Recently syzkaller reported a UAF in LINE6 driver, and it's likely
because we call cancel_delayed_work() at the disconnect callback
instead of cancel_delayed_work_sync().  Let's use the correct one
instead.

Reported-by: syzbot+145012a46658ac00fc9e@syzkaller.appspotmail.com
Suggested-by: Alan Stern <stern@rowland.harvard.edu>
Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/s5hlfjr4gio.wl-tiwai@suse.deSigned-off-by: Takashi Iwai <tiwai@suse.de>