this_cpu_cmpxchg was modified to atomic version, which can be more costly than non-atomic version. Switch users of this_cpu_cmpxchg to this_cpu_cmpxchg_local (which preserves pre-non-atomic this_cpu_cmpxchg behaviour). Acked-by: Peter Xu <peterx@xxxxxxxxxx> Signed-off-by: Marcelo Tosatti <mtosatti@xxxxxxxxxx> Index: linux-vmstat-remote/kernel/fork.c =================================================================== --- linux-vmstat-remote.orig/kernel/fork.c +++ linux-vmstat-remote/kernel/fork.c @@ -203,7 +203,7 @@ static bool try_release_thread_stack_to_ unsigned int i; for (i = 0; i < NR_CACHED_STACKS; i++) { - if (this_cpu_cmpxchg(cached_stacks[i], NULL, vm) != NULL) + if (this_cpu_cmpxchg_local(cached_stacks[i], NULL, vm) != NULL) continue; return true; } Index: linux-vmstat-remote/kernel/scs.c =================================================================== --- linux-vmstat-remote.orig/kernel/scs.c +++ linux-vmstat-remote/kernel/scs.c @@ -83,7 +83,7 @@ void scs_free(void *s) */ for (i = 0; i < NR_CACHED_SCS; i++) - if (this_cpu_cmpxchg(scs_cache[i], 0, s) == NULL) + if (this_cpu_cmpxchg_local(scs_cache[i], 0, s) == NULL) return; kasan_unpoison_vmalloc(s, SCS_SIZE, KASAN_VMALLOC_PROT_NORMAL);