1 files changed, 589 insertions, 74 deletions

diff --git a/fs/xfs/libxfs/xfs_ag.c b/fs/xfs/libxfs/xfs_ag.c
index 999ad8d00d43..e6ba914f6d06 100644
--- a/fs/xfs/libxfs/xfs_ag.c
+++ b/fs/xfs/libxfs/xfs_ag.c
@@ -10,6 +10,7 @@
 #include "xfs_shared.h"
 #include "xfs_format.h"
 #include "xfs_trans_resv.h"
+#include "xfs_bit.h"
 #include "xfs_sb.h"
 #include "xfs_mount.h"
 #include "xfs_btree.h"
@@ -19,27 +20,296 @@
 #include "xfs_ialloc.h"
 #include "xfs_rmap.h"
 #include "xfs_ag.h"
+#include "xfs_ag_resv.h"
+#include "xfs_health.h"
+#include "xfs_error.h"
+#include "xfs_bmap.h"
+#include "xfs_defer.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_trace.h"
+#include "xfs_inode.h"
+#include "xfs_icache.h"
+#include "xfs_group.h"
 
-static struct xfs_buf *
+/*
+ * xfs_initialize_perag_data
+ *
+ * Read in each per-ag structure so we can count up the number of
+ * allocated inodes, free inodes and used filesystem blocks as this
+ * information is no longer persistent in the superblock. Once we have
+ * this information, write it into the in-core superblock structure.
+ */
+int
+xfs_initialize_perag_data(
+	struct xfs_mount	*mp,
+	xfs_agnumber_t		agcount)
+{
+	xfs_agnumber_t		index;
+	struct xfs_perag	*pag;
+	struct xfs_sb		*sbp = &mp->m_sb;
+	uint64_t		ifree = 0;
+	uint64_t		ialloc = 0;
+	uint64_t		bfree = 0;
+	uint64_t		bfreelst = 0;
+	uint64_t		btree = 0;
+	uint64_t		fdblocks;
+	int			error = 0;
+
+	for (index = 0; index < agcount; index++) {
+		/*
+		 * Read the AGF and AGI buffers to populate the per-ag
+		 * structures for us.
+		 */
+		pag = xfs_perag_get(mp, index);
+		error = xfs_alloc_read_agf(pag, NULL, 0, NULL);
+		if (!error)
+			error = xfs_ialloc_read_agi(pag, NULL, 0, NULL);
+		if (error) {
+			xfs_perag_put(pag);
+			return error;
+		}
+
+		ifree += pag->pagi_freecount;
+		ialloc += pag->pagi_count;
+		bfree += pag->pagf_freeblks;
+		bfreelst += pag->pagf_flcount;
+		btree += pag->pagf_btreeblks;
+		xfs_perag_put(pag);
+	}
+	fdblocks = bfree + bfreelst + btree;
+
+	/*
+	 * If the new summary counts are obviously incorrect, fail the
+	 * mount operation because that implies the AGFs are also corrupt.
+	 * Clear FS_COUNTERS so that we don't unmount with a dirty log, which
+	 * will prevent xfs_repair from fixing anything.
+	 */
+	if (fdblocks > sbp->sb_dblocks || ifree > ialloc) {
+		xfs_alert(mp, "AGF corruption. Please run xfs_repair.");
+		xfs_fs_mark_sick(mp, XFS_SICK_FS_COUNTERS);
+		error = -EFSCORRUPTED;
+		goto out;
+	}
+
+	/* Overwrite incore superblock counters with just-read data */
+	spin_lock(&mp->m_sb_lock);
+	sbp->sb_ifree = ifree;
+	sbp->sb_icount = ialloc;
+	sbp->sb_fdblocks = fdblocks;
+	spin_unlock(&mp->m_sb_lock);
+
+	xfs_reinit_percpu_counters(mp);
+out:
+	xfs_fs_mark_healthy(mp, XFS_SICK_FS_COUNTERS);
+	return error;
+}
+
+static void
+xfs_perag_uninit(
+	struct xfs_group	*xg)
+{
+#ifdef __KERNEL__
+	struct xfs_perag	*pag = to_perag(xg);
+
+	cancel_delayed_work_sync(&pag->pag_blockgc_work);
+	xfs_buf_cache_destroy(&pag->pag_bcache);
+#endif
+}
+
+/*
+ * Free up the per-ag resources  within the specified AG range.
+ */
+void
+xfs_free_perag_range(
+	struct xfs_mount	*mp,
+	xfs_agnumber_t		first_agno,
+	xfs_agnumber_t		end_agno)
+
+{
+	xfs_agnumber_t		agno;
+
+	for (agno = first_agno; agno < end_agno; agno++)
+		xfs_group_free(mp, agno, XG_TYPE_AG, xfs_perag_uninit);
+}
+
+/* Find the size of the AG, in blocks. */
+static xfs_agblock_t
+__xfs_ag_block_count(
+	struct xfs_mount	*mp,
+	xfs_agnumber_t		agno,
+	xfs_agnumber_t		agcount,
+	xfs_rfsblock_t		dblocks)
+{
+	ASSERT(agno < agcount);
+
+	if (agno < agcount - 1)
+		return mp->m_sb.sb_agblocks;
+	return dblocks - (agno * mp->m_sb.sb_agblocks);
+}
+
+xfs_agblock_t
+xfs_ag_block_count(
+	struct xfs_mount	*mp,
+	xfs_agnumber_t		agno)
+{
+	return __xfs_ag_block_count(mp, agno, mp->m_sb.sb_agcount,
+			mp->m_sb.sb_dblocks);
+}
+
+/* Calculate the first and last possible inode number in an AG. */
+static void
+__xfs_agino_range(
+	struct xfs_mount	*mp,
+	xfs_agblock_t		eoag,
+	xfs_agino_t		*first,
+	xfs_agino_t		*last)
+{
+	xfs_agblock_t		bno;
+
+	/*
+	 * Calculate the first inode, which will be in the first
+	 * cluster-aligned block after the AGFL.
+	 */
+	bno = round_up(XFS_AGFL_BLOCK(mp) + 1, M_IGEO(mp)->cluster_align);
+	*first = XFS_AGB_TO_AGINO(mp, bno);
+
+	/*
+	 * Calculate the last inode, which will be at the end of the
+	 * last (aligned) cluster that can be allocated in the AG.
+	 */
+	bno = round_down(eoag, M_IGEO(mp)->cluster_align);
+	*last = XFS_AGB_TO_AGINO(mp, bno) - 1;
+}
+
+void
+xfs_agino_range(
+	struct xfs_mount	*mp,
+	xfs_agnumber_t		agno,
+	xfs_agino_t		*first,
+	xfs_agino_t		*last)
+{
+	return __xfs_agino_range(mp, xfs_ag_block_count(mp, agno), first, last);
+}
+
+/*
+ * Update the perag of the previous tail AG if it has been changed during
+ * recovery (i.e. recovery of a growfs).
+ */
+int
+xfs_update_last_ag_size(
+	struct xfs_mount	*mp,
+	xfs_agnumber_t		prev_agcount)
+{
+	struct xfs_perag	*pag = xfs_perag_grab(mp, prev_agcount - 1);
+
+	if (!pag)
+		return -EFSCORRUPTED;
+	pag_group(pag)->xg_block_count = __xfs_ag_block_count(mp,
+			prev_agcount - 1, mp->m_sb.sb_agcount,
+			mp->m_sb.sb_dblocks);
+	__xfs_agino_range(mp, pag_group(pag)->xg_block_count, &pag->agino_min,
+			&pag->agino_max);
+	xfs_perag_rele(pag);
+	return 0;
+}
+
+static int
+xfs_perag_alloc(
+	struct xfs_mount	*mp,
+	xfs_agnumber_t		index,
+	xfs_agnumber_t		agcount,
+	xfs_rfsblock_t		dblocks)
+{
+	struct xfs_perag	*pag;
+	int			error;
+
+	pag = kzalloc(sizeof(*pag), GFP_KERNEL);
+	if (!pag)
+		return -ENOMEM;
+
+#ifdef __KERNEL__
+	/* Place kernel structure only init below this point. */
+	spin_lock_init(&pag->pag_ici_lock);
+	INIT_DELAYED_WORK(&pag->pag_blockgc_work, xfs_blockgc_worker);
+	INIT_RADIX_TREE(&pag->pag_ici_root, GFP_ATOMIC);
+#endif /* __KERNEL__ */
+
+	error = xfs_buf_cache_init(&pag->pag_bcache);
+	if (error)
+		goto out_free_perag;
+
+	/*
+	 * Pre-calculated geometry
+	 */
+	pag_group(pag)->xg_block_count = __xfs_ag_block_count(mp, index, agcount,
+				dblocks);
+	pag_group(pag)->xg_min_gbno = XFS_AGFL_BLOCK(mp) + 1;
+	__xfs_agino_range(mp, pag_group(pag)->xg_block_count, &pag->agino_min,
+			&pag->agino_max);
+
+	error = xfs_group_insert(mp, pag_group(pag), index, XG_TYPE_AG);
+	if (error)
+		goto out_buf_cache_destroy;
+
+	return 0;
+
+out_buf_cache_destroy:
+	xfs_buf_cache_destroy(&pag->pag_bcache);
+out_free_perag:
+	kfree(pag);
+	return error;
+}
+
+int
+xfs_initialize_perag(
+	struct xfs_mount	*mp,
+	xfs_agnumber_t		orig_agcount,
+	xfs_agnumber_t		new_agcount,
+	xfs_rfsblock_t		dblocks,
+	xfs_agnumber_t		*maxagi)
+{
+	xfs_agnumber_t		index;
+	int			error;
+
+	if (orig_agcount >= new_agcount)
+		return 0;
+
+	for (index = orig_agcount; index < new_agcount; index++) {
+		error = xfs_perag_alloc(mp, index, new_agcount, dblocks);
+		if (error)
+			goto out_unwind_new_pags;
+	}
+
+	*maxagi = xfs_set_inode_alloc(mp, new_agcount);
+	mp->m_ag_prealloc_blocks = xfs_prealloc_blocks(mp);
+	return 0;
+
+out_unwind_new_pags:
+	xfs_free_perag_range(mp, orig_agcount, index);
+	return error;
+}
+
+static int
 xfs_get_aghdr_buf(
 	struct xfs_mount	*mp,
 	xfs_daddr_t		blkno,
 	size_t			numblks,
-	int			flags,
+	struct xfs_buf		**bpp,
 	const struct xfs_buf_ops *ops)
 {
 	struct xfs_buf		*bp;
+	int			error;
 
-	bp = xfs_buf_get_uncached(mp->m_ddev_targp, numblks, flags);
-	if (!bp)
-		return NULL;
+	error = xfs_buf_get_uncached(mp->m_ddev_targp, numblks, &bp);
+	if (error)
+		return error;
 
-	xfs_buf_zero(bp, 0, BBTOB(bp->b_length));
-	bp->b_bn = blkno;
 	bp->b_maps[0].bm_bn = blkno;
 	bp->b_ops = ops;
 
-	return bp;
+	*bpp = bp;
+	return 0;
 }
 
 /*
@@ -51,40 +321,74 @@ xfs_btroot_init(
 	struct xfs_buf		*bp,
 	struct aghdr_init_data	*id)
 {
-	xfs_btree_init_block(mp, bp, id->type, 0, 0, id->agno, 0);
+	xfs_btree_init_buf(mp, bp, id->bc_ops, 0, 0, id->agno);
 }
 
-/*
- * Alloc btree root block init functions
- */
+/* Finish initializing a free space btree. */
 static void
-xfs_bnoroot_init(
+xfs_freesp_init_recs(
 	struct xfs_mount	*mp,
 	struct xfs_buf		*bp,
 	struct aghdr_init_data	*id)
 {
 	struct xfs_alloc_rec	*arec;
+	struct xfs_btree_block	*block = XFS_BUF_TO_BLOCK(bp);
 
-	xfs_btree_init_block(mp, bp, XFS_BTNUM_BNO, 0, 1, id->agno, 0);
 	arec = XFS_ALLOC_REC_ADDR(mp, XFS_BUF_TO_BLOCK(bp), 1);
 	arec->ar_startblock = cpu_to_be32(mp->m_ag_prealloc_blocks);
+
+	if (xfs_ag_contains_log(mp, id->agno)) {
+		struct xfs_alloc_rec	*nrec;
+		xfs_agblock_t		start = XFS_FSB_TO_AGBNO(mp,
+							mp->m_sb.sb_logstart);
+
+		ASSERT(start >= mp->m_ag_prealloc_blocks);
+		if (start != mp->m_ag_prealloc_blocks) {
+			/*
+			 * Modify first record to pad stripe align of log and
+			 * bump the record count.
+			 */
+			arec->ar_blockcount = cpu_to_be32(start -
+						mp->m_ag_prealloc_blocks);
+			be16_add_cpu(&block->bb_numrecs, 1);
+			nrec = arec + 1;
+
+			/*
+			 * Insert second record at start of internal log
+			 * which then gets trimmed.
+			 */
+			nrec->ar_startblock = cpu_to_be32(
+					be32_to_cpu(arec->ar_startblock) +
+					be32_to_cpu(arec->ar_blockcount));
+			arec = nrec;
+		}
+		/*
+		 * Change record start to after the internal log
+		 */
+		be32_add_cpu(&arec->ar_startblock, mp->m_sb.sb_logblocks);
+	}
+
+	/*
+	 * Calculate the block count of this record; if it is nonzero,
+	 * increment the record count.
+	 */
 	arec->ar_blockcount = cpu_to_be32(id->agsize -
 					  be32_to_cpu(arec->ar_startblock));
+	if (arec->ar_blockcount)
+		be16_add_cpu(&block->bb_numrecs, 1);
 }
 
+/*
+ * bnobt/cntbt btree root block init functions
+ */
 static void
-xfs_cntroot_init(
+xfs_bnoroot_init(
 	struct xfs_mount	*mp,
 	struct xfs_buf		*bp,
 	struct aghdr_init_data	*id)
 {
-	struct xfs_alloc_rec	*arec;
-
-	xfs_btree_init_block(mp, bp, XFS_BTNUM_CNT, 0, 1, id->agno, 0);
-	arec = XFS_ALLOC_REC_ADDR(mp, XFS_BUF_TO_BLOCK(bp), 1);
-	arec->ar_startblock = cpu_to_be32(mp->m_ag_prealloc_blocks);
-	arec->ar_blockcount = cpu_to_be32(id->agsize -
-					  be32_to_cpu(arec->ar_startblock));
+	xfs_btree_init_buf(mp, bp, id->bc_ops, 0, 0, id->agno);
+	xfs_freesp_init_recs(mp, bp, id);
 }
 
 /*
@@ -99,7 +403,7 @@ xfs_rmaproot_init(
 	struct xfs_btree_block	*block = XFS_BUF_TO_BLOCK(bp);
 	struct xfs_rmap_rec	*rrec;
 
-	xfs_btree_init_block(mp, bp, XFS_BTNUM_RMAP, 0, 4, id->agno, 0);
+	xfs_btree_init_buf(mp, bp, id->bc_ops, 0, 4, id->agno);
 
 	/*
 	 * mark the AG header regions as static metadata The BNO
@@ -139,7 +443,7 @@ xfs_rmaproot_init(
 	rrec->rm_offset = 0;
 
 	/* account for refc btree root */
-	if (xfs_sb_version_hasreflink(&mp->m_sb)) {
+	if (xfs_has_reflink(mp)) {
 		rrec = XFS_RMAP_REC_ADDR(block, 5);
 		rrec->rm_startblock = cpu_to_be32(xfs_refc_block(mp));
 		rrec->rm_blockcount = cpu_to_be32(1);
@@ -147,6 +451,18 @@ xfs_rmaproot_init(
 		rrec->rm_offset = 0;
 		be16_add_cpu(&block->bb_numrecs, 1);
 	}
+
+	/* account for the log space */
+	if (xfs_ag_contains_log(mp, id->agno)) {
+		rrec = XFS_RMAP_REC_ADDR(block,
+				be16_to_cpu(block->bb_numrecs) + 1);
+		rrec->rm_startblock = cpu_to_be32(
+				XFS_FSB_TO_AGBNO(mp, mp->m_sb.sb_logstart));
+		rrec->rm_blockcount = cpu_to_be32(mp->m_sb.sb_logblocks);
+		rrec->rm_owner = cpu_to_be64(XFS_RMAP_OWN_LOG);
+		rrec->rm_offset = 0;
+		be16_add_cpu(&block->bb_numrecs, 1);
+	}
 }
 
 /*
@@ -163,7 +479,7 @@ xfs_sbblock_init(
 	struct xfs_buf		*bp,
 	struct aghdr_init_data	*id)
 {
-	struct xfs_dsb		*dsb = XFS_BUF_TO_SBP(bp);
+	struct xfs_dsb		*dsb = bp->b_addr;
 
 	xfs_sb_to_disk(dsb, &mp->m_sb);
 	dsb->sb_inprogress = 1;
@@ -175,21 +491,20 @@ xfs_agfblock_init(
 	struct xfs_buf		*bp,
 	struct aghdr_init_data	*id)
 {
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(bp);
+	struct xfs_agf		*agf = bp->b_addr;
 	xfs_extlen_t		tmpsize;
 
 	agf->agf_magicnum = cpu_to_be32(XFS_AGF_MAGIC);
 	agf->agf_versionnum = cpu_to_be32(XFS_AGF_VERSION);
 	agf->agf_seqno = cpu_to_be32(id->agno);
 	agf->agf_length = cpu_to_be32(id->agsize);
-	agf->agf_roots[XFS_BTNUM_BNOi] = cpu_to_be32(XFS_BNO_BLOCK(mp));
-	agf->agf_roots[XFS_BTNUM_CNTi] = cpu_to_be32(XFS_CNT_BLOCK(mp));
-	agf->agf_levels[XFS_BTNUM_BNOi] = cpu_to_be32(1);
-	agf->agf_levels[XFS_BTNUM_CNTi] = cpu_to_be32(1);
-	if (xfs_sb_version_hasrmapbt(&mp->m_sb)) {
-		agf->agf_roots[XFS_BTNUM_RMAPi] =
-					cpu_to_be32(XFS_RMAP_BLOCK(mp));
-		agf->agf_levels[XFS_BTNUM_RMAPi] = cpu_to_be32(1);
+	agf->agf_bno_root = cpu_to_be32(XFS_BNO_BLOCK(mp));
+	agf->agf_cnt_root = cpu_to_be32(XFS_CNT_BLOCK(mp));
+	agf->agf_bno_level = cpu_to_be32(1);
+	agf->agf_cnt_level = cpu_to_be32(1);
+	if (xfs_has_rmapbt(mp)) {
+		agf->agf_rmap_root = cpu_to_be32(XFS_RMAP_BLOCK(mp));
+		agf->agf_rmap_level = cpu_to_be32(1);
 		agf->agf_rmap_blocks = cpu_to_be32(1);
 	}
 
@@ -199,14 +514,22 @@ xfs_agfblock_init(
 	tmpsize = id->agsize - mp->m_ag_prealloc_blocks;
 	agf->agf_freeblks = cpu_to_be32(tmpsize);
 	agf->agf_longest = cpu_to_be32(tmpsize);
-	if (xfs_sb_version_hascrc(&mp->m_sb))
+	if (xfs_has_crc(mp))
 		uuid_copy(&agf->agf_uuid, &mp->m_sb.sb_meta_uuid);
-	if (xfs_sb_version_hasreflink(&mp->m_sb)) {
+	if (xfs_has_reflink(mp)) {
 		agf->agf_refcount_root = cpu_to_be32(
 				xfs_refc_block(mp));
 		agf->agf_refcount_level = cpu_to_be32(1);
 		agf->agf_refcount_blocks = cpu_to_be32(1);
 	}
+
+	if (xfs_ag_contains_log(mp, id->agno)) {
+		int64_t	logblocks = mp->m_sb.sb_logblocks;
+
+		be32_add_cpu(&agf->agf_freeblks, -logblocks);
+		agf->agf_longest = cpu_to_be32(id->agsize -
+			XFS_FSB_TO_AGBNO(mp, mp->m_sb.sb_logstart) - logblocks);
+	}
 }
 
 static void
@@ -219,13 +542,13 @@ xfs_agflblock_init(
 	__be32			*agfl_bno;
 	int			bucket;
 
-	if (xfs_sb_version_hascrc(&mp->m_sb)) {
+	if (xfs_has_crc(mp)) {
 		agfl->agfl_magicnum = cpu_to_be32(XFS_AGFL_MAGIC);
 		agfl->agfl_seqno = cpu_to_be32(id->agno);
 		uuid_copy(&agfl->agfl_uuid, &mp->m_sb.sb_meta_uuid);
 	}
 
-	agfl_bno = XFS_BUF_TO_AGFL_BNO(mp, bp);
+	agfl_bno = xfs_buf_to_agfl_bno(bp);
 	for (bucket = 0; bucket < xfs_agfl_size(mp); bucket++)
 		agfl_bno[bucket] = cpu_to_be32(NULLAGBLOCK);
 }
@@ -236,7 +559,7 @@ xfs_agiblock_init(
 	struct xfs_buf		*bp,
 	struct aghdr_init_data	*id)
 {
-	struct xfs_agi		*agi = XFS_BUF_TO_AGI(bp);
+	struct xfs_agi		*agi = bp->b_addr;
 	int			bucket;
 
 	agi->agi_magicnum = cpu_to_be32(XFS_AGI_MAGIC);
@@ -249,14 +572,19 @@ xfs_agiblock_init(
 	agi->agi_freecount = 0;
 	agi->agi_newino = cpu_to_be32(NULLAGINO);
 	agi->agi_dirino = cpu_to_be32(NULLAGINO);
-	if (xfs_sb_version_hascrc(&mp->m_sb))
+	if (xfs_has_crc(mp))
 		uuid_copy(&agi->agi_uuid, &mp->m_sb.sb_meta_uuid);
-	if (xfs_sb_version_hasfinobt(&mp->m_sb)) {
+	if (xfs_has_finobt(mp)) {
 		agi->agi_free_root = cpu_to_be32(XFS_FIBT_BLOCK(mp));
 		agi->agi_free_level = cpu_to_be32(1);
 	}
 	for (bucket = 0; bucket < XFS_AGI_UNLINKED_BUCKETS; bucket++)
 		agi->agi_unlinked[bucket] = cpu_to_be32(NULLAGINO);
+	if (xfs_has_inobtcounts(mp)) {
+		agi->agi_iblocks = cpu_to_be32(1);
+		if (xfs_has_finobt(mp))
+			agi->agi_fblocks = cpu_to_be32(1);
+	}
 }
 
 typedef void (*aghdr_init_work_f)(struct xfs_mount *mp, struct xfs_buf *bp,
@@ -267,13 +595,13 @@ xfs_ag_init_hdr(
 	struct aghdr_init_data	*id,
 	aghdr_init_work_f	work,
 	const struct xfs_buf_ops *ops)
-
 {
 	struct xfs_buf		*bp;
+	int			error;
 
-	bp = xfs_get_aghdr_buf(mp, id->daddr, id->numblks, 0, ops);
-	if (!bp)
-		return -ENOMEM;
+	error = xfs_get_aghdr_buf(mp, id->daddr, id->numblks, &bp, ops);
+	if (error)
+		return error;
 
 	(*work)(mp, bp, id);
 
@@ -287,7 +615,7 @@ struct xfs_aghdr_grow_data {
 	size_t			numblks;
 	const struct xfs_buf_ops *ops;
 	aghdr_init_work_f	work;
-	xfs_btnum_t		type;
+	const struct xfs_btree_ops *bc_ops;
 	bool			need_init;
 };
 
@@ -339,15 +667,17 @@ xfs_ag_init_headers(
 	{ /* BNO root block */
 		.daddr = XFS_AGB_TO_DADDR(mp, id->agno, XFS_BNO_BLOCK(mp)),
 		.numblks = BTOBB(mp->m_sb.sb_blocksize),
-		.ops = &xfs_allocbt_buf_ops,
+		.ops = &xfs_bnobt_buf_ops,
 		.work = &xfs_bnoroot_init,
+		.bc_ops = &xfs_bnobt_ops,
 		.need_init = true
 	},
 	{ /* CNT root block */
 		.daddr = XFS_AGB_TO_DADDR(mp, id->agno, XFS_CNT_BLOCK(mp)),
 		.numblks = BTOBB(mp->m_sb.sb_blocksize),
-		.ops = &xfs_allocbt_buf_ops,
-		.work = &xfs_cntroot_init,
+		.ops = &xfs_cntbt_buf_ops,
+		.work = &xfs_bnoroot_init,
+		.bc_ops = &xfs_cntbt_ops,
 		.need_init = true
 	},
 	{ /* INO root block */
@@ -355,31 +685,32 @@ xfs_ag_init_headers(
 		.numblks = BTOBB(mp->m_sb.sb_blocksize),
 		.ops = &xfs_inobt_buf_ops,
 		.work = &xfs_btroot_init,
-		.type = XFS_BTNUM_INO,
+		.bc_ops = &xfs_inobt_ops,
 		.need_init = true
 	},
 	{ /* FINO root block */
 		.daddr = XFS_AGB_TO_DADDR(mp, id->agno, XFS_FIBT_BLOCK(mp)),
 		.numblks = BTOBB(mp->m_sb.sb_blocksize),
-		.ops = &xfs_inobt_buf_ops,
+		.ops = &xfs_finobt_buf_ops,
 		.work = &xfs_btroot_init,
-		.type = XFS_BTNUM_FINO,
-		.need_init =  xfs_sb_version_hasfinobt(&mp->m_sb)
+		.bc_ops = &xfs_finobt_ops,
+		.need_init =  xfs_has_finobt(mp)
 	},
 	{ /* RMAP root block */
 		.daddr = XFS_AGB_TO_DADDR(mp, id->agno, XFS_RMAP_BLOCK(mp)),
 		.numblks = BTOBB(mp->m_sb.sb_blocksize),
 		.ops = &xfs_rmapbt_buf_ops,
 		.work = &xfs_rmaproot_init,
-		.need_init = xfs_sb_version_hasrmapbt(&mp->m_sb)
+		.bc_ops = &xfs_rmapbt_ops,
+		.need_init = xfs_has_rmapbt(mp)
 	},
 	{ /* REFC root block */
 		.daddr = XFS_AGB_TO_DADDR(mp, id->agno, xfs_refc_block(mp)),
 		.numblks = BTOBB(mp->m_sb.sb_blocksize),
 		.ops = &xfs_refcountbt_buf_ops,
 		.work = &xfs_btroot_init,
-		.type = XFS_BTNUM_REFC,
-		.need_init = xfs_sb_version_hasreflink(&mp->m_sb)
+		.bc_ops = &xfs_refcountbt_ops,
+		.need_init = xfs_has_reflink(mp)
 	},
 	{ /* NULL terminating block */
 		.daddr = XFS_BUF_DADDR_NULL,
@@ -396,7 +727,7 @@ xfs_ag_init_headers(
 
 		id->daddr = dp->daddr;
 		id->numblks = dp->numblks;
-		id->type = dp->type;
+		id->bc_ops = dp->bc_ops;
 		error = xfs_ag_init_hdr(mp, id, dp->work, dp->ops);
 		if (error)
 			break;
@@ -404,42 +735,176 @@ xfs_ag_init_headers(
 	return error;
 }
 
+int
+xfs_ag_shrink_space(
+	struct xfs_perag	*pag,
+	struct xfs_trans	**tpp,
+	xfs_extlen_t		delta)
+{
+	struct xfs_mount	*mp = pag_mount(pag);
+	struct xfs_alloc_arg	args = {
+		.tp	= *tpp,
+		.mp	= mp,
+		.pag	= pag,
+		.minlen = delta,
+		.maxlen = delta,
+		.oinfo	= XFS_RMAP_OINFO_SKIP_UPDATE,
+		.resv	= XFS_AG_RESV_NONE,
+		.prod	= 1
+	};
+	struct xfs_buf		*agibp, *agfbp;
+	struct xfs_agi		*agi;
+	struct xfs_agf		*agf;
+	xfs_agblock_t		aglen;
+	int			error, err2;
+
+	ASSERT(pag_agno(pag) == mp->m_sb.sb_agcount - 1);
+	error = xfs_ialloc_read_agi(pag, *tpp, 0, &agibp);
+	if (error)
+		return error;
+
+	agi = agibp->b_addr;
+
+	error = xfs_alloc_read_agf(pag, *tpp, 0, &agfbp);
+	if (error)
+		return error;
+
+	agf = agfbp->b_addr;
+	aglen = be32_to_cpu(agi->agi_length);
+	/* some extra paranoid checks before we shrink the ag */
+	if (XFS_IS_CORRUPT(mp, agf->agf_length != agi->agi_length)) {
+		xfs_ag_mark_sick(pag, XFS_SICK_AG_AGF);
+		return -EFSCORRUPTED;
+	}
+	if (delta >= aglen)
+		return -EINVAL;
+
+	/*
+	 * Make sure that the last inode cluster cannot overlap with the new
+	 * end of the AG, even if it's sparse.
+	 */
+	error = xfs_ialloc_check_shrink(pag, *tpp, agibp, aglen - delta);
+	if (error)
+		return error;
+
+	/*
+	 * Disable perag reservations so it doesn't cause the allocation request
+	 * to fail. We'll reestablish reservation before we return.
+	 */
+	xfs_ag_resv_free(pag);
+
+	/* internal log shouldn't also show up in the free space btrees */
+	error = xfs_alloc_vextent_exact_bno(&args,
+			xfs_agbno_to_fsb(pag, aglen - delta));
+	if (!error && args.agbno == NULLAGBLOCK)
+		error = -ENOSPC;
+
+	if (error) {
+		/*
+		 * If extent allocation fails, need to roll the transaction to
+		 * ensure that the AGFL fixup has been committed anyway.
+		 *
+		 * We need to hold the AGF across the roll to ensure nothing can
+		 * access the AG for allocation until the shrink is fully
+		 * cleaned up. And due to the resetting of the AG block
+		 * reservation space needing to lock the AGI, we also have to
+		 * hold that so we don't get AGI/AGF lock order inversions in
+		 * the error handling path.
+		 */
+		xfs_trans_bhold(*tpp, agfbp);
+		xfs_trans_bhold(*tpp, agibp);
+		err2 = xfs_trans_roll(tpp);
+		if (err2)
+			return err2;
+		xfs_trans_bjoin(*tpp, agfbp);
+		xfs_trans_bjoin(*tpp, agibp);
+		goto resv_init_out;
+	}
+
+	/*
+	 * if successfully deleted from freespace btrees, need to confirm
+	 * per-AG reservation works as expected.
+	 */
+	be32_add_cpu(&agi->agi_length, -delta);
+	be32_add_cpu(&agf->agf_length, -delta);
+
+	err2 = xfs_ag_resv_init(pag, *tpp);
+	if (err2) {
+		be32_add_cpu(&agi->agi_length, delta);
+		be32_add_cpu(&agf->agf_length, delta);
+		if (err2 != -ENOSPC)
+			goto resv_err;
+
+		err2 = xfs_free_extent_later(*tpp, args.fsbno, delta, NULL,
+				XFS_AG_RESV_NONE, XFS_FREE_EXTENT_SKIP_DISCARD);
+		if (err2)
+			goto resv_err;
+
+		/*
+		 * Roll the transaction before trying to re-init the per-ag
+		 * reservation. The new transaction is clean so it will cancel
+		 * without any side effects.
+		 */
+		error = xfs_defer_finish(tpp);
+		if (error)
+			return error;
+
+		error = -ENOSPC;
+		goto resv_init_out;
+	}
+
+	/* Update perag geometry */
+	pag_group(pag)->xg_block_count -= delta;
+	__xfs_agino_range(mp, pag_group(pag)->xg_block_count, &pag->agino_min,
+			&pag->agino_max);
+
+	xfs_ialloc_log_agi(*tpp, agibp, XFS_AGI_LENGTH);
+	xfs_alloc_log_agf(*tpp, agfbp, XFS_AGF_LENGTH);
+	return 0;
+
+resv_init_out:
+	err2 = xfs_ag_resv_init(pag, *tpp);
+	if (!err2)
+		return error;
+resv_err:
+	xfs_warn(mp, "Error %d reserving per-AG metadata reserve pool.", err2);
+	xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
+	return err2;
+}
+
 /*
  * Extent the AG indicated by the @id by the length passed in
  */
 int
 xfs_ag_extend_space(
-	struct xfs_mount	*mp,
+	struct xfs_perag	*pag,
 	struct xfs_trans	*tp,
-	struct aghdr_init_data	*id,
 	xfs_extlen_t		len)
 {
+	struct xfs_mount	*mp = pag_mount(pag);
 	struct xfs_buf		*bp;
 	struct xfs_agi		*agi;
 	struct xfs_agf		*agf;
 	int			error;
 
-	/*
-	 * Change the agi length.
-	 */
-	error = xfs_ialloc_read_agi(mp, tp, id->agno, &bp);
+	ASSERT(pag_agno(pag) == mp->m_sb.sb_agcount - 1);
+
+	error = xfs_ialloc_read_agi(pag, tp, 0, &bp);
 	if (error)
 		return error;
 
-	agi = XFS_BUF_TO_AGI(bp);
+	agi = bp->b_addr;
 	be32_add_cpu(&agi->agi_length, len);
-	ASSERT(id->agno == mp->m_sb.sb_agcount - 1 ||
-	       be32_to_cpu(agi->agi_length) == mp->m_sb.sb_agblocks);
 	xfs_ialloc_log_agi(tp, bp, XFS_AGI_LENGTH);
 
 	/*
 	 * Change agf length.
 	 */
-	error = xfs_alloc_read_agf(mp, tp, id->agno, 0, &bp);
+	error = xfs_alloc_read_agf(pag, tp, 0, &bp);
 	if (error)
 		return error;
 
-	agf = XFS_BUF_TO_AGF(bp);
+	agf = bp->b_addr;
 	be32_add_cpu(&agf->agf_length, len);
 	ASSERT(agf->agf_length == agi->agi_length);
 	xfs_alloc_log_agf(tp, bp, XFS_AGF_LENGTH);
@@ -450,14 +915,64 @@ xfs_ag_extend_space(
 	 * XFS_RMAP_OINFO_SKIP_UPDATE is used here to tell the rmap btree that
 	 * this doesn't actually exist in the rmap btree.
 	 */
-	error = xfs_rmap_free(tp, bp, id->agno,
-				be32_to_cpu(agf->agf_length) - len,
+	error = xfs_rmap_free(tp, bp, pag, be32_to_cpu(agf->agf_length) - len,
 				len, &XFS_RMAP_OINFO_SKIP_UPDATE);
 	if (error)
 		return error;
 
-	return  xfs_free_extent(tp, XFS_AGB_TO_FSB(mp, id->agno,
-					be32_to_cpu(agf->agf_length) - len),
-				len, &XFS_RMAP_OINFO_SKIP_UPDATE,
-				XFS_AG_RESV_NONE);
+	error = xfs_free_extent(tp, pag, be32_to_cpu(agf->agf_length) - len,
+			len, &XFS_RMAP_OINFO_SKIP_UPDATE, XFS_AG_RESV_NONE);
+	if (error)
+		return error;
+
+	/* Update perag geometry */
+	pag_group(pag)->xg_block_count = be32_to_cpu(agf->agf_length);
+	__xfs_agino_range(mp, pag_group(pag)->xg_block_count, &pag->agino_min,
+			&pag->agino_max);
+	return 0;
+}
+
+/* Retrieve AG geometry. */
+int
+xfs_ag_get_geometry(
+	struct xfs_perag	*pag,
+	struct xfs_ag_geometry	*ageo)
+{
+	struct xfs_buf		*agi_bp;
+	struct xfs_buf		*agf_bp;
+	struct xfs_agi		*agi;
+	struct xfs_agf		*agf;
+	unsigned int		freeblks;
+	int			error;
+
+	/* Lock the AG headers. */
+	error = xfs_ialloc_read_agi(pag, NULL, 0, &agi_bp);
+	if (error)
+		return error;
+	error = xfs_alloc_read_agf(pag, NULL, 0, &agf_bp);
+	if (error)
+		goto out_agi;
+
+	/* Fill out form. */
+	memset(ageo, 0, sizeof(*ageo));
+	ageo->ag_number = pag_agno(pag);
+
+	agi = agi_bp->b_addr;
+	ageo->ag_icount = be32_to_cpu(agi->agi_count);
+	ageo->ag_ifree = be32_to_cpu(agi->agi_freecount);
+
+	agf = agf_bp->b_addr;
+	ageo->ag_length = be32_to_cpu(agf->agf_length);
+	freeblks = pag->pagf_freeblks +
+		   pag->pagf_flcount +
+		   pag->pagf_btreeblks -
+		   xfs_ag_resv_needed(pag, XFS_AG_RESV_NONE);
+	ageo->ag_freeblks = freeblks;
+	xfs_ag_geom_health(pag, ageo);
+
+	/* Release resources. */
+	xfs_buf_relse(agf_bp);
+out_agi:
+	xfs_buf_relse(agi_bp);
+	return error;
 }