[Vortex] vortex_channel_wait_reply deadlocks until vortex_connection_new is called on remote

[Chris Yourch]

-----Original Message-----
From: Jens Alfke [mailto:jens at mooseyard.com]
Sent: Friday, May 02, 2008 5:43 PM
To: Chris Yourch
Cc: Vortex
Subject: Re: [Vortex] vortex_channel_wait_reply deadlocks until vortex_connection_new is called on remote


On 2 May '08, at 4:55 PM, Chris Yourch wrote:

> 3)     Remote calls vortex_connection_new() during MSG frame
> processing to connect to the local machine using a different channel.

It's opening a new _connection_, not a new channel? That seems odd;
usually there's no need for an application to open multiple BEEP
connections to the same peer, since BEEP already lets you multiplex
using channels.

Hmmm.... On the remote, can I use the connection that is passed to my listener instead of calling vortex_connection_new()?


> 4)     vortex_connection_new() hangs for a while.
> vortex_connection_new() stops hanging when the
> vortex_channel_wait_reply() call on the local machine times out.

I would guess that you're using the synchronous mode of
vortex_connection_new, and that the local peer's client code is
getting stuck responding to the incoming connection request while it's
still blocked waiting for the reply to the message. You'd have to
examine the stacks on the local machine (and turn on logging) to
figure out what's going on.

I am using the synchronous version of vortex_connection_new() so maybe that's the issue. I'll look into to it.



-Jens