Commit a4244454 authored by Tejun Heo's avatar Tejun Heo

percpu-refcount: use RCU-sched insted of normal RCU

percpu-refcount was incorrectly using preempt_disable/enable() for RCU
critical sections against call_rcu().  6a24474d ("percpu-refcount:
consistently use plain (non-sched) RCU") fixed it by converting the
preepmtion operations with rcu_read_[un]lock() citing that there isn't
any advantage in using sched-RCU over using the usual one; however,
rcu_read_[un]lock() for the preemptible RCU implementation -
CONFIG_TREE_PREEMPT_RCU, chosen when CONFIG_PREEMPT - are slightly
more expensive than preempt_disable/enable().

In a contrived microbench which repeats the followings,

 - percpu_ref_get()
 - copy 32 bytes of data into percpu buffer
 - percpu_put_get()
 - copy 32 bytes of data into percpu buffer

rcu_read_[un]lock() used in percpu_ref_get/put() makes it go slower by
about 15% when compared to using sched-RCU.

As the RCU critical sections are extremely short, using sched-RCU
shouldn't have any latency implications.  Convert to RCU-sched.
Signed-off-by: default avatarTejun Heo <tj@kernel.org>
Acked-by: default avatarKent Overstreet <koverstreet@google.com>
Acked-by: default avatar"Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Rusty Russell <rusty@rustcorp.com.au>
parent dbece3a0
...@@ -105,7 +105,7 @@ static inline void percpu_ref_get(struct percpu_ref *ref) ...@@ -105,7 +105,7 @@ static inline void percpu_ref_get(struct percpu_ref *ref)
{ {
unsigned __percpu *pcpu_count; unsigned __percpu *pcpu_count;
rcu_read_lock(); rcu_read_lock_sched();
pcpu_count = ACCESS_ONCE(ref->pcpu_count); pcpu_count = ACCESS_ONCE(ref->pcpu_count);
...@@ -114,7 +114,7 @@ static inline void percpu_ref_get(struct percpu_ref *ref) ...@@ -114,7 +114,7 @@ static inline void percpu_ref_get(struct percpu_ref *ref)
else else
atomic_inc(&ref->count); atomic_inc(&ref->count);
rcu_read_unlock(); rcu_read_unlock_sched();
} }
/** /**
...@@ -134,7 +134,7 @@ static inline bool percpu_ref_tryget(struct percpu_ref *ref) ...@@ -134,7 +134,7 @@ static inline bool percpu_ref_tryget(struct percpu_ref *ref)
unsigned __percpu *pcpu_count; unsigned __percpu *pcpu_count;
int ret = false; int ret = false;
rcu_read_lock(); rcu_read_lock_sched();
pcpu_count = ACCESS_ONCE(ref->pcpu_count); pcpu_count = ACCESS_ONCE(ref->pcpu_count);
...@@ -143,7 +143,7 @@ static inline bool percpu_ref_tryget(struct percpu_ref *ref) ...@@ -143,7 +143,7 @@ static inline bool percpu_ref_tryget(struct percpu_ref *ref)
ret = true; ret = true;
} }
rcu_read_unlock(); rcu_read_unlock_sched();
return ret; return ret;
} }
...@@ -159,7 +159,7 @@ static inline void percpu_ref_put(struct percpu_ref *ref) ...@@ -159,7 +159,7 @@ static inline void percpu_ref_put(struct percpu_ref *ref)
{ {
unsigned __percpu *pcpu_count; unsigned __percpu *pcpu_count;
rcu_read_lock(); rcu_read_lock_sched();
pcpu_count = ACCESS_ONCE(ref->pcpu_count); pcpu_count = ACCESS_ONCE(ref->pcpu_count);
...@@ -168,7 +168,7 @@ static inline void percpu_ref_put(struct percpu_ref *ref) ...@@ -168,7 +168,7 @@ static inline void percpu_ref_put(struct percpu_ref *ref)
else if (unlikely(atomic_dec_and_test(&ref->count))) else if (unlikely(atomic_dec_and_test(&ref->count)))
ref->release(ref); ref->release(ref);
rcu_read_unlock(); rcu_read_unlock_sched();
} }
#endif #endif
...@@ -154,5 +154,5 @@ void percpu_ref_kill_and_confirm(struct percpu_ref *ref, ...@@ -154,5 +154,5 @@ void percpu_ref_kill_and_confirm(struct percpu_ref *ref,
(((unsigned long) ref->pcpu_count)|PCPU_REF_DEAD); (((unsigned long) ref->pcpu_count)|PCPU_REF_DEAD);
ref->confirm_kill = confirm_kill; ref->confirm_kill = confirm_kill;
call_rcu(&ref->rcu, percpu_ref_kill_rcu); call_rcu_sched(&ref->rcu, percpu_ref_kill_rcu);
} }
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment