[SPARK-23484][SS] Fix possible race condition in KafkaContinuousReader

## What changes were proposed in this pull request?

var `KafkaContinuousReader.knownPartitions` should be threadsafe as it is accessed from multiple threads - the query thread at the time of reader factory creation, and the epoch tracking thread at the time of `needsReconfiguration`.

## How was this patch tested?

Existing tests.

Author: Tathagata Das <tathagata.das1565@gmail.com>

Closes apache#20655 from tdas/SPARK-23484.

external/kafka-0-10-sql/src/main/scala/org/apache/spark/sql/kafka010/KafkaContinuousReader.scala

@@ -66,7 +66,7 @@ class KafkaContinuousReader(
   // Initialized when creating reader factories. If this diverges from the partitions at the latest
   // offsets, we need to reconfigure.
   // Exposed outside this object only for unit tests.
-  private[sql] var knownPartitions: Set[TopicPartition] = _
+  @volatile private[sql] var knownPartitions: Set[TopicPartition] = _
 
   override def readSchema: StructType = KafkaOffsetReader.kafkaSchema