Fix double-loading of extension

Previously, there was a case involving cache invalidation when
the extension could be preloaded. Namely the following happened:

1) extension_check (outer) was called and found loaded = true.
2) While the extension_check was checking the state of the extension
   a cache invalidation callback happened.
3) The cache invalidation callback called extension_check (inner).
4) The inner extension_check loads the extension and sets
   loaded = true.
5) The outer extension_check also loads the extension after the
   cache invalidation callback returns. This happens because it
   never re-checks the value of loaded.

This commit adds another check of loaded just before it is changed
to prevent this situation.

src/loader/loader.c

@@ -273,6 +273,22 @@ do_load()
 	StrNCpy(soversion, version, MAX_VERSION_LEN);
 	snprintf(soname, MAX_SO_NAME_LEN, "%s-%s", EXTENSION_NAME, version);
 
+	/*
+	 * An inval_relcache callback can be called after previous checks of
+	 * loaded had found it to be false. But the inval_relcache callback may
+	 * load the extension setting it to true. Thus it needs to be rechecked
+	 * here again by the outer call after inval_relcache completes. This is
+	 * double-check locking, in effect.
+	 */
+	if (loaded)
+		return;
+
+	/*
+	 * Set to true whether or not the load succeeds to prevent reloading if
+	 * failure happened after partial load.
+	 */
+	loaded = true;
+
 	/*
 	 * we need to capture the loaded extension's post analyze hook, giving it
 	 * a NULL as previous
@@ -283,13 +299,9 @@ do_load()
 	PG_TRY();
 	{
 		load_file(soname, false);
-		loaded = true;
 	}
 	PG_CATCH();
 	{
-		/* Assume the extension was loaded to prevent re-loading another .so */
-		loaded = true;
-
 		extension_post_parse_analyze_hook = post_parse_analyze_hook;
 		post_parse_analyze_hook = old_hook;
 		PG_RE_THROW();