Merge branch 'ENETC-mqprio-taprio-cleanup'

Vladimir Oltean says:

====================
net: ENETC mqprio/taprio cleanup

Please excuse the increased patch set size compared to v4's 15 patches,
but Claudiu stirred up the pot :) when he pointed out that the mqprio
TXQ validation procedure is still incorrect, so I had to fix that, and
then do some consolidation work so that taprio doesn't duplicate
mqprio's bugs. Compared to v4, 3 patches are new and 1 was dropped for now
("net/sched: taprio: mask off bits in gate mask that exceed number of TCs"),
since there's not really much to gain from it. Since the previous patch
set has largely been reviewed, I hope that a delta overview will help
and make up for the large size.

v4->v5:
- new patches:
  "[08/17] net/sched: mqprio: allow reverse TC:TXQ mappings"
  "[11/17] net/sched: taprio: centralize mqprio qopt validation"
  "[12/17] net/sched: refactor mqprio qopt reconstruction to a library function"
- changed patches worth revisiting:
  "[09/17] net/sched: mqprio: allow offloading drivers to request queue
  count validation"
v4 at:
https://patchwork.kernel.org/project/netdevbpf/cover/20230130173145.475943-1-vladimir.oltean@nxp.com/

v3->v4:
- adjusted patch 07/15 to not remove "#include <net/pkt_sched.h>" from
  ti cpsw
https://patchwork.kernel.org/project/netdevbpf/cover/20230127001516.592984-1-vladimir.oltean@nxp.com/

v2->v3:
- move min_num_stack_tx_queues definition so it doesn't conflict with
  the ethtool mm patches I haven't submitted yet for enetc (and also to
  make use of a 4 byte hole)
- warn and mask off excess TCs in gate mask instead of failing
- finally CC qdisc maintainers
v2 at:
https://patchwork.kernel.org/project/netdevbpf/patch/20230126125308.1199404-16-vladimir.oltean@nxp.com/

v1->v2:
- patches 1->4 are new
- update some header inclusions in drivers
- fix typo (said "taprio" instead of "mqprio")
- better enetc mqprio error handling
- dynamically reconstruct mqprio configuration in taprio offload
- also let stmmac and tsnep use per-TXQ gate_mask
v1 (RFC) at:
https://patchwork.kernel.org/project/netdevbpf/cover/20230120141537.1350744-1-vladimir.oltean@nxp.com/

The main goal of this patch set is to make taprio pass the mqprio queue
configuration structure down to ndo_setup_tc() - patch 13/17. But mqprio
itself is not in the best shape currently, so there are some
consolidation patches on that as well.

Next, there are some consolidation patches in the enetc driver's
handling of TX queues and their traffic class assignment. Then, there is
a consolidation between the TX queue configuration for mqprio and
taprio.

Finally, there is a change in the meaning of the gate_mask passed by
taprio through ndo_setup_tc(). We introduce a capability through which
drivers can request the gate mask to be per TXQ. The default is changed
so that it is per TC.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/ethernet/aquantia/atlantic/aq_main.c

@@ -21,6 +21,7 @@
 #include <linux/ip.h>
 #include <linux/udp.h>
 #include <net/pkt_cls.h>
+#include <net/pkt_sched.h>
 #include <linux/filter.h>
 
 MODULE_LICENSE("GPL v2");

drivers/net/ethernet/chelsio/cxgb4/cxgb4_tc_mqprio.h

@@ -4,7 +4,7 @@
 #ifndef __CXGB4_TC_MQPRIO_H__
 #define __CXGB4_TC_MQPRIO_H__
 
-#include <net/pkt_cls.h>
+#include <net/pkt_sched.h>
 
 #define CXGB4_EOSW_TXQ_DEFAULT_DESC_NUM 128
 

drivers/net/ethernet/engleder/tsnep_tc.c

@@ -403,12 +403,33 @@ static int tsnep_taprio(struct tsnep_adapter *adapter,
 	return 0;
 }
 
+static int tsnep_tc_query_caps(struct tsnep_adapter *adapter,
+			       struct tc_query_caps_base *base)
+{
+	switch (base->type) {
+	case TC_SETUP_QDISC_TAPRIO: {
+		struct tc_taprio_caps *caps = base->caps;
+
+		if (!adapter->gate_control)
+			return -EOPNOTSUPP;
+
+		caps->gate_mask_per_txq = true;
+
+		return 0;
+	}
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
 int tsnep_tc_setup(struct net_device *netdev, enum tc_setup_type type,
 		   void *type_data)
 {
 	struct tsnep_adapter *adapter = netdev_priv(netdev);
 
 	switch (type) {
+	case TC_QUERY_CAPS:
+		return tsnep_tc_query_caps(adapter, type_data);
 	case TC_SETUP_QDISC_TAPRIO:
 		return tsnep_taprio(adapter, type_data);
 	default:

drivers/net/ethernet/freescale/enetc/enetc.c

@@ -2609,63 +2609,95 @@ static int enetc_reconfigure(struct enetc_ndev_priv *priv, bool extended,
 	return err;
 }
 
-int enetc_setup_tc_mqprio(struct net_device *ndev, void *type_data)
+static void enetc_debug_tx_ring_prios(struct enetc_ndev_priv *priv)
+{
+	int i;
+
+	for (i = 0; i < priv->num_tx_rings; i++)
+		netdev_dbg(priv->ndev, "TX ring %d prio %d\n", i,
+			   priv->tx_ring[i]->prio);
+}
+
+static void enetc_reset_tc_mqprio(struct net_device *ndev)
 {
 	struct enetc_ndev_priv *priv = netdev_priv(ndev);
-	struct tc_mqprio_qopt *mqprio = type_data;
 	struct enetc_hw *hw = &priv->si->hw;
 	struct enetc_bdr *tx_ring;
 	int num_stack_tx_queues;
-	u8 num_tc;
 	int i;
 
 	num_stack_tx_queues = enetc_num_stack_tx_queues(priv);
-	mqprio->hw = TC_MQPRIO_HW_OFFLOAD_TCS;
-	num_tc = mqprio->num_tc;
 
-	if (!num_tc) {
-		netdev_reset_tc(ndev);
-		netif_set_real_num_tx_queues(ndev, num_stack_tx_queues);
-		priv->min_num_stack_tx_queues = num_possible_cpus();
-
-		/* Reset all ring priorities to 0 */
-		for (i = 0; i < priv->num_tx_rings; i++) {
-			tx_ring = priv->tx_ring[i];
-			tx_ring->prio = 0;
-			enetc_set_bdr_prio(hw, tx_ring->index, tx_ring->prio);
-		}
+	netdev_reset_tc(ndev);
+	netif_set_real_num_tx_queues(ndev, num_stack_tx_queues);
+	priv->min_num_stack_tx_queues = num_possible_cpus();
 
-		return 0;
+	/* Reset all ring priorities to 0 */
+	for (i = 0; i < priv->num_tx_rings; i++) {
+		tx_ring = priv->tx_ring[i];
+		tx_ring->prio = 0;
+		enetc_set_bdr_prio(hw, tx_ring->index, tx_ring->prio);
 	}
 
-	/* Check if we have enough BD rings available to accommodate all TCs */
-	if (num_tc > num_stack_tx_queues) {
-		netdev_err(ndev, "Max %d traffic classes supported\n",
-			   priv->num_tx_rings);
-		return -EINVAL;
+	enetc_debug_tx_ring_prios(priv);
+}
+
+int enetc_setup_tc_mqprio(struct net_device *ndev, void *type_data)
+{
+	struct enetc_ndev_priv *priv = netdev_priv(ndev);
+	struct tc_mqprio_qopt *mqprio = type_data;
+	struct enetc_hw *hw = &priv->si->hw;
+	int num_stack_tx_queues = 0;
+	u8 num_tc = mqprio->num_tc;
+	struct enetc_bdr *tx_ring;
+	int offset, count;
+	int err, tc, q;
+
+	if (!num_tc) {
+		enetc_reset_tc_mqprio(ndev);
+		return 0;
 	}
 
-	/* For the moment, we use only one BD ring per TC.
-	 *
-	 * Configure num_tc BD rings with increasing priorities.
-	 */
-	for (i = 0; i < num_tc; i++) {
-		tx_ring = priv->tx_ring[i];
-		tx_ring->prio = i;
-		enetc_set_bdr_prio(hw, tx_ring->index, tx_ring->prio);
+	err = netdev_set_num_tc(ndev, num_tc);
+	if (err)
+		return err;
+
+	for (tc = 0; tc < num_tc; tc++) {
+		offset = mqprio->offset[tc];
+		count = mqprio->count[tc];
+		num_stack_tx_queues += count;
+
+		err = netdev_set_tc_queue(ndev, tc, count, offset);
+		if (err)
+			goto err_reset_tc;
+
+		for (q = offset; q < offset + count; q++) {
+			tx_ring = priv->tx_ring[q];
+			/* The prio_tc_map is skb_tx_hash()'s way of selecting
+			 * between TX queues based on skb->priority. As such,
+			 * there's nothing to offload based on it.
+			 * Make the mqprio "traffic class" be the priority of
+			 * this ring group, and leave the Tx IPV to traffic
+			 * class mapping as its default mapping value of 1:1.
+			 */
+			tx_ring->prio = tc;
+			enetc_set_bdr_prio(hw, tx_ring->index, tx_ring->prio);
+		}
 	}
 
-	/* Reset the number of netdev queues based on the TC count */
-	netif_set_real_num_tx_queues(ndev, num_tc);
-	priv->min_num_stack_tx_queues = num_tc;
+	err = netif_set_real_num_tx_queues(ndev, num_stack_tx_queues);
+	if (err)
+		goto err_reset_tc;
 
-	netdev_set_num_tc(ndev, num_tc);
+	priv->min_num_stack_tx_queues = num_stack_tx_queues;
 
-	/* Each TC is associated with one netdev queue */
-	for (i = 0; i < num_tc; i++)
-		netdev_set_tc_queue(ndev, i, 1, i);
+	enetc_debug_tx_ring_prios(priv);
 
 	return 0;
+
+err_reset_tc:
+	enetc_reset_tc_mqprio(ndev);
+	return err;
 }
 EXPORT_SYMBOL_GPL(enetc_setup_tc_mqprio);
 

drivers/net/ethernet/freescale/enetc/enetc_qos.c

@@ -136,29 +136,21 @@ int enetc_setup_tc_taprio(struct net_device *ndev, void *type_data)
 {
 	struct tc_taprio_qopt_offload *taprio = type_data;
 	struct enetc_ndev_priv *priv = netdev_priv(ndev);
-	struct enetc_hw *hw = &priv->si->hw;
-	struct enetc_bdr *tx_ring;
-	int err;
-	int i;
+	int err, i;
 
 	/* TSD and Qbv are mutually exclusive in hardware */
 	for (i = 0; i < priv->num_tx_rings; i++)
 		if (priv->tx_ring[i]->tsd_enable)
 			return -EBUSY;
 
-	for (i = 0; i < priv->num_tx_rings; i++) {
-		tx_ring = priv->tx_ring[i];
-		tx_ring->prio = taprio->enable ? i : 0;
-		enetc_set_bdr_prio(hw, tx_ring->index, tx_ring->prio);
-	}
+	err = enetc_setup_tc_mqprio(ndev, &taprio->mqprio);
+	if (err)
+		return err;
 
 	err = enetc_setup_taprio(ndev, taprio);
 	if (err) {
-		for (i = 0; i < priv->num_tx_rings; i++) {
-			tx_ring = priv->tx_ring[i];
-			tx_ring->prio = taprio->enable ? 0 : i;
-			enetc_set_bdr_prio(hw, tx_ring->index, tx_ring->prio);
-		}
+		taprio->mqprio.qopt.num_tc = 0;
+		enetc_setup_tc_mqprio(ndev, &taprio->mqprio);
 	}
 
 	return err;
@@ -1611,6 +1603,13 @@ int enetc_qos_query_caps(struct net_device *ndev, void *type_data)
 	struct enetc_si *si = priv->si;
 
 	switch (base->type) {
+	case TC_SETUP_QDISC_MQPRIO: {
+		struct tc_mqprio_caps *caps = base->caps;
+
+		caps->validate_queue_counts = true;
+
+		return 0;
+	}
 	case TC_SETUP_QDISC_TAPRIO: {
 		struct tc_taprio_caps *caps = base->caps;
 

drivers/net/ethernet/hisilicon/hns3/hnae3.h

@@ -32,6 +32,7 @@
 #include <linux/pkt_sched.h>
 #include <linux/types.h>
 #include <net/pkt_cls.h>
+#include <net/pkt_sched.h>
 
 #define HNAE3_MOD_VERSION "1.0"
 

drivers/net/ethernet/hisilicon/hns3/hns3_enet.c

@@ -20,6 +20,7 @@
 #include <net/gro.h>
 #include <net/ip6_checksum.h>
 #include <net/pkt_cls.h>
+#include <net/pkt_sched.h>
 #include <net/tcp.h>
 #include <net/vxlan.h>
 #include <net/geneve.h>

drivers/net/ethernet/intel/i40e/i40e.h

@@ -33,6 +33,7 @@
 #include <linux/net_tstamp.h>
 #include <linux/ptp_clock_kernel.h>
 #include <net/pkt_cls.h>
+#include <net/pkt_sched.h>
 #include <net/tc_act/tc_gact.h>
 #include <net/tc_act/tc_mirred.h>
 #include <net/udp_tunnel.h>

drivers/net/ethernet/intel/iavf/iavf.h

@@ -30,6 +30,7 @@
 #include <linux/jiffies.h>
 #include <net/ip6_checksum.h>
 #include <net/pkt_cls.h>
+#include <net/pkt_sched.h>
 #include <net/udp.h>
 #include <net/tc_act/tc_gact.h>
 #include <net/tc_act/tc_mirred.h>

drivers/net/ethernet/intel/ice/ice.h

@@ -41,6 +41,7 @@
 #include <linux/dim.h>
 #include <linux/gnss.h>
 #include <net/pkt_cls.h>
+#include <net/pkt_sched.h>
 #include <net/tc_act/tc_mirred.h>
 #include <net/tc_act/tc_gact.h>
 #include <net/ip.h>

drivers/net/ethernet/intel/igc/igc_main.c

@@ -6205,12 +6205,35 @@ static int igc_tsn_enable_cbs(struct igc_adapter *adapter,
 	return igc_tsn_offload_apply(adapter);
 }
 
+static int igc_tc_query_caps(struct igc_adapter *adapter,
+			     struct tc_query_caps_base *base)
+{
+	struct igc_hw *hw = &adapter->hw;
+
+	switch (base->type) {
+	case TC_SETUP_QDISC_TAPRIO: {
+		struct tc_taprio_caps *caps = base->caps;
+
+		if (hw->mac.type != igc_i225)
+			return -EOPNOTSUPP;
+
+		caps->gate_mask_per_txq = true;
+
+		return 0;
+	}
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
 static int igc_setup_tc(struct net_device *dev, enum tc_setup_type type,
 			void *type_data)
 {
 	struct igc_adapter *adapter = netdev_priv(dev);
 
 	switch (type) {
+	case TC_QUERY_CAPS:
+		return igc_tc_query_caps(adapter, type_data);
 	case TC_SETUP_QDISC_TAPRIO:
 		return igc_tsn_enable_qbv_scheduling(adapter, type_data);
 

drivers/net/ethernet/marvell/mvneta.c

@@ -38,7 +38,7 @@
 #include <net/ipv6.h>
 #include <net/tso.h>
 #include <net/page_pool.h>
-#include <net/pkt_cls.h>
+#include <net/pkt_sched.h>
 #include <linux/bpf_trace.h>
 
 /* Registers */

drivers/net/ethernet/mellanox/mlx5/core/en_main.c

@@ -39,6 +39,7 @@
 #include <linux/if_bridge.h>
 #include <linux/filter.h>
 #include <net/page_pool.h>
+#include <net/pkt_sched.h>
 #include <net/xdp_sock_drv.h>
 #include "eswitch.h"
 #include "en.h"

drivers/net/ethernet/microchip/lan966x/lan966x_tc.c

 // SPDX-License-Identifier: GPL-2.0+
 
 #include <net/pkt_cls.h>
+#include <net/pkt_sched.h>
 
 #include "lan966x_main.h"
 

drivers/net/ethernet/microchip/sparx5/sparx5_tc.c

@@ -5,6 +5,7 @@
  */
 
 #include <net/pkt_cls.h>
+#include <net/pkt_sched.h>
 
 #include "sparx5_tc.h"
 #include "sparx5_main.h"

drivers/net/ethernet/stmicro/stmmac/hwif.h

@@ -567,6 +567,7 @@ struct tc_cbs_qopt_offload;
 struct flow_cls_offload;
 struct tc_taprio_qopt_offload;
 struct tc_etf_qopt_offload;
+struct tc_query_caps_base;
 
 struct stmmac_tc_ops {
 	int (*init)(struct stmmac_priv *priv);
@@ -580,6 +581,8 @@ struct stmmac_tc_ops {
 			    struct tc_taprio_qopt_offload *qopt);
 	int (*setup_etf)(struct stmmac_priv *priv,
 			 struct tc_etf_qopt_offload *qopt);
+	int (*query_caps)(struct stmmac_priv *priv,
+			  struct tc_query_caps_base *base);
 };
 
 #define stmmac_tc_init(__priv, __args...) \
@@ -594,6 +597,8 @@ struct stmmac_tc_ops {
 	stmmac_do_callback(__priv, tc, setup_taprio, __args)
 #define stmmac_tc_setup_etf(__priv, __args...) \
 	stmmac_do_callback(__priv, tc, setup_etf, __args)
+#define stmmac_tc_query_caps(__priv, __args...) \
+	stmmac_do_callback(__priv, tc, query_caps, __args)
 
 struct stmmac_counters;
 

drivers/net/ethernet/stmicro/stmmac/stmmac_main.c

@@ -5992,6 +5992,8 @@ static int stmmac_setup_tc(struct net_device *ndev, enum tc_setup_type type,
 	struct stmmac_priv *priv = netdev_priv(ndev);
 
 	switch (type) {
+	case TC_QUERY_CAPS:
+		return stmmac_tc_query_caps(priv, priv, type_data);
 	case TC_SETUP_BLOCK:
 		return flow_block_cb_setup_simple(type_data,
 						  &stmmac_block_cb_list,

drivers/net/ethernet/stmicro/stmmac/stmmac_tc.c

@@ -1107,6 +1107,25 @@ static int tc_setup_etf(struct stmmac_priv *priv,
 	return 0;
 }
 
+static int tc_query_caps(struct stmmac_priv *priv,
+			 struct tc_query_caps_base *base)
+{
+	switch (base->type) {
+	case TC_SETUP_QDISC_TAPRIO: {
+		struct tc_taprio_caps *caps = base->caps;
+
+		if (!priv->dma_cap.estsel)
+			return -EOPNOTSUPP;
+
+		caps->gate_mask_per_txq = true;
+
+		return 0;
+	}
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
 const struct stmmac_tc_ops dwmac510_tc_ops = {
 	.init = tc_init,
 	.setup_cls_u32 = tc_setup_cls_u32,
@@ -1114,4 +1133,5 @@ const struct stmmac_tc_ops dwmac510_tc_ops = {
 	.setup_cls = tc_setup_cls,
 	.setup_taprio = tc_setup_taprio,
 	.setup_etf = tc_setup_etf,
+	.query_caps = tc_query_caps,
 };

drivers/net/ethernet/ti/am65-cpsw-qos.c

@@ -585,6 +585,26 @@ static int am65_cpsw_setup_taprio(struct net_device *ndev, void *type_data)
 	return am65_cpsw_set_taprio(ndev, type_data);
 }
 
+static int am65_cpsw_tc_query_caps(struct net_device *ndev, void *type_data)
+{
+	struct tc_query_caps_base *base = type_data;
+
+	switch (base->type) {
+	case TC_SETUP_QDISC_TAPRIO: {
+		struct tc_taprio_caps *caps = base->caps;
+
+		if (!IS_ENABLED(CONFIG_TI_AM65_CPSW_TAS))
+			return -EOPNOTSUPP;
+
+		caps->gate_mask_per_txq = true;
+
+		return 0;
+	}
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
 static int am65_cpsw_qos_clsflower_add_policer(struct am65_cpsw_port *port,
 					       struct netlink_ext_ack *extack,
 					       struct flow_cls_offload *cls,
@@ -765,6 +785,8 @@ int am65_cpsw_qos_ndo_setup_tc(struct net_device *ndev, enum tc_setup_type type,
 			       void *type_data)
 {
 	switch (type) {
+	case TC_QUERY_CAPS:
+		return am65_cpsw_tc_query_caps(ndev, type_data);
 	case TC_SETUP_QDISC_TAPRIO:
 		return am65_cpsw_setup_taprio(ndev, type_data);
 	case TC_SETUP_BLOCK:

drivers/net/ethernet/ti/cpsw_priv.c

@@ -20,6 +20,7 @@
 #include <linux/skbuff.h>
 #include <net/page_pool.h>
 #include <net/pkt_cls.h>
+#include <net/pkt_sched.h>
 
 #include "cpsw.h"
 #include "cpts.h"

include/net/pkt_cls.h

@@ -788,16 +788,6 @@ struct tc_cls_bpf_offload {
 	bool exts_integrated;
 };
 
-struct tc_mqprio_qopt_offload {
-	/* struct tc_mqprio_qopt must always be the first element */
-	struct tc_mqprio_qopt qopt;
-	u16 mode;
-	u16 shaper;
-	u32 flags;
-	u64 min_rate[TC_QOPT_MAX_QUEUE];
-	u64 max_rate[TC_QOPT_MAX_QUEUE];
-};
-
 /* This structure holds cookie structure that is passed from user
  * to the kernel for actions and classifiers
  */

include/net/pkt_sched.h

@@ -160,8 +160,23 @@ struct tc_etf_qopt_offload {
 	s32 queue;
 };
 
+struct tc_mqprio_caps {
+	bool validate_queue_counts:1;
+};
+
+struct tc_mqprio_qopt_offload {
+	/* struct tc_mqprio_qopt must always be the first element */
+	struct tc_mqprio_qopt qopt;
+	u16 mode;
+	u16 shaper;
+	u32 flags;
+	u64 min_rate[TC_QOPT_MAX_QUEUE];
+	u64 max_rate[TC_QOPT_MAX_QUEUE];
+};
+
 struct tc_taprio_caps {
 	bool supports_queue_max_sdu:1;
+	bool gate_mask_per_txq:1;
 };
 
 struct tc_taprio_sched_entry {
@@ -173,6 +188,7 @@ struct tc_taprio_sched_entry {
 };
 
 struct tc_taprio_qopt_offload {
+	struct tc_mqprio_qopt_offload mqprio;
 	u8 enable;
 	ktime_t base_time;
 	u64 cycle_time;

net/sched/Kconfig

@@ -195,8 +195,14 @@ config NET_SCH_ETF
 	  To compile this code as a module, choose M here: the
 	  module will be called sch_etf.
 
+config NET_SCH_MQPRIO_LIB
+	tristate
+	help
+	  Common library for manipulating mqprio queue configurations.
+
 config NET_SCH_TAPRIO
 	tristate "Time Aware Priority (taprio) Scheduler"
+	select NET_SCH_MQPRIO_LIB
 	help
 	  Say Y here if you want to use the Time Aware Priority (taprio) packet
 	  scheduling algorithm.
@@ -253,6 +259,7 @@ config NET_SCH_DRR
 
 config NET_SCH_MQPRIO
 	tristate "Multi-queue priority scheduler (MQPRIO)"
+	select NET_SCH_MQPRIO_LIB
 	help
 	  Say Y here if you want to use the Multi-queue Priority scheduler.
 	  This scheduler allows QOS to be offloaded on NICs that have support

net/sched/Makefile

@@ -52,6 +52,7 @@ obj-$(CONFIG_NET_SCH_DRR)	+= sch_drr.o
 obj-$(CONFIG_NET_SCH_PLUG)	+= sch_plug.o
 obj-$(CONFIG_NET_SCH_ETS)	+= sch_ets.o
 obj-$(CONFIG_NET_SCH_MQPRIO)	+= sch_mqprio.o
+obj-$(CONFIG_NET_SCH_MQPRIO_LIB) += sch_mqprio_lib.o
 obj-$(CONFIG_NET_SCH_SKBPRIO)	+= sch_skbprio.o
 obj-$(CONFIG_NET_SCH_CHOKE)	+= sch_choke.o
 obj-$(CONFIG_NET_SCH_QFQ)	+= sch_qfq.o

net/sched/sch_mqprio_lib.c

+// SPDX-License-Identifier: GPL-2.0-only
+
+#include <linux/net.h>
+#include <linux/netdevice.h>
+#include <linux/netlink.h>
+#include <linux/types.h>
+#include <net/pkt_sched.h>
+
+#include "sch_mqprio_lib.h"
+
+/* Returns true if the intervals [a, b) and [c, d) overlap. */
+static bool intervals_overlap(int a, int b, int c, int d)
+{
+	int left = max(a, c), right = min(b, d);
+
+	return left < right;
+}
+
+static int mqprio_validate_queue_counts(struct net_device *dev,
+					const struct tc_mqprio_qopt *qopt,
+					bool allow_overlapping_txqs,
+					struct netlink_ext_ack *extack)
+{
+	int i, j;
+
+	for (i = 0; i < qopt->num_tc; i++) {
+		unsigned int last = qopt->offset[i] + qopt->count[i];
+
+		if (!qopt->count[i]) {
+			NL_SET_ERR_MSG_FMT_MOD(extack, "No queues for TC %d",
+					       i);
+			return -EINVAL;
+		}
+
+		/* Verify the queue count is in tx range being equal to the
+		 * real_num_tx_queues indicates the last queue is in use.
+		 */
+		if (qopt->offset[i] >= dev->real_num_tx_queues ||
+		    last > dev->real_num_tx_queues) {
+			NL_SET_ERR_MSG_FMT_MOD(extack,
+					       "Queues %d:%d for TC %d exceed the %d TX queues available",
+					       qopt->count[i], qopt->offset[i],
+					       i, dev->real_num_tx_queues);
+			return -EINVAL;
+		}
+
+		if (allow_overlapping_txqs)
+			continue;
+
+		/* Verify that the offset and counts do not overlap */
+		for (j = i + 1; j < qopt->num_tc; j++) {
+			if (intervals_overlap(qopt->offset[i], last,
+					      qopt->offset[j],
+					      qopt->offset[j] +
+					      qopt->count[j])) {
+				NL_SET_ERR_MSG_FMT_MOD(extack,
+						       "TC %d queues %d@%d overlap with TC %d queues %d@%d",
+						       i, qopt->count[i], qopt->offset[i],
+						       j, qopt->count[j], qopt->offset[j]);
+				return -EINVAL;
+			}
+		}
+	}
+
+	return 0;
+}
+
+int mqprio_validate_qopt(struct net_device *dev, struct tc_mqprio_qopt *qopt,
+			 bool validate_queue_counts,
+			 bool allow_overlapping_txqs,
+			 struct netlink_ext_ack *extack)
+{
+	int i, err;
+
+	/* Verify num_tc is not out of max range */
+	if (qopt->num_tc > TC_MAX_QUEUE) {
+		NL_SET_ERR_MSG(extack,
+			       "Number of traffic classes is outside valid range");
+		return -EINVAL;
+	}
+
+	/* Verify priority mapping uses valid tcs */
+	for (i = 0; i <= TC_BITMASK; i++) {
+		if (qopt->prio_tc_map[i] >= qopt->num_tc) {
+			NL_SET_ERR_MSG(extack,
+				       "Invalid traffic class in priority to traffic class mapping");
+			return -EINVAL;
+		}
+	}
+
+	if (validate_queue_counts) {
+		err = mqprio_validate_queue_counts(dev, qopt,
+						   allow_overlapping_txqs,
+						   extack);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mqprio_validate_qopt);
+
+void mqprio_qopt_reconstruct(struct net_device *dev, struct tc_mqprio_qopt *qopt)
+{
+	int tc, num_tc = netdev_get_num_tc(dev);
+
+	qopt->num_tc = num_tc;
+	memcpy(qopt->prio_tc_map, dev->prio_tc_map, sizeof(qopt->prio_tc_map));
+
+	for (tc = 0; tc < num_tc; tc++) {
+		qopt->count[tc] = dev->tc_to_txq[tc].count;
+		qopt->offset[tc] = dev->tc_to_txq[tc].offset;
+	}
+}
+EXPORT_SYMBOL_GPL(mqprio_qopt_reconstruct);
+
+MODULE_LICENSE("GPL");

net/sched/sch_mqprio_lib.h

+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __SCH_MQPRIO_LIB_H
+#define __SCH_MQPRIO_LIB_H
+
+#include <linux/types.h>
+
+struct net_device;
+struct netlink_ext_ack;
+struct tc_mqprio_qopt;
+
+int mqprio_validate_qopt(struct net_device *dev, struct tc_mqprio_qopt *qopt,
+			 bool validate_queue_counts,
+			 bool allow_overlapping_txqs,
+			 struct netlink_ext_ack *extack);
+void mqprio_qopt_reconstruct(struct net_device *dev,
+			     struct tc_mqprio_qopt *qopt);
+
+#endif

net/sched/sch_taprio.c

@@ -26,6 +26,8 @@
 #include <net/sock.h>
 #include <net/tcp.h>
 
+#include "sch_mqprio_lib.h"
+
 static LIST_HEAD(taprio_list);
 
 #define TAPRIO_ALL_GATES_OPEN -1
@@ -924,7 +926,7 @@ static int taprio_parse_mqprio_opt(struct net_device *dev,
 				   struct netlink_ext_ack *extack,
 				   u32 taprio_flags)
 {
-	int i, j;
+	bool allow_overlapping_txqs = TXTIME_ASSIST_IS_ENABLED(taprio_flags);
 
 	if (!qopt && !dev->num_tc) {
 		NL_SET_ERR_MSG(extack, "'mqprio' configuration is necessary");
@@ -937,52 +939,17 @@ static int taprio_parse_mqprio_opt(struct net_device *dev,
 	if (dev->num_tc)
 		return 0;
 
-	/* Verify num_tc is not out of max range */
-	if (qopt->num_tc > TC_MAX_QUEUE) {
-		NL_SET_ERR_MSG(extack, "Number of traffic classes is outside valid range");
-		return -EINVAL;
-	}
-
 	/* taprio imposes that traffic classes map 1:n to tx queues */
 	if (qopt->num_tc > dev->num_tx_queues) {
 		NL_SET_ERR_MSG(extack, "Number of traffic classes is greater than number of HW queues");
 		return -EINVAL;
 	}
 
-	/* Verify priority mapping uses valid tcs */
-	for (i = 0; i <= TC_BITMASK; i++) {
-		if (qopt->prio_tc_map[i] >= qopt->num_tc) {
-			NL_SET_ERR_MSG(extack, "Invalid traffic class in priority to traffic class mapping");
-			return -EINVAL;
-		}
-	}
-
-	for (i = 0; i < qopt->num_tc; i++) {
-		unsigned int last = qopt->offset[i] + qopt->count[i];
-
-		/* Verify the queue count is in tx range being equal to the
-		 * real_num_tx_queues indicates the last queue is in use.
-		 */
-		if (qopt->offset[i] >= dev->num_tx_queues ||
-		    !qopt->count[i] ||
-		    last > dev->real_num_tx_queues) {
-			NL_SET_ERR_MSG(extack, "Invalid queue in traffic class to queue mapping");
-			return -EINVAL;
-		}
-
-		if (TXTIME_ASSIST_IS_ENABLED(taprio_flags))
-			continue;
-
-		/* Verify that the offset and counts do not overlap */
-		for (j = i + 1; j < qopt->num_tc; j++) {
-			if (last > qopt->offset[j]) {
-				NL_SET_ERR_MSG(extack, "Detected overlap in the traffic class to queue mapping");
-				return -EINVAL;
-			}
-		}
-	}
-
-	return 0;
+	/* For some reason, in txtime-assist mode, we allow TXQ ranges for
+	 * different TCs to overlap, and just validate the TXQ ranges.
+	 */
+	return mqprio_validate_qopt(dev, qopt, true, allow_overlapping_txqs,
+				    extack);
 }
 
 static int taprio_get_start_time(struct Qdisc *sch,
@@ -1203,7 +1170,8 @@ static u32 tc_map_to_queue_mask(struct net_device *dev, u32 tc_mask)
 
 static void taprio_sched_to_offload(struct net_device *dev,
 				    struct sched_gate_list *sched,
-				    struct tc_taprio_qopt_offload *offload)
+				    struct tc_taprio_qopt_offload *offload,
+				    const struct tc_taprio_caps *caps)
 {
 	struct sched_entry *entry;
 	int i = 0;
@@ -1217,7 +1185,11 @@ static void taprio_sched_to_offload(struct net_device *dev,
 
 		e->command = entry->command;
 		e->interval = entry->interval;
-		e->gate_mask = tc_map_to_queue_mask(dev, entry->gate_mask);
+		if (caps->gate_mask_per_txq)
+			e->gate_mask = tc_map_to_queue_mask(dev,
+							    entry->gate_mask);
+		else
+			e->gate_mask = entry->gate_mask;
 
 		i++;
 	}
@@ -1261,7 +1233,8 @@ static int taprio_enable_offload(struct net_device *dev,
 		return -ENOMEM;
 	}
 	offload->enable = 1;
-	taprio_sched_to_offload(dev, sched, offload);
+	mqprio_qopt_reconstruct(dev, &offload->mqprio.qopt);
+	taprio_sched_to_offload(dev, sched, offload, &caps);
 
 	for (tc = 0; tc < TC_MAX_QUEUE; tc++)
 		offload->max_sdu[tc] = q->max_sdu[tc];
@@ -1981,18 +1954,11 @@ static int taprio_dump(struct Qdisc *sch, struct sk_buff *skb)
 	struct sched_gate_list *oper, *admin;
 	struct tc_mqprio_qopt opt = { 0 };
 	struct nlattr *nest, *sched_nest;
-	unsigned int i;
 
 	oper = rtnl_dereference(q->oper_sched);
 	admin = rtnl_dereference(q->admin_sched);
 
-	opt.num_tc = netdev_get_num_tc(dev);
-	memcpy(opt.prio_tc_map, dev->prio_tc_map, sizeof(opt.prio_tc_map));
-
-	for (i = 0; i < netdev_get_num_tc(dev); i++) {
-		opt.count[i] = dev->tc_to_txq[i].count;
-		opt.offset[i] = dev->tc_to_txq[i].offset;
-	}
+	mqprio_qopt_reconstruct(dev, &opt);
 
 	nest = nla_nest_start_noflag(skb, TCA_OPTIONS);
 	if (!nest)