NetServerEnum2 from libsmbclient on port 445.

[Christopher R. Hertel]

Stephan Kulow wrote:
> 
> On Tuesday 08 April 2003 14:31, Derrell.Lipman at UnwiredUniverse.com wrote:
> > "Christopher R. Hertel" <crh at ubiqx.mn.org> writes:
> >
> > > This is an interesting area of study.  The NetServerEnum2() function
> > > call is supposed to return a list of NetBIOS names.  SMB on port 445
> > > is NetBIOS-less SMB.  In theory, it makes some sort of sense that a
> > > NetServerEnum2() call via port 445 would not return any NetBIOS server
> > > names.
> >
> > I had this issue in smbwrapper as well.  After some discussion with people on
> > this list, I came to the following solution which has worked well for me so
> > far.
> >
> >         /* Force use of port 139 for first try, so browse lists can work*/
> >         c.port = 139;
> >
> >         /* Attempt to connect */
> >         if (!cli_connect(&c, server_n, &ip)) {
> >                 /*
> >                  * Port 139 connection failed.  Try port 445 to handle
> >                  * connections to newer (e.g. XP) hosts with NetBIOS disabled.
> >                  */
> >                 c.port = 445;
> >                 if (!cli_connect(&c, server_n, &ip)) {
> >                         errno = ENOENT;
> >                         free(username);
> >                         free(workgroup);
> >                         free(password);
> >                         return NULL;
> >                 }
> >         }
> >
> Which arises the question why 445 is tried first. Reading http://www.ubiqx.org/cifs/SMB.html

:)

> the only hint I could find was "Running naked is probably the easiest
> transport test to try first", but I can't really find a reason not to try
> 139 first instead.

Since the NetServerEnum2 call returns a list of NetBIOS names it makes sense
that you'd only try it on the NBT port (unless the user specifically asks
for a different port, eg: smb://workgroup:445/)

Otherwise, I have found that there is a lot of argument regarding the
'right' way to go through the list of options.  Some are quick, but fail
badly, others are difficult to set up.  It's all a matter of taste, really,
so I tried to provide a guide in that section, not a rule.  There's no
really good rule in that case.

Thanks for reading!

Chris -)-----

PS... Can someone else on Team take a look at the patch?  I've pulled two
all-nighters already this week trying to finish my book.

> Otherwise the attached patch might be the cleaner version.
> 
> Greetings, Stephan
> 
> --- libsmb/cliconnect.c 4 Apr 2003 23:39:16 -0000       1.135
> +++ libsmb/cliconnect.c 8 Apr 2003 13:52:25 -0000
> @@ -1109,12 +1109,12 @@ BOOL cli_connect(struct cli_state *cli,
>         if (getenv("LIBSMB_PROG")) {
>                 cli->fd = sock_exec(getenv("LIBSMB_PROG"));
>         } else {
> -               /* try 445 first, then 139 */
> -               int port = cli->port?cli->port:445;
> +               /* try 139 first, then 445 */
> +               int port = cli->port?cli->port:139;
>                 cli->fd = open_socket_out(SOCK_STREAM, &cli->dest_ip,
>                                           port, cli->timeout);
>                 if (cli->fd == -1 && cli->port == 0) {
> -                       port = 139;
> +                       port = 445;
>                         cli->fd = open_socket_out(SOCK_STREAM, &cli->dest_ip,
>                                                   port, cli->timeout);
>                 }

-- 
Samba Team -- http://www.samba.org/     -)-----   Christopher R. Hertel
jCIFS Team -- http://jcifs.samba.org/   -)-----   ubiqx development, uninq.
ubiqx Team -- http://www.ubiqx.org/     -)-----   crh at ubiqx.mn.org
OnLineBook -- http://ubiqx.org/cifs/    -)-----   crh at ubiqx.org