Fix missing increment of the instance count on every call to startup()

Change how the implicit startup is tracked.
The extra change is needed since the startup() calls in
srt::CUDT::socket()
and
srt::CUDT::createGroup
would endlessly increment the startup counter as sockets were created
otherwise.

Force a clean() when the singleton is getting destructed.
Previously it just did a single decrement of the instance count,
which could result in attempting to exit without actually cleaning up.

fixes #3098

srtcore/api.cpp

@@ -198,13 +198,8 @@ srt::CUDTUnited::CUDTUnited()
 
 srt::CUDTUnited::~CUDTUnited()
 {
-    // Call it if it wasn't called already.
-    // This will happen at the end of main() of the application,
-    // when the user didn't call srt_cleanup().
-    if (m_bGCStatus)
-    {
-        cleanup();
-    }
+    // force cleanup, even if the instance count is > 1
+    cleanup(true);
 
     releaseMutex(m_GlobControlLock);
     releaseMutex(m_IDLock);
@@ -233,15 +228,24 @@ string srt::CUDTUnited::CONID(SRTSOCKET sock)
     return os.str();
 }
 
-int srt::CUDTUnited::startup()
+void srt::CUDTUnited::implicitStartup()
+{
+    startup(true);
+}
+
+int srt::CUDTUnited::startup(bool implicit)
 {
     ScopedLock gcinit(m_InitLock);
-    if (m_bGCStatus)
+
+    if (implicit && m_iInstanceCount > 0)
         return 1;
 
     if (m_iInstanceCount++ > 0)
         return 1;
 
+    if (m_bGCStatus)
+        return 1;
+
         // Global initialization code
 #ifdef _WIN32
     WORD    wVersionRequested;
@@ -268,7 +272,7 @@ int srt::CUDTUnited::startup()
     return 0;
 }
 
-int srt::CUDTUnited::cleanup()
+int srt::CUDTUnited::cleanup(bool force)
 {
     // IMPORTANT!!!
     // In this function there must be NO LOGGING AT ALL.  This function may
@@ -280,8 +284,14 @@ int srt::CUDTUnited::cleanup()
     // executing this procedure.
     ScopedLock gcinit(m_InitLock);
 
-    if (--m_iInstanceCount > 0)
-        return 0;
+    if (!force)
+    {
+        if (m_iInstanceCount == 0)
+            return 0;
+
+        if (--m_iInstanceCount > 0)
+            return 0;
+    }
 
     if (!m_bGCStatus)
         return 0;
@@ -298,6 +308,8 @@ int srt::CUDTUnited::cleanup()
     CSync::notify_one_relaxed(m_GCStopCond);
     m_GCThread.join();
 
+    if (force)
+        m_iInstanceCount = 0;
     m_bGCStatus = false;
 
     // Global destruction code
@@ -3469,7 +3481,7 @@ int srt::CUDT::cleanup()
 
 SRTSOCKET srt::CUDT::socket()
 {
-    uglobal().startup();
+    uglobal().implicitStartup();
 
     try
     {
@@ -3519,7 +3531,7 @@ srt::CUDTGroup& srt::CUDT::newGroup(const int type)
 SRTSOCKET srt::CUDT::createGroup(SRT_GROUP_TYPE gt)
 {
     // Doing the same lazy-startup as with srt_create_socket()
-    uglobal().startup();
+    uglobal().implicitStartup();
 
     try
     {

srtcore/api.h

@@ -257,12 +257,14 @@ class CUDTUnited
     static std::string CONID(SRTSOCKET sock);
 
     /// initialize the UDT library.
+    /// @param [in] implicit should be set to true for internal implicit startup() calls.
     /// @return 0 if success, otherwise -1 is returned.
-    int startup();
+    int startup(bool implicit = false);
 
     /// release the UDT library.
+    /// @param [in] force cleanup regardless of the instance count.
     /// @return 0 if success, otherwise -1 is returned.
-    int cleanup();
+    int cleanup(bool force = false);
 
     /// Create a new UDT socket.
     /// @param [out] pps Variable (optional) to which the new socket will be written, if succeeded
@@ -428,6 +430,7 @@ class CUDTUnited
     CUDTSocket* locateSocket_LOCKED(SRTSOCKET u);
     CUDTSocket* locatePeer(const sockaddr_any& peer, const SRTSOCKET id, int32_t isn);
 
+    void implicitStartup();
 #if ENABLE_BONDING
     CUDTGroup* locateAcquireGroup(SRTSOCKET u, ErrorHandling erh = ERH_RETURN);
     CUDTGroup* acquireSocketsGroup(CUDTSocket* s);