Is it still possible to remove bond in CTS?

RE: Is it still possible to remove bond in CTS?

Edvin — Fri, 10 Jul 2020 09:17:21 GMT

There probably is. Try to debug into pm_sec_params_set() and see why it returns 7. The Peer Manager is separated from the softdevice, so everything in there is open source. Which check is it that fails, so that it returns 7?

BR,

Edvin

RE: Is it still possible to remove bond in CTS?

Simon — Thu, 09 Jul 2020 16:17:58 GMT

I'm looking at trying to reproduce INSUF_AUTHENTICATION.
Just toggling SEC_PARAM_BOND 0 in the cts example is giving me trouble,

00> <info> app_timer: RTC: initialized.
00>
00> <error> app: ERROR 7 [NRF_ERROR_INVALID_PARAM] at .....\nRF5_SDK_17.0.0\examples\ble_peripheral\ble_app_cts_c-2\main.c:894
00>
00> PC at: 0x0002BB7F
00>
00> <error> app: End of error report
00>

which is this spot,
pm_sec_params_set() returns != 0 (NRF_SUCCESS)

Is some other setting needed to change for SEC_PARAM_BOND 0 to work in the 17.0 example?
That is the only setting I'm changing in the example.

thank-you,

RE: Is it still possible to remove bond in CTS?

Edvin — Thu, 09 Jul 2020 13:15:48 GMT

Hello,

[quote user="foxble"](p.s. the reply seems to have been posted 2 days ago, but I only received notification today)[/quote]

I'll forward this internally. Thank you for reporting.

The description from Sigurd is correct. That is how you remove the bonding in an example. And it may be correct that pairing (short term bonding) is not required by the Current Time Service by the BLE specification, since I can't find that requirement on Bluetooth's webpage.

However, when I test with my iPhone, I am not able to read the CTS without encrypting (pairing or bonding) the connection (as you can see in the screenshot in my previous reply). This is something they can add themselves. So perhaps the user in Sigurds ticket, which you linked to, was not testing against an iPhone, but something else that doesn't put the encryption requirement on the service. He/she doesn't mention what the central in the connection is, as far as I can tell.

Now, I am not sure whether you got that far that you actually tried to read the CTS characteristic or not, or if you stopped at the SEC_PARAM_BOND 0, since it threw an error. If you define DEBUG in your preprocessor definitions. Then the log should point to the APP_ERROR_CHECK(err_code) that caught the error.

I guess the pm_sec_params_set() returns != 0 (NRF_SUCCESS). Try debugging into it with SEC_PARAM_BOND set to 0.

You should be able to disable bonding, but please be aware that at least on my iphone, I wasn't able to read the CTS without at least pairing.

If you are not sure about the difference between bonding and pairing, please check out Einar's answer in this ticket.

Best regards,

Edvin

RE: Is it still possible to remove bond in CTS?

Simon — Thu, 09 Jul 2020 12:16:50 GMT

luckxiaoer asked "how to remove the bonding procedure completely？",
Sigurd replied "Try this",
luckxiaoer answered back "It solved my problem".
https://devzone.nordicsemi.com/f/nordic-q-a/38214/remove-bond-in-cts/165249

Now after a couple of years it seems by looking at your nRF Connect test,
that it is now no longer possible to access CTS without bonding.

I was looking for a double check to be certain.

thank-you,

(p.s. the reply was posted 2 days ago, but I only received notification today)

RE: Is it still possible to remove bond in CTS?

Edvin — Tue, 07 Jul 2020 15:03:19 GMT

Hello,

I have not seen anywhere that you don't need bonding to access the current time service. Where is that information coming from?

I tried to access this in nRF Connect without success, due to insufficient authentication:

I don't understand how to read it from the bluetooth web pages anymore (they changed the layout), but perhaps you can point me to where it says, if you know where to find this information.

Remember that even though it is not mandatory to protect a service with security, they may still do it as an addon.

best regards,

Edvin