rcu: Fix note_new_gpnum() uses of ->gpnum

Impose a clear locking design on the note_new_gpnum()
function's use of the ->gpnum counter.  This is done by updating
rdp->gpnum only from the corresponding leaf rcu_node structure's
rnp->gpnum field, and even then only under the protection of
that same rcu_node structure's ->lock field.  Performance and
scalability are maintained using a form of double-checked
locking, and excessive spinning is avoided by use of the
spin_trylock() function.  The use of spin_trylock() is safe due
to the fact that CPUs who fail to acquire this lock will try
again later. The hierarchical nature of the rcu_node data
structure limits contention (which could be limited further if
need be using the RCU_FANOUT kernel parameter).

Without this patch, obscure but quite possible races could
result in a quiescent state that occurred during one grace
period to be accounted to the following grace period, causing
this following grace period to end prematurely.  Not good!
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: laijs@cn.fujitsu.com
Cc: dipankar@in.ibm.com
Cc: mathieu.desnoyers@polymtl.ca
Cc: josh@joshtriplett.org
Cc: dvhltc@us.ibm.com
Cc: niv@us.ibm.com
Cc: peterz@infradead.org
Cc: rostedt@goodmis.org
Cc: Valdis.Kletnieks@vt.edu
Cc: dhowells@redhat.com
Cc: <stable@kernel.org> # .32.x
LKML-Reference: <12571987492350-git-send-email->
Signed-off-by: Ingo Molnar <mingo@elte.hu>

kernel/rcutree.c

@@ -540,13 +540,33 @@ static void check_cpu_stall(struct rcu_state *rsp, struct rcu_data *rdp)
 /*
  * Update CPU-local rcu_data state to record the newly noticed grace period.
  * This is used both when we started the grace period and when we notice
- * that someone else started the grace period.
+ * that someone else started the grace period.  The caller must hold the
+ * ->lock of the leaf rcu_node structure corresponding to the current CPU,
+ *  and must have irqs disabled.
  */
-static void note_new_gpnum(struct rcu_state *rsp, struct rcu_data *rdp)
+static void __note_new_gpnum(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_data *rdp)
 {
+	if (rdp->gpnum != rnp->gpnum) {
 		rdp->qs_pending = 1;
 		rdp->passed_quiesc = 0;
-	rdp->gpnum = rsp->gpnum;
+		rdp->gpnum = rnp->gpnum;
+	}
+}
+
+static void note_new_gpnum(struct rcu_state *rsp, struct rcu_data *rdp)
+{
+	unsigned long flags;
+	struct rcu_node *rnp;
+
+	local_irq_save(flags);
+	rnp = rdp->mynode;
+	if (rdp->gpnum == ACCESS_ONCE(rnp->gpnum) || /* outside lock. */
+	    !spin_trylock(&rnp->lock)) { /* irqs already off, retry later. */
+		local_irq_restore(flags);
+		return;
+	}
+	__note_new_gpnum(rsp, rnp, rdp);
+	spin_unlock_irqrestore(&rnp->lock, flags);
 }
 
 /*
@@ -637,6 +657,9 @@ rcu_start_gp_per_cpu(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_dat
 	 */
 	rdp->nxttail[RCU_NEXT_READY_TAIL] = rdp->nxttail[RCU_NEXT_TAIL];
 	rdp->nxttail[RCU_WAIT_TAIL] = rdp->nxttail[RCU_NEXT_TAIL];
+
+	/* Set state so that this CPU will detect the next quiescent state. */
+	__note_new_gpnum(rsp, rnp, rdp);
 }
 
 /*
@@ -664,7 +687,6 @@ rcu_start_gp(struct rcu_state *rsp, unsigned long flags)
 	rsp->jiffies_force_qs = jiffies + RCU_JIFFIES_TILL_FORCE_QS;
 	record_gp_stall_check_time(rsp);
 	dyntick_record_completed(rsp, rsp->completed - 1);
-	note_new_gpnum(rsp, rdp);
 
 	/* Special-case the common single-level case. */
 	if (NUM_RCU_NODES == 1) {