A few days back we reported on a Meta engineer [1]uncovering an architectural issue with RDSEED usage on AMD EPYC 9005 "Turin" CPUs . It ended up being found to affect more CPU models than originally anticipated and a new patch posted to the Linux kernel mailing list would disable RDSEED usage across all AMD Zen 5 processors.

Gregory Price of Meta posted a new patch to the LKML on Friday night that would disable RDSEED now for all AMD Zen 5 processors due to the problematic behavior.

Price commented on the [2]updated patch :

"Under unknown conditions, Zen5 chips running rdseed can produce (val=0,CF=1) over 10% of the time (when rdseed is successful). CF=1 indicates success, while val=0 is typically only produced when rdseed fails (CF=0).

This suggests there is a bug which causes rdseed to silently fail.

This was reproduced reliably by launching 2-threads per available core, 1-thread per for hamming on RDSEED, and 1-thread per core collectively eating and hammering on ~90% of memory.

This was observed on more than 1 Zen5 model, so it should be disabled for all of Zen5 until/unless a comprehensive blacklist can be built."

AMD Linux engineer Borislav Petkov responded that they are looking into it at the company. In an earlier message it sounds like they will be working on at least sorting out a proper blacklist/denylist for affected CPU models rather than blanket disabling RDSEED for all Zen 5 CPUs. We'll see what ultimately is decided, stay tuned.



[1] https://www.phoronix.com/news/AMD-EPYC-Turin-RDSEED-Bug

[2] https://lore.kernel.org/lkml/20251018024010.4112396-1-gourry@gourry.net/