mlx5_core output concerns: Error cqe on cqn 0x##, ci 0x##, qn 0x##, opcode 0x##, syndrome 0x##, vendor syndrome 0x##

Issue

• The following are the mlx5_core kernel log messages of concern:

mlx5_core 0000:ae:00.0 mlx5_ib2: Error cqe on cqn 0x4bb, ci 0xec, qn 0x11d4, opcode 0xd, syndrome 0x2, vendor syndrome 0x68
00000000: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00000010: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00000020: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00000030: 00 00 00 00 67 00 68 02 0a 00 11 d4 64 0c 18 d2
00000000: 00 64 0c 0a 00 11 d4 08 00 00 00 08 00 00 00 00
00000010: 00 00 00 00 00 00 00 00 80 00 00 00 60 00 04 04
00000030: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00000040: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00000050: 00 00 00 00 c0 00 00 00 00 00 00 00 00 12 08 00
00000060: 00 00 45 00 00 8c 3e 4a 40 00 40 11 a3 bb ac 10
00000070: 00 00 00 7e 00 18 19 00 00 00 00 25 23 03 ec 2e
mlx5_core 0000:ae:00.0 mlx5_ib2: ERR CQE on SQ: 0x11d4

• These messages seem to notify that the mlx5_core driver detected that an invalid opcode (0xd: MLX5_CQE_REQ_ERR) is set in the completion queue entry (CQE) that is connected to the completion queue (CQ) after transmission. However, the opcode does not seem to be set in the driver but is set from the IB-HCA hardware information. We assume from the Connect-X4 Adapters Programmer’s Reference
  Manual (PRM) that this is the event described in 7.12.7 Completion With Error, but as we cannot refer to the Connect-X5/X6 data sheet, but is difficult to investigate further.
• What might have caused these messages to appear?
• Do these messages mean that the system is being affected?
• What should be done when these messages appear? Can they be ignored?