You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@commons.apache.org by "Archie Cobbs (JIRA)" <ji...@apache.org> on 2010/11/22 21:03:14 UTC
[jira] Commented: (NET-345) Telnet client: not properly handling
IAC bytes within subnegotiation messages
[ https://issues.apache.org/jira/browse/NET-345?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12934573#action_12934573 ]
Archie Cobbs commented on NET-345:
----------------------------------
FYI, the latest version of my combined patch to fix NET-343, NET-344, and NET-345 can be found [here|http://jvser.googlecode.com/svn/trunk/src/commons-net-telnet/patch.txt].
Please use this version as I'll be keeping it up to date.
> Telnet client: not properly handling IAC bytes within subnegotiation messages
> -----------------------------------------------------------------------------
>
> Key: NET-345
> URL: https://issues.apache.org/jira/browse/NET-345
> Project: Commons Net
> Issue Type: Bug
> Components: Telnet
> Affects Versions: 2.0
> Reporter: Archie Cobbs
> Attachments: patch3.txt, patch4.txt
>
>
> Subnegotiation messages in telnet are sent using the sequence {{IAC SB ... IAC SE}}.
> Although it's not clearly spelled out in [RFC 854|http://tools.ietf.org/html/rfc854], any {{IAC}} ({{0xff}}) bytes inside these messages must be escaped by doubling. Other clients do this and this is the only behavior that makes sense.
> The commons-net telnet client is failing both to escape and to unescape {{IAC}} bytes within subnegotiation messages. Moreover, if it does receive a valid {{IAC IAC}} sequence within a subnegotiation message, it will incorrectly jump back to "data" input mode, discarding the message and introducing its remainder as garbage in the data stream.
> In addition, the code fails to check for an overflow of the subnegotiation buffer, which would cause an {{ArrayIndexOutOfBounds}} exception if a malicious peer triggered this condition.
> Finally, a {{IAC SE}} sequence appearing by itself should probably be discarded, rather than passing as a command to the handler.
> I'm attaching a patch to fix these issues.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.