At 05:06 08/08/13, Mingming Cao wrote:

quoted

With that said, I don't have strong feelings against falling back to
buffered IO when the invalidate fails.  

It seems a little odd that we have to back to buffered IO in this case.
The pages are all flushed,  DIO just want to make sure the
journaltransactions who still keep those buffers are removed from their
list. It did that, the only reason to keep DIO fail is someone else
hasn't release the bh.

Current code enforce all the buffers have to be freed and pages are
removed from page cache, in order to force later read are from disk.  I
am not sure why can't we just leave the page in the cache, just clear it
uptodate flag, without reduce the page ref count?   I think DIO should
proceed it's IO in this case...

Like this?
Following patch just clears PG_uptodate when try_to_release_page fails.

diff -Nrup linux-2.6.27-rc3.org/mm/truncate.c linux-2.6.27-rc3/mm/truncate.c

--- linux-2.6.27-rc3.org/mm/truncate.c	2008-08-13 13:48:48.000000000 +0900
+++ linux-2.6.27-rc3/mm/truncate.c	2008-08-13 14:24:47.000000000 +0900

@@ -345,8 +345,10 @@ invalidate_complete_page2(struct address
 	if (page->mapping != mapping)
 		return 0;
 
-	if (PagePrivate(page) && !try_to_release_page(page, GFP_KERNEL))
-		return 0;
+	if (PagePrivate(page) && !try_to_release_page(page, GFP_KERNEL)) {
+		ClearPageUptodate(page);
+		return 1;
+	}
 
 	spin_lock_irq(&mapping->tree_lock);
 	if (PageDirty(page))

But this patch may cause some issue.
See, 
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=84209e02de48d72289650cc5a7ae8dd18223620f;hp=2b12a4c524812fb3f6ee590a02e65b95c8c32229

ClearPageUptodate in truncate.c was removed at 2.6.27-rc2 because this triggers some 
problems around NFS,FUSE,madvice.