MDEV-30422 Merge new release of InnoDB 5.7.41 to 10.3

MySQL 5.7.41 includes one InnoDB change
mysql/mysql-server@d2d6b2d
that seems to be applicable to MariaDB Server 10.3 and 10.4.
Even though commit 5b9ee8d
seems to have fixed sporadic failures on our CI systems, it is
theoretically possible that another race condition remained.

buf_flush_page_cleaner_coordinator(): In the final loop,
wait also for buf_get_n_pending_read_ios() to reach 0.
In this way, if a secondary index leaf page was read into the
buffer pool and ibuf_merge_or_delete_for_page() modified that
page or some change buffer pages, the flush loop would execute
until the buffer pool really is in a clean state.

This potential data corruption bug does not affect MariaDB Server 10.5
or later, thanks to commit b42294b
which removed change buffer merges that are not explicitly requested.

Author: dr-m

storage/innobase/buf/buf0flu.cc

@@ -3324,20 +3324,27 @@ DECLARE_THREAD(buf_flush_page_cleaner_coordinator)(void*)
 	bool	success;
 
 	do {
+		/* In case an asynchronous read request was posted by
+		any thread (other than something invoking
+		ibuf_merge_in_background()), it is possible that the
+		change buffer will be merged to the page once the read
+		completes. To avoid race conditions and corruption due
+		to that, we will loop here until there are no pending
+		page read operations. */
+		success = !buf_get_n_pending_read_ios();
 		pc_request(ULINT_MAX, LSN_MAX);
 
 		while (pc_flush_slot() > 0) {}
 
 		ulint	n_flushed_lru = 0;
 		ulint	n_flushed_list = 0;
-		success = pc_wait_finished(&n_flushed_lru, &n_flushed_list);
-
-		n_flushed = n_flushed_lru + n_flushed_list;
+		success = pc_wait_finished(&n_flushed_lru, &n_flushed_list)
+			&& success && !n_flushed_lru && !n_flushed_list;
 
 		buf_flush_wait_batch_end(NULL, BUF_FLUSH_LIST);
 		buf_flush_wait_LRU_batch_end();
 
-	} while (!success || n_flushed > 0);
+	} while (!success);
 
 	/* Some sanity checks */
 	ut_a(srv_get_active_thread_type() == SRV_NONE);