Re: [PATCH v2] x86/microcode/intel: Refresh old_microcode defines with the May 2026 release

[Dave Hansen]

On 9/29/26 06:08, Andrew Cooper wrote:
> This a known-buggy SRF ucode.  The only reason there wasn't an explicit
> retraction of it was because a newer IPU had already been published,
> with 0x030003b2 being believed to be good again.

That's good to know, thanks.

But, for now, we've chosen to trail the ucode repo. On some level, the
kernel is always going to be "wrong" to some degree. In this case, the
cost of being "wrong" is that we're giving an answer to the old/new
question that's from May and not August or September.

The only real way to fix this is to be more aggressive about how far we
trail the official repo. Is this SRF issue a strong enough reason to
change that temporarily or permanently?