Fix data race in on_switch_shutdown_request() (#2931)

The data race in on_switch_shutdown_request() is the following:

* the on_switch_shutdown_request() calls exit() which calls the destructors for global static variables (e.g.BufferOrch::m_buffer_type_maps) * in parallel to that, orchagent accesses the global static variables

A fix is to avoid calling the destructors by using quick_exit() instead of exit()

orchagent/notifications.cpp

@@ -5,6 +5,10 @@ extern "C" {
 #include "logger.h"
 #include "notifications.h"
 
+#ifdef ASAN_ENABLED
+#include <sanitizer/lsan_interface.h>
+#endif
+
 void on_fdb_event(uint32_t count, sai_fdb_event_notification_data_t *data)
 {
     // don't use this event handler, because it runs by libsairedis in a separate thread
@@ -30,5 +34,16 @@ void on_switch_shutdown_request(sai_object_id_t switch_id)
     /* TODO: Later a better restart story will be told here */
     SWSS_LOG_ERROR("Syncd stopped");
 
-    exit(EXIT_FAILURE);
+    /*
+        The quick_exit() is used instead of the exit() to avoid a following data race:
+            * the exit() calls the destructors for global static variables (e.g.BufferOrch::m_buffer_type_maps)
+            * in parallel to that, orchagent accesses the global static variables
+        Since quick_exit doesn't call atexit() flows, the LSAN check is called explicitly via __lsan_do_leak_check()
+    */
+
+#ifdef ASAN_ENABLED
+    __lsan_do_leak_check();
+#endif
+
+    quick_exit(EXIT_FAILURE);
 }