When to call pm_conn_secure()

[Bruno Randolf]

I am making a central application which connects to a peripheral and uses Peer Manager. The first time I connect, I have to call pm_conn_secure() to secure the link, basically any time after connecting, but before reading a (protected) characteristic, that works fine.

When the peer has already bonded before, then Peer manager automatically secures the link and I do not have to call pm_conn_secure(). That also works fine with the function commented out. If I call pm_conn_secure() it looks like the security proceduer is executed twice (I get the PM_EVT_CONN_SEC_SUCCEEDED event twice, on the central and on the peripheral).

Now I wonder how to handle these different cases, so that the procedure is executed only once. What would be the best event to start pm_conn_secure() only when needed?

[Bruno Randolf]

Ok, figured it out: I will receive PM_EVT_CONN_SEC_START before BLE_GAP_EVT_CONNECTED when the PM has started the security procedure if the my own SDH_BLE_OBSERVER prio is lower than the observer prio of peer manager.