HCIComm thread safety, resolve deadlock; Renaming: MgmtHandler -> DBTManager, revert DBTSession -> HCISession

HCIComm thread safety

With the java DBTAdapter, we have introduces multithreading access
to the HCIComm instance: start/stop etc + discovery-thread.

We have experienced a deadlock issuing stopDiscovery()
while the discovery-thread is in progress.

HCIComm and its dd is now locked using a recursive mutex,
which is also being exposed to HCISession allowing
DBTAdapter's discovery to use it as well.

+++

Renaming: MgmtHandler -> DBTManager, revert DBTSession -> HCISession

HCISession is not a major DBT (direct-bt) (java) instance
and actually describing the HCI(Comm) session.

DBTManager is a major DBT java instance, mapped.
Hence also splitting MgmtComm.* -> MgmtTypes.* + DBTManager.*

1 files changed, 1 insertions, 1 deletions

diff --git a/examples/direct_bt_scanner/dbt_scanner.cpp b/examples/direct_bt_scanner/dbt_scanner.cpp
index fcce3a22..81611565 100644
--- a/examples/direct_bt_scanner/dbt_scanner.cpp
+++ b/examples/direct_bt_scanner/dbt_scanner.cpp
@@ -142,7 +142,7 @@ int main(int argc, char *argv[])
 
     const int64_t t0 = direct_bt::getCurrentMilliseconds();
 
-    std::shared_ptr<direct_bt::DBTSession> session = adapter.open();
+    std::shared_ptr<direct_bt::HCISession> session = adapter.open();
 
     while( ok && !done && nullptr != session ) {
         ok = adapter.startDiscovery(*session);