[Intel-wired-lan] [PATCH net-next v1 0/9] advanced switch rule management

[Intel-wired-lan] [PATCH net-next v1 0/9] advanced switch rule management

[Wojciech Drewek]

This series adds support for adding/removing advanced switch filters
in ice driver. Advanced filters are building blocks for HW acceleration
of TC orchestration. Add ndo_setup_tc callback implementation for PF and
VF port representors (when device is configured in switchdev mode).

Define dummy packet headers to allow adding advanced rules in HW.
Supported headers, and thus filters, are:
- MAC + IPv4 + UDP
- MAC + VLAN + IPv4 + UDP
- MAC + IPv4 + TCP
- MAC + VLAN + IPv4 + TCP
- MAC + IPv6 + UDP
- MAC + VLAN + IPv6 + UDP
- MAC + IPv6 + TCP
- MAC + VLAN + IPv6 + TCP

Dan Nowlin (2):
  ice: manage profiles and field vectors
  ice: create advanced switch recipe

Grishma Kotecha (2):
  ice: implement low level recipes functions
  ice: allow adding advanced rules

Kiran Patil (1):
  ice: ndo_setup_tc implementation for PF

Michal Swiatkowski (2):
  ice: Allow changing lan_en and lb_en on all kinds of filters
  ice: ndo_setup_tc implementation for PR

Shivanshu Shukla (1):
  ice: allow deleting advanced rules

Victor Raj (1):
  ice: cleanup rules info

 drivers/net/ethernet/intel/ice/Makefile            |    3 +-
 drivers/net/ethernet/intel/ice/ice.h               |    4 +
 drivers/net/ethernet/intel/ice/ice_adminq_cmd.h    |   57 +
 drivers/net/ethernet/intel/ice/ice_common.c        |   42 +-
 drivers/net/ethernet/intel/ice/ice_eswitch.c       |    3 +
 drivers/net/ethernet/intel/ice/ice_flex_pipe.c     |  278 ++-
 drivers/net/ethernet/intel/ice/ice_flex_pipe.h     |   14 +
 drivers/net/ethernet/intel/ice/ice_flex_type.h     |   13 +
 drivers/net/ethernet/intel/ice/ice_fltr.c          |  127 +
 drivers/net/ethernet/intel/ice/ice_main.c          |   68 +
 drivers/net/ethernet/intel/ice/ice_protocol_type.h |  169 ++
 drivers/net/ethernet/intel/ice/ice_repr.c          |   53 +
 drivers/net/ethernet/intel/ice/ice_switch.c        | 2541 +++++++++++++++++++-
 drivers/net/ethernet/intel/ice/ice_switch.h        |  139 +-
 drivers/net/ethernet/intel/ice/ice_tc_lib.c        |  855 +++++++
 drivers/net/ethernet/intel/ice/ice_tc_lib.h        |  130 +
 drivers/net/ethernet/intel/ice/ice_type.h          |    4 +
 17 files changed, 4402 insertions(+), 98 deletions(-)
 create mode 100644 drivers/net/ethernet/intel/ice/ice_tc_lib.c
 create mode 100644 drivers/net/ethernet/intel/ice/ice_tc_lib.h

-- 
1.8.3.1

[Intel-wired-lan] [PATCH net-next v6 1/9] ice: implement low level recipes functions

[Wojciech Drewek]

From: Grishma Kotecha <grishma.kotecha@intel.com>

Add code to manage recipes and profiles on admin queue layer.

Allow the driver to add a new recipe and update an existing one. Get a
recipe and get a recipe to profile association is mostly used in update
existing recipes code.

Only default recipes can be updated. An update is done by reading recipes
from HW, changing their params and calling add recipe command.

Support following admin queue commands:
- ice_aqc_opc_add_recipe (0x0290) - create a recipe with protocol
header information and other details that determine how this recipe
filter works
- ice_aqc_opc_recipe_to_profile (0x0291) - associate a switch recipe
to a profile
- ice_aqc_opc_get_recipe (0x0292) - get details of an exsisting recipe
- ice_aqc_opc_get_recipe_to_profile (0x0293) - get a recipe associated
with profile ID

Define ICE_AQC_RES_TYPE_RECIPE resource type to hold a switch
recipe. It is needed when a new switch recipe needs to be created.

Co-developed-by: Dan Nowlin <dan.nowlin@intel.com>
Signed-off-by: Dan Nowlin <dan.nowlin@intel.com>
Signed-off-by: Grishma Kotecha  <grishma.kotecha@intel.com>
Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_adminq_cmd.h |  57 +++++++
 drivers/net/ethernet/intel/ice/ice_switch.c     | 213 ++++++++++++++++++++++++
 drivers/net/ethernet/intel/ice/ice_switch.h     |   9 +
 3 files changed, 279 insertions(+)

diff --git a/drivers/net/ethernet/intel/ice/ice_adminq_cmd.h b/drivers/net/ethernet/intel/ice/ice_adminq_cmd.h
index f05d78319d11..6c727745bb29 100644
--- a/drivers/net/ethernet/intel/ice/ice_adminq_cmd.h
+++ b/drivers/net/ethernet/intel/ice/ice_adminq_cmd.h
@@ -233,6 +233,7 @@ struct ice_aqc_get_sw_cfg_resp_elem {
  */
 #define ICE_AQC_RES_TYPE_VSI_LIST_REP			0x03
 #define ICE_AQC_RES_TYPE_VSI_LIST_PRUNE			0x04
+#define ICE_AQC_RES_TYPE_RECIPE				0x05
 #define ICE_AQC_RES_TYPE_FDIR_COUNTER_BLOCK		0x21
 #define ICE_AQC_RES_TYPE_FDIR_GUARANTEED_ENTRIES	0x22
 #define ICE_AQC_RES_TYPE_FDIR_SHARED_ENTRIES		0x23
@@ -241,6 +242,7 @@ struct ice_aqc_get_sw_cfg_resp_elem {
 #define ICE_AQC_RES_TYPE_HASH_PROF_BLDR_PROFID		0x60
 #define ICE_AQC_RES_TYPE_HASH_PROF_BLDR_TCAM		0x61
 
+#define ICE_AQC_RES_TYPE_FLAG_SHARED			BIT(7)
 #define ICE_AQC_RES_TYPE_FLAG_SCAN_BOTTOM		BIT(12)
 #define ICE_AQC_RES_TYPE_FLAG_IGNORE_INDEX		BIT(13)
 
@@ -474,6 +476,53 @@ struct ice_aqc_vsi_props {
 
 #define ICE_MAX_NUM_RECIPES 64
 
+/* Add/Get Recipe (indirect 0x0290/0x0292) */
+struct ice_aqc_add_get_recipe {
+	__le16 num_sub_recipes;	/* Input in Add cmd, Output in Get cmd */
+	__le16 return_index;	/* Input, used for Get cmd only */
+	u8 reserved[4];
+	__le32 addr_high;
+	__le32 addr_low;
+};
+
+struct ice_aqc_recipe_content {
+	u8 rid;
+#define ICE_AQ_RECIPE_ID_IS_ROOT	BIT(7)
+#define ICE_AQ_SW_ID_LKUP_IDX		0
+	u8 lkup_indx[5];
+#define ICE_AQ_RECIPE_LKUP_IGNORE	BIT(7)
+#define ICE_AQ_SW_ID_LKUP_MASK		0x00FF
+	__le16 mask[5];
+	u8 result_indx;
+#define ICE_AQ_RECIPE_RESULT_DATA_S	0
+#define ICE_AQ_RECIPE_RESULT_DATA_M	(0x3F << ICE_AQ_RECIPE_RESULT_DATA_S)
+#define ICE_AQ_RECIPE_RESULT_EN		BIT(7)
+	u8 rsvd0[3];
+	u8 act_ctrl_join_priority;
+	u8 act_ctrl_fwd_priority;
+	u8 act_ctrl;
+#define ICE_AQ_RECIPE_ACT_INV_ACT	BIT(2)
+	u8 rsvd1;
+	__le32 dflt_act;
+};
+
+struct ice_aqc_recipe_data_elem {
+	u8 recipe_indx;
+	u8 resp_bits;
+	u8 rsvd0[2];
+	u8 recipe_bitmap[8];
+	u8 rsvd1[4];
+	struct ice_aqc_recipe_content content;
+	u8 rsvd2[20];
+};
+
+/* Set/Get Recipes to Profile Association (direct 0x0291/0x0293) */
+struct ice_aqc_recipe_to_profile {
+	__le16 profile_id;
+	u8 rsvd[6];
+	DECLARE_BITMAP(recipe_assoc, ICE_MAX_NUM_RECIPES);
+};
+
 /* Add/Update/Remove/Get switch rules (indirect 0x02A0, 0x02A1, 0x02A2, 0x02A3)
  */
 struct ice_aqc_sw_rules {
@@ -1940,6 +1989,8 @@ struct ice_aq_desc {
 		struct ice_aqc_set_port_id_led set_port_id_led;
 		struct ice_aqc_get_sw_cfg get_sw_conf;
 		struct ice_aqc_sw_rules sw_rules;
+		struct ice_aqc_add_get_recipe add_get_recipe;
+		struct ice_aqc_recipe_to_profile recipe_to_profile;
 		struct ice_aqc_get_topo get_topo;
 		struct ice_aqc_sched_elem_cmd sched_elem_cmd;
 		struct ice_aqc_query_txsched_res query_sched_res;
@@ -2048,6 +2099,12 @@ enum ice_adminq_opc {
 	ice_aqc_opc_update_vsi				= 0x0211,
 	ice_aqc_opc_free_vsi				= 0x0213,
 
+	/* recipe commands */
+	ice_aqc_opc_add_recipe				= 0x0290,
+	ice_aqc_opc_recipe_to_profile			= 0x0291,
+	ice_aqc_opc_get_recipe				= 0x0292,
+	ice_aqc_opc_get_recipe_to_profile		= 0x0293,
+
 	/* switch rules population commands */
 	ice_aqc_opc_add_sw_rules			= 0x02A0,
 	ice_aqc_opc_update_sw_rules			= 0x02A1,
diff --git a/drivers/net/ethernet/intel/ice/ice_switch.c b/drivers/net/ethernet/intel/ice/ice_switch.c
index 1e86a6dba454..88ee213cf816 100644
--- a/drivers/net/ethernet/intel/ice/ice_switch.c
+++ b/drivers/net/ethernet/intel/ice/ice_switch.c
@@ -543,6 +543,219 @@ enum ice_status
 	return status;
 }
 
+/**
+ * ice_aq_add_recipe - add switch recipe
+ * @hw: pointer to the HW struct
+ * @s_recipe_list: pointer to switch rule population list
+ * @num_recipes: number of switch recipes in the list
+ * @cd: pointer to command details structure or NULL
+ *
+ * Add(0x0290)
+ */
+static enum ice_status
+ice_aq_add_recipe(struct ice_hw *hw,
+		  struct ice_aqc_recipe_data_elem *s_recipe_list,
+		  u16 num_recipes, struct ice_sq_cd *cd)
+{
+	struct ice_aqc_add_get_recipe *cmd;
+	struct ice_aq_desc desc;
+	u16 buf_size;
+
+	cmd = &desc.params.add_get_recipe;
+	ice_fill_dflt_direct_cmd_desc(&desc, ice_aqc_opc_add_recipe);
+
+	cmd->num_sub_recipes = cpu_to_le16(num_recipes);
+	desc.flags |= cpu_to_le16(ICE_AQ_FLAG_RD);
+
+	buf_size = num_recipes * sizeof(*s_recipe_list);
+
+	return ice_aq_send_cmd(hw, &desc, s_recipe_list, buf_size, cd);
+}
+
+/**
+ * ice_aq_get_recipe - get switch recipe
+ * @hw: pointer to the HW struct
+ * @s_recipe_list: pointer to switch rule population list
+ * @num_recipes: pointer to the number of recipes (input and output)
+ * @recipe_root: root recipe number of recipe(s) to retrieve
+ * @cd: pointer to command details structure or NULL
+ *
+ * Get(0x0292)
+ *
+ * On input, *num_recipes should equal the number of entries in s_recipe_list.
+ * On output, *num_recipes will equal the number of entries returned in
+ * s_recipe_list.
+ *
+ * The caller must supply enough space in s_recipe_list to hold all possible
+ * recipes and *num_recipes must equal ICE_MAX_NUM_RECIPES.
+ */
+static enum ice_status
+ice_aq_get_recipe(struct ice_hw *hw,
+		  struct ice_aqc_recipe_data_elem *s_recipe_list,
+		  u16 *num_recipes, u16 recipe_root, struct ice_sq_cd *cd)
+{
+	struct ice_aqc_add_get_recipe *cmd;
+	struct ice_aq_desc desc;
+	enum ice_status status;
+	u16 buf_size;
+
+	if (*num_recipes != ICE_MAX_NUM_RECIPES)
+		return ICE_ERR_PARAM;
+
+	cmd = &desc.params.add_get_recipe;
+	ice_fill_dflt_direct_cmd_desc(&desc, ice_aqc_opc_get_recipe);
+
+	cmd->return_index = cpu_to_le16(recipe_root);
+	cmd->num_sub_recipes = 0;
+
+	buf_size = *num_recipes * sizeof(*s_recipe_list);
+
+	status = ice_aq_send_cmd(hw, &desc, s_recipe_list, buf_size, cd);
+	*num_recipes = le16_to_cpu(cmd->num_sub_recipes);
+
+	return status;
+}
+
+/**
+ * ice_update_recipe_lkup_idx - update a default recipe based on the lkup_idx
+ * @hw: pointer to the HW struct
+ * @params: parameters used to update the default recipe
+ *
+ * This function only supports updating default recipes and it only supports
+ * updating a single recipe based on the lkup_idx at a time.
+ *
+ * This is done as a read-modify-write operation. First, get the current recipe
+ * contents based on the recipe's ID. Then modify the field vector index and
+ * mask if it's valid at the lkup_idx. Finally, use the add recipe AQ to update
+ * the pre-existing recipe with the modifications.
+ */
+static enum ice_status __maybe_unused
+ice_update_recipe_lkup_idx(struct ice_hw *hw,
+			   struct ice_update_recipe_lkup_idx_params *params)
+{
+	struct ice_aqc_recipe_data_elem *rcp_list;
+	u16 num_recps = ICE_MAX_NUM_RECIPES;
+	enum ice_status status;
+
+	rcp_list = kcalloc(num_recps, sizeof(*rcp_list), GFP_KERNEL);
+	if (!rcp_list)
+		return ICE_ERR_NO_MEMORY;
+
+	/* read current recipe list from firmware */
+	rcp_list->recipe_indx = params->rid;
+	status = ice_aq_get_recipe(hw, rcp_list, &num_recps, params->rid, NULL);
+	if (status) {
+		ice_debug(hw, ICE_DBG_SW, "Failed to get recipe %d, status %d\n",
+			  params->rid, status);
+		goto error_out;
+	}
+
+	/* only modify existing recipe's lkup_idx and mask if valid, while
+	 * leaving all other fields the same, then update the recipe firmware
+	 */
+	rcp_list->content.lkup_indx[params->lkup_idx] = params->fv_idx;
+	if (params->mask_valid)
+		rcp_list->content.mask[params->lkup_idx] =
+			cpu_to_le16(params->mask);
+
+	if (params->ignore_valid)
+		rcp_list->content.lkup_indx[params->lkup_idx] |=
+			ICE_AQ_RECIPE_LKUP_IGNORE;
+
+	status = ice_aq_add_recipe(hw, &rcp_list[0], 1, NULL);
+	if (status)
+		ice_debug(hw, ICE_DBG_SW, "Failed to update recipe %d lkup_idx %d fv_idx %d mask %d mask_valid %s, status %d\n",
+			  params->rid, params->lkup_idx, params->fv_idx,
+			  params->mask, params->mask_valid ? "true" : "false",
+			  status);
+
+error_out:
+	kfree(rcp_list);
+	return status;
+}
+
+/**
+ * ice_aq_map_recipe_to_profile - Map recipe to packet profile
+ * @hw: pointer to the HW struct
+ * @profile_id: package profile ID to associate the recipe with
+ * @r_bitmap: Recipe bitmap filled in and need to be returned as response
+ * @cd: pointer to command details structure or NULL
+ * Recipe to profile association (0x0291)
+ */
+static enum ice_status __maybe_unused
+ice_aq_map_recipe_to_profile(struct ice_hw *hw, u32 profile_id, u8 *r_bitmap,
+			     struct ice_sq_cd *cd)
+{
+	struct ice_aqc_recipe_to_profile *cmd;
+	struct ice_aq_desc desc;
+
+	cmd = &desc.params.recipe_to_profile;
+	ice_fill_dflt_direct_cmd_desc(&desc, ice_aqc_opc_recipe_to_profile);
+	cmd->profile_id = cpu_to_le16(profile_id);
+	/* Set the recipe ID bit in the bitmask to let the device know which
+	 * profile we are associating the recipe to
+	 */
+	memcpy(cmd->recipe_assoc, r_bitmap, sizeof(cmd->recipe_assoc));
+
+	return ice_aq_send_cmd(hw, &desc, NULL, 0, cd);
+}
+
+/**
+ * ice_aq_get_recipe_to_profile - Map recipe to packet profile
+ * @hw: pointer to the HW struct
+ * @profile_id: package profile ID to associate the recipe with
+ * @r_bitmap: Recipe bitmap filled in and need to be returned as response
+ * @cd: pointer to command details structure or NULL
+ * Associate profile ID with given recipe (0x0293)
+ */
+static enum ice_status __maybe_unused
+ice_aq_get_recipe_to_profile(struct ice_hw *hw, u32 profile_id, u8 *r_bitmap,
+			     struct ice_sq_cd *cd)
+{
+	struct ice_aqc_recipe_to_profile *cmd;
+	struct ice_aq_desc desc;
+	enum ice_status status;
+
+	cmd = &desc.params.recipe_to_profile;
+	ice_fill_dflt_direct_cmd_desc(&desc, ice_aqc_opc_get_recipe_to_profile);
+	cmd->profile_id = cpu_to_le16(profile_id);
+
+	status = ice_aq_send_cmd(hw, &desc, NULL, 0, cd);
+	if (!status)
+		memcpy(r_bitmap, cmd->recipe_assoc, sizeof(cmd->recipe_assoc));
+
+	return status;
+}
+
+/**
+ * ice_alloc_recipe - add recipe resource
+ * @hw: pointer to the hardware structure
+ * @rid: recipe ID returned as response to AQ call
+ */
+static enum ice_status __maybe_unused ice_alloc_recipe(struct ice_hw *hw, u16 *rid)
+{
+	struct ice_aqc_alloc_free_res_elem *sw_buf;
+	enum ice_status status;
+	u16 buf_len;
+
+	buf_len = struct_size(sw_buf, elem, 1);
+	sw_buf = kzalloc(buf_len, GFP_KERNEL);
+	if (!sw_buf)
+		return ICE_ERR_NO_MEMORY;
+
+	sw_buf->num_elems = cpu_to_le16(1);
+	sw_buf->res_type = cpu_to_le16((ICE_AQC_RES_TYPE_RECIPE <<
+					ICE_AQC_RES_TYPE_S) |
+					ICE_AQC_RES_TYPE_FLAG_SHARED);
+	status = ice_aq_alloc_free_res(hw, 1, sw_buf, buf_len,
+				       ice_aqc_opc_alloc_res, NULL);
+	if (!status)
+		*rid = le16_to_cpu(sw_buf->elem[0].e.sw_resp);
+	kfree(sw_buf);
+
+	return status;
+}
+
 /* ice_init_port_info - Initialize port_info with switch configuration data
  * @pi: pointer to port_info
  * @vsi_port_num: VSI number or port number
diff --git a/drivers/net/ethernet/intel/ice/ice_switch.h b/drivers/net/ethernet/intel/ice/ice_switch.h
index e6eeffb3dde9..592c564c6de8 100644
--- a/drivers/net/ethernet/intel/ice/ice_switch.h
+++ b/drivers/net/ethernet/intel/ice/ice_switch.h
@@ -125,6 +125,15 @@ struct ice_fltr_info {
 	u8 lan_en;	/* Indicate if packet can be forwarded to the uplink */
 };
 
+struct ice_update_recipe_lkup_idx_params {
+	u16 rid;
+	u16 fv_idx;
+	bool ignore_valid;
+	u16 mask;
+	bool mask_valid;
+	u8 lkup_idx;
+};
+
 struct ice_sw_recipe {
 	struct list_head l_entry;
 
-- 
1.8.3.1

[Intel-wired-lan] [PATCH net-next v6 2/9] ice: manage profiles and field vectors

[Wojciech Drewek]

From: Dan Nowlin <dan.nowlin@intel.com>

Implement functions to manage profiles and field vectors in hardware.

In hardware, there are up to 256 profiles and each of these profiles can
have 48 field vector words. Each field vector word is described by
protocol id and offset in the packet. To add a new recipe all used
profiles need to be searched. If the profile contains all required
protocol ids and offsets from the recipe it can be used. The driver has
to add this profile to recipe association to tell hardware that newly
added recipe is going to be associated with this profile.

The amount of used profiles depend on the package. To avoid searching
across not used profile, max profile id value is calculated at init flow.
The profile is considered as unused when all field vector words in the
profile are invalid (protocol id 0xff and offset 0x1ff).

Profiles are read from the package section ICE_SID_FLD_VEC_SW. Empty
field vector words can be used for recipe results. Store all unused field
vector words in prof_res_bm. It is a 256 elements array (max number of
profiles) each element is a 48 bit bitmap (max number of field vector
words).

For now, support only non-tunnel profiles type.

Co-developed-by: Grishma Kotecha <grishma.kotecha@intel.com>
Signed-off-by: Grishma Kotecha <grishma.kotecha@intel.com>
Signed-off-by: Dan Nowlin <dan.nowlin@intel.com>
Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_flex_pipe.c | 245 +++++++++++++++++++++++++
 drivers/net/ethernet/intel/ice/ice_flex_pipe.h |   8 +
 drivers/net/ethernet/intel/ice/ice_flex_type.h |  13 ++
 drivers/net/ethernet/intel/ice/ice_switch.c    |   3 +-
 drivers/net/ethernet/intel/ice/ice_type.h      |   3 +
 5 files changed, 271 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_flex_pipe.c b/drivers/net/ethernet/intel/ice/ice_flex_pipe.c
index 915a00f73d34..19cbab69ca4c 100644
--- a/drivers/net/ethernet/intel/ice/ice_flex_pipe.c
+++ b/drivers/net/ethernet/intel/ice/ice_flex_pipe.c
@@ -1405,6 +1405,86 @@ static enum ice_status ice_chk_pkg_version(struct ice_pkg_ver *pkg_ver)
 }
 
 /**
+ * ice_sw_fv_handler
+ * @sect_type: section type
+ * @section: pointer to section
+ * @index: index of the field vector entry to be returned
+ * @offset: ptr to variable that receives the offset in the field vector table
+ *
+ * This is a callback function that can be passed to ice_pkg_enum_entry.
+ * This function treats the given section as of type ice_sw_fv_section and
+ * enumerates offset field. "offset" is an index into the field vector table.
+ */
+static void *
+ice_sw_fv_handler(u32 sect_type, void *section, u32 index, u32 *offset)
+{
+	struct ice_sw_fv_section *fv_section = section;
+
+	if (!section || sect_type != ICE_SID_FLD_VEC_SW)
+		return NULL;
+	if (index >= le16_to_cpu(fv_section->count))
+		return NULL;
+	if (offset)
+		/* "index" passed in to this function is relative to a given
+		 * 4k block. To get to the true index into the field vector
+		 * table need to add the relative index to the base_offset
+		 * field of this section
+		 */
+		*offset = le16_to_cpu(fv_section->base_offset) + index;
+	return fv_section->fv + index;
+}
+
+/**
+ * ice_get_prof_index_max - get the max profile index for used profile
+ * @hw: pointer to the HW struct
+ *
+ * Calling this function will get the max profile index for used profile
+ * and store the index number in struct ice_switch_info *switch_info
+ * in HW for following use.
+ */
+static enum ice_status ice_get_prof_index_max(struct ice_hw *hw)
+{
+	u16 prof_index = 0, j, max_prof_index = 0;
+	struct ice_pkg_enum state;
+	struct ice_seg *ice_seg;
+	bool flag = false;
+	struct ice_fv *fv;
+	u32 offset;
+
+	memset(&state, 0, sizeof(state));
+
+	if (!hw->seg)
+		return ICE_ERR_PARAM;
+
+	ice_seg = hw->seg;
+
+	do {
+		fv = ice_pkg_enum_entry(ice_seg, &state, ICE_SID_FLD_VEC_SW,
+					&offset, ice_sw_fv_handler);
+		if (!fv)
+			break;
+		ice_seg = NULL;
+
+		/* in the profile that not be used, the prot_id is set to 0xff
+		 * and the off is set to 0x1ff for all the field vectors.
+		 */
+		for (j = 0; j < hw->blk[ICE_BLK_SW].es.fvw; j++)
+			if (fv->ew[j].prot_id != ICE_PROT_INVALID ||
+			    fv->ew[j].off != ICE_FV_OFFSET_INVAL)
+				flag = true;
+		if (flag && prof_index > max_prof_index)
+			max_prof_index = prof_index;
+
+		prof_index++;
+		flag = false;
+	} while (fv);
+
+	hw->switch_info->max_used_prof_index = max_prof_index;
+
+	return 0;
+}
+
+/**
  * ice_init_pkg - initialize/download package
  * @hw: pointer to the hardware structure
  * @buf: pointer to the package buffer
@@ -1484,6 +1564,7 @@ enum ice_status ice_init_pkg(struct ice_hw *hw, u8 *buf, u32 len)
 		ice_init_pkg_regs(hw);
 		ice_fill_blk_tbls(hw);
 		ice_fill_hw_ptype(hw);
+		ice_get_prof_index_max(hw);
 	} else {
 		ice_debug(hw, ICE_DBG_INIT, "package load failed, %d\n",
 			  status);
@@ -1561,6 +1642,170 @@ static struct ice_buf_build *ice_pkg_buf_alloc(struct ice_hw *hw)
 }
 
 /**
+ * ice_get_sw_fv_bitmap - Get switch field vector bitmap based on profile type
+ * @hw: pointer to hardware structure
+ * @req_profs: type of profiles requested
+ * @bm: pointer to memory for returning the bitmap of field vectors
+ */
+void
+ice_get_sw_fv_bitmap(struct ice_hw *hw, enum ice_prof_type req_profs,
+		     unsigned long *bm)
+{
+	struct ice_pkg_enum state;
+	struct ice_seg *ice_seg;
+	struct ice_fv *fv;
+
+	if (req_profs == ICE_PROF_ALL) {
+		bitmap_set(bm, 0, ICE_MAX_NUM_PROFILES);
+		return;
+	}
+
+	memset(&state, 0, sizeof(state));
+	bitmap_zero(bm, ICE_MAX_NUM_PROFILES);
+	ice_seg = hw->seg;
+	do {
+		enum ice_prof_type prof_type;
+		u32 offset;
+
+		fv = ice_pkg_enum_entry(ice_seg, &state, ICE_SID_FLD_VEC_SW,
+					&offset, ice_sw_fv_handler);
+		ice_seg = NULL;
+
+		if (fv) {
+			prof_type = ICE_PROF_NON_TUN;
+
+			if (req_profs & prof_type)
+				set_bit((u16)offset, bm);
+		}
+	} while (fv);
+}
+
+/**
+ * ice_get_sw_fv_list
+ * @hw: pointer to the HW structure
+ * @prot_ids: field vector to search for with a given protocol ID
+ * @ids_cnt: lookup/protocol count
+ * @bm: bitmap of field vectors to consider
+ * @fv_list: Head of a list
+ *
+ * Finds all the field vector entries from switch block that contain
+ * a given protocol ID and returns a list of structures of type
+ * "ice_sw_fv_list_entry". Every structure in the list has a field vector
+ * definition and profile ID information
+ * NOTE: The caller of the function is responsible for freeing the memory
+ * allocated for every list entry.
+ */
+enum ice_status
+ice_get_sw_fv_list(struct ice_hw *hw, u8 *prot_ids, u16 ids_cnt,
+		   unsigned long *bm, struct list_head *fv_list)
+{
+	struct ice_sw_fv_list_entry *fvl;
+	struct ice_sw_fv_list_entry *tmp;
+	struct ice_pkg_enum state;
+	struct ice_seg *ice_seg;
+	struct ice_fv *fv;
+	u32 offset;
+
+	memset(&state, 0, sizeof(state));
+
+	if (!ids_cnt || !hw->seg)
+		return ICE_ERR_PARAM;
+
+	ice_seg = hw->seg;
+	do {
+		u16 i;
+
+		fv = ice_pkg_enum_entry(ice_seg, &state, ICE_SID_FLD_VEC_SW,
+					&offset, ice_sw_fv_handler);
+		if (!fv)
+			break;
+		ice_seg = NULL;
+
+		/* If field vector is not in the bitmap list, then skip this
+		 * profile.
+		 */
+		if (!test_bit((u16)offset, bm))
+			continue;
+
+		for (i = 0; i < ids_cnt; i++) {
+			int j;
+
+			/* This code assumes that if a switch field vector line
+			 * has a matching protocol, then this line will contain
+			 * the entries necessary to represent every field in
+			 * that protocol header.
+			 */
+			for (j = 0; j < hw->blk[ICE_BLK_SW].es.fvw; j++)
+				if (fv->ew[j].prot_id == prot_ids[i])
+					break;
+			if (j >= hw->blk[ICE_BLK_SW].es.fvw)
+				break;
+			if (i + 1 == ids_cnt) {
+				fvl = devm_kzalloc(ice_hw_to_dev(hw),
+						   sizeof(*fvl), GFP_KERNEL);
+				if (!fvl)
+					goto err;
+				fvl->fv_ptr = fv;
+				fvl->profile_id = offset;
+				list_add(&fvl->list_entry, fv_list);
+				break;
+			}
+		}
+	} while (fv);
+	if (list_empty(fv_list))
+		return ICE_ERR_CFG;
+	return 0;
+
+err:
+	list_for_each_entry_safe(fvl, tmp, fv_list, list_entry) {
+		list_del(&fvl->list_entry);
+		devm_kfree(ice_hw_to_dev(hw), fvl);
+	}
+
+	return ICE_ERR_NO_MEMORY;
+}
+
+/**
+ * ice_init_prof_result_bm - Initialize the profile result index bitmap
+ * @hw: pointer to hardware structure
+ */
+void ice_init_prof_result_bm(struct ice_hw *hw)
+{
+	struct ice_pkg_enum state;
+	struct ice_seg *ice_seg;
+	struct ice_fv *fv;
+
+	memset(&state, 0, sizeof(state));
+
+	if (!hw->seg)
+		return;
+
+	ice_seg = hw->seg;
+	do {
+		u32 off;
+		u16 i;
+
+		fv = ice_pkg_enum_entry(ice_seg, &state, ICE_SID_FLD_VEC_SW,
+					&off, ice_sw_fv_handler);
+		ice_seg = NULL;
+		if (!fv)
+			break;
+
+		bitmap_zero(hw->switch_info->prof_res_bm[off],
+			    ICE_MAX_FV_WORDS);
+
+		/* Determine empty field vector indices, these can be
+		 * used for recipe results. Skip index 0, since it is
+		 * always used for Switch ID.
+		 */
+		for (i = 1; i < ICE_MAX_FV_WORDS; i++)
+			if (fv->ew[i].prot_id == ICE_PROT_INVALID &&
+			    fv->ew[i].off == ICE_FV_OFFSET_INVAL)
+				set_bit(i, hw->switch_info->prof_res_bm[off]);
+	} while (fv);
+}
+
+/**
  * ice_pkg_buf_free
  * @hw: pointer to the HW structure
  * @bld: pointer to pkg build (allocated by ice_pkg_buf_alloc())
diff --git a/drivers/net/ethernet/intel/ice/ice_flex_pipe.h b/drivers/net/ethernet/intel/ice/ice_flex_pipe.h
index 995b33d37b67..2dcb11937205 100644
--- a/drivers/net/ethernet/intel/ice/ice_flex_pipe.h
+++ b/drivers/net/ethernet/intel/ice/ice_flex_pipe.h
@@ -18,6 +18,14 @@
 
 #define ICE_PKG_CNT 4
 
+void
+ice_get_sw_fv_bitmap(struct ice_hw *hw, enum ice_prof_type type,
+		     unsigned long *bm);
+void
+ice_init_prof_result_bm(struct ice_hw *hw);
+enum ice_status
+ice_get_sw_fv_list(struct ice_hw *hw, u8 *prot_ids, u16 ids_cnt,
+		   unsigned long *bm, struct list_head *fv_list);
 bool
 ice_get_open_tunnel_port(struct ice_hw *hw, u16 *port);
 int ice_udp_tunnel_set_port(struct net_device *netdev, unsigned int table,
diff --git a/drivers/net/ethernet/intel/ice/ice_flex_type.h b/drivers/net/ethernet/intel/ice/ice_flex_type.h
index 6358e5776e02..c73fde45b266 100644
--- a/drivers/net/ethernet/intel/ice/ice_flex_type.h
+++ b/drivers/net/ethernet/intel/ice/ice_flex_type.h
@@ -13,6 +13,8 @@ struct ice_fv_word {
 	u8 resvrd;
 } __packed;
 
+#define ICE_MAX_NUM_PROFILES 256
+
 #define ICE_MAX_FV_WORDS 48
 struct ice_fv {
 	struct ice_fv_word ew[ICE_MAX_FV_WORDS];
@@ -302,6 +304,12 @@ struct ice_sw_fv_section {
 	struct ice_fv fv[];
 };
 
+struct ice_sw_fv_list_entry {
+	struct list_head list_entry;
+	u32 profile_id;
+	struct ice_fv *fv_ptr;
+};
+
 /* The BOOST TCAM stores the match packet header in reverse order, meaning
  * the fields are reversed; in addition, this means that the normally big endian
  * fields of the packet are now little endian.
@@ -645,4 +653,9 @@ struct ice_chs_chg {
 };
 
 #define ICE_FLOW_PTYPE_MAX		ICE_XLT1_CNT
+
+enum ice_prof_type {
+	ICE_PROF_NON_TUN = 0x1,
+	ICE_PROF_ALL = 0xFF,
+};
 #endif /* _ICE_FLEX_TYPE_H_ */
diff --git a/drivers/net/ethernet/intel/ice/ice_switch.c b/drivers/net/ethernet/intel/ice/ice_switch.c
index 88ee213cf816..04023d71c8f7 100644
--- a/drivers/net/ethernet/intel/ice/ice_switch.c
+++ b/drivers/net/ethernet/intel/ice/ice_switch.c
@@ -59,10 +59,11 @@ enum ice_status ice_init_def_sw_recp(struct ice_hw *hw)
 	if (!recps)
 		return ICE_ERR_NO_MEMORY;
 
-	for (i = 0; i < ICE_SW_LKUP_LAST; i++) {
+	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);
+		INIT_LIST_HEAD(&recps[i].rg_list);
 		mutex_init(&recps[i].filt_rule_lock);
 	}
 
diff --git a/drivers/net/ethernet/intel/ice/ice_type.h b/drivers/net/ethernet/intel/ice/ice_type.h
index dcaa6bfab484..6b4f064f6c0d 100644
--- a/drivers/net/ethernet/intel/ice/ice_type.h
+++ b/drivers/net/ethernet/intel/ice/ice_type.h
@@ -698,6 +698,9 @@ struct ice_port_info {
 struct ice_switch_info {
 	struct list_head vsi_list_map_head;
 	struct ice_sw_recipe *recp_list;
+	u16 max_used_prof_index;
+
+	DECLARE_BITMAP(prof_res_bm[ICE_MAX_NUM_PROFILES], ICE_MAX_FV_WORDS);
 };
 
 /* FW logging configuration */
-- 
1.8.3.1

[Intel-wired-lan] [PATCH net-next v6 3/9] ice: create advanced switch recipe

[Wojciech Drewek]

From: Dan Nowlin <dan.nowlin@intel.com>

These changes introduce code for creating advanced recipes for the
switch in hardware.

There are a couple of recipes already defined in the HW. They apply to
matching on basic protocol headers, like MAC, VLAN, MACVLAN,
ethertype or direction (promiscuous), etc.. If the user wants to match on
other protocol headers (eg. ip address, src/dst port etc.) or different
variation of already supported protocols, there is a need to create
new, more complex recipe. That new recipe is referred as
'advanced recipe', and the filtering rule created on top of that recipe
is called 'advanced rule'.

One recipe can have up to 5 words, but the first word is always reserved
for match on switch id, so the driver can define up to 4 words for one
recipe. To support recipes with more words up to 5 recipes can be
chained, so 20 words can be programmed for look up.

Input for adding recipe function is a list of protocols to support. Based
on this list correct profile is being chosen. Correct profile means
that it contains all protocol types from a list. Each profile have up to
48 field vector words and each of this word have protocol id and offset.
These two fields need to match with input data for adding recipe
function. If the correct profile can't be found the function returns an
error.

The next step after finding the correct profile is grouping words into
groups. One group can have up to 4 words. This is done to simplify
sending recipes to HW (because recipe also can have up to 4 words).

In case of chaining (so when look up consists of more than 4 words) last
recipe will always have results from the previous recipes used as words.

A recipe to profile map is used to store information about which profile
is associate with this recipe. This map is an array of 64 elements (max
number of recipes) and each element is a 256 bits bitmap (max number of
profiles)

Profile to recipe map is used to store information about which recipe is
associate with this profile. This map is an array of 256 elements (max
number of profiles) and each element is a 64 bits bitmap (max number of
recipes)

Signed-off-by: Dan Nowlin <dan.nowlin@intel.com>
Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_flex_pipe.c     |   33 +-
 drivers/net/ethernet/intel/ice/ice_flex_pipe.h     |    6 +
 drivers/net/ethernet/intel/ice/ice_protocol_type.h |  169 +++
 drivers/net/ethernet/intel/ice/ice_switch.c        | 1112 +++++++++++++++++++-
 drivers/net/ethernet/intel/ice/ice_switch.h        |  110 +-
 5 files changed, 1410 insertions(+), 20 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_flex_pipe.c b/drivers/net/ethernet/intel/ice/ice_flex_pipe.c
index 19cbab69ca4c..df8ed0b40f00 100644
--- a/drivers/net/ethernet/intel/ice/ice_flex_pipe.c
+++ b/drivers/net/ethernet/intel/ice/ice_flex_pipe.c
@@ -810,7 +810,7 @@ static void ice_release_global_cfg_lock(struct ice_hw *hw)
  *
  * This function will request ownership of the change lock.
  */
-static enum ice_status
+enum ice_status
 ice_acquire_change_lock(struct ice_hw *hw, enum ice_aq_res_access_type access)
 {
 	return ice_acquire_res(hw, ICE_CHANGE_LOCK_RES_ID, access,
@@ -823,7 +823,7 @@ static void ice_release_global_cfg_lock(struct ice_hw *hw)
  *
  * This function will release the change lock using the proper Admin Command.
  */
-static void ice_release_change_lock(struct ice_hw *hw)
+void ice_release_change_lock(struct ice_hw *hw)
 {
 	ice_release_res(hw, ICE_CHANGE_LOCK_RES_ID);
 }
@@ -2184,6 +2184,35 @@ int ice_udp_tunnel_unset_port(struct net_device *netdev, unsigned int table,
 	return 0;
 }
 
+/**
+ * ice_find_prot_off - find prot ID and offset pair, based on prof and FV index
+ * @hw: pointer to the hardware structure
+ * @blk: hardware block
+ * @prof: profile ID
+ * @fv_idx: field vector word index
+ * @prot: variable to receive the protocol ID
+ * @off: variable to receive the protocol offset
+ */
+enum ice_status
+ice_find_prot_off(struct ice_hw *hw, enum ice_block blk, u8 prof, u16 fv_idx,
+		  u8 *prot, u16 *off)
+{
+	struct ice_fv_word *fv_ext;
+
+	if (prof >= hw->blk[blk].es.count)
+		return ICE_ERR_PARAM;
+
+	if (fv_idx >= hw->blk[blk].es.fvw)
+		return ICE_ERR_PARAM;
+
+	fv_ext = hw->blk[blk].es.t + (prof * hw->blk[blk].es.fvw);
+
+	*prot = fv_ext[fv_idx].prot_id;
+	*off = fv_ext[fv_idx].off;
+
+	return 0;
+}
+
 /* PTG Management */
 
 /**
diff --git a/drivers/net/ethernet/intel/ice/ice_flex_pipe.h b/drivers/net/ethernet/intel/ice/ice_flex_pipe.h
index 2dcb11937205..e8c16ea672fb 100644
--- a/drivers/net/ethernet/intel/ice/ice_flex_pipe.h
+++ b/drivers/net/ethernet/intel/ice/ice_flex_pipe.h
@@ -18,6 +18,12 @@
 
 #define ICE_PKG_CNT 4
 
+enum ice_status
+ice_acquire_change_lock(struct ice_hw *hw, enum ice_aq_res_access_type access);
+void ice_release_change_lock(struct ice_hw *hw);
+enum ice_status
+ice_find_prot_off(struct ice_hw *hw, enum ice_block blk, u8 prof, u16 fv_idx,
+		  u8 *prot, u16 *off);
 void
 ice_get_sw_fv_bitmap(struct ice_hw *hw, enum ice_prof_type type,
 		     unsigned long *bm);
diff --git a/drivers/net/ethernet/intel/ice/ice_protocol_type.h b/drivers/net/ethernet/intel/ice/ice_protocol_type.h
index 199aa5b71540..0b220dfa7457 100644
--- a/drivers/net/ethernet/intel/ice/ice_protocol_type.h
+++ b/drivers/net/ethernet/intel/ice/ice_protocol_type.h
@@ -3,6 +3,44 @@
 
 #ifndef _ICE_PROTOCOL_TYPE_H_
 #define _ICE_PROTOCOL_TYPE_H_
+#define ICE_IPV6_ADDR_LENGTH 16
+
+/* Each recipe can match up to 5 different fields. Fields to match can be meta-
+ * data, values extracted from packet headers, or results from other recipes.
+ * One of the 5 fields is reserved for matching the switch ID. So, up to 4
+ * recipes can provide intermediate results to another one through chaining,
+ * e.g. recipes 0, 1, 2, and 3 can provide intermediate results to recipe 4.
+ */
+#define ICE_NUM_WORDS_RECIPE 4
+
+/* Max recipes that can be chained */
+#define ICE_MAX_CHAIN_RECIPE 5
+
+/* 1 word reserved for switch ID from allowed 5 words.
+ * So a recipe can have max 4 words. And you can chain 5 such recipes
+ * together. So maximum words that can be programmed for look up is 5 * 4.
+ */
+#define ICE_MAX_CHAIN_WORDS (ICE_NUM_WORDS_RECIPE * ICE_MAX_CHAIN_RECIPE)
+
+/* Field vector index corresponding to chaining */
+#define ICE_CHAIN_FV_INDEX_START 47
+
+enum ice_protocol_type {
+	ICE_MAC_OFOS = 0,
+	ICE_MAC_IL,
+	ICE_ETYPE_OL,
+	ICE_VLAN_OFOS,
+	ICE_IPV4_OFOS,
+	ICE_IPV4_IL,
+	ICE_IPV6_OFOS,
+	ICE_IPV6_IL,
+	ICE_TCP_IL,
+	ICE_UDP_OF,
+	ICE_UDP_ILOS,
+	ICE_SCTP_IL,
+	ICE_PROTOCOL_LAST
+};
+
 /* Decoders for ice_prot_id:
  * - F: First
  * - I: Inner
@@ -35,4 +73,135 @@ enum ice_prot_id {
 	ICE_PROT_META_ID	= 255, /* when offset == metadata */
 	ICE_PROT_INVALID	= 255  /* when offset == ICE_FV_OFFSET_INVAL */
 };
+
+#define ICE_MAC_OFOS_HW		1
+#define ICE_MAC_IL_HW		4
+#define ICE_ETYPE_OL_HW		9
+#define ICE_VLAN_OF_HW		16
+#define ICE_VLAN_OL_HW		17
+#define ICE_IPV4_OFOS_HW	32
+#define ICE_IPV4_IL_HW		33
+#define ICE_IPV6_OFOS_HW	40
+#define ICE_IPV6_IL_HW		41
+#define ICE_TCP_IL_HW		49
+#define ICE_UDP_ILOS_HW		53
+
+#define ICE_UDP_OF_HW	52 /* UDP Tunnels */
+
+#define ICE_TUN_FLAG_FV_IND 2
+
+/* Mapping of software defined protocol ID to hardware defined protocol ID */
+struct ice_protocol_entry {
+	enum ice_protocol_type type;
+	u8 protocol_id;
+};
+
+struct ice_ether_hdr {
+	u8 dst_addr[ETH_ALEN];
+	u8 src_addr[ETH_ALEN];
+};
+
+struct ice_ethtype_hdr {
+	__be16 ethtype_id;
+};
+
+struct ice_ether_vlan_hdr {
+	u8 dst_addr[ETH_ALEN];
+	u8 src_addr[ETH_ALEN];
+	__be32 vlan_id;
+};
+
+struct ice_vlan_hdr {
+	__be16 type;
+	__be16 vlan;
+};
+
+struct ice_ipv4_hdr {
+	u8 version;
+	u8 tos;
+	__be16 total_length;
+	__be16 id;
+	__be16 frag_off;
+	u8 time_to_live;
+	u8 protocol;
+	__be16 check;
+	__be32 src_addr;
+	__be32 dst_addr;
+};
+
+struct ice_ipv6_hdr {
+	__be32 be_ver_tc_flow;
+	__be16 payload_len;
+	u8 next_hdr;
+	u8 hop_limit;
+	u8 src_addr[ICE_IPV6_ADDR_LENGTH];
+	u8 dst_addr[ICE_IPV6_ADDR_LENGTH];
+};
+
+struct ice_sctp_hdr {
+	__be16 src_port;
+	__be16 dst_port;
+	__be32 verification_tag;
+	__be32 check;
+};
+
+struct ice_l4_hdr {
+	__be16 src_port;
+	__be16 dst_port;
+	__be16 len;
+	__be16 check;
+};
+
+union ice_prot_hdr {
+	struct ice_ether_hdr eth_hdr;
+	struct ice_ethtype_hdr ethertype;
+	struct ice_vlan_hdr vlan_hdr;
+	struct ice_ipv4_hdr ipv4_hdr;
+	struct ice_ipv6_hdr ipv6_hdr;
+	struct ice_l4_hdr l4_hdr;
+	struct ice_sctp_hdr sctp_hdr;
+};
+
+/* This is mapping table entry that maps every word within a given protocol
+ * structure to the real byte offset as per the specification of that
+ * protocol header.
+ * for e.g. dst address is 3 words in ethertype header and corresponding bytes
+ * are 0, 2, 3 in the actual packet header and src address is at 4, 6, 8
+ */
+struct ice_prot_ext_tbl_entry {
+	enum ice_protocol_type prot_type;
+	/* Byte offset into header of given protocol type */
+	u8 offs[sizeof(union ice_prot_hdr)];
+};
+
+/* Extractions to be looked up for a given recipe */
+struct ice_prot_lkup_ext {
+	u16 prot_type;
+	u8 n_val_words;
+	/* create a buffer to hold max words per recipe */
+	u16 field_off[ICE_MAX_CHAIN_WORDS];
+	u16 field_mask[ICE_MAX_CHAIN_WORDS];
+
+	struct ice_fv_word fv_words[ICE_MAX_CHAIN_WORDS];
+
+	/* Indicate field offsets that have field vector indices assigned */
+	DECLARE_BITMAP(done, ICE_MAX_CHAIN_WORDS);
+};
+
+struct ice_pref_recipe_group {
+	u8 n_val_pairs;		/* Number of valid pairs */
+	struct ice_fv_word pairs[ICE_NUM_WORDS_RECIPE];
+	u16 mask[ICE_NUM_WORDS_RECIPE];
+};
+
+struct ice_recp_grp_entry {
+	struct list_head l_entry;
+
+#define ICE_INVAL_CHAIN_IND 0xFF
+	u16 rid;
+	u8 chain_idx;
+	u16 fv_idx[ICE_NUM_WORDS_RECIPE];
+	u16 fv_mask[ICE_NUM_WORDS_RECIPE];
+	struct ice_pref_recipe_group r_group;
+};
 #endif /* _ICE_PROTOCOL_TYPE_H_ */
diff --git a/drivers/net/ethernet/intel/ice/ice_switch.c b/drivers/net/ethernet/intel/ice/ice_switch.c
index 04023d71c8f7..8274f2e8df38 100644
--- a/drivers/net/ethernet/intel/ice/ice_switch.c
+++ b/drivers/net/ethernet/intel/ice/ice_switch.c
@@ -42,6 +42,14 @@
 	(offsetof(struct ice_aqc_sw_rules_elem, pdata.vsi_list.vsi) + \
 	 ((n) * sizeof(((struct ice_sw_rule_vsi_list *)0)->vsi[0])))
 
+/* this is a recipe to profile association bitmap */
+static DECLARE_BITMAP(recipe_to_profile[ICE_MAX_NUM_RECIPES],
+			  ICE_MAX_NUM_PROFILES);
+
+/* this is a profile to recipe association bitmap */
+static DECLARE_BITMAP(profile_to_recipe[ICE_MAX_NUM_PROFILES],
+			  ICE_MAX_NUM_RECIPES);
+
 /**
  * ice_init_def_sw_recp - initialize the recipe book keeping tables
  * @hw: pointer to the HW struct
@@ -683,7 +691,7 @@ enum ice_status
  * @cd: pointer to command details structure or NULL
  * Recipe to profile association (0x0291)
  */
-static enum ice_status __maybe_unused
+static enum ice_status
 ice_aq_map_recipe_to_profile(struct ice_hw *hw, u32 profile_id, u8 *r_bitmap,
 			     struct ice_sq_cd *cd)
 {
@@ -709,7 +717,7 @@ enum ice_status
  * @cd: pointer to command details structure or NULL
  * Associate profile ID with given recipe (0x0293)
  */
-static enum ice_status __maybe_unused
+static enum ice_status
 ice_aq_get_recipe_to_profile(struct ice_hw *hw, u32 profile_id, u8 *r_bitmap,
 			     struct ice_sq_cd *cd)
 {
@@ -733,7 +741,7 @@ enum ice_status
  * @hw: pointer to the hardware structure
  * @rid: recipe ID returned as response to AQ call
  */
-static enum ice_status __maybe_unused ice_alloc_recipe(struct ice_hw *hw, u16 *rid)
+static enum ice_status ice_alloc_recipe(struct ice_hw *hw, u16 *rid)
 {
 	struct ice_aqc_alloc_free_res_elem *sw_buf;
 	enum ice_status status;
@@ -757,6 +765,203 @@ static enum ice_status __maybe_unused ice_alloc_recipe(struct ice_hw *hw, u16 *r
 	return status;
 }
 
+/**
+ * ice_get_recp_to_prof_map - updates recipe to profile mapping
+ * @hw: pointer to hardware structure
+ *
+ * This function is used to populate recipe_to_profile matrix where index to
+ * this array is the recipe ID and the element is the mapping of which profiles
+ * is this recipe mapped to.
+ */
+static void ice_get_recp_to_prof_map(struct ice_hw *hw)
+{
+	DECLARE_BITMAP(r_bitmap, ICE_MAX_NUM_RECIPES);
+	u16 i;
+
+	for (i = 0; i < hw->switch_info->max_used_prof_index + 1; i++) {
+		u16 j;
+
+		bitmap_zero(profile_to_recipe[i], ICE_MAX_NUM_RECIPES);
+		bitmap_zero(r_bitmap, ICE_MAX_NUM_RECIPES);
+		if (ice_aq_get_recipe_to_profile(hw, i, (u8 *)r_bitmap, NULL))
+			continue;
+		bitmap_copy(profile_to_recipe[i], r_bitmap,
+			    ICE_MAX_NUM_RECIPES);
+		for_each_set_bit(j, r_bitmap, ICE_MAX_NUM_RECIPES)
+			set_bit(i, recipe_to_profile[j]);
+	}
+}
+
+/**
+ * ice_collect_result_idx - copy result index values
+ * @buf: buffer that contains the result index
+ * @recp: the recipe struct to copy data into
+ */
+static void
+ice_collect_result_idx(struct ice_aqc_recipe_data_elem *buf,
+		       struct ice_sw_recipe *recp)
+{
+	if (buf->content.result_indx & ICE_AQ_RECIPE_RESULT_EN)
+		set_bit(buf->content.result_indx & ~ICE_AQ_RECIPE_RESULT_EN,
+			recp->res_idxs);
+}
+
+/**
+ * ice_get_recp_frm_fw - update SW bookkeeping from FW recipe entries
+ * @hw: pointer to hardware structure
+ * @recps: struct that we need to populate
+ * @rid: recipe ID that we are populating
+ * @refresh_required: true if we should get recipe to profile mapping from FW
+ *
+ * This function is used to populate all the necessary entries into our
+ * bookkeeping so that we have a current list of all the recipes that are
+ * programmed in the firmware.
+ */
+static enum ice_status
+ice_get_recp_frm_fw(struct ice_hw *hw, struct ice_sw_recipe *recps, u8 rid,
+		    bool *refresh_required)
+{
+	DECLARE_BITMAP(result_bm, ICE_MAX_FV_WORDS);
+	struct ice_aqc_recipe_data_elem *tmp;
+	u16 num_recps = ICE_MAX_NUM_RECIPES;
+	struct ice_prot_lkup_ext *lkup_exts;
+	enum ice_status status;
+	u8 fv_word_idx = 0;
+	u16 sub_recps;
+
+	bitmap_zero(result_bm, ICE_MAX_FV_WORDS);
+
+	/* we need a buffer big enough to accommodate all the recipes */
+	tmp = kcalloc(ICE_MAX_NUM_RECIPES, sizeof(*tmp), GFP_KERNEL);
+	if (!tmp)
+		return ICE_ERR_NO_MEMORY;
+
+	tmp[0].recipe_indx = rid;
+	status = ice_aq_get_recipe(hw, tmp, &num_recps, rid, NULL);
+	/* non-zero status meaning recipe doesn't exist */
+	if (status)
+		goto err_unroll;
+
+	/* Get recipe to profile map so that we can get the fv from lkups that
+	 * we read for a recipe from FW. Since we want to minimize the number of
+	 * times we make this FW call, just make one call and cache the copy
+	 * until a new recipe is added. This operation is only required the
+	 * first time to get the changes from FW. Then to search existing
+	 * entries we don't need to update the cache again until another recipe
+	 * gets added.
+	 */
+	if (*refresh_required) {
+		ice_get_recp_to_prof_map(hw);
+		*refresh_required = false;
+	}
+
+	/* Start populating all the entries for recps[rid] based on lkups from
+	 * firmware. Note that we are only creating the root recipe in our
+	 * database.
+	 */
+	lkup_exts = &recps[rid].lkup_exts;
+
+	for (sub_recps = 0; sub_recps < num_recps; sub_recps++) {
+		struct ice_aqc_recipe_data_elem root_bufs = tmp[sub_recps];
+		struct ice_recp_grp_entry *rg_entry;
+		u8 i, prof, idx, prot = 0;
+		bool is_root;
+		u16 off = 0;
+
+		rg_entry = devm_kzalloc(ice_hw_to_dev(hw), sizeof(*rg_entry),
+					GFP_KERNEL);
+		if (!rg_entry) {
+			status = ICE_ERR_NO_MEMORY;
+			goto err_unroll;
+		}
+
+		idx = root_bufs.recipe_indx;
+		is_root = root_bufs.content.rid & ICE_AQ_RECIPE_ID_IS_ROOT;
+
+		/* Mark all result indices in this chain */
+		if (root_bufs.content.result_indx & ICE_AQ_RECIPE_RESULT_EN)
+			set_bit(root_bufs.content.result_indx & ~ICE_AQ_RECIPE_RESULT_EN,
+				result_bm);
+
+		/* get the first profile that is associated with rid */
+		prof = find_first_bit(recipe_to_profile[idx],
+				      ICE_MAX_NUM_PROFILES);
+		for (i = 0; i < ICE_NUM_WORDS_RECIPE; i++) {
+			u8 lkup_indx = root_bufs.content.lkup_indx[i + 1];
+
+			rg_entry->fv_idx[i] = lkup_indx;
+			rg_entry->fv_mask[i] =
+				le16_to_cpu(root_bufs.content.mask[i + 1]);
+
+			/* If the recipe is a chained recipe then all its
+			 * child recipe's result will have a result index.
+			 * To fill fv_words we should not use those result
+			 * index, we only need the protocol ids and offsets.
+			 * We will skip all the fv_idx which stores result
+			 * index in them. We also need to skip any fv_idx which
+			 * has ICE_AQ_RECIPE_LKUP_IGNORE or 0 since it isn't a
+			 * valid offset value.
+			 */
+			if (test_bit(rg_entry->fv_idx[i], hw->switch_info->prof_res_bm[prof]) ||
+			    rg_entry->fv_idx[i] & ICE_AQ_RECIPE_LKUP_IGNORE ||
+			    rg_entry->fv_idx[i] == 0)
+				continue;
+
+			ice_find_prot_off(hw, ICE_BLK_SW, prof,
+					  rg_entry->fv_idx[i], &prot, &off);
+			lkup_exts->fv_words[fv_word_idx].prot_id = prot;
+			lkup_exts->fv_words[fv_word_idx].off = off;
+			lkup_exts->field_mask[fv_word_idx] =
+				rg_entry->fv_mask[i];
+			fv_word_idx++;
+		}
+		/* populate rg_list with the data from the child entry of this
+		 * recipe
+		 */
+		list_add(&rg_entry->l_entry, &recps[rid].rg_list);
+
+		/* Propagate some data to the recipe database */
+		recps[idx].is_root = !!is_root;
+		recps[idx].priority = root_bufs.content.act_ctrl_fwd_priority;
+		bitmap_zero(recps[idx].res_idxs, ICE_MAX_FV_WORDS);
+		if (root_bufs.content.result_indx & ICE_AQ_RECIPE_RESULT_EN) {
+			recps[idx].chain_idx = root_bufs.content.result_indx &
+				~ICE_AQ_RECIPE_RESULT_EN;
+			set_bit(recps[idx].chain_idx, recps[idx].res_idxs);
+		} else {
+			recps[idx].chain_idx = ICE_INVAL_CHAIN_IND;
+		}
+
+		if (!is_root)
+			continue;
+
+		/* Only do the following for root recipes entries */
+		memcpy(recps[idx].r_bitmap, root_bufs.recipe_bitmap,
+		       sizeof(recps[idx].r_bitmap));
+		recps[idx].root_rid = root_bufs.content.rid &
+			~ICE_AQ_RECIPE_ID_IS_ROOT;
+		recps[idx].priority = root_bufs.content.act_ctrl_fwd_priority;
+	}
+
+	/* Complete initialization of the root recipe entry */
+	lkup_exts->n_val_words = fv_word_idx;
+	recps[rid].big_recp = (num_recps > 1);
+	recps[rid].n_grp_count = (u8)num_recps;
+	recps[rid].root_buf = devm_kmemdup(ice_hw_to_dev(hw), tmp,
+					   recps[rid].n_grp_count * sizeof(*recps[rid].root_buf),
+					   GFP_KERNEL);
+	if (!recps[rid].root_buf)
+		goto err_unroll;
+
+	/* Copy result indexes */
+	bitmap_copy(recps[rid].res_idxs, result_bm, ICE_MAX_FV_WORDS);
+	recps[rid].recp_created = true;
+
+err_unroll:
+	kfree(tmp);
+	return status;
+}
+
 /* ice_init_port_info - Initialize port_info with switch configuration data
  * @pi: pointer to port_info
  * @vsi_port_num: VSI number or port number
@@ -2987,6 +3192,907 @@ enum ice_status
 	return status;
 }
 
+/* This is mapping table entry that maps every word within a given protocol
+ * structure to the real byte offset as per the specification of that
+ * protocol header.
+ * for example dst address is 3 words in ethertype header and corresponding
+ * bytes are 0, 2, 3 in the actual packet header and src address is at 4, 6, 8
+ * IMPORTANT: Every structure part of "ice_prot_hdr" union should have a
+ * matching entry describing its field. This needs to be updated if new
+ * structure is added to that union.
+ */
+static const struct ice_prot_ext_tbl_entry ice_prot_ext[ICE_PROTOCOL_LAST] = {
+	{ ICE_MAC_OFOS,		{ 0, 2, 4, 6, 8, 10, 12 } },
+	{ ICE_MAC_IL,		{ 0, 2, 4, 6, 8, 10, 12 } },
+	{ ICE_ETYPE_OL,		{ 0 } },
+	{ ICE_VLAN_OFOS,	{ 2, 0 } },
+	{ ICE_IPV4_OFOS,	{ 0, 2, 4, 6, 8, 10, 12, 14, 16, 18 } },
+	{ ICE_IPV4_IL,		{ 0, 2, 4, 6, 8, 10, 12, 14, 16, 18 } },
+	{ ICE_IPV6_OFOS,	{ 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24,
+				 26, 28, 30, 32, 34, 36, 38 } },
+	{ ICE_IPV6_IL,		{ 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24,
+				 26, 28, 30, 32, 34, 36, 38 } },
+	{ ICE_TCP_IL,		{ 0, 2 } },
+	{ ICE_UDP_OF,		{ 0, 2 } },
+	{ ICE_UDP_ILOS,		{ 0, 2 } },
+};
+
+static struct ice_protocol_entry ice_prot_id_tbl[ICE_PROTOCOL_LAST] = {
+	{ ICE_MAC_OFOS,		ICE_MAC_OFOS_HW },
+	{ ICE_MAC_IL,		ICE_MAC_IL_HW },
+	{ ICE_ETYPE_OL,		ICE_ETYPE_OL_HW },
+	{ ICE_VLAN_OFOS,	ICE_VLAN_OL_HW },
+	{ ICE_IPV4_OFOS,	ICE_IPV4_OFOS_HW },
+	{ ICE_IPV4_IL,		ICE_IPV4_IL_HW },
+	{ ICE_IPV6_OFOS,	ICE_IPV6_OFOS_HW },
+	{ ICE_IPV6_IL,		ICE_IPV6_IL_HW },
+	{ ICE_TCP_IL,		ICE_TCP_IL_HW },
+	{ ICE_UDP_OF,		ICE_UDP_OF_HW },
+	{ ICE_UDP_ILOS,		ICE_UDP_ILOS_HW },
+};
+
+/**
+ * ice_find_recp - find a recipe
+ * @hw: pointer to the hardware structure
+ * @lkup_exts: extension sequence to match
+ *
+ * Returns index of matching recipe, or ICE_MAX_NUM_RECIPES if not found.
+ */
+static u16 ice_find_recp(struct ice_hw *hw, struct ice_prot_lkup_ext *lkup_exts)
+{
+	bool refresh_required = true;
+	struct ice_sw_recipe *recp;
+	u8 i;
+
+	/* Walk through existing recipes to find a match */
+	recp = hw->switch_info->recp_list;
+	for (i = 0; i < ICE_MAX_NUM_RECIPES; i++) {
+		/* If recipe was not created for this ID, in SW bookkeeping,
+		 * check if FW has an entry for this recipe. If the FW has an
+		 * entry update it in our SW bookkeeping and continue with the
+		 * matching.
+		 */
+		if (!recp[i].recp_created)
+			if (ice_get_recp_frm_fw(hw,
+						hw->switch_info->recp_list, i,
+						&refresh_required))
+				continue;
+
+		/* Skip inverse action recipes */
+		if (recp[i].root_buf && recp[i].root_buf->content.act_ctrl &
+		    ICE_AQ_RECIPE_ACT_INV_ACT)
+			continue;
+
+		/* if number of words we are looking for match */
+		if (lkup_exts->n_val_words == recp[i].lkup_exts.n_val_words) {
+			struct ice_fv_word *ar = recp[i].lkup_exts.fv_words;
+			struct ice_fv_word *be = lkup_exts->fv_words;
+			u16 *cr = recp[i].lkup_exts.field_mask;
+			u16 *de = lkup_exts->field_mask;
+			bool found = true;
+			u8 pe, qr;
+
+			/* ar, cr, and qr are related to the recipe words, while
+			 * be, de, and pe are related to the lookup words
+			 */
+			for (pe = 0; pe < lkup_exts->n_val_words; pe++) {
+				for (qr = 0; qr < recp[i].lkup_exts.n_val_words;
+				     qr++) {
+					if (ar[qr].off == be[pe].off &&
+					    ar[qr].prot_id == be[pe].prot_id &&
+					    cr[qr] == de[pe])
+						/* Found the "pe"th word in the
+						 * given recipe
+						 */
+						break;
+				}
+				/* After walking through all the words in the
+				 * "i"th recipe if "p"th word was not found then
+				 * this recipe is not what we are looking for.
+				 * So break out from this loop and try the next
+				 * recipe
+				 */
+				if (qr >= recp[i].lkup_exts.n_val_words) {
+					found = false;
+					break;
+				}
+			}
+			/* If for "i"th recipe the found was never set to false
+			 * then it means we found our match
+			 */
+			if (found)
+				return i; /* Return the recipe ID */
+		}
+	}
+	return ICE_MAX_NUM_RECIPES;
+}
+
+/**
+ * ice_prot_type_to_id - get protocol ID from protocol type
+ * @type: protocol type
+ * @id: pointer to variable that will receive the ID
+ *
+ * Returns true if found, false otherwise
+ */
+static bool ice_prot_type_to_id(enum ice_protocol_type type, u8 *id)
+{
+	u8 i;
+
+	for (i = 0; i < ARRAY_SIZE(ice_prot_id_tbl); i++)
+		if (ice_prot_id_tbl[i].type == type) {
+			*id = ice_prot_id_tbl[i].protocol_id;
+			return true;
+		}
+	return false;
+}
+
+/**
+ * ice_fill_valid_words - count valid words
+ * @rule: advanced rule with lookup information
+ * @lkup_exts: byte offset extractions of the words that are valid
+ *
+ * calculate valid words in a lookup rule using mask value
+ */
+static u8
+ice_fill_valid_words(struct ice_adv_lkup_elem *rule,
+		     struct ice_prot_lkup_ext *lkup_exts)
+{
+	u8 j, word, prot_id, ret_val;
+
+	if (!ice_prot_type_to_id(rule->type, &prot_id))
+		return 0;
+
+	word = lkup_exts->n_val_words;
+
+	for (j = 0; j < sizeof(rule->m_u) / sizeof(u16); j++)
+		if (((u16 *)&rule->m_u)[j] &&
+		    rule->type < ARRAY_SIZE(ice_prot_ext)) {
+			/* No more space to accommodate */
+			if (word >= ICE_MAX_CHAIN_WORDS)
+				return 0;
+			lkup_exts->fv_words[word].off =
+				ice_prot_ext[rule->type].offs[j];
+			lkup_exts->fv_words[word].prot_id =
+				ice_prot_id_tbl[rule->type].protocol_id;
+			lkup_exts->field_mask[word] =
+				be16_to_cpu(((__force __be16 *)&rule->m_u)[j]);
+			word++;
+		}
+
+	ret_val = word - lkup_exts->n_val_words;
+	lkup_exts->n_val_words = word;
+
+	return ret_val;
+}
+
+/**
+ * ice_create_first_fit_recp_def - Create a recipe grouping
+ * @hw: pointer to the hardware structure
+ * @lkup_exts: an array of protocol header extractions
+ * @rg_list: pointer to a list that stores new recipe groups
+ * @recp_cnt: pointer to a variable that stores returned number of recipe groups
+ *
+ * Using first fit algorithm, take all the words that are still not done
+ * and start grouping them in 4-word groups. Each group makes up one
+ * recipe.
+ */
+static enum ice_status
+ice_create_first_fit_recp_def(struct ice_hw *hw,
+			      struct ice_prot_lkup_ext *lkup_exts,
+			      struct list_head *rg_list,
+			      u8 *recp_cnt)
+{
+	struct ice_pref_recipe_group *grp = NULL;
+	u8 j;
+
+	*recp_cnt = 0;
+
+	/* Walk through every word in the rule to check if it is not done. If so
+	 * then this word needs to be part of a new recipe.
+	 */
+	for (j = 0; j < lkup_exts->n_val_words; j++)
+		if (!test_bit(j, lkup_exts->done)) {
+			if (!grp ||
+			    grp->n_val_pairs == ICE_NUM_WORDS_RECIPE) {
+				struct ice_recp_grp_entry *entry;
+
+				entry = devm_kzalloc(ice_hw_to_dev(hw),
+						     sizeof(*entry),
+						     GFP_KERNEL);
+				if (!entry)
+					return ICE_ERR_NO_MEMORY;
+				list_add(&entry->l_entry, rg_list);
+				grp = &entry->r_group;
+				(*recp_cnt)++;
+			}
+
+			grp->pairs[grp->n_val_pairs].prot_id =
+				lkup_exts->fv_words[j].prot_id;
+			grp->pairs[grp->n_val_pairs].off =
+				lkup_exts->fv_words[j].off;
+			grp->mask[grp->n_val_pairs] = lkup_exts->field_mask[j];
+			grp->n_val_pairs++;
+		}
+
+	return 0;
+}
+
+/**
+ * ice_fill_fv_word_index - fill in the field vector indices for a recipe group
+ * @hw: pointer to the hardware structure
+ * @fv_list: field vector with the extraction sequence information
+ * @rg_list: recipe groupings with protocol-offset pairs
+ *
+ * Helper function to fill in the field vector indices for protocol-offset
+ * pairs. These indexes are then ultimately programmed into a recipe.
+ */
+static enum ice_status
+ice_fill_fv_word_index(struct ice_hw *hw, struct list_head *fv_list,
+		       struct list_head *rg_list)
+{
+	struct ice_sw_fv_list_entry *fv;
+	struct ice_recp_grp_entry *rg;
+	struct ice_fv_word *fv_ext;
+
+	if (list_empty(fv_list))
+		return 0;
+
+	fv = list_first_entry(fv_list, struct ice_sw_fv_list_entry,
+			      list_entry);
+	fv_ext = fv->fv_ptr->ew;
+
+	list_for_each_entry(rg, rg_list, l_entry) {
+		u8 i;
+
+		for (i = 0; i < rg->r_group.n_val_pairs; i++) {
+			struct ice_fv_word *pr;
+			bool found = false;
+			u16 mask;
+			u8 j;
+
+			pr = &rg->r_group.pairs[i];
+			mask = rg->r_group.mask[i];
+
+			for (j = 0; j < hw->blk[ICE_BLK_SW].es.fvw; j++)
+				if (fv_ext[j].prot_id == pr->prot_id &&
+				    fv_ext[j].off == pr->off) {
+					found = true;
+
+					/* Store index of field vector */
+					rg->fv_idx[i] = j;
+					rg->fv_mask[i] = mask;
+					break;
+				}
+
+			/* Protocol/offset could not be found, caller gave an
+			 * invalid pair
+			 */
+			if (!found)
+				return ICE_ERR_PARAM;
+		}
+	}
+
+	return 0;
+}
+
+/**
+ * ice_find_free_recp_res_idx - find free result indexes for recipe
+ * @hw: pointer to hardware structure
+ * @profiles: bitmap of profiles that will be associated with the new recipe
+ * @free_idx: pointer to variable to receive the free index bitmap
+ *
+ * The algorithm used here is:
+ *	1. When creating a new recipe, create a set P which contains all
+ *	   Profiles that will be associated with our new recipe
+ *
+ *	2. For each Profile p in set P:
+ *	    a. Add all recipes associated with Profile p into set R
+ *	    b. Optional : PossibleIndexes &= profile[p].possibleIndexes
+ *		[initially PossibleIndexes should be 0xFFFFFFFFFFFFFFFF]
+ *		i. Or just assume they all have the same possible indexes:
+ *			44, 45, 46, 47
+ *			i.e., PossibleIndexes = 0x0000F00000000000
+ *
+ *	3. For each Recipe r in set R:
+ *	    a. UsedIndexes |= (bitwise or ) recipe[r].res_indexes
+ *	    b. FreeIndexes = UsedIndexes ^ PossibleIndexes
+ *
+ *	FreeIndexes will contain the bits indicating the indexes free for use,
+ *      then the code needs to update the recipe[r].used_result_idx_bits to
+ *      indicate which indexes were selected for use by this recipe.
+ */
+static u16
+ice_find_free_recp_res_idx(struct ice_hw *hw, const unsigned long *profiles,
+			   unsigned long *free_idx)
+{
+	DECLARE_BITMAP(possible_idx, ICE_MAX_FV_WORDS);
+	DECLARE_BITMAP(recipes, ICE_MAX_NUM_RECIPES);
+	DECLARE_BITMAP(used_idx, ICE_MAX_FV_WORDS);
+	u16 bit;
+
+	bitmap_zero(possible_idx, ICE_MAX_FV_WORDS);
+	bitmap_zero(recipes, ICE_MAX_NUM_RECIPES);
+	bitmap_zero(used_idx, ICE_MAX_FV_WORDS);
+	bitmap_zero(free_idx, ICE_MAX_FV_WORDS);
+
+	bitmap_set(possible_idx, 0, ICE_MAX_FV_WORDS);
+
+	/* For each profile we are going to associate the recipe with, add the
+	 * recipes that are associated with that profile. This will give us
+	 * the set of recipes that our recipe may collide with. Also, determine
+	 * what possible result indexes are usable given this set of profiles.
+	 */
+	for_each_set_bit(bit, profiles, ICE_MAX_NUM_PROFILES) {
+		bitmap_or(recipes, recipes, profile_to_recipe[bit],
+			  ICE_MAX_NUM_RECIPES);
+		bitmap_and(possible_idx, possible_idx,
+			   hw->switch_info->prof_res_bm[bit],
+			   ICE_MAX_FV_WORDS);
+	}
+
+	/* For each recipe that our new recipe may collide with, determine
+	 * which indexes have been used.
+	 */
+	for_each_set_bit(bit, recipes, ICE_MAX_NUM_RECIPES)
+		bitmap_or(used_idx, used_idx,
+			  hw->switch_info->recp_list[bit].res_idxs,
+			  ICE_MAX_FV_WORDS);
+
+	bitmap_xor(free_idx, used_idx, possible_idx, ICE_MAX_FV_WORDS);
+
+	/* return number of free indexes */
+	return (u16)bitmap_weight(free_idx, ICE_MAX_FV_WORDS);
+}
+
+/**
+ * ice_add_sw_recipe - function to call AQ calls to create switch recipe
+ * @hw: pointer to hardware structure
+ * @rm: recipe management list entry
+ * @match_tun_mask: tunnel mask that needs to be programmed
+ * @profiles: bitmap of profiles that will be associated.
+ */
+static enum ice_status
+ice_add_sw_recipe(struct ice_hw *hw, struct ice_sw_recipe *rm,
+		  u16 match_tun_mask, unsigned long *profiles)
+{
+	DECLARE_BITMAP(result_idx_bm, ICE_MAX_FV_WORDS);
+	struct ice_aqc_recipe_data_elem *tmp;
+	struct ice_aqc_recipe_data_elem *buf;
+	struct ice_recp_grp_entry *entry;
+	enum ice_status status;
+	u16 free_res_idx;
+	u16 recipe_count;
+	u8 chain_idx;
+	u8 recps = 0;
+
+	/* When more than one recipe are required, another recipe is needed to
+	 * chain them together. Matching a tunnel metadata ID takes up one of
+	 * the match fields in the chaining recipe reducing the number of
+	 * chained recipes by one.
+	 */
+	 /* check number of free result indices */
+	bitmap_zero(result_idx_bm, ICE_MAX_FV_WORDS);
+	free_res_idx = ice_find_free_recp_res_idx(hw, profiles, result_idx_bm);
+
+	ice_debug(hw, ICE_DBG_SW, "Result idx slots: %d, need %d\n",
+		  free_res_idx, rm->n_grp_count);
+
+	if (rm->n_grp_count > 1) {
+		if (rm->n_grp_count > free_res_idx)
+			return ICE_ERR_MAX_LIMIT;
+
+		rm->n_grp_count++;
+	}
+
+	if (rm->n_grp_count > ICE_MAX_CHAIN_RECIPE)
+		return ICE_ERR_MAX_LIMIT;
+
+	tmp = kcalloc(ICE_MAX_NUM_RECIPES, sizeof(*tmp), GFP_KERNEL);
+	if (!tmp)
+		return ICE_ERR_NO_MEMORY;
+
+	buf = devm_kcalloc(ice_hw_to_dev(hw), rm->n_grp_count, sizeof(*buf),
+			   GFP_KERNEL);
+	if (!buf) {
+		status = ICE_ERR_NO_MEMORY;
+		goto err_mem;
+	}
+
+	bitmap_zero(rm->r_bitmap, ICE_MAX_NUM_RECIPES);
+	recipe_count = ICE_MAX_NUM_RECIPES;
+	status = ice_aq_get_recipe(hw, tmp, &recipe_count, ICE_SW_LKUP_MAC,
+				   NULL);
+	if (status || recipe_count == 0)
+		goto err_unroll;
+
+	/* Allocate the recipe resources, and configure them according to the
+	 * match fields from protocol headers and extracted field vectors.
+	 */
+	chain_idx = find_first_bit(result_idx_bm, ICE_MAX_FV_WORDS);
+	list_for_each_entry(entry, &rm->rg_list, l_entry) {
+		u8 i;
+
+		status = ice_alloc_recipe(hw, &entry->rid);
+		if (status)
+			goto err_unroll;
+
+		/* Clear the result index of the located recipe, as this will be
+		 * updated, if needed, later in the recipe creation process.
+		 */
+		tmp[0].content.result_indx = 0;
+
+		buf[recps] = tmp[0];
+		buf[recps].recipe_indx = (u8)entry->rid;
+		/* if the recipe is a non-root recipe RID should be programmed
+		 * as 0 for the rules to be applied correctly.
+		 */
+		buf[recps].content.rid = 0;
+		memset(&buf[recps].content.lkup_indx, 0,
+		       sizeof(buf[recps].content.lkup_indx));
+
+		/* All recipes use look-up index 0 to match switch ID. */
+		buf[recps].content.lkup_indx[0] = ICE_AQ_SW_ID_LKUP_IDX;
+		buf[recps].content.mask[0] =
+			cpu_to_le16(ICE_AQ_SW_ID_LKUP_MASK);
+		/* Setup lkup_indx 1..4 to INVALID/ignore and set the mask
+		 * to be 0
+		 */
+		for (i = 1; i <= ICE_NUM_WORDS_RECIPE; i++) {
+			buf[recps].content.lkup_indx[i] = 0x80;
+			buf[recps].content.mask[i] = 0;
+		}
+
+		for (i = 0; i < entry->r_group.n_val_pairs; i++) {
+			buf[recps].content.lkup_indx[i + 1] = entry->fv_idx[i];
+			buf[recps].content.mask[i + 1] =
+				cpu_to_le16(entry->fv_mask[i]);
+		}
+
+		if (rm->n_grp_count > 1) {
+			/* Checks to see if there really is a valid result index
+			 * that can be used.
+			 */
+			if (chain_idx >= ICE_MAX_FV_WORDS) {
+				ice_debug(hw, ICE_DBG_SW, "No chain index available\n");
+				status = ICE_ERR_MAX_LIMIT;
+				goto err_unroll;
+			}
+
+			entry->chain_idx = chain_idx;
+			buf[recps].content.result_indx =
+				ICE_AQ_RECIPE_RESULT_EN |
+				((chain_idx << ICE_AQ_RECIPE_RESULT_DATA_S) &
+				 ICE_AQ_RECIPE_RESULT_DATA_M);
+			clear_bit(chain_idx, result_idx_bm);
+			chain_idx = find_first_bit(result_idx_bm,
+						   ICE_MAX_FV_WORDS);
+		}
+
+		/* fill recipe dependencies */
+		bitmap_zero((unsigned long *)buf[recps].recipe_bitmap,
+			    ICE_MAX_NUM_RECIPES);
+		set_bit(buf[recps].recipe_indx,
+			(unsigned long *)buf[recps].recipe_bitmap);
+		buf[recps].content.act_ctrl_fwd_priority = rm->priority;
+		recps++;
+	}
+
+	if (rm->n_grp_count == 1) {
+		rm->root_rid = buf[0].recipe_indx;
+		set_bit(buf[0].recipe_indx, rm->r_bitmap);
+		buf[0].content.rid = rm->root_rid | ICE_AQ_RECIPE_ID_IS_ROOT;
+		if (sizeof(buf[0].recipe_bitmap) >= sizeof(rm->r_bitmap)) {
+			memcpy(buf[0].recipe_bitmap, rm->r_bitmap,
+			       sizeof(buf[0].recipe_bitmap));
+		} else {
+			status = ICE_ERR_BAD_PTR;
+			goto err_unroll;
+		}
+		/* Applicable only for ROOT_RECIPE, set the fwd_priority for
+		 * the recipe which is getting created if specified
+		 * by user. Usually any advanced switch filter, which results
+		 * into new extraction sequence, ended up creating a new recipe
+		 * of type ROOT and usually recipes are associated with profiles
+		 * Switch rule referreing newly created recipe, needs to have
+		 * either/or 'fwd' or 'join' priority, otherwise switch rule
+		 * evaluation will not happen correctly. In other words, if
+		 * switch rule to be evaluated on priority basis, then recipe
+		 * needs to have priority, otherwise it will be evaluated last.
+		 */
+		buf[0].content.act_ctrl_fwd_priority = rm->priority;
+	} else {
+		struct ice_recp_grp_entry *last_chain_entry;
+		u16 rid, i;
+
+		/* Allocate the last recipe that will chain the outcomes of the
+		 * other recipes together
+		 */
+		status = ice_alloc_recipe(hw, &rid);
+		if (status)
+			goto err_unroll;
+
+		buf[recps].recipe_indx = (u8)rid;
+		buf[recps].content.rid = (u8)rid;
+		buf[recps].content.rid |= ICE_AQ_RECIPE_ID_IS_ROOT;
+		/* the new entry created should also be part of rg_list to
+		 * make sure we have complete recipe
+		 */
+		last_chain_entry = devm_kzalloc(ice_hw_to_dev(hw),
+						sizeof(*last_chain_entry),
+						GFP_KERNEL);
+		if (!last_chain_entry) {
+			status = ICE_ERR_NO_MEMORY;
+			goto err_unroll;
+		}
+		last_chain_entry->rid = rid;
+		memset(&buf[recps].content.lkup_indx, 0,
+		       sizeof(buf[recps].content.lkup_indx));
+		/* All recipes use look-up index 0 to match switch ID. */
+		buf[recps].content.lkup_indx[0] = ICE_AQ_SW_ID_LKUP_IDX;
+		buf[recps].content.mask[0] =
+			cpu_to_le16(ICE_AQ_SW_ID_LKUP_MASK);
+		for (i = 1; i <= ICE_NUM_WORDS_RECIPE; i++) {
+			buf[recps].content.lkup_indx[i] =
+				ICE_AQ_RECIPE_LKUP_IGNORE;
+			buf[recps].content.mask[i] = 0;
+		}
+
+		i = 1;
+		/* update r_bitmap with the recp that is used for chaining */
+		set_bit(rid, rm->r_bitmap);
+		/* this is the recipe that chains all the other recipes so it
+		 * should not have a chaining ID to indicate the same
+		 */
+		last_chain_entry->chain_idx = ICE_INVAL_CHAIN_IND;
+		list_for_each_entry(entry, &rm->rg_list, l_entry) {
+			last_chain_entry->fv_idx[i] = entry->chain_idx;
+			buf[recps].content.lkup_indx[i] = entry->chain_idx;
+			buf[recps].content.mask[i++] = cpu_to_le16(0xFFFF);
+			set_bit(entry->rid, rm->r_bitmap);
+		}
+		list_add(&last_chain_entry->l_entry, &rm->rg_list);
+		if (sizeof(buf[recps].recipe_bitmap) >=
+		    sizeof(rm->r_bitmap)) {
+			memcpy(buf[recps].recipe_bitmap, rm->r_bitmap,
+			       sizeof(buf[recps].recipe_bitmap));
+		} else {
+			status = ICE_ERR_BAD_PTR;
+			goto err_unroll;
+		}
+		buf[recps].content.act_ctrl_fwd_priority = rm->priority;
+
+		/* To differentiate among different UDP tunnels, a meta data ID
+		 * flag is used.
+		 */
+		if (match_tun_mask) {
+			buf[recps].content.lkup_indx[i] = ICE_TUN_FLAG_FV_IND;
+			buf[recps].content.mask[i] =
+				cpu_to_le16(match_tun_mask);
+		}
+
+		recps++;
+		rm->root_rid = (u8)rid;
+	}
+	status = ice_acquire_change_lock(hw, ICE_RES_WRITE);
+	if (status)
+		goto err_unroll;
+
+	status = ice_aq_add_recipe(hw, buf, rm->n_grp_count, NULL);
+	ice_release_change_lock(hw);
+	if (status)
+		goto err_unroll;
+
+	/* Every recipe that just got created add it to the recipe
+	 * book keeping list
+	 */
+	list_for_each_entry(entry, &rm->rg_list, l_entry) {
+		struct ice_switch_info *sw = hw->switch_info;
+		bool is_root, idx_found = false;
+		struct ice_sw_recipe *recp;
+		u16 idx, buf_idx = 0;
+
+		/* find buffer index for copying some data */
+		for (idx = 0; idx < rm->n_grp_count; idx++)
+			if (buf[idx].recipe_indx == entry->rid) {
+				buf_idx = idx;
+				idx_found = true;
+			}
+
+		if (!idx_found) {
+			status = ICE_ERR_OUT_OF_RANGE;
+			goto err_unroll;
+		}
+
+		recp = &sw->recp_list[entry->rid];
+		is_root = (rm->root_rid == entry->rid);
+		recp->is_root = is_root;
+
+		recp->root_rid = entry->rid;
+		recp->big_recp = (is_root && rm->n_grp_count > 1);
+
+		memcpy(&recp->ext_words, entry->r_group.pairs,
+		       entry->r_group.n_val_pairs * sizeof(struct ice_fv_word));
+
+		memcpy(recp->r_bitmap, buf[buf_idx].recipe_bitmap,
+		       sizeof(recp->r_bitmap));
+
+		/* Copy non-result fv index values and masks to recipe. This
+		 * call will also update the result recipe bitmask.
+		 */
+		ice_collect_result_idx(&buf[buf_idx], recp);
+
+		/* for non-root recipes, also copy to the root, this allows
+		 * easier matching of a complete chained recipe
+		 */
+		if (!is_root)
+			ice_collect_result_idx(&buf[buf_idx],
+					       &sw->recp_list[rm->root_rid]);
+
+		recp->n_ext_words = entry->r_group.n_val_pairs;
+		recp->chain_idx = entry->chain_idx;
+		recp->priority = buf[buf_idx].content.act_ctrl_fwd_priority;
+		recp->n_grp_count = rm->n_grp_count;
+		recp->recp_created = true;
+	}
+	rm->root_buf = buf;
+	kfree(tmp);
+	return status;
+
+err_unroll:
+err_mem:
+	kfree(tmp);
+	devm_kfree(ice_hw_to_dev(hw), buf);
+	return status;
+}
+
+/**
+ * ice_create_recipe_group - creates recipe group
+ * @hw: pointer to hardware structure
+ * @rm: recipe management list entry
+ * @lkup_exts: lookup elements
+ */
+static enum ice_status
+ice_create_recipe_group(struct ice_hw *hw, struct ice_sw_recipe *rm,
+			struct ice_prot_lkup_ext *lkup_exts)
+{
+	enum ice_status status;
+	u8 recp_count = 0;
+
+	rm->n_grp_count = 0;
+
+	/* Create recipes for words that are marked not done by packing them
+	 * as best fit.
+	 */
+	status = ice_create_first_fit_recp_def(hw, lkup_exts,
+					       &rm->rg_list, &recp_count);
+	if (!status) {
+		rm->n_grp_count += recp_count;
+		rm->n_ext_words = lkup_exts->n_val_words;
+		memcpy(&rm->ext_words, lkup_exts->fv_words,
+		       sizeof(rm->ext_words));
+		memcpy(rm->word_masks, lkup_exts->field_mask,
+		       sizeof(rm->word_masks));
+	}
+
+	return status;
+}
+
+/**
+ * ice_get_fv - get field vectors/extraction sequences for spec. lookup types
+ * @hw: pointer to hardware structure
+ * @lkups: lookup elements or match criteria for the advanced recipe, one
+ *	   structure per protocol header
+ * @lkups_cnt: number of protocols
+ * @bm: bitmap of field vectors to consider
+ * @fv_list: pointer to a list that holds the returned field vectors
+ */
+static enum ice_status
+ice_get_fv(struct ice_hw *hw, struct ice_adv_lkup_elem *lkups, u16 lkups_cnt,
+	   unsigned long *bm, struct list_head *fv_list)
+{
+	enum ice_status status;
+	u8 *prot_ids;
+	u16 i;
+
+	prot_ids = kcalloc(lkups_cnt, sizeof(*prot_ids), GFP_KERNEL);
+	if (!prot_ids)
+		return ICE_ERR_NO_MEMORY;
+
+	for (i = 0; i < lkups_cnt; i++)
+		if (!ice_prot_type_to_id(lkups[i].type, &prot_ids[i])) {
+			status = ICE_ERR_CFG;
+			goto free_mem;
+		}
+
+	/* Find field vectors that include all specified protocol types */
+	status = ice_get_sw_fv_list(hw, prot_ids, lkups_cnt, bm, fv_list);
+
+free_mem:
+	kfree(prot_ids);
+	return status;
+}
+
+/* ice_get_compat_fv_bitmap - Get compatible field vector bitmap for rule
+ * @hw: pointer to hardware structure
+ * @rinfo: other information regarding the rule e.g. priority and action info
+ * @bm: pointer to memory for returning the bitmap of field vectors
+ */
+static void
+ice_get_compat_fv_bitmap(struct ice_hw *hw, struct ice_adv_rule_info *rinfo,
+			 unsigned long *bm)
+{
+	bitmap_zero(bm, ICE_MAX_NUM_PROFILES);
+
+	ice_get_sw_fv_bitmap(hw, ICE_PROF_NON_TUN, bm);
+}
+
+/**
+ * ice_add_adv_recipe - Add an advanced recipe that is not part of the default
+ * @hw: pointer to hardware structure
+ * @lkups: lookup elements or match criteria for the advanced recipe, one
+ *  structure per protocol header
+ * @lkups_cnt: number of protocols
+ * @rinfo: other information regarding the rule e.g. priority and action info
+ * @rid: return the recipe ID of the recipe created
+ */
+static enum ice_status __maybe_unused
+ice_add_adv_recipe(struct ice_hw *hw, struct ice_adv_lkup_elem *lkups,
+		   u16 lkups_cnt, struct ice_adv_rule_info *rinfo, u16 *rid)
+{
+	DECLARE_BITMAP(fv_bitmap, ICE_MAX_NUM_PROFILES);
+	DECLARE_BITMAP(profiles, ICE_MAX_NUM_PROFILES);
+	struct ice_prot_lkup_ext *lkup_exts;
+	struct ice_recp_grp_entry *r_entry;
+	struct ice_sw_fv_list_entry *fvit;
+	struct ice_recp_grp_entry *r_tmp;
+	struct ice_sw_fv_list_entry *tmp;
+	enum ice_status status = 0;
+	struct ice_sw_recipe *rm;
+	u16 match_tun_mask = 0;
+	u8 i;
+
+	if (!lkups_cnt)
+		return ICE_ERR_PARAM;
+
+	lkup_exts = kzalloc(sizeof(*lkup_exts), GFP_KERNEL);
+	if (!lkup_exts)
+		return ICE_ERR_NO_MEMORY;
+
+	/* Determine the number of words to be matched and if it exceeds a
+	 * recipe's restrictions
+	 */
+	for (i = 0; i < lkups_cnt; i++) {
+		u16 count;
+
+		if (lkups[i].type >= ICE_PROTOCOL_LAST) {
+			status = ICE_ERR_CFG;
+			goto err_free_lkup_exts;
+		}
+
+		count = ice_fill_valid_words(&lkups[i], lkup_exts);
+		if (!count) {
+			status = ICE_ERR_CFG;
+			goto err_free_lkup_exts;
+		}
+	}
+
+	rm = kzalloc(sizeof(*rm), GFP_KERNEL);
+	if (!rm) {
+		status = ICE_ERR_NO_MEMORY;
+		goto err_free_lkup_exts;
+	}
+
+	/* Get field vectors that contain fields extracted from all the protocol
+	 * headers being programmed.
+	 */
+	INIT_LIST_HEAD(&rm->fv_list);
+	INIT_LIST_HEAD(&rm->rg_list);
+
+	/* Get bitmap of field vectors (profiles) that are compatible with the
+	 * rule request; only these will be searched in the subsequent call to
+	 * ice_get_fv.
+	 */
+	ice_get_compat_fv_bitmap(hw, rinfo, fv_bitmap);
+
+	status = ice_get_fv(hw, lkups, lkups_cnt, fv_bitmap, &rm->fv_list);
+	if (status)
+		goto err_unroll;
+
+	/* Group match words into recipes using preferred recipe grouping
+	 * criteria.
+	 */
+	status = ice_create_recipe_group(hw, rm, lkup_exts);
+	if (status)
+		goto err_unroll;
+
+	/* set the recipe priority if specified */
+	rm->priority = (u8)rinfo->priority;
+
+	/* Find offsets from the field vector. Pick the first one for all the
+	 * recipes.
+	 */
+	status = ice_fill_fv_word_index(hw, &rm->fv_list, &rm->rg_list);
+	if (status)
+		goto err_unroll;
+
+	/* get bitmap of all profiles the recipe will be associated with */
+	bitmap_zero(profiles, ICE_MAX_NUM_PROFILES);
+	list_for_each_entry(fvit, &rm->fv_list, list_entry) {
+		ice_debug(hw, ICE_DBG_SW, "profile: %d\n", fvit->profile_id);
+		set_bit((u16)fvit->profile_id, profiles);
+	}
+
+	/* Look for a recipe which matches our requested fv / mask list */
+	*rid = ice_find_recp(hw, lkup_exts);
+	if (*rid < ICE_MAX_NUM_RECIPES)
+		/* Success if found a recipe that match the existing criteria */
+		goto err_unroll;
+
+	/* Recipe we need does not exist, add a recipe */
+	status = ice_add_sw_recipe(hw, rm, match_tun_mask, profiles);
+	if (status)
+		goto err_unroll;
+
+	/* Associate all the recipes created with all the profiles in the
+	 * common field vector.
+	 */
+	list_for_each_entry(fvit, &rm->fv_list, list_entry) {
+		DECLARE_BITMAP(r_bitmap, ICE_MAX_NUM_RECIPES);
+		u16 j;
+
+		status = ice_aq_get_recipe_to_profile(hw, fvit->profile_id,
+						      (u8 *)r_bitmap, NULL);
+		if (status)
+			goto err_unroll;
+
+		bitmap_or(r_bitmap, r_bitmap, rm->r_bitmap,
+			  ICE_MAX_NUM_RECIPES);
+		status = ice_acquire_change_lock(hw, ICE_RES_WRITE);
+		if (status)
+			goto err_unroll;
+
+		status = ice_aq_map_recipe_to_profile(hw, fvit->profile_id,
+						      (u8 *)r_bitmap,
+						      NULL);
+		ice_release_change_lock(hw);
+
+		if (status)
+			goto err_unroll;
+
+		/* Update profile to recipe bitmap array */
+		bitmap_copy(profile_to_recipe[fvit->profile_id], r_bitmap,
+			    ICE_MAX_NUM_RECIPES);
+
+		/* Update recipe to profile bitmap array */
+		for_each_set_bit(j, rm->r_bitmap, ICE_MAX_NUM_RECIPES)
+			set_bit((u16)fvit->profile_id, recipe_to_profile[j]);
+	}
+
+	*rid = rm->root_rid;
+	memcpy(&hw->switch_info->recp_list[*rid].lkup_exts, lkup_exts,
+	       sizeof(*lkup_exts));
+err_unroll:
+	list_for_each_entry_safe(r_entry, r_tmp, &rm->rg_list, l_entry) {
+		list_del(&r_entry->l_entry);
+		devm_kfree(ice_hw_to_dev(hw), r_entry);
+	}
+
+	list_for_each_entry_safe(fvit, tmp, &rm->fv_list, list_entry) {
+		list_del(&fvit->list_entry);
+		devm_kfree(ice_hw_to_dev(hw), fvit);
+	}
+
+	if (rm->root_buf)
+		devm_kfree(ice_hw_to_dev(hw), rm->root_buf);
+
+	kfree(rm);
+
+err_free_lkup_exts:
+	kfree(lkup_exts);
+
+	return status;
+}
+
 /**
  * ice_replay_vsi_fltr - Replay filters for requested VSI
  * @hw: pointer to the hardware structure
diff --git a/drivers/net/ethernet/intel/ice/ice_switch.h b/drivers/net/ethernet/intel/ice/ice_switch.h
index 592c564c6de8..56bf54fe462f 100644
--- a/drivers/net/ethernet/intel/ice/ice_switch.h
+++ b/drivers/net/ethernet/intel/ice/ice_switch.h
@@ -134,30 +134,110 @@ struct ice_update_recipe_lkup_idx_params {
 	u8 lkup_idx;
 };
 
+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;
+	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_rule_query_data {
+	/* Recipe ID for which the requested rule was added */
+	u16 rid;
+	/* Rule ID that was added or is supposed to be removed */
+	u16 rule_id;
+	/* vsi_handle for which Rule was added or is supposed to be removed */
+	u16 vsi_handle;
+};
+
+struct ice_adv_rule_info {
+	struct ice_sw_act_ctrl sw_act;
+	u32 priority;
+	u8 rx; /* true means LOOKUP_RX otherwise LOOKUP_TX */
+	u16 fltr_rule_id;
+};
+
+/* A collection of one or more four word recipe */
 struct ice_sw_recipe {
-	struct list_head l_entry;
+	/* For a chained recipe the root recipe is what should be used for
+	 * programming rules
+	 */
+	u8 is_root;
+	u8 root_rid;
+	u8 recp_created;
 
-	/* To protect modification of filt_rule list
-	 * defined below
+	/* Number of extraction words */
+	u8 n_ext_words;
+	/* Protocol ID and Offset pair (extraction word) to describe the
+	 * recipe
 	 */
-	struct mutex filt_rule_lock;
+	struct ice_fv_word ext_words[ICE_MAX_CHAIN_WORDS];
+	u16 word_masks[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 */
+	DECLARE_BITMAP(r_bitmap, ICE_MAX_NUM_RECIPES);
 
-	/* List of type ice_fltr_mgmt_list_entry */
+	/* List of type ice_fltr_mgmt_list_entry or adv_rule */
+	u8 adv_rule;
 	struct list_head filt_rules;
 	struct list_head filt_replay_rules;
 
-	/* linked list of type recipe_list_entry */
-	struct list_head rg_list;
-	/* linked list of type ice_sw_fv_list_entry*/
+	struct mutex filt_rule_lock;	/* protect filter rule structure */
+
+	/* Profiles this recipe should be associated with */
 	struct list_head fv_list;
-	struct ice_aqc_recipe_data_elem *r_buf;
-	u8 recp_count;
-	u8 root_rid;
-	u8 num_profs;
-	u8 *prof_ids;
 
-	/* recipe bitmap: what all recipes makes this recipe */
-	DECLARE_BITMAP(r_bitmap, ICE_MAX_NUM_RECIPES);
+	/* Profiles this recipe is associated with */
+	u8 num_profs, *prof_ids;
+
+	/* Bit map for possible result indexes */
+	DECLARE_BITMAP(res_idxs, ICE_MAX_FV_WORDS);
+
+	/* 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 rg_list;
+
+	/* AQ buffer associated with this recipe */
+	struct ice_aqc_recipe_data_elem *root_buf;
+	/* This struct saves the fv_words for a given lookup */
+	struct ice_prot_lkup_ext lkup_exts;
 };
 
 /* Bookkeeping structure to hold bitmap of VSIs corresponding to VSI list ID */
-- 
1.8.3.1

[Intel-wired-lan] [PATCH net-next v6 4/9] ice: allow adding advanced rules

[Wojciech Drewek]

From: Grishma Kotecha <grishma.kotecha@intel.com>

Define dummy packet headers to allow adding advanced rules in HW. This
header is used as admin queue command parameter for adding a rule.
The firmware will extract correct fields and will use them in look ups.

Define each supported packets header and offsets to words used in recipe.
Supported headers:
- MAC + IPv4 + UDP
- MAC + VLAN + IPv4 + UDP
- MAC + IPv4 + TCP
- MAC + VLAN + IPv4 + TCP
- MAC + IPv6 + UDP
- MAC + VLAN + IPv6 + UDP
- MAC + IPv6 + TCP
- MAC + VLAN + IPv6 + TCP

Add code for creating an advanced rule. Rule needs to match defined
dummy packet, if not return error, which means that this type of rule
isn't currently supported.

The first step in adding advanced rule is searching for an advanced
recipe matching this kind of rule. If it doesn't exist new recipe is
created. Dummy packet has to be filled with the correct header field
value from the rule definition. It will be used to do look up in HW.

Support searching for existing advance rule entry. It is used in case
of adding the same rule on different VSI. In this case, instead of
creating new rule, the existing one should be updated with refreshed VSI
list.

Add initialization for prof_res_bm_init flag to zero so that
the possible resource for fv in the files can be initialized.

Co-developed-by: Dan Nowlin <dan.nowlin@intel.com>
Signed-off-by: Dan Nowlin <dan.nowlin@intel.com>
Signed-off-by: Grishma Kotecha <grishma.kotecha@intel.com>
Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_common.c |   1 +
 drivers/net/ethernet/intel/ice/ice_switch.c | 832 +++++++++++++++++++++++++++-
 drivers/net/ethernet/intel/ice/ice_switch.h |  14 +
 drivers/net/ethernet/intel/ice/ice_type.h   |   1 +
 4 files changed, 847 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_common.c b/drivers/net/ethernet/intel/ice/ice_common.c
index f63912797548..9d32eb970028 100644
--- a/drivers/net/ethernet/intel/ice/ice_common.c
+++ b/drivers/net/ethernet/intel/ice/ice_common.c
@@ -568,6 +568,7 @@ static enum ice_status ice_init_fltr_mgmt_struct(struct ice_hw *hw)
 		return ICE_ERR_NO_MEMORY;
 
 	INIT_LIST_HEAD(&sw->vsi_list_map_head);
+	sw->prof_res_bm_init = 0;
 
 	status = ice_init_def_sw_recp(hw);
 	if (status) {
diff --git a/drivers/net/ethernet/intel/ice/ice_switch.c b/drivers/net/ethernet/intel/ice/ice_switch.c
index 8274f2e8df38..04cbd9d417fb 100644
--- a/drivers/net/ethernet/intel/ice/ice_switch.c
+++ b/drivers/net/ethernet/intel/ice/ice_switch.c
@@ -8,6 +8,7 @@
 #define ICE_ETH_ETHTYPE_OFFSET		12
 #define ICE_ETH_VLAN_TCI_OFFSET		14
 #define ICE_MAX_VLAN_ID			0xFFF
+#define ICE_IPV6_ETHER_ID		0x86DD
 
 /* Dummy ethernet header needed in the ice_aqc_sw_rules_elem
  * struct to configure any switch filter rules.
@@ -29,6 +30,290 @@
 							0x2, 0, 0, 0, 0, 0,
 							0x81, 0, 0, 0};
 
+struct ice_dummy_pkt_offsets {
+	enum ice_protocol_type type;
+	u16 offset; /* ICE_PROTOCOL_LAST indicates end of list */
+};
+
+/* offset info for MAC + IPv4 + UDP dummy packet */
+static const struct ice_dummy_pkt_offsets dummy_udp_packet_offsets[] = {
+	{ ICE_MAC_OFOS,		0 },
+	{ ICE_ETYPE_OL,		12 },
+	{ ICE_IPV4_OFOS,	14 },
+	{ ICE_UDP_ILOS,		34 },
+	{ ICE_PROTOCOL_LAST,	0 },
+};
+
+/* Dummy packet for MAC + IPv4 + UDP */
+static const u8 dummy_udp_packet[] = {
+	0x00, 0x00, 0x00, 0x00, /* ICE_MAC_OFOS 0 */
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+
+	0x08, 0x00,		/* ICE_ETYPE_OL 12 */
+
+	0x45, 0x00, 0x00, 0x1c, /* ICE_IPV4_OFOS 14 */
+	0x00, 0x01, 0x00, 0x00,
+	0x00, 0x11, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+
+	0x00, 0x00, 0x00, 0x00, /* ICE_UDP_ILOS 34 */
+	0x00, 0x08, 0x00, 0x00,
+
+	0x00, 0x00,	/* 2 bytes for 4 byte alignment */
+};
+
+/* offset info for MAC + VLAN + IPv4 + UDP dummy packet */
+static const struct ice_dummy_pkt_offsets dummy_vlan_udp_packet_offsets[] = {
+	{ ICE_MAC_OFOS,		0 },
+	{ ICE_VLAN_OFOS,	12 },
+	{ ICE_ETYPE_OL,		16 },
+	{ ICE_IPV4_OFOS,	18 },
+	{ ICE_UDP_ILOS,		38 },
+	{ ICE_PROTOCOL_LAST,	0 },
+};
+
+/* C-tag (801.1Q), IPv4:UDP dummy packet */
+static const u8 dummy_vlan_udp_packet[] = {
+	0x00, 0x00, 0x00, 0x00, /* ICE_MAC_OFOS 0 */
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+
+	0x81, 0x00, 0x00, 0x00, /* ICE_VLAN_OFOS 12 */
+
+	0x08, 0x00,		/* ICE_ETYPE_OL 16 */
+
+	0x45, 0x00, 0x00, 0x1c, /* ICE_IPV4_OFOS 18 */
+	0x00, 0x01, 0x00, 0x00,
+	0x00, 0x11, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+
+	0x00, 0x00, 0x00, 0x00, /* ICE_UDP_ILOS 38 */
+	0x00, 0x08, 0x00, 0x00,
+
+	0x00, 0x00,	/* 2 bytes for 4 byte alignment */
+};
+
+/* offset info for MAC + IPv4 + TCP dummy packet */
+static const struct ice_dummy_pkt_offsets dummy_tcp_packet_offsets[] = {
+	{ ICE_MAC_OFOS,		0 },
+	{ ICE_ETYPE_OL,		12 },
+	{ ICE_IPV4_OFOS,	14 },
+	{ ICE_TCP_IL,		34 },
+	{ ICE_PROTOCOL_LAST,	0 },
+};
+
+/* Dummy packet for MAC + IPv4 + TCP */
+static const u8 dummy_tcp_packet[] = {
+	0x00, 0x00, 0x00, 0x00, /* ICE_MAC_OFOS 0 */
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+
+	0x08, 0x00,		/* ICE_ETYPE_OL 12 */
+
+	0x45, 0x00, 0x00, 0x28, /* ICE_IPV4_OFOS 14 */
+	0x00, 0x01, 0x00, 0x00,
+	0x00, 0x06, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+
+	0x00, 0x00, 0x00, 0x00, /* ICE_TCP_IL 34 */
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x50, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+
+	0x00, 0x00,	/* 2 bytes for 4 byte alignment */
+};
+
+/* offset info for MAC + VLAN (C-tag, 802.1Q) + IPv4 + TCP dummy packet */
+static const struct ice_dummy_pkt_offsets dummy_vlan_tcp_packet_offsets[] = {
+	{ ICE_MAC_OFOS,		0 },
+	{ ICE_VLAN_OFOS,	12 },
+	{ ICE_ETYPE_OL,		16 },
+	{ ICE_IPV4_OFOS,	18 },
+	{ ICE_TCP_IL,		38 },
+	{ ICE_PROTOCOL_LAST,	0 },
+};
+
+/* C-tag (801.1Q), IPv4:TCP dummy packet */
+static const u8 dummy_vlan_tcp_packet[] = {
+	0x00, 0x00, 0x00, 0x00, /* ICE_MAC_OFOS 0 */
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+
+	0x81, 0x00, 0x00, 0x00,	/* ICE_VLAN_OFOS 12 */
+
+	0x08, 0x00,		/* ICE_ETYPE_OL 16 */
+
+	0x45, 0x00, 0x00, 0x28, /* ICE_IPV4_OFOS 18 */
+	0x00, 0x01, 0x00, 0x00,
+	0x00, 0x06, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+
+	0x00, 0x00, 0x00, 0x00, /* ICE_TCP_IL 38 */
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x50, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+
+	0x00, 0x00,	/* 2 bytes for 4 byte alignment */
+};
+
+static const struct ice_dummy_pkt_offsets dummy_tcp_ipv6_packet_offsets[] = {
+	{ ICE_MAC_OFOS,		0 },
+	{ ICE_ETYPE_OL,		12 },
+	{ ICE_IPV6_OFOS,	14 },
+	{ ICE_TCP_IL,		54 },
+	{ ICE_PROTOCOL_LAST,	0 },
+};
+
+static const u8 dummy_tcp_ipv6_packet[] = {
+	0x00, 0x00, 0x00, 0x00, /* ICE_MAC_OFOS 0 */
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+
+	0x86, 0xDD,		/* ICE_ETYPE_OL 12 */
+
+	0x60, 0x00, 0x00, 0x00, /* ICE_IPV6_OFOS 40 */
+	0x00, 0x14, 0x06, 0x00, /* Next header is TCP */
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+
+	0x00, 0x00, 0x00, 0x00, /* ICE_TCP_IL 54 */
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x50, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+
+	0x00, 0x00, /* 2 bytes for 4 byte alignment */
+};
+
+/* C-tag (802.1Q): IPv6 + TCP */
+static const struct ice_dummy_pkt_offsets
+dummy_vlan_tcp_ipv6_packet_offsets[] = {
+	{ ICE_MAC_OFOS,		0 },
+	{ ICE_VLAN_OFOS,	12 },
+	{ ICE_ETYPE_OL,		16 },
+	{ ICE_IPV6_OFOS,	18 },
+	{ ICE_TCP_IL,		58 },
+	{ ICE_PROTOCOL_LAST,	0 },
+};
+
+/* C-tag (802.1Q), IPv6 + TCP dummy packet */
+static const u8 dummy_vlan_tcp_ipv6_packet[] = {
+	0x00, 0x00, 0x00, 0x00, /* ICE_MAC_OFOS 0 */
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+
+	0x81, 0x00, 0x00, 0x00,	/* ICE_VLAN_OFOS 12 */
+
+	0x86, 0xDD,		/* ICE_ETYPE_OL 16 */
+
+	0x60, 0x00, 0x00, 0x00, /* ICE_IPV6_OFOS 18 */
+	0x00, 0x14, 0x06, 0x00, /* Next header is TCP */
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+
+	0x00, 0x00, 0x00, 0x00, /* ICE_TCP_IL 58 */
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x50, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+
+	0x00, 0x00, /* 2 bytes for 4 byte alignment */
+};
+
+/* IPv6 + UDP */
+static const struct ice_dummy_pkt_offsets dummy_udp_ipv6_packet_offsets[] = {
+	{ ICE_MAC_OFOS,		0 },
+	{ ICE_ETYPE_OL,		12 },
+	{ ICE_IPV6_OFOS,	14 },
+	{ ICE_UDP_ILOS,		54 },
+	{ ICE_PROTOCOL_LAST,	0 },
+};
+
+/* IPv6 + UDP dummy packet */
+static const u8 dummy_udp_ipv6_packet[] = {
+	0x00, 0x00, 0x00, 0x00, /* ICE_MAC_OFOS 0 */
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+
+	0x86, 0xDD,		/* ICE_ETYPE_OL 12 */
+
+	0x60, 0x00, 0x00, 0x00, /* ICE_IPV6_OFOS 40 */
+	0x00, 0x10, 0x11, 0x00, /* Next header UDP */
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+
+	0x00, 0x00, 0x00, 0x00, /* ICE_UDP_ILOS 54 */
+	0x00, 0x10, 0x00, 0x00,
+
+	0x00, 0x00, 0x00, 0x00, /* needed for ESP packets */
+	0x00, 0x00, 0x00, 0x00,
+
+	0x00, 0x00, /* 2 bytes for 4 byte alignment */
+};
+
+/* C-tag (802.1Q): IPv6 + UDP */
+static const struct ice_dummy_pkt_offsets
+dummy_vlan_udp_ipv6_packet_offsets[] = {
+	{ ICE_MAC_OFOS,		0 },
+	{ ICE_VLAN_OFOS,	12 },
+	{ ICE_ETYPE_OL,		16 },
+	{ ICE_IPV6_OFOS,	18 },
+	{ ICE_UDP_ILOS,		58 },
+	{ ICE_PROTOCOL_LAST,	0 },
+};
+
+/* C-tag (802.1Q), IPv6 + UDP dummy packet */
+static const u8 dummy_vlan_udp_ipv6_packet[] = {
+	0x00, 0x00, 0x00, 0x00, /* ICE_MAC_OFOS 0 */
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+
+	0x81, 0x00, 0x00, 0x00,/* ICE_VLAN_OFOS 12 */
+
+	0x86, 0xDD,		/* ICE_ETYPE_OL 16 */
+
+	0x60, 0x00, 0x00, 0x00, /* ICE_IPV6_OFOS 18 */
+	0x00, 0x08, 0x11, 0x00, /* Next header UDP */
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00,
+
+	0x00, 0x00, 0x00, 0x00, /* ICE_UDP_ILOS 58 */
+	0x00, 0x08, 0x00, 0x00,
+
+	0x00, 0x00, /* 2 bytes for 4 byte alignment */
+};
+
 #define ICE_SW_RULE_RX_TX_ETH_HDR_SIZE \
 	(offsetof(struct ice_aqc_sw_rules_elem, pdata.lkup_tx_rx.hdr) + \
 	 (DUMMY_ETH_HDR_LEN * \
@@ -3935,7 +4220,7 @@ static bool ice_prot_type_to_id(enum ice_protocol_type type, u8 *id)
  * @rinfo: other information regarding the rule e.g. priority and action info
  * @rid: return the recipe ID of the recipe created
  */
-static enum ice_status __maybe_unused
+static enum ice_status
 ice_add_adv_recipe(struct ice_hw *hw, struct ice_adv_lkup_elem *lkups,
 		   u16 lkups_cnt, struct ice_adv_rule_info *rinfo, u16 *rid)
 {
@@ -4094,6 +4379,551 @@ static bool ice_prot_type_to_id(enum ice_protocol_type type, u8 *id)
 }
 
 /**
+ * ice_find_dummy_packet - find dummy packet
+ *
+ * @lkups: lookup elements or match criteria for the advanced recipe, one
+ *	   structure per protocol header
+ * @lkups_cnt: number of protocols
+ * @pkt: dummy packet to fill according to filter match criteria
+ * @pkt_len: packet length of dummy packet
+ * @offsets: pointer to receive the pointer to the offsets for the packet
+ */
+static void
+ice_find_dummy_packet(struct ice_adv_lkup_elem *lkups, u16 lkups_cnt,
+		      const u8 **pkt, u16 *pkt_len,
+		      const struct ice_dummy_pkt_offsets **offsets)
+{
+	bool tcp = false, udp = false, ipv6 = false, vlan = false;
+	u16 i;
+
+	for (i = 0; i < lkups_cnt; i++) {
+		if (lkups[i].type == ICE_UDP_ILOS)
+			udp = true;
+		else if (lkups[i].type == ICE_TCP_IL)
+			tcp = true;
+		else if (lkups[i].type == ICE_IPV6_OFOS)
+			ipv6 = true;
+		else if (lkups[i].type == ICE_VLAN_OFOS)
+			vlan = true;
+		else if (lkups[i].type == ICE_ETYPE_OL &&
+			 lkups[i].h_u.ethertype.ethtype_id ==
+				cpu_to_be16(ICE_IPV6_ETHER_ID) &&
+			 lkups[i].m_u.ethertype.ethtype_id ==
+					cpu_to_be16(0xFFFF))
+			ipv6 = true;
+	}
+
+	if (udp && !ipv6) {
+		if (vlan) {
+			*pkt = dummy_vlan_udp_packet;
+			*pkt_len = sizeof(dummy_vlan_udp_packet);
+			*offsets = dummy_vlan_udp_packet_offsets;
+			return;
+		}
+		*pkt = dummy_udp_packet;
+		*pkt_len = sizeof(dummy_udp_packet);
+		*offsets = dummy_udp_packet_offsets;
+		return;
+	} else if (udp && ipv6) {
+		if (vlan) {
+			*pkt = dummy_vlan_udp_ipv6_packet;
+			*pkt_len = sizeof(dummy_vlan_udp_ipv6_packet);
+			*offsets = dummy_vlan_udp_ipv6_packet_offsets;
+			return;
+		}
+		*pkt = dummy_udp_ipv6_packet;
+		*pkt_len = sizeof(dummy_udp_ipv6_packet);
+		*offsets = dummy_udp_ipv6_packet_offsets;
+		return;
+	} else if ((tcp && ipv6) || ipv6) {
+		if (vlan) {
+			*pkt = dummy_vlan_tcp_ipv6_packet;
+			*pkt_len = sizeof(dummy_vlan_tcp_ipv6_packet);
+			*offsets = dummy_vlan_tcp_ipv6_packet_offsets;
+			return;
+		}
+		*pkt = dummy_tcp_ipv6_packet;
+		*pkt_len = sizeof(dummy_tcp_ipv6_packet);
+		*offsets = dummy_tcp_ipv6_packet_offsets;
+		return;
+	}
+
+	if (vlan) {
+		*pkt = dummy_vlan_tcp_packet;
+		*pkt_len = sizeof(dummy_vlan_tcp_packet);
+		*offsets = dummy_vlan_tcp_packet_offsets;
+	} else {
+		*pkt = dummy_tcp_packet;
+		*pkt_len = sizeof(dummy_tcp_packet);
+		*offsets = dummy_tcp_packet_offsets;
+	}
+}
+
+/**
+ * ice_fill_adv_dummy_packet - fill a dummy packet with given match criteria
+ *
+ * @lkups: lookup elements or match criteria for the advanced recipe, one
+ *	   structure per protocol header
+ * @lkups_cnt: number of protocols
+ * @s_rule: stores rule information from the match criteria
+ * @dummy_pkt: dummy packet to fill according to filter match criteria
+ * @pkt_len: packet length of dummy packet
+ * @offsets: offset info for the dummy packet
+ */
+static enum ice_status
+ice_fill_adv_dummy_packet(struct ice_adv_lkup_elem *lkups, u16 lkups_cnt,
+			  struct ice_aqc_sw_rules_elem *s_rule,
+			  const u8 *dummy_pkt, u16 pkt_len,
+			  const struct ice_dummy_pkt_offsets *offsets)
+{
+	u8 *pkt;
+	u16 i;
+
+	/* Start with a packet with a pre-defined/dummy content. Then, fill
+	 * in the header values to be looked up or matched.
+	 */
+	pkt = s_rule->pdata.lkup_tx_rx.hdr;
+
+	memcpy(pkt, dummy_pkt, pkt_len);
+
+	for (i = 0; i < lkups_cnt; i++) {
+		enum ice_protocol_type type;
+		u16 offset = 0, len = 0, j;
+		bool found = false;
+
+		/* find the start of this layer; it should be found since this
+		 * was already checked when search for the dummy packet
+		 */
+		type = lkups[i].type;
+		for (j = 0; offsets[j].type != ICE_PROTOCOL_LAST; j++) {
+			if (type == offsets[j].type) {
+				offset = offsets[j].offset;
+				found = true;
+				break;
+			}
+		}
+		/* this should never happen in a correct calling sequence */
+		if (!found)
+			return ICE_ERR_PARAM;
+
+		switch (lkups[i].type) {
+		case ICE_MAC_OFOS:
+		case ICE_MAC_IL:
+			len = sizeof(struct ice_ether_hdr);
+			break;
+		case ICE_ETYPE_OL:
+			len = sizeof(struct ice_ethtype_hdr);
+			break;
+		case ICE_VLAN_OFOS:
+			len = sizeof(struct ice_vlan_hdr);
+			break;
+		case ICE_IPV4_OFOS:
+		case ICE_IPV4_IL:
+			len = sizeof(struct ice_ipv4_hdr);
+			break;
+		case ICE_IPV6_OFOS:
+		case ICE_IPV6_IL:
+			len = sizeof(struct ice_ipv6_hdr);
+			break;
+		case ICE_TCP_IL:
+		case ICE_UDP_OF:
+		case ICE_UDP_ILOS:
+			len = sizeof(struct ice_l4_hdr);
+			break;
+		case ICE_SCTP_IL:
+			len = sizeof(struct ice_sctp_hdr);
+			break;
+		default:
+			return ICE_ERR_PARAM;
+		}
+
+		/* the length should be a word multiple */
+		if (len % ICE_BYTES_PER_WORD)
+			return ICE_ERR_CFG;
+
+		/* We have the offset to the header start, the length, the
+		 * caller's header values and mask. Use this information to
+		 * copy the data into the dummy packet appropriately based on
+		 * the mask. Note that we need to only write the bits as
+		 * indicated by the mask to make sure we don't improperly write
+		 * over any significant packet data.
+		 */
+		for (j = 0; j < len / sizeof(u16); j++)
+			if (((u16 *)&lkups[i].m_u)[j])
+				((u16 *)(pkt + offset))[j] =
+					(((u16 *)(pkt + offset))[j] &
+					 ~((u16 *)&lkups[i].m_u)[j]) |
+					(((u16 *)&lkups[i].h_u)[j] &
+					 ((u16 *)&lkups[i].m_u)[j]);
+	}
+
+	s_rule->pdata.lkup_tx_rx.hdr_len = cpu_to_le16(pkt_len);
+
+	return 0;
+}
+
+/**
+ * ice_find_adv_rule_entry - Search a rule entry
+ * @hw: pointer to the hardware structure
+ * @lkups: lookup elements or match criteria for the advanced recipe, one
+ *	   structure per protocol header
+ * @lkups_cnt: number of protocols
+ * @recp_id: recipe ID for which we are finding the rule
+ * @rinfo: other information regarding the rule e.g. priority and action info
+ *
+ * Helper function to search for a given advance rule entry
+ * Returns pointer to entry storing the rule if found
+ */
+static struct ice_adv_fltr_mgmt_list_entry *
+ice_find_adv_rule_entry(struct ice_hw *hw, struct ice_adv_lkup_elem *lkups,
+			u16 lkups_cnt, u16 recp_id,
+			struct ice_adv_rule_info *rinfo)
+{
+	struct ice_adv_fltr_mgmt_list_entry *list_itr;
+	struct ice_switch_info *sw = hw->switch_info;
+	int i;
+
+	list_for_each_entry(list_itr, &sw->recp_list[recp_id].filt_rules,
+			    list_entry) {
+		bool lkups_matched = true;
+
+		if (lkups_cnt != list_itr->lkups_cnt)
+			continue;
+		for (i = 0; i < list_itr->lkups_cnt; i++)
+			if (memcmp(&list_itr->lkups[i], &lkups[i],
+				   sizeof(*lkups))) {
+				lkups_matched = false;
+				break;
+			}
+		if (rinfo->sw_act.flag == list_itr->rule_info.sw_act.flag &&
+		    lkups_matched)
+			return list_itr;
+	}
+	return NULL;
+}
+
+/**
+ * ice_adv_add_update_vsi_list
+ * @hw: pointer to the hardware structure
+ * @m_entry: pointer to current adv 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 booking keeping is described below :
+ * When a VSI needs to subscribe to a given advanced filter
+ *	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_adv_add_update_vsi_list(struct ice_hw *hw,
+			    struct ice_adv_fltr_mgmt_list_entry *m_entry,
+			    struct ice_adv_rule_info *cur_fltr,
+			    struct ice_adv_rule_info *new_fltr)
+{
+	enum ice_status status;
+	u16 vsi_list_id = 0;
+
+	if (cur_fltr->sw_act.fltr_act == ICE_FWD_TO_Q ||
+	    cur_fltr->sw_act.fltr_act == ICE_FWD_TO_QGRP ||
+	    cur_fltr->sw_act.fltr_act == ICE_DROP_PACKET)
+		return ICE_ERR_NOT_IMPL;
+
+	if ((new_fltr->sw_act.fltr_act == ICE_FWD_TO_Q ||
+	     new_fltr->sw_act.fltr_act == ICE_FWD_TO_QGRP) &&
+	    (cur_fltr->sw_act.fltr_act == ICE_FWD_TO_VSI ||
+	     cur_fltr->sw_act.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->sw_act.fwd_id.hw_vsi_id ==
+		    new_fltr->sw_act.fwd_id.hw_vsi_id)
+			return ICE_ERR_ALREADY_EXISTS;
+
+		vsi_handle_arr[0] = cur_fltr->sw_act.vsi_handle;
+		vsi_handle_arr[1] = new_fltr->sw_act.vsi_handle;
+		status = ice_create_vsi_list_rule(hw, &vsi_handle_arr[0], 2,
+						  &vsi_list_id,
+						  ICE_SW_LKUP_LAST);
+		if (status)
+			return status;
+
+		memset(&tmp_fltr, 0, sizeof(tmp_fltr));
+		tmp_fltr.flag = m_entry->rule_info.sw_act.flag;
+		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;
+		tmp_fltr.lkup_type = ICE_SW_LKUP_LAST;
+
+		/* Update the previous switch rule of "forward to VSI" to
+		 * "fwd to VSI list"
+		 */
+		status = ice_update_pkt_fwd_rule(hw, &tmp_fltr);
+		if (status)
+			return status;
+
+		cur_fltr->sw_act.fwd_id.vsi_list_id = vsi_list_id;
+		cur_fltr->sw_act.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);
+	} else {
+		u16 vsi_handle = new_fltr->sw_act.vsi_handle;
+
+		if (!m_entry->vsi_list_info)
+			return ICE_ERR_CFG;
+
+		/* A rule already exists with the new VSI being added */
+		if (test_bit(vsi_handle, m_entry->vsi_list_info->vsi_map))
+			return 0;
+
+		/* Update the previously created VSI list set with
+		 * the new VSI ID passed in
+		 */
+		vsi_list_id = cur_fltr->sw_act.fwd_id.vsi_list_id;
+
+		status = ice_update_vsi_list_rule(hw, &vsi_handle, 1,
+						  vsi_list_id, false,
+						  ice_aqc_opc_update_sw_rules,
+						  ICE_SW_LKUP_LAST);
+		/* update VSI list mapping info with new VSI ID */
+		if (!status)
+			set_bit(vsi_handle, m_entry->vsi_list_info->vsi_map);
+	}
+	if (!status)
+		m_entry->vsi_count++;
+	return status;
+}
+
+/**
+ * ice_add_adv_rule - helper function to create an advanced switch rule
+ * @hw: pointer to the hardware structure
+ * @lkups: information on the words that needs to be looked up. All words
+ * together makes one recipe
+ * @lkups_cnt: num of entries in the lkups array
+ * @rinfo: other information related to the rule that needs to be programmed
+ * @added_entry: this will return recipe_id, rule_id and vsi_handle. should be
+ *               ignored is case of error.
+ *
+ * This function can program only 1 rule at a time. The lkups is used to
+ * describe the all the words that forms the "lookup" portion of the recipe.
+ * These words can span multiple protocols. Callers to this function need to
+ * pass in a list of protocol headers with lookup information along and mask
+ * that determines which words are valid from the given protocol header.
+ * rinfo describes other information related to this rule such as forwarding
+ * IDs, priority of this rule, etc.
+ */
+enum ice_status
+ice_add_adv_rule(struct ice_hw *hw, struct ice_adv_lkup_elem *lkups,
+		 u16 lkups_cnt, struct ice_adv_rule_info *rinfo,
+		 struct ice_rule_query_data *added_entry)
+{
+	struct ice_adv_fltr_mgmt_list_entry *m_entry, *adv_fltr = NULL;
+	u16 rid = 0, i, pkt_len, rule_buf_sz, vsi_handle;
+	const struct ice_dummy_pkt_offsets *pkt_offsets;
+	struct ice_aqc_sw_rules_elem *s_rule = NULL;
+	struct list_head *rule_head;
+	struct ice_switch_info *sw;
+	enum ice_status status;
+	const u8 *pkt = NULL;
+	u16 word_cnt;
+	u32 act = 0;
+	u8 q_rgn;
+
+	/* Initialize profile to result index bitmap */
+	if (!hw->switch_info->prof_res_bm_init) {
+		hw->switch_info->prof_res_bm_init = 1;
+		ice_init_prof_result_bm(hw);
+	}
+
+	if (!lkups_cnt)
+		return ICE_ERR_PARAM;
+
+	/* get # of words we need to match */
+	word_cnt = 0;
+	for (i = 0; i < lkups_cnt; i++) {
+		u16 j, *ptr;
+
+		ptr = (u16 *)&lkups[i].m_u;
+		for (j = 0; j < sizeof(lkups->m_u) / sizeof(u16); j++)
+			if (ptr[j] != 0)
+				word_cnt++;
+	}
+
+	if (!word_cnt || word_cnt > ICE_MAX_CHAIN_WORDS)
+		return ICE_ERR_PARAM;
+
+	/* make sure that we can locate a dummy packet */
+	ice_find_dummy_packet(lkups, lkups_cnt, &pkt, &pkt_len,
+			      &pkt_offsets);
+	if (!pkt) {
+		status = ICE_ERR_PARAM;
+		goto err_ice_add_adv_rule;
+	}
+
+	if (!(rinfo->sw_act.fltr_act == ICE_FWD_TO_VSI ||
+	      rinfo->sw_act.fltr_act == ICE_FWD_TO_Q ||
+	      rinfo->sw_act.fltr_act == ICE_FWD_TO_QGRP ||
+	      rinfo->sw_act.fltr_act == ICE_DROP_PACKET))
+		return ICE_ERR_CFG;
+
+	vsi_handle = rinfo->sw_act.vsi_handle;
+	if (!ice_is_vsi_valid(hw, vsi_handle))
+		return ICE_ERR_PARAM;
+
+	if (rinfo->sw_act.fltr_act == ICE_FWD_TO_VSI)
+		rinfo->sw_act.fwd_id.hw_vsi_id =
+			ice_get_hw_vsi_num(hw, vsi_handle);
+	if (rinfo->sw_act.flag & ICE_FLTR_TX)
+		rinfo->sw_act.src = ice_get_hw_vsi_num(hw, vsi_handle);
+
+	status = ice_add_adv_recipe(hw, lkups, lkups_cnt, rinfo, &rid);
+	if (status)
+		return status;
+	m_entry = ice_find_adv_rule_entry(hw, lkups, lkups_cnt, rid, rinfo);
+	if (m_entry) {
+		/* we have to add VSI to VSI_LIST and increment vsi_count.
+		 * Also Update VSI list so that we can change forwarding rule
+		 * if the rule already exists, we will check if it exists with
+		 * same vsi_id, if not then add it to the VSI list if it already
+		 * exists if not then create a VSI list and add the existing VSI
+		 * ID and the new VSI ID to the list
+		 * We will add that VSI to the list
+		 */
+		status = ice_adv_add_update_vsi_list(hw, m_entry,
+						     &m_entry->rule_info,
+						     rinfo);
+		if (added_entry) {
+			added_entry->rid = rid;
+			added_entry->rule_id = m_entry->rule_info.fltr_rule_id;
+			added_entry->vsi_handle = rinfo->sw_act.vsi_handle;
+		}
+		return status;
+	}
+	rule_buf_sz = ICE_SW_RULE_RX_TX_NO_HDR_SIZE + pkt_len;
+	s_rule = kzalloc(rule_buf_sz, GFP_KERNEL);
+	if (!s_rule)
+		return ICE_ERR_NO_MEMORY;
+	act |= ICE_SINGLE_ACT_LB_ENABLE | ICE_SINGLE_ACT_LAN_ENABLE;
+	switch (rinfo->sw_act.fltr_act) {
+	case ICE_FWD_TO_VSI:
+		act |= (rinfo->sw_act.fwd_id.hw_vsi_id <<
+			ICE_SINGLE_ACT_VSI_ID_S) & ICE_SINGLE_ACT_VSI_ID_M;
+		act |= ICE_SINGLE_ACT_VSI_FORWARDING | ICE_SINGLE_ACT_VALID_BIT;
+		break;
+	case ICE_FWD_TO_Q:
+		act |= ICE_SINGLE_ACT_TO_Q;
+		act |= (rinfo->sw_act.fwd_id.q_id << ICE_SINGLE_ACT_Q_INDEX_S) &
+		       ICE_SINGLE_ACT_Q_INDEX_M;
+		break;
+	case ICE_FWD_TO_QGRP:
+		q_rgn = rinfo->sw_act.qgrp_size > 0 ?
+			(u8)ilog2(rinfo->sw_act.qgrp_size) : 0;
+		act |= ICE_SINGLE_ACT_TO_Q;
+		act |= (rinfo->sw_act.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;
+	case ICE_DROP_PACKET:
+		act |= ICE_SINGLE_ACT_VSI_FORWARDING | ICE_SINGLE_ACT_DROP |
+		       ICE_SINGLE_ACT_VALID_BIT;
+		break;
+	default:
+		status = ICE_ERR_CFG;
+		goto err_ice_add_adv_rule;
+	}
+
+	/* set the rule LOOKUP type based on caller specified 'Rx'
+	 * instead of hardcoding it to be either LOOKUP_TX/RX
+	 *
+	 * for 'Rx' set the source to be the port number
+	 * for 'Tx' set the source to be the source HW VSI number (determined
+	 * by caller)
+	 */
+	if (rinfo->rx) {
+		s_rule->type = cpu_to_le16(ICE_AQC_SW_RULES_T_LKUP_RX);
+		s_rule->pdata.lkup_tx_rx.src =
+			cpu_to_le16(hw->port_info->lport);
+	} else {
+		s_rule->type = cpu_to_le16(ICE_AQC_SW_RULES_T_LKUP_TX);
+		s_rule->pdata.lkup_tx_rx.src = cpu_to_le16(rinfo->sw_act.src);
+	}
+
+	s_rule->pdata.lkup_tx_rx.recipe_id = cpu_to_le16(rid);
+	s_rule->pdata.lkup_tx_rx.act = cpu_to_le32(act);
+
+	status = ice_fill_adv_dummy_packet(lkups, lkups_cnt, s_rule, pkt,
+					   pkt_len, pkt_offsets);
+	if (status)
+		goto err_ice_add_adv_rule;
+
+	status = ice_aq_sw_rules(hw, (struct ice_aqc_sw_rules *)s_rule,
+				 rule_buf_sz, 1, ice_aqc_opc_add_sw_rules,
+				 NULL);
+	if (status)
+		goto err_ice_add_adv_rule;
+	adv_fltr = devm_kzalloc(ice_hw_to_dev(hw),
+				sizeof(struct ice_adv_fltr_mgmt_list_entry),
+				GFP_KERNEL);
+	if (!adv_fltr) {
+		status = ICE_ERR_NO_MEMORY;
+		goto err_ice_add_adv_rule;
+	}
+
+	adv_fltr->lkups = devm_kmemdup(ice_hw_to_dev(hw), lkups,
+				       lkups_cnt * sizeof(*lkups), GFP_KERNEL);
+	if (!adv_fltr->lkups) {
+		status = ICE_ERR_NO_MEMORY;
+		goto err_ice_add_adv_rule;
+	}
+
+	adv_fltr->lkups_cnt = lkups_cnt;
+	adv_fltr->rule_info = *rinfo;
+	adv_fltr->rule_info.fltr_rule_id =
+		le16_to_cpu(s_rule->pdata.lkup_tx_rx.index);
+	sw = hw->switch_info;
+	sw->recp_list[rid].adv_rule = true;
+	rule_head = &sw->recp_list[rid].filt_rules;
+
+	if (rinfo->sw_act.fltr_act == ICE_FWD_TO_VSI)
+		adv_fltr->vsi_count = 1;
+
+	/* Add rule entry to book keeping list */
+	list_add(&adv_fltr->list_entry, rule_head);
+	if (added_entry) {
+		added_entry->rid = rid;
+		added_entry->rule_id = adv_fltr->rule_info.fltr_rule_id;
+		added_entry->vsi_handle = rinfo->sw_act.vsi_handle;
+	}
+err_ice_add_adv_rule:
+	if (status && adv_fltr) {
+		devm_kfree(ice_hw_to_dev(hw), adv_fltr->lkups);
+		devm_kfree(ice_hw_to_dev(hw), adv_fltr);
+	}
+
+	kfree(s_rule);
+
+	return status;
+}
+
+/**
  * ice_replay_vsi_fltr - Replay filters for requested VSI
  * @hw: pointer to the hardware structure
  * @vsi_handle: driver VSI handle
diff --git a/drivers/net/ethernet/intel/ice/ice_switch.h b/drivers/net/ethernet/intel/ice/ice_switch.h
index 56bf54fe462f..082b710077c2 100644
--- a/drivers/net/ethernet/intel/ice/ice_switch.h
+++ b/drivers/net/ethernet/intel/ice/ice_switch.h
@@ -275,6 +275,16 @@ struct ice_fltr_mgmt_list_entry {
 	u8 counter_index;
 };
 
+struct ice_adv_fltr_mgmt_list_entry {
+	struct list_head list_entry;
+
+	struct ice_adv_lkup_elem *lkups;
+	struct ice_adv_rule_info rule_info;
+	u16 lkups_cnt;
+	struct ice_vsi_list_map_info *vsi_list_info;
+	u16 vsi_count;
+};
+
 enum ice_promisc_flags {
 	ICE_PROMISC_UCAST_RX = 0x1,
 	ICE_PROMISC_UCAST_TX = 0x2,
@@ -310,6 +320,10 @@ enum ice_status
 		  u16 counter_id);
 
 /* Switch/bridge related commands */
+enum ice_status
+ice_add_adv_rule(struct ice_hw *hw, struct ice_adv_lkup_elem *lkups,
+		 u16 lkups_cnt, struct ice_adv_rule_info *rinfo,
+		 struct ice_rule_query_data *added_entry);
 enum ice_status ice_update_sw_rule_bridge_mode(struct ice_hw *hw);
 enum ice_status ice_add_mac(struct ice_hw *hw, struct list_head *m_lst);
 enum ice_status ice_remove_mac(struct ice_hw *hw, struct list_head *m_lst);
diff --git a/drivers/net/ethernet/intel/ice/ice_type.h b/drivers/net/ethernet/intel/ice/ice_type.h
index 6b4f064f6c0d..582225df46f4 100644
--- a/drivers/net/ethernet/intel/ice/ice_type.h
+++ b/drivers/net/ethernet/intel/ice/ice_type.h
@@ -698,6 +698,7 @@ struct ice_port_info {
 struct ice_switch_info {
 	struct list_head vsi_list_map_head;
 	struct ice_sw_recipe *recp_list;
+	u16 prof_res_bm_init;
 	u16 max_used_prof_index;
 
 	DECLARE_BITMAP(prof_res_bm[ICE_MAX_NUM_PROFILES], ICE_MAX_FV_WORDS);
-- 
1.8.3.1

[Intel-wired-lan] [PATCH net-next v6 5/9] ice: allow deleting advanced rules

[Wojciech Drewek]

From: Shivanshu Shukla <shivanshu.shukla@intel.com>

To remove advanced rule the same protocols list like in adding should be
send to function. Based on this information list of advanced rules is
searched to find the correct rule id.

Remove advanced rule if it forwards to only one VSI. If it forwards
to list of VSI remove only input VSI from this list.

Introduce function to remove rule by id. It is used in case rule needs to
be removed even if it forwards to the list of VSI.

Allow removing all advanced rules from a particular VSI. It is useful in
rebuilding VSI path.

Co-developed-by: Dan Nowlin <dan.nowlin@intel.com>
Signed-off-by: Dan Nowlin <dan.nowlin@intel.com>
Signed-off-by: Shivanshu Shukla <shivanshu.shukla@intel.com>
Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_switch.c | 223 ++++++++++++++++++++++++++++
 drivers/net/ethernet/intel/ice/ice_switch.h |   6 +
 2 files changed, 229 insertions(+)

diff --git a/drivers/net/ethernet/intel/ice/ice_switch.c b/drivers/net/ethernet/intel/ice/ice_switch.c
index 04cbd9d417fb..316c41cd87e8 100644
--- a/drivers/net/ethernet/intel/ice/ice_switch.c
+++ b/drivers/net/ethernet/intel/ice/ice_switch.c
@@ -4981,6 +4981,229 @@ enum ice_status
 }
 
 /**
+ * ice_adv_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_adv_rem_update_vsi_list(struct ice_hw *hw, u16 vsi_handle,
+			    struct ice_adv_fltr_mgmt_list_entry *fm_list)
+{
+	struct ice_vsi_list_map_info *vsi_list_info;
+	enum ice_sw_lkup_type lkup_type;
+	enum ice_status status;
+	u16 vsi_list_id;
+
+	if (fm_list->rule_info.sw_act.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 (!test_bit(vsi_handle, fm_list->vsi_list_info->vsi_map))
+		return ICE_ERR_DOES_NOT_EXIST;
+
+	lkup_type = ICE_SW_LKUP_LAST;
+	vsi_list_id = fm_list->rule_info.sw_act.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--;
+	clear_bit(vsi_handle, fm_list->vsi_list_info->vsi_map);
+	vsi_list_info = fm_list->vsi_list_info;
+	if (fm_list->vsi_count == 1) {
+		struct ice_fltr_info tmp_fltr;
+		u16 rem_vsi_handle;
+
+		rem_vsi_handle = 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;
+
+		memset(&tmp_fltr, 0, sizeof(tmp_fltr));
+		tmp_fltr.flag = fm_list->rule_info.sw_act.flag;
+		tmp_fltr.fltr_rule_id = fm_list->rule_info.fltr_rule_id;
+		fm_list->rule_info.sw_act.fltr_act = ICE_FWD_TO_VSI;
+		tmp_fltr.fltr_act = ICE_FWD_TO_VSI;
+		tmp_fltr.fwd_id.hw_vsi_id =
+			ice_get_hw_vsi_num(hw, rem_vsi_handle);
+		fm_list->rule_info.sw_act.fwd_id.hw_vsi_id =
+			ice_get_hw_vsi_num(hw, rem_vsi_handle);
+		fm_list->rule_info.sw_act.vsi_handle = rem_vsi_handle;
+
+		/* 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) {
+			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.fwd_id.hw_vsi_id, status);
+			return status;
+		}
+		fm_list->vsi_list_info->ref_cnt--;
+
+		/* 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);
+		devm_kfree(ice_hw_to_dev(hw), vsi_list_info);
+		fm_list->vsi_list_info = NULL;
+	}
+
+	return status;
+}
+
+/**
+ * ice_rem_adv_rule - removes existing advanced switch rule
+ * @hw: pointer to the hardware structure
+ * @lkups: information on the words that needs to be looked up. All words
+ *         together makes one recipe
+ * @lkups_cnt: num of entries in the lkups array
+ * @rinfo: Its the pointer to the rule information for the rule
+ *
+ * This function can be used to remove 1 rule at a time. The lkups is
+ * used to describe all the words that forms the "lookup" portion of the
+ * rule. These words can span multiple protocols. Callers to this function
+ * need to pass in a list of protocol headers with lookup information along
+ * and mask that determines which words are valid from the given protocol
+ * header. rinfo describes other information related to this rule such as
+ * forwarding IDs, priority of this rule, etc.
+ */
+static enum ice_status
+ice_rem_adv_rule(struct ice_hw *hw, struct ice_adv_lkup_elem *lkups,
+		 u16 lkups_cnt, struct ice_adv_rule_info *rinfo)
+{
+	struct ice_adv_fltr_mgmt_list_entry *list_elem;
+	struct ice_prot_lkup_ext lkup_exts;
+	enum ice_status status = 0;
+	bool remove_rule = false;
+	struct mutex *rule_lock; /* Lock to protect filter rule list */
+	u16 i, rid, vsi_handle;
+
+	memset(&lkup_exts, 0, sizeof(lkup_exts));
+	for (i = 0; i < lkups_cnt; i++) {
+		u16 count;
+
+		if (lkups[i].type >= ICE_PROTOCOL_LAST)
+			return ICE_ERR_CFG;
+
+		count = ice_fill_valid_words(&lkups[i], &lkup_exts);
+		if (!count)
+			return ICE_ERR_CFG;
+	}
+
+	rid = ice_find_recp(hw, &lkup_exts);
+	/* If did not find a recipe that match the existing criteria */
+	if (rid == ICE_MAX_NUM_RECIPES)
+		return ICE_ERR_PARAM;
+
+	rule_lock = &hw->switch_info->recp_list[rid].filt_rule_lock;
+	list_elem = ice_find_adv_rule_entry(hw, lkups, lkups_cnt, rid, rinfo);
+	/* the rule is already removed */
+	if (!list_elem)
+		return 0;
+	mutex_lock(rule_lock);
+	if (list_elem->rule_info.sw_act.fltr_act != ICE_FWD_TO_VSI_LIST) {
+		remove_rule = true;
+	} else if (list_elem->vsi_count > 1) {
+		remove_rule = false;
+		vsi_handle = rinfo->sw_act.vsi_handle;
+		status = ice_adv_rem_update_vsi_list(hw, vsi_handle, list_elem);
+	} else {
+		vsi_handle = rinfo->sw_act.vsi_handle;
+		status = ice_adv_rem_update_vsi_list(hw, vsi_handle, list_elem);
+		if (status) {
+			mutex_unlock(rule_lock);
+			return status;
+		}
+		if (list_elem->vsi_count == 0)
+			remove_rule = true;
+	}
+	mutex_unlock(rule_lock);
+	if (remove_rule) {
+		struct ice_aqc_sw_rules_elem *s_rule;
+		u16 rule_buf_sz;
+
+		rule_buf_sz = ICE_SW_RULE_RX_TX_NO_HDR_SIZE;
+		s_rule = kzalloc(rule_buf_sz, GFP_KERNEL);
+		if (!s_rule)
+			return ICE_ERR_NO_MEMORY;
+		s_rule->pdata.lkup_tx_rx.act = 0;
+		s_rule->pdata.lkup_tx_rx.index =
+			cpu_to_le16(list_elem->rule_info.fltr_rule_id);
+		s_rule->pdata.lkup_tx_rx.hdr_len = 0;
+		status = ice_aq_sw_rules(hw, (struct ice_aqc_sw_rules *)s_rule,
+					 rule_buf_sz, 1,
+					 ice_aqc_opc_remove_sw_rules, NULL);
+		if (!status || status == ICE_ERR_DOES_NOT_EXIST) {
+			struct ice_switch_info *sw = hw->switch_info;
+
+			mutex_lock(rule_lock);
+			list_del(&list_elem->list_entry);
+			devm_kfree(ice_hw_to_dev(hw), list_elem->lkups);
+			devm_kfree(ice_hw_to_dev(hw), list_elem);
+			mutex_unlock(rule_lock);
+			if (list_empty(&sw->recp_list[rid].filt_rules))
+				sw->recp_list[rid].adv_rule = false;
+		}
+		kfree(s_rule);
+	}
+	return status;
+}
+
+/**
+ * ice_rem_adv_rule_by_id - removes existing advanced switch rule by ID
+ * @hw: pointer to the hardware structure
+ * @remove_entry: data struct which holds rule_id, VSI handle and recipe ID
+ *
+ * This function is used to remove 1 rule at a time. The removal is based on
+ * the remove_entry parameter. This function will remove rule for a given
+ * vsi_handle with a given rule_id which is passed as parameter in remove_entry
+ */
+enum ice_status
+ice_rem_adv_rule_by_id(struct ice_hw *hw,
+		       struct ice_rule_query_data *remove_entry)
+{
+	struct ice_adv_fltr_mgmt_list_entry *list_itr;
+	struct list_head *list_head;
+	struct ice_adv_rule_info rinfo;
+	struct ice_switch_info *sw;
+
+	sw = hw->switch_info;
+	if (!sw->recp_list[remove_entry->rid].recp_created)
+		return ICE_ERR_PARAM;
+	list_head = &sw->recp_list[remove_entry->rid].filt_rules;
+	list_for_each_entry(list_itr, list_head, list_entry) {
+		if (list_itr->rule_info.fltr_rule_id ==
+		    remove_entry->rule_id) {
+			rinfo = list_itr->rule_info;
+			rinfo.sw_act.vsi_handle = remove_entry->vsi_handle;
+			return ice_rem_adv_rule(hw, list_itr->lkups,
+						list_itr->lkups_cnt, &rinfo);
+		}
+	}
+	/* either list is empty or unable to find rule */
+	return ICE_ERR_DOES_NOT_EXIST;
+}
+
+/**
  * ice_replay_vsi_all_fltr - replay all filters stored in bookkeeping lists
  * @hw: pointer to the hardware structure
  * @vsi_handle: driver VSI handle
diff --git a/drivers/net/ethernet/intel/ice/ice_switch.h b/drivers/net/ethernet/intel/ice/ice_switch.h
index 082b710077c2..cd5421a41002 100644
--- a/drivers/net/ethernet/intel/ice/ice_switch.h
+++ b/drivers/net/ethernet/intel/ice/ice_switch.h
@@ -351,6 +351,12 @@ 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_rem_adv_rule_for_vsi(struct ice_hw *hw, u16 vsi_handle);
+enum ice_status
+ice_rem_adv_rule_by_id(struct ice_hw *hw,
+		       struct ice_rule_query_data *remove_entry);
+
 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);
 
-- 
1.8.3.1

[Intel-wired-lan] [PATCH net-next v6 6/9] ice: cleanup rules info

[Wojciech Drewek]

From: Victor Raj <victor.raj@intel.com>

Change ICE_SW_LKUP_LAST to ICE_MAX_NUM_RECIPES as for now there also can
be recipes other than the default.

Free all structures created for advanced recipes in cleanup function.
Write a function to clean allocated structures on advanced rule info.

Signed-off-by: Victor Raj <victor.raj@intel.com>
Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_common.c | 41 +++++++++++++++++++++++------
 drivers/net/ethernet/intel/ice/ice_switch.c | 28 ++++++++++++++++++--
 2 files changed, 59 insertions(+), 10 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_common.c b/drivers/net/ethernet/intel/ice/ice_common.c
index 9d32eb970028..1a3c6b60fdca 100644
--- a/drivers/net/ethernet/intel/ice/ice_common.c
+++ b/drivers/net/ethernet/intel/ice/ice_common.c
@@ -595,17 +595,42 @@ static void ice_cleanup_fltr_mgmt_struct(struct ice_hw *hw)
 		list_del(&v_pos_map->list_entry);
 		devm_kfree(ice_hw_to_dev(hw), v_pos_map);
 	}
-	recps = hw->switch_info->recp_list;
-	for (i = 0; i < ICE_SW_LKUP_LAST; i++) {
-		struct ice_fltr_mgmt_list_entry *lst_itr, *tmp_entry;
+	recps = sw->recp_list;
+	for (i = 0; i < ICE_MAX_NUM_RECIPES; i++) {
+		struct ice_recp_grp_entry *rg_entry, *tmprg_entry;
 
 		recps[i].root_rid = i;
-		mutex_destroy(&recps[i].filt_rule_lock);
-		list_for_each_entry_safe(lst_itr, tmp_entry,
-					 &recps[i].filt_rules, list_entry) {
-			list_del(&lst_itr->list_entry);
-			devm_kfree(ice_hw_to_dev(hw), lst_itr);
+		list_for_each_entry_safe(rg_entry, tmprg_entry,
+					 &recps[i].rg_list, l_entry) {
+			list_del(&rg_entry->l_entry);
+			devm_kfree(ice_hw_to_dev(hw), rg_entry);
 		}
+
+		if (recps[i].adv_rule) {
+			struct ice_adv_fltr_mgmt_list_entry *tmp_entry;
+			struct ice_adv_fltr_mgmt_list_entry *lst_itr;
+
+			mutex_destroy(&recps[i].filt_rule_lock);
+			list_for_each_entry_safe(lst_itr, tmp_entry,
+						 &recps[i].filt_rules,
+						 list_entry) {
+				list_del(&lst_itr->list_entry);
+				devm_kfree(ice_hw_to_dev(hw), lst_itr->lkups);
+				devm_kfree(ice_hw_to_dev(hw), lst_itr);
+			}
+		} else {
+			struct ice_fltr_mgmt_list_entry *lst_itr, *tmp_entry;
+
+			mutex_destroy(&recps[i].filt_rule_lock);
+			list_for_each_entry_safe(lst_itr, tmp_entry,
+						 &recps[i].filt_rules,
+						 list_entry) {
+				list_del(&lst_itr->list_entry);
+				devm_kfree(ice_hw_to_dev(hw), lst_itr);
+			}
+		}
+		if (recps[i].root_buf)
+			devm_kfree(ice_hw_to_dev(hw), recps[i].root_buf);
 	}
 	ice_rm_all_sw_replay_rule_info(hw);
 	devm_kfree(ice_hw_to_dev(hw), sw->recp_list);
diff --git a/drivers/net/ethernet/intel/ice/ice_switch.c b/drivers/net/ethernet/intel/ice/ice_switch.c
index 316c41cd87e8..ca039a4c3a79 100644
--- a/drivers/net/ethernet/intel/ice/ice_switch.c
+++ b/drivers/net/ethernet/intel/ice/ice_switch.c
@@ -2741,6 +2741,27 @@ enum ice_status
 }
 
 /**
+ * ice_rem_adv_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_adv_rule_info(struct ice_hw *hw, struct list_head *rule_head)
+{
+	struct ice_adv_fltr_mgmt_list_entry *tmp_entry;
+	struct ice_adv_fltr_mgmt_list_entry *lst_itr;
+
+	if (list_empty(rule_head))
+		return;
+
+	list_for_each_entry_safe(lst_itr, tmp_entry, rule_head, list_entry) {
+		list_del(&lst_itr->list_entry);
+		devm_kfree(ice_hw_to_dev(hw), lst_itr->lkups);
+		devm_kfree(ice_hw_to_dev(hw), lst_itr);
+	}
+}
+
+/**
  * ice_cfg_dflt_vsi - change state of VSI to set/clear default
  * @hw: pointer to the hardware structure
  * @vsi_handle: VSI handle to set as default
@@ -5241,12 +5262,15 @@ void ice_rm_all_sw_replay_rule_info(struct ice_hw *hw)
 	if (!sw)
 		return;
 
-	for (i = 0; i < ICE_SW_LKUP_LAST; i++) {
+	for (i = 0; i < ICE_MAX_NUM_RECIPES; i++) {
 		if (!list_empty(&sw->recp_list[i].filt_replay_rules)) {
 			struct list_head *l_head;
 
 			l_head = &sw->recp_list[i].filt_replay_rules;
-			ice_rem_sw_rule_info(hw, l_head);
+			if (!sw->recp_list[i].adv_rule)
+				ice_rem_sw_rule_info(hw, l_head);
+			else
+				ice_rem_adv_rule_info(hw, l_head);
 		}
 	}
 }
-- 
1.8.3.1

[Intel-wired-lan] [PATCH net-next v6 7/9] ice: Allow changing lan_en and lb_en on all kinds of filters

[Wojciech Drewek]

From: Michal Swiatkowski <michal.swiatkowski@intel.com>

There is no way to change default lan_en and lb_en flags while
adding new rule. Add function that allows changing these flags
on rule determined by rule id and recipe id.

Function checks if the rule is presented on regular rules list or
advance rules list and call the appropriate function to update
rule entry.

As rules with ICE_SW_LKUP_DFLT recipe aren't tracked in a list,
implement function which updates flags without searching for rules
based only on rule id.

Signed-off-by: Michal Swiatkowski <michal.swiatkowski@intel.com>
Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_fltr.c | 127 ++++++++++++++++++++++++++++++
 1 file changed, 127 insertions(+)

diff --git a/drivers/net/ethernet/intel/ice/ice_fltr.c b/drivers/net/ethernet/intel/ice/ice_fltr.c
index e49378893390..f22ca4354f84 100644
--- a/drivers/net/ethernet/intel/ice/ice_fltr.c
+++ b/drivers/net/ethernet/intel/ice/ice_fltr.c
@@ -510,6 +510,133 @@ static u32 ice_fltr_build_action(u16 vsi_id)
 }
 
 /**
+ * ice_fltr_find_adv_entry - find advanced rule
+ * @rules: list of rules
+ * @rule_id: id of wanted rule
+ */
+static struct ice_adv_fltr_mgmt_list_entry *
+ice_fltr_find_adv_entry(struct list_head *rules, u16 rule_id)
+{
+	struct ice_adv_fltr_mgmt_list_entry *entry;
+
+	list_for_each_entry(entry, rules, list_entry) {
+		if (entry->rule_info.fltr_rule_id == rule_id)
+			return entry;
+	}
+
+	return NULL;
+}
+
+/**
+ * ice_fltr_update_adv_rule_flags - update flags on advanced rule
+ * @vsi: pointer to VSI
+ * @recipe_id: id of recipe
+ * @entry: advanced rule entry
+ * @new_flags: flags to update
+ */
+static enum ice_status
+ice_fltr_update_adv_rule_flags(struct ice_vsi *vsi, u16 recipe_id,
+			       struct ice_adv_fltr_mgmt_list_entry *entry,
+			       u32 new_flags)
+{
+	struct ice_adv_rule_info *info = &entry->rule_info;
+	struct ice_sw_act_ctrl *act = &info->sw_act;
+	u32 action;
+
+	if (act->fltr_act != ICE_FWD_TO_VSI)
+		return ICE_ERR_NOT_SUPPORTED;
+
+	action = ice_fltr_build_action(act->fwd_id.hw_vsi_id);
+
+	return ice_fltr_update_rule_flags(&vsi->back->hw, info->fltr_rule_id,
+					  recipe_id, action, info->sw_act.flag,
+					  act->src, new_flags);
+}
+
+/**
+ * ice_fltr_find_regular_entry - find regular rule
+ * @rules: list of rules
+ * @rule_id: id of wanted rule
+ */
+static struct ice_fltr_mgmt_list_entry *
+ice_fltr_find_regular_entry(struct list_head *rules, u16 rule_id)
+{
+	struct ice_fltr_mgmt_list_entry *entry;
+
+	list_for_each_entry(entry, rules, list_entry) {
+		if (entry->fltr_info.fltr_rule_id == rule_id)
+			return entry;
+	}
+
+	return NULL;
+}
+
+/**
+ * ice_fltr_update_regular_rule - update flags on regular rule
+ * @vsi: pointer to VSI
+ * @recipe_id: id of recipe
+ * @entry: regular rule entry
+ * @new_flags: flags to update
+ */
+static enum ice_status
+ice_fltr_update_regular_rule(struct ice_vsi *vsi, u16 recipe_id,
+			     struct ice_fltr_mgmt_list_entry *entry,
+			     u32 new_flags)
+{
+	struct ice_fltr_info *info = &entry->fltr_info;
+	u32 action;
+
+	if (info->fltr_act != ICE_FWD_TO_VSI)
+		return ICE_ERR_NOT_SUPPORTED;
+
+	action = ice_fltr_build_action(info->fwd_id.hw_vsi_id);
+
+	return ice_fltr_update_rule_flags(&vsi->back->hw, info->fltr_rule_id,
+					  recipe_id, action, info->flag,
+					  info->src, new_flags);
+}
+
+/**
+ * ice_fltr_update_flags - update flags on rule
+ * @vsi: pointer to VSI
+ * @rule_id: id of rule
+ * @recipe_id: id of recipe
+ * @new_flags: flags to update
+ *
+ * Function updates flags on regular and advance rule.
+ *
+ * Flags should be a combination of ICE_SINGLE_ACT_LB_ENABLE and
+ * ICE_SINGLE_ACT_LAN_ENABLE.
+ */
+enum ice_status
+ice_fltr_update_flags(struct ice_vsi *vsi, u16 rule_id, u16 recipe_id,
+		      u32 new_flags)
+{
+	struct ice_adv_fltr_mgmt_list_entry *adv_entry;
+	struct ice_fltr_mgmt_list_entry *regular_entry;
+	struct ice_hw *hw = &vsi->back->hw;
+	struct ice_sw_recipe *recp_list;
+	struct list_head *fltr_rules;
+
+	recp_list = &hw->switch_info->recp_list[recipe_id];
+	if (!recp_list)
+		return ICE_ERR_DOES_NOT_EXIST;
+
+	fltr_rules = &recp_list->filt_rules;
+	regular_entry = ice_fltr_find_regular_entry(fltr_rules, rule_id);
+	if (regular_entry)
+		return ice_fltr_update_regular_rule(vsi, recipe_id,
+						    regular_entry, new_flags);
+
+	adv_entry = ice_fltr_find_adv_entry(fltr_rules, rule_id);
+	if (adv_entry)
+		return ice_fltr_update_adv_rule_flags(vsi, recipe_id,
+						      adv_entry, new_flags);
+
+	return ICE_ERR_DOES_NOT_EXIST;
+}
+
+/**
  * ice_fltr_update_flags_dflt_rule - update flags on default rule
  * @vsi: pointer to VSI
  * @rule_id: id of rule
-- 
1.8.3.1

[Intel-wired-lan] [PATCH net-next v6 8/9] ice: ndo_setup_tc implementation for PF

[Wojciech Drewek]

From: Kiran Patil <kiran.patil@intel.com>

Implement ndo_setup_tc net device callback for TC HW offload on PF device.

ndo_setup_tc provides support for HW offloading various TC filters.
Add support for configuring the following filter with tc-flower:
- default L2 filters (src/dst mac addresses, ethertype, VLAN)
- variations of L3, L3+L4, L2+L3+L4 filters using advanced filters
(including ipv4 and ipv6 addresses).

Allow for adding/removing TC flows when PF device is configured in
eswitch switchdev mode. Two types of actions are supported at the
moment: FLOW_ACTION_DROP and FLOW_ACTION_REDIRECT.

Co-developed-by: Priyalee Kushwaha <priyalee.kushwaha@intel.com>
Signed-off-by: Priyalee Kushwaha <priyalee.kushwaha@intel.com>
Signed-off-by: Kiran Patil <kiran.patil@intel.com>
Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com>
---
 drivers/net/ethernet/intel/ice/Makefile     |   3 +-
 drivers/net/ethernet/intel/ice/ice.h        |   4 +
 drivers/net/ethernet/intel/ice/ice_main.c   |  68 +++
 drivers/net/ethernet/intel/ice/ice_tc_lib.c | 838 ++++++++++++++++++++++++++++
 drivers/net/ethernet/intel/ice/ice_tc_lib.h | 129 +++++
 5 files changed, 1041 insertions(+), 1 deletion(-)
 create mode 100644 drivers/net/ethernet/intel/ice/ice_tc_lib.c
 create mode 100644 drivers/net/ethernet/intel/ice/ice_tc_lib.h

diff --git a/drivers/net/ethernet/intel/ice/Makefile b/drivers/net/ethernet/intel/ice/Makefile
index 1866be50095d..c36faa7d1471 100644
--- a/drivers/net/ethernet/intel/ice/Makefile
+++ b/drivers/net/ethernet/intel/ice/Makefile
@@ -27,7 +27,8 @@ ice-y := ice_main.o	\
 	 ice_fw_update.o \
 	 ice_lag.o	\
 	 ice_ethtool.o  \
-	 ice_repr.o
+	 ice_repr.o	\
+	 ice_tc_lib.o
 ice-$(CONFIG_PCI_IOV) += ice_virtchnl_allowlist.o
 ice-$(CONFIG_PCI_IOV) += ice_virtchnl_pf.o ice_sriov.o ice_virtchnl_fdir.o
 ice-$(CONFIG_PTP_1588_CLOCK) += ice_ptp.o ice_ptp_hw.o
diff --git a/drivers/net/ethernet/intel/ice/ice.h b/drivers/net/ethernet/intel/ice/ice.h
index 5517e13bcba1..9bf70e6cd0de 100644
--- a/drivers/net/ethernet/intel/ice/ice.h
+++ b/drivers/net/ethernet/intel/ice/ice.h
@@ -64,6 +64,7 @@
 #include "ice_xsk.h"
 #include "ice_arfs.h"
 #include "ice_repr.h"
+#include "ice_eswitch.h"
 #include "ice_lag.h"
 
 #define ICE_BAR0		0
@@ -398,6 +399,7 @@ enum ice_pf_flags {
 	ICE_FLAG_PTP,			/* PTP is enabled by software */
 	ICE_FLAG_AUX_ENA,
 	ICE_FLAG_ADV_FEATURES,
+	ICE_FLAG_CLS_FLOWER,
 	ICE_FLAG_LINK_DOWN_ON_CLOSE_ENA,
 	ICE_FLAG_TOTAL_PORT_SHUTDOWN_ENA,
 	ICE_FLAG_NO_MEDIA,
@@ -511,6 +513,8 @@ struct ice_pf {
 	int aux_idx;
 	u32 sw_int_count;
 
+	struct hlist_head tc_flower_fltr_list;
+
 	__le64 nvm_phy_type_lo; /* NVM PHY type low */
 	__le64 nvm_phy_type_hi; /* NVM PHY type high */
 	struct ice_link_default_override_tlv link_dflt_override;
diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index 856249c49098..3cf8f24c77fc 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -20,6 +20,7 @@
 #define CREATE_TRACE_POINTS
 #include "ice_trace.h"
 #include "ice_eswitch.h"
+#include "ice_tc_lib.h"
 
 #define DRV_SUMMARY	"Intel(R) Ethernet Connection E800 Series Linux Driver"
 static const char ice_driver_string[] = DRV_SUMMARY;
@@ -7175,6 +7176,72 @@ static void ice_tx_timeout(struct net_device *netdev, unsigned int txqueue)
 }
 
 /**
+ * ice_setup_tc_cls_flower - flower classifier offloads
+ * @np: net device to configure
+ * @filter_dev: device on which filter is added
+ * @cls_flower: offload data
+ */
+static int
+ice_setup_tc_cls_flower(struct ice_netdev_priv *np,
+			struct net_device *filter_dev,
+			struct flow_cls_offload *cls_flower)
+{
+	struct ice_vsi *vsi = np->vsi;
+
+	if (cls_flower->common.chain_index)
+		return -EOPNOTSUPP;
+
+	switch (cls_flower->command) {
+	case FLOW_CLS_REPLACE:
+		return ice_add_cls_flower(filter_dev, vsi, cls_flower);
+	case FLOW_CLS_DESTROY:
+		return ice_del_cls_flower(vsi, cls_flower);
+	default:
+		return -EINVAL;
+	}
+}
+
+/**
+ * ice_setup_tc_block_cb - callback handler registered for TC block
+ * @type: TC SETUP type
+ * @type_data: TC flower offload data that contains user input
+ * @cb_priv: netdev private data
+ */
+static int
+ice_setup_tc_block_cb(enum tc_setup_type type, void *type_data, void *cb_priv)
+{
+	struct ice_netdev_priv *np = cb_priv;
+
+	switch (type) {
+	case TC_SETUP_CLSFLOWER:
+		return ice_setup_tc_cls_flower(np, np->vsi->netdev,
+					       type_data);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+static LIST_HEAD(ice_block_cb_list);
+
+static int
+ice_setup_tc(struct net_device *netdev, enum tc_setup_type type,
+	     void *type_data)
+{
+	struct ice_netdev_priv *np = netdev_priv(netdev);
+
+	switch (type) {
+	case TC_SETUP_BLOCK:
+		return flow_block_cb_setup_simple(type_data,
+						  &ice_block_cb_list,
+						  ice_setup_tc_block_cb,
+						  np, np, true);
+	default:
+		return -EOPNOTSUPP;
+	}
+	return -EOPNOTSUPP;
+}
+
+/**
  * ice_open - Called when a network interface becomes active
  * @netdev: network interface device structure
  *
@@ -7377,6 +7444,7 @@ int ice_stop(struct net_device *netdev)
 	.ndo_get_vf_stats = ice_get_vf_stats,
 	.ndo_vlan_rx_add_vid = ice_vlan_rx_add_vid,
 	.ndo_vlan_rx_kill_vid = ice_vlan_rx_kill_vid,
+	.ndo_setup_tc = ice_setup_tc,
 	.ndo_set_features = ice_set_features,
 	.ndo_bridge_getlink = ice_bridge_getlink,
 	.ndo_bridge_setlink = ice_bridge_setlink,
diff --git a/drivers/net/ethernet/intel/ice/ice_tc_lib.c b/drivers/net/ethernet/intel/ice/ice_tc_lib.c
new file mode 100644
index 000000000000..6a61888f80aa
--- /dev/null
+++ b/drivers/net/ethernet/intel/ice/ice_tc_lib.c
@@ -0,0 +1,838 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (C) 2019-2021, Intel Corporation. */
+
+#include "ice.h"
+#include "ice_tc_lib.h"
+#include "ice_lib.h"
+#include "ice_fltr.h"
+
+/**
+ * ice_tc_count_lkups - determine lookup count for switch filter
+ * @flags: TC-flower flags
+ * @headers: Pointer to TC flower filter header structure
+ * @fltr: Pointer to outer TC filter structure
+ *
+ * Determine lookup count based on TC flower input for switch filter.
+ */
+static int
+ice_tc_count_lkups(u32 flags, struct ice_tc_flower_lyr_2_4_hdrs *headers,
+		   struct ice_tc_flower_fltr *fltr)
+{
+	int lkups_cnt = 0;
+
+	if (flags & ICE_TC_FLWR_FIELD_ETH_TYPE_ID)
+		lkups_cnt++;
+
+	/* is MAC fields specified? */
+	if (flags & (ICE_TC_FLWR_FIELD_DST_MAC | ICE_TC_FLWR_FIELD_SRC_MAC))
+		lkups_cnt++;
+
+	/* is VLAN specified? */
+	if (flags & ICE_TC_FLWR_FIELD_VLAN)
+		lkups_cnt++;
+
+	/* are IPv[4|6] fields specified? */
+	if (flags & (ICE_TC_FLWR_FIELD_DEST_IPV4 | ICE_TC_FLWR_FIELD_SRC_IPV4))
+		lkups_cnt++;
+	else if (flags & (ICE_TC_FLWR_FIELD_DEST_IPV6 |
+			  ICE_TC_FLWR_FIELD_SRC_IPV6))
+		lkups_cnt++;
+
+	/* is L4 (TCP/UDP/any other L4 protocol fields) specified? */
+	if (flags & (ICE_TC_FLWR_FIELD_DEST_L4_PORT |
+		     ICE_TC_FLWR_FIELD_SRC_L4_PORT))
+		lkups_cnt++;
+
+	return lkups_cnt;
+}
+
+/**
+ * ice_tc_fill_rules - fill filter rules based on TC fltr
+ * @hw: pointer to HW structure
+ * @flags: tc flower field flags
+ * @tc_fltr: pointer to TC flower filter
+ * @list: list of advance rule elements
+ * @rule_info: pointer to information about rule
+ * @l4_proto: pointer to information such as L4 proto type
+ *
+ * Fill ice_adv_lkup_elem list based on TC flower flags and
+ * TC flower headers. This list should be used to add
+ * advance filter in hardware.
+ */
+static int
+ice_tc_fill_rules(struct ice_hw *hw, u32 flags,
+		  struct ice_tc_flower_fltr *tc_fltr,
+		  struct ice_adv_lkup_elem *list,
+		  struct ice_adv_rule_info *rule_info,
+		  u16 *l4_proto)
+{
+	struct ice_tc_flower_lyr_2_4_hdrs *headers = &tc_fltr->outer_headers;
+	int i = 0;
+
+	if (flags & ICE_TC_FLWR_FIELD_ETH_TYPE_ID) {
+		list[i].type = ICE_ETYPE_OL;
+		list[i].h_u.ethertype.ethtype_id = headers->l2_key.n_proto;
+		list[i].m_u.ethertype.ethtype_id = headers->l2_mask.n_proto;
+		i++;
+	}
+
+	if (flags & (ICE_TC_FLWR_FIELD_DST_MAC |
+		     ICE_TC_FLWR_FIELD_SRC_MAC)) {
+		struct ice_tc_l2_hdr *l2_key, *l2_mask;
+
+		l2_key = &headers->l2_key;
+		l2_mask = &headers->l2_mask;
+
+		list[i].type = ICE_MAC_OFOS;
+		if (flags & ICE_TC_FLWR_FIELD_DST_MAC) {
+			ether_addr_copy(list[i].h_u.eth_hdr.dst_addr,
+					l2_key->dst_mac);
+			ether_addr_copy(list[i].m_u.eth_hdr.dst_addr,
+					l2_mask->dst_mac);
+		}
+		if (flags & ICE_TC_FLWR_FIELD_SRC_MAC) {
+			ether_addr_copy(list[i].h_u.eth_hdr.src_addr,
+					l2_key->src_mac);
+			ether_addr_copy(list[i].m_u.eth_hdr.src_addr,
+					l2_mask->src_mac);
+		}
+		i++;
+	}
+
+	/* copy VLAN info */
+	if (flags & ICE_TC_FLWR_FIELD_VLAN) {
+		list[i].type = ICE_VLAN_OFOS;
+		list[i].h_u.vlan_hdr.vlan = headers->vlan_hdr.vlan_id;
+		list[i].m_u.vlan_hdr.vlan = cpu_to_be16(0xFFFF);
+		i++;
+	}
+
+	/* copy L3 (IPv[4|6]: src, dest) address */
+	if (flags & (ICE_TC_FLWR_FIELD_DEST_IPV4 |
+		     ICE_TC_FLWR_FIELD_SRC_IPV4)) {
+		struct ice_tc_l3_hdr *l3_key, *l3_mask;
+
+		list[i].type = ICE_IPV4_OFOS;
+		l3_key = &headers->l3_key;
+		l3_mask = &headers->l3_mask;
+		if (flags & ICE_TC_FLWR_FIELD_DEST_IPV4) {
+			list[i].h_u.ipv4_hdr.dst_addr = l3_key->dst_ipv4;
+			list[i].m_u.ipv4_hdr.dst_addr = l3_mask->dst_ipv4;
+		}
+		if (flags & ICE_TC_FLWR_FIELD_SRC_IPV4) {
+			list[i].h_u.ipv4_hdr.src_addr = l3_key->src_ipv4;
+			list[i].m_u.ipv4_hdr.src_addr = l3_mask->src_ipv4;
+		}
+		i++;
+	} else if (flags & (ICE_TC_FLWR_FIELD_DEST_IPV6 |
+			    ICE_TC_FLWR_FIELD_SRC_IPV6)) {
+		struct ice_ipv6_hdr *ipv6_hdr, *ipv6_mask;
+		struct ice_tc_l3_hdr *l3_key, *l3_mask;
+
+		list[i].type = ICE_IPV6_OFOS;
+		ipv6_hdr = &list[i].h_u.ipv6_hdr;
+		ipv6_mask = &list[i].m_u.ipv6_hdr;
+		l3_key = &headers->l3_key;
+		l3_mask = &headers->l3_mask;
+
+		if (flags & ICE_TC_FLWR_FIELD_DEST_IPV6) {
+			memcpy(&ipv6_hdr->dst_addr, &l3_key->dst_ipv6_addr,
+			       sizeof(l3_key->dst_ipv6_addr));
+			memcpy(&ipv6_mask->dst_addr, &l3_mask->dst_ipv6_addr,
+			       sizeof(l3_mask->dst_ipv6_addr));
+		}
+		if (flags & ICE_TC_FLWR_FIELD_SRC_IPV6) {
+			memcpy(&ipv6_hdr->src_addr, &l3_key->src_ipv6_addr,
+			       sizeof(l3_key->src_ipv6_addr));
+			memcpy(&ipv6_mask->src_addr, &l3_mask->src_ipv6_addr,
+			       sizeof(l3_mask->src_ipv6_addr));
+		}
+		i++;
+	}
+
+	/* copy L4 (src, dest) port */
+	if (flags & (ICE_TC_FLWR_FIELD_DEST_L4_PORT |
+		     ICE_TC_FLWR_FIELD_SRC_L4_PORT)) {
+		struct ice_tc_l4_hdr *l4_key, *l4_mask;
+
+		l4_key = &headers->l4_key;
+		l4_mask = &headers->l4_mask;
+		if (headers->l3_key.ip_proto == IPPROTO_TCP) {
+			list[i].type = ICE_TCP_IL;
+			/* detected L4 proto is TCP */
+			if (l4_proto)
+				*l4_proto = IPPROTO_TCP;
+		} else if (headers->l3_key.ip_proto == IPPROTO_UDP) {
+			list[i].type = ICE_UDP_ILOS;
+			/* detected L4 proto is UDP */
+			if (l4_proto)
+				*l4_proto = IPPROTO_UDP;
+		}
+		if (flags & ICE_TC_FLWR_FIELD_DEST_L4_PORT) {
+			list[i].h_u.l4_hdr.dst_port = l4_key->dst_port;
+			list[i].m_u.l4_hdr.dst_port = l4_mask->dst_port;
+		}
+		if (flags & ICE_TC_FLWR_FIELD_SRC_L4_PORT) {
+			list[i].h_u.l4_hdr.src_port = l4_key->src_port;
+			list[i].m_u.l4_hdr.src_port = l4_mask->src_port;
+		}
+		i++;
+	}
+
+	return i;
+}
+
+static int
+ice_eswitch_tc_parse_action(struct ice_tc_flower_fltr *fltr,
+			    struct flow_action_entry *act)
+{
+	struct ice_repr *repr;
+
+	switch (act->id) {
+	case FLOW_ACTION_DROP:
+		fltr->action.fltr_act = ICE_DROP_PACKET;
+		break;
+
+	case FLOW_ACTION_REDIRECT:
+		fltr->action.fltr_act = ICE_FWD_TO_VSI;
+
+		if (ice_is_port_repr_netdev(act->dev)) {
+			repr = ice_netdev_to_repr(act->dev);
+
+			fltr->dest_vsi = repr->src_vsi;
+			fltr->direction = ICE_ESWITCH_FLTR_INGRESS;
+		} else if (netif_is_ice(act->dev)) {
+			struct ice_netdev_priv *np = netdev_priv(act->dev);
+
+			fltr->dest_vsi = np->vsi;
+			fltr->direction = ICE_ESWITCH_FLTR_EGRESS;
+		} else {
+			NL_SET_ERR_MSG_MOD(fltr->extack, "Unsupported netdevice in switchdev mode");
+			return -EINVAL;
+		}
+
+		break;
+
+	default:
+		NL_SET_ERR_MSG_MOD(fltr->extack, "Unsupported action in switchdev mode");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int
+ice_eswitch_add_tc_fltr(struct ice_vsi *vsi, struct ice_tc_flower_fltr *fltr)
+{
+	struct ice_tc_flower_lyr_2_4_hdrs *headers = &fltr->outer_headers;
+	struct ice_adv_rule_info rule_info = { 0 };
+	struct ice_rule_query_data rule_added;
+	struct ice_hw *hw = &vsi->back->hw;
+	struct ice_adv_lkup_elem *list;
+	u32 flags = fltr->flags;
+	enum ice_status status;
+	int lkups_cnt;
+	int ret = 0;
+	int i;
+
+	if (!flags || (flags & (ICE_TC_FLWR_FIELD_ENC_DEST_IPV4 |
+				ICE_TC_FLWR_FIELD_ENC_SRC_IPV4 |
+				ICE_TC_FLWR_FIELD_ENC_DEST_IPV6 |
+				ICE_TC_FLWR_FIELD_ENC_SRC_IPV6 |
+				ICE_TC_FLWR_FIELD_ENC_SRC_L4_PORT))) {
+		NL_SET_ERR_MSG_MOD(fltr->extack, "Unsupported encap field(s)");
+		return -EOPNOTSUPP;
+	}
+
+	lkups_cnt = ice_tc_count_lkups(flags, headers, fltr);
+	list = kcalloc(lkups_cnt, sizeof(*list), GFP_ATOMIC);
+	if (!list)
+		return -ENOMEM;
+
+	i = ice_tc_fill_rules(hw, flags, fltr, list, &rule_info, NULL);
+	if (i != lkups_cnt) {
+		ret = -EINVAL;
+		goto exit;
+	}
+
+	rule_info.sw_act.fltr_act = fltr->action.fltr_act;
+	if (fltr->action.fltr_act != ICE_DROP_PACKET)
+		rule_info.sw_act.vsi_handle = fltr->dest_vsi->idx;
+	/* For now, making priority to be highest, and it also becomes
+	 * the priority for recipe which will get created as a result of
+	 * new extraction sequence based on input set.
+	 * Priority '7' is max val for switch recipe, higher the number
+	 * results into order of switch rule evaluation.
+	 */
+	rule_info.priority = 7;
+
+	if (fltr->direction == ICE_ESWITCH_FLTR_INGRESS) {
+		rule_info.sw_act.flag |= ICE_FLTR_RX;
+		rule_info.sw_act.src = hw->pf_id;
+		rule_info.rx = true;
+	} else {
+		rule_info.sw_act.flag |= ICE_FLTR_TX;
+		rule_info.sw_act.src = vsi->idx;
+		rule_info.rx = false;
+	}
+
+	/* specify the cookie as filter_rule_id */
+	rule_info.fltr_rule_id = fltr->cookie;
+
+	status = ice_add_adv_rule(hw, list, lkups_cnt, &rule_info, &rule_added);
+	if (status == ICE_ERR_ALREADY_EXISTS) {
+		NL_SET_ERR_MSG_MOD(fltr->extack, "Unable to add filter because it already exist");
+		ret = -EINVAL;
+		goto exit;
+	} else if (status) {
+		NL_SET_ERR_MSG_MOD(fltr->extack, "Unable to add filter due to error");
+		ret = -EIO;
+		goto exit;
+	}
+
+	/* store the output params, which are needed later for removing
+	 * advanced switch filter
+	 */
+	fltr->rid = rule_added.rid;
+	fltr->rule_id = rule_added.rule_id;
+
+	if (fltr->direction == ICE_ESWITCH_FLTR_EGRESS) {
+		if (ice_fltr_update_flags(vsi, fltr->rule_id, fltr->rid,
+					  ICE_SINGLE_ACT_LAN_ENABLE))
+			ice_rem_adv_rule_by_id(hw, &rule_added);
+	}
+
+exit:
+	kfree(list);
+	return ret;
+}
+
+/**
+ * ice_tc_set_ipv4 - Parse IPv4 addresses from TC flower filter
+ * @match: Pointer to flow match structure
+ * @fltr: Pointer to filter structure
+ * @headers: inner or outer header fields
+ */
+static int
+ice_tc_set_ipv4(struct flow_match_ipv4_addrs *match,
+		struct ice_tc_flower_fltr *fltr,
+		struct ice_tc_flower_lyr_2_4_hdrs *headers)
+{
+	if (match->key->dst) {
+		fltr->flags |= ICE_TC_FLWR_FIELD_DEST_IPV4;
+		headers->l3_key.dst_ipv4 = match->key->dst;
+		headers->l3_mask.dst_ipv4 = match->mask->dst;
+	}
+	if (match->key->src) {
+		fltr->flags |= ICE_TC_FLWR_FIELD_SRC_IPV4;
+		headers->l3_key.src_ipv4 = match->key->src;
+		headers->l3_mask.src_ipv4 = match->mask->src;
+	}
+	return 0;
+}
+
+/**
+ * ice_tc_set_ipv6 - Parse IPv6 addresses from TC flower filter
+ * @match: Pointer to flow match structure
+ * @fltr: Pointer to filter structure
+ * @headers: inner or outer header fields
+ */
+static int
+ice_tc_set_ipv6(struct flow_match_ipv6_addrs *match,
+		struct ice_tc_flower_fltr *fltr,
+		struct ice_tc_flower_lyr_2_4_hdrs *headers)
+{
+	struct ice_tc_l3_hdr *l3_key, *l3_mask;
+
+	/* src and dest IPV6 address should not be LOOPBACK
+	 * (0:0:0:0:0:0:0:1), which can be represented as ::1
+	 */
+	if (ipv6_addr_loopback(&match->key->dst) ||
+	    ipv6_addr_loopback(&match->key->src)) {
+		NL_SET_ERR_MSG_MOD(fltr->extack, "Bad IPv6, addr is LOOPBACK");
+		return -EINVAL;
+	}
+	/* if src/dest IPv6 address is *,* error */
+	if (ipv6_addr_any(&match->mask->dst) &&
+	    ipv6_addr_any(&match->mask->src)) {
+		NL_SET_ERR_MSG_MOD(fltr->extack, "Bad src/dest IPv6, addr is any");
+		return -EINVAL;
+	}
+	if (!ipv6_addr_any(&match->mask->dst))
+		fltr->flags |= ICE_TC_FLWR_FIELD_DEST_IPV6;
+	if (!ipv6_addr_any(&match->mask->src))
+		fltr->flags |= ICE_TC_FLWR_FIELD_SRC_IPV6;
+
+	l3_key = &headers->l3_key;
+	l3_mask = &headers->l3_mask;
+
+	if (fltr->flags & ICE_TC_FLWR_FIELD_SRC_IPV6) {
+		memcpy(&l3_key->src_ipv6_addr, &match->key->src.s6_addr,
+		       sizeof(match->key->src.s6_addr));
+		memcpy(&l3_mask->src_ipv6_addr, &match->mask->src.s6_addr,
+		       sizeof(match->mask->src.s6_addr));
+	}
+	if (fltr->flags & ICE_TC_FLWR_FIELD_DEST_IPV6) {
+		memcpy(&l3_key->dst_ipv6_addr, &match->key->dst.s6_addr,
+		       sizeof(match->key->dst.s6_addr));
+		memcpy(&l3_mask->dst_ipv6_addr, &match->mask->dst.s6_addr,
+		       sizeof(match->mask->dst.s6_addr));
+	}
+
+	return 0;
+}
+
+/**
+ * ice_tc_set_port - Parse ports from TC flower filter
+ * @match: Flow match structure
+ * @fltr: Pointer to filter structure
+ * @headers: inner or outer header fields
+ */
+static int
+ice_tc_set_port(struct flow_match_ports match,
+		struct ice_tc_flower_fltr *fltr,
+		struct ice_tc_flower_lyr_2_4_hdrs *headers)
+{
+	if (match.key->dst) {
+		fltr->flags |= ICE_TC_FLWR_FIELD_DEST_L4_PORT;
+		headers->l4_key.dst_port = match.key->dst;
+		headers->l4_mask.dst_port = match.mask->dst;
+	}
+	if (match.key->src) {
+		fltr->flags |= ICE_TC_FLWR_FIELD_SRC_L4_PORT;
+		headers->l4_key.src_port = match.key->src;
+		headers->l4_mask.src_port = match.mask->src;
+	}
+	return 0;
+}
+
+/**
+ * ice_parse_cls_flower - Parse TC flower filters provided by kernel
+ * @vsi: Pointer to the VSI
+ * @filter_dev: Pointer to device on which filter is being added
+ * @f: Pointer to struct flow_cls_offload
+ * @fltr: Pointer to filter structure
+ */
+static int
+ice_parse_cls_flower(struct net_device *filter_dev, struct ice_vsi *vsi,
+		     struct flow_cls_offload *f,
+		     struct ice_tc_flower_fltr *fltr)
+{
+	struct ice_tc_flower_lyr_2_4_hdrs *headers = &fltr->outer_headers;
+	struct flow_rule *rule = flow_cls_offload_flow_rule(f);
+	u16 n_proto_mask = 0, n_proto_key = 0, addr_type = 0;
+	struct flow_dissector *dissector;
+
+	dissector = rule->match.dissector;
+
+	if (dissector->used_keys &
+	    ~(BIT(FLOW_DISSECTOR_KEY_CONTROL) |
+	      BIT(FLOW_DISSECTOR_KEY_BASIC) |
+	      BIT(FLOW_DISSECTOR_KEY_ETH_ADDRS) |
+	      BIT(FLOW_DISSECTOR_KEY_VLAN) |
+	      BIT(FLOW_DISSECTOR_KEY_IPV4_ADDRS) |
+	      BIT(FLOW_DISSECTOR_KEY_IPV6_ADDRS) |
+	      BIT(FLOW_DISSECTOR_KEY_ENC_CONTROL) |
+	      BIT(FLOW_DISSECTOR_KEY_ENC_IP) |
+	      BIT(FLOW_DISSECTOR_KEY_PORTS))) {
+		NL_SET_ERR_MSG_MOD(fltr->extack, "Unsupported key used");
+		return -EOPNOTSUPP;
+	}
+
+	if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_BASIC)) {
+		struct flow_match_basic match;
+
+		flow_rule_match_basic(rule, &match);
+
+		n_proto_key = ntohs(match.key->n_proto);
+		n_proto_mask = ntohs(match.mask->n_proto);
+
+		if (n_proto_key == ETH_P_ALL || n_proto_key == 0) {
+			n_proto_key = 0;
+			n_proto_mask = 0;
+		} else {
+			fltr->flags |= ICE_TC_FLWR_FIELD_ETH_TYPE_ID;
+		}
+
+		headers->l2_key.n_proto = cpu_to_be16(n_proto_key);
+		headers->l2_mask.n_proto = cpu_to_be16(n_proto_mask);
+		headers->l3_key.ip_proto = match.key->ip_proto;
+	}
+
+	if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ETH_ADDRS)) {
+		struct flow_match_eth_addrs match;
+
+		flow_rule_match_eth_addrs(rule, &match);
+
+		if (!is_zero_ether_addr(match.key->dst)) {
+			ether_addr_copy(headers->l2_key.dst_mac,
+					match.key->dst);
+			ether_addr_copy(headers->l2_mask.dst_mac,
+					match.mask->dst);
+			fltr->flags |= ICE_TC_FLWR_FIELD_DST_MAC;
+		}
+
+		if (!is_zero_ether_addr(match.key->src)) {
+			ether_addr_copy(headers->l2_key.src_mac,
+					match.key->src);
+			ether_addr_copy(headers->l2_mask.src_mac,
+					match.mask->src);
+			fltr->flags |= ICE_TC_FLWR_FIELD_SRC_MAC;
+		}
+	}
+
+	if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_VLAN) ||
+	    is_vlan_dev(filter_dev)) {
+		struct flow_dissector_key_vlan mask;
+		struct flow_dissector_key_vlan key;
+		struct flow_match_vlan match;
+
+		if (is_vlan_dev(filter_dev)) {
+			match.key = &key;
+			match.key->vlan_id = vlan_dev_vlan_id(filter_dev);
+			match.key->vlan_priority = 0;
+			match.mask = &mask;
+			memset(match.mask, 0xff, sizeof(*match.mask));
+			match.mask->vlan_priority = 0;
+		} else {
+			flow_rule_match_vlan(rule, &match);
+		}
+
+		if (match.mask->vlan_id) {
+			if (match.mask->vlan_id == VLAN_VID_MASK) {
+				fltr->flags |= ICE_TC_FLWR_FIELD_VLAN;
+			} else {
+				NL_SET_ERR_MSG_MOD(fltr->extack, "Bad VLAN mask");
+				return -EINVAL;
+			}
+		}
+
+		headers->vlan_hdr.vlan_id =
+				cpu_to_be16(match.key->vlan_id & VLAN_VID_MASK);
+		if (match.mask->vlan_priority)
+			headers->vlan_hdr.vlan_prio = match.key->vlan_priority;
+	}
+
+	if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_CONTROL)) {
+		struct flow_match_control match;
+
+		flow_rule_match_control(rule, &match);
+
+		addr_type = match.key->addr_type;
+	}
+
+	if (addr_type == FLOW_DISSECTOR_KEY_IPV4_ADDRS) {
+		struct flow_match_ipv4_addrs match;
+
+		flow_rule_match_ipv4_addrs(rule, &match);
+		if (ice_tc_set_ipv4(&match, fltr, headers))
+			return -EINVAL;
+	}
+
+	if (addr_type == FLOW_DISSECTOR_KEY_IPV6_ADDRS) {
+		struct flow_match_ipv6_addrs match;
+
+		flow_rule_match_ipv6_addrs(rule, &match);
+		if (ice_tc_set_ipv6(&match, fltr, headers))
+			return -EINVAL;
+	}
+
+	if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_PORTS)) {
+		struct flow_match_ports match;
+
+		flow_rule_match_ports(rule, &match);
+		if (ice_tc_set_port(match, fltr, headers))
+			return -EINVAL;
+		switch (headers->l3_key.ip_proto) {
+		case IPPROTO_TCP:
+		case IPPROTO_UDP:
+			break;
+		default:
+			NL_SET_ERR_MSG_MOD(fltr->extack, "Only UDP and TCP transport are supported");
+			return -EINVAL;
+		}
+	}
+	return 0;
+}
+
+/**
+ * ice_add_switch_fltr - Add TC flower filters
+ * @vsi: Pointer to VSI
+ * @fltr: Pointer to struct ice_tc_flower_fltr
+ *
+ * Add filter in HW switch block
+ */
+static int
+ice_add_switch_fltr(struct ice_vsi *vsi, struct ice_tc_flower_fltr *fltr)
+{
+	if (ice_is_eswitch_mode_switchdev(vsi->back))
+		return ice_eswitch_add_tc_fltr(vsi, fltr);
+
+	return -EOPNOTSUPP;
+}
+
+/**
+ * ice_handle_tclass_action - Support directing to a traffic class
+ * @vsi: Pointer to VSI
+ * @cls_flower: Pointer to TC flower offload structure
+ * @fltr: Pointer to TC flower filter structure
+ *
+ * Support directing traffic to a traffic class
+ */
+static int
+ice_handle_tclass_action(struct ice_vsi *vsi,
+			 struct flow_cls_offload *cls_flower,
+			 struct ice_tc_flower_fltr *fltr)
+{
+	int tc = tc_classid_to_hwtc(vsi->netdev, cls_flower->classid);
+
+	if (tc < 0) {
+		NL_SET_ERR_MSG_MOD(fltr->extack, "Unable to add filter because specified destination is invalid");
+		return -EINVAL;
+	}
+	if (!tc) {
+		NL_SET_ERR_MSG_MOD(fltr->extack, "Unable to add filter because of invalid destination");
+		return -EINVAL;
+	}
+
+	if (!(vsi->tc_cfg.ena_tc & BIT(tc))) {
+		NL_SET_ERR_MSG_MOD(fltr->extack, "Unable to add filter because of non-existence destination");
+		return -EINVAL;
+	}
+
+	/* Redirect to a TC class or Queue Group */
+	fltr->action.fltr_act = ICE_FWD_TO_QGRP;
+	fltr->action.tc_class = tc;
+
+	return 0;
+}
+
+/**
+ * ice_parse_tc_flower_actions - Parse the actions for a TC filter
+ * @vsi: Pointer to VSI
+ * @cls_flower: Pointer to TC flower offload structure
+ * @fltr: Pointer to TC flower filter structure
+ *
+ * Parse the actions for a TC filter
+ */
+static int
+ice_parse_tc_flower_actions(struct ice_vsi *vsi,
+			    struct flow_cls_offload *cls_flower,
+			    struct ice_tc_flower_fltr *fltr)
+{
+	struct flow_rule *rule = flow_cls_offload_flow_rule(cls_flower);
+	struct flow_action *flow_action = &rule->action;
+	struct flow_action_entry *act;
+	int i;
+
+	if (cls_flower->classid)
+		return ice_handle_tclass_action(vsi, cls_flower, fltr);
+
+	if (!flow_action_has_entries(flow_action))
+		return -EINVAL;
+
+	flow_action_for_each(i, act, flow_action) {
+		if (ice_is_eswitch_mode_switchdev(vsi->back)) {
+			int err = ice_eswitch_tc_parse_action(fltr, act);
+
+			if (err)
+				return err;
+			continue;
+		}
+		/* Allow only one rule per filter */
+
+		/* Drop action */
+		if (act->id == FLOW_ACTION_DROP) {
+			fltr->action.fltr_act = ICE_DROP_PACKET;
+			return 0;
+		}
+		fltr->action.fltr_act = ICE_FWD_TO_VSI;
+	}
+	return 0;
+}
+
+/**
+ * ice_del_tc_fltr - deletes a filter from HW table
+ * @vsi: Pointer to VSI
+ * @fltr: Pointer to struct ice_tc_flower_fltr
+ *
+ * This function deletes a filter from HW table and manages book-keeping
+ */
+static int ice_del_tc_fltr(struct ice_vsi *vsi, struct ice_tc_flower_fltr *fltr)
+{
+	struct ice_rule_query_data rule_rem;
+	struct ice_pf *pf = vsi->back;
+	int err;
+
+	rule_rem.rid = fltr->rid;
+	rule_rem.rule_id = fltr->rule_id;
+	rule_rem.vsi_handle = fltr->dest_id;
+	err = ice_rem_adv_rule_by_id(&pf->hw, &rule_rem);
+	if (err) {
+		if (err == ICE_ERR_DOES_NOT_EXIST) {
+			NL_SET_ERR_MSG_MOD(fltr->extack, "filter does not exist\n");
+			return -ENOENT;
+		}
+		NL_SET_ERR_MSG_MOD(fltr->extack, "Failed to delete TC flower filter");
+		return -EIO;
+	}
+
+	return 0;
+}
+
+/**
+ * ice_add_tc_fltr - adds a TC flower filter
+ * @netdev: Pointer to netdev
+ * @vsi: Pointer to VSI
+ * @f: Pointer to flower offload structure
+ * @__fltr: Pointer to struct ice_tc_flower_fltr
+ *
+ * This function parses TC-flower input fields, parses action,
+ * and adds a filter.
+ */
+static int
+ice_add_tc_fltr(struct net_device *netdev, struct ice_vsi *vsi,
+		struct flow_cls_offload *f,
+		struct ice_tc_flower_fltr **__fltr)
+{
+	struct ice_tc_flower_fltr *fltr;
+	int err;
+
+	/* by default, set output to be INVALID */
+	*__fltr = NULL;
+
+	fltr = kzalloc(sizeof(*fltr), GFP_KERNEL);
+	if (!fltr)
+		return -ENOMEM;
+
+	fltr->cookie = f->cookie;
+	fltr->extack = f->common.extack;
+	fltr->src_vsi = vsi;
+	INIT_HLIST_NODE(&fltr->tc_flower_node);
+
+	err = ice_parse_cls_flower(netdev, vsi, f, fltr);
+	if (err < 0)
+		goto err;
+
+	err = ice_parse_tc_flower_actions(vsi, f, fltr);
+	if (err < 0)
+		goto err;
+
+	err = ice_add_switch_fltr(vsi, fltr);
+	if (err < 0)
+		goto err;
+
+	/* return the newly created filter */
+	*__fltr = fltr;
+
+	return 0;
+err:
+	kfree(fltr);
+	return err;
+}
+
+/**
+ * ice_find_tc_flower_fltr - Find the TC flower filter in the list
+ * @pf: Pointer to PF
+ * @cookie: filter specific cookie
+ */
+static struct ice_tc_flower_fltr *
+ice_find_tc_flower_fltr(struct ice_pf *pf, unsigned long cookie)
+{
+	struct ice_tc_flower_fltr *fltr;
+
+	hlist_for_each_entry(fltr, &pf->tc_flower_fltr_list, tc_flower_node)
+		if (cookie == fltr->cookie)
+			return fltr;
+
+	return NULL;
+}
+
+/**
+ * ice_add_cls_flower - add TC flower filters
+ * @netdev: Pointer to filter device
+ * @vsi: Pointer to VSI
+ * @cls_flower: Pointer to flower offload structure
+ */
+int
+ice_add_cls_flower(struct net_device *netdev, struct ice_vsi *vsi,
+		   struct flow_cls_offload *cls_flower)
+{
+	struct netlink_ext_ack *extack = cls_flower->common.extack;
+	struct net_device *vsi_netdev = vsi->netdev;
+	struct ice_tc_flower_fltr *fltr;
+	struct ice_pf *pf = vsi->back;
+	int err;
+
+	if (ice_is_reset_in_progress(pf->state))
+		return -EBUSY;
+	if (test_bit(ICE_FLAG_FW_LLDP_AGENT, pf->flags))
+		return -EINVAL;
+
+	if (ice_is_port_repr_netdev(netdev))
+		vsi_netdev = netdev;
+
+	if (!(vsi_netdev->features & NETIF_F_HW_TC) &&
+	    !test_bit(ICE_FLAG_CLS_FLOWER, pf->flags)) {
+		/* Based on TC indirect notifications from kernel, all ice
+		 * devices get an instance of rule from higher level device.
+		 * Avoid triggering explicit error in this case.
+		 */
+		if (netdev == vsi_netdev)
+			NL_SET_ERR_MSG_MOD(extack, "can't apply TC flower filters, turn ON hw-tc-offload and try again");
+		return -EINVAL;
+	}
+
+	/* avoid duplicate entries, if exists - return error */
+	fltr = ice_find_tc_flower_fltr(pf, cls_flower->cookie);
+	if (fltr) {
+		NL_SET_ERR_MSG_MOD(extack, "filter cookie already exists, ignoring");
+		return -EEXIST;
+	}
+
+	/* prep and add TC-flower filter in HW */
+	err = ice_add_tc_fltr(netdev, vsi, cls_flower, &fltr);
+	if (err)
+		return err;
+
+	/* add filter into an ordered list */
+	hlist_add_head(&fltr->tc_flower_node, &pf->tc_flower_fltr_list);
+	return 0;
+}
+
+/**
+ * ice_del_cls_flower - delete TC flower filters
+ * @vsi: Pointer to VSI
+ * @cls_flower: Pointer to struct flow_cls_offload
+ */
+int
+ice_del_cls_flower(struct ice_vsi *vsi, struct flow_cls_offload *cls_flower)
+{
+	struct ice_tc_flower_fltr *fltr;
+	struct ice_pf *pf = vsi->back;
+	int err;
+
+	/* find filter */
+	fltr = ice_find_tc_flower_fltr(pf, cls_flower->cookie);
+	if (!fltr) {
+		if (hlist_empty(&pf->tc_flower_fltr_list))
+			return 0;
+
+		NL_SET_ERR_MSG_MOD(cls_flower->common.extack, "failed to delete TC flower filter because unable to find it");
+		return -EINVAL;
+	}
+
+	fltr->extack = cls_flower->common.extack;
+	/* delete filter from HW */
+	err = ice_del_tc_fltr(vsi, fltr);
+	if (err)
+		return err;
+
+	/* delete filter from an ordered list */
+	hlist_del(&fltr->tc_flower_node);
+
+	/* free the filter node */
+	kfree(fltr);
+
+	return 0;
+}
diff --git a/drivers/net/ethernet/intel/ice/ice_tc_lib.h b/drivers/net/ethernet/intel/ice/ice_tc_lib.h
new file mode 100644
index 000000000000..883d2891f92c
--- /dev/null
+++ b/drivers/net/ethernet/intel/ice/ice_tc_lib.h
@@ -0,0 +1,129 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright (C) 2019-2021, Intel Corporation. */
+
+#ifndef _ICE_TC_LIB_H_
+#define _ICE_TC_LIB_H_
+
+#define ICE_TC_FLWR_FIELD_DST_MAC		BIT(0)
+#define ICE_TC_FLWR_FIELD_SRC_MAC		BIT(1)
+#define ICE_TC_FLWR_FIELD_VLAN			BIT(2)
+#define ICE_TC_FLWR_FIELD_DEST_IPV4		BIT(3)
+#define ICE_TC_FLWR_FIELD_SRC_IPV4		BIT(4)
+#define ICE_TC_FLWR_FIELD_DEST_IPV6		BIT(5)
+#define ICE_TC_FLWR_FIELD_SRC_IPV6		BIT(6)
+#define ICE_TC_FLWR_FIELD_DEST_L4_PORT		BIT(7)
+#define ICE_TC_FLWR_FIELD_SRC_L4_PORT		BIT(8)
+#define ICE_TC_FLWR_FIELD_TENANT_ID		BIT(9)
+#define ICE_TC_FLWR_FIELD_ENC_DEST_IPV4		BIT(10)
+#define ICE_TC_FLWR_FIELD_ENC_SRC_IPV4		BIT(11)
+#define ICE_TC_FLWR_FIELD_ENC_DEST_IPV6		BIT(12)
+#define ICE_TC_FLWR_FIELD_ENC_SRC_IPV6		BIT(13)
+#define ICE_TC_FLWR_FIELD_ENC_DEST_L4_PORT	BIT(14)
+#define ICE_TC_FLWR_FIELD_ENC_SRC_L4_PORT	BIT(15)
+#define ICE_TC_FLWR_FIELD_ENC_DST_MAC		BIT(16)
+#define ICE_TC_FLWR_FIELD_ETH_TYPE_ID		BIT(17)
+
+struct ice_tc_flower_action {
+	u32 tc_class;
+	enum ice_sw_fwd_act_type fltr_act;
+};
+
+struct ice_tc_vlan_hdr {
+	__be16 vlan_id; /* Only last 12 bits valid */
+	u16 vlan_prio; /* Only last 3 bits valid (valid values: 0..7) */
+};
+
+struct ice_tc_l2_hdr {
+	u8 dst_mac[ETH_ALEN];
+	u8 src_mac[ETH_ALEN];
+	__be16 n_proto;    /* Ethernet Protocol */
+};
+
+struct ice_tc_l3_hdr {
+	u8 ip_proto;    /* IPPROTO value */
+	union {
+		struct {
+			struct in_addr dst_ip;
+			struct in_addr src_ip;
+		} v4;
+		struct {
+			struct in6_addr dst_ip6;
+			struct in6_addr src_ip6;
+		} v6;
+	} ip;
+#define dst_ipv6	ip.v6.dst_ip6.s6_addr32
+#define dst_ipv6_addr	ip.v6.dst_ip6.s6_addr
+#define src_ipv6	ip.v6.src_ip6.s6_addr32
+#define src_ipv6_addr	ip.v6.src_ip6.s6_addr
+#define dst_ipv4	ip.v4.dst_ip.s_addr
+#define src_ipv4	ip.v4.src_ip.s_addr
+
+	u8 tos;
+	u8 ttl;
+};
+
+struct ice_tc_l4_hdr {
+	__be16 dst_port;
+	__be16 src_port;
+};
+
+struct ice_tc_flower_lyr_2_4_hdrs {
+	/* L2 layer fields with their mask */
+	struct ice_tc_l2_hdr l2_key;
+	struct ice_tc_l2_hdr l2_mask;
+	struct ice_tc_vlan_hdr vlan_hdr;
+	/* L3 (IPv4[6]) layer fields with their mask */
+	struct ice_tc_l3_hdr l3_key;
+	struct ice_tc_l3_hdr l3_mask;
+
+	/* L4 layer fields with their mask */
+	struct ice_tc_l4_hdr l4_key;
+	struct ice_tc_l4_hdr l4_mask;
+};
+
+enum ice_eswitch_fltr_direction {
+	ICE_ESWITCH_FLTR_INGRESS,
+	ICE_ESWITCH_FLTR_EGRESS,
+};
+
+struct ice_tc_flower_fltr {
+	struct hlist_node tc_flower_node;
+
+	/* cookie becomes filter_rule_id if rule is added successfully */
+	unsigned long cookie;
+
+	/* add_adv_rule returns information like recipe ID, rule_id. Store
+	 * those values since they are needed to remove advanced rule
+	 */
+	u16 rid;
+	u16 rule_id;
+	/* this could be queue/vsi_idx (sw handle)/queue_group, depending upon
+	 * destination type
+	 */
+	u16 dest_id;
+	/* if dest_id is vsi_idx, then need to store destination VSI ptr */
+	struct ice_vsi *dest_vsi;
+	/* direction of fltr for eswitch use case */
+	enum ice_eswitch_fltr_direction direction;
+
+	/* Parsed TC flower configuration params */
+	struct ice_tc_flower_lyr_2_4_hdrs outer_headers;
+	struct ice_tc_flower_lyr_2_4_hdrs inner_headers;
+	struct ice_vsi *src_vsi;
+	__be32 tenant_id;
+	u32 flags;
+	struct ice_tc_flower_action	action;
+
+	/* cache ptr which is used wherever needed to communicate netlink
+	 * messages
+	 */
+	struct netlink_ext_ack *extack;
+};
+
+int
+ice_add_cls_flower(struct net_device *netdev, struct ice_vsi *vsi,
+		   struct flow_cls_offload *cls_flower);
+int
+ice_del_cls_flower(struct ice_vsi *vsi, struct flow_cls_offload *cls_flower);
+
+#endif /* _ICE_TC_LIB_H_ */
-- 
1.8.3.1

[Intel-wired-lan] [PATCH net-next v6 9/9] ice: ndo_setup_tc implementation for PR

[Wojciech Drewek]

From: Michal Swiatkowski <michal.swiatkowski@intel.com>

Add tc-flower support for VF port representor devices.

Implement ndo_setup_tc callback for TC HW offload on VF port representors
devices. Implemented both methods: add and delete tc-flower flows.

Mark NETIF_F_HW_TC bit in net device's feature set to enable offload TC
infrastructure for port representor.

Implement TC filters replay function required to restore filters settings
while switchdev configuration is rebuilt.

Signed-off-by: Michal Swiatkowski <michal.swiatkowski@intel.com>
Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_eswitch.c |  3 ++
 drivers/net/ethernet/intel/ice/ice_repr.c    | 53 ++++++++++++++++++++++++++++
 drivers/net/ethernet/intel/ice/ice_tc_lib.c  | 17 +++++++++
 drivers/net/ethernet/intel/ice/ice_tc_lib.h  |  1 +
 4 files changed, 74 insertions(+)

diff --git a/drivers/net/ethernet/intel/ice/ice_eswitch.c b/drivers/net/ethernet/intel/ice/ice_eswitch.c
index ec2aa789e214..ef6ca61c9859 100644
--- a/drivers/net/ethernet/intel/ice/ice_eswitch.c
+++ b/drivers/net/ethernet/intel/ice/ice_eswitch.c
@@ -7,6 +7,7 @@
 #include "ice_fltr.h"
 #include "ice_repr.h"
 #include "ice_devlink.h"
+#include "ice_tc_lib.h"
 
 /**
  * ice_eswitch_setup_env - configure switchdev HW filters
@@ -644,6 +645,8 @@ int ice_eswitch_rebuild(struct ice_pf *pf)
 
 	ice_eswitch_remap_rings_to_vectors(pf);
 
+	ice_replay_tc_fltrs(pf);
+
 	status = ice_vsi_open(ctrl_vsi);
 	if (status)
 		return status;
diff --git a/drivers/net/ethernet/intel/ice/ice_repr.c b/drivers/net/ethernet/intel/ice/ice_repr.c
index 23e27bec4a16..8fd90ac661b1 100644
--- a/drivers/net/ethernet/intel/ice/ice_repr.c
+++ b/drivers/net/ethernet/intel/ice/ice_repr.c
@@ -5,6 +5,7 @@
 #include "ice_eswitch.h"
 #include "ice_devlink.h"
 #include "ice_virtchnl_pf.h"
+#include "ice_tc_lib.h"
 
 /**
  * ice_repr_get_sw_port_id - get port ID associated with representor
@@ -141,6 +142,55 @@ static int ice_repr_stop(struct net_device *netdev)
 	return &repr->vf->devlink_port;
 }
 
+static int
+ice_repr_setup_tc_cls_flower(struct ice_repr *repr,
+			     struct flow_cls_offload *flower)
+{
+	switch (flower->command) {
+	case FLOW_CLS_REPLACE:
+		return ice_add_cls_flower(repr->netdev, repr->src_vsi, flower);
+	case FLOW_CLS_DESTROY:
+		return ice_del_cls_flower(repr->src_vsi, flower);
+	default:
+		return -EINVAL;
+	}
+}
+
+static int
+ice_repr_setup_tc_block_cb(enum tc_setup_type type, void *type_data,
+			   void *cb_priv)
+{
+	struct flow_cls_offload *flower = (struct flow_cls_offload *)type_data;
+	struct ice_netdev_priv *np = (struct ice_netdev_priv *)cb_priv;
+
+	switch (type) {
+	case TC_SETUP_CLSFLOWER:
+		return ice_repr_setup_tc_cls_flower(np->repr, flower);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+static LIST_HEAD(ice_repr_block_cb_list);
+
+static int
+ice_repr_setup_tc(struct net_device *netdev, enum tc_setup_type type,
+		  void *type_data)
+{
+	struct ice_netdev_priv *np = netdev_priv(netdev);
+
+	switch (type) {
+	case TC_SETUP_BLOCK:
+		return flow_block_cb_setup_simple((struct flow_block_offload *)
+						  type_data,
+						  &ice_repr_block_cb_list,
+						  ice_repr_setup_tc_block_cb,
+						  np, np, true);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
 static const struct net_device_ops ice_repr_netdev_ops = {
 	.ndo_get_phys_port_name = ice_repr_get_phys_port_name,
 	.ndo_get_stats64 = ice_repr_get_stats64,
@@ -148,6 +198,7 @@ static int ice_repr_stop(struct net_device *netdev)
 	.ndo_stop = ice_repr_stop,
 	.ndo_start_xmit = ice_eswitch_port_start_xmit,
 	.ndo_get_devlink_port = ice_repr_get_devlink_port,
+	.ndo_setup_tc = ice_repr_setup_tc,
 };
 
 /**
@@ -170,6 +221,8 @@ bool ice_is_port_repr_netdev(struct net_device *netdev)
 	netdev->netdev_ops = &ice_repr_netdev_ops;
 	ice_set_ethtool_repr_ops(netdev);
 
+	netdev->hw_features |= NETIF_F_HW_TC;
+
 	netif_carrier_off(netdev);
 	netif_tx_stop_all_queues(netdev);
 
diff --git a/drivers/net/ethernet/intel/ice/ice_tc_lib.c b/drivers/net/ethernet/intel/ice/ice_tc_lib.c
index 6a61888f80aa..34fe725b4a53 100644
--- a/drivers/net/ethernet/intel/ice/ice_tc_lib.c
+++ b/drivers/net/ethernet/intel/ice/ice_tc_lib.c
@@ -836,3 +836,20 @@ static int ice_del_tc_fltr(struct ice_vsi *vsi, struct ice_tc_flower_fltr *fltr)
 
 	return 0;
 }
+
+/**
+ * ice_replay_tc_fltrs - replay TC filters
+ * @pf: pointer to PF struct
+ */
+void ice_replay_tc_fltrs(struct ice_pf *pf)
+{
+	struct ice_tc_flower_fltr *fltr;
+	struct hlist_node *node;
+
+	hlist_for_each_entry_safe(fltr, node,
+				  &pf->tc_flower_fltr_list,
+				  tc_flower_node) {
+		fltr->extack = NULL;
+		ice_add_switch_fltr(fltr->src_vsi, fltr);
+	}
+}
diff --git a/drivers/net/ethernet/intel/ice/ice_tc_lib.h b/drivers/net/ethernet/intel/ice/ice_tc_lib.h
index 883d2891f92c..d90e9e37ae25 100644
--- a/drivers/net/ethernet/intel/ice/ice_tc_lib.h
+++ b/drivers/net/ethernet/intel/ice/ice_tc_lib.h
@@ -125,5 +125,6 @@ struct ice_tc_flower_fltr {
 		   struct flow_cls_offload *cls_flower);
 int
 ice_del_cls_flower(struct ice_vsi *vsi, struct flow_cls_offload *cls_flower);
+void ice_replay_tc_fltrs(struct ice_pf *pf);
 
 #endif /* _ICE_TC_LIB_H_ */
-- 
1.8.3.1

[Intel-wired-lan] [PATCH net-next v6 3/9] ice: create advanced switch recipe

[Penigalapati, Sandeep]

>-----Original Message-----
>From: Intel-wired-lan <intel-wired-lan-bounces@osuosl.org> On Behalf Of
>Wojciech Drewek
>Sent: Friday, August 6, 2021 2:19 PM
To: intel-wired-lan@lists.osuosl.org
>Subject: [Intel-wired-lan] [PATCH net-next v6 3/9] ice: create advanced switch
>recipe
>
>From: Dan Nowlin <dan.nowlin@intel.com>
>
>These changes introduce code for creating advanced recipes for the
>switch in hardware.
>
>There are a couple of recipes already defined in the HW. They apply to
>matching on basic protocol headers, like MAC, VLAN, MACVLAN,
>ethertype or direction (promiscuous), etc.. If the user wants to match on
>other protocol headers (eg. ip address, src/dst port etc.) or different
>variation of already supported protocols, there is a need to create
>new, more complex recipe. That new recipe is referred as
>'advanced recipe', and the filtering rule created on top of that recipe
>is called 'advanced rule'.
>
>One recipe can have up to 5 words, but the first word is always reserved
>for match on switch id, so the driver can define up to 4 words for one
>recipe. To support recipes with more words up to 5 recipes can be
>chained, so 20 words can be programmed for look up.
>
>Input for adding recipe function is a list of protocols to support. Based
>on this list correct profile is being chosen. Correct profile means
>that it contains all protocol types from a list. Each profile have up to
>48 field vector words and each of this word have protocol id and offset.
>These two fields need to match with input data for adding recipe
>function. If the correct profile can't be found the function returns an
>error.
>
>The next step after finding the correct profile is grouping words into
>groups. One group can have up to 4 words. This is done to simplify
>sending recipes to HW (because recipe also can have up to 4 words).
>
>In case of chaining (so when look up consists of more than 4 words) last
>recipe will always have results from the previous recipes used as words.
>
>A recipe to profile map is used to store information about which profile
>is associate with this recipe. This map is an array of 64 elements (max
>number of recipes) and each element is a 256 bits bitmap (max number of
>profiles)
>
>Profile to recipe map is used to store information about which recipe is
>associate with this profile. This map is an array of 256 elements (max
>number of profiles) and each element is a 64 bits bitmap (max number of
>recipes)
>
>Signed-off-by: Dan Nowlin <dan.nowlin@intel.com>
>Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com>
>---
> drivers/net/ethernet/intel/ice/ice_flex_pipe.c     |   33 +-
> drivers/net/ethernet/intel/ice/ice_flex_pipe.h     |    6 +
> drivers/net/ethernet/intel/ice/ice_protocol_type.h |  169 +++
> drivers/net/ethernet/intel/ice/ice_switch.c        | 1112
>+++++++++++++++++++-
> drivers/net/ethernet/intel/ice/ice_switch.h        |  110 +-
> 5 files changed, 1410 insertions(+), 20 deletions(-)
>
Tested-by: Sandeep Penigalapati <sandeep.penigalapati@intel.com>

[Intel-wired-lan] [PATCH net-next v6 8/9] ice: ndo_setup_tc implementation for PF

[Penigalapati, Sandeep]

>-----Original Message-----
>From: Intel-wired-lan <intel-wired-lan-bounces@osuosl.org> On Behalf Of
>Wojciech Drewek
>Sent: Friday, August 6, 2021 2:19 PM
To: intel-wired-lan@lists.osuosl.org
>Subject: [Intel-wired-lan] [PATCH net-next v6 8/9] ice: ndo_setup_tc
>implementation for PF
>
>From: Kiran Patil <kiran.patil@intel.com>
>
>Implement ndo_setup_tc net device callback for TC HW offload on PF device.
>
>ndo_setup_tc provides support for HW offloading various TC filters.
>Add support for configuring the following filter with tc-flower:
>- default L2 filters (src/dst mac addresses, ethertype, VLAN)
>- variations of L3, L3+L4, L2+L3+L4 filters using advanced filters (including ipv4
>and ipv6 addresses).
>
>Allow for adding/removing TC flows when PF device is configured in eswitch
>switchdev mode. Two types of actions are supported at the
>moment: FLOW_ACTION_DROP and FLOW_ACTION_REDIRECT.
>
>Co-developed-by: Priyalee Kushwaha <priyalee.kushwaha@intel.com>
>Signed-off-by: Priyalee Kushwaha <priyalee.kushwaha@intel.com>
>Signed-off-by: Kiran Patil <kiran.patil@intel.com>
>Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com>
>---
> drivers/net/ethernet/intel/ice/Makefile     |   3 +-
> drivers/net/ethernet/intel/ice/ice.h        |   4 +
> drivers/net/ethernet/intel/ice/ice_main.c   |  68 +++
> drivers/net/ethernet/intel/ice/ice_tc_lib.c | 838
>++++++++++++++++++++++++++++  drivers/net/ethernet/intel/ice/ice_tc_lib.h
>| 129 +++++
> 5 files changed, 1041 insertions(+), 1 deletion(-)  create mode 100644
>drivers/net/ethernet/intel/ice/ice_tc_lib.c
> create mode 100644 drivers/net/ethernet/intel/ice/ice_tc_lib.h
>
Tested-by: Sandeep Penigalapati <sandeep.penigalapati@intel.com>

[Intel-wired-lan] [PATCH net-next v6 9/9] ice: ndo_setup_tc implementation for PR

[Penigalapati, Sandeep]

>-----Original Message-----
>From: Intel-wired-lan <intel-wired-lan-bounces@osuosl.org> On Behalf Of
>Wojciech Drewek
>Sent: Friday, August 6, 2021 2:19 PM
To: intel-wired-lan@lists.osuosl.org
>Subject: [Intel-wired-lan] [PATCH net-next v6 9/9] ice: ndo_setup_tc
>implementation for PR
>
>From: Michal Swiatkowski <michal.swiatkowski@intel.com>
>
>Add tc-flower support for VF port representor devices.
>
>Implement ndo_setup_tc callback for TC HW offload on VF port representors
>devices. Implemented both methods: add and delete tc-flower flows.
>
>Mark NETIF_F_HW_TC bit in net device's feature set to enable offload TC
>infrastructure for port representor.
>
>Implement TC filters replay function required to restore filters settings while
>switchdev configuration is rebuilt.
>
>Signed-off-by: Michal Swiatkowski <michal.swiatkowski@intel.com>
>Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com>
>---
> drivers/net/ethernet/intel/ice/ice_eswitch.c |  3 ++
> drivers/net/ethernet/intel/ice/ice_repr.c    | 53
>++++++++++++++++++++++++++++
> drivers/net/ethernet/intel/ice/ice_tc_lib.c  | 17 +++++++++
>drivers/net/ethernet/intel/ice/ice_tc_lib.h  |  1 +
> 4 files changed, 74 insertions(+)
>
Tested-by: Sandeep Penigalapati <sandeep.penigalapati@intel.com>

[Intel-wired-lan] [PATCH net-next v6 7/9] ice: Allow changing lan_en and lb_en on all kinds of filters

[Penigalapati, Sandeep]

>-----Original Message-----
>From: Intel-wired-lan <intel-wired-lan-bounces@osuosl.org> On Behalf Of
>Wojciech Drewek
>Sent: Friday, August 6, 2021 2:19 PM
To: intel-wired-lan@lists.osuosl.org
>Subject: [Intel-wired-lan] [PATCH net-next v6 7/9] ice: Allow changing lan_en
>and lb_en on all kinds of filters
>
>From: Michal Swiatkowski <michal.swiatkowski@intel.com>
>
>There is no way to change default lan_en and lb_en flags while adding new
>rule. Add function that allows changing these flags on rule determined by rule
>id and recipe id.
>
>Function checks if the rule is presented on regular rules list or advance rules
>list and call the appropriate function to update rule entry.
>
>As rules with ICE_SW_LKUP_DFLT recipe aren't tracked in a list, implement
>function which updates flags without searching for rules based only on rule id.
>
>Signed-off-by: Michal Swiatkowski <michal.swiatkowski@intel.com>
>Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com>
>---
> drivers/net/ethernet/intel/ice/ice_fltr.c | 127
>++++++++++++++++++++++++++++++
> 1 file changed, 127 insertions(+)
>
Tested-by: Sandeep Penigalapati <sandeep.penigalapati@intel.com>

[Intel-wired-lan] [PATCH net-next v6 4/9] ice: allow adding advanced rules

[Penigalapati, Sandeep]

>-----Original Message-----
>From: Intel-wired-lan <intel-wired-lan-bounces@osuosl.org> On Behalf Of
>Wojciech Drewek
>Sent: Friday, August 6, 2021 2:19 PM
To: intel-wired-lan@lists.osuosl.org
>Subject: [Intel-wired-lan] [PATCH net-next v6 4/9] ice: allow adding advanced
>rules
>
>From: Grishma Kotecha <grishma.kotecha@intel.com>
>
>Define dummy packet headers to allow adding advanced rules in HW. This
>header is used as admin queue command parameter for adding a rule.
>The firmware will extract correct fields and will use them in look ups.
>
>Define each supported packets header and offsets to words used in recipe.
>Supported headers:
>- MAC + IPv4 + UDP
>- MAC + VLAN + IPv4 + UDP
>- MAC + IPv4 + TCP
>- MAC + VLAN + IPv4 + TCP
>- MAC + IPv6 + UDP
>- MAC + VLAN + IPv6 + UDP
>- MAC + IPv6 + TCP
>- MAC + VLAN + IPv6 + TCP
>
>Add code for creating an advanced rule. Rule needs to match defined dummy
>packet, if not return error, which means that this type of rule isn't currently
>supported.
>
>The first step in adding advanced rule is searching for an advanced recipe
>matching this kind of rule. If it doesn't exist new recipe is created. Dummy
>packet has to be filled with the correct header field value from the rule
>definition. It will be used to do look up in HW.
>
>Support searching for existing advance rule entry. It is used in case of adding
>the same rule on different VSI. In this case, instead of creating new rule, the
>existing one should be updated with refreshed VSI list.
>
>Add initialization for prof_res_bm_init flag to zero so that the possible
>resource for fv in the files can be initialized.
>
>Co-developed-by: Dan Nowlin <dan.nowlin@intel.com>
>Signed-off-by: Dan Nowlin <dan.nowlin@intel.com>
>Signed-off-by: Grishma Kotecha <grishma.kotecha@intel.com>
>Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com>
>---
> drivers/net/ethernet/intel/ice/ice_common.c |   1 +
> drivers/net/ethernet/intel/ice/ice_switch.c | 832
>+++++++++++++++++++++++++++-  drivers/net/ethernet/intel/ice/ice_switch.h
>|  14 +
> drivers/net/ethernet/intel/ice/ice_type.h   |   1 +
> 4 files changed, 847 insertions(+), 1 deletion(-)
>
Tested-by: Sandeep Penigalapati <sandeep.penigalapati@intel.com>

[Intel-wired-lan] [PATCH net-next v6 5/9] ice: allow deleting advanced rules

[Penigalapati, Sandeep]

>-----Original Message-----
>From: Intel-wired-lan <intel-wired-lan-bounces@osuosl.org> On Behalf Of
>Wojciech Drewek
>Sent: Friday, August 6, 2021 2:19 PM
To: intel-wired-lan@lists.osuosl.org
>Subject: [Intel-wired-lan] [PATCH net-next v6 5/9] ice: allow deleting advanced
>rules
>
>From: Shivanshu Shukla <shivanshu.shukla@intel.com>
>
>To remove advanced rule the same protocols list like in adding should be send
>to function. Based on this information list of advanced rules is searched to
>find the correct rule id.
>
>Remove advanced rule if it forwards to only one VSI. If it forwards to list of VSI
>remove only input VSI from this list.
>
>Introduce function to remove rule by id. It is used in case rule needs to be
>removed even if it forwards to the list of VSI.
>
>Allow removing all advanced rules from a particular VSI. It is useful in
>rebuilding VSI path.
>
>Co-developed-by: Dan Nowlin <dan.nowlin@intel.com>
>Signed-off-by: Dan Nowlin <dan.nowlin@intel.com>
>Signed-off-by: Shivanshu Shukla <shivanshu.shukla@intel.com>
>Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com>
>---
> drivers/net/ethernet/intel/ice/ice_switch.c | 223
>++++++++++++++++++++++++++++
> drivers/net/ethernet/intel/ice/ice_switch.h |   6 +
> 2 files changed, 229 insertions(+)
>
Tested-by: Sandeep Penigalapati <sandeep.penigalapati@intel.com>

[Intel-wired-lan] [PATCH net-next v6 6/9] ice: cleanup rules info

[Penigalapati, Sandeep]

>-----Original Message-----
>From: Intel-wired-lan <intel-wired-lan-bounces@osuosl.org> On Behalf Of
>Wojciech Drewek
>Sent: Friday, August 6, 2021 2:19 PM
To: intel-wired-lan@lists.osuosl.org
>Subject: [Intel-wired-lan] [PATCH net-next v6 6/9] ice: cleanup rules info
>
>From: Victor Raj <victor.raj@intel.com>
>
>Change ICE_SW_LKUP_LAST to ICE_MAX_NUM_RECIPES as for now there also
>can be recipes other than the default.
>
>Free all structures created for advanced recipes in cleanup function.
>Write a function to clean allocated structures on advanced rule info.
>
>Signed-off-by: Victor Raj <victor.raj@intel.com>
>Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com>
>---
> drivers/net/ethernet/intel/ice/ice_common.c | 41
>+++++++++++++++++++++++------  drivers/net/ethernet/intel/ice/ice_switch.c |
>28 ++++++++++++++++++--
> 2 files changed, 59 insertions(+), 10 deletions(-)
>
Tested-by: Sandeep Penigalapati <sandeep.penigalapati@intel.com>

[Intel-wired-lan] [PATCH net-next v6 2/9] ice: manage profiles and field vectors

[Penigalapati, Sandeep]

>-----Original Message-----
>From: Intel-wired-lan <intel-wired-lan-bounces@osuosl.org> On Behalf Of
>Wojciech Drewek
>Sent: Friday, August 6, 2021 2:19 PM
To: intel-wired-lan@lists.osuosl.org
>Subject: [Intel-wired-lan] [PATCH net-next v6 2/9] ice: manage profiles and
>field vectors
>
>From: Dan Nowlin <dan.nowlin@intel.com>
>
>Implement functions to manage profiles and field vectors in hardware.
>
>In hardware, there are up to 256 profiles and each of these profiles can have
>48 field vector words. Each field vector word is described by protocol id and
>offset in the packet. To add a new recipe all used profiles need to be
>searched. If the profile contains all required protocol ids and offsets from the
>recipe it can be used. The driver has to add this profile to recipe association to
>tell hardware that newly added recipe is going to be associated with this
>profile.
>
>The amount of used profiles depend on the package. To avoid searching
>across not used profile, max profile id value is calculated at init flow.
>The profile is considered as unused when all field vector words in the profile
>are invalid (protocol id 0xff and offset 0x1ff).
>
>Profiles are read from the package section ICE_SID_FLD_VEC_SW. Empty field
>vector words can be used for recipe results. Store all unused field vector
>words in prof_res_bm. It is a 256 elements array (max number of
>profiles) each element is a 48 bit bitmap (max number of field vector words).
>
>For now, support only non-tunnel profiles type.
>
>Co-developed-by: Grishma Kotecha <grishma.kotecha@intel.com>
>Signed-off-by: Grishma Kotecha <grishma.kotecha@intel.com>
>Signed-off-by: Dan Nowlin <dan.nowlin@intel.com>
>Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com>
>---
> drivers/net/ethernet/intel/ice/ice_flex_pipe.c | 245
>+++++++++++++++++++++++++
> drivers/net/ethernet/intel/ice/ice_flex_pipe.h |   8 +
> drivers/net/ethernet/intel/ice/ice_flex_type.h |  13 ++
> drivers/net/ethernet/intel/ice/ice_switch.c    |   3 +-
> drivers/net/ethernet/intel/ice/ice_type.h      |   3 +
> 5 files changed, 271 insertions(+), 1 deletion(-)
>
Tested-by: Sandeep Penigalapati <sandeep.penigalapati@intel.com>

[Intel-wired-lan] [PATCH net-next v6 1/9] ice: implement low level recipes functions

[Penigalapati, Sandeep]

>-----Original Message-----
>From: Intel-wired-lan <intel-wired-lan-bounces@osuosl.org> On Behalf Of
>Wojciech Drewek
>Sent: Friday, August 6, 2021 2:19 PM
To: intel-wired-lan@lists.osuosl.org
>Subject: [Intel-wired-lan] [PATCH net-next v6 1/9] ice: implement low level
>recipes functions
>
>From: Grishma Kotecha <grishma.kotecha@intel.com>
>
>Add code to manage recipes and profiles on admin queue layer.
>
>Allow the driver to add a new recipe and update an existing one. Get a recipe
>and get a recipe to profile association is mostly used in update existing recipes
>code.
>
>Only default recipes can be updated. An update is done by reading recipes
>from HW, changing their params and calling add recipe command.
>
>Support following admin queue commands:
>- ice_aqc_opc_add_recipe (0x0290) - create a recipe with protocol header
>information and other details that determine how this recipe filter works
>- ice_aqc_opc_recipe_to_profile (0x0291) - associate a switch recipe to a
>profile
>- ice_aqc_opc_get_recipe (0x0292) - get details of an exsisting recipe
>- ice_aqc_opc_get_recipe_to_profile (0x0293) - get a recipe associated with
>profile ID
>
>Define ICE_AQC_RES_TYPE_RECIPE resource type to hold a switch recipe. It is
>needed when a new switch recipe needs to be created.
>
>Co-developed-by: Dan Nowlin <dan.nowlin@intel.com>
>Signed-off-by: Dan Nowlin <dan.nowlin@intel.com>
>Signed-off-by: Grishma Kotecha  <grishma.kotecha@intel.com>
>Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com>
>---
> drivers/net/ethernet/intel/ice/ice_adminq_cmd.h |  57 +++++++
> drivers/net/ethernet/intel/ice/ice_switch.c     | 213
>++++++++++++++++++++++++
> drivers/net/ethernet/intel/ice/ice_switch.h     |   9 +
> 3 files changed, 279 insertions(+)
>
Tested-by: Sandeep Penigalapati <sandeep.penigalapati@intel.com>