Re: [PATCH 3/3 v6] tty/serial/8250: use mctrl_gpio helpers

[Stefan Roese <sr@xxxxxxx>]

On 17.06.19 11:51, Yegor Yefremov wrote:

<snip>

> > > @@ -1944,11 +1948,15 @@ unsigned int serial8250_do_get_mctrl(struct uart_port *port)
> > >   {
> > >          struct uart_8250_port *up = up_to_u8250p(port);
> > >          unsigned int status;
> > > +       unsigned int val = 0;
> > >
> > >          serial8250_rpm_get(up);
> > >          status = serial8250_modem_status(up);
> > >          serial8250_rpm_put(up);
> > >
> > > +       if (up->gpios)
> > > +               return mctrl_gpio_get(up->gpios, &val);
> > > +
> >
> > What happens when you have a mixed setup i.e. CTS controlled by UART
> > but other status pins controlled by GPIO? In this case CTS status
> > won't be returned. Do I see it right?

Yes, your analysis does seem to be correct. Please note that I did
not intentionally did change it this way. I was not thinking about
such a "mixed design".
 
> What about something like this:
>
> unsigned int serial8250_do_get_mctrl(struct uart_port *port)
>    {
>            struct uart_8250_port *up = up_to_u8250p(port);
>            unsigned int status;
>            unsigned int val;
>
>            serial8250_rpm_get(up);
>            status = serial8250_modem_status(up);
>            serial8250_rpm_put(up);
>
>            val = serial8250_MSR_to_TIOCM(status);
>            if (up->gpios)
>                    mctrl_gpio_get(up->gpios, &val);
>
>            return val;
>    }

Looks good to me, thanks. Do you have such a setup with some modem
control signal handled via GPIO and some via the UART? Could you
test such a change?

Thanks,
Stefan