[SPARK-22721] BytesToBytesMap peak memory usage not accurate after reset()

## What changes were proposed in this pull request?

BytesToBytesMap doesn't update peak memory usage before shrinking back to initial capacity in reset(), so after a disk spill one never knows what was the size of hash table was before spilling.

## How was this patch tested?

Checked manually.

Author: Juliusz Sompolski <[email protected]>

Closes #19915 from juliuszsompolski/SPARK-22721.

Author: juliuszsompolski

core/src/main/java/org/apache/spark/unsafe/map/BytesToBytesMap.java

@@ -879,6 +879,7 @@ public LongArray getArray() {
    * Reset this map to initialized state.
    */
   public void reset() {
+    updatePeakMemoryUsed();
     numKeys = 0;
     numValues = 0;
     freeArray(longArray);