| Message ID | 520646F0.9010801@newflow.co.uk (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Sorry for the late reply, I've been thinking about this for some time and was sad to see it didn't really evoke any sort of discussion :(. On Sat, Aug 10, 2013 at 02:58:08PM +0100, Mark Jackson wrote: > When a UART transmitter is connected to (eg) a RS485 driver, it is > necessary to turn the driver on/off as quickly as possible. This is > best achieved in the serial driver itself (rather than in userspace > where the latency can be quite large). > > This patch allows the GPIO pin to be defined (via DT) that controls > the enabling of the driver at the start of a message, and disables > the driver when the message has been completed. > > Still to do:- > Allow userspace to turn this feature on/off > Do the same for the receiver (useful for 2 wire RS485) I've been wondering about this as well but I have a slightly different situation. On my board the RTS line controls the RS485 transmit/receive direction. I don't know if you've found the Documentation/serial/serial-rs485.txt file but it describes, at the very least, a standard API For controlling several parameters related to RS485 including configurable delay between rts->start of data/end of data->rts. Unfortunately it seems like only one driver really implements the full API (Atmel AT91) and I guess it needs to be bolted onto each serial driver individually (although it seems like a fairly general concept that could be handled at another level). That being said, maybe this patch would better be rethought as a way to specify a GPIO for the RTS line (I don't know enough about OMAP and whether or not it already provides for hardware flow control in its builtin UARTs and you just aren't using it for RS485 flow control?) and then in a separate patch implement this already documented user->kernel API? - Andy -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 13/08/13 21:12, Andrew Ruder wrote: > Sorry for the late reply, I've been thinking about this for some time > and was sad to see it didn't really evoke any sort of discussion :(. > > On Sat, Aug 10, 2013 at 02:58:08PM +0100, Mark Jackson wrote: >> When a UART transmitter is connected to (eg) a RS485 driver, it is >> necessary to turn the driver on/off as quickly as possible. This is >> best achieved in the serial driver itself (rather than in userspace >> where the latency can be quite large). >> >> This patch allows the GPIO pin to be defined (via DT) that controls >> the enabling of the driver at the start of a message, and disables >> the driver when the message has been completed. >> >> Still to do:- >> Allow userspace to turn this feature on/off >> Do the same for the receiver (useful for 2 wire RS485) > > I've been wondering about this as well but I have a slightly different > situation. On my board the RTS line controls the RS485 transmit/receive > direction. > > I don't know if you've found the Documentation/serial/serial-rs485.txt > file but it describes, at the very least, a standard API For controlling > several parameters related to RS485 including configurable delay between > rts->start of data/end of data->rts. Unfortunately it seems like only > one driver really implements the full API (Atmel AT91) and I guess it > needs to be bolted onto each serial driver individually (although it > seems like a fairly general concept that could be handled at another > level). > > That being said, maybe this patch would better be rethought as a way to > specify a GPIO for the RTS line (I don't know enough about OMAP and > whether or not it already provides for hardware flow control in its > builtin UARTs and you just aren't using it for RS485 flow control?) and > then in a separate patch implement this already documented user->kernel > API? I've actually submitted a newer version that does support the documented API. See http://permalink.gmane.org/gmane.linux.ports.arm.omap/102765 Does this address some of your questions ? -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/drivers/tty/serial/omap-serial.c b/drivers/tty/serial/omap-serial.c index b6d1728..1d3d117 100644 --- a/drivers/tty/serial/omap-serial.c +++ b/drivers/tty/serial/omap-serial.c @@ -40,9 +40,12 @@ #include <linux/pm_runtime.h> #include <linux/of.h> #include <linux/gpio.h> +#include <linux/of_gpio.h> #include <linux/pinctrl/consumer.h> #include <linux/platform_data/serial-omap.h> +#include <dt-bindings/gpio/gpio.h> + #define OMAP_MAX_HSUART_PORTS 6 #define UART_BUILD_REVISION(x, y) (((x) << 8) | (y)) @@ -156,6 +159,10 @@ struct uart_omap_port { int DTR_inverted; int DTR_active; + int txen_gpio; + int txen_inverted; + int txen_active; + struct pm_qos_request pm_qos_request; u32 latency; u32 calc_latency; @@ -272,8 +279,25 @@ static void serial_omap_enable_ms(struct uart_port *port) static void serial_omap_stop_tx(struct uart_port *port) { struct uart_omap_port *up = to_uart_omap_port(port); + struct circ_buf *xmit = &up->port.state->xmit; pm_runtime_get_sync(up->dev); + + // if txen currently active + if (up->txen_active) { + // do nothing if current tx not yet completed + int res = serial_in(up, UART_LSR) & UART_LSR_TEMT; + if (!res) + return; + + // if there's no more data to send, turn off txen + if (uart_circ_empty(xmit)) { + gpio_set_value(up->txen_gpio, + up->txen_active == up->txen_inverted); + up->txen_active = 0; + } + } + if (up->ier & UART_IER_THRI) { up->ier &= ~UART_IER_THRI; serial_out(up, UART_IER, up->ier); @@ -300,6 +324,16 @@ static void transmit_chars(struct uart_omap_port *up, unsigned int lsr) struct circ_buf *xmit = &up->port.state->xmit; int count; + // if txen in use + if (gpio_is_valid(up->txen_gpio)) { + // turn on txen ? + if (!uart_circ_empty(xmit) && (!up->txen_active)) { + gpio_set_value(up->txen_gpio, + up->txen_active == up->txen_inverted); + up->txen_active = 1; + } + } + if (up->port.x_char) { serial_out(up, UART_TX, up->port.x_char); up->port.icount.tx++; @@ -1468,6 +1502,28 @@ static int serial_omap_probe(struct platform_device *pdev) goto err_port_line; } + // init tx enable signal + up->txen_gpio = -EINVAL; + up->txen_inverted = 0; + up->txen_active = 0; + if (pdev->dev.of_node) + { + enum of_gpio_flags flags; + up->txen_gpio = of_get_named_gpio_flags(pdev->dev.of_node, + "txen-gpio", 0, &flags); + up->txen_inverted = (flags == GPIO_ACTIVE_LOW); + if (gpio_is_valid(up->txen_gpio)) { + ret = gpio_request(up->txen_gpio, "omap-serial"); + if (ret < 0) + goto err_txen; + ret = gpio_direction_output(up->txen_gpio, + up->txen_inverted); + if (ret < 0) + goto err_txen; + } else + up->txen_gpio = -EINVAL; + } + up->pins = devm_pinctrl_get_select_default(&pdev->dev); if (IS_ERR(up->pins)) { dev_warn(&pdev->dev, "did not get pins for uart%i error: %li\n", @@ -1529,6 +1585,7 @@ err_add_port: pm_runtime_put(&pdev->dev); pm_runtime_disable(&pdev->dev); err_ioremap: +err_txen: err_port_line: dev_err(&pdev->dev, "[UART%d]: failure [%s]: %d\n", pdev->id, __func__, ret);
When a UART transmitter is connected to (eg) a RS485 driver, it is necessary to turn the driver on/off as quickly as possible. This is best achieved in the serial driver itself (rather than in userspace where the latency can be quite large). This patch allows the GPIO pin to be defined (via DT) that controls the enabling of the driver at the start of a message, and disables the driver when the message has been completed. Still to do:- Allow userspace to turn this feature on/off Do the same for the receiver (useful for 2 wire RS485) Signed-off-by: Mark Jackson <mpfj@newflow.co.uk> --- drivers/tty/serial/omap-serial.c | 57 ++++++++++++++++++++++++++++++++++++++ 1 file changed, 57 insertions(+)