This is the mail archive of the
ecos-discuss@sources.redhat.com
mailing list for the eCos project.
Re: server_test dosn't work
- From: Gary Thomas <gthomas at redhat dot com>
- To: Jonathan Larmour <jlarmour at redhat dot com>
- Cc: Roland Caßebohm <roland dot cassebohm at visionsystems dot de>,eCos Discussion <ecos-discuss at sources dot redhat dot com>
- Date: 08 Jan 2002 11:38:03 -0700
- Subject: Re: [ECOS] server_test dosn't work
- References: <200201081725.g08HPXg31178@hhlx01.visionsystems.de> <3C3B3291.FF126A11@redhat.com>
On Tue, 2002-01-08 at 10:55, Jonathan Larmour wrote:
> Roland Caßebohm wrote:
> >
> > I have now a i82559 ethernet card and use redboot on a floppy.
> > This works fine but. Also I can upload an program with gdb and run it.
> >
> > Now I want to run the "server_test" program from
> > "packages/net/tcpip/current/tests":
> > First: shouldn't this line
> > local.sin_port = ntohs(7734);
> > be
> > local.sin_port = htons(7734);
> > ?
>
> Yes you're right! Shows how few of our customers who request network stack
> support have big-endian targets evidently. I'll fix it.
a) This really doesn't matter since ntohs() and htons() are either null
for Big Endian targets, or defined as synonyms on Little Endian ones.
b) network order _is_ Big Endian, thus errors would only make a
difference on Little Endian targets, like the PC.
Changing it doesn't hurt [in fact it makes the usage clearer], but it
also won't change the behaviour.
>
> > After loading and starting the program it gets its ip-address from a
> > dhcp-server and I can ping this ecos station.
> > But I can't connect with a telnet client to the port 7734.
>
> I assume you mean it gives connection refused rather than anything else.
> I'm afraid I don't know what it could be straight away. This has worked
> before....
... and did when I tested it just a few days ago.