On 2018/10/1 15:19, Dan Carpenter wrote:
> On Sun, Sep 30, 2018 at 06:02:24PM +0200, Johan Hovold wrote:
>> On Sat, Sep 29, 2018 at 09:54:03AM +0000, YueHaibing wrote:
>>> Fixes gcc '-Wunused-but-set-variable' warning:
>>>
>>> drivers/usb/serial/cypress_m8.c: In function 'cypress_send':
>>> drivers/usb/serial/cypress_m8.c:689:33: warning:
>>> variable 'actual_size' set but not used [-Wunused-but-set-variable]
>>>
>>> drivers/usb/serial/cypress_m8.c: In function 'cypress_set_termios':
>>> drivers/usb/serial/cypress_m8.c:866:18: warning:
>>> variable 'iflag' set but not used [-Wunused-but-set-variable]
>>>
>>> Signed-off-by: YueHaibing <[email protected]>
>>> ---
>>> drivers/usb/serial/cypress_m8.c | 11 ++---------
>>> 1 file changed, 2 insertions(+), 9 deletions(-)
>>>
>>> diff --git a/drivers/usb/serial/cypress_m8.c
>>> b/drivers/usb/serial/cypress_m8.c
>>> index 31c6091..98dff12 100644
>>> --- a/drivers/usb/serial/cypress_m8.c
>>> +++ b/drivers/usb/serial/cypress_m8.c
>>> @@ -686,7 +686,7 @@ static int cypress_write(struct tty_struct *tty, struct
>>> usb_serial_port *port,
>>>
>>> static void cypress_send(struct usb_serial_port *port)
>>> {
>>> - int count = 0, result, offset, actual_size;
>>> + int count = 0, result, offset;
>>> struct cypress_private *priv = usb_get_serial_port_data(port);
>>> struct device *dev = &port->dev;
>>> unsigned long flags;
>>> @@ -758,12 +758,6 @@ static void cypress_send(struct usb_serial_port *port)
>>> priv->write_urb_in_use = 1;
>>> spin_unlock_irqrestore(&priv->lock, flags);
>>>
>>> - if (priv->cmd_ctrl)
>>> - actual_size = 1;
>>> - else
>>> - actual_size = count +
>>> - (priv->pkt_fmt == packet_format_1 ? 2 : 1);
>>> -
>>
>> This looks like we have a bug in the driver, and sure enough we do.
>> Commit 9aa8dae7b1fa ("cypress_m8: use usb_fill_int_urb where
>> appropriate") incorrectly started setting the transfer length to the
>> size of the buffer.
>>
>
> This is like the third time recently where we've had found a real bug.
>
> YueHaibing, could you mention in the commit message where the variable
> became unused? It would help reviewing. It's not a Fixes tag but it
> would look like this:
Ok, will do it.
>
> "After commit 9aa8dae7b1fa ("cypress_m8: use usb_fill_int_urb where
> appropriate") then we don't use actual_size any more so that can
> be removed."
>
>
> regards,
> dan carpenter
>
>
> .
>