Correctly clean up `keeping_mutexes` before resuming any other threads. (#7460)

It's possible (but very rare) to have a race condition between setting
`mutex->fiber = NULL` and `thread_mutex_remove(th, mutex)` which results
in the following bug:

```
[BUG] invalid keeping_mutexes: Attempt to unlock a mutex which is not locked
```

Fixes <https://bugs.ruby-lang.org/issues/19480>.

1 files changed, 1 insertions, 1 deletions

diff --git a/thread.c b/thread.c
index cf8b9ed87b..4e662f6b66 100644
--- a/thread.c
+++ b/thread.c
@@ -428,7 +428,7 @@ rb_threadptr_unlock_all_locking_mutexes(rb_thread_t *th)
         rb_mutex_t *mutex = th->keeping_mutexes;
         th->keeping_mutexes = mutex->next_mutex;
 
-        /* rb_warn("mutex #<%p> remains to be locked by terminated thread", (void *)mutexes); */
+        // rb_warn("mutex #<%p> was not unlocked by thread #<%p>", (void *)mutex, (void*)th);
 
         const char *error_message = rb_mutex_unlock_th(mutex, th, mutex->fiber);
         if (error_message) rb_bug("invalid keeping_mutexes: %s", error_message);