Re: usbserial not working/oops on removal

[Andreas Laumann <al@authentidate.de>]

Am Samstag, den 03.03.2007, 21:43 +0100 schrieb Oliver Neukum:
> Am Samstag, 3. März 2007 14:27 schrieb Andreas Laumann:
> > Unfortunately your fix  does not work.  The first one seems registered
> > correctly, but after that it fails again for the next one.  
> 
> Unfortunately the last patch I sent you introduced a race. Could you test
> this alternative patch?
> 
> 	Regards
> 		Oliver
> 
> --- a/drivers/usb/serial/usb-serial.c	2007-03-03 16:28:22.000000000 +0100
> +++ b/drivers/usb/serial/usb-serial.c	2007-03-03 21:37:27.000000000 +0100
> @@ -99,9 +99,12 @@
>  			continue;
>  
>  		*minor = i;
> +		j = 0;
>  		dbg("%s - minor base = %d", __FUNCTION__, *minor);
> -		for (i = *minor; (i < (*minor + num_ports)) && (i < SERIAL_TTY_MINORS); ++i)
> +		for (i = *minor; (i < (*minor + num_ports)) && (i < SERIAL_TTY_MINORS); ++i) {
>  			serial_table[i] = serial;
> +			serial->port[j++]->number = i;
> +		}
>  		spin_unlock(&table_lock);
>  		return serial;
>  	}
> @@ -847,7 +850,6 @@
>  		port = kzalloc(sizeof(struct usb_serial_port), GFP_KERNEL);
>  		if (!port)
>  			goto probe_error;
> -		port->number = i + serial->minor;
>  		port->serial = serial;
>  		spin_lock_init(&port->lock);
>  		mutex_init(&port->mutex);

On top of the first patch ? Sure, I'll try it in a couple of minutes

Btw, I got rejects from the last patch (the racy one) you've sent and
had to apply them manually.  Which version do you use ? 

Regards
	Andreas