sched, modules: Fix nested sleep in add_unformed_module()

This is a genuine bug in add_unformed_module(), we cannot use blocking
primitives inside a wait loop.

So rewrite the wait_event_interruptible() usage to use the fresh
wait_woken() stuff.

Reported-by: Fengguang Wu <[email protected]>
Signed-off-by: Peter Zijlstra (Intel) <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Cc: Rusty Russell <[email protected]>
Cc: [email protected]
Cc: Linus Torvalds <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: Greg Kroah-Hartman <[email protected]>
Link: http://lkml.kernel.org/r/[email protected]
[ So this is probably complex to backport and the race wasn't reported AFAIK,
  so not marked for -stable. ]
Signed-off-by: Ingo Molnar <[email protected]>

Signed-off-by: Ingo Molnar <[email protected]>

Author: Peter Zijlstra

kernel/module.c

@@ -3096,6 +3096,32 @@ static int may_init_module(void)
 	return 0;
 }
 
+/*
+ * Can't use wait_event_interruptible() because our condition
+ * 'finished_loading()' contains a blocking primitive itself (mutex_lock).
+ */
+static int wait_finished_loading(struct module *mod)
+{
+	DEFINE_WAIT_FUNC(wait, woken_wake_function);
+	int ret = 0;
+
+	add_wait_queue(&module_wq, &wait);
+	for (;;) {
+		if (finished_loading(mod->name))
+			break;
+
+		if (signal_pending(current)) {
+			ret = -ERESTARTSYS;
+			break;
+		}
+
+		wait_woken(&wait, TASK_INTERRUPTIBLE, MAX_SCHEDULE_TIMEOUT);
+	}
+	remove_wait_queue(&module_wq, &wait);
+
+	return ret;
+}
+
 /*
  * We try to place it in the list now to make sure it's unique before
  * we dedicate too many resources.  In particular, temporary percpu
@@ -3116,8 +3142,8 @@ static int add_unformed_module(struct module *mod)
 		    || old->state == MODULE_STATE_UNFORMED) {
 			/* Wait in case it fails to load. */
 			mutex_unlock(&module_mutex);
-			err = wait_event_interruptible(module_wq,
-					       finished_loading(mod->name));
+
+			err = wait_finished_loading(mod);
 			if (err)
 				goto out_unlocked;
 			goto again;