[dm-devel] dm-snap: optimize track_chunk

[Mikulas Patocka]

On Tue, 16 Oct 2012, Mike Snitzer wrote:

> On Tue, Oct 16 2012 at  7:04pm -0400,
> Mikulas Patocka <mpatocka at redhat.com> wrote:
> 
> > dm-snap: optimize track_chunk
> > 
> > track_chunk is always called with interrupts enabled. Consequently, we
> > do not need to save and restore interrupt state in "flags" variable.
> > This patch changes spin_lock_irqsave to spin_lock_irq and
> > spin_unlock_irqrestore to spin_unlock_irq.
> > 
> > Signed-off-by: Mikulas Patocka <mpatocka at redhat.com>
> > 
> > ---
> >  drivers/md/dm-snap.c |    5 ++---
> >  1 file changed, 2 insertions(+), 3 deletions(-)
> > 
> > Index: linux-3.6.2-fast/drivers/md/dm-snap.c
> > ===================================================================
> > --- linux-3.6.2-fast.orig/drivers/md/dm-snap.c	2012-10-17 00:43:01.000000000 +0200
> > +++ linux-3.6.2-fast/drivers/md/dm-snap.c	2012-10-17 00:43:22.000000000 +0200
> > @@ -201,14 +201,13 @@ static struct dm_snap_tracked_chunk *tra
> >  {
> >  	struct dm_snap_tracked_chunk *c = mempool_alloc(s->tracked_chunk_pool,
> >  							GFP_NOIO);
> > -	unsigned long flags;
> >  
> >  	c->chunk = chunk;
> >  
> > -	spin_lock_irqsave(&s->tracked_chunk_lock, flags);
> > +	spin_lock_irq(&s->tracked_chunk_lock);
> 
> Does is make any sense to add "BUG_ON(irqs_disabled());" before the 
> spin_lock_irq?

You can add it, but I think there is not high risk that someone calls it 
from interrupt context in the future.

Mikulas