pythongh-136870: fix data race in PyThreadState_Clear on sys_tracing_threads (python#136951)

In free-threading, multiple threads can be cleared concurrently as such the modifications on `sys_tracing_threads` should be done while holding the profile lock, otherwise it can race with other threads setting up profiling.

Author: kumaraditya303

Python/pystate.c

@@ -1682,6 +1682,10 @@ PyThreadState_Clear(PyThreadState *tstate)
           "PyThreadState_Clear: warning: thread still has a generator\n");
     }
 
+#ifdef Py_GIL_DISABLED
+    PyMutex_Lock(&_PyRuntime.ceval.sys_trace_profile_mutex);
+#endif
+
     if (tstate->c_profilefunc != NULL) {
         tstate->interp->sys_profiling_threads--;
         tstate->c_profilefunc = NULL;
@@ -1690,6 +1694,11 @@ PyThreadState_Clear(PyThreadState *tstate)
         tstate->interp->sys_tracing_threads--;
         tstate->c_tracefunc = NULL;
     }
+
+#ifdef Py_GIL_DISABLED
+    PyMutex_Unlock(&_PyRuntime.ceval.sys_trace_profile_mutex);
+#endif
+
     Py_CLEAR(tstate->c_profileobj);
     Py_CLEAR(tstate->c_traceobj);