On Wed, 2019-08-21 at 07:45 -0700, [email protected] wrote: > Btw, for the next version it also might make sense to do one > optimization at a time. E.g. the empty cpumask one as the > first patch, the local cpu directly one next, and the threshold > based full flush as a third.
ok sure. I will refactor my optimization patch and remove the base patch(moving the functions from header to tlbflush.c) as you have already sent it out. -- Regards, Atish
