Re: [RFC] Bluetooth: Fix using address type from events

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Marcel,

On Mon, Aug 30, 2021 at 7:57 AM Marcel Holtmann <marcel@xxxxxxxxxxxx> wrote:
>
> Hi Luiz,
>
> > Address types ADDR_LE_DEV_PUBLIC_RESOLVED and
> > ADDR_LE_DEV_RANDOM_RESOLVED shall be converted to ADDR_LE_PUBLIC and
> > ADDR_LE_RANDOM repectively since they are not safe to be used beyond
> > the scope of the events themselves.
> >
> > Also don't connect using RPA as peer address when LL Privacy is
> > enabled since the controller will be resolving the addresses it should
> > be able to generate the RPA.
> >
> > Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@xxxxxxxxx>
> > ---
> > net/bluetooth/hci_conn.c  | 10 +++++---
> > net/bluetooth/hci_core.c  |  9 -------
> > net/bluetooth/hci_event.c | 52 ++++++++++++++++++---------------------
> > 3 files changed, 30 insertions(+), 41 deletions(-)
> >
> > diff --git a/net/bluetooth/hci_conn.c b/net/bluetooth/hci_conn.c
> > index bdaa64cb2f3a..28f23654b8a6 100644
> > --- a/net/bluetooth/hci_conn.c
> > +++ b/net/bluetooth/hci_conn.c
> > @@ -1087,10 +1087,12 @@ struct hci_conn *hci_connect_le(struct hci_dev *hdev, bdaddr_t *dst,
> >        * be resolved back into the original identity address
> >        * from the connect request.
> >        */
> > -     irk = hci_find_irk_by_addr(hdev, dst, dst_type);
> > -     if (irk && bacmp(&irk->rpa, BDADDR_ANY)) {
> > -             dst = &irk->rpa;
> > -             dst_type = ADDR_LE_DEV_RANDOM;
> > +     if (!hci_dev_test_flag(hdev, HCI_LL_RPA_RESOLUTION)) {
>
> I do not understand this pre-caution with HCI_LL_RPA_RESOLUTION. What is this for?

At least with Intel controllers I could not reconnect if we keep using
the RPA when the IRK has been added to the resolving list, I suspect
it may be a bug in the firmware but in the other hand it may not be
necessary to use the original RPA as the controller has the IRK as
well it can generate the RPA before connecting.

> > +             irk = hci_find_irk_by_addr(hdev, dst, dst_type);
> > +             if (irk && bacmp(&irk->rpa, BDADDR_ANY)) {
> > +                     dst = &irk->rpa;
> > +                     dst_type = ADDR_LE_DEV_RANDOM;
> > +             }
> >       }
> >
> >       if (conn) {
> > diff --git a/net/bluetooth/hci_core.c b/net/bluetooth/hci_core.c
> > index 6ab0ff3cd4bc..e83a3accfc33 100644
> > --- a/net/bluetooth/hci_core.c
> > +++ b/net/bluetooth/hci_core.c
> > @@ -3562,15 +3562,6 @@ struct hci_conn_params *hci_pend_le_action_lookup(struct list_head *list,
> > {
> >       struct hci_conn_params *param;
> >
> > -     switch (addr_type) {
> > -     case ADDR_LE_DEV_PUBLIC_RESOLVED:
> > -             addr_type = ADDR_LE_DEV_PUBLIC;
> > -             break;
> > -     case ADDR_LE_DEV_RANDOM_RESOLVED:
> > -             addr_type = ADDR_LE_DEV_RANDOM;
> > -             break;
> > -     }
> > -
> >       list_for_each_entry(param, list, action) {
> >               if (bacmp(&param->addr, addr) == 0 &&
> >                   param->addr_type == addr_type)
> > diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c
> > index 24d0773b861a..028cdf7ef075 100644
> > --- a/net/bluetooth/hci_event.c
> > +++ b/net/bluetooth/hci_event.c
> > @@ -2411,6 +2411,24 @@ static void hci_cs_disconnect(struct hci_dev *hdev, u8 status)
> >       hci_dev_unlock(hdev);
> > }
> >
> > +static u8 ev_bdaddr_type(struct hci_dev *hdev, u8 type)
> > +{
> > +     /* When using controller based address resolution, then the new
> > +      * address types 0x02 and 0x03 are used. These types need to be
> > +      * converted back into either public address or random address type
> > +      */
> > +     if (hci_dev_test_flag(hdev, HCI_LL_RPA_RESOLUTION)) {
> > +             switch (type) {
> > +             case ADDR_LE_DEV_PUBLIC_RESOLVED:
> > +                     return ADDR_LE_DEV_PUBLIC;
> > +             case ADDR_LE_DEV_RANDOM_RESOLVED:
> > +                     return ADDR_LE_DEV_RANDOM;
> > +             }
> > +     }
> > +
>
> The check for HCI_LL_RPA_RESOLUTION makes no sense to me. It is always safe to convert addr_type 0x02 and 0x03. You really only see them if resolving list usage is active.

I will remove it, it was part of the original code but you are right
we should convert these types regardless.

> > +     return type;
> > +}
> > +
> > static void cs_le_create_conn(struct hci_dev *hdev, bdaddr_t *peer_addr,
> >                             u8 peer_addr_type, u8 own_address_type,
> >                             u8 filter_policy)
> > @@ -2422,20 +2440,7 @@ static void cs_le_create_conn(struct hci_dev *hdev, bdaddr_t *peer_addr,
> >       if (!conn)
> >               return;
> >
> > -     /* When using controller based address resolution, then the new
> > -      * address types 0x02 and 0x03 are used. These types need to be
> > -      * converted back into either public address or random address type
> > -      */
> > -     if (hci_dev_test_flag(hdev, HCI_LL_RPA_RESOLUTION)) {
> > -             switch (own_address_type) {
> > -             case ADDR_LE_DEV_PUBLIC_RESOLVED:
> > -                     own_address_type = ADDR_LE_DEV_PUBLIC;
> > -                     break;
> > -             case ADDR_LE_DEV_RANDOM_RESOLVED:
> > -                     own_address_type = ADDR_LE_DEV_RANDOM;
> > -                     break;
> > -             }
> > -     }
> > +     own_address_type = ev_bdaddr_type(hdev, own_address_type);
> >
> >       /* Store the initiator and responder address information which
> >        * is needed for SMP. These values will not change during the
> > @@ -5333,20 +5338,7 @@ static void le_conn_complete_evt(struct hci_dev *hdev, u8 status,
> >               conn->dst_type = irk->addr_type;
> >       }
> >
> > -     /* When using controller based address resolution, then the new
> > -      * address types 0x02 and 0x03 are used. These types need to be
> > -      * converted back into either public address or random address type
> > -      */
> > -     if (hci_dev_test_flag(hdev, HCI_LL_RPA_RESOLUTION)) {
> > -             switch (conn->dst_type) {
> > -             case ADDR_LE_DEV_PUBLIC_RESOLVED:
> > -                     conn->dst_type = ADDR_LE_DEV_PUBLIC;
> > -                     break;
> > -             case ADDR_LE_DEV_RANDOM_RESOLVED:
> > -                     conn->dst_type = ADDR_LE_DEV_RANDOM;
> > -                     break;
> > -             }
> > -     }
> > +     conn->dst_type = ev_bdaddr_type(hdev, conn->dst_type);
> >
>
> I see, it comes from the original code.
>
> Frankly, I fail to see why any of these checks are needed. When we get that value, we need to convert it. So we better remove these pointless checks unless someone tells me the reason for it. Greping through the code seems we have even more and even more complex ones.

Yep, I tried making all of them use the same function so 0x02 and 0x03
are never passed around without being converted first.

> >       if (status) {
> >               hci_le_conn_failed(conn, status);
> > @@ -5663,6 +5655,8 @@ static void process_adv_report(struct hci_dev *hdev, u8 type, bdaddr_t *bdaddr,
> >        * controller address.
> >        */
> >       if (direct_addr) {
> > +             direct_addr_type = ev_bdaddr_type(hdev, direct_addr_type);
> > +
> >               /* Only resolvable random addresses are valid for these
> >                * kind of reports and others can be ignored.
> >                */
> > @@ -5690,6 +5684,8 @@ static void process_adv_report(struct hci_dev *hdev, u8 type, bdaddr_t *bdaddr,
> >               bdaddr_type = irk->addr_type;
> >       }
> >
> > +     bdaddr_type = ev_bdaddr_type(hdev, bdaddr_type);
> > +
> >       /* Check if we have been requested to connect to this device.
> >        *
> >        * direct_addr is set only for directed advertising reports (it is NULL
>
> Regards
>
> Marcel
>


-- 
Luiz Augusto von Dentz




[Index of Archives]     [Bluez Devel]     [Linux Wireless Networking]     [Linux Wireless Personal Area Networking]     [Linux ATH6KL]     [Linux USB Devel]     [Linux Media Drivers]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [Big List of Linux Books]

  Powered by Linux