On Sun, 24 Jan 2021 at 16:11, Hillf Danton <hdan...@sina.com> wrote: > > If it is supposed due to the race between pool->stale_lock and > pool->lock that are both protecting the buddy list_head then adding > another one can be a cure. The diff below is not for any test.
Thanks for looking at the issue. Why the proposed patch not intended for testing? It is not the final (optimal) variant? -- Best Regards, Mike Gavrilov.
