Bluetooth: Host: Revert sending Host Number Of Completed Packets as early as possible

[PavelVPV]

This PR reverts #75779 as SoftDevice Controller does not any longer send the disconnection event only after Host confirms all received ACL data throguh Host Number Of Completed Packets command.

[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]

What if other controllers do this?

[PavelVPV]

What if other controllers do this?

This was specifically added for SDC. I don't know if other controllers do this.

Also, this approach showed that regardless of number of internal ACL buffers Host ends up in a situation where it has less buffers than Controller thinks Host has. IOW, it always ends up in desynchronization of number of ACL buffers on Host side between Host and Controller. It is neither possible to predict future and stop reporting HNCP to avoid desynchronization, because this in turns will lead to a deadlock between Host and Controller when one needs buffers to complete recombination and the other needs HNCP to to send the disconnection event.