[SPARK-26776][PYTHON] Reduce Py4J communication cost in PySpark's execution barrier check

## What changes were proposed in this pull request?

I am investigating flaky tests. I realised that:

```
      File "/home/jenkins/workspace/SparkPullRequestBuilder/python/pyspark/rdd.py", line 2512, in __init__
        self.is_barrier = prev._is_barrier() or isFromBarrier
      File "/home/jenkins/workspace/SparkPullRequestBuilder/python/pyspark/rdd.py", line 2412, in _is_barrier
        return self._jrdd.rdd().isBarrier()
      File "/home/jenkins/workspace/SparkPullRequestBuilder/python/lib/py4j-0.10.8.1-src.zip/py4j/java_gateway.py", line 1286, in __call__
        answer, self.gateway_client, self.target_id, self.name)
      File "/home/jenkins/workspace/SparkPullRequestBuilder/python/lib/py4j-0.10.8.1-src.zip/py4j/protocol.py", line 342, in get_return_value
        return OUTPUT_CONVERTER[type](answer[2:], gateway_client)
      File "/home/jenkins/workspace/SparkPullRequestBuilder/python/lib/py4j-0.10.8.1-src.zip/py4j/java_gateway.py", line 2492, in <lambda>
        lambda target_id, gateway_client: JavaObject(target_id, gateway_client))
      File "/home/jenkins/workspace/SparkPullRequestBuilder/python/lib/py4j-0.10.8.1-src.zip/py4j/java_gateway.py", line 1324, in __init__
        ThreadSafeFinalizer.add_finalizer(key, value)
      File "/home/jenkins/workspace/SparkPullRequestBuilder/python/lib/py4j-0.10.8.1-src.zip/py4j/finalizer.py", line 43, in add_finalizer
        cls.finalizers[id] = weak_ref
      File "/usr/lib64/pypy-2.5.1/lib-python/2.7/threading.py", line 216, in __exit__
        self.release()
      File "/usr/lib64/pypy-2.5.1/lib-python/2.7/threading.py", line 208, in release
        self.__block.release()
    error: release unlocked lock
```

I assume it might not be directly related with the test itself but I noticed that it `prev._is_barrier()` attempts to access via Py4J.

Accessing via Py4J is expensive. Therefore, this PR proposes to avoid Py4J access when `isFromBarrier` is `True`.

## How was this patch tested?

Unittests should cover this.

Closes #23690 from HyukjinKwon/minor-barrier.

Authored-by: Hyukjin Kwon <gurwls223@apache.org>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>

Author: HyukjinKwon

python/pyspark/rdd.py

@@ -2509,7 +2509,7 @@ def pipeline_func(split, iterator):
         self._jrdd_deserializer = self.ctx.serializer
         self._bypass_serializer = False
         self.partitioner = prev.partitioner if self.preservesPartitioning else None
-        self.is_barrier = prev._is_barrier() or isFromBarrier
+        self.is_barrier = isFromBarrier or prev._is_barrier()
 
     def getNumPartitions(self):
         return self._prev_jrdd.partitions().size()