Re: [RFC 2/4] tty: fix DTR/RTS not being dropped on hang up

[Peter Hurley <peter-WaGBZJeGNqdsbIuE7sb01tBPR1lH4CV8@public.gmane.org>]

On Wed, 2013-02-13 at 14:32 -0500, Peter Hurley wrote:
> On Wed, 2013-02-13 at 18:27 +0100, Johan Hovold wrote:
> > Move HUPCL handling to port shutdown so that DTR/RTS is dropped also on
> > hang up.
> > 
> > Currently a hung up port will return immediately from
> > tty_port_close_start leaving DTR/RTS unchanged.
> > ---
> >  drivers/tty/tty_port.c | 22 ++++++++++++----------
> >  1 file changed, 12 insertions(+), 10 deletions(-)
> > 
> > diff --git a/drivers/tty/tty_port.c b/drivers/tty/tty_port.c
> > index 57a061e..ffe3689 100644
> > --- a/drivers/tty/tty_port.c
> > +++ b/drivers/tty/tty_port.c
> > @@ -198,11 +198,20 @@ EXPORT_SYMBOL(tty_port_tty_set);
> >  
> >  static void tty_port_shutdown(struct tty_port *port)
> >  {
> > +	struct tty_struct *tty = port->tty;
> > +
> >  	mutex_lock(&port->mutex);
> >  	if (port->console)
> >  		goto out;
> >  
> >  	if (test_and_clear_bit(ASYNCB_INITIALIZED, &port->flags)) {
> > +		/*
> > +		 * Drop DTR/RTS if HUPCL is set. This causes any attached
> > +		 * modem to hang up the line.
> > +		 */
> > +		if (!tty || tty->termios.c_cflag & HUPCL)
> > +			tty_port_lower_dtr_rts(port);
> > +
> 
> port->ops->shutdown() requires the hardware to reset anyway, including
> the DTR/RTS state.

Ok, after reviewing this further, I like this more.

For one, there is more flexibility for tty drivers that have a custom
close() routine (ie, one that doesn't call tty_port_close()). At the
same time, every call site of tty_port_close_start() needs to be checked
to ensure it does not assume (or does not require) DTR to be dropped:

char/pcmcia/synclink_cs.c:	if (tty_port_close_start(port, tty, filp) == 0)
tty/synclinkmp.c:	if (tty_port_close_start(&info->port, tty, filp) == 0)
tty/rocket.c:	if (tty_port_close_start(port, tty, filp) == 0)
tty/amiserial.c:	if (tty_port_close_start(port, tty, filp) == 0)
tty/n_gsm.c:	if (tty_port_close_start(&dlci->port, tty, filp) == 0)
tty/mxser.c:	if (tty_port_close_start(port, tty, filp) == 0)
tty/synclink_gt.c:	if (tty_port_close_start(&info->port, tty, filp) == 0)
tty/serial/serial_core.c:	if (tty_port_close_start(port, tty, filp) == 0)
tty/synclink.c:	if (tty_port_close_start(&info->port, tty, filp) == 0)			 

In the case of the serial core, this actually fixes the code excerpt
below where tty_port_close_start() prematurely drops DTR before the uart
can stop_rx() or wait_until_sent().

static void uart_close(struct tty_struct *tty, struct file *filp)
{
	struct uart_state *state = tty->driver_data;
	struct tty_port *port;
	struct uart_port *uport;
	unsigned long flags;

	if (!state)
		return;

	uport = state->uart_port;
	port = &state->port;

	pr_debug("uart_close(%d) called\n", uport->line);

	if (tty_port_close_start(port, tty, filp) == 0)
		return;

	/*
	 * At this point, we stop accepting input.  To do this, we
	 * disable the receive line status interrupts.
	 */
	if (port->flags & ASYNC_INITIALIZED) {
		unsigned long flags;
		spin_lock_irqsave(&uport->lock, flags);
		uport->ops->stop_rx(uport);
		spin_unlock_irqrestore(&uport->lock, flags);
		/*
		 * Before we drop DTR, make sure the UART transmitter
		 * has completely drained; this is especially
		 * important if there is a transmit FIFO!
		 */
		uart_wait_until_sent(tty, uport->timeout);
	}

	mutex_lock(&port->mutex);
	uart_shutdown(tty, state);
	uart_flush_buffer(tty);


Regards,
Peter Hurley

--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html