http://bugs.dpdk.org/show_bug.cgi?id=1868

            Bug ID: 1868
           Summary: bnxt: RTE_MBUF_F_RX_VLAN_STRIPPED set but vlan_tci is
                    0
           Product: DPDK
           Version: 26.03
          Hardware: All
                OS: All
            Status: UNCONFIRMED
          Severity: normal
          Priority: Normal
         Component: ethdev
          Assignee: [email protected]
          Reporter: [email protected]
  Target Milestone: ---

BCM957608-P2100GQF20 (235.1.157.0) with TruFlow disabled (via niccli) and
RTE_ETH_RX_OFFLOAD_VLAN_STRIP enabled doesn't set vlan_tci - it always remains
0, even though the VLAN itself is removed from the packet, and
RTE_MBUF_F_TX_VLAN and RTE_MBUF_F_RX_VLAN_STRIPPED flags are set.


- With testpmd, vlan_tci is 0 as well, type is 0x8000, so vlan is not there
anymore.

  src=00:C4:7A:4C:B0:F8 - dst=00:BE:BE:BE:BE:BE - pool=mb_pool_1 - type=0x0800
- length=70 - nb_segs=1 - RSS hash=0xd65b3fd5 - RSS queue=0x5 - VLAN tci=0x0 -
hw ptype: L2_ETHER_VLAN L3_IPV4_EXT_UNKNOWN L4_UDP  - sw ptype: L2_ETHER
L3_IPV4 L4_UDP  - l2_len=14 - l3_len=20 - l4_len=8 - Destination UDP port=53 -
Receive queue=0x5
  ol_flags: RTE_MBUF_F_RX_VLAN RTE_MBUF_F_RX_RSS_HASH
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

- If I do "port config 0 rx_offload vlan_strip off" RTE_MBUF_F_RX_VLAN_STRIPPED
added anyway, even though the type is 0x8100, so vlan wasn't stripped.

  src=00:C4:7A:4C:B0:F8 - dst=00:BE:BE:BE:BE:BE - pool=mb_pool_1 - type=0x8100
- length=1514 - nb_segs=1 - RSS hash=0xe3688ad3 - RSS queue=0x3 - VLAN tci=0x0
- hw ptype: L2_ETHER_VLAN L3_IPV4_EXT_UNKNOWN L4_UDP  - sw ptype: L2_ETHER_VLAN
L3_IPV4 L4_UDP  - l2_len=18 - l3_len=20 - l4_len=8 - Receive queue=0x3
  ol_flags: RTE_MBUF_F_RX_VLAN RTE_MBUF_F_RX_RSS_HASH
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

-- 
You are receiving this mail because:
You are the assignee for the bug.

Reply via email to