]> git.openfabrics.org - ~shefty/rdma-dev.git/commit
percpu-rw-semaphores: use rcu_read_lock_sched
authorMikulas Patocka <mpatocka@redhat.com>
Mon, 22 Oct 2012 23:39:16 +0000 (19:39 -0400)
committerLinus Torvalds <torvalds@linux-foundation.org>
Sun, 28 Oct 2012 17:59:36 +0000 (10:59 -0700)
commit1bf11c53535ab87e3bf14ecdf6747bf46f601c5d
tree9170fa956be0f564b988a88685eb82584347c2cf
parent5c1eabe68501d1e1b1586c7f4c46cc531828c4ab
percpu-rw-semaphores: use rcu_read_lock_sched

Use rcu_read_lock_sched / rcu_read_unlock_sched / synchronize_sched
instead of rcu_read_lock / rcu_read_unlock / synchronize_rcu.

This is an optimization. The RCU-protected region is very small, so
there will be no latency problems if we disable preempt in this region.

So we use rcu_read_lock_sched / rcu_read_unlock_sched that translates
to preempt_disable / preempt_disable. It is smaller (and supposedly
faster) than preemptible rcu_read_lock / rcu_read_unlock.

Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
include/linux/percpu-rwsem.h