Intel LAM Patches Updated After Its Rejection From Linux 6.2

Intel engineers had submitted support for Linear Address Masking (LAM) with the recently-closed Linux 6.2 merge window but it was rejected by Linus Torvalds. In working toward re-submitted it for the v6.3 cycle or later, an updated Linux LAM patch series was posted today.

Linus Torvalds rejected the Intel LAM patches from Linux 6.2 on the basis of changes to the untagged_addr() function believed to be "fundamentally broken" and making invalid assumptions. He also wasn't fond of the LAM name when Arm already provides this functionality as TBI, but that's something the Intel engineers would have a tougher time changing at Intel.

Intel LAM is about using the untranslated address bits of user-space addresses so it can be used for metadata. The LAM metadata can be used for purposes such like user-space memory sanitizers and tagging and is similar in nature to AMD's Upper Address Ignore "UAI" with Zen 4 and Arm's Top Byte Ignore "TBI" features.

In any event this morning marked the v13 patches for Linear Address Masking being sent out for review. With this v13 series is a fix for a race between the untagged_addr() and LAM enabling as part of Linus Torvalds' criticism.

With the v13 changes it does not allow for enabling LAM after the process spawned a second thread and additional changes to the untagged address functionality. See the v13 series for more details on the changes.

At least with Intel engineers being quick to iterate on their Linux patches, it gives hope this will be all squared away for the Linux 6.3 cycle that should kick off in February.