locking/rwsem: Don't hog RCU read lock while optimistically spinning

There's no reason to hold an RCU read lock the entire time while
optimistically spinning for a rwsem. This can needlessly lengthen RCU
grace periods and slow down synchronize_rcu() when it doesn't brute
force the RCU grace period via rcupdate.rcu_expedited=1.

Signed-off-by: Sultan Alsawaf <sultan@kerneltoast.com>
Signed-off-by: Ruchit <ruchitmarathe@gmail.com>

kernel/locking/rwsem-xadd.c

@@ -400,32 +400,36 @@ static noinline bool rwsem_spin_on_owner(struct rw_semaphore *sem)
 {
 	struct task_struct *owner = READ_ONCE(sem->owner);
 
-	if (!is_rwsem_owner_spinnable(owner))
+	if (!owner || !is_rwsem_owner_spinnable(owner))
 		return false;
 
-	rcu_read_lock();
-	while (owner && (READ_ONCE(sem->owner) == owner)) {
+	while (true) {
+		bool on_cpu, same_owner;
+
 		/*
-		 * Ensure we emit the owner->on_cpu, dereference _after_
-		 * checking sem->owner still matches owner, if that fails,
+		 * Ensure sem->owner still matches owner. If that fails,
 		 * owner might point to free()d memory, if it still matches,
 		 * the rcu_read_lock() ensures the memory stays valid.
 		 */
-		barrier();
+		rcu_read_lock();
+		same_owner = sem->owner == owner;
+		if (same_owner)
+			on_cpu = owner->on_cpu &&
+				 !vcpu_is_preempted(task_cpu(owner));
+		rcu_read_unlock();
+
+		if (!same_owner)
+			break;
 
 		/*
 		 * abort spinning when need_resched or owner is not running or
 		 * owner's cpu is preempted.
 		 */
-		if (!owner->on_cpu || need_resched() ||
-				vcpu_is_preempted(task_cpu(owner))) {
-			rcu_read_unlock();
+		if (!on_cpu || need_resched())
 			return false;
-		}
 
 		cpu_relax();
 	}
-	rcu_read_unlock();
 
 	/*
 	 * If there is a new owner or the owner is not set, we continue