The Linux 5.14 RAS (Reliability, Availability and Serviceability) and EDAC (Error Detection And Correction) adjustments enjoy landed with several enhancements this time round on the Intel aspect.
RAS/EDAC adjustments this time round consist of the reinforce for on-equipment excessive bandwidth memory (HBM) for Xeon Sapphire Rapids as previously covered on Phoronix and officially confirmed earlier this week by Intel that varied Sapphire Rapids SKUs will certainly enjoy HBM. The Xeon HBM bits listed below are around the EDAC reinforce for memory error checking/reporting.
This pull interrogate additionally entails including Ice Lake Neural Community Processor for Deep Studying Inference (ICL-NNPI) reinforce to the igen6_edac driver. Furthermore, Tiger Lake and Alder Lake are added as properly for supporting situations the place the CPUs reinforce in-band ECC. These SoCs including Alder Lake all enjoy the identical in-band ECC (IBECC) capabilities as the Elkhart Lake SoC nonetheless with the exception TGL / ADL now has two memory controllers in space of one.
Also on the Intel entrance is a fix for Xeon Ice Lake and Sapphire Rapids for reporting “shut to” and “some distance” devices for errors in 2LM configurations. There could be additionally a fix to no longer attempt to load the Intel EDAC drivers when working as a virtualized guest.
The finest AMD RAS/EDAC trade this cycle rate noting is officially marking AMD’s Yazen Ghannam as the maintainer for his or her EDAC drivers. Yazen has been working on the AMD RAS/EDAC code for years nonetheless factual lacked the formality / good recognition of being smartly is notion as the upstream driver maintainer.
As talked about the earlier day there are AMD EDAC preparations for AMD heterogeneous servers with Aldebaran multi-die GPUs nonetheless that work became published too slack that could no longer land now till Linux 5.15.
The checklist of RAS/EDAC patches for Linux 5.14 could presumably well be stumbled on from this PR.