fix `native_thread_destroy()` timing

With M:N thread scheduler, the native thread (NT) related resources
should be freed when the NT is no longer needed. So the calling
`native_thread_destroy()` at the end of `is will be freed when
`thread_cleanup_func()` (at the end of Ruby thread) is not correct
timing. Call it when the corresponding Ruby thread is collected.

1 files changed, 3 insertions, 6 deletions

diff --git a/thread.c b/thread.c
index d59e3d8773..71e33b164a 100644
--- a/thread.c
+++ b/thread.c
@@ -508,16 +508,13 @@ thread_cleanup_func(void *th_ptr, int atfork)
      * Unfortunately, we can't release native threading resource at fork
      * because libc may have unstable locking state therefore touching
      * a threading resource may cause a deadlock.
-     *
-     * FIXME: Skipping native_mutex_destroy(pthread_mutex_destroy) is safe
-     * with NPTL, but native_thread_destroy calls pthread_cond_destroy
-     * which calls free(3), so there is a small memory leak atfork, here.
      */
-    if (atfork)
+    if (atfork) {
+        th->nt = NULL;
         return;
+    }
 
     rb_native_mutex_destroy(&th->interrupt_lock);
-    native_thread_destroy(th);
 }
 
 static VALUE rb_threadptr_raise(rb_thread_t *, int, VALUE *);