On Wed, Sep 23, 2020 at 09:11:43AM -0400, Mikulas Patocka wrote: > I also don't know how to implement journling on persistent memory :) On > EXT4 or XFS you can pin dirty buffers in memory until the journal is > flushed. This is obviously impossible on persistent memory. So, I'm > considering implementing only some lightweight journaling that will > guarantee atomicity between just a few writes.
That's a bit disappointing considering people have been publishing papers on how to do umpteen different variations on persistent memory journalling for the last five years. https://www.google.com/search?q=intel+persistent+memory+atomic+updates for example
