Intel Revs Its Linear Address Masking Patches For Linux

Added to Intel's documentation in late 2020 and initial kernel patches out since early 2021, Intel has been slowly working on Linear Address Masking (LAM) support for the Linux kernel. Out this past week was finally the latest iteration of this work for leveraging untranslated address bits of 64-bit linear addresses to be used for storing arbitrary software metadata.

Intel LAM v5 has been published as the newest version of these patches that with upcoming Intel CPUs allow making use of untranslated address bits of user-space addresses to be used for arbitrary metadata.

Intel documentation on LAM has been public since late 2020.

The v5 patches re-base against the latest Linux kernel state, add more code comments, added a self-test, and various low-level code tweaks for this CPU hardware feature. Enabling LAM for the Linux kernel is just over one thousand lines of new code.

We'll see if these v5 LAM patches are now good enough for mainlining or will continue slowly working its way to the mainline kernel.

Meanwhile coming out with AMD Zen 4 is a similar feature called Upper Address Ignore (UAI) for user-space address tagging using the unused bits of a memory address.