[PATCH 1/2] pch_uart: Fix hw-flow control issue

[PATCH 1/2] pch_uart: Fix hw-flow control issue

[Tomoya MORINAGA]

Using hardware flow control,
currently, register of the control-bit(AFE) is not set.
This patch fixes the issue.

Signed-off-by: Tomoya MORINAGA <tomoya-linux@dsn.lapis-semi.com>
---
 drivers/tty/serial/pch_uart.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/drivers/tty/serial/pch_uart.c b/drivers/tty/serial/pch_uart.c
index b950d05..b1aabe8 100644
--- a/drivers/tty/serial/pch_uart.c
+++ b/drivers/tty/serial/pch_uart.c
@@ -1290,6 +1290,7 @@ static void pch_uart_set_termios(struct uart_port *port,
 	if (rtn)
 		goto out;
 
+	pch_uart_set_mctrl(&priv->port, priv->port.mctrl);
 	/* Don't rewrite B0 */
 	if (tty_termios_baud_rate(termios))
 		tty_termios_encode_baud_rate(termios, baud, baud);
-- 
1.7.4.4

[PATCH 2/2] pch_uart: Delete modem status interrupt enable code

[Tomoya MORINAGA]

Currently, modem status interrupt software processing is not implemented.
So, if enabling modem status interrupt, the interrupt processing is not
processed correctly.
(pch_uart device has auto hardware flow control function. So, pch_uart driver
doesn't have to control modem status interrupt.)
This patch deletes the code.

Signed-off-by: Tomoya MORINAGA <tomoya-linux@dsn.lapis-semi.com>
---
 drivers/tty/serial/pch_uart.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/drivers/tty/serial/pch_uart.c b/drivers/tty/serial/pch_uart.c
index b1aabe8..1a076cd 100644
--- a/drivers/tty/serial/pch_uart.c
+++ b/drivers/tty/serial/pch_uart.c
@@ -1121,7 +1121,8 @@ static void pch_uart_enable_ms(struct uart_port *port)
 {
 	struct eg20t_port *priv;
 	priv = container_of(port, struct eg20t_port, port);
-	pch_uart_hal_enable_interrupt(priv, PCH_UART_HAL_MS_INT);
+	dev_info(priv->port.dev,
+		"PCH UART : Modem status interrupt is not supported.\n");
 }
 
 /* Control the transmission of a break signal. */
-- 
1.7.4.4

Re: [PATCH 2/2] pch_uart: Delete modem status interrupt enable code

[Alan Cox]

> (pch_uart device has auto hardware flow control function. So,
> pch_uart driver doesn't have to control modem status interrupt.)

This is true for flow control bits but not for incoming carrier detect ?

> This patch deletes the code.

> +	dev_info(priv->port.dev,
> +		"PCH UART : Modem status interrupt is not
> supported.\n"); }

dev_info seems wrong here. Does the user need to know this ?

Re: [PATCH 2/2] pch_uart: Delete modem status interrupt enable code

[Tomoya MORINAGA]

(2011/11/01 21:57), Alan Cox wrote:
>> (pch_uart device has auto hardware flow control function. So,
>> pch_uart driver doesn't have to control modem status interrupt.)
>
> This is true for flow control bits but not for incoming carrier detect ?

Called ".get_mctrl" function, other software can know incoming carrier 
detection.
Or you mean a uart driver must detect incoming carrier with interrupt ?

>> This patch deletes the code.
>
>> +	dev_info(priv->port.dev,
>> +		"PCH UART : Modem status interrupt is not
>> supported.\n"); }
>
> dev_info seems wrong here. Does the user need to know this ?
OK, a user doesn't have to know the information.
I will delete the dev_info.

Thanks,
-- 
tomoya
ROHM Co., Ltd.

Re: [PATCH 2/2] pch_uart: Delete modem status interrupt enable code

[Alan Cox]

> Called ".get_mctrl" function, other software can know incoming carrier 
> detection.
> Or you mean a uart driver must detect incoming carrier with interrupt ?

If a user opens a tty port which has no carrier signal then the open is
supposed to block until the carrier is raised unless either clocal
termios or O_NDELAY open flags are set.

That usually means you need to take the carrier raise interrupt in order
to wake out of open in this case.

Alan

Re: [PATCH 2/2] pch_uart: Delete modem status interrupt enable code

[Tomoya MORINAGA]

Hi Alan,

(2011/11/02 20:00), Alan Cox wrote:
>> Called ".get_mctrl" function, other software can know incoming carrier
>> detection.
>> Or you mean a uart driver must detect incoming carrier with interrupt ?
>
> If a user opens a tty port which has no carrier signal then the open is
> supposed to block until the carrier is raised unless either clocal
> termios or O_NDELAY open flags are set.
>
> That usually means you need to take the carrier raise interrupt in order
> to wake out of open in this case.
>

I understand.
We need to add modem status interrupt processing.

Thanks,
-- 
tomoya
ROHM Co., Ltd.

Re: [PATCH 1/2] pch_uart: Fix hw-flow control issue

[Alan Cox]

On Thu, 27 Oct 2011 15:45:18 +0900
Tomoya MORINAGA <tomoya-linux@dsn.lapis-semi.com> wrote:

> Using hardware flow control,
> currently, register of the control-bit(AFE) is not set.
> This patch fixes the issue.
> 
> Signed-off-by: Tomoya MORINAGA <tomoya-linux@dsn.lapis-semi.com>

Acked-by: Alan Cox <alan@linux.intel.com>