rcu: Reduce rcutorture tracing
authorPaul E. McKenney <paul.mckenney@linaro.org>
Thu, 15 Nov 2012 00:26:40 +0000 (16:26 -0800)
committerPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Tue, 8 Jan 2013 22:14:55 +0000 (14:14 -0800)
commit52494535103986dbbf689b44d8c2c7efe2132b16
tree88dd5a9d064608e9711c36f05b3b4aab8e255a4f
parentdc975e94f322e60fa8fcc44dec1820fde4de174c
rcu: Reduce rcutorture tracing

Currently, rcutorture traces every read-side access.  This can be
problematic because even a two-minute rcutorture run on a two-CPU system
can generate 28,853,363 reads.  Normally, only a failing read is of
interest, so this commit traces adjusts rcutorture's tracing to only
trace failing reads.  The resulting event tracing records the time
and the ->completed value captured at the beginning of the RCU read-side
critical section, allowing correlation with other event-tracing messages.

Signed-off-by: Paul E. McKenney <paul.mckenney@linaro.org>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
[ paulmck: Add fix to build problem located by Randy Dunlap based on
  diagnosis by Steven Rostedt. ]
include/linux/rcupdate.h
include/trace/events/rcu.h
kernel/rcupdate.c
kernel/rcutorture.c
lib/Kconfig.debug