[RFC PATCH v9.1 00/18] xfs: add realtime reverse-mapping support

[RFC PATCH v9.1 00/18] xfs: add realtime reverse-mapping support

[Darrick J. Wong]

Hi all,

This is the ninth revision of a patchset that adds to XFS kernel support
for reverse mapping for the realtime device.  The patches haven't
changed significantly since the previous posting in August 2017.  This
time around I've forward ported this to 4.19-rc1.

At the beginning of the set are patches to implement storing B+tree
leaves in an inode root, since the realtime rmapbt is rooted in an
inode, unlike the regular rmapbt which is rooted in an AG block.
Prior to this, the only btree that could be rooted in the inode fork
was the block mapping btree; if all the extent records fit in the
inode, format would be switched from 'btree' to 'extents'.

The next few patches widen the reverse mapping routines to fit the
64-bit numbers required to store information about the realtime
device and establish a new b+tree type (rtrmapbt) for the realtime
variant of the rmapbt.  After that are a few patches to handle rooting
the rtrmapbt in a specific inode that's referenced by the superblock.

Finally, there are patches to implement GETFSMAP with the rtrmapbt and
scrub functionality for the rtrmapbt and rtbitmap; and then wire up the
online scrub functionality.

If you're going to start using this mess, you probably ought to just
pull from my github trees for kernel[1], xfsprogs[2], xfstests[3].  The
kernel patches in the git trees should apply to 4.19-rc1; xfsprogs
patches to for-next; and xfstest to master.

This is an extraordinary way to eat your data.  Enjoy!
Comments and questions are, as always, welcome.

--D

[1] https://git.kernel.org/cgit/linux/kernel/git/djwong/xfs-linux.git/log/?h=djwong-wtf
[2] https://git.kernel.org/cgit/linux/kernel/git/djwong/xfsprogs-dev.git/log/?h=djwong-wtf
[3] https://git.kernel.org/cgit/linux/kernel/git/djwong/xfstests-dev.git/log/?h=djwong-wtf

[PATCH 01/18] xfs: make iroot_realloc a btree function

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

For btrees that are rooted in the inode core, we have to have a
function to resize the root.  This is fairly specific to each
btree type, so make xfs_iroot_realloc a per-btree function.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 fs/xfs/libxfs/xfs_bmap.c       |    6 +-
 fs/xfs/libxfs/xfs_bmap_btree.c |  137 ++++++++++++++++++++++++++++++++++++++++
 fs/xfs/libxfs/xfs_bmap_btree.h |    2 +
 fs/xfs/libxfs/xfs_btree.c      |   12 +---
 fs/xfs/libxfs/xfs_btree.h      |    7 ++
 fs/xfs/libxfs/xfs_inode_fork.c |  130 --------------------------------------
 fs/xfs/libxfs/xfs_inode_fork.h |    1 
 7 files changed, 153 insertions(+), 142 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c
index 1580ac995690..f61a0afabd88 100644
--- a/fs/xfs/libxfs/xfs_bmap.c
+++ b/fs/xfs/libxfs/xfs_bmap.c
@@ -631,7 +631,7 @@ xfs_bmap_btree_to_extents(
 	xfs_trans_binval(tp, cbp);
 	if (cur->bc_bufs[0] == cbp)
 		cur->bc_bufs[0] = NULL;
-	xfs_iroot_realloc(ip, -1, whichfork);
+	cur->bc_ops->iroot_realloc(cur, -1);
 	ASSERT(ifp->if_broot == NULL);
 	ASSERT((ifp->if_flags & XFS_IFBROOT) == 0);
 	XFS_IFORK_FMT_SET(ip, whichfork, XFS_DINODE_FMT_EXTENTS);
@@ -675,7 +675,7 @@ xfs_bmap_extents_to_btree(
 	/*
 	 * Make space in the inode incore.
 	 */
-	xfs_iroot_realloc(ip, 1, whichfork);
+	xfs_bmbt_iroot_realloc(ip, 1, whichfork);
 	ifp->if_flags |= XFS_IFBROOT;
 
 	/*
@@ -778,7 +778,7 @@ xfs_bmap_extents_to_btree(
 err2:
 	xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT, -1L);
 err1:
-	xfs_iroot_realloc(ip, -1, whichfork);
+	cur->bc_ops->iroot_realloc(cur, -1);
 	XFS_IFORK_FMT_SET(ip, whichfork, XFS_DINODE_FMT_EXTENTS);
 	xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
 
diff --git a/fs/xfs/libxfs/xfs_bmap_btree.c b/fs/xfs/libxfs/xfs_bmap_btree.c
index cdb74d2e2a43..4f7e566dd4bd 100644
--- a/fs/xfs/libxfs/xfs_bmap_btree.c
+++ b/fs/xfs/libxfs/xfs_bmap_btree.c
@@ -407,6 +407,142 @@ xfs_bmbt_diff_two_keys(
 			  be64_to_cpu(k2->bmbt.br_startoff);
 }
 
+/*
+ * Reallocate the space for if_broot based on the number of records
+ * being added or deleted as indicated in rec_diff.  Move the records
+ * and pointers in if_broot to fit the new size.  When shrinking this
+ * will eliminate holes between the records and pointers created by
+ * the caller.  When growing this will create holes to be filled in
+ * by the caller.
+ *
+ * The caller must not request to add more records than would fit in
+ * the on-disk inode root.  If the if_broot is currently NULL, then
+ * if we are adding records, one will be allocated.  The caller must also
+ * not request that the number of records go below zero, although
+ * it can go to zero.
+ *
+ * ip -- the inode whose if_broot area is changing
+ * ext_diff -- the change in the number of records, positive or negative,
+ *	 requested for the if_broot array.
+ */
+void
+xfs_bmbt_iroot_realloc(
+	struct xfs_inode	*ip,
+	int			rec_diff,
+	int			whichfork)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	int			cur_max;
+	struct xfs_ifork	*ifp;
+	struct xfs_btree_block	*new_broot;
+	int			new_max;
+	size_t			new_size;
+	char			*np;
+	char			*op;
+
+	/*
+	 * Handle the degenerate case quietly.
+	 */
+	if (rec_diff == 0)
+		return;
+
+	ifp = XFS_IFORK_PTR(ip, whichfork);
+	if (rec_diff > 0) {
+		/*
+		 * If there wasn't any memory allocated before, just
+		 * allocate it now and get out.
+		 */
+		if (ifp->if_broot_bytes == 0) {
+			new_size = XFS_BMAP_BROOT_SPACE_CALC(mp, rec_diff);
+			ifp->if_broot = kmem_alloc(new_size, KM_SLEEP | KM_NOFS);
+			ifp->if_broot_bytes = (int)new_size;
+			return;
+		}
+
+		/*
+		 * If there is already an existing if_broot, then we need
+		 * to realloc() it and shift the pointers to their new
+		 * location.  The records don't change location because
+		 * they are kept butted up against the btree block header.
+		 */
+		cur_max = xfs_bmbt_maxrecs(mp, ifp->if_broot_bytes, 0);
+		new_max = cur_max + rec_diff;
+		new_size = XFS_BMAP_BROOT_SPACE_CALC(mp, new_max);
+		ifp->if_broot = kmem_realloc(ifp->if_broot, new_size,
+				KM_SLEEP | KM_NOFS);
+		op = (char *)XFS_BMAP_BROOT_PTR_ADDR(mp, ifp->if_broot, 1,
+						     ifp->if_broot_bytes);
+		np = (char *)XFS_BMAP_BROOT_PTR_ADDR(mp, ifp->if_broot, 1,
+						     (int)new_size);
+		ifp->if_broot_bytes = (int)new_size;
+		ASSERT(XFS_BMAP_BMDR_SPACE(ifp->if_broot) <=
+			XFS_IFORK_SIZE(ip, whichfork));
+		memmove(np, op, cur_max * (uint)sizeof(xfs_fsblock_t));
+		return;
+	}
+
+	/*
+	 * rec_diff is less than 0.  In this case, we are shrinking the
+	 * if_broot buffer.  It must already exist.  If we go to zero
+	 * records, just get rid of the root and clear the status bit.
+	 */
+	ASSERT((ifp->if_broot != NULL) && (ifp->if_broot_bytes > 0));
+	cur_max = xfs_bmbt_maxrecs(mp, ifp->if_broot_bytes, 0);
+	new_max = cur_max + rec_diff;
+	ASSERT(new_max >= 0);
+	if (new_max > 0)
+		new_size = XFS_BMAP_BROOT_SPACE_CALC(mp, new_max);
+	else
+		new_size = 0;
+	if (new_size > 0) {
+		new_broot = kmem_alloc(new_size, KM_SLEEP | KM_NOFS);
+		/*
+		 * First copy over the btree block header.
+		 */
+		memcpy(new_broot, ifp->if_broot,
+			XFS_BMBT_BLOCK_LEN(ip->i_mount));
+	} else {
+		new_broot = NULL;
+		ifp->if_flags &= ~XFS_IFBROOT;
+	}
+
+	/*
+	 * Only copy the records and pointers if there are any.
+	 */
+	if (new_max > 0) {
+		/*
+		 * First copy the records.
+		 */
+		op = (char *)XFS_BMBT_REC_ADDR(mp, ifp->if_broot, 1);
+		np = (char *)XFS_BMBT_REC_ADDR(mp, new_broot, 1);
+		memcpy(np, op, new_max * (uint)sizeof(xfs_bmbt_rec_t));
+
+		/*
+		 * Then copy the pointers.
+		 */
+		op = (char *)XFS_BMAP_BROOT_PTR_ADDR(mp, ifp->if_broot, 1,
+						     ifp->if_broot_bytes);
+		np = (char *)XFS_BMAP_BROOT_PTR_ADDR(mp, new_broot, 1,
+						     (int)new_size);
+		memcpy(np, op, new_max * (uint)sizeof(xfs_fsblock_t));
+	}
+	kmem_free(ifp->if_broot);
+	ifp->if_broot = new_broot;
+	ifp->if_broot_bytes = (int)new_size;
+	if (ifp->if_broot)
+		ASSERT(XFS_BMAP_BMDR_SPACE(ifp->if_broot) <=
+			XFS_IFORK_SIZE(ip, whichfork));
+}
+
+STATIC void
+__xfs_bmbt_iroot_realloc(
+	struct xfs_btree_cur	*cur,
+	int			rec_diff)
+{
+	return xfs_bmbt_iroot_realloc(cur->bc_private.b.ip, rec_diff,
+			cur->bc_private.b.whichfork);
+}
+
 static xfs_failaddr_t
 xfs_bmbt_verify(
 	struct xfs_buf		*bp)
@@ -526,6 +662,7 @@ static const struct xfs_btree_ops xfs_bmbt_ops = {
 	.key_diff		= xfs_bmbt_key_diff,
 	.diff_two_keys		= xfs_bmbt_diff_two_keys,
 	.buf_ops		= &xfs_bmbt_buf_ops,
+	.iroot_realloc		= __xfs_bmbt_iroot_realloc,
 	.keys_inorder		= xfs_bmbt_keys_inorder,
 	.recs_inorder		= xfs_bmbt_recs_inorder,
 };
diff --git a/fs/xfs/libxfs/xfs_bmap_btree.h b/fs/xfs/libxfs/xfs_bmap_btree.h
index 29b407d053b4..ae347f4c4b60 100644
--- a/fs/xfs/libxfs/xfs_bmap_btree.h
+++ b/fs/xfs/libxfs/xfs_bmap_btree.h
@@ -108,5 +108,7 @@ extern struct xfs_btree_cur *xfs_bmbt_init_cursor(struct xfs_mount *,
 
 extern unsigned long long xfs_bmbt_calc_size(struct xfs_mount *mp,
 		unsigned long long len);
+extern void xfs_bmbt_iroot_realloc(struct xfs_inode *ip, int rec_diff,
+		int whichfork);
 
 #endif	/* __XFS_BMAP_BTREE_H__ */
diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c
index 34c6d7bd4d18..81b2fedefb95 100644
--- a/fs/xfs/libxfs/xfs_btree.c
+++ b/fs/xfs/libxfs/xfs_btree.c
@@ -3001,9 +3001,7 @@ xfs_btree_new_iroot(
 
 	xfs_btree_copy_ptrs(cur, pp, &nptr, 1);
 
-	xfs_iroot_realloc(cur->bc_private.b.ip,
-			  1 - xfs_btree_get_numrecs(cblock),
-			  cur->bc_private.b.whichfork);
+	cur->bc_ops->iroot_realloc(cur, 1 - xfs_btree_get_numrecs(cblock));
 
 	xfs_btree_setbuf(cur, level, cbp);
 
@@ -3172,7 +3170,7 @@ xfs_btree_make_block_unfull(
 
 		if (numrecs < cur->bc_ops->get_dmaxrecs(cur, level)) {
 			/* A root block that can be made bigger. */
-			xfs_iroot_realloc(ip, 1, cur->bc_private.b.whichfork);
+			cur->bc_ops->iroot_realloc(cur, 1);
 			*stat = 1;
 		} else {
 			/* A root block that needs replacing */
@@ -3574,8 +3572,7 @@ xfs_btree_kill_iroot(
 
 	index = numrecs - cur->bc_ops->get_maxrecs(cur, level);
 	if (index) {
-		xfs_iroot_realloc(cur->bc_private.b.ip, index,
-				  cur->bc_private.b.whichfork);
+		cur->bc_ops->iroot_realloc(cur, index);
 		block = ifp->if_broot;
 	}
 
@@ -3772,8 +3769,7 @@ xfs_btree_delrec(
 	 */
 	if (level == cur->bc_nlevels - 1) {
 		if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) {
-			xfs_iroot_realloc(cur->bc_private.b.ip, -1,
-					  cur->bc_private.b.whichfork);
+			cur->bc_ops->iroot_realloc(cur, -1);
 
 			error = xfs_btree_kill_iroot(cur);
 			if (error)
diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h
index e3b3e9dce5da..59f3e42397fb 100644
--- a/fs/xfs/libxfs/xfs_btree.h
+++ b/fs/xfs/libxfs/xfs_btree.h
@@ -148,6 +148,13 @@ struct xfs_btree_ops {
 				   union xfs_btree_key *key1,
 				   union xfs_btree_key *key2);
 
+	/*
+	 * Reallocate the space for if_broot based on the number of records
+	 * being added or deleted as indicated in rec_diff.
+	 */
+	void (*iroot_realloc)(struct xfs_btree_cur *cur,
+			      int rec_diff);
+
 	const struct xfs_buf_ops	*buf_ops;
 
 	/* check that k1 is lower than k2 */
diff --git a/fs/xfs/libxfs/xfs_inode_fork.c b/fs/xfs/libxfs/xfs_inode_fork.c
index f9acf1d436f6..cfb68d4befc4 100644
--- a/fs/xfs/libxfs/xfs_inode_fork.c
+++ b/fs/xfs/libxfs/xfs_inode_fork.c
@@ -321,136 +321,6 @@ xfs_iformat_btree(
 	return 0;
 }
 
-/*
- * Reallocate the space for if_broot based on the number of records
- * being added or deleted as indicated in rec_diff.  Move the records
- * and pointers in if_broot to fit the new size.  When shrinking this
- * will eliminate holes between the records and pointers created by
- * the caller.  When growing this will create holes to be filled in
- * by the caller.
- *
- * The caller must not request to add more records than would fit in
- * the on-disk inode root.  If the if_broot is currently NULL, then
- * if we are adding records, one will be allocated.  The caller must also
- * not request that the number of records go below zero, although
- * it can go to zero.
- *
- * ip -- the inode whose if_broot area is changing
- * ext_diff -- the change in the number of records, positive or negative,
- *	 requested for the if_broot array.
- */
-void
-xfs_iroot_realloc(
-	xfs_inode_t		*ip,
-	int			rec_diff,
-	int			whichfork)
-{
-	struct xfs_mount	*mp = ip->i_mount;
-	int			cur_max;
-	struct xfs_ifork	*ifp;
-	struct xfs_btree_block	*new_broot;
-	int			new_max;
-	size_t			new_size;
-	char			*np;
-	char			*op;
-
-	/*
-	 * Handle the degenerate case quietly.
-	 */
-	if (rec_diff == 0) {
-		return;
-	}
-
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	if (rec_diff > 0) {
-		/*
-		 * If there wasn't any memory allocated before, just
-		 * allocate it now and get out.
-		 */
-		if (ifp->if_broot_bytes == 0) {
-			new_size = XFS_BMAP_BROOT_SPACE_CALC(mp, rec_diff);
-			ifp->if_broot = kmem_alloc(new_size, KM_SLEEP | KM_NOFS);
-			ifp->if_broot_bytes = (int)new_size;
-			return;
-		}
-
-		/*
-		 * If there is already an existing if_broot, then we need
-		 * to realloc() it and shift the pointers to their new
-		 * location.  The records don't change location because
-		 * they are kept butted up against the btree block header.
-		 */
-		cur_max = xfs_bmbt_maxrecs(mp, ifp->if_broot_bytes, 0);
-		new_max = cur_max + rec_diff;
-		new_size = XFS_BMAP_BROOT_SPACE_CALC(mp, new_max);
-		ifp->if_broot = kmem_realloc(ifp->if_broot, new_size,
-				KM_SLEEP | KM_NOFS);
-		op = (char *)XFS_BMAP_BROOT_PTR_ADDR(mp, ifp->if_broot, 1,
-						     ifp->if_broot_bytes);
-		np = (char *)XFS_BMAP_BROOT_PTR_ADDR(mp, ifp->if_broot, 1,
-						     (int)new_size);
-		ifp->if_broot_bytes = (int)new_size;
-		ASSERT(XFS_BMAP_BMDR_SPACE(ifp->if_broot) <=
-			XFS_IFORK_SIZE(ip, whichfork));
-		memmove(np, op, cur_max * (uint)sizeof(xfs_fsblock_t));
-		return;
-	}
-
-	/*
-	 * rec_diff is less than 0.  In this case, we are shrinking the
-	 * if_broot buffer.  It must already exist.  If we go to zero
-	 * records, just get rid of the root and clear the status bit.
-	 */
-	ASSERT((ifp->if_broot != NULL) && (ifp->if_broot_bytes > 0));
-	cur_max = xfs_bmbt_maxrecs(mp, ifp->if_broot_bytes, 0);
-	new_max = cur_max + rec_diff;
-	ASSERT(new_max >= 0);
-	if (new_max > 0)
-		new_size = XFS_BMAP_BROOT_SPACE_CALC(mp, new_max);
-	else
-		new_size = 0;
-	if (new_size > 0) {
-		new_broot = kmem_alloc(new_size, KM_SLEEP | KM_NOFS);
-		/*
-		 * First copy over the btree block header.
-		 */
-		memcpy(new_broot, ifp->if_broot,
-			XFS_BMBT_BLOCK_LEN(ip->i_mount));
-	} else {
-		new_broot = NULL;
-		ifp->if_flags &= ~XFS_IFBROOT;
-	}
-
-	/*
-	 * Only copy the records and pointers if there are any.
-	 */
-	if (new_max > 0) {
-		/*
-		 * First copy the records.
-		 */
-		op = (char *)XFS_BMBT_REC_ADDR(mp, ifp->if_broot, 1);
-		np = (char *)XFS_BMBT_REC_ADDR(mp, new_broot, 1);
-		memcpy(np, op, new_max * (uint)sizeof(xfs_bmbt_rec_t));
-
-		/*
-		 * Then copy the pointers.
-		 */
-		op = (char *)XFS_BMAP_BROOT_PTR_ADDR(mp, ifp->if_broot, 1,
-						     ifp->if_broot_bytes);
-		np = (char *)XFS_BMAP_BROOT_PTR_ADDR(mp, new_broot, 1,
-						     (int)new_size);
-		memcpy(np, op, new_max * (uint)sizeof(xfs_fsblock_t));
-	}
-	kmem_free(ifp->if_broot);
-	ifp->if_broot = new_broot;
-	ifp->if_broot_bytes = (int)new_size;
-	if (ifp->if_broot)
-		ASSERT(XFS_BMAP_BMDR_SPACE(ifp->if_broot) <=
-			XFS_IFORK_SIZE(ip, whichfork));
-	return;
-}
-
-
 /*
  * This is called when the amount of space needed for if_data
  * is increased or decreased.  The change in size is indicated by
diff --git a/fs/xfs/libxfs/xfs_inode_fork.h b/fs/xfs/libxfs/xfs_inode_fork.h
index 60361d2d74a1..8de3dc0fffb3 100644
--- a/fs/xfs/libxfs/xfs_inode_fork.h
+++ b/fs/xfs/libxfs/xfs_inode_fork.h
@@ -94,7 +94,6 @@ void		xfs_iflush_fork(struct xfs_inode *, struct xfs_dinode *,
 				struct xfs_inode_log_item *, int);
 void		xfs_idestroy_fork(struct xfs_inode *, int);
 void		xfs_idata_realloc(struct xfs_inode *, int, int);
-void		xfs_iroot_realloc(struct xfs_inode *, int, int);
 int		xfs_iread_extents(struct xfs_trans *, struct xfs_inode *, int);
 int		xfs_iextents_copy(struct xfs_inode *, struct xfs_bmbt_rec *,
 				  int);

[PATCH 02/18] xfs: support storing records in the inode core root

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Make it so that we can actually store btree records in the inode
core (i.e. enable bb_level == 0) so that the rtrmapbt can do this.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 fs/xfs/libxfs/xfs_btree.c |  196 ++++++++++++++++++++++++++++++++++-----------
 fs/xfs/libxfs/xfs_btree.h |    1 
 2 files changed, 148 insertions(+), 49 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c
index 81b2fedefb95..4ac96c3ccb8b 100644
--- a/fs/xfs/libxfs/xfs_btree.c
+++ b/fs/xfs/libxfs/xfs_btree.c
@@ -192,6 +192,11 @@ xfs_btree_check_block(
 	int			level,	/* level of the btree block */
 	struct xfs_buf		*bp)	/* buffer containing block, if any */
 {
+	/* Don't check the inode-core root. */
+	if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
+	    level == cur->bc_nlevels - 1)
+		return 0;
+
 	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
 		return xfs_btree_check_lblock(cur, block, level, bp);
 	else
@@ -1478,10 +1483,15 @@ xfs_btree_log_recs(
 	int			last)
 {
 
-	xfs_trans_buf_set_type(cur->bc_tp, bp, XFS_BLFT_BTREE_BUF);
-	xfs_trans_log_buf(cur->bc_tp, bp,
-			  xfs_btree_rec_offset(cur, first),
-			  xfs_btree_rec_offset(cur, last + 1) - 1);
+	if (bp) {
+		xfs_trans_buf_set_type(cur->bc_tp, bp, XFS_BLFT_BTREE_BUF);
+		xfs_trans_log_buf(cur->bc_tp, bp,
+				  xfs_btree_rec_offset(cur, first),
+				  xfs_btree_rec_offset(cur, last + 1) - 1);
+	} else {
+		xfs_trans_log_inode(cur->bc_tp, cur->bc_private.b.ip,
+				xfs_ilog_fbroot(cur->bc_private.b.whichfork));
+	}
 
 }
 
@@ -2935,8 +2945,11 @@ xfs_btree_new_iroot(
 	struct xfs_btree_block	*cblock;	/* child btree block */
 	union xfs_btree_key	*ckp;		/* child key pointer */
 	union xfs_btree_ptr	*cpp;		/* child ptr pointer */
+	union xfs_btree_rec	*crp;
 	union xfs_btree_key	*kp;		/* pointer to btree key */
 	union xfs_btree_ptr	*pp;		/* pointer to block addr */
+	union xfs_btree_rec	*rp;
+	union xfs_btree_ptr	aptr;
 	union xfs_btree_ptr	nptr;		/* new block addr */
 	int			level;		/* btree level */
 	int			error;		/* error return code */
@@ -2949,10 +2962,15 @@ xfs_btree_new_iroot(
 	level = cur->bc_nlevels - 1;
 
 	block = xfs_btree_get_iroot(cur);
-	pp = xfs_btree_ptr_addr(cur, 1, block);
+	ASSERT(level > 0 || (cur->bc_flags & XFS_BTREE_IROOT_RECORDS));
+	if (level > 0)
+		aptr = *xfs_btree_ptr_addr(cur, 1, block);
+	else
+		aptr.l = cpu_to_be64(XFS_INO_TO_FSB(cur->bc_mp,
+				cur->bc_private.b.ip->i_ino));
 
 	/* Allocate the new block. If we can't do it, we're toast. Give up. */
-	error = cur->bc_ops->alloc_block(cur, pp, &nptr, stat);
+	error = cur->bc_ops->alloc_block(cur, &aptr, &nptr, stat);
 	if (error)
 		goto error0;
 	if (*stat == 0)
@@ -2977,41 +2995,92 @@ xfs_btree_new_iroot(
 			cblock->bb_u.s.bb_blkno = cpu_to_be64(cbp->b_bn);
 	}
 
-	be16_add_cpu(&block->bb_level, 1);
 	xfs_btree_set_numrecs(block, 1);
 	cur->bc_nlevels++;
 	cur->bc_ptrs[level + 1] = 1;
 
-	kp = xfs_btree_key_addr(cur, 1, block);
-	ckp = xfs_btree_key_addr(cur, 1, cblock);
-	xfs_btree_copy_keys(cur, ckp, kp, xfs_btree_get_numrecs(cblock));
+	if (level > 0) {
+		/*
+		 * We already incremented nlevels, so we have to do the
+		 * same to bb_level or else pp will be calculated with the
+		 * maxrecs for regular blocks and point at the wrong place.
+		 */
+		be16_add_cpu(&block->bb_level, 1);
+
+		kp = xfs_btree_key_addr(cur, 1, block);
+		ckp = xfs_btree_key_addr(cur, 1, cblock);
+		xfs_btree_copy_keys(cur, ckp, kp,
+				xfs_btree_get_numrecs(cblock));
+
+		pp = xfs_btree_ptr_addr(cur, 1, block);
+		cpp = xfs_btree_ptr_addr(cur, 1, cblock);
 
-	cpp = xfs_btree_ptr_addr(cur, 1, cblock);
-	for (i = 0; i < be16_to_cpu(cblock->bb_numrecs); i++) {
-		error = xfs_btree_debug_check_ptr(cur, pp, i, level);
+		for (i = 0; i < be16_to_cpu(cblock->bb_numrecs); i++) {
+			error = xfs_btree_debug_check_ptr(cur, pp, i, level);
+			if (error)
+				goto error0;
+		}
+
+		xfs_btree_copy_ptrs(cur, cpp, pp,
+				xfs_btree_get_numrecs(cblock));
+
+		error = xfs_btree_debug_check_ptr(cur, &nptr, 0, level);
 		if (error)
 			goto error0;
-	}
 
-	xfs_btree_copy_ptrs(cur, cpp, pp, xfs_btree_get_numrecs(cblock));
+		xfs_btree_copy_ptrs(cur, pp, &nptr, 1);
 
-	error = xfs_btree_debug_check_ptr(cur, &nptr, 0, level);
-	if (error)
-		goto error0;
+		cur->bc_ops->iroot_realloc(cur,
+				1 - xfs_btree_get_numrecs(cblock));
+		block = xfs_btree_get_iroot(cur);
+
+		xfs_btree_setbuf(cur, level, cbp);
+
+		/*
+		 * Do all this logging at the end so that
+		 * the root is at the right level.
+		 */
+		xfs_btree_log_block(cur, cbp, XFS_BB_ALL_BITS);
+		xfs_btree_log_keys(cur, cbp, 1,
+				be16_to_cpu(cblock->bb_numrecs));
+		xfs_btree_log_ptrs(cur, cbp, 1,
+				be16_to_cpu(cblock->bb_numrecs));
+	} else {
+		rp = xfs_btree_rec_addr(cur, 1, block);
+		crp = xfs_btree_rec_addr(cur, 1, cblock);
+		xfs_btree_copy_recs(cur, crp, rp,
+				xfs_btree_get_numrecs(cblock));
 
-	xfs_btree_copy_ptrs(cur, pp, &nptr, 1);
+		/*
+		 * Trickery here: The amount of memory we need for the root
+		 * changes when we convert a leaf to a node.  Therefore,
+		 * set the length to zero, increment the level, and set
+		 * the length to 1 record.
+		 */
+		cur->bc_ops->iroot_realloc(cur, -xfs_btree_get_numrecs(cblock));
+		block = xfs_btree_get_iroot(cur);
+		be16_add_cpu(&block->bb_level, 1);
+		cur->bc_ops->iroot_realloc(cur, 1);
+		block = xfs_btree_get_iroot(cur);
 
-	cur->bc_ops->iroot_realloc(cur, 1 - xfs_btree_get_numrecs(cblock));
+		/* Copy pointer into the block. */
+		xfs_btree_copy_ptrs(cur, xfs_btree_ptr_addr(cur, 1, block),
+				&nptr, 1);
 
-	xfs_btree_setbuf(cur, level, cbp);
+		xfs_btree_setbuf(cur, level, cbp);
 
-	/*
-	 * Do all this logging at the end so that
-	 * the root is at the right level.
-	 */
-	xfs_btree_log_block(cur, cbp, XFS_BB_ALL_BITS);
-	xfs_btree_log_keys(cur, cbp, 1, be16_to_cpu(cblock->bb_numrecs));
-	xfs_btree_log_ptrs(cur, cbp, 1, be16_to_cpu(cblock->bb_numrecs));
+		/*
+		 * Do all this logging at the end so that
+		 * the root is at the right level.
+		 */
+		xfs_btree_log_block(cur, cbp, XFS_BB_ALL_BITS);
+		xfs_btree_log_recs(cur, cbp, 1,
+				be16_to_cpu(cblock->bb_numrecs));
+
+		/* Write the new keys into the root block. */
+	}
+	/* Get the keys for the new block and put them into the root. */
+	xfs_btree_get_keys(cur, cblock, xfs_btree_key_addr(cur, 1, block));
 
 	*logflags |=
 		XFS_ILOG_CORE | xfs_ilog_fbroot(cur->bc_private.b.whichfork);
@@ -3513,15 +3582,15 @@ STATIC int
 xfs_btree_kill_iroot(
 	struct xfs_btree_cur	*cur)
 {
-	int			whichfork = cur->bc_private.b.whichfork;
 	struct xfs_inode	*ip = cur->bc_private.b.ip;
-	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, whichfork);
 	struct xfs_btree_block	*block;
 	struct xfs_btree_block	*cblock;
 	union xfs_btree_key	*kp;
 	union xfs_btree_key	*ckp;
 	union xfs_btree_ptr	*pp;
 	union xfs_btree_ptr	*cpp;
+	union xfs_btree_rec	*rp;
+	union xfs_btree_rec	*crp;
 	struct xfs_buf		*cbp;
 	int			level;
 	int			index;
@@ -3533,14 +3602,19 @@ xfs_btree_kill_iroot(
 	int			i;
 
 	ASSERT(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE);
-	ASSERT(cur->bc_nlevels > 1);
+	ASSERT((cur->bc_flags & XFS_BTREE_IROOT_RECORDS) ||
+	       cur->bc_nlevels > 1);
 
 	/*
 	 * Don't deal with the root block needs to be a leaf case.
 	 * We're just going to turn the thing back into extents anyway.
 	 */
 	level = cur->bc_nlevels - 1;
-	if (level == 1)
+	if (level == 1 && !(cur->bc_flags & XFS_BTREE_IROOT_RECORDS))
+		goto out0;
+
+	/* If we're already a leaf, jump out. */
+	if (level == 0)
 		goto out0;
 
 	/*
@@ -3571,35 +3645,59 @@ xfs_btree_kill_iroot(
 #endif
 
 	index = numrecs - cur->bc_ops->get_maxrecs(cur, level);
-	if (index) {
-		cur->bc_ops->iroot_realloc(cur, index);
-		block = ifp->if_broot;
-	}
-
 	be16_add_cpu(&block->bb_numrecs, index);
 	ASSERT(block->bb_numrecs == cblock->bb_numrecs);
 
-	kp = xfs_btree_key_addr(cur, 1, block);
-	ckp = xfs_btree_key_addr(cur, 1, cblock);
-	xfs_btree_copy_keys(cur, kp, ckp, numrecs);
+	if (be16_to_cpu(cblock->bb_level) > 0) {
+		if (index) {
+			cur->bc_ops->iroot_realloc(cur, index);
+			block = xfs_btree_get_iroot(cur);
+		}
 
-	pp = xfs_btree_ptr_addr(cur, 1, block);
-	cpp = xfs_btree_ptr_addr(cur, 1, cblock);
+		kp = xfs_btree_key_addr(cur, 1, block);
+		ckp = xfs_btree_key_addr(cur, 1, cblock);
+		xfs_btree_copy_keys(cur, kp, ckp, numrecs);
 
-	for (i = 0; i < numrecs; i++) {
-		error = xfs_btree_debug_check_ptr(cur, cpp, i, level - 1);
-		if (error)
-			return error;
-	}
+		pp = xfs_btree_ptr_addr(cur, 1, block);
+		cpp = xfs_btree_ptr_addr(cur, 1, cblock);
 
-	xfs_btree_copy_ptrs(cur, pp, cpp, numrecs);
+		for (i = 0; i < numrecs; i++) {
+			error = xfs_btree_debug_check_ptr(cur, cpp, i, level - 1);
+			if (error)
+				return error;
+		}
+
+		xfs_btree_copy_ptrs(cur, pp, cpp, numrecs);
+		/*
+		 * Decrement the (root) block's level after copying the
+		 * pointers or else pp will be calculated using maxrecs
+		 * for a regular block and won't point to the right place.
+		 * Notice how we don't adjust nlevels until later.
+		 */
+		be16_add_cpu(&block->bb_level, -1);
+	} else {
+		/*
+		 * Trickery here: The amount of memory we need for the root
+		 * changes when we convert a leaf to a node.  Therefore,
+		 * set the length to zero, change the level, and set
+		 * the length to however many records we're getting.
+		 */
+		cur->bc_ops->iroot_realloc(cur, -xfs_btree_get_numrecs(block));
+		block = xfs_btree_get_iroot(cur);
+		be16_add_cpu(&block->bb_level, -1);
+		cur->bc_ops->iroot_realloc(cur, numrecs);
+		block = xfs_btree_get_iroot(cur);
+
+		rp = xfs_btree_rec_addr(cur, 1, block);
+		crp = xfs_btree_rec_addr(cur, 1, cblock);
+		xfs_btree_copy_recs(cur, rp, crp, numrecs);
+	}
 
 	error = xfs_btree_free_block(cur, cbp);
 	if (error)
 		return error;
 
 	cur->bc_bufs[level - 1] = NULL;
-	be16_add_cpu(&block->bb_level, -1);
 	xfs_trans_log_inode(cur->bc_tp, ip,
 		XFS_ILOG_CORE | xfs_ilog_fbroot(cur->bc_private.b.whichfork));
 	cur->bc_nlevels--;
diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h
index 59f3e42397fb..d20e15d14ab2 100644
--- a/fs/xfs/libxfs/xfs_btree.h
+++ b/fs/xfs/libxfs/xfs_btree.h
@@ -236,6 +236,7 @@ typedef struct xfs_btree_cur
 #define XFS_BTREE_LASTREC_UPDATE	(1<<2)	/* track last rec externally */
 #define XFS_BTREE_CRC_BLOCKS		(1<<3)	/* uses extended btree blocks */
 #define XFS_BTREE_OVERLAPPING		(1<<4)	/* overlapping intervals */
+#define XFS_BTREE_IROOT_RECORDS		(1<<5)	/* iroot can store records */
 
 
 #define	XFS_BTREE_NOERROR	0

Re: [PATCH 02/18] xfs: support storing records in the inode core root

[Dave Chinner]

On Thu, Aug 30, 2018 at 11:31:48AM -0700, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
> 
> Make it so that we can actually store btree records in the inode
> core (i.e. enable bb_level == 0) so that the rtrmapbt can do this.
> 
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>

So this is so you don't need XFS_DINODE_FMT_RMAP_EXTENTS/BTREE 
inode fork types, right? i.e. there just XFS_DINODE_FMT_RMAP to
indicate it's an rmap tree in the fork?

Could this code be used by the data fork, too? Sure, we need to
keep the on-disk fork format info the same, but we could use this to
kill the special BMDR formatting code it currently uses, right?

(just trying to get an idea of whether we'll have to maintain two
bits of code that do essentially the same thing forever)

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

Re: [PATCH 02/18] xfs: support storing records in the inode core root

[Darrick J. Wong]

On Fri, Aug 31, 2018 at 12:28:13PM +1000, Dave Chinner wrote:
> On Thu, Aug 30, 2018 at 11:31:48AM -0700, Darrick J. Wong wrote:
> > From: Darrick J. Wong <darrick.wong@oracle.com>
> > 
> > Make it so that we can actually store btree records in the inode
> > core (i.e. enable bb_level == 0) so that the rtrmapbt can do this.
> > 
> > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> 
> So this is so you don't need XFS_DINODE_FMT_RMAP_EXTENTS/BTREE 
> inode fork types, right? i.e. there just XFS_DINODE_FMT_RMAP to
> indicate it's an rmap tree in the fork?

Yep.

> Could this code be used by the data fork, too? Sure, we need to
> keep the on-disk fork format info the same, but we could use this to
> kill the special BMDR formatting code it currently uses, right?

Hmm.  FMT_EXTENTS uses di_nextents to count the number of records in the
data fork area, whereas FMT_BTREE stores the number of records in the
bmbt root in the data fork area, followed by bmbt keys/ptrs.  I imagine
you could make the formats share code, but you'd have to special case
how much btree header gets written into the data fork area.

The maximum number of bmbt records that can fit in the data fork area
(factoring in alignment requirements) actually does depend on whether or
not we have to allocate space for the root block record count.

Might be worth deeper investigation...

(FWIW, FMT_RMAP always stores the number of root block records in the
data fork area, since it doesn't affect the max record count.)

--D

> (just trying to get an idea of whether we'll have to maintain two
> bits of code that do essentially the same thing forever)
> 
> Cheers,
> 
> Dave.
> -- 
> Dave Chinner
> david@fromorbit.com

[PATCH 03/18] xfs: widen xfs_rmap_irec fields to handle realtime rmapbt

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Change the startblock and blockcount fields of xfs_rmap_irec to be 64
bits wide.  This enables us to use the same high level rmap code for
either tree.  We'll also collect all the resulting breakage fixes here.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 fs/xfs/libxfs/xfs_format.h   |    4 +-
 fs/xfs/libxfs/xfs_refcount.c |    6 +--
 fs/xfs/libxfs/xfs_rmap.c     |   83 +++++++++++++++++++++---------------------
 fs/xfs/libxfs/xfs_rmap.h     |   36 +++++++++---------
 fs/xfs/xfs_trace.h           |   38 ++++++++++---------
 5 files changed, 83 insertions(+), 84 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index d4ebf1a4f3e8..e837a8cde8af 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -1419,8 +1419,8 @@ struct xfs_rmap_rec {
 					 XFS_RMAP_BMBT_BLOCK)
 #define XFS_RMAP_REC_FLAGS		(XFS_RMAP_UNWRITTEN)
 struct xfs_rmap_irec {
-	xfs_agblock_t	rm_startblock;	/* extent start block */
-	xfs_extlen_t	rm_blockcount;	/* extent length */
+	xfs_fsblock_t	rm_startblock;	/* extent start block */
+	xfs_filblks_t	rm_blockcount;	/* extent length */
 	uint64_t	rm_owner;	/* extent owner */
 	uint64_t	rm_offset;	/* offset within the owner */
 	unsigned int	rm_flags;	/* state flags */
diff --git a/fs/xfs/libxfs/xfs_refcount.c b/fs/xfs/libxfs/xfs_refcount.c
index 542aa1475b5f..198f0c96ccae 100644
--- a/fs/xfs/libxfs/xfs_refcount.c
+++ b/fs/xfs/libxfs/xfs_refcount.c
@@ -1560,8 +1560,7 @@ xfs_refcount_alloc_cow_extent(
 		return error;
 
 	/* Add rmap entry */
-	return xfs_rmap_alloc_extent(tp, XFS_FSB_TO_AGNO(mp, fsb),
-			XFS_FSB_TO_AGBNO(mp, fsb), len, XFS_RMAP_OWN_COW);
+	return xfs_rmap_alloc_extent(tp, fsb, len, XFS_RMAP_OWN_COW);
 }
 
 /* Forget a CoW staging event in the refcount btree. */
@@ -1578,8 +1577,7 @@ xfs_refcount_free_cow_extent(
 		return 0;
 
 	/* Remove rmap entry */
-	error = xfs_rmap_free_extent(tp, XFS_FSB_TO_AGNO(mp, fsb),
-			XFS_FSB_TO_AGBNO(mp, fsb), len, XFS_RMAP_OWN_COW);
+	error = xfs_rmap_free_extent(tp, fsb, len, XFS_RMAP_OWN_COW);
 	if (error)
 		return error;
 
diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c
index 245af452840e..0372f77e019b 100644
--- a/fs/xfs/libxfs/xfs_rmap.c
+++ b/fs/xfs/libxfs/xfs_rmap.c
@@ -36,8 +36,8 @@
 int
 xfs_rmap_lookup_le(
 	struct xfs_btree_cur	*cur,
-	xfs_agblock_t		bno,
-	xfs_extlen_t		len,
+	xfs_fsblock_t		bno,
+	xfs_filblks_t		len,
 	uint64_t		owner,
 	uint64_t		offset,
 	unsigned int		flags,
@@ -58,8 +58,8 @@ xfs_rmap_lookup_le(
 int
 xfs_rmap_lookup_eq(
 	struct xfs_btree_cur	*cur,
-	xfs_agblock_t		bno,
-	xfs_extlen_t		len,
+	xfs_fsblock_t		bno,
+	xfs_filblks_t		len,
 	uint64_t		owner,
 	uint64_t		offset,
 	unsigned int		flags,
@@ -105,8 +105,8 @@ xfs_rmap_update(
 int
 xfs_rmap_insert(
 	struct xfs_btree_cur	*rcur,
-	xfs_agblock_t		agbno,
-	xfs_extlen_t		len,
+	xfs_fsblock_t		agbno,
+	xfs_filblks_t		len,
 	uint64_t		owner,
 	uint64_t		offset,
 	unsigned int		flags)
@@ -141,8 +141,8 @@ xfs_rmap_insert(
 STATIC int
 xfs_rmap_delete(
 	struct xfs_btree_cur	*rcur,
-	xfs_agblock_t		agbno,
-	xfs_extlen_t		len,
+	xfs_fsblock_t		agbno,
+	xfs_filblks_t		len,
 	uint64_t		owner,
 	uint64_t		offset,
 	unsigned int		flags)
@@ -193,10 +193,13 @@ xfs_rmap_get_rec(
 	int			*stat)
 {
 	struct xfs_mount	*mp = cur->bc_mp;
-	xfs_agnumber_t		agno = cur->bc_private.a.agno;
+	xfs_agnumber_t		agno;
 	union xfs_btree_rec	*rec;
 	int			error;
 
+	if (cur->bc_btnum != XFS_BTNUM_RMAP)
+		goto out_bad_rec;
+
 	error = xfs_btree_get_rec(cur, &rec, stat);
 	if (error || !*stat)
 		return error;
@@ -234,7 +237,7 @@ xfs_rmap_get_rec(
 		"Reverse Mapping BTree record corruption in AG %d detected!",
 		agno);
 	xfs_warn(mp,
-		"Owner 0x%llx, flags 0x%x, start block 0x%x block count 0x%x",
+		"Owner 0x%llx, flags 0x%x, start block 0x%llx block count 0x%llx",
 		irec->rm_owner, irec->rm_flags, irec->rm_startblock,
 		irec->rm_blockcount);
 	return -EFSCORRUPTED;
@@ -280,7 +283,7 @@ xfs_rmap_find_left_neighbor_helper(
 int
 xfs_rmap_find_left_neighbor(
 	struct xfs_btree_cur	*cur,
-	xfs_agblock_t		bno,
+	xfs_fsblock_t		bno,
 	uint64_t		owner,
 	uint64_t		offset,
 	unsigned int		flags,
@@ -358,7 +361,7 @@ xfs_rmap_lookup_le_range_helper(
 int
 xfs_rmap_lookup_le_range(
 	struct xfs_btree_cur	*cur,
-	xfs_agblock_t		bno,
+	xfs_fsblock_t		bno,
 	uint64_t		owner,
 	uint64_t		offset,
 	unsigned int		flags,
@@ -459,8 +462,8 @@ xfs_rmap_free_check_owner(
 STATIC int
 xfs_rmap_unmap(
 	struct xfs_btree_cur	*cur,
-	xfs_agblock_t		bno,
-	xfs_extlen_t		len,
+	xfs_fsblock_t		bno,
+	xfs_filblks_t		len,
 	bool			unwritten,
 	struct xfs_owner_info	*oinfo)
 {
@@ -607,7 +610,7 @@ xfs_rmap_unmap(
 		 * Result:  |rrrrr|         |rrrr|
 		 *               bno       len
 		 */
-		xfs_extlen_t	orig_len = ltrec.rm_blockcount;
+		xfs_filblks_t	orig_len = ltrec.rm_blockcount;
 
 		ltrec.rm_blockcount = bno - ltrec.rm_startblock;
 		error = xfs_rmap_update(cur, &ltrec);
@@ -711,8 +714,8 @@ xfs_rmap_is_mergeable(
 STATIC int
 xfs_rmap_map(
 	struct xfs_btree_cur	*cur,
-	xfs_agblock_t		bno,
-	xfs_extlen_t		len,
+	xfs_fsblock_t		bno,
+	xfs_filblks_t		len,
 	bool			unwritten,
 	struct xfs_owner_info	*oinfo)
 {
@@ -930,8 +933,8 @@ xfs_rmap_alloc(
 STATIC int
 xfs_rmap_convert(
 	struct xfs_btree_cur	*cur,
-	xfs_agblock_t		bno,
-	xfs_extlen_t		len,
+	xfs_fsblock_t		bno,
+	xfs_filblks_t		len,
 	bool			unwritten,
 	struct xfs_owner_info	*oinfo)
 {
@@ -1355,8 +1358,8 @@ xfs_rmap_convert(
 STATIC int
 xfs_rmap_convert_shared(
 	struct xfs_btree_cur	*cur,
-	xfs_agblock_t		bno,
-	xfs_extlen_t		len,
+	xfs_fsblock_t		bno,
+	xfs_filblks_t		len,
 	bool			unwritten,
 	struct xfs_owner_info	*oinfo)
 {
@@ -1744,8 +1747,8 @@ xfs_rmap_convert_shared(
 STATIC int
 xfs_rmap_unmap_shared(
 	struct xfs_btree_cur	*cur,
-	xfs_agblock_t		bno,
-	xfs_extlen_t		len,
+	xfs_fsblock_t		bno,
+	xfs_filblks_t		len,
 	bool			unwritten,
 	struct xfs_owner_info	*oinfo)
 {
@@ -1862,7 +1865,7 @@ xfs_rmap_unmap_shared(
 		 * Result:  |rrrrr|         |rrrr|
 		 *               bno       len
 		 */
-		xfs_extlen_t	orig_len = ltrec.rm_blockcount;
+		xfs_filblks_t	orig_len = ltrec.rm_blockcount;
 
 		/* Shrink the left side of the rmap */
 		error = xfs_rmap_lookup_eq(cur, ltrec.rm_startblock,
@@ -1906,8 +1909,8 @@ xfs_rmap_unmap_shared(
 STATIC int
 xfs_rmap_map_shared(
 	struct xfs_btree_cur	*cur,
-	xfs_agblock_t		bno,
-	xfs_extlen_t		len,
+	xfs_fsblock_t		bno,
+	xfs_filblks_t		len,
 	bool			unwritten,
 	struct xfs_owner_info	*oinfo)
 {
@@ -2174,7 +2177,7 @@ xfs_rmap_finish_one(
 	int				error = 0;
 	xfs_agnumber_t			agno;
 	struct xfs_owner_info		oinfo;
-	xfs_agblock_t			bno;
+	xfs_fsblock_t			bno;
 	bool				unwritten;
 
 	agno = XFS_FSB_TO_AGNO(mp, startblock);
@@ -2363,9 +2366,8 @@ xfs_rmap_convert_extent(
 int
 xfs_rmap_alloc_extent(
 	struct xfs_trans	*tp,
-	xfs_agnumber_t		agno,
-	xfs_agblock_t		bno,
-	xfs_extlen_t		len,
+	xfs_fsblock_t		fsbno,
+	xfs_filblks_t		len,
 	uint64_t		owner)
 {
 	struct xfs_bmbt_irec	bmap;
@@ -2373,7 +2375,7 @@ xfs_rmap_alloc_extent(
 	if (!xfs_rmap_update_is_needed(tp->t_mountp, XFS_DATA_FORK))
 		return 0;
 
-	bmap.br_startblock = XFS_AGB_TO_FSB(tp->t_mountp, agno, bno);
+	bmap.br_startblock = fsbno;
 	bmap.br_blockcount = len;
 	bmap.br_startoff = 0;
 	bmap.br_state = XFS_EXT_NORM;
@@ -2385,9 +2387,8 @@ xfs_rmap_alloc_extent(
 int
 xfs_rmap_free_extent(
 	struct xfs_trans	*tp,
-	xfs_agnumber_t		agno,
-	xfs_agblock_t		bno,
-	xfs_extlen_t		len,
+	xfs_fsblock_t		fsbno,
+	xfs_filblks_t		len,
 	uint64_t		owner)
 {
 	struct xfs_bmbt_irec	bmap;
@@ -2395,7 +2396,7 @@ xfs_rmap_free_extent(
 	if (!xfs_rmap_update_is_needed(tp->t_mountp, XFS_DATA_FORK))
 		return 0;
 
-	bmap.br_startblock = XFS_AGB_TO_FSB(tp->t_mountp, agno, bno);
+	bmap.br_startblock = fsbno;
 	bmap.br_blockcount = len;
 	bmap.br_startoff = 0;
 	bmap.br_state = XFS_EXT_NORM;
@@ -2435,8 +2436,8 @@ xfs_rmap_compare(
 int
 xfs_rmap_has_record(
 	struct xfs_btree_cur	*cur,
-	xfs_agblock_t		bno,
-	xfs_extlen_t		len,
+	xfs_fsblock_t		bno,
+	xfs_filblks_t		len,
 	bool			*exists)
 {
 	union xfs_btree_irec	low;
@@ -2460,8 +2461,8 @@ xfs_rmap_has_record(
 int
 xfs_rmap_record_exists(
 	struct xfs_btree_cur	*cur,
-	xfs_agblock_t		bno,
-	xfs_extlen_t		len,
+	xfs_fsblock_t		bno,
+	xfs_filblks_t		len,
 	struct xfs_owner_info	*oinfo,
 	bool			*has_rmap)
 {
@@ -2528,8 +2529,8 @@ xfs_rmap_has_other_keys_helper(
 int
 xfs_rmap_has_other_keys(
 	struct xfs_btree_cur		*cur,
-	xfs_agblock_t			bno,
-	xfs_extlen_t			len,
+	xfs_fsblock_t			bno,
+	xfs_filblks_t			len,
 	struct xfs_owner_info		*oinfo,
 	bool				*has_rmap)
 {
diff --git a/fs/xfs/libxfs/xfs_rmap.h b/fs/xfs/libxfs/xfs_rmap.h
index 157dc722ad35..f92b59c6adda 100644
--- a/fs/xfs/libxfs/xfs_rmap.h
+++ b/fs/xfs/libxfs/xfs_rmap.h
@@ -142,14 +142,14 @@ int xfs_rmap_free(struct xfs_trans *tp, struct xfs_buf *agbp,
 		  xfs_agnumber_t agno, xfs_agblock_t bno, xfs_extlen_t len,
 		  struct xfs_owner_info *oinfo);
 
-int xfs_rmap_lookup_le(struct xfs_btree_cur *cur, xfs_agblock_t bno,
-		xfs_extlen_t len, uint64_t owner, uint64_t offset,
+int xfs_rmap_lookup_le(struct xfs_btree_cur *cur, xfs_fsblock_t bno,
+		xfs_filblks_t len, uint64_t owner, uint64_t offset,
 		unsigned int flags, int *stat);
-int xfs_rmap_lookup_eq(struct xfs_btree_cur *cur, xfs_agblock_t bno,
-		xfs_extlen_t len, uint64_t owner, uint64_t offset,
+int xfs_rmap_lookup_eq(struct xfs_btree_cur *cur, xfs_fsblock_t bno,
+		xfs_filblks_t len, uint64_t owner, uint64_t offset,
 		unsigned int flags, int *stat);
-int xfs_rmap_insert(struct xfs_btree_cur *rcur, xfs_agblock_t agbno,
-		xfs_extlen_t len, uint64_t owner, uint64_t offset,
+int xfs_rmap_insert(struct xfs_btree_cur *rcur, xfs_fsblock_t agbno,
+		xfs_filblks_t len, uint64_t owner, uint64_t offset,
 		unsigned int flags);
 int xfs_rmap_get_rec(struct xfs_btree_cur *cur, struct xfs_rmap_irec *irec,
 		int *stat);
@@ -192,10 +192,10 @@ int xfs_rmap_unmap_extent(struct xfs_trans *tp, struct xfs_inode *ip,
 int xfs_rmap_convert_extent(struct xfs_mount *mp, struct xfs_trans *tp,
 		struct xfs_inode *ip, int whichfork,
 		struct xfs_bmbt_irec *imap);
-int xfs_rmap_alloc_extent(struct xfs_trans *tp, xfs_agnumber_t agno,
-		xfs_agblock_t bno, xfs_extlen_t len, uint64_t owner);
-int xfs_rmap_free_extent(struct xfs_trans *tp, xfs_agnumber_t agno,
-		xfs_agblock_t bno, xfs_extlen_t len, uint64_t owner);
+int xfs_rmap_alloc_extent(struct xfs_trans *tp, xfs_fsblock_t fsbno,
+		xfs_filblks_t len, uint64_t owner);
+int xfs_rmap_free_extent(struct xfs_trans *tp, xfs_fsblock_t fsbno,
+		xfs_filblks_t len, uint64_t owner);
 
 void xfs_rmap_finish_one_cleanup(struct xfs_trans *tp,
 		struct xfs_btree_cur *rcur, int error);
@@ -204,10 +204,10 @@ int xfs_rmap_finish_one(struct xfs_trans *tp, enum xfs_rmap_intent_type type,
 		xfs_fsblock_t startblock, xfs_filblks_t blockcount,
 		xfs_exntst_t state, struct xfs_btree_cur **pcur);
 
-int xfs_rmap_find_left_neighbor(struct xfs_btree_cur *cur, xfs_agblock_t bno,
+int xfs_rmap_find_left_neighbor(struct xfs_btree_cur *cur, xfs_fsblock_t bno,
 		uint64_t owner, uint64_t offset, unsigned int flags,
 		struct xfs_rmap_irec *irec, int	*stat);
-int xfs_rmap_lookup_le_range(struct xfs_btree_cur *cur, xfs_agblock_t bno,
+int xfs_rmap_lookup_le_range(struct xfs_btree_cur *cur, xfs_fsblock_t bno,
 		uint64_t owner, uint64_t offset, unsigned int flags,
 		struct xfs_rmap_irec *irec, int	*stat);
 int xfs_rmap_compare(const struct xfs_rmap_irec *a,
@@ -215,13 +215,13 @@ int xfs_rmap_compare(const struct xfs_rmap_irec *a,
 union xfs_btree_rec;
 int xfs_rmap_btrec_to_irec(union xfs_btree_rec *rec,
 		struct xfs_rmap_irec *irec);
-int xfs_rmap_has_record(struct xfs_btree_cur *cur, xfs_agblock_t bno,
-		xfs_extlen_t len, bool *exists);
-int xfs_rmap_record_exists(struct xfs_btree_cur *cur, xfs_agblock_t bno,
-		xfs_extlen_t len, struct xfs_owner_info *oinfo,
+int xfs_rmap_has_record(struct xfs_btree_cur *cur, xfs_fsblock_t bno,
+		xfs_filblks_t len, bool *exists);
+int xfs_rmap_record_exists(struct xfs_btree_cur *cur, xfs_fsblock_t bno,
+		xfs_filblks_t len, struct xfs_owner_info *oinfo,
 		bool *has_rmap);
-int xfs_rmap_has_other_keys(struct xfs_btree_cur *cur, xfs_agblock_t bno,
-		xfs_extlen_t len, struct xfs_owner_info *oinfo,
+int xfs_rmap_has_other_keys(struct xfs_btree_cur *cur, xfs_fsblock_t bno,
+		xfs_filblks_t len, struct xfs_owner_info *oinfo,
 		bool *has_rmap);
 int xfs_rmap_map_raw(struct xfs_btree_cur *cur, struct xfs_rmap_irec *rmap);
 
diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h
index d86f4e6a9cec..e87fa49ec2b7 100644
--- a/fs/xfs/xfs_trace.h
+++ b/fs/xfs/xfs_trace.h
@@ -2410,14 +2410,14 @@ DEFINE_BMAP_FREE_DEFERRED_EVENT(xfs_agfl_free_deferred);
 /* rmap tracepoints */
 DECLARE_EVENT_CLASS(xfs_rmap_class,
 	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
-		 xfs_agblock_t agbno, xfs_extlen_t len, bool unwritten,
+		 xfs_fsblock_t bno, xfs_filblks_t len, bool unwritten,
 		 struct xfs_owner_info *oinfo),
-	TP_ARGS(mp, agno, agbno, len, unwritten, oinfo),
+	TP_ARGS(mp, agno, bno, len, unwritten, oinfo),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
 		__field(xfs_agnumber_t, agno)
-		__field(xfs_agblock_t, agbno)
-		__field(xfs_extlen_t, len)
+		__field(xfs_fsblock_t, bno)
+		__field(xfs_filblks_t, len)
 		__field(uint64_t, owner)
 		__field(uint64_t, offset)
 		__field(unsigned long, flags)
@@ -2425,7 +2425,7 @@ DECLARE_EVENT_CLASS(xfs_rmap_class,
 	TP_fast_assign(
 		__entry->dev = mp->m_super->s_dev;
 		__entry->agno = agno;
-		__entry->agbno = agbno;
+		__entry->bno = bno;
 		__entry->len = len;
 		__entry->owner = oinfo->oi_owner;
 		__entry->offset = oinfo->oi_offset;
@@ -2433,10 +2433,10 @@ DECLARE_EVENT_CLASS(xfs_rmap_class,
 		if (unwritten)
 			__entry->flags |= XFS_RMAP_UNWRITTEN;
 	),
-	TP_printk("dev %d:%d agno %u agbno %u len %u owner %lld offset %llu flags 0x%lx",
+	TP_printk("dev %d:%d agno %d bno %llu len %llu owner %lld offset %llu flags 0x%lx",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		  __entry->agno,
-		  __entry->agbno,
+		  __entry->bno,
 		  __entry->len,
 		  __entry->owner,
 		  __entry->offset,
@@ -2445,9 +2445,9 @@ DECLARE_EVENT_CLASS(xfs_rmap_class,
 #define DEFINE_RMAP_EVENT(name) \
 DEFINE_EVENT(xfs_rmap_class, name, \
 	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \
-		 xfs_agblock_t agbno, xfs_extlen_t len, bool unwritten, \
+		 xfs_fsblock_t bno, xfs_filblks_t len, bool unwritten, \
 		 struct xfs_owner_info *oinfo), \
-	TP_ARGS(mp, agno, agbno, len, unwritten, oinfo))
+	TP_ARGS(mp, agno, bno, len, unwritten, oinfo))
 
 /* simple AG-based error/%ip tracepoint class */
 DECLARE_EVENT_CLASS(xfs_ag_error_class,
@@ -2466,7 +2466,7 @@ DECLARE_EVENT_CLASS(xfs_ag_error_class,
 		__entry->error = error;
 		__entry->caller_ip = caller_ip;
 	),
-	TP_printk("dev %d:%d agno %u error %d caller %pS",
+	TP_printk("dev %d:%d agno %d error %d caller %pS",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		  __entry->agno,
 		  __entry->error,
@@ -2492,14 +2492,14 @@ DEFINE_AG_ERROR_EVENT(xfs_rmap_convert_state);
 
 DECLARE_EVENT_CLASS(xfs_rmapbt_class,
 	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
-		 xfs_agblock_t agbno, xfs_extlen_t len,
+		 xfs_fsblock_t bno, xfs_filblks_t len,
 		 uint64_t owner, uint64_t offset, unsigned int flags),
-	TP_ARGS(mp, agno, agbno, len, owner, offset, flags),
+	TP_ARGS(mp, agno, bno, len, owner, offset, flags),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
 		__field(xfs_agnumber_t, agno)
-		__field(xfs_agblock_t, agbno)
-		__field(xfs_extlen_t, len)
+		__field(xfs_fsblock_t, bno)
+		__field(xfs_filblks_t, len)
 		__field(uint64_t, owner)
 		__field(uint64_t, offset)
 		__field(unsigned int, flags)
@@ -2507,16 +2507,16 @@ DECLARE_EVENT_CLASS(xfs_rmapbt_class,
 	TP_fast_assign(
 		__entry->dev = mp->m_super->s_dev;
 		__entry->agno = agno;
-		__entry->agbno = agbno;
+		__entry->bno = bno;
 		__entry->len = len;
 		__entry->owner = owner;
 		__entry->offset = offset;
 		__entry->flags = flags;
 	),
-	TP_printk("dev %d:%d agno %u agbno %u len %u owner %lld offset %llu flags 0x%x",
+	TP_printk("dev %d:%d agno %d bno %llu len %llu owner %lld offset %llu flags 0x%x",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		  __entry->agno,
-		  __entry->agbno,
+		  __entry->bno,
 		  __entry->len,
 		  __entry->owner,
 		  __entry->offset,
@@ -2525,9 +2525,9 @@ DECLARE_EVENT_CLASS(xfs_rmapbt_class,
 #define DEFINE_RMAPBT_EVENT(name) \
 DEFINE_EVENT(xfs_rmapbt_class, name, \
 	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \
-		 xfs_agblock_t agbno, xfs_extlen_t len, \
+		 xfs_fsblock_t bno, xfs_filblks_t len, \
 		 uint64_t owner, uint64_t offset, unsigned int flags), \
-	TP_ARGS(mp, agno, agbno, len, owner, offset, flags))
+	TP_ARGS(mp, agno, bno, len, owner, offset, flags))
 
 #define DEFINE_RMAP_DEFERRED_EVENT DEFINE_MAP_EXTENT_DEFERRED_EVENT
 DEFINE_RMAP_DEFERRED_EVENT(xfs_rmap_defer);

[PATCH 04/18] xfs: introduce realtime rmap btree definitions

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Add new realtime rmap btree definitions. The realtime rmap btree will
be rooted from a hidden inode, but has its own shape and therefore
needs to have most of its own separate types.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 fs/xfs/libxfs/xfs_btree.c  |    4 ++--
 fs/xfs/libxfs/xfs_btree.h  |    1 +
 fs/xfs/libxfs/xfs_format.h |    7 +++++++
 fs/xfs/libxfs/xfs_types.h  |    3 ++-
 4 files changed, 12 insertions(+), 3 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c
index 4ac96c3ccb8b..f77762777aa6 100644
--- a/fs/xfs/libxfs/xfs_btree.c
+++ b/fs/xfs/libxfs/xfs_btree.c
@@ -34,10 +34,10 @@ 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, 0 },
+	  XFS_FIBT_MAGIC, 0, 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_REFC_CRC_MAGIC }
+	  XFS_REFC_CRC_MAGIC, XFS_RTRMAP_CRC_MAGIC }
 };
 
 uint32_t
diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h
index d20e15d14ab2..23806757dba2 100644
--- a/fs/xfs/libxfs/xfs_btree.h
+++ b/fs/xfs/libxfs/xfs_btree.h
@@ -62,6 +62,7 @@ union xfs_btree_rec {
 #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)
+#define	XFS_BTNUM_RTRMAP	((xfs_btnum_t)XFS_BTNUM_RTRMAPi)
 
 uint32_t xfs_btree_magic(int crc, xfs_btnum_t btnum);
 
diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index e837a8cde8af..6770dac9c82d 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -1445,6 +1445,13 @@ typedef __be32 xfs_rmap_ptr_t;
 	 XFS_FIBT_BLOCK(mp) + 1 : \
 	 XFS_IBT_BLOCK(mp) + 1)
 
+/*
+ * Realtime Reverse mapping btree format definitions
+ *
+ * There is a btree for the reverse map per allocation group
+ */
+#define	XFS_RTRMAP_CRC_MAGIC	0x4d415052	/* 'MAPR' */
+
 /*
  * Reference Count Btree format definitions
  *
diff --git a/fs/xfs/libxfs/xfs_types.h b/fs/xfs/libxfs/xfs_types.h
index b9e6c89284c3..aacd52f44c56 100644
--- a/fs/xfs/libxfs/xfs_types.h
+++ b/fs/xfs/libxfs/xfs_types.h
@@ -106,7 +106,8 @@ typedef enum {
 
 typedef enum {
 	XFS_BTNUM_BNOi, XFS_BTNUM_CNTi, XFS_BTNUM_RMAPi, XFS_BTNUM_BMAPi,
-	XFS_BTNUM_INOi, XFS_BTNUM_FINOi, XFS_BTNUM_REFCi, XFS_BTNUM_MAX
+	XFS_BTNUM_INOi, XFS_BTNUM_FINOi, XFS_BTNUM_REFCi, XFS_BTNUM_RTRMAPi,
+	XFS_BTNUM_MAX
 } xfs_btnum_t;
 
 struct xfs_name {

[PATCH 05/18] xfs: define the on-disk realtime rmap btree format

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Start filling out the rtrmap btree implementation. Start with the
on-disk btree format; add everything needed to read, write and
manipulate rmap btree blocks. This prepares the way for connecting the
btree operations implementation.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 fs/xfs/Makefile                  |    1 
 fs/xfs/libxfs/xfs_btree.c        |    1 
 fs/xfs/libxfs/xfs_btree.h        |    3 +
 fs/xfs/libxfs/xfs_format.h       |   48 ++++++++++
 fs/xfs/libxfs/xfs_rmap.c         |   21 ++++
 fs/xfs/libxfs/xfs_rtrmap_btree.c |  184 ++++++++++++++++++++++++++++++++++++++
 fs/xfs/libxfs/xfs_rtrmap_btree.h |   48 ++++++++++
 fs/xfs/libxfs/xfs_sb.c           |    6 +
 fs/xfs/libxfs/xfs_shared.h       |    1 
 fs/xfs/xfs_mount.c               |    2 
 fs/xfs/xfs_mount.h               |    3 +
 fs/xfs/xfs_ondisk.h              |    2 
 12 files changed, 316 insertions(+), 4 deletions(-)
 create mode 100644 fs/xfs/libxfs/xfs_rtrmap_btree.c
 create mode 100644 fs/xfs/libxfs/xfs_rtrmap_btree.h


diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile
index 5543cadd46db..c9b452646699 100644
--- a/fs/xfs/Makefile
+++ b/fs/xfs/Makefile
@@ -51,6 +51,7 @@ xfs-y				+= $(addprefix libxfs/, \
 				   xfs_ag_resv.o \
 				   xfs_rmap.o \
 				   xfs_rmap_btree.o \
+				   xfs_rtrmap_btree.o \
 				   xfs_refcount.o \
 				   xfs_refcount_btree.o \
 				   xfs_sb.o \
diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c
index f77762777aa6..90f9ce04b64d 100644
--- a/fs/xfs/libxfs/xfs_btree.c
+++ b/fs/xfs/libxfs/xfs_btree.c
@@ -1279,6 +1279,7 @@ xfs_btree_set_refs(
 		xfs_buf_set_ref(bp, XFS_BMAP_BTREE_REF);
 		break;
 	case XFS_BTNUM_RMAP:
+	case XFS_BTNUM_RTRMAP:
 		xfs_buf_set_ref(bp, XFS_RMAP_BTREE_REF);
 		break;
 	case XFS_BTNUM_REFC:
diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h
index 23806757dba2..96b0073abfa2 100644
--- a/fs/xfs/libxfs/xfs_btree.h
+++ b/fs/xfs/libxfs/xfs_btree.h
@@ -37,6 +37,8 @@ union xfs_btree_key {
 	struct xfs_rmap_key		rmap;
 	struct xfs_rmap_key		__rmap_bigkey[2];
 	struct xfs_refcount_key		refc;
+	struct xfs_rtrmap_key		rtrmap;
+	struct xfs_rtrmap_key		__rtrmap_bigkey[2];
 };
 
 union xfs_btree_rec {
@@ -46,6 +48,7 @@ union xfs_btree_rec {
 	struct xfs_inobt_rec		inobt;
 	struct xfs_rmap_rec		rmap;
 	struct xfs_refcount_rec		refc;
+	struct xfs_rtrmap_rec		rtrmap;
 };
 
 /*
diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index 6770dac9c82d..ad767b27dbb3 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -1452,6 +1452,54 @@ typedef __be32 xfs_rmap_ptr_t;
  */
 #define	XFS_RTRMAP_CRC_MAGIC	0x4d415052	/* 'MAPR' */
 
+/*
+ * Data record structure
+ */
+struct xfs_rtrmap_rec {
+	__be64		rm_startblock;	/* extent start block */
+	__be64		rm_blockcount;	/* extent length */
+	__be64		rm_owner;	/* extent owner */
+	__be64		rm_offset;	/* offset within the owner */
+};
+
+/* rm_offset has the same values as the regular rmapbt. */
+#define XFS_RTRMAP_OFF_ATTR_FORK	XFS_RMAP_OFF_ATTR_FORK
+#define XFS_RTRMAP_OFF_BMBT_BLOCK	XFS_RMAP_OFF_BMBT_BLOCK
+#define XFS_RTRMAP_OFF_UNWRITTEN	XFS_RMAP_OFF_UNWRITTEN
+
+#define XFS_RTRMAP_LEN_MAX		((uint64_t)~0U)
+#define XFS_RTRMAP_OFF_FLAGS		XFS_RMAP_OFF_FLAGS
+#define XFS_RTRMAP_OFF_MASK		XFS_RMAP_OFF_MASK
+
+#define XFS_RTRMAP_OFF			XFS_RMAP_OFF
+
+#define XFS_RTRMAP_IS_BMBT_BLOCK(off)	XFS_RMAP_IS_BMBT_BLOCK
+#define XFS_RTRMAP_IS_ATTR_FORK(off)	XFS_RMAP_IS_ATTR_FORK
+#define XFS_RTRMAP_IS_UNWRITTEN(len)	XFS_RMAP_IS_UNWRITTEN
+
+#define RTRMAPBT_STARTBLOCK_BITLEN	64
+#define RTRMAPBT_BLOCKCOUNT_BITLEN	64
+#define RTRMAPBT_OWNER_BITLEN		RMAPBT_OWNER_BITLEN
+#define RTRMAPBT_ATTRFLAG_BITLEN	RMAPBT_ATTRFLAG_BITLEN
+#define RTRMAPBT_BMBTFLAG_BITLEN	RMAPBT_BMBTFLAG_BITLEN
+#define RTRMAPBT_EXNTFLAG_BITLEN	RMAPBT_EXNTFLAG_BITLEN
+#define RTRMAPBT_UNUSED_OFFSET_BITLEN	RMAPBT_UNUSED_OFFSET_BITLEN
+#define RTRMAPBT_OFFSET_BITLEN		RMAPBT_OFFSET_BITLEN
+
+/*
+ * Key structure
+ *
+ * We don't use the length for lookups
+ */
+struct xfs_rtrmap_key {
+	__be64		rm_startblock;	/* extent start block */
+	__be64		rm_owner;	/* extent owner */
+	__be64		rm_offset;	/* offset within the owner */
+} __packed;
+
+/* btree pointer type */
+typedef __be64 xfs_rtrmap_ptr_t;
+
 /*
  * Reference Count Btree format definitions
  *
diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c
index 0372f77e019b..5294fd3464f0 100644
--- a/fs/xfs/libxfs/xfs_rmap.c
+++ b/fs/xfs/libxfs/xfs_rmap.c
@@ -197,19 +197,32 @@ xfs_rmap_get_rec(
 	union xfs_btree_rec	*rec;
 	int			error;
 
-	if (cur->bc_btnum != XFS_BTNUM_RMAP)
-		goto out_bad_rec;
-
 	error = xfs_btree_get_rec(cur, &rec, stat);
 	if (error || !*stat)
 		return error;
 
+	if (cur->bc_btnum == XFS_BTNUM_RTRMAP)
+		agno = -1;
+	else
+		agno = cur->bc_private.a.agno;
+
 	if (xfs_rmap_btrec_to_irec(rec, irec))
 		goto out_bad_rec;
 
 	if (irec->rm_blockcount == 0)
 		goto out_bad_rec;
-	if (irec->rm_startblock <= XFS_AGFL_BLOCK(mp)) {
+	if (cur->bc_btnum == XFS_BTNUM_RTRMAP) {
+		if (!xfs_verify_rtbno(mp, irec->rm_startblock))
+			goto out_bad_rec;
+		if (irec->rm_startblock >
+				irec->rm_startblock + irec->rm_blockcount)
+			goto out_bad_rec;
+		if (!xfs_verify_rtbno(mp,
+				irec->rm_startblock + irec->rm_blockcount - 1))
+			goto out_bad_rec;
+		if (XFS_RMAP_NON_INODE_OWNER(irec->rm_owner))
+			goto out_bad_rec;
+	} else if (irec->rm_startblock <= XFS_AGFL_BLOCK(mp)) {
 		if (irec->rm_owner != XFS_RMAP_OWN_FS)
 			goto out_bad_rec;
 		if (irec->rm_blockcount != XFS_AGFL_BLOCK(mp) + 1)
diff --git a/fs/xfs/libxfs/xfs_rtrmap_btree.c b/fs/xfs/libxfs/xfs_rtrmap_btree.c
new file mode 100644
index 000000000000..62ff99278b3d
--- /dev/null
+++ b/fs/xfs/libxfs/xfs_rtrmap_btree.c
@@ -0,0 +1,184 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Copyright (C) 2018 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ */
+#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_bit.h"
+#include "xfs_sb.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_inode.h"
+#include "xfs_trans.h"
+#include "xfs_alloc.h"
+#include "xfs_btree.h"
+#include "xfs_rtrmap_btree.h"
+#include "xfs_trace.h"
+#include "xfs_cksum.h"
+#include "xfs_error.h"
+#include "xfs_extent_busy.h"
+#include "xfs_ag_resv.h"
+
+/*
+ * Realtime Reverse map btree.
+ *
+ * This is a per-ag tree used to track the owner(s) of a given extent
+ * in the realtime device.  See the comments in xfs_rmap_btree.c for
+ * more information.
+ *
+ * This tree is basically the same as the regular rmap btree except that
+ * it doesn't live in free space, and the startblock and blockcount
+ * fields have been widened to 64 bits.
+ */
+
+static struct xfs_btree_cur *
+xfs_rtrmapbt_dup_cursor(
+	struct xfs_btree_cur	*cur)
+{
+	struct xfs_btree_cur	*new;
+
+	new = xfs_rtrmapbt_init_cursor(cur->bc_mp, cur->bc_tp,
+			cur->bc_private.b.ip);
+
+	/* Copy the flags values since init cursor doesn't get them. */
+	new->bc_private.b.flags = cur->bc_private.b.flags;
+
+	return new;
+}
+
+static xfs_failaddr_t
+xfs_rtrmapbt_verify(
+	struct xfs_buf		*bp)
+{
+	struct xfs_mount	*mp = bp->b_target->bt_mount;
+	struct xfs_btree_block	*block = XFS_BUF_TO_BLOCK(bp);
+	xfs_failaddr_t		fa;
+	int			level;
+
+	if (block->bb_magic != cpu_to_be32(XFS_RTRMAP_CRC_MAGIC))
+		return __this_address;
+
+	if (!xfs_sb_version_hasrmapbt(&mp->m_sb))
+		return __this_address;
+	fa = xfs_btree_lblock_v5hdr_verify(bp, XFS_RMAP_OWN_UNKNOWN);
+	if (fa)
+		return fa;
+	level = be16_to_cpu(block->bb_level);
+	if (level > mp->m_rtrmap_maxlevels)
+		return __this_address;
+
+	return xfs_btree_lblock_verify(bp, mp->m_rtrmap_mxr[level != 0]);
+}
+
+static void
+xfs_rtrmapbt_read_verify(
+	struct xfs_buf	*bp)
+{
+	xfs_failaddr_t	fa;
+
+	if (!xfs_btree_lblock_verify_crc(bp))
+		xfs_verifier_error(bp, -EFSBADCRC, __this_address);
+	else {
+		fa = xfs_rtrmapbt_verify(bp);
+		if (fa)
+			xfs_verifier_error(bp, -EFSCORRUPTED, fa);
+	}
+
+	if (bp->b_error)
+		trace_xfs_btree_corrupt(bp, _RET_IP_);
+}
+
+static void
+xfs_rtrmapbt_write_verify(
+	struct xfs_buf	*bp)
+{
+	xfs_failaddr_t	fa;
+
+	fa = xfs_rtrmapbt_verify(bp);
+	if (fa) {
+		trace_xfs_btree_corrupt(bp, _RET_IP_);
+		xfs_verifier_error(bp, -EFSCORRUPTED, fa);
+		return;
+	}
+	xfs_btree_lblock_calc_crc(bp);
+
+}
+
+const struct xfs_buf_ops xfs_rtrmapbt_buf_ops = {
+	.name			= "xfs_rtrmapbt",
+	.verify_read		= xfs_rtrmapbt_read_verify,
+	.verify_write		= xfs_rtrmapbt_write_verify,
+	.verify_struct		= xfs_rtrmapbt_verify,
+};
+
+static const struct xfs_btree_ops xfs_rtrmapbt_ops = {
+	.rec_len		= sizeof(struct xfs_rtrmap_rec),
+	.key_len		= 2 * sizeof(struct xfs_rtrmap_key),
+
+	.dup_cursor		= xfs_rtrmapbt_dup_cursor,
+	.buf_ops		= &xfs_rtrmapbt_buf_ops,
+};
+
+/*
+ * Allocate a new allocation btree cursor.
+ */
+struct xfs_btree_cur *
+xfs_rtrmapbt_init_cursor(
+	struct xfs_mount	*mp,
+	struct xfs_trans	*tp,
+	struct xfs_inode	*ip)
+{
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
+	struct xfs_btree_cur	*cur;
+
+	cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_NOFS);
+	cur->bc_tp = tp;
+	cur->bc_mp = mp;
+	cur->bc_btnum = XFS_BTNUM_RTRMAP;
+	cur->bc_flags = XFS_BTREE_LONG_PTRS | XFS_BTREE_ROOT_IN_INODE |
+			XFS_BTREE_CRC_BLOCKS | XFS_BTREE_IROOT_RECORDS |
+			XFS_BTREE_OVERLAPPING;
+	cur->bc_blocklog = mp->m_sb.sb_blocklog;
+	cur->bc_ops = &xfs_rtrmapbt_ops;
+	cur->bc_nlevels = be16_to_cpu(ifp->if_broot->bb_level) + 1;
+	cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_rmap_2);
+
+	cur->bc_private.b.forksize = XFS_IFORK_SIZE(ip, XFS_DATA_FORK);
+	cur->bc_private.b.ip = ip;
+	cur->bc_private.b.allocated = 0;
+	cur->bc_private.b.flags = 0;
+	cur->bc_private.b.whichfork = XFS_DATA_FORK;
+
+	return cur;
+}
+
+/*
+ * Calculate number of records in an rmap btree block.
+ */
+int
+xfs_rtrmapbt_maxrecs(
+	int			blocklen,
+	bool			leaf)
+{
+	blocklen -= XFS_RTRMAP_BLOCK_LEN;
+
+	if (leaf)
+		return blocklen / sizeof(struct xfs_rtrmap_rec);
+	return blocklen /
+		(2 * sizeof(struct xfs_rtrmap_key) + sizeof(xfs_rtrmap_ptr_t));
+}
+
+/* Compute the maximum height of an rmap btree. */
+void
+xfs_rtrmapbt_compute_maxlevels(
+	struct xfs_mount		*mp)
+{
+	mp->m_rtrmap_maxlevels = xfs_btree_compute_maxlevels(mp->m_rtrmap_mnr,
+			mp->m_sb.sb_rblocks);
+	ASSERT(mp->m_rtrmap_maxlevels <= XFS_BTREE_MAXLEVELS);
+}
diff --git a/fs/xfs/libxfs/xfs_rtrmap_btree.h b/fs/xfs/libxfs/xfs_rtrmap_btree.h
new file mode 100644
index 000000000000..4b32823a2ec6
--- /dev/null
+++ b/fs/xfs/libxfs/xfs_rtrmap_btree.h
@@ -0,0 +1,48 @@
+/* SPDX-License-Identifier: GPL-2.0+ */
+/*
+ * Copyright (C) 2018 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ */
+#ifndef __XFS_RTRMAP_BTREE_H__
+#define	__XFS_RTRMAP_BTREE_H__
+
+struct xfs_buf;
+struct xfs_btree_cur;
+struct xfs_mount;
+
+/* rmaps only exist on crc enabled filesystems */
+#define XFS_RTRMAP_BLOCK_LEN	XFS_BTREE_LBLOCK_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_RTRMAP_REC_ADDR(block, index) \
+	((struct xfs_rtrmap_rec *) \
+		((char *)(block) + XFS_RTRMAP_BLOCK_LEN + \
+		 (((index) - 1) * sizeof(struct xfs_rtrmap_rec))))
+
+#define XFS_RTRMAP_KEY_ADDR(block, index) \
+	((struct xfs_rtrmap_key *) \
+		((char *)(block) + XFS_RTRMAP_BLOCK_LEN + \
+		 ((index) - 1) * 2 * sizeof(struct xfs_rtrmap_key)))
+
+#define XFS_RTRMAP_HIGH_KEY_ADDR(block, index) \
+	((struct xfs_rtrmap_key *) \
+		((char *)(block) + XFS_RTRMAP_BLOCK_LEN + \
+		 sizeof(struct xfs_rtrmap_key) + \
+		 ((index) - 1) * 2 * sizeof(struct xfs_rtrmap_key)))
+
+#define XFS_RTRMAP_PTR_ADDR(block, index, maxrecs) \
+	((xfs_rtrmap_ptr_t *) \
+		((char *)(block) + XFS_RTRMAP_BLOCK_LEN + \
+		 (maxrecs) * 2 * sizeof(struct xfs_rtrmap_key) + \
+		 ((index) - 1) * sizeof(xfs_rtrmap_ptr_t)))
+
+struct xfs_btree_cur *xfs_rtrmapbt_init_cursor(struct xfs_mount *mp,
+				struct xfs_trans *tp, struct xfs_inode *ip);
+int xfs_rtrmapbt_maxrecs(int blocklen, bool leaf);
+void xfs_rtrmapbt_compute_maxlevels(struct xfs_mount *mp);
+
+#endif	/* __XFS_RTRMAP_BTREE_H__ */
diff --git a/fs/xfs/libxfs/xfs_sb.c b/fs/xfs/libxfs/xfs_sb.c
index 081f46e30556..733bb7e5409c 100644
--- a/fs/xfs/libxfs/xfs_sb.c
+++ b/fs/xfs/libxfs/xfs_sb.c
@@ -30,6 +30,7 @@
 #include "xfs_refcount_btree.h"
 #include "xfs_da_format.h"
 #include "xfs_da_btree.h"
+#include "xfs_rtrmap_btree.h"
 
 /*
  * Physical superblock buffer manipulations. Shared with libxfs in userspace.
@@ -834,6 +835,11 @@ 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_rtrmap_mxr[0] = xfs_rtrmapbt_maxrecs(sbp->sb_blocksize, true);
+	mp->m_rtrmap_mxr[1] = xfs_rtrmapbt_maxrecs(sbp->sb_blocksize, false);
+	mp->m_rtrmap_mnr[0] = mp->m_rtrmap_mxr[0] / 2;
+	mp->m_rtrmap_mnr[1] = mp->m_rtrmap_mxr[1] / 2;
+
 	mp->m_refc_mxr[0] = xfs_refcountbt_maxrecs(sbp->sb_blocksize, true);
 	mp->m_refc_mxr[1] = xfs_refcountbt_maxrecs(sbp->sb_blocksize, false);
 	mp->m_refc_mnr[0] = mp->m_refc_mxr[0] / 2;
diff --git a/fs/xfs/libxfs/xfs_shared.h b/fs/xfs/libxfs/xfs_shared.h
index 1c5debe748f0..f34e2740d9dd 100644
--- a/fs/xfs/libxfs/xfs_shared.h
+++ b/fs/xfs/libxfs/xfs_shared.h
@@ -27,6 +27,7 @@ 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_rtrmapbt_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;
diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c
index 611f8e188170..bd70a67ec3c6 100644
--- a/fs/xfs/xfs_mount.c
+++ b/fs/xfs/xfs_mount.c
@@ -34,6 +34,7 @@
 #include "xfs_refcount_btree.h"
 #include "xfs_reflink.h"
 #include "xfs_extent_busy.h"
+#include "xfs_rtrmap_btree.h"
 
 
 static DEFINE_MUTEX(xfs_uuid_table_mutex);
@@ -744,6 +745,7 @@ xfs_mountfs(
 	xfs_bmap_compute_maxlevels(mp, XFS_ATTR_FORK);
 	xfs_ialloc_compute_maxlevels(mp);
 	xfs_rmapbt_compute_maxlevels(mp);
+	xfs_rtrmapbt_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 c12e59d3425d..5a4df6b860e2 100644
--- a/fs/xfs/xfs_mount.h
+++ b/fs/xfs/xfs_mount.h
@@ -112,12 +112,15 @@ 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_rtrmap_mxr[2]; /* max rtrmap btree records */
+	uint			m_rtrmap_mnr[2]; /* min rtrmap 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_rtrmap_maxlevels; /* max rtrmap btree level */
 	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 */
diff --git a/fs/xfs/xfs_ondisk.h b/fs/xfs/xfs_ondisk.h
index d3e04d20d8d4..23c9b6f047c5 100644
--- a/fs/xfs/xfs_ondisk.h
+++ b/fs/xfs/xfs_ondisk.h
@@ -41,6 +41,8 @@ xfs_check_ondisk_structs(void)
 	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_rtrmap_key,		24);
+	XFS_CHECK_STRUCT_SIZE(struct xfs_rtrmap_rec,		32);
 	XFS_CHECK_STRUCT_SIZE(struct xfs_timestamp,		8);
 	XFS_CHECK_STRUCT_SIZE(xfs_alloc_key_t,			8);
 	XFS_CHECK_STRUCT_SIZE(xfs_alloc_ptr_t,			4);

[PATCH 06/18] xfs: realtime rmap btree transaction reservations

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Make sure that there's enough log reservation to handle mapping
and unmapping realtime extents.  We have to reserve enough space
to handle a split in the rtrmapbt to add the record and a second
split in the regular rmapbt to record the rtrmapbt split.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 fs/xfs/libxfs/xfs_trans_resv.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)


diff --git a/fs/xfs/libxfs/xfs_trans_resv.c b/fs/xfs/libxfs/xfs_trans_resv.c
index f99a7aefe418..33922646ec65 100644
--- a/fs/xfs/libxfs/xfs_trans_resv.c
+++ b/fs/xfs/libxfs/xfs_trans_resv.c
@@ -74,7 +74,8 @@ 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);
+		blocks += max(num_ops * (2 * mp->m_rmap_maxlevels - 1),
+			      num_ops * (2 * mp->m_rtrmap_maxlevels - 1));
 	if (xfs_sb_version_hasreflink(&mp->m_sb))
 		blocks += num_ops * (2 * mp->m_refc_maxlevels - 1);
 

[PATCH 07/18] xfs: add realtime rmap btree operations

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Implement the generic btree operations needed to manipulate rtrmap
btree blocks. This is different from the regular rmapbt in that we
allocate space from the filesystem at large, and are neither
constrained to the free space nor any particular AG.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 fs/xfs/libxfs/xfs_rtrmap_btree.c |  314 ++++++++++++++++++++++++++++++++++++++
 1 file changed, 314 insertions(+)


diff --git a/fs/xfs/libxfs/xfs_rtrmap_btree.c b/fs/xfs/libxfs/xfs_rtrmap_btree.c
index 62ff99278b3d..4529dbc6545c 100644
--- a/fs/xfs/libxfs/xfs_rtrmap_btree.c
+++ b/fs/xfs/libxfs/xfs_rtrmap_btree.c
@@ -17,12 +17,14 @@
 #include "xfs_trans.h"
 #include "xfs_alloc.h"
 #include "xfs_btree.h"
+#include "xfs_rmap.h"
 #include "xfs_rtrmap_btree.h"
 #include "xfs_trace.h"
 #include "xfs_cksum.h"
 #include "xfs_error.h"
 #include "xfs_extent_busy.h"
 #include "xfs_ag_resv.h"
+#include "xfs_bmap.h"
 
 /*
  * Realtime Reverse map btree.
@@ -51,6 +53,270 @@ xfs_rtrmapbt_dup_cursor(
 	return new;
 }
 
+STATIC int
+xfs_rtrmapbt_alloc_block(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*start,
+	union xfs_btree_ptr	*new,
+	int			*stat)
+{
+	struct xfs_alloc_arg	args;
+	int			error;
+
+	memset(&args, 0, sizeof(args));
+	args.tp = cur->bc_tp;
+	args.mp = cur->bc_mp;
+	args.fsbno = cur->bc_tp->t_firstblock;
+	xfs_rmap_ino_bmbt_owner(&args.oinfo, cur->bc_private.b.ip->i_ino,
+			cur->bc_private.b.whichfork);
+
+	if (args.fsbno == NULLFSBLOCK) {
+		args.fsbno = be64_to_cpu(start->l);
+		args.type = XFS_ALLOCTYPE_START_BNO;
+		/*
+		 * Make sure there is sufficient room left in the AG to
+		 * complete a full tree split for an extent insert.  If
+		 * we are converting the middle part of an extent then
+		 * we may need space for two tree splits.
+		 *
+		 * We are relying on the caller to make the correct block
+		 * reservation for this operation to succeed.  If the
+		 * reservation amount is insufficient then we may fail a
+		 * block allocation here and corrupt the filesystem.
+		 */
+		args.minleft = args.tp->t_blk_res;
+	} else if (cur->bc_tp->t_flags & XFS_TRANS_LOWMODE) {
+		args.type = XFS_ALLOCTYPE_START_BNO;
+	} else {
+		args.type = XFS_ALLOCTYPE_NEAR_BNO;
+	}
+
+	args.minlen = args.maxlen = args.prod = 1;
+	args.wasdel = 0;
+	error = xfs_alloc_vextent(&args);
+	if (error)
+		goto error0;
+
+	if (args.fsbno == NULLFSBLOCK && args.minleft) {
+		/*
+		 * Could not find an AG with enough free space to satisfy
+		 * a full btree split.  Try again without minleft and if
+		 * successful activate the lowspace algorithm.
+		 */
+		args.fsbno = 0;
+		args.type = XFS_ALLOCTYPE_FIRST_AG;
+		args.minleft = 0;
+		error = xfs_alloc_vextent(&args);
+		if (error)
+			goto error0;
+		cur->bc_tp->t_flags |= XFS_TRANS_LOWMODE;
+	}
+	if (args.fsbno == NULLFSBLOCK) {
+		*stat = 0;
+		return 0;
+	}
+	ASSERT(args.len == 1);
+	cur->bc_private.b.allocated++;
+	cur->bc_private.b.ip->i_d.di_nblocks++;
+	xfs_trans_log_inode(args.tp, cur->bc_private.b.ip, XFS_ILOG_CORE);
+
+	new->l = cpu_to_be64(args.fsbno);
+
+	*stat = 1;
+	return 0;
+
+ error0:
+	return error;
+}
+
+STATIC int
+xfs_rtrmapbt_free_block(
+	struct xfs_btree_cur	*cur,
+	struct xfs_buf		*bp)
+{
+	struct xfs_mount	*mp = cur->bc_mp;
+	struct xfs_inode	*ip = cur->bc_private.b.ip;
+	struct xfs_trans	*tp = cur->bc_tp;
+	xfs_fsblock_t		fsbno = XFS_DADDR_TO_FSB(mp, XFS_BUF_ADDR(bp));
+	struct xfs_owner_info	oinfo;
+
+	xfs_rmap_ino_bmbt_owner(&oinfo, ip->i_ino, cur->bc_private.b.whichfork);
+	xfs_bmap_add_free(cur->bc_tp, fsbno, 1, &oinfo);
+	ip->i_d.di_nblocks--;
+
+	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+	return 0;
+}
+
+/*
+ * Calculate number of records in the in-core realtime rmap btree inode root.
+ */
+STATIC int
+xfs_rtrmapbt_broot_maxrecs(
+	int			blocklen,
+	bool			leaf)
+{
+	blocklen -= XFS_RTRMAP_BLOCK_LEN;
+
+	if (leaf)
+		return blocklen / sizeof(struct xfs_rtrmap_rec);
+	return blocklen / (2 * sizeof(struct xfs_rtrmap_key) +
+			sizeof(xfs_rtrmap_ptr_t));
+}
+
+STATIC int
+xfs_rtrmapbt_get_minrecs(
+	struct xfs_btree_cur	*cur,
+	int			level)
+{
+	struct xfs_ifork	*ifp;
+
+	if (level == cur->bc_nlevels - 1) {
+		ifp = XFS_IFORK_PTR(cur->bc_private.b.ip,
+				    cur->bc_private.b.whichfork);
+
+		return xfs_rtrmapbt_broot_maxrecs(ifp->if_broot_bytes,
+				level == 0) / 2;
+	}
+
+	return cur->bc_mp->m_rtrmap_mnr[level != 0];
+}
+
+STATIC int
+xfs_rtrmapbt_get_maxrecs(
+	struct xfs_btree_cur	*cur,
+	int			level)
+{
+	struct xfs_ifork	*ifp;
+
+	if (level == cur->bc_nlevels - 1) {
+		ifp = XFS_IFORK_PTR(cur->bc_private.b.ip,
+				    cur->bc_private.b.whichfork);
+
+		return xfs_rtrmapbt_broot_maxrecs(ifp->if_broot_bytes,
+				level == 0);
+	}
+
+	return cur->bc_mp->m_rtrmap_mxr[level != 0];
+}
+
+STATIC void
+xfs_rtrmapbt_init_key_from_rec(
+	union xfs_btree_key	*key,
+	union xfs_btree_rec	*rec)
+{
+	key->rtrmap.rm_startblock = rec->rtrmap.rm_startblock;
+	key->rtrmap.rm_owner = rec->rtrmap.rm_owner;
+	key->rtrmap.rm_offset = rec->rtrmap.rm_offset;
+}
+
+STATIC void
+xfs_rtrmapbt_init_high_key_from_rec(
+	union xfs_btree_key	*key,
+	union xfs_btree_rec	*rec)
+{
+	uint64_t		off;
+	int			adj;
+
+	adj = be64_to_cpu(rec->rtrmap.rm_blockcount) - 1;
+
+	key->rtrmap.rm_startblock = rec->rtrmap.rm_startblock;
+	be64_add_cpu(&key->rtrmap.rm_startblock, adj);
+	key->rtrmap.rm_owner = rec->rtrmap.rm_owner;
+	key->rtrmap.rm_offset = rec->rtrmap.rm_offset;
+	if (XFS_RMAP_NON_INODE_OWNER(be64_to_cpu(rec->rtrmap.rm_owner)) ||
+	    XFS_RMAP_IS_BMBT_BLOCK(be64_to_cpu(rec->rtrmap.rm_offset)))
+		return;
+	off = be64_to_cpu(key->rtrmap.rm_offset);
+	off = (XFS_RMAP_OFF(off) + adj) | (off & ~XFS_RMAP_OFF_MASK);
+	key->rtrmap.rm_offset = cpu_to_be64(off);
+}
+
+STATIC void
+xfs_rtrmapbt_init_rec_from_cur(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*rec)
+{
+	rec->rtrmap.rm_startblock = cpu_to_be64(cur->bc_rec.r.rm_startblock);
+	rec->rtrmap.rm_blockcount = cpu_to_be64(cur->bc_rec.r.rm_blockcount);
+	rec->rtrmap.rm_owner = cpu_to_be64(cur->bc_rec.r.rm_owner);
+	rec->rtrmap.rm_offset = cpu_to_be64(
+			xfs_rmap_irec_offset_pack(&cur->bc_rec.r));
+}
+
+STATIC void
+xfs_rtrmapbt_init_ptr_from_cur(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr)
+{
+	ptr->l = 0;
+}
+
+STATIC int64_t
+xfs_rtrmapbt_key_diff(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_key	*key)
+{
+	struct xfs_rmap_irec	*rec = &cur->bc_rec.r;
+	struct xfs_rtrmap_key	*kp = &key->rtrmap;
+	__u64			x, y;
+
+	x = be64_to_cpu(kp->rm_startblock);
+	y = rec->rm_startblock;
+	if (x > y)
+		return 1;
+	else if (y > x)
+		return -1;
+
+	x = be64_to_cpu(kp->rm_owner);
+	y = rec->rm_owner;
+	if (x > y)
+		return 1;
+	else if (y > x)
+		return -1;
+
+	x = XFS_RMAP_OFF(be64_to_cpu(kp->rm_offset));
+	y = rec->rm_offset;
+	if (x > y)
+		return 1;
+	else if (y > x)
+		return -1;
+	return 0;
+}
+
+STATIC int64_t
+xfs_rtrmapbt_diff_two_keys(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_key	*k1,
+	union xfs_btree_key	*k2)
+{
+	struct xfs_rtrmap_key	*kp1 = &k1->rtrmap;
+	struct xfs_rtrmap_key	*kp2 = &k2->rtrmap;
+	__u64			x, y;
+
+	x = be64_to_cpu(kp1->rm_startblock);
+	y = be64_to_cpu(kp2->rm_startblock);
+	if (x > y)
+		return 1;
+	else if (y > x)
+		return -1;
+
+	x = be64_to_cpu(kp1->rm_owner);
+	y = be64_to_cpu(kp2->rm_owner);
+	if (x > y)
+		return 1;
+	else if (y > x)
+		return -1;
+
+	x = XFS_RMAP_OFF(be64_to_cpu(kp1->rm_offset));
+	y = XFS_RMAP_OFF(be64_to_cpu(kp2->rm_offset));
+	if (x > y)
+		return 1;
+	else if (y > x)
+		return -1;
+	return 0;
+}
+
 static xfs_failaddr_t
 xfs_rtrmapbt_verify(
 	struct xfs_buf		*bp)
@@ -116,12 +382,60 @@ const struct xfs_buf_ops xfs_rtrmapbt_buf_ops = {
 	.verify_struct		= xfs_rtrmapbt_verify,
 };
 
+STATIC int
+xfs_rtrmapbt_keys_inorder(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_key	*k1,
+	union xfs_btree_key	*k2)
+{
+	if (be64_to_cpu(k1->rtrmap.rm_startblock) <
+	    be64_to_cpu(k2->rtrmap.rm_startblock))
+		return 1;
+	if (be64_to_cpu(k1->rtrmap.rm_owner) <
+	    be64_to_cpu(k2->rtrmap.rm_owner))
+		return 1;
+	if (XFS_RMAP_OFF(be64_to_cpu(k1->rtrmap.rm_offset)) <=
+	    XFS_RMAP_OFF(be64_to_cpu(k2->rtrmap.rm_offset)))
+		return 1;
+	return 0;
+}
+
+STATIC int
+xfs_rtrmapbt_recs_inorder(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*r1,
+	union xfs_btree_rec	*r2)
+{
+	if (be64_to_cpu(r1->rtrmap.rm_startblock) <
+	    be64_to_cpu(r2->rtrmap.rm_startblock))
+		return 1;
+	if (XFS_RMAP_OFF(be64_to_cpu(r1->rtrmap.rm_offset)) <
+	    XFS_RMAP_OFF(be64_to_cpu(r2->rtrmap.rm_offset)))
+		return 1;
+	if (be64_to_cpu(r1->rtrmap.rm_owner) <=
+	    be64_to_cpu(r2->rtrmap.rm_owner))
+		return 1;
+	return 0;
+}
+
 static const struct xfs_btree_ops xfs_rtrmapbt_ops = {
 	.rec_len		= sizeof(struct xfs_rtrmap_rec),
 	.key_len		= 2 * sizeof(struct xfs_rtrmap_key),
 
 	.dup_cursor		= xfs_rtrmapbt_dup_cursor,
+	.alloc_block		= xfs_rtrmapbt_alloc_block,
+	.free_block		= xfs_rtrmapbt_free_block,
+	.get_minrecs		= xfs_rtrmapbt_get_minrecs,
+	.get_maxrecs		= xfs_rtrmapbt_get_maxrecs,
+	.init_key_from_rec	= xfs_rtrmapbt_init_key_from_rec,
+	.init_high_key_from_rec	= xfs_rtrmapbt_init_high_key_from_rec,
+	.init_rec_from_cur	= xfs_rtrmapbt_init_rec_from_cur,
+	.init_ptr_from_cur	= xfs_rtrmapbt_init_ptr_from_cur,
+	.key_diff		= xfs_rtrmapbt_key_diff,
 	.buf_ops		= &xfs_rtrmapbt_buf_ops,
+	.diff_two_keys		= xfs_rtrmapbt_diff_two_keys,
+	.keys_inorder		= xfs_rtrmapbt_keys_inorder,
+	.recs_inorder		= xfs_rtrmapbt_recs_inorder,
 };
 
 /*

[PATCH 08/18] xfs: prepare rmap functions to deal with rtrmapbt

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Prepare the high-level rmap functions to deal with the new realtime
rmapbt and its slightly different conventions.  Provide the ability
to talk to either rmapbt or rtrmapbt formats from the same high
level code.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 fs/xfs/libxfs/xfs_rmap.c |  178 +++++++++++++++++++++++++++-------------------
 fs/xfs/libxfs/xfs_rmap.h |    2 -
 fs/xfs/scrub/rmap.c      |    2 -
 3 files changed, 108 insertions(+), 74 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c
index 5294fd3464f0..c75d3ba34613 100644
--- a/fs/xfs/libxfs/xfs_rmap.c
+++ b/fs/xfs/libxfs/xfs_rmap.c
@@ -29,6 +29,24 @@
 #include "xfs_inode.h"
 #include "xfs_ialloc.h"
 
+/* By convention, the rtrmapbt's "AG" number is NULLAGNUMBER. */
+static xfs_agnumber_t
+xfs_rmap_cur_agno(
+	struct xfs_btree_cur	*cur)
+{
+	return (cur->bc_flags & XFS_BTREE_LONG_PTRS) ?
+			NULLAGNUMBER : cur->bc_private.a.agno;
+}
+
+/* Return the maximum length of an rmap record. */
+static xfs_filblks_t
+xfs_rmap_len_max(
+	struct xfs_btree_cur	*cur)
+{
+	return (cur->bc_flags & XFS_BTREE_LONG_PTRS) ?
+			XFS_RTRMAP_LEN_MAX : XFS_RMAP_LEN_MAX;
+}
+
 /*
  * Lookup the first record less than or equal to [bno, len, owner, offset]
  * in the btree given by cur.
@@ -86,19 +104,27 @@ xfs_rmap_update(
 	union xfs_btree_rec	rec;
 	int			error;
 
-	trace_xfs_rmap_update(cur->bc_mp, cur->bc_private.a.agno,
+	trace_xfs_rmap_update(cur->bc_mp, xfs_rmap_cur_agno(cur),
 			irec->rm_startblock, irec->rm_blockcount,
 			irec->rm_owner, irec->rm_offset, irec->rm_flags);
 
-	rec.rmap.rm_startblock = cpu_to_be32(irec->rm_startblock);
-	rec.rmap.rm_blockcount = cpu_to_be32(irec->rm_blockcount);
-	rec.rmap.rm_owner = cpu_to_be64(irec->rm_owner);
-	rec.rmap.rm_offset = cpu_to_be64(
-			xfs_rmap_irec_offset_pack(irec));
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
+		rec.rtrmap.rm_startblock = cpu_to_be64(irec->rm_startblock);
+		rec.rtrmap.rm_blockcount = cpu_to_be64(irec->rm_blockcount);
+		rec.rtrmap.rm_owner = cpu_to_be64(irec->rm_owner);
+		rec.rtrmap.rm_offset = cpu_to_be64(
+				xfs_rmap_irec_offset_pack(irec));
+	} else {
+		rec.rmap.rm_startblock = cpu_to_be32(irec->rm_startblock);
+		rec.rmap.rm_blockcount = cpu_to_be32(irec->rm_blockcount);
+		rec.rmap.rm_owner = cpu_to_be64(irec->rm_owner);
+		rec.rmap.rm_offset = cpu_to_be64(
+				xfs_rmap_irec_offset_pack(irec));
+	}
 	error = xfs_btree_update(cur, &rec);
 	if (error)
 		trace_xfs_rmap_update_error(cur->bc_mp,
-				cur->bc_private.a.agno, error, _RET_IP_);
+				xfs_rmap_cur_agno(cur), error, _RET_IP_);
 	return error;
 }
 
@@ -114,7 +140,7 @@ xfs_rmap_insert(
 	int			i;
 	int			error;
 
-	trace_xfs_rmap_insert(rcur->bc_mp, rcur->bc_private.a.agno, agbno,
+	trace_xfs_rmap_insert(rcur->bc_mp, xfs_rmap_cur_agno(rcur), agbno,
 			len, owner, offset, flags);
 
 	error = xfs_rmap_lookup_eq(rcur, agbno, len, owner, offset, flags, &i);
@@ -134,7 +160,7 @@ xfs_rmap_insert(
 done:
 	if (error)
 		trace_xfs_rmap_insert_error(rcur->bc_mp,
-				rcur->bc_private.a.agno, error, _RET_IP_);
+				xfs_rmap_cur_agno(rcur), error, _RET_IP_);
 	return error;
 }
 
@@ -150,7 +176,7 @@ xfs_rmap_delete(
 	int			i;
 	int			error;
 
-	trace_xfs_rmap_delete(rcur->bc_mp, rcur->bc_private.a.agno, agbno,
+	trace_xfs_rmap_delete(rcur->bc_mp, xfs_rmap_cur_agno(rcur), agbno,
 			len, owner, offset, flags);
 
 	error = xfs_rmap_lookup_eq(rcur, agbno, len, owner, offset, flags, &i);
@@ -165,22 +191,31 @@ xfs_rmap_delete(
 done:
 	if (error)
 		trace_xfs_rmap_delete_error(rcur->bc_mp,
-				rcur->bc_private.a.agno, error, _RET_IP_);
+				xfs_rmap_cur_agno(rcur), error, _RET_IP_);
 	return error;
 }
 
 /* Convert an internal btree record to an rmap record. */
 int
 xfs_rmap_btrec_to_irec(
+	struct xfs_btree_cur	*cur,
 	union xfs_btree_rec	*rec,
 	struct xfs_rmap_irec	*irec)
 {
 	irec->rm_flags = 0;
-	irec->rm_startblock = be32_to_cpu(rec->rmap.rm_startblock);
-	irec->rm_blockcount = be32_to_cpu(rec->rmap.rm_blockcount);
-	irec->rm_owner = be64_to_cpu(rec->rmap.rm_owner);
-	return xfs_rmap_irec_offset_unpack(be64_to_cpu(rec->rmap.rm_offset),
-			irec);
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
+		irec->rm_startblock = be64_to_cpu(rec->rtrmap.rm_startblock);
+		irec->rm_blockcount = be64_to_cpu(rec->rtrmap.rm_blockcount);
+		irec->rm_owner = be64_to_cpu(rec->rtrmap.rm_owner);
+		return xfs_rmap_irec_offset_unpack(
+				be64_to_cpu(rec->rtrmap.rm_offset), irec);
+	} else {
+		irec->rm_startblock = be32_to_cpu(rec->rmap.rm_startblock);
+		irec->rm_blockcount = be32_to_cpu(rec->rmap.rm_blockcount);
+		irec->rm_owner = be64_to_cpu(rec->rmap.rm_owner);
+		return xfs_rmap_irec_offset_unpack(
+				be64_to_cpu(rec->rmap.rm_offset), irec);
+	}
 }
 
 /*
@@ -206,7 +241,7 @@ xfs_rmap_get_rec(
 	else
 		agno = cur->bc_private.a.agno;
 
-	if (xfs_rmap_btrec_to_irec(rec, irec))
+	if (xfs_rmap_btrec_to_irec(cur, rec, irec))
 		goto out_bad_rec;
 
 	if (irec->rm_blockcount == 0)
@@ -272,7 +307,7 @@ xfs_rmap_find_left_neighbor_helper(
 	struct xfs_find_left_neighbor_info	*info = priv;
 
 	trace_xfs_rmap_find_left_neighbor_candidate(cur->bc_mp,
-			cur->bc_private.a.agno, rec->rm_startblock,
+			xfs_rmap_cur_agno(cur), rec->rm_startblock,
 			rec->rm_blockcount, rec->rm_owner, rec->rm_offset,
 			rec->rm_flags);
 
@@ -324,7 +359,7 @@ xfs_rmap_find_left_neighbor(
 	info.stat = stat;
 
 	trace_xfs_rmap_find_left_neighbor_query(cur->bc_mp,
-			cur->bc_private.a.agno, bno, 0, owner, offset, flags);
+			xfs_rmap_cur_agno(cur), bno, 0, owner, offset, flags);
 
 	error = xfs_rmap_query_range(cur, &info.high, &info.high,
 			xfs_rmap_find_left_neighbor_helper, &info);
@@ -332,7 +367,7 @@ xfs_rmap_find_left_neighbor(
 		error = 0;
 	if (*stat)
 		trace_xfs_rmap_find_left_neighbor_result(cur->bc_mp,
-				cur->bc_private.a.agno, irec->rm_startblock,
+				xfs_rmap_cur_agno(cur), irec->rm_startblock,
 				irec->rm_blockcount, irec->rm_owner,
 				irec->rm_offset, irec->rm_flags);
 	return error;
@@ -348,7 +383,7 @@ xfs_rmap_lookup_le_range_helper(
 	struct xfs_find_left_neighbor_info	*info = priv;
 
 	trace_xfs_rmap_lookup_le_range_candidate(cur->bc_mp,
-			cur->bc_private.a.agno, rec->rm_startblock,
+			xfs_rmap_cur_agno(cur), rec->rm_startblock,
 			rec->rm_blockcount, rec->rm_owner, rec->rm_offset,
 			rec->rm_flags);
 
@@ -397,14 +432,14 @@ xfs_rmap_lookup_le_range(
 	info.stat = stat;
 
 	trace_xfs_rmap_lookup_le_range(cur->bc_mp,
-			cur->bc_private.a.agno, bno, 0, owner, offset, flags);
+			xfs_rmap_cur_agno(cur), bno, 0, owner, offset, flags);
 	error = xfs_rmap_query_range(cur, &info.high, &info.high,
 			xfs_rmap_lookup_le_range_helper, &info);
 	if (error == XFS_BTREE_QUERY_RANGE_ABORT)
 		error = 0;
 	if (*stat)
 		trace_xfs_rmap_lookup_le_range_result(cur->bc_mp,
-				cur->bc_private.a.agno, irec->rm_startblock,
+				xfs_rmap_cur_agno(cur), irec->rm_startblock,
 				irec->rm_blockcount, irec->rm_owner,
 				irec->rm_offset, irec->rm_flags);
 	return error;
@@ -495,7 +530,7 @@ xfs_rmap_unmap(
 			(flags & XFS_RMAP_BMBT_BLOCK);
 	if (unwritten)
 		flags |= XFS_RMAP_UNWRITTEN;
-	trace_xfs_rmap_unmap(mp, cur->bc_private.a.agno, bno, len,
+	trace_xfs_rmap_unmap(mp, xfs_rmap_cur_agno(cur), bno, len,
 			unwritten, oinfo);
 
 	/*
@@ -513,7 +548,7 @@ xfs_rmap_unmap(
 		goto out_error;
 	XFS_WANT_CORRUPTED_GOTO(mp, i == 1, out_error);
 	trace_xfs_rmap_lookup_le_range_result(cur->bc_mp,
-			cur->bc_private.a.agno, ltrec.rm_startblock,
+			xfs_rmap_cur_agno(cur), ltrec.rm_startblock,
 			ltrec.rm_blockcount, ltrec.rm_owner,
 			ltrec.rm_offset, ltrec.rm_flags);
 	ltoff = ltrec.rm_offset;
@@ -568,7 +603,7 @@ xfs_rmap_unmap(
 
 	if (ltrec.rm_startblock == bno && ltrec.rm_blockcount == len) {
 		/* exact match, simply remove the record from rmap tree */
-		trace_xfs_rmap_delete(mp, cur->bc_private.a.agno,
+		trace_xfs_rmap_delete(mp, xfs_rmap_cur_agno(cur),
 				ltrec.rm_startblock, ltrec.rm_blockcount,
 				ltrec.rm_owner, ltrec.rm_offset,
 				ltrec.rm_flags);
@@ -643,7 +678,7 @@ xfs_rmap_unmap(
 		else
 			cur->bc_rec.r.rm_offset = offset + len;
 		cur->bc_rec.r.rm_flags = flags;
-		trace_xfs_rmap_insert(mp, cur->bc_private.a.agno,
+		trace_xfs_rmap_insert(mp, xfs_rmap_cur_agno(cur),
 				cur->bc_rec.r.rm_startblock,
 				cur->bc_rec.r.rm_blockcount,
 				cur->bc_rec.r.rm_owner,
@@ -655,11 +690,11 @@ xfs_rmap_unmap(
 	}
 
 out_done:
-	trace_xfs_rmap_unmap_done(mp, cur->bc_private.a.agno, bno, len,
+	trace_xfs_rmap_unmap_done(mp, xfs_rmap_cur_agno(cur), bno, len,
 			unwritten, oinfo);
 out_error:
 	if (error)
-		trace_xfs_rmap_unmap_error(mp, cur->bc_private.a.agno,
+		trace_xfs_rmap_unmap_error(mp, xfs_rmap_cur_agno(cur),
 				error, _RET_IP_);
 	return error;
 }
@@ -750,7 +785,7 @@ xfs_rmap_map(
 			(flags & XFS_RMAP_BMBT_BLOCK);
 	if (unwritten)
 		flags |= XFS_RMAP_UNWRITTEN;
-	trace_xfs_rmap_map(mp, cur->bc_private.a.agno, bno, len,
+	trace_xfs_rmap_map(mp, xfs_rmap_cur_agno(cur), bno, len,
 			unwritten, oinfo);
 	ASSERT(!xfs_rmap_should_skip_owner_update(oinfo));
 
@@ -769,7 +804,7 @@ xfs_rmap_map(
 			goto out_error;
 		XFS_WANT_CORRUPTED_GOTO(mp, have_lt == 1, out_error);
 		trace_xfs_rmap_lookup_le_range_result(cur->bc_mp,
-				cur->bc_private.a.agno, ltrec.rm_startblock,
+				xfs_rmap_cur_agno(cur), ltrec.rm_startblock,
 				ltrec.rm_blockcount, ltrec.rm_owner,
 				ltrec.rm_offset, ltrec.rm_flags);
 
@@ -797,7 +832,7 @@ xfs_rmap_map(
 		XFS_WANT_CORRUPTED_GOTO(mp, bno + len <= gtrec.rm_startblock,
 					out_error);
 		trace_xfs_rmap_find_right_neighbor_result(cur->bc_mp,
-			cur->bc_private.a.agno, gtrec.rm_startblock,
+			xfs_rmap_cur_agno(cur), gtrec.rm_startblock,
 			gtrec.rm_blockcount, gtrec.rm_owner,
 			gtrec.rm_offset, gtrec.rm_flags);
 		if (!xfs_rmap_is_mergeable(&gtrec, owner, flags))
@@ -824,8 +859,8 @@ xfs_rmap_map(
 		if (have_gt &&
 		    bno + len == gtrec.rm_startblock &&
 		    (ignore_off || offset + len == gtrec.rm_offset) &&
-		    (unsigned long)ltrec.rm_blockcount + len +
-				gtrec.rm_blockcount <= XFS_RMAP_LEN_MAX) {
+		    ltrec.rm_blockcount + len + gtrec.rm_blockcount <=
+		    xfs_rmap_len_max(cur)) {
 			/*
 			 * right edge also contiguous, delete right record
 			 * and merge into left record.
@@ -836,7 +871,7 @@ xfs_rmap_map(
 			 * result: |rrrrrrrrrrrrrrrrrrrrrrrrrrrrr|
 			 */
 			ltrec.rm_blockcount += gtrec.rm_blockcount;
-			trace_xfs_rmap_delete(mp, cur->bc_private.a.agno,
+			trace_xfs_rmap_delete(mp, xfs_rmap_cur_agno(cur),
 					gtrec.rm_startblock,
 					gtrec.rm_blockcount,
 					gtrec.rm_owner,
@@ -884,7 +919,7 @@ xfs_rmap_map(
 		cur->bc_rec.r.rm_owner = owner;
 		cur->bc_rec.r.rm_offset = offset;
 		cur->bc_rec.r.rm_flags = flags;
-		trace_xfs_rmap_insert(mp, cur->bc_private.a.agno, bno, len,
+		trace_xfs_rmap_insert(mp, xfs_rmap_cur_agno(cur), bno, len,
 			owner, offset, flags);
 		error = xfs_btree_insert(cur, &i);
 		if (error)
@@ -892,11 +927,11 @@ xfs_rmap_map(
 		XFS_WANT_CORRUPTED_GOTO(mp, i == 1, out_error);
 	}
 
-	trace_xfs_rmap_map_done(mp, cur->bc_private.a.agno, bno, len,
+	trace_xfs_rmap_map_done(mp, xfs_rmap_cur_agno(cur), bno, len,
 			unwritten, oinfo);
 out_error:
 	if (error)
-		trace_xfs_rmap_map_error(mp, cur->bc_private.a.agno,
+		trace_xfs_rmap_map_error(mp, xfs_rmap_cur_agno(cur),
 				error, _RET_IP_);
 	return error;
 }
@@ -970,7 +1005,7 @@ xfs_rmap_convert(
 			(flags & (XFS_RMAP_ATTR_FORK | XFS_RMAP_BMBT_BLOCK))));
 	oldext = unwritten ? XFS_RMAP_UNWRITTEN : 0;
 	new_endoff = offset + len;
-	trace_xfs_rmap_convert(mp, cur->bc_private.a.agno, bno, len,
+	trace_xfs_rmap_convert(mp, xfs_rmap_cur_agno(cur), bno, len,
 			unwritten, oinfo);
 
 	/*
@@ -988,7 +1023,7 @@ xfs_rmap_convert(
 		goto done;
 	XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
 	trace_xfs_rmap_lookup_le_range_result(cur->bc_mp,
-			cur->bc_private.a.agno, PREV.rm_startblock,
+			xfs_rmap_cur_agno(cur), PREV.rm_startblock,
 			PREV.rm_blockcount, PREV.rm_owner,
 			PREV.rm_offset, PREV.rm_flags);
 
@@ -1024,7 +1059,7 @@ xfs_rmap_convert(
 				LEFT.rm_startblock + LEFT.rm_blockcount <= bno,
 				done);
 		trace_xfs_rmap_find_left_neighbor_result(cur->bc_mp,
-				cur->bc_private.a.agno, LEFT.rm_startblock,
+				xfs_rmap_cur_agno(cur), LEFT.rm_startblock,
 				LEFT.rm_blockcount, LEFT.rm_owner,
 				LEFT.rm_offset, LEFT.rm_flags);
 		if (LEFT.rm_startblock + LEFT.rm_blockcount == bno &&
@@ -1054,7 +1089,7 @@ xfs_rmap_convert(
 		XFS_WANT_CORRUPTED_GOTO(mp, bno + len <= RIGHT.rm_startblock,
 					done);
 		trace_xfs_rmap_find_right_neighbor_result(cur->bc_mp,
-				cur->bc_private.a.agno, RIGHT.rm_startblock,
+				xfs_rmap_cur_agno(cur), RIGHT.rm_startblock,
 				RIGHT.rm_blockcount, RIGHT.rm_owner,
 				RIGHT.rm_offset, RIGHT.rm_flags);
 		if (bno + len == RIGHT.rm_startblock &&
@@ -1068,11 +1103,11 @@ xfs_rmap_convert(
 			 RMAP_RIGHT_FILLING | RMAP_RIGHT_CONTIG)) ==
 	    (RMAP_LEFT_FILLING | RMAP_LEFT_CONTIG |
 	     RMAP_RIGHT_FILLING | RMAP_RIGHT_CONTIG) &&
-	    (unsigned long)LEFT.rm_blockcount + len +
-	     RIGHT.rm_blockcount > XFS_RMAP_LEN_MAX)
+	    LEFT.rm_blockcount + len + RIGHT.rm_blockcount >
+	    xfs_rmap_len_max(cur))
 		state &= ~RMAP_RIGHT_CONTIG;
 
-	trace_xfs_rmap_convert_state(mp, cur->bc_private.a.agno, state,
+	trace_xfs_rmap_convert_state(mp, xfs_rmap_cur_agno(cur), state,
 			_RET_IP_);
 
 	/* reset the cursor back to PREV */
@@ -1096,7 +1131,7 @@ xfs_rmap_convert(
 		if (error)
 			goto done;
 		XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
-		trace_xfs_rmap_delete(mp, cur->bc_private.a.agno,
+		trace_xfs_rmap_delete(mp, xfs_rmap_cur_agno(cur),
 				RIGHT.rm_startblock, RIGHT.rm_blockcount,
 				RIGHT.rm_owner, RIGHT.rm_offset,
 				RIGHT.rm_flags);
@@ -1108,7 +1143,7 @@ xfs_rmap_convert(
 		if (error)
 			goto done;
 		XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
-		trace_xfs_rmap_delete(mp, cur->bc_private.a.agno,
+		trace_xfs_rmap_delete(mp, xfs_rmap_cur_agno(cur),
 				PREV.rm_startblock, PREV.rm_blockcount,
 				PREV.rm_owner, PREV.rm_offset,
 				PREV.rm_flags);
@@ -1132,7 +1167,7 @@ xfs_rmap_convert(
 		 * Setting all of a previous oldext extent to newext.
 		 * The left neighbor is contiguous, the right is not.
 		 */
-		trace_xfs_rmap_delete(mp, cur->bc_private.a.agno,
+		trace_xfs_rmap_delete(mp, xfs_rmap_cur_agno(cur),
 				PREV.rm_startblock, PREV.rm_blockcount,
 				PREV.rm_owner, PREV.rm_offset,
 				PREV.rm_flags);
@@ -1160,7 +1195,7 @@ xfs_rmap_convert(
 		if (error)
 			goto done;
 		XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
-		trace_xfs_rmap_delete(mp, cur->bc_private.a.agno,
+		trace_xfs_rmap_delete(mp, xfs_rmap_cur_agno(cur),
 				RIGHT.rm_startblock, RIGHT.rm_blockcount,
 				RIGHT.rm_owner, RIGHT.rm_offset,
 				RIGHT.rm_flags);
@@ -1233,7 +1268,7 @@ xfs_rmap_convert(
 		NEW.rm_blockcount = len;
 		NEW.rm_flags = newext;
 		cur->bc_rec.r = NEW;
-		trace_xfs_rmap_insert(mp, cur->bc_private.a.agno, bno,
+		trace_xfs_rmap_insert(mp, xfs_rmap_cur_agno(cur), bno,
 				len, owner, offset, newext);
 		error = xfs_btree_insert(cur, &i);
 		if (error)
@@ -1284,7 +1319,7 @@ xfs_rmap_convert(
 		NEW.rm_blockcount = len;
 		NEW.rm_flags = newext;
 		cur->bc_rec.r = NEW;
-		trace_xfs_rmap_insert(mp, cur->bc_private.a.agno, bno,
+		trace_xfs_rmap_insert(mp, xfs_rmap_cur_agno(cur), bno,
 				len, owner, offset, newext);
 		error = xfs_btree_insert(cur, &i);
 		if (error)
@@ -1312,7 +1347,7 @@ xfs_rmap_convert(
 		NEW = PREV;
 		NEW.rm_blockcount = offset - PREV.rm_offset;
 		cur->bc_rec.r = NEW;
-		trace_xfs_rmap_insert(mp, cur->bc_private.a.agno,
+		trace_xfs_rmap_insert(mp, xfs_rmap_cur_agno(cur),
 				NEW.rm_startblock, NEW.rm_blockcount,
 				NEW.rm_owner, NEW.rm_offset,
 				NEW.rm_flags);
@@ -1333,7 +1368,7 @@ xfs_rmap_convert(
 		/* new middle extent - newext */
 		cur->bc_rec.r.rm_flags &= ~XFS_RMAP_UNWRITTEN;
 		cur->bc_rec.r.rm_flags |= newext;
-		trace_xfs_rmap_insert(mp, cur->bc_private.a.agno, bno, len,
+		trace_xfs_rmap_insert(mp, xfs_rmap_cur_agno(cur), bno, len,
 				owner, offset, newext);
 		error = xfs_btree_insert(cur, &i);
 		if (error)
@@ -1354,12 +1389,12 @@ xfs_rmap_convert(
 		ASSERT(0);
 	}
 
-	trace_xfs_rmap_convert_done(mp, cur->bc_private.a.agno, bno, len,
+	trace_xfs_rmap_convert_done(mp, xfs_rmap_cur_agno(cur), bno, len,
 			unwritten, oinfo);
 done:
 	if (error)
 		trace_xfs_rmap_convert_error(cur->bc_mp,
-				cur->bc_private.a.agno, error, _RET_IP_);
+				xfs_rmap_cur_agno(cur), error, _RET_IP_);
 	return error;
 }
 
@@ -1395,7 +1430,7 @@ xfs_rmap_convert_shared(
 			(flags & (XFS_RMAP_ATTR_FORK | XFS_RMAP_BMBT_BLOCK))));
 	oldext = unwritten ? XFS_RMAP_UNWRITTEN : 0;
 	new_endoff = offset + len;
-	trace_xfs_rmap_convert(mp, cur->bc_private.a.agno, bno, len,
+	trace_xfs_rmap_convert(mp, xfs_rmap_cur_agno(cur), bno, len,
 			unwritten, oinfo);
 
 	/*
@@ -1451,7 +1486,7 @@ xfs_rmap_convert_shared(
 		XFS_WANT_CORRUPTED_GOTO(mp, bno + len <= RIGHT.rm_startblock,
 				done);
 		trace_xfs_rmap_find_right_neighbor_result(cur->bc_mp,
-				cur->bc_private.a.agno, RIGHT.rm_startblock,
+				xfs_rmap_cur_agno(cur), RIGHT.rm_startblock,
 				RIGHT.rm_blockcount, RIGHT.rm_owner,
 				RIGHT.rm_offset, RIGHT.rm_flags);
 		if (xfs_rmap_is_mergeable(&RIGHT, owner, newext))
@@ -1463,11 +1498,11 @@ xfs_rmap_convert_shared(
 			 RMAP_RIGHT_FILLING | RMAP_RIGHT_CONTIG)) ==
 	    (RMAP_LEFT_FILLING | RMAP_LEFT_CONTIG |
 	     RMAP_RIGHT_FILLING | RMAP_RIGHT_CONTIG) &&
-	    (unsigned long)LEFT.rm_blockcount + len +
-	     RIGHT.rm_blockcount > XFS_RMAP_LEN_MAX)
+	    LEFT.rm_blockcount + len + RIGHT.rm_blockcount >
+	    xfs_rmap_len_max(cur))
 		state &= ~RMAP_RIGHT_CONTIG;
 
-	trace_xfs_rmap_convert_state(mp, cur->bc_private.a.agno, state,
+	trace_xfs_rmap_convert_state(mp, xfs_rmap_cur_agno(cur), state,
 			_RET_IP_);
 	/*
 	 * Switch out based on the FILLING and CONTIG state bits.
@@ -1734,12 +1769,12 @@ xfs_rmap_convert_shared(
 		ASSERT(0);
 	}
 
-	trace_xfs_rmap_convert_done(mp, cur->bc_private.a.agno, bno, len,
+	trace_xfs_rmap_convert_done(mp, xfs_rmap_cur_agno(cur), bno, len,
 			unwritten, oinfo);
 done:
 	if (error)
 		trace_xfs_rmap_convert_error(cur->bc_mp,
-				cur->bc_private.a.agno, error, _RET_IP_);
+				xfs_rmap_cur_agno(cur), error, _RET_IP_);
 	return error;
 }
 
@@ -1777,7 +1812,7 @@ xfs_rmap_unmap_shared(
 	xfs_owner_info_unpack(oinfo, &owner, &offset, &flags);
 	if (unwritten)
 		flags |= XFS_RMAP_UNWRITTEN;
-	trace_xfs_rmap_unmap(mp, cur->bc_private.a.agno, bno, len,
+	trace_xfs_rmap_unmap(mp, xfs_rmap_cur_agno(cur), bno, len,
 			unwritten, oinfo);
 
 	/*
@@ -1901,12 +1936,12 @@ xfs_rmap_unmap_shared(
 			goto out_error;
 	}
 
-	trace_xfs_rmap_unmap_done(mp, cur->bc_private.a.agno, bno, len,
+	trace_xfs_rmap_unmap_done(mp, xfs_rmap_cur_agno(cur), bno, len,
 			unwritten, oinfo);
 out_error:
 	if (error)
 		trace_xfs_rmap_unmap_error(cur->bc_mp,
-				cur->bc_private.a.agno, error, _RET_IP_);
+				xfs_rmap_cur_agno(cur), error, _RET_IP_);
 	return error;
 }
 
@@ -1941,7 +1976,7 @@ xfs_rmap_map_shared(
 	xfs_owner_info_unpack(oinfo, &owner, &offset, &flags);
 	if (unwritten)
 		flags |= XFS_RMAP_UNWRITTEN;
-	trace_xfs_rmap_map(mp, cur->bc_private.a.agno, bno, len,
+	trace_xfs_rmap_map(mp, xfs_rmap_cur_agno(cur), bno, len,
 			unwritten, oinfo);
 
 	/* Is there a left record that abuts our range? */
@@ -1964,7 +1999,7 @@ xfs_rmap_map_shared(
 			goto out_error;
 		XFS_WANT_CORRUPTED_GOTO(mp, have_gt == 1, out_error);
 		trace_xfs_rmap_find_right_neighbor_result(cur->bc_mp,
-			cur->bc_private.a.agno, gtrec.rm_startblock,
+			xfs_rmap_cur_agno(cur), gtrec.rm_startblock,
 			gtrec.rm_blockcount, gtrec.rm_owner,
 			gtrec.rm_offset, gtrec.rm_flags);
 
@@ -2054,12 +2089,12 @@ xfs_rmap_map_shared(
 			goto out_error;
 	}
 
-	trace_xfs_rmap_map_done(mp, cur->bc_private.a.agno, bno, len,
+	trace_xfs_rmap_map_done(mp, xfs_rmap_cur_agno(cur), bno, len,
 			unwritten, oinfo);
 out_error:
 	if (error)
 		trace_xfs_rmap_map_error(cur->bc_mp,
-				cur->bc_private.a.agno, error, _RET_IP_);
+				xfs_rmap_cur_agno(cur), error, _RET_IP_);
 	return error;
 }
 
@@ -2107,7 +2142,7 @@ xfs_rmap_query_range_helper(
 	struct xfs_rmap_irec			irec;
 	int					error;
 
-	error = xfs_rmap_btrec_to_irec(rec, &irec);
+	error = xfs_rmap_btrec_to_irec(cur, rec, &irec);
 	if (error)
 		return error;
 	return query->fn(cur, &irec, query->priv);
@@ -2203,13 +2238,12 @@ xfs_rmap_finish_one(
 	if (XFS_TEST_ERROR(false, mp,
 			XFS_ERRTAG_RMAP_FINISH_ONE))
 		return -EIO;
-
 	/*
 	 * If we haven't gotten a cursor or the cursor AG doesn't match
 	 * the startblock, get one now.
 	 */
 	rcur = *pcur;
-	if (rcur != NULL && rcur->bc_private.a.agno != agno) {
+	if (rcur != NULL && xfs_rmap_cur_agno(rcur) != agno) {
 		xfs_rmap_finish_one_cleanup(tp, rcur, 0);
 		rcur = NULL;
 		*pcur = NULL;
diff --git a/fs/xfs/libxfs/xfs_rmap.h b/fs/xfs/libxfs/xfs_rmap.h
index f92b59c6adda..74f02830edb3 100644
--- a/fs/xfs/libxfs/xfs_rmap.h
+++ b/fs/xfs/libxfs/xfs_rmap.h
@@ -213,7 +213,7 @@ int xfs_rmap_lookup_le_range(struct xfs_btree_cur *cur, xfs_fsblock_t bno,
 int xfs_rmap_compare(const struct xfs_rmap_irec *a,
 		const struct xfs_rmap_irec *b);
 union xfs_btree_rec;
-int xfs_rmap_btrec_to_irec(union xfs_btree_rec *rec,
+int xfs_rmap_btrec_to_irec(struct xfs_btree_cur *cur, union xfs_btree_rec *rec,
 		struct xfs_rmap_irec *irec);
 int xfs_rmap_has_record(struct xfs_btree_cur *cur, xfs_fsblock_t bno,
 		xfs_filblks_t len, bool *exists);
diff --git a/fs/xfs/scrub/rmap.c b/fs/xfs/scrub/rmap.c
index c6a1c59ef25e..32dc329e5411 100644
--- a/fs/xfs/scrub/rmap.c
+++ b/fs/xfs/scrub/rmap.c
@@ -112,7 +112,7 @@ xchk_rmapbt_rec(
 	bool			is_attr;
 	int			error;
 
-	error = xfs_rmap_btrec_to_irec(rec, &irec);
+	error = xfs_rmap_btrec_to_irec(bs->cur, rec, &irec);
 	if (!xchk_btree_process_error(bs->sc, bs->cur, 0, &error))
 		goto out;
 

[PATCH 09/18] xfs: add a realtime flag to the rmap update log redo items

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Extend the rmap update (RUI) log items with a new realtime flag that
indicates that the updates apply against the realtime rmapbt.  We'll
wire up the actual rmap code later.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 fs/xfs/libxfs/xfs_log_format.h |    4 +++-
 fs/xfs/libxfs/xfs_refcount.c   |    4 ++--
 fs/xfs/libxfs/xfs_rmap.c       |   24 +++++++++++++++---------
 fs/xfs/libxfs/xfs_rmap.h       |    5 +++--
 fs/xfs/xfs_rmap_item.c         |    4 +++-
 fs/xfs/xfs_trans.h             |    2 +-
 fs/xfs/xfs_trans_rmap.c        |    9 +++++++--
 7 files changed, 34 insertions(+), 18 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_log_format.h b/fs/xfs/libxfs/xfs_log_format.h
index e5f97c69b320..c8184a0d9290 100644
--- a/fs/xfs/libxfs/xfs_log_format.h
+++ b/fs/xfs/libxfs/xfs_log_format.h
@@ -637,11 +637,13 @@ struct xfs_map_extent {
 #define XFS_RMAP_EXTENT_ATTR_FORK	(1U << 31)
 #define XFS_RMAP_EXTENT_BMBT_BLOCK	(1U << 30)
 #define XFS_RMAP_EXTENT_UNWRITTEN	(1U << 29)
+#define XFS_RMAP_EXTENT_REALTIME	(1U << 28)
 
 #define XFS_RMAP_EXTENT_FLAGS		(XFS_RMAP_EXTENT_TYPE_MASK | \
 					 XFS_RMAP_EXTENT_ATTR_FORK | \
 					 XFS_RMAP_EXTENT_BMBT_BLOCK | \
-					 XFS_RMAP_EXTENT_UNWRITTEN)
+					 XFS_RMAP_EXTENT_UNWRITTEN | \
+					 XFS_RMAP_EXTENT_REALTIME)
 
 /*
  * This is the structure used to lay out an rui log item in the
diff --git a/fs/xfs/libxfs/xfs_refcount.c b/fs/xfs/libxfs/xfs_refcount.c
index 198f0c96ccae..9e07350159bd 100644
--- a/fs/xfs/libxfs/xfs_refcount.c
+++ b/fs/xfs/libxfs/xfs_refcount.c
@@ -1560,7 +1560,7 @@ xfs_refcount_alloc_cow_extent(
 		return error;
 
 	/* Add rmap entry */
-	return xfs_rmap_alloc_extent(tp, fsb, len, XFS_RMAP_OWN_COW);
+	return xfs_rmap_alloc_extent(tp, fsb, len, XFS_RMAP_OWN_COW, false);
 }
 
 /* Forget a CoW staging event in the refcount btree. */
@@ -1577,7 +1577,7 @@ xfs_refcount_free_cow_extent(
 		return 0;
 
 	/* Remove rmap entry */
-	error = xfs_rmap_free_extent(tp, fsb, len, XFS_RMAP_OWN_COW);
+	error = xfs_rmap_free_extent(tp, fsb, len, XFS_RMAP_OWN_COW, false);
 	if (error)
 		return error;
 
diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c
index c75d3ba34613..377d4c31d69f 100644
--- a/fs/xfs/libxfs/xfs_rmap.c
+++ b/fs/xfs/libxfs/xfs_rmap.c
@@ -2331,11 +2331,12 @@ __xfs_rmap_add(
 	enum xfs_rmap_intent_type	type,
 	uint64_t			owner,
 	int				whichfork,
-	struct xfs_bmbt_irec		*bmap)
+	struct xfs_bmbt_irec		*bmap,
+	bool				realtime)
 {
 	struct xfs_rmap_intent		*ri;
 
-	trace_xfs_rmap_defer(tp->t_mountp,
+	trace_xfs_rmap_defer(tp->t_mountp, realtime ? NULLAGNUMBER :
 			XFS_FSB_TO_AGNO(tp->t_mountp, bmap->br_startblock),
 			type,
 			XFS_FSB_TO_AGBNO(tp->t_mountp, bmap->br_startblock),
@@ -2350,6 +2351,7 @@ __xfs_rmap_add(
 	ri->ri_owner = owner;
 	ri->ri_whichfork = whichfork;
 	ri->ri_bmap = *bmap;
+	ri->ri_realtime = realtime;
 
 	xfs_defer_add(tp, XFS_DEFER_OPS_TYPE_RMAP, &ri->ri_list);
 	return 0;
@@ -2368,7 +2370,7 @@ xfs_rmap_map_extent(
 
 	return __xfs_rmap_add(tp, xfs_is_reflink_inode(ip) ?
 			XFS_RMAP_MAP_SHARED : XFS_RMAP_MAP, ip->i_ino,
-			whichfork, PREV);
+			whichfork, PREV, XFS_IS_REALTIME_INODE(ip));
 }
 
 /* Unmap an extent out of a file. */
@@ -2384,7 +2386,7 @@ xfs_rmap_unmap_extent(
 
 	return __xfs_rmap_add(tp, xfs_is_reflink_inode(ip) ?
 			XFS_RMAP_UNMAP_SHARED : XFS_RMAP_UNMAP, ip->i_ino,
-			whichfork, PREV);
+			whichfork, PREV, XFS_IS_REALTIME_INODE(ip));
 }
 
 /*
@@ -2406,7 +2408,7 @@ xfs_rmap_convert_extent(
 
 	return __xfs_rmap_add(tp, xfs_is_reflink_inode(ip) ?
 			XFS_RMAP_CONVERT_SHARED : XFS_RMAP_CONVERT, ip->i_ino,
-			whichfork, PREV);
+			whichfork, PREV, XFS_IS_REALTIME_INODE(ip));
 }
 
 /* Schedule the creation of an rmap for non-file data. */
@@ -2415,7 +2417,8 @@ xfs_rmap_alloc_extent(
 	struct xfs_trans	*tp,
 	xfs_fsblock_t		fsbno,
 	xfs_filblks_t		len,
-	uint64_t		owner)
+	uint64_t		owner,
+	bool			isrt)
 {
 	struct xfs_bmbt_irec	bmap;
 
@@ -2427,7 +2430,8 @@ xfs_rmap_alloc_extent(
 	bmap.br_startoff = 0;
 	bmap.br_state = XFS_EXT_NORM;
 
-	return __xfs_rmap_add(tp, XFS_RMAP_ALLOC, owner, XFS_DATA_FORK, &bmap);
+	return __xfs_rmap_add(tp, XFS_RMAP_ALLOC, owner, XFS_DATA_FORK, &bmap,
+			isrt);
 }
 
 /* Schedule the deletion of an rmap for non-file data. */
@@ -2436,7 +2440,8 @@ xfs_rmap_free_extent(
 	struct xfs_trans	*tp,
 	xfs_fsblock_t		fsbno,
 	xfs_filblks_t		len,
-	uint64_t		owner)
+	uint64_t		owner,
+	bool			isrt)
 {
 	struct xfs_bmbt_irec	bmap;
 
@@ -2448,7 +2453,8 @@ xfs_rmap_free_extent(
 	bmap.br_startoff = 0;
 	bmap.br_state = XFS_EXT_NORM;
 
-	return __xfs_rmap_add(tp, XFS_RMAP_FREE, owner, XFS_DATA_FORK, &bmap);
+	return __xfs_rmap_add(tp, XFS_RMAP_FREE, owner, XFS_DATA_FORK, &bmap,
+			isrt);
 }
 
 /* Compare rmap records.  Returns -1 if a < b, 1 if a > b, and 0 if equal. */
diff --git a/fs/xfs/libxfs/xfs_rmap.h b/fs/xfs/libxfs/xfs_rmap.h
index 74f02830edb3..a4d3146e5619 100644
--- a/fs/xfs/libxfs/xfs_rmap.h
+++ b/fs/xfs/libxfs/xfs_rmap.h
@@ -182,6 +182,7 @@ struct xfs_rmap_intent {
 	uint64_t				ri_owner;
 	int					ri_whichfork;
 	struct xfs_bmbt_irec			ri_bmap;
+	bool					ri_realtime;
 };
 
 /* functions for updating the rmapbt based on bmbt map/unmap operations */
@@ -193,9 +194,9 @@ int xfs_rmap_convert_extent(struct xfs_mount *mp, struct xfs_trans *tp,
 		struct xfs_inode *ip, int whichfork,
 		struct xfs_bmbt_irec *imap);
 int xfs_rmap_alloc_extent(struct xfs_trans *tp, xfs_fsblock_t fsbno,
-		xfs_filblks_t len, uint64_t owner);
+		xfs_filblks_t len, uint64_t owner, bool isrt);
 int xfs_rmap_free_extent(struct xfs_trans *tp, xfs_fsblock_t fsbno,
-		xfs_filblks_t len, uint64_t owner);
+		xfs_filblks_t len, uint64_t owner, bool isrt);
 
 void xfs_rmap_finish_one_cleanup(struct xfs_trans *tp,
 		struct xfs_btree_cur *rcur, int error);
diff --git a/fs/xfs/xfs_rmap_item.c b/fs/xfs/xfs_rmap_item.c
index 127dc9c32a54..2fe98a300d0e 100644
--- a/fs/xfs/xfs_rmap_item.c
+++ b/fs/xfs/xfs_rmap_item.c
@@ -415,6 +415,7 @@ xfs_rui_recover(
 	xfs_exntst_t			state;
 	struct xfs_trans		*tp;
 	struct xfs_btree_cur		*rcur = NULL;
+	bool				rt;
 
 	ASSERT(!test_bit(XFS_RUI_RECOVERED, &ruip->rui_flags));
 
@@ -469,6 +470,7 @@ xfs_rui_recover(
 				XFS_EXT_UNWRITTEN : XFS_EXT_NORM;
 		whichfork = (rmap->me_flags & XFS_RMAP_EXTENT_ATTR_FORK) ?
 				XFS_ATTR_FORK : XFS_DATA_FORK;
+		rt = !!(rmap->me_flags & XFS_RMAP_EXTENT_REALTIME);
 		switch (rmap->me_flags & XFS_RMAP_EXTENT_TYPE_MASK) {
 		case XFS_RMAP_EXTENT_MAP:
 			type = XFS_RMAP_MAP;
@@ -501,7 +503,7 @@ xfs_rui_recover(
 		error = xfs_trans_log_finish_rmap_update(tp, rudp, type,
 				rmap->me_owner, whichfork,
 				rmap->me_startoff, rmap->me_startblock,
-				rmap->me_len, state, &rcur);
+				rmap->me_len, state, rt, &rcur);
 		if (error)
 			goto abort_error;
 
diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h
index aded200c59a8..a2b83c6d71ab 100644
--- a/fs/xfs/xfs_trans.h
+++ b/fs/xfs/xfs_trans.h
@@ -256,7 +256,7 @@ int xfs_trans_log_finish_rmap_update(struct xfs_trans *tp,
 		struct xfs_rud_log_item *rudp, enum xfs_rmap_intent_type type,
 		uint64_t owner, int whichfork, xfs_fileoff_t startoff,
 		xfs_fsblock_t startblock, xfs_filblks_t blockcount,
-		xfs_exntst_t state, struct xfs_btree_cur **pcur);
+		xfs_exntst_t state, bool rt, struct xfs_btree_cur **pcur);
 
 /* refcount updates */
 enum xfs_refcount_intent_type;
diff --git a/fs/xfs/xfs_trans_rmap.c b/fs/xfs/xfs_trans_rmap.c
index 05b00e40251f..abc673c63dfb 100644
--- a/fs/xfs/xfs_trans_rmap.c
+++ b/fs/xfs/xfs_trans_rmap.c
@@ -23,13 +23,16 @@ xfs_trans_set_rmap_flags(
 	struct xfs_map_extent		*rmap,
 	enum xfs_rmap_intent_type	type,
 	int				whichfork,
-	xfs_exntst_t			state)
+	xfs_exntst_t			state,
+	bool				rt)
 {
 	rmap->me_flags = 0;
 	if (state == XFS_EXT_UNWRITTEN)
 		rmap->me_flags |= XFS_RMAP_EXTENT_UNWRITTEN;
 	if (whichfork == XFS_ATTR_FORK)
 		rmap->me_flags |= XFS_RMAP_EXTENT_ATTR_FORK;
+	if (rt)
+		rmap->me_flags |= XFS_RMAP_EXTENT_REALTIME;
 	switch (type) {
 	case XFS_RMAP_MAP:
 		rmap->me_flags |= XFS_RMAP_EXTENT_MAP;
@@ -88,6 +91,7 @@ xfs_trans_log_finish_rmap_update(
 	xfs_fsblock_t			startblock,
 	xfs_filblks_t			blockcount,
 	xfs_exntst_t			state,
+	bool				rt,
 	struct xfs_btree_cur		**pcur)
 {
 	int				error;
@@ -176,7 +180,7 @@ xfs_rmap_update_log_item(
 	map->me_startoff = rmap->ri_bmap.br_startoff;
 	map->me_len = rmap->ri_bmap.br_blockcount;
 	xfs_trans_set_rmap_flags(map, rmap->ri_type, rmap->ri_whichfork,
-			rmap->ri_bmap.br_state);
+			rmap->ri_bmap.br_state, rmap->ri_realtime);
 }
 
 /* Get an RUD so we can process all the deferred rmap updates. */
@@ -208,6 +212,7 @@ xfs_rmap_update_finish_item(
 			rmap->ri_bmap.br_startblock,
 			rmap->ri_bmap.br_blockcount,
 			rmap->ri_bmap.br_state,
+			rmap->ri_realtime,
 			(struct xfs_btree_cur **)state);
 	kmem_free(rmap);
 	return error;

[PATCH 10/18] xfs: add realtime rmap btree block detection to log recovery

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Identify rtrmapbt blocks in the log correctly so that we can
validate them during log recovery.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 fs/xfs/xfs_log_recover.c |    4 ++++
 1 file changed, 4 insertions(+)


diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index db6d380ffa8d..f269c96cb0ab 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -2302,6 +2302,7 @@ xlog_recover_get_buf_lsn(
 		uuid = &btb->bb_u.s.bb_uuid;
 		break;
 	}
+	case XFS_RTRMAP_CRC_MAGIC:
 	case XFS_BMAP_CRC_MAGIC:
 	case XFS_BMAP_MAGIC: {
 		struct xfs_btree_block *btb = blk;
@@ -2464,6 +2465,9 @@ xlog_recover_validate_buf_type(
 		case XFS_BMAP_MAGIC:
 			bp->b_ops = &xfs_bmbt_buf_ops;
 			break;
+		case XFS_RTRMAP_CRC_MAGIC:
+			bp->b_ops = &xfs_rtrmapbt_buf_ops;
+			break;
 		case XFS_RMAP_CRC_MAGIC:
 			bp->b_ops = &xfs_rmapbt_buf_ops;
 			break;

[PATCH 11/18] xfs: add realtime reverse map inode to superblock

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Add a field to the superblock to record the rt rmapbt inode and load
it at mount time.  The rtrmapbt inode will have a unique extent format
code, which means that we also have to update the inode validation and
flush routines to look for it.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 fs/xfs/libxfs/xfs_format.h       |   10 +++++++++-
 fs/xfs/libxfs/xfs_inode_buf.c    |    6 ++++++
 fs/xfs/libxfs/xfs_inode_fork.c   |   11 +++++++++++
 fs/xfs/libxfs/xfs_rtrmap_btree.c |    2 +-
 fs/xfs/libxfs/xfs_sb.c           |    2 ++
 fs/xfs/libxfs/xfs_types.c        |    1 +
 fs/xfs/xfs_inode.c               |    9 ++++++++-
 fs/xfs/xfs_inode_item.c          |    2 ++
 fs/xfs/xfs_log_recover.c         |   14 +++++++++++++-
 fs/xfs/xfs_mount.h               |    1 +
 fs/xfs/xfs_ondisk.h              |    2 +-
 fs/xfs/xfs_rtalloc.c             |   21 +++++++++++++++++++++
 12 files changed, 76 insertions(+), 5 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index ad767b27dbb3..bb5ea9d46c07 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -174,6 +174,7 @@ typedef struct xfs_sb {
 	xfs_ino_t	sb_pquotino;	/* project quota inode */
 	xfs_lsn_t	sb_lsn;		/* last write sequence */
 	uuid_t		sb_meta_uuid;	/* metadata file system unique id */
+	xfs_ino_t	sb_rrmapino;	/* realtime reverse map inode */
 
 	/* must be padded to 64 bit alignment */
 } xfs_sb_t;
@@ -261,6 +262,7 @@ typedef struct xfs_dsb {
 	__be64		sb_pquotino;	/* project quota inode */
 	__be64		sb_lsn;		/* last write sequence */
 	uuid_t		sb_meta_uuid;	/* metadata file system unique id */
+	__be64		sb_rrmapino;	/* realtime reverse map inode */
 
 	/* must be padded to 64 bit alignment */
 } xfs_dsb_t;
@@ -544,6 +546,11 @@ 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_hasrtrmapbt(struct xfs_sb *sbp)
+{
+	return sbp->sb_rblocks > 0 && xfs_sb_version_hasrmapbt(sbp);
+}
+
 static inline bool xfs_sb_version_hasreflink(struct xfs_sb *sbp)
 {
 	return XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5 &&
@@ -926,7 +933,8 @@ typedef enum xfs_dinode_fmt {
 	XFS_DINODE_FMT_LOCAL,		/* bulk data */
 	XFS_DINODE_FMT_EXTENTS,		/* struct xfs_bmbt_rec */
 	XFS_DINODE_FMT_BTREE,		/* struct xfs_bmdr_block */
-	XFS_DINODE_FMT_UUID		/* added long ago, but never used */
+	XFS_DINODE_FMT_UUID,		/* added long ago, but never used */
+	XFS_DINODE_FMT_RMAP,		/* reverse mapping btree */
 } xfs_dinode_fmt_t;
 
 /*
diff --git a/fs/xfs/libxfs/xfs_inode_buf.c b/fs/xfs/libxfs/xfs_inode_buf.c
index 30d1d60f1d46..e33e3fa22b14 100644
--- a/fs/xfs/libxfs/xfs_inode_buf.c
+++ b/fs/xfs/libxfs/xfs_inode_buf.c
@@ -480,6 +480,12 @@ xfs_dinode_verify(
 			return __this_address;
 		break;
 	case S_IFREG:
+		if (dip->di_format == XFS_DINODE_FMT_RMAP) {
+			if (ino != mp->m_sb.sb_rrmapino)
+				return __this_address;
+			break;
+		}
+		/* fall through */
 	case S_IFLNK:
 	case S_IFDIR:
 		fa = xfs_dinode_verify_fork(dip, mp, XFS_DATA_FORK);
diff --git a/fs/xfs/libxfs/xfs_inode_fork.c b/fs/xfs/libxfs/xfs_inode_fork.c
index cfb68d4befc4..d1401cbdd266 100644
--- a/fs/xfs/libxfs/xfs_inode_fork.c
+++ b/fs/xfs/libxfs/xfs_inode_fork.c
@@ -76,6 +76,13 @@ xfs_iformat_fork(
 		case XFS_DINODE_FMT_BTREE:
 			error = xfs_iformat_btree(ip, dip, XFS_DATA_FORK);
 			break;
+		case XFS_DINODE_FMT_RMAP:
+			if (!xfs_sb_version_hasrtrmapbt(&ip->i_mount->m_sb))
+				return -EFSCORRUPTED;
+			if (ip->i_ino != ip->i_mount->m_sb.sb_rrmapino)
+				return -EFSCORRUPTED;
+			/* to be implemented later */
+			break;
 		default:
 			return -EFSCORRUPTED;
 		}
@@ -525,6 +532,10 @@ xfs_iflush_fork(
 		}
 		break;
 
+	case XFS_DINODE_FMT_RMAP:
+		/* to be implemented later */
+		break;
+
 	default:
 		ASSERT(0);
 		break;
diff --git a/fs/xfs/libxfs/xfs_rtrmap_btree.c b/fs/xfs/libxfs/xfs_rtrmap_btree.c
index 4529dbc6545c..ce9932397aa1 100644
--- a/fs/xfs/libxfs/xfs_rtrmap_btree.c
+++ b/fs/xfs/libxfs/xfs_rtrmap_btree.c
@@ -331,7 +331,7 @@ xfs_rtrmapbt_verify(
 
 	if (!xfs_sb_version_hasrmapbt(&mp->m_sb))
 		return __this_address;
-	fa = xfs_btree_lblock_v5hdr_verify(bp, XFS_RMAP_OWN_UNKNOWN);
+	fa = xfs_btree_lblock_v5hdr_verify(bp, mp->m_sb.sb_rrmapino);
 	if (fa)
 		return fa;
 	level = be16_to_cpu(block->bb_level);
diff --git a/fs/xfs/libxfs/xfs_sb.c b/fs/xfs/libxfs/xfs_sb.c
index 733bb7e5409c..e21fc6e0d2aa 100644
--- a/fs/xfs/libxfs/xfs_sb.c
+++ b/fs/xfs/libxfs/xfs_sb.c
@@ -522,6 +522,7 @@ __xfs_sb_from_disk(
 		uuid_copy(&to->sb_meta_uuid, &from->sb_meta_uuid);
 	else
 		uuid_copy(&to->sb_meta_uuid, &from->sb_uuid);
+	to->sb_rrmapino = be64_to_cpu(from->sb_rrmapino);
 	/* Convert on-disk flags to in-memory flags? */
 	if (convert_xquota)
 		xfs_sb_quota_from_disk(to);
@@ -665,6 +666,7 @@ xfs_sb_to_disk(
 		to->sb_lsn = cpu_to_be64(from->sb_lsn);
 		if (xfs_sb_version_hasmetauuid(from))
 			uuid_copy(&to->sb_meta_uuid, &from->sb_meta_uuid);
+		to->sb_rrmapino = cpu_to_be64(from->sb_rrmapino);
 	}
 }
 
diff --git a/fs/xfs/libxfs/xfs_types.c b/fs/xfs/libxfs/xfs_types.c
index 33a5ca346baf..3679ca947954 100644
--- a/fs/xfs/libxfs/xfs_types.c
+++ b/fs/xfs/libxfs/xfs_types.c
@@ -142,6 +142,7 @@ xfs_internal_inum(
 	xfs_ino_t		ino)
 {
 	return ino == mp->m_sb.sb_rbmino || ino == mp->m_sb.sb_rsumino ||
+	       ino == mp->m_sb.sb_rrmapino ||
 		(xfs_sb_version_hasquota(&mp->m_sb) &&
 		 xfs_is_quota_inode(&mp->m_sb, ino));
 }
diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
index 0e142787138c..f53e9488e796 100644
--- a/fs/xfs/xfs_inode.c
+++ b/fs/xfs/xfs_inode.c
@@ -3524,7 +3524,14 @@ xfs_iflush_int(
 			__func__, ip->i_ino, be16_to_cpu(dip->di_magic), dip);
 		goto corrupt_out;
 	}
-	if (S_ISREG(VFS_I(ip)->i_mode)) {
+	if (ip->i_ino == mp->m_sb.sb_rrmapino) {
+		if (ip->i_d.di_format != XFS_DINODE_FMT_RMAP) {
+			xfs_alert_tag(mp, XFS_PTAG_IFLUSH,
+				"%s: Bad rtrmap inode %llu, ptr 0x%p",
+				__func__, ip->i_ino, ip);
+			goto corrupt_out;
+		}
+	} else if (S_ISREG(VFS_I(ip)->i_mode)) {
 		if (XFS_TEST_ERROR(
 		    (ip->i_d.di_format != XFS_DINODE_FMT_EXTENTS) &&
 		    (ip->i_d.di_format != XFS_DINODE_FMT_BTREE),
diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c
index fa1c4fe2ffbf..1d2eb7bf9e04 100644
--- a/fs/xfs/xfs_inode_item.c
+++ b/fs/xfs/xfs_inode_item.c
@@ -46,6 +46,7 @@ xfs_inode_item_data_fork_size(
 		}
 		break;
 	case XFS_DINODE_FMT_BTREE:
+	case XFS_DINODE_FMT_RMAP:
 		if ((iip->ili_fields & XFS_ILOG_DBROOT) &&
 		    ip->i_df.if_broot_bytes > 0) {
 			*nbytes += ip->i_df.if_broot_bytes;
@@ -166,6 +167,7 @@ xfs_inode_item_format_data_fork(
 		}
 		break;
 	case XFS_DINODE_FMT_BTREE:
+	case XFS_DINODE_FMT_RMAP:
 		iip->ili_fields &=
 			~(XFS_ILOG_DDATA | XFS_ILOG_DEXT | XFS_ILOG_DEV);
 
diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index f269c96cb0ab..95af44487b6c 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -3123,7 +3123,19 @@ xlog_recover_inode_pass2(
 	/* Take the opportunity to reset the flush iteration count */
 	ldip->di_flushiter = 0;
 
-	if (unlikely(S_ISREG(ldip->di_mode))) {
+	if (in_f->ilf_ino == mp->m_sb.sb_rrmapino) {
+		if (ldip->di_format != XFS_DINODE_FMT_RMAP) {
+			XFS_CORRUPTION_ERROR("xlog_recover_inode_pass2(3)",
+					 XFS_ERRLEVEL_LOW, mp, ldip,
+					 sizeof(*ldip));
+			xfs_alert(mp,
+		"%s: Bad rtrmapbt inode log record, rec ptr 0x%p, "
+		"ino ptr = 0x%p, ino bp = 0x%p, ino %Ld",
+				__func__, item, dip, bp, in_f->ilf_ino);
+			error = -EFSCORRUPTED;
+			goto out_release;
+		}
+	} else if (unlikely(S_ISREG(ldip->di_mode))) {
 		if ((ldip->di_format != XFS_DINODE_FMT_EXTENTS) &&
 		    (ldip->di_format != XFS_DINODE_FMT_BTREE)) {
 			XFS_CORRUPTION_ERROR("xlog_recover_inode_pass2(3)",
diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h
index 5a4df6b860e2..712bec475c5d 100644
--- a/fs/xfs/xfs_mount.h
+++ b/fs/xfs/xfs_mount.h
@@ -91,6 +91,7 @@ typedef struct xfs_mount {
 	uint			m_rsumsize;	/* size of rt summary, bytes */
 	struct xfs_inode	*m_rbmip;	/* pointer to bitmap inode */
 	struct xfs_inode	*m_rsumip;	/* pointer to summary inode */
+	struct xfs_inode	*m_rrmapip;	/* pointer to rmap inode */
 	struct xfs_inode	*m_rootip;	/* pointer to root directory */
 	struct xfs_quotainfo	*m_quotainfo;	/* disk quota information */
 	xfs_buftarg_t		*m_ddev_targp;	/* saves taking the address */
diff --git a/fs/xfs/xfs_ondisk.h b/fs/xfs/xfs_ondisk.h
index 23c9b6f047c5..0840ec7e58aa 100644
--- a/fs/xfs/xfs_ondisk.h
+++ b/fs/xfs/xfs_ondisk.h
@@ -33,7 +33,7 @@ xfs_check_ondisk_structs(void)
 	XFS_CHECK_STRUCT_SIZE(struct xfs_dinode,		176);
 	XFS_CHECK_STRUCT_SIZE(struct xfs_disk_dquot,		104);
 	XFS_CHECK_STRUCT_SIZE(struct xfs_dqblk,			136);
-	XFS_CHECK_STRUCT_SIZE(struct xfs_dsb,			264);
+	XFS_CHECK_STRUCT_SIZE(struct xfs_dsb,			272);
 	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);
diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c
index 926ed314ffba..1ac69ff79a22 100644
--- a/fs/xfs/xfs_rtalloc.c
+++ b/fs/xfs/xfs_rtalloc.c
@@ -1211,6 +1211,25 @@ xfs_rtmount_inodes(
 		return error;
 	}
 	ASSERT(mp->m_rsumip != NULL);
+
+	/* If we have rmap and a realtime device, look for the rtrmapbt. */
+	if (xfs_sb_version_hasrmapbt(&mp->m_sb) && mp->m_sb.sb_rblocks > 0) {
+		ASSERT(sbp->sb_rrmapino != NULLFSINO);
+		error = xfs_iget(mp, NULL, sbp->sb_rrmapino, 0, 0,
+				&mp->m_rrmapip);
+		if (error) {
+			if (mp->m_rrmapip)
+				xfs_irele(mp->m_rrmapip);
+			return error;
+		}
+		ASSERT(mp->m_rrmapip != NULL);
+		if (mp->m_rrmapip->i_d.di_format != XFS_DINODE_FMT_RMAP) {
+			xfs_irele(mp->m_rrmapip);
+			mp->m_rrmapip = NULL;
+			return -EFSCORRUPTED;
+		}
+	}
+
 	return 0;
 }
 
@@ -1218,6 +1237,8 @@ void
 xfs_rtunmount_inodes(
 	struct xfs_mount	*mp)
 {
+	if (mp->m_rrmapip)
+		xfs_irele(mp->m_rrmapip);
 	if (mp->m_rbmip)
 		xfs_irele(mp->m_rbmip);
 	if (mp->m_rsumip)

[PATCH 12/18] xfs: wire up a new inode fork type for the realtime rmap

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Plumb in the pieces we need to embed the root of the realtime rmap
btree in an inode's data fork, complete with new fork type and
on-disk interpretation functions.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 fs/xfs/libxfs/xfs_format.h       |    8 +
 fs/xfs/libxfs/xfs_inode_fork.c   |   48 +++++++
 fs/xfs/libxfs/xfs_rtrmap_btree.c |  246 ++++++++++++++++++++++++++++++++++++++
 fs/xfs/libxfs/xfs_rtrmap_btree.h |   51 ++++++++
 4 files changed, 351 insertions(+), 2 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index bb5ea9d46c07..dcd790f61895 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -1460,6 +1460,14 @@ typedef __be32 xfs_rmap_ptr_t;
  */
 #define	XFS_RTRMAP_CRC_MAGIC	0x4d415052	/* 'MAPR' */
 
+/*
+ * rtrmap root header, on-disk form only.
+ */
+struct xfs_rtrmap_root {
+	__be16		bb_level;	/* 0 is a leaf */
+	__be16		bb_numrecs;	/* current # of data records */
+};
+
 /*
  * Data record structure
  */
diff --git a/fs/xfs/libxfs/xfs_inode_fork.c b/fs/xfs/libxfs/xfs_inode_fork.c
index d1401cbdd266..65859a518def 100644
--- a/fs/xfs/libxfs/xfs_inode_fork.c
+++ b/fs/xfs/libxfs/xfs_inode_fork.c
@@ -25,12 +25,14 @@
 #include "xfs_dir2_priv.h"
 #include "xfs_attr_leaf.h"
 #include "xfs_shared.h"
+#include "xfs_rtrmap_btree.h"
 
 kmem_zone_t *xfs_ifork_zone;
 
 STATIC int xfs_iformat_local(xfs_inode_t *, xfs_dinode_t *, int, int);
 STATIC int xfs_iformat_extents(xfs_inode_t *, xfs_dinode_t *, int);
 STATIC int xfs_iformat_btree(xfs_inode_t *, xfs_dinode_t *, int);
+STATIC int xfs_iformat_rmap(struct xfs_inode *ip, struct xfs_dinode *dip);
 
 /*
  * Copy inode type and data and attr format specific information from the
@@ -81,7 +83,7 @@ xfs_iformat_fork(
 				return -EFSCORRUPTED;
 			if (ip->i_ino != ip->i_mount->m_sb.sb_rrmapino)
 				return -EFSCORRUPTED;
-			/* to be implemented later */
+			error = xfs_iformat_rmap(ip, dip);
 			break;
 		default:
 			return -EFSCORRUPTED;
@@ -328,6 +330,37 @@ xfs_iformat_btree(
 	return 0;
 }
 
+/* The file is a reverse mapping tree. */
+STATIC int
+xfs_iformat_rmap(
+	struct xfs_inode	*ip,
+	struct xfs_dinode	*dip)
+{
+	struct xfs_rtrmap_root	*dfp;
+	struct xfs_ifork	*ifp;
+	/* REFERENCED */
+	int			size;
+	int			whichfork = XFS_DATA_FORK;
+
+	ifp = XFS_IFORK_PTR(ip, whichfork);
+	dfp = (struct xfs_rtrmap_root *)XFS_DFORK_PTR(dip, whichfork);
+	size = XFS_RTRMAP_BROOT_SPACE(dfp);
+
+	ifp->if_broot_bytes = size;
+	ifp->if_broot = kmem_alloc(size, KM_SLEEP | KM_NOFS);
+	ASSERT(ifp->if_broot != NULL);
+	/*
+	 * Copy and convert from the on-disk structure
+	 * to the in-memory structure.
+	 */
+	xfs_rtrmapbt_from_disk(ip, dfp,
+			XFS_DFORK_SIZE(dip, ip->i_mount, whichfork),
+			ifp->if_broot, size);
+	ifp->if_flags = XFS_IFBROOT;
+
+	return 0;
+}
+
 /*
  * This is called when the amount of space needed for if_data
  * is increased or decreased.  The change in size is indicated by
@@ -533,7 +566,18 @@ xfs_iflush_fork(
 		break;
 
 	case XFS_DINODE_FMT_RMAP:
-		/* to be implemented later */
+		ASSERT(whichfork == XFS_DATA_FORK);
+		ASSERT(ip->i_ino == mp->m_sb.sb_rrmapino);
+		if ((iip->ili_fields & brootflag[whichfork]) &&
+		    (ifp->if_broot_bytes > 0)) {
+			ASSERT(ifp->if_broot != NULL);
+			ASSERT(XFS_RTRMAP_ROOT_SPACE(ifp->if_broot) <=
+				XFS_IFORK_SIZE(ip, whichfork));
+			xfs_rtrmapbt_to_disk(mp, ifp->if_broot,
+				ifp->if_broot_bytes,
+				(struct xfs_rtrmap_root *)cp,
+				XFS_DFORK_SIZE(dip, mp, whichfork));
+		}
 		break;
 
 	default:
diff --git a/fs/xfs/libxfs/xfs_rtrmap_btree.c b/fs/xfs/libxfs/xfs_rtrmap_btree.c
index ce9932397aa1..d04630cb9fd0 100644
--- a/fs/xfs/libxfs/xfs_rtrmap_btree.c
+++ b/fs/xfs/libxfs/xfs_rtrmap_btree.c
@@ -200,6 +200,42 @@ xfs_rtrmapbt_get_maxrecs(
 	return cur->bc_mp->m_rtrmap_mxr[level != 0];
 }
 
+/*
+ * Calculate number of records in a realtime rmap btree inode root.
+ */
+STATIC int
+xfs_rtrmapbt_root_maxrecs(
+	int			blocklen,
+	bool			leaf)
+{
+	blocklen -= sizeof(struct xfs_rtrmap_root);
+
+	if (leaf)
+		return blocklen / sizeof(struct xfs_rtrmap_rec);
+	return blocklen / (2 * sizeof(struct xfs_rtrmap_key) +
+			sizeof(xfs_rtrmap_ptr_t));
+}
+
+/*
+ * Get the maximum records we could store in the on-disk format.
+ *
+ * For non-root nodes this is equivalent to xfs_bmbt_get_maxrecs, but
+ * for the root node this checks the available space in the dinode fork
+ * so that we can resize the in-memory buffer to match it.  After a
+ * resize to the maximum size this function returns the same value
+ * as xfs_bmbt_get_maxrecs for the root node, too.
+ */
+STATIC int
+xfs_rtrmapbt_get_dmaxrecs(
+	struct xfs_btree_cur	*cur,
+	int			level)
+{
+	if (level != cur->bc_nlevels - 1)
+		return cur->bc_mp->m_rtrmap_mxr[level != 0];
+	return xfs_rtrmapbt_root_maxrecs(cur->bc_private.b.forksize,
+			level == 0);
+}
+
 STATIC void
 xfs_rtrmapbt_init_key_from_rec(
 	union xfs_btree_key	*key,
@@ -317,6 +353,125 @@ xfs_rtrmapbt_diff_two_keys(
 	return 0;
 }
 
+/*
+ * Reallocate the space for if_broot based on the number of records
+ * being added or deleted as indicated in rec_diff.  Move the records
+ * and pointers in if_broot to fit the new size.  When shrinking this
+ * will eliminate holes between the records and pointers created by
+ * the caller.  When growing this will create holes to be filled in
+ * by the caller.
+ *
+ * The caller must not request to add more records than would fit in
+ * the on-disk inode root.  If the if_broot is currently NULL, then
+ * if we are adding records, one will be allocated.  The caller must also
+ * not request that the number of records go below zero, although
+ * it can go to zero.
+ */
+STATIC void
+xfs_rtrmapbt_iroot_realloc(
+	struct xfs_btree_cur	*cur,
+	int			rec_diff)
+{
+	struct xfs_inode	*ip = cur->bc_private.b.ip;
+	int			whichfork = cur->bc_private.b.whichfork;
+	int			cur_max;
+	struct xfs_ifork	*ifp;
+	struct xfs_btree_block	*new_broot;
+	struct xfs_btree_block	*broot;
+	int			new_max;
+	size_t			new_size;
+	char			*np;
+	char			*op;
+	int			level;
+
+	/*
+	 * Handle the degenerate case quietly.
+	 */
+	if (rec_diff == 0)
+		return;
+
+	ifp = XFS_IFORK_PTR(ip, whichfork);
+	if (rec_diff > 0) {
+		/*
+		 * If there wasn't any memory allocated before, just
+		 * allocate it now and get out.
+		 */
+		if (ifp->if_broot_bytes == 0) {
+			new_size = XFS_RTRMAP_BROOT_SPACE_CALC(rec_diff,
+					cur->bc_nlevels - 1);
+			ifp->if_broot = kmem_alloc(new_size,
+					KM_SLEEP | KM_NOFS);
+			ifp->if_broot_bytes = (int)new_size;
+			return;
+		}
+
+		/*
+		 * If there is already an existing if_broot, then we need
+		 * to realloc() it and shift the pointers to their new
+		 * location.  The records don't change location because
+		 * they are kept butted up against the btree block header.
+		 */
+		broot = (struct xfs_btree_block *)ifp->if_broot;
+		level = be16_to_cpu(broot->bb_level);
+		cur_max = xfs_rtrmapbt_maxrecs(ifp->if_broot_bytes, level == 0);
+		new_max = cur_max + rec_diff;
+		new_size = XFS_RTRMAP_BROOT_SPACE_CALC(new_max, level);
+		ifp->if_broot = kmem_realloc(ifp->if_broot, new_size,
+				KM_SLEEP | KM_NOFS);
+		if (level > 0) {
+			op = (char *)XFS_RTRMAP_BROOT_PTR_ADDR(ifp->if_broot,
+					1, ifp->if_broot_bytes);
+			np = (char *)XFS_RTRMAP_BROOT_PTR_ADDR(ifp->if_broot,
+					1, (int)new_size);
+			memmove(np, op, cur_max * sizeof(xfs_fsblock_t));
+		}
+		ifp->if_broot_bytes = (int)new_size;
+		ASSERT(XFS_RTRMAP_ROOT_SPACE(ifp->if_broot) <=
+				XFS_IFORK_SIZE(ip, whichfork));
+		return;
+	}
+
+	/*
+	 * rec_diff is less than 0.  In this case, we are shrinking the
+	 * if_broot buffer.  It must already exist.  If we go to zero
+	 * records, just get rid of the root and clear the status bit.
+	 */
+	ASSERT((ifp->if_broot != NULL) && (ifp->if_broot_bytes > 0));
+	broot = (struct xfs_btree_block *)ifp->if_broot;
+	level = be16_to_cpu(broot->bb_level);
+	cur_max = xfs_rtrmapbt_maxrecs(ifp->if_broot_bytes, level == 0);
+	new_max = cur_max + rec_diff;
+	if (new_max < 0)
+		new_max = 0;
+	new_size = XFS_RTRMAP_BROOT_SPACE_CALC(new_max, level);
+	new_broot = kmem_alloc(new_size, KM_SLEEP | KM_NOFS);
+	memcpy(new_broot, ifp->if_broot, XFS_RTRMAP_BLOCK_LEN);
+
+	/* Copy the records or keys and pointers. */
+	if (level > 0) {
+		op = (char *)XFS_RTRMAP_KEY_ADDR(ifp->if_broot, 1);
+		np = (char *)XFS_RTRMAP_KEY_ADDR(new_broot, 1);
+		memcpy(np, op, new_max * 2 * sizeof(struct xfs_rtrmap_key));
+
+		op = (char *)XFS_RTRMAP_BROOT_PTR_ADDR(ifp->if_broot, 1,
+				ifp->if_broot_bytes);
+		np = (char *)XFS_RTRMAP_BROOT_PTR_ADDR(new_broot, 1,
+				(int)new_size);
+		memcpy(np, op, new_max * sizeof(xfs_fsblock_t));
+	} else {
+		op = (char *)XFS_RTRMAP_REC_ADDR(ifp->if_broot, 1);
+		np = (char *)XFS_RTRMAP_REC_ADDR(new_broot, 1);
+		memcpy(np, op, new_max * sizeof(struct xfs_rtrmap_rec));
+	}
+
+	kmem_free(ifp->if_broot);
+	ifp->if_broot = new_broot;
+	ifp->if_broot_bytes = (int)new_size;
+	if (ifp->if_broot)
+		ASSERT(XFS_RTRMAP_ROOT_SPACE(ifp->if_broot) <=
+				XFS_IFORK_SIZE(ip, whichfork));
+}
+
 static xfs_failaddr_t
 xfs_rtrmapbt_verify(
 	struct xfs_buf		*bp)
@@ -427,12 +582,14 @@ static const struct xfs_btree_ops xfs_rtrmapbt_ops = {
 	.free_block		= xfs_rtrmapbt_free_block,
 	.get_minrecs		= xfs_rtrmapbt_get_minrecs,
 	.get_maxrecs		= xfs_rtrmapbt_get_maxrecs,
+	.get_dmaxrecs		= xfs_rtrmapbt_get_dmaxrecs,
 	.init_key_from_rec	= xfs_rtrmapbt_init_key_from_rec,
 	.init_high_key_from_rec	= xfs_rtrmapbt_init_high_key_from_rec,
 	.init_rec_from_cur	= xfs_rtrmapbt_init_rec_from_cur,
 	.init_ptr_from_cur	= xfs_rtrmapbt_init_ptr_from_cur,
 	.key_diff		= xfs_rtrmapbt_key_diff,
 	.buf_ops		= &xfs_rtrmapbt_buf_ops,
+	.iroot_realloc		= xfs_rtrmapbt_iroot_realloc,
 	.diff_two_keys		= xfs_rtrmapbt_diff_two_keys,
 	.keys_inorder		= xfs_rtrmapbt_keys_inorder,
 	.recs_inorder		= xfs_rtrmapbt_recs_inorder,
@@ -496,3 +653,92 @@ xfs_rtrmapbt_compute_maxlevels(
 			mp->m_sb.sb_rblocks);
 	ASSERT(mp->m_rtrmap_maxlevels <= XFS_BTREE_MAXLEVELS);
 }
+
+/*
+ * Convert on-disk form of btree root to in-memory form.
+ */
+void
+xfs_rtrmapbt_from_disk(
+	struct xfs_inode	*ip,
+	struct xfs_rtrmap_root	*dblock,
+	int			dblocklen,
+	struct xfs_btree_block	*rblock,
+	int			rblocklen)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	int			dmxr;
+	struct xfs_rtrmap_key	*fkp;
+	__be64			*fpp;
+	struct xfs_rtrmap_key	*tkp;
+	__be64			*tpp;
+	struct xfs_rtrmap_rec	*frp;
+	struct xfs_rtrmap_rec	*trp;
+
+	xfs_btree_init_block_int(mp, rblock, XFS_BUF_DADDR_NULL,
+			 XFS_BTNUM_RTRMAP, 0, 0, ip->i_ino,
+			 XFS_BTREE_LONG_PTRS | XFS_BTREE_CRC_BLOCKS);
+
+	rblock->bb_level = dblock->bb_level;
+	rblock->bb_numrecs = dblock->bb_numrecs;
+
+	if (be16_to_cpu(rblock->bb_level) > 0) {
+		dmxr = xfs_rtrmapbt_maxrecs(dblocklen, 0);
+		fkp = XFS_RTRMAP_ROOT_KEY_ADDR(dblock, 1);
+		tkp = XFS_RTRMAP_KEY_ADDR(rblock, 1);
+		fpp = XFS_RTRMAP_ROOT_PTR_ADDR(dblock, 1, dmxr);
+		tpp = XFS_RTRMAP_BROOT_PTR_ADDR(rblock, 1, rblocklen);
+		dmxr = be16_to_cpu(dblock->bb_numrecs);
+		memcpy(tkp, fkp, 2 * sizeof(*fkp) * dmxr);
+		memcpy(tpp, fpp, sizeof(*fpp) * dmxr);
+	} else {
+		frp = XFS_RTRMAP_ROOT_REC_ADDR(dblock, 1);
+		trp = XFS_RTRMAP_REC_ADDR(rblock, 1);
+		dmxr = be16_to_cpu(dblock->bb_numrecs);
+		memcpy(trp, frp, sizeof(*frp) * dmxr);
+	}
+}
+
+/*
+ * Convert in-memory form of btree root to on-disk form.
+ */
+void
+xfs_rtrmapbt_to_disk(
+	struct xfs_mount	*mp,
+	struct xfs_btree_block	*rblock,
+	int			rblocklen,
+	struct xfs_rtrmap_root	*dblock,
+	int			dblocklen)
+{
+	int			dmxr;
+	struct xfs_rtrmap_key	*fkp;
+	__be64			*fpp;
+	struct xfs_rtrmap_key	*tkp;
+	__be64			*tpp;
+	struct xfs_rtrmap_rec	*frp;
+	struct xfs_rtrmap_rec	*trp;
+
+	ASSERT(rblock->bb_magic == cpu_to_be32(XFS_RTRMAP_CRC_MAGIC));
+	ASSERT(uuid_equal(&rblock->bb_u.l.bb_uuid, &mp->m_sb.sb_meta_uuid));
+	ASSERT(rblock->bb_u.l.bb_blkno == cpu_to_be64(XFS_BUF_DADDR_NULL));
+	ASSERT(rblock->bb_u.l.bb_leftsib == cpu_to_be64(NULLFSBLOCK));
+	ASSERT(rblock->bb_u.l.bb_rightsib == cpu_to_be64(NULLFSBLOCK));
+
+	dblock->bb_level = rblock->bb_level;
+	dblock->bb_numrecs = rblock->bb_numrecs;
+
+	if (be16_to_cpu(rblock->bb_level) > 0) {
+		dmxr = xfs_rtrmapbt_maxrecs(dblocklen, 0);
+		fkp = XFS_RTRMAP_KEY_ADDR(rblock, 1);
+		tkp = XFS_RTRMAP_ROOT_KEY_ADDR(dblock, 1);
+		fpp = XFS_RTRMAP_BROOT_PTR_ADDR(rblock, 1, rblocklen);
+		tpp = XFS_RTRMAP_ROOT_PTR_ADDR(dblock, 1, dmxr);
+		dmxr = be16_to_cpu(rblock->bb_numrecs);
+		memcpy(tkp, fkp, 2 * sizeof(*fkp) * dmxr);
+		memcpy(tpp, fpp, sizeof(*fpp) * dmxr);
+	} else {
+		frp = XFS_RTRMAP_REC_ADDR(rblock, 1);
+		trp = XFS_RTRMAP_ROOT_REC_ADDR(dblock, 1);
+		dmxr = be16_to_cpu(rblock->bb_numrecs);
+		memcpy(trp, frp, sizeof(*frp) * dmxr);
+	}
+}
diff --git a/fs/xfs/libxfs/xfs_rtrmap_btree.h b/fs/xfs/libxfs/xfs_rtrmap_btree.h
index 4b32823a2ec6..321be07d1cd1 100644
--- a/fs/xfs/libxfs/xfs_rtrmap_btree.h
+++ b/fs/xfs/libxfs/xfs_rtrmap_btree.h
@@ -40,9 +40,60 @@ struct xfs_mount;
 		 (maxrecs) * 2 * sizeof(struct xfs_rtrmap_key) + \
 		 ((index) - 1) * sizeof(xfs_rtrmap_ptr_t)))
 
+/* Macros for handling the inode root */
+
+#define XFS_RTRMAP_ROOT_REC_ADDR(block, index) \
+	((struct xfs_rtrmap_rec *) \
+		((char *)(block) + \
+		 sizeof(struct xfs_rtrmap_root) + \
+		 ((index) - 1) * sizeof(struct xfs_rtrmap_rec)))
+
+#define XFS_RTRMAP_ROOT_KEY_ADDR(block, index) \
+	((struct xfs_rtrmap_key *) \
+		((char *)(block) + \
+		 sizeof(struct xfs_rtrmap_root) + \
+		 ((index) - 1) * 2 * sizeof(struct xfs_rtrmap_key)))
+
+#define XFS_RTRMAP_ROOT_PTR_ADDR(block, index, maxrecs) \
+	((xfs_rtrmap_ptr_t *) \
+		((char *)(block) + \
+		 sizeof(struct xfs_rtrmap_root) + \
+		 (maxrecs) * 2 * sizeof(struct xfs_rtrmap_key) + \
+		 ((index) - 1) * sizeof(xfs_rtrmap_ptr_t)))
+
+#define XFS_RTRMAP_BROOT_PTR_ADDR(bb, i, sz) \
+	XFS_RTRMAP_PTR_ADDR(bb, i, xfs_rtrmapbt_maxrecs(sz, 0))
+
+#define XFS_RTRMAP_BROOT_SPACE_CALC(nrecs, level) \
+	(int)(XFS_RTRMAP_BLOCK_LEN + ((level) > 0 ? \
+	       ((nrecs) * (2 * sizeof(struct xfs_rtrmap_key) + \
+			   sizeof(xfs_rtrmap_ptr_t))) : \
+	       ((nrecs) * sizeof(struct xfs_rtrmap_rec))))
+
+#define XFS_RTRMAP_BROOT_SPACE(bb) \
+	(XFS_RTRMAP_BROOT_SPACE_CALC(be16_to_cpu((bb)->bb_numrecs), \
+				     be16_to_cpu((bb)->bb_level)))
+
+#define XFS_RTRMAP_ROOT_SPACE_CALC(nrecs, level) \
+	(int)(sizeof(struct xfs_rtrmap_root) + ((level) > 0 ? \
+	       ((nrecs) * (2 * sizeof(struct xfs_rtrmap_key) + \
+			   sizeof(xfs_rtrmap_ptr_t))) : \
+	       ((nrecs) * sizeof(struct xfs_rtrmap_rec))))
+
+#define XFS_RTRMAP_ROOT_SPACE(bb) \
+	(XFS_RTRMAP_ROOT_SPACE_CALC(be16_to_cpu((bb)->bb_numrecs), \
+				    be16_to_cpu((bb)->bb_level)))
+
 struct xfs_btree_cur *xfs_rtrmapbt_init_cursor(struct xfs_mount *mp,
 				struct xfs_trans *tp, struct xfs_inode *ip);
 int xfs_rtrmapbt_maxrecs(int blocklen, bool leaf);
 void xfs_rtrmapbt_compute_maxlevels(struct xfs_mount *mp);
 
+void xfs_rtrmapbt_from_disk(struct xfs_inode *ip,
+		struct xfs_rtrmap_root *dblock, int dblocklen,
+		struct xfs_btree_block *rblock, int rblocklen);
+void xfs_rtrmapbt_to_disk(struct xfs_mount *mp,
+		struct xfs_btree_block *rblock, int rblocklen,
+		struct xfs_rtrmap_root *dblock, int dblocklen);
+
 #endif	/* __XFS_RTRMAP_BTREE_H__ */

[PATCH 13/18] xfs: wire up rmap map and unmap to the realtime rmapbt

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Connect the map and unmap reverse-mapping operations to the realtime
rmapbt via the deferred operation callbacks.  This enables us to
perform rmap operations against the correct btree.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 fs/xfs/libxfs/xfs_rmap.c |   62 ++++++++++++++++++++++++++++++----------------
 fs/xfs/libxfs/xfs_rmap.h |    2 +
 fs/xfs/xfs_log_recover.c |    6 ++--
 fs/xfs/xfs_rmap_item.c   |   11 +++++++-
 fs/xfs/xfs_rmap_item.h   |    2 +
 fs/xfs/xfs_trans_rmap.c  |    2 +
 6 files changed, 56 insertions(+), 29 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c
index 377d4c31d69f..7151c284a415 100644
--- a/fs/xfs/libxfs/xfs_rmap.c
+++ b/fs/xfs/libxfs/xfs_rmap.c
@@ -28,6 +28,7 @@
 #include "xfs_bmap.h"
 #include "xfs_inode.h"
 #include "xfs_ialloc.h"
+#include "xfs_rtrmap_btree.h"
 
 /* By convention, the rtrmapbt's "AG" number is NULLAGNUMBER. */
 static xfs_agnumber_t
@@ -2190,13 +2191,14 @@ xfs_rmap_finish_one_cleanup(
 	struct xfs_btree_cur	*rcur,
 	int			error)
 {
-	struct xfs_buf		*agbp;
+	struct xfs_buf		*agbp = NULL;
 
 	if (rcur == NULL)
 		return;
-	agbp = rcur->bc_private.a.agbp;
+	if (!(rcur->bc_flags & XFS_BTREE_LONG_PTRS))
+		agbp = rcur->bc_private.a.agbp;
 	xfs_btree_del_cursor(rcur, error);
-	if (error)
+	if (error && agbp)
 		xfs_trans_brelse(tp, agbp);
 }
 
@@ -2217,20 +2219,21 @@ xfs_rmap_finish_one(
 	xfs_fsblock_t			startblock,
 	xfs_filblks_t			blockcount,
 	xfs_exntst_t			state,
+	bool				realtime,
 	struct xfs_btree_cur		**pcur)
 {
 	struct xfs_mount		*mp = tp->t_mountp;
 	struct xfs_btree_cur		*rcur;
 	struct xfs_buf			*agbp = NULL;
+	int				lockmode;
 	int				error = 0;
 	xfs_agnumber_t			agno;
 	struct xfs_owner_info		oinfo;
 	xfs_fsblock_t			bno;
 	bool				unwritten;
 
-	agno = XFS_FSB_TO_AGNO(mp, startblock);
-	ASSERT(agno != NULLAGNUMBER);
-	bno = XFS_FSB_TO_AGBNO(mp, startblock);
+	agno = realtime ? NULLAGNUMBER : XFS_FSB_TO_AGNO(mp, startblock);
+	bno = realtime ? startblock : XFS_FSB_TO_AGBNO(mp, startblock);
 
 	trace_xfs_rmap_deferred(mp, agno, type, bno, owner, whichfork,
 			startoff, blockcount, state);
@@ -2249,31 +2252,44 @@ xfs_rmap_finish_one(
 		*pcur = NULL;
 	}
 	if (rcur == NULL) {
-		/*
-		 * Refresh the freelist before we start changing the
-		 * rmapbt, because a shape change could cause us to
-		 * allocate blocks.
-		 */
-		error = xfs_free_extent_fix_freelist(tp, agno, &agbp);
-		if (error)
-			return error;
-		if (!agbp)
-			return -EFSCORRUPTED;
-
-		rcur = xfs_rmapbt_init_cursor(mp, tp, agbp, agno);
-		if (!rcur) {
-			error = -ENOMEM;
-			goto out_cur;
+		if (realtime) {
+			lockmode = XFS_ILOCK_EXCL | XFS_ILOCK_RTBITMAP;
+			xfs_ilock(mp->m_rrmapip, lockmode);
+			xfs_trans_ijoin(tp, mp->m_rrmapip, lockmode);
+			rcur = xfs_rtrmapbt_init_cursor(mp, tp, mp->m_rrmapip);
+			if (!rcur) {
+				error = -ENOMEM;
+				goto out_cur;
+			}
+			rcur->bc_private.b.flags = 0;
+		} else {
+			/*
+			 * Refresh the freelist before we start changing the
+			 * rmapbt, because a shape change could cause us to
+			 * allocate blocks.
+			 */
+			error = xfs_free_extent_fix_freelist(tp, agno, &agbp);
+			if (error)
+				return error;
+			if (!agbp)
+				return -EFSCORRUPTED;
+
+			rcur = xfs_rmapbt_init_cursor(mp, tp, agbp, agno);
+			if (!rcur) {
+				error = -ENOMEM;
+				goto out_cur;
+			}
 		}
 	}
 	*pcur = rcur;
 
 	xfs_rmap_ino_owner(&oinfo, owner, whichfork, startoff);
 	unwritten = state == XFS_EXT_UNWRITTEN;
-	bno = XFS_FSB_TO_AGBNO(rcur->bc_mp, startblock);
 
 	switch (type) {
 	case XFS_RMAP_ALLOC:
+		ASSERT(!realtime);
+		/* fall through */
 	case XFS_RMAP_MAP:
 		error = xfs_rmap_map(rcur, bno, blockcount, unwritten, &oinfo);
 		break;
@@ -2282,6 +2298,8 @@ xfs_rmap_finish_one(
 				&oinfo);
 		break;
 	case XFS_RMAP_FREE:
+		ASSERT(!realtime);
+		/* fall through */
 	case XFS_RMAP_UNMAP:
 		error = xfs_rmap_unmap(rcur, bno, blockcount, unwritten,
 				&oinfo);
diff --git a/fs/xfs/libxfs/xfs_rmap.h b/fs/xfs/libxfs/xfs_rmap.h
index a4d3146e5619..88075a8db153 100644
--- a/fs/xfs/libxfs/xfs_rmap.h
+++ b/fs/xfs/libxfs/xfs_rmap.h
@@ -203,7 +203,7 @@ void xfs_rmap_finish_one_cleanup(struct xfs_trans *tp,
 int xfs_rmap_finish_one(struct xfs_trans *tp, enum xfs_rmap_intent_type type,
 		uint64_t owner, int whichfork, xfs_fileoff_t startoff,
 		xfs_fsblock_t startblock, xfs_filblks_t blockcount,
-		xfs_exntst_t state, struct xfs_btree_cur **pcur);
+		xfs_exntst_t state, bool realtime, struct xfs_btree_cur **pcur);
 
 int xfs_rmap_find_left_neighbor(struct xfs_btree_cur *cur, xfs_fsblock_t bno,
 		uint64_t owner, uint64_t offset, unsigned int flags,
diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index 95af44487b6c..34451a777709 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -4709,7 +4709,7 @@ xlog_recover_cancel_efi(
 /* Recover the RUI if necessary. */
 STATIC int
 xlog_recover_process_rui(
-	struct xfs_mount		*mp,
+	struct xfs_trans		*parent_tp,
 	struct xfs_ail			*ailp,
 	struct xfs_log_item		*lip)
 {
@@ -4724,7 +4724,7 @@ xlog_recover_process_rui(
 		return 0;
 
 	spin_unlock(&ailp->ail_lock);
-	error = xfs_rui_recover(mp, ruip);
+	error = xfs_rui_recover(parent_tp, ruip);
 	spin_lock(&ailp->ail_lock);
 
 	return error;
@@ -4953,7 +4953,7 @@ xlog_recover_process_intents(
 			error = xlog_recover_process_efi(log->l_mp, ailp, lip);
 			break;
 		case XFS_LI_RUI:
-			error = xlog_recover_process_rui(log->l_mp, ailp, lip);
+			error = xlog_recover_process_rui(parent_tp, ailp, lip);
 			break;
 		case XFS_LI_CUI:
 			error = xlog_recover_process_cui(parent_tp, ailp, lip);
diff --git a/fs/xfs/xfs_rmap_item.c b/fs/xfs/xfs_rmap_item.c
index 2fe98a300d0e..bc4d1e7ae602 100644
--- a/fs/xfs/xfs_rmap_item.c
+++ b/fs/xfs/xfs_rmap_item.c
@@ -401,7 +401,7 @@ xfs_rud_init(
  */
 int
 xfs_rui_recover(
-	struct xfs_mount		*mp,
+	struct xfs_trans		*parent_tp,
 	struct xfs_rui_log_item		*ruip)
 {
 	int				i;
@@ -415,6 +415,7 @@ xfs_rui_recover(
 	xfs_exntst_t			state;
 	struct xfs_trans		*tp;
 	struct xfs_btree_cur		*rcur = NULL;
+	struct xfs_mount		*mp = parent_tp->t_mountp;
 	bool				rt;
 
 	ASSERT(!test_bit(XFS_RUI_RECOVERED, &ruip->rui_flags));
@@ -462,6 +463,12 @@ xfs_rui_recover(
 			mp->m_rmap_maxlevels, 0, XFS_TRANS_RESERVE, &tp);
 	if (error)
 		return error;
+	/*
+	 * Recovery stashes all deferred ops during intent processing and
+	 * finishes them on completion. Transfer current dfops state to this
+	 * transaction and transfer the result back before we return.
+	 */
+	xfs_defer_move(tp, parent_tp);
 	rudp = xfs_trans_get_rud(tp, ruip);
 
 	for (i = 0; i < ruip->rui_format.rui_nextents; i++) {
@@ -511,10 +518,12 @@ xfs_rui_recover(
 
 	xfs_rmap_finish_one_cleanup(tp, rcur, error);
 	set_bit(XFS_RUI_RECOVERED, &ruip->rui_flags);
+	xfs_defer_move(parent_tp, tp);
 	error = xfs_trans_commit(tp);
 	return error;
 
 abort_error:
+	xfs_defer_move(parent_tp, tp);
 	xfs_rmap_finish_one_cleanup(tp, rcur, error);
 	xfs_trans_cancel(tp);
 	return error;
diff --git a/fs/xfs/xfs_rmap_item.h b/fs/xfs/xfs_rmap_item.h
index 7e482baa27f5..6e61d13f3c3c 100644
--- a/fs/xfs/xfs_rmap_item.h
+++ b/fs/xfs/xfs_rmap_item.h
@@ -84,6 +84,6 @@ int xfs_rui_copy_format(struct xfs_log_iovec *buf,
 		struct xfs_rui_log_format *dst_rui_fmt);
 void xfs_rui_item_free(struct xfs_rui_log_item *);
 void xfs_rui_release(struct xfs_rui_log_item *);
-int xfs_rui_recover(struct xfs_mount *mp, struct xfs_rui_log_item *ruip);
+int xfs_rui_recover(struct xfs_trans *parent_tp, struct xfs_rui_log_item *ruip);
 
 #endif	/* __XFS_RMAP_ITEM_H__ */
diff --git a/fs/xfs/xfs_trans_rmap.c b/fs/xfs/xfs_trans_rmap.c
index abc673c63dfb..6aed8841a63f 100644
--- a/fs/xfs/xfs_trans_rmap.c
+++ b/fs/xfs/xfs_trans_rmap.c
@@ -97,7 +97,7 @@ xfs_trans_log_finish_rmap_update(
 	int				error;
 
 	error = xfs_rmap_finish_one(tp, type, owner, whichfork, startoff,
-			startblock, blockcount, state, pcur);
+			startblock, blockcount, state, rt, pcur);
 
 	/*
 	 * Mark the transaction dirty, even on error. This ensures the

[PATCH 14/18] xfs: enable realtime rmap btree

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 fs/xfs/xfs_super.c |    7 -------
 1 file changed, 7 deletions(-)


diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c
index b73eb01f95f3..0b54e74b87ef 100644
--- a/fs/xfs/xfs_super.c
+++ b/fs/xfs/xfs_super.c
@@ -1798,13 +1798,6 @@ xfs_fs_fill_super(
 		goto out_filestream_unmount;
 	}
 
-	if (xfs_sb_version_hasrmapbt(&mp->m_sb) && mp->m_sb.sb_rblocks) {
-		xfs_alert(mp,
-	"reverse mapping btree not compatible with realtime device!");
-		error = -EINVAL;
-		goto out_filestream_unmount;
-	}
-
 	error = xfs_mountfs(mp);
 	if (error)
 		goto out_filestream_unmount;

[PATCH 15/18] xfs: wire up getfsmap to the realtime reverse mapping btree

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Connect the getfsmap ioctl to the realtime rmapbt.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 fs/xfs/xfs_fsmap.c |   66 ++++++++++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 64 insertions(+), 2 deletions(-)


diff --git a/fs/xfs/xfs_fsmap.c b/fs/xfs/xfs_fsmap.c
index 3d76a9e35870..b63435b1df2e 100644
--- a/fs/xfs/xfs_fsmap.c
+++ b/fs/xfs/xfs_fsmap.c
@@ -28,6 +28,7 @@
 #include "xfs_refcount_btree.h"
 #include "xfs_alloc_btree.h"
 #include "xfs_rtalloc.h"
+#include "xfs_rtrmap_btree.h"
 
 /* Convert an xfs_fsmap to an fsmap. */
 void
@@ -550,7 +551,65 @@ xfs_getfsmap_rtdev_rtbitmap(
 	return __xfs_getfsmap_rtdev(tp, keys, xfs_getfsmap_rtdev_rtbitmap_query,
 			info);
 }
-#endif /* CONFIG_XFS_RT */
+
+/* Transform a absolute-startblock rmap (rtdev, logdev) into a fsmap */
+STATIC int
+xfs_getfsmap_rtdev_helper(
+	struct xfs_btree_cur		*cur,
+	struct xfs_rmap_irec		*rec,
+	void				*priv)
+{
+	struct xfs_mount		*mp = cur->bc_mp;
+	struct xfs_getfsmap_info	*info = priv;
+	xfs_daddr_t			rec_daddr;
+
+	rec_daddr = XFS_FSB_TO_BB(mp, rec->rm_startblock);
+
+	return xfs_getfsmap_helper(cur->bc_tp, info, rec, rec_daddr);
+}
+
+/* Actually query the rtrmap btree. */
+STATIC int
+xfs_getfsmap_rtdev_rmapbt_query(
+	struct xfs_trans		*tp,
+	struct xfs_getfsmap_info	*info)
+{
+	struct xfs_mount		*mp = tp->t_mountp;
+	struct xfs_btree_cur		*bt_cur;
+	int				error;
+
+	/* Query the rtrmapbt */
+	xfs_ilock(mp->m_rrmapip, XFS_ILOCK_SHARED);
+	bt_cur = xfs_rtrmapbt_init_cursor(mp, tp, mp->m_rrmapip);
+	error = xfs_rmap_query_range(bt_cur, &info->low, &info->high,
+			xfs_getfsmap_rtdev_helper, info);
+	if (error)
+		goto err;
+
+	/* Report any gaps at the end of the rtrmapbt */
+	info->last = true;
+	error = xfs_getfsmap_rtdev_helper(bt_cur, &info->high, info);
+	if (error)
+		goto err;
+
+err:
+	xfs_btree_del_cursor(bt_cur, error);
+	xfs_iunlock(mp->m_rrmapip, XFS_ILOCK_SHARED);
+	return error;
+}
+
+/* Execute a getfsmap query against the realtime device rmapbt. */
+STATIC int
+xfs_getfsmap_rtdev_rmapbt(
+	struct xfs_trans		*tp,
+	struct xfs_fsmap		*keys,
+	struct xfs_getfsmap_info	*info)
+{
+	info->missing_owner = XFS_FMR_OWN_FREE;
+	return __xfs_getfsmap_rtdev(tp, keys, xfs_getfsmap_rtdev_rmapbt_query,
+			info);
+}
+#endif
 
 /* Execute a getfsmap query against the regular data device. */
 STATIC int
@@ -854,7 +913,10 @@ xfs_getfsmap(
 #ifdef CONFIG_XFS_RT
 	if (mp->m_rtdev_targp) {
 		handlers[2].dev = new_encode_dev(mp->m_rtdev_targp->bt_dev);
-		handlers[2].fn = xfs_getfsmap_rtdev_rtbitmap;
+		if (use_rmap)
+			handlers[2].fn = xfs_getfsmap_rtdev_rmapbt;
+		else
+			handlers[2].fn = xfs_getfsmap_rtdev_rtbitmap;
 	}
 #endif /* CONFIG_XFS_RT */
 

[PATCH 16/18] xfs: scrub the realtime rmapbt

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Check the realtime reverse mapping btree against the rtbitmap, and
modify the rtbitmap scrub to check against the rtrmapbt.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 fs/xfs/Makefile        |    5 ++
 fs/xfs/libxfs/xfs_fs.h |    3 +
 fs/xfs/scrub/bmap.c    |    1 
 fs/xfs/scrub/common.h  |    6 +++
 fs/xfs/scrub/rtrmap.c  |  109 ++++++++++++++++++++++++++++++++++++++++++++++++
 fs/xfs/scrub/scrub.c   |    7 +++
 fs/xfs/scrub/scrub.h   |    6 +++
 7 files changed, 135 insertions(+), 2 deletions(-)
 create mode 100644 fs/xfs/scrub/rtrmap.c


diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile
index c9b452646699..8aa72dc8be83 100644
--- a/fs/xfs/Makefile
+++ b/fs/xfs/Makefile
@@ -159,7 +159,10 @@ xfs-y				+= $(addprefix scrub/, \
 				   symlink.o \
 				   )
 
-xfs-$(CONFIG_XFS_RT)		+= scrub/rtbitmap.o
+xfs-$(CONFIG_XFS_RT)		+= $(addprefix scrub/, \
+				   rtbitmap.o \
+				   rtrmap.o \
+				   )
 xfs-$(CONFIG_XFS_QUOTA)		+= scrub/quota.o
 
 # online repair
diff --git a/fs/xfs/libxfs/xfs_fs.h b/fs/xfs/libxfs/xfs_fs.h
index 0f0e2948866c..11d8217cf005 100644
--- a/fs/xfs/libxfs/xfs_fs.h
+++ b/fs/xfs/libxfs/xfs_fs.h
@@ -503,9 +503,10 @@ struct xfs_scrub_metadata {
 #define XFS_SCRUB_TYPE_GQUOTA	22	/* group quotas */
 #define XFS_SCRUB_TYPE_PQUOTA	23	/* project quotas */
 #define XFS_SCRUB_TYPE_FSCOUNTERS 24	/* fs summary counters */
+#define XFS_SCRUB_TYPE_RTRMAPBT	25	/* realtime reverse mapping btree */
 
 /* Number of scrub subcommands. */
-#define XFS_SCRUB_TYPE_NR	25
+#define XFS_SCRUB_TYPE_NR	26
 
 /* i: Repair this metadata. */
 #define XFS_SCRUB_IFLAG_REPAIR		(1 << 0)
diff --git a/fs/xfs/scrub/bmap.c b/fs/xfs/scrub/bmap.c
index 6659f41e7b4c..b62845d016d2 100644
--- a/fs/xfs/scrub/bmap.c
+++ b/fs/xfs/scrub/bmap.c
@@ -648,6 +648,7 @@ xchk_bmap(
 	case XFS_DINODE_FMT_UUID:
 	case XFS_DINODE_FMT_DEV:
 	case XFS_DINODE_FMT_LOCAL:
+	case XFS_DINODE_FMT_RMAP:
 		/* No mappings to check. */
 		goto out;
 	case XFS_DINODE_FMT_EXTENTS:
diff --git a/fs/xfs/scrub/common.h b/fs/xfs/scrub/common.h
index 97da298abedd..6bccffddc832 100644
--- a/fs/xfs/scrub/common.h
+++ b/fs/xfs/scrub/common.h
@@ -89,12 +89,18 @@ int xchk_setup_parent(struct xfs_scrub *sc,
 			   struct xfs_inode *ip);
 #ifdef CONFIG_XFS_RT
 int xchk_setup_rt(struct xfs_scrub *sc, struct xfs_inode *ip);
+int xchk_setup_rtrmapbt(struct xfs_scrub *sc, struct xfs_inode *ip);
 #else
 static inline int
 xchk_setup_rt(struct xfs_scrub *sc, struct xfs_inode *ip)
 {
 	return -ENOENT;
 }
+static inline int
+xchk_setup_rtrmapbt(struct xfs_scrub *sc, struct xfs_inode *ip)
+{
+	return -ENOENT;
+}
 #endif
 #ifdef CONFIG_XFS_QUOTA
 int xchk_setup_quota(struct xfs_scrub *sc, struct xfs_inode *ip);
diff --git a/fs/xfs/scrub/rtrmap.c b/fs/xfs/scrub/rtrmap.c
new file mode 100644
index 000000000000..6dde3c026918
--- /dev/null
+++ b/fs/xfs/scrub/rtrmap.c
@@ -0,0 +1,109 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Copyright (C) 2018 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_rmap.h"
+#include "xfs_rmap_btree.h"
+#include "xfs_rtrmap_btree.h"
+#include "xfs_inode.h"
+#include "xfs_rtalloc.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/btree.h"
+#include "scrub/trace.h"
+
+/* Set us up with the realtime metadata and AG headers locked. */
+int
+xchk_setup_rtrmapbt(
+	struct xfs_scrub	*sc,
+	struct xfs_inode	*ip)
+{
+	struct xfs_mount	*mp = sc->mp;
+	int			lockmode;
+	int			error = 0;
+
+	if (sc->sm->sm_agno || sc->sm->sm_ino || sc->sm->sm_gen)
+		return -EINVAL;
+
+	error = xchk_setup_fs(sc, ip);
+	if (error)
+		return error;
+
+	lockmode = XFS_ILOCK_EXCL;
+	xfs_ilock(mp->m_rrmapip, lockmode);
+	xfs_trans_ijoin(sc->tp, mp->m_rrmapip, lockmode);
+
+	lockmode = XFS_ILOCK_EXCL | XFS_ILOCK_RTBITMAP;
+	xfs_ilock(mp->m_rbmip, lockmode);
+	xfs_trans_ijoin(sc->tp, mp->m_rbmip, lockmode);
+
+	return 0;
+}
+
+/* Realtime reverse mapping. */
+
+/* Scrub a realtime rmapbt record. */
+STATIC int
+xchk_rtrmapbt_helper(
+	struct xchk_btree	*bs,
+	union xfs_btree_rec	*rec)
+{
+	struct xfs_mount	*mp = bs->cur->bc_mp;
+	struct xfs_rmap_irec	irec;
+	bool			non_inode;
+	bool			is_bmbt;
+	bool			is_attr;
+	int			error;
+
+	error = xfs_rmap_btrec_to_irec(bs->cur, rec, &irec);
+	if (!xchk_btree_process_error(bs->sc, bs->cur, 0, &error))
+		goto out;
+
+	if (irec.rm_startblock + irec.rm_blockcount <= irec.rm_startblock ||
+	    (!xfs_verify_rtbno(mp, irec.rm_startblock) ||
+	     !xfs_verify_rtbno(mp, irec.rm_startblock +
+				irec.rm_blockcount - 1)))
+		xchk_btree_set_corrupt(bs->sc, bs->cur, 0);
+
+	non_inode = XFS_RMAP_NON_INODE_OWNER(irec.rm_owner);
+	is_bmbt = irec.rm_flags & XFS_RMAP_BMBT_BLOCK;
+	is_attr = irec.rm_flags & XFS_RMAP_ATTR_FORK;
+
+	if (is_bmbt || non_inode || is_attr)
+		xchk_btree_set_corrupt(bs->sc, bs->cur, 0);
+
+out:
+	return error;
+}
+
+/* Scrub the realtime rmap btree. */
+int
+xchk_rtrmapbt(
+	struct xfs_scrub	*sc)
+{
+	struct xfs_owner_info	oinfo;
+	struct xfs_mount	*mp = sc->mp;
+	struct xfs_btree_cur	*cur;
+	int			error;
+
+	cur = xfs_rtrmapbt_init_cursor(mp, sc->tp, mp->m_rrmapip);
+	xfs_rmap_ino_bmbt_owner(&oinfo, mp->m_sb.sb_rrmapino, XFS_DATA_FORK);
+	error = xchk_btree(sc, cur, xchk_rtrmapbt_helper, &oinfo, NULL);
+	xfs_btree_del_cursor(cur, error);
+
+	return error;
+}
diff --git a/fs/xfs/scrub/scrub.c b/fs/xfs/scrub/scrub.c
index 047261cedd4b..b7910238d4a4 100644
--- a/fs/xfs/scrub/scrub.c
+++ b/fs/xfs/scrub/scrub.c
@@ -361,6 +361,13 @@ static const struct xchk_meta_ops meta_scrub_ops[] = {
 		.scrub	= xchk_fscounters,
 		.repair	= xrep_fscounters,
 	},
+	[XFS_SCRUB_TYPE_RTRMAPBT] = {	/* realtime rmapbt */
+		.type	= ST_FS,
+		.setup	= xchk_setup_rtrmapbt,
+		.scrub	= xchk_rtrmapbt,
+		.has	= xfs_sb_version_hasrtrmapbt,
+		.repair	= xrep_notsupported,
+	},
 };
 
 /* This isn't a stable feature, warn once per day. */
diff --git a/fs/xfs/scrub/scrub.h b/fs/xfs/scrub/scrub.h
index 1ce261f648f3..470785292b46 100644
--- a/fs/xfs/scrub/scrub.h
+++ b/fs/xfs/scrub/scrub.h
@@ -99,6 +99,7 @@ int xchk_parent(struct xfs_scrub *sc);
 #ifdef CONFIG_XFS_RT
 int xchk_rtbitmap(struct xfs_scrub *sc);
 int xchk_rtsummary(struct xfs_scrub *sc);
+int xchk_rtrmapbt(struct xfs_scrub *sc);
 #else
 static inline int
 xchk_rtbitmap(struct xfs_scrub *sc)
@@ -110,6 +111,11 @@ xchk_rtsummary(struct xfs_scrub *sc)
 {
 	return -ENOENT;
 }
+static inline int
+xchk_rtrmapbt(struct xfs_scrub *sc)
+{
+	return -ENOENT;
+}
 #endif
 #ifdef CONFIG_XFS_QUOTA
 int xchk_quota(struct xfs_scrub *sc);

Re: [PATCH 16/18] xfs: scrub the realtime rmapbt

[Christoph Hellwig]

FYI, this fails in xfs_fs.h when applying to 4.19-rc1.

Maybe just push out the branch you are using somewhere if you get
around it.

[PATCH 17/18] xfs: cross-reference realtime bitmap to realtime rmapbt scrubber

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

When we're checking the realtime rmapbt, cross-reference the entries
with the realtime bitmap too.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 fs/xfs/scrub/rtrmap.c |   14 ++++++++++++++
 1 file changed, 14 insertions(+)


diff --git a/fs/xfs/scrub/rtrmap.c b/fs/xfs/scrub/rtrmap.c
index 6dde3c026918..0685a647d373 100644
--- a/fs/xfs/scrub/rtrmap.c
+++ b/fs/xfs/scrub/rtrmap.c
@@ -56,6 +56,16 @@ xchk_setup_rtrmapbt(
 
 /* Realtime reverse mapping. */
 
+/* Cross-reference with other metadata. */
+STATIC void
+xchk_rtrmapbt_xref(
+	struct xfs_scrub	*sc,
+	struct xfs_rmap_irec	*irec)
+{
+	xchk_xref_is_used_rt_space(sc, irec->rm_startblock,
+			irec->rm_blockcount);
+}
+
 /* Scrub a realtime rmapbt record. */
 STATIC int
 xchk_rtrmapbt_helper(
@@ -86,6 +96,10 @@ xchk_rtrmapbt_helper(
 	if (is_bmbt || non_inode || is_attr)
 		xchk_btree_set_corrupt(bs->sc, bs->cur, 0);
 
+	if (bs->sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		goto out;
+
+	xchk_rtrmapbt_xref(bs->sc, &irec);
 out:
 	return error;
 }

[PATCH 18/18] xfs: cross-reference the realtime rmapbt

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

When we're scrubbing the realtime metadata, cross-reference
the rtrmapt.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 fs/xfs/scrub/bmap.c     |   10 ++++++++++
 fs/xfs/scrub/common.c   |   22 ++++++++++++++++++++++
 fs/xfs/scrub/common.h   |    1 +
 fs/xfs/scrub/rmap.c     |   16 ++++++++--------
 fs/xfs/scrub/rtbitmap.c |   30 ++++++++++++++++++++++++++++++
 fs/xfs/scrub/scrub.h    |   12 ++++++------
 6 files changed, 77 insertions(+), 14 deletions(-)


diff --git a/fs/xfs/scrub/bmap.c b/fs/xfs/scrub/bmap.c
index b62845d016d2..2a7eff7a9b96 100644
--- a/fs/xfs/scrub/bmap.c
+++ b/fs/xfs/scrub/bmap.c
@@ -249,11 +249,21 @@ xchk_bmap_rt_extent_xref(
 	struct xfs_btree_cur	*cur,
 	struct xfs_bmbt_irec	*irec)
 {
+	int				error;
+
 	if (info->sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
 		return;
 
+	error = xchk_rt_init(info->sc, &info->sc->sa);
+	if (!xchk_fblock_process_error(info->sc, info->whichfork,
+			irec->br_startoff, &error))
+		return;
+
 	xchk_xref_is_used_rt_space(info->sc, irec->br_startblock,
 			irec->br_blockcount);
+	xchk_bmap_xref_rmap(info, irec, irec->br_startblock);
+
+	xchk_ag_free(info->sc, &info->sc->sa);
 }
 
 /* Cross-reference a single datadev extent record. */
diff --git a/fs/xfs/scrub/common.c b/fs/xfs/scrub/common.c
index d0f9b53a7ceb..204b572d0ce4 100644
--- a/fs/xfs/scrub/common.c
+++ b/fs/xfs/scrub/common.c
@@ -32,6 +32,7 @@
 #include "xfs_trans_priv.h"
 #include "xfs_attr.h"
 #include "xfs_reflink.h"
+#include "xfs_rtrmap_btree.h"
 #include "scrub/xfs_scrub.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
@@ -574,6 +575,27 @@ xchk_perag_get(
 		sa->pag = xfs_perag_get(mp, sa->agno);
 }
 
+/*
+ * For scrubbing a realtime file, grab the rtrmapt.  We follow the same
+ * resource release rules as xfs_scrub_ag_init.
+ */
+int
+xchk_rt_init(
+	struct xfs_scrub	*sc,
+	struct xchk_ag			*sa)
+{
+	memset(sa, 0, sizeof(*sa));
+	sa->agno = NULLAGNUMBER;
+	if (xfs_sb_version_hasrmapbt(&sc->mp->m_sb)) {
+		ASSERT(xfs_isilocked(sc->mp->m_rrmapip,
+				XFS_ILOCK_EXCL | XFS_ILOCK_SHARED));
+		sa->rmap_cur = xfs_rtrmapbt_init_cursor(sc->mp, sc->tp,
+				sc->mp->m_rrmapip);
+	}
+
+	return 0;
+}
+
 /* Per-scrubber setup functions */
 
 /*
diff --git a/fs/xfs/scrub/common.h b/fs/xfs/scrub/common.h
index 6bccffddc832..4fcc2ccd2401 100644
--- a/fs/xfs/scrub/common.h
+++ b/fs/xfs/scrub/common.h
@@ -113,6 +113,7 @@ xchk_setup_quota(struct xfs_scrub *sc, struct xfs_inode *ip)
 #endif
 int xchk_setup_fscounters(struct xfs_scrub *sc, struct xfs_inode *ip);
 
+int xchk_rt_init(struct xfs_scrub *sc, struct xchk_ag *sa);
 void xchk_ag_free(struct xfs_scrub *sc, struct xchk_ag *sa);
 int xchk_ag_init(struct xfs_scrub *sc, xfs_agnumber_t agno,
 		struct xchk_ag *sa);
diff --git a/fs/xfs/scrub/rmap.c b/fs/xfs/scrub/rmap.c
index 32dc329e5411..32df9ed77b74 100644
--- a/fs/xfs/scrub/rmap.c
+++ b/fs/xfs/scrub/rmap.c
@@ -189,8 +189,8 @@ xchk_rmapbt(
 static inline void
 xchk_xref_check_owner(
 	struct xfs_scrub	*sc,
-	xfs_agblock_t		bno,
-	xfs_extlen_t		len,
+	xfs_fsblock_t		bno,
+	xfs_filblks_t		len,
 	struct xfs_owner_info	*oinfo,
 	bool			should_have_rmap)
 {
@@ -212,8 +212,8 @@ xchk_xref_check_owner(
 void
 xchk_xref_is_owned_by(
 	struct xfs_scrub	*sc,
-	xfs_agblock_t		bno,
-	xfs_extlen_t		len,
+	xfs_fsblock_t		bno,
+	xfs_filblks_t		len,
 	struct xfs_owner_info	*oinfo)
 {
 	xchk_xref_check_owner(sc, bno, len, oinfo, true);
@@ -223,8 +223,8 @@ xchk_xref_is_owned_by(
 void
 xchk_xref_is_not_owned_by(
 	struct xfs_scrub	*sc,
-	xfs_agblock_t		bno,
-	xfs_extlen_t		len,
+	xfs_fsblock_t		bno,
+	xfs_filblks_t		len,
 	struct xfs_owner_info	*oinfo)
 {
 	xchk_xref_check_owner(sc, bno, len, oinfo, false);
@@ -234,8 +234,8 @@ xchk_xref_is_not_owned_by(
 void
 xchk_xref_has_no_owner(
 	struct xfs_scrub	*sc,
-	xfs_agblock_t		bno,
-	xfs_extlen_t		len)
+	xfs_fsblock_t		bno,
+	xfs_filblks_t		len)
 {
 	bool			has_rmap;
 	int			error;
diff --git a/fs/xfs/scrub/rtbitmap.c b/fs/xfs/scrub/rtbitmap.c
index 665d4bbb17cc..08041fc51fa1 100644
--- a/fs/xfs/scrub/rtbitmap.c
+++ b/fs/xfs/scrub/rtbitmap.c
@@ -15,12 +15,16 @@
 #include "xfs_log_format.h"
 #include "xfs_trans.h"
 #include "xfs_sb.h"
+#include "xfs_inode.h"
 #include "xfs_alloc.h"
 #include "xfs_rtalloc.h"
 #include "xfs_inode.h"
+#include "xfs_rmap.h"
+#include "xfs_rtrmap_btree.h"
 #include "scrub/xfs_scrub.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
+#include "scrub/btree.h"
 #include "scrub/trace.h"
 
 /* Set us up with the realtime metadata locked. */
@@ -39,11 +43,31 @@ xchk_setup_rt(
 	sc->ip = sc->mp->m_rbmip;
 	xfs_ilock(sc->ip, sc->ilock_flags);
 
+	if (xfs_sb_version_hasrmapbt(&sc->mp->m_sb)) {
+		unsigned int	lockmode = XFS_ILOCK_EXCL;
+
+		xfs_ilock(sc->mp->m_rrmapip, lockmode);
+		xfs_trans_ijoin(sc->tp, sc->mp->m_rrmapip, lockmode);
+	}
+
 	return 0;
 }
 
 /* Realtime bitmap. */
 
+/* Cross-reference rtbitmap entries with other metadata. */
+STATIC void
+xchk_rtbitmap_xref(
+	struct xfs_scrub	*sc,
+	xfs_rtblock_t		startblock,
+	xfs_rtblock_t		blockcount)
+{
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return;
+
+	xchk_xref_has_no_owner(sc, startblock, blockcount);
+}
+
 /* Scrub a free extent record from the realtime bitmap. */
 STATIC int
 xchk_rtbitmap_rec(
@@ -62,6 +86,8 @@ xchk_rtbitmap_rec(
 	    !xfs_verify_rtbno(sc->mp, startblock) ||
 	    !xfs_verify_rtbno(sc->mp, startblock + blockcount - 1))
 		xchk_fblock_set_corrupt(sc, XFS_DATA_FORK, 0);
+
+	xchk_rtbitmap_xref(sc, startblock, blockcount);
 	return 0;
 }
 
@@ -77,6 +103,10 @@ xchk_rtbitmap(
 	if (error || (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT))
 		return error;
 
+	error = xchk_rt_init(sc, &sc->sa);
+	if (error)
+		return error;
+
 	error = xfs_rtalloc_query_all(sc->tp, xchk_rtbitmap_rec, sc);
 	if (!xchk_fblock_process_error(sc, XFS_DATA_FORK, 0, &error))
 		goto out;
diff --git a/fs/xfs/scrub/scrub.h b/fs/xfs/scrub/scrub.h
index 470785292b46..87eed5bc6c80 100644
--- a/fs/xfs/scrub/scrub.h
+++ b/fs/xfs/scrub/scrub.h
@@ -135,12 +135,12 @@ void xchk_xref_is_not_inode_chunk(struct xfs_scrub *sc, xfs_agblock_t agbno,
 		xfs_extlen_t len);
 void xchk_xref_is_inode_chunk(struct xfs_scrub *sc, xfs_agblock_t agbno,
 		xfs_extlen_t len);
-void xchk_xref_is_owned_by(struct xfs_scrub *sc, xfs_agblock_t agbno,
-		xfs_extlen_t len, struct xfs_owner_info *oinfo);
-void xchk_xref_is_not_owned_by(struct xfs_scrub *sc, xfs_agblock_t agbno,
-		xfs_extlen_t len, struct xfs_owner_info *oinfo);
-void xchk_xref_has_no_owner(struct xfs_scrub *sc, xfs_agblock_t agbno,
-		xfs_extlen_t len);
+void xchk_xref_is_owned_by(struct xfs_scrub *sc, xfs_fsblock_t bno,
+		xfs_filblks_t len, struct xfs_owner_info *oinfo);
+void xchk_xref_is_not_owned_by(struct xfs_scrub *sc, xfs_fsblock_t bno,
+		xfs_filblks_t len, struct xfs_owner_info *oinfo);
+void xchk_xref_has_no_owner(struct xfs_scrub *sc, xfs_fsblock_t bno,
+		xfs_filblks_t len);
 void xchk_xref_is_cow_staging(struct xfs_scrub *sc, xfs_agblock_t bno,
 		xfs_extlen_t len);
 void xchk_xref_is_not_shared(struct xfs_scrub *sc, xfs_agblock_t bno,

Re: [RFC PATCH v9.1 00/18] xfs: add realtime reverse-mapping support

[Christoph Hellwig]

On Thu, Aug 30, 2018 at 11:31:36AM -0700, Darrick J. Wong wrote:
> If you're going to start using this mess, you probably ought to just
> pull from my github trees for kernel[1], xfsprogs[2], xfstests[3].  The
> kernel patches in the git trees should apply to 4.19-rc1; xfsprogs
> patches to for-next; and xfstest to master.

Ok, I should have read this before..