Intel Releases Updated CPU Microcode For Xeon 6 SoCs "Granite Rapids D"
Catching me by surprise today was a new Intel CPU microcode drop "20260227" for Linux users/administrators outside of their typical Patch Tuesday alignment for CPU microcode releases.
Today's Intel CPU Microcode 20260227 update is just for addressing functional issues affecting Xeon 6 "Granite Rapids D" SoCs. Only the Xeon 6700P-B/6500P-B Series SoC with P-Cores (Granite Rapids D) processors are affected with no other CPU cores/families seeing firmware changes with today's releases.
In going off the Xeon 6 SoC specification updates with the errata most recently published in January, it appears that today's "functional issues" corrected by the new microcode include some or all of:
The GitHub release announcement doesn't elaborate on the "functional issues" while the current Intel documentation notes no workarounds for the most recent errata. Unfortunately I don't have any Granite Rapids D hardware so unable to confirm any of the changes myself or if there is any performance impact from the updated microcode.
Intel Granite Rapids D customers on Linux wanting the new microcode can find the files on GitHub.
Today's Intel CPU Microcode 20260227 update is just for addressing functional issues affecting Xeon 6 "Granite Rapids D" SoCs. Only the Xeon 6700P-B/6500P-B Series SoC with P-Cores (Granite Rapids D) processors are affected with no other CPU cores/families seeing firmware changes with today's releases.
In going off the Xeon 6 SoC specification updates with the errata most recently published in January, it appears that today's "functional issues" corrected by the new microcode include some or all of:
GNRD40. Remapping Hardware with Major Version Number 6 Incorrectly Advertises the ESRTPS Support
GNRD41. Reserved(0) Check For a PASID Table Entry May Not Happen For a DMA Request
GNRD42. Remapping Hardware Will Not Report The PASID Value For RTA.2 Faults in Modes Other Than Scalable Mode
GNRD43. Remapping Hardware Does Not Perform a Reserved(0) Check in Interrupt Remap Table Entry
The GitHub release announcement doesn't elaborate on the "functional issues" while the current Intel documentation notes no workarounds for the most recent errata. Unfortunately I don't have any Granite Rapids D hardware so unable to confirm any of the changes myself or if there is any performance impact from the updated microcode.
Intel Granite Rapids D customers on Linux wanting the new microcode can find the files on GitHub.