Re: [PATCH] serial: core: Fix kmemleak issue for serial core device remove

[Greg Kroah-Hartman <gregkh@linuxfoundation.org>]

On Fri, Aug 04, 2023 at 12:21:05PM +0300, Tony Lindgren wrote:
> * Jiri Slaby <jirislaby@kernel.org> [230804 09:16]:
> > On 04. 08. 23, 11:09, Tony Lindgren wrote:
> > > Kmemleak reports issues for serial8250 ports after the hardware specific
> > > driver takes over on boot as noted by Tomi.
> > > 
> > > The kerneldoc for device_initialize() says we must call device_put()
> > > after calling device_initialize(). We are calling device_put() on the
> > > error path, but are missing it from the device remove path. This causes
> > > release() to never get called for the devices on remove.
> > > 
> > > Let's add the missing put_device() calls for both serial ctrl and
> > > port devices.
> > > 
> > > Fixes: 84a9582fd203 ("serial: core: Start managing serial controllers to enable runtime PM")
> > > Reported-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
> > > Signed-off-by: Tony Lindgren <tony@atomide.com>
> > > ---
> > >   drivers/tty/serial/serial_base_bus.c | 2 ++
> > >   1 file changed, 2 insertions(+)
> > > 
> > > diff --git a/drivers/tty/serial/serial_base_bus.c b/drivers/tty/serial/serial_base_bus.c
> > > --- a/drivers/tty/serial/serial_base_bus.c
> > > +++ b/drivers/tty/serial/serial_base_bus.c
> > > @@ -99,6 +99,7 @@ void serial_base_ctrl_device_remove(struct serial_ctrl_device *ctrl_dev)
> > >   		return;
> > >   	device_del(&ctrl_dev->dev);
> > > +	put_device(&ctrl_dev->dev);
> > >   }
> > >   struct serial_ctrl_device *serial_base_ctrl_add(struct uart_port *port,
> > > @@ -174,6 +175,7 @@ void serial_base_port_device_remove(struct serial_port_device *port_dev)
> > >   		return;
> > >   	device_del(&port_dev->dev);
> > > +	put_device(&port_dev->dev);
> > 
> > I didn't check the code, but device_unregister()?
> 
> I thought about that as it does the same, but since we're not calling
> device_register() I felt it would be and unpaired call. No objections to
> changing to use device_unregister() naturally if folks prefer that.

This is fine as device_register() isn't happening, thanks.

greg k-h