On Tue, 10 May 2011, Christoph Hellwig wrote:
> On Sun, May 08, 2011 at 08:44:19PM +0200, Mikulas Patocka wrote:
> > Index: linux-2.6.39-rc5-fast/fs/hpfs/file.c
> > ===================================================================
> > --- linux-2.6.39-rc5-fast.orig/fs/hpfs/file.c 2011-05-05 01:02:43.000000000 +0200
> > +++ linux-2.6.39-rc5-fast/fs/hpfs/file.c 2011-05-05 01:02:56.000000000 +0200
> > @@ -20,8 +20,8 @@ static int hpfs_file_release(struct inod
> >
> > int hpfs_file_fsync(struct file *file, int datasync)
> > {
> > - /*return file_fsync(file, datasync);*/
> > - return 0; /* Don't fsync :-) */
> > + struct inode *inode = file->f_mapping->host;
> > + return sync_blockdev(inode->i_sb->s_bdev);
>
> So you push all dirty data into buffers immediately and don't have
> anything else to write back?
Data are written in vfs_fsync_range ... so fsync should write just
metadata. I write all metadata, HPFS is not performance-sensitive anyway,
so there is no need to differentiate metadata for a specific file.
Mikulas
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
