pnowojski commented on a change in pull request #18173:
URL: https://github.com/apache/flink/pull/18173#discussion_r776622362
##########
File path:
flink-runtime/src/main/java/org/apache/flink/runtime/io/network/buffer/NetworkBufferPool.java
##########
@@ -158,7 +159,8 @@ public MemorySegment requestMemorySegment() {
public List<MemorySegment> requestMemorySegmentsBlocking(int
numberOfSegmentsToRequest)
throws IOException {
- return internalRequestMemorySegments(numberOfSegmentsToRequest);
+ return internalRequestMemorySegments(
+ numberOfSegmentsToRequest,
this::internalRecycleMemorySegments);
Review comment:
Maybe lets rename `requestMemorySegments` to something like
`requestUnpooledMemorySegments()` (but please do so as a separate commit - can
be in this PR), with the following java doc:
```
Unpooled memory segments are requested directly from {@link
NetworkBufferPool}, as opposed to pooled
segments, that are requested through {@link BufferPool} that was created
from this {@link NetworkBufferPool}
(see {@link NetworkBufferPool#createBufferPool(...)}). They are used for
example for exclusive
{@link RemoteInputChannel} credits, that are permanently assigned to that
channel, and never returned to
any {@link BufferPool}. As opposed to pooled segments, when requested,
unpooled segments needs to be accounted against {@link
#numTotalRequiredBuffers}, which might require redistribution of the segments.
```
+ renaming other methods by adding "pooled" suffixes?
I don't like that "unpooled" is ambiguous, because we have two levels of
pooling, but I don't know how to express it differently. Or do you have some
better proposal?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscr...@flink.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
