openssh bugs - Jul 2013 - [Bug 2124] New: TCP_NODELAY not set by sshd for non-interactive non-exec sessions

https://bugzilla.mindrot.org/show_bug.cgi?id=2124

            Bug ID: 2124
           Summary: TCP_NODELAY not set by sshd for non-interactive
                    non-exec sessions
           Product: Portable OpenSSH
           Version: -current
          Hardware: All
                OS: All
            Status: NEW
          Severity: trivial
          Priority: P5
         Component: sshd
          Assignee: unassigned-bugs at mindrot.org
          Reporter: arthurhsliu at gmail.com

Bugs #556 and #981 describe situations where TCP_NODELAY is not set. A
fix is provided in bug #556. However, the patch commit for packet.c in
revision 1.120 does not set TCP_NODELAY for all connections.

If the ssh client is invoked with:
ssh -N host -R port          # TCP_NODELAY is not set
ssh -n host -R port sleep 1d # TCP_NODELAY is set - this is a
workaround

There is a difference in TCP_NODELAY being set or not depending on
whether there is a remote command being executed. This is because
packet_set_interactive() is only called in do_exec() on the server
side.
This bug has the unfortunate effect of not setting TCP_NODELAY for port
forwarding only connections.

The proper fix is probably to use attachment 1099 by Darren Tucker that
sets it for all connections on the server side.
https://bugzilla.mindrot.org/show_bug.cgi?id=556#c10
https://bugzilla.mindrot.org/attachment.cgi?id=1099&action=diff

Issue present is in OpenSSH 4.4 through to 6.2 and -current.

-- 
You are receiving this mail because:
You are watching the assignee of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=2124

Damien Miller <djm at mindrot.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |djm at mindrot.org,
                   |                            |dtucker at zip.com.au
           Assignee|unassigned-bugs at mindrot.org |djm at mindrot.org
   Attachment #2311|                            |ok?(dtucker at zip.com.au)
              Flags|                            |

--- Comment #1 from Damien Miller <djm at mindrot.org> ---
Created attachment 2311
  --> https://bugzilla.mindrot.org/attachment.cgi?id=2311&action=edit
call packet_set_interactive() when -N set

I think this is better - it sets the QoS too and uses the presence of a
multiplexing socket to infer interactivity.

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=2124

Darren Tucker <dtucker at zip.com.au> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Attachment #2311|ok?(dtucker at zip.com.au)     |ok+
              Flags|                            |

-- 
You are receiving this mail because:
You are watching someone on the CC list of the bug.
You are watching the assignee of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=2124

--- Comment #2 from Arthur Liu <arthurhsliu at gmail.com> ---
Thanks for the fix. It fixes the client side socket, which has the
issue as well (we worked around it by using -tt flag). With your fix it
fixes traffic ssh->sshd.

However, TCP_NODELAY is still not set for the TCP socket on sshd's side
(for traffic sshd->ssh), which the original attachment 1099 fixes.

-- 
You are receiving this mail because:
You are watching someone on the CC list of the bug.
You are watching the assignee of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=2124

Damien Miller <djm at mindrot.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|---                         |FIXED
             Blocks|                            |2076

--- Comment #3 from Damien Miller <djm at mindrot.org> ---
fix applied. This will be in openssh-6.3. Thanks!

-- 
You are receiving this mail because:
You are watching someone on the CC list of the bug.
You are watching the assignee of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=2124

Damien Miller <djm at mindrot.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |CLOSED

--- Comment #4 from Damien Miller <djm at mindrot.org> ---
Set all RESOLVED bugs to CLOSED with release of OpenSSH 7.1

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.