Unchecked MSR access error, when applying the update

[ac130kz]

Do I need to enable msr with msr.allow_writes=on?

Config:

  services.ucodenix = {
    enable = true;
    cpuModelId = "00A60F12"; # 7800x3d
  };

Logs:

[    0.000000] unchecked MSR access error: WRMSR to 0xc0010020 (tried to write 0xffff8880732380c8) at rIP: 0xffffffffa206e59a (__apply_microcode_amd+0x1a/0x80)
[    0.000000] Call Trace:
[    0.000000]  <TASK>
[    0.000000]  ? ex_handler_msr.isra.0.cold+0x5b/0x60
[    0.000000]  ? fixup_exception+0x2a0/0x300
[    0.000000]  ? early_fixup_exception+0x45/0xe0
[    0.000000]  ? early_idt_handler_common+0x2f/0x3a
[    0.000000]  ? __apply_microcode_amd+0x1a/0x80
[    0.000000]  ? load_ucode_amd_bsp+0x89/0xd0
[    0.000000]  ? x86_64_start_kernel+0x80/0xa0
[    0.000000]  ? common_startup_64+0x13e/0x141
[    0.000000]  </TASK>

[    0.002346] microcode: CPU1: update failed for patch_level=0x0a60120c
[    0.002346] microcode: CPU4: update failed for patch_level=0x0a60120c
[    0.002346] microcode: CPU2: update failed for patch_level=0x0a60120c
[    0.002346] microcode: CPU6: update failed for patch_level=0x0a60120c
[    0.002346] microcode: CPU3: update failed for patch_level=0x0a60120c
[    0.002346] microcode: CPU5: update failed for patch_level=0x0a60120c
[    0.002346] microcode: CPU7: update failed for patch_level=0x0a60120c
[    0.002346] microcode: CPU8: update failed for patch_level=0x0a60120c
[    0.002346] microcode: CPU9: update failed for patch_level=0x0a60120c
[    0.002346] microcode: CPU10: update failed for patch_level=0x0a60120c
[    0.002346] microcode: CPU11: update failed for patch_level=0x0a60120c
[    0.002346] microcode: CPU12: update failed for patch_level=0x0a60120c
[    0.002346] microcode: CPU13: update failed for patch_level=0x0a60120c
[    0.002346] microcode: CPU14: update failed for patch_level=0x0a60120c
[    0.002346] microcode: CPU15: update failed for patch_level=0x0a60120c
[    0.607011] microcode: Current revision: 0x0a601209

[mightyiam]

Same CPU. Same logs. I'm not sure what MSR is, but the following didn't seem to prevent those log lines:

boot.kernelParams = [ "msr.allow_writes=on" ];

[e-tho]

I don’t have access to a CPU from this family for investigation. Could you please determine whether the issue is caused by a kernel update or a microcode update? You can set the microcode repository input to a previous revision to verify.

[mightyiam]

github:platomav/CPUMicrocodes/ec5200961ecdf78cf00e55d73902683e835edefd

With above, no error.

Only change is allowing it to bump up to c0a15766a4bc8c4697da34f9c8972d5db94dca91 and I get the error.

[e-tho]

The issue might stem from an incorrect microcode binary format, which would suggest a problem at the source, or an unexpected format update that would require an adjustment to amd-ucodegen. Without access to a CPU from this family, I’m unable to reproduce or diagnose the issue myself. I’d appreciate it if someone with the relevant hardware could investigate further and share their findings.

[mightyiam]

Well, I can share my screen and you'd tell me exactly what to do, if you're interested, @e-tho.

[e-tho]

@mightyiam Try reaching out to the author of the CPUMicrocodes repository to find out if there have been any changes regarding the format of the 00A60F12 microcode binary in the latest update, mentioning this issue.

[mightyiam]

@e-tho I received a response there that there were no changes to the format. Is there anything I can do towards gaining a better understanding of this error?