ice: Add code for DCB initialization part 3/4

This patch adds a new function ice_pf_dcb_cfg (and related helpers)
which applies the DCB configuration obtained from the firmware. As
part of this, VSIs/netdevs are updated with traffic class information.

This patch requires a bit of a refactor of existing code.

1. For a MIB change event, the associated VSI is closed and brought up
   again. The gap between closing and opening the VSI can cause a race
   condition. Fix this by grabbing the rtnl_lock prior to closing the
   VSI and then only free it after re-opening the VSI during a MIB
   change event.

2. ice_sched_query_elem is used in ice_sched.c and with this patch, in
   ice_dcb.c as well. However, ice_dcb.c is not built when CONFIG_DCB is
   unset. This results in namespace warnings (ice_sched.o: Externally
   defined symbols with no external references) when CONFIG_DCB is unset.
   To avoid this move ice_sched_query_elem from ice_sched.c to
   ice_common.c.
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

drivers/net/ethernet/intel/ice/ice.h

@@ -378,6 +378,9 @@ struct ice_pf {
 	struct ice_hw_port_stats stats_prev;
 	struct ice_hw hw;
 	u8 stat_prev_loaded;	/* has previous stats been loaded */
+#ifdef CONFIG_DCB
+	u16 dcbx_cap;
+#endif /* CONFIG_DCB */
 	u32 tx_timeout_count;
 	unsigned long tx_timeout_last_recovery;
 	u32 tx_timeout_recovery_level;
@@ -414,12 +417,6 @@ ice_irq_dynamic_ena(struct ice_hw *hw, struct ice_vsi *vsi,
 	wr32(hw, GLINT_DYN_CTL(vector), val);
 }
 
-static inline void ice_vsi_set_tc_cfg(struct ice_vsi *vsi)
-{
-	vsi->tc_cfg.ena_tc =  ICE_DFLT_TRAFFIC_CLASS;
-	vsi->tc_cfg.numtc = 1;
-}
-
 void ice_set_ethtool_ops(struct net_device *netdev);
 int ice_up(struct ice_vsi *vsi);
 int ice_down(struct ice_vsi *vsi);
@@ -428,5 +425,9 @@ int ice_get_rss(struct ice_vsi *vsi, u8 *seed, u8 *lut, u16 lut_size);
 void ice_fill_rss_lut(u8 *lut, u16 rss_table_size, u16 rss_size);
 void ice_print_link_msg(struct ice_vsi *vsi, bool isup);
 void ice_napi_del(struct ice_vsi *vsi);
+#ifdef CONFIG_DCB
+int ice_pf_ena_all_vsi(struct ice_pf *pf, bool locked);
+void ice_pf_dis_all_vsi(struct ice_pf *pf, bool locked);
+#endif /* CONFIG_DCB */
 
 #endif /* _ICE_H_ */

drivers/net/ethernet/intel/ice/ice_adminq_cmd.h

@@ -747,6 +747,32 @@ struct ice_aqc_delete_elem {
 	__le32 teid[1];
 };
 
+/* Query Port ETS (indirect 0x040E)
+ *
+ * This indirect command is used to query port TC node configuration.
+ */
+struct ice_aqc_query_port_ets {
+	__le32 port_teid;
+	__le32 reserved;
+	__le32 addr_high;
+	__le32 addr_low;
+};
+
+struct ice_aqc_port_ets_elem {
+	u8 tc_valid_bits;
+	u8 reserved[3];
+	/* 3 bits for UP per TC 0-7, 4th byte reserved */
+	__le32 up2tc;
+	u8 tc_bw_share[8];
+	__le32 port_eir_prof_id;
+	__le32 port_cir_prof_id;
+	/* 3 bits per Node priority to TC 0-7, 4th byte reserved */
+	__le32 tc_node_prio;
+#define ICE_TC_NODE_PRIO_S	0x4
+	u8 reserved1[4];
+	__le32 tc_node_teid[8]; /* Used for response, reserved in command */
+};
+
 /* Query Scheduler Resource Allocation (indirect 0x0412)
  * This indirect command retrieves the scheduler resources allocated by
  * EMP Firmware to the given PF.
@@ -1212,6 +1238,23 @@ struct ice_aqc_get_cee_dcb_cfg_resp {
 	u8 reserved[12];
 };
 
+/* Set Local LLDP MIB (indirect 0x0A08)
+ * Used to replace the local MIB of a given LLDP agent. e.g. DCBx
+ */
+struct ice_aqc_lldp_set_local_mib {
+	u8 type;
+#define SET_LOCAL_MIB_TYPE_DCBX_M		BIT(0)
+#define SET_LOCAL_MIB_TYPE_LOCAL_MIB		0
+#define SET_LOCAL_MIB_TYPE_CEE_M		BIT(1)
+#define SET_LOCAL_MIB_TYPE_CEE_WILLING		0
+#define SET_LOCAL_MIB_TYPE_CEE_NON_WILLING	SET_LOCAL_MIB_TYPE_CEE_M
+	u8 reserved0;
+	__le16 length;
+	u8 reserved1[4];
+	__le32 addr_high;
+	__le32 addr_low;
+};
+
 /* Stop/Start LLDP Agent (direct 0x0A09)
  * Used for stopping/starting specific LLDP agent. e.g. DCBx.
  * The same structure is used for the response, with the command field
@@ -1481,11 +1524,13 @@ struct ice_aq_desc {
 		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;
+		struct ice_aqc_query_port_ets port_ets;
 		struct ice_aqc_nvm nvm;
 		struct ice_aqc_pf_vf_msg virt;
 		struct ice_aqc_lldp_get_mib lldp_get_mib;
 		struct ice_aqc_lldp_set_mib_change lldp_set_event;
 		struct ice_aqc_lldp_start lldp_start;
+		struct ice_aqc_lldp_set_local_mib lldp_set_mib;
 		struct ice_aqc_lldp_stop_start_specific_agent lldp_agent_ctrl;
 		struct ice_aqc_get_set_rss_lut get_set_rss_lut;
 		struct ice_aqc_get_set_rss_key get_set_rss_key;
@@ -1573,6 +1618,7 @@ enum ice_adminq_opc {
 	ice_aqc_opc_get_sched_elems			= 0x0404,
 	ice_aqc_opc_suspend_sched_elems			= 0x0409,
 	ice_aqc_opc_resume_sched_elems			= 0x040A,
+	ice_aqc_opc_query_port_ets			= 0x040E,
 	ice_aqc_opc_delete_sched_elems			= 0x040F,
 	ice_aqc_opc_query_sched_res			= 0x0412,
 
@@ -1595,6 +1641,7 @@ enum ice_adminq_opc {
 	ice_aqc_opc_lldp_set_mib_change			= 0x0A01,
 	ice_aqc_opc_lldp_start				= 0x0A06,
 	ice_aqc_opc_get_cee_dcb_cfg			= 0x0A07,
+	ice_aqc_opc_lldp_set_local_mib			= 0x0A08,
 	ice_aqc_opc_lldp_stop_start_specific_agent	= 0x0A09,
 
 	/* RSS commands */

drivers/net/ethernet/intel/ice/ice_common.c

@@ -3106,3 +3106,28 @@ ice_stat_update32(struct ice_hw *hw, u32 reg, bool prev_stat_loaded,
 		/* to manage the potential roll-over */
 		*cur_stat = (new_data + BIT_ULL(32)) - *prev_stat;
 }
+
+/**
+ * ice_sched_query_elem - query element information from HW
+ * @hw: pointer to the HW struct
+ * @node_teid: node TEID to be queried
+ * @buf: buffer to element information
+ *
+ * This function queries HW element information
+ */
+enum ice_status
+ice_sched_query_elem(struct ice_hw *hw, u32 node_teid,
+		     struct ice_aqc_get_elem *buf)
+{
+	u16 buf_size, num_elem_ret = 0;
+	enum ice_status status;
+
+	buf_size = sizeof(*buf);
+	memset(buf, 0, buf_size);
+	buf->generic[0].node_teid = cpu_to_le32(node_teid);
+	status = ice_aq_query_sched_elems(hw, 1, buf, buf_size, &num_elem_ret,
+					  NULL);
+	if (status || num_elem_ret != 1)
+		ice_debug(hw, ICE_DBG_SCHED, "query element failed\n");
+	return status;
+}

drivers/net/ethernet/intel/ice/ice_common.h

@@ -118,4 +118,7 @@ ice_stat_update40(struct ice_hw *hw, u32 hireg, u32 loreg,
 void
 ice_stat_update32(struct ice_hw *hw, u32 reg, bool prev_stat_loaded,
 		  u64 *prev_stat, u64 *cur_stat);
+enum ice_status
+ice_sched_query_elem(struct ice_hw *hw, u32 node_teid,
+		     struct ice_aqc_get_elem *buf);
 #endif /* _ICE_COMMON_H_ */

drivers/net/ethernet/intel/ice/ice_dcb.c

@@ -98,6 +98,39 @@ enum ice_status ice_aq_start_lldp(struct ice_hw *hw, struct ice_sq_cd *cd)
 	return ice_aq_send_cmd(hw, &desc, NULL, 0, cd);
 }
 
+/**
+ * ice_aq_set_lldp_mib - Set the LLDP MIB
+ * @hw: pointer to the HW struct
+ * @mib_type: Local, Remote or both Local and Remote MIBs
+ * @buf: pointer to the caller-supplied buffer to store the MIB block
+ * @buf_size: size of the buffer (in bytes)
+ * @cd: pointer to command details structure or NULL
+ *
+ * Set the LLDP MIB. (0x0A08)
+ */
+static enum ice_status
+ice_aq_set_lldp_mib(struct ice_hw *hw, u8 mib_type, void *buf, u16 buf_size,
+		    struct ice_sq_cd *cd)
+{
+	struct ice_aqc_lldp_set_local_mib *cmd;
+	struct ice_aq_desc desc;
+
+	cmd = &desc.params.lldp_set_mib;
+
+	if (buf_size == 0 || !buf)
+		return ICE_ERR_PARAM;
+
+	ice_fill_dflt_direct_cmd_desc(&desc, ice_aqc_opc_lldp_set_local_mib);
+
+	desc.flags |= cpu_to_le16((u16)ICE_AQ_FLAG_RD);
+	desc.datalen = cpu_to_le16(buf_size);
+
+	cmd->type = mib_type;
+	cmd->length = cpu_to_le16(buf_size);
+
+	return ice_aq_send_cmd(hw, &desc, buf, buf_size, cd);
+}
+
 /**
  * ice_get_dcbx_status
  * @hw: pointer to the HW struct
@@ -902,3 +935,433 @@ enum ice_status ice_init_dcb(struct ice_hw *hw)
 
 	return ret;
 }
+
+/**
+ * ice_add_ieee_ets_common_tlv
+ * @buf: Data buffer to be populated with ice_dcb_ets_cfg data
+ * @ets_cfg: Container for ice_dcb_ets_cfg data
+ *
+ * Populate the TLV buffer with ice_dcb_ets_cfg data
+ */
+static void
+ice_add_ieee_ets_common_tlv(u8 *buf, struct ice_dcb_ets_cfg *ets_cfg)
+{
+	u8 priority0, priority1;
+	u8 offset = 0;
+	int i;
+
+	/* Priority Assignment Table (4 octets)
+	 * Octets:|    1    |    2    |    3    |    4    |
+	 *        -----------------------------------------
+	 *        |pri0|pri1|pri2|pri3|pri4|pri5|pri6|pri7|
+	 *        -----------------------------------------
+	 *   Bits:|7  4|3  0|7  4|3  0|7  4|3  0|7  4|3  0|
+	 *        -----------------------------------------
+	 */
+	for (i = 0; i < ICE_MAX_TRAFFIC_CLASS / 2; i++) {
+		priority0 = ets_cfg->prio_table[i * 2] & 0xF;
+		priority1 = ets_cfg->prio_table[i * 2 + 1] & 0xF;
+		buf[offset] = (priority0 << ICE_IEEE_ETS_PRIO_1_S) | priority1;
+		offset++;
+	}
+
+	/* TC Bandwidth Table (8 octets)
+	 * Octets:| 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 |
+	 *        ---------------------------------
+	 *        |tc0|tc1|tc2|tc3|tc4|tc5|tc6|tc7|
+	 *        ---------------------------------
+	 *
+	 * TSA Assignment Table (8 octets)
+	 * Octets:| 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 |
+	 *        ---------------------------------
+	 *        |tc0|tc1|tc2|tc3|tc4|tc5|tc6|tc7|
+	 *        ---------------------------------
+	 */
+	ice_for_each_traffic_class(i) {
+		buf[offset] = ets_cfg->tcbwtable[i];
+		buf[ICE_MAX_TRAFFIC_CLASS + offset] = ets_cfg->tsatable[i];
+		offset++;
+	}
+}
+
+/**
+ * ice_add_ieee_ets_tlv - Prepare ETS TLV in IEEE format
+ * @tlv: Fill the ETS config data in IEEE format
+ * @dcbcfg: Local store which holds the DCB Config
+ *
+ * Prepare IEEE 802.1Qaz ETS CFG TLV
+ */
+static void
+ice_add_ieee_ets_tlv(struct ice_lldp_org_tlv *tlv, struct ice_dcbx_cfg *dcbcfg)
+{
+	struct ice_dcb_ets_cfg *etscfg;
+	u8 *buf = tlv->tlvinfo;
+	u8 maxtcwilling = 0;
+	u32 ouisubtype;
+	u16 typelen;
+
+	typelen = ((ICE_TLV_TYPE_ORG << ICE_LLDP_TLV_TYPE_S) |
+		   ICE_IEEE_ETS_TLV_LEN);
+	tlv->typelen = htons(typelen);
+
+	ouisubtype = ((ICE_IEEE_8021QAZ_OUI << ICE_LLDP_TLV_OUI_S) |
+		      ICE_IEEE_SUBTYPE_ETS_CFG);
+	tlv->ouisubtype = htonl(ouisubtype);
+
+	/* First Octet post subtype
+	 * --------------------------
+	 * |will-|CBS  | Re-  | Max |
+	 * |ing  |     |served| TCs |
+	 * --------------------------
+	 * |1bit | 1bit|3 bits|3bits|
+	 */
+	etscfg = &dcbcfg->etscfg;
+	if (etscfg->willing)
+		maxtcwilling = BIT(ICE_IEEE_ETS_WILLING_S);
+	maxtcwilling |= etscfg->maxtcs & ICE_IEEE_ETS_MAXTC_M;
+	buf[0] = maxtcwilling;
+
+	/* Begin adding at Priority Assignment Table (offset 1 in buf) */
+	ice_add_ieee_ets_common_tlv(&buf[1], etscfg);
+}
+
+/**
+ * ice_add_ieee_etsrec_tlv - Prepare ETS Recommended TLV in IEEE format
+ * @tlv: Fill ETS Recommended TLV in IEEE format
+ * @dcbcfg: Local store which holds the DCB Config
+ *
+ * Prepare IEEE 802.1Qaz ETS REC TLV
+ */
+static void
+ice_add_ieee_etsrec_tlv(struct ice_lldp_org_tlv *tlv,
+			struct ice_dcbx_cfg *dcbcfg)
+{
+	struct ice_dcb_ets_cfg *etsrec;
+	u8 *buf = tlv->tlvinfo;
+	u32 ouisubtype;
+	u16 typelen;
+
+	typelen = ((ICE_TLV_TYPE_ORG << ICE_LLDP_TLV_TYPE_S) |
+		   ICE_IEEE_ETS_TLV_LEN);
+	tlv->typelen = htons(typelen);
+
+	ouisubtype = ((ICE_IEEE_8021QAZ_OUI << ICE_LLDP_TLV_OUI_S) |
+		      ICE_IEEE_SUBTYPE_ETS_REC);
+	tlv->ouisubtype = htonl(ouisubtype);
+
+	etsrec = &dcbcfg->etsrec;
+
+	/* First Octet is reserved */
+	/* Begin adding at Priority Assignment Table (offset 1 in buf) */
+	ice_add_ieee_ets_common_tlv(&buf[1], etsrec);
+}
+
+/**
+ * ice_add_ieee_pfc_tlv - Prepare PFC TLV in IEEE format
+ * @tlv: Fill PFC TLV in IEEE format
+ * @dcbcfg: Local store which holds the PFC CFG data
+ *
+ * Prepare IEEE 802.1Qaz PFC CFG TLV
+ */
+static void
+ice_add_ieee_pfc_tlv(struct ice_lldp_org_tlv *tlv, struct ice_dcbx_cfg *dcbcfg)
+{
+	u8 *buf = tlv->tlvinfo;
+	u32 ouisubtype;
+	u16 typelen;
+
+	typelen = ((ICE_TLV_TYPE_ORG << ICE_LLDP_TLV_TYPE_S) |
+		   ICE_IEEE_PFC_TLV_LEN);
+	tlv->typelen = htons(typelen);
+
+	ouisubtype = ((ICE_IEEE_8021QAZ_OUI << ICE_LLDP_TLV_OUI_S) |
+		      ICE_IEEE_SUBTYPE_PFC_CFG);
+	tlv->ouisubtype = htonl(ouisubtype);
+
+	/* ----------------------------------------
+	 * |will-|MBC  | Re-  | PFC |  PFC Enable  |
+	 * |ing  |     |served| cap |              |
+	 * -----------------------------------------
+	 * |1bit | 1bit|2 bits|4bits| 1 octet      |
+	 */
+	if (dcbcfg->pfc.willing)
+		buf[0] = BIT(ICE_IEEE_PFC_WILLING_S);
+
+	if (dcbcfg->pfc.mbc)
+		buf[0] |= BIT(ICE_IEEE_PFC_MBC_S);
+
+	buf[0] |= dcbcfg->pfc.pfccap & 0xF;
+	buf[1] = dcbcfg->pfc.pfcena;
+}
+
+/**
+ * ice_add_ieee_app_pri_tlv -  Prepare APP TLV in IEEE format
+ * @tlv: Fill APP TLV in IEEE format
+ * @dcbcfg: Local store which holds the APP CFG data
+ *
+ * Prepare IEEE 802.1Qaz APP CFG TLV
+ */
+static void
+ice_add_ieee_app_pri_tlv(struct ice_lldp_org_tlv *tlv,
+			 struct ice_dcbx_cfg *dcbcfg)
+{
+	u16 typelen, len, offset = 0;
+	u8 priority, selector, i = 0;
+	u8 *buf = tlv->tlvinfo;
+	u32 ouisubtype;
+
+	/* No APP TLVs then just return */
+	if (dcbcfg->numapps == 0)
+		return;
+	ouisubtype = ((ICE_IEEE_8021QAZ_OUI << ICE_LLDP_TLV_OUI_S) |
+		      ICE_IEEE_SUBTYPE_APP_PRI);
+	tlv->ouisubtype = htonl(ouisubtype);
+
+	/* Move offset to App Priority Table */
+	offset++;
+	/* Application Priority Table (3 octets)
+	 * Octets:|         1          |    2    |    3    |
+	 *        -----------------------------------------
+	 *        |Priority|Rsrvd| Sel |    Protocol ID    |
+	 *        -----------------------------------------
+	 *   Bits:|23    21|20 19|18 16|15                0|
+	 *        -----------------------------------------
+	 */
+	while (i < dcbcfg->numapps) {
+		priority = dcbcfg->app[i].priority & 0x7;
+		selector = dcbcfg->app[i].selector & 0x7;
+		buf[offset] = (priority << ICE_IEEE_APP_PRIO_S) | selector;
+		buf[offset + 1] = (dcbcfg->app[i].prot_id >> 0x8) & 0xFF;
+		buf[offset + 2] = dcbcfg->app[i].prot_id & 0xFF;
+		/* Move to next app */
+		offset += 3;
+		i++;
+		if (i >= ICE_DCBX_MAX_APPS)
+			break;
+	}
+	/* len includes size of ouisubtype + 1 reserved + 3*numapps */
+	len = sizeof(tlv->ouisubtype) + 1 + (i * 3);
+	typelen = ((ICE_TLV_TYPE_ORG << ICE_LLDP_TLV_TYPE_S) | (len & 0x1FF));
+	tlv->typelen = htons(typelen);
+}
+
+/**
+ * ice_add_dcb_tlv - Add all IEEE TLVs
+ * @tlv: Fill TLV data in IEEE format
+ * @dcbcfg: Local store which holds the DCB Config
+ * @tlvid: Type of IEEE TLV
+ *
+ * Add tlv information
+ */
+static void
+ice_add_dcb_tlv(struct ice_lldp_org_tlv *tlv, struct ice_dcbx_cfg *dcbcfg,
+		u16 tlvid)
+{
+	switch (tlvid) {
+	case ICE_IEEE_TLV_ID_ETS_CFG:
+		ice_add_ieee_ets_tlv(tlv, dcbcfg);
+		break;
+	case ICE_IEEE_TLV_ID_ETS_REC:
+		ice_add_ieee_etsrec_tlv(tlv, dcbcfg);
+		break;
+	case ICE_IEEE_TLV_ID_PFC_CFG:
+		ice_add_ieee_pfc_tlv(tlv, dcbcfg);
+		break;
+	case ICE_IEEE_TLV_ID_APP_PRI:
+		ice_add_ieee_app_pri_tlv(tlv, dcbcfg);
+		break;
+	default:
+		break;
+	}
+}
+
+/**
+ * ice_dcb_cfg_to_lldp - Convert DCB configuration to MIB format
+ * @lldpmib: pointer to the HW struct
+ * @miblen: length of LLDP MIB
+ * @dcbcfg: Local store which holds the DCB Config
+ *
+ * Convert the DCB configuration to MIB format
+ */
+static void
+ice_dcb_cfg_to_lldp(u8 *lldpmib, u16 *miblen, struct ice_dcbx_cfg *dcbcfg)
+{
+	u16 len, offset = 0, tlvid = ICE_TLV_ID_START;
+	struct ice_lldp_org_tlv *tlv;
+	u16 typelen;
+
+	tlv = (struct ice_lldp_org_tlv *)lldpmib;
+	while (1) {
+		ice_add_dcb_tlv(tlv, dcbcfg, tlvid++);
+		typelen = ntohs(tlv->typelen);
+		len = (typelen & ICE_LLDP_TLV_LEN_M) >> ICE_LLDP_TLV_LEN_S;
+		if (len)
+			offset += len + 2;
+		/* END TLV or beyond LLDPDU size */
+		if (tlvid >= ICE_TLV_ID_END_OF_LLDPPDU ||
+		    offset > ICE_LLDPDU_SIZE)
+			break;
+		/* Move to next TLV */
+		if (len)
+			tlv = (struct ice_lldp_org_tlv *)
+				((char *)tlv + sizeof(tlv->typelen) + len);
+	}
+	*miblen = offset;
+}
+
+/**
+ * ice_set_dcb_cfg - Set the local LLDP MIB to FW
+ * @pi: port information structure
+ *
+ * Set DCB configuration to the Firmware
+ */
+enum ice_status ice_set_dcb_cfg(struct ice_port_info *pi)
+{
+	u8 mib_type, *lldpmib = NULL;
+	struct ice_dcbx_cfg *dcbcfg;
+	enum ice_status ret;
+	struct ice_hw *hw;
+	u16 miblen;
+
+	if (!pi)
+		return ICE_ERR_PARAM;
+
+	hw = pi->hw;
+
+	/* update the HW local config */
+	dcbcfg = &pi->local_dcbx_cfg;
+	/* Allocate the LLDPDU */
+	lldpmib = devm_kzalloc(ice_hw_to_dev(hw), ICE_LLDPDU_SIZE, GFP_KERNEL);
+	if (!lldpmib)
+		return ICE_ERR_NO_MEMORY;
+
+	mib_type = SET_LOCAL_MIB_TYPE_LOCAL_MIB;
+	if (dcbcfg->app_mode == ICE_DCBX_APPS_NON_WILLING)
+		mib_type |= SET_LOCAL_MIB_TYPE_CEE_NON_WILLING;
+
+	ice_dcb_cfg_to_lldp(lldpmib, &miblen, dcbcfg);
+	ret = ice_aq_set_lldp_mib(hw, mib_type, (void *)lldpmib, miblen,
+				  NULL);
+
+	devm_kfree(ice_hw_to_dev(hw), lldpmib);
+
+	return ret;
+}
+
+/**
+ * ice_aq_query_port_ets - query port ets configuration
+ * @pi: port information structure
+ * @buf: pointer to buffer
+ * @buf_size: buffer size in bytes
+ * @cd: pointer to command details structure or NULL
+ *
+ * query current port ets configuration
+ */
+static enum ice_status
+ice_aq_query_port_ets(struct ice_port_info *pi,
+		      struct ice_aqc_port_ets_elem *buf, u16 buf_size,
+		      struct ice_sq_cd *cd)
+{
+	struct ice_aqc_query_port_ets *cmd;
+	struct ice_aq_desc desc;
+	enum ice_status status;
+
+	if (!pi)
+		return ICE_ERR_PARAM;
+	cmd = &desc.params.port_ets;
+	ice_fill_dflt_direct_cmd_desc(&desc, ice_aqc_opc_query_port_ets);
+	cmd->port_teid = pi->root->info.node_teid;
+
+	status = ice_aq_send_cmd(pi->hw, &desc, buf, buf_size, cd);
+	return status;
+}
+
+/**
+ * ice_update_port_tc_tree_cfg - update TC tree configuration
+ * @pi: port information structure
+ * @buf: pointer to buffer
+ *
+ * update the SW DB with the new TC changes
+ */
+static enum ice_status
+ice_update_port_tc_tree_cfg(struct ice_port_info *pi,
+			    struct ice_aqc_port_ets_elem *buf)
+{
+	struct ice_sched_node *node, *tc_node;
+	struct ice_aqc_get_elem elem;
+	enum ice_status status = 0;
+	u32 teid1, teid2;
+	u8 i, j;
+
+	if (!pi)
+		return ICE_ERR_PARAM;
+	/* suspend the missing TC nodes */
+	for (i = 0; i < pi->root->num_children; i++) {
+		teid1 = le32_to_cpu(pi->root->children[i]->info.node_teid);
+		ice_for_each_traffic_class(j) {
+			teid2 = le32_to_cpu(buf->tc_node_teid[j]);
+			if (teid1 == teid2)
+				break;
+		}
+		if (j < ICE_MAX_TRAFFIC_CLASS)
+			continue;
+		/* TC is missing */
+		pi->root->children[i]->in_use = false;
+	}
+	/* add the new TC nodes */
+	ice_for_each_traffic_class(j) {
+		teid2 = le32_to_cpu(buf->tc_node_teid[j]);
+		if (teid2 == ICE_INVAL_TEID)
+			continue;
+		/* Is it already present in the tree ? */
+		for (i = 0; i < pi->root->num_children; i++) {
+			tc_node = pi->root->children[i];
+			if (!tc_node)
+				continue;
+			teid1 = le32_to_cpu(tc_node->info.node_teid);
+			if (teid1 == teid2) {
+				tc_node->tc_num = j;
+				tc_node->in_use = true;
+				break;
+			}
+		}
+		if (i < pi->root->num_children)
+			continue;
+		/* new TC */
+		status = ice_sched_query_elem(pi->hw, teid2, &elem);
+		if (!status)
+			status = ice_sched_add_node(pi, 1, &elem.generic[0]);
+		if (status)
+			break;
+		/* update the TC number */
+		node = ice_sched_find_node_by_teid(pi->root, teid2);
+		if (node)
+			node->tc_num = j;
+	}
+	return status;
+}
+
+/**
+ * ice_query_port_ets - query port ets configuration
+ * @pi: port information structure
+ * @buf: pointer to buffer
+ * @buf_size: buffer size in bytes
+ * @cd: pointer to command details structure or NULL
+ *
+ * query current port ets configuration and update the
+ * SW DB with the TC changes
+ */
+enum ice_status
+ice_query_port_ets(struct ice_port_info *pi,
+		   struct ice_aqc_port_ets_elem *buf, u16 buf_size,
+		   struct ice_sq_cd *cd)
+{
+	enum ice_status status;
+
+	mutex_lock(&pi->sched_lock);
+	status = ice_aq_query_port_ets(pi, buf, buf_size, cd);
+	if (!status)
+		status = ice_update_port_tc_tree_cfg(pi, buf);
+	mutex_unlock(&pi->sched_lock);
+	return status;
+}

drivers/net/ethernet/intel/ice/ice_dcb.h

@@ -70,6 +70,18 @@
 #define ICE_IEEE_APP_PRIO_S		5
 #define ICE_IEEE_APP_PRIO_M		(0x7 << ICE_IEEE_APP_PRIO_S)
 
+/* TLV definitions for preparing MIB */
+#define ICE_IEEE_TLV_ID_ETS_CFG		3
+#define ICE_IEEE_TLV_ID_ETS_REC		4
+#define ICE_IEEE_TLV_ID_PFC_CFG		5
+#define ICE_IEEE_TLV_ID_APP_PRI		6
+#define ICE_TLV_ID_END_OF_LLDPPDU	7
+#define ICE_TLV_ID_START		ICE_IEEE_TLV_ID_ETS_CFG
+
+#define ICE_IEEE_ETS_TLV_LEN		25
+#define ICE_IEEE_PFC_TLV_LEN		6
+#define ICE_IEEE_APP_TLV_LEN		11
+
 /* IEEE 802.1AB LLDP Organization specific TLV */
 struct ice_lldp_org_tlv {
 	__be16 typelen;
@@ -108,7 +120,12 @@ struct ice_cee_app_prio {
 } __packed;
 
 u8 ice_get_dcbx_status(struct ice_hw *hw);
+enum ice_status ice_set_dcb_cfg(struct ice_port_info *pi);
 enum ice_status ice_init_dcb(struct ice_hw *hw);
+enum ice_status
+ice_query_port_ets(struct ice_port_info *pi,
+		   struct ice_aqc_port_ets_elem *buf, u16 buf_size,
+		   struct ice_sq_cd *cmd_details);
 #ifdef CONFIG_DCB
 enum ice_status ice_aq_start_lldp(struct ice_hw *hw, struct ice_sq_cd *cd);
 enum ice_status

drivers/net/ethernet/intel/ice/ice_dcb_lib.c

@@ -3,6 +3,189 @@
 
 #include "ice_dcb_lib.h"
 
+/**
+ * ice_dcb_get_ena_tc - return bitmap of enabled TCs
+ * @dcbcfg: DCB config to evaluate for enabled TCs
+ */
+u8 ice_dcb_get_ena_tc(struct ice_dcbx_cfg *dcbcfg)
+{
+	u8 i, num_tc, ena_tc = 1;
+
+	num_tc = ice_dcb_get_num_tc(dcbcfg);
+
+	for (i = 0; i < num_tc; i++)
+		ena_tc |= BIT(i);
+
+	return ena_tc;
+}
+
+/**
+ * ice_dcb_get_num_tc - Get the number of TCs from DCBX config
+ * @dcbcfg: config to retrieve number of TCs from
+ */
+u8 ice_dcb_get_num_tc(struct ice_dcbx_cfg *dcbcfg)
+{
+	bool tc_unused = false;
+	u8 num_tc = 0;
+	u8 ret = 0;
+	int i;
+
+	/* Scan the ETS Config Priority Table to find traffic classes
+	 * enabled and create a bitmask of enabled TCs
+	 */
+	for (i = 0; i < CEE_DCBX_MAX_PRIO; i++)
+		num_tc |= BIT(dcbcfg->etscfg.prio_table[i]);
+
+	/* Scan bitmask for contiguous TCs starting with TC0 */
+	for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++) {
+		if (num_tc & BIT(i)) {
+			if (!tc_unused) {
+				ret++;
+			} else {
+				pr_err("Non-contiguous TCs - Disabling DCB\n");
+				return 1;
+			}
+		} else {
+			tc_unused = true;
+		}
+	}
+
+	/* There is always at least 1 TC */
+	if (!ret)
+		ret = 1;
+
+	return ret;
+}
+
+/**
+ * ice_pf_dcb_recfg - Reconfigure all VEBs and VSIs
+ * @pf: pointer to the PF struct
+ *
+ * Assumed caller has already disabled all VSIs before
+ * calling this function. Reconfiguring DCB based on
+ * local_dcbx_cfg.
+ */
+static void ice_pf_dcb_recfg(struct ice_pf *pf)
+{
+	struct ice_dcbx_cfg *dcbcfg = &pf->hw.port_info->local_dcbx_cfg;
+	u8 tc_map = 0;
+	int v, ret;
+
+	/* Update each VSI */
+	ice_for_each_vsi(pf, v) {
+		if (!pf->vsi[v])
+			continue;
+
+		if (pf->vsi[v]->type == ICE_VSI_PF)
+			tc_map = ice_dcb_get_ena_tc(dcbcfg);
+		else
+			tc_map = ICE_DFLT_TRAFFIC_CLASS;
+
+		ret = ice_vsi_cfg_tc(pf->vsi[v], tc_map);
+		if (ret)
+			dev_err(&pf->pdev->dev,
+				"Failed to config TC for VSI index: %d\n",
+				pf->vsi[v]->idx);
+		else
+			ice_vsi_map_rings_to_vectors(pf->vsi[v]);
+	}
+}
+
+/**
+ * ice_pf_dcb_cfg - Apply new DCB configuration
+ * @pf: pointer to the PF struct
+ * @new_cfg: DCBX config to apply
+ */
+static int ice_pf_dcb_cfg(struct ice_pf *pf, struct ice_dcbx_cfg *new_cfg)
+{
+	struct ice_dcbx_cfg *old_cfg, *curr_cfg;
+	struct ice_aqc_port_ets_elem buf = { 0 };
+	int ret = 0;
+
+	curr_cfg = &pf->hw.port_info->local_dcbx_cfg;
+
+	/* Enable DCB tagging only when more than one TC */
+	if (ice_dcb_get_num_tc(new_cfg) > 1) {
+		dev_dbg(&pf->pdev->dev, "DCB tagging enabled (num TC > 1)\n");
+		set_bit(ICE_FLAG_DCB_ENA, pf->flags);
+	} else {
+		dev_dbg(&pf->pdev->dev, "DCB tagging disabled (num TC = 1)\n");
+		clear_bit(ICE_FLAG_DCB_ENA, pf->flags);
+	}
+
+	if (!memcmp(new_cfg, curr_cfg, sizeof(*new_cfg))) {
+		dev_dbg(&pf->pdev->dev, "No change in DCB config required\n");
+		return ret;
+	}
+
+	/* Store old config in case FW config fails */
+	old_cfg = devm_kzalloc(&pf->pdev->dev, sizeof(*old_cfg), GFP_KERNEL);
+	memcpy(old_cfg, curr_cfg, sizeof(*old_cfg));
+
+	/* avoid race conditions by holding the lock while disabling and
+	 * re-enabling the VSI
+	 */
+	rtnl_lock();
+	ice_pf_dis_all_vsi(pf, true);
+
+	memcpy(curr_cfg, new_cfg, sizeof(*curr_cfg));
+	memcpy(&curr_cfg->etsrec, &curr_cfg->etscfg, sizeof(curr_cfg->etsrec));
+
+	/* Only send new config to HW if we are in SW LLDP mode. Otherwise,
+	 * the new config came from the HW in the first place.
+	 */
+	if (pf->hw.port_info->is_sw_lldp) {
+		ret = ice_set_dcb_cfg(pf->hw.port_info);
+		if (ret) {
+			dev_err(&pf->pdev->dev, "Set DCB Config failed\n");
+			/* Restore previous settings to local config */
+			memcpy(curr_cfg, old_cfg, sizeof(*curr_cfg));
+			goto out;
+		}
+	}
+
+	ret = ice_query_port_ets(pf->hw.port_info, &buf, sizeof(buf), NULL);
+	if (ret) {
+		dev_err(&pf->pdev->dev, "Query Port ETS failed\n");
+		goto out;
+	}
+
+	ice_pf_dcb_recfg(pf);
+
+out:
+	ice_pf_ena_all_vsi(pf, true);
+	rtnl_unlock();
+	devm_kfree(&pf->pdev->dev, old_cfg);
+	return ret;
+}
+
+/**
+ * ice_dcb_init_cfg - set the initial DCB config in SW
+ * @pf: pf to apply config to
+ */
+static int ice_dcb_init_cfg(struct ice_pf *pf)
+{
+	struct ice_dcbx_cfg *newcfg;
+	struct ice_port_info *pi;
+	int ret = 0;
+
+	pi = pf->hw.port_info;
+	newcfg = devm_kzalloc(&pf->pdev->dev, sizeof(*newcfg), GFP_KERNEL);
+	if (!newcfg)
+		return -ENOMEM;
+
+	memcpy(newcfg, &pi->local_dcbx_cfg, sizeof(*newcfg));
+	memset(&pi->local_dcbx_cfg, 0, sizeof(*newcfg));
+
+	dev_info(&pf->pdev->dev, "Configuring initial DCB values\n");
+	if (ice_pf_dcb_cfg(pf, newcfg))
+		ret = -EINVAL;
+
+	devm_kfree(&pf->pdev->dev, newcfg);
+
+	return ret;
+}
+
 /**
  * ice_init_pf_dcb - initialize DCB for a PF
  * @pf: pf to initiialize DCB for
@@ -12,6 +195,7 @@ int ice_init_pf_dcb(struct ice_pf *pf)
 	struct device *dev = &pf->pdev->dev;
 	struct ice_port_info *port_info;
 	struct ice_hw *hw = &pf->hw;
+	int err;
 
 	port_info = hw->port_info;
 
@@ -38,5 +222,23 @@ int ice_init_pf_dcb(struct ice_pf *pf)
 		ice_aq_start_stop_dcbx(hw, true, &dcbx_status, NULL);
 	}
 
-	return ice_init_dcb(hw);
+	err = ice_init_dcb(hw);
+	if (err)
+		goto dcb_init_err;
+
+	/* DCBX in FW and LLDP enabled in FW */
+	pf->dcbx_cap = DCB_CAP_DCBX_LLD_MANAGED | DCB_CAP_DCBX_VER_IEEE;
+
+	set_bit(ICE_FLAG_DCB_CAPABLE, pf->flags);
+
+	err = ice_dcb_init_cfg(pf);
+	if (err)
+		goto dcb_init_err;
+
+	dev_info(&pf->pdev->dev, "DCBX offload supported\n");
+	return err;
+
+dcb_init_err:
+	dev_err(dev, "DCB init failed\n");
+	return err;
 }

drivers/net/ethernet/intel/ice/ice_dcb_lib.h

@@ -8,12 +8,25 @@
 #include "ice_lib.h"
 
 #ifdef CONFIG_DCB
+u8 ice_dcb_get_ena_tc(struct ice_dcbx_cfg *dcbcfg);
+u8 ice_dcb_get_num_tc(struct ice_dcbx_cfg *dcbcfg);
 int ice_init_pf_dcb(struct ice_pf *pf);
 #else
+static inline u8 ice_dcb_get_ena_tc(struct ice_dcbx_cfg __always_unused *dcbcfg)
+{
+	return ICE_DFLT_TRAFFIC_CLASS;
+}
+
+static inline u8 ice_dcb_get_num_tc(struct ice_dcbx_cfg __always_unused *dcbcfg)
+{
+	return 1;
+}
+
 static inline int ice_init_pf_dcb(struct ice_pf *pf)
 {
 	dev_dbg(&pf->pdev->dev, "DCB not supported\n");
 	return -EOPNOTSUPP;
 }
+
 #endif /* CONFIG_DCB */
 #endif /* _ICE_DCB_LIB_H_ */

drivers/net/ethernet/intel/ice/ice_lib.c

@@ -3,6 +3,7 @@
 
 #include "ice.h"
 #include "ice_lib.h"
+#include "ice_dcb_lib.h"
 
 /**
  * ice_setup_rx_ctx - Configure a receive ring context
@@ -1301,7 +1302,11 @@ static int ice_vsi_alloc_rings(struct ice_vsi *vsi)
  * through the MSI-X enabling code. On a constrained vector budget, we map Tx
  * and Rx rings to the vector as "efficiently" as possible.
  */
+#ifdef CONFIG_DCB
+void ice_vsi_map_rings_to_vectors(struct ice_vsi *vsi)
+#else
 static void ice_vsi_map_rings_to_vectors(struct ice_vsi *vsi)
+#endif /* CONFIG_DCB */
 {
 	int q_vectors = vsi->num_q_vectors;
 	int tx_rings_rem, rx_rings_rem;
@@ -2172,6 +2177,14 @@ int ice_cfg_vlan_pruning(struct ice_vsi *vsi, bool ena, bool vlan_promisc)
 	return -EIO;
 }
 
+static void ice_vsi_set_tc_cfg(struct ice_vsi *vsi)
+{
+	struct ice_dcbx_cfg *cfg = &vsi->port_info->local_dcbx_cfg;
+
+	vsi->tc_cfg.ena_tc = ice_dcb_get_ena_tc(cfg);
+	vsi->tc_cfg.numtc = ice_dcb_get_num_tc(cfg);
+}
+
 /**
  * ice_vsi_setup - Set up a VSI by a given type
  * @pf: board private structure
@@ -2815,3 +2828,125 @@ bool ice_is_reset_in_progress(unsigned long *state)
 	       test_bit(__ICE_CORER_REQ, state) ||
 	       test_bit(__ICE_GLOBR_REQ, state);
 }
+
+#ifdef CONFIG_DCB
+/**
+ * ice_vsi_update_q_map - update our copy of the VSI info with new queue map
+ * @vsi: VSI being configured
+ * @ctx: the context buffer returned from AQ VSI update command
+ */
+static void ice_vsi_update_q_map(struct ice_vsi *vsi, struct ice_vsi_ctx *ctx)
+{
+	vsi->info.mapping_flags = ctx->info.mapping_flags;
+	memcpy(&vsi->info.q_mapping, &ctx->info.q_mapping,
+	       sizeof(vsi->info.q_mapping));
+	memcpy(&vsi->info.tc_mapping, ctx->info.tc_mapping,
+	       sizeof(vsi->info.tc_mapping));
+}
+
+/**
+ * ice_vsi_cfg_netdev_tc - Setup the netdev TC configuration
+ * @vsi: the VSI being configured
+ * @ena_tc: TC map to be enabled
+ */
+static void ice_vsi_cfg_netdev_tc(struct ice_vsi *vsi, u8 ena_tc)
+{
+	struct net_device *netdev = vsi->netdev;
+	struct ice_pf *pf = vsi->back;
+	struct ice_dcbx_cfg *dcbcfg;
+	u8 netdev_tc;
+	int i;
+
+	if (!netdev)
+		return;
+
+	if (!ena_tc) {
+		netdev_reset_tc(netdev);
+		return;
+	}
+
+	if (netdev_set_num_tc(netdev, vsi->tc_cfg.numtc))
+		return;
+
+	dcbcfg = &pf->hw.port_info->local_dcbx_cfg;
+
+	ice_for_each_traffic_class(i)
+		if (vsi->tc_cfg.ena_tc & BIT(i))
+			netdev_set_tc_queue(netdev,
+					    vsi->tc_cfg.tc_info[i].netdev_tc,
+					    vsi->tc_cfg.tc_info[i].qcount_tx,
+					    vsi->tc_cfg.tc_info[i].qoffset);
+
+	for (i = 0; i < ICE_MAX_USER_PRIORITY; i++) {
+		u8 ets_tc = dcbcfg->etscfg.prio_table[i];
+
+		/* Get the mapped netdev TC# for the UP */
+		netdev_tc = vsi->tc_cfg.tc_info[ets_tc].netdev_tc;
+		netdev_set_prio_tc_map(netdev, i, netdev_tc);
+	}
+}
+
+/**
+ * ice_vsi_cfg_tc - Configure VSI Tx Sched for given TC map
+ * @vsi: VSI to be configured
+ * @ena_tc: TC bitmap
+ *
+ * VSI queues expected to be quiesced before calling this function
+ */
+int ice_vsi_cfg_tc(struct ice_vsi *vsi, u8 ena_tc)
+{
+	u16 max_txqs[ICE_MAX_TRAFFIC_CLASS] = { 0 };
+	struct ice_vsi_ctx *ctx;
+	struct ice_pf *pf = vsi->back;
+	enum ice_status status;
+	int i, ret = 0;
+	u8 num_tc = 0;
+
+	ice_for_each_traffic_class(i) {
+		/* build bitmap of enabled TCs */
+		if (ena_tc & BIT(i))
+			num_tc++;
+		/* populate max_txqs per TC */
+		max_txqs[i] = pf->num_lan_tx;
+	}
+
+	vsi->tc_cfg.ena_tc = ena_tc;
+	vsi->tc_cfg.numtc = num_tc;
+
+	ctx = devm_kzalloc(&pf->pdev->dev, sizeof(*ctx), GFP_KERNEL);
+	if (!ctx)
+		return -ENOMEM;
+
+	ctx->vf_num = 0;
+	ctx->info = vsi->info;
+
+	ice_vsi_setup_q_map(vsi, ctx);
+
+	/* must to indicate which section of VSI context are being modified */
+	ctx->info.valid_sections = cpu_to_le16(ICE_AQ_VSI_PROP_RXQ_MAP_VALID);
+	status = ice_update_vsi(&pf->hw, vsi->idx, ctx, NULL);
+	if (status) {
+		dev_info(&pf->pdev->dev, "Failed VSI Update\n");
+		ret = -EIO;
+		goto out;
+	}
+
+	status = ice_cfg_vsi_lan(vsi->port_info, vsi->idx, vsi->tc_cfg.ena_tc,
+				 max_txqs);
+
+	if (status) {
+		dev_err(&pf->pdev->dev,
+			"VSI %d failed TC config, error %d\n",
+			vsi->vsi_num, status);
+		ret = -EIO;
+		goto out;
+	}
+	ice_vsi_update_q_map(vsi, ctx);
+	vsi->info.valid_sections = 0;
+
+	ice_vsi_cfg_netdev_tc(vsi, ena_tc);
+out:
+	devm_kfree(&pf->pdev->dev, ctx);
+	return ret;
+}
+#endif /* CONFIG_DCB */

drivers/net/ethernet/intel/ice/ice_lib.h

@@ -41,6 +41,10 @@ void ice_vsi_delete(struct ice_vsi *vsi);
 
 int ice_vsi_clear(struct ice_vsi *vsi);
 
+#ifdef CONFIG_DCB
+int ice_vsi_cfg_tc(struct ice_vsi *vsi, u8 ena_tc);
+#endif /* CONFIG_DCB */
+
 struct ice_vsi *
 ice_vsi_setup(struct ice_pf *pf, struct ice_port_info *pi,
 	      enum ice_vsi_type type, u16 vf_id);
@@ -62,6 +66,10 @@ void ice_vsi_free_q_vectors(struct ice_vsi *vsi);
 
 void ice_vsi_put_qs(struct ice_vsi *vsi);
 
+#ifdef CONFIG_DCB
+void ice_vsi_map_rings_to_vectors(struct ice_vsi *vsi);
+#endif /* CONFIG_DCB */
+
 void ice_vsi_dis_irq(struct ice_vsi *vsi);
 
 void ice_vsi_free_irq(struct ice_vsi *vsi);

drivers/net/ethernet/intel/ice/ice_main.c

@@ -31,7 +31,6 @@ MODULE_PARM_DESC(debug, "netif level (0=none,...,16=all)");
 static struct workqueue_struct *ice_wq;
 static const struct net_device_ops ice_netdev_ops;
 
-static void ice_pf_dis_all_vsi(struct ice_pf *pf);
 static void ice_rebuild(struct ice_pf *pf);
 
 static void ice_vsi_release_all(struct ice_pf *pf);
@@ -397,6 +396,51 @@ static void ice_sync_fltr_subtask(struct ice_pf *pf)
 		}
 }
 
+/**
+ * ice_dis_vsi - pause a VSI
+ * @vsi: the VSI being paused
+ * @locked: is the rtnl_lock already held
+ */
+static void ice_dis_vsi(struct ice_vsi *vsi, bool locked)
+{
+	if (test_bit(__ICE_DOWN, vsi->state))
+		return;
+
+	set_bit(__ICE_NEEDS_RESTART, vsi->state);
+
+	if (vsi->type == ICE_VSI_PF && vsi->netdev) {
+		if (netif_running(vsi->netdev)) {
+			if (!locked) {
+				rtnl_lock();
+				vsi->netdev->netdev_ops->ndo_stop(vsi->netdev);
+				rtnl_unlock();
+			} else {
+				vsi->netdev->netdev_ops->ndo_stop(vsi->netdev);
+			}
+		} else {
+			ice_vsi_close(vsi);
+		}
+	}
+}
+
+/**
+ * ice_pf_dis_all_vsi - Pause all VSIs on a PF
+ * @pf: the PF
+ * @locked: is the rtnl_lock already held
+ */
+#ifdef CONFIG_DCB
+void ice_pf_dis_all_vsi(struct ice_pf *pf, bool locked)
+#else
+static void ice_pf_dis_all_vsi(struct ice_pf *pf, bool locked)
+#endif /* CONFIG_DCB */
+{
+	int v;
+
+	ice_for_each_vsi(pf, v)
+		if (pf->vsi[v])
+			ice_dis_vsi(pf->vsi[v], locked);
+}
+
 /**
  * ice_prepare_for_reset - prep for the core to reset
  * @pf: board private structure
@@ -417,7 +461,7 @@ ice_prepare_for_reset(struct ice_pf *pf)
 		ice_vc_notify_reset(pf);
 
 	/* disable the VSIs and their queues that are not already DOWN */
-	ice_pf_dis_all_vsi(pf);
+	ice_pf_dis_all_vsi(pf, false);
 
 	if (hw->port_info)
 		ice_sched_clear_port(hw->port_info);
@@ -3581,46 +3625,30 @@ static void ice_vsi_release_all(struct ice_pf *pf)
 }
 
 /**
- * ice_dis_vsi - pause a VSI
- * @vsi: the VSI being paused
+ * ice_ena_vsi - resume a VSI
+ * @vsi: the VSI being resume
  * @locked: is the rtnl_lock already held
  */
-static void ice_dis_vsi(struct ice_vsi *vsi, bool locked)
+static int ice_ena_vsi(struct ice_vsi *vsi, bool locked)
 {
-	if (test_bit(__ICE_DOWN, vsi->state))
-		return;
+	int err = 0;
 
-	set_bit(__ICE_NEEDS_RESTART, vsi->state);
+	if (!test_bit(__ICE_NEEDS_RESTART, vsi->state))
+		return err;
+
+	clear_bit(__ICE_NEEDS_RESTART, vsi->state);
+
+	if (vsi->netdev && vsi->type == ICE_VSI_PF) {
+		struct net_device *netd = vsi->netdev;
 
-	if (vsi->type == ICE_VSI_PF && vsi->netdev) {
 		if (netif_running(vsi->netdev)) {
-			if (!locked) {
+			if (locked) {
+				err = netd->netdev_ops->ndo_open(netd);
+			} else {
 				rtnl_lock();
-				vsi->netdev->netdev_ops->ndo_stop(vsi->netdev);
+				err = netd->netdev_ops->ndo_open(netd);
 				rtnl_unlock();
-			} else {
-				vsi->netdev->netdev_ops->ndo_stop(vsi->netdev);
 			}
-		} else {
-			ice_vsi_close(vsi);
-		}
-	}
-}
-
-/**
- * ice_ena_vsi - resume a VSI
- * @vsi: the VSI being resume
- */
-static int ice_ena_vsi(struct ice_vsi *vsi)
-{
-	int err = 0;
-
-	if (test_and_clear_bit(__ICE_NEEDS_RESTART, vsi->state) &&
-	    vsi->netdev) {
-		if (netif_running(vsi->netdev)) {
-			rtnl_lock();
-			err = vsi->netdev->netdev_ops->ndo_open(vsi->netdev);
-			rtnl_unlock();
 		} else {
 			err = ice_vsi_open(vsi);
 		}
@@ -3629,30 +3657,22 @@ static int ice_ena_vsi(struct ice_vsi *vsi)
 	return err;
 }
 
-/**
- * ice_pf_dis_all_vsi - Pause all VSIs on a PF
- * @pf: the PF
- */
-static void ice_pf_dis_all_vsi(struct ice_pf *pf)
-{
-	int v;
-
-	ice_for_each_vsi(pf, v)
-		if (pf->vsi[v])
-			ice_dis_vsi(pf->vsi[v], false);
-}
-
 /**
  * ice_pf_ena_all_vsi - Resume all VSIs on a PF
  * @pf: the PF
+ * @locked: is the rtnl_lock already held
  */
-static int ice_pf_ena_all_vsi(struct ice_pf *pf)
+#ifdef CONFIG_DCB
+int ice_pf_ena_all_vsi(struct ice_pf *pf, bool locked)
+#else
+static int ice_pf_ena_all_vsi(struct ice_pf *pf, bool locked)
+#endif /* CONFIG_DCB */
 {
 	int v;
 
 	ice_for_each_vsi(pf, v)
 		if (pf->vsi[v])
-			if (ice_ena_vsi(pf->vsi[v]))
+			if (ice_ena_vsi(pf->vsi[v], locked))
 				return -EIO;
 
 	return 0;
@@ -3800,7 +3820,7 @@ static void ice_rebuild(struct ice_pf *pf)
 	}
 
 	/* restart the VSIs that were rebuilt and running before the reset */
-	err = ice_pf_ena_all_vsi(pf);
+	err = ice_pf_ena_all_vsi(pf, false);
 	if (err) {
 		dev_err(&pf->pdev->dev, "error enabling VSIs\n");
 		/* no need to disable VSIs in tear down path in ice_rebuild()

drivers/net/ethernet/intel/ice/ice_sched.c

@@ -127,7 +127,7 @@ ice_aqc_send_sched_elem_cmd(struct ice_hw *hw, enum ice_adminq_opc cmd_opc,
  *
  * Query scheduling elements (0x0404)
  */
-static enum ice_status
+enum ice_status
 ice_aq_query_sched_elems(struct ice_hw *hw, u16 elems_req,
 			 struct ice_aqc_get_elem *buf, u16 buf_size,
 			 u16 *elems_ret, struct ice_sq_cd *cd)
@@ -137,31 +137,6 @@ ice_aq_query_sched_elems(struct ice_hw *hw, u16 elems_req,
 					   elems_ret, cd);
 }
 
-/**
- * ice_sched_query_elem - query element information from HW
- * @hw: pointer to the HW struct
- * @node_teid: node TEID to be queried
- * @buf: buffer to element information
- *
- * This function queries HW element information
- */
-static enum ice_status
-ice_sched_query_elem(struct ice_hw *hw, u32 node_teid,
-		     struct ice_aqc_get_elem *buf)
-{
-	u16 buf_size, num_elem_ret = 0;
-	enum ice_status status;
-
-	buf_size = sizeof(*buf);
-	memset(buf, 0, buf_size);
-	buf->generic[0].node_teid = cpu_to_le32(node_teid);
-	status = ice_aq_query_sched_elems(hw, 1, buf, buf_size, &num_elem_ret,
-					  NULL);
-	if (status || num_elem_ret != 1)
-		ice_debug(hw, ICE_DBG_SCHED, "query element failed\n");
-	return status;
-}
-
 /**
  * ice_sched_add_node - Insert the Tx scheduler node in SW DB
  * @pi: port information structure

drivers/net/ethernet/intel/ice/ice_sched.h

@@ -24,6 +24,10 @@ struct ice_sched_agg_info {
 };
 
 /* FW AQ command calls */
+enum ice_status
+ice_aq_query_sched_elems(struct ice_hw *hw, u16 elems_req,
+			 struct ice_aqc_get_elem *buf, u16 buf_size,
+			 u16 *elems_ret, struct ice_sq_cd *cd);
 enum ice_status ice_sched_init_port(struct ice_port_info *pi);
 enum ice_status ice_sched_query_res_alloc(struct ice_hw *hw);
 void ice_sched_clear_port(struct ice_port_info *pi);

drivers/net/ethernet/intel/ice/ice_type.h

@@ -215,6 +215,8 @@ struct ice_nvm_info {
 #define ice_for_each_traffic_class(_i)	\
 	for ((_i) = 0; (_i) < ICE_MAX_TRAFFIC_CLASS; (_i)++)
 
+#define ICE_INVAL_TEID 0xFFFFFFFF
+
 struct ice_sched_node {
 	struct ice_sched_node *parent;
 	struct ice_sched_node *sibling; /* next sibling in the same layer */