Re: [PATCH V9 4/4] USB: serial: f81232: implement break control

[Ji-Ze Hong (Peter Hong)]

Hi Johan,

Johan Hovold æ 2019/5/3 äå 02:59 åé:
> > @@ -519,6 +530,7 @@ static int f81232_port_disable(struct usb_serial_port *port)
> >   static void f81232_set_termios(struct tty_struct *tty,
> >   		struct usb_serial_port *port, struct ktermios *old_termios)
> >   {
> > +	struct f81232_private *priv = usb_get_serial_port_data(port);
> >   	u8 new_lcr = 0;
> >   	int status = 0;
> >   	speed_t baudrate;
> > @@ -572,11 +584,16 @@ static void f81232_set_termios(struct tty_struct *tty,
> >   		break;
> >   	}
> >   
> > +	mutex_lock(&priv->lock);
> > +
> > +	new_lcr |= (priv->shadow_lcr & UART_LCR_SBC);
> >   	status = f81232_set_register(port, LINE_CONTROL_REGISTER, new_lcr);
> >   	if (status) {
> >   		dev_err(&port->dev, "%s failed to set LCR: %d\n",
> >   			__func__, status);
> >   	}
> > +
> > +	mutex_unlock(&priv->lock);
>
> You forgot to update shadow_lcr here so the line settings would be
> corrupted (e.g set to 5 bit words) whenever a break is signalled.
>
> Did you not test this version of the patch?
>
> I added the missing update, but please test the code that's in my
> usb-next branch in a minute and confirm everything works as expected.

Thanks for mention and patch the lost parts. I had only tested the
break signal and forgot to test full LCR. I'll add it to my test items
and it tested well on new usb-next branch after your fix.

BTW, Should I disable break control when port close ?

Thanks.

--
With Best Regards,
Peter Hong