Merge branch 'Generic-adjustment-for-flow-dissector-in-DSA'

Vladimir Oltean says:

====================
Generic adjustment for flow dissector in DSA

This is the v2 of a series initially submitted in May:
https://www.spinics.net/lists/netdev/msg651866.html

The end goal is to get rid of the unintuitive code for the flow
dissector that currently exists in the taggers. It can all be replaced
by a single, common function.

Some background work needs to be done for that. Especially the ocelot
driver poses some problems, since it has a different tag length between
RX and TX, and I didn't want to make DSA aware of that, since I could
instead make the tag lengths equal.

Changes in v3:
- Added an optimization (08/15) that makes the generic case not need to
  call the .flow_dissect function pointer. Basically .flow_dissect now
  currently only exists for sja1105.
- Moved the .promisc_on_master property to the tagger structure.
- Added the .tail_tag property to the tagger structure.
- Disabled "suppresscc = all" from my .gitconfig.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/dsa/ocelot/felix.c

@@ -439,6 +439,8 @@ static int felix_init_structs(struct felix *felix, int num_phys_ports)
 	ocelot->vcap_is2_actions= felix->info->vcap_is2_actions;
 	ocelot->vcap		= felix->info->vcap;
 	ocelot->ops		= felix->info->ops;
+	ocelot->inj_prefix	= OCELOT_TAG_PREFIX_SHORT;
+	ocelot->xtr_prefix	= OCELOT_TAG_PREFIX_SHORT;
 
 	port_phy_modes = kcalloc(num_phys_ports, sizeof(phy_interface_t),
 				 GFP_KERNEL);
@@ -509,7 +511,7 @@ static int felix_init_structs(struct felix *felix, int num_phys_ports)
 			return PTR_ERR(target);
 		}
 
-		template = devm_kzalloc(ocelot->dev, OCELOT_TAG_LEN,
+		template = devm_kzalloc(ocelot->dev, OCELOT_TOTAL_TAG_LEN,
 					GFP_KERNEL);
 		if (!template) {
 			dev_err(ocelot->dev,
@@ -538,6 +540,28 @@ static int felix_init_structs(struct felix *felix, int num_phys_ports)
 	return 0;
 }
 
+/* The CPU port module is connected to the Node Processor Interface (NPI). This
+ * is the mode through which frames can be injected from and extracted to an
+ * external CPU, over Ethernet.
+ */
+static void felix_npi_port_init(struct ocelot *ocelot, int port)
+{
+	ocelot->npi = port;
+
+	ocelot_write(ocelot, QSYS_EXT_CPU_CFG_EXT_CPUQ_MSK_M |
+		     QSYS_EXT_CPU_CFG_EXT_CPU_PORT(port),
+		     QSYS_EXT_CPU_CFG);
+
+	/* NPI port Injection/Extraction configuration */
+	ocelot_fields_write(ocelot, port, SYS_PORT_MODE_INCL_XTR_HDR,
+			    ocelot->xtr_prefix);
+	ocelot_fields_write(ocelot, port, SYS_PORT_MODE_INCL_INJ_HDR,
+			    ocelot->inj_prefix);
+
+	/* Disable transmission of pause frames */
+	ocelot_fields_write(ocelot, port, SYS_PAUSE_CFG_PAUSE_ENA, 0);
+}
+
 /* Hardware initialization done here so that we can allocate structures with
  * devm without fear of dsa_register_switch returning -EPROBE_DEFER and causing
  * us to allocate structures twice (leak memory) and map PCI memory twice
@@ -570,11 +594,8 @@ static int felix_setup(struct dsa_switch *ds)
 	for (port = 0; port < ds->num_ports; port++) {
 		ocelot_init_port(ocelot, port);
 
-		/* Bring up the CPU port module and configure the NPI port */
 		if (dsa_is_cpu_port(ds, port))
-			ocelot_configure_cpu(ocelot, port,
-					     OCELOT_TAG_PREFIX_NONE,
-					     OCELOT_TAG_PREFIX_LONG);
+			felix_npi_port_init(ocelot, port);
 
 		/* Set the default QoS Classification based on PCP and DEI
 		 * bits of vlan tag.

drivers/net/dsa/ocelot/felix_vsc9959.c

@@ -1155,6 +1155,8 @@ static void vsc9959_xmit_template_populate(struct ocelot *ocelot, int port)
 	struct ocelot_port *ocelot_port = ocelot->ports[port];
 	u8 *template = ocelot_port->xmit_template;
 	u64 bypass, dest, src;
+	__be32 *prefix;
+	u8 *injection;
 
 	/* Set the source port as the CPU port module and not the
 	 * NPI port
@@ -1163,9 +1165,14 @@ static void vsc9959_xmit_template_populate(struct ocelot *ocelot, int port)
 	dest = BIT(port);
 	bypass = true;
 
-	packing(template, &bypass, 127, 127, OCELOT_TAG_LEN, PACK, 0);
-	packing(template, &dest,    68,  56, OCELOT_TAG_LEN, PACK, 0);
-	packing(template, &src,     46,  43, OCELOT_TAG_LEN, PACK, 0);
+	injection = template + OCELOT_SHORT_PREFIX_LEN;
+	prefix = (__be32 *)template;
+
+	packing(injection, &bypass, 127, 127, OCELOT_TAG_LEN, PACK, 0);
+	packing(injection, &dest,    68,  56, OCELOT_TAG_LEN, PACK, 0);
+	packing(injection, &src,     46,  43, OCELOT_TAG_LEN, PACK, 0);
+
+	*prefix = cpu_to_be32(0x8880000a);
 }
 
 static const struct felix_info felix_info_vsc9959 = {

drivers/net/dsa/ocelot/seville_vsc9953.c

@@ -1003,6 +1003,8 @@ static void vsc9953_xmit_template_populate(struct ocelot *ocelot, int port)
 	struct ocelot_port *ocelot_port = ocelot->ports[port];
 	u8 *template = ocelot_port->xmit_template;
 	u64 bypass, dest, src;
+	__be32 *prefix;
+	u8 *injection;
 
 	/* Set the source port as the CPU port module and not the
 	 * NPI port
@@ -1011,9 +1013,14 @@ static void vsc9953_xmit_template_populate(struct ocelot *ocelot, int port)
 	dest = BIT(port);
 	bypass = true;
 
-	packing(template, &bypass, 127, 127, OCELOT_TAG_LEN, PACK, 0);
-	packing(template, &dest,    67,  57, OCELOT_TAG_LEN, PACK, 0);
-	packing(template, &src,     46,  43, OCELOT_TAG_LEN, PACK, 0);
+	injection = template + OCELOT_SHORT_PREFIX_LEN;
+	prefix = (__be32 *)template;
+
+	packing(injection, &bypass, 127, 127, OCELOT_TAG_LEN, PACK, 0);
+	packing(injection, &dest,    67,  57, OCELOT_TAG_LEN, PACK, 0);
+	packing(injection, &src,     46,  43, OCELOT_TAG_LEN, PACK, 0);
+
+	*prefix = cpu_to_be32(0x88800005);
 }
 
 static const struct felix_info seville_info_vsc9953 = {

drivers/net/ethernet/mscc/ocelot.c

@@ -1346,22 +1346,14 @@ void ocelot_init_port(struct ocelot *ocelot, int port)
 }
 EXPORT_SYMBOL(ocelot_init_port);
 
-/* Configure and enable the CPU port module, which is a set of queues.
- * If @npi contains a valid port index, the CPU port module is connected
- * to the Node Processor Interface (NPI). This is the mode through which
- * frames can be injected from and extracted to an external CPU,
- * over Ethernet.
+/* Configure and enable the CPU port module, which is a set of queues
+ * accessible through register MMIO, frame DMA or Ethernet (in case
+ * NPI mode is used).
  */
-void ocelot_configure_cpu(struct ocelot *ocelot, int npi,
-			  enum ocelot_tag_prefix injection,
-			  enum ocelot_tag_prefix extraction)
+static void ocelot_cpu_port_init(struct ocelot *ocelot)
 {
 	int cpu = ocelot->num_phys_ports;
 
-	ocelot->npi = npi;
-	ocelot->inj_prefix = injection;
-	ocelot->xtr_prefix = extraction;
-
 	/* The unicast destination PGID for the CPU port module is unused */
 	ocelot_write_rix(ocelot, 0, ANA_PGID_PGID, cpu);
 	/* Instead set up a multicast destination PGID for traffic copied to
@@ -1373,39 +1365,20 @@ void ocelot_configure_cpu(struct ocelot *ocelot, int npi,
 			 ANA_PORT_PORT_CFG_PORTID_VAL(cpu),
 			 ANA_PORT_PORT_CFG, cpu);
 
-	if (npi >= 0 && npi < ocelot->num_phys_ports) {
-		ocelot_write(ocelot, QSYS_EXT_CPU_CFG_EXT_CPUQ_MSK_M |
-			     QSYS_EXT_CPU_CFG_EXT_CPU_PORT(npi),
-			     QSYS_EXT_CPU_CFG);
-
-		/* Enable NPI port */
-		ocelot_fields_write(ocelot, npi,
-				    QSYS_SWITCH_PORT_MODE_PORT_ENA, 1);
-		/* NPI port Injection/Extraction configuration */
-		ocelot_fields_write(ocelot, npi, SYS_PORT_MODE_INCL_XTR_HDR,
-				    extraction);
-		ocelot_fields_write(ocelot, npi, SYS_PORT_MODE_INCL_INJ_HDR,
-				    injection);
-
-		/* Disable transmission of pause frames */
-		ocelot_fields_write(ocelot, npi, SYS_PAUSE_CFG_PAUSE_ENA, 0);
-	}
-
 	/* Enable CPU port module */
 	ocelot_fields_write(ocelot, cpu, QSYS_SWITCH_PORT_MODE_PORT_ENA, 1);
 	/* CPU port Injection/Extraction configuration */
 	ocelot_fields_write(ocelot, cpu, SYS_PORT_MODE_INCL_XTR_HDR,
-			    extraction);
+			    ocelot->xtr_prefix);
 	ocelot_fields_write(ocelot, cpu, SYS_PORT_MODE_INCL_INJ_HDR,
-			    injection);
+			    ocelot->inj_prefix);
 
 	/* Configure the CPU port to be VLAN aware */
 	ocelot_write_gix(ocelot, ANA_PORT_VLAN_CFG_VLAN_VID(0) |
 				 ANA_PORT_VLAN_CFG_VLAN_AWARE_ENA |
 				 ANA_PORT_VLAN_CFG_VLAN_POP_CNT(1),
 			 ANA_PORT_VLAN_CFG, cpu);
 }
-EXPORT_SYMBOL(ocelot_configure_cpu);
 
 int ocelot_init(struct ocelot *ocelot)
 {
@@ -1445,6 +1418,7 @@ int ocelot_init(struct ocelot *ocelot)
 	ocelot_mact_init(ocelot);
 	ocelot_vlan_init(ocelot);
 	ocelot_vcap_init(ocelot);
+	ocelot_cpu_port_init(ocelot);
 
 	for (port = 0; port < ocelot->num_phys_ports; port++) {
 		/* Clear all counters (5 groups) */

drivers/net/ethernet/mscc/ocelot_vsc7514.c

@@ -930,10 +930,6 @@ static int mscc_ocelot_init_ports(struct platform_device *pdev,
 	if (!ocelot->ports)
 		return -ENOMEM;
 
-	/* No NPI port */
-	ocelot_configure_cpu(ocelot, -1, OCELOT_TAG_PREFIX_NONE,
-			     OCELOT_TAG_PREFIX_NONE);
-
 	for_each_available_child_of_node(ports, portnp) {
 		struct ocelot_port_private *priv;
 		struct ocelot_port *ocelot_port;
@@ -1120,6 +1116,9 @@ static int mscc_ocelot_probe(struct platform_device *pdev)
 	ocelot->vcap_is2_keys = vsc7514_vcap_is2_keys;
 	ocelot->vcap_is2_actions = vsc7514_vcap_is2_actions;
 	ocelot->vcap = vsc7514_vcap_props;
+	ocelot->inj_prefix = OCELOT_TAG_PREFIX_NONE;
+	ocelot->xtr_prefix = OCELOT_TAG_PREFIX_NONE;
+	ocelot->npi = -1;
 
 	err = ocelot_init(ocelot);
 	if (err)

include/net/dsa.h

@@ -74,8 +74,8 @@ struct dsa_device_ops {
 	struct sk_buff *(*xmit)(struct sk_buff *skb, struct net_device *dev);
 	struct sk_buff *(*rcv)(struct sk_buff *skb, struct net_device *dev,
 			       struct packet_type *pt);
-	int (*flow_dissect)(const struct sk_buff *skb, __be16 *proto,
-			    int *offset);
+	void (*flow_dissect)(const struct sk_buff *skb, __be16 *proto,
+			     int *offset);
 	/* Used to determine which traffic should match the DSA filter in
 	 * eth_type_trans, and which, if any, should bypass it and be processed
 	 * as regular on the master net device.
@@ -84,6 +84,13 @@ struct dsa_device_ops {
 	unsigned int overhead;
 	const char *name;
 	enum dsa_tag_protocol proto;
+	/* Some tagging protocols either mangle or shift the destination MAC
+	 * address, in which case the DSA master would drop packets on ingress
+	 * if what it understands out of the destination MAC address is not in
+	 * its RX filter.
+	 */
+	bool promisc_on_master;
+	bool tail_tag;
 };
 
 /* This structure defines the control interfaces that are overlayed by the
@@ -705,6 +712,32 @@ static inline bool dsa_can_decode(const struct sk_buff *skb,
 	return false;
 }
 
+/* All DSA tags that push the EtherType to the right (basically all except tail
+ * tags, which don't break dissection) can be treated the same from the
+ * perspective of the flow dissector.
+ *
+ * We need to return:
+ *  - offset: the (B - A) difference between:
+ *    A. the position of the real EtherType and
+ *    B. the current skb->data (aka ETH_HLEN bytes into the frame, aka 2 bytes
+ *       after the normal EtherType was supposed to be)
+ *    The offset in bytes is exactly equal to the tagger overhead (and half of
+ *    that, in __be16 shorts).
+ *
+ *  - proto: the value of the real EtherType.
+ */
+static inline void dsa_tag_generic_flow_dissect(const struct sk_buff *skb,
+						__be16 *proto, int *offset)
+{
+#if IS_ENABLED(CONFIG_NET_DSA)
+	const struct dsa_device_ops *ops = skb->dev->dsa_ptr->tag_ops;
+	int tag_len = ops->overhead;
+
+	*offset = tag_len;
+	*proto = ((__be16 *)skb->data)[(tag_len / 2) - 1];
+#endif
+}
+
 #if IS_ENABLED(CONFIG_NET_DSA)
 static inline int __dsa_netdevice_ops_check(struct net_device *dev)
 {

include/soc/mscc/ocelot.h

@@ -101,6 +101,7 @@
 #define OCELOT_TAG_LEN			16
 #define OCELOT_SHORT_PREFIX_LEN		4
 #define OCELOT_LONG_PREFIX_LEN		16
+#define OCELOT_TOTAL_TAG_LEN	(OCELOT_SHORT_PREFIX_LEN + OCELOT_TAG_LEN)
 
 #define OCELOT_SPEED_2500		0
 #define OCELOT_SPEED_1000		1
@@ -672,9 +673,6 @@ void __ocelot_rmw_ix(struct ocelot *ocelot, u32 val, u32 mask, u32 reg,
 int ocelot_regfields_init(struct ocelot *ocelot,
 			  const struct reg_field *const regfields);
 struct regmap *ocelot_regmap_init(struct ocelot *ocelot, struct resource *res);
-void ocelot_configure_cpu(struct ocelot *ocelot, int npi,
-			  enum ocelot_tag_prefix injection,
-			  enum ocelot_tag_prefix extraction);
 int ocelot_init(struct ocelot *ocelot);
 void ocelot_deinit(struct ocelot *ocelot);
 void ocelot_init_port(struct ocelot *ocelot, int port);

net/core/flow_dissector.c

@@ -932,8 +932,14 @@ bool __skb_flow_dissect(const struct net *net,
 			int offset = 0;
 
 			ops = skb->dev->dsa_ptr->tag_ops;
-			if (ops->flow_dissect &&
-			    !ops->flow_dissect(skb, &proto, &offset)) {
+			/* Tail taggers don't break flow dissection */
+			if (!ops->tail_tag) {
+				if (ops->flow_dissect)
+					ops->flow_dissect(skb, &proto, &offset);
+				else
+					dsa_tag_generic_flow_dissect(skb,
+								     &proto,
+								     &offset);
 				hlen -= offset;
 				nhoff += offset;
 			}

net/dsa/master.c

@@ -259,6 +259,18 @@ static void dsa_netdev_ops_set(struct net_device *dev,
 	dev->dsa_ptr->netdev_ops = ops;
 }
 
+static void dsa_master_set_promiscuity(struct net_device *dev, int inc)
+{
+	const struct dsa_device_ops *ops = dev->dsa_ptr->tag_ops;
+
+	if (!ops->promisc_on_master)
+		return;
+
+	rtnl_lock();
+	dev_set_promiscuity(dev, inc);
+	rtnl_unlock();
+}
+
 static ssize_t tagging_show(struct device *d, struct device_attribute *attr,
 			    char *buf)
 {
@@ -314,9 +326,12 @@ int dsa_master_setup(struct net_device *dev, struct dsa_port *cpu_dp)
 	dev->dsa_ptr = cpu_dp;
 	lockdep_set_class(&dev->addr_list_lock,
 			  &dsa_master_addr_list_lock_key);
+
+	dsa_master_set_promiscuity(dev, 1);
+
 	ret = dsa_master_ethtool_setup(dev);
 	if (ret)
-		return ret;
+		goto out_err_reset_promisc;
 
 	dsa_netdev_ops_set(dev, &dsa_netdev_ops);
 
@@ -329,6 +344,8 @@ int dsa_master_setup(struct net_device *dev, struct dsa_port *cpu_dp)
 out_err_ndo_teardown:
 	dsa_netdev_ops_set(dev, NULL);
 	dsa_master_ethtool_teardown(dev);
+out_err_reset_promisc:
+	dsa_master_set_promiscuity(dev, -1);
 	return ret;
 }
 
@@ -338,6 +355,7 @@ void dsa_master_teardown(struct net_device *dev)
 	dsa_netdev_ops_set(dev, NULL);
 	dsa_master_ethtool_teardown(dev);
 	dsa_master_reset_mtu(dev);
+	dsa_master_set_promiscuity(dev, -1);
 
 	dev->dsa_ptr = NULL;