Re: [PATCH 2/3] usb console,usb-serial: fix regression use of serial->console

[Jason Wessel]

Alan Stern wrote:
> On Wed, 16 Sep 2009, Jason Wessel wrote:
>
>   
>> During the 2.6.32 development usb serial tty interface was changed.
>> This change caused the usb serial driver to crash when used as a
>> system console.
>>
>> The open and close of the raw hardware has to be protected by the
>> serial->console variable so either the console code does the open and
>> close or the higher level tty based driver executes the open and
>> close.
>>
>> * In serial_open() the raw open must be protected by serial->console
>> * In serial_down() the test_and_clear_bit must be executed before
>>   leaving the function if the device is a serial console
>> * serial_release() should no longer protect for the console condition
>>   due to the 2.6.32 interface changes.  The actual low level close
>>   protection is accounted for in serial_down()
>>
>> Signed-off-by: Jason Wessel <jason.wessel@xxxxxxxxxxxxx>
>> Cc: Greg KH <gregkh@xxxxxxx>
>> Cc: Alan Stern <stern@xxxxxxxxxxxxxxxxxxx>
>>
>> ---
>>  drivers/usb/serial/usb-serial.c |   28 +++++++++++++---------------
>>  1 files changed, 13 insertions(+), 15 deletions(-)
>>
>> diff --git a/drivers/usb/serial/usb-serial.c b/drivers/usb/serial/usb-serial.c
>> index ff75a35..3d35ed2 100644
>> --- a/drivers/usb/serial/usb-serial.c
>> +++ b/drivers/usb/serial/usb-serial.c
>> @@ -267,10 +267,14 @@ static int serial_open(struct tty_struct *tty, struct file *filp)
>>  		if (mutex_lock_interruptible(&port->mutex))
>>  			return -ERESTARTSYS;
>>  		mutex_lock(&serial->disc_mutex);
>> -		if (serial->disconnected)
>> +		if (serial->disconnected) {
>>  			retval = -ENODEV;
>> -		else
>> -			retval = port->serial->type->open(tty, port);
>> +		} else {
>> +			if (port->console)
>> +				retval = 0;
>> +			else
>> +				retval = port->serial->type->open(tty, port);
>> +		}
>>  		mutex_unlock(&serial->disc_mutex);
>>  		mutex_unlock(&port->mutex);
>>  		if (retval)
>>     
>
> This part of the patch shouldn't be necessary.  The console device
> should always be open.
>
>   

With your other patch you just mentioned, I absolutely agree.  This does
not need to be tracked in this manner and I don't believe this patch is
needed either.


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/