On Mon, Dec 09, 2013 at 10:16:12AM -0800, Christoph Hellwig wrote:
> On Mon, Dec 09, 2013 at 09:36:10AM +1100, Dave Chinner wrote:
> > > + /*
> > > + * Even after flushing the inode, there can still be
> > > + * delalloc blocks on the inode beyond EOF due to
> > > + * speculative reallocation. These are not removed
> > "speculative preallocation"
> I just re-indented the comment, the wording is the original one. Maybe
> we'll get a commit that cares enought to fix it on the fly..
Actually, I checked that before commenting on it - the original
comment is correct:
627 * even after flushing the inode, there can still be
628 * blocks on the inode beyond EOF due to speculative
629 * preallocation. These are not removed until the release
630 * function is called or the inode is inactivated. Hence we
631 * cannot assert here that ip->i_delayed_blks == 0.
I wouldn't have pointed at it if the original code had that