can private address be used to stop connection

[andrewjfox]

Hi,

I've read the BT specs covering addressing several times and still don't really understand how private addresses work in real life.

Lets say my peripheral starts out with no IRKs stored and has never been bonded. During user setup, a known central (iPhone) bonds with my peripheral and we share IRKs. At some point a malicious central (sniffer, hacker etc) comes along, and tries to spoof the known centrals address, however they won't have the IRKs. Will my peripheral deny a connection?

For non-resolvable address, what does peripheral advertise with, a random number, and when central conencts to peripheral with non-resolvable address which actual address does it use in connection request?

Thanks in advance Andrew

[RK]

If you don't have the IRK you can't try to spoof the address, the private resolvable address is made up of a random part plus the same random part encrypted with the IRK.

If the peripheral is set up only to accept connections from whitelisted peers then it will deny connection from anything not on the list. If that list contains some resolvable private addresses, ie basically IRKs as that's the important part, it will deny connection to any resolvable private address it sees which was not constructed using an IRK in that list.

[andrewjfox]

How does a new, legitimate, central bond with the device, if it's previously bonded with another central? Is there a way of 2 centrals sharing the IRKs so the new central can also connect to the peripheral?