[PATCH net-next 0/9] mlxsw: Add support for 802.1ad bridging

[PATCH net-next 0/9] mlxsw: Add support for 802.1ad bridging

[Ido Schimmel]

From: Ido Schimmel <idosch@nvidia.com>

802.1ad, also known as QinQ, is an extension to the 802.1q standard,
which is concerned with passing possibly 802.1q-tagged packets through
another VLAN-like tunnel. The format of 802.1ad tag is the same as
802.1q, except it uses the EtherType of 0x88a8, unlike 802.1q's 0x8100.

Currently, mlxsw supports bridging with VLAN-unaware (802.1d) bridges
and with VLAN-aware bridges whose VLAN protocol is 802.1q. This set adds
support for VLAN-aware bridges whose VLAN protocol is 802.1ad.

From mlxsw perspective, 802.1ad support entails two main changes:

1. Ports member in an 802.1ad bridge need to be configured to classify
802.1ad packets as tagged and all other packets as untagged

2. When pushing a VLAN at ingress (PVID), its EtherType needs to be
0x88a8 instead of 802.1q's 0x8100

The rest stays the same as with 802.1q bridges.

A follow-up patch set will add support for QinQ with VXLAN, also known
as QinVNI. Currently, linking of a VXLAN netdev to an 802.1ad bridge is
vetoed and an error is returned to user space.

Patch set overview:

Patches #1-#2 add the registers required to configure the two changes
described above.

Patch #3 changes the device to only treat 802.1q packets as tagged by
default, as opposed to both 802.1q and 802.1ad packets. This is more
inline with the behavior supported by the driver.

Patch #4 adds the ability to configure the EtherType when pushing a PVID
at ingress.

Patch #5 performs small refactoring to allow for code re-use in the next
patch.

Patch #6 adds support for 802.1ad bridging and allows mlxsw ports and
their uppers to join such a bridge.

Patch #7 changes the bridge driver to notify about changes to its VLAN
protocol, so that these could be vetoed by mlxsw in the next patch.

Patches #8-#9 teach mlxsw to veto unsupported 802.1ad configurations and
add a corresponding selftest to make sure such configurations are indeed
vetoed.

Amit Cohen (6):
  mlxsw: reg: Add Switch Port VLAN Classification Register
  mlxsw: reg: Add et_vlan field to SPVID register
  mlxsw: spectrum: Only treat 802.1q packets as tagged packets
  mlxsw: Make EtherType configurable when pushing VLAN at ingress
  mlxsw: spectrum_switchdev: Create common functions for VLAN-aware
    bridge
  mlxsw: spectrum_switchdev: Add support of QinQ traffic

Danielle Ratson (3):
  bridge: switchdev: Notify about VLAN protocol changes
  mlxsw: Add QinQ configuration vetoes
  selftests: forwarding: Add QinQ veto testing

 drivers/net/ethernet/mellanox/mlxsw/reg.h     | 114 ++++++-
 .../net/ethernet/mellanox/mlxsw/spectrum.c    | 111 ++++++-
 .../net/ethernet/mellanox/mlxsw/spectrum.h    |   7 +-
 .../ethernet/mellanox/mlxsw/spectrum_router.c |   9 +
 .../mellanox/mlxsw/spectrum_switchdev.c       | 123 +++++++-
 include/net/switchdev.h                       |   2 +
 net/bridge/br_vlan.c                          |  16 +-
 .../drivers/net/mlxsw/q_in_q_veto.sh          | 296 ++++++++++++++++++
 8 files changed, 657 insertions(+), 21 deletions(-)
 create mode 100755 tools/testing/selftests/drivers/net/mlxsw/q_in_q_veto.sh

-- 
2.28.0

[PATCH net-next 1/9] mlxsw: reg: Add Switch Port VLAN Classification Register

[Ido Schimmel]

From: Amit Cohen <amcohen@nvidia.com>

SPVC configures the port to identify packets as untagged / single tagged /
double tagged packets based on the packet EtherTypes.

It will be used to classify 802.1q packets as untagged and 802.1ad packets
as tagged when received by ports member in a 802.1ad bridge.

Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlxsw/reg.h | 99 +++++++++++++++++++++++
 1 file changed, 99 insertions(+)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/reg.h b/drivers/net/ethernet/mellanox/mlxsw/reg.h
index fcf9095b3f55..5248532c3999 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/reg.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/reg.h
@@ -1857,6 +1857,104 @@ static inline void mlxsw_reg_spvmlr_pack(char *payload, u8 local_port,
 	}
 }
 
+/* SPVC - Switch Port VLAN Classification Register
+ * -----------------------------------------------
+ * Configures the port to identify packets as untagged / single tagged /
+ * double packets based on the packet EtherTypes.
+ * Ethertype IDs are configured by SVER.
+ */
+#define MLXSW_REG_SPVC_ID 0x2026
+#define MLXSW_REG_SPVC_LEN 0x0C
+
+MLXSW_REG_DEFINE(spvc, MLXSW_REG_SPVC_ID, MLXSW_REG_SPVC_LEN);
+
+/* reg_spvc_local_port
+ * Local port.
+ * Access: Index
+ *
+ * Note: applies both to Rx port and Tx port, so if a packet traverses
+ * through Rx port i and a Tx port j then port i and port j must have the
+ * same configuration.
+ */
+MLXSW_ITEM32(reg, spvc, local_port, 0x00, 16, 8);
+
+/* reg_spvc_inner_et2
+ * Vlan Tag1 EtherType2 enable.
+ * Packet is initially classified as double VLAN Tag if in addition to
+ * being classified with a tag0 VLAN Tag its tag1 EtherType value is
+ * equal to ether_type2.
+ * 0: disable (default)
+ * 1: enable
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, spvc, inner_et2, 0x08, 17, 1);
+
+/* reg_spvc_et2
+ * Vlan Tag0 EtherType2 enable.
+ * Packet is initially classified as VLAN Tag if its tag0 EtherType is
+ * equal to ether_type2.
+ * 0: disable (default)
+ * 1: enable
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, spvc, et2, 0x08, 16, 1);
+
+/* reg_spvc_inner_et1
+ * Vlan Tag1 EtherType1 enable.
+ * Packet is initially classified as double VLAN Tag if in addition to
+ * being classified with a tag0 VLAN Tag its tag1 EtherType value is
+ * equal to ether_type1.
+ * 0: disable
+ * 1: enable (default)
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, spvc, inner_et1, 0x08, 9, 1);
+
+/* reg_spvc_et1
+ * Vlan Tag0 EtherType1 enable.
+ * Packet is initially classified as VLAN Tag if its tag0 EtherType is
+ * equal to ether_type1.
+ * 0: disable
+ * 1: enable (default)
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, spvc, et1, 0x08, 8, 1);
+
+/* reg_inner_et0
+ * Vlan Tag1 EtherType0 enable.
+ * Packet is initially classified as double VLAN Tag if in addition to
+ * being classified with a tag0 VLAN Tag its tag1 EtherType value is
+ * equal to ether_type0.
+ * 0: disable
+ * 1: enable (default)
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, spvc, inner_et0, 0x08, 1, 1);
+
+/* reg_et0
+ * Vlan Tag0 EtherType0 enable.
+ * Packet is initially classified as VLAN Tag if its tag0 EtherType is
+ * equal to ether_type0.
+ * 0: disable
+ * 1: enable (default)
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, spvc, et0, 0x08, 0, 1);
+
+static inline void mlxsw_reg_spvc_pack(char *payload, u8 local_port, bool et1,
+				       bool et0)
+{
+	MLXSW_REG_ZERO(spvc, payload);
+	mlxsw_reg_spvc_local_port_set(payload, local_port);
+	/* Enable inner_et1 and inner_et0 to enable identification of double
+	 * tagged packets.
+	 */
+	mlxsw_reg_spvc_inner_et1_set(payload, 1);
+	mlxsw_reg_spvc_inner_et0_set(payload, 1);
+	mlxsw_reg_spvc_et1_set(payload, et1);
+	mlxsw_reg_spvc_et0_set(payload, et0);
+}
+
 /* CWTP - Congetion WRED ECN TClass Profile
  * ----------------------------------------
  * Configures the profiles for queues of egress port and traffic class
@@ -11212,6 +11310,7 @@ static const struct mlxsw_reg_info *mlxsw_reg_infos[] = {
 	MLXSW_REG(svpe),
 	MLXSW_REG(sfmr),
 	MLXSW_REG(spvmlr),
+	MLXSW_REG(spvc),
 	MLXSW_REG(cwtp),
 	MLXSW_REG(cwtpm),
 	MLXSW_REG(pgcr),
-- 
2.28.0

[PATCH net-next 2/9] mlxsw: reg: Add et_vlan field to SPVID register

[Ido Schimmel]

From: Amit Cohen <amcohen@nvidia.com>

et_vlan field is used to configure which EtherType is used when VLAN is
pushed at ingress (for untagged packets or for QinQ push mode).

It will be used to configure tagging with ether_type1 (i.e., 0x88A8) for
QinQ mode.

Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlxsw/reg.h | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/reg.h b/drivers/net/ethernet/mellanox/mlxsw/reg.h
index 5248532c3999..bea919b92f76 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/reg.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/reg.h
@@ -834,6 +834,17 @@ MLXSW_ITEM32(reg, spvid, local_port, 0x00, 16, 8);
  */
 MLXSW_ITEM32(reg, spvid, sub_port, 0x00, 8, 8);
 
+/* reg_spvid_et_vlan
+ * EtherType used for when VLAN is pushed at ingress (for untagged
+ * packets or for QinQ push mode).
+ * 0: ether_type0 - (default)
+ * 1: ether_type1
+ * 2: ether_type2 - Reserved when Spectrum-1, supported by Spectrum-2
+ * Ethertype IDs are configured by SVER.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, spvid, et_vlan, 0x04, 16, 2);
+
 /* reg_spvid_pvid
  * Port default VID
  * Access: RW
-- 
2.28.0

[PATCH net-next 3/9] mlxsw: spectrum: Only treat 802.1q packets as tagged packets

[Ido Schimmel]

From: Amit Cohen <amcohen@nvidia.com>

By default, the device considers both 802.1ad and 802.1q packets as tagged,
but this is not supported by the driver. It only supports VLAN and bridge
devices that use 802.1q protocol.

Instead, configure the device to only treat 802.1q packets as tagged
packets.

Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
---
 .../net/ethernet/mellanox/mlxsw/spectrum.c    | 26 +++++++++++++++++++
 1 file changed, 26 insertions(+)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
index b08853f71b2b..ee0c4d098c78 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
@@ -1386,6 +1386,19 @@ static int mlxsw_sp_port_overheat_init_val_set(struct mlxsw_sp_port *mlxsw_sp_po
 	return 0;
 }
 
+static int
+mlxsw_sp_port_vlan_classification_set(struct mlxsw_sp_port *mlxsw_sp_port,
+				      bool is_8021ad_tagged,
+				      bool is_8021q_tagged)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	char spvc_pl[MLXSW_REG_SPVC_LEN];
+
+	mlxsw_reg_spvc_pack(spvc_pl, mlxsw_sp_port->local_port,
+			    is_8021ad_tagged, is_8021q_tagged);
+	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(spvc), spvc_pl);
+}
+
 static int mlxsw_sp_port_create(struct mlxsw_sp *mlxsw_sp, u8 local_port,
 				u8 split_base_local_port,
 				struct mlxsw_sp_port_mapping *port_mapping)
@@ -1592,6 +1605,16 @@ static int mlxsw_sp_port_create(struct mlxsw_sp *mlxsw_sp, u8 local_port,
 	}
 	mlxsw_sp_port->default_vlan = mlxsw_sp_port_vlan;
 
+	/* Set SPVC.et0=true and SPVC.et1=false to make the local port to treat
+	 * only packets with 802.1q header as tagged packets.
+	 */
+	err = mlxsw_sp_port_vlan_classification_set(mlxsw_sp_port, false, true);
+	if (err) {
+		dev_err(mlxsw_sp->bus_info->dev, "Port %d: Failed to set default VLAN classification\n",
+			local_port);
+		goto err_port_vlan_classification_set;
+	}
+
 	INIT_DELAYED_WORK(&mlxsw_sp_port->ptp.shaper_dw,
 			  mlxsw_sp->ptp_ops->shaper_work);
 
@@ -1618,6 +1641,8 @@ static int mlxsw_sp_port_create(struct mlxsw_sp *mlxsw_sp, u8 local_port,
 
 err_register_netdev:
 err_port_overheat_init_val_set:
+	mlxsw_sp_port_vlan_classification_set(mlxsw_sp_port, true, true);
+err_port_vlan_classification_set:
 	mlxsw_sp->ports[local_port] = NULL;
 	mlxsw_sp_port_vlan_destroy(mlxsw_sp_port_vlan);
 err_port_vlan_create:
@@ -1664,6 +1689,7 @@ static void mlxsw_sp_port_remove(struct mlxsw_sp *mlxsw_sp, u8 local_port)
 	mlxsw_sp_port_ptp_clear(mlxsw_sp_port);
 	mlxsw_core_port_clear(mlxsw_sp->core, local_port, mlxsw_sp);
 	unregister_netdev(mlxsw_sp_port->dev); /* This calls ndo_stop */
+	mlxsw_sp_port_vlan_classification_set(mlxsw_sp_port, true, true);
 	mlxsw_sp->ports[local_port] = NULL;
 	mlxsw_sp_port_vlan_flush(mlxsw_sp_port, true);
 	mlxsw_sp_port_nve_fini(mlxsw_sp_port);
-- 
2.28.0

[PATCH net-next 4/9] mlxsw: Make EtherType configurable when pushing VLAN at ingress

[Ido Schimmel]

From: Amit Cohen <amcohen@nvidia.com>

Currently, when pushing a PVID at ingress, mlxsw always uses 802.1q
EtherType.

Make this EtherType configurable by extending mlxsw_sp_port_pvid_set()
with an EtherType argument.

This is a preparation for QinQ support, that needs to push a PVID with
802.1ad EtherType.

Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlxsw/reg.h     |  4 +-
 .../net/ethernet/mellanox/mlxsw/spectrum.c    | 41 +++++++++++++++----
 .../net/ethernet/mellanox/mlxsw/spectrum.h    |  3 +-
 .../mellanox/mlxsw/spectrum_switchdev.c       | 13 ++++--
 4 files changed, 48 insertions(+), 13 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/reg.h b/drivers/net/ethernet/mellanox/mlxsw/reg.h
index bea919b92f76..1077ed2046fe 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/reg.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/reg.h
@@ -851,11 +851,13 @@ MLXSW_ITEM32(reg, spvid, et_vlan, 0x04, 16, 2);
  */
 MLXSW_ITEM32(reg, spvid, pvid, 0x04, 0, 12);
 
-static inline void mlxsw_reg_spvid_pack(char *payload, u8 local_port, u16 pvid)
+static inline void mlxsw_reg_spvid_pack(char *payload, u8 local_port, u16 pvid,
+					u8 et_vlan)
 {
 	MLXSW_REG_ZERO(spvid, payload);
 	mlxsw_reg_spvid_local_port_set(payload, local_port);
 	mlxsw_reg_spvid_pvid_set(payload, pvid);
+	mlxsw_reg_spvid_et_vlan_set(payload, et_vlan);
 }
 
 /* SPVM - Switch Port VLAN Membership
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
index ee0c4d098c78..6ecd9a4dceee 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
@@ -384,13 +384,37 @@ int mlxsw_sp_port_vid_learning_set(struct mlxsw_sp_port *mlxsw_sp_port, u16 vid,
 	return err;
 }
 
+static int mlxsw_sp_ethtype_to_sver_type(u16 ethtype, u8 *p_sver_type)
+{
+	switch (ethtype) {
+	case ETH_P_8021Q:
+		*p_sver_type = 0;
+		break;
+	case ETH_P_8021AD:
+		*p_sver_type = 1;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
 static int __mlxsw_sp_port_pvid_set(struct mlxsw_sp_port *mlxsw_sp_port,
-				    u16 vid)
+				    u16 vid, u16 ethtype)
 {
 	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
 	char spvid_pl[MLXSW_REG_SPVID_LEN];
+	u8 sver_type;
+	int err;
+
+	err = mlxsw_sp_ethtype_to_sver_type(ethtype, &sver_type);
+	if (err)
+		return err;
+
+	mlxsw_reg_spvid_pack(spvid_pl, mlxsw_sp_port->local_port, vid,
+			     sver_type);
 
-	mlxsw_reg_spvid_pack(spvid_pl, mlxsw_sp_port->local_port, vid);
 	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(spvid), spvid_pl);
 }
 
@@ -404,7 +428,8 @@ static int mlxsw_sp_port_allow_untagged_set(struct mlxsw_sp_port *mlxsw_sp_port,
 	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(spaft), spaft_pl);
 }
 
-int mlxsw_sp_port_pvid_set(struct mlxsw_sp_port *mlxsw_sp_port, u16 vid)
+int mlxsw_sp_port_pvid_set(struct mlxsw_sp_port *mlxsw_sp_port, u16 vid,
+			   u16 ethtype)
 {
 	int err;
 
@@ -413,7 +438,7 @@ int mlxsw_sp_port_pvid_set(struct mlxsw_sp_port *mlxsw_sp_port, u16 vid)
 		if (err)
 			return err;
 	} else {
-		err = __mlxsw_sp_port_pvid_set(mlxsw_sp_port, vid);
+		err = __mlxsw_sp_port_pvid_set(mlxsw_sp_port, vid, ethtype);
 		if (err)
 			return err;
 		err = mlxsw_sp_port_allow_untagged_set(mlxsw_sp_port, true);
@@ -425,7 +450,7 @@ int mlxsw_sp_port_pvid_set(struct mlxsw_sp_port *mlxsw_sp_port, u16 vid)
 	return 0;
 
 err_port_allow_untagged_set:
-	__mlxsw_sp_port_pvid_set(mlxsw_sp_port, mlxsw_sp_port->pvid);
+	__mlxsw_sp_port_pvid_set(mlxsw_sp_port, mlxsw_sp_port->pvid, ethtype);
 	return err;
 }
 
@@ -1588,7 +1613,8 @@ static int mlxsw_sp_port_create(struct mlxsw_sp *mlxsw_sp, u8 local_port,
 		goto err_port_nve_init;
 	}
 
-	err = mlxsw_sp_port_pvid_set(mlxsw_sp_port, MLXSW_SP_DEFAULT_VID);
+	err = mlxsw_sp_port_pvid_set(mlxsw_sp_port, MLXSW_SP_DEFAULT_VID,
+				     ETH_P_8021Q);
 	if (err) {
 		dev_err(mlxsw_sp->bus_info->dev, "Port %d: Failed to set PVID\n",
 			mlxsw_sp_port->local_port);
@@ -3644,7 +3670,8 @@ static void mlxsw_sp_port_lag_leave(struct mlxsw_sp_port *mlxsw_sp_port,
 	lag->ref_count--;
 
 	/* Make sure untagged frames are allowed to ingress */
-	mlxsw_sp_port_pvid_set(mlxsw_sp_port, MLXSW_SP_DEFAULT_VID);
+	mlxsw_sp_port_pvid_set(mlxsw_sp_port, MLXSW_SP_DEFAULT_VID,
+			       ETH_P_8021Q);
 }
 
 static int mlxsw_sp_lag_dist_port_add(struct mlxsw_sp_port *mlxsw_sp_port,
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
index 642099fee380..338a4c9e329c 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
@@ -580,7 +580,8 @@ int mlxsw_sp_port_vid_stp_set(struct mlxsw_sp_port *mlxsw_sp_port, u16 vid,
 int mlxsw_sp_port_vp_mode_set(struct mlxsw_sp_port *mlxsw_sp_port, bool enable);
 int mlxsw_sp_port_vid_learning_set(struct mlxsw_sp_port *mlxsw_sp_port, u16 vid,
 				   bool learn_enable);
-int mlxsw_sp_port_pvid_set(struct mlxsw_sp_port *mlxsw_sp_port, u16 vid);
+int mlxsw_sp_port_pvid_set(struct mlxsw_sp_port *mlxsw_sp_port, u16 vid,
+			   u16 ethtype);
 struct mlxsw_sp_port_vlan *
 mlxsw_sp_port_vlan_create(struct mlxsw_sp_port *mlxsw_sp_port, u16 vid);
 void mlxsw_sp_port_vlan_destroy(struct mlxsw_sp_port_vlan *mlxsw_sp_port_vlan);
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
index 6501ce94ace5..a4aa2f620066 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
@@ -1129,6 +1129,7 @@ mlxsw_sp_bridge_port_vlan_add(struct mlxsw_sp_port *mlxsw_sp_port,
 	u16 pvid = mlxsw_sp_port_pvid_determine(mlxsw_sp_port, vid, is_pvid);
 	struct mlxsw_sp_port_vlan *mlxsw_sp_port_vlan;
 	u16 old_pvid = mlxsw_sp_port->pvid;
+	u16 proto;
 	int err;
 
 	/* The only valid scenario in which a port-vlan already exists, is if
@@ -1152,7 +1153,8 @@ mlxsw_sp_bridge_port_vlan_add(struct mlxsw_sp_port *mlxsw_sp_port,
 	if (err)
 		goto err_port_vlan_set;
 
-	err = mlxsw_sp_port_pvid_set(mlxsw_sp_port, pvid);
+	br_vlan_get_proto(bridge_port->bridge_device->dev, &proto);
+	err = mlxsw_sp_port_pvid_set(mlxsw_sp_port, pvid, proto);
 	if (err)
 		goto err_port_pvid_set;
 
@@ -1164,7 +1166,7 @@ mlxsw_sp_bridge_port_vlan_add(struct mlxsw_sp_port *mlxsw_sp_port,
 	return 0;
 
 err_port_vlan_bridge_join:
-	mlxsw_sp_port_pvid_set(mlxsw_sp_port, old_pvid);
+	mlxsw_sp_port_pvid_set(mlxsw_sp_port, old_pvid, proto);
 err_port_pvid_set:
 	mlxsw_sp_port_vlan_set(mlxsw_sp_port, vid, vid, false, false);
 err_port_vlan_set:
@@ -1821,13 +1823,15 @@ mlxsw_sp_bridge_port_vlan_del(struct mlxsw_sp_port *mlxsw_sp_port,
 {
 	u16 pvid = mlxsw_sp_port->pvid == vid ? 0 : mlxsw_sp_port->pvid;
 	struct mlxsw_sp_port_vlan *mlxsw_sp_port_vlan;
+	u16 proto;
 
 	mlxsw_sp_port_vlan = mlxsw_sp_port_vlan_find_by_vid(mlxsw_sp_port, vid);
 	if (WARN_ON(!mlxsw_sp_port_vlan))
 		return;
 
 	mlxsw_sp_port_vlan_bridge_leave(mlxsw_sp_port_vlan);
-	mlxsw_sp_port_pvid_set(mlxsw_sp_port, pvid);
+	br_vlan_get_proto(bridge_port->bridge_device->dev, &proto);
+	mlxsw_sp_port_pvid_set(mlxsw_sp_port, pvid, proto);
 	mlxsw_sp_port_vlan_set(mlxsw_sp_port, vid, vid, false, false);
 	mlxsw_sp_port_vlan_destroy(mlxsw_sp_port_vlan);
 }
@@ -1998,7 +2002,8 @@ mlxsw_sp_bridge_8021q_port_leave(struct mlxsw_sp_bridge_device *bridge_device,
 				 struct mlxsw_sp_port *mlxsw_sp_port)
 {
 	/* Make sure untagged frames are allowed to ingress */
-	mlxsw_sp_port_pvid_set(mlxsw_sp_port, MLXSW_SP_DEFAULT_VID);
+	mlxsw_sp_port_pvid_set(mlxsw_sp_port, MLXSW_SP_DEFAULT_VID,
+			       ETH_P_8021Q);
 }
 
 static int
-- 
2.28.0

[PATCH net-next 5/9] mlxsw: spectrum_switchdev: Create common functions for VLAN-aware bridge

[Ido Schimmel]

From: Amit Cohen <amcohen@nvidia.com>

The code in mlxsw_sp_bridge_8021q_port_{join, leave}() can be used also
for 802.1ad bridge.

Move the code to functions called
mlxsw_sp_bridge_vlan_aware_port_{join, leave}() and call them from
mlxsw_sp_bridge_8021q_port_{join, leave}() respectively to enable code
reuse.

Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
---
 .../mellanox/mlxsw/spectrum_switchdev.c       | 29 ++++++++++++++-----
 1 file changed, 22 insertions(+), 7 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
index a4aa2f620066..dfb97a847efc 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
@@ -1979,10 +1979,9 @@ static struct mlxsw_sp_port *mlxsw_sp_lag_rep_port(struct mlxsw_sp *mlxsw_sp,
 }
 
 static int
-mlxsw_sp_bridge_8021q_port_join(struct mlxsw_sp_bridge_device *bridge_device,
-				struct mlxsw_sp_bridge_port *bridge_port,
-				struct mlxsw_sp_port *mlxsw_sp_port,
-				struct netlink_ext_ack *extack)
+mlxsw_sp_bridge_vlan_aware_port_join(struct mlxsw_sp_bridge_port *bridge_port,
+				     struct mlxsw_sp_port *mlxsw_sp_port,
+				     struct netlink_ext_ack *extack)
 {
 	if (is_vlan_dev(bridge_port->dev)) {
 		NL_SET_ERR_MSG_MOD(extack, "Can not enslave a VLAN device to a VLAN-aware bridge");
@@ -1996,16 +1995,32 @@ mlxsw_sp_bridge_8021q_port_join(struct mlxsw_sp_bridge_device *bridge_device,
 	return 0;
 }
 
+static int
+mlxsw_sp_bridge_8021q_port_join(struct mlxsw_sp_bridge_device *bridge_device,
+				struct mlxsw_sp_bridge_port *bridge_port,
+				struct mlxsw_sp_port *mlxsw_sp_port,
+				struct netlink_ext_ack *extack)
+{
+	return mlxsw_sp_bridge_vlan_aware_port_join(bridge_port, mlxsw_sp_port,
+						    extack);
+}
+
 static void
-mlxsw_sp_bridge_8021q_port_leave(struct mlxsw_sp_bridge_device *bridge_device,
-				 struct mlxsw_sp_bridge_port *bridge_port,
-				 struct mlxsw_sp_port *mlxsw_sp_port)
+mlxsw_sp_bridge_vlan_aware_port_leave(struct mlxsw_sp_port *mlxsw_sp_port)
 {
 	/* Make sure untagged frames are allowed to ingress */
 	mlxsw_sp_port_pvid_set(mlxsw_sp_port, MLXSW_SP_DEFAULT_VID,
 			       ETH_P_8021Q);
 }
 
+static void
+mlxsw_sp_bridge_8021q_port_leave(struct mlxsw_sp_bridge_device *bridge_device,
+				 struct mlxsw_sp_bridge_port *bridge_port,
+				 struct mlxsw_sp_port *mlxsw_sp_port)
+{
+	mlxsw_sp_bridge_vlan_aware_port_leave(mlxsw_sp_port);
+}
+
 static int
 mlxsw_sp_bridge_8021q_vxlan_join(struct mlxsw_sp_bridge_device *bridge_device,
 				 const struct net_device *vxlan_dev, u16 vid,
-- 
2.28.0

[PATCH net-next 6/9] mlxsw: spectrum_switchdev: Add support of QinQ traffic

[Ido Schimmel]

From: Amit Cohen <amcohen@nvidia.com>

802.1ad, also known as QinQ is an extension to the 802.1q standard, which
is concerned with passing possibly 802.1q-tagged packets through another
VLAN-like tunnel. The format of 802.1ad tag is the same as 802.1q, except
it uses the EtherType of 0x88a8, unlike 802.1q's 0x8100.

Add support for 802.1ad protocol. Most of the configuration is the same
as 802.1q. The difference is that before a port joins an 802.1ad bridge it
needs to be configured to recognize 802.1ad packets as tagged and other
packets (e.g., 802.1q) as untagged.

VXLAN is not currently supported with 802.1ad bridge, so return an error
with an appropriate extack message.

Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
---
 .../net/ethernet/mellanox/mlxsw/spectrum.c    |  2 +-
 .../net/ethernet/mellanox/mlxsw/spectrum.h    |  4 ++
 .../mellanox/mlxsw/spectrum_switchdev.c       | 61 ++++++++++++++++++-
 3 files changed, 65 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
index 6ecd9a4dceee..fe954e87c5a7 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
@@ -1411,7 +1411,7 @@ static int mlxsw_sp_port_overheat_init_val_set(struct mlxsw_sp_port *mlxsw_sp_po
 	return 0;
 }
 
-static int
+int
 mlxsw_sp_port_vlan_classification_set(struct mlxsw_sp_port *mlxsw_sp_port,
 				      bool is_8021ad_tagged,
 				      bool is_8021q_tagged)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
index 338a4c9e329c..ce26cc41831f 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
@@ -428,6 +428,10 @@ int mlxsw_sp_port_get_stats_raw(struct net_device *dev, int grp,
 				int prio, char *ppcnt_pl);
 int mlxsw_sp_port_admin_status_set(struct mlxsw_sp_port *mlxsw_sp_port,
 				   bool is_up);
+int
+mlxsw_sp_port_vlan_classification_set(struct mlxsw_sp_port *mlxsw_sp_port,
+				      bool is_8021ad_tagged,
+				      bool is_8021q_tagged);
 
 /* spectrum_buffers.c */
 struct mlxsw_sp_hdroom_prio {
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
index dfb97a847efc..d8ee8801331c 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
@@ -41,6 +41,7 @@ struct mlxsw_sp_bridge {
 	DECLARE_BITMAP(mids_bitmap, MLXSW_SP_MID_MAX);
 	const struct mlxsw_sp_bridge_ops *bridge_8021q_ops;
 	const struct mlxsw_sp_bridge_ops *bridge_8021d_ops;
+	const struct mlxsw_sp_bridge_ops *bridge_8021ad_ops;
 };
 
 struct mlxsw_sp_bridge_device {
@@ -228,8 +229,14 @@ mlxsw_sp_bridge_device_create(struct mlxsw_sp_bridge *bridge,
 	bridge_device->mrouter = br_multicast_router(br_dev);
 	INIT_LIST_HEAD(&bridge_device->ports_list);
 	if (vlan_enabled) {
+		u16 proto;
+
 		bridge->vlan_enabled_exists = true;
-		bridge_device->ops = bridge->bridge_8021q_ops;
+		br_vlan_get_proto(br_dev, &proto);
+		if (proto == ETH_P_8021AD)
+			bridge_device->ops = bridge->bridge_8021ad_ops;
+		else
+			bridge_device->ops = bridge->bridge_8021q_ops;
 	} else {
 		bridge_device->ops = bridge->bridge_8021d_ops;
 	}
@@ -2266,6 +2273,57 @@ static const struct mlxsw_sp_bridge_ops mlxsw_sp_bridge_8021d_ops = {
 	.fid_vid	= mlxsw_sp_bridge_8021d_fid_vid,
 };
 
+static int
+mlxsw_sp_bridge_8021ad_port_join(struct mlxsw_sp_bridge_device *bridge_device,
+				 struct mlxsw_sp_bridge_port *bridge_port,
+				 struct mlxsw_sp_port *mlxsw_sp_port,
+				 struct netlink_ext_ack *extack)
+{
+	int err;
+
+	err = mlxsw_sp_port_vlan_classification_set(mlxsw_sp_port, true, false);
+	if (err)
+		return err;
+
+	err = mlxsw_sp_bridge_vlan_aware_port_join(bridge_port, mlxsw_sp_port,
+						   extack);
+	if (err)
+		goto err_bridge_vlan_aware_port_join;
+
+	return 0;
+
+err_bridge_vlan_aware_port_join:
+	mlxsw_sp_port_vlan_classification_set(mlxsw_sp_port, false, true);
+	return err;
+}
+
+static void
+mlxsw_sp_bridge_8021ad_port_leave(struct mlxsw_sp_bridge_device *bridge_device,
+				  struct mlxsw_sp_bridge_port *bridge_port,
+				  struct mlxsw_sp_port *mlxsw_sp_port)
+{
+	mlxsw_sp_bridge_vlan_aware_port_leave(mlxsw_sp_port);
+	mlxsw_sp_port_vlan_classification_set(mlxsw_sp_port, false, true);
+}
+
+static int
+mlxsw_sp_bridge_8021ad_vxlan_join(struct mlxsw_sp_bridge_device *bridge_device,
+				  const struct net_device *vxlan_dev, u16 vid,
+				  struct netlink_ext_ack *extack)
+{
+	NL_SET_ERR_MSG_MOD(extack, "VXLAN is not supported with 802.1ad");
+	return -EOPNOTSUPP;
+}
+
+static const struct mlxsw_sp_bridge_ops mlxsw_sp_bridge_8021ad_ops = {
+	.port_join	= mlxsw_sp_bridge_8021ad_port_join,
+	.port_leave	= mlxsw_sp_bridge_8021ad_port_leave,
+	.vxlan_join	= mlxsw_sp_bridge_8021ad_vxlan_join,
+	.fid_get	= mlxsw_sp_bridge_8021q_fid_get,
+	.fid_lookup	= mlxsw_sp_bridge_8021q_fid_lookup,
+	.fid_vid	= mlxsw_sp_bridge_8021q_fid_vid,
+};
+
 int mlxsw_sp_port_bridge_join(struct mlxsw_sp_port *mlxsw_sp_port,
 			      struct net_device *brport_dev,
 			      struct net_device *br_dev,
@@ -3527,6 +3585,7 @@ int mlxsw_sp_switchdev_init(struct mlxsw_sp *mlxsw_sp)
 
 	bridge->bridge_8021q_ops = &mlxsw_sp_bridge_8021q_ops;
 	bridge->bridge_8021d_ops = &mlxsw_sp_bridge_8021d_ops;
+	bridge->bridge_8021ad_ops = &mlxsw_sp_bridge_8021ad_ops;
 
 	return mlxsw_sp_fdb_init(mlxsw_sp);
 }
-- 
2.28.0

[PATCH net-next 7/9] bridge: switchdev: Notify about VLAN protocol changes

[Ido Schimmel]

From: Danielle Ratson <danieller@nvidia.com>

Drivers that support bridge offload need to be notified about changes to
the bridge's VLAN protocol so that they could react accordingly and
potentially veto the change.

Add a new switchdev attribute to communicate the change to drivers.

Signed-off-by: Danielle Ratson <danieller@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Acked-by: Nikolay Aleksandrov <nikolay@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
---
 include/net/switchdev.h |  2 ++
 net/bridge/br_vlan.c    | 16 ++++++++++++++--
 2 files changed, 16 insertions(+), 2 deletions(-)

diff --git a/include/net/switchdev.h b/include/net/switchdev.h
index 53e8b4994296..99cd538d6519 100644
--- a/include/net/switchdev.h
+++ b/include/net/switchdev.h
@@ -38,6 +38,7 @@ enum switchdev_attr_id {
 	SWITCHDEV_ATTR_ID_PORT_MROUTER,
 	SWITCHDEV_ATTR_ID_BRIDGE_AGEING_TIME,
 	SWITCHDEV_ATTR_ID_BRIDGE_VLAN_FILTERING,
+	SWITCHDEV_ATTR_ID_BRIDGE_VLAN_PROTOCOL,
 	SWITCHDEV_ATTR_ID_BRIDGE_MC_DISABLED,
 	SWITCHDEV_ATTR_ID_BRIDGE_MROUTER,
 #if IS_ENABLED(CONFIG_BRIDGE_MRP)
@@ -58,6 +59,7 @@ struct switchdev_attr {
 		bool mrouter;				/* PORT_MROUTER */
 		clock_t ageing_time;			/* BRIDGE_AGEING_TIME */
 		bool vlan_filtering;			/* BRIDGE_VLAN_FILTERING */
+		u16 vlan_protocol;			/* BRIDGE_VLAN_PROTOCOL */
 		bool mc_disabled;			/* MC_DISABLED */
 #if IS_ENABLED(CONFIG_BRIDGE_MRP)
 		u8 mrp_port_state;			/* MRP_PORT_STATE */
diff --git a/net/bridge/br_vlan.c b/net/bridge/br_vlan.c
index 11f54a7c0d1d..d07008678d32 100644
--- a/net/bridge/br_vlan.c
+++ b/net/bridge/br_vlan.c
@@ -854,15 +854,25 @@ EXPORT_SYMBOL_GPL(br_vlan_get_proto);
 
 int __br_vlan_set_proto(struct net_bridge *br, __be16 proto)
 {
+	struct switchdev_attr attr = {
+		.orig_dev = br->dev,
+		.id = SWITCHDEV_ATTR_ID_BRIDGE_VLAN_PROTOCOL,
+		.flags = SWITCHDEV_F_SKIP_EOPNOTSUPP,
+		.u.vlan_protocol = ntohs(proto),
+	};
 	int err = 0;
 	struct net_bridge_port *p;
 	struct net_bridge_vlan *vlan;
 	struct net_bridge_vlan_group *vg;
-	__be16 oldproto;
+	__be16 oldproto = br->vlan_proto;
 
 	if (br->vlan_proto == proto)
 		return 0;
 
+	err = switchdev_port_attr_set(br->dev, &attr);
+	if (err && err != -EOPNOTSUPP)
+		return err;
+
 	/* Add VLANs for the new proto to the device filter. */
 	list_for_each_entry(p, &br->port_list, list) {
 		vg = nbp_vlan_group(p);
@@ -873,7 +883,6 @@ int __br_vlan_set_proto(struct net_bridge *br, __be16 proto)
 		}
 	}
 
-	oldproto = br->vlan_proto;
 	br->vlan_proto = proto;
 
 	recalculate_group_addr(br);
@@ -889,6 +898,9 @@ int __br_vlan_set_proto(struct net_bridge *br, __be16 proto)
 	return 0;
 
 err_filt:
+	attr.u.vlan_protocol = ntohs(oldproto);
+	switchdev_port_attr_set(br->dev, &attr);
+
 	list_for_each_entry_continue_reverse(vlan, &vg->vlan_list, vlist)
 		vlan_vid_del(p->dev, proto, vlan->vid);
 
-- 
2.28.0

Re: [PATCH net-next 7/9] bridge: switchdev: Notify about VLAN protocol changes

[Ivan Vecera]

On Sun, 29 Nov 2020 14:54:05 +0200
Ido Schimmel <idosch@idosch.org> wrote:

> From: Danielle Ratson <danieller@nvidia.com>
> 
> Drivers that support bridge offload need to be notified about changes to
> the bridge's VLAN protocol so that they could react accordingly and
> potentially veto the change.
> 
> Add a new switchdev attribute to communicate the change to drivers.
> 
> Signed-off-by: Danielle Ratson <danieller@nvidia.com>
> Reviewed-by: Petr Machata <petrm@nvidia.com>
> Acked-by: Nikolay Aleksandrov <nikolay@nvidia.com>
> Signed-off-by: Ido Schimmel <idosch@nvidia.com>
> ---
>  include/net/switchdev.h |  2 ++
>  net/bridge/br_vlan.c    | 16 ++++++++++++++--
>  2 files changed, 16 insertions(+), 2 deletions(-)
> 
> diff --git a/include/net/switchdev.h b/include/net/switchdev.h
> index 53e8b4994296..99cd538d6519 100644
> --- a/include/net/switchdev.h
> +++ b/include/net/switchdev.h
> @@ -38,6 +38,7 @@ enum switchdev_attr_id {
>  	SWITCHDEV_ATTR_ID_PORT_MROUTER,
>  	SWITCHDEV_ATTR_ID_BRIDGE_AGEING_TIME,
>  	SWITCHDEV_ATTR_ID_BRIDGE_VLAN_FILTERING,
> +	SWITCHDEV_ATTR_ID_BRIDGE_VLAN_PROTOCOL,
>  	SWITCHDEV_ATTR_ID_BRIDGE_MC_DISABLED,
>  	SWITCHDEV_ATTR_ID_BRIDGE_MROUTER,
>  #if IS_ENABLED(CONFIG_BRIDGE_MRP)
> @@ -58,6 +59,7 @@ struct switchdev_attr {
>  		bool mrouter;				/* PORT_MROUTER */
>  		clock_t ageing_time;			/* BRIDGE_AGEING_TIME */
>  		bool vlan_filtering;			/* BRIDGE_VLAN_FILTERING */
> +		u16 vlan_protocol;			/* BRIDGE_VLAN_PROTOCOL */
>  		bool mc_disabled;			/* MC_DISABLED */
>  #if IS_ENABLED(CONFIG_BRIDGE_MRP)
>  		u8 mrp_port_state;			/* MRP_PORT_STATE */
> diff --git a/net/bridge/br_vlan.c b/net/bridge/br_vlan.c
> index 11f54a7c0d1d..d07008678d32 100644
> --- a/net/bridge/br_vlan.c
> +++ b/net/bridge/br_vlan.c
> @@ -854,15 +854,25 @@ EXPORT_SYMBOL_GPL(br_vlan_get_proto);
>  
>  int __br_vlan_set_proto(struct net_bridge *br, __be16 proto)
>  {
> +	struct switchdev_attr attr = {
> +		.orig_dev = br->dev,
> +		.id = SWITCHDEV_ATTR_ID_BRIDGE_VLAN_PROTOCOL,
> +		.flags = SWITCHDEV_F_SKIP_EOPNOTSUPP,
> +		.u.vlan_protocol = ntohs(proto),
> +	};
>  	int err = 0;
>  	struct net_bridge_port *p;
>  	struct net_bridge_vlan *vlan;
>  	struct net_bridge_vlan_group *vg;
> -	__be16 oldproto;
> +	__be16 oldproto = br->vlan_proto;
>  
>  	if (br->vlan_proto == proto)
>  		return 0;
>  
> +	err = switchdev_port_attr_set(br->dev, &attr);
> +	if (err && err != -EOPNOTSUPP)
> +		return err;
> +
>  	/* Add VLANs for the new proto to the device filter. */
>  	list_for_each_entry(p, &br->port_list, list) {
>  		vg = nbp_vlan_group(p);
> @@ -873,7 +883,6 @@ int __br_vlan_set_proto(struct net_bridge *br, __be16 proto)
>  		}
>  	}
>  
> -	oldproto = br->vlan_proto;
>  	br->vlan_proto = proto;
>  
>  	recalculate_group_addr(br);
> @@ -889,6 +898,9 @@ int __br_vlan_set_proto(struct net_bridge *br, __be16 proto)
>  	return 0;
>  
>  err_filt:
> +	attr.u.vlan_protocol = ntohs(oldproto);
> +	switchdev_port_attr_set(br->dev, &attr);
> +
>  	list_for_each_entry_continue_reverse(vlan, &vg->vlan_list, vlist)
>  		vlan_vid_del(p->dev, proto, vlan->vid);
>  

Reviewed-by: Ivan Vecera <ivecera@redhat.com>

[PATCH net-next 8/9] mlxsw: Add QinQ configuration vetoes

[Ido Schimmel]

From: Danielle Ratson <danieller@nvidia.com>

After adding support for QinQ, a.k.a 802.1ad protocol, there are a few
scenarios that should be vetoed.

The vetoes are motivated by various ASIC limitations.
For example, a port that is member in a 802.1ad bridge cannot have 802.1q
uppers as the port needs to be configured to treat 802.1q packets as
untagged packets.

Veto all those unsupported scenarios and return suitable messages.

Signed-off-by: Danielle Ratson <danieller@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
---
 .../net/ethernet/mellanox/mlxsw/spectrum.c    | 44 +++++++++++++++++++
 .../ethernet/mellanox/mlxsw/spectrum_router.c |  9 ++++
 .../mellanox/mlxsw/spectrum_switchdev.c       | 24 ++++++++++
 3 files changed, 77 insertions(+)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
index fe954e87c5a7..385eb3c3b362 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
@@ -3893,6 +3893,7 @@ static int mlxsw_sp_netdevice_port_upper_event(struct net_device *lower_dev,
 	struct net_device *upper_dev;
 	struct mlxsw_sp *mlxsw_sp;
 	int err = 0;
+	u16 proto;
 
 	mlxsw_sp_port = netdev_priv(dev);
 	mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
@@ -3950,6 +3951,36 @@ static int mlxsw_sp_netdevice_port_upper_event(struct net_device *lower_dev,
 			NL_SET_ERR_MSG_MOD(extack, "Can not put a VLAN on an OVS port");
 			return -EINVAL;
 		}
+		if (netif_is_bridge_master(upper_dev)) {
+			br_vlan_get_proto(upper_dev, &proto);
+			if (br_vlan_enabled(upper_dev) &&
+			    proto != ETH_P_8021Q && proto != ETH_P_8021AD) {
+				NL_SET_ERR_MSG_MOD(extack, "Enslaving a port to a bridge with unknown VLAN protocol is not supported");
+				return -EOPNOTSUPP;
+			}
+			if (vlan_uses_dev(lower_dev) &&
+			    br_vlan_enabled(upper_dev) &&
+			    proto == ETH_P_8021AD) {
+				NL_SET_ERR_MSG_MOD(extack, "Enslaving a port that already has a VLAN upper to an 802.1ad bridge is not supported");
+				return -EOPNOTSUPP;
+			}
+		}
+		if (netif_is_bridge_port(lower_dev) && is_vlan_dev(upper_dev)) {
+			struct net_device *br_dev = netdev_master_upper_dev_get(lower_dev);
+
+			if (br_vlan_enabled(br_dev)) {
+				br_vlan_get_proto(br_dev, &proto);
+				if (proto == ETH_P_8021AD) {
+					NL_SET_ERR_MSG_MOD(extack, "VLAN uppers are not supported on a port enslaved to an 802.1ad bridge");
+					return -EOPNOTSUPP;
+				}
+			}
+		}
+		if (is_vlan_dev(upper_dev) &&
+		    ntohs(vlan_dev_vlan_proto(upper_dev)) != ETH_P_8021Q) {
+			NL_SET_ERR_MSG_MOD(extack, "VLAN uppers are only supported with 802.1q VLAN protocol");
+			return -EOPNOTSUPP;
+		}
 		break;
 	case NETDEV_CHANGEUPPER:
 		upper_dev = info->upper_dev;
@@ -4215,6 +4246,7 @@ static int mlxsw_sp_netdevice_bridge_event(struct net_device *br_dev,
 	struct netdev_notifier_changeupper_info *info = ptr;
 	struct netlink_ext_ack *extack;
 	struct net_device *upper_dev;
+	u16 proto;
 
 	if (!mlxsw_sp)
 		return 0;
@@ -4230,6 +4262,18 @@ static int mlxsw_sp_netdevice_bridge_event(struct net_device *br_dev,
 		}
 		if (!info->linking)
 			break;
+		if (br_vlan_enabled(br_dev)) {
+			br_vlan_get_proto(br_dev, &proto);
+			if (proto == ETH_P_8021AD) {
+				NL_SET_ERR_MSG_MOD(extack, "Uppers are not supported on top of an 802.1ad bridge");
+				return -EOPNOTSUPP;
+			}
+		}
+		if (is_vlan_dev(upper_dev) &&
+		    ntohs(vlan_dev_vlan_proto(upper_dev)) != ETH_P_8021Q) {
+			NL_SET_ERR_MSG_MOD(extack, "VLAN uppers are only supported with 802.1q VLAN protocol");
+			return -EOPNOTSUPP;
+		}
 		if (netif_is_macvlan(upper_dev) &&
 		    !mlxsw_sp_rif_exists(mlxsw_sp, br_dev)) {
 			NL_SET_ERR_MSG_MOD(extack, "macvlan is only supported on top of router interfaces");
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index 12b5d7fbe1e2..85223647fdb6 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -7857,6 +7857,15 @@ static int mlxsw_sp_inetaddr_bridge_event(struct mlxsw_sp *mlxsw_sp,
 
 	switch (event) {
 	case NETDEV_UP:
+		if (netif_is_bridge_master(l3_dev) && br_vlan_enabled(l3_dev)) {
+			u16 proto;
+
+			br_vlan_get_proto(l3_dev, &proto);
+			if (proto == ETH_P_8021AD) {
+				NL_SET_ERR_MSG_MOD(extack, "Adding an IP address to 802.1ad bridge is not supported");
+				return -EOPNOTSUPP;
+			}
+		}
 		rif = mlxsw_sp_rif_create(mlxsw_sp, &params, extack);
 		if (IS_ERR(rif))
 			return PTR_ERR(rif);
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
index d8ee8801331c..9c4e17607e6a 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
@@ -764,6 +764,25 @@ static int mlxsw_sp_port_attr_br_vlan_set(struct mlxsw_sp_port *mlxsw_sp_port,
 	return -EINVAL;
 }
 
+static int mlxsw_sp_port_attr_br_vlan_proto_set(struct mlxsw_sp_port *mlxsw_sp_port,
+						struct switchdev_trans *trans,
+						struct net_device *orig_dev,
+						u16 vlan_proto)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	struct mlxsw_sp_bridge_device *bridge_device;
+
+	if (!switchdev_trans_ph_prepare(trans))
+		return 0;
+
+	bridge_device = mlxsw_sp_bridge_device_find(mlxsw_sp->bridge, orig_dev);
+	if (WARN_ON(!bridge_device))
+		return -EINVAL;
+
+	netdev_err(bridge_device->dev, "VLAN protocol can't be changed on existing bridge\n");
+	return -EINVAL;
+}
+
 static int mlxsw_sp_port_attr_mrouter_set(struct mlxsw_sp_port *mlxsw_sp_port,
 					  struct switchdev_trans *trans,
 					  struct net_device *orig_dev,
@@ -933,6 +952,11 @@ static int mlxsw_sp_port_attr_set(struct net_device *dev,
 						     attr->orig_dev,
 						     attr->u.vlan_filtering);
 		break;
+	case SWITCHDEV_ATTR_ID_BRIDGE_VLAN_PROTOCOL:
+		err = mlxsw_sp_port_attr_br_vlan_proto_set(mlxsw_sp_port, trans,
+							   attr->orig_dev,
+							   attr->u.vlan_protocol);
+		break;
 	case SWITCHDEV_ATTR_ID_PORT_MROUTER:
 		err = mlxsw_sp_port_attr_mrouter_set(mlxsw_sp_port, trans,
 						     attr->orig_dev,
-- 
2.28.0

[PATCH net-next 9/9] selftests: forwarding: Add QinQ veto testing

[Ido Schimmel]

From: Danielle Ratson <danieller@nvidia.com>

Test that each veto that was added in the previous patch, is indeed
vetoed.

$ ./q_in_q_veto.sh

TEST: create 802.1ad vlan upper on top of a front panel             [ OK ]
TEST: create 802.1ad vlan upper on top of a bridge port             [ OK ]
TEST: create 802.1ad vlan upper on top of a lag                     [ OK ]
TEST: create 802.1ad vlan upper on top 802.1q bridge                [ OK ]
TEST: create 802.1ad vlan upper on top 802.1ad bridge               [ OK ]
TEST: create 802.1q vlan upper on top 802.1ad bridge                [ OK ]
TEST: create vlan upper on top of front panel enslaved to 802.1ad bridge
[ OK ]
TEST: create vlan upper on top of lag enslaved to 802.1ad bridge    [ OK ]
TEST: enslave front panel with vlan upper to 802.1ad bridge         [ OK ]
TEST: enslave lag with vlan upper to 802.1ad bridge                 [ OK ]
TEST: IP address addition to 802.1ad bridge                         [ OK ]
TEST: switch bridge protocol                                        [ OK ]

Signed-off-by: Danielle Ratson <danieller@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
---
 .../drivers/net/mlxsw/q_in_q_veto.sh          | 296 ++++++++++++++++++
 1 file changed, 296 insertions(+)
 create mode 100755 tools/testing/selftests/drivers/net/mlxsw/q_in_q_veto.sh

diff --git a/tools/testing/selftests/drivers/net/mlxsw/q_in_q_veto.sh b/tools/testing/selftests/drivers/net/mlxsw/q_in_q_veto.sh
new file mode 100755
index 000000000000..7edaed8eb86a
--- /dev/null
+++ b/tools/testing/selftests/drivers/net/mlxsw/q_in_q_veto.sh
@@ -0,0 +1,296 @@
+#!/bin/bash
+# SPDX-License-Identifier: GPL-2.0
+
+lib_dir=$(dirname $0)/../../../net/forwarding
+
+ALL_TESTS="
+	create_8021ad_vlan_upper_on_top_front_panel_port
+	create_8021ad_vlan_upper_on_top_bridge_port
+	create_8021ad_vlan_upper_on_top_lag
+	create_8021ad_vlan_upper_on_top_bridge
+	create_8021ad_vlan_upper_on_top_8021ad_bridge
+	create_vlan_upper_on_top_8021ad_bridge
+	create_vlan_upper_on_top_front_panel_enslaved_to_8021ad_bridge
+	create_vlan_upper_on_top_lag_enslaved_to_8021ad_bridge
+	enslave_front_panel_with_vlan_upper_to_8021ad_bridge
+	enslave_lag_with_vlan_upper_to_8021ad_bridge
+	add_ip_address_to_8021ad_bridge
+	switch_bridge_protocol_from_8021q_to_8021ad
+"
+NUM_NETIFS=2
+source $lib_dir/lib.sh
+
+setup_prepare()
+{
+	swp1=${NETIFS[p1]}
+	swp2=${NETIFS[p2]}
+
+	ip link set dev $swp1 up
+	ip link set dev $swp2 up
+
+	sleep 10
+}
+
+cleanup()
+{
+	pre_cleanup
+
+	ip link set dev $swp2 down
+	ip link set dev $swp1 down
+}
+
+create_vlan_upper_on_top_of_bridge()
+{
+	RET=0
+
+	local bridge_proto=$1; shift
+	local netdev_proto=$1; shift
+
+	ip link add dev br0 type bridge vlan_filtering 1 \
+		vlan_protocol $bridge_proto vlan_default_pvid 0 mcast_snooping 0
+
+	ip link set dev br0 up
+	ip link set dev $swp1 master br0
+
+	ip link add name br0.100 link br0 type vlan \
+		protocol $netdev_proto id 100 2>/dev/null
+	check_fail $? "$netdev_proto vlan upper creation on top of an $bridge_proto bridge not rejected"
+
+	ip link add name br0.100 link br0 type vlan \
+		protocol $netdev_proto id 100 2>&1 >/dev/null \
+		| grep -q mlxsw_spectrum
+	check_err $? "$netdev_proto vlan upper creation on top of an $bridge_proto bridge rejected without extack"
+
+	log_test "create $netdev_proto vlan upper on top $bridge_proto bridge"
+
+	ip link del dev br0
+}
+
+create_8021ad_vlan_upper_on_top_front_panel_port()
+{
+	RET=0
+
+	ip link add name $swp1.100 link $swp1 type vlan \
+		protocol 802.1ad id 100 2>/dev/null
+	check_fail $? "802.1ad vlan upper creation on top of a front panel not rejected"
+
+	ip link add name $swp1.100 link $swp1 type vlan \
+		protocol 802.1ad id 100 2>&1 >/dev/null \
+		| grep -q mlxsw_spectrum
+	check_err $? "802.1ad vlan upper creation on top of a front panel rejected without extack"
+
+	log_test "create 802.1ad vlan upper on top of a front panel"
+}
+
+create_8021ad_vlan_upper_on_top_bridge_port()
+{
+	RET=0
+
+	ip link add dev br0 type bridge vlan_filtering 1 \
+		vlan_default_pvid 0 mcast_snooping 0
+
+	ip link set dev $swp1 master br0
+	ip link set dev br0 up
+
+	ip link add name $swp1.100 link $swp1 type vlan \
+		protocol 802.1ad id 100 2>/dev/null
+	check_fail $? "802.1ad vlan upper creation on top of a bridge port not rejected"
+
+	ip link add name $swp1.100 link $swp1 type vlan \
+		protocol 802.1ad id 100 2>&1 >/dev/null \
+		| grep -q mlxsw_spectrum
+	check_err $? "802.1ad vlan upper creation on top of a bridge port rejected without extack"
+
+	log_test "create 802.1ad vlan upper on top of a bridge port"
+
+	ip link del dev br0
+}
+
+create_8021ad_vlan_upper_on_top_lag()
+{
+	RET=0
+
+	ip link add name bond1 type bond mode 802.3ad
+	ip link set dev $swp1 down
+	ip link set dev $swp1 master bond1
+
+	ip link add name bond1.100 link bond1 type vlan \
+		protocol 802.1ad id 100 2>/dev/null
+	check_fail $? "802.1ad vlan upper creation on top of a lag not rejected"
+
+	ip link add name bond1.100 link bond1 type vlan \
+		protocol 802.1ad id 100 2>&1 >/dev/null \
+		| grep -q mlxsw_spectrum
+	check_err $? "802.1ad vlan upper creation on top of a lag rejected without extack"
+
+	log_test "create 802.1ad vlan upper on top of a lag"
+
+	ip link del dev bond1
+}
+
+create_8021ad_vlan_upper_on_top_bridge()
+{
+	RET=0
+
+	create_vlan_upper_on_top_of_bridge "802.1q" "802.1ad"
+}
+
+create_8021ad_vlan_upper_on_top_8021ad_bridge()
+{
+	RET=0
+
+	create_vlan_upper_on_top_of_bridge "802.1ad" "802.1ad"
+}
+
+create_vlan_upper_on_top_8021ad_bridge()
+{
+	RET=0
+
+	create_vlan_upper_on_top_of_bridge "802.1ad" "802.1q"
+}
+
+create_vlan_upper_on_top_front_panel_enslaved_to_8021ad_bridge()
+{
+	RET=0
+
+	ip link add dev br0 type bridge vlan_filtering 1 \
+		vlan_protocol 802.1ad vlan_default_pvid 0 mcast_snooping 0
+	ip link set dev br0 up
+
+	ip link set dev $swp1 master br0
+
+	ip link add name $swp1.100 link $swp1 type vlan id 100 2>/dev/null
+	check_fail $? "vlan upper creation on top of front panel enslaved to 802.1ad bridge not rejected"
+
+	ip link add name $swp1.100 link $swp1 type vlan id 100 2>&1 >/dev/null \
+		| grep -q mlxsw_spectrum
+	check_err $? "vlan upper creation on top of front panel enslaved to 802.1ad bridge rejected without extack"
+
+	log_test "create vlan upper on top of front panel enslaved to 802.1ad bridge"
+
+	ip link del dev br0
+}
+
+create_vlan_upper_on_top_lag_enslaved_to_8021ad_bridge()
+{
+	RET=0
+
+	ip link add dev br0 type bridge vlan_filtering 1 \
+		vlan_protocol 802.1ad vlan_default_pvid 0 mcast_snooping 0
+	ip link set dev br0 up
+
+	ip link add name bond1 type bond mode 802.3ad
+	ip link set dev $swp1 down
+	ip link set dev $swp1 master bond1
+	ip link set dev bond1 master br0
+
+	ip link add name bond1.100 link bond1 type vlan id 100 2>/dev/null
+	check_fail $? "vlan upper creation on top of lag enslaved to 802.1ad bridge not rejected"
+
+	ip link add name bond1.100 link bond1 type vlan id 100 2>&1 >/dev/null \
+		| grep -q mlxsw_spectrum
+	check_err $? "vlan upper creation on top of lag enslaved to 802.1ad bridge rejected without extack"
+
+	log_test "create vlan upper on top of lag enslaved to 802.1ad bridge"
+
+	ip link del dev bond1
+	ip link del dev br0
+}
+
+enslave_front_panel_with_vlan_upper_to_8021ad_bridge()
+{
+	RET=0
+
+	ip link add dev br0 type bridge vlan_filtering 1 \
+		vlan_protocol 802.1ad vlan_default_pvid 0 mcast_snooping 0
+	ip link set dev br0 up
+
+	ip link add name $swp1.100 link $swp1 type vlan id 100
+
+	ip link set dev $swp1 master br0 2>/dev/null
+	check_fail $? "front panel with vlan upper enslavemnt to 802.1ad bridge not rejected"
+
+	ip link set dev $swp1 master br0 2>&1 >/dev/null | grep -q mlxsw_spectrum
+	check_err $? "front panel with vlan upper enslavemnt to 802.1ad bridge rejected without extack"
+
+	log_test "enslave front panel with vlan upper to 802.1ad bridge"
+
+	ip link del dev $swp1.100
+	ip link del dev br0
+}
+
+enslave_lag_with_vlan_upper_to_8021ad_bridge()
+{
+	RET=0
+
+	ip link add dev br0 type bridge vlan_filtering 1 \
+		vlan_protocol 802.1ad vlan_default_pvid 0 mcast_snooping 0
+	ip link set dev br0 up
+
+	ip link add name bond1 type bond mode 802.3ad
+	ip link set dev $swp1 down
+	ip link set dev $swp1 master bond1
+	ip link add name bond1.100 link bond1 type vlan id 100
+
+	ip link set dev bond1 master br0 2>/dev/null
+	check_fail $? "lag with vlan upper enslavemnt to 802.1ad bridge not rejected"
+
+	ip link set dev bond1 master br0 2>&1 >/dev/null \
+		| grep -q mlxsw_spectrum
+	check_err $? "lag with vlan upper enslavemnt to 802.1ad bridge rejected without extack"
+
+	log_test "enslave lag with vlan upper to 802.1ad bridge"
+
+	ip link del dev bond1
+	ip link del dev br0
+}
+
+
+add_ip_address_to_8021ad_bridge()
+{
+	RET=0
+
+	ip link add dev br0 type bridge vlan_filtering 1 \
+		vlan_protocol 802.1ad vlan_default_pvid 0 mcast_snooping 0
+
+	ip link set dev br0 up
+	ip link set dev $swp1 master br0
+
+	ip addr add dev br0 192.0.2.17/28 2>/dev/null
+	check_fail $? "IP address addition to 802.1ad bridge not rejected"
+
+	ip addr add dev br0 192.0.2.17/28 2>&1 >/dev/null | grep -q mlxsw_spectrum
+	check_err $? "IP address addition to 802.1ad bridge rejected without extack"
+
+	log_test "IP address addition to 802.1ad bridge"
+
+	ip link del dev br0
+}
+
+switch_bridge_protocol_from_8021q_to_8021ad()
+{
+	RET=0
+
+	ip link add dev br0 type bridge vlan_filtering 1 \
+		vlan_protocol 802.1ad vlan_default_pvid 0 mcast_snooping 0
+
+	ip link set dev br0 up
+	ip link set dev $swp1 master br0
+
+	ip link set dev br0 type bridge vlan_protocol 802.1q 2>/dev/null
+	check_fail $? "switching bridge protocol from 802.1q to 802.1ad not rejected"
+
+	log_test "switch bridge protocol"
+
+	ip link del dev br0
+}
+
+
+trap cleanup EXIT
+
+setup_prepare
+setup_wait
+
+tests_run
+
+exit $EXIT_STATUS
-- 
2.28.0

Re: [PATCH net-next 0/9] mlxsw: Add support for 802.1ad bridging

[patchwork-bot+netdevbpf]

Hello:

This series was applied to netdev/net-next.git (refs/heads/master):

On Sun, 29 Nov 2020 14:53:58 +0200 you wrote:
> From: Ido Schimmel <idosch@nvidia.com>
> 
> 802.1ad, also known as QinQ, is an extension to the 802.1q standard,
> which is concerned with passing possibly 802.1q-tagged packets through
> another VLAN-like tunnel. The format of 802.1ad tag is the same as
> 802.1q, except it uses the EtherType of 0x88a8, unlike 802.1q's 0x8100.
> 
> [...]

Here is the summary with links:
  - [net-next,1/9] mlxsw: reg: Add Switch Port VLAN Classification Register
    https://git.kernel.org/netdev/net-next/c/7e9a6620d5c3
  - [net-next,2/9] mlxsw: reg: Add et_vlan field to SPVID register
    https://git.kernel.org/netdev/net-next/c/2a5a290d6d94
  - [net-next,3/9] mlxsw: spectrum: Only treat 802.1q packets as tagged packets
    https://git.kernel.org/netdev/net-next/c/a2ef3ae15834
  - [net-next,4/9] mlxsw: Make EtherType configurable when pushing VLAN at ingress
    https://git.kernel.org/netdev/net-next/c/3ae7a65b6424
  - [net-next,5/9] mlxsw: spectrum_switchdev: Create common functions for VLAN-aware bridge
    https://git.kernel.org/netdev/net-next/c/773ce33a4860
  - [net-next,6/9] mlxsw: spectrum_switchdev: Add support of QinQ traffic
    https://git.kernel.org/netdev/net-next/c/80dfeafd6479
  - [net-next,7/9] bridge: switchdev: Notify about VLAN protocol changes
    https://git.kernel.org/netdev/net-next/c/22ec19f3aee3
  - [net-next,8/9] mlxsw: Add QinQ configuration vetoes
    https://git.kernel.org/netdev/net-next/c/09139f67d346
  - [net-next,9/9] selftests: forwarding: Add QinQ veto testing
    https://git.kernel.org/netdev/net-next/c/008cb2ec4354

You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html