Skip to contentSkip to conversation

mlx5 updates 2023-04-11

GIT PULL Applied Saeed Mahameed 3 participants 18 messages Apr 12, 2023 – Apr 14, 2023

Subsystems (4)
lore
Share & tools
lore
#1 lore raw

PATCH 1/15 net/mlx5: Add mlx5_ifc definitions for bridge multicast support

In reply to Saeed Mahameed

From: Vlad Buslov <[email protected]>

Add the required hardware definitions to mlx5_ifc: fdb_uplink_hairpin, fdb_multi_path_any_table_limit_regc, fdb_multi_path_any_table.

---
 include/linux/mlx5/mlx5_ifc.h | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
include/linux/mlx5/mlx5_ifc.h +6 -1
index e47d6c58da35..02c628f4fe26 100644
@@ -880,7 +880,12 @@ enum {
 
 struct mlx5_ifc_flow_table_eswitch_cap_bits {
 	u8      fdb_to_vport_reg_c_id[0x8];
-	u8      reserved_at_8[0xd];
+	u8      reserved_at_8[0x5];
+	u8      fdb_uplink_hairpin[0x1];
+	u8      fdb_multi_path_any_table_limit_regc[0x1];
+	u8      reserved_at_f[0x3];
+	u8      fdb_multi_path_any_table[0x1];
+	u8      reserved_at_13[0x2];
 	u8      fdb_modify_header_fwd_to_table[0x1];
 	u8      fdb_ipv4_ttl_modify[0x1];
 	u8      flow_source[0x1];
Signature · 1 line
2.39.2
#0 lore raw

From: Saeed Mahameed <[email protected]>

This series provides some updates to mlx5 driver For more information please see detailed tag log below.

Please pull and let me know if there is any problem.

Thanks,
Saeed.

The following changes since commit 4de00f0acc722f43046dca06fe1336597d1250ab:

gve: Unify duplicate GQ min pkt desc size constants (2023-04-11 15:47:14 +0200)

are available in the Git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux.git tags/mlx5-updates-2023-04-11

for you to fetch changes up to 108ff8215b55903545abafa198f83624a20f44c8:

net/mlx5: DR, Add modify-header-pattern ICM pool (2023-04-11 20:57:38 -0700)

----------------------------------------------------------------
mlx5-updates-2023-04-11

1) Vlad adds the support for linux bridge multicast offload support Patches #1 through #9
Synopsis

Vlad Says:
==============
Implement support of bridge multicast offload in mlx5. Handle port object attribute SWITCHDEV_ATTR_ID_BRIDGE_MC_DISABLED notification to toggle multicast offload and bridge snooping support on bridge. Handle port object SWITCHDEV_OBJ_ID_PORT_MDB notification to attach a bridge port to MDB.

Steering architecture

Existing offload infrastructure relies on two levels of flow tables - bridge ingress and egress. For multicast offload the architecture is extended with additional layer of per-port multicast replication tables. Such tables filter loopback traffic (so packets are not replicated to their source port) and pop VLAN headers for "untagged" VLANs. The tables are referenced by the MDB rules in egress table. MDB egress rule can point to multiple per-port multicast tables, which causes matching multicast traffic to be replicated to all of them, and, consecutively, to several bridge ports:

                                                                                                                            +--------+--+
                                                                                    +---------------------------------------> Port 1 |  |
                                                                                    |                                       +-^------+--+
                                                                                    |                                         |
                                                                                    |                                         |
                                       +-----------------------------------------+  |     +---------------------------+       |
                                       | EGRESS table                            |  |  +--> PORT 1 multicast table    |       |
+----------------------------------+   +-----------------------------------------+  |  |  +---------------------------+       |
| INGRESS table                    |   |                                         |  |  |  |                           |       |
+----------------------------------+   | dst_mac=P1,vlan=X -> pop vlan, goto P1  +--+  |  | FG0:                      |       |
|                                  |   | dst_mac=P1,vlan=Y -> pop vlan, goto P1  |     |  | src_port=dst_port -> drop |       |
| src_mac=M1,vlan=X -> goto egress +---> dst_mac=P2,vlan=X -> pop vlan, goto P2  +--+  |  | FG1:                      |       |
| ...                              |   | dst_mac=P2,vlan=Y -> goto P2            |  |  |  | VLAN X -> pop, goto port  |       |
|                                  |   | dst_mac=MDB1,vlan=Y -> goto mcast P1,P2 +-----+  | ...                       |       |
+----------------------------------+   |                                         |  |  |  | VLAN Y -> pop, goto port  +-------+
                                       +-----------------------------------------+  |  |  | FG3:                      |
                                                                                    |  |  | matchall -> goto port     |
                                                                                    |  |  |                           |
                                                                                    |  |  +---------------------------+
                                                                                    |  |
                                                                                    |  |
                                                                                    |  |                                    +--------+--+
                                                                                    +---------------------------------------> Port 2 |  |
                                                                                       |                                    +-^------+--+
                                                                                       |                                      |
                                                                                       |                                      |
                                                                                       |  +---------------------------+       |
                                                                                       +--> PORT 2 multicast table    |       |
                                                                                          +---------------------------+       |
                                                                                          |                           |       |
                                                                                          | FG0:                      |       |
                                                                                          | src_port=dst_port -> drop |       |
                                                                                          | FG1:                      |       |
                                                                                          | VLAN X -> pop, goto port  |       |
                                                                                          | ...                       |       |
                                                                                          |                           |       |
                                                                                          | FG3:                      |       |
                                                                                          | matchall -> goto port     +-------+
                                                                                          |                           |
                                                                                          +---------------------------+

Patches overview:

- Patch 1 adds hardware definition bits for capabilities required to replicate multicast packets to multiple per-port tables. These bits are used by following patches to only attempt multicast offload if firmware and hardware provide necessary support.

- Pathces 2-4 patches are preparations and refactoring.

- Patch 5 implements necessary infrastructure to toggle multicast offload via SWITCHDEV_ATTR_ID_BRIDGE_MC_DISABLED port object attribute notification. This also enabled IGMP and MLD snooping.

- Patch 6 implements per-port multicast replication tables. It only supports filtering of loopback packets.

- Patch 7 extends per-port multicast tables with VLAN pop support for 'untagged' VLANs.

- Patch 8 handles SWITCHDEV_OBJ_ID_PORT_MDB port object notifications. It creates MDB replication rules in egress table that can replicate packets to multiple per-port multicast tables.

- Patch 9 adds tracepoints for MDB events.

==============

2) Parav Create a new allocation profile for SFs, to save on memory

3) Yevgeny provides some initial patches for upcoming software steering support new pattern/arguments type of modify_header actions.

Starting with ConnectX-6 DX, we use a new design of modify_header FW object. The current modify_header object allows for having only limited number of these FW objects, which means that we are limited in the number of offloaded flows that require modify_header action.

As a preparation Yevgeny provides the following 4 patches:
- Patch 1: Add required mlx5_ifc HW bits
- Patch 2, 3: Add new WQE type and opcode that is required for pattern/arg support and adds appropriate support in dr_send.c
- Patch 4: Add ICM pool for modify-header-pattern objects and implement patterns cache, allowing patterns reuse for different flows

----------------------------------------------------------------
Parav Pandit (1):
net/mlx5: Create a new profile for SFs

Vlad Buslov (9):
net/mlx5: Add mlx5_ifc definitions for bridge multicast support net/mlx5: Bridge, increase bridge tables sizes net/mlx5: Bridge, move additional data structures to priv header net/mlx5: Bridge, extract code to lookup parent bridge of port net/mlx5: Bridge, snoop igmp/mld packets
net/mlx5: Bridge, add per-port multicast replication tables net/mlx5: Bridge, support multicast VLAN pop net/mlx5: Bridge, implement mdb offload
net/mlx5: Bridge, add tracepoints for multicast

Yevgeny Kliteynik (5):
net/mlx5: DR, Set counter ID on the last STE for STEv1 TX net/mlx5: Add mlx5_ifc bits for modify header argument net/mlx5: Add new WQE for updating flow table net/mlx5: DR, Prepare sending new WQE type net/mlx5: DR, Add modify-header-pattern ICM pool

 drivers/net/ethernet/mellanox/mlx5/core/Makefile   |    4 +-
 drivers/net/ethernet/mellanox/mlx5/core/cmd.c      |    6 +-
 .../ethernet/mellanox/mlx5/core/en/rep/bridge.c    |   16 +
 .../net/ethernet/mellanox/mlx5/core/esw/bridge.c   |  287 +++--
 .../net/ethernet/mellanox/mlx5/core/esw/bridge.h   |   17 +
 .../ethernet/mellanox/mlx5/core/esw/bridge_mcast.c | 1126 ++++++++++++++++++++
 .../ethernet/mellanox/mlx5/core/esw/bridge_priv.h  |  181 ++++
 .../mlx5/core/esw/diag/bridge_tracepoint.h         |   35 +
 drivers/net/ethernet/mellanox/mlx5/core/fs_core.c  |    2 +-
 drivers/net/ethernet/mellanox/mlx5/core/main.c     |    9 +
 .../net/ethernet/mellanox/mlx5/core/mlx5_core.h    |    1 +
 .../ethernet/mellanox/mlx5/core/sf/dev/driver.c    |    2 +-
 .../ethernet/mellanox/mlx5/core/steering/dr_cmd.c  |    6 +
 .../mellanox/mlx5/core/steering/dr_domain.c        |   45 +-
 .../mellanox/mlx5/core/steering/dr_icm_pool.c      |   41 +-
 .../ethernet/mellanox/mlx5/core/steering/dr_ptrn.c |   43 +
 .../ethernet/mellanox/mlx5/core/steering/dr_send.c |   60 +-
 .../mellanox/mlx5/core/steering/dr_ste_v1.c        |    7 +-
 .../mellanox/mlx5/core/steering/dr_types.h         |   11 +
 include/linux/mlx5/device.h                        |    2 +
 include/linux/mlx5/driver.h                        |    1 +
 include/linux/mlx5/mlx5_ifc.h                      |   35 +-
 include/linux/mlx5/qp.h                            |   10 +
 23 files changed, 1783 insertions(+), 164 deletions(-)
 create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/esw/bridge_mcast.c
 create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ptrn.c
#3 lore raw

PATCH 2/15 net/mlx5: Bridge, increase bridge tables sizes

From: Vlad Buslov <[email protected]>

Bridge ingress and egress tables got more flow groups recently for QinQ support and will get more in following patches of this series. Increase the sizes of the tables to allow offloading more flows in each mode.

---
 drivers/net/ethernet/mellanox/mlx5/core/esw/bridge.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/esw/bridge.c +6 -6
index 3cdcb0e0b20f..e45f9bb80535 100644
@@ -13,8 +13,8 @@
 #define CREATE_TRACE_POINTS
 #include "diag/bridge_tracepoint.h"
 
-#define MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_GRP_SIZE 12000
-#define MLX5_ESW_BRIDGE_INGRESS_TABLE_UNTAGGED_GRP_SIZE 16000
+#define MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_GRP_SIZE 131072
+#define MLX5_ESW_BRIDGE_INGRESS_TABLE_UNTAGGED_GRP_SIZE 524288
 #define MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_GRP_IDX_FROM 0
 #define MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_GRP_IDX_TO		\
 	(MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_GRP_SIZE - 1)
@@ -40,10 +40,10 @@
 	 MLX5_ESW_BRIDGE_INGRESS_TABLE_UNTAGGED_GRP_SIZE - 1)
 #define MLX5_ESW_BRIDGE_INGRESS_TABLE_SIZE			\
 	(MLX5_ESW_BRIDGE_INGRESS_TABLE_MAC_GRP_IDX_TO + 1)
-static_assert(MLX5_ESW_BRIDGE_INGRESS_TABLE_SIZE == 64000);
+static_assert(MLX5_ESW_BRIDGE_INGRESS_TABLE_SIZE == 1048576);
 
-#define MLX5_ESW_BRIDGE_EGRESS_TABLE_VLAN_GRP_SIZE 16000
-#define MLX5_ESW_BRIDGE_EGRESS_TABLE_MAC_GRP_SIZE (32000 - 1)
+#define MLX5_ESW_BRIDGE_EGRESS_TABLE_VLAN_GRP_SIZE 131072
+#define MLX5_ESW_BRIDGE_EGRESS_TABLE_MAC_GRP_SIZE (262144 - 1)
 #define MLX5_ESW_BRIDGE_EGRESS_TABLE_VLAN_GRP_IDX_FROM 0
 #define MLX5_ESW_BRIDGE_EGRESS_TABLE_VLAN_GRP_IDX_TO		\
 	(MLX5_ESW_BRIDGE_EGRESS_TABLE_VLAN_GRP_SIZE - 1)
@@ -63,7 +63,7 @@ static_assert(MLX5_ESW_BRIDGE_INGRESS_TABLE_SIZE == 64000);
 	MLX5_ESW_BRIDGE_EGRESS_TABLE_MISS_GRP_IDX_FROM
 #define MLX5_ESW_BRIDGE_EGRESS_TABLE_SIZE			\
 	(MLX5_ESW_BRIDGE_EGRESS_TABLE_MISS_GRP_IDX_TO + 1)
-static_assert(MLX5_ESW_BRIDGE_EGRESS_TABLE_SIZE == 64000);
+static_assert(MLX5_ESW_BRIDGE_EGRESS_TABLE_SIZE == 524288);
 
 #define MLX5_ESW_BRIDGE_SKIP_TABLE_SIZE 0
 
Signature · 1 line
2.39.2
#4 lore raw

PATCH 3/15 net/mlx5: Bridge, move additional data structures to priv header

In reply to Saeed Mahameed

From: Vlad Buslov <[email protected]>

Following patches in series will require accessing flow tables and groups sizes, table levels and struct mlx5_esw_bridge from new the new source file dedicated to multicast code. Expose these data in bridge_priv.h to reduce clutter in following patches that will implement the actual functionality.

---
 .../ethernet/mellanox/mlx5/core/esw/bridge.c  | 85 -------------------
 .../mellanox/mlx5/core/esw/bridge_priv.h      | 85 +++++++++++++++++++
 2 files changed, 85 insertions(+), 85 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/esw/bridge.c +0 -85
index e45f9bb80535..ec052fff7712 100644
@@ -13,66 +13,6 @@
 #define CREATE_TRACE_POINTS
 #include "diag/bridge_tracepoint.h"
 
-#define MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_GRP_SIZE 131072
-#define MLX5_ESW_BRIDGE_INGRESS_TABLE_UNTAGGED_GRP_SIZE 524288
-#define MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_GRP_IDX_FROM 0
-#define MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_GRP_IDX_TO		\
-	(MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_GRP_SIZE - 1)
-#define MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_FILTER_GRP_IDX_FROM	\
-	(MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_GRP_IDX_TO + 1)
-#define MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_FILTER_GRP_IDX_TO		\
-	(MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_FILTER_GRP_IDX_FROM +	\
-	 MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_GRP_SIZE - 1)
-#define MLX5_ESW_BRIDGE_INGRESS_TABLE_QINQ_GRP_IDX_FROM			\
-	(MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_FILTER_GRP_IDX_TO + 1)
-#define MLX5_ESW_BRIDGE_INGRESS_TABLE_QINQ_GRP_IDX_TO			\
-	(MLX5_ESW_BRIDGE_INGRESS_TABLE_QINQ_GRP_IDX_FROM +		\
-	 MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_GRP_SIZE - 1)
-#define MLX5_ESW_BRIDGE_INGRESS_TABLE_QINQ_FILTER_GRP_IDX_FROM	\
-	(MLX5_ESW_BRIDGE_INGRESS_TABLE_QINQ_GRP_IDX_TO + 1)
-#define MLX5_ESW_BRIDGE_INGRESS_TABLE_QINQ_FILTER_GRP_IDX_TO		\
-	(MLX5_ESW_BRIDGE_INGRESS_TABLE_QINQ_FILTER_GRP_IDX_FROM +	\
-	 MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_GRP_SIZE - 1)
-#define MLX5_ESW_BRIDGE_INGRESS_TABLE_MAC_GRP_IDX_FROM			\
-	(MLX5_ESW_BRIDGE_INGRESS_TABLE_QINQ_FILTER_GRP_IDX_TO + 1)
-#define MLX5_ESW_BRIDGE_INGRESS_TABLE_MAC_GRP_IDX_TO			\
-	(MLX5_ESW_BRIDGE_INGRESS_TABLE_MAC_GRP_IDX_FROM +		\
-	 MLX5_ESW_BRIDGE_INGRESS_TABLE_UNTAGGED_GRP_SIZE - 1)
-#define MLX5_ESW_BRIDGE_INGRESS_TABLE_SIZE			\
-	(MLX5_ESW_BRIDGE_INGRESS_TABLE_MAC_GRP_IDX_TO + 1)
-static_assert(MLX5_ESW_BRIDGE_INGRESS_TABLE_SIZE == 1048576);
-
-#define MLX5_ESW_BRIDGE_EGRESS_TABLE_VLAN_GRP_SIZE 131072
-#define MLX5_ESW_BRIDGE_EGRESS_TABLE_MAC_GRP_SIZE (262144 - 1)
-#define MLX5_ESW_BRIDGE_EGRESS_TABLE_VLAN_GRP_IDX_FROM 0
-#define MLX5_ESW_BRIDGE_EGRESS_TABLE_VLAN_GRP_IDX_TO		\
-	(MLX5_ESW_BRIDGE_EGRESS_TABLE_VLAN_GRP_SIZE - 1)
-#define MLX5_ESW_BRIDGE_EGRESS_TABLE_QINQ_GRP_IDX_FROM		\
-	(MLX5_ESW_BRIDGE_EGRESS_TABLE_VLAN_GRP_IDX_TO + 1)
-#define MLX5_ESW_BRIDGE_EGRESS_TABLE_QINQ_GRP_IDX_TO			\
-	(MLX5_ESW_BRIDGE_EGRESS_TABLE_QINQ_GRP_IDX_FROM +		\
-	 MLX5_ESW_BRIDGE_EGRESS_TABLE_VLAN_GRP_SIZE - 1)
-#define MLX5_ESW_BRIDGE_EGRESS_TABLE_MAC_GRP_IDX_FROM \
-	(MLX5_ESW_BRIDGE_EGRESS_TABLE_QINQ_GRP_IDX_TO + 1)
-#define MLX5_ESW_BRIDGE_EGRESS_TABLE_MAC_GRP_IDX_TO			\
-	(MLX5_ESW_BRIDGE_EGRESS_TABLE_MAC_GRP_IDX_FROM +		\
-	 MLX5_ESW_BRIDGE_EGRESS_TABLE_MAC_GRP_SIZE - 1)
-#define MLX5_ESW_BRIDGE_EGRESS_TABLE_MISS_GRP_IDX_FROM \
-	(MLX5_ESW_BRIDGE_EGRESS_TABLE_MAC_GRP_IDX_TO + 1)
-#define MLX5_ESW_BRIDGE_EGRESS_TABLE_MISS_GRP_IDX_TO	\
-	MLX5_ESW_BRIDGE_EGRESS_TABLE_MISS_GRP_IDX_FROM
-#define MLX5_ESW_BRIDGE_EGRESS_TABLE_SIZE			\
-	(MLX5_ESW_BRIDGE_EGRESS_TABLE_MISS_GRP_IDX_TO + 1)
-static_assert(MLX5_ESW_BRIDGE_EGRESS_TABLE_SIZE == 524288);
-
-#define MLX5_ESW_BRIDGE_SKIP_TABLE_SIZE 0
-
-enum {
-	MLX5_ESW_BRIDGE_LEVEL_INGRESS_TABLE,
-	MLX5_ESW_BRIDGE_LEVEL_EGRESS_TABLE,
-	MLX5_ESW_BRIDGE_LEVEL_SKIP_TABLE,
-};
-
 static const struct rhashtable_params fdb_ht_params = {
 	.key_offset = offsetof(struct mlx5_esw_bridge_fdb_entry, key),
 	.key_len = sizeof(struct mlx5_esw_bridge_fdb_key),
@@ -80,31 +20,6 @@ static const struct rhashtable_params fdb_ht_params = {
 	.automatic_shrinking = true,
 };
 
-enum {
-	MLX5_ESW_BRIDGE_VLAN_FILTERING_FLAG = BIT(0),
-};
-
-struct mlx5_esw_bridge {
-	int ifindex;
-	int refcnt;
-	struct list_head list;
-	struct mlx5_esw_bridge_offloads *br_offloads;
-
-	struct list_head fdb_list;
-	struct rhashtable fdb_ht;
-
-	struct mlx5_flow_table *egress_ft;
-	struct mlx5_flow_group *egress_vlan_fg;
-	struct mlx5_flow_group *egress_qinq_fg;
-	struct mlx5_flow_group *egress_mac_fg;
-	struct mlx5_flow_group *egress_miss_fg;
-	struct mlx5_pkt_reformat *egress_miss_pkt_reformat;
-	struct mlx5_flow_handle *egress_miss_handle;
-	unsigned long ageing_time;
-	u32 flags;
-	u16 vlan_proto;
-};
-
 static void
 mlx5_esw_bridge_fdb_offload_notify(struct net_device *dev, const unsigned char *addr, u16 vid,
 				   unsigned long val)
drivers/net/ethernet/mellanox/mlx5/core/esw/bridge_priv.h +85 -0

95 lines not on this page · raw message

Signature · 1 line
2.39.2
#5 lore raw

PATCH 4/15 net/mlx5: Bridge, extract code to lookup parent bridge of port

In reply to Saeed Mahameed

From: Vlad Buslov <[email protected]>

The pattern when function looks up a port by vport_num+vhca_id tuple in order to just obtain its parent bridge is repeated multiple times in bridge.c file. Further commits in this series use the pattern even more. Extract the pattern to standalone mlx5_esw_bridge_from_port_lookup() function to improve code readability.

This commits doesn't change functionality.

---
 .../ethernet/mellanox/mlx5/core/esw/bridge.c  | 47 ++++++++++---------
 1 file changed, 26 insertions(+), 21 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/esw/bridge.c +26 -21
index ec052fff7712..bbbf982bbbc0 100644
@@ -933,6 +933,19 @@ static void mlx5_esw_bridge_port_erase(struct mlx5_esw_bridge_port *port,
 	xa_erase(&br_offloads->ports, mlx5_esw_bridge_port_key(port));
 }
 
+static struct mlx5_esw_bridge *
+mlx5_esw_bridge_from_port_lookup(u16 vport_num, u16 esw_owner_vhca_id,
+				 struct mlx5_esw_bridge_offloads *br_offloads)
+{
+	struct mlx5_esw_bridge_port *port;
+
+	port = mlx5_esw_bridge_port_lookup(vport_num, esw_owner_vhca_id, br_offloads);
+	if (!port)
+		return NULL;
+
+	return port->bridge;
+}
+
 static void mlx5_esw_bridge_fdb_entry_refresh(struct mlx5_esw_bridge_fdb_entry *entry)
 {
 	trace_mlx5_esw_bridge_fdb_entry_refresh(entry);
@@ -1388,28 +1401,26 @@ mlx5_esw_bridge_fdb_entry_init(struct net_device *dev, u16 vport_num, u16 esw_ow
 int mlx5_esw_bridge_ageing_time_set(u16 vport_num, u16 esw_owner_vhca_id, unsigned long ageing_time,
 				    struct mlx5_esw_bridge_offloads *br_offloads)
 {
-	struct mlx5_esw_bridge_port *port;
+	struct mlx5_esw_bridge *bridge;
 
-	port = mlx5_esw_bridge_port_lookup(vport_num, esw_owner_vhca_id, br_offloads);
-	if (!port)
+	bridge = mlx5_esw_bridge_from_port_lookup(vport_num, esw_owner_vhca_id, br_offloads);
+	if (!bridge)
 		return -EINVAL;
 
-	port->bridge->ageing_time = clock_t_to_jiffies(ageing_time);
+	bridge->ageing_time = clock_t_to_jiffies(ageing_time);
 	return 0;
 }
 
 int mlx5_esw_bridge_vlan_filtering_set(u16 vport_num, u16 esw_owner_vhca_id, bool enable,
 				       struct mlx5_esw_bridge_offloads *br_offloads)
 {
-	struct mlx5_esw_bridge_port *port;
 	struct mlx5_esw_bridge *bridge;
 	bool filtering;
 
-	port = mlx5_esw_bridge_port_lookup(vport_num, esw_owner_vhca_id, br_offloads);
-	if (!port)
+	bridge = mlx5_esw_bridge_from_port_lookup(vport_num, esw_owner_vhca_id, br_offloads);
+	if (!bridge)
 		return -EINVAL;
 
-	bridge = port->bridge;
 	filtering = bridge->flags & MLX5_ESW_BRIDGE_VLAN_FILTERING_FLAG;
 	if (filtering == enable)
 		return 0;
@@ -1426,15 +1437,13 @@ int mlx5_esw_bridge_vlan_filtering_set(u16 vport_num, u16 esw_owner_vhca_id, boo
 int mlx5_esw_bridge_vlan_proto_set(u16 vport_num, u16 esw_owner_vhca_id, u16 proto,
 				   struct mlx5_esw_bridge_offloads *br_offloads)
 {
-	struct mlx5_esw_bridge_port *port;
 	struct mlx5_esw_bridge *bridge;
 
-	port = mlx5_esw_bridge_port_lookup(vport_num, esw_owner_vhca_id,
-					   br_offloads);
-	if (!port)
+	bridge = mlx5_esw_bridge_from_port_lookup(vport_num, esw_owner_vhca_id,
+						  br_offloads);
+	if (!bridge)
 		return -EINVAL;
 
-	bridge = port->bridge;
 	if (bridge->vlan_proto == proto)
 		return 0;
 	if (proto != ETH_P_8021Q && proto != ETH_P_8021AD) {
@@ -1626,14 +1635,12 @@ void mlx5_esw_bridge_fdb_update_used(struct net_device *dev, u16 vport_num, u16
 				     struct switchdev_notifier_fdb_info *fdb_info)
 {
 	struct mlx5_esw_bridge_fdb_entry *entry;
-	struct mlx5_esw_bridge_port *port;
 	struct mlx5_esw_bridge *bridge;
 
-	port = mlx5_esw_bridge_port_lookup(vport_num, esw_owner_vhca_id, br_offloads);
-	if (!port)
+	bridge = mlx5_esw_bridge_from_port_lookup(vport_num, esw_owner_vhca_id, br_offloads);
+	if (!bridge)
 		return;
 
-	bridge = port->bridge;
 	entry = mlx5_esw_bridge_fdb_lookup(bridge, fdb_info->addr, fdb_info->vid);
 	if (!entry) {
 		esw_debug(br_offloads->esw->dev,
@@ -1680,14 +1687,12 @@ void mlx5_esw_bridge_fdb_remove(struct net_device *dev, u16 vport_num, u16 esw_o
 {
 	struct mlx5_eswitch *esw = br_offloads->esw;
 	struct mlx5_esw_bridge_fdb_entry *entry;
-	struct mlx5_esw_bridge_port *port;
 	struct mlx5_esw_bridge *bridge;
 
-	port = mlx5_esw_bridge_port_lookup(vport_num, esw_owner_vhca_id, br_offloads);
-	if (!port)
+	bridge = mlx5_esw_bridge_from_port_lookup(vport_num, esw_owner_vhca_id, br_offloads);
+	if (!bridge)
 		return;
 
-	bridge = port->bridge;
 	entry = mlx5_esw_bridge_fdb_lookup(bridge, fdb_info->addr, fdb_info->vid);
 	if (!entry) {
 		esw_debug(esw->dev,
Signature · 1 line
2.39.2
#6 lore raw

PATCH 5/15 net/mlx5: Bridge, snoop igmp/mld packets

In reply to Saeed Mahameed

From: Vlad Buslov <[email protected]>

Handle SWITCHDEV_ATTR_ID_BRIDGE_MC_DISABLED attribute notification to dynamically toggle bridge multicast offload. Set new MLX5_ESW_BRIDGE_MCAST_FLAG bridge flag when multicast offload is enabled. Put multicast-specific code into new bridge_mcast.c file.

When initializing bridge multicast pipeline create a static rule for snooping on IGMP traffic and three rules for snooping on MLD traffic (for query, report and done message types). Note that matching MLD traffic requires having flexparser MLX5_FLEX_PROTO_ICMPV6 capability enabled.

By default Linux bridge is created with multicast enabled which can be modified by 'mcast_snooping' argument:

$ ip link set name my_bridge type bridge mcast_snooping 0
---
 .../net/ethernet/mellanox/mlx5/core/Makefile  |   2 +-
 .../mellanox/mlx5/core/en/rep/bridge.c        |   4 +
 .../ethernet/mellanox/mlx5/core/esw/bridge.c  |  31 ++
 .../ethernet/mellanox/mlx5/core/esw/bridge.h  |   9 +
 .../mellanox/mlx5/core/esw/bridge_mcast.c     | 316 ++++++++++++++++++
 .../mellanox/mlx5/core/esw/bridge_priv.h      |  27 +-
 6 files changed, 384 insertions(+), 5 deletions(-)
 create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/esw/bridge_mcast.c
drivers/net/ethernet/mellanox/mlx5/core/Makefile +1 -1
index 6c2f1d4a58ab..4a009b720bee 100644
@@ -75,7 +75,7 @@ mlx5_core-$(CONFIG_MLX5_ESWITCH)   += esw/acl/helper.o \
 				      esw/acl/egress_lgcy.o esw/acl/egress_ofld.o \
 				      esw/acl/ingress_lgcy.o esw/acl/ingress_ofld.o
 
-mlx5_core-$(CONFIG_MLX5_BRIDGE)    += esw/bridge.o en/rep/bridge.o
+mlx5_core-$(CONFIG_MLX5_BRIDGE)    += esw/bridge.o esw/bridge_mcast.o en/rep/bridge.o
 
 mlx5_core-$(CONFIG_THERMAL)        += thermal.o
 mlx5_core-$(CONFIG_MLX5_MPFS)      += lib/mpfs.o
drivers/net/ethernet/mellanox/mlx5/core/en/rep/bridge.c +4 -0
index ce85b48d327d..9701eb3c5f1b 100644
@@ -306,6 +306,10 @@ mlx5_esw_bridge_port_obj_attr_set(struct net_device *dev,
 						     attr->u.vlan_protocol,
 						     br_offloads);
 		break;
+	case SWITCHDEV_ATTR_ID_BRIDGE_MC_DISABLED:
+		err = mlx5_esw_bridge_mcast_set(vport_num, esw_owner_vhca_id,
+						!attr->u.mc_disabled, br_offloads);
+		break;
 	default:
 		err = -EOPNOTSUPP;
 	}
drivers/net/ethernet/mellanox/mlx5/core/esw/bridge.c +31 -0
index bbbf982bbbc0..35436aa9548d 100644
@@ -868,6 +868,7 @@ static void mlx5_esw_bridge_put(struct mlx5_esw_bridge_offloads *br_offloads,
 		return;
 
 	mlx5_esw_bridge_egress_table_cleanup(bridge);
+	mlx5_esw_bridge_mcast_disable(bridge);
 	list_del(&bridge->list);
 	rhashtable_destroy(&bridge->fdb_ht);
 	kvfree(bridge);
@@ -1458,6 +1459,36 @@ int mlx5_esw_bridge_vlan_proto_set(u16 vport_num, u16 esw_owner_vhca_id, u16 pro
 	return 0;
 }
 
+int mlx5_esw_bridge_mcast_set(u16 vport_num, u16 esw_owner_vhca_id, bool enable,
+			      struct mlx5_esw_bridge_offloads *br_offloads)
+{
+	struct mlx5_eswitch *esw = br_offloads->esw;
+	struct mlx5_esw_bridge *bridge;
+	int err = 0;
+	bool mcast;
+
+	if (!(MLX5_CAP_ESW_FLOWTABLE((esw)->dev, fdb_multi_path_any_table) ||
+	      MLX5_CAP_ESW_FLOWTABLE((esw)->dev, fdb_multi_path_any_table_limit_regc)) ||
+	    !MLX5_CAP_ESW_FLOWTABLE((esw)->dev, fdb_uplink_hairpin) ||
+	    !MLX5_CAP_ESW_FLOWTABLE_FDB((esw)->dev, ignore_flow_level))
+		return -EOPNOTSUPP;
+
+	bridge = mlx5_esw_bridge_from_port_lookup(vport_num, esw_owner_vhca_id, br_offloads);
+	if (!bridge)
+		return -EINVAL;
+
+	mcast = bridge->flags & MLX5_ESW_BRIDGE_MCAST_FLAG;
+	if (mcast == enable)
+		return 0;
+
+	if (enable)
+		err = mlx5_esw_bridge_mcast_enable(bridge);
+	else
+		mlx5_esw_bridge_mcast_disable(bridge);
+
+	return err;
+}
+
 static int mlx5_esw_bridge_vport_init(u16 vport_num, u16 esw_owner_vhca_id, u16 flags,
 				      struct mlx5_esw_bridge_offloads *br_offloads,
 				      struct mlx5_esw_bridge *bridge)
drivers/net/ethernet/mellanox/mlx5/core/esw/bridge.h +9 -0
index 10851a515bca..b18f137173d9 100644
@@ -25,12 +25,19 @@ struct mlx5_esw_bridge_offloads {
 	struct delayed_work update_work;
 
 	struct mlx5_flow_table *ingress_ft;
+	struct mlx5_flow_group *ingress_igmp_fg;
+	struct mlx5_flow_group *ingress_mld_fg;
 	struct mlx5_flow_group *ingress_vlan_fg;
 	struct mlx5_flow_group *ingress_vlan_filter_fg;
 	struct mlx5_flow_group *ingress_qinq_fg;
 	struct mlx5_flow_group *ingress_qinq_filter_fg;
 	struct mlx5_flow_group *ingress_mac_fg;
 
+	struct mlx5_flow_handle *igmp_handle;
+	struct mlx5_flow_handle *mld_query_handle;
+	struct mlx5_flow_handle *mld_report_handle;
+	struct mlx5_flow_handle *mld_done_handle;
+
 	struct mlx5_flow_table *skip_ft;
 };
 
@@ -64,6 +71,8 @@ int mlx5_esw_bridge_vlan_filtering_set(u16 vport_num, u16 esw_owner_vhca_id, boo
 				       struct mlx5_esw_bridge_offloads *br_offloads);
 int mlx5_esw_bridge_vlan_proto_set(u16 vport_num, u16 esw_owner_vhca_id, u16 proto,
 				   struct mlx5_esw_bridge_offloads *br_offloads);
+int mlx5_esw_bridge_mcast_set(u16 vport_num, u16 esw_owner_vhca_id, bool enable,
+			      struct mlx5_esw_bridge_offloads *br_offloads);
 int mlx5_esw_bridge_port_vlan_add(u16 vport_num, u16 esw_owner_vhca_id, u16 vid, u16 flags,
 				  struct mlx5_esw_bridge_offloads *br_offloads,
 				  struct netlink_ext_ack *extack);
drivers/net/ethernet/mellanox/mlx5/core/esw/bridge_mcast.c new file +316 -0

316 lines not on this page · raw message

drivers/net/ethernet/mellanox/mlx5/core/esw/bridge_priv.h +23 -4
index b99761e73c1b..dbb935db1b3c 100644
@@ -12,11 +12,26 @@
 #include <linux/xarray.h>
 #include "fs_core.h"
 
+#define MLX5_ESW_BRIDGE_INGRESS_TABLE_IGMP_GRP_SIZE 1
+#define MLX5_ESW_BRIDGE_INGRESS_TABLE_MLD_GRP_SIZE 3
 #define MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_GRP_SIZE 131072
-#define MLX5_ESW_BRIDGE_INGRESS_TABLE_UNTAGGED_GRP_SIZE 524288
-#define MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_GRP_IDX_FROM 0
-#define MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_GRP_IDX_TO		\
-	(MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_GRP_SIZE - 1)
+#define MLX5_ESW_BRIDGE_INGRESS_TABLE_UNTAGGED_GRP_SIZE			\
+	(524288 - MLX5_ESW_BRIDGE_INGRESS_TABLE_IGMP_GRP_SIZE -		\
+	 MLX5_ESW_BRIDGE_INGRESS_TABLE_MLD_GRP_SIZE)
+
+#define MLX5_ESW_BRIDGE_INGRESS_TABLE_IGMP_GRP_IDX_FROM 0
+#define MLX5_ESW_BRIDGE_INGRESS_TABLE_IGMP_GRP_IDX_TO		\
+	(MLX5_ESW_BRIDGE_INGRESS_TABLE_IGMP_GRP_SIZE - 1)
+#define MLX5_ESW_BRIDGE_INGRESS_TABLE_MLD_GRP_IDX_FROM	\
+	(MLX5_ESW_BRIDGE_INGRESS_TABLE_IGMP_GRP_IDX_TO + 1)
+#define MLX5_ESW_BRIDGE_INGRESS_TABLE_MLD_GRP_IDX_TO		\
+	(MLX5_ESW_BRIDGE_INGRESS_TABLE_MLD_GRP_IDX_FROM +	\
+	 MLX5_ESW_BRIDGE_INGRESS_TABLE_MLD_GRP_SIZE - 1)
+#define MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_GRP_IDX_FROM			\
+	(MLX5_ESW_BRIDGE_INGRESS_TABLE_MLD_GRP_IDX_TO + 1)
+#define MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_GRP_IDX_TO			\
+	(MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_GRP_IDX_FROM +		\
+	 MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_GRP_SIZE - 1)
 #define MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_FILTER_GRP_IDX_FROM	\
 	(MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_GRP_IDX_TO + 1)
 #define MLX5_ESW_BRIDGE_INGRESS_TABLE_VLAN_FILTER_GRP_IDX_TO		\
@@ -74,6 +89,7 @@ enum {
 
 enum {
 	MLX5_ESW_BRIDGE_VLAN_FILTERING_FLAG = BIT(0),
+	MLX5_ESW_BRIDGE_MCAST_FLAG = BIT(1),
 };
 
 struct mlx5_esw_bridge_fdb_key {
@@ -145,4 +161,7 @@ struct mlx5_esw_bridge {
 	u16 vlan_proto;
 };
 
+int mlx5_esw_bridge_mcast_enable(struct mlx5_esw_bridge *bridge);
+void mlx5_esw_bridge_mcast_disable(struct mlx5_esw_bridge *bridge);
+
 #endif /* _MLX5_ESW_BRIDGE_PRIVATE_ */
Signature · 1 line
2.39.2
#7 lore raw

PATCH 6/15 net/mlx5: Bridge, add per-port multicast replication tables

In reply to Saeed Mahameed

From: Vlad Buslov <[email protected]>

Multicast replication requires adding one more level of FDB_BR_OFFLOAD priority flow tables. The new level is used for per-port multicast-specific tables that have following flow groups structure (flow highest to lowest priority):

- Flow group of size one that matches on source port metadata. This will have a static single rule that prevent packets from being replicated to their source port.

- Flow group of size one that matches all packets and forwards them to the port that owns the table.

Initialize the table dynamically on all bridge ports when adding a port to the bridge that has multicast enabled and on all existing bridge ports when receiving multicast enable notification.

---
 .../ethernet/mellanox/mlx5/core/esw/bridge.c  |  14 +-
 .../mellanox/mlx5/core/esw/bridge_mcast.c     | 329 +++++++++++++++++-
 .../mellanox/mlx5/core/esw/bridge_priv.h      |  28 ++
 .../net/ethernet/mellanox/mlx5/core/fs_core.c |   2 +-
 4 files changed, 370 insertions(+), 3 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/esw/bridge.c +13 -1
index 35436aa9548d..4bc8c6fc394b 100644
@@ -61,7 +61,7 @@ mlx5_esw_bridge_pkt_reformat_vlan_pop_create(struct mlx5_eswitch *esw)
 	return mlx5_packet_reformat_alloc(esw->dev, &reformat_params, MLX5_FLOW_NAMESPACE_FDB);
 }
 
-static struct mlx5_flow_table *
+struct mlx5_flow_table *
 mlx5_esw_bridge_table_create(int max_fte, u32 level, struct mlx5_eswitch *esw)
 {
 	struct mlx5_flow_table_attr ft_attr = {};
@@ -1506,6 +1506,15 @@ static int mlx5_esw_bridge_vport_init(u16 vport_num, u16 esw_owner_vhca_id, u16
 	port->bridge = bridge;
 	port->flags |= flags;
 	xa_init(&port->vlans);
+
+	err = mlx5_esw_bridge_port_mcast_init(port);
+	if (err) {
+		esw_warn(esw->dev,
+			 "Failed to initialize port multicast (vport=%u,esw_owner_vhca_id=%u,err=%d)\n",
+			 port->vport_num, port->esw_owner_vhca_id, err);
+		goto err_port_mcast;
+	}
+
 	err = mlx5_esw_bridge_port_insert(port, br_offloads);
 	if (err) {
 		esw_warn(esw->dev,
@@ -1518,6 +1527,8 @@ static int mlx5_esw_bridge_vport_init(u16 vport_num, u16 esw_owner_vhca_id, u16
 	return 0;
 
 err_port_insert:
+	mlx5_esw_bridge_port_mcast_cleanup(port);
+err_port_mcast:
 	kvfree(port);
 	return err;
 }
@@ -1535,6 +1546,7 @@ static int mlx5_esw_bridge_vport_cleanup(struct mlx5_esw_bridge_offloads *br_off
 
 	trace_mlx5_esw_bridge_vport_cleanup(port);
 	mlx5_esw_bridge_port_vlans_flush(port, bridge);
+	mlx5_esw_bridge_port_mcast_cleanup(port);
 	mlx5_esw_bridge_port_erase(port, br_offloads);
 	kvfree(port);
 	mlx5_esw_bridge_put(br_offloads, bridge);
drivers/net/ethernet/mellanox/mlx5/core/esw/bridge_mcast.c +328 -1

355 lines not on this page · raw message

drivers/net/ethernet/mellanox/mlx5/core/esw/bridge_priv.h +28 -0
index dbb935db1b3c..7fdd719f363c 100644
@@ -81,9 +81,23 @@ static_assert(MLX5_ESW_BRIDGE_EGRESS_TABLE_SIZE == 524288);
 
 #define MLX5_ESW_BRIDGE_SKIP_TABLE_SIZE 0
 
+#define MLX5_ESW_BRIDGE_MCAST_TABLE_FILTER_GRP_SIZE 1
+#define MLX5_ESW_BRIDGE_MCAST_TABLE_FWD_GRP_SIZE 1
+#define MLX5_ESW_BRIDGE_MCAST_TABLE_FILTER_GRP_IDX_FROM 0
+#define MLX5_ESW_BRIDGE_MCAST_TABLE_FILTER_GRP_IDX_TO		\
+	(MLX5_ESW_BRIDGE_MCAST_TABLE_FILTER_GRP_SIZE - 1)
+#define MLX5_ESW_BRIDGE_MCAST_TABLE_FWD_GRP_IDX_FROM		\
+	(MLX5_ESW_BRIDGE_MCAST_TABLE_FILTER_GRP_IDX_TO + 1)
+#define MLX5_ESW_BRIDGE_MCAST_TABLE_FWD_GRP_IDX_TO			\
+	(MLX5_ESW_BRIDGE_MCAST_TABLE_FWD_GRP_IDX_FROM +			\
+	 MLX5_ESW_BRIDGE_MCAST_TABLE_FWD_GRP_SIZE - 1)
+
+#define MLX5_ESW_BRIDGE_MCAST_TABLE_SIZE			\
+	(MLX5_ESW_BRIDGE_MCAST_TABLE_FWD_GRP_IDX_TO + 1)
 enum {
 	MLX5_ESW_BRIDGE_LEVEL_INGRESS_TABLE,
 	MLX5_ESW_BRIDGE_LEVEL_EGRESS_TABLE,
+	MLX5_ESW_BRIDGE_LEVEL_MCAST_TABLE,
 	MLX5_ESW_BRIDGE_LEVEL_SKIP_TABLE,
 };
 
@@ -138,6 +152,14 @@ struct mlx5_esw_bridge_port {
 	u16 flags;
 	struct mlx5_esw_bridge *bridge;
 	struct xarray vlans;
+	struct {
+		struct mlx5_flow_table *ft;
+		struct mlx5_flow_group *filter_fg;
+		struct mlx5_flow_group *fwd_fg;
+
+		struct mlx5_flow_handle *filter_handle;
+		struct mlx5_flow_handle *fwd_handle;
+	} mcast;
 };
 
 struct mlx5_esw_bridge {
@@ -161,6 +183,12 @@ struct mlx5_esw_bridge {
 	u16 vlan_proto;
 };
 
+struct mlx5_flow_table *mlx5_esw_bridge_table_create(int max_fte, u32 level,
+						     struct mlx5_eswitch *esw);
+
+int mlx5_esw_bridge_port_mcast_init(struct mlx5_esw_bridge_port *port);
+void mlx5_esw_bridge_port_mcast_cleanup(struct mlx5_esw_bridge_port *port);
+
 int mlx5_esw_bridge_mcast_enable(struct mlx5_esw_bridge *bridge);
 void mlx5_esw_bridge_mcast_disable(struct mlx5_esw_bridge *bridge);
 
drivers/net/ethernet/mellanox/mlx5/core/fs_core.c +1 -1
index 8e3da9d4fe1c..19da02c41616 100644
@@ -2997,7 +2997,7 @@ static int init_fdb_root_ns(struct mlx5_flow_steering *steering)
 		goto out_err;
 	}
 
-	maj_prio = fs_create_prio(&steering->fdb_root_ns->ns, FDB_BR_OFFLOAD, 3);
+	maj_prio = fs_create_prio(&steering->fdb_root_ns->ns, FDB_BR_OFFLOAD, 4);
 	if (IS_ERR(maj_prio)) {
 		err = PTR_ERR(maj_prio);
 		goto out_err;
Signature · 1 line
2.39.2
#8 lore raw

PATCH 7/15 net/mlx5: Bridge, support multicast VLAN pop

In reply to Saeed Mahameed

From: Vlad Buslov <[email protected]>

When VLAN with 'untagged' flag is created on port also provision the per-port multicast table rule to pop the VLAN during packet replication. This functionality must be in per-port table because some subset of ports that are member of multicast group can require just a match on VLAN (trunk mode) while other subset can be configured to remove the VLAN tag from packets received on the ports (access mode).

---
 .../ethernet/mellanox/mlx5/core/esw/bridge.c  |  33 ++-
 .../mellanox/mlx5/core/esw/bridge_mcast.c     | 190 +++++++++++++++++-
 .../mellanox/mlx5/core/esw/bridge_priv.h      |  22 +-
 3 files changed, 236 insertions(+), 9 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/esw/bridge.c +28 -5
index 4bc8c6fc394b..52c976135397 100644
@@ -1095,8 +1095,21 @@ mlx5_esw_bridge_vlan_push_mark_cleanup(struct mlx5_esw_bridge_vlan *vlan, struct
 }
 
 static int
-mlx5_esw_bridge_vlan_push_pop_create(u16 vlan_proto, u16 flags, struct mlx5_esw_bridge_vlan *vlan,
-				     struct mlx5_eswitch *esw)
+mlx5_esw_bridge_vlan_push_pop_fhs_create(u16 vlan_proto, struct mlx5_esw_bridge_port *port,
+					 struct mlx5_esw_bridge_vlan *vlan)
+{
+	return mlx5_esw_bridge_vlan_mcast_init(vlan_proto, port, vlan);
+}
+
+static void
+mlx5_esw_bridge_vlan_push_pop_fhs_cleanup(struct mlx5_esw_bridge_vlan *vlan)
+{
+	mlx5_esw_bridge_vlan_mcast_cleanup(vlan);
+}
+
+static int
+mlx5_esw_bridge_vlan_push_pop_create(u16 vlan_proto, u16 flags, struct mlx5_esw_bridge_port *port,
+				     struct mlx5_esw_bridge_vlan *vlan, struct mlx5_eswitch *esw)
 {
 	int err;
 
@@ -1114,10 +1127,16 @@ mlx5_esw_bridge_vlan_push_pop_create(u16 vlan_proto, u16 flags, struct mlx5_esw_
 		err = mlx5_esw_bridge_vlan_pop_create(vlan, esw);
 		if (err)
 			goto err_vlan_pop;
+
+		err = mlx5_esw_bridge_vlan_push_pop_fhs_create(vlan_proto, port, vlan);
+		if (err)
+			goto err_vlan_pop_fhs;
 	}
 
 	return 0;
 
+err_vlan_pop_fhs:
+	mlx5_esw_bridge_vlan_pop_cleanup(vlan, esw);
 err_vlan_pop:
 	if (vlan->pkt_mod_hdr_push_mark)
 		mlx5_esw_bridge_vlan_push_mark_cleanup(vlan, esw);
@@ -1142,7 +1161,7 @@ mlx5_esw_bridge_vlan_create(u16 vlan_proto, u16 vid, u16 flags, struct mlx5_esw_
 	vlan->flags = flags;
 	INIT_LIST_HEAD(&vlan->fdb_list);
 
-	err = mlx5_esw_bridge_vlan_push_pop_create(vlan_proto, flags, vlan, esw);
+	err = mlx5_esw_bridge_vlan_push_pop_create(vlan_proto, flags, port, vlan, esw);
 	if (err)
 		goto err_vlan_push_pop;
 
@@ -1154,6 +1173,8 @@ mlx5_esw_bridge_vlan_create(u16 vlan_proto, u16 vid, u16 flags, struct mlx5_esw_
 	return vlan;
 
 err_xa_insert:
+	if (vlan->mcast_handle)
+		mlx5_esw_bridge_vlan_push_pop_fhs_cleanup(vlan);
 	if (vlan->pkt_reformat_pop)
 		mlx5_esw_bridge_vlan_pop_cleanup(vlan, esw);
 	if (vlan->pkt_mod_hdr_push_mark)
@@ -1180,6 +1201,8 @@ static void mlx5_esw_bridge_vlan_flush(struct mlx5_esw_bridge_vlan *vlan,
 	list_for_each_entry_safe(entry, tmp, &vlan->fdb_list, vlan_list)
 		mlx5_esw_bridge_fdb_entry_notify_and_cleanup(entry, bridge);
 
+	if (vlan->mcast_handle)
+		mlx5_esw_bridge_vlan_push_pop_fhs_cleanup(vlan);
 	if (vlan->pkt_reformat_pop)
 		mlx5_esw_bridge_vlan_pop_cleanup(vlan, esw);
 	if (vlan->pkt_mod_hdr_push_mark)
@@ -1218,8 +1241,8 @@ static int mlx5_esw_bridge_port_vlans_recreate(struct mlx5_esw_bridge_port *port
 
 	xa_for_each(&port->vlans, i, vlan) {
 		mlx5_esw_bridge_vlan_flush(vlan, bridge);
-		err = mlx5_esw_bridge_vlan_push_pop_create(bridge->vlan_proto, vlan->flags, vlan,
-							   br_offloads->esw);
+		err = mlx5_esw_bridge_vlan_push_pop_create(bridge->vlan_proto, vlan->flags, port,
+							   vlan, br_offloads->esw);
 		if (err) {
 			esw_warn(br_offloads->esw->dev,
 				 "Failed to create VLAN=%u(proto=%x) push/pop actions (vport=%u,err=%d)\n",
drivers/net/ethernet/mellanox/mlx5/core/esw/bridge_mcast.c +187 -3

255 lines not on this page · raw message

drivers/net/ethernet/mellanox/mlx5/core/esw/bridge_priv.h +21 -1
index 7fdd719f363c..36ff32001ce8 100644
@@ -83,17 +83,31 @@ static_assert(MLX5_ESW_BRIDGE_EGRESS_TABLE_SIZE == 524288);
 
 #define MLX5_ESW_BRIDGE_MCAST_TABLE_FILTER_GRP_SIZE 1
 #define MLX5_ESW_BRIDGE_MCAST_TABLE_FWD_GRP_SIZE 1
+#define MLX5_ESW_BRIDGE_MCAST_TABLE_VLAN_GRP_SIZE 4095
+#define MLX5_ESW_BRIDGE_MCAST_TABLE_QINQ_GRP_SIZE MLX5_ESW_BRIDGE_MCAST_TABLE_VLAN_GRP_SIZE
 #define MLX5_ESW_BRIDGE_MCAST_TABLE_FILTER_GRP_IDX_FROM 0
 #define MLX5_ESW_BRIDGE_MCAST_TABLE_FILTER_GRP_IDX_TO		\
 	(MLX5_ESW_BRIDGE_MCAST_TABLE_FILTER_GRP_SIZE - 1)
-#define MLX5_ESW_BRIDGE_MCAST_TABLE_FWD_GRP_IDX_FROM		\
+#define MLX5_ESW_BRIDGE_MCAST_TABLE_VLAN_GRP_IDX_FROM		\
 	(MLX5_ESW_BRIDGE_MCAST_TABLE_FILTER_GRP_IDX_TO + 1)
+#define MLX5_ESW_BRIDGE_MCAST_TABLE_VLAN_GRP_IDX_TO			\
+	(MLX5_ESW_BRIDGE_MCAST_TABLE_VLAN_GRP_IDX_FROM +		\
+	 MLX5_ESW_BRIDGE_MCAST_TABLE_VLAN_GRP_SIZE - 1)
+#define MLX5_ESW_BRIDGE_MCAST_TABLE_QINQ_GRP_IDX_FROM		\
+	(MLX5_ESW_BRIDGE_MCAST_TABLE_VLAN_GRP_IDX_TO + 1)
+#define MLX5_ESW_BRIDGE_MCAST_TABLE_QINQ_GRP_IDX_TO			\
+	(MLX5_ESW_BRIDGE_MCAST_TABLE_QINQ_GRP_IDX_FROM +		\
+	 MLX5_ESW_BRIDGE_MCAST_TABLE_QINQ_GRP_SIZE - 1)
+#define MLX5_ESW_BRIDGE_MCAST_TABLE_FWD_GRP_IDX_FROM		\
+	(MLX5_ESW_BRIDGE_MCAST_TABLE_QINQ_GRP_IDX_TO + 1)
 #define MLX5_ESW_BRIDGE_MCAST_TABLE_FWD_GRP_IDX_TO			\
 	(MLX5_ESW_BRIDGE_MCAST_TABLE_FWD_GRP_IDX_FROM +			\
 	 MLX5_ESW_BRIDGE_MCAST_TABLE_FWD_GRP_SIZE - 1)
 
 #define MLX5_ESW_BRIDGE_MCAST_TABLE_SIZE			\
 	(MLX5_ESW_BRIDGE_MCAST_TABLE_FWD_GRP_IDX_TO + 1)
+static_assert(MLX5_ESW_BRIDGE_MCAST_TABLE_SIZE == 8192);
+
 enum {
 	MLX5_ESW_BRIDGE_LEVEL_INGRESS_TABLE,
 	MLX5_ESW_BRIDGE_LEVEL_EGRESS_TABLE,
@@ -144,6 +158,7 @@ struct mlx5_esw_bridge_vlan {
 	struct mlx5_pkt_reformat *pkt_reformat_push;
 	struct mlx5_pkt_reformat *pkt_reformat_pop;
 	struct mlx5_modify_hdr *pkt_mod_hdr_push_mark;
+	struct mlx5_flow_handle *mcast_handle;
 };
 
 struct mlx5_esw_bridge_port {
@@ -155,6 +170,8 @@ struct mlx5_esw_bridge_port {
 	struct {
 		struct mlx5_flow_table *ft;
 		struct mlx5_flow_group *filter_fg;
+		struct mlx5_flow_group *vlan_fg;
+		struct mlx5_flow_group *qinq_fg;
 		struct mlx5_flow_group *fwd_fg;
 
 		struct mlx5_flow_handle *filter_handle;
@@ -188,6 +205,9 @@ struct mlx5_flow_table *mlx5_esw_bridge_table_create(int max_fte, u32 level,
 
 int mlx5_esw_bridge_port_mcast_init(struct mlx5_esw_bridge_port *port);
 void mlx5_esw_bridge_port_mcast_cleanup(struct mlx5_esw_bridge_port *port);
+int mlx5_esw_bridge_vlan_mcast_init(u16 vlan_proto, struct mlx5_esw_bridge_port *port,
+				    struct mlx5_esw_bridge_vlan *vlan);
+void mlx5_esw_bridge_vlan_mcast_cleanup(struct mlx5_esw_bridge_vlan *vlan);
 
 int mlx5_esw_bridge_mcast_enable(struct mlx5_esw_bridge *bridge);
 void mlx5_esw_bridge_mcast_disable(struct mlx5_esw_bridge *bridge);
Signature · 1 line
2.39.2
#10 lore raw

PATCH 9/15 net/mlx5: Bridge, add tracepoints for multicast

In reply to Saeed Mahameed

From: Vlad Buslov <[email protected]>

Pass target struct net_device to mdb attach/detach handler in order to expose the port name to the new tracepoints. Implemented following tracepoints:

- Attach mdb to port.
- Detach mdb from port.

Usage example:

Quoted 3 lines
# cd /sys/kernel/debug/tracing
# echo mlx5:mlx5_esw_bridge_port_mdb_attach >> set_event
# cat trace

...

     kworker/0:0-19071   [000] ..... 259004.253848: mlx5_esw_bridge_port_mdb_attach: net_device=enp8s0f0_0 addr=33:33:ff:00:00:01 vid=0 num_ports=1 offloaded=1
---
 .../mellanox/mlx5/core/en/rep/bridge.c        |  4 +--
 .../ethernet/mellanox/mlx5/core/esw/bridge.c  | 14 ++++----
 .../ethernet/mellanox/mlx5/core/esw/bridge.h  | 10 +++---
 .../mellanox/mlx5/core/esw/bridge_mcast.c     | 12 ++++---
 .../mellanox/mlx5/core/esw/bridge_priv.h      |  8 ++---
 .../mlx5/core/esw/diag/bridge_tracepoint.h    | 35 +++++++++++++++++++
 6 files changed, 63 insertions(+), 20 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en/rep/bridge.c +2 -2
index dd0dd3f028a3..fd191925ab4b 100644
@@ -238,7 +238,7 @@ mlx5_esw_bridge_port_obj_add(struct net_device *dev,
 		break;
 	case SWITCHDEV_OBJ_ID_PORT_MDB:
 		mdb = SWITCHDEV_OBJ_PORT_MDB(obj);
-		err = mlx5_esw_bridge_port_mdb_add(vport_num, esw_owner_vhca_id, mdb->addr,
+		err = mlx5_esw_bridge_port_mdb_add(dev, vport_num, esw_owner_vhca_id, mdb->addr,
 						   mdb->vid, br_offloads, extack);
 		break;
 	default:
@@ -270,7 +270,7 @@ mlx5_esw_bridge_port_obj_del(struct net_device *dev,
 		break;
 	case SWITCHDEV_OBJ_ID_PORT_MDB:
 		mdb = SWITCHDEV_OBJ_PORT_MDB(obj);
-		mlx5_esw_bridge_port_mdb_del(vport_num, esw_owner_vhca_id, mdb->addr, mdb->vid,
+		mlx5_esw_bridge_port_mdb_del(dev, vport_num, esw_owner_vhca_id, mdb->addr, mdb->vid,
 					     br_offloads);
 		break;
 	default:
drivers/net/ethernet/mellanox/mlx5/core/esw/bridge.c +8 -6
index be4787539c6c..1ba03e219111 100644
@@ -1803,8 +1803,9 @@ void mlx5_esw_bridge_update(struct mlx5_esw_bridge_offloads *br_offloads)
 	}
 }
 
-int mlx5_esw_bridge_port_mdb_add(u16 vport_num, u16 esw_owner_vhca_id, const unsigned char *addr,
-				 u16 vid, struct mlx5_esw_bridge_offloads *br_offloads,
+int mlx5_esw_bridge_port_mdb_add(struct net_device *dev, u16 vport_num, u16 esw_owner_vhca_id,
+				 const unsigned char *addr, u16 vid,
+				 struct mlx5_esw_bridge_offloads *br_offloads,
 				 struct netlink_ext_ack *extack)
 {
 	struct mlx5_esw_bridge_vlan *vlan;
@@ -1837,7 +1838,7 @@ int mlx5_esw_bridge_port_mdb_add(u16 vport_num, u16 esw_owner_vhca_id, const uns
 		}
 	}
 
-	err = mlx5_esw_bridge_port_mdb_attach(port, addr, vid);
+	err = mlx5_esw_bridge_port_mdb_attach(dev, port, addr, vid);
 	if (err) {
 		NL_SET_ERR_MSG_FMT_MOD(extack, "Failed to add MDB (MAC=%pM,vid=%u,vport=%u)\n",
 				       addr, vid, vport_num);
@@ -1847,8 +1848,9 @@ int mlx5_esw_bridge_port_mdb_add(u16 vport_num, u16 esw_owner_vhca_id, const uns
 	return 0;
 }
 
-void mlx5_esw_bridge_port_mdb_del(u16 vport_num, u16 esw_owner_vhca_id, const unsigned char *addr,
-				  u16 vid, struct mlx5_esw_bridge_offloads *br_offloads)
+void mlx5_esw_bridge_port_mdb_del(struct net_device *dev, u16 vport_num, u16 esw_owner_vhca_id,
+				  const unsigned char *addr, u16 vid,
+				  struct mlx5_esw_bridge_offloads *br_offloads)
 {
 	struct mlx5_esw_bridge_port *port;
 
@@ -1856,7 +1858,7 @@ void mlx5_esw_bridge_port_mdb_del(u16 vport_num, u16 esw_owner_vhca_id, const un
 	if (!port)
 		return;
 
-	mlx5_esw_bridge_port_mdb_detach(port, addr, vid);
+	mlx5_esw_bridge_port_mdb_detach(dev, port, addr, vid);
 }
 
 static void mlx5_esw_bridge_flush(struct mlx5_esw_bridge_offloads *br_offloads)
drivers/net/ethernet/mellanox/mlx5/core/esw/bridge.h +6 -4
index 9cab66467289..a9dd18c73d6a 100644
@@ -79,10 +79,12 @@ int mlx5_esw_bridge_port_vlan_add(u16 vport_num, u16 esw_owner_vhca_id, u16 vid,
 void mlx5_esw_bridge_port_vlan_del(u16 vport_num, u16 esw_owner_vhca_id, u16 vid,
 				   struct mlx5_esw_bridge_offloads *br_offloads);
 
-int mlx5_esw_bridge_port_mdb_add(u16 vport_num, u16 esw_owner_vhca_id, const unsigned char *addr,
-				 u16 vid, struct mlx5_esw_bridge_offloads *br_offloads,
+int mlx5_esw_bridge_port_mdb_add(struct net_device *dev, u16 vport_num, u16 esw_owner_vhca_id,
+				 const unsigned char *addr, u16 vid,
+				 struct mlx5_esw_bridge_offloads *br_offloads,
 				 struct netlink_ext_ack *extack);
-void mlx5_esw_bridge_port_mdb_del(u16 vport_num, u16 esw_owner_vhca_id, const unsigned char *addr,
-				  u16 vid, struct mlx5_esw_bridge_offloads *br_offloads);
+void mlx5_esw_bridge_port_mdb_del(struct net_device *dev, u16 vport_num, u16 esw_owner_vhca_id,
+				  const unsigned char *addr, u16 vid,
+				  struct mlx5_esw_bridge_offloads *br_offloads);
 
 #endif /* __MLX5_ESW_BRIDGE_H__ */
drivers/net/ethernet/mellanox/mlx5/core/esw/bridge_mcast.c +8 -4
index d17fe6d374b5..2eae594a5e80 100644
@@ -5,6 +5,7 @@
 #include "bridge.h"
 #include "eswitch.h"
 #include "bridge_priv.h"
+#include "diag/bridge_tracepoint.h"
 
 static const struct rhashtable_params mdb_ht_params = {
 	.key_offset = offsetof(struct mlx5_esw_bridge_mdb_entry, key),
@@ -180,8 +181,8 @@ static void mlx5_esw_bridge_port_mdb_entry_cleanup(struct mlx5_esw_bridge *bridg
 	kvfree(entry);
 }
 
-int mlx5_esw_bridge_port_mdb_attach(struct mlx5_esw_bridge_port *port, const unsigned char *addr,
-				    u16 vid)
+int mlx5_esw_bridge_port_mdb_attach(struct net_device *dev, struct mlx5_esw_bridge_port *port,
+				    const unsigned char *addr, u16 vid)
 {
 	struct mlx5_esw_bridge *bridge = port->bridge;
 	struct mlx5_esw_bridge_mdb_entry *entry;
@@ -224,6 +225,8 @@ int mlx5_esw_bridge_port_mdb_attach(struct mlx5_esw_bridge_port *port, const uns
 		 */
 		esw_warn(bridge->br_offloads->esw->dev, "MDB attach failed to offload (MAC=%pM,vid=%u,vport=%u,err=%d)\n",
 			 addr, vid, port->vport_num, err);
+
+	trace_mlx5_esw_bridge_port_mdb_attach(dev, entry);
 	return 0;
 }
 
@@ -248,8 +251,8 @@ static void mlx5_esw_bridge_port_mdb_entry_detach(struct mlx5_esw_bridge_port *p
 			 entry->key.addr, entry->key.vid, port->vport_num);
 }
 
-void mlx5_esw_bridge_port_mdb_detach(struct mlx5_esw_bridge_port *port, const unsigned char *addr,
-				     u16 vid)
+void mlx5_esw_bridge_port_mdb_detach(struct net_device *dev, struct mlx5_esw_bridge_port *port,
+				     const unsigned char *addr, u16 vid)
 {
 	struct mlx5_esw_bridge *bridge = port->bridge;
 	struct mlx5_esw_bridge_mdb_entry *entry;
@@ -269,6 +272,7 @@ void mlx5_esw_bridge_port_mdb_detach(struct mlx5_esw_bridge_port *port, const un
 		return;
 	}
 
+	trace_mlx5_esw_bridge_port_mdb_detach(dev, entry);
 	mlx5_esw_bridge_port_mdb_entry_detach(port, entry);
 }
 
drivers/net/ethernet/mellanox/mlx5/core/esw/bridge_priv.h +4 -4
index 849028f94be2..c9595801bdb4 100644
@@ -233,10 +233,10 @@ void mlx5_esw_bridge_mcast_disable(struct mlx5_esw_bridge *bridge);
 
 int mlx5_esw_bridge_mdb_init(struct mlx5_esw_bridge *bridge);
 void mlx5_esw_bridge_mdb_cleanup(struct mlx5_esw_bridge *bridge);
-int mlx5_esw_bridge_port_mdb_attach(struct mlx5_esw_bridge_port *port, const unsigned char *addr,
-				    u16 vid);
-void mlx5_esw_bridge_port_mdb_detach(struct mlx5_esw_bridge_port *port, const unsigned char *addr,
-				     u16 vid);
+int mlx5_esw_bridge_port_mdb_attach(struct net_device *dev, struct mlx5_esw_bridge_port *port,
+				    const unsigned char *addr, u16 vid);
+void mlx5_esw_bridge_port_mdb_detach(struct net_device *dev, struct mlx5_esw_bridge_port *port,
+				     const unsigned char *addr, u16 vid);
 void mlx5_esw_bridge_port_mdb_vlan_flush(struct mlx5_esw_bridge_port *port,
 					 struct mlx5_esw_bridge_vlan *vlan);
 void mlx5_esw_bridge_mdb_flush(struct mlx5_esw_bridge *bridge);
drivers/net/ethernet/mellanox/mlx5/core/esw/diag/bridge_tracepoint.h +35 -0
index 51ac24e6ec3c..1808da214094 100644
@@ -110,6 +110,41 @@ DEFINE_EVENT(mlx5_esw_bridge_port_template,
 	     TP_ARGS(port)
 	);
 
+DECLARE_EVENT_CLASS(mlx5_esw_bridge_mdb_port_change_template,
+		    TP_PROTO(const struct net_device *dev,
+			     const struct mlx5_esw_bridge_mdb_entry *mdb),
+		    TP_ARGS(dev, mdb),
+		    TP_STRUCT__entry(
+			    __array(char, dev_name, IFNAMSIZ)
+			    __array(unsigned char, addr, ETH_ALEN)
+			    __field(u16, vid)
+			    __field(int, num_ports)
+			    __field(bool, offloaded)),
+		    TP_fast_assign(
+			    strscpy(__entry->dev_name, netdev_name(dev), IFNAMSIZ);
+			    memcpy(__entry->addr, mdb->key.addr, ETH_ALEN);
+			    __entry->vid = mdb->key.vid;
+			    __entry->num_ports = mdb->num_ports;
+			    __entry->offloaded = mdb->egress_handle;),
+		    TP_printk("net_device=%s addr=%pM vid=%u num_ports=%d offloaded=%d",
+			      __entry->dev_name,
+			      __entry->addr,
+			      __entry->vid,
+			      __entry->num_ports,
+			      __entry->offloaded));
+
+DEFINE_EVENT(mlx5_esw_bridge_mdb_port_change_template,
+	     mlx5_esw_bridge_port_mdb_attach,
+	     TP_PROTO(const struct net_device *dev,
+		      const struct mlx5_esw_bridge_mdb_entry *mdb),
+	     TP_ARGS(dev, mdb));
+
+DEFINE_EVENT(mlx5_esw_bridge_mdb_port_change_template,
+	     mlx5_esw_bridge_port_mdb_detach,
+	     TP_PROTO(const struct net_device *dev,
+		      const struct mlx5_esw_bridge_mdb_entry *mdb),
+	     TP_ARGS(dev, mdb));
+
 #endif
 
 /* This part must be outside protection */
Signature · 1 line
2.39.2
#9 lore raw

PATCH 8/15 net/mlx5: Bridge, implement mdb offload

In reply to Saeed Mahameed

From: Vlad Buslov <[email protected]>

Implement support for add/del SWITCHDEV_OBJ_ID_PORT_MDB events. For mdb destination addresses configure egress table rules to replicate to per-port multicast tables of all ports that are member of the multicast group as illustrated by 'MDB1' rule in the following diagram:

                                                                                                                            +--------+--+
                                                                                    +---------------------------------------> Port 1 |  |
                                                                                    |                                       +-^------+--+
                                                                                    |                                         |
                                                                                    |                                         |
                                       +-----------------------------------------+  |     +---------------------------+       |
                                       | EGRESS table                            |  |  +--> PORT 1 multicast table    |       |
+----------------------------------+   +-----------------------------------------+  |  |  +---------------------------+       |
| INGRESS table                    |   |                                         |  |  |  |                           |       |
+----------------------------------+   | dst_mac=P1,vlan=X -> pop vlan, goto P1  +--+  |  | FG0:                      |       |
|                                  |   | dst_mac=P1,vlan=Y -> pop vlan, goto P1  |     |  | src_port=dst_port -> drop |       |
| src_mac=M1,vlan=X -> goto egress +---> dst_mac=P2,vlan=X -> pop vlan, goto P2  +--+  |  | FG1:                      |       |
| ...                              |   | dst_mac=P2,vlan=Y -> goto P2            |  |  |  | VLAN X -> pop, goto port  |       |
|                                  |   | dst_mac=MDB1,vlan=Y -> goto mcast P1,P2 +-----+  | ...                       |       |
+----------------------------------+   |                                         |  |  |  | VLAN Y -> pop, goto port  +-------+
                                       +-----------------------------------------+  |  |  | FG3:                      |
                                                                                    |  |  | matchall -> goto port     |
                                                                                    |  |  |                           |
                                                                                    |  |  +---------------------------+
                                                                                    |  |
                                                                                    |  |
                                                                                    |  |                                    +--------+--+
                                                                                    +---------------------------------------> Port 2 |  |
                                                                                       |                                    +-^------+--+
                                                                                       |                                      |
                                                                                       |                                      |
                                                                                       |  +---------------------------+       |
                                                                                       +--> PORT 2 multicast table    |       |
                                                                                          +---------------------------+       |
                                                                                          |                           |       |
                                                                                          | FG0:                      |       |
                                                                                          | src_port=dst_port -> drop |       |
                                                                                          | FG1:                      |       |
                                                                                          | VLAN X -> pop, goto port  |       |
                                                                                          | ...                       |       |
                                                                                          |                           |       |
                                                                                          | FG3:                      |       |
                                                                                          | matchall -> goto port     +-------+
                                                                                          |                           |
                                                                                          +---------------------------+

MDB is managed by extending mlx5 bridge to store an entry in mlx5_esw_bridge->mdb_list linked list (used to iterate over all offloaded MDBs) and mlx5_esw_bridge->mdb_ht hash table (used to lookup existing MDB by MAC+VLAN). Every MDB entry can be attached to arbitrary amount of bridge ports that are stored in mlx5_esw_bridge_mdb_entry->ports xarray in order to allow both efficient lookup of the port and also iteration over all ports that the entry is attached to. Every time MDB is attached/detached to/from a port, the hardware rule is recreated with list of destinations corresponding to all attached ports. When the entry is detached from the last port it is removed from mdb and destroyed which means that the ports xarray also acts as implicit reference counting mechanism.

---
 .../mellanox/mlx5/core/en/rep/bridge.c        |  12 +
 .../ethernet/mellanox/mlx5/core/esw/bridge.c  |  75 ++++-
 .../ethernet/mellanox/mlx5/core/esw/bridge.h  |   6 +
 .../mellanox/mlx5/core/esw/bridge_mcast.c     | 295 ++++++++++++++++++
 .../mellanox/mlx5/core/esw/bridge_priv.h      |  29 ++
 5 files changed, 413 insertions(+), 4 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en/rep/bridge.c +12 -0
index 9701eb3c5f1b..dd0dd3f028a3 100644
@@ -220,6 +220,7 @@ mlx5_esw_bridge_port_obj_add(struct net_device *dev,
 	struct netlink_ext_ack *extack = switchdev_notifier_info_to_extack(&port_obj_info->info);
 	const struct switchdev_obj *obj = port_obj_info->obj;
 	const struct switchdev_obj_port_vlan *vlan;
+	const struct switchdev_obj_port_mdb *mdb;
 	u16 vport_num, esw_owner_vhca_id;
 	int err;
 
@@ -235,6 +236,11 @@ mlx5_esw_bridge_port_obj_add(struct net_device *dev,
 		err = mlx5_esw_bridge_port_vlan_add(vport_num, esw_owner_vhca_id, vlan->vid,
 						    vlan->flags, br_offloads, extack);
 		break;
+	case SWITCHDEV_OBJ_ID_PORT_MDB:
+		mdb = SWITCHDEV_OBJ_PORT_MDB(obj);
+		err = mlx5_esw_bridge_port_mdb_add(vport_num, esw_owner_vhca_id, mdb->addr,
+						   mdb->vid, br_offloads, extack);
+		break;
 	default:
 		return -EOPNOTSUPP;
 	}
@@ -248,6 +254,7 @@ mlx5_esw_bridge_port_obj_del(struct net_device *dev,
 {
 	const struct switchdev_obj *obj = port_obj_info->obj;
 	const struct switchdev_obj_port_vlan *vlan;
+	const struct switchdev_obj_port_mdb *mdb;
 	u16 vport_num, esw_owner_vhca_id;
 
 	if (!mlx5_esw_bridge_rep_vport_num_vhca_id_get(dev, br_offloads->esw, &vport_num,
@@ -261,6 +268,11 @@ mlx5_esw_bridge_port_obj_del(struct net_device *dev,
 		vlan = SWITCHDEV_OBJ_PORT_VLAN(obj);
 		mlx5_esw_bridge_port_vlan_del(vport_num, esw_owner_vhca_id, vlan->vid, br_offloads);
 		break;
+	case SWITCHDEV_OBJ_ID_PORT_MDB:
+		mdb = SWITCHDEV_OBJ_PORT_MDB(obj);
+		mlx5_esw_bridge_port_mdb_del(vport_num, esw_owner_vhca_id, mdb->addr, mdb->vid,
+					     br_offloads);
+		break;
 	default:
 		return -EOPNOTSUPP;
 	}
drivers/net/ethernet/mellanox/mlx5/core/esw/bridge.c +71 -4

141 lines not on this page · raw message

drivers/net/ethernet/mellanox/mlx5/core/esw/bridge.h +6 -0
index b18f137173d9..9cab66467289 100644
@@ -79,4 +79,10 @@ int mlx5_esw_bridge_port_vlan_add(u16 vport_num, u16 esw_owner_vhca_id, u16 vid,
 void mlx5_esw_bridge_port_vlan_del(u16 vport_num, u16 esw_owner_vhca_id, u16 vid,
 				   struct mlx5_esw_bridge_offloads *br_offloads);
 
+int mlx5_esw_bridge_port_mdb_add(u16 vport_num, u16 esw_owner_vhca_id, const unsigned char *addr,
+				 u16 vid, struct mlx5_esw_bridge_offloads *br_offloads,
+				 struct netlink_ext_ack *extack);
+void mlx5_esw_bridge_port_mdb_del(u16 vport_num, u16 esw_owner_vhca_id, const unsigned char *addr,
+				  u16 vid, struct mlx5_esw_bridge_offloads *br_offloads);
+
 #endif /* __MLX5_ESW_BRIDGE_H__ */
drivers/net/ethernet/mellanox/mlx5/core/esw/bridge_mcast.c +295 -0

307 lines not on this page · raw message

drivers/net/ethernet/mellanox/mlx5/core/esw/bridge_priv.h +29 -0
index 36ff32001ce8..849028f94be2 100644
@@ -125,6 +125,11 @@ struct mlx5_esw_bridge_fdb_key {
 	u16 vid;
 };
 
+struct mlx5_esw_bridge_mdb_key {
+	unsigned char addr[ETH_ALEN];
+	u16 vid;
+};
+
 enum {
 	MLX5_ESW_BRIDGE_FLAG_ADDED_BY_USER = BIT(0),
 	MLX5_ESW_BRIDGE_FLAG_PEER = BIT(1),
@@ -151,6 +156,16 @@ struct mlx5_esw_bridge_fdb_entry {
 	struct mlx5_flow_handle *filter_handle;
 };
 
+struct mlx5_esw_bridge_mdb_entry {
+	struct mlx5_esw_bridge_mdb_key key;
+	struct rhash_head ht_node;
+	struct list_head list;
+	struct xarray ports;
+	int num_ports;
+
+	struct mlx5_flow_handle *egress_handle;
+};
+
 struct mlx5_esw_bridge_vlan {
 	u16 vid;
 	u16 flags;
@@ -188,6 +203,9 @@ struct mlx5_esw_bridge {
 	struct list_head fdb_list;
 	struct rhashtable fdb_ht;
 
+	struct list_head mdb_list;
+	struct rhashtable mdb_ht;
+
 	struct mlx5_flow_table *egress_ft;
 	struct mlx5_flow_group *egress_vlan_fg;
 	struct mlx5_flow_group *egress_qinq_fg;
@@ -202,6 +220,7 @@ struct mlx5_esw_bridge {
 
 struct mlx5_flow_table *mlx5_esw_bridge_table_create(int max_fte, u32 level,
 						     struct mlx5_eswitch *esw);
+unsigned long mlx5_esw_bridge_port_key(struct mlx5_esw_bridge_port *port);
 
 int mlx5_esw_bridge_port_mcast_init(struct mlx5_esw_bridge_port *port);
 void mlx5_esw_bridge_port_mcast_cleanup(struct mlx5_esw_bridge_port *port);
@@ -212,4 +231,14 @@ void mlx5_esw_bridge_vlan_mcast_cleanup(struct mlx5_esw_bridge_vlan *vlan);
 int mlx5_esw_bridge_mcast_enable(struct mlx5_esw_bridge *bridge);
 void mlx5_esw_bridge_mcast_disable(struct mlx5_esw_bridge *bridge);
 
+int mlx5_esw_bridge_mdb_init(struct mlx5_esw_bridge *bridge);
+void mlx5_esw_bridge_mdb_cleanup(struct mlx5_esw_bridge *bridge);
+int mlx5_esw_bridge_port_mdb_attach(struct mlx5_esw_bridge_port *port, const unsigned char *addr,
+				    u16 vid);
+void mlx5_esw_bridge_port_mdb_detach(struct mlx5_esw_bridge_port *port, const unsigned char *addr,
+				     u16 vid);
+void mlx5_esw_bridge_port_mdb_vlan_flush(struct mlx5_esw_bridge_port *port,
+					 struct mlx5_esw_bridge_vlan *vlan);
+void mlx5_esw_bridge_mdb_flush(struct mlx5_esw_bridge *bridge);
+
 #endif /* _MLX5_ESW_BRIDGE_PRIVATE_ */
Signature · 1 line
2.39.2
#12 lore raw

PATCH 11/15 net/mlx5: DR, Set counter ID on the last STE for STEv1 TX

In reply to Saeed Mahameed

From: Yevgeny Kliteynik <[email protected]>

In STEv1 counter action can be set either by filling counter ID on STE, in which case it is executed before other actions on this STE, or as a single action, in which case it is executed in accordance with the actions order. FW steering on STEv1 devices implements counter as counter ID on STE, and this counter is set on the last STE.
Fix SMFS to be consistent with this behaviour - move TX counter to the last STE, this way the counter will include all actions of the previous STEs that might have changed packet headers length, e.g. encap, vlan push, etc.

---
 .../net/ethernet/mellanox/mlx5/core/steering/dr_ste_v1.c   | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste_v1.c +4 -3
index 084145f18084..27cc6931bbde 100644
@@ -604,9 +604,6 @@ void dr_ste_v1_set_actions_tx(struct mlx5dr_domain *dmn,
 			allow_modify_hdr = false;
 	}
 
-	if (action_type_set[DR_ACTION_TYP_CTR])
-		dr_ste_v1_set_counter_id(last_ste, attr->ctr_id);
-
 	if (action_type_set[DR_ACTION_TYP_MODIFY_HDR]) {
 		if (!allow_modify_hdr || action_sz < DR_STE_ACTION_DOUBLE_SZ) {
 			dr_ste_v1_arr_init_next_match(&last_ste, added_stes,
@@ -724,6 +721,10 @@ void dr_ste_v1_set_actions_tx(struct mlx5dr_domain *dmn,
 						  attr->range.max);
 	}
 
+	/* set counter ID on the last STE to adhere to DMFS behavior */
+	if (action_type_set[DR_ACTION_TYP_CTR])
+		dr_ste_v1_set_counter_id(last_ste, attr->ctr_id);
+
 	dr_ste_v1_set_hit_gvmi(last_ste, attr->hit_gvmi);
 	dr_ste_v1_set_hit_addr(last_ste, attr->final_icm_addr, 1);
 }
Signature · 1 line
2.39.2
#13 lore raw

PATCH 12/15 net/mlx5: Add mlx5_ifc bits for modify header argument

In reply to Saeed Mahameed

From: Yevgeny Kliteynik <[email protected]>

Add enum value for modify-header argument object and mlx5_bits for the related capabilities.

---
 include/linux/mlx5/mlx5_ifc.h | 28 +++++++++++++++++++++++++++-
 1 file changed, 27 insertions(+), 1 deletion(-)
include/linux/mlx5/mlx5_ifc.h +27 -1
index 02c628f4fe26..6c84bf6eec85 100644
@@ -78,12 +78,15 @@ enum {
 
 enum {
 	MLX5_OBJ_TYPE_SW_ICM = 0x0008,
+	MLX5_OBJ_TYPE_HEADER_MODIFY_ARGUMENT  = 0x23,
 };
 
 enum {
 	MLX5_GENERAL_OBJ_TYPES_CAP_SW_ICM = (1ULL << MLX5_OBJ_TYPE_SW_ICM),
 	MLX5_GENERAL_OBJ_TYPES_CAP_GENEVE_TLV_OPT = (1ULL << 11),
 	MLX5_GENERAL_OBJ_TYPES_CAP_VIRTIO_NET_Q = (1ULL << 13),
+	MLX5_GENERAL_OBJ_TYPES_CAP_HEADER_MODIFY_ARGUMENT =
+		(1ULL << MLX5_OBJ_TYPE_HEADER_MODIFY_ARGUMENT),
 	MLX5_GENERAL_OBJ_TYPES_CAP_MACSEC_OFFLOAD = (1ULL << 39),
 };
 
@@ -321,6 +324,10 @@ enum {
 	MLX5_FT_NIC_TX_RDMA_2_NIC_TX = BIT(1),
 };
 
+enum {
+	MLX5_CMD_OP_MOD_UPDATE_HEADER_MODIFY_ARGUMENT = 0x1,
+};
+
 struct mlx5_ifc_flow_table_fields_supported_bits {
 	u8         outer_dmac[0x1];
 	u8         outer_smac[0x1];
@@ -1927,7 +1934,14 @@ struct mlx5_ifc_cmd_hca_cap_bits {
 	u8	   reserved_at_750[0x4];
 	u8	   max_dynamic_vf_msix_table_size[0xc];
 
-	u8	   reserved_at_760[0x20];
+	u8         reserved_at_760[0x3];
+	u8         log_max_num_header_modify_argument[0x5];
+	u8         reserved_at_768[0x4];
+	u8         log_header_modify_argument_granularity[0x4];
+	u8         reserved_at_770[0x3];
+	u8         log_header_modify_argument_max_alloc[0x5];
+	u8         reserved_at_778[0x8];
+
 	u8	   vhca_tunnel_commands[0x40];
 	u8         match_definer_format_supported[0x40];
 };
@@ -6361,6 +6375,18 @@ struct mlx5_ifc_general_obj_out_cmd_hdr_bits {
 	u8         reserved_at_60[0x20];
 };
 
+struct mlx5_ifc_modify_header_arg_bits {
+	u8         reserved_at_0[0x80];
+
+	u8         reserved_at_80[0x8];
+	u8         access_pd[0x18];
+};
+
+struct mlx5_ifc_create_modify_header_arg_in_bits {
+	struct mlx5_ifc_general_obj_in_cmd_hdr_bits hdr;
+	struct mlx5_ifc_modify_header_arg_bits arg;
+};
+
 struct mlx5_ifc_create_match_definer_in_bits {
 	struct mlx5_ifc_general_obj_in_cmd_hdr_bits general_obj_in_cmd_hdr;
 
Signature · 1 line
2.39.2
#14 lore raw

PATCH 13/15 net/mlx5: Add new WQE for updating flow table

In reply to Saeed Mahameed

From: Yevgeny Kliteynik <[email protected]>

Add new WQE type: FLOW_TBL_ACCESS, which will be used for writing modify header arguments.
This type has specific control segment and special data segment.

---
 include/linux/mlx5/device.h |  2 ++
 include/linux/mlx5/qp.h     | 10 ++++++++++
 2 files changed, 12 insertions(+)
include/linux/mlx5/device.h +2 -0
index af4dd536a52c..e4aa147ab390 100644
@@ -442,6 +442,8 @@ enum {
 
 	MLX5_OPCODE_UMR			= 0x25,
 
+	MLX5_OPCODE_FLOW_TBL_ACCESS	= 0x2c,
+
 	MLX5_OPCODE_ACCESS_ASO		= 0x2d,
 };
 
include/linux/mlx5/qp.h +10 -0
index df55fbb65717..bd53cf4be7bd 100644
@@ -499,6 +499,16 @@ struct mlx5_stride_block_ctrl_seg {
 	__be16		num_entries;
 };
 
+struct mlx5_wqe_flow_update_ctrl_seg {
+	__be32		flow_idx_update;
+	__be32		dest_handle;
+	u8		reserved0[40];
+};
+
+struct mlx5_wqe_header_modify_argument_update_seg {
+	u8		argument_list[64];
+};
+
 struct mlx5_core_qp {
 	struct mlx5_core_rsc_common	common; /* must be first */
 	void (*event)		(struct mlx5_core_qp *, int);
Signature · 1 line
2.39.2
#11 lore raw

PATCH 10/15 net/mlx5: Create a new profile for SFs

In reply to Saeed Mahameed

From: Parav Pandit <[email protected]>

Create a new profile for SFs in order to disable the command cache. Each function command cache consumes ~500KB of memory, when using a large number of SFs this savings is notable on memory constarined systems.

Use a new profile to provide for future differences between SFs and PFs.

The mr_cache not used for non-PF functions, so it is excluded from the new profile.

---
 drivers/net/ethernet/mellanox/mlx5/core/cmd.c           | 6 +++---
 drivers/net/ethernet/mellanox/mlx5/core/main.c          | 9 +++++++++
 drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h     | 1 +
 drivers/net/ethernet/mellanox/mlx5/core/sf/dev/driver.c | 2 +-
 include/linux/mlx5/driver.h                             | 1 +
 5 files changed, 15 insertions(+), 4 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/cmd.c +3 -3
index b00e33ed05e9..d53de39539a8 100644
@@ -1802,7 +1802,7 @@ static struct mlx5_cmd_msg *alloc_msg(struct mlx5_core_dev *dev, int in_size,
 	if (in_size <= 16)
 		goto cache_miss;
 
-	for (i = 0; i < MLX5_NUM_COMMAND_CACHES; i++) {
+	for (i = 0; i < dev->profile.num_cmd_caches; i++) {
 		ch = &cmd->cache[i];
 		if (in_size > ch->max_inbox_size)
 			continue;
@@ -2097,7 +2097,7 @@ static void destroy_msg_cache(struct mlx5_core_dev *dev)
 	struct mlx5_cmd_msg *n;
 	int i;
 
-	for (i = 0; i < MLX5_NUM_COMMAND_CACHES; i++) {
+	for (i = 0; i < dev->profile.num_cmd_caches; i++) {
 		ch = &dev->cmd.cache[i];
 		list_for_each_entry_safe(msg, n, &ch->head, list) {
 			list_del(&msg->list);
@@ -2127,7 +2127,7 @@ static void create_msg_cache(struct mlx5_core_dev *dev)
 	int k;
 
 	/* Initialize and fill the caches with initial entries */
-	for (k = 0; k < MLX5_NUM_COMMAND_CACHES; k++) {
+	for (k = 0; k < dev->profile.num_cmd_caches; k++) {
 		ch = &cmd->cache[k];
 		spin_lock_init(&ch->lock);
 		INIT_LIST_HEAD(&ch->head);
drivers/net/ethernet/mellanox/mlx5/core/main.c +9 -0
index f95df73d1089..a95d1218def9 100644
@@ -100,15 +100,19 @@ enum {
 static struct mlx5_profile profile[] = {
 	[0] = {
 		.mask           = 0,
+		.num_cmd_caches = MLX5_NUM_COMMAND_CACHES,
 	},
 	[1] = {
 		.mask		= MLX5_PROF_MASK_QP_SIZE,
 		.log_max_qp	= 12,
+		.num_cmd_caches = MLX5_NUM_COMMAND_CACHES,
+
 	},
 	[2] = {
 		.mask		= MLX5_PROF_MASK_QP_SIZE |
 				  MLX5_PROF_MASK_MR_CACHE,
 		.log_max_qp	= LOG_MAX_SUPPORTED_QPS,
+		.num_cmd_caches = MLX5_NUM_COMMAND_CACHES,
 		.mr_cache[0]	= {
 			.size	= 500,
 			.limit	= 250
@@ -174,6 +178,11 @@ static struct mlx5_profile profile[] = {
 			.limit	= 4
 		},
 	},
+	[3] = {
+		.mask		= MLX5_PROF_MASK_QP_SIZE,
+		.log_max_qp	= LOG_MAX_SUPPORTED_QPS,
+		.num_cmd_caches = 0,
+	},
 };
 
 static int wait_fw_init(struct mlx5_core_dev *dev, u32 max_wait_mili,
drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h +1 -0
index be0785f83083..5eaab99678ee 100644
@@ -142,6 +142,7 @@ enum mlx5_semaphore_space_address {
 };
 
 #define MLX5_DEFAULT_PROF       2
+#define MLX5_SF_PROF		3
 
 static inline int mlx5_flexible_inlen(struct mlx5_core_dev *dev, size_t fixed,
 				      size_t item_size, size_t num_items,
drivers/net/ethernet/mellanox/mlx5/core/sf/dev/driver.c +1 -1
index a7377619ba6f..e2f26d0bc615 100644
@@ -28,7 +28,7 @@ static int mlx5_sf_dev_probe(struct auxiliary_device *adev, const struct auxilia
 	mdev->priv.adev_idx = adev->id;
 	sf_dev->mdev = mdev;
 
-	err = mlx5_mdev_init(mdev, MLX5_DEFAULT_PROF);
+	err = mlx5_mdev_init(mdev, MLX5_SF_PROF);
 	if (err) {
 		mlx5_core_warn(mdev, "mlx5_mdev_init on err=%d\n", err);
 		goto mdev_err;
include/linux/mlx5/driver.h +1 -0
index f243bd10a5e1..135a3c8d8237 100644
@@ -751,6 +751,7 @@ enum {
 struct mlx5_profile {
 	u64	mask;
 	u8	log_max_qp;
+	u8	num_cmd_caches;
 	struct {
 		int	size;
 		int	limit;
Signature · 1 line
2.39.2
#15 lore raw

PATCH 14/15 net/mlx5: DR, Prepare sending new WQE type

In reply to Saeed Mahameed

From: Yevgeny Kliteynik <[email protected]>

The send engine should be ready to handle more opcodes in addition to RDMA_WRITE/RDMA_READ.

---
 .../mellanox/mlx5/core/steering/dr_send.c     | 60 ++++++++++++-------
 1 file changed, 39 insertions(+), 21 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/steering/dr_send.c +39 -21
index fd2d31cdbcf9..00bb65613300 100644
@@ -18,7 +18,12 @@ struct dr_data_seg {
 	unsigned int send_flags;
 };
 
+enum send_info_type {
+	WRITE_ICM = 0,
+};
+
 struct postsend_info {
+	enum send_info_type type;
 	struct dr_data_seg write;
 	struct dr_data_seg read;
 	u64 remote_addr;
@@ -402,10 +407,12 @@ static void dr_rdma_segments(struct mlx5dr_qp *dr_qp, u64 remote_addr,
 
 static void dr_post_send(struct mlx5dr_qp *dr_qp, struct postsend_info *send_info)
 {
-	dr_rdma_segments(dr_qp, send_info->remote_addr, send_info->rkey,
-			 &send_info->write, MLX5_OPCODE_RDMA_WRITE, false);
-	dr_rdma_segments(dr_qp, send_info->remote_addr, send_info->rkey,
-			 &send_info->read, MLX5_OPCODE_RDMA_READ, true);
+	if (send_info->type == WRITE_ICM) {
+		dr_rdma_segments(dr_qp, send_info->remote_addr, send_info->rkey,
+				 &send_info->write, MLX5_OPCODE_RDMA_WRITE, false);
+		dr_rdma_segments(dr_qp, send_info->remote_addr, send_info->rkey,
+				 &send_info->read, MLX5_OPCODE_RDMA_READ, true);
+	}
 }
 
 /**
@@ -476,9 +483,26 @@ static int dr_handle_pending_wc(struct mlx5dr_domain *dmn,
 	return 0;
 }
 
-static void dr_fill_data_segs(struct mlx5dr_send_ring *send_ring,
-			      struct postsend_info *send_info)
+static void dr_fill_write_icm_segs(struct mlx5dr_domain *dmn,
+				   struct mlx5dr_send_ring *send_ring,
+				   struct postsend_info *send_info)
 {
+	u32 buff_offset;
+
+	if (send_info->write.length > dmn->info.max_inline_size) {
+		buff_offset = (send_ring->tx_head &
+			       (dmn->send_ring->signal_th - 1)) *
+			      send_ring->max_post_send_size;
+		/* Copy to ring mr */
+		memcpy(send_ring->buf + buff_offset,
+		       (void *)(uintptr_t)send_info->write.addr,
+		       send_info->write.length);
+		send_info->write.addr = (uintptr_t)send_ring->mr->dma_addr + buff_offset;
+		send_info->write.lkey = send_ring->mr->mkey;
+
+		send_ring->tx_head++;
+	}
+
 	send_ring->pending_wqe++;
 
 	if (send_ring->pending_wqe % send_ring->signal_th == 0)
@@ -496,11 +520,18 @@ static void dr_fill_data_segs(struct mlx5dr_send_ring *send_ring,
 		send_info->read.send_flags = 0;
 }
 
+static void dr_fill_data_segs(struct mlx5dr_domain *dmn,
+			      struct mlx5dr_send_ring *send_ring,
+			      struct postsend_info *send_info)
+{
+	if (send_info->type == WRITE_ICM)
+		dr_fill_write_icm_segs(dmn, send_ring, send_info);
+}
+
 static int dr_postsend_icm_data(struct mlx5dr_domain *dmn,
 				struct postsend_info *send_info)
 {
 	struct mlx5dr_send_ring *send_ring = dmn->send_ring;
-	u32 buff_offset;
 	int ret;
 
 	if (unlikely(dmn->mdev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR ||
@@ -517,20 +548,7 @@ static int dr_postsend_icm_data(struct mlx5dr_domain *dmn,
 	if (ret)
 		goto out_unlock;
 
-	if (send_info->write.length > dmn->info.max_inline_size) {
-		buff_offset = (send_ring->tx_head &
-			       (dmn->send_ring->signal_th - 1)) *
-			send_ring->max_post_send_size;
-		/* Copy to ring mr */
-		memcpy(send_ring->buf + buff_offset,
-		       (void *)(uintptr_t)send_info->write.addr,
-		       send_info->write.length);
-		send_info->write.addr = (uintptr_t)send_ring->mr->dma_addr + buff_offset;
-		send_info->write.lkey = send_ring->mr->mkey;
-	}
-
-	send_ring->tx_head++;
-	dr_fill_data_segs(send_ring, send_info);
+	dr_fill_data_segs(dmn, send_ring, send_info);
 	dr_post_send(send_ring->qp, send_info);
 
 out_unlock:
Signature · 1 line
2.39.2
#16 lore raw

PATCH 15/15 net/mlx5: DR, Add modify-header-pattern ICM pool

In reply to Saeed Mahameed

From: Yevgeny Kliteynik <[email protected]>

There is a new ICM area for that memory, so we need to handle it as we did for the others ICM types.
The patch added that specific pool with its requirements and management.

---
 .../net/ethernet/mellanox/mlx5/core/Makefile  |  2 +-
 .../mellanox/mlx5/core/steering/dr_cmd.c      |  6 +++
 .../mellanox/mlx5/core/steering/dr_domain.c   | 45 +++++++++++++++++--
 .../mellanox/mlx5/core/steering/dr_icm_pool.c | 41 ++++++++++++-----
 .../mellanox/mlx5/core/steering/dr_ptrn.c     | 43 ++++++++++++++++++
 .../mellanox/mlx5/core/steering/dr_types.h    | 11 +++++
 6 files changed, 132 insertions(+), 16 deletions(-)
 create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ptrn.c
drivers/net/ethernet/mellanox/mlx5/core/Makefile +1 -1
index 4a009b720bee..39c2c8dc7e07 100644
@@ -112,7 +112,7 @@ mlx5_core-$(CONFIG_MLX5_SW_STEERING) += steering/dr_domain.o steering/dr_table.o
 					steering/dr_ste_v2.o \
 					steering/dr_cmd.o steering/dr_fw.o \
 					steering/dr_action.o steering/fs_dr.o \
-					steering/dr_definer.o \
+					steering/dr_definer.o steering/dr_ptrn.o \
 					steering/dr_dbg.o lib/smfs.o
 #
 # SF device
drivers/net/ethernet/mellanox/mlx5/core/steering/dr_cmd.c +6 -0
index 07b6a6dcb92f..229f3684100c 100644
@@ -200,6 +200,12 @@ int mlx5dr_cmd_query_device(struct mlx5_core_dev *mdev,
 	caps->hdr_modify_icm_addr =
 		MLX5_CAP64_DEV_MEM(mdev, header_modify_sw_icm_start_address);
 
+	caps->log_modify_pattern_icm_size =
+		MLX5_CAP_DEV_MEM(mdev, log_header_modify_pattern_sw_icm_size);
+
+	caps->hdr_modify_pattern_icm_addr =
+		MLX5_CAP64_DEV_MEM(mdev, header_modify_pattern_sw_icm_start_address);
+
 	caps->roce_min_src_udp = MLX5_CAP_ROCE(mdev, r_roce_min_src_udp_port);
 
 	caps->is_ecpf = mlx5_core_is_ecpf_esw_manager(mdev);
drivers/net/ethernet/mellanox/mlx5/core/steering/dr_domain.c +42 -3

88 lines not on this page · raw message

drivers/net/ethernet/mellanox/mlx5/core/steering/dr_icm_pool.c +29 -12

79 lines not on this page · raw message

drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ptrn.c new file +43 -0

43 lines not on this page · raw message

drivers/net/ethernet/mellanox/mlx5/core/steering/dr_types.h +11 -0
index 2b769dcbd453..5b9faa714f42 100644
@@ -26,6 +26,8 @@
 #define mlx5dr_info(dmn, arg...) mlx5_core_info((dmn)->mdev, ##arg)
 #define mlx5dr_dbg(dmn, arg...) mlx5_core_dbg((dmn)->mdev, ##arg)
 
+struct mlx5dr_ptrn_mgr;
+
 static inline bool dr_is_flex_parser_0_id(u8 parser_id)
 {
 	return parser_id <= DR_STE_MAX_FLEX_0_ID;
@@ -66,6 +68,7 @@ enum mlx5dr_icm_chunk_size {
 enum mlx5dr_icm_type {
 	DR_ICM_TYPE_STE,
 	DR_ICM_TYPE_MODIFY_ACTION,
+	DR_ICM_TYPE_MODIFY_HDR_PTRN,
 };
 
 static inline enum mlx5dr_icm_chunk_size
@@ -861,6 +864,8 @@ struct mlx5dr_cmd_caps {
 	u64 esw_tx_drop_address;
 	u32 log_icm_size;
 	u64 hdr_modify_icm_addr;
+	u32 log_modify_pattern_icm_size;
+	u64 hdr_modify_pattern_icm_addr;
 	u32 flex_protocols;
 	u8 flex_parser_id_icmp_dw0;
 	u8 flex_parser_id_icmp_dw1;
@@ -910,6 +915,7 @@ struct mlx5dr_domain_info {
 	u32 max_send_wr;
 	u32 max_log_sw_icm_sz;
 	u32 max_log_action_icm_sz;
+	u32 max_log_modify_hdr_pattern_icm_sz;
 	struct mlx5dr_domain_rx_tx rx;
 	struct mlx5dr_domain_rx_tx tx;
 	struct mlx5dr_cmd_caps caps;
@@ -928,6 +934,7 @@ struct mlx5dr_domain {
 	struct mlx5dr_send_info_pool *send_info_pool_tx;
 	struct kmem_cache *chunks_kmem_cache;
 	struct kmem_cache *htbls_kmem_cache;
+	struct mlx5dr_ptrn_mgr *ptrn_mgr;
 	struct mlx5dr_send_ring *send_ring;
 	struct mlx5dr_domain_info info;
 	struct xarray csum_fts_xa;
@@ -1526,4 +1533,8 @@ static inline bool mlx5dr_supp_match_ranges(struct mlx5_core_dev *dev)
 			(1ULL << MLX5_IFC_DEFINER_FORMAT_ID_SELECT));
 }
 
+bool mlx5dr_domain_is_support_ptrn_arg(struct mlx5dr_domain *dmn);
+struct mlx5dr_ptrn_mgr *mlx5dr_ptrn_mgr_create(struct mlx5dr_domain *dmn);
+void mlx5dr_ptrn_mgr_destroy(struct mlx5dr_ptrn_mgr *mgr);
+
 #endif  /* _DR_TYPES_H_ */
Signature · 1 line
2.39.2
Reviewed-by
#17 lore raw
In reply to Saeed Mahameed
Quoted 10 lines · Saeed Mahameed

From: Saeed Mahameed <[email protected]>

This series provides some updates to mlx5 driver For more information please see detailed tag log below.

Please pull and let me know if there is any problem.

Thanks,
Saeed.

The whole series looks good to me, though I don't have a full grasp on 1-9's multicast offloading support, I didn't see any problems while reading through it. The other changes are straight forward and obvious improvements.

Thanks,
Jake

#2 lore raw

Re: [net-next 01/15] net/mlx5: Add mlx5_ifc definitions for bridge multicast support

In reply to Saeed Mahameed
This series was applied to netdev/net-next.git (main) by Saeed Mahameed <[email protected]>:

Hello:

This series was applied to netdev/net-next.git (main) by Saeed Mahameed <[email protected]>:

Quoted 11 lines · On Tue, 11 Apr 2023 21:07:38 -0700 you wrote:

From: Vlad Buslov <[email protected]>

Add the required hardware definitions to mlx5_ifc: fdb_uplink_hairpin, fdb_multi_path_any_table_limit_regc, fdb_multi_path_any_table.

[...]

Here is the summary with links:
- [net-next,01/15] net/mlx5: Add mlx5_ifc definitions for bridge multicast support

    https://git.kernel.org/netdev/net-next/c/e5688f6fb9e3
  - [net-next,02/15] net/mlx5: Bridge, increase bridge tables sizes
    https://git.kernel.org/netdev/net-next/c/9071b423c302
  - [net-next,03/15] net/mlx5: Bridge, move additional data structures to priv header
    https://git.kernel.org/netdev/net-next/c/6767c97d7adc
  - [net-next,04/15] net/mlx5: Bridge, extract code to lookup parent bridge of port
    https://git.kernel.org/netdev/net-next/c/b99c4ef29e27
  - [net-next,05/15] net/mlx5: Bridge, snoop igmp/mld packets
    https://git.kernel.org/netdev/net-next/c/18c2916cee12
  - [net-next,06/15] net/mlx5: Bridge, add per-port multicast replication tables
    https://git.kernel.org/netdev/net-next/c/272ecfc92f6f
  - [net-next,07/15] net/mlx5: Bridge, support multicast VLAN pop
    https://git.kernel.org/netdev/net-next/c/b5e80625d168
  - [net-next,08/15] net/mlx5: Bridge, implement mdb offload
    https://git.kernel.org/netdev/net-next/c/70f0302b3f20
  - [net-next,09/15] net/mlx5: Bridge, add tracepoints for multicast
    https://git.kernel.org/netdev/net-next/c/55f3e740f7f6
  - [net-next,10/15] net/mlx5: Create a new profile for SFs
    https://git.kernel.org/netdev/net-next/c/9df839a711ae
  - [net-next,11/15] net/mlx5: DR, Set counter ID on the last STE for STEv1 TX
    https://git.kernel.org/netdev/net-next/c/cee6484eddc1
  - [net-next,12/15] net/mlx5: Add mlx5_ifc bits for modify header argument
    https://git.kernel.org/netdev/net-next/c/9fa7f1de3dda
  - [net-next,13/15] net/mlx5: Add new WQE for updating flow table
    https://git.kernel.org/netdev/net-next/c/977c4a3e7c89
  - [net-next,14/15] net/mlx5: DR, Prepare sending new WQE type
    https://git.kernel.org/netdev/net-next/c/1e5cc7369bb0
  - [net-next,15/15] net/mlx5: DR, Add modify-header-pattern ICM pool
    https://git.kernel.org/netdev/net-next/c/108ff8215b55

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html