[dpdk-dev] [PATCH v12 2/2] i40e: add floating VEB support in i40e

[Zhe Tao <zhe.tao@intel.com>]

This patch add the support for floating VEB in i40e.
All the VFs VSIs can decide whether to connect to the legacy VEB/VEPA or
the floating VEB. When connect to the floating VEB a new floating VEB is
created. Now all the VFs need to connect to floating VEB or legacy VEB,
cannot connect to both of them. The PF and VMDQ,FD VSIs still connect to
the old legacy VEB/VEPA.

All the VEB/VEPA concepts are not specific for FVL, they are defined in
the 802.1Qbg spec.

The floating VEB feature is only available for the FW version which
newer than 5.0 (FW major version number > 5).

To enable this feature, the user should pass a devargs parameter to the
EAL like "-w 84:00.0,enable_floating_veb=1", and the application use
this method to tell PMD to connect all the VFs created by
this PF device to the floating VEB.

Also you can specify which VF need to connect to this floating veb using
"floating_veb_list".
Like "-w 84:00.0,enable_floating_veb=1,floating_veb_list=1/3-4", means VF1, VF3,
VF4 connect to the floating VEB, other VFs connect to the legacy VEB.The "/"
is used for delimiter of the floating VEB list.

Signed-off-by: Zhe Tao <zhe.tao@intel.com>
---
 doc/guides/nics/i40e.rst               |  27 ++++++++
 doc/guides/rel_notes/release_16_07.rst |   4 ++
 drivers/net/i40e/i40e_ethdev.c         | 112 ++++++++++++++++++++++++++-------
 drivers/net/i40e/i40e_ethdev.h         |   2 +
 drivers/net/i40e/i40e_pf.c             |  12 +++-
 5 files changed, 134 insertions(+), 23 deletions(-)

diff --git a/doc/guides/nics/i40e.rst b/doc/guides/nics/i40e.rst
index 934eb02..1ce60ab 100644
--- a/doc/guides/nics/i40e.rst
+++ b/doc/guides/nics/i40e.rst
@@ -366,3 +366,30 @@ Delete all flow director rules on a port:
 
    testpmd> flush_flow_director 0
 
+Floating VEB
+~~~~~~~~~~~~~
+FVL can support floating VEB feature.
+
+The floating VEB means the VEB doesn't has some uplink connection to the outside
+world, so all the switching belong to this VEB cannot go to outside, the
+security can be assured. Because the floating VEB doesn't need to connect to
+the MAC port, so when the physical port link is down, all the switching within
+this VEB still works fine, but for legacy VEB when the physical port is down
+the VEB cannot forward packets anymore.
+
+With this feature, VFs can communicate with each other, but cannot access
+outside network. When PF is down, and VFs can still forward pkts between each
+other.
+
+To enable this feature, the user should pass a devargs parameter to the EAL
+like "-w 84:00.0,enable_floating_veb=1", and then the PMD will use the floating
+VEB feature for all the VFs created by this PF device.
+Also you can specify which VF need to connect to this floating veb using
+"floating_veb_list".
+Like "-w 84:00.0,enable_floating_veb=1,floating_veb_list=1/3-4", means VF1, VF3,
+VF4 connect to the floating VEB, other VFs connect to the legacy VEB.The "/"
+is used for delimiter of the floating VEB list.
+
+The current implementation only support one floating VEB and one legacy
+VEB. VF can connect to floating VEB or legacy VEB according to the
+configuration.
diff --git a/doc/guides/rel_notes/release_16_07.rst b/doc/guides/rel_notes/release_16_07.rst
index 30e78d4..1752c40 100644
--- a/doc/guides/rel_notes/release_16_07.rst
+++ b/doc/guides/rel_notes/release_16_07.rst
@@ -47,6 +47,10 @@ New Features
   * Dropped specific Xen Dom0 code.
   * Dropped specific anonymous mempool code in testpmd.
 
+* **Added floating VEB support for i40e PF driver.**
+
+  More details please see floating VEB part in the document
+  doc/guides/nics/i40e.rst.
 
 Resolved Issues
 ---------------
diff --git a/drivers/net/i40e/i40e_ethdev.c b/drivers/net/i40e/i40e_ethdev.c
index b304fc3..4673619 100644
--- a/drivers/net/i40e/i40e_ethdev.c
+++ b/drivers/net/i40e/i40e_ethdev.c
@@ -3860,21 +3860,27 @@ i40e_veb_release(struct i40e_veb *veb)
 	struct i40e_vsi *vsi;
 	struct i40e_hw *hw;
 
-	if (veb == NULL || veb->associate_vsi == NULL)
+	if (veb == NULL)
 		return -EINVAL;
 
 	if (!TAILQ_EMPTY(&veb->head)) {
 		PMD_DRV_LOG(ERR, "VEB still has VSI attached, can't remove");
 		return -EACCES;
 	}
+	/* associate_vsi field is NULL for floating VEB */
+	if (veb->associate_vsi != NULL) {
+		vsi = veb->associate_vsi;
+		hw = I40E_VSI_TO_HW(vsi);
 
-	vsi = veb->associate_vsi;
-	hw = I40E_VSI_TO_HW(vsi);
+		vsi->uplink_seid = veb->uplink_seid;
+		vsi->veb = NULL;
+	} else {
+		veb->associate_pf->main_vsi->floating_veb = NULL;
+		hw = I40E_VSI_TO_HW(veb->associate_pf->main_vsi);
+	}
 
-	vsi->uplink_seid = veb->uplink_seid;
 	i40e_aq_delete_element(hw, veb->seid, NULL);
 	rte_free(veb);
-	vsi->veb = NULL;
 	return I40E_SUCCESS;
 }
 
@@ -3886,9 +3892,9 @@ i40e_veb_setup(struct i40e_pf *pf, struct i40e_vsi *vsi)
 	int ret;
 	struct i40e_hw *hw;
 
-	if (NULL == pf || vsi == NULL) {
+	if (NULL == pf) {
 		PMD_DRV_LOG(ERR, "veb setup failed, "
-			    "associated VSI shouldn't null");
+			    "associated PF shouldn't null");
 		return NULL;
 	}
 	hw = I40E_PF_TO_HW(pf);
@@ -3900,11 +3906,19 @@ i40e_veb_setup(struct i40e_pf *pf, struct i40e_vsi *vsi)
 	}
 
 	veb->associate_vsi = vsi;
+	veb->associate_pf = pf;
 	TAILQ_INIT(&veb->head);
-	veb->uplink_seid = vsi->uplink_seid;
+	veb->uplink_seid = vsi ? vsi->uplink_seid : 0;
 
-	ret = i40e_aq_add_veb(hw, veb->uplink_seid, vsi->seid,
-		I40E_DEFAULT_TCMAP, false, &veb->seid, false, NULL);
+	/* create floating veb if vsi is NULL */
+	if (vsi != NULL) {
+		ret = i40e_aq_add_veb(hw, veb->uplink_seid, vsi->seid,
+				      I40E_DEFAULT_TCMAP, false,
+				      &veb->seid, false, NULL);
+	} else {
+		ret = i40e_aq_add_veb(hw, 0, 0, I40E_DEFAULT_TCMAP,
+				      true, &veb->seid, false, NULL);
+	}
 
 	if (ret != I40E_SUCCESS) {
 		PMD_DRV_LOG(ERR, "Add veb failed, aq_err: %d",
@@ -3920,10 +3934,10 @@ i40e_veb_setup(struct i40e_pf *pf, struct i40e_vsi *vsi)
 			    hw->aq.asq_last_status);
 		goto fail;
 	}
-
 	/* Get VEB bandwidth, to be implemented */
 	/* Now associated vsi binding to the VEB, set uplink to this VEB */
-	vsi->uplink_seid = veb->seid;
+	if (vsi)
+		vsi->uplink_seid = veb->seid;
 
 	return veb;
 fail:
@@ -3939,6 +3953,7 @@ i40e_vsi_release(struct i40e_vsi *vsi)
 	struct i40e_vsi_list *vsi_list;
 	int ret;
 	struct i40e_mac_filter *f;
+	uint16_t user_param = vsi->user_param;
 
 	if (!vsi)
 		return I40E_SUCCESS;
@@ -3956,12 +3971,22 @@ i40e_vsi_release(struct i40e_vsi *vsi)
 		i40e_veb_release(vsi->veb);
 	}
 
+	if (vsi->floating_veb) {
+		TAILQ_FOREACH(vsi_list, &vsi->floating_veb->head, list) {
+			if (i40e_vsi_release(vsi_list->vsi) != I40E_SUCCESS)
+				return -1;
+			TAILQ_REMOVE(&vsi->floating_veb->head, vsi_list, list);
+		}
+	}
+
 	/* Remove all macvlan filters of the VSI */
 	i40e_vsi_remove_all_macvlan_filter(vsi);
 	TAILQ_FOREACH(f, &vsi->mac_list, next)
 		rte_free(f);
 
-	if (vsi->type != I40E_VSI_MAIN) {
+	if (vsi->type != I40E_VSI_MAIN &&
+	    ((vsi->type != I40E_VSI_SRIOV) ||
+	    !pf->vf_floating_veb[user_param])) {
 		/* Remove vsi from parent's sibling list */
 		if (vsi->parent_vsi == NULL || vsi->parent_vsi->veb == NULL) {
 			PMD_DRV_LOG(ERR, "VSI's parent VSI is NULL");
@@ -3975,6 +4000,24 @@ i40e_vsi_release(struct i40e_vsi *vsi)
 		if (ret != I40E_SUCCESS)
 			PMD_DRV_LOG(ERR, "Failed to delete element");
 	}
+
+	if ((vsi->type == I40E_VSI_SRIOV) &&
+	    pf->vf_floating_veb[user_param]) {
+		/* Remove vsi from parent's sibling list */
+		if (vsi->parent_vsi == NULL ||
+		    vsi->parent_vsi->floating_veb == NULL) {
+			PMD_DRV_LOG(ERR, "VSI's parent VSI is NULL");
+			return I40E_ERR_PARAM;
+		}
+		TAILQ_REMOVE(&vsi->parent_vsi->floating_veb->head,
+			     &vsi->sib_vsi_list, list);
+
+		/* Remove all switch element of the VSI */
+		ret = i40e_aq_delete_element(hw, vsi->seid, NULL);
+		if (ret != I40E_SUCCESS)
+			PMD_DRV_LOG(ERR, "Failed to delete element");
+	}
+
 	i40e_res_pool_free(&pf->qp_pool, vsi->base_queue);
 
 	if (vsi->type != I40E_VSI_SRIOV)
@@ -4143,7 +4186,8 @@ i40e_vsi_setup(struct i40e_pf *pf,
 	struct ether_addr broadcast =
 		{.addr_bytes = {0xff, 0xff, 0xff, 0xff, 0xff, 0xff}};
 
-	if (type != I40E_VSI_MAIN && uplink_vsi == NULL) {
+	if (type != I40E_VSI_MAIN && type != I40E_VSI_SRIOV &&
+	    uplink_vsi == NULL) {
 		PMD_DRV_LOG(ERR, "VSI setup failed, "
 			    "VSI link shouldn't be NULL");
 		return NULL;
@@ -4155,11 +4199,18 @@ i40e_vsi_setup(struct i40e_pf *pf,
 		return NULL;
 	}
 
-	/* If uplink vsi didn't setup VEB, create one first */
-	if (type != I40E_VSI_MAIN && uplink_vsi->veb == NULL) {
+	/* two situations
+	 * 1.type is not MAIN and uplink vsi is not NULL
+	 * If uplink vsi didn't setup VEB, create one first under veb field
+	 * 2.type is SRIOV and the uplink is NULL
+	 * If floating VEB is NULL, create one veb under floating veb field
+	 */
+
+	if (type != I40E_VSI_MAIN && uplink_vsi != NULL &&
+	    uplink_vsi->veb == NULL) {
 		uplink_vsi->veb = i40e_veb_setup(pf, uplink_vsi);
 
-		if (NULL == uplink_vsi->veb) {
+		if (uplink_vsi->veb == NULL) {
 			PMD_DRV_LOG(ERR, "VEB setup failed");
 			return NULL;
 		}
@@ -4167,6 +4218,16 @@ i40e_vsi_setup(struct i40e_pf *pf,
 		i40e_enable_pf_lb(pf);
 	}
 
+	if (type == I40E_VSI_SRIOV && uplink_vsi == NULL &&
+	    pf->main_vsi->floating_veb == NULL) {
+		pf->main_vsi->floating_veb = i40e_veb_setup(pf, uplink_vsi);
+
+		if (pf->main_vsi->floating_veb == NULL) {
+			PMD_DRV_LOG(ERR, "VEB setup failed");
+			return NULL;
+		}
+	}
+
 	vsi = rte_zmalloc("i40e_vsi", sizeof(struct i40e_vsi), 0);
 	if (!vsi) {
 		PMD_DRV_LOG(ERR, "Failed to allocate memory for vsi");
@@ -4176,7 +4237,7 @@ i40e_vsi_setup(struct i40e_pf *pf,
 	vsi->type = type;
 	vsi->adapter = I40E_PF_TO_ADAPTER(pf);
 	vsi->max_macaddrs = I40E_NUM_MACADDR_MAX;
-	vsi->parent_vsi = uplink_vsi;
+	vsi->parent_vsi = uplink_vsi ? uplink_vsi : pf->main_vsi;
 	vsi->user_param = user_param;
 	/* Allocate queues */
 	switch (vsi->type) {
@@ -4330,7 +4391,11 @@ i40e_vsi_setup(struct i40e_pf *pf,
 		 * For other VSI, the uplink_seid equals to uplink VSI's
 		 * uplink_seid since they share same VEB
 		 */
-		vsi->uplink_seid = uplink_vsi->uplink_seid;
+		if (uplink_vsi == NULL) {
+			vsi->uplink_seid = pf->main_vsi->floating_veb->seid;
+		} else {
+			vsi->uplink_seid = uplink_vsi->uplink_seid;
+		}
 		ctxt.pf_num = hw->pf_id;
 		ctxt.vf_num = hw->func_caps.vf_base_id + user_param;
 		ctxt.uplink_seid = vsi->uplink_seid;
@@ -4438,8 +4503,13 @@ i40e_vsi_setup(struct i40e_pf *pf,
 		vsi->seid = ctxt.seid;
 		vsi->vsi_id = ctxt.vsi_number;
 		vsi->sib_vsi_list.vsi = vsi;
-		TAILQ_INSERT_TAIL(&uplink_vsi->veb->head,
-				&vsi->sib_vsi_list, list);
+		if (vsi->type == I40E_VSI_SRIOV && uplink_vsi == NULL) {
+			TAILQ_INSERT_TAIL(&pf->main_vsi->floating_veb->head,
+					  &vsi->sib_vsi_list, list);
+		} else {
+			TAILQ_INSERT_TAIL(&uplink_vsi->veb->head,
+					  &vsi->sib_vsi_list, list);
+		}
 	}
 
 	/* MAC/VLAN configuration */
diff --git a/drivers/net/i40e/i40e_ethdev.h b/drivers/net/i40e/i40e_ethdev.h
index 71bca65..d8a7832 100644
--- a/drivers/net/i40e/i40e_ethdev.h
+++ b/drivers/net/i40e/i40e_ethdev.h
@@ -224,6 +224,7 @@ struct i40e_bw_info {
 struct i40e_veb {
 	struct i40e_vsi_list_head head;
 	struct i40e_vsi *associate_vsi; /* Associate VSI who owns the VEB */
+	struct i40e_pf *associate_pf; /* Associate PF who owns the VEB */
 	uint16_t seid; /* The seid of VEB itself */
 	uint16_t uplink_seid; /* The uplink seid of this VEB */
 	uint16_t stats_idx;
@@ -264,6 +265,7 @@ struct i40e_vsi {
 	struct i40e_vsi_list sib_vsi_list; /* sibling vsi list */
 	struct i40e_vsi *parent_vsi;
 	struct i40e_veb *veb;    /* Associated veb, could be null */
+	struct i40e_veb *floating_veb; /* Associated floating veb */
 	bool offset_loaded;
 	enum i40e_vsi_type type; /* VSI types */
 	uint16_t vlan_num;       /* Total VLAN number */
diff --git a/drivers/net/i40e/i40e_pf.c b/drivers/net/i40e/i40e_pf.c
index 5afd61a..ac27070 100644
--- a/drivers/net/i40e/i40e_pf.c
+++ b/drivers/net/i40e/i40e_pf.c
@@ -124,6 +124,7 @@ i40e_pf_host_vf_reset(struct i40e_pf_vf *vf, bool do_hw_reset)
 {
 	uint32_t val, i;
 	struct i40e_hw *hw = I40E_PF_TO_HW(vf->pf);
+	struct i40e_pf *pf = vf->pf;
 	uint16_t vf_id, abs_vf_id, vf_msix_num;
 	int ret;
 	struct i40e_virtchnl_queue_select qsel;
@@ -223,9 +224,16 @@ i40e_pf_host_vf_reset(struct i40e_pf_vf *vf, bool do_hw_reset)
 	vf->reset_cnt++;
 	I40E_WRITE_FLUSH(hw);
 
+	if (pf->floating_veb == true &&
+	    pf->vf_floating_veb[vf_id]) {
+		vf->vsi = i40e_vsi_setup(vf->pf, I40E_VSI_SRIOV,
+			NULL, vf->vf_idx);
+	} else {
 	/* Allocate resource again */
-	vf->vsi = i40e_vsi_setup(vf->pf, I40E_VSI_SRIOV,
-			vf->pf->main_vsi, vf->vf_idx);
+		vf->vsi = i40e_vsi_setup(vf->pf, I40E_VSI_SRIOV,
+				vf->pf->main_vsi, vf->vf_idx);
+	}
+
 	if (vf->vsi == NULL) {
 		PMD_DRV_LOG(ERR, "Add vsi failed");
 		return -EFAULT;
-- 
2.1.4