On Mon, Oct 5, 2015 at 1:21 AM, Johan Hedberg <johan.hedberg@xxxxxxxxx> wrote:
> Hi Jakub,
>
> On Wed, Sep 02, 2015, Jakub Pawlowski wrote:
>> This patch adds handling of BT_HCI_CMD_LE_CREATE_CONN_CANCEL command.
>>
>> If btdev_set_le_noresp_conn_request is called on btdev, other devices that
>> try to connect to it will stuck on BT_HCI_CMD_LE_CREATE_CONN, by not
>> sending bt_hci_evt_le_conn_complete event back. Thanks to that,
>> BT_HCI_CMD_LE_CREATE_CONN_CANCEL can be triggered to cancel connect
>> attempt.
>> ---
>> emulator/btdev.c | 45 +++++++++++++++++++++++++++++++++++++++++++++
>> emulator/btdev.h | 2 ++
>> emulator/hciemu.c | 8 ++++++++
>> emulator/hciemu.h | 3 +++
>> 4 files changed, 58 insertions(+)
>>
>> diff --git a/emulator/btdev.c b/emulator/btdev.c
>> index 60e8e8c..1645776 100644
>> --- a/emulator/btdev.c
>> +++ b/emulator/btdev.c
>> @@ -145,6 +145,11 @@ struct btdev {
>> uint16_t sync_train_interval;
>> uint32_t sync_train_timeout;
>> uint8_t sync_train_service_data;
>> +
>> + bool le_noresp_conn_request;
>> + bool pending_le_conn;
>> + uint8_t pending_le_conn_addr[6];
>> + uint8_t pending_le_conn_addr_type;
>> };
>>
>> struct inquiry_data {
>> @@ -674,6 +679,11 @@ bool btdev_is_le_scan_enabled(struct btdev *btdev)
>> return btdev->le_scan_enable;
>> }
>>
>> +void btdev_set_le_noresp_conn_request(struct btdev *btdev, bool value)
>> +{
>> + btdev->le_noresp_conn_request = value;
>> +}
>> +
>> static bool use_ssp(struct btdev *btdev1, struct btdev *btdev2)
>> {
>> if (btdev1->auth_enable || btdev2->auth_enable)
>> @@ -1134,6 +1144,9 @@ static void le_conn_complete(struct btdev *btdev,
>> cc->handle = cpu_to_le16(42);
>> }
>>
>> + if (btdev->pending_le_conn)
>> + btdev->pending_le_conn = false;
>> +
>> cc->status = status;
>> cc->peer_addr_type = bdaddr_type;
>> memcpy(cc->peer_addr, bdaddr, 6);
>> @@ -1183,6 +1196,13 @@ static void le_conn_request(struct btdev *btdev, const uint8_t *bdaddr,
>> {
>> struct btdev *remote = find_btdev_by_bdaddr_type(bdaddr, bdaddr_type);
>>
>> + if (remote && remote->le_noresp_conn_request) {
>> + btdev->pending_le_conn = true;
>> + memcpy(btdev->pending_le_conn_addr, bdaddr, 6);
>> + btdev->pending_le_conn_addr_type = bdaddr_type;
>> + return;
>> + }
>
> btdev.c has a hooks mechanism to avoid adding test-specific complexity
> from creeping in by letting the exception handling be in the test code
> itself. In this case, couldn't you use command hooks to deal with the
> le_create_conn & le_create_conn_cancel commands in the tester and
> prevent the default handlers for them from running in btdev.c (by
> returning 'false' from the hook callback)?
Thanks for pointing me to hooks. It actually made this code shorter.
>
> Johan
--
To unsubscribe from this list: send the line "unsubscribe linux-bluetooth" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
