Merge tag 'xfs-4.18-merge-3' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux

Pull xfs updates from Darrick Wong:
 "New features this cycle include the ability to relabel mounted
  filesystems, support for fallocated swapfiles, and using FUA for pure
  data O_DSYNC directio writes. With this cycle we begin to integrate
  online filesystem repair and refactor the growfs code in preparation
  for eventual subvolume support, though the road ahead for both
  features is quite long.

  There are also numerous refactorings of the iomap code to remove
  unnecessary log overhead, to disentangle some of the quota code, and
  to prepare for buffer head removal in a future upstream kernel.

  Metadata validation continues to improve, both in the hot path
  veifiers and the online filesystem check code. I anticipate sending a
  second pull request in a few days with more metadata validation
  improvements.

  This series has been run through a full xfstests run over the weekend
  and through a quick xfstests run against this morning's master, with
  no major failures reported.

  Summary:

   - Strengthen inode number and structure validation when allocating
     inodes.

   - Reduce pointless buffer allocations during cache miss

   - Use FUA for pure data O_DSYNC directio writes

   - Various iomap refactorings

   - Strengthen quota metadata verification to avoid unfixable broken
     quota

   - Make AGFL block freeing a deferred operation to avoid blowing out
     transaction reservations when running complex operations

   - Get rid of the log item descriptors to reduce log overhead

   - Fix various reflink bugs where inodes were double-joined to
     transactions

   - Don't issue discards when trimming unwritten extents

   - Refactor incore dquot initialization and retrieval interfaces

   - Fix some locking problmes in the quota scrub code

   - Strengthen btree structure checks in scrub code

   - Rewrite swapfile activation to use iomap and support unwritten
     extents

   - Make scrub exit to userspace sooner when corruptions or
     cross-referencing problems are found

   - Make scrub invoke the data fork scrubber directly on metadata
     inodes

   - Don't do background reclamation of post-eof and cow blocks when the
     fs is suspended

   - Fix secondary superblock buffer lifespan hinting

   - Refactor growfs to use table-dispatched functions instead of long
     stringy functions

   - Move growfs code to libxfs

   - Implement online fs label getting and setting

   - Introduce online filesystem repair (in a very limited capacity)

   - Fix unit conversion problems in the realtime freemap iteration
     functions

   - Various refactorings and cleanups in preparation to remove buffer
     heads in a future release

   - Reimplement the old bmap call with iomap

   - Remove direct buffer head accesses from seek hole/data

   - Various bug fixes"

* tag 'xfs-4.18-merge-3' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux: (121 commits)
  fs: use ->is_partially_uptodate in page_cache_seek_hole_data
  fs: remove the buffer_unwritten check in page_seek_hole_data
  fs: move page_cache_seek_hole_data to iomap.c
  xfs: use iomap_bmap
  iomap: add an iomap-based bmap implementation
  iomap: add a iomap_sector helper
  iomap: use __bio_add_page in iomap_dio_zero
  iomap: move IOMAP_F_BOUNDARY to gfs2
  iomap: fix the comment describing IOMAP_NOWAIT
  iomap: inline data should be an iomap type, not a flag
  mm: split ->readpages calls to avoid non-contiguous pages lists
  mm: return an unsigned int from __do_page_cache_readahead
  mm: give the 'ret' variable a better name __do_page_cache_readahead
  block: add a lower-level bio_add_page interface
  xfs: fix error handling in xfs_refcount_insert()
  xfs: fix xfs_rtalloc_rec units
  xfs: strengthen rtalloc query range checks
  xfs: xfs_rtbuf_get should check the bmapi_read results
  xfs: xfs_rtword_t should be unsigned, not signed
  dax: change bdev_dax_supported() to support boolean returns
  ...

127 files changed, 5543 insertions, 1980 deletions

diff --git a/Documentation/ioctl/ioctl-number.txt b/Documentation/ioctl/ioctl-number.txt
index 7f7413e597f3..27c1b7b78504 100644
--- a/Documentation/ioctl/ioctl-number.txt
+++ b/Documentation/ioctl/ioctl-number.txt
@@ -296,7 +296,8 @@ Code  Seq#(hex)	Include File		Comments
 0x90	00	drivers/cdrom/sbpcd.h
 0x92	00-0F	drivers/usb/mon/mon_bin.c
 0x93	60-7F	linux/auto_fs.h
-0x94	all	fs/btrfs/ioctl.h
+0x94	all	fs/btrfs/ioctl.h	Btrfs filesystem
+		and linux/fs.h		some lifted to vfs/generic
 0x97	00-7F	fs/ceph/ioctl.h		Ceph file system
 0x99	00-0F				537-Addinboard driver
 					<mailto:buk@buks.ipn.de>
diff --git a/block/bio.c b/block/bio.c
index 595663e0281a..5f7563598b1c 100644
--- a/block/bio.c
+++ b/block/bio.c
@@ -774,7 +774,7 @@ int bio_add_pc_page(struct request_queue *q, struct bio *bio, struct page
 			return 0;
 	}
 
-	if (bio->bi_vcnt >= bio->bi_max_vecs)
+	if (bio_full(bio))
 		return 0;
 
 	/*
@@ -822,52 +822,82 @@ int bio_add_pc_page(struct request_queue *q, struct bio *bio, struct page
 EXPORT_SYMBOL(bio_add_pc_page);
 
 /**
- *	bio_add_page	-	attempt to add page to bio
- *	@bio: destination bio
- *	@page: page to add
- *	@len: vec entry length
- *	@offset: vec entry offset
+ * __bio_try_merge_page - try appending data to an existing bvec.
+ * @bio: destination bio
+ * @page: page to add
+ * @len: length of the data to add
+ * @off: offset of the data in @page
  *
- *	Attempt to add a page to the bio_vec maplist. This will only fail
- *	if either bio->bi_vcnt == bio->bi_max_vecs or it's a cloned bio.
+ * Try to add the data at @page + @off to the last bvec of @bio.  This is a
+ * a useful optimisation for file systems with a block size smaller than the
+ * page size.
+ *
+ * Return %true on success or %false on failure.
  */
-int bio_add_page(struct bio *bio, struct page *page,
-		 unsigned int len, unsigned int offset)
+bool __bio_try_merge_page(struct bio *bio, struct page *page,
+		unsigned int len, unsigned int off)
 {
-	struct bio_vec *bv;
-
-	/*
-	 * cloned bio must not modify vec list
-	 */
 	if (WARN_ON_ONCE(bio_flagged(bio, BIO_CLONED)))
-		return 0;
+		return false;
 
-	/*
-	 * For filesystems with a blocksize smaller than the pagesize
-	 * we will often be called with the same page as last time and
-	 * a consecutive offset.  Optimize this special case.
-	 */
 	if (bio->bi_vcnt > 0) {
-		bv = &bio->bi_io_vec[bio->bi_vcnt - 1];
+		struct bio_vec *bv = &bio->bi_io_vec[bio->bi_vcnt - 1];
 
-		if (page == bv->bv_page &&
-		    offset == bv->bv_offset + bv->bv_len) {
+		if (page == bv->bv_page && off == bv->bv_offset + bv->bv_len) {
 			bv->bv_len += len;
-			goto done;
+			bio->bi_iter.bi_size += len;
+			return true;
 		}
 	}
+	return false;
+}
+EXPORT_SYMBOL_GPL(__bio_try_merge_page);
 
-	if (bio->bi_vcnt >= bio->bi_max_vecs)
-		return 0;
+/**
+ * __bio_add_page - add page to a bio in a new segment
+ * @bio: destination bio
+ * @page: page to add
+ * @len: length of the data to add
+ * @off: offset of the data in @page
+ *
+ * Add the data at @page + @off to @bio as a new bvec.  The caller must ensure
+ * that @bio has space for another bvec.
+ */
+void __bio_add_page(struct bio *bio, struct page *page,
+		unsigned int len, unsigned int off)
+{
+	struct bio_vec *bv = &bio->bi_io_vec[bio->bi_vcnt];
 
-	bv		= &bio->bi_io_vec[bio->bi_vcnt];
-	bv->bv_page	= page;
-	bv->bv_len	= len;
-	bv->bv_offset	= offset;
+	WARN_ON_ONCE(bio_flagged(bio, BIO_CLONED));
+	WARN_ON_ONCE(bio_full(bio));
+
+	bv->bv_page = page;
+	bv->bv_offset = off;
+	bv->bv_len = len;
 
-	bio->bi_vcnt++;
-done:
 	bio->bi_iter.bi_size += len;
+	bio->bi_vcnt++;
+}
+EXPORT_SYMBOL_GPL(__bio_add_page);
+
+/**
+ *	bio_add_page	-	attempt to add page to bio
+ *	@bio: destination bio
+ *	@page: page to add
+ *	@len: vec entry length
+ *	@offset: vec entry offset
+ *
+ *	Attempt to add a page to the bio_vec maplist. This will only fail
+ *	if either bio->bi_vcnt == bio->bi_max_vecs or it's a cloned bio.
+ */
+int bio_add_page(struct bio *bio, struct page *page,
+		 unsigned int len, unsigned int offset)
+{
+	if (!__bio_try_merge_page(bio, page, len, offset)) {
+		if (bio_full(bio))
+			return 0;
+		__bio_add_page(bio, page, len, offset);
+	}
 	return len;
 }
 EXPORT_SYMBOL(bio_add_page);
diff --git a/drivers/dax/super.c b/drivers/dax/super.c
index 2b2332b605e4..1d7bd96511f0 100644
--- a/drivers/dax/super.c
+++ b/drivers/dax/super.c
@@ -74,42 +74,42 @@ EXPORT_SYMBOL_GPL(fs_dax_get_by_bdev);
 
 /**
  * __bdev_dax_supported() - Check if the device supports dax for filesystem
- * @sb: The superblock of the device
+ * @bdev: block device to check
  * @blocksize: The block size of the device
  *
  * This is a library function for filesystems to check if the block device
  * can be mounted with dax option.
  *
- * Return: negative errno if unsupported, 0 if supported.
+ * Return: true if supported, false if unsupported
  */
-int __bdev_dax_supported(struct super_block *sb, int blocksize)
+bool __bdev_dax_supported(struct block_device *bdev, int blocksize)
 {
-	struct block_device *bdev = sb->s_bdev;
 	struct dax_device *dax_dev;
 	pgoff_t pgoff;
 	int err, id;
 	void *kaddr;
 	pfn_t pfn;
 	long len;
+	char buf[BDEVNAME_SIZE];
 
 	if (blocksize != PAGE_SIZE) {
-		pr_debug("VFS (%s): error: unsupported blocksize for dax\n",
-				sb->s_id);
-		return -EINVAL;
+		pr_debug("%s: error: unsupported blocksize for dax\n",
+				bdevname(bdev, buf));
+		return false;
 	}
 
 	err = bdev_dax_pgoff(bdev, 0, PAGE_SIZE, &pgoff);
 	if (err) {
-		pr_debug("VFS (%s): error: unaligned partition for dax\n",
-				sb->s_id);
-		return err;
+		pr_debug("%s: error: unaligned partition for dax\n",
+				bdevname(bdev, buf));
+		return false;
 	}
 
 	dax_dev = dax_get_by_host(bdev->bd_disk->disk_name);
 	if (!dax_dev) {
-		pr_debug("VFS (%s): error: device does not support dax\n",
-				sb->s_id);
-		return -EOPNOTSUPP;
+		pr_debug("%s: error: device does not support dax\n",
+				bdevname(bdev, buf));
+		return false;
 	}
 
 	id = dax_read_lock();
@@ -119,9 +119,9 @@ int __bdev_dax_supported(struct super_block *sb, int blocksize)
 	put_dax(dax_dev);
 
 	if (len < 1) {
-		pr_debug("VFS (%s): error: dax access failed (%ld)\n",
-				sb->s_id, len);
-		return len < 0 ? len : -EIO;
+		pr_debug("%s: error: dax access failed (%ld)\n",
+				bdevname(bdev, buf), len);
+		return false;
 	}
 
 	if (IS_ENABLED(CONFIG_FS_DAX_LIMITED) && pfn_t_special(pfn)) {
@@ -137,12 +137,12 @@ int __bdev_dax_supported(struct super_block *sb, int blocksize)
 	} else if (pfn_t_devmap(pfn)) {
 		/* pass */;
 	} else {
-		pr_debug("VFS (%s): error: dax support not enabled\n",
-				sb->s_id);
-		return -EOPNOTSUPP;
+		pr_debug("%s: error: dax support not enabled\n",
+				bdevname(bdev, buf));
+		return false;
 	}
 
-	return 0;
+	return true;
 }
 EXPORT_SYMBOL_GPL(__bdev_dax_supported);
 #endif
diff --git a/fs/buffer.c b/fs/buffer.c
index 249b83fafe48..cabc045f483d 100644
--- a/fs/buffer.c
+++ b/fs/buffer.c
@@ -3427,120 +3427,6 @@ int bh_submit_read(struct buffer_head *bh)
 }
 EXPORT_SYMBOL(bh_submit_read);
 
-/*
- * Seek for SEEK_DATA / SEEK_HOLE within @page, starting at @lastoff.
- *
- * Returns the offset within the file on success, and -ENOENT otherwise.
- */
-static loff_t
-page_seek_hole_data(struct page *page, loff_t lastoff, int whence)
-{
-	loff_t offset = page_offset(page);
-	struct buffer_head *bh, *head;
-	bool seek_data = whence == SEEK_DATA;
-
-	if (lastoff < offset)
-		lastoff = offset;
-
-	bh = head = page_buffers(page);
-	do {
-		offset += bh->b_size;
-		if (lastoff >= offset)
-			continue;
-
-		/*
-		 * Unwritten extents that have data in the page cache covering
-		 * them can be identified by the BH_Unwritten state flag.
-		 * Pages with multiple buffers might have a mix of holes, data
-		 * and unwritten extents - any buffer with valid data in it
-		 * should have BH_Uptodate flag set on it.
-		 */
-
-		if ((buffer_unwritten(bh) || buffer_uptodate(bh)) == seek_data)
-			return lastoff;
-
-		lastoff = offset;
-	} while ((bh = bh->b_this_page) != head);
-	return -ENOENT;
-}
-
-/*
- * Seek for SEEK_DATA / SEEK_HOLE in the page cache.
- *
- * Within unwritten extents, the page cache determines which parts are holes
- * and which are data: unwritten and uptodate buffer heads count as data;
- * everything else counts as a hole.
- *
- * Returns the resulting offset on successs, and -ENOENT otherwise.
- */
-loff_t
-page_cache_seek_hole_data(struct inode *inode, loff_t offset, loff_t length,
-			  int whence)
-{
-	pgoff_t index = offset >> PAGE_SHIFT;
-	pgoff_t end = DIV_ROUND_UP(offset + length, PAGE_SIZE);
-	loff_t lastoff = offset;
-	struct pagevec pvec;
-
-	if (length <= 0)
-		return -ENOENT;
-
-	pagevec_init(&pvec);
-
-	do {
-		unsigned nr_pages, i;
-
-		nr_pages = pagevec_lookup_range(&pvec, inode->i_mapping, &index,
-						end - 1);
-		if (nr_pages == 0)
-			break;
-
-		for (i = 0; i < nr_pages; i++) {
-			struct page *page = pvec.pages[i];
-
-			/*
-			 * At this point, the page may be truncated or
-			 * invalidated (changing page->mapping to NULL), or
-			 * even swizzled back from swapper_space to tmpfs file
-			 * mapping.  However, page->index will not change
-			 * because we have a reference on the page.
-                         *
-			 * If current page offset is beyond where we've ended,
-			 * we've found a hole.
-                         */
-			if (whence == SEEK_HOLE &&
-			    lastoff < page_offset(page))
-				goto check_range;
-
-			lock_page(page);
-			if (likely(page->mapping == inode->i_mapping) &&
-			    page_has_buffers(page)) {
-				lastoff = page_seek_hole_data(page, lastoff, whence);
-				if (lastoff >= 0) {
-					unlock_page(page);
-					goto check_range;
-				}
-			}
-			unlock_page(page);
-			lastoff = page_offset(page) + PAGE_SIZE;
-		}
-		pagevec_release(&pvec);
-	} while (index < end);
-
-	/* When no page at lastoff and we are not done, we found a hole. */
-	if (whence != SEEK_HOLE)
-		goto not_found;
-
-check_range:
-	if (lastoff < offset + length)
-		goto out;
-not_found:
-	lastoff = -ENOENT;
-out:
-	pagevec_release(&pvec);
-	return lastoff;
-}
-
 void __init buffer_init(void)
 {
 	unsigned long nrpages;
diff --git a/fs/ext2/super.c b/fs/ext2/super.c
index de1694512f1f..c09289a42dc5 100644
--- a/fs/ext2/super.c
+++ b/fs/ext2/super.c
@@ -961,8 +961,7 @@ static int ext2_fill_super(struct super_block *sb, void *data, int silent)
 	blocksize = BLOCK_SIZE << le32_to_cpu(sbi->s_es->s_log_block_size);
 
 	if (sbi->s_mount_opt & EXT2_MOUNT_DAX) {
-		err = bdev_dax_supported(sb, blocksize);
-		if (err) {
+		if (!bdev_dax_supported(sb->s_bdev, blocksize)) {
 			ext2_msg(sb, KERN_ERR,
 				"DAX unsupported by block device. Turning off DAX.");
 			sbi->s_mount_opt &= ~EXT2_MOUNT_DAX;
diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c
index 44b4fcdc3755..285ed1588730 100644
--- a/fs/ext4/inline.c
+++ b/fs/ext4/inline.c
@@ -1841,8 +1841,8 @@ int ext4_inline_data_iomap(struct inode *inode, struct iomap *iomap)
 	iomap->offset = 0;
 	iomap->length = min_t(loff_t, ext4_get_inline_size(inode),
 			      i_size_read(inode));
-	iomap->type = 0;
-	iomap->flags = IOMAP_F_DATA_INLINE;
+	iomap->type = IOMAP_INLINE;
+	iomap->flags = 0;
 
 out:
 	up_read(&EXT4_I(inode)->xattr_sem);
diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index c1c5c8775ae7..7022d843c667 100644
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c
@@ -3762,8 +3762,7 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 					" that may contain inline data");
 			sbi->s_mount_opt &= ~EXT4_MOUNT_DAX;
 		}
-		err = bdev_dax_supported(sb, blocksize);
-		if (err) {
+		if (!bdev_dax_supported(sb->s_bdev, blocksize)) {
 			ext4_msg(sb, KERN_ERR,
 				"DAX unsupported by block device. Turning off DAX.");
 			sbi->s_mount_opt &= ~EXT4_MOUNT_DAX;
diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c
index a7b586e02693..ed6699705c13 100644
--- a/fs/gfs2/bmap.c
+++ b/fs/gfs2/bmap.c
@@ -767,10 +767,11 @@ static void gfs2_stuffed_iomap(struct inode *inode, struct iomap *iomap)
 		      sizeof(struct gfs2_dinode);
 	iomap->offset = 0;
 	iomap->length = i_size_read(inode);
-	iomap->type = IOMAP_MAPPED;
-	iomap->flags = IOMAP_F_DATA_INLINE;
+	iomap->type = IOMAP_INLINE;
 }
 
+#define IOMAP_F_GFS2_BOUNDARY IOMAP_F_PRIVATE
+
 /**
  * gfs2_iomap_get - Map blocks from an inode to disk blocks
  * @inode: The inode
@@ -846,7 +847,7 @@ static int gfs2_iomap_get(struct inode *inode, loff_t pos, loff_t length,
 	iomap->type = IOMAP_MAPPED;
 	iomap->flags = IOMAP_F_MERGED;
 	if (eob)
-		iomap->flags |= IOMAP_F_BOUNDARY;
+		iomap->flags |= IOMAP_F_GFS2_BOUNDARY;
 
 out:
 	iomap->bdev = inode->i_sb->s_bdev;
@@ -952,12 +953,12 @@ int gfs2_block_map(struct inode *inode, sector_t lblock,
 
 	if (iomap.length > bh_map->b_size) {
 		iomap.length = bh_map->b_size;
-		iomap.flags &= ~IOMAP_F_BOUNDARY;
+		iomap.flags &= ~IOMAP_F_GFS2_BOUNDARY;
 	}
 	if (iomap.addr != IOMAP_NULL_ADDR)
 		map_bh(bh_map, inode->i_sb, iomap.addr >> inode->i_blkbits);
 	bh_map->b_size = iomap.length;
-	if (iomap.flags & IOMAP_F_BOUNDARY)
+	if (iomap.flags & IOMAP_F_GFS2_BOUNDARY)
 		set_buffer_boundary(bh_map);
 	if (iomap.flags & IOMAP_F_NEW)
 		set_buffer_new(bh_map);
diff --git a/fs/iomap.c b/fs/iomap.c
index afd163586aa0..206539d369a8 100644
--- a/fs/iomap.c
+++ b/fs/iomap.c
@@ -20,6 +20,7 @@
 #include <linux/mm.h>
 #include <linux/swap.h>
 #include <linux/pagemap.h>
+#include <linux/pagevec.h>
 #include <linux/file.h>
 #include <linux/uio.h>
 #include <linux/backing-dev.h>
@@ -27,6 +28,7 @@
 #include <linux/task_io_accounting_ops.h>
 #include <linux/dax.h>
 #include <linux/sched/signal.h>
+#include <linux/swap.h>
 
 #include "internal.h"
 
@@ -95,6 +97,12 @@ iomap_apply(struct inode *inode, loff_t pos, loff_t length, unsigned flags,
 	return written ? written : ret;
 }
 
+static sector_t
+iomap_sector(struct iomap *iomap, loff_t pos)
+{
+	return (iomap->addr + pos - iomap->offset) >> SECTOR_SHIFT;
+}
+
 static void
 iomap_write_failed(struct inode *inode, loff_t pos, unsigned len)
 {
@@ -352,11 +360,8 @@ static int iomap_zero(struct inode *inode, loff_t pos, unsigned offset,
 static int iomap_dax_zero(loff_t pos, unsigned offset, unsigned bytes,
 		struct iomap *iomap)
 {
-	sector_t sector = (iomap->addr +
-			   (pos & PAGE_MASK) - iomap->offset) >> 9;
-
-	return __dax_zero_page_range(iomap->bdev, iomap->dax_dev, sector,
-			offset, bytes);
+	return __dax_zero_page_range(iomap->bdev, iomap->dax_dev,
+			iomap_sector(iomap, pos & PAGE_MASK), offset, bytes);
 }
 
 static loff_t
@@ -501,10 +506,13 @@ static int iomap_to_fiemap(struct fiemap_extent_info *fi,
 	case IOMAP_DELALLOC:
 		flags |= FIEMAP_EXTENT_DELALLOC | FIEMAP_EXTENT_UNKNOWN;
 		break;
+	case IOMAP_MAPPED:
+		break;
 	case IOMAP_UNWRITTEN:
 		flags |= FIEMAP_EXTENT_UNWRITTEN;
 		break;
-	case IOMAP_MAPPED:
+	case IOMAP_INLINE:
+		flags |= FIEMAP_EXTENT_DATA_INLINE;
 		break;
 	}
 
@@ -512,8 +520,6 @@ static int iomap_to_fiemap(struct fiemap_extent_info *fi,
 		flags |= FIEMAP_EXTENT_MERGED;
 	if (iomap->flags & IOMAP_F_SHARED)
 		flags |= FIEMAP_EXTENT_SHARED;
-	if (iomap->flags & IOMAP_F_DATA_INLINE)
-		flags |= FIEMAP_EXTENT_DATA_INLINE;
 
 	return fiemap_fill_next_extent(fi, iomap->offset,
 			iomap->addr != IOMAP_NULL_ADDR ? iomap->addr : 0,
@@ -587,6 +593,113 @@ int iomap_fiemap(struct inode *inode, struct fiemap_extent_info *fi,
 }
 EXPORT_SYMBOL_GPL(iomap_fiemap);
 
+/*
+ * Seek for SEEK_DATA / SEEK_HOLE within @page, starting at @lastoff.
+ * Returns true if found and updates @lastoff to the offset in file.
+ */
+static bool
+page_seek_hole_data(struct inode *inode, struct page *page, loff_t *lastoff,
+		int whence)
+{
+	const struct address_space_operations *ops = inode->i_mapping->a_ops;
+	unsigned int bsize = i_blocksize(inode), off;
+	bool seek_data = whence == SEEK_DATA;
+	loff_t poff = page_offset(page);
+
+	if (WARN_ON_ONCE(*lastoff >= poff + PAGE_SIZE))
+		return false;
+
+	if (*lastoff < poff) {
+		/*
+		 * Last offset smaller than the start of the page means we found
+		 * a hole:
+		 */
+		if (whence == SEEK_HOLE)
+			return true;
+		*lastoff = poff;
+	}
+
+	/*
+	 * Just check the page unless we can and should check block ranges:
+	 */
+	if (bsize == PAGE_SIZE || !ops->is_partially_uptodate)
+		return PageUptodate(page) == seek_data;
+
+	lock_page(page);
+	if (unlikely(page->mapping != inode->i_mapping))
+		goto out_unlock_not_found;
+
+	for (off = 0; off < PAGE_SIZE; off += bsize) {
+		if ((*lastoff & ~PAGE_MASK) >= off + bsize)
+			continue;
+		if (ops->is_partially_uptodate(page, off, bsize) == seek_data) {
+			unlock_page(page);
+			return true;
+		}
+		*lastoff = poff + off + bsize;
+	}
+
+out_unlock_not_found:
+	unlock_page(page);
+	return false;
+}
+
+/*
+ * Seek for SEEK_DATA / SEEK_HOLE in the page cache.
+ *
+ * Within unwritten extents, the page cache determines which parts are holes
+ * and which are data: uptodate buffer heads count as data; everything else
+ * counts as a hole.
+ *
+ * Returns the resulting offset on successs, and -ENOENT otherwise.
+ */
+static loff_t
+page_cache_seek_hole_data(struct inode *inode, loff_t offset, loff_t length,
+		int whence)
+{
+	pgoff_t index = offset >> PAGE_SHIFT;
+	pgoff_t end = DIV_ROUND_UP(offset + length, PAGE_SIZE);
+	loff_t lastoff = offset;
+	struct pagevec pvec;
+
+	if (length <= 0)
+		return -ENOENT;
+
+	pagevec_init(&pvec);
+
+	do {
+		unsigned nr_pages, i;
+
+		nr_pages = pagevec_lookup_range(&pvec, inode->i_mapping, &index,
+						end - 1);
+		if (nr_pages == 0)
+			break;
+
+		for (i = 0; i < nr_pages; i++) {
+			struct page *page = pvec.pages[i];
+
+			if (page_seek_hole_data(inode, page, &lastoff, whence))
+				goto check_range;
+			lastoff = page_offset(page) + PAGE_SIZE;
+		}
+		pagevec_release(&pvec);
+	} while (index < end);
+
+	/* When no page at lastoff and we are not done, we found a hole. */
+	if (whence != SEEK_HOLE)
+		goto not_found;
+
+check_range:
+	if (lastoff < offset + length)
+		goto out;
+not_found:
+	lastoff = -ENOENT;
+out:
+	pagevec_release(&pvec);
+	return lastoff;
+}
+
+
 static loff_t
 iomap_seek_hole_actor(struct inode *inode, loff_t offset, loff_t length,
 		      void *data, struct iomap *iomap)
@@ -685,6 +798,8 @@ EXPORT_SYMBOL_GPL(iomap_seek_data);
  * Private flags for iomap_dio, must not overlap with the public ones in
  * iomap.h:
  */
+#define IOMAP_DIO_WRITE_FUA	(1 << 28)
+#define IOMAP_DIO_NEED_SYNC	(1 << 29)
 #define IOMAP_DIO_WRITE		(1 << 30)
 #define IOMAP_DIO_DIRTY		(1 << 31)
 
@@ -759,6 +874,13 @@ static ssize_t iomap_dio_complete(struct iomap_dio *dio)
 			dio_warn_stale_pagecache(iocb->ki_filp);
 	}
 
+	/*
+	 * If this is a DSYNC write, make sure we push it to stable storage now
+	 * that we've written data.
+	 */
+	if (ret > 0 && (dio->flags & IOMAP_DIO_NEED_SYNC))
+		ret = generic_write_sync(iocb, ret);
+
 	inode_dio_end(file_inode(iocb->ki_filp));
 	kfree(dio);
 
@@ -769,13 +891,8 @@ static void iomap_dio_complete_work(struct work_struct *work)
 {
 	struct iomap_dio *dio = container_of(work, struct iomap_dio, aio.work);
 	struct kiocb *iocb = dio->iocb;
-	bool is_write = (dio->flags & IOMAP_DIO_WRITE);
-	ssize_t ret;
 
-	ret = iomap_dio_complete(dio);
-	if (is_write && ret > 0)
-		ret = generic_write_sync(iocb, ret);
-	iocb->ki_complete(iocb, ret, 0);
+	iocb->ki_complete(iocb, iomap_dio_complete(dio), 0);
 }
 
 /*
@@ -833,14 +950,12 @@ iomap_dio_zero(struct iomap_dio *dio, struct iomap *iomap, loff_t pos,
 
 	bio = bio_alloc(GFP_KERNEL, 1);
 	bio_set_dev(bio, iomap->bdev);
-	bio->bi_iter.bi_sector =
-		(iomap->addr + pos - iomap->offset) >> 9;
+	bio->bi_iter.bi_sector = iomap_sector(iomap, pos);
 	bio->bi_private = dio;
 	bio->bi_end_io = iomap_dio_bio_end_io;
 
 	get_page(page);
-	if (bio_add_page(bio, page, len, 0) != len)
-		BUG();
+	__bio_add_page(bio, page, len, 0);
 	bio_set_op_attrs(bio, REQ_OP_WRITE, REQ_SYNC | REQ_IDLE);
 
 	atomic_inc(&dio->ref);
@@ -858,6 +973,7 @@ iomap_dio_actor(struct inode *inode, loff_t pos, loff_t length,
 	struct iov_iter iter;
 	struct bio *bio;
 	bool need_zeroout = false;
+	bool use_fua = false;
 	int nr_pages, ret;
 	size_t copied = 0;
 
@@ -881,8 +997,20 @@ iomap_dio_actor(struct inode *inode, loff_t pos, loff_t length,
 	case IOMAP_MAPPED:
 		if (iomap->flags & IOMAP_F_SHARED)
 			dio->flags |= IOMAP_DIO_COW;
-		if (iomap->flags & IOMAP_F_NEW)
+		if (iomap->flags & IOMAP_F_NEW) {
 			need_zeroout = true;
+		} else {
+			/*
+			 * Use a FUA write if we need datasync semantics, this
+			 * is a pure data IO that doesn't require any metadata
+			 * updates and the underlying device supports FUA. This
+			 * allows us to avoid cache flushes on IO completion.
+			 */
+			if (!(iomap->flags & (IOMAP_F_SHARED|IOMAP_F_DIRTY)) &&
+			    (dio->flags & IOMAP_DIO_WRITE_FUA) &&
+			    blk_queue_fua(bdev_get_queue(iomap->bdev)))
+				use_fua = true;
+		}
 		break;
 	default:
 		WARN_ON_ONCE(1);
@@ -916,8 +1044,7 @@ iomap_dio_actor(struct inode *inode, loff_t pos, loff_t length,
 
 		bio = bio_alloc(GFP_KERNEL, nr_pages);
 		bio_set_dev(bio, iomap->bdev);
-		bio->bi_iter.bi_sector =
-			(iomap->addr + pos - iomap->offset) >> 9;
+		bio->bi_iter.bi_sector = iomap_sector(iomap, pos);
 		bio->bi_write_hint = dio->iocb->ki_hint;
 		bio->bi_private = dio;
 		bio->bi_end_io = iomap_dio_bio_end_io;
@@ -930,10 +1057,14 @@ iomap_dio_actor(struct inode *inode, loff_t pos, loff_t length,
 
 		n = bio->bi_iter.bi_size;
 		if (dio->flags & IOMAP_DIO_WRITE) {
-			bio_set_op_attrs(bio, REQ_OP_WRITE, REQ_SYNC | REQ_IDLE);
+			bio->bi_opf = REQ_OP_WRITE | REQ_SYNC | REQ_IDLE;
+			if (use_fua)
+				bio->bi_opf |= REQ_FUA;
+			else
+				dio->flags &= ~IOMAP_DIO_WRITE_FUA;
 			task_io_account_write(n);
 		} else {
-			bio_set_op_attrs(bio, REQ_OP_READ, 0);
+			bio->bi_opf = REQ_OP_READ;
 			if (dio->flags & IOMAP_DIO_DIRTY)
 				bio_set_pages_dirty(bio);
 		}
@@ -961,6 +1092,15 @@ iomap_dio_actor(struct inode *inode, loff_t pos, loff_t length,
 	return copied;
 }
 
+/*
+ * iomap_dio_rw() always completes O_[D]SYNC writes regardless of whether the IO
+ * is being issued as AIO or not.  This allows us to optimise pure data writes
+ * to use REQ_FUA rather than requiring generic_write_sync() to issue a
+ * REQ_FLUSH post write. This is slightly tricky because a single request here
+ * can be mapped into multiple disjoint IOs and only a subset of the IOs issued
+ * may be pure data writes. In that case, we still need to do a full data sync
+ * completion.
+ */
 ssize_t
 iomap_dio_rw(struct kiocb *iocb, struct iov_iter *iter,
 		const struct iomap_ops *ops, iomap_dio_end_io_t end_io)
@@ -1005,8 +1145,21 @@ iomap_dio_rw(struct kiocb *iocb, struct iov_iter *iter,
 		if (iter->type == ITER_IOVEC)
 			dio->flags |= IOMAP_DIO_DIRTY;
 	} else {
-		dio->flags |= IOMAP_DIO_WRITE;
 		flags |= IOMAP_WRITE;
+		dio->flags |= IOMAP_DIO_WRITE;
+
+		/* for data sync or sync, we need sync completion processing */
+		if (iocb->ki_flags & IOCB_DSYNC)
+			dio->flags |= IOMAP_DIO_NEED_SYNC;
+
+		/*
+		 * For datasync only writes, we optimistically try using FUA for
+		 * this IO.  Any non-FUA write that occurs will clear this flag,
+		 * hence we know before completion whether a cache flush is
+		 * necessary.
+		 */
+		if ((iocb->ki_flags & (IOCB_DSYNC | IOCB_SYNC)) == IOCB_DSYNC)
+			dio->flags |= IOMAP_DIO_WRITE_FUA;
 	}
 
 	if (iocb->ki_flags & IOCB_NOWAIT) {
@@ -1062,6 +1215,13 @@ iomap_dio_rw(struct kiocb *iocb, struct iov_iter *iter,
 	if (ret < 0)
 		iomap_dio_set_error(dio, ret);
 
+	/*
+	 * If all the writes we issued were FUA, we don't need to flush the
+	 * cache on IO completion. Clear the sync flag for this case.
+	 */
+	if (dio->flags & IOMAP_DIO_WRITE_FUA)
+		dio->flags &= ~IOMAP_DIO_NEED_SYNC;
+
 	if (!atomic_dec_and_test(&dio->ref)) {
 		if (!is_sync_kiocb(iocb))
 			return -EIOCBQUEUED;
@@ -1089,3 +1249,203 @@ out_free_dio:
 	return ret;
 }
 EXPORT_SYMBOL_GPL(iomap_dio_rw);
+
+/* Swapfile activation */
+
+#ifdef CONFIG_SWAP
+struct iomap_swapfile_info {
+	struct iomap iomap;		/* accumulated iomap */
+	struct swap_info_struct *sis;
+	uint64_t lowest_ppage;		/* lowest physical addr seen (pages) */
+	uint64_t highest_ppage;		/* highest physical addr seen (pages) */
+	unsigned long nr_pages;		/* number of pages collected */
+	int nr_extents;			/* extent count */
+};
+
+/*
+ * Collect physical extents for this swap file.  Physical extents reported to
+ * the swap code must be trimmed to align to a page boundary.  The logical
+ * offset within the file is irrelevant since the swapfile code maps logical
+ * page numbers of the swap device to the physical page-aligned extents.
+ */
+static int iomap_swapfile_add_extent(struct iomap_swapfile_info *isi)
+{
+	struct iomap *iomap = &isi->iomap;
+	unsigned long nr_pages;
+	uint64_t first_ppage;
+	uint64_t first_ppage_reported;
+	uint64_t next_ppage;
+	int error;
+
+	/*
+	 * Round the start up and the end down so that the physical
+	 * extent aligns to a page boundary.
+	 */
+	first_ppage = ALIGN(iomap->addr, PAGE_SIZE) >> PAGE_SHIFT;
+	next_ppage = ALIGN_DOWN(iomap->addr + iomap->length, PAGE_SIZE) >>
+			PAGE_SHIFT;
+
+	/* Skip too-short physical extents. */
+	if (first_ppage >= next_ppage)
+		return 0;
+	nr_pages = next_ppage - first_ppage;
+
+	/*
+	 * Calculate how much swap space we're adding; the first page contains
+	 * the swap header and doesn't count.  The mm still wants that first
+	 * page fed to add_swap_extent, however.
+	 */
+	first_ppage_reported = first_ppage;
+	if (iomap->offset == 0)
+		first_ppage_reported++;
+	if (isi->lowest_ppage > first_ppage_reported)
+		isi->lowest_ppage = first_ppage_reported;
+	if (isi->highest_ppage < (next_ppage - 1))
+		isi->highest_ppage = next_ppage - 1;
+
+	/* Add extent, set up for the next call. */
+	error = add_swap_extent(isi->sis, isi->nr_pages, nr_pages, first_ppage);
+	if (error < 0)
+		return error;
+	isi->nr_extents += error;
+	isi->nr_pages += nr_pages;
+	return 0;
+}
+
+/*
+ * Accumulate iomaps for this swap file.  We have to accumulate iomaps because
+ * swap only cares about contiguous page-aligned physical extents and makes no
+ * distinction between written and unwritten extents.
+ */
+static loff_t iomap_swapfile_activate_actor(struct inode *inode, loff_t pos,
+		loff_t count, void *data, struct iomap *iomap)
+{
+	struct iomap_swapfile_info *isi = data;
+	int error;
+
+	switch (iomap->type) {
+	case IOMAP_MAPPED:
+	case IOMAP_UNWRITTEN:
+		/* Only real or unwritten extents. */
+		break;
+	case IOMAP_INLINE:
+		/* No inline data. */
+		pr_err("swapon: file is inline\n");
+		return -EINVAL;
+	default:
+		pr_err("swapon: file has unallocated extents\n");
+		return -EINVAL;
+	}
+
+	/* No uncommitted metadata or shared blocks. */
+	if (iomap->flags & IOMAP_F_DIRTY) {
+		pr_err("swapon: file is not committed\n");
+		return -EINVAL;
+	}
+	if (iomap->flags & IOMAP_F_SHARED) {
+		pr_err("swapon: file has shared extents\n");
+		return -EINVAL;
+	}
+
+	/* Only one bdev per swap file. */
+	if (iomap->bdev != isi->sis->bdev) {
+		pr_err("swapon: file is on multiple devices\n");
+		return -EINVAL;
+	}
+
+	if (isi->iomap.length == 0) {
+		/* No accumulated extent, so just store it. */
+		memcpy(&isi->iomap, iomap, sizeof(isi->iomap));
+	} else if (isi->iomap.addr + isi->iomap.length == iomap->addr) {
+		/* Append this to the accumulated extent. */
+		isi->iomap.length += iomap->length;
+	} else {
+		/* Otherwise, add the retained iomap and store this one. */
+		error = iomap_swapfile_add_extent(isi);
+		if (error)
+			return error;
+		memcpy(&isi->iomap, iomap, sizeof(isi->iomap));
+	}
+	return count;
+}
+
+/*
+ * Iterate a swap file's iomaps to construct physical extents that can be
+ * passed to the swapfile subsystem.
+ */
+int iomap_swapfile_activate(struct swap_info_struct *sis,
+		struct file *swap_file, sector_t *pagespan,
+		const struct iomap_ops *ops)
+{
+	struct iomap_swapfile_info isi = {
+		.sis = sis,
+		.lowest_ppage = (sector_t)-1ULL,
+	};
+	struct address_space *mapping = swap_file->f_mapping;
+	struct inode *inode = mapping->host;
+	loff_t pos = 0;
+	loff_t len = ALIGN_DOWN(i_size_read(inode), PAGE_SIZE);
+	loff_t ret;
+
+	ret = filemap_write_and_wait(inode->i_mapping);
+	if (ret)
+		return ret;
+
+	while (len > 0) {
+		ret = iomap_apply(inode, pos, len, IOMAP_REPORT,
+				ops, &isi, iomap_swapfile_activate_actor);
+		if (ret <= 0)
+			return ret;
+
+		pos += ret;
+		len -= ret;
+	}
+
+	if (isi.iomap.length) {
+		ret = iomap_swapfile_add_extent(&isi);
+		if (ret)
+			return ret;
+	}
+
+	*pagespan = 1 + isi.highest_ppage - isi.lowest_ppage;
+	sis->max = isi.nr_pages;
+	sis->pages = isi.nr_pages - 1;
+	sis->highest_bit = isi.nr_pages - 1;
+	return isi.nr_extents;
+}
+EXPORT_SYMBOL_GPL(iomap_swapfile_activate);
+#endif /* CONFIG_SWAP */
+
+static loff_t
+iomap_bmap_actor(struct inode *inode, loff_t pos, loff_t length,
+		void *data, struct iomap *iomap)
+{
+	sector_t *bno = data, addr;
+
+	if (iomap->type == IOMAP_MAPPED) {
+		addr = (pos - iomap->offset + iomap->addr) >> inode->i_blkbits;
+		if (addr > INT_MAX)
+			WARN(1, "would truncate bmap result\n");
+		else
+			*bno = addr;
+	}
+	return 0;
+}
+
+/* legacy ->bmap interface.  0 is the error return (!) */
+sector_t
+iomap_bmap(struct address_space *mapping, sector_t bno,
+		const struct iomap_ops *ops)
+{
+	struct inode *inode = mapping->host;
+	loff_t pos = bno >> inode->i_blkbits;
+	unsigned blocksize = i_blocksize(inode);
+
+	if (filemap_write_and_wait(mapping))
+		return 0;
+
+	bno = 0;
+	iomap_apply(inode, pos, blocksize, 0, ops, &bno, iomap_bmap_actor);
+	return bno;
+}
+EXPORT_SYMBOL_GPL(iomap_bmap);
diff --git a/fs/xfs/Kconfig b/fs/xfs/Kconfig
index 46bcf0e649f5..457ac9f97377 100644
--- a/fs/xfs/Kconfig
+++ b/fs/xfs/Kconfig
@@ -85,6 +85,24 @@ config XFS_ONLINE_SCRUB
 
 	  If unsure, say N.
 
+config XFS_ONLINE_REPAIR
+	bool "XFS online metadata repair support"
+	default n
+	depends on XFS_FS && XFS_ONLINE_SCRUB
+	help
+	  If you say Y here you will be able to repair metadata on a
+	  mounted XFS filesystem.  This feature is intended to reduce
+	  filesystem downtime by fixing minor problems before they cause the
+	  filesystem to go down.  However, it requires that the filesystem be
+	  formatted with secondary metadata, such as reverse mappings and inode
+	  parent pointers.
+
+	  This feature is considered EXPERIMENTAL.  Use with caution!
+
+	  See the xfs_scrub man page in section 8 for additional information.
+
+	  If unsure, say N.
+
 config XFS_WARN
 	bool "XFS Verbose Warnings"
 	depends on XFS_FS && !XFS_DEBUG
diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile
index 7ceb41a9786a..e8d67a443bd7 100644
--- a/fs/xfs/Makefile
+++ b/fs/xfs/Makefile
@@ -28,6 +28,7 @@ xfs-y				+= xfs_trace.o
 
 # build the libxfs code first
 xfs-y				+= $(addprefix libxfs/, \
+				   xfs_ag.o \
 				   xfs_alloc.o \
 				   xfs_alloc_btree.o \
 				   xfs_attr.o \
@@ -163,4 +164,12 @@ xfs-y				+= $(addprefix scrub/, \
 
 xfs-$(CONFIG_XFS_RT)		+= scrub/rtbitmap.o
 xfs-$(CONFIG_XFS_QUOTA)		+= scrub/quota.o
+
+# online repair
+ifeq ($(CONFIG_XFS_ONLINE_REPAIR),y)
+xfs-y				+= $(addprefix scrub/, \
+				   agheader_repair.o \
+				   repair.o \
+				   )
+endif
 endif
diff --git a/fs/xfs/libxfs/xfs_ag.c b/fs/xfs/libxfs/xfs_ag.c
new file mode 100644
index 000000000000..9345802c99f7
--- /dev/null
+++ b/fs/xfs/libxfs/xfs_ag.c
@@ -0,0 +1,464 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * Copyright (c) 2018 Red Hat, Inc.
+ * All rights reserved.
+ */
+
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_sb.h"
+#include "xfs_mount.h"
+#include "xfs_btree.h"
+#include "xfs_alloc_btree.h"
+#include "xfs_rmap_btree.h"
+#include "xfs_alloc.h"
+#include "xfs_ialloc.h"
+#include "xfs_rmap.h"
+#include "xfs_ag.h"
+
+static struct xfs_buf *
+xfs_get_aghdr_buf(
+	struct xfs_mount	*mp,
+	xfs_daddr_t		blkno,
+	size_t			numblks,
+	int			flags,
+	const struct xfs_buf_ops *ops)
+{
+	struct xfs_buf		*bp;
+
+	bp = xfs_buf_get_uncached(mp->m_ddev_targp, numblks, flags);
+	if (!bp)
+		return NULL;
+
+	xfs_buf_zero(bp, 0, BBTOB(bp->b_length));
+	bp->b_bn = blkno;
+	bp->b_maps[0].bm_bn = blkno;
+	bp->b_ops = ops;
+
+	return bp;
+}
+
+/*
+ * Generic btree root block init function
+ */
+static void
+xfs_btroot_init(
+	struct xfs_mount	*mp,
+	struct xfs_buf		*bp,
+	struct aghdr_init_data	*id)
+{
+	xfs_btree_init_block(mp, bp, id->type, 0, 0, id->agno, 0);
+}
+
+/*
+ * Alloc btree root block init functions
+ */
+static void
+xfs_bnoroot_init(
+	struct xfs_mount	*mp,
+	struct xfs_buf		*bp,
+	struct aghdr_init_data	*id)
+{
+	struct xfs_alloc_rec	*arec;
+
+	xfs_btree_init_block(mp, bp, XFS_BTNUM_BNO, 0, 1, id->agno, 0);
+	arec = XFS_ALLOC_REC_ADDR(mp, XFS_BUF_TO_BLOCK(bp), 1);
+	arec->ar_startblock = cpu_to_be32(mp->m_ag_prealloc_blocks);
+	arec->ar_blockcount = cpu_to_be32(id->agsize -
+					  be32_to_cpu(arec->ar_startblock));
+}
+
+static void
+xfs_cntroot_init(
+	struct xfs_mount	*mp,
+	struct xfs_buf		*bp,
+	struct aghdr_init_data	*id)
+{
+	struct xfs_alloc_rec	*arec;
+
+	xfs_btree_init_block(mp, bp, XFS_BTNUM_CNT, 0, 1, id->agno, 0);
+	arec = XFS_ALLOC_REC_ADDR(mp, XFS_BUF_TO_BLOCK(bp), 1);
+	arec->ar_startblock = cpu_to_be32(mp->m_ag_prealloc_blocks);
+	arec->ar_blockcount = cpu_to_be32(id->agsize -
+					  be32_to_cpu(arec->ar_startblock));
+}
+
+/*
+ * Reverse map root block init
+ */
+static void
+xfs_rmaproot_init(
+	struct xfs_mount	*mp,
+	struct xfs_buf		*bp,
+	struct aghdr_init_data	*id)
+{
+	struct xfs_btree_block	*block = XFS_BUF_TO_BLOCK(bp);
+	struct xfs_rmap_rec	*rrec;
+
+	xfs_btree_init_block(mp, bp, XFS_BTNUM_RMAP, 0, 4, id->agno, 0);
+
+	/*
+	 * mark the AG header regions as static metadata The BNO
+	 * btree block is the first block after the headers, so
+	 * it's location defines the size of region the static
+	 * metadata consumes.
+	 *
+	 * Note: unlike mkfs, we never have to account for log
+	 * space when growing the data regions
+	 */
+	rrec = XFS_RMAP_REC_ADDR(block, 1);
+	rrec->rm_startblock = 0;
+	rrec->rm_blockcount = cpu_to_be32(XFS_BNO_BLOCK(mp));
+	rrec->rm_owner = cpu_to_be64(XFS_RMAP_OWN_FS);
+	rrec->rm_offset = 0;
+
+	/* account freespace btree root blocks */
+	rrec = XFS_RMAP_REC_ADDR(block, 2);
+	rrec->rm_startblock = cpu_to_be32(XFS_BNO_BLOCK(mp));
+	rrec->rm_blockcount = cpu_to_be32(2);
+	rrec->rm_owner = cpu_to_be64(XFS_RMAP_OWN_AG);
+	rrec->rm_offset = 0;
+
+	/* account inode btree root blocks */
+	rrec = XFS_RMAP_REC_ADDR(block, 3);
+	rrec->rm_startblock = cpu_to_be32(XFS_IBT_BLOCK(mp));
+	rrec->rm_blockcount = cpu_to_be32(XFS_RMAP_BLOCK(mp) -
+					  XFS_IBT_BLOCK(mp));
+	rrec->rm_owner = cpu_to_be64(XFS_RMAP_OWN_INOBT);
+	rrec->rm_offset = 0;
+
+	/* account for rmap btree root */
+	rrec = XFS_RMAP_REC_ADDR(block, 4);
+	rrec->rm_startblock = cpu_to_be32(XFS_RMAP_BLOCK(mp));
+	rrec->rm_blockcount = cpu_to_be32(1);
+	rrec->rm_owner = cpu_to_be64(XFS_RMAP_OWN_AG);
+	rrec->rm_offset = 0;
+
+	/* account for refc btree root */
+	if (xfs_sb_version_hasreflink(&mp->m_sb)) {
+		rrec = XFS_RMAP_REC_ADDR(block, 5);
+		rrec->rm_startblock = cpu_to_be32(xfs_refc_block(mp));
+		rrec->rm_blockcount = cpu_to_be32(1);
+		rrec->rm_owner = cpu_to_be64(XFS_RMAP_OWN_REFC);
+		rrec->rm_offset = 0;
+		be16_add_cpu(&block->bb_numrecs, 1);
+	}
+}
+
+/*
+ * Initialise new secondary superblocks with the pre-grow geometry, but mark
+ * them as "in progress" so we know they haven't yet been activated. This will
+ * get cleared when the update with the new geometry information is done after
+ * changes to the primary are committed. This isn't strictly necessary, but we
+ * get it for free with the delayed buffer write lists and it means we can tell
+ * if a grow operation didn't complete properly after the fact.
+ */
+static void
+xfs_sbblock_init(
+	struct xfs_mount	*mp,
+	struct xfs_buf		*bp,
+	struct aghdr_init_data	*id)
+{
+	struct xfs_dsb		*dsb = XFS_BUF_TO_SBP(bp);
+
+	xfs_sb_to_disk(dsb, &mp->m_sb);
+	dsb->sb_inprogress = 1;
+}
+
+static void
+xfs_agfblock_init(
+	struct xfs_mount	*mp,
+	struct xfs_buf		*bp,
+	struct aghdr_init_data	*id)
+{
+	struct xfs_agf		*agf = XFS_BUF_TO_AGF(bp);
+	xfs_extlen_t		tmpsize;
+
+	agf->agf_magicnum = cpu_to_be32(XFS_AGF_MAGIC);
+	agf->agf_versionnum = cpu_to_be32(XFS_AGF_VERSION);
+	agf->agf_seqno = cpu_to_be32(id->agno);
+	agf->agf_length = cpu_to_be32(id->agsize);
+	agf->agf_roots[XFS_BTNUM_BNOi] = cpu_to_be32(XFS_BNO_BLOCK(mp));
+	agf->agf_roots[XFS_BTNUM_CNTi] = cpu_to_be32(XFS_CNT_BLOCK(mp));
+	agf->agf_levels[XFS_BTNUM_BNOi] = cpu_to_be32(1);
+	agf->agf_levels[XFS_BTNUM_CNTi] = cpu_to_be32(1);
+	if (xfs_sb_version_hasrmapbt(&mp->m_sb)) {
+		agf->agf_roots[XFS_BTNUM_RMAPi] =
+					cpu_to_be32(XFS_RMAP_BLOCK(mp));
+		agf->agf_levels[XFS_BTNUM_RMAPi] = cpu_to_be32(1);
+		agf->agf_rmap_blocks = cpu_to_be32(1);
+	}
+
+	agf->agf_flfirst = cpu_to_be32(1);
+	agf->agf_fllast = 0;
+	agf->agf_flcount = 0;
+	tmpsize = id->agsize - mp->m_ag_prealloc_blocks;
+	agf->agf_freeblks = cpu_to_be32(tmpsize);
+	agf->agf_longest = cpu_to_be32(tmpsize);
+	if (xfs_sb_version_hascrc(&mp->m_sb))
+		uuid_copy(&agf->agf_uuid, &mp->m_sb.sb_meta_uuid);
+	if (xfs_sb_version_hasreflink(&mp->m_sb)) {
+		agf->agf_refcount_root = cpu_to_be32(
+				xfs_refc_block(mp));
+		agf->agf_refcount_level = cpu_to_be32(1);
+		agf->agf_refcount_blocks = cpu_to_be32(1);
+	}
+}
+
+static void
+xfs_agflblock_init(
+	struct xfs_mount	*mp,
+	struct xfs_buf		*bp,
+	struct aghdr_init_data	*id)
+{
+	struct xfs_agfl		*agfl = XFS_BUF_TO_AGFL(bp);
+	__be32			*agfl_bno;
+	int			bucket;
+
+	if (xfs_sb_version_hascrc(&mp->m_sb)) {
+		agfl->agfl_magicnum = cpu_to_be32(XFS_AGFL_MAGIC);
+		agfl->agfl_seqno = cpu_to_be32(id->agno);
+		uuid_copy(&agfl->agfl_uuid, &mp->m_sb.sb_meta_uuid);
+	}
+
+	agfl_bno = XFS_BUF_TO_AGFL_BNO(mp, bp);
+	for (bucket = 0; bucket < xfs_agfl_size(mp); bucket++)
+		agfl_bno[bucket] = cpu_to_be32(NULLAGBLOCK);
+}
+
+static void
+xfs_agiblock_init(
+	struct xfs_mount	*mp,
+	struct xfs_buf		*bp,
+	struct aghdr_init_data	*id)
+{
+	struct xfs_agi		*agi = XFS_BUF_TO_AGI(bp);
+	int			bucket;
+
+	agi->agi_magicnum = cpu_to_be32(XFS_AGI_MAGIC);
+	agi->agi_versionnum = cpu_to_be32(XFS_AGI_VERSION);
+	agi->agi_seqno = cpu_to_be32(id->agno);
+	agi->agi_length = cpu_to_be32(id->agsize);
+	agi->agi_count = 0;
+	agi->agi_root = cpu_to_be32(XFS_IBT_BLOCK(mp));
+	agi->agi_level = cpu_to_be32(1);
+	agi->agi_freecount = 0;
+	agi->agi_newino = cpu_to_be32(NULLAGINO);
+	agi->agi_dirino = cpu_to_be32(NULLAGINO);
+	if (xfs_sb_version_hascrc(&mp->m_sb))
+		uuid_copy(&agi->agi_uuid, &mp->m_sb.sb_meta_uuid);
+	if (xfs_sb_version_hasfinobt(&mp->m_sb)) {
+		agi->agi_free_root = cpu_to_be32(XFS_FIBT_BLOCK(mp));
+		agi->agi_free_level = cpu_to_be32(1);
+	}
+	for (bucket = 0; bucket < XFS_AGI_UNLINKED_BUCKETS; bucket++)
+		agi->agi_unlinked[bucket] = cpu_to_be32(NULLAGINO);
+}
+
+typedef void (*aghdr_init_work_f)(struct xfs_mount *mp, struct xfs_buf *bp,
+				  struct aghdr_init_data *id);
+static int
+xfs_ag_init_hdr(
+	struct xfs_mount	*mp,
+	struct aghdr_init_data	*id,
+	aghdr_init_work_f	work,
+	const struct xfs_buf_ops *ops)
+
+{
+	struct xfs_buf		*bp;
+
+	bp = xfs_get_aghdr_buf(mp, id->daddr, id->numblks, 0, ops);
+	if (!bp)
+		return -ENOMEM;
+
+	(*work)(mp, bp, id);
+
+	xfs_buf_delwri_queue(bp, &id->buffer_list);
+	xfs_buf_relse(bp);
+	return 0;
+}
+
+struct xfs_aghdr_grow_data {
+	xfs_daddr_t		daddr;
+	size_t			numblks;
+	const struct xfs_buf_ops *ops;
+	aghdr_init_work_f	work;
+	xfs_btnum_t		type;
+	bool			need_init;
+};
+
+/*
+ * Prepare new AG headers to be written to disk. We use uncached buffers here,
+ * as it is assumed these new AG headers are currently beyond the currently
+ * valid filesystem address space. Using cached buffers would trip over EOFS
+ * corruption detection alogrithms in the buffer cache lookup routines.
+ *
+ * This is a non-transactional function, but the prepared buffers are added to a
+ * delayed write buffer list supplied by the caller so they can submit them to
+ * disk and wait on them as required.
+ */
+int
+xfs_ag_init_headers(
+	struct xfs_mount	*mp,
+	struct aghdr_init_data	*id)
+
+{
+	struct xfs_aghdr_grow_data aghdr_data[] = {
+	{ /* SB */
+		.daddr = XFS_AG_DADDR(mp, id->agno, XFS_SB_DADDR),
+		.numblks = XFS_FSS_TO_BB(mp, 1),
+		.ops = &xfs_sb_buf_ops,
+		.work = &xfs_sbblock_init,
+		.need_init = true
+	},
+	{ /* AGF */
+		.daddr = XFS_AG_DADDR(mp, id->agno, XFS_AGF_DADDR(mp)),
+		.numblks = XFS_FSS_TO_BB(mp, 1),
+		.ops = &xfs_agf_buf_ops,
+		.work = &xfs_agfblock_init,
+		.need_init = true
+	},
+	{ /* AGFL */
+		.daddr = XFS_AG_DADDR(mp, id->agno, XFS_AGFL_DADDR(mp)),
+		.numblks = XFS_FSS_TO_BB(mp, 1),
+		.ops = &xfs_agfl_buf_ops,
+		.work = &xfs_agflblock_init,
+		.need_init = true
+	},
+	{ /* AGI */
+		.daddr = XFS_AG_DADDR(mp, id->agno, XFS_AGI_DADDR(mp)),
+		.numblks = XFS_FSS_TO_BB(mp, 1),
+		.ops = &xfs_agi_buf_ops,
+		.work = &xfs_agiblock_init,
+		.need_init = true
+	},
+	{ /* BNO root block */
+		.daddr = XFS_AGB_TO_DADDR(mp, id->agno, XFS_BNO_BLOCK(mp)),
+		.numblks = BTOBB(mp->m_sb.sb_blocksize),
+		.ops = &xfs_allocbt_buf_ops,
+		.work = &xfs_bnoroot_init,
+		.need_init = true
+	},
+	{ /* CNT root block */
+		.daddr = XFS_AGB_TO_DADDR(mp, id->agno, XFS_CNT_BLOCK(mp)),
+		.numblks = BTOBB(mp->m_sb.sb_blocksize),
+		.ops = &xfs_allocbt_buf_ops,
+		.work = &xfs_cntroot_init,
+		.need_init = true
+	},
+	{ /* INO root block */
+		.daddr = XFS_AGB_TO_DADDR(mp, id->agno, XFS_IBT_BLOCK(mp)),
+		.numblks = BTOBB(mp->m_sb.sb_blocksize),
+		.ops = &xfs_inobt_buf_ops,
+		.work = &xfs_btroot_init,
+		.type = XFS_BTNUM_INO,
+		.need_init = true
+	},
+	{ /* FINO root block */
+		.daddr = XFS_AGB_TO_DADDR(mp, id->agno, XFS_FIBT_BLOCK(mp)),
+		.numblks = BTOBB(mp->m_sb.sb_blocksize),
+		.ops = &xfs_inobt_buf_ops,
+		.work = &xfs_btroot_init,
+		.type = XFS_BTNUM_FINO,
+		.need_init =  xfs_sb_version_hasfinobt(&mp->m_sb)
+	},
+	{ /* RMAP root block */
+		.daddr = XFS_AGB_TO_DADDR(mp, id->agno, XFS_RMAP_BLOCK(mp)),
+		.numblks = BTOBB(mp->m_sb.sb_blocksize),
+		.ops = &xfs_rmapbt_buf_ops,
+		.work = &xfs_rmaproot_init,
+		.need_init = xfs_sb_version_hasrmapbt(&mp->m_sb)
+	},
+	{ /* REFC root block */
+		.daddr = XFS_AGB_TO_DADDR(mp, id->agno, xfs_refc_block(mp)),
+		.numblks = BTOBB(mp->m_sb.sb_blocksize),
+		.ops = &xfs_refcountbt_buf_ops,
+		.work = &xfs_btroot_init,
+		.type = XFS_BTNUM_REFC,
+		.need_init = xfs_sb_version_hasreflink(&mp->m_sb)
+	},
+	{ /* NULL terminating block */
+		.daddr = XFS_BUF_DADDR_NULL,
+	}
+	};
+	struct  xfs_aghdr_grow_data *dp;
+	int			error = 0;
+
+	/* Account for AG free space in new AG */
+	id->nfree += id->agsize - mp->m_ag_prealloc_blocks;
+	for (dp = &aghdr_data[0]; dp->daddr != XFS_BUF_DADDR_NULL; dp++) {
+		if (!dp->need_init)
+			continue;
+
+		id->daddr = dp->daddr;
+		id->numblks = dp->numblks;
+		id->type = dp->type;
+		error = xfs_ag_init_hdr(mp, id, dp->work, dp->ops);
+		if (error)
+			break;
+	}
+	return error;
+}
+
+/*
+ * Extent the AG indicated by the @id by the length passed in
+ */
+int
+xfs_ag_extend_space(
+	struct xfs_mount	*mp,
+	struct xfs_trans	*tp,
+	struct aghdr_init_data	*id,
+	xfs_extlen_t		len)
+{
+	struct xfs_owner_info	oinfo;
+	struct xfs_buf		*bp;
+	struct xfs_agi		*agi;
+	struct xfs_agf		*agf;
+	int			error;
+
+	/*
+	 * Change the agi length.
+	 */
+	error = xfs_ialloc_read_agi(mp, tp, id->agno, &bp);
+	if (error)
+		return error;
+
+	agi = XFS_BUF_TO_AGI(bp);
+	be32_add_cpu(&agi->agi_length, len);
+	ASSERT(id->agno == mp->m_sb.sb_agcount - 1 ||
+	       be32_to_cpu(agi->agi_length) == mp->m_sb.sb_agblocks);
+	xfs_ialloc_log_agi(tp, bp, XFS_AGI_LENGTH);
+
+	/*
+	 * Change agf length.
+	 */
+	error = xfs_alloc_read_agf(mp, tp, id->agno, 0, &bp);
+	if (error)
+		return error;
+
+	agf = XFS_BUF_TO_AGF(bp);
+	be32_add_cpu(&agf->agf_length, len);
+	ASSERT(agf->agf_length == agi->agi_length);
+	xfs_alloc_log_agf(tp, bp, XFS_AGF_LENGTH);
+
+	/*
+	 * Free the new space.
+	 *
+	 * XFS_RMAP_OWN_NULL is used here to tell the rmap btree that
+	 * this doesn't actually exist in the rmap btree.
+	 */
+	xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_NULL);
+	error = xfs_rmap_free(tp, bp, id->agno,
+				be32_to_cpu(agf->agf_length) - len,
+				len, &oinfo);
+	if (error)
+		return error;
+
+	return  xfs_free_extent(tp, XFS_AGB_TO_FSB(mp, id->agno,
+					be32_to_cpu(agf->agf_length) - len),
+				len, &oinfo, XFS_AG_RESV_NONE);
+}
diff --git a/fs/xfs/libxfs/xfs_ag.h b/fs/xfs/libxfs/xfs_ag.h
new file mode 100644
index 000000000000..412702e23f61
--- /dev/null
+++ b/fs/xfs/libxfs/xfs_ag.h
@@ -0,0 +1,30 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (c) 2018 Red Hat, Inc.
+ * All rights reserved.
+ */
+
+#ifndef __LIBXFS_AG_H
+#define __LIBXFS_AG_H 1
+
+struct xfs_mount;
+struct xfs_trans;
+
+struct aghdr_init_data {
+	/* per ag data */
+	xfs_agblock_t		agno;		/* ag to init */
+	xfs_extlen_t		agsize;		/* new AG size */
+	struct list_head	buffer_list;	/* buffer writeback list */
+	xfs_rfsblock_t		nfree;		/* cumulative new free space */
+
+	/* per header data */
+	xfs_daddr_t		daddr;		/* header location */
+	size_t			numblks;	/* size of header */
+	xfs_btnum_t		type;		/* type of btree root block */
+};
+
+int xfs_ag_init_headers(struct xfs_mount *mp, struct aghdr_init_data *id);
+int xfs_ag_extend_space(struct xfs_mount *mp, struct xfs_trans *tp,
+			struct aghdr_init_data *id, xfs_extlen_t len);
+
+#endif /* __LIBXFS_AG_H */
diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c
index 4bcc095fe44a..dc9dd3805d97 100644
--- a/fs/xfs/libxfs/xfs_alloc.c
+++ b/fs/xfs/libxfs/xfs_alloc.c
@@ -39,6 +39,9 @@
 #include "xfs_buf_item.h"
 #include "xfs_log.h"
 #include "xfs_ag_resv.h"
+#include "xfs_bmap.h"
+
+extern kmem_zone_t	*xfs_bmap_free_item_zone;
 
 struct workqueue_struct *xfs_alloc_wq;
 
@@ -2060,6 +2063,30 @@ xfs_alloc_space_available(
 	return true;
 }
 
+int
+xfs_free_agfl_block(
+	struct xfs_trans	*tp,
+	xfs_agnumber_t		agno,
+	xfs_agblock_t		agbno,
+	struct xfs_buf		*agbp,
+	struct xfs_owner_info	*oinfo)
+{
+	int			error;
+	struct xfs_buf		*bp;
+
+	error = xfs_free_ag_extent(tp, agbp, agno, agbno, 1, oinfo,
+				   XFS_AG_RESV_AGFL);
+	if (error)
+		return error;
+
+	bp = xfs_btree_get_bufs(tp->t_mountp, tp, agno, agbno, 0);
+	if (!bp)
+		return -EFSCORRUPTED;
+	xfs_trans_binval(tp, bp);
+
+	return 0;
+}
+
 /*
  * Check the agfl fields of the agf for inconsistency or corruption. The purpose
  * is to detect an agfl header padding mismatch between current and early v5
@@ -2148,6 +2175,40 @@ xfs_agfl_reset(
 }
 
 /*
+ * Defer an AGFL block free. This is effectively equivalent to
+ * xfs_bmap_add_free() with some special handling particular to AGFL blocks.
+ *
+ * Deferring AGFL frees helps prevent log reservation overruns due to too many
+ * allocation operations in a transaction. AGFL frees are prone to this problem
+ * because for one they are always freed one at a time. Further, an immediate
+ * AGFL block free can cause a btree join and require another block free before
+ * the real allocation can proceed. Deferring the free disconnects freeing up
+ * the AGFL slot from freeing the block.
+ */
+STATIC void
+xfs_defer_agfl_block(
+	struct xfs_mount		*mp,
+	struct xfs_defer_ops		*dfops,
+	xfs_agnumber_t			agno,
+	xfs_fsblock_t			agbno,
+	struct xfs_owner_info		*oinfo)
+{
+	struct xfs_extent_free_item	*new;		/* new element */
+
+	ASSERT(xfs_bmap_free_item_zone != NULL);
+	ASSERT(oinfo != NULL);
+
+	new = kmem_zone_alloc(xfs_bmap_free_item_zone, KM_SLEEP);
+	new->xefi_startblock = XFS_AGB_TO_FSB(mp, agno, agbno);
+	new->xefi_blockcount = 1;
+	new->xefi_oinfo = *oinfo;
+
+	trace_xfs_agfl_free_defer(mp, agno, 0, agbno, 1);
+
+	xfs_defer_add(dfops, XFS_DEFER_OPS_TYPE_AGFL_FREE, &new->xefi_list);
+}
+
+/*
  * Decide whether to use this allocation group for this allocation.
  * If so, fix up the btree freelist's size.
  */
@@ -2247,21 +2308,20 @@ xfs_alloc_fix_freelist(
 	else
 		xfs_rmap_ag_owner(&targs.oinfo, XFS_RMAP_OWN_AG);
 	while (!(flags & XFS_ALLOC_FLAG_NOSHRINK) && pag->pagf_flcount > need) {
-		struct xfs_buf	*bp;
-
 		error = xfs_alloc_get_freelist(tp, agbp, &bno, 0);
 		if (error)
 			goto out_agbp_relse;
-		error = xfs_free_ag_extent(tp, agbp, args->agno, bno, 1,
-					   &targs.oinfo, XFS_AG_RESV_AGFL);
-		if (error)
-			goto out_agbp_relse;
-		bp = xfs_btree_get_bufs(mp, tp, args->agno, bno, 0);
-		if (!bp) {
-			error = -EFSCORRUPTED;
-			goto out_agbp_relse;
+
+		/* defer agfl frees if dfops is provided */
+		if (tp->t_agfl_dfops) {
+			xfs_defer_agfl_block(mp, tp->t_agfl_dfops, args->agno,
+					     bno, &targs.oinfo);
+		} else {
+			error = xfs_free_agfl_block(tp, args->agno, bno, agbp,
+						    &targs.oinfo);
+			if (error)
+				goto out_agbp_relse;
 		}
-		xfs_trans_binval(tp, bp);
 	}
 
 	targs.tp = tp;
@@ -2949,18 +3009,20 @@ out:
  * after fixing up the freelist.
  */
 int				/* error */
-xfs_free_extent(
+__xfs_free_extent(
 	struct xfs_trans	*tp,	/* transaction pointer */
 	xfs_fsblock_t		bno,	/* starting block number of extent */
 	xfs_extlen_t		len,	/* length of extent */
 	struct xfs_owner_info	*oinfo,	/* extent owner */
-	enum xfs_ag_resv_type	type)	/* block reservation type */
+	enum xfs_ag_resv_type	type,	/* block reservation type */
+	bool			skip_discard)
 {
 	struct xfs_mount	*mp = tp->t_mountp;
 	struct xfs_buf		*agbp;
 	xfs_agnumber_t		agno = XFS_FSB_TO_AGNO(mp, bno);
 	xfs_agblock_t		agbno = XFS_FSB_TO_AGBNO(mp, bno);
 	int			error;
+	unsigned int		busy_flags = 0;
 
 	ASSERT(len != 0);
 	ASSERT(type != XFS_AG_RESV_AGFL);
@@ -2984,7 +3046,9 @@ xfs_free_extent(
 	if (error)
 		goto err;
 
-	xfs_extent_busy_insert(tp, agno, agbno, len, 0);
+	if (skip_discard)
+		busy_flags |= XFS_EXTENT_BUSY_SKIP_DISCARD;
+	xfs_extent_busy_insert(tp, agno, agbno, len, busy_flags);
 	return 0;
 
 err:
@@ -3116,3 +3180,40 @@ xfs_alloc_has_record(
 
 	return xfs_btree_has_record(cur, &low, &high, exists);
 }
+
+/*
+ * Walk all the blocks in the AGFL.  The @walk_fn can return any negative
+ * error code or XFS_BTREE_QUERY_RANGE_ABORT.
+ */
+int
+xfs_agfl_walk(
+	struct xfs_mount	*mp,
+	struct xfs_agf		*agf,
+	struct xfs_buf		*agflbp,
+	xfs_agfl_walk_fn	walk_fn,
+	void			*priv)
+{
+	__be32			*agfl_bno;
+	unsigned int		i;
+	int			error;
+
+	agfl_bno = XFS_BUF_TO_AGFL_BNO(mp, agflbp);
+	i = be32_to_cpu(agf->agf_flfirst);
+
+	/* Nothing to walk in an empty AGFL. */
+	if (agf->agf_flcount == cpu_to_be32(0))
+		return 0;
+
+	/* Otherwise, walk from first to last, wrapping as needed. */
+	for (;;) {
+		error = walk_fn(mp, be32_to_cpu(agfl_bno[i]), priv);
+		if (error)
+			return error;
+		if (i == be32_to_cpu(agf->agf_fllast))
+			break;
+		if (++i == xfs_agfl_size(mp))
+			i = 0;
+	}
+
+	return 0;
+}
diff --git a/fs/xfs/libxfs/xfs_alloc.h b/fs/xfs/libxfs/xfs_alloc.h
index cbf789ea5a4e..0747adcd57d6 100644
--- a/fs/xfs/libxfs/xfs_alloc.h
+++ b/fs/xfs/libxfs/xfs_alloc.h
@@ -191,12 +191,24 @@ xfs_alloc_vextent(
  * Free an extent.
  */
 int				/* error */
-xfs_free_extent(
+__xfs_free_extent(
 	struct xfs_trans	*tp,	/* transaction pointer */
 	xfs_fsblock_t		bno,	/* starting block number of extent */
 	xfs_extlen_t		len,	/* length of extent */
 	struct xfs_owner_info	*oinfo,	/* extent owner */
-	enum xfs_ag_resv_type	type);	/* block reservation type */
+	enum xfs_ag_resv_type	type,	/* block reservation type */
+	bool			skip_discard);
+
+static inline int
+xfs_free_extent(
+	struct xfs_trans	*tp,
+	xfs_fsblock_t		bno,
+	xfs_extlen_t		len,
+	struct xfs_owner_info	*oinfo,
+	enum xfs_ag_resv_type	type)
+{
+	return __xfs_free_extent(tp, bno, len, oinfo, type, false);
+}
 
 int				/* error */
 xfs_alloc_lookup_le(
@@ -223,6 +235,8 @@ int xfs_read_agf(struct xfs_mount *mp, struct xfs_trans *tp,
 			xfs_agnumber_t agno, int flags, struct xfs_buf **bpp);
 int xfs_alloc_read_agfl(struct xfs_mount *mp, struct xfs_trans *tp,
 			xfs_agnumber_t agno, struct xfs_buf **bpp);
+int xfs_free_agfl_block(struct xfs_trans *, xfs_agnumber_t, xfs_agblock_t,
+			struct xfs_buf *, struct xfs_owner_info *);
 int xfs_alloc_fix_freelist(struct xfs_alloc_arg *args, int flags);
 int xfs_free_extent_fix_freelist(struct xfs_trans *tp, xfs_agnumber_t agno,
 		struct xfs_buf **agbp);
@@ -248,4 +262,9 @@ bool xfs_verify_fsbno(struct xfs_mount *mp, xfs_fsblock_t fsbno);
 int xfs_alloc_has_record(struct xfs_btree_cur *cur, xfs_agblock_t bno,
 		xfs_extlen_t len, bool *exist);
 
+typedef int (*xfs_agfl_walk_fn)(struct xfs_mount *mp, xfs_agblock_t bno,
+		void *priv);
+int xfs_agfl_walk(struct xfs_mount *mp, struct xfs_agf *agf,
+		struct xfs_buf *agflbp, xfs_agfl_walk_fn walk_fn, void *priv);
+
 #endif	/* __XFS_ALLOC_H__ */
diff --git a/fs/xfs/libxfs/xfs_alloc_btree.c b/fs/xfs/libxfs/xfs_alloc_btree.c
index b451649ba176..18aec7a0e599 100644
--- a/fs/xfs/libxfs/xfs_alloc_btree.c
+++ b/fs/xfs/libxfs/xfs_alloc_btree.c
@@ -547,3 +547,12 @@ xfs_allocbt_maxrecs(
 		return blocklen / sizeof(xfs_alloc_rec_t);
 	return blocklen / (sizeof(xfs_alloc_key_t) + sizeof(xfs_alloc_ptr_t));
 }
+
+/* Calculate the freespace btree size for some records. */
+xfs_extlen_t
+xfs_allocbt_calc_size(
+	struct xfs_mount	*mp,
+	unsigned long long	len)
+{
+	return xfs_btree_calc_size(mp->m_alloc_mnr, len);
+}
diff --git a/fs/xfs/libxfs/xfs_alloc_btree.h b/fs/xfs/libxfs/xfs_alloc_btree.h
index 45e189e7e81c..2fd54728871c 100644
--- a/fs/xfs/libxfs/xfs_alloc_btree.h
+++ b/fs/xfs/libxfs/xfs_alloc_btree.h
@@ -61,5 +61,7 @@ extern struct xfs_btree_cur *xfs_allocbt_init_cursor(struct xfs_mount *,
 		struct xfs_trans *, struct xfs_buf *,
 		xfs_agnumber_t, xfs_btnum_t);
 extern int xfs_allocbt_maxrecs(struct xfs_mount *, int, int);
+extern xfs_extlen_t xfs_allocbt_calc_size(struct xfs_mount *mp,
+		unsigned long long len);
 
 #endif	/* __XFS_ALLOC_BTREE_H__ */
diff --git a/fs/xfs/libxfs/xfs_attr.c b/fs/xfs/libxfs/xfs_attr.c
index 35a124400d60..c3d02a66d39d 100644
--- a/fs/xfs/libxfs/xfs_attr.c
+++ b/fs/xfs/libxfs/xfs_attr.c
@@ -236,7 +236,7 @@ xfs_attr_set(
 	args.op_flags = XFS_DA_OP_ADDNAME | XFS_DA_OP_OKNOENT;
 	args.total = xfs_attr_calc_size(&args, &local);
 
-	error = xfs_qm_dqattach(dp, 0);
+	error = xfs_qm_dqattach(dp);
 	if (error)
 		return error;
 
@@ -427,7 +427,7 @@ xfs_attr_remove(
 	 */
 	args.op_flags = XFS_DA_OP_OKNOENT;
 
-	error = xfs_qm_dqattach(dp, 0);
+	error = xfs_qm_dqattach(dp);
 	if (error)
 		return error;
 
diff --git a/fs/xfs/libxfs/xfs_attr_remote.c b/fs/xfs/libxfs/xfs_attr_remote.c
index 21be186067a2..83a6d3c7f872 100644
--- a/fs/xfs/libxfs/xfs_attr_remote.c
+++ b/fs/xfs/libxfs/xfs_attr_remote.c
@@ -620,7 +620,7 @@ xfs_attr_rmtval_remove(
 		/*
 		 * If the "remote" value is in the cache, remove it.
 		 */
-		bp = xfs_incore(mp->m_ddev_targp, dblkno, dblkcnt, XBF_TRYLOCK);
+		bp = xfs_buf_incore(mp->m_ddev_targp, dblkno, dblkcnt, XBF_TRYLOCK);
 		if (bp) {
 			xfs_buf_stale(bp);
 			xfs_buf_relse(bp);
diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c
index 040eeda8426f..7b0e2b551e23 100644
--- a/fs/xfs/libxfs/xfs_bmap.c
+++ b/fs/xfs/libxfs/xfs_bmap.c
@@ -246,7 +246,7 @@ xfs_bmap_get_bp(
 	struct xfs_btree_cur	*cur,
 	xfs_fsblock_t		bno)
 {
-	struct xfs_log_item_desc *lidp;
+	struct xfs_log_item	*lip;
 	int			i;
 
 	if (!cur)
@@ -260,9 +260,9 @@ xfs_bmap_get_bp(
 	}
 
 	/* Chase down all the log items to see if the bp is there */
-	list_for_each_entry(lidp, &cur->bc_tp->t_items, lid_trans) {
-		struct xfs_buf_log_item	*bip;
-		bip = (struct xfs_buf_log_item *)lidp->lid_item;
+	list_for_each_entry(lip, &cur->bc_tp->t_items, li_trans) {
+		struct xfs_buf_log_item	*bip = (struct xfs_buf_log_item *)lip;
+
 		if (bip->bli_item.li_type == XFS_LI_BUF &&
 		    XFS_BUF_ADDR(bip->bli_buf) == bno)
 			return bip->bli_buf;
@@ -312,8 +312,9 @@ xfs_check_block(
 				xfs_warn(mp, "%s: thispa(%d) == pp(%d) %Ld",
 					__func__, j, i,
 					(unsigned long long)be64_to_cpu(*thispa));
-				panic("%s: ptrs are equal in node\n",
+				xfs_err(mp, "%s: ptrs are equal in node\n",
 					__func__);
+				xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
 			}
 		}
 	}
@@ -483,7 +484,8 @@ error0:
 error_norelse:
 	xfs_warn(mp, "%s: BAD after btree leaves for %d extents",
 		__func__, i);
-	panic("%s: CORRUPTED BTREE OR SOMETHING", __func__);
+	xfs_err(mp, "%s: CORRUPTED BTREE OR SOMETHING", __func__);
+	xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
 	return;
 }
 
@@ -542,12 +544,13 @@ xfs_bmap_validate_ret(
  * The list is maintained sorted (by block number).
  */
 void
-xfs_bmap_add_free(
+__xfs_bmap_add_free(
 	struct xfs_mount		*mp,
 	struct xfs_defer_ops		*dfops,
 	xfs_fsblock_t			bno,
 	xfs_filblks_t			len,
-	struct xfs_owner_info		*oinfo)
+	struct xfs_owner_info		*oinfo,
+	bool				skip_discard)
 {
 	struct xfs_extent_free_item	*new;		/* new element */
 #ifdef DEBUG
@@ -574,6 +577,7 @@ xfs_bmap_add_free(
 		new->xefi_oinfo = *oinfo;
 	else
 		xfs_rmap_skip_owner_update(&new->xefi_oinfo);
+	new->xefi_skip_discard = skip_discard;
 	trace_xfs_bmap_free_defer(mp, XFS_FSB_TO_AGNO(mp, bno), 0,
 			XFS_FSB_TO_AGBNO(mp, bno), len);
 	xfs_defer_add(dfops, XFS_DEFER_OPS_TYPE_FREE, &new->xefi_list);
@@ -2001,10 +2005,13 @@ xfs_bmap_add_extent_delay_real(
 		ASSERT(0);
 	}
 
-	/* add reverse mapping */
-	error = xfs_rmap_map_extent(mp, bma->dfops, bma->ip, whichfork, new);
-	if (error)
-		goto done;
+	/* add reverse mapping unless caller opted out */
+	if (!(bma->flags & XFS_BMAPI_NORMAP)) {
+		error = xfs_rmap_map_extent(mp, bma->dfops, bma->ip,
+				whichfork, new);
+		if (error)
+			goto done;
+	}
 
 	/* convert to a btree if necessary */
 	if (xfs_bmap_needs_btree(bma->ip, whichfork)) {
@@ -2668,7 +2675,8 @@ xfs_bmap_add_extent_hole_real(
 	struct xfs_bmbt_irec	*new,
 	xfs_fsblock_t		*first,
 	struct xfs_defer_ops	*dfops,
-	int			*logflagsp)
+	int			*logflagsp,
+	int			flags)
 {
 	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, whichfork);
 	struct xfs_mount	*mp = ip->i_mount;
@@ -2845,10 +2853,12 @@ xfs_bmap_add_extent_hole_real(
 		break;
 	}
 
-	/* add reverse mapping */
-	error = xfs_rmap_map_extent(mp, dfops, ip, whichfork, new);
-	if (error)
-		goto done;
+	/* add reverse mapping unless caller opted out */
+	if (!(flags & XFS_BMAPI_NORMAP)) {
+		error = xfs_rmap_map_extent(mp, dfops, ip, whichfork, new);
+		if (error)
+			goto done;
+	}
 
 	/* convert to a btree if necessary */
 	if (xfs_bmap_needs_btree(ip, whichfork)) {
@@ -4123,7 +4133,8 @@ xfs_bmapi_allocate(
 	else
 		error = xfs_bmap_add_extent_hole_real(bma->tp, bma->ip,
 				whichfork, &bma->icur, &bma->cur, &bma->got,
-				bma->firstblock, bma->dfops, &bma->logflags);
+				bma->firstblock, bma->dfops, &bma->logflags,
+				bma->flags);
 
 	bma->logflags |= tmp_logflags;
 	if (error)
@@ -4509,30 +4520,37 @@ error0:
 	return error;
 }
 
-static int
+int
 xfs_bmapi_remap(
 	struct xfs_trans	*tp,
 	struct xfs_inode	*ip,
 	xfs_fileoff_t		bno,
 	xfs_filblks_t		len,
 	xfs_fsblock_t		startblock,
-	struct xfs_defer_ops	*dfops)
+	struct xfs_defer_ops	*dfops,
+	int			flags)
 {
 	struct xfs_mount	*mp = ip->i_mount;
-	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
+	struct xfs_ifork	*ifp;
 	struct xfs_btree_cur	*cur = NULL;
 	xfs_fsblock_t		firstblock = NULLFSBLOCK;
 	struct xfs_bmbt_irec	got;
 	struct xfs_iext_cursor	icur;
+	int			whichfork = xfs_bmapi_whichfork(flags);
 	int			logflags = 0, error;
 
+	ifp = XFS_IFORK_PTR(ip, whichfork);
 	ASSERT(len > 0);
 	ASSERT(len <= (xfs_filblks_t)MAXEXTLEN);
 	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	ASSERT(!(flags & ~(XFS_BMAPI_ATTRFORK | XFS_BMAPI_PREALLOC |
+			   XFS_BMAPI_NORMAP)));
+	ASSERT((flags & (XFS_BMAPI_ATTRFORK | XFS_BMAPI_PREALLOC)) !=
+			(XFS_BMAPI_ATTRFORK | XFS_BMAPI_PREALLOC));
 
 	if (unlikely(XFS_TEST_ERROR(
-	    (XFS_IFORK_FORMAT(ip, XFS_DATA_FORK) != XFS_DINODE_FMT_EXTENTS &&
-	     XFS_IFORK_FORMAT(ip, XFS_DATA_FORK) != XFS_DINODE_FMT_BTREE),
+	    (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS &&
+	     XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE),
 	     mp, XFS_ERRTAG_BMAPIFORMAT))) {
 		XFS_ERROR_REPORT("xfs_bmapi_remap", XFS_ERRLEVEL_LOW, mp);
 		return -EFSCORRUPTED;
@@ -4542,7 +4560,7 @@ xfs_bmapi_remap(
 		return -EIO;
 
 	if (!(ifp->if_flags & XFS_IFEXTENTS)) {
-		error = xfs_iread_extents(NULL, ip, XFS_DATA_FORK);
+		error = xfs_iread_extents(tp, ip, whichfork);
 		if (error)
 			return error;
 	}
@@ -4557,7 +4575,7 @@ xfs_bmapi_remap(
 	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
 
 	if (ifp->if_flags & XFS_IFBROOT) {
-		cur = xfs_bmbt_init_cursor(mp, tp, ip, XFS_DATA_FORK);
+		cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork);
 		cur->bc_private.b.firstblock = firstblock;
 		cur->bc_private.b.dfops = dfops;
 		cur->bc_private.b.flags = 0;
@@ -4566,18 +4584,21 @@ xfs_bmapi_remap(
 	got.br_startoff = bno;
 	got.br_startblock = startblock;
 	got.br_blockcount = len;
-	got.br_state = XFS_EXT_NORM;
+	if (flags & XFS_BMAPI_PREALLOC)
+		got.br_state = XFS_EXT_UNWRITTEN;
+	else
+		got.br_state = XFS_EXT_NORM;
 
-	error = xfs_bmap_add_extent_hole_real(tp, ip, XFS_DATA_FORK, &icur,
-			&cur, &got, &firstblock, dfops, &logflags);
+	error = xfs_bmap_add_extent_hole_real(tp, ip, whichfork, &icur,
+			&cur, &got, &firstblock, dfops, &logflags, flags);
 	if (error)
 		goto error0;
 
-	if (xfs_bmap_wants_extents(ip, XFS_DATA_FORK)) {
+	if (xfs_bmap_wants_extents(ip, whichfork)) {
 		int		tmp_logflags = 0;
 
 		error = xfs_bmap_btree_to_extents(tp, ip, cur,
-			&tmp_logflags, XFS_DATA_FORK);
+			&tmp_logflags, whichfork);
 		logflags |= tmp_logflags;
 	}
 
@@ -5104,9 +5125,12 @@ xfs_bmap_del_extent_real(
 			error = xfs_refcount_decrease_extent(mp, dfops, del);
 			if (error)
 				goto done;
-		} else
-			xfs_bmap_add_free(mp, dfops, del->br_startblock,
-					del->br_blockcount, NULL);
+		} else {
+			__xfs_bmap_add_free(mp, dfops, del->br_startblock,
+					del->br_blockcount, NULL,
+					(bflags & XFS_BMAPI_NODISCARD) ||
+					del->br_state == XFS_EXT_UNWRITTEN);
+		}
 	}
 
 	/*
@@ -6148,7 +6172,7 @@ xfs_bmap_finish_one(
 	switch (type) {
 	case XFS_BMAP_MAP:
 		error = xfs_bmapi_remap(tp, ip, startoff, *blockcount,
-				startblock, dfops);
+				startblock, dfops, 0);
 		*blockcount = 0;
 		break;
 	case XFS_BMAP_UNMAP:
diff --git a/fs/xfs/libxfs/xfs_bmap.h b/fs/xfs/libxfs/xfs_bmap.h
index 2b766b37096d..2c233f9f1a26 100644
--- a/fs/xfs/libxfs/xfs_bmap.h
+++ b/fs/xfs/libxfs/xfs_bmap.h
@@ -68,6 +68,7 @@ struct xfs_extent_free_item
 	xfs_extlen_t		xefi_blockcount;/* number of blocks in extent */
 	struct list_head	xefi_list;
 	struct xfs_owner_info	xefi_oinfo;	/* extent owner */
+	bool			xefi_skip_discard;
 };
 
 #define	XFS_BMAP_MAX_NMAP	4
@@ -116,6 +117,12 @@ struct xfs_extent_free_item
 /* Only convert unwritten extents, don't allocate new blocks */
 #define XFS_BMAPI_CONVERT_ONLY	0x800
 
+/* Skip online discard of freed extents */
+#define XFS_BMAPI_NODISCARD	0x1000
+
+/* Do not update the rmap btree.  Used for reconstructing bmbt from rmapbt. */
+#define XFS_BMAPI_NORMAP	0x2000
+
 #define XFS_BMAPI_FLAGS \
 	{ XFS_BMAPI_ENTIRE,	"ENTIRE" }, \
 	{ XFS_BMAPI_METADATA,	"METADATA" }, \
@@ -128,7 +135,9 @@ struct xfs_extent_free_item
 	{ XFS_BMAPI_REMAP,	"REMAP" }, \
 	{ XFS_BMAPI_COWFORK,	"COWFORK" }, \
 	{ XFS_BMAPI_DELALLOC,	"DELALLOC" }, \
-	{ XFS_BMAPI_CONVERT_ONLY, "CONVERT_ONLY" }
+	{ XFS_BMAPI_CONVERT_ONLY, "CONVERT_ONLY" }, \
+	{ XFS_BMAPI_NODISCARD,	"NODISCARD" }, \
+	{ XFS_BMAPI_NORMAP,	"NORMAP" }
 
 
 static inline int xfs_bmapi_aflag(int w)
@@ -192,9 +201,9 @@ void	xfs_trim_extent(struct xfs_bmbt_irec *irec, xfs_fileoff_t bno,
 void	xfs_trim_extent_eof(struct xfs_bmbt_irec *, struct xfs_inode *);
 int	xfs_bmap_add_attrfork(struct xfs_inode *ip, int size, int rsvd);
 void	xfs_bmap_local_to_extents_empty(struct xfs_inode *ip, int whichfork);
-void	xfs_bmap_add_free(struct xfs_mount *mp, struct xfs_defer_ops *dfops,
+void	__xfs_bmap_add_free(struct xfs_mount *mp, struct xfs_defer_ops *dfops,
 			  xfs_fsblock_t bno, xfs_filblks_t len,
-			  struct xfs_owner_info *oinfo);
+			  struct xfs_owner_info *oinfo, bool skip_discard);
 void	xfs_bmap_compute_maxlevels(struct xfs_mount *mp, int whichfork);
 int	xfs_bmap_first_unused(struct xfs_trans *tp, struct xfs_inode *ip,
 		xfs_extlen_t len, xfs_fileoff_t *unused, int whichfork);
@@ -240,6 +249,17 @@ int	xfs_bmapi_reserve_delalloc(struct xfs_inode *ip, int whichfork,
 		struct xfs_bmbt_irec *got, struct xfs_iext_cursor *cur,
 		int eof);
 
+static inline void
+xfs_bmap_add_free(
+	struct xfs_mount		*mp,
+	struct xfs_defer_ops		*dfops,
+	xfs_fsblock_t			bno,
+	xfs_filblks_t			len,
+	struct xfs_owner_info		*oinfo)
+{
+	__xfs_bmap_add_free(mp, dfops, bno, len, oinfo, false);
+}
+
 enum xfs_bmap_intent_type {
 	XFS_BMAP_MAP = 1,
 	XFS_BMAP_UNMAP,
@@ -277,4 +297,8 @@ static inline int xfs_bmap_fork_to_state(int whichfork)
 xfs_failaddr_t xfs_bmap_validate_extent(struct xfs_inode *ip, int whichfork,
 		struct xfs_bmbt_irec *irec);
 
+int	xfs_bmapi_remap(struct xfs_trans *tp, struct xfs_inode *ip,
+		xfs_fileoff_t bno, xfs_filblks_t len, xfs_fsblock_t startblock,
+		struct xfs_defer_ops *dfops, int flags);
+
 #endif	/* __XFS_BMAP_H__ */
diff --git a/fs/xfs/libxfs/xfs_bmap_btree.c b/fs/xfs/libxfs/xfs_bmap_btree.c
index d89d06bea6e3..ac9d4aeedb09 100644
--- a/fs/xfs/libxfs/xfs_bmap_btree.c
+++ b/fs/xfs/libxfs/xfs_bmap_btree.c
@@ -660,3 +660,12 @@ xfs_bmbt_change_owner(
 	xfs_btree_del_cursor(cur, error ? XFS_BTREE_ERROR : XFS_BTREE_NOERROR);
 	return error;
 }
+
+/* Calculate the bmap btree size for some records. */
+unsigned long long
+xfs_bmbt_calc_size(
+	struct xfs_mount	*mp,
+	unsigned long long	len)
+{
+	return xfs_btree_calc_size(mp->m_bmap_dmnr, len);
+}
diff --git a/fs/xfs/libxfs/xfs_bmap_btree.h b/fs/xfs/libxfs/xfs_bmap_btree.h
index e4505746ccaa..fb3cd2d9e0f8 100644
--- a/fs/xfs/libxfs/xfs_bmap_btree.h
+++ b/fs/xfs/libxfs/xfs_bmap_btree.h
@@ -118,4 +118,7 @@ extern int xfs_bmbt_change_owner(struct xfs_trans *tp, struct xfs_inode *ip,
 extern struct xfs_btree_cur *xfs_bmbt_init_cursor(struct xfs_mount *,
 		struct xfs_trans *, struct xfs_inode *, int);
 
+extern unsigned long long xfs_bmbt_calc_size(struct xfs_mount *mp,
+		unsigned long long len);
+
 #endif	/* __XFS_BMAP_BTREE_H__ */
diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c
index ac7d66427e42..c825c8182b30 100644
--- a/fs/xfs/libxfs/xfs_btree.c
+++ b/fs/xfs/libxfs/xfs_btree.c
@@ -4836,14 +4836,14 @@ xfs_btree_query_all(
  * Calculate the number of blocks needed to store a given number of records
  * in a short-format (per-AG metadata) btree.
  */
-xfs_extlen_t
+unsigned long long
 xfs_btree_calc_size(
 	uint			*limits,
 	unsigned long long	len)
 {
 	int			level;
 	int			maxrecs;
-	xfs_extlen_t		rval;
+	unsigned long long	rval;
 
 	maxrecs = limits[0];
 	for (level = 0, rval = 0; len > 1; level++) {
@@ -4919,3 +4919,24 @@ xfs_btree_has_record(
 	*exists = false;
 	return error;
 }
+
+/* Are there more records in this btree? */
+bool
+xfs_btree_has_more_records(
+	struct xfs_btree_cur	*cur)
+{
+	struct xfs_btree_block	*block;
+	struct xfs_buf		*bp;
+
+	block = xfs_btree_get_block(cur, 0, &bp);
+
+	/* There are still records in this block. */
+	if (cur->bc_ptrs[0] < xfs_btree_get_numrecs(block))
+		return true;
+
+	/* There are more record blocks. */
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
+		return block->bb_u.l.bb_rightsib != cpu_to_be64(NULLFSBLOCK);
+	else
+		return block->bb_u.s.bb_rightsib != cpu_to_be32(NULLAGBLOCK);
+}
diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h
index 9227159a751e..d7911efee6dc 100644
--- a/fs/xfs/libxfs/xfs_btree.h
+++ b/fs/xfs/libxfs/xfs_btree.h
@@ -482,7 +482,7 @@ xfs_failaddr_t xfs_btree_lblock_verify(struct xfs_buf *bp,
 		unsigned int max_recs);
 
 uint xfs_btree_compute_maxlevels(uint *limits, unsigned long len);
-xfs_extlen_t xfs_btree_calc_size(uint *limits, unsigned long long len);
+unsigned long long xfs_btree_calc_size(uint *limits, unsigned long long len);
 
 /* return codes */
 #define XFS_BTREE_QUERY_RANGE_CONTINUE	0	/* keep iterating */
@@ -528,5 +528,6 @@ union xfs_btree_key *xfs_btree_high_key_from_key(struct xfs_btree_cur *cur,
 		union xfs_btree_key *key);
 int xfs_btree_has_record(struct xfs_btree_cur *cur, union xfs_btree_irec *low,
 		union xfs_btree_irec *high, bool *exists);
+bool xfs_btree_has_more_records(struct xfs_btree_cur *cur);
 
 #endif	/* __XFS_BTREE_H__ */
diff --git a/fs/xfs/libxfs/xfs_defer.c b/fs/xfs/libxfs/xfs_defer.c
index 087fea02c389..3daf175e2535 100644
--- a/fs/xfs/libxfs/xfs_defer.c
+++ b/fs/xfs/libxfs/xfs_defer.c
@@ -220,7 +220,7 @@ xfs_defer_trans_abort(
 {
 	struct xfs_defer_pending	*dfp;
 
-	trace_xfs_defer_trans_abort(tp->t_mountp, dop);
+	trace_xfs_defer_trans_abort(tp->t_mountp, dop, _RET_IP_);
 
 	/* Abort intent items that don't have a done item. */
 	list_for_each_entry(dfp, &dop->dop_pending, dfp_list) {
@@ -253,7 +253,7 @@ xfs_defer_trans_roll(
 	for (i = 0; i < XFS_DEFER_OPS_NR_BUFS && dop->dop_bufs[i]; i++)
 		xfs_trans_dirty_buf(*tp, dop->dop_bufs[i]);
 
-	trace_xfs_defer_trans_roll((*tp)->t_mountp, dop);
+	trace_xfs_defer_trans_roll((*tp)->t_mountp, dop, _RET_IP_);
 
 	/* Roll the transaction. */
 	error = xfs_trans_roll(tp);
@@ -352,10 +352,21 @@ xfs_defer_finish(
 	void				*state;
 	int				error = 0;
 	void				(*cleanup_fn)(struct xfs_trans *, void *, int);
+	struct xfs_defer_ops		*orig_dop;
 
 	ASSERT((*tp)->t_flags & XFS_TRANS_PERM_LOG_RES);
 
-	trace_xfs_defer_finish((*tp)->t_mountp, dop);
+	trace_xfs_defer_finish((*tp)->t_mountp, dop, _RET_IP_);
+
+	/*
+	 * Attach dfops to the transaction during deferred ops processing. This
+	 * explicitly causes calls into the allocator to defer AGFL block frees.
+	 * Note that this code can go away once all dfops users attach to the
+	 * associated tp.
+	 */
+	ASSERT(!(*tp)->t_agfl_dfops || ((*tp)->t_agfl_dfops == dop));
+	orig_dop = (*tp)->t_agfl_dfops;
+	(*tp)->t_agfl_dfops = dop;
 
 	/* Until we run out of pending work to finish... */
 	while (xfs_defer_has_unfinished_work(dop)) {
@@ -428,10 +439,11 @@ xfs_defer_finish(
 	}
 
 out:
+	(*tp)->t_agfl_dfops = orig_dop;
 	if (error)
 		trace_xfs_defer_finish_error((*tp)->t_mountp, dop, error);
 	else
-		trace_xfs_defer_finish_done((*tp)->t_mountp, dop);
+		trace_xfs_defer_finish_done((*tp)->t_mountp, dop, _RET_IP_);
 	return error;
 }
 
@@ -447,7 +459,7 @@ xfs_defer_cancel(
 	struct list_head		*pwi;
 	struct list_head		*n;
 
-	trace_xfs_defer_cancel(NULL, dop);
+	trace_xfs_defer_cancel(NULL, dop, _RET_IP_);
 
 	/*
 	 * Free the pending items.  Caller should already have arranged
@@ -532,5 +544,5 @@ xfs_defer_init(
 	*fbp = NULLFSBLOCK;
 	INIT_LIST_HEAD(&dop->dop_intake);
 	INIT_LIST_HEAD(&dop->dop_pending);
-	trace_xfs_defer_init(NULL, dop);
+	trace_xfs_defer_init(NULL, dop, _RET_IP_);
 }
diff --git a/fs/xfs/libxfs/xfs_defer.h b/fs/xfs/libxfs/xfs_defer.h
index 045beacdd37d..e70725ba1f5f 100644
--- a/fs/xfs/libxfs/xfs_defer.h
+++ b/fs/xfs/libxfs/xfs_defer.h
@@ -55,6 +55,7 @@ enum xfs_defer_ops_type {
 	XFS_DEFER_OPS_TYPE_REFCOUNT,
 	XFS_DEFER_OPS_TYPE_RMAP,
 	XFS_DEFER_OPS_TYPE_FREE,
+	XFS_DEFER_OPS_TYPE_AGFL_FREE,
 	XFS_DEFER_OPS_TYPE_MAX,
 };
 
diff --git a/fs/xfs/libxfs/xfs_dquot_buf.c b/fs/xfs/libxfs/xfs_dquot_buf.c
index 8b7a6c3cb599..cce520becee4 100644
--- a/fs/xfs/libxfs/xfs_dquot_buf.c
+++ b/fs/xfs/libxfs/xfs_dquot_buf.c
@@ -41,14 +41,18 @@ xfs_calc_dquots_per_chunk(
 
 /*
  * Do some primitive error checking on ondisk dquot data structures.
+ *
+ * The xfs_dqblk structure /contains/ the xfs_disk_dquot structure;
+ * we verify them separately because at some points we have only the
+ * smaller xfs_disk_dquot structure available.
  */
+
 xfs_failaddr_t
 xfs_dquot_verify(
 	struct xfs_mount *mp,
 	xfs_disk_dquot_t *ddq,
 	xfs_dqid_t	 id,
-	uint		 type,	  /* used only when IO_dorepair is true */
-	uint		 flags)
+	uint		 type)	  /* used only during quotacheck */
 {
 	/*
 	 * We can encounter an uninitialized dquot buffer for 2 reasons:
@@ -70,6 +74,8 @@ xfs_dquot_verify(
 	if (ddq->d_version != XFS_DQUOT_VERSION)
 		return __this_address;
 
+	if (type && ddq->d_flags != type)
+		return __this_address;
 	if (ddq->d_flags != XFS_DQ_USER &&
 	    ddq->d_flags != XFS_DQ_PROJ &&
 	    ddq->d_flags != XFS_DQ_GROUP)
@@ -99,33 +105,44 @@ xfs_dquot_verify(
 	return NULL;
 }
 
+xfs_failaddr_t
+xfs_dqblk_verify(
+	struct xfs_mount	*mp,
+	struct xfs_dqblk	*dqb,
+	xfs_dqid_t	 	id,
+	uint		 	type)	/* used only during quotacheck */
+{
+	if (xfs_sb_version_hascrc(&mp->m_sb) &&
+	    !uuid_equal(&dqb->dd_uuid, &mp->m_sb.sb_meta_uuid))
+		return __this_address;
+
+	return xfs_dquot_verify(mp, &dqb->dd_diskdq, id, type);
+}
+
 /*
  * Do some primitive error checking on ondisk dquot data structures.
  */
 int
-xfs_dquot_repair(
+xfs_dqblk_repair(
 	struct xfs_mount	*mp,
-	struct xfs_disk_dquot	*ddq,
+	struct xfs_dqblk	*dqb,
 	xfs_dqid_t		id,
 	uint			type)
 {
-	struct xfs_dqblk	*d = (struct xfs_dqblk *)ddq;
-
-
 	/*
 	 * Typically, a repair is only requested by quotacheck.
 	 */
 	ASSERT(id != -1);
-	memset(d, 0, sizeof(xfs_dqblk_t));
+	memset(dqb, 0, sizeof(xfs_dqblk_t));
 
-	d->dd_diskdq.d_magic = cpu_to_be16(XFS_DQUOT_MAGIC);
-	d->dd_diskdq.d_version = XFS_DQUOT_VERSION;
-	d->dd_diskdq.d_flags = type;
-	d->dd_diskdq.d_id = cpu_to_be32(id);
+	dqb->dd_diskdq.d_magic = cpu_to_be16(XFS_DQUOT_MAGIC);
+	dqb->dd_diskdq.d_version = XFS_DQUOT_VERSION;
+	dqb->dd_diskdq.d_flags = type;
+	dqb->dd_diskdq.d_id = cpu_to_be32(id);
 
 	if (xfs_sb_version_hascrc(&mp->m_sb)) {
-		uuid_copy(&d->dd_uuid, &mp->m_sb.sb_meta_uuid);
-		xfs_update_cksum((char *)d, sizeof(struct xfs_dqblk),
+		uuid_copy(&dqb->dd_uuid, &mp->m_sb.sb_meta_uuid);
+		xfs_update_cksum((char *)dqb, sizeof(struct xfs_dqblk),
 				 XFS_DQUOT_CRC_OFF);
 	}
 
@@ -135,7 +152,8 @@ xfs_dquot_repair(
 STATIC bool
 xfs_dquot_buf_verify_crc(
 	struct xfs_mount	*mp,
-	struct xfs_buf		*bp)
+	struct xfs_buf		*bp,
+	bool			readahead)
 {
 	struct xfs_dqblk	*d = (struct xfs_dqblk *)bp->b_addr;
 	int			ndquots;
@@ -156,10 +174,12 @@ xfs_dquot_buf_verify_crc(
 
 	for (i = 0; i < ndquots; i++, d++) {
 		if (!xfs_verify_cksum((char *)d, sizeof(struct xfs_dqblk),
-				 XFS_DQUOT_CRC_OFF))
-			return false;
-		if (!uuid_equal(&d->dd_uuid, &mp->m_sb.sb_meta_uuid))
+				 XFS_DQUOT_CRC_OFF)) {
+			if (!readahead)
+				xfs_buf_verifier_error(bp, -EFSBADCRC, __func__,
+					d, sizeof(*d), __this_address);
 			return false;
+		}
 	}
 	return true;
 }
@@ -167,9 +187,10 @@ xfs_dquot_buf_verify_crc(
 STATIC xfs_failaddr_t
 xfs_dquot_buf_verify(
 	struct xfs_mount	*mp,
-	struct xfs_buf		*bp)
+	struct xfs_buf		*bp,
+	bool			readahead)
 {
-	struct xfs_dqblk	*d = (struct xfs_dqblk *)bp->b_addr;
+	struct xfs_dqblk	*dqb = bp->b_addr;
 	xfs_failaddr_t		fa;
 	xfs_dqid_t		id = 0;
 	int			ndquots;
@@ -195,14 +216,19 @@ xfs_dquot_buf_verify(
 	for (i = 0; i < ndquots; i++) {
 		struct xfs_disk_dquot	*ddq;
 
-		ddq = &d[i].dd_diskdq;
+		ddq = &dqb[i].dd_diskdq;
 
 		if (i == 0)
 			id = be32_to_cpu(ddq->d_id);
 
-		fa = xfs_dquot_verify(mp, ddq, id + i, 0, 0);
-		if (fa)
+		fa = xfs_dqblk_verify(mp, &dqb[i], id + i, 0);
+		if (fa) {
+			if (!readahead)
+				xfs_buf_verifier_error(bp, -EFSCORRUPTED,
+					__func__, &dqb[i],
+					sizeof(struct xfs_dqblk), fa);
 			return fa;
+		}
 	}
 
 	return NULL;
@@ -214,7 +240,7 @@ xfs_dquot_buf_verify_struct(
 {
 	struct xfs_mount	*mp = bp->b_target->bt_mount;
 
-	return xfs_dquot_buf_verify(mp, bp);
+	return xfs_dquot_buf_verify(mp, bp, false);
 }
 
 static void
@@ -222,15 +248,10 @@ xfs_dquot_buf_read_verify(
 	struct xfs_buf		*bp)
 {
 	struct xfs_mount	*mp = bp->b_target->bt_mount;
-	xfs_failaddr_t		fa;
 
-	if (!xfs_dquot_buf_verify_crc(mp, bp))
-		xfs_verifier_error(bp, -EFSBADCRC, __this_address);
-	else {
-		fa = xfs_dquot_buf_verify(mp, bp);
-		if (fa)
-			xfs_verifier_error(bp, -EFSCORRUPTED, __this_address);
-	}
+	if (!xfs_dquot_buf_verify_crc(mp, bp, false))
+		return;
+	xfs_dquot_buf_verify(mp, bp, false);
 }
 
 /*
@@ -245,8 +266,8 @@ xfs_dquot_buf_readahead_verify(
 {
 	struct xfs_mount	*mp = bp->b_target->bt_mount;
 
-	if (!xfs_dquot_buf_verify_crc(mp, bp) ||
-	    xfs_dquot_buf_verify(mp, bp) != NULL) {
+	if (!xfs_dquot_buf_verify_crc(mp, bp, true) ||
+	    xfs_dquot_buf_verify(mp, bp, true) != NULL) {
 		xfs_buf_ioerror(bp, -EIO);
 		bp->b_flags &= ~XBF_DONE;
 	}
@@ -262,11 +283,8 @@ xfs_dquot_buf_write_verify(
 	struct xfs_buf		*bp)
 {
 	struct xfs_mount	*mp = bp->b_target->bt_mount;
-	xfs_failaddr_t		fa;
 
-	fa = xfs_dquot_buf_verify(mp, bp);
-	if (fa)
-		xfs_verifier_error(bp, -EFSCORRUPTED, __this_address);
+	xfs_dquot_buf_verify(mp, bp, false);
 }
 
 const struct xfs_buf_ops xfs_dquot_buf_ops = {
diff --git a/fs/xfs/libxfs/xfs_errortag.h b/fs/xfs/libxfs/xfs_errortag.h
index bc1789d95152..d47b91625945 100644
--- a/fs/xfs/libxfs/xfs_errortag.h
+++ b/fs/xfs/libxfs/xfs_errortag.h
@@ -65,7 +65,8 @@
 #define XFS_ERRTAG_LOG_BAD_CRC				29
 #define XFS_ERRTAG_LOG_ITEM_PIN				30
 #define XFS_ERRTAG_BUF_LRU_REF				31
-#define XFS_ERRTAG_MAX					32
+#define XFS_ERRTAG_FORCE_SCRUB_REPAIR			32
+#define XFS_ERRTAG_MAX					33
 
 /*
  * Random factors for above tags, 1 means always, 2 means 1/2 time, etc.
@@ -102,5 +103,6 @@
 #define XFS_RANDOM_LOG_BAD_CRC				1
 #define XFS_RANDOM_LOG_ITEM_PIN				1
 #define XFS_RANDOM_BUF_LRU_REF				2
+#define XFS_RANDOM_FORCE_SCRUB_REPAIR			1
 
 #endif /* __XFS_ERRORTAG_H_ */
diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index 42956d8d95ed..c1cb29a5f4f6 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -98,6 +98,9 @@ struct xfs_ifork;
 	 XFS_SB_VERSION2_PROJID32BIT	| \
 	 XFS_SB_VERSION2_FTYPE)
 
+/* Maximum size of the xfs filesystem label, no terminating NULL */
+#define XFSLABEL_MAX			12
+
 /*
  * Superblock - in core version.  Must match the ondisk version below.
  * Must be padded to 64 bit alignment.
@@ -122,7 +125,7 @@ typedef struct xfs_sb {
 	uint16_t	sb_sectsize;	/* volume sector size, bytes */
 	uint16_t	sb_inodesize;	/* inode size, bytes */
 	uint16_t	sb_inopblock;	/* inodes per block */
-	char		sb_fname[12];	/* file system name */
+	char		sb_fname[XFSLABEL_MAX]; /* file system name */
 	uint8_t		sb_blocklog;	/* log2 of sb_blocksize */
 	uint8_t		sb_sectlog;	/* log2 of sb_sectsize */
 	uint8_t		sb_inodelog;	/* log2 of sb_inodesize */
@@ -213,7 +216,7 @@ typedef struct xfs_dsb {
 	__be16		sb_sectsize;	/* volume sector size, bytes */
 	__be16		sb_inodesize;	/* inode size, bytes */
 	__be16		sb_inopblock;	/* inodes per block */
-	char		sb_fname[12];	/* file system name */
+	char		sb_fname[XFSLABEL_MAX]; /* file system name */
 	__u8		sb_blocklog;	/* log2 of sb_blocksize */
 	__u8		sb_sectlog;	/* log2 of sb_sectsize */
 	__u8		sb_inodelog;	/* log2 of sb_inodesize */
diff --git a/fs/xfs/libxfs/xfs_fs.h b/fs/xfs/libxfs/xfs_fs.h
index faf1a4edd618..dddc75e4f1f6 100644
--- a/fs/xfs/libxfs/xfs_fs.h
+++ b/fs/xfs/libxfs/xfs_fs.h
@@ -542,13 +542,20 @@ struct xfs_scrub_metadata {
 /* o: Metadata object looked funny but isn't corrupt. */
 #define XFS_SCRUB_OFLAG_WARNING		(1 << 6)
 
+/*
+ * o: IFLAG_REPAIR was set but metadata object did not need fixing or
+ *    optimization and has therefore not been altered.
+ */
+#define XFS_SCRUB_OFLAG_NO_REPAIR_NEEDED (1 << 7)
+
 #define XFS_SCRUB_FLAGS_IN	(XFS_SCRUB_IFLAG_REPAIR)
 #define XFS_SCRUB_FLAGS_OUT	(XFS_SCRUB_OFLAG_CORRUPT | \
 				 XFS_SCRUB_OFLAG_PREEN | \
 				 XFS_SCRUB_OFLAG_XFAIL | \
 				 XFS_SCRUB_OFLAG_XCORRUPT | \
 				 XFS_SCRUB_OFLAG_INCOMPLETE | \
-				 XFS_SCRUB_OFLAG_WARNING)
+				 XFS_SCRUB_OFLAG_WARNING | \
+				 XFS_SCRUB_OFLAG_NO_REPAIR_NEEDED)
 #define XFS_SCRUB_FLAGS_ALL	(XFS_SCRUB_FLAGS_IN | XFS_SCRUB_FLAGS_OUT)
 
 /*
diff --git a/fs/xfs/libxfs/xfs_ialloc.c b/fs/xfs/libxfs/xfs_ialloc.c
index de627fa19168..4ca4ff7a757d 100644
--- a/fs/xfs/libxfs/xfs_ialloc.c
+++ b/fs/xfs/libxfs/xfs_ialloc.c
@@ -148,7 +148,7 @@ xfs_inobt_get_rec(
 /*
  * Insert a single inobt record. Cursor must already point to desired location.
  */
-STATIC int
+int
 xfs_inobt_insert_rec(
 	struct xfs_btree_cur	*cur,
 	uint16_t		holemask,
diff --git a/fs/xfs/libxfs/xfs_ialloc.h b/fs/xfs/libxfs/xfs_ialloc.h
index c5402bb4ce0c..77fffced8bac 100644
--- a/fs/xfs/libxfs/xfs_ialloc.h
+++ b/fs/xfs/libxfs/xfs_ialloc.h
@@ -176,6 +176,9 @@ int xfs_ialloc_has_inode_record(struct xfs_btree_cur *cur, xfs_agino_t low,
 		xfs_agino_t high, bool *exists);
 int xfs_ialloc_count_inodes(struct xfs_btree_cur *cur, xfs_agino_t *count,
 		xfs_agino_t *freecount);
+int xfs_inobt_insert_rec(struct xfs_btree_cur *cur, uint16_t holemask,
+		uint8_t count, int32_t freecount, xfs_inofree_t free,
+		int *stat);
 
 int xfs_ialloc_cluster_alignment(struct xfs_mount *mp);
 void xfs_ialloc_agino_range(struct xfs_mount *mp, xfs_agnumber_t agno,
diff --git a/fs/xfs/libxfs/xfs_ialloc_btree.c b/fs/xfs/libxfs/xfs_ialloc_btree.c
index 367e9a0726e6..b04c55512159 100644
--- a/fs/xfs/libxfs/xfs_ialloc_btree.c
+++ b/fs/xfs/libxfs/xfs_ialloc_btree.c
@@ -296,7 +296,7 @@ xfs_inobt_verify(
 	case cpu_to_be32(XFS_FIBT_MAGIC):
 		break;
 	default:
-		return NULL;
+		return __this_address;
 	}
 
 	/* level verification */
@@ -608,3 +608,12 @@ xfs_finobt_calc_reserves(
 	*used += tree_len;
 	return 0;
 }
+
+/* Calculate the inobt btree size for some records. */
+xfs_extlen_t
+xfs_iallocbt_calc_size(
+	struct xfs_mount	*mp,
+	unsigned long long	len)
+{
+	return xfs_btree_calc_size(mp->m_inobt_mnr, len);
+}
diff --git a/fs/xfs/libxfs/xfs_ialloc_btree.h b/fs/xfs/libxfs/xfs_ialloc_btree.h
index aa81e2e63f3f..4acdd5458d59 100644
--- a/fs/xfs/libxfs/xfs_ialloc_btree.h
+++ b/fs/xfs/libxfs/xfs_ialloc_btree.h
@@ -74,5 +74,7 @@ int xfs_inobt_rec_check_count(struct xfs_mount *,
 
 int xfs_finobt_calc_reserves(struct xfs_mount *mp, xfs_agnumber_t agno,
 		xfs_extlen_t *ask, xfs_extlen_t *used);
+extern xfs_extlen_t xfs_iallocbt_calc_size(struct xfs_mount *mp,
+		unsigned long long len);
 
 #endif	/* __XFS_IALLOC_BTREE_H__ */
diff --git a/fs/xfs/libxfs/xfs_quota_defs.h b/fs/xfs/libxfs/xfs_quota_defs.h
index bb1b13a9b5f4..d4af2804b178 100644
--- a/fs/xfs/libxfs/xfs_quota_defs.h
+++ b/fs/xfs/libxfs/xfs_quota_defs.h
@@ -107,14 +107,12 @@ typedef uint16_t	xfs_qwarncnt_t;
  * to a single function. None of these XFS_QMOPT_* flags are meant to have
  * persistent values (ie. their values can and will change between versions)
  */
-#define XFS_QMOPT_DQALLOC	0x0000002 /* alloc dquot ondisk if needed */
 #define XFS_QMOPT_UQUOTA	0x0000004 /* user dquot requested */
 #define XFS_QMOPT_PQUOTA	0x0000008 /* project dquot requested */
 #define XFS_QMOPT_FORCE_RES	0x0000010 /* ignore quota limits */
 #define XFS_QMOPT_SBVERSION	0x0000040 /* change superblock version num */
 #define XFS_QMOPT_GQUOTA	0x0002000 /* group dquot requested */
 #define XFS_QMOPT_ENOSPC	0x0004000 /* enospc instead of edquot (prj) */
-#define XFS_QMOPT_DQNEXT	0x0008000 /* return next dquot >= this ID */
 
 /*
  * flags to xfs_trans_mod_dquot to indicate which field needs to be
@@ -152,10 +150,11 @@ typedef uint16_t	xfs_qwarncnt_t;
 #define XFS_QMOPT_RESBLK_MASK	(XFS_QMOPT_RES_REGBLKS | XFS_QMOPT_RES_RTBLKS)
 
 extern xfs_failaddr_t xfs_dquot_verify(struct xfs_mount *mp,
-		struct xfs_disk_dquot *ddq, xfs_dqid_t id, uint type,
-		uint flags);
+		struct xfs_disk_dquot *ddq, xfs_dqid_t id, uint type);
+extern xfs_failaddr_t xfs_dqblk_verify(struct xfs_mount *mp,
+		struct xfs_dqblk *dqb, xfs_dqid_t id, uint type);
 extern int xfs_calc_dquots_per_chunk(unsigned int nbblks);
-extern int xfs_dquot_repair(struct xfs_mount *mp, struct xfs_disk_dquot *ddq,
+extern int xfs_dqblk_repair(struct xfs_mount *mp, struct xfs_dqblk *dqb,
 		xfs_dqid_t id, uint type);
 
 #endif	/* __XFS_QUOTA_H__ */
diff --git a/fs/xfs/libxfs/xfs_refcount.c b/fs/xfs/libxfs/xfs_refcount.c
index 560e28473024..418d53295893 100644
--- a/fs/xfs/libxfs/xfs_refcount.c
+++ b/fs/xfs/libxfs/xfs_refcount.c
@@ -88,8 +88,25 @@ xfs_refcount_lookup_ge(
 	return xfs_btree_lookup(cur, XFS_LOOKUP_GE, stat);
 }
 
+/*
+ * Look up the first record equal to [bno, len] in the btree
+ * given by cur.
+ */
+int
+xfs_refcount_lookup_eq(
+	struct xfs_btree_cur	*cur,
+	xfs_agblock_t		bno,
+	int			*stat)
+{
+	trace_xfs_refcount_lookup(cur->bc_mp, cur->bc_private.a.agno, bno,
+			XFS_LOOKUP_LE);
+	cur->bc_rec.rc.rc_startblock = bno;
+	cur->bc_rec.rc.rc_blockcount = 0;
+	return xfs_btree_lookup(cur, XFS_LOOKUP_EQ, stat);
+}
+
 /* Convert on-disk record to in-core format. */
-static inline void
+void
 xfs_refcount_btrec_to_irec(
 	union xfs_btree_rec		*rec,
 	struct xfs_refcount_irec	*irec)
@@ -149,7 +166,7 @@ xfs_refcount_update(
  * by [bno, len, refcount].
  * This either works (return 0) or gets an EFSCORRUPTED error.
  */
-STATIC int
+int
 xfs_refcount_insert(
 	struct xfs_btree_cur		*cur,
 	struct xfs_refcount_irec	*irec,
@@ -162,7 +179,10 @@ xfs_refcount_insert(
 	cur->bc_rec.rc.rc_blockcount = irec->rc_blockcount;
 	cur->bc_rec.rc.rc_refcount = irec->rc_refcount;
 	error = xfs_btree_insert(cur, i);
+	if (error)
+		goto out_error;
 	XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, *i == 1, out_error);
+
 out_error:
 	if (error)
 		trace_xfs_refcount_insert_error(cur->bc_mp,
diff --git a/fs/xfs/libxfs/xfs_refcount.h b/fs/xfs/libxfs/xfs_refcount.h
index 2a731ac68fe4..a92ad9078bc1 100644
--- a/fs/xfs/libxfs/xfs_refcount.h
+++ b/fs/xfs/libxfs/xfs_refcount.h
@@ -24,6 +24,8 @@ extern int xfs_refcount_lookup_le(struct xfs_btree_cur *cur,
 		xfs_agblock_t bno, int *stat);
 extern int xfs_refcount_lookup_ge(struct xfs_btree_cur *cur,
 		xfs_agblock_t bno, int *stat);
+extern int xfs_refcount_lookup_eq(struct xfs_btree_cur *cur,
+		xfs_agblock_t bno, int *stat);
 extern int xfs_refcount_get_rec(struct xfs_btree_cur *cur,
 		struct xfs_refcount_irec *irec, int *stat);
 
@@ -85,5 +87,10 @@ static inline xfs_fileoff_t xfs_refcount_max_unmap(int log_res)
 
 extern int xfs_refcount_has_record(struct xfs_btree_cur *cur,
 		xfs_agblock_t bno, xfs_extlen_t len, bool *exists);
+union xfs_btree_rec;
+extern void xfs_refcount_btrec_to_irec(union xfs_btree_rec *rec,
+		struct xfs_refcount_irec *irec);
+extern int xfs_refcount_insert(struct xfs_btree_cur *cur,
+		struct xfs_refcount_irec *irec, int *stat);
 
 #endif	/* __XFS_REFCOUNT_H__ */
diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c
index fba8d2718017..c0644f1be8a8 100644
--- a/fs/xfs/libxfs/xfs_rmap.c
+++ b/fs/xfs/libxfs/xfs_rmap.c
@@ -1374,6 +1374,8 @@ xfs_rmap_convert_shared(
 	 */
 	error = xfs_rmap_lookup_le_range(cur, bno, owner, offset, flags,
 			&PREV, &i);
+	if (error)
+		goto done;
 	XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
 
 	ASSERT(PREV.rm_offset <= offset);
@@ -2030,6 +2032,34 @@ out_error:
 	return error;
 }
 
+/* Insert a raw rmap into the rmapbt. */
+int
+xfs_rmap_map_raw(
+	struct xfs_btree_cur	*cur,
+	struct xfs_rmap_irec	*rmap)
+{
+	struct xfs_owner_info	oinfo;
+
+	oinfo.oi_owner = rmap->rm_owner;
+	oinfo.oi_offset = rmap->rm_offset;
+	oinfo.oi_flags = 0;
+	if (rmap->rm_flags & XFS_RMAP_ATTR_FORK)
+		oinfo.oi_flags |= XFS_OWNER_INFO_ATTR_FORK;
+	if (rmap->rm_flags & XFS_RMAP_BMBT_BLOCK)
+		oinfo.oi_flags |= XFS_OWNER_INFO_BMBT_BLOCK;
+
+	if (rmap->rm_flags || XFS_RMAP_NON_INODE_OWNER(rmap->rm_owner))
+		return xfs_rmap_map(cur, rmap->rm_startblock,
+				rmap->rm_blockcount,
+				rmap->rm_flags & XFS_RMAP_UNWRITTEN,
+				&oinfo);
+
+	return xfs_rmap_map_shared(cur, rmap->rm_startblock,
+			rmap->rm_blockcount,
+			rmap->rm_flags & XFS_RMAP_UNWRITTEN,
+			&oinfo);
+}
+
 struct xfs_rmap_query_range_info {
 	xfs_rmap_query_range_fn	fn;
 	void				*priv;
@@ -2453,3 +2483,56 @@ xfs_rmap_record_exists(
 		     irec.rm_startblock + irec.rm_blockcount >= bno + len);
 	return 0;
 }
+
+struct xfs_rmap_key_state {
+	uint64_t			owner;
+	uint64_t			offset;
+	unsigned int			flags;
+	bool				has_rmap;
+};
+
+/* For each rmap given, figure out if it doesn't match the key we want. */
+STATIC int
+xfs_rmap_has_other_keys_helper(
+	struct xfs_btree_cur		*cur,
+	struct xfs_rmap_irec		*rec,
+	void				*priv)
+{
+	struct xfs_rmap_key_state	*rks = priv;
+
+	if (rks->owner == rec->rm_owner && rks->offset == rec->rm_offset &&
+	    ((rks->flags & rec->rm_flags) & XFS_RMAP_KEY_FLAGS) == rks->flags)
+		return 0;
+	rks->has_rmap = true;
+	return XFS_BTREE_QUERY_RANGE_ABORT;
+}
+
+/*
+ * Given an extent and some owner info, can we find records overlapping
+ * the extent whose owner info does not match the given owner?
+ */
+int
+xfs_rmap_has_other_keys(
+	struct xfs_btree_cur		*cur,
+	xfs_agblock_t			bno,
+	xfs_extlen_t			len,
+	struct xfs_owner_info		*oinfo,
+	bool				*has_rmap)
+{
+	struct xfs_rmap_irec		low = {0};
+	struct xfs_rmap_irec		high;
+	struct xfs_rmap_key_state	rks;
+	int				error;
+
+	xfs_owner_info_unpack(oinfo, &rks.owner, &rks.offset, &rks.flags);
+	rks.has_rmap = false;
+
+	low.rm_startblock = bno;
+	memset(&high, 0xFF, sizeof(high));
+	high.rm_startblock = bno + len - 1;
+
+	error = xfs_rmap_query_range(cur, &low, &high,
+			xfs_rmap_has_other_keys_helper, &rks);
+	*has_rmap = rks.has_rmap;
+	return error;
+}
diff --git a/fs/xfs/libxfs/xfs_rmap.h b/fs/xfs/libxfs/xfs_rmap.h
index 380e53be98d5..43e506f67680 100644
--- a/fs/xfs/libxfs/xfs_rmap.h
+++ b/fs/xfs/libxfs/xfs_rmap.h
@@ -238,5 +238,9 @@ int xfs_rmap_has_record(struct xfs_btree_cur *cur, xfs_agblock_t bno,
 int xfs_rmap_record_exists(struct xfs_btree_cur *cur, xfs_agblock_t bno,
 		xfs_extlen_t len, struct xfs_owner_info *oinfo,
 		bool *has_rmap);
+int xfs_rmap_has_other_keys(struct xfs_btree_cur *cur, xfs_agblock_t bno,
+		xfs_extlen_t len, struct xfs_owner_info *oinfo,
+		bool *has_rmap);
+int xfs_rmap_map_raw(struct xfs_btree_cur *cur, struct xfs_rmap_irec *rmap);
 
 #endif	/* __XFS_RMAP_H__ */
diff --git a/fs/xfs/libxfs/xfs_rtbitmap.c b/fs/xfs/libxfs/xfs_rtbitmap.c
index 106be2d0bb88..369eeb7a52ec 100644
--- a/fs/xfs/libxfs/xfs_rtbitmap.c
+++ b/fs/xfs/libxfs/xfs_rtbitmap.c
@@ -90,6 +90,9 @@ xfs_rtbuf_get(
 	if (error)
 		return error;
 
+	if (nmap == 0 || !xfs_bmap_is_real_extent(&map))
+		return -EFSCORRUPTED;
+
 	ASSERT(map.br_startblock != NULLFSBLOCK);
 	error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp,
 				   XFS_FSB_TO_DADDR(mp, map.br_startblock),
@@ -1033,14 +1036,17 @@ xfs_rtalloc_query_range(
 	int				is_free;
 	int				error = 0;
 
-	if (low_rec->ar_startblock > high_rec->ar_startblock)
+	if (low_rec->ar_startext > high_rec->ar_startext)
 		return -EINVAL;
-	else if (low_rec->ar_startblock == high_rec->ar_startblock)
+	if (low_rec->ar_startext >= mp->m_sb.sb_rextents ||
+	    low_rec->ar_startext == high_rec->ar_startext)
 		return 0;
+	if (high_rec->ar_startext >= mp->m_sb.sb_rextents)
+		high_rec->ar_startext = mp->m_sb.sb_rextents - 1;
 
 	/* Iterate the bitmap, looking for discrepancies. */
-	rtstart = low_rec->ar_startblock;
-	rem = high_rec->ar_startblock - rtstart;
+	rtstart = low_rec->ar_startext;
+	rem = high_rec->ar_startext - rtstart;
 	while (rem) {
 		/* Is the first block free? */
 		error = xfs_rtcheck_range(mp, tp, rtstart, 1, 1, &rtend,
@@ -1050,13 +1056,13 @@ xfs_rtalloc_query_range(
 
 		/* How long does the extent go for? */
 		error = xfs_rtfind_forw(mp, tp, rtstart,
-				high_rec->ar_startblock - 1, &rtend);
+				high_rec->ar_startext - 1, &rtend);
 		if (error)
 			break;
 
 		if (is_free) {
-			rec.ar_startblock = rtstart;
-			rec.ar_blockcount = rtend - rtstart + 1;
+			rec.ar_startext = rtstart;
+			rec.ar_extcount = rtend - rtstart + 1;
 
 			error = fn(tp, &rec, priv);
 			if (error)
@@ -1079,9 +1085,9 @@ xfs_rtalloc_query_all(
 {
 	struct xfs_rtalloc_rec		keys[2];
 
-	keys[0].ar_startblock = 0;
-	keys[1].ar_startblock = tp->t_mountp->m_sb.sb_rblocks;
-	keys[0].ar_blockcount = keys[1].ar_blockcount = 0;
+	keys[0].ar_startext = 0;
+	keys[1].ar_startext = tp->t_mountp->m_sb.sb_rextents - 1;
+	keys[0].ar_extcount = keys[1].ar_extcount = 0;
 
 	return xfs_rtalloc_query_range(tp, &keys[0], &keys[1], fn, priv);
 }
diff --git a/fs/xfs/libxfs/xfs_sb.c b/fs/xfs/libxfs/xfs_sb.c
index d9b94bd5f689..d485e14313c6 100644
--- a/fs/xfs/libxfs/xfs_sb.c
+++ b/fs/xfs/libxfs/xfs_sb.c
@@ -888,6 +888,109 @@ xfs_sync_sb(
 	return xfs_trans_commit(tp);
 }
 
+/*
+ * Update all the secondary superblocks to match the new state of the primary.
+ * Because we are completely overwriting all the existing fields in the
+ * secondary superblock buffers, there is no need to read them in from disk.
+ * Just get a new buffer, stamp it and write it.
+ *
+ * The sb buffers need to be cached here so that we serialise against other
+ * operations that access the secondary superblocks, but we don't want to keep
+ * them in memory once it is written so we mark it as a one-shot buffer.
+ */
+int
+xfs_update_secondary_sbs(
+	struct xfs_mount	*mp)
+{
+	xfs_agnumber_t		agno;
+	int			saved_error = 0;
+	int			error = 0;
+	LIST_HEAD		(buffer_list);
+
+	/* update secondary superblocks. */
+	for (agno = 1; agno < mp->m_sb.sb_agcount; agno++) {
+		struct xfs_buf		*bp;
+
+		bp = xfs_buf_get(mp->m_ddev_targp,
+				 XFS_AG_DADDR(mp, agno, XFS_SB_DADDR),
+				 XFS_FSS_TO_BB(mp, 1), 0);
+		/*
+		 * If we get an error reading or writing alternate superblocks,
+		 * continue.  xfs_repair chooses the "best" superblock based
+		 * on most matches; if we break early, we'll leave more
+		 * superblocks un-updated than updated, and xfs_repair may
+		 * pick them over the properly-updated primary.
+		 */
+		if (!bp) {
+			xfs_warn(mp,
+		"error allocating secondary superblock for ag %d",
+				agno);
+			if (!saved_error)
+				saved_error = -ENOMEM;
+			continue;
+		}
+
+		bp->b_ops = &xfs_sb_buf_ops;
+		xfs_buf_oneshot(bp);
+		xfs_buf_zero(bp, 0, BBTOB(bp->b_length));
+		xfs_sb_to_disk(XFS_BUF_TO_SBP(bp), &mp->m_sb);
+		xfs_buf_delwri_queue(bp, &buffer_list);
+		xfs_buf_relse(bp);
+
+		/* don't hold too many buffers at once */
+		if (agno % 16)
+			continue;
+
+		error = xfs_buf_delwri_submit(&buffer_list);
+		if (error) {
+			xfs_warn(mp,
+		"write error %d updating a secondary superblock near ag %d",
+				error, agno);
+			if (!saved_error)
+				saved_error = error;
+			continue;
+		}
+	}
+	error = xfs_buf_delwri_submit(&buffer_list);
+	if (error) {
+		xfs_warn(mp,
+		"write error %d updating a secondary superblock near ag %d",
+			error, agno);
+	}
+
+	return saved_error ? saved_error : error;
+}
+
+/*
+ * Same behavior as xfs_sync_sb, except that it is always synchronous and it
+ * also writes the superblock buffer to disk sector 0 immediately.
+ */
+int
+xfs_sync_sb_buf(
+	struct xfs_mount	*mp)
+{
+	struct xfs_trans	*tp;
+	int			error;
+
+	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_sb, 0, 0, 0, &tp);
+	if (error)
+		return error;
+
+	xfs_log_sb(tp);
+	xfs_trans_bhold(tp, mp->m_sb_bp);
+	xfs_trans_set_sync(tp);
+	error = xfs_trans_commit(tp);
+	if (error)
+		goto out;
+	/*
+	 * write out the sb buffer to get the changes to disk
+	 */
+	error = xfs_bwrite(mp->m_sb_bp);
+out:
+	xfs_buf_relse(mp->m_sb_bp);
+	return error;
+}
+
 int
 xfs_fs_geometry(
 	struct xfs_sb		*sbp,
@@ -972,3 +1075,47 @@ xfs_fs_geometry(
 
 	return 0;
 }
+
+/* Read a secondary superblock. */
+int
+xfs_sb_read_secondary(
+	struct xfs_mount	*mp,
+	struct xfs_trans	*tp,
+	xfs_agnumber_t		agno,
+	struct xfs_buf		**bpp)
+{
+	struct xfs_buf		*bp;
+	int			error;
+
+	ASSERT(agno != 0 && agno != NULLAGNUMBER);
+	error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp,
+			XFS_AG_DADDR(mp, agno, XFS_SB_BLOCK(mp)),
+			XFS_FSS_TO_BB(mp, 1), 0, &bp, &xfs_sb_buf_ops);
+	if (error)
+		return error;
+	xfs_buf_set_ref(bp, XFS_SSB_REF);
+	*bpp = bp;
+	return 0;
+}
+
+/* Get an uninitialised secondary superblock buffer. */
+int
+xfs_sb_get_secondary(
+	struct xfs_mount	*mp,
+	struct xfs_trans	*tp,
+	xfs_agnumber_t		agno,
+	struct xfs_buf		**bpp)
+{
+	struct xfs_buf		*bp;
+
+	ASSERT(agno != 0 && agno != NULLAGNUMBER);
+	bp = xfs_trans_get_buf(tp, mp->m_ddev_targp,
+			XFS_AG_DADDR(mp, agno, XFS_SB_BLOCK(mp)),
+			XFS_FSS_TO_BB(mp, 1), 0);
+	if (!bp)
+		return -ENOMEM;
+	bp->b_ops = &xfs_sb_buf_ops;
+	xfs_buf_oneshot(bp);
+	*bpp = bp;
+	return 0;
+}
diff --git a/fs/xfs/libxfs/xfs_sb.h b/fs/xfs/libxfs/xfs_sb.h
index 63dcd2a1a657..244e0162c49e 100644
--- a/fs/xfs/libxfs/xfs_sb.h
+++ b/fs/xfs/libxfs/xfs_sb.h
@@ -18,6 +18,13 @@
 #ifndef __XFS_SB_H__
 #define	__XFS_SB_H__
 
+struct xfs_mount;
+struct xfs_sb;
+struct xfs_dsb;
+struct xfs_trans;
+struct xfs_fsop_geom;
+struct xfs_perag;
+
 /*
  * perag get/put wrappers for ref counting
  */
@@ -29,13 +36,22 @@ extern int	xfs_initialize_perag_data(struct xfs_mount *, xfs_agnumber_t);
 
 extern void	xfs_log_sb(struct xfs_trans *tp);
 extern int	xfs_sync_sb(struct xfs_mount *mp, bool wait);
+extern int	xfs_sync_sb_buf(struct xfs_mount *mp);
 extern void	xfs_sb_mount_common(struct xfs_mount *mp, struct xfs_sb *sbp);
 extern void	xfs_sb_from_disk(struct xfs_sb *to, struct xfs_dsb *from);
 extern void	xfs_sb_to_disk(struct xfs_dsb *to, struct xfs_sb *from);
 extern void	xfs_sb_quota_from_disk(struct xfs_sb *sbp);
 
+extern int	xfs_update_secondary_sbs(struct xfs_mount *mp);
+
 #define XFS_FS_GEOM_MAX_STRUCT_VER	(4)
 extern int	xfs_fs_geometry(struct xfs_sb *sbp, struct xfs_fsop_geom *geo,
 				int struct_version);
+extern int	xfs_sb_read_secondary(struct xfs_mount *mp,
+				struct xfs_trans *tp, xfs_agnumber_t agno,
+				struct xfs_buf **bpp);
+extern int	xfs_sb_get_secondary(struct xfs_mount *mp,
+				struct xfs_trans *tp, xfs_agnumber_t agno,
+				struct xfs_buf **bpp);
 
 #endif	/* __XFS_SB_H__ */
diff --git a/fs/xfs/libxfs/xfs_shared.h b/fs/xfs/libxfs/xfs_shared.h
index d0b84da0cb1e..ae99c260adb1 100644
--- a/fs/xfs/libxfs/xfs_shared.h
+++ b/fs/xfs/libxfs/xfs_shared.h
@@ -57,21 +57,6 @@ extern const struct xfs_buf_ops xfs_sb_quiet_buf_ops;
 extern const struct xfs_buf_ops xfs_symlink_buf_ops;
 extern const struct xfs_buf_ops xfs_rtbuf_ops;
 
-/*
- * This structure is used to track log items associated with
- * a transaction.  It points to the log item and keeps some
- * flags to track the state of the log item.  It also tracks
- * the amount of space needed to log the item it describes
- * once we get to commit processing (see xfs_trans_commit()).
- */
-struct xfs_log_item_desc {
-	struct xfs_log_item	*lid_item;
-	struct list_head	lid_trans;
-	unsigned char		lid_flags;
-};
-
-#define XFS_LID_DIRTY		0x1
-
 /* 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 *);
@@ -127,6 +112,7 @@ void	xfs_log_get_max_trans_res(struct xfs_mount *mp,
 #define	XFS_ATTR_BTREE_REF	1
 #define	XFS_DQUOT_REF		1
 #define	XFS_REFC_BTREE_REF	1
+#define	XFS_SSB_REF		0
 
 /*
  * Flags for xfs_trans_ichgtime().
diff --git a/fs/xfs/libxfs/xfs_types.h b/fs/xfs/libxfs/xfs_types.h
index 3c560695c546..ea18449bd732 100644
--- a/fs/xfs/libxfs/xfs_types.h
+++ b/fs/xfs/libxfs/xfs_types.h
@@ -30,7 +30,7 @@ typedef int64_t		xfs_fsize_t;	/* bytes in a file */
 typedef uint64_t	xfs_ufsize_t;	/* unsigned bytes in a file */
 
 typedef int32_t		xfs_suminfo_t;	/* type of bitmap summary info */
-typedef int32_t		xfs_rtword_t;	/* word type for bitmap manipulations */
+typedef uint32_t	xfs_rtword_t;	/* word type for bitmap manipulations */
 
 typedef int64_t		xfs_lsn_t;	/* log sequence number */
 typedef int32_t		xfs_tid_t;	/* transaction identifier */
diff --git a/fs/xfs/scrub/agheader.c b/fs/xfs/scrub/agheader.c
index 018aabbd9394..1f71793f7db4 100644
--- a/fs/xfs/scrub/agheader.c
+++ b/fs/xfs/scrub/agheader.c
@@ -38,68 +38,6 @@
 #include "scrub/common.h"
 #include "scrub/trace.h"
 
-/*
- * Walk all the blocks in the AGFL.  The fn function can return any negative
- * error code or XFS_BTREE_QUERY_RANGE_ABORT.
- */
-int
-xfs_scrub_walk_agfl(
-	struct xfs_scrub_context	*sc,
-	int				(*fn)(struct xfs_scrub_context *,
-					      xfs_agblock_t bno, void *),
-	void				*priv)
-{
-	struct xfs_agf			*agf;
-	__be32				*agfl_bno;
-	struct xfs_mount		*mp = sc->mp;
-	unsigned int			flfirst;
-	unsigned int			fllast;
-	int				i;
-	int				error;
-
-	agf = XFS_BUF_TO_AGF(sc->sa.agf_bp);
-	agfl_bno = XFS_BUF_TO_AGFL_BNO(mp, sc->sa.agfl_bp);
-	flfirst = be32_to_cpu(agf->agf_flfirst);
-	fllast = be32_to_cpu(agf->agf_fllast);
-
-	/* Nothing to walk in an empty AGFL. */
-	if (agf->agf_flcount == cpu_to_be32(0))
-		return 0;
-
-	/* first to last is a consecutive list. */
-	if (fllast >= flfirst) {
-		for (i = flfirst; i <= fllast; i++) {
-			error = fn(sc, be32_to_cpu(agfl_bno[i]), priv);
-			if (error)
-				return error;
-			if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
-				return error;
-		}
-
-		return 0;
-	}
-
-	/* first to the end */
-	for (i = flfirst; i < xfs_agfl_size(mp); i++) {
-		error = fn(sc, be32_to_cpu(agfl_bno[i]), priv);
-		if (error)
-			return error;
-		if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
-			return error;
-	}
-
-	/* the start to last. */
-	for (i = 0; i <= fllast; i++) {
-		error = fn(sc, be32_to_cpu(agfl_bno[i]), priv);
-		if (error)
-			return error;
-		if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
-			return error;
-	}
-
-	return 0;
-}
-
 /* Superblock */
 
 /* Cross-reference with the other btrees. */
@@ -157,9 +95,7 @@ xfs_scrub_superblock(
 	if (agno == 0)
 		return 0;
 
-	error = xfs_trans_read_buf(mp, sc->tp, mp->m_ddev_targp,
-		  XFS_AGB_TO_DADDR(mp, agno, XFS_SB_BLOCK(mp)),
-		  XFS_FSS_TO_BB(mp, 1), 0, &bp, &xfs_sb_buf_ops);
+	error = xfs_sb_read_secondary(mp, sc->tp, agno, &bp);
 	/*
 	 * The superblock verifier can return several different error codes
 	 * if it thinks the superblock doesn't look right.  For a mount these
@@ -680,6 +616,7 @@ struct xfs_scrub_agfl_info {
 	unsigned int			sz_entries;
 	unsigned int			nr_entries;
 	xfs_agblock_t			*entries;
+	struct xfs_scrub_context	*sc;
 };
 
 /* Cross-reference with the other btrees. */
@@ -701,12 +638,12 @@ xfs_scrub_agfl_block_xref(
 /* Scrub an AGFL block. */
 STATIC int
 xfs_scrub_agfl_block(
-	struct xfs_scrub_context	*sc,
+	struct xfs_mount		*mp,
 	xfs_agblock_t			agbno,
 	void				*priv)
 {
-	struct xfs_mount		*mp = sc->mp;
 	struct xfs_scrub_agfl_info	*sai = priv;
+	struct xfs_scrub_context	*sc = sai->sc;
 	xfs_agnumber_t			agno = sc->sa.agno;
 
 	if (xfs_verify_agbno(mp, agno, agbno) &&
@@ -717,6 +654,9 @@ xfs_scrub_agfl_block(
 
 	xfs_scrub_agfl_block_xref(sc, agbno, priv);
 
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return XFS_BTREE_QUERY_RANGE_ABORT;
+
 	return 0;
 }
 
@@ -796,8 +736,10 @@ xfs_scrub_agfl(
 		goto out;
 	}
 	memset(&sai, 0, sizeof(sai));
+	sai.sc = sc;
 	sai.sz_entries = agflcount;
-	sai.entries = kmem_zalloc(sizeof(xfs_agblock_t) * agflcount, KM_NOFS);
+	sai.entries = kmem_zalloc(sizeof(xfs_agblock_t) * agflcount,
+			KM_MAYFAIL);
 	if (!sai.entries) {
 		error = -ENOMEM;
 		goto out;
@@ -805,7 +747,12 @@ xfs_scrub_agfl(
 
 	/* Check the blocks in the AGFL. */
 	xfs_rmap_ag_owner(&sai.oinfo, XFS_RMAP_OWN_AG);
-	error = xfs_scrub_walk_agfl(sc, xfs_scrub_agfl_block, &sai);
+	error = xfs_agfl_walk(sc->mp, XFS_BUF_TO_AGF(sc->sa.agf_bp),
+			sc->sa.agfl_bp, xfs_scrub_agfl_block, &sai);
+	if (error == XFS_BTREE_QUERY_RANGE_ABORT) {
+		error = 0;
+		goto out_free;
+	}
 	if (error)
 		goto out_free;
 
diff --git a/fs/xfs/scrub/agheader_repair.c b/fs/xfs/scrub/agheader_repair.c
new file mode 100644
index 000000000000..8b91e9ebe1e7
--- /dev/null
+++ b/fs/xfs/scrub/agheader_repair.c
@@ -0,0 +1,70 @@
+/*
+ * Copyright (C) 2018 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_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_inode.h"
+#include "xfs_alloc.h"
+#include "xfs_ialloc.h"
+#include "xfs_rmap.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/trace.h"
+
+/* Superblock */
+
+/* Repair the superblock. */
+int
+xfs_repair_superblock(
+	struct xfs_scrub_context	*sc)
+{
+	struct xfs_mount		*mp = sc->mp;
+	struct xfs_buf			*bp;
+	xfs_agnumber_t			agno;
+	int				error;
+
+	/* Don't try to repair AG 0's sb; let xfs_repair deal with it. */
+	agno = sc->sm->sm_agno;
+	if (agno == 0)
+		return -EOPNOTSUPP;
+
+	error = xfs_sb_get_secondary(mp, sc->tp, agno, &bp);
+	if (error)
+		return error;
+
+	/* Copy AG 0's superblock to this one. */
+	xfs_buf_zero(bp, 0, BBTOB(bp->b_length));
+	xfs_sb_to_disk(XFS_BUF_TO_SBP(bp), &mp->m_sb);
+
+	/* Write this to disk. */
+	xfs_trans_buf_set_type(sc->tp, bp, XFS_BLFT_SB_BUF);
+	xfs_trans_log_buf(sc->tp, bp, 0, BBTOB(bp->b_length) - 1);
+	return error;
+}
diff --git a/fs/xfs/scrub/alloc.c b/fs/xfs/scrub/alloc.c
index 517c079d3f68..941a0a55224e 100644
--- a/fs/xfs/scrub/alloc.c
+++ b/fs/xfs/scrub/alloc.c
@@ -70,7 +70,7 @@ xfs_scrub_allocbt_xref_other(
 		pcur = &sc->sa.cnt_cur;
 	else
 		pcur = &sc->sa.bno_cur;
-	if (!*pcur)
+	if (!*pcur || xfs_scrub_skip_xref(sc->sm))
 		return;
 
 	error = xfs_alloc_lookup_le(*pcur, agbno, len, &has_otherrec);
@@ -172,7 +172,7 @@ xfs_scrub_xref_is_used_space(
 	bool				is_freesp;
 	int				error;
 
-	if (!sc->sa.bno_cur)
+	if (!sc->sa.bno_cur || xfs_scrub_skip_xref(sc->sm))
 		return;
 
 	error = xfs_alloc_has_record(sc->sa.bno_cur, agbno, len, &is_freesp);
diff --git a/fs/xfs/scrub/attr.c b/fs/xfs/scrub/attr.c
index 127575f0abfb..84b6d6b66578 100644
--- a/fs/xfs/scrub/attr.c
+++ b/fs/xfs/scrub/attr.c
@@ -126,8 +126,9 @@ xfs_scrub_xattr_listent(
 	if (args.valuelen != valuelen)
 		xfs_scrub_fblock_set_corrupt(sx->sc, XFS_ATTR_FORK,
 					     args.blkno);
-
 fail_xref:
+	if (sx->sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		context->seen_enough = 1;
 	return;
 }
 
diff --git a/fs/xfs/scrub/bmap.c b/fs/xfs/scrub/bmap.c
index 639d14b51e90..eeadb33a701c 100644
--- a/fs/xfs/scrub/bmap.c
+++ b/fs/xfs/scrub/bmap.c
@@ -51,7 +51,6 @@ xfs_scrub_setup_inode_bmap(
 	struct xfs_scrub_context	*sc,
 	struct xfs_inode		*ip)
 {
-	struct xfs_mount		*mp = sc->mp;
 	int				error;
 
 	error = xfs_scrub_get_inode(sc, ip);
@@ -75,7 +74,7 @@ xfs_scrub_setup_inode_bmap(
 	}
 
 	/* Got the inode, lock it and we're ready to go. */
-	error = xfs_scrub_trans_alloc(sc->sm, mp, &sc->tp);
+	error = xfs_scrub_trans_alloc(sc, 0);
 	if (error)
 		goto out;
 	sc->ilock_flags |= XFS_ILOCK_EXCL;
@@ -175,7 +174,7 @@ xfs_scrub_bmap_xref_rmap(
 	unsigned long long		rmap_end;
 	uint64_t			owner;
 
-	if (!info->sc->sa.rmap_cur)
+	if (!info->sc->sa.rmap_cur || xfs_scrub_skip_xref(info->sc->sm))
 		return;
 
 	if (info->whichfork == XFS_COW_FORK)
@@ -684,7 +683,8 @@ xfs_scrub_bmap(
 	info.lastoff = 0;
 	ifp = XFS_IFORK_PTR(ip, whichfork);
 	for_each_xfs_iext(ifp, &icur, &irec) {
-		if (xfs_scrub_should_terminate(sc, &error))
+		if (xfs_scrub_should_terminate(sc, &error) ||
+		    (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT))
 			break;
 		if (isnullstartblock(irec.br_startblock))
 			continue;
diff --git a/fs/xfs/scrub/btree.c b/fs/xfs/scrub/btree.c
index 54218168c8f9..2d29dceaa00e 100644
--- a/fs/xfs/scrub/btree.c
+++ b/fs/xfs/scrub/btree.c
@@ -442,7 +442,7 @@ xfs_scrub_btree_check_owner(
 	 */
 	if (cur->bc_btnum == XFS_BTNUM_BNO || cur->bc_btnum == XFS_BTNUM_RMAP) {
 		co = kmem_alloc(sizeof(struct check_owner),
-				KM_MAYFAIL | KM_NOFS);
+				KM_MAYFAIL);
 		if (!co)
 			return -ENOMEM;
 		co->level = level;
@@ -455,6 +455,44 @@ xfs_scrub_btree_check_owner(
 }
 
 /*
+ * Check that this btree block has at least minrecs records or is one of the
+ * special blocks that don't require that.
+ */
+STATIC void
+xfs_scrub_btree_check_minrecs(
+	struct xfs_scrub_btree	*bs,
+	int			level,
+	struct xfs_btree_block	*block)
+{
+	unsigned int		numrecs;
+	int			ok_level;
+
+	numrecs = be16_to_cpu(block->bb_numrecs);
+
+	/* More records than minrecs means the block is ok. */
+	if (numrecs >= bs->cur->bc_ops->get_minrecs(bs->cur, level))
+		return;
+
+	/*
+	 * Certain btree blocks /can/ have fewer than minrecs records.  Any
+	 * level greater than or equal to the level of the highest dedicated
+	 * btree block are allowed to violate this constraint.
+	 *
+	 * For a btree rooted in a block, the btree root can have fewer than
+	 * minrecs records.  If the btree is rooted in an inode and does not
+	 * store records in the root, the direct children of the root and the
+	 * root itself can have fewer than minrecs records.
+	 */
+	ok_level = bs->cur->bc_nlevels - 1;
+	if (bs->cur->bc_flags & XFS_BTREE_ROOT_IN_INODE)
+		ok_level--;
+	if (level >= ok_level)
+		return;
+
+	xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, level);
+}
+
+/*
  * Grab and scrub a btree block given a btree pointer.  Returns block
  * and buffer pointers (if applicable) if they're ok to use.
  */
@@ -491,6 +529,8 @@ xfs_scrub_btree_get_block(
 	if (*pbp)
 		xfs_scrub_buffer_recheck(bs->sc, *pbp);
 
+	xfs_scrub_btree_check_minrecs(bs, level, *pblock);
+
 	/*
 	 * Check the block's owner; this function absorbs error codes
 	 * for us.
diff --git a/fs/xfs/scrub/common.c b/fs/xfs/scrub/common.c
index 8ed91d5c868d..41198a5f872c 100644
--- a/fs/xfs/scrub/common.c
+++ b/fs/xfs/scrub/common.c
@@ -44,11 +44,14 @@
 #include "xfs_rmap_btree.h"
 #include "xfs_log.h"
 #include "xfs_trans_priv.h"
+#include "xfs_attr.h"
+#include "xfs_reflink.h"
 #include "scrub/xfs_scrub.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/trace.h"
 #include "scrub/btree.h"
+#include "scrub/repair.h"
 
 /* Common code for the metadata scrubbers. */
 
@@ -539,6 +542,10 @@ xfs_scrub_ag_free(
 		xfs_trans_brelse(sc->tp, sa->agi_bp);
 		sa->agi_bp = NULL;
 	}
+	if (sa->pag) {
+		xfs_perag_put(sa->pag);
+		sa->pag = NULL;
+	}
 	sa->agno = NULLAGNUMBER;
 }
 
@@ -566,15 +573,53 @@ xfs_scrub_ag_init(
 	return xfs_scrub_ag_btcur_init(sc, sa);
 }
 
+/*
+ * Grab the per-ag structure if we haven't already gotten it.  Teardown of the
+ * xfs_scrub_ag will release it for us.
+ */
+void
+xfs_scrub_perag_get(
+	struct xfs_mount	*mp,
+	struct xfs_scrub_ag	*sa)
+{
+	if (!sa->pag)
+		sa->pag = xfs_perag_get(mp, sa->agno);
+}
+
 /* Per-scrubber setup functions */
 
+/*
+ * Grab an empty transaction so that we can re-grab locked buffers if
+ * one of our btrees turns out to be cyclic.
+ *
+ * If we're going to repair something, we need to ask for the largest possible
+ * log reservation so that we can handle the worst case scenario for metadata
+ * updates while rebuilding a metadata item.  We also need to reserve as many
+ * blocks in the head transaction as we think we're going to need to rebuild
+ * the metadata object.
+ */
+int
+xfs_scrub_trans_alloc(
+	struct xfs_scrub_context	*sc,
+	uint				resblks)
+{
+	if (sc->sm->sm_flags & XFS_SCRUB_IFLAG_REPAIR)
+		return xfs_trans_alloc(sc->mp, &M_RES(sc->mp)->tr_itruncate,
+				resblks, 0, 0, &sc->tp);
+
+	return xfs_trans_alloc_empty(sc->mp, &sc->tp);
+}
+
 /* Set us up with a transaction and an empty context. */
 int
 xfs_scrub_setup_fs(
 	struct xfs_scrub_context	*sc,
 	struct xfs_inode		*ip)
 {
-	return xfs_scrub_trans_alloc(sc->sm, sc->mp, &sc->tp);
+	uint				resblks;
+
+	resblks = xfs_repair_calc_ag_resblks(sc);
+	return xfs_scrub_trans_alloc(sc, resblks);
 }
 
 /* Set us up with AG headers and btree cursors. */
@@ -695,7 +740,6 @@ xfs_scrub_setup_inode_contents(
 	struct xfs_inode		*ip,
 	unsigned int			resblks)
 {
-	struct xfs_mount		*mp = sc->mp;
 	int				error;
 
 	error = xfs_scrub_get_inode(sc, ip);
@@ -705,7 +749,7 @@ xfs_scrub_setup_inode_contents(
 	/* Got the inode, lock it and we're ready to go. */
 	sc->ilock_flags = XFS_IOLOCK_EXCL | XFS_MMAPLOCK_EXCL;
 	xfs_ilock(sc->ip, sc->ilock_flags);
-	error = xfs_scrub_trans_alloc(sc->sm, mp, &sc->tp);
+	error = xfs_scrub_trans_alloc(sc, resblks);
 	if (error)
 		goto out;
 	sc->ilock_flags |= XFS_ILOCK_EXCL;
@@ -727,6 +771,10 @@ xfs_scrub_should_check_xref(
 	int				*error,
 	struct xfs_btree_cur		**curpp)
 {
+	/* No point in xref if we already know we're corrupt. */
+	if (xfs_scrub_skip_xref(sc->sm))
+		return false;
+
 	if (*error == 0)
 		return true;
 
@@ -773,3 +821,80 @@ xfs_scrub_buffer_recheck(
 	sc->sm->sm_flags |= XFS_SCRUB_OFLAG_CORRUPT;
 	trace_xfs_scrub_block_error(sc, bp->b_bn, fa);
 }
+
+/*
+ * Scrub the attr/data forks of a metadata inode.  The metadata inode must be
+ * pointed to by sc->ip and the ILOCK must be held.
+ */
+int
+xfs_scrub_metadata_inode_forks(
+	struct xfs_scrub_context	*sc)
+{
+	__u32				smtype;
+	bool				shared;
+	int				error;
+
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return 0;
+
+	/* Metadata inodes don't live on the rt device. */
+	if (sc->ip->i_d.di_flags & XFS_DIFLAG_REALTIME) {
+		xfs_scrub_ino_set_corrupt(sc, sc->ip->i_ino);
+		return 0;
+	}
+
+	/* They should never participate in reflink. */
+	if (xfs_is_reflink_inode(sc->ip)) {
+		xfs_scrub_ino_set_corrupt(sc, sc->ip->i_ino);
+		return 0;
+	}
+
+	/* They also should never have extended attributes. */
+	if (xfs_inode_hasattr(sc->ip)) {
+		xfs_scrub_ino_set_corrupt(sc, sc->ip->i_ino);
+		return 0;
+	}
+
+	/* Invoke the data fork scrubber. */
+	smtype = sc->sm->sm_type;
+	sc->sm->sm_type = XFS_SCRUB_TYPE_BMBTD;
+	error = xfs_scrub_bmap_data(sc);
+	sc->sm->sm_type = smtype;
+	if (error || (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT))
+		return error;
+
+	/* Look for incorrect shared blocks. */
+	if (xfs_sb_version_hasreflink(&sc->mp->m_sb)) {
+		error = xfs_reflink_inode_has_shared_extents(sc->tp, sc->ip,
+				&shared);
+		if (!xfs_scrub_fblock_process_error(sc, XFS_DATA_FORK, 0,
+				&error))
+			return error;
+		if (shared)
+			xfs_scrub_ino_set_corrupt(sc, sc->ip->i_ino);
+	}
+
+	return error;
+}
+
+/*
+ * Try to lock an inode in violation of the usual locking order rules.  For
+ * example, trying to get the IOLOCK while in transaction context, or just
+ * plain breaking AG-order or inode-order inode locking rules.  Either way,
+ * the only way to avoid an ABBA deadlock is to use trylock and back off if
+ * we can't.
+ */
+int
+xfs_scrub_ilock_inverted(
+	struct xfs_inode	*ip,
+	uint			lock_mode)
+{
+	int			i;
+
+	for (i = 0; i < 20; i++) {
+		if (xfs_ilock_nowait(ip, lock_mode))
+			return 0;
+		delay(1);
+	}
+	return -EDEADLOCK;
+}
diff --git a/fs/xfs/scrub/common.h b/fs/xfs/scrub/common.h
index deaf60400981..76bb2d1d808c 100644
--- a/fs/xfs/scrub/common.h
+++ b/fs/xfs/scrub/common.h
@@ -38,19 +38,7 @@ xfs_scrub_should_terminate(
 	return false;
 }
 
-/*
- * Grab an empty transaction so that we can re-grab locked buffers if
- * one of our btrees turns out to be cyclic.
- */
-static inline int
-xfs_scrub_trans_alloc(
-	struct xfs_scrub_metadata	*sm,
-	struct xfs_mount		*mp,
-	struct xfs_trans		**tpp)
-{
-	return xfs_trans_alloc_empty(mp, tpp);
-}
-
+int xfs_scrub_trans_alloc(struct xfs_scrub_context *sc, uint resblks);
 bool xfs_scrub_process_error(struct xfs_scrub_context *sc, xfs_agnumber_t agno,
 		xfs_agblock_t bno, int *error);
 bool xfs_scrub_fblock_process_error(struct xfs_scrub_context *sc, int whichfork,
@@ -135,16 +123,13 @@ xfs_scrub_setup_quota(struct xfs_scrub_context *sc, struct xfs_inode *ip)
 void xfs_scrub_ag_free(struct xfs_scrub_context *sc, struct xfs_scrub_ag *sa);
 int xfs_scrub_ag_init(struct xfs_scrub_context *sc, xfs_agnumber_t agno,
 		      struct xfs_scrub_ag *sa);
+void xfs_scrub_perag_get(struct xfs_mount *mp, struct xfs_scrub_ag *sa);
 int xfs_scrub_ag_read_headers(struct xfs_scrub_context *sc, xfs_agnumber_t agno,
 			      struct xfs_buf **agi, struct xfs_buf **agf,
 			      struct xfs_buf **agfl);
 void xfs_scrub_ag_btcur_free(struct xfs_scrub_ag *sa);
 int xfs_scrub_ag_btcur_init(struct xfs_scrub_context *sc,
 			    struct xfs_scrub_ag *sa);
-int xfs_scrub_walk_agfl(struct xfs_scrub_context *sc,
-			int (*fn)(struct xfs_scrub_context *, xfs_agblock_t bno,
-				  void *),
-			void *priv);
 int xfs_scrub_count_rmap_ownedby_ag(struct xfs_scrub_context *sc,
 				    struct xfs_btree_cur *cur,
 				    struct xfs_owner_info *oinfo,
@@ -157,4 +142,17 @@ int xfs_scrub_setup_inode_contents(struct xfs_scrub_context *sc,
 				   struct xfs_inode *ip, unsigned int resblks);
 void xfs_scrub_buffer_recheck(struct xfs_scrub_context *sc, struct xfs_buf *bp);
 
+/*
+ * Don't bother cross-referencing if we already found corruption or cross
+ * referencing discrepancies.
+ */
+static inline bool xfs_scrub_skip_xref(struct xfs_scrub_metadata *sm)
+{
+	return sm->sm_flags & (XFS_SCRUB_OFLAG_CORRUPT |
+			       XFS_SCRUB_OFLAG_XCORRUPT);
+}
+
+int xfs_scrub_metadata_inode_forks(struct xfs_scrub_context *sc);
+int xfs_scrub_ilock_inverted(struct xfs_inode *ip, uint lock_mode);
+
 #endif	/* __XFS_SCRUB_COMMON_H__ */
diff --git a/fs/xfs/scrub/dir.c b/fs/xfs/scrub/dir.c
index 38f29806eb54..1a4309b3e786 100644
--- a/fs/xfs/scrub/dir.c
+++ b/fs/xfs/scrub/dir.c
@@ -172,7 +172,7 @@ xfs_scrub_dir_actor(
 	error = xfs_dir_lookup(sdc->sc->tp, ip, &xname, &lookup_ino, NULL);
 	if (!xfs_scrub_fblock_process_error(sdc->sc, XFS_DATA_FORK, offset,
 			&error))
-		goto fail_xref;
+		goto out;
 	if (lookup_ino != ino) {
 		xfs_scrub_fblock_set_corrupt(sdc->sc, XFS_DATA_FORK, offset);
 		goto out;
@@ -183,8 +183,13 @@ xfs_scrub_dir_actor(
 	if (error)
 		goto out;
 out:
-	return error;
-fail_xref:
+	/*
+	 * A negative error code returned here is supposed to cause the
+	 * dir_emit caller (xfs_readdir) to abort the directory iteration
+	 * and return zero to xfs_scrub_directory.
+	 */
+	if (error == 0 && sdc->sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return -EFSCORRUPTED;
 	return error;
 }
 
@@ -240,6 +245,9 @@ xfs_scrub_dir_rec(
 	}
 	xfs_scrub_buffer_recheck(ds->sc, bp);
 
+	if (ds->sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		goto out_relse;
+
 	dent = (struct xfs_dir2_data_entry *)(((char *)bp->b_addr) + off);
 
 	/* Make sure we got a real directory entry. */
@@ -357,6 +365,9 @@ xfs_scrub_directory_data_bestfree(
 
 	/* XXX: Check xfs_dir3_data_hdr.pad is zero once we start setting it. */
 
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		goto out_buf;
+
 	/* Do the bestfrees correspond to actual free space? */
 	bf = d_ops->data_bestfree_p(bp->b_addr);
 	smallest_bestfree = UINT_MAX;
@@ -413,14 +424,18 @@ xfs_scrub_directory_data_bestfree(
 
 		/* Spot check this free entry */
 		tag = be16_to_cpu(*xfs_dir2_data_unused_tag_p(dup));
-		if (tag != ((char *)dup - (char *)bp->b_addr))
+		if (tag != ((char *)dup - (char *)bp->b_addr)) {
 			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
+			goto out_buf;
+		}
 
 		/*
 		 * Either this entry is a bestfree or it's smaller than
 		 * any of the bestfrees.
 		 */
 		xfs_scrub_directory_check_free_entry(sc, lblk, bf, dup);
+		if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+			goto out_buf;
 
 		/* Move on. */
 		newlen = be16_to_cpu(dup->length);
@@ -546,6 +561,8 @@ xfs_scrub_directory_leaf1_bestfree(
 	}
 	if (leafhdr.stale != stale)
 		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		goto out;
 
 	/* Check all the bestfree entries. */
 	for (i = 0; i < bestcount; i++, bestp++) {
@@ -556,9 +573,11 @@ xfs_scrub_directory_leaf1_bestfree(
 				i * args->geo->fsbcount, -1, &dbp);
 		if (!xfs_scrub_fblock_process_error(sc, XFS_DATA_FORK, lblk,
 				&error))
-			continue;
+			break;
 		xfs_scrub_directory_check_freesp(sc, lblk, dbp, best);
 		xfs_trans_brelse(sc->tp, dbp);
+		if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+			goto out;
 	}
 out:
 	return error;
@@ -607,7 +626,7 @@ xfs_scrub_directory_free_bestfree(
 				-1, &dbp);
 		if (!xfs_scrub_fblock_process_error(sc, XFS_DATA_FORK, lblk,
 				&error))
-			continue;
+			break;
 		xfs_scrub_directory_check_freesp(sc, lblk, dbp, best);
 		xfs_trans_brelse(sc->tp, dbp);
 	}
@@ -656,7 +675,7 @@ xfs_scrub_directory_blocks(
 
 	/* Iterate all the data extents in the directory... */
 	found = xfs_iext_lookup_extent(sc->ip, ifp, lblk, &icur, &got);
-	while (found) {
+	while (found && !(sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)) {
 		/* Block directories only have a single block at offset 0. */
 		if (is_block &&
 		    (got.br_startoff > 0 ||
@@ -719,7 +738,7 @@ xfs_scrub_directory_blocks(
 	/* Scan for free blocks */
 	lblk = free_lblk;
 	found = xfs_iext_lookup_extent(sc->ip, ifp, lblk, &icur, &got);
-	while (found) {
+	while (found && !(sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)) {
 		/*
 		 * Dirs can't have blocks mapped above 2^32.
 		 * Single-block dirs shouldn't even be here.
diff --git a/fs/xfs/scrub/ialloc.c b/fs/xfs/scrub/ialloc.c
index 106ca4bd753f..00a834d3b56d 100644
--- a/fs/xfs/scrub/ialloc.c
+++ b/fs/xfs/scrub/ialloc.c
@@ -387,7 +387,8 @@ xfs_scrub_iallocbt_xref_rmap_btreeblks(
 	int				error;
 
 	if (!sc->sa.ino_cur || !sc->sa.rmap_cur ||
-	    (xfs_sb_version_hasfinobt(&sc->mp->m_sb) && !sc->sa.fino_cur))
+	    (xfs_sb_version_hasfinobt(&sc->mp->m_sb) && !sc->sa.fino_cur) ||
+	    xfs_scrub_skip_xref(sc->sm))
 		return;
 
 	/* Check that we saw as many inobt blocks as the rmap says. */
@@ -424,7 +425,7 @@ xfs_scrub_iallocbt_xref_rmap_inodes(
 	xfs_filblks_t			blocks;
 	int				error;
 
-	if (!sc->sa.rmap_cur)
+	if (!sc->sa.rmap_cur || xfs_scrub_skip_xref(sc->sm))
 		return;
 
 	/* Check that we saw as many inode blocks as the rmap knows about. */
@@ -496,7 +497,7 @@ xfs_scrub_xref_inode_check(
 	bool				has_inodes;
 	int				error;
 
-	if (!(*icur))
+	if (!(*icur) || xfs_scrub_skip_xref(sc->sm))
 		return;
 
 	error = xfs_ialloc_has_inodes_at_extent(*icur, agbno, len, &has_inodes);
diff --git a/fs/xfs/scrub/inode.c b/fs/xfs/scrub/inode.c
index df14930e4fc5..0c696f7018de 100644
--- a/fs/xfs/scrub/inode.c
+++ b/fs/xfs/scrub/inode.c
@@ -55,7 +55,6 @@ xfs_scrub_setup_inode(
 	struct xfs_scrub_context	*sc,
 	struct xfs_inode		*ip)
 {
-	struct xfs_mount		*mp = sc->mp;
 	int				error;
 
 	/*
@@ -68,7 +67,7 @@ xfs_scrub_setup_inode(
 		break;
 	case -EFSCORRUPTED:
 	case -EFSBADCRC:
-		return xfs_scrub_trans_alloc(sc->sm, mp, &sc->tp);
+		return xfs_scrub_trans_alloc(sc, 0);
 	default:
 		return error;
 	}
@@ -76,7 +75,7 @@ xfs_scrub_setup_inode(
 	/* Got the inode, lock it and we're ready to go. */
 	sc->ilock_flags = XFS_IOLOCK_EXCL | XFS_MMAPLOCK_EXCL;
 	xfs_ilock(sc->ip, sc->ilock_flags);
-	error = xfs_scrub_trans_alloc(sc->sm, mp, &sc->tp);
+	error = xfs_scrub_trans_alloc(sc, 0);
 	if (error)
 		goto out;
 	sc->ilock_flags |= XFS_ILOCK_EXCL;
@@ -449,7 +448,7 @@ xfs_scrub_inode_xref_finobt(
 	int				has_record;
 	int				error;
 
-	if (!sc->sa.fino_cur)
+	if (!sc->sa.fino_cur || xfs_scrub_skip_xref(sc->sm))
 		return;
 
 	agino = XFS_INO_TO_AGINO(sc->mp, ino);
@@ -492,6 +491,9 @@ xfs_scrub_inode_xref_bmap(
 	xfs_filblks_t			acount;
 	int				error;
 
+	if (xfs_scrub_skip_xref(sc->sm))
+		return;
+
 	/* Walk all the extents to check nextents/naextents/nblocks. */
 	error = xfs_bmap_count_blocks(sc->tp, sc->ip, XFS_DATA_FORK,
 			&nextents, &count);
diff --git a/fs/xfs/scrub/parent.c b/fs/xfs/scrub/parent.c
index 1fb88c18d455..77c6b22c6bfd 100644
--- a/fs/xfs/scrub/parent.c
+++ b/fs/xfs/scrub/parent.c
@@ -147,6 +147,9 @@ xfs_scrub_parent_validate(
 
 	*try_again = false;
 
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		goto out;
+
 	/* '..' must not point to ourselves. */
 	if (sc->ip->i_ino == dnum) {
 		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, 0);
@@ -211,7 +214,9 @@ xfs_scrub_parent_validate(
 	 */
 	xfs_iunlock(sc->ip, sc->ilock_flags);
 	sc->ilock_flags = 0;
-	xfs_ilock(dp, XFS_IOLOCK_SHARED);
+	error = xfs_scrub_ilock_inverted(dp, XFS_IOLOCK_SHARED);
+	if (error)
+		goto out_rele;
 
 	/* Go looking for our dentry. */
 	error = xfs_scrub_parent_count_parent_dentries(sc, dp, &nlink);
@@ -220,8 +225,10 @@ xfs_scrub_parent_validate(
 
 	/* Drop the parent lock, relock this inode. */
 	xfs_iunlock(dp, XFS_IOLOCK_SHARED);
+	error = xfs_scrub_ilock_inverted(sc->ip, XFS_IOLOCK_EXCL);
+	if (error)
+		goto out_rele;
 	sc->ilock_flags = XFS_IOLOCK_EXCL;
-	xfs_ilock(sc->ip, sc->ilock_flags);
 
 	/*
 	 * If we're an unlinked directory, the parent /won't/ have a link
@@ -323,5 +330,13 @@ xfs_scrub_parent(
 	if (try_again && tries == 20)
 		xfs_scrub_set_incomplete(sc);
 out:
+	/*
+	 * If we failed to lock the parent inode even after a retry, just mark
+	 * this scrub incomplete and return.
+	 */
+	if (sc->try_harder && error == -EDEADLOCK) {
+		error = 0;
+		xfs_scrub_set_incomplete(sc);
+	}
 	return error;
 }
diff --git a/fs/xfs/scrub/quota.c b/fs/xfs/scrub/quota.c
index 6ba465e6c885..15ae4d23d6ac 100644
--- a/fs/xfs/scrub/quota.c
+++ b/fs/xfs/scrub/quota.c
@@ -66,25 +66,43 @@ xfs_scrub_setup_quota(
 	struct xfs_inode		*ip)
 {
 	uint				dqtype;
+	int				error;
+
+	if (!XFS_IS_QUOTA_RUNNING(sc->mp) || !XFS_IS_QUOTA_ON(sc->mp))
+		return -ENOENT;
 
 	dqtype = xfs_scrub_quota_to_dqtype(sc);
 	if (dqtype == 0)
 		return -EINVAL;
+	sc->has_quotaofflock = true;
+	mutex_lock(&sc->mp->m_quotainfo->qi_quotaofflock);
 	if (!xfs_this_quota_on(sc->mp, dqtype))
 		return -ENOENT;
+	error = xfs_scrub_setup_fs(sc, ip);
+	if (error)
+		return error;
+	sc->ip = xfs_quota_inode(sc->mp, dqtype);
+	xfs_ilock(sc->ip, XFS_ILOCK_EXCL);
+	sc->ilock_flags = XFS_ILOCK_EXCL;
 	return 0;
 }
 
 /* Quotas. */
 
+struct xfs_scrub_quota_info {
+	struct xfs_scrub_context	*sc;
+	xfs_dqid_t			last_id;
+};
+
 /* Scrub the fields in an individual quota item. */
-STATIC void
+STATIC int
 xfs_scrub_quota_item(
-	struct xfs_scrub_context	*sc,
-	uint				dqtype,
 	struct xfs_dquot		*dq,
-	xfs_dqid_t			id)
+	uint				dqtype,
+	void				*priv)
 {
+	struct xfs_scrub_quota_info	*sqi = priv;
+	struct xfs_scrub_context	*sc = sqi->sc;
 	struct xfs_mount		*mp = sc->mp;
 	struct xfs_disk_dquot		*d = &dq->q_core;
 	struct xfs_quotainfo		*qi = mp->m_quotainfo;
@@ -99,17 +117,18 @@ xfs_scrub_quota_item(
 	unsigned long long		icount;
 	unsigned long long		rcount;
 	xfs_ino_t			fs_icount;
-
-	offset = id / qi->qi_dqperchunk;
+	xfs_dqid_t			id = be32_to_cpu(d->d_id);
 
 	/*
-	 * We fed $id and DQNEXT into the xfs_qm_dqget call, which means
-	 * that the actual dquot we got must either have the same id or
-	 * the next higher id.
+	 * Except for the root dquot, the actual dquot we got must either have
+	 * the same or higher id as we saw before.
 	 */
-	if (id > be32_to_cpu(d->d_id))
+	offset = id / qi->qi_dqperchunk;
+	if (id && id <= sqi->last_id)
 		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, offset);
 
+	sqi->last_id = id;
+
 	/* Did we get the dquot type we wanted? */
 	if (dqtype != (d->d_flags & XFS_DQ_ALLTYPES))
 		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, offset);
@@ -183,115 +202,85 @@ xfs_scrub_quota_item(
 		xfs_scrub_fblock_set_warning(sc, XFS_DATA_FORK, offset);
 	if (id != 0 && rhard != 0 && rcount > rhard)
 		xfs_scrub_fblock_set_warning(sc, XFS_DATA_FORK, offset);
+
+	return 0;
 }
 
-/* Scrub all of a quota type's items. */
-int
-xfs_scrub_quota(
+/* Check the quota's data fork. */
+STATIC int
+xfs_scrub_quota_data_fork(
 	struct xfs_scrub_context	*sc)
 {
 	struct xfs_bmbt_irec		irec = { 0 };
-	struct xfs_mount		*mp = sc->mp;
-	struct xfs_inode		*ip;
-	struct xfs_quotainfo		*qi = mp->m_quotainfo;
-	struct xfs_dquot		*dq;
+	struct xfs_iext_cursor		icur;
+	struct xfs_quotainfo		*qi = sc->mp->m_quotainfo;
+	struct xfs_ifork		*ifp;
 	xfs_fileoff_t			max_dqid_off;
-	xfs_fileoff_t			off = 0;
-	xfs_dqid_t			id = 0;
-	uint				dqtype;
-	int				nimaps;
 	int				error = 0;
 
-	if (!XFS_IS_QUOTA_RUNNING(mp) || !XFS_IS_QUOTA_ON(mp))
-		return -ENOENT;
-
-	mutex_lock(&qi->qi_quotaofflock);
-	dqtype = xfs_scrub_quota_to_dqtype(sc);
-	if (!xfs_this_quota_on(sc->mp, dqtype)) {
-		error = -ENOENT;
-		goto out_unlock_quota;
-	}
-
-	/* Attach to the quota inode and set sc->ip so that reporting works. */
-	ip = xfs_quota_inode(sc->mp, dqtype);
-	sc->ip = ip;
+	/* Invoke the fork scrubber. */
+	error = xfs_scrub_metadata_inode_forks(sc);
+	if (error || (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT))
+		return error;
 
-	/* Look for problem extents. */
-	xfs_ilock(ip, XFS_ILOCK_EXCL);
-	if (ip->i_d.di_flags & XFS_DIFLAG_REALTIME) {
-		xfs_scrub_ino_set_corrupt(sc, sc->ip->i_ino);
-		goto out_unlock_inode;
-	}
+	/* Check for data fork problems that apply only to quota files. */
 	max_dqid_off = ((xfs_dqid_t)-1) / qi->qi_dqperchunk;
-	while (1) {
+	ifp = XFS_IFORK_PTR(sc->ip, XFS_DATA_FORK);
+	for_each_xfs_iext(ifp, &icur, &irec) {
 		if (xfs_scrub_should_terminate(sc, &error))
 			break;
-
-		off = irec.br_startoff + irec.br_blockcount;
-		nimaps = 1;
-		error = xfs_bmapi_read(ip, off, -1, &irec, &nimaps,
-				XFS_BMAPI_ENTIRE);
-		if (!xfs_scrub_fblock_process_error(sc, XFS_DATA_FORK, off,
-				&error))
-			goto out_unlock_inode;
-		if (!nimaps)
-			break;
-		if (irec.br_startblock == HOLESTARTBLOCK)
-			continue;
-
-		/* Check the extent record doesn't point to crap. */
-		if (irec.br_startblock + irec.br_blockcount <=
-		    irec.br_startblock)
-			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK,
-					irec.br_startoff);
-		if (!xfs_verify_fsbno(mp, irec.br_startblock) ||
-		    !xfs_verify_fsbno(mp, irec.br_startblock +
-					irec.br_blockcount - 1))
-			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK,
-					irec.br_startoff);
-
 		/*
-		 * Unwritten extents or blocks mapped above the highest
+		 * delalloc extents or blocks mapped above the highest
 		 * quota id shouldn't happen.
 		 */
 		if (isnullstartblock(irec.br_startblock) ||
 		    irec.br_startoff > max_dqid_off ||
-		    irec.br_startoff + irec.br_blockcount > max_dqid_off + 1)
-			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, off);
+		    irec.br_startoff + irec.br_blockcount - 1 > max_dqid_off) {
+			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK,
+					irec.br_startoff);
+			break;
+		}
 	}
-	xfs_iunlock(ip, XFS_ILOCK_EXCL);
-	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
-		goto out;
 
-	/* Check all the quota items. */
-	while (id < ((xfs_dqid_t)-1ULL)) {
-		if (xfs_scrub_should_terminate(sc, &error))
-			break;
+	return error;
+}
 
-		error = xfs_qm_dqget(mp, NULL, id, dqtype, XFS_QMOPT_DQNEXT,
-				&dq);
-		if (error == -ENOENT)
-			break;
-		if (!xfs_scrub_fblock_process_error(sc, XFS_DATA_FORK,
-				id * qi->qi_dqperchunk, &error))
-			break;
+/* Scrub all of a quota type's items. */
+int
+xfs_scrub_quota(
+	struct xfs_scrub_context	*sc)
+{
+	struct xfs_scrub_quota_info	sqi;
+	struct xfs_mount		*mp = sc->mp;
+	struct xfs_quotainfo		*qi = mp->m_quotainfo;
+	uint				dqtype;
+	int				error = 0;
 
-		xfs_scrub_quota_item(sc, dqtype, dq, id);
+	dqtype = xfs_scrub_quota_to_dqtype(sc);
 
-		id = be32_to_cpu(dq->q_core.d_id) + 1;
-		xfs_qm_dqput(dq);
-		if (!id)
-			break;
-	}
+	/* Look for problem extents. */
+	error = xfs_scrub_quota_data_fork(sc);
+	if (error)
+		goto out;
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		goto out;
+
+	/*
+	 * Check all the quota items.  Now that we've checked the quota inode
+	 * data fork we have to drop ILOCK_EXCL to use the regular dquot
+	 * functions.
+	 */
+	xfs_iunlock(sc->ip, sc->ilock_flags);
+	sc->ilock_flags = 0;
+	sqi.sc = sc;
+	sqi.last_id = 0;
+	error = xfs_qm_dqiterate(mp, dqtype, xfs_scrub_quota_item, &sqi);
+	sc->ilock_flags = XFS_ILOCK_EXCL;
+	xfs_ilock(sc->ip, sc->ilock_flags);
+	if (!xfs_scrub_fblock_process_error(sc, XFS_DATA_FORK,
+			sqi.last_id * qi->qi_dqperchunk, &error))
+		goto out;
 
 out:
-	/* We set sc->ip earlier, so make sure we clear it now. */
-	sc->ip = NULL;
-out_unlock_quota:
-	mutex_unlock(&qi->qi_quotaofflock);
 	return error;
-
-out_unlock_inode:
-	xfs_iunlock(ip, XFS_ILOCK_EXCL);
-	goto out;
 }
diff --git a/fs/xfs/scrub/refcount.c b/fs/xfs/scrub/refcount.c
index 400f1561cd3d..324a5f159145 100644
--- a/fs/xfs/scrub/refcount.c
+++ b/fs/xfs/scrub/refcount.c
@@ -150,7 +150,7 @@ xfs_scrub_refcountbt_rmap_check(
 		 * so we don't need insertion sort here.
 		 */
 		frag = kmem_alloc(sizeof(struct xfs_scrub_refcnt_frag),
-				KM_MAYFAIL | KM_NOFS);
+				KM_MAYFAIL);
 		if (!frag)
 			return -ENOMEM;
 		memcpy(&frag->rm, rec, sizeof(frag->rm));
@@ -310,7 +310,7 @@ xfs_scrub_refcountbt_xref_rmap(
 	struct xfs_scrub_refcnt_frag	*n;
 	int				error;
 
-	if (!sc->sa.rmap_cur)
+	if (!sc->sa.rmap_cur || xfs_scrub_skip_xref(sc->sm))
 		return;
 
 	/* Cross-reference with the rmapbt to confirm the refcount. */
@@ -404,7 +404,7 @@ xfs_scrub_refcount_xref_rmap(
 	xfs_filblks_t			blocks;
 	int				error;
 
-	if (!sc->sa.rmap_cur)
+	if (!sc->sa.rmap_cur || xfs_scrub_skip_xref(sc->sm))
 		return;
 
 	/* Check that we saw as many refcbt blocks as the rmap knows about. */
@@ -460,7 +460,7 @@ xfs_scrub_xref_is_cow_staging(
 	int				has_refcount;
 	int				error;
 
-	if (!sc->sa.refc_cur)
+	if (!sc->sa.refc_cur || xfs_scrub_skip_xref(sc->sm))
 		return;
 
 	/* Find the CoW staging extent. */
@@ -504,7 +504,7 @@ xfs_scrub_xref_is_not_shared(
 	bool				shared;
 	int				error;
 
-	if (!sc->sa.refc_cur)
+	if (!sc->sa.refc_cur || xfs_scrub_skip_xref(sc->sm))
 		return;
 
 	error = xfs_refcount_has_record(sc->sa.refc_cur, agbno, len, &shared);
diff --git a/fs/xfs/scrub/repair.c b/fs/xfs/scrub/repair.c
new file mode 100644
index 000000000000..e3e8fba1c99c
--- /dev/null
+++ b/fs/xfs/scrub/repair.c
@@ -0,0 +1,1089 @@
+/*
+ * Copyright (C) 2018 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_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_inode.h"
+#include "xfs_icache.h"
+#include "xfs_alloc.h"
+#include "xfs_alloc_btree.h"
+#include "xfs_ialloc.h"
+#include "xfs_ialloc_btree.h"
+#include "xfs_rmap.h"
+#include "xfs_rmap_btree.h"
+#include "xfs_refcount.h"
+#include "xfs_refcount_btree.h"
+#include "xfs_extent_busy.h"
+#include "xfs_ag_resv.h"
+#include "xfs_trans_space.h"
+#include "xfs_quota.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/trace.h"
+#include "scrub/repair.h"
+
+/*
+ * Attempt to repair some metadata, if the metadata is corrupt and userspace
+ * told us to fix it.  This function returns -EAGAIN to mean "re-run scrub",
+ * and will set *fixed to true if it thinks it repaired anything.
+ */
+int
+xfs_repair_attempt(
+	struct xfs_inode		*ip,
+	struct xfs_scrub_context	*sc,
+	bool				*fixed)
+{
+	int				error = 0;
+
+	trace_xfs_repair_attempt(ip, sc->sm, error);
+
+	xfs_scrub_ag_btcur_free(&sc->sa);
+
+	/* Repair whatever's broken. */
+	ASSERT(sc->ops->repair);
+	error = sc->ops->repair(sc);
+	trace_xfs_repair_done(ip, sc->sm, error);
+	switch (error) {
+	case 0:
+		/*
+		 * Repair succeeded.  Commit the fixes and perform a second
+		 * scrub so that we can tell userspace if we fixed the problem.
+		 */
+		sc->sm->sm_flags &= ~XFS_SCRUB_FLAGS_OUT;
+		*fixed = true;
+		return -EAGAIN;
+	case -EDEADLOCK:
+	case -EAGAIN:
+		/* Tell the caller to try again having grabbed all the locks. */
+		if (!sc->try_harder) {
+			sc->try_harder = true;
+			return -EAGAIN;
+		}
+		/*
+		 * We tried harder but still couldn't grab all the resources
+		 * we needed to fix it.  The corruption has not been fixed,
+		 * so report back to userspace.
+		 */
+		return -EFSCORRUPTED;
+	default:
+		return error;
+	}
+}
+
+/*
+ * Complain about unfixable problems in the filesystem.  We don't log
+ * corruptions when IFLAG_REPAIR wasn't set on the assumption that the driver
+ * program is xfs_scrub, which will call back with IFLAG_REPAIR set if the
+ * administrator isn't running xfs_scrub in no-repairs mode.
+ *
+ * Use this helper function because _ratelimited silently declares a static
+ * structure to track rate limiting information.
+ */
+void
+xfs_repair_failure(
+	struct xfs_mount		*mp)
+{
+	xfs_alert_ratelimited(mp,
+"Corruption not fixed during online repair.  Unmount and run xfs_repair.");
+}
+
+/*
+ * Repair probe -- userspace uses this to probe if we're willing to repair a
+ * given mountpoint.
+ */
+int
+xfs_repair_probe(
+	struct xfs_scrub_context	*sc)
+{
+	int				error = 0;
+
+	if (xfs_scrub_should_terminate(sc, &error))
+		return error;
+
+	return 0;
+}
+
+/*
+ * Roll a transaction, keeping the AG headers locked and reinitializing
+ * the btree cursors.
+ */
+int
+xfs_repair_roll_ag_trans(
+	struct xfs_scrub_context	*sc)
+{
+	int				error;
+
+	/* Keep the AG header buffers locked so we can keep going. */
+	xfs_trans_bhold(sc->tp, sc->sa.agi_bp);
+	xfs_trans_bhold(sc->tp, sc->sa.agf_bp);
+	xfs_trans_bhold(sc->tp, sc->sa.agfl_bp);
+
+	/* Roll the transaction. */
+	error = xfs_trans_roll(&sc->tp);
+	if (error)
+		goto out_release;
+
+	/* Join AG headers to the new transaction. */
+	xfs_trans_bjoin(sc->tp, sc->sa.agi_bp);
+	xfs_trans_bjoin(sc->tp, sc->sa.agf_bp);
+	xfs_trans_bjoin(sc->tp, sc->sa.agfl_bp);
+
+	return 0;
+
+out_release:
+	/*
+	 * Rolling failed, so release the hold on the buffers.  The
+	 * buffers will be released during teardown on our way out
+	 * of the kernel.
+	 */
+	xfs_trans_bhold_release(sc->tp, sc->sa.agi_bp);
+	xfs_trans_bhold_release(sc->tp, sc->sa.agf_bp);
+	xfs_trans_bhold_release(sc->tp, sc->sa.agfl_bp);
+
+	return error;
+}
+
+/*
+ * Does the given AG have enough space to rebuild a btree?  Neither AG
+ * reservation can be critical, and we must have enough space (factoring
+ * in AG reservations) to construct a whole btree.
+ */
+bool
+xfs_repair_ag_has_space(
+	struct xfs_perag		*pag,
+	xfs_extlen_t			nr_blocks,
+	enum xfs_ag_resv_type		type)
+{
+	return  !xfs_ag_resv_critical(pag, XFS_AG_RESV_RMAPBT) &&
+		!xfs_ag_resv_critical(pag, XFS_AG_RESV_METADATA) &&
+		pag->pagf_freeblks > xfs_ag_resv_needed(pag, type) + nr_blocks;
+}
+
+/*
+ * Figure out how many blocks to reserve for an AG repair.  We calculate the
+ * worst case estimate for the number of blocks we'd need to rebuild one of
+ * any type of per-AG btree.
+ */
+xfs_extlen_t
+xfs_repair_calc_ag_resblks(
+	struct xfs_scrub_context	*sc)
+{
+	struct xfs_mount		*mp = sc->mp;
+	struct xfs_scrub_metadata	*sm = sc->sm;
+	struct xfs_perag		*pag;
+	struct xfs_buf			*bp;
+	xfs_agino_t			icount = 0;
+	xfs_extlen_t			aglen = 0;
+	xfs_extlen_t			usedlen;
+	xfs_extlen_t			freelen;
+	xfs_extlen_t			bnobt_sz;
+	xfs_extlen_t			inobt_sz;
+	xfs_extlen_t			rmapbt_sz;
+	xfs_extlen_t			refcbt_sz;
+	int				error;
+
+	if (!(sm->sm_flags & XFS_SCRUB_IFLAG_REPAIR))
+		return 0;
+
+	/* Use in-core counters if possible. */
+	pag = xfs_perag_get(mp, sm->sm_agno);
+	if (pag->pagi_init)
+		icount = pag->pagi_count;
+
+	/*
+	 * Otherwise try to get the actual counters from disk; if not, make
+	 * some worst case assumptions.
+	 */
+	if (icount == 0) {
+		error = xfs_ialloc_read_agi(mp, NULL, sm->sm_agno, &bp);
+		if (error) {
+			icount = mp->m_sb.sb_agblocks / mp->m_sb.sb_inopblock;
+		} else {
+			icount = pag->pagi_count;
+			xfs_buf_relse(bp);
+		}
+	}
+
+	/* Now grab the block counters from the AGF. */
+	error = xfs_alloc_read_agf(mp, NULL, sm->sm_agno, 0, &bp);
+	if (error) {
+		aglen = mp->m_sb.sb_agblocks;
+		freelen = aglen;
+		usedlen = aglen;
+	} else {
+		aglen = be32_to_cpu(XFS_BUF_TO_AGF(bp)->agf_length);
+		freelen = pag->pagf_freeblks;
+		usedlen = aglen - freelen;
+		xfs_buf_relse(bp);
+	}
+	xfs_perag_put(pag);
+
+	trace_xfs_repair_calc_ag_resblks(mp, sm->sm_agno, icount, aglen,
+			freelen, usedlen);
+
+	/*
+	 * Figure out how many blocks we'd need worst case to rebuild
+	 * each type of btree.  Note that we can only rebuild the
+	 * bnobt/cntbt or inobt/finobt as pairs.
+	 */
+	bnobt_sz = 2 * xfs_allocbt_calc_size(mp, freelen);
+	if (xfs_sb_version_hassparseinodes(&mp->m_sb))
+		inobt_sz = xfs_iallocbt_calc_size(mp, icount /
+				XFS_INODES_PER_HOLEMASK_BIT);
+	else
+		inobt_sz = xfs_iallocbt_calc_size(mp, icount /
+				XFS_INODES_PER_CHUNK);
+	if (xfs_sb_version_hasfinobt(&mp->m_sb))
+		inobt_sz *= 2;
+	if (xfs_sb_version_hasreflink(&mp->m_sb))
+		refcbt_sz = xfs_refcountbt_calc_size(mp, usedlen);
+	else
+		refcbt_sz = 0;
+	if (xfs_sb_version_hasrmapbt(&mp->m_sb)) {
+		/*
+		 * Guess how many blocks we need to rebuild the rmapbt.
+		 * For non-reflink filesystems we can't have more records than
+		 * used blocks.  However, with reflink it's possible to have
+		 * more than one rmap record per AG block.  We don't know how
+		 * many rmaps there could be in the AG, so we start off with
+		 * what we hope is an generous over-estimation.
+		 */
+		if (xfs_sb_version_hasreflink(&mp->m_sb))
+			rmapbt_sz = xfs_rmapbt_calc_size(mp,
+					(unsigned long long)aglen * 2);
+		else
+			rmapbt_sz = xfs_rmapbt_calc_size(mp, usedlen);
+	} else {
+		rmapbt_sz = 0;
+	}
+
+	trace_xfs_repair_calc_ag_resblks_btsize(mp, sm->sm_agno, bnobt_sz,
+			inobt_sz, rmapbt_sz, refcbt_sz);
+
+	return max(max(bnobt_sz, inobt_sz), max(rmapbt_sz, refcbt_sz));
+}
+
+/* Allocate a block in an AG. */
+int
+xfs_repair_alloc_ag_block(
+	struct xfs_scrub_context	*sc,
+	struct xfs_owner_info		*oinfo,
+	xfs_fsblock_t			*fsbno,
+	enum xfs_ag_resv_type		resv)
+{
+	struct xfs_alloc_arg		args = {0};
+	xfs_agblock_t			bno;
+	int				error;
+
+	switch (resv) {
+	case XFS_AG_RESV_AGFL:
+	case XFS_AG_RESV_RMAPBT:
+		error = xfs_alloc_get_freelist(sc->tp, sc->sa.agf_bp, &bno, 1);
+		if (error)
+			return error;
+		if (bno == NULLAGBLOCK)
+			return -ENOSPC;
+		xfs_extent_busy_reuse(sc->mp, sc->sa.agno, bno,
+				1, false);
+		*fsbno = XFS_AGB_TO_FSB(sc->mp, sc->sa.agno, bno);
+		if (resv == XFS_AG_RESV_RMAPBT)
+			xfs_ag_resv_rmapbt_alloc(sc->mp, sc->sa.agno);
+		return 0;
+	default:
+		break;
+	}
+
+	args.tp = sc->tp;
+	args.mp = sc->mp;
+	args.oinfo = *oinfo;
+	args.fsbno = XFS_AGB_TO_FSB(args.mp, sc->sa.agno, 0);
+	args.minlen = 1;
+	args.maxlen = 1;
+	args.prod = 1;
+	args.type = XFS_ALLOCTYPE_THIS_AG;
+	args.resv = resv;
+
+	error = xfs_alloc_vextent(&args);
+	if (error)
+		return error;
+	if (args.fsbno == NULLFSBLOCK)
+		return -ENOSPC;
+	ASSERT(args.len == 1);
+	*fsbno = args.fsbno;
+
+	return 0;
+}
+
+/* Initialize a new AG btree root block with zero entries. */
+int
+xfs_repair_init_btblock(
+	struct xfs_scrub_context	*sc,
+	xfs_fsblock_t			fsb,
+	struct xfs_buf			**bpp,
+	xfs_btnum_t			btnum,
+	const struct xfs_buf_ops	*ops)
+{
+	struct xfs_trans		*tp = sc->tp;
+	struct xfs_mount		*mp = sc->mp;
+	struct xfs_buf			*bp;
+
+	trace_xfs_repair_init_btblock(mp, XFS_FSB_TO_AGNO(mp, fsb),
+			XFS_FSB_TO_AGBNO(mp, fsb), btnum);
+
+	ASSERT(XFS_FSB_TO_AGNO(mp, fsb) == sc->sa.agno);
+	bp = xfs_trans_get_buf(tp, mp->m_ddev_targp, XFS_FSB_TO_DADDR(mp, fsb),
+			XFS_FSB_TO_BB(mp, 1), 0);
+	xfs_buf_zero(bp, 0, BBTOB(bp->b_length));
+	xfs_btree_init_block(mp, bp, btnum, 0, 0, sc->sa.agno, 0);
+	xfs_trans_buf_set_type(tp, bp, XFS_BLFT_BTREE_BUF);
+	xfs_trans_log_buf(tp, bp, 0, bp->b_length);
+	bp->b_ops = ops;
+	*bpp = bp;
+
+	return 0;
+}
+
+/*
+ * Reconstructing per-AG Btrees
+ *
+ * When a space btree is corrupt, we don't bother trying to fix it.  Instead,
+ * we scan secondary space metadata to derive the records that should be in
+ * the damaged btree, initialize a fresh btree root, and insert the records.
+ * Note that for rebuilding the rmapbt we scan all the primary data to
+ * generate the new records.
+ *
+ * However, that leaves the matter of removing all the metadata describing the
+ * old broken structure.  For primary metadata we use the rmap data to collect
+ * every extent with a matching rmap owner (exlist); we then iterate all other
+ * metadata structures with the same rmap owner to collect the extents that
+ * cannot be removed (sublist).  We then subtract sublist from exlist to
+ * derive the blocks that were used by the old btree.  These blocks can be
+ * reaped.
+ *
+ * For rmapbt reconstructions we must use different tactics for extent
+ * collection.  First we iterate all primary metadata (this excludes the old
+ * rmapbt, obviously) to generate new rmap records.  The gaps in the rmap
+ * records are collected as exlist.  The bnobt records are collected as
+ * sublist.  As with the other btrees we subtract sublist from exlist, and the
+ * result (since the rmapbt lives in the free space) are the blocks from the
+ * old rmapbt.
+ */
+
+/* Collect a dead btree extent for later disposal. */
+int
+xfs_repair_collect_btree_extent(
+	struct xfs_scrub_context	*sc,
+	struct xfs_repair_extent_list	*exlist,
+	xfs_fsblock_t			fsbno,
+	xfs_extlen_t			len)
+{
+	struct xfs_repair_extent	*rex;
+
+	trace_xfs_repair_collect_btree_extent(sc->mp,
+			XFS_FSB_TO_AGNO(sc->mp, fsbno),
+			XFS_FSB_TO_AGBNO(sc->mp, fsbno), len);
+
+	rex = kmem_alloc(sizeof(struct xfs_repair_extent), KM_MAYFAIL);
+	if (!rex)
+		return -ENOMEM;
+
+	INIT_LIST_HEAD(&rex->list);
+	rex->fsbno = fsbno;
+	rex->len = len;
+	list_add_tail(&rex->list, &exlist->list);
+
+	return 0;
+}
+
+/*
+ * An error happened during the rebuild so the transaction will be cancelled.
+ * The fs will shut down, and the administrator has to unmount and run repair.
+ * Therefore, free all the memory associated with the list so we can die.
+ */
+void
+xfs_repair_cancel_btree_extents(
+	struct xfs_scrub_context	*sc,
+	struct xfs_repair_extent_list	*exlist)
+{
+	struct xfs_repair_extent	*rex;
+	struct xfs_repair_extent	*n;
+
+	for_each_xfs_repair_extent_safe(rex, n, exlist) {
+		list_del(&rex->list);
+		kmem_free(rex);
+	}
+}
+
+/* Compare two btree extents. */
+static int
+xfs_repair_btree_extent_cmp(
+	void				*priv,
+	struct list_head		*a,
+	struct list_head		*b)
+{
+	struct xfs_repair_extent	*ap;
+	struct xfs_repair_extent	*bp;
+
+	ap = container_of(a, struct xfs_repair_extent, list);
+	bp = container_of(b, struct xfs_repair_extent, list);
+
+	if (ap->fsbno > bp->fsbno)
+		return 1;
+	if (ap->fsbno < bp->fsbno)
+		return -1;
+	return 0;
+}
+
+/*
+ * Remove all the blocks mentioned in @sublist from the extents in @exlist.
+ *
+ * The intent is that callers will iterate the rmapbt for all of its records
+ * for a given owner to generate @exlist; and iterate all the blocks of the
+ * metadata structures that are not being rebuilt and have the same rmapbt
+ * owner to generate @sublist.  This routine subtracts all the extents
+ * mentioned in sublist from all the extents linked in @exlist, which leaves
+ * @exlist as the list of blocks that are not accounted for, which we assume
+ * are the dead blocks of the old metadata structure.  The blocks mentioned in
+ * @exlist can be reaped.
+ */
+#define LEFT_ALIGNED	(1 << 0)
+#define RIGHT_ALIGNED	(1 << 1)
+int
+xfs_repair_subtract_extents(
+	struct xfs_scrub_context	*sc,
+	struct xfs_repair_extent_list	*exlist,
+	struct xfs_repair_extent_list	*sublist)
+{
+	struct list_head		*lp;
+	struct xfs_repair_extent	*ex;
+	struct xfs_repair_extent	*newex;
+	struct xfs_repair_extent	*subex;
+	xfs_fsblock_t			sub_fsb;
+	xfs_extlen_t			sub_len;
+	int				state;
+	int				error = 0;
+
+	if (list_empty(&exlist->list) || list_empty(&sublist->list))
+		return 0;
+	ASSERT(!list_empty(&sublist->list));
+
+	list_sort(NULL, &exlist->list, xfs_repair_btree_extent_cmp);
+	list_sort(NULL, &sublist->list, xfs_repair_btree_extent_cmp);
+
+	/*
+	 * Now that we've sorted both lists, we iterate exlist once, rolling
+	 * forward through sublist and/or exlist as necessary until we find an
+	 * overlap or reach the end of either list.  We do not reset lp to the
+	 * head of exlist nor do we reset subex to the head of sublist.  The
+	 * list traversal is similar to merge sort, but we're deleting
+	 * instead.  In this manner we avoid O(n^2) operations.
+	 */
+	subex = list_first_entry(&sublist->list, struct xfs_repair_extent,
+			list);
+	lp = exlist->list.next;
+	while (lp != &exlist->list) {
+		ex = list_entry(lp, struct xfs_repair_extent, list);
+
+		/*
+		 * Advance subex and/or ex until we find a pair that
+		 * intersect or we run out of extents.
+		 */
+		while (subex->fsbno + subex->len <= ex->fsbno) {
+			if (list_is_last(&subex->list, &sublist->list))
+				goto out;
+			subex = list_next_entry(subex, list);
+		}
+		if (subex->fsbno >= ex->fsbno + ex->len) {
+			lp = lp->next;
+			continue;
+		}
+
+		/* trim subex to fit the extent we have */
+		sub_fsb = subex->fsbno;
+		sub_len = subex->len;
+		if (subex->fsbno < ex->fsbno) {
+			sub_len -= ex->fsbno - subex->fsbno;
+			sub_fsb = ex->fsbno;
+		}
+		if (sub_len > ex->len)
+			sub_len = ex->len;
+
+		state = 0;
+		if (sub_fsb == ex->fsbno)
+			state |= LEFT_ALIGNED;
+		if (sub_fsb + sub_len == ex->fsbno + ex->len)
+			state |= RIGHT_ALIGNED;
+		switch (state) {
+		case LEFT_ALIGNED:
+			/* Coincides with only the left. */
+			ex->fsbno += sub_len;
+			ex->len -= sub_len;
+			break;
+		case RIGHT_ALIGNED:
+			/* Coincides with only the right. */
+			ex->len -= sub_len;
+			lp = lp->next;
+			break;
+		case LEFT_ALIGNED | RIGHT_ALIGNED:
+			/* Total overlap, just delete ex. */
+			lp = lp->next;
+			list_del(&ex->list);
+			kmem_free(ex);
+			break;
+		case 0:
+			/*
+			 * Deleting from the middle: add the new right extent
+			 * and then shrink the left extent.
+			 */
+			newex = kmem_alloc(sizeof(struct xfs_repair_extent),
+					KM_MAYFAIL);
+			if (!newex) {
+				error = -ENOMEM;
+				goto out;
+			}
+			INIT_LIST_HEAD(&newex->list);
+			newex->fsbno = sub_fsb + sub_len;
+			newex->len = ex->fsbno + ex->len - newex->fsbno;
+			list_add(&newex->list, &ex->list);
+			ex->len = sub_fsb - ex->fsbno;
+			lp = lp->next;
+			break;
+		default:
+			ASSERT(0);
+			break;
+		}
+	}
+
+out:
+	return error;
+}
+#undef LEFT_ALIGNED
+#undef RIGHT_ALIGNED
+
+/*
+ * Disposal of Blocks from Old per-AG Btrees
+ *
+ * Now that we've constructed a new btree to replace the damaged one, we want
+ * to dispose of the blocks that (we think) the old btree was using.
+ * Previously, we used the rmapbt to collect the extents (exlist) with the
+ * rmap owner corresponding to the tree we rebuilt, collected extents for any
+ * blocks with the same rmap owner that are owned by another data structure
+ * (sublist), and subtracted sublist from exlist.  In theory the extents
+ * remaining in exlist are the old btree's blocks.
+ *
+ * Unfortunately, it's possible that the btree was crosslinked with other
+ * blocks on disk.  The rmap data can tell us if there are multiple owners, so
+ * if the rmapbt says there is an owner of this block other than @oinfo, then
+ * the block is crosslinked.  Remove the reverse mapping and continue.
+ *
+ * If there is one rmap record, we can free the block, which removes the
+ * reverse mapping but doesn't add the block to the free space.  Our repair
+ * strategy is to hope the other metadata objects crosslinked on this block
+ * will be rebuilt (atop different blocks), thereby removing all the cross
+ * links.
+ *
+ * If there are no rmap records at all, we also free the block.  If the btree
+ * being rebuilt lives in the free space (bnobt/cntbt/rmapbt) then there isn't
+ * supposed to be a rmap record and everything is ok.  For other btrees there
+ * had to have been an rmap entry for the block to have ended up on @exlist,
+ * so if it's gone now there's something wrong and the fs will shut down.
+ *
+ * Note: If there are multiple rmap records with only the same rmap owner as
+ * the btree we're trying to rebuild and the block is indeed owned by another
+ * data structure with the same rmap owner, then the block will be in sublist
+ * and therefore doesn't need disposal.  If there are multiple rmap records
+ * with only the same rmap owner but the block is not owned by something with
+ * the same rmap owner, the block will be freed.
+ *
+ * The caller is responsible for locking the AG headers for the entire rebuild
+ * operation so that nothing else can sneak in and change the AG state while
+ * we're not looking.  We also assume that the caller already invalidated any
+ * buffers associated with @exlist.
+ */
+
+/*
+ * Invalidate buffers for per-AG btree blocks we're dumping.  This function
+ * is not intended for use with file data repairs; we have bunmapi for that.
+ */
+int
+xfs_repair_invalidate_blocks(
+	struct xfs_scrub_context	*sc,
+	struct xfs_repair_extent_list	*exlist)
+{
+	struct xfs_repair_extent	*rex;
+	struct xfs_repair_extent	*n;
+	struct xfs_buf			*bp;
+	xfs_fsblock_t			fsbno;
+	xfs_agblock_t			i;
+
+	/*
+	 * For each block in each extent, see if there's an incore buffer for
+	 * exactly that block; if so, invalidate it.  The buffer cache only
+	 * lets us look for one buffer at a time, so we have to look one block
+	 * at a time.  Avoid invalidating AG headers and post-EOFS blocks
+	 * because we never own those; and if we can't TRYLOCK the buffer we
+	 * assume it's owned by someone else.
+	 */
+	for_each_xfs_repair_extent_safe(rex, n, exlist) {
+		for (fsbno = rex->fsbno, i = rex->len; i > 0; fsbno++, i--) {
+			/* Skip AG headers and post-EOFS blocks */
+			if (!xfs_verify_fsbno(sc->mp, fsbno))
+				continue;
+			bp = xfs_buf_incore(sc->mp->m_ddev_targp,
+					XFS_FSB_TO_DADDR(sc->mp, fsbno),
+					XFS_FSB_TO_BB(sc->mp, 1), XBF_TRYLOCK);
+			if (bp) {
+				xfs_trans_bjoin(sc->tp, bp);
+				xfs_trans_binval(sc->tp, bp);
+			}
+		}
+	}
+
+	return 0;
+}
+
+/* Ensure the freelist is the correct size. */
+int
+xfs_repair_fix_freelist(
+	struct xfs_scrub_context	*sc,
+	bool				can_shrink)
+{
+	struct xfs_alloc_arg		args = {0};
+
+	args.mp = sc->mp;
+	args.tp = sc->tp;
+	args.agno = sc->sa.agno;
+	args.alignment = 1;
+	args.pag = sc->sa.pag;
+
+	return xfs_alloc_fix_freelist(&args,
+			can_shrink ? 0 : XFS_ALLOC_FLAG_NOSHRINK);
+}
+
+/*
+ * Put a block back on the AGFL.
+ */
+STATIC int
+xfs_repair_put_freelist(
+	struct xfs_scrub_context	*sc,
+	xfs_agblock_t			agbno)
+{
+	struct xfs_owner_info		oinfo;
+	int				error;
+
+	/* Make sure there's space on the freelist. */
+	error = xfs_repair_fix_freelist(sc, true);
+	if (error)
+		return error;
+
+	/*
+	 * Since we're "freeing" a lost block onto the AGFL, we have to
+	 * create an rmap for the block prior to merging it or else other
+	 * parts will break.
+	 */
+	xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_AG);
+	error = xfs_rmap_alloc(sc->tp, sc->sa.agf_bp, sc->sa.agno, agbno, 1,
+			&oinfo);
+	if (error)
+		return error;
+
+	/* Put the block on the AGFL. */
+	error = xfs_alloc_put_freelist(sc->tp, sc->sa.agf_bp, sc->sa.agfl_bp,
+			agbno, 0);
+	if (error)
+		return error;
+	xfs_extent_busy_insert(sc->tp, sc->sa.agno, agbno, 1,
+			XFS_EXTENT_BUSY_SKIP_DISCARD);
+
+	return 0;
+}
+
+/* Dispose of a single metadata block. */
+STATIC int
+xfs_repair_dispose_btree_block(
+	struct xfs_scrub_context	*sc,
+	xfs_fsblock_t			fsbno,
+	struct xfs_owner_info		*oinfo,
+	enum xfs_ag_resv_type		resv)
+{
+	struct xfs_btree_cur		*cur;
+	struct xfs_buf			*agf_bp = NULL;
+	xfs_agnumber_t			agno;
+	xfs_agblock_t			agbno;
+	bool				has_other_rmap;
+	int				error;
+
+	agno = XFS_FSB_TO_AGNO(sc->mp, fsbno);
+	agbno = XFS_FSB_TO_AGBNO(sc->mp, fsbno);
+
+	/*
+	 * If we are repairing per-inode metadata, we need to read in the AGF
+	 * buffer.  Otherwise, we're repairing a per-AG structure, so reuse
+	 * the AGF buffer that the setup functions already grabbed.
+	 */
+	if (sc->ip) {
+		error = xfs_alloc_read_agf(sc->mp, sc->tp, agno, 0, &agf_bp);
+		if (error)
+			return error;
+		if (!agf_bp)
+			return -ENOMEM;
+	} else {
+		agf_bp = sc->sa.agf_bp;
+	}
+	cur = xfs_rmapbt_init_cursor(sc->mp, sc->tp, agf_bp, agno);
+
+	/* Can we find any other rmappings? */
+	error = xfs_rmap_has_other_keys(cur, agbno, 1, oinfo, &has_other_rmap);
+	if (error)
+		goto out_cur;
+	xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
+
+	/*
+	 * If there are other rmappings, this block is cross linked and must
+	 * not be freed.  Remove the reverse mapping and move on.  Otherwise,
+	 * we were the only owner of the block, so free the extent, which will
+	 * also remove the rmap.
+	 *
+	 * XXX: XFS doesn't support detecting the case where a single block
+	 * metadata structure is crosslinked with a multi-block structure
+	 * because the buffer cache doesn't detect aliasing problems, so we
+	 * can't fix 100% of crosslinking problems (yet).  The verifiers will
+	 * blow on writeout, the filesystem will shut down, and the admin gets
+	 * to run xfs_repair.
+	 */
+	if (has_other_rmap)
+		error = xfs_rmap_free(sc->tp, agf_bp, agno, agbno, 1, oinfo);
+	else if (resv == XFS_AG_RESV_AGFL)
+		error = xfs_repair_put_freelist(sc, agbno);
+	else
+		error = xfs_free_extent(sc->tp, fsbno, 1, oinfo, resv);
+	if (agf_bp != sc->sa.agf_bp)
+		xfs_trans_brelse(sc->tp, agf_bp);
+	if (error)
+		return error;
+
+	if (sc->ip)
+		return xfs_trans_roll_inode(&sc->tp, sc->ip);
+	return xfs_repair_roll_ag_trans(sc);
+
+out_cur:
+	xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
+	if (agf_bp != sc->sa.agf_bp)
+		xfs_trans_brelse(sc->tp, agf_bp);
+	return error;
+}
+
+/* Dispose of btree blocks from an old per-AG btree. */
+int
+xfs_repair_reap_btree_extents(
+	struct xfs_scrub_context	*sc,
+	struct xfs_repair_extent_list	*exlist,
+	struct xfs_owner_info		*oinfo,
+	enum xfs_ag_resv_type		type)
+{
+	struct xfs_repair_extent	*rex;
+	struct xfs_repair_extent	*n;
+	int				error = 0;
+
+	ASSERT(xfs_sb_version_hasrmapbt(&sc->mp->m_sb));
+
+	/* Dispose of every block from the old btree. */
+	for_each_xfs_repair_extent_safe(rex, n, exlist) {
+		ASSERT(sc->ip != NULL ||
+		       XFS_FSB_TO_AGNO(sc->mp, rex->fsbno) == sc->sa.agno);
+
+		trace_xfs_repair_dispose_btree_extent(sc->mp,
+				XFS_FSB_TO_AGNO(sc->mp, rex->fsbno),
+				XFS_FSB_TO_AGBNO(sc->mp, rex->fsbno), rex->len);
+
+		for (; rex->len > 0; rex->len--, rex->fsbno++) {
+			error = xfs_repair_dispose_btree_block(sc, rex->fsbno,
+					oinfo, type);
+			if (error)
+				goto out;
+		}
+		list_del(&rex->list);
+		kmem_free(rex);
+	}
+
+out:
+	xfs_repair_cancel_btree_extents(sc, exlist);
+	return error;
+}
+
+/*
+ * Finding per-AG Btree Roots for AGF/AGI Reconstruction
+ *
+ * If the AGF or AGI become slightly corrupted, it may be necessary to rebuild
+ * the AG headers by using the rmap data to rummage through the AG looking for
+ * btree roots.  This is not guaranteed to work if the AG is heavily damaged
+ * or the rmap data are corrupt.
+ *
+ * Callers of xfs_repair_find_ag_btree_roots must lock the AGF and AGFL
+ * buffers if the AGF is being rebuilt; or the AGF and AGI buffers if the
+ * AGI is being rebuilt.  It must maintain these locks until it's safe for
+ * other threads to change the btrees' shapes.  The caller provides
+ * information about the btrees to look for by passing in an array of
+ * xfs_repair_find_ag_btree with the (rmap owner, buf_ops, magic) fields set.
+ * The (root, height) fields will be set on return if anything is found.  The
+ * last element of the array should have a NULL buf_ops to mark the end of the
+ * array.
+ *
+ * For every rmapbt record matching any of the rmap owners in btree_info,
+ * read each block referenced by the rmap record.  If the block is a btree
+ * block from this filesystem matching any of the magic numbers and has a
+ * level higher than what we've already seen, remember the block and the
+ * height of the tree required to have such a block.  When the call completes,
+ * we return the highest block we've found for each btree description; those
+ * should be the roots.
+ */
+
+struct xfs_repair_findroot {
+	struct xfs_scrub_context	*sc;
+	struct xfs_buf			*agfl_bp;
+	struct xfs_agf			*agf;
+	struct xfs_repair_find_ag_btree	*btree_info;
+};
+
+/* See if our block is in the AGFL. */
+STATIC int
+xfs_repair_findroot_agfl_walk(
+	struct xfs_mount		*mp,
+	xfs_agblock_t			bno,
+	void				*priv)
+{
+	xfs_agblock_t			*agbno = priv;
+
+	return (*agbno == bno) ? XFS_BTREE_QUERY_RANGE_ABORT : 0;
+}
+
+/* Does this block match the btree information passed in? */
+STATIC int
+xfs_repair_findroot_block(
+	struct xfs_repair_findroot	*ri,
+	struct xfs_repair_find_ag_btree	*fab,
+	uint64_t			owner,
+	xfs_agblock_t			agbno,
+	bool				*found_it)
+{
+	struct xfs_mount		*mp = ri->sc->mp;
+	struct xfs_buf			*bp;
+	struct xfs_btree_block		*btblock;
+	xfs_daddr_t			daddr;
+	int				error;
+
+	daddr = XFS_AGB_TO_DADDR(mp, ri->sc->sa.agno, agbno);
+
+	/*
+	 * Blocks in the AGFL have stale contents that might just happen to
+	 * have a matching magic and uuid.  We don't want to pull these blocks
+	 * in as part of a tree root, so we have to filter out the AGFL stuff
+	 * here.  If the AGFL looks insane we'll just refuse to repair.
+	 */
+	if (owner == XFS_RMAP_OWN_AG) {
+		error = xfs_agfl_walk(mp, ri->agf, ri->agfl_bp,
+				xfs_repair_findroot_agfl_walk, &agbno);
+		if (error == XFS_BTREE_QUERY_RANGE_ABORT)
+			return 0;
+		if (error)
+			return error;
+	}
+
+	error = xfs_trans_read_buf(mp, ri->sc->tp, mp->m_ddev_targp, daddr,
+			mp->m_bsize, 0, &bp, NULL);
+	if (error)
+		return error;
+
+	/*
+	 * Does this look like a block matching our fs and higher than any
+	 * other block we've found so far?  If so, reattach buffer verifiers
+	 * so the AIL won't complain if the buffer is also dirty.
+	 */
+	btblock = XFS_BUF_TO_BLOCK(bp);
+	if (be32_to_cpu(btblock->bb_magic) != fab->magic)
+		goto out;
+	if (xfs_sb_version_hascrc(&mp->m_sb) &&
+	    !uuid_equal(&btblock->bb_u.s.bb_uuid, &mp->m_sb.sb_meta_uuid))
+		goto out;
+	bp->b_ops = fab->buf_ops;
+
+	/* Ignore this block if it's lower in the tree than we've seen. */
+	if (fab->root != NULLAGBLOCK &&
+	    xfs_btree_get_level(btblock) < fab->height)
+		goto out;
+
+	/* Make sure we pass the verifiers. */
+	bp->b_ops->verify_read(bp);
+	if (bp->b_error)
+		goto out;
+	fab->root = agbno;
+	fab->height = xfs_btree_get_level(btblock) + 1;
+	*found_it = true;
+
+	trace_xfs_repair_findroot_block(mp, ri->sc->sa.agno, agbno,
+			be32_to_cpu(btblock->bb_magic), fab->height - 1);
+out:
+	xfs_trans_brelse(ri->sc->tp, bp);
+	return error;
+}
+
+/*
+ * Do any of the blocks in this rmap record match one of the btrees we're
+ * looking for?
+ */
+STATIC int
+xfs_repair_findroot_rmap(
+	struct xfs_btree_cur		*cur,
+	struct xfs_rmap_irec		*rec,
+	void				*priv)
+{
+	struct xfs_repair_findroot	*ri = priv;
+	struct xfs_repair_find_ag_btree	*fab;
+	xfs_agblock_t			b;
+	bool				found_it;
+	int				error = 0;
+
+	/* Ignore anything that isn't AG metadata. */
+	if (!XFS_RMAP_NON_INODE_OWNER(rec->rm_owner))
+		return 0;
+
+	/* Otherwise scan each block + btree type. */
+	for (b = 0; b < rec->rm_blockcount; b++) {
+		found_it = false;
+		for (fab = ri->btree_info; fab->buf_ops; fab++) {
+			if (rec->rm_owner != fab->rmap_owner)
+				continue;
+			error = xfs_repair_findroot_block(ri, fab,
+					rec->rm_owner, rec->rm_startblock + b,
+					&found_it);
+			if (error)
+				return error;
+			if (found_it)
+				break;
+		}
+	}
+
+	return 0;
+}
+
+/* Find the roots of the per-AG btrees described in btree_info. */
+int
+xfs_repair_find_ag_btree_roots(
+	struct xfs_scrub_context	*sc,
+	struct xfs_buf			*agf_bp,
+	struct xfs_repair_find_ag_btree	*btree_info,
+	struct xfs_buf			*agfl_bp)
+{
+	struct xfs_mount		*mp = sc->mp;
+	struct xfs_repair_findroot	ri;
+	struct xfs_repair_find_ag_btree	*fab;
+	struct xfs_btree_cur		*cur;
+	int				error;
+
+	ASSERT(xfs_buf_islocked(agf_bp));
+	ASSERT(agfl_bp == NULL || xfs_buf_islocked(agfl_bp));
+
+	ri.sc = sc;
+	ri.btree_info = btree_info;
+	ri.agf = XFS_BUF_TO_AGF(agf_bp);
+	ri.agfl_bp = agfl_bp;
+	for (fab = btree_info; fab->buf_ops; fab++) {
+		ASSERT(agfl_bp || fab->rmap_owner != XFS_RMAP_OWN_AG);
+		ASSERT(XFS_RMAP_NON_INODE_OWNER(fab->rmap_owner));
+		fab->root = NULLAGBLOCK;
+		fab->height = 0;
+	}
+
+	cur = xfs_rmapbt_init_cursor(mp, sc->tp, agf_bp, sc->sa.agno);
+	error = xfs_rmap_query_all(cur, xfs_repair_findroot_rmap, &ri);
+	xfs_btree_del_cursor(cur, error ? XFS_BTREE_ERROR : XFS_BTREE_NOERROR);
+
+	return error;
+}
+
+/* Force a quotacheck the next time we mount. */
+void
+xfs_repair_force_quotacheck(
+	struct xfs_scrub_context	*sc,
+	uint				dqtype)
+{
+	uint				flag;
+
+	flag = xfs_quota_chkd_flag(dqtype);
+	if (!(flag & sc->mp->m_qflags))
+		return;
+
+	sc->mp->m_qflags &= ~flag;
+	spin_lock(&sc->mp->m_sb_lock);
+	sc->mp->m_sb.sb_qflags &= ~flag;
+	spin_unlock(&sc->mp->m_sb_lock);
+	xfs_log_sb(sc->tp);
+}
+
+/*
+ * Attach dquots to this inode, or schedule quotacheck to fix them.
+ *
+ * This function ensures that the appropriate dquots are attached to an inode.
+ * We cannot allow the dquot code to allocate an on-disk dquot block here
+ * because we're already in transaction context with the inode locked.  The
+ * on-disk dquot should already exist anyway.  If the quota code signals
+ * corruption or missing quota information, schedule quotacheck, which will
+ * repair corruptions in the quota metadata.
+ */
+int
+xfs_repair_ino_dqattach(
+	struct xfs_scrub_context	*sc)
+{
+	int				error;
+
+	error = xfs_qm_dqattach_locked(sc->ip, false);
+	switch (error) {
+	case -EFSBADCRC:
+	case -EFSCORRUPTED:
+	case -ENOENT:
+		xfs_err_ratelimited(sc->mp,
+"inode %llu repair encountered quota error %d, quotacheck forced.",
+				(unsigned long long)sc->ip->i_ino, error);
+		if (XFS_IS_UQUOTA_ON(sc->mp) && !sc->ip->i_udquot)
+			xfs_repair_force_quotacheck(sc, XFS_DQ_USER);
+		if (XFS_IS_GQUOTA_ON(sc->mp) && !sc->ip->i_gdquot)
+			xfs_repair_force_quotacheck(sc, XFS_DQ_GROUP);
+		if (XFS_IS_PQUOTA_ON(sc->mp) && !sc->ip->i_pdquot)
+			xfs_repair_force_quotacheck(sc, XFS_DQ_PROJ);
+		/* fall through */
+	case -ESRCH:
+		error = 0;
+		break;
+	default:
+		break;
+	}
+
+	return error;
+}
diff --git a/fs/xfs/scrub/repair.h b/fs/xfs/scrub/repair.h
new file mode 100644
index 000000000000..f2b0895294db
--- /dev/null
+++ b/fs/xfs/scrub/repair.h
@@ -0,0 +1,132 @@
+/*
+ * Copyright (C) 2018 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#ifndef __XFS_SCRUB_REPAIR_H__
+#define __XFS_SCRUB_REPAIR_H__
+
+static inline int xfs_repair_notsupported(struct xfs_scrub_context *sc)
+{
+	return -EOPNOTSUPP;
+}
+
+#ifdef CONFIG_XFS_ONLINE_REPAIR
+
+/* Repair helpers */
+
+int xfs_repair_attempt(struct xfs_inode *ip, struct xfs_scrub_context *sc,
+		bool *fixed);
+void xfs_repair_failure(struct xfs_mount *mp);
+int xfs_repair_roll_ag_trans(struct xfs_scrub_context *sc);
+bool xfs_repair_ag_has_space(struct xfs_perag *pag, xfs_extlen_t nr_blocks,
+		enum xfs_ag_resv_type type);
+xfs_extlen_t xfs_repair_calc_ag_resblks(struct xfs_scrub_context *sc);
+int xfs_repair_alloc_ag_block(struct xfs_scrub_context *sc,
+		struct xfs_owner_info *oinfo, xfs_fsblock_t *fsbno,
+		enum xfs_ag_resv_type resv);
+int xfs_repair_init_btblock(struct xfs_scrub_context *sc, xfs_fsblock_t fsb,
+		struct xfs_buf **bpp, xfs_btnum_t btnum,
+		const struct xfs_buf_ops *ops);
+
+struct xfs_repair_extent {
+	struct list_head		list;
+	xfs_fsblock_t			fsbno;
+	xfs_extlen_t			len;
+};
+
+struct xfs_repair_extent_list {
+	struct list_head		list;
+};
+
+static inline void
+xfs_repair_init_extent_list(
+	struct xfs_repair_extent_list	*exlist)
+{
+	INIT_LIST_HEAD(&exlist->list);
+}
+
+#define for_each_xfs_repair_extent_safe(rbe, n, exlist) \
+	list_for_each_entry_safe((rbe), (n), &(exlist)->list, list)
+int xfs_repair_collect_btree_extent(struct xfs_scrub_context *sc,
+		struct xfs_repair_extent_list *btlist, xfs_fsblock_t fsbno,
+		xfs_extlen_t len);
+void xfs_repair_cancel_btree_extents(struct xfs_scrub_context *sc,
+		struct xfs_repair_extent_list *btlist);
+int xfs_repair_subtract_extents(struct xfs_scrub_context *sc,
+		struct xfs_repair_extent_list *exlist,
+		struct xfs_repair_extent_list *sublist);
+int xfs_repair_fix_freelist(struct xfs_scrub_context *sc, bool can_shrink);
+int xfs_repair_invalidate_blocks(struct xfs_scrub_context *sc,
+		struct xfs_repair_extent_list *btlist);
+int xfs_repair_reap_btree_extents(struct xfs_scrub_context *sc,
+		struct xfs_repair_extent_list *exlist,
+		struct xfs_owner_info *oinfo, enum xfs_ag_resv_type type);
+
+struct xfs_repair_find_ag_btree {
+	/* in: rmap owner of the btree we're looking for */
+	uint64_t			rmap_owner;
+
+	/* in: buffer ops */
+	const struct xfs_buf_ops	*buf_ops;
+
+	/* in: magic number of the btree */
+	uint32_t			magic;
+
+	/* out: the highest btree block found and the tree height */
+	xfs_agblock_t			root;
+	unsigned int			height;
+};
+
+int xfs_repair_find_ag_btree_roots(struct xfs_scrub_context *sc,
+		struct xfs_buf *agf_bp,
+		struct xfs_repair_find_ag_btree *btree_info,
+		struct xfs_buf *agfl_bp);
+void xfs_repair_force_quotacheck(struct xfs_scrub_context *sc, uint dqtype);
+int xfs_repair_ino_dqattach(struct xfs_scrub_context *sc);
+
+/* Metadata repairers */
+
+int xfs_repair_probe(struct xfs_scrub_context *sc);
+int xfs_repair_superblock(struct xfs_scrub_context *sc);
+
+#else
+
+static inline int xfs_repair_attempt(
+	struct xfs_inode		*ip,
+	struct xfs_scrub_context	*sc,
+	bool				*fixed)
+{
+	return -EOPNOTSUPP;
+}
+
+static inline void xfs_repair_failure(struct xfs_mount *mp) {}
+
+static inline xfs_extlen_t
+xfs_repair_calc_ag_resblks(
+	struct xfs_scrub_context	*sc)
+{
+	ASSERT(!(sc->sm->sm_flags & XFS_SCRUB_IFLAG_REPAIR));
+	return 0;
+}
+
+#define xfs_repair_probe		xfs_repair_notsupported
+#define xfs_repair_superblock		xfs_repair_notsupported
+
+#endif /* CONFIG_XFS_ONLINE_REPAIR */
+
+#endif	/* __XFS_SCRUB_REPAIR_H__ */
diff --git a/fs/xfs/scrub/rmap.c b/fs/xfs/scrub/rmap.c
index 8f2a7c3ff455..b376a9a77c04 100644
--- a/fs/xfs/scrub/rmap.c
+++ b/fs/xfs/scrub/rmap.c
@@ -66,7 +66,7 @@ xfs_scrub_rmapbt_xref_refc(
 	bool				is_unwritten;
 	int				error;
 
-	if (!sc->sa.refc_cur)
+	if (!sc->sa.refc_cur || xfs_scrub_skip_xref(sc->sm))
 		return;
 
 	non_inode = XFS_RMAP_NON_INODE_OWNER(irec->rm_owner);
@@ -207,7 +207,7 @@ xfs_scrub_xref_check_owner(
 	bool				has_rmap;
 	int				error;
 
-	if (!sc->sa.rmap_cur)
+	if (!sc->sa.rmap_cur || xfs_scrub_skip_xref(sc->sm))
 		return;
 
 	error = xfs_rmap_record_exists(sc->sa.rmap_cur, bno, len, oinfo,
@@ -250,7 +250,7 @@ xfs_scrub_xref_has_no_owner(
 	bool				has_rmap;
 	int				error;
 
-	if (!sc->sa.rmap_cur)
+	if (!sc->sa.rmap_cur || xfs_scrub_skip_xref(sc->sm))
 		return;
 
 	error = xfs_rmap_has_record(sc->sa.rmap_cur, bno, len, &has_rmap);
diff --git a/fs/xfs/scrub/rtbitmap.c b/fs/xfs/scrub/rtbitmap.c
index 39c41dfe08ee..40f462a11ea5 100644
--- a/fs/xfs/scrub/rtbitmap.c
+++ b/fs/xfs/scrub/rtbitmap.c
@@ -66,11 +66,15 @@ xfs_scrub_rtbitmap_rec(
 	void				*priv)
 {
 	struct xfs_scrub_context	*sc = priv;
+	xfs_rtblock_t			startblock;
+	xfs_rtblock_t			blockcount;
 
-	if (rec->ar_startblock + rec->ar_blockcount <= rec->ar_startblock ||
-	    !xfs_verify_rtbno(sc->mp, rec->ar_startblock) ||
-	    !xfs_verify_rtbno(sc->mp, rec->ar_startblock +
-			rec->ar_blockcount - 1))
+	startblock = rec->ar_startext * tp->t_mountp->m_sb.sb_rextsize;
+	blockcount = rec->ar_extcount * tp->t_mountp->m_sb.sb_rextsize;
+
+	if (startblock + blockcount <= startblock ||
+	    !xfs_verify_rtbno(sc->mp, startblock) ||
+	    !xfs_verify_rtbno(sc->mp, startblock + blockcount - 1))
 		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, 0);
 	return 0;
 }
@@ -82,6 +86,11 @@ xfs_scrub_rtbitmap(
 {
 	int				error;
 
+	/* Invoke the fork scrubber. */
+	error = xfs_scrub_metadata_inode_forks(sc);
+	if (error || (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT))
+		return error;
+
 	error = xfs_rtalloc_query_all(sc->tp, xfs_scrub_rtbitmap_rec, sc);
 	if (!xfs_scrub_fblock_process_error(sc, XFS_DATA_FORK, 0, &error))
 		goto out;
@@ -95,8 +104,35 @@ int
 xfs_scrub_rtsummary(
 	struct xfs_scrub_context	*sc)
 {
+	struct xfs_inode		*rsumip = sc->mp->m_rsumip;
+	struct xfs_inode		*old_ip = sc->ip;
+	uint				old_ilock_flags = sc->ilock_flags;
+	int				error = 0;
+
+	/*
+	 * We ILOCK'd the rt bitmap ip in the setup routine, now lock the
+	 * rt summary ip in compliance with the rt inode locking rules.
+	 *
+	 * Since we switch sc->ip to rsumip we have to save the old ilock
+	 * flags so that we don't mix up the inode state that @sc tracks.
+	 */
+	sc->ip = rsumip;
+	sc->ilock_flags = XFS_ILOCK_EXCL | XFS_ILOCK_RTSUM;
+	xfs_ilock(sc->ip, sc->ilock_flags);
+
+	/* Invoke the fork scrubber. */
+	error = xfs_scrub_metadata_inode_forks(sc);
+	if (error || (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT))
+		goto out;
+
 	/* XXX: implement this some day */
-	return -ENOENT;
+	xfs_scrub_set_incomplete(sc);
+out:
+	/* Switch back to the rtbitmap inode and lock flags. */
+	xfs_iunlock(sc->ip, sc->ilock_flags);
+	sc->ilock_flags = old_ilock_flags;
+	sc->ip = old_ip;
+	return error;
 }
 
 
@@ -107,11 +143,23 @@ xfs_scrub_xref_is_used_rt_space(
 	xfs_rtblock_t			fsbno,
 	xfs_extlen_t			len)
 {
+	xfs_rtblock_t			startext;
+	xfs_rtblock_t			endext;
+	xfs_rtblock_t			extcount;
 	bool				is_free;
 	int				error;
 
+	if (xfs_scrub_skip_xref(sc->sm))
+		return;
+
+	startext = fsbno;
+	endext = fsbno + len - 1;
+	do_div(startext, sc->mp->m_sb.sb_rextsize);
+	if (do_div(endext, sc->mp->m_sb.sb_rextsize))
+		endext++;
+	extcount = endext - startext;
 	xfs_ilock(sc->mp->m_rbmip, XFS_ILOCK_SHARED | XFS_ILOCK_RTBITMAP);
-	error = xfs_rtalloc_extent_is_free(sc->mp, sc->tp, fsbno, len,
+	error = xfs_rtalloc_extent_is_free(sc->mp, sc->tp, startext, extcount,
 			&is_free);
 	if (!xfs_scrub_should_check_xref(sc, &error, NULL))
 		goto out_unlock;
diff --git a/fs/xfs/scrub/scrub.c b/fs/xfs/scrub/scrub.c
index 26c75967a072..36db098ba583 100644
--- a/fs/xfs/scrub/scrub.c
+++ b/fs/xfs/scrub/scrub.c
@@ -42,11 +42,18 @@
 #include "xfs_refcount_btree.h"
 #include "xfs_rmap.h"
 #include "xfs_rmap_btree.h"
+#include "xfs_quota.h"
+#include "xfs_qm.h"
+#include "xfs_errortag.h"
+#include "xfs_error.h"
+#include "xfs_log.h"
+#include "xfs_trans_priv.h"
 #include "scrub/xfs_scrub.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 #include "scrub/trace.h"
 #include "scrub/btree.h"
+#include "scrub/repair.h"
 
 /*
  * Online Scrub and Repair
@@ -120,6 +127,24 @@
  * XCORRUPT flag; btree query function errors are noted by setting the
  * XFAIL flag and deleting the cursor to prevent further attempts to
  * cross-reference with a defective btree.
+ *
+ * If a piece of metadata proves corrupt or suboptimal, the userspace
+ * program can ask the kernel to apply some tender loving care (TLC) to
+ * the metadata object by setting the REPAIR flag and re-calling the
+ * scrub ioctl.  "Corruption" is defined by metadata violating the
+ * on-disk specification; operations cannot continue if the violation is
+ * left untreated.  It is possible for XFS to continue if an object is
+ * "suboptimal", however performance may be degraded.  Repairs are
+ * usually performed by rebuilding the metadata entirely out of
+ * redundant metadata.  Optimizing, on the other hand, can sometimes be
+ * done without rebuilding entire structures.
+ *
+ * Generally speaking, the repair code has the following code structure:
+ * Lock -> scrub -> repair -> commit -> re-lock -> re-scrub -> unlock.
+ * The first check helps us figure out if we need to rebuild or simply
+ * optimize the structure so that the rebuild knows what to do.  The
+ * second check evaluates the completeness of the repair; that is what
+ * is reported to userspace.
  */
 
 /*
@@ -155,7 +180,10 @@ xfs_scrub_teardown(
 {
 	xfs_scrub_ag_free(sc, &sc->sa);
 	if (sc->tp) {
-		xfs_trans_cancel(sc->tp);
+		if (error == 0 && (sc->sm->sm_flags & XFS_SCRUB_IFLAG_REPAIR))
+			error = xfs_trans_commit(sc->tp);
+		else
+			xfs_trans_cancel(sc->tp);
 		sc->tp = NULL;
 	}
 	if (sc->ip) {
@@ -166,6 +194,8 @@ xfs_scrub_teardown(
 			iput(VFS_I(sc->ip));
 		sc->ip = NULL;
 	}
+	if (sc->has_quotaofflock)
+		mutex_unlock(&sc->mp->m_quotainfo->qi_quotaofflock);
 	if (sc->buf) {
 		kmem_free(sc->buf);
 		sc->buf = NULL;
@@ -180,126 +210,150 @@ static const struct xfs_scrub_meta_ops meta_scrub_ops[] = {
 		.type	= ST_NONE,
 		.setup	= xfs_scrub_setup_fs,
 		.scrub	= xfs_scrub_probe,
+		.repair = xfs_repair_probe,
 	},
 	[XFS_SCRUB_TYPE_SB] = {		/* superblock */
 		.type	= ST_PERAG,
 		.setup	= xfs_scrub_setup_fs,
 		.scrub	= xfs_scrub_superblock,
+		.repair	= xfs_repair_superblock,
 	},
 	[XFS_SCRUB_TYPE_AGF] = {	/* agf */
 		.type	= ST_PERAG,
 		.setup	= xfs_scrub_setup_fs,
 		.scrub	= xfs_scrub_agf,
+		.repair	= xfs_repair_notsupported,
 	},
 	[XFS_SCRUB_TYPE_AGFL]= {	/* agfl */
 		.type	= ST_PERAG,
 		.setup	= xfs_scrub_setup_fs,
 		.scrub	= xfs_scrub_agfl,
+		.repair	= xfs_repair_notsupported,
 	},
 	[XFS_SCRUB_TYPE_AGI] = {	/* agi */
 		.type	= ST_PERAG,
 		.setup	= xfs_scrub_setup_fs,
 		.scrub	= xfs_scrub_agi,
+		.repair	= xfs_repair_notsupported,
 	},
 	[XFS_SCRUB_TYPE_BNOBT] = {	/* bnobt */
 		.type	= ST_PERAG,
 		.setup	= xfs_scrub_setup_ag_allocbt,
 		.scrub	= xfs_scrub_bnobt,
+		.repair	= xfs_repair_notsupported,
 	},
 	[XFS_SCRUB_TYPE_CNTBT] = {	/* cntbt */
 		.type	= ST_PERAG,
 		.setup	= xfs_scrub_setup_ag_allocbt,
 		.scrub	= xfs_scrub_cntbt,
+		.repair	= xfs_repair_notsupported,
 	},
 	[XFS_SCRUB_TYPE_INOBT] = {	/* inobt */
 		.type	= ST_PERAG,
 		.setup	= xfs_scrub_setup_ag_iallocbt,
 		.scrub	= xfs_scrub_inobt,
+		.repair	= xfs_repair_notsupported,
 	},
 	[XFS_SCRUB_TYPE_FINOBT] = {	/* finobt */
 		.type	= ST_PERAG,
 		.setup	= xfs_scrub_setup_ag_iallocbt,
 		.scrub	= xfs_scrub_finobt,
 		.has	= xfs_sb_version_hasfinobt,
+		.repair	= xfs_repair_notsupported,
 	},
 	[XFS_SCRUB_TYPE_RMAPBT] = {	/* rmapbt */
 		.type	= ST_PERAG,
 		.setup	= xfs_scrub_setup_ag_rmapbt,
 		.scrub	= xfs_scrub_rmapbt,
 		.has	= xfs_sb_version_hasrmapbt,
+		.repair	= xfs_repair_notsupported,
 	},
 	[XFS_SCRUB_TYPE_REFCNTBT] = {	/* refcountbt */
 		.type	= ST_PERAG,
 		.setup	= xfs_scrub_setup_ag_refcountbt,
 		.scrub	= xfs_scrub_refcountbt,
 		.has	= xfs_sb_version_hasreflink,
+		.repair	= xfs_repair_notsupported,
 	},
 	[XFS_SCRUB_TYPE_INODE] = {	/* inode record */
 		.type	= ST_INODE,
 		.setup	= xfs_scrub_setup_inode,
 		.scrub	= xfs_scrub_inode,
+		.repair	= xfs_repair_notsupported,
 	},
 	[XFS_SCRUB_TYPE_BMBTD] = {	/* inode data fork */
 		.type	= ST_INODE,
 		.setup	= xfs_scrub_setup_inode_bmap,
 		.scrub	= xfs_scrub_bmap_data,
+		.repair	= xfs_repair_notsupported,
 	},
 	[XFS_SCRUB_TYPE_BMBTA] = {	/* inode attr fork */
 		.type	= ST_INODE,
 		.setup	= xfs_scrub_setup_inode_bmap,
 		.scrub	= xfs_scrub_bmap_attr,
+		.repair	= xfs_repair_notsupported,
 	},
 	[XFS_SCRUB_TYPE_BMBTC] = {	/* inode CoW fork */
 		.type	= ST_INODE,
 		.setup	= xfs_scrub_setup_inode_bmap,
 		.scrub	= xfs_scrub_bmap_cow,
+		.repair	= xfs_repair_notsupported,
 	},
 	[XFS_SCRUB_TYPE_DIR] = {	/* directory */
 		.type	= ST_INODE,
 		.setup	= xfs_scrub_setup_directory,
 		.scrub	= xfs_scrub_directory,
+		.repair	= xfs_repair_notsupported,
 	},
 	[XFS_SCRUB_TYPE_XATTR] = {	/* extended attributes */
 		.type	= ST_INODE,
 		.setup	= xfs_scrub_setup_xattr,
 		.scrub	= xfs_scrub_xattr,
+		.repair	= xfs_repair_notsupported,
 	},
 	[XFS_SCRUB_TYPE_SYMLINK] = {	/* symbolic link */
 		.type	= ST_INODE,
 		.setup	= xfs_scrub_setup_symlink,
 		.scrub	= xfs_scrub_symlink,
+		.repair	= xfs_repair_notsupported,
 	},
 	[XFS_SCRUB_TYPE_PARENT] = {	/* parent pointers */
 		.type	= ST_INODE,
 		.setup	= xfs_scrub_setup_parent,
 		.scrub	= xfs_scrub_parent,
+		.repair	= xfs_repair_notsupported,
 	},
 	[XFS_SCRUB_TYPE_RTBITMAP] = {	/* realtime bitmap */
 		.type	= ST_FS,
 		.setup	= xfs_scrub_setup_rt,
 		.scrub	= xfs_scrub_rtbitmap,
 		.has	= xfs_sb_version_hasrealtime,
+		.repair	= xfs_repair_notsupported,
 	},
 	[XFS_SCRUB_TYPE_RTSUM] = {	/* realtime summary */
 		.type	= ST_FS,
 		.setup	= xfs_scrub_setup_rt,
 		.scrub	= xfs_scrub_rtsummary,
 		.has	= xfs_sb_version_hasrealtime,
+		.repair	= xfs_repair_notsupported,
 	},
 	[XFS_SCRUB_TYPE_UQUOTA] = {	/* user quota */
 		.type	= ST_FS,
 		.setup	= xfs_scrub_setup_quota,
 		.scrub	= xfs_scrub_quota,
+		.repair	= xfs_repair_notsupported,
 	},
 	[XFS_SCRUB_TYPE_GQUOTA] = {	/* group quota */
 		.type	= ST_FS,
 		.setup	= xfs_scrub_setup_quota,
 		.scrub	= xfs_scrub_quota,
+		.repair	= xfs_repair_notsupported,
 	},
 	[XFS_SCRUB_TYPE_PQUOTA] = {	/* project quota */
 		.type	= ST_FS,
 		.setup	= xfs_scrub_setup_quota,
 		.scrub	= xfs_scrub_quota,
+		.repair	= xfs_repair_notsupported,
 	},
 };
 
@@ -379,15 +433,54 @@ xfs_scrub_validate_inputs(
 	if (!xfs_sb_version_hasextflgbit(&mp->m_sb))
 		goto out;
 
-	/* We don't know how to repair anything yet. */
-	if (sm->sm_flags & XFS_SCRUB_IFLAG_REPAIR)
-		goto out;
+	/*
+	 * We only want to repair read-write v5+ filesystems.  Defer the check
+	 * for ops->repair until after our scrub confirms that we need to
+	 * perform repairs so that we avoid failing due to not supporting
+	 * repairing an object that doesn't need repairs.
+	 */
+	if (sm->sm_flags & XFS_SCRUB_IFLAG_REPAIR) {
+		error = -EOPNOTSUPP;
+		if (!xfs_sb_version_hascrc(&mp->m_sb))
+			goto out;
+
+		error = -EROFS;
+		if (mp->m_flags & XFS_MOUNT_RDONLY)
+			goto out;
+	}
 
 	error = 0;
 out:
 	return error;
 }
 
+#ifdef CONFIG_XFS_ONLINE_REPAIR
+static inline void xfs_scrub_postmortem(struct xfs_scrub_context *sc)
+{
+	/*
+	 * Userspace asked us to repair something, we repaired it, rescanned
+	 * it, and the rescan says it's still broken.  Scream about this in
+	 * the system logs.
+	 */
+	if ((sc->sm->sm_flags & XFS_SCRUB_IFLAG_REPAIR) &&
+	    (sc->sm->sm_flags & (XFS_SCRUB_OFLAG_CORRUPT |
+				 XFS_SCRUB_OFLAG_XCORRUPT)))
+		xfs_repair_failure(sc->mp);
+}
+#else
+static inline void xfs_scrub_postmortem(struct xfs_scrub_context *sc)
+{
+	/*
+	 * Userspace asked us to scrub something, it's broken, and we have no
+	 * way of fixing it.  Scream in the logs.
+	 */
+	if (sc->sm->sm_flags & (XFS_SCRUB_OFLAG_CORRUPT |
+				XFS_SCRUB_OFLAG_XCORRUPT))
+		xfs_alert_ratelimited(sc->mp,
+				"Corruption detected during scrub.");
+}
+#endif /* CONFIG_XFS_ONLINE_REPAIR */
+
 /* Dispatch metadata scrubbing. */
 int
 xfs_scrub_metadata(
@@ -397,6 +490,7 @@ xfs_scrub_metadata(
 	struct xfs_scrub_context	sc;
 	struct xfs_mount		*mp = ip->i_mount;
 	bool				try_harder = false;
+	bool				already_fixed = false;
 	int				error = 0;
 
 	BUILD_BUG_ON(sizeof(meta_scrub_ops) !=
@@ -446,10 +540,44 @@ retry_op:
 	} else if (error)
 		goto out_teardown;
 
-	if (sc.sm->sm_flags & (XFS_SCRUB_OFLAG_CORRUPT |
-			       XFS_SCRUB_OFLAG_XCORRUPT))
-		xfs_alert_ratelimited(mp, "Corruption detected during scrub.");
+	if ((sc.sm->sm_flags & XFS_SCRUB_IFLAG_REPAIR) && !already_fixed) {
+		bool needs_fix;
+
+		/* Let debug users force us into the repair routines. */
+		if (XFS_TEST_ERROR(false, mp, XFS_ERRTAG_FORCE_SCRUB_REPAIR))
+			sc.sm->sm_flags |= XFS_SCRUB_OFLAG_CORRUPT;
+
+		needs_fix = (sc.sm->sm_flags & (XFS_SCRUB_OFLAG_CORRUPT |
+						XFS_SCRUB_OFLAG_XCORRUPT |
+						XFS_SCRUB_OFLAG_PREEN));
+		/*
+		 * If userspace asked for a repair but it wasn't necessary,
+		 * report that back to userspace.
+		 */
+		if (!needs_fix) {
+			sc.sm->sm_flags |= XFS_SCRUB_OFLAG_NO_REPAIR_NEEDED;
+			goto out_nofix;
+		}
+
+		/*
+		 * If it's broken, userspace wants us to fix it, and we haven't
+		 * already tried to fix it, then attempt a repair.
+		 */
+		error = xfs_repair_attempt(ip, &sc, &already_fixed);
+		if (error == -EAGAIN) {
+			if (sc.try_harder)
+				try_harder = true;
+			error = xfs_scrub_teardown(&sc, ip, 0);
+			if (error) {
+				xfs_repair_failure(mp);
+				goto out;
+			}
+			goto retry_op;
+		}
+	}
 
+out_nofix:
+	xfs_scrub_postmortem(&sc);
 out_teardown:
 	error = xfs_scrub_teardown(&sc, ip, error);
 out:
diff --git a/fs/xfs/scrub/scrub.h b/fs/xfs/scrub/scrub.h
index 0d92af86f67a..636424d5e2ee 100644
--- a/fs/xfs/scrub/scrub.h
+++ b/fs/xfs/scrub/scrub.h
@@ -38,6 +38,9 @@ struct xfs_scrub_meta_ops {
 	/* Examine metadata for errors. */
 	int		(*scrub)(struct xfs_scrub_context *);
 
+	/* Repair or optimize the metadata. */
+	int		(*repair)(struct xfs_scrub_context *);
+
 	/* Decide if we even have this piece of metadata. */
 	bool		(*has)(struct xfs_sb *);
 
@@ -48,6 +51,7 @@ struct xfs_scrub_meta_ops {
 /* Buffer pointers and btree cursors for an entire AG. */
 struct xfs_scrub_ag {
 	xfs_agnumber_t			agno;
+	struct xfs_perag		*pag;
 
 	/* AG btree roots */
 	struct xfs_buf			*agf_bp;
@@ -73,6 +77,7 @@ struct xfs_scrub_context {
 	void				*buf;
 	uint				ilock_flags;
 	bool				try_harder;
+	bool				has_quotaofflock;
 
 	/* State tracking for single-AG operations. */
 	struct xfs_scrub_ag		sa;
diff --git a/fs/xfs/scrub/trace.h b/fs/xfs/scrub/trace.h
index 5d2b1c241be5..794d56bb1af8 100644
--- a/fs/xfs/scrub/trace.h
+++ b/fs/xfs/scrub/trace.h
@@ -69,6 +69,8 @@ DEFINE_EVENT(xfs_scrub_class, name, \
 DEFINE_SCRUB_EVENT(xfs_scrub_start);
 DEFINE_SCRUB_EVENT(xfs_scrub_done);
 DEFINE_SCRUB_EVENT(xfs_scrub_deadlock_retry);
+DEFINE_SCRUB_EVENT(xfs_repair_attempt);
+DEFINE_SCRUB_EVENT(xfs_repair_done);
 
 TRACE_EVENT(xfs_scrub_op_error,
 	TP_PROTO(struct xfs_scrub_context *sc, xfs_agnumber_t agno,
@@ -492,6 +494,262 @@ TRACE_EVENT(xfs_scrub_xref_error,
 		  __entry->ret_ip)
 );
 
+/* repair tracepoints */
+#if IS_ENABLED(CONFIG_XFS_ONLINE_REPAIR)
+
+DECLARE_EVENT_CLASS(xfs_repair_extent_class,
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
+		 xfs_agblock_t agbno, xfs_extlen_t len),
+	TP_ARGS(mp, agno, agbno, len),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, agbno)
+		__field(xfs_extlen_t, len)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->agno = agno;
+		__entry->agbno = agbno;
+		__entry->len = len;
+	),
+	TP_printk("dev %d:%d agno %u agbno %u len %u",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->agbno,
+		  __entry->len)
+);
+#define DEFINE_REPAIR_EXTENT_EVENT(name) \
+DEFINE_EVENT(xfs_repair_extent_class, name, \
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \
+		 xfs_agblock_t agbno, xfs_extlen_t len), \
+	TP_ARGS(mp, agno, agbno, len))
+DEFINE_REPAIR_EXTENT_EVENT(xfs_repair_dispose_btree_extent);
+DEFINE_REPAIR_EXTENT_EVENT(xfs_repair_collect_btree_extent);
+DEFINE_REPAIR_EXTENT_EVENT(xfs_repair_agfl_insert);
+
+DECLARE_EVENT_CLASS(xfs_repair_rmap_class,
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
+		 xfs_agblock_t agbno, xfs_extlen_t len,
+		 uint64_t owner, uint64_t offset, unsigned int flags),
+	TP_ARGS(mp, agno, agbno, len, owner, offset, flags),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, agbno)
+		__field(xfs_extlen_t, len)
+		__field(uint64_t, owner)
+		__field(uint64_t, offset)
+		__field(unsigned int, flags)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->agno = agno;
+		__entry->agbno = agbno;
+		__entry->len = len;
+		__entry->owner = owner;
+		__entry->offset = offset;
+		__entry->flags = flags;
+	),
+	TP_printk("dev %d:%d agno %u agbno %u len %u owner %lld offset %llu flags 0x%x",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->agbno,
+		  __entry->len,
+		  __entry->owner,
+		  __entry->offset,
+		  __entry->flags)
+);
+#define DEFINE_REPAIR_RMAP_EVENT(name) \
+DEFINE_EVENT(xfs_repair_rmap_class, name, \
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \
+		 xfs_agblock_t agbno, xfs_extlen_t len, \
+		 uint64_t owner, uint64_t offset, unsigned int flags), \
+	TP_ARGS(mp, agno, agbno, len, owner, offset, flags))
+DEFINE_REPAIR_RMAP_EVENT(xfs_repair_alloc_extent_fn);
+DEFINE_REPAIR_RMAP_EVENT(xfs_repair_ialloc_extent_fn);
+DEFINE_REPAIR_RMAP_EVENT(xfs_repair_rmap_extent_fn);
+DEFINE_REPAIR_RMAP_EVENT(xfs_repair_bmap_extent_fn);
+
+TRACE_EVENT(xfs_repair_refcount_extent_fn,
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
+		 struct xfs_refcount_irec *irec),
+	TP_ARGS(mp, agno, irec),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, startblock)
+		__field(xfs_extlen_t, blockcount)
+		__field(xfs_nlink_t, refcount)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->agno = agno;
+		__entry->startblock = irec->rc_startblock;
+		__entry->blockcount = irec->rc_blockcount;
+		__entry->refcount = irec->rc_refcount;
+	),
+	TP_printk("dev %d:%d agno %u agbno %u len %u refcount %u",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->startblock,
+		  __entry->blockcount,
+		  __entry->refcount)
+)
+
+TRACE_EVENT(xfs_repair_init_btblock,
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, xfs_agblock_t agbno,
+		 xfs_btnum_t btnum),
+	TP_ARGS(mp, agno, agbno, btnum),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, agbno)
+		__field(uint32_t, btnum)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->agno = agno;
+		__entry->agbno = agbno;
+		__entry->btnum = btnum;
+	),
+	TP_printk("dev %d:%d agno %u agbno %u btnum %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->agbno,
+		  __entry->btnum)
+)
+TRACE_EVENT(xfs_repair_findroot_block,
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, xfs_agblock_t agbno,
+		 uint32_t magic, uint16_t level),
+	TP_ARGS(mp, agno, agbno, magic, level),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, agbno)
+		__field(uint32_t, magic)
+		__field(uint16_t, level)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->agno = agno;
+		__entry->agbno = agbno;
+		__entry->magic = magic;
+		__entry->level = level;
+	),
+	TP_printk("dev %d:%d agno %u agbno %u magic 0x%x level %u",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->agbno,
+		  __entry->magic,
+		  __entry->level)
+)
+TRACE_EVENT(xfs_repair_calc_ag_resblks,
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
+		 xfs_agino_t icount, xfs_agblock_t aglen, xfs_agblock_t freelen,
+		 xfs_agblock_t usedlen),
+	TP_ARGS(mp, agno, icount, aglen, freelen, usedlen),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agino_t, icount)
+		__field(xfs_agblock_t, aglen)
+		__field(xfs_agblock_t, freelen)
+		__field(xfs_agblock_t, usedlen)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->agno = agno;
+		__entry->icount = icount;
+		__entry->aglen = aglen;
+		__entry->freelen = freelen;
+		__entry->usedlen = usedlen;
+	),
+	TP_printk("dev %d:%d agno %d icount %u aglen %u freelen %u usedlen %u",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->icount,
+		  __entry->aglen,
+		  __entry->freelen,
+		  __entry->usedlen)
+)
+TRACE_EVENT(xfs_repair_calc_ag_resblks_btsize,
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
+		 xfs_agblock_t bnobt_sz, xfs_agblock_t inobt_sz,
+		 xfs_agblock_t rmapbt_sz, xfs_agblock_t refcbt_sz),
+	TP_ARGS(mp, agno, bnobt_sz, inobt_sz, rmapbt_sz, refcbt_sz),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, bnobt_sz)
+		__field(xfs_agblock_t, inobt_sz)
+		__field(xfs_agblock_t, rmapbt_sz)
+		__field(xfs_agblock_t, refcbt_sz)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->agno = agno;
+		__entry->bnobt_sz = bnobt_sz;
+		__entry->inobt_sz = inobt_sz;
+		__entry->rmapbt_sz = rmapbt_sz;
+		__entry->refcbt_sz = refcbt_sz;
+	),
+	TP_printk("dev %d:%d agno %d bno %u ino %u rmap %u refcount %u",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->bnobt_sz,
+		  __entry->inobt_sz,
+		  __entry->rmapbt_sz,
+		  __entry->refcbt_sz)
+)
+TRACE_EVENT(xfs_repair_reset_counters,
+	TP_PROTO(struct xfs_mount *mp),
+	TP_ARGS(mp),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+	),
+	TP_printk("dev %d:%d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev))
+)
+
+TRACE_EVENT(xfs_repair_ialloc_insert,
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
+		 xfs_agino_t startino, uint16_t holemask, uint8_t count,
+		 uint8_t freecount, uint64_t freemask),
+	TP_ARGS(mp, agno, startino, holemask, count, freecount, freemask),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agino_t, startino)
+		__field(uint16_t, holemask)
+		__field(uint8_t, count)
+		__field(uint8_t, freecount)
+		__field(uint64_t, freemask)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->agno = agno;
+		__entry->startino = startino;
+		__entry->holemask = holemask;
+		__entry->count = count;
+		__entry->freecount = freecount;
+		__entry->freemask = freemask;
+	),
+	TP_printk("dev %d:%d agno %d startino %u holemask 0x%x count %u freecount %u freemask 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->startino,
+		  __entry->holemask,
+		  __entry->count,
+		  __entry->freecount,
+		  __entry->freemask)
+)
+
+#endif /* IS_ENABLED(CONFIG_XFS_ONLINE_REPAIR) */
+
 #endif /* _TRACE_XFS_SCRUB_TRACE_H */
 
 #undef TRACE_INCLUDE_PATH
diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c
index 102463543db3..ca6903726689 100644
--- a/fs/xfs/xfs_aops.c
+++ b/fs/xfs/xfs_aops.c
@@ -1378,10 +1378,9 @@ xfs_vm_bmap(
 	struct address_space	*mapping,
 	sector_t		block)
 {
-	struct inode		*inode = (struct inode *)mapping->host;
-	struct xfs_inode	*ip = XFS_I(inode);
+	struct xfs_inode	*ip = XFS_I(mapping->host);
 
-	trace_xfs_vm_bmap(XFS_I(inode));
+	trace_xfs_vm_bmap(ip);
 
 	/*
 	 * The swap code (ab-)uses ->bmap to get a block mapping and then
@@ -1394,9 +1393,7 @@ xfs_vm_bmap(
 	 */
 	if (xfs_is_reflink_inode(ip) || XFS_IS_REALTIME_INODE(ip))
 		return 0;
-
-	filemap_write_and_wait(mapping);
-	return generic_block_bmap(mapping, block, xfs_get_blocks);
+	return iomap_bmap(mapping, block, &xfs_iomap_ops);
 }
 
 STATIC int
@@ -1475,6 +1472,16 @@ xfs_vm_set_page_dirty(
 	return newly_dirty;
 }
 
+static int
+xfs_iomap_swapfile_activate(
+	struct swap_info_struct		*sis,
+	struct file			*swap_file,
+	sector_t			*span)
+{
+	sis->bdev = xfs_find_bdev_for_inode(file_inode(swap_file));
+	return iomap_swapfile_activate(sis, swap_file, span, &xfs_iomap_ops);
+}
+
 const struct address_space_operations xfs_address_space_operations = {
 	.readpage		= xfs_vm_readpage,
 	.readpages		= xfs_vm_readpages,
@@ -1488,6 +1495,7 @@ const struct address_space_operations xfs_address_space_operations = {
 	.migratepage		= buffer_migrate_page,
 	.is_partially_uptodate  = block_is_partially_uptodate,
 	.error_remove_page	= generic_error_remove_page,
+	.swap_activate		= xfs_iomap_swapfile_activate,
 };
 
 const struct address_space_operations xfs_dax_aops = {
@@ -1495,4 +1503,5 @@ const struct address_space_operations xfs_dax_aops = {
 	.direct_IO		= noop_direct_IO,
 	.set_page_dirty		= noop_set_page_dirty,
 	.invalidatepage		= noop_invalidatepage,
+	.swap_activate		= xfs_iomap_swapfile_activate,
 };
diff --git a/fs/xfs/xfs_bmap_item.c b/fs/xfs/xfs_bmap_item.c
index 2203465e63ea..618bb71535c8 100644
--- a/fs/xfs/xfs_bmap_item.c
+++ b/fs/xfs/xfs_bmap_item.c
@@ -160,7 +160,7 @@ STATIC void
 xfs_bui_item_unlock(
 	struct xfs_log_item	*lip)
 {
-	if (lip->li_flags & XFS_LI_ABORTED)
+	if (test_bit(XFS_LI_ABORTED, &lip->li_flags))
 		xfs_bui_release(BUI_ITEM(lip));
 }
 
@@ -305,7 +305,7 @@ xfs_bud_item_unlock(
 {
 	struct xfs_bud_log_item	*budp = BUD_ITEM(lip);
 
-	if (lip->li_flags & XFS_LI_ABORTED) {
+	if (test_bit(XFS_LI_ABORTED, &lip->li_flags)) {
 		xfs_bui_release(budp->bud_buip);
 		kmem_zone_free(xfs_bud_zone, budp);
 	}
diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c
index 8cd8c412f52d..06badcbadeb4 100644
--- a/fs/xfs/xfs_bmap_util.c
+++ b/fs/xfs/xfs_bmap_util.c
@@ -848,7 +848,7 @@ xfs_free_eofblocks(
 		/*
 		 * Attach the dquots to the inode up front.
 		 */
-		error = xfs_qm_dqattach(ip, 0);
+		error = xfs_qm_dqattach(ip);
 		if (error)
 			return error;
 
@@ -871,8 +871,8 @@ xfs_free_eofblocks(
 		 * contents of the file are flushed to disk then the files
 		 * may be full of holes (ie NULL files bug).
 		 */
-		error = xfs_itruncate_extents(&tp, ip, XFS_DATA_FORK,
-					      XFS_ISIZE(ip));
+		error = xfs_itruncate_extents_flags(&tp, ip, XFS_DATA_FORK,
+					XFS_ISIZE(ip), XFS_BMAPI_NODISCARD);
 		if (error) {
 			/*
 			 * If we get an error at this point we simply don't
@@ -918,7 +918,7 @@ xfs_alloc_file_space(
 	if (XFS_FORCED_SHUTDOWN(mp))
 		return -EIO;
 
-	error = xfs_qm_dqattach(ip, 0);
+	error = xfs_qm_dqattach(ip);
 	if (error)
 		return error;
 
@@ -1169,7 +1169,7 @@ xfs_free_file_space(
 
 	trace_xfs_free_file_space(ip);
 
-	error = xfs_qm_dqattach(ip, 0);
+	error = xfs_qm_dqattach(ip);
 	if (error)
 		return error;
 
diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c
index 55661cbdb51b..5179ab9e3d6a 100644
--- a/fs/xfs/xfs_buf.c
+++ b/fs/xfs/xfs_buf.c
@@ -549,17 +549,31 @@ xfs_buf_hash_destroy(
 }
 
 /*
- *	Look up, and creates if absent, a lockable buffer for
- *	a given range of an inode.  The buffer is returned
- *	locked.	No I/O is implied by this call.
+ * Look up a buffer in the buffer cache and return it referenced and locked
+ * in @found_bp.
+ *
+ * If @new_bp is supplied and we have a lookup miss, insert @new_bp into the
+ * cache.
+ *
+ * If XBF_TRYLOCK is set in @flags, only try to lock the buffer and return
+ * -EAGAIN if we fail to lock it.
+ *
+ * Return values are:
+ *	-EFSCORRUPTED if have been supplied with an invalid address
+ *	-EAGAIN on trylock failure
+ *	-ENOENT if we fail to find a match and @new_bp was NULL
+ *	0, with @found_bp:
+ *		- @new_bp if we inserted it into the cache
+ *		- the buffer we found and locked.
  */
-xfs_buf_t *
-_xfs_buf_find(
+static int
+xfs_buf_find(
 	struct xfs_buftarg	*btp,
 	struct xfs_buf_map	*map,
 	int			nmaps,
 	xfs_buf_flags_t		flags,
-	xfs_buf_t		*new_bp)
+	struct xfs_buf		*new_bp,
+	struct xfs_buf		**found_bp)
 {
 	struct xfs_perag	*pag;
 	xfs_buf_t		*bp;
@@ -567,6 +581,8 @@ _xfs_buf_find(
 	xfs_daddr_t		eofs;
 	int			i;
 
+	*found_bp = NULL;
+
 	for (i = 0; i < nmaps; i++)
 		cmap.bm_len += map[i].bm_len;
 
@@ -580,16 +596,11 @@ _xfs_buf_find(
 	 */
 	eofs = XFS_FSB_TO_BB(btp->bt_mount, btp->bt_mount->m_sb.sb_dblocks);
 	if (cmap.bm_bn < 0 || cmap.bm_bn >= eofs) {
-		/*
-		 * XXX (dgc): we should really be returning -EFSCORRUPTED here,
-		 * but none of the higher level infrastructure supports
-		 * returning a specific error on buffer lookup failures.
-		 */
 		xfs_alert(btp->bt_mount,
 			  "%s: daddr 0x%llx out of range, EOFS 0x%llx",
 			  __func__, cmap.bm_bn, eofs);
 		WARN_ON(1);
-		return NULL;
+		return -EFSCORRUPTED;
 	}
 
 	pag = xfs_perag_get(btp->bt_mount,
@@ -604,19 +615,20 @@ _xfs_buf_find(
 	}
 
 	/* No match found */
-	if (new_bp) {
-		/* the buffer keeps the perag reference until it is freed */
-		new_bp->b_pag = pag;
-		rhashtable_insert_fast(&pag->pag_buf_hash,
-				       &new_bp->b_rhash_head,
-				       xfs_buf_hash_params);
-		spin_unlock(&pag->pag_buf_lock);
-	} else {
+	if (!new_bp) {
 		XFS_STATS_INC(btp->bt_mount, xb_miss_locked);
 		spin_unlock(&pag->pag_buf_lock);
 		xfs_perag_put(pag);
+		return -ENOENT;
 	}
-	return new_bp;
+
+	/* the buffer keeps the perag reference until it is freed */
+	new_bp->b_pag = pag;
+	rhashtable_insert_fast(&pag->pag_buf_hash, &new_bp->b_rhash_head,
+			       xfs_buf_hash_params);
+	spin_unlock(&pag->pag_buf_lock);
+	*found_bp = new_bp;
+	return 0;
 
 found:
 	spin_unlock(&pag->pag_buf_lock);
@@ -626,7 +638,7 @@ found:
 		if (flags & XBF_TRYLOCK) {
 			xfs_buf_rele(bp);
 			XFS_STATS_INC(btp->bt_mount, xb_busy_locked);
-			return NULL;
+			return -EAGAIN;
 		}
 		xfs_buf_lock(bp);
 		XFS_STATS_INC(btp->bt_mount, xb_get_locked_waited);
@@ -646,6 +658,24 @@ found:
 
 	trace_xfs_buf_find(bp, flags, _RET_IP_);
 	XFS_STATS_INC(btp->bt_mount, xb_get_locked);
+	*found_bp = bp;
+	return 0;
+}
+
+struct xfs_buf *
+xfs_buf_incore(
+	struct xfs_buftarg	*target,
+	xfs_daddr_t		blkno,
+	size_t			numblks,
+	xfs_buf_flags_t		flags)
+{
+	struct xfs_buf		*bp;
+	int			error;
+	DEFINE_SINGLE_BUF_MAP(map, blkno, numblks);
+
+	error = xfs_buf_find(target, &map, 1, flags, NULL, &bp);
+	if (error)
+		return NULL;
 	return bp;
 }
 
@@ -665,9 +695,27 @@ xfs_buf_get_map(
 	struct xfs_buf		*new_bp;
 	int			error = 0;
 
-	bp = _xfs_buf_find(target, map, nmaps, flags, NULL);
-	if (likely(bp))
+	error = xfs_buf_find(target, map, nmaps, flags, NULL, &bp);
+
+	switch (error) {
+	case 0:
+		/* cache hit */
 		goto found;
+	case -EAGAIN:
+		/* cache hit, trylock failure, caller handles failure */
+		ASSERT(flags & XBF_TRYLOCK);
+		return NULL;
+	case -ENOENT:
+		/* cache miss, go for insert */
+		break;
+	case -EFSCORRUPTED:
+	default:
+		/*
+		 * None of the higher layers understand failure types
+		 * yet, so return NULL to signal a fatal lookup error.
+		 */
+		return NULL;
+	}
 
 	new_bp = _xfs_buf_alloc(target, map, nmaps, flags);
 	if (unlikely(!new_bp))
@@ -679,8 +727,8 @@ xfs_buf_get_map(
 		return NULL;
 	}
 
-	bp = _xfs_buf_find(target, map, nmaps, flags, new_bp);
-	if (!bp) {
+	error = xfs_buf_find(target, map, nmaps, flags, new_bp, &bp);
+	if (error) {
 		xfs_buf_free(new_bp);
 		return NULL;
 	}
diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h
index edced162a674..f5f2b71c2fde 100644
--- a/fs/xfs/xfs_buf.h
+++ b/fs/xfs/xfs_buf.h
@@ -218,20 +218,9 @@ typedef struct xfs_buf {
 } xfs_buf_t;
 
 /* Finding and Reading Buffers */
-struct xfs_buf *_xfs_buf_find(struct xfs_buftarg *target,
-			      struct xfs_buf_map *map, int nmaps,
-			      xfs_buf_flags_t flags, struct xfs_buf *new_bp);
-
-static inline struct xfs_buf *
-xfs_incore(
-	struct xfs_buftarg	*target,
-	xfs_daddr_t		blkno,
-	size_t			numblks,
-	xfs_buf_flags_t		flags)
-{
-	DEFINE_SINGLE_BUF_MAP(map, blkno, numblks);
-	return _xfs_buf_find(target, &map, 1, flags, NULL);
-}
+struct xfs_buf *xfs_buf_incore(struct xfs_buftarg *target,
+			   xfs_daddr_t blkno, size_t numblks,
+			   xfs_buf_flags_t flags);
 
 struct xfs_buf *_xfs_buf_alloc(struct xfs_buftarg *target,
 			       struct xfs_buf_map *map, int nmaps,
@@ -358,6 +347,18 @@ extern void xfs_buf_terminate(void);
 
 void xfs_buf_set_ref(struct xfs_buf *bp, int lru_ref);
 
+/*
+ * If the buffer is already on the LRU, do nothing. Otherwise set the buffer
+ * up with a reference count of 0 so it will be tossed from the cache when
+ * released.
+ */
+static inline void xfs_buf_oneshot(struct xfs_buf *bp)
+{
+	if (!list_empty(&bp->b_lru) || atomic_read(&bp->b_lru_ref) > 1)
+		return;
+	atomic_set(&bp->b_lru_ref, 0);
+}
+
 static inline int xfs_buf_ispinned(struct xfs_buf *bp)
 {
 	return atomic_read(&bp->b_pin_count);
diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c
index 82ad270e390e..c2311379d1c3 100644
--- a/fs/xfs/xfs_buf_item.c
+++ b/fs/xfs/xfs_buf_item.c
@@ -438,7 +438,7 @@ xfs_buf_item_unpin(
 			 * xfs_trans_uncommit() will try to reference the
 			 * buffer which we no longer have a hold on.
 			 */
-			if (lip->li_desc)
+			if (!list_empty(&lip->li_trans))
 				xfs_trans_del_item(lip);
 
 			/*
@@ -568,13 +568,15 @@ xfs_buf_item_unlock(
 {
 	struct xfs_buf_log_item	*bip = BUF_ITEM(lip);
 	struct xfs_buf		*bp = bip->bli_buf;
-	bool			aborted = !!(lip->li_flags & XFS_LI_ABORTED);
+	bool			aborted;
 	bool			hold = !!(bip->bli_flags & XFS_BLI_HOLD);
 	bool			dirty = !!(bip->bli_flags & XFS_BLI_DIRTY);
 #if defined(DEBUG) || defined(XFS_WARN)
 	bool			ordered = !!(bip->bli_flags & XFS_BLI_ORDERED);
 #endif
 
+	aborted = test_bit(XFS_LI_ABORTED, &lip->li_flags);
+
 	/* Clear the buffer's association with this transaction. */
 	bp->b_transp = NULL;
 
@@ -743,8 +745,10 @@ xfs_buf_item_init(
 	 * nothing to do here so return.
 	 */
 	ASSERT(bp->b_target->bt_mount == mp);
-	if (bip != NULL) {
+	if (bip) {
 		ASSERT(bip->bli_item.li_type == XFS_LI_BUF);
+		ASSERT(!bp->b_transp);
+		ASSERT(bip->bli_buf == bp);
 		return 0;
 	}
 
diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c
index a7daef9e16bf..2567391489bd 100644
--- a/fs/xfs/xfs_dquot.c
+++ b/fs/xfs/xfs_dquot.c
@@ -288,49 +288,43 @@ xfs_dquot_set_prealloc_limits(struct xfs_dquot *dqp)
 }
 
 /*
- * Allocate a block and fill it with dquots.
- * This is called when the bmapi finds a hole.
+ * Ensure that the given in-core dquot has a buffer on disk backing it, and
+ * return the buffer. This is called when the bmapi finds a hole.
  */
 STATIC int
-xfs_qm_dqalloc(
-	xfs_trans_t	**tpp,
-	xfs_mount_t	*mp,
-	xfs_dquot_t	*dqp,
-	xfs_inode_t	*quotip,
-	xfs_fileoff_t	offset_fsb,
-	xfs_buf_t	**O_bpp)
+xfs_dquot_disk_alloc(
+	struct xfs_trans	**tpp,
+	struct xfs_dquot	*dqp,
+	struct xfs_buf		**bpp)
 {
-	xfs_fsblock_t	firstblock;
-	struct xfs_defer_ops dfops;
-	xfs_bmbt_irec_t map;
-	int		nmaps, error;
-	xfs_buf_t	*bp;
-	xfs_trans_t	*tp = *tpp;
-
-	ASSERT(tp != NULL);
+	struct xfs_bmbt_irec	map;
+	struct xfs_defer_ops	dfops;
+	struct xfs_mount	*mp = (*tpp)->t_mountp;
+	struct xfs_buf		*bp;
+	struct xfs_inode	*quotip = xfs_quota_inode(mp, dqp->dq_flags);
+	xfs_fsblock_t		firstblock;
+	int			nmaps = 1;
+	int			error;
 
 	trace_xfs_dqalloc(dqp);
 
-	/*
-	 * Initialize the bmap freelist prior to calling bmapi code.
-	 */
 	xfs_defer_init(&dfops, &firstblock);
 	xfs_ilock(quotip, XFS_ILOCK_EXCL);
-	/*
-	 * Return if this type of quotas is turned off while we didn't
-	 * have an inode lock
-	 */
 	if (!xfs_this_quota_on(dqp->q_mount, dqp->dq_flags)) {
+		/*
+		 * Return if this type of quotas is turned off while we didn't
+		 * have an inode lock
+		 */
 		xfs_iunlock(quotip, XFS_ILOCK_EXCL);
 		return -ESRCH;
 	}
 
-	xfs_trans_ijoin(tp, quotip, XFS_ILOCK_EXCL);
-	nmaps = 1;
-	error = xfs_bmapi_write(tp, quotip, offset_fsb,
-				XFS_DQUOT_CLUSTER_SIZE_FSB, XFS_BMAPI_METADATA,
-				&firstblock, XFS_QM_DQALLOC_SPACE_RES(mp),
-				&map, &nmaps, &dfops);
+	/* Create the block mapping. */
+	xfs_trans_ijoin(*tpp, quotip, XFS_ILOCK_EXCL);
+	error = xfs_bmapi_write(*tpp, quotip, dqp->q_fileoffset,
+			XFS_DQUOT_CLUSTER_SIZE_FSB, XFS_BMAPI_METADATA,
+			&firstblock, XFS_QM_DQALLOC_SPACE_RES(mp),
+			&map, &nmaps, &dfops);
 	if (error)
 		goto error0;
 	ASSERT(map.br_blockcount == XFS_DQUOT_CLUSTER_SIZE_FSB);
@@ -344,10 +338,8 @@ xfs_qm_dqalloc(
 	dqp->q_blkno = XFS_FSB_TO_DADDR(mp, map.br_startblock);
 
 	/* now we can just get the buffer (there's nothing to read yet) */
-	bp = xfs_trans_get_buf(tp, mp->m_ddev_targp,
-			       dqp->q_blkno,
-			       mp->m_quotainfo->qi_dqchunklen,
-			       0);
+	bp = xfs_trans_get_buf(*tpp, mp->m_ddev_targp, dqp->q_blkno,
+			mp->m_quotainfo->qi_dqchunklen, 0);
 	if (!bp) {
 		error = -ENOMEM;
 		goto error1;
@@ -358,37 +350,45 @@ xfs_qm_dqalloc(
 	 * Make a chunk of dquots out of this buffer and log
 	 * the entire thing.
 	 */
-	xfs_qm_init_dquot_blk(tp, mp, be32_to_cpu(dqp->q_core.d_id),
+	xfs_qm_init_dquot_blk(*tpp, mp, be32_to_cpu(dqp->q_core.d_id),
 			      dqp->dq_flags & XFS_DQ_ALLTYPES, bp);
+	xfs_buf_set_ref(bp, XFS_DQUOT_REF);
 
 	/*
-	 * xfs_defer_finish() may commit the current transaction and
-	 * start a second transaction if the freelist is not empty.
+	 * Hold the buffer and join it to the dfops so that we'll still own
+	 * the buffer when we return to the caller.  The buffer disposal on
+	 * error must be paid attention to very carefully, as it has been
+	 * broken since commit efa092f3d4c6 "[XFS] Fixes a bug in the quota
+	 * code when allocating a new dquot record" in 2005, and the later
+	 * conversion to xfs_defer_ops in commit 310a75a3c6c747 failed to keep
+	 * the buffer locked across the _defer_finish call.  We can now do
+	 * this correctly with xfs_defer_bjoin.
 	 *
-	 * Since we still want to modify this buffer, we need to
-	 * ensure that the buffer is not released on commit of
-	 * the first transaction and ensure the buffer is added to the
-	 * second transaction.
+	 * Above, we allocated a disk block for the dquot information and
+	 * used get_buf to initialize the dquot.  If the _defer_bjoin fails,
+	 * the buffer is still locked to *tpp, so we must _bhold_release and
+	 * then _trans_brelse the buffer.  If the _defer_finish fails, the old
+	 * transaction is gone but the new buffer is not joined or held to any
+	 * transaction, so we must _buf_relse it.
 	 *
-	 * If there is only one transaction then don't stop the buffer
-	 * from being released when it commits later on.
+	 * If everything succeeds, the caller of this function is returned a
+	 * buffer that is locked and held to the transaction.  The caller
+	 * is responsible for unlocking any buffer passed back, either
+	 * manually or by committing the transaction.
 	 */
-
-	xfs_trans_bhold(tp, bp);
-
+	xfs_trans_bhold(*tpp, bp);
+	error = xfs_defer_bjoin(&dfops, bp);
+	if (error) {
+		xfs_trans_bhold_release(*tpp, bp);
+		xfs_trans_brelse(*tpp, bp);
+		goto error1;
+	}
 	error = xfs_defer_finish(tpp, &dfops);
-	if (error)
+	if (error) {
+		xfs_buf_relse(bp);
 		goto error1;
-
-	/* Transaction was committed? */
-	if (*tpp != tp) {
-		tp = *tpp;
-		xfs_trans_bjoin(tp, bp);
-	} else {
-		xfs_trans_bhold_release(tp, bp);
 	}
-
-	*O_bpp = bp;
+	*bpp = bp;
 	return 0;
 
 error1:
@@ -398,32 +398,24 @@ error0:
 }
 
 /*
- * Maps a dquot to the buffer containing its on-disk version.
- * This returns a ptr to the buffer containing the on-disk dquot
- * in the bpp param, and a ptr to the on-disk dquot within that buffer
+ * Read in the in-core dquot's on-disk metadata and return the buffer.
+ * Returns ENOENT to signal a hole.
  */
 STATIC int
-xfs_qm_dqtobp(
-	xfs_trans_t		**tpp,
-	xfs_dquot_t		*dqp,
-	xfs_disk_dquot_t	**O_ddpp,
-	xfs_buf_t		**O_bpp,
-	uint			flags)
+xfs_dquot_disk_read(
+	struct xfs_mount	*mp,
+	struct xfs_dquot	*dqp,
+	struct xfs_buf		**bpp)
 {
 	struct xfs_bmbt_irec	map;
-	int			nmaps = 1, error;
 	struct xfs_buf		*bp;
-	struct xfs_inode	*quotip;
-	struct xfs_mount	*mp = dqp->q_mount;
-	xfs_dqid_t		id = be32_to_cpu(dqp->q_core.d_id);
-	struct xfs_trans	*tp = (tpp ? *tpp : NULL);
+	struct xfs_inode	*quotip = xfs_quota_inode(mp, dqp->dq_flags);
 	uint			lock_mode;
-
-	quotip = xfs_quota_inode(dqp->q_mount, dqp->dq_flags);
-	dqp->q_fileoffset = (xfs_fileoff_t)id / mp->m_quotainfo->qi_dqperchunk;
+	int			nmaps = 1;
+	int			error;
 
 	lock_mode = xfs_ilock_data_map_shared(quotip);
-	if (!xfs_this_quota_on(dqp->q_mount, dqp->dq_flags)) {
+	if (!xfs_this_quota_on(mp, dqp->dq_flags)) {
 		/*
 		 * Return if this type of quotas is turned off while we
 		 * didn't have the quota inode lock.
@@ -436,81 +428,48 @@ xfs_qm_dqtobp(
 	 * Find the block map; no allocations yet
 	 */
 	error = xfs_bmapi_read(quotip, dqp->q_fileoffset,
-			       XFS_DQUOT_CLUSTER_SIZE_FSB, &map, &nmaps, 0);
-
+			XFS_DQUOT_CLUSTER_SIZE_FSB, &map, &nmaps, 0);
 	xfs_iunlock(quotip, lock_mode);
 	if (error)
 		return error;
 
 	ASSERT(nmaps == 1);
-	ASSERT(map.br_blockcount == 1);
+	ASSERT(map.br_blockcount >= 1);
+	ASSERT(map.br_startblock != DELAYSTARTBLOCK);
+	if (map.br_startblock == HOLESTARTBLOCK)
+		return -ENOENT;
+
+	trace_xfs_dqtobp_read(dqp);
 
 	/*
-	 * Offset of dquot in the (fixed sized) dquot chunk.
+	 * store the blkno etc so that we don't have to do the
+	 * mapping all the time
 	 */
-	dqp->q_bufoffset = (id % mp->m_quotainfo->qi_dqperchunk) *
-		sizeof(xfs_dqblk_t);
-
-	ASSERT(map.br_startblock != DELAYSTARTBLOCK);
-	if (map.br_startblock == HOLESTARTBLOCK) {
-		/*
-		 * We don't allocate unless we're asked to
-		 */
-		if (!(flags & XFS_QMOPT_DQALLOC))
-			return -ENOENT;
-
-		ASSERT(tp);
-		error = xfs_qm_dqalloc(tpp, mp, dqp, quotip,
-					dqp->q_fileoffset, &bp);
-		if (error)
-			return error;
-		tp = *tpp;
-	} else {
-		trace_xfs_dqtobp_read(dqp);
+	dqp->q_blkno = XFS_FSB_TO_DADDR(mp, map.br_startblock);
 
-		/*
-		 * store the blkno etc so that we don't have to do the
-		 * mapping all the time
-		 */
-		dqp->q_blkno = XFS_FSB_TO_DADDR(mp, map.br_startblock);
-
-		error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp,
-					   dqp->q_blkno,
-					   mp->m_quotainfo->qi_dqchunklen,
-					   0, &bp, &xfs_dquot_buf_ops);
-		if (error) {
-			ASSERT(bp == NULL);
-			return error;
-		}
+	error = xfs_trans_read_buf(mp, NULL, mp->m_ddev_targp, dqp->q_blkno,
+			mp->m_quotainfo->qi_dqchunklen, 0, &bp,
+			&xfs_dquot_buf_ops);
+	if (error) {
+		ASSERT(bp == NULL);
+		return error;
 	}
 
 	ASSERT(xfs_buf_islocked(bp));
-	*O_bpp = bp;
-	*O_ddpp = bp->b_addr + dqp->q_bufoffset;
+	xfs_buf_set_ref(bp, XFS_DQUOT_REF);
+	*bpp = bp;
 
 	return 0;
 }
 
-
-/*
- * Read in the ondisk dquot using dqtobp() then copy it to an incore version,
- * and release the buffer immediately.
- *
- * If XFS_QMOPT_DQALLOC is set, allocate a dquot on disk if it needed.
- */
-int
-xfs_qm_dqread(
+/* Allocate and initialize everything we need for an incore dquot. */
+STATIC struct xfs_dquot *
+xfs_dquot_alloc(
 	struct xfs_mount	*mp,
 	xfs_dqid_t		id,
-	uint			type,
-	uint			flags,
-	struct xfs_dquot	**O_dqpp)
+	uint			type)
 {
 	struct xfs_dquot	*dqp;
-	struct xfs_disk_dquot	*ddqp;
-	struct xfs_buf		*bp;
-	struct xfs_trans	*tp = NULL;
-	int			error;
 
 	dqp = kmem_zone_zalloc(xfs_qm_dqzone, KM_SLEEP);
 
@@ -520,6 +479,12 @@ xfs_qm_dqread(
 	INIT_LIST_HEAD(&dqp->q_lru);
 	mutex_init(&dqp->q_qlock);
 	init_waitqueue_head(&dqp->q_pinwait);
+	dqp->q_fileoffset = (xfs_fileoff_t)id / mp->m_quotainfo->qi_dqperchunk;
+	/*
+	 * Offset of dquot in the (fixed sized) dquot chunk.
+	 */
+	dqp->q_bufoffset = (id % mp->m_quotainfo->qi_dqperchunk) *
+			sizeof(xfs_dqblk_t);
 
 	/*
 	 * Because we want to use a counting completion, complete
@@ -548,35 +513,22 @@ xfs_qm_dqread(
 		break;
 	}
 
-	XFS_STATS_INC(mp, xs_qm_dquot);
-
-	trace_xfs_dqread(dqp);
+	xfs_qm_dquot_logitem_init(dqp);
 
-	if (flags & XFS_QMOPT_DQALLOC) {
-		error = xfs_trans_alloc(mp, &M_RES(mp)->tr_qm_dqalloc,
-				XFS_QM_DQALLOC_SPACE_RES(mp), 0, 0, &tp);
-		if (error)
-			goto error0;
-	}
+	XFS_STATS_INC(mp, xs_qm_dquot);
+	return dqp;
+}
 
-	/*
-	 * get a pointer to the on-disk dquot and the buffer containing it
-	 * dqp already knows its own type (GROUP/USER).
-	 */
-	error = xfs_qm_dqtobp(&tp, dqp, &ddqp, &bp, flags);
-	if (error) {
-		/*
-		 * This can happen if quotas got turned off (ESRCH),
-		 * or if the dquot didn't exist on disk and we ask to
-		 * allocate (ENOENT).
-		 */
-		trace_xfs_dqread_fail(dqp);
-		goto error1;
-	}
+/* Copy the in-core quota fields in from the on-disk buffer. */
+STATIC void
+xfs_dquot_from_disk(
+	struct xfs_dquot	*dqp,
+	struct xfs_buf		*bp)
+{
+	struct xfs_disk_dquot	*ddqp = bp->b_addr + dqp->q_bufoffset;
 
 	/* copy everything from disk dquot to the incore dquot */
 	memcpy(&dqp->q_core, ddqp, sizeof(xfs_disk_dquot_t));
-	xfs_qm_dquot_logitem_init(dqp);
 
 	/*
 	 * Reservation counters are defined as reservation plus current usage
@@ -588,40 +540,90 @@ xfs_qm_dqread(
 
 	/* initialize the dquot speculative prealloc thresholds */
 	xfs_dquot_set_prealloc_limits(dqp);
+}
 
-	/* Mark the buf so that this will stay incore a little longer */
-	xfs_buf_set_ref(bp, XFS_DQUOT_REF);
+/* Allocate and initialize the dquot buffer for this in-core dquot. */
+static int
+xfs_qm_dqread_alloc(
+	struct xfs_mount	*mp,
+	struct xfs_dquot	*dqp,
+	struct xfs_buf		**bpp)
+{
+	struct xfs_trans	*tp;
+	struct xfs_buf		*bp;
+	int			error;
+
+	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_qm_dqalloc,
+			XFS_QM_DQALLOC_SPACE_RES(mp), 0, 0, &tp);
+	if (error)
+		goto err;
+
+	error = xfs_dquot_disk_alloc(&tp, dqp, &bp);
+	if (error)
+		goto err_cancel;
+
+	error = xfs_trans_commit(tp);
+	if (error) {
+		/*
+		 * Buffer was held to the transaction, so we have to unlock it
+		 * manually here because we're not passing it back.
+		 */
+		xfs_buf_relse(bp);
+		goto err;
+	}
+	*bpp = bp;
+	return 0;
+
+err_cancel:
+	xfs_trans_cancel(tp);
+err:
+	return error;
+}
+
+/*
+ * Read in the ondisk dquot using dqtobp() then copy it to an incore version,
+ * and release the buffer immediately.  If @can_alloc is true, fill any
+ * holes in the on-disk metadata.
+ */
+static int
+xfs_qm_dqread(
+	struct xfs_mount	*mp,
+	xfs_dqid_t		id,
+	uint			type,
+	bool			can_alloc,
+	struct xfs_dquot	**dqpp)
+{
+	struct xfs_dquot	*dqp;
+	struct xfs_buf		*bp;
+	int			error;
+
+	dqp = xfs_dquot_alloc(mp, id, type);
+	trace_xfs_dqread(dqp);
+
+	/* Try to read the buffer, allocating if necessary. */
+	error = xfs_dquot_disk_read(mp, dqp, &bp);
+	if (error == -ENOENT && can_alloc)
+		error = xfs_qm_dqread_alloc(mp, dqp, &bp);
+	if (error)
+		goto err;
 
 	/*
-	 * We got the buffer with a xfs_trans_read_buf() (in dqtobp())
-	 * So we need to release with xfs_trans_brelse().
-	 * The strategy here is identical to that of inodes; we lock
-	 * the dquot in xfs_qm_dqget() before making it accessible to
-	 * others. This is because dquots, like inodes, need a good level of
-	 * concurrency, and we don't want to take locks on the entire buffers
-	 * for dquot accesses.
-	 * Note also that the dquot buffer may even be dirty at this point, if
-	 * this particular dquot was repaired. We still aren't afraid to
-	 * brelse it because we have the changes incore.
+	 * At this point we should have a clean locked buffer.  Copy the data
+	 * to the incore dquot and release the buffer since the incore dquot
+	 * has its own locking protocol so we needn't tie up the buffer any
+	 * further.
 	 */
 	ASSERT(xfs_buf_islocked(bp));
-	xfs_trans_brelse(tp, bp);
+	xfs_dquot_from_disk(dqp, bp);
 
-	if (tp) {
-		error = xfs_trans_commit(tp);
-		if (error)
-			goto error0;
-	}
-
-	*O_dqpp = dqp;
+	xfs_buf_relse(bp);
+	*dqpp = dqp;
 	return error;
 
-error1:
-	if (tp)
-		xfs_trans_cancel(tp);
-error0:
+err:
+	trace_xfs_dqread_fail(dqp);
 	xfs_qm_dqdestroy(dqp);
-	*O_dqpp = NULL;
+	*dqpp = NULL;
 	return error;
 }
 
@@ -679,77 +681,230 @@ xfs_dq_get_next_id(
 }
 
 /*
- * Given the file system, inode OR id, and type (UDQUOT/GDQUOT), return a
- * a locked dquot, doing an allocation (if requested) as needed.
- * When both an inode and an id are given, the inode's id takes precedence.
- * That is, if the id changes while we don't hold the ilock inside this
- * function, the new dquot is returned, not necessarily the one requested
- * in the id argument.
+ * Look up the dquot in the in-core cache.  If found, the dquot is returned
+ * locked and ready to go.
+ */
+static struct xfs_dquot *
+xfs_qm_dqget_cache_lookup(
+	struct xfs_mount	*mp,
+	struct xfs_quotainfo	*qi,
+	struct radix_tree_root	*tree,
+	xfs_dqid_t		id)
+{
+	struct xfs_dquot	*dqp;
+
+restart:
+	mutex_lock(&qi->qi_tree_lock);
+	dqp = radix_tree_lookup(tree, id);
+	if (!dqp) {
+		mutex_unlock(&qi->qi_tree_lock);
+		XFS_STATS_INC(mp, xs_qm_dqcachemisses);
+		return NULL;
+	}
+
+	xfs_dqlock(dqp);
+	if (dqp->dq_flags & XFS_DQ_FREEING) {
+		xfs_dqunlock(dqp);
+		mutex_unlock(&qi->qi_tree_lock);
+		trace_xfs_dqget_freeing(dqp);
+		delay(1);
+		goto restart;
+	}
+
+	dqp->q_nrefs++;
+	mutex_unlock(&qi->qi_tree_lock);
+
+	trace_xfs_dqget_hit(dqp);
+	XFS_STATS_INC(mp, xs_qm_dqcachehits);
+	return dqp;
+}
+
+/*
+ * Try to insert a new dquot into the in-core cache.  If an error occurs the
+ * caller should throw away the dquot and start over.  Otherwise, the dquot
+ * is returned locked (and held by the cache) as if there had been a cache
+ * hit.
+ */
+static int
+xfs_qm_dqget_cache_insert(
+	struct xfs_mount	*mp,
+	struct xfs_quotainfo	*qi,
+	struct radix_tree_root	*tree,
+	xfs_dqid_t		id,
+	struct xfs_dquot	*dqp)
+{
+	int			error;
+
+	mutex_lock(&qi->qi_tree_lock);
+	error = radix_tree_insert(tree, id, dqp);
+	if (unlikely(error)) {
+		/* Duplicate found!  Caller must try again. */
+		WARN_ON(error != -EEXIST);
+		mutex_unlock(&qi->qi_tree_lock);
+		trace_xfs_dqget_dup(dqp);
+		return error;
+	}
+
+	/* Return a locked dquot to the caller, with a reference taken. */
+	xfs_dqlock(dqp);
+	dqp->q_nrefs = 1;
+
+	qi->qi_dquots++;
+	mutex_unlock(&qi->qi_tree_lock);
+
+	return 0;
+}
+
+/* Check our input parameters. */
+static int
+xfs_qm_dqget_checks(
+	struct xfs_mount	*mp,
+	uint			type)
+{
+	if (WARN_ON_ONCE(!XFS_IS_QUOTA_RUNNING(mp)))
+		return -ESRCH;
+
+	switch (type) {
+	case XFS_DQ_USER:
+		if (!XFS_IS_UQUOTA_ON(mp))
+			return -ESRCH;
+		return 0;
+	case XFS_DQ_GROUP:
+		if (!XFS_IS_GQUOTA_ON(mp))
+			return -ESRCH;
+		return 0;
+	case XFS_DQ_PROJ:
+		if (!XFS_IS_PQUOTA_ON(mp))
+			return -ESRCH;
+		return 0;
+	default:
+		WARN_ON_ONCE(0);
+		return -EINVAL;
+	}
+}
+
+/*
+ * Given the file system, id, and type (UDQUOT/GDQUOT), return a a locked
+ * dquot, doing an allocation (if requested) as needed.
  */
 int
 xfs_qm_dqget(
-	xfs_mount_t	*mp,
-	xfs_inode_t	*ip,	  /* locked inode (optional) */
-	xfs_dqid_t	id,	  /* uid/projid/gid depending on type */
-	uint		type,	  /* XFS_DQ_USER/XFS_DQ_PROJ/XFS_DQ_GROUP */
-	uint		flags,	  /* DQALLOC, DQSUSER, DQREPAIR, DOWARN */
-	xfs_dquot_t	**O_dqpp) /* OUT : locked incore dquot */
+	struct xfs_mount	*mp,
+	xfs_dqid_t		id,
+	uint			type,
+	bool			can_alloc,
+	struct xfs_dquot	**O_dqpp)
 {
 	struct xfs_quotainfo	*qi = mp->m_quotainfo;
-	struct radix_tree_root *tree = xfs_dquot_tree(qi, type);
+	struct radix_tree_root	*tree = xfs_dquot_tree(qi, type);
 	struct xfs_dquot	*dqp;
 	int			error;
 
-	ASSERT(XFS_IS_QUOTA_RUNNING(mp));
-	if ((! XFS_IS_UQUOTA_ON(mp) && type == XFS_DQ_USER) ||
-	    (! XFS_IS_PQUOTA_ON(mp) && type == XFS_DQ_PROJ) ||
-	    (! XFS_IS_GQUOTA_ON(mp) && type == XFS_DQ_GROUP)) {
-		return -ESRCH;
+	error = xfs_qm_dqget_checks(mp, type);
+	if (error)
+		return error;
+
+restart:
+	dqp = xfs_qm_dqget_cache_lookup(mp, qi, tree, id);
+	if (dqp) {
+		*O_dqpp = dqp;
+		return 0;
 	}
 
-	ASSERT(type == XFS_DQ_USER ||
-	       type == XFS_DQ_PROJ ||
-	       type == XFS_DQ_GROUP);
-	if (ip) {
-		ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
-		ASSERT(xfs_inode_dquot(ip, type) == NULL);
+	error = xfs_qm_dqread(mp, id, type, can_alloc, &dqp);
+	if (error)
+		return error;
+
+	error = xfs_qm_dqget_cache_insert(mp, qi, tree, id, dqp);
+	if (error) {
+		/*
+		 * Duplicate found. Just throw away the new dquot and start
+		 * over.
+		 */
+		xfs_qm_dqdestroy(dqp);
+		XFS_STATS_INC(mp, xs_qm_dquot_dups);
+		goto restart;
 	}
 
-restart:
-	mutex_lock(&qi->qi_tree_lock);
-	dqp = radix_tree_lookup(tree, id);
-	if (dqp) {
-		xfs_dqlock(dqp);
-		if (dqp->dq_flags & XFS_DQ_FREEING) {
-			xfs_dqunlock(dqp);
-			mutex_unlock(&qi->qi_tree_lock);
-			trace_xfs_dqget_freeing(dqp);
-			delay(1);
-			goto restart;
-		}
+	trace_xfs_dqget_miss(dqp);
+	*O_dqpp = dqp;
+	return 0;
+}
 
-		/* uninit / unused quota found in radix tree, keep looking  */
-		if (flags & XFS_QMOPT_DQNEXT) {
-			if (XFS_IS_DQUOT_UNINITIALIZED(dqp)) {
-				xfs_dqunlock(dqp);
-				mutex_unlock(&qi->qi_tree_lock);
-				error = xfs_dq_get_next_id(mp, type, &id);
-				if (error)
-					return error;
-				goto restart;
-			}
-		}
+/*
+ * Given a dquot id and type, read and initialize a dquot from the on-disk
+ * metadata.  This function is only for use during quota initialization so
+ * it ignores the dquot cache assuming that the dquot shrinker isn't set up.
+ * The caller is responsible for _qm_dqdestroy'ing the returned dquot.
+ */
+int
+xfs_qm_dqget_uncached(
+	struct xfs_mount	*mp,
+	xfs_dqid_t		id,
+	uint			type,
+	struct xfs_dquot	**dqpp)
+{
+	int			error;
 
-		dqp->q_nrefs++;
-		mutex_unlock(&qi->qi_tree_lock);
+	error = xfs_qm_dqget_checks(mp, type);
+	if (error)
+		return error;
+
+	return xfs_qm_dqread(mp, id, type, 0, dqpp);
+}
+
+/* Return the quota id for a given inode and type. */
+xfs_dqid_t
+xfs_qm_id_for_quotatype(
+	struct xfs_inode	*ip,
+	uint			type)
+{
+	switch (type) {
+	case XFS_DQ_USER:
+		return ip->i_d.di_uid;
+	case XFS_DQ_GROUP:
+		return ip->i_d.di_gid;
+	case XFS_DQ_PROJ:
+		return xfs_get_projid(ip);
+	}
+	ASSERT(0);
+	return 0;
+}
+
+/*
+ * Return the dquot for a given inode and type.  If @can_alloc is true, then
+ * allocate blocks if needed.  The inode's ILOCK must be held and it must not
+ * have already had an inode attached.
+ */
+int
+xfs_qm_dqget_inode(
+	struct xfs_inode	*ip,
+	uint			type,
+	bool			can_alloc,
+	struct xfs_dquot	**O_dqpp)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_quotainfo	*qi = mp->m_quotainfo;
+	struct radix_tree_root	*tree = xfs_dquot_tree(qi, type);
+	struct xfs_dquot	*dqp;
+	xfs_dqid_t		id;
+	int			error;
 
-		trace_xfs_dqget_hit(dqp);
-		XFS_STATS_INC(mp, xs_qm_dqcachehits);
+	error = xfs_qm_dqget_checks(mp, type);
+	if (error)
+		return error;
+
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	ASSERT(xfs_inode_dquot(ip, type) == NULL);
+
+	id = xfs_qm_id_for_quotatype(ip, type);
+
+restart:
+	dqp = xfs_qm_dqget_cache_lookup(mp, qi, tree, id);
+	if (dqp) {
 		*O_dqpp = dqp;
 		return 0;
 	}
-	mutex_unlock(&qi->qi_tree_lock);
-	XFS_STATS_INC(mp, xs_qm_dqcachemisses);
 
 	/*
 	 * Dquot cache miss. We don't want to keep the inode lock across
@@ -758,87 +913,81 @@ restart:
 	 * lock here means dealing with a chown that can happen before
 	 * we re-acquire the lock.
 	 */
-	if (ip)
-		xfs_iunlock(ip, XFS_ILOCK_EXCL);
-
-	error = xfs_qm_dqread(mp, id, type, flags, &dqp);
-
-	if (ip)
-		xfs_ilock(ip, XFS_ILOCK_EXCL);
-
-	/* If we are asked to find next active id, keep looking */
-	if (error == -ENOENT && (flags & XFS_QMOPT_DQNEXT)) {
-		error = xfs_dq_get_next_id(mp, type, &id);
-		if (!error)
-			goto restart;
-	}
-
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	error = xfs_qm_dqread(mp, id, type, can_alloc, &dqp);
+	xfs_ilock(ip, XFS_ILOCK_EXCL);
 	if (error)
 		return error;
 
-	if (ip) {
-		/*
-		 * A dquot could be attached to this inode by now, since
-		 * we had dropped the ilock.
-		 */
-		if (xfs_this_quota_on(mp, type)) {
-			struct xfs_dquot	*dqp1;
-
-			dqp1 = xfs_inode_dquot(ip, type);
-			if (dqp1) {
-				xfs_qm_dqdestroy(dqp);
-				dqp = dqp1;
-				xfs_dqlock(dqp);
-				goto dqret;
-			}
-		} else {
-			/* inode stays locked on return */
+	/*
+	 * A dquot could be attached to this inode by now, since we had
+	 * dropped the ilock.
+	 */
+	if (xfs_this_quota_on(mp, type)) {
+		struct xfs_dquot	*dqp1;
+
+		dqp1 = xfs_inode_dquot(ip, type);
+		if (dqp1) {
 			xfs_qm_dqdestroy(dqp);
-			return -ESRCH;
+			dqp = dqp1;
+			xfs_dqlock(dqp);
+			goto dqret;
 		}
+	} else {
+		/* inode stays locked on return */
+		xfs_qm_dqdestroy(dqp);
+		return -ESRCH;
 	}
 
-	mutex_lock(&qi->qi_tree_lock);
-	error = radix_tree_insert(tree, id, dqp);
-	if (unlikely(error)) {
-		WARN_ON(error != -EEXIST);
-
+	error = xfs_qm_dqget_cache_insert(mp, qi, tree, id, dqp);
+	if (error) {
 		/*
 		 * Duplicate found. Just throw away the new dquot and start
 		 * over.
 		 */
-		mutex_unlock(&qi->qi_tree_lock);
-		trace_xfs_dqget_dup(dqp);
 		xfs_qm_dqdestroy(dqp);
 		XFS_STATS_INC(mp, xs_qm_dquot_dups);
 		goto restart;
 	}
 
-	/*
-	 * We return a locked dquot to the caller, with a reference taken
-	 */
-	xfs_dqlock(dqp);
-	dqp->q_nrefs = 1;
+dqret:
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	trace_xfs_dqget_miss(dqp);
+	*O_dqpp = dqp;
+	return 0;
+}
 
-	qi->qi_dquots++;
-	mutex_unlock(&qi->qi_tree_lock);
+/*
+ * Starting at @id and progressing upwards, look for an initialized incore
+ * dquot, lock it, and return it.
+ */
+int
+xfs_qm_dqget_next(
+	struct xfs_mount	*mp,
+	xfs_dqid_t		id,
+	uint			type,
+	struct xfs_dquot	**dqpp)
+{
+	struct xfs_dquot	*dqp;
+	int			error = 0;
 
-	/* If we are asked to find next active id, keep looking */
-	if (flags & XFS_QMOPT_DQNEXT) {
-		if (XFS_IS_DQUOT_UNINITIALIZED(dqp)) {
-			xfs_qm_dqput(dqp);
-			error = xfs_dq_get_next_id(mp, type, &id);
-			if (error)
-				return error;
-			goto restart;
+	*dqpp = NULL;
+	for (; !error; error = xfs_dq_get_next_id(mp, type, &id)) {
+		error = xfs_qm_dqget(mp, id, type, false, &dqp);
+		if (error == -ENOENT)
+			continue;
+		else if (error != 0)
+			break;
+
+		if (!XFS_IS_DQUOT_UNINITIALIZED(dqp)) {
+			*dqpp = dqp;
+			return 0;
 		}
+
+		xfs_qm_dqput(dqp);
 	}
 
- dqret:
-	ASSERT((ip == NULL) || xfs_isilocked(ip, XFS_ILOCK_EXCL));
-	trace_xfs_dqget_miss(dqp);
-	*O_dqpp = dqp;
-	return 0;
+	return error;
 }
 
 /*
@@ -913,9 +1062,9 @@ xfs_qm_dqflush_done(
 	 * since it's cheaper, and then we recheck while
 	 * holding the lock before removing the dquot from the AIL.
 	 */
-	if ((lip->li_flags & XFS_LI_IN_AIL) &&
+	if (test_bit(XFS_LI_IN_AIL, &lip->li_flags) &&
 	    ((lip->li_lsn == qip->qli_flush_lsn) ||
-	     (lip->li_flags & XFS_LI_FAILED))) {
+	     test_bit(XFS_LI_FAILED, &lip->li_flags))) {
 
 		/* xfs_trans_ail_delete() drops the AIL lock. */
 		spin_lock(&ailp->ail_lock);
@@ -926,8 +1075,7 @@ xfs_qm_dqflush_done(
 			 * Clear the failed state since we are about to drop the
 			 * flush lock
 			 */
-			if (lip->li_flags & XFS_LI_FAILED)
-				xfs_clear_li_failed(lip);
+			xfs_clear_li_failed(lip);
 			spin_unlock(&ailp->ail_lock);
 		}
 	}
@@ -953,6 +1101,7 @@ xfs_qm_dqflush(
 {
 	struct xfs_mount	*mp = dqp->q_mount;
 	struct xfs_buf		*bp;
+	struct xfs_dqblk	*dqb;
 	struct xfs_disk_dquot	*ddqp;
 	xfs_failaddr_t		fa;
 	int			error;
@@ -996,12 +1145,13 @@ xfs_qm_dqflush(
 	/*
 	 * Calculate the location of the dquot inside the buffer.
 	 */
-	ddqp = bp->b_addr + dqp->q_bufoffset;
+	dqb = bp->b_addr + dqp->q_bufoffset;
+	ddqp = &dqb->dd_diskdq;
 
 	/*
-	 * A simple sanity check in case we got a corrupted dquot..
+	 * A simple sanity check in case we got a corrupted dquot.
 	 */
-	fa = xfs_dquot_verify(mp, &dqp->q_core, be32_to_cpu(ddqp->d_id), 0, 0);
+	fa = xfs_dqblk_verify(mp, dqb, be32_to_cpu(ddqp->d_id), 0);
 	if (fa) {
 		xfs_alert(mp, "corrupt dquot ID 0x%x in memory at %pS",
 				be32_to_cpu(ddqp->d_id), fa);
@@ -1032,8 +1182,6 @@ xfs_qm_dqflush(
 	 * of a dquot without an up-to-date CRC getting to disk.
 	 */
 	if (xfs_sb_version_hascrc(&mp->m_sb)) {
-		struct xfs_dqblk *dqb = (struct xfs_dqblk *)ddqp;
-
 		dqb->dd_lsn = cpu_to_be64(dqp->q_logitem.qli_item.li_lsn);
 		xfs_update_cksum((char *)dqb, sizeof(struct xfs_dqblk),
 				 XFS_DQUOT_CRC_OFF);
@@ -1119,3 +1267,35 @@ xfs_qm_exit(void)
 	kmem_zone_destroy(xfs_qm_dqtrxzone);
 	kmem_zone_destroy(xfs_qm_dqzone);
 }
+
+/*
+ * Iterate every dquot of a particular type.  The caller must ensure that the
+ * particular quota type is active.  iter_fn can return negative error codes,
+ * or XFS_BTREE_QUERY_RANGE_ABORT to indicate that it wants to stop iterating.
+ */
+int
+xfs_qm_dqiterate(
+	struct xfs_mount	*mp,
+	uint			dqtype,
+	xfs_qm_dqiterate_fn	iter_fn,
+	void			*priv)
+{
+	struct xfs_dquot	*dq;
+	xfs_dqid_t		id = 0;
+	int			error;
+
+	do {
+		error = xfs_qm_dqget_next(mp, id, dqtype, &dq);
+		if (error == -ENOENT)
+			return 0;
+		if (error)
+			return error;
+
+		error = iter_fn(dq, dqtype, priv);
+		id = be32_to_cpu(dq->q_core.d_id);
+		xfs_qm_dqput(dq);
+		id++;
+	} while (error == 0 && id != 0);
+
+	return error;
+}
diff --git a/fs/xfs/xfs_dquot.h b/fs/xfs/xfs_dquot.h
index 2f536f33cd26..bdd6bd921528 100644
--- a/fs/xfs/xfs_dquot.h
+++ b/fs/xfs/xfs_dquot.h
@@ -160,8 +160,6 @@ static inline bool xfs_dquot_lowsp(struct xfs_dquot *dqp)
 #define XFS_QM_ISPDQ(dqp)	((dqp)->dq_flags & XFS_DQ_PROJ)
 #define XFS_QM_ISGDQ(dqp)	((dqp)->dq_flags & XFS_DQ_GROUP)
 
-extern int		xfs_qm_dqread(struct xfs_mount *, xfs_dqid_t, uint,
-					uint, struct xfs_dquot	**);
 extern void		xfs_qm_dqdestroy(xfs_dquot_t *);
 extern int		xfs_qm_dqflush(struct xfs_dquot *, struct xfs_buf **);
 extern void		xfs_qm_dqunpin_wait(xfs_dquot_t *);
@@ -169,8 +167,19 @@ extern void		xfs_qm_adjust_dqtimers(xfs_mount_t *,
 					xfs_disk_dquot_t *);
 extern void		xfs_qm_adjust_dqlimits(struct xfs_mount *,
 					       struct xfs_dquot *);
-extern int		xfs_qm_dqget(xfs_mount_t *, xfs_inode_t *,
-					xfs_dqid_t, uint, uint, xfs_dquot_t **);
+extern xfs_dqid_t	xfs_qm_id_for_quotatype(struct xfs_inode *ip,
+					uint type);
+extern int		xfs_qm_dqget(struct xfs_mount *mp, xfs_dqid_t id,
+					uint type, bool can_alloc,
+					struct xfs_dquot **dqpp);
+extern int		xfs_qm_dqget_inode(struct xfs_inode *ip, uint type,
+					bool can_alloc,
+					struct xfs_dquot **dqpp);
+extern int		xfs_qm_dqget_next(struct xfs_mount *mp, xfs_dqid_t id,
+					uint type, struct xfs_dquot **dqpp);
+extern int		xfs_qm_dqget_uncached(struct xfs_mount *mp,
+					xfs_dqid_t id, uint type,
+					struct xfs_dquot **dqpp);
 extern void		xfs_qm_dqput(xfs_dquot_t *);
 
 extern void		xfs_dqlock2(struct xfs_dquot *, struct xfs_dquot *);
@@ -185,4 +194,9 @@ static inline struct xfs_dquot *xfs_qm_dqhold(struct xfs_dquot *dqp)
 	return dqp;
 }
 
+typedef int (*xfs_qm_dqiterate_fn)(struct xfs_dquot *dq, uint dqtype,
+		void *priv);
+int xfs_qm_dqiterate(struct xfs_mount *mp, uint dqtype,
+		xfs_qm_dqiterate_fn iter_fn, void *priv);
+
 #endif /* __XFS_DQUOT_H__ */
diff --git a/fs/xfs/xfs_dquot_item.c b/fs/xfs/xfs_dquot_item.c
index 4b331e354da7..8eb7415474d6 100644
--- a/fs/xfs/xfs_dquot_item.c
+++ b/fs/xfs/xfs_dquot_item.c
@@ -173,7 +173,7 @@ xfs_qm_dquot_logitem_push(
 	 * The buffer containing this item failed to be written back
 	 * previously. Resubmit the buffer for IO
 	 */
-	if (lip->li_flags & XFS_LI_FAILED) {
+	if (test_bit(XFS_LI_FAILED, &lip->li_flags)) {
 		if (!xfs_buf_trylock(bp))
 			return XFS_ITEM_LOCKED;
 
@@ -209,10 +209,7 @@ xfs_qm_dquot_logitem_push(
 	spin_unlock(&lip->li_ailp->ail_lock);
 
 	error = xfs_qm_dqflush(dqp, &bp);
-	if (error) {
-		xfs_warn(dqp->q_mount, "%s: push error %d on dqp "PTR_FMT,
-			__func__, error, dqp);
-	} else {
+	if (!error) {
 		if (!xfs_buf_delwri_queue(bp, buffer_list))
 			rval = XFS_ITEM_FLUSHING;
 		xfs_buf_relse(bp);
diff --git a/fs/xfs/xfs_error.c b/fs/xfs/xfs_error.c
index a63f5083f497..7975634cb8fe 100644
--- a/fs/xfs/xfs_error.c
+++ b/fs/xfs/xfs_error.c
@@ -61,6 +61,7 @@ static unsigned int xfs_errortag_random_default[] = {
 	XFS_RANDOM_LOG_BAD_CRC,
 	XFS_RANDOM_LOG_ITEM_PIN,
 	XFS_RANDOM_BUF_LRU_REF,
+	XFS_RANDOM_FORCE_SCRUB_REPAIR,
 };
 
 struct xfs_errortag_attr {
@@ -167,6 +168,7 @@ XFS_ERRORTAG_ATTR_RW(drop_writes,	XFS_ERRTAG_DROP_WRITES);
 XFS_ERRORTAG_ATTR_RW(log_bad_crc,	XFS_ERRTAG_LOG_BAD_CRC);
 XFS_ERRORTAG_ATTR_RW(log_item_pin,	XFS_ERRTAG_LOG_ITEM_PIN);
 XFS_ERRORTAG_ATTR_RW(buf_lru_ref,	XFS_ERRTAG_BUF_LRU_REF);
+XFS_ERRORTAG_ATTR_RW(force_repair,	XFS_ERRTAG_FORCE_SCRUB_REPAIR);
 
 static struct attribute *xfs_errortag_attrs[] = {
 	XFS_ERRORTAG_ATTR_LIST(noerror),
@@ -201,6 +203,7 @@ static struct attribute *xfs_errortag_attrs[] = {
 	XFS_ERRORTAG_ATTR_LIST(log_bad_crc),
 	XFS_ERRORTAG_ATTR_LIST(log_item_pin),
 	XFS_ERRORTAG_ATTR_LIST(buf_lru_ref),
+	XFS_ERRORTAG_ATTR_LIST(force_repair),
 	NULL,
 };
 
diff --git a/fs/xfs/xfs_extfree_item.c b/fs/xfs/xfs_extfree_item.c
index b5b1e567b9f4..a889b550979a 100644
--- a/fs/xfs/xfs_extfree_item.c
+++ b/fs/xfs/xfs_extfree_item.c
@@ -168,7 +168,7 @@ STATIC void
 xfs_efi_item_unlock(
 	struct xfs_log_item	*lip)
 {
-	if (lip->li_flags & XFS_LI_ABORTED)
+	if (test_bit(XFS_LI_ABORTED, &lip->li_flags))
 		xfs_efi_release(EFI_ITEM(lip));
 }
 
@@ -402,7 +402,7 @@ xfs_efd_item_unlock(
 {
 	struct xfs_efd_log_item	*efdp = EFD_ITEM(lip);
 
-	if (lip->li_flags & XFS_LI_ABORTED) {
+	if (test_bit(XFS_LI_ABORTED, &lip->li_flags)) {
 		xfs_efi_release(efdp->efd_efip);
 		xfs_efd_item_free(efdp);
 	}
@@ -542,7 +542,7 @@ xfs_efi_recover(
 	for (i = 0; i < efip->efi_format.efi_nextents; i++) {
 		extp = &efip->efi_format.efi_extents[i];
 		error = xfs_trans_free_extent(tp, efdp, extp->ext_start,
-					      extp->ext_len, &oinfo);
+					      extp->ext_len, &oinfo, false);
 		if (error)
 			goto abort_error;
 
diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c
index e70fb8ccecea..0e3fb8978344 100644
--- a/fs/xfs/xfs_file.c
+++ b/fs/xfs/xfs_file.c
@@ -414,6 +414,12 @@ xfs_dio_write_end_io(
 	if (size <= 0)
 		return size;
 
+	/*
+	 * Capture amount written on completion as we can't reliably account
+	 * for it on submission.
+	 */
+	XFS_STATS_ADD(ip->i_mount, xs_write_bytes, size);
+
 	if (flags & IOMAP_DIO_COW) {
 		error = xfs_reflink_end_cow(ip, offset, size);
 		if (error)
@@ -599,7 +605,16 @@ xfs_file_dax_write(
 	}
 out:
 	xfs_iunlock(ip, iolock);
-	return error ? error : ret;
+	if (error)
+		return error;
+
+	if (ret > 0) {
+		XFS_STATS_ADD(ip->i_mount, xs_write_bytes, ret);
+
+		/* Handle various SYNC-type writes */
+		ret = generic_write_sync(iocb, ret);
+	}
+	return ret;
 }
 
 STATIC ssize_t
@@ -669,6 +684,12 @@ write_retry:
 out:
 	if (iolock)
 		xfs_iunlock(ip, iolock);
+
+	if (ret > 0) {
+		XFS_STATS_ADD(ip->i_mount, xs_write_bytes, ret);
+		/* Handle various SYNC-type writes */
+		ret = generic_write_sync(iocb, ret);
+	}
 	return ret;
 }
 
@@ -693,8 +714,9 @@ xfs_file_write_iter(
 		return -EIO;
 
 	if (IS_DAX(inode))
-		ret = xfs_file_dax_write(iocb, from);
-	else if (iocb->ki_flags & IOCB_DIRECT) {
+		return xfs_file_dax_write(iocb, from);
+
+	if (iocb->ki_flags & IOCB_DIRECT) {
 		/*
 		 * Allow a directio write to fall back to a buffered
 		 * write *only* in the case that we're doing a reflink
@@ -702,20 +724,11 @@ xfs_file_write_iter(
 		 * allow an operation to fall back to buffered mode.
 		 */
 		ret = xfs_file_dio_aio_write(iocb, from);
-		if (ret == -EREMCHG)
-			goto buffered;
-	} else {
-buffered:
-		ret = xfs_file_buffered_aio_write(iocb, from);
+		if (ret != -EREMCHG)
+			return ret;
 	}
 
-	if (ret > 0) {
-		XFS_STATS_ADD(ip->i_mount, xs_write_bytes, ret);
-
-		/* Handle various SYNC-type writes */
-		ret = generic_write_sync(iocb, ret);
-	}
-	return ret;
+	return xfs_file_buffered_aio_write(iocb, from);
 }
 
 #define	XFS_FALLOC_FL_SUPPORTED						\
@@ -1007,7 +1020,7 @@ xfs_file_llseek(
  *       page_lock (MM)
  *         i_lock (XFS - extent map serialisation)
  */
-static int
+static vm_fault_t
 __xfs_filemap_fault(
 	struct vm_fault		*vmf,
 	enum page_entry_size	pe_size,
@@ -1015,7 +1028,7 @@ __xfs_filemap_fault(
 {
 	struct inode		*inode = file_inode(vmf->vma->vm_file);
 	struct xfs_inode	*ip = XFS_I(inode);
-	int			ret;
+	vm_fault_t		ret;
 
 	trace_xfs_filemap_fault(ip, pe_size, write_fault);
 
@@ -1044,7 +1057,7 @@ __xfs_filemap_fault(
 	return ret;
 }
 
-static int
+static vm_fault_t
 xfs_filemap_fault(
 	struct vm_fault		*vmf)
 {
@@ -1054,7 +1067,7 @@ xfs_filemap_fault(
 			(vmf->flags & FAULT_FLAG_WRITE));
 }
 
-static int
+static vm_fault_t
 xfs_filemap_huge_fault(
 	struct vm_fault		*vmf,
 	enum page_entry_size	pe_size)
@@ -1067,7 +1080,7 @@ xfs_filemap_huge_fault(
 			(vmf->flags & FAULT_FLAG_WRITE));
 }
 
-static int
+static vm_fault_t
 xfs_filemap_page_mkwrite(
 	struct vm_fault		*vmf)
 {
@@ -1079,7 +1092,7 @@ xfs_filemap_page_mkwrite(
  * on write faults. In reality, it needs to serialise against truncate and
  * prepare memory for writing so handle is as standard write fault.
  */
-static int
+static vm_fault_t
 xfs_filemap_pfn_mkwrite(
 	struct vm_fault		*vmf)
 {
diff --git a/fs/xfs/xfs_fsmap.c b/fs/xfs/xfs_fsmap.c
index 43cfc07996a4..0299febece9c 100644
--- a/fs/xfs/xfs_fsmap.c
+++ b/fs/xfs/xfs_fsmap.c
@@ -465,10 +465,9 @@ xfs_getfsmap_rtdev_rtbitmap_helper(
 	struct xfs_rmap_irec		irec;
 	xfs_daddr_t			rec_daddr;
 
-	rec_daddr = XFS_FSB_TO_BB(mp, rec->ar_startblock);
-
-	irec.rm_startblock = rec->ar_startblock;
-	irec.rm_blockcount = rec->ar_blockcount;
+	irec.rm_startblock = rec->ar_startext * mp->m_sb.sb_rextsize;
+	rec_daddr = XFS_FSB_TO_BB(mp, irec.rm_startblock);
+	irec.rm_blockcount = rec->ar_extcount * mp->m_sb.sb_rextsize;
 	irec.rm_owner = XFS_RMAP_OWN_NULL;	/* "free" */
 	irec.rm_offset = 0;
 	irec.rm_flags = 0;
@@ -534,8 +533,11 @@ xfs_getfsmap_rtdev_rtbitmap_query(
 
 	xfs_ilock(tp->t_mountp->m_rbmip, XFS_ILOCK_SHARED);
 
-	alow.ar_startblock = info->low.rm_startblock;
-	ahigh.ar_startblock = info->high.rm_startblock;
+	alow.ar_startext = info->low.rm_startblock;
+	ahigh.ar_startext = info->high.rm_startblock;
+	do_div(alow.ar_startext, tp->t_mountp->m_sb.sb_rextsize);
+	if (do_div(ahigh.ar_startext, tp->t_mountp->m_sb.sb_rextsize))
+		ahigh.ar_startext++;
 	error = xfs_rtalloc_query_range(tp, &alow, &ahigh,
 			xfs_getfsmap_rtdev_rtbitmap_helper, info);
 	if (error)
diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c
index 523792768080..bc7ef18da243 100644
--- a/fs/xfs/xfs_fsops.c
+++ b/fs/xfs/xfs_fsops.c
@@ -24,85 +24,42 @@
 #include "xfs_sb.h"
 #include "xfs_mount.h"
 #include "xfs_defer.h"
-#include "xfs_da_format.h"
-#include "xfs_da_btree.h"
-#include "xfs_inode.h"
 #include "xfs_trans.h"
-#include "xfs_inode_item.h"
 #include "xfs_error.h"
 #include "xfs_btree.h"
-#include "xfs_alloc_btree.h"
 #include "xfs_alloc.h"
-#include "xfs_rmap_btree.h"
-#include "xfs_ialloc.h"
 #include "xfs_fsops.h"
-#include "xfs_itable.h"
 #include "xfs_trans_space.h"
 #include "xfs_rtalloc.h"
 #include "xfs_trace.h"
 #include "xfs_log.h"
-#include "xfs_filestream.h"
-#include "xfs_rmap.h"
+#include "xfs_ag.h"
 #include "xfs_ag_resv.h"
 
 /*
- * File system operations
+ * growfs operations
  */
-
-static struct xfs_buf *
-xfs_growfs_get_hdr_buf(
-	struct xfs_mount	*mp,
-	xfs_daddr_t		blkno,
-	size_t			numblks,
-	int			flags,
-	const struct xfs_buf_ops *ops)
-{
-	struct xfs_buf		*bp;
-
-	bp = xfs_buf_get_uncached(mp->m_ddev_targp, numblks, flags);
-	if (!bp)
-		return NULL;
-
-	xfs_buf_zero(bp, 0, BBTOB(bp->b_length));
-	bp->b_bn = blkno;
-	bp->b_maps[0].bm_bn = blkno;
-	bp->b_ops = ops;
-
-	return bp;
-}
-
 static int
 xfs_growfs_data_private(
 	xfs_mount_t		*mp,		/* mount point for filesystem */
 	xfs_growfs_data_t	*in)		/* growfs data input struct */
 {
-	xfs_agf_t		*agf;
-	struct xfs_agfl		*agfl;
-	xfs_agi_t		*agi;
-	xfs_agnumber_t		agno;
-	xfs_extlen_t		agsize;
-	xfs_extlen_t		tmpsize;
-	xfs_alloc_rec_t		*arec;
 	xfs_buf_t		*bp;
-	int			bucket;
-	int			dpct;
-	int			error, saved_error = 0;
+	int			error;
 	xfs_agnumber_t		nagcount;
 	xfs_agnumber_t		nagimax = 0;
 	xfs_rfsblock_t		nb, nb_mod;
 	xfs_rfsblock_t		new;
-	xfs_rfsblock_t		nfree;
 	xfs_agnumber_t		oagcount;
-	int			pct;
 	xfs_trans_t		*tp;
+	LIST_HEAD		(buffer_list);
+	struct aghdr_init_data	id = {};
 
 	nb = in->newblocks;
-	pct = in->imaxpct;
-	if (nb < mp->m_sb.sb_dblocks || pct < 0 || pct > 100)
+	if (nb < mp->m_sb.sb_dblocks)
 		return -EINVAL;
 	if ((error = xfs_sb_validate_fsb_count(&mp->m_sb, nb)))
 		return error;
-	dpct = pct - mp->m_sb.sb_imax_pct;
 	error = xfs_buf_read_uncached(mp->m_ddev_targp,
 				XFS_FSB_TO_BB(mp, nb) - XFS_FSS_TO_BB(mp, 1),
 				XFS_FSS_TO_BB(mp, 1), 0, &bp, NULL);
@@ -135,376 +92,45 @@ xfs_growfs_data_private(
 		return error;
 
 	/*
-	 * Write new AG headers to disk. Non-transactional, but written
-	 * synchronously so they are completed prior to the growfs transaction
-	 * being logged.
+	 * Write new AG headers to disk. Non-transactional, but need to be
+	 * written and completed prior to the growfs transaction being logged.
+	 * To do this, we use a delayed write buffer list and wait for
+	 * submission and IO completion of the list as a whole. This allows the
+	 * IO subsystem to merge all the AG headers in a single AG into a single
+	 * IO and hide most of the latency of the IO from us.
+	 *
+	 * This also means that if we get an error whilst building the buffer
+	 * list to write, we can cancel the entire list without having written
+	 * anything.
 	 */
-	nfree = 0;
-	for (agno = nagcount - 1; agno >= oagcount; agno--, new -= agsize) {
-		__be32	*agfl_bno;
-
-		/*
-		 * AG freespace header block
-		 */
-		bp = xfs_growfs_get_hdr_buf(mp,
-				XFS_AG_DADDR(mp, agno, XFS_AGF_DADDR(mp)),
-				XFS_FSS_TO_BB(mp, 1), 0,
-				&xfs_agf_buf_ops);
-		if (!bp) {
-			error = -ENOMEM;
-			goto error0;
-		}
-
-		agf = XFS_BUF_TO_AGF(bp);
-		agf->agf_magicnum = cpu_to_be32(XFS_AGF_MAGIC);
-		agf->agf_versionnum = cpu_to_be32(XFS_AGF_VERSION);
-		agf->agf_seqno = cpu_to_be32(agno);
-		if (agno == nagcount - 1)
-			agsize =
-				nb -
-				(agno * (xfs_rfsblock_t)mp->m_sb.sb_agblocks);
+	INIT_LIST_HEAD(&id.buffer_list);
+	for (id.agno = nagcount - 1;
+	     id.agno >= oagcount;
+	     id.agno--, new -= id.agsize) {
+
+		if (id.agno == nagcount - 1)
+			id.agsize = nb -
+				(id.agno * (xfs_rfsblock_t)mp->m_sb.sb_agblocks);
 		else
-			agsize = mp->m_sb.sb_agblocks;
-		agf->agf_length = cpu_to_be32(agsize);
-		agf->agf_roots[XFS_BTNUM_BNOi] = cpu_to_be32(XFS_BNO_BLOCK(mp));
-		agf->agf_roots[XFS_BTNUM_CNTi] = cpu_to_be32(XFS_CNT_BLOCK(mp));
-		agf->agf_levels[XFS_BTNUM_BNOi] = cpu_to_be32(1);
-		agf->agf_levels[XFS_BTNUM_CNTi] = cpu_to_be32(1);
-		if (xfs_sb_version_hasrmapbt(&mp->m_sb)) {
-			agf->agf_roots[XFS_BTNUM_RMAPi] =
-						cpu_to_be32(XFS_RMAP_BLOCK(mp));
-			agf->agf_levels[XFS_BTNUM_RMAPi] = cpu_to_be32(1);
-			agf->agf_rmap_blocks = cpu_to_be32(1);
-		}
-
-		agf->agf_flfirst = cpu_to_be32(1);
-		agf->agf_fllast = 0;
-		agf->agf_flcount = 0;
-		tmpsize = agsize - mp->m_ag_prealloc_blocks;
-		agf->agf_freeblks = cpu_to_be32(tmpsize);
-		agf->agf_longest = cpu_to_be32(tmpsize);
-		if (xfs_sb_version_hascrc(&mp->m_sb))
-			uuid_copy(&agf->agf_uuid, &mp->m_sb.sb_meta_uuid);
-		if (xfs_sb_version_hasreflink(&mp->m_sb)) {
-			agf->agf_refcount_root = cpu_to_be32(
-					xfs_refc_block(mp));
-			agf->agf_refcount_level = cpu_to_be32(1);
-			agf->agf_refcount_blocks = cpu_to_be32(1);
-		}
-
-		error = xfs_bwrite(bp);
-		xfs_buf_relse(bp);
-		if (error)
-			goto error0;
-
-		/*
-		 * AG freelist header block
-		 */
-		bp = xfs_growfs_get_hdr_buf(mp,
-				XFS_AG_DADDR(mp, agno, XFS_AGFL_DADDR(mp)),
-				XFS_FSS_TO_BB(mp, 1), 0,
-				&xfs_agfl_buf_ops);
-		if (!bp) {
-			error = -ENOMEM;
-			goto error0;
-		}
-
-		agfl = XFS_BUF_TO_AGFL(bp);
-		if (xfs_sb_version_hascrc(&mp->m_sb)) {
-			agfl->agfl_magicnum = cpu_to_be32(XFS_AGFL_MAGIC);
-			agfl->agfl_seqno = cpu_to_be32(agno);
-			uuid_copy(&agfl->agfl_uuid, &mp->m_sb.sb_meta_uuid);
-		}
-
-		agfl_bno = XFS_BUF_TO_AGFL_BNO(mp, bp);
-		for (bucket = 0; bucket < xfs_agfl_size(mp); bucket++)
-			agfl_bno[bucket] = cpu_to_be32(NULLAGBLOCK);
-
-		error = xfs_bwrite(bp);
-		xfs_buf_relse(bp);
-		if (error)
-			goto error0;
-
-		/*
-		 * AG inode header block
-		 */
-		bp = xfs_growfs_get_hdr_buf(mp,
-				XFS_AG_DADDR(mp, agno, XFS_AGI_DADDR(mp)),
-				XFS_FSS_TO_BB(mp, 1), 0,
-				&xfs_agi_buf_ops);
-		if (!bp) {
-			error = -ENOMEM;
-			goto error0;
-		}
-
-		agi = XFS_BUF_TO_AGI(bp);
-		agi->agi_magicnum = cpu_to_be32(XFS_AGI_MAGIC);
-		agi->agi_versionnum = cpu_to_be32(XFS_AGI_VERSION);
-		agi->agi_seqno = cpu_to_be32(agno);
-		agi->agi_length = cpu_to_be32(agsize);
-		agi->agi_count = 0;
-		agi->agi_root = cpu_to_be32(XFS_IBT_BLOCK(mp));
-		agi->agi_level = cpu_to_be32(1);
-		agi->agi_freecount = 0;
-		agi->agi_newino = cpu_to_be32(NULLAGINO);
-		agi->agi_dirino = cpu_to_be32(NULLAGINO);
-		if (xfs_sb_version_hascrc(&mp->m_sb))
-			uuid_copy(&agi->agi_uuid, &mp->m_sb.sb_meta_uuid);
-		if (xfs_sb_version_hasfinobt(&mp->m_sb)) {
-			agi->agi_free_root = cpu_to_be32(XFS_FIBT_BLOCK(mp));
-			agi->agi_free_level = cpu_to_be32(1);
-		}
-		for (bucket = 0; bucket < XFS_AGI_UNLINKED_BUCKETS; bucket++)
-			agi->agi_unlinked[bucket] = cpu_to_be32(NULLAGINO);
-
-		error = xfs_bwrite(bp);
-		xfs_buf_relse(bp);
-		if (error)
-			goto error0;
+			id.agsize = mp->m_sb.sb_agblocks;
 
-		/*
-		 * BNO btree root block
-		 */
-		bp = xfs_growfs_get_hdr_buf(mp,
-				XFS_AGB_TO_DADDR(mp, agno, XFS_BNO_BLOCK(mp)),
-				BTOBB(mp->m_sb.sb_blocksize), 0,
-				&xfs_allocbt_buf_ops);
-
-		if (!bp) {
-			error = -ENOMEM;
-			goto error0;
-		}
-
-		xfs_btree_init_block(mp, bp, XFS_BTNUM_BNO, 0, 1, agno, 0);
-
-		arec = XFS_ALLOC_REC_ADDR(mp, XFS_BUF_TO_BLOCK(bp), 1);
-		arec->ar_startblock = cpu_to_be32(mp->m_ag_prealloc_blocks);
-		arec->ar_blockcount = cpu_to_be32(
-			agsize - be32_to_cpu(arec->ar_startblock));
-
-		error = xfs_bwrite(bp);
-		xfs_buf_relse(bp);
-		if (error)
-			goto error0;
-
-		/*
-		 * CNT btree root block
-		 */
-		bp = xfs_growfs_get_hdr_buf(mp,
-				XFS_AGB_TO_DADDR(mp, agno, XFS_CNT_BLOCK(mp)),
-				BTOBB(mp->m_sb.sb_blocksize), 0,
-				&xfs_allocbt_buf_ops);
-		if (!bp) {
-			error = -ENOMEM;
-			goto error0;
-		}
-
-		xfs_btree_init_block(mp, bp, XFS_BTNUM_CNT, 0, 1, agno, 0);
-
-		arec = XFS_ALLOC_REC_ADDR(mp, XFS_BUF_TO_BLOCK(bp), 1);
-		arec->ar_startblock = cpu_to_be32(mp->m_ag_prealloc_blocks);
-		arec->ar_blockcount = cpu_to_be32(
-			agsize - be32_to_cpu(arec->ar_startblock));
-		nfree += be32_to_cpu(arec->ar_blockcount);
-
-		error = xfs_bwrite(bp);
-		xfs_buf_relse(bp);
-		if (error)
-			goto error0;
-
-		/* RMAP btree root block */
-		if (xfs_sb_version_hasrmapbt(&mp->m_sb)) {
-			struct xfs_rmap_rec	*rrec;
-			struct xfs_btree_block	*block;
-
-			bp = xfs_growfs_get_hdr_buf(mp,
-				XFS_AGB_TO_DADDR(mp, agno, XFS_RMAP_BLOCK(mp)),
-				BTOBB(mp->m_sb.sb_blocksize), 0,
-				&xfs_rmapbt_buf_ops);
-			if (!bp) {
-				error = -ENOMEM;
-				goto error0;
-			}
-
-			xfs_btree_init_block(mp, bp, XFS_BTNUM_RMAP, 0, 0,
-						agno, 0);
-			block = XFS_BUF_TO_BLOCK(bp);
-
-
-			/*
-			 * mark the AG header regions as static metadata The BNO
-			 * btree block is the first block after the headers, so
-			 * it's location defines the size of region the static
-			 * metadata consumes.
-			 *
-			 * Note: unlike mkfs, we never have to account for log
-			 * space when growing the data regions
-			 */
-			rrec = XFS_RMAP_REC_ADDR(block, 1);
-			rrec->rm_startblock = 0;
-			rrec->rm_blockcount = cpu_to_be32(XFS_BNO_BLOCK(mp));
-			rrec->rm_owner = cpu_to_be64(XFS_RMAP_OWN_FS);
-			rrec->rm_offset = 0;
-			be16_add_cpu(&block->bb_numrecs, 1);
-
-			/* account freespace btree root blocks */
-			rrec = XFS_RMAP_REC_ADDR(block, 2);
-			rrec->rm_startblock = cpu_to_be32(XFS_BNO_BLOCK(mp));
-			rrec->rm_blockcount = cpu_to_be32(2);
-			rrec->rm_owner = cpu_to_be64(XFS_RMAP_OWN_AG);
-			rrec->rm_offset = 0;
-			be16_add_cpu(&block->bb_numrecs, 1);
-
-			/* account inode btree root blocks */
-			rrec = XFS_RMAP_REC_ADDR(block, 3);
-			rrec->rm_startblock = cpu_to_be32(XFS_IBT_BLOCK(mp));
-			rrec->rm_blockcount = cpu_to_be32(XFS_RMAP_BLOCK(mp) -
-							XFS_IBT_BLOCK(mp));
-			rrec->rm_owner = cpu_to_be64(XFS_RMAP_OWN_INOBT);
-			rrec->rm_offset = 0;
-			be16_add_cpu(&block->bb_numrecs, 1);
-
-			/* account for rmap btree root */
-			rrec = XFS_RMAP_REC_ADDR(block, 4);
-			rrec->rm_startblock = cpu_to_be32(XFS_RMAP_BLOCK(mp));
-			rrec->rm_blockcount = cpu_to_be32(1);
-			rrec->rm_owner = cpu_to_be64(XFS_RMAP_OWN_AG);
-			rrec->rm_offset = 0;
-			be16_add_cpu(&block->bb_numrecs, 1);
-
-			/* account for refc btree root */
-			if (xfs_sb_version_hasreflink(&mp->m_sb)) {
-				rrec = XFS_RMAP_REC_ADDR(block, 5);
-				rrec->rm_startblock = cpu_to_be32(
-						xfs_refc_block(mp));
-				rrec->rm_blockcount = cpu_to_be32(1);
-				rrec->rm_owner = cpu_to_be64(XFS_RMAP_OWN_REFC);
-				rrec->rm_offset = 0;
-				be16_add_cpu(&block->bb_numrecs, 1);
-			}
-
-			error = xfs_bwrite(bp);
-			xfs_buf_relse(bp);
-			if (error)
-				goto error0;
-		}
-
-		/*
-		 * INO btree root block
-		 */
-		bp = xfs_growfs_get_hdr_buf(mp,
-				XFS_AGB_TO_DADDR(mp, agno, XFS_IBT_BLOCK(mp)),
-				BTOBB(mp->m_sb.sb_blocksize), 0,
-				&xfs_inobt_buf_ops);
-		if (!bp) {
-			error = -ENOMEM;
-			goto error0;
-		}
-
-		xfs_btree_init_block(mp, bp, XFS_BTNUM_INO , 0, 0, agno, 0);
-
-		error = xfs_bwrite(bp);
-		xfs_buf_relse(bp);
-		if (error)
-			goto error0;
-
-		/*
-		 * FINO btree root block
-		 */
-		if (xfs_sb_version_hasfinobt(&mp->m_sb)) {
-			bp = xfs_growfs_get_hdr_buf(mp,
-				XFS_AGB_TO_DADDR(mp, agno, XFS_FIBT_BLOCK(mp)),
-				BTOBB(mp->m_sb.sb_blocksize), 0,
-				&xfs_inobt_buf_ops);
-			if (!bp) {
-				error = -ENOMEM;
-				goto error0;
-			}
-
-			xfs_btree_init_block(mp, bp, XFS_BTNUM_FINO,
-						     0, 0, agno, 0);
-
-			error = xfs_bwrite(bp);
-			xfs_buf_relse(bp);
-			if (error)
-				goto error0;
-		}
-
-		/*
-		 * refcount btree root block
-		 */
-		if (xfs_sb_version_hasreflink(&mp->m_sb)) {
-			bp = xfs_growfs_get_hdr_buf(mp,
-				XFS_AGB_TO_DADDR(mp, agno, xfs_refc_block(mp)),
-				BTOBB(mp->m_sb.sb_blocksize), 0,
-				&xfs_refcountbt_buf_ops);
-			if (!bp) {
-				error = -ENOMEM;
-				goto error0;
-			}
-
-			xfs_btree_init_block(mp, bp, XFS_BTNUM_REFC,
-					     0, 0, agno, 0);
-
-			error = xfs_bwrite(bp);
-			xfs_buf_relse(bp);
-			if (error)
-				goto error0;
-		}
-	}
-	xfs_trans_agblocks_delta(tp, nfree);
-	/*
-	 * There are new blocks in the old last a.g.
-	 */
-	if (new) {
-		struct xfs_owner_info	oinfo;
-
-		/*
-		 * Change the agi length.
-		 */
-		error = xfs_ialloc_read_agi(mp, tp, agno, &bp);
-		if (error) {
-			goto error0;
-		}
-		ASSERT(bp);
-		agi = XFS_BUF_TO_AGI(bp);
-		be32_add_cpu(&agi->agi_length, new);
-		ASSERT(nagcount == oagcount ||
-		       be32_to_cpu(agi->agi_length) == mp->m_sb.sb_agblocks);
-		xfs_ialloc_log_agi(tp, bp, XFS_AGI_LENGTH);
-		/*
-		 * Change agf length.
-		 */
-		error = xfs_alloc_read_agf(mp, tp, agno, 0, &bp);
+		error = xfs_ag_init_headers(mp, &id);
 		if (error) {
-			goto error0;
+			xfs_buf_delwri_cancel(&id.buffer_list);
+			goto out_trans_cancel;
 		}
-		ASSERT(bp);
-		agf = XFS_BUF_TO_AGF(bp);
-		be32_add_cpu(&agf->agf_length, new);
-		ASSERT(be32_to_cpu(agf->agf_length) ==
-		       be32_to_cpu(agi->agi_length));
+	}
+	error = xfs_buf_delwri_submit(&id.buffer_list);
+	if (error)
+		goto out_trans_cancel;
 
-		xfs_alloc_log_agf(tp, bp, XFS_AGF_LENGTH);
+	xfs_trans_agblocks_delta(tp, id.nfree);
 
-		/*
-		 * Free the new space.
-		 *
-		 * XFS_RMAP_OWN_NULL is used here to tell the rmap btree that
-		 * this doesn't actually exist in the rmap btree.
-		 */
-		xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_NULL);
-		error = xfs_rmap_free(tp, bp, agno,
-				be32_to_cpu(agf->agf_length) - new,
-				new, &oinfo);
-		if (error)
-			goto error0;
-		error = xfs_free_extent(tp,
-				XFS_AGB_TO_FSB(mp, agno,
-					be32_to_cpu(agf->agf_length) - new),
-				new, &oinfo, XFS_AG_RESV_NONE);
+	/* If there are new blocks in the old last AG, extend it. */
+	if (new) {
+		error = xfs_ag_extend_space(mp, tp, &id, new);
 		if (error)
-			goto error0;
+			goto out_trans_cancel;
 	}
 
 	/*
@@ -517,10 +143,8 @@ xfs_growfs_data_private(
 	if (nb > mp->m_sb.sb_dblocks)
 		xfs_trans_mod_sb(tp, XFS_TRANS_SB_DBLOCKS,
 				 nb - mp->m_sb.sb_dblocks);
-	if (nfree)
-		xfs_trans_mod_sb(tp, XFS_TRANS_SB_FDBLOCKS, nfree);
-	if (dpct)
-		xfs_trans_mod_sb(tp, XFS_TRANS_SB_IMAXPCT, dpct);
+	if (id.nfree)
+		xfs_trans_mod_sb(tp, XFS_TRANS_SB_FDBLOCKS, id.nfree);
 	xfs_trans_set_sync(tp);
 	error = xfs_trans_commit(tp);
 	if (error)
@@ -529,12 +153,6 @@ xfs_growfs_data_private(
 	/* New allocation groups fully initialized, so update mount struct */
 	if (nagimax)
 		mp->m_maxagi = nagimax;
-	if (mp->m_sb.sb_imax_pct) {
-		uint64_t icount = mp->m_sb.sb_dblocks * mp->m_sb.sb_imax_pct;
-		do_div(icount, 100);
-		mp->m_maxicount = icount << mp->m_sb.sb_inopblog;
-	} else
-		mp->m_maxicount = 0;
 	xfs_set_low_space_thresholds(mp);
 	mp->m_alloc_set_aside = xfs_alloc_set_aside(mp);
 
@@ -545,73 +163,24 @@ xfs_growfs_data_private(
 	if (new) {
 		struct xfs_perag	*pag;
 
-		pag = xfs_perag_get(mp, agno);
+		pag = xfs_perag_get(mp, id.agno);
 		error = xfs_ag_resv_free(pag);
 		xfs_perag_put(pag);
 		if (error)
-			goto out;
+			return error;
 	}
 
-	/* Reserve AG metadata blocks. */
+	/*
+	 * Reserve AG metadata blocks. ENOSPC here does not mean there was a
+	 * growfs failure, just that there still isn't space for new user data
+	 * after the grow has been run.
+	 */
 	error = xfs_fs_reserve_ag_blocks(mp);
-	if (error && error != -ENOSPC)
-		goto out;
-
-	/* update secondary superblocks. */
-	for (agno = 1; agno < nagcount; agno++) {
+	if (error == -ENOSPC)
 		error = 0;
-		/*
-		 * new secondary superblocks need to be zeroed, not read from
-		 * disk as the contents of the new area we are growing into is
-		 * completely unknown.
-		 */
-		if (agno < oagcount) {
-			error = xfs_trans_read_buf(mp, NULL, mp->m_ddev_targp,
-				  XFS_AGB_TO_DADDR(mp, agno, XFS_SB_BLOCK(mp)),
-				  XFS_FSS_TO_BB(mp, 1), 0, &bp,
-				  &xfs_sb_buf_ops);
-		} else {
-			bp = xfs_trans_get_buf(NULL, mp->m_ddev_targp,
-				  XFS_AGB_TO_DADDR(mp, agno, XFS_SB_BLOCK(mp)),
-				  XFS_FSS_TO_BB(mp, 1), 0);
-			if (bp) {
-				bp->b_ops = &xfs_sb_buf_ops;
-				xfs_buf_zero(bp, 0, BBTOB(bp->b_length));
-			} else
-				error = -ENOMEM;
-		}
-
-		/*
-		 * If we get an error reading or writing alternate superblocks,
-		 * continue.  xfs_repair chooses the "best" superblock based
-		 * on most matches; if we break early, we'll leave more
-		 * superblocks un-updated than updated, and xfs_repair may
-		 * pick them over the properly-updated primary.
-		 */
-		if (error) {
-			xfs_warn(mp,
-		"error %d reading secondary superblock for ag %d",
-				error, agno);
-			saved_error = error;
-			continue;
-		}
-		xfs_sb_to_disk(XFS_BUF_TO_SBP(bp), &mp->m_sb);
-
-		error = xfs_bwrite(bp);
-		xfs_buf_relse(bp);
-		if (error) {
-			xfs_warn(mp,
-		"write error %d updating secondary superblock for ag %d",
-				error, agno);
-			saved_error = error;
-			continue;
-		}
-	}
-
- out:
-	return saved_error ? saved_error : error;
+	return error;
 
- error0:
+out_trans_cancel:
 	xfs_trans_cancel(tp);
 	return error;
 }
@@ -638,25 +207,71 @@ xfs_growfs_log_private(
 	return -ENOSYS;
 }
 
+static int
+xfs_growfs_imaxpct(
+	struct xfs_mount	*mp,
+	__u32			imaxpct)
+{
+	struct xfs_trans	*tp;
+	int			dpct;
+	int			error;
+
+	if (imaxpct > 100)
+		return -EINVAL;
+
+	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_growdata,
+			XFS_GROWFS_SPACE_RES(mp), 0, XFS_TRANS_RESERVE, &tp);
+	if (error)
+		return error;
+
+	dpct = imaxpct - mp->m_sb.sb_imax_pct;
+	xfs_trans_mod_sb(tp, XFS_TRANS_SB_IMAXPCT, dpct);
+	xfs_trans_set_sync(tp);
+	return xfs_trans_commit(tp);
+}
+
 /*
  * protected versions of growfs function acquire and release locks on the mount
  * point - exported through ioctls: XFS_IOC_FSGROWFSDATA, XFS_IOC_FSGROWFSLOG,
  * XFS_IOC_FSGROWFSRT
  */
-
-
 int
 xfs_growfs_data(
-	xfs_mount_t		*mp,
-	xfs_growfs_data_t	*in)
+	struct xfs_mount	*mp,
+	struct xfs_growfs_data	*in)
 {
-	int error;
+	int			error = 0;
 
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
 	if (!mutex_trylock(&mp->m_growlock))
 		return -EWOULDBLOCK;
-	error = xfs_growfs_data_private(mp, in);
+
+	/* update imaxpct separately to the physical grow of the filesystem */
+	if (in->imaxpct != mp->m_sb.sb_imax_pct) {
+		error = xfs_growfs_imaxpct(mp, in->imaxpct);
+		if (error)
+			goto out_error;
+	}
+
+	if (in->newblocks != mp->m_sb.sb_dblocks) {
+		error = xfs_growfs_data_private(mp, in);
+		if (error)
+			goto out_error;
+	}
+
+	/* Post growfs calculations needed to reflect new state in operations */
+	if (mp->m_sb.sb_imax_pct) {
+		uint64_t icount = mp->m_sb.sb_dblocks * mp->m_sb.sb_imax_pct;
+		do_div(icount, 100);
+		mp->m_maxicount = icount << mp->m_sb.sb_inopblog;
+	} else
+		mp->m_maxicount = 0;
+
+	/* Update secondary superblocks now the physical grow has completed */
+	error = xfs_update_secondary_sbs(mp);
+
+out_error:
 	/*
 	 * Increment the generation unconditionally, the error could be from
 	 * updating the secondary superblocks, in which case the new size
diff --git a/fs/xfs/xfs_globals.c b/fs/xfs/xfs_globals.c
index 3e1cc3001bcb..fdde17a2333c 100644
--- a/fs/xfs/xfs_globals.c
+++ b/fs/xfs/xfs_globals.c
@@ -47,6 +47,7 @@ xfs_param_t xfs_params = {
 
 struct xfs_globals xfs_globals = {
 	.log_recovery_delay	=	0,	/* no delay by default */
+	.mount_delay		=	0,	/* no delay by default */
 #ifdef XFS_ASSERT_FATAL
 	.bug_on_assert		=	true,	/* assert failures BUG() */
 #else
diff --git a/fs/xfs/xfs_icache.c b/fs/xfs/xfs_icache.c
index 9a18f69f6e96..164350d91efc 100644
--- a/fs/xfs/xfs_icache.c
+++ b/fs/xfs/xfs_icache.c
@@ -107,7 +107,8 @@ xfs_inode_free_callback(
 		xfs_idestroy_fork(ip, XFS_COW_FORK);
 
 	if (ip->i_itemp) {
-		ASSERT(!(ip->i_itemp->ili_item.li_flags & XFS_LI_IN_AIL));
+		ASSERT(!test_bit(XFS_LI_IN_AIL,
+				 &ip->i_itemp->ili_item.li_flags));
 		xfs_inode_item_destroy(ip);
 		ip->i_itemp = NULL;
 	}
@@ -309,6 +310,46 @@ xfs_reinit_inode(
 }
 
 /*
+ * If we are allocating a new inode, then check what was returned is
+ * actually a free, empty inode. If we are not allocating an inode,
+ * then check we didn't find a free inode.
+ *
+ * Returns:
+ *	0		if the inode free state matches the lookup context
+ *	-ENOENT		if the inode is free and we are not allocating
+ *	-EFSCORRUPTED	if there is any state mismatch at all
+ */
+static int
+xfs_iget_check_free_state(
+	struct xfs_inode	*ip,
+	int			flags)
+{
+	if (flags & XFS_IGET_CREATE) {
+		/* should be a free inode */
+		if (VFS_I(ip)->i_mode != 0) {
+			xfs_warn(ip->i_mount,
+"Corruption detected! Free inode 0x%llx not marked free! (mode 0x%x)",
+				ip->i_ino, VFS_I(ip)->i_mode);
+			return -EFSCORRUPTED;
+		}
+
+		if (ip->i_d.di_nblocks != 0) {
+			xfs_warn(ip->i_mount,
+"Corruption detected! Free inode 0x%llx has blocks allocated!",
+				ip->i_ino);
+			return -EFSCORRUPTED;
+		}
+		return 0;
+	}
+
+	/* should be an allocated inode */
+	if (VFS_I(ip)->i_mode == 0)
+		return -ENOENT;
+
+	return 0;
+}
+
+/*
  * Check the validity of the inode we just found it the cache
  */
 static int
@@ -357,12 +398,12 @@ xfs_iget_cache_hit(
 	}
 
 	/*
-	 * If lookup is racing with unlink return an error immediately.
+	 * Check the inode free state is valid. This also detects lookup
+	 * racing with unlinks.
 	 */
-	if (VFS_I(ip)->i_mode == 0 && !(flags & XFS_IGET_CREATE)) {
-		error = -ENOENT;
+	error = xfs_iget_check_free_state(ip, flags);
+	if (error)
 		goto out_error;
-	}
 
 	/*
 	 * If IRECLAIMABLE is set, we've torn down the VFS inode already.
@@ -485,29 +526,12 @@ xfs_iget_cache_miss(
 
 
 	/*
-	 * If we are allocating a new inode, then check what was returned is
-	 * actually a free, empty inode. If we are not allocating an inode,
-	 * the check we didn't find a free inode.
+	 * Check the inode free state is valid. This also detects lookup
+	 * racing with unlinks.
 	 */
-	if (flags & XFS_IGET_CREATE) {
-		if (VFS_I(ip)->i_mode != 0) {
-			xfs_warn(mp,
-"Corruption detected! Free inode 0x%llx not marked free on disk",
-				ino);
-			error = -EFSCORRUPTED;
-			goto out_destroy;
-		}
-		if (ip->i_d.di_nblocks != 0) {
-			xfs_warn(mp,
-"Corruption detected! Free inode 0x%llx has blocks allocated!",
-				ino);
-			error = -EFSCORRUPTED;
-			goto out_destroy;
-		}
-	} else if (VFS_I(ip)->i_mode == 0) {
-		error = -ENOENT;
+	error = xfs_iget_check_free_state(ip, flags);
+	if (error)
 		goto out_destroy;
-	}
 
 	/*
 	 * Preload the radix tree so we can insert safely under the
@@ -1802,3 +1826,21 @@ xfs_inode_clear_cowblocks_tag(
 	return __xfs_inode_clear_blocks_tag(ip,
 			trace_xfs_perag_clear_cowblocks, XFS_ICI_COWBLOCKS_TAG);
 }
+
+/* Disable post-EOF and CoW block auto-reclamation. */
+void
+xfs_icache_disable_reclaim(
+	struct xfs_mount	*mp)
+{
+	cancel_delayed_work_sync(&mp->m_eofblocks_work);
+	cancel_delayed_work_sync(&mp->m_cowblocks_work);
+}
+
+/* Enable post-EOF and CoW block auto-reclamation. */
+void
+xfs_icache_enable_reclaim(
+	struct xfs_mount	*mp)
+{
+	xfs_queue_eofblocks(mp);
+	xfs_queue_cowblocks(mp);
+}
diff --git a/fs/xfs/xfs_icache.h b/fs/xfs/xfs_icache.h
index d4a77588eca1..d69a0f5a6a73 100644
--- a/fs/xfs/xfs_icache.h
+++ b/fs/xfs/xfs_icache.h
@@ -131,4 +131,7 @@ xfs_fs_eofblocks_from_user(
 int xfs_icache_inode_is_allocated(struct xfs_mount *mp, struct xfs_trans *tp,
 				  xfs_ino_t ino, bool *inuse);
 
+void xfs_icache_disable_reclaim(struct xfs_mount *mp);
+void xfs_icache_enable_reclaim(struct xfs_mount *mp);
+
 #endif
diff --git a/fs/xfs/xfs_icreate_item.c b/fs/xfs/xfs_icreate_item.c
index 865ad1373e5e..5da9599156ed 100644
--- a/fs/xfs/xfs_icreate_item.c
+++ b/fs/xfs/xfs_icreate_item.c
@@ -91,7 +91,7 @@ xfs_icreate_item_unlock(
 {
 	struct xfs_icreate_item	*icp = ICR_ITEM(lip);
 
-	if (icp->ic_item.li_flags & XFS_LI_ABORTED)
+	if (test_bit(XFS_LI_ABORTED, &lip->li_flags))
 		kmem_zone_free(xfs_icreate_zone, icp);
 	return;
 }
@@ -184,5 +184,5 @@ xfs_icreate_log(
 
 	xfs_trans_add_item(tp, &icp->ic_item);
 	tp->t_flags |= XFS_TRANS_DIRTY;
-	icp->ic_item.li_desc->lid_flags |= XFS_LID_DIRTY;
+	set_bit(XFS_LI_DIRTY, &icp->ic_item.li_flags);
 }
diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
index 2b70c8b4cee2..05207a64dd53 100644
--- a/fs/xfs/xfs_inode.c
+++ b/fs/xfs/xfs_inode.c
@@ -498,7 +498,7 @@ again:
 		if (!try_lock) {
 			for (j = (i - 1); j >= 0 && !try_lock; j--) {
 				lp = (xfs_log_item_t *)ips[j]->i_itemp;
-				if (lp && (lp->li_flags & XFS_LI_IN_AIL))
+				if (lp && test_bit(XFS_LI_IN_AIL, &lp->li_flags))
 					try_lock++;
 			}
 		}
@@ -598,7 +598,7 @@ xfs_lock_two_inodes(
 	 * and try again.
 	 */
 	lp = (xfs_log_item_t *)ip0->i_itemp;
-	if (lp && (lp->li_flags & XFS_LI_IN_AIL)) {
+	if (lp && test_bit(XFS_LI_IN_AIL, &lp->li_flags)) {
 		if (!xfs_ilock_nowait(ip1, xfs_lock_inumorder(ip1_mode, 1))) {
 			xfs_iunlock(ip0, ip0_mode);
 			if ((++attempts % 5) == 0)
@@ -791,6 +791,18 @@ xfs_ialloc(
 	ASSERT(*ialloc_context == NULL);
 
 	/*
+	 * Protect against obviously corrupt allocation btree records. Later
+	 * xfs_iget checks will catch re-allocation of other active in-memory
+	 * and on-disk inodes. If we don't catch reallocating the parent inode
+	 * here we will deadlock in xfs_iget() so we have to do these checks
+	 * first.
+	 */
+	if ((pip && ino == pip->i_ino) || !xfs_verify_dir_ino(mp, ino)) {
+		xfs_alert(mp, "Allocated a known in-use inode 0x%llx!", ino);
+		return -EFSCORRUPTED;
+	}
+
+	/*
 	 * Get the in-core inode with the lock held exclusively.
 	 * This is because we're setting fields here we need
 	 * to prevent others from looking at until we're done.
@@ -1196,6 +1208,7 @@ xfs_create(
 	unlock_dp_on_error = true;
 
 	xfs_defer_init(&dfops, &first_block);
+	tp->t_agfl_dfops = &dfops;
 
 	/*
 	 * Reserve disk quota and the inode.
@@ -1411,11 +1424,11 @@ xfs_link(
 	if (XFS_FORCED_SHUTDOWN(mp))
 		return -EIO;
 
-	error = xfs_qm_dqattach(sip, 0);
+	error = xfs_qm_dqattach(sip);
 	if (error)
 		goto std_return;
 
-	error = xfs_qm_dqattach(tdp, 0);
+	error = xfs_qm_dqattach(tdp);
 	if (error)
 		goto std_return;
 
@@ -1451,6 +1464,7 @@ xfs_link(
 	}
 
 	xfs_defer_init(&dfops, &first_block);
+	tp->t_agfl_dfops = &dfops;
 
 	/*
 	 * Handle initial link state of O_TMPFILE inode
@@ -1534,11 +1548,12 @@ xfs_itruncate_clear_reflink_flags(
  * dirty on error so that transactions can be easily aborted if possible.
  */
 int
-xfs_itruncate_extents(
+xfs_itruncate_extents_flags(
 	struct xfs_trans	**tpp,
 	struct xfs_inode	*ip,
 	int			whichfork,
-	xfs_fsize_t		new_size)
+	xfs_fsize_t		new_size,
+	int			flags)
 {
 	struct xfs_mount	*mp = ip->i_mount;
 	struct xfs_trans	*tp = *tpp;
@@ -1561,6 +1576,8 @@ xfs_itruncate_extents(
 
 	trace_xfs_itruncate_extents_start(ip, new_size);
 
+	flags |= xfs_bmapi_aflag(whichfork);
+
 	/*
 	 * Since it is possible for space to become allocated beyond
 	 * the end of the file (in a crash where the space is allocated
@@ -1579,12 +1596,9 @@ xfs_itruncate_extents(
 	unmap_len = last_block - first_unmap_block + 1;
 	while (!done) {
 		xfs_defer_init(&dfops, &first_block);
-		error = xfs_bunmapi(tp, ip,
-				    first_unmap_block, unmap_len,
-				    xfs_bmapi_aflag(whichfork),
-				    XFS_ITRUNC_MAX_EXTENTS,
-				    &first_block, &dfops,
-				    &done);
+		error = xfs_bunmapi(tp, ip, first_unmap_block, unmap_len, flags,
+				    XFS_ITRUNC_MAX_EXTENTS, &first_block,
+				    &dfops, &done);
 		if (error)
 			goto out_bmap_cancel;
 
@@ -1811,6 +1825,7 @@ xfs_inactive_ifree(
 	xfs_trans_ijoin(tp, ip, 0);
 
 	xfs_defer_init(&dfops, &first_block);
+	tp->t_agfl_dfops = &dfops;
 	error = xfs_ifree(tp, ip, &dfops);
 	if (error) {
 		/*
@@ -1911,7 +1926,7 @@ xfs_inactive(
 	     ip->i_d.di_nextents > 0 || ip->i_delayed_blks > 0))
 		truncate = 1;
 
-	error = xfs_qm_dqattach(ip, 0);
+	error = xfs_qm_dqattach(ip);
 	if (error)
 		return;
 
@@ -2574,11 +2589,11 @@ xfs_remove(
 	if (XFS_FORCED_SHUTDOWN(mp))
 		return -EIO;
 
-	error = xfs_qm_dqattach(dp, 0);
+	error = xfs_qm_dqattach(dp);
 	if (error)
 		goto std_return;
 
-	error = xfs_qm_dqattach(ip, 0);
+	error = xfs_qm_dqattach(ip);
 	if (error)
 		goto std_return;
 
@@ -2647,6 +2662,7 @@ xfs_remove(
 		goto out_trans_cancel;
 
 	xfs_defer_init(&dfops, &first_block);
+	tp->t_agfl_dfops = &dfops;
 	error = xfs_dir_removename(tp, dp, name, ip->i_ino,
 					&first_block, &dfops, resblks);
 	if (error) {
@@ -3014,6 +3030,7 @@ xfs_rename(
 	}
 
 	xfs_defer_init(&dfops, &first_block);
+	tp->t_agfl_dfops = &dfops;
 
 	/* RENAME_EXCHANGE is unique from here on. */
 	if (flags & RENAME_EXCHANGE)
diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h
index 1eebc53df7d7..00fee6824745 100644
--- a/fs/xfs/xfs_inode.h
+++ b/fs/xfs/xfs_inode.h
@@ -415,8 +415,8 @@ uint		xfs_ilock_attr_map_shared(struct xfs_inode *);
 uint		xfs_ip2xflags(struct xfs_inode *);
 int		xfs_ifree(struct xfs_trans *, xfs_inode_t *,
 			   struct xfs_defer_ops *);
-int		xfs_itruncate_extents(struct xfs_trans **, struct xfs_inode *,
-				      int, xfs_fsize_t);
+int		xfs_itruncate_extents_flags(struct xfs_trans **,
+				struct xfs_inode *, int, xfs_fsize_t, int);
 void		xfs_iext_realloc(xfs_inode_t *, int, int);
 
 void		xfs_iunpin_wait(xfs_inode_t *);
@@ -433,6 +433,16 @@ int		xfs_dir_ialloc(struct xfs_trans **, struct xfs_inode *, umode_t,
 			       xfs_nlink_t, dev_t, prid_t,
 			       struct xfs_inode **);
 
+static inline int
+xfs_itruncate_extents(
+	struct xfs_trans	**tpp,
+	struct xfs_inode	*ip,
+	int			whichfork,
+	xfs_fsize_t		new_size)
+{
+	return xfs_itruncate_extents_flags(tpp, ip, whichfork, new_size, 0);
+}
+
 /* from xfs_file.c */
 enum xfs_prealloc_flags {
 	XFS_PREALLOC_SET	= (1 << 1),
diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c
index 34b91b789702..3e5b8574818e 100644
--- a/fs/xfs/xfs_inode_item.c
+++ b/fs/xfs/xfs_inode_item.c
@@ -518,7 +518,7 @@ xfs_inode_item_push(
 	 * The buffer containing this item failed to be written back
 	 * previously. Resubmit the buffer for IO.
 	 */
-	if (lip->li_flags & XFS_LI_FAILED) {
+	if (test_bit(XFS_LI_FAILED, &lip->li_flags)) {
 		if (!xfs_buf_trylock(bp))
 			return XFS_ITEM_LOCKED;
 
@@ -729,14 +729,14 @@ xfs_iflush_done(
 		 */
 		iip = INODE_ITEM(blip);
 		if ((iip->ili_logged && blip->li_lsn == iip->ili_flush_lsn) ||
-		    (blip->li_flags & XFS_LI_FAILED))
+		    test_bit(XFS_LI_FAILED, &blip->li_flags))
 			need_ail++;
 	}
 
 	/* make sure we capture the state of the initial inode. */
 	iip = INODE_ITEM(lip);
 	if ((iip->ili_logged && lip->li_lsn == iip->ili_flush_lsn) ||
-	    lip->li_flags & XFS_LI_FAILED)
+	    test_bit(XFS_LI_FAILED, &lip->li_flags))
 		need_ail++;
 
 	/*
@@ -803,7 +803,7 @@ xfs_iflush_abort(
 	xfs_inode_log_item_t	*iip = ip->i_itemp;
 
 	if (iip) {
-		if (iip->ili_item.li_flags & XFS_LI_IN_AIL) {
+		if (test_bit(XFS_LI_IN_AIL, &iip->ili_item.li_flags)) {
 			xfs_trans_ail_remove(&iip->ili_item,
 					     stale ? SHUTDOWN_LOG_IO_ERROR :
 						     SHUTDOWN_CORRUPT_INCORE);
diff --git a/fs/xfs/xfs_ioctl.c b/fs/xfs/xfs_ioctl.c
index 89fb1eb80aae..5dd9e22b4a4c 100644
--- a/fs/xfs/xfs_ioctl.c
+++ b/fs/xfs/xfs_ioctl.c
@@ -1103,7 +1103,8 @@ xfs_ioctl_setattr_dax_invalidate(
 	if (fa->fsx_xflags & FS_XFLAG_DAX) {
 		if (!(S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode)))
 			return -EINVAL;
-		if (bdev_dax_supported(sb, sb->s_blocksize) < 0)
+		if (!bdev_dax_supported(xfs_find_bdev_for_inode(VFS_I(ip)),
+				sb->s_blocksize))
 			return -EINVAL;
 	}
 
@@ -1811,6 +1812,88 @@ xfs_ioc_swapext(
 	return error;
 }
 
+static int
+xfs_ioc_getlabel(
+	struct xfs_mount	*mp,
+	char			__user *user_label)
+{
+	struct xfs_sb		*sbp = &mp->m_sb;
+	char			label[XFSLABEL_MAX + 1];
+
+	/* Paranoia */
+	BUILD_BUG_ON(sizeof(sbp->sb_fname) > FSLABEL_MAX);
+
+	spin_lock(&mp->m_sb_lock);
+	strncpy(label, sbp->sb_fname, sizeof(sbp->sb_fname));
+	spin_unlock(&mp->m_sb_lock);
+
+	/* xfs on-disk label is 12 chars, be sure we send a null to user */
+	label[XFSLABEL_MAX] = '\0';
+	if (copy_to_user(user_label, label, sizeof(sbp->sb_fname)))
+		return -EFAULT;
+	return 0;
+}
+
+static int
+xfs_ioc_setlabel(
+	struct file		*filp,
+	struct xfs_mount	*mp,
+	char			__user *newlabel)
+{
+	struct xfs_sb		*sbp = &mp->m_sb;
+	char			label[XFSLABEL_MAX + 1];
+	size_t			len;
+	int			error;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+	/*
+	 * The generic ioctl allows up to FSLABEL_MAX chars, but XFS is much
+	 * smaller, at 12 bytes.  We copy one more to be sure we find the
+	 * (required) NULL character to test the incoming label length.
+	 * NB: The on disk label doesn't need to be null terminated.
+	 */
+	if (copy_from_user(label, newlabel, XFSLABEL_MAX + 1))
+		return -EFAULT;
+	len = strnlen(label, XFSLABEL_MAX + 1);
+	if (len > sizeof(sbp->sb_fname))
+		return -EINVAL;
+
+	error = mnt_want_write_file(filp);
+	if (error)
+		return error;
+
+	spin_lock(&mp->m_sb_lock);
+	memset(sbp->sb_fname, 0, sizeof(sbp->sb_fname));
+	strncpy(sbp->sb_fname, label, sizeof(sbp->sb_fname));
+	spin_unlock(&mp->m_sb_lock);
+
+	/*
+	 * Now we do several things to satisfy userspace.
+	 * In addition to normal logging of the primary superblock, we also
+	 * immediately write these changes to sector zero for the primary, then
+	 * update all backup supers (as xfs_db does for a label change), then
+	 * invalidate the block device page cache.  This is so that any prior
+	 * buffered reads from userspace (i.e. from blkid) are invalidated,
+	 * and userspace will see the newly-written label.
+	 */
+	error = xfs_sync_sb_buf(mp);
+	if (error)
+		goto out;
+	/*
+	 * growfs also updates backup supers so lock against that.
+	 */
+	mutex_lock(&mp->m_growlock);
+	error = xfs_update_secondary_sbs(mp);
+	mutex_unlock(&mp->m_growlock);
+
+	invalidate_bdev(mp->m_ddev_targp->bt_bdev);
+
+out:
+	mnt_drop_write_file(filp);
+	return error;
+}
+
 /*
  * Note: some of the ioctl's return positive numbers as a
  * byte count indicating success, such as readlink_by_handle.
@@ -1834,6 +1917,10 @@ xfs_file_ioctl(
 	switch (cmd) {
 	case FITRIM:
 		return xfs_ioc_trim(mp, arg);
+	case FS_IOC_GETFSLABEL:
+		return xfs_ioc_getlabel(mp, arg);
+	case FS_IOC_SETFSLABEL:
+		return xfs_ioc_setlabel(filp, mp, arg);
 	case XFS_IOC_ALLOCSP:
 	case XFS_IOC_FREESP:
 	case XFS_IOC_RESVSP:
diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c
index 046469fcc1b8..c6ce6f9335b6 100644
--- a/fs/xfs/xfs_iomap.c
+++ b/fs/xfs/xfs_iomap.c
@@ -224,7 +224,7 @@ xfs_iomap_write_direct(
 	 * necessary and move on to transaction setup.
 	 */
 	xfs_iunlock(ip, lockmode);
-	error = xfs_qm_dqattach(ip, 0);
+	error = xfs_qm_dqattach(ip);
 	if (error)
 		return error;
 
@@ -576,7 +576,7 @@ xfs_file_iomap_begin_delay(
 		goto done;
 	}
 
-	error = xfs_qm_dqattach_locked(ip, 0);
+	error = xfs_qm_dqattach_locked(ip, false);
 	if (error)
 		goto out_unlock;
 
@@ -692,7 +692,7 @@ xfs_iomap_write_allocate(
 	/*
 	 * Make sure that the dquots are there.
 	 */
-	error = xfs_qm_dqattach(ip, 0);
+	error = xfs_qm_dqattach(ip);
 	if (error)
 		return error;
 
@@ -946,8 +946,11 @@ error_on_bmapi_transaction:
 	return error;
 }
 
-static inline bool imap_needs_alloc(struct inode *inode,
-		struct xfs_bmbt_irec *imap, int nimaps)
+static inline bool
+imap_needs_alloc(
+	struct inode		*inode,
+	struct xfs_bmbt_irec	*imap,
+	int			nimaps)
 {
 	return !nimaps ||
 		imap->br_startblock == HOLESTARTBLOCK ||
@@ -955,31 +958,58 @@ static inline bool imap_needs_alloc(struct inode *inode,
 		(IS_DAX(inode) && imap->br_state == XFS_EXT_UNWRITTEN);
 }
 
-static inline bool needs_cow_for_zeroing(struct xfs_bmbt_irec *imap, int nimaps)
+static inline bool
+needs_cow_for_zeroing(
+	struct xfs_bmbt_irec	*imap,
+	int			nimaps)
 {
 	return nimaps &&
 		imap->br_startblock != HOLESTARTBLOCK &&
 		imap->br_state != XFS_EXT_UNWRITTEN;
 }
 
-static inline bool need_excl_ilock(struct xfs_inode *ip, unsigned flags)
+static int
+xfs_ilock_for_iomap(
+	struct xfs_inode	*ip,
+	unsigned		flags,
+	unsigned		*lockmode)
 {
+	unsigned		mode = XFS_ILOCK_SHARED;
+
 	/*
 	 * COW writes may allocate delalloc space or convert unwritten COW
 	 * extents, so we need to make sure to take the lock exclusively here.
 	 */
-	if (xfs_is_reflink_inode(ip) && (flags & (IOMAP_WRITE | IOMAP_ZERO)))
-		return true;
+	if (xfs_is_reflink_inode(ip) && (flags & (IOMAP_WRITE | IOMAP_ZERO))) {
+		/*
+		 * FIXME: It could still overwrite on unshared extents and not
+		 * need allocation.
+		 */
+		if (flags & IOMAP_NOWAIT)
+			return -EAGAIN;
+		mode = XFS_ILOCK_EXCL;
+	}
 
 	/*
-	 * Extents not yet cached requires exclusive access, don't block.
-	 * This is an opencoded xfs_ilock_data_map_shared() to cater for the
+	 * Extents not yet cached requires exclusive access, don't block.  This
+	 * is an opencoded xfs_ilock_data_map_shared() call but with
 	 * non-blocking behaviour.
 	 */
-	if (ip->i_d.di_format == XFS_DINODE_FMT_BTREE &&
-	    !(ip->i_df.if_flags & XFS_IFEXTENTS))
-		return true;
-	return false;
+	if (!(ip->i_df.if_flags & XFS_IFEXTENTS)) {
+		if (flags & IOMAP_NOWAIT)
+			return -EAGAIN;
+		mode = XFS_ILOCK_EXCL;
+	}
+
+	if (flags & IOMAP_NOWAIT) {
+		if (!xfs_ilock_nowait(ip, mode))
+			return -EAGAIN;
+	} else {
+		xfs_ilock(ip, mode);
+	}
+
+	*lockmode = mode;
+	return 0;
 }
 
 static int
@@ -1007,19 +1037,15 @@ xfs_file_iomap_begin(
 		return xfs_file_iomap_begin_delay(inode, offset, length, iomap);
 	}
 
-	if (need_excl_ilock(ip, flags))
-		lockmode = XFS_ILOCK_EXCL;
-	else
-		lockmode = XFS_ILOCK_SHARED;
-
-	if (flags & IOMAP_NOWAIT) {
-		if (!(ip->i_df.if_flags & XFS_IFEXTENTS))
-			return -EAGAIN;
-		if (!xfs_ilock_nowait(ip, lockmode))
-			return -EAGAIN;
-	} else {
-		xfs_ilock(ip, lockmode);
-	}
+	/*
+	 * Lock the inode in the manner required for the specified operation and
+	 * check for as many conditions that would result in blocking as
+	 * possible. This removes most of the non-blocking checks from the
+	 * mapping code below.
+	 */
+	error = xfs_ilock_for_iomap(ip, flags, &lockmode);
+	if (error)
+		return error;
 
 	ASSERT(offset <= mp->m_super->s_maxbytes);
 	if (offset > mp->m_super->s_maxbytes - length)
@@ -1040,19 +1066,21 @@ xfs_file_iomap_begin(
 			goto out_unlock;
 	}
 
-	if (xfs_is_reflink_inode(ip) &&
-	    ((flags & IOMAP_WRITE) ||
-	     ((flags & IOMAP_ZERO) && needs_cow_for_zeroing(&imap, nimaps)))) {
+	/* Non-modifying mapping requested, so we are done */
+	if (!(flags & (IOMAP_WRITE | IOMAP_ZERO)))
+		goto out_found;
+
+	/*
+	 * Break shared extents if necessary. Checks for non-blocking IO have
+	 * been done up front, so we don't need to do them here.
+	 */
+	if (xfs_is_reflink_inode(ip)) {
+		/* if zeroing doesn't need COW allocation, then we are done. */
+		if ((flags & IOMAP_ZERO) &&
+		    !needs_cow_for_zeroing(&imap, nimaps))
+			goto out_found;
+
 		if (flags & IOMAP_DIRECT) {
-			/*
-			 * A reflinked inode will result in CoW alloc.
-			 * FIXME: It could still overwrite on unshared extents
-			 * and not need allocation.
-			 */
-			if (flags & IOMAP_NOWAIT) {
-				error = -EAGAIN;
-				goto out_unlock;
-			}
 			/* may drop and re-acquire the ilock */
 			error = xfs_reflink_allocate_cow(ip, &imap, &shared,
 					&lockmode);
@@ -1068,46 +1096,45 @@ xfs_file_iomap_begin(
 		length = XFS_FSB_TO_B(mp, end_fsb) - offset;
 	}
 
-	if ((flags & IOMAP_WRITE) && imap_needs_alloc(inode, &imap, nimaps)) {
-		/*
-		 * If nowait is set bail since we are going to make
-		 * allocations.
-		 */
-		if (flags & IOMAP_NOWAIT) {
-			error = -EAGAIN;
-			goto out_unlock;
-		}
-		/*
-		 * We cap the maximum length we map here to MAX_WRITEBACK_PAGES
-		 * pages to keep the chunks of work done where somewhat symmetric
-		 * with the work writeback does. This is a completely arbitrary
-		 * number pulled out of thin air as a best guess for initial
-		 * testing.
-		 *
-		 * Note that the values needs to be less than 32-bits wide until
-		 * the lower level functions are updated.
-		 */
-		length = min_t(loff_t, length, 1024 * PAGE_SIZE);
-		/*
-		 * xfs_iomap_write_direct() expects the shared lock. It
-		 * is unlocked on return.
-		 */
-		if (lockmode == XFS_ILOCK_EXCL)
-			xfs_ilock_demote(ip, lockmode);
-		error = xfs_iomap_write_direct(ip, offset, length, &imap,
-				nimaps);
-		if (error)
-			return error;
+	/* Don't need to allocate over holes when doing zeroing operations. */
+	if (flags & IOMAP_ZERO)
+		goto out_found;
 
-		iomap->flags = IOMAP_F_NEW;
-		trace_xfs_iomap_alloc(ip, offset, length, 0, &imap);
-	} else {
-		ASSERT(nimaps);
+	if (!imap_needs_alloc(inode, &imap, nimaps))
+		goto out_found;
 
-		xfs_iunlock(ip, lockmode);
-		trace_xfs_iomap_found(ip, offset, length, 0, &imap);
+	/* If nowait is set bail since we are going to make allocations. */
+	if (flags & IOMAP_NOWAIT) {
+		error = -EAGAIN;
+		goto out_unlock;
 	}
 
+	/*
+	 * We cap the maximum length we map to a sane size  to keep the chunks
+	 * of work done where somewhat symmetric with the work writeback does.
+	 * This is a completely arbitrary number pulled out of thin air as a
+	 * best guess for initial testing.
+	 *
+	 * Note that the values needs to be less than 32-bits wide until the
+	 * lower level functions are updated.
+	 */
+	length = min_t(loff_t, length, 1024 * PAGE_SIZE);
+
+	/*
+	 * xfs_iomap_write_direct() expects the shared lock. It is unlocked on
+	 * return.
+	 */
+	if (lockmode == XFS_ILOCK_EXCL)
+		xfs_ilock_demote(ip, lockmode);
+	error = xfs_iomap_write_direct(ip, offset, length, &imap,
+			nimaps);
+	if (error)
+		return error;
+
+	iomap->flags = IOMAP_F_NEW;
+	trace_xfs_iomap_alloc(ip, offset, length, 0, &imap);
+
+out_finish:
 	if (xfs_ipincount(ip) && (ip->i_itemp->ili_fsync_fields
 				& ~XFS_ILOG_TIMESTAMP))
 		iomap->flags |= IOMAP_F_DIRTY;
@@ -1117,6 +1144,13 @@ xfs_file_iomap_begin(
 	if (shared)
 		iomap->flags |= IOMAP_F_SHARED;
 	return 0;
+
+out_found:
+	ASSERT(nimaps);
+	xfs_iunlock(ip, lockmode);
+	trace_xfs_iomap_found(ip, offset, length, 0, &imap);
+	goto out_finish;
+
 out_unlock:
 	xfs_iunlock(ip, lockmode);
 	return error;
diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c
index df42e4cb4dc4..b0eb49bb4918 100644
--- a/fs/xfs/xfs_iops.c
+++ b/fs/xfs/xfs_iops.c
@@ -855,7 +855,7 @@ xfs_setattr_size(
 	/*
 	 * Make sure that the dquots are attached to the inode.
 	 */
-	error = xfs_qm_dqattach(ip, 0);
+	error = xfs_qm_dqattach(ip);
 	if (error)
 		return error;
 
@@ -1195,6 +1195,30 @@ static const struct inode_operations xfs_inline_symlink_inode_operations = {
 	.update_time		= xfs_vn_update_time,
 };
 
+/* Figure out if this file actually supports DAX. */
+static bool
+xfs_inode_supports_dax(
+	struct xfs_inode	*ip)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+
+	/* Only supported on non-reflinked files. */
+	if (!S_ISREG(VFS_I(ip)->i_mode) || xfs_is_reflink_inode(ip))
+		return false;
+
+	/* DAX mount option or DAX iflag must be set. */
+	if (!(mp->m_flags & XFS_MOUNT_DAX) &&
+	    !(ip->i_d.di_flags2 & XFS_DIFLAG2_DAX))
+		return false;
+
+	/* Block size must match page size */
+	if (mp->m_sb.sb_blocksize != PAGE_SIZE)
+		return false;
+
+	/* Device has to support DAX too. */
+	return xfs_find_daxdev_for_inode(VFS_I(ip)) != NULL;
+}
+
 STATIC void
 xfs_diflags_to_iflags(
 	struct inode		*inode,
@@ -1213,11 +1237,7 @@ xfs_diflags_to_iflags(
 		inode->i_flags |= S_SYNC;
 	if (flags & XFS_DIFLAG_NOATIME)
 		inode->i_flags |= S_NOATIME;
-	if (S_ISREG(inode->i_mode) &&
-	    ip->i_mount->m_sb.sb_blocksize == PAGE_SIZE &&
-	    !xfs_is_reflink_inode(ip) &&
-	    (ip->i_mount->m_flags & XFS_MOUNT_DAX ||
-	     ip->i_d.di_flags2 & XFS_DIFLAG2_DAX))
+	if (xfs_inode_supports_dax(ip))
 		inode->i_flags |= S_DAX;
 }
 
diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c
index 2fcd9ed5d075..c21039f27e39 100644
--- a/fs/xfs/xfs_log.c
+++ b/fs/xfs/xfs_log.c
@@ -1047,6 +1047,7 @@ xfs_log_item_init(
 	INIT_LIST_HEAD(&item->li_ail);
 	INIT_LIST_HEAD(&item->li_cil);
 	INIT_LIST_HEAD(&item->li_bio_list);
+	INIT_LIST_HEAD(&item->li_trans);
 }
 
 /*
@@ -2110,10 +2111,10 @@ xlog_print_tic_res(
  */
 void
 xlog_print_trans(
-	struct xfs_trans		*tp)
+	struct xfs_trans	*tp)
 {
-	struct xfs_mount		*mp = tp->t_mountp;
-	struct xfs_log_item_desc	*lidp;
+	struct xfs_mount	*mp = tp->t_mountp;
+	struct xfs_log_item	*lip;
 
 	/* dump core transaction and ticket info */
 	xfs_warn(mp, "transaction summary:");
@@ -2124,15 +2125,14 @@ xlog_print_trans(
 	xlog_print_tic_res(mp, tp->t_ticket);
 
 	/* dump each log item */
-	list_for_each_entry(lidp, &tp->t_items, lid_trans) {
-		struct xfs_log_item	*lip = lidp->lid_item;
+	list_for_each_entry(lip, &tp->t_items, li_trans) {
 		struct xfs_log_vec	*lv = lip->li_lv;
 		struct xfs_log_iovec	*vec;
 		int			i;
 
 		xfs_warn(mp, "log item: ");
 		xfs_warn(mp, "  type	= 0x%x", lip->li_type);
-		xfs_warn(mp, "  flags	= 0x%x", lip->li_flags);
+		xfs_warn(mp, "  flags	= 0x%lx", lip->li_flags);
 		if (!lv)
 			continue;
 		xfs_warn(mp, "  niovecs	= %d", lv->lv_niovecs);
diff --git a/fs/xfs/xfs_log_cil.c b/fs/xfs/xfs_log_cil.c
index 4668403b1741..c15687724728 100644
--- a/fs/xfs/xfs_log_cil.c
+++ b/fs/xfs/xfs_log_cil.c
@@ -141,10 +141,9 @@ xlog_cil_alloc_shadow_bufs(
 	struct xlog		*log,
 	struct xfs_trans	*tp)
 {
-	struct xfs_log_item_desc *lidp;
+	struct xfs_log_item	*lip;
 
-	list_for_each_entry(lidp, &tp->t_items, lid_trans) {
-		struct xfs_log_item *lip = lidp->lid_item;
+	list_for_each_entry(lip, &tp->t_items, li_trans) {
 		struct xfs_log_vec *lv;
 		int	niovecs = 0;
 		int	nbytes = 0;
@@ -152,7 +151,7 @@ xlog_cil_alloc_shadow_bufs(
 		bool	ordered = false;
 
 		/* Skip items which aren't dirty in this transaction. */
-		if (!(lidp->lid_flags & XFS_LID_DIRTY))
+		if (!test_bit(XFS_LI_DIRTY, &lip->li_flags))
 			continue;
 
 		/* get number of vecs and size of data to be stored */
@@ -317,7 +316,7 @@ xlog_cil_insert_format_items(
 	int			*diff_len,
 	int			*diff_iovecs)
 {
-	struct xfs_log_item_desc *lidp;
+	struct xfs_log_item	*lip;
 
 
 	/* Bail out if we didn't find a log item.  */
@@ -326,15 +325,14 @@ xlog_cil_insert_format_items(
 		return;
 	}
 
-	list_for_each_entry(lidp, &tp->t_items, lid_trans) {
-		struct xfs_log_item *lip = lidp->lid_item;
+	list_for_each_entry(lip, &tp->t_items, li_trans) {
 		struct xfs_log_vec *lv;
 		struct xfs_log_vec *old_lv = NULL;
 		struct xfs_log_vec *shadow;
 		bool	ordered = false;
 
 		/* Skip items which aren't dirty in this transaction. */
-		if (!(lidp->lid_flags & XFS_LID_DIRTY))
+		if (!test_bit(XFS_LI_DIRTY, &lip->li_flags))
 			continue;
 
 		/*
@@ -406,7 +404,7 @@ xlog_cil_insert_items(
 {
 	struct xfs_cil		*cil = log->l_cilp;
 	struct xfs_cil_ctx	*ctx = cil->xc_ctx;
-	struct xfs_log_item_desc *lidp;
+	struct xfs_log_item	*lip;
 	int			len = 0;
 	int			diff_iovecs = 0;
 	int			iclog_space;
@@ -479,11 +477,10 @@ xlog_cil_insert_items(
 	 * We do this here so we only need to take the CIL lock once during
 	 * the transaction commit.
 	 */
-	list_for_each_entry(lidp, &tp->t_items, lid_trans) {
-		struct xfs_log_item	*lip = lidp->lid_item;
+	list_for_each_entry(lip, &tp->t_items, li_trans) {
 
 		/* Skip items which aren't dirty in this transaction. */
-		if (!(lidp->lid_flags & XFS_LID_DIRTY))
+		if (!test_bit(XFS_LI_DIRTY, &lip->li_flags))
 			continue;
 
 		/*
@@ -1013,6 +1010,7 @@ xfs_log_commit_cil(
 		*commit_lsn = xc_commit_lsn;
 
 	xfs_log_done(mp, tp->t_ticket, NULL, regrant);
+	tp->t_ticket = NULL;
 	xfs_trans_unreserve_and_mod_sb(tp);
 
 	/*
diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index 2b2383f1895e..06a09cb948b5 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -2702,7 +2702,7 @@ xlog_recover_do_reg_buffer(
 				goto next;
 			}
 			fa = xfs_dquot_verify(mp, item->ri_buf[i].i_addr,
-					       -1, 0, 0);
+					       -1, 0);
 			if (fa) {
 				xfs_alert(mp,
 	"dquot corrupt at %pS trying to replay into block 0x%llx",
@@ -3348,7 +3348,7 @@ xlog_recover_dquot_pass2(
 	 */
 	dq_f = item->ri_buf[0].i_addr;
 	ASSERT(dq_f);
-	fa = xfs_dquot_verify(mp, recddq, dq_f->qlf_id, 0, 0);
+	fa = xfs_dquot_verify(mp, recddq, dq_f->qlf_id, 0);
 	if (fa) {
 		xfs_alert(mp, "corrupt dquot ID 0x%x in log at %pS",
 				dq_f->qlf_id, fa);
diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c
index a901b86772f8..73ed8fec0328 100644
--- a/fs/xfs/xfs_mount.c
+++ b/fs/xfs/xfs_mount.c
@@ -1072,9 +1072,7 @@ xfs_unmountfs(
 	uint64_t		resblks;
 	int			error;
 
-	cancel_delayed_work_sync(&mp->m_eofblocks_work);
-	cancel_delayed_work_sync(&mp->m_cowblocks_work);
-
+	xfs_icache_disable_reclaim(mp);
 	xfs_fs_unreserve_ag_blocks(mp);
 	xfs_qm_unmount_quotas(mp);
 	xfs_rtunmount_inodes(mp);
diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c
index ec39ae274c78..c3e014bfc848 100644
--- a/fs/xfs/xfs_qm.c
+++ b/fs/xfs/xfs_qm.c
@@ -161,10 +161,7 @@ xfs_qm_dqpurge(
 		 * to purge this dquot anyway, so we go ahead regardless.
 		 */
 		error = xfs_qm_dqflush(dqp, &bp);
-		if (error) {
-			xfs_warn(mp, "%s: dquot "PTR_FMT" flush failed",
-				__func__, dqp);
-		} else {
+		if (!error) {
 			error = xfs_bwrite(bp);
 			xfs_buf_relse(bp);
 		}
@@ -173,7 +170,7 @@ xfs_qm_dqpurge(
 
 	ASSERT(atomic_read(&dqp->q_pincount) == 0);
 	ASSERT(XFS_FORCED_SHUTDOWN(mp) ||
-	       !(dqp->q_logitem.qli_item.li_flags & XFS_LI_IN_AIL));
+		!test_bit(XFS_LI_IN_AIL, &dqp->q_logitem.qli_item.li_flags));
 
 	xfs_dqfunlock(dqp);
 	xfs_dqunlock(dqp);
@@ -265,7 +262,7 @@ xfs_qm_dqattach_one(
 	xfs_inode_t	*ip,
 	xfs_dqid_t	id,
 	uint		type,
-	uint		doalloc,
+	bool		doalloc,
 	xfs_dquot_t	**IO_idqpp)
 {
 	xfs_dquot_t	*dqp;
@@ -291,7 +288,7 @@ xfs_qm_dqattach_one(
 	 * exist on disk and we didn't ask it to allocate; ESRCH if quotas got
 	 * turned off suddenly.
 	 */
-	error = xfs_qm_dqget(ip->i_mount, ip, id, type, doalloc, &dqp);
+	error = xfs_qm_dqget_inode(ip, type, doalloc, &dqp);
 	if (error)
 		return error;
 
@@ -326,14 +323,14 @@ xfs_qm_need_dqattach(
 /*
  * Given a locked inode, attach dquot(s) to it, taking U/G/P-QUOTAON
  * into account.
- * If XFS_QMOPT_DQALLOC, the dquot(s) will be allocated if needed.
+ * If @doalloc is true, the dquot(s) will be allocated if needed.
  * Inode may get unlocked and relocked in here, and the caller must deal with
  * the consequences.
  */
 int
 xfs_qm_dqattach_locked(
 	xfs_inode_t	*ip,
-	uint		flags)
+	bool		doalloc)
 {
 	xfs_mount_t	*mp = ip->i_mount;
 	int		error = 0;
@@ -345,8 +342,7 @@ xfs_qm_dqattach_locked(
 
 	if (XFS_IS_UQUOTA_ON(mp) && !ip->i_udquot) {
 		error = xfs_qm_dqattach_one(ip, ip->i_d.di_uid, XFS_DQ_USER,
-						flags & XFS_QMOPT_DQALLOC,
-						&ip->i_udquot);
+				doalloc, &ip->i_udquot);
 		if (error)
 			goto done;
 		ASSERT(ip->i_udquot);
@@ -354,8 +350,7 @@ xfs_qm_dqattach_locked(
 
 	if (XFS_IS_GQUOTA_ON(mp) && !ip->i_gdquot) {
 		error = xfs_qm_dqattach_one(ip, ip->i_d.di_gid, XFS_DQ_GROUP,
-						flags & XFS_QMOPT_DQALLOC,
-						&ip->i_gdquot);
+				doalloc, &ip->i_gdquot);
 		if (error)
 			goto done;
 		ASSERT(ip->i_gdquot);
@@ -363,8 +358,7 @@ xfs_qm_dqattach_locked(
 
 	if (XFS_IS_PQUOTA_ON(mp) && !ip->i_pdquot) {
 		error = xfs_qm_dqattach_one(ip, xfs_get_projid(ip), XFS_DQ_PROJ,
-						flags & XFS_QMOPT_DQALLOC,
-						&ip->i_pdquot);
+				doalloc, &ip->i_pdquot);
 		if (error)
 			goto done;
 		ASSERT(ip->i_pdquot);
@@ -381,8 +375,7 @@ done:
 
 int
 xfs_qm_dqattach(
-	struct xfs_inode	*ip,
-	uint			flags)
+	struct xfs_inode	*ip)
 {
 	int			error;
 
@@ -390,7 +383,7 @@ xfs_qm_dqattach(
 		return 0;
 
 	xfs_ilock(ip, XFS_ILOCK_EXCL);
-	error = xfs_qm_dqattach_locked(ip, flags);
+	error = xfs_qm_dqattach_locked(ip, false);
 	xfs_iunlock(ip, XFS_ILOCK_EXCL);
 
 	return error;
@@ -479,11 +472,8 @@ xfs_qm_dquot_isolate(
 		spin_unlock(lru_lock);
 
 		error = xfs_qm_dqflush(dqp, &bp);
-		if (error) {
-			xfs_warn(dqp->q_mount, "%s: dquot "PTR_FMT" flush failed",
-				 __func__, dqp);
+		if (error)
 			goto out_unlock_dirty;
-		}
 
 		xfs_buf_delwri_queue(bp, &isol->buffers);
 		xfs_buf_relse(bp);
@@ -571,27 +561,88 @@ xfs_qm_set_defquota(
 {
 	xfs_dquot_t		*dqp;
 	struct xfs_def_quota    *defq;
+	struct xfs_disk_dquot	*ddqp;
 	int			error;
 
-	error = xfs_qm_dqread(mp, 0, type, 0, &dqp);
+	error = xfs_qm_dqget_uncached(mp, 0, type, &dqp);
+	if (error)
+		return;
 
-	if (!error) {
-		xfs_disk_dquot_t        *ddqp = &dqp->q_core;
+	ddqp = &dqp->q_core;
+	defq = xfs_get_defquota(dqp, qinf);
 
-		defq = xfs_get_defquota(dqp, qinf);
+	/*
+	 * Timers and warnings have been already set, let's just set the
+	 * default limits for this quota type
+	 */
+	defq->bhardlimit = be64_to_cpu(ddqp->d_blk_hardlimit);
+	defq->bsoftlimit = be64_to_cpu(ddqp->d_blk_softlimit);
+	defq->ihardlimit = be64_to_cpu(ddqp->d_ino_hardlimit);
+	defq->isoftlimit = be64_to_cpu(ddqp->d_ino_softlimit);
+	defq->rtbhardlimit = be64_to_cpu(ddqp->d_rtb_hardlimit);
+	defq->rtbsoftlimit = be64_to_cpu(ddqp->d_rtb_softlimit);
+	xfs_qm_dqdestroy(dqp);
+}
 
-		/*
-		 * Timers and warnings have been already set, let's just set the
-		 * default limits for this quota type
-		 */
-		defq->bhardlimit = be64_to_cpu(ddqp->d_blk_hardlimit);
-		defq->bsoftlimit = be64_to_cpu(ddqp->d_blk_softlimit);
-		defq->ihardlimit = be64_to_cpu(ddqp->d_ino_hardlimit);
-		defq->isoftlimit = be64_to_cpu(ddqp->d_ino_softlimit);
-		defq->rtbhardlimit = be64_to_cpu(ddqp->d_rtb_hardlimit);
-		defq->rtbsoftlimit = be64_to_cpu(ddqp->d_rtb_softlimit);
-		xfs_qm_dqdestroy(dqp);
-	}
+/* Initialize quota time limits from the root dquot. */
+static void
+xfs_qm_init_timelimits(
+	struct xfs_mount	*mp,
+	struct xfs_quotainfo	*qinf)
+{
+	struct xfs_disk_dquot	*ddqp;
+	struct xfs_dquot	*dqp;
+	uint			type;
+	int			error;
+
+	qinf->qi_btimelimit = XFS_QM_BTIMELIMIT;
+	qinf->qi_itimelimit = XFS_QM_ITIMELIMIT;
+	qinf->qi_rtbtimelimit = XFS_QM_RTBTIMELIMIT;
+	qinf->qi_bwarnlimit = XFS_QM_BWARNLIMIT;
+	qinf->qi_iwarnlimit = XFS_QM_IWARNLIMIT;
+	qinf->qi_rtbwarnlimit = XFS_QM_RTBWARNLIMIT;
+
+	/*
+	 * We try to get the limits from the superuser's limits fields.
+	 * This is quite hacky, but it is standard quota practice.
+	 *
+	 * Since we may not have done a quotacheck by this point, just read
+	 * the dquot without attaching it to any hashtables or lists.
+	 *
+	 * Timers and warnings are globally set by the first timer found in
+	 * user/group/proj quota types, otherwise a default value is used.
+	 * This should be split into different fields per quota type.
+	 */
+	if (XFS_IS_UQUOTA_RUNNING(mp))
+		type = XFS_DQ_USER;
+	else if (XFS_IS_GQUOTA_RUNNING(mp))
+		type = XFS_DQ_GROUP;
+	else
+		type = XFS_DQ_PROJ;
+	error = xfs_qm_dqget_uncached(mp, 0, type, &dqp);
+	if (error)
+		return;
+
+	ddqp = &dqp->q_core;
+	/*
+	 * The warnings and timers set the grace period given to
+	 * a user or group before he or she can not perform any
+	 * more writing. If it is zero, a default is used.
+	 */
+	if (ddqp->d_btimer)
+		qinf->qi_btimelimit = be32_to_cpu(ddqp->d_btimer);
+	if (ddqp->d_itimer)
+		qinf->qi_itimelimit = be32_to_cpu(ddqp->d_itimer);
+	if (ddqp->d_rtbtimer)
+		qinf->qi_rtbtimelimit = be32_to_cpu(ddqp->d_rtbtimer);
+	if (ddqp->d_bwarns)
+		qinf->qi_bwarnlimit = be16_to_cpu(ddqp->d_bwarns);
+	if (ddqp->d_iwarns)
+		qinf->qi_iwarnlimit = be16_to_cpu(ddqp->d_iwarns);
+	if (ddqp->d_rtbwarns)
+		qinf->qi_rtbwarnlimit = be16_to_cpu(ddqp->d_rtbwarns);
+
+	xfs_qm_dqdestroy(dqp);
 }
 
 /*
@@ -600,11 +651,10 @@ xfs_qm_set_defquota(
  */
 STATIC int
 xfs_qm_init_quotainfo(
-	xfs_mount_t	*mp)
+	struct xfs_mount	*mp)
 {
-	xfs_quotainfo_t *qinf;
-	int		error;
-	xfs_dquot_t	*dqp;
+	struct xfs_quotainfo	*qinf;
+	int			error;
 
 	ASSERT(XFS_IS_QUOTA_RUNNING(mp));
 
@@ -636,52 +686,7 @@ xfs_qm_init_quotainfo(
 
 	mp->m_qflags |= (mp->m_sb.sb_qflags & XFS_ALL_QUOTA_CHKD);
 
-	/*
-	 * We try to get the limits from the superuser's limits fields.
-	 * This is quite hacky, but it is standard quota practice.
-	 *
-	 * Since we may not have done a quotacheck by this point, just read
-	 * the dquot without attaching it to any hashtables or lists.
-	 *
-	 * Timers and warnings are globally set by the first timer found in
-	 * user/group/proj quota types, otherwise a default value is used.
-	 * This should be split into different fields per quota type.
-	 */
-	error = xfs_qm_dqread(mp, 0,
-			XFS_IS_UQUOTA_RUNNING(mp) ? XFS_DQ_USER :
-			 (XFS_IS_GQUOTA_RUNNING(mp) ? XFS_DQ_GROUP :
-			  XFS_DQ_PROJ),
-			0, &dqp);
-
-	if (!error) {
-		xfs_disk_dquot_t	*ddqp = &dqp->q_core;
-
-		/*
-		 * The warnings and timers set the grace period given to
-		 * a user or group before he or she can not perform any
-		 * more writing. If it is zero, a default is used.
-		 */
-		qinf->qi_btimelimit = ddqp->d_btimer ?
-			be32_to_cpu(ddqp->d_btimer) : XFS_QM_BTIMELIMIT;
-		qinf->qi_itimelimit = ddqp->d_itimer ?
-			be32_to_cpu(ddqp->d_itimer) : XFS_QM_ITIMELIMIT;
-		qinf->qi_rtbtimelimit = ddqp->d_rtbtimer ?
-			be32_to_cpu(ddqp->d_rtbtimer) : XFS_QM_RTBTIMELIMIT;
-		qinf->qi_bwarnlimit = ddqp->d_bwarns ?
-			be16_to_cpu(ddqp->d_bwarns) : XFS_QM_BWARNLIMIT;
-		qinf->qi_iwarnlimit = ddqp->d_iwarns ?
-			be16_to_cpu(ddqp->d_iwarns) : XFS_QM_IWARNLIMIT;
-		qinf->qi_rtbwarnlimit = ddqp->d_rtbwarns ?
-			be16_to_cpu(ddqp->d_rtbwarns) : XFS_QM_RTBWARNLIMIT;
-		xfs_qm_dqdestroy(dqp);
-	} else {
-		qinf->qi_btimelimit = XFS_QM_BTIMELIMIT;
-		qinf->qi_itimelimit = XFS_QM_ITIMELIMIT;
-		qinf->qi_rtbtimelimit = XFS_QM_RTBTIMELIMIT;
-		qinf->qi_bwarnlimit = XFS_QM_BWARNLIMIT;
-		qinf->qi_iwarnlimit = XFS_QM_IWARNLIMIT;
-		qinf->qi_rtbwarnlimit = XFS_QM_RTBWARNLIMIT;
-	}
+	xfs_qm_init_timelimits(mp, qinf);
 
 	if (XFS_IS_UQUOTA_RUNNING(mp))
 		xfs_qm_set_defquota(mp, XFS_DQ_USER, qinf);
@@ -865,9 +870,9 @@ xfs_qm_reset_dqcounts(
 		 * find uninitialised dquot blks. See comment in
 		 * xfs_dquot_verify.
 		 */
-		fa = xfs_dquot_verify(mp, ddq, id + j, type, 0);
+		fa = xfs_dqblk_verify(mp, &dqb[j], id + j, type);
 		if (fa)
-			xfs_dquot_repair(mp, ddq, id + j, type);
+			xfs_dqblk_repair(mp, &dqb[j], id + j, type);
 
 		/*
 		 * Reset type in case we are reusing group quota file for
@@ -893,7 +898,7 @@ xfs_qm_reset_dqcounts(
 }
 
 STATIC int
-xfs_qm_dqiter_bufs(
+xfs_qm_reset_dqcounts_all(
 	struct xfs_mount	*mp,
 	xfs_dqid_t		firstid,
 	xfs_fsblock_t		bno,
@@ -961,11 +966,11 @@ xfs_qm_dqiter_bufs(
 }
 
 /*
- * Iterate over all allocated USR/GRP/PRJ dquots in the system, calling a
- * caller supplied function for every chunk of dquots that we find.
+ * Iterate over all allocated dquot blocks in this quota inode, zeroing all
+ * counters for every chunk of dquots that we find.
  */
 STATIC int
-xfs_qm_dqiterate(
+xfs_qm_reset_dqcounts_buf(
 	struct xfs_mount	*mp,
 	struct xfs_inode	*qip,
 	uint			flags,
@@ -1041,7 +1046,7 @@ xfs_qm_dqiterate(
 			 * Iterate thru all the blks in the extent and
 			 * reset the counters of all the dquots inside them.
 			 */
-			error = xfs_qm_dqiter_bufs(mp, firstid,
+			error = xfs_qm_reset_dqcounts_all(mp, firstid,
 						   map[i].br_startblock,
 						   map[i].br_blockcount,
 						   flags, buffer_list);
@@ -1066,16 +1071,17 @@ out:
 STATIC int
 xfs_qm_quotacheck_dqadjust(
 	struct xfs_inode	*ip,
-	xfs_dqid_t		id,
 	uint			type,
 	xfs_qcnt_t		nblks,
 	xfs_qcnt_t		rtblks)
 {
 	struct xfs_mount	*mp = ip->i_mount;
 	struct xfs_dquot	*dqp;
+	xfs_dqid_t		id;
 	int			error;
 
-	error = xfs_qm_dqget(mp, ip, id, type, XFS_QMOPT_DQALLOC, &dqp);
+	id = xfs_qm_id_for_quotatype(ip, type);
+	error = xfs_qm_dqget(mp, id, type, true, &dqp);
 	if (error) {
 		/*
 		 * Shouldn't be able to turn off quotas here.
@@ -1148,13 +1154,10 @@ xfs_qm_dqusage_adjust(
 	}
 
 	/*
-	 * We don't _need_ to take the ilock EXCL. However, the xfs_qm_dqget
-	 * interface expects the inode to be exclusively locked because that's
-	 * the case in all other instances. It's OK that we do this because
-	 * quotacheck is done only at mount time.
+	 * We don't _need_ to take the ilock EXCL here because quotacheck runs
+	 * at mount time and therefore nobody will be racing chown/chproj.
 	 */
-	error = xfs_iget(mp, NULL, ino, XFS_IGET_DONTCACHE, XFS_ILOCK_EXCL,
-			 &ip);
+	error = xfs_iget(mp, NULL, ino, XFS_IGET_DONTCACHE, 0, &ip);
 	if (error) {
 		*res = BULKSTAT_RV_NOTHING;
 		return error;
@@ -1189,33 +1192,31 @@ xfs_qm_dqusage_adjust(
 	 * and quotaoffs don't race. (Quotachecks happen at mount time only).
 	 */
 	if (XFS_IS_UQUOTA_ON(mp)) {
-		error = xfs_qm_quotacheck_dqadjust(ip, ip->i_d.di_uid,
-						   XFS_DQ_USER, nblks, rtblks);
+		error = xfs_qm_quotacheck_dqadjust(ip, XFS_DQ_USER, nblks,
+				rtblks);
 		if (error)
 			goto error0;
 	}
 
 	if (XFS_IS_GQUOTA_ON(mp)) {
-		error = xfs_qm_quotacheck_dqadjust(ip, ip->i_d.di_gid,
-						   XFS_DQ_GROUP, nblks, rtblks);
+		error = xfs_qm_quotacheck_dqadjust(ip, XFS_DQ_GROUP, nblks,
+				rtblks);
 		if (error)
 			goto error0;
 	}
 
 	if (XFS_IS_PQUOTA_ON(mp)) {
-		error = xfs_qm_quotacheck_dqadjust(ip, xfs_get_projid(ip),
-						   XFS_DQ_PROJ, nblks, rtblks);
+		error = xfs_qm_quotacheck_dqadjust(ip, XFS_DQ_PROJ, nblks,
+				rtblks);
 		if (error)
 			goto error0;
 	}
 
-	xfs_iunlock(ip, XFS_ILOCK_EXCL);
 	IRELE(ip);
 	*res = BULKSTAT_RV_DIDONE;
 	return 0;
 
 error0:
-	xfs_iunlock(ip, XFS_ILOCK_EXCL);
 	IRELE(ip);
 	*res = BULKSTAT_RV_GIVEUP;
 	return error;
@@ -1247,9 +1248,8 @@ xfs_qm_flush_one(
 	 */
 	if (!xfs_dqflock_nowait(dqp)) {
 		/* buf is pinned in-core by delwri list */
-		DEFINE_SINGLE_BUF_MAP(map, dqp->q_blkno,
-				      mp->m_quotainfo->qi_dqchunklen);
-		bp = _xfs_buf_find(mp->m_ddev_targp, &map, 1, 0, NULL);
+		bp = xfs_buf_incore(mp->m_ddev_targp, dqp->q_blkno,
+				mp->m_quotainfo->qi_dqchunklen, 0);
 		if (!bp) {
 			error = -EINVAL;
 			goto out_unlock;
@@ -1307,7 +1307,7 @@ xfs_qm_quotacheck(
 	 * We don't log our changes till later.
 	 */
 	if (uip) {
-		error = xfs_qm_dqiterate(mp, uip, XFS_QMOPT_UQUOTA,
+		error = xfs_qm_reset_dqcounts_buf(mp, uip, XFS_QMOPT_UQUOTA,
 					 &buffer_list);
 		if (error)
 			goto error_return;
@@ -1315,7 +1315,7 @@ xfs_qm_quotacheck(
 	}
 
 	if (gip) {
-		error = xfs_qm_dqiterate(mp, gip, XFS_QMOPT_GQUOTA,
+		error = xfs_qm_reset_dqcounts_buf(mp, gip, XFS_QMOPT_GQUOTA,
 					 &buffer_list);
 		if (error)
 			goto error_return;
@@ -1323,7 +1323,7 @@ xfs_qm_quotacheck(
 	}
 
 	if (pip) {
-		error = xfs_qm_dqiterate(mp, pip, XFS_QMOPT_PQUOTA,
+		error = xfs_qm_reset_dqcounts_buf(mp, pip, XFS_QMOPT_PQUOTA,
 					 &buffer_list);
 		if (error)
 			goto error_return;
@@ -1675,7 +1675,7 @@ xfs_qm_vop_dqalloc(
 	 * if necessary. The dquot(s) will not be locked.
 	 */
 	if (XFS_NOT_DQATTACHED(mp, ip)) {
-		error = xfs_qm_dqattach_locked(ip, XFS_QMOPT_DQALLOC);
+		error = xfs_qm_dqattach_locked(ip, true);
 		if (error) {
 			xfs_iunlock(ip, lockflags);
 			return error;
@@ -1694,10 +1694,7 @@ xfs_qm_vop_dqalloc(
 			 * holding ilock.
 			 */
 			xfs_iunlock(ip, lockflags);
-			error = xfs_qm_dqget(mp, NULL, uid,
-						 XFS_DQ_USER,
-						 XFS_QMOPT_DQALLOC,
-						 &uq);
+			error = xfs_qm_dqget(mp, uid, XFS_DQ_USER, true, &uq);
 			if (error) {
 				ASSERT(error != -ENOENT);
 				return error;
@@ -1720,10 +1717,7 @@ xfs_qm_vop_dqalloc(
 	if ((flags & XFS_QMOPT_GQUOTA) && XFS_IS_GQUOTA_ON(mp)) {
 		if (ip->i_d.di_gid != gid) {
 			xfs_iunlock(ip, lockflags);
-			error = xfs_qm_dqget(mp, NULL, gid,
-						 XFS_DQ_GROUP,
-						 XFS_QMOPT_DQALLOC,
-						 &gq);
+			error = xfs_qm_dqget(mp, gid, XFS_DQ_GROUP, true, &gq);
 			if (error) {
 				ASSERT(error != -ENOENT);
 				goto error_rele;
@@ -1739,10 +1733,8 @@ xfs_qm_vop_dqalloc(
 	if ((flags & XFS_QMOPT_PQUOTA) && XFS_IS_PQUOTA_ON(mp)) {
 		if (xfs_get_projid(ip) != prid) {
 			xfs_iunlock(ip, lockflags);
-			error = xfs_qm_dqget(mp, NULL, (xfs_dqid_t)prid,
-						 XFS_DQ_PROJ,
-						 XFS_QMOPT_DQALLOC,
-						 &pq);
+			error = xfs_qm_dqget(mp, (xfs_dqid_t)prid, XFS_DQ_PROJ,
+					true, &pq);
 			if (error) {
 				ASSERT(error != -ENOENT);
 				goto error_rele;
@@ -1933,7 +1925,7 @@ xfs_qm_vop_rename_dqattach(
 		 */
 		if (i == 0 || ip != i_tab[i-1]) {
 			if (XFS_NOT_DQATTACHED(mp, ip)) {
-				error = xfs_qm_dqattach(ip, 0);
+				error = xfs_qm_dqattach(ip);
 				if (error)
 					return error;
 			}
diff --git a/fs/xfs/xfs_qm.h b/fs/xfs/xfs_qm.h
index 2975a822e9f0..e3129b280423 100644
--- a/fs/xfs/xfs_qm.h
+++ b/fs/xfs/xfs_qm.h
@@ -170,8 +170,10 @@ extern void		xfs_qm_dqrele_all_inodes(struct xfs_mount *, uint);
 
 /* quota ops */
 extern int		xfs_qm_scall_trunc_qfiles(struct xfs_mount *, uint);
-extern int		xfs_qm_scall_getquota(struct xfs_mount *, xfs_dqid_t *,
-					uint, struct qc_dqblk *, uint);
+extern int		xfs_qm_scall_getquota(struct xfs_mount *, xfs_dqid_t,
+					uint, struct qc_dqblk *);
+extern int		xfs_qm_scall_getquota_next(struct xfs_mount *,
+					xfs_dqid_t *, uint, struct qc_dqblk *);
 extern int		xfs_qm_scall_setqlim(struct xfs_mount *, xfs_dqid_t, uint,
 					struct qc_dqblk *);
 extern int		xfs_qm_scall_quotaon(struct xfs_mount *, uint);
diff --git a/fs/xfs/xfs_qm_bhv.c b/fs/xfs/xfs_qm_bhv.c
index 2be6d2735ca9..36b89e2c5eb9 100644
--- a/fs/xfs/xfs_qm_bhv.c
+++ b/fs/xfs/xfs_qm_bhv.c
@@ -72,7 +72,7 @@ xfs_qm_statvfs(
 	xfs_mount_t		*mp = ip->i_mount;
 	xfs_dquot_t		*dqp;
 
-	if (!xfs_qm_dqget(mp, NULL, xfs_get_projid(ip), XFS_DQ_PROJ, 0, &dqp)) {
+	if (!xfs_qm_dqget(mp, xfs_get_projid(ip), XFS_DQ_PROJ, false, &dqp)) {
 		xfs_fill_statvfs_from_dquot(statp, dqp);
 		xfs_qm_dqput(dqp);
 	}
diff --git a/fs/xfs/xfs_qm_syscalls.c b/fs/xfs/xfs_qm_syscalls.c
index 9cb5c381b01c..3e05d300b14e 100644
--- a/fs/xfs/xfs_qm_syscalls.c
+++ b/fs/xfs/xfs_qm_syscalls.c
@@ -425,7 +425,7 @@ xfs_qm_scall_setqlim(
 	 * a reference to the dquot, so it's safe to do this unlock/lock without
 	 * it being reclaimed in the mean time.
 	 */
-	error = xfs_qm_dqget(mp, NULL, id, type, XFS_QMOPT_DQALLOC, &dqp);
+	error = xfs_qm_dqget(mp, id, type, true, &dqp);
 	if (error) {
 		ASSERT(error != -ENOENT);
 		goto out_unlock;
@@ -622,39 +622,14 @@ out:
 	return error;
 }
 
-
-int
-xfs_qm_scall_getquota(
+/* Fill out the quota context. */
+static void
+xfs_qm_scall_getquota_fill_qc(
 	struct xfs_mount	*mp,
-	xfs_dqid_t		*id,
 	uint			type,
-	struct qc_dqblk		*dst,
-	uint			dqget_flags)
+	const struct xfs_dquot	*dqp,
+	struct qc_dqblk		*dst)
 {
-	struct xfs_dquot	*dqp;
-	int			error;
-
-	/*
-	 * Try to get the dquot. We don't want it allocated on disk, so
-	 * we aren't passing the XFS_QMOPT_DOALLOC flag. If it doesn't
-	 * exist, we'll get ENOENT back.
-	 */
-	error = xfs_qm_dqget(mp, NULL, *id, type, dqget_flags, &dqp);
-	if (error)
-		return error;
-
-	/*
-	 * If everything's NULL, this dquot doesn't quite exist as far as
-	 * our utility programs are concerned.
-	 */
-	if (XFS_IS_DQUOT_UNINITIALIZED(dqp)) {
-		error = -ENOENT;
-		goto out_put;
-	}
-
-	/* Fill in the ID we actually read from disk */
-	*id = be32_to_cpu(dqp->q_core.d_id);
-
 	memset(dst, 0, sizeof(*dst));
 	dst->d_spc_hardlimit =
 		XFS_FSB_TO_B(mp, be64_to_cpu(dqp->q_core.d_blk_hardlimit));
@@ -696,7 +671,7 @@ xfs_qm_scall_getquota(
 	if (((XFS_IS_UQUOTA_ENFORCED(mp) && type == XFS_DQ_USER) ||
 	     (XFS_IS_GQUOTA_ENFORCED(mp) && type == XFS_DQ_GROUP) ||
 	     (XFS_IS_PQUOTA_ENFORCED(mp) && type == XFS_DQ_PROJ)) &&
-	    *id != 0) {
+	    dqp->q_core.d_id != 0) {
 		if ((dst->d_space > dst->d_spc_softlimit) &&
 		    (dst->d_spc_softlimit > 0)) {
 			ASSERT(dst->d_spc_timer != 0);
@@ -707,11 +682,69 @@ xfs_qm_scall_getquota(
 		}
 	}
 #endif
+}
+
+/* Return the quota information for the dquot matching id. */
+int
+xfs_qm_scall_getquota(
+	struct xfs_mount	*mp,
+	xfs_dqid_t		id,
+	uint			type,
+	struct qc_dqblk		*dst)
+{
+	struct xfs_dquot	*dqp;
+	int			error;
+
+	/*
+	 * Try to get the dquot. We don't want it allocated on disk, so don't
+	 * set doalloc. If it doesn't exist, we'll get ENOENT back.
+	 */
+	error = xfs_qm_dqget(mp, id, type, false, &dqp);
+	if (error)
+		return error;
+
+	/*
+	 * If everything's NULL, this dquot doesn't quite exist as far as
+	 * our utility programs are concerned.
+	 */
+	if (XFS_IS_DQUOT_UNINITIALIZED(dqp)) {
+		error = -ENOENT;
+		goto out_put;
+	}
+
+	xfs_qm_scall_getquota_fill_qc(mp, type, dqp, dst);
+
 out_put:
 	xfs_qm_dqput(dqp);
 	return error;
 }
 
+/*
+ * Return the quota information for the first initialized dquot whose id
+ * is at least as high as id.
+ */
+int
+xfs_qm_scall_getquota_next(
+	struct xfs_mount	*mp,
+	xfs_dqid_t		*id,
+	uint			type,
+	struct qc_dqblk		*dst)
+{
+	struct xfs_dquot	*dqp;
+	int			error;
+
+	error = xfs_qm_dqget_next(mp, *id, type, &dqp);
+	if (error)
+		return error;
+
+	/* Fill in the ID we actually read from disk */
+	*id = be32_to_cpu(dqp->q_core.d_id);
+
+	xfs_qm_scall_getquota_fill_qc(mp, type, dqp, dst);
+
+	xfs_qm_dqput(dqp);
+	return error;
+}
 
 STATIC int
 xfs_dqrele_inode(
diff --git a/fs/xfs/xfs_quota.h b/fs/xfs/xfs_quota.h
index ce6506adab7b..3edf52b14919 100644
--- a/fs/xfs/xfs_quota.h
+++ b/fs/xfs/xfs_quota.h
@@ -48,6 +48,22 @@ struct xfs_trans;
 	 (XFS_IS_PQUOTA_ON(mp) && \
 		(mp->m_sb.sb_qflags & XFS_PQUOTA_CHKD) == 0))
 
+static inline uint
+xfs_quota_chkd_flag(
+	uint		dqtype)
+{
+	switch (dqtype) {
+	case XFS_DQ_USER:
+		return XFS_UQUOTA_CHKD;
+	case XFS_DQ_GROUP:
+		return XFS_GQUOTA_CHKD;
+	case XFS_DQ_PROJ:
+		return XFS_PQUOTA_CHKD;
+	default:
+		return 0;
+	}
+}
+
 /*
  * The structure kept inside the xfs_trans_t keep track of dquot changes
  * within a transaction and apply them later.
@@ -90,8 +106,8 @@ extern struct xfs_dquot *xfs_qm_vop_chown(struct xfs_trans *,
 extern int xfs_qm_vop_chown_reserve(struct xfs_trans *, struct xfs_inode *,
 		struct xfs_dquot *, struct xfs_dquot *,
 		struct xfs_dquot *, uint);
-extern int xfs_qm_dqattach(struct xfs_inode *, uint);
-extern int xfs_qm_dqattach_locked(struct xfs_inode *, uint);
+extern int xfs_qm_dqattach(struct xfs_inode *);
+extern int xfs_qm_dqattach_locked(struct xfs_inode *ip, bool doalloc);
 extern void xfs_qm_dqdetach(struct xfs_inode *);
 extern void xfs_qm_dqrele(struct xfs_dquot *);
 extern void xfs_qm_statvfs(struct xfs_inode *, struct kstatfs *);
@@ -132,7 +148,7 @@ static inline int xfs_trans_reserve_quota_bydquots(struct xfs_trans *tp,
 #define xfs_qm_vop_rename_dqattach(it)					(0)
 #define xfs_qm_vop_chown(tp, ip, old, new)				(NULL)
 #define xfs_qm_vop_chown_reserve(tp, ip, u, g, p, fl)			(0)
-#define xfs_qm_dqattach(ip, fl)						(0)
+#define xfs_qm_dqattach(ip)						(0)
 #define xfs_qm_dqattach_locked(ip, fl)					(0)
 #define xfs_qm_dqdetach(ip)
 #define xfs_qm_dqrele(d)
diff --git a/fs/xfs/xfs_quotaops.c b/fs/xfs/xfs_quotaops.c
index a65108594a07..c93fc913dffb 100644
--- a/fs/xfs/xfs_quotaops.c
+++ b/fs/xfs/xfs_quotaops.c
@@ -239,8 +239,7 @@ xfs_fs_get_dqblk(
 		return -ESRCH;
 
 	id = from_kqid(&init_user_ns, qid);
-	return xfs_qm_scall_getquota(mp, &id,
-				      xfs_quota_type(qid.type), qdq, 0);
+	return xfs_qm_scall_getquota(mp, id, xfs_quota_type(qid.type), qdq);
 }
 
 /* Return quota info for active quota >= this qid */
@@ -260,9 +259,8 @@ xfs_fs_get_nextdqblk(
 		return -ESRCH;
 
 	id = from_kqid(&init_user_ns, *qid);
-	ret = xfs_qm_scall_getquota(mp, &id,
-				    xfs_quota_type(qid->type), qdq,
-				    XFS_QMOPT_DQNEXT);
+	ret = xfs_qm_scall_getquota_next(mp, &id, xfs_quota_type(qid->type),
+			qdq);
 	if (ret)
 		return ret;
 
diff --git a/fs/xfs/xfs_refcount_item.c b/fs/xfs/xfs_refcount_item.c
index 15c9393dd7a7..e5866b714d5f 100644
--- a/fs/xfs/xfs_refcount_item.c
+++ b/fs/xfs/xfs_refcount_item.c
@@ -159,7 +159,7 @@ STATIC void
 xfs_cui_item_unlock(
 	struct xfs_log_item	*lip)
 {
-	if (lip->li_flags & XFS_LI_ABORTED)
+	if (test_bit(XFS_LI_ABORTED, &lip->li_flags))
 		xfs_cui_release(CUI_ITEM(lip));
 }
 
@@ -310,7 +310,7 @@ xfs_cud_item_unlock(
 {
 	struct xfs_cud_log_item	*cudp = CUD_ITEM(lip);
 
-	if (lip->li_flags & XFS_LI_ABORTED) {
+	if (test_bit(XFS_LI_ABORTED, &lip->li_flags)) {
 		xfs_cui_release(cudp->cud_cuip);
 		kmem_zone_free(xfs_cud_zone, cudp);
 	}
diff --git a/fs/xfs/xfs_reflink.c b/fs/xfs/xfs_reflink.c
index cdbd342a5249..713e857d9ffa 100644
--- a/fs/xfs/xfs_reflink.c
+++ b/fs/xfs/xfs_reflink.c
@@ -305,7 +305,7 @@ xfs_reflink_reserve_cow(
 	 * Fork all the shared blocks from our write offset until the end of
 	 * the extent.
 	 */
-	error = xfs_qm_dqattach_locked(ip, 0);
+	error = xfs_qm_dqattach_locked(ip, false);
 	if (error)
 		return error;
 
@@ -431,7 +431,7 @@ retry:
 		if (error)
 			return error;
 
-		error = xfs_qm_dqattach_locked(ip, 0);
+		error = xfs_qm_dqattach_locked(ip, false);
 		if (error)
 			goto out;
 		goto retry;
@@ -552,6 +552,9 @@ xfs_reflink_trim_irec_to_next_cow(
  *
  * If cancel_real is true this function cancels all COW fork extents for the
  * inode; if cancel_real is false, real extents are not cleared.
+ *
+ * Caller must have already joined the inode to the current transaction. The
+ * inode will be joined to the transaction returned to the caller.
  */
 int
 xfs_reflink_cancel_cow_blocks(
@@ -592,7 +595,6 @@ xfs_reflink_cancel_cow_blocks(
 			if (error)
 				break;
 		} else if (del.br_state == XFS_EXT_UNWRITTEN || cancel_real) {
-			xfs_trans_ijoin(*tpp, ip, 0);
 			xfs_defer_init(&dfops, &firstfsb);
 
 			/* Free the CoW orphan record. */
@@ -1359,7 +1361,7 @@ xfs_reflink_remap_range(
 		goto out_unlock;
 
 	/* Attach dquots to dest inode before changing block map */
-	ret = xfs_qm_dqattach(dest, 0);
+	ret = xfs_qm_dqattach(dest);
 	if (ret)
 		goto out_unlock;
 
@@ -1551,7 +1553,12 @@ next:
 	return 0;
 }
 
-/* Clear the inode reflink flag if there are no shared extents. */
+/*
+ * Clear the inode reflink flag if there are no shared extents.
+ *
+ * The caller is responsible for joining the inode to the transaction passed in.
+ * The inode will be joined to the transaction that is returned to the caller.
+ */
 int
 xfs_reflink_clear_inode_flag(
 	struct xfs_inode	*ip,
@@ -1578,7 +1585,6 @@ xfs_reflink_clear_inode_flag(
 	trace_xfs_reflink_unset_inode_flag(ip);
 	ip->i_d.di_flags2 &= ~XFS_DIFLAG2_REFLINK;
 	xfs_inode_clear_cowblocks_tag(ip);
-	xfs_trans_ijoin(*tpp, ip, 0);
 	xfs_trans_log_inode(*tpp, ip, XFS_ILOG_CORE);
 
 	return error;
diff --git a/fs/xfs/xfs_rmap_item.c b/fs/xfs/xfs_rmap_item.c
index 06a07846c9b3..e5b5b3e7ef82 100644
--- a/fs/xfs/xfs_rmap_item.c
+++ b/fs/xfs/xfs_rmap_item.c
@@ -158,7 +158,7 @@ STATIC void
 xfs_rui_item_unlock(
 	struct xfs_log_item	*lip)
 {
-	if (lip->li_flags & XFS_LI_ABORTED)
+	if (test_bit(XFS_LI_ABORTED, &lip->li_flags))
 		xfs_rui_release(RUI_ITEM(lip));
 }
 
@@ -331,7 +331,7 @@ xfs_rud_item_unlock(
 {
 	struct xfs_rud_log_item	*rudp = RUD_ITEM(lip);
 
-	if (lip->li_flags & XFS_LI_ABORTED) {
+	if (test_bit(XFS_LI_ABORTED, &lip->li_flags)) {
 		xfs_rui_release(rudp->rud_ruip);
 		kmem_zone_free(xfs_rud_zone, rudp);
 	}
diff --git a/fs/xfs/xfs_rtalloc.h b/fs/xfs/xfs_rtalloc.h
index dfee3c991155..52632ab727f7 100644
--- a/fs/xfs/xfs_rtalloc.h
+++ b/fs/xfs/xfs_rtalloc.h
@@ -23,9 +23,14 @@
 struct xfs_mount;
 struct xfs_trans;
 
+/*
+ * XXX: Most of the realtime allocation functions deal in units of realtime
+ * extents, not realtime blocks.  This looks funny when paired with the type
+ * name and screams for a larger cleanup.
+ */
 struct xfs_rtalloc_rec {
-	xfs_rtblock_t		ar_startblock;
-	xfs_rtblock_t		ar_blockcount;
+	xfs_rtblock_t		ar_startext;
+	xfs_rtblock_t		ar_extcount;
 };
 
 typedef int (*xfs_rtalloc_query_range_fn)(
diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c
index f643d76db516..ed67389f4948 100644
--- a/fs/xfs/xfs_super.c
+++ b/fs/xfs/xfs_super.c
@@ -1372,7 +1372,6 @@ xfs_fs_remount(
 		 */
 		xfs_restore_resvblks(mp);
 		xfs_log_work_queue(mp);
-		xfs_queue_eofblocks(mp);
 
 		/* Recover any CoW blocks that never got remapped. */
 		error = xfs_reflink_recover_cow(mp);
@@ -1382,7 +1381,7 @@ xfs_fs_remount(
 			xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
 			return error;
 		}
-		xfs_queue_cowblocks(mp);
+		xfs_icache_enable_reclaim(mp);
 
 		/* Create the per-AG metadata reservation pool .*/
 		error = xfs_fs_reserve_ag_blocks(mp);
@@ -1392,8 +1391,13 @@ xfs_fs_remount(
 
 	/* rw -> ro */
 	if (!(mp->m_flags & XFS_MOUNT_RDONLY) && (*flags & SB_RDONLY)) {
+		/*
+		 * Cancel background eofb scanning so it cannot race with the
+		 * final log force+buftarg wait and deadlock the remount.
+		 */
+		xfs_icache_disable_reclaim(mp);
+
 		/* Get rid of any leftover CoW reservations... */
-		cancel_delayed_work_sync(&mp->m_cowblocks_work);
 		error = xfs_icache_free_cowblocks(mp, NULL);
 		if (error) {
 			xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
@@ -1416,12 +1420,6 @@ xfs_fs_remount(
 		 */
 		xfs_save_resvblks(mp);
 
-		/*
-		 * Cancel background eofb scanning so it cannot race with the
-		 * final log force+buftarg wait and deadlock the remount.
-		 */
-		cancel_delayed_work_sync(&mp->m_eofblocks_work);
-
 		xfs_quiesce_attr(mp);
 		mp->m_flags |= XFS_MOUNT_RDONLY;
 	}
@@ -1441,6 +1439,7 @@ xfs_fs_freeze(
 {
 	struct xfs_mount	*mp = XFS_M(sb);
 
+	xfs_icache_disable_reclaim(mp);
 	xfs_save_resvblks(mp);
 	xfs_quiesce_attr(mp);
 	return xfs_sync_sb(mp, true);
@@ -1454,6 +1453,7 @@ xfs_fs_unfreeze(
 
 	xfs_restore_resvblks(mp);
 	xfs_log_work_queue(mp);
+	xfs_icache_enable_reclaim(mp);
 	return 0;
 }
 
@@ -1635,6 +1635,17 @@ xfs_fs_fill_super(
 #endif
 	sb->s_op = &xfs_super_operations;
 
+	/*
+	 * Delay mount work if the debug hook is set. This is debug
+	 * instrumention to coordinate simulation of xfs mount failures with
+	 * VFS superblock operations
+	 */
+	if (xfs_globals.mount_delay) {
+		xfs_notice(mp, "Delaying mount for %d seconds.",
+			xfs_globals.mount_delay);
+		msleep(xfs_globals.mount_delay * 1000);
+	}
+
 	if (silent)
 		flags |= XFS_MFSI_QUIET;
 
@@ -1690,11 +1701,17 @@ xfs_fs_fill_super(
 		sb->s_flags |= SB_I_VERSION;
 
 	if (mp->m_flags & XFS_MOUNT_DAX) {
+		bool rtdev_is_dax = false, datadev_is_dax;
+
 		xfs_warn(mp,
 		"DAX enabled. Warning: EXPERIMENTAL, use at your own risk");
 
-		error = bdev_dax_supported(sb, sb->s_blocksize);
-		if (error) {
+		datadev_is_dax = bdev_dax_supported(mp->m_ddev_targp->bt_bdev,
+			sb->s_blocksize);
+		if (mp->m_rtdev_targp)
+			rtdev_is_dax = bdev_dax_supported(
+				mp->m_rtdev_targp->bt_bdev, sb->s_blocksize);
+		if (!rtdev_is_dax && !datadev_is_dax) {
 			xfs_alert(mp,
 			"DAX unsupported by block device. Turning off DAX.");
 			mp->m_flags &= ~XFS_MOUNT_DAX;
@@ -1761,6 +1778,7 @@ xfs_fs_fill_super(
  out_close_devices:
 	xfs_close_devices(mp);
  out_free_fsname:
+	sb->s_fs_info = NULL;
 	xfs_free_fsname(mp);
 	kfree(mp);
  out:
@@ -1778,6 +1796,10 @@ xfs_fs_put_super(
 {
 	struct xfs_mount	*mp = XFS_M(sb);
 
+	/* if ->fill_super failed, we have no mount to tear down */
+	if (!sb->s_fs_info)
+		return;
+
 	xfs_notice(mp, "Unmounting Filesystem");
 	xfs_filestream_unmount(mp);
 	xfs_unmountfs(mp);
@@ -1787,6 +1809,8 @@ xfs_fs_put_super(
 	xfs_destroy_percpu_counters(mp);
 	xfs_destroy_mount_workqueues(mp);
 	xfs_close_devices(mp);
+
+	sb->s_fs_info = NULL;
 	xfs_free_fsname(mp);
 	kfree(mp);
 }
@@ -1806,6 +1830,9 @@ xfs_fs_nr_cached_objects(
 	struct super_block	*sb,
 	struct shrink_control	*sc)
 {
+	/* Paranoia: catch incorrect calls during mount setup or teardown */
+	if (WARN_ON_ONCE(!sb->s_fs_info))
+		return 0;
 	return xfs_reclaim_inodes_count(XFS_M(sb));
 }
 
@@ -1879,11 +1906,6 @@ xfs_init_zones(void)
 	if (!xfs_trans_zone)
 		goto out_destroy_ifork_zone;
 
-	xfs_log_item_desc_zone =
-		kmem_zone_init(sizeof(struct xfs_log_item_desc),
-			       "xfs_log_item_desc");
-	if (!xfs_log_item_desc_zone)
-		goto out_destroy_trans_zone;
 
 	/*
 	 * The size of the zone allocated buf log item is the maximum
@@ -1893,7 +1915,7 @@ xfs_init_zones(void)
 	xfs_buf_item_zone = kmem_zone_init(sizeof(struct xfs_buf_log_item),
 					   "xfs_buf_item");
 	if (!xfs_buf_item_zone)
-		goto out_destroy_log_item_desc_zone;
+		goto out_destroy_trans_zone;
 
 	xfs_efd_zone = kmem_zone_init((sizeof(xfs_efd_log_item_t) +
 			((XFS_EFD_MAX_FAST_EXTENTS - 1) *
@@ -1981,8 +2003,6 @@ xfs_init_zones(void)
 	kmem_zone_destroy(xfs_efd_zone);
  out_destroy_buf_item_zone:
 	kmem_zone_destroy(xfs_buf_item_zone);
- out_destroy_log_item_desc_zone:
-	kmem_zone_destroy(xfs_log_item_desc_zone);
  out_destroy_trans_zone:
 	kmem_zone_destroy(xfs_trans_zone);
  out_destroy_ifork_zone:
@@ -2021,7 +2041,6 @@ xfs_destroy_zones(void)
 	kmem_zone_destroy(xfs_efi_zone);
 	kmem_zone_destroy(xfs_efd_zone);
 	kmem_zone_destroy(xfs_buf_item_zone);
-	kmem_zone_destroy(xfs_log_item_desc_zone);
 	kmem_zone_destroy(xfs_trans_zone);
 	kmem_zone_destroy(xfs_ifork_zone);
 	kmem_zone_destroy(xfs_da_state_zone);
diff --git a/fs/xfs/xfs_symlink.c b/fs/xfs/xfs_symlink.c
index 5b66ac12913c..aed03da637d4 100644
--- a/fs/xfs/xfs_symlink.c
+++ b/fs/xfs/xfs_symlink.c
@@ -259,6 +259,7 @@ xfs_symlink(
 	 * bmapi or the directory create code.
 	 */
 	xfs_defer_init(&dfops, &first_block);
+	tp->t_agfl_dfops = &dfops;
 
 	/*
 	 * Allocate an inode for the symlink.
@@ -488,16 +489,11 @@ xfs_inactive_symlink_rmt(
 	error = xfs_defer_finish(&tp, &dfops);
 	if (error)
 		goto error_bmap_cancel;
-	/*
-	 * The first xact was committed, so add the inode to the new one.
-	 * Mark it dirty so it will be logged and moved forward in the log as
-	 * part of every commit.
-	 */
-	xfs_trans_ijoin(tp, ip, 0);
-	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+
 	/*
 	 * Commit the transaction containing extent freeing and EFDs.
 	 */
+	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
 	error = xfs_trans_commit(tp);
 	if (error) {
 		ASSERT(XFS_FORCED_SHUTDOWN(mp));
diff --git a/fs/xfs/xfs_sysctl.h b/fs/xfs/xfs_sysctl.h
index 82afee005140..b53a33e69932 100644
--- a/fs/xfs/xfs_sysctl.h
+++ b/fs/xfs/xfs_sysctl.h
@@ -95,6 +95,7 @@ extern xfs_param_t	xfs_params;
 
 struct xfs_globals {
 	int	log_recovery_delay;	/* log recovery delay (secs) */
+	int	mount_delay;		/* mount setup delay (secs) */
 	bool	bug_on_assert;		/* BUG() the kernel on assert failure */
 };
 extern struct xfs_globals	xfs_globals;
diff --git a/fs/xfs/xfs_sysfs.c b/fs/xfs/xfs_sysfs.c
index 8b2ccc234f36..2d5cd2529f8e 100644
--- a/fs/xfs/xfs_sysfs.c
+++ b/fs/xfs/xfs_sysfs.c
@@ -165,9 +165,40 @@ log_recovery_delay_show(
 }
 XFS_SYSFS_ATTR_RW(log_recovery_delay);
 
+STATIC ssize_t
+mount_delay_store(
+	struct kobject	*kobject,
+	const char	*buf,
+	size_t		count)
+{
+	int		ret;
+	int		val;
+
+	ret = kstrtoint(buf, 0, &val);
+	if (ret)
+		return ret;
+
+	if (val < 0 || val > 60)
+		return -EINVAL;
+
+	xfs_globals.mount_delay = val;
+
+	return count;
+}
+
+STATIC ssize_t
+mount_delay_show(
+	struct kobject	*kobject,
+	char		*buf)
+{
+	return snprintf(buf, PAGE_SIZE, "%d\n", xfs_globals.mount_delay);
+}
+XFS_SYSFS_ATTR_RW(mount_delay);
+
 static struct attribute *xfs_dbg_attrs[] = {
 	ATTR_LIST(bug_on_assert),
 	ATTR_LIST(log_recovery_delay),
+	ATTR_LIST(mount_delay),
 	NULL,
 };
 
diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h
index 8955254b900e..9d4c4ca24fe6 100644
--- a/fs/xfs/xfs_trace.h
+++ b/fs/xfs/xfs_trace.h
@@ -441,8 +441,7 @@ DECLARE_EVENT_CLASS(xfs_buf_item_class,
 		__field(unsigned, bli_recur)
 		__field(int, bli_refcount)
 		__field(unsigned, bli_flags)
-		__field(void *, li_desc)
-		__field(unsigned, li_flags)
+		__field(unsigned long, li_flags)
 	),
 	TP_fast_assign(
 		__entry->dev = bip->bli_buf->b_target->bt_dev;
@@ -455,12 +454,11 @@ DECLARE_EVENT_CLASS(xfs_buf_item_class,
 		__entry->buf_hold = atomic_read(&bip->bli_buf->b_hold);
 		__entry->buf_pincount = atomic_read(&bip->bli_buf->b_pin_count);
 		__entry->buf_lockval = bip->bli_buf->b_sema.count;
-		__entry->li_desc = bip->bli_item.li_desc;
 		__entry->li_flags = bip->bli_item.li_flags;
 	),
 	TP_printk("dev %d:%d bno 0x%llx len 0x%zx hold %d pincount %d "
 		  "lock %d flags %s recur %d refcount %d bliflags %s "
-		  "lidesc %p liflags %s",
+		  "liflags %s",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		  (unsigned long long)__entry->buf_bno,
 		  __entry->buf_len,
@@ -471,7 +469,6 @@ DECLARE_EVENT_CLASS(xfs_buf_item_class,
 		  __entry->bli_recur,
 		  __entry->bli_refcount,
 		  __print_flags(__entry->bli_flags, "|", XFS_BLI_FLAGS),
-		  __entry->li_desc,
 		  __print_flags(__entry->li_flags, "|", XFS_LI_FLAGS))
 )
 
@@ -1018,7 +1015,7 @@ DECLARE_EVENT_CLASS(xfs_log_item_class,
 		__field(dev_t, dev)
 		__field(void *, lip)
 		__field(uint, type)
-		__field(uint, flags)
+		__field(unsigned long, flags)
 		__field(xfs_lsn_t, lsn)
 	),
 	TP_fast_assign(
@@ -1070,7 +1067,7 @@ DECLARE_EVENT_CLASS(xfs_ail_class,
 		__field(dev_t, dev)
 		__field(void *, lip)
 		__field(uint, type)
-		__field(uint, flags)
+		__field(unsigned long, flags)
 		__field(xfs_lsn_t, old_lsn)
 		__field(xfs_lsn_t, new_lsn)
 	),
@@ -1750,6 +1747,7 @@ DECLARE_EVENT_CLASS(xfs_attr_class,
 		__field(int, namelen)
 		__field(int, valuelen)
 		__field(xfs_dahash_t, hashval)
+		__field(int, flags)
 		__field(int, op_flags)
 	),
 	TP_fast_assign(
@@ -1760,10 +1758,11 @@ DECLARE_EVENT_CLASS(xfs_attr_class,
 		__entry->namelen = args->namelen;
 		__entry->valuelen = args->valuelen;
 		__entry->hashval = args->hashval;
+		__entry->flags = args->flags;
 		__entry->op_flags = args->op_flags;
 	),
 	TP_printk("dev %d:%d ino 0x%llx name %.*s namelen %d valuelen %d "
-		  "hashval 0x%x op_flags %s",
+		  "hashval 0x%x flags %s op_flags %s",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		  __entry->ino,
 		  __entry->namelen,
@@ -1771,6 +1770,7 @@ DECLARE_EVENT_CLASS(xfs_attr_class,
 		  __entry->namelen,
 		  __entry->valuelen,
 		  __entry->hashval,
+		  __print_flags(__entry->flags, "|", XFS_ATTR_FLAGS),
 		  __print_flags(__entry->op_flags, "|", XFS_DA_OP_FLAGS))
 )
 
@@ -2243,30 +2243,35 @@ struct xfs_defer_pending;
 struct xfs_defer_ops;
 
 DECLARE_EVENT_CLASS(xfs_defer_class,
-	TP_PROTO(struct xfs_mount *mp, struct xfs_defer_ops *dop),
-	TP_ARGS(mp, dop),
+	TP_PROTO(struct xfs_mount *mp, struct xfs_defer_ops *dop,
+		 unsigned long caller_ip),
+	TP_ARGS(mp, dop, caller_ip),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
 		__field(void *, dop)
 		__field(char, committed)
 		__field(char, low)
+		__field(unsigned long, caller_ip)
 	),
 	TP_fast_assign(
 		__entry->dev = mp ? mp->m_super->s_dev : 0;
 		__entry->dop = dop;
 		__entry->committed = dop->dop_committed;
 		__entry->low = dop->dop_low;
+		__entry->caller_ip = caller_ip;
 	),
-	TP_printk("dev %d:%d ops %p committed %d low %d",
+	TP_printk("dev %d:%d ops %p committed %d low %d, caller %pS",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		  __entry->dop,
 		  __entry->committed,
-		  __entry->low)
+		  __entry->low,
+		  (char *)__entry->caller_ip)
 )
 #define DEFINE_DEFER_EVENT(name) \
 DEFINE_EVENT(xfs_defer_class, name, \
-	TP_PROTO(struct xfs_mount *mp, struct xfs_defer_ops *dop), \
-	TP_ARGS(mp, dop))
+	TP_PROTO(struct xfs_mount *mp, struct xfs_defer_ops *dop, \
+		 unsigned long caller_ip), \
+	TP_ARGS(mp, dop, caller_ip))
 
 DECLARE_EVENT_CLASS(xfs_defer_error_class,
 	TP_PROTO(struct xfs_mount *mp, struct xfs_defer_ops *dop, int error),
@@ -2433,6 +2438,8 @@ DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_abort);
 #define DEFINE_BMAP_FREE_DEFERRED_EVENT DEFINE_PHYS_EXTENT_DEFERRED_EVENT
 DEFINE_BMAP_FREE_DEFERRED_EVENT(xfs_bmap_free_defer);
 DEFINE_BMAP_FREE_DEFERRED_EVENT(xfs_bmap_free_deferred);
+DEFINE_BMAP_FREE_DEFERRED_EVENT(xfs_agfl_free_defer);
+DEFINE_BMAP_FREE_DEFERRED_EVENT(xfs_agfl_free_deferred);
 
 /* rmap tracepoints */
 DECLARE_EVENT_CLASS(xfs_rmap_class,
@@ -3346,6 +3353,43 @@ TRACE_EVENT(xfs_trans_resv_calc,
 		  __entry->logflags)
 );
 
+DECLARE_EVENT_CLASS(xfs_trans_class,
+	TP_PROTO(struct xfs_trans *tp, unsigned long caller_ip),
+	TP_ARGS(tp, caller_ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(uint32_t, tid)
+		__field(uint32_t, flags)
+		__field(unsigned long, caller_ip)
+	),
+	TP_fast_assign(
+		__entry->dev = tp->t_mountp->m_super->s_dev;
+		__entry->tid = 0;
+		if (tp->t_ticket)
+			__entry->tid = tp->t_ticket->t_tid;
+		__entry->flags = tp->t_flags;
+		__entry->caller_ip = caller_ip;
+	),
+	TP_printk("dev %d:%d trans %x flags 0x%x caller %pS",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->tid,
+		  __entry->flags,
+		  (char *)__entry->caller_ip)
+)
+
+#define DEFINE_TRANS_EVENT(name) \
+DEFINE_EVENT(xfs_trans_class, name, \
+	TP_PROTO(struct xfs_trans *tp, unsigned long caller_ip), \
+	TP_ARGS(tp, caller_ip))
+DEFINE_TRANS_EVENT(xfs_trans_alloc);
+DEFINE_TRANS_EVENT(xfs_trans_cancel);
+DEFINE_TRANS_EVENT(xfs_trans_commit);
+DEFINE_TRANS_EVENT(xfs_trans_dup);
+DEFINE_TRANS_EVENT(xfs_trans_free);
+DEFINE_TRANS_EVENT(xfs_trans_roll);
+DEFINE_TRANS_EVENT(xfs_trans_add_item);
+DEFINE_TRANS_EVENT(xfs_trans_free_items);
+
 #endif /* _TRACE_XFS_H */
 
 #undef TRACE_INCLUDE_PATH
diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c
index d6d8f9d129a7..fc7ba75b8b69 100644
--- a/fs/xfs/xfs_trans.c
+++ b/fs/xfs/xfs_trans.c
@@ -31,9 +31,9 @@
 #include "xfs_log.h"
 #include "xfs_trace.h"
 #include "xfs_error.h"
+#include "xfs_defer.h"
 
 kmem_zone_t	*xfs_trans_zone;
-kmem_zone_t	*xfs_log_item_desc_zone;
 
 #if defined(CONFIG_TRACEPOINTS)
 static void
@@ -79,6 +79,7 @@ xfs_trans_free(
 	xfs_extent_busy_sort(&tp->t_busy);
 	xfs_extent_busy_clear(tp->t_mountp, &tp->t_busy, false);
 
+	trace_xfs_trans_free(tp, _RET_IP_);
 	atomic_dec(&tp->t_mountp->m_active_trans);
 	if (!(tp->t_flags & XFS_TRANS_NO_WRITECOUNT))
 		sb_end_intwrite(tp->t_mountp->m_super);
@@ -94,11 +95,13 @@ xfs_trans_free(
  * blocks.  Locks and log items, however, are no inherited.  They must
  * be added to the new transaction explicitly.
  */
-STATIC xfs_trans_t *
+STATIC struct xfs_trans *
 xfs_trans_dup(
-	xfs_trans_t	*tp)
+	struct xfs_trans	*tp)
 {
-	xfs_trans_t	*ntp;
+	struct xfs_trans	*ntp;
+
+	trace_xfs_trans_dup(tp, _RET_IP_);
 
 	ntp = kmem_zone_zalloc(xfs_trans_zone, KM_SLEEP);
 
@@ -127,6 +130,7 @@ xfs_trans_dup(
 	ntp->t_rtx_res = tp->t_rtx_res - tp->t_rtx_res_used;
 	tp->t_rtx_res = tp->t_rtx_res_used;
 	ntp->t_pflags = tp->t_pflags;
+	ntp->t_agfl_dfops = tp->t_agfl_dfops;
 
 	xfs_trans_dup_dqinfo(tp, ntp);
 
@@ -283,6 +287,8 @@ xfs_trans_alloc(
 		return error;
 	}
 
+	trace_xfs_trans_alloc(tp, _RET_IP_);
+
 	*tpp = tp;
 	return 0;
 }
@@ -727,73 +733,52 @@ out:
 	return;
 }
 
-/*
- * Add the given log item to the transaction's list of log items.
- *
- * The log item will now point to its new descriptor with its li_desc field.
- */
+/* Add the given log item to the transaction's list of log items. */
 void
 xfs_trans_add_item(
 	struct xfs_trans	*tp,
 	struct xfs_log_item	*lip)
 {
-	struct xfs_log_item_desc *lidp;
-
 	ASSERT(lip->li_mountp == tp->t_mountp);
 	ASSERT(lip->li_ailp == tp->t_mountp->m_ail);
+	ASSERT(list_empty(&lip->li_trans));
+	ASSERT(!test_bit(XFS_LI_DIRTY, &lip->li_flags));
 
-	lidp = kmem_zone_zalloc(xfs_log_item_desc_zone, KM_SLEEP | KM_NOFS);
-
-	lidp->lid_item = lip;
-	lidp->lid_flags = 0;
-	list_add_tail(&lidp->lid_trans, &tp->t_items);
-
-	lip->li_desc = lidp;
-}
-
-STATIC void
-xfs_trans_free_item_desc(
-	struct xfs_log_item_desc *lidp)
-{
-	list_del_init(&lidp->lid_trans);
-	kmem_zone_free(xfs_log_item_desc_zone, lidp);
+	list_add_tail(&lip->li_trans, &tp->t_items);
+	trace_xfs_trans_add_item(tp, _RET_IP_);
 }
 
 /*
- * Unlink and free the given descriptor.
+ * Unlink the log item from the transaction. the log item is no longer
+ * considered dirty in this transaction, as the linked transaction has
+ * finished, either by abort or commit completion.
  */
 void
 xfs_trans_del_item(
 	struct xfs_log_item	*lip)
 {
-	xfs_trans_free_item_desc(lip->li_desc);
-	lip->li_desc = NULL;
+	clear_bit(XFS_LI_DIRTY, &lip->li_flags);
+	list_del_init(&lip->li_trans);
 }
 
-/*
- * Unlock all of the items of a transaction and free all the descriptors
- * of that transaction.
- */
+/* Detach and unlock all of the items in a transaction */
 void
 xfs_trans_free_items(
 	struct xfs_trans	*tp,
 	xfs_lsn_t		commit_lsn,
 	bool			abort)
 {
-	struct xfs_log_item_desc *lidp, *next;
-
-	list_for_each_entry_safe(lidp, next, &tp->t_items, lid_trans) {
-		struct xfs_log_item	*lip = lidp->lid_item;
+	struct xfs_log_item	*lip, *next;
 
-		lip->li_desc = NULL;
+	trace_xfs_trans_free_items(tp, _RET_IP_);
 
+	list_for_each_entry_safe(lip, next, &tp->t_items, li_trans) {
+		xfs_trans_del_item(lip);
 		if (commit_lsn != NULLCOMMITLSN)
 			lip->li_ops->iop_committing(lip, commit_lsn);
 		if (abort)
-			lip->li_flags |= XFS_LI_ABORTED;
+			set_bit(XFS_LI_ABORTED, &lip->li_flags);
 		lip->li_ops->iop_unlock(lip);
-
-		xfs_trans_free_item_desc(lidp);
 	}
 }
 
@@ -861,7 +846,7 @@ xfs_trans_committed_bulk(
 		xfs_lsn_t		item_lsn;
 
 		if (aborted)
-			lip->li_flags |= XFS_LI_ABORTED;
+			set_bit(XFS_LI_ABORTED, &lip->li_flags);
 		item_lsn = lip->li_ops->iop_committed(lip, commit_lsn);
 
 		/* item_lsn of -1 means the item needs no further processing */
@@ -936,6 +921,11 @@ __xfs_trans_commit(
 	int			error = 0;
 	int			sync = tp->t_flags & XFS_TRANS_SYNC;
 
+	ASSERT(!tp->t_agfl_dfops ||
+	       !xfs_defer_has_unfinished_work(tp->t_agfl_dfops) || regrant);
+
+	trace_xfs_trans_commit(tp, _RET_IP_);
+
 	/*
 	 * If there is nothing to be logged by the transaction,
 	 * then unlock all of the items associated with the
@@ -991,6 +981,7 @@ out_unreserve:
 		commit_lsn = xfs_log_done(mp, tp->t_ticket, NULL, regrant);
 		if (commit_lsn == -1 && !error)
 			error = -EIO;
+		tp->t_ticket = NULL;
 	}
 	current_restore_flags_nested(&tp->t_pflags, PF_MEMALLOC_NOFS);
 	xfs_trans_free_items(tp, NULLCOMMITLSN, !!error);
@@ -1022,6 +1013,8 @@ xfs_trans_cancel(
 	struct xfs_mount	*mp = tp->t_mountp;
 	bool			dirty = (tp->t_flags & XFS_TRANS_DIRTY);
 
+	trace_xfs_trans_cancel(tp, _RET_IP_);
+
 	/*
 	 * See if the caller is relying on us to shut down the
 	 * filesystem.  This happens in paths where we detect
@@ -1033,17 +1026,19 @@ xfs_trans_cancel(
 	}
 #ifdef DEBUG
 	if (!dirty && !XFS_FORCED_SHUTDOWN(mp)) {
-		struct xfs_log_item_desc *lidp;
+		struct xfs_log_item *lip;
 
-		list_for_each_entry(lidp, &tp->t_items, lid_trans)
-			ASSERT(!(lidp->lid_item->li_type == XFS_LI_EFD));
+		list_for_each_entry(lip, &tp->t_items, li_trans)
+			ASSERT(!(lip->li_type == XFS_LI_EFD));
 	}
 #endif
 	xfs_trans_unreserve_and_mod_sb(tp);
 	xfs_trans_unreserve_and_mod_dquots(tp);
 
-	if (tp->t_ticket)
+	if (tp->t_ticket) {
 		xfs_log_done(mp, tp->t_ticket, NULL, false);
+		tp->t_ticket = NULL;
+	}
 
 	/* mark this thread as no longer being in a transaction */
 	current_restore_flags_nested(&tp->t_pflags, PF_MEMALLOC_NOFS);
@@ -1067,6 +1062,8 @@ xfs_trans_roll(
 	struct xfs_trans_res	tres;
 	int			error;
 
+	trace_xfs_trans_roll(trans, _RET_IP_);
+
 	/*
 	 * Copy the critical parameters from one trans to the next.
 	 */
diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h
index 9d542dfe0052..29706b8b3bd4 100644
--- a/fs/xfs/xfs_trans.h
+++ b/fs/xfs/xfs_trans.h
@@ -27,7 +27,6 @@ struct xfs_efi_log_item;
 struct xfs_inode;
 struct xfs_item_ops;
 struct xfs_log_iovec;
-struct xfs_log_item_desc;
 struct xfs_mount;
 struct xfs_trans;
 struct xfs_trans_res;
@@ -43,12 +42,12 @@ struct xfs_bud_log_item;
 
 typedef struct xfs_log_item {
 	struct list_head		li_ail;		/* AIL pointers */
+	struct list_head		li_trans;	/* transaction list */
 	xfs_lsn_t			li_lsn;		/* last on-disk lsn */
-	struct xfs_log_item_desc	*li_desc;	/* ptr to current desc*/
 	struct xfs_mount		*li_mountp;	/* ptr to fs mount */
 	struct xfs_ail			*li_ailp;	/* ptr to AIL */
 	uint				li_type;	/* item type */
-	uint				li_flags;	/* misc flags */
+	unsigned long			li_flags;	/* misc flags */
 	struct xfs_buf			*li_buf;	/* real buffer pointer */
 	struct list_head		li_bio_list;	/* buffer item list */
 	void				(*li_cb)(struct xfs_buf *,
@@ -64,14 +63,21 @@ typedef struct xfs_log_item {
 	xfs_lsn_t			li_seq;		/* CIL commit seq */
 } xfs_log_item_t;
 
-#define	XFS_LI_IN_AIL	0x1
-#define	XFS_LI_ABORTED	0x2
-#define	XFS_LI_FAILED	0x4
+/*
+ * li_flags use the (set/test/clear)_bit atomic interfaces because updates can
+ * race with each other and we don't want to have to use the AIL lock to
+ * serialise all updates.
+ */
+#define	XFS_LI_IN_AIL	0
+#define	XFS_LI_ABORTED	1
+#define	XFS_LI_FAILED	2
+#define	XFS_LI_DIRTY	3	/* log item dirty in transaction */
 
 #define XFS_LI_FLAGS \
-	{ XFS_LI_IN_AIL,	"IN_AIL" }, \
-	{ XFS_LI_ABORTED,	"ABORTED" }, \
-	{ XFS_LI_FAILED,	"FAILED" }
+	{ (1 << XFS_LI_IN_AIL),		"IN_AIL" }, \
+	{ (1 << XFS_LI_ABORTED),	"ABORTED" }, \
+	{ (1 << XFS_LI_FAILED),		"FAILED" }, \
+	{ (1 << XFS_LI_DIRTY),		"DIRTY" }
 
 struct xfs_item_ops {
 	void (*iop_size)(xfs_log_item_t *, int *, int *);
@@ -111,6 +117,7 @@ typedef struct xfs_trans {
 	struct xlog_ticket	*t_ticket;	/* log mgr ticket */
 	struct xfs_mount	*t_mountp;	/* ptr to fs mount struct */
 	struct xfs_dquot_acct   *t_dqinfo;	/* acctg info for dquots */
+	struct xfs_defer_ops	*t_agfl_dfops;	/* optional agfl fixup dfops */
 	unsigned int		t_flags;	/* misc flags */
 	int64_t			t_icount_delta;	/* superblock icount change */
 	int64_t			t_ifree_delta;	/* superblock ifree change */
@@ -228,7 +235,8 @@ struct xfs_efd_log_item	*xfs_trans_get_efd(struct xfs_trans *,
 				  uint);
 int		xfs_trans_free_extent(struct xfs_trans *,
 				      struct xfs_efd_log_item *, xfs_fsblock_t,
-				      xfs_extlen_t, struct xfs_owner_info *);
+				      xfs_extlen_t, struct xfs_owner_info *,
+				      bool);
 int		xfs_trans_commit(struct xfs_trans *);
 int		xfs_trans_roll(struct xfs_trans **);
 int		xfs_trans_roll_inode(struct xfs_trans **, struct xfs_inode *);
@@ -242,7 +250,6 @@ void		xfs_trans_buf_copy_type(struct xfs_buf *dst_bp,
 					struct xfs_buf *src_bp);
 
 extern kmem_zone_t	*xfs_trans_zone;
-extern kmem_zone_t	*xfs_log_item_desc_zone;
 
 /* rmap updates */
 enum xfs_rmap_intent_type;
diff --git a/fs/xfs/xfs_trans_ail.c b/fs/xfs/xfs_trans_ail.c
index d4a2445215e6..41e280ef1483 100644
--- a/fs/xfs/xfs_trans_ail.c
+++ b/fs/xfs/xfs_trans_ail.c
@@ -32,30 +32,51 @@
 #ifdef DEBUG
 /*
  * Check that the list is sorted as it should be.
+ *
+ * Called with the ail lock held, but we don't want to assert fail with it
+ * held otherwise we'll lock everything up and won't be able to debug the
+ * cause. Hence we sample and check the state under the AIL lock and return if
+ * everything is fine, otherwise we drop the lock and run the ASSERT checks.
+ * Asserts may not be fatal, so pick the lock back up and continue onwards.
  */
 STATIC void
 xfs_ail_check(
-	struct xfs_ail	*ailp,
-	xfs_log_item_t	*lip)
+	struct xfs_ail		*ailp,
+	struct xfs_log_item	*lip)
 {
-	xfs_log_item_t	*prev_lip;
+	struct xfs_log_item	*prev_lip;
+	struct xfs_log_item	*next_lip;
+	xfs_lsn_t		prev_lsn = NULLCOMMITLSN;
+	xfs_lsn_t		next_lsn = NULLCOMMITLSN;
+	xfs_lsn_t		lsn;
+	bool			in_ail;
+
 
 	if (list_empty(&ailp->ail_head))
 		return;
 
 	/*
-	 * Check the next and previous entries are valid.
+	 * Sample then check the next and previous entries are valid.
 	 */
-	ASSERT((lip->li_flags & XFS_LI_IN_AIL) != 0);
-	prev_lip = list_entry(lip->li_ail.prev, xfs_log_item_t, li_ail);
+	in_ail = test_bit(XFS_LI_IN_AIL, &lip->li_flags);
+	prev_lip = list_entry(lip->li_ail.prev, struct xfs_log_item, li_ail);
 	if (&prev_lip->li_ail != &ailp->ail_head)
-		ASSERT(XFS_LSN_CMP(prev_lip->li_lsn, lip->li_lsn) <= 0);
-
-	prev_lip = list_entry(lip->li_ail.next, xfs_log_item_t, li_ail);
-	if (&prev_lip->li_ail != &ailp->ail_head)
-		ASSERT(XFS_LSN_CMP(prev_lip->li_lsn, lip->li_lsn) >= 0);
+		prev_lsn = prev_lip->li_lsn;
+	next_lip = list_entry(lip->li_ail.next, struct xfs_log_item, li_ail);
+	if (&next_lip->li_ail != &ailp->ail_head)
+		next_lsn = next_lip->li_lsn;
+	lsn = lip->li_lsn;
 
+	if (in_ail &&
+	    (prev_lsn == NULLCOMMITLSN || XFS_LSN_CMP(prev_lsn, lsn) <= 0) &&
+	    (next_lsn == NULLCOMMITLSN || XFS_LSN_CMP(next_lsn, lsn) >= 0))
+		return;
 
+	spin_unlock(&ailp->ail_lock);
+	ASSERT(in_ail);
+	ASSERT(prev_lsn == NULLCOMMITLSN || XFS_LSN_CMP(prev_lsn, lsn) <= 0);
+	ASSERT(next_lsn == NULLCOMMITLSN || XFS_LSN_CMP(next_lsn, lsn) >= 0);
+	spin_lock(&ailp->ail_lock);
 }
 #else /* !DEBUG */
 #define	xfs_ail_check(a,l)
@@ -684,7 +705,7 @@ xfs_trans_ail_update_bulk(
 
 	for (i = 0; i < nr_items; i++) {
 		struct xfs_log_item *lip = log_items[i];
-		if (lip->li_flags & XFS_LI_IN_AIL) {
+		if (test_and_set_bit(XFS_LI_IN_AIL, &lip->li_flags)) {
 			/* check if we really need to move the item */
 			if (XFS_LSN_CMP(lsn, lip->li_lsn) <= 0)
 				continue;
@@ -694,7 +715,6 @@ xfs_trans_ail_update_bulk(
 			if (mlip == lip)
 				mlip_changed = 1;
 		} else {
-			lip->li_flags |= XFS_LI_IN_AIL;
 			trace_xfs_ail_insert(lip, 0, lsn);
 		}
 		lip->li_lsn = lsn;
@@ -725,7 +745,7 @@ xfs_ail_delete_one(
 	trace_xfs_ail_delete(lip, mlip->li_lsn, lip->li_lsn);
 	xfs_ail_delete(ailp, lip);
 	xfs_clear_li_failed(lip);
-	lip->li_flags &= ~XFS_LI_IN_AIL;
+	clear_bit(XFS_LI_IN_AIL, &lip->li_flags);
 	lip->li_lsn = 0;
 
 	return mlip == lip;
@@ -761,7 +781,7 @@ xfs_trans_ail_delete(
 	struct xfs_mount	*mp = ailp->ail_mount;
 	bool			mlip_changed;
 
-	if (!(lip->li_flags & XFS_LI_IN_AIL)) {
+	if (!test_bit(XFS_LI_IN_AIL, &lip->li_flags)) {
 		spin_unlock(&ailp->ail_lock);
 		if (!XFS_FORCED_SHUTDOWN(mp)) {
 			xfs_alert_tag(mp, XFS_PTAG_AILDELETE,
diff --git a/fs/xfs/xfs_trans_bmap.c b/fs/xfs/xfs_trans_bmap.c
index 14543d93cd4b..230a21df4b12 100644
--- a/fs/xfs/xfs_trans_bmap.c
+++ b/fs/xfs/xfs_trans_bmap.c
@@ -79,7 +79,7 @@ xfs_trans_log_finish_bmap_update(
 	 * 2.) shuts down the filesystem
 	 */
 	tp->t_flags |= XFS_TRANS_DIRTY;
-	budp->bud_item.li_desc->lid_flags |= XFS_LID_DIRTY;
+	set_bit(XFS_LI_DIRTY, &budp->bud_item.li_flags);
 
 	return error;
 }
@@ -158,7 +158,7 @@ xfs_bmap_update_log_item(
 	bmap = container_of(item, struct xfs_bmap_intent, bi_list);
 
 	tp->t_flags |= XFS_TRANS_DIRTY;
-	buip->bui_item.li_desc->lid_flags |= XFS_LID_DIRTY;
+	set_bit(XFS_LI_DIRTY, &buip->bui_item.li_flags);
 
 	/*
 	 * atomic_inc_return gives us the value after the increment;
diff --git a/fs/xfs/xfs_trans_buf.c b/fs/xfs/xfs_trans_buf.c
index a5d9dfc45d98..a8ddb4eed279 100644
--- a/fs/xfs/xfs_trans_buf.c
+++ b/fs/xfs/xfs_trans_buf.c
@@ -40,7 +40,7 @@ xfs_trans_buf_item_match(
 	struct xfs_buf_map	*map,
 	int			nmaps)
 {
-	struct xfs_log_item_desc *lidp;
+	struct xfs_log_item	*lip;
 	struct xfs_buf_log_item	*blip;
 	int			len = 0;
 	int			i;
@@ -48,8 +48,8 @@ xfs_trans_buf_item_match(
 	for (i = 0; i < nmaps; i++)
 		len += map[i].bm_len;
 
-	list_for_each_entry(lidp, &tp->t_items, lid_trans) {
-		blip = (struct xfs_buf_log_item *)lidp->lid_item;
+	list_for_each_entry(lip, &tp->t_items, li_trans) {
+		blip = (struct xfs_buf_log_item *)lip;
 		if (blip->bli_item.li_type == XFS_LI_BUF &&
 		    blip->bli_buf->b_target == target &&
 		    XFS_BUF_ADDR(blip->bli_buf) == map[0].bm_bn &&
@@ -100,14 +100,10 @@ _xfs_trans_bjoin(
 	atomic_inc(&bip->bli_refcount);
 
 	/*
-	 * Get a log_item_desc to point at the new item.
+	 * Attach the item to the transaction so we can find it in
+	 * xfs_trans_get_buf() and friends.
 	 */
 	xfs_trans_add_item(tp, &bip->bli_item);
-
-	/*
-	 * Initialize b_fsprivate2 so we can find it with incore_match()
-	 * in xfs_trans_get_buf() and friends above.
-	 */
 	bp->b_transp = tp;
 
 }
@@ -391,7 +387,7 @@ xfs_trans_brelse(
 	 * If the buffer is dirty within this transaction, we can't
 	 * release it until we commit.
 	 */
-	if (bip->bli_item.li_desc->lid_flags & XFS_LID_DIRTY)
+	if (test_bit(XFS_LI_DIRTY, &bip->bli_item.li_flags))
 		return;
 
 	/*
@@ -442,7 +438,7 @@ xfs_trans_brelse(
 		ASSERT(bp->b_pincount == 0);
 ***/
 		ASSERT(atomic_read(&bip->bli_refcount) == 0);
-		ASSERT(!(bip->bli_item.li_flags & XFS_LI_IN_AIL));
+		ASSERT(!test_bit(XFS_LI_IN_AIL, &bip->bli_item.li_flags));
 		ASSERT(!(bip->bli_flags & XFS_BLI_INODE_ALLOC_BUF));
 		xfs_buf_item_relse(bp);
 	}
@@ -542,7 +538,7 @@ xfs_trans_dirty_buf(
 	bip->bli_flags |= XFS_BLI_DIRTY | XFS_BLI_LOGGED;
 
 	tp->t_flags |= XFS_TRANS_DIRTY;
-	bip->bli_item.li_desc->lid_flags |= XFS_LID_DIRTY;
+	set_bit(XFS_LI_DIRTY, &bip->bli_item.li_flags);
 }
 
 /*
@@ -626,7 +622,7 @@ xfs_trans_binval(
 		ASSERT(!(bip->__bli_format.blf_flags & XFS_BLF_INODE_BUF));
 		ASSERT(!(bip->__bli_format.blf_flags & XFS_BLFT_MASK));
 		ASSERT(bip->__bli_format.blf_flags & XFS_BLF_CANCEL);
-		ASSERT(bip->bli_item.li_desc->lid_flags & XFS_LID_DIRTY);
+		ASSERT(test_bit(XFS_LI_DIRTY, &bip->bli_item.li_flags));
 		ASSERT(tp->t_flags & XFS_TRANS_DIRTY);
 		return;
 	}
@@ -642,7 +638,7 @@ xfs_trans_binval(
 		memset(bip->bli_formats[i].blf_data_map, 0,
 		       (bip->bli_formats[i].blf_map_size * sizeof(uint)));
 	}
-	bip->bli_item.li_desc->lid_flags |= XFS_LID_DIRTY;
+	set_bit(XFS_LI_DIRTY, &bip->bli_item.li_flags);
 	tp->t_flags |= XFS_TRANS_DIRTY;
 }
 
diff --git a/fs/xfs/xfs_trans_dquot.c b/fs/xfs/xfs_trans_dquot.c
index c3d547211d16..c381c02cca45 100644
--- a/fs/xfs/xfs_trans_dquot.c
+++ b/fs/xfs/xfs_trans_dquot.c
@@ -77,7 +77,7 @@ xfs_trans_log_dquot(
 	ASSERT(XFS_DQ_IS_LOCKED(dqp));
 
 	tp->t_flags |= XFS_TRANS_DIRTY;
-	dqp->q_logitem.qli_item.li_desc->lid_flags |= XFS_LID_DIRTY;
+	set_bit(XFS_LI_DIRTY, &dqp->q_logitem.qli_item.li_flags);
 }
 
 /*
@@ -879,7 +879,7 @@ xfs_trans_log_quotaoff_item(
 	xfs_qoff_logitem_t	*qlp)
 {
 	tp->t_flags |= XFS_TRANS_DIRTY;
-	qlp->qql_item.li_desc->lid_flags |= XFS_LID_DIRTY;
+	set_bit(XFS_LI_DIRTY, &qlp->qql_item.li_flags);
 }
 
 STATIC void
diff --git a/fs/xfs/xfs_trans_extfree.c b/fs/xfs/xfs_trans_extfree.c
index ab438647592a..2f44a08bdf65 100644
--- a/fs/xfs/xfs_trans_extfree.c
+++ b/fs/xfs/xfs_trans_extfree.c
@@ -68,7 +68,8 @@ xfs_trans_free_extent(
 	struct xfs_efd_log_item	*efdp,
 	xfs_fsblock_t		start_block,
 	xfs_extlen_t		ext_len,
-	struct xfs_owner_info	*oinfo)
+	struct xfs_owner_info	*oinfo,
+	bool			skip_discard)
 {
 	struct xfs_mount	*mp = tp->t_mountp;
 	uint			next_extent;
@@ -79,9 +80,8 @@ xfs_trans_free_extent(
 
 	trace_xfs_bmap_free_deferred(tp->t_mountp, agno, 0, agbno, ext_len);
 
-	error = xfs_free_extent(tp, start_block, ext_len, oinfo,
-			XFS_AG_RESV_NONE);
-
+	error = __xfs_free_extent(tp, start_block, ext_len,
+				  oinfo, XFS_AG_RESV_NONE, skip_discard);
 	/*
 	 * Mark the transaction dirty, even on error. This ensures the
 	 * transaction is aborted, which:
@@ -90,7 +90,7 @@ xfs_trans_free_extent(
 	 * 2.) shuts down the filesystem
 	 */
 	tp->t_flags |= XFS_TRANS_DIRTY;
-	efdp->efd_item.li_desc->lid_flags |= XFS_LID_DIRTY;
+	set_bit(XFS_LI_DIRTY, &efdp->efd_item.li_flags);
 
 	next_extent = efdp->efd_next_extent;
 	ASSERT(next_extent < efdp->efd_format.efd_nextents);
@@ -155,7 +155,7 @@ xfs_extent_free_log_item(
 	free = container_of(item, struct xfs_extent_free_item, xefi_list);
 
 	tp->t_flags |= XFS_TRANS_DIRTY;
-	efip->efi_item.li_desc->lid_flags |= XFS_LID_DIRTY;
+	set_bit(XFS_LI_DIRTY, &efip->efi_item.li_flags);
 
 	/*
 	 * atomic_inc_return gives us the value after the increment;
@@ -195,7 +195,7 @@ xfs_extent_free_finish_item(
 	error = xfs_trans_free_extent(tp, done_item,
 			free->xefi_startblock,
 			free->xefi_blockcount,
-			&free->xefi_oinfo);
+			&free->xefi_oinfo, free->xefi_skip_discard);
 	kmem_free(free);
 	return error;
 }
@@ -231,9 +231,79 @@ static const struct xfs_defer_op_type xfs_extent_free_defer_type = {
 	.cancel_item	= xfs_extent_free_cancel_item,
 };
 
+/*
+ * AGFL blocks are accounted differently in the reserve pools and are not
+ * inserted into the busy extent list.
+ */
+STATIC int
+xfs_agfl_free_finish_item(
+	struct xfs_trans		*tp,
+	struct xfs_defer_ops		*dop,
+	struct list_head		*item,
+	void				*done_item,
+	void				**state)
+{
+	struct xfs_mount		*mp = tp->t_mountp;
+	struct xfs_efd_log_item		*efdp = done_item;
+	struct xfs_extent_free_item	*free;
+	struct xfs_extent		*extp;
+	struct xfs_buf			*agbp;
+	int				error;
+	xfs_agnumber_t			agno;
+	xfs_agblock_t			agbno;
+	uint				next_extent;
+
+	free = container_of(item, struct xfs_extent_free_item, xefi_list);
+	ASSERT(free->xefi_blockcount == 1);
+	agno = XFS_FSB_TO_AGNO(mp, free->xefi_startblock);
+	agbno = XFS_FSB_TO_AGBNO(mp, free->xefi_startblock);
+
+	trace_xfs_agfl_free_deferred(mp, agno, 0, agbno, free->xefi_blockcount);
+
+	error = xfs_alloc_read_agf(mp, tp, agno, 0, &agbp);
+	if (!error)
+		error = xfs_free_agfl_block(tp, agno, agbno, agbp,
+					    &free->xefi_oinfo);
+
+	/*
+	 * Mark the transaction dirty, even on error. This ensures the
+	 * transaction is aborted, which:
+	 *
+	 * 1.) releases the EFI and frees the EFD
+	 * 2.) shuts down the filesystem
+	 */
+	tp->t_flags |= XFS_TRANS_DIRTY;
+	set_bit(XFS_LI_DIRTY, &efdp->efd_item.li_flags);
+
+	next_extent = efdp->efd_next_extent;
+	ASSERT(next_extent < efdp->efd_format.efd_nextents);
+	extp = &(efdp->efd_format.efd_extents[next_extent]);
+	extp->ext_start = free->xefi_startblock;
+	extp->ext_len = free->xefi_blockcount;
+	efdp->efd_next_extent++;
+
+	kmem_free(free);
+	return error;
+}
+
+
+/* sub-type with special handling for AGFL deferred frees */
+static const struct xfs_defer_op_type xfs_agfl_free_defer_type = {
+	.type		= XFS_DEFER_OPS_TYPE_AGFL_FREE,
+	.max_items	= XFS_EFI_MAX_FAST_EXTENTS,
+	.diff_items	= xfs_extent_free_diff_items,
+	.create_intent	= xfs_extent_free_create_intent,
+	.abort_intent	= xfs_extent_free_abort_intent,
+	.log_item	= xfs_extent_free_log_item,
+	.create_done	= xfs_extent_free_create_done,
+	.finish_item	= xfs_agfl_free_finish_item,
+	.cancel_item	= xfs_extent_free_cancel_item,
+};
+
 /* Register the deferred op type. */
 void
 xfs_extent_free_init_defer_op(void)
 {
 	xfs_defer_init_op_type(&xfs_extent_free_defer_type);
+	xfs_defer_init_op_type(&xfs_agfl_free_defer_type);
 }
diff --git a/fs/xfs/xfs_trans_inode.c b/fs/xfs/xfs_trans_inode.c
index 07cea592dc01..f7bd7960a90f 100644
--- a/fs/xfs/xfs_trans_inode.c
+++ b/fs/xfs/xfs_trans_inode.c
@@ -133,14 +133,13 @@ xfs_trans_log_inode(
 	 * set however, then go ahead and bump the i_version counter
 	 * unconditionally.
 	 */
-	if (!(ip->i_itemp->ili_item.li_desc->lid_flags & XFS_LID_DIRTY) &&
+	if (!test_and_set_bit(XFS_LI_DIRTY, &ip->i_itemp->ili_item.li_flags) &&
 	    IS_I_VERSION(VFS_I(ip))) {
 		if (inode_maybe_inc_iversion(VFS_I(ip), flags & XFS_ILOG_CORE))
 			flags |= XFS_ILOG_CORE;
 	}
 
 	tp->t_flags |= XFS_TRANS_DIRTY;
-	ip->i_itemp->ili_item.li_desc->lid_flags |= XFS_LID_DIRTY;
 
 	/*
 	 * Always OR in the bits from the ili_last_fields field.
diff --git a/fs/xfs/xfs_trans_priv.h b/fs/xfs/xfs_trans_priv.h
index be24b0c8a332..9717ae74b36d 100644
--- a/fs/xfs/xfs_trans_priv.h
+++ b/fs/xfs/xfs_trans_priv.h
@@ -19,7 +19,6 @@
 #define	__XFS_TRANS_PRIV_H__
 
 struct xfs_log_item;
-struct xfs_log_item_desc;
 struct xfs_mount;
 struct xfs_trans;
 struct xfs_ail;
@@ -119,7 +118,7 @@ xfs_trans_ail_remove(
 
 	spin_lock(&ailp->ail_lock);
 	/* xfs_trans_ail_delete() drops the AIL lock */
-	if (lip->li_flags & XFS_LI_IN_AIL)
+	if (test_bit(XFS_LI_IN_AIL, &lip->li_flags))
 		xfs_trans_ail_delete(ailp, lip, shutdown_type);
 	else
 		spin_unlock(&ailp->ail_lock);
@@ -171,11 +170,10 @@ xfs_clear_li_failed(
 {
 	struct xfs_buf	*bp = lip->li_buf;
 
-	ASSERT(lip->li_flags & XFS_LI_IN_AIL);
+	ASSERT(test_bit(XFS_LI_IN_AIL, &lip->li_flags));
 	lockdep_assert_held(&lip->li_ailp->ail_lock);
 
-	if (lip->li_flags & XFS_LI_FAILED) {
-		lip->li_flags &= ~XFS_LI_FAILED;
+	if (test_and_clear_bit(XFS_LI_FAILED, &lip->li_flags)) {
 		lip->li_buf = NULL;
 		xfs_buf_rele(bp);
 	}
@@ -188,9 +186,8 @@ xfs_set_li_failed(
 {
 	lockdep_assert_held(&lip->li_ailp->ail_lock);
 
-	if (!(lip->li_flags & XFS_LI_FAILED)) {
+	if (!test_and_set_bit(XFS_LI_FAILED, &lip->li_flags)) {
 		xfs_buf_hold(bp);
-		lip->li_flags |= XFS_LI_FAILED;
 		lip->li_buf = bp;
 	}
 }
diff --git a/fs/xfs/xfs_trans_refcount.c b/fs/xfs/xfs_trans_refcount.c
index 94c1877af834..c7f8e82f5bda 100644
--- a/fs/xfs/xfs_trans_refcount.c
+++ b/fs/xfs/xfs_trans_refcount.c
@@ -77,7 +77,7 @@ xfs_trans_log_finish_refcount_update(
 	 * 2.) shuts down the filesystem
 	 */
 	tp->t_flags |= XFS_TRANS_DIRTY;
-	cudp->cud_item.li_desc->lid_flags |= XFS_LID_DIRTY;
+	set_bit(XFS_LI_DIRTY, &cudp->cud_item.li_flags);
 
 	return error;
 }
@@ -154,7 +154,7 @@ xfs_refcount_update_log_item(
 	refc = container_of(item, struct xfs_refcount_intent, ri_list);
 
 	tp->t_flags |= XFS_TRANS_DIRTY;
-	cuip->cui_item.li_desc->lid_flags |= XFS_LID_DIRTY;
+	set_bit(XFS_LI_DIRTY, &cuip->cui_item.li_flags);
 
 	/*
 	 * atomic_inc_return gives us the value after the increment;
diff --git a/fs/xfs/xfs_trans_rmap.c b/fs/xfs/xfs_trans_rmap.c
index 9b577beb43d7..5831ca0c270b 100644
--- a/fs/xfs/xfs_trans_rmap.c
+++ b/fs/xfs/xfs_trans_rmap.c
@@ -117,7 +117,7 @@ xfs_trans_log_finish_rmap_update(
 	 * 2.) shuts down the filesystem
 	 */
 	tp->t_flags |= XFS_TRANS_DIRTY;
-	rudp->rud_item.li_desc->lid_flags |= XFS_LID_DIRTY;
+	set_bit(XFS_LI_DIRTY, &rudp->rud_item.li_flags);
 
 	return error;
 }
@@ -175,7 +175,7 @@ xfs_rmap_update_log_item(
 	rmap = container_of(item, struct xfs_rmap_intent, ri_list);
 
 	tp->t_flags |= XFS_TRANS_DIRTY;
-	ruip->rui_item.li_desc->lid_flags |= XFS_LID_DIRTY;
+	set_bit(XFS_LI_DIRTY, &ruip->rui_item.li_flags);
 
 	/*
 	 * atomic_inc_return gives us the value after the increment;
diff --git a/include/linux/bio.h b/include/linux/bio.h
index 810a8bee8f85..397a38aca182 100644
--- a/include/linux/bio.h
+++ b/include/linux/bio.h
@@ -127,6 +127,11 @@ static inline void *bio_data(struct bio *bio)
 	return NULL;
 }
 
+static inline bool bio_full(struct bio *bio)
+{
+	return bio->bi_vcnt >= bio->bi_max_vecs;
+}
+
 /*
  * will die
  */
@@ -474,6 +479,10 @@ void bio_chain(struct bio *, struct bio *);
 extern int bio_add_page(struct bio *, struct page *, unsigned int,unsigned int);
 extern int bio_add_pc_page(struct request_queue *, struct bio *, struct page *,
 			   unsigned int, unsigned int);
+bool __bio_try_merge_page(struct bio *bio, struct page *page,
+		unsigned int len, unsigned int off);
+void __bio_add_page(struct bio *bio, struct page *page,
+		unsigned int len, unsigned int off);
 int bio_iov_iter_get_pages(struct bio *bio, struct iov_iter *iter);
 struct rq_map_data;
 extern struct bio *bio_map_user_iov(struct request_queue *,
diff --git a/include/linux/buffer_head.h b/include/linux/buffer_head.h
index 894e5d125de6..96225a77c112 100644
--- a/include/linux/buffer_head.h
+++ b/include/linux/buffer_head.h
@@ -205,8 +205,6 @@ void write_boundary_block(struct block_device *bdev,
 			sector_t bblock, unsigned blocksize);
 int bh_uptodate_or_lock(struct buffer_head *bh);
 int bh_submit_read(struct buffer_head *bh);
-loff_t page_cache_seek_hole_data(struct inode *inode, loff_t offset,
-				 loff_t length, int whence);
 
 extern int buffer_heads_over_limit;
 
diff --git a/include/linux/dax.h b/include/linux/dax.h
index f9eb22ad341e..c99692ddd4b5 100644
--- a/include/linux/dax.h
+++ b/include/linux/dax.h
@@ -64,10 +64,10 @@ static inline bool dax_write_cache_enabled(struct dax_device *dax_dev)
 struct writeback_control;
 int bdev_dax_pgoff(struct block_device *, sector_t, size_t, pgoff_t *pgoff);
 #if IS_ENABLED(CONFIG_FS_DAX)
-int __bdev_dax_supported(struct super_block *sb, int blocksize);
-static inline int bdev_dax_supported(struct super_block *sb, int blocksize)
+bool __bdev_dax_supported(struct block_device *bdev, int blocksize);
+static inline bool bdev_dax_supported(struct block_device *bdev, int blocksize)
 {
-	return __bdev_dax_supported(sb, blocksize);
+	return __bdev_dax_supported(bdev, blocksize);
 }
 
 static inline struct dax_device *fs_dax_get_by_host(const char *host)
@@ -84,9 +84,10 @@ struct dax_device *fs_dax_get_by_bdev(struct block_device *bdev);
 int dax_writeback_mapping_range(struct address_space *mapping,
 		struct block_device *bdev, struct writeback_control *wbc);
 #else
-static inline int bdev_dax_supported(struct super_block *sb, int blocksize)
+static inline bool bdev_dax_supported(struct block_device *bdev,
+		int blocksize)
 {
-	return -EOPNOTSUPP;
+	return false;
 }
 
 static inline struct dax_device *fs_dax_get_by_host(const char *host)
diff --git a/include/linux/iomap.h b/include/linux/iomap.h
index 19a07de28212..a044a824da85 100644
--- a/include/linux/iomap.h
+++ b/include/linux/iomap.h
@@ -4,6 +4,7 @@
 
 #include <linux/types.h>
 
+struct address_space;
 struct fiemap_extent_info;
 struct inode;
 struct iov_iter;
@@ -18,6 +19,7 @@ struct vm_fault;
 #define IOMAP_DELALLOC	0x02	/* delayed allocation blocks */
 #define IOMAP_MAPPED	0x03	/* blocks allocated at @addr */
 #define IOMAP_UNWRITTEN	0x04	/* blocks allocated at @addr in unwritten state */
+#define IOMAP_INLINE	0x05	/* data inline in the inode */
 
 /*
  * Flags for all iomap mappings:
@@ -26,15 +28,19 @@ struct vm_fault;
  * written data and requires fdatasync to commit them to persistent storage.
  */
 #define IOMAP_F_NEW		0x01	/* blocks have been newly allocated */
-#define IOMAP_F_BOUNDARY	0x02	/* mapping ends at metadata boundary */
-#define IOMAP_F_DIRTY		0x04	/* uncommitted metadata */
+#define IOMAP_F_DIRTY		0x02	/* uncommitted metadata */
 
 /*
  * Flags that only need to be reported for IOMAP_REPORT requests:
  */
 #define IOMAP_F_MERGED		0x10	/* contains multiple blocks/extents */
 #define IOMAP_F_SHARED		0x20	/* block shared with another file */
-#define IOMAP_F_DATA_INLINE	0x40	/* data inline in the inode */
+
+/*
+ * Flags from 0x1000 up are for file system specific usage:
+ */
+#define IOMAP_F_PRIVATE		0x1000
+
 
 /*
  * Magic value for addr:
@@ -59,7 +65,7 @@ struct iomap {
 #define IOMAP_REPORT		(1 << 2) /* report extent status, e.g. FIEMAP */
 #define IOMAP_FAULT		(1 << 3) /* mapping for page fault */
 #define IOMAP_DIRECT		(1 << 4) /* direct I/O */
-#define IOMAP_NOWAIT		(1 << 5) /* Don't wait for writeback */
+#define IOMAP_NOWAIT		(1 << 5) /* do not block */
 
 struct iomap_ops {
 	/*
@@ -95,6 +101,8 @@ loff_t iomap_seek_hole(struct inode *inode, loff_t offset,
 		const struct iomap_ops *ops);
 loff_t iomap_seek_data(struct inode *inode, loff_t offset,
 		const struct iomap_ops *ops);
+sector_t iomap_bmap(struct address_space *mapping, sector_t bno,
+		const struct iomap_ops *ops);
 
 /*
  * Flags for direct I/O ->end_io:
@@ -106,4 +114,15 @@ typedef int (iomap_dio_end_io_t)(struct kiocb *iocb, ssize_t ret,
 ssize_t iomap_dio_rw(struct kiocb *iocb, struct iov_iter *iter,
 		const struct iomap_ops *ops, iomap_dio_end_io_t end_io);
 
+#ifdef CONFIG_SWAP
+struct file;
+struct swap_info_struct;
+
+int iomap_swapfile_activate(struct swap_info_struct *sis,
+		struct file *swap_file, sector_t *pagespan,
+		const struct iomap_ops *ops);
+#else
+# define iomap_swapfile_activate(sis, swapfile, pagespan, ops)	(-EIO)
+#endif /* CONFIG_SWAP */
+
 #endif /* LINUX_IOMAP_H */
diff --git a/include/uapi/linux/fs.h b/include/uapi/linux/fs.h
index d2a8313fabd7..9d132f8f2df8 100644
--- a/include/uapi/linux/fs.h
+++ b/include/uapi/linux/fs.h
@@ -242,6 +242,8 @@ struct fsxattr {
 #define FICLONERANGE	_IOW(0x94, 13, struct file_clone_range)
 #define FIDEDUPERANGE	_IOWR(0x94, 54, struct file_dedupe_range)
 
+#define FSLABEL_MAX 256	/* Max chars for the interface; each fs may differ */
+
 #define	FS_IOC_GETFLAGS			_IOR('f', 1, long)
 #define	FS_IOC_SETFLAGS			_IOW('f', 2, long)
 #define	FS_IOC_GETVERSION		_IOR('v', 1, long)
@@ -251,8 +253,10 @@ struct fsxattr {
 #define FS_IOC32_SETFLAGS		_IOW('f', 2, int)
 #define FS_IOC32_GETVERSION		_IOR('v', 1, int)
 #define FS_IOC32_SETVERSION		_IOW('v', 2, int)
-#define FS_IOC_FSGETXATTR		_IOR ('X', 31, struct fsxattr)
-#define FS_IOC_FSSETXATTR		_IOW ('X', 32, struct fsxattr)
+#define FS_IOC_FSGETXATTR		_IOR('X', 31, struct fsxattr)
+#define FS_IOC_FSSETXATTR		_IOW('X', 32, struct fsxattr)
+#define FS_IOC_GETFSLABEL		_IOR(0x94, 49, char[FSLABEL_MAX])
+#define FS_IOC_SETFSLABEL		_IOW(0x94, 50, char[FSLABEL_MAX])
 
 /*
  * File system encryption support
diff --git a/mm/internal.h b/mm/internal.h
index 502d14189794..9e3654d70289 100644
--- a/mm/internal.h
+++ b/mm/internal.h
@@ -53,7 +53,7 @@ void unmap_page_range(struct mmu_gather *tlb,
 			     unsigned long addr, unsigned long end,
 			     struct zap_details *details);
 
-extern int __do_page_cache_readahead(struct address_space *mapping,
+extern unsigned int __do_page_cache_readahead(struct address_space *mapping,
 		struct file *filp, pgoff_t offset, unsigned long nr_to_read,
 		unsigned long lookahead_size);
 
diff --git a/mm/readahead.c b/mm/readahead.c
index 539bbb6c1fad..e273f0de3376 100644
--- a/mm/readahead.c
+++ b/mm/readahead.c
@@ -140,23 +140,23 @@ out:
 }
 
 /*
- * __do_page_cache_readahead() actually reads a chunk of disk.  It allocates all
- * the pages first, then submits them all for I/O. This avoids the very bad
+ * __do_page_cache_readahead() actually reads a chunk of disk.  It allocates
+ * the pages first, then submits them for I/O. This avoids the very bad
  * behaviour which would occur if page allocations are causing VM writeback.
  * We really don't want to intermingle reads and writes like that.
  *
  * Returns the number of pages requested, or the maximum amount of I/O allowed.
  */
-int __do_page_cache_readahead(struct address_space *mapping, struct file *filp,
-			pgoff_t offset, unsigned long nr_to_read,
-			unsigned long lookahead_size)
+unsigned int __do_page_cache_readahead(struct address_space *mapping,
+		struct file *filp, pgoff_t offset, unsigned long nr_to_read,
+		unsigned long lookahead_size)
 {
 	struct inode *inode = mapping->host;
 	struct page *page;
 	unsigned long end_index;	/* The last page we want to read */
 	LIST_HEAD(page_pool);
 	int page_idx;
-	int ret = 0;
+	unsigned int nr_pages = 0;
 	loff_t isize = i_size_read(inode);
 	gfp_t gfp_mask = readahead_gfp_mask(mapping);
 
@@ -177,8 +177,18 @@ int __do_page_cache_readahead(struct address_space *mapping, struct file *filp,
 		rcu_read_lock();
 		page = radix_tree_lookup(&mapping->i_pages, page_offset);
 		rcu_read_unlock();
-		if (page && !radix_tree_exceptional_entry(page))
+		if (page && !radix_tree_exceptional_entry(page)) {
+			/*
+			 * Page already present?  Kick off the current batch of
+			 * contiguous pages before continuing with the next
+			 * batch.
+			 */
+			if (nr_pages)
+				read_pages(mapping, filp, &page_pool, nr_pages,
+						gfp_mask);
+			nr_pages = 0;
 			continue;
+		}
 
 		page = __page_cache_alloc(gfp_mask);
 		if (!page)
@@ -187,7 +197,7 @@ int __do_page_cache_readahead(struct address_space *mapping, struct file *filp,
 		list_add(&page->lru, &page_pool);
 		if (page_idx == nr_to_read - lookahead_size)
 			SetPageReadahead(page);
-		ret++;
+		nr_pages++;
 	}
 
 	/*
@@ -195,11 +205,11 @@ int __do_page_cache_readahead(struct address_space *mapping, struct file *filp,
 	 * uptodate then the caller will launch readpage again, and
 	 * will then handle the error.
 	 */
-	if (ret)
-		read_pages(mapping, filp, &page_pool, ret, gfp_mask);
+	if (nr_pages)
+		read_pages(mapping, filp, &page_pool, nr_pages, gfp_mask);
 	BUG_ON(!list_empty(&page_pool));
 out:
-	return ret;
+	return nr_pages;
 }
 
 /*
@@ -223,16 +233,11 @@ int force_page_cache_readahead(struct address_space *mapping, struct file *filp,
 	max_pages = max_t(unsigned long, bdi->io_pages, ra->ra_pages);
 	nr_to_read = min(nr_to_read, max_pages);
 	while (nr_to_read) {
-		int err;
-
 		unsigned long this_chunk = (2 * 1024 * 1024) / PAGE_SIZE;
 
 		if (this_chunk > nr_to_read)
 			this_chunk = nr_to_read;
-		err = __do_page_cache_readahead(mapping, filp,
-						offset, this_chunk, 0);
-		if (err < 0)
-			return err;
+		__do_page_cache_readahead(mapping, filp, offset, this_chunk, 0);
 
 		offset += this_chunk;
 		nr_to_read -= this_chunk;