Bug 1328

Summary: Support multiple X11 forwarding in multiplexing
Product: Portable OpenSSH Reporter: David Woodhouse <dwmw2>
Component: sshAssignee: Assigned to nobody <unassigned-bugs>
Status: CLOSED WONTFIX    
Severity: enhancement CC: djm
Priority: P2    
Version: -current   
Hardware: All   
OS: All   
Attachments:
Description Flags
Handle multiple $DISPLAYs
none
pass display information through controlclient socket
none
Updated multi-display patch against 5.5p1
none
Updated muxclient display patch against 5.5p1 none

Description David Woodhouse 2007-07-06 01:15:52 AEST
Created attachment 1316 [details]
Handle multiple $DISPLAYs

When using the ControlClient feature, X programs get forwarded to the $DISPLAY of the _master_, not the client.

The X forwarding support in the SSH protocol doesn't allow for specifying a display for forwarded connections, but we can achieve it by generating different 'fake authentication data' for each one, then using that to infer which display should be used.

(We have a similar bug with agent forwarding, and no 'easy' fix for that one without a protocol change. It has been suggested that because we can't fix _that_ bug, we shouldn't fix _this_ bug either, for 'consistency'. That logic seems a little flawed to me. Besides, I've tripped over the 'wrong display' thing _many_ times and I have not _once_ been bothered by the fact that I'm using the 'wrong' agent.)
Comment 1 David Woodhouse 2007-07-06 01:17:12 AEST
Created attachment 1317 [details]
pass display information through controlclient socket

If we're going to use the correct $DISPLAY it's probably a good idea to let the controlclient tell the master which one to use.
Comment 2 David Woodhouse 2010-06-05 12:03:54 AEST
Created attachment 1856 [details]
Updated multi-display patch against 5.5p1
Comment 3 David Woodhouse 2010-06-05 12:05:35 AEST
Created attachment 1857 [details]
Updated muxclient display patch against 5.5p1
Comment 4 Damien Miller 2020-01-25 23:00:57 AEDT
Sorry, but X11 forwarding is enough of a niche feature these days that we don't want to support protocol workarounds to support additional corner cases atop it.
Comment 5 Damien Miller 2021-04-23 14:59:58 AEST
closing resolved bugs as of 8.6p1 release