Bug 2417 - SOCKS5 should respond with appropriate error reply in error situations
Summary: SOCKS5 should respond with appropriate error reply in error situations
Status: ASSIGNED
Alias: None
Product: Portable OpenSSH
Classification: Unclassified
Component: ssh (show other bugs)
Version: 6.8p1
Hardware: All All
: P5 enhancement
Assignee: Damien Miller
URL:
Keywords:
Depends on:
Blocks: 2250
  Show dependency treegraph
 
Reported: 2015-06-25 22:15 AEST by Jonas Berlin
Modified: 2016-06-17 14:52 AEST (History)
1 user (show)

See Also:


Attachments
Initial implementation proposal with a few TODOs that I don't know how to implement (3.43 KB, patch)
2015-06-25 22:15 AEST, Jonas Berlin
no flags Details | Diff
Initial implementation proposal with a few TODOs that I don't know how to implement (3.22 KB, patch)
2015-06-25 22:22 AEST, Jonas Berlin
no flags Details | Diff
tidied diff (8.07 KB, patch)
2015-10-31 09:18 AEDT, Damien Miller
no flags Details | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Jonas Berlin 2015-06-25 22:15:34 AEST
Created attachment 2656 [details]
Initial implementation proposal with a few TODOs that I don't know how to implement

- report "bad address type" error if requested address type is not supported
- report "ruleset block" error if requested hostname too long
- report "ruleset block", "connection refused" or "generic error" if server-side connection attempt failure result is "administratively prohibited", "connect failed" or something else, respectively.
Comment 1 Jonas Berlin 2015-06-25 22:17:12 AEST
This bug complements bug #2250 which handles one additional error situation
Comment 2 Jonas Berlin 2015-06-25 22:22:59 AEST
Created attachment 2657 [details]
Initial implementation proposal with a few TODOs that I don't know how to implement

removes #define that is already added by bug #2250 and not actually needed for this bug
Comment 3 Damien Miller 2015-10-31 09:18:27 AEDT
Created attachment 2744 [details]
tidied diff

I've tidied the diff up a bit, but I think we need some extra support in the channels code to allow the reply to connfailed requests to be sent in a timely manner.
Comment 4 Damien Miller 2016-06-17 14:52:18 AEST
Fixing this is required for Bug 2250 too, but I'm not going to attempt it until the channels code has been refactored a bit.