taprio: Add support for hardware offloading

This allows taprio to offload the schedule enforcement to capable
network cards, resulting in more precise windows and less CPU usage.

The gate mask acts on traffic classes (groups of queues of same
priority), as specified in IEEE 802.1Q-2018, and following the existing
taprio and mqprio semantics.
It is up to the driver to perform conversion between tc and individual
netdev queues if for some reason it needs to make that distinction.

Full offload is requested from the network interface by specifying
"flags 2" in the tc qdisc creation command, which in turn corresponds to
the TCA_TAPRIO_ATTR_FLAG_FULL_OFFLOAD bit.

The important detail here is the clockid which is implicitly /dev/ptpN
for full offload, and hence not configurable.

A reference counting API is added to support the use case where Ethernet
drivers need to keep the taprio offload structure locally (i.e. they are
a multi-port switch driver, and configuring a port depends on the
settings of other ports as well). The refcount_t variable is kept in a
private structure (__tc_taprio_qopt_offload) and not exposed to drivers.

In the future, the private structure might also be expanded with a
backpointer to taprio_sched *q, to implement the notification system
described in the patch (of when admin became oper, or an error occurred,
etc, so the offload can be monitored with 'tc qdisc show').
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Signed-off-by: Voon Weifeng <weifeng.voon@intel.com>
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

include/linux/netdevice.h

@@ -847,6 +847,7 @@ enum tc_setup_type {
 	TC_SETUP_QDISC_ETF,
 	TC_SETUP_ROOT_QDISC,
 	TC_SETUP_QDISC_GRED,
+	TC_SETUP_QDISC_TAPRIO,
 };
 
 /* These structures hold the attributes of bpf state that are being passed

include/net/pkt_sched.h

@@ -161,4 +161,27 @@ struct tc_etf_qopt_offload {
 	s32 queue;
 };
 
+struct tc_taprio_sched_entry {
+	u8 command; /* TC_TAPRIO_CMD_* */
+
+	/* The gate_mask in the offloading side refers to traffic classes */
+	u32 gate_mask;
+	u32 interval;
+};
+
+struct tc_taprio_qopt_offload {
+	u8 enable;
+	ktime_t base_time;
+	u64 cycle_time;
+	u64 cycle_time_extension;
+
+	size_t num_entries;
+	struct tc_taprio_sched_entry entries[0];
+};
+
+/* Reference counting */
+struct tc_taprio_qopt_offload *taprio_offload_get(struct tc_taprio_qopt_offload
+						  *offload);
+void taprio_offload_free(struct tc_taprio_qopt_offload *offload);
+
 #endif

include/uapi/linux/pkt_sched.h

@@ -1160,7 +1160,8 @@ enum {
  *       [TCA_TAPRIO_ATTR_SCHED_ENTRY_INTERVAL]
  */
 
-#define TCA_TAPRIO_ATTR_FLAG_TXTIME_ASSIST 0x1
+#define TCA_TAPRIO_ATTR_FLAG_TXTIME_ASSIST	BIT(0)
+#define TCA_TAPRIO_ATTR_FLAG_FULL_OFFLOAD	BIT(1)
 
 enum {
 	TCA_TAPRIO_ATTR_UNSPEC,

net/sched/sch_taprio.c

@@ -29,8 +29,8 @@ static DEFINE_SPINLOCK(taprio_list_lock);
 
 #define TAPRIO_ALL_GATES_OPEN -1
 
-#define FLAGS_VALID(flags) (!((flags) & ~TCA_TAPRIO_ATTR_FLAG_TXTIME_ASSIST))
 #define TXTIME_ASSIST_IS_ENABLED(flags) ((flags) & TCA_TAPRIO_ATTR_FLAG_TXTIME_ASSIST)
+#define FULL_OFFLOAD_IS_ENABLED(flags) ((flags) & TCA_TAPRIO_ATTR_FLAG_FULL_OFFLOAD)
 
 struct sched_entry {
 	struct list_head list;
@@ -75,9 +75,16 @@ struct taprio_sched {
 	struct sched_gate_list __rcu *admin_sched;
 	struct hrtimer advance_timer;
 	struct list_head taprio_list;
+	struct sk_buff *(*dequeue)(struct Qdisc *sch);
+	struct sk_buff *(*peek)(struct Qdisc *sch);
 	u32 txtime_delay;
 };
 
+struct __tc_taprio_qopt_offload {
+	refcount_t users;
+	struct tc_taprio_qopt_offload offload;
+};
+
 static ktime_t sched_base_time(const struct sched_gate_list *sched)
 {
 	if (!sched)
@@ -268,6 +275,19 @@ static bool is_valid_interval(struct sk_buff *skb, struct Qdisc *sch)
 	return entry;
 }
 
+static bool taprio_flags_valid(u32 flags)
+{
+	/* Make sure no other flag bits are set. */
+	if (flags & ~(TCA_TAPRIO_ATTR_FLAG_TXTIME_ASSIST |
+		      TCA_TAPRIO_ATTR_FLAG_FULL_OFFLOAD))
+		return false;
+	/* txtime-assist and full offload are mutually exclusive */
+	if ((flags & TCA_TAPRIO_ATTR_FLAG_TXTIME_ASSIST) &&
+	    (flags & TCA_TAPRIO_ATTR_FLAG_FULL_OFFLOAD))
+		return false;
+	return true;
+}
+
 /* This returns the tstamp value set by TCP in terms of the set clock. */
 static ktime_t get_tcp_tstamp(struct taprio_sched *q, struct sk_buff *skb)
 {
@@ -417,7 +437,7 @@ static int taprio_enqueue(struct sk_buff *skb, struct Qdisc *sch,
 	return qdisc_enqueue(skb, child, to_free);
 }
 
-static struct sk_buff *taprio_peek(struct Qdisc *sch)
+static struct sk_buff *taprio_peek_soft(struct Qdisc *sch)
 {
 	struct taprio_sched *q = qdisc_priv(sch);
 	struct net_device *dev = qdisc_dev(sch);
@@ -461,6 +481,36 @@ static struct sk_buff *taprio_peek(struct Qdisc *sch)
 	return NULL;
 }
 
+static struct sk_buff *taprio_peek_offload(struct Qdisc *sch)
+{
+	struct taprio_sched *q = qdisc_priv(sch);
+	struct net_device *dev = qdisc_dev(sch);
+	struct sk_buff *skb;
+	int i;
+
+	for (i = 0; i < dev->num_tx_queues; i++) {
+		struct Qdisc *child = q->qdiscs[i];
+
+		if (unlikely(!child))
+			continue;
+
+		skb = child->ops->peek(child);
+		if (!skb)
+			continue;
+
+		return skb;
+	}
+
+	return NULL;
+}
+
+static struct sk_buff *taprio_peek(struct Qdisc *sch)
+{
+	struct taprio_sched *q = qdisc_priv(sch);
+
+	return q->peek(sch);
+}
+
 static void taprio_set_budget(struct taprio_sched *q, struct sched_entry *entry)
 {
 	atomic_set(&entry->budget,
@@ -468,7 +518,7 @@ static void taprio_set_budget(struct taprio_sched *q, struct sched_entry *entry)
 			     atomic64_read(&q->picos_per_byte)));
 }
 
-static struct sk_buff *taprio_dequeue(struct Qdisc *sch)
+static struct sk_buff *taprio_dequeue_soft(struct Qdisc *sch)
 {
 	struct taprio_sched *q = qdisc_priv(sch);
 	struct net_device *dev = qdisc_dev(sch);
@@ -550,6 +600,40 @@ static struct sk_buff *taprio_dequeue(struct Qdisc *sch)
 	return skb;
 }
 
+static struct sk_buff *taprio_dequeue_offload(struct Qdisc *sch)
+{
+	struct taprio_sched *q = qdisc_priv(sch);
+	struct net_device *dev = qdisc_dev(sch);
+	struct sk_buff *skb;
+	int i;
+
+	for (i = 0; i < dev->num_tx_queues; i++) {
+		struct Qdisc *child = q->qdiscs[i];
+
+		if (unlikely(!child))
+			continue;
+
+		skb = child->ops->dequeue(child);
+		if (unlikely(!skb))
+			continue;
+
+		qdisc_bstats_update(sch, skb);
+		qdisc_qstats_backlog_dec(sch, skb);
+		sch->q.qlen--;
+
+		return skb;
+	}
+
+	return NULL;
+}
+
+static struct sk_buff *taprio_dequeue(struct Qdisc *sch)
+{
+	struct taprio_sched *q = qdisc_priv(sch);
+
+	return q->dequeue(sch);
+}
+
 static bool should_restart_cycle(const struct sched_gate_list *oper,
 				 const struct sched_entry *entry)
 {
@@ -932,6 +1016,9 @@ static void taprio_start_sched(struct Qdisc *sch,
 	struct taprio_sched *q = qdisc_priv(sch);
 	ktime_t expires;
 
+	if (FULL_OFFLOAD_IS_ENABLED(q->flags))
+		return;
+
 	expires = hrtimer_get_expires(&q->advance_timer);
 	if (expires == 0)
 		expires = KTIME_MAX;
@@ -1011,6 +1098,254 @@ static void setup_txtime(struct taprio_sched *q,
 	}
 }
 
+static struct tc_taprio_qopt_offload *taprio_offload_alloc(int num_entries)
+{
+	size_t size = sizeof(struct tc_taprio_sched_entry) * num_entries +
+		      sizeof(struct __tc_taprio_qopt_offload);
+	struct __tc_taprio_qopt_offload *__offload;
+
+	__offload = kzalloc(size, GFP_KERNEL);
+	if (!__offload)
+		return NULL;
+
+	refcount_set(&__offload->users, 1);
+
+	return &__offload->offload;
+}
+
+struct tc_taprio_qopt_offload *taprio_offload_get(struct tc_taprio_qopt_offload
+						  *offload)
+{
+	struct __tc_taprio_qopt_offload *__offload;
+
+	__offload = container_of(offload, struct __tc_taprio_qopt_offload,
+				 offload);
+
+	refcount_inc(&__offload->users);
+
+	return offload;
+}
+EXPORT_SYMBOL_GPL(taprio_offload_get);
+
+void taprio_offload_free(struct tc_taprio_qopt_offload *offload)
+{
+	struct __tc_taprio_qopt_offload *__offload;
+
+	__offload = container_of(offload, struct __tc_taprio_qopt_offload,
+				 offload);
+
+	if (!refcount_dec_and_test(&__offload->users))
+		return;
+
+	kfree(__offload);
+}
+EXPORT_SYMBOL_GPL(taprio_offload_free);
+
+/* The function will only serve to keep the pointers to the "oper" and "admin"
+ * schedules valid in relation to their base times, so when calling dump() the
+ * users looks at the right schedules.
+ * When using full offload, the admin configuration is promoted to oper at the
+ * base_time in the PHC time domain.  But because the system time is not
+ * necessarily in sync with that, we can't just trigger a hrtimer to call
+ * switch_schedules at the right hardware time.
+ * At the moment we call this by hand right away from taprio, but in the future
+ * it will be useful to create a mechanism for drivers to notify taprio of the
+ * offload state (PENDING, ACTIVE, INACTIVE) so it can be visible in dump().
+ * This is left as TODO.
+ */
+void taprio_offload_config_changed(struct taprio_sched *q)
+{
+	struct sched_gate_list *oper, *admin;
+
+	spin_lock(&q->current_entry_lock);
+
+	oper = rcu_dereference_protected(q->oper_sched,
+					 lockdep_is_held(&q->current_entry_lock));
+	admin = rcu_dereference_protected(q->admin_sched,
+					  lockdep_is_held(&q->current_entry_lock));
+
+	switch_schedules(q, &admin, &oper);
+
+	spin_unlock(&q->current_entry_lock);
+}
+
+static void taprio_sched_to_offload(struct taprio_sched *q,
+				    struct sched_gate_list *sched,
+				    const struct tc_mqprio_qopt *mqprio,
+				    struct tc_taprio_qopt_offload *offload)
+{
+	struct sched_entry *entry;
+	int i = 0;
+
+	offload->base_time = sched->base_time;
+	offload->cycle_time = sched->cycle_time;
+	offload->cycle_time_extension = sched->cycle_time_extension;
+
+	list_for_each_entry(entry, &sched->entries, list) {
+		struct tc_taprio_sched_entry *e = &offload->entries[i];
+
+		e->command = entry->command;
+		e->interval = entry->interval;
+		e->gate_mask = entry->gate_mask;
+		i++;
+	}
+
+	offload->num_entries = i;
+}
+
+static int taprio_enable_offload(struct net_device *dev,
+				 struct tc_mqprio_qopt *mqprio,
+				 struct taprio_sched *q,
+				 struct sched_gate_list *sched,
+				 struct netlink_ext_ack *extack)
+{
+	const struct net_device_ops *ops = dev->netdev_ops;
+	struct tc_taprio_qopt_offload *offload;
+	int err = 0;
+
+	if (!ops->ndo_setup_tc) {
+		NL_SET_ERR_MSG(extack,
+			       "Device does not support taprio offload");
+		return -EOPNOTSUPP;
+	}
+
+	offload = taprio_offload_alloc(sched->num_entries);
+	if (!offload) {
+		NL_SET_ERR_MSG(extack,
+			       "Not enough memory for enabling offload mode");
+		return -ENOMEM;
+	}
+	offload->enable = 1;
+	taprio_sched_to_offload(q, sched, mqprio, offload);
+
+	err = ops->ndo_setup_tc(dev, TC_SETUP_QDISC_TAPRIO, offload);
+	if (err < 0) {
+		NL_SET_ERR_MSG(extack,
+			       "Device failed to setup taprio offload");
+		goto done;
+	}
+
+	taprio_offload_config_changed(q);
+
+done:
+	taprio_offload_free(offload);
+
+	return err;
+}
+
+static int taprio_disable_offload(struct net_device *dev,
+				  struct taprio_sched *q,
+				  struct netlink_ext_ack *extack)
+{
+	const struct net_device_ops *ops = dev->netdev_ops;
+	struct tc_taprio_qopt_offload *offload;
+	int err;
+
+	if (!FULL_OFFLOAD_IS_ENABLED(q->flags))
+		return 0;
+
+	if (!ops->ndo_setup_tc)
+		return -EOPNOTSUPP;
+
+	offload = taprio_offload_alloc(0);
+	if (!offload) {
+		NL_SET_ERR_MSG(extack,
+			       "Not enough memory to disable offload mode");
+		return -ENOMEM;
+	}
+	offload->enable = 0;
+
+	err = ops->ndo_setup_tc(dev, TC_SETUP_QDISC_TAPRIO, offload);
+	if (err < 0) {
+		NL_SET_ERR_MSG(extack,
+			       "Device failed to disable offload");
+		goto out;
+	}
+
+out:
+	taprio_offload_free(offload);
+
+	return err;
+}
+
+/* If full offload is enabled, the only possible clockid is the net device's
+ * PHC. For that reason, specifying a clockid through netlink is incorrect.
+ * For txtime-assist, it is implicitly assumed that the device's PHC is kept
+ * in sync with the specified clockid via a user space daemon such as phc2sys.
+ * For both software taprio and txtime-assist, the clockid is used for the
+ * hrtimer that advances the schedule and hence mandatory.
+ */
+static int taprio_parse_clockid(struct Qdisc *sch, struct nlattr **tb,
+				struct netlink_ext_ack *extack)
+{
+	struct taprio_sched *q = qdisc_priv(sch);
+	struct net_device *dev = qdisc_dev(sch);
+	int err = -EINVAL;
+
+	if (FULL_OFFLOAD_IS_ENABLED(q->flags)) {
+		const struct ethtool_ops *ops = dev->ethtool_ops;
+		struct ethtool_ts_info info = {
+			.cmd = ETHTOOL_GET_TS_INFO,
+			.phc_index = -1,
+		};
+
+		if (tb[TCA_TAPRIO_ATTR_SCHED_CLOCKID]) {
+			NL_SET_ERR_MSG(extack,
+				       "The 'clockid' cannot be specified for full offload");
+			goto out;
+		}
+
+		if (ops && ops->get_ts_info)
+			err = ops->get_ts_info(dev, &info);
+
+		if (err || info.phc_index < 0) {
+			NL_SET_ERR_MSG(extack,
+				       "Device does not have a PTP clock");
+			err = -ENOTSUPP;
+			goto out;
+		}
+	} else if (tb[TCA_TAPRIO_ATTR_SCHED_CLOCKID]) {
+		int clockid = nla_get_s32(tb[TCA_TAPRIO_ATTR_SCHED_CLOCKID]);
+
+		/* We only support static clockids and we don't allow
+		 * for it to be modified after the first init.
+		 */
+		if (clockid < 0 ||
+		    (q->clockid != -1 && q->clockid != clockid)) {
+			NL_SET_ERR_MSG(extack,
+				       "Changing the 'clockid' of a running schedule is not supported");
+			err = -ENOTSUPP;
+			goto out;
+		}
+
+		switch (clockid) {
+		case CLOCK_REALTIME:
+			q->tk_offset = TK_OFFS_REAL;
+			break;
+		case CLOCK_MONOTONIC:
+			q->tk_offset = TK_OFFS_MAX;
+			break;
+		case CLOCK_BOOTTIME:
+			q->tk_offset = TK_OFFS_BOOT;
+			break;
+		case CLOCK_TAI:
+			q->tk_offset = TK_OFFS_TAI;
+			break;
+		default:
+			NL_SET_ERR_MSG(extack, "Invalid 'clockid'");
+			err = -EINVAL;
+			goto out;
+		}
+
+		q->clockid = clockid;
+	} else {
+		NL_SET_ERR_MSG(extack, "Specifying a 'clockid' is mandatory");
+		goto out;
+	}
+out:
+	return err;
+}
+
 static int taprio_change(struct Qdisc *sch, struct nlattr *opt,
 			 struct netlink_ext_ack *extack)
 {
@@ -1020,9 +1355,9 @@ static int taprio_change(struct Qdisc *sch, struct nlattr *opt,
 	struct net_device *dev = qdisc_dev(sch);
 	struct tc_mqprio_qopt *mqprio = NULL;
 	u32 taprio_flags = 0;
-	int i, err, clockid;
 	unsigned long flags;
 	ktime_t start;
+	int i, err;
 
 	err = nla_parse_nested_deprecated(tb, TCA_TAPRIO_ATTR_MAX, opt,
 					  taprio_policy, extack);
@@ -1038,7 +1373,7 @@ static int taprio_change(struct Qdisc *sch, struct nlattr *opt,
 		if (q->flags != 0 && q->flags != taprio_flags) {
 			NL_SET_ERR_MSG_MOD(extack, "Changing 'flags' of a running schedule is not supported");
 			return -EOPNOTSUPP;
-		} else if (!FLAGS_VALID(taprio_flags)) {
+		} else if (!taprio_flags_valid(taprio_flags)) {
 			NL_SET_ERR_MSG_MOD(extack, "Specified 'flags' are not valid");
 			return -EINVAL;
 		}
@@ -1078,30 +1413,19 @@ static int taprio_change(struct Qdisc *sch, struct nlattr *opt,
 		goto free_sched;
 	}
 
-	if (tb[TCA_TAPRIO_ATTR_SCHED_CLOCKID]) {
-		clockid = nla_get_s32(tb[TCA_TAPRIO_ATTR_SCHED_CLOCKID]);
-
-		/* We only support static clockids and we don't allow
-		 * for it to be modified after the first init.
-		 */
-		if (clockid < 0 ||
-		    (q->clockid != -1 && q->clockid != clockid)) {
-			NL_SET_ERR_MSG(extack, "Changing the 'clockid' of a running schedule is not supported");
-			err = -ENOTSUPP;
-			goto free_sched;
-		}
-
-		q->clockid = clockid;
-	}
-
-	if (q->clockid == -1 && !tb[TCA_TAPRIO_ATTR_SCHED_CLOCKID]) {
-		NL_SET_ERR_MSG(extack, "Specifying a 'clockid' is mandatory");
-		err = -EINVAL;
+	err = taprio_parse_clockid(sch, tb, extack);
+	if (err < 0)
 		goto free_sched;
-	}
 
 	taprio_set_picos_per_byte(dev, q);
 
+	if (FULL_OFFLOAD_IS_ENABLED(taprio_flags))
+		err = taprio_enable_offload(dev, mqprio, q, new_admin, extack);
+	else
+		err = taprio_disable_offload(dev, q, extack);
+	if (err)
+		goto free_sched;
+
 	/* Protects against enqueue()/dequeue() */
 	spin_lock_bh(qdisc_lock(sch));
 
@@ -1116,6 +1440,7 @@ static int taprio_change(struct Qdisc *sch, struct nlattr *opt,
 	}
 
 	if (!TXTIME_ASSIST_IS_ENABLED(taprio_flags) &&
+	    !FULL_OFFLOAD_IS_ENABLED(taprio_flags) &&
 	    !hrtimer_active(&q->advance_timer)) {
 		hrtimer_init(&q->advance_timer, q->clockid, HRTIMER_MODE_ABS);
 		q->advance_timer.function = advance_sched;
@@ -1134,23 +1459,15 @@ static int taprio_change(struct Qdisc *sch, struct nlattr *opt,
 					       mqprio->prio_tc_map[i]);
 	}
 
-	switch (q->clockid) {
-	case CLOCK_REALTIME:
-		q->tk_offset = TK_OFFS_REAL;
-		break;
-	case CLOCK_MONOTONIC:
-		q->tk_offset = TK_OFFS_MAX;
-		break;
-	case CLOCK_BOOTTIME:
-		q->tk_offset = TK_OFFS_BOOT;
-		break;
-	case CLOCK_TAI:
-		q->tk_offset = TK_OFFS_TAI;
-		break;
-	default:
-		NL_SET_ERR_MSG(extack, "Invalid 'clockid'");
-		err = -EINVAL;
-		goto unlock;
+	if (FULL_OFFLOAD_IS_ENABLED(taprio_flags)) {
+		q->dequeue = taprio_dequeue_offload;
+		q->peek = taprio_peek_offload;
+	} else {
+		/* Be sure to always keep the function pointers
+		 * in a consistent state.
+		 */
+		q->dequeue = taprio_dequeue_soft;
+		q->peek = taprio_peek_soft;
 	}
 
 	err = taprio_get_start_time(sch, new_admin, &start);
@@ -1212,6 +1529,8 @@ static void taprio_destroy(struct Qdisc *sch)
 
 	hrtimer_cancel(&q->advance_timer);
 
+	taprio_disable_offload(dev, q, NULL);
+
 	if (q->qdiscs) {
 		for (i = 0; i < dev->num_tx_queues && q->qdiscs[i]; i++)
 			qdisc_put(q->qdiscs[i]);
@@ -1241,6 +1560,9 @@ static int taprio_init(struct Qdisc *sch, struct nlattr *opt,
 	hrtimer_init(&q->advance_timer, CLOCK_TAI, HRTIMER_MODE_ABS);
 	q->advance_timer.function = advance_sched;
 
+	q->dequeue = taprio_dequeue_soft;
+	q->peek = taprio_peek_soft;
+
 	q->root = sch;
 
 	/* We only support static clockids. Use an invalid value as default
@@ -1423,7 +1745,8 @@ static int taprio_dump(struct Qdisc *sch, struct sk_buff *skb)
 	if (nla_put(skb, TCA_TAPRIO_ATTR_PRIOMAP, sizeof(opt), &opt))
 		goto options_error;
 
-	if (nla_put_s32(skb, TCA_TAPRIO_ATTR_SCHED_CLOCKID, q->clockid))
+	if (!FULL_OFFLOAD_IS_ENABLED(q->flags) &&
+	    nla_put_s32(skb, TCA_TAPRIO_ATTR_SCHED_CLOCKID, q->clockid))
 		goto options_error;
 
 	if (q->flags && nla_put_u32(skb, TCA_TAPRIO_ATTR_FLAGS, q->flags))