rcu: increase synchronize_sched_expedited() batching

The fix in commit #6a0cc49 requires more than three concurrent instances
of synchronize_sched_expedited() before batching is possible.  This
patch uses a ticket-counter-like approach that is also not unrelated to
Lai Jiangshan's Ring RCU to allow sharing of expedited grace periods even
when there are only two concurrent instances of synchronize_sched_expedited().

This commit builds on Tejun's original posting, which may be found at
http://lkml.org/lkml/2010/11/9/204, adding memory barriers, avoiding
overflow of signed integers (other than via atomic_t), and fixing the
detection of batching.

Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>

1 files changed, 2 insertions, 0 deletions

diff --git a/include/linux/rcupdate.h b/include/linux/rcupdate.h
index 49e8e16308e1..af5614856285 100644
--- a/include/linux/rcupdate.h
+++ b/include/linux/rcupdate.h
@@ -47,6 +47,8 @@
 extern int rcutorture_runnable; /* for sysctl */
 #endif /* #ifdef CONFIG_RCU_TORTURE_TEST */
 
+#define UINT_CMP_GE(a, b)       (UINT_MAX / 2 >= (a) - (b))
+#define UINT_CMP_LT(a, b)       (UINT_MAX / 2 < (a) - (b))
 #define ULONG_CMP_GE(a, b)      (ULONG_MAX / 2 >= (a) - (b))
 #define ULONG_CMP_LT(a, b)      (ULONG_MAX / 2 < (a) - (b))