Re: connect() hangs on a listen()ing AF_UNIX socket

[Corinna Vinschen <corinna-cygwin at cygwin dot com>]

On Aug 22 20:32, Christian Franke wrote:
> Corinna Vinschen wrote:
> >On Aug 21 21:14, Christian Franke wrote:
> >>Easier and may work for Postfix: Add a Cygwin specific socket option like
> >>SO_DONT_NEED_PEERCRED which is set immediately after Postfix calls
> >>socket(AF_UNIX, SOCK_STREAM). If set, no handshake occurs on
> >>connect()/accept(). getpeerid()/SO_PEERCRED should fail then.
> >Well, it's not *only* SO_PEERCRED.  Another, the older part of the
> >handshake, is about recognizing the peer.  Since AF_UNIX sockets don't
> >exist on Windows, Cygwin is using AF_INET sockets under the hood, and
> >so *any* Windows process could accidentally connect to a Cygwin AF_UNIX
> >socket.  The handshake also aims to avoid this scenario.  Only if the
> >handshake worked, the peers can be sure to talk to another Cygwin
> >process assuming an AF_UNIX socket.
> >
> >A Cygwin-specific socket option which switches off the handshake would
> >disallow this peer recognition.  How bad is that?  I'm not sure.
> 
> Good question.
> 
> >Another potential solution might be to defer the AF_UNIX handshake to
> >the first send/recv:
> >
> >Whatever the peers do, there is a certain protocol used.  That means,
> >there's an implicit understanding who's going to do the first send and
> >who's doing the first recv.  So, after connect/accept, both sides of the
> >sockets go into "connected_but_handshake_missing" mode.  On the first
> >send/recv, the handshake gets started and if it fails, send/recv
> >return ECONNRESET.
> 
> Is an actual handshake really required? It would possibly be sufficient that
> each peer sends its secret+credential and then expects a correct
> secret+credential from the other peer before sending anything.
> 
> After actual connect()/accept():
> 
> send our secret+cred (should not block due to TCP queuing).

So both peers send their credentials...

> if (! nonblocking recv peer secret+cred)
>   set_state(connected_but_secret_missing)
> else
>   set_state(connected)

This will almost always result in connected_but_secret_missing.  It's
probably ok to drop the recv attempt here entirely.

> Before actual send()/recv()/getpeerid():
> 
> if (state == connected_but_secret_missing) {
>   if (! recv peer secret+cred)
>     abort_connection(ECONNRESET)
>   else
>     set_state(connected)
> }

Sounds like a nice idea.  We should try that.  I'm just not sure how
much time I have left to work on this before my vaca next month.  Do you
have fun to look into that?  We have waited so long for postfix, I guess
a couple more weeks won't really hurt.

Otherwise the easy solution you suggested before would be rather quickly
implemented...

> AFAICS this should provide the behavior required for postfix: client
> connect() succeeds before server accept().
> It adds the following unusual behavior: client send() and getpeereid() wait
> for server accept().

Same with recv.  Well, that might be unusual, but in most cases send
recv and getpeereid will be called after a connect/accept.  It's as
much a trade-off as the connect/accept requirement today.  As a resort
we'd still have the "easy" solution removing the credential exchange
entirely.


Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Maintainer                 cygwin AT cygwin DOT com
Red Hat

Attachment: pgpbOMb1zkr2g.pgp
Description: PGP signature