From 71be6b4942dd64bc17728f82f787be98fd8afed7 Mon Sep 17 00:00:00 2001
From: "Darrick J. Wong" <darrick.wong@oracle.com>
Date: Mon, 3 Oct 2016 09:11:14 -0700
Subject: vfs: add a FALLOC_FL_UNSHARE mode to fallocate to unshare a range of
 blocks

Add a new fallocate mode flag that explicitly unshares blocks on
filesystems that support such features.  The new flag can only
be used with an allocate-mode fallocate call.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 fs/open.c | 5 +++++
 1 file changed, 5 insertions(+)

(limited to 'fs')

diff --git a/fs/open.c b/fs/open.c
index 4fd6e256f4f4..d58525dda28d 100644
--- a/fs/open.c
+++ b/fs/open.c
@@ -256,6 +256,11 @@ int vfs_fallocate(struct file *file, int mode, loff_t offset, loff_t len)
 	    (mode & ~FALLOC_FL_INSERT_RANGE))
 		return -EINVAL;
 
+	/* Unshare range should only be used with allocate mode. */
+	if ((mode & FALLOC_FL_UNSHARE_RANGE) &&
+	    (mode & ~(FALLOC_FL_UNSHARE_RANGE | FALLOC_FL_KEEP_SIZE)))
+		return -EINVAL;
+
 	if (!(file->f_mode & FMODE_WRITE))
 		return -EBADF;
 
-- 
cgit v1.2.3


From 9cdafd8a769b7c3a54e474393fb69d1fc2c42185 Mon Sep 17 00:00:00 2001
From: "Darrick J. Wong" <darrick.wong@oracle.com>
Date: Mon, 3 Oct 2016 09:11:15 -0700
Subject: xfs: return an error when an inline directory is too small

If the size of an inline directory is so small that it doesn't
even cover the required header size, return an error to userspace
instead of ASSERTing and returning 0 like everything's ok.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reported-by: Jan Kara <jack@suse.cz>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/xfs_dir2_readdir.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

(limited to 'fs')

diff --git a/fs/xfs/xfs_dir2_readdir.c b/fs/xfs/xfs_dir2_readdir.c
index f44f79996978..29816981b50a 100644
--- a/fs/xfs/xfs_dir2_readdir.c
+++ b/fs/xfs/xfs_dir2_readdir.c
@@ -84,7 +84,8 @@ xfs_dir2_sf_getdents(
 
 	sfp = (xfs_dir2_sf_hdr_t *)dp->i_df.if_u1.if_data;
 
-	ASSERT(dp->i_d.di_size >= xfs_dir2_sf_hdr_size(sfp->i8count));
+	if (dp->i_d.di_size < xfs_dir2_sf_hdr_size(sfp->i8count))
+		return -EFSCORRUPTED;
 
 	/*
 	 * If the block number in the offset is out of range, we're done.
-- 
cgit v1.2.3


From c75c752d03741a0ebda3ea4c5020ba22c7570561 Mon Sep 17 00:00:00 2001
From: "Darrick J. Wong" <darrick.wong@oracle.com>
Date: Mon, 3 Oct 2016 09:11:15 -0700
Subject: xfs: define tracepoints for refcount btree activities

Define all the tracepoints we need to inspect the refcount btree
runtime operation.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/xfs_trace.h | 301 +++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 301 insertions(+)

(limited to 'fs')

diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h
index c6b2b1dcde75..8446338929f6 100644
--- a/fs/xfs/xfs_trace.h
+++ b/fs/xfs/xfs_trace.h
@@ -40,6 +40,16 @@ struct xfs_inode_log_format;
 struct xfs_bmbt_irec;
 struct xfs_btree_cur;
 
+#ifndef XFS_REFCOUNT_IREC_PLACEHOLDER
+#define XFS_REFCOUNT_IREC_PLACEHOLDER
+/* Placeholder definition to avoid breaking bisectability. */
+struct xfs_refcount_irec {
+	xfs_agblock_t	rc_startblock;	/* starting block number */
+	xfs_extlen_t	rc_blockcount;	/* count of free blocks */
+	xfs_nlink_t	rc_refcount;	/* number of inodes linked here */
+};
+#endif
+
 DECLARE_EVENT_CLASS(xfs_attr_list_class,
 	TP_PROTO(struct xfs_attr_list_context *ctx),
 	TP_ARGS(ctx),
@@ -2640,6 +2650,297 @@ DEFINE_AG_RESV_EVENT(xfs_ag_resv_needed);
 DEFINE_AG_ERROR_EVENT(xfs_ag_resv_free_error);
 DEFINE_AG_ERROR_EVENT(xfs_ag_resv_init_error);
 
+/* refcount tracepoint classes */
+
+/* reuse the discard trace class for agbno/aglen-based traces */
+#define DEFINE_AG_EXTENT_EVENT(name) DEFINE_DISCARD_EVENT(name)
+
+/* ag btree lookup tracepoint class */
+#define XFS_AG_BTREE_CMP_FORMAT_STR \
+	{ XFS_LOOKUP_EQ,	"eq" }, \
+	{ XFS_LOOKUP_LE,	"le" }, \
+	{ XFS_LOOKUP_GE,	"ge" }
+DECLARE_EVENT_CLASS(xfs_ag_btree_lookup_class,
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
+		 xfs_agblock_t agbno, xfs_lookup_t dir),
+	TP_ARGS(mp, agno, agbno, dir),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, agbno)
+		__field(xfs_lookup_t, dir)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->agno = agno;
+		__entry->agbno = agbno;
+		__entry->dir = dir;
+	),
+	TP_printk("dev %d:%d agno %u agbno %u cmp %s(%d)\n",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->agbno,
+		  __print_symbolic(__entry->dir, XFS_AG_BTREE_CMP_FORMAT_STR),
+		  __entry->dir)
+)
+
+#define DEFINE_AG_BTREE_LOOKUP_EVENT(name) \
+DEFINE_EVENT(xfs_ag_btree_lookup_class, name, \
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \
+		 xfs_agblock_t agbno, xfs_lookup_t dir), \
+	TP_ARGS(mp, agno, agbno, dir))
+
+/* single-rcext tracepoint class */
+DECLARE_EVENT_CLASS(xfs_refcount_extent_class,
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
+		 struct xfs_refcount_irec *irec),
+	TP_ARGS(mp, agno, irec),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, startblock)
+		__field(xfs_extlen_t, blockcount)
+		__field(xfs_nlink_t, refcount)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->agno = agno;
+		__entry->startblock = irec->rc_startblock;
+		__entry->blockcount = irec->rc_blockcount;
+		__entry->refcount = irec->rc_refcount;
+	),
+	TP_printk("dev %d:%d agno %u agbno %u len %u refcount %u\n",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->startblock,
+		  __entry->blockcount,
+		  __entry->refcount)
+)
+
+#define DEFINE_REFCOUNT_EXTENT_EVENT(name) \
+DEFINE_EVENT(xfs_refcount_extent_class, name, \
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \
+		 struct xfs_refcount_irec *irec), \
+	TP_ARGS(mp, agno, irec))
+
+/* single-rcext and an agbno tracepoint class */
+DECLARE_EVENT_CLASS(xfs_refcount_extent_at_class,
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
+		 struct xfs_refcount_irec *irec, xfs_agblock_t agbno),
+	TP_ARGS(mp, agno, irec, agbno),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, startblock)
+		__field(xfs_extlen_t, blockcount)
+		__field(xfs_nlink_t, refcount)
+		__field(xfs_agblock_t, agbno)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->agno = agno;
+		__entry->startblock = irec->rc_startblock;
+		__entry->blockcount = irec->rc_blockcount;
+		__entry->refcount = irec->rc_refcount;
+		__entry->agbno = agbno;
+	),
+	TP_printk("dev %d:%d agno %u agbno %u len %u refcount %u @ agbno %u\n",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->startblock,
+		  __entry->blockcount,
+		  __entry->refcount,
+		  __entry->agbno)
+)
+
+#define DEFINE_REFCOUNT_EXTENT_AT_EVENT(name) \
+DEFINE_EVENT(xfs_refcount_extent_at_class, name, \
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \
+		 struct xfs_refcount_irec *irec, xfs_agblock_t agbno), \
+	TP_ARGS(mp, agno, irec, agbno))
+
+/* double-rcext tracepoint class */
+DECLARE_EVENT_CLASS(xfs_refcount_double_extent_class,
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
+		 struct xfs_refcount_irec *i1, struct xfs_refcount_irec *i2),
+	TP_ARGS(mp, agno, i1, i2),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, i1_startblock)
+		__field(xfs_extlen_t, i1_blockcount)
+		__field(xfs_nlink_t, i1_refcount)
+		__field(xfs_agblock_t, i2_startblock)
+		__field(xfs_extlen_t, i2_blockcount)
+		__field(xfs_nlink_t, i2_refcount)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->agno = agno;
+		__entry->i1_startblock = i1->rc_startblock;
+		__entry->i1_blockcount = i1->rc_blockcount;
+		__entry->i1_refcount = i1->rc_refcount;
+		__entry->i2_startblock = i2->rc_startblock;
+		__entry->i2_blockcount = i2->rc_blockcount;
+		__entry->i2_refcount = i2->rc_refcount;
+	),
+	TP_printk("dev %d:%d agno %u agbno %u len %u refcount %u -- "
+		  "agbno %u len %u refcount %u\n",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->i1_startblock,
+		  __entry->i1_blockcount,
+		  __entry->i1_refcount,
+		  __entry->i2_startblock,
+		  __entry->i2_blockcount,
+		  __entry->i2_refcount)
+)
+
+#define DEFINE_REFCOUNT_DOUBLE_EXTENT_EVENT(name) \
+DEFINE_EVENT(xfs_refcount_double_extent_class, name, \
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \
+		 struct xfs_refcount_irec *i1, struct xfs_refcount_irec *i2), \
+	TP_ARGS(mp, agno, i1, i2))
+
+/* double-rcext and an agbno tracepoint class */
+DECLARE_EVENT_CLASS(xfs_refcount_double_extent_at_class,
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
+		 struct xfs_refcount_irec *i1, struct xfs_refcount_irec *i2,
+		 xfs_agblock_t agbno),
+	TP_ARGS(mp, agno, i1, i2, agbno),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, i1_startblock)
+		__field(xfs_extlen_t, i1_blockcount)
+		__field(xfs_nlink_t, i1_refcount)
+		__field(xfs_agblock_t, i2_startblock)
+		__field(xfs_extlen_t, i2_blockcount)
+		__field(xfs_nlink_t, i2_refcount)
+		__field(xfs_agblock_t, agbno)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->agno = agno;
+		__entry->i1_startblock = i1->rc_startblock;
+		__entry->i1_blockcount = i1->rc_blockcount;
+		__entry->i1_refcount = i1->rc_refcount;
+		__entry->i2_startblock = i2->rc_startblock;
+		__entry->i2_blockcount = i2->rc_blockcount;
+		__entry->i2_refcount = i2->rc_refcount;
+		__entry->agbno = agbno;
+	),
+	TP_printk("dev %d:%d agno %u agbno %u len %u refcount %u -- "
+		  "agbno %u len %u refcount %u @ agbno %u\n",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->i1_startblock,
+		  __entry->i1_blockcount,
+		  __entry->i1_refcount,
+		  __entry->i2_startblock,
+		  __entry->i2_blockcount,
+		  __entry->i2_refcount,
+		  __entry->agbno)
+)
+
+#define DEFINE_REFCOUNT_DOUBLE_EXTENT_AT_EVENT(name) \
+DEFINE_EVENT(xfs_refcount_double_extent_at_class, name, \
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \
+		 struct xfs_refcount_irec *i1, struct xfs_refcount_irec *i2, \
+		 xfs_agblock_t agbno), \
+	TP_ARGS(mp, agno, i1, i2, agbno))
+
+/* triple-rcext tracepoint class */
+DECLARE_EVENT_CLASS(xfs_refcount_triple_extent_class,
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
+		 struct xfs_refcount_irec *i1, struct xfs_refcount_irec *i2,
+		 struct xfs_refcount_irec *i3),
+	TP_ARGS(mp, agno, i1, i2, i3),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, i1_startblock)
+		__field(xfs_extlen_t, i1_blockcount)
+		__field(xfs_nlink_t, i1_refcount)
+		__field(xfs_agblock_t, i2_startblock)
+		__field(xfs_extlen_t, i2_blockcount)
+		__field(xfs_nlink_t, i2_refcount)
+		__field(xfs_agblock_t, i3_startblock)
+		__field(xfs_extlen_t, i3_blockcount)
+		__field(xfs_nlink_t, i3_refcount)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->agno = agno;
+		__entry->i1_startblock = i1->rc_startblock;
+		__entry->i1_blockcount = i1->rc_blockcount;
+		__entry->i1_refcount = i1->rc_refcount;
+		__entry->i2_startblock = i2->rc_startblock;
+		__entry->i2_blockcount = i2->rc_blockcount;
+		__entry->i2_refcount = i2->rc_refcount;
+		__entry->i3_startblock = i3->rc_startblock;
+		__entry->i3_blockcount = i3->rc_blockcount;
+		__entry->i3_refcount = i3->rc_refcount;
+	),
+	TP_printk("dev %d:%d agno %u agbno %u len %u refcount %u -- "
+		  "agbno %u len %u refcount %u -- "
+		  "agbno %u len %u refcount %u\n",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->i1_startblock,
+		  __entry->i1_blockcount,
+		  __entry->i1_refcount,
+		  __entry->i2_startblock,
+		  __entry->i2_blockcount,
+		  __entry->i2_refcount,
+		  __entry->i3_startblock,
+		  __entry->i3_blockcount,
+		  __entry->i3_refcount)
+);
+
+#define DEFINE_REFCOUNT_TRIPLE_EXTENT_EVENT(name) \
+DEFINE_EVENT(xfs_refcount_triple_extent_class, name, \
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \
+		 struct xfs_refcount_irec *i1, struct xfs_refcount_irec *i2, \
+		 struct xfs_refcount_irec *i3), \
+	TP_ARGS(mp, agno, i1, i2, i3))
+
+/* refcount btree tracepoints */
+DEFINE_BUSY_EVENT(xfs_refcountbt_alloc_block);
+DEFINE_BUSY_EVENT(xfs_refcountbt_free_block);
+DEFINE_AG_BTREE_LOOKUP_EVENT(xfs_refcount_lookup);
+DEFINE_REFCOUNT_EXTENT_EVENT(xfs_refcount_get);
+DEFINE_REFCOUNT_EXTENT_EVENT(xfs_refcount_update);
+DEFINE_REFCOUNT_EXTENT_EVENT(xfs_refcount_insert);
+DEFINE_REFCOUNT_EXTENT_EVENT(xfs_refcount_delete);
+DEFINE_AG_ERROR_EVENT(xfs_refcount_insert_error);
+DEFINE_AG_ERROR_EVENT(xfs_refcount_delete_error);
+DEFINE_AG_ERROR_EVENT(xfs_refcount_update_error);
+
+/* refcount adjustment tracepoints */
+DEFINE_AG_EXTENT_EVENT(xfs_refcount_increase);
+DEFINE_AG_EXTENT_EVENT(xfs_refcount_decrease);
+DEFINE_REFCOUNT_TRIPLE_EXTENT_EVENT(xfs_refcount_merge_center_extents);
+DEFINE_REFCOUNT_EXTENT_EVENT(xfs_refcount_modify_extent);
+DEFINE_REFCOUNT_EXTENT_AT_EVENT(xfs_refcount_split_extent);
+DEFINE_REFCOUNT_DOUBLE_EXTENT_EVENT(xfs_refcount_merge_left_extent);
+DEFINE_REFCOUNT_DOUBLE_EXTENT_EVENT(xfs_refcount_merge_right_extent);
+DEFINE_REFCOUNT_DOUBLE_EXTENT_AT_EVENT(xfs_refcount_find_left_extent);
+DEFINE_REFCOUNT_DOUBLE_EXTENT_AT_EVENT(xfs_refcount_find_right_extent);
+DEFINE_AG_ERROR_EVENT(xfs_refcount_adjust_error);
+DEFINE_AG_ERROR_EVENT(xfs_refcount_merge_center_extents_error);
+DEFINE_AG_ERROR_EVENT(xfs_refcount_modify_extent_error);
+DEFINE_AG_ERROR_EVENT(xfs_refcount_split_extent_error);
+DEFINE_AG_ERROR_EVENT(xfs_refcount_merge_left_extent_error);
+DEFINE_AG_ERROR_EVENT(xfs_refcount_merge_right_extent_error);
+DEFINE_AG_ERROR_EVENT(xfs_refcount_find_left_extent_error);
+DEFINE_AG_ERROR_EVENT(xfs_refcount_find_right_extent_error);
+
+/* reflink helpers */
+DEFINE_AG_EXTENT_EVENT(xfs_refcount_find_shared);
+DEFINE_AG_EXTENT_EVENT(xfs_refcount_find_shared_result);
+DEFINE_AG_ERROR_EVENT(xfs_refcount_find_shared_error);
+
 #endif /* _TRACE_XFS_H */
 
 #undef TRACE_INCLUDE_PATH
-- 
cgit v1.2.3


From 46eeb521b95247170d2db773bb4cc8fb3de1d85c Mon Sep 17 00:00:00 2001
From: "Darrick J. Wong" <darrick.wong@oracle.com>
Date: Mon, 3 Oct 2016 09:11:16 -0700
Subject: xfs: introduce refcount btree definitions

Add new per-AG refcount btree definitions to the per-AG structures.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_alloc.c      |  5 +++++
 fs/xfs/libxfs/xfs_btree.c      |  5 +++--
 fs/xfs/libxfs/xfs_btree.h      |  4 ++++
 fs/xfs/libxfs/xfs_format.h     | 31 +++++++++++++++++++++++++++----
 fs/xfs/libxfs/xfs_rmap_btree.c | 22 ++++++++++++++++++++--
 fs/xfs/libxfs/xfs_types.h      |  2 +-
 fs/xfs/xfs_inode.h             |  5 +++++
 fs/xfs/xfs_mount.h             |  3 +++
 fs/xfs/xfs_pnfs.c              |  7 +++++++
 fs/xfs/xfs_stats.c             |  1 +
 fs/xfs/xfs_stats.h             | 18 +++++++++++++++++-
 11 files changed, 93 insertions(+), 10 deletions(-)

(limited to 'fs')

diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c
index ca75dc90ebe0..275d3459e9e8 100644
--- a/fs/xfs/libxfs/xfs_alloc.c
+++ b/fs/xfs/libxfs/xfs_alloc.c
@@ -2458,6 +2458,10 @@ xfs_agf_verify(
 	    be32_to_cpu(agf->agf_btreeblks) > be32_to_cpu(agf->agf_length))
 		return false;
 
+	if (xfs_sb_version_hasreflink(&mp->m_sb) &&
+	    be32_to_cpu(agf->agf_refcount_level) > XFS_BTREE_MAXLEVELS)
+		return false;
+
 	return true;;
 
 }
@@ -2578,6 +2582,7 @@ xfs_alloc_read_agf(
 			be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNTi]);
 		pag->pagf_levels[XFS_BTNUM_RMAPi] =
 			be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAPi]);
+		pag->pagf_refcount_level = be32_to_cpu(agf->agf_refcount_level);
 		spin_lock_init(&pag->pagb_lock);
 		pag->pagb_count = 0;
 		pag->pagb_tree = RB_ROOT;
diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c
index aa1752f918b8..f8bab9b22755 100644
--- a/fs/xfs/libxfs/xfs_btree.c
+++ b/fs/xfs/libxfs/xfs_btree.c
@@ -45,9 +45,10 @@ kmem_zone_t	*xfs_btree_cur_zone;
  */
 static const __uint32_t xfs_magics[2][XFS_BTNUM_MAX] = {
 	{ XFS_ABTB_MAGIC, XFS_ABTC_MAGIC, 0, XFS_BMAP_MAGIC, XFS_IBT_MAGIC,
-	  XFS_FIBT_MAGIC },
+	  XFS_FIBT_MAGIC, 0 },
 	{ XFS_ABTB_CRC_MAGIC, XFS_ABTC_CRC_MAGIC, XFS_RMAP_CRC_MAGIC,
-	  XFS_BMAP_CRC_MAGIC, XFS_IBT_CRC_MAGIC, XFS_FIBT_CRC_MAGIC }
+	  XFS_BMAP_CRC_MAGIC, XFS_IBT_CRC_MAGIC, XFS_FIBT_CRC_MAGIC,
+	  XFS_REFC_CRC_MAGIC }
 };
 #define xfs_btree_magic(cur) \
 	xfs_magics[!!((cur)->bc_flags & XFS_BTREE_CRC_BLOCKS)][cur->bc_btnum]
diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h
index 3f8556a5c2ad..e7ef1d9dbe6e 100644
--- a/fs/xfs/libxfs/xfs_btree.h
+++ b/fs/xfs/libxfs/xfs_btree.h
@@ -72,6 +72,7 @@ union xfs_btree_rec {
 #define	XFS_BTNUM_INO	((xfs_btnum_t)XFS_BTNUM_INOi)
 #define	XFS_BTNUM_FINO	((xfs_btnum_t)XFS_BTNUM_FINOi)
 #define	XFS_BTNUM_RMAP	((xfs_btnum_t)XFS_BTNUM_RMAPi)
+#define	XFS_BTNUM_REFC	((xfs_btnum_t)XFS_BTNUM_REFCi)
 
 /*
  * For logging record fields.
@@ -105,6 +106,7 @@ do {    \
 	case XFS_BTNUM_INO: __XFS_BTREE_STATS_INC(__mp, ibt, stat); break; \
 	case XFS_BTNUM_FINO: __XFS_BTREE_STATS_INC(__mp, fibt, stat); break; \
 	case XFS_BTNUM_RMAP: __XFS_BTREE_STATS_INC(__mp, rmap, stat); break; \
+	case XFS_BTNUM_REFC: __XFS_BTREE_STATS_INC(__mp, refcbt, stat); break; \
 	case XFS_BTNUM_MAX: ASSERT(0); /* fucking gcc */ ; break;	\
 	}       \
 } while (0)
@@ -127,6 +129,8 @@ do {    \
 		__XFS_BTREE_STATS_ADD(__mp, fibt, stat, val); break; \
 	case XFS_BTNUM_RMAP:	\
 		__XFS_BTREE_STATS_ADD(__mp, rmap, stat, val); break; \
+	case XFS_BTNUM_REFC:	\
+		__XFS_BTREE_STATS_ADD(__mp, refcbt, stat, val); break; \
 	case XFS_BTNUM_MAX: ASSERT(0); /* fucking gcc */ ; break; \
 	}       \
 } while (0)
diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index 270fb5cf4fa1..57c52a6c7bb8 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -456,6 +456,7 @@ xfs_sb_has_compat_feature(
 
 #define XFS_SB_FEAT_RO_COMPAT_FINOBT   (1 << 0)		/* free inode btree */
 #define XFS_SB_FEAT_RO_COMPAT_RMAPBT   (1 << 1)		/* reverse map btree */
+#define XFS_SB_FEAT_RO_COMPAT_REFLINK  (1 << 2)		/* reflinked files */
 #define XFS_SB_FEAT_RO_COMPAT_ALL \
 		(XFS_SB_FEAT_RO_COMPAT_FINOBT | \
 		 XFS_SB_FEAT_RO_COMPAT_RMAPBT)
@@ -546,6 +547,12 @@ static inline bool xfs_sb_version_hasrmapbt(struct xfs_sb *sbp)
 		(sbp->sb_features_ro_compat & XFS_SB_FEAT_RO_COMPAT_RMAPBT);
 }
 
+static inline bool xfs_sb_version_hasreflink(struct xfs_sb *sbp)
+{
+	return XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5 &&
+		(sbp->sb_features_ro_compat & XFS_SB_FEAT_RO_COMPAT_REFLINK);
+}
+
 /*
  * end of superblock version macros
  */
@@ -641,14 +648,17 @@ typedef struct xfs_agf {
 	uuid_t		agf_uuid;	/* uuid of filesystem */
 
 	__be32		agf_rmap_blocks;	/* rmapbt blocks used */
-	__be32		agf_padding;		/* padding */
+	__be32		agf_refcount_blocks;	/* refcountbt blocks used */
+
+	__be32		agf_refcount_root;	/* refcount tree root block */
+	__be32		agf_refcount_level;	/* refcount btree levels */
 
 	/*
 	 * reserve some contiguous space for future logged fields before we add
 	 * the unlogged fields. This makes the range logging via flags and
 	 * structure offsets much simpler.
 	 */
-	__be64		agf_spare64[15];
+	__be64		agf_spare64[14];
 
 	/* unlogged fields, written during buffer writeback. */
 	__be64		agf_lsn;	/* last write sequence */
@@ -1041,9 +1051,14 @@ static inline void xfs_dinode_put_rdev(struct xfs_dinode *dip, xfs_dev_t rdev)
  * 16 bits of the XFS_XFLAG_s range.
  */
 #define XFS_DIFLAG2_DAX_BIT	0	/* use DAX for this inode */
+#define XFS_DIFLAG2_REFLINK_BIT	1	/* file's blocks may be shared */
+#define XFS_DIFLAG2_COWEXTSIZE_BIT   2  /* copy on write extent size hint */
 #define XFS_DIFLAG2_DAX		(1 << XFS_DIFLAG2_DAX_BIT)
+#define XFS_DIFLAG2_REFLINK     (1 << XFS_DIFLAG2_REFLINK_BIT)
+#define XFS_DIFLAG2_COWEXTSIZE  (1 << XFS_DIFLAG2_COWEXTSIZE_BIT)
 
-#define XFS_DIFLAG2_ANY		(XFS_DIFLAG2_DAX)
+#define XFS_DIFLAG2_ANY \
+	(XFS_DIFLAG2_DAX | XFS_DIFLAG2_REFLINK | XFS_DIFLAG2_COWEXTSIZE)
 
 /*
  * Inode number format:
@@ -1353,7 +1368,8 @@ struct xfs_owner_info {
 #define XFS_RMAP_OWN_AG		(-5ULL)	/* AG freespace btree blocks */
 #define XFS_RMAP_OWN_INOBT	(-6ULL)	/* Inode btree blocks */
 #define XFS_RMAP_OWN_INODES	(-7ULL)	/* Inode chunk */
-#define XFS_RMAP_OWN_MIN	(-8ULL) /* guard */
+#define XFS_RMAP_OWN_REFC	(-8ULL) /* refcount tree */
+#define XFS_RMAP_OWN_MIN	(-9ULL) /* guard */
 
 #define XFS_RMAP_NON_INODE_OWNER(owner)	(!!((owner) & (1ULL << 63)))
 
@@ -1433,6 +1449,13 @@ typedef __be32 xfs_rmap_ptr_t;
 	 XFS_FIBT_BLOCK(mp) + 1 : \
 	 XFS_IBT_BLOCK(mp) + 1)
 
+/*
+ * Reference Count Btree format definitions
+ *
+ */
+#define	XFS_REFC_CRC_MAGIC	0x52334643	/* 'R3FC' */
+
+
 /*
  * BMAP Btree format definitions
  *
diff --git a/fs/xfs/libxfs/xfs_rmap_btree.c b/fs/xfs/libxfs/xfs_rmap_btree.c
index 17b8eeb34ac8..9c0585e7db1f 100644
--- a/fs/xfs/libxfs/xfs_rmap_btree.c
+++ b/fs/xfs/libxfs/xfs_rmap_btree.c
@@ -512,6 +512,24 @@ void
 xfs_rmapbt_compute_maxlevels(
 	struct xfs_mount		*mp)
 {
-	mp->m_rmap_maxlevels = xfs_btree_compute_maxlevels(mp,
-			mp->m_rmap_mnr, mp->m_sb.sb_agblocks);
+	/*
+	 * On a non-reflink filesystem, the maximum number of rmap
+	 * records is the number of blocks in the AG, hence the max
+	 * rmapbt height is log_$maxrecs($agblocks).  However, with
+	 * reflink each AG block can have up to 2^32 (per the refcount
+	 * record format) owners, which means that theoretically we
+	 * could face up to 2^64 rmap records.
+	 *
+	 * That effectively means that the max rmapbt height must be
+	 * XFS_BTREE_MAXLEVELS.  "Fortunately" we'll run out of AG
+	 * blocks to feed the rmapbt long before the rmapbt reaches
+	 * maximum height.  The reflink code uses ag_resv_critical to
+	 * disallow reflinking when less than 10% of the per-AG metadata
+	 * block reservation since the fallback is a regular file copy.
+	 */
+	if (xfs_sb_version_hasreflink(&mp->m_sb))
+		mp->m_rmap_maxlevels = XFS_BTREE_MAXLEVELS;
+	else
+		mp->m_rmap_maxlevels = xfs_btree_compute_maxlevels(mp,
+				mp->m_rmap_mnr, mp->m_sb.sb_agblocks);
 }
diff --git a/fs/xfs/libxfs/xfs_types.h b/fs/xfs/libxfs/xfs_types.h
index 3d503647f26b..be7b6de5137d 100644
--- a/fs/xfs/libxfs/xfs_types.h
+++ b/fs/xfs/libxfs/xfs_types.h
@@ -109,7 +109,7 @@ typedef enum {
 
 typedef enum {
 	XFS_BTNUM_BNOi, XFS_BTNUM_CNTi, XFS_BTNUM_RMAPi, XFS_BTNUM_BMAPi,
-	XFS_BTNUM_INOi, XFS_BTNUM_FINOi, XFS_BTNUM_MAX
+	XFS_BTNUM_INOi, XFS_BTNUM_FINOi, XFS_BTNUM_REFCi, XFS_BTNUM_MAX
 } xfs_btnum_t;
 
 struct xfs_name {
diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h
index 8f30d2533b48..a8658e68041a 100644
--- a/fs/xfs/xfs_inode.h
+++ b/fs/xfs/xfs_inode.h
@@ -202,6 +202,11 @@ xfs_get_initial_prid(struct xfs_inode *dp)
 	return XFS_PROJID_DEFAULT;
 }
 
+static inline bool xfs_is_reflink_inode(struct xfs_inode *ip)
+{
+	return ip->i_d.di_flags2 & XFS_DIFLAG2_REFLINK;
+}
+
 /*
  * In-core inode flags.
  */
diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h
index 041d9493e798..8fab4962c4f1 100644
--- a/fs/xfs/xfs_mount.h
+++ b/fs/xfs/xfs_mount.h
@@ -399,6 +399,9 @@ typedef struct xfs_perag {
 	struct xfs_ag_resv	pag_meta_resv;
 	/* Blocks reserved for just AGFL-based metadata. */
 	struct xfs_ag_resv	pag_agfl_resv;
+
+	/* reference count */
+	__uint8_t		pagf_refcount_level;
 } xfs_perag_t;
 
 static inline struct xfs_ag_resv *
diff --git a/fs/xfs/xfs_pnfs.c b/fs/xfs/xfs_pnfs.c
index 0f14b2e4bf6c..93a7aafa56d6 100644
--- a/fs/xfs/xfs_pnfs.c
+++ b/fs/xfs/xfs_pnfs.c
@@ -113,6 +113,13 @@ xfs_fs_map_blocks(
 	if (XFS_IS_REALTIME_INODE(ip))
 		return -ENXIO;
 
+	/*
+	 * The pNFS block layout spec actually supports reflink like
+	 * functionality, but the Linux pNFS server doesn't implement it yet.
+	 */
+	if (xfs_is_reflink_inode(ip))
+		return -ENXIO;
+
 	/*
 	 * Lock out any other I/O before we flush and invalidate the pagecache,
 	 * and then hand out a layout to the remote system.  This is very
diff --git a/fs/xfs/xfs_stats.c b/fs/xfs/xfs_stats.c
index 6e812fe0fd43..12d48cd8f8a4 100644
--- a/fs/xfs/xfs_stats.c
+++ b/fs/xfs/xfs_stats.c
@@ -62,6 +62,7 @@ int xfs_stats_format(struct xfsstats __percpu *stats, char *buf)
 		{ "ibt2",		XFSSTAT_END_IBT_V2		},
 		{ "fibt2",		XFSSTAT_END_FIBT_V2		},
 		{ "rmapbt",		XFSSTAT_END_RMAP_V2		},
+		{ "refcntbt",		XFSSTAT_END_REFCOUNT		},
 		/* we print both series of quota information together */
 		{ "qm",			XFSSTAT_END_QM			},
 	};
diff --git a/fs/xfs/xfs_stats.h b/fs/xfs/xfs_stats.h
index 657865f51e78..79ad2e69fc33 100644
--- a/fs/xfs/xfs_stats.h
+++ b/fs/xfs/xfs_stats.h
@@ -213,7 +213,23 @@ struct xfsstats {
 	__uint32_t		xs_rmap_2_alloc;
 	__uint32_t		xs_rmap_2_free;
 	__uint32_t		xs_rmap_2_moves;
-#define XFSSTAT_END_XQMSTAT		(XFSSTAT_END_RMAP_V2+6)
+#define XFSSTAT_END_REFCOUNT		(XFSSTAT_END_RMAP_V2 + 15)
+	__uint32_t		xs_refcbt_2_lookup;
+	__uint32_t		xs_refcbt_2_compare;
+	__uint32_t		xs_refcbt_2_insrec;
+	__uint32_t		xs_refcbt_2_delrec;
+	__uint32_t		xs_refcbt_2_newroot;
+	__uint32_t		xs_refcbt_2_killroot;
+	__uint32_t		xs_refcbt_2_increment;
+	__uint32_t		xs_refcbt_2_decrement;
+	__uint32_t		xs_refcbt_2_lshift;
+	__uint32_t		xs_refcbt_2_rshift;
+	__uint32_t		xs_refcbt_2_split;
+	__uint32_t		xs_refcbt_2_join;
+	__uint32_t		xs_refcbt_2_alloc;
+	__uint32_t		xs_refcbt_2_free;
+	__uint32_t		xs_refcbt_2_moves;
+#define XFSSTAT_END_XQMSTAT		(XFSSTAT_END_REFCOUNT + 6)
 	__uint32_t		xs_qm_dqreclaims;
 	__uint32_t		xs_qm_dqreclaim_misses;
 	__uint32_t		xs_qm_dquot_dups;
-- 
cgit v1.2.3


From af30dfa14411e9df0e69c6e46e8c6c467b88229d Mon Sep 17 00:00:00 2001
From: "Darrick J. Wong" <darrick.wong@oracle.com>
Date: Mon, 3 Oct 2016 09:11:17 -0700
Subject: xfs: refcount btree add more reserved blocks

Since XFS reserves a small amount of space in each AG as the minimum
free space needed for an operation, save some more space in case we
touch the refcount btree.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_alloc.c  | 13 +++++++++++++
 fs/xfs/libxfs/xfs_format.h |  2 ++
 2 files changed, 15 insertions(+)

(limited to 'fs')

diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c
index 275d3459e9e8..aa0e1ca7b6a5 100644
--- a/fs/xfs/libxfs/xfs_alloc.c
+++ b/fs/xfs/libxfs/xfs_alloc.c
@@ -52,10 +52,23 @@ STATIC int xfs_alloc_ag_vextent_size(xfs_alloc_arg_t *);
 STATIC int xfs_alloc_ag_vextent_small(xfs_alloc_arg_t *,
 		xfs_btree_cur_t *, xfs_agblock_t *, xfs_extlen_t *, int *);
 
+unsigned int
+xfs_refc_block(
+	struct xfs_mount	*mp)
+{
+	if (xfs_sb_version_hasrmapbt(&mp->m_sb))
+		return XFS_RMAP_BLOCK(mp) + 1;
+	if (xfs_sb_version_hasfinobt(&mp->m_sb))
+		return XFS_FIBT_BLOCK(mp) + 1;
+	return XFS_IBT_BLOCK(mp) + 1;
+}
+
 xfs_extlen_t
 xfs_prealloc_blocks(
 	struct xfs_mount	*mp)
 {
+	if (xfs_sb_version_hasreflink(&mp->m_sb))
+		return xfs_refc_block(mp) + 1;
 	if (xfs_sb_version_hasrmapbt(&mp->m_sb))
 		return XFS_RMAP_BLOCK(mp) + 1;
 	if (xfs_sb_version_hasfinobt(&mp->m_sb))
diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index 57c52a6c7bb8..622055bb283b 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -1455,6 +1455,8 @@ typedef __be32 xfs_rmap_ptr_t;
  */
 #define	XFS_REFC_CRC_MAGIC	0x52334643	/* 'R3FC' */
 
+unsigned int xfs_refc_block(struct xfs_mount *mp);
+
 
 /*
  * BMAP Btree format definitions
-- 
cgit v1.2.3


From 1946b91cee4fc8ae25450673e4d4f35e9b462e9e Mon Sep 17 00:00:00 2001
From: "Darrick J. Wong" <darrick.wong@oracle.com>
Date: Mon, 3 Oct 2016 09:11:18 -0700
Subject: xfs: define the on-disk refcount btree format

Start constructing the refcount btree implementation by establishing
the on-disk format and everything needed to read, write, and
manipulate the refcount btree blocks.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/Makefile                    |   1 +
 fs/xfs/libxfs/xfs_btree.c          |   3 +
 fs/xfs/libxfs/xfs_btree.h          |  12 +++
 fs/xfs/libxfs/xfs_format.h         |  36 ++++++++
 fs/xfs/libxfs/xfs_refcount_btree.c | 178 +++++++++++++++++++++++++++++++++++++
 fs/xfs/libxfs/xfs_refcount_btree.h |  67 ++++++++++++++
 fs/xfs/libxfs/xfs_sb.c             |   9 ++
 fs/xfs/libxfs/xfs_shared.h         |   2 +
 fs/xfs/libxfs/xfs_trans_resv.c     |   2 +-
 fs/xfs/libxfs/xfs_trans_resv.h     |   1 +
 fs/xfs/xfs_mount.c                 |   2 +
 fs/xfs/xfs_mount.h                 |   3 +
 fs/xfs/xfs_ondisk.h                |   3 +
 fs/xfs/xfs_trace.h                 |  11 +--
 14 files changed, 319 insertions(+), 11 deletions(-)
 create mode 100644 fs/xfs/libxfs/xfs_refcount_btree.c
 create mode 100644 fs/xfs/libxfs/xfs_refcount_btree.h

(limited to 'fs')

diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile
index 584e87e11cb6..8d749f2df99e 100644
--- a/fs/xfs/Makefile
+++ b/fs/xfs/Makefile
@@ -55,6 +55,7 @@ xfs-y				+= $(addprefix libxfs/, \
 				   xfs_ag_resv.o \
 				   xfs_rmap.o \
 				   xfs_rmap_btree.o \
+				   xfs_refcount_btree.o \
 				   xfs_sb.o \
 				   xfs_symlink_remote.o \
 				   xfs_trans_resv.o \
diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c
index f8bab9b22755..5c8e6f2ce44f 100644
--- a/fs/xfs/libxfs/xfs_btree.c
+++ b/fs/xfs/libxfs/xfs_btree.c
@@ -1217,6 +1217,9 @@ xfs_btree_set_refs(
 	case XFS_BTNUM_RMAP:
 		xfs_buf_set_ref(bp, XFS_RMAP_BTREE_REF);
 		break;
+	case XFS_BTNUM_REFC:
+		xfs_buf_set_ref(bp, XFS_REFC_BTREE_REF);
+		break;
 	default:
 		ASSERT(0);
 	}
diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h
index e7ef1d9dbe6e..c2b01d1c79ee 100644
--- a/fs/xfs/libxfs/xfs_btree.h
+++ b/fs/xfs/libxfs/xfs_btree.h
@@ -49,6 +49,7 @@ union xfs_btree_key {
 	struct xfs_inobt_key		inobt;
 	struct xfs_rmap_key		rmap;
 	struct xfs_rmap_key		__rmap_bigkey[2];
+	struct xfs_refcount_key		refc;
 };
 
 union xfs_btree_rec {
@@ -57,6 +58,7 @@ union xfs_btree_rec {
 	struct xfs_alloc_rec		alloc;
 	struct xfs_inobt_rec		inobt;
 	struct xfs_rmap_rec		rmap;
+	struct xfs_refcount_rec		refc;
 };
 
 /*
@@ -221,6 +223,15 @@ union xfs_btree_irec {
 	struct xfs_bmbt_irec		b;
 	struct xfs_inobt_rec_incore	i;
 	struct xfs_rmap_irec		r;
+	struct xfs_refcount_irec	rc;
+};
+
+/* Per-AG btree private information. */
+union xfs_btree_cur_private {
+	struct {
+		unsigned long	nr_ops;		/* # record updates */
+		int		shape_changes;	/* # of extent splits */
+	} refc;
 };
 
 /*
@@ -247,6 +258,7 @@ typedef struct xfs_btree_cur
 			struct xfs_buf	*agbp;	/* agf/agi buffer pointer */
 			struct xfs_defer_ops *dfops;	/* deferred updates */
 			xfs_agnumber_t	agno;	/* ag number */
+			union xfs_btree_cur_private	priv;
 		} a;
 		struct {			/* needed for BMAP */
 			struct xfs_inode *ip;	/* pointer to our inode */
diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index 622055bb283b..97c74f4a22b5 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -1457,6 +1457,42 @@ typedef __be32 xfs_rmap_ptr_t;
 
 unsigned int xfs_refc_block(struct xfs_mount *mp);
 
+/*
+ * Data record/key structure
+ *
+ * Each record associates a range of physical blocks (starting at
+ * rc_startblock and ending rc_blockcount blocks later) with a reference
+ * count (rc_refcount).  Extents that are being used to stage a copy on
+ * write (CoW) operation are recorded in the refcount btree with a
+ * refcount of 1.  All other records must have a refcount > 1 and must
+ * track an extent mapped only by file data forks.
+ *
+ * Extents with a single owner (attributes, metadata, non-shared file
+ * data) are not tracked here.  Free space is also not tracked here.
+ * This is consistent with pre-reflink XFS.
+ */
+struct xfs_refcount_rec {
+	__be32		rc_startblock;	/* starting block number */
+	__be32		rc_blockcount;	/* count of blocks */
+	__be32		rc_refcount;	/* number of inodes linked here */
+};
+
+struct xfs_refcount_key {
+	__be32		rc_startblock;	/* starting block number */
+};
+
+struct xfs_refcount_irec {
+	xfs_agblock_t	rc_startblock;	/* starting block number */
+	xfs_extlen_t	rc_blockcount;	/* count of free blocks */
+	xfs_nlink_t	rc_refcount;	/* number of inodes linked here */
+};
+
+#define MAXREFCOUNT	((xfs_nlink_t)~0U)
+#define MAXREFCEXTLEN	((xfs_extlen_t)~0U)
+
+/* btree pointer type */
+typedef __be32 xfs_refcount_ptr_t;
+
 
 /*
  * BMAP Btree format definitions
diff --git a/fs/xfs/libxfs/xfs_refcount_btree.c b/fs/xfs/libxfs/xfs_refcount_btree.c
new file mode 100644
index 000000000000..359cf0c3a6b4
--- /dev/null
+++ b/fs/xfs/libxfs/xfs_refcount_btree.c
@@ -0,0 +1,178 @@
+/*
+ * Copyright (C) 2016 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_sb.h"
+#include "xfs_mount.h"
+#include "xfs_btree.h"
+#include "xfs_bmap.h"
+#include "xfs_refcount_btree.h"
+#include "xfs_alloc.h"
+#include "xfs_error.h"
+#include "xfs_trace.h"
+#include "xfs_cksum.h"
+#include "xfs_trans.h"
+#include "xfs_bit.h"
+
+static struct xfs_btree_cur *
+xfs_refcountbt_dup_cursor(
+	struct xfs_btree_cur	*cur)
+{
+	return xfs_refcountbt_init_cursor(cur->bc_mp, cur->bc_tp,
+			cur->bc_private.a.agbp, cur->bc_private.a.agno,
+			cur->bc_private.a.dfops);
+}
+
+STATIC bool
+xfs_refcountbt_verify(
+	struct xfs_buf		*bp)
+{
+	struct xfs_mount	*mp = bp->b_target->bt_mount;
+	struct xfs_btree_block	*block = XFS_BUF_TO_BLOCK(bp);
+	struct xfs_perag	*pag = bp->b_pag;
+	unsigned int		level;
+
+	if (block->bb_magic != cpu_to_be32(XFS_REFC_CRC_MAGIC))
+		return false;
+
+	if (!xfs_sb_version_hasreflink(&mp->m_sb))
+		return false;
+	if (!xfs_btree_sblock_v5hdr_verify(bp))
+		return false;
+
+	level = be16_to_cpu(block->bb_level);
+	if (pag && pag->pagf_init) {
+		if (level >= pag->pagf_refcount_level)
+			return false;
+	} else if (level >= mp->m_refc_maxlevels)
+		return false;
+
+	return xfs_btree_sblock_verify(bp, mp->m_refc_mxr[level != 0]);
+}
+
+STATIC void
+xfs_refcountbt_read_verify(
+	struct xfs_buf	*bp)
+{
+	if (!xfs_btree_sblock_verify_crc(bp))
+		xfs_buf_ioerror(bp, -EFSBADCRC);
+	else if (!xfs_refcountbt_verify(bp))
+		xfs_buf_ioerror(bp, -EFSCORRUPTED);
+
+	if (bp->b_error) {
+		trace_xfs_btree_corrupt(bp, _RET_IP_);
+		xfs_verifier_error(bp);
+	}
+}
+
+STATIC void
+xfs_refcountbt_write_verify(
+	struct xfs_buf	*bp)
+{
+	if (!xfs_refcountbt_verify(bp)) {
+		trace_xfs_btree_corrupt(bp, _RET_IP_);
+		xfs_buf_ioerror(bp, -EFSCORRUPTED);
+		xfs_verifier_error(bp);
+		return;
+	}
+	xfs_btree_sblock_calc_crc(bp);
+
+}
+
+const struct xfs_buf_ops xfs_refcountbt_buf_ops = {
+	.name			= "xfs_refcountbt",
+	.verify_read		= xfs_refcountbt_read_verify,
+	.verify_write		= xfs_refcountbt_write_verify,
+};
+
+static const struct xfs_btree_ops xfs_refcountbt_ops = {
+	.rec_len		= sizeof(struct xfs_refcount_rec),
+	.key_len		= sizeof(struct xfs_refcount_key),
+
+	.dup_cursor		= xfs_refcountbt_dup_cursor,
+	.buf_ops		= &xfs_refcountbt_buf_ops,
+};
+
+/*
+ * Allocate a new refcount btree cursor.
+ */
+struct xfs_btree_cur *
+xfs_refcountbt_init_cursor(
+	struct xfs_mount	*mp,
+	struct xfs_trans	*tp,
+	struct xfs_buf		*agbp,
+	xfs_agnumber_t		agno,
+	struct xfs_defer_ops	*dfops)
+{
+	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
+	struct xfs_btree_cur	*cur;
+
+	ASSERT(agno != NULLAGNUMBER);
+	ASSERT(agno < mp->m_sb.sb_agcount);
+	cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_NOFS);
+
+	cur->bc_tp = tp;
+	cur->bc_mp = mp;
+	cur->bc_btnum = XFS_BTNUM_REFC;
+	cur->bc_blocklog = mp->m_sb.sb_blocklog;
+	cur->bc_ops = &xfs_refcountbt_ops;
+
+	cur->bc_nlevels = be32_to_cpu(agf->agf_refcount_level);
+
+	cur->bc_private.a.agbp = agbp;
+	cur->bc_private.a.agno = agno;
+	cur->bc_private.a.dfops = dfops;
+	cur->bc_flags |= XFS_BTREE_CRC_BLOCKS;
+
+	cur->bc_private.a.priv.refc.nr_ops = 0;
+	cur->bc_private.a.priv.refc.shape_changes = 0;
+
+	return cur;
+}
+
+/*
+ * Calculate the number of records in a refcount btree block.
+ */
+int
+xfs_refcountbt_maxrecs(
+	struct xfs_mount	*mp,
+	int			blocklen,
+	bool			leaf)
+{
+	blocklen -= XFS_REFCOUNT_BLOCK_LEN;
+
+	if (leaf)
+		return blocklen / sizeof(struct xfs_refcount_rec);
+	return blocklen / (sizeof(struct xfs_refcount_key) +
+			   sizeof(xfs_refcount_ptr_t));
+}
+
+/* Compute the maximum height of a refcount btree. */
+void
+xfs_refcountbt_compute_maxlevels(
+	struct xfs_mount		*mp)
+{
+	mp->m_refc_maxlevels = xfs_btree_compute_maxlevels(mp,
+			mp->m_refc_mnr, mp->m_sb.sb_agblocks);
+}
diff --git a/fs/xfs/libxfs/xfs_refcount_btree.h b/fs/xfs/libxfs/xfs_refcount_btree.h
new file mode 100644
index 000000000000..9e9ad7c6d267
--- /dev/null
+++ b/fs/xfs/libxfs/xfs_refcount_btree.h
@@ -0,0 +1,67 @@
+/*
+ * Copyright (C) 2016 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#ifndef __XFS_REFCOUNT_BTREE_H__
+#define	__XFS_REFCOUNT_BTREE_H__
+
+/*
+ * Reference Count Btree on-disk structures
+ */
+
+struct xfs_buf;
+struct xfs_btree_cur;
+struct xfs_mount;
+
+/*
+ * Btree block header size
+ */
+#define XFS_REFCOUNT_BLOCK_LEN	XFS_BTREE_SBLOCK_CRC_LEN
+
+/*
+ * Record, key, and pointer address macros for btree blocks.
+ *
+ * (note that some of these may appear unused, but they are used in userspace)
+ */
+#define XFS_REFCOUNT_REC_ADDR(block, index) \
+	((struct xfs_refcount_rec *) \
+		((char *)(block) + \
+		 XFS_REFCOUNT_BLOCK_LEN + \
+		 (((index) - 1) * sizeof(struct xfs_refcount_rec))))
+
+#define XFS_REFCOUNT_KEY_ADDR(block, index) \
+	((struct xfs_refcount_key *) \
+		((char *)(block) + \
+		 XFS_REFCOUNT_BLOCK_LEN + \
+		 ((index) - 1) * sizeof(struct xfs_refcount_key)))
+
+#define XFS_REFCOUNT_PTR_ADDR(block, index, maxrecs) \
+	((xfs_refcount_ptr_t *) \
+		((char *)(block) + \
+		 XFS_REFCOUNT_BLOCK_LEN + \
+		 (maxrecs) * sizeof(struct xfs_refcount_key) + \
+		 ((index) - 1) * sizeof(xfs_refcount_ptr_t)))
+
+extern struct xfs_btree_cur *xfs_refcountbt_init_cursor(struct xfs_mount *mp,
+		struct xfs_trans *tp, struct xfs_buf *agbp, xfs_agnumber_t agno,
+		struct xfs_defer_ops *dfops);
+extern int xfs_refcountbt_maxrecs(struct xfs_mount *mp, int blocklen,
+		bool leaf);
+extern void xfs_refcountbt_compute_maxlevels(struct xfs_mount *mp);
+
+#endif	/* __XFS_REFCOUNT_BTREE_H__ */
diff --git a/fs/xfs/libxfs/xfs_sb.c b/fs/xfs/libxfs/xfs_sb.c
index 4aecc5fefe96..a70aec910626 100644
--- a/fs/xfs/libxfs/xfs_sb.c
+++ b/fs/xfs/libxfs/xfs_sb.c
@@ -38,6 +38,8 @@
 #include "xfs_ialloc_btree.h"
 #include "xfs_log.h"
 #include "xfs_rmap_btree.h"
+#include "xfs_bmap.h"
+#include "xfs_refcount_btree.h"
 
 /*
  * Physical superblock buffer manipulations. Shared with libxfs in userspace.
@@ -737,6 +739,13 @@ xfs_sb_mount_common(
 	mp->m_rmap_mnr[0] = mp->m_rmap_mxr[0] / 2;
 	mp->m_rmap_mnr[1] = mp->m_rmap_mxr[1] / 2;
 
+	mp->m_refc_mxr[0] = xfs_refcountbt_maxrecs(mp, sbp->sb_blocksize,
+			true);
+	mp->m_refc_mxr[1] = xfs_refcountbt_maxrecs(mp, sbp->sb_blocksize,
+			false);
+	mp->m_refc_mnr[0] = mp->m_refc_mxr[0] / 2;
+	mp->m_refc_mnr[1] = mp->m_refc_mxr[1] / 2;
+
 	mp->m_bsize = XFS_FSB_TO_BB(mp, 1);
 	mp->m_ialloc_inos = (int)MAX((__uint16_t)XFS_INODES_PER_CHUNK,
 					sbp->sb_inopblock);
diff --git a/fs/xfs/libxfs/xfs_shared.h b/fs/xfs/libxfs/xfs_shared.h
index 0c5b30bd884c..c6f4eb46fe26 100644
--- a/fs/xfs/libxfs/xfs_shared.h
+++ b/fs/xfs/libxfs/xfs_shared.h
@@ -39,6 +39,7 @@ extern const struct xfs_buf_ops xfs_agf_buf_ops;
 extern const struct xfs_buf_ops xfs_agfl_buf_ops;
 extern const struct xfs_buf_ops xfs_allocbt_buf_ops;
 extern const struct xfs_buf_ops xfs_rmapbt_buf_ops;
+extern const struct xfs_buf_ops xfs_refcountbt_buf_ops;
 extern const struct xfs_buf_ops xfs_attr3_leaf_buf_ops;
 extern const struct xfs_buf_ops xfs_attr3_rmt_buf_ops;
 extern const struct xfs_buf_ops xfs_bmbt_buf_ops;
@@ -122,6 +123,7 @@ int	xfs_log_calc_minimum_size(struct xfs_mount *);
 #define	XFS_INO_REF		2
 #define	XFS_ATTR_BTREE_REF	1
 #define	XFS_DQUOT_REF		1
+#define	XFS_REFC_BTREE_REF	1
 
 /*
  * Flags for xfs_trans_ichgtime().
diff --git a/fs/xfs/libxfs/xfs_trans_resv.c b/fs/xfs/libxfs/xfs_trans_resv.c
index 301ef2f4dbd6..7c840e1c8bb6 100644
--- a/fs/xfs/libxfs/xfs_trans_resv.c
+++ b/fs/xfs/libxfs/xfs_trans_resv.c
@@ -73,7 +73,7 @@ xfs_calc_buf_res(
  *
  * Keep in mind that max depth is calculated separately for each type of tree.
  */
-static uint
+uint
 xfs_allocfree_log_count(
 	struct xfs_mount *mp,
 	uint		num_ops)
diff --git a/fs/xfs/libxfs/xfs_trans_resv.h b/fs/xfs/libxfs/xfs_trans_resv.h
index 0eb46ed6d404..36a15110c1fe 100644
--- a/fs/xfs/libxfs/xfs_trans_resv.h
+++ b/fs/xfs/libxfs/xfs_trans_resv.h
@@ -102,5 +102,6 @@ struct xfs_trans_resv {
 #define	XFS_ATTRRM_LOG_COUNT		3
 
 void xfs_trans_resv_calc(struct xfs_mount *mp, struct xfs_trans_resv *resp);
+uint xfs_allocfree_log_count(struct xfs_mount *mp, uint num_ops);
 
 #endif	/* __XFS_TRANS_RESV_H__ */
diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c
index 56e85a6c85c7..3f64615f58db 100644
--- a/fs/xfs/xfs_mount.c
+++ b/fs/xfs/xfs_mount.c
@@ -43,6 +43,7 @@
 #include "xfs_icache.h"
 #include "xfs_sysfs.h"
 #include "xfs_rmap_btree.h"
+#include "xfs_refcount_btree.h"
 
 
 static DEFINE_MUTEX(xfs_uuid_table_mutex);
@@ -684,6 +685,7 @@ xfs_mountfs(
 	xfs_bmap_compute_maxlevels(mp, XFS_ATTR_FORK);
 	xfs_ialloc_compute_maxlevels(mp);
 	xfs_rmapbt_compute_maxlevels(mp);
+	xfs_refcountbt_compute_maxlevels(mp);
 
 	xfs_set_maxicount(mp);
 
diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h
index 8fab4962c4f1..0be14a76216a 100644
--- a/fs/xfs/xfs_mount.h
+++ b/fs/xfs/xfs_mount.h
@@ -124,10 +124,13 @@ typedef struct xfs_mount {
 	uint			m_inobt_mnr[2];	/* min inobt btree records */
 	uint			m_rmap_mxr[2];	/* max rmap btree records */
 	uint			m_rmap_mnr[2];	/* min rmap btree records */
+	uint			m_refc_mxr[2];	/* max refc btree records */
+	uint			m_refc_mnr[2];	/* min refc btree records */
 	uint			m_ag_maxlevels;	/* XFS_AG_MAXLEVELS */
 	uint			m_bm_maxlevels[2]; /* XFS_BM_MAXLEVELS */
 	uint			m_in_maxlevels;	/* max inobt btree levels. */
 	uint			m_rmap_maxlevels; /* max rmap btree levels */
+	uint			m_refc_maxlevels; /* max refcount btree level */
 	xfs_extlen_t		m_ag_prealloc_blocks; /* reserved ag blocks */
 	uint			m_alloc_set_aside; /* space we can't use */
 	uint			m_ag_max_usable; /* max space per AG */
diff --git a/fs/xfs/xfs_ondisk.h b/fs/xfs/xfs_ondisk.h
index 69e2986a3776..0c381d71b242 100644
--- a/fs/xfs/xfs_ondisk.h
+++ b/fs/xfs/xfs_ondisk.h
@@ -49,6 +49,8 @@ xfs_check_ondisk_structs(void)
 	XFS_CHECK_STRUCT_SIZE(struct xfs_dsymlink_hdr,		56);
 	XFS_CHECK_STRUCT_SIZE(struct xfs_inobt_key,		4);
 	XFS_CHECK_STRUCT_SIZE(struct xfs_inobt_rec,		16);
+	XFS_CHECK_STRUCT_SIZE(struct xfs_refcount_key,		4);
+	XFS_CHECK_STRUCT_SIZE(struct xfs_refcount_rec,		12);
 	XFS_CHECK_STRUCT_SIZE(struct xfs_rmap_key,		20);
 	XFS_CHECK_STRUCT_SIZE(struct xfs_rmap_rec,		24);
 	XFS_CHECK_STRUCT_SIZE(struct xfs_timestamp,		8);
@@ -56,6 +58,7 @@ xfs_check_ondisk_structs(void)
 	XFS_CHECK_STRUCT_SIZE(xfs_alloc_ptr_t,			4);
 	XFS_CHECK_STRUCT_SIZE(xfs_alloc_rec_t,			8);
 	XFS_CHECK_STRUCT_SIZE(xfs_inobt_ptr_t,			4);
+	XFS_CHECK_STRUCT_SIZE(xfs_refcount_ptr_t,		4);
 	XFS_CHECK_STRUCT_SIZE(xfs_rmap_ptr_t,			4);
 
 	/* dir/attr trees */
diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h
index 8446338929f6..c7b9853c7ae9 100644
--- a/fs/xfs/xfs_trace.h
+++ b/fs/xfs/xfs_trace.h
@@ -39,16 +39,7 @@ struct xfs_buf_log_format;
 struct xfs_inode_log_format;
 struct xfs_bmbt_irec;
 struct xfs_btree_cur;
-
-#ifndef XFS_REFCOUNT_IREC_PLACEHOLDER
-#define XFS_REFCOUNT_IREC_PLACEHOLDER
-/* Placeholder definition to avoid breaking bisectability. */
-struct xfs_refcount_irec {
-	xfs_agblock_t	rc_startblock;	/* starting block number */
-	xfs_extlen_t	rc_blockcount;	/* count of free blocks */
-	xfs_nlink_t	rc_refcount;	/* number of inodes linked here */
-};
-#endif
+struct xfs_refcount_irec;
 
 DECLARE_EVENT_CLASS(xfs_attr_list_class,
 	TP_PROTO(struct xfs_attr_list_context *ctx),
-- 
cgit v1.2.3


From ac4fef69386afb1c25d6b900fe1fcf77b96ea567 Mon Sep 17 00:00:00 2001
From: "Darrick J. Wong" <darrick.wong@oracle.com>
Date: Mon, 3 Oct 2016 09:11:18 -0700
Subject: xfs: add refcount btree support to growfs

Modify the growfs code to initialize new refcount btree blocks.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/xfs_fsops.c | 39 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 39 insertions(+)

(limited to 'fs')

diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c
index 94ac06f3d908..4b4059b5e501 100644
--- a/fs/xfs/xfs_fsops.c
+++ b/fs/xfs/xfs_fsops.c
@@ -259,6 +259,12 @@ xfs_growfs_data_private(
 		agf->agf_longest = cpu_to_be32(tmpsize);
 		if (xfs_sb_version_hascrc(&mp->m_sb))
 			uuid_copy(&agf->agf_uuid, &mp->m_sb.sb_meta_uuid);
+		if (xfs_sb_version_hasreflink(&mp->m_sb)) {
+			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);
+		}
 
 		error = xfs_bwrite(bp);
 		xfs_buf_relse(bp);
@@ -450,6 +456,17 @@ xfs_growfs_data_private(
 			rrec->rm_offset = 0;
 			be16_add_cpu(&block->bb_numrecs, 1);
 
+			/* account for refc btree root */
+			if (xfs_sb_version_hasreflink(&mp->m_sb)) {
+				rrec = XFS_RMAP_REC_ADDR(block, 5);
+				rrec->rm_startblock = cpu_to_be32(
+						xfs_refc_block(mp));
+				rrec->rm_blockcount = cpu_to_be32(1);
+				rrec->rm_owner = cpu_to_be64(XFS_RMAP_OWN_REFC);
+				rrec->rm_offset = 0;
+				be16_add_cpu(&block->bb_numrecs, 1);
+			}
+
 			error = xfs_bwrite(bp);
 			xfs_buf_relse(bp);
 			if (error)
@@ -507,6 +524,28 @@ xfs_growfs_data_private(
 				goto error0;
 		}
 
+		/*
+		 * refcount btree root block
+		 */
+		if (xfs_sb_version_hasreflink(&mp->m_sb)) {
+			bp = xfs_growfs_get_hdr_buf(mp,
+				XFS_AGB_TO_DADDR(mp, agno, xfs_refc_block(mp)),
+				BTOBB(mp->m_sb.sb_blocksize), 0,
+				&xfs_refcountbt_buf_ops);
+			if (!bp) {
+				error = -ENOMEM;
+				goto error0;
+			}
+
+			xfs_btree_init_block(mp, bp, XFS_REFC_CRC_MAGIC,
+					     0, 0, agno,
+					     XFS_BTREE_CRC_BLOCKS);
+
+			error = xfs_bwrite(bp);
+			xfs_buf_relse(bp);
+			if (error)
+				goto error0;
+		}
 	}
 	xfs_trans_agblocks_delta(tp, nfree);
 	/*
-- 
cgit v1.2.3


From f310bd2ecd37b17bf0042c9d1595329057970eb6 Mon Sep 17 00:00:00 2001
From: "Darrick J. Wong" <darrick.wong@oracle.com>
Date: Mon, 3 Oct 2016 09:11:19 -0700
Subject: xfs: account for the refcount btree in the alloc/free log reservation

Every time we allocate or free a data extent, we might need to split
the refcount btree.  Reserve some blocks in the transaction to handle
this possibility.  Even though the deferred refcount code can roll a
transaction to avoid overloading the transaction, we can still exceed
the reservation.

Certain pathological workloads (1k blocks, no cowextsize hint, random
directio writes), cause a perfect storm wherein a refcount adjustment
of a large range of blocks causes full tree splits in two separate
extents in two separate refcount tree blocks; allocating new refcount
tree blocks causes rmap btree splits; and all the allocation activity
causes the freespace btrees to split, blowing the reservation.

(Reproduced by generic/167 over NFS atop XFS)

Signed-off-by: Christoph Hellwig <hch@lst.de>
[darrick.wong@oracle.com: add commit message]
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 fs/xfs/libxfs/xfs_trans_resv.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

(limited to 'fs')

diff --git a/fs/xfs/libxfs/xfs_trans_resv.c b/fs/xfs/libxfs/xfs_trans_resv.c
index 7c840e1c8bb6..a59838f1627b 100644
--- a/fs/xfs/libxfs/xfs_trans_resv.c
+++ b/fs/xfs/libxfs/xfs_trans_resv.c
@@ -67,7 +67,8 @@ xfs_calc_buf_res(
  * Per-extent log reservation for the btree changes involved in freeing or
  * allocating an extent.  In classic XFS there were two trees that will be
  * modified (bnobt + cntbt).  With rmap enabled, there are three trees
- * (rmapbt).  The number of blocks reserved is based on the formula:
+ * (rmapbt).  With reflink, there are four trees (refcountbt).  The number of
+ * blocks reserved is based on the formula:
  *
  * num trees * ((2 blocks/level * max depth) - 1)
  *
@@ -83,6 +84,8 @@ xfs_allocfree_log_count(
 	blocks = num_ops * 2 * (2 * mp->m_ag_maxlevels - 1);
 	if (xfs_sb_version_hasrmapbt(&mp->m_sb))
 		blocks += num_ops * (2 * mp->m_rmap_maxlevels - 1);
+	if (xfs_sb_version_hasreflink(&mp->m_sb))
+		blocks += num_ops * (2 * mp->m_refc_maxlevels - 1);
 
 	return blocks;
 }
-- 
cgit v1.2.3


From bdf28630b72154e5766cbad5874576b6f22e7237 Mon Sep 17 00:00:00 2001
From: "Darrick J. Wong" <darrick.wong@oracle.com>
Date: Mon, 3 Oct 2016 09:11:19 -0700
Subject: xfs: add refcount btree operations

Implement the generic btree operations required to manipulate refcount
btree blocks.  The implementation is similar to the bmapbt, though it
will only allocate and free blocks from the AG.

Since the refcount root and level fields are separate from the
existing roots and levels array, they need a separate logging flag.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
[hch: fix logging of AGF refcount btree fields]
Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/Makefile                    |   1 +
 fs/xfs/libxfs/xfs_alloc.c          |   3 +
 fs/xfs/libxfs/xfs_format.h         |  10 +-
 fs/xfs/libxfs/xfs_refcount.c       | 177 +++++++++++++++++++++++++++++++
 fs/xfs/libxfs/xfs_refcount.h       |  30 ++++++
 fs/xfs/libxfs/xfs_refcount_btree.c | 211 +++++++++++++++++++++++++++++++++++++
 6 files changed, 430 insertions(+), 2 deletions(-)
 create mode 100644 fs/xfs/libxfs/xfs_refcount.c
 create mode 100644 fs/xfs/libxfs/xfs_refcount.h

(limited to 'fs')

diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile
index 8d749f2df99e..98b24279b75f 100644
--- a/fs/xfs/Makefile
+++ b/fs/xfs/Makefile
@@ -55,6 +55,7 @@ xfs-y				+= $(addprefix libxfs/, \
 				   xfs_ag_resv.o \
 				   xfs_rmap.o \
 				   xfs_rmap_btree.o \
+				   xfs_refcount.o \
 				   xfs_refcount_btree.o \
 				   xfs_sb.o \
 				   xfs_symlink_remote.o \
diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c
index aa0e1ca7b6a5..be7e3fcefd73 100644
--- a/fs/xfs/libxfs/xfs_alloc.c
+++ b/fs/xfs/libxfs/xfs_alloc.c
@@ -2334,6 +2334,9 @@ xfs_alloc_log_agf(
 		offsetof(xfs_agf_t, agf_btreeblks),
 		offsetof(xfs_agf_t, agf_uuid),
 		offsetof(xfs_agf_t, agf_rmap_blocks),
+		offsetof(xfs_agf_t, agf_refcount_blocks),
+		offsetof(xfs_agf_t, agf_refcount_root),
+		offsetof(xfs_agf_t, agf_refcount_level),
 		/* needed so that we don't log the whole rest of the structure: */
 		offsetof(xfs_agf_t, agf_spare64),
 		sizeof(xfs_agf_t)
diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index 97c74f4a22b5..8b826102dfe3 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -684,8 +684,11 @@ typedef struct xfs_agf {
 #define	XFS_AGF_BTREEBLKS	0x00000800
 #define	XFS_AGF_UUID		0x00001000
 #define	XFS_AGF_RMAP_BLOCKS	0x00002000
-#define	XFS_AGF_SPARE64		0x00004000
-#define	XFS_AGF_NUM_BITS	15
+#define	XFS_AGF_REFCOUNT_BLOCKS	0x00004000
+#define	XFS_AGF_REFCOUNT_ROOT	0x00008000
+#define	XFS_AGF_REFCOUNT_LEVEL	0x00010000
+#define	XFS_AGF_SPARE64		0x00020000
+#define	XFS_AGF_NUM_BITS	18
 #define	XFS_AGF_ALL_BITS	((1 << XFS_AGF_NUM_BITS) - 1)
 
 #define XFS_AGF_FLAGS \
@@ -703,6 +706,9 @@ typedef struct xfs_agf {
 	{ XFS_AGF_BTREEBLKS,	"BTREEBLKS" }, \
 	{ XFS_AGF_UUID,		"UUID" }, \
 	{ XFS_AGF_RMAP_BLOCKS,	"RMAP_BLOCKS" }, \
+	{ XFS_AGF_REFCOUNT_BLOCKS,	"REFCOUNT_BLOCKS" }, \
+	{ XFS_AGF_REFCOUNT_ROOT,	"REFCOUNT_ROOT" }, \
+	{ XFS_AGF_REFCOUNT_LEVEL,	"REFCOUNT_LEVEL" }, \
 	{ XFS_AGF_SPARE64,	"SPARE64" }
 
 /* disk block (xfs_daddr_t) in the AG */
diff --git a/fs/xfs/libxfs/xfs_refcount.c b/fs/xfs/libxfs/xfs_refcount.c
new file mode 100644
index 000000000000..de1340652933
--- /dev/null
+++ b/fs/xfs/libxfs/xfs_refcount.c
@@ -0,0 +1,177 @@
+/*
+ * Copyright (C) 2016 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_sb.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_bmap.h"
+#include "xfs_refcount_btree.h"
+#include "xfs_alloc.h"
+#include "xfs_error.h"
+#include "xfs_trace.h"
+#include "xfs_cksum.h"
+#include "xfs_trans.h"
+#include "xfs_bit.h"
+#include "xfs_refcount.h"
+
+/*
+ * Look up the first record less than or equal to [bno, len] in the btree
+ * given by cur.
+ */
+int
+xfs_refcount_lookup_le(
+	struct xfs_btree_cur	*cur,
+	xfs_agblock_t		bno,
+	int			*stat)
+{
+	trace_xfs_refcount_lookup(cur->bc_mp, cur->bc_private.a.agno, bno,
+			XFS_LOOKUP_LE);
+	cur->bc_rec.rc.rc_startblock = bno;
+	cur->bc_rec.rc.rc_blockcount = 0;
+	return xfs_btree_lookup(cur, XFS_LOOKUP_LE, stat);
+}
+
+/*
+ * Look up the first record greater than or equal to [bno, len] in the btree
+ * given by cur.
+ */
+int
+xfs_refcount_lookup_ge(
+	struct xfs_btree_cur	*cur,
+	xfs_agblock_t		bno,
+	int			*stat)
+{
+	trace_xfs_refcount_lookup(cur->bc_mp, cur->bc_private.a.agno, bno,
+			XFS_LOOKUP_GE);
+	cur->bc_rec.rc.rc_startblock = bno;
+	cur->bc_rec.rc.rc_blockcount = 0;
+	return xfs_btree_lookup(cur, XFS_LOOKUP_GE, stat);
+}
+
+/*
+ * Get the data from the pointed-to record.
+ */
+int
+xfs_refcount_get_rec(
+	struct xfs_btree_cur		*cur,
+	struct xfs_refcount_irec	*irec,
+	int				*stat)
+{
+	union xfs_btree_rec	*rec;
+	int			error;
+
+	error = xfs_btree_get_rec(cur, &rec, stat);
+	if (!error && *stat == 1) {
+		irec->rc_startblock = be32_to_cpu(rec->refc.rc_startblock);
+		irec->rc_blockcount = be32_to_cpu(rec->refc.rc_blockcount);
+		irec->rc_refcount = be32_to_cpu(rec->refc.rc_refcount);
+		trace_xfs_refcount_get(cur->bc_mp, cur->bc_private.a.agno,
+				irec);
+	}
+	return error;
+}
+
+/*
+ * Update the record referred to by cur to the value given
+ * by [bno, len, refcount].
+ * This either works (return 0) or gets an EFSCORRUPTED error.
+ */
+STATIC int
+xfs_refcount_update(
+	struct xfs_btree_cur		*cur,
+	struct xfs_refcount_irec	*irec)
+{
+	union xfs_btree_rec	rec;
+	int			error;
+
+	trace_xfs_refcount_update(cur->bc_mp, cur->bc_private.a.agno, irec);
+	rec.refc.rc_startblock = cpu_to_be32(irec->rc_startblock);
+	rec.refc.rc_blockcount = cpu_to_be32(irec->rc_blockcount);
+	rec.refc.rc_refcount = cpu_to_be32(irec->rc_refcount);
+	error = xfs_btree_update(cur, &rec);
+	if (error)
+		trace_xfs_refcount_update_error(cur->bc_mp,
+				cur->bc_private.a.agno, error, _RET_IP_);
+	return error;
+}
+
+/*
+ * Insert the record referred to by cur to the value given
+ * by [bno, len, refcount].
+ * This either works (return 0) or gets an EFSCORRUPTED error.
+ */
+STATIC int
+xfs_refcount_insert(
+	struct xfs_btree_cur		*cur,
+	struct xfs_refcount_irec	*irec,
+	int				*i)
+{
+	int				error;
+
+	trace_xfs_refcount_insert(cur->bc_mp, cur->bc_private.a.agno, irec);
+	cur->bc_rec.rc.rc_startblock = irec->rc_startblock;
+	cur->bc_rec.rc.rc_blockcount = irec->rc_blockcount;
+	cur->bc_rec.rc.rc_refcount = irec->rc_refcount;
+	error = xfs_btree_insert(cur, i);
+	XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, *i == 1, out_error);
+out_error:
+	if (error)
+		trace_xfs_refcount_insert_error(cur->bc_mp,
+				cur->bc_private.a.agno, error, _RET_IP_);
+	return error;
+}
+
+/*
+ * Remove the record referred to by cur, then set the pointer to the spot
+ * where the record could be re-inserted, in case we want to increment or
+ * decrement the cursor.
+ * This either works (return 0) or gets an EFSCORRUPTED error.
+ */
+STATIC int
+xfs_refcount_delete(
+	struct xfs_btree_cur	*cur,
+	int			*i)
+{
+	struct xfs_refcount_irec	irec;
+	int			found_rec;
+	int			error;
+
+	error = xfs_refcount_get_rec(cur, &irec, &found_rec);
+	if (error)
+		goto out_error;
+	XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, found_rec == 1, out_error);
+	trace_xfs_refcount_delete(cur->bc_mp, cur->bc_private.a.agno, &irec);
+	error = xfs_btree_delete(cur, i);
+	XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, *i == 1, out_error);
+	if (error)
+		goto out_error;
+	error = xfs_refcount_lookup_ge(cur, irec.rc_startblock, &found_rec);
+out_error:
+	if (error)
+		trace_xfs_refcount_delete_error(cur->bc_mp,
+				cur->bc_private.a.agno, error, _RET_IP_);
+	return error;
+}
diff --git a/fs/xfs/libxfs/xfs_refcount.h b/fs/xfs/libxfs/xfs_refcount.h
new file mode 100644
index 000000000000..4dc335a1dc73
--- /dev/null
+++ b/fs/xfs/libxfs/xfs_refcount.h
@@ -0,0 +1,30 @@
+/*
+ * Copyright (C) 2016 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#ifndef __XFS_REFCOUNT_H__
+#define __XFS_REFCOUNT_H__
+
+extern int xfs_refcount_lookup_le(struct xfs_btree_cur *cur,
+		xfs_agblock_t bno, int *stat);
+extern int xfs_refcount_lookup_ge(struct xfs_btree_cur *cur,
+		xfs_agblock_t bno, int *stat);
+extern int xfs_refcount_get_rec(struct xfs_btree_cur *cur,
+		struct xfs_refcount_irec *irec, int *stat);
+
+#endif	/* __XFS_REFCOUNT_H__ */
diff --git a/fs/xfs/libxfs/xfs_refcount_btree.c b/fs/xfs/libxfs/xfs_refcount_btree.c
index 359cf0c3a6b4..81d58b09af71 100644
--- a/fs/xfs/libxfs/xfs_refcount_btree.c
+++ b/fs/xfs/libxfs/xfs_refcount_btree.c
@@ -34,6 +34,7 @@
 #include "xfs_cksum.h"
 #include "xfs_trans.h"
 #include "xfs_bit.h"
+#include "xfs_rmap.h"
 
 static struct xfs_btree_cur *
 xfs_refcountbt_dup_cursor(
@@ -44,6 +45,178 @@ xfs_refcountbt_dup_cursor(
 			cur->bc_private.a.dfops);
 }
 
+STATIC void
+xfs_refcountbt_set_root(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr,
+	int			inc)
+{
+	struct xfs_buf		*agbp = cur->bc_private.a.agbp;
+	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
+	xfs_agnumber_t		seqno = be32_to_cpu(agf->agf_seqno);
+	struct xfs_perag	*pag = xfs_perag_get(cur->bc_mp, seqno);
+
+	ASSERT(ptr->s != 0);
+
+	agf->agf_refcount_root = ptr->s;
+	be32_add_cpu(&agf->agf_refcount_level, inc);
+	pag->pagf_refcount_level += inc;
+	xfs_perag_put(pag);
+
+	xfs_alloc_log_agf(cur->bc_tp, agbp,
+			XFS_AGF_REFCOUNT_ROOT | XFS_AGF_REFCOUNT_LEVEL);
+}
+
+STATIC int
+xfs_refcountbt_alloc_block(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*start,
+	union xfs_btree_ptr	*new,
+	int			*stat)
+{
+	struct xfs_buf		*agbp = cur->bc_private.a.agbp;
+	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
+	struct xfs_alloc_arg	args;		/* block allocation args */
+	int			error;		/* error return value */
+
+	memset(&args, 0, sizeof(args));
+	args.tp = cur->bc_tp;
+	args.mp = cur->bc_mp;
+	args.type = XFS_ALLOCTYPE_NEAR_BNO;
+	args.fsbno = XFS_AGB_TO_FSB(cur->bc_mp, cur->bc_private.a.agno,
+			xfs_refc_block(args.mp));
+	args.firstblock = args.fsbno;
+	xfs_rmap_ag_owner(&args.oinfo, XFS_RMAP_OWN_REFC);
+	args.minlen = args.maxlen = args.prod = 1;
+
+	error = xfs_alloc_vextent(&args);
+	if (error)
+		goto out_error;
+	trace_xfs_refcountbt_alloc_block(cur->bc_mp, cur->bc_private.a.agno,
+			args.agbno, 1);
+	if (args.fsbno == NULLFSBLOCK) {
+		XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+		*stat = 0;
+		return 0;
+	}
+	ASSERT(args.agno == cur->bc_private.a.agno);
+	ASSERT(args.len == 1);
+
+	new->s = cpu_to_be32(args.agbno);
+	be32_add_cpu(&agf->agf_refcount_blocks, 1);
+	xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_REFCOUNT_BLOCKS);
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 1;
+	return 0;
+
+out_error:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+}
+
+STATIC int
+xfs_refcountbt_free_block(
+	struct xfs_btree_cur	*cur,
+	struct xfs_buf		*bp)
+{
+	struct xfs_mount	*mp = cur->bc_mp;
+	struct xfs_buf		*agbp = cur->bc_private.a.agbp;
+	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
+	xfs_fsblock_t		fsbno = XFS_DADDR_TO_FSB(mp, XFS_BUF_ADDR(bp));
+	struct xfs_owner_info	oinfo;
+
+	trace_xfs_refcountbt_free_block(cur->bc_mp, cur->bc_private.a.agno,
+			XFS_FSB_TO_AGBNO(cur->bc_mp, fsbno), 1);
+	xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_REFC);
+	be32_add_cpu(&agf->agf_refcount_blocks, -1);
+	xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_REFCOUNT_BLOCKS);
+	xfs_bmap_add_free(mp, cur->bc_private.a.dfops, fsbno, 1,
+			&oinfo);
+
+	return 0;
+}
+
+STATIC int
+xfs_refcountbt_get_minrecs(
+	struct xfs_btree_cur	*cur,
+	int			level)
+{
+	return cur->bc_mp->m_refc_mnr[level != 0];
+}
+
+STATIC int
+xfs_refcountbt_get_maxrecs(
+	struct xfs_btree_cur	*cur,
+	int			level)
+{
+	return cur->bc_mp->m_refc_mxr[level != 0];
+}
+
+STATIC void
+xfs_refcountbt_init_key_from_rec(
+	union xfs_btree_key	*key,
+	union xfs_btree_rec	*rec)
+{
+	key->refc.rc_startblock = rec->refc.rc_startblock;
+}
+
+STATIC void
+xfs_refcountbt_init_high_key_from_rec(
+	union xfs_btree_key	*key,
+	union xfs_btree_rec	*rec)
+{
+	__u32			x;
+
+	x = be32_to_cpu(rec->refc.rc_startblock);
+	x += be32_to_cpu(rec->refc.rc_blockcount) - 1;
+	key->refc.rc_startblock = cpu_to_be32(x);
+}
+
+STATIC void
+xfs_refcountbt_init_rec_from_cur(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*rec)
+{
+	rec->refc.rc_startblock = cpu_to_be32(cur->bc_rec.rc.rc_startblock);
+	rec->refc.rc_blockcount = cpu_to_be32(cur->bc_rec.rc.rc_blockcount);
+	rec->refc.rc_refcount = cpu_to_be32(cur->bc_rec.rc.rc_refcount);
+}
+
+STATIC void
+xfs_refcountbt_init_ptr_from_cur(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr)
+{
+	struct xfs_agf		*agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
+
+	ASSERT(cur->bc_private.a.agno == be32_to_cpu(agf->agf_seqno));
+	ASSERT(agf->agf_refcount_root != 0);
+
+	ptr->s = agf->agf_refcount_root;
+}
+
+STATIC __int64_t
+xfs_refcountbt_key_diff(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_key	*key)
+{
+	struct xfs_refcount_irec	*rec = &cur->bc_rec.rc;
+	struct xfs_refcount_key		*kp = &key->refc;
+
+	return (__int64