[PATCH 3/8] xfs: bulk AIL insertion during transaction commit

Dave Chinner david at fromorbit.com
Wed Dec 1 19:32:13 CST 2010


On Tue, Nov 30, 2010 at 05:40:57PM -0500, Christoph Hellwig wrote:
> Generally looks good, a few minor comments below:
> 
> > +void
> > +xfs_trans_committed_bulk(
> > +	struct xfs_ail		*ailp,
> > +	struct xfs_log_vec	*log_vector,
> > +	xfs_lsn_t		commit_lsn,
> > +	int			aborted)
> > +{
> > +#define LGIA_SIZE	32
> > +	struct xfs_log_item	*lgia[LGIA_SIZE];
> 
> The lgia name seems a bit to obscure to me.

lgia == "log item array"

> What about just log_items
> as variable name, and LOG_ITEM_BATCH_SIZE or similar for the size?

Ok, that's easy to do.

> > +			spin_lock(&ailp->xa_lock);
> > +			xfs_trans_ail_update_bulk(ailp, lgia, LGIA_SIZE,
> > +							commit_lsn);
> > +			for (i = 0; i < LGIA_SIZE; i++)
> > +				IOP_UNPIN(lgia[i], 0);
> 
> > +		spin_lock(&ailp->xa_lock);
> > +		xfs_trans_ail_update_bulk(ailp, lgia, i, commit_lsn);
> > +		for (j = 0; j < i; j++)
> > +			IOP_UNPIN(lgia[j], 0);
> 
> It might be worth to factor these two little sniplets into a little
> helper.

I thought about it, but didn't because there was more code in
writing a helper. I'll do it anyway...

> 
> > + * Bulk update version of xfs_trans_ail_update.
> 
> That won't be a very useful comment anymore once xfs_trans_ail_update is
> implemented as a wrapper around this function..

Yup, but when the wrapper is introduced the comment is updated
appropriately.

> 
> > +	struct xfs_log_item	**lgia,
> 
> Same naming comment here.

Will fix.

Cheers,

Dave.
-- 
Dave Chinner
david at fromorbit.com




More information about the xfs mailing list