Free TCB (and thread-local storage) area on thread destruction.

We forgot to free the TCB allocated buffer on thread destruction, causing
a leak of around 1000 bytes per thread creation. We still have another
leak (exactly one page per thread object creation) that I'm trying to find :(

arch/x64/arch-switch.hh

@@ -93,6 +93,11 @@ void thread::setup_tcb()
     _tcb->tls_base = p;
 }
 
+void thread::free_tcb()
+{
+    free(_tcb->tls_base);
+}
+
 void thread_main_c(thread* t)
 {
     s_current = t;

core/sched.cc

@@ -276,6 +276,7 @@ thread::~thread()
     if (_attr.stack.deleter) {
         _attr.stack.deleter(_attr.stack.begin);
     }
+    free_tcb();
 }
 
 void thread::start()

include/sched.hh

@@ -175,6 +175,7 @@ private:
     void stop_wait();
     void init_stack();
     void setup_tcb();
+    void free_tcb();
     void complete();
     void suspend_timers();
     void resume_timers();