Hi.

As far as I know, that is a big detail expecting to be
fixed on the modem_run driver. Every time the USB plug
is detached, the driver hangs. And the process can't
be  iniciated again in 1.3.1 (due the fact that only
one registered instance of modem_run is allowed). And
the easy way to get connection again is rebooting.

The way in Linux to know when a device is dettached or
attached is usbdevfs. That implementation is used on
modem_run to catch the events on the interface, right?

The manual page of usb(4) on NetBSD say:

<-snip
USB EVENT INTERFACE
All USB events are reported via the /dev/usb device. 
This devices can be opened for reading and each
read(2) will yield an event record (if something has
happened).  The poll(2) system call can be used to
determine if an event record is available for reading.
...
Note that that there is a separation between device
and device driver events.  A device event is generated
when a physical USB device is attached or detached.  A
single USB device may have zero, one, or many device
drivers associated with it.
->snip

So, this could be associated in some way with
libpusb-bsd.

I don't know about other BSD's, but it has to be done
on a similar way.

When the connection dies: native ppp on BSD works a
lot better than userppp. It is able to reconnect and
set the default route and MTU properly.

Since NetBSD 2.0, there are in-kernel threads
available, but I don't know if is related to pppoa2.

The rest is to elaborate a good rc script that can be
integrated to the /etc/ifconfig.* files, and to solve
a misterious murder (signal 1) of the father or the
son after the boot at starting of pppoa2.

Regards

--
  César Catrián C.
  http://mioficina.cjc.cl

 --- Francois Rogler <[EMAIL PROTECTED]> escribió: 
> 
> Hi Guys,
> 
> I guess I have to give you some explanations.
> I am still the maintainer of the BSD part (yes, it's
> written on the
> website).. Unfortunatly, the last 2 years, I didn't
> do any dev on the
> driver, just a few corrections on docs and scripts.
> For the personal part, I have been moving around
> with no internet
> connection at home for about 2 years. I just settled
> down 1 month ago and
> sent the signup form for the internet connection 3
> days ago ! Hopefuly
> I'll be connected in a few weeks.
> 
> > Do you know if there are any plans in the future
> for further *BSD support
> > in
> > the speedtouch software, as in:
> Right now, there is no big announcement from me.
> 
> > 1. Support for when the connection dies: In Linux
> this was handled by some
> > system call which allowed the modem (USB flavour)
> to do a hardware reset
> > without having to reboot?
> >
> > 2. Any params to modem_run to allow a hardware
> reset of the USB modem? Is
> > that
> > possible?
> >
> > I'm using v 1.2 of the package as a lot of the new
> functionality doesn't
> > work
> > on *BSD. (so the features I'm mentioning may
> already be in 1.3.x, but are
> > not
> > supported yet on *BSD)
> 
> Anyway let's try to handle thoses problems. The open
> points that needs to
> be reviwed are :
> - 1.3 working on every BSD ? if not why ? (btw now
> gmake is mandatory to
> compile the driver)
> - hard reset of the modem (The 2 points you
> mentionned deals with it or
> did I miss something?)
> 
> Is there anyother problem with the driver under BSD
> ? Open and Net still
> use pppoa2 cauz of the lack of good thread library ?
> 
> Anyway, I promise I'll have a look at your request
> very soon.
> 
> > Many thanks,
> >
> > Simon.
> 
> About the futur of the project (BSD part), now as I
> am working, I don't
> have much free time. So if someone is interested in
> taking over that job,
> there is no problem. I can also give him some hints
> and help. Just let me
> know !
> 
> Have Fun
> 
> --
> Francois Rogler

_________________________________________________________
Do You Yahoo!?
Información de Estados Unidos y América Latina, en Yahoo! Noticias.
Visítanos en http://noticias.espanol.yahoo.com

Liste de diffusion modem ALCATEL SpeedTouch USB
Pour se désinscrire : mailto:[EMAIL PROTECTED]

        

Reply via email to