[Oberon] Ctrl/Break, tasking, ser-port ?

eas-lab at absamail.co.za eas-lab at absamail.co.za
Wed Apr 9 04:43:41 CEST 2003


Hi,
  I've invested considerable effort in chasing this bug and would like
to complete the war  soon !!

There's a complex interaction between a failed/aborted ppp connection,
the next attempt at ppp (which calls the dialer > V24) and Ctrl/Break.

I need to know how critically timed ( needs to be repeated several
times until it 'hits the target') Ctrl/Break could 'unlock the bad-state'
and allow ppp/dial to work.
( The symptom is that the modem lights activate but no dial tone.
eg. as if the bytes sent were bad. )

Is the source for Ctrl/Break response available ?
Is it in the kernel ?

-- Chris Glur.

PS.

Extract from NewsGroup comp.protocols.ppp:

> > That's almost what I suspect. Except that no other task is
> > supposed to use the 2nd serial port. I believe that with a good
> > ppp 'flight' the ser-port 'lands' after it's finnished.
> > But when the connection aborts abnormally, the ser-port
> > doesn't land - ready for the next take-off.  It stays 'up'.
> 
> Then that narrows the search.  It's the system-dependent code in your
> PPP implementation that opens the serial port for use with PPP that's
> most likely to be broken.

{BTW the 'man who wrote the book' contributes to this NewsGroup,
so I'm taking advantage to fully understand ppp and will try to write
a minimalist/Wirthian description. }

PSS. I never imagined that using the serial port was so complicated:
Installing and de-installing interrupts and checking for various
chip families and ....!   
The code base of n-o represents a big investment for humanity.







More information about the Oberon mailing list