]> asedeno.scripts.mit.edu Git - linux.git/commitdiff
rcu: Suppress lockdep false-positive ->boost_mtx complaints
authorPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Tue, 19 Sep 2017 22:36:42 +0000 (15:36 -0700)
committerPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Thu, 19 Oct 2017 19:13:36 +0000 (12:13 -0700)
RCU priority boosting uses rt_mutex_init_proxy_locked() to initialize an
rt_mutex structure in locked state held by some other task.  When that
other task releases it, lockdep complains (quite accurately, but a bit
uselessly) that the other task never acquired it.  This complaint can
suppress other, more helpful, lockdep complaints, and in any case it is
a false positive.

This commit therefore switches from rt_mutex_unlock() to
rt_mutex_futex_unlock(), thereby avoiding the lockdep annotations.
Of course, if lockdep ever learns about rt_mutex_init_proxy_locked(),
addtional adjustments will be required.

Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
kernel/rcu/tree_plugin.h

index 969eae45f05dbff7491f9ecffe3205f3ae5339d8..1eaab96d1a3cdb07d55ed378b28a143c32502553 100644 (file)
@@ -531,7 +531,7 @@ void rcu_read_unlock_special(struct task_struct *t)
 
                /* Unboost if we were boosted. */
                if (IS_ENABLED(CONFIG_RCU_BOOST) && drop_boost_mutex)
-                       rt_mutex_unlock(&rnp->boost_mtx);
+                       rt_mutex_futex_unlock(&rnp->boost_mtx);
 
                /*
                 * If this was the last task on the expedited lists,