On Thu, Mar 28, 2019 at 02:59:30PM -0700, Yang Shi wrote: > Yes, it still could fail. I can't tell which way is better for now. I > just thought scanning another round then migrating should be still > faster than swapping off the top of my head. I think it depends on the relative capacities between your primary and migration tiers and how it's used. Applications may allocate and pin directly out of pmem if they wish, so it's not a dedicated fallback memory space like swap.
