Re: [ECOS] Problem with TCP Urgent data (old stack).

[Grant Edwards <grante at visi dot com>]

On Fri, Apr 25, 2003 at 08:03:36PM +0100, Jonathan Larmour wrote:

> >>When my eCos app sends a Ctrl-C to telnetd, it kills the
> >>running shell command and inserts a telnet data-mark (0xff
> >>0xf2) as urgent data in the stream.
> >>
> >>The eCos TCP/IP stack only removes the first byte of the
> >>sequence so that my app sees the 0xf2 but not the 0xff.
> > 
> > Setting the OOBINLINE option on the socket fixes the problem,
> > but I'm still curious about whether linux telnetd is only
> > marking the first byte of the data-mark tuple as urgent or if
> > there's an RFC961 dis-connect between the Linux stack and the
> > FreeBSD stack (that's the old one, right?).
> 
> OpenBSD is the old one.

Doh! You'd think I'd be able to remember that by now.

> I would be surprised Linux would diverge from BSD when BSD
> virtually defined the de facto standards!

Me too.

> You could also have a look to see if the FreeBSD stack differs. If it 
> does, you may not have so many qualms about patching the OpenBSD stack.

I'm going to look at the telnetd sources to see if, in fact,
it's only sending the first of the two bytes as urgent data.  I
suspect that all the stacks agree with each other and not with
the RFC.

I have verified that linux telnet code sets the OOBINLINE
option on the socket -- which means that you would get the same
results either way, so it's a moot point.

-- 
Grant Edwards
grante at visi dot com

-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss