lhotari commented on code in PR #22760:
URL: https://github.com/apache/pulsar/pull/22760#discussion_r1610760483
##########
pulsar-common/src/main/java/org/apache/pulsar/common/protocol/ByteBufPair.java:
##########
@@ -122,38 +123,22 @@ public void write(ChannelHandlerContext ctx, Object msg,
ChannelPromise promise)
// ByteBuf are automatically released after a write. If the
ByteBufPair ref count is increased and it
// gets written multiple times, the individual buffers
refcount should be reflected as well.
try {
- ctx.write(b.getFirst().retainedDuplicate(),
ctx.voidPromise());
- ctx.write(b.getSecond().retainedDuplicate(), promise);
+ ctx.write(readOnlyRetainedDuplicate(b.getFirst()),
ctx.voidPromise());
+ ctx.write(readOnlyRetainedDuplicate(b.getSecond()),
promise);
} finally {
ReferenceCountUtil.safeRelease(b);
}
} else {
ctx.write(msg, promise);
}
}
- }
-
- @Sharable
- @SuppressWarnings("checkstyle:JavadocType")
- public static class CopyingEncoder extends ChannelOutboundHandlerAdapter {
- @Override
- public void write(ChannelHandlerContext ctx, Object msg,
ChannelPromise promise) throws Exception {
- if (msg instanceof ByteBufPair) {
- ByteBufPair b = (ByteBufPair) msg;
- // Some handlers in the pipeline will modify the bytebufs
passed in to them (i.e. SslHandler).
- // For these handlers, we need to pass a copy of the buffers
as the source buffers may be cached
- // for multiple requests.
- try {
- ctx.write(b.getFirst().copy(), ctx.voidPromise());
- ctx.write(b.getSecond().copy(), promise);
- } finally {
- ReferenceCountUtil.safeRelease(b);
- }
- } else {
- ctx.write(msg, promise);
- }
+ // .asReadOnly() is needed to prevent SslHandler from modifying the
input buffers.
+ private static ByteBuf readOnlyRetainedDuplicate(ByteBuf buf) {
+ // If the buffer is already read-only, .asReadOnly() will return
the same buffer.
+ // That's why the additional .retainedDuplicate() is needed to
ensure that the returned buffer
+ // has independent readIndex and writeIndex.
+ return buf.asReadOnly().retainedDuplicate();
Review Comment:
> It sounds very strange, can you please point me where?
Yes, it's very surprising behavior. You can try it out with a debugger by
running for example TlsProducerConsumerTest and checking what happens in the
attemptCopyToCumulation if the `.asReadOnly().retainedDuplicate()` solution is
used.
The io.netty.buffer.AbstractByteBuf#isWritable(int) method gets called:
https://github.com/netty/netty/blob/70d6a3f40d7e8fd3f5ced7600ed209c58944f673/buffer/src/main/java/io/netty/buffer/AbstractByteBuf.java#L171-L174
It will evaluate to true since it's implementation doesn't check for
`isReadOnly()`:
```java
@Override
public boolean isWritable(int numBytes) {
return capacity() - writerIndex >= numBytes;
}
```
--
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: commits-unsubscr...@pulsar.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
