[PATCHBOMB 6.14 v6.1] xfs: realtime rmap and reflink

[PATCHBOMB 6.14 v6.1] xfs: realtime rmap and reflink

[Darrick J. Wong]

[resend, this time not as a reply to the v6.0 series]

Hi everyone,

Christoph and I have been working on getting the long-delayed port of
reverse mapping and reflink to the realtime device into mergeable shape.
With these changes, the realtime volume finally reaches feature parity
with the data device.  This is the base for building more functionality
into xfs, such as the zoned storage support that Christoph posted last
week.

Since v6.0 I've added Christoph's review tags, and folded in the
documentation updates requested during review.

The first patchset are bug fixes.  The final patchset are a few xfsprogs
patches that complete the metadir/rtgroups changes.  Both of these are
technically 6.13 material.

The second and third patchsets are all cleanups and refactoring so that
we can fully support having btrees rooted in an inode's data fork.  This
is necessary because the generic btree code only supports using the
immediate area as an internal tree node -- conversion from extents to
bmbt format only happens when there are too many leaf records to fit in
the immediate area.  Therefore, we need to remodel it to support storing
records in the immediate area.  We also need to be able to reserve space
for future btree expansion, so the second patchset enables tracking
per-inode reservations from the free space.

The fourth patchset ports reverse mapping btree to the realtime device,
which mainly consists of constructing a btree in an inode, linking the
inode into the metadata directory tree, and updating the log items to
handle rt rmap update log intent items.

The fifth patchset ports the refcount btree, block sharing, and copy on
write to the realtime device.

I dropped the fifth patchset from v6.0 because rtextsize>1 is a fringe
feature and doesn't need to be added right now.

Please have a look at the git tree links for code changes:

https://git.kernel.org/pub/scm/linux/kernel/git/djwong/xfs-linux.git/log/?h=realtime-reflink_2024-12-19
https://git.kernel.org/pub/scm/linux/kernel/git/djwong/xfsprogs-dev.git/log/?h=realtime-reflink_2024-12-19
https://git.kernel.org/pub/scm/linux/kernel/git/djwong/xfstests-dev.git/log/?h=realtime-reflink_2024-12-19

(fstests is still behind because I haven't rebased atop the parallel
fstests work)

These are the patches that haven't passed review yet.

[PATCHSET 1/5] xfs: bug fixes for 6.13
  [PATCH 2/2] xfs: release the dquot buf outside of qli_lock
[PATCHSET v6.1 3/5] xfs: enable in-core block reservation for rt
  [PATCH 1/2] xfs: prepare to reuse the dquot pointer space in struct
  [PATCH 2/2] xfs: allow inode-based btrees to reserve space in the
[PATCHSET v6.1 4/5] xfs: realtime reverse-mapping support
  [PATCH 27/37] xfs: online repair of realtime file bmaps
  [PATCH 32/37] xfs: online repair of the realtime rmap btree
  [PATCH 35/37] xfs: don't shut down the filesystem for media failures
  [PATCH 36/37] xfs: react to fsdax failure notifications on the rt
  [PATCH 37/37] xfs: enable realtime rmap btree
[PATCHSET v6.1 5/5] xfs: reflink on the realtime device
  [PATCH 28/43] xfs: scrub the realtime refcount btree
  [PATCH 42/43] xfs: fix CoW forks for realtime files
  [PATCH 43/43] xfs: enable realtime reflink
[PATCHSET v6.1] xfsprogs: last few bits of rtgroups patches
  [PATCH 1/3] xfs_db: drop the metadata checking code from blockget
  [PATCH 2/3] xfs_mdrestore: refactor open-coded fd/is_file into a
  [PATCH 3/3] xfs_mdrestore: restore rt group superblocks to realtime

--D

[PATCHSET 1/5] xfs: bug fixes for 6.13

[Darrick J. Wong]

Hi all,

Bug fixes for 6.13.

If you're going to start using this code, I strongly recommend pulling
from my git trees, which are linked below.

This has been running on the djcloud for months with no problems.  Enjoy!
Comments and questions are, as always, welcome.

--D

kernel git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfs-linux.git/log/?h=xfs-6.13-fixes

xfsprogs git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfsprogs-dev.git/log/?h=xfs-6.13-fixes
---
Commits in this patchset:
 * xfs: don't over-report free space or inodes in statvfs
 * xfs: release the dquot buf outside of qli_lock
---
 fs/xfs/xfs_dquot.c  |   12 ++++++++----
 fs/xfs/xfs_qm_bhv.c |   27 +++++++++++++++++----------
 2 files changed, 25 insertions(+), 14 deletions(-)

[PATCH 1/2] xfs: don't over-report free space or inodes in statvfs

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Emmanual Florac reports a strange occurrence when project quota limits
are enabled, free space is lower than the remaining quota, and someone
runs statvfs:

  # mkfs.xfs -f /dev/sda
  # mount /dev/sda /mnt -o prjquota
  # xfs_quota  -x -c 'limit -p bhard=2G 55' /mnt
  # mkdir /mnt/dir
  # xfs_io -c 'chproj 55' -c 'chattr +P' -c 'stat -vvvv' /mnt/dir
  # fallocate -l 19g /mnt/a
  # df /mnt /mnt/dir
  Filesystem      Size  Used Avail Use% Mounted on
  /dev/sda         20G   20G  345M  99% /mnt
  /dev/sda        2.0G     0  2.0G   0% /mnt

I think the bug here is that xfs_fill_statvfs_from_dquot unconditionally
assigns to f_bfree without checking that the filesystem has enough free
space to fill the remaining project quota.  However, this is a
longstanding behavior of xfs so it's unclear what to do here.

Cc: <stable@vger.kernel.org> # v2.6.18
Fixes: 932f2c323196c2 ("[XFS] statvfs component of directory/project quota support, code originally by Glen.")
Reported-by: Emmanuel Florac <eflorac@intellique.com>
Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/xfs_qm_bhv.c |   27 +++++++++++++++++----------
 1 file changed, 17 insertions(+), 10 deletions(-)


diff --git a/fs/xfs/xfs_qm_bhv.c b/fs/xfs/xfs_qm_bhv.c
index 847ba29630e9d8..db5b8afd9d1b97 100644
--- a/fs/xfs/xfs_qm_bhv.c
+++ b/fs/xfs/xfs_qm_bhv.c
@@ -32,21 +32,28 @@ xfs_fill_statvfs_from_dquot(
 	limit = blkres->softlimit ?
 		blkres->softlimit :
 		blkres->hardlimit;
-	if (limit && statp->f_blocks > limit) {
-		statp->f_blocks = limit;
-		statp->f_bfree = statp->f_bavail =
-			(statp->f_blocks > blkres->reserved) ?
-			 (statp->f_blocks - blkres->reserved) : 0;
+	if (limit) {
+		uint64_t	remaining = 0;
+
+		if (limit > blkres->reserved)
+			remaining = limit - blkres->reserved;
+
+		statp->f_blocks = min(statp->f_blocks, limit);
+		statp->f_bfree = min(statp->f_bfree, remaining);
+		statp->f_bavail = min(statp->f_bavail, remaining);
 	}
 
 	limit = dqp->q_ino.softlimit ?
 		dqp->q_ino.softlimit :
 		dqp->q_ino.hardlimit;
-	if (limit && statp->f_files > limit) {
-		statp->f_files = limit;
-		statp->f_ffree =
-			(statp->f_files > dqp->q_ino.reserved) ?
-			 (statp->f_files - dqp->q_ino.reserved) : 0;
+	if (limit) {
+		uint64_t	remaining = 0;
+
+		if (limit > dqp->q_ino.reserved)
+			remaining = limit - dqp->q_ino.reserved;
+
+		statp->f_files = min(statp->f_files, limit);
+		statp->f_ffree = min(statp->f_ffree, remaining);
 	}
 }
 

[PATCH 2/2] xfs: release the dquot buf outside of qli_lock

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Lai Yi reported a lockdep complaint about circular locking:

 Chain exists of:
   &lp->qli_lock --> &bch->bc_lock --> &l->lock

  Possible unsafe locking scenario:

        CPU0                    CPU1
        ----                    ----
   lock(&l->lock);
                                lock(&bch->bc_lock);
                                lock(&l->lock);
   lock(&lp->qli_lock);

I /think/ the problem here is that xfs_dquot_attach_buf during
quotacheck will release the buffer while it's holding the qli_lock.
Because this is a cached buffer, xfs_buf_rele_cached takes b_lock before
decrementing b_hold.  Other threads have taught lockdep that a locking
dependency chain is bp->b_lock -> bch->bc_lock -> l(ru)->lock; and that
another chain is l(ru)->lock -> lp->qli_lock.  Hence we do not want to
take b_lock while holding qli_lock.

Reported-by: syzbot+3126ab3db03db42e7a31@syzkaller.appspotmail.com
Cc: <stable@vger.kernel.org> # v6.13-rc3
Fixes: ca378189fdfa89 ("xfs: convert quotacheck to attach dquot buffers")
Tested-by: syzbot+3126ab3db03db42e7a31@syzkaller.appspotmail.com
Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
---
 fs/xfs/xfs_dquot.c |   12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)


diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c
index f11d475898f280..576b7755b1f1fc 100644
--- a/fs/xfs/xfs_dquot.c
+++ b/fs/xfs/xfs_dquot.c
@@ -1315,7 +1315,8 @@ xfs_dquot_read_buf(
 
 /*
  * Attach a dquot buffer to this dquot to avoid allocating a buffer during a
- * dqflush, since dqflush can be called from reclaim context.
+ * dqflush, since dqflush can be called from reclaim context.  Caller must hold
+ * the dqlock.
  */
 int
 xfs_dquot_attach_buf(
@@ -1336,13 +1337,16 @@ xfs_dquot_attach_buf(
 			return error;
 
 		/*
-		 * Attach the dquot to the buffer so that the AIL does not have
-		 * to read the dquot buffer to push this item.
+		 * Hold the dquot buffer so that we retain our ref to it after
+		 * detaching it from the transaction, then give that ref to the
+		 * dquot log item so that the AIL does not have to read the
+		 * dquot buffer to push this item.
 		 */
 		xfs_buf_hold(bp);
+		xfs_trans_brelse(tp, bp);
+
 		spin_lock(&qlip->qli_lock);
 		lip->li_buf = bp;
-		xfs_trans_brelse(tp, bp);
 	}
 	qlip->qli_dirty = true;
 	spin_unlock(&qlip->qli_lock);

Re: [PATCH 2/2] xfs: release the dquot buf outside of qli_lock

[Christoph Hellwig]

Yes, doing a xfs_buf_rele under any kind of low-level lock is probably
a bad idea:

Reviewed-by: Christoph Hellwig <hch@lst.de>

[PATCHSET v6.1 2/5] xfs: refactor btrees to support records in inode root

[Darrick J. Wong]

Hi all,

Amend the btree code to support storing btree rcords in the inode root,
because the current bmbt code does not support this.

If you're going to start using this code, I strongly recommend pulling
from my git trees, which are linked below.

This has been running on the djcloud for months with no problems.  Enjoy!
Comments and questions are, as always, welcome.

--D

kernel git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfs-linux.git/log/?h=btree-ifork-records

xfsprogs git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfsprogs-dev.git/log/?h=btree-ifork-records
---
Commits in this patchset:
 * xfs: tidy up xfs_iroot_realloc
 * xfs: refactor the inode fork memory allocation functions
 * xfs: make xfs_iroot_realloc take the new numrecs instead of deltas
 * xfs: make xfs_iroot_realloc a bmap btree function
 * xfs: tidy up xfs_bmap_broot_realloc a bit
 * xfs: hoist the node iroot update code out of xfs_btree_new_iroot
 * xfs: hoist the node iroot update code out of xfs_btree_kill_iroot
 * xfs: support storing records in the inode core root
---
 fs/xfs/libxfs/xfs_bmap.c          |    7 -
 fs/xfs/libxfs/xfs_bmap_btree.c    |  111 ++++++++++++
 fs/xfs/libxfs/xfs_bmap_btree.h    |    3 
 fs/xfs/libxfs/xfs_btree.c         |  333 ++++++++++++++++++++++++++++---------
 fs/xfs/libxfs/xfs_btree.h         |   18 ++
 fs/xfs/libxfs/xfs_btree_staging.c |    9 +
 fs/xfs/libxfs/xfs_inode_fork.c    |  170 ++++++-------------
 fs/xfs/libxfs/xfs_inode_fork.h    |    6 +
 8 files changed, 445 insertions(+), 212 deletions(-)

[PATCH 1/8] xfs: tidy up xfs_iroot_realloc

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Tidy up this function a bit before we start refactoring the memory
handling and move the function to the bmbt code.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_inode_fork.c |   83 +++++++++++++++++++---------------------
 1 file changed, 40 insertions(+), 43 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_inode_fork.c b/fs/xfs/libxfs/xfs_inode_fork.c
index 1158ca48626b71..7f865479c4159f 100644
--- a/fs/xfs/libxfs/xfs_inode_fork.c
+++ b/fs/xfs/libxfs/xfs_inode_fork.c
@@ -382,33 +382,32 @@ xfs_iformat_attr_fork(
  */
 void
 xfs_iroot_realloc(
-	xfs_inode_t		*ip,
+	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_ifork	*ifp = xfs_ifork_ptr(ip, whichfork);
 	struct xfs_btree_block	*new_broot;
-	int			new_max;
-	size_t			new_size;
 	char			*np;
 	char			*op;
+	size_t			new_size;
+	short			old_size = ifp->if_broot_bytes;
+	int			cur_max;
+	int			new_max;
 
 	/*
 	 * Handle the degenerate case quietly.
 	 */
-	if (rec_diff == 0) {
+	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) {
+		if (old_size == 0) {
 			new_size = xfs_bmap_broot_space_calc(mp, rec_diff);
 			ifp->if_broot = kmalloc(new_size,
 						GFP_KERNEL | __GFP_NOFAIL);
@@ -422,13 +421,13 @@ xfs_iroot_realloc(
 		 * 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, false);
+		cur_max = xfs_bmbt_maxrecs(mp, old_size, false);
 		new_max = cur_max + rec_diff;
 		new_size = xfs_bmap_broot_space_calc(mp, new_max);
 		ifp->if_broot = krealloc(ifp->if_broot, new_size,
 					 GFP_KERNEL | __GFP_NOFAIL);
 		op = (char *)xfs_bmap_broot_ptr_addr(mp, ifp->if_broot, 1,
-						     ifp->if_broot_bytes);
+						     old_size);
 		np = (char *)xfs_bmap_broot_ptr_addr(mp, ifp->if_broot, 1,
 						     (int)new_size);
 		ifp->if_broot_bytes = (int)new_size;
@@ -443,52 +442,50 @@ xfs_iroot_realloc(
 	 * 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, false);
+	ASSERT(ifp->if_broot != NULL && old_size > 0);
+	cur_max = xfs_bmbt_maxrecs(mp, old_size, false);
 	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 = kmalloc(new_size, GFP_KERNEL | __GFP_NOFAIL);
-		/*
-		 * First copy over the btree block header.
-		 */
-		memcpy(new_broot, ifp->if_broot,
-			xfs_bmbt_block_len(ip->i_mount));
-	} else {
-		new_broot = NULL;
+	if (new_size == 0) {
+		ifp->if_broot = NULL;
+		ifp->if_broot_bytes = 0;
+		return;
 	}
 
 	/*
-	 * Only copy the keys and pointers if there are any.
+	 * Shrink the btree root by allocating a smaller object and copying the
+	 * fields from the old object to the new object.  krealloc does nothing
+	 * if we realloc downwards.
 	 */
-	if (new_max > 0) {
-		/*
-		 * First copy the keys.
-		 */
-		op = (char *)xfs_bmbt_key_addr(mp, ifp->if_broot, 1);
-		np = (char *)xfs_bmbt_key_addr(mp, new_broot, 1);
-		memcpy(np, op, new_max * (uint)sizeof(xfs_bmbt_key_t));
+	new_broot = kmalloc(new_size, GFP_KERNEL | __GFP_NOFAIL);
+	/*
+	 * First copy over the btree block header.
+	 */
+	memcpy(new_broot, ifp->if_broot, xfs_bmbt_block_len(ip->i_mount));
+
+	/*
+	 * First copy the keys.
+	 */
+	op = (char *)xfs_bmbt_key_addr(mp, ifp->if_broot, 1);
+	np = (char *)xfs_bmbt_key_addr(mp, new_broot, 1);
+	memcpy(np, op, new_max * (uint)sizeof(xfs_bmbt_key_t));
+
+	/*
+	 * Then copy the pointers.
+	 */
+	op = (char *)xfs_bmap_broot_ptr_addr(mp, ifp->if_broot, 1, old_size);
+	np = (char *)xfs_bmap_broot_ptr_addr(mp, new_broot, 1, (int)new_size);
+	memcpy(np, op, new_max * (uint)sizeof(xfs_fsblock_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));
-	}
 	kfree(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_inode_fork_size(ip, whichfork));
-	return;
+	ASSERT(xfs_bmap_bmdr_space(ifp->if_broot) <=
+	       xfs_inode_fork_size(ip, whichfork));
 }
 
 

[PATCH 2/8] xfs: refactor the inode fork memory allocation functions

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Hoist the code that allocates, frees, and reallocates if_broot into a
single xfs_iroot_krealloc function.  Eventually we're going to push
xfs_iroot_realloc into the btree ops structure to handle multiple
inode-rooted btrees, but first let's separate out the bits that should
stay in xfs_inode_fork.c.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_inode_fork.c |  116 +++++++++++++++++++++++++++-------------
 fs/xfs/libxfs/xfs_inode_fork.h |    5 ++
 2 files changed, 82 insertions(+), 39 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_inode_fork.c b/fs/xfs/libxfs/xfs_inode_fork.c
index 7f865479c4159f..294c3c5556836e 100644
--- a/fs/xfs/libxfs/xfs_inode_fork.c
+++ b/fs/xfs/libxfs/xfs_inode_fork.c
@@ -178,7 +178,7 @@ xfs_iformat_btree(
 	struct xfs_mount	*mp = ip->i_mount;
 	xfs_bmdr_block_t	*dfp;
 	struct xfs_ifork	*ifp;
-	/* REFERENCED */
+	struct xfs_btree_block	*broot;
 	int			nrecs;
 	int			size;
 	int			level;
@@ -211,16 +211,13 @@ xfs_iformat_btree(
 		return -EFSCORRUPTED;
 	}
 
-	ifp->if_broot_bytes = size;
-	ifp->if_broot = kmalloc(size,
-				GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL);
-	ASSERT(ifp->if_broot != NULL);
+	broot = xfs_broot_alloc(ifp, size);
 	/*
 	 * Copy and convert from the on-disk structure
 	 * to the in-memory structure.
 	 */
 	xfs_bmdr_to_bmbt(ip, dfp, XFS_DFORK_SIZE(dip, ip->i_mount, whichfork),
-			 ifp->if_broot, size);
+			 broot, size);
 
 	ifp->if_bytes = 0;
 	ifp->if_data = NULL;
@@ -362,6 +359,69 @@ xfs_iformat_attr_fork(
 	return error;
 }
 
+/*
+ * Allocate the if_broot component of an inode fork so that it is @new_size
+ * bytes in size, using __GFP_NOLOCKDEP like all the other code that
+ * initializes a broot during inode load.  Returns if_broot.
+ */
+struct xfs_btree_block *
+xfs_broot_alloc(
+	struct xfs_ifork	*ifp,
+	size_t			new_size)
+{
+	ASSERT(ifp->if_broot == NULL);
+
+	ifp->if_broot = kmalloc(new_size,
+				GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL);
+	ifp->if_broot_bytes = new_size;
+	return ifp->if_broot;
+}
+
+/*
+ * Reallocate the if_broot component of an inode fork so that it is @new_size
+ * bytes in size.  Returns if_broot.
+ */
+struct xfs_btree_block *
+xfs_broot_realloc(
+	struct xfs_ifork	*ifp,
+	size_t			new_size)
+{
+	/* No size change?  No action needed. */
+	if (new_size == ifp->if_broot_bytes)
+		return ifp->if_broot;
+
+	/* New size is zero, free it. */
+	if (new_size == 0) {
+		ifp->if_broot_bytes = 0;
+		kfree(ifp->if_broot);
+		ifp->if_broot = NULL;
+		return NULL;
+	}
+
+	/*
+	 * Shrinking the iroot means we allocate a new smaller object and copy
+	 * it.  We don't trust krealloc not to nop on realloc-down.
+	 */
+	if (ifp->if_broot_bytes > 0 && ifp->if_broot_bytes > new_size) {
+		struct xfs_btree_block	*old_broot = ifp->if_broot;
+
+		ifp->if_broot = kmalloc(new_size, GFP_KERNEL | __GFP_NOFAIL);
+		ifp->if_broot_bytes = new_size;
+		memcpy(ifp->if_broot, old_broot, new_size);
+		kfree(old_broot);
+		return ifp->if_broot;
+	}
+
+	/*
+	 * Growing the iroot means we can krealloc.  This may get us the same
+	 * object.
+	 */
+	ifp->if_broot = krealloc(ifp->if_broot, new_size,
+			GFP_KERNEL | __GFP_NOFAIL);
+	ifp->if_broot_bytes = new_size;
+	return ifp->if_broot;
+}
+
 /*
  * Reallocate the space for if_broot based on the number of records
  * being added or deleted as indicated in rec_diff.  Move the records
@@ -388,7 +448,6 @@ xfs_iroot_realloc(
 {
 	struct xfs_mount	*mp = ip->i_mount;
 	struct xfs_ifork	*ifp = xfs_ifork_ptr(ip, whichfork);
-	struct xfs_btree_block	*new_broot;
 	char			*np;
 	char			*op;
 	size_t			new_size;
@@ -409,9 +468,7 @@ xfs_iroot_realloc(
 		 */
 		if (old_size == 0) {
 			new_size = xfs_bmap_broot_space_calc(mp, rec_diff);
-			ifp->if_broot = kmalloc(new_size,
-						GFP_KERNEL | __GFP_NOFAIL);
-			ifp->if_broot_bytes = (int)new_size;
+			xfs_broot_realloc(ifp, new_size);
 			return;
 		}
 
@@ -424,13 +481,12 @@ xfs_iroot_realloc(
 		cur_max = xfs_bmbt_maxrecs(mp, old_size, false);
 		new_max = cur_max + rec_diff;
 		new_size = xfs_bmap_broot_space_calc(mp, new_max);
-		ifp->if_broot = krealloc(ifp->if_broot, new_size,
-					 GFP_KERNEL | __GFP_NOFAIL);
+
+		xfs_broot_realloc(ifp, new_size);
 		op = (char *)xfs_bmap_broot_ptr_addr(mp, ifp->if_broot, 1,
 						     old_size);
 		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_inode_fork_size(ip, whichfork));
 		memmove(np, op, cur_max * (uint)sizeof(xfs_fsblock_t));
@@ -451,39 +507,21 @@ xfs_iroot_realloc(
 	else
 		new_size = 0;
 	if (new_size == 0) {
-		ifp->if_broot = NULL;
-		ifp->if_broot_bytes = 0;
+		xfs_broot_realloc(ifp, 0);
 		return;
 	}
 
 	/*
-	 * Shrink the btree root by allocating a smaller object and copying the
-	 * fields from the old object to the new object.  krealloc does nothing
-	 * if we realloc downwards.
-	 */
-	new_broot = kmalloc(new_size, GFP_KERNEL | __GFP_NOFAIL);
-	/*
-	 * First copy over the btree block header.
-	 */
-	memcpy(new_broot, ifp->if_broot, xfs_bmbt_block_len(ip->i_mount));
-
-	/*
-	 * First copy the keys.
-	 */
-	op = (char *)xfs_bmbt_key_addr(mp, ifp->if_broot, 1);
-	np = (char *)xfs_bmbt_key_addr(mp, new_broot, 1);
-	memcpy(np, op, new_max * (uint)sizeof(xfs_bmbt_key_t));
-
-	/*
-	 * Then copy the pointers.
+	 * Shrink the btree root by moving the bmbt pointers, since they are
+	 * not butted up against the btree block header, then reallocating
+	 * broot.
 	 */
 	op = (char *)xfs_bmap_broot_ptr_addr(mp, ifp->if_broot, 1, old_size);
-	np = (char *)xfs_bmap_broot_ptr_addr(mp, new_broot, 1, (int)new_size);
-	memcpy(np, op, new_max * (uint)sizeof(xfs_fsblock_t));
+	np = (char *)xfs_bmap_broot_ptr_addr(mp, ifp->if_broot, 1,
+					     (int)new_size);
+	memmove(np, op, new_max * (uint)sizeof(xfs_fsblock_t));
 
-	kfree(ifp->if_broot);
-	ifp->if_broot = new_broot;
-	ifp->if_broot_bytes = (int)new_size;
+	xfs_broot_realloc(ifp, new_size);
 	ASSERT(xfs_bmap_bmdr_space(ifp->if_broot) <=
 	       xfs_inode_fork_size(ip, whichfork));
 }
diff --git a/fs/xfs/libxfs/xfs_inode_fork.h b/fs/xfs/libxfs/xfs_inode_fork.h
index 2373d12fd474f0..e3c5c9121044fd 100644
--- a/fs/xfs/libxfs/xfs_inode_fork.h
+++ b/fs/xfs/libxfs/xfs_inode_fork.h
@@ -170,6 +170,11 @@ void		xfs_iflush_fork(struct xfs_inode *, struct xfs_dinode *,
 void		xfs_idestroy_fork(struct xfs_ifork *ifp);
 void *		xfs_idata_realloc(struct xfs_inode *ip, int64_t byte_diff,
 				int whichfork);
+struct xfs_btree_block *xfs_broot_alloc(struct xfs_ifork *ifp,
+				size_t new_size);
+struct xfs_btree_block *xfs_broot_realloc(struct xfs_ifork *ifp,
+				size_t new_size);
+
 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 *,

[PATCH 3/8] xfs: make xfs_iroot_realloc take the new numrecs instead of deltas

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Change the calling signature of xfs_iroot_realloc to take the ifork and
the new number of records in the btree block, not a diff against the
current number.  This will make the callsites easier to understand.

Note that this function is misnamed because it is very specific to the
single type of inode-rooted btree supported.  This will be addressed in
a subsequent patch.

Return the new btree root to reduce the amount of code clutter.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_bmap.c       |    7 +--
 fs/xfs/libxfs/xfs_btree.c      |   25 ++++--------
 fs/xfs/libxfs/xfs_inode_fork.c |   83 ++++++++++++++++++----------------------
 fs/xfs/libxfs/xfs_inode_fork.h |    3 +
 4 files changed, 51 insertions(+), 67 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c
index 5255f93bae31f3..8ab38f07cb78dd 100644
--- a/fs/xfs/libxfs/xfs_bmap.c
+++ b/fs/xfs/libxfs/xfs_bmap.c
@@ -615,7 +615,7 @@ xfs_bmap_btree_to_extents(
 	xfs_trans_binval(tp, cbp);
 	if (cur->bc_levels[0].bp == cbp)
 		cur->bc_levels[0].bp = NULL;
-	xfs_iroot_realloc(ip, -1, whichfork);
+	xfs_iroot_realloc(ip, whichfork, 0);
 	ASSERT(ifp->if_broot == NULL);
 	ifp->if_format = XFS_DINODE_FMT_EXTENTS;
 	*logflagsp |= XFS_ILOG_CORE | xfs_ilog_fext(whichfork);
@@ -659,12 +659,11 @@ xfs_bmap_extents_to_btree(
 	 * Make space in the inode incore. This needs to be undone if we fail
 	 * to expand the root.
 	 */
-	xfs_iroot_realloc(ip, 1, whichfork);
+	block = xfs_iroot_realloc(ip, whichfork, 1);
 
 	/*
 	 * Fill in the root.
 	 */
-	block = ifp->if_broot;
 	xfs_bmbt_init_block(ip, block, NULL, 1, 1);
 	/*
 	 * Need a cursor.  Can't allocate until bb_level is filled in.
@@ -746,7 +745,7 @@ xfs_bmap_extents_to_btree(
 out_unreserve_dquot:
 	xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT, -1L);
 out_root_realloc:
-	xfs_iroot_realloc(ip, -1, whichfork);
+	xfs_iroot_realloc(ip, whichfork, 0);
 	ifp->if_format = XFS_DINODE_FMT_EXTENTS;
 	ASSERT(ifp->if_broot == NULL);
 	xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c
index 68ee1c299c25fd..5714bec26c2084 100644
--- a/fs/xfs/libxfs/xfs_btree.c
+++ b/fs/xfs/libxfs/xfs_btree.c
@@ -3161,9 +3161,7 @@ xfs_btree_new_iroot(
 
 	xfs_btree_copy_ptrs(cur, pp, &nptr, 1);
 
-	xfs_iroot_realloc(cur->bc_ino.ip,
-			  1 - xfs_btree_get_numrecs(cblock),
-			  cur->bc_ino.whichfork);
+	xfs_iroot_realloc(cur->bc_ino.ip, cur->bc_ino.whichfork, 1);
 
 	xfs_btree_setbuf(cur, level, cbp);
 
@@ -3347,7 +3345,8 @@ 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_ino.whichfork);
+			xfs_iroot_realloc(ip, cur->bc_ino.whichfork,
+					numrecs + 1);
 			*stat = 1;
 		} else {
 			/* A root block that needs replacing */
@@ -3705,9 +3704,7 @@ STATIC int
 xfs_btree_kill_iroot(
 	struct xfs_btree_cur	*cur)
 {
-	int			whichfork = cur->bc_ino.whichfork;
 	struct xfs_inode	*ip = cur->bc_ino.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;
@@ -3716,7 +3713,6 @@ xfs_btree_kill_iroot(
 	union xfs_btree_ptr	*cpp;
 	struct xfs_buf		*cbp;
 	int			level;
-	int			index;
 	int			numrecs;
 	int			error;
 #ifdef DEBUG
@@ -3762,14 +3758,10 @@ xfs_btree_kill_iroot(
 	ASSERT(xfs_btree_ptr_is_null(cur, &ptr));
 #endif
 
-	index = numrecs - cur->bc_ops->get_maxrecs(cur, level);
-	if (index) {
-		xfs_iroot_realloc(cur->bc_ino.ip, index,
-				  cur->bc_ino.whichfork);
-		block = ifp->if_broot;
-	}
+	block = xfs_iroot_realloc(cur->bc_ino.ip, cur->bc_ino.whichfork,
+			numrecs);
 
-	be16_add_cpu(&block->bb_numrecs, index);
+	block->bb_numrecs = be16_to_cpu(numrecs);
 	ASSERT(block->bb_numrecs == cblock->bb_numrecs);
 
 	kp = xfs_btree_key_addr(cur, 1, block);
@@ -3949,10 +3941,11 @@ xfs_btree_delrec(
 	/*
 	 * We're at the root level.  First, shrink the root block in-memory.
 	 * Try to get rid of the next level down.  If we can't then there's
-	 * nothing left to do.
+	 * nothing left to do.  numrecs was decremented above.
 	 */
 	if (xfs_btree_at_iroot(cur, level)) {
-		xfs_iroot_realloc(cur->bc_ino.ip, -1, cur->bc_ino.whichfork);
+		xfs_iroot_realloc(cur->bc_ino.ip, cur->bc_ino.whichfork,
+				numrecs);
 
 		error = xfs_btree_kill_iroot(cur);
 		if (error)
diff --git a/fs/xfs/libxfs/xfs_inode_fork.c b/fs/xfs/libxfs/xfs_inode_fork.c
index 294c3c5556836e..53bfdf422ad820 100644
--- a/fs/xfs/libxfs/xfs_inode_fork.c
+++ b/fs/xfs/libxfs/xfs_inode_fork.c
@@ -423,12 +423,10 @@ xfs_broot_realloc(
 }
 
 /*
- * 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.
+ * Reallocate the space for if_broot based on the number of records.  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
@@ -437,40 +435,47 @@ xfs_broot_realloc(
  * 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.
+ * whichfork -- which inode fork to change
+ * new_numrecs -- the new number of records requested for the if_broot array
+ *
+ * Returns the incore btree root block.
  */
-void
+struct xfs_btree_block *
 xfs_iroot_realloc(
 	struct xfs_inode	*ip,
-	int			rec_diff,
-	int			whichfork)
+	int			whichfork,
+	unsigned int		new_numrecs)
 {
 	struct xfs_mount	*mp = ip->i_mount;
 	struct xfs_ifork	*ifp = xfs_ifork_ptr(ip, whichfork);
 	char			*np;
 	char			*op;
-	size_t			new_size;
-	short			old_size = ifp->if_broot_bytes;
-	int			cur_max;
-	int			new_max;
+	unsigned int		new_size;
+	unsigned int		old_size = ifp->if_broot_bytes;
 
 	/*
-	 * Handle the degenerate case quietly.
+	 * Block mapping btrees do not support storing zero records; if this
+	 * happens, the fork is being changed to FMT_EXTENTS.  Free the broot
+	 * and get out.
 	 */
-	if (rec_diff == 0)
-		return;
+	if (new_numrecs == 0)
+		return xfs_broot_realloc(ifp, 0);
+
+	new_size = xfs_bmap_broot_space_calc(mp, new_numrecs);
+
+	/* Handle the nop case quietly. */
+	if (new_size == old_size)
+		return ifp->if_broot;
+
+	if (new_size > old_size) {
+		unsigned int	old_numrecs;
 
-	if (rec_diff > 0) {
 		/*
 		 * If there wasn't any memory allocated before, just
 		 * allocate it now and get out.
 		 */
-		if (old_size == 0) {
-			new_size = xfs_bmap_broot_space_calc(mp, rec_diff);
-			xfs_broot_realloc(ifp, new_size);
-			return;
-		}
+		if (old_size == 0)
+			return xfs_broot_realloc(ifp, new_size);
 
 		/*
 		 * If there is already an existing if_broot, then we need
@@ -478,10 +483,7 @@ xfs_iroot_realloc(
 		 * location.  The records don't change location because
 		 * they are kept butted up against the btree block header.
 		 */
-		cur_max = xfs_bmbt_maxrecs(mp, old_size, false);
-		new_max = cur_max + rec_diff;
-		new_size = xfs_bmap_broot_space_calc(mp, new_max);
-
+		old_numrecs = xfs_bmbt_maxrecs(mp, old_size, false);
 		xfs_broot_realloc(ifp, new_size);
 		op = (char *)xfs_bmap_broot_ptr_addr(mp, ifp->if_broot, 1,
 						     old_size);
@@ -489,27 +491,15 @@ xfs_iroot_realloc(
 						     (int)new_size);
 		ASSERT(xfs_bmap_bmdr_space(ifp->if_broot) <=
 			xfs_inode_fork_size(ip, whichfork));
-		memmove(np, op, cur_max * (uint)sizeof(xfs_fsblock_t));
-		return;
+		memmove(np, op, old_numrecs * (uint)sizeof(xfs_fsblock_t));
+		return ifp->if_broot;
 	}
 
 	/*
-	 * 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.
+	 * We're reducing, but not totally eliminating, numrecs.  In this case,
+	 * we are shrinking the if_broot buffer, so it must already exist.
 	 */
-	ASSERT(ifp->if_broot != NULL && old_size > 0);
-	cur_max = xfs_bmbt_maxrecs(mp, old_size, false);
-	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) {
-		xfs_broot_realloc(ifp, 0);
-		return;
-	}
+	ASSERT(ifp->if_broot != NULL && old_size > 0 && new_size > 0);
 
 	/*
 	 * Shrink the btree root by moving the bmbt pointers, since they are
@@ -519,11 +509,12 @@ xfs_iroot_realloc(
 	op = (char *)xfs_bmap_broot_ptr_addr(mp, ifp->if_broot, 1, old_size);
 	np = (char *)xfs_bmap_broot_ptr_addr(mp, ifp->if_broot, 1,
 					     (int)new_size);
-	memmove(np, op, new_max * (uint)sizeof(xfs_fsblock_t));
+	memmove(np, op, new_numrecs * (uint)sizeof(xfs_fsblock_t));
 
 	xfs_broot_realloc(ifp, new_size);
 	ASSERT(xfs_bmap_bmdr_space(ifp->if_broot) <=
 	       xfs_inode_fork_size(ip, whichfork));
+	return ifp->if_broot;
 }
 
 
diff --git a/fs/xfs/libxfs/xfs_inode_fork.h b/fs/xfs/libxfs/xfs_inode_fork.h
index e3c5c9121044fd..d05eb0bad864e1 100644
--- a/fs/xfs/libxfs/xfs_inode_fork.h
+++ b/fs/xfs/libxfs/xfs_inode_fork.h
@@ -175,7 +175,8 @@ struct xfs_btree_block *xfs_broot_alloc(struct xfs_ifork *ifp,
 struct xfs_btree_block *xfs_broot_realloc(struct xfs_ifork *ifp,
 				size_t new_size);
 
-void		xfs_iroot_realloc(struct xfs_inode *, int, int);
+struct xfs_btree_block *xfs_iroot_realloc(struct xfs_inode *ip, int whichfork,
+				unsigned int new_numrecs);
 int		xfs_iread_extents(struct xfs_trans *, struct xfs_inode *, int);
 int		xfs_iextents_copy(struct xfs_inode *, struct xfs_bmbt_rec *,
 				  int);

[PATCH 4/8] xfs: make xfs_iroot_realloc a bmap btree function

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Move the inode fork btree root reallocation function part of the btree
ops because it's now mostly bmbt-specific code.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_bmap.c       |    6 +-
 fs/xfs/libxfs/xfs_bmap_btree.c |  104 ++++++++++++++++++++++++++++++++++++++++
 fs/xfs/libxfs/xfs_bmap_btree.h |    3 +
 fs/xfs/libxfs/xfs_btree.c      |   11 ++--
 fs/xfs/libxfs/xfs_btree.h      |   16 ++++++
 fs/xfs/libxfs/xfs_inode_fork.c |   96 -------------------------------------
 fs/xfs/libxfs/xfs_inode_fork.h |    2 -
 7 files changed, 130 insertions(+), 108 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c
index 8ab38f07cb78dd..0842577755f7bb 100644
--- a/fs/xfs/libxfs/xfs_bmap.c
+++ b/fs/xfs/libxfs/xfs_bmap.c
@@ -615,7 +615,7 @@ xfs_bmap_btree_to_extents(
 	xfs_trans_binval(tp, cbp);
 	if (cur->bc_levels[0].bp == cbp)
 		cur->bc_levels[0].bp = NULL;
-	xfs_iroot_realloc(ip, whichfork, 0);
+	xfs_bmap_broot_realloc(ip, whichfork, 0);
 	ASSERT(ifp->if_broot == NULL);
 	ifp->if_format = XFS_DINODE_FMT_EXTENTS;
 	*logflagsp |= XFS_ILOG_CORE | xfs_ilog_fext(whichfork);
@@ -659,7 +659,7 @@ xfs_bmap_extents_to_btree(
 	 * Make space in the inode incore. This needs to be undone if we fail
 	 * to expand the root.
 	 */
-	block = xfs_iroot_realloc(ip, whichfork, 1);
+	block = xfs_bmap_broot_realloc(ip, whichfork, 1);
 
 	/*
 	 * Fill in the root.
@@ -745,7 +745,7 @@ xfs_bmap_extents_to_btree(
 out_unreserve_dquot:
 	xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT, -1L);
 out_root_realloc:
-	xfs_iroot_realloc(ip, whichfork, 0);
+	xfs_bmap_broot_realloc(ip, whichfork, 0);
 	ifp->if_format = XFS_DINODE_FMT_EXTENTS;
 	ASSERT(ifp->if_broot == NULL);
 	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 3464be771f95d8..22cf2059d54dd4 100644
--- a/fs/xfs/libxfs/xfs_bmap_btree.c
+++ b/fs/xfs/libxfs/xfs_bmap_btree.c
@@ -516,6 +516,109 @@ xfs_bmbt_keys_contiguous(
 				 be64_to_cpu(key2->bmbt.br_startoff));
 }
 
+/*
+ * Reallocate the space for if_broot based on the number of records.  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
+ * whichfork -- which inode fork to change
+ * new_numrecs -- the new number of records requested for the if_broot array
+ *
+ * Returns the incore btree root block.
+ */
+struct xfs_btree_block *
+xfs_bmap_broot_realloc(
+	struct xfs_inode	*ip,
+	int			whichfork,
+	unsigned int		new_numrecs)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_ifork	*ifp = xfs_ifork_ptr(ip, whichfork);
+	char			*np;
+	char			*op;
+	unsigned int		new_size;
+	unsigned int		old_size = ifp->if_broot_bytes;
+
+	/*
+	 * Block mapping btrees do not support storing zero records; if this
+	 * happens, the fork is being changed to FMT_EXTENTS.  Free the broot
+	 * and get out.
+	 */
+	if (new_numrecs == 0)
+		return xfs_broot_realloc(ifp, 0);
+
+	new_size = xfs_bmap_broot_space_calc(mp, new_numrecs);
+
+	/* Handle the nop case quietly. */
+	if (new_size == old_size)
+		return ifp->if_broot;
+
+	if (new_size > old_size) {
+		unsigned int	old_numrecs;
+
+		/*
+		 * If there wasn't any memory allocated before, just
+		 * allocate it now and get out.
+		 */
+		if (old_size == 0)
+			return xfs_broot_realloc(ifp, new_size);
+
+		/*
+		 * 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.
+		 */
+		old_numrecs = xfs_bmbt_maxrecs(mp, old_size, false);
+		xfs_broot_realloc(ifp, new_size);
+		op = (char *)xfs_bmap_broot_ptr_addr(mp, ifp->if_broot, 1,
+						     old_size);
+		np = (char *)xfs_bmap_broot_ptr_addr(mp, ifp->if_broot, 1,
+						     (int)new_size);
+		ASSERT(xfs_bmap_bmdr_space(ifp->if_broot) <=
+			xfs_inode_fork_size(ip, whichfork));
+		memmove(np, op, old_numrecs * (uint)sizeof(xfs_fsblock_t));
+		return ifp->if_broot;
+	}
+
+	/*
+	 * We're reducing, but not totally eliminating, numrecs.  In this case,
+	 * we are shrinking the if_broot buffer, so it must already exist.
+	 */
+	ASSERT(ifp->if_broot != NULL && old_size > 0 && new_size > 0);
+
+	/*
+	 * Shrink the btree root by moving the bmbt pointers, since they are
+	 * not butted up against the btree block header, then reallocating
+	 * broot.
+	 */
+	op = (char *)xfs_bmap_broot_ptr_addr(mp, ifp->if_broot, 1, old_size);
+	np = (char *)xfs_bmap_broot_ptr_addr(mp, ifp->if_broot, 1,
+					     (int)new_size);
+	memmove(np, op, new_numrecs * (uint)sizeof(xfs_fsblock_t));
+
+	xfs_broot_realloc(ifp, new_size);
+	ASSERT(xfs_bmap_bmdr_space(ifp->if_broot) <=
+	       xfs_inode_fork_size(ip, whichfork));
+	return ifp->if_broot;
+}
+
+static struct xfs_btree_block *
+xfs_bmbt_broot_realloc(
+	struct xfs_btree_cur	*cur,
+	unsigned int		new_numrecs)
+{
+	return xfs_bmap_broot_realloc(cur->bc_ino.ip, cur->bc_ino.whichfork,
+			new_numrecs);
+}
+
 const struct xfs_btree_ops xfs_bmbt_ops = {
 	.name			= "bmap",
 	.type			= XFS_BTREE_TYPE_INODE,
@@ -543,6 +646,7 @@ const struct xfs_btree_ops xfs_bmbt_ops = {
 	.keys_inorder		= xfs_bmbt_keys_inorder,
 	.recs_inorder		= xfs_bmbt_recs_inorder,
 	.keys_contiguous	= xfs_bmbt_keys_contiguous,
+	.broot_realloc		= xfs_bmbt_broot_realloc,
 };
 
 /*
diff --git a/fs/xfs/libxfs/xfs_bmap_btree.h b/fs/xfs/libxfs/xfs_bmap_btree.h
index 49a3bae3f6ecec..b238d559ab0369 100644
--- a/fs/xfs/libxfs/xfs_bmap_btree.h
+++ b/fs/xfs/libxfs/xfs_bmap_btree.h
@@ -198,4 +198,7 @@ xfs_bmap_bmdr_space(struct xfs_btree_block *bb)
 	return xfs_bmdr_space_calc(be16_to_cpu(bb->bb_numrecs));
 }
 
+struct xfs_btree_block *xfs_bmap_broot_realloc(struct xfs_inode *ip,
+		int whichfork, unsigned int new_numrecs);
+
 #endif	/* __XFS_BMAP_BTREE_H__ */
diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c
index 5714bec26c2084..672746f7217cff 100644
--- a/fs/xfs/libxfs/xfs_btree.c
+++ b/fs/xfs/libxfs/xfs_btree.c
@@ -3161,7 +3161,7 @@ xfs_btree_new_iroot(
 
 	xfs_btree_copy_ptrs(cur, pp, &nptr, 1);
 
-	xfs_iroot_realloc(cur->bc_ino.ip, cur->bc_ino.whichfork, 1);
+	cur->bc_ops->broot_realloc(cur, 1);
 
 	xfs_btree_setbuf(cur, level, cbp);
 
@@ -3345,8 +3345,7 @@ 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, cur->bc_ino.whichfork,
-					numrecs + 1);
+			cur->bc_ops->broot_realloc(cur, numrecs + 1);
 			*stat = 1;
 		} else {
 			/* A root block that needs replacing */
@@ -3758,8 +3757,7 @@ xfs_btree_kill_iroot(
 	ASSERT(xfs_btree_ptr_is_null(cur, &ptr));
 #endif
 
-	block = xfs_iroot_realloc(cur->bc_ino.ip, cur->bc_ino.whichfork,
-			numrecs);
+	block = cur->bc_ops->broot_realloc(cur, numrecs);
 
 	block->bb_numrecs = be16_to_cpu(numrecs);
 	ASSERT(block->bb_numrecs == cblock->bb_numrecs);
@@ -3944,8 +3942,7 @@ xfs_btree_delrec(
 	 * nothing left to do.  numrecs was decremented above.
 	 */
 	if (xfs_btree_at_iroot(cur, level)) {
-		xfs_iroot_realloc(cur->bc_ino.ip, cur->bc_ino.whichfork,
-				numrecs);
+		cur->bc_ops->broot_realloc(cur, numrecs);
 
 		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 c5bff273cae255..8380ae0a64dd5e 100644
--- a/fs/xfs/libxfs/xfs_btree.h
+++ b/fs/xfs/libxfs/xfs_btree.h
@@ -213,6 +213,22 @@ struct xfs_btree_ops {
 			       const union xfs_btree_key *key1,
 			       const union xfs_btree_key *key2,
 			       const union xfs_btree_key *mask);
+
+	/*
+	 * Reallocate the space for if_broot to fit the number of records.
+	 * 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.
+	 */
+	struct xfs_btree_block *(*broot_realloc)(struct xfs_btree_cur *cur,
+				unsigned int new_numrecs);
 };
 
 /* btree geometry flags */
diff --git a/fs/xfs/libxfs/xfs_inode_fork.c b/fs/xfs/libxfs/xfs_inode_fork.c
index 53bfdf422ad820..60853bac289a39 100644
--- a/fs/xfs/libxfs/xfs_inode_fork.c
+++ b/fs/xfs/libxfs/xfs_inode_fork.c
@@ -422,102 +422,6 @@ xfs_broot_realloc(
 	return ifp->if_broot;
 }
 
-/*
- * Reallocate the space for if_broot based on the number of records.  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
- * whichfork -- which inode fork to change
- * new_numrecs -- the new number of records requested for the if_broot array
- *
- * Returns the incore btree root block.
- */
-struct xfs_btree_block *
-xfs_iroot_realloc(
-	struct xfs_inode	*ip,
-	int			whichfork,
-	unsigned int		new_numrecs)
-{
-	struct xfs_mount	*mp = ip->i_mount;
-	struct xfs_ifork	*ifp = xfs_ifork_ptr(ip, whichfork);
-	char			*np;
-	char			*op;
-	unsigned int		new_size;
-	unsigned int		old_size = ifp->if_broot_bytes;
-
-	/*
-	 * Block mapping btrees do not support storing zero records; if this
-	 * happens, the fork is being changed to FMT_EXTENTS.  Free the broot
-	 * and get out.
-	 */
-	if (new_numrecs == 0)
-		return xfs_broot_realloc(ifp, 0);
-
-	new_size = xfs_bmap_broot_space_calc(mp, new_numrecs);
-
-	/* Handle the nop case quietly. */
-	if (new_size == old_size)
-		return ifp->if_broot;
-
-	if (new_size > old_size) {
-		unsigned int	old_numrecs;
-
-		/*
-		 * If there wasn't any memory allocated before, just
-		 * allocate it now and get out.
-		 */
-		if (old_size == 0)
-			return xfs_broot_realloc(ifp, new_size);
-
-		/*
-		 * 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.
-		 */
-		old_numrecs = xfs_bmbt_maxrecs(mp, old_size, false);
-		xfs_broot_realloc(ifp, new_size);
-		op = (char *)xfs_bmap_broot_ptr_addr(mp, ifp->if_broot, 1,
-						     old_size);
-		np = (char *)xfs_bmap_broot_ptr_addr(mp, ifp->if_broot, 1,
-						     (int)new_size);
-		ASSERT(xfs_bmap_bmdr_space(ifp->if_broot) <=
-			xfs_inode_fork_size(ip, whichfork));
-		memmove(np, op, old_numrecs * (uint)sizeof(xfs_fsblock_t));
-		return ifp->if_broot;
-	}
-
-	/*
-	 * We're reducing, but not totally eliminating, numrecs.  In this case,
-	 * we are shrinking the if_broot buffer, so it must already exist.
-	 */
-	ASSERT(ifp->if_broot != NULL && old_size > 0 && new_size > 0);
-
-	/*
-	 * Shrink the btree root by moving the bmbt pointers, since they are
-	 * not butted up against the btree block header, then reallocating
-	 * broot.
-	 */
-	op = (char *)xfs_bmap_broot_ptr_addr(mp, ifp->if_broot, 1, old_size);
-	np = (char *)xfs_bmap_broot_ptr_addr(mp, ifp->if_broot, 1,
-					     (int)new_size);
-	memmove(np, op, new_numrecs * (uint)sizeof(xfs_fsblock_t));
-
-	xfs_broot_realloc(ifp, new_size);
-	ASSERT(xfs_bmap_bmdr_space(ifp->if_broot) <=
-	       xfs_inode_fork_size(ip, whichfork));
-	return ifp->if_broot;
-}
-
-
 /*
  * 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 d05eb0bad864e1..69ed0919d60b12 100644
--- a/fs/xfs/libxfs/xfs_inode_fork.h
+++ b/fs/xfs/libxfs/xfs_inode_fork.h
@@ -175,8 +175,6 @@ struct xfs_btree_block *xfs_broot_alloc(struct xfs_ifork *ifp,
 struct xfs_btree_block *xfs_broot_realloc(struct xfs_ifork *ifp,
 				size_t new_size);
 
-struct xfs_btree_block *xfs_iroot_realloc(struct xfs_inode *ip, int whichfork,
-				unsigned int new_numrecs);
 int		xfs_iread_extents(struct xfs_trans *, struct xfs_inode *, int);
 int		xfs_iextents_copy(struct xfs_inode *, struct xfs_bmbt_rec *,
 				  int);

[PATCH 5/8] xfs: tidy up xfs_bmap_broot_realloc a bit

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Hoist out the code that migrates broot pointers during a resize
operation to avoid code duplication and streamline the caller.  Also
use the correct bmbt pointer type for the sizeof operation.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_bmap_btree.c |   43 +++++++++++++++++++++++-----------------
 1 file changed, 25 insertions(+), 18 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_bmap_btree.c b/fs/xfs/libxfs/xfs_bmap_btree.c
index 22cf2059d54dd4..908d7b050e9ce0 100644
--- a/fs/xfs/libxfs/xfs_bmap_btree.c
+++ b/fs/xfs/libxfs/xfs_bmap_btree.c
@@ -516,6 +516,22 @@ xfs_bmbt_keys_contiguous(
 				 be64_to_cpu(key2->bmbt.br_startoff));
 }
 
+static inline void
+xfs_bmbt_move_ptrs(
+	struct xfs_mount	*mp,
+	struct xfs_btree_block	*broot,
+	short			old_size,
+	size_t			new_size,
+	unsigned int		numrecs)
+{
+	void			*dptr;
+	void			*sptr;
+
+	sptr = xfs_bmap_broot_ptr_addr(mp, broot, 1, old_size);
+	dptr = xfs_bmap_broot_ptr_addr(mp, broot, 1, new_size);
+	memmove(dptr, sptr, numrecs * sizeof(xfs_bmbt_ptr_t));
+}
+
 /*
  * Reallocate the space for if_broot based on the number of records.  Move the
  * records and pointers in if_broot to fit the new size.  When shrinking this
@@ -541,8 +557,7 @@ xfs_bmap_broot_realloc(
 {
 	struct xfs_mount	*mp = ip->i_mount;
 	struct xfs_ifork	*ifp = xfs_ifork_ptr(ip, whichfork);
-	char			*np;
-	char			*op;
+	struct xfs_btree_block	*broot;
 	unsigned int		new_size;
 	unsigned int		old_size = ifp->if_broot_bytes;
 
@@ -577,15 +592,11 @@ xfs_bmap_broot_realloc(
 		 * they are kept butted up against the btree block header.
 		 */
 		old_numrecs = xfs_bmbt_maxrecs(mp, old_size, false);
-		xfs_broot_realloc(ifp, new_size);
-		op = (char *)xfs_bmap_broot_ptr_addr(mp, ifp->if_broot, 1,
-						     old_size);
-		np = (char *)xfs_bmap_broot_ptr_addr(mp, ifp->if_broot, 1,
-						     (int)new_size);
-		ASSERT(xfs_bmap_bmdr_space(ifp->if_broot) <=
+		broot = xfs_broot_realloc(ifp, new_size);
+		ASSERT(xfs_bmap_bmdr_space(broot) <=
 			xfs_inode_fork_size(ip, whichfork));
-		memmove(np, op, old_numrecs * (uint)sizeof(xfs_fsblock_t));
-		return ifp->if_broot;
+		xfs_bmbt_move_ptrs(mp, broot, old_size, new_size, old_numrecs);
+		return broot;
 	}
 
 	/*
@@ -599,15 +610,11 @@ xfs_bmap_broot_realloc(
 	 * not butted up against the btree block header, then reallocating
 	 * broot.
 	 */
-	op = (char *)xfs_bmap_broot_ptr_addr(mp, ifp->if_broot, 1, old_size);
-	np = (char *)xfs_bmap_broot_ptr_addr(mp, ifp->if_broot, 1,
-					     (int)new_size);
-	memmove(np, op, new_numrecs * (uint)sizeof(xfs_fsblock_t));
-
-	xfs_broot_realloc(ifp, new_size);
-	ASSERT(xfs_bmap_bmdr_space(ifp->if_broot) <=
+	xfs_bmbt_move_ptrs(mp, ifp->if_broot, old_size, new_size, new_numrecs);
+	broot = xfs_broot_realloc(ifp, new_size);
+	ASSERT(xfs_bmap_bmdr_space(broot) <=
 	       xfs_inode_fork_size(ip, whichfork));
-	return ifp->if_broot;
+	return broot;
 }
 
 static struct xfs_btree_block *

[PATCH 6/8] xfs: hoist the node iroot update code out of xfs_btree_new_iroot

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

In preparation for allowing records in an inode btree root, hoist the
code that copies keyptrs from an existing node root into a child block
to a separate function.  Note that the new function explicitly computes
the keys of the new child block and stores that in the root block; while
the bmap btree could rely on leaving the key alone, realtime rmap needs
to set the new high key.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_btree.c |  117 +++++++++++++++++++++++++++++----------------
 1 file changed, 76 insertions(+), 41 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c
index 672746f7217cff..ed09eeee916160 100644
--- a/fs/xfs/libxfs/xfs_btree.c
+++ b/fs/xfs/libxfs/xfs_btree.c
@@ -3078,6 +3078,78 @@ xfs_btree_split(
 #define xfs_btree_split	__xfs_btree_split
 #endif /* __KERNEL__ */
 
+/*
+ * Move the keys and pointers from a root block to a separate block.
+ *
+ * Since the keyptr size does not change, all we have to do is increase the
+ * tree height, copy the keyptrs to the new internal node (cblock), shrink
+ * the root, and copy the pointers there.
+ */
+STATIC int
+xfs_btree_promote_node_iroot(
+	struct xfs_btree_cur	*cur,
+	struct xfs_btree_block	*block,
+	int			level,
+	struct xfs_buf		*cbp,
+	union xfs_btree_ptr	*cptr,
+	struct xfs_btree_block	*cblock)
+{
+	union xfs_btree_key	*ckp;
+	union xfs_btree_key	*kp;
+	union xfs_btree_ptr	*cpp;
+	union xfs_btree_ptr	*pp;
+	int			i;
+	int			error;
+	int			numrecs = xfs_btree_get_numrecs(block);
+
+	/*
+	 * Increase tree height, adjusting the root block level to match.
+	 * We cannot change the root btree node size until we've copied the
+	 * block contents to the new child block.
+	 */
+	be16_add_cpu(&block->bb_level, 1);
+	cur->bc_nlevels++;
+	cur->bc_levels[level + 1].ptr = 1;
+
+	/*
+	 * Adjust the root btree record count, then copy the keys from the old
+	 * root to the new child block.
+	 */
+	xfs_btree_set_numrecs(block, 1);
+	kp = xfs_btree_key_addr(cur, 1, block);
+	ckp = xfs_btree_key_addr(cur, 1, cblock);
+	xfs_btree_copy_keys(cur, ckp, kp, numrecs);
+
+	/* Check the pointers and copy them to the new child block. */
+	pp = xfs_btree_ptr_addr(cur, 1, block);
+	cpp = xfs_btree_ptr_addr(cur, 1, cblock);
+	for (i = 0; i < numrecs; i++) {
+		error = xfs_btree_debug_check_ptr(cur, pp, i, level);
+		if (error)
+			return error;
+	}
+	xfs_btree_copy_ptrs(cur, cpp, pp, numrecs);
+
+	/*
+	 * Set the first keyptr to point to the new child block, then shrink
+	 * the memory buffer for the root block.
+	 */
+	error = xfs_btree_debug_check_ptr(cur, cptr, 0, level);
+	if (error)
+		return error;
+	xfs_btree_copy_ptrs(cur, pp, cptr, 1);
+	xfs_btree_get_keys(cur, cblock, kp);
+
+	cur->bc_ops->broot_realloc(cur, 1);
+
+	/* Attach the new block to the cursor and log it. */
+	xfs_btree_setbuf(cur, level, cbp);
+	xfs_btree_log_block(cur, cbp, XFS_BB_ALL_BITS);
+	xfs_btree_log_keys(cur, cbp, 1, numrecs);
+	xfs_btree_log_ptrs(cur, cbp, 1, numrecs);
+	return 0;
+}
+
 /*
  * Copy the old inode root contents into a real block and make the
  * broot point to it.
@@ -3091,14 +3163,10 @@ xfs_btree_new_iroot(
 	struct xfs_buf		*cbp;		/* buffer for cblock */
 	struct xfs_btree_block	*block;		/* btree block */
 	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_key	*kp;		/* pointer to btree key */
-	union xfs_btree_ptr	*pp;		/* pointer to block addr */
+	union xfs_btree_ptr	*pp;
 	union xfs_btree_ptr	nptr;		/* new block addr */
 	int			level;		/* btree level */
 	int			error;		/* error return code */
-	int			i;		/* loop counter */
 
 	XFS_BTREE_STATS_INC(cur, newroot);
 
@@ -3136,45 +3204,12 @@ xfs_btree_new_iroot(
 			cblock->bb_u.s.bb_blkno = bno;
 	}
 
-	be16_add_cpu(&block->bb_level, 1);
-	xfs_btree_set_numrecs(block, 1);
-	cur->bc_nlevels++;
-	ASSERT(cur->bc_nlevels <= cur->bc_maxlevels);
-	cur->bc_levels[level + 1].ptr = 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));
-
-	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);
-		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);
+	error = xfs_btree_promote_node_iroot(cur, block, level, cbp, &nptr,
+			cblock);
 	if (error)
 		goto error0;
 
-	xfs_btree_copy_ptrs(cur, pp, &nptr, 1);
-
-	cur->bc_ops->broot_realloc(cur, 1);
-
-	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));
-
-	*logflags |=
-		XFS_ILOG_CORE | xfs_ilog_fbroot(cur->bc_ino.whichfork);
+	*logflags |= XFS_ILOG_CORE | xfs_ilog_fbroot(cur->bc_ino.whichfork);
 	*stat = 1;
 	return 0;
 error0:

[PATCH 7/8] xfs: hoist the node iroot update code out of xfs_btree_kill_iroot

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

In preparation for allowing records in an inode btree root, hoist the
code that copies keyptrs from an existing node child into the root block
to a separate function.  Remove some unnecessary conditionals and clean
up a few function calls in the new function.  Note that this change
reorders the ->free_block call with respect to the change in bc_nlevels
to make it easier to support inode root leaf blocks in the next patch.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_btree.c |   84 +++++++++++++++++++++++++++++++--------------
 1 file changed, 57 insertions(+), 27 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c
index ed09eeee916160..e83a8de5fb8746 100644
--- a/fs/xfs/libxfs/xfs_btree.c
+++ b/fs/xfs/libxfs/xfs_btree.c
@@ -3726,6 +3726,60 @@ xfs_btree_insert(
 	return error;
 }
 
+/*
+ * Move the keyptrs from a child node block to the root block.
+ *
+ * Since the keyptr size does not change, all we have to do is increase the
+ * tree height, copy the keyptrs to the new internal node (cblock), shrink
+ * the root, and copy the pointers there.
+ */
+STATIC int
+xfs_btree_demote_node_child(
+	struct xfs_btree_cur	*cur,
+	struct xfs_btree_block	*cblock,
+	int			level,
+	int			numrecs)
+{
+	struct xfs_btree_block	*block;
+	union xfs_btree_key	*ckp;
+	union xfs_btree_key	*kp;
+	union xfs_btree_ptr	*cpp;
+	union xfs_btree_ptr	*pp;
+	int			i;
+	int			error;
+
+	/*
+	 * Adjust the root btree node size and the record count to match the
+	 * doomed child so that we can copy the keyptrs ahead of changing the
+	 * tree shape.
+	 */
+	block = cur->bc_ops->broot_realloc(cur, numrecs);
+
+	xfs_btree_set_numrecs(block, numrecs);
+	ASSERT(block->bb_numrecs == cblock->bb_numrecs);
+
+	/* Copy keys from the doomed block. */
+	kp = xfs_btree_key_addr(cur, 1, block);
+	ckp = xfs_btree_key_addr(cur, 1, cblock);
+	xfs_btree_copy_keys(cur, kp, ckp, numrecs);
+
+	/* Copy pointers from the doomed block. */
+	pp = xfs_btree_ptr_addr(cur, 1, block);
+	cpp = xfs_btree_ptr_addr(cur, 1, cblock);
+	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);
+
+	/* Decrease tree height, adjusting the root block level to match. */
+	cur->bc_levels[level - 1].bp = NULL;
+	be16_add_cpu(&block->bb_level, -1);
+	cur->bc_nlevels--;
+	return 0;
+}
+
 /*
  * Try to merge a non-leaf block back into the inode root.
  *
@@ -3741,10 +3795,6 @@ xfs_btree_kill_iroot(
 	struct xfs_inode	*ip = cur->bc_ino.ip;
 	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;
 	struct xfs_buf		*cbp;
 	int			level;
 	int			numrecs;
@@ -3752,7 +3802,6 @@ xfs_btree_kill_iroot(
 #ifdef DEBUG
 	union xfs_btree_ptr	ptr;
 #endif
-	int			i;
 
 	ASSERT(cur->bc_ops->type == XFS_BTREE_TYPE_INODE);
 	ASSERT(cur->bc_nlevels > 1);
@@ -3792,35 +3841,16 @@ xfs_btree_kill_iroot(
 	ASSERT(xfs_btree_ptr_is_null(cur, &ptr));
 #endif
 
-	block = cur->bc_ops->broot_realloc(cur, numrecs);
-
-	block->bb_numrecs = be16_to_cpu(numrecs);
-	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);
-
-	pp = xfs_btree_ptr_addr(cur, 1, block);
-	cpp = xfs_btree_ptr_addr(cur, 1, cblock);
-
-	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);
+	error = xfs_btree_demote_node_child(cur, cblock, level, numrecs);
+	if (error)
+		return error;
 
 	error = xfs_btree_free_block(cur, cbp);
 	if (error)
 		return error;
 
-	cur->bc_levels[level - 1].bp = NULL;
-	be16_add_cpu(&block->bb_level, -1);
 	xfs_trans_log_inode(cur->bc_tp, ip,
 		XFS_ILOG_CORE | xfs_ilog_fbroot(cur->bc_ino.whichfork));
-	cur->bc_nlevels--;
 out0:
 	return 0;
 }

[PATCH 8/8] xfs: support storing records in the inode core root

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Add the necessary flags and code so that we can support storing leaf
records in the inode root block of a btree.  This hasn't been necessary
before, but the realtime rmapbt will need to be able to do this.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_btree.c         |  138 ++++++++++++++++++++++++++++++++++---
 fs/xfs/libxfs/xfs_btree.h         |    2 -
 fs/xfs/libxfs/xfs_btree_staging.c |    9 ++
 3 files changed, 132 insertions(+), 17 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c
index e83a8de5fb8746..5ab201ef041e7d 100644
--- a/fs/xfs/libxfs/xfs_btree.c
+++ b/fs/xfs/libxfs/xfs_btree.c
@@ -1537,12 +1537,16 @@ xfs_btree_log_recs(
 	int			first,
 	int			last)
 {
+	if (!bp) {
+		xfs_trans_log_inode(cur->bc_tp, cur->bc_ino.ip,
+				xfs_ilog_fbroot(cur->bc_ino.whichfork));
+		return;
+	}
 
 	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);
-
 }
 
 /*
@@ -3078,6 +3082,59 @@ xfs_btree_split(
 #define xfs_btree_split	__xfs_btree_split
 #endif /* __KERNEL__ */
 
+/* Move the records from a root leaf block to a separate block. */
+STATIC void
+xfs_btree_promote_leaf_iroot(
+	struct xfs_btree_cur	*cur,
+	struct xfs_btree_block	*block,
+	struct xfs_buf		*cbp,
+	union xfs_btree_ptr	*cptr,
+	struct xfs_btree_block	*cblock)
+{
+	union xfs_btree_rec	*rp;
+	union xfs_btree_rec	*crp;
+	union xfs_btree_key	*kp;
+	union xfs_btree_ptr	*pp;
+	struct xfs_btree_block	*broot;
+	int			numrecs = xfs_btree_get_numrecs(block);
+
+	/* Copy the records from the leaf broot into the new child block. */
+	rp = xfs_btree_rec_addr(cur, 1, block);
+	crp = xfs_btree_rec_addr(cur, 1, cblock);
+	xfs_btree_copy_recs(cur, crp, rp, numrecs);
+
+	/*
+	 * Increment the tree height.
+	 *
+	 * Trickery here: The amount of memory that we need per record for the
+	 * ifork's btree root block may change when we convert the broot from a
+	 * leaf to a node block.  Free the existing leaf broot so that nobody
+	 * thinks we need to migrate node pointers when we realloc the broot
+	 * buffer after bumping nlevels.
+	 */
+	cur->bc_ops->broot_realloc(cur, 0);
+	cur->bc_nlevels++;
+	cur->bc_levels[1].ptr = 1;
+
+	/*
+	 * Allocate a new node broot and initialize it to point to the new
+	 * child block.
+	 */
+	broot = cur->bc_ops->broot_realloc(cur, 1);
+	xfs_btree_init_block(cur->bc_mp, broot, cur->bc_ops,
+			cur->bc_nlevels - 1, 1, cur->bc_ino.ip->i_ino);
+
+	pp = xfs_btree_ptr_addr(cur, 1, broot);
+	kp = xfs_btree_key_addr(cur, 1, broot);
+	xfs_btree_copy_ptrs(cur, pp, cptr, 1);
+	xfs_btree_get_keys(cur, cblock, kp);
+
+	/* Attach the new block to the cursor and log it. */
+	xfs_btree_setbuf(cur, 0, cbp);
+	xfs_btree_log_block(cur, cbp, XFS_BB_ALL_BITS);
+	xfs_btree_log_recs(cur, cbp, 1, numrecs);
+}
+
 /*
  * Move the keys and pointers from a root block to a separate block.
  *
@@ -3163,7 +3220,7 @@ xfs_btree_new_iroot(
 	struct xfs_buf		*cbp;		/* buffer for cblock */
 	struct xfs_btree_block	*block;		/* btree block */
 	struct xfs_btree_block	*cblock;	/* child btree block */
-	union xfs_btree_ptr	*pp;
+	union xfs_btree_ptr	aptr;
 	union xfs_btree_ptr	nptr;		/* new block addr */
 	int			level;		/* btree level */
 	int			error;		/* error return code */
@@ -3175,10 +3232,15 @@ 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_ops->geom_flags & XFS_BTGEO_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_ino.ip->i_ino));
 
 	/* Allocate the new block. If we can't do it, we're toast. Give up. */
-	error = xfs_btree_alloc_block(cur, pp, &nptr, stat);
+	error = xfs_btree_alloc_block(cur, &aptr, &nptr, stat);
 	if (error)
 		goto error0;
 	if (*stat == 0)
@@ -3204,10 +3266,14 @@ xfs_btree_new_iroot(
 			cblock->bb_u.s.bb_blkno = bno;
 	}
 
-	error = xfs_btree_promote_node_iroot(cur, block, level, cbp, &nptr,
-			cblock);
-	if (error)
-		goto error0;
+	if (level > 0) {
+		error = xfs_btree_promote_node_iroot(cur, block, level, cbp,
+				&nptr, cblock);
+		if (error)
+			goto error0;
+	} else {
+		xfs_btree_promote_leaf_iroot(cur, block, cbp, &nptr, cblock);
+	}
 
 	*logflags |= XFS_ILOG_CORE | xfs_ilog_fbroot(cur->bc_ino.whichfork);
 	*stat = 1;
@@ -3726,6 +3792,43 @@ xfs_btree_insert(
 	return error;
 }
 
+/* Move the records from a child leaf block to the root block. */
+STATIC void
+xfs_btree_demote_leaf_child(
+	struct xfs_btree_cur	*cur,
+	struct xfs_btree_block	*cblock,
+	int			numrecs)
+{
+	union xfs_btree_rec	*rp;
+	union xfs_btree_rec	*crp;
+	struct xfs_btree_block	*broot;
+
+	/*
+	 * Decrease the tree height.
+	 *
+	 * Trickery here: The amount of memory that we need per record for the
+	 * ifork's btree root block may change when we convert the broot from a
+	 * node to a leaf.  Free the old node broot so that we can get a fresh
+	 * leaf broot.
+	 */
+	cur->bc_ops->broot_realloc(cur, 0);
+	cur->bc_nlevels--;
+
+	/*
+	 * Allocate a new leaf broot and copy the records from the old child.
+	 * Detach the old child from the cursor.
+	 */
+	broot = cur->bc_ops->broot_realloc(cur, numrecs);
+	xfs_btree_init_block(cur->bc_mp, broot, cur->bc_ops, 0, numrecs,
+			cur->bc_ino.ip->i_ino);
+
+	rp = xfs_btree_rec_addr(cur, 1, broot);
+	crp = xfs_btree_rec_addr(cur, 1, cblock);
+	xfs_btree_copy_recs(cur, rp, crp, numrecs);
+
+	cur->bc_levels[0].bp = NULL;
+}
+
 /*
  * Move the keyptrs from a child node block to the root block.
  *
@@ -3804,14 +3907,19 @@ xfs_btree_kill_iroot(
 #endif
 
 	ASSERT(cur->bc_ops->type == XFS_BTREE_TYPE_INODE);
-	ASSERT(cur->bc_nlevels > 1);
+	ASSERT((cur->bc_ops->geom_flags & XFS_BTGEO_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_ops->geom_flags & XFS_BTGEO_IROOT_RECORDS))
+		goto out0;
+
+	/* If we're already a leaf, jump out. */
+	if (level == 0)
 		goto out0;
 
 	/*
@@ -3841,9 +3949,13 @@ xfs_btree_kill_iroot(
 	ASSERT(xfs_btree_ptr_is_null(cur, &ptr));
 #endif
 
-	error = xfs_btree_demote_node_child(cur, cblock, level, numrecs);
-	if (error)
-		return error;
+	if (level > 1) {
+		error = xfs_btree_demote_node_child(cur, cblock, level,
+				numrecs);
+		if (error)
+			return error;
+	} else
+		xfs_btree_demote_leaf_child(cur, cblock, numrecs);
 
 	error = xfs_btree_free_block(cur, cbp);
 	if (error)
diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h
index 8380ae0a64dd5e..3b8c2ccad90847 100644
--- a/fs/xfs/libxfs/xfs_btree.h
+++ b/fs/xfs/libxfs/xfs_btree.h
@@ -233,7 +233,7 @@ struct xfs_btree_ops {
 
 /* btree geometry flags */
 #define XFS_BTGEO_OVERLAPPING		(1U << 0) /* overlapping intervals */
-
+#define XFS_BTGEO_IROOT_RECORDS		(1U << 1) /* iroot can store records */
 
 union xfs_btree_irec {
 	struct xfs_alloc_rec_incore	a;
diff --git a/fs/xfs/libxfs/xfs_btree_staging.c b/fs/xfs/libxfs/xfs_btree_staging.c
index 6949297031529e..58c146b5c9d479 100644
--- a/fs/xfs/libxfs/xfs_btree_staging.c
+++ b/fs/xfs/libxfs/xfs_btree_staging.c
@@ -573,6 +573,7 @@ xfs_btree_bload_compute_geometry(
 	struct xfs_btree_bload	*bbl,
 	uint64_t		nr_records)
 {
+	const struct xfs_btree_ops *ops = cur->bc_ops;
 	uint64_t		nr_blocks = 0;
 	uint64_t		nr_this_level;
 
@@ -599,7 +600,7 @@ xfs_btree_bload_compute_geometry(
 		xfs_btree_bload_level_geometry(cur, bbl, level, nr_this_level,
 				&avg_per_block, &level_blocks, &dontcare64);
 
-		if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE) {
+		if (ops->type == XFS_BTREE_TYPE_INODE) {
 			/*
 			 * If all the items we want to store at this level
 			 * would fit in the inode root block, then we have our
@@ -607,7 +608,9 @@ xfs_btree_bload_compute_geometry(
 			 *
 			 * Note that bmap btrees forbid records in the root.
 			 */
-			if (level != 0 && nr_this_level <= avg_per_block) {
+			if ((level != 0 ||
+			     (ops->geom_flags & XFS_BTGEO_IROOT_RECORDS)) &&
+			    nr_this_level <= avg_per_block) {
 				nr_blocks++;
 				break;
 			}
@@ -658,7 +661,7 @@ xfs_btree_bload_compute_geometry(
 		return -EOVERFLOW;
 
 	bbl->btree_height = cur->bc_nlevels;
-	if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE)
+	if (ops->type == XFS_BTREE_TYPE_INODE)
 		bbl->nr_blocks = nr_blocks - 1;
 	else
 		bbl->nr_blocks = nr_blocks;

[PATCHSET v6.1 3/5] xfs: enable in-core block reservation for rt metadata

[Darrick J. Wong]

Hi all,

In preparation for adding reverse mapping and refcounting to the
realtime device, enhance the metadir code to reserve free space for
btree shape changes as delayed allocation blocks.

This enables us to pre-allocate space for the rmap and refcount btrees
in the same manner as we do for the data device counterparts, which is
how we avoid ENOSPC failures when space is low but we've already
committed to a COW operation.

If you're going to start using this code, I strongly recommend pulling
from my git trees, which are linked below.

This has been running on the djcloud for months with no problems.  Enjoy!
Comments and questions are, as always, welcome.

--D

kernel git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfs-linux.git/log/?h=reserve-rt-metadata-space

xfsprogs git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfsprogs-dev.git/log/?h=reserve-rt-metadata-space
---
Commits in this patchset:
 * xfs: prepare to reuse the dquot pointer space in struct xfs_inode
 * xfs: allow inode-based btrees to reserve space in the data device
---
 fs/xfs/libxfs/xfs_ag_resv.c  |    3 +
 fs/xfs/libxfs/xfs_attr.c     |    4 -
 fs/xfs/libxfs/xfs_bmap.c     |    4 -
 fs/xfs/libxfs/xfs_errortag.h |    4 +
 fs/xfs/libxfs/xfs_metadir.c  |    4 +
 fs/xfs/libxfs/xfs_metafile.c |  205 ++++++++++++++++++++++++++++++++++++++++++
 fs/xfs/libxfs/xfs_metafile.h |   11 ++
 fs/xfs/libxfs/xfs_types.h    |    7 +
 fs/xfs/scrub/tempfile.c      |    1 
 fs/xfs/xfs_dquot.h           |    3 +
 fs/xfs/xfs_error.c           |    3 +
 fs/xfs/xfs_exchrange.c       |    3 +
 fs/xfs/xfs_fsops.c           |   17 +++
 fs/xfs/xfs_inode.h           |   16 +++
 fs/xfs/xfs_mount.c           |   10 ++
 fs/xfs/xfs_mount.h           |    1 
 fs/xfs/xfs_qm.c              |    2 
 fs/xfs/xfs_quota.h           |    5 -
 fs/xfs/xfs_rtalloc.c         |   21 ++++
 fs/xfs/xfs_rtalloc.h         |    5 +
 fs/xfs/xfs_trace.h           |   45 +++++++++
 fs/xfs/xfs_trans.c           |    4 +
 fs/xfs/xfs_trans_dquot.c     |    8 +-
 23 files changed, 367 insertions(+), 19 deletions(-)

[PATCH 1/2] xfs: prepare to reuse the dquot pointer space in struct xfs_inode

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Files participating in the metadata directory tree are not accounted to
the quota subsystem.  Therefore, the i_[ugp]dquot pointers in struct
xfs_inode are never used and should always be NULL.

In the next patch we want to add a u64 count of fs blocks reserved for
metadata btree expansion, but we don't want every inode in the fs to pay
the memory price for this feature.  The intent is to union those three
pointers with the u64 counter, but for that to work we must guard
against all access to the dquot pointers for metadata files.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
---
 fs/xfs/libxfs/xfs_attr.c |    4 +---
 fs/xfs/libxfs/xfs_bmap.c |    4 +---
 fs/xfs/scrub/tempfile.c  |    1 +
 fs/xfs/xfs_dquot.h       |    3 +++
 fs/xfs/xfs_exchrange.c   |    3 +++
 fs/xfs/xfs_inode.h       |   10 +++++++---
 fs/xfs/xfs_qm.c          |    2 ++
 fs/xfs/xfs_quota.h       |    5 -----
 fs/xfs/xfs_trans.c       |    4 ++++
 fs/xfs/xfs_trans_dquot.c |    8 ++++----
 10 files changed, 26 insertions(+), 18 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_attr.c b/fs/xfs/libxfs/xfs_attr.c
index 17875ad865f5d6..8c04acd30d489c 100644
--- a/fs/xfs/libxfs/xfs_attr.c
+++ b/fs/xfs/libxfs/xfs_attr.c
@@ -1004,9 +1004,7 @@ xfs_attr_add_fork(
 	unsigned int		blks;		/* space reservation */
 	int			error;		/* error return value */
 
-	if (xfs_is_metadir_inode(ip))
-		ASSERT(XFS_IS_DQDETACHED(ip));
-	else
+	if (!xfs_is_metadir_inode(ip))
 		ASSERT(!XFS_NOT_DQATTACHED(mp, ip));
 
 	blks = XFS_ADDAFORK_SPACE_RES(mp);
diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c
index 0842577755f7bb..02323936cc9b20 100644
--- a/fs/xfs/libxfs/xfs_bmap.c
+++ b/fs/xfs/libxfs/xfs_bmap.c
@@ -1042,9 +1042,7 @@ xfs_bmap_add_attrfork(
 	int			error;		/* error return value */
 
 	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
-	if (xfs_is_metadir_inode(ip))
-		ASSERT(XFS_IS_DQDETACHED(ip));
-	else
+	if (!xfs_is_metadir_inode(ip))
 		ASSERT(!XFS_NOT_DQATTACHED(mp, ip));
 	ASSERT(!xfs_inode_has_attr_fork(ip));
 
diff --git a/fs/xfs/scrub/tempfile.c b/fs/xfs/scrub/tempfile.c
index 2d7ca7e1bbca0f..4ebb5f8459e8f3 100644
--- a/fs/xfs/scrub/tempfile.c
+++ b/fs/xfs/scrub/tempfile.c
@@ -749,6 +749,7 @@ xrep_tempexch_reserve_quota(
 	 * or the two inodes have the same dquots.
 	 */
 	if (!XFS_IS_QUOTA_ON(tp->t_mountp) || req->ip1 == req->ip2 ||
+	    xfs_is_metadir_inode(req->ip1) ||
 	    (req->ip1->i_udquot == req->ip2->i_udquot &&
 	     req->ip1->i_gdquot == req->ip2->i_gdquot &&
 	     req->ip1->i_pdquot == req->ip2->i_pdquot))
diff --git a/fs/xfs/xfs_dquot.h b/fs/xfs/xfs_dquot.h
index c617bac75361b2..61217adf5ba551 100644
--- a/fs/xfs/xfs_dquot.h
+++ b/fs/xfs/xfs_dquot.h
@@ -160,6 +160,9 @@ static inline struct xfs_dquot *xfs_inode_dquot(
 	struct xfs_inode	*ip,
 	xfs_dqtype_t		type)
 {
+	if (xfs_is_metadir_inode(ip))
+		return NULL;
+
 	switch (type) {
 	case XFS_DQTYPE_USER:
 		return ip->i_udquot;
diff --git a/fs/xfs/xfs_exchrange.c b/fs/xfs/xfs_exchrange.c
index 265c424498933e..f340a2015c4c71 100644
--- a/fs/xfs/xfs_exchrange.c
+++ b/fs/xfs/xfs_exchrange.c
@@ -119,6 +119,9 @@ xfs_exchrange_reserve_quota(
 	int				ip1_error = 0;
 	int				error;
 
+	ASSERT(!xfs_is_metadir_inode(req->ip1));
+	ASSERT(!xfs_is_metadir_inode(req->ip2));
+
 	/*
 	 * Don't bother with a quota reservation if we're not enforcing them
 	 * or the two inodes have the same dquots.
diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h
index 1648dc5a806882..1141c2e8e123ae 100644
--- a/fs/xfs/xfs_inode.h
+++ b/fs/xfs/xfs_inode.h
@@ -25,9 +25,13 @@ struct xfs_dquot;
 typedef struct xfs_inode {
 	/* Inode linking and identification information. */
 	struct xfs_mount	*i_mount;	/* fs mount struct ptr */
-	struct xfs_dquot	*i_udquot;	/* user dquot */
-	struct xfs_dquot	*i_gdquot;	/* group dquot */
-	struct xfs_dquot	*i_pdquot;	/* project dquot */
+	union {
+		struct {
+			struct xfs_dquot *i_udquot;	/* user dquot */
+			struct xfs_dquot *i_gdquot;	/* group dquot */
+			struct xfs_dquot *i_pdquot;	/* project dquot */
+		};
+	};
 
 	/* Inode location stuff */
 	xfs_ino_t		i_ino;		/* inode number (agno/agino)*/
diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c
index dc8b1010d4d332..3abab5fb593e37 100644
--- a/fs/xfs/xfs_qm.c
+++ b/fs/xfs/xfs_qm.c
@@ -428,6 +428,8 @@ void
 xfs_qm_dqdetach(
 	xfs_inode_t	*ip)
 {
+	if (xfs_is_metadir_inode(ip))
+		return;
 	if (!(ip->i_udquot || ip->i_gdquot || ip->i_pdquot))
 		return;
 
diff --git a/fs/xfs/xfs_quota.h b/fs/xfs/xfs_quota.h
index d7565462af3dc4..105e6eb5762011 100644
--- a/fs/xfs/xfs_quota.h
+++ b/fs/xfs/xfs_quota.h
@@ -29,11 +29,6 @@ struct xfs_buf;
 	 (XFS_IS_GQUOTA_ON(mp) && (ip)->i_gdquot == NULL) || \
 	 (XFS_IS_PQUOTA_ON(mp) && (ip)->i_pdquot == NULL))
 
-#define XFS_IS_DQDETACHED(ip) \
-	((ip)->i_udquot == NULL && \
-	 (ip)->i_gdquot == NULL && \
-	 (ip)->i_pdquot == NULL)
-
 #define XFS_QM_NEED_QUOTACHECK(mp) \
 	((XFS_IS_UQUOTA_ON(mp) && \
 		(mp->m_sb.sb_qflags & XFS_UQUOTA_CHKD) == 0) || \
diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c
index 4cd25717c9d130..f53f82456288e5 100644
--- a/fs/xfs/xfs_trans.c
+++ b/fs/xfs/xfs_trans.c
@@ -1266,6 +1266,9 @@ xfs_trans_alloc_ichange(
 	xfs_ilock(ip, XFS_ILOCK_EXCL);
 	xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
 
+	if (xfs_is_metadir_inode(ip))
+		goto out;
+
 	error = xfs_qm_dqattach_locked(ip, false);
 	if (error) {
 		/* Caller should have allocated the dquots! */
@@ -1334,6 +1337,7 @@ xfs_trans_alloc_ichange(
 			goto out_cancel;
 	}
 
+out:
 	*tpp = tp;
 	return 0;
 
diff --git a/fs/xfs/xfs_trans_dquot.c b/fs/xfs/xfs_trans_dquot.c
index 713b6d243e5631..765456bf342851 100644
--- a/fs/xfs/xfs_trans_dquot.c
+++ b/fs/xfs/xfs_trans_dquot.c
@@ -156,7 +156,8 @@ xfs_trans_mod_ino_dquot(
 	unsigned int			field,
 	int64_t				delta)
 {
-	ASSERT(!xfs_is_metadir_inode(ip) || XFS_IS_DQDETACHED(ip));
+	if (xfs_is_metadir_inode(ip))
+		return;
 
 	xfs_trans_mod_dquot(tp, dqp, field, delta);
 
@@ -246,11 +247,10 @@ xfs_trans_mod_dquot_byino(
 	xfs_mount_t	*mp = tp->t_mountp;
 
 	if (!XFS_IS_QUOTA_ON(mp) ||
-	    xfs_is_quota_inode(&mp->m_sb, ip->i_ino))
+	    xfs_is_quota_inode(&mp->m_sb, ip->i_ino) ||
+	    xfs_is_metadir_inode(ip))
 		return;
 
-	ASSERT(!xfs_is_metadir_inode(ip) || XFS_IS_DQDETACHED(ip));
-
 	if (XFS_IS_UQUOTA_ON(mp) && ip->i_udquot)
 		xfs_trans_mod_ino_dquot(tp, ip, ip->i_udquot, field, delta);
 	if (XFS_IS_GQUOTA_ON(mp) && ip->i_gdquot)

Re: [PATCH 1/2] xfs: prepare to reuse the dquot pointer space in struct xfs_inode

[Christoph Hellwig]

Looks good:

Reviewed-by: Christoph Hellwig <hch@lst.de>

[PATCH 2/2] xfs: allow inode-based btrees to reserve space in the data device

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Create a new space reservation scheme so that btree metadata for the
realtime volume can reserve space in the data device to avoid space
underruns.

Back when we were testing the rmap and refcount btrees for the data
device, people observed occasional shutdowns when xfs_btree_split was
called for either of those two btrees.  This happened when certain
operations (mostly writeback ioends) created new rmap or refcount
records, which would expand the size of the btree.  If there were no
free blocks available the allocation would fail and the split would shut
down the filesystem.

I considered pre-reserving blocks for btree expansion at the time of a
write() call, but there wasn't any good way to attach the reservations
to an inode and keep them there all the way to ioend processing.  Unlike
delalloc reservations which have that indlen mechanism, there's no way
to do that for mapped extents; and indlen blocks are given back during
the delalloc -> unwritten transition.

The solution was to reserve sufficient blocks for rmap/refcount btree
expansion at mount time.  This is what the XFS_AG_RESV_* flags provide;
any expansion of those two btrees can come from the pre-reserved space.

This patch brings that pre-reservation ability to inode-rooted btrees so
that the rt rmap and refcount btrees can also save room for future
expansion.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
---
 fs/xfs/libxfs/xfs_ag_resv.c  |    3 +
 fs/xfs/libxfs/xfs_errortag.h |    4 +
 fs/xfs/libxfs/xfs_metadir.c  |    4 +
 fs/xfs/libxfs/xfs_metafile.c |  205 ++++++++++++++++++++++++++++++++++++++++++
 fs/xfs/libxfs/xfs_metafile.h |   11 ++
 fs/xfs/libxfs/xfs_types.h    |    7 +
 fs/xfs/xfs_error.c           |    3 +
 fs/xfs/xfs_fsops.c           |   17 +++
 fs/xfs/xfs_inode.h           |    6 +
 fs/xfs/xfs_mount.c           |   10 ++
 fs/xfs/xfs_mount.h           |    1 
 fs/xfs/xfs_rtalloc.c         |   21 ++++
 fs/xfs/xfs_rtalloc.h         |    5 +
 fs/xfs/xfs_trace.h           |   45 +++++++++
 14 files changed, 341 insertions(+), 1 deletion(-)


diff --git a/fs/xfs/libxfs/xfs_ag_resv.c b/fs/xfs/libxfs/xfs_ag_resv.c
index f5d853089019f0..fb79215a509d21 100644
--- a/fs/xfs/libxfs/xfs_ag_resv.c
+++ b/fs/xfs/libxfs/xfs_ag_resv.c
@@ -114,6 +114,7 @@ xfs_ag_resv_needed(
 	case XFS_AG_RESV_RMAPBT:
 		len -= xfs_perag_resv(pag, type)->ar_reserved;
 		break;
+	case XFS_AG_RESV_METAFILE:
 	case XFS_AG_RESV_NONE:
 		/* empty */
 		break;
@@ -347,6 +348,7 @@ xfs_ag_resv_alloc_extent(
 
 	switch (type) {
 	case XFS_AG_RESV_AGFL:
+	case XFS_AG_RESV_METAFILE:
 		return;
 	case XFS_AG_RESV_METADATA:
 	case XFS_AG_RESV_RMAPBT:
@@ -389,6 +391,7 @@ xfs_ag_resv_free_extent(
 
 	switch (type) {
 	case XFS_AG_RESV_AGFL:
+	case XFS_AG_RESV_METAFILE:
 		return;
 	case XFS_AG_RESV_METADATA:
 	case XFS_AG_RESV_RMAPBT:
diff --git a/fs/xfs/libxfs/xfs_errortag.h b/fs/xfs/libxfs/xfs_errortag.h
index 7002d7676a7884..a53c5d40e084dc 100644
--- a/fs/xfs/libxfs/xfs_errortag.h
+++ b/fs/xfs/libxfs/xfs_errortag.h
@@ -64,7 +64,8 @@
 #define XFS_ERRTAG_WB_DELAY_MS				42
 #define XFS_ERRTAG_WRITE_DELAY_MS			43
 #define XFS_ERRTAG_EXCHMAPS_FINISH_ONE			44
-#define XFS_ERRTAG_MAX					45
+#define XFS_ERRTAG_METAFILE_RESV_CRITICAL		45
+#define XFS_ERRTAG_MAX					46
 
 /*
  * Random factors for above tags, 1 means always, 2 means 1/2 time, etc.
@@ -113,5 +114,6 @@
 #define XFS_RANDOM_WB_DELAY_MS				3000
 #define XFS_RANDOM_WRITE_DELAY_MS			3000
 #define XFS_RANDOM_EXCHMAPS_FINISH_ONE			1
+#define XFS_RANDOM_METAFILE_RESV_CRITICAL		4
 
 #endif /* __XFS_ERRORTAG_H_ */
diff --git a/fs/xfs/libxfs/xfs_metadir.c b/fs/xfs/libxfs/xfs_metadir.c
index bae7377c0f228c..178e89711cb7c6 100644
--- a/fs/xfs/libxfs/xfs_metadir.c
+++ b/fs/xfs/libxfs/xfs_metadir.c
@@ -29,6 +29,10 @@
 #include "xfs_dir2_priv.h"
 #include "xfs_parent.h"
 #include "xfs_health.h"
+#include "xfs_errortag.h"
+#include "xfs_error.h"
+#include "xfs_btree.h"
+#include "xfs_alloc.h"
 
 /*
  * Metadata Directory Tree
diff --git a/fs/xfs/libxfs/xfs_metafile.c b/fs/xfs/libxfs/xfs_metafile.c
index adeb25d1a444ca..e151663cc9efd6 100644
--- a/fs/xfs/libxfs/xfs_metafile.c
+++ b/fs/xfs/libxfs/xfs_metafile.c
@@ -17,6 +17,10 @@
 #include "xfs_metafile.h"
 #include "xfs_trace.h"
 #include "xfs_inode.h"
+#include "xfs_quota.h"
+#include "xfs_errortag.h"
+#include "xfs_error.h"
+#include "xfs_alloc.h"
 
 /* Set up an inode to be recognized as a metadata directory inode. */
 void
@@ -50,3 +54,204 @@ xfs_metafile_clear_iflag(
 	ip->i_diflags2 &= ~XFS_DIFLAG2_METADATA;
 	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
 }
+
+/*
+ * Is the amount of space that could be allocated towards a given metadata
+ * file at or beneath a certain threshold?
+ */
+static inline bool
+xfs_metafile_resv_can_cover(
+	struct xfs_inode	*ip,
+	int64_t			rhs)
+{
+	/*
+	 * The amount of space that can be allocated to this metadata file is
+	 * the remaining reservation for the particular metadata file + the
+	 * global free block count.  Take care of the first case to avoid
+	 * touching the per-cpu counter.
+	 */
+	if (ip->i_delayed_blks >= rhs)
+		return true;
+
+	/*
+	 * There aren't enough blocks left in the inode's reservation, but it
+	 * isn't critical unless there also isn't enough free space.
+	 */
+	return __percpu_counter_compare(&ip->i_mount->m_fdblocks,
+			rhs - ip->i_delayed_blks, 2048) >= 0;
+}
+
+/*
+ * Is this metadata file critically low on blocks?  For now we'll define that
+ * as the number of blocks we can get our hands on being less than 10% of what
+ * we reserved or less than some arbitrary number (maximum btree height).
+ */
+bool
+xfs_metafile_resv_critical(
+	struct xfs_inode	*ip)
+{
+	uint64_t		asked_low_water;
+
+	if (!ip)
+		return false;
+
+	ASSERT(xfs_is_metadir_inode(ip));
+	trace_xfs_metafile_resv_critical(ip, 0);
+
+	if (!xfs_metafile_resv_can_cover(ip, ip->i_mount->m_rtbtree_maxlevels))
+		return true;
+
+	asked_low_water = div_u64(ip->i_meta_resv_asked, 10);
+	if (!xfs_metafile_resv_can_cover(ip, asked_low_water))
+		return true;
+
+	return XFS_TEST_ERROR(false, ip->i_mount,
+			XFS_ERRTAG_METAFILE_RESV_CRITICAL);
+}
+
+/* Allocate a block from the metadata file's reservation. */
+void
+xfs_metafile_resv_alloc_space(
+	struct xfs_inode	*ip,
+	struct xfs_alloc_arg	*args)
+{
+	int64_t			len = args->len;
+
+	ASSERT(xfs_is_metadir_inode(ip));
+	ASSERT(args->resv == XFS_AG_RESV_METAFILE);
+
+	trace_xfs_metafile_resv_alloc_space(ip, args->len);
+
+	/*
+	 * Allocate the blocks from the metadata inode's block reservation
+	 * and update the ondisk sb counter.
+	 */
+	if (ip->i_delayed_blks > 0) {
+		int64_t		from_resv;
+
+		from_resv = min_t(int64_t, len, ip->i_delayed_blks);
+		ip->i_delayed_blks -= from_resv;
+		xfs_mod_delalloc(ip, 0, -from_resv);
+		xfs_trans_mod_sb(args->tp, XFS_TRANS_SB_RES_FDBLOCKS,
+				-from_resv);
+		len -= from_resv;
+	}
+
+	/*
+	 * Any allocation in excess of the reservation requires in-core and
+	 * on-disk fdblocks updates.  If we can grab @len blocks from the
+	 * in-core fdblocks then all we need to do is update the on-disk
+	 * superblock; if not, then try to steal some from the transaction's
+	 * block reservation.  Overruns are only expected for rmap btrees.
+	 */
+	if (len) {
+		unsigned int	field;
+		int		error;
+
+		error = xfs_dec_fdblocks(ip->i_mount, len, true);
+		if (error)
+			field = XFS_TRANS_SB_FDBLOCKS;
+		else
+			field = XFS_TRANS_SB_RES_FDBLOCKS;
+
+		xfs_trans_mod_sb(args->tp, field, -len);
+	}
+
+	ip->i_nblocks += args->len;
+	xfs_trans_log_inode(args->tp, ip, XFS_ILOG_CORE);
+}
+
+/* Free a block to the metadata file's reservation. */
+void
+xfs_metafile_resv_free_space(
+	struct xfs_inode	*ip,
+	struct xfs_trans	*tp,
+	xfs_filblks_t		len)
+{
+	int64_t			to_resv;
+
+	ASSERT(xfs_is_metadir_inode(ip));
+	trace_xfs_metafile_resv_free_space(ip, len);
+
+	ip->i_nblocks -= len;
+	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+
+	/*
+	 * Add the freed blocks back into the inode's delalloc reservation
+	 * until it reaches the maximum size.  Update the ondisk fdblocks only.
+	 */
+	to_resv = ip->i_meta_resv_asked - (ip->i_nblocks + ip->i_delayed_blks);
+	if (to_resv > 0) {
+		to_resv = min_t(int64_t, to_resv, len);
+		ip->i_delayed_blks += to_resv;
+		xfs_mod_delalloc(ip, 0, to_resv);
+		xfs_trans_mod_sb(tp, XFS_TRANS_SB_RES_FDBLOCKS, to_resv);
+		len -= to_resv;
+	}
+
+	/*
+	 * Everything else goes back to the filesystem, so update the in-core
+	 * and on-disk counters.
+	 */
+	if (len)
+		xfs_trans_mod_sb(tp, XFS_TRANS_SB_FDBLOCKS, len);
+}
+
+/* Release a metadata file's space reservation. */
+void
+xfs_metafile_resv_free(
+	struct xfs_inode	*ip)
+{
+	/* Non-btree metadata inodes don't need space reservations. */
+	if (!ip || !ip->i_meta_resv_asked)
+		return;
+
+	ASSERT(xfs_is_metadir_inode(ip));
+	trace_xfs_metafile_resv_free(ip, 0);
+
+	if (ip->i_delayed_blks) {
+		xfs_mod_delalloc(ip, 0, -ip->i_delayed_blks);
+		xfs_add_fdblocks(ip->i_mount, ip->i_delayed_blks);
+		ip->i_delayed_blks = 0;
+	}
+	ip->i_meta_resv_asked = 0;
+}
+
+/* Set up a metadata file's space reservation. */
+int
+xfs_metafile_resv_init(
+	struct xfs_inode	*ip,
+	xfs_filblks_t		ask)
+{
+	xfs_filblks_t		hidden_space;
+	xfs_filblks_t		used;
+	int			error;
+
+	if (!ip || ip->i_meta_resv_asked > 0)
+		return 0;
+
+	ASSERT(xfs_is_metadir_inode(ip));
+
+	/*
+	 * Space taken by all other metadata btrees are accounted on-disk as
+	 * used space.  We therefore only hide the space that is reserved but
+	 * not used by the trees.
+	 */
+	used = ip->i_nblocks;
+	if (used > ask)
+		ask = used;
+	hidden_space = ask - used;
+
+	error = xfs_dec_fdblocks(ip->i_mount, hidden_space, true);
+	if (error) {
+		trace_xfs_metafile_resv_init_error(ip, error, _RET_IP_);
+		return error;
+	}
+
+	xfs_mod_delalloc(ip, 0, hidden_space);
+	ip->i_delayed_blks = hidden_space;
+	ip->i_meta_resv_asked = ask;
+
+	trace_xfs_metafile_resv_init(ip, ask);
+	return 0;
+}
diff --git a/fs/xfs/libxfs/xfs_metafile.h b/fs/xfs/libxfs/xfs_metafile.h
index acec400123db05..8d8f08a6071c23 100644
--- a/fs/xfs/libxfs/xfs_metafile.h
+++ b/fs/xfs/libxfs/xfs_metafile.h
@@ -21,6 +21,17 @@ void xfs_metafile_set_iflag(struct xfs_trans *tp, struct xfs_inode *ip,
 		enum xfs_metafile_type metafile_type);
 void xfs_metafile_clear_iflag(struct xfs_trans *tp, struct xfs_inode *ip);
 
+/* Space reservations for metadata inodes. */
+struct xfs_alloc_arg;
+
+bool xfs_metafile_resv_critical(struct xfs_inode *ip);
+void xfs_metafile_resv_alloc_space(struct xfs_inode *ip,
+		struct xfs_alloc_arg *args);
+void xfs_metafile_resv_free_space(struct xfs_inode *ip, struct xfs_trans *tp,
+		xfs_filblks_t len);
+void xfs_metafile_resv_free(struct xfs_inode *ip);
+int xfs_metafile_resv_init(struct xfs_inode *ip, xfs_filblks_t ask);
+
 /* Code specific to kernel/userspace; must be provided externally. */
 
 int xfs_trans_metafile_iget(struct xfs_trans *tp, xfs_ino_t ino,
diff --git a/fs/xfs/libxfs/xfs_types.h b/fs/xfs/libxfs/xfs_types.h
index bf33c2b1e43e5f..ca2401c1facda7 100644
--- a/fs/xfs/libxfs/xfs_types.h
+++ b/fs/xfs/libxfs/xfs_types.h
@@ -202,6 +202,13 @@ enum xfs_ag_resv_type {
 	 * altering fdblocks.  If you think you need this you're wrong.
 	 */
 	XFS_AG_RESV_IGNORE,
+
+	/*
+	 * This allocation activity is being done on behalf of a metadata file.
+	 * These files maintain their own permanent space reservations and are
+	 * required to adjust fdblocks using the xfs_metafile_resv_* helpers.
+	 */
+	XFS_AG_RESV_METAFILE,
 };
 
 /* Results of scanning a btree keyspace to check occupancy. */
diff --git a/fs/xfs/xfs_error.c b/fs/xfs/xfs_error.c
index 78cdc5064a8c47..dbd87e1376943a 100644
--- a/fs/xfs/xfs_error.c
+++ b/fs/xfs/xfs_error.c
@@ -63,6 +63,7 @@ static unsigned int xfs_errortag_random_default[] = {
 	XFS_RANDOM_WB_DELAY_MS,
 	XFS_RANDOM_WRITE_DELAY_MS,
 	XFS_RANDOM_EXCHMAPS_FINISH_ONE,
+	XFS_RANDOM_METAFILE_RESV_CRITICAL,
 };
 
 struct xfs_errortag_attr {
@@ -181,6 +182,7 @@ XFS_ERRORTAG_ATTR_RW(attr_leaf_to_node,	XFS_ERRTAG_ATTR_LEAF_TO_NODE);
 XFS_ERRORTAG_ATTR_RW(wb_delay_ms,	XFS_ERRTAG_WB_DELAY_MS);
 XFS_ERRORTAG_ATTR_RW(write_delay_ms,	XFS_ERRTAG_WRITE_DELAY_MS);
 XFS_ERRORTAG_ATTR_RW(exchmaps_finish_one, XFS_ERRTAG_EXCHMAPS_FINISH_ONE);
+XFS_ERRORTAG_ATTR_RW(metafile_resv_crit, XFS_ERRTAG_METAFILE_RESV_CRITICAL);
 
 static struct attribute *xfs_errortag_attrs[] = {
 	XFS_ERRORTAG_ATTR_LIST(noerror),
@@ -227,6 +229,7 @@ static struct attribute *xfs_errortag_attrs[] = {
 	XFS_ERRORTAG_ATTR_LIST(wb_delay_ms),
 	XFS_ERRORTAG_ATTR_LIST(write_delay_ms),
 	XFS_ERRORTAG_ATTR_LIST(exchmaps_finish_one),
+	XFS_ERRORTAG_ATTR_LIST(metafile_resv_crit),
 	NULL,
 };
 ATTRIBUTE_GROUPS(xfs_errortag);
diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c
index 28dde215c89953..e1145107d8cbd1 100644
--- a/fs/xfs/xfs_fsops.c
+++ b/fs/xfs/xfs_fsops.c
@@ -21,6 +21,7 @@
 #include "xfs_ag.h"
 #include "xfs_ag_resv.h"
 #include "xfs_trace.h"
+#include "xfs_rtalloc.h"
 
 /*
  * Write new AG headers to disk. Non-transactional, but need to be
@@ -541,6 +542,19 @@ xfs_fs_reserve_ag_blocks(
 		xfs_warn(mp,
 	"Error %d reserving per-AG metadata reserve pool.", error);
 		xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
+		return error;
+	}
+
+	if (xfs_has_realtime(mp)) {
+		err2 = xfs_rt_resv_init(mp);
+		if (err2 && err2 != -ENOSPC) {
+			xfs_warn(mp,
+		"Error %d reserving realtime metadata reserve pool.", err2);
+			xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
+		}
+
+		if (err2 && !error)
+			error = err2;
 	}
 
 	return error;
@@ -555,6 +569,9 @@ xfs_fs_unreserve_ag_blocks(
 {
 	struct xfs_perag	*pag = NULL;
 
+	if (xfs_has_realtime(mp))
+		xfs_rt_resv_free(mp);
+
 	while ((pag = xfs_perag_next(mp, pag)))
 		xfs_ag_resv_free(pag);
 }
diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h
index 1141c2e8e123ae..c08093a65352ec 100644
--- a/fs/xfs/xfs_inode.h
+++ b/fs/xfs/xfs_inode.h
@@ -31,6 +31,12 @@ typedef struct xfs_inode {
 			struct xfs_dquot *i_gdquot;	/* group dquot */
 			struct xfs_dquot *i_pdquot;	/* project dquot */
 		};
+
+		/*
+		 * Space that has been set aside to accomodate expansions of a
+		 * metadata btree rooted in this file.
+		 */
+		uint64_t	i_meta_resv_asked;
 	};
 
 	/* Inode location stuff */
diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c
index 5918f433dba754..97137126b16f5a 100644
--- a/fs/xfs/xfs_mount.c
+++ b/fs/xfs/xfs_mount.c
@@ -650,6 +650,15 @@ xfs_agbtree_compute_maxlevels(
 	mp->m_agbtree_maxlevels = max(levels, mp->m_refc_maxlevels);
 }
 
+/* Compute maximum possible height for realtime btree types for this fs. */
+static inline void
+xfs_rtbtree_compute_maxlevels(
+	struct xfs_mount	*mp)
+{
+	/* This will be filled in later. */
+	mp->m_rtbtree_maxlevels = 0;
+}
+
 /*
  * This function does the following on an initial mount of a file system:
  *	- reads the superblock from disk and init the mount struct
@@ -721,6 +730,7 @@ xfs_mountfs(
 	xfs_refcountbt_compute_maxlevels(mp);
 
 	xfs_agbtree_compute_maxlevels(mp);
+	xfs_rtbtree_compute_maxlevels(mp);
 
 	/*
 	 * Check if sb_agblocks is aligned at stripe boundary.  If sb_agblocks
diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h
index db9dade7d22a19..ddb9d19a3a3d53 100644
--- a/fs/xfs/xfs_mount.h
+++ b/fs/xfs/xfs_mount.h
@@ -165,6 +165,7 @@ typedef struct xfs_mount {
 	uint			m_rmap_maxlevels; /* max rmap btree levels */
 	uint			m_refc_maxlevels; /* max refcount btree level */
 	unsigned int		m_agbtree_maxlevels; /* max level of all AG btrees */
+	unsigned int		m_rtbtree_maxlevels; /* max level of all rt btrees */
 	xfs_extlen_t		m_ag_prealloc_blocks; /* reserved ag blocks */
 	uint			m_alloc_set_aside; /* space we can't use */
 	uint			m_ag_max_usable; /* max space per AG */
diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c
index fcfa6e0eb3ad2a..5128c5ad72f5da 100644
--- a/fs/xfs/xfs_rtalloc.c
+++ b/fs/xfs/xfs_rtalloc.c
@@ -1344,6 +1344,12 @@ xfs_growfs_rt(
 
 		if (!error)
 			error = error2;
+
+		/* Reset the rt metadata btree space reservations. */
+		xfs_rt_resv_free(mp);
+		error2 = xfs_rt_resv_init(mp);
+		if (error2 && error2 != -ENOSPC)
+			error = error2;
 	}
 
 out_unlock:
@@ -1487,6 +1493,21 @@ xfs_rtalloc_reinit_frextents(
 	return 0;
 }
 
+/* Free space reservations for rt metadata inodes. */
+void
+xfs_rt_resv_free(
+	struct xfs_mount	*mp)
+{
+}
+
+/* Reserve space for rt metadata inodes' space expansion. */
+int
+xfs_rt_resv_init(
+	struct xfs_mount	*mp)
+{
+	return 0;
+}
+
 /*
  * Read in the bmbt of an rt metadata inode so that we never have to load them
  * at runtime.  This enables the use of shared ILOCKs for rtbitmap scans.  Use
diff --git a/fs/xfs/xfs_rtalloc.h b/fs/xfs/xfs_rtalloc.h
index 8e2a07b8174b74..d87523e6a55006 100644
--- a/fs/xfs/xfs_rtalloc.h
+++ b/fs/xfs/xfs_rtalloc.h
@@ -34,6 +34,9 @@ int					/* error */
 xfs_rtmount_inodes(
 	struct xfs_mount	*mp);	/* file system mount structure */
 
+void xfs_rt_resv_free(struct xfs_mount *mp);
+int xfs_rt_resv_init(struct xfs_mount *mp);
+
 /*
  * Grow the realtime area of the filesystem.
  */
@@ -60,6 +63,8 @@ xfs_rtmount_init(
 }
 # define xfs_rtmount_inodes(m)  (((mp)->m_sb.sb_rblocks == 0)? 0 : (-ENOSYS))
 # define xfs_rtunmount_inodes(m)
+# define xfs_rt_resv_free(mp)				((void)0)
+# define xfs_rt_resv_init(mp)				(0)
 #endif	/* CONFIG_XFS_RT */
 
 #endif	/* __XFS_RTALLOC_H__ */
diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h
index 7b16cdd72e9ddd..cbda663fe6e817 100644
--- a/fs/xfs/xfs_trace.h
+++ b/fs/xfs/xfs_trace.h
@@ -5574,6 +5574,51 @@ DEFINE_EVENT(xfs_metadir_class, name, \
 	TP_ARGS(dp, name, ino))
 DEFINE_METADIR_EVENT(xfs_metadir_lookup);
 
+/* metadata inode space reservations */
+
+DECLARE_EVENT_CLASS(xfs_metafile_resv_class,
+	TP_PROTO(struct xfs_inode *ip, xfs_filblks_t len),
+	TP_ARGS(ip, len),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(unsigned long long, freeblks)
+		__field(unsigned long long, reserved)
+		__field(unsigned long long, asked)
+		__field(unsigned long long, used)
+		__field(unsigned long long, len)
+	),
+	TP_fast_assign(
+		struct xfs_mount *mp = ip->i_mount;
+
+		__entry->dev = mp->m_super->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->freeblks = percpu_counter_sum(&mp->m_fdblocks);
+		__entry->reserved = ip->i_delayed_blks;
+		__entry->asked = ip->i_meta_resv_asked;
+		__entry->used = ip->i_nblocks;
+		__entry->len = len;
+	),
+	TP_printk("dev %d:%d ino 0x%llx freeblks %llu resv %llu ask %llu used %llu len %llu",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->freeblks,
+		  __entry->reserved,
+		  __entry->asked,
+		  __entry->used,
+		  __entry->len)
+)
+#define DEFINE_METAFILE_RESV_EVENT(name) \
+DEFINE_EVENT(xfs_metafile_resv_class, name, \
+	TP_PROTO(struct xfs_inode *ip, xfs_filblks_t len), \
+	TP_ARGS(ip, len))
+DEFINE_METAFILE_RESV_EVENT(xfs_metafile_resv_init);
+DEFINE_METAFILE_RESV_EVENT(xfs_metafile_resv_free);
+DEFINE_METAFILE_RESV_EVENT(xfs_metafile_resv_alloc_space);
+DEFINE_METAFILE_RESV_EVENT(xfs_metafile_resv_free_space);
+DEFINE_METAFILE_RESV_EVENT(xfs_metafile_resv_critical);
+DEFINE_INODE_ERROR_EVENT(xfs_metafile_resv_init_error);
+
 #endif /* _TRACE_XFS_H */
 
 #undef TRACE_INCLUDE_PATH

Re: [PATCH 2/2] xfs: allow inode-based btrees to reserve space in the data device

[Christoph Hellwig]

Looks good:

Reviewed-by: Christoph Hellwig <hch@lst.de>

[PATCHSET v6.1 4/5] xfs: realtime reverse-mapping support

[Darrick J. Wong]

Hi all,

This is the latest revision of a patchset that adds to XFS kernel
support for reverse mapping for the realtime device.  This time around
I've fixed some of the bitrot that I've noticed over the past few
months, and most notably have converted rtrmapbt to use the metadata
inode directory feature instead of burning more space in the superblock.

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 enhance the reverse mapping routines to handle
the parts that are specific to rtgroups -- adding the new btree type,
adding a new log intent item type, and wiring up the metadata directory
tree entries.

Finally, implement GETFSMAP with the rtrmapbt and scrub functionality
for the rtrmapbt and rtbitmap and online fsck functionality.

If you're going to start using this code, I strongly recommend pulling
from my git trees, which are linked below.

This has been running on the djcloud for months with no problems.  Enjoy!
Comments and questions are, as always, welcome.

--D

kernel git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfs-linux.git/log/?h=realtime-rmap

xfsprogs git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfsprogs-dev.git/log/?h=realtime-rmap

fstests git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfstests-dev.git/log/?h=realtime-rmap

xfsdocs git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfs-documentation.git/log/?h=realtime-rmap
---
Commits in this patchset:
 * xfs: add some rtgroup inode helpers
 * xfs: prepare rmap btree cursor tracepoints for realtime
 * xfs: simplify the xfs_rmap_{alloc,free}_extent calling conventions
 * xfs: introduce realtime rmap btree ondisk definitions
 * xfs: realtime rmap btree transaction reservations
 * xfs: add realtime rmap btree operations
 * xfs: prepare rmap functions to deal with rtrmapbt
 * xfs: add a realtime flag to the rmap update log redo items
 * xfs: support recovering rmap intent items targetting realtime extents
 * xfs: pretty print metadata file types in error messages
 * xfs: support file data forks containing metadata btrees
 * xfs: add realtime reverse map inode to metadata directory
 * xfs: add metadata reservations for realtime rmap btrees
 * xfs: wire up a new metafile type for the realtime rmap
 * xfs: wire up rmap map and unmap to the realtime rmapbt
 * xfs: create routine to allocate and initialize a realtime rmap btree inode
 * xfs: wire up getfsmap to the realtime reverse mapping btree
 * xfs: check that the rtrmapbt maxlevels doesn't increase when growing fs
 * xfs: report realtime rmap btree corruption errors to the health system
 * xfs: allow queued realtime intents to drain before scrubbing
 * xfs: scrub the realtime rmapbt
 * xfs: cross-reference realtime bitmap to realtime rmapbt scrubber
 * xfs: cross-reference the realtime rmapbt
 * xfs: scan rt rmap when we're doing an intense rmap check of bmbt mappings
 * xfs: scrub the metadir path of rt rmap btree files
 * xfs: walk the rt reverse mapping tree when rebuilding rmap
 * xfs: online repair of realtime file bmaps
 * xfs: repair inodes that have realtime extents
 * xfs: repair rmap btree inodes
 * xfs: online repair of realtime bitmaps for a realtime group
 * xfs: support repairing metadata btrees rooted in metadir inodes
 * xfs: online repair of the realtime rmap btree
 * xfs: create a shadow rmap btree during realtime rmap repair
 * xfs: hook live realtime rmap operations during a repair operation
 * xfs: don't shut down the filesystem for media failures beyond end of log
 * xfs: react to fsdax failure notifications on the rt device
 * xfs: enable realtime rmap btree
---
 fs/xfs/Makefile                   |    3 
 fs/xfs/libxfs/xfs_btree.c         |   73 +++
 fs/xfs/libxfs/xfs_btree.h         |    8 
 fs/xfs/libxfs/xfs_btree_mem.c     |    1 
 fs/xfs/libxfs/xfs_btree_staging.c |    1 
 fs/xfs/libxfs/xfs_defer.h         |    1 
 fs/xfs/libxfs/xfs_exchmaps.c      |    4 
 fs/xfs/libxfs/xfs_format.h        |   28 +
 fs/xfs/libxfs/xfs_fs.h            |    7 
 fs/xfs/libxfs/xfs_health.h        |    4 
 fs/xfs/libxfs/xfs_inode_buf.c     |   32 +
 fs/xfs/libxfs/xfs_inode_fork.c    |   25 +
 fs/xfs/libxfs/xfs_log_format.h    |    6 
 fs/xfs/libxfs/xfs_log_recover.h   |    2 
 fs/xfs/libxfs/xfs_metafile.c      |   18 +
 fs/xfs/libxfs/xfs_metafile.h      |    2 
 fs/xfs/libxfs/xfs_ondisk.h        |    2 
 fs/xfs/libxfs/xfs_refcount.c      |    6 
 fs/xfs/libxfs/xfs_rmap.c          |  171 +++++-
 fs/xfs/libxfs/xfs_rmap.h          |   12 
 fs/xfs/libxfs/xfs_rtbitmap.c      |    2 
 fs/xfs/libxfs/xfs_rtbitmap.h      |    9 
 fs/xfs/libxfs/xfs_rtgroup.c       |   53 +-
 fs/xfs/libxfs/xfs_rtgroup.h       |   49 ++
 fs/xfs/libxfs/xfs_rtrmap_btree.c  | 1011 +++++++++++++++++++++++++++++++++++++
 fs/xfs/libxfs/xfs_rtrmap_btree.h  |  210 ++++++++
 fs/xfs/libxfs/xfs_sb.c            |    6 
 fs/xfs/libxfs/xfs_shared.h        |   14 +
 fs/xfs/libxfs/xfs_trans_resv.c    |   12 
 fs/xfs/libxfs/xfs_trans_space.h   |   13 
 fs/xfs/scrub/alloc_repair.c       |    5 
 fs/xfs/scrub/bmap.c               |  108 +++-
 fs/xfs/scrub/bmap_repair.c        |  129 +++++
 fs/xfs/scrub/common.c             |  160 ++++++
 fs/xfs/scrub/common.h             |   23 +
 fs/xfs/scrub/health.c             |    1 
 fs/xfs/scrub/inode.c              |   10 
 fs/xfs/scrub/inode_repair.c       |  136 +++++
 fs/xfs/scrub/metapath.c           |    3 
 fs/xfs/scrub/newbt.c              |   42 ++
 fs/xfs/scrub/newbt.h              |    1 
 fs/xfs/scrub/reap.c               |   41 ++
 fs/xfs/scrub/reap.h               |    2 
 fs/xfs/scrub/repair.c             |  191 +++++++
 fs/xfs/scrub/repair.h             |   17 +
 fs/xfs/scrub/rgsuper.c            |    6 
 fs/xfs/scrub/rmap_repair.c        |   84 +++
 fs/xfs/scrub/rtbitmap.c           |   75 ++-
 fs/xfs/scrub/rtbitmap.h           |   55 ++
 fs/xfs/scrub/rtbitmap_repair.c    |  429 +++++++++++++++-
 fs/xfs/scrub/rtrmap.c             |  271 ++++++++++
 fs/xfs/scrub/rtrmap_repair.c      |  903 +++++++++++++++++++++++++++++++++
 fs/xfs/scrub/rtsummary.c          |   17 -
 fs/xfs/scrub/rtsummary_repair.c   |    3 
 fs/xfs/scrub/scrub.c              |   11 
 fs/xfs/scrub/scrub.h              |   14 +
 fs/xfs/scrub/stats.c              |    1 
 fs/xfs/scrub/tempexch.h           |    2 
 fs/xfs/scrub/tempfile.c           |   20 -
 fs/xfs/scrub/trace.c              |    1 
 fs/xfs/scrub/trace.h              |  228 ++++++++
 fs/xfs/xfs_buf.c                  |    1 
 fs/xfs/xfs_buf_item_recover.c     |    4 
 fs/xfs/xfs_drain.c                |   20 -
 fs/xfs/xfs_drain.h                |    7 
 fs/xfs/xfs_fsmap.c                |  174 ++++++
 fs/xfs/xfs_fsops.c                |   11 
 fs/xfs/xfs_health.c               |    1 
 fs/xfs/xfs_inode.c                |   19 +
 fs/xfs/xfs_inode_item.c           |    2 
 fs/xfs/xfs_inode_item_recover.c   |   44 +-
 fs/xfs/xfs_log_recover.c          |    2 
 fs/xfs/xfs_mount.c                |    5 
 fs/xfs/xfs_mount.h                |    9 
 fs/xfs/xfs_notify_failure.c       |  230 +++++---
 fs/xfs/xfs_notify_failure.h       |   11 
 fs/xfs/xfs_qm.c                   |    8 
 fs/xfs/xfs_rmap_item.c            |  216 +++++++-
 fs/xfs/xfs_rtalloc.c              |   82 ++-
 fs/xfs/xfs_rtalloc.h              |   10 
 fs/xfs/xfs_stats.c                |    4 
 fs/xfs/xfs_stats.h                |    2 
 fs/xfs/xfs_super.c                |    6 
 fs/xfs/xfs_super.h                |    1 
 fs/xfs/xfs_trace.h                |  104 ++--
 85 files changed, 5381 insertions(+), 366 deletions(-)
 create mode 100644 fs/xfs/libxfs/xfs_rtrmap_btree.c
 create mode 100644 fs/xfs/libxfs/xfs_rtrmap_btree.h
 create mode 100644 fs/xfs/scrub/rtrmap.c
 create mode 100644 fs/xfs/scrub/rtrmap_repair.c
 create mode 100644 fs/xfs/xfs_notify_failure.h

[PATCH 01/37] xfs: add some rtgroup inode helpers

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Create some simple helpers to reduce the amount of typing whenever we
access rtgroup inodes.  Conversion was done with this spatch and some
minor reformatting:

@@
expression rtg;
@@

- rtg->rtg_inodes[XFS_RTGI_BITMAP]
+ rtg_bitmap(rtg)

@@
expression rtg;
@@

- rtg->rtg_inodes[XFS_RTGI_SUMMARY]
+ rtg_summary(rtg)

and the CLI command:

$ spatch --sp-file /tmp/moo.cocci --dir fs/xfs/ --use-gitgrep --in-place

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_rtbitmap.c |    2 +-
 fs/xfs/libxfs/xfs_rtgroup.c  |   18 ++++++++----------
 fs/xfs/libxfs/xfs_rtgroup.h  |   10 ++++++++++
 fs/xfs/scrub/rtbitmap.c      |    7 +++----
 fs/xfs/scrub/rtsummary.c     |   12 +++++-------
 fs/xfs/xfs_qm.c              |    8 ++++----
 fs/xfs/xfs_rtalloc.c         |   10 +++++-----
 7 files changed, 36 insertions(+), 31 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_rtbitmap.c b/fs/xfs/libxfs/xfs_rtbitmap.c
index 4ddfb7e395b38a..770adf60dd7392 100644
--- a/fs/xfs/libxfs/xfs_rtbitmap.c
+++ b/fs/xfs/libxfs/xfs_rtbitmap.c
@@ -1055,7 +1055,7 @@ xfs_rtfree_extent(
 	xfs_rtxlen_t		len)	/* length of extent freed */
 {
 	struct xfs_mount	*mp = tp->t_mountp;
-	struct xfs_inode	*rbmip = rtg->rtg_inodes[XFS_RTGI_BITMAP];
+	struct xfs_inode	*rbmip = rtg_bitmap(rtg);
 	struct xfs_rtalloc_args	args = {
 		.mp		= mp,
 		.tp		= tp,
diff --git a/fs/xfs/libxfs/xfs_rtgroup.c b/fs/xfs/libxfs/xfs_rtgroup.c
index 4f3bfc884aff29..a79b734e70440d 100644
--- a/fs/xfs/libxfs/xfs_rtgroup.c
+++ b/fs/xfs/libxfs/xfs_rtgroup.c
@@ -197,10 +197,10 @@ xfs_rtgroup_lock(
 		 * Lock both realtime free space metadata inodes for a freespace
 		 * update.
 		 */
-		xfs_ilock(rtg->rtg_inodes[XFS_RTGI_BITMAP], XFS_ILOCK_EXCL);
-		xfs_ilock(rtg->rtg_inodes[XFS_RTGI_SUMMARY], XFS_ILOCK_EXCL);
+		xfs_ilock(rtg_bitmap(rtg), XFS_ILOCK_EXCL);
+		xfs_ilock(rtg_summary(rtg), XFS_ILOCK_EXCL);
 	} else if (rtglock_flags & XFS_RTGLOCK_BITMAP_SHARED) {
-		xfs_ilock(rtg->rtg_inodes[XFS_RTGI_BITMAP], XFS_ILOCK_SHARED);
+		xfs_ilock(rtg_bitmap(rtg), XFS_ILOCK_SHARED);
 	}
 }
 
@@ -215,10 +215,10 @@ xfs_rtgroup_unlock(
 	       !(rtglock_flags & XFS_RTGLOCK_BITMAP));
 
 	if (rtglock_flags & XFS_RTGLOCK_BITMAP) {
-		xfs_iunlock(rtg->rtg_inodes[XFS_RTGI_SUMMARY], XFS_ILOCK_EXCL);
-		xfs_iunlock(rtg->rtg_inodes[XFS_RTGI_BITMAP], XFS_ILOCK_EXCL);
+		xfs_iunlock(rtg_summary(rtg), XFS_ILOCK_EXCL);
+		xfs_iunlock(rtg_bitmap(rtg), XFS_ILOCK_EXCL);
 	} else if (rtglock_flags & XFS_RTGLOCK_BITMAP_SHARED) {
-		xfs_iunlock(rtg->rtg_inodes[XFS_RTGI_BITMAP], XFS_ILOCK_SHARED);
+		xfs_iunlock(rtg_bitmap(rtg), XFS_ILOCK_SHARED);
 	}
 }
 
@@ -236,10 +236,8 @@ xfs_rtgroup_trans_join(
 	ASSERT(!(rtglock_flags & XFS_RTGLOCK_BITMAP_SHARED));
 
 	if (rtglock_flags & XFS_RTGLOCK_BITMAP) {
-		xfs_trans_ijoin(tp, rtg->rtg_inodes[XFS_RTGI_BITMAP],
-				XFS_ILOCK_EXCL);
-		xfs_trans_ijoin(tp, rtg->rtg_inodes[XFS_RTGI_SUMMARY],
-				XFS_ILOCK_EXCL);
+		xfs_trans_ijoin(tp, rtg_bitmap(rtg), XFS_ILOCK_EXCL);
+		xfs_trans_ijoin(tp, rtg_summary(rtg), XFS_ILOCK_EXCL);
 	}
 }
 
diff --git a/fs/xfs/libxfs/xfs_rtgroup.h b/fs/xfs/libxfs/xfs_rtgroup.h
index 7e7e491ff06fa5..19f8d302b9aa3f 100644
--- a/fs/xfs/libxfs/xfs_rtgroup.h
+++ b/fs/xfs/libxfs/xfs_rtgroup.h
@@ -64,6 +64,16 @@ static inline xfs_rgnumber_t rtg_rgno(const struct xfs_rtgroup *rtg)
 	return rtg->rtg_group.xg_gno;
 }
 
+static inline struct xfs_inode *rtg_bitmap(const struct xfs_rtgroup *rtg)
+{
+	return rtg->rtg_inodes[XFS_RTGI_BITMAP];
+}
+
+static inline struct xfs_inode *rtg_summary(const struct xfs_rtgroup *rtg)
+{
+	return rtg->rtg_inodes[XFS_RTGI_SUMMARY];
+}
+
 /* Passive rtgroup references */
 static inline struct xfs_rtgroup *
 xfs_rtgroup_get(
diff --git a/fs/xfs/scrub/rtbitmap.c b/fs/xfs/scrub/rtbitmap.c
index 376a36fd9a9cdd..fb4970c877abd3 100644
--- a/fs/xfs/scrub/rtbitmap.c
+++ b/fs/xfs/scrub/rtbitmap.c
@@ -49,8 +49,7 @@ xchk_setup_rtbitmap(
 	if (error)
 		return error;
 
-	error = xchk_install_live_inode(sc,
-			sc->sr.rtg->rtg_inodes[XFS_RTGI_BITMAP]);
+	error = xchk_install_live_inode(sc, rtg_bitmap(sc->sr.rtg));
 	if (error)
 		return error;
 
@@ -146,7 +145,7 @@ xchk_rtbitmap(
 {
 	struct xfs_mount	*mp = sc->mp;
 	struct xfs_rtgroup	*rtg = sc->sr.rtg;
-	struct xfs_inode	*rbmip = rtg->rtg_inodes[XFS_RTGI_BITMAP];
+	struct xfs_inode	*rbmip = rtg_bitmap(rtg);
 	struct xchk_rtbitmap	*rtb = sc->buf;
 	int			error;
 
@@ -215,7 +214,7 @@ xchk_xref_is_used_rt_space(
 	xfs_extlen_t		len)
 {
 	struct xfs_rtgroup	*rtg = sc->sr.rtg;
-	struct xfs_inode	*rbmip = rtg->rtg_inodes[XFS_RTGI_BITMAP];
+	struct xfs_inode	*rbmip = rtg_bitmap(rtg);
 	xfs_rtxnum_t		startext;
 	xfs_rtxnum_t		endext;
 	bool			is_free;
diff --git a/fs/xfs/scrub/rtsummary.c b/fs/xfs/scrub/rtsummary.c
index 49fc6250bafcaa..f1af5431b38856 100644
--- a/fs/xfs/scrub/rtsummary.c
+++ b/fs/xfs/scrub/rtsummary.c
@@ -81,8 +81,7 @@ xchk_setup_rtsummary(
 	if (error)
 		return error;
 
-	error = xchk_install_live_inode(sc,
-			sc->sr.rtg->rtg_inodes[XFS_RTGI_SUMMARY]);
+	error = xchk_install_live_inode(sc, rtg_summary(sc->sr.rtg));
 	if (error)
 		return error;
 
@@ -191,8 +190,7 @@ xchk_rtsum_record_free(
 	rtlen = xfs_rtxlen_to_extlen(mp, rec->ar_extcount);
 
 	if (!xfs_verify_rtbext(mp, rtbno, rtlen)) {
-		xchk_ino_xref_set_corrupt(sc,
-				rtg->rtg_inodes[XFS_RTGI_BITMAP]->i_ino);
+		xchk_ino_xref_set_corrupt(sc, rtg_bitmap(rtg)->i_ino);
 		return -EFSCORRUPTED;
 	}
 
@@ -218,7 +216,7 @@ xchk_rtsum_compute(
 
 	/* If the bitmap size doesn't match the computed size, bail. */
 	if (XFS_FSB_TO_B(mp, xfs_rtbitmap_blockcount(mp)) !=
-	    rtg->rtg_inodes[XFS_RTGI_BITMAP]->i_disk_size)
+	    rtg_bitmap(rtg)->i_disk_size)
 		return -EFSCORRUPTED;
 
 	return xfs_rtalloc_query_all(rtg, sc->tp, xchk_rtsum_record_free, sc);
@@ -310,8 +308,8 @@ xchk_rtsummary(
 {
 	struct xfs_mount	*mp = sc->mp;
 	struct xfs_rtgroup	*rtg = sc->sr.rtg;
-	struct xfs_inode	*rbmip = rtg->rtg_inodes[XFS_RTGI_BITMAP];
-	struct xfs_inode	*rsumip = rtg->rtg_inodes[XFS_RTGI_SUMMARY];
+	struct xfs_inode	*rbmip = rtg_bitmap(rtg);
+	struct xfs_inode	*rsumip = rtg_summary(rtg);
 	struct xchk_rtsummary	*rts = sc->buf;
 	int			error;
 
diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c
index 3abab5fb593e37..e1ba5af6250f0b 100644
--- a/fs/xfs/xfs_qm.c
+++ b/fs/xfs/xfs_qm.c
@@ -230,10 +230,10 @@ xfs_qm_unmount_rt(
 
 	if (!rtg)
 		return;
-	if (rtg->rtg_inodes[XFS_RTGI_BITMAP])
-		xfs_qm_dqdetach(rtg->rtg_inodes[XFS_RTGI_BITMAP]);
-	if (rtg->rtg_inodes[XFS_RTGI_SUMMARY])
-		xfs_qm_dqdetach(rtg->rtg_inodes[XFS_RTGI_SUMMARY]);
+	if (rtg_bitmap(rtg))
+		xfs_qm_dqdetach(rtg_bitmap(rtg));
+	if (rtg_summary(rtg))
+		xfs_qm_dqdetach(rtg_summary(rtg));
 	xfs_rtgroup_rele(rtg);
 }
 
diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c
index 5128c5ad72f5da..4cd2f32aa70a0a 100644
--- a/fs/xfs/xfs_rtalloc.c
+++ b/fs/xfs/xfs_rtalloc.c
@@ -856,8 +856,8 @@ xfs_growfs_rt_bmblock(
 	xfs_fileoff_t		bmbno)
 {
 	struct xfs_mount	*mp = rtg_mount(rtg);
-	struct xfs_inode	*rbmip = rtg->rtg_inodes[XFS_RTGI_BITMAP];
-	struct xfs_inode	*rsumip = rtg->rtg_inodes[XFS_RTGI_SUMMARY];
+	struct xfs_inode	*rbmip = rtg_bitmap(rtg);
+	struct xfs_inode	*rsumip = rtg_summary(rtg);
 	struct xfs_rtalloc_args	args = {
 		.mp		= mp,
 		.rtg		= rtg,
@@ -1041,8 +1041,8 @@ xfs_growfs_rt_alloc_blocks(
 	xfs_extlen_t		*nrbmblocks)
 {
 	struct xfs_mount	*mp = rtg_mount(rtg);
-	struct xfs_inode	*rbmip = rtg->rtg_inodes[XFS_RTGI_BITMAP];
-	struct xfs_inode	*rsumip = rtg->rtg_inodes[XFS_RTGI_SUMMARY];
+	struct xfs_inode	*rbmip = rtg_bitmap(rtg);
+	struct xfs_inode	*rsumip = rtg_summary(rtg);
 	xfs_extlen_t		orbmblocks = 0;
 	xfs_extlen_t		orsumblocks = 0;
 	struct xfs_mount	*nmp;
@@ -1622,7 +1622,7 @@ xfs_rtpick_extent(
 	xfs_rtxlen_t		len)		/* allocation length (rtextents) */
 {
 	struct xfs_mount	*mp = rtg_mount(rtg);
-	struct xfs_inode	*rbmip = rtg->rtg_inodes[XFS_RTGI_BITMAP];
+	struct xfs_inode	*rbmip = rtg_bitmap(rtg);
 	xfs_rtxnum_t		b = 0;		/* result rtext */
 	int			log2;		/* log of sequence number */
 	uint64_t		resid;		/* residual after log removed */

[PATCH 02/37] xfs: prepare rmap btree cursor tracepoints for realtime

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Rework the rmap btree cursor tracepoints in preparation to handle the
realtime rmap btree cursor.  Mostly this involves renaming the field to
"gbno" and extracting the group number from the cursor.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/xfs_rmap_item.c |    4 +-
 fs/xfs/xfs_trace.h     |   82 +++++++++++++++++++++++++-----------------------
 2 files changed, 44 insertions(+), 42 deletions(-)


diff --git a/fs/xfs/xfs_rmap_item.c b/fs/xfs/xfs_rmap_item.c
index 76b3c0ed3b4f63..ac2913a7335871 100644
--- a/fs/xfs/xfs_rmap_item.c
+++ b/fs/xfs/xfs_rmap_item.c
@@ -351,10 +351,10 @@ xfs_rmap_defer_add(
 {
 	struct xfs_mount	*mp = tp->t_mountp;
 
-	trace_xfs_rmap_defer(mp, ri);
-
 	ri->ri_group = xfs_group_intent_get(mp, ri->ri_bmap.br_startblock,
 			XG_TYPE_AG);
+
+	trace_xfs_rmap_defer(mp, ri);
 	xfs_defer_add(tp, &ri->ri_list, &xfs_rmap_update_defer_type);
 }
 
diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h
index cbda663fe6e817..8b7bb1f5ae3c6f 100644
--- a/fs/xfs/xfs_trace.h
+++ b/fs/xfs/xfs_trace.h
@@ -14,11 +14,15 @@
  * ino: filesystem inode number
  *
  * agbno: per-AG block number in fs blocks
+ * rgbno: per-rtgroup block number in fs blocks
  * startblock: physical block number for file mappings.  This is either a
  *             segmented fsblock for data device mappings, or a rfsblock
  *             for realtime device mappings
  * fsbcount: number of blocks in an extent, in fs blocks
  *
+ * gbno: generic allocation group block number.  This is an agbno for
+ *       space in a per-AG or a rgbno for space in a realtime group.
+ *
  * daddr: physical block number in 512b blocks
  * bbcount: number of blocks in a physical extent, in 512b blocks
  *
@@ -2918,13 +2922,14 @@ DEFINE_DEFER_PENDING_ITEM_EVENT(xfs_defer_finish_item);
 /* rmap tracepoints */
 DECLARE_EVENT_CLASS(xfs_rmap_class,
 	TP_PROTO(struct xfs_btree_cur *cur,
-		 xfs_agblock_t agbno, xfs_extlen_t len, bool unwritten,
+		 xfs_agblock_t gbno, xfs_extlen_t len, bool unwritten,
 		 const struct xfs_owner_info *oinfo),
-	TP_ARGS(cur, agbno, len, unwritten, oinfo),
+	TP_ARGS(cur, gbno, len, unwritten, oinfo),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
+		__field(enum xfs_group_type, type)
 		__field(xfs_agnumber_t, agno)
-		__field(xfs_agblock_t, agbno)
+		__field(xfs_agblock_t, gbno)
 		__field(xfs_extlen_t, len)
 		__field(uint64_t, owner)
 		__field(uint64_t, offset)
@@ -2932,8 +2937,9 @@ DECLARE_EVENT_CLASS(xfs_rmap_class,
 	),
 	TP_fast_assign(
 		__entry->dev = cur->bc_mp->m_super->s_dev;
+		__entry->type = cur->bc_group->xg_type;
 		__entry->agno = cur->bc_group->xg_gno;
-		__entry->agbno = agbno;
+		__entry->gbno = gbno;
 		__entry->len = len;
 		__entry->owner = oinfo->oi_owner;
 		__entry->offset = oinfo->oi_offset;
@@ -2941,10 +2947,11 @@ DECLARE_EVENT_CLASS(xfs_rmap_class,
 		if (unwritten)
 			__entry->flags |= XFS_RMAP_UNWRITTEN;
 	),
-	TP_printk("dev %d:%d agno 0x%x agbno 0x%x fsbcount 0x%x owner 0x%llx fileoff 0x%llx flags 0x%lx",
+	TP_printk("dev %d:%d %sno 0x%x gbno 0x%x fsbcount 0x%x owner 0x%llx fileoff 0x%llx flags 0x%lx",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __print_symbolic(__entry->type, XG_TYPE_STRINGS),
 		  __entry->agno,
-		  __entry->agbno,
+		  __entry->gbno,
 		  __entry->len,
 		  __entry->owner,
 		  __entry->offset,
@@ -2953,9 +2960,9 @@ DECLARE_EVENT_CLASS(xfs_rmap_class,
 #define DEFINE_RMAP_EVENT(name) \
 DEFINE_EVENT(xfs_rmap_class, name, \
 	TP_PROTO(struct xfs_btree_cur *cur, \
-		 xfs_agblock_t agbno, xfs_extlen_t len, bool unwritten, \
+		 xfs_agblock_t gbno, xfs_extlen_t len, bool unwritten, \
 		 const struct xfs_owner_info *oinfo), \
-	TP_ARGS(cur, agbno, len, unwritten, oinfo))
+	TP_ARGS(cur, gbno, len, unwritten, oinfo))
 
 /* btree cursor error/%ip tracepoint class */
 DECLARE_EVENT_CLASS(xfs_btree_error_class,
@@ -3018,47 +3025,36 @@ TRACE_EVENT(xfs_rmap_convert_state,
 	TP_ARGS(cur, state, caller_ip),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
+		__field(enum xfs_group_type, type)
 		__field(xfs_agnumber_t, agno)
-		__field(xfs_ino_t, ino)
 		__field(int, state)
 		__field(unsigned long, caller_ip)
 	),
 	TP_fast_assign(
 		__entry->dev = cur->bc_mp->m_super->s_dev;
-		switch (cur->bc_ops->type) {
-		case XFS_BTREE_TYPE_INODE:
-			__entry->agno = 0;
-			__entry->ino = cur->bc_ino.ip->i_ino;
-			break;
-		case XFS_BTREE_TYPE_AG:
-			__entry->agno = cur->bc_group->xg_gno;
-			__entry->ino = 0;
-			break;
-		case XFS_BTREE_TYPE_MEM:
-			__entry->agno = 0;
-			__entry->ino = 0;
-			break;
-		}
+		__entry->type = cur->bc_group->xg_type;
+		__entry->agno = cur->bc_group->xg_gno;
 		__entry->state = state;
 		__entry->caller_ip = caller_ip;
 	),
-	TP_printk("dev %d:%d agno 0x%x ino 0x%llx state %d caller %pS",
+	TP_printk("dev %d:%d %sno 0x%x state %d caller %pS",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __print_symbolic(__entry->type, XG_TYPE_STRINGS),
 		  __entry->agno,
-		  __entry->ino,
 		  __entry->state,
 		  (char *)__entry->caller_ip)
 );
 
 DECLARE_EVENT_CLASS(xfs_rmapbt_class,
 	TP_PROTO(struct xfs_btree_cur *cur,
-		 xfs_agblock_t agbno, xfs_extlen_t len,
+		 xfs_agblock_t gbno, xfs_extlen_t len,
 		 uint64_t owner, uint64_t offset, unsigned int flags),
-	TP_ARGS(cur, agbno, len, owner, offset, flags),
+	TP_ARGS(cur, gbno, len, owner, offset, flags),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
+		__field(enum xfs_group_type, type)
 		__field(xfs_agnumber_t, agno)
-		__field(xfs_agblock_t, agbno)
+		__field(xfs_agblock_t, gbno)
 		__field(xfs_extlen_t, len)
 		__field(uint64_t, owner)
 		__field(uint64_t, offset)
@@ -3066,17 +3062,19 @@ DECLARE_EVENT_CLASS(xfs_rmapbt_class,
 	),
 	TP_fast_assign(
 		__entry->dev = cur->bc_mp->m_super->s_dev;
+		__entry->type = cur->bc_group->xg_type;
 		__entry->agno = cur->bc_group->xg_gno;
-		__entry->agbno = agbno;
+		__entry->gbno = gbno;
 		__entry->len = len;
 		__entry->owner = owner;
 		__entry->offset = offset;
 		__entry->flags = flags;
 	),
-	TP_printk("dev %d:%d agno 0x%x agbno 0x%x fsbcount 0x%x owner 0x%llx fileoff 0x%llx flags 0x%x",
+	TP_printk("dev %d:%d %sno 0x%x gbno 0x%x fsbcount 0x%x owner 0x%llx fileoff 0x%llx flags 0x%x",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __print_symbolic(__entry->type, XG_TYPE_STRINGS),
 		  __entry->agno,
-		  __entry->agbno,
+		  __entry->gbno,
 		  __entry->len,
 		  __entry->owner,
 		  __entry->offset,
@@ -3085,9 +3083,9 @@ DECLARE_EVENT_CLASS(xfs_rmapbt_class,
 #define DEFINE_RMAPBT_EVENT(name) \
 DEFINE_EVENT(xfs_rmapbt_class, name, \
 	TP_PROTO(struct xfs_btree_cur *cur, \
-		 xfs_agblock_t agbno, xfs_extlen_t len, \
+		 xfs_agblock_t gbno, xfs_extlen_t len, \
 		 uint64_t owner, uint64_t offset, unsigned int flags), \
-	TP_ARGS(cur, agbno, len, owner, offset, flags))
+	TP_ARGS(cur, gbno, len, owner, offset, flags))
 
 TRACE_DEFINE_ENUM(XFS_RMAP_MAP);
 TRACE_DEFINE_ENUM(XFS_RMAP_MAP_SHARED);
@@ -3104,8 +3102,9 @@ DECLARE_EVENT_CLASS(xfs_rmap_deferred_class,
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
 		__field(unsigned long long, owner)
+		__field(enum xfs_group_type, type)
 		__field(xfs_agnumber_t, agno)
-		__field(xfs_agblock_t, agbno)
+		__field(xfs_agblock_t, gbno)
 		__field(int, whichfork)
 		__field(xfs_fileoff_t, l_loff)
 		__field(xfs_filblks_t, l_len)
@@ -3114,9 +3113,11 @@ DECLARE_EVENT_CLASS(xfs_rmap_deferred_class,
 	),
 	TP_fast_assign(
 		__entry->dev = mp->m_super->s_dev;
-		__entry->agno = XFS_FSB_TO_AGNO(mp, ri->ri_bmap.br_startblock);
-		__entry->agbno = XFS_FSB_TO_AGBNO(mp,
-					ri->ri_bmap.br_startblock);
+		__entry->type = ri->ri_group->xg_type;
+		__entry->agno = ri->ri_group->xg_gno;
+		__entry->gbno = xfs_fsb_to_gbno(mp,
+						ri->ri_bmap.br_startblock,
+						ri->ri_group->xg_type);
 		__entry->owner = ri->ri_owner;
 		__entry->whichfork = ri->ri_whichfork;
 		__entry->l_loff = ri->ri_bmap.br_startoff;
@@ -3124,11 +3125,12 @@ DECLARE_EVENT_CLASS(xfs_rmap_deferred_class,
 		__entry->l_state = ri->ri_bmap.br_state;
 		__entry->op = ri->ri_type;
 	),
-	TP_printk("dev %d:%d op %s agno 0x%x agbno 0x%x owner 0x%llx %s fileoff 0x%llx fsbcount 0x%llx state %d",
+	TP_printk("dev %d:%d op %s %sno 0x%x gbno 0x%x owner 0x%llx %s fileoff 0x%llx fsbcount 0x%llx state %d",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		  __print_symbolic(__entry->op, XFS_RMAP_INTENT_STRINGS),
+		  __print_symbolic(__entry->type, XG_TYPE_STRINGS),
 		  __entry->agno,
-		  __entry->agbno,
+		  __entry->gbno,
 		  __entry->owner,
 		  __print_symbolic(__entry->whichfork, XFS_WHICHFORK_STRINGS),
 		  __entry->l_loff,
@@ -3993,7 +3995,7 @@ TRACE_EVENT(xfs_fsmap_mapping,
 		__entry->offset = frec->offset;
 		__entry->flags = frec->rm_flags;
 	),
-	TP_printk("dev %d:%d keydev %d:%d agno 0x%x rmapbno 0x%x start_daddr 0x%llx len_daddr 0x%llx owner 0x%llx fileoff 0x%llx flags 0x%x",
+	TP_printk("dev %d:%d keydev %d:%d agno 0x%x gbno 0x%x start_daddr 0x%llx len_daddr 0x%llx owner 0x%llx fileoff 0x%llx flags 0x%x",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		  MAJOR(__entry->keydev), MINOR(__entry->keydev),
 		  __entry->agno,

[PATCH 03/37] xfs: simplify the xfs_rmap_{alloc,free}_extent calling conventions

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Simplify the calling conventions by allowing callers to pass a fsbno
(xfs_fsblock_t) directly into these functions, since we're just going to
set it in a struct anyway.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_refcount.c |    6 ++----
 fs/xfs/libxfs/xfs_rmap.c     |   12 +++++-------
 fs/xfs/libxfs/xfs_rmap.h     |    8 ++++----
 fs/xfs/scrub/alloc_repair.c  |    5 +++--
 4 files changed, 14 insertions(+), 17 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_refcount.c b/fs/xfs/libxfs/xfs_refcount.c
index 2dbab68b4fe69f..26d3d7956e069d 100644
--- a/fs/xfs/libxfs/xfs_refcount.c
+++ b/fs/xfs/libxfs/xfs_refcount.c
@@ -1831,8 +1831,7 @@ xfs_refcount_alloc_cow_extent(
 	__xfs_refcount_add(tp, XFS_REFCOUNT_ALLOC_COW, fsb, len);
 
 	/* Add rmap entry */
-	xfs_rmap_alloc_extent(tp, XFS_FSB_TO_AGNO(mp, fsb),
-			XFS_FSB_TO_AGBNO(mp, fsb), len, XFS_RMAP_OWN_COW);
+	xfs_rmap_alloc_extent(tp, fsb, len, XFS_RMAP_OWN_COW);
 }
 
 /* Forget a CoW staging event in the refcount btree. */
@@ -1848,8 +1847,7 @@ xfs_refcount_free_cow_extent(
 		return;
 
 	/* Remove rmap entry */
-	xfs_rmap_free_extent(tp, XFS_FSB_TO_AGNO(mp, fsb),
-			XFS_FSB_TO_AGBNO(mp, fsb), len, XFS_RMAP_OWN_COW);
+	xfs_rmap_free_extent(tp, fsb, len, XFS_RMAP_OWN_COW);
 	__xfs_refcount_add(tp, XFS_REFCOUNT_FREE_COW, fsb, len);
 }
 
diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c
index d0df68dc313185..57dbf99ce00453 100644
--- a/fs/xfs/libxfs/xfs_rmap.c
+++ b/fs/xfs/libxfs/xfs_rmap.c
@@ -525,7 +525,7 @@ xfs_rmap_free_check_owner(
 	struct xfs_btree_cur	*cur,
 	uint64_t		ltoff,
 	struct xfs_rmap_irec	*rec,
-	xfs_filblks_t		len,
+	xfs_extlen_t		len,
 	uint64_t		owner,
 	uint64_t		offset,
 	unsigned int		flags)
@@ -2729,8 +2729,7 @@ xfs_rmap_convert_extent(
 void
 xfs_rmap_alloc_extent(
 	struct xfs_trans	*tp,
-	xfs_agnumber_t		agno,
-	xfs_agblock_t		bno,
+	xfs_fsblock_t		fsbno,
 	xfs_extlen_t		len,
 	uint64_t		owner)
 {
@@ -2739,7 +2738,7 @@ xfs_rmap_alloc_extent(
 	if (!xfs_rmap_update_is_needed(tp->t_mountp, XFS_DATA_FORK))
 		return;
 
-	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;
@@ -2751,8 +2750,7 @@ xfs_rmap_alloc_extent(
 void
 xfs_rmap_free_extent(
 	struct xfs_trans	*tp,
-	xfs_agnumber_t		agno,
-	xfs_agblock_t		bno,
+	xfs_fsblock_t		fsbno,
 	xfs_extlen_t		len,
 	uint64_t		owner)
 {
@@ -2761,7 +2759,7 @@ xfs_rmap_free_extent(
 	if (!xfs_rmap_update_is_needed(tp->t_mountp, XFS_DATA_FORK))
 		return;
 
-	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;
diff --git a/fs/xfs/libxfs/xfs_rmap.h b/fs/xfs/libxfs/xfs_rmap.h
index 96b4321d831007..8e2657af038e9e 100644
--- a/fs/xfs/libxfs/xfs_rmap.h
+++ b/fs/xfs/libxfs/xfs_rmap.h
@@ -184,10 +184,10 @@ void xfs_rmap_unmap_extent(struct xfs_trans *tp, struct xfs_inode *ip,
 void xfs_rmap_convert_extent(struct xfs_mount *mp, struct xfs_trans *tp,
 		struct xfs_inode *ip, int whichfork,
 		struct xfs_bmbt_irec *imap);
-void xfs_rmap_alloc_extent(struct xfs_trans *tp, xfs_agnumber_t agno,
-		xfs_agblock_t bno, xfs_extlen_t len, uint64_t owner);
-void xfs_rmap_free_extent(struct xfs_trans *tp, xfs_agnumber_t agno,
-		xfs_agblock_t bno, xfs_extlen_t len, uint64_t owner);
+void xfs_rmap_alloc_extent(struct xfs_trans *tp, xfs_fsblock_t fsbno,
+		xfs_extlen_t len, uint64_t owner);
+void xfs_rmap_free_extent(struct xfs_trans *tp, xfs_fsblock_t fsbno,
+		xfs_extlen_t len, uint64_t owner);
 
 int xfs_rmap_finish_one(struct xfs_trans *tp, struct xfs_rmap_intent *ri,
 		struct xfs_btree_cur **pcur);
diff --git a/fs/xfs/scrub/alloc_repair.c b/fs/xfs/scrub/alloc_repair.c
index 0433363a90b616..11e1e5404fc6dc 100644
--- a/fs/xfs/scrub/alloc_repair.c
+++ b/fs/xfs/scrub/alloc_repair.c
@@ -542,8 +542,9 @@ xrep_abt_dispose_one(
 
 	/* Add a deferred rmap for each extent we used. */
 	if (resv->used > 0)
-		xfs_rmap_alloc_extent(sc->tp, pag_agno(pag), resv->agbno,
-				resv->used, XFS_RMAP_OWN_AG);
+		xfs_rmap_alloc_extent(sc->tp,
+				xfs_agbno_to_fsb(pag, resv->agbno), resv->used,
+				XFS_RMAP_OWN_AG);
 
 	/*
 	 * For each reserved btree block we didn't use, add it to the free

[PATCH 04/37] xfs: introduce realtime rmap btree ondisk definitions

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Add the ondisk structure definitions for realtime rmap btrees. The
realtime rmap btree will be rooted from a hidden inode so it needs to
have a separate btree block magic and pointer format.

Next, add everything needed to read, write and manipulate rmap btree
blocks. This prepares the way for connecting the btree operations
implementation, though embedding the rtrmap btree root in the inode
comes later in the series.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/Makefile                  |    1 
 fs/xfs/libxfs/xfs_btree.c        |    5 +
 fs/xfs/libxfs/xfs_format.h       |   10 +
 fs/xfs/libxfs/xfs_ondisk.h       |    1 
 fs/xfs/libxfs/xfs_rtrmap_btree.c |  271 ++++++++++++++++++++++++++++++++++++++
 fs/xfs/libxfs/xfs_rtrmap_btree.h |   82 +++++++++++
 fs/xfs/libxfs/xfs_sb.c           |    6 +
 fs/xfs/libxfs/xfs_shared.h       |    7 +
 fs/xfs/xfs_mount.c               |    5 -
 fs/xfs/xfs_mount.h               |    9 +
 fs/xfs/xfs_stats.c               |    3 
 fs/xfs/xfs_stats.h               |    1 
 12 files changed, 398 insertions(+), 3 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 ed9b0dabc1f11d..ff45efb2463f73 100644
--- a/fs/xfs/Makefile
+++ b/fs/xfs/Makefile
@@ -51,6 +51,7 @@ xfs-y				+= $(addprefix libxfs/, \
 				   xfs_rmap_btree.o \
 				   xfs_refcount.o \
 				   xfs_refcount_btree.o \
+				   xfs_rtrmap_btree.o \
 				   xfs_sb.o \
 				   xfs_symlink_remote.o \
 				   xfs_trans_inode.o \
diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c
index 5ab201ef041e7d..0e271919374780 100644
--- a/fs/xfs/libxfs/xfs_btree.c
+++ b/fs/xfs/libxfs/xfs_btree.c
@@ -30,6 +30,7 @@
 #include "xfs_health.h"
 #include "xfs_buf_mem.h"
 #include "xfs_btree_mem.h"
+#include "xfs_rtrmap_btree.h"
 
 /*
  * Btree magic numbers.
@@ -5525,6 +5526,9 @@ xfs_btree_init_cur_caches(void)
 	if (error)
 		goto err;
 	error = xfs_refcountbt_init_cur_cache();
+	if (error)
+		goto err;
+	error = xfs_rtrmapbt_init_cur_cache();
 	if (error)
 		goto err;
 
@@ -5543,6 +5547,7 @@ xfs_btree_destroy_cur_caches(void)
 	xfs_bmbt_destroy_cur_cache();
 	xfs_rmapbt_destroy_cur_cache();
 	xfs_refcountbt_destroy_cur_cache();
+	xfs_rtrmapbt_destroy_cur_cache();
 }
 
 /* Move the btree cursor before the first record. */
diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index 4d47a3e723aa13..469fc7afa591b4 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -1725,6 +1725,16 @@ typedef __be32 xfs_rmap_ptr_t;
 	 XFS_FIBT_BLOCK(mp) + 1 : \
 	 XFS_IBT_BLOCK(mp) + 1)
 
+/*
+ * Realtime Reverse mapping btree format definitions
+ *
+ * This is a btree for reverse mapping records for realtime volumes
+ */
+#define	XFS_RTRMAP_CRC_MAGIC	0x4d415052	/* 'MAPR' */
+
+/* inode-based btree pointer type */
+typedef __be64 xfs_rtrmap_ptr_t;
+
 /*
  * Reference Count Btree format definitions
  *
diff --git a/fs/xfs/libxfs/xfs_ondisk.h b/fs/xfs/libxfs/xfs_ondisk.h
index ad0dedf00f1898..2c50877a1a2f0b 100644
--- a/fs/xfs/libxfs/xfs_ondisk.h
+++ b/fs/xfs/libxfs/xfs_ondisk.h
@@ -83,6 +83,7 @@ xfs_check_ondisk_structs(void)
 	XFS_CHECK_STRUCT_SIZE(union xfs_rtword_raw,		4);
 	XFS_CHECK_STRUCT_SIZE(union xfs_suminfo_raw,		4);
 	XFS_CHECK_STRUCT_SIZE(struct xfs_rtbuf_blkinfo,		48);
+	XFS_CHECK_STRUCT_SIZE(xfs_rtrmap_ptr_t,			8);
 
 	/*
 	 * m68k has problems with struct xfs_attr_leaf_name_remote, but we pad
diff --git a/fs/xfs/libxfs/xfs_rtrmap_btree.c b/fs/xfs/libxfs/xfs_rtrmap_btree.c
new file mode 100644
index 00000000000000..d3e4c52dcaa9d0
--- /dev/null
+++ b/fs/xfs/libxfs/xfs_rtrmap_btree.c
@@ -0,0 +1,271 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2018-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#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_btree_staging.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_rtgroup.h"
+
+static struct kmem_cache	*xfs_rtrmapbt_cur_cache;
+
+/*
+ * Realtime Reverse Map btree.
+ *
+ * This is a btree 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
+ * is rooted in an inode and does not live in free space.
+ */
+
+static struct xfs_btree_cur *
+xfs_rtrmapbt_dup_cursor(
+	struct xfs_btree_cur	*cur)
+{
+	return xfs_rtrmapbt_init_cursor(cur->bc_tp, to_rtg(cur->bc_group));
+}
+
+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 (!xfs_verify_magic(bp, block->bb_magic))
+		return __this_address;
+
+	if (!xfs_has_rmapbt(mp))
+		return __this_address;
+	fa = xfs_btree_fsblock_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_fsblock_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_fsblock_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_fsblock_calc_crc(bp);
+
+}
+
+const struct xfs_buf_ops xfs_rtrmapbt_buf_ops = {
+	.name			= "xfs_rtrmapbt",
+	.magic			= { 0, cpu_to_be32(XFS_RTRMAP_CRC_MAGIC) },
+	.verify_read		= xfs_rtrmapbt_read_verify,
+	.verify_write		= xfs_rtrmapbt_write_verify,
+	.verify_struct		= xfs_rtrmapbt_verify,
+};
+
+const struct xfs_btree_ops xfs_rtrmapbt_ops = {
+	.name			= "rtrmap",
+	.type			= XFS_BTREE_TYPE_INODE,
+	.geom_flags		= XFS_BTGEO_OVERLAPPING |
+				  XFS_BTGEO_IROOT_RECORDS,
+
+	.rec_len		= sizeof(struct xfs_rmap_rec),
+	/* Overlapping btree; 2 keys per pointer. */
+	.key_len		= 2 * sizeof(struct xfs_rmap_key),
+	.ptr_len		= XFS_BTREE_LONG_PTR_LEN,
+
+	.lru_refs		= XFS_RMAP_BTREE_REF,
+	.statoff		= XFS_STATS_CALC_INDEX(xs_rtrmap_2),
+
+	.dup_cursor		= xfs_rtrmapbt_dup_cursor,
+	.buf_ops		= &xfs_rtrmapbt_buf_ops,
+};
+
+/* Allocate a new rt rmap btree cursor. */
+struct xfs_btree_cur *
+xfs_rtrmapbt_init_cursor(
+	struct xfs_trans	*tp,
+	struct xfs_rtgroup	*rtg)
+{
+	struct xfs_inode	*ip = NULL;
+	struct xfs_mount	*mp = rtg_mount(rtg);
+	struct xfs_btree_cur	*cur;
+
+	return NULL; /* XXX */
+
+	xfs_assert_ilocked(ip, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL);
+
+	cur = xfs_btree_alloc_cursor(mp, tp, &xfs_rtrmapbt_ops,
+			mp->m_rtrmap_maxlevels, xfs_rtrmapbt_cur_cache);
+
+	cur->bc_ino.ip = ip;
+	cur->bc_group = xfs_group_hold(rtg_group(rtg));
+	cur->bc_ino.whichfork = XFS_DATA_FORK;
+	cur->bc_nlevels = be16_to_cpu(ip->i_df.if_broot->bb_level) + 1;
+	cur->bc_ino.forksize = xfs_inode_fork_size(ip, XFS_DATA_FORK);
+
+	return cur;
+}
+
+/*
+ * Install a new rt reverse mapping btree root.  Caller is responsible for
+ * invalidating and freeing the old btree blocks.
+ */
+void
+xfs_rtrmapbt_commit_staged_btree(
+	struct xfs_btree_cur	*cur,
+	struct xfs_trans	*tp)
+{
+	struct xbtree_ifakeroot	*ifake = cur->bc_ino.ifake;
+	struct xfs_ifork	*ifp;
+	int			flags = XFS_ILOG_CORE | XFS_ILOG_DBROOT;
+
+	ASSERT(cur->bc_flags & XFS_BTREE_STAGING);
+
+	/*
+	 * Free any resources hanging off the real fork, then shallow-copy the
+	 * staging fork's contents into the real fork to transfer everything
+	 * we just built.
+	 */
+	ifp = xfs_ifork_ptr(cur->bc_ino.ip, XFS_DATA_FORK);
+	xfs_idestroy_fork(ifp);
+	memcpy(ifp, ifake->if_fork, sizeof(struct xfs_ifork));
+
+	cur->bc_ino.ip->i_projid = cur->bc_group->xg_gno;
+	xfs_trans_log_inode(tp, cur->bc_ino.ip, flags);
+	xfs_btree_commit_ifakeroot(cur, tp, XFS_DATA_FORK);
+}
+
+/* Calculate number of records in a rt reverse mapping btree block. */
+static inline unsigned int
+xfs_rtrmapbt_block_maxrecs(
+	unsigned int		blocklen,
+	bool			leaf)
+{
+	if (leaf)
+		return blocklen / sizeof(struct xfs_rmap_rec);
+	return blocklen /
+		(2 * sizeof(struct xfs_rmap_key) + sizeof(xfs_rtrmap_ptr_t));
+}
+
+/*
+ * Calculate number of records in an rt reverse mapping btree block.
+ */
+unsigned int
+xfs_rtrmapbt_maxrecs(
+	struct xfs_mount	*mp,
+	unsigned int		blocklen,
+	bool			leaf)
+{
+	blocklen -= XFS_RTRMAP_BLOCK_LEN;
+	return xfs_rtrmapbt_block_maxrecs(blocklen, leaf);
+}
+
+/* Compute the max possible height for realtime reverse mapping btrees. */
+unsigned int
+xfs_rtrmapbt_maxlevels_ondisk(void)
+{
+	unsigned int		minrecs[2];
+	unsigned int		blocklen;
+
+	blocklen = XFS_MIN_CRC_BLOCKSIZE - XFS_BTREE_LBLOCK_CRC_LEN;
+
+	minrecs[0] = xfs_rtrmapbt_block_maxrecs(blocklen, true) / 2;
+	minrecs[1] = xfs_rtrmapbt_block_maxrecs(blocklen, false) / 2;
+
+	/* We need at most one record for every block in an rt group. */
+	return xfs_btree_compute_maxlevels(minrecs, XFS_MAX_RGBLOCKS);
+}
+
+int __init
+xfs_rtrmapbt_init_cur_cache(void)
+{
+	xfs_rtrmapbt_cur_cache = kmem_cache_create("xfs_rtrmapbt_cur",
+			xfs_btree_cur_sizeof(xfs_rtrmapbt_maxlevels_ondisk()),
+			0, 0, NULL);
+
+	if (!xfs_rtrmapbt_cur_cache)
+		return -ENOMEM;
+	return 0;
+}
+
+void
+xfs_rtrmapbt_destroy_cur_cache(void)
+{
+	kmem_cache_destroy(xfs_rtrmapbt_cur_cache);
+	xfs_rtrmapbt_cur_cache = NULL;
+}
+
+/* Compute the maximum height of an rt reverse mapping btree. */
+void
+xfs_rtrmapbt_compute_maxlevels(
+	struct xfs_mount	*mp)
+{
+	unsigned int		d_maxlevels, r_maxlevels;
+
+	if (!xfs_has_rtrmapbt(mp)) {
+		mp->m_rtrmap_maxlevels = 0;
+		return;
+	}
+
+	/*
+	 * The realtime rmapbt lives on the data device, which means that its
+	 * maximum height is constrained by the size of the data device and
+	 * the height required to store one rmap record for each block in an
+	 * rt group.
+	 */
+	d_maxlevels = xfs_btree_space_to_height(mp->m_rtrmap_mnr,
+				mp->m_sb.sb_dblocks);
+	r_maxlevels = xfs_btree_compute_maxlevels(mp->m_rtrmap_mnr,
+				mp->m_groups[XG_TYPE_RTG].blocks);
+
+	/* Add one level to handle the inode root level. */
+	mp->m_rtrmap_maxlevels = min(d_maxlevels, r_maxlevels) + 1;
+}
diff --git a/fs/xfs/libxfs/xfs_rtrmap_btree.h b/fs/xfs/libxfs/xfs_rtrmap_btree.h
new file mode 100644
index 00000000000000..7d1a3a49a2d69b
--- /dev/null
+++ b/fs/xfs/libxfs/xfs_rtrmap_btree.h
@@ -0,0 +1,82 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright (c) 2018-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_RTRMAP_BTREE_H__
+#define __XFS_RTRMAP_BTREE_H__
+
+struct xfs_buf;
+struct xfs_btree_cur;
+struct xfs_mount;
+struct xbtree_ifakeroot;
+struct xfs_rtgroup;
+
+/* rmaps only exist on crc enabled filesystems */
+#define XFS_RTRMAP_BLOCK_LEN	XFS_BTREE_LBLOCK_CRC_LEN
+
+struct xfs_btree_cur *xfs_rtrmapbt_init_cursor(struct xfs_trans *tp,
+		struct xfs_rtgroup *rtg);
+struct xfs_btree_cur *xfs_rtrmapbt_stage_cursor(struct xfs_mount *mp,
+		struct xfs_rtgroup *rtg, struct xfs_inode *ip,
+		struct xbtree_ifakeroot *ifake);
+void xfs_rtrmapbt_commit_staged_btree(struct xfs_btree_cur *cur,
+		struct xfs_trans *tp);
+unsigned int xfs_rtrmapbt_maxrecs(struct xfs_mount *mp, unsigned int blocklen,
+		bool leaf);
+void xfs_rtrmapbt_compute_maxlevels(struct xfs_mount *mp);
+
+/*
+ * Addresses of records, keys, and pointers within an incore rtrmapbt block.
+ *
+ * (note that some of these may appear unused, but they are used in userspace)
+ */
+static inline struct xfs_rmap_rec *
+xfs_rtrmap_rec_addr(
+	struct xfs_btree_block	*block,
+	unsigned int		index)
+{
+	return (struct xfs_rmap_rec *)
+		((char *)block + XFS_RTRMAP_BLOCK_LEN +
+		 (index - 1) * sizeof(struct xfs_rmap_rec));
+}
+
+static inline struct xfs_rmap_key *
+xfs_rtrmap_key_addr(
+	struct xfs_btree_block	*block,
+	unsigned int		index)
+{
+	return (struct xfs_rmap_key *)
+		((char *)block + XFS_RTRMAP_BLOCK_LEN +
+		 (index - 1) * 2 * sizeof(struct xfs_rmap_key));
+}
+
+static inline struct xfs_rmap_key *
+xfs_rtrmap_high_key_addr(
+	struct xfs_btree_block	*block,
+	unsigned int		index)
+{
+	return (struct xfs_rmap_key *)
+		((char *)block + XFS_RTRMAP_BLOCK_LEN +
+		 sizeof(struct xfs_rmap_key) +
+		 (index - 1) * 2 * sizeof(struct xfs_rmap_key));
+}
+
+static inline xfs_rtrmap_ptr_t *
+xfs_rtrmap_ptr_addr(
+	struct xfs_btree_block	*block,
+	unsigned int		index,
+	unsigned int		maxrecs)
+{
+	return (xfs_rtrmap_ptr_t *)
+		((char *)block + XFS_RTRMAP_BLOCK_LEN +
+		 maxrecs * 2 * sizeof(struct xfs_rmap_key) +
+		 (index - 1) * sizeof(xfs_rtrmap_ptr_t));
+}
+
+unsigned int xfs_rtrmapbt_maxlevels_ondisk(void);
+
+int __init xfs_rtrmapbt_init_cur_cache(void);
+void xfs_rtrmapbt_destroy_cur_cache(void);
+
+#endif /* __XFS_RTRMAP_BTREE_H__ */
diff --git a/fs/xfs/libxfs/xfs_sb.c b/fs/xfs/libxfs/xfs_sb.c
index 3b5623611eba02..83fb14b4074c8d 100644
--- a/fs/xfs/libxfs/xfs_sb.c
+++ b/fs/xfs/libxfs/xfs_sb.c
@@ -28,6 +28,7 @@
 #include "xfs_rtbitmap.h"
 #include "xfs_exchrange.h"
 #include "xfs_rtgroup.h"
+#include "xfs_rtrmap_btree.h"
 
 /*
  * Physical superblock buffer manipulations. Shared with libxfs in userspace.
@@ -1215,6 +1216,11 @@ 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(mp, sbp->sb_blocksize, true);
+	mp->m_rtrmap_mxr[1] = xfs_rtrmapbt_maxrecs(mp, 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(mp, sbp->sb_blocksize, true);
 	mp->m_refc_mxr[1] = xfs_refcountbt_maxrecs(mp, sbp->sb_blocksize, false);
 	mp->m_refc_mnr[0] = mp->m_refc_mxr[0] / 2;
diff --git a/fs/xfs/libxfs/xfs_shared.h b/fs/xfs/libxfs/xfs_shared.h
index e7efdb9ceaf382..da23dac22c3f08 100644
--- a/fs/xfs/libxfs/xfs_shared.h
+++ b/fs/xfs/libxfs/xfs_shared.h
@@ -42,6 +42,7 @@ extern const struct xfs_buf_ops xfs_rtbitmap_buf_ops;
 extern const struct xfs_buf_ops xfs_rtsummary_buf_ops;
 extern const struct xfs_buf_ops xfs_rtbuf_ops;
 extern const struct xfs_buf_ops xfs_rtsb_buf_ops;
+extern const struct xfs_buf_ops xfs_rtrmapbt_buf_ops;
 extern const struct xfs_buf_ops xfs_sb_buf_ops;
 extern const struct xfs_buf_ops xfs_sb_quiet_buf_ops;
 extern const struct xfs_buf_ops xfs_symlink_buf_ops;
@@ -55,6 +56,7 @@ extern const struct xfs_btree_ops xfs_bmbt_ops;
 extern const struct xfs_btree_ops xfs_refcountbt_ops;
 extern const struct xfs_btree_ops xfs_rmapbt_ops;
 extern const struct xfs_btree_ops xfs_rmapbt_mem_ops;
+extern const struct xfs_btree_ops xfs_rtrmapbt_ops;
 
 static inline bool xfs_btree_is_bno(const struct xfs_btree_ops *ops)
 {
@@ -100,6 +102,11 @@ static inline bool xfs_btree_is_mem_rmap(const struct xfs_btree_ops *ops)
 # define xfs_btree_is_mem_rmap(...)	(false)
 #endif
 
+static inline bool xfs_btree_is_rtrmap(const struct xfs_btree_ops *ops)
+{
+	return ops == &xfs_rtrmapbt_ops;
+}
+
 /* log size calculation functions */
 int	xfs_log_calc_unit_res(struct xfs_mount *mp, int unit_bytes);
 int	xfs_log_calc_minimum_size(struct xfs_mount *);
diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c
index 97137126b16f5a..7b7d21b50d5409 100644
--- a/fs/xfs/xfs_mount.c
+++ b/fs/xfs/xfs_mount.c
@@ -37,6 +37,7 @@
 #include "xfs_rtbitmap.h"
 #include "xfs_metafile.h"
 #include "xfs_rtgroup.h"
+#include "xfs_rtrmap_btree.h"
 #include "scrub/stats.h"
 
 static DEFINE_MUTEX(xfs_uuid_table_mutex);
@@ -655,8 +656,7 @@ static inline void
 xfs_rtbtree_compute_maxlevels(
 	struct xfs_mount	*mp)
 {
-	/* This will be filled in later. */
-	mp->m_rtbtree_maxlevels = 0;
+	mp->m_rtbtree_maxlevels = mp->m_rtrmap_maxlevels;
 }
 
 /*
@@ -727,6 +727,7 @@ xfs_mountfs(
 	xfs_bmap_compute_maxlevels(mp, XFS_ATTR_FORK);
 	xfs_mount_setup_inode_geom(mp);
 	xfs_rmapbt_compute_maxlevels(mp);
+	xfs_rtrmapbt_compute_maxlevels(mp);
 	xfs_refcountbt_compute_maxlevels(mp);
 
 	xfs_agbtree_compute_maxlevels(mp);
diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h
index ddb9d19a3a3d53..1bc95fb170db61 100644
--- a/fs/xfs/xfs_mount.h
+++ b/fs/xfs/xfs_mount.h
@@ -158,11 +158,14 @@ typedef struct xfs_mount {
 	uint			m_bmap_dmnr[2];	/* min bmap 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_alloc_maxlevels; /* max alloc btree levels */
 	uint			m_bm_maxlevels[2]; /* max bmap 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 */
 	unsigned int		m_agbtree_maxlevels; /* max level of all AG btrees */
 	unsigned int		m_rtbtree_maxlevels; /* max level of all rt btrees */
@@ -399,6 +402,12 @@ static inline bool xfs_has_rtsb(struct xfs_mount *mp)
 	return xfs_has_rtgroups(mp) && xfs_has_realtime(mp);
 }
 
+static inline bool xfs_has_rtrmapbt(struct xfs_mount *mp)
+{
+	return xfs_has_rtgroups(mp) && xfs_has_realtime(mp) &&
+	       xfs_has_rmapbt(mp);
+}
+
 /*
  * Some features are always on for v5 file systems, allow the compiler to
  * eliminiate dead code when building without v4 support.
diff --git a/fs/xfs/xfs_stats.c b/fs/xfs/xfs_stats.c
index ffb52725c2a8e8..f94fb70b524ffb 100644
--- a/fs/xfs/xfs_stats.c
+++ b/fs/xfs/xfs_stats.c
@@ -52,7 +52,8 @@ int xfs_stats_format(struct xfsstats __percpu *stats, char *buf)
 		{ "rmapbt",		xfsstats_offset(xs_refcbt_2)	},
 		{ "refcntbt",		xfsstats_offset(xs_rmap_mem_2)	},
 		{ "rmapbt_mem",		xfsstats_offset(xs_rcbag_2)	},
-		{ "rcbagbt",		xfsstats_offset(xs_qm_dqreclaims)},
+		{ "rcbagbt",		xfsstats_offset(xs_rtrmap_2)	},
+		{ "rtrmapbt",		xfsstats_offset(xs_qm_dqreclaims)},
 		/* we print both series of quota information together */
 		{ "qm",			xfsstats_offset(xs_xstrat_bytes)},
 	};
diff --git a/fs/xfs/xfs_stats.h b/fs/xfs/xfs_stats.h
index a61fb56ed2e66c..05dc69c6d94906 100644
--- a/fs/xfs/xfs_stats.h
+++ b/fs/xfs/xfs_stats.h
@@ -127,6 +127,7 @@ struct __xfsstats {
 	uint32_t		xs_refcbt_2[__XBTS_MAX];
 	uint32_t		xs_rmap_mem_2[__XBTS_MAX];
 	uint32_t		xs_rcbag_2[__XBTS_MAX];
+	uint32_t		xs_rtrmap_2[__XBTS_MAX];
 	uint32_t		xs_qm_dqreclaims;
 	uint32_t		xs_qm_dqreclaim_misses;
 	uint32_t		xs_qm_dquot_dups;

[PATCH 05/37] xfs: realtime rmap btree transaction reservations

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

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" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_exchmaps.c    |    4 +++-
 fs/xfs/libxfs/xfs_trans_resv.c  |   12 ++++++++++--
 fs/xfs/libxfs/xfs_trans_space.h |   13 +++++++++++++
 3 files changed, 26 insertions(+), 3 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_exchmaps.c b/fs/xfs/libxfs/xfs_exchmaps.c
index 2021396651de27..3f1d6a98c11819 100644
--- a/fs/xfs/libxfs/xfs_exchmaps.c
+++ b/fs/xfs/libxfs/xfs_exchmaps.c
@@ -662,7 +662,9 @@ xfs_exchmaps_rmapbt_blocks(
 	if (!xfs_has_rmapbt(mp))
 		return 0;
 	if (XFS_IS_REALTIME_INODE(req->ip1))
-		return 0;
+		return howmany_64(req->nr_exchanges,
+					XFS_MAX_CONTIG_RTRMAPS_PER_BLOCK(mp)) *
+			XFS_RTRMAPADD_SPACE_RES(mp);
 
 	return howmany_64(req->nr_exchanges,
 					XFS_MAX_CONTIG_RMAPS_PER_BLOCK(mp)) *
diff --git a/fs/xfs/libxfs/xfs_trans_resv.c b/fs/xfs/libxfs/xfs_trans_resv.c
index bab402340b5da8..f3392eb2d7f41f 100644
--- a/fs/xfs/libxfs/xfs_trans_resv.c
+++ b/fs/xfs/libxfs/xfs_trans_resv.c
@@ -213,7 +213,9 @@ xfs_calc_inode_chunk_res(
  * Per-extent log reservation for the btree changes involved in freeing or
  * allocating a realtime extent.  We have to be able to log as many rtbitmap
  * blocks as needed to mark inuse XFS_BMBT_MAX_EXTLEN blocks' worth of realtime
- * extents, as well as the realtime summary block.
+ * extents, as well as the realtime summary block (t1).  Realtime rmap btree
+ * operations happen in a second transaction, so factor in a couple of rtrmapbt
+ * splits (t2).
  */
 static unsigned int
 xfs_rtalloc_block_count(
@@ -222,10 +224,16 @@ xfs_rtalloc_block_count(
 {
 	unsigned int		rtbmp_blocks;
 	xfs_rtxlen_t		rtxlen;
+	unsigned int		t1, t2 = 0;
 
 	rtxlen = xfs_extlen_to_rtxlen(mp, XFS_MAX_BMBT_EXTLEN);
 	rtbmp_blocks = xfs_rtbitmap_blockcount_len(mp, rtxlen);
-	return (rtbmp_blocks + 1) * num_ops;
+	t1 = (rtbmp_blocks + 1) * num_ops;
+
+	if (xfs_has_rmapbt(mp))
+		t2 = num_ops * (2 * mp->m_rtrmap_maxlevels - 1);
+
+	return max(t1, t2);
 }
 
 /*
diff --git a/fs/xfs/libxfs/xfs_trans_space.h b/fs/xfs/libxfs/xfs_trans_space.h
index 1155ff2d37e29f..d89b570aafcc64 100644
--- a/fs/xfs/libxfs/xfs_trans_space.h
+++ b/fs/xfs/libxfs/xfs_trans_space.h
@@ -14,6 +14,19 @@
 #define XFS_MAX_CONTIG_BMAPS_PER_BLOCK(mp)    \
 		(((mp)->m_bmap_dmxr[0]) - ((mp)->m_bmap_dmnr[0]))
 
+/* Worst case number of realtime rmaps that can be held in a block. */
+#define XFS_MAX_CONTIG_RTRMAPS_PER_BLOCK(mp)    \
+		(((mp)->m_rtrmap_mxr[0]) - ((mp)->m_rtrmap_mnr[0]))
+
+/* Adding one realtime rmap could split every level to the top of the tree. */
+#define XFS_RTRMAPADD_SPACE_RES(mp) ((mp)->m_rtrmap_maxlevels)
+
+/* Blocks we might need to add "b" realtime rmaps to a tree. */
+#define XFS_NRTRMAPADD_SPACE_RES(mp, b) \
+	((((b) + XFS_MAX_CONTIG_RTRMAPS_PER_BLOCK(mp) - 1) / \
+	  XFS_MAX_CONTIG_RTRMAPS_PER_BLOCK(mp)) * \
+	  XFS_RTRMAPADD_SPACE_RES(mp))
+
 /* Worst case number of rmaps that can be held in a block. */
 #define XFS_MAX_CONTIG_RMAPS_PER_BLOCK(mp)    \
 		(((mp)->m_rmap_mxr[0]) - ((mp)->m_rmap_mnr[0]))

[PATCH 06/37] xfs: add realtime rmap btree operations

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

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" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_btree.c        |   68 ++++++++++
 fs/xfs/libxfs/xfs_btree.h        |    6 +
 fs/xfs/libxfs/xfs_rtrmap_btree.c |  271 ++++++++++++++++++++++++++++++++++++++
 3 files changed, 345 insertions(+)


diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c
index 0e271919374780..36ab06f8a3bc99 100644
--- a/fs/xfs/libxfs/xfs_btree.c
+++ b/fs/xfs/libxfs/xfs_btree.c
@@ -31,6 +31,10 @@
 #include "xfs_buf_mem.h"
 #include "xfs_btree_mem.h"
 #include "xfs_rtrmap_btree.h"
+#include "xfs_bmap.h"
+#include "xfs_rmap.h"
+#include "xfs_quota.h"
+#include "xfs_metafile.h"
 
 /*
  * Btree magic numbers.
@@ -5576,3 +5580,67 @@ xfs_btree_goto_left_edge(
 
 	return 0;
 }
+
+/* Allocate a block for an inode-rooted metadata btree. */
+int
+xfs_btree_alloc_metafile_block(
+	struct xfs_btree_cur		*cur,
+	const union xfs_btree_ptr	*start,
+	union xfs_btree_ptr		*new,
+	int				*stat)
+{
+	struct xfs_alloc_arg		args = {
+		.mp			= cur->bc_mp,
+		.tp			= cur->bc_tp,
+		.resv			= XFS_AG_RESV_METAFILE,
+		.minlen			= 1,
+		.maxlen			= 1,
+		.prod			= 1,
+	};
+	struct xfs_inode		*ip = cur->bc_ino.ip;
+	int				error;
+
+	ASSERT(xfs_is_metadir_inode(ip));
+
+	xfs_rmap_ino_bmbt_owner(&args.oinfo, ip->i_ino, cur->bc_ino.whichfork);
+	error = xfs_alloc_vextent_start_ag(&args,
+			XFS_INO_TO_FSB(cur->bc_mp, ip->i_ino));
+	if (error)
+		return error;
+	if (args.fsbno == NULLFSBLOCK) {
+		*stat = 0;
+		return 0;
+	}
+	ASSERT(args.len == 1);
+
+	xfs_metafile_resv_alloc_space(ip, &args);
+
+	new->l = cpu_to_be64(args.fsbno);
+	*stat = 1;
+	return 0;
+}
+
+/* Free a block from an inode-rooted metadata btree. */
+int
+xfs_btree_free_metafile_block(
+	struct xfs_btree_cur	*cur,
+	struct xfs_buf		*bp)
+{
+	struct xfs_owner_info	oinfo;
+	struct xfs_mount	*mp = cur->bc_mp;
+	struct xfs_inode	*ip = cur->bc_ino.ip;
+	struct xfs_trans	*tp = cur->bc_tp;
+	xfs_fsblock_t		fsbno = XFS_DADDR_TO_FSB(mp, xfs_buf_daddr(bp));
+	int			error;
+
+	ASSERT(xfs_is_metadir_inode(ip));
+
+	xfs_rmap_ino_bmbt_owner(&oinfo, ip->i_ino, cur->bc_ino.whichfork);
+	error = xfs_free_extent_later(tp, fsbno, 1, &oinfo, XFS_AG_RESV_METAFILE,
+			0);
+	if (error)
+		return error;
+
+	xfs_metafile_resv_free_space(ip, tp, 1);
+	return 0;
+}
diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h
index 3b8c2ccad90847..ee82dc777d6d5b 100644
--- a/fs/xfs/libxfs/xfs_btree.h
+++ b/fs/xfs/libxfs/xfs_btree.h
@@ -703,4 +703,10 @@ xfs_btree_at_iroot(
 	       level == cur->bc_nlevels - 1;
 }
 
+int xfs_btree_alloc_metafile_block(struct xfs_btree_cur *cur,
+		const union xfs_btree_ptr *start, union xfs_btree_ptr *newp,
+		int *stat);
+int xfs_btree_free_metafile_block(struct xfs_btree_cur *cur,
+		struct xfs_buf *bp);
+
 #endif	/* __XFS_BTREE_H__ */
diff --git a/fs/xfs/libxfs/xfs_rtrmap_btree.c b/fs/xfs/libxfs/xfs_rtrmap_btree.c
index d3e4c52dcaa9d0..99d828bb5fe7c3 100644
--- a/fs/xfs/libxfs/xfs_rtrmap_btree.c
+++ b/fs/xfs/libxfs/xfs_rtrmap_btree.c
@@ -18,12 +18,14 @@
 #include "xfs_alloc.h"
 #include "xfs_btree.h"
 #include "xfs_btree_staging.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_rtgroup.h"
+#include "xfs_bmap.h"
 
 static struct kmem_cache	*xfs_rtrmapbt_cur_cache;
 
@@ -44,6 +46,182 @@ xfs_rtrmapbt_dup_cursor(
 	return xfs_rtrmapbt_init_cursor(cur->bc_tp, to_rtg(cur->bc_group));
 }
 
+STATIC int
+xfs_rtrmapbt_get_minrecs(
+	struct xfs_btree_cur	*cur,
+	int			level)
+{
+	if (level == cur->bc_nlevels - 1) {
+		struct xfs_ifork	*ifp = xfs_btree_ifork_ptr(cur);
+
+		return xfs_rtrmapbt_maxrecs(cur->bc_mp, 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)
+{
+	if (level == cur->bc_nlevels - 1) {
+		struct xfs_ifork	*ifp = xfs_btree_ifork_ptr(cur);
+
+		return xfs_rtrmapbt_maxrecs(cur->bc_mp, ifp->if_broot_bytes,
+				level == 0);
+	}
+
+	return cur->bc_mp->m_rtrmap_mxr[level != 0];
+}
+
+/*
+ * Convert the ondisk record's offset field into the ondisk key's offset field.
+ * Fork and bmbt are significant parts of the rmap record key, but written
+ * status is merely a record attribute.
+ */
+static inline __be64 ondisk_rec_offset_to_key(const union xfs_btree_rec *rec)
+{
+	return rec->rmap.rm_offset & ~cpu_to_be64(XFS_RMAP_OFF_UNWRITTEN);
+}
+
+STATIC void
+xfs_rtrmapbt_init_key_from_rec(
+	union xfs_btree_key		*key,
+	const union xfs_btree_rec	*rec)
+{
+	key->rmap.rm_startblock = rec->rmap.rm_startblock;
+	key->rmap.rm_owner = rec->rmap.rm_owner;
+	key->rmap.rm_offset = ondisk_rec_offset_to_key(rec);
+}
+
+STATIC void
+xfs_rtrmapbt_init_high_key_from_rec(
+	union xfs_btree_key		*key,
+	const union xfs_btree_rec	*rec)
+{
+	uint64_t			off;
+	int				adj;
+
+	adj = be32_to_cpu(rec->rmap.rm_blockcount) - 1;
+
+	key->rmap.rm_startblock = rec->rmap.rm_startblock;
+	be32_add_cpu(&key->rmap.rm_startblock, adj);
+	key->rmap.rm_owner = rec->rmap.rm_owner;
+	key->rmap.rm_offset = ondisk_rec_offset_to_key(rec);
+	if (XFS_RMAP_NON_INODE_OWNER(be64_to_cpu(rec->rmap.rm_owner)) ||
+	    XFS_RMAP_IS_BMBT_BLOCK(be64_to_cpu(rec->rmap.rm_offset)))
+		return;
+	off = be64_to_cpu(key->rmap.rm_offset);
+	off = (XFS_RMAP_OFF(off) + adj) | (off & ~XFS_RMAP_OFF_MASK);
+	key->rmap.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->rmap.rm_startblock = cpu_to_be32(cur->bc_rec.r.rm_startblock);
+	rec->rmap.rm_blockcount = cpu_to_be32(cur->bc_rec.r.rm_blockcount);
+	rec->rmap.rm_owner = cpu_to_be64(cur->bc_rec.r.rm_owner);
+	rec->rmap.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;
+}
+
+/*
+ * Mask the appropriate parts of the ondisk key field for a key comparison.
+ * Fork and bmbt are significant parts of the rmap record key, but written
+ * status is merely a record attribute.
+ */
+static inline uint64_t offset_keymask(uint64_t offset)
+{
+	return offset & ~XFS_RMAP_OFF_UNWRITTEN;
+}
+
+STATIC int64_t
+xfs_rtrmapbt_key_diff(
+	struct xfs_btree_cur		*cur,
+	const union xfs_btree_key	*key)
+{
+	struct xfs_rmap_irec		*rec = &cur->bc_rec.r;
+	const struct xfs_rmap_key	*kp = &key->rmap;
+	__u64				x, y;
+	int64_t				d;
+
+	d = (int64_t)be32_to_cpu(kp->rm_startblock) - rec->rm_startblock;
+	if (d)
+		return d;
+
+	x = be64_to_cpu(kp->rm_owner);
+	y = rec->rm_owner;
+	if (x > y)
+		return 1;
+	else if (y > x)
+		return -1;
+
+	x = offset_keymask(be64_to_cpu(kp->rm_offset));
+	y = offset_keymask(xfs_rmap_irec_offset_pack(rec));
+	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,
+	const union xfs_btree_key	*k1,
+	const union xfs_btree_key	*k2,
+	const union xfs_btree_key	*mask)
+{
+	const struct xfs_rmap_key	*kp1 = &k1->rmap;
+	const struct xfs_rmap_key	*kp2 = &k2->rmap;
+	int64_t				d;
+	__u64				x, y;
+
+	/* Doesn't make sense to mask off the physical space part */
+	ASSERT(!mask || mask->rmap.rm_startblock);
+
+	d = (int64_t)be32_to_cpu(kp1->rm_startblock) -
+		     be32_to_cpu(kp2->rm_startblock);
+	if (d)
+		return d;
+
+	if (!mask || mask->rmap.rm_owner) {
+		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;
+	}
+
+	if (!mask || mask->rmap.rm_offset) {
+		/* Doesn't make sense to allow offset but not owner */
+		ASSERT(!mask || mask->rmap.rm_owner);
+
+		x = offset_keymask(be64_to_cpu(kp1->rm_offset));
+		y = offset_keymask(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)
@@ -110,6 +288,86 @@ 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,
+	const union xfs_btree_key	*k1,
+	const union xfs_btree_key	*k2)
+{
+	uint32_t			x;
+	uint32_t			y;
+	uint64_t			a;
+	uint64_t			b;
+
+	x = be32_to_cpu(k1->rmap.rm_startblock);
+	y = be32_to_cpu(k2->rmap.rm_startblock);
+	if (x < y)
+		return 1;
+	else if (x > y)
+		return 0;
+	a = be64_to_cpu(k1->rmap.rm_owner);
+	b = be64_to_cpu(k2->rmap.rm_owner);
+	if (a < b)
+		return 1;
+	else if (a > b)
+		return 0;
+	a = offset_keymask(be64_to_cpu(k1->rmap.rm_offset));
+	b = offset_keymask(be64_to_cpu(k2->rmap.rm_offset));
+	if (a <= b)
+		return 1;
+	return 0;
+}
+
+STATIC int
+xfs_rtrmapbt_recs_inorder(
+	struct xfs_btree_cur		*cur,
+	const union xfs_btree_rec	*r1,
+	const union xfs_btree_rec	*r2)
+{
+	uint32_t			x;
+	uint32_t			y;
+	uint64_t			a;
+	uint64_t			b;
+
+	x = be32_to_cpu(r1->rmap.rm_startblock);
+	y = be32_to_cpu(r2->rmap.rm_startblock);
+	if (x < y)
+		return 1;
+	else if (x > y)
+		return 0;
+	a = be64_to_cpu(r1->rmap.rm_owner);
+	b = be64_to_cpu(r2->rmap.rm_owner);
+	if (a < b)
+		return 1;
+	else if (a > b)
+		return 0;
+	a = offset_keymask(be64_to_cpu(r1->rmap.rm_offset));
+	b = offset_keymask(be64_to_cpu(r2->rmap.rm_offset));
+	if (a <= b)
+		return 1;
+	return 0;
+}
+
+STATIC enum xbtree_key_contig
+xfs_rtrmapbt_keys_contiguous(
+	struct xfs_btree_cur		*cur,
+	const union xfs_btree_key	*key1,
+	const union xfs_btree_key	*key2,
+	const union xfs_btree_key	*mask)
+{
+	ASSERT(!mask || mask->rmap.rm_startblock);
+
+	/*
+	 * We only support checking contiguity of the physical space component.
+	 * If any callers ever need more specificity than that, they'll have to
+	 * implement it here.
+	 */
+	ASSERT(!mask || (!mask->rmap.rm_owner && !mask->rmap.rm_offset));
+
+	return xbtree_key_contig(be32_to_cpu(key1->rmap.rm_startblock),
+				 be32_to_cpu(key2->rmap.rm_startblock));
+}
+
 const struct xfs_btree_ops xfs_rtrmapbt_ops = {
 	.name			= "rtrmap",
 	.type			= XFS_BTREE_TYPE_INODE,
@@ -125,7 +383,20 @@ const struct xfs_btree_ops xfs_rtrmapbt_ops = {
 	.statoff		= XFS_STATS_CALC_INDEX(xs_rtrmap_2),
 
 	.dup_cursor		= xfs_rtrmapbt_dup_cursor,
+	.alloc_block		= xfs_btree_alloc_metafile_block,
+	.free_block		= xfs_btree_free_metafile_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,
+	.keys_contiguous	= xfs_rtrmapbt_keys_contiguous,
 };
 
 /* Allocate a new rt rmap btree cursor. */

[PATCH 07/37] xfs: prepare rmap functions to deal with rtrmapbt

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

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" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_rmap.c    |   63 +++++++++++++++++++++++++++++++++++++++++++
 fs/xfs/libxfs/xfs_rmap.h    |    3 ++
 fs/xfs/libxfs/xfs_rtgroup.h |   26 ++++++++++++++++++
 3 files changed, 92 insertions(+)


diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c
index 57dbf99ce00453..da1b004837d3ad 100644
--- a/fs/xfs/libxfs/xfs_rmap.c
+++ b/fs/xfs/libxfs/xfs_rmap.c
@@ -25,6 +25,7 @@
 #include "xfs_ag.h"
 #include "xfs_health.h"
 #include "xfs_rmap_item.h"
+#include "xfs_rtgroup.h"
 
 struct kmem_cache	*xfs_rmap_intent_cache;
 
@@ -264,11 +265,69 @@ xfs_rmap_check_irec(
 	return NULL;
 }
 
+static xfs_failaddr_t
+xfs_rtrmap_check_meta_irec(
+	struct xfs_rtgroup		*rtg,
+	const struct xfs_rmap_irec	*irec)
+{
+	struct xfs_mount		*mp = rtg_mount(rtg);
+
+	if (irec->rm_offset != 0)
+		return __this_address;
+	if (irec->rm_flags & XFS_RMAP_UNWRITTEN)
+		return __this_address;
+
+	switch (irec->rm_owner) {
+	case XFS_RMAP_OWN_FS:
+		if (irec->rm_startblock != 0)
+			return __this_address;
+		if (irec->rm_blockcount != mp->m_sb.sb_rextsize)
+			return __this_address;
+		return NULL;
+	default:
+		return __this_address;
+	}
+
+	return NULL;
+}
+
+static xfs_failaddr_t
+xfs_rtrmap_check_inode_irec(
+	struct xfs_rtgroup		*rtg,
+	const struct xfs_rmap_irec	*irec)
+{
+	struct xfs_mount		*mp = rtg_mount(rtg);
+
+	if (!xfs_verify_ino(mp, irec->rm_owner))
+		return __this_address;
+	if (!xfs_verify_rgbext(rtg, irec->rm_startblock, irec->rm_blockcount))
+		return __this_address;
+	if (!xfs_verify_fileext(mp, irec->rm_offset, irec->rm_blockcount))
+		return __this_address;
+	return NULL;
+}
+
+xfs_failaddr_t
+xfs_rtrmap_check_irec(
+	struct xfs_rtgroup		*rtg,
+	const struct xfs_rmap_irec	*irec)
+{
+	if (irec->rm_blockcount == 0)
+		return __this_address;
+	if (irec->rm_flags & (XFS_RMAP_BMBT_BLOCK | XFS_RMAP_ATTR_FORK))
+		return __this_address;
+	if (XFS_RMAP_NON_INODE_OWNER(irec->rm_owner))
+		return xfs_rtrmap_check_meta_irec(rtg, irec);
+	return xfs_rtrmap_check_inode_irec(rtg, irec);
+}
+
 static inline xfs_failaddr_t
 xfs_rmap_check_btrec(
 	struct xfs_btree_cur		*cur,
 	const struct xfs_rmap_irec	*irec)
 {
+	if (xfs_btree_is_rtrmap(cur->bc_ops))
+		return xfs_rtrmap_check_irec(to_rtg(cur->bc_group), irec);
 	return xfs_rmap_check_irec(to_perag(cur->bc_group), irec);
 }
 
@@ -283,6 +342,10 @@ xfs_rmap_complain_bad_rec(
 	if (xfs_btree_is_mem_rmap(cur->bc_ops))
 		xfs_warn(mp,
  "In-Memory Reverse Mapping BTree record corruption detected at %pS!", fa);
+	else if (xfs_btree_is_rtrmap(cur->bc_ops))
+		xfs_warn(mp,
+ "RT Reverse Mapping BTree record corruption in rtgroup %u detected at %pS!",
+				cur->bc_group->xg_gno, fa);
 	else
 		xfs_warn(mp,
  "Reverse Mapping BTree record corruption in AG %d detected at %pS!",
diff --git a/fs/xfs/libxfs/xfs_rmap.h b/fs/xfs/libxfs/xfs_rmap.h
index 8e2657af038e9e..1b19f54b65047f 100644
--- a/fs/xfs/libxfs/xfs_rmap.h
+++ b/fs/xfs/libxfs/xfs_rmap.h
@@ -7,6 +7,7 @@
 #define __XFS_RMAP_H__
 
 struct xfs_perag;
+struct xfs_rtgroup;
 
 static inline void
 xfs_rmap_ino_bmbt_owner(
@@ -206,6 +207,8 @@ xfs_failaddr_t xfs_rmap_btrec_to_irec(const union xfs_btree_rec *rec,
 		struct xfs_rmap_irec *irec);
 xfs_failaddr_t xfs_rmap_check_irec(struct xfs_perag *pag,
 		const struct xfs_rmap_irec *irec);
+xfs_failaddr_t xfs_rtrmap_check_irec(struct xfs_rtgroup *rtg,
+		const struct xfs_rmap_irec *irec);
 
 int xfs_rmap_has_records(struct xfs_btree_cur *cur, xfs_agblock_t bno,
 		xfs_extlen_t len, enum xbtree_recpacking *outcome);
diff --git a/fs/xfs/libxfs/xfs_rtgroup.h b/fs/xfs/libxfs/xfs_rtgroup.h
index 19f8d302b9aa3f..dc3ce660a01307 100644
--- a/fs/xfs/libxfs/xfs_rtgroup.h
+++ b/fs/xfs/libxfs/xfs_rtgroup.h
@@ -132,6 +132,32 @@ xfs_rtgroup_next(
 	return xfs_rtgroup_next_range(mp, rtg, 0, mp->m_sb.sb_rgcount - 1);
 }
 
+static inline bool
+xfs_verify_rgbno(
+	struct xfs_rtgroup	*rtg,
+	xfs_rgblock_t		rgbno)
+{
+	ASSERT(xfs_has_rtgroups(rtg_mount(rtg)));
+
+	return xfs_verify_gbno(rtg_group(rtg), rgbno);
+}
+
+/*
+ * Check that [@rgbno,@len] is a valid extent range in @rtg.
+ *
+ * Must only be used for RTG-enabled file systems.
+ */
+static inline bool
+xfs_verify_rgbext(
+	struct xfs_rtgroup	*rtg,
+	xfs_rgblock_t		rgbno,
+	xfs_extlen_t		len)
+{
+	ASSERT(xfs_has_rtgroups(rtg_mount(rtg)));
+
+	return xfs_verify_gbext(rtg_group(rtg), rgbno, len);
+}
+
 static inline xfs_rtblock_t
 xfs_rgbno_to_rtb(
 	struct xfs_rtgroup	*rtg,

[PATCH 08/37] xfs: add a realtime flag to the rmap update log redo items

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Extend the rmap update (RUI) log items to handle realtime volumes by
adding a new log intent item type.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_defer.h       |    1 
 fs/xfs/libxfs/xfs_log_format.h  |    6 +
 fs/xfs/libxfs/xfs_log_recover.h |    2 
 fs/xfs/libxfs/xfs_refcount.c    |    4 -
 fs/xfs/libxfs/xfs_rmap.c        |   17 ++-
 fs/xfs/libxfs/xfs_rmap.h        |    5 +
 fs/xfs/scrub/alloc_repair.c     |    2 
 fs/xfs/xfs_log_recover.c        |    2 
 fs/xfs/xfs_rmap_item.c          |  197 +++++++++++++++++++++++++++++++++++++--
 9 files changed, 213 insertions(+), 23 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_defer.h b/fs/xfs/libxfs/xfs_defer.h
index ec51b8465e61cb..1e2477eaa5a844 100644
--- a/fs/xfs/libxfs/xfs_defer.h
+++ b/fs/xfs/libxfs/xfs_defer.h
@@ -69,6 +69,7 @@ struct xfs_defer_op_type {
 extern const struct xfs_defer_op_type xfs_bmap_update_defer_type;
 extern const struct xfs_defer_op_type xfs_refcount_update_defer_type;
 extern const struct xfs_defer_op_type xfs_rmap_update_defer_type;
+extern const struct xfs_defer_op_type xfs_rtrmap_update_defer_type;
 extern const struct xfs_defer_op_type xfs_extent_free_defer_type;
 extern const struct xfs_defer_op_type xfs_agfl_free_defer_type;
 extern const struct xfs_defer_op_type xfs_rtextent_free_defer_type;
diff --git a/fs/xfs/libxfs/xfs_log_format.h b/fs/xfs/libxfs/xfs_log_format.h
index 15dec19b6c32ad..a7e0e479454d3d 100644
--- a/fs/xfs/libxfs/xfs_log_format.h
+++ b/fs/xfs/libxfs/xfs_log_format.h
@@ -250,6 +250,8 @@ typedef struct xfs_trans_header {
 #define	XFS_LI_XMD		0x1249  /* mapping exchange done */
 #define	XFS_LI_EFI_RT		0x124a	/* realtime extent free intent */
 #define	XFS_LI_EFD_RT		0x124b	/* realtime extent free done */
+#define	XFS_LI_RUI_RT		0x124c	/* realtime rmap update intent */
+#define	XFS_LI_RUD_RT		0x124d	/* realtime rmap update done */
 
 #define XFS_LI_TYPE_DESC \
 	{ XFS_LI_EFI,		"XFS_LI_EFI" }, \
@@ -271,7 +273,9 @@ typedef struct xfs_trans_header {
 	{ XFS_LI_XMI,		"XFS_LI_XMI" }, \
 	{ XFS_LI_XMD,		"XFS_LI_XMD" }, \
 	{ XFS_LI_EFI_RT,	"XFS_LI_EFI_RT" }, \
-	{ XFS_LI_EFD_RT,	"XFS_LI_EFD_RT" }
+	{ XFS_LI_EFD_RT,	"XFS_LI_EFD_RT" }, \
+	{ XFS_LI_RUI_RT,	"XFS_LI_RUI_RT" }, \
+	{ XFS_LI_RUD_RT,	"XFS_LI_RUD_RT" }
 
 /*
  * Inode Log Item Format definitions.
diff --git a/fs/xfs/libxfs/xfs_log_recover.h b/fs/xfs/libxfs/xfs_log_recover.h
index 5397a8ff004df8..abc705aff26dfe 100644
--- a/fs/xfs/libxfs/xfs_log_recover.h
+++ b/fs/xfs/libxfs/xfs_log_recover.h
@@ -79,6 +79,8 @@ extern const struct xlog_recover_item_ops xlog_xmi_item_ops;
 extern const struct xlog_recover_item_ops xlog_xmd_item_ops;
 extern const struct xlog_recover_item_ops xlog_rtefi_item_ops;
 extern const struct xlog_recover_item_ops xlog_rtefd_item_ops;
+extern const struct xlog_recover_item_ops xlog_rtrui_item_ops;
+extern const struct xlog_recover_item_ops xlog_rtrud_item_ops;
 
 /*
  * Macros, structures, prototypes for internal log manager use.
diff --git a/fs/xfs/libxfs/xfs_refcount.c b/fs/xfs/libxfs/xfs_refcount.c
index 26d3d7956e069d..bbb86dc9a25c7f 100644
--- a/fs/xfs/libxfs/xfs_refcount.c
+++ b/fs/xfs/libxfs/xfs_refcount.c
@@ -1831,7 +1831,7 @@ xfs_refcount_alloc_cow_extent(
 	__xfs_refcount_add(tp, XFS_REFCOUNT_ALLOC_COW, fsb, len);
 
 	/* Add rmap entry */
-	xfs_rmap_alloc_extent(tp, fsb, len, XFS_RMAP_OWN_COW);
+	xfs_rmap_alloc_extent(tp, false, fsb, len, XFS_RMAP_OWN_COW);
 }
 
 /* Forget a CoW staging event in the refcount btree. */
@@ -1847,7 +1847,7 @@ xfs_refcount_free_cow_extent(
 		return;
 
 	/* Remove rmap entry */
-	xfs_rmap_free_extent(tp, fsb, len, XFS_RMAP_OWN_COW);
+	xfs_rmap_free_extent(tp, false, fsb, len, XFS_RMAP_OWN_COW);
 	__xfs_refcount_add(tp, XFS_REFCOUNT_FREE_COW, fsb, len);
 }
 
diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c
index da1b004837d3ad..8d3cea90c7cd04 100644
--- a/fs/xfs/libxfs/xfs_rmap.c
+++ b/fs/xfs/libxfs/xfs_rmap.c
@@ -2710,6 +2710,7 @@ __xfs_rmap_add(
 	struct xfs_trans		*tp,
 	enum xfs_rmap_intent_type	type,
 	uint64_t			owner,
+	bool				isrt,
 	int				whichfork,
 	struct xfs_bmbt_irec		*bmap)
 {
@@ -2721,6 +2722,7 @@ __xfs_rmap_add(
 	ri->ri_owner = owner;
 	ri->ri_whichfork = whichfork;
 	ri->ri_bmap = *bmap;
+	ri->ri_realtime = isrt;
 
 	xfs_rmap_defer_add(tp, ri);
 }
@@ -2734,6 +2736,7 @@ xfs_rmap_map_extent(
 	struct xfs_bmbt_irec	*PREV)
 {
 	enum xfs_rmap_intent_type type = XFS_RMAP_MAP;
+	bool			isrt = xfs_ifork_is_realtime(ip, whichfork);
 
 	if (!xfs_rmap_update_is_needed(tp->t_mountp, whichfork))
 		return;
@@ -2741,7 +2744,7 @@ xfs_rmap_map_extent(
 	if (whichfork != XFS_ATTR_FORK && xfs_is_reflink_inode(ip))
 		type = XFS_RMAP_MAP_SHARED;
 
-	__xfs_rmap_add(tp, type, ip->i_ino, whichfork, PREV);
+	__xfs_rmap_add(tp, type, ip->i_ino, isrt, whichfork, PREV);
 }
 
 /* Unmap an extent out of a file. */
@@ -2753,6 +2756,7 @@ xfs_rmap_unmap_extent(
 	struct xfs_bmbt_irec	*PREV)
 {
 	enum xfs_rmap_intent_type type = XFS_RMAP_UNMAP;
+	bool			isrt = xfs_ifork_is_realtime(ip, whichfork);
 
 	if (!xfs_rmap_update_is_needed(tp->t_mountp, whichfork))
 		return;
@@ -2760,7 +2764,7 @@ xfs_rmap_unmap_extent(
 	if (whichfork != XFS_ATTR_FORK && xfs_is_reflink_inode(ip))
 		type = XFS_RMAP_UNMAP_SHARED;
 
-	__xfs_rmap_add(tp, type, ip->i_ino, whichfork, PREV);
+	__xfs_rmap_add(tp, type, ip->i_ino, isrt, whichfork, PREV);
 }
 
 /*
@@ -2778,6 +2782,7 @@ xfs_rmap_convert_extent(
 	struct xfs_bmbt_irec	*PREV)
 {
 	enum xfs_rmap_intent_type type = XFS_RMAP_CONVERT;
+	bool			isrt = xfs_ifork_is_realtime(ip, whichfork);
 
 	if (!xfs_rmap_update_is_needed(mp, whichfork))
 		return;
@@ -2785,13 +2790,14 @@ xfs_rmap_convert_extent(
 	if (whichfork != XFS_ATTR_FORK && xfs_is_reflink_inode(ip))
 		type = XFS_RMAP_CONVERT_SHARED;
 
-	__xfs_rmap_add(tp, type, ip->i_ino, whichfork, PREV);
+	__xfs_rmap_add(tp, type, ip->i_ino, isrt, whichfork, PREV);
 }
 
 /* Schedule the creation of an rmap for non-file data. */
 void
 xfs_rmap_alloc_extent(
 	struct xfs_trans	*tp,
+	bool			isrt,
 	xfs_fsblock_t		fsbno,
 	xfs_extlen_t		len,
 	uint64_t		owner)
@@ -2806,13 +2812,14 @@ xfs_rmap_alloc_extent(
 	bmap.br_startoff = 0;
 	bmap.br_state = XFS_EXT_NORM;
 
-	__xfs_rmap_add(tp, XFS_RMAP_ALLOC, owner, XFS_DATA_FORK, &bmap);
+	__xfs_rmap_add(tp, XFS_RMAP_ALLOC, owner, isrt, XFS_DATA_FORK, &bmap);
 }
 
 /* Schedule the deletion of an rmap for non-file data. */
 void
 xfs_rmap_free_extent(
 	struct xfs_trans	*tp,
+	bool			isrt,
 	xfs_fsblock_t		fsbno,
 	xfs_extlen_t		len,
 	uint64_t		owner)
@@ -2827,7 +2834,7 @@ xfs_rmap_free_extent(
 	bmap.br_startoff = 0;
 	bmap.br_state = XFS_EXT_NORM;
 
-	__xfs_rmap_add(tp, XFS_RMAP_FREE, owner, XFS_DATA_FORK, &bmap);
+	__xfs_rmap_add(tp, XFS_RMAP_FREE, owner, isrt, XFS_DATA_FORK, &bmap);
 }
 
 /* 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 1b19f54b65047f..5f39f6e53cd19a 100644
--- a/fs/xfs/libxfs/xfs_rmap.h
+++ b/fs/xfs/libxfs/xfs_rmap.h
@@ -175,6 +175,7 @@ struct xfs_rmap_intent {
 	uint64_t				ri_owner;
 	struct xfs_bmbt_irec			ri_bmap;
 	struct xfs_group			*ri_group;
+	bool					ri_realtime;
 };
 
 /* functions for updating the rmapbt based on bmbt map/unmap operations */
@@ -185,9 +186,9 @@ void xfs_rmap_unmap_extent(struct xfs_trans *tp, struct xfs_inode *ip,
 void xfs_rmap_convert_extent(struct xfs_mount *mp, struct xfs_trans *tp,
 		struct xfs_inode *ip, int whichfork,
 		struct xfs_bmbt_irec *imap);
-void xfs_rmap_alloc_extent(struct xfs_trans *tp, xfs_fsblock_t fsbno,
+void xfs_rmap_alloc_extent(struct xfs_trans *tp, bool isrt, xfs_fsblock_t fsbno,
 		xfs_extlen_t len, uint64_t owner);
-void xfs_rmap_free_extent(struct xfs_trans *tp, xfs_fsblock_t fsbno,
+void xfs_rmap_free_extent(struct xfs_trans *tp, bool isrt, xfs_fsblock_t fsbno,
 		xfs_extlen_t len, uint64_t owner);
 
 int xfs_rmap_finish_one(struct xfs_trans *tp, struct xfs_rmap_intent *ri,
diff --git a/fs/xfs/scrub/alloc_repair.c b/fs/xfs/scrub/alloc_repair.c
index 11e1e5404fc6dc..bed6a09aa79112 100644
--- a/fs/xfs/scrub/alloc_repair.c
+++ b/fs/xfs/scrub/alloc_repair.c
@@ -542,7 +542,7 @@ xrep_abt_dispose_one(
 
 	/* Add a deferred rmap for each extent we used. */
 	if (resv->used > 0)
-		xfs_rmap_alloc_extent(sc->tp,
+		xfs_rmap_alloc_extent(sc->tp, false,
 				xfs_agbno_to_fsb(pag, resv->agbno), resv->used,
 				XFS_RMAP_OWN_AG);
 
diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index 0af3d477197b24..5c95c97519c767 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -1820,6 +1820,8 @@ static const struct xlog_recover_item_ops *xlog_recover_item_ops[] = {
 	&xlog_xmd_item_ops,
 	&xlog_rtefi_item_ops,
 	&xlog_rtefd_item_ops,
+	&xlog_rtrui_item_ops,
+	&xlog_rtrud_item_ops,
 };
 
 static const struct xlog_recover_item_ops *
diff --git a/fs/xfs/xfs_rmap_item.c b/fs/xfs/xfs_rmap_item.c
index ac2913a7335871..e8caa600a95cae 100644
--- a/fs/xfs/xfs_rmap_item.c
+++ b/fs/xfs/xfs_rmap_item.c
@@ -23,6 +23,7 @@
 #include "xfs_ag.h"
 #include "xfs_btree.h"
 #include "xfs_trace.h"
+#include "xfs_rtgroup.h"
 
 struct kmem_cache	*xfs_rui_cache;
 struct kmem_cache	*xfs_rud_cache;
@@ -94,7 +95,9 @@ xfs_rui_item_format(
 	ASSERT(atomic_read(&ruip->rui_next_extent) ==
 			ruip->rui_format.rui_nextents);
 
-	ruip->rui_format.rui_type = XFS_LI_RUI;
+	ASSERT(lip->li_type == XFS_LI_RUI || lip->li_type == XFS_LI_RUI_RT);
+
+	ruip->rui_format.rui_type = lip->li_type;
 	ruip->rui_format.rui_size = 1;
 
 	xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_RUI_FORMAT, &ruip->rui_format,
@@ -137,12 +140,15 @@ xfs_rui_item_release(
 STATIC struct xfs_rui_log_item *
 xfs_rui_init(
 	struct xfs_mount		*mp,
+	unsigned short			item_type,
 	uint				nextents)
 
 {
 	struct xfs_rui_log_item		*ruip;
 
 	ASSERT(nextents > 0);
+	ASSERT(item_type == XFS_LI_RUI || item_type == XFS_LI_RUI_RT);
+
 	if (nextents > XFS_RUI_MAX_FAST_EXTENTS)
 		ruip = kzalloc(xfs_rui_log_item_sizeof(nextents),
 				GFP_KERNEL | __GFP_NOFAIL);
@@ -150,7 +156,7 @@ xfs_rui_init(
 		ruip = kmem_cache_zalloc(xfs_rui_cache,
 					 GFP_KERNEL | __GFP_NOFAIL);
 
-	xfs_log_item_init(mp, &ruip->rui_item, XFS_LI_RUI, &xfs_rui_item_ops);
+	xfs_log_item_init(mp, &ruip->rui_item, item_type, &xfs_rui_item_ops);
 	ruip->rui_format.rui_nextents = nextents;
 	ruip->rui_format.rui_id = (uintptr_t)(void *)ruip;
 	atomic_set(&ruip->rui_next_extent, 0);
@@ -189,7 +195,9 @@ xfs_rud_item_format(
 	struct xfs_rud_log_item	*rudp = RUD_ITEM(lip);
 	struct xfs_log_iovec	*vecp = NULL;
 
-	rudp->rud_format.rud_type = XFS_LI_RUD;
+	ASSERT(lip->li_type == XFS_LI_RUD || lip->li_type == XFS_LI_RUD_RT);
+
+	rudp->rud_format.rud_type = lip->li_type;
 	rudp->rud_format.rud_size = 1;
 
 	xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_RUD_FORMAT, &rudp->rud_format,
@@ -233,6 +241,14 @@ static inline struct xfs_rmap_intent *ri_entry(const struct list_head *e)
 	return list_entry(e, struct xfs_rmap_intent, ri_list);
 }
 
+static inline bool
+xfs_rui_item_isrt(const struct xfs_log_item *lip)
+{
+	ASSERT(lip->li_type == XFS_LI_RUI || lip->li_type == XFS_LI_RUI_RT);
+
+	return lip->li_type == XFS_LI_RUI_RT;
+}
+
 /* Sort rmap intents by AG. */
 static int
 xfs_rmap_update_diff_items(
@@ -305,18 +321,20 @@ xfs_rmap_update_log_item(
 }
 
 static struct xfs_log_item *
-xfs_rmap_update_create_intent(
+__xfs_rmap_update_create_intent(
 	struct xfs_trans		*tp,
 	struct list_head		*items,
 	unsigned int			count,
-	bool				sort)
+	bool				sort,
+	unsigned short			item_type)
 {
 	struct xfs_mount		*mp = tp->t_mountp;
-	struct xfs_rui_log_item		*ruip = xfs_rui_init(mp, count);
+	struct xfs_rui_log_item		*ruip;
 	struct xfs_rmap_intent		*ri;
 
 	ASSERT(count > 0);
 
+	ruip = xfs_rui_init(mp, item_type, count);
 	if (sort)
 		list_sort(mp, items, xfs_rmap_update_diff_items);
 	list_for_each_entry(ri, items, ri_list)
@@ -324,6 +342,23 @@ xfs_rmap_update_create_intent(
 	return &ruip->rui_item;
 }
 
+static struct xfs_log_item *
+xfs_rmap_update_create_intent(
+	struct xfs_trans		*tp,
+	struct list_head		*items,
+	unsigned int			count,
+	bool				sort)
+{
+	return __xfs_rmap_update_create_intent(tp, items, count, sort,
+			XFS_LI_RUI);
+}
+
+static inline unsigned short
+xfs_rud_type_from_rui(const struct xfs_rui_log_item *ruip)
+{
+	return xfs_rui_item_isrt(&ruip->rui_item) ? XFS_LI_RUD_RT : XFS_LI_RUD;
+}
+
 /* Get an RUD so we can process all the deferred rmap updates. */
 static struct xfs_log_item *
 xfs_rmap_update_create_done(
@@ -335,8 +370,8 @@ xfs_rmap_update_create_done(
 	struct xfs_rud_log_item		*rudp;
 
 	rudp = kmem_cache_zalloc(xfs_rud_cache, GFP_KERNEL | __GFP_NOFAIL);
-	xfs_log_item_init(tp->t_mountp, &rudp->rud_item, XFS_LI_RUD,
-			  &xfs_rud_item_ops);
+	xfs_log_item_init(tp->t_mountp, &rudp->rud_item,
+			xfs_rud_type_from_rui(ruip), &xfs_rud_item_ops);
 	rudp->rud_ruip = ruip;
 	rudp->rud_format.rud_rui_id = ruip->rui_format.rui_id;
 
@@ -351,11 +386,20 @@ xfs_rmap_defer_add(
 {
 	struct xfs_mount	*mp = tp->t_mountp;
 
+	/*
+	 * Deferred rmap updates for the realtime and data sections must use
+	 * separate transactions to finish deferred work because updates to
+	 * realtime metadata files can lock AGFs to allocate btree blocks and
+	 * we don't want that mixing with the AGF locks taken to finish data
+	 * section updates.
+	 */
 	ri->ri_group = xfs_group_intent_get(mp, ri->ri_bmap.br_startblock,
-			XG_TYPE_AG);
+			ri->ri_realtime ? XG_TYPE_RTG : XG_TYPE_AG);
 
 	trace_xfs_rmap_defer(mp, ri);
-	xfs_defer_add(tp, &ri->ri_list, &xfs_rmap_update_defer_type);
+	xfs_defer_add(tp, &ri->ri_list, ri->ri_realtime ?
+			&xfs_rtrmap_update_defer_type :
+			&xfs_rmap_update_defer_type);
 }
 
 /* Cancel a deferred rmap update. */
@@ -566,10 +610,13 @@ xfs_rmap_relog_intent(
 	struct xfs_map_extent		*map;
 	unsigned int			count;
 
+	ASSERT(intent->li_type == XFS_LI_RUI ||
+	       intent->li_type == XFS_LI_RUI_RT);
+
 	count = RUI_ITEM(intent)->rui_format.rui_nextents;
 	map = RUI_ITEM(intent)->rui_format.rui_extents;
 
-	ruip = xfs_rui_init(tp->t_mountp, count);
+	ruip = xfs_rui_init(tp->t_mountp, intent->li_type, count);
 	memcpy(ruip->rui_format.rui_extents, map, count * sizeof(*map));
 	atomic_set(&ruip->rui_next_extent, count);
 
@@ -589,6 +636,47 @@ const struct xfs_defer_op_type xfs_rmap_update_defer_type = {
 	.relog_intent	= xfs_rmap_relog_intent,
 };
 
+#ifdef CONFIG_XFS_RT
+static struct xfs_log_item *
+xfs_rtrmap_update_create_intent(
+	struct xfs_trans		*tp,
+	struct list_head		*items,
+	unsigned int			count,
+	bool				sort)
+{
+	return __xfs_rmap_update_create_intent(tp, items, count, sort,
+			XFS_LI_RUI_RT);
+}
+
+/* Clean up after calling xfs_rmap_finish_one. */
+STATIC void
+xfs_rtrmap_finish_one_cleanup(
+	struct xfs_trans	*tp,
+	struct xfs_btree_cur	*rcur,
+	int			error)
+{
+	if (rcur)
+		xfs_btree_del_cursor(rcur, error);
+}
+
+const struct xfs_defer_op_type xfs_rtrmap_update_defer_type = {
+	.name		= "rtrmap",
+	.max_items	= XFS_RUI_MAX_FAST_EXTENTS,
+	.create_intent	= xfs_rtrmap_update_create_intent,
+	.abort_intent	= xfs_rmap_update_abort_intent,
+	.create_done	= xfs_rmap_update_create_done,
+	.finish_item	= xfs_rmap_update_finish_item,
+	.finish_cleanup = xfs_rtrmap_finish_one_cleanup,
+	.cancel_item	= xfs_rmap_update_cancel_item,
+	.recover_work	= xfs_rmap_recover_work,
+	.relog_intent	= xfs_rmap_relog_intent,
+};
+#else
+const struct xfs_defer_op_type xfs_rtrmap_update_defer_type = {
+	.name		= "rtrmap",
+};
+#endif
+
 STATIC bool
 xfs_rui_item_match(
 	struct xfs_log_item	*lip,
@@ -654,7 +742,7 @@ xlog_recover_rui_commit_pass2(
 		return -EFSCORRUPTED;
 	}
 
-	ruip = xfs_rui_init(mp, rui_formatp->rui_nextents);
+	ruip = xfs_rui_init(mp, ITEM_TYPE(item), rui_formatp->rui_nextents);
 	xfs_rui_copy_format(&ruip->rui_format, rui_formatp);
 	atomic_set(&ruip->rui_next_extent, rui_formatp->rui_nextents);
 
@@ -668,6 +756,61 @@ const struct xlog_recover_item_ops xlog_rui_item_ops = {
 	.commit_pass2		= xlog_recover_rui_commit_pass2,
 };
 
+#ifdef CONFIG_XFS_RT
+STATIC int
+xlog_recover_rtrui_commit_pass2(
+	struct xlog			*log,
+	struct list_head		*buffer_list,
+	struct xlog_recover_item	*item,
+	xfs_lsn_t			lsn)
+{
+	struct xfs_mount		*mp = log->l_mp;
+	struct xfs_rui_log_item		*ruip;
+	struct xfs_rui_log_format	*rui_formatp;
+	size_t				len;
+
+	rui_formatp = item->ri_buf[0].i_addr;
+
+	if (item->ri_buf[0].i_len < xfs_rui_log_format_sizeof(0)) {
+		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
+				item->ri_buf[0].i_addr, item->ri_buf[0].i_len);
+		return -EFSCORRUPTED;
+	}
+
+	len = xfs_rui_log_format_sizeof(rui_formatp->rui_nextents);
+	if (item->ri_buf[0].i_len != len) {
+		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
+				item->ri_buf[0].i_addr, item->ri_buf[0].i_len);
+		return -EFSCORRUPTED;
+	}
+
+	ruip = xfs_rui_init(mp, ITEM_TYPE(item), rui_formatp->rui_nextents);
+	xfs_rui_copy_format(&ruip->rui_format, rui_formatp);
+	atomic_set(&ruip->rui_next_extent, rui_formatp->rui_nextents);
+
+	xlog_recover_intent_item(log, &ruip->rui_item, lsn,
+			&xfs_rtrmap_update_defer_type);
+	return 0;
+}
+#else
+STATIC int
+xlog_recover_rtrui_commit_pass2(
+	struct xlog			*log,
+	struct list_head		*buffer_list,
+	struct xlog_recover_item	*item,
+	xfs_lsn_t			lsn)
+{
+	XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, log->l_mp,
+			item->ri_buf[0].i_addr, item->ri_buf[0].i_len);
+	return -EFSCORRUPTED;
+}
+#endif
+
+const struct xlog_recover_item_ops xlog_rtrui_item_ops = {
+	.item_type		= XFS_LI_RUI_RT,
+	.commit_pass2		= xlog_recover_rtrui_commit_pass2,
+};
+
 /*
  * This routine is called when an RUD format structure is found in a committed
  * transaction in the log. Its purpose is to cancel the corresponding RUI if it
@@ -699,3 +842,33 @@ const struct xlog_recover_item_ops xlog_rud_item_ops = {
 	.item_type		= XFS_LI_RUD,
 	.commit_pass2		= xlog_recover_rud_commit_pass2,
 };
+
+#ifdef CONFIG_XFS_RT
+STATIC int
+xlog_recover_rtrud_commit_pass2(
+	struct xlog			*log,
+	struct list_head		*buffer_list,
+	struct xlog_recover_item	*item,
+	xfs_lsn_t			lsn)
+{
+	struct xfs_rud_log_format	*rud_formatp;
+
+	rud_formatp = item->ri_buf[0].i_addr;
+	if (item->ri_buf[0].i_len != sizeof(struct xfs_rud_log_format)) {
+		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, log->l_mp,
+				rud_formatp, item->ri_buf[0].i_len);
+		return -EFSCORRUPTED;
+	}
+
+	xlog_recover_release_intent(log, XFS_LI_RUI_RT,
+			rud_formatp->rud_rui_id);
+	return 0;
+}
+#else
+# define xlog_recover_rtrud_commit_pass2	xlog_recover_rtrui_commit_pass2
+#endif
+
+const struct xlog_recover_item_ops xlog_rtrud_item_ops = {
+	.item_type		= XFS_LI_RUD_RT,
+	.commit_pass2		= xlog_recover_rtrud_commit_pass2,
+};

[PATCH 09/37] xfs: support recovering rmap intent items targetting realtime extents

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Now that we have rmap on the realtime device and rmap intent items that
target the realtime device, log recovery has to support remapping
extents on the realtime volume.  Make this work.  Identify rtrmapbt
blocks in the log correctly so that we can validate them during log
recovery.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/xfs_buf_item_recover.c |    4 ++++
 fs/xfs/xfs_rmap_item.c        |   15 ++++++++++++---
 2 files changed, 16 insertions(+), 3 deletions(-)


diff --git a/fs/xfs/xfs_buf_item_recover.c b/fs/xfs/xfs_buf_item_recover.c
index 3d0c6402cb3634..4f2e4ea29e1f57 100644
--- a/fs/xfs/xfs_buf_item_recover.c
+++ b/fs/xfs/xfs_buf_item_recover.c
@@ -262,6 +262,9 @@ 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;
@@ -855,6 +858,7 @@ 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;
diff --git a/fs/xfs/xfs_rmap_item.c b/fs/xfs/xfs_rmap_item.c
index e8caa600a95cae..89decffe76c8b5 100644
--- a/fs/xfs/xfs_rmap_item.c
+++ b/fs/xfs/xfs_rmap_item.c
@@ -459,6 +459,7 @@ xfs_rmap_update_abort_intent(
 static inline bool
 xfs_rui_validate_map(
 	struct xfs_mount		*mp,
+	bool				isrt,
 	struct xfs_map_extent		*map)
 {
 	if (!xfs_has_rmapbt(mp))
@@ -488,6 +489,9 @@ xfs_rui_validate_map(
 	if (!xfs_verify_fileext(mp, map->me_startoff, map->me_len))
 		return false;
 
+	if (isrt)
+		return xfs_verify_rtbext(mp, map->me_startblock, map->me_len);
+
 	return xfs_verify_fsbext(mp, map->me_startblock, map->me_len);
 }
 
@@ -495,6 +499,7 @@ static inline void
 xfs_rui_recover_work(
 	struct xfs_mount		*mp,
 	struct xfs_defer_pending	*dfp,
+	bool				isrt,
 	const struct xfs_map_extent	*map)
 {
 	struct xfs_rmap_intent		*ri;
@@ -539,7 +544,9 @@ xfs_rui_recover_work(
 	ri->ri_bmap.br_blockcount = map->me_len;
 	ri->ri_bmap.br_state = (map->me_flags & XFS_RMAP_EXTENT_UNWRITTEN) ?
 			XFS_EXT_UNWRITTEN : XFS_EXT_NORM;
-	ri->ri_group = xfs_group_intent_get(mp, map->me_startblock, XG_TYPE_AG);
+	ri->ri_group = xfs_group_intent_get(mp, map->me_startblock,
+			isrt ? XG_TYPE_RTG : XG_TYPE_AG);
+	ri->ri_realtime = isrt;
 
 	xfs_defer_add_item(dfp, &ri->ri_list);
 }
@@ -558,6 +565,7 @@ xfs_rmap_recover_work(
 	struct xfs_rui_log_item		*ruip = RUI_ITEM(lip);
 	struct xfs_trans		*tp;
 	struct xfs_mount		*mp = lip->li_log->l_mp;
+	bool				isrt = xfs_rui_item_isrt(lip);
 	int				i;
 	int				error = 0;
 
@@ -567,7 +575,7 @@ xfs_rmap_recover_work(
 	 * just toss the RUI.
 	 */
 	for (i = 0; i < ruip->rui_format.rui_nextents; i++) {
-		if (!xfs_rui_validate_map(mp,
+		if (!xfs_rui_validate_map(mp, isrt,
 					&ruip->rui_format.rui_extents[i])) {
 			XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
 					&ruip->rui_format,
@@ -575,7 +583,8 @@ xfs_rmap_recover_work(
 			return -EFSCORRUPTED;
 		}
 
-		xfs_rui_recover_work(mp, dfp, &ruip->rui_format.rui_extents[i]);
+		xfs_rui_recover_work(mp, dfp, isrt,
+				&ruip->rui_format.rui_extents[i]);
 	}
 
 	resv = xlog_recover_resv(&M_RES(mp)->tr_itruncate);

[PATCH 10/37] xfs: pretty print metadata file types in error messages

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Create a helper function to turn a metadata file type code into a
printable string, and use this to complain about lockdep problems with
rtgroup inodes.  We'll use this more in the next patch.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_metafile.c |   18 ++++++++++++++++++
 fs/xfs/libxfs/xfs_metafile.h |    2 ++
 fs/xfs/libxfs/xfs_rtgroup.c  |    3 ++-
 3 files changed, 22 insertions(+), 1 deletion(-)


diff --git a/fs/xfs/libxfs/xfs_metafile.c b/fs/xfs/libxfs/xfs_metafile.c
index e151663cc9efd6..2f5f554a36d4c9 100644
--- a/fs/xfs/libxfs/xfs_metafile.c
+++ b/fs/xfs/libxfs/xfs_metafile.c
@@ -22,6 +22,24 @@
 #include "xfs_error.h"
 #include "xfs_alloc.h"
 
+static const struct {
+	enum xfs_metafile_type	mtype;
+	const char		*name;
+} xfs_metafile_type_strs[] = { XFS_METAFILE_TYPE_STR };
+
+const char *
+xfs_metafile_type_str(enum xfs_metafile_type metatype)
+{
+	unsigned int	i;
+
+	for (i = 0; i < ARRAY_SIZE(xfs_metafile_type_strs); i++) {
+		if (xfs_metafile_type_strs[i].mtype == metatype)
+			return xfs_metafile_type_strs[i].name;
+	}
+
+	return NULL;
+}
+
 /* Set up an inode to be recognized as a metadata directory inode. */
 void
 xfs_metafile_set_iflag(
diff --git a/fs/xfs/libxfs/xfs_metafile.h b/fs/xfs/libxfs/xfs_metafile.h
index 8d8f08a6071c23..95af4b52e5a75f 100644
--- a/fs/xfs/libxfs/xfs_metafile.h
+++ b/fs/xfs/libxfs/xfs_metafile.h
@@ -6,6 +6,8 @@
 #ifndef __XFS_METAFILE_H__
 #define __XFS_METAFILE_H__
 
+const char *xfs_metafile_type_str(enum xfs_metafile_type metatype);
+
 /* All metadata files must have these flags set. */
 #define XFS_METAFILE_DIFLAGS	(XFS_DIFLAG_IMMUTABLE | \
 				 XFS_DIFLAG_SYNC | \
diff --git a/fs/xfs/libxfs/xfs_rtgroup.c b/fs/xfs/libxfs/xfs_rtgroup.c
index a79b734e70440d..9e5fdc0dc55cef 100644
--- a/fs/xfs/libxfs/xfs_rtgroup.c
+++ b/fs/xfs/libxfs/xfs_rtgroup.c
@@ -282,7 +282,8 @@ xfs_rtginode_ilock_print_fn(
 	const struct xfs_inode *ip =
 		container_of(m, struct xfs_inode, i_lock.dep_map);
 
-	printk(KERN_CONT " rgno=%u", ip->i_projid);
+	printk(KERN_CONT " rgno=%u metatype=%s", ip->i_projid,
+			xfs_metafile_type_str(ip->i_metatype));
 }
 
 /*

[PATCH 11/37] xfs: support file data forks containing metadata btrees

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Create a new fork format type for metadata btrees.  This fork type
requires that the inode is in the metadata directory tree, and only
applies to the data fork.  The actual type of the metadata btree itself
is determined by the di_metatype field.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_format.h      |    6 ++++--
 fs/xfs/libxfs/xfs_inode_buf.c   |   23 ++++++++++++++++++++--
 fs/xfs/libxfs/xfs_inode_fork.c  |   19 +++++++++++++++++++
 fs/xfs/scrub/bmap.c             |    1 +
 fs/xfs/scrub/bmap_repair.c      |    1 +
 fs/xfs/scrub/inode.c            |    4 ++++
 fs/xfs/scrub/inode_repair.c     |   36 +++++++++++++++++++++++++++++++++++
 fs/xfs/scrub/rmap_repair.c      |   31 ++++++++++++++++++++----------
 fs/xfs/xfs_inode.c              |   19 ++++++++++++++++++-
 fs/xfs/xfs_inode_item.c         |    2 ++
 fs/xfs/xfs_inode_item_recover.c |   40 ++++++++++++++++++++++++++++++++++-----
 fs/xfs/xfs_trace.h              |    1 +
 12 files changed, 162 insertions(+), 21 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index 469fc7afa591b4..41ea4283c43cb4 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -997,7 +997,8 @@ 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_META_BTREE,	/* metadata btree */
 };
 
 #define XFS_INODE_FORMAT_STR \
@@ -1005,7 +1006,8 @@ enum xfs_dinode_fmt {
 	{ XFS_DINODE_FMT_LOCAL,		"local" }, \
 	{ XFS_DINODE_FMT_EXTENTS,	"extent" }, \
 	{ XFS_DINODE_FMT_BTREE,		"btree" }, \
-	{ XFS_DINODE_FMT_UUID,		"uuid" }
+	{ XFS_DINODE_FMT_UUID,		"uuid" }, \
+	{ XFS_DINODE_FMT_META_BTREE,	"meta_btree" }
 
 /*
  * Max values for extnum and aextnum.
diff --git a/fs/xfs/libxfs/xfs_inode_buf.c b/fs/xfs/libxfs/xfs_inode_buf.c
index 424861fbf1bd49..1648d72d6ed95a 100644
--- a/fs/xfs/libxfs/xfs_inode_buf.c
+++ b/fs/xfs/libxfs/xfs_inode_buf.c
@@ -441,6 +441,16 @@ xfs_dinode_verify_fork(
 		if (di_nextents > max_extents)
 			return __this_address;
 		break;
+	case XFS_DINODE_FMT_META_BTREE:
+		if (!xfs_has_metadir(mp))
+			return __this_address;
+		if (!(dip->di_flags2 & cpu_to_be64(XFS_DIFLAG2_METADATA)))
+			return __this_address;
+		switch (be16_to_cpu(dip->di_metatype)) {
+		default:
+			return __this_address;
+		}
+		break;
 	default:
 		return __this_address;
 	}
@@ -460,6 +470,10 @@ xfs_dinode_verify_forkoff(
 		if (dip->di_forkoff != (roundup(sizeof(xfs_dev_t), 8) >> 3))
 			return __this_address;
 		break;
+	case XFS_DINODE_FMT_META_BTREE:
+		if (!xfs_has_metadir(mp) || !xfs_has_parent(mp))
+			return __this_address;
+		fallthrough;
 	case XFS_DINODE_FMT_LOCAL:	/* fall through ... */
 	case XFS_DINODE_FMT_EXTENTS:    /* fall through ... */
 	case XFS_DINODE_FMT_BTREE:
@@ -637,9 +651,6 @@ xfs_dinode_verify(
 	if (mode && nextents + naextents > nblocks)
 		return __this_address;
 
-	if (nextents + naextents == 0 && nblocks != 0)
-		return __this_address;
-
 	if (S_ISDIR(mode) && nextents > mp->m_dir_geo->max_extents)
 		return __this_address;
 
@@ -743,6 +754,12 @@ xfs_dinode_verify(
 			return fa;
 	}
 
+	/* metadata inodes containing btrees always have zero extent count */
+	if (XFS_DFORK_FORMAT(dip, XFS_DATA_FORK) != XFS_DINODE_FMT_META_BTREE) {
+		if (nextents + naextents == 0 && nblocks != 0)
+			return __this_address;
+	}
+
 	return NULL;
 }
 
diff --git a/fs/xfs/libxfs/xfs_inode_fork.c b/fs/xfs/libxfs/xfs_inode_fork.c
index 60853bac289a39..1a782339396dc3 100644
--- a/fs/xfs/libxfs/xfs_inode_fork.c
+++ b/fs/xfs/libxfs/xfs_inode_fork.c
@@ -267,6 +267,12 @@ xfs_iformat_data_fork(
 			return xfs_iformat_extents(ip, dip, XFS_DATA_FORK);
 		case XFS_DINODE_FMT_BTREE:
 			return xfs_iformat_btree(ip, dip, XFS_DATA_FORK);
+		case XFS_DINODE_FMT_META_BTREE:
+			switch (ip->i_metatype) {
+			default:
+				break;
+			}
+			fallthrough;
 		default:
 			xfs_inode_verifier_error(ip, -EFSCORRUPTED, __func__,
 					dip, sizeof(*dip), __this_address);
@@ -601,6 +607,19 @@ xfs_iflush_fork(
 		}
 		break;
 
+	case XFS_DINODE_FMT_META_BTREE:
+		ASSERT(whichfork == XFS_DATA_FORK);
+
+		if (!(iip->ili_fields & brootflag[whichfork]))
+			break;
+
+		switch (ip->i_metatype) {
+		default:
+			ASSERT(0);
+			break;
+		}
+		break;
+
 	default:
 		ASSERT(0);
 		break;
diff --git a/fs/xfs/scrub/bmap.c b/fs/xfs/scrub/bmap.c
index 7e00312225ed10..0d7ad692822d48 100644
--- a/fs/xfs/scrub/bmap.c
+++ b/fs/xfs/scrub/bmap.c
@@ -983,6 +983,7 @@ xchk_bmap(
 	case XFS_DINODE_FMT_UUID:
 	case XFS_DINODE_FMT_DEV:
 	case XFS_DINODE_FMT_LOCAL:
+	case XFS_DINODE_FMT_META_BTREE:
 		/* No mappings to check. */
 		if (whichfork == XFS_COW_FORK)
 			xchk_fblock_set_corrupt(sc, whichfork, 0);
diff --git a/fs/xfs/scrub/bmap_repair.c b/fs/xfs/scrub/bmap_repair.c
index 7c4955482641f7..141d36f1da9a71 100644
--- a/fs/xfs/scrub/bmap_repair.c
+++ b/fs/xfs/scrub/bmap_repair.c
@@ -731,6 +731,7 @@ xrep_bmap_check_inputs(
 	case XFS_DINODE_FMT_DEV:
 	case XFS_DINODE_FMT_LOCAL:
 	case XFS_DINODE_FMT_UUID:
+	case XFS_DINODE_FMT_META_BTREE:
 		return -ECANCELED;
 	case XFS_DINODE_FMT_EXTENTS:
 	case XFS_DINODE_FMT_BTREE:
diff --git a/fs/xfs/scrub/inode.c b/fs/xfs/scrub/inode.c
index 25ee66e7649d40..2e911f38deaebe 100644
--- a/fs/xfs/scrub/inode.c
+++ b/fs/xfs/scrub/inode.c
@@ -502,6 +502,10 @@ xchk_dinode(
 		if (!S_ISREG(mode) && !S_ISDIR(mode))
 			xchk_ino_set_corrupt(sc, ino);
 		break;
+	case XFS_DINODE_FMT_META_BTREE:
+		if (!S_ISREG(mode))
+			xchk_ino_set_corrupt(sc, ino);
+		break;
 	case XFS_DINODE_FMT_UUID:
 	default:
 		xchk_ino_set_corrupt(sc, ino);
diff --git a/fs/xfs/scrub/inode_repair.c b/fs/xfs/scrub/inode_repair.c
index 5a58ddd27bd2f5..7faa27472b9129 100644
--- a/fs/xfs/scrub/inode_repair.c
+++ b/fs/xfs/scrub/inode_repair.c
@@ -888,6 +888,25 @@ xrep_dinode_bad_bmbt_fork(
 	return false;
 }
 
+/* Check a metadata-btree fork. */
+STATIC bool
+xrep_dinode_bad_metabt_fork(
+	struct xfs_scrub	*sc,
+	struct xfs_dinode	*dip,
+	unsigned int		dfork_size,
+	int			whichfork)
+{
+	if (whichfork != XFS_DATA_FORK)
+		return true;
+
+	switch (be16_to_cpu(dip->di_metatype)) {
+	default:
+		return true;
+	}
+
+	return false;
+}
+
 /*
  * Check the data fork for things that will fail the ifork verifiers or the
  * ifork formatters.
@@ -968,6 +987,11 @@ xrep_dinode_check_dfork(
 				XFS_DATA_FORK))
 			return true;
 		break;
+	case XFS_DINODE_FMT_META_BTREE:
+		if (xrep_dinode_bad_metabt_fork(sc, dip, dfork_size,
+				XFS_DATA_FORK))
+			return true;
+		break;
 	default:
 		return true;
 	}
@@ -1088,6 +1112,11 @@ xrep_dinode_check_afork(
 					XFS_ATTR_FORK))
 			return true;
 		break;
+	case XFS_DINODE_FMT_META_BTREE:
+		if (xrep_dinode_bad_metabt_fork(sc, dip, afork_size,
+					XFS_ATTR_FORK))
+			return true;
+		break;
 	default:
 		return true;
 	}
@@ -1241,6 +1270,13 @@ xrep_dinode_ensure_forkoff(
 		bmdr = XFS_DFORK_PTR(dip, XFS_DATA_FORK);
 		dfork_min = xfs_bmap_broot_space(sc->mp, bmdr);
 		break;
+	case XFS_DINODE_FMT_META_BTREE:
+		switch (be16_to_cpu(dip->di_metatype)) {
+		default:
+			dfork_min = 0;
+			break;
+		}
+		break;
 	default:
 		dfork_min = 0;
 		break;
diff --git a/fs/xfs/scrub/rmap_repair.c b/fs/xfs/scrub/rmap_repair.c
index a0a227d183d28d..2a0b9e3d0fbaee 100644
--- a/fs/xfs/scrub/rmap_repair.c
+++ b/fs/xfs/scrub/rmap_repair.c
@@ -499,6 +499,14 @@ xrep_rmap_scan_iext(
 	return xrep_rmap_stash_accumulated(rf);
 }
 
+static int
+xrep_rmap_scan_meta_btree(
+	struct xrep_rmap_ifork	*rf,
+	struct xfs_inode	*ip)
+{
+	return -EFSCORRUPTED; /* XXX placeholder */
+}
+
 /* Find all the extents from a given AG in an inode fork. */
 STATIC int
 xrep_rmap_scan_ifork(
@@ -512,14 +520,14 @@ xrep_rmap_scan_ifork(
 		.whichfork	= whichfork,
 	};
 	struct xfs_ifork	*ifp = xfs_ifork_ptr(ip, whichfork);
+	bool			mappings_done;
 	int			error = 0;
 
 	if (!ifp)
 		return 0;
 
-	if (ifp->if_format == XFS_DINODE_FMT_BTREE) {
-		bool		mappings_done;
-
+	switch (ifp->if_format) {
+	case XFS_DINODE_FMT_BTREE:
 		/*
 		 * Scan the bmap btree for data device mappings.  This includes
 		 * the btree blocks themselves, even if this is a realtime
@@ -528,15 +536,18 @@ xrep_rmap_scan_ifork(
 		error = xrep_rmap_scan_bmbt(&rf, ip, &mappings_done);
 		if (error || mappings_done)
 			return error;
-	} else if (ifp->if_format != XFS_DINODE_FMT_EXTENTS) {
-		return 0;
+		fallthrough;
+	case XFS_DINODE_FMT_EXTENTS:
+		/* Scan incore extent cache if this isn't a realtime file. */
+		if (xfs_ifork_is_realtime(ip, whichfork))
+			return 0;
+
+		return xrep_rmap_scan_iext(&rf, ifp);
+	case XFS_DINODE_FMT_META_BTREE:
+		return xrep_rmap_scan_meta_btree(&rf, ip);
 	}
 
-	/* Scan incore extent cache if this isn't a realtime file. */
-	if (xfs_ifork_is_realtime(ip, whichfork))
-		return 0;
-
-	return xrep_rmap_scan_iext(&rf, ifp);
+	return 0;
 }
 
 /*
diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
index c8ad2606f928b2..c95fe1b1de4e6f 100644
--- a/fs/xfs/xfs_inode.c
+++ b/fs/xfs/xfs_inode.c
@@ -2382,7 +2382,16 @@ xfs_iflush(
 			__func__, ip->i_ino, be16_to_cpu(dip->di_magic), dip);
 		goto flush_out;
 	}
-	if (S_ISREG(VFS_I(ip)->i_mode)) {
+	if (ip->i_df.if_format == XFS_DINODE_FMT_META_BTREE) {
+		if (!S_ISREG(VFS_I(ip)->i_mode) ||
+		    !(ip->i_diflags2 & XFS_DIFLAG2_METADATA)) {
+			xfs_alert_tag(mp, XFS_PTAG_IFLUSH,
+				"%s: Bad %s meta btree inode %Lu, ptr "PTR_FMT,
+				__func__, xfs_metafile_type_str(ip->i_metatype),
+				ip->i_ino, ip);
+			goto flush_out;
+		}
+	} else if (S_ISREG(VFS_I(ip)->i_mode)) {
 		if (XFS_TEST_ERROR(
 		    ip->i_df.if_format != XFS_DINODE_FMT_EXTENTS &&
 		    ip->i_df.if_format != XFS_DINODE_FMT_BTREE,
@@ -2422,6 +2431,14 @@ xfs_iflush(
 		goto flush_out;
 	}
 
+	if (xfs_inode_has_attr_fork(ip) &&
+	    ip->i_af.if_format == XFS_DINODE_FMT_META_BTREE) {
+		xfs_alert_tag(mp, XFS_PTAG_IFLUSH,
+			"%s: meta btree in inode %Lu attr fork, ptr "PTR_FMT,
+			__func__, ip->i_ino, ip);
+		goto flush_out;
+	}
+
 	/*
 	 * Inode item log recovery for v2 inodes are dependent on the flushiter
 	 * count for correct sequencing.  We bump the flush iteration count so
diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c
index 912f0b1bc3cb70..a174f64b8bb250 100644
--- a/fs/xfs/xfs_inode_item.c
+++ b/fs/xfs/xfs_inode_item.c
@@ -242,6 +242,7 @@ xfs_inode_item_data_fork_size(
 		}
 		break;
 	case XFS_DINODE_FMT_BTREE:
+	case XFS_DINODE_FMT_META_BTREE:
 		if ((iip->ili_fields & XFS_ILOG_DBROOT) &&
 		    ip->i_df.if_broot_bytes > 0) {
 			*nbytes += ip->i_df.if_broot_bytes;
@@ -362,6 +363,7 @@ xfs_inode_item_format_data_fork(
 		}
 		break;
 	case XFS_DINODE_FMT_BTREE:
+	case XFS_DINODE_FMT_META_BTREE:
 		iip->ili_fields &=
 			~(XFS_ILOG_DDATA | XFS_ILOG_DEXT | XFS_ILOG_DEV);
 
diff --git a/fs/xfs/xfs_inode_item_recover.c b/fs/xfs/xfs_inode_item_recover.c
index e70d2611456bc9..6e9b3bfc718c0b 100644
--- a/fs/xfs/xfs_inode_item_recover.c
+++ b/fs/xfs/xfs_inode_item_recover.c
@@ -266,6 +266,35 @@ xlog_dinode_verify_extent_counts(
 	return 0;
 }
 
+static inline int
+xlog_recover_inode_dbroot(
+	struct xfs_mount	*mp,
+	void			*src,
+	unsigned int		len,
+	struct xfs_dinode	*dip)
+{
+	void			*dfork = XFS_DFORK_DPTR(dip);
+	unsigned int		dsize = XFS_DFORK_DSIZE(dip, mp);
+
+	switch (dip->di_format) {
+	case XFS_DINODE_FMT_BTREE:
+		xfs_bmbt_to_bmdr(mp, src, len, dfork, dsize);
+		break;
+	case XFS_DINODE_FMT_META_BTREE:
+		switch (be16_to_cpu(dip->di_metatype)) {
+		default:
+			ASSERT(0);
+			return -EFSCORRUPTED;
+		}
+		break;
+	default:
+		ASSERT(0);
+		return -EFSCORRUPTED;
+	}
+
+	return 0;
+}
+
 STATIC int
 xlog_recover_inode_commit_pass2(
 	struct xlog			*log,
@@ -393,8 +422,9 @@ xlog_recover_inode_commit_pass2(
 
 
 	if (unlikely(S_ISREG(ldip->di_mode))) {
-		if ((ldip->di_format != XFS_DINODE_FMT_EXTENTS) &&
-		    (ldip->di_format != XFS_DINODE_FMT_BTREE)) {
+		if (ldip->di_format != XFS_DINODE_FMT_EXTENTS &&
+		    ldip->di_format != XFS_DINODE_FMT_BTREE &&
+		    ldip->di_format != XFS_DINODE_FMT_META_BTREE) {
 			XFS_CORRUPTION_ERROR(
 				"Bad log dinode data fork format for regular file",
 				XFS_ERRLEVEL_LOW, mp, ldip, sizeof(*ldip));
@@ -475,9 +505,9 @@ xlog_recover_inode_commit_pass2(
 		break;
 
 	case XFS_ILOG_DBROOT:
-		xfs_bmbt_to_bmdr(mp, (struct xfs_btree_block *)src, len,
-				 (struct xfs_bmdr_block *)XFS_DFORK_DPTR(dip),
-				 XFS_DFORK_DSIZE(dip, mp));
+		error = xlog_recover_inode_dbroot(mp, src, len, dip);
+		if (error)
+			goto out_release;
 		break;
 
 	default:
diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h
index 8b7bb1f5ae3c6f..a098935163b7c2 100644
--- a/fs/xfs/xfs_trace.h
+++ b/fs/xfs/xfs_trace.h
@@ -2299,6 +2299,7 @@ TRACE_DEFINE_ENUM(XFS_DINODE_FMT_LOCAL);
 TRACE_DEFINE_ENUM(XFS_DINODE_FMT_EXTENTS);
 TRACE_DEFINE_ENUM(XFS_DINODE_FMT_BTREE);
 TRACE_DEFINE_ENUM(XFS_DINODE_FMT_UUID);
+TRACE_DEFINE_ENUM(XFS_DINODE_FMT_META_BTREE);
 
 DECLARE_EVENT_CLASS(xfs_swap_extent_class,
 	TP_PROTO(struct xfs_inode *ip, int which),

[PATCH 12/37] xfs: add realtime reverse map inode to metadata directory

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Add a metadir path to select the realtime rmap btree 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" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_format.h       |    4 +++-
 fs/xfs/libxfs/xfs_inode_buf.c    |    9 +++++++++
 fs/xfs/libxfs/xfs_inode_fork.c   |    6 ++++++
 fs/xfs/libxfs/xfs_rtgroup.c      |   20 ++++++++++++++++++--
 fs/xfs/libxfs/xfs_rtgroup.h      |    8 ++++++++
 fs/xfs/libxfs/xfs_rtrmap_btree.c |    6 +++---
 6 files changed, 47 insertions(+), 6 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index 41ea4283c43cb4..f32c9fda5a195f 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -857,6 +857,7 @@ enum xfs_metafile_type {
 	XFS_METAFILE_PRJQUOTA,		/* project quota */
 	XFS_METAFILE_RTBITMAP,		/* rt bitmap */
 	XFS_METAFILE_RTSUMMARY,		/* rt summary */
+	XFS_METAFILE_RTRMAP,		/* rt rmap */
 
 	XFS_METAFILE_MAX
 } __packed;
@@ -868,7 +869,8 @@ enum xfs_metafile_type {
 	{ XFS_METAFILE_GRPQUOTA,	"grpquota" }, \
 	{ XFS_METAFILE_PRJQUOTA,	"prjquota" }, \
 	{ XFS_METAFILE_RTBITMAP,	"rtbitmap" }, \
-	{ XFS_METAFILE_RTSUMMARY,	"rtsummary" }
+	{ XFS_METAFILE_RTSUMMARY,	"rtsummary" }, \
+	{ XFS_METAFILE_RTRMAP,		"rtrmap" }
 
 /*
  * On-disk inode structure.
diff --git a/fs/xfs/libxfs/xfs_inode_buf.c b/fs/xfs/libxfs/xfs_inode_buf.c
index 1648d72d6ed95a..17cb91b89fcaa1 100644
--- a/fs/xfs/libxfs/xfs_inode_buf.c
+++ b/fs/xfs/libxfs/xfs_inode_buf.c
@@ -447,6 +447,15 @@ xfs_dinode_verify_fork(
 		if (!(dip->di_flags2 & cpu_to_be64(XFS_DIFLAG2_METADATA)))
 			return __this_address;
 		switch (be16_to_cpu(dip->di_metatype)) {
+		case XFS_METAFILE_RTRMAP:
+			/*
+			 * growfs must create the rtrmap inodes before adding a
+			 * realtime volume to the filesystem, so we cannot use
+			 * the rtrmapbt predicate here.
+			 */
+			if (!xfs_has_rmapbt(mp))
+				return __this_address;
+			break;
 		default:
 			return __this_address;
 		}
diff --git a/fs/xfs/libxfs/xfs_inode_fork.c b/fs/xfs/libxfs/xfs_inode_fork.c
index 1a782339396dc3..7c2b071a58384d 100644
--- a/fs/xfs/libxfs/xfs_inode_fork.c
+++ b/fs/xfs/libxfs/xfs_inode_fork.c
@@ -269,6 +269,9 @@ xfs_iformat_data_fork(
 			return xfs_iformat_btree(ip, dip, XFS_DATA_FORK);
 		case XFS_DINODE_FMT_META_BTREE:
 			switch (ip->i_metatype) {
+			case XFS_METAFILE_RTRMAP:
+				ASSERT(0); /* to be implemented later */
+				return -EFSCORRUPTED;
 			default:
 				break;
 			}
@@ -614,6 +617,9 @@ xfs_iflush_fork(
 			break;
 
 		switch (ip->i_metatype) {
+		case XFS_METAFILE_RTRMAP:
+			ASSERT(0); /* to be implemented later */
+			break;
 		default:
 			ASSERT(0);
 			break;
diff --git a/fs/xfs/libxfs/xfs_rtgroup.c b/fs/xfs/libxfs/xfs_rtgroup.c
index 9e5fdc0dc55cef..1b56c13b282788 100644
--- a/fs/xfs/libxfs/xfs_rtgroup.c
+++ b/fs/xfs/libxfs/xfs_rtgroup.c
@@ -315,6 +315,8 @@ struct xfs_rtginode_ops {
 
 	unsigned int		sick;	/* rtgroup sickness flag */
 
+	unsigned int		fmt_mask; /* all valid data fork formats */
+
 	/* Does the fs have this feature? */
 	bool			(*enabled)(struct xfs_mount *mp);
 
@@ -330,14 +332,29 @@ static const struct xfs_rtginode_ops xfs_rtginode_ops[XFS_RTGI_MAX] = {
 		.name		= "bitmap",
 		.metafile_type	= XFS_METAFILE_RTBITMAP,
 		.sick		= XFS_SICK_RG_BITMAP,
+		.fmt_mask	= (1U << XFS_DINODE_FMT_EXTENTS) |
+				  (1U << XFS_DINODE_FMT_BTREE),
 		.create		= xfs_rtbitmap_create,
 	},
 	[XFS_RTGI_SUMMARY] = {
 		.name		= "summary",
 		.metafile_type	= XFS_METAFILE_RTSUMMARY,
 		.sick		= XFS_SICK_RG_SUMMARY,
+		.fmt_mask	= (1U << XFS_DINODE_FMT_EXTENTS) |
+				  (1U << XFS_DINODE_FMT_BTREE),
 		.create		= xfs_rtsummary_create,
 	},
+	[XFS_RTGI_RMAP] = {
+		.name		= "rmap",
+		.metafile_type	= XFS_METAFILE_RTRMAP,
+		.fmt_mask	= 1U << XFS_DINODE_FMT_META_BTREE,
+		/*
+		 * growfs must create the rtrmap inodes before adding a
+		 * realtime volume to the filesystem, so we cannot use the
+		 * rtrmapbt predicate here.
+		 */
+		.enabled	= xfs_has_rmapbt,
+	},
 };
 
 /* Return the shortname of this rtgroup inode. */
@@ -434,8 +451,7 @@ xfs_rtginode_load(
 		return error;
 	}
 
-	if (XFS_IS_CORRUPT(mp, ip->i_df.if_format != XFS_DINODE_FMT_EXTENTS &&
-			       ip->i_df.if_format != XFS_DINODE_FMT_BTREE)) {
+	if (XFS_IS_CORRUPT(mp, !((1U << ip->i_df.if_format) & ops->fmt_mask))) {
 		xfs_irele(ip);
 		xfs_rtginode_mark_sick(rtg, type);
 		return -EFSCORRUPTED;
diff --git a/fs/xfs/libxfs/xfs_rtgroup.h b/fs/xfs/libxfs/xfs_rtgroup.h
index dc3ce660a01307..5b61291d26691f 100644
--- a/fs/xfs/libxfs/xfs_rtgroup.h
+++ b/fs/xfs/libxfs/xfs_rtgroup.h
@@ -14,6 +14,7 @@ struct xfs_trans;
 enum xfs_rtg_inodes {
 	XFS_RTGI_BITMAP,	/* allocation bitmap */
 	XFS_RTGI_SUMMARY,	/* allocation summary */
+	XFS_RTGI_RMAP,		/* rmap btree inode */
 
 	XFS_RTGI_MAX,
 };
@@ -74,6 +75,11 @@ static inline struct xfs_inode *rtg_summary(const struct xfs_rtgroup *rtg)
 	return rtg->rtg_inodes[XFS_RTGI_SUMMARY];
 }
 
+static inline struct xfs_inode *rtg_rmap(const struct xfs_rtgroup *rtg)
+{
+	return rtg->rtg_inodes[XFS_RTGI_RMAP];
+}
+
 /* Passive rtgroup references */
 static inline struct xfs_rtgroup *
 xfs_rtgroup_get(
@@ -284,6 +290,8 @@ int xfs_rtginode_create(struct xfs_rtgroup *rtg, enum xfs_rtg_inodes type,
 		bool init);
 void xfs_rtginode_irele(struct xfs_inode **ipp);
 
+void xfs_rtginode_irele(struct xfs_inode **ipp);
+
 static inline const char *xfs_rtginode_path(xfs_rgnumber_t rgno,
 		enum xfs_rtg_inodes type)
 {
diff --git a/fs/xfs/libxfs/xfs_rtrmap_btree.c b/fs/xfs/libxfs/xfs_rtrmap_btree.c
index 99d828bb5fe7c3..22aabf326b2ccd 100644
--- a/fs/xfs/libxfs/xfs_rtrmap_btree.c
+++ b/fs/xfs/libxfs/xfs_rtrmap_btree.c
@@ -18,6 +18,7 @@
 #include "xfs_alloc.h"
 #include "xfs_btree.h"
 #include "xfs_btree_staging.h"
+#include "xfs_metafile.h"
 #include "xfs_rmap.h"
 #include "xfs_rtrmap_btree.h"
 #include "xfs_trace.h"
@@ -405,12 +406,10 @@ xfs_rtrmapbt_init_cursor(
 	struct xfs_trans	*tp,
 	struct xfs_rtgroup	*rtg)
 {
-	struct xfs_inode	*ip = NULL;
+	struct xfs_inode	*ip = rtg_rmap(rtg);
 	struct xfs_mount	*mp = rtg_mount(rtg);
 	struct xfs_btree_cur	*cur;
 
-	return NULL; /* XXX */
-
 	xfs_assert_ilocked(ip, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL);
 
 	cur = xfs_btree_alloc_cursor(mp, tp, &xfs_rtrmapbt_ops,
@@ -439,6 +438,7 @@ xfs_rtrmapbt_commit_staged_btree(
 	int			flags = XFS_ILOG_CORE | XFS_ILOG_DBROOT;
 
 	ASSERT(cur->bc_flags & XFS_BTREE_STAGING);
+	ASSERT(ifake->if_fork->if_format == XFS_DINODE_FMT_META_BTREE);
 
 	/*
 	 * Free any resources hanging off the real fork, then shallow-copy the

[PATCH 13/37] xfs: add metadata reservations for realtime rmap btrees

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Reserve some free blocks so that we will always have enough free blocks
in the data volume to handle expansion of the realtime rmap btree.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_rtrmap_btree.c |   41 ++++++++++++++++++++++++++++++++++++++
 fs/xfs/libxfs/xfs_rtrmap_btree.h |    2 ++
 fs/xfs/xfs_rtalloc.c             |   23 ++++++++++++++++++++-
 3 files changed, 65 insertions(+), 1 deletion(-)


diff --git a/fs/xfs/libxfs/xfs_rtrmap_btree.c b/fs/xfs/libxfs/xfs_rtrmap_btree.c
index 22aabf326b2ccd..08c4014a75a42c 100644
--- a/fs/xfs/libxfs/xfs_rtrmap_btree.c
+++ b/fs/xfs/libxfs/xfs_rtrmap_btree.c
@@ -540,3 +540,44 @@ xfs_rtrmapbt_compute_maxlevels(
 	/* Add one level to handle the inode root level. */
 	mp->m_rtrmap_maxlevels = min(d_maxlevels, r_maxlevels) + 1;
 }
+
+/* Calculate the rtrmap btree size for some records. */
+static unsigned long long
+xfs_rtrmapbt_calc_size(
+	struct xfs_mount	*mp,
+	unsigned long long	len)
+{
+	return xfs_btree_calc_size(mp->m_rtrmap_mnr, len);
+}
+
+/*
+ * Calculate the maximum rmap btree size.
+ */
+static unsigned long long
+xfs_rtrmapbt_max_size(
+	struct xfs_mount	*mp,
+	xfs_rtblock_t		rtblocks)
+{
+	/* Bail out if we're uninitialized, which can happen in mkfs. */
+	if (mp->m_rtrmap_mxr[0] == 0)
+		return 0;
+
+	return xfs_rtrmapbt_calc_size(mp, rtblocks);
+}
+
+/*
+ * Figure out how many blocks to reserve and how many are used by this btree.
+ */
+xfs_filblks_t
+xfs_rtrmapbt_calc_reserves(
+	struct xfs_mount	*mp)
+{
+	uint32_t		blocks = mp->m_groups[XG_TYPE_RTG].blocks;
+
+	if (!xfs_has_rtrmapbt(mp))
+		return 0;
+
+	/* 1/64th (~1.5%) of the space, and enough for 1 record per block. */
+	return max_t(xfs_filblks_t, blocks >> 6,
+			xfs_rtrmapbt_max_size(mp, blocks));
+}
diff --git a/fs/xfs/libxfs/xfs_rtrmap_btree.h b/fs/xfs/libxfs/xfs_rtrmap_btree.h
index 7d1a3a49a2d69b..eaa2942297e20c 100644
--- a/fs/xfs/libxfs/xfs_rtrmap_btree.h
+++ b/fs/xfs/libxfs/xfs_rtrmap_btree.h
@@ -79,4 +79,6 @@ unsigned int xfs_rtrmapbt_maxlevels_ondisk(void);
 int __init xfs_rtrmapbt_init_cur_cache(void);
 void xfs_rtrmapbt_destroy_cur_cache(void);
 
+xfs_filblks_t xfs_rtrmapbt_calc_reserves(struct xfs_mount *mp);
+
 #endif /* __XFS_RTRMAP_BTREE_H__ */
diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c
index 4cd2f32aa70a0a..2245f9ecaa3398 100644
--- a/fs/xfs/xfs_rtalloc.c
+++ b/fs/xfs/xfs_rtalloc.c
@@ -22,6 +22,7 @@
 #include "xfs_rtalloc.h"
 #include "xfs_sb.h"
 #include "xfs_rtbitmap.h"
+#include "xfs_rtrmap_btree.h"
 #include "xfs_quota.h"
 #include "xfs_log_priv.h"
 #include "xfs_health.h"
@@ -1498,6 +1499,13 @@ void
 xfs_rt_resv_free(
 	struct xfs_mount	*mp)
 {
+	struct xfs_rtgroup	*rtg = NULL;
+	unsigned int		i;
+
+	while ((rtg = xfs_rtgroup_next(mp, rtg))) {
+		for (i = 0; i < XFS_RTGI_MAX; i++)
+			xfs_metafile_resv_free(rtg->rtg_inodes[i]);
+	}
 }
 
 /* Reserve space for rt metadata inodes' space expansion. */
@@ -1505,7 +1513,20 @@ int
 xfs_rt_resv_init(
 	struct xfs_mount	*mp)
 {
-	return 0;
+	struct xfs_rtgroup	*rtg = NULL;
+	xfs_filblks_t		ask;
+	int			error = 0;
+
+	while ((rtg = xfs_rtgroup_next(mp, rtg))) {
+		int		err2;
+
+		ask = xfs_rtrmapbt_calc_reserves(mp);
+		err2 = xfs_metafile_resv_init(rtg_rmap(rtg), ask);
+		if (err2 && !error)
+			error = err2;
+	}
+
+	return error;
 }
 
 /*

[PATCH 14/37] xfs: wire up a new metafile type for the realtime rmap

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

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

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_format.h       |    8 +
 fs/xfs/libxfs/xfs_inode_fork.c   |    6 -
 fs/xfs/libxfs/xfs_ondisk.h       |    1 
 fs/xfs/libxfs/xfs_rtrmap_btree.c |  251 ++++++++++++++++++++++++++++++++++++++
 fs/xfs/libxfs/xfs_rtrmap_btree.h |  112 +++++++++++++++++
 fs/xfs/xfs_inode_item_recover.c  |    4 +
 6 files changed, 379 insertions(+), 3 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index f32c9fda5a195f..fba4e59aded4a0 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -1736,6 +1736,14 @@ 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 */
+};
+
 /* inode-based btree pointer type */
 typedef __be64 xfs_rtrmap_ptr_t;
 
diff --git a/fs/xfs/libxfs/xfs_inode_fork.c b/fs/xfs/libxfs/xfs_inode_fork.c
index 7c2b071a58384d..d9b3c182cb400b 100644
--- a/fs/xfs/libxfs/xfs_inode_fork.c
+++ b/fs/xfs/libxfs/xfs_inode_fork.c
@@ -27,6 +27,7 @@
 #include "xfs_errortag.h"
 #include "xfs_health.h"
 #include "xfs_symlink_remote.h"
+#include "xfs_rtrmap_btree.h"
 
 struct kmem_cache *xfs_ifork_cache;
 
@@ -270,8 +271,7 @@ xfs_iformat_data_fork(
 		case XFS_DINODE_FMT_META_BTREE:
 			switch (ip->i_metatype) {
 			case XFS_METAFILE_RTRMAP:
-				ASSERT(0); /* to be implemented later */
-				return -EFSCORRUPTED;
+				return xfs_iformat_rtrmap(ip, dip);
 			default:
 				break;
 			}
@@ -618,7 +618,7 @@ xfs_iflush_fork(
 
 		switch (ip->i_metatype) {
 		case XFS_METAFILE_RTRMAP:
-			ASSERT(0); /* to be implemented later */
+			xfs_iflush_rtrmap(ip, dip);
 			break;
 		default:
 			ASSERT(0);
diff --git a/fs/xfs/libxfs/xfs_ondisk.h b/fs/xfs/libxfs/xfs_ondisk.h
index 2c50877a1a2f0b..07e2f5fb3a94ae 100644
--- a/fs/xfs/libxfs/xfs_ondisk.h
+++ b/fs/xfs/libxfs/xfs_ondisk.h
@@ -84,6 +84,7 @@ xfs_check_ondisk_structs(void)
 	XFS_CHECK_STRUCT_SIZE(union xfs_suminfo_raw,		4);
 	XFS_CHECK_STRUCT_SIZE(struct xfs_rtbuf_blkinfo,		48);
 	XFS_CHECK_STRUCT_SIZE(xfs_rtrmap_ptr_t,			8);
+	XFS_CHECK_STRUCT_SIZE(struct xfs_rtrmap_root,		4);
 
 	/*
 	 * m68k has problems with struct xfs_attr_leaf_name_remote, but we pad
diff --git a/fs/xfs/libxfs/xfs_rtrmap_btree.c b/fs/xfs/libxfs/xfs_rtrmap_btree.c
index 08c4014a75a42c..d90189a1ef10b5 100644
--- a/fs/xfs/libxfs/xfs_rtrmap_btree.c
+++ b/fs/xfs/libxfs/xfs_rtrmap_btree.c
@@ -77,6 +77,39 @@ xfs_rtrmapbt_get_maxrecs(
 	return cur->bc_mp->m_rtrmap_mxr[level != 0];
 }
 
+/* Calculate number of records in the ondisk realtime rmap btree inode root. */
+unsigned int
+xfs_rtrmapbt_droot_maxrecs(
+	unsigned int		blocklen,
+	bool			leaf)
+{
+	blocklen -= sizeof(struct xfs_rtrmap_root);
+
+	if (leaf)
+		return blocklen / sizeof(struct xfs_rmap_rec);
+	return blocklen / (2 * sizeof(struct xfs_rmap_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_rtrmapbt_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_rtrmapbt_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_droot_maxrecs(cur->bc_ino.forksize, level == 0);
+}
+
 /*
  * Convert the ondisk record's offset field into the ondisk key's offset field.
  * Fork and bmbt are significant parts of the rmap record key, but written
@@ -369,6 +402,87 @@ xfs_rtrmapbt_keys_contiguous(
 				 be32_to_cpu(key2->rmap.rm_startblock));
 }
 
+static inline void
+xfs_rtrmapbt_move_ptrs(
+	struct xfs_mount	*mp,
+	struct xfs_btree_block	*broot,
+	short			old_size,
+	size_t			new_size,
+	unsigned int		numrecs)
+{
+	void			*dptr;
+	void			*sptr;
+
+	sptr = xfs_rtrmap_broot_ptr_addr(mp, broot, 1, old_size);
+	dptr = xfs_rtrmap_broot_ptr_addr(mp, broot, 1, new_size);
+	memmove(dptr, sptr, numrecs * sizeof(xfs_rtrmap_ptr_t));
+}
+
+static struct xfs_btree_block *
+xfs_rtrmapbt_broot_realloc(
+	struct xfs_btree_cur	*cur,
+	unsigned int		new_numrecs)
+{
+	struct xfs_mount	*mp = cur->bc_mp;
+	struct xfs_ifork	*ifp = xfs_btree_ifork_ptr(cur);
+	struct xfs_btree_block	*broot;
+	unsigned int		new_size;
+	unsigned int		old_size = ifp->if_broot_bytes;
+	const unsigned int	level = cur->bc_nlevels - 1;
+
+	new_size = xfs_rtrmap_broot_space_calc(mp, level, new_numrecs);
+
+	/* Handle the nop case quietly. */
+	if (new_size == old_size)
+		return ifp->if_broot;
+
+	if (new_size > old_size) {
+		unsigned int	old_numrecs;
+
+		/*
+		 * If there wasn't any memory allocated before, just allocate
+		 * it now and get out.
+		 */
+		if (old_size == 0)
+			return xfs_broot_realloc(ifp, new_size);
+
+		/*
+		 * If there is already an existing if_broot, then we need to
+		 * realloc it and possibly move the node block pointers because
+		 * those are not butted up against the btree block header.
+		 */
+		old_numrecs = xfs_rtrmapbt_maxrecs(mp, old_size, level == 0);
+		broot = xfs_broot_realloc(ifp, new_size);
+		if (level > 0)
+			xfs_rtrmapbt_move_ptrs(mp, broot, old_size, new_size,
+					old_numrecs);
+		goto out_broot;
+	}
+
+	/*
+	 * We're reducing numrecs.  If we're going all the way to zero, just
+	 * free the block.
+	 */
+	ASSERT(ifp->if_broot != NULL && old_size > 0);
+	if (new_size == 0)
+		return xfs_broot_realloc(ifp, 0);
+
+	/*
+	 * Shrink the btree root by possibly moving the rtrmapbt pointers,
+	 * since they are not butted up against the btree block header.  Then
+	 * reallocate broot.
+	 */
+	if (level > 0)
+		xfs_rtrmapbt_move_ptrs(mp, ifp->if_broot, old_size, new_size,
+				new_numrecs);
+	broot = xfs_broot_realloc(ifp, new_size);
+
+out_broot:
+	ASSERT(xfs_rtrmap_droot_space(broot) <=
+	       xfs_inode_fork_size(cur->bc_ino.ip, cur->bc_ino.whichfork));
+	return broot;
+}
+
 const struct xfs_btree_ops xfs_rtrmapbt_ops = {
 	.name			= "rtrmap",
 	.type			= XFS_BTREE_TYPE_INODE,
@@ -388,6 +502,7 @@ const struct xfs_btree_ops xfs_rtrmapbt_ops = {
 	.free_block		= xfs_btree_free_metafile_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,
@@ -398,6 +513,7 @@ const struct xfs_btree_ops xfs_rtrmapbt_ops = {
 	.keys_inorder		= xfs_rtrmapbt_keys_inorder,
 	.recs_inorder		= xfs_rtrmapbt_recs_inorder,
 	.keys_contiguous	= xfs_rtrmapbt_keys_contiguous,
+	.broot_realloc		= xfs_rtrmapbt_broot_realloc,
 };
 
 /* Allocate a new rt rmap btree cursor. */
@@ -581,3 +697,138 @@ xfs_rtrmapbt_calc_reserves(
 	return max_t(xfs_filblks_t, blocks >> 6,
 			xfs_rtrmapbt_max_size(mp, blocks));
 }
+
+/* Convert on-disk form of btree root to in-memory form. */
+STATIC void
+xfs_rtrmapbt_from_disk(
+	struct xfs_inode	*ip,
+	struct xfs_rtrmap_root	*dblock,
+	unsigned int		dblocklen,
+	struct xfs_btree_block	*rblock)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_rmap_key	*fkp;
+	__be64			*fpp;
+	struct xfs_rmap_key	*tkp;
+	__be64			*tpp;
+	struct xfs_rmap_rec	*frp;
+	struct xfs_rmap_rec	*trp;
+	unsigned int		rblocklen = xfs_rtrmap_broot_space(mp, dblock);
+	unsigned int		numrecs;
+	unsigned int		maxrecs;
+
+	xfs_btree_init_block(mp, rblock, &xfs_rtrmapbt_ops, 0, 0, ip->i_ino);
+
+	rblock->bb_level = dblock->bb_level;
+	rblock->bb_numrecs = dblock->bb_numrecs;
+	numrecs = be16_to_cpu(dblock->bb_numrecs);
+
+	if (be16_to_cpu(rblock->bb_level) > 0) {
+		maxrecs = xfs_rtrmapbt_droot_maxrecs(dblocklen, false);
+		fkp = xfs_rtrmap_droot_key_addr(dblock, 1);
+		tkp = xfs_rtrmap_key_addr(rblock, 1);
+		fpp = xfs_rtrmap_droot_ptr_addr(dblock, 1, maxrecs);
+		tpp = xfs_rtrmap_broot_ptr_addr(mp, rblock, 1, rblocklen);
+		memcpy(tkp, fkp, 2 * sizeof(*fkp) * numrecs);
+		memcpy(tpp, fpp, sizeof(*fpp) * numrecs);
+	} else {
+		frp = xfs_rtrmap_droot_rec_addr(dblock, 1);
+		trp = xfs_rtrmap_rec_addr(rblock, 1);
+		memcpy(trp, frp, sizeof(*frp) * numrecs);
+	}
+}
+
+/* Load a realtime reverse mapping btree root in from disk. */
+int
+xfs_iformat_rtrmap(
+	struct xfs_inode	*ip,
+	struct xfs_dinode	*dip)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_rtrmap_root	*dfp = XFS_DFORK_PTR(dip, XFS_DATA_FORK);
+	struct xfs_btree_block	*broot;
+	unsigned int		numrecs;
+	unsigned int		level;
+	int			dsize;
+
+	/*
+	 * growfs must create the rtrmap inodes before adding a realtime volume
+	 * to the filesystem, so we cannot use the rtrmapbt predicate here.
+	 */
+	if (!xfs_has_rmapbt(ip->i_mount))
+		return -EFSCORRUPTED;
+
+	dsize = XFS_DFORK_SIZE(dip, mp, XFS_DATA_FORK);
+	numrecs = be16_to_cpu(dfp->bb_numrecs);
+	level = be16_to_cpu(dfp->bb_level);
+
+	if (level > mp->m_rtrmap_maxlevels ||
+	    xfs_rtrmap_droot_space_calc(level, numrecs) > dsize)
+		return -EFSCORRUPTED;
+
+	broot = xfs_broot_alloc(xfs_ifork_ptr(ip, XFS_DATA_FORK),
+			xfs_rtrmap_broot_space_calc(mp, level, numrecs));
+	if (broot)
+		xfs_rtrmapbt_from_disk(ip, dfp, dsize, broot);
+	return 0;
+}
+
+/* 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,
+	unsigned int		rblocklen,
+	struct xfs_rtrmap_root	*dblock,
+	unsigned int		dblocklen)
+{
+	struct xfs_rmap_key	*fkp;
+	__be64			*fpp;
+	struct xfs_rmap_key	*tkp;
+	__be64			*tpp;
+	struct xfs_rmap_rec	*frp;
+	struct xfs_rmap_rec	*trp;
+	unsigned int		numrecs;
+	unsigned int		maxrecs;
+
+	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;
+	numrecs = be16_to_cpu(rblock->bb_numrecs);
+
+	if (be16_to_cpu(rblock->bb_level) > 0) {
+		maxrecs = xfs_rtrmapbt_droot_maxrecs(dblocklen, false);
+		fkp = xfs_rtrmap_key_addr(rblock, 1);
+		tkp = xfs_rtrmap_droot_key_addr(dblock, 1);
+		fpp = xfs_rtrmap_broot_ptr_addr(mp, rblock, 1, rblocklen);
+		tpp = xfs_rtrmap_droot_ptr_addr(dblock, 1, maxrecs);
+		memcpy(tkp, fkp, 2 * sizeof(*fkp) * numrecs);
+		memcpy(tpp, fpp, sizeof(*fpp) * numrecs);
+	} else {
+		frp = xfs_rtrmap_rec_addr(rblock, 1);
+		trp = xfs_rtrmap_droot_rec_addr(dblock, 1);
+		memcpy(trp, frp, sizeof(*frp) * numrecs);
+	}
+}
+
+/* Flush a realtime reverse mapping btree root out to disk. */
+void
+xfs_iflush_rtrmap(
+	struct xfs_inode	*ip,
+	struct xfs_dinode	*dip)
+{
+	struct xfs_ifork	*ifp = xfs_ifork_ptr(ip, XFS_DATA_FORK);
+	struct xfs_rtrmap_root	*dfp = XFS_DFORK_PTR(dip, XFS_DATA_FORK);
+
+	ASSERT(ifp->if_broot != NULL);
+	ASSERT(ifp->if_broot_bytes > 0);
+	ASSERT(xfs_rtrmap_droot_space(ifp->if_broot) <=
+			xfs_inode_fork_size(ip, XFS_DATA_FORK));
+	xfs_rtrmapbt_to_disk(ip->i_mount, ifp->if_broot, ifp->if_broot_bytes,
+			dfp, XFS_DFORK_SIZE(dip, ip->i_mount, XFS_DATA_FORK));
+}
diff --git a/fs/xfs/libxfs/xfs_rtrmap_btree.h b/fs/xfs/libxfs/xfs_rtrmap_btree.h
index eaa2942297e20c..e97695066920ee 100644
--- a/fs/xfs/libxfs/xfs_rtrmap_btree.h
+++ b/fs/xfs/libxfs/xfs_rtrmap_btree.h
@@ -25,6 +25,7 @@ void xfs_rtrmapbt_commit_staged_btree(struct xfs_btree_cur *cur,
 unsigned int xfs_rtrmapbt_maxrecs(struct xfs_mount *mp, unsigned int blocklen,
 		bool leaf);
 void xfs_rtrmapbt_compute_maxlevels(struct xfs_mount *mp);
+unsigned int xfs_rtrmapbt_droot_maxrecs(unsigned int blocklen, bool leaf);
 
 /*
  * Addresses of records, keys, and pointers within an incore rtrmapbt block.
@@ -81,4 +82,115 @@ void xfs_rtrmapbt_destroy_cur_cache(void);
 
 xfs_filblks_t xfs_rtrmapbt_calc_reserves(struct xfs_mount *mp);
 
+/* Addresses of key, pointers, and records within an ondisk rtrmapbt block. */
+
+static inline struct xfs_rmap_rec *
+xfs_rtrmap_droot_rec_addr(
+	struct xfs_rtrmap_root	*block,
+	unsigned int		index)
+{
+	return (struct xfs_rmap_rec *)
+		((char *)(block + 1) +
+		 (index - 1) * sizeof(struct xfs_rmap_rec));
+}
+
+static inline struct xfs_rmap_key *
+xfs_rtrmap_droot_key_addr(
+	struct xfs_rtrmap_root	*block,
+	unsigned int		index)
+{
+	return (struct xfs_rmap_key *)
+		((char *)(block + 1) +
+		 (index - 1) * 2 * sizeof(struct xfs_rmap_key));
+}
+
+static inline xfs_rtrmap_ptr_t *
+xfs_rtrmap_droot_ptr_addr(
+	struct xfs_rtrmap_root	*block,
+	unsigned int		index,
+	unsigned int		maxrecs)
+{
+	return (xfs_rtrmap_ptr_t *)
+		((char *)(block + 1) +
+		 maxrecs * 2 * sizeof(struct xfs_rmap_key) +
+		 (index - 1) * sizeof(xfs_rtrmap_ptr_t));
+}
+
+/*
+ * Address of pointers within the incore btree root.
+ *
+ * These are to be used when we know the size of the block and
+ * we don't have a cursor.
+ */
+static inline xfs_rtrmap_ptr_t *
+xfs_rtrmap_broot_ptr_addr(
+	struct xfs_mount	*mp,
+	struct xfs_btree_block	*bb,
+	unsigned int		index,
+	unsigned int		block_size)
+{
+	return xfs_rtrmap_ptr_addr(bb, index,
+			xfs_rtrmapbt_maxrecs(mp, block_size, false));
+}
+
+/*
+ * Compute the space required for the incore btree root containing the given
+ * number of records.
+ */
+static inline size_t
+xfs_rtrmap_broot_space_calc(
+	struct xfs_mount	*mp,
+	unsigned int		level,
+	unsigned int		nrecs)
+{
+	size_t			sz = XFS_RTRMAP_BLOCK_LEN;
+
+	if (level > 0)
+		return sz + nrecs * (2 * sizeof(struct xfs_rmap_key) +
+					 sizeof(xfs_rtrmap_ptr_t));
+	return sz + nrecs * sizeof(struct xfs_rmap_rec);
+}
+
+/*
+ * Compute the space required for the incore btree root given the ondisk
+ * btree root block.
+ */
+static inline size_t
+xfs_rtrmap_broot_space(struct xfs_mount *mp, struct xfs_rtrmap_root *bb)
+{
+	return xfs_rtrmap_broot_space_calc(mp, be16_to_cpu(bb->bb_level),
+			be16_to_cpu(bb->bb_numrecs));
+}
+
+/* Compute the space required for the ondisk root block. */
+static inline size_t
+xfs_rtrmap_droot_space_calc(
+	unsigned int		level,
+	unsigned int		nrecs)
+{
+	size_t			sz = sizeof(struct xfs_rtrmap_root);
+
+	if (level > 0)
+		return sz + nrecs * (2 * sizeof(struct xfs_rmap_key) +
+					 sizeof(xfs_rtrmap_ptr_t));
+	return sz + nrecs * sizeof(struct xfs_rmap_rec);
+}
+
+/*
+ * Compute the space required for the ondisk root block given an incore root
+ * block.
+ */
+static inline size_t
+xfs_rtrmap_droot_space(struct xfs_btree_block *bb)
+{
+	return xfs_rtrmap_droot_space_calc(be16_to_cpu(bb->bb_level),
+			be16_to_cpu(bb->bb_numrecs));
+}
+
+int xfs_iformat_rtrmap(struct xfs_inode *ip, struct xfs_dinode *dip);
+void xfs_rtrmapbt_to_disk(struct xfs_mount *mp, struct xfs_btree_block *rblock,
+		unsigned int rblocklen, struct xfs_rtrmap_root *dblock,
+		unsigned int dblocklen);
+void xfs_iflush_rtrmap(struct xfs_inode *ip, struct xfs_dinode *dip);
+
 #endif /* __XFS_RTRMAP_BTREE_H__ */
diff --git a/fs/xfs/xfs_inode_item_recover.c b/fs/xfs/xfs_inode_item_recover.c
index 6e9b3bfc718c0b..5de1d3563b7686 100644
--- a/fs/xfs/xfs_inode_item_recover.c
+++ b/fs/xfs/xfs_inode_item_recover.c
@@ -22,6 +22,7 @@
 #include "xfs_log_recover.h"
 #include "xfs_icache.h"
 #include "xfs_bmap_btree.h"
+#include "xfs_rtrmap_btree.h"
 
 STATIC void
 xlog_recover_inode_ra_pass2(
@@ -282,6 +283,9 @@ xlog_recover_inode_dbroot(
 		break;
 	case XFS_DINODE_FMT_META_BTREE:
 		switch (be16_to_cpu(dip->di_metatype)) {
+		case XFS_METAFILE_RTRMAP:
+			xfs_rtrmapbt_to_disk(mp, src, len, dfork, dsize);
+			return 0;
 		default:
 			ASSERT(0);
 			return -EFSCORRUPTED;

[PATCH 15/37] xfs: wire up rmap map and unmap to the realtime rmapbt

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

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" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_rmap.c    |   78 +++++++++++++++++++++++++++++--------------
 fs/xfs/libxfs/xfs_rtgroup.c |    9 +++++
 fs/xfs/libxfs/xfs_rtgroup.h |    5 ++-
 3 files changed, 66 insertions(+), 26 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c
index 8d3cea90c7cd04..2f0688a57991cc 100644
--- a/fs/xfs/libxfs/xfs_rmap.c
+++ b/fs/xfs/libxfs/xfs_rmap.c
@@ -26,6 +26,7 @@
 #include "xfs_health.h"
 #include "xfs_rmap_item.h"
 #include "xfs_rtgroup.h"
+#include "xfs_rtrmap_btree.h"
 
 struct kmem_cache	*xfs_rmap_intent_cache;
 
@@ -2619,6 +2620,47 @@ __xfs_rmap_finish_intent(
 	}
 }
 
+static int
+xfs_rmap_finish_init_cursor(
+	struct xfs_trans		*tp,
+	struct xfs_rmap_intent		*ri,
+	struct xfs_btree_cur		**pcur)
+{
+	struct xfs_perag		*pag = to_perag(ri->ri_group);
+	struct xfs_buf			*agbp = NULL;
+	int				error;
+
+	/*
+	 * 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, pag, &agbp);
+	if (error) {
+		xfs_ag_mark_sick(pag, XFS_SICK_AG_AGFL);
+		return error;
+	}
+	if (XFS_IS_CORRUPT(tp->t_mountp, !agbp)) {
+		xfs_ag_mark_sick(pag, XFS_SICK_AG_AGFL);
+		return -EFSCORRUPTED;
+	}
+	*pcur = xfs_rmapbt_init_cursor(tp->t_mountp, tp, agbp, pag);
+	return 0;
+}
+
+static int
+xfs_rtrmap_finish_init_cursor(
+	struct xfs_trans		*tp,
+	struct xfs_rmap_intent		*ri,
+	struct xfs_btree_cur		**pcur)
+{
+	struct xfs_rtgroup		*rtg = to_rtg(ri->ri_group);
+
+	xfs_rtgroup_lock(rtg, XFS_RTGLOCK_RMAP);
+	xfs_rtgroup_trans_join(tp, rtg, XFS_RTGLOCK_RMAP);
+	*pcur = xfs_rtrmapbt_init_cursor(tp, rtg);
+	return 0;
+}
+
 /*
  * Process one of the deferred rmap operations.  We pass back the
  * btree cursor to maintain our lock on the rmapbt between calls.
@@ -2634,8 +2676,6 @@ xfs_rmap_finish_one(
 {
 	struct xfs_owner_info		oinfo;
 	struct xfs_mount		*mp = tp->t_mountp;
-	struct xfs_btree_cur		*rcur = *pcur;
-	struct xfs_buf			*agbp = NULL;
 	xfs_agblock_t			bno;
 	bool				unwritten;
 	int				error = 0;
@@ -2649,38 +2689,26 @@ xfs_rmap_finish_one(
 	 * If we haven't gotten a cursor or the cursor AG doesn't match
 	 * the startblock, get one now.
 	 */
-	if (rcur != NULL && rcur->bc_group != ri->ri_group) {
-		xfs_btree_del_cursor(rcur, 0);
-		rcur = NULL;
+	if (*pcur != NULL && (*pcur)->bc_group != ri->ri_group) {
+		xfs_btree_del_cursor(*pcur, 0);
 		*pcur = NULL;
 	}
-	if (rcur == NULL) {
-		struct xfs_perag	*pag = to_perag(ri->ri_group);
-
-		/*
-		 * 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, pag, &agbp);
-		if (error) {
-			xfs_ag_mark_sick(pag, XFS_SICK_AG_AGFL);
+	if (*pcur == NULL) {
+		if (ri->ri_group->xg_type == XG_TYPE_RTG)
+			error = xfs_rtrmap_finish_init_cursor(tp, ri, pcur);
+		else
+			error = xfs_rmap_finish_init_cursor(tp, ri, pcur);
+		if (error)
 			return error;
-		}
-		if (XFS_IS_CORRUPT(tp->t_mountp, !agbp)) {
-			xfs_ag_mark_sick(pag, XFS_SICK_AG_AGFL);
-			return -EFSCORRUPTED;
-		}
-
-		*pcur = rcur = xfs_rmapbt_init_cursor(mp, tp, agbp, pag);
 	}
 
 	xfs_rmap_ino_owner(&oinfo, ri->ri_owner, ri->ri_whichfork,
 			ri->ri_bmap.br_startoff);
 	unwritten = ri->ri_bmap.br_state == XFS_EXT_UNWRITTEN;
-	bno = XFS_FSB_TO_AGBNO(rcur->bc_mp, ri->ri_bmap.br_startblock);
 
-	error = __xfs_rmap_finish_intent(rcur, ri->ri_type, bno,
+	bno = xfs_fsb_to_gbno(mp, ri->ri_bmap.br_startblock,
+			ri->ri_group->xg_type);
+	error = __xfs_rmap_finish_intent(*pcur, ri->ri_type, bno,
 			ri->ri_bmap.br_blockcount, &oinfo, unwritten);
 	if (error)
 		return error;
diff --git a/fs/xfs/libxfs/xfs_rtgroup.c b/fs/xfs/libxfs/xfs_rtgroup.c
index 1b56c13b282788..af1716ec0691a4 100644
--- a/fs/xfs/libxfs/xfs_rtgroup.c
+++ b/fs/xfs/libxfs/xfs_rtgroup.c
@@ -202,6 +202,9 @@ xfs_rtgroup_lock(
 	} else if (rtglock_flags & XFS_RTGLOCK_BITMAP_SHARED) {
 		xfs_ilock(rtg_bitmap(rtg), XFS_ILOCK_SHARED);
 	}
+
+	if ((rtglock_flags & XFS_RTGLOCK_RMAP) && rtg_rmap(rtg))
+		xfs_ilock(rtg_rmap(rtg), XFS_ILOCK_EXCL);
 }
 
 /* Unlock metadata inodes associated with this rt group. */
@@ -214,6 +217,9 @@ xfs_rtgroup_unlock(
 	ASSERT(!(rtglock_flags & XFS_RTGLOCK_BITMAP_SHARED) ||
 	       !(rtglock_flags & XFS_RTGLOCK_BITMAP));
 
+	if ((rtglock_flags & XFS_RTGLOCK_RMAP) && rtg_rmap(rtg))
+		xfs_iunlock(rtg_rmap(rtg), XFS_ILOCK_EXCL);
+
 	if (rtglock_flags & XFS_RTGLOCK_BITMAP) {
 		xfs_iunlock(rtg_summary(rtg), XFS_ILOCK_EXCL);
 		xfs_iunlock(rtg_bitmap(rtg), XFS_ILOCK_EXCL);
@@ -239,6 +245,9 @@ xfs_rtgroup_trans_join(
 		xfs_trans_ijoin(tp, rtg_bitmap(rtg), XFS_ILOCK_EXCL);
 		xfs_trans_ijoin(tp, rtg_summary(rtg), XFS_ILOCK_EXCL);
 	}
+
+	if ((rtglock_flags & XFS_RTGLOCK_RMAP) && rtg_rmap(rtg))
+		xfs_trans_ijoin(tp, rtg_rmap(rtg), XFS_ILOCK_EXCL);
 }
 
 /* Retrieve rt group geometry. */
diff --git a/fs/xfs/libxfs/xfs_rtgroup.h b/fs/xfs/libxfs/xfs_rtgroup.h
index 5b61291d26691f..733da7417c9cd7 100644
--- a/fs/xfs/libxfs/xfs_rtgroup.h
+++ b/fs/xfs/libxfs/xfs_rtgroup.h
@@ -265,9 +265,12 @@ int xfs_update_last_rtgroup_size(struct xfs_mount *mp,
 #define XFS_RTGLOCK_BITMAP		(1U << 0)
 /* Lock the rt bitmap inode in shared mode */
 #define XFS_RTGLOCK_BITMAP_SHARED	(1U << 1)
+/* Lock the rt rmap inode in exclusive mode */
+#define XFS_RTGLOCK_RMAP		(1U << 2)
 
 #define XFS_RTGLOCK_ALL_FLAGS	(XFS_RTGLOCK_BITMAP | \
-				 XFS_RTGLOCK_BITMAP_SHARED)
+				 XFS_RTGLOCK_BITMAP_SHARED | \
+				 XFS_RTGLOCK_RMAP)
 
 void xfs_rtgroup_lock(struct xfs_rtgroup *rtg, unsigned int rtglock_flags);
 void xfs_rtgroup_unlock(struct xfs_rtgroup *rtg, unsigned int rtglock_flags);

[PATCH 16/37] xfs: create routine to allocate and initialize a realtime rmap btree inode

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Create a library routine to allocate and initialize an empty realtime
rmapbt inode.  We'll use this for mkfs and repair.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_rtgroup.c      |    2 +
 fs/xfs/libxfs/xfs_rtrmap_btree.c |   54 ++++++++++++++++++++++++++++++++++++++
 fs/xfs/libxfs/xfs_rtrmap_btree.h |    5 ++++
 fs/xfs/xfs_rtalloc.c             |   12 +++++++-
 4 files changed, 71 insertions(+), 2 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_rtgroup.c b/fs/xfs/libxfs/xfs_rtgroup.c
index af1716ec0691a4..5f31b6e65d5d17 100644
--- a/fs/xfs/libxfs/xfs_rtgroup.c
+++ b/fs/xfs/libxfs/xfs_rtgroup.c
@@ -33,6 +33,7 @@
 #include "xfs_rtbitmap.h"
 #include "xfs_metafile.h"
 #include "xfs_metadir.h"
+#include "xfs_rtrmap_btree.h"
 
 /* Find the first usable fsblock in this rtgroup. */
 static inline uint32_t
@@ -363,6 +364,7 @@ static const struct xfs_rtginode_ops xfs_rtginode_ops[XFS_RTGI_MAX] = {
 		 * rtrmapbt predicate here.
 		 */
 		.enabled	= xfs_has_rmapbt,
+		.create		= xfs_rtrmapbt_create,
 	},
 };
 
diff --git a/fs/xfs/libxfs/xfs_rtrmap_btree.c b/fs/xfs/libxfs/xfs_rtrmap_btree.c
index d90189a1ef10b5..7654661f4f5823 100644
--- a/fs/xfs/libxfs/xfs_rtrmap_btree.c
+++ b/fs/xfs/libxfs/xfs_rtrmap_btree.c
@@ -832,3 +832,57 @@ xfs_iflush_rtrmap(
 	xfs_rtrmapbt_to_disk(ip->i_mount, ifp->if_broot, ifp->if_broot_bytes,
 			dfp, XFS_DFORK_SIZE(dip, ip->i_mount, XFS_DATA_FORK));
 }
+
+/*
+ * Create a realtime rmap btree inode.
+ */
+int
+xfs_rtrmapbt_create(
+	struct xfs_rtgroup	*rtg,
+	struct xfs_inode	*ip,
+	struct xfs_trans	*tp,
+	bool			init)
+{
+	struct xfs_ifork	*ifp = xfs_ifork_ptr(ip, XFS_DATA_FORK);
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_btree_block	*broot;
+
+	ifp->if_format = XFS_DINODE_FMT_META_BTREE;
+	ASSERT(ifp->if_broot_bytes == 0);
+	ASSERT(ifp->if_bytes == 0);
+
+	/* Initialize the empty incore btree root. */
+	broot = xfs_broot_realloc(ifp, xfs_rtrmap_broot_space_calc(mp, 0, 0));
+	if (broot)
+		xfs_btree_init_block(mp, broot, &xfs_rtrmapbt_ops, 0, 0,
+				ip->i_ino);
+	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE | XFS_ILOG_DBROOT);
+
+	return 0;
+}
+
+/*
+ * Initialize an rmap for a realtime superblock using the potentially updated
+ * rt geometry in the provided @mp.
+ */
+int
+xfs_rtrmapbt_init_rtsb(
+	struct xfs_mount	*mp,
+	struct xfs_rtgroup	*rtg,
+	struct xfs_trans	*tp)
+{
+	struct xfs_rmap_irec	rmap = {
+		.rm_blockcount	= mp->m_sb.sb_rextsize,
+		.rm_owner	= XFS_RMAP_OWN_FS,
+	};
+	struct xfs_btree_cur	*cur;
+	int			error;
+
+	ASSERT(xfs_has_rtsb(mp));
+	ASSERT(rtg_rgno(rtg) == 0);
+
+	cur = xfs_rtrmapbt_init_cursor(tp, rtg);
+	error = xfs_rmap_map_raw(cur, &rmap);
+	xfs_btree_del_cursor(cur, error);
+	return error;
+}
diff --git a/fs/xfs/libxfs/xfs_rtrmap_btree.h b/fs/xfs/libxfs/xfs_rtrmap_btree.h
index e97695066920ee..bf73460be274d1 100644
--- a/fs/xfs/libxfs/xfs_rtrmap_btree.h
+++ b/fs/xfs/libxfs/xfs_rtrmap_btree.h
@@ -193,4 +193,9 @@ void xfs_rtrmapbt_to_disk(struct xfs_mount *mp, struct xfs_btree_block *rblock,
 		unsigned int dblocklen);
 void xfs_iflush_rtrmap(struct xfs_inode *ip, struct xfs_dinode *dip);
 
+int xfs_rtrmapbt_create(struct xfs_rtgroup *rtg, struct xfs_inode *ip,
+		struct xfs_trans *tp, bool init);
+int xfs_rtrmapbt_init_rtsb(struct xfs_mount *mp, struct xfs_rtgroup *rtg,
+		struct xfs_trans *tp);
+
 #endif /* __XFS_RTRMAP_BTREE_H__ */
diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c
index 2245f9ecaa3398..c7efd926413981 100644
--- a/fs/xfs/xfs_rtalloc.c
+++ b/fs/xfs/xfs_rtalloc.c
@@ -846,6 +846,13 @@ xfs_growfs_rt_init_rtsb(
 	mp->m_rtsb_bp = rtsb_bp;
 	error = xfs_bwrite(rtsb_bp);
 	xfs_buf_unlock(rtsb_bp);
+	if (error)
+		return error;
+
+	/* Initialize the rtrmap to reflect the rtsb. */
+	if (rtg_rmap(args->rtg) != NULL)
+		error = xfs_rtrmapbt_init_rtsb(nargs->mp, args->rtg, args->tp);
+
 	return error;
 }
 
@@ -894,8 +901,9 @@ xfs_growfs_rt_bmblock(
 		goto out_free;
 	nargs.tp = args.tp;
 
-	xfs_rtgroup_lock(args.rtg, XFS_RTGLOCK_BITMAP);
-	xfs_rtgroup_trans_join(args.tp, args.rtg, XFS_RTGLOCK_BITMAP);
+	xfs_rtgroup_lock(args.rtg, XFS_RTGLOCK_BITMAP | XFS_RTGLOCK_RMAP);
+	xfs_rtgroup_trans_join(args.tp, args.rtg,
+			XFS_RTGLOCK_BITMAP | XFS_RTGLOCK_RMAP);
 
 	/*
 	 * Update the bitmap inode's size ondisk and incore.  We need to update

[PATCH 17/37] xfs: wire up getfsmap to the realtime reverse mapping btree

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Connect the getfsmap ioctl to the realtime rmapbt.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/xfs_fsmap.c |  174 ++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 173 insertions(+), 1 deletion(-)


diff --git a/fs/xfs/xfs_fsmap.c b/fs/xfs/xfs_fsmap.c
index 3290dd8524a69a..3e3ef16f65a335 100644
--- a/fs/xfs/xfs_fsmap.c
+++ b/fs/xfs/xfs_fsmap.c
@@ -26,6 +26,7 @@
 #include "xfs_rtbitmap.h"
 #include "xfs_ag.h"
 #include "xfs_rtgroup.h"
+#include "xfs_rtrmap_btree.h"
 
 /* Convert an xfs_fsmap to an fsmap. */
 static void
@@ -832,6 +833,174 @@ xfs_getfsmap_rtdev_rtbitmap(
 
 	return error;
 }
+
+/* Transform a realtime rmapbt record into a fsmap */
+STATIC int
+xfs_getfsmap_rtdev_rmapbt_helper(
+	struct xfs_btree_cur		*cur,
+	const struct xfs_rmap_irec	*rec,
+	void				*priv)
+{
+	struct xfs_fsmap_irec		frec = {
+		.owner			= rec->rm_owner,
+		.offset			= rec->rm_offset,
+		.rm_flags		= rec->rm_flags,
+		.rec_key		= rec->rm_startblock,
+	};
+	struct xfs_getfsmap_info	*info = priv;
+
+	return xfs_getfsmap_group_helper(info, cur->bc_tp, cur->bc_group,
+			rec->rm_startblock, rec->rm_blockcount, &frec);
+}
+
+/* Actually query the rtrmap btree. */
+STATIC int
+xfs_getfsmap_rtdev_rmapbt_query(
+	struct xfs_trans		*tp,
+	struct xfs_getfsmap_info	*info,
+	struct xfs_btree_cur		**curpp)
+{
+	struct xfs_rtgroup		*rtg = to_rtg(info->group);
+
+	/* Query the rtrmapbt */
+	xfs_rtgroup_lock(rtg, XFS_RTGLOCK_RMAP);
+	*curpp = xfs_rtrmapbt_init_cursor(tp, rtg);
+	return xfs_rmap_query_range(*curpp, &info->low, &info->high,
+			xfs_getfsmap_rtdev_rmapbt_helper, info);
+}
+
+/* Execute a getfsmap query against the realtime device rmapbt. */
+STATIC int
+xfs_getfsmap_rtdev_rmapbt(
+	struct xfs_trans		*tp,
+	const struct xfs_fsmap		*keys,
+	struct xfs_getfsmap_info	*info)
+{
+	struct xfs_mount		*mp = tp->t_mountp;
+	struct xfs_rtgroup		*rtg = NULL;
+	struct xfs_btree_cur		*bt_cur = NULL;
+	xfs_rtblock_t			start_rtb;
+	xfs_rtblock_t			end_rtb;
+	xfs_rgnumber_t			start_rg, end_rg;
+	uint64_t			eofs;
+	int				error = 0;
+
+	eofs = XFS_FSB_TO_BB(mp, mp->m_sb.sb_rblocks);
+	if (keys[0].fmr_physical >= eofs)
+		return 0;
+	start_rtb = xfs_daddr_to_rtb(mp, keys[0].fmr_physical);
+	end_rtb = xfs_daddr_to_rtb(mp, min(eofs - 1, keys[1].fmr_physical));
+
+	info->missing_owner = XFS_FMR_OWN_FREE;
+
+	/*
+	 * Convert the fsmap low/high keys to rtgroup based keys.  Initialize
+	 * low to the fsmap low key and max out the high key to the end
+	 * of the rtgroup.
+	 */
+	info->low.rm_offset = XFS_BB_TO_FSBT(mp, keys[0].fmr_offset);
+	error = xfs_fsmap_owner_to_rmap(&info->low, &keys[0]);
+	if (error)
+		return error;
+	info->low.rm_blockcount = XFS_BB_TO_FSBT(mp, keys[0].fmr_length);
+	xfs_getfsmap_set_irec_flags(&info->low, &keys[0]);
+
+	/* Adjust the low key if we are continuing from where we left off. */
+	if (info->low.rm_blockcount == 0) {
+		/* No previous record from which to continue */
+	} else if (rmap_not_shareable(mp, &info->low)) {
+		/* Last record seen was an unshareable extent */
+		info->low.rm_owner = 0;
+		info->low.rm_offset = 0;
+
+		start_rtb += info->low.rm_blockcount;
+		if (xfs_rtb_to_daddr(mp, start_rtb) >= eofs)
+			return 0;
+	} else {
+		/* Last record seen was a shareable file data extent */
+		info->low.rm_offset += info->low.rm_blockcount;
+	}
+	info->low.rm_startblock = xfs_rtb_to_rgbno(mp, start_rtb);
+
+	info->high.rm_startblock = -1U;
+	info->high.rm_owner = ULLONG_MAX;
+	info->high.rm_offset = ULLONG_MAX;
+	info->high.rm_blockcount = 0;
+	info->high.rm_flags = XFS_RMAP_KEY_FLAGS | XFS_RMAP_REC_FLAGS;
+
+	start_rg = xfs_rtb_to_rgno(mp, start_rtb);
+	end_rg = xfs_rtb_to_rgno(mp, end_rtb);
+
+	while ((rtg = xfs_rtgroup_next_range(mp, rtg, start_rg, end_rg))) {
+		/*
+		 * Set the rtgroup high key from the fsmap high key if this
+		 * is the last rtgroup that we're querying.
+		 */
+		info->group = rtg_group(rtg);
+		if (rtg_rgno(rtg) == end_rg) {
+			info->high.rm_startblock =
+				xfs_rtb_to_rgbno(mp, end_rtb);
+			info->high.rm_offset =
+				XFS_BB_TO_FSBT(mp, keys[1].fmr_offset);
+			error = xfs_fsmap_owner_to_rmap(&info->high, &keys[1]);
+			if (error)
+				break;
+			xfs_getfsmap_set_irec_flags(&info->high, &keys[1]);
+		}
+
+		if (bt_cur) {
+			xfs_rtgroup_unlock(to_rtg(bt_cur->bc_group),
+					XFS_RTGLOCK_RMAP);
+			xfs_btree_del_cursor(bt_cur, XFS_BTREE_NOERROR);
+			bt_cur = NULL;
+		}
+
+		trace_xfs_fsmap_low_group_key(mp, info->dev, rtg_rgno(rtg),
+				&info->low);
+		trace_xfs_fsmap_high_group_key(mp, info->dev, rtg_rgno(rtg),
+				&info->high);
+
+		error = xfs_getfsmap_rtdev_rmapbt_query(tp, info, &bt_cur);
+		if (error)
+			break;
+
+		/*
+		 * Set the rtgroup low key to the start of the rtgroup prior to
+		 * moving on to the next rtgroup.
+		 */
+		if (rtg_rgno(rtg) == start_rg)
+			memset(&info->low, 0, sizeof(info->low));
+
+		/*
+		 * If this is the last rtgroup, report any gap at the end of it
+		 * before we drop the reference to the perag when the loop
+		 * terminates.
+		 */
+		if (rtg_rgno(rtg) == end_rg) {
+			info->last = true;
+			error = xfs_getfsmap_rtdev_rmapbt_helper(bt_cur,
+					&info->high, info);
+			if (error)
+				break;
+		}
+		info->group = NULL;
+	}
+
+	if (bt_cur) {
+		xfs_rtgroup_unlock(to_rtg(bt_cur->bc_group),
+				XFS_RTGLOCK_RMAP);
+		xfs_btree_del_cursor(bt_cur, error < 0 ? XFS_BTREE_ERROR :
+							 XFS_BTREE_NOERROR);
+	}
+
+	/* loop termination case */
+	if (rtg) {
+		info->group = NULL;
+		xfs_rtgroup_rele(rtg);
+	}
+
+	return error;
+}
 #endif /* CONFIG_XFS_RT */
 
 /* Do we recognize the device? */
@@ -971,7 +1140,10 @@ xfs_getfsmap(
 	if (mp->m_rtdev_targp) {
 		handlers[2].nr_sectors = XFS_FSB_TO_BB(mp, mp->m_sb.sb_rblocks);
 		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 18/37] xfs: check that the rtrmapbt maxlevels doesn't increase when growing fs

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

The size of filesystem transaction reservations depends on the maximum
height (maxlevels) of the realtime btrees.  Since we don't want a grow
operation to increase the reservation size enough that we'll fail the
minimum log size checks on the next mount, constrain growfs operations
if they would cause an increase in those maxlevels.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/xfs_fsops.c   |   11 +++++++++++
 fs/xfs/xfs_rtalloc.c |   25 ++++++++++++++++++-------
 fs/xfs/xfs_rtalloc.h |   10 ++++++++++
 fs/xfs/xfs_trace.h   |   21 +++++++++++++++++++++
 4 files changed, 60 insertions(+), 7 deletions(-)


diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c
index e1145107d8cbd1..9df5a09c0acd3b 100644
--- a/fs/xfs/xfs_fsops.c
+++ b/fs/xfs/xfs_fsops.c
@@ -22,6 +22,7 @@
 #include "xfs_ag_resv.h"
 #include "xfs_trace.h"
 #include "xfs_rtalloc.h"
+#include "xfs_rtrmap_btree.h"
 
 /*
  * Write new AG headers to disk. Non-transactional, but need to be
@@ -114,6 +115,12 @@ xfs_growfs_data_private(
 		xfs_buf_relse(bp);
 	}
 
+	/* Make sure the new fs size won't cause problems with the log. */
+	error = xfs_growfs_check_rtgeom(mp, nb, mp->m_sb.sb_rblocks,
+			mp->m_sb.sb_rextsize);
+	if (error)
+		return error;
+
 	nb_div = nb;
 	nb_mod = do_div(nb_div, mp->m_sb.sb_agblocks);
 	if (nb_mod && nb_mod >= XFS_MIN_AG_BLOCKS)
@@ -221,7 +228,11 @@ xfs_growfs_data_private(
 		error = xfs_fs_reserve_ag_blocks(mp);
 		if (error == -ENOSPC)
 			error = 0;
+
+		/* Compute new maxlevels for rt btrees. */
+		xfs_rtrmapbt_compute_maxlevels(mp);
 	}
+
 	return error;
 
 out_trans_cancel:
diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c
index c7efd926413981..3c1bce5a4855f2 100644
--- a/fs/xfs/xfs_rtalloc.c
+++ b/fs/xfs/xfs_rtalloc.c
@@ -989,9 +989,11 @@ xfs_growfs_rt_bmblock(
 		goto out_free;
 
 	/*
-	 * Ensure the mount RT feature flag is now set.
+	 * Ensure the mount RT feature flag is now set, and compute new
+	 * maxlevels for rt btrees.
 	 */
 	mp->m_features |= XFS_FEAT_REALTIME;
+	xfs_rtrmapbt_compute_maxlevels(mp);
 
 	kfree(nmp);
 	return 0;
@@ -1159,29 +1161,37 @@ xfs_growfs_rtg(
 	return error;
 }
 
-static int
+int
 xfs_growfs_check_rtgeom(
 	const struct xfs_mount	*mp,
+	xfs_rfsblock_t		dblocks,
 	xfs_rfsblock_t		rblocks,
 	xfs_extlen_t		rextsize)
 {
+	xfs_extlen_t		min_logfsbs;
 	struct xfs_mount	*nmp;
-	int			error = 0;
 
 	nmp = xfs_growfs_rt_alloc_fake_mount(mp, rblocks, rextsize);
 	if (!nmp)
 		return -ENOMEM;
+	nmp->m_sb.sb_dblocks = dblocks;
+
+	xfs_rtrmapbt_compute_maxlevels(nmp);
+	xfs_trans_resv_calc(nmp, M_RES(nmp));
 
 	/*
 	 * New summary size can't be more than half the size of the log.  This
 	 * prevents us from getting a log overflow, since we'll log basically
 	 * the whole summary file at once.
 	 */
-	if (nmp->m_rsumblocks > (mp->m_sb.sb_logblocks >> 1))
-		error = -EINVAL;
+	min_logfsbs = min_t(xfs_extlen_t, xfs_log_calc_minimum_size(nmp),
+			nmp->m_rsumblocks * 2);
 
 	kfree(nmp);
-	return error;
+
+	if (min_logfsbs > mp->m_sb.sb_logblocks)
+		return -EINVAL;
+	return 0;
 }
 
 /*
@@ -1300,7 +1310,8 @@ xfs_growfs_rt(
 		goto out_unlock;
 
 	/* Make sure the new fs size won't cause problems with the log. */
-	error = xfs_growfs_check_rtgeom(mp, in->newblocks, in->extsize);
+	error = xfs_growfs_check_rtgeom(mp, mp->m_sb.sb_dblocks, in->newblocks,
+			in->extsize);
 	if (error)
 		goto out_unlock;
 
diff --git a/fs/xfs/xfs_rtalloc.h b/fs/xfs/xfs_rtalloc.h
index d87523e6a55006..9044f7226ab6fc 100644
--- a/fs/xfs/xfs_rtalloc.h
+++ b/fs/xfs/xfs_rtalloc.h
@@ -46,6 +46,8 @@ xfs_growfs_rt(
 	xfs_growfs_rt_t		*in);	/* user supplied growfs struct */
 
 int xfs_rtalloc_reinit_frextents(struct xfs_mount *mp);
+int xfs_growfs_check_rtgeom(const struct xfs_mount *mp, xfs_rfsblock_t dblocks,
+		xfs_rfsblock_t rblocks, xfs_agblock_t rextsize);
 #else
 # define xfs_growfs_rt(mp,in)				(-ENOSYS)
 # define xfs_rtalloc_reinit_frextents(m)		(0)
@@ -65,6 +67,14 @@ xfs_rtmount_init(
 # define xfs_rtunmount_inodes(m)
 # define xfs_rt_resv_free(mp)				((void)0)
 # define xfs_rt_resv_init(mp)				(0)
+
+static inline int
+xfs_growfs_check_rtgeom(const struct xfs_mount *mp,
+		xfs_rfsblock_t dblocks, xfs_rfsblock_t rblocks,
+		xfs_extlen_t rextsize)
+{
+	return 0;
+}
 #endif	/* CONFIG_XFS_RT */
 
 #endif	/* __XFS_RTALLOC_H__ */
diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h
index a098935163b7c2..84cdc145e2d96a 100644
--- a/fs/xfs/xfs_trace.h
+++ b/fs/xfs/xfs_trace.h
@@ -5622,6 +5622,27 @@ DEFINE_METAFILE_RESV_EVENT(xfs_metafile_resv_free_space);
 DEFINE_METAFILE_RESV_EVENT(xfs_metafile_resv_critical);
 DEFINE_INODE_ERROR_EVENT(xfs_metafile_resv_init_error);
 
+#ifdef CONFIG_XFS_RT
+TRACE_EVENT(xfs_growfs_check_rtgeom,
+	TP_PROTO(const struct xfs_mount *mp, unsigned int min_logfsbs),
+	TP_ARGS(mp, min_logfsbs),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(unsigned int, logblocks)
+		__field(unsigned int, min_logfsbs)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->logblocks = mp->m_sb.sb_logblocks;
+		__entry->min_logfsbs = min_logfsbs;
+	),
+	TP_printk("dev %d:%d logblocks %u min_logfsbs %u",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->logblocks,
+		  __entry->min_logfsbs)
+);
+#endif /* CONFIG_XFS_RT */
+
 #endif /* _TRACE_XFS_H */
 
 #undef TRACE_INCLUDE_PATH

[PATCH 19/37] xfs: report realtime rmap btree corruption errors to the health system

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Whenever we encounter corrupt realtime rmap btree blocks, we should
report that to the health monitoring system for later reporting.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_btree.h        |    2 +-
 fs/xfs/libxfs/xfs_fs.h           |    1 +
 fs/xfs/libxfs/xfs_health.h       |    4 +++-
 fs/xfs/libxfs/xfs_rtgroup.c      |    1 +
 fs/xfs/libxfs/xfs_rtrmap_btree.c |   10 ++++++++--
 fs/xfs/xfs_health.c              |    1 +
 6 files changed, 15 insertions(+), 4 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h
index ee82dc777d6d5b..dbc047b2fb2cf5 100644
--- a/fs/xfs/libxfs/xfs_btree.h
+++ b/fs/xfs/libxfs/xfs_btree.h
@@ -135,7 +135,7 @@ struct xfs_btree_ops {
 	/* offset of btree stats array */
 	unsigned int		statoff;
 
-	/* sick mask for health reporting (only for XFS_BTREE_TYPE_AG) */
+	/* sick mask for health reporting (not for bmap btrees) */
 	unsigned int		sick_mask;
 
 	/* cursor operations */
diff --git a/fs/xfs/libxfs/xfs_fs.h b/fs/xfs/libxfs/xfs_fs.h
index 41ce4d3d650ec7..7cca458ff81245 100644
--- a/fs/xfs/libxfs/xfs_fs.h
+++ b/fs/xfs/libxfs/xfs_fs.h
@@ -993,6 +993,7 @@ struct xfs_rtgroup_geometry {
 #define XFS_RTGROUP_GEOM_SICK_SUPER	(1U << 0)  /* superblock */
 #define XFS_RTGROUP_GEOM_SICK_BITMAP	(1U << 1)  /* rtbitmap */
 #define XFS_RTGROUP_GEOM_SICK_SUMMARY	(1U << 2)  /* rtsummary */
+#define XFS_RTGROUP_GEOM_SICK_RMAPBT	(1U << 3)  /* reverse mappings */
 
 /*
  * ioctl commands that are used by Linux filesystems
diff --git a/fs/xfs/libxfs/xfs_health.h b/fs/xfs/libxfs/xfs_health.h
index d34986ac18c3fa..5c8a0aff6ba6e9 100644
--- a/fs/xfs/libxfs/xfs_health.h
+++ b/fs/xfs/libxfs/xfs_health.h
@@ -70,6 +70,7 @@ struct xfs_rtgroup;
 #define XFS_SICK_RG_SUPER	(1 << 0)  /* rt group superblock */
 #define XFS_SICK_RG_BITMAP	(1 << 1)  /* rt group bitmap */
 #define XFS_SICK_RG_SUMMARY	(1 << 2)  /* rt groups summary */
+#define XFS_SICK_RG_RMAPBT	(1 << 3)  /* reverse mappings */
 
 /* Observable health issues for AG metadata. */
 #define XFS_SICK_AG_SB		(1 << 0)  /* superblock */
@@ -115,7 +116,8 @@ struct xfs_rtgroup;
 
 #define XFS_SICK_RG_PRIMARY	(XFS_SICK_RG_SUPER | \
 				 XFS_SICK_RG_BITMAP | \
-				 XFS_SICK_RG_SUMMARY)
+				 XFS_SICK_RG_SUMMARY | \
+				 XFS_SICK_RG_RMAPBT)
 
 #define XFS_SICK_AG_PRIMARY	(XFS_SICK_AG_SB | \
 				 XFS_SICK_AG_AGF | \
diff --git a/fs/xfs/libxfs/xfs_rtgroup.c b/fs/xfs/libxfs/xfs_rtgroup.c
index 5f31b6e65d5d17..b7ed2d27d54553 100644
--- a/fs/xfs/libxfs/xfs_rtgroup.c
+++ b/fs/xfs/libxfs/xfs_rtgroup.c
@@ -357,6 +357,7 @@ static const struct xfs_rtginode_ops xfs_rtginode_ops[XFS_RTGI_MAX] = {
 	[XFS_RTGI_RMAP] = {
 		.name		= "rmap",
 		.metafile_type	= XFS_METAFILE_RTRMAP,
+		.sick		= XFS_SICK_RG_RMAPBT,
 		.fmt_mask	= 1U << XFS_DINODE_FMT_META_BTREE,
 		/*
 		 * growfs must create the rtrmap inodes before adding a
diff --git a/fs/xfs/libxfs/xfs_rtrmap_btree.c b/fs/xfs/libxfs/xfs_rtrmap_btree.c
index 7654661f4f5823..0a78dee01b1b2e 100644
--- a/fs/xfs/libxfs/xfs_rtrmap_btree.c
+++ b/fs/xfs/libxfs/xfs_rtrmap_btree.c
@@ -27,6 +27,7 @@
 #include "xfs_extent_busy.h"
 #include "xfs_rtgroup.h"
 #include "xfs_bmap.h"
+#include "xfs_health.h"
 
 static struct kmem_cache	*xfs_rtrmapbt_cur_cache;
 
@@ -496,6 +497,7 @@ const struct xfs_btree_ops xfs_rtrmapbt_ops = {
 
 	.lru_refs		= XFS_RMAP_BTREE_REF,
 	.statoff		= XFS_STATS_CALC_INDEX(xs_rtrmap_2),
+	.sick_mask		= XFS_SICK_RG_RMAPBT,
 
 	.dup_cursor		= xfs_rtrmapbt_dup_cursor,
 	.alloc_block		= xfs_btree_alloc_metafile_block,
@@ -755,16 +757,20 @@ xfs_iformat_rtrmap(
 	 * growfs must create the rtrmap inodes before adding a realtime volume
 	 * to the filesystem, so we cannot use the rtrmapbt predicate here.
 	 */
-	if (!xfs_has_rmapbt(ip->i_mount))
+	if (!xfs_has_rmapbt(ip->i_mount)) {
+		xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE);
 		return -EFSCORRUPTED;
+	}
 
 	dsize = XFS_DFORK_SIZE(dip, mp, XFS_DATA_FORK);
 	numrecs = be16_to_cpu(dfp->bb_numrecs);
 	level = be16_to_cpu(dfp->bb_level);
 
 	if (level > mp->m_rtrmap_maxlevels ||
-	    xfs_rtrmap_droot_space_calc(level, numrecs) > dsize)
+	    xfs_rtrmap_droot_space_calc(level, numrecs) > dsize) {
+		xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE);
 		return -EFSCORRUPTED;
+	}
 
 	broot = xfs_broot_alloc(xfs_ifork_ptr(ip, XFS_DATA_FORK),
 			xfs_rtrmap_broot_space_calc(mp, level, numrecs));
diff --git a/fs/xfs/xfs_health.c b/fs/xfs/xfs_health.c
index c7c2e656199862..d438c3c001c829 100644
--- a/fs/xfs/xfs_health.c
+++ b/fs/xfs/xfs_health.c
@@ -447,6 +447,7 @@ static const struct ioctl_sick_map rtgroup_map[] = {
 	{ XFS_SICK_RG_SUPER,	XFS_RTGROUP_GEOM_SICK_SUPER },
 	{ XFS_SICK_RG_BITMAP,	XFS_RTGROUP_GEOM_SICK_BITMAP },
 	{ XFS_SICK_RG_SUMMARY,	XFS_RTGROUP_GEOM_SICK_SUMMARY },
+	{ XFS_SICK_RG_RMAPBT,	XFS_RTGROUP_GEOM_SICK_RMAPBT },
 };
 
 /* Fill out rtgroup geometry health info. */

[PATCH 20/37] xfs: allow queued realtime intents to drain before scrubbing

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

When a writer thread executes a chain of log intent items for the
realtime volume, the ILOCKs taken during each step are for each rt
metadata file, not the entire rt volume itself.  Although scrub takes
all rt metadata ILOCKs, this isn't sufficient to guard against scrub
checking the rt volume while that writer thread is in the middle of
finishing a chain because there's no higher level locking primitive
guarding the realtime volume.

When there's a collision, cross-referencing between data structures
(e.g. rtrmapbt and rtrefcountbt) yields false corruption events; if
repair is running, this results in incorrect repairs, which is
catastrophic.

Fix this by adding to the mount structure the same drain that we use to
protect scrub against concurrent AG updates, but this time for the
realtime volume.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/scrub/bmap.c      |    7 ++++
 fs/xfs/scrub/common.c    |   72 ++++++++++++++++++++++++++++++++++++++++++++--
 fs/xfs/scrub/common.h    |    5 ++-
 fs/xfs/scrub/rgsuper.c   |    4 ++-
 fs/xfs/scrub/rtbitmap.c  |    8 ++++-
 fs/xfs/scrub/rtsummary.c |    5 +++
 fs/xfs/scrub/scrub.c     |    2 +
 fs/xfs/xfs_drain.c       |   20 ++++++-------
 fs/xfs/xfs_drain.h       |    7 +++-
 9 files changed, 108 insertions(+), 22 deletions(-)


diff --git a/fs/xfs/scrub/bmap.c b/fs/xfs/scrub/bmap.c
index 0d7ad692822d48..dd99366643f832 100644
--- a/fs/xfs/scrub/bmap.c
+++ b/fs/xfs/scrub/bmap.c
@@ -324,10 +324,15 @@ xchk_bmap_rt_iextent_xref(
 			irec->br_startoff, &error))
 		return;
 
-	xchk_rtgroup_lock(&info->sc->sr, XCHK_RTGLOCK_ALL);
+	error = xchk_rtgroup_lock(info->sc, &info->sc->sr, XCHK_RTGLOCK_ALL);
+	if (!xchk_fblock_process_error(info->sc, info->whichfork,
+			irec->br_startoff, &error))
+		goto out_free;
+
 	xchk_xref_is_used_rt_space(info->sc, irec->br_startblock,
 			irec->br_blockcount);
 
+out_free:
 	xchk_rtgroup_free(info->sc, &info->sc->sr);
 }
 
diff --git a/fs/xfs/scrub/common.c b/fs/xfs/scrub/common.c
index 5cbd94b56582a4..613fb54e723ede 100644
--- a/fs/xfs/scrub/common.c
+++ b/fs/xfs/scrub/common.c
@@ -719,13 +719,79 @@ xchk_rtgroup_init(
 	return 0;
 }
 
-void
+/* Lock all the rt group metadata inode ILOCKs and wait for intents. */
+int
 xchk_rtgroup_lock(
+	struct xfs_scrub	*sc,
 	struct xchk_rt		*sr,
 	unsigned int		rtglock_flags)
 {
-	xfs_rtgroup_lock(sr->rtg, rtglock_flags);
+	int			error = 0;
+
+	ASSERT(sr->rtg != NULL);
+
+	/*
+	 * If we're /only/ locking the rtbitmap in shared mode, then we're
+	 * obviously not trying to compare records in two metadata inodes.
+	 * There's no need to drain intents here because the caller (most
+	 * likely the rgsuper scanner) doesn't need that level of consistency.
+	 */
+	if (rtglock_flags == XFS_RTGLOCK_BITMAP_SHARED) {
+		xfs_rtgroup_lock(sr->rtg, rtglock_flags);
+		sr->rtlock_flags = rtglock_flags;
+		return 0;
+	}
+
+	do {
+		if (xchk_should_terminate(sc, &error))
+			return error;
+
+		xfs_rtgroup_lock(sr->rtg, rtglock_flags);
+
+		/*
+		 * If we've grabbed a non-metadata file for scrubbing, we
+		 * assume that holding its ILOCK will suffice to coordinate
+		 * with any rt intent chains involving this inode.
+		 */
+		if (sc->ip && !xfs_is_internal_inode(sc->ip))
+			break;
+
+		/*
+		 * Decide if the rt group is quiet enough for all metadata to
+		 * be consistent with each other.  Regular file IO doesn't get
+		 * to lock all the rt inodes at the same time, which means that
+		 * there could be other threads in the middle of processing a
+		 * chain of deferred ops.
+		 *
+		 * We just locked all the metadata inodes for this rt group;
+		 * now take a look to see if there are any intents in progress.
+		 * If there are, drop the rt group inode locks and wait for the
+		 * intents to drain.  Since we hold the rt group inode locks
+		 * for the duration of the scrub, this is the only time we have
+		 * to sample the intents counter; any threads increasing it
+		 * after this point can't possibly be in the middle of a chain
+		 * of rt metadata updates.
+		 *
+		 * Obviously, this should be slanted against scrub and in favor
+		 * of runtime threads.
+		 */
+		if (!xfs_group_intent_busy(rtg_group(sr->rtg)))
+			break;
+
+		xfs_rtgroup_unlock(sr->rtg, rtglock_flags);
+
+		if (!(sc->flags & XCHK_FSGATES_DRAIN))
+			return -ECHRNG;
+		error = xfs_group_intent_drain(rtg_group(sr->rtg));
+		if (error) {
+			if (error == -ERESTARTSYS)
+				error = -EINTR;
+			return error;
+		}
+	} while (1);
+
 	sr->rtlock_flags = rtglock_flags;
+	return 0;
 }
 
 /*
@@ -1379,7 +1445,7 @@ xchk_fsgates_enable(
 	trace_xchk_fsgates_enable(sc, scrub_fsgates);
 
 	if (scrub_fsgates & XCHK_FSGATES_DRAIN)
-		xfs_drain_wait_enable();
+		xfs_defer_drain_wait_enable();
 
 	if (scrub_fsgates & XCHK_FSGATES_QUOTA)
 		xfs_dqtrx_hook_enable();
diff --git a/fs/xfs/scrub/common.h b/fs/xfs/scrub/common.h
index 9ff3cafd867962..e734572a8dd6ec 100644
--- a/fs/xfs/scrub/common.h
+++ b/fs/xfs/scrub/common.h
@@ -141,12 +141,13 @@ xchk_rtgroup_init_existing(
 	return error == -ENOENT ? -EFSCORRUPTED : error;
 }
 
-void xchk_rtgroup_lock(struct xchk_rt *sr, unsigned int rtglock_flags);
+int xchk_rtgroup_lock(struct xfs_scrub *sc, struct xchk_rt *sr,
+		unsigned int rtglock_flags);
 void xchk_rtgroup_free(struct xfs_scrub *sc, struct xchk_rt *sr);
 #else
 # define xchk_rtgroup_init(sc, rgno, sr)		(-EFSCORRUPTED)
 # define xchk_rtgroup_init_existing(sc, rgno, sr)	(-EFSCORRUPTED)
-# define xchk_rtgroup_lock(sc, lockflags)		do { } while (0)
+# define xchk_rtgroup_lock(sc, sr, lockflags)		(-EFSCORRUPTED)
 # define xchk_rtgroup_free(sc, sr)			do { } while (0)
 #endif /* CONFIG_XFS_RT */
 
diff --git a/fs/xfs/scrub/rgsuper.c b/fs/xfs/scrub/rgsuper.c
index 463b3573bb761b..e062c7d12565cd 100644
--- a/fs/xfs/scrub/rgsuper.c
+++ b/fs/xfs/scrub/rgsuper.c
@@ -61,7 +61,9 @@ xchk_rgsuperblock(
 	if (!xchk_xref_process_error(sc, 0, 0, &error))
 		return error;
 
-	xchk_rtgroup_lock(&sc->sr, XFS_RTGLOCK_BITMAP_SHARED);
+	error = xchk_rtgroup_lock(sc, &sc->sr, XFS_RTGLOCK_BITMAP_SHARED);
+	if (error)
+		return error;
 
 	/*
 	 * Since we already validated the rt superblock at mount time, we don't
diff --git a/fs/xfs/scrub/rtbitmap.c b/fs/xfs/scrub/rtbitmap.c
index fb4970c877abd3..819026ea2d741f 100644
--- a/fs/xfs/scrub/rtbitmap.c
+++ b/fs/xfs/scrub/rtbitmap.c
@@ -30,6 +30,9 @@ xchk_setup_rtbitmap(
 	struct xchk_rtbitmap	*rtb;
 	int			error;
 
+	if (xchk_need_intent_drain(sc))
+		xchk_fsgates_enable(sc, XCHK_FSGATES_DRAIN);
+
 	rtb = kzalloc(sizeof(struct xchk_rtbitmap), XCHK_GFP_FLAGS);
 	if (!rtb)
 		return -ENOMEM;
@@ -57,12 +60,15 @@ xchk_setup_rtbitmap(
 	if (error)
 		return error;
 
+	error = xchk_rtgroup_lock(sc, &sc->sr, XCHK_RTGLOCK_ALL);
+	if (error)
+		return error;
+
 	/*
 	 * Now that we've locked the rtbitmap, we can't race with growfsrt
 	 * trying to expand the bitmap or change the size of the rt volume.
 	 * Hence it is safe to compute and check the geometry values.
 	 */
-	xchk_rtgroup_lock(&sc->sr, XFS_RTGLOCK_BITMAP);
 	if (mp->m_sb.sb_rblocks) {
 		rtb->rextents = xfs_blen_to_rtbxlen(mp, mp->m_sb.sb_rblocks);
 		rtb->rextslog = xfs_compute_rextslog(rtb->rextents);
diff --git a/fs/xfs/scrub/rtsummary.c b/fs/xfs/scrub/rtsummary.c
index f1af5431b38856..4ac679c1bd29cd 100644
--- a/fs/xfs/scrub/rtsummary.c
+++ b/fs/xfs/scrub/rtsummary.c
@@ -89,6 +89,10 @@ xchk_setup_rtsummary(
 	if (error)
 		return error;
 
+	error = xchk_rtgroup_lock(sc, &sc->sr, XFS_RTGLOCK_BITMAP);
+	if (error)
+		return error;
+
 	/*
 	 * Now that we've locked the rtbitmap and rtsummary, we can't race with
 	 * growfsrt trying to expand the summary or change the size of the rt
@@ -99,7 +103,6 @@ xchk_setup_rtsummary(
 	 * exclusively here.  If we ever start caring about running concurrent
 	 * fsmap with scrub this could be changed.
 	 */
-	xchk_rtgroup_lock(&sc->sr, XFS_RTGLOCK_BITMAP);
 	if (mp->m_sb.sb_rblocks) {
 		rts->rextents = xfs_blen_to_rtbxlen(mp, mp->m_sb.sb_rblocks);
 		rts->rbmblocks = xfs_rtbitmap_blockcount(mp);
diff --git a/fs/xfs/scrub/scrub.c b/fs/xfs/scrub/scrub.c
index 950f5a58dcd967..652d347cee9929 100644
--- a/fs/xfs/scrub/scrub.c
+++ b/fs/xfs/scrub/scrub.c
@@ -164,7 +164,7 @@ xchk_fsgates_disable(
 	trace_xchk_fsgates_disable(sc, sc->flags & XCHK_FSGATES_ALL);
 
 	if (sc->flags & XCHK_FSGATES_DRAIN)
-		xfs_drain_wait_disable();
+		xfs_defer_drain_wait_disable();
 
 	if (sc->flags & XCHK_FSGATES_QUOTA)
 		xfs_dqtrx_hook_disable();
diff --git a/fs/xfs/xfs_drain.c b/fs/xfs/xfs_drain.c
index 5ede81fadbd8ca..fa5f31931efdb5 100644
--- a/fs/xfs/xfs_drain.c
+++ b/fs/xfs/xfs_drain.c
@@ -13,28 +13,28 @@
 #include "xfs_trace.h"
 
 /*
- * Use a static key here to reduce the overhead of xfs_drain_rele.  If the
- * compiler supports jump labels, the static branch will be replaced by a nop
- * sled when there are no xfs_drain_wait callers.  Online fsck is currently
- * the only caller, so this is a reasonable tradeoff.
+ * Use a static key here to reduce the overhead of xfs_defer_drain_rele.  If
+ * the compiler supports jump labels, the static branch will be replaced by a
+ * nop sled when there are no xfs_defer_drain_wait callers.  Online fsck is
+ * currently the only caller, so this is a reasonable tradeoff.
  *
  * Note: Patching the kernel code requires taking the cpu hotplug lock.  Other
  * parts of the kernel allocate memory with that lock held, which means that
  * XFS callers cannot hold any locks that might be used by memory reclaim or
  * writeback when calling the static_branch_{inc,dec} functions.
  */
-static DEFINE_STATIC_KEY_FALSE(xfs_drain_waiter_gate);
+static DEFINE_STATIC_KEY_FALSE(xfs_defer_drain_waiter_gate);
 
 void
-xfs_drain_wait_disable(void)
+xfs_defer_drain_wait_disable(void)
 {
-	static_branch_dec(&xfs_drain_waiter_gate);
+	static_branch_dec(&xfs_defer_drain_waiter_gate);
 }
 
 void
-xfs_drain_wait_enable(void)
+xfs_defer_drain_wait_enable(void)
 {
-	static_branch_inc(&xfs_drain_waiter_gate);
+	static_branch_inc(&xfs_defer_drain_waiter_gate);
 }
 
 void
@@ -71,7 +71,7 @@ static inline bool has_waiters(struct wait_queue_head *wq_head)
 static inline void xfs_defer_drain_rele(struct xfs_defer_drain *dr)
 {
 	if (atomic_dec_and_test(&dr->dr_count) &&
-	    static_branch_unlikely(&xfs_drain_waiter_gate) &&
+	    static_branch_unlikely(&xfs_defer_drain_waiter_gate) &&
 	    has_waiters(&dr->dr_waiters))
 		wake_up(&dr->dr_waiters);
 }
diff --git a/fs/xfs/xfs_drain.h b/fs/xfs/xfs_drain.h
index efcf88df9a5e70..4d446dbf65e519 100644
--- a/fs/xfs/xfs_drain.h
+++ b/fs/xfs/xfs_drain.h
@@ -26,8 +26,8 @@ struct xfs_defer_drain {
 void xfs_defer_drain_init(struct xfs_defer_drain *dr);
 void xfs_defer_drain_free(struct xfs_defer_drain *dr);
 
-void xfs_drain_wait_disable(void);
-void xfs_drain_wait_enable(void);
+void xfs_defer_drain_wait_disable(void);
+void xfs_defer_drain_wait_enable(void);
 
 /*
  * Deferred Work Intent Drains
@@ -61,6 +61,9 @@ void xfs_drain_wait_enable(void);
  * All functions that create work items must increment the intent counter as
  * soon as the item is added to the transaction and cannot drop the counter
  * until the item is finished or cancelled.
+ *
+ * The same principles apply to realtime groups because the rt metadata inode
+ * ILOCKs are not held across transaction rolls.
  */
 struct xfs_group *xfs_group_intent_get(struct xfs_mount *mp,
 		xfs_fsblock_t fsbno, enum xfs_group_type type);

[PATCH 21/37] xfs: scrub the realtime rmapbt

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

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" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/Makefile             |    1 
 fs/xfs/libxfs/xfs_fs.h      |    3 -
 fs/xfs/scrub/common.c       |   86 ++++++++++++++++++++
 fs/xfs/scrub/common.h       |   10 ++
 fs/xfs/scrub/health.c       |    1 
 fs/xfs/scrub/inode.c        |    6 -
 fs/xfs/scrub/inode_repair.c |    7 +-
 fs/xfs/scrub/repair.c       |    1 
 fs/xfs/scrub/rtrmap.c       |  184 +++++++++++++++++++++++++++++++++++++++++++
 fs/xfs/scrub/scrub.c        |    9 ++
 fs/xfs/scrub/scrub.h        |    5 +
 fs/xfs/scrub/stats.c        |    1 
 fs/xfs/scrub/trace.h        |    4 +
 13 files changed, 307 insertions(+), 11 deletions(-)
 create mode 100644 fs/xfs/scrub/rtrmap.c


diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile
index ff45efb2463f73..136a465e00d2b1 100644
--- a/fs/xfs/Makefile
+++ b/fs/xfs/Makefile
@@ -194,6 +194,7 @@ xfs-$(CONFIG_XFS_ONLINE_SCRUB_STATS) += scrub/stats.o
 xfs-$(CONFIG_XFS_RT)		+= $(addprefix scrub/, \
 				   rgsuper.o \
 				   rtbitmap.o \
+				   rtrmap.o \
 				   rtsummary.o \
 				   )
 
diff --git a/fs/xfs/libxfs/xfs_fs.h b/fs/xfs/libxfs/xfs_fs.h
index 7cca458ff81245..34fcbcd0bcd5e3 100644
--- a/fs/xfs/libxfs/xfs_fs.h
+++ b/fs/xfs/libxfs/xfs_fs.h
@@ -737,9 +737,10 @@ struct xfs_scrub_metadata {
 #define XFS_SCRUB_TYPE_DIRTREE	28	/* directory tree structure */
 #define XFS_SCRUB_TYPE_METAPATH	29	/* metadata directory tree paths */
 #define XFS_SCRUB_TYPE_RGSUPER	30	/* realtime superblock */
+#define XFS_SCRUB_TYPE_RTRMAPBT	31	/* rtgroup reverse mapping btree */
 
 /* Number of scrub subcommands. */
-#define XFS_SCRUB_TYPE_NR	31
+#define XFS_SCRUB_TYPE_NR	32
 
 /*
  * This special type code only applies to the vectored scrub implementation.
diff --git a/fs/xfs/scrub/common.c b/fs/xfs/scrub/common.c
index 613fb54e723ede..ca43dd4f52b2d6 100644
--- a/fs/xfs/scrub/common.c
+++ b/fs/xfs/scrub/common.c
@@ -35,6 +35,8 @@
 #include "xfs_exchmaps.h"
 #include "xfs_rtbitmap.h"
 #include "xfs_rtgroup.h"
+#include "xfs_rtrmap_btree.h"
+#include "xfs_bmap_util.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/trace.h"
@@ -791,9 +793,28 @@ xchk_rtgroup_lock(
 	} while (1);
 
 	sr->rtlock_flags = rtglock_flags;
+
+	if (xfs_has_rtrmapbt(sc->mp) && (rtglock_flags & XFS_RTGLOCK_RMAP))
+		sr->rmap_cur = xfs_rtrmapbt_init_cursor(sc->tp, sr->rtg);
+
 	return 0;
 }
 
+/*
+ * Free all the btree cursors and other incore data relating to the realtime
+ * group.  This has to be done /before/ committing (or cancelling) the scrub
+ * transaction.
+ */
+void
+xchk_rtgroup_btcur_free(
+	struct xchk_rt		*sr)
+{
+	if (sr->rmap_cur)
+		xfs_btree_del_cursor(sr->rmap_cur, XFS_BTREE_ERROR);
+
+	sr->rmap_cur = NULL;
+}
+
 /*
  * Unlock the realtime group.  This must be done /after/ committing (or
  * cancelling) the scrub transaction.
@@ -878,6 +899,14 @@ xchk_setup_fs(
 	return xchk_trans_alloc(sc, resblks);
 }
 
+/* Set us up with a transaction and an empty context to repair rt metadata. */
+int
+xchk_setup_rt(
+	struct xfs_scrub	*sc)
+{
+	return xchk_trans_alloc(sc, 0);
+}
+
 /* Set us up with AG headers and btree cursors. */
 int
 xchk_setup_ag_btree(
@@ -1639,3 +1668,60 @@ xchk_inode_rootdir_inum(const struct xfs_inode *ip)
 		return mp->m_metadirip->i_ino;
 	return mp->m_rootip->i_ino;
 }
+
+static int
+xchk_meta_btree_count_blocks(
+	struct xfs_scrub	*sc,
+	xfs_extnum_t		*nextents,
+	xfs_filblks_t		*count)
+{
+	struct xfs_btree_cur	*cur;
+	int			error;
+
+	if (!sc->sr.rtg) {
+		ASSERT(0);
+		return -EFSCORRUPTED;
+	}
+
+	switch (sc->ip->i_metatype) {
+	case XFS_METAFILE_RTRMAP:
+		cur = xfs_rtrmapbt_init_cursor(sc->tp, sc->sr.rtg);
+		break;
+	default:
+		ASSERT(0);
+		return -EFSCORRUPTED;
+	}
+
+	error = xfs_btree_count_blocks(cur, count);
+	xfs_btree_del_cursor(cur, error);
+	if (!error) {
+		*nextents = 0;
+		(*count)--;	/* don't count the btree iroot */
+	}
+	return error;
+}
+
+/* Count the blocks used by a file, even if it's a metadata inode. */
+int
+xchk_inode_count_blocks(
+	struct xfs_scrub	*sc,
+	int			whichfork,
+	xfs_extnum_t		*nextents,
+	xfs_filblks_t		*count)
+{
+	struct xfs_ifork	*ifp = xfs_ifork_ptr(sc->ip, whichfork);
+
+	if (!ifp) {
+		*nextents = 0;
+		*count = 0;
+		return 0;
+	}
+
+	if (ifp->if_format == XFS_DINODE_FMT_META_BTREE) {
+		ASSERT(whichfork == XFS_DATA_FORK);
+		return xchk_meta_btree_count_blocks(sc, nextents, count);
+	}
+
+	return xfs_bmap_count_blocks(sc->tp, sc->ip, whichfork, nextents,
+			count);
+}
diff --git a/fs/xfs/scrub/common.h b/fs/xfs/scrub/common.h
index e734572a8dd6ec..1576467f724431 100644
--- a/fs/xfs/scrub/common.h
+++ b/fs/xfs/scrub/common.h
@@ -63,6 +63,7 @@ static inline int xchk_setup_nothing(struct xfs_scrub *sc)
 /* Setup functions */
 int xchk_setup_agheader(struct xfs_scrub *sc);
 int xchk_setup_fs(struct xfs_scrub *sc);
+int xchk_setup_rt(struct xfs_scrub *sc);
 int xchk_setup_ag_allocbt(struct xfs_scrub *sc);
 int xchk_setup_ag_iallocbt(struct xfs_scrub *sc);
 int xchk_setup_ag_rmapbt(struct xfs_scrub *sc);
@@ -80,10 +81,12 @@ int xchk_setup_metapath(struct xfs_scrub *sc);
 int xchk_setup_rtbitmap(struct xfs_scrub *sc);
 int xchk_setup_rtsummary(struct xfs_scrub *sc);
 int xchk_setup_rgsuperblock(struct xfs_scrub *sc);
+int xchk_setup_rtrmapbt(struct xfs_scrub *sc);
 #else
 # define xchk_setup_rtbitmap		xchk_setup_nothing
 # define xchk_setup_rtsummary		xchk_setup_nothing
 # define xchk_setup_rgsuperblock	xchk_setup_nothing
+# define xchk_setup_rtrmapbt		xchk_setup_nothing
 #endif
 #ifdef CONFIG_XFS_QUOTA
 int xchk_ino_dqattach(struct xfs_scrub *sc);
@@ -125,7 +128,8 @@ xchk_ag_init_existing(
 #ifdef CONFIG_XFS_RT
 
 /* All the locks we need to check an rtgroup. */
-#define XCHK_RTGLOCK_ALL	(XFS_RTGLOCK_BITMAP)
+#define XCHK_RTGLOCK_ALL	(XFS_RTGLOCK_BITMAP | \
+				 XFS_RTGLOCK_RMAP)
 
 int xchk_rtgroup_init(struct xfs_scrub *sc, xfs_rgnumber_t rgno,
 		struct xchk_rt *sr);
@@ -143,11 +147,13 @@ xchk_rtgroup_init_existing(
 
 int xchk_rtgroup_lock(struct xfs_scrub *sc, struct xchk_rt *sr,
 		unsigned int rtglock_flags);
+void xchk_rtgroup_btcur_free(struct xchk_rt *sr);
 void xchk_rtgroup_free(struct xfs_scrub *sc, struct xchk_rt *sr);
 #else
 # define xchk_rtgroup_init(sc, rgno, sr)		(-EFSCORRUPTED)
 # define xchk_rtgroup_init_existing(sc, rgno, sr)	(-EFSCORRUPTED)
 # define xchk_rtgroup_lock(sc, sr, lockflags)		(-EFSCORRUPTED)
+# define xchk_rtgroup_btcur_free(sr)			do { } while (0)
 # define xchk_rtgroup_free(sc, sr)			do { } while (0)
 #endif /* CONFIG_XFS_RT */
 
@@ -275,6 +281,8 @@ void xchk_fsgates_enable(struct xfs_scrub *sc, unsigned int scrub_fshooks);
 
 int xchk_inode_is_allocated(struct xfs_scrub *sc, xfs_agino_t agino,
 		bool *inuse);
+int xchk_inode_count_blocks(struct xfs_scrub *sc, int whichfork,
+		xfs_extnum_t *nextents, xfs_filblks_t *count);
 
 bool xchk_inode_is_dirtree_root(const struct xfs_inode *ip);
 bool xchk_inode_is_sb_rooted(const struct xfs_inode *ip);
diff --git a/fs/xfs/scrub/health.c b/fs/xfs/scrub/health.c
index ccc6ca5934ca6a..bcc4244e3b55db 100644
--- a/fs/xfs/scrub/health.c
+++ b/fs/xfs/scrub/health.c
@@ -114,6 +114,7 @@ static const struct xchk_health_map type_to_health_flag[XFS_SCRUB_TYPE_NR] = {
 	[XFS_SCRUB_TYPE_DIRTREE]	= { XHG_INO, XFS_SICK_INO_DIRTREE },
 	[XFS_SCRUB_TYPE_METAPATH]	= { XHG_FS,  XFS_SICK_FS_METAPATH },
 	[XFS_SCRUB_TYPE_RGSUPER]	= { XHG_RTGROUP, XFS_SICK_RG_SUPER },
+	[XFS_SCRUB_TYPE_RTRMAPBT]	= { XHG_RTGROUP, XFS_SICK_RG_RMAPBT },
 };
 
 /* Return the health status mask for this scrub type. */
diff --git a/fs/xfs/scrub/inode.c b/fs/xfs/scrub/inode.c
index 2e911f38deaebe..8e702121dc8699 100644
--- a/fs/xfs/scrub/inode.c
+++ b/fs/xfs/scrub/inode.c
@@ -690,15 +690,13 @@ xchk_inode_xref_bmap(
 		return;
 
 	/* Walk all the extents to check nextents/naextents/nblocks. */
-	error = xfs_bmap_count_blocks(sc->tp, sc->ip, XFS_DATA_FORK,
-			&nextents, &count);
+	error = xchk_inode_count_blocks(sc, XFS_DATA_FORK, &nextents, &count);
 	if (!xchk_should_check_xref(sc, &error, NULL))
 		return;
 	if (nextents < xfs_dfork_data_extents(dip))
 		xchk_ino_xref_set_corrupt(sc, sc->ip->i_ino);
 
-	error = xfs_bmap_count_blocks(sc->tp, sc->ip, XFS_ATTR_FORK,
-			&nextents, &acount);
+	error = xchk_inode_count_blocks(sc, XFS_ATTR_FORK, &nextents, &acount);
 	if (!xchk_should_check_xref(sc, &error, NULL))
 		return;
 	if (nextents != xfs_dfork_attr_extents(dip))
diff --git a/fs/xfs/scrub/inode_repair.c b/fs/xfs/scrub/inode_repair.c
index 7faa27472b9129..a94f9df0ca78f6 100644
--- a/fs/xfs/scrub/inode_repair.c
+++ b/fs/xfs/scrub/inode_repair.c
@@ -1536,8 +1536,7 @@ xrep_inode_blockcounts(
 	trace_xrep_inode_blockcounts(sc);
 
 	/* Set data fork counters from the data fork mappings. */
-	error = xfs_bmap_count_blocks(sc->tp, sc->ip, XFS_DATA_FORK,
-			&nextents, &count);
+	error = xchk_inode_count_blocks(sc, XFS_DATA_FORK, &nextents, &count);
 	if (error)
 		return error;
 	if (xfs_is_reflink_inode(sc->ip)) {
@@ -1561,8 +1560,8 @@ xrep_inode_blockcounts(
 	/* Set attr fork counters from the attr fork mappings. */
 	ifp = xfs_ifork_ptr(sc->ip, XFS_ATTR_FORK);
 	if (ifp) {
-		error = xfs_bmap_count_blocks(sc->tp, sc->ip, XFS_ATTR_FORK,
-				&nextents, &acount);
+		error = xchk_inode_count_blocks(sc, XFS_ATTR_FORK, &nextents,
+				&acount);
 		if (error)
 			return error;
 		if (count >= sc->mp->m_sb.sb_dblocks)
diff --git a/fs/xfs/scrub/repair.c b/fs/xfs/scrub/repair.c
index 91c8bc055a4fd7..e788e3032f8e33 100644
--- a/fs/xfs/scrub/repair.c
+++ b/fs/xfs/scrub/repair.c
@@ -62,6 +62,7 @@ xrep_attempt(
 	trace_xrep_attempt(XFS_I(file_inode(sc->file)), sc->sm, error);
 
 	xchk_ag_btcur_free(&sc->sa);
+	xchk_rtgroup_btcur_free(&sc->sr);
 
 	/* Repair whatever's broken. */
 	ASSERT(sc->ops->repair);
diff --git a/fs/xfs/scrub/rtrmap.c b/fs/xfs/scrub/rtrmap.c
new file mode 100644
index 00000000000000..7b5f932bcd947f
--- /dev/null
+++ b/fs/xfs/scrub/rtrmap.c
@@ -0,0 +1,184 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2018-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#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 "xfs_rtgroup.h"
+#include "xfs_metafile.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. */
+int
+xchk_setup_rtrmapbt(
+	struct xfs_scrub	*sc)
+{
+	int			error;
+
+	if (xchk_need_intent_drain(sc))
+		xchk_fsgates_enable(sc, XCHK_FSGATES_DRAIN);
+
+	error = xchk_rtgroup_init(sc, sc->sm->sm_agno, &sc->sr);
+	if (error)
+		return error;
+
+	error = xchk_setup_rt(sc);
+	if (error)
+		return error;
+
+	error = xchk_install_live_inode(sc, rtg_rmap(sc->sr.rtg));
+	if (error)
+		return error;
+
+	return xchk_rtgroup_lock(sc, &sc->sr, XCHK_RTGLOCK_ALL);
+}
+
+/* Realtime reverse mapping. */
+
+struct xchk_rtrmap {
+	/*
+	 * The furthest-reaching of the rmapbt records that we've already
+	 * processed.  This enables us to detect overlapping records for space
+	 * allocations that cannot be shared.
+	 */
+	struct xfs_rmap_irec	overlap_rec;
+
+	/*
+	 * The previous rmapbt record, so that we can check for two records
+	 * that could be one.
+	 */
+	struct xfs_rmap_irec	prev_rec;
+};
+
+/* Flag failures for records that overlap but cannot. */
+STATIC void
+xchk_rtrmapbt_check_overlapping(
+	struct xchk_btree		*bs,
+	struct xchk_rtrmap		*cr,
+	const struct xfs_rmap_irec	*irec)
+{
+	xfs_rtblock_t			pnext, inext;
+
+	if (bs->sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return;
+
+	/* No previous record? */
+	if (cr->overlap_rec.rm_blockcount == 0)
+		goto set_prev;
+
+	/* Do overlap_rec and irec overlap? */
+	pnext = cr->overlap_rec.rm_startblock + cr->overlap_rec.rm_blockcount;
+	if (pnext <= irec->rm_startblock)
+		goto set_prev;
+
+	xchk_btree_set_corrupt(bs->sc, bs->cur, 0);
+
+	/* Save whichever rmap record extends furthest. */
+	inext = irec->rm_startblock + irec->rm_blockcount;
+	if (pnext > inext)
+		return;
+
+set_prev:
+	memcpy(&cr->overlap_rec, irec, sizeof(struct xfs_rmap_irec));
+}
+
+/* Decide if two reverse-mapping records can be merged. */
+static inline bool
+xchk_rtrmap_mergeable(
+	struct xchk_rtrmap		*cr,
+	const struct xfs_rmap_irec	*r2)
+{
+	const struct xfs_rmap_irec	*r1 = &cr->prev_rec;
+
+	/* Ignore if prev_rec is not yet initialized. */
+	if (cr->prev_rec.rm_blockcount == 0)
+		return false;
+
+	if (r1->rm_owner != r2->rm_owner)
+		return false;
+	if (r1->rm_startblock + r1->rm_blockcount != r2->rm_startblock)
+		return false;
+	if ((unsigned long long)r1->rm_blockcount + r2->rm_blockcount >
+	    XFS_RMAP_LEN_MAX)
+		return false;
+	if (r1->rm_flags != r2->rm_flags)
+		return false;
+	return r1->rm_offset + r1->rm_blockcount == r2->rm_offset;
+}
+
+/* Flag failures for records that could be merged. */
+STATIC void
+xchk_rtrmapbt_check_mergeable(
+	struct xchk_btree		*bs,
+	struct xchk_rtrmap		*cr,
+	const struct xfs_rmap_irec	*irec)
+{
+	if (bs->sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return;
+
+	if (xchk_rtrmap_mergeable(cr, irec))
+		xchk_btree_set_corrupt(bs->sc, bs->cur, 0);
+
+	memcpy(&cr->prev_rec, irec, sizeof(struct xfs_rmap_irec));
+}
+
+/* Scrub a realtime rmapbt record. */
+STATIC int
+xchk_rtrmapbt_rec(
+	struct xchk_btree		*bs,
+	const union xfs_btree_rec	*rec)
+{
+	struct xchk_rtrmap		*cr = bs->private;
+	struct xfs_rmap_irec		irec;
+
+	if (xfs_rmap_btrec_to_irec(rec, &irec) != NULL ||
+	    xfs_rtrmap_check_irec(to_rtg(bs->cur->bc_group), &irec) != NULL) {
+		xchk_btree_set_corrupt(bs->sc, bs->cur, 0);
+		return 0;
+	}
+
+	if (bs->sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return 0;
+
+	xchk_rtrmapbt_check_mergeable(bs, cr, &irec);
+	xchk_rtrmapbt_check_overlapping(bs, cr, &irec);
+	return 0;
+}
+
+/* Scrub the realtime rmap btree. */
+int
+xchk_rtrmapbt(
+	struct xfs_scrub	*sc)
+{
+	struct xfs_inode	*ip = rtg_rmap(sc->sr.rtg);
+	struct xfs_owner_info	oinfo;
+	struct xchk_rtrmap	cr = { };
+	int			error;
+
+	error = xchk_metadata_inode_forks(sc);
+	if (error || (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT))
+		return error;
+
+	xfs_rmap_ino_bmbt_owner(&oinfo, ip->i_ino, XFS_DATA_FORK);
+	return xchk_btree(sc, sc->sr.rmap_cur, xchk_rtrmapbt_rec, &oinfo, &cr);
+}
diff --git a/fs/xfs/scrub/scrub.c b/fs/xfs/scrub/scrub.c
index 652d347cee9929..09983899c34164 100644
--- a/fs/xfs/scrub/scrub.c
+++ b/fs/xfs/scrub/scrub.c
@@ -218,6 +218,8 @@ xchk_teardown(
 	int			error)
 {
 	xchk_ag_free(sc, &sc->sa);
+	xchk_rtgroup_btcur_free(&sc->sr);
+
 	if (sc->tp) {
 		if (error == 0 && (sc->sm->sm_flags & XFS_SCRUB_IFLAG_REPAIR))
 			error = xfs_trans_commit(sc->tp);
@@ -458,6 +460,13 @@ static const struct xchk_meta_ops meta_scrub_ops[] = {
 		.has	= xfs_has_rtsb,
 		.repair = xrep_rgsuperblock,
 	},
+	[XFS_SCRUB_TYPE_RTRMAPBT] = {	/* realtime group rmapbt */
+		.type	= ST_RTGROUP,
+		.setup	= xchk_setup_rtrmapbt,
+		.scrub	= xchk_rtrmapbt,
+		.has	= xfs_has_rtrmapbt,
+		.repair	= xrep_notsupported,
+	},
 };
 
 static int
diff --git a/fs/xfs/scrub/scrub.h b/fs/xfs/scrub/scrub.h
index 5dbbe93cb49bfa..0ad5122af486e1 100644
--- a/fs/xfs/scrub/scrub.h
+++ b/fs/xfs/scrub/scrub.h
@@ -126,6 +126,9 @@ struct xchk_rt {
 
 	/* XFS_RTGLOCK_* lock state if locked */
 	unsigned int		rtlock_flags;
+
+	/* rtgroup btrees */
+	struct xfs_btree_cur	*rmap_cur;
 };
 
 struct xfs_scrub {
@@ -280,10 +283,12 @@ int xchk_metapath(struct xfs_scrub *sc);
 int xchk_rtbitmap(struct xfs_scrub *sc);
 int xchk_rtsummary(struct xfs_scrub *sc);
 int xchk_rgsuperblock(struct xfs_scrub *sc);
+int xchk_rtrmapbt(struct xfs_scrub *sc);
 #else
 # define xchk_rtbitmap		xchk_nothing
 # define xchk_rtsummary		xchk_nothing
 # define xchk_rgsuperblock	xchk_nothing
+# define xchk_rtrmapbt		xchk_nothing
 #endif
 #ifdef CONFIG_XFS_QUOTA
 int xchk_quota(struct xfs_scrub *sc);
diff --git a/fs/xfs/scrub/stats.c b/fs/xfs/scrub/stats.c
index a476c7b2ab7597..eb6bb170c902b3 100644
--- a/fs/xfs/scrub/stats.c
+++ b/fs/xfs/scrub/stats.c
@@ -82,6 +82,7 @@ static const char *name_map[XFS_SCRUB_TYPE_NR] = {
 	[XFS_SCRUB_TYPE_DIRTREE]	= "dirtree",
 	[XFS_SCRUB_TYPE_METAPATH]	= "metapath",
 	[XFS_SCRUB_TYPE_RGSUPER]	= "rgsuper",
+	[XFS_SCRUB_TYPE_RTRMAPBT]	= "rtrmapbt",
 };
 
 /* Format the scrub stats into a text buffer, similar to pcp style. */
diff --git a/fs/xfs/scrub/trace.h b/fs/xfs/scrub/trace.h
index d2ae7e93acb08e..5afc440f22f56c 100644
--- a/fs/xfs/scrub/trace.h
+++ b/fs/xfs/scrub/trace.h
@@ -72,6 +72,7 @@ TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_DIRTREE);
 TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_BARRIER);
 TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_METAPATH);
 TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_RGSUPER);
+TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_RTRMAPBT);
 
 #define XFS_SCRUB_TYPE_STRINGS \
 	{ XFS_SCRUB_TYPE_PROBE,		"probe" }, \
@@ -105,7 +106,8 @@ TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_RGSUPER);
 	{ XFS_SCRUB_TYPE_DIRTREE,	"dirtree" }, \
 	{ XFS_SCRUB_TYPE_BARRIER,	"barrier" }, \
 	{ XFS_SCRUB_TYPE_METAPATH,	"metapath" }, \
-	{ XFS_SCRUB_TYPE_RGSUPER,	"rgsuper" }
+	{ XFS_SCRUB_TYPE_RGSUPER,	"rgsuper" }, \
+	{ XFS_SCRUB_TYPE_RTRMAPBT,	"rtrmapbt" }
 
 #define XFS_SCRUB_FLAG_STRINGS \
 	{ XFS_SCRUB_IFLAG_REPAIR,		"repair" }, \

[PATCH 22/37] xfs: cross-reference realtime bitmap to realtime rmapbt scrubber

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

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

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/scrub/rtrmap.c |   15 +++++++++++++++
 1 file changed, 15 insertions(+)


diff --git a/fs/xfs/scrub/rtrmap.c b/fs/xfs/scrub/rtrmap.c
index 7b5f932bcd947f..515c2a9b02cdae 100644
--- a/fs/xfs/scrub/rtrmap.c
+++ b/fs/xfs/scrub/rtrmap.c
@@ -142,6 +142,20 @@ xchk_rtrmapbt_check_mergeable(
 	memcpy(&cr->prev_rec, irec, sizeof(struct xfs_rmap_irec));
 }
 
+/* Cross-reference with other metadata. */
+STATIC void
+xchk_rtrmapbt_xref(
+	struct xfs_scrub	*sc,
+	struct xfs_rmap_irec	*irec)
+{
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return;
+
+	xchk_xref_is_used_rt_space(sc,
+			xfs_rgbno_to_rtb(sc->sr.rtg, irec->rm_startblock),
+			irec->rm_blockcount);
+}
+
 /* Scrub a realtime rmapbt record. */
 STATIC int
 xchk_rtrmapbt_rec(
@@ -162,6 +176,7 @@ xchk_rtrmapbt_rec(
 
 	xchk_rtrmapbt_check_mergeable(bs, cr, &irec);
 	xchk_rtrmapbt_check_overlapping(bs, cr, &irec);
+	xchk_rtrmapbt_xref(bs->sc, &irec);
 	return 0;
 }
 

[PATCH 23/37] xfs: cross-reference the realtime rmapbt

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Teach the data fork and realtime bitmap scrubbers to cross-reference
information with the realtime rmap btree.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/scrub/bmap.c     |   52 +++++++++++++++++++++++++++-----------
 fs/xfs/scrub/rgsuper.c  |    2 +
 fs/xfs/scrub/rtbitmap.c |   55 +++++++++++++++++++++++++++++++++++++---
 fs/xfs/scrub/rtbitmap.h |    5 ++++
 fs/xfs/scrub/rtrmap.c   |   65 +++++++++++++++++++++++++++++++++++++++++++++++
 fs/xfs/scrub/scrub.h    |    9 +++++++
 6 files changed, 169 insertions(+), 19 deletions(-)


diff --git a/fs/xfs/scrub/bmap.c b/fs/xfs/scrub/bmap.c
index dd99366643f832..b7f9f3b3d81a3a 100644
--- a/fs/xfs/scrub/bmap.c
+++ b/fs/xfs/scrub/bmap.c
@@ -143,15 +143,22 @@ static inline bool
 xchk_bmap_get_rmap(
 	struct xchk_bmap_info	*info,
 	struct xfs_bmbt_irec	*irec,
-	xfs_agblock_t		agbno,
+	xfs_agblock_t		bno,
 	uint64_t		owner,
 	struct xfs_rmap_irec	*rmap)
 {
+	struct xfs_btree_cur	**curp = &info->sc->sa.rmap_cur;
 	xfs_fileoff_t		offset;
 	unsigned int		rflags = 0;
 	int			has_rmap;
 	int			error;
 
+	if (xfs_ifork_is_realtime(info->sc->ip, info->whichfork))
+		curp = &info->sc->sr.rmap_cur;
+
+	if (*curp == NULL)
+		return false;
+
 	if (info->whichfork == XFS_ATTR_FORK)
 		rflags |= XFS_RMAP_ATTR_FORK;
 	if (irec->br_state == XFS_EXT_UNWRITTEN)
@@ -172,13 +179,13 @@ xchk_bmap_get_rmap(
 	 * range rmap lookup to make sure we get the correct owner/offset.
 	 */
 	if (info->is_shared) {
-		error = xfs_rmap_lookup_le_range(info->sc->sa.rmap_cur, agbno,
-				owner, offset, rflags, rmap, &has_rmap);
+		error = xfs_rmap_lookup_le_range(*curp, bno, owner, offset,
+				rflags, rmap, &has_rmap);
 	} else {
-		error = xfs_rmap_lookup_le(info->sc->sa.rmap_cur, agbno,
-				owner, offset, rflags, rmap, &has_rmap);
+		error = xfs_rmap_lookup_le(*curp, bno, owner, offset,
+				rflags, rmap, &has_rmap);
 	}
-	if (!xchk_should_check_xref(info->sc, &error, &info->sc->sa.rmap_cur))
+	if (!xchk_should_check_xref(info->sc, &error, curp))
 		return false;
 
 	if (!has_rmap)
@@ -192,29 +199,29 @@ STATIC void
 xchk_bmap_xref_rmap(
 	struct xchk_bmap_info	*info,
 	struct xfs_bmbt_irec	*irec,
-	xfs_agblock_t		agbno)
+	xfs_agblock_t		bno)
 {
 	struct xfs_rmap_irec	rmap;
 	unsigned long long	rmap_end;
 	uint64_t		owner = info->sc->ip->i_ino;
 
-	if (!info->sc->sa.rmap_cur || xchk_skip_xref(info->sc->sm))
+	if (xchk_skip_xref(info->sc->sm))
 		return;
 
 	/* Find the rmap record for this irec. */
-	if (!xchk_bmap_get_rmap(info, irec, agbno, owner, &rmap))
+	if (!xchk_bmap_get_rmap(info, irec, bno, owner, &rmap))
 		return;
 
 	/*
 	 * The rmap must be an exact match for this incore file mapping record,
 	 * which may have arisen from multiple ondisk records.
 	 */
-	if (rmap.rm_startblock != agbno)
+	if (rmap.rm_startblock != bno)
 		xchk_fblock_xref_set_corrupt(info->sc, info->whichfork,
 				irec->br_startoff);
 
 	rmap_end = (unsigned long long)rmap.rm_startblock + rmap.rm_blockcount;
-	if (rmap_end != agbno + irec->br_blockcount)
+	if (rmap_end != bno + irec->br_blockcount)
 		xchk_fblock_xref_set_corrupt(info->sc, info->whichfork,
 				irec->br_startoff);
 
@@ -259,7 +266,7 @@ STATIC void
 xchk_bmap_xref_rmap_cow(
 	struct xchk_bmap_info	*info,
 	struct xfs_bmbt_irec	*irec,
-	xfs_agblock_t		agbno)
+	xfs_agblock_t		bno)
 {
 	struct xfs_rmap_irec	rmap;
 	unsigned long long	rmap_end;
@@ -269,7 +276,7 @@ xchk_bmap_xref_rmap_cow(
 		return;
 
 	/* Find the rmap record for this irec. */
-	if (!xchk_bmap_get_rmap(info, irec, agbno, owner, &rmap))
+	if (!xchk_bmap_get_rmap(info, irec, bno, owner, &rmap))
 		return;
 
 	/*
@@ -277,12 +284,12 @@ xchk_bmap_xref_rmap_cow(
 	 * can start before and end after the physical space allocated to this
 	 * mapping.  There are no offsets to check.
 	 */
-	if (rmap.rm_startblock > agbno)
+	if (rmap.rm_startblock > bno)
 		xchk_fblock_xref_set_corrupt(info->sc, info->whichfork,
 				irec->br_startoff);
 
 	rmap_end = (unsigned long long)rmap.rm_startblock + rmap.rm_blockcount;
-	if (rmap_end < agbno + irec->br_blockcount)
+	if (rmap_end < bno + irec->br_blockcount)
 		xchk_fblock_xref_set_corrupt(info->sc, info->whichfork,
 				irec->br_startoff);
 
@@ -315,6 +322,8 @@ xchk_bmap_rt_iextent_xref(
 	struct xchk_bmap_info	*info,
 	struct xfs_bmbt_irec	*irec)
 {
+	struct xfs_owner_info	oinfo;
+	xfs_rgblock_t		rgbno;
 	int			error;
 
 	error = xchk_rtgroup_init_existing(info->sc,
@@ -332,6 +341,19 @@ xchk_bmap_rt_iextent_xref(
 	xchk_xref_is_used_rt_space(info->sc, irec->br_startblock,
 			irec->br_blockcount);
 
+	if (!xfs_has_rtrmapbt(info->sc->mp))
+		goto out_cur;
+
+	rgbno = xfs_rtb_to_rgbno(info->sc->mp, irec->br_startblock);
+	xchk_bmap_xref_rmap(info, irec, rgbno);
+
+	xfs_rmap_ino_owner(&oinfo, info->sc->ip->i_ino, info->whichfork,
+			irec->br_startoff);
+	xchk_xref_is_only_rt_owned_by(info->sc, rgbno,
+			irec->br_blockcount, &oinfo);
+
+out_cur:
+	xchk_rtgroup_btcur_free(&info->sc->sr);
 out_free:
 	xchk_rtgroup_free(info->sc, &info->sc->sr);
 }
diff --git a/fs/xfs/scrub/rgsuper.c b/fs/xfs/scrub/rgsuper.c
index e062c7d12565cd..d189732d0e24fb 100644
--- a/fs/xfs/scrub/rgsuper.c
+++ b/fs/xfs/scrub/rgsuper.c
@@ -13,6 +13,7 @@
 #include "xfs_log_format.h"
 #include "xfs_trans.h"
 #include "xfs_sb.h"
+#include "xfs_rmap.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/repair.h"
@@ -34,6 +35,7 @@ xchk_rgsuperblock_xref(
 		return;
 
 	xchk_xref_is_used_rt_space(sc, xfs_rgbno_to_rtb(sc->sr.rtg, 0), 1);
+	xchk_xref_is_only_rt_owned_by(sc, 0, 1, &XFS_RMAP_OINFO_FS);
 }
 
 int
diff --git a/fs/xfs/scrub/rtbitmap.c b/fs/xfs/scrub/rtbitmap.c
index 819026ea2d741f..675f4fdd1e675f 100644
--- a/fs/xfs/scrub/rtbitmap.c
+++ b/fs/xfs/scrub/rtbitmap.c
@@ -9,17 +9,22 @@
 #include "xfs_format.h"
 #include "xfs_trans_resv.h"
 #include "xfs_mount.h"
+#include "xfs_btree.h"
 #include "xfs_log_format.h"
 #include "xfs_trans.h"
 #include "xfs_rtbitmap.h"
 #include "xfs_inode.h"
 #include "xfs_bmap.h"
 #include "xfs_bit.h"
+#include "xfs_rtgroup.h"
 #include "xfs_sb.h"
+#include "xfs_rmap.h"
+#include "xfs_rtrmap_btree.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/repair.h"
 #include "scrub/rtbitmap.h"
+#include "scrub/btree.h"
 
 /* Set us up with the realtime metadata locked. */
 int
@@ -37,6 +42,7 @@ xchk_setup_rtbitmap(
 	if (!rtb)
 		return -ENOMEM;
 	sc->buf = rtb;
+	rtb->sc = sc;
 
 	error = xchk_rtgroup_init(sc, sc->sm->sm_agno, &sc->sr);
 	if (error)
@@ -78,7 +84,30 @@ xchk_setup_rtbitmap(
 	return 0;
 }
 
-/* Realtime bitmap. */
+/* Per-rtgroup bitmap contents. */
+
+/* Cross-reference rtbitmap entries with other metadata. */
+STATIC void
+xchk_rtbitmap_xref(
+	struct xchk_rtbitmap	*rtb,
+	xfs_rtblock_t		startblock,
+	xfs_rtblock_t		blockcount)
+{
+	struct xfs_scrub	*sc = rtb->sc;
+	xfs_rgblock_t		rgbno = xfs_rtb_to_rgbno(sc->mp, startblock);
+
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return;
+	if (!sc->sr.rmap_cur)
+		return;
+
+	xchk_xref_has_no_rt_owner(sc, rgbno, blockcount);
+
+	if (rtb->next_free_rgbno < rgbno)
+		xchk_xref_has_rt_owner(sc, rtb->next_free_rgbno,
+				rgbno - rtb->next_free_rgbno);
+	rtb->next_free_rgbno = rgbno + blockcount;
+}
 
 /* Scrub a free extent record from the realtime bitmap. */
 STATIC int
@@ -88,7 +117,8 @@ xchk_rtbitmap_rec(
 	const struct xfs_rtalloc_rec *rec,
 	void			*priv)
 {
-	struct xfs_scrub	*sc = priv;
+	struct xchk_rtbitmap	*rtb = priv;
+	struct xfs_scrub	*sc = rtb->sc;
 	xfs_rtblock_t		startblock;
 	xfs_filblks_t		blockcount;
 
@@ -97,6 +127,12 @@ xchk_rtbitmap_rec(
 
 	if (!xfs_verify_rtbext(rtg_mount(rtg), startblock, blockcount))
 		xchk_fblock_set_corrupt(sc, XFS_DATA_FORK, 0);
+
+	xchk_rtbitmap_xref(rtb, startblock, blockcount);
+
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return -ECANCELED;
+
 	return 0;
 }
 
@@ -144,7 +180,7 @@ xchk_rtbitmap_check_extents(
 	return error;
 }
 
-/* Scrub the realtime bitmap. */
+/* Scrub this group's realtime bitmap. */
 int
 xchk_rtbitmap(
 	struct xfs_scrub	*sc)
@@ -153,6 +189,7 @@ xchk_rtbitmap(
 	struct xfs_rtgroup	*rtg = sc->sr.rtg;
 	struct xfs_inode	*rbmip = rtg_bitmap(rtg);
 	struct xchk_rtbitmap	*rtb = sc->buf;
+	xfs_rgblock_t		last_rgbno;
 	int			error;
 
 	/* Is sb_rextents correct? */
@@ -205,10 +242,20 @@ xchk_rtbitmap(
 	if (error || (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT))
 		return error;
 
-	error = xfs_rtalloc_query_all(rtg, sc->tp, xchk_rtbitmap_rec, sc);
+	rtb->next_free_rgbno = 0;
+	error = xfs_rtalloc_query_all(rtg, sc->tp, xchk_rtbitmap_rec, rtb);
 	if (!xchk_fblock_process_error(sc, XFS_DATA_FORK, 0, &error))
 		return error;
 
+	/*
+	 * Check that the are rmappings for all rt extents between the end of
+	 * the last free extent we saw and the last possible extent in the rt
+	 * group.
+	 */
+	last_rgbno = rtg->rtg_extents * mp->m_sb.sb_rextsize - 1;
+	if (rtb->next_free_rgbno < last_rgbno)
+		xchk_xref_has_rt_owner(sc, rtb->next_free_rgbno,
+				last_rgbno - rtb->next_free_rgbno);
 	return 0;
 }
 
diff --git a/fs/xfs/scrub/rtbitmap.h b/fs/xfs/scrub/rtbitmap.h
index 85304ff019e1dc..dd5b394d9697d2 100644
--- a/fs/xfs/scrub/rtbitmap.h
+++ b/fs/xfs/scrub/rtbitmap.h
@@ -7,10 +7,15 @@
 #define __XFS_SCRUB_RTBITMAP_H__
 
 struct xchk_rtbitmap {
+	struct xfs_scrub	*sc;
+
 	uint64_t		rextents;
 	uint64_t		rbmblocks;
 	unsigned int		rextslog;
 	unsigned int		resblks;
+
+	/* The next free rt group block number that we expect to see. */
+	xfs_rgblock_t		next_free_rgbno;
 };
 
 #ifdef CONFIG_XFS_ONLINE_REPAIR
diff --git a/fs/xfs/scrub/rtrmap.c b/fs/xfs/scrub/rtrmap.c
index 515c2a9b02cdae..764fa296792234 100644
--- a/fs/xfs/scrub/rtrmap.c
+++ b/fs/xfs/scrub/rtrmap.c
@@ -197,3 +197,68 @@ xchk_rtrmapbt(
 	xfs_rmap_ino_bmbt_owner(&oinfo, ip->i_ino, XFS_DATA_FORK);
 	return xchk_btree(sc, sc->sr.rmap_cur, xchk_rtrmapbt_rec, &oinfo, &cr);
 }
+
+/* xref check that the extent has no realtime reverse mapping at all */
+void
+xchk_xref_has_no_rt_owner(
+	struct xfs_scrub	*sc,
+	xfs_rgblock_t		bno,
+	xfs_extlen_t		len)
+{
+	enum xbtree_recpacking	outcome;
+	int			error;
+
+	if (!sc->sr.rmap_cur || xchk_skip_xref(sc->sm))
+		return;
+
+	error = xfs_rmap_has_records(sc->sr.rmap_cur, bno, len, &outcome);
+	if (!xchk_should_check_xref(sc, &error, &sc->sr.rmap_cur))
+		return;
+	if (outcome != XBTREE_RECPACKING_EMPTY)
+		xchk_btree_xref_set_corrupt(sc, sc->sr.rmap_cur, 0);
+}
+
+/* xref check that the extent is completely mapped */
+void
+xchk_xref_has_rt_owner(
+	struct xfs_scrub	*sc,
+	xfs_rgblock_t		bno,
+	xfs_extlen_t		len)
+{
+	enum xbtree_recpacking	outcome;
+	int			error;
+
+	if (!sc->sr.rmap_cur || xchk_skip_xref(sc->sm))
+		return;
+
+	error = xfs_rmap_has_records(sc->sr.rmap_cur, bno, len, &outcome);
+	if (!xchk_should_check_xref(sc, &error, &sc->sr.rmap_cur))
+		return;
+	if (outcome != XBTREE_RECPACKING_FULL)
+		xchk_btree_xref_set_corrupt(sc, sc->sr.rmap_cur, 0);
+}
+
+/* xref check that the extent is only owned by a given owner */
+void
+xchk_xref_is_only_rt_owned_by(
+	struct xfs_scrub		*sc,
+	xfs_agblock_t			bno,
+	xfs_extlen_t			len,
+	const struct xfs_owner_info	*oinfo)
+{
+	struct xfs_rmap_matches		res;
+	int				error;
+
+	if (!sc->sr.rmap_cur || xchk_skip_xref(sc->sm))
+		return;
+
+	error = xfs_rmap_count_owners(sc->sr.rmap_cur, bno, len, oinfo, &res);
+	if (!xchk_should_check_xref(sc, &error, &sc->sr.rmap_cur))
+		return;
+	if (res.matches != 1)
+		xchk_btree_xref_set_corrupt(sc, sc->sr.rmap_cur, 0);
+	if (res.bad_non_owner_matches)
+		xchk_btree_xref_set_corrupt(sc, sc->sr.rmap_cur, 0);
+	if (res.non_owner_matches)
+		xchk_btree_xref_set_corrupt(sc, sc->sr.rmap_cur, 0);
+}
diff --git a/fs/xfs/scrub/scrub.h b/fs/xfs/scrub/scrub.h
index 0ad5122af486e1..cba4e89a3a627b 100644
--- a/fs/xfs/scrub/scrub.h
+++ b/fs/xfs/scrub/scrub.h
@@ -322,8 +322,17 @@ void xchk_xref_is_not_cow_staging(struct xfs_scrub *sc, xfs_agblock_t bno,
 #ifdef CONFIG_XFS_RT
 void xchk_xref_is_used_rt_space(struct xfs_scrub *sc, xfs_rtblock_t rtbno,
 		xfs_extlen_t len);
+void xchk_xref_has_no_rt_owner(struct xfs_scrub *sc, xfs_rgblock_t rgbno,
+		xfs_extlen_t len);
+void xchk_xref_has_rt_owner(struct xfs_scrub *sc, xfs_rgblock_t rgbno,
+		xfs_extlen_t len);
+void xchk_xref_is_only_rt_owned_by(struct xfs_scrub *sc, xfs_rgblock_t rgbno,
+		xfs_extlen_t len, const struct xfs_owner_info *oinfo);
 #else
 # define xchk_xref_is_used_rt_space(sc, rtbno, len) do { } while (0)
+# define xchk_xref_has_no_rt_owner(sc, rtbno, len) do { } while (0)
+# define xchk_xref_has_rt_owner(sc, rtbno, len) do { } while (0)
+# define xchk_xref_is_only_rt_owned_by(sc, bno, len, oinfo) do { } while (0)
 #endif
 
 #endif	/* __XFS_SCRUB_SCRUB_H__ */

[PATCH 24/37] xfs: scan rt rmap when we're doing an intense rmap check of bmbt mappings

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Teach the bmbt scrubber how to perform a comprehensive check that the
rmapbt does not contain /any/ mappings that are not described by bmbt
records when it's dealing with a realtime file.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/scrub/bmap.c |   48 ++++++++++++++++++++++++++++++++++++++++++------
 1 file changed, 42 insertions(+), 6 deletions(-)


diff --git a/fs/xfs/scrub/bmap.c b/fs/xfs/scrub/bmap.c
index b7f9f3b3d81a3a..f6077b0cba8a14 100644
--- a/fs/xfs/scrub/bmap.c
+++ b/fs/xfs/scrub/bmap.c
@@ -21,6 +21,8 @@
 #include "xfs_rmap_btree.h"
 #include "xfs_rtgroup.h"
 #include "xfs_health.h"
+#include "xfs_rtalloc.h"
+#include "xfs_rtrmap_btree.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/btree.h"
@@ -641,8 +643,7 @@ xchk_bmap_check_rmap(
 			xchk_fblock_set_corrupt(sc, sbcri->whichfork,
 					check_rec.rm_offset);
 		if (irec.br_startblock !=
-		    xfs_agbno_to_fsb(to_perag(cur->bc_group),
-				check_rec.rm_startblock))
+		    xfs_gbno_to_fsb(cur->bc_group, check_rec.rm_startblock))
 			xchk_fblock_set_corrupt(sc, sbcri->whichfork,
 					check_rec.rm_offset);
 		if (irec.br_blockcount > check_rec.rm_blockcount)
@@ -696,6 +697,30 @@ xchk_bmap_check_ag_rmaps(
 	return error;
 }
 
+/* Make sure each rt rmap has a corresponding bmbt entry. */
+STATIC int
+xchk_bmap_check_rt_rmaps(
+	struct xfs_scrub		*sc,
+	struct xfs_rtgroup		*rtg)
+{
+	struct xchk_bmap_check_rmap_info sbcri;
+	struct xfs_btree_cur		*cur;
+	int				error;
+
+	xfs_rtgroup_lock(rtg, XFS_RTGLOCK_RMAP);
+	cur = xfs_rtrmapbt_init_cursor(sc->tp, rtg);
+
+	sbcri.sc = sc;
+	sbcri.whichfork = XFS_DATA_FORK;
+	error = xfs_rmap_query_all(cur, xchk_bmap_check_rmap, &sbcri);
+	if (error == -ECANCELED)
+		error = 0;
+
+	xfs_btree_del_cursor(cur, error);
+	xfs_rtgroup_unlock(rtg, XFS_RTGLOCK_RMAP);
+	return error;
+}
+
 /*
  * Decide if we want to scan the reverse mappings to determine if the attr
  * fork /really/ has zero space mappings.
@@ -750,10 +775,6 @@ xchk_bmap_check_empty_datafork(
 {
 	struct xfs_ifork	*ifp = &ip->i_df;
 
-	/* Don't support realtime rmap checks yet. */
-	if (XFS_IS_REALTIME_INODE(ip))
-		return false;
-
 	/*
 	 * If the dinode repair found a bad data fork, it will reset the fork
 	 * to extents format with zero records and wait for the this scrubber
@@ -804,6 +825,21 @@ xchk_bmap_check_rmaps(
 	struct xfs_perag	*pag = NULL;
 	int			error;
 
+	if (xfs_ifork_is_realtime(sc->ip, whichfork)) {
+		struct xfs_rtgroup	*rtg = NULL;
+
+		while ((rtg = xfs_rtgroup_next(sc->mp, rtg))) {
+			error = xchk_bmap_check_rt_rmaps(sc, rtg);
+			if (error ||
+			    (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)) {
+				xfs_rtgroup_rele(rtg);
+				return error;
+			}
+		}
+
+		return 0;
+	}
+
 	while ((pag = xfs_perag_next(sc->mp, pag))) {
 		error = xchk_bmap_check_ag_rmaps(sc, whichfork, pag);
 		if (error ||

[PATCH 25/37] xfs: scrub the metadir path of rt rmap btree files

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Add a new XFS_SCRUB_METAPATH subtype so that we can scrub the metadata
directory tree path to the rmap btree file for each rt group.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_fs.h  |    3 ++-
 fs/xfs/scrub/metapath.c |    3 +++
 2 files changed, 5 insertions(+), 1 deletion(-)


diff --git a/fs/xfs/libxfs/xfs_fs.h b/fs/xfs/libxfs/xfs_fs.h
index 34fcbcd0bcd5e3..d42d3a5617e314 100644
--- a/fs/xfs/libxfs/xfs_fs.h
+++ b/fs/xfs/libxfs/xfs_fs.h
@@ -830,9 +830,10 @@ struct xfs_scrub_vec_head {
 #define XFS_SCRUB_METAPATH_USRQUOTA	(5)  /* user quota */
 #define XFS_SCRUB_METAPATH_GRPQUOTA	(6)  /* group quota */
 #define XFS_SCRUB_METAPATH_PRJQUOTA	(7)  /* project quota */
+#define XFS_SCRUB_METAPATH_RTRMAPBT	(8)  /* realtime reverse mapping */
 
 /* Number of metapath sm_ino values */
-#define XFS_SCRUB_METAPATH_NR		(8)
+#define XFS_SCRUB_METAPATH_NR		(9)
 
 /*
  * ioctl limits
diff --git a/fs/xfs/scrub/metapath.c b/fs/xfs/scrub/metapath.c
index c678cba1ffc3f7..74d71373e7edf1 100644
--- a/fs/xfs/scrub/metapath.c
+++ b/fs/xfs/scrub/metapath.c
@@ -21,6 +21,7 @@
 #include "xfs_trans_space.h"
 #include "xfs_attr.h"
 #include "xfs_rtgroup.h"
+#include "xfs_rtrmap_btree.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/trace.h"
@@ -246,6 +247,8 @@ xchk_setup_metapath(
 		return xchk_setup_metapath_dqinode(sc, XFS_DQTYPE_GROUP);
 	case XFS_SCRUB_METAPATH_PRJQUOTA:
 		return xchk_setup_metapath_dqinode(sc, XFS_DQTYPE_PROJ);
+	case XFS_SCRUB_METAPATH_RTRMAPBT:
+		return xchk_setup_metapath_rtginode(sc, XFS_RTGI_RMAP);
 	default:
 		return -ENOENT;
 	}

[PATCH 26/37] xfs: walk the rt reverse mapping tree when rebuilding rmap

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

When we're rebuilding the data device rmap, if we encounter an "rmap"
format fork, we have to walk the (realtime) rmap btree inode to build
the appropriate mappings.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/scrub/rmap_repair.c |   53 +++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 52 insertions(+), 1 deletion(-)


diff --git a/fs/xfs/scrub/rmap_repair.c b/fs/xfs/scrub/rmap_repair.c
index 2a0b9e3d0fbaee..91c17feb49768b 100644
--- a/fs/xfs/scrub/rmap_repair.c
+++ b/fs/xfs/scrub/rmap_repair.c
@@ -31,6 +31,8 @@
 #include "xfs_refcount.h"
 #include "xfs_refcount_btree.h"
 #include "xfs_ag.h"
+#include "xfs_rtrmap_btree.h"
+#include "xfs_rtgroup.h"
 #include "scrub/xfs_scrub.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
@@ -504,7 +506,56 @@ xrep_rmap_scan_meta_btree(
 	struct xrep_rmap_ifork	*rf,
 	struct xfs_inode	*ip)
 {
-	return -EFSCORRUPTED; /* XXX placeholder */
+	struct xfs_scrub	*sc = rf->rr->sc;
+	struct xfs_rtgroup	*rtg = NULL;
+	struct xfs_btree_cur	*cur = NULL;
+	enum xfs_rtg_inodes	type;
+	int			error;
+
+	if (rf->whichfork != XFS_DATA_FORK)
+		return -EFSCORRUPTED;
+
+	switch (ip->i_metatype) {
+	case XFS_METAFILE_RTRMAP:
+		type = XFS_RTGI_RMAP;
+		break;
+	default:
+		ASSERT(0);
+		return -EFSCORRUPTED;
+	}
+
+	while ((rtg = xfs_rtgroup_next(sc->mp, rtg))) {
+		if (ip == rtg->rtg_inodes[type])
+			goto found;
+	}
+
+	/*
+	 * We should never find an rt metadata btree inode that isn't
+	 * associated with an rtgroup yet has ondisk blocks allocated to it.
+	 */
+	if (ip->i_nblocks) {
+		ASSERT(0);
+		return -EFSCORRUPTED;
+	}
+
+	return 0;
+
+found:
+	switch (ip->i_metatype) {
+	case XFS_METAFILE_RTRMAP:
+		cur = xfs_rtrmapbt_init_cursor(sc->tp, rtg);
+		break;
+	default:
+		ASSERT(0);
+		error = -EFSCORRUPTED;
+		goto out_rtg;
+	}
+
+	error = xrep_rmap_scan_iroot_btree(rf, cur);
+	xfs_btree_del_cursor(cur, error);
+out_rtg:
+	xfs_rtgroup_rele(rtg);
+	return error;
 }
 
 /* Find all the extents from a given AG in an inode fork. */

[PATCH 27/37] xfs: online repair of realtime file bmaps

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Now that we have a reverse-mapping index of the realtime device, we can
rebuild the data fork forward-mappings of any realtime file.  Enhance
the existing bmbt repair code to walk the rtrmap btrees to gather this
information.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
---
 fs/xfs/scrub/bmap_repair.c |  128 +++++++++++++++++++++++++++++++++++++++++++-
 fs/xfs/scrub/repair.c      |   46 ++++++++++++++++
 fs/xfs/scrub/repair.h      |    2 +
 3 files changed, 172 insertions(+), 4 deletions(-)


diff --git a/fs/xfs/scrub/bmap_repair.c b/fs/xfs/scrub/bmap_repair.c
index 141d36f1da9a71..fd64bdf4e13887 100644
--- a/fs/xfs/scrub/bmap_repair.c
+++ b/fs/xfs/scrub/bmap_repair.c
@@ -25,11 +25,13 @@
 #include "xfs_bmap_btree.h"
 #include "xfs_rmap.h"
 #include "xfs_rmap_btree.h"
+#include "xfs_rtrmap_btree.h"
 #include "xfs_refcount.h"
 #include "xfs_quota.h"
 #include "xfs_ialloc.h"
 #include "xfs_ag.h"
 #include "xfs_reflink.h"
+#include "xfs_rtgroup.h"
 #include "scrub/xfs_scrub.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
@@ -359,6 +361,112 @@ xrep_bmap_scan_ag(
 	return error;
 }
 
+#ifdef CONFIG_XFS_RT
+/* Check for any obvious errors or conflicts in the file mapping. */
+STATIC int
+xrep_bmap_check_rtfork_rmap(
+	struct xfs_scrub		*sc,
+	struct xfs_btree_cur		*cur,
+	const struct xfs_rmap_irec	*rec)
+{
+	/* xattr extents are never stored on realtime devices */
+	if (rec->rm_flags & XFS_RMAP_ATTR_FORK)
+		return -EFSCORRUPTED;
+
+	/* bmbt blocks are never stored on realtime devices */
+	if (rec->rm_flags & XFS_RMAP_BMBT_BLOCK)
+		return -EFSCORRUPTED;
+
+	/* Data extents for non-rt files are never stored on the rt device. */
+	if (!XFS_IS_REALTIME_INODE(sc->ip))
+		return -EFSCORRUPTED;
+
+	/* Check the file offsets and physical extents. */
+	if (!xfs_verify_fileext(sc->mp, rec->rm_offset, rec->rm_blockcount))
+		return -EFSCORRUPTED;
+
+	/* Check that this is within the rtgroup. */
+	if (!xfs_verify_rgbext(to_rtg(cur->bc_group), rec->rm_startblock,
+				rec->rm_blockcount))
+		return -EFSCORRUPTED;
+
+	/* Make sure this isn't free space. */
+	return xrep_require_rtext_inuse(sc, rec->rm_startblock,
+			rec->rm_blockcount);
+}
+
+/* Record realtime extents that belong to this inode's fork. */
+STATIC int
+xrep_bmap_walk_rtrmap(
+	struct xfs_btree_cur		*cur,
+	const struct xfs_rmap_irec	*rec,
+	void				*priv)
+{
+	struct xrep_bmap		*rb = priv;
+	int				error = 0;
+
+	if (xchk_should_terminate(rb->sc, &error))
+		return error;
+
+	/* Skip extents which are not owned by this inode and fork. */
+	if (rec->rm_owner != rb->sc->ip->i_ino)
+		return 0;
+
+	error = xrep_bmap_check_rtfork_rmap(rb->sc, cur, rec);
+	if (error)
+		return error;
+
+	/*
+	 * Record all blocks allocated to this file even if the extent isn't
+	 * for the fork we're rebuilding so that we can reset di_nblocks later.
+	 */
+	rb->nblocks += rec->rm_blockcount;
+
+	/* If this rmap isn't for the fork we want, we're done. */
+	if (rb->whichfork == XFS_DATA_FORK &&
+	    (rec->rm_flags & XFS_RMAP_ATTR_FORK))
+		return 0;
+	if (rb->whichfork == XFS_ATTR_FORK &&
+	    !(rec->rm_flags & XFS_RMAP_ATTR_FORK))
+		return 0;
+
+	return xrep_bmap_from_rmap(rb, rec->rm_offset,
+			xfs_rgbno_to_rtb(to_rtg(cur->bc_group),
+				rec->rm_startblock),
+			rec->rm_blockcount,
+			rec->rm_flags & XFS_RMAP_UNWRITTEN);
+}
+
+/* Scan the realtime reverse mappings to build the new extent map. */
+STATIC int
+xrep_bmap_scan_rtgroup(
+	struct xrep_bmap	*rb,
+	struct xfs_rtgroup	*rtg)
+{
+	struct xfs_scrub	*sc = rb->sc;
+	int			error;
+
+	if (!xfs_has_rtrmapbt(sc->mp))
+		return 0;
+
+	error = xrep_rtgroup_init(sc, rtg, &sc->sr,
+			XFS_RTGLOCK_RMAP | XFS_RTGLOCK_BITMAP_SHARED);
+	if (error)
+		return error;
+
+	error = xfs_rmap_query_all(sc->sr.rmap_cur, xrep_bmap_walk_rtrmap, rb);
+	xchk_rtgroup_btcur_free(&sc->sr);
+	xchk_rtgroup_free(sc, &sc->sr);
+	return error;
+}
+#else
+static inline int
+xrep_bmap_scan_rtgroup(struct xrep_bmap *rb, struct xfs_rtgroup *rtg)
+{
+	return -EFSCORRUPTED;
+}
+#endif
+
 /* Find the delalloc extents from the old incore extent tree. */
 STATIC int
 xrep_bmap_find_delalloc(
@@ -410,6 +518,22 @@ xrep_bmap_find_mappings(
 	struct xfs_perag	*pag = NULL;
 	int			error = 0;
 
+	/*
+	 * Iterate the rtrmaps for extents.  Metadata files never have content
+	 * on the realtime device, so there's no need to scan them.
+	 */
+	if (!xfs_is_metadir_inode(sc->ip)) {
+		struct xfs_rtgroup	*rtg = NULL;
+
+		while ((rtg = xfs_rtgroup_next(sc->mp, rtg))) {
+			error = xrep_bmap_scan_rtgroup(rb, rtg);
+			if (error) {
+				xfs_rtgroup_rele(rtg);
+				return error;
+			}
+		}
+	}
+
 	/* Iterate the rmaps for extents. */
 	while ((pag = xfs_perag_next(sc->mp, pag))) {
 		error = xrep_bmap_scan_ag(rb, pag);
@@ -754,10 +878,6 @@ xrep_bmap_check_inputs(
 		return -EINVAL;
 	}
 
-	/* Don't know how to rebuild realtime data forks. */
-	if (XFS_IS_REALTIME_INODE(sc->ip))
-		return -EOPNOTSUPP;
-
 	return 0;
 }
 
diff --git a/fs/xfs/scrub/repair.c b/fs/xfs/scrub/repair.c
index e788e3032f8e33..18946dd46fa745 100644
--- a/fs/xfs/scrub/repair.c
+++ b/fs/xfs/scrub/repair.c
@@ -37,6 +37,9 @@
 #include "xfs_da_btree.h"
 #include "xfs_attr.h"
 #include "xfs_dir2.h"
+#include "xfs_rtrmap_btree.h"
+#include "xfs_rtbitmap.h"
+#include "xfs_rtgroup.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/trace.h"
@@ -955,6 +958,22 @@ xrep_ag_init(
 }
 
 #ifdef CONFIG_XFS_RT
+/* Initialize all the btree cursors for a RT repair. */
+static void
+xrep_rtgroup_btcur_init(
+	struct xfs_scrub	*sc,
+	struct xchk_rt		*sr)
+{
+	struct xfs_mount	*mp = sc->mp;
+
+	ASSERT(sr->rtg != NULL);
+
+	if (sc->sm->sm_type != XFS_SCRUB_TYPE_RTRMAPBT &&
+	    (sr->rtlock_flags & XFS_RTGLOCK_RMAP) &&
+	    xfs_has_rtrmapbt(mp))
+		sr->rmap_cur = xfs_rtrmapbt_init_cursor(sc->tp, sr->rtg);
+}
+
 /*
  * Given a reference to a rtgroup structure, lock rtgroup btree inodes and
  * create btree cursors.  Must only be called to repair a regular rt file.
@@ -973,6 +992,33 @@ xrep_rtgroup_init(
 
 	/* Grab our own passive reference from the caller's ref. */
 	sr->rtg = xfs_rtgroup_hold(rtg);
+	xrep_rtgroup_btcur_init(sc, sr);
+	return 0;
+}
+
+/* Ensure that all rt blocks in the given range are not marked free. */
+int
+xrep_require_rtext_inuse(
+	struct xfs_scrub	*sc,
+	xfs_rgblock_t		rgbno,
+	xfs_filblks_t		len)
+{
+	struct xfs_mount	*mp = sc->mp;
+	xfs_rtxnum_t		startrtx;
+	xfs_rtxnum_t		endrtx;
+	bool			is_free = false;
+	int			error;
+
+	startrtx = xfs_rgbno_to_rtx(mp, rgbno);
+	endrtx = xfs_rgbno_to_rtx(mp, rgbno + len - 1);
+
+	error = xfs_rtalloc_extent_is_free(sc->sr.rtg, sc->tp, startrtx,
+			endrtx - startrtx + 1, &is_free);
+	if (error)
+		return error;
+	if (is_free)
+		return -EFSCORRUPTED;
+
 	return 0;
 }
 #endif /* CONFIG_XFS_RT */
diff --git a/fs/xfs/scrub/repair.h b/fs/xfs/scrub/repair.h
index b649da1a93eb8c..584135042d9aa9 100644
--- a/fs/xfs/scrub/repair.h
+++ b/fs/xfs/scrub/repair.h
@@ -110,6 +110,8 @@ int xrep_ag_init(struct xfs_scrub *sc, struct xfs_perag *pag,
 #ifdef CONFIG_XFS_RT
 int xrep_rtgroup_init(struct xfs_scrub *sc, struct xfs_rtgroup *rtg,
 		struct xchk_rt *sr, unsigned int rtglock_flags);
+int xrep_require_rtext_inuse(struct xfs_scrub *sc, xfs_rgblock_t rgbno,
+		xfs_filblks_t len);
 #else
 # define xrep_rtgroup_init(sc, rtg, sr, lockflags)	(-ENOSYS)
 #endif /* CONFIG_XFS_RT */

Re: [PATCH 27/37] xfs: online repair of realtime file bmaps

[Christoph Hellwig]

Looks good:

Reviewed-by: Christoph Hellwig <hch@lst.de>

[PATCH 28/37] xfs: repair inodes that have realtime extents

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Plumb into the inode core repair code the ability to search for extents
on realtime devices.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/scrub/inode_repair.c |   58 ++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 57 insertions(+), 1 deletion(-)


diff --git a/fs/xfs/scrub/inode_repair.c b/fs/xfs/scrub/inode_repair.c
index a94f9df0ca78f6..816e81330ffc99 100644
--- a/fs/xfs/scrub/inode_repair.c
+++ b/fs/xfs/scrub/inode_repair.c
@@ -38,6 +38,8 @@
 #include "xfs_log_priv.h"
 #include "xfs_health.h"
 #include "xfs_symlink_remote.h"
+#include "xfs_rtgroup.h"
+#include "xfs_rtrmap_btree.h"
 #include "scrub/xfs_scrub.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
@@ -773,17 +775,71 @@ xrep_dinode_count_ag_rmaps(
 	return error;
 }
 
+/* Count extents and blocks for an inode given an rt rmap. */
+STATIC int
+xrep_dinode_walk_rtrmap(
+	struct xfs_btree_cur		*cur,
+	const struct xfs_rmap_irec	*rec,
+	void				*priv)
+{
+	struct xrep_inode		*ri = priv;
+	int				error = 0;
+
+	if (xchk_should_terminate(ri->sc, &error))
+		return error;
+
+	/* We only care about this inode. */
+	if (rec->rm_owner != ri->sc->sm->sm_ino)
+		return 0;
+
+	if (rec->rm_flags & (XFS_RMAP_ATTR_FORK | XFS_RMAP_BMBT_BLOCK))
+		return -EFSCORRUPTED;
+
+	ri->rt_blocks += rec->rm_blockcount;
+	ri->rt_extents++;
+	return 0;
+}
+
+/* Count extents and blocks for an inode from all realtime rmap data. */
+STATIC int
+xrep_dinode_count_rtgroup_rmaps(
+	struct xrep_inode	*ri,
+	struct xfs_rtgroup	*rtg)
+{
+	struct xfs_scrub	*sc = ri->sc;
+	int			error;
+
+	error = xrep_rtgroup_init(sc, rtg, &sc->sr, XFS_RTGLOCK_RMAP);
+	if (error)
+		return error;
+
+	error = xfs_rmap_query_all(sc->sr.rmap_cur, xrep_dinode_walk_rtrmap,
+			ri);
+	xchk_rtgroup_btcur_free(&sc->sr);
+	xchk_rtgroup_free(sc, &sc->sr);
+	return error;
+}
+
 /* Count extents and blocks for a given inode from all rmap data. */
 STATIC int
 xrep_dinode_count_rmaps(
 	struct xrep_inode	*ri)
 {
 	struct xfs_perag	*pag = NULL;
+	struct xfs_rtgroup	*rtg = NULL;
 	int			error;
 
-	if (!xfs_has_rmapbt(ri->sc->mp) || xfs_has_realtime(ri->sc->mp))
+	if (!xfs_has_rmapbt(ri->sc->mp))
 		return -EOPNOTSUPP;
 
+	while ((rtg = xfs_rtgroup_next(ri->sc->mp, rtg))) {
+		error = xrep_dinode_count_rtgroup_rmaps(ri, rtg);
+		if (error) {
+			xfs_rtgroup_rele(rtg);
+			return error;
+		}
+	}
+
 	while ((pag = xfs_perag_next(ri->sc->mp, pag))) {
 		error = xrep_dinode_count_ag_rmaps(ri, pag);
 		if (error) {

[PATCH 29/37] xfs: repair rmap btree inodes

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Teach the inode repair code how to deal with realtime rmap btree inodes
that won't load properly.  This is most likely moot since the filesystem
generally won't mount without the rtrmapbt inodes being usable, but
we'll add this for completeness.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/scrub/inode_repair.c |   35 +++++++++++++++++++++++++++++++++++
 1 file changed, 35 insertions(+)


diff --git a/fs/xfs/scrub/inode_repair.c b/fs/xfs/scrub/inode_repair.c
index 816e81330ffc99..d7e3f033b16073 100644
--- a/fs/xfs/scrub/inode_repair.c
+++ b/fs/xfs/scrub/inode_repair.c
@@ -944,6 +944,34 @@ xrep_dinode_bad_bmbt_fork(
 	return false;
 }
 
+/* Return true if this rmap-format ifork looks like garbage. */
+STATIC bool
+xrep_dinode_bad_rtrmapbt_fork(
+	struct xfs_scrub	*sc,
+	struct xfs_dinode	*dip,
+	unsigned int		dfork_size)
+{
+	struct xfs_rtrmap_root	*dfp;
+	unsigned int		nrecs;
+	unsigned int		level;
+
+	if (dfork_size < sizeof(struct xfs_rtrmap_root))
+		return true;
+
+	dfp = XFS_DFORK_PTR(dip, XFS_DATA_FORK);
+	nrecs = be16_to_cpu(dfp->bb_numrecs);
+	level = be16_to_cpu(dfp->bb_level);
+
+	if (level > sc->mp->m_rtrmap_maxlevels)
+		return true;
+	if (xfs_rtrmap_droot_space_calc(level, nrecs) > dfork_size)
+		return true;
+	if (level > 0 && nrecs == 0)
+		return true;
+
+	return false;
+}
+
 /* Check a metadata-btree fork. */
 STATIC bool
 xrep_dinode_bad_metabt_fork(
@@ -956,6 +984,8 @@ xrep_dinode_bad_metabt_fork(
 		return true;
 
 	switch (be16_to_cpu(dip->di_metatype)) {
+	case XFS_METAFILE_RTRMAP:
+		return xrep_dinode_bad_rtrmapbt_fork(sc, dip, dfork_size);
 	default:
 		return true;
 	}
@@ -1220,6 +1250,7 @@ xrep_dinode_ensure_forkoff(
 	uint16_t		mode)
 {
 	struct xfs_bmdr_block	*bmdr;
+	struct xfs_rtrmap_root	*rmdr;
 	struct xfs_scrub	*sc = ri->sc;
 	xfs_extnum_t		attr_extents, data_extents;
 	size_t			bmdr_minsz = xfs_bmdr_space_calc(1);
@@ -1328,6 +1359,10 @@ xrep_dinode_ensure_forkoff(
 		break;
 	case XFS_DINODE_FMT_META_BTREE:
 		switch (be16_to_cpu(dip->di_metatype)) {
+		case XFS_METAFILE_RTRMAP:
+			rmdr = XFS_DFORK_PTR(dip, XFS_DATA_FORK);
+			dfork_min = xfs_rtrmap_broot_space(sc->mp, rmdr);
+			break;
 		default:
 			dfork_min = 0;
 			break;

[PATCH 30/37] xfs: online repair of realtime bitmaps for a realtime group

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

For a given rt group, regenerate the bitmap contents from the group's
realtime rmap btree.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_rtbitmap.h    |    9 +
 fs/xfs/scrub/common.h           |    6 +
 fs/xfs/scrub/repair.c           |    2 
 fs/xfs/scrub/repair.h           |    1 
 fs/xfs/scrub/rtbitmap.c         |    5 
 fs/xfs/scrub/rtbitmap.h         |   50 +++++
 fs/xfs/scrub/rtbitmap_repair.c  |  429 ++++++++++++++++++++++++++++++++++++++-
 fs/xfs/scrub/rtsummary_repair.c |    3 
 fs/xfs/scrub/tempexch.h         |    2 
 fs/xfs/scrub/tempfile.c         |   20 +-
 fs/xfs/scrub/trace.c            |    1 
 fs/xfs/scrub/trace.h            |  150 ++++++++++++++
 12 files changed, 659 insertions(+), 19 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_rtbitmap.h b/fs/xfs/libxfs/xfs_rtbitmap.h
index 16563a44bd138a..22e5d9cd95f47c 100644
--- a/fs/xfs/libxfs/xfs_rtbitmap.h
+++ b/fs/xfs/libxfs/xfs_rtbitmap.h
@@ -135,6 +135,15 @@ xfs_rtb_to_rtx(
 	return div_u64(rtbno, mp->m_sb.sb_rextsize);
 }
 
+/* Return the offset of a rtgroup block number within an rt extent. */
+static inline xfs_extlen_t
+xfs_rgbno_to_rtxoff(
+	struct xfs_mount	*mp,
+	xfs_rgblock_t		rgbno)
+{
+	return rgbno % mp->m_sb.sb_rextsize;
+}
+
 /* Return the offset of an rt block number within an rt extent. */
 static inline xfs_extlen_t
 xfs_rtb_to_rtxoff(
diff --git a/fs/xfs/scrub/common.h b/fs/xfs/scrub/common.h
index 1576467f724431..e5891609af2740 100644
--- a/fs/xfs/scrub/common.h
+++ b/fs/xfs/scrub/common.h
@@ -264,6 +264,12 @@ int xchk_metadata_inode_forks(struct xfs_scrub *sc);
 			(sc)->mp->m_super->s_id, \
 			(sc)->ip ? (sc)->ip->i_ino : (sc)->sm->sm_ino, \
 			##__VA_ARGS__)
+#define xchk_xfile_rtgroup_descr(sc, fmt, ...) \
+	kasprintf(XCHK_GFP_FLAGS, "XFS (%s): rtgroup 0x%x " fmt, \
+			(sc)->mp->m_super->s_id, \
+			(sc)->sa.pag ? \
+				rtg_rgno((sc)->sr.rtg) : (sc)->sm->sm_agno, \
+			##__VA_ARGS__)
 
 /*
  * Setting up a hook to wait for intents to drain is costly -- we have to take
diff --git a/fs/xfs/scrub/repair.c b/fs/xfs/scrub/repair.c
index 18946dd46fa745..82fe01d78cb08d 100644
--- a/fs/xfs/scrub/repair.c
+++ b/fs/xfs/scrub/repair.c
@@ -959,7 +959,7 @@ xrep_ag_init(
 
 #ifdef CONFIG_XFS_RT
 /* Initialize all the btree cursors for a RT repair. */
-static void
+void
 xrep_rtgroup_btcur_init(
 	struct xfs_scrub	*sc,
 	struct xchk_rt		*sr)
diff --git a/fs/xfs/scrub/repair.h b/fs/xfs/scrub/repair.h
index 584135042d9aa9..7f493752ea78e6 100644
--- a/fs/xfs/scrub/repair.h
+++ b/fs/xfs/scrub/repair.h
@@ -110,6 +110,7 @@ int xrep_ag_init(struct xfs_scrub *sc, struct xfs_perag *pag,
 #ifdef CONFIG_XFS_RT
 int xrep_rtgroup_init(struct xfs_scrub *sc, struct xfs_rtgroup *rtg,
 		struct xchk_rt *sr, unsigned int rtglock_flags);
+void xrep_rtgroup_btcur_init(struct xfs_scrub *sc, struct xchk_rt *sr);
 int xrep_require_rtext_inuse(struct xfs_scrub *sc, xfs_rgblock_t rgbno,
 		xfs_filblks_t len);
 #else
diff --git a/fs/xfs/scrub/rtbitmap.c b/fs/xfs/scrub/rtbitmap.c
index 675f4fdd1e675f..28c90a31f4c32b 100644
--- a/fs/xfs/scrub/rtbitmap.c
+++ b/fs/xfs/scrub/rtbitmap.c
@@ -20,9 +20,11 @@
 #include "xfs_sb.h"
 #include "xfs_rmap.h"
 #include "xfs_rtrmap_btree.h"
+#include "xfs_exchmaps.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/repair.h"
+#include "scrub/tempexch.h"
 #include "scrub/rtbitmap.h"
 #include "scrub/btree.h"
 
@@ -38,7 +40,8 @@ xchk_setup_rtbitmap(
 	if (xchk_need_intent_drain(sc))
 		xchk_fsgates_enable(sc, XCHK_FSGATES_DRAIN);
 
-	rtb = kzalloc(sizeof(struct xchk_rtbitmap), XCHK_GFP_FLAGS);
+	rtb = kzalloc(struct_size(rtb, words, xchk_rtbitmap_wordcnt(sc)),
+			XCHK_GFP_FLAGS);
 	if (!rtb)
 		return -ENOMEM;
 	sc->buf = rtb;
diff --git a/fs/xfs/scrub/rtbitmap.h b/fs/xfs/scrub/rtbitmap.h
index dd5b394d9697d2..fe52b877253d35 100644
--- a/fs/xfs/scrub/rtbitmap.h
+++ b/fs/xfs/scrub/rtbitmap.h
@@ -6,6 +6,20 @@
 #ifndef __XFS_SCRUB_RTBITMAP_H__
 #define __XFS_SCRUB_RTBITMAP_H__
 
+/*
+ * We use an xfile to construct new bitmap blocks for the portion of the
+ * rtbitmap file that we're replacing.  Whereas the ondisk bitmap must be
+ * accessed through the buffer cache, the xfile bitmap supports direct
+ * word-level accesses.  Therefore, we create a small abstraction for linear
+ * access.
+ */
+typedef unsigned long long xrep_wordoff_t;
+typedef unsigned int xrep_wordcnt_t;
+
+/* Mask to round an rtx down to the nearest bitmap word. */
+#define XREP_RTBMP_WORDMASK	((1ULL << XFS_NBWORDLOG) - 1)
+
+
 struct xchk_rtbitmap {
 	struct xfs_scrub	*sc;
 
@@ -16,12 +30,48 @@ struct xchk_rtbitmap {
 
 	/* The next free rt group block number that we expect to see. */
 	xfs_rgblock_t		next_free_rgbno;
+
+#ifdef CONFIG_XFS_ONLINE_REPAIR
+	/* stuff for staging a new bitmap */
+	struct xfs_rtalloc_args	args;
+	struct xrep_tempexch	tempexch;
+#endif
+
+	/* The next rtgroup block we expect to see during our rtrmapbt walk. */
+	xfs_rgblock_t		next_rgbno;
+
+	/* rtgroup lock flags */
+	unsigned int		rtglock_flags;
+
+	/* rtword position of xfile as we write buffers to disk. */
+	xrep_wordoff_t		prep_wordoff;
+
+	/* In-Memory rtbitmap for repair. */
+	union xfs_rtword_raw	words[];
 };
 
 #ifdef CONFIG_XFS_ONLINE_REPAIR
 int xrep_setup_rtbitmap(struct xfs_scrub *sc, struct xchk_rtbitmap *rtb);
+
+/*
+ * How big should the words[] buffer be?
+ *
+ * For repairs, we want a full fsblock worth of space so that we can memcpy a
+ * buffer full of 1s into the xfile bitmap.  The xfile bitmap doesn't have
+ * rtbitmap block headers, so we don't use blockwsize.  Scrub doesn't use the
+ * words buffer at all.
+ */
+static inline unsigned int
+xchk_rtbitmap_wordcnt(
+	struct xfs_scrub	*sc)
+{
+	if (xchk_could_repair(sc))
+		return sc->mp->m_sb.sb_blocksize >> XFS_WORDLOG;
+	return 0;
+}
 #else
 # define xrep_setup_rtbitmap(sc, rtb)	(0)
+# define xchk_rtbitmap_wordcnt(sc)	(0)
 #endif /* CONFIG_XFS_ONLINE_REPAIR */
 
 #endif /* __XFS_SCRUB_RTBITMAP_H__ */
diff --git a/fs/xfs/scrub/rtbitmap_repair.c b/fs/xfs/scrub/rtbitmap_repair.c
index 0fef98e9f83409..c6e33834c5ae98 100644
--- a/fs/xfs/scrub/rtbitmap_repair.c
+++ b/fs/xfs/scrub/rtbitmap_repair.c
@@ -12,32 +12,65 @@
 #include "xfs_btree.h"
 #include "xfs_log_format.h"
 #include "xfs_trans.h"
+#include "xfs_rtalloc.h"
 #include "xfs_inode.h"
 #include "xfs_bit.h"
 #include "xfs_bmap.h"
 #include "xfs_bmap_btree.h"
+#include "xfs_rmap.h"
+#include "xfs_rtrmap_btree.h"
+#include "xfs_exchmaps.h"
+#include "xfs_rtbitmap.h"
+#include "xfs_rtgroup.h"
+#include "xfs_extent_busy.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/trace.h"
 #include "scrub/repair.h"
 #include "scrub/xfile.h"
+#include "scrub/tempfile.h"
+#include "scrub/tempexch.h"
+#include "scrub/reap.h"
 #include "scrub/rtbitmap.h"
 
-/* Set up to repair the realtime bitmap file metadata. */
+/* rt bitmap content repairs */
+
+/* Set up to repair the realtime bitmap for this group. */
 int
 xrep_setup_rtbitmap(
 	struct xfs_scrub	*sc,
 	struct xchk_rtbitmap	*rtb)
 {
 	struct xfs_mount	*mp = sc->mp;
-	unsigned long long	blocks = 0;
+	char			*descr;
+	unsigned long long	blocks = mp->m_sb.sb_rbmblocks;
+	int			error;
+
+	error = xrep_tempfile_create(sc, S_IFREG);
+	if (error)
+		return error;
+
+	/* Create an xfile to hold our reconstructed bitmap. */
+	descr = xchk_xfile_rtgroup_descr(sc, "bitmap file");
+	error = xfile_create(descr, blocks * mp->m_sb.sb_blocksize, &sc->xfile);
+	kfree(descr);
+	if (error)
+		return error;
 
 	/*
-	 * Reserve enough blocks to write out a completely new bmbt for a
-	 * maximally fragmented bitmap file.  We do not hold the rtbitmap
-	 * ILOCK yet, so this is entirely speculative.
+	 * Reserve enough blocks to write out a completely new bitmap file,
+	 * plus twice as many blocks as we would need if we can only allocate
+	 * one block per data fork mapping.  This should cover the
+	 * preallocation of the temporary file and exchanging the extent
+	 * mappings.
+	 *
+	 * We cannot use xfs_exchmaps_estimate because we have not yet
+	 * constructed the replacement bitmap and therefore do not know how
+	 * many extents it will use.  By the time we do, we will have a dirty
+	 * transaction (which we cannot drop because we cannot drop the
+	 * rtbitmap ILOCK) and cannot ask for more reservation.
 	 */
-	blocks = xfs_bmbt_calc_size(mp, mp->m_sb.sb_rbmblocks);
+	blocks += xfs_bmbt_calc_size(mp, blocks) * 2;
 	if (blocks > UINT_MAX)
 		return -EOPNOTSUPP;
 
@@ -45,6 +78,304 @@ xrep_setup_rtbitmap(
 	return 0;
 }
 
+static inline xrep_wordoff_t
+rtx_to_wordoff(
+	struct xfs_mount	*mp,
+	xfs_rtxnum_t		rtx)
+{
+	return rtx >> XFS_NBWORDLOG;
+}
+
+static inline xrep_wordcnt_t
+rtxlen_to_wordcnt(
+	xfs_rtxlen_t	rtxlen)
+{
+	return rtxlen >> XFS_NBWORDLOG;
+}
+
+/* Helper functions to record rtwords in an xfile. */
+
+static inline int
+xfbmp_load(
+	struct xchk_rtbitmap	*rtb,
+	xrep_wordoff_t		wordoff,
+	xfs_rtword_t		*word)
+{
+	union xfs_rtword_raw	urk;
+	int			error;
+
+	ASSERT(xfs_has_rtgroups(rtb->sc->mp));
+
+	error = xfile_load(rtb->sc->xfile, &urk,
+			sizeof(union xfs_rtword_raw),
+			wordoff << XFS_WORDLOG);
+	if (error)
+		return error;
+
+	*word = be32_to_cpu(urk.rtg);
+	return 0;
+}
+
+static inline int
+xfbmp_store(
+	struct xchk_rtbitmap	*rtb,
+	xrep_wordoff_t		wordoff,
+	const xfs_rtword_t	word)
+{
+	union xfs_rtword_raw	urk;
+
+	ASSERT(xfs_has_rtgroups(rtb->sc->mp));
+
+	urk.rtg = cpu_to_be32(word);
+	return xfile_store(rtb->sc->xfile, &urk,
+			sizeof(union xfs_rtword_raw),
+			wordoff << XFS_WORDLOG);
+}
+
+static inline int
+xfbmp_copyin(
+	struct xchk_rtbitmap	*rtb,
+	xrep_wordoff_t		wordoff,
+	const union xfs_rtword_raw	*word,
+	xrep_wordcnt_t		nr_words)
+{
+	return xfile_store(rtb->sc->xfile, word, nr_words << XFS_WORDLOG,
+			wordoff << XFS_WORDLOG);
+}
+
+static inline int
+xfbmp_copyout(
+	struct xchk_rtbitmap	*rtb,
+	xrep_wordoff_t		wordoff,
+	union xfs_rtword_raw	*word,
+	xrep_wordcnt_t		nr_words)
+{
+	return xfile_load(rtb->sc->xfile, word, nr_words << XFS_WORDLOG,
+			wordoff << XFS_WORDLOG);
+}
+
+/* Perform a logical OR operation on an rtword in the incore bitmap. */
+static int
+xrep_rtbitmap_or(
+	struct xchk_rtbitmap	*rtb,
+	xrep_wordoff_t		wordoff,
+	xfs_rtword_t		mask)
+{
+	xfs_rtword_t		word;
+	int			error;
+
+	error = xfbmp_load(rtb, wordoff, &word);
+	if (error)
+		return error;
+
+	trace_xrep_rtbitmap_or(rtb->sc->mp, wordoff, mask, word);
+
+	return xfbmp_store(rtb, wordoff, word | mask);
+}
+
+/*
+ * Mark as free every rt extent between the next rt block we expected to see
+ * in the rtrmap records and the given rt block.
+ */
+STATIC int
+xrep_rtbitmap_mark_free(
+	struct xchk_rtbitmap	*rtb,
+	xfs_rgblock_t		rgbno)
+{
+	struct xfs_mount	*mp = rtb->sc->mp;
+	struct xfs_rtgroup	*rtg = rtb->sc->sr.rtg;
+	xfs_rtxnum_t		startrtx;
+	xfs_rtxnum_t		nextrtx;
+	xrep_wordoff_t		wordoff, nextwordoff;
+	unsigned int		bit;
+	unsigned int		bufwsize;
+	xfs_extlen_t		mod;
+	xfs_rtword_t		mask;
+	int			error;
+
+	if (!xfs_verify_rgbext(rtg, rtb->next_rgbno, rgbno - rtb->next_rgbno))
+		return -EFSCORRUPTED;
+
+	/*
+	 * Convert rt blocks to rt extents  The block range we find must be
+	 * aligned to an rtextent boundary on both ends.
+	 */
+	startrtx = xfs_rgbno_to_rtx(mp, rtb->next_rgbno);
+	mod = xfs_rgbno_to_rtxoff(mp, rtb->next_rgbno);
+	if (mod)
+		return -EFSCORRUPTED;
+
+	nextrtx = xfs_rgbno_to_rtx(mp, rgbno - 1) + 1;
+	mod = xfs_rgbno_to_rtxoff(mp, rgbno - 1);
+	if (mod != mp->m_sb.sb_rextsize - 1)
+		return -EFSCORRUPTED;
+
+	trace_xrep_rtbitmap_record_free(mp, startrtx, nextrtx - 1);
+
+	/* Set bits as needed to round startrtx up to the nearest word. */
+	bit = startrtx & XREP_RTBMP_WORDMASK;
+	if (bit) {
+		xfs_rtblock_t	len = nextrtx - startrtx;
+		unsigned int	lastbit;
+
+		lastbit = min(bit + len, XFS_NBWORD);
+		mask = (((xfs_rtword_t)1 << (lastbit - bit)) - 1) << bit;
+
+		error = xrep_rtbitmap_or(rtb, rtx_to_wordoff(mp, startrtx),
+				mask);
+		if (error || lastbit - bit == len)
+			return error;
+		startrtx += XFS_NBWORD - bit;
+	}
+
+	/* Set bits as needed to round nextrtx down to the nearest word. */
+	bit = nextrtx & XREP_RTBMP_WORDMASK;
+	if (bit) {
+		mask = ((xfs_rtword_t)1 << bit) - 1;
+
+		error = xrep_rtbitmap_or(rtb, rtx_to_wordoff(mp, nextrtx),
+				mask);
+		if (error || startrtx + bit == nextrtx)
+			return error;
+		nextrtx -= bit;
+	}
+
+	trace_xrep_rtbitmap_record_free_bulk(mp, startrtx, nextrtx - 1);
+
+	/* Set all the words in between, up to a whole fs block at once. */
+	wordoff = rtx_to_wordoff(mp, startrtx);
+	nextwordoff = rtx_to_wordoff(mp, nextrtx);
+	bufwsize = mp->m_sb.sb_blocksize >> XFS_WORDLOG;
+
+	while (wordoff < nextwordoff) {
+		xrep_wordoff_t	rem;
+		xrep_wordcnt_t	wordcnt;
+
+		wordcnt = min_t(xrep_wordcnt_t, nextwordoff - wordoff,
+				bufwsize);
+
+		/*
+		 * Try to keep us aligned to the rtwords buffer to reduce the
+		 * number of xfile writes.
+		 */
+		rem = wordoff & (bufwsize - 1);
+		if (rem)
+			wordcnt = min_t(xrep_wordcnt_t, wordcnt,
+					bufwsize - rem);
+
+		error = xfbmp_copyin(rtb, wordoff, rtb->words, wordcnt);
+		if (error)
+			return error;
+
+		wordoff += wordcnt;
+	}
+
+	return 0;
+}
+
+/* Set free space in the rtbitmap based on rtrmapbt records. */
+STATIC int
+xrep_rtbitmap_walk_rtrmap(
+	struct xfs_btree_cur		*cur,
+	const struct xfs_rmap_irec	*rec,
+	void				*priv)
+{
+	struct xchk_rtbitmap		*rtb = priv;
+	int				error = 0;
+
+	if (xchk_should_terminate(rtb->sc, &error))
+		return error;
+
+	if (rtb->next_rgbno < rec->rm_startblock) {
+		error = xrep_rtbitmap_mark_free(rtb, rec->rm_startblock);
+		if (error)
+			return error;
+	}
+
+	rtb->next_rgbno = max(rtb->next_rgbno,
+			      rec->rm_startblock + rec->rm_blockcount);
+	return 0;
+}
+
+/*
+ * Walk the rtrmapbt to find all the gaps between records, and mark the gaps
+ * in the realtime bitmap that we're computing.
+ */
+STATIC int
+xrep_rtbitmap_find_freespace(
+	struct xchk_rtbitmap	*rtb)
+{
+	struct xfs_scrub	*sc = rtb->sc;
+	struct xfs_mount	*mp = sc->mp;
+	struct xfs_rtgroup	*rtg = sc->sr.rtg;
+	uint64_t		blockcount;
+	int			error;
+
+	/* Prepare a buffer of ones so that we can accelerate bulk setting. */
+	memset(rtb->words, 0xFF, mp->m_sb.sb_blocksize);
+
+	xrep_rtgroup_btcur_init(sc, &sc->sr);
+	error = xfs_rmap_query_all(sc->sr.rmap_cur, xrep_rtbitmap_walk_rtrmap,
+			rtb);
+	if (error)
+		goto out;
+
+	/*
+	 * Mark as free every possible rt extent from the last one we saw to
+	 * the end of the rt group.
+	 */
+	blockcount = rtg->rtg_extents * mp->m_sb.sb_rextsize;
+	if (rtb->next_rgbno < blockcount) {
+		error = xrep_rtbitmap_mark_free(rtb, blockcount);
+		if (error)
+			goto out;
+	}
+
+out:
+	xchk_rtgroup_btcur_free(&sc->sr);
+	return error;
+}
+
+static int
+xrep_rtbitmap_prep_buf(
+	struct xfs_scrub	*sc,
+	struct xfs_buf		*bp,
+	void			*data)
+{
+	struct xchk_rtbitmap	*rtb = data;
+	struct xfs_mount	*mp = sc->mp;
+	union xfs_rtword_raw	*ondisk;
+	int			error;
+
+	rtb->args.mp = sc->mp;
+	rtb->args.tp = sc->tp;
+	rtb->args.rbmbp = bp;
+	ondisk = xfs_rbmblock_wordptr(&rtb->args, 0);
+	rtb->args.rbmbp = NULL;
+
+	error = xfbmp_copyout(rtb, rtb->prep_wordoff, ondisk,
+			mp->m_blockwsize);
+	if (error)
+		return error;
+
+	if (xfs_has_rtgroups(sc->mp)) {
+		struct xfs_rtbuf_blkinfo	*hdr = bp->b_addr;
+
+		hdr->rt_magic = cpu_to_be32(XFS_RTBITMAP_MAGIC);
+		hdr->rt_owner = cpu_to_be64(sc->ip->i_ino);
+		hdr->rt_blkno = cpu_to_be64(xfs_buf_daddr(bp));
+		hdr->rt_lsn = 0;
+		uuid_copy(&hdr->rt_uuid, &sc->mp->m_sb.sb_meta_uuid);
+		bp->b_ops = &xfs_rtbitmap_buf_ops;
+	} else {
+		bp->b_ops = &xfs_rtbuf_ops;
+	}
+
+	rtb->prep_wordoff += mp->m_blockwsize;
+	xfs_trans_buf_set_type(sc->tp, bp, XFS_BLFT_RTBITMAP_BUF);
+	return 0;
+}
+
 /*
  * Make sure that the given range of the data fork of the realtime file is
  * mapped to written blocks.  The caller must ensure that the inode is joined
@@ -160,9 +491,18 @@ xrep_rtbitmap(
 {
 	struct xchk_rtbitmap	*rtb = sc->buf;
 	struct xfs_mount	*mp = sc->mp;
+	struct xfs_group	*xg = rtg_group(sc->sr.rtg);
 	unsigned long long	blocks = 0;
+	unsigned int		busy_gen;
 	int			error;
 
+	/* We require the realtime rmapbt to rebuild anything. */
+	if (!xfs_has_rtrmapbt(sc->mp))
+		return -EOPNOTSUPP;
+	/* We require atomic file exchange range to rebuild anything. */
+	if (!xfs_has_exchange_range(sc->mp))
+		return -EOPNOTSUPP;
+
 	/* Impossibly large rtbitmap means we can't touch the filesystem. */
 	if (rtb->rbmblocks > U32_MAX)
 		return 0;
@@ -195,6 +535,79 @@ xrep_rtbitmap(
 	if (error)
 		return error;
 
-	/* Fix inconsistent bitmap geometry */
-	return xrep_rtbitmap_geometry(sc, rtb);
+	/*
+	 * Fix inconsistent bitmap geometry.  This function returns with a
+	 * clean scrub transaction.
+	 */
+	error = xrep_rtbitmap_geometry(sc, rtb);
+	if (error)
+		return error;
+
+	/*
+	 * Make sure the busy extent list is clear because we can't put extents
+	 * on there twice.
+	 */
+	if (!xfs_extent_busy_list_empty(xg, &busy_gen)) {
+		error = xfs_extent_busy_flush(sc->tp, xg, busy_gen, 0);
+		if (error)
+			return error;
+	}
+
+	/*
+	 * Generate the new rtbitmap data.  We don't need the rtbmp information
+	 * once this call is finished.
+	 */
+	error = xrep_rtbitmap_find_freespace(rtb);
+	if (error)
+		return error;
+
+	/*
+	 * Try to take ILOCK_EXCL of the temporary file.  We had better be the
+	 * only ones holding onto this inode, but we can't block while holding
+	 * the rtbitmap file's ILOCK_EXCL.
+	 */
+	while (!xrep_tempfile_ilock_nowait(sc)) {
+		if (xchk_should_terminate(sc, &error))
+			return error;
+		delay(1);
+	}
+
+	/*
+	 * Make sure we have space allocated for the part of the bitmap
+	 * file that corresponds to this group.  We already joined sc->ip.
+	 */
+	xfs_trans_ijoin(sc->tp, sc->tempip, 0);
+	error = xrep_tempfile_prealloc(sc, 0, rtb->rbmblocks);
+	if (error)
+		return error;
+
+	/* Last chance to abort before we start committing fixes. */
+	if (xchk_should_terminate(sc, &error))
+		return error;
+
+	/* Copy the bitmap file that we generated. */
+	error = xrep_tempfile_copyin(sc, 0, rtb->rbmblocks,
+			xrep_rtbitmap_prep_buf, rtb);
+	if (error)
+		return error;
+	error = xrep_tempfile_set_isize(sc,
+			XFS_FSB_TO_B(sc->mp, sc->mp->m_sb.sb_rbmblocks));
+	if (error)
+		return error;
+
+	/*
+	 * Now exchange the data fork contents.  We're done with the temporary
+	 * buffer, so we can reuse it for the tempfile exchmaps information.
+	 */
+	error = xrep_tempexch_trans_reserve(sc, XFS_DATA_FORK, 0,
+			rtb->rbmblocks, &rtb->tempexch);
+	if (error)
+		return error;
+
+	error = xrep_tempexch_contents(sc, &rtb->tempexch);
+	if (error)
+		return error;
+
+	/* Free the old rtbitmap blocks if they're not in use. */
+	return xrep_reap_ifork(sc, sc->tempip, XFS_DATA_FORK);
 }
diff --git a/fs/xfs/scrub/rtsummary_repair.c b/fs/xfs/scrub/rtsummary_repair.c
index 8198ea84ad70e5..d593977d70df21 100644
--- a/fs/xfs/scrub/rtsummary_repair.c
+++ b/fs/xfs/scrub/rtsummary_repair.c
@@ -165,7 +165,8 @@ xrep_rtsummary(
 	 * Now exchange the contents.  Nothing in repair uses the temporary
 	 * buffer, so we can reuse it for the tempfile exchrange information.
 	 */
-	error = xrep_tempexch_trans_reserve(sc, XFS_DATA_FORK, &rts->tempexch);
+	error = xrep_tempexch_trans_reserve(sc, XFS_DATA_FORK, 0,
+			rts->rsumblocks, &rts->tempexch);
 	if (error)
 		return error;
 
diff --git a/fs/xfs/scrub/tempexch.h b/fs/xfs/scrub/tempexch.h
index 995ba187c5aa62..eccda720c2ca40 100644
--- a/fs/xfs/scrub/tempexch.h
+++ b/fs/xfs/scrub/tempexch.h
@@ -12,7 +12,7 @@ struct xrep_tempexch {
 };
 
 int xrep_tempexch_trans_reserve(struct xfs_scrub *sc, int whichfork,
-		struct xrep_tempexch *ti);
+		xfs_fileoff_t off, xfs_filblks_t len, struct xrep_tempexch *ti);
 int xrep_tempexch_trans_alloc(struct xfs_scrub *sc, int whichfork,
 		struct xrep_tempexch *ti);
 
diff --git a/fs/xfs/scrub/tempfile.c b/fs/xfs/scrub/tempfile.c
index 4ebb5f8459e8f3..cf99e0ca51b008 100644
--- a/fs/xfs/scrub/tempfile.c
+++ b/fs/xfs/scrub/tempfile.c
@@ -606,6 +606,8 @@ STATIC int
 xrep_tempexch_prep_request(
 	struct xfs_scrub	*sc,
 	int			whichfork,
+	xfs_fileoff_t		off,
+	xfs_filblks_t		len,
 	struct xrep_tempexch	*tx)
 {
 	struct xfs_exchmaps_req	*req = &tx->req;
@@ -629,18 +631,19 @@ xrep_tempexch_prep_request(
 	/* Exchange all mappings in both forks. */
 	req->ip1 = sc->tempip;
 	req->ip2 = sc->ip;
-	req->startoff1 = 0;
-	req->startoff2 = 0;
+	req->startoff1 = off;
+	req->startoff2 = off;
 	switch (whichfork) {
 	case XFS_ATTR_FORK:
 		req->flags |= XFS_EXCHMAPS_ATTR_FORK;
 		break;
 	case XFS_DATA_FORK:
-		/* Always exchange sizes when exchanging data fork mappings. */
-		req->flags |= XFS_EXCHMAPS_SET_SIZES;
+		/* Exchange sizes when exchanging all data fork mappings. */
+		if (off == 0 && len == XFS_MAX_FILEOFF)
+			req->flags |= XFS_EXCHMAPS_SET_SIZES;
 		break;
 	}
-	req->blockcount = XFS_MAX_FILEOFF;
+	req->blockcount = len;
 
 	return 0;
 }
@@ -796,6 +799,8 @@ int
 xrep_tempexch_trans_reserve(
 	struct xfs_scrub	*sc,
 	int			whichfork,
+	xfs_fileoff_t		off,
+	xfs_filblks_t		len,
 	struct xrep_tempexch	*tx)
 {
 	int			error;
@@ -804,7 +809,7 @@ xrep_tempexch_trans_reserve(
 	xfs_assert_ilocked(sc->ip, XFS_ILOCK_EXCL);
 	xfs_assert_ilocked(sc->tempip, XFS_ILOCK_EXCL);
 
-	error = xrep_tempexch_prep_request(sc, whichfork, tx);
+	error = xrep_tempexch_prep_request(sc, whichfork, off, len, tx);
 	if (error)
 		return error;
 
@@ -842,7 +847,8 @@ xrep_tempexch_trans_alloc(
 	ASSERT(sc->tp == NULL);
 	ASSERT(xfs_has_exchange_range(sc->mp));
 
-	error = xrep_tempexch_prep_request(sc, whichfork, tx);
+	error = xrep_tempexch_prep_request(sc, whichfork, 0, XFS_MAX_FILEOFF,
+			tx);
 	if (error)
 		return error;
 
diff --git a/fs/xfs/scrub/trace.c b/fs/xfs/scrub/trace.c
index 98f923ae664d0e..2450e214103fed 100644
--- a/fs/xfs/scrub/trace.c
+++ b/fs/xfs/scrub/trace.c
@@ -21,6 +21,7 @@
 #include "xfs_rmap.h"
 #include "xfs_parent.h"
 #include "xfs_metafile.h"
+#include "xfs_rtgroup.h"
 #include "scrub/scrub.h"
 #include "scrub/xfile.h"
 #include "scrub/xfarray.h"
diff --git a/fs/xfs/scrub/trace.h b/fs/xfs/scrub/trace.h
index 5afc440f22f56c..3b661e4443453c 100644
--- a/fs/xfs/scrub/trace.h
+++ b/fs/xfs/scrub/trace.h
@@ -17,6 +17,7 @@
 #include "xfs_bit.h"
 #include "xfs_quota_defs.h"
 
+struct xfs_rtgroup;
 struct xfs_scrub;
 struct xfile;
 struct xfarray;
@@ -3607,6 +3608,155 @@ DEFINE_XCHK_METAPATH_EVENT(xrep_metapath_try_unlink);
 DEFINE_XCHK_METAPATH_EVENT(xrep_metapath_unlink);
 DEFINE_XCHK_METAPATH_EVENT(xrep_metapath_link);
 
+#ifdef CONFIG_XFS_RT
+DECLARE_EVENT_CLASS(xrep_rtbitmap_class,
+	TP_PROTO(struct xfs_mount *mp, xfs_rtxnum_t start, xfs_rtxnum_t end),
+	TP_ARGS(mp, start, end),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(dev_t, rtdev)
+		__field(xfs_rtxnum_t, start)
+		__field(xfs_rtxnum_t, end)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->rtdev = mp->m_rtdev_targp->bt_dev;
+		__entry->start = start;
+		__entry->end = end;
+	),
+	TP_printk("dev %d:%d rtdev %d:%d startrtx 0x%llx endrtx 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  MAJOR(__entry->rtdev), MINOR(__entry->rtdev),
+		  __entry->start,
+		  __entry->end)
+);
+#define DEFINE_REPAIR_RGBITMAP_EVENT(name) \
+DEFINE_EVENT(xrep_rtbitmap_class, name, \
+	TP_PROTO(struct xfs_mount *mp, xfs_rtxnum_t start, \
+		 xfs_rtxnum_t end), \
+	TP_ARGS(mp, start, end))
+DEFINE_REPAIR_RGBITMAP_EVENT(xrep_rtbitmap_record_free);
+DEFINE_REPAIR_RGBITMAP_EVENT(xrep_rtbitmap_record_free_bulk);
+
+TRACE_EVENT(xrep_rtbitmap_or,
+	TP_PROTO(struct xfs_mount *mp, unsigned long long wordoff,
+		 xfs_rtword_t mask, xfs_rtword_t word),
+	TP_ARGS(mp, wordoff, mask, word),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(dev_t, rtdev)
+		__field(unsigned long long, wordoff)
+		__field(unsigned int, mask)
+		__field(unsigned int, word)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->rtdev = mp->m_rtdev_targp->bt_dev;
+		__entry->wordoff = wordoff;
+		__entry->mask = mask;
+		__entry->word = word;
+	),
+	TP_printk("dev %d:%d rtdev %d:%d wordoff 0x%llx mask 0x%x word 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  MAJOR(__entry->rtdev), MINOR(__entry->rtdev),
+		  __entry->wordoff,
+		  __entry->mask,
+		  __entry->word)
+);
+
+TRACE_EVENT(xrep_rtbitmap_load,
+	TP_PROTO(struct xfs_rtgroup *rtg, xfs_fileoff_t rbmoff,
+		 xfs_rtxnum_t rtx, xfs_rtxnum_t len),
+	TP_ARGS(rtg, rbmoff, rtx, len),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(dev_t, rtdev)
+		__field(xfs_rgnumber_t, rgno)
+		__field(xfs_fileoff_t, rbmoff)
+		__field(xfs_rtxnum_t, rtx)
+		__field(xfs_rtxnum_t, len)
+	),
+	TP_fast_assign(
+		__entry->dev = rtg_mount(rtg)->m_super->s_dev;
+		__entry->rtdev = rtg_mount(rtg)->m_rtdev_targp->bt_dev;
+		__entry->rgno = rtg_rgno(rtg);
+		__entry->rbmoff = rbmoff;
+		__entry->rtx = rtx;
+		__entry->len = len;
+	),
+	TP_printk("dev %d:%d rtdev %d:%d rgno 0x%x rbmoff 0x%llx rtx 0x%llx rtxcount 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  MAJOR(__entry->rtdev), MINOR(__entry->rtdev),
+		  __entry->rgno,
+		  __entry->rbmoff,
+		  __entry->rtx,
+		  __entry->len)
+);
+
+TRACE_EVENT(xrep_rtbitmap_load_words,
+	TP_PROTO(struct xfs_mount *mp, xfs_fileoff_t rbmoff,
+		 unsigned long long wordoff, unsigned int wordcnt),
+	TP_ARGS(mp, rbmoff, wordoff, wordcnt),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(dev_t, rtdev)
+		__field(xfs_fileoff_t, rbmoff)
+		__field(unsigned long long, wordoff)
+		__field(unsigned int, wordcnt)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->rtdev = mp->m_rtdev_targp->bt_dev;
+		__entry->rbmoff = rbmoff;
+		__entry->wordoff = wordoff;
+		__entry->wordcnt = wordcnt;
+	),
+	TP_printk("dev %d:%d rtdev %d:%d rbmoff 0x%llx wordoff 0x%llx wordcnt 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  MAJOR(__entry->rtdev), MINOR(__entry->rtdev),
+		  __entry->rbmoff,
+		  __entry->wordoff,
+		  __entry->wordcnt)
+);
+
+TRACE_EVENT(xrep_rtbitmap_load_word,
+	TP_PROTO(struct xfs_mount *mp, unsigned long long wordoff,
+		 unsigned int bit, xfs_rtword_t ondisk_word,
+		 xfs_rtword_t xfile_word, xfs_rtword_t word_mask),
+	TP_ARGS(mp, wordoff, bit, ondisk_word, xfile_word, word_mask),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(dev_t, rtdev)
+		__field(unsigned long long, wordoff)
+		__field(unsigned int, bit)
+		__field(xfs_rtword_t, ondisk_word)
+		__field(xfs_rtword_t, xfile_word)
+		__field(xfs_rtword_t, word_mask)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->rtdev = mp->m_rtdev_targp->bt_dev;
+		__entry->wordoff = wordoff;
+		__entry->bit = bit;
+		__entry->ondisk_word = ondisk_word;
+		__entry->xfile_word = xfile_word;
+		__entry->word_mask = word_mask;
+	),
+	TP_printk("dev %d:%d rtdev %d:%d wordoff 0x%llx bit %u ondisk 0x%x(0x%x) inmem 0x%x(0x%x) result 0x%x mask 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  MAJOR(__entry->rtdev), MINOR(__entry->rtdev),
+		  __entry->wordoff,
+		  __entry->bit,
+		  __entry->ondisk_word,
+		  __entry->ondisk_word & __entry->word_mask,
+		  __entry->xfile_word,
+		  __entry->xfile_word & ~__entry->word_mask,
+		  (__entry->xfile_word & ~__entry->word_mask) |
+		  (__entry->ondisk_word & __entry->word_mask),
+		  __entry->word_mask)
+);
+#endif /* CONFIG_XFS_RT */
+
 #endif /* IS_ENABLED(CONFIG_XFS_ONLINE_REPAIR) */
 
 #endif /* _TRACE_XFS_SCRUB_TRACE_H */

[PATCH 31/37] xfs: support repairing metadata btrees rooted in metadir inodes

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Adapt the repair code so that we can stage a new btree in the data fork
area of a metadir inode and reap the old blocks.  We already have nearly
all of the infrastructure; the only parts that were missing were the
metadata inode reservation handling.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/scrub/newbt.c |   42 ++++++++++++++++++++++++++++++++++++++++++
 fs/xfs/scrub/newbt.h |    1 +
 fs/xfs/scrub/reap.c  |   41 +++++++++++++++++++++++++++++++++++++++++
 fs/xfs/scrub/reap.h  |    2 ++
 4 files changed, 86 insertions(+)


diff --git a/fs/xfs/scrub/newbt.c b/fs/xfs/scrub/newbt.c
index 70af27d987342f..ac38f584309029 100644
--- a/fs/xfs/scrub/newbt.c
+++ b/fs/xfs/scrub/newbt.c
@@ -19,6 +19,8 @@
 #include "xfs_rmap.h"
 #include "xfs_ag.h"
 #include "xfs_defer.h"
+#include "xfs_metafile.h"
+#include "xfs_quota.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/trace.h"
@@ -120,6 +122,43 @@ xrep_newbt_init_inode(
 	return 0;
 }
 
+/*
+ * Initialize accounting resources for staging a new metadata inode btree.
+ * If the metadata file has a space reservation, the caller must adjust that
+ * reservation when committing the new ondisk btree.
+ */
+int
+xrep_newbt_init_metadir_inode(
+	struct xrep_newbt		*xnr,
+	struct xfs_scrub		*sc)
+{
+	struct xfs_owner_info		oinfo;
+	struct xfs_ifork		*ifp;
+
+	ASSERT(xfs_is_metadir_inode(sc->ip));
+
+	xfs_rmap_ino_bmbt_owner(&oinfo, sc->ip->i_ino, XFS_DATA_FORK);
+
+	ifp = kmem_cache_zalloc(xfs_ifork_cache, XCHK_GFP_FLAGS);
+	if (!ifp)
+		return -ENOMEM;
+
+	/*
+	 * Allocate new metadir btree blocks with XFS_AG_RESV_NONE because the
+	 * inode metadata space reservations can only account allocated space
+	 * to the i_nblocks.  We do not want to change the inode core fields
+	 * until we're ready to commit the new tree, so we allocate the blocks
+	 * as if they were regular file blocks.  This exposes us to a higher
+	 * risk of the repair being cancelled due to ENOSPC.
+	 */
+	xrep_newbt_init_ag(xnr, sc, &oinfo,
+			XFS_INO_TO_FSB(sc->mp, sc->ip->i_ino),
+			XFS_AG_RESV_NONE);
+	xnr->ifake.if_fork = ifp;
+	xnr->ifake.if_fork_size = xfs_inode_fork_size(sc->ip, XFS_DATA_FORK);
+	return 0;
+}
+
 /*
  * Initialize accounting resources for staging a new btree.  Callers are
  * expected to add their own reservations (and clean them up) manually.
@@ -224,6 +263,7 @@ xrep_newbt_alloc_ag_blocks(
 	int			error = 0;
 
 	ASSERT(sc->sa.pag != NULL);
+	ASSERT(xnr->resv != XFS_AG_RESV_METAFILE);
 
 	while (nr_blocks > 0) {
 		struct xfs_alloc_arg	args = {
@@ -297,6 +337,8 @@ xrep_newbt_alloc_file_blocks(
 	struct xfs_mount	*mp = sc->mp;
 	int			error = 0;
 
+	ASSERT(xnr->resv != XFS_AG_RESV_METAFILE);
+
 	while (nr_blocks > 0) {
 		struct xfs_alloc_arg	args = {
 			.tp		= sc->tp,
diff --git a/fs/xfs/scrub/newbt.h b/fs/xfs/scrub/newbt.h
index 3d804d31af24a8..5ce785599287be 100644
--- a/fs/xfs/scrub/newbt.h
+++ b/fs/xfs/scrub/newbt.h
@@ -63,6 +63,7 @@ void xrep_newbt_init_ag(struct xrep_newbt *xnr, struct xfs_scrub *sc,
 		enum xfs_ag_resv_type resv);
 int xrep_newbt_init_inode(struct xrep_newbt *xnr, struct xfs_scrub *sc,
 		int whichfork, const struct xfs_owner_info *oinfo);
+int xrep_newbt_init_metadir_inode(struct xrep_newbt *xnr, struct xfs_scrub *sc);
 int xrep_newbt_alloc_blocks(struct xrep_newbt *xnr, uint64_t nr_blocks);
 int xrep_newbt_add_extent(struct xrep_newbt *xnr, struct xfs_perag *pag,
 		xfs_agblock_t agbno, xfs_extlen_t len);
diff --git a/fs/xfs/scrub/reap.c b/fs/xfs/scrub/reap.c
index 08230952053b7d..4d7f1b82dc559d 100644
--- a/fs/xfs/scrub/reap.c
+++ b/fs/xfs/scrub/reap.c
@@ -33,6 +33,7 @@
 #include "xfs_attr.h"
 #include "xfs_attr_remote.h"
 #include "xfs_defer.h"
+#include "xfs_metafile.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/trace.h"
@@ -390,6 +391,8 @@ xreap_agextent_iter(
 	xfs_fsblock_t		fsbno;
 	int			error = 0;
 
+	ASSERT(rs->resv != XFS_AG_RESV_METAFILE);
+
 	fsbno = xfs_agbno_to_fsb(sc->sa.pag, agbno);
 
 	/*
@@ -675,6 +678,44 @@ xrep_reap_fsblocks(
 	return 0;
 }
 
+/*
+ * Dispose of every block of an old metadata btree that used to be rooted in a
+ * metadata directory file.
+ */
+int
+xrep_reap_metadir_fsblocks(
+	struct xfs_scrub		*sc,
+	struct xfsb_bitmap		*bitmap)
+{
+	/*
+	 * Reap old metadir btree blocks with XFS_AG_RESV_NONE because the old
+	 * blocks are no longer mapped by the inode, and inode metadata space
+	 * reservations can only account freed space to the i_nblocks.
+	 */
+	struct xfs_owner_info		oinfo;
+	struct xreap_state		rs = {
+		.sc			= sc,
+		.oinfo			= &oinfo,
+		.resv			= XFS_AG_RESV_NONE,
+	};
+	int				error;
+
+	ASSERT(xfs_has_rmapbt(sc->mp));
+	ASSERT(sc->ip != NULL);
+	ASSERT(xfs_is_metadir_inode(sc->ip));
+
+	xfs_rmap_ino_bmbt_owner(&oinfo, sc->ip->i_ino, XFS_DATA_FORK);
+
+	error = xfsb_bitmap_walk(bitmap, xreap_fsmeta_extent, &rs);
+	if (error)
+		return error;
+
+	if (xreap_dirty(&rs))
+		return xrep_defer_finish(sc);
+
+	return 0;
+}
+
 /*
  * Metadata files are not supposed to share blocks with anything else.
  * If blocks are shared, we remove the reverse mapping (thus reducing the
diff --git a/fs/xfs/scrub/reap.h b/fs/xfs/scrub/reap.h
index 3f2f1775e29db4..70e5e6bbb8d38d 100644
--- a/fs/xfs/scrub/reap.h
+++ b/fs/xfs/scrub/reap.h
@@ -14,6 +14,8 @@ int xrep_reap_agblocks(struct xfs_scrub *sc, struct xagb_bitmap *bitmap,
 int xrep_reap_fsblocks(struct xfs_scrub *sc, struct xfsb_bitmap *bitmap,
 		const struct xfs_owner_info *oinfo);
 int xrep_reap_ifork(struct xfs_scrub *sc, struct xfs_inode *ip, int whichfork);
+int xrep_reap_metadir_fsblocks(struct xfs_scrub *sc,
+		struct xfsb_bitmap *bitmap);
 
 /* Buffer cache scan context. */
 struct xrep_bufscan {

[PATCH 32/37] xfs: online repair of the realtime rmap btree

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Repair the realtime rmap btree while mounted.  Similar to the regular
rmap btree repair code, we walk the data fork mappings of every realtime
file in the filesystem to collect reverse-mapping records in an xfarray.
Then we sort the xfarray, and use the btree bulk loader to create a new
rtrmap btree ondisk.  Finally, we swap the btree roots, and reap the old
blocks in the usual way.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
---
 fs/xfs/Makefile                   |    1 
 fs/xfs/libxfs/xfs_btree_staging.c |    1 
 fs/xfs/libxfs/xfs_rtrmap_btree.c  |    2 
 fs/xfs/libxfs/xfs_rtrmap_btree.h  |    3 
 fs/xfs/scrub/common.c             |    4 
 fs/xfs/scrub/common.h             |    2 
 fs/xfs/scrub/repair.c             |  144 +++++++
 fs/xfs/scrub/repair.h             |   14 +
 fs/xfs/scrub/rtrmap.c             |    7 
 fs/xfs/scrub/rtrmap_repair.c      |  733 +++++++++++++++++++++++++++++++++++++
 fs/xfs/scrub/scrub.c              |    2 
 fs/xfs/scrub/trace.h              |   57 +++
 12 files changed, 966 insertions(+), 4 deletions(-)
 create mode 100644 fs/xfs/scrub/rtrmap_repair.c


diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile
index 136a465e00d2b1..338e10f81b7b71 100644
--- a/fs/xfs/Makefile
+++ b/fs/xfs/Makefile
@@ -234,6 +234,7 @@ xfs-y				+= $(addprefix scrub/, \
 
 xfs-$(CONFIG_XFS_RT)		+= $(addprefix scrub/, \
 				   rtbitmap_repair.o \
+				   rtrmap_repair.o \
 				   rtsummary_repair.o \
 				   )
 
diff --git a/fs/xfs/libxfs/xfs_btree_staging.c b/fs/xfs/libxfs/xfs_btree_staging.c
index 58c146b5c9d479..5ed84f9cc877ef 100644
--- a/fs/xfs/libxfs/xfs_btree_staging.c
+++ b/fs/xfs/libxfs/xfs_btree_staging.c
@@ -134,6 +134,7 @@ xfs_btree_stage_ifakeroot(
 	cur->bc_ino.ifake = ifake;
 	cur->bc_nlevels = ifake->if_levels;
 	cur->bc_ino.forksize = ifake->if_fork_size;
+	cur->bc_ino.whichfork = XFS_STAGING_FORK;
 	cur->bc_flags |= XFS_BTREE_STAGING;
 }
 
diff --git a/fs/xfs/libxfs/xfs_rtrmap_btree.c b/fs/xfs/libxfs/xfs_rtrmap_btree.c
index 0a78dee01b1b2e..571a9e1b956099 100644
--- a/fs/xfs/libxfs/xfs_rtrmap_btree.c
+++ b/fs/xfs/libxfs/xfs_rtrmap_btree.c
@@ -660,7 +660,7 @@ xfs_rtrmapbt_compute_maxlevels(
 }
 
 /* Calculate the rtrmap btree size for some records. */
-static unsigned long long
+unsigned long long
 xfs_rtrmapbt_calc_size(
 	struct xfs_mount	*mp,
 	unsigned long long	len)
diff --git a/fs/xfs/libxfs/xfs_rtrmap_btree.h b/fs/xfs/libxfs/xfs_rtrmap_btree.h
index bf73460be274d1..ad76ac7938b602 100644
--- a/fs/xfs/libxfs/xfs_rtrmap_btree.h
+++ b/fs/xfs/libxfs/xfs_rtrmap_btree.h
@@ -198,4 +198,7 @@ int xfs_rtrmapbt_create(struct xfs_rtgroup *rtg, struct xfs_inode *ip,
 int xfs_rtrmapbt_init_rtsb(struct xfs_mount *mp, struct xfs_rtgroup *rtg,
 		struct xfs_trans *tp);
 
+unsigned long long xfs_rtrmapbt_calc_size(struct xfs_mount *mp,
+		unsigned long long len);
+
 #endif /* __XFS_RTRMAP_BTREE_H__ */
diff --git a/fs/xfs/scrub/common.c b/fs/xfs/scrub/common.c
index ca43dd4f52b2d6..06cb61e6349827 100644
--- a/fs/xfs/scrub/common.c
+++ b/fs/xfs/scrub/common.c
@@ -819,7 +819,7 @@ xchk_rtgroup_btcur_free(
  * Unlock the realtime group.  This must be done /after/ committing (or
  * cancelling) the scrub transaction.
  */
-static void
+void
 xchk_rtgroup_unlock(
 	struct xchk_rt		*sr)
 {
@@ -904,7 +904,7 @@ int
 xchk_setup_rt(
 	struct xfs_scrub	*sc)
 {
-	return xchk_trans_alloc(sc, 0);
+	return xchk_trans_alloc(sc, xrep_calc_rtgroup_resblks(sc));
 }
 
 /* Set us up with AG headers and btree cursors. */
diff --git a/fs/xfs/scrub/common.h b/fs/xfs/scrub/common.h
index e5891609af2740..50ac6cca18fe45 100644
--- a/fs/xfs/scrub/common.h
+++ b/fs/xfs/scrub/common.h
@@ -147,12 +147,14 @@ xchk_rtgroup_init_existing(
 
 int xchk_rtgroup_lock(struct xfs_scrub *sc, struct xchk_rt *sr,
 		unsigned int rtglock_flags);
+void xchk_rtgroup_unlock(struct xchk_rt *sr);
 void xchk_rtgroup_btcur_free(struct xchk_rt *sr);
 void xchk_rtgroup_free(struct xfs_scrub *sc, struct xchk_rt *sr);
 #else
 # define xchk_rtgroup_init(sc, rgno, sr)		(-EFSCORRUPTED)
 # define xchk_rtgroup_init_existing(sc, rgno, sr)	(-EFSCORRUPTED)
 # define xchk_rtgroup_lock(sc, sr, lockflags)		(-EFSCORRUPTED)
+# define xchk_rtgroup_unlock(sr)			do { } while (0)
 # define xchk_rtgroup_btcur_free(sr)			do { } while (0)
 # define xchk_rtgroup_free(sc, sr)			do { } while (0)
 #endif /* CONFIG_XFS_RT */
diff --git a/fs/xfs/scrub/repair.c b/fs/xfs/scrub/repair.c
index 82fe01d78cb08d..61e414c81253af 100644
--- a/fs/xfs/scrub/repair.c
+++ b/fs/xfs/scrub/repair.c
@@ -40,6 +40,8 @@
 #include "xfs_rtrmap_btree.h"
 #include "xfs_rtbitmap.h"
 #include "xfs_rtgroup.h"
+#include "xfs_rtalloc.h"
+#include "xfs_metafile.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/trace.h"
@@ -382,6 +384,41 @@ xrep_calc_ag_resblks(
 	return max(max(bnobt_sz, inobt_sz), max(rmapbt_sz, refcbt_sz));
 }
 
+#ifdef CONFIG_XFS_RT
+/*
+ * Figure out how many blocks to reserve for a rtgroup repair.  We calculate
+ * the worst case estimate for the number of blocks we'd need to rebuild one of
+ * any type of per-rtgroup btree.
+ */
+xfs_extlen_t
+xrep_calc_rtgroup_resblks(
+	struct xfs_scrub		*sc)
+{
+	struct xfs_mount		*mp = sc->mp;
+	struct xfs_scrub_metadata	*sm = sc->sm;
+	uint64_t			usedlen;
+	xfs_extlen_t			rmapbt_sz = 0;
+
+	if (!(sm->sm_flags & XFS_SCRUB_IFLAG_REPAIR))
+		return 0;
+	if (!xfs_has_rtgroups(mp)) {
+		ASSERT(0);
+		return -EFSCORRUPTED;
+	}
+
+	usedlen = xfs_rtbxlen_to_blen(mp, xfs_rtgroup_extents(mp, sm->sm_agno));
+	ASSERT(usedlen <= XFS_MAX_RGBLOCKS);
+
+	if (xfs_has_rmapbt(mp))
+		rmapbt_sz = xfs_rtrmapbt_calc_size(mp, usedlen);
+
+	trace_xrep_calc_rtgroup_resblks_btsize(mp, sm->sm_agno, usedlen,
+			rmapbt_sz);
+
+	return rmapbt_sz;
+}
+#endif /* CONFIG_XFS_RT */
+
 /*
  * Reconstructing per-AG Btrees
  *
@@ -1284,3 +1321,110 @@ xrep_buf_verify_struct(
 
 	return fa == NULL;
 }
+
+/* Check the sanity of a rmap record for a metadata btree inode. */
+int
+xrep_check_ino_btree_mapping(
+	struct xfs_scrub		*sc,
+	const struct xfs_rmap_irec	*rec)
+{
+	enum xbtree_recpacking		outcome;
+	int				error;
+
+	/*
+	 * Metadata btree inodes never have extended attributes, and all blocks
+	 * should have the bmbt block flag set.
+	 */
+	if ((rec->rm_flags & XFS_RMAP_ATTR_FORK) ||
+	    !(rec->rm_flags & XFS_RMAP_BMBT_BLOCK))
+		return -EFSCORRUPTED;
+
+	/* Make sure the block is within the AG. */
+	if (!xfs_verify_agbext(sc->sa.pag, rec->rm_startblock,
+				rec->rm_blockcount))
+		return -EFSCORRUPTED;
+
+	/* Make sure this isn't free space. */
+	error = xfs_alloc_has_records(sc->sa.bno_cur, rec->rm_startblock,
+			rec->rm_blockcount, &outcome);
+	if (error)
+		return error;
+	if (outcome != XBTREE_RECPACKING_EMPTY)
+		return -EFSCORRUPTED;
+
+	return 0;
+}
+
+/*
+ * Reset the block count of the inode being repaired, and adjust the dquot
+ * block usage to match.  The inode must not have an xattr fork.
+ */
+void
+xrep_inode_set_nblocks(
+	struct xfs_scrub	*sc,
+	int64_t			new_blocks)
+{
+	int64_t			delta =
+		new_blocks - sc->ip->i_nblocks;
+
+	sc->ip->i_nblocks = new_blocks;
+
+	xfs_trans_log_inode(sc->tp, sc->ip, XFS_ILOG_CORE);
+	if (delta != 0)
+		xfs_trans_mod_dquot_byino(sc->tp, sc->ip, XFS_TRANS_DQ_BCOUNT,
+				delta);
+}
+
+/* Reset the block reservation for a metadata inode. */
+int
+xrep_reset_metafile_resv(
+	struct xfs_scrub	*sc)
+{
+	struct xfs_inode	*ip = sc->ip;
+	int64_t			delta;
+	int			error;
+
+	delta = ip->i_nblocks + ip->i_delayed_blks - ip->i_meta_resv_asked;
+	if (delta == 0)
+		return 0;
+
+	/*
+	 * Too many blocks have been reserved, transfer some from the incore
+	 * reservation back to the filesystem.
+	 */
+	if (delta > 0) {
+		int64_t		give_back;
+
+		give_back = min_t(uint64_t, delta, ip->i_delayed_blks);
+		if (give_back > 0) {
+			xfs_mod_delalloc(ip, 0, -give_back);
+			xfs_add_fdblocks(ip->i_mount, give_back);
+			ip->i_delayed_blks -= give_back;
+		}
+
+		return 0;
+	}
+
+	/*
+	 * Not enough reservation; try to take some blocks from the filesystem
+	 * to the metadata inode.  @delta is negative here, so invert the sign.
+	 */
+	delta = -delta;
+	error = xfs_dec_fdblocks(sc->mp, delta, true);
+	while (error == -ENOSPC) {
+		delta--;
+		if (delta == 0) {
+			xfs_warn(sc->mp,
+"Insufficient free space to reset space reservation for inode 0x%llx after repair.",
+					ip->i_ino);
+			return 0;
+		}
+		error = xfs_dec_fdblocks(sc->mp, delta, true);
+	}
+	if (error)
+		return error;
+
+	xfs_mod_delalloc(ip, 0, delta);
+	ip->i_delayed_blks += delta;
+	return 0;
+}
diff --git a/fs/xfs/scrub/repair.h b/fs/xfs/scrub/repair.h
index 7f493752ea78e6..ac5962732d269d 100644
--- a/fs/xfs/scrub/repair.h
+++ b/fs/xfs/scrub/repair.h
@@ -97,6 +97,7 @@ int xrep_setup_parent(struct xfs_scrub *sc);
 int xrep_setup_nlinks(struct xfs_scrub *sc);
 int xrep_setup_symlink(struct xfs_scrub *sc, unsigned int *resblks);
 int xrep_setup_dirtree(struct xfs_scrub *sc);
+int xrep_setup_rtrmapbt(struct xfs_scrub *sc);
 
 /* Repair setup functions */
 int xrep_setup_ag_allocbt(struct xfs_scrub *sc);
@@ -113,10 +114,15 @@ int xrep_rtgroup_init(struct xfs_scrub *sc, struct xfs_rtgroup *rtg,
 void xrep_rtgroup_btcur_init(struct xfs_scrub *sc, struct xchk_rt *sr);
 int xrep_require_rtext_inuse(struct xfs_scrub *sc, xfs_rgblock_t rgbno,
 		xfs_filblks_t len);
+xfs_extlen_t xrep_calc_rtgroup_resblks(struct xfs_scrub *sc);
 #else
 # define xrep_rtgroup_init(sc, rtg, sr, lockflags)	(-ENOSYS)
+# define xrep_calc_rtgroup_resblks(sc)			(0)
 #endif /* CONFIG_XFS_RT */
 
+int xrep_check_ino_btree_mapping(struct xfs_scrub *sc,
+		const struct xfs_rmap_irec *rec);
+
 /* Metadata revalidators */
 
 int xrep_revalidate_allocbt(struct xfs_scrub *sc);
@@ -150,10 +156,12 @@ int xrep_metapath(struct xfs_scrub *sc);
 int xrep_rtbitmap(struct xfs_scrub *sc);
 int xrep_rtsummary(struct xfs_scrub *sc);
 int xrep_rgsuperblock(struct xfs_scrub *sc);
+int xrep_rtrmapbt(struct xfs_scrub *sc);
 #else
 # define xrep_rtbitmap			xrep_notsupported
 # define xrep_rtsummary			xrep_notsupported
 # define xrep_rgsuperblock		xrep_notsupported
+# define xrep_rtrmapbt			xrep_notsupported
 #endif /* CONFIG_XFS_RT */
 
 #ifdef CONFIG_XFS_QUOTA
@@ -172,6 +180,8 @@ int xrep_trans_alloc_hook_dummy(struct xfs_mount *mp, void **cookiep,
 void xrep_trans_cancel_hook_dummy(void **cookiep, struct xfs_trans *tp);
 
 bool xrep_buf_verify_struct(struct xfs_buf *bp, const struct xfs_buf_ops *ops);
+void xrep_inode_set_nblocks(struct xfs_scrub *sc, int64_t new_blocks);
+int xrep_reset_metafile_resv(struct xfs_scrub *sc);
 
 #else
 
@@ -195,6 +205,8 @@ xrep_calc_ag_resblks(
 	return 0;
 }
 
+#define xrep_calc_rtgroup_resblks	xrep_calc_ag_resblks
+
 static inline int
 xrep_reset_perag_resv(
 	struct xfs_scrub	*sc)
@@ -222,6 +234,7 @@ xrep_setup_nothing(
 #define xrep_setup_nlinks		xrep_setup_nothing
 #define xrep_setup_dirtree		xrep_setup_nothing
 #define xrep_setup_metapath		xrep_setup_nothing
+#define xrep_setup_rtrmapbt		xrep_setup_nothing
 
 #define xrep_setup_inode(sc, imap)	((void)0)
 
@@ -259,6 +272,7 @@ static inline int xrep_setup_symlink(struct xfs_scrub *sc, unsigned int *x)
 #define xrep_dirtree			xrep_notsupported
 #define xrep_metapath			xrep_notsupported
 #define xrep_rgsuperblock		xrep_notsupported
+#define xrep_rtrmapbt			xrep_notsupported
 
 #endif /* CONFIG_XFS_ONLINE_REPAIR */
 
diff --git a/fs/xfs/scrub/rtrmap.c b/fs/xfs/scrub/rtrmap.c
index 764fa296792234..300a1e85b3d625 100644
--- a/fs/xfs/scrub/rtrmap.c
+++ b/fs/xfs/scrub/rtrmap.c
@@ -27,6 +27,7 @@
 #include "scrub/common.h"
 #include "scrub/btree.h"
 #include "scrub/trace.h"
+#include "scrub/repair.h"
 
 /* Set us up with the realtime metadata locked. */
 int
@@ -38,6 +39,12 @@ xchk_setup_rtrmapbt(
 	if (xchk_need_intent_drain(sc))
 		xchk_fsgates_enable(sc, XCHK_FSGATES_DRAIN);
 
+	if (xchk_could_repair(sc)) {
+		error = xrep_setup_rtrmapbt(sc);
+		if (error)
+			return error;
+	}
+
 	error = xchk_rtgroup_init(sc, sc->sm->sm_agno, &sc->sr);
 	if (error)
 		return error;
diff --git a/fs/xfs/scrub/rtrmap_repair.c b/fs/xfs/scrub/rtrmap_repair.c
new file mode 100644
index 00000000000000..60e317725dea86
--- /dev/null
+++ b/fs/xfs/scrub/rtrmap_repair.c
@@ -0,0 +1,733 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2020-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#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_btree_staging.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_alloc.h"
+#include "xfs_rmap.h"
+#include "xfs_rmap_btree.h"
+#include "xfs_rtrmap_btree.h"
+#include "xfs_inode.h"
+#include "xfs_icache.h"
+#include "xfs_bmap.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_quota.h"
+#include "xfs_rtalloc.h"
+#include "xfs_ag.h"
+#include "xfs_rtgroup.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/btree.h"
+#include "scrub/trace.h"
+#include "scrub/repair.h"
+#include "scrub/bitmap.h"
+#include "scrub/fsb_bitmap.h"
+#include "scrub/xfile.h"
+#include "scrub/xfarray.h"
+#include "scrub/iscan.h"
+#include "scrub/newbt.h"
+#include "scrub/reap.h"
+
+/*
+ * Realtime Reverse Mapping Btree Repair
+ * =====================================
+ *
+ * This isn't quite as difficult as repairing the rmap btree on the data
+ * device, since we only store the data fork extents of realtime files on the
+ * realtime device.  We still have to freeze the filesystem and stop the
+ * background threads like we do for the rmap repair, but we only have to scan
+ * realtime inodes.
+ *
+ * Collecting entries for the new realtime rmap btree is easy -- all we have
+ * to do is generate rtrmap entries from the data fork mappings of all realtime
+ * files in the filesystem.  We then scan the rmap btrees of the data device
+ * looking for extents belonging to the old btree and note them in a bitmap.
+ *
+ * To rebuild the realtime rmap btree, we bulk-load the collected mappings into
+ * a new btree cursor and atomically swap that into the realtime inode.  Then
+ * we can free the blocks from the old btree.
+ *
+ * We use the 'xrep_rtrmap' prefix for all the rmap functions.
+ */
+
+/*
+ * Packed rmap record.  The UNWRITTEN flags are hidden in the upper bits of
+ * offset, just like the on-disk record.
+ */
+struct xrep_rtrmap_extent {
+	xfs_rgblock_t	startblock;
+	xfs_extlen_t	blockcount;
+	uint64_t	owner;
+	uint64_t	offset;
+} __packed;
+
+/* Context for collecting rmaps */
+struct xrep_rtrmap {
+	/* new rtrmapbt information */
+	struct xrep_newbt	new_btree;
+
+	/* rmap records generated from primary metadata */
+	struct xfarray		*rtrmap_records;
+
+	struct xfs_scrub	*sc;
+
+	/* bitmap of old rtrmapbt blocks */
+	struct xfsb_bitmap	old_rtrmapbt_blocks;
+
+	/* inode scan cursor */
+	struct xchk_iscan	iscan;
+
+	/* get_records()'s position in the free space record array. */
+	xfarray_idx_t		array_cur;
+};
+
+/* Set us up to repair rt reverse mapping btrees. */
+int
+xrep_setup_rtrmapbt(
+	struct xfs_scrub	*sc)
+{
+	struct xrep_rtrmap	*rr;
+
+	rr = kzalloc(sizeof(struct xrep_rtrmap), XCHK_GFP_FLAGS);
+	if (!rr)
+		return -ENOMEM;
+
+	rr->sc = sc;
+	sc->buf = rr;
+	return 0;
+}
+
+/* Make sure there's nothing funny about this mapping. */
+STATIC int
+xrep_rtrmap_check_mapping(
+	struct xfs_scrub	*sc,
+	const struct xfs_rmap_irec *rec)
+{
+	if (xfs_rtrmap_check_irec(sc->sr.rtg, rec) != NULL)
+		return -EFSCORRUPTED;
+
+	/* Make sure this isn't free space. */
+	return xrep_require_rtext_inuse(sc, rec->rm_startblock,
+			rec->rm_blockcount);
+}
+
+/* Store a reverse-mapping record. */
+static inline int
+xrep_rtrmap_stash(
+	struct xrep_rtrmap	*rr,
+	xfs_rgblock_t		startblock,
+	xfs_extlen_t		blockcount,
+	uint64_t		owner,
+	uint64_t		offset,
+	unsigned int		flags)
+{
+	struct xrep_rtrmap_extent	rre = {
+		.startblock	= startblock,
+		.blockcount	= blockcount,
+		.owner		= owner,
+	};
+	struct xfs_rmap_irec	rmap = {
+		.rm_startblock	= startblock,
+		.rm_blockcount	= blockcount,
+		.rm_owner	= owner,
+		.rm_offset	= offset,
+		.rm_flags	= flags,
+	};
+	struct xfs_scrub	*sc = rr->sc;
+	int			error = 0;
+
+	if (xchk_should_terminate(sc, &error))
+		return error;
+
+	trace_xrep_rtrmap_found(sc->mp, &rmap);
+
+	rre.offset = xfs_rmap_irec_offset_pack(&rmap);
+	return xfarray_append(rr->rtrmap_records, &rre);
+}
+
+/* Finding all file and bmbt extents. */
+
+/* Context for accumulating rmaps for an inode fork. */
+struct xrep_rtrmap_ifork {
+	/*
+	 * Accumulate rmap data here to turn multiple adjacent bmaps into a
+	 * single rmap.
+	 */
+	struct xfs_rmap_irec	accum;
+
+	struct xrep_rtrmap	*rr;
+};
+
+/* Stash an rmap that we accumulated while walking an inode fork. */
+STATIC int
+xrep_rtrmap_stash_accumulated(
+	struct xrep_rtrmap_ifork	*rf)
+{
+	if (rf->accum.rm_blockcount == 0)
+		return 0;
+
+	return xrep_rtrmap_stash(rf->rr, rf->accum.rm_startblock,
+			rf->accum.rm_blockcount, rf->accum.rm_owner,
+			rf->accum.rm_offset, rf->accum.rm_flags);
+}
+
+/* Accumulate a bmbt record. */
+STATIC int
+xrep_rtrmap_visit_bmbt(
+	struct xfs_btree_cur	*cur,
+	struct xfs_bmbt_irec	*rec,
+	void			*priv)
+{
+	struct xrep_rtrmap_ifork *rf = priv;
+	struct xfs_rmap_irec	*accum = &rf->accum;
+	struct xfs_mount	*mp = rf->rr->sc->mp;
+	xfs_rgblock_t		rgbno;
+	unsigned int		rmap_flags = 0;
+	int			error;
+
+	if (xfs_rtb_to_rgno(mp, rec->br_startblock) !=
+	    rtg_rgno(rf->rr->sc->sr.rtg))
+		return 0;
+
+	if (rec->br_state == XFS_EXT_UNWRITTEN)
+		rmap_flags |= XFS_RMAP_UNWRITTEN;
+
+	/* If this bmap is adjacent to the previous one, just add it. */
+	rgbno = xfs_rtb_to_rgbno(mp, rec->br_startblock);
+	if (accum->rm_blockcount > 0 &&
+	    rec->br_startoff == accum->rm_offset + accum->rm_blockcount &&
+	    rgbno == accum->rm_startblock + accum->rm_blockcount &&
+	    rmap_flags == accum->rm_flags) {
+		accum->rm_blockcount += rec->br_blockcount;
+		return 0;
+	}
+
+	/* Otherwise stash the old rmap and start accumulating a new one. */
+	error = xrep_rtrmap_stash_accumulated(rf);
+	if (error)
+		return error;
+
+	accum->rm_startblock = rgbno;
+	accum->rm_blockcount = rec->br_blockcount;
+	accum->rm_offset = rec->br_startoff;
+	accum->rm_flags = rmap_flags;
+	return 0;
+}
+
+/*
+ * Iterate the block mapping btree to collect rmap records for anything in this
+ * fork that maps to the rt volume.  Sets @mappings_done to true if we've
+ * scanned the block mappings in this fork.
+ */
+STATIC int
+xrep_rtrmap_scan_bmbt(
+	struct xrep_rtrmap_ifork *rf,
+	struct xfs_inode	*ip,
+	bool			*mappings_done)
+{
+	struct xrep_rtrmap	*rr = rf->rr;
+	struct xfs_btree_cur	*cur;
+	struct xfs_ifork	*ifp = xfs_ifork_ptr(ip, XFS_DATA_FORK);
+	int			error = 0;
+
+	*mappings_done = false;
+
+	/*
+	 * If the incore extent cache is already loaded, we'll just use the
+	 * incore extent scanner to record mappings.  Don't bother walking the
+	 * ondisk extent tree.
+	 */
+	if (!xfs_need_iread_extents(ifp))
+		return 0;
+
+	/* Accumulate all the mappings in the bmap btree. */
+	cur = xfs_bmbt_init_cursor(rr->sc->mp, rr->sc->tp, ip, XFS_DATA_FORK);
+	error = xfs_bmap_query_all(cur, xrep_rtrmap_visit_bmbt, rf);
+	xfs_btree_del_cursor(cur, error);
+	if (error)
+		return error;
+
+	/* Stash any remaining accumulated rmaps and exit. */
+	*mappings_done = true;
+	return xrep_rtrmap_stash_accumulated(rf);
+}
+
+/*
+ * Iterate the in-core extent cache to collect rmap records for anything in
+ * this fork that matches the AG.
+ */
+STATIC int
+xrep_rtrmap_scan_iext(
+	struct xrep_rtrmap_ifork *rf,
+	struct xfs_ifork	*ifp)
+{
+	struct xfs_bmbt_irec	rec;
+	struct xfs_iext_cursor	icur;
+	int			error;
+
+	for_each_xfs_iext(ifp, &icur, &rec) {
+		if (isnullstartblock(rec.br_startblock))
+			continue;
+		error = xrep_rtrmap_visit_bmbt(NULL, &rec, rf);
+		if (error)
+			return error;
+	}
+
+	return xrep_rtrmap_stash_accumulated(rf);
+}
+
+/* Find all the extents on the realtime device mapped by an inode fork. */
+STATIC int
+xrep_rtrmap_scan_dfork(
+	struct xrep_rtrmap	*rr,
+	struct xfs_inode	*ip)
+{
+	struct xrep_rtrmap_ifork rf = {
+		.accum		= { .rm_owner = ip->i_ino, },
+		.rr		= rr,
+	};
+	struct xfs_ifork	*ifp = xfs_ifork_ptr(ip, XFS_DATA_FORK);
+	int			error = 0;
+
+	if (ifp->if_format == XFS_DINODE_FMT_BTREE) {
+		bool		mappings_done;
+
+		/*
+		 * Scan the bmbt for mappings.  If the incore extent tree is
+		 * loaded, we want to scan the cached mappings since that's
+		 * faster when the extent counts are very high.
+		 */
+		error = xrep_rtrmap_scan_bmbt(&rf, ip, &mappings_done);
+		if (error || mappings_done)
+			return error;
+	} else if (ifp->if_format != XFS_DINODE_FMT_EXTENTS) {
+		/* realtime data forks should only be extents or btree */
+		return -EFSCORRUPTED;
+	}
+
+	/* Scan incore extent cache. */
+	return xrep_rtrmap_scan_iext(&rf, ifp);
+}
+
+/* Record reverse mappings for a file. */
+STATIC int
+xrep_rtrmap_scan_inode(
+	struct xrep_rtrmap	*rr,
+	struct xfs_inode	*ip)
+{
+	unsigned int		lock_mode;
+	int			error = 0;
+
+	/* Skip the rt rmap btree inode. */
+	if (rr->sc->ip == ip)
+		return 0;
+
+	lock_mode = xfs_ilock_data_map_shared(ip);
+
+	/* Check the data fork if it's on the realtime device. */
+	if (XFS_IS_REALTIME_INODE(ip)) {
+		error = xrep_rtrmap_scan_dfork(rr, ip);
+		if (error)
+			goto out_unlock;
+	}
+
+	xchk_iscan_mark_visited(&rr->iscan, ip);
+out_unlock:
+	xfs_iunlock(ip, lock_mode);
+	return error;
+}
+
+/* Record extents that belong to the realtime rmap inode. */
+STATIC int
+xrep_rtrmap_walk_rmap(
+	struct xfs_btree_cur		*cur,
+	const struct xfs_rmap_irec	*rec,
+	void				*priv)
+{
+	struct xrep_rtrmap		*rr = priv;
+	int				error = 0;
+
+	if (xchk_should_terminate(rr->sc, &error))
+		return error;
+
+	/* Skip extents which are not owned by this inode and fork. */
+	if (rec->rm_owner != rr->sc->ip->i_ino)
+		return 0;
+
+	error = xrep_check_ino_btree_mapping(rr->sc, rec);
+	if (error)
+		return error;
+
+	return xfsb_bitmap_set(&rr->old_rtrmapbt_blocks,
+			xfs_gbno_to_fsb(cur->bc_group, rec->rm_startblock),
+			rec->rm_blockcount);
+}
+
+/* Scan one AG for reverse mappings for the realtime rmap btree. */
+STATIC int
+xrep_rtrmap_scan_ag(
+	struct xrep_rtrmap	*rr,
+	struct xfs_perag	*pag)
+{
+	struct xfs_scrub	*sc = rr->sc;
+	int			error;
+
+	error = xrep_ag_init(sc, pag, &sc->sa);
+	if (error)
+		return error;
+
+	error = xfs_rmap_query_all(sc->sa.rmap_cur, xrep_rtrmap_walk_rmap, rr);
+	xchk_ag_free(sc, &sc->sa);
+	return error;
+}
+
+/* Generate all the reverse-mappings for the realtime device. */
+STATIC int
+xrep_rtrmap_find_rmaps(
+	struct xrep_rtrmap	*rr)
+{
+	struct xfs_scrub	*sc = rr->sc;
+	struct xfs_perag	*pag = NULL;
+	struct xfs_inode	*ip;
+	int			error;
+
+	/* Generate rmaps for the realtime superblock */
+	if (xfs_has_rtsb(sc->mp) && rtg_rgno(rr->sc->sr.rtg) == 0) {
+		error = xrep_rtrmap_stash(rr, 0, sc->mp->m_sb.sb_rextsize,
+				XFS_RMAP_OWN_FS, 0, 0);
+		if (error)
+			return error;
+	}
+
+	/*
+	 * Set up for a potentially lengthy filesystem scan by reducing our
+	 * transaction resource usage for the duration.  Specifically:
+	 *
+	 * Unlock the realtime metadata inodes and cancel the transaction to
+	 * release the log grant space while we scan the filesystem.
+	 *
+	 * Create a new empty transaction to eliminate the possibility of the
+	 * inode scan deadlocking on cyclical metadata.
+	 *
+	 * We pass the empty transaction to the file scanning function to avoid
+	 * repeatedly cycling empty transactions.  This can be done even though
+	 * we take the IOLOCK to quiesce the file because empty transactions
+	 * do not take sb_internal.
+	 */
+	xchk_trans_cancel(sc);
+	xchk_rtgroup_unlock(&sc->sr);
+	error = xchk_trans_alloc_empty(sc);
+	if (error)
+		return error;
+
+	while ((error = xchk_iscan_iter(&rr->iscan, &ip)) == 1) {
+		error = xrep_rtrmap_scan_inode(rr, ip);
+		xchk_irele(sc, ip);
+		if (error)
+			break;
+
+		if (xchk_should_terminate(sc, &error))
+			break;
+	}
+	xchk_iscan_iter_finish(&rr->iscan);
+	if (error)
+		return error;
+
+	/*
+	 * Switch out for a real transaction and lock the RT metadata in
+	 * preparation for building a new tree.
+	 */
+	xchk_trans_cancel(sc);
+	error = xchk_setup_rt(sc);
+	if (error)
+		return error;
+	error = xchk_rtgroup_lock(sc, &sc->sr, XCHK_RTGLOCK_ALL);
+	if (error)
+		return error;
+
+	/* Scan for old rtrmap blocks. */
+	while ((pag = xfs_perag_next(sc->mp, pag))) {
+		error = xrep_rtrmap_scan_ag(rr, pag);
+		if (error) {
+			xfs_perag_rele(pag);
+			return error;
+		}
+	}
+
+	return 0;
+}
+
+/* Building the new rtrmap btree. */
+
+/* Retrieve rtrmapbt data for bulk load. */
+STATIC int
+xrep_rtrmap_get_records(
+	struct xfs_btree_cur		*cur,
+	unsigned int			idx,
+	struct xfs_btree_block		*block,
+	unsigned int			nr_wanted,
+	void				*priv)
+{
+	struct xrep_rtrmap_extent	rec;
+	struct xfs_rmap_irec		*irec = &cur->bc_rec.r;
+	struct xrep_rtrmap		*rr = priv;
+	union xfs_btree_rec		*block_rec;
+	unsigned int			loaded;
+	int				error;
+
+	for (loaded = 0; loaded < nr_wanted; loaded++, idx++) {
+		error = xfarray_load_next(rr->rtrmap_records, &rr->array_cur,
+				&rec);
+		if (error)
+			return error;
+
+		irec->rm_startblock = rec.startblock;
+		irec->rm_blockcount = rec.blockcount;
+		irec->rm_owner = rec.owner;
+
+		if (xfs_rmap_irec_offset_unpack(rec.offset, irec) != NULL)
+			return -EFSCORRUPTED;
+
+		error = xrep_rtrmap_check_mapping(rr->sc, irec);
+		if (error)
+			return error;
+
+		block_rec = xfs_btree_rec_addr(cur, idx, block);
+		cur->bc_ops->init_rec_from_cur(cur, block_rec);
+	}
+
+	return loaded;
+}
+
+/* Feed one of the new btree blocks to the bulk loader. */
+STATIC int
+xrep_rtrmap_claim_block(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr,
+	void			*priv)
+{
+	struct xrep_rtrmap	*rr = priv;
+
+	return xrep_newbt_claim_block(cur, &rr->new_btree, ptr);
+}
+
+/* Figure out how much space we need to create the incore btree root block. */
+STATIC size_t
+xrep_rtrmap_iroot_size(
+	struct xfs_btree_cur	*cur,
+	unsigned int		level,
+	unsigned int		nr_this_level,
+	void			*priv)
+{
+	return xfs_rtrmap_broot_space_calc(cur->bc_mp, level, nr_this_level);
+}
+
+/*
+ * Use the collected rmap information to stage a new rmap btree.  If this is
+ * successful we'll return with the new btree root information logged to the
+ * repair transaction but not yet committed.  This implements section (III)
+ * above.
+ */
+STATIC int
+xrep_rtrmap_build_new_tree(
+	struct xrep_rtrmap	*rr)
+{
+	struct xfs_scrub	*sc = rr->sc;
+	struct xfs_rtgroup	*rtg = sc->sr.rtg;
+	struct xfs_btree_cur	*rmap_cur;
+	uint64_t		nr_records;
+	int			error;
+
+	/*
+	 * Prepare to construct the new btree by reserving disk space for the
+	 * new btree and setting up all the accounting information we'll need
+	 * to root the new btree while it's under construction and before we
+	 * attach it to the realtime rmapbt inode.
+	 */
+	error = xrep_newbt_init_metadir_inode(&rr->new_btree, sc);
+	if (error)
+		return error;
+
+	rr->new_btree.bload.get_records = xrep_rtrmap_get_records;
+	rr->new_btree.bload.claim_block = xrep_rtrmap_claim_block;
+	rr->new_btree.bload.iroot_size = xrep_rtrmap_iroot_size;
+
+	rmap_cur = xfs_rtrmapbt_init_cursor(NULL, rtg);
+	xfs_btree_stage_ifakeroot(rmap_cur, &rr->new_btree.ifake);
+
+	nr_records = xfarray_length(rr->rtrmap_records);
+
+	/* Compute how many blocks we'll need for the rmaps collected. */
+	error = xfs_btree_bload_compute_geometry(rmap_cur,
+			&rr->new_btree.bload, nr_records);
+	if (error)
+		goto err_cur;
+
+	/* Last chance to abort before we start committing fixes. */
+	if (xchk_should_terminate(sc, &error))
+		goto err_cur;
+
+	/*
+	 * Guess how many blocks we're going to need to rebuild an entire
+	 * rtrmapbt from the number of extents we found, and pump up our
+	 * transaction to have sufficient block reservation.  We're allowed
+	 * to exceed quota to repair inconsistent metadata, though this is
+	 * unlikely.
+	 */
+	error = xfs_trans_reserve_more_inode(sc->tp, rtg_rmap(rtg),
+			rr->new_btree.bload.nr_blocks, 0, true);
+	if (error)
+		goto err_cur;
+
+	/* Reserve the space we'll need for the new btree. */
+	error = xrep_newbt_alloc_blocks(&rr->new_btree,
+			rr->new_btree.bload.nr_blocks);
+	if (error)
+		goto err_cur;
+
+	/* Add all observed rmap records. */
+	rr->new_btree.ifake.if_fork->if_format = XFS_DINODE_FMT_META_BTREE;
+	rr->array_cur = XFARRAY_CURSOR_INIT;
+	error = xfs_btree_bload(rmap_cur, &rr->new_btree.bload, rr);
+	if (error)
+		goto err_cur;
+
+	/*
+	 * Install the new rtrmap btree in the inode.  After this point the old
+	 * btree is no longer accessible, the new tree is live, and we can
+	 * delete the cursor.
+	 */
+	xfs_rtrmapbt_commit_staged_btree(rmap_cur, sc->tp);
+	xrep_inode_set_nblocks(rr->sc, rr->new_btree.ifake.if_blocks);
+	xfs_btree_del_cursor(rmap_cur, 0);
+
+	/* Dispose of any unused blocks and the accounting information. */
+	error = xrep_newbt_commit(&rr->new_btree);
+	if (error)
+		return error;
+
+	return xrep_roll_trans(sc);
+
+err_cur:
+	xfs_btree_del_cursor(rmap_cur, error);
+	xrep_newbt_cancel(&rr->new_btree);
+	return error;
+}
+
+/* Reaping the old btree. */
+
+/* Reap the old rtrmapbt blocks. */
+STATIC int
+xrep_rtrmap_remove_old_tree(
+	struct xrep_rtrmap	*rr)
+{
+	int			error;
+
+	/*
+	 * Free all the extents that were allocated to the former rtrmapbt and
+	 * aren't cross-linked with something else.
+	 */
+	error = xrep_reap_metadir_fsblocks(rr->sc, &rr->old_rtrmapbt_blocks);
+	if (error)
+		return error;
+
+	/*
+	 * Ensure the proper reservation for the rtrmap inode so that we don't
+	 * fail to expand the new btree.
+	 */
+	return xrep_reset_metafile_resv(rr->sc);
+}
+
+/* Set up the filesystem scan components. */
+STATIC int
+xrep_rtrmap_setup_scan(
+	struct xrep_rtrmap	*rr)
+{
+	struct xfs_scrub	*sc = rr->sc;
+	char			*descr;
+	int			error;
+
+	xfsb_bitmap_init(&rr->old_rtrmapbt_blocks);
+
+	/* Set up some storage */
+	descr = xchk_xfile_rtgroup_descr(sc, "reverse mapping records");
+	error = xfarray_create(descr, 0, sizeof(struct xrep_rtrmap_extent),
+			&rr->rtrmap_records);
+	kfree(descr);
+	if (error)
+		goto out_bitmap;
+
+	/* Retry iget every tenth of a second for up to 30 seconds. */
+	xchk_iscan_start(sc, 30000, 100, &rr->iscan);
+	return 0;
+
+out_bitmap:
+	xfsb_bitmap_destroy(&rr->old_rtrmapbt_blocks);
+	return error;
+}
+
+/* Tear down scan components. */
+STATIC void
+xrep_rtrmap_teardown(
+	struct xrep_rtrmap	*rr)
+{
+	xchk_iscan_teardown(&rr->iscan);
+	xfarray_destroy(rr->rtrmap_records);
+	xfsb_bitmap_destroy(&rr->old_rtrmapbt_blocks);
+}
+
+/* Repair the realtime rmap btree. */
+int
+xrep_rtrmapbt(
+	struct xfs_scrub	*sc)
+{
+	struct xrep_rtrmap	*rr = sc->buf;
+	int			error;
+
+	/* Functionality is not yet complete. */
+	return xrep_notsupported(sc);
+
+	/* Make sure any problems with the fork are fixed. */
+	error = xrep_metadata_inode_forks(sc);
+	if (error)
+		return error;
+
+	error = xrep_rtrmap_setup_scan(rr);
+	if (error)
+		return error;
+
+	/* Collect rmaps for realtime files. */
+	error = xrep_rtrmap_find_rmaps(rr);
+	if (error)
+		goto out_records;
+
+	xfs_trans_ijoin(sc->tp, sc->ip, 0);
+
+	/* Rebuild the rtrmap information. */
+	error = xrep_rtrmap_build_new_tree(rr);
+	if (error)
+		goto out_records;
+
+	/* Kill the old tree. */
+	error = xrep_rtrmap_remove_old_tree(rr);
+	if (error)
+		goto out_records;
+
+out_records:
+	xrep_rtrmap_teardown(rr);
+	return error;
+}
diff --git a/fs/xfs/scrub/scrub.c b/fs/xfs/scrub/scrub.c
index 09983899c34164..16da054b2eb0dc 100644
--- a/fs/xfs/scrub/scrub.c
+++ b/fs/xfs/scrub/scrub.c
@@ -465,7 +465,7 @@ static const struct xchk_meta_ops meta_scrub_ops[] = {
 		.setup	= xchk_setup_rtrmapbt,
 		.scrub	= xchk_rtrmapbt,
 		.has	= xfs_has_rtrmapbt,
-		.repair	= xrep_notsupported,
+		.repair	= xrep_rtrmapbt,
 	},
 };
 
diff --git a/fs/xfs/scrub/trace.h b/fs/xfs/scrub/trace.h
index 3b661e4443453c..3f2a8695ef5cb5 100644
--- a/fs/xfs/scrub/trace.h
+++ b/fs/xfs/scrub/trace.h
@@ -2285,6 +2285,32 @@ TRACE_EVENT(xrep_calc_ag_resblks_btsize,
 		  __entry->rmapbt_sz,
 		  __entry->refcbt_sz)
 )
+
+#ifdef CONFIG_XFS_RT
+TRACE_EVENT(xrep_calc_rtgroup_resblks_btsize,
+	TP_PROTO(struct xfs_mount *mp, xfs_rgnumber_t rgno,
+		 xfs_rgblock_t usedlen, xfs_rgblock_t rmapbt_sz),
+	TP_ARGS(mp, rgno, usedlen, rmapbt_sz),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_rgnumber_t, rgno)
+		__field(xfs_rgblock_t, usedlen)
+		__field(xfs_rgblock_t, rmapbt_sz)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->rgno = rgno;
+		__entry->usedlen = usedlen;
+		__entry->rmapbt_sz = rmapbt_sz;
+	),
+	TP_printk("dev %d:%d rgno 0x%x usedlen %u rmapbt %u",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->rgno,
+		  __entry->usedlen,
+		  __entry->rmapbt_sz)
+);
+#endif /* CONFIG_XFS_RT */
+
 TRACE_EVENT(xrep_reset_counters,
 	TP_PROTO(struct xfs_mount *mp, struct xchk_fscounters *fsc),
 	TP_ARGS(mp, fsc),
@@ -3755,6 +3781,37 @@ TRACE_EVENT(xrep_rtbitmap_load_word,
 		  (__entry->ondisk_word & __entry->word_mask),
 		  __entry->word_mask)
 );
+
+TRACE_EVENT(xrep_rtrmap_found,
+	TP_PROTO(struct xfs_mount *mp, const struct xfs_rmap_irec *rec),
+	TP_ARGS(mp, rec),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(dev_t, rtdev)
+		__field(xfs_rgblock_t, rgbno)
+		__field(xfs_extlen_t, len)
+		__field(uint64_t, owner)
+		__field(uint64_t, offset)
+		__field(unsigned int, flags)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->rtdev = mp->m_rtdev_targp->bt_dev;
+		__entry->rgbno = rec->rm_startblock;
+		__entry->len = rec->rm_blockcount;
+		__entry->owner = rec->rm_owner;
+		__entry->offset = rec->rm_offset;
+		__entry->flags = rec->rm_flags;
+	),
+	TP_printk("dev %d:%d rtdev %d:%d rgbno 0x%x fsbcount 0x%x owner 0x%llx fileoff 0x%llx flags 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  MAJOR(__entry->rtdev), MINOR(__entry->rtdev),
+		  __entry->rgbno,
+		  __entry->len,
+		  __entry->owner,
+		  __entry->offset,
+		  __entry->flags)
+);
 #endif /* CONFIG_XFS_RT */
 
 #endif /* IS_ENABLED(CONFIG_XFS_ONLINE_REPAIR) */

Re: [PATCH 32/37] xfs: online repair of the realtime rmap btree

[Christoph Hellwig]

Looks good:

Reviewed-by: Christoph Hellwig <hch@lst.de>

[PATCH 33/37] xfs: create a shadow rmap btree during realtime rmap repair

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Create an in-memory btree of rmap records instead of an array.  This
enables us to do live record collection instead of freezing the fs.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_btree_mem.c    |    1 
 fs/xfs/libxfs/xfs_rmap.c         |    3 +
 fs/xfs/libxfs/xfs_rtrmap_btree.c |  117 ++++++++++++++++++++++++++++++++
 fs/xfs/libxfs/xfs_rtrmap_btree.h |    6 ++
 fs/xfs/libxfs/xfs_shared.h       |    7 ++
 fs/xfs/scrub/rtrmap_repair.c     |  137 ++++++++++++++++++++++++++------------
 fs/xfs/xfs_stats.c               |    3 +
 fs/xfs/xfs_stats.h               |    1 
 8 files changed, 228 insertions(+), 47 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_btree_mem.c b/fs/xfs/libxfs/xfs_btree_mem.c
index df3d613675a15a..f2f7b4305413e9 100644
--- a/fs/xfs/libxfs/xfs_btree_mem.c
+++ b/fs/xfs/libxfs/xfs_btree_mem.c
@@ -18,6 +18,7 @@
 #include "xfs_ag.h"
 #include "xfs_buf_item.h"
 #include "xfs_trace.h"
+#include "xfs_rtgroup.h"
 
 /* Set the root of an in-memory btree. */
 void
diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c
index 2f0688a57991cc..f8415fd96cc2aa 100644
--- a/fs/xfs/libxfs/xfs_rmap.c
+++ b/fs/xfs/libxfs/xfs_rmap.c
@@ -327,7 +327,8 @@ xfs_rmap_check_btrec(
 	struct xfs_btree_cur		*cur,
 	const struct xfs_rmap_irec	*irec)
 {
-	if (xfs_btree_is_rtrmap(cur->bc_ops))
+	if (xfs_btree_is_rtrmap(cur->bc_ops) ||
+	    xfs_btree_is_mem_rtrmap(cur->bc_ops))
 		return xfs_rtrmap_check_irec(to_rtg(cur->bc_group), irec);
 	return xfs_rmap_check_irec(to_perag(cur->bc_group), irec);
 }
diff --git a/fs/xfs/libxfs/xfs_rtrmap_btree.c b/fs/xfs/libxfs/xfs_rtrmap_btree.c
index 571a9e1b956099..3cb8f126b9ce16 100644
--- a/fs/xfs/libxfs/xfs_rtrmap_btree.c
+++ b/fs/xfs/libxfs/xfs_rtrmap_btree.c
@@ -28,6 +28,8 @@
 #include "xfs_rtgroup.h"
 #include "xfs_bmap.h"
 #include "xfs_health.h"
+#include "xfs_buf_mem.h"
+#include "xfs_btree_mem.h"
 
 static struct kmem_cache	*xfs_rtrmapbt_cur_cache;
 
@@ -542,6 +544,121 @@ xfs_rtrmapbt_init_cursor(
 	return cur;
 }
 
+#ifdef CONFIG_XFS_BTREE_IN_MEM
+/*
+ * Validate an in-memory realtime rmap btree block.  Callers are allowed to
+ * generate an in-memory btree even if the ondisk feature is not enabled.
+ */
+static xfs_failaddr_t
+xfs_rtrmapbt_mem_verify(
+	struct xfs_buf		*bp)
+{
+	struct xfs_mount	*mp = bp->b_mount;
+	struct xfs_btree_block	*block = XFS_BUF_TO_BLOCK(bp);
+	xfs_failaddr_t		fa;
+	unsigned int		level;
+	unsigned int		maxrecs;
+
+	if (!xfs_verify_magic(bp, block->bb_magic))
+		return __this_address;
+
+	fa = xfs_btree_fsblock_v5hdr_verify(bp, XFS_RMAP_OWN_UNKNOWN);
+	if (fa)
+		return fa;
+
+	level = be16_to_cpu(block->bb_level);
+	if (xfs_has_rmapbt(mp)) {
+		if (level >= mp->m_rtrmap_maxlevels)
+			return __this_address;
+	} else {
+		if (level >= xfs_rtrmapbt_maxlevels_ondisk())
+			return __this_address;
+	}
+
+	maxrecs = xfs_rtrmapbt_maxrecs(mp, XFBNO_BLOCKSIZE, level == 0);
+	return xfs_btree_memblock_verify(bp, maxrecs);
+}
+
+static void
+xfs_rtrmapbt_mem_rw_verify(
+	struct xfs_buf	*bp)
+{
+	xfs_failaddr_t	fa = xfs_rtrmapbt_mem_verify(bp);
+
+	if (fa)
+		xfs_verifier_error(bp, -EFSCORRUPTED, fa);
+}
+
+/* skip crc checks on in-memory btrees to save time */
+static const struct xfs_buf_ops xfs_rtrmapbt_mem_buf_ops = {
+	.name			= "xfs_rtrmapbt_mem",
+	.magic			= { 0, cpu_to_be32(XFS_RTRMAP_CRC_MAGIC) },
+	.verify_read		= xfs_rtrmapbt_mem_rw_verify,
+	.verify_write		= xfs_rtrmapbt_mem_rw_verify,
+	.verify_struct		= xfs_rtrmapbt_mem_verify,
+};
+
+const struct xfs_btree_ops xfs_rtrmapbt_mem_ops = {
+	.type			= XFS_BTREE_TYPE_MEM,
+	.geom_flags		= XFS_BTGEO_OVERLAPPING,
+
+	.rec_len		= sizeof(struct xfs_rmap_rec),
+	/* Overlapping btree; 2 keys per pointer. */
+	.key_len		= 2 * sizeof(struct xfs_rmap_key),
+	.ptr_len		= XFS_BTREE_LONG_PTR_LEN,
+
+	.lru_refs		= XFS_RMAP_BTREE_REF,
+	.statoff		= XFS_STATS_CALC_INDEX(xs_rtrmap_mem_2),
+
+	.dup_cursor		= xfbtree_dup_cursor,
+	.set_root		= xfbtree_set_root,
+	.alloc_block		= xfbtree_alloc_block,
+	.free_block		= xfbtree_free_block,
+	.get_minrecs		= xfbtree_get_minrecs,
+	.get_maxrecs		= xfbtree_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	= xfbtree_init_ptr_from_cur,
+	.key_diff		= xfs_rtrmapbt_key_diff,
+	.buf_ops		= &xfs_rtrmapbt_mem_buf_ops,
+	.diff_two_keys		= xfs_rtrmapbt_diff_two_keys,
+	.keys_inorder		= xfs_rtrmapbt_keys_inorder,
+	.recs_inorder		= xfs_rtrmapbt_recs_inorder,
+	.keys_contiguous	= xfs_rtrmapbt_keys_contiguous,
+};
+
+/* Create a cursor for an in-memory btree. */
+struct xfs_btree_cur *
+xfs_rtrmapbt_mem_cursor(
+	struct xfs_rtgroup	*rtg,
+	struct xfs_trans	*tp,
+	struct xfbtree		*xfbt)
+{
+	struct xfs_mount	*mp = rtg_mount(rtg);
+	struct xfs_btree_cur	*cur;
+
+	cur = xfs_btree_alloc_cursor(mp, tp, &xfs_rtrmapbt_mem_ops,
+			mp->m_rtrmap_maxlevels, xfs_rtrmapbt_cur_cache);
+	cur->bc_mem.xfbtree = xfbt;
+	cur->bc_nlevels = xfbt->nlevels;
+	cur->bc_group = xfs_group_hold(rtg_group(rtg));
+	return cur;
+}
+
+/* Create an in-memory realtime rmap btree. */
+int
+xfs_rtrmapbt_mem_init(
+	struct xfs_mount	*mp,
+	struct xfbtree		*xfbt,
+	struct xfs_buftarg	*btp,
+	xfs_rgnumber_t		rgno)
+{
+	xfbt->owner = rgno;
+	return xfbtree_init(mp, xfbt, btp, &xfs_rtrmapbt_mem_ops);
+}
+#endif /* CONFIG_XFS_BTREE_IN_MEM */
+
 /*
  * Install a new rt reverse mapping btree root.  Caller is responsible for
  * invalidating and freeing the old btree blocks.
diff --git a/fs/xfs/libxfs/xfs_rtrmap_btree.h b/fs/xfs/libxfs/xfs_rtrmap_btree.h
index ad76ac7938b602..9d0915089891a5 100644
--- a/fs/xfs/libxfs/xfs_rtrmap_btree.h
+++ b/fs/xfs/libxfs/xfs_rtrmap_btree.h
@@ -11,6 +11,7 @@ struct xfs_btree_cur;
 struct xfs_mount;
 struct xbtree_ifakeroot;
 struct xfs_rtgroup;
+struct xfbtree;
 
 /* rmaps only exist on crc enabled filesystems */
 #define XFS_RTRMAP_BLOCK_LEN	XFS_BTREE_LBLOCK_CRC_LEN
@@ -201,4 +202,9 @@ int xfs_rtrmapbt_init_rtsb(struct xfs_mount *mp, struct xfs_rtgroup *rtg,
 unsigned long long xfs_rtrmapbt_calc_size(struct xfs_mount *mp,
 		unsigned long long len);
 
+struct xfs_btree_cur *xfs_rtrmapbt_mem_cursor(struct xfs_rtgroup *rtg,
+		struct xfs_trans *tp, struct xfbtree *xfbtree);
+int xfs_rtrmapbt_mem_init(struct xfs_mount *mp, struct xfbtree *xfbtree,
+		struct xfs_buftarg *btp, xfs_rgnumber_t rgno);
+
 #endif /* __XFS_RTRMAP_BTREE_H__ */
diff --git a/fs/xfs/libxfs/xfs_shared.h b/fs/xfs/libxfs/xfs_shared.h
index da23dac22c3f08..960716c387cc2b 100644
--- a/fs/xfs/libxfs/xfs_shared.h
+++ b/fs/xfs/libxfs/xfs_shared.h
@@ -57,6 +57,7 @@ extern const struct xfs_btree_ops xfs_refcountbt_ops;
 extern const struct xfs_btree_ops xfs_rmapbt_ops;
 extern const struct xfs_btree_ops xfs_rmapbt_mem_ops;
 extern const struct xfs_btree_ops xfs_rtrmapbt_ops;
+extern const struct xfs_btree_ops xfs_rtrmapbt_mem_ops;
 
 static inline bool xfs_btree_is_bno(const struct xfs_btree_ops *ops)
 {
@@ -98,8 +99,14 @@ static inline bool xfs_btree_is_mem_rmap(const struct xfs_btree_ops *ops)
 {
 	return ops == &xfs_rmapbt_mem_ops;
 }
+
+static inline bool xfs_btree_is_mem_rtrmap(const struct xfs_btree_ops *ops)
+{
+	return ops == &xfs_rtrmapbt_mem_ops;
+}
 #else
 # define xfs_btree_is_mem_rmap(...)	(false)
+# define xfs_btree_is_mem_rtrmap(...)	(false)
 #endif
 
 static inline bool xfs_btree_is_rtrmap(const struct xfs_btree_ops *ops)
diff --git a/fs/xfs/scrub/rtrmap_repair.c b/fs/xfs/scrub/rtrmap_repair.c
index 60e317725dea86..b376bcc8d1d2ed 100644
--- a/fs/xfs/scrub/rtrmap_repair.c
+++ b/fs/xfs/scrub/rtrmap_repair.c
@@ -12,6 +12,8 @@
 #include "xfs_defer.h"
 #include "xfs_btree.h"
 #include "xfs_btree_staging.h"
+#include "xfs_buf_mem.h"
+#include "xfs_btree_mem.h"
 #include "xfs_bit.h"
 #include "xfs_log_format.h"
 #include "xfs_trans.h"
@@ -64,24 +66,13 @@
  * We use the 'xrep_rtrmap' prefix for all the rmap functions.
  */
 
-/*
- * Packed rmap record.  The UNWRITTEN flags are hidden in the upper bits of
- * offset, just like the on-disk record.
- */
-struct xrep_rtrmap_extent {
-	xfs_rgblock_t	startblock;
-	xfs_extlen_t	blockcount;
-	uint64_t	owner;
-	uint64_t	offset;
-} __packed;
-
 /* Context for collecting rmaps */
 struct xrep_rtrmap {
 	/* new rtrmapbt information */
 	struct xrep_newbt	new_btree;
 
 	/* rmap records generated from primary metadata */
-	struct xfarray		*rtrmap_records;
+	struct xfbtree		rtrmap_btree;
 
 	struct xfs_scrub	*sc;
 
@@ -91,8 +82,11 @@ struct xrep_rtrmap {
 	/* inode scan cursor */
 	struct xchk_iscan	iscan;
 
-	/* get_records()'s position in the free space record array. */
-	xfarray_idx_t		array_cur;
+	/* in-memory btree cursor for the ->get_blocks walk */
+	struct xfs_btree_cur	*mcur;
+
+	/* Number of records we're staging in the new btree. */
+	uint64_t		nr_records;
 };
 
 /* Set us up to repair rt reverse mapping btrees. */
@@ -101,6 +95,14 @@ xrep_setup_rtrmapbt(
 	struct xfs_scrub	*sc)
 {
 	struct xrep_rtrmap	*rr;
+	char			*descr;
+	int			error;
+
+	descr = xchk_xfile_rtgroup_descr(sc, "reverse mapping records");
+	error = xrep_setup_xfbtree(sc, descr);
+	kfree(descr);
+	if (error)
+		return error;
 
 	rr = kzalloc(sizeof(struct xrep_rtrmap), XCHK_GFP_FLAGS);
 	if (!rr)
@@ -135,11 +137,6 @@ xrep_rtrmap_stash(
 	uint64_t		offset,
 	unsigned int		flags)
 {
-	struct xrep_rtrmap_extent	rre = {
-		.startblock	= startblock,
-		.blockcount	= blockcount,
-		.owner		= owner,
-	};
 	struct xfs_rmap_irec	rmap = {
 		.rm_startblock	= startblock,
 		.rm_blockcount	= blockcount,
@@ -148,6 +145,7 @@ xrep_rtrmap_stash(
 		.rm_flags	= flags,
 	};
 	struct xfs_scrub	*sc = rr->sc;
+	struct xfs_btree_cur	*mcur;
 	int			error = 0;
 
 	if (xchk_should_terminate(sc, &error))
@@ -155,8 +153,18 @@ xrep_rtrmap_stash(
 
 	trace_xrep_rtrmap_found(sc->mp, &rmap);
 
-	rre.offset = xfs_rmap_irec_offset_pack(&rmap);
-	return xfarray_append(rr->rtrmap_records, &rre);
+	/* Add entry to in-memory btree. */
+	mcur = xfs_rtrmapbt_mem_cursor(sc->sr.rtg, sc->tp, &rr->rtrmap_btree);
+	error = xfs_rmap_map_raw(mcur, &rmap);
+	xfs_btree_del_cursor(mcur, error);
+	if (error)
+		goto out_cancel;
+
+	return xfbtree_trans_commit(&rr->rtrmap_btree, sc->tp);
+
+out_cancel:
+	xfbtree_trans_cancel(&rr->rtrmap_btree, sc->tp);
+	return error;
 }
 
 /* Finding all file and bmbt extents. */
@@ -395,6 +403,24 @@ xrep_rtrmap_scan_ag(
 	return error;
 }
 
+/* Count and check all collected records. */
+STATIC int
+xrep_rtrmap_check_record(
+	struct xfs_btree_cur		*cur,
+	const struct xfs_rmap_irec	*rec,
+	void				*priv)
+{
+	struct xrep_rtrmap		*rr = priv;
+	int				error;
+
+	error = xrep_rtrmap_check_mapping(rr->sc, rec);
+	if (error)
+		return error;
+
+	rr->nr_records++;
+	return 0;
+}
+
 /* Generate all the reverse-mappings for the realtime device. */
 STATIC int
 xrep_rtrmap_find_rmaps(
@@ -403,6 +429,7 @@ xrep_rtrmap_find_rmaps(
 	struct xfs_scrub	*sc = rr->sc;
 	struct xfs_perag	*pag = NULL;
 	struct xfs_inode	*ip;
+	struct xfs_btree_cur	*mcur;
 	int			error;
 
 	/* Generate rmaps for the realtime superblock */
@@ -468,7 +495,19 @@ xrep_rtrmap_find_rmaps(
 		}
 	}
 
-	return 0;
+	/*
+	 * Now that we have everything locked again, we need to count the
+	 * number of rmap records stashed in the btree.  This should reflect
+	 * all actively-owned rt files in the filesystem.  At the same time,
+	 * check all our records before we start building a new btree, which
+	 * requires the rtbitmap lock.
+	 */
+	mcur = xfs_rtrmapbt_mem_cursor(rr->sc->sr.rtg, NULL, &rr->rtrmap_btree);
+	rr->nr_records = 0;
+	error = xfs_rmap_query_all(mcur, xrep_rtrmap_check_record, rr);
+	xfs_btree_del_cursor(mcur, error);
+
+	return error;
 }
 
 /* Building the new rtrmap btree. */
@@ -482,29 +521,25 @@ xrep_rtrmap_get_records(
 	unsigned int			nr_wanted,
 	void				*priv)
 {
-	struct xrep_rtrmap_extent	rec;
-	struct xfs_rmap_irec		*irec = &cur->bc_rec.r;
 	struct xrep_rtrmap		*rr = priv;
 	union xfs_btree_rec		*block_rec;
 	unsigned int			loaded;
 	int				error;
 
 	for (loaded = 0; loaded < nr_wanted; loaded++, idx++) {
-		error = xfarray_load_next(rr->rtrmap_records, &rr->array_cur,
-				&rec);
+		int			stat = 0;
+
+		error = xfs_btree_increment(rr->mcur, 0, &stat);
 		if (error)
 			return error;
-
-		irec->rm_startblock = rec.startblock;
-		irec->rm_blockcount = rec.blockcount;
-		irec->rm_owner = rec.owner;
-
-		if (xfs_rmap_irec_offset_unpack(rec.offset, irec) != NULL)
+		if (!stat)
 			return -EFSCORRUPTED;
 
-		error = xrep_rtrmap_check_mapping(rr->sc, irec);
+		error = xfs_rmap_get_rec(rr->mcur, &cur->bc_rec.r, &stat);
 		if (error)
 			return error;
+		if (!stat)
+			return -EFSCORRUPTED;
 
 		block_rec = xfs_btree_rec_addr(cur, idx, block);
 		cur->bc_ops->init_rec_from_cur(cur, block_rec);
@@ -549,7 +584,6 @@ xrep_rtrmap_build_new_tree(
 	struct xfs_scrub	*sc = rr->sc;
 	struct xfs_rtgroup	*rtg = sc->sr.rtg;
 	struct xfs_btree_cur	*rmap_cur;
-	uint64_t		nr_records;
 	int			error;
 
 	/*
@@ -569,11 +603,9 @@ xrep_rtrmap_build_new_tree(
 	rmap_cur = xfs_rtrmapbt_init_cursor(NULL, rtg);
 	xfs_btree_stage_ifakeroot(rmap_cur, &rr->new_btree.ifake);
 
-	nr_records = xfarray_length(rr->rtrmap_records);
-
 	/* Compute how many blocks we'll need for the rmaps collected. */
 	error = xfs_btree_bload_compute_geometry(rmap_cur,
-			&rr->new_btree.bload, nr_records);
+			&rr->new_btree.bload, rr->nr_records);
 	if (error)
 		goto err_cur;
 
@@ -599,12 +631,20 @@ xrep_rtrmap_build_new_tree(
 	if (error)
 		goto err_cur;
 
+	/*
+	 * Create a cursor to the in-memory btree so that we can bulk load the
+	 * new btree.
+	 */
+	rr->mcur = xfs_rtrmapbt_mem_cursor(sc->sr.rtg, NULL, &rr->rtrmap_btree);
+	error = xfs_btree_goto_left_edge(rr->mcur);
+	if (error)
+		goto err_mcur;
+
 	/* Add all observed rmap records. */
 	rr->new_btree.ifake.if_fork->if_format = XFS_DINODE_FMT_META_BTREE;
-	rr->array_cur = XFARRAY_CURSOR_INIT;
 	error = xfs_btree_bload(rmap_cur, &rr->new_btree.bload, rr);
 	if (error)
-		goto err_cur;
+		goto err_mcur;
 
 	/*
 	 * Install the new rtrmap btree in the inode.  After this point the old
@@ -614,6 +654,14 @@ xrep_rtrmap_build_new_tree(
 	xfs_rtrmapbt_commit_staged_btree(rmap_cur, sc->tp);
 	xrep_inode_set_nblocks(rr->sc, rr->new_btree.ifake.if_blocks);
 	xfs_btree_del_cursor(rmap_cur, 0);
+	xfs_btree_del_cursor(rr->mcur, 0);
+	rr->mcur = NULL;
+
+	/*
+	 * Now that we've written the new btree to disk, we don't need to keep
+	 * updating the in-memory btree.  Abort the scan to stop live updates.
+	 */
+	xchk_iscan_abort(&rr->iscan);
 
 	/* Dispose of any unused blocks and the accounting information. */
 	error = xrep_newbt_commit(&rr->new_btree);
@@ -622,6 +670,8 @@ xrep_rtrmap_build_new_tree(
 
 	return xrep_roll_trans(sc);
 
+err_mcur:
+	xfs_btree_del_cursor(rr->mcur, error);
 err_cur:
 	xfs_btree_del_cursor(rmap_cur, error);
 	xrep_newbt_cancel(&rr->new_btree);
@@ -658,16 +708,13 @@ xrep_rtrmap_setup_scan(
 	struct xrep_rtrmap	*rr)
 {
 	struct xfs_scrub	*sc = rr->sc;
-	char			*descr;
 	int			error;
 
 	xfsb_bitmap_init(&rr->old_rtrmapbt_blocks);
 
 	/* Set up some storage */
-	descr = xchk_xfile_rtgroup_descr(sc, "reverse mapping records");
-	error = xfarray_create(descr, 0, sizeof(struct xrep_rtrmap_extent),
-			&rr->rtrmap_records);
-	kfree(descr);
+	error = xfs_rtrmapbt_mem_init(sc->mp, &rr->rtrmap_btree, sc->xmbtp,
+			rtg_rgno(sc->sr.rtg));
 	if (error)
 		goto out_bitmap;
 
@@ -686,7 +733,7 @@ xrep_rtrmap_teardown(
 	struct xrep_rtrmap	*rr)
 {
 	xchk_iscan_teardown(&rr->iscan);
-	xfarray_destroy(rr->rtrmap_records);
+	xfbtree_destroy(&rr->rtrmap_btree);
 	xfsb_bitmap_destroy(&rr->old_rtrmapbt_blocks);
 }
 
diff --git a/fs/xfs/xfs_stats.c b/fs/xfs/xfs_stats.c
index f94fb70b524ffb..b7f2988bc03bb7 100644
--- a/fs/xfs/xfs_stats.c
+++ b/fs/xfs/xfs_stats.c
@@ -53,7 +53,8 @@ int xfs_stats_format(struct xfsstats __percpu *stats, char *buf)
 		{ "refcntbt",		xfsstats_offset(xs_rmap_mem_2)	},
 		{ "rmapbt_mem",		xfsstats_offset(xs_rcbag_2)	},
 		{ "rcbagbt",		xfsstats_offset(xs_rtrmap_2)	},
-		{ "rtrmapbt",		xfsstats_offset(xs_qm_dqreclaims)},
+		{ "rtrmapbt",		xfsstats_offset(xs_rtrmap_mem_2)},
+		{ "rtrmapbt_mem",	xfsstats_offset(xs_qm_dqreclaims)},
 		/* we print both series of quota information together */
 		{ "qm",			xfsstats_offset(xs_xstrat_bytes)},
 	};
diff --git a/fs/xfs/xfs_stats.h b/fs/xfs/xfs_stats.h
index 05dc69c6d94906..9c47de5dff2dd6 100644
--- a/fs/xfs/xfs_stats.h
+++ b/fs/xfs/xfs_stats.h
@@ -128,6 +128,7 @@ struct __xfsstats {
 	uint32_t		xs_rmap_mem_2[__XBTS_MAX];
 	uint32_t		xs_rcbag_2[__XBTS_MAX];
 	uint32_t		xs_rtrmap_2[__XBTS_MAX];
+	uint32_t		xs_rtrmap_mem_2[__XBTS_MAX];
 	uint32_t		xs_qm_dqreclaims;
 	uint32_t		xs_qm_dqreclaim_misses;
 	uint32_t		xs_qm_dquot_dups;

[PATCH 34/37] xfs: hook live realtime rmap operations during a repair operation

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Hook the regular realtime rmap code when an rtrmapbt repair operation is
running so that we can unlock the AGF buffer to scan the filesystem and
keep the in-memory btree up to date during the scan.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/scrub/rmap_repair.c   |    2 -
 fs/xfs/scrub/rtrmap_repair.c |  131 +++++++++++++++++++++++++++++++++++++++++-
 fs/xfs/scrub/trace.h         |   17 ++++-
 3 files changed, 140 insertions(+), 10 deletions(-)


diff --git a/fs/xfs/scrub/rmap_repair.c b/fs/xfs/scrub/rmap_repair.c
index 91c17feb49768b..c2c7b76cc25ab8 100644
--- a/fs/xfs/scrub/rmap_repair.c
+++ b/fs/xfs/scrub/rmap_repair.c
@@ -1614,7 +1614,7 @@ xrep_rmapbt_live_update(
 	if (!xrep_rmapbt_want_live_update(&rr->iscan, &p->oinfo))
 		goto out_unlock;
 
-	trace_xrep_rmap_live_update(rr->sc->sa.pag, action, p);
+	trace_xrep_rmap_live_update(pag_group(rr->sc->sa.pag), action, p);
 
 	error = xrep_trans_alloc_hook_dummy(mp, &txcookie, &tp);
 	if (error)
diff --git a/fs/xfs/scrub/rtrmap_repair.c b/fs/xfs/scrub/rtrmap_repair.c
index b376bcc8d1d2ed..49de8bc2dd17f5 100644
--- a/fs/xfs/scrub/rtrmap_repair.c
+++ b/fs/xfs/scrub/rtrmap_repair.c
@@ -71,6 +71,9 @@ struct xrep_rtrmap {
 	/* new rtrmapbt information */
 	struct xrep_newbt	new_btree;
 
+	/* lock for the xfbtree and xfile */
+	struct mutex		lock;
+
 	/* rmap records generated from primary metadata */
 	struct xfbtree		rtrmap_btree;
 
@@ -79,6 +82,9 @@ struct xrep_rtrmap {
 	/* bitmap of old rtrmapbt blocks */
 	struct xfsb_bitmap	old_rtrmapbt_blocks;
 
+	/* Hooks into rtrmap update code. */
+	struct xfs_rmap_hook	rhook;
+
 	/* inode scan cursor */
 	struct xchk_iscan	iscan;
 
@@ -98,6 +104,8 @@ xrep_setup_rtrmapbt(
 	char			*descr;
 	int			error;
 
+	xchk_fsgates_enable(sc, XCHK_FSGATES_RMAP);
+
 	descr = xchk_xfile_rtgroup_descr(sc, "reverse mapping records");
 	error = xrep_setup_xfbtree(sc, descr);
 	kfree(descr);
@@ -151,19 +159,31 @@ xrep_rtrmap_stash(
 	if (xchk_should_terminate(sc, &error))
 		return error;
 
+	if (xchk_iscan_aborted(&rr->iscan))
+		return -EFSCORRUPTED;
+
 	trace_xrep_rtrmap_found(sc->mp, &rmap);
 
 	/* Add entry to in-memory btree. */
+	mutex_lock(&rr->lock);
 	mcur = xfs_rtrmapbt_mem_cursor(sc->sr.rtg, sc->tp, &rr->rtrmap_btree);
 	error = xfs_rmap_map_raw(mcur, &rmap);
 	xfs_btree_del_cursor(mcur, error);
 	if (error)
 		goto out_cancel;
 
-	return xfbtree_trans_commit(&rr->rtrmap_btree, sc->tp);
+	error = xfbtree_trans_commit(&rr->rtrmap_btree, sc->tp);
+	if (error)
+		goto out_abort;
+
+	mutex_unlock(&rr->lock);
+	return 0;
 
 out_cancel:
 	xfbtree_trans_cancel(&rr->rtrmap_btree, sc->tp);
+out_abort:
+	xchk_iscan_abort(&rr->iscan);
+	mutex_unlock(&rr->lock);
 	return error;
 }
 
@@ -486,6 +506,13 @@ xrep_rtrmap_find_rmaps(
 	if (error)
 		return error;
 
+	/*
+	 * If a hook failed to update the in-memory btree, we lack the data to
+	 * continue the repair.
+	 */
+	if (xchk_iscan_aborted(&rr->iscan))
+		return -EFSCORRUPTED;
+
 	/* Scan for old rtrmap blocks. */
 	while ((pag = xfs_perag_next(sc->mp, pag))) {
 		error = xrep_rtrmap_scan_ag(rr, pag);
@@ -702,6 +729,83 @@ xrep_rtrmap_remove_old_tree(
 	return xrep_reset_metafile_resv(rr->sc);
 }
 
+static inline bool
+xrep_rtrmapbt_want_live_update(
+	struct xchk_iscan		*iscan,
+	const struct xfs_owner_info	*oi)
+{
+	if (xchk_iscan_aborted(iscan))
+		return false;
+
+	/*
+	 * We scanned the CoW staging extents before we started the iscan, so
+	 * we need all the updates.
+	 */
+	if (XFS_RMAP_NON_INODE_OWNER(oi->oi_owner))
+		return true;
+
+	/* Ignore updates to files that the scanner hasn't visited yet. */
+	return xchk_iscan_want_live_update(iscan, oi->oi_owner);
+}
+
+/*
+ * Apply a rtrmapbt update from the regular filesystem into our shadow btree.
+ * We're running from the thread that owns the rtrmap ILOCK and is generating
+ * the update, so we must be careful about which parts of the struct
+ * xrep_rtrmap that we change.
+ */
+static int
+xrep_rtrmapbt_live_update(
+	struct notifier_block		*nb,
+	unsigned long			action,
+	void				*data)
+{
+	struct xfs_rmap_update_params	*p = data;
+	struct xrep_rtrmap		*rr;
+	struct xfs_mount		*mp;
+	struct xfs_btree_cur		*mcur;
+	struct xfs_trans		*tp;
+	void				*txcookie;
+	int				error;
+
+	rr = container_of(nb, struct xrep_rtrmap, rhook.rmap_hook.nb);
+	mp = rr->sc->mp;
+
+	if (!xrep_rtrmapbt_want_live_update(&rr->iscan, &p->oinfo))
+		goto out_unlock;
+
+	trace_xrep_rmap_live_update(rtg_group(rr->sc->sr.rtg), action, p);
+
+	error = xrep_trans_alloc_hook_dummy(mp, &txcookie, &tp);
+	if (error)
+		goto out_abort;
+
+	mutex_lock(&rr->lock);
+	mcur = xfs_rtrmapbt_mem_cursor(rr->sc->sr.rtg, tp, &rr->rtrmap_btree);
+	error = __xfs_rmap_finish_intent(mcur, action, p->startblock,
+			p->blockcount, &p->oinfo, p->unwritten);
+	xfs_btree_del_cursor(mcur, error);
+	if (error)
+		goto out_cancel;
+
+	error = xfbtree_trans_commit(&rr->rtrmap_btree, tp);
+	if (error)
+		goto out_cancel;
+
+	xrep_trans_cancel_hook_dummy(&txcookie, tp);
+	mutex_unlock(&rr->lock);
+	return NOTIFY_DONE;
+
+out_cancel:
+	xfbtree_trans_cancel(&rr->rtrmap_btree, tp);
+	xrep_trans_cancel_hook_dummy(&txcookie, tp);
+out_abort:
+	xchk_iscan_abort(&rr->iscan);
+	mutex_unlock(&rr->lock);
+out_unlock:
+	return NOTIFY_DONE;
+}
+
 /* Set up the filesystem scan components. */
 STATIC int
 xrep_rtrmap_setup_scan(
@@ -710,6 +814,7 @@ xrep_rtrmap_setup_scan(
 	struct xfs_scrub	*sc = rr->sc;
 	int			error;
 
+	mutex_init(&rr->lock);
 	xfsb_bitmap_init(&rr->old_rtrmapbt_blocks);
 
 	/* Set up some storage */
@@ -720,10 +825,26 @@ xrep_rtrmap_setup_scan(
 
 	/* Retry iget every tenth of a second for up to 30 seconds. */
 	xchk_iscan_start(sc, 30000, 100, &rr->iscan);
+
+	/*
+	 * Hook into live rtrmap operations so that we can update our in-memory
+	 * btree to reflect live changes on the filesystem.  Since we drop the
+	 * rtrmap ILOCK to scan all the inodes, we need this piece to avoid
+	 * installing a stale btree.
+	 */
+	ASSERT(sc->flags & XCHK_FSGATES_RMAP);
+	xfs_rmap_hook_setup(&rr->rhook, xrep_rtrmapbt_live_update);
+	error = xfs_rmap_hook_add(rtg_group(sc->sr.rtg), &rr->rhook);
+	if (error)
+		goto out_iscan;
 	return 0;
 
+out_iscan:
+	xchk_iscan_teardown(&rr->iscan);
+	xfbtree_destroy(&rr->rtrmap_btree);
 out_bitmap:
 	xfsb_bitmap_destroy(&rr->old_rtrmapbt_blocks);
+	mutex_destroy(&rr->lock);
 	return error;
 }
 
@@ -732,9 +853,14 @@ STATIC void
 xrep_rtrmap_teardown(
 	struct xrep_rtrmap	*rr)
 {
+	struct xfs_scrub	*sc = rr->sc;
+
+	xchk_iscan_abort(&rr->iscan);
+	xfs_rmap_hook_del(rtg_group(sc->sr.rtg), &rr->rhook);
 	xchk_iscan_teardown(&rr->iscan);
 	xfbtree_destroy(&rr->rtrmap_btree);
 	xfsb_bitmap_destroy(&rr->old_rtrmapbt_blocks);
+	mutex_destroy(&rr->lock);
 }
 
 /* Repair the realtime rmap btree. */
@@ -745,9 +871,6 @@ xrep_rtrmapbt(
 	struct xrep_rtrmap	*rr = sc->buf;
 	int			error;
 
-	/* Functionality is not yet complete. */
-	return xrep_notsupported(sc);
-
 	/* Make sure any problems with the fork are fixed. */
 	error = xrep_metadata_inode_forks(sc);
 	if (error)
diff --git a/fs/xfs/scrub/trace.h b/fs/xfs/scrub/trace.h
index 3f2a8695ef5cb5..fb86b746bc174a 100644
--- a/fs/xfs/scrub/trace.h
+++ b/fs/xfs/scrub/trace.h
@@ -41,6 +41,9 @@ struct xchk_dirtree_outcomes;
 TRACE_DEFINE_ENUM(XFS_REFC_DOMAIN_SHARED);
 TRACE_DEFINE_ENUM(XFS_REFC_DOMAIN_COW);
 
+TRACE_DEFINE_ENUM(XG_TYPE_AG);
+TRACE_DEFINE_ENUM(XG_TYPE_RTG);
+
 TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_PROBE);
 TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_SB);
 TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_AGF);
@@ -2709,11 +2712,12 @@ DEFINE_SCRUB_NLINKS_DIFF_EVENT(xrep_nlinks_update_inode);
 DEFINE_SCRUB_NLINKS_DIFF_EVENT(xrep_nlinks_unfixable_inode);
 
 TRACE_EVENT(xrep_rmap_live_update,
-	TP_PROTO(const struct xfs_perag *pag, unsigned int op,
+	TP_PROTO(const struct xfs_group *xg, unsigned int op,
 		 const struct xfs_rmap_update_params *p),
-	TP_ARGS(pag, op, p),
+	TP_ARGS(xg, op, p),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
+		__field(enum xfs_group_type, type)
 		__field(xfs_agnumber_t, agno)
 		__field(unsigned int, op)
 		__field(xfs_agblock_t, agbno)
@@ -2723,8 +2727,9 @@ TRACE_EVENT(xrep_rmap_live_update,
 		__field(unsigned int, flags)
 	),
 	TP_fast_assign(
-		__entry->dev = pag_mount(pag)->m_super->s_dev;
-		__entry->agno = pag_agno(pag);
+		__entry->dev = xg->xg_mount->m_super->s_dev;
+		__entry->type = xg->xg_type;
+		__entry->agno = xg->xg_gno;
 		__entry->op = op;
 		__entry->agbno = p->startblock;
 		__entry->len = p->blockcount;
@@ -2733,10 +2738,12 @@ TRACE_EVENT(xrep_rmap_live_update,
 		if (p->unwritten)
 			__entry->flags |= XFS_RMAP_UNWRITTEN;
 	),
-	TP_printk("dev %d:%d agno 0x%x op %d agbno 0x%x fsbcount 0x%x owner 0x%llx fileoff 0x%llx flags 0x%x",
+	TP_printk("dev %d:%d %sno 0x%x op %d %sbno 0x%x fsbcount 0x%x owner 0x%llx fileoff 0x%llx flags 0x%x",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __print_symbolic(__entry->type, XG_TYPE_STRINGS),
 		  __entry->agno,
 		  __entry->op,
+		  __print_symbolic(__entry->type, XG_TYPE_STRINGS),
 		  __entry->agbno,
 		  __entry->len,
 		  __entry->owner,

[PATCH 35/37] xfs: don't shut down the filesystem for media failures beyond end of log

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

If the filesystem has an external log device on pmem and the pmem
reports a media error beyond the end of the log area, don't shut down
the filesystem because we don't use that space.

Cc: <stable@vger.kernel.org> # v6.0
Fixes: 6f643c57d57c56 ("xfs: implement ->notify_failure() for XFS")
Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
---
 fs/xfs/xfs_notify_failure.c |  121 +++++++++++++++++++++++++++++--------------
 1 file changed, 82 insertions(+), 39 deletions(-)


diff --git a/fs/xfs/xfs_notify_failure.c b/fs/xfs/xfs_notify_failure.c
index fa50e5308292d3..0b0b0f31aca274 100644
--- a/fs/xfs/xfs_notify_failure.c
+++ b/fs/xfs/xfs_notify_failure.c
@@ -153,6 +153,79 @@ xfs_dax_notify_failure_thaw(
 	thaw_super(sb, FREEZE_HOLDER_USERSPACE);
 }
 
+static int
+xfs_dax_translate_range(
+	struct xfs_buftarg	*btp,
+	u64			offset,
+	u64			len,
+	xfs_daddr_t		*daddr,
+	uint64_t		*bblen)
+{
+	u64			dev_start = btp->bt_dax_part_off;
+	u64			dev_len = bdev_nr_bytes(btp->bt_bdev);
+	u64			dev_end = dev_start + dev_len - 1;
+
+	/* Notify failure on the whole device. */
+	if (offset == 0 && len == U64_MAX) {
+		offset = dev_start;
+		len = dev_len;
+	}
+
+	/* Ignore the range out of filesystem area */
+	if (offset + len - 1 < dev_start)
+		return -ENXIO;
+	if (offset > dev_end)
+		return -ENXIO;
+
+	/* Calculate the real range when it touches the boundary */
+	if (offset > dev_start)
+		offset -= dev_start;
+	else {
+		len -= dev_start - offset;
+		offset = 0;
+	}
+	if (offset + len - 1 > dev_end)
+		len = dev_end - offset + 1;
+
+	*daddr = BTOBB(offset);
+	*bblen = BTOBB(len);
+	return 0;
+}
+
+static int
+xfs_dax_notify_logdev_failure(
+	struct xfs_mount	*mp,
+	u64			offset,
+	u64			len,
+	int			mf_flags)
+{
+	xfs_daddr_t		daddr;
+	uint64_t		bblen;
+	int			error;
+
+	/*
+	 * Return ENXIO instead of shutting down the filesystem if the failed
+	 * region is beyond the end of the log.
+	 */
+	error = xfs_dax_translate_range(mp->m_logdev_targp,
+			offset, len, &daddr, &bblen);
+	if (error)
+		return error;
+
+	/*
+	 * In the pre-remove case the failure notification is attempting to
+	 * trigger a force unmount.  The expectation is that the device is
+	 * still present, but its removal is in progress and can not be
+	 * cancelled, proceed with accessing the log device.
+	 */
+	if (mf_flags & MF_MEM_PRE_REMOVE)
+		return 0;
+
+	xfs_err(mp, "ondisk log corrupt, shutting down fs!");
+	xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_ONDISK);
+	return -EFSCORRUPTED;
+}
+
 static int
 xfs_dax_notify_ddev_failure(
 	struct xfs_mount	*mp,
@@ -263,8 +336,9 @@ xfs_dax_notify_failure(
 	int			mf_flags)
 {
 	struct xfs_mount	*mp = dax_holder(dax_dev);
-	u64			ddev_start;
-	u64			ddev_end;
+	xfs_daddr_t		daddr;
+	uint64_t		bblen;
+	int			error;
 
 	if (!(mp->m_super->s_flags & SB_BORN)) {
 		xfs_warn(mp, "filesystem is not ready for notify_failure()!");
@@ -279,17 +353,7 @@ xfs_dax_notify_failure(
 
 	if (mp->m_logdev_targp && mp->m_logdev_targp->bt_daxdev == dax_dev &&
 	    mp->m_logdev_targp != mp->m_ddev_targp) {
-		/*
-		 * In the pre-remove case the failure notification is attempting
-		 * to trigger a force unmount.  The expectation is that the
-		 * device is still present, but its removal is in progress and
-		 * can not be cancelled, proceed with accessing the log device.
-		 */
-		if (mf_flags & MF_MEM_PRE_REMOVE)
-			return 0;
-		xfs_err(mp, "ondisk log corrupt, shutting down fs!");
-		xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_ONDISK);
-		return -EFSCORRUPTED;
+		return xfs_dax_notify_logdev_failure(mp, offset, len, mf_flags);
 	}
 
 	if (!xfs_has_rmapbt(mp)) {
@@ -297,33 +361,12 @@ xfs_dax_notify_failure(
 		return -EOPNOTSUPP;
 	}
 
-	ddev_start = mp->m_ddev_targp->bt_dax_part_off;
-	ddev_end = ddev_start + bdev_nr_bytes(mp->m_ddev_targp->bt_bdev) - 1;
+	error = xfs_dax_translate_range(mp->m_ddev_targp, offset, len, &daddr,
+			&bblen);
+	if (error)
+		return error;
 
-	/* Notify failure on the whole device. */
-	if (offset == 0 && len == U64_MAX) {
-		offset = ddev_start;
-		len = bdev_nr_bytes(mp->m_ddev_targp->bt_bdev);
-	}
-
-	/* Ignore the range out of filesystem area */
-	if (offset + len - 1 < ddev_start)
-		return -ENXIO;
-	if (offset > ddev_end)
-		return -ENXIO;
-
-	/* Calculate the real range when it touches the boundary */
-	if (offset > ddev_start)
-		offset -= ddev_start;
-	else {
-		len -= ddev_start - offset;
-		offset = 0;
-	}
-	if (offset + len - 1 > ddev_end)
-		len = ddev_end - offset + 1;
-
-	return xfs_dax_notify_ddev_failure(mp, BTOBB(offset), BTOBB(len),
-			mf_flags);
+	return xfs_dax_notify_ddev_failure(mp, daddr, bblen, mf_flags);
 }
 
 const struct dax_holder_operations xfs_dax_holder_operations = {

Re: [PATCH 35/37] xfs: don't shut down the filesystem for media failures beyond end of log

[Christoph Hellwig]

Looks good:

Reviewed-by: Christoph Hellwig <hch@lst.de>

[PATCH 36/37] xfs: react to fsdax failure notifications on the rt device

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Now that we have reverse mapping for the realtime device, use the
information to kill processes that have mappings to bad pmem.  This
requires refactoring the existing routines to handle rtgroups or AGs;
and splitting out the translation function to improve cohesion.
Also make a proper header file for the dax holder ops.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
---
 fs/xfs/xfs_buf.c            |    1 
 fs/xfs/xfs_notify_failure.c |  127 ++++++++++++++++++++++++-------------------
 fs/xfs/xfs_notify_failure.h |   11 ++++
 fs/xfs/xfs_super.h          |    1 
 4 files changed, 84 insertions(+), 56 deletions(-)
 create mode 100644 fs/xfs/xfs_notify_failure.h


diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c
index aa63b8efd78228..6f313fbf766910 100644
--- a/fs/xfs/xfs_buf.c
+++ b/fs/xfs/xfs_buf.c
@@ -22,6 +22,7 @@
 #include "xfs_error.h"
 #include "xfs_ag.h"
 #include "xfs_buf_mem.h"
+#include "xfs_notify_failure.h"
 
 struct kmem_cache *xfs_buf_cache;
 
diff --git a/fs/xfs/xfs_notify_failure.c b/fs/xfs/xfs_notify_failure.c
index 0b0b0f31aca274..ed8d8ed42f0a2c 100644
--- a/fs/xfs/xfs_notify_failure.c
+++ b/fs/xfs/xfs_notify_failure.c
@@ -19,6 +19,9 @@
 #include "xfs_rtalloc.h"
 #include "xfs_trans.h"
 #include "xfs_ag.h"
+#include "xfs_notify_failure.h"
+#include "xfs_rtgroup.h"
+#include "xfs_rtrmap_btree.h"
 
 #include <linux/mm.h>
 #include <linux/dax.h>
@@ -227,23 +230,42 @@ xfs_dax_notify_logdev_failure(
 }
 
 static int
-xfs_dax_notify_ddev_failure(
+xfs_dax_notify_dev_failure(
 	struct xfs_mount	*mp,
-	xfs_daddr_t		daddr,
-	xfs_daddr_t		bblen,
-	int			mf_flags)
+	u64			offset,
+	u64			len,
+	int			mf_flags,
+	enum xfs_group_type	type)
 {
 	struct xfs_failure_info	notify = { .mf_flags = mf_flags };
 	struct xfs_trans	*tp = NULL;
 	struct xfs_btree_cur	*cur = NULL;
-	struct xfs_buf		*agf_bp = NULL;
 	int			error = 0;
 	bool			kernel_frozen = false;
-	xfs_fsblock_t		fsbno = XFS_DADDR_TO_FSB(mp, daddr);
-	xfs_agnumber_t		agno = XFS_FSB_TO_AGNO(mp, fsbno);
-	xfs_fsblock_t		end_fsbno = XFS_DADDR_TO_FSB(mp,
-							     daddr + bblen - 1);
-	xfs_agnumber_t		end_agno = XFS_FSB_TO_AGNO(mp, end_fsbno);
+	uint32_t		start_gno, end_gno;
+	xfs_fsblock_t		start_bno, end_bno;
+	xfs_daddr_t		daddr;
+	uint64_t		bblen;
+	struct xfs_group	*xg = NULL;
+
+	if (!xfs_has_rmapbt(mp)) {
+		xfs_debug(mp, "notify_failure() needs rmapbt enabled!");
+		return -EOPNOTSUPP;
+	}
+
+	error = xfs_dax_translate_range(type == XG_TYPE_RTG ?
+			mp->m_rtdev_targp : mp->m_ddev_targp,
+			offset, len, &daddr, &bblen);
+	if (error)
+		return error;
+
+	if (type == XG_TYPE_RTG) {
+		start_bno = xfs_daddr_to_rtb(mp, daddr);
+		end_bno = xfs_daddr_to_rtb(mp, daddr + bblen - 1);
+	} else {
+		start_bno = XFS_DADDR_TO_FSB(mp, daddr);
+		end_bno = XFS_DADDR_TO_FSB(mp, daddr + bblen - 1);
+	}
 
 	if (mf_flags & MF_MEM_PRE_REMOVE) {
 		xfs_info(mp, "Device is about to be removed!");
@@ -262,46 +284,58 @@ xfs_dax_notify_ddev_failure(
 	if (error)
 		goto out;
 
-	for (; agno <= end_agno; agno++) {
+	start_gno = xfs_fsb_to_gno(mp, start_bno, type);
+	end_gno = xfs_fsb_to_gno(mp, end_bno, type);
+	while ((xg = xfs_group_next_range(mp, xg, start_gno, end_gno, type))) {
+		struct xfs_buf		*agf_bp = NULL;
+		struct xfs_rtgroup	*rtg = NULL;
 		struct xfs_rmap_irec	ri_low = { };
 		struct xfs_rmap_irec	ri_high;
-		struct xfs_agf		*agf;
-		struct xfs_perag	*pag;
-		xfs_agblock_t		range_agend;
 
-		pag = xfs_perag_get(mp, agno);
-		error = xfs_alloc_read_agf(pag, tp, 0, &agf_bp);
-		if (error) {
-			xfs_perag_put(pag);
-			break;
+		if (type == XG_TYPE_AG) {
+			struct xfs_perag	*pag = to_perag(xg);
+
+			error = xfs_alloc_read_agf(pag, tp, 0, &agf_bp);
+			if (error) {
+				xfs_perag_put(pag);
+				break;
+			}
+
+			cur = xfs_rmapbt_init_cursor(mp, tp, agf_bp, pag);
+		} else {
+			rtg = to_rtg(xg);
+			xfs_rtgroup_lock(rtg, XFS_RTGLOCK_RMAP);
+			cur = xfs_rtrmapbt_init_cursor(tp, rtg);
 		}
 
-		cur = xfs_rmapbt_init_cursor(mp, tp, agf_bp, pag);
-
 		/*
 		 * Set the rmap range from ri_low to ri_high, which represents
 		 * a [start, end] where we looking for the files or metadata.
 		 */
 		memset(&ri_high, 0xFF, sizeof(ri_high));
-		ri_low.rm_startblock = XFS_FSB_TO_AGBNO(mp, fsbno);
-		if (agno == end_agno)
-			ri_high.rm_startblock = XFS_FSB_TO_AGBNO(mp, end_fsbno);
+		if (xg->xg_gno == start_gno)
+			ri_low.rm_startblock =
+				xfs_fsb_to_gbno(mp, start_bno, type);
+		if (xg->xg_gno == end_gno)
+			ri_high.rm_startblock =
+				xfs_fsb_to_gbno(mp, end_bno, type);
 
-		agf = agf_bp->b_addr;
-		range_agend = min(be32_to_cpu(agf->agf_length) - 1,
-				ri_high.rm_startblock);
 		notify.startblock = ri_low.rm_startblock;
-		notify.blockcount = range_agend + 1 - ri_low.rm_startblock;
+		notify.blockcount = min(xg->xg_block_count,
+					ri_high.rm_startblock + 1) -
+					ri_low.rm_startblock;
 
 		error = xfs_rmap_query_range(cur, &ri_low, &ri_high,
 				xfs_dax_failure_fn, &notify);
 		xfs_btree_del_cursor(cur, error);
-		xfs_trans_brelse(tp, agf_bp);
-		xfs_perag_put(pag);
-		if (error)
+		if (agf_bp)
+			xfs_trans_brelse(tp, agf_bp);
+		if (rtg)
+			xfs_rtgroup_unlock(rtg, XFS_RTGLOCK_RMAP);
+		if (error) {
+			xfs_group_put(xg);
 			break;
-
-		fsbno = XFS_AGB_TO_FSB(mp, agno + 1, 0);
+		}
 	}
 
 	xfs_trans_cancel(tp);
@@ -336,37 +370,20 @@ xfs_dax_notify_failure(
 	int			mf_flags)
 {
 	struct xfs_mount	*mp = dax_holder(dax_dev);
-	xfs_daddr_t		daddr;
-	uint64_t		bblen;
-	int			error;
 
 	if (!(mp->m_super->s_flags & SB_BORN)) {
 		xfs_warn(mp, "filesystem is not ready for notify_failure()!");
 		return -EIO;
 	}
 
-	if (mp->m_rtdev_targp && mp->m_rtdev_targp->bt_daxdev == dax_dev) {
-		xfs_debug(mp,
-			 "notify_failure() not supported on realtime device!");
-		return -EOPNOTSUPP;
-	}
-
-	if (mp->m_logdev_targp && mp->m_logdev_targp->bt_daxdev == dax_dev &&
-	    mp->m_logdev_targp != mp->m_ddev_targp) {
+	if (mp->m_logdev_targp != mp->m_ddev_targp &&
+	    mp->m_logdev_targp->bt_daxdev == dax_dev) {
 		return xfs_dax_notify_logdev_failure(mp, offset, len, mf_flags);
 	}
 
-	if (!xfs_has_rmapbt(mp)) {
-		xfs_debug(mp, "notify_failure() needs rmapbt enabled!");
-		return -EOPNOTSUPP;
-	}
-
-	error = xfs_dax_translate_range(mp->m_ddev_targp, offset, len, &daddr,
-			&bblen);
-	if (error)
-		return error;
-
-	return xfs_dax_notify_ddev_failure(mp, daddr, bblen, mf_flags);
+	return xfs_dax_notify_dev_failure(mp, offset, len, mf_flags,
+		(mp->m_rtdev_targp && mp->m_rtdev_targp->bt_daxdev == dax_dev) ?
+				XG_TYPE_RTG : XG_TYPE_AG);
 }
 
 const struct dax_holder_operations xfs_dax_holder_operations = {
diff --git a/fs/xfs/xfs_notify_failure.h b/fs/xfs/xfs_notify_failure.h
new file mode 100644
index 00000000000000..8d08ec29dd2949
--- /dev/null
+++ b/fs/xfs/xfs_notify_failure.h
@@ -0,0 +1,11 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright (C) 2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_NOTIFY_FAILURE_H__
+#define __XFS_NOTIFY_FAILURE_H__
+
+extern const struct dax_holder_operations xfs_dax_holder_operations;
+
+#endif /* __XFS_NOTIFY_FAILURE_H__ */
diff --git a/fs/xfs/xfs_super.h b/fs/xfs/xfs_super.h
index 302e6e5d6c7e20..c0e85c1e42f27d 100644
--- a/fs/xfs/xfs_super.h
+++ b/fs/xfs/xfs_super.h
@@ -92,7 +92,6 @@ extern xfs_agnumber_t xfs_set_inode_alloc(struct xfs_mount *,
 
 extern const struct export_operations xfs_export_operations;
 extern const struct quotactl_ops xfs_quotactl_operations;
-extern const struct dax_holder_operations xfs_dax_holder_operations;
 
 extern void xfs_reinit_percpu_counters(struct xfs_mount *mp);
 

Re: [PATCH 36/37] xfs: react to fsdax failure notifications on the rt device

[Christoph Hellwig]

On Thu, Dec 19, 2024 at 11:32:46AM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <djwong@kernel.org>
> 
> Now that we have reverse mapping for the realtime device, use the
> information to kill processes that have mappings to bad pmem.  This
> requires refactoring the existing routines to handle rtgroups or AGs;
> and splitting out the translation function to improve cohesion.
> Also make a proper header file for the dax holder ops.

Looks good:

Reviewed-by: Christoph Hellwig <hch@lst.de>

[PATCH 37/37] xfs: enable realtime rmap btree

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Permit mounting filesystems with realtime rmap btrees.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
---
 fs/xfs/xfs_rtalloc.c |   12 ++++++++----
 fs/xfs/xfs_super.c   |    6 ------
 2 files changed, 8 insertions(+), 10 deletions(-)


diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c
index 3c1bce5a4855f2..a69967f9d88ead 100644
--- a/fs/xfs/xfs_rtalloc.c
+++ b/fs/xfs/xfs_rtalloc.c
@@ -1282,11 +1282,15 @@ xfs_growfs_rt(
 	    XFS_FSB_TO_B(mp, in->extsize) < XFS_MIN_RTEXTSIZE)
 		goto out_unlock;
 
-	/* Unsupported realtime features. */
+	/* Check for features supported only on rtgroups filesystems. */
 	error = -EOPNOTSUPP;
-	if (xfs_has_quota(mp) && !xfs_has_rtgroups(mp))
-		goto out_unlock;
-	if (xfs_has_rmapbt(mp) || xfs_has_reflink(mp))
+	if (!xfs_has_rtgroups(mp)) {
+		if (xfs_has_rmapbt(mp))
+			goto out_unlock;
+		if (xfs_has_quota(mp))
+			goto out_unlock;
+	}
+	if (xfs_has_reflink(mp))
 		goto out_unlock;
 
 	error = xfs_sb_validate_fsb_count(&mp->m_sb, in->newblocks);
diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c
index 394fdf3bb53531..ecd5a9f444d862 100644
--- a/fs/xfs/xfs_super.c
+++ b/fs/xfs/xfs_super.c
@@ -1767,12 +1767,6 @@ xfs_fs_fill_super(
 		}
 	}
 
-	if (xfs_has_rmapbt(mp) && mp->m_sb.sb_rblocks) {
-		xfs_alert(mp,
-	"reverse mapping btree not compatible with realtime device!");
-		error = -EINVAL;
-		goto out_filestream_unmount;
-	}
 
 	if (xfs_has_exchange_range(mp))
 		xfs_warn_experimental(mp, XFS_EXPERIMENTAL_EXCHRANGE);

Re: [PATCH 37/37] xfs: enable realtime rmap btree

[Christoph Hellwig]

Looks good:

Reviewed-by: Christoph Hellwig <hch@lst.de>

[PATCHSET v6.1 5/5] xfs: reflink on the realtime device

[Darrick J. Wong]

Hi all,

This patchset enables use of the file data block sharing feature (i.e.
reflink) on the realtime device.  It follows the same basic sequence as
the realtime rmap series -- first a few cleanups; then  introduction of
the new btree format and inode fork format.  Next comes enabling CoW and
remapping for the rt device; new scrub, repair, and health reporting
code; and at the end we implement some code to lengthen write requests
so that rt extents are always CoWed fully.

If you're going to start using this code, I strongly recommend pulling
from my git trees, which are linked below.

This has been running on the djcloud for months with no problems.  Enjoy!
Comments and questions are, as always, welcome.

--D

kernel git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfs-linux.git/log/?h=realtime-reflink

xfsprogs git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfsprogs-dev.git/log/?h=realtime-reflink

fstests git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfstests-dev.git/log/?h=realtime-reflink

xfsdocs git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfs-documentation.git/log/?h=realtime-reflink
---
Commits in this patchset:
 * xfs: prepare refcount btree cursor tracepoints for realtime
 * xfs: namespace the maximum length/refcount symbols
 * xfs: introduce realtime refcount btree ondisk definitions
 * xfs: realtime refcount btree transaction reservations
 * xfs: add realtime refcount btree operations
 * xfs: prepare refcount functions to deal with rtrefcountbt
 * xfs: add a realtime flag to the refcount update log redo items
 * xfs: support recovering refcount intent items targetting realtime extents
 * xfs: add realtime refcount btree block detection to log recovery
 * xfs: add realtime refcount btree inode to metadata directory
 * xfs: add metadata reservations for realtime refcount btree
 * xfs: wire up a new metafile type for the realtime refcount
 * xfs: refactor xfs_reflink_find_shared
 * xfs: wire up realtime refcount btree cursors
 * xfs: create routine to allocate and initialize a realtime refcount btree inode
 * xfs: update rmap to allow cow staging extents in the rt rmap
 * xfs: compute rtrmap btree max levels when reflink enabled
 * xfs: refactor reflink quota updates
 * xfs: enable CoW for realtime data
 * xfs: enable sharing of realtime file blocks
 * xfs: allow inodes to have the realtime and reflink flags
 * xfs: recover CoW leftovers in the realtime volume
 * xfs: fix xfs_get_extsz_hint behavior with realtime alwayscow files
 * xfs: apply rt extent alignment constraints to CoW extsize hint
 * xfs: enable extent size hints for CoW operations
 * xfs: check that the rtrefcount maxlevels doesn't increase when growing fs
 * xfs: report realtime refcount btree corruption errors to the health system
 * xfs: scrub the realtime refcount btree
 * xfs: cross-reference checks with the rt refcount btree
 * xfs: allow overlapping rtrmapbt records for shared data extents
 * xfs: check reference counts of gaps between rt refcount records
 * xfs: allow dquot rt block count to exceed rt blocks on reflink fs
 * xfs: detect and repair misaligned rtinherit directory cowextsize hints
 * xfs: scrub the metadir path of rt refcount btree files
 * xfs: don't flag quota rt block usage on rtreflink filesystems
 * xfs: check new rtbitmap records against rt refcount btree
 * xfs: walk the rt reference count tree when rebuilding rmap
 * xfs: capture realtime CoW staging extents when rebuilding rt rmapbt
 * xfs: online repair of the realtime refcount btree
 * xfs: repair inodes that have a refcount btree in the data fork
 * xfs: check for shared rt extents when rebuilding rt file's data fork
 * xfs: fix CoW forks for realtime files
 * xfs: enable realtime reflink
---
 fs/xfs/Makefile                      |    3 
 fs/xfs/libxfs/xfs_bmap.c             |   23 +
 fs/xfs/libxfs/xfs_btree.c            |    5 
 fs/xfs/libxfs/xfs_btree.h            |    2 
 fs/xfs/libxfs/xfs_defer.h            |    1 
 fs/xfs/libxfs/xfs_format.h           |   25 +
 fs/xfs/libxfs/xfs_fs.h               |    7 
 fs/xfs/libxfs/xfs_health.h           |    4 
 fs/xfs/libxfs/xfs_inode_buf.c        |   33 +
 fs/xfs/libxfs/xfs_inode_fork.c       |    6 
 fs/xfs/libxfs/xfs_log_format.h       |    6 
 fs/xfs/libxfs/xfs_log_recover.h      |    2 
 fs/xfs/libxfs/xfs_ondisk.h           |    2 
 fs/xfs/libxfs/xfs_refcount.c         |  276 ++++++++++--
 fs/xfs/libxfs/xfs_refcount.h         |   23 +
 fs/xfs/libxfs/xfs_rmap.c             |    7 
 fs/xfs/libxfs/xfs_rtgroup.c          |   19 +
 fs/xfs/libxfs/xfs_rtgroup.h          |   11 
 fs/xfs/libxfs/xfs_rtrefcount_btree.c |  757 +++++++++++++++++++++++++++++++++
 fs/xfs/libxfs/xfs_rtrefcount_btree.h |  189 ++++++++
 fs/xfs/libxfs/xfs_rtrmap_btree.c     |   28 +
 fs/xfs/libxfs/xfs_sb.c               |    8 
 fs/xfs/libxfs/xfs_shared.h           |    7 
 fs/xfs/libxfs/xfs_trans_resv.c       |   25 +
 fs/xfs/scrub/agheader_repair.c       |    2 
 fs/xfs/scrub/bmap.c                  |   30 +
 fs/xfs/scrub/bmap_repair.c           |   21 +
 fs/xfs/scrub/common.c                |   10 
 fs/xfs/scrub/common.h                |    5 
 fs/xfs/scrub/cow_repair.c            |  180 +++++++-
 fs/xfs/scrub/health.c                |    1 
 fs/xfs/scrub/inode.c                 |   31 +
 fs/xfs/scrub/inode_repair.c          |   57 ++
 fs/xfs/scrub/metapath.c              |    3 
 fs/xfs/scrub/quota.c                 |    8 
 fs/xfs/scrub/quota_repair.c          |    2 
 fs/xfs/scrub/reap.c                  |  247 ++++++++++-
 fs/xfs/scrub/reap.h                  |    7 
 fs/xfs/scrub/refcount.c              |    2 
 fs/xfs/scrub/refcount_repair.c       |    6 
 fs/xfs/scrub/repair.c                |    6 
 fs/xfs/scrub/repair.h                |    7 
 fs/xfs/scrub/rgb_bitmap.h            |   37 ++
 fs/xfs/scrub/rmap_repair.c           |    7 
 fs/xfs/scrub/rtb_bitmap.h            |   37 ++
 fs/xfs/scrub/rtbitmap.c              |    2 
 fs/xfs/scrub/rtbitmap_repair.c       |   24 +
 fs/xfs/scrub/rtrefcount.c            |  661 +++++++++++++++++++++++++++++
 fs/xfs/scrub/rtrefcount_repair.c     |  783 ++++++++++++++++++++++++++++++++++
 fs/xfs/scrub/rtrmap.c                |   54 ++
 fs/xfs/scrub/rtrmap_repair.c         |  103 ++++
 fs/xfs/scrub/scrub.c                 |    7 
 fs/xfs/scrub/scrub.h                 |   12 +
 fs/xfs/scrub/stats.c                 |    1 
 fs/xfs/scrub/trace.h                 |   54 +-
 fs/xfs/xfs_buf_item_recover.c        |    4 
 fs/xfs/xfs_fsmap.c                   |   25 +
 fs/xfs/xfs_fsops.c                   |    2 
 fs/xfs/xfs_health.c                  |    1 
 fs/xfs/xfs_inode_item.c              |   14 +
 fs/xfs/xfs_inode_item_recover.c      |    4 
 fs/xfs/xfs_ioctl.c                   |   21 +
 fs/xfs/xfs_log_recover.c             |    2 
 fs/xfs/xfs_mount.c                   |    5 
 fs/xfs/xfs_mount.h                   |    9 
 fs/xfs/xfs_refcount_item.c           |  240 ++++++++++
 fs/xfs/xfs_reflink.c                 |  325 ++++++++++----
 fs/xfs/xfs_reflink.h                 |    4 
 fs/xfs/xfs_rtalloc.c                 |   24 +
 fs/xfs/xfs_rtalloc.h                 |    5 
 fs/xfs/xfs_stats.c                   |    3 
 fs/xfs/xfs_stats.h                   |    1 
 fs/xfs/xfs_super.c                   |    6 
 fs/xfs/xfs_trace.h                   |  111 +++--
 74 files changed, 4330 insertions(+), 352 deletions(-)
 create mode 100644 fs/xfs/libxfs/xfs_rtrefcount_btree.c
 create mode 100644 fs/xfs/libxfs/xfs_rtrefcount_btree.h
 create mode 100644 fs/xfs/scrub/rgb_bitmap.h
 create mode 100644 fs/xfs/scrub/rtb_bitmap.h
 create mode 100644 fs/xfs/scrub/rtrefcount.c
 create mode 100644 fs/xfs/scrub/rtrefcount_repair.c

[PATCH 01/43] xfs: prepare refcount btree cursor tracepoints for realtime

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Rework the refcount btree cursor tracepoints in preparation to handle the
realtime refcount btree cursor.  Mostly this involves renaming the field to
"refcbno" and extracting the group number from the cursor when possible.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/xfs_refcount_item.c |    4 +-
 fs/xfs/xfs_trace.h         |  111 +++++++++++++++++++++++++++-----------------
 2 files changed, 70 insertions(+), 45 deletions(-)


diff --git a/fs/xfs/xfs_refcount_item.c b/fs/xfs/xfs_refcount_item.c
index bede1c96c33011..c807c4b90c44e3 100644
--- a/fs/xfs/xfs_refcount_item.c
+++ b/fs/xfs/xfs_refcount_item.c
@@ -328,9 +328,9 @@ xfs_refcount_defer_add(
 {
 	struct xfs_mount		*mp = tp->t_mountp;
 
-	trace_xfs_refcount_defer(mp, ri);
-
 	ri->ri_group = xfs_group_intent_get(mp, ri->ri_startblock, XG_TYPE_AG);
+
+	trace_xfs_refcount_defer(mp, ri);
 	xfs_defer_add(tp, &ri->ri_list, &xfs_refcount_update_defer_type);
 }
 
diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h
index 84cdc145e2d96a..4fe689410eb6ae 100644
--- a/fs/xfs/xfs_trace.h
+++ b/fs/xfs/xfs_trace.h
@@ -3305,56 +3305,62 @@ TRACE_EVENT(xfs_ag_resv_init_error,
 /* refcount tracepoint classes */
 
 DECLARE_EVENT_CLASS(xfs_refcount_class,
-	TP_PROTO(struct xfs_btree_cur *cur, xfs_agblock_t agbno,
+	TP_PROTO(struct xfs_btree_cur *cur, xfs_agblock_t gbno,
 		xfs_extlen_t len),
-	TP_ARGS(cur, agbno, len),
+	TP_ARGS(cur, gbno, len),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
+		__field(enum xfs_group_type, type)
 		__field(xfs_agnumber_t, agno)
-		__field(xfs_agblock_t, agbno)
+		__field(xfs_agblock_t, gbno)
 		__field(xfs_extlen_t, len)
 	),
 	TP_fast_assign(
 		__entry->dev = cur->bc_mp->m_super->s_dev;
+		__entry->type = cur->bc_group->xg_type;
 		__entry->agno = cur->bc_group->xg_gno;
-		__entry->agbno = agbno;
+		__entry->gbno = gbno;
 		__entry->len = len;
 	),
-	TP_printk("dev %d:%d agno 0x%x agbno 0x%x fsbcount 0x%x",
+	TP_printk("dev %d:%d %sno 0x%x gbno 0x%x fsbcount 0x%x",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __print_symbolic(__entry->type, XG_TYPE_STRINGS),
 		  __entry->agno,
-		  __entry->agbno,
+		  __entry->gbno,
 		  __entry->len)
 );
 #define DEFINE_REFCOUNT_EVENT(name) \
 DEFINE_EVENT(xfs_refcount_class, name, \
-	TP_PROTO(struct xfs_btree_cur *cur, xfs_agblock_t agbno, \
+	TP_PROTO(struct xfs_btree_cur *cur, xfs_agblock_t gbno, \
 		xfs_extlen_t len), \
-	TP_ARGS(cur, agbno, len))
+	TP_ARGS(cur, gbno, len))
 
 TRACE_DEFINE_ENUM(XFS_LOOKUP_EQi);
 TRACE_DEFINE_ENUM(XFS_LOOKUP_LEi);
 TRACE_DEFINE_ENUM(XFS_LOOKUP_GEi);
 TRACE_EVENT(xfs_refcount_lookup,
-	TP_PROTO(struct xfs_btree_cur *cur, xfs_agblock_t agbno,
+	TP_PROTO(struct xfs_btree_cur *cur, xfs_agblock_t gbno,
 		xfs_lookup_t dir),
-	TP_ARGS(cur, agbno, dir),
+	TP_ARGS(cur, gbno, dir),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
+		__field(enum xfs_group_type, type)
 		__field(xfs_agnumber_t, agno)
-		__field(xfs_agblock_t, agbno)
+		__field(xfs_agblock_t, gbno)
 		__field(xfs_lookup_t, dir)
 	),
 	TP_fast_assign(
 		__entry->dev = cur->bc_mp->m_super->s_dev;
+		__entry->type = cur->bc_group->xg_type;
 		__entry->agno = cur->bc_group->xg_gno;
-		__entry->agbno = agbno;
+		__entry->gbno = gbno;
 		__entry->dir = dir;
 	),
-	TP_printk("dev %d:%d agno 0x%x agbno 0x%x cmp %s(%d)",
+	TP_printk("dev %d:%d %sno 0x%x gbno 0x%x cmp %s(%d)",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __print_symbolic(__entry->type, XG_TYPE_STRINGS),
 		  __entry->agno,
-		  __entry->agbno,
+		  __entry->gbno,
 		  __print_symbolic(__entry->dir, XFS_AG_BTREE_CMP_FORMAT_STR),
 		  __entry->dir)
 )
@@ -3365,6 +3371,7 @@ DECLARE_EVENT_CLASS(xfs_refcount_extent_class,
 	TP_ARGS(cur, irec),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
+		__field(enum xfs_group_type, type)
 		__field(xfs_agnumber_t, agno)
 		__field(enum xfs_refc_domain, domain)
 		__field(xfs_agblock_t, startblock)
@@ -3373,14 +3380,16 @@ DECLARE_EVENT_CLASS(xfs_refcount_extent_class,
 	),
 	TP_fast_assign(
 		__entry->dev = cur->bc_mp->m_super->s_dev;
+		__entry->type = cur->bc_group->xg_type;
 		__entry->agno = cur->bc_group->xg_gno;
 		__entry->domain = irec->rc_domain;
 		__entry->startblock = irec->rc_startblock;
 		__entry->blockcount = irec->rc_blockcount;
 		__entry->refcount = irec->rc_refcount;
 	),
-	TP_printk("dev %d:%d agno 0x%x dom %s agbno 0x%x fsbcount 0x%x refcount %u",
+	TP_printk("dev %d:%d %sno 0x%x dom %s gbno 0x%x fsbcount 0x%x refcount %u",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __print_symbolic(__entry->type, XG_TYPE_STRINGS),
 		  __entry->agno,
 		  __print_symbolic(__entry->domain, XFS_REFC_DOMAIN_STRINGS),
 		  __entry->startblock,
@@ -3396,49 +3405,53 @@ DEFINE_EVENT(xfs_refcount_extent_class, name, \
 /* single-rcext and an agbno tracepoint class */
 DECLARE_EVENT_CLASS(xfs_refcount_extent_at_class,
 	TP_PROTO(struct xfs_btree_cur *cur, struct xfs_refcount_irec *irec,
-		 xfs_agblock_t agbno),
-	TP_ARGS(cur, irec, agbno),
+		 xfs_agblock_t gbno),
+	TP_ARGS(cur, irec, gbno),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
+		__field(enum xfs_group_type, type)
 		__field(xfs_agnumber_t, agno)
 		__field(enum xfs_refc_domain, domain)
 		__field(xfs_agblock_t, startblock)
 		__field(xfs_extlen_t, blockcount)
 		__field(xfs_nlink_t, refcount)
-		__field(xfs_agblock_t, agbno)
+		__field(xfs_agblock_t, gbno)
 	),
 	TP_fast_assign(
 		__entry->dev = cur->bc_mp->m_super->s_dev;
+		__entry->type = cur->bc_group->xg_type;
 		__entry->agno = cur->bc_group->xg_gno;
 		__entry->domain = irec->rc_domain;
 		__entry->startblock = irec->rc_startblock;
 		__entry->blockcount = irec->rc_blockcount;
 		__entry->refcount = irec->rc_refcount;
-		__entry->agbno = agbno;
+		__entry->gbno = gbno;
 	),
-	TP_printk("dev %d:%d agno 0x%x dom %s agbno 0x%x fsbcount 0x%x refcount %u @ agbno 0x%x",
+	TP_printk("dev %d:%d %sno 0x%x dom %s gbno 0x%x fsbcount 0x%x refcount %u @ gbno 0x%x",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __print_symbolic(__entry->type, XG_TYPE_STRINGS),
 		  __entry->agno,
 		  __print_symbolic(__entry->domain, XFS_REFC_DOMAIN_STRINGS),
 		  __entry->startblock,
 		  __entry->blockcount,
 		  __entry->refcount,
-		  __entry->agbno)
+		  __entry->gbno)
 )
 
 #define DEFINE_REFCOUNT_EXTENT_AT_EVENT(name) \
 DEFINE_EVENT(xfs_refcount_extent_at_class, name, \
 	TP_PROTO(struct xfs_btree_cur *cur, struct xfs_refcount_irec *irec, \
-		 xfs_agblock_t agbno), \
-	TP_ARGS(cur, irec, agbno))
+		 xfs_agblock_t gbno), \
+	TP_ARGS(cur, irec, gbno))
 
 /* double-rcext tracepoint class */
 DECLARE_EVENT_CLASS(xfs_refcount_double_extent_class,
 	TP_PROTO(struct xfs_btree_cur *cur, struct xfs_refcount_irec *i1,
-		struct xfs_refcount_irec *i2),
+		 struct xfs_refcount_irec *i2),
 	TP_ARGS(cur, i1, i2),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
+		__field(enum xfs_group_type, type)
 		__field(xfs_agnumber_t, agno)
 		__field(enum xfs_refc_domain, i1_domain)
 		__field(xfs_agblock_t, i1_startblock)
@@ -3451,6 +3464,7 @@ DECLARE_EVENT_CLASS(xfs_refcount_double_extent_class,
 	),
 	TP_fast_assign(
 		__entry->dev = cur->bc_mp->m_super->s_dev;
+		__entry->type = cur->bc_group->xg_type;
 		__entry->agno = cur->bc_group->xg_gno;
 		__entry->i1_domain = i1->rc_domain;
 		__entry->i1_startblock = i1->rc_startblock;
@@ -3461,9 +3475,10 @@ DECLARE_EVENT_CLASS(xfs_refcount_double_extent_class,
 		__entry->i2_blockcount = i2->rc_blockcount;
 		__entry->i2_refcount = i2->rc_refcount;
 	),
-	TP_printk("dev %d:%d agno 0x%x dom %s agbno 0x%x fsbcount 0x%x refcount %u -- "
-		  "dom %s agbno 0x%x fsbcount 0x%x refcount %u",
+	TP_printk("dev %d:%d %sno 0x%x dom %s gbno 0x%x fsbcount 0x%x refcount %u -- "
+		  "dom %s gbno 0x%x fsbcount 0x%x refcount %u",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __print_symbolic(__entry->type, XG_TYPE_STRINGS),
 		  __entry->agno,
 		  __print_symbolic(__entry->i1_domain, XFS_REFC_DOMAIN_STRINGS),
 		  __entry->i1_startblock,
@@ -3484,10 +3499,11 @@ DEFINE_EVENT(xfs_refcount_double_extent_class, name, \
 /* double-rcext and an agbno tracepoint class */
 DECLARE_EVENT_CLASS(xfs_refcount_double_extent_at_class,
 	TP_PROTO(struct xfs_btree_cur *cur, struct xfs_refcount_irec *i1,
-		 struct xfs_refcount_irec *i2, xfs_agblock_t agbno),
-	TP_ARGS(cur, i1, i2, agbno),
+		 struct xfs_refcount_irec *i2, xfs_agblock_t gbno),
+	TP_ARGS(cur, i1, i2, gbno),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
+		__field(enum xfs_group_type, type)
 		__field(xfs_agnumber_t, agno)
 		__field(enum xfs_refc_domain, i1_domain)
 		__field(xfs_agblock_t, i1_startblock)
@@ -3497,10 +3513,11 @@ DECLARE_EVENT_CLASS(xfs_refcount_double_extent_at_class,
 		__field(xfs_agblock_t, i2_startblock)
 		__field(xfs_extlen_t, i2_blockcount)
 		__field(xfs_nlink_t, i2_refcount)
-		__field(xfs_agblock_t, agbno)
+		__field(xfs_agblock_t, gbno)
 	),
 	TP_fast_assign(
 		__entry->dev = cur->bc_mp->m_super->s_dev;
+		__entry->type = cur->bc_group->xg_type;
 		__entry->agno = cur->bc_group->xg_gno;
 		__entry->i1_domain = i1->rc_domain;
 		__entry->i1_startblock = i1->rc_startblock;
@@ -3510,11 +3527,12 @@ DECLARE_EVENT_CLASS(xfs_refcount_double_extent_at_class,
 		__entry->i2_startblock = i2->rc_startblock;
 		__entry->i2_blockcount = i2->rc_blockcount;
 		__entry->i2_refcount = i2->rc_refcount;
-		__entry->agbno = agbno;
+		__entry->gbno = gbno;
 	),
-	TP_printk("dev %d:%d agno 0x%x dom %s agbno 0x%x fsbcount 0x%x refcount %u -- "
-		  "dom %s agbno 0x%x fsbcount 0x%x refcount %u @ agbno 0x%x",
+	TP_printk("dev %d:%d %sno 0x%x dom %s gbno 0x%x fsbcount 0x%x refcount %u -- "
+		  "dom %s gbno 0x%x fsbcount 0x%x refcount %u @ gbno 0x%x",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __print_symbolic(__entry->type, XG_TYPE_STRINGS),
 		  __entry->agno,
 		  __print_symbolic(__entry->i1_domain, XFS_REFC_DOMAIN_STRINGS),
 		  __entry->i1_startblock,
@@ -3524,14 +3542,14 @@ DECLARE_EVENT_CLASS(xfs_refcount_double_extent_at_class,
 		  __entry->i2_startblock,
 		  __entry->i2_blockcount,
 		  __entry->i2_refcount,
-		  __entry->agbno)
+		  __entry->gbno)
 )
 
 #define DEFINE_REFCOUNT_DOUBLE_EXTENT_AT_EVENT(name) \
 DEFINE_EVENT(xfs_refcount_double_extent_at_class, name, \
 	TP_PROTO(struct xfs_btree_cur *cur, struct xfs_refcount_irec *i1, \
-		struct xfs_refcount_irec *i2, xfs_agblock_t agbno), \
-	TP_ARGS(cur, i1, i2, agbno))
+		struct xfs_refcount_irec *i2, xfs_agblock_t gbno), \
+	TP_ARGS(cur, i1, i2, gbno))
 
 /* triple-rcext tracepoint class */
 DECLARE_EVENT_CLASS(xfs_refcount_triple_extent_class,
@@ -3540,6 +3558,7 @@ DECLARE_EVENT_CLASS(xfs_refcount_triple_extent_class,
 	TP_ARGS(cur, i1, i2, i3),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
+		__field(enum xfs_group_type, type)
 		__field(xfs_agnumber_t, agno)
 		__field(enum xfs_refc_domain, i1_domain)
 		__field(xfs_agblock_t, i1_startblock)
@@ -3556,6 +3575,7 @@ DECLARE_EVENT_CLASS(xfs_refcount_triple_extent_class,
 	),
 	TP_fast_assign(
 		__entry->dev = cur->bc_mp->m_super->s_dev;
+		__entry->type = cur->bc_group->xg_type;
 		__entry->agno = cur->bc_group->xg_gno;
 		__entry->i1_domain = i1->rc_domain;
 		__entry->i1_startblock = i1->rc_startblock;
@@ -3570,10 +3590,11 @@ DECLARE_EVENT_CLASS(xfs_refcount_triple_extent_class,
 		__entry->i3_blockcount = i3->rc_blockcount;
 		__entry->i3_refcount = i3->rc_refcount;
 	),
-	TP_printk("dev %d:%d agno 0x%x dom %s agbno 0x%x fsbcount 0x%x refcount %u -- "
-		  "dom %s agbno 0x%x fsbcount 0x%x refcount %u -- "
-		  "dom %s agbno 0x%x fsbcount 0x%x refcount %u",
+	TP_printk("dev %d:%d %sno 0x%x dom %s gbno 0x%x fsbcount 0x%x refcount %u -- "
+		  "dom %s gbno 0x%x fsbcount 0x%x refcount %u -- "
+		  "dom %s gbno 0x%x fsbcount 0x%x refcount %u",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __print_symbolic(__entry->type, XG_TYPE_STRINGS),
 		  __entry->agno,
 		  __print_symbolic(__entry->i1_domain, XFS_REFC_DOMAIN_STRINGS),
 		  __entry->i1_startblock,
@@ -3641,23 +3662,27 @@ DECLARE_EVENT_CLASS(xfs_refcount_deferred_class,
 	TP_ARGS(mp, refc),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
+		__field(enum xfs_group_type, type)
 		__field(xfs_agnumber_t, agno)
 		__field(int, op)
-		__field(xfs_agblock_t, agbno)
+		__field(xfs_agblock_t, gbno)
 		__field(xfs_extlen_t, len)
 	),
 	TP_fast_assign(
 		__entry->dev = mp->m_super->s_dev;
-		__entry->agno = XFS_FSB_TO_AGNO(mp, refc->ri_startblock);
+		__entry->type = refc->ri_group->xg_type;
+		__entry->agno = refc->ri_group->xg_gno;
 		__entry->op = refc->ri_type;
-		__entry->agbno = XFS_FSB_TO_AGBNO(mp, refc->ri_startblock);
+		__entry->gbno = xfs_fsb_to_gbno(mp, refc->ri_startblock,
+						   refc->ri_group->xg_type);
 		__entry->len = refc->ri_blockcount;
 	),
-	TP_printk("dev %d:%d op %s agno 0x%x agbno 0x%x fsbcount 0x%x",
+	TP_printk("dev %d:%d op %s %sno 0x%x gbno 0x%x fsbcount 0x%x",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		  __print_symbolic(__entry->op, XFS_REFCOUNT_INTENT_STRINGS),
+		  __print_symbolic(__entry->type, XG_TYPE_STRINGS),
 		  __entry->agno,
-		  __entry->agbno,
+		  __entry->gbno,
 		  __entry->len)
 );
 #define DEFINE_REFCOUNT_DEFERRED_EVENT(name) \

[PATCH 02/43] xfs: namespace the maximum length/refcount symbols

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Actually namespace these variables properly, so that readers can tell
that this is an XFS symbol, and that it's for the refcount
functionality.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_format.h     |    4 ++--
 fs/xfs/libxfs/xfs_refcount.c   |   18 +++++++++---------
 fs/xfs/scrub/refcount.c        |    2 +-
 fs/xfs/scrub/refcount_repair.c |    4 ++--
 4 files changed, 14 insertions(+), 14 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index fba4e59aded4a0..16696bc3ff9445 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -1790,8 +1790,8 @@ struct xfs_refcount_key {
 	__be32		rc_startblock;	/* starting block number */
 };
 
-#define MAXREFCOUNT	((xfs_nlink_t)~0U)
-#define MAXREFCEXTLEN	((xfs_extlen_t)~0U)
+#define XFS_REFC_REFCOUNT_MAX	((xfs_nlink_t)~0U)
+#define XFS_REFC_LEN_MAX	((xfs_extlen_t)~0U)
 
 /* btree pointer type */
 typedef __be32 xfs_refcount_ptr_t;
diff --git a/fs/xfs/libxfs/xfs_refcount.c b/fs/xfs/libxfs/xfs_refcount.c
index bbb86dc9a25c7f..faace12fe2e383 100644
--- a/fs/xfs/libxfs/xfs_refcount.c
+++ b/fs/xfs/libxfs/xfs_refcount.c
@@ -128,7 +128,7 @@ xfs_refcount_check_irec(
 	struct xfs_perag		*pag,
 	const struct xfs_refcount_irec	*irec)
 {
-	if (irec->rc_blockcount == 0 || irec->rc_blockcount > MAXREFCEXTLEN)
+	if (irec->rc_blockcount == 0 || irec->rc_blockcount > XFS_REFC_LEN_MAX)
 		return __this_address;
 
 	if (!xfs_refcount_check_domain(irec))
@@ -138,7 +138,7 @@ xfs_refcount_check_irec(
 	if (!xfs_verify_agbext(pag, irec->rc_startblock, irec->rc_blockcount))
 		return __this_address;
 
-	if (irec->rc_refcount == 0 || irec->rc_refcount > MAXREFCOUNT)
+	if (irec->rc_refcount == 0 || irec->rc_refcount > XFS_REFC_REFCOUNT_MAX)
 		return __this_address;
 
 	return NULL;
@@ -853,9 +853,9 @@ xfs_refc_merge_refcount(
 	const struct xfs_refcount_irec	*irec,
 	enum xfs_refc_adjust_op		adjust)
 {
-	/* Once a record hits MAXREFCOUNT, it is pinned there forever */
-	if (irec->rc_refcount == MAXREFCOUNT)
-		return MAXREFCOUNT;
+	/* Once a record hits XFS_REFC_REFCOUNT_MAX, it is pinned forever */
+	if (irec->rc_refcount == XFS_REFC_REFCOUNT_MAX)
+		return XFS_REFC_REFCOUNT_MAX;
 	return irec->rc_refcount + adjust;
 }
 
@@ -898,7 +898,7 @@ xfs_refc_want_merge_center(
 	 * hence we need to catch u32 addition overflows here.
 	 */
 	ulen += cleft->rc_blockcount + right->rc_blockcount;
-	if (ulen >= MAXREFCEXTLEN)
+	if (ulen >= XFS_REFC_LEN_MAX)
 		return false;
 
 	*ulenp = ulen;
@@ -933,7 +933,7 @@ xfs_refc_want_merge_left(
 	 * hence we need to catch u32 addition overflows here.
 	 */
 	ulen += cleft->rc_blockcount;
-	if (ulen >= MAXREFCEXTLEN)
+	if (ulen >= XFS_REFC_LEN_MAX)
 		return false;
 
 	return true;
@@ -967,7 +967,7 @@ xfs_refc_want_merge_right(
 	 * hence we need to catch u32 addition overflows here.
 	 */
 	ulen += cright->rc_blockcount;
-	if (ulen >= MAXREFCEXTLEN)
+	if (ulen >= XFS_REFC_LEN_MAX)
 		return false;
 
 	return true;
@@ -1196,7 +1196,7 @@ xfs_refcount_adjust_extents(
 		 * Adjust the reference count and either update the tree
 		 * (incr) or free the blocks (decr).
 		 */
-		if (ext.rc_refcount == MAXREFCOUNT)
+		if (ext.rc_refcount == XFS_REFC_REFCOUNT_MAX)
 			goto skip;
 		ext.rc_refcount += adj;
 		trace_xfs_refcount_modify_extent(cur, &ext);
diff --git a/fs/xfs/scrub/refcount.c b/fs/xfs/scrub/refcount.c
index 1c5e45cc64190c..d465280230154f 100644
--- a/fs/xfs/scrub/refcount.c
+++ b/fs/xfs/scrub/refcount.c
@@ -421,7 +421,7 @@ xchk_refcount_mergeable(
 	if (r1->rc_refcount != r2->rc_refcount)
 		return false;
 	if ((unsigned long long)r1->rc_blockcount + r2->rc_blockcount >
-			MAXREFCEXTLEN)
+			XFS_REFC_LEN_MAX)
 		return false;
 
 	return true;
diff --git a/fs/xfs/scrub/refcount_repair.c b/fs/xfs/scrub/refcount_repair.c
index 4e572b81c98669..1ee6d4aeb308f5 100644
--- a/fs/xfs/scrub/refcount_repair.c
+++ b/fs/xfs/scrub/refcount_repair.c
@@ -183,7 +183,7 @@ xrep_refc_stash(
 	if (xchk_should_terminate(sc, &error))
 		return error;
 
-	irec.rc_refcount = min_t(uint64_t, MAXREFCOUNT, refcount);
+	irec.rc_refcount = min_t(uint64_t, XFS_REFC_REFCOUNT_MAX, refcount);
 
 	error = xrep_refc_check_ext(rr->sc, &irec);
 	if (error)
@@ -422,7 +422,7 @@ xrep_refc_find_refcounts(
 	/*
 	 * Set up a bag to store all the rmap records that we're tracking to
 	 * generate a reference count record.  If the size of the bag exceeds
-	 * MAXREFCOUNT, we clamp rc_refcount.
+	 * XFS_REFC_REFCOUNT_MAX, we clamp rc_refcount.
 	 */
 	error = rcbag_init(sc->mp, sc->xmbtp, &rcstack);
 	if (error)

[PATCH 03/43] xfs: introduce realtime refcount btree ondisk definitions

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Add the ondisk structure definitions for realtime refcount btrees. The
realtime refcount btree will be rooted from a hidden inode so it needs
to have a separate btree block magic and pointer format.

Next, add everything needed to read, write and manipulate refcount btree
blocks. This prepares the way for connecting the btree operations
implementation, though the changes to actually root the rtrefcount btree
in an inode come later.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/Makefile                      |    1 
 fs/xfs/libxfs/xfs_btree.c            |    5 +
 fs/xfs/libxfs/xfs_format.h           |    9 +
 fs/xfs/libxfs/xfs_ondisk.h           |    1 
 fs/xfs/libxfs/xfs_rtrefcount_btree.c |  273 ++++++++++++++++++++++++++++++++++
 fs/xfs/libxfs/xfs_rtrefcount_btree.h |   70 +++++++++
 fs/xfs/libxfs/xfs_sb.c               |    8 +
 fs/xfs/libxfs/xfs_shared.h           |    7 +
 fs/xfs/xfs_mount.c                   |    5 -
 fs/xfs/xfs_mount.h                   |    9 +
 fs/xfs/xfs_stats.c                   |    3 
 fs/xfs/xfs_stats.h                   |    1 
 12 files changed, 390 insertions(+), 2 deletions(-)
 create mode 100644 fs/xfs/libxfs/xfs_rtrefcount_btree.c
 create mode 100644 fs/xfs/libxfs/xfs_rtrefcount_btree.h


diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile
index 338e10f81b7b71..b9356d01416e16 100644
--- a/fs/xfs/Makefile
+++ b/fs/xfs/Makefile
@@ -51,6 +51,7 @@ xfs-y				+= $(addprefix libxfs/, \
 				   xfs_rmap_btree.o \
 				   xfs_refcount.o \
 				   xfs_refcount_btree.o \
+				   xfs_rtrefcount_btree.o \
 				   xfs_rtrmap_btree.o \
 				   xfs_sb.o \
 				   xfs_symlink_remote.o \
diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c
index 36ab06f8a3bc99..299ce7fd11b0a9 100644
--- a/fs/xfs/libxfs/xfs_btree.c
+++ b/fs/xfs/libxfs/xfs_btree.c
@@ -35,6 +35,7 @@
 #include "xfs_rmap.h"
 #include "xfs_quota.h"
 #include "xfs_metafile.h"
+#include "xfs_rtrefcount_btree.h"
 
 /*
  * Btree magic numbers.
@@ -5533,6 +5534,9 @@ xfs_btree_init_cur_caches(void)
 	if (error)
 		goto err;
 	error = xfs_rtrmapbt_init_cur_cache();
+	if (error)
+		goto err;
+	error = xfs_rtrefcountbt_init_cur_cache();
 	if (error)
 		goto err;
 
@@ -5552,6 +5556,7 @@ xfs_btree_destroy_cur_caches(void)
 	xfs_rmapbt_destroy_cur_cache();
 	xfs_refcountbt_destroy_cur_cache();
 	xfs_rtrmapbt_destroy_cur_cache();
+	xfs_rtrefcountbt_destroy_cur_cache();
 }
 
 /* Move the btree cursor before the first record. */
diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index 16696bc3ff9445..17f7c0d1aaa452 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -1796,6 +1796,15 @@ struct xfs_refcount_key {
 /* btree pointer type */
 typedef __be32 xfs_refcount_ptr_t;
 
+/*
+ * Realtime Reference Count btree format definitions
+ *
+ * This is a btree for reference count records for realtime volumes
+ */
+#define	XFS_RTREFC_CRC_MAGIC	0x52434e54	/* 'RCNT' */
+
+/* inode-rooted btree pointer type */
+typedef __be64 xfs_rtrefcount_ptr_t;
 
 /*
  * BMAP Btree format definitions
diff --git a/fs/xfs/libxfs/xfs_ondisk.h b/fs/xfs/libxfs/xfs_ondisk.h
index 07e2f5fb3a94ae..efb035050c009c 100644
--- a/fs/xfs/libxfs/xfs_ondisk.h
+++ b/fs/xfs/libxfs/xfs_ondisk.h
@@ -85,6 +85,7 @@ xfs_check_ondisk_structs(void)
 	XFS_CHECK_STRUCT_SIZE(struct xfs_rtbuf_blkinfo,		48);
 	XFS_CHECK_STRUCT_SIZE(xfs_rtrmap_ptr_t,			8);
 	XFS_CHECK_STRUCT_SIZE(struct xfs_rtrmap_root,		4);
+	XFS_CHECK_STRUCT_SIZE(xfs_rtrefcount_ptr_t,		8);
 
 	/*
 	 * m68k has problems with struct xfs_attr_leaf_name_remote, but we pad
diff --git a/fs/xfs/libxfs/xfs_rtrefcount_btree.c b/fs/xfs/libxfs/xfs_rtrefcount_btree.c
new file mode 100644
index 00000000000000..d07d3b1e78f730
--- /dev/null
+++ b/fs/xfs/libxfs/xfs_rtrefcount_btree.c
@@ -0,0 +1,273 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2021-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#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_btree_staging.h"
+#include "xfs_rtrefcount_btree.h"
+#include "xfs_trace.h"
+#include "xfs_cksum.h"
+#include "xfs_error.h"
+#include "xfs_extent_busy.h"
+#include "xfs_rtgroup.h"
+#include "xfs_rtbitmap.h"
+
+static struct kmem_cache	*xfs_rtrefcountbt_cur_cache;
+
+/*
+ * Realtime Reference Count btree.
+ *
+ * This is a btree used to track the owner(s) of a given extent in the realtime
+ * device.  See the comments in xfs_refcount_btree.c for more information.
+ *
+ * This tree is basically the same as the regular refcount btree except that
+ * it's rooted in an inode.
+ */
+
+static struct xfs_btree_cur *
+xfs_rtrefcountbt_dup_cursor(
+	struct xfs_btree_cur	*cur)
+{
+	return xfs_rtrefcountbt_init_cursor(cur->bc_tp, to_rtg(cur->bc_group));
+}
+
+static xfs_failaddr_t
+xfs_rtrefcountbt_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 (!xfs_verify_magic(bp, block->bb_magic))
+		return __this_address;
+
+	if (!xfs_has_reflink(mp))
+		return __this_address;
+	fa = xfs_btree_fsblock_v5hdr_verify(bp, XFS_RMAP_OWN_UNKNOWN);
+	if (fa)
+		return fa;
+	level = be16_to_cpu(block->bb_level);
+	if (level > mp->m_rtrefc_maxlevels)
+		return __this_address;
+
+	return xfs_btree_fsblock_verify(bp, mp->m_rtrefc_mxr[level != 0]);
+}
+
+static void
+xfs_rtrefcountbt_read_verify(
+	struct xfs_buf	*bp)
+{
+	xfs_failaddr_t	fa;
+
+	if (!xfs_btree_fsblock_verify_crc(bp))
+		xfs_verifier_error(bp, -EFSBADCRC, __this_address);
+	else {
+		fa = xfs_rtrefcountbt_verify(bp);
+		if (fa)
+			xfs_verifier_error(bp, -EFSCORRUPTED, fa);
+	}
+
+	if (bp->b_error)
+		trace_xfs_btree_corrupt(bp, _RET_IP_);
+}
+
+static void
+xfs_rtrefcountbt_write_verify(
+	struct xfs_buf	*bp)
+{
+	xfs_failaddr_t	fa;
+
+	fa = xfs_rtrefcountbt_verify(bp);
+	if (fa) {
+		trace_xfs_btree_corrupt(bp, _RET_IP_);
+		xfs_verifier_error(bp, -EFSCORRUPTED, fa);
+		return;
+	}
+	xfs_btree_fsblock_calc_crc(bp);
+
+}
+
+const struct xfs_buf_ops xfs_rtrefcountbt_buf_ops = {
+	.name			= "xfs_rtrefcountbt",
+	.magic			= { 0, cpu_to_be32(XFS_RTREFC_CRC_MAGIC) },
+	.verify_read		= xfs_rtrefcountbt_read_verify,
+	.verify_write		= xfs_rtrefcountbt_write_verify,
+	.verify_struct		= xfs_rtrefcountbt_verify,
+};
+
+const struct xfs_btree_ops xfs_rtrefcountbt_ops = {
+	.name			= "rtrefcount",
+	.type			= XFS_BTREE_TYPE_INODE,
+	.geom_flags		= XFS_BTGEO_IROOT_RECORDS,
+
+	.rec_len		= sizeof(struct xfs_refcount_rec),
+	.key_len		= sizeof(struct xfs_refcount_key),
+	.ptr_len		= XFS_BTREE_LONG_PTR_LEN,
+
+	.lru_refs		= XFS_REFC_BTREE_REF,
+	.statoff		= XFS_STATS_CALC_INDEX(xs_rtrefcbt_2),
+
+	.dup_cursor		= xfs_rtrefcountbt_dup_cursor,
+	.buf_ops		= &xfs_rtrefcountbt_buf_ops,
+};
+
+/* Allocate a new rt refcount btree cursor. */
+struct xfs_btree_cur *
+xfs_rtrefcountbt_init_cursor(
+	struct xfs_trans	*tp,
+	struct xfs_rtgroup	*rtg)
+{
+	struct xfs_inode	*ip = NULL;
+	struct xfs_mount	*mp = rtg_mount(rtg);
+	struct xfs_btree_cur	*cur;
+
+	return NULL; /* XXX */
+
+	xfs_assert_ilocked(ip, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL);
+
+	cur = xfs_btree_alloc_cursor(mp, tp, &xfs_rtrefcountbt_ops,
+			mp->m_rtrefc_maxlevels, xfs_rtrefcountbt_cur_cache);
+
+	cur->bc_ino.ip = ip;
+	cur->bc_refc.nr_ops = 0;
+	cur->bc_refc.shape_changes = 0;
+	cur->bc_group = xfs_group_hold(rtg_group(rtg));
+	cur->bc_nlevels = be16_to_cpu(ip->i_df.if_broot->bb_level) + 1;
+	cur->bc_ino.forksize = xfs_inode_fork_size(ip, XFS_DATA_FORK);
+	cur->bc_ino.whichfork = XFS_DATA_FORK;
+	return cur;
+}
+
+/*
+ * Install a new rt reverse mapping btree root.  Caller is responsible for
+ * invalidating and freeing the old btree blocks.
+ */
+void
+xfs_rtrefcountbt_commit_staged_btree(
+	struct xfs_btree_cur	*cur,
+	struct xfs_trans	*tp)
+{
+	struct xbtree_ifakeroot	*ifake = cur->bc_ino.ifake;
+	struct xfs_ifork	*ifp;
+	int			flags = XFS_ILOG_CORE | XFS_ILOG_DBROOT;
+
+	ASSERT(cur->bc_flags & XFS_BTREE_STAGING);
+
+	/*
+	 * Free any resources hanging off the real fork, then shallow-copy the
+	 * staging fork's contents into the real fork to transfer everything
+	 * we just built.
+	 */
+	ifp = xfs_ifork_ptr(cur->bc_ino.ip, XFS_DATA_FORK);
+	xfs_idestroy_fork(ifp);
+	memcpy(ifp, ifake->if_fork, sizeof(struct xfs_ifork));
+
+	cur->bc_ino.ip->i_projid = cur->bc_group->xg_gno;
+	xfs_trans_log_inode(tp, cur->bc_ino.ip, flags);
+	xfs_btree_commit_ifakeroot(cur, tp, XFS_DATA_FORK);
+}
+
+/* Calculate number of records in a realtime refcount btree block. */
+static inline unsigned int
+xfs_rtrefcountbt_block_maxrecs(
+	unsigned int		blocklen,
+	bool			leaf)
+{
+
+	if (leaf)
+		return blocklen / sizeof(struct xfs_refcount_rec);
+	return blocklen / (sizeof(struct xfs_refcount_key) +
+			   sizeof(xfs_rtrefcount_ptr_t));
+}
+
+/*
+ * Calculate number of records in an refcount btree block.
+ */
+unsigned int
+xfs_rtrefcountbt_maxrecs(
+	struct xfs_mount	*mp,
+	unsigned int		blocklen,
+	bool			leaf)
+{
+	blocklen -= XFS_RTREFCOUNT_BLOCK_LEN;
+	return xfs_rtrefcountbt_block_maxrecs(blocklen, leaf);
+}
+
+/* Compute the max possible height for realtime refcount btrees. */
+unsigned int
+xfs_rtrefcountbt_maxlevels_ondisk(void)
+{
+	unsigned int		minrecs[2];
+	unsigned int		blocklen;
+
+	blocklen = XFS_MIN_CRC_BLOCKSIZE - XFS_BTREE_LBLOCK_CRC_LEN;
+
+	minrecs[0] = xfs_rtrefcountbt_block_maxrecs(blocklen, true) / 2;
+	minrecs[1] = xfs_rtrefcountbt_block_maxrecs(blocklen, false) / 2;
+
+	/* We need at most one record for every block in an rt group. */
+	return xfs_btree_compute_maxlevels(minrecs, XFS_MAX_RGBLOCKS);
+}
+
+int __init
+xfs_rtrefcountbt_init_cur_cache(void)
+{
+	xfs_rtrefcountbt_cur_cache = kmem_cache_create("xfs_rtrefcountbt_cur",
+			xfs_btree_cur_sizeof(
+					xfs_rtrefcountbt_maxlevels_ondisk()),
+			0, 0, NULL);
+
+	if (!xfs_rtrefcountbt_cur_cache)
+		return -ENOMEM;
+	return 0;
+}
+
+void
+xfs_rtrefcountbt_destroy_cur_cache(void)
+{
+	kmem_cache_destroy(xfs_rtrefcountbt_cur_cache);
+	xfs_rtrefcountbt_cur_cache = NULL;
+}
+
+/* Compute the maximum height of a realtime refcount btree. */
+void
+xfs_rtrefcountbt_compute_maxlevels(
+	struct xfs_mount	*mp)
+{
+	unsigned int		d_maxlevels, r_maxlevels;
+
+	if (!xfs_has_rtreflink(mp)) {
+		mp->m_rtrefc_maxlevels = 0;
+		return;
+	}
+
+	/*
+	 * The realtime refcountbt lives on the data device, which means that
+	 * its maximum height is constrained by the size of the data device and
+	 * the height required to store one refcount record for each rtextent
+	 * in an rt group.
+	 */
+	d_maxlevels = xfs_btree_space_to_height(mp->m_rtrefc_mnr,
+				mp->m_sb.sb_dblocks);
+	r_maxlevels = xfs_btree_compute_maxlevels(mp->m_rtrefc_mnr,
+				mp->m_sb.sb_rgextents);
+
+	/* Add one level to handle the inode root level. */
+	mp->m_rtrefc_maxlevels = min(d_maxlevels, r_maxlevels) + 1;
+}
diff --git a/fs/xfs/libxfs/xfs_rtrefcount_btree.h b/fs/xfs/libxfs/xfs_rtrefcount_btree.h
new file mode 100644
index 00000000000000..b713b33818800c
--- /dev/null
+++ b/fs/xfs/libxfs/xfs_rtrefcount_btree.h
@@ -0,0 +1,70 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright (c) 2021-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_RTREFCOUNT_BTREE_H__
+#define __XFS_RTREFCOUNT_BTREE_H__
+
+struct xfs_buf;
+struct xfs_btree_cur;
+struct xfs_mount;
+struct xbtree_ifakeroot;
+struct xfs_rtgroup;
+
+/* refcounts only exist on crc enabled filesystems */
+#define XFS_RTREFCOUNT_BLOCK_LEN	XFS_BTREE_LBLOCK_CRC_LEN
+
+struct xfs_btree_cur *xfs_rtrefcountbt_init_cursor(struct xfs_trans *tp,
+		struct xfs_rtgroup *rtg);
+struct xfs_btree_cur *xfs_rtrefcountbt_stage_cursor(struct xfs_mount *mp,
+		struct xfs_rtgroup *rtg, struct xfs_inode *ip,
+		struct xbtree_ifakeroot *ifake);
+void xfs_rtrefcountbt_commit_staged_btree(struct xfs_btree_cur *cur,
+		struct xfs_trans *tp);
+unsigned int xfs_rtrefcountbt_maxrecs(struct xfs_mount *mp,
+		unsigned int blocklen, bool leaf);
+void xfs_rtrefcountbt_compute_maxlevels(struct xfs_mount *mp);
+
+/*
+ * Addresses of records, keys, and pointers within an incore rtrefcountbt block.
+ *
+ * (note that some of these may appear unused, but they are used in userspace)
+ */
+static inline struct xfs_refcount_rec *
+xfs_rtrefcount_rec_addr(
+	struct xfs_btree_block	*block,
+	unsigned int		index)
+{
+	return (struct xfs_refcount_rec *)
+		((char *)block + XFS_RTREFCOUNT_BLOCK_LEN +
+		 (index - 1) * sizeof(struct xfs_refcount_rec));
+}
+
+static inline struct xfs_refcount_key *
+xfs_rtrefcount_key_addr(
+	struct xfs_btree_block	*block,
+	unsigned int		index)
+{
+	return (struct xfs_refcount_key *)
+		((char *)block + XFS_RTREFCOUNT_BLOCK_LEN +
+		 (index - 1) * sizeof(struct xfs_refcount_key));
+}
+
+static inline xfs_rtrefcount_ptr_t *
+xfs_rtrefcount_ptr_addr(
+	struct xfs_btree_block	*block,
+	unsigned int		index,
+	unsigned int		maxrecs)
+{
+	return (xfs_rtrefcount_ptr_t *)
+		((char *)block + XFS_RTREFCOUNT_BLOCK_LEN +
+		 maxrecs * sizeof(struct xfs_refcount_key) +
+		 (index - 1) * sizeof(xfs_rtrefcount_ptr_t));
+}
+
+unsigned int xfs_rtrefcountbt_maxlevels_ondisk(void);
+int __init xfs_rtrefcountbt_init_cur_cache(void);
+void xfs_rtrefcountbt_destroy_cur_cache(void);
+
+#endif	/* __XFS_RTREFCOUNT_BTREE_H__ */
diff --git a/fs/xfs/libxfs/xfs_sb.c b/fs/xfs/libxfs/xfs_sb.c
index 83fb14b4074c8d..3dc5f5dba162d0 100644
--- a/fs/xfs/libxfs/xfs_sb.c
+++ b/fs/xfs/libxfs/xfs_sb.c
@@ -29,6 +29,7 @@
 #include "xfs_exchrange.h"
 #include "xfs_rtgroup.h"
 #include "xfs_rtrmap_btree.h"
+#include "xfs_rtrefcount_btree.h"
 
 /*
  * Physical superblock buffer manipulations. Shared with libxfs in userspace.
@@ -1226,6 +1227,13 @@ xfs_sb_mount_common(
 	mp->m_refc_mnr[0] = mp->m_refc_mxr[0] / 2;
 	mp->m_refc_mnr[1] = mp->m_refc_mxr[1] / 2;
 
+	mp->m_rtrefc_mxr[0] = xfs_rtrefcountbt_maxrecs(mp, sbp->sb_blocksize,
+			true);
+	mp->m_rtrefc_mxr[1] = xfs_rtrefcountbt_maxrecs(mp, sbp->sb_blocksize,
+			false);
+	mp->m_rtrefc_mnr[0] = mp->m_rtrefc_mxr[0] / 2;
+	mp->m_rtrefc_mnr[1] = mp->m_rtrefc_mxr[1] / 2;
+
 	mp->m_bsize = XFS_FSB_TO_BB(mp, 1);
 	mp->m_alloc_set_aside = xfs_alloc_set_aside(mp);
 	mp->m_ag_max_usable = xfs_alloc_ag_max_usable(mp);
diff --git a/fs/xfs/libxfs/xfs_shared.h b/fs/xfs/libxfs/xfs_shared.h
index 960716c387cc2b..b1e0d9bc1f7d70 100644
--- a/fs/xfs/libxfs/xfs_shared.h
+++ b/fs/xfs/libxfs/xfs_shared.h
@@ -42,6 +42,7 @@ extern const struct xfs_buf_ops xfs_rtbitmap_buf_ops;
 extern const struct xfs_buf_ops xfs_rtsummary_buf_ops;
 extern const struct xfs_buf_ops xfs_rtbuf_ops;
 extern const struct xfs_buf_ops xfs_rtsb_buf_ops;
+extern const struct xfs_buf_ops xfs_rtrefcountbt_buf_ops;
 extern const struct xfs_buf_ops xfs_rtrmapbt_buf_ops;
 extern const struct xfs_buf_ops xfs_sb_buf_ops;
 extern const struct xfs_buf_ops xfs_sb_quiet_buf_ops;
@@ -58,6 +59,7 @@ extern const struct xfs_btree_ops xfs_rmapbt_ops;
 extern const struct xfs_btree_ops xfs_rmapbt_mem_ops;
 extern const struct xfs_btree_ops xfs_rtrmapbt_ops;
 extern const struct xfs_btree_ops xfs_rtrmapbt_mem_ops;
+extern const struct xfs_btree_ops xfs_rtrefcountbt_ops;
 
 static inline bool xfs_btree_is_bno(const struct xfs_btree_ops *ops)
 {
@@ -114,6 +116,11 @@ static inline bool xfs_btree_is_rtrmap(const struct xfs_btree_ops *ops)
 	return ops == &xfs_rtrmapbt_ops;
 }
 
+static inline bool xfs_btree_is_rtrefcount(const struct xfs_btree_ops *ops)
+{
+	return ops == &xfs_rtrefcountbt_ops;
+}
+
 /* log size calculation functions */
 int	xfs_log_calc_unit_res(struct xfs_mount *mp, int unit_bytes);
 int	xfs_log_calc_minimum_size(struct xfs_mount *);
diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c
index 7b7d21b50d5409..477c5262cf9120 100644
--- a/fs/xfs/xfs_mount.c
+++ b/fs/xfs/xfs_mount.c
@@ -38,6 +38,7 @@
 #include "xfs_metafile.h"
 #include "xfs_rtgroup.h"
 #include "xfs_rtrmap_btree.h"
+#include "xfs_rtrefcount_btree.h"
 #include "scrub/stats.h"
 
 static DEFINE_MUTEX(xfs_uuid_table_mutex);
@@ -656,7 +657,8 @@ static inline void
 xfs_rtbtree_compute_maxlevels(
 	struct xfs_mount	*mp)
 {
-	mp->m_rtbtree_maxlevels = mp->m_rtrmap_maxlevels;
+	mp->m_rtbtree_maxlevels = max(mp->m_rtrmap_maxlevels,
+				      mp->m_rtrefc_maxlevels);
 }
 
 /*
@@ -729,6 +731,7 @@ xfs_mountfs(
 	xfs_rmapbt_compute_maxlevels(mp);
 	xfs_rtrmapbt_compute_maxlevels(mp);
 	xfs_refcountbt_compute_maxlevels(mp);
+	xfs_rtrefcountbt_compute_maxlevels(mp);
 
 	xfs_agbtree_compute_maxlevels(mp);
 	xfs_rtbtree_compute_maxlevels(mp);
diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h
index 1bc95fb170db61..9a1516080e6361 100644
--- a/fs/xfs/xfs_mount.h
+++ b/fs/xfs/xfs_mount.h
@@ -162,11 +162,14 @@ typedef struct xfs_mount {
 	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_rtrefc_mxr[2]; /* max rtrefc btree records */
+	uint			m_rtrefc_mnr[2]; /* min rtrefc btree records */
 	uint			m_alloc_maxlevels; /* max alloc btree levels */
 	uint			m_bm_maxlevels[2]; /* max bmap 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 */
+	uint			m_rtrefc_maxlevels; /* max rtrefc btree level */
 	unsigned int		m_agbtree_maxlevels; /* max level of all AG btrees */
 	unsigned int		m_rtbtree_maxlevels; /* max level of all rt btrees */
 	xfs_extlen_t		m_ag_prealloc_blocks; /* reserved ag blocks */
@@ -408,6 +411,12 @@ static inline bool xfs_has_rtrmapbt(struct xfs_mount *mp)
 	       xfs_has_rmapbt(mp);
 }
 
+static inline bool xfs_has_rtreflink(struct xfs_mount *mp)
+{
+	return xfs_has_metadir(mp) && xfs_has_realtime(mp) &&
+	       xfs_has_reflink(mp);
+}
+
 /*
  * Some features are always on for v5 file systems, allow the compiler to
  * eliminiate dead code when building without v4 support.
diff --git a/fs/xfs/xfs_stats.c b/fs/xfs/xfs_stats.c
index b7f2988bc03bb7..35c7fb3ba32448 100644
--- a/fs/xfs/xfs_stats.c
+++ b/fs/xfs/xfs_stats.c
@@ -54,7 +54,8 @@ int xfs_stats_format(struct xfsstats __percpu *stats, char *buf)
 		{ "rmapbt_mem",		xfsstats_offset(xs_rcbag_2)	},
 		{ "rcbagbt",		xfsstats_offset(xs_rtrmap_2)	},
 		{ "rtrmapbt",		xfsstats_offset(xs_rtrmap_mem_2)},
-		{ "rtrmapbt_mem",	xfsstats_offset(xs_qm_dqreclaims)},
+		{ "rtrmapbt_mem",	xfsstats_offset(xs_rtrefcbt_2)	},
+		{ "rtrefcntbt",		xfsstats_offset(xs_qm_dqreclaims)},
 		/* we print both series of quota information together */
 		{ "qm",			xfsstats_offset(xs_xstrat_bytes)},
 	};
diff --git a/fs/xfs/xfs_stats.h b/fs/xfs/xfs_stats.h
index 9c47de5dff2dd6..15ba1abcf2536f 100644
--- a/fs/xfs/xfs_stats.h
+++ b/fs/xfs/xfs_stats.h
@@ -129,6 +129,7 @@ struct __xfsstats {
 	uint32_t		xs_rcbag_2[__XBTS_MAX];
 	uint32_t		xs_rtrmap_2[__XBTS_MAX];
 	uint32_t		xs_rtrmap_mem_2[__XBTS_MAX];
+	uint32_t		xs_rtrefcbt_2[__XBTS_MAX];
 	uint32_t		xs_qm_dqreclaims;
 	uint32_t		xs_qm_dqreclaim_misses;
 	uint32_t		xs_qm_dquot_dups;

[PATCH 04/43] xfs: realtime refcount btree transaction reservations

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

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 rtrefcountbt to add the record and a second
split in the regular refcountbt to record the rtrefcountbt split.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_trans_resv.c |   25 ++++++++++++++++++++++---
 1 file changed, 22 insertions(+), 3 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_trans_resv.c b/fs/xfs/libxfs/xfs_trans_resv.c
index f3392eb2d7f41f..13d00c7166e178 100644
--- a/fs/xfs/libxfs/xfs_trans_resv.c
+++ b/fs/xfs/libxfs/xfs_trans_resv.c
@@ -92,6 +92,14 @@ xfs_refcountbt_block_count(
 	return num_ops * (2 * mp->m_refc_maxlevels - 1);
 }
 
+static unsigned int
+xfs_rtrefcountbt_block_count(
+	struct xfs_mount	*mp,
+	unsigned int		num_ops)
+{
+	return num_ops * (2 * mp->m_rtrefc_maxlevels - 1);
+}
+
 /*
  * Logging inodes is really tricksy. They are logged in memory format,
  * which means that what we write into the log doesn't directly translate into
@@ -259,10 +267,13 @@ xfs_rtalloc_block_count(
  * Compute the log reservation required to handle the refcount update
  * transaction.  Refcount updates are always done via deferred log items.
  *
- * This is calculated as:
+ * This is calculated as the max of:
  * Data device refcount updates (t1):
  *    the agfs of the ags containing the blocks: nr_ops * sector size
  *    the refcount btrees: nr_ops * 1 trees * (2 * max depth - 1) * block size
+ * Realtime refcount updates (t2);
+ *    the rt refcount inode
+ *    the rtrefcount btrees: nr_ops * 1 trees * (2 * max depth - 1) * block size
  */
 static unsigned int
 xfs_calc_refcountbt_reservation(
@@ -270,12 +281,20 @@ xfs_calc_refcountbt_reservation(
 	unsigned int		nr_ops)
 {
 	unsigned int		blksz = XFS_FSB_TO_B(mp, 1);
+	unsigned int		t1, t2 = 0;
 
 	if (!xfs_has_reflink(mp))
 		return 0;
 
-	return xfs_calc_buf_res(nr_ops, mp->m_sb.sb_sectsize) +
-	       xfs_calc_buf_res(xfs_refcountbt_block_count(mp, nr_ops), blksz);
+	t1 = xfs_calc_buf_res(nr_ops, mp->m_sb.sb_sectsize) +
+	     xfs_calc_buf_res(xfs_refcountbt_block_count(mp, nr_ops), blksz);
+
+	if (xfs_has_realtime(mp))
+		t2 = xfs_calc_inode_res(mp, 1) +
+		     xfs_calc_buf_res(xfs_rtrefcountbt_block_count(mp, nr_ops),
+				     blksz);
+
+	return max(t1, t2);
 }
 
 /*

[PATCH 05/43] xfs: add realtime refcount btree operations

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Implement the generic btree operations needed to manipulate rtrefcount
btree blocks. This is different from the regular refcountbt 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" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_rtrefcount_btree.c |  148 ++++++++++++++++++++++++++++++++++
 1 file changed, 148 insertions(+)


diff --git a/fs/xfs/libxfs/xfs_rtrefcount_btree.c b/fs/xfs/libxfs/xfs_rtrefcount_btree.c
index d07d3b1e78f730..e30af941581651 100644
--- a/fs/xfs/libxfs/xfs_rtrefcount_btree.c
+++ b/fs/xfs/libxfs/xfs_rtrefcount_btree.c
@@ -19,6 +19,7 @@
 #include "xfs_btree.h"
 #include "xfs_btree_staging.h"
 #include "xfs_rtrefcount_btree.h"
+#include "xfs_refcount.h"
 #include "xfs_trace.h"
 #include "xfs_cksum.h"
 #include "xfs_error.h"
@@ -45,6 +46,106 @@ xfs_rtrefcountbt_dup_cursor(
 	return xfs_rtrefcountbt_init_cursor(cur->bc_tp, to_rtg(cur->bc_group));
 }
 
+STATIC int
+xfs_rtrefcountbt_get_minrecs(
+	struct xfs_btree_cur	*cur,
+	int			level)
+{
+	if (level == cur->bc_nlevels - 1) {
+		struct xfs_ifork	*ifp = xfs_btree_ifork_ptr(cur);
+
+		return xfs_rtrefcountbt_maxrecs(cur->bc_mp, ifp->if_broot_bytes,
+				level == 0) / 2;
+	}
+
+	return cur->bc_mp->m_rtrefc_mnr[level != 0];
+}
+
+STATIC int
+xfs_rtrefcountbt_get_maxrecs(
+	struct xfs_btree_cur	*cur,
+	int			level)
+{
+	if (level == cur->bc_nlevels - 1) {
+		struct xfs_ifork	*ifp = xfs_btree_ifork_ptr(cur);
+
+		return xfs_rtrefcountbt_maxrecs(cur->bc_mp, ifp->if_broot_bytes,
+				level == 0);
+	}
+
+	return cur->bc_mp->m_rtrefc_mxr[level != 0];
+}
+
+STATIC void
+xfs_rtrefcountbt_init_key_from_rec(
+	union xfs_btree_key		*key,
+	const union xfs_btree_rec	*rec)
+{
+	key->refc.rc_startblock = rec->refc.rc_startblock;
+}
+
+STATIC void
+xfs_rtrefcountbt_init_high_key_from_rec(
+	union xfs_btree_key		*key,
+	const union xfs_btree_rec	*rec)
+{
+	__u32				x;
+
+	x = be32_to_cpu(rec->refc.rc_startblock);
+	x += be32_to_cpu(rec->refc.rc_blockcount) - 1;
+	key->refc.rc_startblock = cpu_to_be32(x);
+}
+
+STATIC void
+xfs_rtrefcountbt_init_rec_from_cur(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*rec)
+{
+	const struct xfs_refcount_irec *irec = &cur->bc_rec.rc;
+	uint32_t		start;
+
+	start = xfs_refcount_encode_startblock(irec->rc_startblock,
+			irec->rc_domain);
+	rec->refc.rc_startblock = cpu_to_be32(start);
+	rec->refc.rc_blockcount = cpu_to_be32(cur->bc_rec.rc.rc_blockcount);
+	rec->refc.rc_refcount = cpu_to_be32(cur->bc_rec.rc.rc_refcount);
+}
+
+STATIC void
+xfs_rtrefcountbt_init_ptr_from_cur(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr)
+{
+	ptr->l = 0;
+}
+
+STATIC int64_t
+xfs_rtrefcountbt_key_diff(
+	struct xfs_btree_cur		*cur,
+	const union xfs_btree_key	*key)
+{
+	const struct xfs_refcount_key	*kp = &key->refc;
+	const struct xfs_refcount_irec	*irec = &cur->bc_rec.rc;
+	uint32_t			start;
+
+	start = xfs_refcount_encode_startblock(irec->rc_startblock,
+			irec->rc_domain);
+	return (int64_t)be32_to_cpu(kp->rc_startblock) - start;
+}
+
+STATIC int64_t
+xfs_rtrefcountbt_diff_two_keys(
+	struct xfs_btree_cur		*cur,
+	const union xfs_btree_key	*k1,
+	const union xfs_btree_key	*k2,
+	const union xfs_btree_key	*mask)
+{
+	ASSERT(!mask || mask->refc.rc_startblock);
+
+	return (int64_t)be32_to_cpu(k1->refc.rc_startblock) -
+			be32_to_cpu(k2->refc.rc_startblock);
+}
+
 static xfs_failaddr_t
 xfs_rtrefcountbt_verify(
 	struct xfs_buf		*bp)
@@ -111,6 +212,40 @@ const struct xfs_buf_ops xfs_rtrefcountbt_buf_ops = {
 	.verify_struct		= xfs_rtrefcountbt_verify,
 };
 
+STATIC int
+xfs_rtrefcountbt_keys_inorder(
+	struct xfs_btree_cur		*cur,
+	const union xfs_btree_key	*k1,
+	const union xfs_btree_key	*k2)
+{
+	return be32_to_cpu(k1->refc.rc_startblock) <
+	       be32_to_cpu(k2->refc.rc_startblock);
+}
+
+STATIC int
+xfs_rtrefcountbt_recs_inorder(
+	struct xfs_btree_cur		*cur,
+	const union xfs_btree_rec	*r1,
+	const union xfs_btree_rec	*r2)
+{
+	return  be32_to_cpu(r1->refc.rc_startblock) +
+		be32_to_cpu(r1->refc.rc_blockcount) <=
+		be32_to_cpu(r2->refc.rc_startblock);
+}
+
+STATIC enum xbtree_key_contig
+xfs_rtrefcountbt_keys_contiguous(
+	struct xfs_btree_cur		*cur,
+	const union xfs_btree_key	*key1,
+	const union xfs_btree_key	*key2,
+	const union xfs_btree_key	*mask)
+{
+	ASSERT(!mask || mask->refc.rc_startblock);
+
+	return xbtree_key_contig(be32_to_cpu(key1->refc.rc_startblock),
+				 be32_to_cpu(key2->refc.rc_startblock));
+}
+
 const struct xfs_btree_ops xfs_rtrefcountbt_ops = {
 	.name			= "rtrefcount",
 	.type			= XFS_BTREE_TYPE_INODE,
@@ -124,7 +259,20 @@ const struct xfs_btree_ops xfs_rtrefcountbt_ops = {
 	.statoff		= XFS_STATS_CALC_INDEX(xs_rtrefcbt_2),
 
 	.dup_cursor		= xfs_rtrefcountbt_dup_cursor,
+	.alloc_block		= xfs_btree_alloc_metafile_block,
+	.free_block		= xfs_btree_free_metafile_block,
+	.get_minrecs		= xfs_rtrefcountbt_get_minrecs,
+	.get_maxrecs		= xfs_rtrefcountbt_get_maxrecs,
+	.init_key_from_rec	= xfs_rtrefcountbt_init_key_from_rec,
+	.init_high_key_from_rec	= xfs_rtrefcountbt_init_high_key_from_rec,
+	.init_rec_from_cur	= xfs_rtrefcountbt_init_rec_from_cur,
+	.init_ptr_from_cur	= xfs_rtrefcountbt_init_ptr_from_cur,
+	.key_diff		= xfs_rtrefcountbt_key_diff,
 	.buf_ops		= &xfs_rtrefcountbt_buf_ops,
+	.diff_two_keys		= xfs_rtrefcountbt_diff_two_keys,
+	.keys_inorder		= xfs_rtrefcountbt_keys_inorder,
+	.recs_inorder		= xfs_rtrefcountbt_recs_inorder,
+	.keys_contiguous	= xfs_rtrefcountbt_keys_contiguous,
 };
 
 /* Allocate a new rt refcount btree cursor. */

[PATCH 06/43] xfs: prepare refcount functions to deal with rtrefcountbt

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

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

Note that we leave the _recover_cow_leftovers functions for a separate
patch so that we can convert it all at once.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_refcount.c |   53 ++++++++++++++++++++++++++++++++++++------
 fs/xfs/libxfs/xfs_refcount.h |    3 ++
 2 files changed, 48 insertions(+), 8 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_refcount.c b/fs/xfs/libxfs/xfs_refcount.c
index faace12fe2e383..9be40ac16c7d56 100644
--- a/fs/xfs/libxfs/xfs_refcount.c
+++ b/fs/xfs/libxfs/xfs_refcount.c
@@ -25,6 +25,7 @@
 #include "xfs_ag.h"
 #include "xfs_health.h"
 #include "xfs_refcount_item.h"
+#include "xfs_rtgroup.h"
 
 struct kmem_cache	*xfs_refcount_intent_cache;
 
@@ -144,6 +145,37 @@ xfs_refcount_check_irec(
 	return NULL;
 }
 
+xfs_failaddr_t
+xfs_rtrefcount_check_irec(
+	struct xfs_rtgroup		*rtg,
+	const struct xfs_refcount_irec	*irec)
+{
+	if (irec->rc_blockcount == 0 || irec->rc_blockcount > XFS_REFC_LEN_MAX)
+		return __this_address;
+
+	if (!xfs_refcount_check_domain(irec))
+		return __this_address;
+
+	/* check for valid extent range, including overflow */
+	if (!xfs_verify_rgbext(rtg, irec->rc_startblock, irec->rc_blockcount))
+		return __this_address;
+
+	if (irec->rc_refcount == 0 || irec->rc_refcount > XFS_REFC_REFCOUNT_MAX)
+		return __this_address;
+
+	return NULL;
+}
+
+static inline xfs_failaddr_t
+xfs_refcount_check_btrec(
+	struct xfs_btree_cur		*cur,
+	const struct xfs_refcount_irec	*irec)
+{
+	if (xfs_btree_is_rtrefcount(cur->bc_ops))
+		return xfs_rtrefcount_check_irec(to_rtg(cur->bc_group), irec);
+	return xfs_refcount_check_irec(to_perag(cur->bc_group), irec);
+}
+
 static inline int
 xfs_refcount_complain_bad_rec(
 	struct xfs_btree_cur		*cur,
@@ -152,9 +184,15 @@ xfs_refcount_complain_bad_rec(
 {
 	struct xfs_mount		*mp = cur->bc_mp;
 
-	xfs_warn(mp,
+	if (xfs_btree_is_rtrefcount(cur->bc_ops)) {
+		xfs_warn(mp,
+ "RT Refcount BTree record corruption in rtgroup %u detected at %pS!",
+				cur->bc_group->xg_gno, fa);
+	} else {
+		xfs_warn(mp,
  "Refcount BTree record corruption in AG %d detected at %pS!",
 				cur->bc_group->xg_gno, fa);
+	}
 	xfs_warn(mp,
 		"Start block 0x%x, block count 0x%x, references 0x%x",
 		irec->rc_startblock, irec->rc_blockcount, irec->rc_refcount);
@@ -180,7 +218,7 @@ xfs_refcount_get_rec(
 		return error;
 
 	xfs_refcount_btrec_to_irec(rec, irec);
-	fa = xfs_refcount_check_irec(to_perag(cur->bc_group), irec);
+	fa = xfs_refcount_check_btrec(cur, irec);
 	if (fa)
 		return xfs_refcount_complain_bad_rec(cur, fa, irec);
 
@@ -1065,7 +1103,7 @@ xfs_refcount_still_have_space(
 	 */
 	overhead = xfs_allocfree_block_count(cur->bc_mp,
 				cur->bc_refc.shape_changes);
-	overhead += cur->bc_mp->m_refc_maxlevels;
+	overhead += cur->bc_maxlevels;
 	overhead *= cur->bc_mp->m_sb.sb_blocksize;
 
 	/*
@@ -1117,7 +1155,7 @@ xfs_refcount_adjust_extents(
 		if (error)
 			goto out_error;
 		if (!found_rec || ext.rc_domain != XFS_REFC_DOMAIN_SHARED) {
-			ext.rc_startblock = cur->bc_mp->m_sb.sb_agblocks;
+			ext.rc_startblock = xfs_group_max_blocks(cur->bc_group);
 			ext.rc_blockcount = 0;
 			ext.rc_refcount = 0;
 			ext.rc_domain = XFS_REFC_DOMAIN_SHARED;
@@ -1666,7 +1704,7 @@ xfs_refcount_adjust_cow_extents(
 		goto out_error;
 	}
 	if (!found_rec) {
-		ext.rc_startblock = cur->bc_mp->m_sb.sb_agblocks;
+		ext.rc_startblock = xfs_group_max_blocks(cur->bc_group);
 		ext.rc_blockcount = 0;
 		ext.rc_refcount = 0;
 		ext.rc_domain = XFS_REFC_DOMAIN_COW;
@@ -1877,8 +1915,7 @@ xfs_refcount_recover_extent(
 	INIT_LIST_HEAD(&rr->rr_list);
 	xfs_refcount_btrec_to_irec(rec, &rr->rr_rrec);
 
-	if (xfs_refcount_check_irec(to_perag(cur->bc_group), &rr->rr_rrec) !=
-			NULL ||
+	if (xfs_refcount_check_btrec(cur, &rr->rr_rrec) != NULL ||
 	    XFS_IS_CORRUPT(cur->bc_mp,
 			   rr->rr_rrec.rc_domain != XFS_REFC_DOMAIN_COW)) {
 		xfs_btree_mark_sick(cur);
@@ -2026,7 +2063,7 @@ xfs_refcount_query_range_helper(
 	xfs_failaddr_t			fa;
 
 	xfs_refcount_btrec_to_irec(rec, &irec);
-	fa = xfs_refcount_check_irec(to_perag(cur->bc_group), &irec);
+	fa = xfs_refcount_check_btrec(cur, &irec);
 	if (fa)
 		return xfs_refcount_complain_bad_rec(cur, fa, &irec);
 
diff --git a/fs/xfs/libxfs/xfs_refcount.h b/fs/xfs/libxfs/xfs_refcount.h
index 62d78afcf1f3ff..9cd58d48716f83 100644
--- a/fs/xfs/libxfs/xfs_refcount.h
+++ b/fs/xfs/libxfs/xfs_refcount.h
@@ -12,6 +12,7 @@ struct xfs_perag;
 struct xfs_btree_cur;
 struct xfs_bmbt_irec;
 struct xfs_refcount_irec;
+struct xfs_rtgroup;
 
 extern int xfs_refcount_lookup_le(struct xfs_btree_cur *cur,
 		enum xfs_refc_domain domain, xfs_agblock_t bno, int *stat);
@@ -120,6 +121,8 @@ extern void xfs_refcount_btrec_to_irec(const union xfs_btree_rec *rec,
 		struct xfs_refcount_irec *irec);
 xfs_failaddr_t xfs_refcount_check_irec(struct xfs_perag *pag,
 		const struct xfs_refcount_irec *irec);
+xfs_failaddr_t xfs_rtrefcount_check_irec(struct xfs_rtgroup *rtg,
+		const struct xfs_refcount_irec *irec);
 extern int xfs_refcount_insert(struct xfs_btree_cur *cur,
 		struct xfs_refcount_irec *irec, int *stat);
 

[PATCH 07/43] xfs: add a realtime flag to the refcount update log redo items

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

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

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_bmap.c        |   10 +-
 fs/xfs/libxfs/xfs_defer.h       |    1 
 fs/xfs/libxfs/xfs_log_format.h  |    6 +
 fs/xfs/libxfs/xfs_log_recover.h |    2 
 fs/xfs/libxfs/xfs_refcount.c    |   64 ++++++++---
 fs/xfs/libxfs/xfs_refcount.h    |   17 ++-
 fs/xfs/scrub/cow_repair.c       |    2 
 fs/xfs/scrub/reap.c             |    5 +
 fs/xfs/xfs_log_recover.c        |    2 
 fs/xfs/xfs_refcount_item.c      |  224 +++++++++++++++++++++++++++++++++++++--
 fs/xfs/xfs_reflink.c            |   19 ++-
 11 files changed, 298 insertions(+), 54 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c
index 02323936cc9b20..d63713630236e7 100644
--- a/fs/xfs/libxfs/xfs_bmap.c
+++ b/fs/xfs/libxfs/xfs_bmap.c
@@ -4564,8 +4564,9 @@ xfs_bmapi_write(
 			 * the refcount btree for orphan recovery.
 			 */
 			if (whichfork == XFS_COW_FORK)
-				xfs_refcount_alloc_cow_extent(tp, bma.blkno,
-						bma.length);
+				xfs_refcount_alloc_cow_extent(tp,
+						XFS_IS_REALTIME_INODE(ip),
+						bma.blkno, bma.length);
 		}
 
 		/* Deal with the allocated space we found.  */
@@ -4740,7 +4741,8 @@ xfs_bmapi_convert_one_delalloc(
 		*seq = READ_ONCE(ifp->if_seq);
 
 	if (whichfork == XFS_COW_FORK)
-		xfs_refcount_alloc_cow_extent(tp, bma.blkno, bma.length);
+		xfs_refcount_alloc_cow_extent(tp, XFS_IS_REALTIME_INODE(ip),
+				bma.blkno, bma.length);
 
 	error = xfs_bmap_btree_to_extents(tp, ip, bma.cur, &bma.logflags,
 			whichfork);
@@ -5388,7 +5390,7 @@ xfs_bmap_del_extent_real(
 		bool	isrt = xfs_ifork_is_realtime(ip, whichfork);
 
 		if (xfs_is_reflink_inode(ip) && whichfork == XFS_DATA_FORK) {
-			xfs_refcount_decrease_extent(tp, del);
+			xfs_refcount_decrease_extent(tp, isrt, del);
 		} else if (isrt && !xfs_has_rtgroups(mp)) {
 			error = xfs_bmap_free_rtblocks(tp, del);
 		} else {
diff --git a/fs/xfs/libxfs/xfs_defer.h b/fs/xfs/libxfs/xfs_defer.h
index 1e2477eaa5a844..9effd95ddcd4ee 100644
--- a/fs/xfs/libxfs/xfs_defer.h
+++ b/fs/xfs/libxfs/xfs_defer.h
@@ -68,6 +68,7 @@ struct xfs_defer_op_type {
 
 extern const struct xfs_defer_op_type xfs_bmap_update_defer_type;
 extern const struct xfs_defer_op_type xfs_refcount_update_defer_type;
+extern const struct xfs_defer_op_type xfs_rtrefcount_update_defer_type;
 extern const struct xfs_defer_op_type xfs_rmap_update_defer_type;
 extern const struct xfs_defer_op_type xfs_rtrmap_update_defer_type;
 extern const struct xfs_defer_op_type xfs_extent_free_defer_type;
diff --git a/fs/xfs/libxfs/xfs_log_format.h b/fs/xfs/libxfs/xfs_log_format.h
index a7e0e479454d3d..ec7157eaba5f68 100644
--- a/fs/xfs/libxfs/xfs_log_format.h
+++ b/fs/xfs/libxfs/xfs_log_format.h
@@ -252,6 +252,8 @@ typedef struct xfs_trans_header {
 #define	XFS_LI_EFD_RT		0x124b	/* realtime extent free done */
 #define	XFS_LI_RUI_RT		0x124c	/* realtime rmap update intent */
 #define	XFS_LI_RUD_RT		0x124d	/* realtime rmap update done */
+#define	XFS_LI_CUI_RT		0x124e	/* realtime refcount update intent */
+#define	XFS_LI_CUD_RT		0x124f	/* realtime refcount update done */
 
 #define XFS_LI_TYPE_DESC \
 	{ XFS_LI_EFI,		"XFS_LI_EFI" }, \
@@ -275,7 +277,9 @@ typedef struct xfs_trans_header {
 	{ XFS_LI_EFI_RT,	"XFS_LI_EFI_RT" }, \
 	{ XFS_LI_EFD_RT,	"XFS_LI_EFD_RT" }, \
 	{ XFS_LI_RUI_RT,	"XFS_LI_RUI_RT" }, \
-	{ XFS_LI_RUD_RT,	"XFS_LI_RUD_RT" }
+	{ XFS_LI_RUD_RT,	"XFS_LI_RUD_RT" }, \
+	{ XFS_LI_CUI_RT,	"XFS_LI_CUI_RT" }, \
+	{ XFS_LI_CUD_RT,	"XFS_LI_CUD_RT" }
 
 /*
  * Inode Log Item Format definitions.
diff --git a/fs/xfs/libxfs/xfs_log_recover.h b/fs/xfs/libxfs/xfs_log_recover.h
index abc705aff26dfe..66c7916fb5cd5e 100644
--- a/fs/xfs/libxfs/xfs_log_recover.h
+++ b/fs/xfs/libxfs/xfs_log_recover.h
@@ -81,6 +81,8 @@ extern const struct xlog_recover_item_ops xlog_rtefi_item_ops;
 extern const struct xlog_recover_item_ops xlog_rtefd_item_ops;
 extern const struct xlog_recover_item_ops xlog_rtrui_item_ops;
 extern const struct xlog_recover_item_ops xlog_rtrud_item_ops;
+extern const struct xlog_recover_item_ops xlog_rtcui_item_ops;
+extern const struct xlog_recover_item_ops xlog_rtcud_item_ops;
 
 /*
  * Macros, structures, prototypes for internal log manager use.
diff --git a/fs/xfs/libxfs/xfs_refcount.c b/fs/xfs/libxfs/xfs_refcount.c
index 9be40ac16c7d56..8007d15856252b 100644
--- a/fs/xfs/libxfs/xfs_refcount.c
+++ b/fs/xfs/libxfs/xfs_refcount.c
@@ -26,6 +26,7 @@
 #include "xfs_health.h"
 #include "xfs_refcount_item.h"
 #include "xfs_rtgroup.h"
+#include "xfs_rtalloc.h"
 
 struct kmem_cache	*xfs_refcount_intent_cache;
 
@@ -1123,6 +1124,22 @@ xfs_refcount_still_have_space(
 		cur->bc_refc.nr_ops * XFS_REFCOUNT_ITEM_OVERHEAD;
 }
 
+/* Schedule an extent free. */
+static int
+xrefc_free_extent(
+	struct xfs_btree_cur		*cur,
+	struct xfs_refcount_irec	*rec)
+{
+	unsigned int			flags = 0;
+
+	if (xfs_btree_is_rtrefcount(cur->bc_ops))
+		flags |= XFS_FREE_EXTENT_REALTIME;
+
+	return xfs_free_extent_later(cur->bc_tp,
+			xfs_gbno_to_fsb(cur->bc_group, rec->rc_startblock),
+			rec->rc_blockcount, NULL, XFS_AG_RESV_NONE, flags);
+}
+
 /*
  * Adjust the refcounts of middle extents.  At this point we should have
  * split extents that crossed the adjustment range; merged with adjacent
@@ -1139,7 +1156,6 @@ xfs_refcount_adjust_extents(
 	struct xfs_refcount_irec	ext, tmp;
 	int				error;
 	int				found_rec, found_tmp;
-	xfs_fsblock_t			fsbno;
 
 	/* Merging did all the work already. */
 	if (*aglen == 0)
@@ -1192,11 +1208,7 @@ xfs_refcount_adjust_extents(
 					goto out_error;
 				}
 			} else {
-				fsbno = xfs_agbno_to_fsb(to_perag(cur->bc_group),
-						tmp.rc_startblock);
-				error = xfs_free_extent_later(cur->bc_tp, fsbno,
-						  tmp.rc_blockcount, NULL,
-						  XFS_AG_RESV_NONE, 0);
+				error = xrefc_free_extent(cur, &tmp);
 				if (error)
 					goto out_error;
 			}
@@ -1254,11 +1266,7 @@ xfs_refcount_adjust_extents(
 			}
 			goto advloop;
 		} else {
-			fsbno = xfs_agbno_to_fsb(to_perag(cur->bc_group),
-					ext.rc_startblock);
-			error = xfs_free_extent_later(cur->bc_tp, fsbno,
-					ext.rc_blockcount, NULL,
-					XFS_AG_RESV_NONE, 0);
+			error = xrefc_free_extent(cur, &ext);
 			if (error)
 				goto out_error;
 		}
@@ -1454,6 +1462,20 @@ xfs_refcount_finish_one(
 	return error;
 }
 
+/*
+ * Process one of the deferred realtime refcount operations.  We pass back the
+ * btree cursor to maintain our lock on the btree between calls.
+ */
+int
+xfs_rtrefcount_finish_one(
+	struct xfs_trans		*tp,
+	struct xfs_refcount_intent	*ri,
+	struct xfs_btree_cur		**pcur)
+{
+	ASSERT(0);
+	return -EFSCORRUPTED;
+}
+
 /*
  * Record a refcount intent for later processing.
  */
@@ -1461,6 +1483,7 @@ static void
 __xfs_refcount_add(
 	struct xfs_trans		*tp,
 	enum xfs_refcount_intent_type	type,
+	bool				isrt,
 	xfs_fsblock_t			startblock,
 	xfs_extlen_t			blockcount)
 {
@@ -1472,6 +1495,7 @@ __xfs_refcount_add(
 	ri->ri_type = type;
 	ri->ri_startblock = startblock;
 	ri->ri_blockcount = blockcount;
+	ri->ri_realtime = isrt;
 
 	xfs_refcount_defer_add(tp, ri);
 }
@@ -1482,12 +1506,13 @@ __xfs_refcount_add(
 void
 xfs_refcount_increase_extent(
 	struct xfs_trans		*tp,
+	bool				isrt,
 	struct xfs_bmbt_irec		*PREV)
 {
 	if (!xfs_has_reflink(tp->t_mountp))
 		return;
 
-	__xfs_refcount_add(tp, XFS_REFCOUNT_INCREASE, PREV->br_startblock,
+	__xfs_refcount_add(tp, XFS_REFCOUNT_INCREASE, isrt, PREV->br_startblock,
 			PREV->br_blockcount);
 }
 
@@ -1497,12 +1522,13 @@ xfs_refcount_increase_extent(
 void
 xfs_refcount_decrease_extent(
 	struct xfs_trans		*tp,
+	bool				isrt,
 	struct xfs_bmbt_irec		*PREV)
 {
 	if (!xfs_has_reflink(tp->t_mountp))
 		return;
 
-	__xfs_refcount_add(tp, XFS_REFCOUNT_DECREASE, PREV->br_startblock,
+	__xfs_refcount_add(tp, XFS_REFCOUNT_DECREASE, isrt, PREV->br_startblock,
 			PREV->br_blockcount);
 }
 
@@ -1858,6 +1884,7 @@ __xfs_refcount_cow_free(
 void
 xfs_refcount_alloc_cow_extent(
 	struct xfs_trans		*tp,
+	bool				isrt,
 	xfs_fsblock_t			fsb,
 	xfs_extlen_t			len)
 {
@@ -1866,16 +1893,17 @@ xfs_refcount_alloc_cow_extent(
 	if (!xfs_has_reflink(mp))
 		return;
 
-	__xfs_refcount_add(tp, XFS_REFCOUNT_ALLOC_COW, fsb, len);
+	__xfs_refcount_add(tp, XFS_REFCOUNT_ALLOC_COW, isrt, fsb, len);
 
 	/* Add rmap entry */
-	xfs_rmap_alloc_extent(tp, false, fsb, len, XFS_RMAP_OWN_COW);
+	xfs_rmap_alloc_extent(tp, isrt, fsb, len, XFS_RMAP_OWN_COW);
 }
 
 /* Forget a CoW staging event in the refcount btree. */
 void
 xfs_refcount_free_cow_extent(
 	struct xfs_trans		*tp,
+	bool				isrt,
 	xfs_fsblock_t			fsb,
 	xfs_extlen_t			len)
 {
@@ -1885,8 +1913,8 @@ xfs_refcount_free_cow_extent(
 		return;
 
 	/* Remove rmap entry */
-	xfs_rmap_free_extent(tp, false, fsb, len, XFS_RMAP_OWN_COW);
-	__xfs_refcount_add(tp, XFS_REFCOUNT_FREE_COW, fsb, len);
+	xfs_rmap_free_extent(tp, isrt, fsb, len, XFS_RMAP_OWN_COW);
+	__xfs_refcount_add(tp, XFS_REFCOUNT_FREE_COW, isrt, fsb, len);
 }
 
 struct xfs_refcount_recovery {
@@ -1992,7 +2020,7 @@ xfs_refcount_recover_cow_leftovers(
 
 		/* Free the orphan record */
 		fsb = xfs_agbno_to_fsb(pag, rr->rr_rrec.rc_startblock);
-		xfs_refcount_free_cow_extent(tp, fsb,
+		xfs_refcount_free_cow_extent(tp, false, fsb,
 				rr->rr_rrec.rc_blockcount);
 
 		/* Free the block. */
diff --git a/fs/xfs/libxfs/xfs_refcount.h b/fs/xfs/libxfs/xfs_refcount.h
index 9cd58d48716f83..be11df25abcc5b 100644
--- a/fs/xfs/libxfs/xfs_refcount.h
+++ b/fs/xfs/libxfs/xfs_refcount.h
@@ -61,6 +61,7 @@ struct xfs_refcount_intent {
 	enum xfs_refcount_intent_type		ri_type;
 	xfs_extlen_t				ri_blockcount;
 	xfs_fsblock_t				ri_startblock;
+	bool					ri_realtime;
 };
 
 /* Check that the refcount is appropriate for the record domain. */
@@ -75,22 +76,24 @@ xfs_refcount_check_domain(
 	return true;
 }
 
-void xfs_refcount_increase_extent(struct xfs_trans *tp,
+void xfs_refcount_increase_extent(struct xfs_trans *tp, bool isrt,
 		struct xfs_bmbt_irec *irec);
-void xfs_refcount_decrease_extent(struct xfs_trans *tp,
+void xfs_refcount_decrease_extent(struct xfs_trans *tp, bool isrt,
 		struct xfs_bmbt_irec *irec);
 
-extern int xfs_refcount_finish_one(struct xfs_trans *tp,
+int xfs_refcount_finish_one(struct xfs_trans *tp,
+		struct xfs_refcount_intent *ri, struct xfs_btree_cur **pcur);
+int xfs_rtrefcount_finish_one(struct xfs_trans *tp,
 		struct xfs_refcount_intent *ri, struct xfs_btree_cur **pcur);
 
 extern int xfs_refcount_find_shared(struct xfs_btree_cur *cur,
 		xfs_agblock_t agbno, xfs_extlen_t aglen, xfs_agblock_t *fbno,
 		xfs_extlen_t *flen, bool find_end_of_shared);
 
-void xfs_refcount_alloc_cow_extent(struct xfs_trans *tp, xfs_fsblock_t fsb,
-		xfs_extlen_t len);
-void xfs_refcount_free_cow_extent(struct xfs_trans *tp, xfs_fsblock_t fsb,
-		xfs_extlen_t len);
+void xfs_refcount_alloc_cow_extent(struct xfs_trans *tp, bool isrt,
+		xfs_fsblock_t fsb, xfs_extlen_t len);
+void xfs_refcount_free_cow_extent(struct xfs_trans *tp, bool isrt,
+		xfs_fsblock_t fsb, xfs_extlen_t len);
 extern int xfs_refcount_recover_cow_leftovers(struct xfs_mount *mp,
 		struct xfs_perag *pag);
 
diff --git a/fs/xfs/scrub/cow_repair.c b/fs/xfs/scrub/cow_repair.c
index 5b6194cef3e5e3..ba695dd21f8b96 100644
--- a/fs/xfs/scrub/cow_repair.c
+++ b/fs/xfs/scrub/cow_repair.c
@@ -343,7 +343,7 @@ xrep_cow_alloc(
 	if (args.fsbno == NULLFSBLOCK)
 		return -ENOSPC;
 
-	xfs_refcount_alloc_cow_extent(sc->tp, args.fsbno, args.len);
+	xfs_refcount_alloc_cow_extent(sc->tp, false, args.fsbno, args.len);
 
 	repl->fsbno = args.fsbno;
 	repl->len = args.len;
diff --git a/fs/xfs/scrub/reap.c b/fs/xfs/scrub/reap.c
index 4d7f1b82dc559d..534c7696e9a1a7 100644
--- a/fs/xfs/scrub/reap.c
+++ b/fs/xfs/scrub/reap.c
@@ -419,7 +419,8 @@ xreap_agextent_iter(
 			 * records from the refcountbt, which will remove the
 			 * rmap record as well.
 			 */
-			xfs_refcount_free_cow_extent(sc->tp, fsbno, *aglenp);
+			xfs_refcount_free_cow_extent(sc->tp, false, fsbno,
+					*aglenp);
 			return 0;
 		}
 
@@ -451,7 +452,7 @@ xreap_agextent_iter(
 	if (rs->oinfo == &XFS_RMAP_OINFO_COW) {
 		ASSERT(rs->resv == XFS_AG_RESV_NONE);
 
-		xfs_refcount_free_cow_extent(sc->tp, fsbno, *aglenp);
+		xfs_refcount_free_cow_extent(sc->tp, false, fsbno, *aglenp);
 		error = xfs_free_extent_later(sc->tp, fsbno, *aglenp, NULL,
 				rs->resv, XFS_FREE_EXTENT_SKIP_DISCARD);
 		if (error)
diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index 5c95c97519c767..b3c27dbccce861 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -1822,6 +1822,8 @@ static const struct xlog_recover_item_ops *xlog_recover_item_ops[] = {
 	&xlog_rtefd_item_ops,
 	&xlog_rtrui_item_ops,
 	&xlog_rtrud_item_ops,
+	&xlog_rtcui_item_ops,
+	&xlog_rtcud_item_ops,
 };
 
 static const struct xlog_recover_item_ops *
diff --git a/fs/xfs/xfs_refcount_item.c b/fs/xfs/xfs_refcount_item.c
index c807c4b90c44e3..2086d40514d086 100644
--- a/fs/xfs/xfs_refcount_item.c
+++ b/fs/xfs/xfs_refcount_item.c
@@ -23,6 +23,7 @@
 #include "xfs_ag.h"
 #include "xfs_btree.h"
 #include "xfs_trace.h"
+#include "xfs_rtgroup.h"
 
 struct kmem_cache	*xfs_cui_cache;
 struct kmem_cache	*xfs_cud_cache;
@@ -94,8 +95,9 @@ xfs_cui_item_format(
 
 	ASSERT(atomic_read(&cuip->cui_next_extent) ==
 			cuip->cui_format.cui_nextents);
+	ASSERT(lip->li_type == XFS_LI_CUI || lip->li_type == XFS_LI_CUI_RT);
 
-	cuip->cui_format.cui_type = XFS_LI_CUI;
+	cuip->cui_format.cui_type = lip->li_type;
 	cuip->cui_format.cui_size = 1;
 
 	xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_CUI_FORMAT, &cuip->cui_format,
@@ -138,12 +140,14 @@ xfs_cui_item_release(
 STATIC struct xfs_cui_log_item *
 xfs_cui_init(
 	struct xfs_mount		*mp,
+	unsigned short			item_type,
 	uint				nextents)
-
 {
 	struct xfs_cui_log_item		*cuip;
 
 	ASSERT(nextents > 0);
+	ASSERT(item_type == XFS_LI_CUI || item_type == XFS_LI_CUI_RT);
+
 	if (nextents > XFS_CUI_MAX_FAST_EXTENTS)
 		cuip = kzalloc(xfs_cui_log_item_sizeof(nextents),
 				GFP_KERNEL | __GFP_NOFAIL);
@@ -151,7 +155,7 @@ xfs_cui_init(
 		cuip = kmem_cache_zalloc(xfs_cui_cache,
 					 GFP_KERNEL | __GFP_NOFAIL);
 
-	xfs_log_item_init(mp, &cuip->cui_item, XFS_LI_CUI, &xfs_cui_item_ops);
+	xfs_log_item_init(mp, &cuip->cui_item, item_type, &xfs_cui_item_ops);
 	cuip->cui_format.cui_nextents = nextents;
 	cuip->cui_format.cui_id = (uintptr_t)(void *)cuip;
 	atomic_set(&cuip->cui_next_extent, 0);
@@ -190,7 +194,9 @@ xfs_cud_item_format(
 	struct xfs_cud_log_item	*cudp = CUD_ITEM(lip);
 	struct xfs_log_iovec	*vecp = NULL;
 
-	cudp->cud_format.cud_type = XFS_LI_CUD;
+	ASSERT(lip->li_type == XFS_LI_CUD || lip->li_type == XFS_LI_CUD_RT);
+
+	cudp->cud_format.cud_type = lip->li_type;
 	cudp->cud_format.cud_size = 1;
 
 	xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_CUD_FORMAT, &cudp->cud_format,
@@ -234,6 +240,14 @@ static inline struct xfs_refcount_intent *ci_entry(const struct list_head *e)
 	return list_entry(e, struct xfs_refcount_intent, ri_list);
 }
 
+static inline bool
+xfs_cui_item_isrt(const struct xfs_log_item *lip)
+{
+	ASSERT(lip->li_type == XFS_LI_CUI || lip->li_type == XFS_LI_CUI_RT);
+
+	return lip->li_type == XFS_LI_CUI_RT;
+}
+
 /* Sort refcount intents by AG. */
 static int
 xfs_refcount_update_diff_items(
@@ -282,18 +296,20 @@ xfs_refcount_update_log_item(
 }
 
 static struct xfs_log_item *
-xfs_refcount_update_create_intent(
+__xfs_refcount_update_create_intent(
 	struct xfs_trans		*tp,
 	struct list_head		*items,
 	unsigned int			count,
-	bool				sort)
+	bool				sort,
+	unsigned short			item_type)
 {
 	struct xfs_mount		*mp = tp->t_mountp;
-	struct xfs_cui_log_item		*cuip = xfs_cui_init(mp, count);
+	struct xfs_cui_log_item		*cuip;
 	struct xfs_refcount_intent	*ri;
 
 	ASSERT(count > 0);
 
+	cuip = xfs_cui_init(mp, item_type, count);
 	if (sort)
 		list_sort(mp, items, xfs_refcount_update_diff_items);
 	list_for_each_entry(ri, items, ri_list)
@@ -301,6 +317,23 @@ xfs_refcount_update_create_intent(
 	return &cuip->cui_item;
 }
 
+static struct xfs_log_item *
+xfs_refcount_update_create_intent(
+	struct xfs_trans		*tp,
+	struct list_head		*items,
+	unsigned int			count,
+	bool				sort)
+{
+	return __xfs_refcount_update_create_intent(tp, items, count, sort,
+			XFS_LI_CUI);
+}
+
+static inline unsigned short
+xfs_cud_type_from_cui(const struct xfs_cui_log_item *cuip)
+{
+	return xfs_cui_item_isrt(&cuip->cui_item) ? XFS_LI_CUD_RT : XFS_LI_CUD;
+}
+
 /* Get an CUD so we can process all the deferred refcount updates. */
 static struct xfs_log_item *
 xfs_refcount_update_create_done(
@@ -312,8 +345,8 @@ xfs_refcount_update_create_done(
 	struct xfs_cud_log_item		*cudp;
 
 	cudp = kmem_cache_zalloc(xfs_cud_cache, GFP_KERNEL | __GFP_NOFAIL);
-	xfs_log_item_init(tp->t_mountp, &cudp->cud_item, XFS_LI_CUD,
-			  &xfs_cud_item_ops);
+	xfs_log_item_init(tp->t_mountp, &cudp->cud_item,
+			xfs_cud_type_from_cui(cuip), &xfs_cud_item_ops);
 	cudp->cud_cuip = cuip;
 	cudp->cud_format.cud_cui_id = cuip->cui_format.cui_id;
 
@@ -328,10 +361,20 @@ xfs_refcount_defer_add(
 {
 	struct xfs_mount		*mp = tp->t_mountp;
 
-	ri->ri_group = xfs_group_intent_get(mp, ri->ri_startblock, XG_TYPE_AG);
+	/*
+	 * Deferred refcount updates for the realtime and data sections must
+	 * use separate transactions to finish deferred work because updates to
+	 * realtime metadata files can lock AGFs to allocate btree blocks and
+	 * we don't want that mixing with the AGF locks taken to finish data
+	 * section updates.
+	 */
+	ri->ri_group = xfs_group_intent_get(mp, ri->ri_startblock,
+			ri->ri_realtime ? XG_TYPE_RTG : XG_TYPE_AG);
 
 	trace_xfs_refcount_defer(mp, ri);
-	xfs_defer_add(tp, &ri->ri_list, &xfs_refcount_update_defer_type);
+	xfs_defer_add(tp, &ri->ri_list, ri->ri_realtime ?
+			&xfs_rtrefcount_update_defer_type :
+			&xfs_refcount_update_defer_type);
 }
 
 /* Cancel a deferred refcount update. */
@@ -381,7 +424,7 @@ xfs_refcount_finish_one_cleanup(
 		return;
 	agbp = rcur->bc_ag.agbp;
 	xfs_btree_del_cursor(rcur, error);
-	if (error)
+	if (error && agbp)
 		xfs_trans_brelse(tp, agbp);
 }
 
@@ -515,10 +558,13 @@ xfs_refcount_relog_intent(
 	struct xfs_phys_extent		*pmap;
 	unsigned int			count;
 
+	ASSERT(intent->li_type == XFS_LI_CUI ||
+	       intent->li_type == XFS_LI_CUI_RT);
+
 	count = CUI_ITEM(intent)->cui_format.cui_nextents;
 	pmap = CUI_ITEM(intent)->cui_format.cui_extents;
 
-	cuip = xfs_cui_init(tp->t_mountp, count);
+	cuip = xfs_cui_init(tp->t_mountp, intent->li_type, count);
 	memcpy(cuip->cui_format.cui_extents, pmap, count * sizeof(*pmap));
 	atomic_set(&cuip->cui_next_extent, count);
 
@@ -538,6 +584,71 @@ const struct xfs_defer_op_type xfs_refcount_update_defer_type = {
 	.relog_intent	= xfs_refcount_relog_intent,
 };
 
+#ifdef CONFIG_XFS_RT
+static struct xfs_log_item *
+xfs_rtrefcount_update_create_intent(
+	struct xfs_trans		*tp,
+	struct list_head		*items,
+	unsigned int			count,
+	bool				sort)
+{
+	return __xfs_refcount_update_create_intent(tp, items, count, sort,
+			XFS_LI_CUI_RT);
+}
+
+/* Process a deferred realtime refcount update. */
+STATIC int
+xfs_rtrefcount_update_finish_item(
+	struct xfs_trans		*tp,
+	struct xfs_log_item		*done,
+	struct list_head		*item,
+	struct xfs_btree_cur		**state)
+{
+	struct xfs_refcount_intent	*ri = ci_entry(item);
+	int				error;
+
+	error = xfs_rtrefcount_finish_one(tp, ri, state);
+
+	/* Did we run out of reservation?  Requeue what we didn't finish. */
+	if (!error && ri->ri_blockcount > 0) {
+		ASSERT(ri->ri_type == XFS_REFCOUNT_INCREASE ||
+		       ri->ri_type == XFS_REFCOUNT_DECREASE);
+		return -EAGAIN;
+	}
+
+	xfs_refcount_update_cancel_item(item);
+	return error;
+}
+
+/* Clean up after calling xfs_rtrefcount_finish_one. */
+STATIC void
+xfs_rtrefcount_finish_one_cleanup(
+	struct xfs_trans	*tp,
+	struct xfs_btree_cur	*rcur,
+	int			error)
+{
+	if (rcur)
+		xfs_btree_del_cursor(rcur, error);
+}
+
+const struct xfs_defer_op_type xfs_rtrefcount_update_defer_type = {
+	.name		= "rtrefcount",
+	.max_items	= XFS_CUI_MAX_FAST_EXTENTS,
+	.create_intent	= xfs_rtrefcount_update_create_intent,
+	.abort_intent	= xfs_refcount_update_abort_intent,
+	.create_done	= xfs_refcount_update_create_done,
+	.finish_item	= xfs_rtrefcount_update_finish_item,
+	.finish_cleanup = xfs_rtrefcount_finish_one_cleanup,
+	.cancel_item	= xfs_refcount_update_cancel_item,
+	.recover_work	= xfs_refcount_recover_work,
+	.relog_intent	= xfs_refcount_relog_intent,
+};
+#else
+const struct xfs_defer_op_type xfs_rtrefcount_update_defer_type = {
+	.name		= "rtrefcount",
+};
+#endif /* CONFIG_XFS_RT */
+
 STATIC bool
 xfs_cui_item_match(
 	struct xfs_log_item	*lip,
@@ -603,7 +714,7 @@ xlog_recover_cui_commit_pass2(
 		return -EFSCORRUPTED;
 	}
 
-	cuip = xfs_cui_init(mp, cui_formatp->cui_nextents);
+	cuip = xfs_cui_init(mp, ITEM_TYPE(item), cui_formatp->cui_nextents);
 	xfs_cui_copy_format(&cuip->cui_format, cui_formatp);
 	atomic_set(&cuip->cui_next_extent, cui_formatp->cui_nextents);
 
@@ -617,6 +728,61 @@ const struct xlog_recover_item_ops xlog_cui_item_ops = {
 	.commit_pass2		= xlog_recover_cui_commit_pass2,
 };
 
+#ifdef CONFIG_XFS_RT
+STATIC int
+xlog_recover_rtcui_commit_pass2(
+	struct xlog			*log,
+	struct list_head		*buffer_list,
+	struct xlog_recover_item	*item,
+	xfs_lsn_t			lsn)
+{
+	struct xfs_mount		*mp = log->l_mp;
+	struct xfs_cui_log_item		*cuip;
+	struct xfs_cui_log_format	*cui_formatp;
+	size_t				len;
+
+	cui_formatp = item->ri_buf[0].i_addr;
+
+	if (item->ri_buf[0].i_len < xfs_cui_log_format_sizeof(0)) {
+		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
+				item->ri_buf[0].i_addr, item->ri_buf[0].i_len);
+		return -EFSCORRUPTED;
+	}
+
+	len = xfs_cui_log_format_sizeof(cui_formatp->cui_nextents);
+	if (item->ri_buf[0].i_len != len) {
+		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
+				item->ri_buf[0].i_addr, item->ri_buf[0].i_len);
+		return -EFSCORRUPTED;
+	}
+
+	cuip = xfs_cui_init(mp, ITEM_TYPE(item), cui_formatp->cui_nextents);
+	xfs_cui_copy_format(&cuip->cui_format, cui_formatp);
+	atomic_set(&cuip->cui_next_extent, cui_formatp->cui_nextents);
+
+	xlog_recover_intent_item(log, &cuip->cui_item, lsn,
+			&xfs_rtrefcount_update_defer_type);
+	return 0;
+}
+#else
+STATIC int
+xlog_recover_rtcui_commit_pass2(
+	struct xlog			*log,
+	struct list_head		*buffer_list,
+	struct xlog_recover_item	*item,
+	xfs_lsn_t			lsn)
+{
+	XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, log->l_mp,
+			item->ri_buf[0].i_addr, item->ri_buf[0].i_len);
+	return -EFSCORRUPTED;
+}
+#endif
+
+const struct xlog_recover_item_ops xlog_rtcui_item_ops = {
+	.item_type		= XFS_LI_CUI_RT,
+	.commit_pass2		= xlog_recover_rtcui_commit_pass2,
+};
+
 /*
  * This routine is called when an CUD format structure is found in a committed
  * transaction in the log. Its purpose is to cancel the corresponding CUI if it
@@ -648,3 +814,33 @@ const struct xlog_recover_item_ops xlog_cud_item_ops = {
 	.item_type		= XFS_LI_CUD,
 	.commit_pass2		= xlog_recover_cud_commit_pass2,
 };
+
+#ifdef CONFIG_XFS_RT
+STATIC int
+xlog_recover_rtcud_commit_pass2(
+	struct xlog			*log,
+	struct list_head		*buffer_list,
+	struct xlog_recover_item	*item,
+	xfs_lsn_t			lsn)
+{
+	struct xfs_cud_log_format	*cud_formatp;
+
+	cud_formatp = item->ri_buf[0].i_addr;
+	if (item->ri_buf[0].i_len != sizeof(struct xfs_cud_log_format)) {
+		XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, log->l_mp,
+				item->ri_buf[0].i_addr, item->ri_buf[0].i_len);
+		return -EFSCORRUPTED;
+	}
+
+	xlog_recover_release_intent(log, XFS_LI_CUI_RT,
+			cud_formatp->cud_cui_id);
+	return 0;
+}
+#else
+# define xlog_recover_rtcud_commit_pass2	xlog_recover_rtcui_commit_pass2
+#endif
+
+const struct xlog_recover_item_ops xlog_rtcud_item_ops = {
+	.item_type		= XFS_LI_CUD_RT,
+	.commit_pass2		= xlog_recover_rtcud_commit_pass2,
+};
diff --git a/fs/xfs/xfs_reflink.c b/fs/xfs/xfs_reflink.c
index b11769c009effc..02457e176c4252 100644
--- a/fs/xfs/xfs_reflink.c
+++ b/fs/xfs/xfs_reflink.c
@@ -571,6 +571,7 @@ xfs_reflink_cancel_cow_blocks(
 	struct xfs_ifork		*ifp = xfs_ifork_ptr(ip, XFS_COW_FORK);
 	struct xfs_bmbt_irec		got, del;
 	struct xfs_iext_cursor		icur;
+	bool				isrt = XFS_IS_REALTIME_INODE(ip);
 	int				error = 0;
 
 	if (!xfs_inode_has_cow_data(ip))
@@ -598,12 +599,13 @@ xfs_reflink_cancel_cow_blocks(
 			ASSERT((*tpp)->t_highest_agno == NULLAGNUMBER);
 
 			/* Free the CoW orphan record. */
-			xfs_refcount_free_cow_extent(*tpp, del.br_startblock,
-					del.br_blockcount);
+			xfs_refcount_free_cow_extent(*tpp, isrt,
+					del.br_startblock, del.br_blockcount);
 
 			error = xfs_free_extent_later(*tpp, del.br_startblock,
 					del.br_blockcount, NULL,
-					XFS_AG_RESV_NONE, 0);
+					XFS_AG_RESV_NONE,
+					isrt ? XFS_FREE_EXTENT_REALTIME : 0);
 			if (error)
 				break;
 
@@ -710,6 +712,7 @@ xfs_reflink_end_cow_extent(
 	struct xfs_ifork	*ifp = xfs_ifork_ptr(ip, XFS_COW_FORK);
 	unsigned int		resblks;
 	int			nmaps;
+	bool			isrt = XFS_IS_REALTIME_INODE(ip);
 	int			error;
 
 	resblks = XFS_EXTENTADD_SPACE_RES(mp, XFS_DATA_FORK);
@@ -779,7 +782,7 @@ xfs_reflink_end_cow_extent(
 		 * or not), unmap the extent and drop its refcount.
 		 */
 		xfs_bmap_unmap_extent(tp, ip, XFS_DATA_FORK, &data);
-		xfs_refcount_decrease_extent(tp, &data);
+		xfs_refcount_decrease_extent(tp, isrt, &data);
 		xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT,
 				-data.br_blockcount);
 	} else if (data.br_startblock == DELAYSTARTBLOCK) {
@@ -799,7 +802,8 @@ xfs_reflink_end_cow_extent(
 	}
 
 	/* Free the CoW orphan record. */
-	xfs_refcount_free_cow_extent(tp, del.br_startblock, del.br_blockcount);
+	xfs_refcount_free_cow_extent(tp, isrt, del.br_startblock,
+			del.br_blockcount);
 
 	/* Map the new blocks into the data fork. */
 	xfs_bmap_map_extent(tp, ip, XFS_DATA_FORK, &del);
@@ -1135,6 +1139,7 @@ xfs_reflink_remap_extent(
 	bool			quota_reserved = true;
 	bool			smap_real;
 	bool			dmap_written = xfs_bmap_is_written_extent(dmap);
+	bool			isrt = XFS_IS_REALTIME_INODE(ip);
 	int			iext_delta = 0;
 	int			nimaps;
 	int			error;
@@ -1264,7 +1269,7 @@ xfs_reflink_remap_extent(
 		 * or not), unmap the extent and drop its refcount.
 		 */
 		xfs_bmap_unmap_extent(tp, ip, XFS_DATA_FORK, &smap);
-		xfs_refcount_decrease_extent(tp, &smap);
+		xfs_refcount_decrease_extent(tp, isrt, &smap);
 		qdelta -= smap.br_blockcount;
 	} else if (smap.br_startblock == DELAYSTARTBLOCK) {
 		int		done;
@@ -1287,7 +1292,7 @@ xfs_reflink_remap_extent(
 	 * its refcount and map it into the file.
 	 */
 	if (dmap_written) {
-		xfs_refcount_increase_extent(tp, dmap);
+		xfs_refcount_increase_extent(tp, isrt, dmap);
 		xfs_bmap_map_extent(tp, ip, XFS_DATA_FORK, dmap);
 		qdelta += dmap->br_blockcount;
 	}

[PATCH 08/43] xfs: support recovering refcount intent items targetting realtime extents

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Now that we have reflink on the realtime device, refcount intent items
have to support remapping extents on the realtime volume.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/xfs_refcount_item.c |   14 +++++++++++---
 1 file changed, 11 insertions(+), 3 deletions(-)


diff --git a/fs/xfs/xfs_refcount_item.c b/fs/xfs/xfs_refcount_item.c
index 2086d40514d086..fe2d7aab8554fc 100644
--- a/fs/xfs/xfs_refcount_item.c
+++ b/fs/xfs/xfs_refcount_item.c
@@ -440,6 +440,7 @@ xfs_refcount_update_abort_intent(
 static inline bool
 xfs_cui_validate_phys(
 	struct xfs_mount		*mp,
+	bool				isrt,
 	struct xfs_phys_extent		*pmap)
 {
 	if (!xfs_has_reflink(mp))
@@ -458,6 +459,9 @@ xfs_cui_validate_phys(
 		return false;
 	}
 
+	if (isrt)
+		return xfs_verify_rtbext(mp, pmap->pe_startblock, pmap->pe_len);
+
 	return xfs_verify_fsbext(mp, pmap->pe_startblock, pmap->pe_len);
 }
 
@@ -465,6 +469,7 @@ static inline void
 xfs_cui_recover_work(
 	struct xfs_mount		*mp,
 	struct xfs_defer_pending	*dfp,
+	bool				isrt,
 	struct xfs_phys_extent		*pmap)
 {
 	struct xfs_refcount_intent	*ri;
@@ -475,7 +480,8 @@ xfs_cui_recover_work(
 	ri->ri_startblock = pmap->pe_startblock;
 	ri->ri_blockcount = pmap->pe_len;
 	ri->ri_group = xfs_group_intent_get(mp, pmap->pe_startblock,
-			XG_TYPE_AG);
+			isrt ? XG_TYPE_RTG : XG_TYPE_AG);
+	ri->ri_realtime = isrt;
 
 	xfs_defer_add_item(dfp, &ri->ri_list);
 }
@@ -494,6 +500,7 @@ xfs_refcount_recover_work(
 	struct xfs_cui_log_item		*cuip = CUI_ITEM(lip);
 	struct xfs_trans		*tp;
 	struct xfs_mount		*mp = lip->li_log->l_mp;
+	bool				isrt = xfs_cui_item_isrt(lip);
 	int				i;
 	int				error = 0;
 
@@ -503,7 +510,7 @@ xfs_refcount_recover_work(
 	 * just toss the CUI.
 	 */
 	for (i = 0; i < cuip->cui_format.cui_nextents; i++) {
-		if (!xfs_cui_validate_phys(mp,
+		if (!xfs_cui_validate_phys(mp, isrt,
 					&cuip->cui_format.cui_extents[i])) {
 			XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
 					&cuip->cui_format,
@@ -511,7 +518,8 @@ xfs_refcount_recover_work(
 			return -EFSCORRUPTED;
 		}
 
-		xfs_cui_recover_work(mp, dfp, &cuip->cui_format.cui_extents[i]);
+		xfs_cui_recover_work(mp, dfp, isrt,
+				&cuip->cui_format.cui_extents[i]);
 	}
 
 	/*

[PATCH 09/43] xfs: add realtime refcount btree block detection to log recovery

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Identify rt refcount btree blocks in the log correctly so that we can
validate them during log recovery.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/xfs_buf_item_recover.c |    4 ++++
 1 file changed, 4 insertions(+)


diff --git a/fs/xfs/xfs_buf_item_recover.c b/fs/xfs/xfs_buf_item_recover.c
index 4f2e4ea29e1f57..b05d5b81f642da 100644
--- a/fs/xfs/xfs_buf_item_recover.c
+++ b/fs/xfs/xfs_buf_item_recover.c
@@ -271,6 +271,9 @@ xlog_recover_validate_buf_type(
 		case XFS_REFC_CRC_MAGIC:
 			bp->b_ops = &xfs_refcountbt_buf_ops;
 			break;
+		case XFS_RTREFC_CRC_MAGIC:
+			bp->b_ops = &xfs_rtrefcountbt_buf_ops;
+			break;
 		default:
 			warnmsg = "Bad btree block magic!";
 			break;
@@ -859,6 +862,7 @@ xlog_recover_get_buf_lsn(
 		break;
 	}
 	case XFS_RTRMAP_CRC_MAGIC:
+	case XFS_RTREFC_CRC_MAGIC:
 	case XFS_BMAP_CRC_MAGIC:
 	case XFS_BMAP_MAGIC: {
 		struct xfs_btree_block *btb = blk;

[PATCH 10/43] xfs: add realtime refcount btree inode to metadata directory

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Add a metadir path to select the realtime refcount btree inode and load
it at mount time.  The rtrefcountbt 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" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_format.h           |    4 +++-
 fs/xfs/libxfs/xfs_inode_buf.c        |    5 +++++
 fs/xfs/libxfs/xfs_inode_fork.c       |    6 ++++++
 fs/xfs/libxfs/xfs_rtgroup.c          |    7 +++++++
 fs/xfs/libxfs/xfs_rtgroup.h          |    6 ++++++
 fs/xfs/libxfs/xfs_rtrefcount_btree.c |    6 +++---
 6 files changed, 30 insertions(+), 4 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index 17f7c0d1aaa452..b6828f92c131fb 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -858,6 +858,7 @@ enum xfs_metafile_type {
 	XFS_METAFILE_RTBITMAP,		/* rt bitmap */
 	XFS_METAFILE_RTSUMMARY,		/* rt summary */
 	XFS_METAFILE_RTRMAP,		/* rt rmap */
+	XFS_METAFILE_RTREFCOUNT,	/* rt refcount */
 
 	XFS_METAFILE_MAX
 } __packed;
@@ -870,7 +871,8 @@ enum xfs_metafile_type {
 	{ XFS_METAFILE_PRJQUOTA,	"prjquota" }, \
 	{ XFS_METAFILE_RTBITMAP,	"rtbitmap" }, \
 	{ XFS_METAFILE_RTSUMMARY,	"rtsummary" }, \
-	{ XFS_METAFILE_RTRMAP,		"rtrmap" }
+	{ XFS_METAFILE_RTRMAP,		"rtrmap" }, \
+	{ XFS_METAFILE_RTREFCOUNT,	"rtrefcount" }
 
 /*
  * On-disk inode structure.
diff --git a/fs/xfs/libxfs/xfs_inode_buf.c b/fs/xfs/libxfs/xfs_inode_buf.c
index 17cb91b89fcaa1..65eec8f60376d3 100644
--- a/fs/xfs/libxfs/xfs_inode_buf.c
+++ b/fs/xfs/libxfs/xfs_inode_buf.c
@@ -456,6 +456,11 @@ xfs_dinode_verify_fork(
 			if (!xfs_has_rmapbt(mp))
 				return __this_address;
 			break;
+		case XFS_METAFILE_RTREFCOUNT:
+			/* same comment about growfs and rmap inodes applies */
+			if (!xfs_has_reflink(mp))
+				return __this_address;
+			break;
 		default:
 			return __this_address;
 		}
diff --git a/fs/xfs/libxfs/xfs_inode_fork.c b/fs/xfs/libxfs/xfs_inode_fork.c
index d9b3c182cb400b..0c4bc12401a151 100644
--- a/fs/xfs/libxfs/xfs_inode_fork.c
+++ b/fs/xfs/libxfs/xfs_inode_fork.c
@@ -272,6 +272,9 @@ xfs_iformat_data_fork(
 			switch (ip->i_metatype) {
 			case XFS_METAFILE_RTRMAP:
 				return xfs_iformat_rtrmap(ip, dip);
+			case XFS_METAFILE_RTREFCOUNT:
+				ASSERT(0); /* to be implemented later */
+				return -EFSCORRUPTED;
 			default:
 				break;
 			}
@@ -620,6 +623,9 @@ xfs_iflush_fork(
 		case XFS_METAFILE_RTRMAP:
 			xfs_iflush_rtrmap(ip, dip);
 			break;
+		case XFS_METAFILE_RTREFCOUNT:
+			ASSERT(0); /* to be implemented later */
+			break;
 		default:
 			ASSERT(0);
 			break;
diff --git a/fs/xfs/libxfs/xfs_rtgroup.c b/fs/xfs/libxfs/xfs_rtgroup.c
index b7ed2d27d54553..6aebe9f484901f 100644
--- a/fs/xfs/libxfs/xfs_rtgroup.c
+++ b/fs/xfs/libxfs/xfs_rtgroup.c
@@ -367,6 +367,13 @@ static const struct xfs_rtginode_ops xfs_rtginode_ops[XFS_RTGI_MAX] = {
 		.enabled	= xfs_has_rmapbt,
 		.create		= xfs_rtrmapbt_create,
 	},
+	[XFS_RTGI_REFCOUNT] = {
+		.name		= "refcount",
+		.metafile_type	= XFS_METAFILE_RTREFCOUNT,
+		.fmt_mask	= 1U << XFS_DINODE_FMT_META_BTREE,
+		/* same comment about growfs and rmap inodes applies here */
+		.enabled	= xfs_has_reflink,
+	},
 };
 
 /* Return the shortname of this rtgroup inode. */
diff --git a/fs/xfs/libxfs/xfs_rtgroup.h b/fs/xfs/libxfs/xfs_rtgroup.h
index 733da7417c9cd7..385ea8e2f28b67 100644
--- a/fs/xfs/libxfs/xfs_rtgroup.h
+++ b/fs/xfs/libxfs/xfs_rtgroup.h
@@ -15,6 +15,7 @@ enum xfs_rtg_inodes {
 	XFS_RTGI_BITMAP,	/* allocation bitmap */
 	XFS_RTGI_SUMMARY,	/* allocation summary */
 	XFS_RTGI_RMAP,		/* rmap btree inode */
+	XFS_RTGI_REFCOUNT,	/* refcount btree inode */
 
 	XFS_RTGI_MAX,
 };
@@ -80,6 +81,11 @@ static inline struct xfs_inode *rtg_rmap(const struct xfs_rtgroup *rtg)
 	return rtg->rtg_inodes[XFS_RTGI_RMAP];
 }
 
+static inline struct xfs_inode *rtg_refcount(const struct xfs_rtgroup *rtg)
+{
+	return rtg->rtg_inodes[XFS_RTGI_REFCOUNT];
+}
+
 /* Passive rtgroup references */
 static inline struct xfs_rtgroup *
 xfs_rtgroup_get(
diff --git a/fs/xfs/libxfs/xfs_rtrefcount_btree.c b/fs/xfs/libxfs/xfs_rtrefcount_btree.c
index e30af941581651..ebbeab112d1412 100644
--- a/fs/xfs/libxfs/xfs_rtrefcount_btree.c
+++ b/fs/xfs/libxfs/xfs_rtrefcount_btree.c
@@ -26,6 +26,7 @@
 #include "xfs_extent_busy.h"
 #include "xfs_rtgroup.h"
 #include "xfs_rtbitmap.h"
+#include "xfs_metafile.h"
 
 static struct kmem_cache	*xfs_rtrefcountbt_cur_cache;
 
@@ -281,12 +282,10 @@ xfs_rtrefcountbt_init_cursor(
 	struct xfs_trans	*tp,
 	struct xfs_rtgroup	*rtg)
 {
-	struct xfs_inode	*ip = NULL;
+	struct xfs_inode	*ip = rtg_refcount(rtg);
 	struct xfs_mount	*mp = rtg_mount(rtg);
 	struct xfs_btree_cur	*cur;
 
-	return NULL; /* XXX */
-
 	xfs_assert_ilocked(ip, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL);
 
 	cur = xfs_btree_alloc_cursor(mp, tp, &xfs_rtrefcountbt_ops,
@@ -316,6 +315,7 @@ xfs_rtrefcountbt_commit_staged_btree(
 	int			flags = XFS_ILOG_CORE | XFS_ILOG_DBROOT;
 
 	ASSERT(cur->bc_flags & XFS_BTREE_STAGING);
+	ASSERT(ifake->if_fork->if_format == XFS_DINODE_FMT_META_BTREE);
 
 	/*
 	 * Free any resources hanging off the real fork, then shallow-copy the

[PATCH 11/43] xfs: add metadata reservations for realtime refcount btree

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Reserve some free blocks so that we will always have enough free blocks
in the data volume to handle expansion of the realtime refcount btree.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_rtrefcount_btree.c |   38 ++++++++++++++++++++++++++++++++++
 fs/xfs/libxfs/xfs_rtrefcount_btree.h |    4 ++++
 fs/xfs/xfs_rtalloc.c                 |    6 +++++
 3 files changed, 48 insertions(+)


diff --git a/fs/xfs/libxfs/xfs_rtrefcount_btree.c b/fs/xfs/libxfs/xfs_rtrefcount_btree.c
index ebbeab112d1412..ff72ed09e75f08 100644
--- a/fs/xfs/libxfs/xfs_rtrefcount_btree.c
+++ b/fs/xfs/libxfs/xfs_rtrefcount_btree.c
@@ -419,3 +419,41 @@ xfs_rtrefcountbt_compute_maxlevels(
 	/* Add one level to handle the inode root level. */
 	mp->m_rtrefc_maxlevels = min(d_maxlevels, r_maxlevels) + 1;
 }
+
+/* Calculate the rtrefcount btree size for some records. */
+unsigned long long
+xfs_rtrefcountbt_calc_size(
+	struct xfs_mount	*mp,
+	unsigned long long	len)
+{
+	return xfs_btree_calc_size(mp->m_rtrefc_mnr, len);
+}
+
+/*
+ * Calculate the maximum refcount btree size.
+ */
+static unsigned long long
+xfs_rtrefcountbt_max_size(
+	struct xfs_mount	*mp,
+	xfs_rtblock_t		rtblocks)
+{
+	/* Bail out if we're uninitialized, which can happen in mkfs. */
+	if (mp->m_rtrefc_mxr[0] == 0)
+		return 0;
+
+	return xfs_rtrefcountbt_calc_size(mp, rtblocks);
+}
+
+/*
+ * Figure out how many blocks to reserve and how many are used by this btree.
+ * We need enough space to hold one record for every rt extent in the rtgroup.
+ */
+xfs_filblks_t
+xfs_rtrefcountbt_calc_reserves(
+	struct xfs_mount	*mp)
+{
+	if (!xfs_has_rtreflink(mp))
+		return 0;
+
+	return xfs_rtrefcountbt_max_size(mp, mp->m_sb.sb_rgextents);
+}
diff --git a/fs/xfs/libxfs/xfs_rtrefcount_btree.h b/fs/xfs/libxfs/xfs_rtrefcount_btree.h
index b713b33818800c..3cd44590c9304c 100644
--- a/fs/xfs/libxfs/xfs_rtrefcount_btree.h
+++ b/fs/xfs/libxfs/xfs_rtrefcount_btree.h
@@ -67,4 +67,8 @@ unsigned int xfs_rtrefcountbt_maxlevels_ondisk(void);
 int __init xfs_rtrefcountbt_init_cur_cache(void);
 void xfs_rtrefcountbt_destroy_cur_cache(void);
 
+xfs_filblks_t xfs_rtrefcountbt_calc_reserves(struct xfs_mount *mp);
+unsigned long long xfs_rtrefcountbt_calc_size(struct xfs_mount *mp,
+		unsigned long long len);
+
 #endif	/* __XFS_RTREFCOUNT_BTREE_H__ */
diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c
index a69967f9d88ead..294aa0739be311 100644
--- a/fs/xfs/xfs_rtalloc.c
+++ b/fs/xfs/xfs_rtalloc.c
@@ -31,6 +31,7 @@
 #include "xfs_rtgroup.h"
 #include "xfs_error.h"
 #include "xfs_trace.h"
+#include "xfs_rtrefcount_btree.h"
 
 /*
  * Return whether there are any free extents in the size range given
@@ -1547,6 +1548,11 @@ xfs_rt_resv_init(
 		err2 = xfs_metafile_resv_init(rtg_rmap(rtg), ask);
 		if (err2 && !error)
 			error = err2;
+
+		ask = xfs_rtrefcountbt_calc_reserves(mp);
+		err2 = xfs_metafile_resv_init(rtg_refcount(rtg), ask);
+		if (err2 && !error)
+			error = err2;
 	}
 
 	return error;

[PATCH 12/43] xfs: wire up a new metafile type for the realtime refcount

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

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

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_format.h           |    8 +
 fs/xfs/libxfs/xfs_inode_fork.c       |    6 -
 fs/xfs/libxfs/xfs_ondisk.h           |    1 
 fs/xfs/libxfs/xfs_rtrefcount_btree.c |  264 ++++++++++++++++++++++++++++++++++
 fs/xfs/libxfs/xfs_rtrefcount_btree.h |  112 ++++++++++++++
 fs/xfs/xfs_inode_item_recover.c      |    4 +
 6 files changed, 392 insertions(+), 3 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index b6828f92c131fb..b1007fb661ba73 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -1805,6 +1805,14 @@ typedef __be32 xfs_refcount_ptr_t;
  */
 #define	XFS_RTREFC_CRC_MAGIC	0x52434e54	/* 'RCNT' */
 
+/*
+ * rt refcount root header, on-disk form only.
+ */
+struct xfs_rtrefcount_root {
+	__be16		bb_level;	/* 0 is a leaf */
+	__be16		bb_numrecs;	/* current # of data records */
+};
+
 /* inode-rooted btree pointer type */
 typedef __be64 xfs_rtrefcount_ptr_t;
 
diff --git a/fs/xfs/libxfs/xfs_inode_fork.c b/fs/xfs/libxfs/xfs_inode_fork.c
index 0c4bc12401a151..4f99b90add5526 100644
--- a/fs/xfs/libxfs/xfs_inode_fork.c
+++ b/fs/xfs/libxfs/xfs_inode_fork.c
@@ -28,6 +28,7 @@
 #include "xfs_health.h"
 #include "xfs_symlink_remote.h"
 #include "xfs_rtrmap_btree.h"
+#include "xfs_rtrefcount_btree.h"
 
 struct kmem_cache *xfs_ifork_cache;
 
@@ -273,8 +274,7 @@ xfs_iformat_data_fork(
 			case XFS_METAFILE_RTRMAP:
 				return xfs_iformat_rtrmap(ip, dip);
 			case XFS_METAFILE_RTREFCOUNT:
-				ASSERT(0); /* to be implemented later */
-				return -EFSCORRUPTED;
+				return xfs_iformat_rtrefcount(ip, dip);
 			default:
 				break;
 			}
@@ -624,7 +624,7 @@ xfs_iflush_fork(
 			xfs_iflush_rtrmap(ip, dip);
 			break;
 		case XFS_METAFILE_RTREFCOUNT:
-			ASSERT(0); /* to be implemented later */
+			xfs_iflush_rtrefcount(ip, dip);
 			break;
 		default:
 			ASSERT(0);
diff --git a/fs/xfs/libxfs/xfs_ondisk.h b/fs/xfs/libxfs/xfs_ondisk.h
index efb035050c009c..a85ecddaa48eed 100644
--- a/fs/xfs/libxfs/xfs_ondisk.h
+++ b/fs/xfs/libxfs/xfs_ondisk.h
@@ -86,6 +86,7 @@ xfs_check_ondisk_structs(void)
 	XFS_CHECK_STRUCT_SIZE(xfs_rtrmap_ptr_t,			8);
 	XFS_CHECK_STRUCT_SIZE(struct xfs_rtrmap_root,		4);
 	XFS_CHECK_STRUCT_SIZE(xfs_rtrefcount_ptr_t,		8);
+	XFS_CHECK_STRUCT_SIZE(struct xfs_rtrefcount_root,	4);
 
 	/*
 	 * m68k has problems with struct xfs_attr_leaf_name_remote, but we pad
diff --git a/fs/xfs/libxfs/xfs_rtrefcount_btree.c b/fs/xfs/libxfs/xfs_rtrefcount_btree.c
index ff72ed09e75f08..6a5bc7ea42fbe6 100644
--- a/fs/xfs/libxfs/xfs_rtrefcount_btree.c
+++ b/fs/xfs/libxfs/xfs_rtrefcount_btree.c
@@ -77,6 +77,41 @@ xfs_rtrefcountbt_get_maxrecs(
 	return cur->bc_mp->m_rtrefc_mxr[level != 0];
 }
 
+/*
+ * Calculate number of records in a realtime refcount btree inode root.
+ */
+unsigned int
+xfs_rtrefcountbt_droot_maxrecs(
+	unsigned int		blocklen,
+	bool			leaf)
+{
+	blocklen -= sizeof(struct xfs_rtrefcount_root);
+
+	if (leaf)
+		return blocklen / sizeof(struct xfs_refcount_rec);
+	return blocklen / (2 * sizeof(struct xfs_refcount_key) +
+			sizeof(xfs_rtrefcount_ptr_t));
+}
+
+/*
+ * Get the maximum records we could store in the on-disk format.
+ *
+ * For non-root nodes this is equivalent to xfs_rtrefcountbt_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_rtrefcountbt_get_maxrecs for the root node, too.
+ */
+STATIC int
+xfs_rtrefcountbt_get_dmaxrecs(
+	struct xfs_btree_cur	*cur,
+	int			level)
+{
+	if (level != cur->bc_nlevels - 1)
+		return cur->bc_mp->m_rtrefc_mxr[level != 0];
+	return xfs_rtrefcountbt_droot_maxrecs(cur->bc_ino.forksize, level == 0);
+}
+
 STATIC void
 xfs_rtrefcountbt_init_key_from_rec(
 	union xfs_btree_key		*key,
@@ -247,6 +282,87 @@ xfs_rtrefcountbt_keys_contiguous(
 				 be32_to_cpu(key2->refc.rc_startblock));
 }
 
+static inline void
+xfs_rtrefcountbt_move_ptrs(
+	struct xfs_mount	*mp,
+	struct xfs_btree_block	*broot,
+	short			old_size,
+	size_t			new_size,
+	unsigned int		numrecs)
+{
+	void			*dptr;
+	void			*sptr;
+
+	sptr = xfs_rtrefcount_broot_ptr_addr(mp, broot, 1, old_size);
+	dptr = xfs_rtrefcount_broot_ptr_addr(mp, broot, 1, new_size);
+	memmove(dptr, sptr, numrecs * sizeof(xfs_rtrefcount_ptr_t));
+}
+
+static struct xfs_btree_block *
+xfs_rtrefcountbt_broot_realloc(
+	struct xfs_btree_cur	*cur,
+	unsigned int		new_numrecs)
+{
+	struct xfs_mount	*mp = cur->bc_mp;
+	struct xfs_ifork	*ifp = xfs_btree_ifork_ptr(cur);
+	struct xfs_btree_block	*broot;
+	unsigned int		new_size;
+	unsigned int		old_size = ifp->if_broot_bytes;
+	const unsigned int	level = cur->bc_nlevels - 1;
+
+	new_size = xfs_rtrefcount_broot_space_calc(mp, level, new_numrecs);
+
+	/* Handle the nop case quietly. */
+	if (new_size == old_size)
+		return ifp->if_broot;
+
+	if (new_size > old_size) {
+		unsigned int	old_numrecs;
+
+		/*
+		 * If there wasn't any memory allocated before, just allocate
+		 * it now and get out.
+		 */
+		if (old_size == 0)
+			return xfs_broot_realloc(ifp, new_size);
+
+		/*
+		 * If there is already an existing if_broot, then we need to
+		 * realloc it and possibly move the node block pointers because
+		 * those are not butted up against the btree block header.
+		 */
+		old_numrecs = xfs_rtrefcountbt_maxrecs(mp, old_size, level);
+		broot = xfs_broot_realloc(ifp, new_size);
+		if (level > 0)
+			xfs_rtrefcountbt_move_ptrs(mp, broot, old_size,
+					new_size, old_numrecs);
+		goto out_broot;
+	}
+
+	/*
+	 * We're reducing numrecs.  If we're going all the way to zero, just
+	 * free the block.
+	 */
+	ASSERT(ifp->if_broot != NULL && old_size > 0);
+	if (new_size == 0)
+		return xfs_broot_realloc(ifp, 0);
+
+	/*
+	 * Shrink the btree root by possibly moving the rtrmapbt pointers,
+	 * since they are not butted up against the btree block header.  Then
+	 * reallocate broot.
+	 */
+	if (level > 0)
+		xfs_rtrefcountbt_move_ptrs(mp, ifp->if_broot, old_size,
+				new_size, new_numrecs);
+	broot = xfs_broot_realloc(ifp, new_size);
+
+out_broot:
+	ASSERT(xfs_rtrefcount_droot_space(broot) <=
+	       xfs_inode_fork_size(cur->bc_ino.ip, cur->bc_ino.whichfork));
+	return broot;
+}
+
 const struct xfs_btree_ops xfs_rtrefcountbt_ops = {
 	.name			= "rtrefcount",
 	.type			= XFS_BTREE_TYPE_INODE,
@@ -264,6 +380,7 @@ const struct xfs_btree_ops xfs_rtrefcountbt_ops = {
 	.free_block		= xfs_btree_free_metafile_block,
 	.get_minrecs		= xfs_rtrefcountbt_get_minrecs,
 	.get_maxrecs		= xfs_rtrefcountbt_get_maxrecs,
+	.get_dmaxrecs		= xfs_rtrefcountbt_get_dmaxrecs,
 	.init_key_from_rec	= xfs_rtrefcountbt_init_key_from_rec,
 	.init_high_key_from_rec	= xfs_rtrefcountbt_init_high_key_from_rec,
 	.init_rec_from_cur	= xfs_rtrefcountbt_init_rec_from_cur,
@@ -274,6 +391,7 @@ const struct xfs_btree_ops xfs_rtrefcountbt_ops = {
 	.keys_inorder		= xfs_rtrefcountbt_keys_inorder,
 	.recs_inorder		= xfs_rtrefcountbt_recs_inorder,
 	.keys_contiguous	= xfs_rtrefcountbt_keys_contiguous,
+	.broot_realloc		= xfs_rtrefcountbt_broot_realloc,
 };
 
 /* Allocate a new rt refcount btree cursor. */
@@ -457,3 +575,149 @@ xfs_rtrefcountbt_calc_reserves(
 
 	return xfs_rtrefcountbt_max_size(mp, mp->m_sb.sb_rgextents);
 }
+
+/*
+ * Convert on-disk form of btree root to in-memory form.
+ */
+STATIC void
+xfs_rtrefcountbt_from_disk(
+	struct xfs_inode		*ip,
+	struct xfs_rtrefcount_root	*dblock,
+	int				dblocklen,
+	struct xfs_btree_block		*rblock)
+{
+	struct xfs_mount		*mp = ip->i_mount;
+	struct xfs_refcount_key	*fkp;
+	__be64				*fpp;
+	struct xfs_refcount_key	*tkp;
+	__be64				*tpp;
+	struct xfs_refcount_rec	*frp;
+	struct xfs_refcount_rec	*trp;
+	unsigned int			numrecs;
+	unsigned int			maxrecs;
+	unsigned int			rblocklen;
+
+	rblocklen = xfs_rtrefcount_broot_space(mp, dblock);
+
+	xfs_btree_init_block(mp, rblock, &xfs_rtrefcountbt_ops, 0, 0,
+			ip->i_ino);
+
+	rblock->bb_level = dblock->bb_level;
+	rblock->bb_numrecs = dblock->bb_numrecs;
+
+	if (be16_to_cpu(rblock->bb_level) > 0) {
+		maxrecs = xfs_rtrefcountbt_droot_maxrecs(dblocklen, false);
+		fkp = xfs_rtrefcount_droot_key_addr(dblock, 1);
+		tkp = xfs_rtrefcount_key_addr(rblock, 1);
+		fpp = xfs_rtrefcount_droot_ptr_addr(dblock, 1, maxrecs);
+		tpp = xfs_rtrefcount_broot_ptr_addr(mp, rblock, 1, rblocklen);
+		numrecs = be16_to_cpu(dblock->bb_numrecs);
+		memcpy(tkp, fkp, 2 * sizeof(*fkp) * numrecs);
+		memcpy(tpp, fpp, sizeof(*fpp) * numrecs);
+	} else {
+		frp = xfs_rtrefcount_droot_rec_addr(dblock, 1);
+		trp = xfs_rtrefcount_rec_addr(rblock, 1);
+		numrecs = be16_to_cpu(dblock->bb_numrecs);
+		memcpy(trp, frp, sizeof(*frp) * numrecs);
+	}
+}
+
+/* Load a realtime reference count btree root in from disk. */
+int
+xfs_iformat_rtrefcount(
+	struct xfs_inode	*ip,
+	struct xfs_dinode	*dip)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_rtrefcount_root *dfp = XFS_DFORK_PTR(dip, XFS_DATA_FORK);
+	struct xfs_btree_block	*broot;
+	unsigned int		numrecs;
+	unsigned int		level;
+	int			dsize;
+
+	/*
+	 * growfs must create the rtrefcount inodes before adding a realtime
+	 * volume to the filesystem, so we cannot use the rtrefcount predicate
+	 * here.
+	 */
+	if (!xfs_has_reflink(ip->i_mount))
+		return -EFSCORRUPTED;
+
+	dsize = XFS_DFORK_SIZE(dip, mp, XFS_DATA_FORK);
+	numrecs = be16_to_cpu(dfp->bb_numrecs);
+	level = be16_to_cpu(dfp->bb_level);
+
+	if (level > mp->m_rtrefc_maxlevels ||
+	    xfs_rtrefcount_droot_space_calc(level, numrecs) > dsize)
+		return -EFSCORRUPTED;
+
+	broot = xfs_broot_alloc(xfs_ifork_ptr(ip, XFS_DATA_FORK),
+			xfs_rtrefcount_broot_space_calc(mp, level, numrecs));
+	if (broot)
+		xfs_rtrefcountbt_from_disk(ip, dfp, dsize, broot);
+	return 0;
+}
+
+/*
+ * Convert in-memory form of btree root to on-disk form.
+ */
+void
+xfs_rtrefcountbt_to_disk(
+	struct xfs_mount		*mp,
+	struct xfs_btree_block		*rblock,
+	int				rblocklen,
+	struct xfs_rtrefcount_root	*dblock,
+	int				dblocklen)
+{
+	struct xfs_refcount_key	*fkp;
+	__be64				*fpp;
+	struct xfs_refcount_key	*tkp;
+	__be64				*tpp;
+	struct xfs_refcount_rec	*frp;
+	struct xfs_refcount_rec	*trp;
+	unsigned int			maxrecs;
+	unsigned int			numrecs;
+
+	ASSERT(rblock->bb_magic == cpu_to_be32(XFS_RTREFC_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) {
+		maxrecs = xfs_rtrefcountbt_droot_maxrecs(dblocklen, false);
+		fkp = xfs_rtrefcount_key_addr(rblock, 1);
+		tkp = xfs_rtrefcount_droot_key_addr(dblock, 1);
+		fpp = xfs_rtrefcount_broot_ptr_addr(mp, rblock, 1, rblocklen);
+		tpp = xfs_rtrefcount_droot_ptr_addr(dblock, 1, maxrecs);
+		numrecs = be16_to_cpu(rblock->bb_numrecs);
+		memcpy(tkp, fkp, 2 * sizeof(*fkp) * numrecs);
+		memcpy(tpp, fpp, sizeof(*fpp) * numrecs);
+	} else {
+		frp = xfs_rtrefcount_rec_addr(rblock, 1);
+		trp = xfs_rtrefcount_droot_rec_addr(dblock, 1);
+		numrecs = be16_to_cpu(rblock->bb_numrecs);
+		memcpy(trp, frp, sizeof(*frp) * numrecs);
+	}
+}
+
+/* Flush a realtime reference count btree root out to disk. */
+void
+xfs_iflush_rtrefcount(
+	struct xfs_inode	*ip,
+	struct xfs_dinode	*dip)
+{
+	struct xfs_ifork	*ifp = xfs_ifork_ptr(ip, XFS_DATA_FORK);
+	struct xfs_rtrefcount_root *dfp = XFS_DFORK_PTR(dip, XFS_DATA_FORK);
+
+	ASSERT(ifp->if_broot != NULL);
+	ASSERT(ifp->if_broot_bytes > 0);
+	ASSERT(xfs_rtrefcount_droot_space(ifp->if_broot) <=
+			xfs_inode_fork_size(ip, XFS_DATA_FORK));
+	xfs_rtrefcountbt_to_disk(ip->i_mount, ifp->if_broot,
+			ifp->if_broot_bytes, dfp,
+			XFS_DFORK_SIZE(dip, ip->i_mount, XFS_DATA_FORK));
+}
diff --git a/fs/xfs/libxfs/xfs_rtrefcount_btree.h b/fs/xfs/libxfs/xfs_rtrefcount_btree.h
index 3cd44590c9304c..e558a10c4744ad 100644
--- a/fs/xfs/libxfs/xfs_rtrefcount_btree.h
+++ b/fs/xfs/libxfs/xfs_rtrefcount_btree.h
@@ -25,6 +25,7 @@ void xfs_rtrefcountbt_commit_staged_btree(struct xfs_btree_cur *cur,
 unsigned int xfs_rtrefcountbt_maxrecs(struct xfs_mount *mp,
 		unsigned int blocklen, bool leaf);
 void xfs_rtrefcountbt_compute_maxlevels(struct xfs_mount *mp);
+unsigned int xfs_rtrefcountbt_droot_maxrecs(unsigned int blocklen, bool leaf);
 
 /*
  * Addresses of records, keys, and pointers within an incore rtrefcountbt block.
@@ -71,4 +72,115 @@ xfs_filblks_t xfs_rtrefcountbt_calc_reserves(struct xfs_mount *mp);
 unsigned long long xfs_rtrefcountbt_calc_size(struct xfs_mount *mp,
 		unsigned long long len);
 
+/* Addresses of key, pointers, and records within an ondisk rtrefcount block. */
+
+static inline struct xfs_refcount_rec *
+xfs_rtrefcount_droot_rec_addr(
+	struct xfs_rtrefcount_root	*block,
+	unsigned int			index)
+{
+	return (struct xfs_refcount_rec *)
+		((char *)(block + 1) +
+		 (index - 1) * sizeof(struct xfs_refcount_rec));
+}
+
+static inline struct xfs_refcount_key *
+xfs_rtrefcount_droot_key_addr(
+	struct xfs_rtrefcount_root	*block,
+	unsigned int			index)
+{
+	return (struct xfs_refcount_key *)
+		((char *)(block + 1) +
+		 (index - 1) * sizeof(struct xfs_refcount_key));
+}
+
+static inline xfs_rtrefcount_ptr_t *
+xfs_rtrefcount_droot_ptr_addr(
+	struct xfs_rtrefcount_root	*block,
+	unsigned int			index,
+	unsigned int			maxrecs)
+{
+	return (xfs_rtrefcount_ptr_t *)
+		((char *)(block + 1) +
+		 maxrecs * sizeof(struct xfs_refcount_key) +
+		 (index - 1) * sizeof(xfs_rtrefcount_ptr_t));
+}
+
+/*
+ * Address of pointers within the incore btree root.
+ *
+ * These are to be used when we know the size of the block and
+ * we don't have a cursor.
+ */
+static inline xfs_rtrefcount_ptr_t *
+xfs_rtrefcount_broot_ptr_addr(
+	struct xfs_mount	*mp,
+	struct xfs_btree_block	*bb,
+	unsigned int		index,
+	unsigned int		block_size)
+{
+	return xfs_rtrefcount_ptr_addr(bb, index,
+			xfs_rtrefcountbt_maxrecs(mp, block_size, false));
+}
+
+/*
+ * Compute the space required for the incore btree root containing the given
+ * number of records.
+ */
+static inline size_t
+xfs_rtrefcount_broot_space_calc(
+	struct xfs_mount	*mp,
+	unsigned int		level,
+	unsigned int		nrecs)
+{
+	size_t			sz = XFS_RTREFCOUNT_BLOCK_LEN;
+
+	if (level > 0)
+		return sz + nrecs * (sizeof(struct xfs_refcount_key) +
+				     sizeof(xfs_rtrefcount_ptr_t));
+	return sz + nrecs * sizeof(struct xfs_refcount_rec);
+}
+
+/*
+ * Compute the space required for the incore btree root given the ondisk
+ * btree root block.
+ */
+static inline size_t
+xfs_rtrefcount_broot_space(struct xfs_mount *mp, struct xfs_rtrefcount_root *bb)
+{
+	return xfs_rtrefcount_broot_space_calc(mp, be16_to_cpu(bb->bb_level),
+			be16_to_cpu(bb->bb_numrecs));
+}
+
+/* Compute the space required for the ondisk root block. */
+static inline size_t
+xfs_rtrefcount_droot_space_calc(
+	unsigned int		level,
+	unsigned int		nrecs)
+{
+	size_t			sz = sizeof(struct xfs_rtrefcount_root);
+
+	if (level > 0)
+		return sz + nrecs * (sizeof(struct xfs_refcount_key) +
+				     sizeof(xfs_rtrefcount_ptr_t));
+	return sz + nrecs * sizeof(struct xfs_refcount_rec);
+}
+
+/*
+ * Compute the space required for the ondisk root block given an incore root
+ * block.
+ */
+static inline size_t
+xfs_rtrefcount_droot_space(struct xfs_btree_block *bb)
+{
+	return xfs_rtrefcount_droot_space_calc(be16_to_cpu(bb->bb_level),
+			be16_to_cpu(bb->bb_numrecs));
+}
+
+int xfs_iformat_rtrefcount(struct xfs_inode *ip, struct xfs_dinode *dip);
+void xfs_rtrefcountbt_to_disk(struct xfs_mount *mp,
+		struct xfs_btree_block *rblock, int rblocklen,
+		struct xfs_rtrefcount_root *dblock, int dblocklen);
+void xfs_iflush_rtrefcount(struct xfs_inode *ip, struct xfs_dinode *dip);
+
 #endif	/* __XFS_RTREFCOUNT_BTREE_H__ */
diff --git a/fs/xfs/xfs_inode_item_recover.c b/fs/xfs/xfs_inode_item_recover.c
index 5de1d3563b7686..f3bfb814378c06 100644
--- a/fs/xfs/xfs_inode_item_recover.c
+++ b/fs/xfs/xfs_inode_item_recover.c
@@ -23,6 +23,7 @@
 #include "xfs_icache.h"
 #include "xfs_bmap_btree.h"
 #include "xfs_rtrmap_btree.h"
+#include "xfs_rtrefcount_btree.h"
 
 STATIC void
 xlog_recover_inode_ra_pass2(
@@ -286,6 +287,9 @@ xlog_recover_inode_dbroot(
 		case XFS_METAFILE_RTRMAP:
 			xfs_rtrmapbt_to_disk(mp, src, len, dfork, dsize);
 			return 0;
+		case XFS_METAFILE_RTREFCOUNT:
+			xfs_rtrefcountbt_to_disk(mp, src, len, dfork, dsize);
+			return 0;
 		default:
 			ASSERT(0);
 			return -EFSCORRUPTED;

[PATCH 13/43] xfs: refactor xfs_reflink_find_shared

[Darrick J. Wong]

From: Christoph Hellwig <hch@lst.de>

Move lookup of the perag structure from the callers into the helpers,
and return the offset into the extent of the shared region instead of
the block number that needs post-processing.  This prepares the
callsites for the creation of an rt-specific variant in the next patch.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: "Darrick J. Wong" <djwong@kernel.org>
[djwong: port to the middle of the rtreflink series for cleanliness]
Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
---
 fs/xfs/xfs_reflink.c |  110 ++++++++++++++++++++++----------------------------
 fs/xfs/xfs_reflink.h |    2 -
 2 files changed, 50 insertions(+), 62 deletions(-)


diff --git a/fs/xfs/xfs_reflink.c b/fs/xfs/xfs_reflink.c
index 02457e176c4252..71b4743ffb7741 100644
--- a/fs/xfs/xfs_reflink.c
+++ b/fs/xfs/xfs_reflink.c
@@ -120,38 +120,46 @@
  */
 
 /*
- * Given an AG extent, find the lowest-numbered run of shared blocks
- * within that range and return the range in fbno/flen.  If
- * find_end_of_shared is true, return the longest contiguous extent of
- * shared blocks.  If there are no shared extents, fbno and flen will
- * be set to NULLAGBLOCK and 0, respectively.
+ * Given a file mapping for the data device, find the lowest-numbered run of
+ * shared blocks within that mapping and return it in shared_offset/shared_len.
+ * The offset is relative to the start of irec.
+ *
+ * If find_end_of_shared is true, return the longest contiguous extent of shared
+ * blocks.  If there are no shared extents, shared_offset and shared_len will be
+ * set to 0;
  */
 static int
 xfs_reflink_find_shared(
-	struct xfs_perag	*pag,
+	struct xfs_mount	*mp,
 	struct xfs_trans	*tp,
-	xfs_agblock_t		agbno,
-	xfs_extlen_t		aglen,
-	xfs_agblock_t		*fbno,
-	xfs_extlen_t		*flen,
+	const struct xfs_bmbt_irec *irec,
+	xfs_extlen_t		*shared_offset,
+	xfs_extlen_t		*shared_len,
 	bool			find_end_of_shared)
 {
 	struct xfs_buf		*agbp;
+	struct xfs_perag	*pag;
 	struct xfs_btree_cur	*cur;
 	int			error;
+	xfs_agblock_t		orig_bno, found_bno;
+
+	pag = xfs_perag_get(mp, XFS_FSB_TO_AGNO(mp, irec->br_startblock));
+	orig_bno = XFS_FSB_TO_AGBNO(mp, irec->br_startblock);
 
 	error = xfs_alloc_read_agf(pag, tp, 0, &agbp);
 	if (error)
-		return error;
-
-	cur = xfs_refcountbt_init_cursor(pag_mount(pag), tp, agbp, pag);
-
-	error = xfs_refcount_find_shared(cur, agbno, aglen, fbno, flen,
-			find_end_of_shared);
+		goto out;
 
+	cur = xfs_refcountbt_init_cursor(mp, tp, agbp, pag);
+	error = xfs_refcount_find_shared(cur, orig_bno, irec->br_blockcount,
+			&found_bno, shared_len, find_end_of_shared);
 	xfs_btree_del_cursor(cur, error);
-
 	xfs_trans_brelse(tp, agbp);
+
+	if (!error && *shared_len)
+		*shared_offset = found_bno - orig_bno;
+out:
+	xfs_perag_put(pag);
 	return error;
 }
 
@@ -172,11 +180,7 @@ xfs_reflink_trim_around_shared(
 	bool			*shared)
 {
 	struct xfs_mount	*mp = ip->i_mount;
-	struct xfs_perag	*pag;
-	xfs_agblock_t		agbno;
-	xfs_extlen_t		aglen;
-	xfs_agblock_t		fbno;
-	xfs_extlen_t		flen;
+	xfs_extlen_t		shared_offset, shared_len;
 	int			error = 0;
 
 	/* Holes, unwritten, and delalloc extents cannot be shared */
@@ -187,41 +191,33 @@ xfs_reflink_trim_around_shared(
 
 	trace_xfs_reflink_trim_around_shared(ip, irec);
 
-	pag = xfs_perag_get(mp, XFS_FSB_TO_AGNO(mp, irec->br_startblock));
-	agbno = XFS_FSB_TO_AGBNO(mp, irec->br_startblock);
-	aglen = irec->br_blockcount;
-
-	error = xfs_reflink_find_shared(pag, NULL, agbno, aglen, &fbno, &flen,
-			true);
-	xfs_perag_put(pag);
+	error = xfs_reflink_find_shared(mp, NULL, irec, &shared_offset,
+			&shared_len, true);
 	if (error)
 		return error;
 
-	*shared = false;
-	if (fbno == NULLAGBLOCK) {
+	if (!shared_len) {
 		/* No shared blocks at all. */
-		return 0;
-	}
-
-	if (fbno == agbno) {
+		*shared = false;
+	} else if (!shared_offset) {
 		/*
-		 * The start of this extent is shared.  Truncate the
-		 * mapping at the end of the shared region so that a
-		 * subsequent iteration starts at the start of the
-		 * unshared region.
+		 * The start of this mapping points to shared space.  Truncate
+		 * the mapping at the end of the shared region so that a
+		 * subsequent iteration starts at the start of the unshared
+		 * region.
 		 */
-		irec->br_blockcount = flen;
+		irec->br_blockcount = shared_len;
 		*shared = true;
-		return 0;
+	} else {
+		/*
+		 * There's a shared region that doesn't start at the beginning
+		 * of the mapping.  Truncate the mapping at the start of the
+		 * shared extent so that a subsequent iteration starts at the
+		 * start of the shared region.
+		 */
+		irec->br_blockcount = shared_offset;
+		*shared = false;
 	}
-
-	/*
-	 * There's a shared extent midway through this extent.
-	 * Truncate the mapping at the start of the shared
-	 * extent so that a subsequent iteration starts at the
-	 * start of the shared region.
-	 */
-	irec->br_blockcount = fbno - agbno;
 	return 0;
 }
 
@@ -1552,27 +1548,19 @@ xfs_reflink_inode_has_shared_extents(
 	*has_shared = false;
 	found = xfs_iext_lookup_extent(ip, ifp, 0, &icur, &got);
 	while (found) {
-		struct xfs_perag	*pag;
-		xfs_agblock_t		agbno;
-		xfs_extlen_t		aglen;
-		xfs_agblock_t		rbno;
-		xfs_extlen_t		rlen;
+		xfs_extlen_t		shared_offset, shared_len;
 
 		if (isnullstartblock(got.br_startblock) ||
 		    got.br_state != XFS_EXT_NORM)
 			goto next;
 
-		pag = xfs_perag_get(mp, XFS_FSB_TO_AGNO(mp, got.br_startblock));
-		agbno = XFS_FSB_TO_AGBNO(mp, got.br_startblock);
-		aglen = got.br_blockcount;
-		error = xfs_reflink_find_shared(pag, tp, agbno, aglen,
-				&rbno, &rlen, false);
-		xfs_perag_put(pag);
+		error = xfs_reflink_find_shared(mp, tp, &got, &shared_offset,
+				&shared_len, false);
 		if (error)
 			return error;
 
 		/* Is there still a shared block here? */
-		if (rbno != NULLAGBLOCK) {
+		if (shared_len) {
 			*has_shared = true;
 			return 0;
 		}
diff --git a/fs/xfs/xfs_reflink.h b/fs/xfs/xfs_reflink.h
index 4a58e4533671c0..3bfd7ab9e1148a 100644
--- a/fs/xfs/xfs_reflink.h
+++ b/fs/xfs/xfs_reflink.h
@@ -25,7 +25,7 @@ xfs_can_free_cowblocks(struct xfs_inode *ip)
 	return true;
 }
 
-extern int xfs_reflink_trim_around_shared(struct xfs_inode *ip,
+int xfs_reflink_trim_around_shared(struct xfs_inode *ip,
 		struct xfs_bmbt_irec *irec, bool *shared);
 int xfs_bmap_trim_cow(struct xfs_inode *ip, struct xfs_bmbt_irec *imap,
 		bool *shared);

[PATCH 14/43] xfs: wire up realtime refcount btree cursors

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Wire up realtime refcount btree cursors wherever they're needed
throughout the code base.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_btree.h    |    2 -
 fs/xfs/libxfs/xfs_refcount.c |  100 +++++++++++++++++++++++++++++++++++++++++-
 fs/xfs/libxfs/xfs_rtgroup.c  |    9 ++++
 fs/xfs/libxfs/xfs_rtgroup.h  |    5 ++
 fs/xfs/xfs_fsmap.c           |   25 +++++------
 fs/xfs/xfs_reflink.c         |   66 ++++++++++++++++++++++++++--
 6 files changed, 187 insertions(+), 20 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h
index dbc047b2fb2cf5..355b304696e6c3 100644
--- a/fs/xfs/libxfs/xfs_btree.h
+++ b/fs/xfs/libxfs/xfs_btree.h
@@ -297,7 +297,7 @@ struct xfs_btree_cur
 		struct {
 			unsigned int	nr_ops;		/* # record updates */
 			unsigned int	shape_changes;	/* # of extent splits */
-		} bc_refc;	/* refcountbt */
+		} bc_refc;	/* refcountbt/rtrefcountbt */
 	};
 
 	/* Must be at the end of the struct! */
diff --git a/fs/xfs/libxfs/xfs_refcount.c b/fs/xfs/libxfs/xfs_refcount.c
index 8007d15856252b..11bff098db2dbb 100644
--- a/fs/xfs/libxfs/xfs_refcount.c
+++ b/fs/xfs/libxfs/xfs_refcount.c
@@ -27,6 +27,7 @@
 #include "xfs_refcount_item.h"
 #include "xfs_rtgroup.h"
 #include "xfs_rtalloc.h"
+#include "xfs_rtrefcount_btree.h"
 
 struct kmem_cache	*xfs_refcount_intent_cache;
 
@@ -1462,6 +1463,32 @@ xfs_refcount_finish_one(
 	return error;
 }
 
+/*
+ * Set up a continuation a deferred rtrefcount operation by updating the
+ * intent.  Checks to make sure we're not going to run off the end of the
+ * rtgroup.
+ */
+static inline int
+xfs_rtrefcount_continue_op(
+	struct xfs_btree_cur		*cur,
+	struct xfs_refcount_intent	*ri,
+	xfs_agblock_t			new_agbno)
+{
+	struct xfs_mount		*mp = cur->bc_mp;
+	struct xfs_rtgroup		*rtg = to_rtg(ri->ri_group);
+
+	if (XFS_IS_CORRUPT(mp, !xfs_verify_rgbext(rtg, new_agbno,
+					ri->ri_blockcount))) {
+		xfs_btree_mark_sick(cur);
+		return -EFSCORRUPTED;
+	}
+
+	ri->ri_startblock = xfs_rgbno_to_rtb(rtg, new_agbno);
+
+	ASSERT(xfs_verify_rtbext(mp, ri->ri_startblock, ri->ri_blockcount));
+	return 0;
+}
+
 /*
  * Process one of the deferred realtime refcount operations.  We pass back the
  * btree cursor to maintain our lock on the btree between calls.
@@ -1472,8 +1499,77 @@ xfs_rtrefcount_finish_one(
 	struct xfs_refcount_intent	*ri,
 	struct xfs_btree_cur		**pcur)
 {
-	ASSERT(0);
-	return -EFSCORRUPTED;
+	struct xfs_mount		*mp = tp->t_mountp;
+	struct xfs_rtgroup		*rtg = to_rtg(ri->ri_group);
+	struct xfs_btree_cur		*rcur = *pcur;
+	int				error = 0;
+	xfs_rgblock_t			bno;
+	unsigned long			nr_ops = 0;
+	int				shape_changes = 0;
+
+	bno = xfs_rtb_to_rgbno(mp, ri->ri_startblock);
+
+	trace_xfs_refcount_deferred(mp, ri);
+
+	if (XFS_TEST_ERROR(false, mp, XFS_ERRTAG_REFCOUNT_FINISH_ONE))
+		return -EIO;
+
+	/*
+	 * If we haven't gotten a cursor or the cursor AG doesn't match
+	 * the startblock, get one now.
+	 */
+	if (rcur != NULL && rcur->bc_group != ri->ri_group) {
+		nr_ops = rcur->bc_refc.nr_ops;
+		shape_changes = rcur->bc_refc.shape_changes;
+		xfs_btree_del_cursor(rcur, 0);
+		rcur = NULL;
+		*pcur = NULL;
+	}
+	if (rcur == NULL) {
+		xfs_rtgroup_lock(rtg, XFS_RTGLOCK_REFCOUNT);
+		xfs_rtgroup_trans_join(tp, rtg, XFS_RTGLOCK_REFCOUNT);
+		*pcur = rcur = xfs_rtrefcountbt_init_cursor(tp, rtg);
+
+		rcur->bc_refc.nr_ops = nr_ops;
+		rcur->bc_refc.shape_changes = shape_changes;
+	}
+
+	switch (ri->ri_type) {
+	case XFS_REFCOUNT_INCREASE:
+		error = xfs_refcount_adjust(rcur, &bno, &ri->ri_blockcount,
+				XFS_REFCOUNT_ADJUST_INCREASE);
+		if (error)
+			return error;
+		if (ri->ri_blockcount > 0)
+			error = xfs_rtrefcount_continue_op(rcur, ri, bno);
+		break;
+	case XFS_REFCOUNT_DECREASE:
+		error = xfs_refcount_adjust(rcur, &bno, &ri->ri_blockcount,
+				XFS_REFCOUNT_ADJUST_DECREASE);
+		if (error)
+			return error;
+		if (ri->ri_blockcount > 0)
+			error = xfs_rtrefcount_continue_op(rcur, ri, bno);
+		break;
+	case XFS_REFCOUNT_ALLOC_COW:
+		error = __xfs_refcount_cow_alloc(rcur, bno, ri->ri_blockcount);
+		if (error)
+			return error;
+		ri->ri_blockcount = 0;
+		break;
+	case XFS_REFCOUNT_FREE_COW:
+		error = __xfs_refcount_cow_free(rcur, bno, ri->ri_blockcount);
+		if (error)
+			return error;
+		ri->ri_blockcount = 0;
+		break;
+	default:
+		ASSERT(0);
+		return -EFSCORRUPTED;
+	}
+	if (!error && ri->ri_blockcount > 0)
+		trace_xfs_refcount_finish_one_leftover(mp, ri);
+	return error;
 }
 
 /*
diff --git a/fs/xfs/libxfs/xfs_rtgroup.c b/fs/xfs/libxfs/xfs_rtgroup.c
index 6aebe9f484901f..d5ecc2f6c5c202 100644
--- a/fs/xfs/libxfs/xfs_rtgroup.c
+++ b/fs/xfs/libxfs/xfs_rtgroup.c
@@ -206,6 +206,9 @@ xfs_rtgroup_lock(
 
 	if ((rtglock_flags & XFS_RTGLOCK_RMAP) && rtg_rmap(rtg))
 		xfs_ilock(rtg_rmap(rtg), XFS_ILOCK_EXCL);
+
+	if ((rtglock_flags & XFS_RTGLOCK_REFCOUNT) && rtg_refcount(rtg))
+		xfs_ilock(rtg_refcount(rtg), XFS_ILOCK_EXCL);
 }
 
 /* Unlock metadata inodes associated with this rt group. */
@@ -218,6 +221,9 @@ xfs_rtgroup_unlock(
 	ASSERT(!(rtglock_flags & XFS_RTGLOCK_BITMAP_SHARED) ||
 	       !(rtglock_flags & XFS_RTGLOCK_BITMAP));
 
+	if ((rtglock_flags & XFS_RTGLOCK_REFCOUNT) && rtg_refcount(rtg))
+		xfs_iunlock(rtg_refcount(rtg), XFS_ILOCK_EXCL);
+
 	if ((rtglock_flags & XFS_RTGLOCK_RMAP) && rtg_rmap(rtg))
 		xfs_iunlock(rtg_rmap(rtg), XFS_ILOCK_EXCL);
 
@@ -249,6 +255,9 @@ xfs_rtgroup_trans_join(
 
 	if ((rtglock_flags & XFS_RTGLOCK_RMAP) && rtg_rmap(rtg))
 		xfs_trans_ijoin(tp, rtg_rmap(rtg), XFS_ILOCK_EXCL);
+
+	if ((rtglock_flags & XFS_RTGLOCK_REFCOUNT) && rtg_refcount(rtg))
+		xfs_trans_ijoin(tp, rtg_refcount(rtg), XFS_ILOCK_EXCL);
 }
 
 /* Retrieve rt group geometry. */
diff --git a/fs/xfs/libxfs/xfs_rtgroup.h b/fs/xfs/libxfs/xfs_rtgroup.h
index 385ea8e2f28b67..de4eeb381fc9fc 100644
--- a/fs/xfs/libxfs/xfs_rtgroup.h
+++ b/fs/xfs/libxfs/xfs_rtgroup.h
@@ -273,10 +273,13 @@ int xfs_update_last_rtgroup_size(struct xfs_mount *mp,
 #define XFS_RTGLOCK_BITMAP_SHARED	(1U << 1)
 /* Lock the rt rmap inode in exclusive mode */
 #define XFS_RTGLOCK_RMAP		(1U << 2)
+/* Lock the rt refcount inode in exclusive mode */
+#define XFS_RTGLOCK_REFCOUNT		(1U << 3)
 
 #define XFS_RTGLOCK_ALL_FLAGS	(XFS_RTGLOCK_BITMAP | \
 				 XFS_RTGLOCK_BITMAP_SHARED | \
-				 XFS_RTGLOCK_RMAP)
+				 XFS_RTGLOCK_RMAP | \
+				 XFS_RTGLOCK_REFCOUNT)
 
 void xfs_rtgroup_lock(struct xfs_rtgroup *rtg, unsigned int rtglock_flags);
 void xfs_rtgroup_unlock(struct xfs_rtgroup *rtg, unsigned int rtglock_flags);
diff --git a/fs/xfs/xfs_fsmap.c b/fs/xfs/xfs_fsmap.c
index 3e3ef16f65a335..1dbd2d75f7ae3e 100644
--- a/fs/xfs/xfs_fsmap.c
+++ b/fs/xfs/xfs_fsmap.c
@@ -27,6 +27,7 @@
 #include "xfs_ag.h"
 #include "xfs_rtgroup.h"
 #include "xfs_rtrmap_btree.h"
+#include "xfs_rtrefcount_btree.h"
 
 /* Convert an xfs_fsmap to an fsmap. */
 static void
@@ -212,21 +213,20 @@ xfs_getfsmap_is_shared(
 	struct xfs_mount		*mp = tp->t_mountp;
 	struct xfs_btree_cur		*cur;
 	xfs_agblock_t			fbno;
-	xfs_extlen_t			flen;
+	xfs_extlen_t			flen = 0;
 	int				error;
 
 	*stat = false;
-	if (!xfs_has_reflink(mp))
-		return 0;
-	/* rt files will have no perag structure */
-	if (!info->group)
+	if (!xfs_has_reflink(mp) || !info->group)
 		return 0;
 
+	if (info->group->xg_type == XG_TYPE_RTG)
+		cur = xfs_rtrefcountbt_init_cursor(tp, to_rtg(info->group));
+	else
+		cur = xfs_refcountbt_init_cursor(mp, tp, info->agf_bp,
+				to_perag(info->group));
+
 	/* Are there any shared blocks here? */
-	flen = 0;
-	cur = xfs_refcountbt_init_cursor(mp, tp, info->agf_bp,
-			to_perag(info->group));
-
 	error = xfs_refcount_find_shared(cur, frec->rec_key,
 			XFS_BB_TO_FSBT(mp, frec->len_daddr), &fbno, &flen,
 			false);
@@ -863,7 +863,7 @@ xfs_getfsmap_rtdev_rmapbt_query(
 	struct xfs_rtgroup		*rtg = to_rtg(info->group);
 
 	/* Query the rtrmapbt */
-	xfs_rtgroup_lock(rtg, XFS_RTGLOCK_RMAP);
+	xfs_rtgroup_lock(rtg, XFS_RTGLOCK_RMAP | XFS_RTGLOCK_REFCOUNT);
 	*curpp = xfs_rtrmapbt_init_cursor(tp, rtg);
 	return xfs_rmap_query_range(*curpp, &info->low, &info->high,
 			xfs_getfsmap_rtdev_rmapbt_helper, info);
@@ -950,7 +950,8 @@ xfs_getfsmap_rtdev_rmapbt(
 
 		if (bt_cur) {
 			xfs_rtgroup_unlock(to_rtg(bt_cur->bc_group),
-					XFS_RTGLOCK_RMAP);
+					XFS_RTGLOCK_RMAP |
+					XFS_RTGLOCK_REFCOUNT);
 			xfs_btree_del_cursor(bt_cur, XFS_BTREE_NOERROR);
 			bt_cur = NULL;
 		}
@@ -988,7 +989,7 @@ xfs_getfsmap_rtdev_rmapbt(
 
 	if (bt_cur) {
 		xfs_rtgroup_unlock(to_rtg(bt_cur->bc_group),
-				XFS_RTGLOCK_RMAP);
+				XFS_RTGLOCK_RMAP | XFS_RTGLOCK_REFCOUNT);
 		xfs_btree_del_cursor(bt_cur, error < 0 ? XFS_BTREE_ERROR :
 							 XFS_BTREE_NOERROR);
 	}
diff --git a/fs/xfs/xfs_reflink.c b/fs/xfs/xfs_reflink.c
index 71b4743ffb7741..66ce29101462cd 100644
--- a/fs/xfs/xfs_reflink.c
+++ b/fs/xfs/xfs_reflink.c
@@ -30,6 +30,9 @@
 #include "xfs_ag.h"
 #include "xfs_ag_resv.h"
 #include "xfs_health.h"
+#include "xfs_rtrefcount_btree.h"
+#include "xfs_rtalloc.h"
+#include "xfs_rtgroup.h"
 
 /*
  * Copy on Write of Shared Blocks
@@ -163,6 +166,53 @@ xfs_reflink_find_shared(
 	return error;
 }
 
+/*
+ * Given a file mapping for the rt device, find the lowest-numbered run of
+ * shared blocks within that mapping and return it in shared_offset/shared_len.
+ * The offset is relative to the start of irec.
+ *
+ * If find_end_of_shared is true, return the longest contiguous extent of shared
+ * blocks.  If there are no shared extents, shared_offset and shared_len will be
+ * set to 0;
+ */
+static int
+xfs_reflink_find_rtshared(
+	struct xfs_mount	*mp,
+	struct xfs_trans	*tp,
+	const struct xfs_bmbt_irec *irec,
+	xfs_extlen_t		*shared_offset,
+	xfs_extlen_t		*shared_len,
+	bool			find_end_of_shared)
+{
+	struct xfs_rtgroup	*rtg;
+	struct xfs_btree_cur	*cur;
+	xfs_rgblock_t		orig_bno;
+	xfs_agblock_t		found_bno;
+	int			error;
+
+	BUILD_BUG_ON(NULLRGBLOCK != NULLAGBLOCK);
+
+	/*
+	 * Note: this uses the not quite correct xfs_agblock_t type because
+	 * xfs_refcount_find_shared is shared between the RT and data device
+	 * refcount code.
+	 */
+	orig_bno = xfs_rtb_to_rgbno(mp, irec->br_startblock);
+	rtg = xfs_rtgroup_get(mp, xfs_rtb_to_rgno(mp, irec->br_startblock));
+
+	xfs_rtgroup_lock(rtg, XFS_RTGLOCK_REFCOUNT);
+	cur = xfs_rtrefcountbt_init_cursor(tp, rtg);
+	error = xfs_refcount_find_shared(cur, orig_bno, irec->br_blockcount,
+			&found_bno, shared_len, find_end_of_shared);
+	xfs_btree_del_cursor(cur, error);
+	xfs_rtgroup_unlock(rtg, XFS_RTGLOCK_REFCOUNT);
+	xfs_rtgroup_put(rtg);
+
+	if (!error && *shared_len)
+		*shared_offset = found_bno - orig_bno;
+	return error;
+}
+
 /*
  * Trim the mapping to the next block where there's a change in the
  * shared/unshared status.  More specifically, this means that we
@@ -191,8 +241,12 @@ xfs_reflink_trim_around_shared(
 
 	trace_xfs_reflink_trim_around_shared(ip, irec);
 
-	error = xfs_reflink_find_shared(mp, NULL, irec, &shared_offset,
-			&shared_len, true);
+	if (XFS_IS_REALTIME_INODE(ip))
+		error = xfs_reflink_find_rtshared(mp, NULL, irec,
+				&shared_offset, &shared_len, true);
+	else
+		error = xfs_reflink_find_shared(mp, NULL, irec,
+				&shared_offset, &shared_len, true);
 	if (error)
 		return error;
 
@@ -1554,8 +1608,12 @@ xfs_reflink_inode_has_shared_extents(
 		    got.br_state != XFS_EXT_NORM)
 			goto next;
 
-		error = xfs_reflink_find_shared(mp, tp, &got, &shared_offset,
-				&shared_len, false);
+		if (XFS_IS_REALTIME_INODE(ip))
+			error = xfs_reflink_find_rtshared(mp, tp, &got,
+					&shared_offset, &shared_len, false);
+		else
+			error = xfs_reflink_find_shared(mp, tp, &got,
+					&shared_offset, &shared_len, false);
 		if (error)
 			return error;
 

[PATCH 15/43] xfs: create routine to allocate and initialize a realtime refcount btree inode

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Create a library routine to allocate and initialize an empty realtime
refcountbt inode.  We'll use this for growfs, mkfs, and repair.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_rtgroup.c          |    2 ++
 fs/xfs/libxfs/xfs_rtrefcount_btree.c |   28 ++++++++++++++++++++++++++++
 fs/xfs/libxfs/xfs_rtrefcount_btree.h |    3 +++
 3 files changed, 33 insertions(+)


diff --git a/fs/xfs/libxfs/xfs_rtgroup.c b/fs/xfs/libxfs/xfs_rtgroup.c
index d5ecc2f6c5c202..eab655a4a9ef5c 100644
--- a/fs/xfs/libxfs/xfs_rtgroup.c
+++ b/fs/xfs/libxfs/xfs_rtgroup.c
@@ -34,6 +34,7 @@
 #include "xfs_metafile.h"
 #include "xfs_metadir.h"
 #include "xfs_rtrmap_btree.h"
+#include "xfs_rtrefcount_btree.h"
 
 /* Find the first usable fsblock in this rtgroup. */
 static inline uint32_t
@@ -382,6 +383,7 @@ static const struct xfs_rtginode_ops xfs_rtginode_ops[XFS_RTGI_MAX] = {
 		.fmt_mask	= 1U << XFS_DINODE_FMT_META_BTREE,
 		/* same comment about growfs and rmap inodes applies here */
 		.enabled	= xfs_has_reflink,
+		.create		= xfs_rtrefcountbt_create,
 	},
 };
 
diff --git a/fs/xfs/libxfs/xfs_rtrefcount_btree.c b/fs/xfs/libxfs/xfs_rtrefcount_btree.c
index 6a5bc7ea42fbe6..151fb1ef7db126 100644
--- a/fs/xfs/libxfs/xfs_rtrefcount_btree.c
+++ b/fs/xfs/libxfs/xfs_rtrefcount_btree.c
@@ -721,3 +721,31 @@ xfs_iflush_rtrefcount(
 			ifp->if_broot_bytes, dfp,
 			XFS_DFORK_SIZE(dip, ip->i_mount, XFS_DATA_FORK));
 }
+
+/*
+ * Create a realtime refcount btree inode.
+ */
+int
+xfs_rtrefcountbt_create(
+	struct xfs_rtgroup	*rtg,
+	struct xfs_inode	*ip,
+	struct xfs_trans	*tp,
+	bool			init)
+{
+	struct xfs_ifork	*ifp = xfs_ifork_ptr(ip, XFS_DATA_FORK);
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_btree_block	*broot;
+
+	ifp->if_format = XFS_DINODE_FMT_META_BTREE;
+	ASSERT(ifp->if_broot_bytes == 0);
+	ASSERT(ifp->if_bytes == 0);
+
+	/* Initialize the empty incore btree root. */
+	broot = xfs_broot_realloc(ifp,
+			xfs_rtrefcount_broot_space_calc(mp, 0, 0));
+	if (broot)
+		xfs_btree_init_block(mp, broot, &xfs_rtrefcountbt_ops, 0, 0,
+				ip->i_ino);
+	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE | XFS_ILOG_DBROOT);
+	return 0;
+}
diff --git a/fs/xfs/libxfs/xfs_rtrefcount_btree.h b/fs/xfs/libxfs/xfs_rtrefcount_btree.h
index e558a10c4744ad..a99b7a8aec8659 100644
--- a/fs/xfs/libxfs/xfs_rtrefcount_btree.h
+++ b/fs/xfs/libxfs/xfs_rtrefcount_btree.h
@@ -183,4 +183,7 @@ void xfs_rtrefcountbt_to_disk(struct xfs_mount *mp,
 		struct xfs_rtrefcount_root *dblock, int dblocklen);
 void xfs_iflush_rtrefcount(struct xfs_inode *ip, struct xfs_dinode *dip);
 
+int xfs_rtrefcountbt_create(struct xfs_rtgroup *rtg, struct xfs_inode *ip,
+		struct xfs_trans *tp, bool init);
+
 #endif	/* __XFS_RTREFCOUNT_BTREE_H__ */

[PATCH 16/43] xfs: update rmap to allow cow staging extents in the rt rmap

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Don't error out on CoW staging extent records when realtime reflink is
enabled.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_rmap.c |    7 +++++++
 1 file changed, 7 insertions(+)


diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c
index f8415fd96cc2aa..3cdf50563fecb9 100644
--- a/fs/xfs/libxfs/xfs_rmap.c
+++ b/fs/xfs/libxfs/xfs_rmap.c
@@ -285,6 +285,13 @@ xfs_rtrmap_check_meta_irec(
 		if (irec->rm_blockcount != mp->m_sb.sb_rextsize)
 			return __this_address;
 		return NULL;
+	case XFS_RMAP_OWN_COW:
+		if (!xfs_has_rtreflink(mp))
+			return __this_address;
+		if (!xfs_verify_rgbext(rtg, irec->rm_startblock,
+					    irec->rm_blockcount))
+			return __this_address;
+		return NULL;
 	default:
 		return __this_address;
 	}

[PATCH 17/43] xfs: compute rtrmap btree max levels when reflink enabled

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Compute the maximum possible height of the realtime rmap btree when
reflink is enabled.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_rtrmap_btree.c |   28 ++++++++++++++++++++++++++--
 1 file changed, 26 insertions(+), 2 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_rtrmap_btree.c b/fs/xfs/libxfs/xfs_rtrmap_btree.c
index 3cb8f126b9ce16..04b9c76380adb0 100644
--- a/fs/xfs/libxfs/xfs_rtrmap_btree.c
+++ b/fs/xfs/libxfs/xfs_rtrmap_btree.c
@@ -718,6 +718,7 @@ xfs_rtrmapbt_maxrecs(
 unsigned int
 xfs_rtrmapbt_maxlevels_ondisk(void)
 {
+	unsigned long long	max_dblocks;
 	unsigned int		minrecs[2];
 	unsigned int		blocklen;
 
@@ -726,8 +727,20 @@ xfs_rtrmapbt_maxlevels_ondisk(void)
 	minrecs[0] = xfs_rtrmapbt_block_maxrecs(blocklen, true) / 2;
 	minrecs[1] = xfs_rtrmapbt_block_maxrecs(blocklen, false) / 2;
 
-	/* We need at most one record for every block in an rt group. */
-	return xfs_btree_compute_maxlevels(minrecs, XFS_MAX_RGBLOCKS);
+	/*
+	 * Compute the asymptotic maxlevels for an rtrmapbt on any rtreflink fs.
+	 *
+	 * On a reflink filesystem, each block in an rtgroup can have up to
+	 * 2^32 (per the refcount record format) owners, which means that
+	 * theoretically we could face up to 2^64 rmap records.  However, we're
+	 * likely to run out of blocks in the data device long before that
+	 * happens, which means that we must compute the max height based on
+	 * what the btree will look like if it consumes almost all the blocks
+	 * in the data device due to maximal sharing factor.
+	 */
+	max_dblocks = -1U; /* max ag count */
+	max_dblocks *= XFS_MAX_CRC_AG_BLOCKS;
+	return xfs_btree_space_to_height(minrecs, max_dblocks);
 }
 
 int __init
@@ -766,9 +779,20 @@ xfs_rtrmapbt_compute_maxlevels(
 	 * maximum height is constrained by the size of the data device and
 	 * the height required to store one rmap record for each block in an
 	 * rt group.
+	 *
+	 * On a reflink filesystem, each rt block can have up to 2^32 (per the
+	 * refcount record format) owners, which means that theoretically we
+	 * could face up to 2^64 rmap records.  This makes the computation of
+	 * maxlevels based on record count meaningless, so we only consider the
+	 * size of the data device.
 	 */
 	d_maxlevels = xfs_btree_space_to_height(mp->m_rtrmap_mnr,
 				mp->m_sb.sb_dblocks);
+	if (xfs_has_rtreflink(mp)) {
+		mp->m_rtrmap_maxlevels = d_maxlevels + 1;
+		return;
+	}
+
 	r_maxlevels = xfs_btree_compute_maxlevels(mp->m_rtrmap_mnr,
 				mp->m_groups[XG_TYPE_RTG].blocks);
 

[PATCH 18/43] xfs: refactor reflink quota updates

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Hoist all quota updates for reflink into a helper function, since things
are about to become more complicated.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/xfs_reflink.c |   37 ++++++++++++++++++++++++++++++++-----
 1 file changed, 32 insertions(+), 5 deletions(-)


diff --git a/fs/xfs/xfs_reflink.c b/fs/xfs/xfs_reflink.c
index 66ce29101462cd..29574b015fddc0 100644
--- a/fs/xfs/xfs_reflink.c
+++ b/fs/xfs/xfs_reflink.c
@@ -739,6 +739,35 @@ xfs_reflink_cancel_cow_range(
 	return error;
 }
 
+#ifdef CONFIG_XFS_QUOTA
+/*
+ * Update quota accounting for a remapping operation.  When we're remapping
+ * something from the CoW fork to the data fork, we must update the quota
+ * accounting for delayed allocations.  For remapping from the data fork to the
+ * data fork, use regular block accounting.
+ */
+static inline void
+xfs_reflink_update_quota(
+	struct xfs_trans	*tp,
+	struct xfs_inode	*ip,
+	bool			is_cow,
+	int64_t			blocks)
+{
+	unsigned int		qflag;
+
+	if (XFS_IS_REALTIME_INODE(ip)) {
+		qflag = is_cow ? XFS_TRANS_DQ_DELRTBCOUNT :
+				 XFS_TRANS_DQ_RTBCOUNT;
+	} else {
+		qflag = is_cow ? XFS_TRANS_DQ_DELBCOUNT :
+				 XFS_TRANS_DQ_BCOUNT;
+	}
+	xfs_trans_mod_dquot_byino(tp, ip, qflag, blocks);
+}
+#else
+# define xfs_reflink_update_quota(tp, ip, is_cow, blocks)	((void)0)
+#endif
+
 /*
  * Remap part of the CoW fork into the data fork.
  *
@@ -833,8 +862,7 @@ xfs_reflink_end_cow_extent(
 		 */
 		xfs_bmap_unmap_extent(tp, ip, XFS_DATA_FORK, &data);
 		xfs_refcount_decrease_extent(tp, isrt, &data);
-		xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT,
-				-data.br_blockcount);
+		xfs_reflink_update_quota(tp, ip, false, -data.br_blockcount);
 	} else if (data.br_startblock == DELAYSTARTBLOCK) {
 		int		done;
 
@@ -859,8 +887,7 @@ xfs_reflink_end_cow_extent(
 	xfs_bmap_map_extent(tp, ip, XFS_DATA_FORK, &del);
 
 	/* Charge this new data fork mapping to the on-disk quota. */
-	xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_DELBCOUNT,
-			(long)del.br_blockcount);
+	xfs_reflink_update_quota(tp, ip, true, del.br_blockcount);
 
 	/* Remove the mapping from the CoW fork. */
 	xfs_bmap_del_extent_cow(ip, &icur, &got, &del);
@@ -1347,7 +1374,7 @@ xfs_reflink_remap_extent(
 		qdelta += dmap->br_blockcount;
 	}
 
-	xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT, qdelta);
+	xfs_reflink_update_quota(tp, ip, false, qdelta);
 
 	/* Update dest isize if needed. */
 	newlen = XFS_FSB_TO_B(mp, dmap->br_startoff + dmap->br_blockcount);

[PATCH 19/43] xfs: enable CoW for realtime data

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Update our write paths to support copy on write on the rt volume.  This
works in more or less the same way as it does on the data device, with
the major exception that we never do delalloc on the rt volume.

Because we consider unwritten CoW fork staging extents to be incore
quota reservation, we update xfs_quota_reserve_blkres to support this
case.  Though xfs doesn't allow rt and quota together, the change is
trivial and we shouldn't leave a logic bomb here.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/xfs_reflink.c |   36 ++++++++++++++++++++++++++++--------
 1 file changed, 28 insertions(+), 8 deletions(-)


diff --git a/fs/xfs/xfs_reflink.c b/fs/xfs/xfs_reflink.c
index 29574b015fddc0..24f545687b8730 100644
--- a/fs/xfs/xfs_reflink.c
+++ b/fs/xfs/xfs_reflink.c
@@ -439,20 +439,26 @@ xfs_reflink_fill_cow_hole(
 	struct xfs_mount	*mp = ip->i_mount;
 	struct xfs_trans	*tp;
 	xfs_filblks_t		resaligned;
-	xfs_extlen_t		resblks;
+	unsigned int		dblocks = 0, rblocks = 0;
 	int			nimaps;
 	int			error;
 	bool			found;
 
 	resaligned = xfs_aligned_fsb_count(imap->br_startoff,
 		imap->br_blockcount, xfs_get_cowextsz_hint(ip));
-	resblks = XFS_DIOSTRAT_SPACE_RES(mp, resaligned);
+	if (XFS_IS_REALTIME_INODE(ip)) {
+		dblocks = XFS_DIOSTRAT_SPACE_RES(mp, 0);
+		rblocks = resaligned;
+	} else {
+		dblocks = XFS_DIOSTRAT_SPACE_RES(mp, resaligned);
+		rblocks = 0;
+	}
 
 	xfs_iunlock(ip, *lockmode);
 	*lockmode = 0;
 
-	error = xfs_trans_alloc_inode(ip, &M_RES(mp)->tr_write, resblks, 0,
-			false, &tp);
+	error = xfs_trans_alloc_inode(ip, &M_RES(mp)->tr_write, dblocks,
+			rblocks, false, &tp);
 	if (error)
 		return error;
 
@@ -1212,7 +1218,7 @@ xfs_reflink_remap_extent(
 	struct xfs_trans	*tp;
 	xfs_off_t		newlen;
 	int64_t			qdelta = 0;
-	unsigned int		resblks;
+	unsigned int		dblocks, rblocks, resblks;
 	bool			quota_reserved = true;
 	bool			smap_real;
 	bool			dmap_written = xfs_bmap_is_written_extent(dmap);
@@ -1243,8 +1249,15 @@ xfs_reflink_remap_extent(
 	 * we're remapping.
 	 */
 	resblks = XFS_EXTENTADD_SPACE_RES(mp, XFS_DATA_FORK);
+	if (XFS_IS_REALTIME_INODE(ip)) {
+		dblocks = resblks;
+		rblocks = dmap->br_blockcount;
+	} else {
+		dblocks = resblks + dmap->br_blockcount;
+		rblocks = 0;
+	}
 	error = xfs_trans_alloc_inode(ip, &M_RES(mp)->tr_write,
-			resblks + dmap->br_blockcount, 0, false, &tp);
+			dblocks, rblocks, false, &tp);
 	if (error == -EDQUOT || error == -ENOSPC) {
 		quota_reserved = false;
 		error = xfs_trans_alloc_inode(ip, &M_RES(mp)->tr_write,
@@ -1324,8 +1337,15 @@ xfs_reflink_remap_extent(
 	 * done.
 	 */
 	if (!quota_reserved && !smap_real && dmap_written) {
-		error = xfs_trans_reserve_quota_nblks(tp, ip,
-				dmap->br_blockcount, 0, false);
+		if (XFS_IS_REALTIME_INODE(ip)) {
+			dblocks = 0;
+			rblocks = dmap->br_blockcount;
+		} else {
+			dblocks = dmap->br_blockcount;
+			rblocks = 0;
+		}
+		error = xfs_trans_reserve_quota_nblks(tp, ip, dblocks, rblocks,
+				false);
 		if (error)
 			goto out_cancel;
 	}

[PATCH 20/43] xfs: enable sharing of realtime file blocks

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Update the remapping routines to be able to handle realtime files.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/xfs_reflink.c |   25 ++++++++++++++++++++-----
 1 file changed, 20 insertions(+), 5 deletions(-)


diff --git a/fs/xfs/xfs_reflink.c b/fs/xfs/xfs_reflink.c
index 24f545687b8730..78b47b2ac12453 100644
--- a/fs/xfs/xfs_reflink.c
+++ b/fs/xfs/xfs_reflink.c
@@ -33,6 +33,7 @@
 #include "xfs_rtrefcount_btree.h"
 #include "xfs_rtalloc.h"
 #include "xfs_rtgroup.h"
+#include "xfs_metafile.h"
 
 /*
  * Copy on Write of Shared Blocks
@@ -1187,14 +1188,28 @@ xfs_reflink_update_dest(
 static int
 xfs_reflink_ag_has_free_space(
 	struct xfs_mount	*mp,
-	xfs_agnumber_t		agno)
+	struct xfs_inode	*ip,
+	xfs_fsblock_t		fsb)
 {
 	struct xfs_perag	*pag;
+	xfs_agnumber_t		agno;
 	int			error = 0;
 
 	if (!xfs_has_rmapbt(mp))
 		return 0;
+	if (XFS_IS_REALTIME_INODE(ip)) {
+		struct xfs_rtgroup	*rtg;
+		xfs_rgnumber_t		rgno;
 
+		rgno = xfs_rtb_to_rgno(mp, fsb);
+		rtg = xfs_rtgroup_get(mp, rgno);
+		if (xfs_metafile_resv_critical(rtg_rmap(rtg)))
+			error = -ENOSPC;
+		xfs_rtgroup_put(rtg);
+		return error;
+	}
+
+	agno = XFS_FSB_TO_AGNO(mp, fsb);
 	pag = xfs_perag_get(mp, agno);
 	if (xfs_ag_resv_critical(pag, XFS_AG_RESV_RMAPBT) ||
 	    xfs_ag_resv_critical(pag, XFS_AG_RESV_METADATA))
@@ -1308,8 +1323,8 @@ xfs_reflink_remap_extent(
 
 	/* No reflinking if the AG of the dest mapping is low on space. */
 	if (dmap_written) {
-		error = xfs_reflink_ag_has_free_space(mp,
-				XFS_FSB_TO_AGNO(mp, dmap->br_startblock));
+		error = xfs_reflink_ag_has_free_space(mp, ip,
+				dmap->br_startblock);
 		if (error)
 			goto out_cancel;
 	}
@@ -1568,8 +1583,8 @@ xfs_reflink_remap_prep(
 
 	/* Check file eligibility and prepare for block sharing. */
 	ret = -EINVAL;
-	/* Don't reflink realtime inodes */
-	if (XFS_IS_REALTIME_INODE(src) || XFS_IS_REALTIME_INODE(dest))
+	/* Can't reflink between data and rt volumes */
+	if (XFS_IS_REALTIME_INODE(src) != XFS_IS_REALTIME_INODE(dest))
 		goto out_unlock;
 
 	/* Don't share DAX file data with non-DAX file. */

[PATCH 21/43] xfs: allow inodes to have the realtime and reflink flags

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Now that we can share blocks between realtime files, allow this
combination.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_inode_buf.c |    3 ++-
 fs/xfs/scrub/inode.c          |    5 +++--
 fs/xfs/scrub/inode_repair.c   |    6 ------
 fs/xfs/xfs_ioctl.c            |    4 ----
 4 files changed, 5 insertions(+), 13 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_inode_buf.c b/fs/xfs/libxfs/xfs_inode_buf.c
index 65eec8f60376d3..4273d096fb0a9c 100644
--- a/fs/xfs/libxfs/xfs_inode_buf.c
+++ b/fs/xfs/libxfs/xfs_inode_buf.c
@@ -748,7 +748,8 @@ xfs_dinode_verify(
 		return __this_address;
 
 	/* don't let reflink and realtime mix */
-	if ((flags2 & XFS_DIFLAG2_REFLINK) && (flags & XFS_DIFLAG_REALTIME))
+	if ((flags2 & XFS_DIFLAG2_REFLINK) && (flags & XFS_DIFLAG_REALTIME) &&
+	    !xfs_has_rtreflink(mp))
 		return __this_address;
 
 	/* COW extent size hint validation */
diff --git a/fs/xfs/scrub/inode.c b/fs/xfs/scrub/inode.c
index 8e702121dc8699..c7bbc3f78e90b1 100644
--- a/fs/xfs/scrub/inode.c
+++ b/fs/xfs/scrub/inode.c
@@ -360,8 +360,9 @@ xchk_inode_flags2(
 	if ((flags2 & XFS_DIFLAG2_REFLINK) && !S_ISREG(mode))
 		goto bad;
 
-	/* realtime and reflink make no sense, currently */
-	if ((flags & XFS_DIFLAG_REALTIME) && (flags2 & XFS_DIFLAG2_REFLINK))
+	/* realtime and reflink don't always go together */
+	if ((flags & XFS_DIFLAG_REALTIME) && (flags2 & XFS_DIFLAG2_REFLINK) &&
+	    !xfs_has_rtreflink(mp))
 		goto bad;
 
 	/* no bigtime iflag without the bigtime feature */
diff --git a/fs/xfs/scrub/inode_repair.c b/fs/xfs/scrub/inode_repair.c
index d7e3f033b16073..938a18721f3697 100644
--- a/fs/xfs/scrub/inode_repair.c
+++ b/fs/xfs/scrub/inode_repair.c
@@ -564,8 +564,6 @@ xrep_dinode_flags(
 		flags2 |= XFS_DIFLAG2_REFLINK;
 	else
 		flags2 &= ~(XFS_DIFLAG2_REFLINK | XFS_DIFLAG2_COWEXTSIZE);
-	if (flags & XFS_DIFLAG_REALTIME)
-		flags2 &= ~XFS_DIFLAG2_REFLINK;
 	if (!xfs_has_bigtime(mp))
 		flags2 &= ~XFS_DIFLAG2_BIGTIME;
 	if (!xfs_has_large_extent_counts(mp))
@@ -1790,10 +1788,6 @@ xrep_inode_flags(
 	/* DAX only applies to files and dirs. */
 	if (!(S_ISREG(mode) || S_ISDIR(mode)))
 		sc->ip->i_diflags2 &= ~XFS_DIFLAG2_DAX;
-
-	/* No reflink files on the realtime device. */
-	if (sc->ip->i_diflags & XFS_DIFLAG_REALTIME)
-		sc->ip->i_diflags2 &= ~XFS_DIFLAG2_REFLINK;
 }
 
 /*
diff --git a/fs/xfs/xfs_ioctl.c b/fs/xfs/xfs_ioctl.c
index 0789c18aaa1871..4caf29cc59b9ef 100644
--- a/fs/xfs/xfs_ioctl.c
+++ b/fs/xfs/xfs_ioctl.c
@@ -541,10 +541,6 @@ xfs_ioctl_setattr_xflags(
 		if (mp->m_sb.sb_rblocks == 0 || mp->m_sb.sb_rextsize == 0 ||
 		    xfs_extlen_to_rtxmod(mp, ip->i_extsize))
 			return -EINVAL;
-
-		/* Clear reflink if we are actually able to set the rt flag. */
-		if (xfs_is_reflink_inode(ip))
-			ip->i_diflags2 &= ~XFS_DIFLAG2_REFLINK;
 	}
 
 	/* diflags2 only valid for v3 inodes. */

[PATCH 22/43] xfs: recover CoW leftovers in the realtime volume

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Scan the realtime refcount tree at mount time to get rid of leftover
CoW staging extents.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_refcount.c |   47 +++++++++++++++++++++++++++++-------------
 fs/xfs/libxfs/xfs_refcount.h |    3 +--
 fs/xfs/xfs_reflink.c         |   15 +++++++++++--
 3 files changed, 46 insertions(+), 19 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_refcount.c b/fs/xfs/libxfs/xfs_refcount.c
index 11bff098db2dbb..cebe83f7842a1c 100644
--- a/fs/xfs/libxfs/xfs_refcount.c
+++ b/fs/xfs/libxfs/xfs_refcount.c
@@ -2054,12 +2054,13 @@ xfs_refcount_recover_extent(
 /* Find and remove leftover CoW reservations. */
 int
 xfs_refcount_recover_cow_leftovers(
-	struct xfs_mount		*mp,
-	struct xfs_perag		*pag)
+	struct xfs_group		*xg)
 {
+	struct xfs_mount		*mp = xg->xg_mount;
+	bool				isrt = xg->xg_type == XG_TYPE_RTG;
 	struct xfs_trans		*tp;
 	struct xfs_btree_cur		*cur;
-	struct xfs_buf			*agbp;
+	struct xfs_buf			*agbp = NULL;
 	struct xfs_refcount_recovery	*rr, *n;
 	struct list_head		debris;
 	union xfs_btree_irec		low = {
@@ -2072,10 +2073,19 @@ xfs_refcount_recover_cow_leftovers(
 	xfs_fsblock_t			fsb;
 	int				error;
 
-	/* reflink filesystems mustn't have AGs larger than 2^31-1 blocks */
+	/* reflink filesystems must not have groups larger than 2^31-1 blocks */
+	BUILD_BUG_ON(XFS_MAX_RGBLOCKS >= XFS_REFC_COWFLAG);
 	BUILD_BUG_ON(XFS_MAX_CRC_AG_BLOCKS >= XFS_REFC_COWFLAG);
-	if (mp->m_sb.sb_agblocks > XFS_MAX_CRC_AG_BLOCKS)
-		return -EOPNOTSUPP;
+
+	if (isrt) {
+		if (!xfs_has_rtgroups(mp))
+			return 0;
+		if (xfs_group_max_blocks(xg) >= XFS_MAX_RGBLOCKS)
+			return -EOPNOTSUPP;
+	} else {
+		if (xfs_group_max_blocks(xg) > XFS_MAX_CRC_AG_BLOCKS)
+			return -EOPNOTSUPP;
+	}
 
 	INIT_LIST_HEAD(&debris);
 
@@ -2093,16 +2103,24 @@ xfs_refcount_recover_cow_leftovers(
 	if (error)
 		return error;
 
-	error = xfs_alloc_read_agf(pag, tp, 0, &agbp);
-	if (error)
-		goto out_trans;
-	cur = xfs_refcountbt_init_cursor(mp, tp, agbp, pag);
+	if (isrt) {
+		xfs_rtgroup_lock(to_rtg(xg), XFS_RTGLOCK_REFCOUNT);
+		cur = xfs_rtrefcountbt_init_cursor(tp, to_rtg(xg));
+	} else {
+		error = xfs_alloc_read_agf(to_perag(xg), tp, 0, &agbp);
+		if (error)
+			goto out_trans;
+		cur = xfs_refcountbt_init_cursor(mp, tp, agbp, to_perag(xg));
+	}
 
 	/* Find all the leftover CoW staging extents. */
 	error = xfs_btree_query_range(cur, &low, &high,
 			xfs_refcount_recover_extent, &debris);
 	xfs_btree_del_cursor(cur, error);
-	xfs_trans_brelse(tp, agbp);
+	if (agbp)
+		xfs_trans_brelse(tp, agbp);
+	else
+		xfs_rtgroup_unlock(to_rtg(xg), XFS_RTGLOCK_REFCOUNT);
 	xfs_trans_cancel(tp);
 	if (error)
 		goto out_free;
@@ -2115,14 +2133,15 @@ xfs_refcount_recover_cow_leftovers(
 			goto out_free;
 
 		/* Free the orphan record */
-		fsb = xfs_agbno_to_fsb(pag, rr->rr_rrec.rc_startblock);
-		xfs_refcount_free_cow_extent(tp, false, fsb,
+		fsb = xfs_gbno_to_fsb(xg, rr->rr_rrec.rc_startblock);
+		xfs_refcount_free_cow_extent(tp, isrt, fsb,
 				rr->rr_rrec.rc_blockcount);
 
 		/* Free the block. */
 		error = xfs_free_extent_later(tp, fsb,
 				rr->rr_rrec.rc_blockcount, NULL,
-				XFS_AG_RESV_NONE, 0);
+				XFS_AG_RESV_NONE,
+				isrt ? XFS_FREE_EXTENT_REALTIME : 0);
 		if (error)
 			goto out_trans;
 
diff --git a/fs/xfs/libxfs/xfs_refcount.h b/fs/xfs/libxfs/xfs_refcount.h
index be11df25abcc5b..f2e299a716a4ee 100644
--- a/fs/xfs/libxfs/xfs_refcount.h
+++ b/fs/xfs/libxfs/xfs_refcount.h
@@ -94,8 +94,7 @@ void xfs_refcount_alloc_cow_extent(struct xfs_trans *tp, bool isrt,
 		xfs_fsblock_t fsb, xfs_extlen_t len);
 void xfs_refcount_free_cow_extent(struct xfs_trans *tp, bool isrt,
 		xfs_fsblock_t fsb, xfs_extlen_t len);
-extern int xfs_refcount_recover_cow_leftovers(struct xfs_mount *mp,
-		struct xfs_perag *pag);
+int xfs_refcount_recover_cow_leftovers(struct xfs_group *xg);
 
 /*
  * While we're adjusting the refcounts records of an extent, we have
diff --git a/fs/xfs/xfs_reflink.c b/fs/xfs/xfs_reflink.c
index 78b47b2ac12453..d9b33e22c17669 100644
--- a/fs/xfs/xfs_reflink.c
+++ b/fs/xfs/xfs_reflink.c
@@ -983,20 +983,29 @@ xfs_reflink_recover_cow(
 	struct xfs_mount	*mp)
 {
 	struct xfs_perag	*pag = NULL;
+	struct xfs_rtgroup	*rtg = NULL;
 	int			error = 0;
 
 	if (!xfs_has_reflink(mp))
 		return 0;
 
 	while ((pag = xfs_perag_next(mp, pag))) {
-		error = xfs_refcount_recover_cow_leftovers(mp, pag);
+		error = xfs_refcount_recover_cow_leftovers(pag_group(pag));
 		if (error) {
 			xfs_perag_rele(pag);
-			break;
+			return error;
 		}
 	}
 
-	return error;
+	while ((rtg = xfs_rtgroup_next(mp, rtg))) {
+		error = xfs_refcount_recover_cow_leftovers(rtg_group(rtg));
+		if (error) {
+			xfs_rtgroup_rele(rtg);
+			return error;
+		}
+	}
+
+	return 0;
 }
 
 /*

[PATCH 23/43] xfs: fix xfs_get_extsz_hint behavior with realtime alwayscow files

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Currently, we (ab)use xfs_get_extsz_hint so that it always returns a
nonzero value for realtime files.  This apparently was done to disable
delayed allocation for realtime files.

However, once we enable realtime reflink, we can also turn on the
alwayscow flag to force CoW writes to realtime files.  In this case, the
logic will incorrectly send the write through the delalloc write path.

Fix this by adjusting the logic slightly.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_bmap.c |    5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c
index d63713630236e7..ae3d33d6076102 100644
--- a/fs/xfs/libxfs/xfs_bmap.c
+++ b/fs/xfs/libxfs/xfs_bmap.c
@@ -6500,9 +6500,8 @@ xfs_get_extsz_hint(
 	 * No point in aligning allocations if we need to COW to actually
 	 * write to them.
 	 */
-	if (xfs_is_always_cow_inode(ip))
-		return 0;
-	if ((ip->i_diflags & XFS_DIFLAG_EXTSIZE) && ip->i_extsize)
+	if (!xfs_is_always_cow_inode(ip) &&
+	    (ip->i_diflags & XFS_DIFLAG_EXTSIZE) && ip->i_extsize)
 		return ip->i_extsize;
 	if (XFS_IS_REALTIME_INODE(ip) &&
 	    ip->i_mount->m_sb.sb_rextsize > 1)

[PATCH 24/43] xfs: apply rt extent alignment constraints to CoW extsize hint

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

The copy-on-write extent size hint is subject to the same alignment
constraints as the regular extent size hint.  Since we're in the process
of adding reflink (and therefore CoW) to the realtime device, we must
apply the same scattered rextsize alignment validation strategies to
both hints to deal with the possibility of rextsize changing.

Therefore, fix the inode validator to perform rextsize alignment checks
on regular realtime files, and to remove misaligned directory hints.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_inode_buf.c |   25 ++++++++++++++++++++-----
 fs/xfs/xfs_inode_item.c       |   14 ++++++++++++++
 fs/xfs/xfs_ioctl.c            |   17 +++++++++++++++--
 3 files changed, 49 insertions(+), 7 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_inode_buf.c b/fs/xfs/libxfs/xfs_inode_buf.c
index 4273d096fb0a9c..f24fa628fecf1e 100644
--- a/fs/xfs/libxfs/xfs_inode_buf.c
+++ b/fs/xfs/libxfs/xfs_inode_buf.c
@@ -910,11 +910,29 @@ xfs_inode_validate_cowextsize(
 	bool				rt_flag;
 	bool				hint_flag;
 	uint32_t			cowextsize_bytes;
+	uint32_t			blocksize_bytes;
 
 	rt_flag = (flags & XFS_DIFLAG_REALTIME);
 	hint_flag = (flags2 & XFS_DIFLAG2_COWEXTSIZE);
 	cowextsize_bytes = XFS_FSB_TO_B(mp, cowextsize);
 
+	/*
+	 * Similar to extent size hints, a directory can be configured to
+	 * propagate realtime status and a CoW extent size hint to newly
+	 * created files even if there is no realtime device, and the hints on
+	 * disk can become misaligned if the sysadmin changes the rt extent
+	 * size while adding the realtime device.
+	 *
+	 * Therefore, we can only enforce the rextsize alignment check against
+	 * regular realtime files, and rely on callers to decide when alignment
+	 * checks are appropriate, and fix things up as needed.
+	 */
+
+	if (rt_flag)
+		blocksize_bytes = XFS_FSB_TO_B(mp, mp->m_sb.sb_rextsize);
+	else
+		blocksize_bytes = mp->m_sb.sb_blocksize;
+
 	if (hint_flag && !xfs_has_reflink(mp))
 		return __this_address;
 
@@ -928,16 +946,13 @@ xfs_inode_validate_cowextsize(
 	if (mode && !hint_flag && cowextsize != 0)
 		return __this_address;
 
-	if (hint_flag && rt_flag)
-		return __this_address;
-
-	if (cowextsize_bytes % mp->m_sb.sb_blocksize)
+	if (cowextsize_bytes % blocksize_bytes)
 		return __this_address;
 
 	if (cowextsize > XFS_MAX_BMBT_EXTLEN)
 		return __this_address;
 
-	if (cowextsize > mp->m_sb.sb_agblocks / 2)
+	if (!rt_flag && cowextsize > mp->m_sb.sb_agblocks / 2)
 		return __this_address;
 
 	return NULL;
diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c
index a174f64b8bb250..70283c6419fd30 100644
--- a/fs/xfs/xfs_inode_item.c
+++ b/fs/xfs/xfs_inode_item.c
@@ -157,6 +157,20 @@ xfs_inode_item_precommit(
 	if (flags & XFS_ILOG_IVERSION)
 		flags = ((flags & ~XFS_ILOG_IVERSION) | XFS_ILOG_CORE);
 
+	/*
+	 * Inode verifiers do not check that the CoW extent size hint is an
+	 * integer multiple of the rt extent size on a directory with both
+	 * rtinherit and cowextsize flags set.  If we're logging a directory
+	 * that is misconfigured in this way, clear the hint.
+	 */
+	if ((ip->i_diflags & XFS_DIFLAG_RTINHERIT) &&
+	    (ip->i_diflags2 & XFS_DIFLAG2_COWEXTSIZE) &&
+	    xfs_extlen_to_rtxmod(ip->i_mount, ip->i_cowextsize) > 0) {
+		ip->i_diflags2 &= ~XFS_DIFLAG2_COWEXTSIZE;
+		ip->i_cowextsize = 0;
+		flags |= XFS_ILOG_CORE;
+	}
+
 	if (!iip->ili_item.li_buf) {
 		struct xfs_buf	*bp;
 		int		error;
diff --git a/fs/xfs/xfs_ioctl.c b/fs/xfs/xfs_ioctl.c
index 4caf29cc59b9ef..726282e74d546d 100644
--- a/fs/xfs/xfs_ioctl.c
+++ b/fs/xfs/xfs_ioctl.c
@@ -469,8 +469,21 @@ xfs_fill_fsxattr(
 		}
 	}
 
-	if (ip->i_diflags2 & XFS_DIFLAG2_COWEXTSIZE)
-		fa->fsx_cowextsize = XFS_FSB_TO_B(mp, ip->i_cowextsize);
+	if (ip->i_diflags2 & XFS_DIFLAG2_COWEXTSIZE) {
+		/*
+		 * Don't let a misaligned CoW extent size hint on a directory
+		 * escape to userspace if it won't pass the setattr checks
+		 * later.
+		 */
+		if ((ip->i_diflags & XFS_DIFLAG_RTINHERIT) &&
+		    ip->i_cowextsize % mp->m_sb.sb_rextsize > 0) {
+			fa->fsx_xflags &= ~FS_XFLAG_COWEXTSIZE;
+			fa->fsx_cowextsize = 0;
+		} else {
+			fa->fsx_cowextsize = XFS_FSB_TO_B(mp, ip->i_cowextsize);
+		}
+	}
+
 	fa->fsx_projid = ip->i_projid;
 	if (ifp && !xfs_need_iread_extents(ifp))
 		fa->fsx_nextents = xfs_iext_count(ifp);

[PATCH 25/43] xfs: enable extent size hints for CoW operations

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Wire up the copy-on-write extent size hint for realtime files, and
connect it to the rt allocator so that we avoid fragmentation on rt
filesystems.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_bmap.c |    8 +++++++-
 fs/xfs/xfs_rtalloc.c     |    5 ++++-
 2 files changed, 11 insertions(+), 2 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c
index ae3d33d6076102..40ad22fb808b95 100644
--- a/fs/xfs/libxfs/xfs_bmap.c
+++ b/fs/xfs/libxfs/xfs_bmap.c
@@ -6524,7 +6524,13 @@ xfs_get_cowextsz_hint(
 	a = 0;
 	if (ip->i_diflags2 & XFS_DIFLAG2_COWEXTSIZE)
 		a = ip->i_cowextsize;
-	b = xfs_get_extsz_hint(ip);
+	if (XFS_IS_REALTIME_INODE(ip)) {
+		b = 0;
+		if (ip->i_diflags & XFS_DIFLAG_EXTSIZE)
+			b = ip->i_extsize;
+	} else {
+		b = xfs_get_extsz_hint(ip);
+	}
 
 	a = max(a, b);
 	if (a == 0)
diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c
index 294aa0739be311..f5a3d5f8c948d8 100644
--- a/fs/xfs/xfs_rtalloc.c
+++ b/fs/xfs/xfs_rtalloc.c
@@ -2021,7 +2021,10 @@ xfs_rtallocate_align(
 	if (*noalign) {
 		align = mp->m_sb.sb_rextsize;
 	} else {
-		align = xfs_get_extsz_hint(ap->ip);
+		if (ap->flags & XFS_BMAPI_COWFORK)
+			align = xfs_get_cowextsz_hint(ap->ip);
+		else
+			align = xfs_get_extsz_hint(ap->ip);
 		if (!align)
 			align = 1;
 		if (align == mp->m_sb.sb_rextsize)

[PATCH 26/43] xfs: check that the rtrefcount maxlevels doesn't increase when growing fs

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

The size of filesystem transaction reservations depends on the maximum
height (maxlevels) of the realtime btrees.  Since we don't want a grow
operation to increase the reservation size enough that we'll fail the
minimum log size checks on the next mount, constrain growfs operations
if they would cause an increase in the rt refcount btree maxlevels.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/xfs_fsops.c   |    2 ++
 fs/xfs/xfs_rtalloc.c |    2 ++
 2 files changed, 4 insertions(+)


diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c
index 9df5a09c0acd3b..455298503d0102 100644
--- a/fs/xfs/xfs_fsops.c
+++ b/fs/xfs/xfs_fsops.c
@@ -23,6 +23,7 @@
 #include "xfs_trace.h"
 #include "xfs_rtalloc.h"
 #include "xfs_rtrmap_btree.h"
+#include "xfs_rtrefcount_btree.h"
 
 /*
  * Write new AG headers to disk. Non-transactional, but need to be
@@ -231,6 +232,7 @@ xfs_growfs_data_private(
 
 		/* Compute new maxlevels for rt btrees. */
 		xfs_rtrmapbt_compute_maxlevels(mp);
+		xfs_rtrefcountbt_compute_maxlevels(mp);
 	}
 
 	return error;
diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c
index f5a3d5f8c948d8..a5de5405800a22 100644
--- a/fs/xfs/xfs_rtalloc.c
+++ b/fs/xfs/xfs_rtalloc.c
@@ -995,6 +995,7 @@ xfs_growfs_rt_bmblock(
 	 */
 	mp->m_features |= XFS_FEAT_REALTIME;
 	xfs_rtrmapbt_compute_maxlevels(mp);
+	xfs_rtrefcountbt_compute_maxlevels(mp);
 
 	kfree(nmp);
 	return 0;
@@ -1178,6 +1179,7 @@ xfs_growfs_check_rtgeom(
 	nmp->m_sb.sb_dblocks = dblocks;
 
 	xfs_rtrmapbt_compute_maxlevels(nmp);
+	xfs_rtrefcountbt_compute_maxlevels(nmp);
 	xfs_trans_resv_calc(nmp, M_RES(nmp));
 
 	/*

[PATCH 27/43] xfs: report realtime refcount btree corruption errors to the health system

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Whenever we encounter corrupt realtime refcount btree blocks, we should
report that to the health monitoring system for later reporting.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_fs.h               |    1 +
 fs/xfs/libxfs/xfs_health.h           |    4 +++-
 fs/xfs/libxfs/xfs_rtgroup.c          |    1 +
 fs/xfs/libxfs/xfs_rtrefcount_btree.c |   10 ++++++++--
 fs/xfs/xfs_health.c                  |    1 +
 5 files changed, 14 insertions(+), 3 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_fs.h b/fs/xfs/libxfs/xfs_fs.h
index d42d3a5617e314..ea9e58a89d92d3 100644
--- a/fs/xfs/libxfs/xfs_fs.h
+++ b/fs/xfs/libxfs/xfs_fs.h
@@ -996,6 +996,7 @@ struct xfs_rtgroup_geometry {
 #define XFS_RTGROUP_GEOM_SICK_BITMAP	(1U << 1)  /* rtbitmap */
 #define XFS_RTGROUP_GEOM_SICK_SUMMARY	(1U << 2)  /* rtsummary */
 #define XFS_RTGROUP_GEOM_SICK_RMAPBT	(1U << 3)  /* reverse mappings */
+#define XFS_RTGROUP_GEOM_SICK_REFCNTBT	(1U << 4)  /* reference counts */
 
 /*
  * ioctl commands that are used by Linux filesystems
diff --git a/fs/xfs/libxfs/xfs_health.h b/fs/xfs/libxfs/xfs_health.h
index 5c8a0aff6ba6e9..b31000f7190ce5 100644
--- a/fs/xfs/libxfs/xfs_health.h
+++ b/fs/xfs/libxfs/xfs_health.h
@@ -71,6 +71,7 @@ struct xfs_rtgroup;
 #define XFS_SICK_RG_BITMAP	(1 << 1)  /* rt group bitmap */
 #define XFS_SICK_RG_SUMMARY	(1 << 2)  /* rt groups summary */
 #define XFS_SICK_RG_RMAPBT	(1 << 3)  /* reverse mappings */
+#define XFS_SICK_RG_REFCNTBT	(1 << 4)  /* reference counts */
 
 /* Observable health issues for AG metadata. */
 #define XFS_SICK_AG_SB		(1 << 0)  /* superblock */
@@ -117,7 +118,8 @@ struct xfs_rtgroup;
 #define XFS_SICK_RG_PRIMARY	(XFS_SICK_RG_SUPER | \
 				 XFS_SICK_RG_BITMAP | \
 				 XFS_SICK_RG_SUMMARY | \
-				 XFS_SICK_RG_RMAPBT)
+				 XFS_SICK_RG_RMAPBT | \
+				 XFS_SICK_RG_REFCNTBT)
 
 #define XFS_SICK_AG_PRIMARY	(XFS_SICK_AG_SB | \
 				 XFS_SICK_AG_AGF | \
diff --git a/fs/xfs/libxfs/xfs_rtgroup.c b/fs/xfs/libxfs/xfs_rtgroup.c
index eab655a4a9ef5c..a6468e591232c3 100644
--- a/fs/xfs/libxfs/xfs_rtgroup.c
+++ b/fs/xfs/libxfs/xfs_rtgroup.c
@@ -380,6 +380,7 @@ static const struct xfs_rtginode_ops xfs_rtginode_ops[XFS_RTGI_MAX] = {
 	[XFS_RTGI_REFCOUNT] = {
 		.name		= "refcount",
 		.metafile_type	= XFS_METAFILE_RTREFCOUNT,
+		.sick		= XFS_SICK_RG_REFCNTBT,
 		.fmt_mask	= 1U << XFS_DINODE_FMT_META_BTREE,
 		/* same comment about growfs and rmap inodes applies here */
 		.enabled	= xfs_has_reflink,
diff --git a/fs/xfs/libxfs/xfs_rtrefcount_btree.c b/fs/xfs/libxfs/xfs_rtrefcount_btree.c
index 151fb1ef7db126..3db5e7a4a94567 100644
--- a/fs/xfs/libxfs/xfs_rtrefcount_btree.c
+++ b/fs/xfs/libxfs/xfs_rtrefcount_btree.c
@@ -27,6 +27,7 @@
 #include "xfs_rtgroup.h"
 #include "xfs_rtbitmap.h"
 #include "xfs_metafile.h"
+#include "xfs_health.h"
 
 static struct kmem_cache	*xfs_rtrefcountbt_cur_cache;
 
@@ -374,6 +375,7 @@ const struct xfs_btree_ops xfs_rtrefcountbt_ops = {
 
 	.lru_refs		= XFS_REFC_BTREE_REF,
 	.statoff		= XFS_STATS_CALC_INDEX(xs_rtrefcbt_2),
+	.sick_mask		= XFS_SICK_RG_REFCNTBT,
 
 	.dup_cursor		= xfs_rtrefcountbt_dup_cursor,
 	.alloc_block		= xfs_btree_alloc_metafile_block,
@@ -640,16 +642,20 @@ xfs_iformat_rtrefcount(
 	 * volume to the filesystem, so we cannot use the rtrefcount predicate
 	 * here.
 	 */
-	if (!xfs_has_reflink(ip->i_mount))
+	if (!xfs_has_reflink(ip->i_mount)) {
+		xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE);
 		return -EFSCORRUPTED;
+	}
 
 	dsize = XFS_DFORK_SIZE(dip, mp, XFS_DATA_FORK);
 	numrecs = be16_to_cpu(dfp->bb_numrecs);
 	level = be16_to_cpu(dfp->bb_level);
 
 	if (level > mp->m_rtrefc_maxlevels ||
-	    xfs_rtrefcount_droot_space_calc(level, numrecs) > dsize)
+	    xfs_rtrefcount_droot_space_calc(level, numrecs) > dsize) {
+		xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE);
 		return -EFSCORRUPTED;
+	}
 
 	broot = xfs_broot_alloc(xfs_ifork_ptr(ip, XFS_DATA_FORK),
 			xfs_rtrefcount_broot_space_calc(mp, level, numrecs));
diff --git a/fs/xfs/xfs_health.c b/fs/xfs/xfs_health.c
index d438c3c001c829..7c541fb373d5b2 100644
--- a/fs/xfs/xfs_health.c
+++ b/fs/xfs/xfs_health.c
@@ -448,6 +448,7 @@ static const struct ioctl_sick_map rtgroup_map[] = {
 	{ XFS_SICK_RG_BITMAP,	XFS_RTGROUP_GEOM_SICK_BITMAP },
 	{ XFS_SICK_RG_SUMMARY,	XFS_RTGROUP_GEOM_SICK_SUMMARY },
 	{ XFS_SICK_RG_RMAPBT,	XFS_RTGROUP_GEOM_SICK_RMAPBT },
+	{ XFS_SICK_RG_REFCNTBT,	XFS_RTGROUP_GEOM_SICK_REFCNTBT },
 };
 
 /* Fill out rtgroup geometry health info. */

[PATCH 28/43] xfs: scrub the realtime refcount btree

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Add code to scrub realtime refcount btrees.  Similar to the refcount
btree checking code for the data device, we walk the rmap btree for each
refcount record to confirm that the reference counts are correct.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
---
 fs/xfs/Makefile           |    1 
 fs/xfs/libxfs/xfs_fs.h    |    3 
 fs/xfs/scrub/common.c     |   10 +
 fs/xfs/scrub/common.h     |    5 
 fs/xfs/scrub/health.c     |    1 
 fs/xfs/scrub/rtrefcount.c |  487 +++++++++++++++++++++++++++++++++++++++++++++
 fs/xfs/scrub/scrub.c      |    7 +
 fs/xfs/scrub/scrub.h      |    3 
 fs/xfs/scrub/stats.c      |    1 
 fs/xfs/scrub/trace.h      |    4 
 10 files changed, 519 insertions(+), 3 deletions(-)
 create mode 100644 fs/xfs/scrub/rtrefcount.c


diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile
index b9356d01416e16..9dd9921e53567c 100644
--- a/fs/xfs/Makefile
+++ b/fs/xfs/Makefile
@@ -195,6 +195,7 @@ xfs-$(CONFIG_XFS_ONLINE_SCRUB_STATS) += scrub/stats.o
 xfs-$(CONFIG_XFS_RT)		+= $(addprefix scrub/, \
 				   rgsuper.o \
 				   rtbitmap.o \
+				   rtrefcount.o \
 				   rtrmap.o \
 				   rtsummary.o \
 				   )
diff --git a/fs/xfs/libxfs/xfs_fs.h b/fs/xfs/libxfs/xfs_fs.h
index ea9e58a89d92d3..a4bd6a39c6ba71 100644
--- a/fs/xfs/libxfs/xfs_fs.h
+++ b/fs/xfs/libxfs/xfs_fs.h
@@ -738,9 +738,10 @@ struct xfs_scrub_metadata {
 #define XFS_SCRUB_TYPE_METAPATH	29	/* metadata directory tree paths */
 #define XFS_SCRUB_TYPE_RGSUPER	30	/* realtime superblock */
 #define XFS_SCRUB_TYPE_RTRMAPBT	31	/* rtgroup reverse mapping btree */
+#define XFS_SCRUB_TYPE_RTREFCBT	32	/* realtime reference count btree */
 
 /* Number of scrub subcommands. */
-#define XFS_SCRUB_TYPE_NR	32
+#define XFS_SCRUB_TYPE_NR	33
 
 /*
  * This special type code only applies to the vectored scrub implementation.
diff --git a/fs/xfs/scrub/common.c b/fs/xfs/scrub/common.c
index 06cb61e6349827..28ad341df8eede 100644
--- a/fs/xfs/scrub/common.c
+++ b/fs/xfs/scrub/common.c
@@ -37,6 +37,7 @@
 #include "xfs_rtgroup.h"
 #include "xfs_rtrmap_btree.h"
 #include "xfs_bmap_util.h"
+#include "xfs_rtrefcount_btree.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/trace.h"
@@ -797,6 +798,9 @@ xchk_rtgroup_lock(
 	if (xfs_has_rtrmapbt(sc->mp) && (rtglock_flags & XFS_RTGLOCK_RMAP))
 		sr->rmap_cur = xfs_rtrmapbt_init_cursor(sc->tp, sr->rtg);
 
+	if (xfs_has_rtreflink(sc->mp) && (rtglock_flags & XFS_RTGLOCK_REFCOUNT))
+		sr->refc_cur = xfs_rtrefcountbt_init_cursor(sc->tp, sr->rtg);
+
 	return 0;
 }
 
@@ -811,7 +815,10 @@ xchk_rtgroup_btcur_free(
 {
 	if (sr->rmap_cur)
 		xfs_btree_del_cursor(sr->rmap_cur, XFS_BTREE_ERROR);
+	if (sr->refc_cur)
+		xfs_btree_del_cursor(sr->refc_cur, XFS_BTREE_ERROR);
 
+	sr->refc_cur = NULL;
 	sr->rmap_cur = NULL;
 }
 
@@ -1687,6 +1694,9 @@ xchk_meta_btree_count_blocks(
 	case XFS_METAFILE_RTRMAP:
 		cur = xfs_rtrmapbt_init_cursor(sc->tp, sc->sr.rtg);
 		break;
+	case XFS_METAFILE_RTREFCOUNT:
+		cur = xfs_rtrefcountbt_init_cursor(sc->tp, sc->sr.rtg);
+		break;
 	default:
 		ASSERT(0);
 		return -EFSCORRUPTED;
diff --git a/fs/xfs/scrub/common.h b/fs/xfs/scrub/common.h
index 50ac6cca18fe45..bdcd40f0ec742c 100644
--- a/fs/xfs/scrub/common.h
+++ b/fs/xfs/scrub/common.h
@@ -82,11 +82,13 @@ int xchk_setup_rtbitmap(struct xfs_scrub *sc);
 int xchk_setup_rtsummary(struct xfs_scrub *sc);
 int xchk_setup_rgsuperblock(struct xfs_scrub *sc);
 int xchk_setup_rtrmapbt(struct xfs_scrub *sc);
+int xchk_setup_rtrefcountbt(struct xfs_scrub *sc);
 #else
 # define xchk_setup_rtbitmap		xchk_setup_nothing
 # define xchk_setup_rtsummary		xchk_setup_nothing
 # define xchk_setup_rgsuperblock	xchk_setup_nothing
 # define xchk_setup_rtrmapbt		xchk_setup_nothing
+# define xchk_setup_rtrefcountbt	xchk_setup_nothing
 #endif
 #ifdef CONFIG_XFS_QUOTA
 int xchk_ino_dqattach(struct xfs_scrub *sc);
@@ -129,7 +131,8 @@ xchk_ag_init_existing(
 
 /* All the locks we need to check an rtgroup. */
 #define XCHK_RTGLOCK_ALL	(XFS_RTGLOCK_BITMAP | \
-				 XFS_RTGLOCK_RMAP)
+				 XFS_RTGLOCK_RMAP | \
+				 XFS_RTGLOCK_REFCOUNT)
 
 int xchk_rtgroup_init(struct xfs_scrub *sc, xfs_rgnumber_t rgno,
 		struct xchk_rt *sr);
diff --git a/fs/xfs/scrub/health.c b/fs/xfs/scrub/health.c
index bcc4244e3b55db..3c0f25098b69f0 100644
--- a/fs/xfs/scrub/health.c
+++ b/fs/xfs/scrub/health.c
@@ -115,6 +115,7 @@ static const struct xchk_health_map type_to_health_flag[XFS_SCRUB_TYPE_NR] = {
 	[XFS_SCRUB_TYPE_METAPATH]	= { XHG_FS,  XFS_SICK_FS_METAPATH },
 	[XFS_SCRUB_TYPE_RGSUPER]	= { XHG_RTGROUP, XFS_SICK_RG_SUPER },
 	[XFS_SCRUB_TYPE_RTRMAPBT]	= { XHG_RTGROUP, XFS_SICK_RG_RMAPBT },
+	[XFS_SCRUB_TYPE_RTREFCBT]	= { XHG_RTGROUP, XFS_SICK_RG_REFCNTBT },
 };
 
 /* Return the health status mask for this scrub type. */
diff --git a/fs/xfs/scrub/rtrefcount.c b/fs/xfs/scrub/rtrefcount.c
new file mode 100644
index 00000000000000..92ae2e15ae9f79
--- /dev/null
+++ b/fs/xfs/scrub/rtrefcount.c
@@ -0,0 +1,487 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2021-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#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_btree.h"
+#include "xfs_rmap.h"
+#include "xfs_refcount.h"
+#include "xfs_inode.h"
+#include "xfs_rtbitmap.h"
+#include "xfs_rtgroup.h"
+#include "xfs_metafile.h"
+#include "xfs_rtrefcount_btree.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/btree.h"
+
+/* Set us up with the realtime refcount metadata locked. */
+int
+xchk_setup_rtrefcountbt(
+	struct xfs_scrub	*sc)
+{
+	int			error;
+
+	if (xchk_need_intent_drain(sc))
+		xchk_fsgates_enable(sc, XCHK_FSGATES_DRAIN);
+
+	error = xchk_rtgroup_init(sc, sc->sm->sm_agno, &sc->sr);
+	if (error)
+		return error;
+
+	error = xchk_setup_rt(sc);
+	if (error)
+		return error;
+
+	error = xchk_install_live_inode(sc, rtg_refcount(sc->sr.rtg));
+	if (error)
+		return error;
+
+	return xchk_rtgroup_lock(sc, &sc->sr, XCHK_RTGLOCK_ALL);
+}
+
+/* Realtime Reference count btree scrubber. */
+
+/*
+ * Confirming Reference Counts via Reverse Mappings
+ *
+ * We want to count the reverse mappings overlapping a refcount record
+ * (bno, len, refcount), allowing for the possibility that some of the
+ * overlap may come from smaller adjoining reverse mappings, while some
+ * comes from single extents which overlap the range entirely.  The
+ * outer loop is as follows:
+ *
+ * 1. For all reverse mappings overlapping the refcount extent,
+ *    a. If a given rmap completely overlaps, mark it as seen.
+ *    b. Otherwise, record the fragment (in agbno order) for later
+ *       processing.
+ *
+ * Once we've seen all the rmaps, we know that for all blocks in the
+ * refcount record we want to find $refcount owners and we've already
+ * visited $seen extents that overlap all the blocks.  Therefore, we
+ * need to find ($refcount - $seen) owners for every block in the
+ * extent; call that quantity $target_nr.  Proceed as follows:
+ *
+ * 2. Pull the first $target_nr fragments from the list; all of them
+ *    should start at or before the start of the extent.
+ *    Call this subset of fragments the working set.
+ * 3. Until there are no more unprocessed fragments,
+ *    a. Find the shortest fragments in the set and remove them.
+ *    b. Note the block number of the end of these fragments.
+ *    c. Pull the same number of fragments from the list.  All of these
+ *       fragments should start at the block number recorded in the
+ *       previous step.
+ *    d. Put those fragments in the set.
+ * 4. Check that there are $target_nr fragments remaining in the list,
+ *    and that they all end at or beyond the end of the refcount extent.
+ *
+ * If the refcount is correct, all the check conditions in the algorithm
+ * should always hold true.  If not, the refcount is incorrect.
+ */
+struct xchk_rtrefcnt_frag {
+	struct list_head	list;
+	struct xfs_rmap_irec	rm;
+};
+
+struct xchk_rtrefcnt_check {
+	struct xfs_scrub	*sc;
+	struct list_head	fragments;
+
+	/* refcount extent we're examining */
+	xfs_rgblock_t		bno;
+	xfs_extlen_t		len;
+	xfs_nlink_t		refcount;
+
+	/* number of owners seen */
+	xfs_nlink_t		seen;
+};
+
+/*
+ * Decide if the given rmap is large enough that we can redeem it
+ * towards refcount verification now, or if it's a fragment, in
+ * which case we'll hang onto it in the hopes that we'll later
+ * discover that we've collected exactly the correct number of
+ * fragments as the rtrefcountbt says we should have.
+ */
+STATIC int
+xchk_rtrefcountbt_rmap_check(
+	struct xfs_btree_cur		*cur,
+	const struct xfs_rmap_irec	*rec,
+	void				*priv)
+{
+	struct xchk_rtrefcnt_check	*refchk = priv;
+	struct xchk_rtrefcnt_frag	*frag;
+	xfs_rgblock_t			rm_last;
+	xfs_rgblock_t			rc_last;
+	int				error = 0;
+
+	if (xchk_should_terminate(refchk->sc, &error))
+		return error;
+
+	rm_last = rec->rm_startblock + rec->rm_blockcount - 1;
+	rc_last = refchk->bno + refchk->len - 1;
+
+	/* Confirm that a single-owner refc extent is a CoW stage. */
+	if (refchk->refcount == 1 && rec->rm_owner != XFS_RMAP_OWN_COW) {
+		xchk_btree_xref_set_corrupt(refchk->sc, cur, 0);
+		return 0;
+	}
+
+	if (rec->rm_startblock <= refchk->bno && rm_last >= rc_last) {
+		/*
+		 * The rmap overlaps the refcount record, so we can confirm
+		 * one refcount owner seen.
+		 */
+		refchk->seen++;
+	} else {
+		/*
+		 * This rmap covers only part of the refcount record, so
+		 * save the fragment for later processing.  If the rmapbt
+		 * is healthy each rmap_irec we see will be in agbno order
+		 * so we don't need insertion sort here.
+		 */
+		frag = kmalloc(sizeof(struct xchk_rtrefcnt_frag),
+				XCHK_GFP_FLAGS);
+		if (!frag)
+			return -ENOMEM;
+		memcpy(&frag->rm, rec, sizeof(frag->rm));
+		list_add_tail(&frag->list, &refchk->fragments);
+	}
+
+	return 0;
+}
+
+/*
+ * Given a bunch of rmap fragments, iterate through them, keeping
+ * a running tally of the refcount.  If this ever deviates from
+ * what we expect (which is the rtrefcountbt's refcount minus the
+ * number of extents that totally covered the rtrefcountbt extent),
+ * we have a rtrefcountbt error.
+ */
+STATIC void
+xchk_rtrefcountbt_process_rmap_fragments(
+	struct xchk_rtrefcnt_check	*refchk)
+{
+	struct list_head		worklist;
+	struct xchk_rtrefcnt_frag	*frag;
+	struct xchk_rtrefcnt_frag	*n;
+	xfs_rgblock_t			bno;
+	xfs_rgblock_t			rbno;
+	xfs_rgblock_t			next_rbno;
+	xfs_nlink_t			nr;
+	xfs_nlink_t			target_nr;
+
+	target_nr = refchk->refcount - refchk->seen;
+	if (target_nr == 0)
+		return;
+
+	/*
+	 * There are (refchk->rc.rc_refcount - refchk->nr refcount)
+	 * references we haven't found yet.  Pull that many off the
+	 * fragment list and figure out where the smallest rmap ends
+	 * (and therefore the next rmap should start).  All the rmaps
+	 * we pull off should start at or before the beginning of the
+	 * refcount record's range.
+	 */
+	INIT_LIST_HEAD(&worklist);
+	rbno = NULLRGBLOCK;
+
+	/* Make sure the fragments actually /are/ in bno order. */
+	bno = 0;
+	list_for_each_entry(frag, &refchk->fragments, list) {
+		if (frag->rm.rm_startblock < bno)
+			goto done;
+		bno = frag->rm.rm_startblock;
+	}
+
+	/*
+	 * Find all the rmaps that start at or before the refc extent,
+	 * and put them on the worklist.
+	 */
+	nr = 0;
+	list_for_each_entry_safe(frag, n, &refchk->fragments, list) {
+		if (frag->rm.rm_startblock > refchk->bno || nr > target_nr)
+			break;
+		bno = frag->rm.rm_startblock + frag->rm.rm_blockcount;
+		if (bno < rbno)
+			rbno = bno;
+		list_move_tail(&frag->list, &worklist);
+		nr++;
+	}
+
+	/*
+	 * We should have found exactly $target_nr rmap fragments starting
+	 * at or before the refcount extent.
+	 */
+	if (nr != target_nr)
+		goto done;
+
+	while (!list_empty(&refchk->fragments)) {
+		/* Discard any fragments ending at rbno from the worklist. */
+		nr = 0;
+		next_rbno = NULLRGBLOCK;
+		list_for_each_entry_safe(frag, n, &worklist, list) {
+			bno = frag->rm.rm_startblock + frag->rm.rm_blockcount;
+			if (bno != rbno) {
+				if (bno < next_rbno)
+					next_rbno = bno;
+				continue;
+			}
+			list_del(&frag->list);
+			kfree(frag);
+			nr++;
+		}
+
+		/* Try to add nr rmaps starting at rbno to the worklist. */
+		list_for_each_entry_safe(frag, n, &refchk->fragments, list) {
+			bno = frag->rm.rm_startblock + frag->rm.rm_blockcount;
+			if (frag->rm.rm_startblock != rbno)
+				goto done;
+			list_move_tail(&frag->list, &worklist);
+			if (next_rbno > bno)
+				next_rbno = bno;
+			nr--;
+			if (nr == 0)
+				break;
+		}
+
+		/*
+		 * If we get here and nr > 0, this means that we added fewer
+		 * items to the worklist than we discarded because the fragment
+		 * list ran out of items.  Therefore, we cannot maintain the
+		 * required refcount.  Something is wrong, so we're done.
+		 */
+		if (nr)
+			goto done;
+
+		rbno = next_rbno;
+	}
+
+	/*
+	 * Make sure the last extent we processed ends at or beyond
+	 * the end of the refcount extent.
+	 */
+	if (rbno < refchk->bno + refchk->len)
+		goto done;
+
+	/* Actually record us having seen the remaining refcount. */
+	refchk->seen = refchk->refcount;
+done:
+	/* Delete fragments and work list. */
+	list_for_each_entry_safe(frag, n, &worklist, list) {
+		list_del(&frag->list);
+		kfree(frag);
+	}
+	list_for_each_entry_safe(frag, n, &refchk->fragments, list) {
+		list_del(&frag->list);
+		kfree(frag);
+	}
+}
+
+/* Use the rmap entries covering this extent to verify the refcount. */
+STATIC void
+xchk_rtrefcountbt_xref_rmap(
+	struct xfs_scrub		*sc,
+	const struct xfs_refcount_irec	*irec)
+{
+	struct xchk_rtrefcnt_check	refchk = {
+		.sc			= sc,
+		.bno			= irec->rc_startblock,
+		.len			= irec->rc_blockcount,
+		.refcount		= irec->rc_refcount,
+		.seen			= 0,
+	};
+	struct xfs_rmap_irec		low;
+	struct xfs_rmap_irec		high;
+	struct xchk_rtrefcnt_frag	*frag;
+	struct xchk_rtrefcnt_frag	*n;
+	int				error;
+
+	if (!sc->sr.rmap_cur || xchk_skip_xref(sc->sm))
+		return;
+
+	/* Cross-reference with the rmapbt to confirm the refcount. */
+	memset(&low, 0, sizeof(low));
+	low.rm_startblock = irec->rc_startblock;
+	memset(&high, 0xFF, sizeof(high));
+	high.rm_startblock = irec->rc_startblock + irec->rc_blockcount - 1;
+
+	INIT_LIST_HEAD(&refchk.fragments);
+	error = xfs_rmap_query_range(sc->sr.rmap_cur, &low, &high,
+			xchk_rtrefcountbt_rmap_check, &refchk);
+	if (!xchk_should_check_xref(sc, &error, &sc->sr.rmap_cur))
+		goto out_free;
+
+	xchk_rtrefcountbt_process_rmap_fragments(&refchk);
+	if (irec->rc_refcount != refchk.seen)
+		xchk_btree_xref_set_corrupt(sc, sc->sr.rmap_cur, 0);
+
+out_free:
+	list_for_each_entry_safe(frag, n, &refchk.fragments, list) {
+		list_del(&frag->list);
+		kfree(frag);
+	}
+}
+
+/* Cross-reference with the other btrees. */
+STATIC void
+xchk_rtrefcountbt_xref(
+	struct xfs_scrub		*sc,
+	const struct xfs_refcount_irec	*irec)
+{
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return;
+
+	xchk_xref_is_used_rt_space(sc,
+			xfs_rgbno_to_rtb(sc->sr.rtg, irec->rc_startblock),
+			irec->rc_blockcount);
+	xchk_rtrefcountbt_xref_rmap(sc, irec);
+}
+
+struct xchk_rtrefcbt_records {
+	/* Previous refcount record. */
+	struct xfs_refcount_irec	prev_rec;
+
+	/* Number of CoW blocks we expect. */
+	xfs_extlen_t			cow_blocks;
+};
+
+static inline bool
+xchk_rtrefcount_mergeable(
+	struct xchk_rtrefcbt_records	*rrc,
+	const struct xfs_refcount_irec	*r2)
+{
+	const struct xfs_refcount_irec	*r1 = &rrc->prev_rec;
+
+	/* Ignore if prev_rec is not yet initialized. */
+	if (r1->rc_blockcount > 0)
+		return false;
+
+	if (r1->rc_startblock + r1->rc_blockcount != r2->rc_startblock)
+		return false;
+	if (r1->rc_refcount != r2->rc_refcount)
+		return false;
+	if ((unsigned long long)r1->rc_blockcount + r2->rc_blockcount >
+			XFS_REFC_LEN_MAX)
+		return false;
+
+	return true;
+}
+
+/* Flag failures for records that could be merged. */
+STATIC void
+xchk_rtrefcountbt_check_mergeable(
+	struct xchk_btree		*bs,
+	struct xchk_rtrefcbt_records	*rrc,
+	const struct xfs_refcount_irec	*irec)
+{
+	if (bs->sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return;
+
+	if (xchk_rtrefcount_mergeable(rrc, irec))
+		xchk_btree_set_corrupt(bs->sc, bs->cur, 0);
+
+	memcpy(&rrc->prev_rec, irec, sizeof(struct xfs_refcount_irec));
+}
+
+/* Scrub a rtrefcountbt record. */
+STATIC int
+xchk_rtrefcountbt_rec(
+	struct xchk_btree		*bs,
+	const union xfs_btree_rec	*rec)
+{
+	struct xfs_mount		*mp = bs->cur->bc_mp;
+	struct xchk_rtrefcbt_records	*rrc = bs->private;
+	struct xfs_refcount_irec	irec;
+	u32				mod;
+
+	xfs_refcount_btrec_to_irec(rec, &irec);
+	if (xfs_rtrefcount_check_irec(to_rtg(bs->cur->bc_group), &irec) !=
+			NULL) {
+		xchk_btree_set_corrupt(bs->sc, bs->cur, 0);
+		return 0;
+	}
+
+	/* We can only share full rt extents. */
+	mod = xfs_rgbno_to_rtxoff(mp, irec.rc_startblock);
+	if (mod)
+		xchk_btree_set_corrupt(bs->sc, bs->cur, 0);
+	mod = xfs_extlen_to_rtxmod(mp, irec.rc_blockcount);
+	if (mod)
+		xchk_btree_set_corrupt(bs->sc, bs->cur, 0);
+
+	if (irec.rc_domain == XFS_REFC_DOMAIN_COW)
+		rrc->cow_blocks += irec.rc_blockcount;
+
+	xchk_rtrefcountbt_check_mergeable(bs, rrc, &irec);
+	xchk_rtrefcountbt_xref(bs->sc, &irec);
+
+	return 0;
+}
+
+/* Make sure we have as many refc blocks as the rmap says. */
+STATIC void
+xchk_refcount_xref_rmap(
+	struct xfs_scrub	*sc,
+	const struct xfs_owner_info *btree_oinfo,
+	xfs_extlen_t		cow_blocks)
+{
+	xfs_filblks_t		refcbt_blocks = 0;
+	xfs_filblks_t		blocks;
+	int			error;
+
+	if (!sc->sr.rmap_cur || !sc->sa.rmap_cur || xchk_skip_xref(sc->sm))
+		return;
+
+	/* Check that we saw as many refcbt blocks as the rmap knows about. */
+	error = xfs_btree_count_blocks(sc->sr.refc_cur, &refcbt_blocks);
+	if (!xchk_btree_process_error(sc, sc->sr.refc_cur, 0, &error))
+		return;
+	error = xchk_count_rmap_ownedby_ag(sc, sc->sa.rmap_cur, btree_oinfo,
+			&blocks);
+	if (!xchk_should_check_xref(sc, &error, &sc->sa.rmap_cur))
+		return;
+	if (blocks != refcbt_blocks)
+		xchk_btree_xref_set_corrupt(sc, sc->sa.rmap_cur, 0);
+
+	/* Check that we saw as many cow blocks as the rmap knows about. */
+	error = xchk_count_rmap_ownedby_ag(sc, sc->sr.rmap_cur,
+			&XFS_RMAP_OINFO_COW, &blocks);
+	if (!xchk_should_check_xref(sc, &error, &sc->sr.rmap_cur))
+		return;
+	if (blocks != cow_blocks)
+		xchk_btree_xref_set_corrupt(sc, sc->sr.rmap_cur, 0);
+}
+
+/* Scrub the refcount btree for some AG. */
+int
+xchk_rtrefcountbt(
+	struct xfs_scrub	*sc)
+{
+	struct xfs_owner_info	btree_oinfo;
+	struct xchk_rtrefcbt_records rrc = {
+		.cow_blocks	= 0,
+	};
+	int			error;
+
+	error = xchk_metadata_inode_forks(sc);
+	if (error || (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT))
+		return error;
+
+	xfs_rmap_ino_bmbt_owner(&btree_oinfo, rtg_refcount(sc->sr.rtg)->i_ino,
+			XFS_DATA_FORK);
+	error = xchk_btree(sc, sc->sr.refc_cur, xchk_rtrefcountbt_rec,
+			&btree_oinfo, &rrc);
+	if (error || (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT))
+		return error;
+
+	xchk_refcount_xref_rmap(sc, &btree_oinfo, rrc.cow_blocks);
+
+	return 0;
+}
diff --git a/fs/xfs/scrub/scrub.c b/fs/xfs/scrub/scrub.c
index 16da054b2eb0dc..6e31f12cef4cc9 100644
--- a/fs/xfs/scrub/scrub.c
+++ b/fs/xfs/scrub/scrub.c
@@ -467,6 +467,13 @@ static const struct xchk_meta_ops meta_scrub_ops[] = {
 		.has	= xfs_has_rtrmapbt,
 		.repair	= xrep_rtrmapbt,
 	},
+	[XFS_SCRUB_TYPE_RTREFCBT] = {	/* realtime refcountbt */
+		.type	= ST_RTGROUP,
+		.setup	= xchk_setup_rtrefcountbt,
+		.scrub	= xchk_rtrefcountbt,
+		.has	= xfs_has_rtreflink,
+		.repair	= xrep_notsupported,
+	},
 };
 
 static int
diff --git a/fs/xfs/scrub/scrub.h b/fs/xfs/scrub/scrub.h
index cba4e89a3a627b..ab3d221dd9ed2d 100644
--- a/fs/xfs/scrub/scrub.h
+++ b/fs/xfs/scrub/scrub.h
@@ -129,6 +129,7 @@ struct xchk_rt {
 
 	/* rtgroup btrees */
 	struct xfs_btree_cur	*rmap_cur;
+	struct xfs_btree_cur	*refc_cur;
 };
 
 struct xfs_scrub {
@@ -284,11 +285,13 @@ int xchk_rtbitmap(struct xfs_scrub *sc);
 int xchk_rtsummary(struct xfs_scrub *sc);
 int xchk_rgsuperblock(struct xfs_scrub *sc);
 int xchk_rtrmapbt(struct xfs_scrub *sc);
+int xchk_rtrefcountbt(struct xfs_scrub *sc);
 #else
 # define xchk_rtbitmap		xchk_nothing
 # define xchk_rtsummary		xchk_nothing
 # define xchk_rgsuperblock	xchk_nothing
 # define xchk_rtrmapbt		xchk_nothing
+# define xchk_rtrefcountbt	xchk_nothing
 #endif
 #ifdef CONFIG_XFS_QUOTA
 int xchk_quota(struct xfs_scrub *sc);
diff --git a/fs/xfs/scrub/stats.c b/fs/xfs/scrub/stats.c
index eb6bb170c902b3..f8a37ea9779192 100644
--- a/fs/xfs/scrub/stats.c
+++ b/fs/xfs/scrub/stats.c
@@ -83,6 +83,7 @@ static const char *name_map[XFS_SCRUB_TYPE_NR] = {
 	[XFS_SCRUB_TYPE_METAPATH]	= "metapath",
 	[XFS_SCRUB_TYPE_RGSUPER]	= "rgsuper",
 	[XFS_SCRUB_TYPE_RTRMAPBT]	= "rtrmapbt",
+	[XFS_SCRUB_TYPE_RTREFCBT]	= "rtrefcountbt",
 };
 
 /* Format the scrub stats into a text buffer, similar to pcp style. */
diff --git a/fs/xfs/scrub/trace.h b/fs/xfs/scrub/trace.h
index fb86b746bc174a..289e39d1f418ba 100644
--- a/fs/xfs/scrub/trace.h
+++ b/fs/xfs/scrub/trace.h
@@ -77,6 +77,7 @@ TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_BARRIER);
 TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_METAPATH);
 TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_RGSUPER);
 TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_RTRMAPBT);
+TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_RTREFCBT);
 
 #define XFS_SCRUB_TYPE_STRINGS \
 	{ XFS_SCRUB_TYPE_PROBE,		"probe" }, \
@@ -111,7 +112,8 @@ TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_RTRMAPBT);
 	{ XFS_SCRUB_TYPE_BARRIER,	"barrier" }, \
 	{ XFS_SCRUB_TYPE_METAPATH,	"metapath" }, \
 	{ XFS_SCRUB_TYPE_RGSUPER,	"rgsuper" }, \
-	{ XFS_SCRUB_TYPE_RTRMAPBT,	"rtrmapbt" }
+	{ XFS_SCRUB_TYPE_RTRMAPBT,	"rtrmapbt" }, \
+	{ XFS_SCRUB_TYPE_RTREFCBT,	"rtrefcountbt" }
 
 #define XFS_SCRUB_FLAG_STRINGS \
 	{ XFS_SCRUB_IFLAG_REPAIR,		"repair" }, \

Re: [PATCH 28/43] xfs: scrub the realtime refcount btree

[Christoph Hellwig]

Looks good:

Reviewed-by: Christoph Hellwig <hch@lst.de>

[PATCH 29/43] xfs: cross-reference checks with the rt refcount btree

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Use the realtime refcount btree to implement cross-reference checks in
other data structures.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/scrub/bmap.c       |   30 +++++++++++++---
 fs/xfs/scrub/rtbitmap.c   |    2 +
 fs/xfs/scrub/rtrefcount.c |   86 +++++++++++++++++++++++++++++++++++++++++++++
 fs/xfs/scrub/rtrmap.c     |   37 +++++++++++++++++++
 fs/xfs/scrub/scrub.h      |    9 +++++
 5 files changed, 158 insertions(+), 6 deletions(-)


diff --git a/fs/xfs/scrub/bmap.c b/fs/xfs/scrub/bmap.c
index f6077b0cba8a14..66da7d4d56ba74 100644
--- a/fs/xfs/scrub/bmap.c
+++ b/fs/xfs/scrub/bmap.c
@@ -347,13 +347,31 @@ xchk_bmap_rt_iextent_xref(
 		goto out_cur;
 
 	rgbno = xfs_rtb_to_rgbno(info->sc->mp, irec->br_startblock);
-	xchk_bmap_xref_rmap(info, irec, rgbno);
-
-	xfs_rmap_ino_owner(&oinfo, info->sc->ip->i_ino, info->whichfork,
-			irec->br_startoff);
-	xchk_xref_is_only_rt_owned_by(info->sc, rgbno,
-			irec->br_blockcount, &oinfo);
 
+	switch (info->whichfork) {
+	case XFS_DATA_FORK:
+		xchk_bmap_xref_rmap(info, irec, rgbno);
+		if (!xfs_is_reflink_inode(info->sc->ip)) {
+			xfs_rmap_ino_owner(&oinfo, info->sc->ip->i_ino,
+					info->whichfork, irec->br_startoff);
+			xchk_xref_is_only_rt_owned_by(info->sc, rgbno,
+					irec->br_blockcount, &oinfo);
+			xchk_xref_is_not_rt_shared(info->sc, rgbno,
+					irec->br_blockcount);
+		}
+		xchk_xref_is_not_rt_cow_staging(info->sc, rgbno,
+				irec->br_blockcount);
+		break;
+	case XFS_COW_FORK:
+		xchk_bmap_xref_rmap_cow(info, irec, rgbno);
+		xchk_xref_is_only_rt_owned_by(info->sc, rgbno,
+				irec->br_blockcount, &XFS_RMAP_OINFO_COW);
+		xchk_xref_is_rt_cow_staging(info->sc, rgbno,
+				irec->br_blockcount);
+		xchk_xref_is_not_rt_shared(info->sc, rgbno,
+				irec->br_blockcount);
+		break;
+	}
 out_cur:
 	xchk_rtgroup_btcur_free(&info->sc->sr);
 out_free:
diff --git a/fs/xfs/scrub/rtbitmap.c b/fs/xfs/scrub/rtbitmap.c
index 28c90a31f4c32b..e8c776a34c1df1 100644
--- a/fs/xfs/scrub/rtbitmap.c
+++ b/fs/xfs/scrub/rtbitmap.c
@@ -105,6 +105,8 @@ xchk_rtbitmap_xref(
 		return;
 
 	xchk_xref_has_no_rt_owner(sc, rgbno, blockcount);
+	xchk_xref_is_not_rt_shared(sc, rgbno, blockcount);
+	xchk_xref_is_not_rt_cow_staging(sc, rgbno, blockcount);
 
 	if (rtb->next_free_rgbno < rgbno)
 		xchk_xref_has_rt_owner(sc, rtb->next_free_rgbno,
diff --git a/fs/xfs/scrub/rtrefcount.c b/fs/xfs/scrub/rtrefcount.c
index 92ae2e15ae9f79..f56f2cb7a9f707 100644
--- a/fs/xfs/scrub/rtrefcount.c
+++ b/fs/xfs/scrub/rtrefcount.c
@@ -485,3 +485,89 @@ xchk_rtrefcountbt(
 
 	return 0;
 }
+
+/* xref check that a cow staging extent is marked in the rtrefcountbt. */
+void
+xchk_xref_is_rt_cow_staging(
+	struct xfs_scrub		*sc,
+	xfs_rgblock_t			bno,
+	xfs_extlen_t			len)
+{
+	struct xfs_refcount_irec	rc;
+	int				has_refcount;
+	int				error;
+
+	if (!sc->sr.refc_cur || xchk_skip_xref(sc->sm))
+		return;
+
+	/* Find the CoW staging extent. */
+	error = xfs_refcount_lookup_le(sc->sr.refc_cur, XFS_REFC_DOMAIN_COW,
+			bno, &has_refcount);
+	if (!xchk_should_check_xref(sc, &error, &sc->sr.refc_cur))
+		return;
+	if (!has_refcount) {
+		xchk_btree_xref_set_corrupt(sc, sc->sr.refc_cur, 0);
+		return;
+	}
+
+	error = xfs_refcount_get_rec(sc->sr.refc_cur, &rc, &has_refcount);
+	if (!xchk_should_check_xref(sc, &error, &sc->sr.refc_cur))
+		return;
+	if (!has_refcount) {
+		xchk_btree_xref_set_corrupt(sc, sc->sr.refc_cur, 0);
+		return;
+	}
+
+	/* CoW lookup returned a shared extent record? */
+	if (rc.rc_domain != XFS_REFC_DOMAIN_COW)
+		xchk_btree_xref_set_corrupt(sc, sc->sa.refc_cur, 0);
+
+	/* Must be at least as long as what was passed in */
+	if (rc.rc_blockcount < len)
+		xchk_btree_xref_set_corrupt(sc, sc->sr.refc_cur, 0);
+}
+
+/*
+ * xref check that the extent is not shared.  Only file data blocks
+ * can have multiple owners.
+ */
+void
+xchk_xref_is_not_rt_shared(
+	struct xfs_scrub	*sc,
+	xfs_rgblock_t		bno,
+	xfs_extlen_t		len)
+{
+	enum xbtree_recpacking	outcome;
+	int			error;
+
+	if (!sc->sr.refc_cur || xchk_skip_xref(sc->sm))
+		return;
+
+	error = xfs_refcount_has_records(sc->sr.refc_cur,
+			XFS_REFC_DOMAIN_SHARED, bno, len, &outcome);
+	if (!xchk_should_check_xref(sc, &error, &sc->sr.refc_cur))
+		return;
+	if (outcome != XBTREE_RECPACKING_EMPTY)
+		xchk_btree_xref_set_corrupt(sc, sc->sr.refc_cur, 0);
+}
+
+/* xref check that the extent is not being used for CoW staging. */
+void
+xchk_xref_is_not_rt_cow_staging(
+	struct xfs_scrub	*sc,
+	xfs_rgblock_t		bno,
+	xfs_extlen_t		len)
+{
+	enum xbtree_recpacking	outcome;
+	int			error;
+
+	if (!sc->sr.refc_cur || xchk_skip_xref(sc->sm))
+		return;
+
+	error = xfs_refcount_has_records(sc->sr.refc_cur, XFS_REFC_DOMAIN_COW,
+			bno, len, &outcome);
+	if (!xchk_should_check_xref(sc, &error, &sc->sr.refc_cur))
+		return;
+	if (outcome != XBTREE_RECPACKING_EMPTY)
+		xchk_btree_xref_set_corrupt(sc, sc->sr.refc_cur, 0);
+}
diff --git a/fs/xfs/scrub/rtrmap.c b/fs/xfs/scrub/rtrmap.c
index 300a1e85b3d625..3d5419682d6528 100644
--- a/fs/xfs/scrub/rtrmap.c
+++ b/fs/xfs/scrub/rtrmap.c
@@ -22,6 +22,7 @@
 #include "xfs_rtalloc.h"
 #include "xfs_rtgroup.h"
 #include "xfs_metafile.h"
+#include "xfs_refcount.h"
 #include "scrub/xfs_scrub.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
@@ -149,6 +150,37 @@ xchk_rtrmapbt_check_mergeable(
 	memcpy(&cr->prev_rec, irec, sizeof(struct xfs_rmap_irec));
 }
 
+/* Cross-reference a rmap against the refcount btree. */
+STATIC void
+xchk_rtrmapbt_xref_rtrefc(
+	struct xfs_scrub	*sc,
+	struct xfs_rmap_irec	*irec)
+{
+	xfs_rgblock_t		fbno;
+	xfs_extlen_t		flen;
+	bool			is_inode;
+	bool			is_bmbt;
+	bool			is_attr;
+	bool			is_unwritten;
+	int			error;
+
+	if (!sc->sr.refc_cur || xchk_skip_xref(sc->sm))
+		return;
+
+	is_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;
+	is_unwritten = irec->rm_flags & XFS_RMAP_UNWRITTEN;
+
+	/* If this is shared, must be a data fork extent. */
+	error = xfs_refcount_find_shared(sc->sr.refc_cur, irec->rm_startblock,
+			irec->rm_blockcount, &fbno, &flen, false);
+	if (!xchk_should_check_xref(sc, &error, &sc->sr.refc_cur))
+		return;
+	if (flen != 0 && (!is_inode || is_attr || is_bmbt || is_unwritten))
+		xchk_btree_xref_set_corrupt(sc, sc->sr.refc_cur, 0);
+}
+
 /* Cross-reference with other metadata. */
 STATIC void
 xchk_rtrmapbt_xref(
@@ -161,6 +193,11 @@ xchk_rtrmapbt_xref(
 	xchk_xref_is_used_rt_space(sc,
 			xfs_rgbno_to_rtb(sc->sr.rtg, irec->rm_startblock),
 			irec->rm_blockcount);
+	if (irec->rm_owner == XFS_RMAP_OWN_COW)
+		xchk_xref_is_cow_staging(sc, irec->rm_startblock,
+				irec->rm_blockcount);
+	else
+		xchk_rtrmapbt_xref_rtrefc(sc, irec);
 }
 
 /* Scrub a realtime rmapbt record. */
diff --git a/fs/xfs/scrub/scrub.h b/fs/xfs/scrub/scrub.h
index ab3d221dd9ed2d..a1086f1f06d0d9 100644
--- a/fs/xfs/scrub/scrub.h
+++ b/fs/xfs/scrub/scrub.h
@@ -331,11 +331,20 @@ void xchk_xref_has_rt_owner(struct xfs_scrub *sc, xfs_rgblock_t rgbno,
 		xfs_extlen_t len);
 void xchk_xref_is_only_rt_owned_by(struct xfs_scrub *sc, xfs_rgblock_t rgbno,
 		xfs_extlen_t len, const struct xfs_owner_info *oinfo);
+void xchk_xref_is_rt_cow_staging(struct xfs_scrub *sc, xfs_rgblock_t rgbno,
+		xfs_extlen_t len);
+void xchk_xref_is_not_rt_shared(struct xfs_scrub *sc, xfs_rgblock_t rgbno,
+		xfs_extlen_t len);
+void xchk_xref_is_not_rt_cow_staging(struct xfs_scrub *sc, xfs_rgblock_t rgbno,
+		xfs_extlen_t len);
 #else
 # define xchk_xref_is_used_rt_space(sc, rtbno, len) do { } while (0)
 # define xchk_xref_has_no_rt_owner(sc, rtbno, len) do { } while (0)
 # define xchk_xref_has_rt_owner(sc, rtbno, len) do { } while (0)
 # define xchk_xref_is_only_rt_owned_by(sc, bno, len, oinfo) do { } while (0)
+# define xchk_xref_is_rt_cow_staging(sc, bno, len) do { } while (0)
+# define xchk_xref_is_not_rt_shared(sc, bno, len) do { } while (0)
+# define xchk_xref_is_not_rt_cow_staging(sc, bno, len) do { } while (0)
 #endif
 
 #endif	/* __XFS_SCRUB_SCRUB_H__ */

[PATCH 30/43] xfs: allow overlapping rtrmapbt records for shared data extents

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Allow overlapping realtime reverse mapping records if they both describe
shared data extents and the fs supports reflink on the realtime volume.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/scrub/rtrmap.c |   17 ++++++++++++++++-
 1 file changed, 16 insertions(+), 1 deletion(-)


diff --git a/fs/xfs/scrub/rtrmap.c b/fs/xfs/scrub/rtrmap.c
index 3d5419682d6528..12989fe80e8bda 100644
--- a/fs/xfs/scrub/rtrmap.c
+++ b/fs/xfs/scrub/rtrmap.c
@@ -78,6 +78,18 @@ struct xchk_rtrmap {
 	struct xfs_rmap_irec	prev_rec;
 };
 
+static inline bool
+xchk_rtrmapbt_is_shareable(
+	struct xfs_scrub		*sc,
+	const struct xfs_rmap_irec	*irec)
+{
+	if (!xfs_has_rtreflink(sc->mp))
+		return false;
+	if (irec->rm_flags & XFS_RMAP_UNWRITTEN)
+		return false;
+	return true;
+}
+
 /* Flag failures for records that overlap but cannot. */
 STATIC void
 xchk_rtrmapbt_check_overlapping(
@@ -99,7 +111,10 @@ xchk_rtrmapbt_check_overlapping(
 	if (pnext <= irec->rm_startblock)
 		goto set_prev;
 
-	xchk_btree_set_corrupt(bs->sc, bs->cur, 0);
+	/* Overlap is only allowed if both records are data fork mappings. */
+	if (!xchk_rtrmapbt_is_shareable(bs->sc, &cr->overlap_rec) ||
+	    !xchk_rtrmapbt_is_shareable(bs->sc, irec))
+		xchk_btree_set_corrupt(bs->sc, bs->cur, 0);
 
 	/* Save whichever rmap record extends furthest. */
 	inext = irec->rm_startblock + irec->rm_blockcount;

[PATCH 31/43] xfs: check reference counts of gaps between rt refcount records

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

If there's a gap between records in the rt refcount btree, we ought to
cross-reference the gap with the rtrmap records to make sure that there
aren't any overlapping records for a region that doesn't have any shared
ownership.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/scrub/rtrefcount.c |   81 ++++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 80 insertions(+), 1 deletion(-)


diff --git a/fs/xfs/scrub/rtrefcount.c b/fs/xfs/scrub/rtrefcount.c
index f56f2cb7a9f707..18c9bcb0e82184 100644
--- a/fs/xfs/scrub/rtrefcount.c
+++ b/fs/xfs/scrub/rtrefcount.c
@@ -17,6 +17,7 @@
 #include "xfs_rtgroup.h"
 #include "xfs_metafile.h"
 #include "xfs_rtrefcount_btree.h"
+#include "xfs_rtalloc.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/btree.h"
@@ -348,8 +349,14 @@ struct xchk_rtrefcbt_records {
 	/* Previous refcount record. */
 	struct xfs_refcount_irec	prev_rec;
 
+	/* The next rtgroup block where we aren't expecting shared extents. */
+	xfs_rgblock_t			next_unshared_rgbno;
+
 	/* Number of CoW blocks we expect. */
 	xfs_extlen_t			cow_blocks;
+
+	/* Was the last record a shared or CoW staging extent? */
+	enum xfs_refc_domain		prev_domain;
 };
 
 static inline bool
@@ -390,6 +397,53 @@ xchk_rtrefcountbt_check_mergeable(
 	memcpy(&rrc->prev_rec, irec, sizeof(struct xfs_refcount_irec));
 }
 
+STATIC int
+xchk_rtrefcountbt_rmap_check_gap(
+	struct xfs_btree_cur		*cur,
+	const struct xfs_rmap_irec	*rec,
+	void				*priv)
+{
+	xfs_rgblock_t			*next_bno = priv;
+
+	if (*next_bno != NULLRGBLOCK && rec->rm_startblock < *next_bno)
+		return -ECANCELED;
+
+	*next_bno = rec->rm_startblock + rec->rm_blockcount;
+	return 0;
+}
+
+/*
+ * Make sure that a gap in the reference count records does not correspond to
+ * overlapping records (i.e. shared extents) in the reverse mappings.
+ */
+static inline void
+xchk_rtrefcountbt_xref_gaps(
+	struct xfs_scrub	*sc,
+	struct xchk_rtrefcbt_records *rrc,
+	xfs_rtblock_t		bno)
+{
+	struct xfs_rmap_irec	low;
+	struct xfs_rmap_irec	high;
+	xfs_rgblock_t		next_bno = NULLRGBLOCK;
+	int			error;
+
+	if (bno <= rrc->next_unshared_rgbno || !sc->sr.rmap_cur ||
+            xchk_skip_xref(sc->sm))
+		return;
+
+	memset(&low, 0, sizeof(low));
+	low.rm_startblock = rrc->next_unshared_rgbno;
+	memset(&high, 0xFF, sizeof(high));
+	high.rm_startblock = bno - 1;
+
+	error = xfs_rmap_query_range(sc->sr.rmap_cur, &low, &high,
+			xchk_rtrefcountbt_rmap_check_gap, &next_bno);
+	if (error == -ECANCELED)
+		xchk_btree_xref_set_corrupt(sc, sc->sr.rmap_cur, 0);
+	else
+		xchk_should_check_xref(sc, &error, &sc->sr.rmap_cur);
+}
+
 /* Scrub a rtrefcountbt record. */
 STATIC int
 xchk_rtrefcountbt_rec(
@@ -419,9 +473,26 @@ xchk_rtrefcountbt_rec(
 	if (irec.rc_domain == XFS_REFC_DOMAIN_COW)
 		rrc->cow_blocks += irec.rc_blockcount;
 
+	/* Shared records always come before CoW records. */
+	if (irec.rc_domain == XFS_REFC_DOMAIN_SHARED &&
+	    rrc->prev_domain == XFS_REFC_DOMAIN_COW)
+		xchk_btree_set_corrupt(bs->sc, bs->cur, 0);
+	rrc->prev_domain = irec.rc_domain;
+
 	xchk_rtrefcountbt_check_mergeable(bs, rrc, &irec);
 	xchk_rtrefcountbt_xref(bs->sc, &irec);
 
+	/*
+	 * If this is a record for a shared extent, check that all blocks
+	 * between the previous record and this one have at most one reverse
+	 * mapping.
+	 */
+	if (irec.rc_domain == XFS_REFC_DOMAIN_SHARED) {
+		xchk_rtrefcountbt_xref_gaps(bs->sc, rrc, irec.rc_startblock);
+		rrc->next_unshared_rgbno = irec.rc_startblock +
+					   irec.rc_blockcount;
+	}
+
 	return 0;
 }
 
@@ -466,7 +537,9 @@ xchk_rtrefcountbt(
 {
 	struct xfs_owner_info	btree_oinfo;
 	struct xchk_rtrefcbt_records rrc = {
-		.cow_blocks	= 0,
+		.cow_blocks		= 0,
+		.next_unshared_rgbno	= 0,
+		.prev_domain		= XFS_REFC_DOMAIN_SHARED,
 	};
 	int			error;
 
@@ -481,6 +554,12 @@ xchk_rtrefcountbt(
 	if (error || (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT))
 		return error;
 
+	/*
+	 * Check that all blocks between the last refcount > 1 record and the
+	 * end of the rt volume have at most one reverse mapping.
+	 */
+	xchk_rtrefcountbt_xref_gaps(sc, &rrc, sc->mp->m_sb.sb_rblocks);
+
 	xchk_refcount_xref_rmap(sc, &btree_oinfo, rrc.cow_blocks);
 
 	return 0;

[PATCH 32/43] xfs: allow dquot rt block count to exceed rt blocks on reflink fs

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Update the quota scrubber to allow dquots where the realtime block count
exceeds the block count of the rt volume if reflink is enabled.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/scrub/quota.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)


diff --git a/fs/xfs/scrub/quota.c b/fs/xfs/scrub/quota.c
index 183d531875eae5..58d6d4ed2853b3 100644
--- a/fs/xfs/scrub/quota.c
+++ b/fs/xfs/scrub/quota.c
@@ -212,12 +212,18 @@ xchk_quota_item(
 		if (mp->m_sb.sb_dblocks < dq->q_blk.count)
 			xchk_fblock_set_warning(sc, XFS_DATA_FORK,
 					offset);
+		if (mp->m_sb.sb_rblocks < dq->q_rtb.count)
+			xchk_fblock_set_warning(sc, XFS_DATA_FORK,
+					offset);
 	} else {
 		if (mp->m_sb.sb_dblocks < dq->q_blk.count)
 			xchk_fblock_set_corrupt(sc, XFS_DATA_FORK,
 					offset);
+		if (mp->m_sb.sb_rblocks < dq->q_rtb.count)
+			xchk_fblock_set_corrupt(sc, XFS_DATA_FORK,
+					offset);
 	}
-	if (dq->q_ino.count > fs_icount || dq->q_rtb.count > mp->m_sb.sb_rblocks)
+	if (dq->q_ino.count > fs_icount)
 		xchk_fblock_set_corrupt(sc, XFS_DATA_FORK, offset);
 
 	/*

[PATCH 33/43] xfs: detect and repair misaligned rtinherit directory cowextsize hints

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

If we encounter a directory that has been configured to pass on a CoW
extent size hint to a new realtime file and the hint isn't an integer
multiple of the rt extent size, we should flag the hint for
administrative review and/or turn it off because that is a
misconfiguration.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/scrub/inode.c        |   26 +++++++++++++++++---------
 fs/xfs/scrub/inode_repair.c |   15 +++++++++++++++
 2 files changed, 32 insertions(+), 9 deletions(-)


diff --git a/fs/xfs/scrub/inode.c b/fs/xfs/scrub/inode.c
index c7bbc3f78e90b1..db6edd5a5fe5d8 100644
--- a/fs/xfs/scrub/inode.c
+++ b/fs/xfs/scrub/inode.c
@@ -260,12 +260,7 @@ xchk_inode_extsize(
 		xchk_ino_set_warning(sc, ino);
 }
 
-/*
- * Validate di_cowextsize hint.
- *
- * The rules are documented at xfs_ioctl_setattr_check_cowextsize().
- * These functions must be kept in sync with each other.
- */
+/* Validate di_cowextsize hint. */
 STATIC void
 xchk_inode_cowextsize(
 	struct xfs_scrub	*sc,
@@ -276,12 +271,25 @@ xchk_inode_cowextsize(
 	uint64_t		flags2)
 {
 	xfs_failaddr_t		fa;
+	uint32_t		value = be32_to_cpu(dip->di_cowextsize);
 
-	fa = xfs_inode_validate_cowextsize(sc->mp,
-			be32_to_cpu(dip->di_cowextsize), mode, flags,
-			flags2);
+	fa = xfs_inode_validate_cowextsize(sc->mp, value, mode, flags, flags2);
 	if (fa)
 		xchk_ino_set_corrupt(sc, ino);
+
+	/*
+	 * XFS allows a sysadmin to change the rt extent size when adding a rt
+	 * section to a filesystem after formatting.  If there are any
+	 * directories with cowextsize and rtinherit set, the hint could become
+	 * misaligned with the new rextsize.  The verifier doesn't check this,
+	 * because we allow rtinherit directories even without an rt device.
+	 * Flag this as an administrative warning since we will clean this up
+	 * eventually.
+	 */
+	if ((flags & XFS_DIFLAG_RTINHERIT) &&
+	    (flags2 & XFS_DIFLAG2_COWEXTSIZE) &&
+	    value % sc->mp->m_sb.sb_rextsize > 0)
+		xchk_ino_set_warning(sc, ino);
 }
 
 /* Make sure the di_flags make sense for the inode. */
diff --git a/fs/xfs/scrub/inode_repair.c b/fs/xfs/scrub/inode_repair.c
index 938a18721f3697..701baee144a66e 100644
--- a/fs/xfs/scrub/inode_repair.c
+++ b/fs/xfs/scrub/inode_repair.c
@@ -1903,6 +1903,20 @@ xrep_inode_pptr(
 			sizeof(struct xfs_attr_sf_hdr), true);
 }
 
+/* Fix COW extent size hint problems. */
+STATIC void
+xrep_inode_cowextsize(
+	struct xfs_scrub	*sc)
+{
+	/* Fix misaligned CoW extent size hints on a directory. */
+	if ((sc->ip->i_diflags & XFS_DIFLAG_RTINHERIT) &&
+	    (sc->ip->i_diflags2 & XFS_DIFLAG2_COWEXTSIZE) &&
+	    sc->ip->i_extsize % sc->mp->m_sb.sb_rextsize > 0) {
+		sc->ip->i_cowextsize = 0;
+		sc->ip->i_diflags2 &= ~XFS_DIFLAG2_COWEXTSIZE;
+	}
+}
+
 /* Fix any irregularities in an inode that the verifiers don't catch. */
 STATIC int
 xrep_inode_problems(
@@ -1926,6 +1940,7 @@ xrep_inode_problems(
 	if (S_ISDIR(VFS_I(sc->ip)->i_mode))
 		xrep_inode_dir_size(sc);
 	xrep_inode_extsize(sc);
+	xrep_inode_cowextsize(sc);
 
 	trace_xrep_inode_fixed(sc);
 	xfs_trans_log_inode(sc->tp, sc->ip, XFS_ILOG_CORE);

[PATCH 34/43] xfs: scrub the metadir path of rt refcount btree files

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Add a new XFS_SCRUB_METAPATH subtype so that we can scrub the metadata
directory tree path to the refcount btree file for each rt group.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_fs.h  |    3 ++-
 fs/xfs/scrub/metapath.c |    3 +++
 2 files changed, 5 insertions(+), 1 deletion(-)


diff --git a/fs/xfs/libxfs/xfs_fs.h b/fs/xfs/libxfs/xfs_fs.h
index a4bd6a39c6ba71..2c3171262b445b 100644
--- a/fs/xfs/libxfs/xfs_fs.h
+++ b/fs/xfs/libxfs/xfs_fs.h
@@ -832,9 +832,10 @@ struct xfs_scrub_vec_head {
 #define XFS_SCRUB_METAPATH_GRPQUOTA	(6)  /* group quota */
 #define XFS_SCRUB_METAPATH_PRJQUOTA	(7)  /* project quota */
 #define XFS_SCRUB_METAPATH_RTRMAPBT	(8)  /* realtime reverse mapping */
+#define XFS_SCRUB_METAPATH_RTREFCOUNTBT	(9)  /* realtime refcount */
 
 /* Number of metapath sm_ino values */
-#define XFS_SCRUB_METAPATH_NR		(9)
+#define XFS_SCRUB_METAPATH_NR		(10)
 
 /*
  * ioctl limits
diff --git a/fs/xfs/scrub/metapath.c b/fs/xfs/scrub/metapath.c
index 74d71373e7edf1..e21c16fbd15d90 100644
--- a/fs/xfs/scrub/metapath.c
+++ b/fs/xfs/scrub/metapath.c
@@ -22,6 +22,7 @@
 #include "xfs_attr.h"
 #include "xfs_rtgroup.h"
 #include "xfs_rtrmap_btree.h"
+#include "xfs_rtrefcount_btree.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/trace.h"
@@ -249,6 +250,8 @@ xchk_setup_metapath(
 		return xchk_setup_metapath_dqinode(sc, XFS_DQTYPE_PROJ);
 	case XFS_SCRUB_METAPATH_RTRMAPBT:
 		return xchk_setup_metapath_rtginode(sc, XFS_RTGI_RMAP);
+	case XFS_SCRUB_METAPATH_RTREFCOUNTBT:
+		return xchk_setup_metapath_rtginode(sc, XFS_RTGI_REFCOUNT);
 	default:
 		return -ENOENT;
 	}

[PATCH 35/43] xfs: don't flag quota rt block usage on rtreflink filesystems

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Quota space usage is allowed to exceed the size of the physical storage
when reflink is enabled.  Now that we have reflink for the realtime
volume, apply this same logic to the rtb repair logic.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/scrub/quota_repair.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)


diff --git a/fs/xfs/scrub/quota_repair.c b/fs/xfs/scrub/quota_repair.c
index cd51f10f29209e..8f4c8d41f3083a 100644
--- a/fs/xfs/scrub/quota_repair.c
+++ b/fs/xfs/scrub/quota_repair.c
@@ -233,7 +233,7 @@ xrep_quota_item(
 		rqi->need_quotacheck = true;
 		dirty = true;
 	}
-	if (dq->q_rtb.count > mp->m_sb.sb_rblocks) {
+	if (!xfs_has_reflink(mp) && dq->q_rtb.count > mp->m_sb.sb_rblocks) {
 		dq->q_rtb.reserved -= dq->q_rtb.count;
 		dq->q_rtb.reserved += mp->m_sb.sb_rblocks;
 		dq->q_rtb.count = mp->m_sb.sb_rblocks;

[PATCH 36/43] xfs: check new rtbitmap records against rt refcount btree

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

When we're rebuilding the realtime bitmap, check the proposed free
extents against the rt refcount btree to make sure we don't commit any
grievous errors.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/scrub/repair.c          |    6 ++++++
 fs/xfs/scrub/rtbitmap_repair.c |   24 +++++++++++++++++++++++-
 2 files changed, 29 insertions(+), 1 deletion(-)


diff --git a/fs/xfs/scrub/repair.c b/fs/xfs/scrub/repair.c
index 61e414c81253af..3b5288d3ef4e34 100644
--- a/fs/xfs/scrub/repair.c
+++ b/fs/xfs/scrub/repair.c
@@ -42,6 +42,7 @@
 #include "xfs_rtgroup.h"
 #include "xfs_rtalloc.h"
 #include "xfs_metafile.h"
+#include "xfs_rtrefcount_btree.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/trace.h"
@@ -1009,6 +1010,11 @@ xrep_rtgroup_btcur_init(
 	    (sr->rtlock_flags & XFS_RTGLOCK_RMAP) &&
 	    xfs_has_rtrmapbt(mp))
 		sr->rmap_cur = xfs_rtrmapbt_init_cursor(sc->tp, sr->rtg);
+
+	if (sc->sm->sm_type != XFS_SCRUB_TYPE_RTREFCBT &&
+	    (sr->rtlock_flags & XFS_RTGLOCK_REFCOUNT) &&
+	    xfs_has_rtreflink(mp))
+		sr->refc_cur = xfs_rtrefcountbt_init_cursor(sc->tp, sr->rtg);
 }
 
 /*
diff --git a/fs/xfs/scrub/rtbitmap_repair.c b/fs/xfs/scrub/rtbitmap_repair.c
index c6e33834c5ae98..203a1a97c5026e 100644
--- a/fs/xfs/scrub/rtbitmap_repair.c
+++ b/fs/xfs/scrub/rtbitmap_repair.c
@@ -23,6 +23,7 @@
 #include "xfs_rtbitmap.h"
 #include "xfs_rtgroup.h"
 #include "xfs_extent_busy.h"
+#include "xfs_refcount.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/trace.h"
@@ -183,7 +184,8 @@ xrep_rtbitmap_mark_free(
 	xfs_rgblock_t		rgbno)
 {
 	struct xfs_mount	*mp = rtb->sc->mp;
-	struct xfs_rtgroup	*rtg = rtb->sc->sr.rtg;
+	struct xchk_rt		*sr = &rtb->sc->sr;
+	struct xfs_rtgroup	*rtg = sr->rtg;
 	xfs_rtxnum_t		startrtx;
 	xfs_rtxnum_t		nextrtx;
 	xrep_wordoff_t		wordoff, nextwordoff;
@@ -191,6 +193,7 @@ xrep_rtbitmap_mark_free(
 	unsigned int		bufwsize;
 	xfs_extlen_t		mod;
 	xfs_rtword_t		mask;
+	enum xbtree_recpacking	outcome;
 	int			error;
 
 	if (!xfs_verify_rgbext(rtg, rtb->next_rgbno, rgbno - rtb->next_rgbno))
@@ -210,6 +213,25 @@ xrep_rtbitmap_mark_free(
 	if (mod != mp->m_sb.sb_rextsize - 1)
 		return -EFSCORRUPTED;
 
+	/* Must not be shared or CoW staging. */
+	if (sr->refc_cur) {
+		error = xfs_refcount_has_records(sr->refc_cur,
+				XFS_REFC_DOMAIN_SHARED, rtb->next_rgbno,
+				rgbno - rtb->next_rgbno, &outcome);
+		if (error)
+			return error;
+		if (outcome != XBTREE_RECPACKING_EMPTY)
+			return -EFSCORRUPTED;
+
+		error = xfs_refcount_has_records(sr->refc_cur,
+				XFS_REFC_DOMAIN_COW, rtb->next_rgbno,
+				rgbno - rtb->next_rgbno, &outcome);
+		if (error)
+			return error;
+		if (outcome != XBTREE_RECPACKING_EMPTY)
+			return -EFSCORRUPTED;
+	}
+
 	trace_xrep_rtbitmap_record_free(mp, startrtx, nextrtx - 1);
 
 	/* Set bits as needed to round startrtx up to the nearest word. */

[PATCH 37/43] xfs: walk the rt reference count tree when rebuilding rmap

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

When we're rebuilding the data device rmap, if we encounter a "refcount"
format fork, we have to walk the (realtime) refcount btree inode to
build the appropriate mappings.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/scrub/rmap_repair.c |    7 +++++++
 1 file changed, 7 insertions(+)


diff --git a/fs/xfs/scrub/rmap_repair.c b/fs/xfs/scrub/rmap_repair.c
index c2c7b76cc25ab8..f5f73078ffe29d 100644
--- a/fs/xfs/scrub/rmap_repair.c
+++ b/fs/xfs/scrub/rmap_repair.c
@@ -33,6 +33,7 @@
 #include "xfs_ag.h"
 #include "xfs_rtrmap_btree.h"
 #include "xfs_rtgroup.h"
+#include "xfs_rtrefcount_btree.h"
 #include "scrub/xfs_scrub.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
@@ -519,6 +520,9 @@ xrep_rmap_scan_meta_btree(
 	case XFS_METAFILE_RTRMAP:
 		type = XFS_RTGI_RMAP;
 		break;
+	case XFS_METAFILE_RTREFCOUNT:
+		type = XFS_RTGI_REFCOUNT;
+		break;
 	default:
 		ASSERT(0);
 		return -EFSCORRUPTED;
@@ -545,6 +549,9 @@ xrep_rmap_scan_meta_btree(
 	case XFS_METAFILE_RTRMAP:
 		cur = xfs_rtrmapbt_init_cursor(sc->tp, rtg);
 		break;
+	case XFS_METAFILE_RTREFCOUNT:
+		cur = xfs_rtrefcountbt_init_cursor(sc->tp, rtg);
+		break;
 	default:
 		ASSERT(0);
 		error = -EFSCORRUPTED;

[PATCH 38/43] xfs: capture realtime CoW staging extents when rebuilding rt rmapbt

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Walk the realtime refcount btree to find the CoW staging extents when
we're rebuilding the realtime rmap btree.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/scrub/repair.h        |    1 
 fs/xfs/scrub/rgb_bitmap.h    |   37 +++++++++++++++
 fs/xfs/scrub/rtrmap_repair.c |  103 ++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 141 insertions(+)
 create mode 100644 fs/xfs/scrub/rgb_bitmap.h


diff --git a/fs/xfs/scrub/repair.h b/fs/xfs/scrub/repair.h
index ac5962732d269d..77343813205375 100644
--- a/fs/xfs/scrub/repair.h
+++ b/fs/xfs/scrub/repair.h
@@ -50,6 +50,7 @@ xrep_trans_commit(
 
 struct xbitmap;
 struct xagb_bitmap;
+struct xrgb_bitmap;
 struct xfsb_bitmap;
 
 int xrep_fix_freelist(struct xfs_scrub *sc, int alloc_flags);
diff --git a/fs/xfs/scrub/rgb_bitmap.h b/fs/xfs/scrub/rgb_bitmap.h
new file mode 100644
index 00000000000000..4c3126b66dcb96
--- /dev/null
+++ b/fs/xfs/scrub/rgb_bitmap.h
@@ -0,0 +1,37 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2020-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_SCRUB_RGB_BITMAP_H__
+#define __XFS_SCRUB_RGB_BITMAP_H__
+
+/* Bitmaps, but for type-checked for xfs_rgblock_t */
+
+struct xrgb_bitmap {
+	struct xbitmap32	rgbitmap;
+};
+
+static inline void xrgb_bitmap_init(struct xrgb_bitmap *bitmap)
+{
+	xbitmap32_init(&bitmap->rgbitmap);
+}
+
+static inline void xrgb_bitmap_destroy(struct xrgb_bitmap *bitmap)
+{
+	xbitmap32_destroy(&bitmap->rgbitmap);
+}
+
+static inline int xrgb_bitmap_set(struct xrgb_bitmap *bitmap,
+		xfs_rgblock_t start, xfs_extlen_t len)
+{
+	return xbitmap32_set(&bitmap->rgbitmap, start, len);
+}
+
+static inline int xrgb_bitmap_walk(struct xrgb_bitmap *bitmap,
+		xbitmap32_walk_fn fn, void *priv)
+{
+	return xbitmap32_walk(&bitmap->rgbitmap, fn, priv);
+}
+
+#endif /* __XFS_SCRUB_RGB_BITMAP_H__ */
diff --git a/fs/xfs/scrub/rtrmap_repair.c b/fs/xfs/scrub/rtrmap_repair.c
index 49de8bc2dd17f5..f2fdd7a9fc2483 100644
--- a/fs/xfs/scrub/rtrmap_repair.c
+++ b/fs/xfs/scrub/rtrmap_repair.c
@@ -30,6 +30,7 @@
 #include "xfs_rtalloc.h"
 #include "xfs_ag.h"
 #include "xfs_rtgroup.h"
+#include "xfs_refcount.h"
 #include "scrub/xfs_scrub.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
@@ -38,6 +39,7 @@
 #include "scrub/repair.h"
 #include "scrub/bitmap.h"
 #include "scrub/fsb_bitmap.h"
+#include "scrub/rgb_bitmap.h"
 #include "scrub/xfile.h"
 #include "scrub/xfarray.h"
 #include "scrub/iscan.h"
@@ -423,6 +425,100 @@ xrep_rtrmap_scan_ag(
 	return error;
 }
 
+struct xrep_rtrmap_stash_run {
+	struct xrep_rtrmap	*rr;
+	uint64_t		owner;
+};
+
+static int
+xrep_rtrmap_stash_run(
+	uint32_t			start,
+	uint32_t			len,
+	void				*priv)
+{
+	struct xrep_rtrmap_stash_run	*rsr = priv;
+	struct xrep_rtrmap		*rr = rsr->rr;
+	xfs_rgblock_t			rgbno = start;
+
+	return xrep_rtrmap_stash(rr, rgbno, len, rsr->owner, 0, 0);
+}
+
+/*
+ * Emit rmaps for every extent of bits set in the bitmap.  Caller must ensure
+ * that the ranges are in units of FS blocks.
+ */
+STATIC int
+xrep_rtrmap_stash_bitmap(
+	struct xrep_rtrmap		*rr,
+	struct xrgb_bitmap		*bitmap,
+	const struct xfs_owner_info	*oinfo)
+{
+	struct xrep_rtrmap_stash_run	rsr = {
+		.rr			= rr,
+		.owner			= oinfo->oi_owner,
+	};
+
+	return xrgb_bitmap_walk(bitmap, xrep_rtrmap_stash_run, &rsr);
+}
+
+/* Record a CoW staging extent. */
+STATIC int
+xrep_rtrmap_walk_cowblocks(
+	struct xfs_btree_cur		*cur,
+	const struct xfs_refcount_irec	*irec,
+	void				*priv)
+{
+	struct xrgb_bitmap		*bitmap = priv;
+
+	if (!xfs_refcount_check_domain(irec) ||
+	    irec->rc_domain != XFS_REFC_DOMAIN_COW)
+		return -EFSCORRUPTED;
+
+	return xrgb_bitmap_set(bitmap, irec->rc_startblock,
+			irec->rc_blockcount);
+}
+
+/*
+ * Collect rmaps for the blocks containing the refcount btree, and all CoW
+ * staging extents.
+ */
+STATIC int
+xrep_rtrmap_find_refcount_rmaps(
+	struct xrep_rtrmap	*rr)
+{
+	struct xrgb_bitmap	cow_blocks;		/* COWBIT */
+	struct xfs_refcount_irec low = {
+		.rc_startblock	= 0,
+		.rc_domain	= XFS_REFC_DOMAIN_COW,
+	};
+	struct xfs_refcount_irec high = {
+		.rc_startblock	= -1U,
+		.rc_domain	= XFS_REFC_DOMAIN_COW,
+	};
+	struct xfs_scrub	*sc = rr->sc;
+	int			error;
+
+	if (!xfs_has_rtreflink(sc->mp))
+		return 0;
+
+	xrgb_bitmap_init(&cow_blocks);
+
+	/* Collect rmaps for CoW staging extents. */
+	error = xfs_refcount_query_range(sc->sr.refc_cur, &low, &high,
+			xrep_rtrmap_walk_cowblocks, &cow_blocks);
+	if (error)
+		goto out_bitmap;
+
+	/* Generate rmaps for everything. */
+	error = xrep_rtrmap_stash_bitmap(rr, &cow_blocks, &XFS_RMAP_OINFO_COW);
+	if (error)
+		goto out_bitmap;
+
+out_bitmap:
+	xrgb_bitmap_destroy(&cow_blocks);
+	return error;
+}
+
 /* Count and check all collected records. */
 STATIC int
 xrep_rtrmap_check_record(
@@ -460,6 +556,13 @@ xrep_rtrmap_find_rmaps(
 			return error;
 	}
 
+	/* Find CoW staging extents. */
+	xrep_rtgroup_btcur_init(sc, &sc->sr);
+	error = xrep_rtrmap_find_refcount_rmaps(rr);
+	xchk_rtgroup_btcur_free(&sc->sr);
+	if (error)
+		return error;
+
 	/*
 	 * Set up for a potentially lengthy filesystem scan by reducing our
 	 * transaction resource usage for the duration.  Specifically:

[PATCH 39/43] xfs: online repair of the realtime refcount btree

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Port the data device's refcount btree repair code to the realtime
refcount btree.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/Makefile                  |    1 
 fs/xfs/scrub/refcount_repair.c   |    2 
 fs/xfs/scrub/repair.h            |    5 
 fs/xfs/scrub/rtrefcount.c        |    9 
 fs/xfs/scrub/rtrefcount_repair.c |  783 ++++++++++++++++++++++++++++++++++++++
 fs/xfs/scrub/scrub.c             |    2 
 fs/xfs/scrub/trace.h             |   14 -
 7 files changed, 809 insertions(+), 7 deletions(-)
 create mode 100644 fs/xfs/scrub/rtrefcount_repair.c


diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile
index 9dd9921e53567c..7afa51e414278e 100644
--- a/fs/xfs/Makefile
+++ b/fs/xfs/Makefile
@@ -236,6 +236,7 @@ xfs-y				+= $(addprefix scrub/, \
 
 xfs-$(CONFIG_XFS_RT)		+= $(addprefix scrub/, \
 				   rtbitmap_repair.o \
+				   rtrefcount_repair.o \
 				   rtrmap_repair.o \
 				   rtsummary_repair.o \
 				   )
diff --git a/fs/xfs/scrub/refcount_repair.c b/fs/xfs/scrub/refcount_repair.c
index 1ee6d4aeb308f5..9c8cb5332da042 100644
--- a/fs/xfs/scrub/refcount_repair.c
+++ b/fs/xfs/scrub/refcount_repair.c
@@ -189,7 +189,7 @@ xrep_refc_stash(
 	if (error)
 		return error;
 
-	trace_xrep_refc_found(sc->sa.pag, &irec);
+	trace_xrep_refc_found(pag_group(sc->sa.pag), &irec);
 
 	return xfarray_append(rr->refcount_records, &irec);
 }
diff --git a/fs/xfs/scrub/repair.h b/fs/xfs/scrub/repair.h
index 77343813205375..8f8f18b48a449d 100644
--- a/fs/xfs/scrub/repair.h
+++ b/fs/xfs/scrub/repair.h
@@ -99,6 +99,7 @@ int xrep_setup_nlinks(struct xfs_scrub *sc);
 int xrep_setup_symlink(struct xfs_scrub *sc, unsigned int *resblks);
 int xrep_setup_dirtree(struct xfs_scrub *sc);
 int xrep_setup_rtrmapbt(struct xfs_scrub *sc);
+int xrep_setup_rtrefcountbt(struct xfs_scrub *sc);
 
 /* Repair setup functions */
 int xrep_setup_ag_allocbt(struct xfs_scrub *sc);
@@ -158,11 +159,13 @@ int xrep_rtbitmap(struct xfs_scrub *sc);
 int xrep_rtsummary(struct xfs_scrub *sc);
 int xrep_rgsuperblock(struct xfs_scrub *sc);
 int xrep_rtrmapbt(struct xfs_scrub *sc);
+int xrep_rtrefcountbt(struct xfs_scrub *sc);
 #else
 # define xrep_rtbitmap			xrep_notsupported
 # define xrep_rtsummary			xrep_notsupported
 # define xrep_rgsuperblock		xrep_notsupported
 # define xrep_rtrmapbt			xrep_notsupported
+# define xrep_rtrefcountbt		xrep_notsupported
 #endif /* CONFIG_XFS_RT */
 
 #ifdef CONFIG_XFS_QUOTA
@@ -236,6 +239,7 @@ xrep_setup_nothing(
 #define xrep_setup_dirtree		xrep_setup_nothing
 #define xrep_setup_metapath		xrep_setup_nothing
 #define xrep_setup_rtrmapbt		xrep_setup_nothing
+#define xrep_setup_rtrefcountbt		xrep_setup_nothing
 
 #define xrep_setup_inode(sc, imap)	((void)0)
 
@@ -274,6 +278,7 @@ static inline int xrep_setup_symlink(struct xfs_scrub *sc, unsigned int *x)
 #define xrep_metapath			xrep_notsupported
 #define xrep_rgsuperblock		xrep_notsupported
 #define xrep_rtrmapbt			xrep_notsupported
+#define xrep_rtrefcountbt		xrep_notsupported
 
 #endif /* CONFIG_XFS_ONLINE_REPAIR */
 
diff --git a/fs/xfs/scrub/rtrefcount.c b/fs/xfs/scrub/rtrefcount.c
index 18c9bcb0e82184..4c5dffc73641b7 100644
--- a/fs/xfs/scrub/rtrefcount.c
+++ b/fs/xfs/scrub/rtrefcount.c
@@ -7,8 +7,10 @@
 #include "xfs_fs.h"
 #include "xfs_shared.h"
 #include "xfs_format.h"
+#include "xfs_log_format.h"
 #include "xfs_trans_resv.h"
 #include "xfs_mount.h"
+#include "xfs_trans.h"
 #include "xfs_btree.h"
 #include "xfs_rmap.h"
 #include "xfs_refcount.h"
@@ -21,6 +23,7 @@
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/btree.h"
+#include "scrub/repair.h"
 
 /* Set us up with the realtime refcount metadata locked. */
 int
@@ -32,6 +35,12 @@ xchk_setup_rtrefcountbt(
 	if (xchk_need_intent_drain(sc))
 		xchk_fsgates_enable(sc, XCHK_FSGATES_DRAIN);
 
+	if (xchk_could_repair(sc)) {
+		error = xrep_setup_rtrefcountbt(sc);
+		if (error)
+			return error;
+	}
+
 	error = xchk_rtgroup_init(sc, sc->sm->sm_agno, &sc->sr);
 	if (error)
 		return error;
diff --git a/fs/xfs/scrub/rtrefcount_repair.c b/fs/xfs/scrub/rtrefcount_repair.c
new file mode 100644
index 00000000000000..257cfb24beb428
--- /dev/null
+++ b/fs/xfs/scrub/rtrefcount_repair.c
@@ -0,0 +1,783 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2021-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#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_btree_staging.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_alloc.h"
+#include "xfs_ialloc.h"
+#include "xfs_rmap.h"
+#include "xfs_rmap_btree.h"
+#include "xfs_rtrmap_btree.h"
+#include "xfs_refcount.h"
+#include "xfs_rtrefcount_btree.h"
+#include "xfs_error.h"
+#include "xfs_health.h"
+#include "xfs_inode.h"
+#include "xfs_quota.h"
+#include "xfs_rtalloc.h"
+#include "xfs_ag.h"
+#include "xfs_rtgroup.h"
+#include "xfs_rtbitmap.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/btree.h"
+#include "scrub/trace.h"
+#include "scrub/repair.h"
+#include "scrub/bitmap.h"
+#include "scrub/fsb_bitmap.h"
+#include "scrub/xfile.h"
+#include "scrub/xfarray.h"
+#include "scrub/newbt.h"
+#include "scrub/reap.h"
+#include "scrub/rcbag.h"
+
+/*
+ * Rebuilding the Reference Count Btree
+ * ====================================
+ *
+ * This algorithm is "borrowed" from xfs_repair.  Imagine the rmap
+ * entries as rectangles representing extents of physical blocks, and
+ * that the rectangles can be laid down to allow them to overlap each
+ * other; then we know that we must emit a refcnt btree entry wherever
+ * the amount of overlap changes, i.e. the emission stimulus is
+ * level-triggered:
+ *
+ *                 -    ---
+ *       --      ----- ----   ---        ------
+ * --   ----     ----------- ----     ---------
+ * -------------------------------- -----------
+ * ^ ^  ^^ ^^    ^ ^^ ^^^  ^^^^  ^ ^^ ^  ^     ^
+ * 2 1  23 21    3 43 234  2123  1 01 2  3     0
+ *
+ * For our purposes, a rmap is a tuple (startblock, len, fileoff, owner).
+ *
+ * Note that in the actual refcnt btree we don't store the refcount < 2
+ * cases because the bnobt tells us which blocks are free; single-use
+ * blocks aren't recorded in the bnobt or the refcntbt.  If the rmapbt
+ * supports storing multiple entries covering a given block we could
+ * theoretically dispense with the refcntbt and simply count rmaps, but
+ * that's inefficient in the (hot) write path, so we'll take the cost of
+ * the extra tree to save time.  Also there's no guarantee that rmap
+ * will be enabled.
+ *
+ * Given an array of rmaps sorted by physical block number, a starting
+ * physical block (sp), a bag to hold rmaps that cover sp, and the next
+ * physical block where the level changes (np), we can reconstruct the
+ * rt refcount btree as follows:
+ *
+ * While there are still unprocessed rmaps in the array,
+ *  - Set sp to the physical block (pblk) of the next unprocessed rmap.
+ *  - Add to the bag all rmaps in the array where startblock == sp.
+ *  - Set np to the physical block where the bag size will change.  This
+ *    is the minimum of (the pblk of the next unprocessed rmap) and
+ *    (startblock + len of each rmap in the bag).
+ *  - Record the bag size as old_bag_size.
+ *
+ *  - While the bag isn't empty,
+ *     - Remove from the bag all rmaps where startblock + len == np.
+ *     - Add to the bag all rmaps in the array where startblock == np.
+ *     - If the bag size isn't old_bag_size, store the refcount entry
+ *       (sp, np - sp, bag_size) in the refcnt btree.
+ *     - If the bag is empty, break out of the inner loop.
+ *     - Set old_bag_size to the bag size
+ *     - Set sp = np.
+ *     - Set np to the physical block where the bag size will change.
+ *       This is the minimum of (the pblk of the next unprocessed rmap)
+ *       and (startblock + len of each rmap in the bag).
+ *
+ * Like all the other repairers, we make a list of all the refcount
+ * records we need, then reinitialize the rt refcount btree root and
+ * insert all the records.
+ */
+
+struct xrep_rtrefc {
+	/* refcount extents */
+	struct xfarray		*refcount_records;
+
+	/* new refcountbt information */
+	struct xrep_newbt	new_btree;
+
+	/* old refcountbt blocks */
+	struct xfsb_bitmap	old_rtrefcountbt_blocks;
+
+	struct xfs_scrub	*sc;
+
+	/* get_records()'s position in the rt refcount record array. */
+	xfarray_idx_t		array_cur;
+
+	/* # of refcountbt blocks */
+	xfs_filblks_t		btblocks;
+};
+
+/* Set us up to repair refcount btrees. */
+int
+xrep_setup_rtrefcountbt(
+	struct xfs_scrub	*sc)
+{
+	char			*descr;
+	int			error;
+
+	descr = xchk_xfile_ag_descr(sc, "rmap record bag");
+	error = xrep_setup_xfbtree(sc, descr);
+	kfree(descr);
+	return error;
+}
+
+/* Check for any obvious conflicts with this shared/CoW staging extent. */
+STATIC int
+xrep_rtrefc_check_ext(
+	struct xfs_scrub		*sc,
+	const struct xfs_refcount_irec	*rec)
+{
+	xfs_rgblock_t			last;
+
+	if (xfs_rtrefcount_check_irec(sc->sr.rtg, rec) != NULL)
+		return -EFSCORRUPTED;
+
+	if (xfs_rgbno_to_rtxoff(sc->mp, rec->rc_startblock) != 0)
+		return -EFSCORRUPTED;
+
+	last = rec->rc_startblock + rec->rc_blockcount - 1;
+	if (xfs_rgbno_to_rtxoff(sc->mp, last) != sc->mp->m_sb.sb_rextsize - 1)
+		return -EFSCORRUPTED;
+
+	/* Make sure this isn't free space or misaligned. */
+	return xrep_require_rtext_inuse(sc, rec->rc_startblock,
+			rec->rc_blockcount);
+}
+
+/* Record a reference count extent. */
+STATIC int
+xrep_rtrefc_stash(
+	struct xrep_rtrefc		*rr,
+	enum xfs_refc_domain		domain,
+	xfs_rgblock_t			bno,
+	xfs_extlen_t			len,
+	uint64_t			refcount)
+{
+	struct xfs_refcount_irec	irec = {
+		.rc_startblock		= bno,
+		.rc_blockcount		= len,
+		.rc_refcount		= refcount,
+		.rc_domain		= domain,
+	};
+	int				error = 0;
+
+	if (xchk_should_terminate(rr->sc, &error))
+		return error;
+
+	irec.rc_refcount = min_t(uint64_t, XFS_REFC_REFCOUNT_MAX, refcount);
+
+	error = xrep_rtrefc_check_ext(rr->sc, &irec);
+	if (error)
+		return error;
+
+	trace_xrep_refc_found(rtg_group(rr->sc->sr.rtg), &irec);
+
+	return xfarray_append(rr->refcount_records, &irec);
+}
+
+/* Record a CoW staging extent. */
+STATIC int
+xrep_rtrefc_stash_cow(
+	struct xrep_rtrefc		*rr,
+	xfs_rgblock_t			bno,
+	xfs_extlen_t			len)
+{
+	return xrep_rtrefc_stash(rr, XFS_REFC_DOMAIN_COW, bno, len, 1);
+}
+
+/* Decide if an rmap could describe a shared extent. */
+static inline bool
+xrep_rtrefc_rmap_shareable(
+	const struct xfs_rmap_irec	*rmap)
+{
+	/* rt metadata are never sharable */
+	if (XFS_RMAP_NON_INODE_OWNER(rmap->rm_owner))
+		return false;
+
+	/* Unwritten file blocks are not shareable. */
+	if (rmap->rm_flags & XFS_RMAP_UNWRITTEN)
+		return false;
+
+	return true;
+}
+
+/* Grab the next (abbreviated) rmap record from the rmapbt. */
+STATIC int
+xrep_rtrefc_walk_rmaps(
+	struct xrep_rtrefc	*rr,
+	struct xfs_rmap_irec	*rmap,
+	bool			*have_rec)
+{
+	struct xfs_btree_cur	*cur = rr->sc->sr.rmap_cur;
+	struct xfs_mount	*mp = cur->bc_mp;
+	int			have_gt;
+	int			error = 0;
+
+	*have_rec = false;
+
+	/*
+	 * Loop through the remaining rmaps.  Remember CoW staging
+	 * extents and the refcountbt blocks from the old tree for later
+	 * disposal.  We can only share written data fork extents, so
+	 * keep looping until we find an rmap for one.
+	 */
+	do {
+		if (xchk_should_terminate(rr->sc, &error))
+			return error;
+
+		error = xfs_btree_increment(cur, 0, &have_gt);
+		if (error)
+			return error;
+		if (!have_gt)
+			return 0;
+
+		error = xfs_rmap_get_rec(cur, rmap, &have_gt);
+		if (error)
+			return error;
+		if (XFS_IS_CORRUPT(mp, !have_gt)) {
+			xfs_btree_mark_sick(cur);
+			return -EFSCORRUPTED;
+		}
+
+		if (rmap->rm_owner == XFS_RMAP_OWN_COW) {
+			error = xrep_rtrefc_stash_cow(rr, rmap->rm_startblock,
+					rmap->rm_blockcount);
+			if (error)
+				return error;
+		} else if (xfs_is_sb_inum(mp, rmap->rm_owner) ||
+			   (rmap->rm_flags & (XFS_RMAP_ATTR_FORK |
+					      XFS_RMAP_BMBT_BLOCK))) {
+			xfs_btree_mark_sick(cur);
+			return -EFSCORRUPTED;
+		}
+	} while (!xrep_rtrefc_rmap_shareable(rmap));
+
+	*have_rec = true;
+	return 0;
+}
+
+static inline uint32_t
+xrep_rtrefc_encode_startblock(
+	const struct xfs_refcount_irec	*irec)
+{
+	uint32_t			start;
+
+	start = irec->rc_startblock & ~XFS_REFC_COWFLAG;
+	if (irec->rc_domain == XFS_REFC_DOMAIN_COW)
+		start |= XFS_REFC_COWFLAG;
+
+	return start;
+}
+
+/*
+ * Compare two refcount records.  We want to sort in order of increasing block
+ * number.
+ */
+static int
+xrep_rtrefc_extent_cmp(
+	const void			*a,
+	const void			*b)
+{
+	const struct xfs_refcount_irec	*ap = a;
+	const struct xfs_refcount_irec	*bp = b;
+	uint32_t			sa, sb;
+
+	sa = xrep_rtrefc_encode_startblock(ap);
+	sb = xrep_rtrefc_encode_startblock(bp);
+
+	if (sa > sb)
+		return 1;
+	if (sa < sb)
+		return -1;
+	return 0;
+}
+
+/*
+ * Sort the refcount extents by startblock or else the btree records will be in
+ * the wrong order.  Make sure the records do not overlap in physical space.
+ */
+STATIC int
+xrep_rtrefc_sort_records(
+	struct xrep_rtrefc		*rr)
+{
+	struct xfs_refcount_irec	irec;
+	xfarray_idx_t			cur;
+	enum xfs_refc_domain		dom = XFS_REFC_DOMAIN_SHARED;
+	xfs_rgblock_t			next_rgbno = 0;
+	int				error;
+
+	error = xfarray_sort(rr->refcount_records, xrep_rtrefc_extent_cmp,
+			XFARRAY_SORT_KILLABLE);
+	if (error)
+		return error;
+
+	foreach_xfarray_idx(rr->refcount_records, cur) {
+		if (xchk_should_terminate(rr->sc, &error))
+			return error;
+
+		error = xfarray_load(rr->refcount_records, cur, &irec);
+		if (error)
+			return error;
+
+		if (dom == XFS_REFC_DOMAIN_SHARED &&
+		    irec.rc_domain == XFS_REFC_DOMAIN_COW) {
+			dom = irec.rc_domain;
+			next_rgbno = 0;
+		}
+
+		if (dom != irec.rc_domain)
+			return -EFSCORRUPTED;
+		if (irec.rc_startblock < next_rgbno)
+			return -EFSCORRUPTED;
+
+		next_rgbno = irec.rc_startblock + irec.rc_blockcount;
+	}
+
+	return error;
+}
+
+/* Record extents that belong to the realtime refcount inode. */
+STATIC int
+xrep_rtrefc_walk_rmap(
+	struct xfs_btree_cur		*cur,
+	const struct xfs_rmap_irec	*rec,
+	void				*priv)
+{
+	struct xrep_rtrefc		*rr = priv;
+	int				error = 0;
+
+	if (xchk_should_terminate(rr->sc, &error))
+		return error;
+
+	/* Skip extents which are not owned by this inode and fork. */
+	if (rec->rm_owner != rr->sc->ip->i_ino)
+		return 0;
+
+	error = xrep_check_ino_btree_mapping(rr->sc, rec);
+	if (error)
+		return error;
+
+	return xfsb_bitmap_set(&rr->old_rtrefcountbt_blocks,
+			xfs_gbno_to_fsb(cur->bc_group, rec->rm_startblock),
+			rec->rm_blockcount);
+}
+
+/*
+ * Walk forward through the rmap btree to collect all rmaps starting at
+ * @bno in @rmap_bag.  These represent the file(s) that share ownership of
+ * the current block.  Upon return, the rmap cursor points to the last record
+ * satisfying the startblock constraint.
+ */
+static int
+xrep_rtrefc_push_rmaps_at(
+	struct xrep_rtrefc	*rr,
+	struct rcbag		*rcstack,
+	xfs_rgblock_t		bno,
+	struct xfs_rmap_irec	*rmap,
+	bool			*have)
+{
+	struct xfs_scrub	*sc = rr->sc;
+	int			have_gt;
+	int			error;
+
+	while (*have && rmap->rm_startblock == bno) {
+		error = rcbag_add(rcstack, rr->sc->tp, rmap);
+		if (error)
+			return error;
+
+		error = xrep_rtrefc_walk_rmaps(rr, rmap, have);
+		if (error)
+			return error;
+	}
+
+	error = xfs_btree_decrement(sc->sr.rmap_cur, 0, &have_gt);
+	if (error)
+		return error;
+	if (XFS_IS_CORRUPT(sc->mp, !have_gt)) {
+		xfs_btree_mark_sick(sc->sr.rmap_cur);
+		return -EFSCORRUPTED;
+	}
+
+	return 0;
+}
+
+/* Scan one AG for reverse mappings for the realtime refcount btree. */
+STATIC int
+xrep_rtrefc_scan_ag(
+	struct xrep_rtrefc	*rr,
+	struct xfs_perag	*pag)
+{
+	struct xfs_scrub	*sc = rr->sc;
+	int			error;
+
+	error = xrep_ag_init(sc, pag, &sc->sa);
+	if (error)
+		return error;
+
+	error = xfs_rmap_query_all(sc->sa.rmap_cur, xrep_rtrefc_walk_rmap, rr);
+	xchk_ag_free(sc, &sc->sa);
+	return error;
+}
+
+/* Iterate all the rmap records to generate reference count data. */
+STATIC int
+xrep_rtrefc_find_refcounts(
+	struct xrep_rtrefc	*rr)
+{
+	struct xfs_scrub	*sc = rr->sc;
+	struct rcbag		*rcstack;
+	struct xfs_perag	*pag = NULL;
+	uint64_t		old_stack_height;
+	xfs_rgblock_t		sbno;
+	xfs_rgblock_t		cbno;
+	xfs_rgblock_t		nbno;
+	bool			have;
+	int			error;
+
+	/* Scan for old rtrefc btree blocks. */
+	while ((pag = xfs_perag_next(sc->mp, pag))) {
+		error = xrep_rtrefc_scan_ag(rr, pag);
+		if (error) {
+			xfs_perag_rele(pag);
+			return error;
+		}
+	}
+
+	xrep_rtgroup_btcur_init(sc, &sc->sr);
+
+	/*
+	 * Set up a bag to store all the rmap records that we're tracking to
+	 * generate a reference count record.  If this exceeds
+	 * XFS_REFC_REFCOUNT_MAX, we clamp rc_refcount.
+	 */
+	error = rcbag_init(sc->mp, sc->xmbtp, &rcstack);
+	if (error)
+		goto out_cur;
+
+	/* Start the rtrmapbt cursor to the left of all records. */
+	error = xfs_btree_goto_left_edge(sc->sr.rmap_cur);
+	if (error)
+		goto out_bag;
+
+	/* Process reverse mappings into refcount data. */
+	while (xfs_btree_has_more_records(sc->sr.rmap_cur)) {
+		struct xfs_rmap_irec	rmap;
+
+		/* Push all rmaps with pblk == sbno onto the stack */
+		error = xrep_rtrefc_walk_rmaps(rr, &rmap, &have);
+		if (error)
+			goto out_bag;
+		if (!have)
+			break;
+		sbno = cbno = rmap.rm_startblock;
+		error = xrep_rtrefc_push_rmaps_at(rr, rcstack, sbno, &rmap,
+				&have);
+		if (error)
+			goto out_bag;
+
+		/* Set nbno to the bno of the next refcount change */
+		error = rcbag_next_edge(rcstack, sc->tp, &rmap, have, &nbno);
+		if (error)
+			goto out_bag;
+
+		ASSERT(nbno > sbno);
+		old_stack_height = rcbag_count(rcstack);
+
+		/* While stack isn't empty... */
+		while (rcbag_count(rcstack) > 0) {
+			/* Pop all rmaps that end at nbno */
+			error = rcbag_remove_ending_at(rcstack, sc->tp, nbno);
+			if (error)
+				goto out_bag;
+
+			/* Push array items that start at nbno */
+			error = xrep_rtrefc_walk_rmaps(rr, &rmap, &have);
+			if (error)
+				goto out_bag;
+			if (have) {
+				error = xrep_rtrefc_push_rmaps_at(rr, rcstack,
+						nbno, &rmap, &have);
+				if (error)
+					goto out_bag;
+			}
+
+			/* Emit refcount if necessary */
+			ASSERT(nbno > cbno);
+			if (rcbag_count(rcstack) != old_stack_height) {
+				if (old_stack_height > 1) {
+					error = xrep_rtrefc_stash(rr,
+							XFS_REFC_DOMAIN_SHARED,
+							cbno, nbno - cbno,
+							old_stack_height);
+					if (error)
+						goto out_bag;
+				}
+				cbno = nbno;
+			}
+
+			/* Stack empty, go find the next rmap */
+			if (rcbag_count(rcstack) == 0)
+				break;
+			old_stack_height = rcbag_count(rcstack);
+			sbno = nbno;
+
+			/* Set nbno to the bno of the next refcount change */
+			error = rcbag_next_edge(rcstack, sc->tp, &rmap, have,
+					&nbno);
+			if (error)
+				goto out_bag;
+
+			ASSERT(nbno > sbno);
+		}
+	}
+
+	ASSERT(rcbag_count(rcstack) == 0);
+out_bag:
+	rcbag_free(&rcstack);
+out_cur:
+	xchk_rtgroup_btcur_free(&sc->sr);
+	return error;
+}
+
+/* Retrieve refcountbt data for bulk load. */
+STATIC int
+xrep_rtrefc_get_records(
+	struct xfs_btree_cur		*cur,
+	unsigned int			idx,
+	struct xfs_btree_block		*block,
+	unsigned int			nr_wanted,
+	void				*priv)
+{
+	struct xrep_rtrefc		*rr = priv;
+	union xfs_btree_rec		*block_rec;
+	unsigned int			loaded;
+	int				error;
+
+	for (loaded = 0; loaded < nr_wanted; loaded++, idx++) {
+		error = xfarray_load(rr->refcount_records, rr->array_cur++,
+				&cur->bc_rec.rc);
+		if (error)
+			return error;
+
+		block_rec = xfs_btree_rec_addr(cur, idx, block);
+		cur->bc_ops->init_rec_from_cur(cur, block_rec);
+	}
+
+	return loaded;
+}
+
+/* Feed one of the new btree blocks to the bulk loader. */
+STATIC int
+xrep_rtrefc_claim_block(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr,
+	void			*priv)
+{
+	struct xrep_rtrefc	*rr = priv;
+
+	return xrep_newbt_claim_block(cur, &rr->new_btree, ptr);
+}
+
+/* Figure out how much space we need to create the incore btree root block. */
+STATIC size_t
+xrep_rtrefc_iroot_size(
+	struct xfs_btree_cur	*cur,
+	unsigned int		level,
+	unsigned int		nr_this_level,
+	void			*priv)
+{
+	return xfs_rtrefcount_broot_space_calc(cur->bc_mp, level,
+			nr_this_level);
+}
+
+/*
+ * Use the collected refcount information to stage a new rt refcount btree.  If
+ * this is successful we'll return with the new btree root information logged
+ * to the repair transaction but not yet committed.
+ */
+STATIC int
+xrep_rtrefc_build_new_tree(
+	struct xrep_rtrefc	*rr)
+{
+	struct xfs_scrub	*sc = rr->sc;
+	struct xfs_rtgroup	*rtg = sc->sr.rtg;
+	struct xfs_btree_cur	*refc_cur;
+	int			error;
+
+	error = xrep_rtrefc_sort_records(rr);
+	if (error)
+		return error;
+
+	/*
+	 * Prepare to construct the new btree by reserving disk space for the
+	 * new btree and setting up all the accounting information we'll need
+	 * to root the new btree while it's under construction and before we
+	 * attach it to the realtime refcount inode.
+	 */
+	error = xrep_newbt_init_metadir_inode(&rr->new_btree, sc);
+	if (error)
+		return error;
+
+	rr->new_btree.bload.get_records = xrep_rtrefc_get_records;
+	rr->new_btree.bload.claim_block = xrep_rtrefc_claim_block;
+	rr->new_btree.bload.iroot_size = xrep_rtrefc_iroot_size;
+
+	refc_cur = xfs_rtrefcountbt_init_cursor(NULL, rtg);
+	xfs_btree_stage_ifakeroot(refc_cur, &rr->new_btree.ifake);
+
+	/* Compute how many blocks we'll need. */
+	error = xfs_btree_bload_compute_geometry(refc_cur, &rr->new_btree.bload,
+			xfarray_length(rr->refcount_records));
+	if (error)
+		goto err_cur;
+
+	/* Last chance to abort before we start committing fixes. */
+	if (xchk_should_terminate(sc, &error))
+		goto err_cur;
+
+	/*
+	 * Guess how many blocks we're going to need to rebuild an entire
+	 * rtrefcountbt from the number of extents we found, and pump up our
+	 * transaction to have sufficient block reservation.  We're allowed
+	 * to exceed quota to repair inconsistent metadata, though this is
+	 * unlikely.
+	 */
+	error = xfs_trans_reserve_more_inode(sc->tp, rtg_refcount(rtg),
+			rr->new_btree.bload.nr_blocks, 0, true);
+	if (error)
+		goto err_cur;
+
+	/* Reserve the space we'll need for the new btree. */
+	error = xrep_newbt_alloc_blocks(&rr->new_btree,
+			rr->new_btree.bload.nr_blocks);
+	if (error)
+		goto err_cur;
+
+	/* Add all observed refcount records. */
+	rr->new_btree.ifake.if_fork->if_format = XFS_DINODE_FMT_META_BTREE;
+	rr->array_cur = XFARRAY_CURSOR_INIT;
+	error = xfs_btree_bload(refc_cur, &rr->new_btree.bload, rr);
+	if (error)
+		goto err_cur;
+
+	/*
+	 * Install the new rtrefc btree in the inode.  After this point the old
+	 * btree is no longer accessible, the new tree is live, and we can
+	 * delete the cursor.
+	 */
+	xfs_rtrefcountbt_commit_staged_btree(refc_cur, sc->tp);
+	xrep_inode_set_nblocks(rr->sc, rr->new_btree.ifake.if_blocks);
+	xfs_btree_del_cursor(refc_cur, 0);
+
+	/* Dispose of any unused blocks and the accounting information. */
+	error = xrep_newbt_commit(&rr->new_btree);
+	if (error)
+		return error;
+
+	return xrep_roll_trans(sc);
+err_cur:
+	xfs_btree_del_cursor(refc_cur, error);
+	xrep_newbt_cancel(&rr->new_btree);
+	return error;
+}
+
+/*
+ * Now that we've logged the roots of the new btrees, invalidate all of the
+ * old blocks and free them.
+ */
+STATIC int
+xrep_rtrefc_remove_old_tree(
+	struct xrep_rtrefc	*rr)
+{
+	int			error;
+
+	/*
+	 * Free all the extents that were allocated to the former rtrefcountbt
+	 * and aren't cross-linked with something else.
+	 */
+	error = xrep_reap_metadir_fsblocks(rr->sc,
+			&rr->old_rtrefcountbt_blocks);
+	if (error)
+		return error;
+
+	/*
+	 * Ensure the proper reservation for the rtrefcount inode so that we
+	 * don't fail to expand the btree.
+	 */
+	return xrep_reset_metafile_resv(rr->sc);
+}
+
+/* Rebuild the rt refcount btree. */
+int
+xrep_rtrefcountbt(
+	struct xfs_scrub	*sc)
+{
+	struct xrep_rtrefc	*rr;
+	struct xfs_mount	*mp = sc->mp;
+	char			*descr;
+	int			error;
+
+	/* We require the rmapbt to rebuild anything. */
+	if (!xfs_has_rtrmapbt(mp))
+		return -EOPNOTSUPP;
+
+	/* Make sure any problems with the fork are fixed. */
+	error = xrep_metadata_inode_forks(sc);
+	if (error)
+		return error;
+
+	rr = kzalloc(sizeof(struct xrep_rtrefc), XCHK_GFP_FLAGS);
+	if (!rr)
+		return -ENOMEM;
+	rr->sc = sc;
+
+	/* Set up enough storage to handle one refcount record per rt extent. */
+	descr = xchk_xfile_ag_descr(sc, "reference count records");
+	error = xfarray_create(descr, mp->m_sb.sb_rextents,
+			sizeof(struct xfs_refcount_irec),
+			&rr->refcount_records);
+	kfree(descr);
+	if (error)
+		goto out_rr;
+
+	/* Collect all reference counts. */
+	xfsb_bitmap_init(&rr->old_rtrefcountbt_blocks);
+	error = xrep_rtrefc_find_refcounts(rr);
+	if (error)
+		goto out_bitmap;
+
+	xfs_trans_ijoin(sc->tp, sc->ip, 0);
+
+	/* Rebuild the refcount information. */
+	error = xrep_rtrefc_build_new_tree(rr);
+	if (error)
+		goto out_bitmap;
+
+	/* Kill the old tree. */
+	error = xrep_rtrefc_remove_old_tree(rr);
+	if (error)
+		goto out_bitmap;
+
+out_bitmap:
+	xfsb_bitmap_destroy(&rr->old_rtrefcountbt_blocks);
+	xfarray_destroy(rr->refcount_records);
+out_rr:
+	kfree(rr);
+	return error;
+}
diff --git a/fs/xfs/scrub/scrub.c b/fs/xfs/scrub/scrub.c
index 6e31f12cef4cc9..7567dd5cad14f4 100644
--- a/fs/xfs/scrub/scrub.c
+++ b/fs/xfs/scrub/scrub.c
@@ -472,7 +472,7 @@ static const struct xchk_meta_ops meta_scrub_ops[] = {
 		.setup	= xchk_setup_rtrefcountbt,
 		.scrub	= xchk_rtrefcountbt,
 		.has	= xfs_has_rtreflink,
-		.repair	= xrep_notsupported,
+		.repair	= xrep_rtrefcountbt,
 	},
 };
 
diff --git a/fs/xfs/scrub/trace.h b/fs/xfs/scrub/trace.h
index 289e39d1f418ba..56811862aa8226 100644
--- a/fs/xfs/scrub/trace.h
+++ b/fs/xfs/scrub/trace.h
@@ -2116,29 +2116,33 @@ TRACE_EVENT(xrep_ibt_found,
 )
 
 TRACE_EVENT(xrep_refc_found,
-	TP_PROTO(const struct xfs_perag *pag,
+	TP_PROTO(const struct xfs_group *xg,
 		 const struct xfs_refcount_irec *rec),
-	TP_ARGS(pag, rec),
+	TP_ARGS(xg, rec),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
 		__field(xfs_agnumber_t, agno)
 		__field(enum xfs_refc_domain, domain)
+		__field(enum xfs_group_type, type)
 		__field(xfs_agblock_t, startblock)
 		__field(xfs_extlen_t, blockcount)
 		__field(xfs_nlink_t, refcount)
 	),
 	TP_fast_assign(
-		__entry->dev = pag_mount(pag)->m_super->s_dev;
-		__entry->agno = pag_agno(pag);
+		__entry->dev = xg->xg_mount->m_super->s_dev;
+		__entry->agno = xg->xg_gno;
+		__entry->type = xg->xg_type;
 		__entry->domain = rec->rc_domain;
 		__entry->startblock = rec->rc_startblock;
 		__entry->blockcount = rec->rc_blockcount;
 		__entry->refcount = rec->rc_refcount;
 	),
-	TP_printk("dev %d:%d agno 0x%x dom %s agbno 0x%x fsbcount 0x%x refcount %u",
+	TP_printk("dev %d:%d %sno 0x%x dom %s %sbno 0x%x fsbcount 0x%x refcount %u",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __print_symbolic(__entry->type, XG_TYPE_STRINGS),
 		  __entry->agno,
 		  __print_symbolic(__entry->domain, XFS_REFC_DOMAIN_STRINGS),
+		  __print_symbolic(__entry->type, XG_TYPE_STRINGS),
 		  __entry->startblock,
 		  __entry->blockcount,
 		  __entry->refcount)

[PATCH 40/43] xfs: repair inodes that have a refcount btree in the data fork

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Plumb knowledge of refcount btrees into the inode core repair code.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/scrub/inode_repair.c |   36 ++++++++++++++++++++++++++++++++++++
 1 file changed, 36 insertions(+)


diff --git a/fs/xfs/scrub/inode_repair.c b/fs/xfs/scrub/inode_repair.c
index 701baee144a66e..2f641b6d663eb2 100644
--- a/fs/xfs/scrub/inode_repair.c
+++ b/fs/xfs/scrub/inode_repair.c
@@ -40,6 +40,7 @@
 #include "xfs_symlink_remote.h"
 #include "xfs_rtgroup.h"
 #include "xfs_rtrmap_btree.h"
+#include "xfs_rtrefcount_btree.h"
 #include "scrub/xfs_scrub.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
@@ -970,6 +971,34 @@ xrep_dinode_bad_rtrmapbt_fork(
 	return false;
 }
 
+/* Return true if this refcount-format ifork looks like garbage. */
+STATIC bool
+xrep_dinode_bad_rtrefcountbt_fork(
+	struct xfs_scrub	*sc,
+	struct xfs_dinode	*dip,
+	unsigned int		dfork_size)
+{
+	struct xfs_rtrefcount_root *dfp;
+	unsigned int		nrecs;
+	unsigned int		level;
+
+	if (dfork_size < sizeof(struct xfs_rtrefcount_root))
+		return true;
+
+	dfp = XFS_DFORK_PTR(dip, XFS_DATA_FORK);
+	nrecs = be16_to_cpu(dfp->bb_numrecs);
+	level = be16_to_cpu(dfp->bb_level);
+
+	if (level > sc->mp->m_rtrefc_maxlevels)
+		return true;
+	if (xfs_rtrefcount_droot_space_calc(level, nrecs) > dfork_size)
+		return true;
+	if (level > 0 && nrecs == 0)
+		return true;
+
+	return false;
+}
+
 /* Check a metadata-btree fork. */
 STATIC bool
 xrep_dinode_bad_metabt_fork(
@@ -984,6 +1013,8 @@ xrep_dinode_bad_metabt_fork(
 	switch (be16_to_cpu(dip->di_metatype)) {
 	case XFS_METAFILE_RTRMAP:
 		return xrep_dinode_bad_rtrmapbt_fork(sc, dip, dfork_size);
+	case XFS_METAFILE_RTREFCOUNT:
+		return xrep_dinode_bad_rtrefcountbt_fork(sc, dip, dfork_size);
 	default:
 		return true;
 	}
@@ -1249,6 +1280,7 @@ xrep_dinode_ensure_forkoff(
 {
 	struct xfs_bmdr_block	*bmdr;
 	struct xfs_rtrmap_root	*rmdr;
+	struct xfs_rtrefcount_root *rcdr;
 	struct xfs_scrub	*sc = ri->sc;
 	xfs_extnum_t		attr_extents, data_extents;
 	size_t			bmdr_minsz = xfs_bmdr_space_calc(1);
@@ -1361,6 +1393,10 @@ xrep_dinode_ensure_forkoff(
 			rmdr = XFS_DFORK_PTR(dip, XFS_DATA_FORK);
 			dfork_min = xfs_rtrmap_broot_space(sc->mp, rmdr);
 			break;
+		case XFS_METAFILE_RTREFCOUNT:
+			rcdr = XFS_DFORK_PTR(dip, XFS_DATA_FORK);
+			dfork_min = xfs_rtrefcount_broot_space(sc->mp, rcdr);
+			break;
 		default:
 			dfork_min = 0;
 			break;

[PATCH 41/43] xfs: check for shared rt extents when rebuilding rt file's data fork

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

When we're rebuilding the data fork of a realtime file, we need to
cross-reference each mapping with the rt refcount btree to ensure that
the reflink flag is set if there are any shared extents found.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/scrub/bmap_repair.c |   21 +++++++++++++--------
 1 file changed, 13 insertions(+), 8 deletions(-)


diff --git a/fs/xfs/scrub/bmap_repair.c b/fs/xfs/scrub/bmap_repair.c
index fd64bdf4e13887..1084213b8e9b88 100644
--- a/fs/xfs/scrub/bmap_repair.c
+++ b/fs/xfs/scrub/bmap_repair.c
@@ -101,14 +101,21 @@ xrep_bmap_discover_shared(
 	xfs_filblks_t		blockcount)
 {
 	struct xfs_scrub	*sc = rb->sc;
+	struct xfs_btree_cur	*cur;
 	xfs_agblock_t		agbno;
 	xfs_agblock_t		fbno;
 	xfs_extlen_t		flen;
 	int			error;
 
-	agbno = XFS_FSB_TO_AGBNO(sc->mp, startblock);
-	error = xfs_refcount_find_shared(sc->sa.refc_cur, agbno, blockcount,
-			&fbno, &flen, false);
+	if (XFS_IS_REALTIME_INODE(sc->ip)) {
+		agbno = xfs_rtb_to_rgbno(sc->mp, startblock);
+		cur = sc->sr.refc_cur;
+	} else {
+		agbno = XFS_FSB_TO_AGBNO(sc->mp, startblock);
+		cur = sc->sa.refc_cur;
+	}
+	error = xfs_refcount_find_shared(cur, agbno, blockcount, &fbno, &flen,
+			false);
 	if (error)
 		return error;
 
@@ -450,7 +457,9 @@ xrep_bmap_scan_rtgroup(
 		return 0;
 
 	error = xrep_rtgroup_init(sc, rtg, &sc->sr,
-			XFS_RTGLOCK_RMAP | XFS_RTGLOCK_BITMAP_SHARED);
+			XFS_RTGLOCK_RMAP |
+			XFS_RTGLOCK_REFCOUNT |
+			XFS_RTGLOCK_BITMAP_SHARED);
 	if (error)
 		return error;
 
@@ -903,10 +912,6 @@ xrep_bmap_init_reflink_scan(
 	if (whichfork != XFS_DATA_FORK)
 		return RLS_IRRELEVANT;
 
-	/* cannot share realtime extents */
-	if (XFS_IS_REALTIME_INODE(sc->ip))
-		return RLS_IRRELEVANT;
-
 	return RLS_UNKNOWN;
 }
 

[PATCH 42/43] xfs: fix CoW forks for realtime files

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Port the copy on write fork repair to realtime files.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
---
 fs/xfs/scrub/agheader_repair.c |    2 
 fs/xfs/scrub/cow_repair.c      |  178 +++++++++++++++++++++++++++--
 fs/xfs/scrub/reap.c            |  242 +++++++++++++++++++++++++++++++++++++++-
 fs/xfs/scrub/reap.h            |    7 +
 fs/xfs/scrub/repair.h          |    1 
 fs/xfs/scrub/rtb_bitmap.h      |   37 ++++++
 fs/xfs/scrub/trace.h           |   36 ++++--
 fs/xfs/xfs_rtalloc.c           |    4 -
 fs/xfs/xfs_rtalloc.h           |    5 +
 9 files changed, 470 insertions(+), 42 deletions(-)
 create mode 100644 fs/xfs/scrub/rtb_bitmap.h


diff --git a/fs/xfs/scrub/agheader_repair.c b/fs/xfs/scrub/agheader_repair.c
index b45d2b32051a63..cd6f0223879f49 100644
--- a/fs/xfs/scrub/agheader_repair.c
+++ b/fs/xfs/scrub/agheader_repair.c
@@ -647,7 +647,7 @@ xrep_agfl_fill(
 	xfs_agblock_t		agbno = start;
 	int			error;
 
-	trace_xrep_agfl_insert(sc->sa.pag, agbno, len);
+	trace_xrep_agfl_insert(pag_group(sc->sa.pag), agbno, len);
 
 	while (agbno < start + len && af->fl_off < af->flcount)
 		af->agfl_bno[af->fl_off++] = cpu_to_be32(agbno++);
diff --git a/fs/xfs/scrub/cow_repair.c b/fs/xfs/scrub/cow_repair.c
index ba695dd21f8b96..38a246b8bf11c9 100644
--- a/fs/xfs/scrub/cow_repair.c
+++ b/fs/xfs/scrub/cow_repair.c
@@ -26,6 +26,9 @@
 #include "xfs_errortag.h"
 #include "xfs_icache.h"
 #include "xfs_refcount_btree.h"
+#include "xfs_rtalloc.h"
+#include "xfs_rtbitmap.h"
+#include "xfs_rtgroup.h"
 #include "scrub/xfs_scrub.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
@@ -34,6 +37,7 @@
 #include "scrub/bitmap.h"
 #include "scrub/off_bitmap.h"
 #include "scrub/fsb_bitmap.h"
+#include "scrub/rtb_bitmap.h"
 #include "scrub/reap.h"
 
 /*
@@ -61,7 +65,10 @@ struct xrep_cow {
 	struct xoff_bitmap	bad_fileoffs;
 
 	/* Bitmap of fsblocks that were removed from the CoW fork. */
-	struct xfsb_bitmap	old_cowfork_fsblocks;
+	union {
+		struct xfsb_bitmap	old_cowfork_fsblocks;
+		struct xrtb_bitmap	old_cowfork_rtblocks;
+	};
 
 	/* CoW fork mappings used to scan for bad CoW staging extents. */
 	struct xfs_bmbt_irec	irec;
@@ -145,8 +152,7 @@ xrep_cow_mark_shared_staging(
 	xrep_cow_trim_refcount(xc, &rrec, rec);
 
 	return xrep_cow_mark_file_range(xc,
-			xfs_agbno_to_fsb(to_perag(cur->bc_group),
-				rrec.rc_startblock),
+			xfs_gbno_to_fsb(cur->bc_group, rrec.rc_startblock),
 			rrec.rc_blockcount);
 }
 
@@ -177,9 +183,8 @@ xrep_cow_mark_missing_staging(
 	if (xc->next_bno >= rrec.rc_startblock)
 		goto next;
 
-
 	error = xrep_cow_mark_file_range(xc,
-			xfs_agbno_to_fsb(to_perag(cur->bc_group), xc->next_bno),
+			xfs_gbno_to_fsb(cur->bc_group, xc->next_bno),
 			rrec.rc_startblock - xc->next_bno);
 	if (error)
 		return error;
@@ -222,8 +227,7 @@ xrep_cow_mark_missing_staging_rmap(
 	}
 
 	return xrep_cow_mark_file_range(xc,
-			xfs_agbno_to_fsb(to_perag(cur->bc_group), rec_bno),
-			rec_len);
+			xfs_gbno_to_fsb(cur->bc_group, rec_bno), rec_len);
 }
 
 /*
@@ -310,6 +314,92 @@ xrep_cow_find_bad(
 	return 0;
 }
 
+/*
+ * Find any part of the CoW fork mapping that isn't a single-owner CoW staging
+ * extent and mark the corresponding part of the file range in the bitmap.
+ */
+STATIC int
+xrep_cow_find_bad_rt(
+	struct xrep_cow			*xc)
+{
+	struct xfs_refcount_irec	rc_low = { 0 };
+	struct xfs_refcount_irec	rc_high = { 0 };
+	struct xfs_rmap_irec		rm_low = { 0 };
+	struct xfs_rmap_irec		rm_high = { 0 };
+	struct xfs_scrub		*sc = xc->sc;
+	struct xfs_rtgroup		*rtg;
+	int				error = 0;
+
+	xc->irec_startbno = xfs_rtb_to_rgbno(sc->mp, xc->irec.br_startblock);
+
+	rtg = xfs_rtgroup_get(sc->mp,
+			xfs_rtb_to_rgno(sc->mp, xc->irec.br_startblock));
+	if (!rtg)
+		return -EFSCORRUPTED;
+
+	error = xrep_rtgroup_init(sc, rtg, &sc->sr,
+			XFS_RTGLOCK_RMAP | XFS_RTGLOCK_REFCOUNT);
+	if (error)
+		goto out_rtg;
+
+	/* Mark any CoW fork extents that are shared. */
+	rc_low.rc_startblock = xc->irec_startbno;
+	rc_high.rc_startblock = xc->irec_startbno + xc->irec.br_blockcount - 1;
+	rc_low.rc_domain = rc_high.rc_domain = XFS_REFC_DOMAIN_SHARED;
+	error = xfs_refcount_query_range(sc->sr.refc_cur, &rc_low, &rc_high,
+			xrep_cow_mark_shared_staging, xc);
+	if (error)
+		goto out_sr;
+
+	/* Make sure there are CoW staging extents for the whole mapping. */
+	rc_low.rc_startblock = xc->irec_startbno;
+	rc_high.rc_startblock = xc->irec_startbno + xc->irec.br_blockcount - 1;
+	rc_low.rc_domain = rc_high.rc_domain = XFS_REFC_DOMAIN_COW;
+	xc->next_bno = xc->irec_startbno;
+	error = xfs_refcount_query_range(sc->sr.refc_cur, &rc_low, &rc_high,
+			xrep_cow_mark_missing_staging, xc);
+	if (error)
+		goto out_sr;
+
+	if (xc->next_bno < xc->irec_startbno + xc->irec.br_blockcount) {
+		error = xrep_cow_mark_file_range(xc,
+				xfs_rgbno_to_rtb(rtg, xc->next_bno),
+				xc->irec_startbno + xc->irec.br_blockcount -
+				xc->next_bno);
+		if (error)
+			goto out_sr;
+	}
+
+	/* Mark any area has an rmap that isn't a COW staging extent. */
+	rm_low.rm_startblock = xc->irec_startbno;
+	memset(&rm_high, 0xFF, sizeof(rm_high));
+	rm_high.rm_startblock = xc->irec_startbno + xc->irec.br_blockcount - 1;
+	error = xfs_rmap_query_range(sc->sr.rmap_cur, &rm_low, &rm_high,
+			xrep_cow_mark_missing_staging_rmap, xc);
+	if (error)
+		goto out_sr;
+
+	/*
+	 * If userspace is forcing us to rebuild the CoW fork or someone
+	 * turned on the debugging knob, replace everything in the
+	 * CoW fork and then scan for staging extents in the refcountbt.
+	 */
+	if ((sc->sm->sm_flags & XFS_SCRUB_IFLAG_FORCE_REBUILD) ||
+	    XFS_TEST_ERROR(false, sc->mp, XFS_ERRTAG_FORCE_SCRUB_REPAIR)) {
+		error = xrep_cow_mark_file_range(xc, xc->irec.br_startblock,
+				xc->irec.br_blockcount);
+		if (error)
+			goto out_rtg;
+	}
+
+out_sr:
+	xchk_rtgroup_btcur_free(&sc->sr);
+	xchk_rtgroup_free(sc, &sc->sr);
+out_rtg:
+	xfs_rtgroup_put(rtg);
+	return error;
+}
+
 /*
  * Allocate a replacement CoW staging extent of up to the given number of
  * blocks, and fill out the mapping.
@@ -350,6 +440,32 @@ xrep_cow_alloc(
 	return 0;
 }
 
+/*
+ * Allocate a replacement rt CoW staging extent of up to the given number of
+ * blocks, and fill out the mapping.
+ */
+STATIC int
+xrep_cow_alloc_rt(
+	struct xfs_scrub	*sc,
+	xfs_extlen_t		maxlen,
+	struct xrep_cow_extent	*repl)
+{
+	xfs_rtxlen_t		maxrtx = xfs_rtb_to_rtx(sc->mp, maxlen);
+	int			error;
+
+	error = xfs_trans_reserve_more(sc->tp, 0, maxrtx);
+	if (error)
+		return error;
+
+	error = xfs_rtallocate_rtgs(sc->tp, NULLRTBLOCK, 1, maxrtx, 1, false,
+			false, &repl->fsbno, &repl->len);
+	if (error)
+		return error;
+
+	xfs_refcount_alloc_cow_extent(sc->tp, true, repl->fsbno, repl->len);
+	return 0;
+}
+
 /*
  * Look up the current CoW fork mapping so that we only allocate enough to
  * replace a single mapping.  If we don't find a mapping that covers the start
@@ -467,7 +583,10 @@ xrep_cow_replace_range(
 	 */
 	alloc_len = min_t(xfs_fileoff_t, XFS_MAX_BMBT_EXTLEN,
 			  nextoff - startoff);
-	error = xrep_cow_alloc(sc, alloc_len, &repl);
+	if (XFS_IS_REALTIME_INODE(sc->ip))
+		error = xrep_cow_alloc_rt(sc, alloc_len, &repl);
+	else
+		error = xrep_cow_alloc(sc, alloc_len, &repl);
 	if (error)
 		return error;
 
@@ -483,8 +602,12 @@ xrep_cow_replace_range(
 		return error;
 
 	/* Note the old CoW staging extents; we'll reap them all later. */
-	error = xfsb_bitmap_set(&xc->old_cowfork_fsblocks, got.br_startblock,
-			repl.len);
+	if (XFS_IS_REALTIME_INODE(sc->ip))
+		error = xrtb_bitmap_set(&xc->old_cowfork_rtblocks,
+				got.br_startblock, repl.len);
+	else
+		error = xfsb_bitmap_set(&xc->old_cowfork_fsblocks,
+				got.br_startblock, repl.len);
 	if (error)
 		return error;
 
@@ -540,8 +663,16 @@ xrep_bmap_cow(
 	if (!ifp)
 		return 0;
 
-	/* realtime files aren't supported yet */
-	if (XFS_IS_REALTIME_INODE(sc->ip))
+	/*
+	 * Realtime files with large extent sizes are not supported because
+	 * we could encounter an CoW mapping that has been partially written
+	 * out *and* requires replacement, and there's no solution to that.
+	 */
+	if (xfs_inode_has_bigrtalloc(sc->ip))
+		return -EOPNOTSUPP;
+
+	/* Metadata inodes aren't supposed to have data on the rt volume. */
+	if (xfs_is_metadir_inode(sc->ip) && XFS_IS_REALTIME_INODE(sc->ip))
 		return -EOPNOTSUPP;
 
 	/*
@@ -562,7 +693,10 @@ xrep_bmap_cow(
 
 	xc->sc = sc;
 	xoff_bitmap_init(&xc->bad_fileoffs);
-	xfsb_bitmap_init(&xc->old_cowfork_fsblocks);
+	if (XFS_IS_REALTIME_INODE(sc->ip))
+		xrtb_bitmap_init(&xc->old_cowfork_rtblocks);
+	else
+		xfsb_bitmap_init(&xc->old_cowfork_fsblocks);
 
 	for_each_xfs_iext(ifp, &icur, &xc->irec) {
 		if (xchk_should_terminate(sc, &error))
@@ -585,7 +719,10 @@ xrep_bmap_cow(
 		if (xfs_bmap_is_written_extent(&xc->irec))
 			continue;
 
-		error = xrep_cow_find_bad(xc);
+		if (XFS_IS_REALTIME_INODE(sc->ip))
+			error = xrep_cow_find_bad_rt(xc);
+		else
+			error = xrep_cow_find_bad(xc);
 		if (error)
 			goto out_bitmap;
 	}
@@ -600,13 +737,20 @@ xrep_bmap_cow(
 	 * by the refcount btree, not the inode, so it is correct to treat them
 	 * like inode metadata.
 	 */
-	error = xrep_reap_fsblocks(sc, &xc->old_cowfork_fsblocks,
-			&XFS_RMAP_OINFO_COW);
+	if (XFS_IS_REALTIME_INODE(sc->ip))
+		error = xrep_reap_rtblocks(sc, &xc->old_cowfork_rtblocks,
+				&XFS_RMAP_OINFO_COW);
+	else
+		error = xrep_reap_fsblocks(sc, &xc->old_cowfork_fsblocks,
+				&XFS_RMAP_OINFO_COW);
 	if (error)
 		goto out_bitmap;
 
 out_bitmap:
-	xfsb_bitmap_destroy(&xc->old_cowfork_fsblocks);
+	if (XFS_IS_REALTIME_INODE(sc->ip))
+		xrtb_bitmap_destroy(&xc->old_cowfork_rtblocks);
+	else
+		xfsb_bitmap_destroy(&xc->old_cowfork_fsblocks);
 	xoff_bitmap_destroy(&xc->bad_fileoffs);
 	kfree(xc);
 	return error;
diff --git a/fs/xfs/scrub/reap.c b/fs/xfs/scrub/reap.c
index 534c7696e9a1a7..b32fb233cf8476 100644
--- a/fs/xfs/scrub/reap.c
+++ b/fs/xfs/scrub/reap.c
@@ -34,6 +34,8 @@
 #include "xfs_attr_remote.h"
 #include "xfs_defer.h"
 #include "xfs_metafile.h"
+#include "xfs_rtgroup.h"
+#include "xfs_rtrmap_btree.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/trace.h"
@@ -41,6 +43,7 @@
 #include "scrub/bitmap.h"
 #include "scrub/agb_bitmap.h"
 #include "scrub/fsb_bitmap.h"
+#include "scrub/rtb_bitmap.h"
 #include "scrub/reap.h"
 
 /*
@@ -311,7 +314,7 @@ xreap_agextent_binval(
 	}
 
 out:
-	trace_xreap_agextent_binval(sc->sa.pag, agbno, *aglenp);
+	trace_xreap_agextent_binval(pag_group(sc->sa.pag), agbno, *aglenp);
 }
 
 /*
@@ -370,7 +373,8 @@ xreap_agextent_select(
 
 out_found:
 	*aglenp = len;
-	trace_xreap_agextent_select(sc->sa.pag, agbno, len, *crosslinked);
+	trace_xreap_agextent_select(pag_group(sc->sa.pag), agbno, len,
+			*crosslinked);
 out_cur:
 	xfs_btree_del_cursor(cur, error);
 	return error;
@@ -409,7 +413,8 @@ xreap_agextent_iter(
 	 * to run xfs_repair.
 	 */
 	if (crosslinked) {
-		trace_xreap_dispose_unmap_extent(sc->sa.pag, agbno, *aglenp);
+		trace_xreap_dispose_unmap_extent(pag_group(sc->sa.pag), agbno,
+				*aglenp);
 
 		rs->force_roll = true;
 
@@ -428,7 +433,7 @@ xreap_agextent_iter(
 				*aglenp, rs->oinfo);
 	}
 
-	trace_xreap_dispose_free_extent(sc->sa.pag, agbno, *aglenp);
+	trace_xreap_dispose_free_extent(pag_group(sc->sa.pag), agbno, *aglenp);
 
 	/*
 	 * Invalidate as many buffers as we can, starting at agbno.  If this
@@ -679,6 +684,225 @@ xrep_reap_fsblocks(
 	return 0;
 }
 
+#ifdef CONFIG_XFS_RT
+/*
+ * Figure out the longest run of blocks that we can dispose of with a single
+ * call.  Cross-linked blocks should have their reverse mappings removed, but
+ * single-owner extents can be freed.  Units are rt blocks, not rt extents.
+ */
+STATIC int
+xreap_rgextent_select(
+	struct xreap_state	*rs,
+	xfs_rgblock_t		rgbno,
+	xfs_rgblock_t		rgbno_next,
+	bool			*crosslinked,
+	xfs_extlen_t		*rglenp)
+{
+	struct xfs_scrub	*sc = rs->sc;
+	struct xfs_btree_cur	*cur;
+	xfs_rgblock_t		bno = rgbno + 1;
+	xfs_extlen_t		len = 1;
+	int			error;
+
+	/*
+	 * Determine if there are any other rmap records covering the first
+	 * block of this extent.  If so, the block is crosslinked.
+	 */
+	cur = xfs_rtrmapbt_init_cursor(sc->tp, sc->sr.rtg);
+	error = xfs_rmap_has_other_keys(cur, rgbno, 1, rs->oinfo,
+			crosslinked);
+	if (error)
+		goto out_cur;
+
+	/*
+	 * Figure out how many of the subsequent blocks have the same crosslink
+	 * status.
+	 */
+	while (bno < rgbno_next) {
+		bool		also_crosslinked;
+
+		error = xfs_rmap_has_other_keys(cur, bno, 1, rs->oinfo,
+				&also_crosslinked);
+		if (error)
+			goto out_cur;
+
+		if (*crosslinked != also_crosslinked)
+			break;
+
+		len++;
+		bno++;
+	}
+
+	*rglenp = len;
+	trace_xreap_agextent_select(rtg_group(sc->sr.rtg), rgbno, len,
+			*crosslinked);
+out_cur:
+	xfs_btree_del_cursor(cur, error);
+	return error;
+}
+
+/*
+ * Dispose of as much of the beginning of this rtgroup extent as possible.
+ * The number of blocks disposed of will be returned in @rglenp.
+ */
+STATIC int
+xreap_rgextent_iter(
+	struct xreap_state	*rs,
+	xfs_rgblock_t		rgbno,
+	xfs_extlen_t		*rglenp,
+	bool			crosslinked)
+{
+	struct xfs_scrub	*sc = rs->sc;
+	xfs_rtblock_t		rtbno;
+	int			error;
+
+	/*
+	 * The only caller so far is CoW fork repair, so we only know how to
+	 * unlink or free CoW staging extents.  Here we don't have to worry
+	 * about invalidating buffers!
+	 */
+	if (rs->oinfo != &XFS_RMAP_OINFO_COW) {
+		ASSERT(rs->oinfo == &XFS_RMAP_OINFO_COW);
+		return -EFSCORRUPTED;
+	}
+	ASSERT(rs->resv == XFS_AG_RESV_NONE);
+
+	rtbno = xfs_rgbno_to_rtb(sc->sr.rtg, rgbno);
+
+	/*
+	 * If there are other rmappings, this block is cross linked and must
+	 * not be freed.  Remove the forward and reverse mapping and move on.
+	 */
+	if (crosslinked) {
+		trace_xreap_dispose_unmap_extent(rtg_group(sc->sr.rtg), rgbno,
+				*rglenp);
+
+		xfs_refcount_free_cow_extent(sc->tp, true, rtbno, *rglenp);
+		rs->deferred++;
+		return 0;
+	}
+
+	trace_xreap_dispose_free_extent(rtg_group(sc->sr.rtg), rgbno, *rglenp);
+
+	/*
+	 * The CoW staging extent is not crosslinked.  Use deferred work items
+	 * to remove the refcountbt records (which removes the rmap records)
+	 * and free the extent.  We're not worried about the system going down
+	 * here because log recovery walks the refcount btree to clean out the
+	 * CoW staging extents.
+	 */
+	xfs_refcount_free_cow_extent(sc->tp, true, rtbno, *rglenp);
+	error = xfs_free_extent_later(sc->tp, rtbno, *rglenp, NULL,
+			rs->resv,
+			XFS_FREE_EXTENT_REALTIME |
+			XFS_FREE_EXTENT_SKIP_DISCARD);
+	if (error)
+		return error;
+
+	rs->deferred++;
+	return 0;
+}
+
+#define XREAP_RTGLOCK_ALL	(XFS_RTGLOCK_BITMAP | \
+				 XFS_RTGLOCK_RMAP | \
+				 XFS_RTGLOCK_REFCOUNT)
+
+/*
+ * Break a rt file metadata extent into sub-extents by fate (crosslinked, not
+ * crosslinked), and dispose of each sub-extent separately.  The extent must
+ * be aligned to a realtime extent.
+ */
+STATIC int
+xreap_rtmeta_extent(
+	uint64_t		rtbno,
+	uint64_t		len,
+	void			*priv)
+{
+	struct xreap_state	*rs = priv;
+	struct xfs_scrub	*sc = rs->sc;
+	xfs_rgblock_t		rgbno = xfs_rtb_to_rgbno(sc->mp, rtbno);
+	xfs_rgblock_t		rgbno_next = rgbno + len;
+	int			error = 0;
+
+	ASSERT(sc->ip != NULL);
+	ASSERT(!sc->sr.rtg);
+
+	/*
+	 * We're reaping blocks after repairing file metadata, which means that
+	 * we have to init the xchk_ag structure ourselves.
+	 */
+	sc->sr.rtg = xfs_rtgroup_get(sc->mp, xfs_rtb_to_rgno(sc->mp, rtbno));
+	if (!sc->sr.rtg)
+		return -EFSCORRUPTED;
+
+	xfs_rtgroup_lock(sc->sr.rtg, XREAP_RTGLOCK_ALL);
+
+	while (rgbno < rgbno_next) {
+		xfs_extlen_t	rglen;
+		bool		crosslinked;
+
+		error = xreap_rgextent_select(rs, rgbno, rgbno_next,
+				&crosslinked, &rglen);
+		if (error)
+			goto out_unlock;
+
+		error = xreap_rgextent_iter(rs, rgbno, &rglen, crosslinked);
+		if (error)
+			goto out_unlock;
+
+		if (xreap_want_defer_finish(rs)) {
+			error = xfs_defer_finish(&sc->tp);
+			if (error)
+				goto out_unlock;
+			xreap_defer_finish_reset(rs);
+		} else if (xreap_want_roll(rs)) {
+			error = xfs_trans_roll_inode(&sc->tp, sc->ip);
+			if (error)
+				goto out_unlock;
+			xreap_reset(rs);
+		}
+
+		rgbno += rglen;
+	}
+
+out_unlock:
+	xfs_rtgroup_unlock(sc->sr.rtg, XREAP_RTGLOCK_ALL);
+	xfs_rtgroup_put(sc->sr.rtg);
+	sc->sr.rtg = NULL;
+	return error;
+}
+
+/*
+ * Dispose of every block of every rt metadata extent in the bitmap.
+ * Do not use this to dispose of the mappings in an ondisk inode fork.
+ */
+int
+xrep_reap_rtblocks(
+	struct xfs_scrub		*sc,
+	struct xrtb_bitmap		*bitmap,
+	const struct xfs_owner_info	*oinfo)
+{
+	struct xreap_state		rs = {
+		.sc			= sc,
+		.oinfo			= oinfo,
+		.resv			= XFS_AG_RESV_NONE,
+	};
+	int				error;
+
+	ASSERT(xfs_has_rmapbt(sc->mp));
+	ASSERT(sc->ip != NULL);
+
+	error = xrtb_bitmap_walk(bitmap, xreap_rtmeta_extent, &rs);
+	if (error)
+		return error;
+
+	if (xreap_dirty(&rs))
+		return xrep_defer_finish(sc);
+
+	return 0;
+}
+#endif /* CONFIG_XFS_RT */
+
 /*
  * Dispose of every block of an old metadata btree that used to be rooted in a
  * metadata directory file.
@@ -771,7 +995,8 @@ xreap_bmapi_select(
 	}
 
 	imap->br_blockcount = len;
-	trace_xreap_bmapi_select(sc->sa.pag, agbno, len, *crosslinked);
+	trace_xreap_bmapi_select(pag_group(sc->sa.pag), agbno, len,
+			*crosslinked);
 out_cur:
 	xfs_btree_del_cursor(cur, error);
 	return error;
@@ -910,7 +1135,8 @@ xreap_bmapi_binval(
 	}
 
 out:
-	trace_xreap_bmapi_binval(sc->sa.pag, agbno, imap->br_blockcount);
+	trace_xreap_bmapi_binval(pag_group(sc->sa.pag), agbno,
+			imap->br_blockcount);
 	return 0;
 }
 
@@ -937,7 +1163,7 @@ xrep_reap_bmapi_iter(
 		 * anybody else who thinks they own the block, even though that
 		 * runs the risk of stale buffer warnings in the future.
 		 */
-		trace_xreap_dispose_unmap_extent(sc->sa.pag,
+		trace_xreap_dispose_unmap_extent(pag_group(sc->sa.pag),
 				XFS_FSB_TO_AGBNO(sc->mp, imap->br_startblock),
 				imap->br_blockcount);
 
@@ -960,7 +1186,7 @@ xrep_reap_bmapi_iter(
 	 * by a block starting before the first block of the extent but overlap
 	 * anyway.
 	 */
-	trace_xreap_dispose_free_extent(sc->sa.pag,
+	trace_xreap_dispose_free_extent(pag_group(sc->sa.pag),
 			XFS_FSB_TO_AGBNO(sc->mp, imap->br_startblock),
 			imap->br_blockcount);
 
diff --git a/fs/xfs/scrub/reap.h b/fs/xfs/scrub/reap.h
index 70e5e6bbb8d38d..4c8f62701fb36b 100644
--- a/fs/xfs/scrub/reap.h
+++ b/fs/xfs/scrub/reap.h
@@ -17,6 +17,13 @@ int xrep_reap_ifork(struct xfs_scrub *sc, struct xfs_inode *ip, int whichfork);
 int xrep_reap_metadir_fsblocks(struct xfs_scrub *sc,
 		struct xfsb_bitmap *bitmap);
 
+#ifdef CONFIG_XFS_RT
+int xrep_reap_rtblocks(struct xfs_scrub *sc, struct xrtb_bitmap *bitmap,
+		const struct xfs_owner_info *oinfo);
+#else
+# define xrep_reap_rtblocks(...)	(-EOPNOTSUPP)
+#endif /* CONFIG_XFS_RT */
+
 /* Buffer cache scan context. */
 struct xrep_bufscan {
 	/* Disk address for the buffers we want to scan. */
diff --git a/fs/xfs/scrub/repair.h b/fs/xfs/scrub/repair.h
index 8f8f18b48a449d..823c00d1a50262 100644
--- a/fs/xfs/scrub/repair.h
+++ b/fs/xfs/scrub/repair.h
@@ -52,6 +52,7 @@ struct xbitmap;
 struct xagb_bitmap;
 struct xrgb_bitmap;
 struct xfsb_bitmap;
+struct xrtb_bitmap;
 
 int xrep_fix_freelist(struct xfs_scrub *sc, int alloc_flags);
 
diff --git a/fs/xfs/scrub/rtb_bitmap.h b/fs/xfs/scrub/rtb_bitmap.h
new file mode 100644
index 00000000000000..1313ef605511ea
--- /dev/null
+++ b/fs/xfs/scrub/rtb_bitmap.h
@@ -0,0 +1,37 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2022-2024 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <djwong@kernel.org>
+ */
+#ifndef __XFS_SCRUB_RTB_BITMAP_H__
+#define __XFS_SCRUB_RTB_BITMAP_H__
+
+/* Bitmaps, but for type-checked for xfs_rtblock_t */
+
+struct xrtb_bitmap {
+	struct xbitmap64	rtbitmap;
+};
+
+static inline void xrtb_bitmap_init(struct xrtb_bitmap *bitmap)
+{
+	xbitmap64_init(&bitmap->rtbitmap);
+}
+
+static inline void xrtb_bitmap_destroy(struct xrtb_bitmap *bitmap)
+{
+	xbitmap64_destroy(&bitmap->rtbitmap);
+}
+
+static inline int xrtb_bitmap_set(struct xrtb_bitmap *bitmap,
+		xfs_rtblock_t start, xfs_filblks_t len)
+{
+	return xbitmap64_set(&bitmap->rtbitmap, start, len);
+}
+
+static inline int xrtb_bitmap_walk(struct xrtb_bitmap *bitmap,
+		xbitmap64_walk_fn fn, void *priv)
+{
+	return xbitmap64_walk(&bitmap->rtbitmap, fn, priv);
+}
+
+#endif	/* __XFS_SCRUB_RTB_BITMAP_H__ */
diff --git a/fs/xfs/scrub/trace.h b/fs/xfs/scrub/trace.h
index 56811862aa8226..d7c4ced47c1567 100644
--- a/fs/xfs/scrub/trace.h
+++ b/fs/xfs/scrub/trace.h
@@ -1964,32 +1964,36 @@ DEFINE_XCHK_METAPATH_EVENT(xchk_metapath_lookup);
 #if IS_ENABLED(CONFIG_XFS_ONLINE_REPAIR)
 
 DECLARE_EVENT_CLASS(xrep_extent_class,
-	TP_PROTO(const struct xfs_perag *pag, xfs_agblock_t agbno,
+	TP_PROTO(const struct xfs_group *xg, xfs_agblock_t agbno,
 		 xfs_extlen_t len),
-	TP_ARGS(pag, agbno, len),
+	TP_ARGS(xg, agbno, len),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
+		__field(enum xfs_group_type, type)
 		__field(xfs_agnumber_t, agno)
 		__field(xfs_agblock_t, agbno)
 		__field(xfs_extlen_t, len)
 	),
 	TP_fast_assign(
-		__entry->dev = pag_mount(pag)->m_super->s_dev;
-		__entry->agno = pag_agno(pag);
+		__entry->dev = xg->xg_mount->m_super->s_dev;
+		__entry->type = xg->xg_type;
+		__entry->agno = xg->xg_gno;
 		__entry->agbno = agbno;
 		__entry->len = len;
 	),
-	TP_printk("dev %d:%d agno 0x%x agbno 0x%x fsbcount 0x%x",
+	TP_printk("dev %d:%d %sno 0x%x %sbno 0x%x fsbcount 0x%x",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __print_symbolic(__entry->type, XG_TYPE_STRINGS),
 		  __entry->agno,
+		  __print_symbolic(__entry->type, XG_TYPE_STRINGS),
 		  __entry->agbno,
 		  __entry->len)
 );
 #define DEFINE_REPAIR_EXTENT_EVENT(name) \
 DEFINE_EVENT(xrep_extent_class, name, \
-	TP_PROTO(const struct xfs_perag *pag, xfs_agblock_t agbno, \
+	TP_PROTO(const struct xfs_group *xg, xfs_agblock_t agbno, \
 		 xfs_extlen_t len), \
-	TP_ARGS(pag, agbno, len))
+	TP_ARGS(xg, agbno, len))
 DEFINE_REPAIR_EXTENT_EVENT(xreap_dispose_unmap_extent);
 DEFINE_REPAIR_EXTENT_EVENT(xreap_dispose_free_extent);
 DEFINE_REPAIR_EXTENT_EVENT(xreap_agextent_binval);
@@ -1997,35 +2001,39 @@ DEFINE_REPAIR_EXTENT_EVENT(xreap_bmapi_binval);
 DEFINE_REPAIR_EXTENT_EVENT(xrep_agfl_insert);
 
 DECLARE_EVENT_CLASS(xrep_reap_find_class,
-	TP_PROTO(const struct xfs_perag *pag, xfs_agblock_t agbno,
+	TP_PROTO(const struct xfs_group *xg, xfs_agblock_t agbno,
 		 xfs_extlen_t len, bool crosslinked),
-	TP_ARGS(pag, agbno, len, crosslinked),
+	TP_ARGS(xg, agbno, len, crosslinked),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
+		__field(enum xfs_group_type, type)
 		__field(xfs_agnumber_t, agno)
 		__field(xfs_agblock_t, agbno)
 		__field(xfs_extlen_t, len)
 		__field(bool, crosslinked)
 	),
 	TP_fast_assign(
-		__entry->dev = pag_mount(pag)->m_super->s_dev;
-		__entry->agno = pag_agno(pag);
+		__entry->dev = xg->xg_mount->m_super->s_dev;
+		__entry->type = xg->xg_type;
+		__entry->agno = xg->xg_gno;
 		__entry->agbno = agbno;
 		__entry->len = len;
 		__entry->crosslinked = crosslinked;
 	),
-	TP_printk("dev %d:%d agno 0x%x agbno 0x%x fsbcount 0x%x crosslinked %d",
+	TP_printk("dev %d:%d %sno 0x%x %sbno 0x%x fsbcount 0x%x crosslinked %d",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __print_symbolic(__entry->type, XG_TYPE_STRINGS),
 		  __entry->agno,
+		  __print_symbolic(__entry->type, XG_TYPE_STRINGS),
 		  __entry->agbno,
 		  __entry->len,
 		  __entry->crosslinked ? 1 : 0)
 );
 #define DEFINE_REPAIR_REAP_FIND_EVENT(name) \
 DEFINE_EVENT(xrep_reap_find_class, name, \
-	TP_PROTO(const struct xfs_perag *pag, xfs_agblock_t agbno, \
+	TP_PROTO(const struct xfs_group *xg, xfs_agblock_t agbno, \
 		 xfs_extlen_t len, bool crosslinked), \
-	TP_ARGS(pag, agbno, len, crosslinked))
+	TP_ARGS(xg, agbno, len, crosslinked))
 DEFINE_REPAIR_REAP_FIND_EVENT(xreap_agextent_select);
 DEFINE_REPAIR_REAP_FIND_EVENT(xreap_bmapi_select);
 
diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c
index a5de5405800a22..7135a6717a9e11 100644
--- a/fs/xfs/xfs_rtalloc.c
+++ b/fs/xfs/xfs_rtalloc.c
@@ -594,7 +594,7 @@ xfs_rtalloc_sumlevel(
  * specified.  If we don't get maxlen then use prod to trim
  * the length, if given.  The lengths are all in rtextents.
  */
-STATIC int
+static int
 xfs_rtallocate_extent_size(
 	struct xfs_rtalloc_args	*args,
 	xfs_rtxlen_t		minlen,	/* minimum length to allocate */
@@ -1958,7 +1958,7 @@ xfs_rtallocate_rtg(
 	goto out_release;
 }
 
-static int
+int
 xfs_rtallocate_rtgs(
 	struct xfs_trans	*tp,
 	xfs_fsblock_t		bno_hint,
diff --git a/fs/xfs/xfs_rtalloc.h b/fs/xfs/xfs_rtalloc.h
index 9044f7226ab6fc..0d95b29092c9f3 100644
--- a/fs/xfs/xfs_rtalloc.h
+++ b/fs/xfs/xfs_rtalloc.h
@@ -77,4 +77,9 @@ xfs_growfs_check_rtgeom(const struct xfs_mount *mp,
 }
 #endif	/* CONFIG_XFS_RT */
 
+int xfs_rtallocate_rtgs(struct xfs_trans *tp, xfs_fsblock_t bno_hint,
+		xfs_rtxlen_t minlen, xfs_rtxlen_t maxlen, xfs_rtxlen_t prod,
+		bool wasdel, bool initial_user_data, xfs_rtblock_t *bno,
+		xfs_extlen_t *blen);
+
 #endif	/* __XFS_RTALLOC_H__ */

[PATCH 43/43] xfs: enable realtime reflink

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Enable reflink for realtime devices, as long as the realtime allocation
unit is a single fsblock.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
---
 fs/xfs/xfs_reflink.c |   25 +++++++++++++++++++++++++
 fs/xfs/xfs_reflink.h |    2 ++
 fs/xfs/xfs_rtalloc.c |    7 +++++--
 fs/xfs/xfs_super.c   |    6 ++++--
 4 files changed, 36 insertions(+), 4 deletions(-)


diff --git a/fs/xfs/xfs_reflink.c b/fs/xfs/xfs_reflink.c
index d9b33e22c17669..59f7fc16eb8093 100644
--- a/fs/xfs/xfs_reflink.c
+++ b/fs/xfs/xfs_reflink.c
@@ -1822,3 +1822,28 @@ xfs_reflink_unshare(
 	trace_xfs_reflink_unshare_error(ip, error, _RET_IP_);
 	return error;
 }
+
+/*
+ * Can we use reflink with this realtime extent size?  Note that we don't check
+ * for rblocks > 0 here because this can be called as part of attaching a new
+ * rt section.
+ */
+bool
+xfs_reflink_supports_rextsize(
+	struct xfs_mount	*mp,
+	unsigned int		rextsize)
+{
+	/* reflink on the realtime device requires rtgroups */
+	if (!xfs_has_rtgroups(mp))
+	       return false;
+
+	/*
+	 * Reflink doesn't support rt extent size larger than a single fsblock
+	 * because we would have to perform CoW-around for unaligned write
+	 * requests to guarantee that we always remap entire rt extents.
+	 */
+	if (rextsize != 1)
+		return false;
+
+	return true;
+}
diff --git a/fs/xfs/xfs_reflink.h b/fs/xfs/xfs_reflink.h
index 3bfd7ab9e1148a..cc4e92278279b6 100644
--- a/fs/xfs/xfs_reflink.h
+++ b/fs/xfs/xfs_reflink.h
@@ -62,4 +62,6 @@ extern int xfs_reflink_remap_blocks(struct xfs_inode *src, loff_t pos_in,
 extern int xfs_reflink_update_dest(struct xfs_inode *dest, xfs_off_t newlen,
 		xfs_extlen_t cowextsize, unsigned int remap_flags);
 
+bool xfs_reflink_supports_rextsize(struct xfs_mount *mp, unsigned int rextsize);
+
 #endif /* __XFS_REFLINK_H */
diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c
index 7135a6717a9e11..d8e6d073d64dc9 100644
--- a/fs/xfs/xfs_rtalloc.c
+++ b/fs/xfs/xfs_rtalloc.c
@@ -32,6 +32,7 @@
 #include "xfs_error.h"
 #include "xfs_trace.h"
 #include "xfs_rtrefcount_btree.h"
+#include "xfs_reflink.h"
 
 /*
  * Return whether there are any free extents in the size range given
@@ -1292,8 +1293,10 @@ xfs_growfs_rt(
 			goto out_unlock;
 		if (xfs_has_quota(mp))
 			goto out_unlock;
-	}
-	if (xfs_has_reflink(mp))
+		if (xfs_has_reflink(mp))
+			goto out_unlock;
+	} else if (xfs_has_reflink(mp) &&
+		   !xfs_reflink_supports_rextsize(mp, in->extsize))
 		goto out_unlock;
 
 	error = xfs_sb_validate_fsb_count(&mp->m_sb, in->newblocks);
diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c
index ecd5a9f444d862..7c3f996cd39e81 100644
--- a/fs/xfs/xfs_super.c
+++ b/fs/xfs/xfs_super.c
@@ -1754,9 +1754,11 @@ xfs_fs_fill_super(
 		xfs_warn_experimental(mp, XFS_EXPERIMENTAL_METADIR);
 
 	if (xfs_has_reflink(mp)) {
-		if (mp->m_sb.sb_rblocks) {
+		if (xfs_has_realtime(mp) &&
+		    !xfs_reflink_supports_rextsize(mp, mp->m_sb.sb_rextsize)) {
 			xfs_alert(mp,
-	"reflink not compatible with realtime device!");
+	"reflink not compatible with realtime extent size %u!",
+					mp->m_sb.sb_rextsize);
 			error = -EINVAL;
 			goto out_filestream_unmount;
 		}

Re: [PATCH 43/43] xfs: enable realtime reflink

[Christoph Hellwig]

Looks good:

Reviewed-by: Christoph Hellwig <hch@lst.de>

[PATCHSET v6.1 6/5] xfsprogs: last few bits of rtgroups patches

[Darrick J. Wong]

Hi all,

Since the rest of the metadir/rtgroups patches are ready to be merged
into xfsprogs 6.13, this is a funny little series with a few straggler
patches that came up during review.

If you're going to start using this code, I strongly recommend pulling
from my git trees, which are linked below.

This has been running on the djcloud for months with no problems.  Enjoy!
Comments and questions are, as always, welcome.

--D

xfsprogs git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfsprogs-dev.git/log/?h=rtgroups-stragglers
---
Commits in this patchset:
 * xfs_db: drop the metadata checking code from blockget
 * xfs_mdrestore: refactor open-coded fd/is_file into a structure
 * xfs_mdrestore: restore rt group superblocks to realtime device
---
 db/check.c                |  294 ---------------------------------------------
 man/man8/xfs_db.8         |   12 --
 man/man8/xfs_mdrestore.8  |   10 ++
 mdrestore/xfs_mdrestore.c |  162 +++++++++++++++----------
 4 files changed, 112 insertions(+), 366 deletions(-)

[PATCH 1/3] xfs_db: drop the metadata checking code from blockget

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Drop the check subcommand and all the metadata checking code from
xfs_db.  We haven't shipped xfs_check in xfsprogs in a decade and the
last known user (fstests) stopped calling it back in July 2024.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
---
 db/check.c        |  294 -----------------------------------------------------
 man/man8/xfs_db.8 |   12 +-
 2 files changed, 5 insertions(+), 301 deletions(-)


diff --git a/db/check.c b/db/check.c
index 37306bd7a6ac2d..4f7785c64f5b49 100644
--- a/db/check.c
+++ b/db/check.c
@@ -236,14 +236,12 @@ static void		check_dbmap(xfs_agnumber_t agno, xfs_agblock_t agbno,
 				    int ignore_reflink);
 static int		check_inomap(xfs_agnumber_t agno, xfs_agblock_t agbno,
 				     xfs_extlen_t len, xfs_ino_t c_ino);
-static void		check_linkcounts(xfs_agnumber_t agno);
 static int		check_range(xfs_agnumber_t agno, xfs_agblock_t agbno,
 				    xfs_extlen_t len);
 static void		check_rdbmap(xfs_rfsblock_t bno, xfs_extlen_t len,
 				     dbm_t type);
 static int		check_rinomap(xfs_rfsblock_t bno, xfs_extlen_t len,
 				      xfs_ino_t c_ino);
-static void		check_rootdir(void);
 static int		check_rrange(xfs_rfsblock_t bno, xfs_extlen_t len);
 static void		check_set_dbmap(xfs_agnumber_t agno,
 					xfs_agblock_t agbno, xfs_extlen_t len,
@@ -252,11 +250,6 @@ static void		check_set_dbmap(xfs_agnumber_t agno,
 					xfs_agblock_t c_agbno);
 static void		check_set_rdbmap(xfs_rfsblock_t bno, xfs_extlen_t len,
 					 dbm_t type1, dbm_t type2);
-static void		check_summary(void);
-static void		checknot_dbmap(xfs_agnumber_t agno, xfs_agblock_t agbno,
-				       xfs_extlen_t len, int typemask);
-static void		checknot_rdbmap(xfs_rfsblock_t bno, xfs_extlen_t len,
-					int typemask);
 static void		dir_hash_add(xfs_dahash_t hash,
 				     xfs_dir2_dataptr_t addr);
 static void		dir_hash_check(inodata_t *id, int v);
@@ -323,7 +316,6 @@ static void		quota_add(xfs_dqid_t *p, xfs_dqid_t *g, xfs_dqid_t *u,
 static void		quota_add1(qdata_t **qt, xfs_dqid_t id, int dq,
 				   xfs_qcnt_t bc, xfs_qcnt_t ic,
 				   xfs_qcnt_t rc);
-static void		quota_check(char *s, qdata_t **qt);
 static void		quota_init(void);
 static void		scan_ag(xfs_agnumber_t agno);
 static void		scan_freelist(xfs_agf_t *agf);
@@ -376,7 +368,7 @@ static const cmdinfo_t	blockfree_cmd =
 	{ "blockfree", NULL, blockfree_f, 0, 0, 0,
 	  NULL, N_("free block usage information"), NULL };
 static const cmdinfo_t	blockget_cmd =
-	{ "blockget", "check", blockget_f, 0, -1, 0,
+	{ "blockget", NULL, blockget_f, 0, -1, 0,
 	  N_("[-s|-v] [-n] [-t] [-b bno]... [-i ino] ..."),
 	  N_("get block usage and check consistency"), NULL };
 static const cmdinfo_t	blocktrash_cmd =
@@ -826,107 +818,9 @@ blockget_f(
 		blist = NULL;
 		blist_size = 0;
 	}
-	if (serious_error) {
+	if (serious_error)
 		exitcode = 2;
-		dbprefix = oldprefix;
-		return 0;
-	}
 
-	if (xfs_has_metadir(mp)) {
-		dbprefix = oldprefix;
-		return 0;
-	}
-
-	check_rootdir();
-	/*
-	 * Check that there are no blocks either
-	 * a) unaccounted for or
-	 * b) bno-free but not cnt-free
-	 */
-	if (!tflag) {	/* are we in test mode, faking out freespace? */
-		for (agno = 0; agno < mp->m_sb.sb_agcount; agno++)
-			checknot_dbmap(agno, 0, mp->m_sb.sb_agblocks,
-				(1 << DBM_UNKNOWN) | (1 << DBM_FREE1));
-	}
-	for (agno = 0; agno < mp->m_sb.sb_agcount; agno++)
-		check_linkcounts(agno);
-	if (mp->m_sb.sb_rblocks) {
-		checknot_rdbmap(0,
-			(xfs_extlen_t)(mp->m_sb.sb_rextents *
-				       mp->m_sb.sb_rextsize),
-			1 << DBM_UNKNOWN);
-		check_summary();
-	}
-	if (mp->m_sb.sb_icount != icount) {
-		if (!sflag)
-			dbprintf(_("sb_icount %lld, counted %lld\n"),
-				mp->m_sb.sb_icount, icount);
-		error++;
-	}
-	if (mp->m_sb.sb_ifree != ifree) {
-		if (!sflag)
-			dbprintf(_("sb_ifree %lld, counted %lld\n"),
-				mp->m_sb.sb_ifree, ifree);
-		error++;
-	}
-	if (mp->m_sb.sb_fdblocks != fdblocks) {
-		if (!sflag)
-			dbprintf(_("sb_fdblocks %lld, counted %lld\n"),
-				mp->m_sb.sb_fdblocks, fdblocks);
-		error++;
-	}
-	if (lazycount && mp->m_sb.sb_fdblocks != agf_aggr_freeblks) {
-		if (!sflag)
-			dbprintf(_("sb_fdblocks %lld, aggregate AGF count %lld\n"),
-				mp->m_sb.sb_fdblocks, agf_aggr_freeblks);
-		error++;
-	}
-	if (mp->m_sb.sb_frextents != frextents) {
-		if (!sflag)
-			dbprintf(_("sb_frextents %lld, counted %lld\n"),
-				mp->m_sb.sb_frextents, frextents);
-		error++;
-	}
-	if (mp->m_sb.sb_bad_features2 != 0 &&
-			mp->m_sb.sb_bad_features2 != mp->m_sb.sb_features2) {
-		if (!sflag)
-			dbprintf(_("sb_features2 (0x%x) not same as "
-				"sb_bad_features2 (0x%x)\n"),
-				mp->m_sb.sb_features2,
-				mp->m_sb.sb_bad_features2);
-		error++;
-	}
-	if ((sbversion & XFS_SB_VERSION_ATTRBIT) &&
-					!xfs_has_attr(mp)) {
-		if (!sflag)
-			dbprintf(_("sb versionnum missing attr bit %x\n"),
-				XFS_SB_VERSION_ATTRBIT);
-		error++;
-	}
-	if ((sbversion & XFS_SB_VERSION_QUOTABIT) &&
-					!xfs_has_quota(mp)) {
-		if (!sflag)
-			dbprintf(_("sb versionnum missing quota bit %x\n"),
-				XFS_SB_VERSION_QUOTABIT);
-		error++;
-	}
-	if (!(sbversion & XFS_SB_VERSION_ALIGNBIT) &&
-					xfs_has_align(mp)) {
-		if (!sflag)
-			dbprintf(_("sb versionnum extra align bit %x\n"),
-				XFS_SB_VERSION_ALIGNBIT);
-		error++;
-	}
-	if (qudo)
-		quota_check("user", qudata);
-	if (qpdo)
-		quota_check("project", qpdata);
-	if (qgdo)
-		quota_check("group", qgdata);
-	if (sbver_err > mp->m_sb.sb_agcount / 2)
-		dbprintf(_("WARNING: this may be a newer XFS filesystem.\n"));
-	if (error)
-		exitcode = 3;
 	dbprefix = oldprefix;
 	return 0;
 }
@@ -1388,58 +1282,6 @@ check_inomap(
 	return rval;
 }
 
-static void
-check_linkcounts(
-	xfs_agnumber_t	agno)
-{
-	inodata_t	*ep;
-	inodata_t	**ht;
-	int		idx;
-	char		*path;
-
-	ht = inodata[agno];
-	for (idx = 0; idx < inodata_hash_size; ht++, idx++) {
-		ep = *ht;
-		while (ep) {
-			if (ep->link_set != ep->link_add || ep->link_set == 0) {
-				path = inode_name(ep->ino, NULL);
-				if (!path && ep->link_add)
-					path = xstrdup("?");
-				if (!sflag || ep->ilist) {
-					if (ep->link_add)
-						dbprintf(_("link count mismatch "
-							 "for inode %lld (name "
-							 "%s), nlink %d, "
-							 "counted %d\n"),
-							ep->ino, path,
-							ep->link_set,
-							ep->link_add);
-					else if (ep->link_set)
-						dbprintf(_("disconnected inode "
-							 "%lld, nlink %d\n"),
-							ep->ino, ep->link_set);
-					else
-						dbprintf(_("allocated inode %lld "
-							 "has 0 link count\n"),
-							ep->ino);
-				}
-				if (path)
-					xfree(path);
-				error++;
-			} else if (verbose || ep->ilist) {
-				path = inode_name(ep->ino, NULL);
-				if (path) {
-					dbprintf(_("inode %lld name %s\n"),
-						ep->ino, path);
-					xfree(path);
-				}
-			}
-			ep = ep->next;
-		}
-	}
-
-}
-
 static int
 check_range(
 	xfs_agnumber_t  agno,
@@ -1556,25 +1398,6 @@ check_rinomap(
 	return rval;
 }
 
-static void
-check_rootdir(void)
-{
-	inodata_t	*id;
-
-	id = find_inode(mp->m_sb.sb_rootino, 0);
-	if (id == NULL) {
-		if (!sflag)
-			dbprintf(_("root inode %lld is missing\n"),
-				mp->m_sb.sb_rootino);
-		error++;
-	} else if (!id->isdir) {
-		if (!sflag || id->ilist)
-			dbprintf(_("root inode %lld is not a directory\n"),
-				mp->m_sb.sb_rootino);
-		error++;
-	}
-}
-
 static inline void
 report_rrange(
 	xfs_rfsblock_t	low,
@@ -1718,77 +1541,6 @@ get_suminfo(
 	return raw->old;
 }
 
-static void
-check_summary(void)
-{
-	xfs_rfsblock_t	bno;
-	union xfs_suminfo_raw *csp;
-	union xfs_suminfo_raw *fsp;
-	int		log;
-
-	csp = sumcompute;
-	fsp = sumfile;
-	for (log = 0; log < mp->m_rsumlevels; log++) {
-		for (bno = 0;
-		     bno < mp->m_sb.sb_rbmblocks;
-		     bno++, csp++, fsp++) {
-			if (csp->old != fsp->old) {
-				if (!sflag)
-					dbprintf(_("rt summary mismatch, size %d "
-						 "block %llu, file: %d, "
-						 "computed: %d\n"),
-						log, bno,
-						get_suminfo(mp, fsp),
-						get_suminfo(mp, csp));
-				error++;
-			}
-		}
-	}
-}
-
-static void
-checknot_dbmap(
-	xfs_agnumber_t	agno,
-	xfs_agblock_t	agbno,
-	xfs_extlen_t	len,
-	int		typemask)
-{
-	xfs_extlen_t	i;
-	char		*p;
-
-	if (!check_range(agno, agbno, len))
-		return;
-	for (i = 0, p = &dbmap[agno][agbno]; i < len; i++, p++) {
-		if ((1 << *p) & typemask) {
-			if (!sflag || CHECK_BLISTA(agno, agbno + i))
-				dbprintf(_("block %u/%u type %s not expected\n"),
-					agno, agbno + i, typename[(dbm_t)*p]);
-			error++;
-		}
-	}
-}
-
-static void
-checknot_rdbmap(
-	xfs_rfsblock_t	bno,
-	xfs_extlen_t	len,
-	int		typemask)
-{
-	xfs_extlen_t	i;
-	char		*p;
-
-	if (!check_rrange(bno, len))
-		return;
-	for (i = 0, p = &dbmap[mp->m_sb.sb_agcount][bno]; i < len; i++, p++) {
-		if ((1 << *p) & typemask) {
-			if (!sflag || CHECK_BLIST(bno + i))
-				dbprintf(_("rtblock %llu type %s not expected\n"),
-					bno + i, typename[(dbm_t)*p]);
-			error++;
-		}
-	}
-}
-
 static void
 dir_hash_add(
 	xfs_dahash_t		hash,
@@ -3923,48 +3675,6 @@ quota_add1(
 	qt[qh] = qe;
 }
 
-static void
-quota_check(
-	char	*s,
-	qdata_t	**qt)
-{
-	int	i;
-	qdata_t	*next;
-	qdata_t	*qp;
-
-	for (i = 0; i < QDATA_HASH_SIZE; i++) {
-		qp = qt[i];
-		while (qp) {
-			next = qp->next;
-			if (qp->count.bc != qp->dq.bc ||
-			    qp->count.ic != qp->dq.ic ||
-			    qp->count.rc != qp->dq.rc) {
-				if (!sflag) {
-					dbprintf(_("%s quota id %u, have/exp"),
-						s, qp->id);
-					if (qp->count.bc != qp->dq.bc)
-						dbprintf(_(" bc %lld/%lld"),
-							qp->dq.bc,
-							qp->count.bc);
-					if (qp->count.ic != qp->dq.ic)
-						dbprintf(_(" ic %lld/%lld"),
-							qp->dq.ic,
-							qp->count.ic);
-					if (qp->count.rc != qp->dq.rc)
-						dbprintf(_(" rc %lld/%lld"),
-							qp->dq.rc,
-							qp->count.rc);
-					dbprintf("\n");
-				}
-				error++;
-			}
-			xfree(qp);
-			qp = next;
-		}
-	}
-	xfree(qt);
-}
-
 static void
 quota_init(void)
 {
diff --git a/man/man8/xfs_db.8 b/man/man8/xfs_db.8
index 5d72de91dd6862..06f4464a928596 100644
--- a/man/man8/xfs_db.8
+++ b/man/man8/xfs_db.8
@@ -345,7 +345,7 @@ .SH COMMANDS
 command can be given, presumably with different arguments than the previous one.
 .TP
 .BI "blockget [\-npvs] [\-b " bno "] ... [\-i " ino "] ..."
-Get block usage and check filesystem consistency.
+Get block usage.
 The information is saved for use by a subsequent
 .BR blockuse ", " ncheck ", or " blocktrash
 command.
@@ -564,11 +564,6 @@ .SH COMMANDS
 half full.
 .RE
 .TP
-.B check
-See the
-.B blockget
-command.
-.TP
 .BI "convert " "type number" " [" "type number" "] ... " type
 Convert from one address form to another.
 The known
@@ -2665,8 +2660,7 @@ .SH TYPES
 and printable ASCII chars.
 .SH DIAGNOSTICS
 Many messages can come from the
-.B check
-.RB ( blockget )
+.B blockget
 command.
 If the filesystem is completely corrupt, a core dump might
 be produced instead of the message
@@ -2676,7 +2670,7 @@ .SH DIAGNOSTICS
 .RE
 .PP
 If the filesystem is very large (has many files) then
-.B check
+.B blockget
 might run out of memory. In this case the message
 .RS
 .B out of memory

Re: [PATCH 1/3] xfs_db: drop the metadata checking code from blockget

[Andrey Albershteyn]

On 2024-12-19 11:44:29, Darrick J. Wong wrote:
> From: Darrick J. Wong <djwong@kernel.org>
> 
> Drop the check subcommand and all the metadata checking code from
> xfs_db.  We haven't shipped xfs_check in xfsprogs in a decade and the
> last known user (fstests) stopped calling it back in July 2024.
> 
> Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>

LGTM
Reviewed-by: Andrey Albershteyn <aalbersh@kernel.org>

> ---
>  db/check.c        |  294 -----------------------------------------------------
>  man/man8/xfs_db.8 |   12 +-
>  2 files changed, 5 insertions(+), 301 deletions(-)
> 
> 
> diff --git a/db/check.c b/db/check.c
> index 37306bd7a6ac2d..4f7785c64f5b49 100644
> --- a/db/check.c
> +++ b/db/check.c
> @@ -236,14 +236,12 @@ static void		check_dbmap(xfs_agnumber_t agno, xfs_agblock_t agbno,
>  				    int ignore_reflink);
>  static int		check_inomap(xfs_agnumber_t agno, xfs_agblock_t agbno,
>  				     xfs_extlen_t len, xfs_ino_t c_ino);
> -static void		check_linkcounts(xfs_agnumber_t agno);
>  static int		check_range(xfs_agnumber_t agno, xfs_agblock_t agbno,
>  				    xfs_extlen_t len);
>  static void		check_rdbmap(xfs_rfsblock_t bno, xfs_extlen_t len,
>  				     dbm_t type);
>  static int		check_rinomap(xfs_rfsblock_t bno, xfs_extlen_t len,
>  				      xfs_ino_t c_ino);
> -static void		check_rootdir(void);
>  static int		check_rrange(xfs_rfsblock_t bno, xfs_extlen_t len);
>  static void		check_set_dbmap(xfs_agnumber_t agno,
>  					xfs_agblock_t agbno, xfs_extlen_t len,
> @@ -252,11 +250,6 @@ static void		check_set_dbmap(xfs_agnumber_t agno,
>  					xfs_agblock_t c_agbno);
>  static void		check_set_rdbmap(xfs_rfsblock_t bno, xfs_extlen_t len,
>  					 dbm_t type1, dbm_t type2);
> -static void		check_summary(void);
> -static void		checknot_dbmap(xfs_agnumber_t agno, xfs_agblock_t agbno,
> -				       xfs_extlen_t len, int typemask);
> -static void		checknot_rdbmap(xfs_rfsblock_t bno, xfs_extlen_t len,
> -					int typemask);
>  static void		dir_hash_add(xfs_dahash_t hash,
>  				     xfs_dir2_dataptr_t addr);
>  static void		dir_hash_check(inodata_t *id, int v);
> @@ -323,7 +316,6 @@ static void		quota_add(xfs_dqid_t *p, xfs_dqid_t *g, xfs_dqid_t *u,
>  static void		quota_add1(qdata_t **qt, xfs_dqid_t id, int dq,
>  				   xfs_qcnt_t bc, xfs_qcnt_t ic,
>  				   xfs_qcnt_t rc);
> -static void		quota_check(char *s, qdata_t **qt);
>  static void		quota_init(void);
>  static void		scan_ag(xfs_agnumber_t agno);
>  static void		scan_freelist(xfs_agf_t *agf);
> @@ -376,7 +368,7 @@ static const cmdinfo_t	blockfree_cmd =
>  	{ "blockfree", NULL, blockfree_f, 0, 0, 0,
>  	  NULL, N_("free block usage information"), NULL };
>  static const cmdinfo_t	blockget_cmd =
> -	{ "blockget", "check", blockget_f, 0, -1, 0,
> +	{ "blockget", NULL, blockget_f, 0, -1, 0,
>  	  N_("[-s|-v] [-n] [-t] [-b bno]... [-i ino] ..."),
>  	  N_("get block usage and check consistency"), NULL };
>  static const cmdinfo_t	blocktrash_cmd =
> @@ -826,107 +818,9 @@ blockget_f(
>  		blist = NULL;
>  		blist_size = 0;
>  	}
> -	if (serious_error) {
> +	if (serious_error)
>  		exitcode = 2;
> -		dbprefix = oldprefix;
> -		return 0;
> -	}
>  
> -	if (xfs_has_metadir(mp)) {
> -		dbprefix = oldprefix;
> -		return 0;
> -	}
> -
> -	check_rootdir();
> -	/*
> -	 * Check that there are no blocks either
> -	 * a) unaccounted for or
> -	 * b) bno-free but not cnt-free
> -	 */
> -	if (!tflag) {	/* are we in test mode, faking out freespace? */
> -		for (agno = 0; agno < mp->m_sb.sb_agcount; agno++)
> -			checknot_dbmap(agno, 0, mp->m_sb.sb_agblocks,
> -				(1 << DBM_UNKNOWN) | (1 << DBM_FREE1));
> -	}
> -	for (agno = 0; agno < mp->m_sb.sb_agcount; agno++)
> -		check_linkcounts(agno);
> -	if (mp->m_sb.sb_rblocks) {
> -		checknot_rdbmap(0,
> -			(xfs_extlen_t)(mp->m_sb.sb_rextents *
> -				       mp->m_sb.sb_rextsize),
> -			1 << DBM_UNKNOWN);
> -		check_summary();
> -	}
> -	if (mp->m_sb.sb_icount != icount) {
> -		if (!sflag)
> -			dbprintf(_("sb_icount %lld, counted %lld\n"),
> -				mp->m_sb.sb_icount, icount);
> -		error++;
> -	}
> -	if (mp->m_sb.sb_ifree != ifree) {
> -		if (!sflag)
> -			dbprintf(_("sb_ifree %lld, counted %lld\n"),
> -				mp->m_sb.sb_ifree, ifree);
> -		error++;
> -	}
> -	if (mp->m_sb.sb_fdblocks != fdblocks) {
> -		if (!sflag)
> -			dbprintf(_("sb_fdblocks %lld, counted %lld\n"),
> -				mp->m_sb.sb_fdblocks, fdblocks);
> -		error++;
> -	}
> -	if (lazycount && mp->m_sb.sb_fdblocks != agf_aggr_freeblks) {
> -		if (!sflag)
> -			dbprintf(_("sb_fdblocks %lld, aggregate AGF count %lld\n"),
> -				mp->m_sb.sb_fdblocks, agf_aggr_freeblks);
> -		error++;
> -	}
> -	if (mp->m_sb.sb_frextents != frextents) {
> -		if (!sflag)
> -			dbprintf(_("sb_frextents %lld, counted %lld\n"),
> -				mp->m_sb.sb_frextents, frextents);
> -		error++;
> -	}
> -	if (mp->m_sb.sb_bad_features2 != 0 &&
> -			mp->m_sb.sb_bad_features2 != mp->m_sb.sb_features2) {
> -		if (!sflag)
> -			dbprintf(_("sb_features2 (0x%x) not same as "
> -				"sb_bad_features2 (0x%x)\n"),
> -				mp->m_sb.sb_features2,
> -				mp->m_sb.sb_bad_features2);
> -		error++;
> -	}
> -	if ((sbversion & XFS_SB_VERSION_ATTRBIT) &&
> -					!xfs_has_attr(mp)) {
> -		if (!sflag)
> -			dbprintf(_("sb versionnum missing attr bit %x\n"),
> -				XFS_SB_VERSION_ATTRBIT);
> -		error++;
> -	}
> -	if ((sbversion & XFS_SB_VERSION_QUOTABIT) &&
> -					!xfs_has_quota(mp)) {
> -		if (!sflag)
> -			dbprintf(_("sb versionnum missing quota bit %x\n"),
> -				XFS_SB_VERSION_QUOTABIT);
> -		error++;
> -	}
> -	if (!(sbversion & XFS_SB_VERSION_ALIGNBIT) &&
> -					xfs_has_align(mp)) {
> -		if (!sflag)
> -			dbprintf(_("sb versionnum extra align bit %x\n"),
> -				XFS_SB_VERSION_ALIGNBIT);
> -		error++;
> -	}
> -	if (qudo)
> -		quota_check("user", qudata);
> -	if (qpdo)
> -		quota_check("project", qpdata);
> -	if (qgdo)
> -		quota_check("group", qgdata);
> -	if (sbver_err > mp->m_sb.sb_agcount / 2)
> -		dbprintf(_("WARNING: this may be a newer XFS filesystem.\n"));
> -	if (error)
> -		exitcode = 3;
>  	dbprefix = oldprefix;
>  	return 0;
>  }
> @@ -1388,58 +1282,6 @@ check_inomap(
>  	return rval;
>  }
>  
> -static void
> -check_linkcounts(
> -	xfs_agnumber_t	agno)
> -{
> -	inodata_t	*ep;
> -	inodata_t	**ht;
> -	int		idx;
> -	char		*path;
> -
> -	ht = inodata[agno];
> -	for (idx = 0; idx < inodata_hash_size; ht++, idx++) {
> -		ep = *ht;
> -		while (ep) {
> -			if (ep->link_set != ep->link_add || ep->link_set == 0) {
> -				path = inode_name(ep->ino, NULL);
> -				if (!path && ep->link_add)
> -					path = xstrdup("?");
> -				if (!sflag || ep->ilist) {
> -					if (ep->link_add)
> -						dbprintf(_("link count mismatch "
> -							 "for inode %lld (name "
> -							 "%s), nlink %d, "
> -							 "counted %d\n"),
> -							ep->ino, path,
> -							ep->link_set,
> -							ep->link_add);
> -					else if (ep->link_set)
> -						dbprintf(_("disconnected inode "
> -							 "%lld, nlink %d\n"),
> -							ep->ino, ep->link_set);
> -					else
> -						dbprintf(_("allocated inode %lld "
> -							 "has 0 link count\n"),
> -							ep->ino);
> -				}
> -				if (path)
> -					xfree(path);
> -				error++;
> -			} else if (verbose || ep->ilist) {
> -				path = inode_name(ep->ino, NULL);
> -				if (path) {
> -					dbprintf(_("inode %lld name %s\n"),
> -						ep->ino, path);
> -					xfree(path);
> -				}
> -			}
> -			ep = ep->next;
> -		}
> -	}
> -
> -}
> -
>  static int
>  check_range(
>  	xfs_agnumber_t  agno,
> @@ -1556,25 +1398,6 @@ check_rinomap(
>  	return rval;
>  }
>  
> -static void
> -check_rootdir(void)
> -{
> -	inodata_t	*id;
> -
> -	id = find_inode(mp->m_sb.sb_rootino, 0);
> -	if (id == NULL) {
> -		if (!sflag)
> -			dbprintf(_("root inode %lld is missing\n"),
> -				mp->m_sb.sb_rootino);
> -		error++;
> -	} else if (!id->isdir) {
> -		if (!sflag || id->ilist)
> -			dbprintf(_("root inode %lld is not a directory\n"),
> -				mp->m_sb.sb_rootino);
> -		error++;
> -	}
> -}
> -
>  static inline void
>  report_rrange(
>  	xfs_rfsblock_t	low,
> @@ -1718,77 +1541,6 @@ get_suminfo(
>  	return raw->old;
>  }
>  
> -static void
> -check_summary(void)
> -{
> -	xfs_rfsblock_t	bno;
> -	union xfs_suminfo_raw *csp;
> -	union xfs_suminfo_raw *fsp;
> -	int		log;
> -
> -	csp = sumcompute;
> -	fsp = sumfile;
> -	for (log = 0; log < mp->m_rsumlevels; log++) {
> -		for (bno = 0;
> -		     bno < mp->m_sb.sb_rbmblocks;
> -		     bno++, csp++, fsp++) {
> -			if (csp->old != fsp->old) {
> -				if (!sflag)
> -					dbprintf(_("rt summary mismatch, size %d "
> -						 "block %llu, file: %d, "
> -						 "computed: %d\n"),
> -						log, bno,
> -						get_suminfo(mp, fsp),
> -						get_suminfo(mp, csp));
> -				error++;
> -			}
> -		}
> -	}
> -}
> -
> -static void
> -checknot_dbmap(
> -	xfs_agnumber_t	agno,
> -	xfs_agblock_t	agbno,
> -	xfs_extlen_t	len,
> -	int		typemask)
> -{
> -	xfs_extlen_t	i;
> -	char		*p;
> -
> -	if (!check_range(agno, agbno, len))
> -		return;
> -	for (i = 0, p = &dbmap[agno][agbno]; i < len; i++, p++) {
> -		if ((1 << *p) & typemask) {
> -			if (!sflag || CHECK_BLISTA(agno, agbno + i))
> -				dbprintf(_("block %u/%u type %s not expected\n"),
> -					agno, agbno + i, typename[(dbm_t)*p]);
> -			error++;
> -		}
> -	}
> -}
> -
> -static void
> -checknot_rdbmap(
> -	xfs_rfsblock_t	bno,
> -	xfs_extlen_t	len,
> -	int		typemask)
> -{
> -	xfs_extlen_t	i;
> -	char		*p;
> -
> -	if (!check_rrange(bno, len))
> -		return;
> -	for (i = 0, p = &dbmap[mp->m_sb.sb_agcount][bno]; i < len; i++, p++) {
> -		if ((1 << *p) & typemask) {
> -			if (!sflag || CHECK_BLIST(bno + i))
> -				dbprintf(_("rtblock %llu type %s not expected\n"),
> -					bno + i, typename[(dbm_t)*p]);
> -			error++;
> -		}
> -	}
> -}
> -
>  static void
>  dir_hash_add(
>  	xfs_dahash_t		hash,
> @@ -3923,48 +3675,6 @@ quota_add1(
>  	qt[qh] = qe;
>  }
>  
> -static void
> -quota_check(
> -	char	*s,
> -	qdata_t	**qt)
> -{
> -	int	i;
> -	qdata_t	*next;
> -	qdata_t	*qp;
> -
> -	for (i = 0; i < QDATA_HASH_SIZE; i++) {
> -		qp = qt[i];
> -		while (qp) {
> -			next = qp->next;
> -			if (qp->count.bc != qp->dq.bc ||
> -			    qp->count.ic != qp->dq.ic ||
> -			    qp->count.rc != qp->dq.rc) {
> -				if (!sflag) {
> -					dbprintf(_("%s quota id %u, have/exp"),
> -						s, qp->id);
> -					if (qp->count.bc != qp->dq.bc)
> -						dbprintf(_(" bc %lld/%lld"),
> -							qp->dq.bc,
> -							qp->count.bc);
> -					if (qp->count.ic != qp->dq.ic)
> -						dbprintf(_(" ic %lld/%lld"),
> -							qp->dq.ic,
> -							qp->count.ic);
> -					if (qp->count.rc != qp->dq.rc)
> -						dbprintf(_(" rc %lld/%lld"),
> -							qp->dq.rc,
> -							qp->count.rc);
> -					dbprintf("\n");
> -				}
> -				error++;
> -			}
> -			xfree(qp);
> -			qp = next;
> -		}
> -	}
> -	xfree(qt);
> -}
> -
>  static void
>  quota_init(void)
>  {
> diff --git a/man/man8/xfs_db.8 b/man/man8/xfs_db.8
> index 5d72de91dd6862..06f4464a928596 100644
> --- a/man/man8/xfs_db.8
> +++ b/man/man8/xfs_db.8
> @@ -345,7 +345,7 @@ .SH COMMANDS
>  command can be given, presumably with different arguments than the previous one.
>  .TP
>  .BI "blockget [\-npvs] [\-b " bno "] ... [\-i " ino "] ..."
> -Get block usage and check filesystem consistency.
> +Get block usage.
>  The information is saved for use by a subsequent
>  .BR blockuse ", " ncheck ", or " blocktrash
>  command.
> @@ -564,11 +564,6 @@ .SH COMMANDS
>  half full.
>  .RE
>  .TP
> -.B check
> -See the
> -.B blockget
> -command.
> -.TP
>  .BI "convert " "type number" " [" "type number" "] ... " type
>  Convert from one address form to another.
>  The known
> @@ -2665,8 +2660,7 @@ .SH TYPES
>  and printable ASCII chars.
>  .SH DIAGNOSTICS
>  Many messages can come from the
> -.B check
> -.RB ( blockget )
> +.B blockget
>  command.
>  If the filesystem is completely corrupt, a core dump might
>  be produced instead of the message
> @@ -2676,7 +2670,7 @@ .SH DIAGNOSTICS
>  .RE
>  .PP
>  If the filesystem is very large (has many files) then
> -.B check
> +.B blockget
>  might run out of memory. In this case the message
>  .RS
>  .B out of memory
> 

Re: [PATCH 1/3] xfs_db: drop the metadata checking code from blockget

[Christoph Hellwig]

Finally..

Reviewed-by: Christoph Hellwig <hch@lst.de>

[PATCH 2/3] xfs_mdrestore: refactor open-coded fd/is_file into a structure

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Create an explicit object to track the fd and flags associated with a
device onto which we are restoring metadata, and use it to reduce the
amount of open-coded arguments to ->restore.  This avoids some grossness
in the next patch.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
---
 mdrestore/xfs_mdrestore.c |  123 +++++++++++++++++++++++----------------------
 1 file changed, 63 insertions(+), 60 deletions(-)


diff --git a/mdrestore/xfs_mdrestore.c b/mdrestore/xfs_mdrestore.c
index c6c00270234442..c5584fec68813e 100644
--- a/mdrestore/xfs_mdrestore.c
+++ b/mdrestore/xfs_mdrestore.c
@@ -15,12 +15,20 @@ union mdrestore_headers {
 	struct xfs_metadump_header	v2;
 };
 
+struct mdrestore_dev {
+	int		fd;
+	bool		is_file;
+};
+
+#define DEFINE_MDRESTORE_DEV(name) \
+	struct mdrestore_dev name = { .fd = -1 }
+
 struct mdrestore_ops {
 	void (*read_header)(union mdrestore_headers *header, FILE *md_fp);
 	void (*show_info)(union mdrestore_headers *header, const char *md_file);
 	void (*restore)(union mdrestore_headers *header, FILE *md_fp,
-			int ddev_fd, bool is_data_target_file, int logdev_fd,
-			bool is_log_target_file);
+			const struct mdrestore_dev *ddev,
+			const struct mdrestore_dev *logdev);
 };
 
 static struct mdrestore {
@@ -108,25 +116,24 @@ fixup_superblock(
 		fatal("error writing primary superblock: %s\n", strerror(errno));
 }
 
-static int
+static void
 open_device(
-	char		*path,
-	bool		*is_file)
+	struct mdrestore_dev	*dev,
+	char			*path)
 {
-	struct stat	statbuf;
-	int		open_flags;
-	int		fd;
+	struct stat		statbuf;
+	int			open_flags;
 
 	open_flags = O_RDWR;
-	*is_file = false;
+	dev->is_file = false;
 
 	if (stat(path, &statbuf) < 0)  {
 		/* ok, assume it's a file and create it */
 		open_flags |= O_CREAT;
-		*is_file = true;
+		dev->is_file = true;
 	} else if (S_ISREG(statbuf.st_mode))  {
 		open_flags |= O_TRUNC;
-		*is_file = true;
+		dev->is_file = true;
 	} else if (platform_check_ismounted(path, NULL, &statbuf, 0)) {
 		/*
 		 * check to make sure a filesystem isn't mounted on the device
@@ -136,23 +143,30 @@ open_device(
 				path);
 	}
 
-	fd = open(path, open_flags, 0644);
-	if (fd < 0)
+	dev->fd = open(path, open_flags, 0644);
+	if (dev->fd < 0)
 		fatal("couldn't open \"%s\"\n", path);
+}
 
-	return fd;
+static void
+close_device(
+	struct mdrestore_dev	*dev)
+{
+	if (dev->fd >= 0)
+		close(dev->fd);
+	dev->fd = -1;
+	dev->is_file = false;
 }
 
 static void
 verify_device_size(
-	int		dev_fd,
-	bool		is_file,
-	xfs_rfsblock_t	nr_blocks,
-	uint32_t	blocksize)
+	const struct mdrestore_dev	*dev,
+	xfs_rfsblock_t			nr_blocks,
+	uint32_t			blocksize)
 {
-	if (is_file) {
+	if (dev->is_file) {
 		/* ensure regular files are correctly sized */
-		if (ftruncate(dev_fd, nr_blocks * blocksize))
+		if (ftruncate(dev->fd, nr_blocks * blocksize))
 			fatal("cannot set filesystem image size: %s\n",
 				strerror(errno));
 	} else {
@@ -161,7 +175,7 @@ verify_device_size(
 		off_t		off;
 
 		off = nr_blocks * blocksize - sizeof(lb);
-		if (pwrite(dev_fd, lb, sizeof(lb), off) < 0)
+		if (pwrite(dev->fd, lb, sizeof(lb), off) < 0)
 			fatal("failed to write last block, is target too "
 				"small? (error: %s)\n", strerror(errno));
 	}
@@ -195,12 +209,10 @@ show_info_v1(
 
 static void
 restore_v1(
-	union mdrestore_headers *h,
-	FILE			*md_fp,
-	int			ddev_fd,
-	bool			is_data_target_file,
-	int			logdev_fd,
-	bool			is_log_target_file)
+	union mdrestore_headers		*h,
+	FILE				*md_fp,
+	const struct mdrestore_dev	*ddev,
+	const struct mdrestore_dev	*logdev)
 {
 	struct xfs_metablock	*metablock;	/* header + index + blocks */
 	__be64			*block_index;
@@ -254,8 +266,7 @@ restore_v1(
 
 	((struct xfs_dsb*)block_buffer)->sb_inprogress = 1;
 
-	verify_device_size(ddev_fd, is_data_target_file, sb.sb_dblocks,
-			sb.sb_blocksize);
+	verify_device_size(ddev, sb.sb_dblocks, sb.sb_blocksize);
 
 	bytes_read = 0;
 
@@ -263,7 +274,7 @@ restore_v1(
 		maybe_print_progress(&mb_read, bytes_read);
 
 		for (cur_index = 0; cur_index < mb_count; cur_index++) {
-			if (pwrite(ddev_fd, &block_buffer[cur_index <<
+			if (pwrite(ddev->fd, &block_buffer[cur_index <<
 					h->v1.mb_blocklog], block_size,
 					be64_to_cpu(block_index[cur_index]) <<
 						BBSHIFT) < 0)
@@ -292,7 +303,7 @@ restore_v1(
 
 	final_print_progress(&mb_read, bytes_read);
 
-	fixup_superblock(ddev_fd, block_buffer, &sb);
+	fixup_superblock(ddev->fd, block_buffer, &sb);
 
 	free(metablock);
 }
@@ -376,12 +387,10 @@ restore_meta_extent(
 
 static void
 restore_v2(
-	union mdrestore_headers	*h,
-	FILE			*md_fp,
-	int			ddev_fd,
-	bool			is_data_target_file,
-	int			logdev_fd,
-	bool			is_log_target_file)
+	union mdrestore_headers		*h,
+	FILE				*md_fp,
+	const struct mdrestore_dev	*ddev,
+	const struct mdrestore_dev	*logdev)
 {
 	struct xfs_sb		sb;
 	struct xfs_meta_extent	xme;
@@ -415,16 +424,14 @@ restore_v2(
 
 	((struct xfs_dsb *)block_buffer)->sb_inprogress = 1;
 
-	verify_device_size(ddev_fd, is_data_target_file, sb.sb_dblocks,
-			sb.sb_blocksize);
+	verify_device_size(ddev, sb.sb_dblocks, sb.sb_blocksize);
 
 	if (sb.sb_logstart == 0) {
 		ASSERT(mdrestore.external_log == true);
-		verify_device_size(logdev_fd, is_log_target_file, sb.sb_logblocks,
-				sb.sb_blocksize);
+		verify_device_size(logdev, sb.sb_logblocks, sb.sb_blocksize);
 	}
 
-	if (pwrite(ddev_fd, block_buffer, len, 0) < 0)
+	if (pwrite(ddev->fd, block_buffer, len, 0) < 0)
 		fatal("error writing primary superblock: %s\n",
 			strerror(errno));
 
@@ -446,11 +453,11 @@ restore_v2(
 		switch (be64_to_cpu(xme.xme_addr) & XME_ADDR_DEVICE_MASK) {
 		case XME_ADDR_DATA_DEVICE:
 			device = "data";
-			fd = ddev_fd;
+			fd = ddev->fd;
 			break;
 		case XME_ADDR_LOG_DEVICE:
 			device = "log";
-			fd = logdev_fd;
+			fd = logdev->fd;
 			break;
 		default:
 			fatal("Invalid device found in metadump\n");
@@ -467,7 +474,7 @@ restore_v2(
 
 	final_print_progress(&mb_read, bytes_read);
 
-	fixup_superblock(ddev_fd, block_buffer, &sb);
+	fixup_superblock(ddev->fd, block_buffer, &sb);
 
 	free(block_buffer);
 }
@@ -492,13 +499,11 @@ main(
 	char			**argv)
 {
 	union mdrestore_headers	headers;
+	DEFINE_MDRESTORE_DEV(ddev);
+	DEFINE_MDRESTORE_DEV(logdev);
 	FILE			*src_f;
-	char			*logdev = NULL;
-	int			data_dev_fd = -1;
-	int			log_dev_fd = -1;
+	char			*logdev_path = NULL;
 	int			c;
-	bool			is_data_dev_file = false;
-	bool			is_log_dev_file = false;
 
 	mdrestore.show_progress = false;
 	mdrestore.show_info = false;
@@ -516,7 +521,7 @@ main(
 				mdrestore.show_info = true;
 				break;
 			case 'l':
-				logdev = optarg;
+				logdev_path = optarg;
 				mdrestore.external_log = true;
 				break;
 			case 'V':
@@ -555,7 +560,7 @@ main(
 
 	switch (be32_to_cpu(headers.magic)) {
 	case XFS_MD_MAGIC_V1:
-		if (logdev != NULL)
+		if (logdev_path != NULL)
 			usage();
 		mdrestore.mdrops = &mdrestore_ops_v1;
 		break;
@@ -581,18 +586,16 @@ main(
 	optind++;
 
 	/* check and open data device */
-	data_dev_fd = open_device(argv[optind], &is_data_dev_file);
+	open_device(&ddev, argv[optind]);
 
+	/* check and open log device */
 	if (mdrestore.external_log)
-		/* check and open log device */
-		log_dev_fd = open_device(logdev, &is_log_dev_file);
+		open_device(&logdev, logdev_path);
 
-	mdrestore.mdrops->restore(&headers, src_f, data_dev_fd,
-			is_data_dev_file, log_dev_fd, is_log_dev_file);
+	mdrestore.mdrops->restore(&headers, src_f, &ddev, &logdev);
 
-	close(data_dev_fd);
-	if (mdrestore.external_log)
-		close(log_dev_fd);
+	close_device(&ddev);
+	close_device(&logdev);
 
 	if (src_f != stdin)
 		fclose(src_f);

Re: [PATCH 2/3] xfs_mdrestore: refactor open-coded fd/is_file into a structure

[Christoph Hellwig]

Looks good:

Reviewed-by: Christoph Hellwig <hch@lst.de>

[PATCH 3/3] xfs_mdrestore: restore rt group superblocks to realtime device

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Support restoring realtime device metadata to the realtime device, if
the dumped filesystem had one.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
---
 man/man8/xfs_mdrestore.8  |   10 ++++++++++
 mdrestore/xfs_mdrestore.c |   47 ++++++++++++++++++++++++++++++++++++---------
 2 files changed, 48 insertions(+), 9 deletions(-)


diff --git a/man/man8/xfs_mdrestore.8 b/man/man8/xfs_mdrestore.8
index f60e7b56ebf0d1..6f6e14e96c6a5c 100644
--- a/man/man8/xfs_mdrestore.8
+++ b/man/man8/xfs_mdrestore.8
@@ -8,6 +8,9 @@ .SH SYNOPSIS
 ] [
 .B \-l
 .I logdev
+] [
+.B \-r
+.I rtdev
 ]
 .I source
 .I target
@@ -17,6 +20,9 @@ .SH SYNOPSIS
 [
 .B \-l
 .I logdev
+] [
+.B \-r
+.I rtdev
 ]
 .I source
 .br
@@ -61,6 +67,10 @@ .SH OPTIONS
 In such a scenario, the user has to provide a device to which the log device
 contents from the metadump file are copied.
 .TP
+.BI \-r " rtdev"
+Restore realtime device metadata to this device.
+This is only required for a metadump in v2 format.
+.TP
 .B \-V
 Prints the version number and exits.
 .SH DIAGNOSTICS
diff --git a/mdrestore/xfs_mdrestore.c b/mdrestore/xfs_mdrestore.c
index c5584fec68813e..d5014981b15a68 100644
--- a/mdrestore/xfs_mdrestore.c
+++ b/mdrestore/xfs_mdrestore.c
@@ -28,7 +28,8 @@ struct mdrestore_ops {
 	void (*show_info)(union mdrestore_headers *header, const char *md_file);
 	void (*restore)(union mdrestore_headers *header, FILE *md_fp,
 			const struct mdrestore_dev *ddev,
-			const struct mdrestore_dev *logdev);
+			const struct mdrestore_dev *logdev,
+			const struct mdrestore_dev *rtdev);
 };
 
 static struct mdrestore {
@@ -37,6 +38,7 @@ static struct mdrestore {
 	bool			show_info;
 	bool			progress_since_warning;
 	bool			external_log;
+	bool			realtime_data;
 } mdrestore;
 
 static void
@@ -212,7 +214,8 @@ restore_v1(
 	union mdrestore_headers		*h,
 	FILE				*md_fp,
 	const struct mdrestore_dev	*ddev,
-	const struct mdrestore_dev	*logdev)
+	const struct mdrestore_dev	*logdev,
+	const struct mdrestore_dev	*rtdev)
 {
 	struct xfs_metablock	*metablock;	/* header + index + blocks */
 	__be64			*block_index;
@@ -336,8 +339,9 @@ read_header_v2(
 	if (!mdrestore.external_log && (compat & XFS_MD2_COMPAT_EXTERNALLOG))
 		fatal("External Log device is required\n");
 
-	if (h->v2.xmh_incompat_flags & cpu_to_be32(XFS_MD2_INCOMPAT_RTDEVICE))
-		fatal("Realtime device not yet supported\n");
+	if ((h->v2.xmh_incompat_flags & cpu_to_be32(XFS_MD2_INCOMPAT_RTDEVICE)) &&
+	    !mdrestore.realtime_data)
+		fatal("Realtime device is required\n");
 }
 
 static void
@@ -346,14 +350,17 @@ show_info_v2(
 	const char		*md_file)
 {
 	uint32_t		compat_flags;
+	uint32_t		incompat_flags;
 
 	compat_flags = be32_to_cpu(h->v2.xmh_compat_flags);
+	incompat_flags = be32_to_cpu(h->v2.xmh_incompat_flags);
 
-	printf("%s: %sobfuscated, %s log, external log contents are %sdumped, %s metadata blocks,\n",
+	printf("%s: %sobfuscated, %s log, external log contents are %sdumped, rt device contents are %sdumped, %s metadata blocks,\n",
 		md_file,
 		compat_flags & XFS_MD2_COMPAT_OBFUSCATED ? "":"not ",
 		compat_flags & XFS_MD2_COMPAT_DIRTYLOG ? "dirty":"clean",
 		compat_flags & XFS_MD2_COMPAT_EXTERNALLOG ? "":"not ",
+		incompat_flags & XFS_MD2_INCOMPAT_RTDEVICE ? "":"not ",
 		compat_flags & XFS_MD2_COMPAT_FULLBLOCKS ? "full":"zeroed");
 }
 
@@ -390,7 +397,8 @@ restore_v2(
 	union mdrestore_headers		*h,
 	FILE				*md_fp,
 	const struct mdrestore_dev	*ddev,
-	const struct mdrestore_dev	*logdev)
+	const struct mdrestore_dev	*logdev,
+	const struct mdrestore_dev	*rtdev)
 {
 	struct xfs_sb		sb;
 	struct xfs_meta_extent	xme;
@@ -431,6 +439,11 @@ restore_v2(
 		verify_device_size(logdev, sb.sb_logblocks, sb.sb_blocksize);
 	}
 
+	if (sb.sb_rblocks > 0) {
+		ASSERT(mdrestore.realtime_data == true);
+		verify_device_size(rtdev, sb.sb_rblocks, sb.sb_blocksize);
+	}
+
 	if (pwrite(ddev->fd, block_buffer, len, 0) < 0)
 		fatal("error writing primary superblock: %s\n",
 			strerror(errno));
@@ -459,6 +472,10 @@ restore_v2(
 			device = "log";
 			fd = logdev->fd;
 			break;
+		case XME_ADDR_RT_DEVICE:
+			device = "rt";
+			fd = rtdev->fd;
+			break;
 		default:
 			fatal("Invalid device found in metadump\n");
 			break;
@@ -488,7 +505,7 @@ static struct mdrestore_ops mdrestore_ops_v2 = {
 static void
 usage(void)
 {
-	fprintf(stderr, "Usage: %s [-V] [-g] [-i] [-l logdev] source target\n",
+	fprintf(stderr, "Usage: %s [-V] [-g] [-i] [-l logdev] [-r rtdev] source target\n",
 		progname);
 	exit(1);
 }
@@ -501,18 +518,21 @@ main(
 	union mdrestore_headers	headers;
 	DEFINE_MDRESTORE_DEV(ddev);
 	DEFINE_MDRESTORE_DEV(logdev);
+	DEFINE_MDRESTORE_DEV(rtdev);
 	FILE			*src_f;
 	char			*logdev_path = NULL;
+	char			*rtdev_path = NULL;
 	int			c;
 
 	mdrestore.show_progress = false;
 	mdrestore.show_info = false;
 	mdrestore.progress_since_warning = false;
 	mdrestore.external_log = false;
+	mdrestore.realtime_data = false;
 
 	progname = basename(argv[0]);
 
-	while ((c = getopt(argc, argv, "gil:V")) != EOF) {
+	while ((c = getopt(argc, argv, "gil:r:V")) != EOF) {
 		switch (c) {
 			case 'g':
 				mdrestore.show_progress = true;
@@ -524,6 +544,10 @@ main(
 				logdev_path = optarg;
 				mdrestore.external_log = true;
 				break;
+			case 'r':
+				rtdev_path = optarg;
+				mdrestore.realtime_data = true;
+				break;
 			case 'V':
 				printf("%s version %s\n", progname, VERSION);
 				exit(0);
@@ -592,10 +616,15 @@ main(
 	if (mdrestore.external_log)
 		open_device(&logdev, logdev_path);
 
-	mdrestore.mdrops->restore(&headers, src_f, &ddev, &logdev);
+	/* check and open realtime device */
+	if (mdrestore.realtime_data)
+		open_device(&rtdev, rtdev_path);
+
+	mdrestore.mdrops->restore(&headers, src_f, &ddev, &logdev, &rtdev);
 
 	close_device(&ddev);
 	close_device(&logdev);
+	close_device(&rtdev);
 
 	if (src_f != stdin)
 		fclose(src_f);

Re: [PATCH 3/3] xfs_mdrestore: restore rt group superblocks to realtime device

[Christoph Hellwig]

Looks good:

Reviewed-by: Christoph Hellwig <hch@lst.de>

[PATCHSET v6.2 3/5] xfs: enable in-core block reservation for rt metadata

[Darrick J. Wong]

Hi all,

In preparation for adding reverse mapping and refcounting to the
realtime device, enhance the metadir code to reserve free space for
btree shape changes as delayed allocation blocks.

This enables us to pre-allocate space for the rmap and refcount btrees
in the same manner as we do for the data device counterparts, which is
how we avoid ENOSPC failures when space is low but we've already
committed to a COW operation.

If you're going to start using this code, I strongly recommend pulling
from my git trees, which are linked below.

This has been running on the djcloud for months with no problems.  Enjoy!
Comments and questions are, as always, welcome.

--D

kernel git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfs-linux.git/log/?h=reserve-rt-metadata-space

xfsprogs git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfsprogs-dev.git/log/?h=reserve-rt-metadata-space
---
Commits in this patchset:
 * xfs: prepare to reuse the dquot pointer space in struct xfs_inode
 * xfs: allow inode-based btrees to reserve space in the data device
---
 fs/xfs/libxfs/xfs_ag_resv.c  |    3 +
 fs/xfs/libxfs/xfs_attr.c     |    4 -
 fs/xfs/libxfs/xfs_bmap.c     |    4 -
 fs/xfs/libxfs/xfs_errortag.h |    4 +
 fs/xfs/libxfs/xfs_metadir.c  |    4 +
 fs/xfs/libxfs/xfs_metafile.c |  205 ++++++++++++++++++++++++++++++++++++++++++
 fs/xfs/libxfs/xfs_metafile.h |   11 ++
 fs/xfs/libxfs/xfs_types.h    |    7 +
 fs/xfs/scrub/tempfile.c      |    1 
 fs/xfs/xfs_dquot.h           |    3 +
 fs/xfs/xfs_error.c           |    3 +
 fs/xfs/xfs_exchrange.c       |    3 +
 fs/xfs/xfs_fsops.c           |   17 +++
 fs/xfs/xfs_inode.h           |   16 +++
 fs/xfs/xfs_mount.c           |   10 ++
 fs/xfs/xfs_mount.h           |    1 
 fs/xfs/xfs_qm.c              |    2 
 fs/xfs/xfs_quota.h           |    5 -
 fs/xfs/xfs_rtalloc.c         |   21 ++++
 fs/xfs/xfs_rtalloc.h         |    5 +
 fs/xfs/xfs_trace.h           |   45 +++++++++
 fs/xfs/xfs_trans.c           |    4 +
 fs/xfs/xfs_trans_dquot.c     |    8 +-
 23 files changed, 367 insertions(+), 19 deletions(-)

[PATCH 1/2] xfs: prepare to reuse the dquot pointer space in struct xfs_inode

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Files participating in the metadata directory tree are not accounted to
the quota subsystem.  Therefore, the i_[ugp]dquot pointers in struct
xfs_inode are never used and should always be NULL.

In the next patch we want to add a u64 count of fs blocks reserved for
metadata btree expansion, but we don't want every inode in the fs to pay
the memory price for this feature.  The intent is to union those three
pointers with the u64 counter, but for that to work we must guard
against all access to the dquot pointers for metadata files.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_attr.c |    4 +---
 fs/xfs/libxfs/xfs_bmap.c |    4 +---
 fs/xfs/scrub/tempfile.c  |    1 +
 fs/xfs/xfs_dquot.h       |    3 +++
 fs/xfs/xfs_exchrange.c   |    3 +++
 fs/xfs/xfs_inode.h       |   10 +++++++---
 fs/xfs/xfs_qm.c          |    2 ++
 fs/xfs/xfs_quota.h       |    5 -----
 fs/xfs/xfs_trans.c       |    4 ++++
 fs/xfs/xfs_trans_dquot.c |    8 ++++----
 10 files changed, 26 insertions(+), 18 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_attr.c b/fs/xfs/libxfs/xfs_attr.c
index 17875ad865f5d6..8c04acd30d489c 100644
--- a/fs/xfs/libxfs/xfs_attr.c
+++ b/fs/xfs/libxfs/xfs_attr.c
@@ -1004,9 +1004,7 @@ xfs_attr_add_fork(
 	unsigned int		blks;		/* space reservation */
 	int			error;		/* error return value */
 
-	if (xfs_is_metadir_inode(ip))
-		ASSERT(XFS_IS_DQDETACHED(ip));
-	else
+	if (!xfs_is_metadir_inode(ip))
 		ASSERT(!XFS_NOT_DQATTACHED(mp, ip));
 
 	blks = XFS_ADDAFORK_SPACE_RES(mp);
diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c
index 0842577755f7bb..02323936cc9b20 100644
--- a/fs/xfs/libxfs/xfs_bmap.c
+++ b/fs/xfs/libxfs/xfs_bmap.c
@@ -1042,9 +1042,7 @@ xfs_bmap_add_attrfork(
 	int			error;		/* error return value */
 
 	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
-	if (xfs_is_metadir_inode(ip))
-		ASSERT(XFS_IS_DQDETACHED(ip));
-	else
+	if (!xfs_is_metadir_inode(ip))
 		ASSERT(!XFS_NOT_DQATTACHED(mp, ip));
 	ASSERT(!xfs_inode_has_attr_fork(ip));
 
diff --git a/fs/xfs/scrub/tempfile.c b/fs/xfs/scrub/tempfile.c
index 2d7ca7e1bbca0f..4ebb5f8459e8f3 100644
--- a/fs/xfs/scrub/tempfile.c
+++ b/fs/xfs/scrub/tempfile.c
@@ -749,6 +749,7 @@ xrep_tempexch_reserve_quota(
 	 * or the two inodes have the same dquots.
 	 */
 	if (!XFS_IS_QUOTA_ON(tp->t_mountp) || req->ip1 == req->ip2 ||
+	    xfs_is_metadir_inode(req->ip1) ||
 	    (req->ip1->i_udquot == req->ip2->i_udquot &&
 	     req->ip1->i_gdquot == req->ip2->i_gdquot &&
 	     req->ip1->i_pdquot == req->ip2->i_pdquot))
diff --git a/fs/xfs/xfs_dquot.h b/fs/xfs/xfs_dquot.h
index c617bac75361b2..61217adf5ba551 100644
--- a/fs/xfs/xfs_dquot.h
+++ b/fs/xfs/xfs_dquot.h
@@ -160,6 +160,9 @@ static inline struct xfs_dquot *xfs_inode_dquot(
 	struct xfs_inode	*ip,
 	xfs_dqtype_t		type)
 {
+	if (xfs_is_metadir_inode(ip))
+		return NULL;
+
 	switch (type) {
 	case XFS_DQTYPE_USER:
 		return ip->i_udquot;
diff --git a/fs/xfs/xfs_exchrange.c b/fs/xfs/xfs_exchrange.c
index 265c424498933e..f340a2015c4c71 100644
--- a/fs/xfs/xfs_exchrange.c
+++ b/fs/xfs/xfs_exchrange.c
@@ -119,6 +119,9 @@ xfs_exchrange_reserve_quota(
 	int				ip1_error = 0;
 	int				error;
 
+	ASSERT(!xfs_is_metadir_inode(req->ip1));
+	ASSERT(!xfs_is_metadir_inode(req->ip2));
+
 	/*
 	 * Don't bother with a quota reservation if we're not enforcing them
 	 * or the two inodes have the same dquots.
diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h
index 1648dc5a806882..1141c2e8e123ae 100644
--- a/fs/xfs/xfs_inode.h
+++ b/fs/xfs/xfs_inode.h
@@ -25,9 +25,13 @@ struct xfs_dquot;
 typedef struct xfs_inode {
 	/* Inode linking and identification information. */
 	struct xfs_mount	*i_mount;	/* fs mount struct ptr */
-	struct xfs_dquot	*i_udquot;	/* user dquot */
-	struct xfs_dquot	*i_gdquot;	/* group dquot */
-	struct xfs_dquot	*i_pdquot;	/* project dquot */
+	union {
+		struct {
+			struct xfs_dquot *i_udquot;	/* user dquot */
+			struct xfs_dquot *i_gdquot;	/* group dquot */
+			struct xfs_dquot *i_pdquot;	/* project dquot */
+		};
+	};
 
 	/* Inode location stuff */
 	xfs_ino_t		i_ino;		/* inode number (agno/agino)*/
diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c
index dc8b1010d4d332..3abab5fb593e37 100644
--- a/fs/xfs/xfs_qm.c
+++ b/fs/xfs/xfs_qm.c
@@ -428,6 +428,8 @@ void
 xfs_qm_dqdetach(
 	xfs_inode_t	*ip)
 {
+	if (xfs_is_metadir_inode(ip))
+		return;
 	if (!(ip->i_udquot || ip->i_gdquot || ip->i_pdquot))
 		return;
 
diff --git a/fs/xfs/xfs_quota.h b/fs/xfs/xfs_quota.h
index d7565462af3dc4..105e6eb5762011 100644
--- a/fs/xfs/xfs_quota.h
+++ b/fs/xfs/xfs_quota.h
@@ -29,11 +29,6 @@ struct xfs_buf;
 	 (XFS_IS_GQUOTA_ON(mp) && (ip)->i_gdquot == NULL) || \
 	 (XFS_IS_PQUOTA_ON(mp) && (ip)->i_pdquot == NULL))
 
-#define XFS_IS_DQDETACHED(ip) \
-	((ip)->i_udquot == NULL && \
-	 (ip)->i_gdquot == NULL && \
-	 (ip)->i_pdquot == NULL)
-
 #define XFS_QM_NEED_QUOTACHECK(mp) \
 	((XFS_IS_UQUOTA_ON(mp) && \
 		(mp->m_sb.sb_qflags & XFS_UQUOTA_CHKD) == 0) || \
diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c
index 4cd25717c9d130..f53f82456288e5 100644
--- a/fs/xfs/xfs_trans.c
+++ b/fs/xfs/xfs_trans.c
@@ -1266,6 +1266,9 @@ xfs_trans_alloc_ichange(
 	xfs_ilock(ip, XFS_ILOCK_EXCL);
 	xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
 
+	if (xfs_is_metadir_inode(ip))
+		goto out;
+
 	error = xfs_qm_dqattach_locked(ip, false);
 	if (error) {
 		/* Caller should have allocated the dquots! */
@@ -1334,6 +1337,7 @@ xfs_trans_alloc_ichange(
 			goto out_cancel;
 	}
 
+out:
 	*tpp = tp;
 	return 0;
 
diff --git a/fs/xfs/xfs_trans_dquot.c b/fs/xfs/xfs_trans_dquot.c
index 713b6d243e5631..765456bf342851 100644
--- a/fs/xfs/xfs_trans_dquot.c
+++ b/fs/xfs/xfs_trans_dquot.c
@@ -156,7 +156,8 @@ xfs_trans_mod_ino_dquot(
 	unsigned int			field,
 	int64_t				delta)
 {
-	ASSERT(!xfs_is_metadir_inode(ip) || XFS_IS_DQDETACHED(ip));
+	if (xfs_is_metadir_inode(ip))
+		return;
 
 	xfs_trans_mod_dquot(tp, dqp, field, delta);
 
@@ -246,11 +247,10 @@ xfs_trans_mod_dquot_byino(
 	xfs_mount_t	*mp = tp->t_mountp;
 
 	if (!XFS_IS_QUOTA_ON(mp) ||
-	    xfs_is_quota_inode(&mp->m_sb, ip->i_ino))
+	    xfs_is_quota_inode(&mp->m_sb, ip->i_ino) ||
+	    xfs_is_metadir_inode(ip))
 		return;
 
-	ASSERT(!xfs_is_metadir_inode(ip) || XFS_IS_DQDETACHED(ip));
-
 	if (XFS_IS_UQUOTA_ON(mp) && ip->i_udquot)
 		xfs_trans_mod_ino_dquot(tp, ip, ip->i_udquot, field, delta);
 	if (XFS_IS_GQUOTA_ON(mp) && ip->i_gdquot)

[PATCHSET v6.0 2/5] xfs: enable in-core block reservation for rt metadata

[Darrick J. Wong]

Hi all,

In preparation for adding reverse mapping and refcounting to the
realtime device, enhance the metadir code to reserve free space for
btree shape changes as delayed allocation blocks.

This enables us to pre-allocate space for the rmap and refcount btrees
in the same manner as we do for the data device counterparts, which is
how we avoid ENOSPC failures when space is low but we've already
committed to a COW operation.

If you're going to start using this code, I strongly recommend pulling
from my git trees, which are linked below.

This has been running on the djcloud for months with no problems.  Enjoy!
Comments and questions are, as always, welcome.

--D

kernel git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfs-linux.git/log/?h=reserve-rt-metadata-space

xfsprogs git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfsprogs-dev.git/log/?h=reserve-rt-metadata-space
---
Commits in this patchset:
 * xfs: prepare to reuse the dquot pointer space in struct xfs_inode
 * xfs: allow inode-based btrees to reserve space in the data device
---
 fs/xfs/libxfs/xfs_ag_resv.c  |    3 +
 fs/xfs/libxfs/xfs_attr.c     |    4 -
 fs/xfs/libxfs/xfs_bmap.c     |    4 -
 fs/xfs/libxfs/xfs_errortag.h |    4 +
 fs/xfs/libxfs/xfs_metadir.c  |    4 +
 fs/xfs/libxfs/xfs_metafile.c |  205 ++++++++++++++++++++++++++++++++++++++++++
 fs/xfs/libxfs/xfs_metafile.h |   11 ++
 fs/xfs/libxfs/xfs_types.h    |    7 +
 fs/xfs/scrub/tempfile.c      |    1 
 fs/xfs/xfs_dquot.h           |    3 +
 fs/xfs/xfs_error.c           |    3 +
 fs/xfs/xfs_exchrange.c       |    3 +
 fs/xfs/xfs_fsops.c           |   17 +++
 fs/xfs/xfs_inode.h           |   16 +++
 fs/xfs/xfs_mount.c           |   10 ++
 fs/xfs/xfs_mount.h           |    1 
 fs/xfs/xfs_qm.c              |    2 
 fs/xfs/xfs_quota.h           |    5 -
 fs/xfs/xfs_rtalloc.c         |   21 ++++
 fs/xfs/xfs_rtalloc.h         |    5 +
 fs/xfs/xfs_trace.h           |   45 +++++++++
 fs/xfs/xfs_trans.c           |    4 +
 fs/xfs/xfs_trans_dquot.c     |    8 +-
 23 files changed, 367 insertions(+), 19 deletions(-)

[PATCH 1/2] xfs: prepare to reuse the dquot pointer space in struct xfs_inode

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

Files participating in the metadata directory tree are not accounted to
the quota subsystem.  Therefore, the i_[ugp]dquot pointers in struct
xfs_inode are never used and should always be NULL.

In the next patch we want to add a u64 count of fs blocks reserved for
metadata btree expansion, but we don't want every inode in the fs to pay
the memory price for this feature.  The intent is to union those three
pointers with the u64 counter, but for that to work we must guard
against all access to the dquot pointers for metadata files.

Signed-off-by: "Darrick J. Wong" <djwong@kernel.org>
---
 fs/xfs/libxfs/xfs_attr.c |    4 +---
 fs/xfs/libxfs/xfs_bmap.c |    4 +---
 fs/xfs/scrub/tempfile.c  |    1 +
 fs/xfs/xfs_dquot.h       |    3 +++
 fs/xfs/xfs_exchrange.c   |    3 +++
 fs/xfs/xfs_inode.h       |   10 +++++++---
 fs/xfs/xfs_qm.c          |    2 ++
 fs/xfs/xfs_quota.h       |    5 -----
 fs/xfs/xfs_trans.c       |    4 ++++
 fs/xfs/xfs_trans_dquot.c |    8 ++++----
 10 files changed, 26 insertions(+), 18 deletions(-)


diff --git a/fs/xfs/libxfs/xfs_attr.c b/fs/xfs/libxfs/xfs_attr.c
index 17875ad865f5d6..8c04acd30d489c 100644
--- a/fs/xfs/libxfs/xfs_attr.c
+++ b/fs/xfs/libxfs/xfs_attr.c
@@ -1004,9 +1004,7 @@ xfs_attr_add_fork(
 	unsigned int		blks;		/* space reservation */
 	int			error;		/* error return value */
 
-	if (xfs_is_metadir_inode(ip))
-		ASSERT(XFS_IS_DQDETACHED(ip));
-	else
+	if (!xfs_is_metadir_inode(ip))
 		ASSERT(!XFS_NOT_DQATTACHED(mp, ip));
 
 	blks = XFS_ADDAFORK_SPACE_RES(mp);
diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c
index 0842577755f7bb..02323936cc9b20 100644
--- a/fs/xfs/libxfs/xfs_bmap.c
+++ b/fs/xfs/libxfs/xfs_bmap.c
@@ -1042,9 +1042,7 @@ xfs_bmap_add_attrfork(
 	int			error;		/* error return value */
 
 	xfs_assert_ilocked(ip, XFS_ILOCK_EXCL);
-	if (xfs_is_metadir_inode(ip))
-		ASSERT(XFS_IS_DQDETACHED(ip));
-	else
+	if (!xfs_is_metadir_inode(ip))
 		ASSERT(!XFS_NOT_DQATTACHED(mp, ip));
 	ASSERT(!xfs_inode_has_attr_fork(ip));
 
diff --git a/fs/xfs/scrub/tempfile.c b/fs/xfs/scrub/tempfile.c
index 2d7ca7e1bbca0f..4ebb5f8459e8f3 100644
--- a/fs/xfs/scrub/tempfile.c
+++ b/fs/xfs/scrub/tempfile.c
@@ -749,6 +749,7 @@ xrep_tempexch_reserve_quota(
 	 * or the two inodes have the same dquots.
 	 */
 	if (!XFS_IS_QUOTA_ON(tp->t_mountp) || req->ip1 == req->ip2 ||
+	    xfs_is_metadir_inode(req->ip1) ||
 	    (req->ip1->i_udquot == req->ip2->i_udquot &&
 	     req->ip1->i_gdquot == req->ip2->i_gdquot &&
 	     req->ip1->i_pdquot == req->ip2->i_pdquot))
diff --git a/fs/xfs/xfs_dquot.h b/fs/xfs/xfs_dquot.h
index c617bac75361b2..61217adf5ba551 100644
--- a/fs/xfs/xfs_dquot.h
+++ b/fs/xfs/xfs_dquot.h
@@ -160,6 +160,9 @@ static inline struct xfs_dquot *xfs_inode_dquot(
 	struct xfs_inode	*ip,
 	xfs_dqtype_t		type)
 {
+	if (xfs_is_metadir_inode(ip))
+		return NULL;
+
 	switch (type) {
 	case XFS_DQTYPE_USER:
 		return ip->i_udquot;
diff --git a/fs/xfs/xfs_exchrange.c b/fs/xfs/xfs_exchrange.c
index dd24de420714ab..d1c1d5819fbfe0 100644
--- a/fs/xfs/xfs_exchrange.c
+++ b/fs/xfs/xfs_exchrange.c
@@ -119,6 +119,9 @@ xfs_exchrange_reserve_quota(
 	int				ip1_error = 0;
 	int				error;
 
+	ASSERT(!xfs_is_metadir_inode(req->ip1));
+	ASSERT(!xfs_is_metadir_inode(req->ip2));
+
 	/*
 	 * Don't bother with a quota reservation if we're not enforcing them
 	 * or the two inodes have the same dquots.
diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h
index 1648dc5a806882..1141c2e8e123ae 100644
--- a/fs/xfs/xfs_inode.h
+++ b/fs/xfs/xfs_inode.h
@@ -25,9 +25,13 @@ struct xfs_dquot;
 typedef struct xfs_inode {
 	/* Inode linking and identification information. */
 	struct xfs_mount	*i_mount;	/* fs mount struct ptr */
-	struct xfs_dquot	*i_udquot;	/* user dquot */
-	struct xfs_dquot	*i_gdquot;	/* group dquot */
-	struct xfs_dquot	*i_pdquot;	/* project dquot */
+	union {
+		struct {
+			struct xfs_dquot *i_udquot;	/* user dquot */
+			struct xfs_dquot *i_gdquot;	/* group dquot */
+			struct xfs_dquot *i_pdquot;	/* project dquot */
+		};
+	};
 
 	/* Inode location stuff */
 	xfs_ino_t		i_ino;		/* inode number (agno/agino)*/
diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c
index dc8b1010d4d332..3abab5fb593e37 100644
--- a/fs/xfs/xfs_qm.c
+++ b/fs/xfs/xfs_qm.c
@@ -428,6 +428,8 @@ void
 xfs_qm_dqdetach(
 	xfs_inode_t	*ip)
 {
+	if (xfs_is_metadir_inode(ip))
+		return;
 	if (!(ip->i_udquot || ip->i_gdquot || ip->i_pdquot))
 		return;
 
diff --git a/fs/xfs/xfs_quota.h b/fs/xfs/xfs_quota.h
index d7565462af3dc4..105e6eb5762011 100644
--- a/fs/xfs/xfs_quota.h
+++ b/fs/xfs/xfs_quota.h
@@ -29,11 +29,6 @@ struct xfs_buf;
 	 (XFS_IS_GQUOTA_ON(mp) && (ip)->i_gdquot == NULL) || \
 	 (XFS_IS_PQUOTA_ON(mp) && (ip)->i_pdquot == NULL))
 
-#define XFS_IS_DQDETACHED(ip) \
-	((ip)->i_udquot == NULL && \
-	 (ip)->i_gdquot == NULL && \
-	 (ip)->i_pdquot == NULL)
-
 #define XFS_QM_NEED_QUOTACHECK(mp) \
 	((XFS_IS_UQUOTA_ON(mp) && \
 		(mp->m_sb.sb_qflags & XFS_UQUOTA_CHKD) == 0) || \
diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c
index 4cd25717c9d130..f53f82456288e5 100644
--- a/fs/xfs/xfs_trans.c
+++ b/fs/xfs/xfs_trans.c
@@ -1266,6 +1266,9 @@ xfs_trans_alloc_ichange(
 	xfs_ilock(ip, XFS_ILOCK_EXCL);
 	xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
 
+	if (xfs_is_metadir_inode(ip))
+		goto out;
+
 	error = xfs_qm_dqattach_locked(ip, false);
 	if (error) {
 		/* Caller should have allocated the dquots! */
@@ -1334,6 +1337,7 @@ xfs_trans_alloc_ichange(
 			goto out_cancel;
 	}
 
+out:
 	*tpp = tp;
 	return 0;
 
diff --git a/fs/xfs/xfs_trans_dquot.c b/fs/xfs/xfs_trans_dquot.c
index 713b6d243e5631..765456bf342851 100644
--- a/fs/xfs/xfs_trans_dquot.c
+++ b/fs/xfs/xfs_trans_dquot.c
@@ -156,7 +156,8 @@ xfs_trans_mod_ino_dquot(
 	unsigned int			field,
 	int64_t				delta)
 {
-	ASSERT(!xfs_is_metadir_inode(ip) || XFS_IS_DQDETACHED(ip));
+	if (xfs_is_metadir_inode(ip))
+		return;
 
 	xfs_trans_mod_dquot(tp, dqp, field, delta);
 
@@ -246,11 +247,10 @@ xfs_trans_mod_dquot_byino(
 	xfs_mount_t	*mp = tp->t_mountp;
 
 	if (!XFS_IS_QUOTA_ON(mp) ||
-	    xfs_is_quota_inode(&mp->m_sb, ip->i_ino))
+	    xfs_is_quota_inode(&mp->m_sb, ip->i_ino) ||
+	    xfs_is_metadir_inode(ip))
 		return;
 
-	ASSERT(!xfs_is_metadir_inode(ip) || XFS_IS_DQDETACHED(ip));
-
 	if (XFS_IS_UQUOTA_ON(mp) && ip->i_udquot)
 		xfs_trans_mod_ino_dquot(tp, ip, ip->i_udquot, field, delta);
 	if (XFS_IS_GQUOTA_ON(mp) && ip->i_gdquot)