Bug#221849: ClientAliveInterval option emits error to screen

[Petr Vandrovec <vandrove@vc.cvut.cz>]

Package: ssh
Version: 3.6.1p2-10

Hi,
  I'm connecting to some system through firewall by using:

    ssh -f -N -L 2022:internalhost:22 user@firewall
    exec ssh -p 2022 -X -L ... -L ... -L ... user@localhost

Recently client keepalives were enabled on the firewall, and it
renders second ssh unusable for interactive work, as after every 
couple of seconds there is

    client_input_channel_req: no channel -1

printed on screen, due to bogus keepalive@openssh.com request.
For now I'll stick '-o loglevel=silent' to the first ssh cmdline,
but it would be nice if keepalive could be somehow fixed in
the client to not generate error message.

With enabled debugging it says

    debug1: client_input_channel_req: channel 1 rtype keepalive@openssh.com reply 1
    client_input_channel_req: no channel -1

so I'm 100% sure that keepalive is a culprit...

					Thanks,
						Petr Vandrovec
						vandrove@vc.cvut.cz