Re: [PATCH 09/12] xfs_repair: rebuild reverse mapping btrees with bulk loader

["Darrick J. Wong" <darrick.wong@oracle.com>]

On Thu, Jun 18, 2020 at 11:25:11AM -0400, Brian Foster wrote:
> On Mon, Jun 01, 2020 at 09:27:51PM -0700, Darrick J. Wong wrote:
> > From: Darrick J. Wong <darrick.wong@oracle.com>
> > 
> > Use the btree bulk loading functions to rebuild the reverse mapping
> > btrees and drop the open-coded implementation.
> > 
> > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> > ---
> >  libxfs/libxfs_api_defs.h |    1 
> >  repair/agbtree.c         |   70 ++++++++
> >  repair/agbtree.h         |    5 +
> >  repair/phase5.c          |  409 ++--------------------------------------------
> >  4 files changed, 96 insertions(+), 389 deletions(-)
> > 
> > 
> ...
> > diff --git a/repair/phase5.c b/repair/phase5.c
> > index e570349d..1c6448f4 100644
> > --- a/repair/phase5.c
> > +++ b/repair/phase5.c
> ...
> > @@ -1244,6 +879,8 @@ build_agf_agfl(
> >  	freelist = xfs_buf_to_agfl_bno(agfl_buf);
> >  	fill_agfl(btr_bno, freelist, &agfl_idx);
> >  	fill_agfl(btr_cnt, freelist, &agfl_idx);
> > +	if (xfs_sb_version_hasrmapbt(&mp->m_sb))
> > +		fill_agfl(btr_rmap, freelist, &agfl_idx);
> 
> Is this new behavior? Either way, I guess it makes sense since the
> rmapbt feeds from/to the agfl:

It's a defensive move to make sure we don't lose the blocks if we
overestimate the size of the rmapbt.  We never did in the past (and we
shouldn't now) but I figured I should throw that in as a defensive
measure so we don't leak the blocks if something goes wrong.

(Granted, I think in the past any overages would have been freed back
into the filesystem...)

Thanks for the review.

--D

> Reviewed-by: Brian Foster <bfoster@redhat.com>
> 
> >  
> >  	/* Set the AGF counters for the AGFL. */
> >  	if (agfl_idx > 0) {
> > @@ -1343,7 +980,7 @@ phase5_func(
> >  	struct bt_rebuild	btr_cnt;
> >  	struct bt_rebuild	btr_ino;
> >  	struct bt_rebuild	btr_fino;
> > -	bt_status_t		rmap_btree_curs;
> > +	struct bt_rebuild	btr_rmap;
> >  	bt_status_t		refcnt_btree_curs;
> >  	int			extra_blocks = 0;
> >  	uint			num_freeblocks;
> > @@ -1378,11 +1015,7 @@ _("unable to rebuild AG %u.  Not enough free space in on-disk AG.\n"),
> >  	init_ino_cursors(&sc, agno, num_freeblocks, &sb_icount_ag[agno],
> >  			&sb_ifree_ag[agno], &btr_ino, &btr_fino);
> >  
> > -	/*
> > -	 * Set up the btree cursors for the on-disk rmap btrees, which includes
> > -	 * pre-allocating all required blocks.
> > -	 */
> > -	init_rmapbt_cursor(mp, agno, &rmap_btree_curs);
> > +	init_rmapbt_cursor(&sc, agno, num_freeblocks, &btr_rmap);
> >  
> >  	/*
> >  	 * Set up the btree cursors for the on-disk refcount btrees,
> > @@ -1448,10 +1081,8 @@ _("unable to rebuild AG %u.  Not enough free space in on-disk AG.\n"),
> >  	ASSERT(btr_bno.freeblks == btr_cnt.freeblks);
> >  
> >  	if (xfs_sb_version_hasrmapbt(&mp->m_sb)) {
> > -		build_rmap_tree(mp, agno, &rmap_btree_curs);
> > -		write_cursor(&rmap_btree_curs);
> > -		sb_fdblocks_ag[agno] += (rmap_btree_curs.num_tot_blocks -
> > -				rmap_btree_curs.num_free_blocks) - 1;
> > +		build_rmap_tree(&sc, agno, &btr_rmap);
> > +		sb_fdblocks_ag[agno] += btr_rmap.newbt.afake.af_blocks - 1;
> >  	}
> >  
> >  	if (xfs_sb_version_hasreflink(&mp->m_sb)) {
> > @@ -1462,7 +1093,7 @@ _("unable to rebuild AG %u.  Not enough free space in on-disk AG.\n"),
> >  	/*
> >  	 * set up agf and agfl
> >  	 */
> > -	build_agf_agfl(mp, agno, &btr_bno, &btr_cnt, &rmap_btree_curs,
> > +	build_agf_agfl(mp, agno, &btr_bno, &btr_cnt, &btr_rmap,
> >  			&refcnt_btree_curs, lost_fsb);
> >  
> >  	build_inode_btrees(&sc, agno, &btr_ino, &btr_fino);
> > @@ -1479,7 +1110,7 @@ _("unable to rebuild AG %u.  Not enough free space in on-disk AG.\n"),
> >  	if (xfs_sb_version_hasfinobt(&mp->m_sb))
> >  		finish_rebuild(mp, &btr_fino, lost_fsb);
> >  	if (xfs_sb_version_hasrmapbt(&mp->m_sb))
> > -		finish_cursor(&rmap_btree_curs);
> > +		finish_rebuild(mp, &btr_rmap, lost_fsb);
> >  	if (xfs_sb_version_hasreflink(&mp->m_sb))
> >  		finish_cursor(&refcnt_btree_curs);
> >  
> > 
>