[dm-devel] dm: don't save and restore bi_private

[Mikulas Patocka]

On Wed, 25 Nov 2015, Mike Snitzer wrote:

> On Wed, Nov 25 2015 at  4:32pm -0500,
> Mike Snitzer <snitzer at redhat.com> wrote:
> 
> > On Wed, Nov 25 2015 at  4:03pm -0500,
> > Mikulas Patocka <mpatocka at redhat.com> wrote:
> > 
> > > Device mapper used the field bi_private to point to dm_target_io. However,
> > > since kernel 3.15, the bi_private field is unused, and so the targets do
> > > not need to save and restore this field.
> > > 
> > > This patch removes code that saves and restores bi_private from dm-cache,
> > > dm-snapshot and dm-verity.
> > > 
> > > Signed-off-by: Mikulas Patocka <mpatocka at redhat.com>
> > > 
> > > ---
> > >  drivers/md/dm-cache-target.c |    3 ---
> > >  drivers/md/dm-snap.c         |    6 +-----
> > >  drivers/md/dm-verity.c       |    3 ---
> > >  3 files changed, 1 insertion(+), 11 deletions(-)
> > > 
> > > Index: linux-4.4-rc2/drivers/md/dm-cache-target.c
> > > ===================================================================
> > > --- linux-4.4-rc2.orig/drivers/md/dm-cache-target.c	2015-11-24 15:33:56.000000000 +0100
> > > +++ linux-4.4-rc2/drivers/md/dm-cache-target.c	2015-11-24 15:34:35.000000000 +0100
> > > @@ -118,14 +118,12 @@ static void iot_io_end(struct io_tracker
> > >   */
> > >  struct dm_hook_info {
> > >  	bio_end_io_t *bi_end_io;
> > > -	void *bi_private;
> > >  };
> > >  
> > >  static void dm_hook_bio(struct dm_hook_info *h, struct bio *bio,
> > >  			bio_end_io_t *bi_end_io, void *bi_private)
> > >  {
> > >  	h->bi_end_io = bio->bi_end_io;
> > > -	h->bi_private = bio->bi_private;
> > >  
> > >  	bio->bi_end_io = bi_end_io;
> > >  	bio->bi_private = bi_private;
> > 
> > As you can see dm_hook_bio() goes on to modify bi_private.
> > dm-cache uses it to get the migration object associated with a bio for
> > the overwrite_endio() case.
> > 
> > So NAK on the dm-cache change.
> > 
> > How closely have you reviewed your change for snapshot and verity?
> 
> Hmm, they follow the same pattern (hooking bi_private for use in bio
> endio).  So you're saying we no longer need to preserve any upper layer
> (e.g. DM core) use of bi_private?

Yes - because dm core doesn't use it.

> I'm really not seeing the harm in
> doing so... but I also don't have a problem with reinstating such
> backup/restore code if/when the need arises.
> 
> So it seems I've talked myself into your patch ;)

Mikulas