On Sun, Aug 16, 2009 at 03:39:29PM -0400, Christoph Hellwig wrote:
> On Sun, Aug 16, 2009 at 08:25:36PM +1000, npiggin@xxxxxxx wrote:
> > }
> >
> > +int simple_setsize(struct inode *inode, loff_t newsize)
> > +{
>
> This wants a kerneldoc comment. Which should also mention that we can't
> easily use this helper if we actuall do need to truncate blocks on disk.
Yep, good idea.
> > ===================================================================
> > --- linux-2.6.orig/fs/buffer.c
> > +++ linux-2.6/fs/buffer.c
> > @@ -1992,9 +1992,14 @@ int block_write_begin(struct file *file,
> > * prepare_write() may have instantiated a few blocks
> > * outside i_size. Trim these off again. Don't need
> > * i_size_read because we hold i_mutex.
> > + *
> > + * Filesystems which set i_op->new_truncate must
> > + * handle this themselves. Eventually this will go
> > + * away because everyone will be converted.
> > */
> > if (pos + len > inode->i_size)
> > - vmtruncate(inode, inode->i_size);
> > + if (!inode->i_op->new_truncate)
> > + vmtruncate(inode, inode->i_size);
>
> Btw, what is you guestimate on how fast we'll get the conversion done?
>
> We have 19 instances of ->truncate currently. Your patch currently
> takes care of two, and I have an XFS patch leaving another 16.
I guess you can do really dumb ones using simple_setsize and then
call their old ->truncate method, basically just moving the vmtruncate
compat code into the filesystems. Then you just need to do a bit of
work to trim blocks after a failed write(2).
I think I will try going through and doing a round of those, and then
send them off to fs maintainers to do anything fancier. Might take
a day or so, which I should be able to get time to do.
> Can the maintainers on this list look into their filesystems to move
> forward?
Yes this would be nice. Thanks.
--
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
