Re: [PATCH v4 2/2] USB: cdc-acm: fix rounding error in TIOCSSERIAL

[Matthias Reichl <hias@horus.com>]

On Thu, Mar 12, 2020 at 02:31:01PM +0100, Anthony Mallet wrote:
> By default, tty_port_init() initializes those parameters to a multiple
> of HZ. For instance in line 69 of tty_port.c:
>    port->close_delay = (50 * HZ) / 100;
> https://github.com/torvalds/linux/blob/master/drivers/tty/tty_port.c#L69
> 
> With e.g. CONFIG_HZ = 250 (as this is the case for Ubuntu 18.04
> linux-image-4.15.0-37-generic), the default setting for close_delay is
> thus 125.
> 
> When ioctl(fd, TIOCGSERIAL, &s) is executed, the setting returned in
> user space is '12' (125/10). When ioctl(fd, TIOCSSERIAL, &s) is then
> executed with the same setting '12', the value is interpreted as '120'
> which is different from the current setting and a EPERM error may be
> raised by set_serial_info() if !CAP_SYS_ADMIN.
> https://github.com/torvalds/linux/blob/master/drivers/usb/class/cdc-acm.c#L919
> 
> Fixes: ba2d8ce9db0a6 ("cdc-acm: implement TIOCSSERIAL to avoid blocking close(2)")
> Signed-off-by: Anthony Mallet <anthony.mallet@laas.fr>
> ---
> Changed in v2: fix typo (extra closing brace)
> Changed in v3: version this patch series
> Changed in v4: diffed against e6e6ec4 from github.com/torvalds/linux.git
> ---
>  drivers/usb/class/cdc-acm.c | 27 +++++++++++++++++----------
>  1 file changed, 17 insertions(+), 10 deletions(-)
> 
> diff --git a/drivers/usb/class/cdc-acm.c b/drivers/usb/class/cdc-acm.c
> index da619176deca..a41a3d27016c 100644
> --- a/drivers/usb/class/cdc-acm.c
> +++ b/drivers/usb/class/cdc-acm.c
> @@ -907,6 +907,7 @@ static int set_serial_info(struct tty_struct *tty, struct serial_struct *ss)
>  {
>  	struct acm *acm = tty->driver_data;
>  	unsigned int closing_wait, close_delay;
> +	unsigned int old_closing_wait, old_close_delay;
>  	int retval = 0;
>  
>  	close_delay = msecs_to_jiffies(ss->close_delay * 10);
> @@ -914,18 +915,24 @@ static int set_serial_info(struct tty_struct *tty, struct serial_struct *ss)
>  			ASYNC_CLOSING_WAIT_NONE :
>  			msecs_to_jiffies(ss->closing_wait * 10);
>  
> +	/* we must redo the rounding here, so that the values match */
> +	old_close_delay	= jiffies_to_msecs(acm->port.close_delay) / 10;
> +	old_closing_wait = acm->port.closing_wait == ASYNC_CLOSING_WAIT_NONE ?
> +				ASYNC_CLOSING_WAIT_NONE :
> +				jiffies_to_msecs(acm->port.closing_wait) / 10;
> +
>  	mutex_lock(&acm->port.mutex);
>  
> -	if (!capable(CAP_SYS_ADMIN)) {
> -		if ((close_delay != acm->port.close_delay) ||
> -		    (closing_wait != acm->port.closing_wait))
> +	if ((ss->close_delay != old_close_delay) ||
> +            (ss->closing_wait != old_closing_wait)) {
> +		if (!capable(CAP_SYS_ADMIN))
>  			retval = -EPERM;
> -		else
> -			retval = -EOPNOTSUPP;
> -	} else {
> -		acm->port.close_delay  = close_delay;
> -		acm->port.closing_wait = closing_wait;
> -	}
> +		else {
> +			acm->port.close_delay  = close_delay;
> +			acm->port.closing_wait = closing_wait;
> +		}
> +	} else
> +		retval = -EOPNOTSUPP;

Swapping the order of the capability / close settings checks introduced
a rather unexpected regression: when running with CAP_SYS_ADMIN we now
get -EOPNOTSUPP when the close settings are unchanged.

This will break programs / scripts that set these settings to fixed
values.

eg (tested on vanilla 5.4.28) changing the settings works, but
setting the same value again fails:

# setserial -a /dev/ttyACM0
/dev/ttyACM0, Line 0, UART: unknown, Port: 0x0000, IRQ: 0
	Baud_base: 9600, close_delay: 50, divisor: 0
	closing_wait: 3000
	Flags: spd_normal

# setserial /dev/ttyACM0 closing_wait 2000
# setserial /dev/ttyACM0 closing_wait 2000
Cannot set serial info: Operation not supported

From the commit message it's not clear why this behaviour was changed
so I suspect this might be a bug.

so long,

Hias

>  
>  	mutex_unlock(&acm->port.mutex);
>  	return retval;
> -- 
> 2.17.1
>