Re: tty: n_gsm: fix closing multiplexer mode

[Sascha Hauer]

On Mon, May 22, 2017 at 07:16:17PM +0100, Alan Cox wrote:
> > disconnect frame before initialising multiplex mode. Since it's not so
> > nice that userspace has to know the layout of a disconnect frame,
> > the second patch introduces a disconnect ioctl which can be issued
> > right before closing the physical port during the first session.
> > This ioctl is only useful when during the second session it is known
> > the the first session has been closed properly, so I'm not sure
> > how useful it is to introduce such an ioctl.
> 
> I don't think it justifies an ioctl and the likely use case is that the
> tty is hung up when the process dies so would already be in N_TTY at the
> point anyone tried to clean up.

Ok, so let's drop the second patch.

> 
> A more interesting question - and I don't have enough hardware to test
> this any more - would be whether we can safely send that sequence if our
> initial connects failed a few times so we try disconnect/reconnect.

In my current userspace test program I now unconditionally send the
sequence right before sending the AT+CMUX command. This works fine, no
matter if the modem was still in CMUX mode or not. I can only speak for
a UBLOX modem though, other modems might behave differently here. For
example my UBLOX needs the CLD message to close down the multiplexer,
just closing DLC0 does not work.

Sascha


-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |