net/ice/base: add virtual switch code
authorPaul M Stillwell Jr <paul.m.stillwell.jr@intel.com>
Tue, 18 Dec 2018 08:46:17 +0000 (16:46 +0800)
committerFerruh Yigit <ferruh.yigit@intel.com>
Fri, 21 Dec 2018 15:22:40 +0000 (16:22 +0100)
Add code to handle the virtual switch within the NIC.

Signed-off-by: Paul M Stillwell Jr <paul.m.stillwell.jr@intel.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
drivers/net/ice/base/ice_switch.c [new file with mode: 0644]
drivers/net/ice/base/ice_switch.h [new file with mode: 0644]

diff --git a/drivers/net/ice/base/ice_switch.c b/drivers/net/ice/base/ice_switch.c
new file mode 100644 (file)
index 0000000..0379cd0
--- /dev/null
@@ -0,0 +1,2812 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2001-2018
+ */
+
+#include "ice_switch.h"
+
+
+#define ICE_ETH_DA_OFFSET              0
+#define ICE_ETH_ETHTYPE_OFFSET         12
+#define ICE_ETH_VLAN_TCI_OFFSET                14
+#define ICE_MAX_VLAN_ID                        0xFFF
+
+/* Dummy ethernet header needed in the ice_aqc_sw_rules_elem
+ * struct to configure any switch filter rules.
+ * {DA (6 bytes), SA(6 bytes),
+ * Ether type (2 bytes for header without VLAN tag) OR
+ * VLAN tag (4 bytes for header with VLAN tag) }
+ *
+ * Word on Hardcoded values
+ * byte 0 = 0x2: to identify it as locally administered DA MAC
+ * byte 6 = 0x2: to identify it as locally administered SA MAC
+ * byte 12 = 0x81 & byte 13 = 0x00:
+ *     In case of VLAN filter first two bytes defines ether type (0x8100)
+ *     and remaining two bytes are placeholder for programming a given VLAN id
+ *     In case of Ether type filter it is treated as header without VLAN tag
+ *     and byte 12 and 13 is used to program a given Ether type instead
+ */
+#define DUMMY_ETH_HDR_LEN              16
+static const u8 dummy_eth_header[DUMMY_ETH_HDR_LEN] = { 0x2, 0, 0, 0, 0, 0,
+                                                       0x2, 0, 0, 0, 0, 0,
+                                                       0x81, 0, 0, 0};
+
+#define ICE_SW_RULE_RX_TX_ETH_HDR_SIZE \
+       (sizeof(struct ice_aqc_sw_rules_elem) - \
+        sizeof(((struct ice_aqc_sw_rules_elem *)0)->pdata) + \
+        sizeof(struct ice_sw_rule_lkup_rx_tx) + DUMMY_ETH_HDR_LEN - 1)
+#define ICE_SW_RULE_RX_TX_NO_HDR_SIZE \
+       (sizeof(struct ice_aqc_sw_rules_elem) - \
+        sizeof(((struct ice_aqc_sw_rules_elem *)0)->pdata) + \
+        sizeof(struct ice_sw_rule_lkup_rx_tx) - 1)
+#define ICE_SW_RULE_LG_ACT_SIZE(n) \
+       (sizeof(struct ice_aqc_sw_rules_elem) - \
+        sizeof(((struct ice_aqc_sw_rules_elem *)0)->pdata) + \
+        sizeof(struct ice_sw_rule_lg_act) - \
+        sizeof(((struct ice_sw_rule_lg_act *)0)->act) + \
+        ((n) * sizeof(((struct ice_sw_rule_lg_act *)0)->act)))
+#define ICE_SW_RULE_VSI_LIST_SIZE(n) \
+       (sizeof(struct ice_aqc_sw_rules_elem) - \
+        sizeof(((struct ice_aqc_sw_rules_elem *)0)->pdata) + \
+        sizeof(struct ice_sw_rule_vsi_list) - \
+        sizeof(((struct ice_sw_rule_vsi_list *)0)->vsi) + \
+        ((n) * sizeof(((struct ice_sw_rule_vsi_list *)0)->vsi)))
+
+
+/**
+ * ice_init_def_sw_recp - initialize the recipe book keeping tables
+ * @hw: pointer to the hw struct
+ *
+ * Allocate memory for the entire recipe table and initialize the structures/
+ * entries corresponding to basic recipes.
+ */
+enum ice_status ice_init_def_sw_recp(struct ice_hw *hw)
+{
+       struct ice_sw_recipe *recps;
+       u8 i;
+
+       recps = (struct ice_sw_recipe *)
+               ice_calloc(hw, ICE_MAX_NUM_RECIPES, sizeof(*recps));
+       if (!recps)
+               return ICE_ERR_NO_MEMORY;
+
+       for (i = 0; i < ICE_MAX_NUM_RECIPES; i++) {
+               recps[i].root_rid = i;
+               INIT_LIST_HEAD(&recps[i].filt_rules);
+               INIT_LIST_HEAD(&recps[i].filt_replay_rules);
+               ice_init_lock(&recps[i].filt_rule_lock);
+       }
+
+       hw->switch_info->recp_list = recps;
+
+       return ICE_SUCCESS;
+}
+
+/**
+ * ice_aq_get_sw_cfg - get switch configuration
+ * @hw: pointer to the hardware structure
+ * @buf: pointer to the result buffer
+ * @buf_size: length of the buffer available for response
+ * @req_desc: pointer to requested descriptor
+ * @num_elems: pointer to number of elements
+ * @cd: pointer to command details structure or NULL
+ *
+ * Get switch configuration (0x0200) to be placed in 'buff'.
+ * This admin command returns information such as initial VSI/port number
+ * and switch ID it belongs to.
+ *
+ * NOTE: *req_desc is both an input/output parameter.
+ * The caller of this function first calls this function with *request_desc set
+ * to 0. If the response from f/w has *req_desc set to 0, all the switch
+ * configuration information has been returned; if non-zero (meaning not all
+ * the information was returned), the caller should call this function again
+ * with *req_desc set to the previous value returned by f/w to get the
+ * next block of switch configuration information.
+ *
+ * *num_elems is output only parameter. This reflects the number of elements
+ * in response buffer. The caller of this function to use *num_elems while
+ * parsing the response buffer.
+ */
+static enum ice_status
+ice_aq_get_sw_cfg(struct ice_hw *hw, struct ice_aqc_get_sw_cfg_resp *buf,
+                 u16 buf_size, u16 *req_desc, u16 *num_elems,
+                 struct ice_sq_cd *cd)
+{
+       struct ice_aqc_get_sw_cfg *cmd;
+       enum ice_status status;
+       struct ice_aq_desc desc;
+
+       ice_fill_dflt_direct_cmd_desc(&desc, ice_aqc_opc_get_sw_cfg);
+       cmd = &desc.params.get_sw_conf;
+       cmd->element = CPU_TO_LE16(*req_desc);
+
+       status = ice_aq_send_cmd(hw, &desc, buf, buf_size, cd);
+       if (!status) {
+               *req_desc = LE16_TO_CPU(cmd->element);
+               *num_elems = LE16_TO_CPU(cmd->num_elems);
+       }
+
+       return status;
+}
+
+
+
+/**
+ * ice_aq_add_vsi
+ * @hw: pointer to the hw struct
+ * @vsi_ctx: pointer to a VSI context struct
+ * @cd: pointer to command details structure or NULL
+ *
+ * Add a VSI context to the hardware (0x0210)
+ */
+static enum ice_status
+ice_aq_add_vsi(struct ice_hw *hw, struct ice_vsi_ctx *vsi_ctx,
+              struct ice_sq_cd *cd)
+{
+       struct ice_aqc_add_update_free_vsi_resp *res;
+       struct ice_aqc_add_get_update_free_vsi *cmd;
+       struct ice_aq_desc desc;
+       enum ice_status status;
+
+       cmd = &desc.params.vsi_cmd;
+       res = &desc.params.add_update_free_vsi_res;
+
+       ice_fill_dflt_direct_cmd_desc(&desc, ice_aqc_opc_add_vsi);
+
+       if (!vsi_ctx->alloc_from_pool)
+               cmd->vsi_num = CPU_TO_LE16(vsi_ctx->vsi_num |
+                                          ICE_AQ_VSI_IS_VALID);
+
+       cmd->vsi_flags = CPU_TO_LE16(vsi_ctx->flags);
+
+       desc.flags |= CPU_TO_LE16(ICE_AQ_FLAG_RD);
+
+       status = ice_aq_send_cmd(hw, &desc, &vsi_ctx->info,
+                                sizeof(vsi_ctx->info), cd);
+
+       if (!status) {
+               vsi_ctx->vsi_num = LE16_TO_CPU(res->vsi_num) & ICE_AQ_VSI_NUM_M;
+               vsi_ctx->vsis_allocd = LE16_TO_CPU(res->vsi_used);
+               vsi_ctx->vsis_unallocated = LE16_TO_CPU(res->vsi_free);
+       }
+
+       return status;
+}
+
+/**
+ * ice_aq_free_vsi
+ * @hw: pointer to the hw struct
+ * @vsi_ctx: pointer to a VSI context struct
+ * @keep_vsi_alloc: keep VSI allocation as part of this PF's resources
+ * @cd: pointer to command details structure or NULL
+ *
+ * Free VSI context info from hardware (0x0213)
+ */
+static enum ice_status
+ice_aq_free_vsi(struct ice_hw *hw, struct ice_vsi_ctx *vsi_ctx,
+               bool keep_vsi_alloc, struct ice_sq_cd *cd)
+{
+       struct ice_aqc_add_update_free_vsi_resp *resp;
+       struct ice_aqc_add_get_update_free_vsi *cmd;
+       struct ice_aq_desc desc;
+       enum ice_status status;
+
+       cmd = &desc.params.vsi_cmd;
+       resp = &desc.params.add_update_free_vsi_res;
+
+       ice_fill_dflt_direct_cmd_desc(&desc, ice_aqc_opc_free_vsi);
+
+       cmd->vsi_num = CPU_TO_LE16(vsi_ctx->vsi_num | ICE_AQ_VSI_IS_VALID);
+       if (keep_vsi_alloc)
+               cmd->cmd_flags = CPU_TO_LE16(ICE_AQ_VSI_KEEP_ALLOC);
+
+       status = ice_aq_send_cmd(hw, &desc, NULL, 0, cd);
+       if (!status) {
+               vsi_ctx->vsis_allocd = LE16_TO_CPU(resp->vsi_used);
+               vsi_ctx->vsis_unallocated = LE16_TO_CPU(resp->vsi_free);
+       }
+
+       return status;
+}
+
+/**
+ * ice_aq_update_vsi
+ * @hw: pointer to the hw struct
+ * @vsi_ctx: pointer to a VSI context struct
+ * @cd: pointer to command details structure or NULL
+ *
+ * Update VSI context in the hardware (0x0211)
+ */
+static enum ice_status
+ice_aq_update_vsi(struct ice_hw *hw, struct ice_vsi_ctx *vsi_ctx,
+                 struct ice_sq_cd *cd)
+{
+       struct ice_aqc_add_update_free_vsi_resp *resp;
+       struct ice_aqc_add_get_update_free_vsi *cmd;
+       struct ice_aq_desc desc;
+       enum ice_status status;
+
+       cmd = &desc.params.vsi_cmd;
+       resp = &desc.params.add_update_free_vsi_res;
+
+       ice_fill_dflt_direct_cmd_desc(&desc, ice_aqc_opc_update_vsi);
+
+       cmd->vsi_num = CPU_TO_LE16(vsi_ctx->vsi_num | ICE_AQ_VSI_IS_VALID);
+
+       desc.flags |= CPU_TO_LE16(ICE_AQ_FLAG_RD);
+
+       status = ice_aq_send_cmd(hw, &desc, &vsi_ctx->info,
+                                sizeof(vsi_ctx->info), cd);
+
+       if (!status) {
+               vsi_ctx->vsis_allocd = LE16_TO_CPU(resp->vsi_used);
+               vsi_ctx->vsis_unallocated = LE16_TO_CPU(resp->vsi_free);
+       }
+
+       return status;
+}
+
+/**
+ * ice_is_vsi_valid - check whether the VSI is valid or not
+ * @hw: pointer to the hw struct
+ * @vsi_handle: VSI handle
+ *
+ * check whether the VSI is valid or not
+ */
+bool ice_is_vsi_valid(struct ice_hw *hw, u16 vsi_handle)
+{
+       return vsi_handle < ICE_MAX_VSI && hw->vsi_ctx[vsi_handle];
+}
+
+/**
+ * ice_get_hw_vsi_num - return the hw VSI number
+ * @hw: pointer to the hw struct
+ * @vsi_handle: VSI handle
+ *
+ * return the hw VSI number
+ * Caution: call this function only if VSI is valid (ice_is_vsi_valid)
+ */
+u16 ice_get_hw_vsi_num(struct ice_hw *hw, u16 vsi_handle)
+{
+       return hw->vsi_ctx[vsi_handle]->vsi_num;
+}
+
+/**
+ * ice_get_vsi_ctx - return the VSI context entry for a given VSI handle
+ * @hw: pointer to the hw struct
+ * @vsi_handle: VSI handle
+ *
+ * return the VSI context entry for a given VSI handle
+ */
+struct ice_vsi_ctx *ice_get_vsi_ctx(struct ice_hw *hw, u16 vsi_handle)
+{
+       return (vsi_handle >= ICE_MAX_VSI) ? NULL : hw->vsi_ctx[vsi_handle];
+}
+
+/**
+ * ice_save_vsi_ctx - save the VSI context for a given VSI handle
+ * @hw: pointer to the hw struct
+ * @vsi_handle: VSI handle
+ * @vsi: VSI context pointer
+ *
+ * save the VSI context entry for a given VSI handle
+ */
+static void
+ice_save_vsi_ctx(struct ice_hw *hw, u16 vsi_handle, struct ice_vsi_ctx *vsi)
+{
+       hw->vsi_ctx[vsi_handle] = vsi;
+}
+
+/**
+ * ice_clear_vsi_ctx - clear the VSI context entry
+ * @hw: pointer to the hw struct
+ * @vsi_handle: VSI handle
+ *
+ * clear the VSI context entry
+ */
+static void ice_clear_vsi_ctx(struct ice_hw *hw, u16 vsi_handle)
+{
+       struct ice_vsi_ctx *vsi;
+
+       vsi = ice_get_vsi_ctx(hw, vsi_handle);
+       if (vsi) {
+               ice_destroy_lock(&vsi->rss_locks);
+               ice_free(hw, vsi);
+               hw->vsi_ctx[vsi_handle] = NULL;
+       }
+}
+
+/**
+ * ice_clear_all_vsi_ctx - clear all the VSI context entries
+ * @hw: pointer to the hw struct
+ */
+void ice_clear_all_vsi_ctx(struct ice_hw *hw)
+{
+       u16 i;
+
+       for (i = 0; i < ICE_MAX_VSI; i++)
+               ice_clear_vsi_ctx(hw, i);
+}
+
+/**
+ * ice_add_vsi - add VSI context to the hardware and VSI handle list
+ * @hw: pointer to the hw struct
+ * @vsi_handle: unique VSI handle provided by drivers
+ * @vsi_ctx: pointer to a VSI context struct
+ * @cd: pointer to command details structure or NULL
+ *
+ * Add a VSI context to the hardware also add it into the VSI handle list.
+ * If this function gets called after reset for exisiting VSIs then update
+ * with the new HW VSI number in the corresponding VSI handle list entry.
+ */
+enum ice_status
+ice_add_vsi(struct ice_hw *hw, u16 vsi_handle, struct ice_vsi_ctx *vsi_ctx,
+           struct ice_sq_cd *cd)
+{
+       struct ice_vsi_ctx *tmp_vsi_ctx;
+       enum ice_status status;
+
+       if (vsi_handle >= ICE_MAX_VSI)
+               return ICE_ERR_PARAM;
+       status = ice_aq_add_vsi(hw, vsi_ctx, cd);
+       if (status)
+               return status;
+       tmp_vsi_ctx = ice_get_vsi_ctx(hw, vsi_handle);
+       if (!tmp_vsi_ctx) {
+               /* Create a new vsi context */
+               tmp_vsi_ctx = (struct ice_vsi_ctx *)
+                       ice_malloc(hw, sizeof(*tmp_vsi_ctx));
+               if (!tmp_vsi_ctx) {
+                       ice_aq_free_vsi(hw, vsi_ctx, false, cd);
+                       return ICE_ERR_NO_MEMORY;
+               }
+               *tmp_vsi_ctx = *vsi_ctx;
+               ice_init_lock(&tmp_vsi_ctx->rss_locks);
+               INIT_LIST_HEAD(&tmp_vsi_ctx->rss_list_head);
+               ice_save_vsi_ctx(hw, vsi_handle, tmp_vsi_ctx);
+       } else {
+               /* update with new HW VSI num */
+               if (tmp_vsi_ctx->vsi_num != vsi_ctx->vsi_num)
+                       tmp_vsi_ctx->vsi_num = vsi_ctx->vsi_num;
+       }
+
+       return status;
+}
+
+/**
+ * ice_free_vsi- free VSI context from hardware and VSI handle list
+ * @hw: pointer to the hw struct
+ * @vsi_handle: unique VSI handle
+ * @vsi_ctx: pointer to a VSI context struct
+ * @keep_vsi_alloc: keep VSI allocation as part of this PF's resources
+ * @cd: pointer to command details structure or NULL
+ *
+ * Free VSI context info from hardware as well as from VSI handle list
+ */
+enum ice_status
+ice_free_vsi(struct ice_hw *hw, u16 vsi_handle, struct ice_vsi_ctx *vsi_ctx,
+            bool keep_vsi_alloc, struct ice_sq_cd *cd)
+{
+       enum ice_status status;
+
+       if (!ice_is_vsi_valid(hw, vsi_handle))
+               return ICE_ERR_PARAM;
+       vsi_ctx->vsi_num = ice_get_hw_vsi_num(hw, vsi_handle);
+       status = ice_aq_free_vsi(hw, vsi_ctx, keep_vsi_alloc, cd);
+       if (!status)
+               ice_clear_vsi_ctx(hw, vsi_handle);
+       return status;
+}
+
+/**
+ * ice_update_vsi
+ * @hw: pointer to the hw struct
+ * @vsi_handle: unique VSI handle
+ * @vsi_ctx: pointer to a VSI context struct
+ * @cd: pointer to command details structure or NULL
+ *
+ * Update VSI context in the hardware
+ */
+enum ice_status
+ice_update_vsi(struct ice_hw *hw, u16 vsi_handle, struct ice_vsi_ctx *vsi_ctx,
+              struct ice_sq_cd *cd)
+{
+       if (!ice_is_vsi_valid(hw, vsi_handle))
+               return ICE_ERR_PARAM;
+       vsi_ctx->vsi_num = ice_get_hw_vsi_num(hw, vsi_handle);
+       return ice_aq_update_vsi(hw, vsi_ctx, cd);
+}
+
+
+
+/**
+ * ice_aq_alloc_free_vsi_list
+ * @hw: pointer to the hw struct
+ * @vsi_list_id: VSI list id returned or used for lookup
+ * @lkup_type: switch rule filter lookup type
+ * @opc: switch rules population command type - pass in the command opcode
+ *
+ * allocates or free a VSI list resource
+ */
+static enum ice_status
+ice_aq_alloc_free_vsi_list(struct ice_hw *hw, u16 *vsi_list_id,
+                          enum ice_sw_lkup_type lkup_type,
+                          enum ice_adminq_opc opc)
+{
+       struct ice_aqc_alloc_free_res_elem *sw_buf;
+       struct ice_aqc_res_elem *vsi_ele;
+       enum ice_status status;
+       u16 buf_len;
+
+       buf_len = sizeof(*sw_buf);
+       sw_buf = (struct ice_aqc_alloc_free_res_elem *)
+               ice_malloc(hw, buf_len);
+       if (!sw_buf)
+               return ICE_ERR_NO_MEMORY;
+       sw_buf->num_elems = CPU_TO_LE16(1);
+
+       if (lkup_type == ICE_SW_LKUP_MAC ||
+           lkup_type == ICE_SW_LKUP_MAC_VLAN ||
+           lkup_type == ICE_SW_LKUP_ETHERTYPE ||
+           lkup_type == ICE_SW_LKUP_ETHERTYPE_MAC ||
+           lkup_type == ICE_SW_LKUP_PROMISC ||
+           lkup_type == ICE_SW_LKUP_PROMISC_VLAN) {
+               sw_buf->res_type = CPU_TO_LE16(ICE_AQC_RES_TYPE_VSI_LIST_REP);
+       } else if (lkup_type == ICE_SW_LKUP_VLAN) {
+               sw_buf->res_type =
+                       CPU_TO_LE16(ICE_AQC_RES_TYPE_VSI_LIST_PRUNE);
+       } else {
+               status = ICE_ERR_PARAM;
+               goto ice_aq_alloc_free_vsi_list_exit;
+       }
+
+       if (opc == ice_aqc_opc_free_res)
+               sw_buf->elem[0].e.sw_resp = CPU_TO_LE16(*vsi_list_id);
+
+       status = ice_aq_alloc_free_res(hw, 1, sw_buf, buf_len, opc, NULL);
+       if (status)
+               goto ice_aq_alloc_free_vsi_list_exit;
+
+       if (opc == ice_aqc_opc_alloc_res) {
+               vsi_ele = &sw_buf->elem[0];
+               *vsi_list_id = LE16_TO_CPU(vsi_ele->e.sw_resp);
+       }
+
+ice_aq_alloc_free_vsi_list_exit:
+       ice_free(hw, sw_buf);
+       return status;
+}
+
+
+/**
+ * ice_aq_sw_rules - add/update/remove switch rules
+ * @hw: pointer to the hw struct
+ * @rule_list: pointer to switch rule population list
+ * @rule_list_sz: total size of the rule list in bytes
+ * @num_rules: number of switch rules in the rule_list
+ * @opc: switch rules population command type - pass in the command opcode
+ * @cd: pointer to command details structure or NULL
+ *
+ * Add(0x02a0)/Update(0x02a1)/Remove(0x02a2) switch rules commands to firmware
+ */
+static enum ice_status
+ice_aq_sw_rules(struct ice_hw *hw, void *rule_list, u16 rule_list_sz,
+               u8 num_rules, enum ice_adminq_opc opc, struct ice_sq_cd *cd)
+{
+       struct ice_aq_desc desc;
+
+       ice_debug(hw, ICE_DBG_TRACE, "ice_aq_sw_rules");
+
+       if (opc != ice_aqc_opc_add_sw_rules &&
+           opc != ice_aqc_opc_update_sw_rules &&
+           opc != ice_aqc_opc_remove_sw_rules)
+               return ICE_ERR_PARAM;
+
+       ice_fill_dflt_direct_cmd_desc(&desc, opc);
+
+       desc.flags |= CPU_TO_LE16(ICE_AQ_FLAG_RD);
+       desc.params.sw_rules.num_rules_fltr_entry_index =
+               CPU_TO_LE16(num_rules);
+       return ice_aq_send_cmd(hw, &desc, rule_list, rule_list_sz, cd);
+}
+
+
+/* ice_init_port_info - Initialize port_info with switch configuration data
+ * @pi: pointer to port_info
+ * @vsi_port_num: VSI number or port number
+ * @type: Type of switch element (port or VSI)
+ * @swid: switch ID of the switch the element is attached to
+ * @pf_vf_num: PF or VF number
+ * @is_vf: true if the element is a VF, false otherwise
+ */
+static void
+ice_init_port_info(struct ice_port_info *pi, u16 vsi_port_num, u8 type,
+                  u16 swid, u16 pf_vf_num, bool is_vf)
+{
+       switch (type) {
+       case ICE_AQC_GET_SW_CONF_RESP_PHYS_PORT:
+               pi->lport = (u8)(vsi_port_num & ICE_LPORT_MASK);
+               pi->sw_id = swid;
+               pi->pf_vf_num = pf_vf_num;
+               pi->is_vf = is_vf;
+               pi->dflt_tx_vsi_num = ICE_DFLT_VSI_INVAL;
+               pi->dflt_rx_vsi_num = ICE_DFLT_VSI_INVAL;
+               break;
+       default:
+               ice_debug(pi->hw, ICE_DBG_SW,
+                         "incorrect VSI/port type received\n");
+               break;
+       }
+}
+
+/* ice_get_initial_sw_cfg - Get initial port and default VSI data
+ * @hw: pointer to the hardware structure
+ */
+enum ice_status ice_get_initial_sw_cfg(struct ice_hw *hw)
+{
+       struct ice_aqc_get_sw_cfg_resp *rbuf;
+       enum ice_status status;
+       u16 num_total_ports;
+       u16 req_desc = 0;
+       u16 num_elems;
+       u16 j = 0;
+       u16 i;
+
+       num_total_ports = 1;
+
+       rbuf = (struct ice_aqc_get_sw_cfg_resp *)
+               ice_malloc(hw, ICE_SW_CFG_MAX_BUF_LEN);
+
+       if (!rbuf)
+               return ICE_ERR_NO_MEMORY;
+
+       /* Multiple calls to ice_aq_get_sw_cfg may be required
+        * to get all the switch configuration information. The need
+        * for additional calls is indicated by ice_aq_get_sw_cfg
+        * writing a non-zero value in req_desc
+        */
+       do {
+               status = ice_aq_get_sw_cfg(hw, rbuf, ICE_SW_CFG_MAX_BUF_LEN,
+                                          &req_desc, &num_elems, NULL);
+
+               if (status)
+                       break;
+
+               for (i = 0; i < num_elems; i++) {
+                       struct ice_aqc_get_sw_cfg_resp_elem *ele;
+                       u16 pf_vf_num, swid, vsi_port_num;
+                       bool is_vf = false;
+                       u8 type;
+
+                       ele = rbuf[i].elements;
+                       vsi_port_num = LE16_TO_CPU(ele->vsi_port_num) &
+                               ICE_AQC_GET_SW_CONF_RESP_VSI_PORT_NUM_M;
+
+                       pf_vf_num = LE16_TO_CPU(ele->pf_vf_num) &
+                               ICE_AQC_GET_SW_CONF_RESP_FUNC_NUM_M;
+
+                       swid = LE16_TO_CPU(ele->swid);
+
+                       if (LE16_TO_CPU(ele->pf_vf_num) &
+                           ICE_AQC_GET_SW_CONF_RESP_IS_VF)
+                               is_vf = true;
+
+                       type = LE16_TO_CPU(ele->vsi_port_num) >>
+                               ICE_AQC_GET_SW_CONF_RESP_TYPE_S;
+
+                       switch (type) {
+                       case ICE_AQC_GET_SW_CONF_RESP_PHYS_PORT:
+                       case ICE_AQC_GET_SW_CONF_RESP_VIRT_PORT:
+                               if (j == num_total_ports) {
+                                       ice_debug(hw, ICE_DBG_SW,
+                                                 "more ports than expected\n");
+                                       status = ICE_ERR_CFG;
+                                       goto out;
+                               }
+                               ice_init_port_info(hw->port_info,
+                                                  vsi_port_num, type, swid,
+                                                  pf_vf_num, is_vf);
+                               j++;
+                               break;
+                       default:
+                               break;
+                       }
+               }
+       } while (req_desc && !status);
+
+
+out:
+       ice_free(hw, (void *)rbuf);
+       return status;
+}
+
+
+/**
+ * ice_fill_sw_info - Helper function to populate lb_en and lan_en
+ * @hw: pointer to the hardware structure
+ * @fi: filter info structure to fill/update
+ *
+ * This helper function populates the lb_en and lan_en elements of the provided
+ * ice_fltr_info struct using the switch's type and characteristics of the
+ * switch rule being configured.
+ */
+static void ice_fill_sw_info(struct ice_hw *hw, struct ice_fltr_info *fi)
+{
+       fi->lb_en = false;
+       fi->lan_en = false;
+       if ((fi->flag & ICE_FLTR_TX) &&
+           (fi->fltr_act == ICE_FWD_TO_VSI ||
+            fi->fltr_act == ICE_FWD_TO_VSI_LIST ||
+            fi->fltr_act == ICE_FWD_TO_Q ||
+            fi->fltr_act == ICE_FWD_TO_QGRP)) {
+               /* Setting LB for prune actions will result in replicated
+                * packets to the internal switch that will be dropped.
+                */
+               if (fi->lkup_type != ICE_SW_LKUP_VLAN)
+                       fi->lb_en = true;
+
+               /* Set lan_en to TRUE if
+                * 1. The switch is a VEB AND
+                * 2
+                * 2.1 The lookup is a directional lookup like ethertype,
+                * promiscuous, ethertype-mac, promiscuous-vlan
+                * and default-port OR
+                * 2.2 The lookup is VLAN, OR
+                * 2.3 The lookup is MAC with mcast or bcast addr for MAC, OR
+                * 2.4 The lookup is MAC_VLAN with mcast or bcast addr for MAC.
+                *
+                * OR
+                *
+                * The switch is a VEPA.
+                *
+                * In all other cases, the LAN enable has to be set to false.
+                */
+               if (hw->evb_veb) {
+                       if (fi->lkup_type == ICE_SW_LKUP_ETHERTYPE ||
+                           fi->lkup_type == ICE_SW_LKUP_PROMISC ||
+                           fi->lkup_type == ICE_SW_LKUP_ETHERTYPE_MAC ||
+                           fi->lkup_type == ICE_SW_LKUP_PROMISC_VLAN ||
+                           fi->lkup_type == ICE_SW_LKUP_DFLT ||
+                           fi->lkup_type == ICE_SW_LKUP_VLAN ||
+                           (fi->lkup_type == ICE_SW_LKUP_MAC &&
+                            !IS_UNICAST_ETHER_ADDR(fi->l_data.mac.mac_addr)) ||
+                           (fi->lkup_type == ICE_SW_LKUP_MAC_VLAN &&
+                            !IS_UNICAST_ETHER_ADDR(fi->l_data.mac.mac_addr)))
+                               fi->lan_en = true;
+               } else {
+                       fi->lan_en = true;
+               }
+       }
+}
+
+/**
+ * ice_ilog2 - Caculates integer log base 2 of a number
+ * @n: number on which to perform operation
+ */
+static int ice_ilog2(u64 n)
+{
+       int i;
+
+       for (i = 63; i >= 0; i--)
+               if (((u64)1 << i) & n)
+                       return i;
+
+       return -1;
+}
+
+
+/**
+ * ice_fill_sw_rule - Helper function to fill switch rule structure
+ * @hw: pointer to the hardware structure
+ * @f_info: entry containing packet forwarding information
+ * @s_rule: switch rule structure to be filled in based on mac_entry
+ * @opc: switch rules population command type - pass in the command opcode
+ */
+static void
+ice_fill_sw_rule(struct ice_hw *hw, struct ice_fltr_info *f_info,
+                struct ice_aqc_sw_rules_elem *s_rule, enum ice_adminq_opc opc)
+{
+       u16 vlan_id = ICE_MAX_VLAN_ID + 1;
+       void *daddr = NULL;
+       u16 eth_hdr_sz;
+       u8 *eth_hdr;
+       u32 act = 0;
+       __be16 *off;
+       u8 q_rgn;
+
+
+       if (opc == ice_aqc_opc_remove_sw_rules) {
+               s_rule->pdata.lkup_tx_rx.act = 0;
+               s_rule->pdata.lkup_tx_rx.index =
+                       CPU_TO_LE16(f_info->fltr_rule_id);
+               s_rule->pdata.lkup_tx_rx.hdr_len = 0;
+               return;
+       }
+
+       eth_hdr_sz = sizeof(dummy_eth_header);
+       eth_hdr = s_rule->pdata.lkup_tx_rx.hdr;
+
+       /* initialize the ether header with a dummy header */
+       ice_memcpy(eth_hdr, dummy_eth_header, eth_hdr_sz, ICE_NONDMA_TO_NONDMA);
+       ice_fill_sw_info(hw, f_info);
+
+       switch (f_info->fltr_act) {
+       case ICE_FWD_TO_VSI:
+               act |= (f_info->fwd_id.hw_vsi_id << ICE_SINGLE_ACT_VSI_ID_S) &
+                       ICE_SINGLE_ACT_VSI_ID_M;
+               if (f_info->lkup_type != ICE_SW_LKUP_VLAN)
+                       act |= ICE_SINGLE_ACT_VSI_FORWARDING |
+                               ICE_SINGLE_ACT_VALID_BIT;
+               break;
+       case ICE_FWD_TO_VSI_LIST:
+               act |= ICE_SINGLE_ACT_VSI_LIST;
+               act |= (f_info->fwd_id.vsi_list_id <<
+                       ICE_SINGLE_ACT_VSI_LIST_ID_S) &
+                       ICE_SINGLE_ACT_VSI_LIST_ID_M;
+               if (f_info->lkup_type != ICE_SW_LKUP_VLAN)
+                       act |= ICE_SINGLE_ACT_VSI_FORWARDING |
+                               ICE_SINGLE_ACT_VALID_BIT;
+               break;
+       case ICE_FWD_TO_Q:
+               act |= ICE_SINGLE_ACT_TO_Q;
+               act |= (f_info->fwd_id.q_id << ICE_SINGLE_ACT_Q_INDEX_S) &
+                       ICE_SINGLE_ACT_Q_INDEX_M;
+               break;
+       case ICE_DROP_PACKET:
+               act |= ICE_SINGLE_ACT_VSI_FORWARDING | ICE_SINGLE_ACT_DROP |
+                       ICE_SINGLE_ACT_VALID_BIT;
+               break;
+       case ICE_FWD_TO_QGRP:
+               q_rgn = f_info->qgrp_size > 0 ?
+                       (u8)ice_ilog2(f_info->qgrp_size) : 0;
+               act |= ICE_SINGLE_ACT_TO_Q;
+               act |= (f_info->fwd_id.q_id << ICE_SINGLE_ACT_Q_INDEX_S) &
+                       ICE_SINGLE_ACT_Q_INDEX_M;
+               act |= (q_rgn << ICE_SINGLE_ACT_Q_REGION_S) &
+                       ICE_SINGLE_ACT_Q_REGION_M;
+               break;
+       default:
+               return;
+       }
+
+       if (f_info->lb_en)
+               act |= ICE_SINGLE_ACT_LB_ENABLE;
+       if (f_info->lan_en)
+               act |= ICE_SINGLE_ACT_LAN_ENABLE;
+
+       switch (f_info->lkup_type) {
+       case ICE_SW_LKUP_MAC:
+               daddr = f_info->l_data.mac.mac_addr;
+               break;
+       case ICE_SW_LKUP_VLAN:
+               vlan_id = f_info->l_data.vlan.vlan_id;
+               if (f_info->fltr_act == ICE_FWD_TO_VSI ||
+                   f_info->fltr_act == ICE_FWD_TO_VSI_LIST) {
+                       act |= ICE_SINGLE_ACT_PRUNE;
+                       act |= ICE_SINGLE_ACT_EGRESS | ICE_SINGLE_ACT_INGRESS;
+               }
+               break;
+       case ICE_SW_LKUP_ETHERTYPE_MAC:
+               daddr = f_info->l_data.ethertype_mac.mac_addr;
+               /* fall-through */
+       case ICE_SW_LKUP_ETHERTYPE:
+               off = (__be16 *)(eth_hdr + ICE_ETH_ETHTYPE_OFFSET);
+               *off = CPU_TO_BE16(f_info->l_data.ethertype_mac.ethertype);
+               break;
+       case ICE_SW_LKUP_MAC_VLAN:
+               daddr = f_info->l_data.mac_vlan.mac_addr;
+               vlan_id = f_info->l_data.mac_vlan.vlan_id;
+               break;
+       case ICE_SW_LKUP_PROMISC_VLAN:
+               vlan_id = f_info->l_data.mac_vlan.vlan_id;
+               /* fall-through */
+       case ICE_SW_LKUP_PROMISC:
+               daddr = f_info->l_data.mac_vlan.mac_addr;
+               break;
+       default:
+               break;
+       }
+
+       s_rule->type = (f_info->flag & ICE_FLTR_RX) ?
+               CPU_TO_LE16(ICE_AQC_SW_RULES_T_LKUP_RX) :
+               CPU_TO_LE16(ICE_AQC_SW_RULES_T_LKUP_TX);
+
+       /* Recipe set depending on lookup type */
+       s_rule->pdata.lkup_tx_rx.recipe_id = CPU_TO_LE16(f_info->lkup_type);
+       s_rule->pdata.lkup_tx_rx.src = CPU_TO_LE16(f_info->src);
+       s_rule->pdata.lkup_tx_rx.act = CPU_TO_LE32(act);
+
+       if (daddr)
+               ice_memcpy(eth_hdr + ICE_ETH_DA_OFFSET, daddr, ETH_ALEN,
+                          ICE_NONDMA_TO_NONDMA);
+
+       if (!(vlan_id > ICE_MAX_VLAN_ID)) {
+               off = (__be16 *)(eth_hdr + ICE_ETH_VLAN_TCI_OFFSET);
+               *off = CPU_TO_BE16(vlan_id);
+       }
+
+       /* Create the switch rule with the final dummy Ethernet header */
+       if (opc != ice_aqc_opc_update_sw_rules)
+               s_rule->pdata.lkup_tx_rx.hdr_len = CPU_TO_LE16(eth_hdr_sz);
+}
+
+/**
+ * ice_add_marker_act
+ * @hw: pointer to the hardware structure
+ * @m_ent: the management entry for which sw marker needs to be added
+ * @sw_marker: sw marker to tag the Rx descriptor with
+ * @l_id: large action resource id
+ *
+ * Create a large action to hold software marker and update the switch rule
+ * entry pointed by m_ent with newly created large action
+ */
+static enum ice_status
+ice_add_marker_act(struct ice_hw *hw, struct ice_fltr_mgmt_list_entry *m_ent,
+                  u16 sw_marker, u16 l_id)
+{
+       struct ice_aqc_sw_rules_elem *lg_act, *rx_tx;
+       /* For software marker we need 3 large actions
+        * 1. FWD action: FWD TO VSI or VSI LIST
+        * 2. GENERIC VALUE action to hold the profile id
+        * 3. GENERIC VALUE action to hold the software marker id
+        */
+       const u16 num_lg_acts = 3;
+       enum ice_status status;
+       u16 lg_act_size;
+       u16 rules_size;
+       u32 act;
+       u16 id;
+
+       if (m_ent->fltr_info.lkup_type != ICE_SW_LKUP_MAC)
+               return ICE_ERR_PARAM;
+
+       /* Create two back-to-back switch rules and submit them to the HW using
+        * one memory buffer:
+        *    1. Large Action
+        *    2. Look up Tx Rx
+        */
+       lg_act_size = (u16)ICE_SW_RULE_LG_ACT_SIZE(num_lg_acts);
+       rules_size = lg_act_size + ICE_SW_RULE_RX_TX_ETH_HDR_SIZE;
+       lg_act = (struct ice_aqc_sw_rules_elem *)ice_malloc(hw, rules_size);
+       if (!lg_act)
+               return ICE_ERR_NO_MEMORY;
+
+       rx_tx = (struct ice_aqc_sw_rules_elem *)((u8 *)lg_act + lg_act_size);
+
+       /* Fill in the first switch rule i.e. large action */
+       lg_act->type = CPU_TO_LE16(ICE_AQC_SW_RULES_T_LG_ACT);
+       lg_act->pdata.lg_act.index = CPU_TO_LE16(l_id);
+       lg_act->pdata.lg_act.size = CPU_TO_LE16(num_lg_acts);
+
+       /* First action VSI forwarding or VSI list forwarding depending on how
+        * many VSIs
+        */
+       id = (m_ent->vsi_count > 1) ? m_ent->fltr_info.fwd_id.vsi_list_id :
+               m_ent->fltr_info.fwd_id.hw_vsi_id;
+
+       act = ICE_LG_ACT_VSI_FORWARDING | ICE_LG_ACT_VALID_BIT;
+       act |= (id << ICE_LG_ACT_VSI_LIST_ID_S) &
+               ICE_LG_ACT_VSI_LIST_ID_M;
+       if (m_ent->vsi_count > 1)
+               act |= ICE_LG_ACT_VSI_LIST;
+       lg_act->pdata.lg_act.act[0] = CPU_TO_LE32(act);
+
+       /* Second action descriptor type */
+       act = ICE_LG_ACT_GENERIC;
+
+       act |= (1 << ICE_LG_ACT_GENERIC_VALUE_S) & ICE_LG_ACT_GENERIC_VALUE_M;
+       lg_act->pdata.lg_act.act[1] = CPU_TO_LE32(act);
+
+       act = (ICE_LG_ACT_GENERIC_OFF_RX_DESC_PROF_IDX <<
+              ICE_LG_ACT_GENERIC_OFFSET_S) & ICE_LG_ACT_GENERIC_OFFSET_M;
+
+       /* Third action Marker value */
+       act |= ICE_LG_ACT_GENERIC;
+       act |= (sw_marker << ICE_LG_ACT_GENERIC_VALUE_S) &
+               ICE_LG_ACT_GENERIC_VALUE_M;
+
+       lg_act->pdata.lg_act.act[2] = CPU_TO_LE32(act);
+
+       /* call the fill switch rule to fill the lookup Tx Rx structure */
+       ice_fill_sw_rule(hw, &m_ent->fltr_info, rx_tx,
+                        ice_aqc_opc_update_sw_rules);
+
+       /* Update the action to point to the large action id */
+       rx_tx->pdata.lkup_tx_rx.act =
+               CPU_TO_LE32(ICE_SINGLE_ACT_PTR |
+                           ((l_id << ICE_SINGLE_ACT_PTR_VAL_S) &
+                            ICE_SINGLE_ACT_PTR_VAL_M));
+
+       /* Use the filter rule id of the previously created rule with single
+        * act. Once the update happens, hardware will treat this as large
+        * action
+        */
+       rx_tx->pdata.lkup_tx_rx.index =
+               CPU_TO_LE16(m_ent->fltr_info.fltr_rule_id);
+
+       status = ice_aq_sw_rules(hw, lg_act, rules_size, 2,
+                                ice_aqc_opc_update_sw_rules, NULL);
+       if (!status) {
+               m_ent->lg_act_idx = l_id;
+               m_ent->sw_marker_id = sw_marker;
+       }
+
+       ice_free(hw, lg_act);
+       return status;
+}
+
+
+/**
+ * ice_create_vsi_list_map
+ * @hw: pointer to the hardware structure
+ * @vsi_handle_arr: array of VSI handles to set in the VSI mapping
+ * @num_vsi: number of VSI handles in the array
+ * @vsi_list_id: VSI list id generated as part of allocate resource
+ *
+ * Helper function to create a new entry of VSI list id to VSI mapping
+ * using the given VSI list id
+ */
+static struct ice_vsi_list_map_info *
+ice_create_vsi_list_map(struct ice_hw *hw, u16 *vsi_handle_arr, u16 num_vsi,
+                       u16 vsi_list_id)
+{
+       struct ice_switch_info *sw = hw->switch_info;
+       struct ice_vsi_list_map_info *v_map;
+       int i;
+
+       v_map = (struct ice_vsi_list_map_info *)ice_calloc(hw, 1,
+               sizeof(*v_map));
+       if (!v_map)
+               return NULL;
+
+       v_map->vsi_list_id = vsi_list_id;
+       v_map->ref_cnt = 1;
+       for (i = 0; i < num_vsi; i++)
+               ice_set_bit(vsi_handle_arr[i], v_map->vsi_map);
+
+       LIST_ADD(&v_map->list_entry, &sw->vsi_list_map_head);
+       return v_map;
+}
+
+/**
+ * ice_update_vsi_list_rule
+ * @hw: pointer to the hardware structure
+ * @vsi_handle_arr: array of VSI handles to form a VSI list
+ * @num_vsi: number of VSI handles in the array
+ * @vsi_list_id: VSI list id generated as part of allocate resource
+ * @remove: Boolean value to indicate if this is a remove action
+ * @opc: switch rules population command type - pass in the command opcode
+ * @lkup_type: lookup type of the filter
+ *
+ * Call AQ command to add a new switch rule or update existing switch rule
+ * using the given VSI list id
+ */
+static enum ice_status
+ice_update_vsi_list_rule(struct ice_hw *hw, u16 *vsi_handle_arr, u16 num_vsi,
+                        u16 vsi_list_id, bool remove, enum ice_adminq_opc opc,
+                        enum ice_sw_lkup_type lkup_type)
+{
+       struct ice_aqc_sw_rules_elem *s_rule;
+       enum ice_status status;
+       u16 s_rule_size;
+       u16 type;
+       int i;
+
+       if (!num_vsi)
+               return ICE_ERR_PARAM;
+
+       if (lkup_type == ICE_SW_LKUP_MAC ||
+           lkup_type == ICE_SW_LKUP_MAC_VLAN ||
+           lkup_type == ICE_SW_LKUP_ETHERTYPE ||
+           lkup_type == ICE_SW_LKUP_ETHERTYPE_MAC ||
+           lkup_type == ICE_SW_LKUP_PROMISC ||
+           lkup_type == ICE_SW_LKUP_PROMISC_VLAN)
+               type = remove ? ICE_AQC_SW_RULES_T_VSI_LIST_CLEAR :
+                               ICE_AQC_SW_RULES_T_VSI_LIST_SET;
+       else if (lkup_type == ICE_SW_LKUP_VLAN)
+               type = remove ? ICE_AQC_SW_RULES_T_PRUNE_LIST_CLEAR :
+                               ICE_AQC_SW_RULES_T_PRUNE_LIST_SET;
+       else
+               return ICE_ERR_PARAM;
+
+       s_rule_size = (u16)ICE_SW_RULE_VSI_LIST_SIZE(num_vsi);
+       s_rule = (struct ice_aqc_sw_rules_elem *)ice_malloc(hw, s_rule_size);
+       if (!s_rule)
+               return ICE_ERR_NO_MEMORY;
+       for (i = 0; i < num_vsi; i++) {
+               if (!ice_is_vsi_valid(hw, vsi_handle_arr[i])) {
+                       status = ICE_ERR_PARAM;
+                       goto exit;
+               }
+               /* AQ call requires hw_vsi_id(s) */
+               s_rule->pdata.vsi_list.vsi[i] =
+                       CPU_TO_LE16(ice_get_hw_vsi_num(hw, vsi_handle_arr[i]));
+       }
+
+       s_rule->type = CPU_TO_LE16(type);
+       s_rule->pdata.vsi_list.number_vsi = CPU_TO_LE16(num_vsi);
+       s_rule->pdata.vsi_list.index = CPU_TO_LE16(vsi_list_id);
+
+       status = ice_aq_sw_rules(hw, s_rule, s_rule_size, 1, opc, NULL);
+
+exit:
+       ice_free(hw, s_rule);
+       return status;
+}
+
+/**
+ * ice_create_vsi_list_rule - Creates and populates a VSI list rule
+ * @hw: pointer to the hw struct
+ * @vsi_handle_arr: array of VSI handles to form a VSI list
+ * @num_vsi: number of VSI handles in the array
+ * @vsi_list_id: stores the ID of the VSI list to be created
+ * @lkup_type: switch rule filter's lookup type
+ */
+static enum ice_status
+ice_create_vsi_list_rule(struct ice_hw *hw, u16 *vsi_handle_arr, u16 num_vsi,
+                        u16 *vsi_list_id, enum ice_sw_lkup_type lkup_type)
+{
+       enum ice_status status;
+
+       status = ice_aq_alloc_free_vsi_list(hw, vsi_list_id, lkup_type,
+                                           ice_aqc_opc_alloc_res);
+       if (status)
+               return status;
+
+       /* Update the newly created VSI list to include the specified VSIs */
+       return ice_update_vsi_list_rule(hw, vsi_handle_arr, num_vsi,
+                                       *vsi_list_id, false,
+                                       ice_aqc_opc_add_sw_rules, lkup_type);
+}
+
+/**
+ * ice_create_pkt_fwd_rule
+ * @hw: pointer to the hardware structure
+ * @f_entry: entry containing packet forwarding information
+ *
+ * Create switch rule with given filter information and add an entry
+ * to the corresponding filter management list to track this switch rule
+ * and VSI mapping
+ */
+static enum ice_status
+ice_create_pkt_fwd_rule(struct ice_hw *hw,
+                       struct ice_fltr_list_entry *f_entry)
+{
+       struct ice_fltr_mgmt_list_entry *fm_entry;
+       struct ice_aqc_sw_rules_elem *s_rule;
+       enum ice_sw_lkup_type l_type;
+       struct ice_sw_recipe *recp;
+       enum ice_status status;
+
+       s_rule = (struct ice_aqc_sw_rules_elem *)
+               ice_malloc(hw, ICE_SW_RULE_RX_TX_ETH_HDR_SIZE);
+       if (!s_rule)
+               return ICE_ERR_NO_MEMORY;
+       fm_entry = (struct ice_fltr_mgmt_list_entry *)
+                  ice_malloc(hw, sizeof(*fm_entry));
+       if (!fm_entry) {
+               status = ICE_ERR_NO_MEMORY;
+               goto ice_create_pkt_fwd_rule_exit;
+       }
+
+       fm_entry->fltr_info = f_entry->fltr_info;
+
+       /* Initialize all the fields for the management entry */
+       fm_entry->vsi_count = 1;
+       fm_entry->lg_act_idx = ICE_INVAL_LG_ACT_INDEX;
+       fm_entry->sw_marker_id = ICE_INVAL_SW_MARKER_ID;
+       fm_entry->counter_index = ICE_INVAL_COUNTER_ID;
+
+       ice_fill_sw_rule(hw, &fm_entry->fltr_info, s_rule,
+                        ice_aqc_opc_add_sw_rules);
+
+       status = ice_aq_sw_rules(hw, s_rule, ICE_SW_RULE_RX_TX_ETH_HDR_SIZE, 1,
+                                ice_aqc_opc_add_sw_rules, NULL);
+       if (status) {
+               ice_free(hw, fm_entry);
+               goto ice_create_pkt_fwd_rule_exit;
+       }
+
+       f_entry->fltr_info.fltr_rule_id =
+               LE16_TO_CPU(s_rule->pdata.lkup_tx_rx.index);
+       fm_entry->fltr_info.fltr_rule_id =
+               LE16_TO_CPU(s_rule->pdata.lkup_tx_rx.index);
+
+       /* The book keeping entries will get removed when base driver
+        * calls remove filter AQ command
+        */
+       l_type = fm_entry->fltr_info.lkup_type;
+       recp = &hw->switch_info->recp_list[l_type];
+       LIST_ADD(&fm_entry->list_entry, &recp->filt_rules);
+
+ice_create_pkt_fwd_rule_exit:
+       ice_free(hw, s_rule);
+       return status;
+}
+
+/**
+ * ice_update_pkt_fwd_rule
+ * @hw: pointer to the hardware structure
+ * @f_info: filter information for switch rule
+ *
+ * Call AQ command to update a previously created switch rule with a
+ * VSI list id
+ */
+static enum ice_status
+ice_update_pkt_fwd_rule(struct ice_hw *hw, struct ice_fltr_info *f_info)
+{
+       struct ice_aqc_sw_rules_elem *s_rule;
+       enum ice_status status;
+
+       s_rule = (struct ice_aqc_sw_rules_elem *)
+               ice_malloc(hw, ICE_SW_RULE_RX_TX_ETH_HDR_SIZE);
+       if (!s_rule)
+               return ICE_ERR_NO_MEMORY;
+
+       ice_fill_sw_rule(hw, f_info, s_rule, ice_aqc_opc_update_sw_rules);
+
+       s_rule->pdata.lkup_tx_rx.index = CPU_TO_LE16(f_info->fltr_rule_id);
+
+       /* Update switch rule with new rule set to forward VSI list */
+       status = ice_aq_sw_rules(hw, s_rule, ICE_SW_RULE_RX_TX_ETH_HDR_SIZE, 1,
+                                ice_aqc_opc_update_sw_rules, NULL);
+
+       ice_free(hw, s_rule);
+       return status;
+}
+
+/**
+ * ice_update_sw_rule_bridge_mode
+ * @hw: pointer to the hw struct
+ *
+ * Updates unicast switch filter rules based on VEB/VEPA mode
+ */
+enum ice_status ice_update_sw_rule_bridge_mode(struct ice_hw *hw)
+{
+       struct ice_switch_info *sw = hw->switch_info;
+       struct ice_fltr_mgmt_list_entry *fm_entry;
+       enum ice_status status = ICE_SUCCESS;
+       struct LIST_HEAD_TYPE *rule_head;
+       struct ice_lock *rule_lock; /* Lock to protect filter rule list */
+
+       rule_lock = &sw->recp_list[ICE_SW_LKUP_MAC].filt_rule_lock;
+       rule_head = &sw->recp_list[ICE_SW_LKUP_MAC].filt_rules;
+
+       ice_acquire_lock(rule_lock);
+       LIST_FOR_EACH_ENTRY(fm_entry, rule_head, ice_fltr_mgmt_list_entry,
+                           list_entry) {
+               struct ice_fltr_info *fi = &fm_entry->fltr_info;
+               u8 *addr = fi->l_data.mac.mac_addr;
+
+               /* Update unicast Tx rules to reflect the selected
+                * VEB/VEPA mode
+                */
+               if ((fi->flag & ICE_FLTR_TX) && IS_UNICAST_ETHER_ADDR(addr) &&
+                   (fi->fltr_act == ICE_FWD_TO_VSI ||
+                    fi->fltr_act == ICE_FWD_TO_VSI_LIST ||
+                    fi->fltr_act == ICE_FWD_TO_Q ||
+                    fi->fltr_act == ICE_FWD_TO_QGRP)) {
+                       status = ice_update_pkt_fwd_rule(hw, fi);
+                       if (status)
+                               break;
+               }
+       }
+
+       ice_release_lock(rule_lock);
+
+       return status;
+}
+
+/**
+ * ice_add_update_vsi_list
+ * @hw: pointer to the hardware structure
+ * @m_entry: pointer to current filter management list entry
+ * @cur_fltr: filter information from the book keeping entry
+ * @new_fltr: filter information with the new VSI to be added
+ *
+ * Call AQ command to add or update previously created VSI list with new VSI.
+ *
+ * Helper function to do book keeping associated with adding filter information
+ * The algorithm to do the book keeping is described below :
+ * When a VSI needs to subscribe to a given filter (MAC/VLAN/Ethtype etc.)
+ *     if only one VSI has been added till now
+ *             Allocate a new VSI list and add two VSIs
+ *             to this list using switch rule command
+ *             Update the previously created switch rule with the
+ *             newly created VSI list id
+ *     if a VSI list was previously created
+ *             Add the new VSI to the previously created VSI list set
+ *             using the update switch rule command
+ */
+static enum ice_status
+ice_add_update_vsi_list(struct ice_hw *hw,
+                       struct ice_fltr_mgmt_list_entry *m_entry,
+                       struct ice_fltr_info *cur_fltr,
+                       struct ice_fltr_info *new_fltr)
+{
+       enum ice_status status = ICE_SUCCESS;
+       u16 vsi_list_id = 0;
+
+       if ((cur_fltr->fltr_act == ICE_FWD_TO_Q ||
+            cur_fltr->fltr_act == ICE_FWD_TO_QGRP))
+               return ICE_ERR_NOT_IMPL;
+
+       if ((new_fltr->fltr_act == ICE_FWD_TO_Q ||
+            new_fltr->fltr_act == ICE_FWD_TO_QGRP) &&
+           (cur_fltr->fltr_act == ICE_FWD_TO_VSI ||
+            cur_fltr->fltr_act == ICE_FWD_TO_VSI_LIST))
+               return ICE_ERR_NOT_IMPL;
+
+       if (m_entry->vsi_count < 2 && !m_entry->vsi_list_info) {
+               /* Only one entry existed in the mapping and it was not already
+                * a part of a VSI list. So, create a VSI list with the old and
+                * new VSIs.
+                */
+               struct ice_fltr_info tmp_fltr;
+               u16 vsi_handle_arr[2];
+
+               /* A rule already exists with the new VSI being added */
+               if (cur_fltr->fwd_id.hw_vsi_id == new_fltr->fwd_id.hw_vsi_id)
+                       return ICE_ERR_ALREADY_EXISTS;
+
+               vsi_handle_arr[0] = cur_fltr->vsi_handle;
+               vsi_handle_arr[1] = new_fltr->vsi_handle;
+               status = ice_create_vsi_list_rule(hw, &vsi_handle_arr[0], 2,
+                                                 &vsi_list_id,
+                                                 new_fltr->lkup_type);
+               if (status)
+                       return status;
+
+               tmp_fltr = *new_fltr;
+               tmp_fltr.fltr_rule_id = cur_fltr->fltr_rule_id;
+               tmp_fltr.fltr_act = ICE_FWD_TO_VSI_LIST;
+               tmp_fltr.fwd_id.vsi_list_id = vsi_list_id;
+               /* Update the previous switch rule of "MAC forward to VSI" to
+                * "MAC fwd to VSI list"
+                */
+               status = ice_update_pkt_fwd_rule(hw, &tmp_fltr);
+               if (status)
+                       return status;
+
+               cur_fltr->fwd_id.vsi_list_id = vsi_list_id;
+               cur_fltr->fltr_act = ICE_FWD_TO_VSI_LIST;
+               m_entry->vsi_list_info =
+                       ice_create_vsi_list_map(hw, &vsi_handle_arr[0], 2,
+                                               vsi_list_id);
+
+               /* If this entry was large action then the large action needs
+                * to be updated to point to FWD to VSI list
+                */
+               if (m_entry->sw_marker_id != ICE_INVAL_SW_MARKER_ID)
+                       status =
+                           ice_add_marker_act(hw, m_entry,
+                                              m_entry->sw_marker_id,
+                                              m_entry->lg_act_idx);
+       } else {
+               u16 vsi_handle = new_fltr->vsi_handle;
+               enum ice_adminq_opc opcode;
+
+               if (!m_entry->vsi_list_info)
+                       return ICE_ERR_CFG;
+
+               /* A rule already exists with the new VSI being added */
+               if (ice_is_bit_set(m_entry->vsi_list_info->vsi_map, vsi_handle))
+                       return ICE_SUCCESS;
+
+               /* Update the previously created VSI list set with
+                * the new VSI id passed in
+                */
+               vsi_list_id = cur_fltr->fwd_id.vsi_list_id;
+               opcode = ice_aqc_opc_update_sw_rules;
+
+               status = ice_update_vsi_list_rule(hw, &vsi_handle, 1,
+                                                 vsi_list_id, false, opcode,
+                                                 new_fltr->lkup_type);
+               /* update VSI list mapping info with new VSI id */
+               if (!status)
+                       ice_set_bit(vsi_handle,
+                                   m_entry->vsi_list_info->vsi_map);
+       }
+       if (!status)
+               m_entry->vsi_count++;
+       return status;
+}
+
+/**
+ * ice_find_rule_entry - Search a rule entry
+ * @hw: pointer to the hardware structure
+ * @recp_id: lookup type for which the specified rule needs to be searched
+ * @f_info: rule information
+ *
+ * Helper function to search for a given rule entry
+ * Returns pointer to entry storing the rule if found
+ */
+static struct ice_fltr_mgmt_list_entry *
+ice_find_rule_entry(struct ice_hw *hw, u8 recp_id, struct ice_fltr_info *f_info)
+{
+       struct ice_fltr_mgmt_list_entry *list_itr, *ret = NULL;
+       struct ice_switch_info *sw = hw->switch_info;
+       struct LIST_HEAD_TYPE *list_head;
+
+       list_head = &sw->recp_list[recp_id].filt_rules;
+       LIST_FOR_EACH_ENTRY(list_itr, list_head, ice_fltr_mgmt_list_entry,
+                           list_entry) {
+               if (!memcmp(&f_info->l_data, &list_itr->fltr_info.l_data,
+                           sizeof(f_info->l_data)) &&
+                   f_info->flag == list_itr->fltr_info.flag) {
+                       ret = list_itr;
+                       break;
+               }
+       }
+       return ret;
+}
+
+/**
+ * ice_find_vsi_list_entry - Search VSI list map with VSI count 1
+ * @hw: pointer to the hardware structure
+ * @recp_id: lookup type for which VSI lists needs to be searched
+ * @vsi_handle: VSI handle to be found in VSI list
+ * @vsi_list_id: VSI list id found contaning vsi_handle
+ *
+ * Helper function to search a VSI list with single entry containing given VSI
+ * handle element. This can be extended further to search VSI list with more
+ * than 1 vsi_count. Returns pointer to VSI list entry if found.
+ */
+static struct ice_vsi_list_map_info *
+ice_find_vsi_list_entry(struct ice_hw *hw, u8 recp_id, u16 vsi_handle,
+                       u16 *vsi_list_id)
+{
+       struct ice_vsi_list_map_info *map_info = NULL;
+       struct ice_switch_info *sw = hw->switch_info;
+       struct ice_fltr_mgmt_list_entry *list_itr;
+       struct LIST_HEAD_TYPE *list_head;
+
+       list_head = &sw->recp_list[recp_id].filt_rules;
+       LIST_FOR_EACH_ENTRY(list_itr, list_head, ice_fltr_mgmt_list_entry,
+                           list_entry) {
+               if (list_itr->vsi_count == 1 && list_itr->vsi_list_info) {
+                       map_info = list_itr->vsi_list_info;
+                       if (ice_is_bit_set(map_info->vsi_map, vsi_handle)) {
+                               *vsi_list_id = map_info->vsi_list_id;
+                               return map_info;
+                       }
+               }
+       }
+       return NULL;
+}
+
+/**
+ * ice_add_rule_internal - add rule for a given lookup type
+ * @hw: pointer to the hardware structure
+ * @recp_id: lookup type (recipe id) for which rule has to be added
+ * @f_entry: structure containing MAC forwarding information
+ *
+ * Adds or updates the rule lists for a given recipe
+ */
+static enum ice_status
+ice_add_rule_internal(struct ice_hw *hw, u8 recp_id,
+                     struct ice_fltr_list_entry *f_entry)
+{
+       struct ice_switch_info *sw = hw->switch_info;
+       struct ice_fltr_info *new_fltr, *cur_fltr;
+       struct ice_fltr_mgmt_list_entry *m_entry;
+       struct ice_lock *rule_lock; /* Lock to protect filter rule list */
+       enum ice_status status = ICE_SUCCESS;
+
+       if (!ice_is_vsi_valid(hw, f_entry->fltr_info.vsi_handle))
+               return ICE_ERR_PARAM;
+
+       /* Load the hw_vsi_id only if the fwd action is fwd to VSI */
+       if (f_entry->fltr_info.fltr_act == ICE_FWD_TO_VSI)
+               f_entry->fltr_info.fwd_id.hw_vsi_id =
+                       ice_get_hw_vsi_num(hw, f_entry->fltr_info.vsi_handle);
+
+       rule_lock = &sw->recp_list[recp_id].filt_rule_lock;
+
+       ice_acquire_lock(rule_lock);
+       new_fltr = &f_entry->fltr_info;
+       if (new_fltr->flag & ICE_FLTR_RX)
+               new_fltr->src = hw->port_info->lport;
+       else if (new_fltr->flag & ICE_FLTR_TX)
+               new_fltr->src =
+                       ice_get_hw_vsi_num(hw, f_entry->fltr_info.vsi_handle);
+
+       m_entry = ice_find_rule_entry(hw, recp_id, new_fltr);
+       if (!m_entry) {
+               ice_release_lock(rule_lock);
+               return ice_create_pkt_fwd_rule(hw, f_entry);
+       }
+
+       cur_fltr = &m_entry->fltr_info;
+       status = ice_add_update_vsi_list(hw, m_entry, cur_fltr, new_fltr);
+       ice_release_lock(rule_lock);
+
+       return status;
+}
+
+/**
+ * ice_remove_vsi_list_rule
+ * @hw: pointer to the hardware structure
+ * @vsi_list_id: VSI list id generated as part of allocate resource
+ * @lkup_type: switch rule filter lookup type
+ *
+ * The VSI list should be emptied before this function is called to remove the
+ * VSI list.
+ */
+static enum ice_status
+ice_remove_vsi_list_rule(struct ice_hw *hw, u16 vsi_list_id,
+                        enum ice_sw_lkup_type lkup_type)
+{
+       struct ice_aqc_sw_rules_elem *s_rule;
+       enum ice_status status;
+       u16 s_rule_size;
+
+       s_rule_size = (u16)ICE_SW_RULE_VSI_LIST_SIZE(0);
+       s_rule = (struct ice_aqc_sw_rules_elem *)ice_malloc(hw, s_rule_size);
+       if (!s_rule)
+               return ICE_ERR_NO_MEMORY;
+
+       s_rule->type = CPU_TO_LE16(ICE_AQC_SW_RULES_T_VSI_LIST_CLEAR);
+       s_rule->pdata.vsi_list.index = CPU_TO_LE16(vsi_list_id);
+
+       /* Free the vsi_list resource that we allocated. It is assumed that the
+        * list is empty at this point.
+        */
+       status = ice_aq_alloc_free_vsi_list(hw, &vsi_list_id, lkup_type,
+                                           ice_aqc_opc_free_res);
+
+       ice_free(hw, s_rule);
+       return status;
+}
+
+/**
+ * ice_rem_update_vsi_list
+ * @hw: pointer to the hardware structure
+ * @vsi_handle: VSI handle of the VSI to remove
+ * @fm_list: filter management entry for which the VSI list management needs to
+ *          be done
+ */
+static enum ice_status
+ice_rem_update_vsi_list(struct ice_hw *hw, u16 vsi_handle,
+                       struct ice_fltr_mgmt_list_entry *fm_list)
+{
+       enum ice_sw_lkup_type lkup_type;
+       enum ice_status status = ICE_SUCCESS;
+       u16 vsi_list_id;
+
+       if (fm_list->fltr_info.fltr_act != ICE_FWD_TO_VSI_LIST ||
+           fm_list->vsi_count == 0)
+               return ICE_ERR_PARAM;
+
+       /* A rule with the VSI being removed does not exist */
+       if (!ice_is_bit_set(fm_list->vsi_list_info->vsi_map, vsi_handle))
+               return ICE_ERR_DOES_NOT_EXIST;
+
+       lkup_type = fm_list->fltr_info.lkup_type;
+       vsi_list_id = fm_list->fltr_info.fwd_id.vsi_list_id;
+       status = ice_update_vsi_list_rule(hw, &vsi_handle, 1, vsi_list_id, true,
+                                         ice_aqc_opc_update_sw_rules,
+                                         lkup_type);
+       if (status)
+               return status;
+
+       fm_list->vsi_count--;
+       ice_clear_bit(vsi_handle, fm_list->vsi_list_info->vsi_map);
+
+       if (fm_list->vsi_count == 1 && lkup_type != ICE_SW_LKUP_VLAN) {
+               struct ice_fltr_info tmp_fltr_info = fm_list->fltr_info;
+               struct ice_vsi_list_map_info *vsi_list_info =
+                       fm_list->vsi_list_info;
+               u16 rem_vsi_handle;
+
+               rem_vsi_handle = ice_find_first_bit(vsi_list_info->vsi_map,
+                                                   ICE_MAX_VSI);
+               if (!ice_is_vsi_valid(hw, rem_vsi_handle))
+                       return ICE_ERR_OUT_OF_RANGE;
+
+               /* Make sure VSI list is empty before removing it below */
+               status = ice_update_vsi_list_rule(hw, &rem_vsi_handle, 1,
+                                                 vsi_list_id, true,
+                                                 ice_aqc_opc_update_sw_rules,
+                                                 lkup_type);
+               if (status)
+                       return status;
+
+               tmp_fltr_info.fltr_act = ICE_FWD_TO_VSI;
+               tmp_fltr_info.fwd_id.hw_vsi_id =
+                       ice_get_hw_vsi_num(hw, rem_vsi_handle);
+               tmp_fltr_info.vsi_handle = rem_vsi_handle;
+               status = ice_update_pkt_fwd_rule(hw, &tmp_fltr_info);
+               if (status) {
+                       ice_debug(hw, ICE_DBG_SW,
+                                 "Failed to update pkt fwd rule to FWD_TO_VSI on HW VSI %d, error %d\n",
+                                 tmp_fltr_info.fwd_id.hw_vsi_id, status);
+                       return status;
+               }
+
+               fm_list->fltr_info = tmp_fltr_info;
+       }
+
+       if ((fm_list->vsi_count == 1 && lkup_type != ICE_SW_LKUP_VLAN) ||
+           (fm_list->vsi_count == 0 && lkup_type == ICE_SW_LKUP_VLAN)) {
+               struct ice_vsi_list_map_info *vsi_list_info =
+                       fm_list->vsi_list_info;
+
+               /* Remove the VSI list since it is no longer used */
+               status = ice_remove_vsi_list_rule(hw, vsi_list_id, lkup_type);
+               if (status) {
+                       ice_debug(hw, ICE_DBG_SW,
+                                 "Failed to remove VSI list %d, error %d\n",
+                                 vsi_list_id, status);
+                       return status;
+               }
+
+               LIST_DEL(&vsi_list_info->list_entry);
+               ice_free(hw, vsi_list_info);
+               fm_list->vsi_list_info = NULL;
+       }
+
+       return status;
+}
+
+/**
+ * ice_remove_rule_internal - Remove a filter rule of a given type
+ *
+ * @hw: pointer to the hardware structure
+ * @recp_id: recipe id for which the rule needs to removed
+ * @f_entry: rule entry containing filter information
+ */
+static enum ice_status
+ice_remove_rule_internal(struct ice_hw *hw, u8 recp_id,
+                        struct ice_fltr_list_entry *f_entry)
+{
+       struct ice_switch_info *sw = hw->switch_info;
+       struct ice_fltr_mgmt_list_entry *list_elem;
+       struct ice_lock *rule_lock; /* Lock to protect filter rule list */
+       enum ice_status status = ICE_SUCCESS;
+       bool remove_rule = false;
+       u16 vsi_handle;
+
+       if (!ice_is_vsi_valid(hw, f_entry->fltr_info.vsi_handle))
+               return ICE_ERR_PARAM;
+       f_entry->fltr_info.fwd_id.hw_vsi_id =
+               ice_get_hw_vsi_num(hw, f_entry->fltr_info.vsi_handle);
+
+       rule_lock = &sw->recp_list[recp_id].filt_rule_lock;
+       ice_acquire_lock(rule_lock);
+       list_elem = ice_find_rule_entry(hw, recp_id, &f_entry->fltr_info);
+       if (!list_elem) {
+               status = ICE_ERR_DOES_NOT_EXIST;
+               goto exit;
+       }
+
+       if (list_elem->fltr_info.fltr_act != ICE_FWD_TO_VSI_LIST) {
+               remove_rule = true;
+       } else if (!list_elem->vsi_list_info) {
+               status = ICE_ERR_DOES_NOT_EXIST;
+               goto exit;
+       } else if (list_elem->vsi_list_info->ref_cnt > 1) {
+               /* a ref_cnt > 1 indicates that the vsi_list is being
+                * shared by multiple rules. Decrement the ref_cnt and
+                * remove this rule, but do not modify the list, as it
+                * is in-use by other rules.
+                */
+               list_elem->vsi_list_info->ref_cnt--;
+               remove_rule = true;
+       } else {
+               /* a ref_cnt of 1 indicates the vsi_list is only used
+                * by one rule. However, the original removal request is only
+                * for a single VSI. Update the vsi_list first, and only
+                * remove the rule if there are no further VSIs in this list.
+                */
+               vsi_handle = f_entry->fltr_info.vsi_handle;
+               status = ice_rem_update_vsi_list(hw, vsi_handle, list_elem);
+               if (status)
+                       goto exit;
+               /* if vsi count goes to zero after updating the vsi list */
+               if (list_elem->vsi_count == 0)
+                       remove_rule = true;
+       }
+
+       if (remove_rule) {
+               /* Remove the lookup rule */
+               struct ice_aqc_sw_rules_elem *s_rule;
+
+               s_rule = (struct ice_aqc_sw_rules_elem *)
+                       ice_malloc(hw, ICE_SW_RULE_RX_TX_NO_HDR_SIZE);
+               if (!s_rule) {
+                       status = ICE_ERR_NO_MEMORY;
+                       goto exit;
+               }
+
+               ice_fill_sw_rule(hw, &list_elem->fltr_info, s_rule,
+                                ice_aqc_opc_remove_sw_rules);
+
+               status = ice_aq_sw_rules(hw, s_rule,
+                                        ICE_SW_RULE_RX_TX_NO_HDR_SIZE, 1,
+                                        ice_aqc_opc_remove_sw_rules, NULL);
+               if (status)
+                       goto exit;
+
+               /* Remove a book keeping from the list */
+               ice_free(hw, s_rule);
+
+               LIST_DEL(&list_elem->list_entry);
+               ice_free(hw, list_elem);
+       }
+exit:
+       ice_release_lock(rule_lock);
+       return status;
+}
+
+
+/**
+ * ice_add_mac - Add a MAC address based filter rule
+ * @hw: pointer to the hardware structure
+ * @m_list: list of MAC addresses and forwarding information
+ *
+ * IMPORTANT: When the ucast_shared flag is set to false and m_list has
+ * multiple unicast addresses, the function assumes that all the
+ * addresses are unique in a given add_mac call. It doesn't
+ * check for duplicates in this case, removing duplicates from a given
+ * list should be taken care of in the caller of this function.
+ */
+enum ice_status
+ice_add_mac(struct ice_hw *hw, struct LIST_HEAD_TYPE *m_list)
+{
+       struct ice_aqc_sw_rules_elem *s_rule, *r_iter;
+       struct ice_fltr_list_entry *m_list_itr;
+       struct LIST_HEAD_TYPE *rule_head;
+       u16 elem_sent, total_elem_left;
+       struct ice_switch_info *sw;
+       struct ice_lock *rule_lock; /* Lock to protect filter rule list */
+       enum ice_status status = ICE_SUCCESS;
+       u16 num_unicast = 0;
+       u16 s_rule_size;
+
+       if (!m_list || !hw)
+               return ICE_ERR_PARAM;
+       s_rule = NULL;
+       sw = hw->switch_info;
+       rule_lock = &sw->recp_list[ICE_SW_LKUP_MAC].filt_rule_lock;
+       LIST_FOR_EACH_ENTRY(m_list_itr, m_list, ice_fltr_list_entry,
+                           list_entry) {
+               u8 *add = &m_list_itr->fltr_info.l_data.mac.mac_addr[0];
+               u16 vsi_handle;
+               u16 hw_vsi_id;
+
+               m_list_itr->fltr_info.flag = ICE_FLTR_TX;
+               vsi_handle = m_list_itr->fltr_info.vsi_handle;
+               if (!ice_is_vsi_valid(hw, vsi_handle))
+                       return ICE_ERR_PARAM;
+               hw_vsi_id = ice_get_hw_vsi_num(hw, vsi_handle);
+               m_list_itr->fltr_info.fwd_id.hw_vsi_id = hw_vsi_id;
+               /* update the src in case it is vsi num */
+               if (m_list_itr->fltr_info.src_id != ICE_SRC_ID_VSI)
+                       return ICE_ERR_PARAM;
+               m_list_itr->fltr_info.src = hw_vsi_id;
+               if (m_list_itr->fltr_info.lkup_type != ICE_SW_LKUP_MAC ||
+                   IS_ZERO_ETHER_ADDR(add))
+                       return ICE_ERR_PARAM;
+               if (IS_UNICAST_ETHER_ADDR(add) && !hw->ucast_shared) {
+                       /* Don't overwrite the unicast address */
+                       ice_acquire_lock(rule_lock);
+                       if (ice_find_rule_entry(hw, ICE_SW_LKUP_MAC,
+                                               &m_list_itr->fltr_info)) {
+                               ice_release_lock(rule_lock);
+                               return ICE_ERR_ALREADY_EXISTS;
+                       }
+                       ice_release_lock(rule_lock);
+                       num_unicast++;
+               } else if (IS_MULTICAST_ETHER_ADDR(add) ||
+                          (IS_UNICAST_ETHER_ADDR(add) && hw->ucast_shared)) {
+                       m_list_itr->status =
+                               ice_add_rule_internal(hw, ICE_SW_LKUP_MAC,
+                                                     m_list_itr);
+                       if (m_list_itr->status)
+                               return m_list_itr->status;
+               }
+       }
+
+       ice_acquire_lock(rule_lock);
+       /* Exit if no suitable entries were found for adding bulk switch rule */
+       if (!num_unicast) {
+               status = ICE_SUCCESS;
+               goto ice_add_mac_exit;
+       }
+
+       rule_head = &sw->recp_list[ICE_SW_LKUP_MAC].filt_rules;
+
+       /* Allocate switch rule buffer for the bulk update for unicast */
+       s_rule_size = ICE_SW_RULE_RX_TX_ETH_HDR_SIZE;
+       s_rule = (struct ice_aqc_sw_rules_elem *)
+               ice_calloc(hw, num_unicast, s_rule_size);
+       if (!s_rule) {
+               status = ICE_ERR_NO_MEMORY;
+               goto ice_add_mac_exit;
+       }
+
+       r_iter = s_rule;
+       LIST_FOR_EACH_ENTRY(m_list_itr, m_list, ice_fltr_list_entry,
+                           list_entry) {
+               struct ice_fltr_info *f_info = &m_list_itr->fltr_info;
+               u8 *mac_addr = &f_info->l_data.mac.mac_addr[0];
+
+               if (IS_UNICAST_ETHER_ADDR(mac_addr)) {
+                       ice_fill_sw_rule(hw, &m_list_itr->fltr_info, r_iter,
+                                        ice_aqc_opc_add_sw_rules);
+                       r_iter = (struct ice_aqc_sw_rules_elem *)
+                               ((u8 *)r_iter + s_rule_size);
+               }
+       }
+
+       /* Call AQ bulk switch rule update for all unicast addresses */
+       r_iter = s_rule;
+       /* Call AQ switch rule in AQ_MAX chunk */
+       for (total_elem_left = num_unicast; total_elem_left > 0;
+            total_elem_left -= elem_sent) {
+               struct ice_aqc_sw_rules_elem *entry = r_iter;
+
+               elem_sent = min(total_elem_left,
+                               (u16)(ICE_AQ_MAX_BUF_LEN / s_rule_size));
+               status = ice_aq_sw_rules(hw, entry, elem_sent * s_rule_size,
+                                        elem_sent, ice_aqc_opc_add_sw_rules,
+                                        NULL);
+               if (status)
+                       goto ice_add_mac_exit;
+               r_iter = (struct ice_aqc_sw_rules_elem *)
+                       ((u8 *)r_iter + (elem_sent * s_rule_size));
+       }
+
+       /* Fill up rule id based on the value returned from FW */
+       r_iter = s_rule;
+       LIST_FOR_EACH_ENTRY(m_list_itr, m_list, ice_fltr_list_entry,
+                           list_entry) {
+               struct ice_fltr_info *f_info = &m_list_itr->fltr_info;
+               u8 *mac_addr = &f_info->l_data.mac.mac_addr[0];
+               struct ice_fltr_mgmt_list_entry *fm_entry;
+
+               if (IS_UNICAST_ETHER_ADDR(mac_addr)) {
+                       f_info->fltr_rule_id =
+                               LE16_TO_CPU(r_iter->pdata.lkup_tx_rx.index);
+                       f_info->fltr_act = ICE_FWD_TO_VSI;
+                       /* Create an entry to track this MAC address */
+                       fm_entry = (struct ice_fltr_mgmt_list_entry *)
+                               ice_malloc(hw, sizeof(*fm_entry));
+                       if (!fm_entry) {
+                               status = ICE_ERR_NO_MEMORY;
+                               goto ice_add_mac_exit;
+                       }
+                       fm_entry->fltr_info = *f_info;
+                       fm_entry->vsi_count = 1;
+                       /* The book keeping entries will get removed when
+                        * base driver calls remove filter AQ command
+                        */
+
+                       LIST_ADD(&fm_entry->list_entry, rule_head);
+                       r_iter = (struct ice_aqc_sw_rules_elem *)
+                               ((u8 *)r_iter + s_rule_size);
+               }
+       }
+
+ice_add_mac_exit:
+       ice_release_lock(rule_lock);
+       if (s_rule)
+               ice_free(hw, s_rule);
+       return status;
+}
+
+/**
+ * ice_add_vlan_internal - Add one VLAN based filter rule
+ * @hw: pointer to the hardware structure
+ * @f_entry: filter entry containing one VLAN information
+ */
+static enum ice_status
+ice_add_vlan_internal(struct ice_hw *hw, struct ice_fltr_list_entry *f_entry)
+{
+       struct ice_switch_info *sw = hw->switch_info;
+       struct ice_fltr_mgmt_list_entry *v_list_itr;
+       struct ice_fltr_info *new_fltr, *cur_fltr;
+       enum ice_sw_lkup_type lkup_type;
+       u16 vsi_list_id = 0, vsi_handle;
+       struct ice_lock *rule_lock; /* Lock to protect filter rule list */
+       enum ice_status status = ICE_SUCCESS;
+
+       if (!ice_is_vsi_valid(hw, f_entry->fltr_info.vsi_handle))
+               return ICE_ERR_PARAM;
+
+       f_entry->fltr_info.fwd_id.hw_vsi_id =
+               ice_get_hw_vsi_num(hw, f_entry->fltr_info.vsi_handle);
+       new_fltr = &f_entry->fltr_info;
+
+       /* VLAN id should only be 12 bits */
+       if (new_fltr->l_data.vlan.vlan_id > ICE_MAX_VLAN_ID)
+               return ICE_ERR_PARAM;
+
+       if (new_fltr->src_id != ICE_SRC_ID_VSI)
+               return ICE_ERR_PARAM;
+
+       new_fltr->src = new_fltr->fwd_id.hw_vsi_id;
+       lkup_type = new_fltr->lkup_type;
+       vsi_handle = new_fltr->vsi_handle;
+       rule_lock = &sw->recp_list[ICE_SW_LKUP_VLAN].filt_rule_lock;
+       ice_acquire_lock(rule_lock);
+       v_list_itr = ice_find_rule_entry(hw, ICE_SW_LKUP_VLAN, new_fltr);
+       if (!v_list_itr) {
+               struct ice_vsi_list_map_info *map_info = NULL;
+
+               if (new_fltr->fltr_act == ICE_FWD_TO_VSI) {
+                       /* All VLAN pruning rules use a VSI list. Check if
+                        * there is already a VSI list containing VSI that we
+                        * want to add. If found, use the same vsi_list_id for
+                        * this new VLAN rule or else create a new list.
+                        */
+                       map_info = ice_find_vsi_list_entry(hw, ICE_SW_LKUP_VLAN,
+                                                          vsi_handle,
+                                                          &vsi_list_id);
+                       if (!map_info) {
+                               status = ice_create_vsi_list_rule(hw,
+                                                                 &vsi_handle,
+                                                                 1,
+                                                                 &vsi_list_id,
+                                                                 lkup_type);
+                               if (status)
+                                       goto exit;
+                       }
+                       /* Convert the action to forwarding to a VSI list. */
+                       new_fltr->fltr_act = ICE_FWD_TO_VSI_LIST;
+                       new_fltr->fwd_id.vsi_list_id = vsi_list_id;
+               }
+
+               status = ice_create_pkt_fwd_rule(hw, f_entry);
+               if (!status) {
+                       v_list_itr = ice_find_rule_entry(hw, ICE_SW_LKUP_VLAN,
+                                                        new_fltr);
+                       if (!v_list_itr) {
+                               status = ICE_ERR_DOES_NOT_EXIST;
+                               goto exit;
+                       }
+                       /* reuse VSI list for new rule and increment ref_cnt */
+                       if (map_info) {
+                               v_list_itr->vsi_list_info = map_info;
+                               map_info->ref_cnt++;
+                       } else {
+                               v_list_itr->vsi_list_info =
+                                       ice_create_vsi_list_map(hw, &vsi_handle,
+                                                               1, vsi_list_id);
+                       }
+               }
+       } else if (v_list_itr->vsi_list_info->ref_cnt == 1) {
+               /* Update existing VSI list to add new VSI id only if it used
+                * by one VLAN rule.
+                */
+               cur_fltr = &v_list_itr->fltr_info;
+               status = ice_add_update_vsi_list(hw, v_list_itr, cur_fltr,
+                                                new_fltr);
+       } else {
+               /* If VLAN rule exists and VSI list being used by this rule is
+                * referenced by more than 1 VLAN rule. Then create a new VSI
+                * list appending previous VSI with new VSI and update existing
+                * VLAN rule to point to new VSI list id
+                */
+               struct ice_fltr_info tmp_fltr;
+               u16 vsi_handle_arr[2];
+               u16 cur_handle;
+
+               /* Current implementation only supports reusing VSI list with
+                * one VSI count. We should never hit below condition
+                */
+               if (v_list_itr->vsi_count > 1 &&
+                   v_list_itr->vsi_list_info->ref_cnt > 1) {
+                       ice_debug(hw, ICE_DBG_SW,
+                                 "Invalid configuration: Optimization to reuse VSI list with more than one VSI is not being done yet\n");
+                       status = ICE_ERR_CFG;
+                       goto exit;
+               }
+
+               cur_handle =
+                       ice_find_first_bit(v_list_itr->vsi_list_info->vsi_map,
+                                          ICE_MAX_VSI);
+
+               /* A rule already exists with the new VSI being added */
+               if (cur_handle == vsi_handle) {
+                       status = ICE_ERR_ALREADY_EXISTS;
+                       goto exit;
+               }
+
+               vsi_handle_arr[0] = cur_handle;
+               vsi_handle_arr[1] = vsi_handle;
+               status = ice_create_vsi_list_rule(hw, &vsi_handle_arr[0], 2,
+                                                 &vsi_list_id, lkup_type);
+               if (status)
+                       goto exit;
+
+               tmp_fltr = v_list_itr->fltr_info;
+               tmp_fltr.fltr_rule_id = v_list_itr->fltr_info.fltr_rule_id;
+               tmp_fltr.fwd_id.vsi_list_id = vsi_list_id;
+               tmp_fltr.fltr_act = ICE_FWD_TO_VSI_LIST;
+               /* Update the previous switch rule to a new VSI list which
+                * includes current VSI that is requested
+                */
+               status = ice_update_pkt_fwd_rule(hw, &tmp_fltr);
+               if (status)
+                       goto exit;
+
+               /* before overriding VSI list map info. decrement ref_cnt of
+                * previous VSI list
+                */
+               v_list_itr->vsi_list_info->ref_cnt--;
+
+               /* now update to newly created list */
+               v_list_itr->fltr_info.fwd_id.vsi_list_id = vsi_list_id;
+               v_list_itr->vsi_list_info =
+                       ice_create_vsi_list_map(hw, &vsi_handle_arr[0], 2,
+                                               vsi_list_id);
+               v_list_itr->vsi_count++;
+       }
+
+exit:
+       ice_release_lock(rule_lock);
+       return status;
+}
+
+/**
+ * ice_add_vlan - Add VLAN based filter rule
+ * @hw: pointer to the hardware structure
+ * @v_list: list of VLAN entries and forwarding information
+ */
+enum ice_status
+ice_add_vlan(struct ice_hw *hw, struct LIST_HEAD_TYPE *v_list)
+{
+       struct ice_fltr_list_entry *v_list_itr;
+
+       if (!v_list || !hw)
+               return ICE_ERR_PARAM;
+
+       LIST_FOR_EACH_ENTRY(v_list_itr, v_list, ice_fltr_list_entry,
+                           list_entry) {
+               if (v_list_itr->fltr_info.lkup_type != ICE_SW_LKUP_VLAN)
+                       return ICE_ERR_PARAM;
+               v_list_itr->fltr_info.flag = ICE_FLTR_TX;
+               v_list_itr->status = ice_add_vlan_internal(hw, v_list_itr);
+               if (v_list_itr->status)
+                       return v_list_itr->status;
+       }
+       return ICE_SUCCESS;
+}
+
+#ifndef NO_MACVLAN_SUPPORT
+/**
+ * ice_add_mac_vlan - Add MAC and VLAN pair based filter rule
+ * @hw: pointer to the hardware structure
+ * @mv_list: list of MAC and VLAN filters
+ *
+ * If the VSI on which the mac-vlan pair has to be added has RX and Tx VLAN
+ * pruning bits enabled, then it is the responsibility of the caller to make
+ * sure to add a vlan only filter on the same VSI. Packets belonging to that
+ * VLAN won't be received on that VSI otherwise.
+ */
+enum ice_status
+ice_add_mac_vlan(struct ice_hw *hw, struct LIST_HEAD_TYPE *mv_list)
+{
+       struct ice_fltr_list_entry *mv_list_itr;
+
+       if (!mv_list || !hw)
+               return ICE_ERR_PARAM;
+
+       LIST_FOR_EACH_ENTRY(mv_list_itr, mv_list, ice_fltr_list_entry,
+                           list_entry) {
+               enum ice_sw_lkup_type l_type =
+                       mv_list_itr->fltr_info.lkup_type;
+
+               if (l_type != ICE_SW_LKUP_MAC_VLAN)
+                       return ICE_ERR_PARAM;
+               mv_list_itr->fltr_info.flag = ICE_FLTR_TX;
+               mv_list_itr->status =
+                       ice_add_rule_internal(hw, ICE_SW_LKUP_MAC_VLAN,
+                                             mv_list_itr);
+               if (mv_list_itr->status)
+                       return mv_list_itr->status;
+       }
+       return ICE_SUCCESS;
+}
+#endif
+
+
+
+/**
+ * ice_rem_sw_rule_info
+ * @hw: pointer to the hardware structure
+ * @rule_head: pointer to the switch list structure that we want to delete
+ */
+static void
+ice_rem_sw_rule_info(struct ice_hw *hw, struct LIST_HEAD_TYPE *rule_head)
+{
+       if (!LIST_EMPTY(rule_head)) {
+               struct ice_fltr_mgmt_list_entry *entry;
+               struct ice_fltr_mgmt_list_entry *tmp;
+
+               LIST_FOR_EACH_ENTRY_SAFE(entry, tmp, rule_head,
+                                        ice_fltr_mgmt_list_entry, list_entry) {
+                       LIST_DEL(&entry->list_entry);
+                       ice_free(hw, entry);
+               }
+       }
+}
+
+
+
+/**
+ * ice_cfg_dflt_vsi - change state of VSI to set/clear default
+ * @pi: pointer to the port_info structure
+ * @vsi_handle: VSI handle to set as default
+ * @set: true to add the above mentioned switch rule, false to remove it
+ * @direction: ICE_FLTR_RX or ICE_FLTR_TX
+ *
+ * add filter rule to set/unset given VSI as default VSI for the switch
+ * (represented by swid)
+ */
+enum ice_status
+ice_cfg_dflt_vsi(struct ice_port_info *pi, u16 vsi_handle, bool set,
+                u8 direction)
+{
+       struct ice_aqc_sw_rules_elem *s_rule;
+       struct ice_fltr_info f_info;
+       struct ice_hw *hw = pi->hw;
+       enum ice_adminq_opc opcode;
+       enum ice_status status;
+       u16 s_rule_size;
+       u16 hw_vsi_id;
+
+       if (!ice_is_vsi_valid(hw, vsi_handle))
+               return ICE_ERR_PARAM;
+       hw_vsi_id = ice_get_hw_vsi_num(hw, vsi_handle);
+
+       s_rule_size = set ? ICE_SW_RULE_RX_TX_ETH_HDR_SIZE :
+                           ICE_SW_RULE_RX_TX_NO_HDR_SIZE;
+       s_rule = (struct ice_aqc_sw_rules_elem *)ice_malloc(hw, s_rule_size);
+       if (!s_rule)
+               return ICE_ERR_NO_MEMORY;
+
+       ice_memset(&f_info, 0, sizeof(f_info), ICE_NONDMA_MEM);
+
+       f_info.lkup_type = ICE_SW_LKUP_DFLT;
+       f_info.flag = direction;
+       f_info.fltr_act = ICE_FWD_TO_VSI;
+       f_info.fwd_id.hw_vsi_id = hw_vsi_id;
+
+       if (f_info.flag & ICE_FLTR_RX) {
+               f_info.src = pi->lport;
+               f_info.src_id = ICE_SRC_ID_LPORT;
+               if (!set)
+                       f_info.fltr_rule_id =
+                               pi->dflt_rx_vsi_rule_id;
+       } else if (f_info.flag & ICE_FLTR_TX) {
+               f_info.src_id = ICE_SRC_ID_VSI;
+               f_info.src = hw_vsi_id;
+               if (!set)
+                       f_info.fltr_rule_id =
+                               pi->dflt_tx_vsi_rule_id;
+       }
+
+       if (set)
+               opcode = ice_aqc_opc_add_sw_rules;
+       else
+               opcode = ice_aqc_opc_remove_sw_rules;
+
+       ice_fill_sw_rule(hw, &f_info, s_rule, opcode);
+
+       status = ice_aq_sw_rules(hw, s_rule, s_rule_size, 1, opcode, NULL);
+       if (status || !(f_info.flag & ICE_FLTR_TX_RX))
+               goto out;
+       if (set) {
+               u16 index = LE16_TO_CPU(s_rule->pdata.lkup_tx_rx.index);
+
+               if (f_info.flag & ICE_FLTR_TX) {
+                       pi->dflt_tx_vsi_num = hw_vsi_id;
+                       pi->dflt_tx_vsi_rule_id = index;
+               } else if (f_info.flag & ICE_FLTR_RX) {
+                       pi->dflt_rx_vsi_num = hw_vsi_id;
+                       pi->dflt_rx_vsi_rule_id = index;
+               }
+       } else {
+               if (f_info.flag & ICE_FLTR_TX) {
+                       pi->dflt_tx_vsi_num = ICE_DFLT_VSI_INVAL;
+                       pi->dflt_tx_vsi_rule_id = ICE_INVAL_ACT;
+               } else if (f_info.flag & ICE_FLTR_RX) {
+                       pi->dflt_rx_vsi_num = ICE_DFLT_VSI_INVAL;
+                       pi->dflt_rx_vsi_rule_id = ICE_INVAL_ACT;
+               }
+       }
+
+out:
+       ice_free(hw, s_rule);
+       return status;
+}
+
+/**
+ * ice_remove_mac - remove a MAC address based filter rule
+ * @hw: pointer to the hardware structure
+ * @m_list: list of MAC addresses and forwarding information
+ *
+ * This function removes either a MAC filter rule or a specific VSI from a
+ * VSI list for a multicast MAC address.
+ *
+ * Returns ICE_ERR_DOES_NOT_EXIST if a given entry was not added by
+ * ice_add_mac. Caller should be aware that this call will only work if all
+ * the entries passed into m_list were added previously. It will not attempt to
+ * do a partial remove of entries that were found.
+ */
+enum ice_status
+ice_remove_mac(struct ice_hw *hw, struct LIST_HEAD_TYPE *m_list)
+{
+       struct ice_fltr_list_entry *list_itr, *tmp;
+
+       if (!m_list)
+               return ICE_ERR_PARAM;
+
+       LIST_FOR_EACH_ENTRY_SAFE(list_itr, tmp, m_list, ice_fltr_list_entry,
+                                list_entry) {
+               enum ice_sw_lkup_type l_type = list_itr->fltr_info.lkup_type;
+
+               if (l_type != ICE_SW_LKUP_MAC)
+                       return ICE_ERR_PARAM;
+               list_itr->status = ice_remove_rule_internal(hw,
+                                                           ICE_SW_LKUP_MAC,
+                                                           list_itr);
+               if (list_itr->status)
+                       return list_itr->status;
+       }
+       return ICE_SUCCESS;
+}
+
+/**
+ * ice_remove_vlan - Remove VLAN based filter rule
+ * @hw: pointer to the hardware structure
+ * @v_list: list of VLAN entries and forwarding information
+ */
+enum ice_status
+ice_remove_vlan(struct ice_hw *hw, struct LIST_HEAD_TYPE *v_list)
+{
+       struct ice_fltr_list_entry *v_list_itr, *tmp;
+
+       if (!v_list || !hw)
+               return ICE_ERR_PARAM;
+
+       LIST_FOR_EACH_ENTRY_SAFE(v_list_itr, tmp, v_list, ice_fltr_list_entry,
+                                list_entry) {
+               enum ice_sw_lkup_type l_type = v_list_itr->fltr_info.lkup_type;
+
+               if (l_type != ICE_SW_LKUP_VLAN)
+                       return ICE_ERR_PARAM;
+               v_list_itr->status = ice_remove_rule_internal(hw,
+                                                             ICE_SW_LKUP_VLAN,
+                                                             v_list_itr);
+               if (v_list_itr->status)
+                       return v_list_itr->status;
+       }
+       return ICE_SUCCESS;
+}
+
+#ifndef NO_MACVLAN_SUPPORT
+/**
+ * ice_remove_mac_vlan - Remove MAC VLAN based filter rule
+ * @hw: pointer to the hardware structure
+ * @v_list: list of MAC VLAN entries and forwarding information
+ */
+enum ice_status
+ice_remove_mac_vlan(struct ice_hw *hw, struct LIST_HEAD_TYPE *v_list)
+{
+       struct ice_fltr_list_entry *v_list_itr, *tmp;
+
+       if (!v_list || !hw)
+               return ICE_ERR_PARAM;
+
+       LIST_FOR_EACH_ENTRY_SAFE(v_list_itr, tmp, v_list, ice_fltr_list_entry,
+                                list_entry) {
+               enum ice_sw_lkup_type l_type = v_list_itr->fltr_info.lkup_type;
+
+               if (l_type != ICE_SW_LKUP_MAC_VLAN)
+                       return ICE_ERR_PARAM;
+               v_list_itr->status =
+                       ice_remove_rule_internal(hw, ICE_SW_LKUP_MAC_VLAN,
+                                                v_list_itr);
+               if (v_list_itr->status)
+                       return v_list_itr->status;
+       }
+       return ICE_SUCCESS;
+}
+#endif /* !NO_MACVLAN_SUPPORT */
+
+/**
+ * ice_vsi_uses_fltr - Determine if given VSI uses specified filter
+ * @fm_entry: filter entry to inspect
+ * @vsi_handle: VSI handle to compare with filter info
+ */
+static bool
+ice_vsi_uses_fltr(struct ice_fltr_mgmt_list_entry *fm_entry, u16 vsi_handle)
+{
+       return ((fm_entry->fltr_info.fltr_act == ICE_FWD_TO_VSI &&
+                fm_entry->fltr_info.vsi_handle == vsi_handle) ||
+               (fm_entry->fltr_info.fltr_act == ICE_FWD_TO_VSI_LIST &&
+                (ice_is_bit_set(fm_entry->vsi_list_info->vsi_map,
+                                vsi_handle))));
+}
+
+/**
+ * ice_add_entry_to_vsi_fltr_list - Add copy of fltr_list_entry to remove list
+ * @hw: pointer to the hardware structure
+ * @vsi_handle: VSI handle to remove filters from
+ * @vsi_list_head: pointer to the list to add entry to
+ * @fi: pointer to fltr_info of filter entry to copy & add
+ *
+ * Helper function, used when creating a list of filters to remove from
+ * a specific VSI. The entry added to vsi_list_head is a COPY of the
+ * original filter entry, with the exception of fltr_info.fltr_act and
+ * fltr_info.fwd_id fields. These are set such that later logic can
+ * extract which VSI to remove the fltr from, and pass on that information.
+ */
+static enum ice_status
+ice_add_entry_to_vsi_fltr_list(struct ice_hw *hw, u16 vsi_handle,
+                              struct LIST_HEAD_TYPE *vsi_list_head,
+                              struct ice_fltr_info *fi)
+{
+       struct ice_fltr_list_entry *tmp;
+
+       /* this memory is freed up in the caller function
+        * once filters for this VSI are removed
+        */
+       tmp = (struct ice_fltr_list_entry *)ice_malloc(hw, sizeof(*tmp));
+       if (!tmp)
+               return ICE_ERR_NO_MEMORY;
+
+       tmp->fltr_info = *fi;
+
+       /* Overwrite these fields to indicate which VSI to remove filter from,
+        * so find and remove logic can extract the information from the
+        * list entries. Note that original entries will still have proper
+        * values.
+        */
+       tmp->fltr_info.fltr_act = ICE_FWD_TO_VSI;
+       tmp->fltr_info.vsi_handle = vsi_handle;
+       tmp->fltr_info.fwd_id.hw_vsi_id = ice_get_hw_vsi_num(hw, vsi_handle);
+
+       LIST_ADD(&tmp->list_entry, vsi_list_head);
+
+       return ICE_SUCCESS;
+}
+
+/**
+ * ice_add_to_vsi_fltr_list - Add VSI filters to the list
+ * @hw: pointer to the hardware structure
+ * @vsi_handle: VSI handle to remove filters from
+ * @lkup_list_head: pointer to the list that has certain lookup type filters
+ * @vsi_list_head: pointer to the list pertaining to VSI with vsi_handle
+ *
+ * Locates all filters in lkup_list_head that are used by the given VSI,
+ * and adds COPIES of those entries to vsi_list_head (intended to be used
+ * to remove the listed filters).
+ * Note that this means all entries in vsi_list_head must be explicitly
+ * deallocated by the caller when done with list.
+ */
+static enum ice_status
+ice_add_to_vsi_fltr_list(struct ice_hw *hw, u16 vsi_handle,
+                        struct LIST_HEAD_TYPE *lkup_list_head,
+                        struct LIST_HEAD_TYPE *vsi_list_head)
+{
+       struct ice_fltr_mgmt_list_entry *fm_entry;
+       enum ice_status status = ICE_SUCCESS;
+
+       /* check to make sure VSI id is valid and within boundary */
+       if (!ice_is_vsi_valid(hw, vsi_handle))
+               return ICE_ERR_PARAM;
+
+       LIST_FOR_EACH_ENTRY(fm_entry, lkup_list_head,
+                           ice_fltr_mgmt_list_entry, list_entry) {
+               struct ice_fltr_info *fi;
+
+               fi = &fm_entry->fltr_info;
+               if (!fi || !ice_vsi_uses_fltr(fm_entry, vsi_handle))
+                       continue;
+
+               status = ice_add_entry_to_vsi_fltr_list(hw, vsi_handle,
+                                                       vsi_list_head, fi);
+               if (status)
+                       return status;
+       }
+       return status;
+}
+
+
+/**
+ * ice_determine_promisc_mask
+ * @fi: filter info to parse
+ *
+ * Helper function to determine which ICE_PROMISC_ mask corresponds
+ * to given filter into.
+ */
+static u8 ice_determine_promisc_mask(struct ice_fltr_info *fi)
+{
+       u16 vid = fi->l_data.mac_vlan.vlan_id;
+       u8 *macaddr = fi->l_data.mac.mac_addr;
+       bool is_tx_fltr = false;
+       u8 promisc_mask = 0;
+
+       if (fi->flag == ICE_FLTR_TX)
+               is_tx_fltr = true;
+
+       if (IS_BROADCAST_ETHER_ADDR(macaddr))
+               promisc_mask |= is_tx_fltr ?
+                       ICE_PROMISC_BCAST_TX : ICE_PROMISC_BCAST_RX;
+       else if (IS_MULTICAST_ETHER_ADDR(macaddr))
+               promisc_mask |= is_tx_fltr ?
+                       ICE_PROMISC_MCAST_TX : ICE_PROMISC_MCAST_RX;
+       else if (IS_UNICAST_ETHER_ADDR(macaddr))
+               promisc_mask |= is_tx_fltr ?
+                       ICE_PROMISC_UCAST_TX : ICE_PROMISC_UCAST_RX;
+       if (vid)
+               promisc_mask |= is_tx_fltr ?
+                       ICE_PROMISC_VLAN_TX : ICE_PROMISC_VLAN_RX;
+
+       return promisc_mask;
+}
+
+
+/**
+ * ice_remove_promisc - Remove promisc based filter rules
+ * @hw: pointer to the hardware structure
+ * @recp_id: recipe id for which the rule needs to removed
+ * @v_list: list of promisc entries
+ */
+static enum ice_status
+ice_remove_promisc(struct ice_hw *hw, u8 recp_id,
+                  struct LIST_HEAD_TYPE *v_list)
+{
+       struct ice_fltr_list_entry *v_list_itr, *tmp;
+
+       LIST_FOR_EACH_ENTRY_SAFE(v_list_itr, tmp, v_list, ice_fltr_list_entry,
+                                list_entry) {
+               v_list_itr->status =
+                       ice_remove_rule_internal(hw, recp_id, v_list_itr);
+               if (v_list_itr->status)
+                       return v_list_itr->status;
+       }
+       return ICE_SUCCESS;
+}
+
+/**
+ * ice_clear_vsi_promisc - clear specified promiscuous mode(s) for given VSI
+ * @hw: pointer to the hardware structure
+ * @vsi_handle: VSI handle to clear mode
+ * @promisc_mask: mask of promiscuous config bits to clear
+ * @vid: VLAN ID to clear VLAN promiscuous
+ */
+enum ice_status
+ice_clear_vsi_promisc(struct ice_hw *hw, u16 vsi_handle, u8 promisc_mask,
+                     u16 vid)
+{
+       struct ice_switch_info *sw = hw->switch_info;
+       struct ice_fltr_list_entry *fm_entry, *tmp;
+       struct LIST_HEAD_TYPE remove_list_head;
+       struct ice_fltr_mgmt_list_entry *itr;
+       struct LIST_HEAD_TYPE *rule_head;
+       struct ice_lock *rule_lock;     /* Lock to protect filter rule list */
+       enum ice_status status = ICE_SUCCESS;
+       u8 recipe_id;
+
+       if (!ice_is_vsi_valid(hw, vsi_handle))
+               return ICE_ERR_PARAM;
+
+       if (vid)
+               recipe_id = ICE_SW_LKUP_PROMISC_VLAN;
+       else
+               recipe_id = ICE_SW_LKUP_PROMISC;
+
+       rule_head = &sw->recp_list[recipe_id].filt_rules;
+       rule_lock = &sw->recp_list[recipe_id].filt_rule_lock;
+
+       INIT_LIST_HEAD(&remove_list_head);
+
+       ice_acquire_lock(rule_lock);
+       LIST_FOR_EACH_ENTRY(itr, rule_head,
+                           ice_fltr_mgmt_list_entry, list_entry) {
+               u8 fltr_promisc_mask = 0;
+
+               if (!ice_vsi_uses_fltr(itr, vsi_handle))
+                       continue;
+
+               fltr_promisc_mask |=
+                       ice_determine_promisc_mask(&itr->fltr_info);
+
+               /* Skip if filter is not completely specified by given mask */
+               if (fltr_promisc_mask & ~promisc_mask)
+                       continue;
+
+               status = ice_add_entry_to_vsi_fltr_list(hw, vsi_handle,
+                                                       &remove_list_head,
+                                                       &itr->fltr_info);
+               if (status) {
+                       ice_release_lock(rule_lock);
+                       goto free_fltr_list;
+               }
+       }
+       ice_release_lock(rule_lock);
+
+       status = ice_remove_promisc(hw, recipe_id, &remove_list_head);
+
+free_fltr_list:
+       LIST_FOR_EACH_ENTRY_SAFE(fm_entry, tmp, &remove_list_head,
+                                ice_fltr_list_entry, list_entry) {
+               LIST_DEL(&fm_entry->list_entry);
+               ice_free(hw, fm_entry);
+       }
+
+       return status;
+}
+
+/**
+ * ice_set_vsi_promisc - set given VSI to given promiscuous mode(s)
+ * @hw: pointer to the hardware structure
+ * @vsi_handle: VSI handle to configure
+ * @promisc_mask: mask of promiscuous config bits
+ * @vid: VLAN ID to set VLAN promiscuous
+ */
+enum ice_status
+ice_set_vsi_promisc(struct ice_hw *hw, u16 vsi_handle, u8 promisc_mask, u16 vid)
+{
+       enum { UCAST_FLTR = 1, MCAST_FLTR, BCAST_FLTR };
+       struct ice_fltr_list_entry f_list_entry;
+       struct ice_fltr_info new_fltr;
+       enum ice_status status = ICE_SUCCESS;
+       bool is_tx_fltr;
+       u16 hw_vsi_id;
+       int pkt_type;
+       u8 recipe_id;
+
+       ice_debug(hw, ICE_DBG_TRACE, "ice_set_vsi_promisc\n");
+
+       if (!ice_is_vsi_valid(hw, vsi_handle))
+               return ICE_ERR_PARAM;
+       hw_vsi_id = ice_get_hw_vsi_num(hw, vsi_handle);
+
+       ice_memset(&new_fltr, 0, sizeof(new_fltr), ICE_NONDMA_MEM);
+
+       if (promisc_mask & (ICE_PROMISC_VLAN_RX | ICE_PROMISC_VLAN_TX)) {
+               new_fltr.lkup_type = ICE_SW_LKUP_PROMISC_VLAN;
+               new_fltr.l_data.mac_vlan.vlan_id = vid;
+               recipe_id = ICE_SW_LKUP_PROMISC_VLAN;
+       } else {
+               new_fltr.lkup_type = ICE_SW_LKUP_PROMISC;
+               recipe_id = ICE_SW_LKUP_PROMISC;
+       }
+
+       /* Separate filters must be set for each direction/packet type
+        * combination, so we will loop over the mask value, store the
+        * individual type, and clear it out in the input mask as it
+        * is found.
+        */
+       while (promisc_mask) {
+               u8 *mac_addr;
+
+               pkt_type = 0;
+               is_tx_fltr = false;
+
+               if (promisc_mask & ICE_PROMISC_UCAST_RX) {
+                       promisc_mask &= ~ICE_PROMISC_UCAST_RX;
+                       pkt_type = UCAST_FLTR;
+               } else if (promisc_mask & ICE_PROMISC_UCAST_TX) {
+                       promisc_mask &= ~ICE_PROMISC_UCAST_TX;
+                       pkt_type = UCAST_FLTR;
+                       is_tx_fltr = true;
+               } else if (promisc_mask & ICE_PROMISC_MCAST_RX) {
+                       promisc_mask &= ~ICE_PROMISC_MCAST_RX;
+                       pkt_type = MCAST_FLTR;
+               } else if (promisc_mask & ICE_PROMISC_MCAST_TX) {
+                       promisc_mask &= ~ICE_PROMISC_MCAST_TX;
+                       pkt_type = MCAST_FLTR;
+                       is_tx_fltr = true;
+               } else if (promisc_mask & ICE_PROMISC_BCAST_RX) {
+                       promisc_mask &= ~ICE_PROMISC_BCAST_RX;
+                       pkt_type = BCAST_FLTR;
+               } else if (promisc_mask & ICE_PROMISC_BCAST_TX) {
+                       promisc_mask &= ~ICE_PROMISC_BCAST_TX;
+                       pkt_type = BCAST_FLTR;
+                       is_tx_fltr = true;
+               }
+
+               /* Check for VLAN promiscuous flag */
+               if (promisc_mask & ICE_PROMISC_VLAN_RX) {
+                       promisc_mask &= ~ICE_PROMISC_VLAN_RX;
+               } else if (promisc_mask & ICE_PROMISC_VLAN_TX) {
+                       promisc_mask &= ~ICE_PROMISC_VLAN_TX;
+                       is_tx_fltr = true;
+               }
+
+               /* Set filter DA based on packet type */
+               mac_addr = new_fltr.l_data.mac.mac_addr;
+               if (pkt_type == BCAST_FLTR) {
+                       ice_memset(mac_addr, 0xff, ETH_ALEN, ICE_NONDMA_MEM);
+               } else if (pkt_type == MCAST_FLTR ||
+                          pkt_type == UCAST_FLTR) {
+                       /* Use the dummy ether header DA */
+                       ice_memcpy(mac_addr, dummy_eth_header, ETH_ALEN,
+                                  ICE_NONDMA_TO_NONDMA);
+                       if (pkt_type == MCAST_FLTR)
+                               mac_addr[0] |= 0x1;     /* Set multicast bit */
+               }
+
+               /* Need to reset this to zero for all iterations */
+               new_fltr.flag = 0;
+               if (is_tx_fltr) {
+                       new_fltr.flag |= ICE_FLTR_TX;
+                       new_fltr.src = hw_vsi_id;
+               } else {
+                       new_fltr.flag |= ICE_FLTR_RX;
+                       new_fltr.src = hw->port_info->lport;
+               }
+
+               new_fltr.fltr_act = ICE_FWD_TO_VSI;
+               new_fltr.vsi_handle = vsi_handle;
+               new_fltr.fwd_id.hw_vsi_id = hw_vsi_id;
+               f_list_entry.fltr_info = new_fltr;
+
+               status = ice_add_rule_internal(hw, recipe_id, &f_list_entry);
+               if (status != ICE_SUCCESS)
+                       goto set_promisc_exit;
+       }
+
+set_promisc_exit:
+       return status;
+}
+
+/**
+ * ice_set_vlan_vsi_promisc
+ * @hw: pointer to the hardware structure
+ * @vsi_handle: VSI handle to configure
+ * @promisc_mask: mask of promiscuous config bits
+ * @rm_vlan_promisc: Clear VLANs VSI promisc mode
+ *
+ * Configure VSI with all associated VLANs to given promiscuous mode(s)
+ */
+enum ice_status
+ice_set_vlan_vsi_promisc(struct ice_hw *hw, u16 vsi_handle, u8 promisc_mask,
+                        bool rm_vlan_promisc)
+{
+       struct ice_switch_info *sw = hw->switch_info;
+       struct ice_fltr_list_entry *list_itr, *tmp;
+       struct LIST_HEAD_TYPE vsi_list_head;
+       struct LIST_HEAD_TYPE *vlan_head;
+       struct ice_lock *vlan_lock; /* Lock to protect filter rule list */
+       enum ice_status status;
+       u16 vlan_id;
+
+       INIT_LIST_HEAD(&vsi_list_head);
+       vlan_lock = &sw->recp_list[ICE_SW_LKUP_VLAN].filt_rule_lock;
+       vlan_head = &sw->recp_list[ICE_SW_LKUP_VLAN].filt_rules;
+       ice_acquire_lock(vlan_lock);
+       status = ice_add_to_vsi_fltr_list(hw, vsi_handle, vlan_head,
+                                         &vsi_list_head);
+       ice_release_lock(vlan_lock);
+       if (status)
+               goto free_fltr_list;
+
+       LIST_FOR_EACH_ENTRY(list_itr, &vsi_list_head, ice_fltr_list_entry,
+                           list_entry) {
+               vlan_id = list_itr->fltr_info.l_data.vlan.vlan_id;
+               if (rm_vlan_promisc)
+                       status = ice_clear_vsi_promisc(hw, vsi_handle,
+                                                      promisc_mask, vlan_id);
+               else
+                       status = ice_set_vsi_promisc(hw, vsi_handle,
+                                                    promisc_mask, vlan_id);
+               if (status)
+                       break;
+       }
+
+free_fltr_list:
+       LIST_FOR_EACH_ENTRY_SAFE(list_itr, tmp, &vsi_list_head,
+                                ice_fltr_list_entry, list_entry) {
+               LIST_DEL(&list_itr->list_entry);
+               ice_free(hw, list_itr);
+       }
+       return status;
+}
+
+/**
+ * ice_remove_vsi_lkup_fltr - Remove lookup type filters for a VSI
+ * @hw: pointer to the hardware structure
+ * @vsi_handle: VSI handle to remove filters from
+ * @lkup: switch rule filter lookup type
+ */
+static void
+ice_remove_vsi_lkup_fltr(struct ice_hw *hw, u16 vsi_handle,
+                        enum ice_sw_lkup_type lkup)
+{
+       struct ice_switch_info *sw = hw->switch_info;
+       struct ice_fltr_list_entry *fm_entry;
+       struct LIST_HEAD_TYPE remove_list_head;
+       struct LIST_HEAD_TYPE *rule_head;
+       struct ice_fltr_list_entry *tmp;
+       struct ice_lock *rule_lock;     /* Lock to protect filter rule list */
+       enum ice_status status;
+
+       INIT_LIST_HEAD(&remove_list_head);
+       rule_lock = &sw->recp_list[lkup].filt_rule_lock;
+       rule_head = &sw->recp_list[lkup].filt_rules;
+       ice_acquire_lock(rule_lock);
+       status = ice_add_to_vsi_fltr_list(hw, vsi_handle, rule_head,
+                                         &remove_list_head);
+       ice_release_lock(rule_lock);
+       if (status)
+               return;
+
+       switch (lkup) {
+       case ICE_SW_LKUP_MAC:
+               ice_remove_mac(hw, &remove_list_head);
+               break;
+       case ICE_SW_LKUP_VLAN:
+               ice_remove_vlan(hw, &remove_list_head);
+               break;
+       case ICE_SW_LKUP_PROMISC:
+       case ICE_SW_LKUP_PROMISC_VLAN:
+               ice_remove_promisc(hw, lkup, &remove_list_head);
+               break;
+       case ICE_SW_LKUP_MAC_VLAN:
+#ifndef NO_MACVLAN_SUPPORT
+               ice_remove_mac_vlan(hw, &remove_list_head);
+#else
+               ice_debug(hw, ICE_DBG_SW, "MAC VLAN look up is not supported yet\n");
+#endif /* !NO_MACVLAN_SUPPORT */
+               break;
+       case ICE_SW_LKUP_ETHERTYPE:
+       case ICE_SW_LKUP_ETHERTYPE_MAC:
+       case ICE_SW_LKUP_DFLT:
+               ice_debug(hw, ICE_DBG_SW,
+                         "Remove filters for this lookup type hasn't been implemented yet\n");
+               break;
+       case ICE_SW_LKUP_LAST:
+               ice_debug(hw, ICE_DBG_SW, "Unsupported lookup type\n");
+               break;
+       }
+
+       LIST_FOR_EACH_ENTRY_SAFE(fm_entry, tmp, &remove_list_head,
+                                ice_fltr_list_entry, list_entry) {
+               LIST_DEL(&fm_entry->list_entry);
+               ice_free(hw, fm_entry);
+       }
+}
+
+/**
+ * ice_remove_vsi_fltr - Remove all filters for a VSI
+ * @hw: pointer to the hardware structure
+ * @vsi_handle: VSI handle to remove filters from
+ */
+void ice_remove_vsi_fltr(struct ice_hw *hw, u16 vsi_handle)
+{
+       ice_debug(hw, ICE_DBG_TRACE, "ice_remove_vsi_fltr\n");
+
+       ice_remove_vsi_lkup_fltr(hw, vsi_handle, ICE_SW_LKUP_MAC);
+       ice_remove_vsi_lkup_fltr(hw, vsi_handle, ICE_SW_LKUP_MAC_VLAN);
+       ice_remove_vsi_lkup_fltr(hw, vsi_handle, ICE_SW_LKUP_PROMISC);
+       ice_remove_vsi_lkup_fltr(hw, vsi_handle, ICE_SW_LKUP_VLAN);
+       ice_remove_vsi_lkup_fltr(hw, vsi_handle, ICE_SW_LKUP_DFLT);
+       ice_remove_vsi_lkup_fltr(hw, vsi_handle, ICE_SW_LKUP_ETHERTYPE);
+       ice_remove_vsi_lkup_fltr(hw, vsi_handle, ICE_SW_LKUP_ETHERTYPE_MAC);
+       ice_remove_vsi_lkup_fltr(hw, vsi_handle, ICE_SW_LKUP_PROMISC_VLAN);
+}
+
+
+
+
+
+/**
+ * ice_replay_vsi_fltr - Replay filters for requested VSI
+ * @hw: pointer to the hardware structure
+ * @vsi_handle: driver vsi handle
+ * @recp_id: Recipe id for which rules need to be replayed
+ * @list_head: list for which filters need to be replayed
+ *
+ * Replays the filter of recipe recp_id for a VSI represented via vsi_handle.
+ * It is required to pass valid VSI handle.
+ */
+static enum ice_status
+ice_replay_vsi_fltr(struct ice_hw *hw, u16 vsi_handle, u8 recp_id,
+                   struct LIST_HEAD_TYPE *list_head)
+{
+       struct ice_fltr_mgmt_list_entry *itr;
+       enum ice_status status = ICE_SUCCESS;
+       u16 hw_vsi_id;
+
+       if (LIST_EMPTY(list_head))
+               return status;
+       hw_vsi_id = ice_get_hw_vsi_num(hw, vsi_handle);
+
+       LIST_FOR_EACH_ENTRY(itr, list_head, ice_fltr_mgmt_list_entry,
+                           list_entry) {
+               struct ice_fltr_list_entry f_entry;
+
+               f_entry.fltr_info = itr->fltr_info;
+               if (itr->vsi_count < 2 && recp_id != ICE_SW_LKUP_VLAN &&
+                   itr->fltr_info.vsi_handle == vsi_handle) {
+                       /* update the src in case it is vsi num */
+                       if (f_entry.fltr_info.src_id == ICE_SRC_ID_VSI)
+                               f_entry.fltr_info.src = hw_vsi_id;
+                       status = ice_add_rule_internal(hw, recp_id, &f_entry);
+                       if (status != ICE_SUCCESS)
+                               goto end;
+                       continue;
+               }
+               if (!itr->vsi_list_info ||
+                   !ice_is_bit_set(itr->vsi_list_info->vsi_map, vsi_handle))
+                       continue;
+               /* Clearing it so that the logic can add it back */
+               ice_clear_bit(vsi_handle, itr->vsi_list_info->vsi_map);
+               f_entry.fltr_info.vsi_handle = vsi_handle;
+               f_entry.fltr_info.fltr_act = ICE_FWD_TO_VSI;
+               /* update the src in case it is vsi num */
+               if (f_entry.fltr_info.src_id == ICE_SRC_ID_VSI)
+                       f_entry.fltr_info.src = hw_vsi_id;
+               if (recp_id == ICE_SW_LKUP_VLAN)
+                       status = ice_add_vlan_internal(hw, &f_entry);
+               else
+                       status = ice_add_rule_internal(hw, recp_id, &f_entry);
+               if (status != ICE_SUCCESS)
+                       goto end;
+       }
+end:
+       return status;
+}
+
+
+/**
+ * ice_replay_vsi_all_fltr - replay all filters stored in bookkeeping lists
+ * @hw: pointer to the hardware structure
+ * @vsi_handle: driver vsi handle
+ *
+ * Replays filters for requested VSI via vsi_handle.
+ */
+enum ice_status ice_replay_vsi_all_fltr(struct ice_hw *hw, u16 vsi_handle)
+{
+       struct ice_switch_info *sw = hw->switch_info;
+       enum ice_status status = ICE_SUCCESS;
+       u8 i;
+
+       for (i = 0; i < ICE_MAX_NUM_RECIPES; i++) {
+               /* Update the default recipe lines and ones that were created */
+               if (i < ICE_MAX_NUM_RECIPES || sw->recp_list[i].recp_created) {
+                       struct LIST_HEAD_TYPE *head;
+
+                       head = &sw->recp_list[i].filt_replay_rules;
+                       if (!sw->recp_list[i].adv_rule)
+                               status = ice_replay_vsi_fltr(hw, vsi_handle, i,
+                                                            head);
+                       if (status != ICE_SUCCESS)
+                               return status;
+               }
+       }
+       return status;
+}
+
+/**
+ * ice_rm_all_sw_replay_rule_info - deletes filter replay rules
+ * @hw: pointer to the hw struct
+ *
+ * Deletes the filter replay rules.
+ */
+void ice_rm_all_sw_replay_rule_info(struct ice_hw *hw)
+{
+       struct ice_switch_info *sw = hw->switch_info;
+       u8 i;
+
+       if (!sw)
+               return;
+
+       for (i = 0; i < ICE_MAX_NUM_RECIPES; i++) {
+               if (!LIST_EMPTY(&sw->recp_list[i].filt_replay_rules)) {
+                       struct LIST_HEAD_TYPE *l_head;
+
+                       l_head = &sw->recp_list[i].filt_replay_rules;
+                       if (!sw->recp_list[i].adv_rule)
+                               ice_rem_sw_rule_info(hw, l_head);
+               }
+       }
+}
diff --git a/drivers/net/ice/base/ice_switch.h b/drivers/net/ice/base/ice_switch.h
new file mode 100644 (file)
index 0000000..66a172f
--- /dev/null
@@ -0,0 +1,333 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2001-2018
+ */
+
+#ifndef _ICE_SWITCH_H_
+#define _ICE_SWITCH_H_
+
+#include "ice_common.h"
+#include "ice_protocol_type.h"
+
+#define ICE_SW_CFG_MAX_BUF_LEN 2048
+#define ICE_MAX_SW 256
+#define ICE_DFLT_VSI_INVAL 0xff
+
+
+
+#define ICE_VSI_INVAL_ID 0xFFFF
+
+/* VSI context structure for add/get/update/free operations */
+struct ice_vsi_ctx {
+       u16 vsi_num;
+       u16 vsis_allocd;
+       u16 vsis_unallocated;
+       u16 flags;
+       struct ice_aqc_vsi_props info;
+       struct ice_sched_vsi_info sched;
+       u8 alloc_from_pool;
+       struct ice_lock rss_locks;      /* protect rss config in VSI ctx */
+       struct LIST_HEAD_TYPE rss_list_head;
+};
+
+
+/* Switch recipe ID enum values are specific to hardware */
+enum ice_sw_lkup_type {
+       ICE_SW_LKUP_ETHERTYPE = 0,
+       ICE_SW_LKUP_MAC = 1,
+       ICE_SW_LKUP_MAC_VLAN = 2,
+       ICE_SW_LKUP_PROMISC = 3,
+       ICE_SW_LKUP_VLAN = 4,
+       ICE_SW_LKUP_DFLT = 5,
+       ICE_SW_LKUP_ETHERTYPE_MAC = 8,
+       ICE_SW_LKUP_PROMISC_VLAN = 9,
+       ICE_SW_LKUP_LAST,
+};
+
+/* type of filter src id */
+enum ice_src_id {
+       ICE_SRC_ID_UNKNOWN = 0,
+       ICE_SRC_ID_VSI,
+       ICE_SRC_ID_QUEUE,
+       ICE_SRC_ID_LPORT,
+};
+
+struct ice_fltr_info {
+       /* Look up information: how to look up packet */
+       enum ice_sw_lkup_type lkup_type;
+       /* Forward action: filter action to do after lookup */
+       enum ice_sw_fwd_act_type fltr_act;
+       /* rule ID returned by firmware once filter rule is created */
+       u16 fltr_rule_id;
+       u16 flag;
+#define ICE_FLTR_RX            BIT(0)
+#define ICE_FLTR_TX            BIT(1)
+#define ICE_FLTR_TX_RX         (ICE_FLTR_RX | ICE_FLTR_TX)
+
+       /* Source VSI for LOOKUP_TX or source port for LOOKUP_RX */
+       u16 src;
+       enum ice_src_id src_id;
+
+       union {
+               struct {
+                       u8 mac_addr[ETH_ALEN];
+               } mac;
+               struct {
+                       u8 mac_addr[ETH_ALEN];
+                       u16 vlan_id;
+               } mac_vlan;
+               struct {
+                       u16 vlan_id;
+               } vlan;
+               /* Set lkup_type as ICE_SW_LKUP_ETHERTYPE
+                * if just using ethertype as filter. Set lkup_type as
+                * ICE_SW_LKUP_ETHERTYPE_MAC if MAC also needs to be
+                * passed in as filter.
+                */
+               struct {
+                       u16 ethertype;
+                       u8 mac_addr[ETH_ALEN]; /* optional */
+               } ethertype_mac;
+       } l_data; /* Make sure to zero out the memory of l_data before using
+                  * it or only set the data associated with lookup match
+                  * rest everything should be zero
+                  */
+
+       /* Depending on filter action */
+       union {
+               /* queue id in case of ICE_FWD_TO_Q and starting
+                * queue id in case of ICE_FWD_TO_QGRP.
+                */
+               u16 q_id:11;
+               u16 hw_vsi_id:10;
+               u16 vsi_id:10;
+               u16 vsi_list_id:10;
+       } fwd_id;
+
+       /* Sw VSI handle */
+       u16 vsi_handle;
+
+       /* Set to num_queues if action is ICE_FWD_TO_QGRP. This field
+        * determines the range of queues the packet needs to be forwarded to.
+        * Note that qgrp_size must be set to a power of 2.
+        */
+       u8 qgrp_size;
+
+       /* Rule creations populate these indicators basing on the switch type */
+       u8 lb_en;       /* Indicate if packet can be looped back */
+       u8 lan_en;      /* Indicate if packet can be forwarded to the uplink */
+};
+
+struct ice_adv_lkup_elem {
+       enum ice_protocol_type type;
+       union ice_prot_hdr h_u; /* Header values */
+       union ice_prot_hdr m_u; /* Mask of header values to match */
+};
+
+struct ice_sw_act_ctrl {
+       /* Source VSI for LOOKUP_TX or source port for LOOKUP_RX */
+       u16 src;
+       u16 flag;
+#define ICE_FLTR_RX             BIT(0)
+#define ICE_FLTR_TX             BIT(1)
+#define ICE_FLTR_TX_RX (ICE_FLTR_RX | ICE_FLTR_TX)
+
+       enum ice_sw_fwd_act_type fltr_act;
+       /* Depending on filter action */
+       union {
+               /* This is a queue id in case of ICE_FWD_TO_Q and starting
+                * queue id in case of ICE_FWD_TO_QGRP.
+                */
+               u16 q_id:11;
+               u16 vsi_id:10;
+               u16 hw_vsi_id:10;
+               u16 vsi_list_id:10;
+       } fwd_id;
+       /* software VSI handle */
+       u16 vsi_handle;
+       u8 qgrp_size;
+};
+
+struct ice_adv_rule_info {
+       enum ice_sw_tunnel_type tun_type;
+       struct ice_sw_act_ctrl sw_act;
+       u32 priority;
+       u8 rx; /* true means LOOKUP_RX otherwise LOOKUP_TX */
+};
+
+/* A collection of one or more four word recipe */
+struct ice_sw_recipe {
+       /* For a chained recipe the root recipe is what should be used for
+        * programming rules
+        */
+       u8 root_rid;
+       u8 recp_created;
+
+       /* Number of extraction words */
+       u8 n_ext_words;
+       /* Protocol ID and Offset pair (extraction word) to describe the
+        * recipe
+        */
+       struct ice_fv_word ext_words[ICE_MAX_CHAIN_WORDS];
+
+       /* if this recipe is a collection of other recipe */
+       u8 big_recp;
+
+       /* if this recipe is part of another bigger recipe then chain index
+        * corresponding to this recipe
+        */
+       u8 chain_idx;
+
+       /* if this recipe is a collection of other recipe then count of other
+        * recipes and recipe ids of those recipes
+        */
+       u8 n_grp_count;
+
+       /* Bit map specifying the IDs associated with this group of recipe */
+       ice_declare_bitmap(r_bitmap, ICE_MAX_NUM_RECIPES);
+
+       enum ice_sw_tunnel_type tun_type;
+
+       /* List of type ice_fltr_mgmt_list_entry or adv_rule */
+       u8 adv_rule;
+       struct LIST_HEAD_TYPE filt_rules;
+       struct LIST_HEAD_TYPE filt_replay_rules;
+
+       struct ice_lock filt_rule_lock; /* protect filter rule structure */
+
+       /* Profiles this recipe should be associated with */
+       struct LIST_HEAD_TYPE fv_list;
+
+       /* Profiles this recipe is associated with */
+       u8 num_profs, *prof_ids;
+
+       /* This allows user to specify the recipe priority.
+        * For now, this becomes 'fwd_priority' when recipe
+        * is created, usually recipes can have 'fwd' and 'join'
+        * priority.
+        */
+       u8 priority;
+
+       struct LIST_HEAD_TYPE rg_list;
+
+       /* AQ buffer associated with this recipe */
+       struct ice_aqc_recipe_data_elem *root_buf;
+};
+
+/* Bookkeeping structure to hold bitmap of VSIs corresponding to VSI list id */
+struct ice_vsi_list_map_info {
+       struct LIST_ENTRY_TYPE list_entry;
+       ice_declare_bitmap(vsi_map, ICE_MAX_VSI);
+       u16 vsi_list_id;
+       /* counter to track how many rules are reusing this VSI list */
+       u16 ref_cnt;
+};
+
+struct ice_fltr_list_entry {
+       struct LIST_ENTRY_TYPE list_entry;
+       enum ice_status status;
+       struct ice_fltr_info fltr_info;
+};
+
+/* This defines an entry in the list that maintains MAC or VLAN membership
+ * to HW list mapping, since multiple VSIs can subscribe to the same MAC or
+ * VLAN. As an optimization the VSI list should be created only when a
+ * second VSI becomes a subscriber to the same MAC address. VSI lists are always
+ * used for VLAN membership.
+ */
+struct ice_fltr_mgmt_list_entry {
+       /* back pointer to VSI list id to VSI list mapping */
+       struct ice_vsi_list_map_info *vsi_list_info;
+       u16 vsi_count;
+#define ICE_INVAL_LG_ACT_INDEX 0xffff
+       u16 lg_act_idx;
+#define ICE_INVAL_SW_MARKER_ID 0xffff
+       u16 sw_marker_id;
+       struct LIST_ENTRY_TYPE list_entry;
+       struct ice_fltr_info fltr_info;
+#define ICE_INVAL_COUNTER_ID 0xff
+       u8 counter_index;
+};
+
+struct ice_adv_fltr_mgmt_list_entry {
+       struct LIST_ENTRY_TYPE list_entry;
+
+       struct ice_adv_lkup_elem *lkups;
+       struct ice_adv_rule_info rule_info;
+       u16 lkups_cnt;
+};
+
+enum ice_promisc_flags {
+       ICE_PROMISC_UCAST_RX = 0x1,
+       ICE_PROMISC_UCAST_TX = 0x2,
+       ICE_PROMISC_MCAST_RX = 0x4,
+       ICE_PROMISC_MCAST_TX = 0x8,
+       ICE_PROMISC_BCAST_RX = 0x10,
+       ICE_PROMISC_BCAST_TX = 0x20,
+       ICE_PROMISC_VLAN_RX = 0x40,
+       ICE_PROMISC_VLAN_TX = 0x80,
+};
+
+/* VSI related commands */
+enum ice_status
+ice_add_vsi(struct ice_hw *hw, u16 vsi_handle, struct ice_vsi_ctx *vsi_ctx,
+           struct ice_sq_cd *cd);
+enum ice_status
+ice_free_vsi(struct ice_hw *hw, u16 vsi_handle, struct ice_vsi_ctx *vsi_ctx,
+            bool keep_vsi_alloc, struct ice_sq_cd *cd);
+enum ice_status
+ice_update_vsi(struct ice_hw *hw, u16 vsi_handle, struct ice_vsi_ctx *vsi_ctx,
+              struct ice_sq_cd *cd);
+struct ice_vsi_ctx *ice_get_vsi_ctx(struct ice_hw *hw, u16 vsi_handle);
+void ice_clear_all_vsi_ctx(struct ice_hw *hw);
+/* Switch config */
+enum ice_status ice_get_initial_sw_cfg(struct ice_hw *hw);
+
+enum ice_status
+ice_alloc_vlan_res_counter(struct ice_hw *hw, u16 *counter_id);
+enum ice_status
+ice_free_vlan_res_counter(struct ice_hw *hw, u16 counter_id);
+
+/* Switch/bridge related commands */
+enum ice_status ice_update_sw_rule_bridge_mode(struct ice_hw *hw);
+enum ice_status
+ice_add_vlan(struct ice_hw *hw, struct LIST_HEAD_TYPE *m_list);
+enum ice_status ice_add_mac(struct ice_hw *hw, struct LIST_HEAD_TYPE *m_lst);
+enum ice_status ice_remove_mac(struct ice_hw *hw, struct LIST_HEAD_TYPE *m_lst);
+enum ice_status
+ice_remove_vlan(struct ice_hw *hw, struct LIST_HEAD_TYPE *v_list);
+#ifndef NO_MACVLAN_SUPPORT
+enum ice_status
+ice_add_mac_vlan(struct ice_hw *hw, struct LIST_HEAD_TYPE *m_list);
+enum ice_status
+ice_remove_mac_vlan(struct ice_hw *hw, struct LIST_HEAD_TYPE *v_list);
+#endif /* !NO_MACVLAN_SUPPORT */
+
+void ice_remove_vsi_fltr(struct ice_hw *hw, u16 vsi_handle);
+
+
+/* Promisc/defport setup for VSIs */
+enum ice_status
+ice_cfg_dflt_vsi(struct ice_port_info *pi, u16 vsi_handle, bool set,
+                u8 direction);
+enum ice_status
+ice_set_vsi_promisc(struct ice_hw *hw, u16 vsi_handle, u8 promisc_mask,
+                   u16 vid);
+enum ice_status
+ice_clear_vsi_promisc(struct ice_hw *hw, u16 vsi_handle, u8 promisc_mask,
+                     u16 vid);
+enum ice_status
+ice_set_vlan_vsi_promisc(struct ice_hw *hw, u16 vsi_handle, u8 promisc_mask,
+                        bool rm_vlan_promisc);
+
+
+
+
+
+enum ice_status ice_init_def_sw_recp(struct ice_hw *hw);
+u16 ice_get_hw_vsi_num(struct ice_hw *hw, u16 vsi_handle);
+bool ice_is_vsi_valid(struct ice_hw *hw, u16 vsi_handle);
+
+enum ice_status ice_replay_vsi_all_fltr(struct ice_hw *hw, u16 vsi_handle);
+void ice_rm_all_sw_replay_rule_info(struct ice_hw *hw);
+
+#endif /* _ICE_SWITCH_H_ */