diff --git a/mm/page-writeback.c b/mm/page-writeback.c
index b3a198c9248d..1d2fc89ca56d 100644
--- a/mm/page-writeback.c
+++ b/mm/page-writeback.c
@@ -862,17 +862,46 @@ int clear_page_dirty_for_io(struct page *page)
 {
 	struct address_space *mapping = page_mapping(page);
 
-	if (!mapping)
-		return TestClearPageDirty(page);
-
-	if (TestClearPageDirty(page)) {
-		if (mapping_cap_account_dirty(mapping)) {
-			page_mkclean(page);
+	if (mapping && mapping_cap_account_dirty(mapping)) {
+		/*
+		 * Yes, Virginia, this is indeed insane.
+		 *
+		 * We use this sequence to make sure that
+		 *  (a) we account for dirty stats properly
+		 *  (b) we tell the low-level filesystem to
+		 *      mark the whole page dirty if it was
+		 *      dirty in a pagetable. Only to then
+		 *  (c) clean the page again and return 1 to
+		 *      cause the writeback.
+		 *
+		 * This way we avoid all nasty races with the
+		 * dirty bit in multiple places and clearing
+		 * them concurrently from different threads.
+		 *
+		 * Note! Normally the "set_page_dirty(page)"
+		 * has no effect on the actual dirty bit - since
+		 * that will already usually be set. But we
+		 * need the side effects, and it can help us
+		 * avoid races.
+		 *
+		 * We basically use the page "master dirty bit"
+		 * as a serialization point for all the different
+		 * threads doing their things.
+		 *
+		 * FIXME! We still have a race here: if somebody
+		 * adds the page back to the page tables in
+		 * between the "page_mkclean()" and the "TestClearPageDirty()",
+		 * we might have it mapped without the dirty bit set.
+		 */
+		if (page_mkclean(page))
+			set_page_dirty(page);
+		if (TestClearPageDirty(page)) {
 			dec_zone_page_state(page, NR_FILE_DIRTY);
+			return 1;
 		}
-		return 1;
+		return 0;
 	}
-	return 0;
+	return TestClearPageDirty(page);
 }
 EXPORT_SYMBOL(clear_page_dirty_for_io);