fix a writeback race

PATCH In review Andrew Morton 1 participant 1 message Jul 4, 2002

Subsystems (3)

Memory management VFS Page cache

lore

Share & tools

lore

Andrew Morton

author

2002-07-04 23:50 UTC #0 lore raw

PATCH 20/27 fix a writeback race

Fixes a bug in generic_writepages() and its cut-n-paste-cousin, mpage_writepages().

The code was clearing PageDirty and then baling out if it discovered the page was nder writeback. Which would cause the dirty bit to be lost.

It's a very small window, but reversing the order so PageDirty is only cleared when we know for-sure that IO will be started fixes it up.

 fs/mpage.c          |    4 ++--
 mm/page-writeback.c |    4 ++--
 2 files changed, 4 insertions(+), 4 deletions(-)

fs/mpage.c renamed from fs/mpage.c~writeback-dirty-fix +2 -2

@@ -515,8 +515,8 @@ mpage_writepages(struct address_space *m
 
 		lock_page(page);
 
-		if (page->mapping && TestClearPageDirty(page) &&
-					!PageWriteback(page)) {
+		if (page->mapping && !PageWriteback(page) &&
+					TestClearPageDirty(page)) {
 			/* FIXME: batch this up */
 			if (!PageActive(page) && PageLRU(page)) {
 				spin_lock(&pagemap_lru_lock);

mm/page-writeback.c renamed from mm/page-writeback.c~writeback-dirty-fix +2 -2

@@ -383,8 +383,8 @@ int generic_writepages(struct address_sp
 		lock_page(page);
 
 		/* It may have been removed from swapcache: check ->mapping */
-		if (page->mapping && TestClearPageDirty(page) &&
-					!PageWriteback(page)) {
+		if (page->mapping && !PageWriteback(page) &&
+					TestClearPageDirty(page)) {
 			/* FIXME: batch this up */
 			if (!PageActive(page) && PageLRU(page)) {
 				spin_lock(&pagemap_lru_lock);