[PATCH] cgroup: avoid synchronize_sched() in __cgroup_procs_write()

[Oleg Nesterov]

From: Peter Zijlstra <peterz@xxxxxxxxxxxxx>

The current percpu-rwsem read side is entirely free of serializing insns
at the cost of having a synchronize_sched() in the write path.

The latency of the synchronize_sched() is too high for cgroups. The
commit 1ed1328792ff talks about the write path being a fairly cold path
but this is not the case for Android which moves task to the foreground
cgroup and back around binder IPC calls from foreground processes to
background processes, so it is significantly hotter than human initiated
operations.

Switch cgroup_threadgroup_rwsem into the slow mode for now to avoid the
problem, hopefully it should not be that slow after another commit
80127a39681b ("locking/percpu-rwsem: Optimize readers and reduce global
impact").

We could just add rcu_sync_enter() into cgroup_init() but we do not want
another synchronize_sched() at boot time, so this patch adds the new helper
which doesn't block but currently can only be called before the first use.

Cc: Tejun Heo <tj@xxxxxxxxxx>
Cc: Paul McKenney <paulmck@xxxxxxxxxxxxxxxxxx>
Reported-by: John Stultz <john.stultz@xxxxxxxxxx>
Reported-by: Dmitry Shmidt <dimitrysh@xxxxxxxxxx>
Signed-off-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx>
Signed-off-by: Oleg Nesterov <oleg@xxxxxxxxxx>
---
 include/linux/rcu_sync.h |  1 +
 kernel/cgroup.c          |  6 ++++++
 kernel/rcu/sync.c        | 12 ++++++++++++
 3 files changed, 19 insertions(+)

diff --git a/include/linux/rcu_sync.h b/include/linux/rcu_sync.h
index a63a33e..ece7ed9 100644
--- a/include/linux/rcu_sync.h
+++ b/include/linux/rcu_sync.h
@@ -59,6 +59,7 @@ static inline bool rcu_sync_is_idle(struct rcu_sync *rsp)
 }
 
 extern void rcu_sync_init(struct rcu_sync *, enum rcu_sync_type);
+extern void rcu_sync_enter_start(struct rcu_sync *);
 extern void rcu_sync_enter(struct rcu_sync *);
 extern void rcu_sync_exit(struct rcu_sync *);
 extern void rcu_sync_dtor(struct rcu_sync *);
diff --git a/kernel/cgroup.c b/kernel/cgroup.c
index 75c0ff0..8b79751 100644
--- a/kernel/cgroup.c
+++ b/kernel/cgroup.c
@@ -5609,6 +5609,12 @@ int __init cgroup_init(void)
 	BUG_ON(cgroup_init_cftypes(NULL, cgroup_dfl_base_files));
 	BUG_ON(cgroup_init_cftypes(NULL, cgroup_legacy_base_files));
 
+	/*
+	 * The latency of the synchronize_sched() is too high for cgroups,
+	 * avoid it at the cost of forcing all readers into the slow path.
+	 */
+	rcu_sync_enter_start(&cgroup_threadgroup_rwsem.rss);
+
 	get_user_ns(init_cgroup_ns.user_ns);
 
 	mutex_lock(&cgroup_mutex);
diff --git a/kernel/rcu/sync.c b/kernel/rcu/sync.c
index be922c9..e358313 100644
--- a/kernel/rcu/sync.c
+++ b/kernel/rcu/sync.c
@@ -83,6 +83,18 @@ void rcu_sync_init(struct rcu_sync *rsp, enum rcu_sync_type type)
 }
 
 /**
+ * Must be called after rcu_sync_init() and before first use.
+ *
+ * Ensures rcu_sync_is_idle() returns false and rcu_sync_{enter,exit}()
+ * pairs turn into NO-OPs.
+ */
+void rcu_sync_enter_start(struct rcu_sync *rsp)
+{
+	rsp->gp_count++;
+	rsp->gp_state = GP_PASSED;
+}
+
+/**
  * rcu_sync_enter() - Force readers onto slowpath
  * @rsp: Pointer to rcu_sync structure to use for synchronization
  *
-- 
2.5.0