rcu: Improve diagnostics for blocked critical sections in irq
authorPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Wed, 21 Jan 2015 23:26:03 +0000 (15:26 -0800)
committerPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Tue, 3 Mar 2015 19:16:00 +0000 (11:16 -0800)
commitd24209bb689e2c7f7418faec9b4a948e922d24da
tree40db37902a30acb68f5733e666f25ea7b52e8485
parent6629240575992a6f0d18c46f5160b34527b0e501
rcu: Improve diagnostics for blocked critical sections in irq

If an RCU read-side critical section occurs within an interrupt handler
or a softirq handler, it cannot have been preempted.  Therefore, there is
a check in rcu_read_unlock_special() checking for this error.  However,
when this check triggers, it lacks diagnostic information.  This commit
therefore moves rcu_read_unlock()'s lockdep annotation to follow the
call to __rcu_read_unlock() and changes rcu_read_unlock_special()'s
WARN_ON_ONCE() to an lockdep_rcu_suspicious() in order to locate where
the offending RCU read-side critical section began.  In addition, the
value of the ->rcu_read_unlock_special field is printed.

Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
include/linux/lockdep.h
include/linux/rcupdate.h
kernel/rcu/tree_plugin.h