MDEV-33009 Server hangs for a long time with innodb_undo_log_truncate=ON

trx_purge_truncate_history(): Release buf_pool.flush_list_mutex
and 'dummily' acquire and release buf_pool.mutex before starting
a rescan of buf_pool.flush_list, to ensure that
the buf_flush_page_cleaner thread (which may be holding buf_pool.mutex)
will be able to proceed.

This fixes up commit 5dbe7a8c (MDEV-32757).

Tested by: Axel Schwenke (on Ubuntu 18.04 and Ubuntu 20.04)
Reviewed by: Vladislav Lesin

storage/innobase/trx/trx0purge.cc

@@ -687,8 +687,8 @@ TRANSACTIONAL_TARGET void trx_purge_truncate_history()
     mini-transaction commit and the server was killed, then
     discarding the to-be-trimmed pages without flushing would
     break crash recovery. */
-    mysql_mutex_lock(&buf_pool.flush_list_mutex);
   rescan:
+    mysql_mutex_lock(&buf_pool.flush_list_mutex);
     for (buf_page_t *bpage= UT_LIST_GET_LAST(buf_pool.flush_list); bpage; )
     {
       ut_ad(bpage->oldest_modification());
@@ -730,7 +730,17 @@ TRANSACTIONAL_TARGET void trx_purge_truncate_history()
         mysql_mutex_lock(&buf_pool.flush_list_mutex);
 
         if (prev != buf_pool.flush_hp.get())
+        {
+          /* The functions buf_pool_t::release_freed_page() or
+          buf_do_flush_list_batch() may be right now holding
+          buf_pool.mutex and waiting to acquire
+          buf_pool.flush_list_mutex. Ensure that they can proceed,
+          to avoid extreme waits. */
+          mysql_mutex_unlock(&buf_pool.flush_list_mutex);
+          mysql_mutex_lock(&buf_pool.mutex);
+          mysql_mutex_unlock(&buf_pool.mutex);
           goto rescan;
+        }
       }
 
       bpage= prev;