On Thu, Sep 04, 2003 at 11:34:09AM -0700, Greg KH wrote:
> On Thu, Sep 04, 2003 at 07:21:54PM +0100, Ian Abbott wrote:
> >
> > I saw a report of a kernel oops on the ftdi-sio-devel list. The
> > oops occurred when the [FTDI] USB serial device was unplugged while
> > doing 'cat /dev/ttyUSB0' and the oops was of the "Unabled to handle
> > kernel NULL pointer dereference" variety and the oops occurred in
> > the usb_serial_disconnect function. To cut to the chase, the
> > offending lines in usb_serial_disconnect are as follows:
> >
> > while (port->open_count > 0) {
> > __serial_close(port, NULL);
> > }
> > port->tty->driver_data = NULL;
> >
> > If port->open_count > 0, the final call to __serial_close sets
> > port->tty = NULL, resulting in the null pointer dereference on the
> > final line above.
>
> Ah, nice catch, thanks.
> So does the following patch solve the problem for you?
Hm, this is probably the better patch for this instead. Can you try
this?
thanks,
greg k-h
--- a/drivers/usb/serial/usbserial.c Sat Aug 30 23:27:18 2003
+++ b/drivers/usb/serial/usbserial.c Thu Sep 4 13:48:45 2003
@@ -556,7 +556,10 @@
else
generic_close(port, filp);
port->open_count = 0;
- port->tty = NULL;
+ if (port->tty) {
+ port->tty->driver_data = NULL;
+ port->tty = NULL;
+ }
}
if (port->serial->type->owner)
@@ -1401,12 +1404,9 @@
for (i = 0; i < serial->num_ports; ++i) {
port = &serial->port[i];
down (&port->sem);
- if (port->tty != NULL) {
- while (port->open_count > 0) {
+ if (port->tty != NULL)
+ while (port->open_count > 0)
__serial_close(port, NULL);
- }
- port->tty->driver_data = NULL;
- }
up (&port->sem);
}
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
