Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next

Pull networking updates from David Miller:

 1) Allow setting bluetooth L2CAP modes via socket option, from Luiz
    Augusto von Dentz.

 2) Add GSO partial support to igc, from Sasha Neftin.

 3) Several cleanups and improvements to r8169 from Heiner Kallweit.

 4) Add IF_OPER_TESTING link state and use it when ethtool triggers a
    device self-test. From Andrew Lunn.

 5) Start moving away from custom driver versions, use the globally
    defined kernel version instead, from Leon Romanovsky.

 6) Support GRO vis gro_cells in DSA layer, from Alexander Lobakin.

 7) Allow hard IRQ deferral during NAPI, from Eric Dumazet.

 8) Add sriov and vf support to hinic, from Luo bin.

 9) Support Media Redundancy Protocol (MRP) in the bridging code, from
    Horatiu Vultur.

10) Support netmap in the nft_nat code, from Pablo Neira Ayuso.

11) Allow UDPv6 encapsulation of ESP in the ipsec code, from Sabrina
    Dubroca. Also add ipv6 support for espintcp.

12) Lots of ReST conversions of the networking documentation, from Mauro
    Carvalho Chehab.

13) Support configuration of ethtool rxnfc flows in bcmgenet driver,
    from Doug Berger.

14) Allow to dump cgroup id and filter by it in inet_diag code, from
    Dmitry Yakunin.

15) Add infrastructure to export netlink attribute policies to
    userspace, from Johannes Berg.

16) Several optimizations to sch_fq scheduler, from Eric Dumazet.

17) Fallback to the default qdisc if qdisc init fails because otherwise
    a packet scheduler init failure will make a device inoperative. From
    Jesper Dangaard Brouer.

18) Several RISCV bpf jit optimizations, from Luke Nelson.

19) Correct the return type of the ->ndo_start_xmit() method in several
    drivers, it's netdev_tx_t but many drivers were using
    'int'. From Yunjian Wang.

20) Add an ethtool interface for PHY master/slave config, from Oleksij
    Rempel.

21) Add BPF iterators, from Yonghang Song.

22) Add cable test infrastructure, including ethool interfaces, from
    Andrew Lunn. Marvell PHY driver is the first to support this
    facility.

23) Remove zero-length arrays all over, from Gustavo A. R. Silva.

24) Calculate and maintain an explicit frame size in XDP, from Jesper
    Dangaard Brouer.

25) Add CAP_BPF, from Alexei Starovoitov.

26) Support terse dumps in the packet scheduler, from Vlad Buslov.

27) Support XDP_TX bulking in dpaa2 driver, from Ioana Ciornei.

28) Add devm_register_netdev(), from Bartosz Golaszewski.

29) Minimize qdisc resets, from Cong Wang.

30) Get rid of kernel_getsockopt and kernel_setsockopt in order to
    eliminate set_fs/get_fs calls. From Christoph Hellwig.

* git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next: (2517 commits)
  selftests: net: ip_defrag: ignore EPERM
  net_failover: fixed rollback in net_failover_open()
  Revert "tipc: Fix potential tipc_aead refcnt leak in tipc_crypto_rcv"
  Revert "tipc: Fix potential tipc_node refcnt leak in tipc_rcv"
  vmxnet3: allow rx flow hash ops only when rss is enabled
  hinic: add set_channels ethtool_ops support
  selftests/bpf: Add a default $(CXX) value
  tools/bpf: Don't use $(COMPILE.c)
  bpf, selftests: Use bpf_probe_read_kernel
  s390/bpf: Use bcr 0,%0 as tail call nop filler
  s390/bpf: Maintain 8-byte stack alignment
  selftests/bpf: Fix verifier test
  selftests/bpf: Fix sample_cnt shared between two threads
  bpf, selftests: Adapt cls_redirect to call csum_level helper
  bpf: Add csum_level helper for fixing up csum levels
  bpf: Fix up bpf_skb_adjust_room helper's skb csum setting
  sfc: add missing annotation for efx_ef10_try_update_nic_stats_vf()
  crypto/chtls: IPv6 support for inline TLS
  Crypto/chcr: Fixes a coccinile check error
  Crypto/chcr: Fixes compilations warnings
  ...

1 files changed, 1175 insertions, 535 deletions

diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
index a758f9ae32be..96bfad52630d 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
@@ -62,14 +62,16 @@ static int hclge_init_vlan_config(struct hclge_dev *hdev);
 static void hclge_sync_vlan_filter(struct hclge_dev *hdev);
 static int hclge_reset_ae_dev(struct hnae3_ae_dev *ae_dev);
 static bool hclge_get_hw_reset_stat(struct hnae3_handle *handle);
-static int hclge_set_umv_space(struct hclge_dev *hdev, u16 space_size,
-			       u16 *allocated_size, bool is_alloc);
 static void hclge_rfs_filter_expire(struct hclge_dev *hdev);
 static void hclge_clear_arfs_rules(struct hnae3_handle *handle);
 static enum hnae3_reset_type hclge_get_reset_level(struct hnae3_ae_dev *ae_dev,
 						   unsigned long *addr);
 static int hclge_set_default_loopback(struct hclge_dev *hdev);
 
+static void hclge_sync_mac_table(struct hclge_dev *hdev);
+static void hclge_restore_hw_table(struct hclge_dev *hdev);
+static void hclge_sync_promisc_mode(struct hclge_dev *hdev);
+
 static struct hnae3_ae_algo ae_algo;
 
 static struct workqueue_struct *hclge_wq;
@@ -550,7 +552,7 @@ static int hclge_tqps_update_stats(struct hnae3_handle *handle)
 		queue = handle->kinfo.tqp[i];
 		tqp = container_of(queue, struct hclge_tqp, q);
 		/* command : HCLGE_OPC_QUERY_IGU_STAT */
-		hclge_cmd_setup_basic_desc(&desc[0], HCLGE_OPC_QUERY_RX_STATUS,
+		hclge_cmd_setup_basic_desc(&desc[0], HCLGE_OPC_QUERY_RX_STATS,
 					   true);
 
 		desc[0].data[0] = cpu_to_le32((tqp->index & 0x1ff));
@@ -570,7 +572,7 @@ static int hclge_tqps_update_stats(struct hnae3_handle *handle)
 		tqp = container_of(queue, struct hclge_tqp, q);
 		/* command : HCLGE_OPC_QUERY_IGU_STAT */
 		hclge_cmd_setup_basic_desc(&desc[0],
-					   HCLGE_OPC_QUERY_TX_STATUS,
+					   HCLGE_OPC_QUERY_TX_STATS,
 					   true);
 
 		desc[0].data[0] = cpu_to_le32((tqp->index & 0x1ff));
@@ -1361,10 +1363,8 @@ static int hclge_configure(struct hclge_dev *hdev)
 	int ret;
 
 	ret = hclge_get_cfg(hdev, &cfg);
-	if (ret) {
-		dev_err(&hdev->pdev->dev, "get mac mode error %d.\n", ret);
+	if (ret)
 		return ret;
-	}
 
 	hdev->num_vmdq_vport = cfg.vmdq_vport_num;
 	hdev->base_tqp_pid = 0;
@@ -1387,7 +1387,8 @@ static int hclge_configure(struct hclge_dev *hdev)
 
 	ret = hclge_parse_speed(cfg.default_speed, &hdev->hw.mac.speed);
 	if (ret) {
-		dev_err(&hdev->pdev->dev, "Get wrong speed ret=%d.\n", ret);
+		dev_err(&hdev->pdev->dev, "failed to parse speed %u, ret = %d\n",
+			cfg.default_speed, ret);
 		return ret;
 	}
 
@@ -1429,26 +1430,17 @@ static int hclge_configure(struct hclge_dev *hdev)
 	return ret;
 }
 
-static int hclge_config_tso(struct hclge_dev *hdev, unsigned int tso_mss_min,
-			    unsigned int tso_mss_max)
+static int hclge_config_tso(struct hclge_dev *hdev, u16 tso_mss_min,
+			    u16 tso_mss_max)
 {
 	struct hclge_cfg_tso_status_cmd *req;
 	struct hclge_desc desc;
-	u16 tso_mss;
 
 	hclge_cmd_setup_basic_desc(&desc, HCLGE_OPC_TSO_GENERIC_CONFIG, false);
 
 	req = (struct hclge_cfg_tso_status_cmd *)desc.data;
-
-	tso_mss = 0;
-	hnae3_set_field(tso_mss, HCLGE_TSO_MSS_MIN_M,
-			HCLGE_TSO_MSS_MIN_S, tso_mss_min);
-	req->tso_mss_min = cpu_to_le16(tso_mss);
-
-	tso_mss = 0;
-	hnae3_set_field(tso_mss, HCLGE_TSO_MSS_MIN_M,
-			HCLGE_TSO_MSS_MIN_S, tso_mss_max);
-	req->tso_mss_max = cpu_to_le16(tso_mss);
+	req->tso_mss_min = cpu_to_le16(tso_mss_min);
+	req->tso_mss_max = cpu_to_le16(tso_mss_max);
 
 	return hclge_cmd_send(&hdev->hw, &desc, 1);
 }
@@ -1465,7 +1457,7 @@ static int hclge_config_gro(struct hclge_dev *hdev, bool en)
 	hclge_cmd_setup_basic_desc(&desc, HCLGE_OPC_GRO_GENERIC_CONFIG, false);
 	req = (struct hclge_cfg_gro_status_cmd *)desc.data;
 
-	req->gro_en = cpu_to_le16(en ? 1 : 0);
+	req->gro_en = en ? 1 : 0;
 
 	ret = hclge_cmd_send(&hdev->hw, &desc, 1);
 	if (ret)
@@ -1687,6 +1679,7 @@ static int hclge_alloc_vport(struct hclge_dev *hdev)
 		INIT_LIST_HEAD(&vport->vlan_list);
 		INIT_LIST_HEAD(&vport->uc_mac_list);
 		INIT_LIST_HEAD(&vport->mc_mac_list);
+		spin_lock_init(&vport->mac_list_lock);
 
 		if (i == 0)
 			ret = hclge_vport_setup(vport, tqp_main_vport);
@@ -2965,13 +2958,11 @@ static int hclge_set_vf_link_state(struct hnae3_handle *handle, int vf,
 
 static u32 hclge_check_event_cause(struct hclge_dev *hdev, u32 *clearval)
 {
-	u32 rst_src_reg, cmdq_src_reg, msix_src_reg;
+	u32 cmdq_src_reg, msix_src_reg;
 
 	/* fetch the events from their corresponding regs */
-	rst_src_reg = hclge_read_dev(&hdev->hw, HCLGE_MISC_VECTOR_INT_STS);
 	cmdq_src_reg = hclge_read_dev(&hdev->hw, HCLGE_VECTOR0_CMDQ_SRC_REG);
-	msix_src_reg = hclge_read_dev(&hdev->hw,
-				      HCLGE_VECTOR0_PF_OTHER_INT_STS_REG);
+	msix_src_reg = hclge_read_dev(&hdev->hw, HCLGE_MISC_VECTOR_INT_STS);
 
 	/* Assumption: If by any chance reset and mailbox events are reported
 	 * together then we will only process reset event in this go and will
@@ -2981,7 +2972,7 @@ static u32 hclge_check_event_cause(struct hclge_dev *hdev, u32 *clearval)
 	 *
 	 * check for vector0 reset event sources
 	 */
-	if (BIT(HCLGE_VECTOR0_IMPRESET_INT_B) & rst_src_reg) {
+	if (BIT(HCLGE_VECTOR0_IMPRESET_INT_B) & msix_src_reg) {
 		dev_info(&hdev->pdev->dev, "IMP reset interrupt\n");
 		set_bit(HNAE3_IMP_RESET, &hdev->reset_pending);
 		set_bit(HCLGE_STATE_CMD_DISABLE, &hdev->state);
@@ -2990,7 +2981,7 @@ static u32 hclge_check_event_cause(struct hclge_dev *hdev, u32 *clearval)
 		return HCLGE_VECTOR0_EVENT_RST;
 	}
 
-	if (BIT(HCLGE_VECTOR0_GLOBALRESET_INT_B) & rst_src_reg) {
+	if (BIT(HCLGE_VECTOR0_GLOBALRESET_INT_B) & msix_src_reg) {
 		dev_info(&hdev->pdev->dev, "global reset interrupt\n");
 		set_bit(HCLGE_STATE_CMD_DISABLE, &hdev->state);
 		set_bit(HNAE3_GLOBAL_RESET, &hdev->reset_pending);
@@ -3480,7 +3471,7 @@ static enum hnae3_reset_type hclge_get_reset_level(struct hnae3_ae_dev *ae_dev,
 	/* first, resolve any unknown reset type to the known type(s) */
 	if (test_bit(HNAE3_UNKNOWN_RESET, addr)) {
 		u32 msix_sts_reg = hclge_read_dev(&hdev->hw,
-					HCLGE_VECTOR0_PF_OTHER_INT_STS_REG);
+					HCLGE_MISC_VECTOR_INT_STS);
 		/* we will intentionally ignore any errors from this function
 		 *  as we will end up in *some* reset request in any case
 		 */
@@ -3729,22 +3720,13 @@ static int hclge_reset_stack(struct hclge_dev *hdev)
 	if (ret)
 		return ret;
 
-	ret = hclge_notify_client(hdev, HNAE3_INIT_CLIENT);
-	if (ret)
-		return ret;
-
-	return hclge_notify_client(hdev, HNAE3_RESTORE_CLIENT);
+	return hclge_notify_client(hdev, HNAE3_INIT_CLIENT);
 }
 
 static int hclge_reset_prepare(struct hclge_dev *hdev)
 {
-	struct hnae3_ae_dev *ae_dev = pci_get_drvdata(hdev->pdev);
 	int ret;
 
-	/* Initialize ae_dev reset status as well, in case enet layer wants to
-	 * know if device is undergoing reset
-	 */
-	ae_dev->reset_type = hdev->reset_type;
 	hdev->rst_stats.reset_cnt++;
 	/* perform reset of the stack & ae device for a client */
 	ret = hclge_notify_roce_client(hdev, HNAE3_DOWN_CLIENT);
@@ -3780,11 +3762,6 @@ static int hclge_reset_rebuild(struct hclge_dev *hdev)
 
 	hclge_clear_reset_cause(hdev);
 
-	ret = hclge_reset_prepare_up(hdev);
-	if (ret)
-		return ret;
-
-
 	ret = hclge_notify_roce_client(hdev, HNAE3_INIT_CLIENT);
 	/* ignore RoCE notify error if it fails HCLGE_RESET_MAX_FAIL_CNT - 1
 	 * times
@@ -3793,6 +3770,10 @@ static int hclge_reset_rebuild(struct hclge_dev *hdev)
 	    hdev->rst_stats.reset_fail_cnt < HCLGE_RESET_MAX_FAIL_CNT - 1)
 		return ret;
 
+	ret = hclge_reset_prepare_up(hdev);
+	if (ret)
+		return ret;
+
 	rtnl_lock();
 	ret = hclge_notify_client(hdev, HNAE3_UP_CLIENT);
 	rtnl_unlock();
@@ -3806,7 +3787,6 @@ static int hclge_reset_rebuild(struct hclge_dev *hdev)
 	hdev->last_reset_time = jiffies;
 	hdev->rst_stats.reset_fail_cnt = 0;
 	hdev->rst_stats.reset_done_cnt++;
-	ae_dev->reset_type = HNAE3_NONE_RESET;
 	clear_bit(HCLGE_STATE_RST_FAIL, &hdev->state);
 
 	/* if default_reset_request has a higher level reset request,
@@ -3973,6 +3953,8 @@ static void hclge_periodic_service_task(struct hclge_dev *hdev)
 	 * updated when it is triggered by mbx.
 	 */
 	hclge_update_link_status(hdev);
+	hclge_sync_mac_table(hdev);
+	hclge_sync_promisc_mode(hdev);
 
 	if (time_is_after_jiffies(hdev->last_serv_processed + HZ)) {
 		delta = jiffies - hdev->last_serv_processed;
@@ -4722,7 +4704,8 @@ static int hclge_cmd_set_promisc_mode(struct hclge_dev *hdev,
 	ret = hclge_cmd_send(&hdev->hw, &desc, 1);
 	if (ret)
 		dev_err(&hdev->pdev->dev,
-			"Set promisc mode fail, status is %d.\n", ret);
+			"failed to set vport %d promisc mode, ret = %d.\n",
+			param->vf_id, ret);
 
 	return ret;
 }
@@ -4772,6 +4755,14 @@ static int hclge_set_promisc_mode(struct hnae3_handle *handle, bool en_uc_pmc,
 					    en_bc_pmc);
 }
 
+static void hclge_request_update_promisc_mode(struct hnae3_handle *handle)
+{
+	struct hclge_vport *vport = hclge_get_vport(handle);
+	struct hclge_dev *hdev = vport->back;
+
+	set_bit(HCLGE_STATE_PROMISC_CHANGED, &hdev->state);
+}
+
 static int hclge_get_fd_mode(struct hclge_dev *hdev, u8 *fd_mode)
 {
 	struct hclge_get_fd_mode_cmd *req;
@@ -4822,7 +4813,8 @@ static int hclge_get_fd_allocation(struct hclge_dev *hdev,
 	return ret;
 }
 
-static int hclge_set_fd_key_config(struct hclge_dev *hdev, int stage_num)
+static int hclge_set_fd_key_config(struct hclge_dev *hdev,
+				   enum HCLGE_FD_STAGE stage_num)
 {
 	struct hclge_set_fd_key_config_cmd *req;
 	struct hclge_fd_key_cfg *stage;
@@ -4876,9 +4868,6 @@ static int hclge_init_fd_config(struct hclge_dev *hdev)
 		return -EOPNOTSUPP;
 	}
 
-	hdev->fd_cfg.proto_support =
-		TCP_V4_FLOW | UDP_V4_FLOW | SCTP_V4_FLOW | TCP_V6_FLOW |
-		UDP_V6_FLOW | SCTP_V6_FLOW | IPV4_USER_FLOW | IPV6_USER_FLOW;
 	key_cfg = &hdev->fd_cfg.key_cfg[HCLGE_FD_STAGE_1];
 	key_cfg->key_sel = HCLGE_FD_KEY_BASE_ON_TUPLE,
 	key_cfg->inner_sipv6_word_en = LOW_2_WORDS;
@@ -4892,11 +4881,9 @@ static int hclge_init_fd_config(struct hclge_dev *hdev)
 				BIT(INNER_SRC_PORT) | BIT(INNER_DST_PORT);
 
 	/* If use max 400bit key, we can support tuples for ether type */
-	if (hdev->fd_cfg.max_key_length == MAX_KEY_LENGTH) {
-		hdev->fd_cfg.proto_support |= ETHER_FLOW;
+	if (hdev->fd_cfg.fd_mode == HCLGE_FD_MODE_DEPTH_2K_WIDTH_400B_STAGE_1)
 		key_cfg->tuple_active |=
 				BIT(INNER_DST_MAC) | BIT(INNER_SRC_MAC);
-	}
 
 	/* roce_type is used to filter roce frames
 	 * dst_vport is used to specify the rule
@@ -5006,8 +4993,6 @@ static bool hclge_fd_convert_tuple(u32 tuple_bit, u8 *key_x, u8 *key_y,
 		return true;
 
 	switch (tuple_bit) {
-	case 0:
-		return false;
 	case BIT(INNER_DST_MAC):
 		for (i = 0; i < ETH_ALEN; i++) {
 			calc_x(key_x[ETH_ALEN - 1 - i], rule->tuples.dst_mac[i],
@@ -5165,9 +5150,10 @@ static int hclge_config_key(struct hclge_dev *hdev, u8 stage,
 	struct hclge_fd_key_cfg *key_cfg = &hdev->fd_cfg.key_cfg[stage];
 	u8 key_x[MAX_KEY_BYTES], key_y[MAX_KEY_BYTES];
 	u8 *cur_key_x, *cur_key_y;
-	unsigned int i;
-	int ret, tuple_size;
 	u8 meta_data_region;
+	u8 tuple_size;
+	int ret;
+	u32 i;
 
 	memset(key_x, 0, sizeof(key_x));
 	memset(key_y, 0, sizeof(key_y));
@@ -5244,172 +5230,255 @@ static int hclge_config_action(struct hclge_dev *hdev, u8 stage,
 	return hclge_fd_ad_config(hdev, stage, ad_data.ad_id, &ad_data);
 }
 
-static int hclge_fd_check_spec(struct hclge_dev *hdev,
-			       struct ethtool_rx_flow_spec *fs, u32 *unused)
+static int hclge_fd_check_tcpip4_tuple(struct ethtool_tcpip4_spec *spec,
+				       u32 *unused_tuple)
 {
-	struct ethtool_tcpip4_spec *tcp_ip4_spec;
-	struct ethtool_usrip4_spec *usr_ip4_spec;
-	struct ethtool_tcpip6_spec *tcp_ip6_spec;
-	struct ethtool_usrip6_spec *usr_ip6_spec;
-	struct ethhdr *ether_spec;
-
-	if (fs->location >= hdev->fd_cfg.rule_num[HCLGE_FD_STAGE_1])
+	if (!spec || !unused_tuple)
 		return -EINVAL;
 
-	if (!(fs->flow_type & hdev->fd_cfg.proto_support))
-		return -EOPNOTSUPP;
-
-	if ((fs->flow_type & FLOW_EXT) &&
-	    (fs->h_ext.data[0] != 0 || fs->h_ext.data[1] != 0)) {
-		dev_err(&hdev->pdev->dev, "user-def bytes are not supported\n");
-		return -EOPNOTSUPP;
-	}
+	*unused_tuple |= BIT(INNER_SRC_MAC) | BIT(INNER_DST_MAC);
 
-	switch (fs->flow_type & ~(FLOW_EXT | FLOW_MAC_EXT)) {
-	case SCTP_V4_FLOW:
-	case TCP_V4_FLOW:
-	case UDP_V4_FLOW:
-		tcp_ip4_spec = &fs->h_u.tcp_ip4_spec;
-		*unused |= BIT(INNER_SRC_MAC) | BIT(INNER_DST_MAC);
+	if (!spec->ip4src)
+		*unused_tuple |= BIT(INNER_SRC_IP);
 
-		if (!tcp_ip4_spec->ip4src)
-			*unused |= BIT(INNER_SRC_IP);
+	if (!spec->ip4dst)
+		*unused_tuple |= BIT(INNER_DST_IP);
 
-		if (!tcp_ip4_spec->ip4dst)
-			*unused |= BIT(INNER_DST_IP);
+	if (!spec->psrc)
+		*unused_tuple |= BIT(INNER_SRC_PORT);
 
-		if (!tcp_ip4_spec->psrc)
-			*unused |= BIT(INNER_SRC_PORT);
+	if (!spec->pdst)
+		*unused_tuple |= BIT(INNER_DST_PORT);
 
-		if (!tcp_ip4_spec->pdst)
-			*unused |= BIT(INNER_DST_PORT);
+	if (!spec->tos)
+		*unused_tuple |= BIT(INNER_IP_TOS);
 
-		if (!tcp_ip4_spec->tos)
-			*unused |= BIT(INNER_IP_TOS);
+	return 0;
+}
 
-		break;
-	case IP_USER_FLOW:
-		usr_ip4_spec = &fs->h_u.usr_ip4_spec;
-		*unused |= BIT(INNER_SRC_MAC) | BIT(INNER_DST_MAC) |
-			BIT(INNER_SRC_PORT) | BIT(INNER_DST_PORT);
+static int hclge_fd_check_ip4_tuple(struct ethtool_usrip4_spec *spec,
+				    u32 *unused_tuple)
+{
+	if (!spec || !unused_tuple)
+		return -EINVAL;
 
-		if (!usr_ip4_spec->ip4src)
-			*unused |= BIT(INNER_SRC_IP);
+	*unused_tuple |= BIT(INNER_SRC_MAC) | BIT(INNER_DST_MAC) |
+		BIT(INNER_SRC_PORT) | BIT(INNER_DST_PORT);
 
-		if (!usr_ip4_spec->ip4dst)
-			*unused |= BIT(INNER_DST_IP);
+	if (!spec->ip4src)
+		*unused_tuple |= BIT(INNER_SRC_IP);
 
-		if (!usr_ip4_spec->tos)
-			*unused |= BIT(INNER_IP_TOS);
+	if (!spec->ip4dst)
+		*unused_tuple |= BIT(INNER_DST_IP);
 
-		if (!usr_ip4_spec->proto)
-			*unused |= BIT(INNER_IP_PROTO);
+	if (!spec->tos)
+		*unused_tuple |= BIT(INNER_IP_TOS);
 
-		if (usr_ip4_spec->l4_4_bytes)
-			return -EOPNOTSUPP;
+	if (!spec->proto)
+		*unused_tuple |= BIT(INNER_IP_PROTO);
 
-		if (usr_ip4_spec->ip_ver != ETH_RX_NFC_IP4)
-			return -EOPNOTSUPP;
+	if (spec->l4_4_bytes)
+		return -EOPNOTSUPP;
 
-		break;
-	case SCTP_V6_FLOW:
-	case TCP_V6_FLOW:
-	case UDP_V6_FLOW:
-		tcp_ip6_spec = &fs->h_u.tcp_ip6_spec;
-		*unused |= BIT(INNER_SRC_MAC) | BIT(INNER_DST_MAC) |
-			BIT(INNER_IP_TOS);
+	if (spec->ip_ver != ETH_RX_NFC_IP4)
+		return -EOPNOTSUPP;
 
-		/* check whether src/dst ip address used */
-		if (!tcp_ip6_spec->ip6src[0] && !tcp_ip6_spec->ip6src[1] &&
-		    !tcp_ip6_spec->ip6src[2] && !tcp_ip6_spec->ip6src[3])
-			*unused |= BIT(INNER_SRC_IP);
+	return 0;
+}
 
-		if (!tcp_ip6_spec->ip6dst[0] && !tcp_ip6_spec->ip6dst[1] &&
-		    !tcp_ip6_spec->ip6dst[2] && !tcp_ip6_spec->ip6dst[3])
-			*unused |= BIT(INNER_DST_IP);
+static int hclge_fd_check_tcpip6_tuple(struct ethtool_tcpip6_spec *spec,
+				       u32 *unused_tuple)
+{
+	if (!spec || !unused_tuple)
+		return -EINVAL;
 
-		if (!tcp_ip6_spec->psrc)
-			*unused |= BIT(INNER_SRC_PORT);
+	*unused_tuple |= BIT(INNER_SRC_MAC) | BIT(INNER_DST_MAC) |
+		BIT(INNER_IP_TOS);
 
-		if (!tcp_ip6_spec->pdst)
-			*unused |= BIT(INNER_DST_PORT);
+	/* check whether src/dst ip address used */
+	if (!spec->ip6src[0] && !spec->ip6src[1] &&
+	    !spec->ip6src[2] && !spec->ip6src[3])
+		*unused_tuple |= BIT(INNER_SRC_IP);
 
-		if (tcp_ip6_spec->tclass)
-			return -EOPNOTSUPP;
+	if (!spec->ip6dst[0] && !spec->ip6dst[1] &&
+	    !spec->ip6dst[2] && !spec->ip6dst[3])
+		*unused_tuple |= BIT(INNER_DST_IP);
 
-		break;
-	case IPV6_USER_FLOW:
-		usr_ip6_spec = &fs->h_u.usr_ip6_spec;
-		*unused |= BIT(INNER_SRC_MAC) | BIT(INNER_DST_MAC) |
-			BIT(INNER_IP_TOS) | BIT(INNER_SRC_PORT) |
-			BIT(INNER_DST_PORT);
+	if (!spec->psrc)
+		*unused_tuple |= BIT(INNER_SRC_PORT);
 
-		/* check whether src/dst ip address used */
-		if (!usr_ip6_spec->ip6src[0] && !usr_ip6_spec->ip6src[1] &&
-		    !usr_ip6_spec->ip6src[2] && !usr_ip6_spec->ip6src[3])
-			*unused |= BIT(INNER_SRC_IP);
+	if (!spec->pdst)
+		*unused_tuple |= BIT(INNER_DST_PORT);
 
-		if (!usr_ip6_spec->ip6dst[0] && !usr_ip6_spec->ip6dst[1] &&
-		    !usr_ip6_spec->ip6dst[2] && !usr_ip6_spec->ip6dst[3])
-			*unused |= BIT(INNER_DST_IP);
+	if (spec->tclass)
+		return -EOPNOTSUPP;
 
-		if (!usr_ip6_spec->l4_proto)
-			*unused |= BIT(INNER_IP_PROTO);
+	return 0;
+}
 
-		if (usr_ip6_spec->tclass)
-			return -EOPNOTSUPP;
+static int hclge_fd_check_ip6_tuple(struct ethtool_usrip6_spec *spec,
+				    u32 *unused_tuple)
+{
+	if (!spec || !unused_tuple)
+		return -EINVAL;
 
-		if (usr_ip6_spec->l4_4_bytes)
-			return -EOPNOTSUPP;
+	*unused_tuple |= BIT(INNER_SRC_MAC) | BIT(INNER_DST_MAC) |
+		BIT(INNER_IP_TOS) | BIT(INNER_SRC_PORT) | BIT(INNER_DST_PORT);
 
-		break;
-	case ETHER_FLOW:
-		ether_spec = &fs->h_u.ether_spec;
-		*unused |= BIT(INNER_SRC_IP) | BIT(INNER_DST_IP) |
-			BIT(INNER_SRC_PORT) | BIT(INNER_DST_PORT) |
-			BIT(INNER_IP_TOS) | BIT(INNER_IP_PROTO);
+	/* check whether src/dst ip address used */
+	if (!spec->ip6src[0] && !spec->ip6src[1] &&
+	    !spec->ip6src[2] && !spec->ip6src[3])
+		*unused_tuple |= BIT(INNER_SRC_IP);
 
-		if (is_zero_ether_addr(ether_spec->h_source))
-			*unused |= BIT(INNER_SRC_MAC);
+	if (!spec->ip6dst[0] && !spec->ip6dst[1] &&
+	    !spec->ip6dst[2] && !spec->ip6dst[3])
+		*unused_tuple |= BIT(INNER_DST_IP);
 
-		if (is_zero_ether_addr(ether_spec->h_dest))
-			*unused |= BIT(INNER_DST_MAC);
+	if (!spec->l4_proto)
+		*unused_tuple |= BIT(INNER_IP_PROTO);
 
-		if (!ether_spec->h_proto)
-			*unused |= BIT(INNER_ETH_TYPE);
+	if (spec->tclass)
+		return -EOPNOTSUPP;
 
-		break;
-	default:
+	if (spec->l4_4_bytes)
 		return -EOPNOTSUPP;
-	}
 
-	if ((fs->flow_type & FLOW_EXT)) {
-		if (fs->h_ext.vlan_etype)
+	return 0;
+}
+
+static int hclge_fd_check_ether_tuple(struct ethhdr *spec, u32 *unused_tuple)
+{
+	if (!spec || !unused_tuple)
+		return -EINVAL;
+
+	*unused_tuple |= BIT(INNER_SRC_IP) | BIT(INNER_DST_IP) |
+		BIT(INNER_SRC_PORT) | BIT(INNER_DST_PORT) |
+		BIT(INNER_IP_TOS) | BIT(INNER_IP_PROTO);
+
+	if (is_zero_ether_addr(spec->h_source))
+		*unused_tuple |= BIT(INNER_SRC_MAC);
+
+	if (is_zero_ether_addr(spec->h_dest))
+		*unused_tuple |= BIT(INNER_DST_MAC);
+
+	if (!spec->h_proto)
+		*unused_tuple |= BIT(INNER_ETH_TYPE);
+
+	return 0;
+}
+
+static int hclge_fd_check_ext_tuple(struct hclge_dev *hdev,
+				    struct ethtool_rx_flow_spec *fs,
+				    u32 *unused_tuple)
+{
+	if (fs->flow_type & FLOW_EXT) {
+		if (fs->h_ext.vlan_etype) {
+			dev_err(&hdev->pdev->dev, "vlan-etype is not supported!\n");
 			return -EOPNOTSUPP;
+		}
+
 		if (!fs->h_ext.vlan_tci)
-			*unused |= BIT(INNER_VLAN_TAG_FST);
+			*unused_tuple |= BIT(INNER_VLAN_TAG_FST);
 
-		if (fs->m_ext.vlan_tci) {
-			if (be16_to_cpu(fs->h_ext.vlan_tci) >= VLAN_N_VID)
-				return -EINVAL;
+		if (fs->m_ext.vlan_tci &&
+		    be16_to_cpu(fs->h_ext.vlan_tci) >= VLAN_N_VID) {
+			dev_err(&hdev->pdev->dev,
+				"failed to config vlan_tci, invalid vlan_tci: %u, max is %u.\n",
+				ntohs(fs->h_ext.vlan_tci), VLAN_N_VID - 1);
+			return -EINVAL;
 		}
 	} else {
-		*unused |= BIT(INNER_VLAN_TAG_FST);
+		*unused_tuple |= BIT(INNER_VLAN_TAG_FST);
 	}
 
 	if (fs->flow_type & FLOW_MAC_EXT) {
-		if (!(hdev->fd_cfg.proto_support & ETHER_FLOW))
+		if (hdev->fd_cfg.fd_mode !=
+		    HCLGE_FD_MODE_DEPTH_2K_WIDTH_400B_STAGE_1) {
+			dev_err(&hdev->pdev->dev,
+				"FLOW_MAC_EXT is not supported in current fd mode!\n");
 			return -EOPNOTSUPP;
+		}
 
 		if (is_zero_ether_addr(fs->h_ext.h_dest))
-			*unused |= BIT(INNER_DST_MAC);
+			*unused_tuple |= BIT(INNER_DST_MAC);
 		else
-			*unused &= ~(BIT(INNER_DST_MAC));
+			*unused_tuple &= ~BIT(INNER_DST_MAC);
 	}
 
 	return 0;
 }
 
+static int hclge_fd_check_spec(struct hclge_dev *hdev,
+			       struct ethtool_rx_flow_spec *fs,
+			       u32 *unused_tuple)
+{
+	u32 flow_type;
+	int ret;
+
+	if (fs->location >= hdev->fd_cfg.rule_num[HCLGE_FD_STAGE_1]) {
+		dev_err(&hdev->pdev->dev,
+			"failed to config fd rules, invalid rule location: %u, max is %u\n.",
+			fs->location,
+			hdev->fd_cfg.rule_num[HCLGE_FD_STAGE_1] - 1);
+		return -EINVAL;
+	}
+
+	if ((fs->flow_type & FLOW_EXT) &&
+	    (fs->h_ext.data[0] != 0 || fs->h_ext.data[1] != 0)) {
+		dev_err(&hdev->pdev->dev, "user-def bytes are not supported\n");
+		return -EOPNOTSUPP;
+	}
+
+	flow_type = fs->flow_type & ~(FLOW_EXT | FLOW_MAC_EXT);
+	switch (flow_type) {
+	case SCTP_V4_FLOW:
+	case TCP_V4_FLOW:
+	case UDP_V4_FLOW:
+		ret = hclge_fd_check_tcpip4_tuple(&fs->h_u.tcp_ip4_spec,
+						  unused_tuple);
+		break;
+	case IP_USER_FLOW:
+		ret = hclge_fd_check_ip4_tuple(&fs->h_u.usr_ip4_spec,
+					       unused_tuple);
+		break;
+	case SCTP_V6_FLOW:
+	case TCP_V6_FLOW:
+	case UDP_V6_FLOW:
+		ret = hclge_fd_check_tcpip6_tuple(&fs->h_u.tcp_ip6_spec,
+						  unused_tuple);
+		break;
+	case IPV6_USER_FLOW:
+		ret = hclge_fd_check_ip6_tuple(&fs->h_u.usr_ip6_spec,
+					       unused_tuple);
+		break;
+	case ETHER_FLOW:
+		if (hdev->fd_cfg.fd_mode !=
+			HCLGE_FD_MODE_DEPTH_2K_WIDTH_400B_STAGE_1) {
+			dev_err(&hdev->pdev->dev,
+				"ETHER_FLOW is not supported in current fd mode!\n");
+			return -EOPNOTSUPP;
+		}
+
+		ret = hclge_fd_check_ether_tuple(&fs->h_u.ether_spec,
+						 unused_tuple);
+		break;
+	default:
+		dev_err(&hdev->pdev->dev,
+			"unsupported protocol type, protocol type = %#x\n",
+			flow_type);
+		return -EOPNOTSUPP;
+	}
+
+	if (ret) {
+		dev_err(&hdev->pdev->dev,
+			"failed to check flow union tuple, ret = %d\n",
+			ret);
+		return ret;
+	}
+
+	return hclge_fd_check_ext_tuple(hdev, fs, unused_tuple);
+}
+
 static bool hclge_fd_rule_exist(struct hclge_dev *hdev, u16 location)
 {
 	struct hclge_fd_rule *rule = NULL;
@@ -5618,7 +5687,7 @@ static int hclge_fd_get_tuple(struct hclge_dev *hdev,
 		break;
 	}
 
-	if ((fs->flow_type & FLOW_EXT)) {
+	if (fs->flow_type & FLOW_EXT) {
 		rule->tuples.vlan_tag1 = be16_to_cpu(fs->h_ext.vlan_tci);
 		rule->tuples_mask.vlan_tag1 = be16_to_cpu(fs->m_ext.vlan_tci);
 	}
@@ -5673,22 +5742,23 @@ static int hclge_add_fd_entry(struct hnae3_handle *handle,
 	u8 action;
 	int ret;
 
-	if (!hnae3_dev_fd_supported(hdev))
+	if (!hnae3_dev_fd_supported(hdev)) {
+		dev_err(&hdev->pdev->dev,
+			"flow table director is not supported\n");
 		return -EOPNOTSUPP;
+	}
 
 	if (!hdev->fd_en) {
-		dev_warn(&hdev->pdev->dev,
-			 "Please enable flow director first\n");
+		dev_err(&hdev->pdev->dev,
+			"please enable flow director first\n");
 		return -EOPNOTSUPP;
 	}
 
 	fs = (struct ethtool_rx_flow_spec *)&cmd->fs;
 
 	ret = hclge_fd_check_spec(hdev, fs, &unused);
-	if (ret) {
-		dev_err(&hdev->pdev->dev, "Check fd spec failed\n");
+	if (ret)
 		return ret;
-	}
 
 	if (fs->ring_cookie == RX_CLS_FLOW_DISC) {
 		action = HCLGE_FD_ACTION_DROP_PACKET;
@@ -5729,7 +5799,6 @@ static int hclge_add_fd_entry(struct hnae3_handle *handle,
 	}
 
 	rule->flow_type = fs->flow_type;
-
 	rule->location = fs->location;
 	rule->unused_tuple = unused;
 	rule->vf_id = dst_vport_id;
@@ -5877,6 +5946,149 @@ static int hclge_get_fd_rule_cnt(struct hnae3_handle *handle,
 	return 0;
 }
 
+static void hclge_fd_get_tcpip4_info(struct hclge_fd_rule *rule,
+				     struct ethtool_tcpip4_spec *spec,
+				     struct ethtool_tcpip4_spec *spec_mask)
+{
+	spec->ip4src = cpu_to_be32(rule->tuples.src_ip[IPV4_INDEX]);
+	spec_mask->ip4src = rule->unused_tuple & BIT(INNER_SRC_IP) ?
+			0 : cpu_to_be32(rule->tuples_mask.src_ip[IPV4_INDEX]);
+
+	spec->ip4dst = cpu_to_be32(rule->tuples.dst_ip[IPV4_INDEX]);
+	spec_mask->ip4dst = rule->unused_tuple & BIT(INNER_DST_IP) ?
+			0 : cpu_to_be32(rule->tuples_mask.dst_ip[IPV4_INDEX]);
+
+	spec->psrc = cpu_to_be16(rule->tuples.src_port);
+	spec_mask->psrc = rule->unused_tuple & BIT(INNER_SRC_PORT) ?
+			0 : cpu_to_be16(rule->tuples_mask.src_port);
+
+	spec->pdst = cpu_to_be16(rule->tuples.dst_port);
+	spec_mask->pdst = rule->unused_tuple & BIT(INNER_DST_PORT) ?
+			0 : cpu_to_be16(rule->tuples_mask.dst_port);
+
+	spec->tos = rule->tuples.ip_tos;
+	spec_mask->tos = rule->unused_tuple & BIT(INNER_IP_TOS) ?
+			0 : rule->tuples_mask.ip_tos;
+}
+
+static void hclge_fd_get_ip4_info(struct hclge_fd_rule *rule,
+				  struct ethtool_usrip4_spec *spec,
+				  struct ethtool_usrip4_spec *spec_mask)
+{
+	spec->ip4src = cpu_to_be32(rule->tuples.src_ip[IPV4_INDEX]);
+	spec_mask->ip4src = rule->unused_tuple & BIT(INNER_SRC_IP) ?
+			0 : cpu_to_be32(rule->tuples_mask.src_ip[IPV4_INDEX]);
+
+	spec->ip4dst = cpu_to_be32(rule->tuples.dst_ip[IPV4_INDEX]);
+	spec_mask->ip4dst = rule->unused_tuple & BIT(INNER_DST_IP) ?
+			0 : cpu_to_be32(rule->tuples_mask.dst_ip[IPV4_INDEX]);
+
+	spec->tos = rule->tuples.ip_tos;
+	spec_mask->tos = rule->unused_tuple & BIT(INNER_IP_TOS) ?
+			0 : rule->tuples_mask.ip_tos;
+
+	spec->proto = rule->tuples.ip_proto;
+	spec_mask->proto = rule->unused_tuple & BIT(INNER_IP_PROTO) ?
+			0 : rule->tuples_mask.ip_proto;
+
+	spec->ip_ver = ETH_RX_NFC_IP4;
+}
+
+static void hclge_fd_get_tcpip6_info(struct hclge_fd_rule *rule,
+				     struct ethtool_tcpip6_spec *spec,
+				     struct ethtool_tcpip6_spec *spec_mask)
+{
+	cpu_to_be32_array(spec->ip6src,
+			  rule->tuples.src_ip, IPV6_SIZE);
+	cpu_to_be32_array(spec->ip6dst,
+			  rule->tuples.dst_ip, IPV6_SIZE);
+	if (rule->unused_tuple & BIT(INNER_SRC_IP))
+		memset(spec_mask->ip6src, 0, sizeof(spec_mask->ip6src));
+	else
+		cpu_to_be32_array(spec_mask->ip6src, rule->tuples_mask.src_ip,
+				  IPV6_SIZE);
+
+	if (rule->unused_tuple & BIT(INNER_DST_IP))
+		memset(spec_mask->ip6dst, 0, sizeof(spec_mask->ip6dst));
+	else
+		cpu_to_be32_array(spec_mask->ip6dst, rule->tuples_mask.dst_ip,
+				  IPV6_SIZE);
+
+	spec->psrc = cpu_to_be16(rule->tuples.src_port);
+	spec_mask->psrc = rule->unused_tuple & BIT(INNER_SRC_PORT) ?
+			0 : cpu_to_be16(rule->tuples_mask.src_port);
+
+	spec->pdst = cpu_to_be16(rule->tuples.dst_port);
+	spec_mask->pdst = rule->unused_tuple & BIT(INNER_DST_PORT) ?
+			0 : cpu_to_be16(rule->tuples_mask.dst_port);
+}
+
+static void hclge_fd_get_ip6_info(struct hclge_fd_rule *rule,
+				  struct ethtool_usrip6_spec *spec,
+				  struct ethtool_usrip6_spec *spec_mask)
+{
+	cpu_to_be32_array(spec->ip6src, rule->tuples.src_ip, IPV6_SIZE);
+	cpu_to_be32_array(spec->ip6dst, rule->tuples.dst_ip, IPV6_SIZE);
+	if (rule->unused_tuple & BIT(INNER_SRC_IP))
+		memset(spec_mask->ip6src, 0, sizeof(spec_mask->ip6src));
+	else
+		cpu_to_be32_array(spec_mask->ip6src,
+				  rule->tuples_mask.src_ip, IPV6_SIZE);
+
+	if (rule->unused_tuple & BIT(INNER_DST_IP))
+		memset(spec_mask->ip6dst, 0, sizeof(spec_mask->ip6dst));
+	else
+		cpu_to_be32_array(spec_mask->ip6dst,
+				  rule->tuples_mask.dst_ip, IPV6_SIZE);
+
+	spec->l4_proto = rule->tuples.ip_proto;
+	spec_mask->l4_proto = rule->unused_tuple & BIT(INNER_IP_PROTO) ?
+			0 : rule->tuples_mask.ip_proto;
+}
+
+static void hclge_fd_get_ether_info(struct hclge_fd_rule *rule,
+				    struct ethhdr *spec,
+				    struct ethhdr *spec_mask)
+{
+	ether_addr_copy(spec->h_source, rule->tuples.src_mac);
+	ether_addr_copy(spec->h_dest, rule->tuples.dst_mac);
+
+	if (rule->unused_tuple & BIT(INNER_SRC_MAC))
+		eth_zero_addr(spec_mask->h_source);
+	else
+		ether_addr_copy(spec_mask->h_source, rule->tuples_mask.src_mac);
+
+	if (rule->unused_tuple & BIT(INNER_DST_MAC))
+		eth_zero_addr(spec_mask->h_dest);
+	else
+		ether_addr_copy(spec_mask->h_dest, rule->tuples_mask.dst_mac);
+
+	spec->h_proto = cpu_to_be16(rule->tuples.ether_proto);
+	spec_mask->h_proto = rule->unused_tuple & BIT(INNER_ETH_TYPE) ?
+			0 : cpu_to_be16(rule->tuples_mask.ether_proto);
+}
+
+static void hclge_fd_get_ext_info(struct ethtool_rx_flow_spec *fs,
+				  struct hclge_fd_rule *rule)
+{
+	if (fs->flow_type & FLOW_EXT) {
+		fs->h_ext.vlan_tci = cpu_to_be16(rule->tuples.vlan_tag1);
+		fs->m_ext.vlan_tci =
+				rule->unused_tuple & BIT(INNER_VLAN_TAG_FST) ?
+				cpu_to_be16(VLAN_VID_MASK) :
+				cpu_to_be16(rule->tuples_mask.vlan_tag1);
+	}
+
+	if (fs->flow_type & FLOW_MAC_EXT) {
+		ether_addr_copy(fs->h_ext.h_dest, rule->tuples.dst_mac);
+		if (rule->unused_tuple & BIT(INNER_DST_MAC))
+			eth_zero_addr(fs->m_u.ether_spec.h_dest);
+		else
+			ether_addr_copy(fs->m_u.ether_spec.h_dest,
+					rule->tuples_mask.dst_mac);
+	}
+}
+
 static int hclge_get_fd_rule_info(struct hnae3_handle *handle,
 				  struct ethtool_rxnfc *cmd)
 {
@@ -5909,162 +6121,34 @@ static int hclge_get_fd_rule_info(struct hnae3_handle *handle,
 	case SCTP_V4_FLOW:
 	case TCP_V4_FLOW:
 	case UDP_V4_FLOW:
-		fs->h_u.tcp_ip4_spec.ip4src =
-				cpu_to_be32(rule->tuples.src_ip[IPV4_INDEX]);
-		fs->m_u.tcp_ip4_spec.ip4src =
-			rule->unused_tuple & BIT(INNER_SRC_IP) ?
-			0 : cpu_to_be32(rule->tuples_mask.src_ip[IPV4_INDEX]);
-
-		fs->h_u.tcp_ip4_spec.ip4dst =
-				cpu_to_be32(rule->tuples.dst_ip[IPV4_INDEX]);
-		fs->m_u.tcp_ip4_spec.ip4dst =
-			rule->unused_tuple & BIT(INNER_DST_IP) ?
-			0 : cpu_to_be32(rule->tuples_mask.dst_ip[IPV4_INDEX]);
-
-		fs->h_u.tcp_ip4_spec.psrc = cpu_to_be16(rule->tuples.src_port);
-		fs->m_u.tcp_ip4_spec.psrc =
-				rule->unused_tuple & BIT(INNER_SRC_PORT) ?
-				0 : cpu_to_be16(rule->tuples_mask.src_port);
-
-		fs->h_u.tcp_ip4_spec.pdst = cpu_to_be16(rule->tuples.dst_port);
-		fs->m_u.tcp_ip4_spec.pdst =
-				rule->unused_tuple & BIT(INNER_DST_PORT) ?
-				0 : cpu_to_be16(rule->tuples_mask.dst_port);
-
-		fs->h_u.tcp_ip4_spec.tos = rule->tuples.ip_tos;
-		fs->m_u.tcp_ip4_spec.tos =
-				rule->unused_tuple & BIT(INNER_IP_TOS) ?
-				0 : rule->tuples_mask.ip_tos;
-
+		hclge_fd_get_tcpip4_info(rule, &fs->h_u.tcp_ip4_spec,
+					 &fs->m_u.tcp_ip4_spec);
 		break;
 	case IP_USER_FLOW:
-		fs->h_u.usr_ip4_spec.ip4src =
-				cpu_to_be32(rule->tuples.src_ip[IPV4_INDEX]);
-		fs->m_u.tcp_ip4_spec.ip4src =
-			rule->unused_tuple & BIT(INNER_SRC_IP) ?
-			0 : cpu_to_be32(rule->tuples_mask.src_ip[IPV4_INDEX]);
-
-		fs->h_u.usr_ip4_spec.ip4dst =
-				cpu_to_be32(rule->tuples.dst_ip[IPV4_INDEX]);
-		fs->m_u.usr_ip4_spec.ip4dst =
-			rule->unused_tuple & BIT(INNER_DST_IP) ?
-			0 : cpu_to_be32(rule->tuples_mask.dst_ip[IPV4_INDEX]);
-
-		fs->h_u.usr_ip4_spec.tos = rule->tuples.ip_tos;
-		fs->m_u.usr_ip4_spec.tos =
-				rule->unused_tuple & BIT(INNER_IP_TOS) ?
-				0 : rule->tuples_mask.ip_tos;
-
-		fs->h_u.usr_ip4_spec.proto = rule->tuples.ip_proto;
-		fs->m_u.usr_ip4_spec.proto =
-				rule->unused_tuple & BIT(INNER_IP_PROTO) ?
-				0 : rule->tuples_mask.ip_proto;
-
-		fs->h_u.usr_ip4_spec.ip_ver = ETH_RX_NFC_IP4;
-
+		hclge_fd_get_ip4_info(rule, &fs->h_u.usr_ip4_spec,
+				      &fs->m_u.usr_ip4_spec);
 		break;
 	case SCTP_V6_FLOW:
 	case TCP_V6_FLOW:
 	case UDP_V6_FLOW:
-		cpu_to_be32_array(fs->h_u.tcp_ip6_spec.ip6src,
-				  rule->tuples.src_ip, IPV6_SIZE);
-		if (rule->unused_tuple & BIT(INNER_SRC_IP))
-			memset(fs->m_u.tcp_ip6_spec.ip6src, 0,
-			       sizeof(int) * IPV6_SIZE);
-		else
-			cpu_to_be32_array(fs->m_u.tcp_ip6_spec.ip6src,
-					  rule->tuples_mask.src_ip, IPV6_SIZE);
-
-		cpu_to_be32_array(fs->h_u.tcp_ip6_spec.ip6dst,
-				  rule->tuples.dst_ip, IPV6_SIZE);
-		if (rule->unused_tuple & BIT(INNER_DST_IP))
-			memset(fs->m_u.tcp_ip6_spec.ip6dst, 0,
-			       sizeof(int) * IPV6_SIZE);
-		else
-			cpu_to_be32_array(fs->m_u.tcp_ip6_spec.ip6dst,
-					  rule->tuples_mask.dst_ip, IPV6_SIZE);
-
-		fs->h_u.tcp_ip6_spec.psrc = cpu_to_be16(rule->tuples.src_port);
-		fs->m_u.tcp_ip6_spec.psrc =
-				rule->unused_tuple & BIT(INNER_SRC_PORT) ?
-				0 : cpu_to_be16(rule->tuples_mask.src_port);
-
-		fs->h_u.tcp_ip6_spec.pdst = cpu_to_be16(rule->tuples.dst_port);
-		fs->m_u.tcp_ip6_spec.pdst =
-				rule->unused_tuple & BIT(INNER_DST_PORT) ?
-				0 : cpu_to_be16(rule->tuples_mask.dst_port);
-
+		hclge_fd_get_tcpip6_info(rule, &fs->h_u.tcp_ip6_spec,
+					 &fs->m_u.tcp_ip6_spec);
 		break;
 	case IPV6_USER_FLOW:
-		cpu_to_be32_array(fs->h_u.usr_ip6_spec.ip6src,
-				  rule->tuples.src_ip, IPV6_SIZE);
-		if (rule->unused_tuple & BIT(INNER_SRC_IP))
-			memset(fs->m_u.usr_ip6_spec.ip6src, 0,
-			       sizeof(int) * IPV6_SIZE);
-		else
-			cpu_to_be32_array(fs->m_u.usr_ip6_spec.ip6src,
-					  rule->tuples_mask.src_ip, IPV6_SIZE);
-
-		cpu_to_be32_array(fs->h_u.usr_ip6_spec.ip6dst,
-				  rule->tuples.dst_ip, IPV6_SIZE);
-		if (rule->unused_tuple & BIT(INNER_DST_IP))
-			memset(fs->m_u.usr_ip6_spec.ip6dst, 0,
-			       sizeof(int) * IPV6_SIZE);
-		else
-			cpu_to_be32_array(fs->m_u.usr_ip6_spec.ip6dst,
-					  rule->tuples_mask.dst_ip, IPV6_SIZE);
-
-		fs->h_u.usr_ip6_spec.l4_proto = rule->tuples.ip_proto;
-		fs->m_u.usr_ip6_spec.l4_proto =
-				rule->unused_tuple & BIT(INNER_IP_PROTO) ?
-				0 : rule->tuples_mask.ip_proto;
-
-		break;
-	case ETHER_FLOW:
-		ether_addr_copy(fs->h_u.ether_spec.h_source,
-				rule->tuples.src_mac);
-		if (rule->unused_tuple & BIT(INNER_SRC_MAC))
-			eth_zero_addr(fs->m_u.ether_spec.h_source);
-		else
-			ether_addr_copy(fs->m_u.ether_spec.h_source,
-					rule->tuples_mask.src_mac);
-
-		ether_addr_copy(fs->h_u.ether_spec.h_dest,
-				rule->tuples.dst_mac);
-		if (rule->unused_tuple & BIT(INNER_DST_MAC))
-			eth_zero_addr(fs->m_u.ether_spec.h_dest);
-		else
-			ether_addr_copy(fs->m_u.ether_spec.h_dest,
-					rule->tuples_mask.dst_mac);
-
-		fs->h_u.ether_spec.h_proto =
-				cpu_to_be16(rule->tuples.ether_proto);
-		fs->m_u.ether_spec.h_proto =
-				rule->unused_tuple & BIT(INNER_ETH_TYPE) ?
-				0 : cpu_to_be16(rule->tuples_mask.ether_proto);
-
+		hclge_fd_get_ip6_info(rule, &fs->h_u.usr_ip6_spec,
+				      &fs->m_u.usr_ip6_spec);
 		break;
+	/* The flow type of fd rule has been checked before adding in to rule
+	 * list. As other flow types have been handled, it must be ETHER_FLOW
+	 * for the default case
+	 */
 	default:
-		spin_unlock_bh(&hdev->fd_rule_lock);
-		return -EOPNOTSUPP;
-	}
-
-	if (fs->flow_type & FLOW_EXT) {
-		fs->h_ext.vlan_tci = cpu_to_be16(rule->tuples.vlan_tag1);
-		fs->m_ext.vlan_tci =
-				rule->unused_tuple & BIT(INNER_VLAN_TAG_FST) ?
-				cpu_to_be16(VLAN_VID_MASK) :
-				cpu_to_be16(rule->tuples_mask.vlan_tag1);
+		hclge_fd_get_ether_info(rule, &fs->h_u.ether_spec,
+					&fs->m_u.ether_spec);
+		break;
 	}
 
-	if (fs->flow_type & FLOW_MAC_EXT) {
-		ether_addr_copy(fs->h_ext.h_dest, rule->tuples.dst_mac);
-		if (rule->unused_tuple & BIT(INNER_DST_MAC))
-			eth_zero_addr(fs->m_u.ether_spec.h_dest);
-		else
-			ether_addr_copy(fs->m_u.ether_spec.h_dest,
-					rule->tuples_mask.dst_mac);
-	}
+	hclge_fd_get_ext_info(fs, rule);
 
 	if (rule->action == HCLGE_FD_ACTION_DROP_PACKET) {
 		fs->ring_cookie = RX_CLS_FLOW_DISC;
@@ -6202,7 +6286,6 @@ static int hclge_add_fd_entry_by_arfs(struct hnae3_handle *handle, u16 queue_id,
 	 */
 	if (hdev->fd_active_type == HCLGE_FD_EP_ACTIVE) {
 		spin_unlock_bh(&hdev->fd_rule_lock);
-
 		return -EOPNOTSUPP;
 	}
 
@@ -6216,14 +6299,12 @@ static int hclge_add_fd_entry_by_arfs(struct hnae3_handle *handle, u16 queue_id,
 		bit_id = find_first_zero_bit(hdev->fd_bmap, MAX_FD_FILTER_NUM);
 		if (bit_id >= hdev->fd_cfg.rule_num[HCLGE_FD_STAGE_1]) {
 			spin_unlock_bh(&hdev->fd_rule_lock);
-
 			return -ENOSPC;
 		}
 
 		rule = kzalloc(sizeof(*rule), GFP_ATOMIC);
 		if (!rule) {
 			spin_unlock_bh(&hdev->fd_rule_lock);
-
 			return -ENOMEM;
 		}
 
@@ -6310,6 +6391,14 @@ static bool hclge_get_hw_reset_stat(struct hnae3_handle *handle)
 	       hclge_read_dev(&hdev->hw, HCLGE_FUN_RST_ING);
 }
 
+static bool hclge_get_cmdq_stat(struct hnae3_handle *handle)
+{
+	struct hclge_vport *vport = hclge_get_vport(handle);
+	struct hclge_dev *hdev = vport->back;
+
+	return test_bit(HCLGE_STATE_CMD_DISABLE, &hdev->state);
+}
+
 static bool hclge_ae_dev_resetting(struct hnae3_handle *handle)
 {
 	struct hclge_vport *vport = hclge_get_vport(handle);
@@ -6486,8 +6575,6 @@ static int hclge_set_app_loopback(struct hclge_dev *hdev, bool en)
 	/* 2 Then setup the loopback flag */
 	loop_en = le32_to_cpu(req->txrx_pad_fcs_loop_en);
 	hnae3_set_bit(loop_en, HCLGE_MAC_APP_LP_B, en ? 1 : 0);
-	hnae3_set_bit(loop_en, HCLGE_MAC_TX_EN_B, en ? 1 : 0);
-	hnae3_set_bit(loop_en, HCLGE_MAC_RX_EN_B, en ? 1 : 0);
 
 	req->txrx_pad_fcs_loop_en = cpu_to_le32(loop_en);
 
@@ -6834,8 +6921,22 @@ static void hclge_ae_stop(struct hnae3_handle *handle)
 
 int hclge_vport_start(struct hclge_vport *vport)
 {
+	struct hclge_dev *hdev = vport->back;
+
 	set_bit(HCLGE_VPORT_STATE_ALIVE, &vport->state);
 	vport->last_active_jiffies = jiffies;
+
+	if (test_bit(vport->vport_id, hdev->vport_config_block)) {
+		if (vport->vport_id) {
+			hclge_restore_mac_table_common(vport);
+			hclge_restore_vport_vlan_table(vport);
+		} else {
+			hclge_restore_hw_table(hdev);
+		}
+	}
+
+	clear_bit(vport->vport_id, hdev->vport_config_block);
+
 	return 0;
 }
 
@@ -6872,17 +6973,11 @@ static int hclge_get_mac_vlan_cmd_status(struct hclge_vport *vport,
 	}
 
 	if (op == HCLGE_MAC_VLAN_ADD) {
-		if ((!resp_code) || (resp_code == 1)) {
+		if (!resp_code || resp_code == 1)
 			return 0;
-		} else if (resp_code == HCLGE_ADD_UC_OVERFLOW) {
-			dev_err(&hdev->pdev->dev,
-				"add mac addr failed for uc_overflow.\n");
-			return -ENOSPC;
-		} else if (resp_code == HCLGE_ADD_MC_OVERFLOW) {
-			dev_err(&hdev->pdev->dev,
-				"add mac addr failed for mc_overflow.\n");
+		else if (resp_code == HCLGE_ADD_UC_OVERFLOW ||
+			 resp_code == HCLGE_ADD_MC_OVERFLOW)
 			return -ENOSPC;
-		}
 
 		dev_err(&hdev->pdev->dev,
 			"add mac addr failed for undefined, code=%u.\n",
@@ -7106,52 +7201,8 @@ static int hclge_add_mac_vlan_tbl(struct hclge_vport *vport,
 	return cfg_status;
 }
 
-static int hclge_init_umv_space(struct hclge_dev *hdev)
-{
-	u16 allocated_size = 0;
-	int ret;
-
-	ret = hclge_set_umv_space(hdev, hdev->wanted_umv_size, &allocated_size,
-				  true);
-	if (ret)
-		return ret;
-
-	if (allocated_size < hdev->wanted_umv_size)
-		dev_warn(&hdev->pdev->dev,
-			 "Alloc umv space failed, want %u, get %u\n",
-			 hdev->wanted_umv_size, allocated_size);
-
-	mutex_init(&hdev->umv_mutex);
-	hdev->max_umv_size = allocated_size;
-	/* divide max_umv_size by (hdev->num_req_vfs + 2), in order to
-	 * preserve some unicast mac vlan table entries shared by pf
-	 * and its vfs.
-	 */
-	hdev->priv_umv_size = hdev->max_umv_size / (hdev->num_req_vfs + 2);
-	hdev->share_umv_size = hdev->priv_umv_size +
-			hdev->max_umv_size % (hdev->num_req_vfs + 2);
-
-	return 0;
-}
-
-static int hclge_uninit_umv_space(struct hclge_dev *hdev)
-{
-	int ret;
-
-	if (hdev->max_umv_size > 0) {
-		ret = hclge_set_umv_space(hdev, hdev->max_umv_size, NULL,
-					  false);
-		if (ret)
-			return ret;
-		hdev->max_umv_size = 0;
-	}
-	mutex_destroy(&hdev->umv_mutex);
-
-	return 0;
-}
-
 static int hclge_set_umv_space(struct hclge_dev *hdev, u16 space_size,
-			       u16 *allocated_size, bool is_alloc)
+			       u16 *allocated_size)
 {
 	struct hclge_umv_spc_alc_cmd *req;
 	struct hclge_desc desc;
@@ -7159,21 +7210,39 @@ static int hclge_set_umv_space(struct hclge_dev *hdev, u16 space_size,
 
 	req = (struct hclge_umv_spc_alc_cmd *)desc.data;
 	hclge_cmd_setup_basic_desc(&desc, HCLGE_OPC_MAC_VLAN_ALLOCATE, false);
-	if (!is_alloc)
-		hnae3_set_bit(req->allocate, HCLGE_UMV_SPC_ALC_B, 1);
 
 	req->space_size = cpu_to_le32(space_size);
 
 	ret = hclge_cmd_send(&hdev->hw, &desc, 1);
 	if (ret) {
-		dev_err(&hdev->pdev->dev,
-			"%s umv space failed for cmd_send, ret =%d\n",
-			is_alloc ? "allocate" : "free", ret);
+		dev_err(&hdev->pdev->dev, "failed to set umv space, ret = %d\n",
+			ret);
 		return ret;
 	}
 
-	if (is_alloc && allocated_size)
-		*allocated_size = le32_to_cpu(desc.data[1]);
+	*allocated_size = le32_to_cpu(desc.data[1]);
+
+	return 0;
+}
+
+static int hclge_init_umv_space(struct hclge_dev *hdev)
+{
+	u16 allocated_size = 0;
+	int ret;
+
+	ret = hclge_set_umv_space(hdev, hdev->wanted_umv_size, &allocated_size);
+	if (ret)
+		return ret;
+
+	if (allocated_size < hdev->wanted_umv_size)
+		dev_warn(&hdev->pdev->dev,
+			 "failed to alloc umv space, want %u, get %u\n",
+			 hdev->wanted_umv_size, allocated_size);
+
+	hdev->max_umv_size = allocated_size;
+	hdev->priv_umv_size = hdev->max_umv_size / (hdev->num_alloc_vport + 1);
+	hdev->share_umv_size = hdev->priv_umv_size +
+			hdev->max_umv_size % (hdev->num_alloc_vport + 1);
 
 	return 0;
 }
@@ -7188,21 +7257,25 @@ static void hclge_reset_umv_space(struct hclge_dev *hdev)
 		vport->used_umv_num = 0;
 	}
 
-	mutex_lock(&hdev->umv_mutex);
+	mutex_lock(&hdev->vport_lock);
 	hdev->share_umv_size = hdev->priv_umv_size +
-			hdev->max_umv_size % (hdev->num_req_vfs + 2);
-	mutex_unlock(&hdev->umv_mutex);
+			hdev->max_umv_size % (hdev->num_alloc_vport + 1);
+	mutex_unlock(&hdev->vport_lock);
 }
 
-static bool hclge_is_umv_space_full(struct hclge_vport *vport)
+static bool hclge_is_umv_space_full(struct hclge_vport *vport, bool need_lock)
 {
 	struct hclge_dev *hdev = vport->back;
 	bool is_full;
 
-	mutex_lock(&hdev->umv_mutex);
+	if (need_lock)
+		mutex_lock(&hdev->vport_lock);
+
 	is_full = (vport->used_umv_num >= hdev->priv_umv_size &&
 		   hdev->share_umv_size == 0);
-	mutex_unlock(&hdev->umv_mutex);
+
+	if (need_lock)
+		mutex_unlock(&hdev->vport_lock);
 
 	return is_full;
 }
@@ -7211,7 +7284,6 @@ static void hclge_update_umv_space(struct hclge_vport *vport, bool is_free)
 {
 	struct hclge_dev *hdev = vport->back;
 
-	mutex_lock(&hdev->umv_mutex);
 	if (is_free) {
 		if (vport->used_umv_num > hdev->priv_umv_size)
 			hdev->share_umv_size++;
@@ -7224,7 +7296,99 @@ static void hclge_update_umv_space(struct hclge_vport *vport, bool is_free)
 			hdev->share_umv_size--;
 		vport->used_umv_num++;
 	}
-	mutex_unlock(&hdev->umv_mutex);
+}
+
+static struct hclge_mac_node *hclge_find_mac_node(struct list_head *list,
+						  const u8 *mac_addr)
+{
+	struct hclge_mac_node *mac_node, *tmp;
+
+	list_for_each_entry_safe(mac_node, tmp, list, node)
+		if (ether_addr_equal(mac_addr, mac_node->mac_addr))
+			return mac_node;
+
+	return NULL;
+}
+
+static void hclge_update_mac_node(struct hclge_mac_node *mac_node,
+				  enum HCLGE_MAC_NODE_STATE state)
+{
+	switch (state) {
+	/* from set_rx_mode or tmp_add_list */
+	case HCLGE_MAC_TO_ADD:
+		if (mac_node->state == HCLGE_MAC_TO_DEL)
+			mac_node->state = HCLGE_MAC_ACTIVE;
+		break;
+	/* only from set_rx_mode */
+	case HCLGE_MAC_TO_DEL:
+		if (mac_node->state == HCLGE_MAC_TO_ADD) {
+			list_del(&mac_node->node);
+			kfree(mac_node);
+		} else {
+			mac_node->state = HCLGE_MAC_TO_DEL;
+		}
+		break;
+	/* only from tmp_add_list, the mac_node->state won't be
+	 * ACTIVE.
+	 */
+	case HCLGE_MAC_ACTIVE:
+		if (mac_node->state == HCLGE_MAC_TO_ADD)
+			mac_node->state = HCLGE_MAC_ACTIVE;
+
+		break;
+	}
+}
+
+int hclge_update_mac_list(struct hclge_vport *vport,
+			  enum HCLGE_MAC_NODE_STATE state,
+			  enum HCLGE_MAC_ADDR_TYPE mac_type,
+			  const unsigned char *addr)
+{
+	struct hclge_dev *hdev = vport->back;
+	struct hclge_mac_node *mac_node;
+	struct list_head *list;
+
+	list = (mac_type == HCLGE_MAC_ADDR_UC) ?
+		&vport->uc_mac_list : &vport->mc_mac_list;
+
+	spin_lock_bh(&vport->mac_list_lock);
+
+	/* if the mac addr is already in the mac list, no need to add a new
+	 * one into it, just check the mac addr state, convert it to a new
+	 * new state, or just remove it, or do nothing.
+	 */
+	mac_node = hclge_find_mac_node(list, addr);
+	if (mac_node) {
+		hclge_update_mac_node(mac_node, state);
+		spin_unlock_bh(&vport->mac_list_lock);
+		set_bit(HCLGE_VPORT_STATE_MAC_TBL_CHANGE, &vport->state);
+		return 0;
+	}
+
+	/* if this address is never added, unnecessary to delete */
+	if (state == HCLGE_MAC_TO_DEL) {
+		spin_unlock_bh(&vport->mac_list_lock);
+		dev_err(&hdev->pdev->dev,
+			"failed to delete address %pM from mac list\n",
+			addr);
+		return -ENOENT;
+	}
+
+	mac_node = kzalloc(sizeof(*mac_node), GFP_ATOMIC);
+	if (!mac_node) {
+		spin_unlock_bh(&vport->mac_list_lock);
+		return -ENOMEM;
+	}
+
+	set_bit(HCLGE_VPORT_STATE_MAC_TBL_CHANGE, &vport->state);
+
+	mac_node->state = state;
+	ether_addr_copy(mac_node->mac_addr, addr);
+	list_add_tail(&mac_node->node, list);
+
+	spin_unlock_bh(&vport->mac_list_lock);
+
+	return 0;
 }
 
 static int hclge_add_uc_addr(struct hnae3_handle *handle,
@@ -7232,7 +7396,8 @@ static int hclge_add_uc_addr(struct hnae3_handle *handle,
 {
 	struct hclge_vport *vport = hclge_get_vport(handle);
 
-	return hclge_add_uc_addr_common(vport, addr);
+	return hclge_update_mac_list(vport, HCLGE_MAC_TO_ADD, HCLGE_MAC_ADDR_UC,
+				     addr);
 }
 
 int hclge_add_uc_addr_common(struct hclge_vport *vport,
@@ -7271,15 +7436,19 @@ int hclge_add_uc_addr_common(struct hclge_vport *vport,
 	 */
 	ret = hclge_lookup_mac_vlan_tbl(vport, &req, &desc, false);
 	if (ret == -ENOENT) {
-		if (!hclge_is_umv_space_full(vport)) {
+		mutex_lock(&hdev->vport_lock);
+		if (!hclge_is_umv_space_full(vport, false)) {
 			ret = hclge_add_mac_vlan_tbl(vport, &req, NULL);
 			if (!ret)
 				hclge_update_umv_space(vport, false);
+			mutex_unlock(&hdev->vport_lock);
 			return ret;
 		}
+		mutex_unlock(&hdev->vport_lock);
 
-		dev_err(&hdev->pdev->dev, "UC MAC table full(%u)\n",
-			hdev->priv_umv_size);
+		if (!(vport->overflow_promisc_flags & HNAE3_OVERFLOW_UPE))
+			dev_err(&hdev->pdev->dev, "UC MAC table full(%u)\n",
+				hdev->priv_umv_size);
 
 		return -ENOSPC;
 	}
@@ -7303,7 +7472,8 @@ static int hclge_rm_uc_addr(struct hnae3_handle *handle,
 {
 	struct hclge_vport *vport = hclge_get_vport(handle);
 
-	return hclge_rm_uc_addr_common(vport, addr);
+	return hclge_update_mac_list(vport, HCLGE_MAC_TO_DEL, HCLGE_MAC_ADDR_UC,
+				     addr);
 }
 
 int hclge_rm_uc_addr_common(struct hclge_vport *vport,
@@ -7326,8 +7496,13 @@ int hclge_rm_uc_addr_common(struct hclge_vport *vport,
 	hnae3_set_bit(req.entry_type, HCLGE_MAC_VLAN_BIT0_EN_B, 0);
 	hclge_prepare_mac_addr(&req, addr, false);
 	ret = hclge_remove_mac_vlan_tbl(vport, &req);
-	if (!ret)
+	if (!ret) {
+		mutex_lock(&hdev->vport_lock);
 		hclge_update_umv_space(vport, true);
+		mutex_unlock(&hdev->vport_lock);
+	} else if (ret == -ENOENT) {
+		ret = 0;
+	}
 
 	return ret;
 }
@@ -7337,7 +7512,8 @@ static int hclge_add_mc_addr(struct hnae3_handle *handle,
 {
 	struct hclge_vport *vport = hclge_get_vport(handle);
 
-	return hclge_add_mc_addr_common(vport, addr);
+	return hclge_update_mac_list(vport, HCLGE_MAC_TO_ADD, HCLGE_MAC_ADDR_MC,
+				     addr);
 }
 
 int hclge_add_mc_addr_common(struct hclge_vport *vport,
@@ -7369,7 +7545,9 @@ int hclge_add_mc_addr_common(struct hclge_vport *vport,
 		return status;
 	status = hclge_add_mac_vlan_tbl(vport, &req, desc);
 
-	if (status == -ENOSPC)
+	/* if already overflow, not to print each time */
+	if (status == -ENOSPC &&
+	    !(vport->overflow_promisc_flags & HNAE3_OVERFLOW_MPE))
 		dev_err(&hdev->pdev->dev, "mc mac vlan table is full\n");
 
 	return status;
@@ -7380,7 +7558,8 @@ static int hclge_rm_mc_addr(struct hnae3_handle *handle,
 {
 	struct hclge_vport *vport = hclge_get_vport(handle);
 
-	return hclge_rm_mc_addr_common(vport, addr);
+	return hclge_update_mac_list(vport, HCLGE_MAC_TO_DEL, HCLGE_MAC_ADDR_MC,
+				     addr);
 }
 
 int hclge_rm_mc_addr_common(struct hclge_vport *vport,
@@ -7415,111 +7594,354 @@ int hclge_rm_mc_addr_common(struct hclge_vport *vport,
 			/* Not all the vfid is zero, update the vfid */
 			status = hclge_add_mac_vlan_tbl(vport, &req, desc);
 
-	} else {
-		/* Maybe this mac address is in mta table, but it cannot be
-		 * deleted here because an entry of mta represents an address
-		 * range rather than a specific address. the delete action to
-		 * all entries will take effect in update_mta_status called by
-		 * hns3_nic_set_rx_mode.
-		 */
+	} else if (status == -ENOENT) {
 		status = 0;
 	}
 
 	return status;
 }
 
-void hclge_add_vport_mac_table(struct hclge_vport *vport, const u8 *mac_addr,
-			       enum HCLGE_MAC_ADDR_TYPE mac_type)
+static void hclge_sync_vport_mac_list(struct hclge_vport *vport,
+				      struct list_head *list,
+				      int (*sync)(struct hclge_vport *,
+						  const unsigned char *))
 {
-	struct hclge_vport_mac_addr_cfg *mac_cfg;
-	struct list_head *list;
+	struct hclge_mac_node *mac_node, *tmp;
+	int ret;
 
-	if (!vport->vport_id)
-		return;
+	list_for_each_entry_safe(mac_node, tmp, list, node) {
+		ret = sync(vport, mac_node->mac_addr);
+		if (!ret) {
+			mac_node->state = HCLGE_MAC_ACTIVE;
+		} else {
+			set_bit(HCLGE_VPORT_STATE_MAC_TBL_CHANGE,
+				&vport->state);
+			break;
+		}
+	}
+}
 
-	mac_cfg = kzalloc(sizeof(*mac_cfg), GFP_KERNEL);
-	if (!mac_cfg)
-		return;
+static void hclge_unsync_vport_mac_list(struct hclge_vport *vport,
+					struct list_head *list,
+					int (*unsync)(struct hclge_vport *,
+						      const unsigned char *))
+{
+	struct hclge_mac_node *mac_node, *tmp;
+	int ret;
+
+	list_for_each_entry_safe(mac_node, tmp, list, node) {
+		ret = unsync(vport, mac_node->mac_addr);
+		if (!ret || ret == -ENOENT) {
+			list_del(&mac_node->node);
+			kfree(mac_node);
+		} else {
+			set_bit(HCLGE_VPORT_STATE_MAC_TBL_CHANGE,
+				&vport->state);
+			break;
+		}
+	}
+}
 
-	mac_cfg->hd_tbl_status = true;
-	memcpy(mac_cfg->mac_addr, mac_addr, ETH_ALEN);
+static bool hclge_sync_from_add_list(struct list_head *add_list,
+				     struct list_head *mac_list)
+{
+	struct hclge_mac_node *mac_node, *tmp, *new_node;
+	bool all_added = true;
 
-	list = (mac_type == HCLGE_MAC_ADDR_UC) ?
-	       &vport->uc_mac_list : &vport->mc_mac_list;
+	list_for_each_entry_safe(mac_node, tmp, add_list, node) {
+		if (mac_node->state == HCLGE_MAC_TO_ADD)
+			all_added = false;
 
-	list_add_tail(&mac_cfg->node, list);
+		/* if the mac address from tmp_add_list is not in the
+		 * uc/mc_mac_list, it means have received a TO_DEL request
+		 * during the time window of adding the mac address into mac
+		 * table. if mac_node state is ACTIVE, then change it to TO_DEL,
+		 * then it will be removed at next time. else it must be TO_ADD,
+		 * this address hasn't been added into mac table,
+		 * so just remove the mac node.
+		 */
+		new_node = hclge_find_mac_node(mac_list, mac_node->mac_addr);
+		if (new_node) {
+			hclge_update_mac_node(new_node, mac_node->state);
+			list_del(&mac_node->node);
+			kfree(mac_node);
+		} else if (mac_node->state == HCLGE_MAC_ACTIVE) {
+			mac_node->state = HCLGE_MAC_TO_DEL;
+			list_del(&mac_node->node);
+			list_add_tail(&mac_node->node, mac_list);
+		} else {
+			list_del(&mac_node->node);
+			kfree(mac_node);
+		}
+	}
+
+	return all_added;
 }
 
-void hclge_rm_vport_mac_table(struct hclge_vport *vport, const u8 *mac_addr,
-			      bool is_write_tbl,
-			      enum HCLGE_MAC_ADDR_TYPE mac_type)
+static void hclge_sync_from_del_list(struct list_head *del_list,
+				     struct list_head *mac_list)
 {
-	struct hclge_vport_mac_addr_cfg *mac_cfg, *tmp;
-	struct list_head *list;
-	bool uc_flag, mc_flag;
+	struct hclge_mac_node *mac_node, *tmp, *new_node;
 
-	list = (mac_type == HCLGE_MAC_ADDR_UC) ?
-	       &vport->uc_mac_list : &vport->mc_mac_list;
+	list_for_each_entry_safe(mac_node, tmp, del_list, node) {
+		new_node = hclge_find_mac_node(mac_list, mac_node->mac_addr);
+		if (new_node) {
+			/* If the mac addr exists in the mac list, it means
+			 * received a new TO_ADD request during the time window
+			 * of configuring the mac address. For the mac node
+			 * state is TO_ADD, and the address is already in the
+			 * in the hardware(due to delete fail), so we just need
+			 * to change the mac node state to ACTIVE.
+			 */
+			new_node->state = HCLGE_MAC_ACTIVE;
+			list_del(&mac_node->node);
+			kfree(mac_node);
+		} else {
+			list_del(&mac_node->node);
+			list_add_tail(&mac_node->node, mac_list);
+		}
+	}
+}
 
-	uc_flag = is_write_tbl && mac_type == HCLGE_MAC_ADDR_UC;
-	mc_flag = is_write_tbl && mac_type == HCLGE_MAC_ADDR_MC;
+static void hclge_update_overflow_flags(struct hclge_vport *vport,
+					enum HCLGE_MAC_ADDR_TYPE mac_type,
+					bool is_all_added)
+{
+	if (mac_type == HCLGE_MAC_ADDR_UC) {
+		if (is_all_added)
+			vport->overflow_promisc_flags &= ~HNAE3_OVERFLOW_UPE;
+		else
+			vport->overflow_promisc_flags |= HNAE3_OVERFLOW_UPE;
+	} else {
+		if (is_all_added)
+			vport->overflow_promisc_flags &= ~HNAE3_OVERFLOW_MPE;
+		else
+			vport->overflow_promisc_flags |= HNAE3_OVERFLOW_MPE;
+	}
+}
 
-	list_for_each_entry_safe(mac_cfg, tmp, list, node) {
-		if (ether_addr_equal(mac_cfg->mac_addr, mac_addr)) {
-			if (uc_flag && mac_cfg->hd_tbl_status)
-				hclge_rm_uc_addr_common(vport, mac_addr);
+static void hclge_sync_vport_mac_table(struct hclge_vport *vport,
+				       enum HCLGE_MAC_ADDR_TYPE mac_type)
+{
+	struct hclge_mac_node *mac_node, *tmp, *new_node;
+	struct list_head tmp_add_list, tmp_del_list;
+	struct list_head *list;
+	bool all_added;
 
-			if (mc_flag && mac_cfg->hd_tbl_status)
-				hclge_rm_mc_addr_common(vport, mac_addr);
+	INIT_LIST_HEAD(&tmp_add_list);
+	INIT_LIST_HEAD(&tmp_del_list);
 
-			list_del(&mac_cfg->node);
-			kfree(mac_cfg);
+	/* move the mac addr to the tmp_add_list and tmp_del_list, then
+	 * we can add/delete these mac addr outside the spin lock
+	 */
+	list = (mac_type == HCLGE_MAC_ADDR_UC) ?
+		&vport->uc_mac_list : &vport->mc_mac_list;
+
+	spin_lock_bh(&vport->mac_list_lock);
+
+	list_for_each_entry_safe(mac_node, tmp, list, node) {
+		switch (mac_node->state) {
+		case HCLGE_MAC_TO_DEL:
+			list_del(&mac_node->node);
+			list_add_tail(&mac_node->node, &tmp_del_list);
+			break;
+		case HCLGE_MAC_TO_ADD:
+			new_node = kzalloc(sizeof(*new_node), GFP_ATOMIC);
+			if (!new_node)
+				goto stop_traverse;
+			ether_addr_copy(new_node->mac_addr, mac_node->mac_addr);
+			new_node->state = mac_node->state;
+			list_add_tail(&new_node->node, &tmp_add_list);
+			break;
+		default:
 			break;
 		}
 	}
+
+stop_traverse:
+	spin_unlock_bh(&vport->mac_list_lock);
+
+	/* delete first, in order to get max mac table space for adding */
+	if (mac_type == HCLGE_MAC_ADDR_UC) {
+		hclge_unsync_vport_mac_list(vport, &tmp_del_list,
+					    hclge_rm_uc_addr_common);
+		hclge_sync_vport_mac_list(vport, &tmp_add_list,
+					  hclge_add_uc_addr_common);
+	} else {
+		hclge_unsync_vport_mac_list(vport, &tmp_del_list,
+					    hclge_rm_mc_addr_common);
+		hclge_sync_vport_mac_list(vport, &tmp_add_list,
+					  hclge_add_mc_addr_common);
+	}
+
+	/* if some mac addresses were added/deleted fail, move back to the
+	 * mac_list, and retry at next time.
+	 */
+	spin_lock_bh(&vport->mac_list_lock);
+
+	hclge_sync_from_del_list(&tmp_del_list, list);
+	all_added = hclge_sync_from_add_list(&tmp_add_list, list);
+
+	spin_unlock_bh(&vport->mac_list_lock);
+
+	hclge_update_overflow_flags(vport, mac_type, all_added);
+}
+
+static bool hclge_need_sync_mac_table(struct hclge_vport *vport)
+{
+	struct hclge_dev *hdev = vport->back;
+
+	if (test_bit(vport->vport_id, hdev->vport_config_block))
+		return false;
+
+	if (test_and_clear_bit(HCLGE_VPORT_STATE_MAC_TBL_CHANGE, &vport->state))
+		return true;
+
+	return false;
+}
+
+static void hclge_sync_mac_table(struct hclge_dev *hdev)
+{
+	int i;
+
+	for (i = 0; i < hdev->num_alloc_vport; i++) {
+		struct hclge_vport *vport = &hdev->vport[i];
+
+		if (!hclge_need_sync_mac_table(vport))
+			continue;
+
+		hclge_sync_vport_mac_table(vport, HCLGE_MAC_ADDR_UC);
+		hclge_sync_vport_mac_table(vport, HCLGE_MAC_ADDR_MC);
+	}
 }
 
 void hclge_rm_vport_all_mac_table(struct hclge_vport *vport, bool is_del_list,
 				  enum HCLGE_MAC_ADDR_TYPE mac_type)
 {
-	struct hclge_vport_mac_addr_cfg *mac_cfg, *tmp;
-	struct list_head *list;
+	int (*unsync)(struct hclge_vport *vport, const unsigned char *addr);
+	struct hclge_mac_node *mac_cfg, *tmp;
+	struct hclge_dev *hdev = vport->back;
+	struct list_head tmp_del_list, *list;
+	int ret;
 
-	list = (mac_type == HCLGE_MAC_ADDR_UC) ?
-	       &vport->uc_mac_list : &vport->mc_mac_list;
+	if (mac_type == HCLGE_MAC_ADDR_UC) {
+		list = &vport->uc_mac_list;
+		unsync = hclge_rm_uc_addr_common;
+	} else {
+		list = &vport->mc_mac_list;
+		unsync = hclge_rm_mc_addr_common;
+	}
 
-	list_for_each_entry_safe(mac_cfg, tmp, list, node) {
-		if (mac_type == HCLGE_MAC_ADDR_UC && mac_cfg->hd_tbl_status)
-			hclge_rm_uc_addr_common(vport, mac_cfg->mac_addr);
+	INIT_LIST_HEAD(&tmp_del_list);
 
-		if (mac_type == HCLGE_MAC_ADDR_MC && mac_cfg->hd_tbl_status)
-			hclge_rm_mc_addr_common(vport, mac_cfg->mac_addr);
+	if (!is_del_list)
+		set_bit(vport->vport_id, hdev->vport_config_block);
 
-		mac_cfg->hd_tbl_status = false;
-		if (is_del_list) {
+	spin_lock_bh(&vport->mac_list_lock);
+
+	list_for_each_entry_safe(mac_cfg, tmp, list, node) {
+		switch (mac_cfg->state) {
+		case HCLGE_MAC_TO_DEL:
+		case HCLGE_MAC_ACTIVE:
 			list_del(&mac_cfg->node);
-			kfree(mac_cfg);
+			list_add_tail(&mac_cfg->node, &tmp_del_list);
+			break;
+		case HCLGE_MAC_TO_ADD:
+			if (is_del_list) {
+				list_del(&mac_cfg->node);
+				kfree(mac_cfg);
+			}
+			break;
+		}
+	}
+
+	spin_unlock_bh(&vport->mac_list_lock);
+
+	list_for_each_entry_safe(mac_cfg, tmp, &tmp_del_list, node) {
+		ret = unsync(vport, mac_cfg->mac_addr);
+		if (!ret || ret == -ENOENT) {
+			/* clear all mac addr from hardware, but remain these
+			 * mac addr in the mac list, and restore them after
+			 * vf reset finished.
+			 */
+			if (!is_del_list &&
+			    mac_cfg->state == HCLGE_MAC_ACTIVE) {
+				mac_cfg->state = HCLGE_MAC_TO_ADD;
+			} else {
+				list_del(&mac_cfg->node);
+				kfree(mac_cfg);
+			}
+		} else if (is_del_list) {
+			mac_cfg->state = HCLGE_MAC_TO_DEL;
+		}
+	}
+
+	spin_lock_bh(&vport->mac_list_lock);
+
+	hclge_sync_from_del_list(&tmp_del_list, list);
+
+	spin_unlock_bh(&vport->mac_list_lock);
+}
+
+/* remove all mac address when uninitailize */
+static void hclge_uninit_vport_mac_list(struct hclge_vport *vport,
+					enum HCLGE_MAC_ADDR_TYPE mac_type)
+{
+	struct hclge_mac_node *mac_node, *tmp;
+	struct hclge_dev *hdev = vport->back;
+	struct list_head tmp_del_list, *list;
+
+	INIT_LIST_HEAD(&tmp_del_list);
+
+	list = (mac_type == HCLGE_MAC_ADDR_UC) ?
+		&vport->uc_mac_list : &vport->mc_mac_list;
+
+	spin_lock_bh(&vport->mac_list_lock);
+
+	list_for_each_entry_safe(mac_node, tmp, list, node) {
+		switch (mac_node->state) {
+		case HCLGE_MAC_TO_DEL:
+		case HCLGE_MAC_ACTIVE:
+			list_del(&mac_node->node);
+			list_add_tail(&mac_node->node, &tmp_del_list);
+			break;
+		case HCLGE_MAC_TO_ADD:
+			list_del(&mac_node->node);
+			kfree(mac_node);
+			break;
 		}
 	}
+
+	spin_unlock_bh(&vport->mac_list_lock);
+
+	if (mac_type == HCLGE_MAC_ADDR_UC)
+		hclge_unsync_vport_mac_list(vport, &tmp_del_list,
+					    hclge_rm_uc_addr_common);
+	else
+		hclge_unsync_vport_mac_list(vport, &tmp_del_list,
+					    hclge_rm_mc_addr_common);
+
+	if (!list_empty(&tmp_del_list))
+		dev_warn(&hdev->pdev->dev,
+			 "uninit %s mac list for vport %u not completely.\n",
+			 mac_type == HCLGE_MAC_ADDR_UC ? "uc" : "mc",
+			 vport->vport_id);
+
+	list_for_each_entry_safe(mac_node, tmp, &tmp_del_list, node) {
+		list_del(&mac_node->node);
+		kfree(mac_node);
+	}
 }
 
-void hclge_uninit_vport_mac_table(struct hclge_dev *hdev)
+static void hclge_uninit_mac_table(struct hclge_dev *hdev)
 {
-	struct hclge_vport_mac_addr_cfg *mac, *tmp;
 	struct hclge_vport *vport;
 	int i;
 
 	for (i = 0; i < hdev->num_alloc_vport; i++) {
 		vport = &hdev->vport[i];
-		list_for_each_entry_safe(mac, tmp, &vport->uc_mac_list, node) {
-			list_del(&mac->node);
-			kfree(mac);
-		}
-
-		list_for_each_entry_safe(mac, tmp, &vport->mc_mac_list, node) {
-			list_del(&mac->node);
-			kfree(mac);
-		}
+		hclge_uninit_vport_mac_list(vport, HCLGE_MAC_ADDR_UC);
+		hclge_uninit_vport_mac_list(vport, HCLGE_MAC_ADDR_MC);
 	}
 }
 
@@ -7683,12 +8105,57 @@ static void hclge_get_mac_addr(struct hnae3_handle *handle, u8 *p)
 	ether_addr_copy(p, hdev->hw.mac.mac_addr);
 }
 
+int hclge_update_mac_node_for_dev_addr(struct hclge_vport *vport,
+				       const u8 *old_addr, const u8 *new_addr)
+{
+	struct list_head *list = &vport->uc_mac_list;
+	struct hclge_mac_node *old_node, *new_node;
+
+	new_node = hclge_find_mac_node(list, new_addr);
+	if (!new_node) {
+		new_node = kzalloc(sizeof(*new_node), GFP_ATOMIC);
+		if (!new_node)
+			return -ENOMEM;
+
+		new_node->state = HCLGE_MAC_TO_ADD;
+		ether_addr_copy(new_node->mac_addr, new_addr);
+		list_add(&new_node->node, list);
+	} else {
+		if (new_node->state == HCLGE_MAC_TO_DEL)
+			new_node->state = HCLGE_MAC_ACTIVE;
+
+		/* make sure the new addr is in the list head, avoid dev
+		 * addr may be not re-added into mac table for the umv space
+		 * limitation after global/imp reset which will clear mac
+		 * table by hardware.
+		 */
+		list_move(&new_node->node, list);
+	}
+
+	if (old_addr && !ether_addr_equal(old_addr, new_addr)) {
+		old_node = hclge_find_mac_node(list, old_addr);
+		if (old_node) {
+			if (old_node->state == HCLGE_MAC_TO_ADD) {
+				list_del(&old_node->node);
+				kfree(old_node);
+			} else {
+				old_node->state = HCLGE_MAC_TO_DEL;
+			}
+		}
+	}
+
+	set_bit(HCLGE_VPORT_STATE_MAC_TBL_CHANGE, &vport->state);
+
+	return 0;
+}
+
 static int hclge_set_mac_addr(struct hnae3_handle *handle, void *p,
 			      bool is_first)
 {
 	const unsigned char *new_addr = (const unsigned char *)p;
 	struct hclge_vport *vport = hclge_get_vport(handle);
 	struct hclge_dev *hdev = vport->back;
+	unsigned char *old_addr = NULL;
 	int ret;
 
 	/* mac addr check */
@@ -7696,39 +8163,42 @@ static int hclge_set_mac_addr(struct hnae3_handle *handle, void *p,
 	    is_broadcast_ether_addr(new_addr) ||
 	    is_multicast_ether_addr(new_addr)) {
 		dev_err(&hdev->pdev->dev,
-			"Change uc mac err! invalid mac:%pM.\n",
+			"change uc mac err! invalid mac: %pM.\n",
 			 new_addr);
 		return -EINVAL;
 	}
 
-	if ((!is_first || is_kdump_kernel()) &&
-	    hclge_rm_uc_addr(handle, hdev->hw.mac.mac_addr))
-		dev_warn(&hdev->pdev->dev,
-			 "remove old uc mac address fail.\n");
-
-	ret = hclge_add_uc_addr(handle, new_addr);
+	ret = hclge_pause_addr_cfg(hdev, new_addr);
 	if (ret) {
 		dev_err(&hdev->pdev->dev,
-			"add uc mac address fail, ret =%d.\n",
+			"failed to configure mac pause address, ret = %d\n",
 			ret);
-
-		if (!is_first &&
-		    hclge_add_uc_addr(handle, hdev->hw.mac.mac_addr))
-			dev_err(&hdev->pdev->dev,
-				"restore uc mac address fail.\n");
-
-		return -EIO;
+		return ret;
 	}
 
-	ret = hclge_pause_addr_cfg(hdev, new_addr);
+	if (!is_first)
+		old_addr = hdev->hw.mac.mac_addr;
+
+	spin_lock_bh(&vport->mac_list_lock);
+	ret = hclge_update_mac_node_for_dev_addr(vport, old_addr, new_addr);
 	if (ret) {
 		dev_err(&hdev->pdev->dev,
-			"configure mac pause address fail, ret =%d.\n",
-			ret);
-		return -EIO;
-	}
+			"failed to change the mac addr:%pM, ret = %d\n",
+			new_addr, ret);
+		spin_unlock_bh(&vport->mac_list_lock);
 
+		if (!is_first)
+			hclge_pause_addr_cfg(hdev, old_addr);
+
+		return ret;
+	}
+	/* we must update dev addr with spin lock protect, preventing dev addr
+	 * being removed by set_rx_mode path.
+	 */
 	ether_addr_copy(hdev->hw.mac.mac_addr, new_addr);
+	spin_unlock_bh(&vport->mac_list_lock);
+
+	hclge_task_schedule(hdev, 0);
 
 	return 0;
 }
@@ -8308,42 +8778,80 @@ void hclge_uninit_vport_vlan_table(struct hclge_dev *hdev)
 	}
 }
 
-static void hclge_restore_vlan_table(struct hnae3_handle *handle)
+void hclge_restore_vport_vlan_table(struct hclge_vport *vport)
 {
-	struct hclge_vport *vport = hclge_get_vport(handle);
 	struct hclge_vport_vlan_cfg *vlan, *tmp;
 	struct hclge_dev *hdev = vport->back;
 	u16 vlan_proto;
-	u16 state, vlan_id;
-	int i;
+	u16 vlan_id;
+	u16 state;
+	int ret;
 
-	for (i = 0; i < hdev->num_alloc_vport; i++) {
-		vport = &hdev->vport[i];
-		vlan_proto = vport->port_base_vlan_cfg.vlan_info.vlan_proto;
-		vlan_id = vport->port_base_vlan_cfg.vlan_info.vlan_tag;
-		state = vport->port_base_vlan_cfg.state;
+	vlan_proto = vport->port_base_vlan_cfg.vlan_info.vlan_proto;
+	vlan_id = vport->port_base_vlan_cfg.vlan_info.vlan_tag;
+	state = vport->port_base_vlan_cfg.state;
 
-		if (state != HNAE3_PORT_BASE_VLAN_DISABLE) {
-			hclge_set_vlan_filter_hw(hdev, htons(vlan_proto),
-						 vport->vport_id, vlan_id,
-						 false);
-			continue;
-		}
+	if (state != HNAE3_PORT_BASE_VLAN_DISABLE) {
+		clear_bit(vport->vport_id, hdev->vlan_table[vlan_id]);
+		hclge_set_vlan_filter_hw(hdev, htons(vlan_proto),
+					 vport->vport_id, vlan_id,
+					 false);
+		return;
+	}
 
-		list_for_each_entry_safe(vlan, tmp, &vport->vlan_list, node) {
-			int ret;
+	list_for_each_entry_safe(vlan, tmp, &vport->vlan_list, node) {
+		ret = hclge_set_vlan_filter_hw(hdev, htons(ETH_P_8021Q),
+					       vport->vport_id,
+					       vlan->vlan_id, false);
+		if (ret)
+			break;
+		vlan->hd_tbl_status = true;
+	}
+}
 
-			if (!vlan->hd_tbl_status)
-				continue;
-			ret = hclge_set_vlan_filter_hw(hdev, htons(ETH_P_8021Q),
-						       vport->vport_id,
-						       vlan->vlan_id, false);
-			if (ret)
-				break;
+/* For global reset and imp reset, hardware will clear the mac table,
+ * so we change the mac address state from ACTIVE to TO_ADD, then they
+ * can be restored in the service task after reset complete. Furtherly,
+ * the mac addresses with state TO_DEL or DEL_FAIL are unnecessary to
+ * be restored after reset, so just remove these mac nodes from mac_list.
+ */
+static void hclge_mac_node_convert_for_reset(struct list_head *list)
+{
+	struct hclge_mac_node *mac_node, *tmp;
+
+	list_for_each_entry_safe(mac_node, tmp, list, node) {
+		if (mac_node->state == HCLGE_MAC_ACTIVE) {
+			mac_node->state = HCLGE_MAC_TO_ADD;
+		} else if (mac_node->state == HCLGE_MAC_TO_DEL) {
+			list_del(&mac_node->node);
+			kfree(mac_node);
 		}
 	}
 }
 
+void hclge_restore_mac_table_common(struct hclge_vport *vport)
+{
+	spin_lock_bh(&vport->mac_list_lock);
+
+	hclge_mac_node_convert_for_reset(&vport->uc_mac_list);
+	hclge_mac_node_convert_for_reset(&vport->mc_mac_list);
+	set_bit(HCLGE_VPORT_STATE_MAC_TBL_CHANGE, &vport->state);
+
+	spin_unlock_bh(&vport->mac_list_lock);
+}
+
+static void hclge_restore_hw_table(struct hclge_dev *hdev)
+{
+	struct hclge_vport *vport = &hdev->vport[0];
+	struct hnae3_handle *handle = &vport->nic;
+
+	hclge_restore_mac_table_common(vport);
+	hclge_restore_vport_vlan_table(vport);
+	set_bit(HCLGE_STATE_PROMISC_CHANGED, &hdev->state);
+
+	hclge_restore_fd_entries(handle);
+}
+
 int hclge_en_hw_strip_rxvtag(struct hnae3_handle *handle, bool enable)
 {
 	struct hclge_vport *vport = hclge_get_vport(handle);
@@ -9412,10 +9920,8 @@ static int hclge_init_ae_dev(struct hnae3_ae_dev *ae_dev)
 	int ret;
 
 	hdev = devm_kzalloc(&pdev->dev, sizeof(*hdev), GFP_KERNEL);
-	if (!hdev) {
-		ret = -ENOMEM;
-		goto out;
-	}
+	if (!hdev)
+		return -ENOMEM;
 
 	hdev->pdev = pdev;
 	hdev->ae_dev = ae_dev;
@@ -9594,6 +10100,7 @@ err_pci_uninit:
 	pci_release_regions(pdev);
 	pci_disable_device(pdev);
 out:
+	mutex_destroy(&hdev->vport_lock);
 	return ret;
 }
 
@@ -9658,7 +10165,7 @@ static int hclge_set_vf_spoofchk(struct hnae3_handle *handle, int vf,
 		dev_warn(&hdev->pdev->dev,
 			 "vf %d vlan table is full, enable spoof check may cause its packet send fail\n",
 			 vf);
-	else if (enable && hclge_is_umv_space_full(vport))
+	else if (enable && hclge_is_umv_space_full(vport, true))
 		dev_warn(&hdev->pdev->dev,
 			 "vf %d mac table is full, enable spoof check may cause its packet send fail\n",
 			 vf);
@@ -9835,8 +10342,16 @@ static int hclge_reset_ae_dev(struct hnae3_ae_dev *ae_dev)
 	set_bit(HCLGE_STATE_DOWN, &hdev->state);
 
 	hclge_stats_clear(hdev);
-	memset(hdev->vlan_table, 0, sizeof(hdev->vlan_table));
-	memset(hdev->vf_vlan_full, 0, sizeof(hdev->vf_vlan_full));
+	/* NOTE: pf reset needn't to clear or restore pf and vf table entry.
+	 * so here should not clean table in memory.
+	 */
+	if (hdev->reset_type == HNAE3_IMP_RESET ||
+	    hdev->reset_type == HNAE3_GLOBAL_RESET) {
+		memset(hdev->vlan_table, 0, sizeof(hdev->vlan_table));
+		memset(hdev->vf_vlan_full, 0, sizeof(hdev->vf_vlan_full));
+		bitmap_set(hdev->vport_config_block, 0, hdev->num_alloc_vport);
+		hclge_reset_umv_space(hdev);
+	}
 
 	ret = hclge_cmd_init(hdev);
 	if (ret) {
@@ -9850,8 +10365,6 @@ static int hclge_reset_ae_dev(struct hnae3_ae_dev *ae_dev)
 		return ret;
 	}
 
-	hclge_reset_umv_space(hdev);
-
 	ret = hclge_mac_init(hdev);
 	if (ret) {
 		dev_err(&pdev->dev, "Mac init error, ret = %d\n", ret);
@@ -9947,12 +10460,11 @@ static void hclge_uninit_ae_dev(struct hnae3_ae_dev *ae_dev)
 	hclge_clear_vf_vlan(hdev);
 	hclge_misc_affinity_teardown(hdev);
 	hclge_state_uninit(hdev);
+	hclge_uninit_mac_table(hdev);
 
 	if (mac->phydev)
 		mdiobus_unregister(mac->mdio_bus);
 
-	hclge_uninit_umv_space(hdev);
-
 	/* Disable MISC vector(vector0) */
 	hclge_enable_vector(&hdev->misc_vector, false);
 	synchronize_irq(hdev->misc_vector.vector_irq);
@@ -9966,7 +10478,6 @@ static void hclge_uninit_ae_dev(struct hnae3_ae_dev *ae_dev)
 	hclge_misc_irq_uninit(hdev);
 	hclge_pci_uninit(hdev);
 	mutex_destroy(&hdev->vport_lock);
-	hclge_uninit_vport_mac_table(hdev);
 	hclge_uninit_vport_vlan_table(hdev);
 	ae_dev->priv = NULL;
 }
@@ -10213,16 +10724,19 @@ static int hclge_get_64_bit_regs(struct hclge_dev *hdev, u32 regs_num,
 
 int hclge_query_bd_num_cmd_send(struct hclge_dev *hdev, struct hclge_desc *desc)
 {
-	/*prepare 4 commands to query DFX BD number*/
-	hclge_cmd_setup_basic_desc(&desc[0], HCLGE_OPC_DFX_BD_NUM, true);
-	desc[0].flag |= cpu_to_le16(HCLGE_CMD_FLAG_NEXT);
-	hclge_cmd_setup_basic_desc(&desc[1], HCLGE_OPC_DFX_BD_NUM, true);
-	desc[1].flag |= cpu_to_le16(HCLGE_CMD_FLAG_NEXT);
-	hclge_cmd_setup_basic_desc(&desc[2], HCLGE_OPC_DFX_BD_NUM, true);
-	desc[2].flag |= cpu_to_le16(HCLGE_CMD_FLAG_NEXT);
-	hclge_cmd_setup_basic_desc(&desc[3], HCLGE_OPC_DFX_BD_NUM, true);
+	int i;
 
-	return hclge_cmd_send(&hdev->hw, desc, 4);
+	/* initialize command BD except the last one */
+	for (i = 0; i < HCLGE_GET_DFX_REG_TYPE_CNT - 1; i++) {
+		hclge_cmd_setup_basic_desc(&desc[i], HCLGE_OPC_DFX_BD_NUM,
+					   true);
+		desc[i].flag |= cpu_to_le16(HCLGE_CMD_FLAG_NEXT);
+	}
+
+	/* initialize the last command BD */
+	hclge_cmd_setup_basic_desc(&desc[i], HCLGE_OPC_DFX_BD_NUM, true);
+
+	return hclge_cmd_send(&hdev->hw, desc, HCLGE_GET_DFX_REG_TYPE_CNT);
 }
 
 static int hclge_get_dfx_reg_bd_num(struct hclge_dev *hdev,
@@ -10576,6 +11090,131 @@ static int hclge_gro_en(struct hnae3_handle *handle, bool enable)
 	return hclge_config_gro(hdev, enable);
 }
 
+static void hclge_sync_promisc_mode(struct hclge_dev *hdev)
+{
+	struct hclge_vport *vport = &hdev->vport[0];
+	struct hnae3_handle *handle = &vport->nic;
+	u8 tmp_flags = 0;
+	int ret;
+
+	if (vport->last_promisc_flags != vport->overflow_promisc_flags) {
+		set_bit(HCLGE_STATE_PROMISC_CHANGED, &hdev->state);
+		vport->last_promisc_flags = vport->overflow_promisc_flags;
+	}
+
+	if (test_bit(HCLGE_STATE_PROMISC_CHANGED, &hdev->state)) {
+		tmp_flags = handle->netdev_flags | vport->last_promisc_flags;
+		ret = hclge_set_promisc_mode(handle, tmp_flags & HNAE3_UPE,
+					     tmp_flags & HNAE3_MPE);
+		if (!ret) {
+			clear_bit(HCLGE_STATE_PROMISC_CHANGED, &hdev->state);
+			hclge_enable_vlan_filter(handle,
+						 tmp_flags & HNAE3_VLAN_FLTR);
+		}
+	}
+}
+
+static bool hclge_module_existed(struct hclge_dev *hdev)
+{
+	struct hclge_desc desc;
+	u32 existed;
+	int ret;
+
+	hclge_cmd_setup_basic_desc(&desc, HCLGE_OPC_GET_SFP_EXIST, true);
+	ret = hclge_cmd_send(&hdev->hw, &desc, 1);
+	if (ret) {
+		dev_err(&hdev->pdev->dev,
+			"failed to get SFP exist state, ret = %d\n", ret);
+		return false;
+	}
+
+	existed = le32_to_cpu(desc.data[0]);
+
+	return existed != 0;
+}
+
+/* need 6 bds(total 140 bytes) in one reading
+ * return the number of bytes actually read, 0 means read failed.
+ */
+static u16 hclge_get_sfp_eeprom_info(struct hclge_dev *hdev, u32 offset,
+				     u32 len, u8 *data)
+{
+	struct hclge_desc desc[HCLGE_SFP_INFO_CMD_NUM];
+	struct hclge_sfp_info_bd0_cmd *sfp_info_bd0;
+	u16 read_len;
+	u16 copy_len;
+	int ret;
+	int i;
+
+	/* setup all 6 bds to read module eeprom info. */
+	for (i = 0; i < HCLGE_SFP_INFO_CMD_NUM; i++) {
+		hclge_cmd_setup_basic_desc(&desc[i], HCLGE_OPC_GET_SFP_EEPROM,
+					   true);
+
+		/* bd0~bd4 need next flag */
+		if (i < HCLGE_SFP_INFO_CMD_NUM - 1)
+			desc[i].flag |= cpu_to_le16(HCLGE_CMD_FLAG_NEXT);
+	}
+
+	/* setup bd0, this bd contains offset and read length. */
+	sfp_info_bd0 = (struct hclge_sfp_info_bd0_cmd *)desc[0].data;
+	sfp_info_bd0->offset = cpu_to_le16((u16)offset);
+	read_len = min_t(u16, len, HCLGE_SFP_INFO_MAX_LEN);
+	sfp_info_bd0->read_len = cpu_to_le16(read_len);
+
+	ret = hclge_cmd_send(&hdev->hw, desc, i);
+	if (ret) {
+		dev_err(&hdev->pdev->dev,
+			"failed to get SFP eeprom info, ret = %d\n", ret);
+		return 0;
+	}
+
+	/* copy sfp info from bd0 to out buffer. */
+	copy_len = min_t(u16, len, HCLGE_SFP_INFO_BD0_LEN);
+	memcpy(data, sfp_info_bd0->data, copy_len);
+	read_len = copy_len;
+
+	/* copy sfp info from bd1~bd5 to out buffer if needed. */
+	for (i = 1; i < HCLGE_SFP_INFO_CMD_NUM; i++) {
+		if (read_len >= len)
+			return read_len;
+
+		copy_len = min_t(u16, len - read_len, HCLGE_SFP_INFO_BDX_LEN);
+		memcpy(data + read_len, desc[i].data, copy_len);
+		read_len += copy_len;
+	}
+
+	return read_len;
+}
+
+static int hclge_get_module_eeprom(struct hnae3_handle *handle, u32 offset,
+				   u32 len, u8 *data)
+{
+	struct hclge_vport *vport = hclge_get_vport(handle);
+	struct hclge_dev *hdev = vport->back;
+	u32 read_len = 0;
+	u16 data_len;
+
+	if (hdev->hw.mac.media_type != HNAE3_MEDIA_TYPE_FIBER)
+		return -EOPNOTSUPP;
+
+	if (!hclge_module_existed(hdev))
+		return -ENXIO;
+
+	while (read_len < len) {
+		data_len = hclge_get_sfp_eeprom_info(hdev,
+						     offset + read_len,
+						     len - read_len,
+						     data + read_len);
+		if (!data_len)
+			return -EIO;
+
+		read_len += data_len;
+	}
+
+	return 0;
+}
+
 static const struct hnae3_ae_ops hclge_ops = {
 	.init_ae_dev = hclge_init_ae_dev,
 	.uninit_ae_dev = hclge_uninit_ae_dev,
@@ -10588,6 +11227,7 @@ static const struct hnae3_ae_ops hclge_ops = {
 	.get_vector = hclge_get_vector,
 	.put_vector = hclge_put_vector,
 	.set_promisc_mode = hclge_set_promisc_mode,
+	.request_update_promisc_mode = hclge_request_update_promisc_mode,
 	.set_loopback = hclge_set_loopback,
 	.start = hclge_ae_start,
 	.stop = hclge_ae_stop,
@@ -10649,7 +11289,6 @@ static const struct hnae3_ae_ops hclge_ops = {
 	.get_fd_rule_cnt = hclge_get_fd_rule_cnt,
 	.get_fd_rule_info = hclge_get_fd_rule_info,
 	.get_fd_all_rules = hclge_get_all_rules,
-	.restore_fd_rules = hclge_restore_fd_entries,
 	.enable_fd = hclge_enable_fd,
 	.add_arfs_entry = hclge_add_fd_entry_by_arfs,
 	.dbg_run_cmd = hclge_dbg_run_cmd,
@@ -10662,13 +11301,14 @@ static const struct hnae3_ae_ops hclge_ops = {
 	.set_timer_task = hclge_set_timer_task,
 	.mac_connect_phy = hclge_mac_connect_phy,
 	.mac_disconnect_phy = hclge_mac_disconnect_phy,
-	.restore_vlan_table = hclge_restore_vlan_table,
 	.get_vf_config = hclge_get_vf_config,
 	.set_vf_link_state = hclge_set_vf_link_state,
 	.set_vf_spoofchk = hclge_set_vf_spoofchk,
 	.set_vf_trust = hclge_set_vf_trust,
 	.set_vf_rate = hclge_set_vf_rate,
 	.set_vf_mac = hclge_set_vf_mac,
+	.get_module_eeprom = hclge_get_module_eeprom,
+	.get_cmdq_stat = hclge_get_cmdq_stat,
 };
 
 static struct hnae3_ae_algo ae_algo = {