Clarify some comments in SMP code

Spotted by Joona.

No functional change.

src/search.cpp

@@ -1547,8 +1547,7 @@ void Thread::idle_loop() {
 
           // After releasing the lock we can't access any SplitPoint related data
           // in a safe way because it could have been released under our feet by
-          // the sp master. Also accessing other Thread objects is unsafe because
-          // if we are exiting there is a chance that they are already freed.
+          // the sp master.
           sp->mutex.unlock();
 
           // Try to late join to another split point if none of its slaves has

src/thread.cpp

@@ -194,7 +194,8 @@ void ThreadPool::init() {
 }
 
 
-// exit() cleanly terminates the threads before the program exits
+// exit() cleanly terminates the threads before the program exits. Cannot be done in
+// d'tor because we have to terminate the threads before to free ThreadPool object.
 
 void ThreadPool::exit() {