Re: [PATCH] net: caif: fix memory leak in ldisc_receive

[Osama Abdelkader <osama.abdelkader@gmail.com>]

On Sun, Jan 18, 2026 at 04:02:44PM +0100, Greg Kroah-Hartman wrote:
> On Sun, Jan 18, 2026 at 03:47:54PM +0100, Osama Abdelkader wrote:
> > Add NULL pointer checks for ser and ser->dev in ldisc_receive() to
> > prevent memory leaks when the function is called during device close
> > or in race conditions where tty->disc_data or ser->dev may be NULL.
> > 
> > The memory leak occurred because netdev_alloc_skb() would allocate an
> > skb, but if ser or ser->dev was NULL, the function would return early
> > without freeing the allocated skb. Additionally, ser->dev was accessed
> > before checking if it was NULL, which could cause a NULL pointer
> > dereference.
> > 
> > Reported-by: syzbot+f9d847b2b84164fa69f3@syzkaller.appspotmail.com
> > Closes:
> > https://syzkaller.appspot.com/bug?extid=f9d847b2b84164fa69f3
> 
> Please do not wrap this line.

OK.

> 
> > Fixes: 9b27105b4a44 ("net-caif-driver: add CAIF serial driver (ldisc)")
> > CC: stable@vger.kernel.org
> > Signed-off-by: Osama Abdelkader <osama.abdelkader@gmail.com>
> > ---
> >  drivers/net/caif/caif_serial.c | 6 ++++--
> >  1 file changed, 4 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/net/caif/caif_serial.c b/drivers/net/caif/caif_serial.c
> > index c398ac42eae9..0ec9670bd35c 100644
> > --- a/drivers/net/caif/caif_serial.c
> > +++ b/drivers/net/caif/caif_serial.c
> > @@ -152,12 +152,16 @@ static void ldisc_receive(struct tty_struct *tty, const u8 *data,
> >  	int ret;
> >  
> >  	ser = tty->disc_data;
> > +	if (!ser)
> > +		return;
> 
> Can this ever be true?

Yes, when the line discipline is changed, tty_set_termios_ldisc() sets tty->disc_data = NULL
> 
> >  	/*
> >  	 * NOTE: flags may contain information about break or overrun.
> >  	 * This is not yet handled.
> >  	 */
> >  
> > +	if (!ser->dev)
> > +		return;
> 
> Why is this check here and not just merged together with the one you
> added above?  And how can ->dev be NULL?

I'm going to combine them in v2.
If ser exists, ser->dev should be non-NULL (they're created together), but the check is defensive.

> 
> And where is the locking to prevent this from changing right after you
> check it?
> 

I'm going to address that in v2.

> thanks,
> 
> greg k-h

Thanks,
Osama