On Sunday, 3 March 2024 19:14:23 GMT Daniel Frey wrote:
> Hi all,
> 
> I've always had problems updating the microcode for my AMD processor. I
> have various other Intel-based PCs and this has never been an issue.
> 
> I have confirmed it's not updating:
> 
> 
> ~ # dmesg | grep -i microcode
> [    0.201619] Zenbleed: please update your microcode for the most
> optimal fix
> [    0.748482] microcode: CPU1: patch_level=0x08701030
> [    0.748482] microcode: CPU0: patch_level=0x08701030
> [    0.748484] microcode: CPU3: patch_level=0x08701030
> [    0.748485] microcode: CPU5: patch_level=0x08701030
> [    0.748485] microcode: CPU4: patch_level=0x08701030
> [    0.748486] microcode: CPU6: patch_level=0x08701030
> [    0.748486] microcode: CPU7: patch_level=0x08701030
> [    0.748487] microcode: CPU8: patch_level=0x08701030
> [    0.748488] microcode: CPU9: patch_level=0x08701030
> [    0.748488] microcode: CPU10: patch_level=0x08701030
> [    0.748488] microcode: CPU11: patch_level=0x08701030
> [    0.748491] microcode: CPU12: patch_level=0x08701030
> [    0.748491] microcode: CPU13: patch_level=0x08701030
> [    0.748492] microcode: CPU14: patch_level=0x08701030
> [    0.748493] microcode: CPU15: patch_level=0x08701030
> [    0.748496] microcode: CPU17: patch_level=0x08701030
> [    0.748496] microcode: CPU18: patch_level=0x08701030
> [    0.748498] microcode: CPU19: patch_level=0x08701030
> [    0.748498] microcode: CPU20: patch_level=0x08701030
> [    0.748500] microcode: CPU21: patch_level=0x08701030
> [    0.748500] microcode: CPU22: patch_level=0x08701030
> [    0.748501] microcode: CPU24: patch_level=0x08701030
> [    0.748501] microcode: CPU23: patch_level=0x08701030
> [    0.748503] microcode: CPU16: patch_level=0x08701030
> [    0.748503] microcode: CPU26: patch_level=0x08701030
> [    0.748503] microcode: CPU27: patch_level=0x08701030
> [    0.748505] microcode: CPU28: patch_level=0x08701030
> [    0.748506] microcode: CPU29: patch_level=0x08701030
> [    0.748507] microcode: CPU30: patch_level=0x08701030
> [    0.748508] microcode: CPU25: patch_level=0x08701030
> [    0.748509] microcode: CPU31: patch_level=0x08701030
> [    0.748511] microcode: CPU2: patch_level=0x08701030
> [    0.748554] microcode: Microcode Update Driver: v2.2.
> 
> I'm pretty sure I wouldn't be getting a zenbleed warning if it was using
> the most recent microcode.
> 
> My processor is this one:
> 
> vendor_id       : AuthenticAMD
> cpu family      : 23
> model           : 113
> model name      : AMD Ryzen 9 3950X 16-Core Processor
> 
> This leads me to the 17h family.
> 
> I do not use an initramfs as my system doesn't require one. I am not
> willing to try an initramfs as my system fully functions without one and
> this is not an issue with the Intel machines I have.
> 
> I have properly configured the kernel (gentoo-sources-6.6.13):
> 
> CONFIG_CPU_SUP_AMD=y
> CONFIG_EXTRA_FIRMWARE="brcm/BCM20702B0-19ff-0239.hcd
> amd-ucode/microcode_amd_fam17h.bin"
> CONFIG_EXTRA_FIRMWARE_DIR="/lib/firmware"
> 
> The firmware loading is working as it does load the firmware for my
> bluetooth adapter with no issues.
> 
> (In the newer kernels microcode loading is enabled by default - no way
> to turn it off. All you have to do is select CPU_SUP_AMD apparently. It
> works on Intel machines.)
> 
> I've even updated the motherboard BIOS firmware, and while that fixed
> all the other issues it apparently does not have patches for zenbleed.
> 
> Does anyone have any idea why this will not update?
> 
> -Dan

It could be AMD have not yet released microcode updates for the community.  
OEMs receive new microcode first and patch it in their MoBo BIOS/UEFI 
firmware.  Eventually the CPU manufacturers release microcode for older CPUs 
no longer supported by OEMs.  Since you have embedded 'amd-ucode/
microcode_amd_fam17h.bin' in your kernel I don't think there's anything else 
you can do at this point in time, beyond emerging the latest sys-kernel/linux-
firmware and rebooting.

PS.  I always place the microcode string first in the CONFIG_EXTRA_FIRMWARE= 
entries, since it should be the fist thing to load by the CPU.  I don't know 
if it would makes any difference, since the whole string of firmwares will be 
parsed in one go.

Attachment: signature.asc
Description: This is a digitally signed message part.

Reply via email to