Re: [PATCH] Documentation about RS485 serial communications

Re: [PATCH] Documentation about RS485 serial communications

[Claudio Scordino]

Russell King - ARM Linux ha scritto:
> On Sun, Aug 15, 2010 at 03:02:57PM -0700, Randy Dunlap wrote:
>> On 08/14/10 05:50, Claudio Scordino wrote:
>>> diff --git a/Documentation/serial/serial-rs485.txt b/Documentation/serial/serial-rs485.txt
>>> new file mode 100644
>>> index 0000000..93b029e
>>> --- /dev/null
>>> +++ b/Documentation/serial/serial-rs485.txt
>>> @@ -0,0 +1,126 @@
>>> +                        RS485 SERIAL COMMUNICATIONS
>>> +
>>> +1. INTRODUCTION
>>> +
>>> +   EIA-485, also known as TIA/EIA-485 or RS-485, is a standard defining the
>>> +   electrical characteristics of drivers and receivers for use in balanced
>>> +   digital multipoint systems.
>>> +   This standard is widely used for communications in industrial automation
>>> +   because it can be used effectively over long distances and in electrically
>>> +   noisy environments.
>>> +   Even though the data is transmitted over a 2-wire twisted pair bus, all
>>> +   EIA-485 transceivers interpret the voltage levels of the differential
>>> +   signals with respect to a third common voltage. Without this common
>>> +   reference, a set of transceivers may interpret the differential signals
>>> +   incorrectly.
>>> +   See [1] for more information.
> 
> There are devices on the market which are fully isolating RS485
> transceivers which just take the A/B connections, measuring the
> differential voltage, and provide you with the TXD/RXD TTL signals
> for your UART.
> 
> Also note that [1] appears a little confused about the number of pins
> required for RS485 - it says two, and then lists three pins.  I don't
> think you can use this as being supportive of the requirement for three
> connections - and as there are these fully isolating transceivers...
> 
>>> +4. USAGE FROM USER-LEVEL
>>> +
>>> +   From user-level, RS485 configuration can be get/set using the previous
>>> +   ioctls. For instance, to set RS485 you can use the following code:
>>> +
>>> +	#include <linux/serial.h>
>>> +
>>> +	/* Driver-specific ioctls: */
>>> +	#define TIOCGRS485      0x542E
>>> +	#define TIOCSRS485      0x542F
>>> +
>>> +	/* Open your specific device (e.g., /dev/mydevice): */
>>> +	int fd = open ("/dev/mydevice", O_RDWR);
>>> +	struct serial_rs485 rs485conf;
>>> +
>>> +	/* Set RS485 mode: */
>>> +	rs485conf.flags |= SER_RS485_ENABLED;
>>> +
>>> +	/* Set rts delay before send, if needed: */
>>> +	rs485conf.flags |= SER_RS485_RTS_BEFORE_SEND;
>>> +	rs485conf.delay_rts_before_send = ...;
>>> +
>>> +	/* Set rts delay after send, if needed: */
>>> +	rs485conf.flags |= SER_RS485_RTS_AFTER_SEND;
>>> +	rs485conf.delay_rts_after_send = ...;
>>> +
>>> +	ioctl (fd, TIOCSRS485, &rs485conf);
> 
> Example code really should do things right, such as check for error
> conditions - otherwise people will copy'n'paste this and assume that
> the ioctl never fails.

Here is the new document containing the changed you've proposed.

If OK, somebody please provide for commit (the merge window is opening...).

Cheers,

	Claudio


Documentation about RS485 serial communications

Signed-off-by: Claudio Scordino <claudio@evidence.eu.com>
---
 Documentation/serial/00-INDEX         |    2 +
 Documentation/serial/serial-rs485.txt |  126 +++++++++++++++++++++++++++++++++
 2 files changed, 128 insertions(+), 0 deletions(-)
 create mode 100644 Documentation/serial/serial-rs485.txt

diff --git a/Documentation/serial/00-INDEX b/Documentation/serial/00-INDEX
index 07dcdb0..e09468a 100644
--- a/Documentation/serial/00-INDEX
+++ b/Documentation/serial/00-INDEX
@@ -14,6 +14,8 @@ riscom8.txt
 	- notes on using the RISCom/8 multi-port serial driver.
 rocket.txt
 	- info on the Comtrol RocketPort multiport serial driver.
+serial-rs485.txt
+	- info about RS485 structures and support in the kernel.
 specialix.txt
 	- info on hardware/driver for specialix IO8+ multiport serial card.
 stallion.txt
diff --git a/Documentation/serial/serial-rs485.txt b/Documentation/serial/serial-rs485.txt
new file mode 100644
index 0000000..82ebc86
--- /dev/null
+++ b/Documentation/serial/serial-rs485.txt
@@ -0,0 +1,126 @@
+                        RS485 SERIAL COMMUNICATIONS
+
+1. INTRODUCTION
+
+   EIA-485, also known as TIA/EIA-485 or RS-485, is a standard defining the
+   electrical characteristics of drivers and receivers for use in balanced
+   digital multipoint systems.
+   This standard is widely used for communications in industrial automation
+   because it can be used effectively over long distances and in electrically
+   noisy environments.
+
+2. HARDWARE-RELATED CONSIDERATIONS
+
+   Some CPUs (e.g., Atmel AT91) contain a transceiver capable of working both
+   as RS232 and RS485. For these microcontrollers, the Linux driver should be
+   made capable of working in both modes, and proper ioctls (see later) should
+   be made available at user-level to allow switching from one mode to the
+   other, and vice versa.
+
+   On some other CPUs (e.g., Freescale imx25) the RS485 transceiver is not
+   integrated inside the microcontroller itself. Therefore, manufacturers who
+   use these microcontrollers to produce embedded boards need to connect an
+   external transceiver to some pin of the CPU.
+   On these architectures, therefore, no assumptions can be done at the
+   CPU-level about the presence of a RS485 transceiver, because the connection
+   (if any) is done outside the microcontroller. Moreover, even in case of
+   RS485 transceiver, the manufacturer is free to choose the CPU pin used for
+   the connection.
+
+
+3. DATA STRUCTURES ALREADY AVAILABLE IN THE KERNEL
+
+   The Linux kernel provides the serial_rs485 structure (see [1]) to handle
+   RS485 communications. This data structure is used to set and configure RS485
+   parameters in the platform data and in ioctls.
+
+   Any driver for devices capable of working both as RS232 and RS485 should
+   provide at least the following ioctls:
+
+    - TIOCSRS485 (typically associated with number 0x542F). This ioctl is used
+      to enable/disable RS485 mode from user-space
+
+    - TIOCGRS485 (typically associated with number 0x542E). This ioctl is used
+      to get RS485 mode from kernel-space (i.e., driver) to user-space.
+
+   In other words, the serial driver should contain a code similar to the next
+   one:
+
+	static struct uart_ops atmel_pops = {
+		/* ... */
+		.ioctl		= handle_ioctl,
+	};
+
+	static int handle_ioctl(struct uart_port *port,
+		unsigned int cmd,
+		unsigned long arg)
+	{
+		struct serial_rs485 rs485conf;
+
+		switch (cmd) {
+		case TIOCSRS485:
+			if (copy_from_user(&rs485conf,
+				(struct serial_rs485 *) arg,
+				sizeof(rs485conf)))
+					return -EFAULT;
+
+			/* ... */
+			break;
+
+		case TIOCGRS485:
+			if (copy_to_user((struct serial_rs485 *) arg,
+				...,
+				sizeof(rs485conf)))
+					return -EFAULT;
+			/* ... */
+			break;
+
+		/* ... */
+		}
+	}
+
+
+4. USAGE FROM USER-LEVEL
+
+   From user-level, RS485 configuration can be get/set using the previous
+   ioctls. For instance, to set RS485 you can use the following code:
+
+	#include <linux/serial.h>
+
+	/* Driver-specific ioctls: */
+	#define TIOCGRS485      0x542E
+	#define TIOCSRS485      0x542F
+
+	/* Open your specific device (e.g., /dev/mydevice): */
+	int fd = open ("/dev/mydevice", O_RDWR);
+	if (fd < 0) {
+		/* Error handling. See errno. */
+	}
+
+	struct serial_rs485 rs485conf;
+
+	/* Set RS485 mode: */
+	rs485conf.flags |= SER_RS485_ENABLED;
+
+	/* Set rts delay before send, if needed: */
+	rs485conf.flags |= SER_RS485_RTS_BEFORE_SEND;
+	rs485conf.delay_rts_before_send = ...;
+
+	/* Set rts delay after send, if needed: */
+	rs485conf.flags |= SER_RS485_RTS_AFTER_SEND;
+	rs485conf.delay_rts_after_send = ...;
+
+	if (ioctl (fd, TIOCSRS485, &rs485conf) < 0) {
+		/* Error handling. See errno. */
+	}
+
+	/* Use read() and write() syscalls here... */
+
+	/* Close the device when finished: */
+	if (close (fd) < 0) {
+		/* Error handling. See errno. */
+	}
+
+5. REFERENCES
+
+ [1]	include/linux/serial.h

Re: [PATCH] Documentation about RS485 serial communications

[Grant Edwards]

On 2010-10-19, Claudio Scordino <claudio@evidence.eu.com> wrote:

> +2. HARDWARE-RELATED CONSIDERATIONS
> +
> +   Some CPUs (e.g., Atmel AT91) contain a transceiver capable of working both
> +   as RS232 and RS485.

Pardon my curiosity, but I've looked at a few AT91 parts, and I've
never seen on with an internal transceiver -- nor does Google seem
able to find any.

Frankly, I'd be pretty surprised, since it would be very difficult
(fab-wise), to build a mircoprocessor with I/O cells capable of
withstanding the voltage levels specified in RS232.

Can you specify to which parts you're referring?

-- 
Grant Edwards               grant.b.edwards        Yow! I have the power to
                                  at               HALT PRODUCTION on all
                              gmail.com            TEENAGE SEX COMEDIES!!

Re: [PATCH] Documentation about RS485 serial communications

[Alexander Stein]

On Tuesday 19 October 2010, 16:29:48 Grant Edwards wrote:
> On 2010-10-19, Claudio Scordino <claudio@evidence.eu.com> wrote:
> > +2. HARDWARE-RELATED CONSIDERATIONS
> > +
> > +   Some CPUs (e.g., Atmel AT91) contain a transceiver capable of working
> > both +   as RS232 and RS485.
> 
> Pardon my curiosity, but I've looked at a few AT91 parts, and I've
> never seen on with an internal transceiver -- nor does Google seem
> able to find any.
> 
> Frankly, I'd be pretty surprised, since it would be very difficult
> (fab-wise), to build a mircoprocessor with I/O cells capable of
> withstanding the voltage levels specified in RS232.
> 
> Can you specify to which parts you're referring?

I doubt the AT91 do have an internal transceiver but it has a dedicated pin 
feature (RTS or CTS or something) which can be used to switch between RS232 
and RS485 by setting a bit in the USART provided the pin is connected 
properly.

Regards,
Alexander

Re: [PATCH] Documentation about RS485 serial communications

[Grant Edwards]

On 2010-10-19, Alexander Stein <alexander.stein@systec-electronic.com> wrote:
> On Tuesday 19 October 2010, 16:29:48 Grant Edwards wrote:
>> On 2010-10-19, Claudio Scordino <claudio@evidence.eu.com> wrote:
>> > +2. HARDWARE-RELATED CONSIDERATIONS
>> > +
>> > +   Some CPUs (e.g., Atmel AT91) contain a transceiver capable of working
>> > both +   as RS232 and RS485.
>> 
>> Pardon my curiosity, but I've looked at a few AT91 parts, and I've
>> never seen on with an internal transceiver -- nor does Google seem
>> able to find any.
>> 
>> Frankly, I'd be pretty surprised, since it would be very difficult
>> (fab-wise), to build a mircoprocessor with I/O cells capable of
>> withstanding the voltage levels specified in RS232.
>> 
>> Can you specify to which parts you're referring?
>
> I doubt the AT91 do have an internal transceiver but it has a
> dedicated pin feature (RTS or CTS or something) which can be used to
> switch between RS232 and RS485 by setting a bit in the USART provided
> the pin is connected properly.

Yes, the AT91 UARTs do have a built-in half-duplex mode that will
automatically control line direction by toggling RTS.

That can used to control external half-duplex hardware like an RS485
transceiver or any RS232-connected half-duplex device like some
modems.

Though the Atmel docs refer to it as "RS485 mode" there is no RS485
transceiver in the AT91 parts (AFAIK), and "RS485 mode" isn't specific
to RS485 - it can be used with any external half-duplex hardware.

-- 
Grant Edwards               grant.b.edwards        Yow! I Know A Joke!!
                                  at               
                              gmail.com            

Re: [PATCH] Documentation about RS485 serial communications

[Alexander Stein]

On Tuesday 19 October 2010, 19:22:32 Grant Edwards wrote:
> Yes, the AT91 UARTs do have a built-in half-duplex mode that will
> automatically control line direction by toggling RTS.
> 
> That can used to control external half-duplex hardware like an RS485
> transceiver or any RS232-connected half-duplex device like some
> modems.
> 
> Though the Atmel docs refer to it as "RS485 mode" there is no RS485
> transceiver in the AT91 parts (AFAIK), and "RS485 mode" isn't specific
> to RS485 - it can be used with any external half-duplex hardware.

Out of curiosity, how can I add RS485 (or whatever) support for other serial 
devices like e.g. i.mx serial? Is there some driver independent interface 
which only gets filled in board configs like setting some GPIO pins?

Alexander

Re: [PATCH] Documentation about RS485 serial communications

[Alan Cox]

On Wed, 20 Oct 2010 10:00:14 +0200
Alexander Stein <alexander.stein@systec-electronic.com> wrote:

> On Tuesday 19 October 2010, 19:22:32 Grant Edwards wrote:
> > Yes, the AT91 UARTs do have a built-in half-duplex mode that will
> > automatically control line direction by toggling RTS.
> > 
> > That can used to control external half-duplex hardware like an RS485
> > transceiver or any RS232-connected half-duplex device like some
> > modems.
> > 
> > Though the Atmel docs refer to it as "RS485 mode" there is no RS485
> > transceiver in the AT91 parts (AFAIK), and "RS485 mode" isn't specific
> > to RS485 - it can be used with any external half-duplex hardware.
> 
> Out of curiosity, how can I add RS485 (or whatever) support for other serial 
> devices like e.g. i.mx serial? Is there some driver independent interface 
> which only gets filled in board configs like setting some GPIO pins?

Several devices support RS485 and the like

There are two interfaces

1.	termiox	which is used in the old Unix world for sync and
other stuff on serial world. Linux now supports a bit of this interface
and has the flags so we can grow it as needed. It covers one way flow
control

2.	TIOCGSRS485/TIOCSRS486 plus struct serial_rs485 which is Linux
specific but also allows for pre/post send delays in hardware that some
chips can do.

And nothing stops you supporting both interfaces when it makes sense

Re: [PATCH] Documentation about RS485 serial communications

[Claudio Scordino]

Grant Edwards ha scritto:
> On 2010-10-19, Alexander Stein <alexander.stein@systec-electronic.com> wrote:
>> On Tuesday 19 October 2010, 16:29:48 Grant Edwards wrote:
>>> On 2010-10-19, Claudio Scordino <claudio@evidence.eu.com> wrote:
>>>> +2. HARDWARE-RELATED CONSIDERATIONS
>>>> +
>>>> +   Some CPUs (e.g., Atmel AT91) contain a transceiver capable of working
>>>> both +   as RS232 and RS485.
>>> Pardon my curiosity, but I've looked at a few AT91 parts, and I've
>>> never seen on with an internal transceiver -- nor does Google seem
>>> able to find any.
>>>
>>> Frankly, I'd be pretty surprised, since it would be very difficult
>>> (fab-wise), to build a mircoprocessor with I/O cells capable of
>>> withstanding the voltage levels specified in RS232.
>>>
>>> Can you specify to which parts you're referring?
>> I doubt the AT91 do have an internal transceiver but it has a
>> dedicated pin feature (RTS or CTS or something) which can be used to
>> switch between RS232 and RS485 by setting a bit in the USART provided
>> the pin is connected properly.
> 
> Yes, the AT91 UARTs do have a built-in half-duplex mode that will
> automatically control line direction by toggling RTS.
> 
> That can used to control external half-duplex hardware like an RS485
> transceiver or any RS232-connected half-duplex device like some
> modems.
> 
> Though the Atmel docs refer to it as "RS485 mode" there is no RS485
> transceiver in the AT91 parts (AFAIK), and "RS485 mode" isn't specific
> to RS485 - it can be used with any external half-duplex hardware.

Hi all,

	sorry for the misuse of the term "transceiver", and thank you
for your corrections.

Hopefully, this new version of the document is slightly better.

Many thanks,

	Claudio	


Documentation about RS485 serial communications

Signed-off-by: Claudio Scordino <claudio@evidence.eu.com>
---
 Documentation/serial/00-INDEX         |    2 +
 Documentation/serial/serial-rs485.txt |  119 +++++++++++++++++++++++++++++++++
 2 files changed, 121 insertions(+), 0 deletions(-)
 create mode 100644 Documentation/serial/serial-rs485.txt

diff --git a/Documentation/serial/00-INDEX b/Documentation/serial/00-INDEX
index 07dcdb0..e09468a 100644
--- a/Documentation/serial/00-INDEX
+++ b/Documentation/serial/00-INDEX
@@ -14,6 +14,8 @@ riscom8.txt
 	- notes on using the RISCom/8 multi-port serial driver.
 rocket.txt
 	- info on the Comtrol RocketPort multiport serial driver.
+serial-rs485.txt
+	- info about RS485 structures and support in the kernel.
 specialix.txt
 	- info on hardware/driver for specialix IO8+ multiport serial card.
 stallion.txt
diff --git a/Documentation/serial/serial-rs485.txt b/Documentation/serial/serial-rs485.txt
new file mode 100644
index 0000000..40f09c6
--- /dev/null
+++ b/Documentation/serial/serial-rs485.txt
@@ -0,0 +1,119 @@
+                        RS485 SERIAL COMMUNICATIONS
+
+1. INTRODUCTION
+
+   EIA-485, also known as TIA/EIA-485 or RS-485, is a standard defining the
+   electrical characteristics of drivers and receivers for use in balanced
+   digital multipoint systems.
+   This standard is widely used for communications in industrial automation
+   because it can be used effectively over long distances and in electrically
+   noisy environments.
+
+2. HARDWARE-RELATED CONSIDERATIONS
+
+   Some CPUs (e.g., Atmel AT91) contain a built-in half-duplex mode capable of
+   automatically controlling line direction by toggling RTS. That can used to
+   control external half-duplex hardware like an RS485 transceiver or any
+   RS232-connected half-duplex device like some modems.
+
+   For these microcontrollers, the Linux driver should be made capable of
+   working in both modes, and proper ioctls (see later) should be made
+   available at user-level to allow switching from one mode to the other, and
+   vice versa.
+
+3. DATA STRUCTURES ALREADY AVAILABLE IN THE KERNEL
+
+   The Linux kernel provides the serial_rs485 structure (see [1]) to handle
+   RS485 communications. This data structure is used to set and configure RS485
+   parameters in the platform data and in ioctls.
+
+   Any driver for devices capable of working both as RS232 and RS485 should
+   provide at least the following ioctls:
+
+    - TIOCSRS485 (typically associated with number 0x542F). This ioctl is used
+      to enable/disable RS485 mode from user-space
+
+    - TIOCGRS485 (typically associated with number 0x542E). This ioctl is used
+      to get RS485 mode from kernel-space (i.e., driver) to user-space.
+
+   In other words, the serial driver should contain a code similar to the next
+   one:
+
+	static struct uart_ops atmel_pops = {
+		/* ... */
+		.ioctl		= handle_ioctl,
+	};
+
+	static int handle_ioctl(struct uart_port *port,
+		unsigned int cmd,
+		unsigned long arg)
+	{
+		struct serial_rs485 rs485conf;
+
+		switch (cmd) {
+		case TIOCSRS485:
+			if (copy_from_user(&rs485conf,
+				(struct serial_rs485 *) arg,
+				sizeof(rs485conf)))
+					return -EFAULT;
+
+			/* ... */
+			break;
+
+		case TIOCGRS485:
+			if (copy_to_user((struct serial_rs485 *) arg,
+				...,
+				sizeof(rs485conf)))
+					return -EFAULT;
+			/* ... */
+			break;
+
+		/* ... */
+		}
+	}
+
+
+4. USAGE FROM USER-LEVEL
+
+   From user-level, RS485 configuration can be get/set using the previous
+   ioctls. For instance, to set RS485 you can use the following code:
+
+	#include <linux/serial.h>
+
+	/* Driver-specific ioctls: */
+	#define TIOCGRS485      0x542E
+	#define TIOCSRS485      0x542F
+
+	/* Open your specific device (e.g., /dev/mydevice): */
+	int fd = open ("/dev/mydevice", O_RDWR);
+	if (fd < 0) {
+		/* Error handling. See errno. */
+	}
+
+	struct serial_rs485 rs485conf;
+
+	/* Set RS485 mode: */
+	rs485conf.flags |= SER_RS485_ENABLED;
+
+	/* Set rts delay before send, if needed: */
+	rs485conf.flags |= SER_RS485_RTS_BEFORE_SEND;
+	rs485conf.delay_rts_before_send = ...;
+
+	/* Set rts delay after send, if needed: */
+	rs485conf.flags |= SER_RS485_RTS_AFTER_SEND;
+	rs485conf.delay_rts_after_send = ...;
+
+	if (ioctl (fd, TIOCSRS485, &rs485conf) < 0) {
+		/* Error handling. See errno. */
+	}
+
+	/* Use read() and write() syscalls here... */
+
+	/* Close the device when finished: */
+	if (close (fd) < 0) {
+		/* Error handling. See errno. */
+	}
+
+5. REFERENCES
+
+ [1]	include/linux/serial.h
-- 
1.6.0.4

Re: [PATCH] Documentation about RS485 serial communications

[Nicolas Ferre]

Le 24/10/2010 13:29, Claudio Scordino :

[..]

> Hi all,
> 
> 	sorry for the misuse of the term "transceiver", and thank you
> for your corrections.
> 
> Hopefully, this new version of the document is slightly better.

Hi all,

This documentation is floating around for some time now. Claudio has
enhanced it and it is maybe time to merge it in mainline...

So, can serial people include it in their tree (Greg?) or should it go
through another path?

Thanks to all of you.

Best regards,

> Many thanks,
> 
> 	Claudio	
> 
> 
> Documentation about RS485 serial communications
> 
> Signed-off-by: Claudio Scordino <claudio@evidence.eu.com>
> ---
>  Documentation/serial/00-INDEX         |    2 +
>  Documentation/serial/serial-rs485.txt |  119 +++++++++++++++++++++++++++++++++
>  2 files changed, 121 insertions(+), 0 deletions(-)
>  create mode 100644 Documentation/serial/serial-rs485.txt
> 
> diff --git a/Documentation/serial/00-INDEX b/Documentation/serial/00-INDEX
> index 07dcdb0..e09468a 100644
> --- a/Documentation/serial/00-INDEX
> +++ b/Documentation/serial/00-INDEX
> @@ -14,6 +14,8 @@ riscom8.txt
>  	- notes on using the RISCom/8 multi-port serial driver.
>  rocket.txt
>  	- info on the Comtrol RocketPort multiport serial driver.
> +serial-rs485.txt
> +	- info about RS485 structures and support in the kernel.
>  specialix.txt
>  	- info on hardware/driver for specialix IO8+ multiport serial card.
>  stallion.txt
> diff --git a/Documentation/serial/serial-rs485.txt b/Documentation/serial/serial-rs485.txt
> new file mode 100644
> index 0000000..40f09c6
> --- /dev/null
> +++ b/Documentation/serial/serial-rs485.txt
> @@ -0,0 +1,119 @@
> +                        RS485 SERIAL COMMUNICATIONS
> +
> +1. INTRODUCTION
> +
> +   EIA-485, also known as TIA/EIA-485 or RS-485, is a standard defining the
> +   electrical characteristics of drivers and receivers for use in balanced
> +   digital multipoint systems.
> +   This standard is widely used for communications in industrial automation
> +   because it can be used effectively over long distances and in electrically
> +   noisy environments.
> +
> +2. HARDWARE-RELATED CONSIDERATIONS
> +
> +   Some CPUs (e.g., Atmel AT91) contain a built-in half-duplex mode capable of
> +   automatically controlling line direction by toggling RTS. That can used to
> +   control external half-duplex hardware like an RS485 transceiver or any
> +   RS232-connected half-duplex device like some modems.
> +
> +   For these microcontrollers, the Linux driver should be made capable of
> +   working in both modes, and proper ioctls (see later) should be made
> +   available at user-level to allow switching from one mode to the other, and
> +   vice versa.
> +
> +3. DATA STRUCTURES ALREADY AVAILABLE IN THE KERNEL
> +
> +   The Linux kernel provides the serial_rs485 structure (see [1]) to handle
> +   RS485 communications. This data structure is used to set and configure RS485
> +   parameters in the platform data and in ioctls.
> +
> +   Any driver for devices capable of working both as RS232 and RS485 should
> +   provide at least the following ioctls:
> +
> +    - TIOCSRS485 (typically associated with number 0x542F). This ioctl is used
> +      to enable/disable RS485 mode from user-space
> +
> +    - TIOCGRS485 (typically associated with number 0x542E). This ioctl is used
> +      to get RS485 mode from kernel-space (i.e., driver) to user-space.
> +
> +   In other words, the serial driver should contain a code similar to the next
> +   one:
> +
> +	static struct uart_ops atmel_pops = {
> +		/* ... */
> +		.ioctl		= handle_ioctl,
> +	};
> +
> +	static int handle_ioctl(struct uart_port *port,
> +		unsigned int cmd,
> +		unsigned long arg)
> +	{
> +		struct serial_rs485 rs485conf;
> +
> +		switch (cmd) {
> +		case TIOCSRS485:
> +			if (copy_from_user(&rs485conf,
> +				(struct serial_rs485 *) arg,
> +				sizeof(rs485conf)))
> +					return -EFAULT;
> +
> +			/* ... */
> +			break;
> +
> +		case TIOCGRS485:
> +			if (copy_to_user((struct serial_rs485 *) arg,
> +				...,
> +				sizeof(rs485conf)))
> +					return -EFAULT;
> +			/* ... */
> +			break;
> +
> +		/* ... */
> +		}
> +	}
> +
> +
> +4. USAGE FROM USER-LEVEL
> +
> +   From user-level, RS485 configuration can be get/set using the previous
> +   ioctls. For instance, to set RS485 you can use the following code:
> +
> +	#include <linux/serial.h>
> +
> +	/* Driver-specific ioctls: */
> +	#define TIOCGRS485      0x542E
> +	#define TIOCSRS485      0x542F
> +
> +	/* Open your specific device (e.g., /dev/mydevice): */
> +	int fd = open ("/dev/mydevice", O_RDWR);
> +	if (fd < 0) {
> +		/* Error handling. See errno. */
> +	}
> +
> +	struct serial_rs485 rs485conf;
> +
> +	/* Set RS485 mode: */
> +	rs485conf.flags |= SER_RS485_ENABLED;
> +
> +	/* Set rts delay before send, if needed: */
> +	rs485conf.flags |= SER_RS485_RTS_BEFORE_SEND;
> +	rs485conf.delay_rts_before_send = ...;
> +
> +	/* Set rts delay after send, if needed: */
> +	rs485conf.flags |= SER_RS485_RTS_AFTER_SEND;
> +	rs485conf.delay_rts_after_send = ...;
> +
> +	if (ioctl (fd, TIOCSRS485, &rs485conf) < 0) {
> +		/* Error handling. See errno. */
> +	}
> +
> +	/* Use read() and write() syscalls here... */
> +
> +	/* Close the device when finished: */
> +	if (close (fd) < 0) {
> +		/* Error handling. See errno. */
> +	}
> +
> +5. REFERENCES
> +
> + [1]	include/linux/serial.h


-- 
Nicolas Ferre

Re: [PATCH] Documentation about RS485 serial communications

[Greg KH]

On Wed, Nov 10, 2010 at 10:17:03AM +0100, Nicolas Ferre wrote:
> Le 24/10/2010 13:29, Claudio Scordino :
> 
> [..]
> 
> > Hi all,
> > 
> > 	sorry for the misuse of the term "transceiver", and thank you
> > for your corrections.
> > 
> > Hopefully, this new version of the document is slightly better.
> 
> Hi all,
> 
> This documentation is floating around for some time now. Claudio has
> enhanced it and it is maybe time to merge it in mainline...
> 
> So, can serial people include it in their tree (Greg?) or should it go
> through another path?

I can take it, if someone resends it with the accumulated acks that I
think it gathered over the review cycle.

thanks,

greg k-h

Re: [PATCH] Documentation about RS485 serial communications

[Claudio Scordino]

Greg KH ha scritto:
> On Wed, Nov 10, 2010 at 10:17:03AM +0100, Nicolas Ferre wrote:
>> Le 24/10/2010 13:29, Claudio Scordino :
>>
>> [..]
>>
>>> Hi all,
>>>
>>> 	sorry for the misuse of the term "transceiver", and thank you
>>> for your corrections.
>>>
>>> Hopefully, this new version of the document is slightly better.
>> Hi all,
>>
>> This documentation is floating around for some time now. Claudio has
>> enhanced it and it is maybe time to merge it in mainline...
>>
>> So, can serial people include it in their tree (Greg?) or should it go
>> through another path?
> 
> I can take it, if someone resends it with the accumulated acks that I
> think it gathered over the review cycle.

Here it is.

Thanks,

	Claudio


Documentation about RS485 serial communications

Signed-off-by: Claudio Scordino <claudio@evidence.eu.com>
Acked-by: Randy Dunlap <randy.dunlap@oracle.com>
Acked-by: Russell King <linux@arm.linux.org.uk>
Acked-by: Grant Edwards <grant.b.edwards@gmail.com>
---
 Documentation/serial/00-INDEX         |    2 +
 Documentation/serial/serial-rs485.txt |  119 +++++++++++++++++++++++++++++++++
 2 files changed, 121 insertions(+), 0 deletions(-)
 create mode 100644 Documentation/serial/serial-rs485.txt

diff --git a/Documentation/serial/00-INDEX b/Documentation/serial/00-INDEX
index 07dcdb0..e09468a 100644
--- a/Documentation/serial/00-INDEX
+++ b/Documentation/serial/00-INDEX
@@ -14,6 +14,8 @@ riscom8.txt
 	- notes on using the RISCom/8 multi-port serial driver.
 rocket.txt
 	- info on the Comtrol RocketPort multiport serial driver.
+serial-rs485.txt
+	- info about RS485 structures and support in the kernel.
 specialix.txt
 	- info on hardware/driver for specialix IO8+ multiport serial card.
 stallion.txt
diff --git a/Documentation/serial/serial-rs485.txt b/Documentation/serial/serial-rs485.txt
new file mode 100644
index 0000000..40f09c6
--- /dev/null
+++ b/Documentation/serial/serial-rs485.txt
@@ -0,0 +1,119 @@
+                        RS485 SERIAL COMMUNICATIONS
+
+1. INTRODUCTION
+
+   EIA-485, also known as TIA/EIA-485 or RS-485, is a standard defining the
+   electrical characteristics of drivers and receivers for use in balanced
+   digital multipoint systems.
+   This standard is widely used for communications in industrial automation
+   because it can be used effectively over long distances and in electrically
+   noisy environments.
+
+2. HARDWARE-RELATED CONSIDERATIONS
+
+   Some CPUs (e.g., Atmel AT91) contain a built-in half-duplex mode capable of
+   automatically controlling line direction by toggling RTS. That can used to
+   control external half-duplex hardware like an RS485 transceiver or any
+   RS232-connected half-duplex device like some modems.
+
+   For these microcontrollers, the Linux driver should be made capable of
+   working in both modes, and proper ioctls (see later) should be made
+   available at user-level to allow switching from one mode to the other, and
+   vice versa.
+
+3. DATA STRUCTURES ALREADY AVAILABLE IN THE KERNEL
+
+   The Linux kernel provides the serial_rs485 structure (see [1]) to handle
+   RS485 communications. This data structure is used to set and configure RS485
+   parameters in the platform data and in ioctls.
+
+   Any driver for devices capable of working both as RS232 and RS485 should
+   provide at least the following ioctls:
+
+    - TIOCSRS485 (typically associated with number 0x542F). This ioctl is used
+      to enable/disable RS485 mode from user-space
+
+    - TIOCGRS485 (typically associated with number 0x542E). This ioctl is used
+      to get RS485 mode from kernel-space (i.e., driver) to user-space.
+
+   In other words, the serial driver should contain a code similar to the next
+   one:
+
+	static struct uart_ops atmel_pops = {
+		/* ... */
+		.ioctl		= handle_ioctl,
+	};
+
+	static int handle_ioctl(struct uart_port *port,
+		unsigned int cmd,
+		unsigned long arg)
+	{
+		struct serial_rs485 rs485conf;
+
+		switch (cmd) {
+		case TIOCSRS485:
+			if (copy_from_user(&rs485conf,
+				(struct serial_rs485 *) arg,
+				sizeof(rs485conf)))
+					return -EFAULT;
+
+			/* ... */
+			break;
+
+		case TIOCGRS485:
+			if (copy_to_user((struct serial_rs485 *) arg,
+				...,
+				sizeof(rs485conf)))
+					return -EFAULT;
+			/* ... */
+			break;
+
+		/* ... */
+		}
+	}
+
+
+4. USAGE FROM USER-LEVEL
+
+   From user-level, RS485 configuration can be get/set using the previous
+   ioctls. For instance, to set RS485 you can use the following code:
+
+	#include <linux/serial.h>
+
+	/* Driver-specific ioctls: */
+	#define TIOCGRS485      0x542E
+	#define TIOCSRS485      0x542F
+
+	/* Open your specific device (e.g., /dev/mydevice): */
+	int fd = open ("/dev/mydevice", O_RDWR);
+	if (fd < 0) {
+		/* Error handling. See errno. */
+	}
+
+	struct serial_rs485 rs485conf;
+
+	/* Set RS485 mode: */
+	rs485conf.flags |= SER_RS485_ENABLED;
+
+	/* Set rts delay before send, if needed: */
+	rs485conf.flags |= SER_RS485_RTS_BEFORE_SEND;
+	rs485conf.delay_rts_before_send = ...;
+
+	/* Set rts delay after send, if needed: */
+	rs485conf.flags |= SER_RS485_RTS_AFTER_SEND;
+	rs485conf.delay_rts_after_send = ...;
+
+	if (ioctl (fd, TIOCSRS485, &rs485conf) < 0) {
+		/* Error handling. See errno. */
+	}
+
+	/* Use read() and write() syscalls here... */
+
+	/* Close the device when finished: */
+	if (close (fd) < 0) {
+		/* Error handling. See errno. */
+	}
+
+5. REFERENCES
+
+ [1]	include/linux/serial.h
-- 
1.6.0.4

Re: [PATCH] Documentation about RS485 serial communications

[Pavel Machek]

Hi!

> Documentation about RS485 serial communications

I have seen hardware (kontron pmc-6l) that was capable of switching
between RS232, RS485 and one other standard by software.

Is such hw common? If so, should we have standard interface?

> +   From user-level, RS485 configuration can be get/set using the previous
> +   ioctls. For instance, to set RS485 you can use the following code:
> +
> +	#include <linux/serial.h>
> +
> +	/* Driver-specific ioctls: */
> +	#define TIOCGRS485      0x542E
> +	#define TIOCSRS485      0x542F
> +
> +	/* Open your specific device (e.g., /dev/mydevice): */
> +	int fd = open ("/dev/mydevice", O_RDWR);
> +	if (fd < 0) {
> +		/* Error handling. See errno. */
> +	}
> +
> +	struct serial_rs485 rs485conf;
> +
> +	/* Set RS485 mode: */
> +	rs485conf.flags |= SER_RS485_ENABLED;

IOW... is this supposed to switch electrical levels?

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

RE: [PATCH] Documentation about RS485 serial communications

[Tosoni]

Hi,

> I have seen hardware (kontron pmc-6l) that was capable of switching
> between RS232, RS485 and one other standard by software.
>
> Is such hw common? If so, should we have standard interface?

I believe it is neither common nor a good thing.

Dynamically switching between electrical levels means that the remote device
must be capable of handling both electrical levels, else it might be damaged
(i.e., when the serial card is configured to RS485 but switches to RS232
when reset).

Or, different pins must be used for RS232 and RS485 (then software switching
is not really useful).

Jean-Pierre Tosoni

> -----Original Message-----
> From: linux-serial-owner@vger.kernel.org
> [mailto:linux-serial-owner@vger.kernel.org]On Behalf Of Pavel Machek
> Sent: Tuesday, November 16, 2010 3:30 PM
> To: Claudio Scordino
> Cc: Greg KH; Nicolas Ferre; linux-serial@vger.kernel.org;
> Grant Edwards;
> linux-arm-kernel@lists.infradead.org;
> linux-kernel@vger.kernel.org; Alan
> Cox
> Subject: Re: [PATCH] Documentation about RS485 serial communications
>
>
> Hi!
>
> > Documentation about RS485 serial communications
>
> I have seen hardware (kontron pmc-6l) that was capable of switching
> between RS232, RS485 and one other standard by software.
>
> Is such hw common? If so, should we have standard interface?
>
> > +   From user-level, RS485 configuration can be get/set
> using the previous
> > +   ioctls. For instance, to set RS485 you can use the
> following code:
> > +
> > +	#include <linux/serial.h>
> > +
> > +	/* Driver-specific ioctls: */
> > +	#define TIOCGRS485      0x542E
> > +	#define TIOCSRS485      0x542F
> > +
> > +	/* Open your specific device (e.g., /dev/mydevice): */
> > +	int fd = open ("/dev/mydevice", O_RDWR);
> > +	if (fd < 0) {
> > +		/* Error handling. See errno. */
> > +	}
> > +
> > +	struct serial_rs485 rs485conf;
> > +
> > +	/* Set RS485 mode: */
> > +	rs485conf.flags |= SER_RS485_ENABLED;
>
> IOW... is this supposed to switch electrical levels?
>
> --
> (english) http://www.livejournal.com/~pavelmachek
> (cesky, pictures)
> http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
> --
> To unsubscribe from this list: send the line "unsubscribe
> linux-serial" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>

Re: [PATCH] Documentation about RS485 serial communications

[Grant Edwards]

On 2010-11-16, Tosoni <jp.tosoni@acksys.fr> wrote:

>> I have seen hardware (kontron pmc-6l) that was capable of switching
>> between RS232, RS485 and one other standard by software.
>>
>> Is such hw common? If so, should we have standard interface?
>
> I believe it is neither common nor a good thing.

People who need multiple interface types disagree with you as to
whether it's a good thing or not.

> Dynamically switching between electrical levels means that the remote
> device must be capable of handling both electrical levels, else it
> might be damaged (i.e., when the serial card is configured to RS485
> but switches to RS232 when reset).

Nobody said anything about doing it "dynamically", and if you read the
manuals for the cards in question, it specifically says not to do
that. [Though if you look at the electrical specs for any reputable
RS485/422/232 driver/receiver you'll see that they're all spec'ed to
withstand that sort of thing without damage.]

The way it's done is that the card is installed, the ports are
configured, and then the cabling is hooked up.

> Or, different pins must be used for RS232 and RS485 (then software
> switching is not really useful).

That's not usually how it's done.

-- 
Grant Edwards               grant.b.edwards        Yow! Don't hit me!!  I'm in
                                  at               the Twilight Zone!!!
                              gmail.com            

Re: [PATCH] Documentation about RS485 serial communications

[Grant Edwards]

On 2010-11-16, Pavel Machek <pavel@ucw.cz> wrote:
> Hi!
>
>> Documentation about RS485 serial communications
>
> I have seen hardware (kontron pmc-6l) that was capable of switching
> between RS232, RS485 and one other standard by software.
>
> Is such hw common? If so, should we have standard interface?

I wouldn't say it's common, but it's not all that rare either.  I
could probably name a half-dozen companies that make serial cards that
have software selectable interface types (my employer makes quite a
few).

-- 
Grant Edwards               grant.b.edwards        Yow! I'm dressing up in
                                  at               an ill-fitting IVY-LEAGUE
                              gmail.com            SUIT!!  Too late...

Re: [PATCH] Documentation about RS485 serial communications

[Alexander Stein]

Hello,

On Tuesday 16 November 2010, 15:30:16 Pavel Machek wrote:
> > Documentation about RS485 serial communications
> 
> I have seen hardware (kontron pmc-6l) that was capable of switching
> between RS232, RS485 and one other standard by software.

Maybe you mean RS422 here.

> Is such hw common? If so, should we have standard interface?

Especially in industrial system, you may have RS485 or RS422 additionally to 
RS232. This also often means switching some GPIO to change the external signal 
levels. Any the drivers must also be capable of turning on and off the 
transmitter.

Best regards,
Alexander

Re: [PATCH] Documentation about RS485 serial communications

[Matt Schulte]

On Tue, Nov 16, 2010 at 8:30 AM, Pavel Machek <pavel@ucw.cz> wrote:
> Hi!
>
>> Documentation about RS485 serial communications
>
> I have seen hardware (kontron pmc-6l) that was capable of switching
> between RS232, RS485 and one other standard by software.
>
> Is such hw common? If so, should we have standard interface?

In my opinion this type of card is not that common.  Generally
speaking the achievable baud rates for this type of multi-protocol
card are very limited because of limitations of the transceiver chips.
 It seems that most of the time people would rather have a faster
serial port than one that does several different voltages.

Matt Schulte

Re: [PATCH] Documentation about RS485 serial communications

[Grant Edwards]

On 2010-11-16, Matt Schulte <matts@commtech-fastcom.com> wrote:
> On Tue, Nov 16, 2010 at 8:30 AM, Pavel Machek <pavel@ucw.cz> wrote:
>> Hi!
>>
>>> Documentation about RS485 serial communications
>>
>> I have seen hardware (kontron pmc-6l) that was capable of switching
>> between RS232, RS485 and one other standard by software.
>>
>> Is such hw common? If so, should we have standard interface?
>
> In my opinion this type of card is not that common.  Generally
> speaking the achievable baud rates for this type of multi-protocol
> card are very limited because of limitations of the transceiver chips.

I'm curious which selectable interface cards you're talking about that
are slow?  The ones I'm familiar with generally support baud rates up
to either 460K bps or 921K bps

> It seems that most of the time people would rather have a faster
> serial port than one that does several different voltages.

Where did you find a selectable interface serial card that couldn't
support high baud rates?

-- 
Grant Edwards               grant.b.edwards        Yow! Well, O.K.
                                  at               I'll compromise with my
                              gmail.com            principles because of
                                                   EXISTENTIAL DESPAIR!

Re: [PATCH] Documentation about RS485 serial communications

[Matt Schulte]

On Tue, Nov 16, 2010 at 10:28 AM, Grant Edwards
<grant.b.edwards@gmail.com> wrote:
> On 2010-11-16, Matt Schulte <matts@commtech-fastcom.com> wrote:
>> On Tue, Nov 16, 2010 at 8:30 AM, Pavel Machek <pavel@ucw.cz> wrote:
>>> Hi!
>>>
>>>> Documentation about RS485 serial communications
>>>
>>> I have seen hardware (kontron pmc-6l) that was capable of switching
>>> between RS232, RS485 and one other standard by software.
>>>
>>> Is such hw common? If so, should we have standard interface?
>>
>> In my opinion this type of card is not that common.  Generally
>> speaking the achievable baud rates for this type of multi-protocol
>> card are very limited because of limitations of the transceiver chips.
>
> I'm curious which selectable interface cards you're talking about that
> are slow?  The ones I'm familiar with generally support baud rates up
> to either 460K bps or 921K bps
>
>> It seems that most of the time people would rather have a faster
>> serial port than one that does several different voltages.
>
> Where did you find a selectable interface serial card that couldn't
> support high baud rates?

In my world 460kbps for an RS422 card is slow.  RS422 cards generally
push multi megabit/s rates.

Matt Schulte
--
To unsubscribe from this list: send the line "unsubscribe linux-serial" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] Documentation about RS485 serial communications

[Grant Edwards]

On 2010-11-16, Matt Schulte <matts@commtech-fastcom.com> wrote:

>>> In my opinion this type of card is not that common. ?Generally
>>> speaking the achievable baud rates for this type of multi-protocol
>>> card are very limited because of limitations of the transceiver
>>> chips.
>>
>> I'm curious which selectable interface cards you're talking about
>> that are slow? ?The ones I'm familiar with generally support baud
>> rates up to either 460K bps or 921K bps
>>
>>> It seems that most of the time people would rather have a faster
>>> serial port than one that does several different voltages.
>>
>> Where did you find a selectable interface serial card that couldn't
>> support high baud rates?
>
> In my world 460kbps for an RS422 card is slow.  RS422 cards generally
> push multi megabit/s rates.

Ah, I see.  That's a completely different market than the markets
served by the multi-interface cards that typically max out at 921K.
I've support multi-interface cards for years and years, and I find it
very rare that anybody uses anything much faster than 19.2K.

-- 
Grant Edwards               grant.b.edwards        Yow! TAILFINS!! ... click
                                  at               ...
                              gmail.com            

Re: [PATCH] Documentation about RS485 serial communications

[Alan Cox]

On Tue, 16 Nov 2010 10:13:22 -0600
Matt Schulte <matts@commtech-fastcom.com> wrote:

> On Tue, Nov 16, 2010 at 8:30 AM, Pavel Machek <pavel@ucw.cz> wrote:
> > Hi!
> >
> >> Documentation about RS485 serial communications
> >
> > I have seen hardware (kontron pmc-6l) that was capable of switching
> > between RS232, RS485 and one other standard by software.
> >
> > Is such hw common? If so, should we have standard interface?
> 
> In my opinion this type of card is not that common.  Generally
> speaking the achievable baud rates for this type of multi-protocol
> card are very limited because of limitations of the transceiver chips.
>  It seems that most of the time people would rather have a faster
> serial port than one that does several different voltages

If there are two types in common use then thats enough to say we should
have a common interface IMHO

Re: [PATCH] Documentation about RS485 serial communications

[Grant Edwards]

On 2010-11-16, Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> On Tue, 16 Nov 2010 10:13:22 -0600
> Matt Schulte <matts@commtech-fastcom.com> wrote:
>
>> On Tue, Nov 16, 2010 at 8:30 AM, Pavel Machek <pavel@ucw.cz> wrote:
>> > Hi!
>> >
>> >> Documentation about RS485 serial communications
>> >
>> > I have seen hardware (kontron pmc-6l) that was capable of switching
>> > between RS232, RS485 and one other standard by software.
>> >
>> > Is such hw common? If so, should we have standard interface?
>> 
>> In my opinion this type of card is not that common.  Generally
>> speaking the achievable baud rates for this type of multi-protocol
>> card are very limited because of limitations of the transceiver chips.
>>  It seems that most of the time people would rather have a faster
>> serial port than one that does several different voltages
>
> If there are two types in common use then thats enough to say we should
> have a common interface IMHO

Comtrol, Moxa, B&B, Sealevel, and others all sell PCI cards and
Ethernet attached serial ports that have selectable interfaces
(typically RS-232/422/485).  Comtrol and Moxa have had drivers in the
kernel tree for ages, but they've always had to use custom ioctl calls
for things like configuring 232/485/422 mode and half/full-duplex
mode.

There are also tons of small Linux-based industrial server appliances
from Comtrol, Silex, Digi, Moxa, and others that have selectable
interface serial ports.

Having a standard API for things like interface mode, half-full
duplex, inter-character timeout, 9-bit mode, and so on would be life a
lot easier for those of us who maintain Linux serial drivers...

-- 
Grant Edwards               grant.b.edwards        Yow! I always have fun
                                  at               because I'm out of my
                              gmail.com            mind!!!

Re: [PATCH] Documentation about RS485 serial communications

[Alan Cox]

> Comtrol, Moxa, B&B, Sealevel, and others all sell PCI cards and
> Ethernet attached serial ports that have selectable interfaces
> (typically RS-232/422/485).  Comtrol and Moxa have had drivers in the
> kernel tree for ages, but they've always had to use custom ioctl calls
> for things like configuring 232/485/422 mode and half/full-duplex
> mode.

Ok so what is needed that isn't covered by the current termiox and
TIOCG/SRS485 ioctls ?

> Having a standard API for things like interface mode, half-full
> duplex, inter-character timeout, 9-bit mode, and so on would be life a
> lot easier for those of us who maintain Linux serial drivers...

9bit mode is a real problem with the tty layer defined in terms of 8bit
streams but yes.

Re: [PATCH] Documentation about RS485 serial communications

[Pavel Machek]

On Tue 2010-11-16 20:04:26, Alan Cox wrote:
> > Comtrol, Moxa, B&B, Sealevel, and others all sell PCI cards and
> > Ethernet attached serial ports that have selectable interfaces
> > (typically RS-232/422/485).  Comtrol and Moxa have had drivers in the
> > kernel tree for ages, but they've always had to use custom ioctl calls
> > for things like configuring 232/485/422 mode and half/full-duplex
> > mode.
> 
> Ok so what is needed that isn't covered by the current termiox and
> TIOCG/SRS485 ioctls ?

IIRC selection of 232/485/422 mode was missing from proposed api.

> > Having a standard API for things like interface mode, half-full
> > duplex, inter-character timeout, 9-bit mode, and so on would be life a
> > lot easier for those of us who maintain Linux serial drivers...
> 
> 9bit mode is a real problem with the tty layer defined in terms of 8bit
> streams but yes.

Well, if you wanted to have some fun... kontron card is able to
timestamp incoming characters, and saves state of the control signals
when character is received...
								Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html