Re: [PATCH v2] Bluetooth: Check for encryption key size on connect

[Archie Pusaka]

Hi Luiz,

On Wed, 23 Sep 2020 at 01:03, Luiz Augusto von Dentz
<luiz.dentz@xxxxxxxxx> wrote:
>
> Hi Archie,
>
> On Tue, Sep 22, 2020 at 12:48 AM Archie Pusaka <apusaka@xxxxxxxxxx> wrote:
> >
> > Hi Luiz,
> >
> > On Tue, 22 Sep 2020 at 01:15, Luiz Augusto von Dentz
> > <luiz.dentz@xxxxxxxxx> wrote:
> > >
> > > Hi Archie,
> > >
> > >
> > > On Mon, Sep 21, 2020 at 12:56 AM Archie Pusaka <apusaka@xxxxxxxxxx> wrote:
> > > >
> > > > From: Archie Pusaka <apusaka@xxxxxxxxxxxx>
> > > >
> > > > When receiving connection, we only check whether the link has been
> > > > encrypted, but not the encryption key size of the link.
> > > >
> > > > This patch adds check for encryption key size, and reject L2CAP
> > > > connection which size is below the specified threshold (default 7)
> > > > with security block.
> > > >
> > > > Here is some btmon trace.
> > > > @ MGMT Event: New Link Key (0x0009) plen 26    {0x0001} [hci0] 5.847722
> > > >         Store hint: No (0x00)
> > > >         BR/EDR Address: 38:00:25:F7:F1:B0 (OUI 38-00-25)
> > > >         Key type: Unauthenticated Combination key from P-192 (0x04)
> > > >         Link key: 7bf2f68c81305d63a6b0ee2c5a7a34bc
> > > >         PIN length: 0
> > > > > HCI Event: Encryption Change (0x08) plen 4        #29 [hci0] 5.871537
> > > >         Status: Success (0x00)
> > > >         Handle: 256
> > > >         Encryption: Enabled with E0 (0x01)
> > > > < HCI Command: Read Encryp... (0x05|0x0008) plen 2  #30 [hci0] 5.871609
> > > >         Handle: 256
> > > > > HCI Event: Command Complete (0x0e) plen 7         #31 [hci0] 5.872524
> > > >       Read Encryption Key Size (0x05|0x0008) ncmd 1
> > > >         Status: Success (0x00)
> > > >         Handle: 256
> > > >         Key size: 3
> > > >
> > > > ////// WITHOUT PATCH //////
> > > > > ACL Data RX: Handle 256 flags 0x02 dlen 12        #42 [hci0] 5.895023
> > > >       L2CAP: Connection Request (0x02) ident 3 len 4
> > > >         PSM: 4097 (0x1001)
> > > >         Source CID: 64
> > > > < ACL Data TX: Handle 256 flags 0x00 dlen 16        #43 [hci0] 5.895213
> > > >       L2CAP: Connection Response (0x03) ident 3 len 8
> > > >         Destination CID: 64
> > > >         Source CID: 64
> > > >         Result: Connection successful (0x0000)
> > > >         Status: No further information available (0x0000)
> > > >
> > > > ////// WITH PATCH //////
> > > > > ACL Data RX: Handle 256 flags 0x02 dlen 12        #42 [hci0] 4.887024
> > > >       L2CAP: Connection Request (0x02) ident 3 len 4
> > > >         PSM: 4097 (0x1001)
> > > >         Source CID: 64
> > > > < ACL Data TX: Handle 256 flags 0x00 dlen 16        #43 [hci0] 4.887127
> > > >       L2CAP: Connection Response (0x03) ident 3 len 8
> > > >         Destination CID: 0
> > > >         Source CID: 64
> > > >         Result: Connection refused - security block (0x0003)
> > > >         Status: No further information available (0x0000)
> > > >
> > > > Signed-off-by: Archie Pusaka <apusaka@xxxxxxxxxxxx>
> > > > Reviewed-by: Alain Michaud <alainm@xxxxxxxxxxxx>
> > > >
> > > > ---
> > > > Btw, it looks like the patch sent by Alex Lu with the title
> > > > [PATCH] Bluetooth: Fix the vulnerable issue on enc key size
> > > > also solves the exact same issue.
> > > >
> > > > Changes in v2:
> > > > * Add btmon trace to the commit message
> > > >
> > > >  net/bluetooth/l2cap_core.c | 3 ++-
> > > >  1 file changed, 2 insertions(+), 1 deletion(-)
> > > >
> > > > diff --git a/net/bluetooth/l2cap_core.c b/net/bluetooth/l2cap_core.c
> > > > index ade83e224567..b4fc0ad38aaa 100644
> > > > --- a/net/bluetooth/l2cap_core.c
> > > > +++ b/net/bluetooth/l2cap_core.c
> > > > @@ -4101,7 +4101,8 @@ static struct l2cap_chan *l2cap_connect(struct l2cap_conn *conn,
> > > >
> > > >         /* Check if the ACL is secure enough (if not SDP) */
> > > >         if (psm != cpu_to_le16(L2CAP_PSM_SDP) &&
> > > > -           !hci_conn_check_link_mode(conn->hcon)) {
> > > > +           (!hci_conn_check_link_mode(conn->hcon) ||
> > > > +           !l2cap_check_enc_key_size(conn->hcon))) {
> > >
> > > I wonder if we couldn't incorporate the check of key size into
> > > hci_conn_check_link_mode, like I said in the first patch checking the
> > > enc key size should not be specific to L2CAP.
> >
> > Yes, I could move the check into hci_conn_check_link_mode.
> > At first look, this function is also called by AMP which I am not
> > familiar with. In addition, I found this patch which moves this check
> > outside hci_conn, so I have my doubts there.
> > https://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth-next.git/commit/?id=693cd8ce3f882524a5d06f7800dd8492411877b3
>
> Right, I think we can have it as part of the hci_conn_check_link_mode,
> that said it is perhaps better to have it as
> hci_conn_check_enc_key_size instead as it is not L2CAP expecific.
> Other than that it looks good to me.

Do you mean we should move l2cap_conn_check_enc_key_size to
hci_conn_check_enc_key_size? I think that is a good idea.
We also have hci_conn_check_secure which I am unsure what the purpose
is. I'll try to merge them together.

>
> > >
> > > >                 conn->disc_reason = HCI_ERROR_AUTH_FAILURE;
> > > >                 result = L2CAP_CR_SEC_BLOCK;
> > > >                 goto response;
> > > > --
> > > > 2.28.0.681.g6f77f65b4e-goog
> > > >
> > >
> > >
> > > --
> > > Luiz Augusto von Dentz
> >
> > Thanks,
> > Archie
>
>
>
> --
> Luiz Augusto von Dentz

Thanks,
Archie