}
+EXPORT_SYMBOL(__gnet_stats_copy_basic);
+
 /**
  * gnet_stats_copy_basic - copy basic statistics into statistic TLV
  * @d: dumping handle
@@ -112,19 +149,25 @@
  * if the room in the socket buffer was not sufficient.
  */
 int
-gnet_stats_copy_basic(struct gnet_dump *d, struct gnet_stats_basic_packed *b)
+gnet_stats_copy_basic(struct gnet_dump *d,
+		      struct gnet_stats_basic_cpu __percpu *cpu,
+		      struct gnet_stats_basic_packed *b)
 {
+	struct gnet_stats_basic_packed bstats = {0};
+
+	__gnet_stats_copy_basic(&bstats, cpu, b);
+
 	if (d->compat_tc_stats) {
-		d->tc_stats.bytes = b->bytes;
-		d->tc_stats.packets = b->packets;
+		d->tc_stats.bytes = bstats.bytes;
+		d->tc_stats.packets = bstats.packets;
 	}
 
 	if (d->tail) {
 		struct gnet_stats_basic sb;
 
 		memset(&sb, 0, sizeof(sb));
-		sb.bytes = b->bytes;
-		sb.packets = b->packets;
+		sb.bytes = bstats.bytes;
+		sb.packets = bstats.packets;
 		return gnet_stats_copy(d, TCA_STATS_BASIC, &sb, sizeof(sb));
 	}
 	return 0;
@@ -146,46 +189,103 @@
 int
 gnet_stats_copy_rate_est(struct gnet_dump *d,
 			 const struct gnet_stats_basic_packed *b,
-			 struct gnet_stats_rate_est *r)
+			 struct gnet_stats_rate_est64 *r)
 {
+	struct gnet_stats_rate_est est;
+	int res;
+
 	if (b && !gen_estimator_active(b, r))
 		return 0;
 
+	est.bps = min_t(u64, UINT_MAX, r->bps);
+	/* we have some time before reaching 2^32 packets per second */
+	est.pps = r->pps;
+
 	if (d->compat_tc_stats) {
-		d->tc_stats.bps = r->bps;
-		d->tc_stats.pps = r->pps;
+		d->tc_stats.bps = est.bps;
+		d->tc_stats.pps = est.pps;
 	}
 
-	if (d->tail)
-		return gnet_stats_copy(d, TCA_STATS_RATE_EST, r, sizeof(*r));
+	if (d->tail) {
+		res = gnet_stats_copy(d, TCA_STATS_RATE_EST, &est, sizeof(est));
+		if (res < 0 || est.bps == r->bps)
+			return res;
+		/* emit 64bit stats only if needed */
+		return gnet_stats_copy(d, TCA_STATS_RATE_EST64, r, sizeof(*r));
+	}
 
 	return 0;
 }
 EXPORT_SYMBOL(gnet_stats_copy_rate_est);
 
+static void
+__gnet_stats_copy_queue_cpu(struct gnet_stats_queue *qstats,
+			    const struct gnet_stats_queue __percpu *q)
+{
+	int i;
+
+	for_each_possible_cpu(i) {
+		const struct gnet_stats_queue *qcpu = per_cpu_ptr(q, i);
+
+		qstats->qlen = 0;
+		qstats->backlog += qcpu->backlog;
+		qstats->drops += qcpu->drops;
+		qstats->requeues += qcpu->requeues;
+		qstats->overlimits += qcpu->overlimits;
+	}
+}
+
+static void __gnet_stats_copy_queue(struct gnet_stats_queue *qstats,
+				    const struct gnet_stats_queue __percpu *cpu,
+				    const struct gnet_stats_queue *q,
+				    __u32 qlen)
+{
+	if (cpu) {
+		__gnet_stats_copy_queue_cpu(qstats, cpu);
+	} else {
+		qstats->qlen = q->qlen;
+		qstats->backlog = q->backlog;
+		qstats->drops = q->drops;
+		qstats->requeues = q->requeues;
+		qstats->overlimits = q->overlimits;
+	}
+
+	qstats->qlen = qlen;
+}
+
 /**
  * gnet_stats_copy_queue - copy queue statistics into statistics TLV
  * @d: dumping handle
+ * @cpu_q: per cpu queue statistics
  * @q: queue statistics
+ * @qlen: queue length statistics
  *
  * Appends the queue statistics to the top level TLV created by
- * gnet_stats_start_copy().
+ * gnet_stats_start_copy(). Using per cpu queue statistics if
+ * they are available.
  *
  * Returns 0 on success or -1 with the statistic lock released
  * if the room in the socket buffer was not sufficient.
  */
 int
-gnet_stats_copy_queue(struct gnet_dump *d, struct gnet_stats_queue *q)
+gnet_stats_copy_queue(struct gnet_dump *d,
+		      struct gnet_stats_queue __percpu *cpu_q,
+		      struct gnet_stats_queue *q, __u32 qlen)
 {
+	struct gnet_stats_queue qstats = {0};
+
+	__gnet_stats_copy_queue(&qstats, cpu_q, q, qlen);
+
 	if (d->compat_tc_stats) {
-		d->tc_stats.drops = q->drops;
-		d->tc_stats.qlen = q->qlen;
-		d->tc_stats.backlog = q->backlog;
-		d->tc_stats.overlimits = q->overlimits;
+		d->tc_stats.drops = qstats.drops;
+		d->tc_stats.qlen = qstats.qlen;
+		d->tc_stats.backlog = qstats.backlog;
+		d->tc_stats.overlimits = qstats.overlimits;
 	}
 
 	if (d->tail)
-		return gnet_stats_copy(d, TCA_STATS_QUEUE, q, sizeof(*q));
+		return gnet_stats_copy(d, TCA_STATS_QUEUE,
+				       &qstats, sizeof(qstats));
 
 	return 0;
 }
@@ -197,7 +297,7 @@
  * @st: application specific statistics data
  * @len: length of data
  *
- * Appends the application sepecific statistics to the top level TLV created by
+ * Appends the application specific statistics to the top level TLV created by
  * gnet_stats_start_copy() and remembers the data for XSTATS if the dumping
  * handle is in backward compatibility mode.
  *