Re: Review: ensure EOF writes into existing extents update filesize

[Nathan Scott <nscott@aconex.com>]

[-- Attachment #1: Type: text/plain, Size: 700 bytes --]

On Tue, 2007-05-22 at 11:03 +1000, David Chinner wrote:
> On Tue, May 22, 2007 at 11:02:59AM +1000, Nathan Scott wrote:
> > On Tue, 2007-05-22 at 10:44 +1000, David Chinner wrote:
> > > On Mon, May 21, 2007 at 04:34:26PM +1000, Timothy Shimmin wrote:
> > > I guess I need to ping Christoph and Nathan on this one....
> > 
> > Could you resend the patch to me please?  I lost the previous copy
> > while ruthlessly ploughing through my mail backlog.  ;)
> 
> Below.

Looks pretty good to me - xfs_convert_page has been overlooked, I
think - attached patch fixes that.  You also initialise iomap_valid
twice inside xfs_page_state_convert now ... I reverted that to just
the once.

cheers.

-- 
Nathan


[-- Attachment #2: nat.patch --]
[-- Type: text/x-patch, Size: 665 bytes --]

Index: linux/fs/xfs/linux-2.6/xfs_aops.c
===================================================================
--- linux.orig/fs/xfs/linux-2.6/xfs_aops.c	2007-05-22 13:47:40.296829500 +1000
+++ linux/fs/xfs/linux-2.6/xfs_aops.c	2007-05-22 13:48:25.987685000 +1000
@@ -810,7 +810,7 @@ xfs_convert_page(
 			page_dirty--;
 			count++;
 		} else {
-			type = 0;
+			type = IOMAP_NEW;
 			if (buffer_mapped(bh) && all_bh && startio) {
 				lock_buffer(bh);
 				xfs_add_to_ioend(inode, bh, offset,
@@ -968,7 +968,6 @@ xfs_page_state_convert(
 
 	bh = head = page_buffers(page);
 	offset = page_offset(page);
-	iomap_valid = 0;
 	flags = BMAPI_READ;
 	type = IOMAP_NEW;