Bluetooth: Classic: Fix LTK cannot be derived issue

[lylezhu2012]

The LTK cannot be derived by LK when LK is not weaker than the old LTK.

Improve the function smp_br_pairing_allowed() to avoid the LTK be overwrote when old LTK has MITM protection but new LK has not MITM protection.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[Thalley]

@alwa-nordic another case where a classic-specific change is not assigned any classic collaborators.

@lylezhu2012 we discussed in the BT meeting yesterday that there are many files (such as this) that isn't under the Bluetooth Classic: area in MAINTAINERS.yml, so they won't get the appropriate reviewers for Classic only changes.

We discussed 2 options:

1. Add shared files between LE and Classic to the classic area (such as conn.c, smp.c, l2cap.c, etc).
2. Try to modify the CI system to assign people to reviewer based on labels, i.e. adding the BT classic label here could possibly add all classic collaborators.

Currently the documentation states that the assignee (e.g. @alwa-nordic ) is responsible for driving the PR to a mergable state (which includes assigning the right reviewers), but that is a manual process I think is too much to handle for BT.

[lylezhu2012]

@Thalley Understood. Thanks a lot.

1. Add shared files between LE and Classic to the classic area (such as conn.c, smp.c, l2cap.c, etc).

Do I need to do this now?

[Thalley]

@Thalley Understood. Thanks a lot.

1. Add shared files between LE and Classic to the classic area (such as conn.c, smp.c, l2cap.c, etc).

Do I need to do this now?

I'll leave that up to you. It will ensure that people interested in classic will get the changes and automatically added as reviewers, but it will also add them as reviewers to a lot more PRs.