serial: fsl_lpuart: Support common rs485 binding for RTS polarity

Invoke the ->rs485_config callback on probe to set UARTMODEM_TXRTSPOL
appropriately based on the UART's device properties.

This implicitly sets UARTMODEM_TXRTSE if rs485 was enabled in the device
properties, so drop the identical code from lpuart_probe().

It also fixes a bug:  If an unsupported rs485 property was specified
(rs485-rx-during-tx or rs485-rts-delay), the driver returns -ENOSYS
without performing any cleanup, in particular without calling
uart_remove_one_port() or clk_disable_unprepare(), thus leaking the
uart_port. But with the invocation of ->rs485_config, the unsupported
properties are now cleared in struct serial_rs485 and thus ignored.
It therefore seems sufficient to just log an error instead of bailing
out.

Cc: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Acked-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This commit is contained in:
Lukas Wunner 2017-11-24 23:26:40 +01:00 committed by Greg Kroah-Hartman
parent f1e5b618c1
commit 01d845352a

View file

@ -2208,21 +2208,14 @@ static int lpuart_probe(struct platform_device *pdev)
uart_get_rs485_mode(&pdev->dev, &sport->port.rs485);
if (sport->port.rs485.flags & SER_RS485_RX_DURING_TX) {
if (sport->port.rs485.flags & SER_RS485_RX_DURING_TX)
dev_err(&pdev->dev, "driver doesn't support RX during TX\n");
return -ENOSYS;
}
if (sport->port.rs485.delay_rts_before_send ||
sport->port.rs485.delay_rts_after_send) {
sport->port.rs485.delay_rts_after_send)
dev_err(&pdev->dev, "driver doesn't support RTS delays\n");
return -ENOSYS;
}
if (sport->port.rs485.flags & SER_RS485_ENABLED) {
sport->port.rs485.flags |= SER_RS485_RTS_ON_SEND;
writeb(UARTMODEM_TXRTSE, sport->port.membase + UARTMODEM);
}
lpuart_config_rs485(&sport->port, &sport->port.rs485);
sport->dma_tx_chan = dma_request_slave_channel(sport->port.dev, "tx");
if (!sport->dma_tx_chan)