linux-kernel - Re: [PATCH v3 3/5] serial: 8250: Support separate rs485 rx-enable GPIO

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Mon, 18 May 2020 06:50:06 +0200
From:   Lukas Wunner <lukas@...ner.de>
To:     Heiko Stuebner <heiko@...ech.de>
Cc:     gregkh@...uxfoundation.org, jslaby@...e.com,
        andriy.shevchenko@...ux.intel.com, matwey.kornilov@...il.com,
        giulio.benetti@...ronovasrl.com, linux-serial@...r.kernel.org,
        linux-kernel@...r.kernel.org,
        christoph.muellner@...obroma-systems.com,
        Heiko Stuebner <heiko.stuebner@...obroma-systems.com>
Subject: Re: [PATCH v3 3/5] serial: 8250: Support separate rs485 rx-enable
 GPIO

On Sun, May 17, 2020 at 11:56:08PM +0200, Heiko Stuebner wrote:
> @@ -1457,6 +1458,7 @@ void serial8250_em485_stop_tx(struct uart_8250_port *p)
>  	 * Enable previously disabled RX interrupts.
>  	 */
>  	if (!(p->port.rs485.flags & SER_RS485_RX_DURING_TX)) {
> +		gpiod_set_value(port->rs485_re_gpio, 1);
>  		serial8250_clear_and_reinit_fifos(p);
>  
>  		p->ier |= UART_IER_RLSI | UART_IER_RDI;

The added line needs to be conditional on if (port->rs485_re_gpio)
because the gpiod could be NULL and gpiod_set_value() doesn't check
for that.

> @@ -1597,9 +1599,12 @@ static inline void __start_tx(struct uart_port *port)
>  void serial8250_em485_start_tx(struct uart_8250_port *up)
>  {
>  	unsigned char mcr = serial8250_in_MCR(up);
> +	struct uart_port *port = &up->port;
>  
> -	if (!(up->port.rs485.flags & SER_RS485_RX_DURING_TX))
> +	if (!(up->port.rs485.flags & SER_RS485_RX_DURING_TX)) {
> +		gpiod_set_value(port->rs485_re_gpio, 0);
>  		serial8250_stop_rx(&up->port);
> +	}

Same here.

> --- a/include/linux/serial_core.h
> +++ b/include/linux/serial_core.h
> @@ -253,6 +253,7 @@ struct uart_port {
>  	const struct attribute_group **tty_groups;	/* all attributes (serial core use only) */
>  	struct serial_rs485     rs485;
>  	struct gpio_desc	*rs485_term_gpio;	/* enable RS485 bus termination */
> +	struct gpio_desc	*rs485_re_gpio;		/* gpio RS485 receive enable */

Nit: I'd probably document this as "enable RS485 receiver" because it's
already apparent from the variable type and name that it's a gpio,
making it unnecessary to repeat that in the code comment.  But I guess
that's a matter of personal preference.

There's something else:  You need to amend serial8250_em485_config()
to toggle the GPIO depending on whether SER_RS485_RX_DURING_TX is
set.  Right now you enable the receiver by default and then disable
it when starting to transmit if half-duplex mode is selected and
likewise re-enable it when stopping to transmit.  But user space
may write some stuff to the tty while in half-duplex mode, then
immediately issue a TIOCSRS485 ioctl to switch to full-duplex mode.
If the ->rs485_config callback is executed while transmitting is
still ongoing, then you'll not re-enable the receiver when transmitting
finally stops.  The ->rs485_config callback is invoked under the
uart port spinlock but the lock may be briefly released and later
re-acquired by the IRQ handler if the TX FIFO is full.  (Unless
I'm missing something.)

Thanks,

Lukas