>
> > Well, yes, I suspected that a the end of the day I will have to do
> > protocol snooping. Ok, thank you for your explanation.
>
> I first suggested another method of gaining more data. You seem to
> have overlooked that. Please do not overlook any advice you get.
>
> But yes, you will need to understand the application protocol better
> to determine if there is a problem with libssh2, and how it should be
> fixed.
>
>
>
I didn't overlook it, I will test linux client as soon as I get hold of a
linux machine. If I can bother you a bit more. I debugged the circuit a bit
and the last thing that happens before disconnect is this:
("TCP" being local network port on which RDC client talks to "SSH" -
channel got from listener on a local ssh server, d_readyRead shows each
clearing of blocking condition in libssh2):
"1. SSH->TCP 627 bytes
QxtSshClientPrivate::d_readyRead----------------------------------------------
"1. SSH->TCP 104 bytes
"2. TCP->SSH: 327 bytes
QxtSshClientPrivate::d_readyRead----------------------------------------------
"1. SSH->TCP 527 bytes
"2. TCP->SSH: 100 bytes
QxtSshClientPrivate::d_readyRead----------------------------------------------
"1. SSH->TCP 226 bytes
"2. TCP->SSH: 104 bytes
"2. TCP->SSH: 1412 bytes
"2. TCP->SSH: 327 bytes
QxtSshClientPrivate::d_readyRead----------------------------------------------
"1. SSH->TCP 118 bytes
QxtSshClientPrivate::d_readyRead----------------------------------------------
TCP disconnects, breaking circuit
I tested it several times and the sequence before disconnect looks very
similar each time. Don't know much about frames, but the number of bytes
exchanged seems... well... so small, are these frames smaller than 1kB?
_______________________________________________
libssh2-devel http://cool.haxx.se/cgi-bin/mailman/listinfo/libssh2-devel
Received on 2011-09-26