Merge branch 'Add-packet-trap-policers-support'

Ido Schimmel says:

====================
Add packet trap policers support

Background
==========

Devices capable of offloading the kernel's datapath and perform
functions such as bridging and routing must also be able to send (trap)
specific packets to the kernel (i.e., the CPU) for processing.

For example, a device acting as a multicast-aware bridge must be able to
trap IGMP membership reports to the kernel for processing by the bridge
module.

Motivation
==========

In most cases, the underlying device is capable of handling packet rates
that are several orders of magnitude higher compared to those that can
be handled by the CPU.

Therefore, in order to prevent the underlying device from overwhelming
the CPU, devices usually include packet trap policers that are able to
police the trapped packets to rates that can be handled by the CPU.

Proposed solution
=================

This patch set allows capable device drivers to register their supported
packet trap policers with devlink. User space can then tune the
parameters of these policers (currently, rate and burst size) and read
from the device the number of packets that were dropped by the policer,
if supported.

These packet trap policers can then be bound to existing packet trap
groups, which are used to aggregate logically related packet traps. As a
result, trapped packets are policed to rates that can be handled the
host CPU.

Example usage
=============

Instantiate netdevsim:

Dump available packet trap policers:
netdevsim/netdevsim10:
  policer 1 rate 1000 burst 128
  policer 2 rate 2000 burst 256
  policer 3 rate 3000 burst 512

Change the parameters of a packet trap policer:

Bind a packet trap policer to a packet trap group:

Dump parameters and statistics of a packet trap policer:
netdevsim/netdevsim10:
  policer 3 rate 100 burst 16
    stats:
        rx:
          dropped 92

Unbind a packet trap policer from a packet trap group:

Patch set overview
==================

Patch #1 adds the core infrastructure in devlink which allows capable
device drivers to register their supported packet trap policers with
devlink.

Patch #2 extends the existing devlink-trap documentation.

Patch #3 extends netdevsim to register a few dummy packet trap policers
with devlink. Used later on to selftests the core infrastructure.

Patches #4-#5 adds infrastructure in devlink to allow binding of packet
trap policers to packet trap groups.

Patch #6 extends netdevsim to allow such binding.

Patch #7 adds a selftest over netdevsim that verifies the core
devlink-trap policers functionality.

Patches #8-#14 gradually add devlink-trap policers support in mlxsw.

Patch #15 adds a selftest over mlxsw. All registered packet trap
policers are verified to handle the configured rate and burst size.

Future plans
============

* Allow changing default association between packet traps and packet
  trap groups
* Add more packet traps. For example, for control packets (e.g., IGMP)

v3:
* Rebase

v2 (address comments from Jiri and Jakub):
* Patch #1: Add 'strict_start_type' in devlink policy
* Patch #1: Have device drivers provide max/min rate/burst size for each
  policer. Use them to check validity of user provided parameters
* Patch #3: Remove check about burst size being a power of 2 and instead
  add a debugfs knob to fail the operation
* Patch #3: Provide max/min rate/burst size when registering policers
  and remove the validity checks from nsim_dev_devlink_trap_policer_set()
* Patch #5: Check for presence of 'DEVLINK_ATTR_TRAP_POLICER_ID' in
  devlink_trap_group_set() and bail if not present
* Patch #5: Add extack error message in case trap group was partially
  modified
* Patch #7: Add test case with new 'fail_trap_policer_set' knob
* Patch #7: Add test case for partially modified trap group
* Patch #10: Provide max/min rate/burst size when registering policers
* Patch #11: Remove the max/min validity checks from
  __mlxsw_sp_trap_policer_set()
====================

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

Documentation/networking/devlink/devlink-trap.rst

@@ -287,6 +287,32 @@ narrow. The description of these groups must be added to the following table:
      - Contains packet traps for packets that were dropped by the device during
        ACL processing
 
+Packet Trap Policers
+====================
+
+As previously explained, the underlying device can trap certain packets to the
+CPU for processing. In most cases, the underlying device is capable of handling
+packet rates that are several orders of magnitude higher compared to those that
+can be handled by the CPU.
+
+Therefore, in order to prevent the underlying device from overwhelming the CPU,
+devices usually include packet trap policers that are able to police the
+trapped packets to rates that can be handled by the CPU.
+
+The ``devlink-trap`` mechanism allows capable device drivers to register their
+supported packet trap policers with ``devlink``. The device driver can choose
+to associate these policers with supported packet trap groups (see
+:ref:`Generic-Packet-Trap-Groups`) during its initialization, thereby exposing
+its default control plane policy to user space.
+
+Device drivers should allow user space to change the parameters of the policers
+(e.g., rate, burst size) as well as the association between the policers and
+trap groups by implementing the relevant callbacks.
+
+If possible, device drivers should implement a callback that allows user space
+to retrieve the number of packets that were dropped by the policer because its
+configured policy was violated.
+
 Testing
 =======
 

drivers/net/ethernet/mellanox/mlxsw/core.c

@@ -1198,6 +1198,72 @@ mlxsw_devlink_trap_group_init(struct devlink *devlink,
 	return mlxsw_driver->trap_group_init(mlxsw_core, group);
 }
 
+static int
+mlxsw_devlink_trap_group_set(struct devlink *devlink,
+			     const struct devlink_trap_group *group,
+			     const struct devlink_trap_policer *policer)
+{
+	struct mlxsw_core *mlxsw_core = devlink_priv(devlink);
+	struct mlxsw_driver *mlxsw_driver = mlxsw_core->driver;
+
+	if (!mlxsw_driver->trap_group_set)
+		return -EOPNOTSUPP;
+	return mlxsw_driver->trap_group_set(mlxsw_core, group, policer);
+}
+
+static int
+mlxsw_devlink_trap_policer_init(struct devlink *devlink,
+				const struct devlink_trap_policer *policer)
+{
+	struct mlxsw_core *mlxsw_core = devlink_priv(devlink);
+	struct mlxsw_driver *mlxsw_driver = mlxsw_core->driver;
+
+	if (!mlxsw_driver->trap_policer_init)
+		return -EOPNOTSUPP;
+	return mlxsw_driver->trap_policer_init(mlxsw_core, policer);
+}
+
+static void
+mlxsw_devlink_trap_policer_fini(struct devlink *devlink,
+				const struct devlink_trap_policer *policer)
+{
+	struct mlxsw_core *mlxsw_core = devlink_priv(devlink);
+	struct mlxsw_driver *mlxsw_driver = mlxsw_core->driver;
+
+	if (!mlxsw_driver->trap_policer_fini)
+		return;
+	mlxsw_driver->trap_policer_fini(mlxsw_core, policer);
+}
+
+static int
+mlxsw_devlink_trap_policer_set(struct devlink *devlink,
+			       const struct devlink_trap_policer *policer,
+			       u64 rate, u64 burst,
+			       struct netlink_ext_ack *extack)
+{
+	struct mlxsw_core *mlxsw_core = devlink_priv(devlink);
+	struct mlxsw_driver *mlxsw_driver = mlxsw_core->driver;
+
+	if (!mlxsw_driver->trap_policer_set)
+		return -EOPNOTSUPP;
+	return mlxsw_driver->trap_policer_set(mlxsw_core, policer, rate, burst,
+					      extack);
+}
+
+static int
+mlxsw_devlink_trap_policer_counter_get(struct devlink *devlink,
+				       const struct devlink_trap_policer *policer,
+				       u64 *p_drops)
+{
+	struct mlxsw_core *mlxsw_core = devlink_priv(devlink);
+	struct mlxsw_driver *mlxsw_driver = mlxsw_core->driver;
+
+	if (!mlxsw_driver->trap_policer_counter_get)
+		return -EOPNOTSUPP;
+	return mlxsw_driver->trap_policer_counter_get(mlxsw_core, policer,
+						      p_drops);
+}
+
 static const struct devlink_ops mlxsw_devlink_ops = {
 	.reload_down		= mlxsw_devlink_core_bus_device_reload_down,
 	.reload_up		= mlxsw_devlink_core_bus_device_reload_up,
@@ -1220,6 +1286,11 @@ static const struct devlink_ops mlxsw_devlink_ops = {
 	.trap_fini			= mlxsw_devlink_trap_fini,
 	.trap_action_set		= mlxsw_devlink_trap_action_set,
 	.trap_group_init		= mlxsw_devlink_trap_group_init,
+	.trap_group_set			= mlxsw_devlink_trap_group_set,
+	.trap_policer_init		= mlxsw_devlink_trap_policer_init,
+	.trap_policer_fini		= mlxsw_devlink_trap_policer_fini,
+	.trap_policer_set		= mlxsw_devlink_trap_policer_set,
+	.trap_policer_counter_get	= mlxsw_devlink_trap_policer_counter_get,
 };
 
 static int

drivers/net/ethernet/mellanox/mlxsw/core.h

@@ -327,6 +327,20 @@ struct mlxsw_driver {
 			       enum devlink_trap_action action);
 	int (*trap_group_init)(struct mlxsw_core *mlxsw_core,
 			       const struct devlink_trap_group *group);
+	int (*trap_group_set)(struct mlxsw_core *mlxsw_core,
+			      const struct devlink_trap_group *group,
+			      const struct devlink_trap_policer *policer);
+	int (*trap_policer_init)(struct mlxsw_core *mlxsw_core,
+				 const struct devlink_trap_policer *policer);
+	void (*trap_policer_fini)(struct mlxsw_core *mlxsw_core,
+				  const struct devlink_trap_policer *policer);
+	int (*trap_policer_set)(struct mlxsw_core *mlxsw_core,
+				const struct devlink_trap_policer *policer,
+				u64 rate, u64 burst,
+				struct netlink_ext_ack *extack);
+	int (*trap_policer_counter_get)(struct mlxsw_core *mlxsw_core,
+					const struct devlink_trap_policer *policer,
+					u64 *p_drops);
 	void (*txhdr_construct)(struct sk_buff *skb,
 				const struct mlxsw_tx_info *tx_info);
 	int (*resources_register)(struct mlxsw_core *mlxsw_core);

drivers/net/ethernet/mellanox/mlxsw/reg.h

@@ -3296,6 +3296,12 @@ MLXSW_ITEM32(reg, qpcr, g, 0x00, 14, 2);
  */
 MLXSW_ITEM32(reg, qpcr, pid, 0x00, 0, 14);
 
+/* reg_qpcr_clear_counter
+ * Clear counters.
+ * Access: OP
+ */
+MLXSW_ITEM32(reg, qpcr, clear_counter, 0x04, 31, 1);
+
 /* reg_qpcr_color_aware
  * Is the policer aware of colors.
  * Must be 0 (unaware) for cpu port.
@@ -3393,6 +3399,17 @@ enum mlxsw_reg_qpcr_action {
  */
 MLXSW_ITEM32(reg, qpcr, violate_action, 0x18, 0, 4);
 
+/* reg_qpcr_violate_count
+ * Counts the number of times violate_action happened on this PID.
+ * Access: RW
+ */
+MLXSW_ITEM64(reg, qpcr, violate_count, 0x20, 0, 64);
+
+#define MLXSW_REG_QPCR_LOWEST_CIR	1
+#define MLXSW_REG_QPCR_HIGHEST_CIR	(2 * 1000 * 1000 * 1000) /* 2Gpps */
+#define MLXSW_REG_QPCR_LOWEST_CBS	4
+#define MLXSW_REG_QPCR_HIGHEST_CBS	24
+
 static inline void mlxsw_reg_qpcr_pack(char *payload, u16 pid,
 				       enum mlxsw_reg_qpcr_ir_units ir_units,
 				       bool bytes, u32 cir, u16 cbs)
@@ -5520,12 +5537,10 @@ enum mlxsw_reg_htgt_trap_group {
 	MLXSW_REG_HTGT_TRAP_GROUP_SP_PIM,
 	MLXSW_REG_HTGT_TRAP_GROUP_SP_MULTICAST,
 	MLXSW_REG_HTGT_TRAP_GROUP_SP_ARP,
-	MLXSW_REG_HTGT_TRAP_GROUP_SP_HOST_MISS,
 	MLXSW_REG_HTGT_TRAP_GROUP_SP_ROUTER_EXP,
 	MLXSW_REG_HTGT_TRAP_GROUP_SP_REMOTE_ROUTE,
 	MLXSW_REG_HTGT_TRAP_GROUP_SP_IP2ME,
 	MLXSW_REG_HTGT_TRAP_GROUP_SP_DHCP,
-	MLXSW_REG_HTGT_TRAP_GROUP_SP_RPF,
 	MLXSW_REG_HTGT_TRAP_GROUP_SP_EVENT,
 	MLXSW_REG_HTGT_TRAP_GROUP_SP_IPV6_MLD,
 	MLXSW_REG_HTGT_TRAP_GROUP_SP_IPV6_ND,

drivers/net/ethernet/mellanox/mlxsw/spectrum.c

@@ -43,6 +43,7 @@
 #include "spectrum_acl_flex_actions.h"
 #include "spectrum_span.h"
 #include "spectrum_ptp.h"
+#include "spectrum_trap.h"
 #include "../mlxfw/mlxfw.h"
 
 #define MLXSW_SP1_FWREV_MAJOR 13
@@ -4556,6 +4557,7 @@ static const struct mlxsw_listener mlxsw_sp1_listener[] = {
 
 static int mlxsw_sp_cpu_policers_set(struct mlxsw_core *mlxsw_core)
 {
+	struct mlxsw_sp *mlxsw_sp = mlxsw_core_driver_priv(mlxsw_core);
 	char qpcr_pl[MLXSW_REG_QPCR_LEN];
 	enum mlxsw_reg_qpcr_ir_units ir_units;
 	int max_cpu_policers;
@@ -4578,7 +4580,6 @@ static int mlxsw_sp_cpu_policers_set(struct mlxsw_core *mlxsw_core)
 		case MLXSW_REG_HTGT_TRAP_GROUP_SP_LLDP:
 		case MLXSW_REG_HTGT_TRAP_GROUP_SP_OSPF:
 		case MLXSW_REG_HTGT_TRAP_GROUP_SP_PIM:
-		case MLXSW_REG_HTGT_TRAP_GROUP_SP_RPF:
 		case MLXSW_REG_HTGT_TRAP_GROUP_SP_LBERROR:
 			rate = 128;
 			burst_size = 7;
@@ -4591,7 +4592,6 @@ static int mlxsw_sp_cpu_policers_set(struct mlxsw_core *mlxsw_core)
 		case MLXSW_REG_HTGT_TRAP_GROUP_SP_BGP:
 		case MLXSW_REG_HTGT_TRAP_GROUP_SP_ARP:
 		case MLXSW_REG_HTGT_TRAP_GROUP_SP_DHCP:
-		case MLXSW_REG_HTGT_TRAP_GROUP_SP_HOST_MISS:
 		case MLXSW_REG_HTGT_TRAP_GROUP_SP_ROUTER_EXP:
 		case MLXSW_REG_HTGT_TRAP_GROUP_SP_REMOTE_ROUTE:
 		case MLXSW_REG_HTGT_TRAP_GROUP_SP_IPV6_ND:
@@ -4619,6 +4619,7 @@ static int mlxsw_sp_cpu_policers_set(struct mlxsw_core *mlxsw_core)
 			continue;
 		}
 
+		__set_bit(i, mlxsw_sp->trap->policers_usage);
 		mlxsw_reg_qpcr_pack(qpcr_pl, i, ir_units, is_bytes, rate,
 				    burst_size);
 		err = mlxsw_reg_write(mlxsw_core, MLXSW_REG(qpcr), qpcr_pl);
@@ -4671,19 +4672,20 @@ static int mlxsw_sp_trap_groups_set(struct mlxsw_core *mlxsw_core)
 			break;
 		case MLXSW_REG_HTGT_TRAP_GROUP_SP_ARP:
 		case MLXSW_REG_HTGT_TRAP_GROUP_SP_IPV6_ND:
-		case MLXSW_REG_HTGT_TRAP_GROUP_SP_RPF:
 		case MLXSW_REG_HTGT_TRAP_GROUP_SP_PTP1:
 			priority = 2;
 			tc = 2;
 			break;
-		case MLXSW_REG_HTGT_TRAP_GROUP_SP_HOST_MISS:
 		case MLXSW_REG_HTGT_TRAP_GROUP_SP_ROUTER_EXP:
 		case MLXSW_REG_HTGT_TRAP_GROUP_SP_REMOTE_ROUTE:
 		case MLXSW_REG_HTGT_TRAP_GROUP_SP_MULTICAST:
-		case MLXSW_REG_HTGT_TRAP_GROUP_SP_LBERROR:
 			priority = 1;
 			tc = 1;
 			break;
+		case MLXSW_REG_HTGT_TRAP_GROUP_SP_LBERROR:
+			priority = 0;
+			tc = 1;
+			break;
 		case MLXSW_REG_HTGT_TRAP_GROUP_SP_EVENT:
 			priority = MLXSW_REG_HTGT_DEFAULT_PRIORITY;
 			tc = MLXSW_REG_HTGT_DEFAULT_TC;
@@ -4747,20 +4749,32 @@ static void mlxsw_sp_traps_unregister(struct mlxsw_sp *mlxsw_sp,
 
 static int mlxsw_sp_traps_init(struct mlxsw_sp *mlxsw_sp)
 {
+	struct mlxsw_sp_trap *trap;
+	u64 max_policers;
 	int err;
 
+	if (!MLXSW_CORE_RES_VALID(mlxsw_sp->core, MAX_CPU_POLICERS))
+		return -EIO;
+	max_policers = MLXSW_CORE_RES_GET(mlxsw_sp->core, MAX_CPU_POLICERS);
+	trap = kzalloc(struct_size(trap, policers_usage,
+				   BITS_TO_LONGS(max_policers)), GFP_KERNEL);
+	if (!trap)
+		return -ENOMEM;
+	trap->max_policers = max_policers;
+	mlxsw_sp->trap = trap;
+
 	err = mlxsw_sp_cpu_policers_set(mlxsw_sp->core);
 	if (err)
-		return err;
+		goto err_cpu_policers_set;
 
 	err = mlxsw_sp_trap_groups_set(mlxsw_sp->core);
 	if (err)
-		return err;
+		goto err_trap_groups_set;
 
 	err = mlxsw_sp_traps_register(mlxsw_sp, mlxsw_sp_listener,
 				      ARRAY_SIZE(mlxsw_sp_listener));
 	if (err)
-		return err;
+		goto err_traps_register;
 
 	err = mlxsw_sp_traps_register(mlxsw_sp, mlxsw_sp->listeners,
 				      mlxsw_sp->listeners_count);
@@ -4772,6 +4786,10 @@ static int mlxsw_sp_traps_init(struct mlxsw_sp *mlxsw_sp)
 err_extra_traps_init:
 	mlxsw_sp_traps_unregister(mlxsw_sp, mlxsw_sp_listener,
 				  ARRAY_SIZE(mlxsw_sp_listener));
+err_traps_register:
+err_trap_groups_set:
+err_cpu_policers_set:
+	kfree(trap);
 	return err;
 }
 
@@ -4781,6 +4799,7 @@ static void mlxsw_sp_traps_fini(struct mlxsw_sp *mlxsw_sp)
 				  mlxsw_sp->listeners_count);
 	mlxsw_sp_traps_unregister(mlxsw_sp, mlxsw_sp_listener,
 				  ARRAY_SIZE(mlxsw_sp_listener));
+	kfree(mlxsw_sp->trap);
 }
 
 #define MLXSW_SP_LAG_SEED_INIT 0xcafecafe
@@ -5655,6 +5674,11 @@ static struct mlxsw_driver mlxsw_sp1_driver = {
 	.trap_fini			= mlxsw_sp_trap_fini,
 	.trap_action_set		= mlxsw_sp_trap_action_set,
 	.trap_group_init		= mlxsw_sp_trap_group_init,
+	.trap_group_set			= mlxsw_sp_trap_group_set,
+	.trap_policer_init		= mlxsw_sp_trap_policer_init,
+	.trap_policer_fini		= mlxsw_sp_trap_policer_fini,
+	.trap_policer_set		= mlxsw_sp_trap_policer_set,
+	.trap_policer_counter_get	= mlxsw_sp_trap_policer_counter_get,
 	.txhdr_construct		= mlxsw_sp_txhdr_construct,
 	.resources_register		= mlxsw_sp1_resources_register,
 	.kvd_sizes_get			= mlxsw_sp_kvd_sizes_get,
@@ -5689,6 +5713,11 @@ static struct mlxsw_driver mlxsw_sp2_driver = {
 	.trap_fini			= mlxsw_sp_trap_fini,
 	.trap_action_set		= mlxsw_sp_trap_action_set,
 	.trap_group_init		= mlxsw_sp_trap_group_init,
+	.trap_group_set			= mlxsw_sp_trap_group_set,
+	.trap_policer_init		= mlxsw_sp_trap_policer_init,
+	.trap_policer_fini		= mlxsw_sp_trap_policer_fini,
+	.trap_policer_set		= mlxsw_sp_trap_policer_set,
+	.trap_policer_counter_get	= mlxsw_sp_trap_policer_counter_get,
 	.txhdr_construct		= mlxsw_sp_txhdr_construct,
 	.resources_register		= mlxsw_sp2_resources_register,
 	.params_register		= mlxsw_sp2_params_register,
@@ -5722,6 +5751,11 @@ static struct mlxsw_driver mlxsw_sp3_driver = {
 	.trap_fini			= mlxsw_sp_trap_fini,
 	.trap_action_set		= mlxsw_sp_trap_action_set,
 	.trap_group_init		= mlxsw_sp_trap_group_init,
+	.trap_group_set			= mlxsw_sp_trap_group_set,
+	.trap_policer_init		= mlxsw_sp_trap_policer_init,
+	.trap_policer_fini		= mlxsw_sp_trap_policer_fini,
+	.trap_policer_set		= mlxsw_sp_trap_policer_set,
+	.trap_policer_counter_get	= mlxsw_sp_trap_policer_counter_get,
 	.txhdr_construct		= mlxsw_sp_txhdr_construct,
 	.resources_register		= mlxsw_sp2_resources_register,
 	.params_register		= mlxsw_sp2_params_register,

drivers/net/ethernet/mellanox/mlxsw/spectrum.h

@@ -176,6 +176,7 @@ struct mlxsw_sp {
 	struct mlxsw_sp_ptp_state *ptp_state;
 	struct mlxsw_sp_counter_pool *counter_pool;
 	struct mlxsw_sp_span *span;
+	struct mlxsw_sp_trap *trap;
 	const struct mlxsw_fw_rev *req_rev;
 	const char *fw_filename;
 	const struct mlxsw_sp_kvdl_ops *kvdl_ops;
@@ -1022,6 +1023,22 @@ int mlxsw_sp_trap_action_set(struct mlxsw_core *mlxsw_core,
 			     enum devlink_trap_action action);
 int mlxsw_sp_trap_group_init(struct mlxsw_core *mlxsw_core,
 			     const struct devlink_trap_group *group);
+int mlxsw_sp_trap_group_set(struct mlxsw_core *mlxsw_core,
+			    const struct devlink_trap_group *group,
+			    const struct devlink_trap_policer *policer);
+int
+mlxsw_sp_trap_policer_init(struct mlxsw_core *mlxsw_core,
+			   const struct devlink_trap_policer *policer);
+void mlxsw_sp_trap_policer_fini(struct mlxsw_core *mlxsw_core,
+				const struct devlink_trap_policer *policer);
+int
+mlxsw_sp_trap_policer_set(struct mlxsw_core *mlxsw_core,
+			  const struct devlink_trap_policer *policer,
+			  u64 rate, u64 burst, struct netlink_ext_ack *extack);
+int
+mlxsw_sp_trap_policer_counter_get(struct mlxsw_core *mlxsw_core,
+				  const struct devlink_trap_policer *policer,
+				  u64 *p_drops);
 
 static inline struct net *mlxsw_sp_net(struct mlxsw_sp *mlxsw_sp)
 {