Re: [PATCH] Bluetooth: Update resolving list when updating whitelist

[Marcel Holtmann <marcel@xxxxxxxxxxxx>]

Hi Luiz,

>> When the whitelist is updated, then also update the entries of the
>> resolving list for devices where IRKs are available.
>> 
>> Signed-off-by: Marcel Holtmann <marcel@xxxxxxxxxxxx>
>> ---
>> net/bluetooth/hci_request.c | 37 +++++++++++++++++++++++++++++++++++--
>> 1 file changed, 35 insertions(+), 2 deletions(-)
>> 
>> diff --git a/net/bluetooth/hci_request.c b/net/bluetooth/hci_request.c
>> index efec2a0bb824..45fbda5323af 100644
>> --- a/net/bluetooth/hci_request.c
>> +++ b/net/bluetooth/hci_request.c
>> @@ -695,6 +695,21 @@ static void del_from_white_list(struct hci_request *req, bdaddr_t *bdaddr,
>>        bt_dev_dbg(req->hdev, "Remove %pMR (0x%x) from whitelist", &cp.bdaddr,
>>                   cp.bdaddr_type);
>>        hci_req_add(req, HCI_OP_LE_DEL_FROM_WHITE_LIST, sizeof(cp), &cp);
>> +
>> +       if (use_ll_privacy(req->hdev)) {
>> +               struct smp_irk *irk;
>> +
>> +               irk = hci_find_irk_by_addr(req->hdev, bdaddr, bdaddr_type);
>> +               if (irk) {
>> +                       struct hci_cp_le_del_from_resolv_list cp;
>> +
>> +                       cp.bdaddr_type = bdaddr_type;
>> +                       bacpy(&cp.bdaddr, bdaddr);
>> +
>> +                       hci_req_add(req, HCI_OP_LE_DEL_FROM_RESOLV_LIST,
>> +                                   sizeof(cp), &cp);
>> +               }
>> +       }
>> }
>> 
>> /* Adds connection to white list if needed. On error, returns -1. */
>> @@ -715,7 +730,7 @@ static int add_to_white_list(struct hci_request *req,
>>                return -1;
>> 
>>        /* White list can not be used with RPAs */
>> -       if (!allow_rpa &&
>> +       if (!allow_rpa && !use_ll_privacy(hdev) &&
>>            hci_find_irk_by_addr(hdev, &params->addr, params->addr_type)) {
>>                return -1;
>>        }
>> @@ -732,6 +747,24 @@ static int add_to_white_list(struct hci_request *req,
>>                   cp.bdaddr_type);
>>        hci_req_add(req, HCI_OP_LE_ADD_TO_WHITE_LIST, sizeof(cp), &cp);
>> 
>> +       if (use_ll_privacy(hdev)) {
>> +               struct smp_irk *irk;
>> +
>> +               irk = hci_find_irk_by_addr(hdev, &params->addr,
>> +                                          params->addr_type);
>> +               if (irk) {
>> +                       struct hci_cp_le_add_to_resolv_list cp;
>> +
>> +                       cp.bdaddr_type = params->addr_type;
>> +                       bacpy(&cp.bdaddr, &params->addr);
>> +                       memcpy(cp.peer_irk, irk->val, 16);
>> +                       memset(cp.local_irk, 0, 16);
>> +
>> +                       hci_req_add(req, HCI_OP_LE_ADD_TO_RESOLV_LIST,
>> +                                   sizeof(cp), &cp);
> 
> Shouldn't we be checking if there is any space left in the list before
> trying to send the command? I wonder what would happen if there is too
> many IRKs, I guess that means we would still have to resolve them in
> the host.

the resolving list size is just a guesstimate and no guarantee that space is available. However my assumption is that if you have x entries for whitespace, then you also get x entries for the resolving list. If we are not putting the device in the whitelist, then it also makes no sense to add it to the resolving list.

Regards

Marcel