From 9a3284fad42f66bb43629c6716709ff791aaa457 Mon Sep 17 00:00:00 2001 From: Michal Koutný Date: Fri, 4 Oct 2019 12:57:40 +0200 Subject: cgroup: Optimize single thread migration MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit There are reports of users who use thread migrations between cgroups and they report performance drop after d59cfc09c32a ("sched, cgroup: replace signal_struct->group_rwsem with a global percpu_rwsem"). The effect is pronounced on machines with more CPUs. The migration is affected by forking noise happening in the background, after the mentioned commit a migrating thread must wait for all (forking) processes on the system, not only of its threadgroup. There are several places that need to synchronize with migration: a) do_exit, b) de_thread, c) copy_process, d) cgroup_update_dfl_csses, e) parallel migration (cgroup_{proc,thread}s_write). In the case of self-migrating thread, we relax the synchronization on cgroup_threadgroup_rwsem to avoid the cost of waiting. d) and e) are excluded with cgroup_mutex, c) does not matter in case of single thread migration and the executing thread cannot exec(2) or exit(2) while it is writing into cgroup.threads. In case of do_exit because of signal delivery, we either exit before the migration or finish the migration (of not yet PF_EXITING thread) and die afterwards. This patch handles only the case of self-migration by writing "0" into cgroup.threads. For simplicity, we always take cgroup_threadgroup_rwsem with numeric PIDs. This change improves migration dependent workload performance similar to per-signal_struct state. Signed-off-by: Michal Koutný Signed-off-by: Tejun Heo --- kernel/cgroup/cgroup-internal.h | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'kernel/cgroup/cgroup-internal.h') diff --git a/kernel/cgroup/cgroup-internal.h b/kernel/cgroup/cgroup-internal.h index 809e34a3c017..90d1710fef6c 100644 --- a/kernel/cgroup/cgroup-internal.h +++ b/kernel/cgroup/cgroup-internal.h @@ -231,9 +231,10 @@ int cgroup_migrate(struct task_struct *leader, bool threadgroup, int cgroup_attach_task(struct cgroup *dst_cgrp, struct task_struct *leader, bool threadgroup); -struct task_struct *cgroup_procs_write_start(char *buf, bool threadgroup) +struct task_struct *cgroup_procs_write_start(char *buf, bool threadgroup, + bool *locked) __acquires(&cgroup_threadgroup_rwsem); -void cgroup_procs_write_finish(struct task_struct *task) +void cgroup_procs_write_finish(struct task_struct *task, bool locked) __releases(&cgroup_threadgroup_rwsem); void cgroup_lock_and_drain_offline(struct cgroup *cgrp); -- cgit