MDEV-32588 InnoDB may hang when running out of buffer pool

buf_flush_LRU_list_batch(): Do not skip pages that are actually clean
but in buf_pool.flush_list due to the "lazy removal" optimization of
commit 22b62eda, but try to evict them.
After acquiring buf_pool.flush_list_mutex, reread oldest_modification
to ensure that the block still remains in buf_pool.flush_list.

In addition to server hangs, this bug could also cause
InnoDB: Failing assertion: list.count > 0
in invocations of UT_LIST_REMOVE(flush_list, ...).

This fixes a regression that was caused by
commit a55b951e
and possibly made more likely to hit due to
commit aa719b50.

storage/innobase/buf/buf0flu.cc

@@ -1304,16 +1304,14 @@ static void buf_flush_LRU_list_batch(ulint max, bool evict,
     ut_ad(state >= buf_page_t::FREED);
     ut_ad(bpage->in_LRU_list);
 
-    switch (bpage->oldest_modification()) {
-    case 0:
+    if (!bpage->oldest_modification())
+    {
     evict:
       if (state != buf_page_t::FREED &&
           (state >= buf_page_t::READ_FIX || (~buf_page_t::LRU_MASK & state)))
         continue;
       buf_LRU_free_page(bpage, true);
       ++n->evicted;
-      /* fall through */
-    case 1:
       if (UNIV_LIKELY(scanned & 31))
         continue;
       mysql_mutex_unlock(&buf_pool.mutex);
@@ -1329,7 +1327,11 @@ static void buf_flush_LRU_list_batch(ulint max, bool evict,
       switch (bpage->oldest_modification()) {
       case 1:
         mysql_mutex_lock(&buf_pool.flush_list_mutex);
-        buf_pool.delete_from_flush_list(bpage);
+        if (ut_d(lsn_t lsn=) bpage->oldest_modification())
+        {
+          ut_ad(lsn == 1); /* It must be clean while we hold bpage->lock */
+          buf_pool.delete_from_flush_list(bpage);
+        }
         mysql_mutex_unlock(&buf_pool.flush_list_mutex);
         /* fall through */
       case 0: