It’s a bug in AMD processors that was mitigated in recent kernels by disabling rdseed32 CPU feature. Framework has nothing to do with it. Hopefully microcode update would eventually resolve it, if at all possible.
Microcode fully fixing the issue began to release in November. For the AI 300 series, that’s version StrixKrackanPI-FP8_1.1.0.0e released on Dec 15. It does not yet seem to be included in linux-firmware. Kernel 6.18 and later (and back-ports) correctly handle disabling RDSEED32 only when running unpatched microcode, so when the update ships from your OS or from FW in a UEFI update (the latest version UEFI 03.04 ships version StrixKrackanPI-FP8_1.1.0.0c) the kernel will handle allowing it to remain enabled again.
Fortunately you shouldn’t see too much direct impact since it’s mostly used by other libraries that handle the error safely already.
You can read the history of the incident in these articles:
As @Korvin said, “it is a bug … mitigated in recent kernels.”
Considering you updated your operating system, you probably got a more recent version of the kernel. The more recent version of the kernel mitigates the bug, and produces the message.
Just to be clear: the bug was always present. The old kernel just didn’t tell you about it.
Well according to some posts bug hasn’t always existed, some say it’s a bios issue and other say it’s a kernel issue.
So I’m confused, if it’s a kernel issue then I guess older kernel has a modification that disabled something so this message didn’t show. While newer kernel doesn’t have it? So then it shows.
Maybe not “always”, but it existed widespread in Zen 5 for a while at least (and some earlier versions of Zen as well actually).
Neither. It is a CPU microcode issue, and it affects all Zen 5 CPUs of the models mentioned in AMD’s buelltin. Later kernels have mitigations for the microcode issue, so you only see the warning now that you have updated to a kernel version that includes the mitigations.
It’s the other way around: older kernels did not have mitigations for the issue, so the issue still existed in the microcode, it was just not handled and reported by the kernel. Newer kernels mitigate and report the issue when they detect unpatched microcode. Once you get the updated microcode, you will no longer see the warning as the kernel will recognize that the issue is no longer present and does not need intervention or reporting.
Well it seems my zorin 18 pro did have something in older kernel since it had kernel 6.14 but was updated to 6.17 recently and now after another update this message appeared, it didn’t show on kernel 6.14 so it must of been a patch/modification in that kernel that disabled the message or something.