DPDK patches and discussions
 help / color / mirror / Atom feed
* [dpdk-dev] [PATCH 0/3] mlx5: LACP handling for bonded devices
@ 2020-06-23  8:41 Shiri Kuzin
  2020-06-23  8:41 ` [dpdk-dev] [PATCH 1/3] common/mlx5: add default miss action Shiri Kuzin
                   ` (3 more replies)
  0 siblings, 4 replies; 5+ messages in thread
From: Shiri Kuzin @ 2020-06-23  8:41 UTC (permalink / raw)
  To: dev; +Cc: matan, viacheslavo

When the driver has a bond configured, LACP traffic should be steered to kernel to manage the bond using a new rdma-core API.
This is done by creating a new internal action and adding a new devarg that will allow the user to manage the bond and get lacp traffic if they wish.  

Shiri Kuzin (3):
  common/mlx5: add default miss action
  net/mlx5: add default miss action to flow engine
  net/mlx5: add new devarg for LACP packets control

 doc/guides/nics/mlx5.rst               |  10 +++
 doc/guides/rel_notes/release_20_08.rst |   1 +
 drivers/common/mlx5/Makefile           |   5 ++
 drivers/common/mlx5/linux/meson.build  |   2 +
 drivers/common/mlx5/linux/mlx5_glue.c  |  13 ++++
 drivers/common/mlx5/linux/mlx5_glue.h  |   1 +
 drivers/net/mlx5/mlx5.c                |   6 ++
 drivers/net/mlx5/mlx5.h                |  11 +++
 drivers/net/mlx5/mlx5_flow.c           |  92 ++++++++++++++++++++++++
 drivers/net/mlx5/mlx5_flow.h           |  11 ++-
 drivers/net/mlx5/mlx5_flow_dv.c        | 127 +++++++++++++++++++++++++++++++--
 drivers/net/mlx5/mlx5_trigger.c        |   9 +++
 12 files changed, 280 insertions(+), 8 deletions(-)

-- 
1.8.3.1


^ permalink raw reply	[flat|nested] 5+ messages in thread

* [dpdk-dev] [PATCH 1/3] common/mlx5: add default miss action
  2020-06-23  8:41 [dpdk-dev] [PATCH 0/3] mlx5: LACP handling for bonded devices Shiri Kuzin
@ 2020-06-23  8:41 ` Shiri Kuzin
  2020-06-23  8:41 ` [dpdk-dev] [PATCH 2/3] net/mlx5: add default miss action to flow engine Shiri Kuzin
                   ` (2 subsequent siblings)
  3 siblings, 0 replies; 5+ messages in thread
From: Shiri Kuzin @ 2020-06-23  8:41 UTC (permalink / raw)
  To: dev; +Cc: matan, viacheslavo

Add dv_create_action_default_miss wrapper function
for the action added to the rdma-core
MLX5DV_FLOW_ACTION_DEFAULT_MISS.

When a packet matches MLX5DV_FLOW_ACTION_DEFAULT_MISS
action it is steered to the default miss of the verbs
steering domain.

Signed-off-by: Shiri Kuzin <shirik@mellanox.com>
Acked-by: Matan Azrad <matan@mellanox.com>
---
 drivers/common/mlx5/Makefile          |  5 +++++
 drivers/common/mlx5/linux/meson.build |  2 ++
 drivers/common/mlx5/linux/mlx5_glue.c | 13 +++++++++++++
 drivers/common/mlx5/linux/mlx5_glue.h |  1 +
 4 files changed, 21 insertions(+)

diff --git a/drivers/common/mlx5/Makefile b/drivers/common/mlx5/Makefile
index 622bde4..f6c762b 100644
--- a/drivers/common/mlx5/Makefile
+++ b/drivers/common/mlx5/Makefile
@@ -157,6 +157,11 @@ mlx5_autoconf.h.new: $(RTE_SDK)/buildtools/auto-config-h.sh
 		enum MLX5DV_FLOW_ACTION_COUNTERS_DEVX \
 		$(AUTOCONF_OUTPUT)
 	$Q sh -- '$<' '$@' \
+		HAVE_MLX5_DR_CREATE_ACTION_DEFAULT_MISS \
+		infiniband/mlx5dv.h \
+		enum MLX5DV_FLOW_ACTION_DEFAULT_MISS \
+		$(AUTOCONF_OUTPUT)
+	$Q sh -- '$<' '$@' \
 		HAVE_IBV_DEVX_ASYNC \
 		infiniband/mlx5dv.h \
 		func mlx5dv_devx_obj_query_async \
diff --git a/drivers/common/mlx5/linux/meson.build b/drivers/common/mlx5/linux/meson.build
index 638bb2b..2294213 100644
--- a/drivers/common/mlx5/linux/meson.build
+++ b/drivers/common/mlx5/linux/meson.build
@@ -95,6 +95,8 @@ has_sym_args = [
 	'mlx5dv_devx_obj_create' ],
 	[ 'HAVE_IBV_FLOW_DEVX_COUNTERS', 'infiniband/mlx5dv.h',
 	'MLX5DV_FLOW_ACTION_COUNTERS_DEVX' ],
+	[ 'HAVE_MLX5_DR_CREATE_ACTION_DEFAULT_MISS', 'infiniband/mlx5dv.h',
+	'MLX5DV_FLOW_ACTION_DEFAULT_MISS' ],
 	[ 'HAVE_IBV_DEVX_ASYNC', 'infiniband/mlx5dv.h',
 	'mlx5dv_devx_obj_query_async' ],
 	[ 'HAVE_IBV_DEVX_QP', 'infiniband/mlx5dv.h',
diff --git a/drivers/common/mlx5/linux/mlx5_glue.c b/drivers/common/mlx5/linux/mlx5_glue.c
index c91ee33..62c4cc7 100644
--- a/drivers/common/mlx5/linux/mlx5_glue.c
+++ b/drivers/common/mlx5/linux/mlx5_glue.c
@@ -795,6 +795,17 @@
 #endif
 }
 
+static void *
+mlx5_glue_dr_create_flow_action_default_miss(void)
+{
+#if defined(HAVE_MLX5DV_DR) && defined(HAVE_MLX5_DR_CREATE_ACTION_DEFAULT_MISS)
+	return mlx5dv_dr_action_create_default_miss();
+#else
+	errno = ENOTSUP;
+	return NULL;
+#endif
+}
+
 static int
 mlx5_glue_dv_destroy_flow(void *flow_id)
 {
@@ -1276,6 +1287,8 @@
 	.dv_create_flow_action_tag =  mlx5_glue_dv_create_flow_action_tag,
 	.dv_create_flow_action_meter = mlx5_glue_dv_create_flow_action_meter,
 	.dv_modify_flow_action_meter = mlx5_glue_dv_modify_flow_action_meter,
+	.dr_create_flow_action_default_miss =
+		mlx5_glue_dr_create_flow_action_default_miss,
 	.dv_destroy_flow = mlx5_glue_dv_destroy_flow,
 	.dv_destroy_flow_matcher = mlx5_glue_dv_destroy_flow_matcher,
 	.dv_open_device = mlx5_glue_dv_open_device,
diff --git a/drivers/common/mlx5/linux/mlx5_glue.h b/drivers/common/mlx5/linux/mlx5_glue.h
index 5d238a4..069d854 100644
--- a/drivers/common/mlx5/linux/mlx5_glue.h
+++ b/drivers/common/mlx5/linux/mlx5_glue.h
@@ -237,6 +237,7 @@ struct mlx5_glue {
 		(struct mlx5dv_dr_flow_meter_attr *attr);
 	int (*dv_modify_flow_action_meter)(void *action,
 		struct mlx5dv_dr_flow_meter_attr *attr, uint64_t modify_bits);
+	void *(*dr_create_flow_action_default_miss)(void);
 	int (*dv_destroy_flow)(void *flow);
 	int (*dv_destroy_flow_matcher)(void *matcher);
 	struct ibv_context *(*dv_open_device)(struct ibv_device *device);
-- 
1.8.3.1


^ permalink raw reply	[flat|nested] 5+ messages in thread

* [dpdk-dev] [PATCH 2/3] net/mlx5: add default miss action to flow engine
  2020-06-23  8:41 [dpdk-dev] [PATCH 0/3] mlx5: LACP handling for bonded devices Shiri Kuzin
  2020-06-23  8:41 ` [dpdk-dev] [PATCH 1/3] common/mlx5: add default miss action Shiri Kuzin
@ 2020-06-23  8:41 ` Shiri Kuzin
  2020-06-23  8:41 ` [dpdk-dev] [PATCH 3/3] net/mlx5: add new devarg for LACP packets control Shiri Kuzin
  2020-06-28 10:19 ` [dpdk-dev] [PATCH 0/3] mlx5: LACP handling for bonded devices Raslan Darawsheh
  3 siblings, 0 replies; 5+ messages in thread
From: Shiri Kuzin @ 2020-06-23  8:41 UTC (permalink / raw)
  To: dev; +Cc: matan, viacheslavo

The new action is an internal mlx5 action that will call
the rdma-core function MLX5DV_FLOW_ACTION_DEFAULT_MISS.

The default miss action will be used when a bond is
configured to allow traffic related to the bond to
be managed in the kernel.

Signed-off-by: Shiri Kuzin <shirik@mellanox.com>
Acked-by: Matan Azrad <matan@mellanox.com>
---
 drivers/net/mlx5/mlx5.h         |   9 +++
 drivers/net/mlx5/mlx5_flow.c    |  92 +++++++++++++++++++++++++++++
 drivers/net/mlx5/mlx5_flow.h    |  11 +++-
 drivers/net/mlx5/mlx5_flow_dv.c | 127 +++++++++++++++++++++++++++++++++++++---
 4 files changed, 231 insertions(+), 8 deletions(-)

diff --git a/drivers/net/mlx5/mlx5.h b/drivers/net/mlx5/mlx5.h
index 8a09ebc..93bde5a 100644
--- a/drivers/net/mlx5/mlx5.h
+++ b/drivers/net/mlx5/mlx5.h
@@ -451,6 +451,12 @@ struct mlx5_flow_counter_mng {
 	LIST_HEAD(stat_raws, mlx5_counter_stats_raw) free_stat_raws;
 };
 
+/* Default miss action resource structure. */
+struct mlx5_flow_default_miss_resource {
+	void *action; /* Pointer to the rdma-core action. */
+	rte_atomic32_t refcnt; /* Default miss action reference counter. */
+};
+
 #define MLX5_AGE_EVENT_NEW		1
 #define MLX5_AGE_TRIGGER		2
 #define MLX5_AGE_SET(age_info, BIT) \
@@ -559,6 +565,8 @@ struct mlx5_dev_ctx_shared {
 	uint32_t port_id_action_list; /* List of port ID actions. */
 	uint32_t push_vlan_action_list; /* List of push VLAN actions. */
 	struct mlx5_flow_counter_mng cmng; /* Counters management structure. */
+	struct mlx5_flow_default_miss_resource default_miss;
+	/* Default miss action resource structure. */
 	struct mlx5_indexed_pool *ipool[MLX5_IPOOL_MAX];
 	/* Memory Pool for mlx5 flow resources. */
 	struct mlx5_l3t_tbl *cnt_id_tbl; /* Shared counter lookup table. */
@@ -872,6 +880,7 @@ int mlx5_ctrl_flow_vlan(struct rte_eth_dev *dev,
 int mlx5_ctrl_flow(struct rte_eth_dev *dev,
 		   struct rte_flow_item_eth *eth_spec,
 		   struct rte_flow_item_eth *eth_mask);
+int mlx5_flow_lacp_miss(struct rte_eth_dev *dev);
 struct rte_flow *mlx5_flow_create_esw_table_zero_flow(struct rte_eth_dev *dev);
 int mlx5_flow_create_drop_queue(struct rte_eth_dev *dev);
 void mlx5_flow_delete_drop_queue(struct rte_eth_dev *dev);
diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index 3a48b89..4700ec1 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -1238,6 +1238,43 @@ uint32_t mlx5_flow_adjust_priority(struct rte_eth_dev *dev, int32_t priority,
 }
 
 /*
+ * Validate the default miss action.
+ *
+ * @param[in] action_flags
+ *   Bit-fields that holds the actions detected until now.
+ * @param[out] error
+ *   Pointer to error structure.
+ *
+ * @return
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
+ */
+int
+mlx5_flow_validate_action_default_miss(uint64_t action_flags,
+				const struct rte_flow_attr *attr,
+				struct rte_flow_error *error)
+{
+	if (action_flags & MLX5_FLOW_FATE_ACTIONS)
+		return rte_flow_error_set(error, EINVAL,
+					  RTE_FLOW_ERROR_TYPE_ACTION, NULL,
+					  "can't have 2 fate actions in"
+					  " same flow");
+	if (attr->egress)
+		return rte_flow_error_set(error, ENOTSUP,
+					  RTE_FLOW_ERROR_TYPE_ATTR_EGRESS, NULL,
+					  "default miss action not supported "
+					  "for egress");
+	if (attr->group)
+		return rte_flow_error_set(error, ENOTSUP,
+					  RTE_FLOW_ERROR_TYPE_ATTR_GROUP, NULL,
+					  "only group 0 is supported");
+	if (attr->transfer)
+		return rte_flow_error_set(error, ENOTSUP,
+					  RTE_FLOW_ERROR_TYPE_ATTR_TRANSFER,
+					  NULL, "transfer is not supported");
+	return 0;
+}
+
+/*
  * Validate the count action.
  *
  * @param[in] dev
@@ -4984,6 +5021,61 @@ struct rte_flow *
 }
 
 /**
+ * Create default miss flow rule matching lacp traffic
+ *
+ * @param dev
+ *   Pointer to Ethernet device.
+ * @param eth_spec
+ *   An Ethernet flow spec to apply.
+ *
+ * @return
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
+ */
+int
+mlx5_flow_lacp_miss(struct rte_eth_dev *dev)
+{
+	struct mlx5_priv *priv = dev->data->dev_private;
+	/*
+	 * The LACP matching is done by only using ether type since using
+	 * a multicast dst mac causes kernel to give low priority to this flow.
+	 */
+	static const struct rte_flow_item_eth lacp_spec = {
+		.type = RTE_BE16(0x8809),
+	};
+	static const struct rte_flow_item_eth lacp_mask = {
+		.type = 0xffff,
+	};
+	const struct rte_flow_attr attr = {
+		.ingress = 1,
+	};
+	struct rte_flow_item items[] = {
+		{
+			.type = RTE_FLOW_ITEM_TYPE_ETH,
+			.spec = &lacp_spec,
+			.mask = &lacp_mask,
+		},
+		{
+			.type = RTE_FLOW_ITEM_TYPE_END,
+		},
+	};
+	struct rte_flow_action actions[] = {
+		{
+			.type = MLX5_RTE_FLOW_ACTION_TYPE_DEFAULT_MISS,
+		},
+		{
+			.type = RTE_FLOW_ACTION_TYPE_END,
+		},
+	};
+	struct rte_flow_error error;
+	uint32_t flow_idx = flow_list_create(dev, &priv->ctrl_flows,
+				&attr, items, actions, false, &error);
+
+	if (!flow_idx)
+		return -rte_errno;
+	return 0;
+}
+
+/**
  * Destroy a flow.
  *
  * @see rte_flow_destroy()
diff --git a/drivers/net/mlx5/mlx5_flow.h b/drivers/net/mlx5/mlx5_flow.h
index 2c96677..50ec741 100644
--- a/drivers/net/mlx5/mlx5_flow.h
+++ b/drivers/net/mlx5/mlx5_flow.h
@@ -43,6 +43,7 @@ enum mlx5_rte_flow_action_type {
 	MLX5_RTE_FLOW_ACTION_TYPE_TAG,
 	MLX5_RTE_FLOW_ACTION_TYPE_MARK,
 	MLX5_RTE_FLOW_ACTION_TYPE_COPY_MREG,
+	MLX5_RTE_FLOW_ACTION_TYPE_DEFAULT_MISS,
 };
 
 /* Matches on selected register. */
@@ -200,10 +201,12 @@ enum mlx5_feature_name {
 #define MLX5_FLOW_ACTION_SET_IPV4_DSCP (1ull << 32)
 #define MLX5_FLOW_ACTION_SET_IPV6_DSCP (1ull << 33)
 #define MLX5_FLOW_ACTION_AGE (1ull << 34)
+#define MLX5_FLOW_ACTION_DEFAULT_MISS (1ull << 35)
 
 #define MLX5_FLOW_FATE_ACTIONS \
 	(MLX5_FLOW_ACTION_DROP | MLX5_FLOW_ACTION_QUEUE | \
-	 MLX5_FLOW_ACTION_RSS | MLX5_FLOW_ACTION_JUMP)
+	 MLX5_FLOW_ACTION_RSS | MLX5_FLOW_ACTION_JUMP | \
+	 MLX5_FLOW_ACTION_DEFAULT_MISS)
 
 #define MLX5_FLOW_FATE_ESWITCH_ACTIONS \
 	(MLX5_FLOW_ACTION_DROP | MLX5_FLOW_ACTION_PORT_ID | \
@@ -364,6 +367,7 @@ enum mlx5_flow_fate_type {
 	MLX5_FLOW_FATE_JUMP,
 	MLX5_FLOW_FATE_PORT_ID,
 	MLX5_FLOW_FATE_DROP,
+	MLX5_FLOW_FATE_DEFAULT_MISS,
 	MLX5_FLOW_FATE_MAX,
 };
 
@@ -545,6 +549,8 @@ struct mlx5_flow_handle {
 		/**< Index to port ID action resource. */
 		uint32_t rix_fate;
 		/**< Generic value indicates the fate action. */
+		uint32_t rix_default_fate;
+		/**< Indicates default miss fate action. */
 	};
 #ifdef HAVE_IBV_FLOW_DV_SUPPORT
 	struct mlx5_flow_handle_dv dvh;
@@ -946,6 +952,9 @@ int mlx5_flow_validate_action_rss(const struct rte_flow_action *action,
 				  const struct rte_flow_attr *attr,
 				  uint64_t item_flags,
 				  struct rte_flow_error *error);
+int mlx5_flow_validate_action_default_miss(uint64_t action_flags,
+				const struct rte_flow_attr *attr,
+				struct rte_flow_error *error);
 int mlx5_flow_validate_attributes(struct rte_eth_dev *dev,
 				  const struct rte_flow_attr *attributes,
 				  struct rte_flow_error *error);
diff --git a/drivers/net/mlx5/mlx5_flow_dv.c b/drivers/net/mlx5/mlx5_flow_dv.c
index f174009..d1eb65b 100644
--- a/drivers/net/mlx5/mlx5_flow_dv.c
+++ b/drivers/net/mlx5/mlx5_flow_dv.c
@@ -79,6 +79,9 @@
 flow_dv_tbl_resource_release(struct rte_eth_dev *dev,
 			     struct mlx5_flow_tbl_resource *tbl);
 
+static int
+flow_dv_default_miss_resource_release(struct rte_eth_dev *dev);
+
 /**
  * Initialize flow attributes structure according to flow items' types.
  *
@@ -2676,6 +2679,42 @@ struct field_modify_info modify_tcp[] = {
 }
 
 /**
+ * Find existing default miss resource or create and register a new one.
+ *
+ * @param[in, out] dev
+ *   Pointer to rte_eth_dev structure.
+ * @param[out] error
+ *   pointer to error structure.
+ *
+ * @return
+ *   0 on success otherwise -errno and errno is set.
+ */
+static int
+flow_dv_default_miss_resource_register(struct rte_eth_dev *dev,
+		struct rte_flow_error *error)
+{
+	struct mlx5_priv *priv = dev->data->dev_private;
+	struct mlx5_dev_ctx_shared *sh = priv->sh;
+	struct mlx5_flow_default_miss_resource *cache_resource =
+			&sh->default_miss;
+	int cnt = rte_atomic32_read(&cache_resource->refcnt);
+
+	if (!cnt) {
+		MLX5_ASSERT(cache_resource->action);
+		cache_resource->action =
+		mlx5_glue->dr_create_flow_action_default_miss();
+		if (!cache_resource->action)
+			return rte_flow_error_set(error, ENOMEM,
+					RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+					"cannot create default miss action");
+		DRV_LOG(DEBUG, "new default miss resource %p: refcnt %d++",
+				(void *)cache_resource->action, cnt);
+	}
+	rte_atomic32_inc(&cache_resource->refcnt);
+	return 0;
+}
+
+/**
  * Find existing table port ID resource or create and register a new one.
  *
  * @param[in, out] dev
@@ -5249,6 +5288,15 @@ struct field_modify_info modify_tcp[] = {
 			action_flags |= MLX5_FLOW_ACTION_RSS;
 			++actions_n;
 			break;
+		case MLX5_RTE_FLOW_ACTION_TYPE_DEFAULT_MISS:
+			ret =
+			mlx5_flow_validate_action_default_miss(action_flags,
+					attr, error);
+			if (ret < 0)
+				return ret;
+			action_flags |= MLX5_FLOW_ACTION_DEFAULT_MISS;
+			++actions_n;
+			break;
 		case RTE_FLOW_ACTION_TYPE_COUNT:
 			ret = flow_dv_validate_action_count(dev, error);
 			if (ret < 0)
@@ -8212,6 +8260,11 @@ struct field_modify_info modify_tcp[] = {
 				return -rte_errno;
 			action_flags |= MLX5_FLOW_ACTION_SET_TAG;
 			break;
+		case MLX5_RTE_FLOW_ACTION_TYPE_DEFAULT_MISS:
+			action_flags |= MLX5_FLOW_ACTION_DEFAULT_MISS;
+			dev_flow->handle->fate_action =
+					MLX5_FLOW_FATE_DEFAULT_MISS;
+			break;
 		case RTE_FLOW_ACTION_TYPE_METER:
 			mtr = actions->conf;
 			if (!flow->meter) {
@@ -8304,7 +8357,11 @@ struct field_modify_info modify_tcp[] = {
 			flow_dv_translate_item_eth(match_mask, match_value,
 						   items, tunnel,
 						   dev_flow->dv.group);
-			matcher.priority = MLX5_PRIORITY_MAP_L2;
+			matcher.priority = action_flags &
+					MLX5_FLOW_ACTION_DEFAULT_MISS &&
+					!dev_flow->external ?
+					MLX5_PRIORITY_MAP_L3 :
+					MLX5_PRIORITY_MAP_L2;
 			last_item = tunnel ? MLX5_FLOW_LAYER_INNER_L2 :
 					     MLX5_FLOW_LAYER_OUTER_L2;
 			break;
@@ -8604,7 +8661,19 @@ struct field_modify_info modify_tcp[] = {
 			}
 			dh->rix_hrxq = hrxq_idx;
 			dv->actions[n++] = hrxq->action;
+		} else if (dh->fate_action == MLX5_FLOW_FATE_DEFAULT_MISS) {
+			if (flow_dv_default_miss_resource_register
+					(dev, error)) {
+				rte_flow_error_set
+					(error, rte_errno,
+					 RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
+					 "cannot create default miss resource");
+				goto error_default_miss;
+			}
+			dh->rix_default_fate =  MLX5_FLOW_FATE_DEFAULT_MISS;
+			dv->actions[n++] = priv->sh->default_miss.action;
 		}
+
 		dh->ib_flow =
 			mlx5_glue->dv_create_flow(dv_h->matcher->matcher_object,
 						  (void *)&dv->value, n,
@@ -8629,6 +8698,9 @@ struct field_modify_info modify_tcp[] = {
 	}
 	return 0;
 error:
+	if (dh->fate_action == MLX5_FLOW_FATE_DEFAULT_MISS)
+		flow_dv_default_miss_resource_release(dev);
+error_default_miss:
 	err = rte_errno; /* Save rte_errno before cleanup. */
 	SILIST_FOREACH(priv->sh->ipool[MLX5_IPOOL_MLX5_FLOW], flow->dev_handles,
 		       handle_idx, dh, next) {
@@ -8766,6 +8838,36 @@ struct field_modify_info modify_tcp[] = {
 }
 
 /**
+ * Release a default miss resource.
+ *
+ * @param dev
+ *   Pointer to Ethernet device.
+ * @return
+ *   1 while a reference on it exists, 0 when freed.
+ */
+static int
+flow_dv_default_miss_resource_release(struct rte_eth_dev *dev)
+{
+	struct mlx5_priv *priv = dev->data->dev_private;
+	struct mlx5_dev_ctx_shared *sh = priv->sh;
+	struct mlx5_flow_default_miss_resource *cache_resource =
+			&sh->default_miss;
+
+	MLX5_ASSERT(cache_resource->action);
+	DRV_LOG(DEBUG, "default miss resource %p: refcnt %d--",
+			(void *)cache_resource->action,
+			rte_atomic32_read(&cache_resource->refcnt));
+	if (rte_atomic32_dec_and_test(&cache_resource->refcnt)) {
+		claim_zero(mlx5_glue->destroy_flow_action
+				(cache_resource->action));
+		DRV_LOG(DEBUG, "default miss resource %p: removed",
+				(void *)cache_resource->action);
+		return 0;
+	}
+	return 1;
+}
+
+/**
  * Release a modify-header resource.
  *
  * @param handle
@@ -8892,16 +8994,26 @@ struct field_modify_info modify_tcp[] = {
 {
 	if (!handle->rix_fate)
 		return;
-	if (handle->fate_action == MLX5_FLOW_FATE_DROP)
+	switch (handle->fate_action) {
+	case MLX5_FLOW_FATE_DROP:
 		mlx5_hrxq_drop_release(dev);
-	else if (handle->fate_action == MLX5_FLOW_FATE_QUEUE)
+		break;
+	case MLX5_FLOW_FATE_QUEUE:
 		mlx5_hrxq_release(dev, handle->rix_hrxq);
-	else if (handle->fate_action == MLX5_FLOW_FATE_JUMP)
+		break;
+	case MLX5_FLOW_FATE_JUMP:
 		flow_dv_jump_tbl_resource_release(dev, handle);
-	else if (handle->fate_action == MLX5_FLOW_FATE_PORT_ID)
+		break;
+	case MLX5_FLOW_FATE_PORT_ID:
 		flow_dv_port_id_action_resource_release(dev, handle);
-	else
+		break;
+	case MLX5_FLOW_FATE_DEFAULT_MISS:
+		flow_dv_default_miss_resource_release(dev);
+		break;
+	default:
 		DRV_LOG(DEBUG, "Incorrect fate action:%d", handle->fate_action);
+		break;
+	}
 	handle->rix_fate = 0;
 }
 
@@ -8934,7 +9046,8 @@ struct field_modify_info modify_tcp[] = {
 			dh->ib_flow = NULL;
 		}
 		if (dh->fate_action == MLX5_FLOW_FATE_DROP ||
-		    dh->fate_action == MLX5_FLOW_FATE_QUEUE)
+		    dh->fate_action == MLX5_FLOW_FATE_QUEUE ||
+		    dh->fate_action == MLX5_FLOW_FATE_DEFAULT_MISS)
 			flow_dv_fate_resource_release(dev, dh);
 		if (dh->vf_vlan.tag && dh->vf_vlan.created)
 			mlx5_vlan_vmwa_release(dev, &dh->vf_vlan);
-- 
1.8.3.1


^ permalink raw reply	[flat|nested] 5+ messages in thread

* [dpdk-dev] [PATCH 3/3] net/mlx5: add new devarg for LACP packets control
  2020-06-23  8:41 [dpdk-dev] [PATCH 0/3] mlx5: LACP handling for bonded devices Shiri Kuzin
  2020-06-23  8:41 ` [dpdk-dev] [PATCH 1/3] common/mlx5: add default miss action Shiri Kuzin
  2020-06-23  8:41 ` [dpdk-dev] [PATCH 2/3] net/mlx5: add default miss action to flow engine Shiri Kuzin
@ 2020-06-23  8:41 ` Shiri Kuzin
  2020-06-28 10:19 ` [dpdk-dev] [PATCH 0/3] mlx5: LACP handling for bonded devices Raslan Darawsheh
  3 siblings, 0 replies; 5+ messages in thread
From: Shiri Kuzin @ 2020-06-23  8:41 UTC (permalink / raw)
  To: dev; +Cc: matan, viacheslavo

The new devarg will control the steering of the lacp traffic.
When setting dv_lacp_by_user = 0 the lacp traffic will be
steered to kernel and managed there.

When setting dv_lacp_by_user = 1 the lacp traffic will
not be steered and the user will need to manage it.

Signed-off-by: Shiri Kuzin <shirik@mellanox.com>
Acked-by: Matan Azrad <matan@mellanox.com>
---
 doc/guides/nics/mlx5.rst               | 10 ++++++++++
 doc/guides/rel_notes/release_20_08.rst |  1 +
 drivers/net/mlx5/mlx5.c                |  6 ++++++
 drivers/net/mlx5/mlx5.h                |  2 ++
 drivers/net/mlx5/mlx5_trigger.c        |  9 +++++++++
 5 files changed, 28 insertions(+)

diff --git a/doc/guides/nics/mlx5.rst b/doc/guides/nics/mlx5.rst
index 0ff3c53..b51aa67 100644
--- a/doc/guides/nics/mlx5.rst
+++ b/doc/guides/nics/mlx5.rst
@@ -796,6 +796,16 @@ Driver options
 
   Enabled by default if supported.
 
+- ``lacp_by_user`` parameter [int]
+
+  A nonzero value enables the control of LACP traffic by the user application.
+  When a bond exists in the driver, by default it should be managed by the
+  kernel and therefore LACP traffic should be steered to the kernel.
+  If this devarg is set to 1 it will allow the user to manage the bond by
+  itself and not steer LACP traffic to the kernel.
+
+  Disabled by default (set to 0).
+
 - ``mr_ext_memseg_en`` parameter [int]
 
   A nonzero value enables extending memseg when registering DMA memory. If
diff --git a/doc/guides/rel_notes/release_20_08.rst b/doc/guides/rel_notes/release_20_08.rst
index cc39984..f03762b 100644
--- a/doc/guides/rel_notes/release_20_08.rst
+++ b/doc/guides/rel_notes/release_20_08.rst
@@ -67,6 +67,7 @@ New Features
   Updated Mellanox mlx5 driver with new features and improvements, including:
 
   * Added new PMD devarg ``reclaim_mem_mode``.
+  * Added new devarg ``lacp_by_user``.
 
 * **Added vDPA device APIs to query virtio queue statistics.**
 
diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index ddbe29d..07c6add 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -140,6 +140,9 @@
 /* Enable extensive flow metadata support. */
 #define MLX5_DV_XMETA_EN "dv_xmeta_en"
 
+/* Device parameter to let the user manage the lacp traffic of bonded device */
+#define MLX5_LACP_BY_USER "lacp_by_user"
+
 /* Activate Netlink support in VF mode. */
 #define MLX5_VF_NL_EN "vf_nl_en"
 
@@ -1352,6 +1355,8 @@ struct mlx5_dev_ctx_shared *
 			return -rte_errno;
 		}
 		config->dv_xmeta_en = tmp;
+	} else if (strcmp(MLX5_LACP_BY_USER, key) == 0) {
+		config->lacp_by_user = !!tmp;
 	} else if (strcmp(MLX5_MR_EXT_MEMSEG_EN, key) == 0) {
 		config->mr_ext_memseg_en = !!tmp;
 	} else if (strcmp(MLX5_MAX_DUMP_FILES_NUM, key) == 0) {
@@ -1419,6 +1424,7 @@ struct mlx5_dev_ctx_shared *
 		MLX5_DV_ESW_EN,
 		MLX5_DV_FLOW_EN,
 		MLX5_DV_XMETA_EN,
+		MLX5_LACP_BY_USER,
 		MLX5_MR_EXT_MEMSEG_EN,
 		MLX5_REPRESENTOR,
 		MLX5_MAX_DUMP_FILES_NUM,
diff --git a/drivers/net/mlx5/mlx5.h b/drivers/net/mlx5/mlx5.h
index 93bde5a..46e66eb 100644
--- a/drivers/net/mlx5/mlx5.h
+++ b/drivers/net/mlx5/mlx5.h
@@ -210,6 +210,8 @@ struct mlx5_dev_config {
 	unsigned int dv_esw_en:1; /* Enable E-Switch DV flow. */
 	unsigned int dv_flow_en:1; /* Enable DV flow. */
 	unsigned int dv_xmeta_en:2; /* Enable extensive flow metadata. */
+	unsigned int lacp_by_user:1;
+	/* Enable user to manage LACP traffic. */
 	unsigned int swp:1; /* Tx generic tunnel checksum and TSO offload. */
 	unsigned int devx:1; /* Whether devx interface is available or not. */
 	unsigned int dest_tir:1; /* Whether advanced DR API is available. */
diff --git a/drivers/net/mlx5/mlx5_trigger.c b/drivers/net/mlx5/mlx5_trigger.c
index c7c2ee6..ef74609 100644
--- a/drivers/net/mlx5/mlx5_trigger.c
+++ b/drivers/net/mlx5/mlx5_trigger.c
@@ -464,6 +464,15 @@
 				" configured - only Eswitch group 0 flows are"
 				" supported.", dev->data->port_id);
 	}
+	if (!priv->config.lacp_by_user && priv->pf_bond >= 0) {
+		ret = mlx5_flow_lacp_miss(dev);
+		if (ret)
+			DRV_LOG(INFO, "port %u LACP rule cannot be created - "
+				"forward LACP to kernel.", dev->data->port_id);
+		else
+			DRV_LOG(INFO, "LACP traffic will be missed in port %u."
+				, dev->data->port_id);
+	}
 	if (priv->isolated)
 		return 0;
 	if (dev->data->promiscuous) {
-- 
1.8.3.1


^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [dpdk-dev] [PATCH 0/3] mlx5: LACP handling for bonded devices
  2020-06-23  8:41 [dpdk-dev] [PATCH 0/3] mlx5: LACP handling for bonded devices Shiri Kuzin
                   ` (2 preceding siblings ...)
  2020-06-23  8:41 ` [dpdk-dev] [PATCH 3/3] net/mlx5: add new devarg for LACP packets control Shiri Kuzin
@ 2020-06-28 10:19 ` Raslan Darawsheh
  3 siblings, 0 replies; 5+ messages in thread
From: Raslan Darawsheh @ 2020-06-28 10:19 UTC (permalink / raw)
  To: Shiri Kuzin, dev; +Cc: Matan Azrad, Slava Ovsiienko

Hi,

> -----Original Message-----
> From: dev <dev-bounces@dpdk.org> On Behalf Of Shiri Kuzin
> Sent: Tuesday, June 23, 2020 11:41 AM
> To: dev@dpdk.org
> Cc: Matan Azrad <matan@mellanox.com>; Slava Ovsiienko
> <viacheslavo@mellanox.com>
> Subject: [dpdk-dev] [PATCH 0/3] mlx5: LACP handling for bonded devices
> 
> When the driver has a bond configured, LACP traffic should be steered to
> kernel to manage the bond using a new rdma-core API.
> This is done by creating a new internal action and adding a new devarg that
> will allow the user to manage the bond and get lacp traffic if they wish.
> 
> Shiri Kuzin (3):
>   common/mlx5: add default miss action
>   net/mlx5: add default miss action to flow engine
>   net/mlx5: add new devarg for LACP packets control
> 
>  doc/guides/nics/mlx5.rst               |  10 +++
>  doc/guides/rel_notes/release_20_08.rst |   1 +
>  drivers/common/mlx5/Makefile           |   5 ++
>  drivers/common/mlx5/linux/meson.build  |   2 +
>  drivers/common/mlx5/linux/mlx5_glue.c  |  13 ++++
>  drivers/common/mlx5/linux/mlx5_glue.h  |   1 +
>  drivers/net/mlx5/mlx5.c                |   6 ++
>  drivers/net/mlx5/mlx5.h                |  11 +++
>  drivers/net/mlx5/mlx5_flow.c           |  92 ++++++++++++++++++++++++
>  drivers/net/mlx5/mlx5_flow.h           |  11 ++-
>  drivers/net/mlx5/mlx5_flow_dv.c        | 127
> +++++++++++++++++++++++++++++++--
>  drivers/net/mlx5/mlx5_trigger.c        |   9 +++
>  12 files changed, 280 insertions(+), 8 deletions(-)
> 
> --
> 1.8.3.1


Series applied to next-net-mlx,

Kindest regards,
Raslan Darawsheh

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2020-06-28 10:19 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-06-23  8:41 [dpdk-dev] [PATCH 0/3] mlx5: LACP handling for bonded devices Shiri Kuzin
2020-06-23  8:41 ` [dpdk-dev] [PATCH 1/3] common/mlx5: add default miss action Shiri Kuzin
2020-06-23  8:41 ` [dpdk-dev] [PATCH 2/3] net/mlx5: add default miss action to flow engine Shiri Kuzin
2020-06-23  8:41 ` [dpdk-dev] [PATCH 3/3] net/mlx5: add new devarg for LACP packets control Shiri Kuzin
2020-06-28 10:19 ` [dpdk-dev] [PATCH 0/3] mlx5: LACP handling for bonded devices Raslan Darawsheh

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).