Github user bkrieger commented on a diff in the pull request:
https://github.com/apache/spark/pull/22995#discussion_r236715555
--- Diff:
core/src/main/scala/org/apache/spark/broadcast/TorrentBroadcast.scala ---
@@ -92,8 +95,15 @@ private[spark] class TorrentBroadcast[T: ClassTag](obj:
T, id: Long)
/** The checksum for all the blocks. */
private var checksums: Array[Int] = _
- override protected def getValue() = {
- _value
+ override protected def getValue() = synchronized {
--- End diff --
I think we can remove the `TorrentBroadcast.synchronized` in
`readBroadcastBlock`, since we're already synchronizing in its only caller?
Though I'm not sure why it was necessary in the first place, as
`readBroadcastBlock` should only have been called once before this PR.
Regardless, I agree that the perf hit should be ok. Let me know if you want
any of this changed.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
