[PATCH] drivers:usb:serial:ftdi_sio.c: ftdi_set_termios make no sense some in a situation

[PATCH] drivers:usb:serial:ftdi_sio.c: ftdi_set_termios make no sense some in a situation

[wangyanqing]

b1ffb4c851f185e9051ba837c16d9b84ef688d26 make a improvement to drop no sense reset
termios. But it miss the situation, when call ftdi_set_termios with the value that
same with the default settings all the time, then termios will take no effect any way,
we need usb_control_msg to make effect. This patch try to resolve this question.

Signed-off-by: Wang YanQing <ukdnight@gmail.com>
---
 drivers/usb/serial/ftdi_sio.c |   10 ++++++++--
 1 files changed, 8 insertions(+), 2 deletions(-)

diff --git a/drivers/usb/serial/ftdi_sio.c b/drivers/usb/serial/ftdi_sio.c
index 486769c..61f1be4 100644
--- a/drivers/usb/serial/ftdi_sio.c
+++ b/drivers/usb/serial/ftdi_sio.c
@@ -2056,6 +2056,7 @@ static void ftdi_set_termios(struct tty_struct *tty,
 	struct ktermios *termios = tty->termios;
 	unsigned int cflag = termios->c_cflag;
 	__u16 urb_value; /* will hold the new flags */
+	static int init_once = 1;
 
 	/* Added for xon/xoff support */
 	unsigned int iflag = termios->c_iflag;
@@ -2082,7 +2083,8 @@ static void ftdi_set_termios(struct tty_struct *tty,
 
 	if (old_termios->c_cflag == termios->c_cflag
 	    && old_termios->c_ispeed == termios->c_ispeed
-	    && old_termios->c_ospeed == termios->c_ospeed)
+	    && old_termios->c_ospeed == termios->c_ospeed
+	    && !init_once)
 		goto no_c_cflag_changes;
 
 	/* NOTE These routines can get interrupted by
@@ -2090,8 +2092,12 @@ static void ftdi_set_termios(struct tty_struct *tty,
 	   don't see any problems yet */
 
 	if ((old_termios->c_cflag & (CSIZE|PARODD|PARENB|CMSPAR|CSTOPB)) ==
-	    (termios->c_cflag & (CSIZE|PARODD|PARENB|CMSPAR|CSTOPB)))
+	    (termios->c_cflag & (CSIZE|PARODD|PARENB|CMSPAR|CSTOPB)) &&
+	    !init_once) {
 		goto no_data_parity_stop_changes;
+	} else {
+		init_once = 0;
+	}
 
 	/* Set number of data bits, parity, stop bits */
 
-- 
1.7.3.4

Re: [PATCH] drivers:usb:serial:ftdi_sio.c: ftdi_set_termios make no sense some in a situation

[Alan Cox]

On Mon, 5 Dec 2011 15:47:30 +0800
wangyanqing <udknight@gmail.com> wrote:

> b1ffb4c851f185e9051ba837c16d9b84ef688d26 make a improvement to drop no sense reset
> termios. But it miss the situation, when call ftdi_set_termios with the value that
> same with the default settings all the time, then termios will take no effect any way,
> we need usb_control_msg to make effect. This patch try to resolve this question.
> 
> Signed-off-by: Wang YanQing <ukdnight@gmail.com>
> ---
>  drivers/usb/serial/ftdi_sio.c |   10 ++++++++--
>  1 files changed, 8 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/usb/serial/ftdi_sio.c b/drivers/usb/serial/ftdi_sio.c
> index 486769c..61f1be4 100644
> --- a/drivers/usb/serial/ftdi_sio.c
> +++ b/drivers/usb/serial/ftdi_sio.c
> @@ -2056,6 +2056,7 @@ static void ftdi_set_termios(struct tty_struct *tty,
>  	struct ktermios *termios = tty->termios;
>  	unsigned int cflag = termios->c_cflag;
>  	__u16 urb_value; /* will hold the new flags */
> +	static int init_once = 1;

NACK

Think about what happens if you have two such ports or remove and re-add
it.

What code path is calling set_termios passing a termios data which has
never been set ?

Alan

Re: [PATCH] drivers:usb:serial:ftdi_sio.c: ftdi_set_termios make no sense some in a situation

[Wang YanQing]

On Mon, Dec 05, 2011 at 11:34:24AM +0000, Alan Cox wrote:
> NACK
> 
> Think about what happens if you have two such ports or remove and re-add
> it.
> 
Alan, you are right! Thanks
> What code path is calling set_termios passing a termios data which has
> never been set ?
> 
> Alan

The question I meet is my usb-convertor base on FT232RL doesn't work, but it 
works well if I revert b1ffb4c851f185e9051ba837c16d9b84ef688d26.It seems that
I should dig deeper.