Revert "sched, cgroup: replace signal_struct->group_rwsem with a global percpu_rwsem"
authorTejun Heo <tj@kernel.org>
Wed, 16 Sep 2015 15:51:12 +0000 (11:51 -0400)
committerTejun Heo <tj@kernel.org>
Wed, 16 Sep 2015 15:51:12 +0000 (11:51 -0400)
commit0c986253b939cc14c69d4adbe2b4121bdf4aa220
tree78c28e9ff07b67baf4c06f72f92bbaa586452c9a
parentf9f9e7b776142fb1c0782cade004cc8e0147a199
Revert "sched, cgroup: replace signal_struct->group_rwsem with a global percpu_rwsem"

This reverts commit d59cfc09c32a2ae31f1c3bc2983a0cd79afb3f14.

d59cfc09c32a ("sched, cgroup: replace signal_struct->group_rwsem with
a global percpu_rwsem") and b5ba75b5fc0e ("cgroup: simplify
threadgroup locking") changed how cgroup synchronizes against task
fork and exits so that it uses global percpu_rwsem instead of
per-process rwsem; unfortunately, the write [un]lock paths of
percpu_rwsem always involve synchronize_rcu_expedited() which turned
out to be too expensive.

Improvements for percpu_rwsem are scheduled to be merged in the coming
v4.4-rc1 merge window which alleviates this issue.  For now, revert
the two commits to restore per-process rwsem.  They will be re-applied
for the v4.4-rc1 merge window.

Signed-off-by: Tejun Heo <tj@kernel.org>
Link: http://lkml.kernel.org/g/55F8097A.7000206@de.ibm.com
Reported-by: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: stable@vger.kernel.org # v4.2+
include/linux/cgroup-defs.h
include/linux/init_task.h
include/linux/sched.h
kernel/cgroup.c
kernel/fork.c