Arch Linux

Starship_Lizard

Member

Registered: 2023-08-09

Posts: 3

[SOLVED] Having some trouble with microcode

Hi everyone, brand new to using Arch and I'm liking it so far. I'm working on debugging some issues on my system and one of the steps is to make sure I have the latest microcode. I'm booting via EFISTUB and used efibootmgr to make the necessary boot config. I saw that to load the microcode for AMD CPUs you have to make sure the initrd options are set correctly. Here's what I ended up doing:

# efibootmgr --create --disk /dev/nvme0n1 --part 1 --label "Arch Linux" --loader /vmlinuz-linux --unicode 'root=nvme0n1p3 rw initrd=\amd-ucode.img initrd=\initramfs-linux.img'

(not before messing it up and ruining my boot the first time I tried this)

However I'm still getting this output:

$ sudo journalctl -k --grep=microcode
Aug 08 20:05:48 PowderKeg kernel: Zenbleed: please update your microcode for the most optimal fix
Aug 08 20:05:49 PowderKeg kernel: microcode: CPU2: patch_level=0x08701013
Aug 08 20:05:49 PowderKeg kernel: microcode: CPU0: patch_level=0x08701013
Aug 08 20:05:49 PowderKeg kernel: microcode: CPU1: patch_level=0x08701013
Aug 08 20:05:49 PowderKeg kernel: microcode: CPU3: patch_level=0x08701013
Aug 08 20:05:49 PowderKeg kernel: microcode: CPU4: patch_level=0x08701013
Aug 08 20:05:49 PowderKeg kernel: microcode: CPU6: patch_level=0x08701013
Aug 08 20:05:49 PowderKeg kernel: microcode: CPU5: patch_level=0x08701013
Aug 08 20:05:49 PowderKeg kernel: microcode: CPU8: patch_level=0x08701013
Aug 08 20:05:49 PowderKeg kernel: microcode: CPU9: patch_level=0x08701013
Aug 08 20:05:49 PowderKeg kernel: microcode: CPU10: patch_level=0x08701013
Aug 08 20:05:49 PowderKeg kernel: microcode: CPU7: patch_level=0x08701013
Aug 08 20:05:49 PowderKeg kernel: microcode: CPU11: patch_level=0x08701013
Aug 08 20:05:49 PowderKeg kernel: microcode: Microcode Update Driver: v2.2.

To me, that seems like the microcode didn't update, considering it's still warning me to update the microcode. I was having some trouble figuring out what the most recent version of the microcode even is for the CPU in question (Ryzen 5 3600x). I did poke around the forums and tried to find any other conversation about it, but I couldn't really find anything definitive.

Thanks for reading and I look forward to being a part of this community

(edited for clarity)

Last edited by Starship_Lizard (2023-08-09 04:24:25)

Scimmia

Fellow

Registered: 2012-09-01

Posts: 13,727

Re: [SOLVED] Having some trouble with microcode

Microcode fixes for Zenbleed are only out for EPYC processors. Everything else is scheduled to be in the next 2-4 months.

Last edited by Scimmia (2023-08-09 04:07:04)

Starship_Lizard

Member

Registered: 2023-08-09

Posts: 3

Re: [SOLVED] Having some trouble with microcode

Microcode fixes for Zenbleed are only out for EPYC processors. Everything else is scheduled to be in the next 2-4 months.

Okay, that's good to know! I'll be extra careful about any links I click on from now 'til then. Thanks for that.

Does my process for getting the latest microcode look correct otherwise you think? Do you know where I can check what the latest microcode version number I should have?

Scimmia

Fellow

Registered: 2012-09-01

Posts: 13,727

Re: [SOLVED] Having some trouble with microcode

Yes, your process looks fine. The issue is mitigated in the kernel, so there's no actual risk, just a potential performance hit.

As for what's included in the microcode archive, see upstream. https://git.kernel.org/pub/scm/linux/ke … ode/README
lscpu will show the info on family/model/stepping, you just have to convert it to hex.

Starship_Lizard

Member

Registered: 2023-08-09

Posts: 3

Re: [SOLVED] Having some trouble with microcode

Yes, your process looks fine. The issue is mitigated in the kernel, so there's no actual risk, just a potential performance hit.

As for what's included in the microcode archive, see upstream. https://git.kernel.org/pub/scm/linux/ke … ode/README
lscpu will show the info on family/model/stepping, you just have to convert it to hex.

Thanks so much for your help! I'll give that a gander. Also good to know that the risk is mitigated in the Kernel. I am a bit more at ease now.