"FAIL" (Error Code -8001) Is Reported During Stream Test or Diagnosis When the Driver Version Is 25.0.RC1 or Later

Symptom

[****@******]# ascend-dmi --dg -i prbs -q
Summary:
    Arch: aarch64
    Mode: *****
    Time: 20250414-15:12:36

Hardware:
    prbs:
        FAIL
        *** error occured on device 1
        *** error occured on device 7

The following error information is displayed in the log:

[WARN][2025-04-14 15:11:03][72878][prbs_thread.cpp ControlTransPrbsCodeStream:86] call dcmi_prbs_operate failed. type=8, direction=1, cardId=7, chipId=0, ret=-8001
[WARN][2025-04-14 15:11:03][72878][prbs_thread.cpp ExecPrbsCodeStream:229] start prbs trans failed on device 7

Possible Causes

During the stream test or diagnosis, the SNR of the NPU or CDR fails to be verified.

Solution

Method 1: Run the msnpureport -f command to collect device logs. If the logs contain "retimer check snr failed.", the SNR verification of the CDR fails. If the logs contain "SNR is below 40W", the SNR verification of the NPU fails.

Method 2: Run the hccn_tool -i ${x} -sdcr -t 5 command to query the CDR SNR and historical SNR information. If the SNR is less than 350000/20dB, the SNR verification fails.

If the SNR does not meet the requirement, replace the corresponding hardware.