[PATCH 6/6] xfs: push the AIL from memory reclaim and periodic sync

[Dave Chinner]

On Thu, Mar 10, 2011 at 04:31:44PM -0500, Christoph Hellwig wrote:
> > -STATIC void xfs_ail_splice(struct xfs_ail *, struct list_head *, xfs_lsn_t);
> > -STATIC void xfs_ail_delete(struct xfs_ail *, xfs_log_item_t *);
> > -STATIC xfs_log_item_t * xfs_ail_min(struct xfs_ail *);
> > -STATIC xfs_log_item_t * xfs_ail_next(struct xfs_ail *, xfs_log_item_t *);
> > -
> 
> Reordering and cleanup of unrelated existing functions should be in a
> separate patch.

OK.

> 
> > @@ -55,16 +93,32 @@ xfs_lsn_t
> >  xfs_trans_ail_tail(
> >  	struct xfs_ail	*ailp)
> >  {
> > -	xfs_lsn_t	lsn;
> > +	xfs_lsn_t	lsn = 0;
> >  	xfs_log_item_t	*lip;
> >  
> >  	spin_lock(&ailp->xa_lock);
> >  	lip = xfs_ail_min(ailp);
> > -	if (lip == NULL) {
> > -		lsn = (xfs_lsn_t)0;
> > -	} else {
> > +	if (lip)
> > +		lsn = lip->li_lsn;
> > +	spin_unlock(&ailp->xa_lock);
> > +
> > +	return lsn;
> > +}
> > +
> > +/*
> > + * Return the maximum lsn held in the AIL, or zero if the AIl is empty.
> > + */
> > +static xfs_lsn_t
> > +xfs_ail_max_lsn(
> > +	struct xfs_ail  *ailp)
> > +{
> > +	xfs_lsn_t       lsn = 0;
> > +	xfs_log_item_t  *lip;
> > +
> > +	spin_lock(&ailp->xa_lock);
> > +	lip = xfs_ail_max(ailp);
> > +	if (lip)
> 
> As this is the counterpart to xfs_trans_ail_tail the naming for both
> should be similar.  I much prefer the descriptive _lsn naming over the
> random trans in xfs_trans_ail_tail.

Ok, will change.

Cheers,

Dave.
-- 
Dave Chinner
david at fromorbit.com