[ 
https://issues.apache.org/jira/browse/QPID-8791?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18136941#comment-18136941
 ] 

ASF GitHub Bot commented on QPID-8791:
--------------------------------------

coolkaito opened a new pull request, #466:
URL: https://github.com/apache/qpid-broker-j/pull/466

   ---
   JIRA Issue
   ---
   
   [QPID-8791](https://issues.apache.org/jira/browse/QPID-8791): [Broker-J] An 
incomplete inbound TLS record can cause repeated rescheduling
   
   ---
   PR Description
   ---
   
   The proposed change tracks when parsing or decryption needs more network 
data, preventing the connection from repeatedly requesting worker time while 
waiting for record completion.




> [Broker-J] An incomplete inbound TLS record can cause repeated rescheduling
> ---------------------------------------------------------------------------
>
>                 Key: QPID-8791
>                 URL: https://issues.apache.org/jira/browse/QPID-8791
>             Project: Qpid
>          Issue Type: Bug
>          Components: Broker-J
>    Affects Versions: qpid-java-broker-10.1.1
>            Reporter: Kato Kaito
>            Priority: Minor
>
> After the TLS handshake has completed, unwrap() can return BUFFER_UNDERFLOW 
> while the handshake status remains NOT_HANDSHAKING. However, 
> NonBlockingConnectionTLSDelegate.needsWork() checks only whether the 
> handshake status differs from NEED_UNWRAP, so it returns true even though 
> processing requires additional network data.
> When pending output has fully drained, NonBlockingConnection.doWork() sees 
> the non-empty input buffer and calls notifyWork(). On subsequent passes, the 
> non-blocking socket read returns 0, leaving the buffered data unchanged, and 
> the connection is rescheduled again. This can continue without considering 
> socket read readiness until the remaining bytes arrive or the connection is 
> closed.



--
This message was sent by Atlassian Jira
(v10.3.26#10030026)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to