Contradictory security parameters for MITM protection?

[Eliot Stock]

Am I right in thinking that these security parameters are contradictory in that for the peripheral (nRF51822) to provide MITM protection, there must be either IO capabilities or out of band capabilities?

#define SEC_PARAM_TIMEOUT                 30
#define SEC_PARAM_BOND                    1
#define SEC_PARAM_MITM                    1
#define SEC_PARAM_IO_CAPABILITIES         BLE_GAP_IO_CAPS_NONE
#define SEC_PARAM_OOB                     0
#define SEC_PARAM_MIN_KEY_SIZE            7
#define SEC_PARAM_MAX_KEY_SIZE            16

The Soft Device seems to have no problem in accepting them:

    err_code = sd_ble_gap_sec_params_reply(m_conn_handle,
                                           BLE_GAP_SEC_STATUS_SUCCESS,
                                           &m_sec_params);
    APP_ERROR_CHECK(err_code);

[Pål Håland]

Hi Eliot,

You are right in thinking that the parameters is contradictory. The softdevice will always select the correct method of pairing/bonding no matter what IO_CAPABILITIES and OOB is set to.

If this is a problem for the application, he can respond with a status other than success in the sd_ble_gap_sec_params_resp and the stack will send a pairing failed instead.

BR Pål

[cocoa]

Hi Pål,

I've not well understood : "The softdevice will always select the correct method of pairing/bonding no matter what IO_CAPABILITIES and OOB is set to."

In which way SD select the correct method? Are the settings defined (SEC_PARAM_xxx) overrided by SD ?

How works this method ?

How manage this to be sure that my settings are mandatory for SD? Thank you.

-c

[Pål Håland]

I should probably rephrase the statement to say that it selects the supported pairing method based on what we send in the pairing response and what we receive in the pairing request. So if the IO is no input no output and there is no OOB, we will always use Just Works.