diff --git a/Modules/posixmodule.c b/Modules/posixmodule.c
index 19e925730a5110a..5bd4e77a6c58142 100644
--- a/Modules/posixmodule.c
+++ b/Modules/posixmodule.c
@@ -613,11 +613,16 @@ PyOS_BeforeFork(void)
     run_at_forkers(interp->before_forkers, 1);
 
     _PyImport_AcquireLock(interp);
+    _PyEval_StopTheWorldAll(&_PyRuntime);
+    HEAD_LOCK(&_PyRuntime);
 }
 
 void
 PyOS_AfterFork_Parent(void)
 {
+    HEAD_UNLOCK(&_PyRuntime);
+    _PyEval_StartTheWorldAll(&_PyRuntime);
+
     PyInterpreterState *interp = _PyInterpreterState_GET();
     if (_PyImport_ReleaseLock(interp) <= 0) {
         Py_FatalError("failed releasing import lock after fork");
@@ -632,6 +637,7 @@ PyOS_AfterFork_Child(void)
     PyStatus status;
     _PyRuntimeState *runtime = &_PyRuntime;
 
+    // re-creates runtime->interpreters.mutex (HEAD_UNLOCK)
     status = _PyRuntimeState_ReInitThreads(runtime);
     if (_PyStatus_EXCEPTION(status)) {
         goto fatal_error;
@@ -7858,9 +7864,9 @@ os_fork1_impl(PyObject *module)
         /* child: this clobbers and resets the import lock. */
         PyOS_AfterFork_Child();
     } else {
-        warn_about_fork_with_threads("fork1");
         /* parent: release the import lock. */
         PyOS_AfterFork_Parent();
+        warn_about_fork_with_threads("fork1");
     }
     if (pid == -1) {
         errno = saved_errno;
@@ -7906,9 +7912,9 @@ os_fork_impl(PyObject *module)
         /* child: this clobbers and resets the import lock. */
         PyOS_AfterFork_Child();
     } else {
-        warn_about_fork_with_threads("fork");
         /* parent: release the import lock. */
         PyOS_AfterFork_Parent();
+        warn_about_fork_with_threads("fork");
     }
     if (pid == -1) {
         errno = saved_errno;
@@ -8737,9 +8743,9 @@ os_forkpty_impl(PyObject *module)
         /* child: this clobbers and resets the import lock. */
         PyOS_AfterFork_Child();
     } else {
-        warn_about_fork_with_threads("forkpty");
         /* parent: release the import lock. */
         PyOS_AfterFork_Parent();
+        warn_about_fork_with_threads("forkpty");
     }
     if (pid == -1) {
         return posix_error();
diff --git a/Python/pylifecycle.c b/Python/pylifecycle.c
index 3a2c0a450ac9d97..bc76822e72c54a9 100644
--- a/Python/pylifecycle.c
+++ b/Python/pylifecycle.c
@@ -1911,6 +1911,9 @@ Py_FinalizeEx(void)
     int malloc_stats = tstate->interp->config.malloc_stats;
 #endif
 
+    /* Ensure that remaining threads are detached */
+    _PyEval_StopTheWorldAll(runtime);
+
     /* Remaining daemon threads will automatically exit
        when they attempt to take the GIL (ex: PyEval_RestoreThread()). */
     _PyInterpreterState_SetFinalizing(tstate->interp, tstate);
diff --git a/Python/pystate.c b/Python/pystate.c
index 9f142223aff340f..69c73f3c5880650 100644
--- a/Python/pystate.c
+++ b/Python/pystate.c
@@ -1693,6 +1693,10 @@ _PyThreadState_DeleteExcept(PyThreadState *tstate)
     PyInterpreterState *interp = tstate->interp;
     _PyRuntimeState *runtime = interp->runtime;
 
+#ifdef Py_GIL_DISABLED
+    assert(runtime->stoptheworld.world_stopped);
+#endif
+
     HEAD_LOCK(runtime);
     /* Remove all thread states, except tstate, from the linked list of
        thread states.  This will allow calling PyThreadState_Clear()
@@ -1711,6 +1715,8 @@ _PyThreadState_DeleteExcept(PyThreadState *tstate)
     interp->threads.head = tstate;
     HEAD_UNLOCK(runtime);
 
+    _PyEval_StartTheWorldAll(runtime);
+
     /* Clear and deallocate all stale thread states.  Even if this
        executes Python code, we should be safe since it executes
        in the current thread, not one of the stale threads. */