Hello, We recently tested QINQ offload for intel E810 nic. [Ethernet controller: Intel Corporation Ethernet Controller E810-C for SFP (rev 02) Device id 8086:1593] relative software versions as below: DPDK: 21.11.1 kernel driver: 1.7.16 OS default DDP: 1.3.27 Firmware: 3.1 We've tested testpmd and the offload setting as below: struct rte_eth_rxmode rx_mode = { .offloads = DEV_RX_OFFLOAD_VLAN_EXTEND | DEV_RX_OFFLOAD_VLAN_STRIP | DEV_RX_OFFLOAD_QINQ_STRIP | DEV_RX_OFFLOAD_RSS_HASH, }; struct rte_eth_txmode tx_mode = { .offloads = RTE_ETH_TX_OFFLOAD_MBUF_FAST_FREE | RTE_ETH_TX_OFFLOAD_VLAN_INSERT | RTE_ETH_TX_OFFLOAD_QINQ_INSERT, }; We use traffic gen some QINQ packet with double vlan tag 0x8100, on rx side: the outer vlan tag has been stripped, but the inner vlan tag not, mbuf->vlan_tci_outer is 0, ol_flags: RTE_MBUF_F_RX_VLAN RTE_MBUF_F_RX_L4_CKSUM_GOOD RTE_MBUF_F_RX_IP_CKSUM_GOOD RTE_MBUF_F_RX_VLAN_STRIPPED RTE_MBUF_F_RX_OUTER_L4_CKSUM_UNKNOWN on tx side: we modified testpmd code, force to insert double vlan, mbuf->vlan_tci_outer = 101; mbuf->vlan_tci = 201; the outer vlan tag (mbuf->vlan_tci_outer) lost, packet send out by testpmd only have one vlan tag (201), ol_flags: RTE_MBUF_F_TX_VLAN | RTE_MBUF_F_TX_QINQ I believe this is a bug at the NVM firmware level or DDP package but still want someone else to confirm. We are trying to reach out to Intel as well to report the issue and get additional assistance. If anyone can assist with that, it would be appreciated. 韩爽 鲲鹏网络平台部门 北京奇安信集团 Beijing Qi An Xin Group 手机 +86 18600206283 邮箱 hanshuang@qianxin.com