Re: [PATCH] Bluetooth: hci_core: Fix poential Use-after-Free bug in hci_remove_adv_monitor

[Zheng Hacker]

Luiz Augusto von Dentz <luiz.dentz@xxxxxxxxx> 于2023年2月11日周六 03:53写道：
>
> Hi Zheng,
>
> On Thu, Feb 9, 2023 at 8:11 PM Zheng Wang <zyytlz.wz@xxxxxxx> wrote:
> >
> > In hci_remove_adv_monitor, if it gets into HCI_ADV_MONITOR_EXT_MSFT case,
> > the function will free the monitor and print its handle after that.
> >
> > Fix it by switch the order.
> >
> > Fixes: 7cf5c2978f23 ("Bluetooth: hci_sync: Refactor remove Adv Monitor")
> > Signed-off-by: Zheng Wang <zyytlz.wz@xxxxxxx>
> > ---
> >  net/bluetooth/hci_core.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/net/bluetooth/hci_core.c b/net/bluetooth/hci_core.c
> > index b65c3aabcd53..db3352c60de6 100644
> > --- a/net/bluetooth/hci_core.c
> > +++ b/net/bluetooth/hci_core.c
> > @@ -1980,9 +1980,9 @@ static int hci_remove_adv_monitor(struct hci_dev *hdev,
> >                 goto free_monitor;
> >
> >         case HCI_ADV_MONITOR_EXT_MSFT:
> > -               status = msft_remove_monitor(hdev, monitor);
> >                 bt_dev_dbg(hdev, "%s remove monitor %d msft status %d",
> >                            hdev->name, monitor->handle, status);
> > +               status = msft_remove_monitor(hdev, monitor);
>
> I wonder if it is not a good idea to move the logging inside
> msft_remove_monitor?
>
> >                 break;
> >         }
> >
Hello Luiz,

Thanks for your quick reply. I think moving it inside msft_remove_monitor
is a good idea. Because the variable status is returned by msft_remove_monitor.
The call chain is
msft_remove_monitor
  ->msft_remove_monitor_sync
    ->msft_le_cancel_monitor_advertisement_cb
      ->hci_free_adv_monitor
        ->kfree(monitor);
We could mov the bt_dev_dbg to the location before
hci_free_adv_monitor calling and delete the releated code in
hci_remove_adv_monitor.

If you agree with that, I'll make another patch.

Best regards,
Zheng Wang