Merge tag 'xfs-for-linus-4.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/dgc/linux-xfs

Pull xfs updates from Dave Chinner:
 "There is quite a varied bunch of stuff in this update, and some of it
  you will have already merged through the ext4 tree which imported the
  dax-4.10-iomap-pmd topic branch from the XFS tree.

  There is also a new direct IO implementation that uses the iomap
  infrastructure. It's much simpler, faster, and has lower IO latency
  than the existing direct IO infrastructure.

  Summary:
   - DAX PMD faults via iomap infrastructure
   - Direct-io support in iomap infrastructure
   - removal of now-redundant XFS inode iolock, replaced with VFS
     i_rwsem
   - synchronisation with fixes and changes in userspace libxfs code
   - extent tree lookup helpers
   - lots of little corruption detection improvements to verifiers
   - optimised CRC calculations
   - faster buffer cache lookups
   - deprecation of barrier/nobarrier mount options - we always use
     REQ_FUA/REQ_FLUSH where appropriate for data integrity now
   - cleanups to speculative preallocation
   - miscellaneous minor bug fixes and cleanups"

* tag 'xfs-for-linus-4.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/dgc/linux-xfs: (63 commits)
  xfs: nuke unused tracepoint definitions
  xfs: use GPF_NOFS when allocating btree cursors
  xfs: use xfs_vn_setattr_size to check on new size
  xfs: deprecate barrier/nobarrier mount option
  xfs: Always flush caches when integrity is required
  xfs: ignore leaf attr ichdr.count in verifier during log replay
  xfs: use rhashtable to track buffer cache
  xfs: optimise CRC updates
  xfs: make xfs btree stats less huge
  xfs: don't cap maximum dedupe request length
  xfs: don't allow di_size with high bit set
  xfs: error out if trying to add attrs and anextents > 0
  xfs: don't crash if reading a directory results in an unexpected hole
  xfs: complain if we don't get nextents bmap records
  xfs: check for bogus values in btree block headers
  xfs: forbid AG btrees with level == 0
  xfs: several xattr functions can be void
  xfs: handle cow fork in xfs_bmap_trace_exlist
  xfs: pass state not whichfork to trace_xfs_extlist
  xfs: Move AGI buffer type setting to xfs_read_agi
  ...

Documentation/filesystems/xfs.txt

@@ -51,13 +51,6 @@ default behaviour.
 	CRC enabled filesystems always use the attr2 format, and so
 	will reject the noattr2 mount option if it is set.
 
-  barrier (*)
-  nobarrier
-	Enables/disables the use of block layer write barriers for
-	writes into the journal and for data integrity operations.
-	This allows for drive level write caching to be enabled, for
-	devices that support write barriers.
-
   discard
   nodiscard (*)
 	Enable/disable the issuing of commands to let the block
@@ -228,7 +221,10 @@ default behaviour.
 Deprecated Mount Options
 ========================
 
-None at present.
+  Name				Removal Schedule
+  ----				----------------
+  barrier			no earlier than v4.15
+  nobarrier			no earlier than v4.15
 
 
 Removed Mount Options

fs/direct-io.c

@@ -554,7 +554,7 @@ static inline int dio_bio_reap(struct dio *dio, struct dio_submit *sdio)
  * filesystems that don't need it and also allows us to create the workqueue
  * late enough so the we can include s_id in the name of the workqueue.
  */
-static int sb_init_dio_done_wq(struct super_block *sb)
+int sb_init_dio_done_wq(struct super_block *sb)
 {
 	struct workqueue_struct *old;
 	struct workqueue_struct *wq = alloc_workqueue("dio/%s",

fs/internal.h

@@ -184,3 +184,6 @@ typedef loff_t (*iomap_actor_t)(struct inode *inode, loff_t pos, loff_t len,
 loff_t iomap_apply(struct inode *inode, loff_t pos, loff_t length,
 		unsigned flags, struct iomap_ops *ops, void *data,
 		iomap_actor_t actor);
+
+/* direct-io.c: */
+int sb_init_dio_done_wq(struct super_block *sb);

fs/iomap.c

@@ -24,6 +24,7 @@
 #include <linux/uio.h>
 #include <linux/backing-dev.h>
 #include <linux/buffer_head.h>
+#include <linux/task_io_accounting_ops.h>
 #include <linux/dax.h>
 #include "internal.h"
 
@@ -584,3 +585,375 @@ int iomap_fiemap(struct inode *inode, struct fiemap_extent_info *fi,
 	return 0;
 }
 EXPORT_SYMBOL_GPL(iomap_fiemap);
+
+/*
+ * Private flags for iomap_dio, must not overlap with the public ones in
+ * iomap.h:
+ */
+#define IOMAP_DIO_WRITE		(1 << 30)
+#define IOMAP_DIO_DIRTY		(1 << 31)
+
+struct iomap_dio {
+	struct kiocb		*iocb;
+	iomap_dio_end_io_t	*end_io;
+	loff_t			i_size;
+	loff_t			size;
+	atomic_t		ref;
+	unsigned		flags;
+	int			error;
+
+	union {
+		/* used during submission and for synchronous completion: */
+		struct {
+			struct iov_iter		*iter;
+			struct task_struct	*waiter;
+			struct request_queue	*last_queue;
+			blk_qc_t		cookie;
+		} submit;
+
+		/* used for aio completion: */
+		struct {
+			struct work_struct	work;
+		} aio;
+	};
+};
+
+static ssize_t iomap_dio_complete(struct iomap_dio *dio)
+{
+	struct kiocb *iocb = dio->iocb;
+	ssize_t ret;
+
+	if (dio->end_io) {
+		ret = dio->end_io(iocb,
+				dio->error ? dio->error : dio->size,
+				dio->flags);
+	} else {
+		ret = dio->error;
+	}
+
+	if (likely(!ret)) {
+		ret = dio->size;
+		/* check for short read */
+		if (iocb->ki_pos + ret > dio->i_size &&
+		    !(dio->flags & IOMAP_DIO_WRITE))
+			ret = dio->i_size - iocb->ki_pos;
+		iocb->ki_pos += ret;
+	}
+
+	inode_dio_end(file_inode(iocb->ki_filp));
+	kfree(dio);
+
+	return ret;
+}
+
+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);
+}
+
+/*
+ * Set an error in the dio if none is set yet.  We have to use cmpxchg
+ * as the submission context and the completion context(s) can race to
+ * update the error.
+ */
+static inline void iomap_dio_set_error(struct iomap_dio *dio, int ret)
+{
+	cmpxchg(&dio->error, 0, ret);
+}
+
+static void iomap_dio_bio_end_io(struct bio *bio)
+{
+	struct iomap_dio *dio = bio->bi_private;
+	bool should_dirty = (dio->flags & IOMAP_DIO_DIRTY);
+
+	if (bio->bi_error)
+		iomap_dio_set_error(dio, bio->bi_error);
+
+	if (atomic_dec_and_test(&dio->ref)) {
+		if (is_sync_kiocb(dio->iocb)) {
+			struct task_struct *waiter = dio->submit.waiter;
+
+			WRITE_ONCE(dio->submit.waiter, NULL);
+			wake_up_process(waiter);
+		} else if (dio->flags & IOMAP_DIO_WRITE) {
+			struct inode *inode = file_inode(dio->iocb->ki_filp);
+
+			INIT_WORK(&dio->aio.work, iomap_dio_complete_work);
+			queue_work(inode->i_sb->s_dio_done_wq, &dio->aio.work);
+		} else {
+			iomap_dio_complete_work(&dio->aio.work);
+		}
+	}
+
+	if (should_dirty) {
+		bio_check_pages_dirty(bio);
+	} else {
+		struct bio_vec *bvec;
+		int i;
+
+		bio_for_each_segment_all(bvec, bio, i)
+			put_page(bvec->bv_page);
+		bio_put(bio);
+	}
+}
+
+static blk_qc_t
+iomap_dio_zero(struct iomap_dio *dio, struct iomap *iomap, loff_t pos,
+		unsigned len)
+{
+	struct page *page = ZERO_PAGE(0);
+	struct bio *bio;
+
+	bio = bio_alloc(GFP_KERNEL, 1);
+	bio->bi_bdev = iomap->bdev;
+	bio->bi_iter.bi_sector =
+		iomap->blkno + ((pos - iomap->offset) >> 9);
+	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_set_op_attrs(bio, REQ_OP_WRITE, REQ_SYNC | REQ_IDLE);
+
+	atomic_inc(&dio->ref);
+	return submit_bio(bio);
+}
+
+static loff_t
+iomap_dio_actor(struct inode *inode, loff_t pos, loff_t length,
+		void *data, struct iomap *iomap)
+{
+	struct iomap_dio *dio = data;
+	unsigned blkbits = blksize_bits(bdev_logical_block_size(iomap->bdev));
+	unsigned fs_block_size = (1 << inode->i_blkbits), pad;
+	unsigned align = iov_iter_alignment(dio->submit.iter);
+	struct iov_iter iter;
+	struct bio *bio;
+	bool need_zeroout = false;
+	int nr_pages, ret;
+
+	if ((pos | length | align) & ((1 << blkbits) - 1))
+		return -EINVAL;
+
+	switch (iomap->type) {
+	case IOMAP_HOLE:
+		if (WARN_ON_ONCE(dio->flags & IOMAP_DIO_WRITE))
+			return -EIO;
+		/*FALLTHRU*/
+	case IOMAP_UNWRITTEN:
+		if (!(dio->flags & IOMAP_DIO_WRITE)) {
+			iov_iter_zero(length, dio->submit.iter);
+			dio->size += length;
+			return length;
+		}
+		dio->flags |= IOMAP_DIO_UNWRITTEN;
+		need_zeroout = true;
+		break;
+	case IOMAP_MAPPED:
+		if (iomap->flags & IOMAP_F_SHARED)
+			dio->flags |= IOMAP_DIO_COW;
+		if (iomap->flags & IOMAP_F_NEW)
+			need_zeroout = true;
+		break;
+	default:
+		WARN_ON_ONCE(1);
+		return -EIO;
+	}
+
+	/*
+	 * Operate on a partial iter trimmed to the extent we were called for.
+	 * We'll update the iter in the dio once we're done with this extent.
+	 */
+	iter = *dio->submit.iter;
+	iov_iter_truncate(&iter, length);
+
+	nr_pages = iov_iter_npages(&iter, BIO_MAX_PAGES);
+	if (nr_pages <= 0)
+		return nr_pages;
+
+	if (need_zeroout) {
+		/* zero out from the start of the block to the write offset */
+		pad = pos & (fs_block_size - 1);
+		if (pad)
+			iomap_dio_zero(dio, iomap, pos - pad, pad);
+	}
+
+	do {
+		if (dio->error)
+			return 0;
+
+		bio = bio_alloc(GFP_KERNEL, nr_pages);
+		bio->bi_bdev = iomap->bdev;
+		bio->bi_iter.bi_sector =
+			iomap->blkno + ((pos - iomap->offset) >> 9);
+		bio->bi_private = dio;
+		bio->bi_end_io = iomap_dio_bio_end_io;
+
+		ret = bio_iov_iter_get_pages(bio, &iter);
+		if (unlikely(ret)) {
+			bio_put(bio);
+			return ret;
+		}
+
+		if (dio->flags & IOMAP_DIO_WRITE) {
+			bio_set_op_attrs(bio, REQ_OP_WRITE, REQ_SYNC | REQ_IDLE);
+			task_io_account_write(bio->bi_iter.bi_size);
+		} else {
+			bio_set_op_attrs(bio, REQ_OP_READ, 0);
+			if (dio->flags & IOMAP_DIO_DIRTY)
+				bio_set_pages_dirty(bio);
+		}
+
+		dio->size += bio->bi_iter.bi_size;
+		pos += bio->bi_iter.bi_size;
+
+		nr_pages = iov_iter_npages(&iter, BIO_MAX_PAGES);
+
+		atomic_inc(&dio->ref);
+
+		dio->submit.last_queue = bdev_get_queue(iomap->bdev);
+		dio->submit.cookie = submit_bio(bio);
+	} while (nr_pages);
+
+	if (need_zeroout) {
+		/* zero out from the end of the write to the end of the block */
+		pad = pos & (fs_block_size - 1);
+		if (pad)
+			iomap_dio_zero(dio, iomap, pos, fs_block_size - pad);
+	}
+
+	iov_iter_advance(dio->submit.iter, length);
+	return length;
+}
+
+ssize_t
+iomap_dio_rw(struct kiocb *iocb, struct iov_iter *iter, struct iomap_ops *ops,
+		iomap_dio_end_io_t end_io)
+{
+	struct address_space *mapping = iocb->ki_filp->f_mapping;
+	struct inode *inode = file_inode(iocb->ki_filp);
+	size_t count = iov_iter_count(iter);
+	loff_t pos = iocb->ki_pos, end = iocb->ki_pos + count - 1, ret = 0;
+	unsigned int flags = IOMAP_DIRECT;
+	struct blk_plug plug;
+	struct iomap_dio *dio;
+
+	lockdep_assert_held(&inode->i_rwsem);
+
+	if (!count)
+		return 0;
+
+	dio = kmalloc(sizeof(*dio), GFP_KERNEL);
+	if (!dio)
+		return -ENOMEM;
+
+	dio->iocb = iocb;
+	atomic_set(&dio->ref, 1);
+	dio->size = 0;
+	dio->i_size = i_size_read(inode);
+	dio->end_io = end_io;
+	dio->error = 0;
+	dio->flags = 0;
+
+	dio->submit.iter = iter;
+	if (is_sync_kiocb(iocb)) {
+		dio->submit.waiter = current;
+		dio->submit.cookie = BLK_QC_T_NONE;
+		dio->submit.last_queue = NULL;
+	}
+
+	if (iov_iter_rw(iter) == READ) {
+		if (pos >= dio->i_size)
+			goto out_free_dio;
+
+		if (iter->type == ITER_IOVEC)
+			dio->flags |= IOMAP_DIO_DIRTY;
+	} else {
+		dio->flags |= IOMAP_DIO_WRITE;
+		flags |= IOMAP_WRITE;
+	}
+
+	if (mapping->nrpages) {
+		ret = filemap_write_and_wait_range(mapping, iocb->ki_pos, end);
+		if (ret)
+			goto out_free_dio;
+
+		ret = invalidate_inode_pages2_range(mapping,
+				iocb->ki_pos >> PAGE_SHIFT, end >> PAGE_SHIFT);
+		WARN_ON_ONCE(ret);
+		ret = 0;
+	}
+
+	inode_dio_begin(inode);
+
+	blk_start_plug(&plug);
+	do {
+		ret = iomap_apply(inode, pos, count, flags, ops, dio,
+				iomap_dio_actor);
+		if (ret <= 0) {
+			/* magic error code to fall back to buffered I/O */
+			if (ret == -ENOTBLK)
+				ret = 0;
+			break;
+		}
+		pos += ret;
+	} while ((count = iov_iter_count(iter)) > 0);
+	blk_finish_plug(&plug);
+
+	if (ret < 0)
+		iomap_dio_set_error(dio, ret);
+
+	if (ret >= 0 && iov_iter_rw(iter) == WRITE && !is_sync_kiocb(iocb) &&
+			!inode->i_sb->s_dio_done_wq) {
+		ret = sb_init_dio_done_wq(inode->i_sb);
+		if (ret < 0)
+			iomap_dio_set_error(dio, ret);
+	}
+
+	if (!atomic_dec_and_test(&dio->ref)) {
+		if (!is_sync_kiocb(iocb))
+			return -EIOCBQUEUED;
+
+		for (;;) {
+			set_current_state(TASK_UNINTERRUPTIBLE);
+			if (!READ_ONCE(dio->submit.waiter))
+				break;
+
+			if (!(iocb->ki_flags & IOCB_HIPRI) ||
+			    !dio->submit.last_queue ||
+			    !blk_mq_poll(dio->submit.last_queue,
+					 dio->submit.cookie))
+				io_schedule();
+		}
+		__set_current_state(TASK_RUNNING);
+	}
+
+	/*
+	 * Try again to invalidate clean pages which might have been cached by
+	 * non-direct readahead, or faulted in by get_user_pages() if the source
+	 * of the write was an mmap'ed region of the file we're writing.  Either
+	 * one is a pretty crazy thing to do, so we don't support it 100%.  If
+	 * this invalidation fails, tough, the write still worked...
+	 */
+	if (iov_iter_rw(iter) == WRITE && mapping->nrpages) {
+		ret = invalidate_inode_pages2_range(mapping,
+				iocb->ki_pos >> PAGE_SHIFT, end >> PAGE_SHIFT);
+		WARN_ON_ONCE(ret);
+	}
+
+	return iomap_dio_complete(dio);
+
+out_free_dio:
+	kfree(dio);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(iomap_dio_rw);

fs/xfs/libxfs/xfs_alloc.c

@@ -2455,12 +2455,15 @@ xfs_agf_verify(
 	      be32_to_cpu(agf->agf_flcount) <= XFS_AGFL_SIZE(mp)))
 		return false;
 
-	if (be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]) > XFS_BTREE_MAXLEVELS ||
+	if (be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]) < 1 ||
+	    be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]) < 1 ||
+	    be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]) > XFS_BTREE_MAXLEVELS ||
 	    be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]) > XFS_BTREE_MAXLEVELS)
 		return false;
 
 	if (xfs_sb_version_hasrmapbt(&mp->m_sb) &&
-	    be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAP]) > XFS_BTREE_MAXLEVELS)
+	    (be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAP]) < 1 ||
+	     be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAP]) > XFS_BTREE_MAXLEVELS))
 		return false;
 
 	/*
@@ -2477,7 +2480,8 @@ xfs_agf_verify(
 		return false;
 
 	if (xfs_sb_version_hasreflink(&mp->m_sb) &&
-	    be32_to_cpu(agf->agf_refcount_level) > XFS_BTREE_MAXLEVELS)
+	    (be32_to_cpu(agf->agf_refcount_level) < 1 ||
+	     be32_to_cpu(agf->agf_refcount_level) > XFS_BTREE_MAXLEVELS))
 		return false;
 
 	return true;;

fs/xfs/libxfs/xfs_alloc_btree.c

@@ -421,13 +421,17 @@ xfs_allocbt_init_cursor(
 
 	ASSERT(btnum == XFS_BTNUM_BNO || btnum == XFS_BTNUM_CNT);
 
-	cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_SLEEP);
+	cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_NOFS);
 
 	cur->bc_tp = tp;
 	cur->bc_mp = mp;
 	cur->bc_btnum = btnum;
 	cur->bc_blocklog = mp->m_sb.sb_blocklog;
 	cur->bc_ops = &xfs_allocbt_ops;
+	if (btnum == XFS_BTNUM_BNO)
+		cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_abtb_2);
+	else
+		cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_abtc_2);
 
 	if (btnum == XFS_BTNUM_CNT) {
 		cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]);

fs/xfs/libxfs/xfs_attr_leaf.c

@@ -253,6 +253,7 @@ xfs_attr3_leaf_verify(
 {
 	struct xfs_mount	*mp = bp->b_target->bt_mount;
 	struct xfs_attr_leafblock *leaf = bp->b_addr;
+	struct xfs_perag *pag = bp->b_pag;
 	struct xfs_attr3_icleaf_hdr ichdr;
 
 	xfs_attr3_leaf_hdr_from_disk(mp->m_attr_geo, &ichdr, leaf);
@@ -273,7 +274,12 @@ xfs_attr3_leaf_verify(
 		if (ichdr.magic != XFS_ATTR_LEAF_MAGIC)
 			return false;
 	}
-	if (ichdr.count == 0)
+	/*
+	 * In recovery there is a transient state where count == 0 is valid
+	 * because we may have transitioned an empty shortform attr to a leaf
+	 * if the attr didn't fit in shortform.
+	 */
+	if (pag && pag->pagf_init && ichdr.count == 0)
 		return false;
 
 	/* XXX: need to range check rest of attr header values */

fs/xfs/libxfs/xfs_attr_leaf.h

@@ -51,7 +51,7 @@ int	xfs_attr_shortform_getvalue(struct xfs_da_args *args);
 int	xfs_attr_shortform_to_leaf(struct xfs_da_args *args);
 int	xfs_attr_shortform_remove(struct xfs_da_args *args);
 int	xfs_attr_shortform_allfit(struct xfs_buf *bp, struct xfs_inode *dp);
-int	xfs_attr_shortform_bytesfit(xfs_inode_t *dp, int bytes);
+int	xfs_attr_shortform_bytesfit(struct xfs_inode *dp, int bytes);
 void	xfs_attr_fork_remove(struct xfs_inode *ip, struct xfs_trans *tp);
 
 /*
@@ -77,7 +77,7 @@ int	xfs_attr3_leaf_add(struct xfs_buf *leaf_buffer,
 				 struct xfs_da_args *args);
 int	xfs_attr3_leaf_remove(struct xfs_buf *leaf_buffer,
 				    struct xfs_da_args *args);
-int	xfs_attr3_leaf_list_int(struct xfs_buf *bp,
+void	xfs_attr3_leaf_list_int(struct xfs_buf *bp,
 				      struct xfs_attr_list_context *context);
 
 /*

fs/xfs/libxfs/xfs_bmap.c

@@ -49,6 +49,8 @@
 #include "xfs_rmap.h"
 #include "xfs_ag_resv.h"
 #include "xfs_refcount.h"
+#include "xfs_rmap_btree.h"
+#include "xfs_icache.h"
 
 
 kmem_zone_t		*xfs_bmap_free_item_zone;
@@ -190,8 +192,12 @@ xfs_bmap_worst_indlen(
 	int		maxrecs;	/* maximum record count at this level */
 	xfs_mount_t	*mp;		/* mount structure */
 	xfs_filblks_t	rval;		/* return value */
+	xfs_filblks_t   orig_len;
 
 	mp = ip->i_mount;
+
+	/* Calculate the worst-case size of the bmbt. */
+	orig_len = len;
 	maxrecs = mp->m_bmap_dmxr[0];
 	for (level = 0, rval = 0;
 	     level < XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK);
@@ -199,12 +205,20 @@ xfs_bmap_worst_indlen(
 		len += maxrecs - 1;
 		do_div(len, maxrecs);
 		rval += len;
-		if (len == 1)
-			return rval + XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK) -
+		if (len == 1) {
+			rval += XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK) -
 				level - 1;
+			break;
+		}
 		if (level == 0)
 			maxrecs = mp->m_bmap_dmxr[1];
 	}
+
+	/* Calculate the worst-case size of the rmapbt. */
+	if (xfs_sb_version_hasrmapbt(&mp->m_sb))
+		rval += 1 + xfs_rmapbt_calc_size(mp, orig_len) +
+				mp->m_rmap_maxlevels;
+
 	return rval;
 }
 
@@ -504,7 +518,7 @@ void
 xfs_bmap_trace_exlist(
 	xfs_inode_t	*ip,		/* incore inode pointer */
 	xfs_extnum_t	cnt,		/* count of entries in the list */
-	int		whichfork,	/* data or attr fork */
+	int		whichfork,	/* data or attr or cow fork */
 	unsigned long	caller_ip)
 {
 	xfs_extnum_t	idx;		/* extent record index */
@@ -513,11 +527,13 @@ xfs_bmap_trace_exlist(
 
 	if (whichfork == XFS_ATTR_FORK)
 		state |= BMAP_ATTRFORK;
+	else if (whichfork == XFS_COW_FORK)
+		state |= BMAP_COWFORK;
 
 	ifp = XFS_IFORK_PTR(ip, whichfork);
-	ASSERT(cnt == (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)));
+	ASSERT(cnt == xfs_iext_count(ifp));
 	for (idx = 0; idx < cnt; idx++)
-		trace_xfs_extlist(ip, idx, whichfork, caller_ip);
+		trace_xfs_extlist(ip, idx, state, caller_ip);
 }
 
 /*
@@ -811,7 +827,7 @@ try_another_ag:
 				XFS_BTREE_LONG_PTRS);
 
 	arp = XFS_BMBT_REC_ADDR(mp, ablock, 1);
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	nextents =  xfs_iext_count(ifp);
 	for (cnt = i = 0; i < nextents; i++) {
 		ep = xfs_iext_get_ext(ifp, i);
 		if (!isnullstartblock(xfs_bmbt_get_startblock(ep))) {
@@ -1137,6 +1153,10 @@ xfs_bmap_add_attrfork(
 		goto trans_cancel;
 	if (XFS_IFORK_Q(ip))
 		goto trans_cancel;
+	if (ip->i_d.di_anextents != 0) {
+		error = -EFSCORRUPTED;
+		goto trans_cancel;
+	}
 	if (ip->i_d.di_aformat != XFS_DINODE_FMT_EXTENTS) {
 		/*
 		 * For inodes coming from pre-6.2 filesystems.
@@ -1144,7 +1164,6 @@ xfs_bmap_add_attrfork(
 		ASSERT(ip->i_d.di_aformat == 0);
 		ip->i_d.di_aformat = XFS_DINODE_FMT_EXTENTS;
 	}
-	ASSERT(ip->i_d.di_anextents == 0);
 
 	xfs_trans_ijoin(tp, ip, 0);
 	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
@@ -1296,7 +1315,7 @@ xfs_bmap_read_extents(
 	/*
 	 * Here with bp and block set to the leftmost leaf node in the tree.
 	 */
-	room = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	room = xfs_iext_count(ifp);
 	i = 0;
 	/*
 	 * Loop over all leaf nodes.  Copy information to the extent records.
@@ -1361,8 +1380,9 @@ xfs_bmap_read_extents(
 			return error;
 		block = XFS_BUF_TO_BLOCK(bp);
 	}
-	ASSERT(i == (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)));
-	ASSERT(i == XFS_IFORK_NEXTENTS(ip, whichfork));
+	if (i != XFS_IFORK_NEXTENTS(ip, whichfork))
+		return -EFSCORRUPTED;
+	ASSERT(i == xfs_iext_count(ifp));
 	XFS_BMAP_TRACE_EXLIST(ip, i, whichfork);
 	return 0;
 error0:
@@ -1370,97 +1390,6 @@ error0:
 	return -EFSCORRUPTED;
 }
 
-
-/*
- * Search the extent records for the entry containing block bno.
- * If bno lies in a hole, point to the next entry.  If bno lies
- * past eof, *eofp will be set, and *prevp will contain the last
- * entry (null if none).  Else, *lastxp will be set to the index
- * of the found entry; *gotp will contain the entry.
- */
-STATIC xfs_bmbt_rec_host_t *		/* pointer to found extent entry */
-xfs_bmap_search_multi_extents(
-	xfs_ifork_t	*ifp,		/* inode fork pointer */
-	xfs_fileoff_t	bno,		/* block number searched for */
-	int		*eofp,		/* out: end of file found */
-	xfs_extnum_t	*lastxp,	/* out: last extent index */
-	xfs_bmbt_irec_t	*gotp,		/* out: extent entry found */
-	xfs_bmbt_irec_t	*prevp)		/* out: previous extent entry found */
-{
-	xfs_bmbt_rec_host_t *ep;		/* extent record pointer */
-	xfs_extnum_t	lastx;		/* last extent index */
-
-	/*
-	 * Initialize the extent entry structure to catch access to
-	 * uninitialized br_startblock field.
-	 */
-	gotp->br_startoff = 0xffa5a5a5a5a5a5a5LL;
-	gotp->br_blockcount = 0xa55a5a5a5a5a5a5aLL;
-	gotp->br_state = XFS_EXT_INVALID;
-	gotp->br_startblock = 0xffffa5a5a5a5a5a5LL;
-	prevp->br_startoff = NULLFILEOFF;
-
-	ep = xfs_iext_bno_to_ext(ifp, bno, &lastx);
-	if (lastx > 0) {
-		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, lastx - 1), prevp);
-	}
-	if (lastx < (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t))) {
-		xfs_bmbt_get_all(ep, gotp);
-		*eofp = 0;
-	} else {
-		if (lastx > 0) {
-			*gotp = *prevp;
-		}
-		*eofp = 1;
-		ep = NULL;
-	}
-	*lastxp = lastx;
-	return ep;
-}
-
-/*
- * Search the extents list for the inode, for the extent containing bno.
- * If bno lies in a hole, point to the next entry.  If bno lies past eof,
- * *eofp will be set, and *prevp will contain the last entry (null if none).
- * Else, *lastxp will be set to the index of the found
- * entry; *gotp will contain the entry.
- */
-xfs_bmbt_rec_host_t *                 /* pointer to found extent entry */
-xfs_bmap_search_extents(
-	xfs_inode_t     *ip,            /* incore inode pointer */
-	xfs_fileoff_t   bno,            /* block number searched for */
-	int             fork,      	/* data or attr fork */
-	int             *eofp,          /* out: end of file found */
-	xfs_extnum_t    *lastxp,        /* out: last extent index */
-	xfs_bmbt_irec_t *gotp,          /* out: extent entry found */
-	xfs_bmbt_irec_t *prevp)         /* out: previous extent entry found */
-{
-	xfs_ifork_t	*ifp;		/* inode fork pointer */
-	xfs_bmbt_rec_host_t  *ep;            /* extent record pointer */
-
-	XFS_STATS_INC(ip->i_mount, xs_look_exlist);
-	ifp = XFS_IFORK_PTR(ip, fork);
-
-	ep = xfs_bmap_search_multi_extents(ifp, bno, eofp, lastxp, gotp, prevp);
-
-	if (unlikely(!(gotp->br_startblock) && (*lastxp != NULLEXTNUM) &&
-		     !(XFS_IS_REALTIME_INODE(ip) && fork == XFS_DATA_FORK))) {
-		xfs_alert_tag(ip->i_mount, XFS_PTAG_FSBLOCK_ZERO,
-				"Access to block zero in inode %llu "
-				"start_block: %llx start_off: %llx "
-				"blkcnt: %llx extent-state: %x lastx: %x",
-			(unsigned long long)ip->i_ino,
-			(unsigned long long)gotp->br_startblock,
-			(unsigned long long)gotp->br_startoff,
-			(unsigned long long)gotp->br_blockcount,
-			gotp->br_state, *lastxp);
-		*lastxp = NULLEXTNUM;
-		*eofp = 1;
-		return NULL;
-	}
-	return ep;
-}
-
 /*
  * Returns the file-relative block number of the first unused block(s)
  * in the file with at least "len" logically contiguous blocks free.
@@ -1497,7 +1426,7 @@ xfs_bmap_first_unused(
 	    (error = xfs_iread_extents(tp, ip, whichfork)))
 		return error;
 	lowest = *first_unused;
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	nextents = xfs_iext_count(ifp);
 	for (idx = 0, lastaddr = 0, max = lowest; idx < nextents; idx++) {
 		xfs_bmbt_rec_host_t *ep = xfs_iext_get_ext(ifp, idx);
 		off = xfs_bmbt_get_startoff(ep);
@@ -1523,44 +1452,44 @@ xfs_bmap_first_unused(
  */
 int						/* error */
 xfs_bmap_last_before(
-	xfs_trans_t	*tp,			/* transaction pointer */
-	xfs_inode_t	*ip,			/* incore inode */
-	xfs_fileoff_t	*last_block,		/* last block */
-	int		whichfork)		/* data or attr fork */
+	struct xfs_trans	*tp,		/* transaction pointer */
+	struct xfs_inode	*ip,		/* incore inode */
+	xfs_fileoff_t		*last_block,	/* last block */
+	int			whichfork)	/* data or attr fork */
 {
-	xfs_fileoff_t	bno;			/* input file offset */
-	int		eof;			/* hit end of file */
-	xfs_bmbt_rec_host_t *ep;		/* pointer to last extent */
-	int		error;			/* error return value */
-	xfs_bmbt_irec_t	got;			/* current extent value */
-	xfs_ifork_t	*ifp;			/* inode fork pointer */
-	xfs_extnum_t	lastx;			/* last extent used */
-	xfs_bmbt_irec_t	prev;			/* previous extent value */
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, whichfork);
+	struct xfs_bmbt_irec	got;
+	xfs_extnum_t		idx;
+	int			error;
 
-	if (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE &&
-	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS &&
-	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_LOCAL)
-	       return -EIO;
-	if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL) {
+	switch (XFS_IFORK_FORMAT(ip, whichfork)) {
+	case XFS_DINODE_FMT_LOCAL:
 		*last_block = 0;
 		return 0;
+	case XFS_DINODE_FMT_BTREE:
+	case XFS_DINODE_FMT_EXTENTS:
+		break;
+	default:
+		return -EIO;
 	}
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	if (!(ifp->if_flags & XFS_IFEXTENTS) &&
-	    (error = xfs_iread_extents(tp, ip, whichfork)))
-		return error;
-	bno = *last_block - 1;
-	ep = xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got,
-		&prev);
-	if (eof || xfs_bmbt_get_startoff(ep) > bno) {
-		if (prev.br_startoff == NULLFILEOFF)
-			*last_block = 0;
-		else
-			*last_block = prev.br_startoff + prev.br_blockcount;
+
+	if (!(ifp->if_flags & XFS_IFEXTENTS)) {
+		error = xfs_iread_extents(tp, ip, whichfork);
+		if (error)
+			return error;
 	}
-	/*
-	 * Otherwise *last_block is already the right answer.
-	 */
+
+	if (xfs_iext_lookup_extent(ip, ifp, *last_block - 1, &idx, &got)) {
+		if (got.br_startoff <= *last_block - 1)
+			return 0;
+	}
+
+	if (xfs_iext_get_extent(ifp, idx - 1, &got)) {
+		*last_block = got.br_startoff + got.br_blockcount;
+		return 0;
+	}
+
+	*last_block = 0;
 	return 0;
 }
 
@@ -1582,7 +1511,7 @@ xfs_bmap_last_extent(
 			return error;
 	}
 
-	nextents = ifp->if_bytes / sizeof(xfs_bmbt_rec_t);
+	nextents = xfs_iext_count(ifp);
 	if (nextents == 0) {
 		*is_empty = 1;
 		return 0;
@@ -1735,7 +1664,7 @@ xfs_bmap_add_extent_delay_real(
 						&bma->ip->i_d.di_nextents);
 
 	ASSERT(bma->idx >= 0);
-	ASSERT(bma->idx <= ifp->if_bytes / sizeof(struct xfs_bmbt_rec));
+	ASSERT(bma->idx <= xfs_iext_count(ifp));
 	ASSERT(!isnullstartblock(new->br_startblock));
 	ASSERT(!bma->cur ||
 	       (bma->cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL));
@@ -1794,7 +1723,7 @@ xfs_bmap_add_extent_delay_real(
 	 * Don't set contiguous if the combined extent would be too large.
 	 * Also check for all-three-contiguous being too large.
 	 */
-	if (bma->idx < ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t) - 1) {
+	if (bma->idx < xfs_iext_count(ifp) - 1) {
 		state |= BMAP_RIGHT_VALID;
 		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx + 1), &RIGHT);
 
@@ -2300,7 +2229,7 @@ xfs_bmap_add_extent_unwritten_real(
 	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
 
 	ASSERT(*idx >= 0);
-	ASSERT(*idx <= ifp->if_bytes / sizeof(struct xfs_bmbt_rec));
+	ASSERT(*idx <= xfs_iext_count(ifp));
 	ASSERT(!isnullstartblock(new->br_startblock));
 
 	XFS_STATS_INC(mp, xs_add_exlist);
@@ -2356,7 +2285,7 @@ xfs_bmap_add_extent_unwritten_real(
 	 * Don't set contiguous if the combined extent would be too large.
 	 * Also check for all-three-contiguous being too large.
 	 */
-	if (*idx < ip->i_df.if_bytes / (uint)sizeof(xfs_bmbt_rec_t) - 1) {
+	if (*idx < xfs_iext_count(&ip->i_df) - 1) {
 		state |= BMAP_RIGHT_VALID;
 		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *idx + 1), &RIGHT);
 		if (isnullstartblock(RIGHT.br_startblock))
@@ -2836,7 +2765,7 @@ xfs_bmap_add_extent_hole_delay(
 	 * Check and set flags if the current (right) segment exists.
 	 * If it doesn't exist, we're converting the hole at end-of-file.
 	 */
-	if (*idx < ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)) {
+	if (*idx < xfs_iext_count(ifp)) {
 		state |= BMAP_RIGHT_VALID;
 		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *idx), &right);
 
@@ -2966,7 +2895,7 @@ xfs_bmap_add_extent_hole_real(
 	ifp = XFS_IFORK_PTR(bma->ip, whichfork);
 
 	ASSERT(bma->idx >= 0);
-	ASSERT(bma->idx <= ifp->if_bytes / sizeof(struct xfs_bmbt_rec));
+	ASSERT(bma->idx <= xfs_iext_count(ifp));
 	ASSERT(!isnullstartblock(new->br_startblock));
 	ASSERT(!bma->cur ||
 	       !(bma->cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL));
@@ -2992,7 +2921,7 @@ xfs_bmap_add_extent_hole_real(
 	 * Check and set flags if this segment has a current value.
 	 * Not true if we're inserting into the "hole" at eof.
 	 */
-	if (bma->idx < ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)) {
+	if (bma->idx < xfs_iext_count(ifp)) {
 		state |= BMAP_RIGHT_VALID;
 		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx), &right);
 		if (isnullstartblock(right.br_startblock))
@@ -4145,12 +4074,11 @@ xfs_bmapi_read(
 	struct xfs_mount	*mp = ip->i_mount;
 	struct xfs_ifork	*ifp;
 	struct xfs_bmbt_irec	got;
-	struct xfs_bmbt_irec	prev;
 	xfs_fileoff_t		obno;
 	xfs_fileoff_t		end;
-	xfs_extnum_t		lastx;
+	xfs_extnum_t		idx;
 	int			error;
-	int			eof;
+	bool			eof = false;
 	int			n = 0;
 	int			whichfork = xfs_bmapi_whichfork(flags);
 
@@ -4190,7 +4118,8 @@ xfs_bmapi_read(
 			return error;
 	}
 
-	xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got, &prev);
+	if (!xfs_iext_lookup_extent(ip, ifp, bno, &idx, &got))
+		eof = true;
 	end = bno + len;
 	obno = bno;
 
@@ -4221,10 +4150,8 @@ xfs_bmapi_read(
 			break;
 
 		/* Else go on to the next record. */
-		if (++lastx < ifp->if_bytes / sizeof(xfs_bmbt_rec_t))
-			xfs_bmbt_get_all(xfs_iext_get_ext(ifp, lastx), &got);
-		else
-			eof = 1;
+		if (!xfs_iext_get_extent(ifp, ++idx, &got))
+			eof = true;
 	}
 	*nmap = n;
 	return 0;
@@ -4234,10 +4161,10 @@ int
 xfs_bmapi_reserve_delalloc(
 	struct xfs_inode	*ip,
 	int			whichfork,
-	xfs_fileoff_t		aoff,
+	xfs_fileoff_t		off,
 	xfs_filblks_t		len,
+	xfs_filblks_t		prealloc,
 	struct xfs_bmbt_irec	*got,
-	struct xfs_bmbt_irec	*prev,
 	xfs_extnum_t		*lastx,
 	int			eof)
 {
@@ -4248,10 +4175,17 @@ xfs_bmapi_reserve_delalloc(
 	char			rt = XFS_IS_REALTIME_INODE(ip);
 	xfs_extlen_t		extsz;
 	int			error;
+	xfs_fileoff_t		aoff = off;
 
-	alen = XFS_FILBLKS_MIN(len, MAXEXTLEN);
+	/*
+	 * Cap the alloc length. Keep track of prealloc so we know whether to
+	 * tag the inode before we return.
+	 */
+	alen = XFS_FILBLKS_MIN(len + prealloc, MAXEXTLEN);
 	if (!eof)
 		alen = XFS_FILBLKS_MIN(alen, got->br_startoff - aoff);
+	if (prealloc && alen >= len)
+		prealloc = alen - len;
 
 	/* Figure out the extent size, adjust alen */
 	if (whichfork == XFS_COW_FORK)
@@ -4259,7 +4193,12 @@ xfs_bmapi_reserve_delalloc(
 	else
 		extsz = xfs_get_extsz_hint(ip);
 	if (extsz) {
-		error = xfs_bmap_extsize_align(mp, got, prev, extsz, rt, eof,
+		struct xfs_bmbt_irec	prev;
+
+		if (!xfs_iext_get_extent(ifp, *lastx - 1, &prev))
+			prev.br_startoff = NULLFILEOFF;
+
+		error = xfs_bmap_extsize_align(mp, got, &prev, extsz, rt, eof,
 					       1, 0, &aoff, &alen);
 		ASSERT(!error);
 	}
@@ -4312,6 +4251,16 @@ xfs_bmapi_reserve_delalloc(
 	 */
 	xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *lastx), got);
 
+	/*
+	 * Tag the inode if blocks were preallocated. Note that COW fork
+	 * preallocation can occur at the start or end of the extent, even when
+	 * prealloc == 0, so we must also check the aligned offset and length.
+	 */
+	if (whichfork == XFS_DATA_FORK && prealloc)
+		xfs_inode_set_eofblocks_tag(ip);
+	if (whichfork == XFS_COW_FORK && (prealloc || aoff < off || alen > len))
+		xfs_inode_set_cowblocks_tag(ip);
+
 	ASSERT(got->br_startoff <= aoff);
 	ASSERT(got->br_startoff + got->br_blockcount >= aoff + alen);
 	ASSERT(isnullstartblock(got->br_startblock));
@@ -4349,7 +4298,7 @@ xfs_bmapi_allocate(
 	if (bma->wasdel) {
 		bma->length = (xfs_extlen_t)bma->got.br_blockcount;
 		bma->offset = bma->got.br_startoff;
-		if (bma->idx != NULLEXTNUM && bma->idx) {
+		if (bma->idx) {
 			xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx - 1),
 					 &bma->prev);
 		}
@@ -4563,7 +4512,7 @@ xfs_bmapi_write(
 	struct xfs_ifork	*ifp;
 	struct xfs_bmalloca	bma = { NULL };	/* args for xfs_bmap_alloc */
 	xfs_fileoff_t		end;		/* end of mapped file region */
-	int			eof;		/* after the end of extents */
+	bool			eof = false;	/* after the end of extents */
 	int			error;		/* error return */
 	int			n;		/* current extent index */
 	xfs_fileoff_t		obno;		/* old block number (offset) */
@@ -4641,12 +4590,14 @@ xfs_bmapi_write(
 			goto error0;
 	}
 
-	xfs_bmap_search_extents(ip, bno, whichfork, &eof, &bma.idx, &bma.got,
-				&bma.prev);
 	n = 0;
 	end = bno + len;
 	obno = bno;
 
+	if (!xfs_iext_lookup_extent(ip, ifp, bno, &bma.idx, &bma.got))
+		eof = true;
+	if (!xfs_iext_get_extent(ifp, bma.idx - 1, &bma.prev))
+		bma.prev.br_startoff = NULLFILEOFF;
 	bma.tp = tp;
 	bma.ip = ip;
 	bma.total = total;
@@ -4733,11 +4684,8 @@ xfs_bmapi_write(
 
 		/* Else go on to the next record. */
 		bma.prev = bma.got;
-		if (++bma.idx < ifp->if_bytes / sizeof(xfs_bmbt_rec_t)) {
-			xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma.idx),
-					 &bma.got);
-		} else
-			eof = 1;
+		if (!xfs_iext_get_extent(ifp, ++bma.idx, &bma.got))
+			eof = true;
 	}
 	*nmap = n;
 
@@ -4885,7 +4833,7 @@ xfs_bmap_del_extent_delay(
 	da_new = 0;
 
 	ASSERT(*idx >= 0);
-	ASSERT(*idx < ifp->if_bytes / sizeof(struct xfs_bmbt_rec));
+	ASSERT(*idx <= xfs_iext_count(ifp));
 	ASSERT(del->br_blockcount > 0);
 	ASSERT(got->br_startoff <= del->br_startoff);
 	ASSERT(got_endoff >= del_endoff);
@@ -4902,8 +4850,11 @@ xfs_bmap_del_extent_delay(
 	 * sb counters as we might have to borrow some blocks for the
 	 * indirect block accounting.
 	 */
-	xfs_trans_reserve_quota_nblks(NULL, ip, -((long)del->br_blockcount), 0,
+	error = xfs_trans_reserve_quota_nblks(NULL, ip,
+			-((long)del->br_blockcount), 0,
 			isrt ? XFS_QMOPT_RES_RTBLKS : XFS_QMOPT_RES_REGBLKS);
+	if (error)
+		return error;
 	ip->i_delayed_blks -= del->br_blockcount;
 
 	if (whichfork == XFS_COW_FORK)
@@ -5013,7 +4964,7 @@ xfs_bmap_del_extent_cow(
 	got_endoff = got->br_startoff + got->br_blockcount;
 
 	ASSERT(*idx >= 0);
-	ASSERT(*idx < ifp->if_bytes / sizeof(struct xfs_bmbt_rec));
+	ASSERT(*idx <= xfs_iext_count(ifp));
 	ASSERT(del->br_blockcount > 0);
 	ASSERT(got->br_startoff <= del->br_startoff);
 	ASSERT(got_endoff >= del_endoff);
@@ -5119,8 +5070,7 @@ xfs_bmap_del_extent(
 		state |= BMAP_COWFORK;
 
 	ifp = XFS_IFORK_PTR(ip, whichfork);
-	ASSERT((*idx >= 0) && (*idx < ifp->if_bytes /
-		(uint)sizeof(xfs_bmbt_rec_t)));
+	ASSERT((*idx >= 0) && (*idx < xfs_iext_count(ifp)));
 	ASSERT(del->br_blockcount > 0);
 	ep = xfs_iext_get_ext(ifp, *idx);
 	xfs_bmbt_get_all(ep, &got);
@@ -5434,8 +5384,6 @@ __xfs_bunmapi(
 {
 	xfs_btree_cur_t		*cur;		/* bmap btree cursor */
 	xfs_bmbt_irec_t		del;		/* extent being deleted */
-	int			eof;		/* is deleting at eof */
-	xfs_bmbt_rec_host_t	*ep;		/* extent record pointer */
 	int			error;		/* error return value */
 	xfs_extnum_t		extno;		/* extent number in list */
 	xfs_bmbt_irec_t		got;		/* current extent record */
@@ -5445,8 +5393,6 @@ __xfs_bunmapi(
 	int			logflags;	/* transaction logging flags */
 	xfs_extlen_t		mod;		/* rt extent offset */
 	xfs_mount_t		*mp;		/* mount structure */
-	xfs_extnum_t		nextents;	/* number of file extents */
-	xfs_bmbt_irec_t		prev;		/* previous extent record */
 	xfs_fileoff_t		start;		/* first file offset deleted */
 	int			tmp_logflags;	/* partial logging flags */
 	int			wasdel;		/* was a delayed alloc extent */
@@ -5477,8 +5423,7 @@ __xfs_bunmapi(
 	if (!(ifp->if_flags & XFS_IFEXTENTS) &&
 	    (error = xfs_iread_extents(tp, ip, whichfork)))
 		return error;
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-	if (nextents == 0) {
+	if (xfs_iext_count(ifp) == 0) {
 		*rlen = 0;
 		return 0;
 	}
@@ -5486,18 +5431,17 @@ __xfs_bunmapi(
 	isrt = (whichfork == XFS_DATA_FORK) && XFS_IS_REALTIME_INODE(ip);
 	start = bno;
 	bno = start + len - 1;
-	ep = xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got,
-		&prev);
 
 	/*
 	 * Check to see if the given block number is past the end of the
 	 * file, back up to the last block if so...
 	 */
-	if (eof) {
-		ep = xfs_iext_get_ext(ifp, --lastx);
-		xfs_bmbt_get_all(ep, &got);
+	if (!xfs_iext_lookup_extent(ip, ifp, bno, &lastx, &got)) {
+		ASSERT(lastx > 0);
+		xfs_iext_get_extent(ifp, --lastx, &got);
 		bno = got.br_startoff + got.br_blockcount - 1;
 	}
+
 	logflags = 0;
 	if (ifp->if_flags & XFS_IFBROOT) {
 		ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE);
@@ -5528,8 +5472,7 @@ __xfs_bunmapi(
 		if (got.br_startoff > bno) {
 			if (--lastx < 0)
 				break;
-			ep = xfs_iext_get_ext(ifp, lastx);
-			xfs_bmbt_get_all(ep, &got);
+			xfs_iext_get_extent(ifp, lastx, &got);
 		}
 		/*
 		 * Is the last block of this extent before the range
@@ -5543,7 +5486,6 @@ __xfs_bunmapi(
 		 * Then deal with the (possibly delayed) allocated space
 		 * we found.
 		 */
-		ASSERT(ep != NULL);
 		del = got;
 		wasdel = isnullstartblock(del.br_startblock);
 		if (got.br_startoff < start) {
@@ -5624,15 +5566,12 @@ __xfs_bunmapi(
 				 */
 				ASSERT(bno >= del.br_blockcount);
 				bno -= del.br_blockcount;
-				if (got.br_startoff > bno) {
-					if (--lastx >= 0) {
-						ep = xfs_iext_get_ext(ifp,
-								      lastx);
-						xfs_bmbt_get_all(ep, &got);
-					}
-				}
+				if (got.br_startoff > bno && --lastx >= 0)
+					xfs_iext_get_extent(ifp, lastx, &got);
 				continue;
 			} else if (del.br_state == XFS_EXT_UNWRITTEN) {
+				struct xfs_bmbt_irec	prev;
+
 				/*
 				 * This one is already unwritten.
 				 * It must have a written left neighbor.
@@ -5640,8 +5579,7 @@ __xfs_bunmapi(
 				 * try again.
 				 */
 				ASSERT(lastx > 0);
-				xfs_bmbt_get_all(xfs_iext_get_ext(ifp,
-						lastx - 1), &prev);
+				xfs_iext_get_extent(ifp, lastx - 1, &prev);
 				ASSERT(prev.br_state == XFS_EXT_NORM);
 				ASSERT(!isnullstartblock(prev.br_startblock));
 				ASSERT(del.br_startblock ==
@@ -5739,13 +5677,9 @@ nodelete:
 		 */
 		if (bno != (xfs_fileoff_t)-1 && bno >= start) {
 			if (lastx >= 0) {
-				ep = xfs_iext_get_ext(ifp, lastx);
-				if (xfs_bmbt_get_startoff(ep) > bno) {
-					if (--lastx >= 0)
-						ep = xfs_iext_get_ext(ifp,
-								      lastx);
-				}
-				xfs_bmbt_get_all(ep, &got);
+				xfs_iext_get_extent(ifp, lastx, &got);
+				if (got.br_startoff > bno && --lastx >= 0)
+					xfs_iext_get_extent(ifp, lastx, &got);
 			}
 			extno++;
 		}
@@ -5963,7 +5897,7 @@ xfs_bmse_shift_one(
 
 	mp = ip->i_mount;
 	ifp = XFS_IFORK_PTR(ip, whichfork);
-	total_extents = ifp->if_bytes / sizeof(xfs_bmbt_rec_t);
+	total_extents = xfs_iext_count(ifp);
 
 	xfs_bmbt_get_all(gotp, &got);
 
@@ -6140,7 +6074,7 @@ xfs_bmap_shift_extents(
 	 * are collapsing out, so we cannot use the count of real extents here.
 	 * Instead we have to calculate it from the incore fork.
 	 */
-	total_extents = ifp->if_bytes / sizeof(xfs_bmbt_rec_t);
+	total_extents = xfs_iext_count(ifp);
 	if (total_extents == 0) {
 		*done = 1;
 		goto del_cursor;
@@ -6200,7 +6134,7 @@ xfs_bmap_shift_extents(
 		 * count can change. Update the total and grade the next record.
 		 */
 		if (direction == SHIFT_LEFT) {
-			total_extents = ifp->if_bytes / sizeof(xfs_bmbt_rec_t);
+			total_extents = xfs_iext_count(ifp);
 			stop_extent = total_extents;
 		}
 

fs/xfs/libxfs/xfs_bmap.h

@@ -237,14 +237,9 @@ int	xfs_bmap_shift_extents(struct xfs_trans *tp, struct xfs_inode *ip,
 		struct xfs_defer_ops *dfops, enum shift_direction direction,
 		int num_exts);
 int	xfs_bmap_split_extent(struct xfs_inode *ip, xfs_fileoff_t split_offset);
-struct xfs_bmbt_rec_host *
-	xfs_bmap_search_extents(struct xfs_inode *ip, xfs_fileoff_t bno,
-		int fork, int *eofp, xfs_extnum_t *lastxp,
-		struct xfs_bmbt_irec *gotp, struct xfs_bmbt_irec *prevp);
 int	xfs_bmapi_reserve_delalloc(struct xfs_inode *ip, int whichfork,
-		xfs_fileoff_t aoff, xfs_filblks_t len,
-		struct xfs_bmbt_irec *got, struct xfs_bmbt_irec *prev,
-		xfs_extnum_t *lastx, int eof);
+		xfs_fileoff_t off, xfs_filblks_t len, xfs_filblks_t prealloc,
+		struct xfs_bmbt_irec *got, xfs_extnum_t *lastx, int eof);
 
 enum xfs_bmap_intent_type {
 	XFS_BMAP_MAP = 1,

fs/xfs/libxfs/xfs_bmap_btree.c

@@ -796,13 +796,14 @@ xfs_bmbt_init_cursor(
 	struct xfs_btree_cur	*cur;
 	ASSERT(whichfork != XFS_COW_FORK);
 
-	cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_SLEEP);
+	cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_NOFS);
 
 	cur->bc_tp = tp;
 	cur->bc_mp = mp;
 	cur->bc_nlevels = be16_to_cpu(ifp->if_broot->bb_level) + 1;
 	cur->bc_btnum = XFS_BTNUM_BMAP;
 	cur->bc_blocklog = mp->m_sb.sb_blocklog;
+	cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_bmbt_2);
 
 	cur->bc_ops = &xfs_bmbt_ops;
 	cur->bc_flags = XFS_BTREE_LONG_PTRS | XFS_BTREE_ROOT_IN_INODE;

fs/xfs/libxfs/xfs_btree.c

@@ -1769,8 +1769,28 @@ xfs_btree_lookup_get_block(
 	if (error)
 		return error;
 
+	/* Check the inode owner since the verifiers don't. */
+	if (xfs_sb_version_hascrc(&cur->bc_mp->m_sb) &&
+	    (cur->bc_flags & XFS_BTREE_LONG_PTRS) &&
+	    be64_to_cpu((*blkp)->bb_u.l.bb_owner) !=
+			cur->bc_private.b.ip->i_ino)
+		goto out_bad;
+
+	/* Did we get the level we were looking for? */
+	if (be16_to_cpu((*blkp)->bb_level) != level)
+		goto out_bad;
+
+	/* Check that internal nodes have at least one record. */
+	if (level != 0 && be16_to_cpu((*blkp)->bb_numrecs) == 0)
+		goto out_bad;
+
 	xfs_btree_setbuf(cur, level, bp);
 	return 0;
+
+out_bad:
+	*blkp = NULL;
+	xfs_trans_brelse(cur->bc_tp, bp);
+	return -EFSCORRUPTED;
 }
 
 /*

fs/xfs/libxfs/xfs_btree.h

@@ -96,46 +96,10 @@ union xfs_btree_rec {
 /*
  * Generic stats interface
  */
-#define __XFS_BTREE_STATS_INC(mp, type, stat) \
-	XFS_STATS_INC(mp, xs_ ## type ## _2_ ## stat)
 #define XFS_BTREE_STATS_INC(cur, stat)	\
-do {    \
-	struct xfs_mount *__mp = cur->bc_mp; \
-	switch (cur->bc_btnum) {  \
-	case XFS_BTNUM_BNO: __XFS_BTREE_STATS_INC(__mp, abtb, stat); break; \
-	case XFS_BTNUM_CNT: __XFS_BTREE_STATS_INC(__mp, abtc, stat); break; \
-	case XFS_BTNUM_BMAP: __XFS_BTREE_STATS_INC(__mp, bmbt, stat); break; \
-	case XFS_BTNUM_INO: __XFS_BTREE_STATS_INC(__mp, ibt, stat); break; \
-	case XFS_BTNUM_FINO: __XFS_BTREE_STATS_INC(__mp, fibt, stat); break; \
-	case XFS_BTNUM_RMAP: __XFS_BTREE_STATS_INC(__mp, rmap, stat); break; \
-	case XFS_BTNUM_REFC: __XFS_BTREE_STATS_INC(__mp, refcbt, stat); break; \
-	case XFS_BTNUM_MAX: ASSERT(0); /* fucking gcc */ ; break;	\
-	}       \
-} while (0)
-
-#define __XFS_BTREE_STATS_ADD(mp, type, stat, val) \
-	XFS_STATS_ADD(mp, xs_ ## type ## _2_ ## stat, val)
-#define XFS_BTREE_STATS_ADD(cur, stat, val)  \
-do {    \
-	struct xfs_mount *__mp = cur->bc_mp; \
-	switch (cur->bc_btnum) {  \
-	case XFS_BTNUM_BNO:	\
-		__XFS_BTREE_STATS_ADD(__mp, abtb, stat, val); break; \
-	case XFS_BTNUM_CNT:	\
-		__XFS_BTREE_STATS_ADD(__mp, abtc, stat, val); break; \
-	case XFS_BTNUM_BMAP:	\
-		__XFS_BTREE_STATS_ADD(__mp, bmbt, stat, val); break; \
-	case XFS_BTNUM_INO:	\
-		__XFS_BTREE_STATS_ADD(__mp, ibt, stat, val); break; \
-	case XFS_BTNUM_FINO:	\
-		__XFS_BTREE_STATS_ADD(__mp, fibt, stat, val); break; \
-	case XFS_BTNUM_RMAP:	\
-		__XFS_BTREE_STATS_ADD(__mp, rmap, stat, val); break; \
-	case XFS_BTNUM_REFC:	\
-		__XFS_BTREE_STATS_ADD(__mp, refcbt, stat, val); break; \
-	case XFS_BTNUM_MAX: ASSERT(0); /* fucking gcc */ ; break; \
-	}       \
-} while (0)
+	XFS_STATS_INC_OFF((cur)->bc_mp, (cur)->bc_statoff + __XBTS_ ## stat)
+#define XFS_BTREE_STATS_ADD(cur, stat, val)	\
+	XFS_STATS_ADD_OFF((cur)->bc_mp, (cur)->bc_statoff + __XBTS_ ## stat, val)
 
 #define	XFS_BTREE_MAXLEVELS	9	/* max of all btrees */
 
@@ -253,6 +217,7 @@ typedef struct xfs_btree_cur
 	__uint8_t	bc_nlevels;	/* number of levels in the tree */
 	__uint8_t	bc_blocklog;	/* log2(blocksize) of btree blocks */
 	xfs_btnum_t	bc_btnum;	/* identifies which btree type */
+	int		bc_statoff;	/* offset of btre stats array */
 	union {
 		struct {			/* needed for BNO, CNT, INO */
 			struct xfs_buf	*agbp;	/* agf/agi buffer pointer */

fs/xfs/libxfs/xfs_cksum.h

@@ -6,10 +6,11 @@
 /*
  * Calculate the intermediate checksum for a buffer that has the CRC field
  * inside it.  The offset of the 32bit crc fields is passed as the
- * cksum_offset parameter.
+ * cksum_offset parameter. We do not modify the buffer during verification,
+ * hence we have to split the CRC calculation across the cksum_offset.
  */
 static inline __uint32_t
-xfs_start_cksum(char *buffer, size_t length, unsigned long cksum_offset)
+xfs_start_cksum_safe(char *buffer, size_t length, unsigned long cksum_offset)
 {
 	__uint32_t zero = 0;
 	__uint32_t crc;
@@ -25,6 +26,20 @@ xfs_start_cksum(char *buffer, size_t length, unsigned long cksum_offset)
 		      length - (cksum_offset + sizeof(__be32)));
 }
 
+/*
+ * Fast CRC method where the buffer is modified. Callers must have exclusive
+ * access to the buffer while the calculation takes place.
+ */
+static inline __uint32_t
+xfs_start_cksum_update(char *buffer, size_t length, unsigned long cksum_offset)
+{
+	/* zero the CRC field */
+	*(__le32 *)(buffer + cksum_offset) = 0;
+
+	/* single pass CRC calculation for the entire buffer */
+	return crc32c(XFS_CRC_SEED, buffer, length);
+}
+
 /*
  * Convert the intermediate checksum to the final ondisk format.
  *
@@ -40,11 +55,14 @@ xfs_end_cksum(__uint32_t crc)
 
 /*
  * Helper to generate the checksum for a buffer.
+ *
+ * This modifies the buffer temporarily - callers must have exclusive
+ * access to the buffer while the calculation takes place.
  */
 static inline void
 xfs_update_cksum(char *buffer, size_t length, unsigned long cksum_offset)
 {
-	__uint32_t crc = xfs_start_cksum(buffer, length, cksum_offset);
+	__uint32_t crc = xfs_start_cksum_update(buffer, length, cksum_offset);
 
 	*(__le32 *)(buffer + cksum_offset) = xfs_end_cksum(crc);
 }
@@ -55,7 +73,7 @@ xfs_update_cksum(char *buffer, size_t length, unsigned long cksum_offset)
 static inline int
 xfs_verify_cksum(char *buffer, size_t length, unsigned long cksum_offset)
 {
-	__uint32_t crc = xfs_start_cksum(buffer, length, cksum_offset);
+	__uint32_t crc = xfs_start_cksum_safe(buffer, length, cksum_offset);
 
 	return *(__le32 *)(buffer + cksum_offset) == xfs_end_cksum(crc);
 }

fs/xfs/libxfs/xfs_dir2.c

@@ -93,7 +93,7 @@ xfs_ascii_ci_compname(
 	return result;
 }
 
-static struct xfs_nameops xfs_ascii_ci_nameops = {
+static const struct xfs_nameops xfs_ascii_ci_nameops = {
 	.hashname	= xfs_ascii_ci_hashname,
 	.compname	= xfs_ascii_ci_compname,
 };

fs/xfs/libxfs/xfs_dir2.h

@@ -157,6 +157,9 @@ extern int xfs_dir2_isleaf(struct xfs_da_args *args, int *r);
 extern int xfs_dir2_shrink_inode(struct xfs_da_args *args, xfs_dir2_db_t db,
 				struct xfs_buf *bp);
 
+extern void xfs_dir2_data_freescan_int(struct xfs_da_geometry *geo,
+		const struct xfs_dir_ops *ops,
+		struct xfs_dir2_data_hdr *hdr, int *loghead);
 extern void xfs_dir2_data_freescan(struct xfs_inode *dp,
 		struct xfs_dir2_data_hdr *hdr, int *loghead);
 extern void xfs_dir2_data_log_entry(struct xfs_da_args *args,
@@ -177,6 +180,8 @@ extern struct xfs_dir2_data_free *xfs_dir2_data_freefind(
 		struct xfs_dir2_data_hdr *hdr, struct xfs_dir2_data_free *bf,
 		struct xfs_dir2_data_unused *dup);
 
+extern int xfs_dir_ino_validate(struct xfs_mount *mp, xfs_ino_t ino);
+
 extern const struct xfs_buf_ops xfs_dir3_block_buf_ops;
 extern const struct xfs_buf_ops xfs_dir3_leafn_buf_ops;
 extern const struct xfs_buf_ops xfs_dir3_leaf1_buf_ops;

fs/xfs/libxfs/xfs_dir2_data.c

@@ -329,7 +329,7 @@ xfs_dir3_data_read(
 
 	err = xfs_da_read_buf(tp, dp, bno, mapped_bno, bpp,
 				XFS_DATA_FORK, &xfs_dir3_data_buf_ops);
-	if (!err && tp)
+	if (!err && tp && *bpp)
 		xfs_trans_buf_set_type(tp, *bpp, XFS_BLFT_DIR_DATA_BUF);
 	return err;
 }
@@ -505,8 +505,9 @@ xfs_dir2_data_freeremove(
  * Given a data block, reconstruct its bestfree map.
  */
 void
-xfs_dir2_data_freescan(
-	struct xfs_inode	*dp,
+xfs_dir2_data_freescan_int(
+	struct xfs_da_geometry	*geo,
+	const struct xfs_dir_ops *ops,
 	struct xfs_dir2_data_hdr *hdr,
 	int			*loghead)
 {
@@ -516,7 +517,6 @@ xfs_dir2_data_freescan(
 	struct xfs_dir2_data_free *bf;
 	char			*endp;		/* end of block's data */
 	char			*p;		/* current entry pointer */
-	struct xfs_da_geometry	*geo = dp->i_mount->m_dir_geo;
 
 	ASSERT(hdr->magic == cpu_to_be32(XFS_DIR2_DATA_MAGIC) ||
 	       hdr->magic == cpu_to_be32(XFS_DIR3_DATA_MAGIC) ||
@@ -526,13 +526,13 @@ xfs_dir2_data_freescan(
 	/*
 	 * Start by clearing the table.
 	 */
-	bf = dp->d_ops->data_bestfree_p(hdr);
+	bf = ops->data_bestfree_p(hdr);
 	memset(bf, 0, sizeof(*bf) * XFS_DIR2_DATA_FD_COUNT);
 	*loghead = 1;
 	/*
 	 * Set up pointers.
 	 */
-	p = (char *)dp->d_ops->data_entry_p(hdr);
+	p = (char *)ops->data_entry_p(hdr);
 	if (hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC) ||
 	    hdr->magic == cpu_to_be32(XFS_DIR3_BLOCK_MAGIC)) {
 		btp = xfs_dir2_block_tail_p(geo, hdr);
@@ -559,12 +559,22 @@ xfs_dir2_data_freescan(
 		else {
 			dep = (xfs_dir2_data_entry_t *)p;
 			ASSERT((char *)dep - (char *)hdr ==
-			       be16_to_cpu(*dp->d_ops->data_entry_tag_p(dep)));
-			p += dp->d_ops->data_entsize(dep->namelen);
+			       be16_to_cpu(*ops->data_entry_tag_p(dep)));
+			p += ops->data_entsize(dep->namelen);
 		}
 	}
 }
 
+void
+xfs_dir2_data_freescan(
+	struct xfs_inode	*dp,
+	struct xfs_dir2_data_hdr *hdr,
+	int			*loghead)
+{
+	return xfs_dir2_data_freescan_int(dp->i_mount->m_dir_geo, dp->d_ops,
+			hdr, loghead);
+}
+
 /*
  * Initialize a data block at the given block number in the directory.
  * Give back the buffer for the created block.

fs/xfs/libxfs/xfs_dir2_priv.h

@@ -21,7 +21,6 @@
 struct dir_context;
 
 /* xfs_dir2.c */
-extern int xfs_dir_ino_validate(struct xfs_mount *mp, xfs_ino_t ino);
 extern int xfs_dir2_grow_inode(struct xfs_da_args *args, int space,
 				xfs_dir2_db_t *dbp);
 extern int xfs_dir_cilookup_result(struct xfs_da_args *args,

fs/xfs/libxfs/xfs_ialloc.c

@@ -2344,7 +2344,8 @@ xfs_imap(
 
 		imap->im_blkno = XFS_AGB_TO_DADDR(mp, agno, agbno);
 		imap->im_len = XFS_FSB_TO_BB(mp, 1);
-		imap->im_boffset = (ushort)(offset << mp->m_sb.sb_inodelog);
+		imap->im_boffset = (unsigned short)(offset <<
+							mp->m_sb.sb_inodelog);
 		return 0;
 	}
 
@@ -2372,7 +2373,7 @@ out_map:
 
 	imap->im_blkno = XFS_AGB_TO_DADDR(mp, agno, cluster_agbno);
 	imap->im_len = XFS_FSB_TO_BB(mp, blks_per_cluster);
-	imap->im_boffset = (ushort)(offset << mp->m_sb.sb_inodelog);
+	imap->im_boffset = (unsigned short)(offset << mp->m_sb.sb_inodelog);
 
 	/*
 	 * If the inode number maps to a block outside the bounds
@@ -2450,8 +2451,6 @@ xfs_ialloc_log_agi(
 	ASSERT(agi->agi_magicnum == cpu_to_be32(XFS_AGI_MAGIC));
 #endif
 
-	xfs_trans_buf_set_type(tp, bp, XFS_BLFT_AGI_BUF);
-
 	/*
 	 * Compute byte offsets for the first and last fields in the first
 	 * region and log the agi buffer. This only logs up through
@@ -2512,8 +2511,15 @@ xfs_agi_verify(
 	if (!XFS_AGI_GOOD_VERSION(be32_to_cpu(agi->agi_versionnum)))
 		return false;
 
-	if (be32_to_cpu(agi->agi_level) > XFS_BTREE_MAXLEVELS)
+	if (be32_to_cpu(agi->agi_level) < 1 ||
+	    be32_to_cpu(agi->agi_level) > XFS_BTREE_MAXLEVELS)
+		return false;
+
+	if (xfs_sb_version_hasfinobt(&mp->m_sb) &&
+	    (be32_to_cpu(agi->agi_free_level) < 1 ||
+	     be32_to_cpu(agi->agi_free_level) > XFS_BTREE_MAXLEVELS))
 		return false;
+
 	/*
 	 * during growfs operations, the perag is not fully initialised,
 	 * so we can't use it for any useful checking. growfs ensures we can't
@@ -2592,6 +2598,8 @@ xfs_read_agi(
 			XFS_FSS_TO_BB(mp, 1), 0, bpp, &xfs_agi_buf_ops);
 	if (error)
 		return error;
+	if (tp)
+		xfs_trans_buf_set_type(tp, *bpp, XFS_BLFT_AGI_BUF);
 
 	xfs_buf_set_ref(*bpp, XFS_AGI_REF);
 	return 0;

fs/xfs/libxfs/xfs_ialloc_btree.c

@@ -357,7 +357,7 @@ xfs_inobt_init_cursor(
 	struct xfs_agi		*agi = XFS_BUF_TO_AGI(agbp);
 	struct xfs_btree_cur	*cur;
 
-	cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_SLEEP);
+	cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_NOFS);
 
 	cur->bc_tp = tp;
 	cur->bc_mp = mp;
@@ -365,9 +365,11 @@ xfs_inobt_init_cursor(
 	if (btnum == XFS_BTNUM_INO) {
 		cur->bc_nlevels = be32_to_cpu(agi->agi_level);
 		cur->bc_ops = &xfs_inobt_ops;
+		cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_ibt_2);
 	} else {
 		cur->bc_nlevels = be32_to_cpu(agi->agi_free_level);
 		cur->bc_ops = &xfs_finobt_ops;
+		cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_fibt_2);
 	}
 
 	cur->bc_blocklog = mp->m_sb.sb_blocklog;

fs/xfs/libxfs/xfs_inode_buf.c

@@ -383,7 +383,7 @@ xfs_log_dinode_to_disk(
 static bool
 xfs_dinode_verify(
 	struct xfs_mount	*mp,
-	struct xfs_inode	*ip,
+	xfs_ino_t		ino,
 	struct xfs_dinode	*dip)
 {
 	uint16_t		flags;
@@ -392,6 +392,14 @@ xfs_dinode_verify(
 	if (dip->di_magic != cpu_to_be16(XFS_DINODE_MAGIC))
 		return false;
 
+	/* don't allow invalid i_size */
+	if (be64_to_cpu(dip->di_size) & (1ULL << 63))
+		return false;
+
+	/* No zero-length symlinks. */
+	if (S_ISLNK(be16_to_cpu(dip->di_mode)) && dip->di_size == 0)
+		return false;
+
 	/* only version 3 or greater inodes are extensively verified here */
 	if (dip->di_version < 3)
 		return true;
@@ -401,7 +409,7 @@ xfs_dinode_verify(
 	if (!xfs_verify_cksum((char *)dip, mp->m_sb.sb_inodesize,
 			      XFS_DINODE_CRC_OFF))
 		return false;
-	if (be64_to_cpu(dip->di_ino) != ip->i_ino)
+	if (be64_to_cpu(dip->di_ino) != ino)
 		return false;
 	if (!uuid_equal(&dip->di_uuid, &mp->m_sb.sb_meta_uuid))
 		return false;
@@ -436,7 +444,7 @@ xfs_dinode_calc_crc(
 		return;
 
 	ASSERT(xfs_sb_version_hascrc(&mp->m_sb));
-	crc = xfs_start_cksum((char *)dip, mp->m_sb.sb_inodesize,
+	crc = xfs_start_cksum_update((char *)dip, mp->m_sb.sb_inodesize,
 			      XFS_DINODE_CRC_OFF);
 	dip->di_crc = xfs_end_cksum(crc);
 }
@@ -493,7 +501,7 @@ xfs_iread(
 		return error;
 
 	/* even unallocated inodes are verified */
-	if (!xfs_dinode_verify(mp, ip, dip)) {
+	if (!xfs_dinode_verify(mp, ip->i_ino, dip)) {
 		xfs_alert(mp, "%s: validation failed for inode %lld failed",
 				__func__, ip->i_ino);
 

fs/xfs/libxfs/xfs_inode_buf.h

@@ -58,8 +58,8 @@ struct xfs_icdinode {
  */
 struct xfs_imap {
 	xfs_daddr_t	im_blkno;	/* starting BB of inode chunk */
-	ushort		im_len;		/* length in BBs of inode chunk */
-	ushort		im_boffset;	/* inode offset in block in bytes */
+	unsigned short	im_len;		/* length in BBs of inode chunk */
+	unsigned short	im_boffset;	/* inode offset in block in bytes */
 };
 
 int	xfs_imap_to_bp(struct xfs_mount *, struct xfs_trans *,

fs/xfs/libxfs/xfs_inode_fork.c

@@ -775,6 +775,13 @@ xfs_idestroy_fork(
 	}
 }
 
+/* Count number of incore extents based on if_bytes */
+xfs_extnum_t
+xfs_iext_count(struct xfs_ifork *ifp)
+{
+	return ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+}
+
 /*
  * Convert in-core extents to on-disk form
  *
@@ -803,7 +810,7 @@ xfs_iextents_copy(
 	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_ILOCK_SHARED));
 	ASSERT(ifp->if_bytes > 0);
 
-	nrecs = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	nrecs = xfs_iext_count(ifp);
 	XFS_BMAP_TRACE_EXLIST(ip, nrecs, whichfork);
 	ASSERT(nrecs > 0);
 
@@ -941,7 +948,7 @@ xfs_iext_get_ext(
 	xfs_extnum_t	idx)		/* index of target extent */
 {
 	ASSERT(idx >= 0);
-	ASSERT(idx < ifp->if_bytes / sizeof(xfs_bmbt_rec_t));
+	ASSERT(idx < xfs_iext_count(ifp));
 
 	if ((ifp->if_flags & XFS_IFEXTIREC) && (idx == 0)) {
 		return ifp->if_u1.if_ext_irec->er_extbuf;
@@ -1017,7 +1024,7 @@ xfs_iext_add(
 	int		new_size;	/* size of extents after adding */
 	xfs_extnum_t	nextents;	/* number of extents in file */
 
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	nextents = xfs_iext_count(ifp);
 	ASSERT((idx >= 0) && (idx <= nextents));
 	byte_diff = ext_diff * sizeof(xfs_bmbt_rec_t);
 	new_size = ifp->if_bytes + byte_diff;
@@ -1241,7 +1248,7 @@ xfs_iext_remove(
 	trace_xfs_iext_remove(ip, idx, state, _RET_IP_);
 
 	ASSERT(ext_diff > 0);
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	nextents = xfs_iext_count(ifp);
 	new_size = (nextents - ext_diff) * sizeof(xfs_bmbt_rec_t);
 
 	if (new_size == 0) {
@@ -1270,7 +1277,7 @@ xfs_iext_remove_inline(
 
 	ASSERT(!(ifp->if_flags & XFS_IFEXTIREC));
 	ASSERT(idx < XFS_INLINE_EXTS);
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	nextents = xfs_iext_count(ifp);
 	ASSERT(((nextents - ext_diff) > 0) &&
 		(nextents - ext_diff) < XFS_INLINE_EXTS);
 
@@ -1309,7 +1316,7 @@ xfs_iext_remove_direct(
 	ASSERT(!(ifp->if_flags & XFS_IFEXTIREC));
 	new_size = ifp->if_bytes -
 		(ext_diff * sizeof(xfs_bmbt_rec_t));
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	nextents = xfs_iext_count(ifp);
 
 	if (new_size == 0) {
 		xfs_iext_destroy(ifp);
@@ -1546,7 +1553,7 @@ xfs_iext_indirect_to_direct(
 	int		size;		/* size of file extents */
 
 	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	nextents = xfs_iext_count(ifp);
 	ASSERT(nextents <= XFS_LINEAR_EXTS);
 	size = nextents * sizeof(xfs_bmbt_rec_t);
 
@@ -1620,7 +1627,7 @@ xfs_iext_bno_to_ext(
 	xfs_extnum_t	nextents;	/* number of file extents */
 	xfs_fileoff_t	startoff = 0;	/* start offset of extent */
 
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	nextents = xfs_iext_count(ifp);
 	if (nextents == 0) {
 		*idxp = 0;
 		return NULL;
@@ -1733,8 +1740,8 @@ xfs_iext_idx_to_irec(
 
 	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
 	ASSERT(page_idx >= 0);
-	ASSERT(page_idx <= ifp->if_bytes / sizeof(xfs_bmbt_rec_t));
-	ASSERT(page_idx < ifp->if_bytes / sizeof(xfs_bmbt_rec_t) || realloc);
+	ASSERT(page_idx <= xfs_iext_count(ifp));
+	ASSERT(page_idx < xfs_iext_count(ifp) || realloc);
 
 	nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
 	erp_idx = 0;
@@ -1782,7 +1789,7 @@ xfs_iext_irec_init(
 	xfs_extnum_t	nextents;	/* number of extents in file */
 
 	ASSERT(!(ifp->if_flags & XFS_IFEXTIREC));
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	nextents = xfs_iext_count(ifp);
 	ASSERT(nextents <= XFS_LINEAR_EXTS);
 
 	erp = kmem_alloc(sizeof(xfs_ext_irec_t), KM_NOFS);
@@ -1906,7 +1913,7 @@ xfs_iext_irec_compact(
 
 	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
 	nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	nextents = xfs_iext_count(ifp);
 
 	if (nextents == 0) {
 		xfs_iext_destroy(ifp);
@@ -1996,3 +2003,49 @@ xfs_ifork_init_cow(
 	ip->i_cformat = XFS_DINODE_FMT_EXTENTS;
 	ip->i_cnextents = 0;
 }
+
+/*
+ * Lookup the extent covering bno.
+ *
+ * If there is an extent covering bno return the extent index, and store the
+ * expanded extent structure in *gotp, and the extent index in *idx.
+ * If there is no extent covering bno, but there is an extent after it (e.g.
+ * it lies in a hole) return that extent in *gotp and its index in *idx
+ * instead.
+ * If bno is beyond the last extent return false, and return the index after
+ * the last valid index in *idxp.
+ */
+bool
+xfs_iext_lookup_extent(
+	struct xfs_inode	*ip,
+	struct xfs_ifork	*ifp,
+	xfs_fileoff_t		bno,
+	xfs_extnum_t		*idxp,
+	struct xfs_bmbt_irec	*gotp)
+{
+	struct xfs_bmbt_rec_host *ep;
+
+	XFS_STATS_INC(ip->i_mount, xs_look_exlist);
+
+	ep = xfs_iext_bno_to_ext(ifp, bno, idxp);
+	if (!ep)
+		return false;
+	xfs_bmbt_get_all(ep, gotp);
+	return true;
+}
+
+/*
+ * Return true if there is an extent at index idx, and return the expanded
+ * extent structure at idx in that case.  Else return false.
+ */
+bool
+xfs_iext_get_extent(
+	struct xfs_ifork	*ifp,
+	xfs_extnum_t		idx,
+	struct xfs_bmbt_irec	*gotp)
+{
+	if (idx < 0 || idx >= xfs_iext_count(ifp))
+		return false;
+	xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx), gotp);
+	return true;
+}

fs/xfs/libxfs/xfs_inode_fork.h

@@ -152,6 +152,7 @@ void		xfs_init_local_fork(struct xfs_inode *, int, const void *, int);
 
 struct xfs_bmbt_rec_host *
 		xfs_iext_get_ext(struct xfs_ifork *, xfs_extnum_t);
+xfs_extnum_t	xfs_iext_count(struct xfs_ifork *);
 void		xfs_iext_insert(struct xfs_inode *, xfs_extnum_t, xfs_extnum_t,
 				struct xfs_bmbt_irec *, int);
 void		xfs_iext_add(struct xfs_ifork *, xfs_extnum_t, int);
@@ -181,6 +182,12 @@ void		xfs_iext_irec_compact_pages(struct xfs_ifork *);
 void		xfs_iext_irec_compact_full(struct xfs_ifork *);
 void		xfs_iext_irec_update_extoffs(struct xfs_ifork *, int, int);
 
+bool		xfs_iext_lookup_extent(struct xfs_inode *ip,
+			struct xfs_ifork *ifp, xfs_fileoff_t bno,
+			xfs_extnum_t *idxp, struct xfs_bmbt_irec *gotp);
+bool		xfs_iext_get_extent(struct xfs_ifork *ifp, xfs_extnum_t idx,
+			struct xfs_bmbt_irec *gotp);
+
 extern struct kmem_zone	*xfs_ifork_zone;
 
 extern void xfs_ifork_init_cow(struct xfs_inode *ip);

fs/xfs/libxfs/xfs_log_format.h

@@ -481,8 +481,8 @@ static inline uint xfs_log_dinode_size(int version)
 typedef struct xfs_buf_log_format {
 	unsigned short	blf_type;	/* buf log item type indicator */
 	unsigned short	blf_size;	/* size of this item */
-	ushort		blf_flags;	/* misc state */
-	ushort		blf_len;	/* number of blocks in this buf */
+	unsigned short	blf_flags;	/* misc state */
+	unsigned short	blf_len;	/* number of blocks in this buf */
 	__int64_t	blf_blkno;	/* starting blkno of this buf */
 	unsigned int	blf_map_size;	/* used size of data bitmap in words */
 	unsigned int	blf_data_map[XFS_BLF_DATAMAP_SIZE]; /* dirty bitmap */

fs/xfs/libxfs/xfs_log_recover.h

@@ -52,7 +52,7 @@ typedef struct xlog_recover {
 	struct list_head	r_itemq;	/* q for items */
 } xlog_recover_t;
 
-#define ITEM_TYPE(i)	(*(ushort *)(i)->ri_buf[0].i_addr)
+#define ITEM_TYPE(i)	(*(unsigned short *)(i)->ri_buf[0].i_addr)
 
 /*
  * This is the number of entries in the l_buf_cancel_table used during

fs/xfs/libxfs/xfs_refcount_btree.c

@@ -354,6 +354,7 @@ xfs_refcountbt_init_cursor(
 	cur->bc_btnum = XFS_BTNUM_REFC;
 	cur->bc_blocklog = mp->m_sb.sb_blocklog;
 	cur->bc_ops = &xfs_refcountbt_ops;
+	cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_refcbt_2);
 
 	cur->bc_nlevels = be32_to_cpu(agf->agf_refcount_level);
 

fs/xfs/libxfs/xfs_rmap_btree.c

@@ -484,6 +484,7 @@ xfs_rmapbt_init_cursor(
 	cur->bc_blocklog = mp->m_sb.sb_blocklog;
 	cur->bc_ops = &xfs_rmapbt_ops;
 	cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAP]);
+	cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_rmap_2);
 
 	cur->bc_private.a.agbp = agbp;
 	cur->bc_private.a.agno = agno;

fs/xfs/libxfs/xfs_rtbitmap.c

@@ -1016,4 +1016,3 @@ xfs_rtfree_extent(
 	}
 	return 0;
 }
-

fs/xfs/libxfs/xfs_sb.c

@@ -262,6 +262,12 @@ xfs_mount_validate_sb(
 		return -EFSCORRUPTED;
 	}
 
+	if (xfs_sb_version_hascrc(&mp->m_sb) &&
+	    sbp->sb_blocksize < XFS_MIN_CRC_BLOCKSIZE) {
+		xfs_notice(mp, "v5 SB sanity check failed");
+		return -EFSCORRUPTED;
+	}
+
 	/*
 	 * Until this is fixed only page-sized or smaller data blocks work.
 	 */
@@ -338,13 +344,16 @@ xfs_sb_quota_from_disk(struct xfs_sb *sbp)
 					XFS_PQUOTA_CHKD : XFS_GQUOTA_CHKD;
 	sbp->sb_qflags &= ~(XFS_OQUOTA_ENFD | XFS_OQUOTA_CHKD);
 
-	if (sbp->sb_qflags & XFS_PQUOTA_ACCT)  {
+	if (sbp->sb_qflags & XFS_PQUOTA_ACCT &&
+	    sbp->sb_gquotino != NULLFSINO)  {
 		/*
 		 * In older version of superblock, on-disk superblock only
 		 * has sb_gquotino, and in-core superblock has both sb_gquotino
 		 * and sb_pquotino. But, only one of them is supported at any
 		 * point of time. So, if PQUOTA is set in disk superblock,
-		 * copy over sb_gquotino to sb_pquotino.
+		 * copy over sb_gquotino to sb_pquotino.  The NULLFSINO test
+		 * above is to make sure we don't do this twice and wipe them
+		 * both out!
 		 */
 		sbp->sb_pquotino = sbp->sb_gquotino;
 		sbp->sb_gquotino = NULLFSINO;

fs/xfs/libxfs/xfs_types.h

@@ -57,7 +57,6 @@ typedef __int64_t	xfs_sfiloff_t;	/* signed block number in a file */
 
 #define	NULLAGBLOCK	((xfs_agblock_t)-1)
 #define	NULLAGNUMBER	((xfs_agnumber_t)-1)
-#define	NULLEXTNUM	((xfs_extnum_t)-1)
 
 #define NULLCOMMITLSN	((xfs_lsn_t)-1)
 
@@ -75,11 +74,14 @@ typedef __int64_t	xfs_sfiloff_t;	/* signed block number in a file */
  * Minimum and maximum blocksize and sectorsize.
  * The blocksize upper limit is pretty much arbitrary.
  * The sectorsize upper limit is due to sizeof(sb_sectsize).
+ * CRC enable filesystems use 512 byte inodes, meaning 512 byte block sizes
+ * cannot be used.
  */
 #define XFS_MIN_BLOCKSIZE_LOG	9	/* i.e. 512 bytes */
 #define XFS_MAX_BLOCKSIZE_LOG	16	/* i.e. 65536 bytes */
 #define XFS_MIN_BLOCKSIZE	(1 << XFS_MIN_BLOCKSIZE_LOG)
 #define XFS_MAX_BLOCKSIZE	(1 << XFS_MAX_BLOCKSIZE_LOG)
+#define XFS_MIN_CRC_BLOCKSIZE	(1 << (XFS_MIN_BLOCKSIZE_LOG + 1))
 #define XFS_MIN_SECTORSIZE_LOG	9	/* i.e. 512 bytes */
 #define XFS_MAX_SECTORSIZE_LOG	15	/* i.e. 32768 bytes */
 #define XFS_MIN_SECTORSIZE	(1 << XFS_MIN_SECTORSIZE_LOG)

fs/xfs/xfs_aops.c

@@ -37,11 +37,6 @@
 #include <linux/pagevec.h>
 #include <linux/writeback.h>
 
-/* flags for direct write completions */
-#define XFS_DIO_FLAG_UNWRITTEN	(1 << 0)
-#define XFS_DIO_FLAG_APPEND	(1 << 1)
-#define XFS_DIO_FLAG_COW	(1 << 2)
-
 /*
  * structure owned by writepages passed to individual writepage calls
  */
@@ -776,7 +771,7 @@ xfs_map_cow(
 {
 	struct xfs_inode	*ip = XFS_I(inode);
 	struct xfs_bmbt_irec	imap;
-	bool			is_cow = false, need_alloc = false;
+	bool			is_cow = false;
 	int			error;
 
 	/*
@@ -794,7 +789,7 @@ xfs_map_cow(
 	 * Else we need to check if there is a COW mapping at this offset.
 	 */
 	xfs_ilock(ip, XFS_ILOCK_SHARED);
-	is_cow = xfs_reflink_find_cow_mapping(ip, offset, &imap, &need_alloc);
+	is_cow = xfs_reflink_find_cow_mapping(ip, offset, &imap);
 	xfs_iunlock(ip, XFS_ILOCK_SHARED);
 
 	if (!is_cow)
@@ -804,7 +799,7 @@ xfs_map_cow(
 	 * And if the COW mapping has a delayed extent here we need to
 	 * allocate real space for it now.
 	 */
-	if (need_alloc) {
+	if (isnullstartblock(imap.br_startblock)) {
 		error = xfs_iomap_write_allocate(ip, XFS_COW_FORK, offset,
 				&imap);
 		if (error)
@@ -1174,45 +1169,6 @@ xfs_vm_releasepage(
 	return try_to_free_buffers(page);
 }
 
-/*
- * When we map a DIO buffer, we may need to pass flags to
- * xfs_end_io_direct_write to tell it what kind of write IO we are doing.
- *
- * Note that for DIO, an IO to the highest supported file block offset (i.e.
- * 2^63 - 1FSB bytes) will result in the offset + count overflowing a signed 64
- * bit variable. Hence if we see this overflow, we have to assume that the IO is
- * extending the file size. We won't know for sure until IO completion is run
- * and the actual max write offset is communicated to the IO completion
- * routine.
- */
-static void
-xfs_map_direct(
-	struct inode		*inode,
-	struct buffer_head	*bh_result,
-	struct xfs_bmbt_irec	*imap,
-	xfs_off_t		offset,
-	bool			is_cow)
-{
-	uintptr_t		*flags = (uintptr_t *)&bh_result->b_private;
-	xfs_off_t		size = bh_result->b_size;
-
-	trace_xfs_get_blocks_map_direct(XFS_I(inode), offset, size,
-		ISUNWRITTEN(imap) ? XFS_IO_UNWRITTEN : is_cow ? XFS_IO_COW :
-		XFS_IO_OVERWRITE, imap);
-
-	if (ISUNWRITTEN(imap)) {
-		*flags |= XFS_DIO_FLAG_UNWRITTEN;
-		set_buffer_defer_completion(bh_result);
-	} else if (is_cow) {
-		*flags |= XFS_DIO_FLAG_COW;
-		set_buffer_defer_completion(bh_result);
-	}
-	if (offset + size > i_size_read(inode) || offset + size < 0) {
-		*flags |= XFS_DIO_FLAG_APPEND;
-		set_buffer_defer_completion(bh_result);
-	}
-}
-
 /*
  * If this is O_DIRECT or the mpage code calling tell them how large the mapping
  * is, so that we can avoid repeated get_blocks calls.
@@ -1253,51 +1209,12 @@ xfs_map_trim_size(
 	bh_result->b_size = mapping_size;
 }
 
-/* Bounce unaligned directio writes to the page cache. */
 static int
-xfs_bounce_unaligned_dio_write(
-	struct xfs_inode	*ip,
-	xfs_fileoff_t		offset_fsb,
-	struct xfs_bmbt_irec	*imap)
-{
-	struct xfs_bmbt_irec	irec;
-	xfs_fileoff_t		delta;
-	bool			shared;
-	bool			x;
-	int			error;
-
-	irec = *imap;
-	if (offset_fsb > irec.br_startoff) {
-		delta = offset_fsb - irec.br_startoff;
-		irec.br_blockcount -= delta;
-		irec.br_startblock += delta;
-		irec.br_startoff = offset_fsb;
-	}
-	error = xfs_reflink_trim_around_shared(ip, &irec, &shared, &x);
-	if (error)
-		return error;
-
-	/*
-	 * We're here because we're trying to do a directio write to a
-	 * region that isn't aligned to a filesystem block.  If any part
-	 * of the extent is shared, fall back to buffered mode to handle
-	 * the RMW.  This is done by returning -EREMCHG ("remote addr
-	 * changed"), which is caught further up the call stack.
-	 */
-	if (shared) {
-		trace_xfs_reflink_bounce_dio_write(ip, imap);
-		return -EREMCHG;
-	}
-	return 0;
-}
-
-STATIC int
-__xfs_get_blocks(
+xfs_get_blocks(
 	struct inode		*inode,
 	sector_t		iblock,
 	struct buffer_head	*bh_result,
-	int			create,
-	bool			direct)
+	int			create)
 {
 	struct xfs_inode	*ip = XFS_I(inode);
 	struct xfs_mount	*mp = ip->i_mount;
@@ -1308,11 +1225,8 @@ __xfs_get_blocks(
 	int			nimaps = 1;
 	xfs_off_t		offset;
 	ssize_t			size;
-	int			new = 0;
-	bool			is_cow = false;
-	bool			need_alloc = false;
 
-	BUG_ON(create && !direct);
+	BUG_ON(create);
 
 	if (XFS_FORCED_SHUTDOWN(mp))
 		return -EIO;
@@ -1321,7 +1235,7 @@ __xfs_get_blocks(
 	ASSERT(bh_result->b_size >= (1 << inode->i_blkbits));
 	size = bh_result->b_size;
 
-	if (!create && offset >= i_size_read(inode))
+	if (offset >= i_size_read(inode))
 		return 0;
 
 	/*
@@ -1336,52 +1250,12 @@ __xfs_get_blocks(
 	end_fsb = XFS_B_TO_FSB(mp, (xfs_ufsize_t)offset + size);
 	offset_fsb = XFS_B_TO_FSBT(mp, offset);
 
-	if (create && direct && xfs_is_reflink_inode(ip))
-		is_cow = xfs_reflink_find_cow_mapping(ip, offset, &imap,
-					&need_alloc);
-	if (!is_cow) {
-		error = xfs_bmapi_read(ip, offset_fsb, end_fsb - offset_fsb,
-					&imap, &nimaps, XFS_BMAPI_ENTIRE);
-		/*
-		 * Truncate an overwrite extent if there's a pending CoW
-		 * reservation before the end of this extent.  This
-		 * forces us to come back to get_blocks to take care of
-		 * the CoW.
-		 */
-		if (create && direct && nimaps &&
-		    imap.br_startblock != HOLESTARTBLOCK &&
-		    imap.br_startblock != DELAYSTARTBLOCK &&
-		    !ISUNWRITTEN(&imap))
-			xfs_reflink_trim_irec_to_next_cow(ip, offset_fsb,
-					&imap);
-	}
-	ASSERT(!need_alloc);
+	error = xfs_bmapi_read(ip, offset_fsb, end_fsb - offset_fsb,
+				&imap, &nimaps, XFS_BMAPI_ENTIRE);
 	if (error)
 		goto out_unlock;
 
-	/* for DAX, we convert unwritten extents directly */
-	if (create &&
-	    (!nimaps ||
-	     (imap.br_startblock == HOLESTARTBLOCK ||
-	      imap.br_startblock == DELAYSTARTBLOCK) ||
-	     (IS_DAX(inode) && ISUNWRITTEN(&imap)))) {
-		/*
-		 * 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, size,
-					       &imap, nimaps);
-		if (error)
-			return error;
-		new = 1;
-
-		trace_xfs_get_blocks_alloc(ip, offset, size,
-				ISUNWRITTEN(&imap) ? XFS_IO_UNWRITTEN
-						   : XFS_IO_DELALLOC, &imap);
-	} else if (nimaps) {
+	if (nimaps) {
 		trace_xfs_get_blocks_found(ip, offset, size,
 				ISUNWRITTEN(&imap) ? XFS_IO_UNWRITTEN
 						   : XFS_IO_OVERWRITE, &imap);
@@ -1391,12 +1265,6 @@ __xfs_get_blocks(
 		goto out_unlock;
 	}
 
-	if (IS_DAX(inode) && create) {
-		ASSERT(!ISUNWRITTEN(&imap));
-		/* zeroing is not needed at a higher layer */
-		new = 0;
-	}
-
 	/* trim mapping down to size requested */
 	xfs_map_trim_size(inode, iblock, bh_result, &imap, offset, size);
 
@@ -1406,45 +1274,14 @@ __xfs_get_blocks(
 	 */
 	if (imap.br_startblock != HOLESTARTBLOCK &&
 	    imap.br_startblock != DELAYSTARTBLOCK &&
-	    (create || !ISUNWRITTEN(&imap))) {
-		if (create && direct && !is_cow) {
-			error = xfs_bounce_unaligned_dio_write(ip, offset_fsb,
-					&imap);
-			if (error)
-				return error;
-		}
-
+	    !ISUNWRITTEN(&imap))
 		xfs_map_buffer(inode, bh_result, &imap, offset);
-		if (ISUNWRITTEN(&imap))
-			set_buffer_unwritten(bh_result);
-		/* direct IO needs special help */
-		if (create)
-			xfs_map_direct(inode, bh_result, &imap, offset, is_cow);
-	}
 
 	/*
 	 * If this is a realtime file, data may be on a different device.
 	 * to that pointed to from the buffer_head b_bdev currently.
 	 */
 	bh_result->b_bdev = xfs_find_bdev_for_inode(inode);
-
-	/*
-	 * If we previously allocated a block out beyond eof and we are now
-	 * coming back to use it then we will need to flag it as new even if it
-	 * has a disk address.
-	 *
-	 * With sub-block writes into unwritten extents we also need to mark
-	 * the buffer as new so that the unwritten parts of the buffer gets
-	 * correctly zeroed.
-	 */
-	if (create &&
-	    ((!buffer_mapped(bh_result) && !buffer_uptodate(bh_result)) ||
-	     (offset >= i_size_read(inode)) ||
-	     (new || ISUNWRITTEN(&imap))))
-		set_buffer_new(bh_result);
-
-	BUG_ON(direct && imap.br_startblock == DELAYSTARTBLOCK);
-
 	return 0;
 
 out_unlock:
@@ -1452,100 +1289,6 @@ out_unlock:
 	return error;
 }
 
-int
-xfs_get_blocks(
-	struct inode		*inode,
-	sector_t		iblock,
-	struct buffer_head	*bh_result,
-	int			create)
-{
-	return __xfs_get_blocks(inode, iblock, bh_result, create, false);
-}
-
-int
-xfs_get_blocks_direct(
-	struct inode		*inode,
-	sector_t		iblock,
-	struct buffer_head	*bh_result,
-	int			create)
-{
-	return __xfs_get_blocks(inode, iblock, bh_result, create, true);
-}
-
-/*
- * Complete a direct I/O write request.
- *
- * xfs_map_direct passes us some flags in the private data to tell us what to
- * do.  If no flags are set, then the write IO is an overwrite wholly within
- * the existing allocated file size and so there is nothing for us to do.
- *
- * Note that in this case the completion can be called in interrupt context,
- * whereas if we have flags set we will always be called in task context
- * (i.e. from a workqueue).
- */
-int
-xfs_end_io_direct_write(
-	struct kiocb		*iocb,
-	loff_t			offset,
-	ssize_t			size,
-	void			*private)
-{
-	struct inode		*inode = file_inode(iocb->ki_filp);
-	struct xfs_inode	*ip = XFS_I(inode);
-	uintptr_t		flags = (uintptr_t)private;
-	int			error = 0;
-
-	trace_xfs_end_io_direct_write(ip, offset, size);
-
-	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
-		return -EIO;
-
-	if (size <= 0)
-		return size;
-
-	/*
-	 * The flags tell us whether we are doing unwritten extent conversions
-	 * or an append transaction that updates the on-disk file size. These
-	 * cases are the only cases where we should *potentially* be needing
-	 * to update the VFS inode size.
-	 */
-	if (flags == 0) {
-		ASSERT(offset + size <= i_size_read(inode));
-		return 0;
-	}
-
-	/*
-	 * We need to update the in-core inode size here so that we don't end up
-	 * with the on-disk inode size being outside the in-core inode size. We
-	 * have no other method of updating EOF for AIO, so always do it here
-	 * if necessary.
-	 *
-	 * We need to lock the test/set EOF update as we can be racing with
-	 * other IO completions here to update the EOF. Failing to serialise
-	 * here can result in EOF moving backwards and Bad Things Happen when
-	 * that occurs.
-	 */
-	spin_lock(&ip->i_flags_lock);
-	if (offset + size > i_size_read(inode))
-		i_size_write(inode, offset + size);
-	spin_unlock(&ip->i_flags_lock);
-
-	if (flags & XFS_DIO_FLAG_COW)
-		error = xfs_reflink_end_cow(ip, offset, size);
-	if (flags & XFS_DIO_FLAG_UNWRITTEN) {
-		trace_xfs_end_io_direct_write_unwritten(ip, offset, size);
-
-		error = xfs_iomap_write_unwritten(ip, offset, size);
-	}
-	if (flags & XFS_DIO_FLAG_APPEND) {
-		trace_xfs_end_io_direct_write_append(ip, offset, size);
-
-		error = xfs_setfilesize(ip, offset, size);
-	}
-
-	return error;
-}
-
 STATIC ssize_t
 xfs_vm_direct_IO(
 	struct kiocb		*iocb,
@@ -1566,7 +1309,6 @@ xfs_vm_bmap(
 	struct xfs_inode	*ip = XFS_I(inode);
 
 	trace_xfs_vm_bmap(XFS_I(inode));
-	xfs_ilock(ip, XFS_IOLOCK_SHARED);
 
 	/*
 	 * The swap code (ab-)uses ->bmap to get a block mapping and then
@@ -1574,12 +1316,10 @@ xfs_vm_bmap(
 	 * that on reflinks inodes, so we have to skip out here.  And yes,
 	 * 0 is the magic code for a bmap error..
 	 */
-	if (xfs_is_reflink_inode(ip)) {
-		xfs_iunlock(ip, XFS_IOLOCK_SHARED);
+	if (xfs_is_reflink_inode(ip))
 		return 0;
-	}
+
 	filemap_write_and_wait(mapping);
-	xfs_iunlock(ip, XFS_IOLOCK_SHARED);
 	return generic_block_bmap(mapping, block, xfs_get_blocks);
 }
 

fs/xfs/xfs_aops.h

@@ -55,12 +55,6 @@ struct xfs_ioend {
 
 extern const struct address_space_operations xfs_address_space_operations;
 
-int	xfs_get_blocks(struct inode *inode, sector_t offset,
-		       struct buffer_head *map_bh, int create);
-int	xfs_get_blocks_direct(struct inode *inode, sector_t offset,
-			      struct buffer_head *map_bh, int create);
-int	xfs_end_io_direct_write(struct kiocb *iocb, loff_t offset,
-		ssize_t size, void *private);
 int	xfs_setfilesize(struct xfs_inode *ip, xfs_off_t offset, size_t size);
 
 extern void xfs_count_page_state(struct page *, int *, int *);

fs/xfs/xfs_attr.h

@@ -112,8 +112,8 @@ typedef struct attrlist_cursor_kern {
  *========================================================================*/
 
 
-/* Return 0 on success, or -errno; other state communicated via *context */
-typedef int (*put_listent_func_t)(struct xfs_attr_list_context *, int,
+/* void; state communicated via *context */
+typedef void (*put_listent_func_t)(struct xfs_attr_list_context *, int,
 			      unsigned char *, int, int);
 
 typedef struct xfs_attr_list_context {

fs/xfs/xfs_attr_list.c

@@ -74,7 +74,6 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context)
 	xfs_attr_sf_entry_t *sfe;
 	xfs_inode_t *dp;
 	int sbsize, nsbuf, count, i;
-	int error;
 
 	ASSERT(context != NULL);
 	dp = context->dp;
@@ -102,13 +101,11 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context)
 	    (XFS_ISRESET_CURSOR(cursor) &&
              (dp->i_afp->if_bytes + sf->hdr.count * 16) < context->bufsize)) {
 		for (i = 0, sfe = &sf->list[0]; i < sf->hdr.count; i++) {
-			error = context->put_listent(context,
-					   sfe->flags,
-					   sfe->nameval,
-					   (int)sfe->namelen,
-					   (int)sfe->valuelen);
-			if (error)
-				return error;
+			context->put_listent(context,
+					     sfe->flags,
+					     sfe->nameval,
+					     (int)sfe->namelen,
+					     (int)sfe->valuelen);
 			/*
 			 * Either search callback finished early or
 			 * didn't fit it all in the buffer after all.
@@ -193,15 +190,11 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context)
 			cursor->hashval = sbp->hash;
 			cursor->offset = 0;
 		}
-		error = context->put_listent(context,
-					sbp->flags,
-					sbp->name,
-					sbp->namelen,
-					sbp->valuelen);
-		if (error) {
-			kmem_free(sbuf);
-			return error;
-		}
+		context->put_listent(context,
+				     sbp->flags,
+				     sbp->name,
+				     sbp->namelen,
+				     sbp->valuelen);
 		if (context->seen_enough)
 			break;
 		cursor->offset++;
@@ -335,11 +328,7 @@ xfs_attr_node_list(xfs_attr_list_context_t *context)
 	 */
 	for (;;) {
 		leaf = bp->b_addr;
-		error = xfs_attr3_leaf_list_int(bp, context);
-		if (error) {
-			xfs_trans_brelse(NULL, bp);
-			return error;
-		}
+		xfs_attr3_leaf_list_int(bp, context);
 		xfs_attr3_leaf_hdr_from_disk(mp->m_attr_geo, &leafhdr, leaf);
 		if (context->seen_enough || leafhdr.forw == 0)
 			break;
@@ -356,7 +345,7 @@ xfs_attr_node_list(xfs_attr_list_context_t *context)
 /*
  * Copy out attribute list entries for attr_list(), for leaf attribute lists.
  */
-int
+void
 xfs_attr3_leaf_list_int(
 	struct xfs_buf			*bp,
 	struct xfs_attr_list_context	*context)
@@ -366,7 +355,6 @@ xfs_attr3_leaf_list_int(
 	struct xfs_attr3_icleaf_hdr	ichdr;
 	struct xfs_attr_leaf_entry	*entries;
 	struct xfs_attr_leaf_entry	*entry;
-	int				retval;
 	int				i;
 	struct xfs_mount		*mp = context->dp->i_mount;
 
@@ -399,7 +387,7 @@ xfs_attr3_leaf_list_int(
 		}
 		if (i == ichdr.count) {
 			trace_xfs_attr_list_notfound(context);
-			return 0;
+			return;
 		}
 	} else {
 		entry = &entries[0];
@@ -410,7 +398,6 @@ xfs_attr3_leaf_list_int(
 	/*
 	 * We have found our place, start copying out the new attributes.
 	 */
-	retval = 0;
 	for (; i < ichdr.count; entry++, i++) {
 		char *name;
 		int namelen, valuelen;
@@ -439,16 +426,14 @@ xfs_attr3_leaf_list_int(
 			valuelen = be32_to_cpu(name_rmt->valuelen);
 		}
 
-		retval = context->put_listent(context, entry->flags,
+		context->put_listent(context, entry->flags,
 					      name, namelen, valuelen);
-		if (retval)
-			break;
 		if (context->seen_enough)
 			break;
 		cursor->offset++;
 	}
 	trace_xfs_attr_list_leaf_end(context);
-	return retval;
+	return;
 }
 
 /*
@@ -467,9 +452,9 @@ xfs_attr_leaf_list(xfs_attr_list_context_t *context)
 	if (error)
 		return error;
 
-	error = xfs_attr3_leaf_list_int(bp, context);
+	xfs_attr3_leaf_list_int(bp, context);
 	xfs_trans_brelse(NULL, bp);
-	return error;
+	return 0;
 }
 
 int
@@ -513,7 +498,7 @@ xfs_attr_list_int(
  * Take care to check values and protect against them changing later,
  * we may be reading them directly out of a user buffer.
  */
-STATIC int
+STATIC void
 xfs_attr_put_listent(
 	xfs_attr_list_context_t *context,
 	int		flags,
@@ -536,10 +521,10 @@ xfs_attr_put_listent(
 	 */
 	if (((context->flags & ATTR_SECURE) == 0) !=
 	    ((flags & XFS_ATTR_SECURE) == 0))
-		return 0;
+		return;
 	if (((context->flags & ATTR_ROOT) == 0) !=
 	    ((flags & XFS_ATTR_ROOT) == 0))
-		return 0;
+		return;
 
 	arraytop = sizeof(*alist) +
 			context->count * sizeof(alist->al_offset[0]);
@@ -548,7 +533,7 @@ xfs_attr_put_listent(
 		trace_xfs_attr_list_full(context);
 		alist->al_more = 1;
 		context->seen_enough = 1;
-		return 0;
+		return;
 	}
 
 	aep = (attrlist_ent_t *)&context->alist[context->firstu];
@@ -558,7 +543,7 @@ xfs_attr_put_listent(
 	alist->al_offset[context->count++] = context->firstu;
 	alist->al_count = context->count;
 	trace_xfs_attr_list_add(context);
-	return 0;
+	return;
 }
 
 /*

fs/xfs/xfs_bmap_util.c

@@ -359,9 +359,7 @@ xfs_bmap_count_blocks(
 	mp = ip->i_mount;
 	ifp = XFS_IFORK_PTR(ip, whichfork);
 	if ( XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS ) {
-		xfs_bmap_count_leaves(ifp, 0,
-			ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t),
-			count);
+		xfs_bmap_count_leaves(ifp, 0, xfs_iext_count(ifp), count);
 		return 0;
 	}
 
@@ -426,7 +424,7 @@ xfs_getbmapx_fix_eof_hole(
 		ifp = XFS_IFORK_PTR(ip, whichfork);
 		if (!moretocome &&
 		    xfs_iext_bno_to_ext(ifp, fileblock, &lastx) &&
-		   (lastx == (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t))-1))
+		   (lastx == xfs_iext_count(ifp) - 1))
 			out->bmv_oflags |= BMV_OF_LAST;
 	}
 
@@ -1792,6 +1790,7 @@ xfs_swap_extent_forks(
 	struct xfs_ifork	tempifp, *ifp, *tifp;
 	int			aforkblks = 0;
 	int			taforkblks = 0;
+	xfs_extnum_t		nextents;
 	__uint64_t		tmp;
 	int			error;
 
@@ -1877,14 +1876,13 @@ xfs_swap_extent_forks(
 
 	switch (ip->i_d.di_format) {
 	case XFS_DINODE_FMT_EXTENTS:
-		/* If the extents fit in the inode, fix the
-		 * pointer.  Otherwise it's already NULL or
-		 * pointing to the extent.
+		/*
+		 * If the extents fit in the inode, fix the pointer.  Otherwise
+		 * it's already NULL or pointing to the extent.
 		 */
-		if (ip->i_d.di_nextents <= XFS_INLINE_EXTS) {
-			ifp->if_u1.if_extents =
-				ifp->if_u2.if_inline_ext;
-		}
+		nextents = xfs_iext_count(&ip->i_df);
+		if (nextents <= XFS_INLINE_EXTS)
+			ifp->if_u1.if_extents = ifp->if_u2.if_inline_ext;
 		(*src_log_flags) |= XFS_ILOG_DEXT;
 		break;
 	case XFS_DINODE_FMT_BTREE:
@@ -1896,14 +1894,13 @@ xfs_swap_extent_forks(
 
 	switch (tip->i_d.di_format) {
 	case XFS_DINODE_FMT_EXTENTS:
-		/* If the extents fit in the inode, fix the
-		 * pointer.  Otherwise it's already NULL or
-		 * pointing to the extent.
+		/*
+		 * If the extents fit in the inode, fix the pointer.  Otherwise
+		 * it's already NULL or pointing to the extent.
 		 */
-		if (tip->i_d.di_nextents <= XFS_INLINE_EXTS) {
-			tifp->if_u1.if_extents =
-				tifp->if_u2.if_inline_ext;
-		}
+		nextents = xfs_iext_count(&tip->i_df);
+		if (nextents <= XFS_INLINE_EXTS)
+			tifp->if_u1.if_extents = tifp->if_u2.if_inline_ext;
 		(*target_log_flags) |= XFS_ILOG_DEXT;
 		break;
 	case XFS_DINODE_FMT_BTREE:
@@ -1938,8 +1935,8 @@ xfs_swap_extents(
 	 * page cache safely. Once we have done this we can take the ilocks and
 	 * do the rest of the checks.
 	 */
-	lock_flags = XFS_IOLOCK_EXCL | XFS_MMAPLOCK_EXCL;
-	xfs_lock_two_inodes(ip, tip, XFS_IOLOCK_EXCL);
+	lock_two_nondirectories(VFS_I(ip), VFS_I(tip));
+	lock_flags = XFS_MMAPLOCK_EXCL;
 	xfs_lock_two_inodes(ip, tip, XFS_MMAPLOCK_EXCL);
 
 	/* Verify that both files have the same format */
@@ -2079,15 +2076,13 @@ xfs_swap_extents(
 	trace_xfs_swap_extent_after(ip, 0);
 	trace_xfs_swap_extent_after(tip, 1);
 
+out_unlock:
 	xfs_iunlock(ip, lock_flags);
 	xfs_iunlock(tip, lock_flags);
+	unlock_two_nondirectories(VFS_I(ip), VFS_I(tip));
 	return error;
 
 out_trans_cancel:
 	xfs_trans_cancel(tp);
-
-out_unlock:
-	xfs_iunlock(ip, lock_flags);
-	xfs_iunlock(tip, lock_flags);
-	return error;
+	goto out_unlock;
 }

fs/xfs/xfs_buf.c

@@ -219,7 +219,6 @@ _xfs_buf_alloc(
 	init_completion(&bp->b_iowait);
 	INIT_LIST_HEAD(&bp->b_lru);
 	INIT_LIST_HEAD(&bp->b_list);
-	RB_CLEAR_NODE(&bp->b_rbnode);
 	sema_init(&bp->b_sema, 0); /* held, no waiters */
 	spin_lock_init(&bp->b_lock);
 	XB_SET_OWNER(bp);
@@ -473,6 +472,62 @@ _xfs_buf_map_pages(
 /*
  *	Finding and Reading Buffers
  */
+static int
+_xfs_buf_obj_cmp(
+	struct rhashtable_compare_arg	*arg,
+	const void			*obj)
+{
+	const struct xfs_buf_map	*map = arg->key;
+	const struct xfs_buf		*bp = obj;
+
+	/*
+	 * The key hashing in the lookup path depends on the key being the
+	 * first element of the compare_arg, make sure to assert this.
+	 */
+	BUILD_BUG_ON(offsetof(struct xfs_buf_map, bm_bn) != 0);
+
+	if (bp->b_bn != map->bm_bn)
+		return 1;
+
+	if (unlikely(bp->b_length != map->bm_len)) {
+		/*
+		 * found a block number match. If the range doesn't
+		 * match, the only way this is allowed is if the buffer
+		 * in the cache is stale and the transaction that made
+		 * it stale has not yet committed. i.e. we are
+		 * reallocating a busy extent. Skip this buffer and
+		 * continue searching for an exact match.
+		 */
+		ASSERT(bp->b_flags & XBF_STALE);
+		return 1;
+	}
+	return 0;
+}
+
+static const struct rhashtable_params xfs_buf_hash_params = {
+	.min_size		= 32,	/* empty AGs have minimal footprint */
+	.nelem_hint		= 16,
+	.key_len		= sizeof(xfs_daddr_t),
+	.key_offset		= offsetof(struct xfs_buf, b_bn),
+	.head_offset		= offsetof(struct xfs_buf, b_rhash_head),
+	.automatic_shrinking	= true,
+	.obj_cmpfn		= _xfs_buf_obj_cmp,
+};
+
+int
+xfs_buf_hash_init(
+	struct xfs_perag	*pag)
+{
+	spin_lock_init(&pag->pag_buf_lock);
+	return rhashtable_init(&pag->pag_buf_hash, &xfs_buf_hash_params);
+}
+
+void
+xfs_buf_hash_destroy(
+	struct xfs_perag	*pag)
+{
+	rhashtable_destroy(&pag->pag_buf_hash);
+}
 
 /*
  *	Look up, and creates if absent, a lockable buffer for
@@ -488,27 +543,24 @@ _xfs_buf_find(
 	xfs_buf_t		*new_bp)
 {
 	struct xfs_perag	*pag;
-	struct rb_node		**rbp;
-	struct rb_node		*parent;
 	xfs_buf_t		*bp;
-	xfs_daddr_t		blkno = map[0].bm_bn;
+	struct xfs_buf_map	cmap = { .bm_bn = map[0].bm_bn };
 	xfs_daddr_t		eofs;
-	int			numblks = 0;
 	int			i;
 
 	for (i = 0; i < nmaps; i++)
-		numblks += map[i].bm_len;
+		cmap.bm_len += map[i].bm_len;
 
 	/* Check for IOs smaller than the sector size / not sector aligned */
-	ASSERT(!(BBTOB(numblks) < btp->bt_meta_sectorsize));
-	ASSERT(!(BBTOB(blkno) & (xfs_off_t)btp->bt_meta_sectormask));
+	ASSERT(!(BBTOB(cmap.bm_len) < btp->bt_meta_sectorsize));
+	ASSERT(!(BBTOB(cmap.bm_bn) & (xfs_off_t)btp->bt_meta_sectormask));
 
 	/*
 	 * Corrupted block numbers can get through to here, unfortunately, so we
 	 * have to check that the buffer falls within the filesystem bounds.
 	 */
 	eofs = XFS_FSB_TO_BB(btp->bt_mount, btp->bt_mount->m_sb.sb_dblocks);
-	if (blkno < 0 || blkno >= eofs) {
+	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
@@ -516,53 +568,29 @@ _xfs_buf_find(
 		 */
 		xfs_alert(btp->bt_mount,
 			  "%s: Block out of range: block 0x%llx, EOFS 0x%llx ",
-			  __func__, blkno, eofs);
+			  __func__, cmap.bm_bn, eofs);
 		WARN_ON(1);
 		return NULL;
 	}
 
-	/* get tree root */
 	pag = xfs_perag_get(btp->bt_mount,
-				xfs_daddr_to_agno(btp->bt_mount, blkno));
+			    xfs_daddr_to_agno(btp->bt_mount, cmap.bm_bn));
 
-	/* walk tree */
 	spin_lock(&pag->pag_buf_lock);
-	rbp = &pag->pag_buf_tree.rb_node;
-	parent = NULL;
-	bp = NULL;
-	while (*rbp) {
-		parent = *rbp;
-		bp = rb_entry(parent, struct xfs_buf, b_rbnode);
-
-		if (blkno < bp->b_bn)
-			rbp = &(*rbp)->rb_left;
-		else if (blkno > bp->b_bn)
-			rbp = &(*rbp)->rb_right;
-		else {
-			/*
-			 * found a block number match. If the range doesn't
-			 * match, the only way this is allowed is if the buffer
-			 * in the cache is stale and the transaction that made
-			 * it stale has not yet committed. i.e. we are
-			 * reallocating a busy extent. Skip this buffer and
-			 * continue searching to the right for an exact match.
-			 */
-			if (bp->b_length != numblks) {
-				ASSERT(bp->b_flags & XBF_STALE);
-				rbp = &(*rbp)->rb_right;
-				continue;
-			}
-			atomic_inc(&bp->b_hold);
-			goto found;
-		}
+	bp = rhashtable_lookup_fast(&pag->pag_buf_hash, &cmap,
+				    xfs_buf_hash_params);
+	if (bp) {
+		atomic_inc(&bp->b_hold);
+		goto found;
 	}
 
 	/* No match found */
 	if (new_bp) {
-		rb_link_node(&new_bp->b_rbnode, parent, rbp);
-		rb_insert_color(&new_bp->b_rbnode, &pag->pag_buf_tree);
 		/* 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 {
 		XFS_STATS_INC(btp->bt_mount, xb_miss_locked);
@@ -930,7 +958,6 @@ xfs_buf_rele(
 
 	if (!pag) {
 		ASSERT(list_empty(&bp->b_lru));
-		ASSERT(RB_EMPTY_NODE(&bp->b_rbnode));
 		if (atomic_dec_and_test(&bp->b_hold)) {
 			xfs_buf_ioacct_dec(bp);
 			xfs_buf_free(bp);
@@ -938,8 +965,6 @@ xfs_buf_rele(
 		return;
 	}
 
-	ASSERT(!RB_EMPTY_NODE(&bp->b_rbnode));
-
 	ASSERT(atomic_read(&bp->b_hold) > 0);
 
 	release = atomic_dec_and_lock(&bp->b_hold, &pag->pag_buf_lock);
@@ -983,7 +1008,8 @@ xfs_buf_rele(
 		}
 
 		ASSERT(!(bp->b_flags & _XBF_DELWRI_Q));
-		rb_erase(&bp->b_rbnode, &pag->pag_buf_tree);
+		rhashtable_remove_fast(&pag->pag_buf_hash, &bp->b_rhash_head,
+				       xfs_buf_hash_params);
 		spin_unlock(&pag->pag_buf_lock);
 		xfs_perag_put(pag);
 		freebuf = true;
@@ -1711,8 +1737,7 @@ xfs_free_buftarg(
 	percpu_counter_destroy(&btp->bt_io_count);
 	list_lru_destroy(&btp->bt_lru);
 
-	if (mp->m_flags & XFS_MOUNT_BARRIER)
-		xfs_blkdev_issue_flush(btp);
+	xfs_blkdev_issue_flush(btp);
 
 	kmem_free(btp);
 }

fs/xfs/xfs_buf.h

@@ -71,6 +71,7 @@ typedef unsigned int xfs_buf_flags_t;
 	{ XBF_READ,		"READ" }, \
 	{ XBF_WRITE,		"WRITE" }, \
 	{ XBF_READ_AHEAD,	"READ_AHEAD" }, \
+	{ XBF_NO_IOACCT,	"NO_IOACCT" }, \
 	{ XBF_ASYNC,		"ASYNC" }, \
 	{ XBF_DONE,		"DONE" }, \
 	{ XBF_STALE,		"STALE" }, \
@@ -150,7 +151,7 @@ typedef struct xfs_buf {
 	 * which is the only bit that is touched if we hit the semaphore
 	 * fast-path on locking.
 	 */
-	struct rb_node		b_rbnode;	/* rbtree node */
+	struct rhash_head	b_rhash_head;	/* pag buffer hash node */
 	xfs_daddr_t		b_bn;		/* block number of buffer */
 	int			b_length;	/* size of buffer in BBs */
 	atomic_t		b_hold;		/* reference count */

fs/xfs/xfs_dir2_readdir.c

@@ -677,7 +677,6 @@ xfs_readdir(
 	args.dp = dp;
 	args.geo = dp->i_mount->m_dir_geo;
 
-	xfs_ilock(dp, XFS_IOLOCK_SHARED);
 	if (dp->i_d.di_format == XFS_DINODE_FMT_LOCAL)
 		rval = xfs_dir2_sf_getdents(&args, ctx);
 	else if ((rval = xfs_dir2_isblock(&args, &v)))
@@ -686,7 +685,6 @@ xfs_readdir(
 		rval = xfs_dir2_block_getdents(&args, ctx);
 	else
 		rval = xfs_dir2_leaf_getdents(&args, ctx, bufsize);
-	xfs_iunlock(dp, XFS_IOLOCK_SHARED);
 
 	return rval;
 }

fs/xfs/xfs_file.c

@@ -47,40 +47,6 @@
 
 static const struct vm_operations_struct xfs_file_vm_ops;
 
-/*
- * Locking primitives for read and write IO paths to ensure we consistently use
- * and order the inode->i_mutex, ip->i_lock and ip->i_iolock.
- */
-static inline void
-xfs_rw_ilock(
-	struct xfs_inode	*ip,
-	int			type)
-{
-	if (type & XFS_IOLOCK_EXCL)
-		inode_lock(VFS_I(ip));
-	xfs_ilock(ip, type);
-}
-
-static inline void
-xfs_rw_iunlock(
-	struct xfs_inode	*ip,
-	int			type)
-{
-	xfs_iunlock(ip, type);
-	if (type & XFS_IOLOCK_EXCL)
-		inode_unlock(VFS_I(ip));
-}
-
-static inline void
-xfs_rw_ilock_demote(
-	struct xfs_inode	*ip,
-	int			type)
-{
-	xfs_ilock_demote(ip, type);
-	if (type & XFS_IOLOCK_EXCL)
-		inode_unlock(VFS_I(ip));
-}
-
 /*
  * Clear the specified ranges to zero through either the pagecache or DAX.
  * Holes and unwritten extents will be left as-is as they already are zeroed.
@@ -183,19 +149,16 @@ xfs_file_fsync(
 
 	xfs_iflags_clear(ip, XFS_ITRUNCATED);
 
-	if (mp->m_flags & XFS_MOUNT_BARRIER) {
-		/*
-		 * If we have an RT and/or log subvolume we need to make sure
-		 * to flush the write cache the device used for file data
-		 * first.  This is to ensure newly written file data make
-		 * it to disk before logging the new inode size in case of
-		 * an extending write.
-		 */
-		if (XFS_IS_REALTIME_INODE(ip))
-			xfs_blkdev_issue_flush(mp->m_rtdev_targp);
-		else if (mp->m_logdev_targp != mp->m_ddev_targp)
-			xfs_blkdev_issue_flush(mp->m_ddev_targp);
-	}
+	/*
+	 * If we have an RT and/or log subvolume we need to make sure to flush
+	 * the write cache the device used for file data first.  This is to
+	 * ensure newly written file data make it to disk before logging the new
+	 * inode size in case of an extending write.
+	 */
+	if (XFS_IS_REALTIME_INODE(ip))
+		xfs_blkdev_issue_flush(mp->m_rtdev_targp);
+	else if (mp->m_logdev_targp != mp->m_ddev_targp)
+		xfs_blkdev_issue_flush(mp->m_ddev_targp);
 
 	/*
 	 * All metadata updates are logged, which means that we just have to
@@ -230,10 +193,8 @@ xfs_file_fsync(
 	 * an already allocated file and thus do not have any metadata to
 	 * commit.
 	 */
-	if ((mp->m_flags & XFS_MOUNT_BARRIER) &&
-	    mp->m_logdev_targp == mp->m_ddev_targp &&
-	    !XFS_IS_REALTIME_INODE(ip) &&
-	    !log_flushed)
+	if (!log_flushed && !XFS_IS_REALTIME_INODE(ip) &&
+	    mp->m_logdev_targp == mp->m_ddev_targp)
 		xfs_blkdev_issue_flush(mp->m_ddev_targp);
 
 	return error;
@@ -244,62 +205,21 @@ xfs_file_dio_aio_read(
 	struct kiocb		*iocb,
 	struct iov_iter		*to)
 {
-	struct address_space	*mapping = iocb->ki_filp->f_mapping;
-	struct inode		*inode = mapping->host;
-	struct xfs_inode	*ip = XFS_I(inode);
-	loff_t			isize = i_size_read(inode);
+	struct xfs_inode	*ip = XFS_I(file_inode(iocb->ki_filp));
 	size_t			count = iov_iter_count(to);
-	loff_t			end = iocb->ki_pos + count - 1;
-	struct iov_iter		data;
-	struct xfs_buftarg	*target;
-	ssize_t			ret = 0;
+	ssize_t			ret;
 
 	trace_xfs_file_direct_read(ip, count, iocb->ki_pos);
 
 	if (!count)
 		return 0; /* skip atime */
 
-	if (XFS_IS_REALTIME_INODE(ip))
-		target = ip->i_mount->m_rtdev_targp;
-	else
-		target = ip->i_mount->m_ddev_targp;
-
-	/* DIO must be aligned to device logical sector size */
-	if ((iocb->ki_pos | count) & target->bt_logical_sectormask) {
-		if (iocb->ki_pos == isize)
-			return 0;
-		return -EINVAL;
-	}
-
 	file_accessed(iocb->ki_filp);
 
-	xfs_rw_ilock(ip, XFS_IOLOCK_SHARED);
-	if (mapping->nrpages) {
-		ret = filemap_write_and_wait_range(mapping, iocb->ki_pos, end);
-		if (ret)
-			goto out_unlock;
-
-		/*
-		 * Invalidate whole pages. This can return an error if we fail
-		 * to invalidate a page, but this should never happen on XFS.
-		 * Warn if it does fail.
-		 */
-		ret = invalidate_inode_pages2_range(mapping,
-				iocb->ki_pos >> PAGE_SHIFT, end >> PAGE_SHIFT);
-		WARN_ON_ONCE(ret);
-		ret = 0;
-	}
+	xfs_ilock(ip, XFS_IOLOCK_SHARED);
+	ret = iomap_dio_rw(iocb, to, &xfs_iomap_ops, NULL);
+	xfs_iunlock(ip, XFS_IOLOCK_SHARED);
 
-	data = *to;
-	ret = __blockdev_direct_IO(iocb, inode, target->bt_bdev, &data,
-			xfs_get_blocks_direct, NULL, NULL, 0);
-	if (ret >= 0) {
-		iocb->ki_pos += ret;
-		iov_iter_advance(to, ret);
-	}
-
-out_unlock:
-	xfs_rw_iunlock(ip, XFS_IOLOCK_SHARED);
 	return ret;
 }
 
@@ -317,9 +237,9 @@ xfs_file_dax_read(
 	if (!count)
 		return 0; /* skip atime */
 
-	xfs_rw_ilock(ip, XFS_IOLOCK_SHARED);
+	xfs_ilock(ip, XFS_IOLOCK_SHARED);
 	ret = dax_iomap_rw(iocb, to, &xfs_iomap_ops);
-	xfs_rw_iunlock(ip, XFS_IOLOCK_SHARED);
+	xfs_iunlock(ip, XFS_IOLOCK_SHARED);
 
 	file_accessed(iocb->ki_filp);
 	return ret;
@@ -335,9 +255,9 @@ xfs_file_buffered_aio_read(
 
 	trace_xfs_file_buffered_read(ip, iov_iter_count(to), iocb->ki_pos);
 
-	xfs_rw_ilock(ip, XFS_IOLOCK_SHARED);
+	xfs_ilock(ip, XFS_IOLOCK_SHARED);
 	ret = generic_file_read_iter(iocb, to);
-	xfs_rw_iunlock(ip, XFS_IOLOCK_SHARED);
+	xfs_iunlock(ip, XFS_IOLOCK_SHARED);
 
 	return ret;
 }
@@ -418,15 +338,18 @@ restart:
 	if (error <= 0)
 		return error;
 
-	error = xfs_break_layouts(inode, iolock, true);
+	error = xfs_break_layouts(inode, iolock);
 	if (error)
 		return error;
 
-	/* For changing security info in file_remove_privs() we need i_mutex */
+	/*
+	 * For changing security info in file_remove_privs() we need i_rwsem
+	 * exclusively.
+	 */
 	if (*iolock == XFS_IOLOCK_SHARED && !IS_NOSEC(inode)) {
-		xfs_rw_iunlock(ip, *iolock);
+		xfs_iunlock(ip, *iolock);
 		*iolock = XFS_IOLOCK_EXCL;
-		xfs_rw_ilock(ip, *iolock);
+		xfs_ilock(ip, *iolock);
 		goto restart;
 	}
 	/*
@@ -451,9 +374,9 @@ restart:
 		spin_unlock(&ip->i_flags_lock);
 		if (!drained_dio) {
 			if (*iolock == XFS_IOLOCK_SHARED) {
-				xfs_rw_iunlock(ip, *iolock);
+				xfs_iunlock(ip, *iolock);
 				*iolock = XFS_IOLOCK_EXCL;
-				xfs_rw_ilock(ip, *iolock);
+				xfs_ilock(ip, *iolock);
 				iov_iter_reexpand(from, count);
 			}
 			/*
@@ -496,6 +419,58 @@ restart:
 	return 0;
 }
 
+static int
+xfs_dio_write_end_io(
+	struct kiocb		*iocb,
+	ssize_t			size,
+	unsigned		flags)
+{
+	struct inode		*inode = file_inode(iocb->ki_filp);
+	struct xfs_inode	*ip = XFS_I(inode);
+	loff_t			offset = iocb->ki_pos;
+	bool			update_size = false;
+	int			error = 0;
+
+	trace_xfs_end_io_direct_write(ip, offset, size);
+
+	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
+		return -EIO;
+
+	if (size <= 0)
+		return size;
+
+	/*
+	 * We need to update the in-core inode size here so that we don't end up
+	 * with the on-disk inode size being outside the in-core inode size. We
+	 * have no other method of updating EOF for AIO, so always do it here
+	 * if necessary.
+	 *
+	 * We need to lock the test/set EOF update as we can be racing with
+	 * other IO completions here to update the EOF. Failing to serialise
+	 * here can result in EOF moving backwards and Bad Things Happen when
+	 * that occurs.
+	 */
+	spin_lock(&ip->i_flags_lock);
+	if (offset + size > i_size_read(inode)) {
+		i_size_write(inode, offset + size);
+		update_size = true;
+	}
+	spin_unlock(&ip->i_flags_lock);
+
+	if (flags & IOMAP_DIO_COW) {
+		error = xfs_reflink_end_cow(ip, offset, size);
+		if (error)
+			return error;
+	}
+
+	if (flags & IOMAP_DIO_UNWRITTEN)
+		error = xfs_iomap_write_unwritten(ip, offset, size);
+	else if (update_size)
+		error = xfs_setfilesize(ip, offset, size);
+
+	return error;
+}
+
 /*
  * xfs_file_dio_aio_write - handle direct IO writes
  *
@@ -535,9 +510,7 @@ xfs_file_dio_aio_write(
 	int			unaligned_io = 0;
 	int			iolock;
 	size_t			count = iov_iter_count(from);
-	loff_t			end;
-	struct iov_iter		data;
-	struct xfs_buftarg	*target = XFS_IS_REALTIME_INODE(ip) ?
+	struct xfs_buftarg      *target = XFS_IS_REALTIME_INODE(ip) ?
 					mp->m_rtdev_targp : mp->m_ddev_targp;
 
 	/* DIO must be aligned to device logical sector size */
@@ -559,29 +532,12 @@ xfs_file_dio_aio_write(
 		iolock = XFS_IOLOCK_SHARED;
 	}
 
-	xfs_rw_ilock(ip, iolock);
+	xfs_ilock(ip, iolock);
 
 	ret = xfs_file_aio_write_checks(iocb, from, &iolock);
 	if (ret)
 		goto out;
 	count = iov_iter_count(from);
-	end = iocb->ki_pos + count - 1;
-
-	if (mapping->nrpages) {
-		ret = filemap_write_and_wait_range(mapping, iocb->ki_pos, end);
-		if (ret)
-			goto out;
-
-		/*
-		 * Invalidate whole pages. This can return an error if we fail
-		 * to invalidate a page, but this should never happen on XFS.
-		 * Warn if it does fail.
-		 */
-		ret = invalidate_inode_pages2_range(mapping,
-				iocb->ki_pos >> PAGE_SHIFT, end >> PAGE_SHIFT);
-		WARN_ON_ONCE(ret);
-		ret = 0;
-	}
 
 	/*
 	 * If we are doing unaligned IO, wait for all other IO to drain,
@@ -591,7 +547,7 @@ xfs_file_dio_aio_write(
 	if (unaligned_io)
 		inode_dio_wait(inode);
 	else if (iolock == XFS_IOLOCK_EXCL) {
-		xfs_rw_ilock_demote(ip, XFS_IOLOCK_EXCL);
+		xfs_ilock_demote(ip, XFS_IOLOCK_EXCL);
 		iolock = XFS_IOLOCK_SHARED;
 	}
 
@@ -604,24 +560,9 @@ xfs_file_dio_aio_write(
 			goto out;
 	}
 
-	data = *from;
-	ret = __blockdev_direct_IO(iocb, inode, target->bt_bdev, &data,
-			xfs_get_blocks_direct, xfs_end_io_direct_write,
-			NULL, DIO_ASYNC_EXTEND);
-
-	/* see generic_file_direct_write() for why this is necessary */
-	if (mapping->nrpages) {
-		invalidate_inode_pages2_range(mapping,
-					      iocb->ki_pos >> PAGE_SHIFT,
-					      end >> PAGE_SHIFT);
-	}
-
-	if (ret > 0) {
-		iocb->ki_pos += ret;
-		iov_iter_advance(from, ret);
-	}
+	ret = iomap_dio_rw(iocb, from, &xfs_iomap_ops, xfs_dio_write_end_io);
 out:
-	xfs_rw_iunlock(ip, iolock);
+	xfs_iunlock(ip, iolock);
 
 	/*
 	 * No fallback to buffered IO on errors for XFS, direct IO will either
@@ -643,7 +584,7 @@ xfs_file_dax_write(
 	size_t			count;
 	loff_t			pos;
 
-	xfs_rw_ilock(ip, iolock);
+	xfs_ilock(ip, iolock);
 	ret = xfs_file_aio_write_checks(iocb, from, &iolock);
 	if (ret)
 		goto out;
@@ -652,15 +593,13 @@ xfs_file_dax_write(
 	count = iov_iter_count(from);
 
 	trace_xfs_file_dax_write(ip, count, pos);
-
 	ret = dax_iomap_rw(iocb, from, &xfs_iomap_ops);
 	if (ret > 0 && iocb->ki_pos > i_size_read(inode)) {
 		i_size_write(inode, iocb->ki_pos);
 		error = xfs_setfilesize(ip, pos, ret);
 	}
-
 out:
-	xfs_rw_iunlock(ip, iolock);
+	xfs_iunlock(ip, iolock);
 	return error ? error : ret;
 }
 
@@ -677,7 +616,7 @@ xfs_file_buffered_aio_write(
 	int			enospc = 0;
 	int			iolock = XFS_IOLOCK_EXCL;
 
-	xfs_rw_ilock(ip, iolock);
+	xfs_ilock(ip, iolock);
 
 	ret = xfs_file_aio_write_checks(iocb, from, &iolock);
 	if (ret)
@@ -721,7 +660,7 @@ write_retry:
 
 	current->backing_dev_info = NULL;
 out:
-	xfs_rw_iunlock(ip, iolock);
+	xfs_iunlock(ip, iolock);
 	return ret;
 }
 
@@ -797,7 +736,7 @@ xfs_file_fallocate(
 		return -EOPNOTSUPP;
 
 	xfs_ilock(ip, iolock);
-	error = xfs_break_layouts(inode, &iolock, false);
+	error = xfs_break_layouts(inode, &iolock);
 	if (error)
 		goto out_unlock;
 
@@ -939,7 +878,6 @@ xfs_file_clone_range(
 				     len, false);
 }
 
-#define XFS_MAX_DEDUPE_LEN	(16 * 1024 * 1024)
 STATIC ssize_t
 xfs_file_dedupe_range(
 	struct file	*src_file,
@@ -950,14 +888,6 @@ xfs_file_dedupe_range(
 {
 	int		error;
 
-	/*
-	 * Limit the total length we will dedupe for each operation.
-	 * This is intended to bound the total time spent in this
-	 * ioctl to something sane.
-	 */
-	if (len > XFS_MAX_DEDUPE_LEN)
-		len = XFS_MAX_DEDUPE_LEN;
-
 	error = xfs_reflink_remap_range(src_file, loff, dst_file, dst_loff,
 				     len, true);
 	if (error)
@@ -1501,15 +1431,9 @@ xfs_filemap_fault(
 		return xfs_filemap_page_mkwrite(vma, vmf);
 
 	xfs_ilock(XFS_I(inode), XFS_MMAPLOCK_SHARED);
-	if (IS_DAX(inode)) {
-		/*
-		 * we do not want to trigger unwritten extent conversion on read
-		 * faults - that is unnecessary overhead and would also require
-		 * changes to xfs_get_blocks_direct() to map unwritten extent
-		 * ioend for conversion on read-only mappings.
-		 */
+	if (IS_DAX(inode))
 		ret = dax_iomap_fault(vma, vmf, &xfs_iomap_ops);
-	} else
+	else
 		ret = filemap_fault(vma, vmf);
 	xfs_iunlock(XFS_I(inode), XFS_MMAPLOCK_SHARED);
 

fs/xfs/xfs_icache.c

@@ -70,8 +70,6 @@ xfs_inode_alloc(
 	ASSERT(!xfs_isiflocked(ip));
 	ASSERT(ip->i_ino == 0);
 
-	mrlock_init(&ip->i_iolock, MRLOCK_BARRIER, "xfsio", ip->i_ino);
-
 	/* initialise the xfs inode */
 	ip->i_ino = ino;
 	ip->i_mount = mp;
@@ -123,7 +121,6 @@ __xfs_inode_free(
 {
 	/* asserts to verify all state is correct here */
 	ASSERT(atomic_read(&ip->i_pincount) == 0);
-	ASSERT(!xfs_isiflocked(ip));
 	XFS_STATS_DEC(ip->i_mount, vn_active);
 
 	call_rcu(&VFS_I(ip)->i_rcu, xfs_inode_free_callback);
@@ -133,6 +130,8 @@ void
 xfs_inode_free(
 	struct xfs_inode	*ip)
 {
+	ASSERT(!xfs_isiflocked(ip));
+
 	/*
 	 * Because we use RCU freeing we need to ensure the inode always
 	 * appears to be reclaimed with an invalid inode number when in the
@@ -393,8 +392,8 @@ xfs_iget_cache_hit(
 		xfs_inode_clear_reclaim_tag(pag, ip->i_ino);
 		inode->i_state = I_NEW;
 
-		ASSERT(!rwsem_is_locked(&ip->i_iolock.mr_lock));
-		mrlock_init(&ip->i_iolock, MRLOCK_BARRIER, "xfsio", ip->i_ino);
+		ASSERT(!rwsem_is_locked(&inode->i_rwsem));
+		init_rwsem(&inode->i_rwsem);
 
 		spin_unlock(&ip->i_flags_lock);
 		spin_unlock(&pag->pag_ici_lock);
@@ -981,6 +980,7 @@ restart:
 
 	if (XFS_FORCED_SHUTDOWN(ip->i_mount)) {
 		xfs_iunpin_wait(ip);
+		/* xfs_iflush_abort() drops the flush lock */
 		xfs_iflush_abort(ip, false);
 		goto reclaim;
 	}
@@ -989,10 +989,10 @@ restart:
 			goto out_ifunlock;
 		xfs_iunpin_wait(ip);
 	}
-	if (xfs_iflags_test(ip, XFS_ISTALE))
-		goto reclaim;
-	if (xfs_inode_clean(ip))
+	if (xfs_iflags_test(ip, XFS_ISTALE) || xfs_inode_clean(ip)) {
+		xfs_ifunlock(ip);
 		goto reclaim;
+	}
 
 	/*
 	 * Never flush out dirty data during non-blocking reclaim, as it would
@@ -1030,25 +1030,24 @@ restart:
 		xfs_buf_relse(bp);
 	}
 
-	xfs_iflock(ip);
 reclaim:
+	ASSERT(!xfs_isiflocked(ip));
+
 	/*
 	 * Because we use RCU freeing we need to ensure the inode always appears
 	 * to be reclaimed with an invalid inode number when in the free state.
-	 * We do this as early as possible under the ILOCK and flush lock so
-	 * that xfs_iflush_cluster() can be guaranteed to detect races with us
-	 * here. By doing this, we guarantee that once xfs_iflush_cluster has
-	 * locked both the XFS_ILOCK and the flush lock that it will see either
-	 * a valid, flushable inode that will serialise correctly against the
-	 * locks below, or it will see a clean (and invalid) inode that it can
-	 * skip.
+	 * We do this as early as possible under the ILOCK so that
+	 * xfs_iflush_cluster() can be guaranteed to detect races with us here.
+	 * By doing this, we guarantee that once xfs_iflush_cluster has locked
+	 * XFS_ILOCK that it will see either a valid, flushable inode that will
+	 * serialise correctly, or it will see a clean (and invalid) inode that
+	 * it can skip.
 	 */
 	spin_lock(&ip->i_flags_lock);
 	ip->i_flags = XFS_IRECLAIM;
 	ip->i_ino = 0;
 	spin_unlock(&ip->i_flags_lock);
 
-	xfs_ifunlock(ip);
 	xfs_iunlock(ip, XFS_ILOCK_EXCL);
 
 	XFS_STATS_INC(ip->i_mount, xs_ig_reclaims);
@@ -1580,10 +1579,15 @@ xfs_inode_free_cowblocks(
 	struct xfs_eofblocks *eofb = args;
 	bool need_iolock = true;
 	int match;
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, XFS_COW_FORK);
 
 	ASSERT(!eofb || (eofb && eofb->eof_scan_owner != 0));
 
-	if (!xfs_reflink_has_real_cow_blocks(ip)) {
+	/*
+	 * Just clear the tag if we have an empty cow fork or none at all. It's
+	 * possible the inode was fully unshared since it was originally tagged.
+	 */
+	if (!xfs_is_reflink_inode(ip) || !ifp->if_bytes) {
 		trace_xfs_inode_free_cowblocks_invalid(ip);
 		xfs_inode_clear_cowblocks_tag(ip);
 		return 0;

fs/xfs/xfs_icreate_item.c

@@ -133,7 +133,7 @@ xfs_icreate_item_committing(
 /*
  * This is the ops vector shared by all buf log items.
  */
-static struct xfs_item_ops xfs_icreate_item_ops = {
+static const struct xfs_item_ops xfs_icreate_item_ops = {
 	.iop_size	= xfs_icreate_item_size,
 	.iop_format	= xfs_icreate_item_format,
 	.iop_pin	= xfs_icreate_item_pin,

fs/xfs/xfs_inode.c

@@ -142,31 +142,31 @@ xfs_ilock_attr_map_shared(
 }
 
 /*
- * The xfs inode contains 3 multi-reader locks: the i_iolock the i_mmap_lock and
- * the i_lock.  This routine allows various combinations of the locks to be
- * obtained.
+ * In addition to i_rwsem in the VFS inode, the xfs inode contains 2
+ * multi-reader locks: i_mmap_lock and the i_lock.  This routine allows
+ * various combinations of the locks to be obtained.
  *
  * The 3 locks should always be ordered so that the IO lock is obtained first,
  * the mmap lock second and the ilock last in order to prevent deadlock.
  *
  * Basic locking order:
  *
- * i_iolock -> i_mmap_lock -> page_lock -> i_ilock
+ * i_rwsem -> i_mmap_lock -> page_lock -> i_ilock
  *
  * mmap_sem locking order:
  *
- * i_iolock -> page lock -> mmap_sem
+ * i_rwsem -> page lock -> mmap_sem
  * mmap_sem -> i_mmap_lock -> page_lock
  *
  * The difference in mmap_sem locking order mean that we cannot hold the
  * i_mmap_lock over syscall based read(2)/write(2) based IO. These IO paths can
  * fault in pages during copy in/out (for buffered IO) or require the mmap_sem
  * in get_user_pages() to map the user pages into the kernel address space for
- * direct IO. Similarly the i_iolock cannot be taken inside a page fault because
+ * direct IO. Similarly the i_rwsem cannot be taken inside a page fault because
  * page faults already hold the mmap_sem.
  *
  * Hence to serialise fully against both syscall and mmap based IO, we need to
- * take both the i_iolock and the i_mmap_lock. These locks should *only* be both
+ * take both the i_rwsem and the i_mmap_lock. These locks should *only* be both
  * taken in places where we need to invalidate the page cache in a race
  * free manner (e.g. truncate, hole punch and other extent manipulation
  * functions).
@@ -191,10 +191,13 @@ xfs_ilock(
 	       (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL));
 	ASSERT((lock_flags & ~(XFS_LOCK_MASK | XFS_LOCK_SUBCLASS_MASK)) == 0);
 
-	if (lock_flags & XFS_IOLOCK_EXCL)
-		mrupdate_nested(&ip->i_iolock, XFS_IOLOCK_DEP(lock_flags));
-	else if (lock_flags & XFS_IOLOCK_SHARED)
-		mraccess_nested(&ip->i_iolock, XFS_IOLOCK_DEP(lock_flags));
+	if (lock_flags & XFS_IOLOCK_EXCL) {
+		down_write_nested(&VFS_I(ip)->i_rwsem,
+				  XFS_IOLOCK_DEP(lock_flags));
+	} else if (lock_flags & XFS_IOLOCK_SHARED) {
+		down_read_nested(&VFS_I(ip)->i_rwsem,
+				 XFS_IOLOCK_DEP(lock_flags));
+	}
 
 	if (lock_flags & XFS_MMAPLOCK_EXCL)
 		mrupdate_nested(&ip->i_mmaplock, XFS_MMAPLOCK_DEP(lock_flags));
@@ -240,10 +243,10 @@ xfs_ilock_nowait(
 	ASSERT((lock_flags & ~(XFS_LOCK_MASK | XFS_LOCK_SUBCLASS_MASK)) == 0);
 
 	if (lock_flags & XFS_IOLOCK_EXCL) {
-		if (!mrtryupdate(&ip->i_iolock))
+		if (!down_write_trylock(&VFS_I(ip)->i_rwsem))
 			goto out;
 	} else if (lock_flags & XFS_IOLOCK_SHARED) {
-		if (!mrtryaccess(&ip->i_iolock))
+		if (!down_read_trylock(&VFS_I(ip)->i_rwsem))
 			goto out;
 	}
 
@@ -271,9 +274,9 @@ out_undo_mmaplock:
 		mrunlock_shared(&ip->i_mmaplock);
 out_undo_iolock:
 	if (lock_flags & XFS_IOLOCK_EXCL)
-		mrunlock_excl(&ip->i_iolock);
+		up_write(&VFS_I(ip)->i_rwsem);
 	else if (lock_flags & XFS_IOLOCK_SHARED)
-		mrunlock_shared(&ip->i_iolock);
+		up_read(&VFS_I(ip)->i_rwsem);
 out:
 	return 0;
 }
@@ -310,9 +313,9 @@ xfs_iunlock(
 	ASSERT(lock_flags != 0);
 
 	if (lock_flags & XFS_IOLOCK_EXCL)
-		mrunlock_excl(&ip->i_iolock);
+		up_write(&VFS_I(ip)->i_rwsem);
 	else if (lock_flags & XFS_IOLOCK_SHARED)
-		mrunlock_shared(&ip->i_iolock);
+		up_read(&VFS_I(ip)->i_rwsem);
 
 	if (lock_flags & XFS_MMAPLOCK_EXCL)
 		mrunlock_excl(&ip->i_mmaplock);
@@ -345,7 +348,7 @@ xfs_ilock_demote(
 	if (lock_flags & XFS_MMAPLOCK_EXCL)
 		mrdemote(&ip->i_mmaplock);
 	if (lock_flags & XFS_IOLOCK_EXCL)
-		mrdemote(&ip->i_iolock);
+		downgrade_write(&VFS_I(ip)->i_rwsem);
 
 	trace_xfs_ilock_demote(ip, lock_flags, _RET_IP_);
 }
@@ -370,8 +373,9 @@ xfs_isilocked(
 
 	if (lock_flags & (XFS_IOLOCK_EXCL|XFS_IOLOCK_SHARED)) {
 		if (!(lock_flags & XFS_IOLOCK_SHARED))
-			return !!ip->i_iolock.mr_writer;
-		return rwsem_is_locked(&ip->i_iolock.mr_lock);
+			return !debug_locks ||
+				lockdep_is_held_type(&VFS_I(ip)->i_rwsem, 0);
+		return rwsem_is_locked(&VFS_I(ip)->i_rwsem);
 	}
 
 	ASSERT(0);
@@ -421,11 +425,7 @@ xfs_lock_inumorder(int lock_mode, int subclass)
 
 	if (lock_mode & (XFS_IOLOCK_SHARED|XFS_IOLOCK_EXCL)) {
 		ASSERT(subclass <= XFS_IOLOCK_MAX_SUBCLASS);
-		ASSERT(xfs_lockdep_subclass_ok(subclass +
-						XFS_IOLOCK_PARENT_VAL));
 		class += subclass << XFS_IOLOCK_SHIFT;
-		if (lock_mode & XFS_IOLOCK_PARENT)
-			class += XFS_IOLOCK_PARENT_VAL << XFS_IOLOCK_SHIFT;
 	}
 
 	if (lock_mode & (XFS_MMAPLOCK_SHARED|XFS_MMAPLOCK_EXCL)) {
@@ -477,8 +477,6 @@ xfs_lock_inodes(
 			    XFS_ILOCK_EXCL));
 	ASSERT(!(lock_mode & (XFS_IOLOCK_SHARED | XFS_MMAPLOCK_SHARED |
 			      XFS_ILOCK_SHARED)));
-	ASSERT(!(lock_mode & XFS_IOLOCK_EXCL) ||
-		inodes <= XFS_IOLOCK_MAX_SUBCLASS + 1);
 	ASSERT(!(lock_mode & XFS_MMAPLOCK_EXCL) ||
 		inodes <= XFS_MMAPLOCK_MAX_SUBCLASS + 1);
 	ASSERT(!(lock_mode & XFS_ILOCK_EXCL) ||
@@ -581,10 +579,8 @@ xfs_lock_two_inodes(
 	int			attempts = 0;
 	xfs_log_item_t		*lp;
 
-	if (lock_mode & (XFS_IOLOCK_SHARED|XFS_IOLOCK_EXCL)) {
-		ASSERT(!(lock_mode & (XFS_MMAPLOCK_SHARED|XFS_MMAPLOCK_EXCL)));
-		ASSERT(!(lock_mode & (XFS_ILOCK_SHARED|XFS_ILOCK_EXCL)));
-	} else if (lock_mode & (XFS_MMAPLOCK_SHARED|XFS_MMAPLOCK_EXCL))
+	ASSERT(!(lock_mode & (XFS_IOLOCK_SHARED|XFS_IOLOCK_EXCL)));
+	if (lock_mode & (XFS_MMAPLOCK_SHARED|XFS_MMAPLOCK_EXCL))
 		ASSERT(!(lock_mode & (XFS_ILOCK_SHARED|XFS_ILOCK_EXCL)));
 
 	ASSERT(ip0->i_ino != ip1->i_ino);
@@ -715,7 +711,6 @@ xfs_lookup(
 	if (XFS_FORCED_SHUTDOWN(dp->i_mount))
 		return -EIO;
 
-	xfs_ilock(dp, XFS_IOLOCK_SHARED);
 	error = xfs_dir_lookup(NULL, dp, name, &inum, ci_name);
 	if (error)
 		goto out_unlock;
@@ -724,14 +719,12 @@ xfs_lookup(
 	if (error)
 		goto out_free_name;
 
-	xfs_iunlock(dp, XFS_IOLOCK_SHARED);
 	return 0;
 
 out_free_name:
 	if (ci_name)
 		kmem_free(ci_name->name);
 out_unlock:
-	xfs_iunlock(dp, XFS_IOLOCK_SHARED);
 	*ipp = NULL;
 	return error;
 }
@@ -1215,8 +1208,7 @@ xfs_create(
 	if (error)
 		goto out_release_inode;
 
-	xfs_ilock(dp, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL |
-		      XFS_IOLOCK_PARENT | XFS_ILOCK_PARENT);
+	xfs_ilock(dp, XFS_ILOCK_EXCL | XFS_ILOCK_PARENT);
 	unlock_dp_on_error = true;
 
 	xfs_defer_init(&dfops, &first_block);
@@ -1252,7 +1244,7 @@ xfs_create(
 	 * the transaction cancel unlocking dp so don't do it explicitly in the
 	 * error path.
 	 */
-	xfs_trans_ijoin(tp, dp, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL);
+	xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL);
 	unlock_dp_on_error = false;
 
 	error = xfs_dir_createname(tp, dp, name, ip->i_ino,
@@ -1325,7 +1317,7 @@ xfs_create(
 	xfs_qm_dqrele(pdqp);
 
 	if (unlock_dp_on_error)
-		xfs_iunlock(dp, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL);
+		xfs_iunlock(dp, XFS_ILOCK_EXCL);
 	return error;
 }
 
@@ -1466,11 +1458,10 @@ xfs_link(
 	if (error)
 		goto std_return;
 
-	xfs_ilock(tdp, XFS_IOLOCK_EXCL | XFS_IOLOCK_PARENT);
 	xfs_lock_two_inodes(sip, tdp, XFS_ILOCK_EXCL);
 
 	xfs_trans_ijoin(tp, sip, XFS_ILOCK_EXCL);
-	xfs_trans_ijoin(tp, tdp, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL);
+	xfs_trans_ijoin(tp, tdp, XFS_ILOCK_EXCL);
 
 	/*
 	 * If we are using project inheritance, we only allow hard link
@@ -2041,7 +2032,6 @@ xfs_iunlink(
 	agi->agi_unlinked[bucket_index] = cpu_to_be32(agino);
 	offset = offsetof(xfs_agi_t, agi_unlinked) +
 		(sizeof(xfs_agino_t) * bucket_index);
-	xfs_trans_buf_set_type(tp, agibp, XFS_BLFT_AGI_BUF);
 	xfs_trans_log_buf(tp, agibp, offset,
 			  (offset + sizeof(xfs_agino_t) - 1));
 	return 0;
@@ -2133,7 +2123,6 @@ xfs_iunlink_remove(
 		agi->agi_unlinked[bucket_index] = cpu_to_be32(next_agino);
 		offset = offsetof(xfs_agi_t, agi_unlinked) +
 			(sizeof(xfs_agino_t) * bucket_index);
-		xfs_trans_buf_set_type(tp, agibp, XFS_BLFT_AGI_BUF);
 		xfs_trans_log_buf(tp, agibp, offset,
 				  (offset + sizeof(xfs_agino_t) - 1));
 	} else {
@@ -2579,10 +2568,9 @@ xfs_remove(
 		goto std_return;
 	}
 
-	xfs_ilock(dp, XFS_IOLOCK_EXCL | XFS_IOLOCK_PARENT);
 	xfs_lock_two_inodes(dp, ip, XFS_ILOCK_EXCL);
 
-	xfs_trans_ijoin(tp, dp, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL);
+	xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL);
 	xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
 
 	/*
@@ -2963,12 +2951,6 @@ xfs_rename(
 	 * whether the target directory is the same as the source
 	 * directory, we can lock from 2 to 4 inodes.
 	 */
-	if (!new_parent)
-		xfs_ilock(src_dp, XFS_IOLOCK_EXCL | XFS_IOLOCK_PARENT);
-	else
-		xfs_lock_two_inodes(src_dp, target_dp,
-				    XFS_IOLOCK_EXCL | XFS_IOLOCK_PARENT);
-
 	xfs_lock_inodes(inodes, num_inodes, XFS_ILOCK_EXCL);
 
 	/*
@@ -2976,9 +2958,9 @@ xfs_rename(
 	 * we can rely on either trans_commit or trans_cancel to unlock
 	 * them.
 	 */
-	xfs_trans_ijoin(tp, src_dp, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL);
+	xfs_trans_ijoin(tp, src_dp, XFS_ILOCK_EXCL);
 	if (new_parent)
-		xfs_trans_ijoin(tp, target_dp, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL);
+		xfs_trans_ijoin(tp, target_dp, XFS_ILOCK_EXCL);
 	xfs_trans_ijoin(tp, src_ip, XFS_ILOCK_EXCL);
 	if (target_ip)
 		xfs_trans_ijoin(tp, target_ip, XFS_ILOCK_EXCL);

fs/xfs/xfs_inode.h

@@ -56,7 +56,6 @@ typedef struct xfs_inode {
 	/* Transaction and locking information. */
 	struct xfs_inode_log_item *i_itemp;	/* logging information */
 	mrlock_t		i_lock;		/* inode lock */
-	mrlock_t		i_iolock;	/* inode IO lock */
 	mrlock_t		i_mmaplock;	/* inode mmap IO lock */
 	atomic_t		i_pincount;	/* inode pin count */
 	spinlock_t		i_flags_lock;	/* inode i_flags lock */
@@ -246,6 +245,11 @@ static inline bool xfs_is_reflink_inode(struct xfs_inode *ip)
  * Synchronize processes attempting to flush the in-core inode back to disk.
  */
 
+static inline int xfs_isiflocked(struct xfs_inode *ip)
+{
+	return xfs_iflags_test(ip, XFS_IFLOCK);
+}
+
 extern void __xfs_iflock(struct xfs_inode *ip);
 
 static inline int xfs_iflock_nowait(struct xfs_inode *ip)
@@ -261,16 +265,12 @@ static inline void xfs_iflock(struct xfs_inode *ip)
 
 static inline void xfs_ifunlock(struct xfs_inode *ip)
 {
+	ASSERT(xfs_isiflocked(ip));
 	xfs_iflags_clear(ip, XFS_IFLOCK);
 	smp_mb();
 	wake_up_bit(&ip->i_flags, __XFS_IFLOCK_BIT);
 }
 
-static inline int xfs_isiflocked(struct xfs_inode *ip)
-{
-	return xfs_iflags_test(ip, XFS_IFLOCK);
-}
-
 /*
  * Flags for inode locking.
  * Bit ranges:	1<<1  - 1<<16-1 -- iolock/ilock modes (bitfield)
@@ -332,7 +332,7 @@ static inline int xfs_isiflocked(struct xfs_inode *ip)
  * IOLOCK values
  *
  * 0-3		subclass value
- * 4-7		PARENT subclass values
+ * 4-7		unused
  *
  * MMAPLOCK values
  *
@@ -347,10 +347,8 @@ static inline int xfs_isiflocked(struct xfs_inode *ip)
  * 
  */
 #define XFS_IOLOCK_SHIFT		16
-#define XFS_IOLOCK_PARENT_VAL		4
-#define XFS_IOLOCK_MAX_SUBCLASS		(XFS_IOLOCK_PARENT_VAL - 1)
+#define XFS_IOLOCK_MAX_SUBCLASS		3
 #define XFS_IOLOCK_DEP_MASK		0x000f0000
-#define	XFS_IOLOCK_PARENT		(XFS_IOLOCK_PARENT_VAL << XFS_IOLOCK_SHIFT)
 
 #define XFS_MMAPLOCK_SHIFT		20
 #define XFS_MMAPLOCK_NUMORDER		0

fs/xfs/xfs_inode_item.c

@@ -164,7 +164,7 @@ xfs_inode_item_format_data_fork(
 			struct xfs_bmbt_rec *p;
 
 			ASSERT(ip->i_df.if_u1.if_extents != NULL);
-			ASSERT(ip->i_df.if_bytes / sizeof(xfs_bmbt_rec_t) > 0);
+			ASSERT(xfs_iext_count(&ip->i_df) > 0);
 
 			p = xlog_prepare_iovec(lv, vecp, XLOG_REG_TYPE_IEXT);
 			data_bytes = xfs_iextents_copy(ip, p, XFS_DATA_FORK);
@@ -261,7 +261,7 @@ xfs_inode_item_format_attr_fork(
 		    ip->i_afp->if_bytes > 0) {
 			struct xfs_bmbt_rec *p;
 
-			ASSERT(ip->i_afp->if_bytes / sizeof(xfs_bmbt_rec_t) ==
+			ASSERT(xfs_iext_count(ip->i_afp) ==
 				ip->i_d.di_anextents);
 			ASSERT(ip->i_afp->if_u1.if_extents != NULL);
 

fs/xfs/xfs_ioctl.c

@@ -639,7 +639,7 @@ xfs_ioc_space(
 		return error;
 
 	xfs_ilock(ip, iolock);
-	error = xfs_break_layouts(inode, &iolock, false);
+	error = xfs_break_layouts(inode, &iolock);
 	if (error)
 		goto out_unlock;
 
@@ -910,16 +910,14 @@ xfs_ioc_fsgetxattr(
 	if (attr) {
 		if (ip->i_afp) {
 			if (ip->i_afp->if_flags & XFS_IFEXTENTS)
-				fa.fsx_nextents = ip->i_afp->if_bytes /
-							sizeof(xfs_bmbt_rec_t);
+				fa.fsx_nextents = xfs_iext_count(ip->i_afp);
 			else
 				fa.fsx_nextents = ip->i_d.di_anextents;
 		} else
 			fa.fsx_nextents = 0;
 	} else {
 		if (ip->i_df.if_flags & XFS_IFEXTENTS)
-			fa.fsx_nextents = ip->i_df.if_bytes /
-						sizeof(xfs_bmbt_rec_t);
+			fa.fsx_nextents = xfs_iext_count(&ip->i_df);
 		else
 			fa.fsx_nextents = ip->i_d.di_nextents;
 	}

fs/xfs/xfs_iomap.c

@@ -395,11 +395,12 @@ xfs_iomap_prealloc_size(
 	struct xfs_inode	*ip,
 	loff_t			offset,
 	loff_t			count,
-	xfs_extnum_t		idx,
-	struct xfs_bmbt_irec	*prev)
+	xfs_extnum_t		idx)
 {
 	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
 	xfs_fileoff_t		offset_fsb = XFS_B_TO_FSBT(mp, offset);
+	struct xfs_bmbt_irec	prev;
 	int			shift = 0;
 	int64_t			freesp;
 	xfs_fsblock_t		qblocks;
@@ -419,8 +420,8 @@ xfs_iomap_prealloc_size(
 	 */
 	if ((mp->m_flags & XFS_MOUNT_DFLT_IOSIZE) ||
 	    XFS_ISIZE(ip) < XFS_FSB_TO_B(mp, mp->m_dalign) ||
-	    idx == 0 ||
-	    prev->br_startoff + prev->br_blockcount < offset_fsb)
+	    !xfs_iext_get_extent(ifp, idx - 1, &prev) ||
+	    prev.br_startoff + prev.br_blockcount < offset_fsb)
 		return mp->m_writeio_blocks;
 
 	/*
@@ -439,8 +440,8 @@ xfs_iomap_prealloc_size(
 	 * always extends to MAXEXTLEN rather than falling short due to things
 	 * like stripe unit/width alignment of real extents.
 	 */
-	if (prev->br_blockcount <= (MAXEXTLEN >> 1))
-		alloc_blocks = prev->br_blockcount << 1;
+	if (prev.br_blockcount <= (MAXEXTLEN >> 1))
+		alloc_blocks = prev.br_blockcount << 1;
 	else
 		alloc_blocks = XFS_B_TO_FSB(mp, offset);
 	if (!alloc_blocks)
@@ -535,11 +536,11 @@ xfs_file_iomap_begin_delay(
 	xfs_fileoff_t		offset_fsb = XFS_B_TO_FSBT(mp, offset);
 	xfs_fileoff_t		maxbytes_fsb =
 		XFS_B_TO_FSB(mp, mp->m_super->s_maxbytes);
-	xfs_fileoff_t		end_fsb, orig_end_fsb;
+	xfs_fileoff_t		end_fsb;
 	int			error = 0, eof = 0;
 	struct xfs_bmbt_irec	got;
-	struct xfs_bmbt_irec	prev;
 	xfs_extnum_t		idx;
+	xfs_fsblock_t		prealloc_blocks = 0;
 
 	ASSERT(!XFS_IS_REALTIME_INODE(ip));
 	ASSERT(!xfs_get_extsz_hint(ip));
@@ -563,8 +564,7 @@ xfs_file_iomap_begin_delay(
 			goto out_unlock;
 	}
 
-	xfs_bmap_search_extents(ip, offset_fsb, XFS_DATA_FORK, &eof, &idx,
-			&got, &prev);
+	eof = !xfs_iext_lookup_extent(ip, ifp, offset_fsb, &idx, &got);
 	if (!eof && got.br_startoff <= offset_fsb) {
 		if (xfs_is_reflink_inode(ip)) {
 			bool		shared;
@@ -595,35 +595,32 @@ xfs_file_iomap_begin_delay(
 	 * the lower level functions are updated.
 	 */
 	count = min_t(loff_t, count, 1024 * PAGE_SIZE);
-	end_fsb = orig_end_fsb =
-		min(XFS_B_TO_FSB(mp, offset + count), maxbytes_fsb);
+	end_fsb = min(XFS_B_TO_FSB(mp, offset + count), maxbytes_fsb);
 
 	if (eof) {
-		xfs_fsblock_t	prealloc_blocks;
-
-		prealloc_blocks =
-			xfs_iomap_prealloc_size(ip, offset, count, idx, &prev);
+		prealloc_blocks = xfs_iomap_prealloc_size(ip, offset, count, idx);
 		if (prealloc_blocks) {
 			xfs_extlen_t	align;
 			xfs_off_t	end_offset;
+			xfs_fileoff_t	p_end_fsb;
 
 			end_offset = XFS_WRITEIO_ALIGN(mp, offset + count - 1);
-			end_fsb = XFS_B_TO_FSBT(mp, end_offset) +
-				prealloc_blocks;
+			p_end_fsb = XFS_B_TO_FSBT(mp, end_offset) +
+					prealloc_blocks;
 
 			align = xfs_eof_alignment(ip, 0);
 			if (align)
-				end_fsb = roundup_64(end_fsb, align);
+				p_end_fsb = roundup_64(p_end_fsb, align);
 
-			end_fsb = min(end_fsb, maxbytes_fsb);
-			ASSERT(end_fsb > offset_fsb);
+			p_end_fsb = min(p_end_fsb, maxbytes_fsb);
+			ASSERT(p_end_fsb > offset_fsb);
+			prealloc_blocks = p_end_fsb - end_fsb;
 		}
 	}
 
 retry:
 	error = xfs_bmapi_reserve_delalloc(ip, XFS_DATA_FORK, offset_fsb,
-			end_fsb - offset_fsb, &got,
-			&prev, &idx, eof);
+			end_fsb - offset_fsb, prealloc_blocks, &got, &idx, eof);
 	switch (error) {
 	case 0:
 		break;
@@ -631,8 +628,8 @@ retry:
 	case -EDQUOT:
 		/* retry without any preallocation */
 		trace_xfs_delalloc_enospc(ip, offset, count);
-		if (end_fsb != orig_end_fsb) {
-			end_fsb = orig_end_fsb;
+		if (prealloc_blocks) {
+			prealloc_blocks = 0;
 			goto retry;
 		}
 		/*FALLTHRU*/
@@ -640,13 +637,6 @@ retry:
 		goto out_unlock;
 	}
 
-	/*
-	 * Tag the inode as speculatively preallocated so we can reclaim this
-	 * space on demand, if necessary.
-	 */
-	if (end_fsb != orig_end_fsb)
-		xfs_inode_set_eofblocks_tag(ip);
-
 	trace_xfs_iomap_alloc(ip, offset, count, 0, &got);
 done:
 	if (isnullstartblock(got.br_startblock))
@@ -960,6 +950,19 @@ static inline bool imap_needs_alloc(struct inode *inode,
 		(IS_DAX(inode) && ISUNWRITTEN(imap));
 }
 
+static inline bool need_excl_ilock(struct xfs_inode *ip, unsigned flags)
+{
+	/*
+	 * COW writes will allocate delalloc space, 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 ((flags & IOMAP_DIRECT) && (flags & IOMAP_WRITE))
+		return true;
+	return false;
+}
+
 static int
 xfs_file_iomap_begin(
 	struct inode		*inode,
@@ -979,18 +982,14 @@ xfs_file_iomap_begin(
 	if (XFS_FORCED_SHUTDOWN(mp))
 		return -EIO;
 
-	if ((flags & IOMAP_WRITE) && !IS_DAX(inode) &&
-		   !xfs_get_extsz_hint(ip)) {
+	if (((flags & (IOMAP_WRITE | IOMAP_DIRECT)) == IOMAP_WRITE) &&
+			!IS_DAX(inode) && !xfs_get_extsz_hint(ip)) {
 		/* Reserve delalloc blocks for regular writeback. */
 		return xfs_file_iomap_begin_delay(inode, offset, length, flags,
 				iomap);
 	}
 
-	/*
-	 * COW writes will allocate delalloc space, so we need to make sure
-	 * to take the lock exclusively here.
-	 */
-	if ((flags & (IOMAP_WRITE | IOMAP_ZERO)) && xfs_is_reflink_inode(ip)) {
+	if (need_excl_ilock(ip, flags)) {
 		lockmode = XFS_ILOCK_EXCL;
 		xfs_ilock(ip, XFS_ILOCK_EXCL);
 	} else {
@@ -1003,17 +1002,41 @@ xfs_file_iomap_begin(
 	offset_fsb = XFS_B_TO_FSBT(mp, offset);
 	end_fsb = XFS_B_TO_FSB(mp, offset + length);
 
+	if (xfs_is_reflink_inode(ip) &&
+	    (flags & IOMAP_WRITE) && (flags & IOMAP_DIRECT)) {
+		shared = xfs_reflink_find_cow_mapping(ip, offset, &imap);
+		if (shared) {
+			xfs_iunlock(ip, lockmode);
+			goto alloc_done;
+		}
+		ASSERT(!isnullstartblock(imap.br_startblock));
+	}
+
 	error = xfs_bmapi_read(ip, offset_fsb, end_fsb - offset_fsb, &imap,
 			       &nimaps, 0);
 	if (error)
 		goto out_unlock;
 
-	if (flags & IOMAP_REPORT) {
+	if ((flags & IOMAP_REPORT) ||
+	    (xfs_is_reflink_inode(ip) &&
+	     (flags & IOMAP_WRITE) && (flags & IOMAP_DIRECT))) {
 		/* Trim the mapping to the nearest shared extent boundary. */
 		error = xfs_reflink_trim_around_shared(ip, &imap, &shared,
 				&trimmed);
 		if (error)
 			goto out_unlock;
+
+		/*
+		 * We're here because we're trying to do a directio write to a
+		 * region that isn't aligned to a filesystem block.  If the
+		 * extent is shared, fall back to buffered mode to handle the
+		 * RMW.
+		 */
+		if (!(flags & IOMAP_REPORT) && shared) {
+			trace_xfs_reflink_bounce_dio_write(ip, &imap);
+			error = -EREMCHG;
+			goto out_unlock;
+		}
 	}
 
 	if ((flags & (IOMAP_WRITE | IOMAP_ZERO)) && xfs_is_reflink_inode(ip)) {
@@ -1048,6 +1071,7 @@ xfs_file_iomap_begin(
 		if (error)
 			return error;
 
+alloc_done:
 		iomap->flags = IOMAP_F_NEW;
 		trace_xfs_iomap_alloc(ip, offset, length, 0, &imap);
 	} else {

fs/xfs/xfs_iops.c

@@ -983,15 +983,13 @@ xfs_vn_setattr(
 		struct xfs_inode	*ip = XFS_I(d_inode(dentry));
 		uint			iolock = XFS_IOLOCK_EXCL;
 
-		xfs_ilock(ip, iolock);
-		error = xfs_break_layouts(d_inode(dentry), &iolock, true);
-		if (!error) {
-			xfs_ilock(ip, XFS_MMAPLOCK_EXCL);
-			iolock |= XFS_MMAPLOCK_EXCL;
+		error = xfs_break_layouts(d_inode(dentry), &iolock);
+		if (error)
+			return error;
 
-			error = xfs_vn_setattr_size(dentry, iattr);
-		}
-		xfs_iunlock(ip, iolock);
+		xfs_ilock(ip, XFS_MMAPLOCK_EXCL);
+		error = xfs_vn_setattr_size(dentry, iattr);
+		xfs_iunlock(ip, XFS_MMAPLOCK_EXCL);
 	} else {
 		error = xfs_vn_setattr_nonsize(dentry, iattr);
 	}

fs/xfs/xfs_linux.h

@@ -78,6 +78,7 @@ typedef __u32			xfs_nlink_t;
 #include <linux/freezer.h>
 #include <linux/list_sort.h>
 #include <linux/ratelimit.h>
+#include <linux/rhashtable.h>
 
 #include <asm/page.h>
 #include <asm/div64.h>

fs/xfs/xfs_log.c

@@ -1668,7 +1668,7 @@ xlog_cksum(
 	__uint32_t		crc;
 
 	/* first generate the crc for the record header ... */
-	crc = xfs_start_cksum((char *)rhead,
+	crc = xfs_start_cksum_update((char *)rhead,
 			      sizeof(struct xlog_rec_header),
 			      offsetof(struct xlog_rec_header, h_crc));
 
@@ -1862,26 +1862,21 @@ xlog_sync(
 
 	bp->b_io_length = BTOBB(count);
 	bp->b_fspriv = iclog;
-	bp->b_flags &= ~(XBF_FUA | XBF_FLUSH);
-	bp->b_flags |= (XBF_ASYNC | XBF_SYNCIO | XBF_WRITE);
+	bp->b_flags &= ~XBF_FLUSH;
+	bp->b_flags |= (XBF_ASYNC | XBF_SYNCIO | XBF_WRITE | XBF_FUA);
 
-	if (log->l_mp->m_flags & XFS_MOUNT_BARRIER) {
-		bp->b_flags |= XBF_FUA;
-
-		/*
-		 * Flush the data device before flushing the log to make
-		 * sure all meta data written back from the AIL actually made
-		 * it to disk before stamping the new log tail LSN into the
-		 * log buffer.  For an external log we need to issue the
-		 * flush explicitly, and unfortunately synchronously here;
-		 * for an internal log we can simply use the block layer
-		 * state machine for preflushes.
-		 */
-		if (log->l_mp->m_logdev_targp != log->l_mp->m_ddev_targp)
-			xfs_blkdev_issue_flush(log->l_mp->m_ddev_targp);
-		else
-			bp->b_flags |= XBF_FLUSH;
-	}
+	/*
+	 * Flush the data device before flushing the log to make sure all meta
+	 * data written back from the AIL actually made it to disk before
+	 * stamping the new log tail LSN into the log buffer.  For an external
+	 * log we need to issue the flush explicitly, and unfortunately
+	 * synchronously here; for an internal log we can simply use the block
+	 * layer state machine for preflushes.
+	 */
+	if (log->l_mp->m_logdev_targp != log->l_mp->m_ddev_targp)
+		xfs_blkdev_issue_flush(log->l_mp->m_ddev_targp);
+	else
+		bp->b_flags |= XBF_FLUSH;
 
 	ASSERT(XFS_BUF_ADDR(bp) <= log->l_logBBsize-1);
 	ASSERT(XFS_BUF_ADDR(bp) + BTOBB(count) <= log->l_logBBsize);
@@ -1906,10 +1901,8 @@ xlog_sync(
 		xfs_buf_associate_memory(bp,
 				(char *)&iclog->ic_header + count, split);
 		bp->b_fspriv = iclog;
-		bp->b_flags &= ~(XBF_FUA | XBF_FLUSH);
-		bp->b_flags |= (XBF_ASYNC | XBF_SYNCIO | XBF_WRITE);
-		if (log->l_mp->m_flags & XFS_MOUNT_BARRIER)
-			bp->b_flags |= XBF_FUA;
+		bp->b_flags &= ~XBF_FLUSH;
+		bp->b_flags |= (XBF_ASYNC | XBF_SYNCIO | XBF_WRITE | XBF_FUA);
 
 		ASSERT(XFS_BUF_ADDR(bp) <= log->l_logBBsize-1);
 		ASSERT(XFS_BUF_ADDR(bp) + BTOBB(count) <= log->l_logBBsize);

fs/xfs/xfs_log_recover.c

@@ -2025,7 +2025,7 @@ xlog_peek_buffer_cancelled(
 	struct xlog		*log,
 	xfs_daddr_t		blkno,
 	uint			len,
-	ushort			flags)
+	unsigned short			flags)
 {
 	struct list_head	*bucket;
 	struct xfs_buf_cancel	*bcp;
@@ -2065,7 +2065,7 @@ xlog_check_buffer_cancelled(
 	struct xlog		*log,
 	xfs_daddr_t		blkno,
 	uint			len,
-	ushort			flags)
+	unsigned short			flags)
 {
 	struct xfs_buf_cancel	*bcp;
 
@@ -5113,19 +5113,21 @@ xlog_recover_process(
 	struct list_head	*buffer_list)
 {
 	int			error;
+	__le32			old_crc = rhead->h_crc;
 	__le32			crc;
 
+
 	crc = xlog_cksum(log, rhead, dp, be32_to_cpu(rhead->h_len));
 
 	/*
 	 * Nothing else to do if this is a CRC verification pass. Just return
 	 * if this a record with a non-zero crc. Unfortunately, mkfs always
-	 * sets h_crc to 0 so we must consider this valid even on v5 supers.
+	 * sets old_crc to 0 so we must consider this valid even on v5 supers.
 	 * Otherwise, return EFSBADCRC on failure so the callers up the stack
 	 * know precisely what failed.
 	 */
 	if (pass == XLOG_RECOVER_CRCPASS) {
-		if (rhead->h_crc && crc != rhead->h_crc)
+		if (old_crc && crc != old_crc)
 			return -EFSBADCRC;
 		return 0;
 	}
@@ -5136,11 +5138,11 @@ xlog_recover_process(
 	 * zero CRC check prevents warnings from being emitted when upgrading
 	 * the kernel from one that does not add CRCs by default.
 	 */
-	if (crc != rhead->h_crc) {
-		if (rhead->h_crc || xfs_sb_version_hascrc(&log->l_mp->m_sb)) {
+	if (crc != old_crc) {
+		if (old_crc || xfs_sb_version_hascrc(&log->l_mp->m_sb)) {
 			xfs_alert(log->l_mp,
 		"log record CRC mismatch: found 0x%x, expected 0x%x.",
-					le32_to_cpu(rhead->h_crc),
+					le32_to_cpu(old_crc),
 					le32_to_cpu(crc));
 			xfs_hex_dump(dp, 32);
 		}

fs/xfs/xfs_mount.c

@@ -157,6 +157,7 @@ xfs_free_perag(
 		spin_unlock(&mp->m_perag_lock);
 		ASSERT(pag);
 		ASSERT(atomic_read(&pag->pag_ref) == 0);
+		xfs_buf_hash_destroy(pag);
 		call_rcu(&pag->rcu_head, __xfs_free_perag);
 	}
 }
@@ -212,8 +213,8 @@ xfs_initialize_perag(
 		spin_lock_init(&pag->pag_ici_lock);
 		mutex_init(&pag->pag_ici_reclaim_lock);
 		INIT_RADIX_TREE(&pag->pag_ici_root, GFP_ATOMIC);
-		spin_lock_init(&pag->pag_buf_lock);
-		pag->pag_buf_tree = RB_ROOT;
+		if (xfs_buf_hash_init(pag))
+			goto out_unwind;
 
 		if (radix_tree_preload(GFP_NOFS))
 			goto out_unwind;
@@ -239,9 +240,11 @@ xfs_initialize_perag(
 	return 0;
 
 out_unwind:
+	xfs_buf_hash_destroy(pag);
 	kmem_free(pag);
 	for (; index > first_initialised; index--) {
 		pag = radix_tree_delete(&mp->m_perag_tree, index);
+		xfs_buf_hash_destroy(pag);
 		kmem_free(pag);
 	}
 	return error;

fs/xfs/xfs_mount.h

@@ -393,8 +393,8 @@ typedef struct xfs_perag {
 	unsigned long	pag_ici_reclaim_cursor;	/* reclaim restart point */
 
 	/* buffer cache index */
-	spinlock_t	pag_buf_lock;	/* lock for pag_buf_tree */
-	struct rb_root	pag_buf_tree;	/* ordered tree of active buffers */
+	spinlock_t	pag_buf_lock;	/* lock for pag_buf_hash */
+	struct rhashtable pag_buf_hash;
 
 	/* for rcu-safe freeing */
 	struct rcu_head	rcu_head;
@@ -424,6 +424,9 @@ xfs_perag_resv(
 	}
 }
 
+int xfs_buf_hash_init(xfs_perag_t *pag);
+void xfs_buf_hash_destroy(xfs_perag_t *pag);
+
 extern void	xfs_uuid_table_free(void);
 extern int	xfs_log_sbcount(xfs_mount_t *);
 extern __uint64_t xfs_default_resblks(xfs_mount_t *mp);

fs/xfs/xfs_pnfs.c

@@ -32,8 +32,7 @@
 int
 xfs_break_layouts(
 	struct inode		*inode,
-	uint			*iolock,
-	bool			with_imutex)
+	uint			*iolock)
 {
 	struct xfs_inode	*ip = XFS_I(inode);
 	int			error;
@@ -42,12 +41,8 @@ xfs_break_layouts(
 
 	while ((error = break_layout(inode, false) == -EWOULDBLOCK)) {
 		xfs_iunlock(ip, *iolock);
-		if (with_imutex && (*iolock & XFS_IOLOCK_EXCL))
-			inode_unlock(inode);
 		error = break_layout(inode, true);
 		*iolock = XFS_IOLOCK_EXCL;
-		if (with_imutex)
-			inode_lock(inode);
 		xfs_ilock(ip, *iolock);
 	}
 

fs/xfs/xfs_pnfs.h

@@ -8,10 +8,10 @@ int xfs_fs_map_blocks(struct inode *inode, loff_t offset, u64 length,
 int xfs_fs_commit_blocks(struct inode *inode, struct iomap *maps, int nr_maps,
 		struct iattr *iattr);
 
-int xfs_break_layouts(struct inode *inode, uint *iolock, bool with_imutex);
+int xfs_break_layouts(struct inode *inode, uint *iolock);
 #else
 static inline int
-xfs_break_layouts(struct inode *inode, uint *iolock, bool with_imutex)
+xfs_break_layouts(struct inode *inode, uint *iolock)
 {
 	return 0;
 }

fs/xfs/xfs_qm.c

@@ -1135,7 +1135,7 @@ xfs_qm_get_rtblks(
 			return error;
 	}
 	rtblks = 0;
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	nextents = xfs_iext_count(ifp);
 	for (idx = 0; idx < nextents; idx++)
 		rtblks += xfs_bmbt_get_blockcount(xfs_iext_get_ext(ifp, idx));
 	*O_rtblks = (xfs_qcnt_t)rtblks;

fs/xfs/xfs_reflink.c

@@ -243,12 +243,11 @@ xfs_reflink_reserve_cow(
 	struct xfs_bmbt_irec	*imap,
 	bool			*shared)
 {
-	struct xfs_bmbt_irec	got, prev;
-	xfs_fileoff_t		end_fsb, orig_end_fsb;
-	int			eof = 0, error = 0;
-	bool			trimmed;
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, XFS_COW_FORK);
+	struct xfs_bmbt_irec	got;
+	int			error = 0;
+	bool			eof = false, trimmed;
 	xfs_extnum_t		idx;
-	xfs_extlen_t		align;
 
 	/*
 	 * Search the COW fork extent list first.  This serves two purposes:
@@ -258,8 +257,9 @@ xfs_reflink_reserve_cow(
 	 * extent list is generally faster than going out to the shared extent
 	 * tree.
 	 */
-	xfs_bmap_search_extents(ip, imap->br_startoff, XFS_COW_FORK, &eof, &idx,
-			&got, &prev);
+
+	if (!xfs_iext_lookup_extent(ip, ifp, imap->br_startoff, &idx, &got))
+		eof = true;
 	if (!eof && got.br_startoff <= imap->br_startoff) {
 		trace_xfs_reflink_cow_found(ip, imap);
 		xfs_trim_extent(imap, got.br_startoff, got.br_blockcount);
@@ -285,33 +285,12 @@ xfs_reflink_reserve_cow(
 	if (error)
 		return error;
 
-	end_fsb = orig_end_fsb = imap->br_startoff + imap->br_blockcount;
-
-	align = xfs_eof_alignment(ip, xfs_get_cowextsz_hint(ip));
-	if (align)
-		end_fsb = roundup_64(end_fsb, align);
-
-retry:
 	error = xfs_bmapi_reserve_delalloc(ip, XFS_COW_FORK, imap->br_startoff,
-			end_fsb - imap->br_startoff, &got, &prev, &idx, eof);
-	switch (error) {
-	case 0:
-		break;
-	case -ENOSPC:
-	case -EDQUOT:
-		/* retry without any preallocation */
+			imap->br_blockcount, 0, &got, &idx, eof);
+	if (error == -ENOSPC || error == -EDQUOT)
 		trace_xfs_reflink_cow_enospc(ip, imap);
-		if (end_fsb != orig_end_fsb) {
-			end_fsb = orig_end_fsb;
-			goto retry;
-		}
-		/*FALLTHRU*/
-	default:
+	if (error)
 		return error;
-	}
-
-	if (end_fsb != orig_end_fsb)
-		xfs_inode_set_cowblocks_tag(ip);
 
 	trace_xfs_reflink_cow_alloc(ip, &got);
 	return 0;
@@ -418,87 +397,65 @@ xfs_reflink_allocate_cow_range(
 }
 
 /*
- * Find the CoW reservation (and whether or not it needs block allocation)
- * for a given byte offset of a file.
+ * Find the CoW reservation for a given byte offset of a file.
  */
 bool
 xfs_reflink_find_cow_mapping(
 	struct xfs_inode		*ip,
 	xfs_off_t			offset,
-	struct xfs_bmbt_irec		*imap,
-	bool				*need_alloc)
+	struct xfs_bmbt_irec		*imap)
 {
-	struct xfs_bmbt_irec		irec;
-	struct xfs_ifork		*ifp;
-	struct xfs_bmbt_rec_host	*gotp;
-	xfs_fileoff_t			bno;
+	struct xfs_ifork		*ifp = XFS_IFORK_PTR(ip, XFS_COW_FORK);
+	xfs_fileoff_t			offset_fsb;
+	struct xfs_bmbt_irec		got;
 	xfs_extnum_t			idx;
 
 	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL | XFS_ILOCK_SHARED));
 	ASSERT(xfs_is_reflink_inode(ip));
 
-	/* Find the extent in the CoW fork. */
-	ifp = XFS_IFORK_PTR(ip, XFS_COW_FORK);
-	bno = XFS_B_TO_FSBT(ip->i_mount, offset);
-	gotp = xfs_iext_bno_to_ext(ifp, bno, &idx);
-	if (!gotp)
+	offset_fsb = XFS_B_TO_FSBT(ip->i_mount, offset);
+	if (!xfs_iext_lookup_extent(ip, ifp, offset_fsb, &idx, &got))
 		return false;
-
-	xfs_bmbt_get_all(gotp, &irec);
-	if (bno >= irec.br_startoff + irec.br_blockcount ||
-	    bno < irec.br_startoff)
+	if (got.br_startoff > offset_fsb)
 		return false;
 
 	trace_xfs_reflink_find_cow_mapping(ip, offset, 1, XFS_IO_OVERWRITE,
-			&irec);
-
-	/* If it's still delalloc, we must allocate later. */
-	*imap = irec;
-	*need_alloc = !!(isnullstartblock(irec.br_startblock));
-
+			&got);
+	*imap = got;
 	return true;
 }
 
 /*
  * Trim an extent to end at the next CoW reservation past offset_fsb.
  */
-int
+void
 xfs_reflink_trim_irec_to_next_cow(
 	struct xfs_inode		*ip,
 	xfs_fileoff_t			offset_fsb,
 	struct xfs_bmbt_irec		*imap)
 {
-	struct xfs_bmbt_irec		irec;
-	struct xfs_ifork		*ifp;
-	struct xfs_bmbt_rec_host	*gotp;
+	struct xfs_ifork		*ifp = XFS_IFORK_PTR(ip, XFS_COW_FORK);
+	struct xfs_bmbt_irec		got;
 	xfs_extnum_t			idx;
 
 	if (!xfs_is_reflink_inode(ip))
-		return 0;
+		return;
 
 	/* Find the extent in the CoW fork. */
-	ifp = XFS_IFORK_PTR(ip, XFS_COW_FORK);
-	gotp = xfs_iext_bno_to_ext(ifp, offset_fsb, &idx);
-	if (!gotp)
-		return 0;
-	xfs_bmbt_get_all(gotp, &irec);
+	if (!xfs_iext_lookup_extent(ip, ifp, offset_fsb, &idx, &got))
+		return;
 
 	/* This is the extent before; try sliding up one. */
-	if (irec.br_startoff < offset_fsb) {
-		idx++;
-		if (idx >= ifp->if_bytes / sizeof(xfs_bmbt_rec_t))
-			return 0;
-		gotp = xfs_iext_get_ext(ifp, idx);
-		xfs_bmbt_get_all(gotp, &irec);
+	if (got.br_startoff < offset_fsb) {
+		if (!xfs_iext_get_extent(ifp, idx + 1, &got))
+			return;
 	}
 
-	if (irec.br_startoff >= imap->br_startoff + imap->br_blockcount)
-		return 0;
+	if (got.br_startoff >= imap->br_startoff + imap->br_blockcount)
+		return;
 
-	imap->br_blockcount = irec.br_startoff - imap->br_startoff;
+	imap->br_blockcount = got.br_startoff - imap->br_startoff;
 	trace_xfs_reflink_trim_irec(ip, imap);
-
-	return 0;
 }
 
 /*
@@ -512,18 +469,15 @@ xfs_reflink_cancel_cow_blocks(
 	xfs_fileoff_t			end_fsb)
 {
 	struct xfs_ifork		*ifp = XFS_IFORK_PTR(ip, XFS_COW_FORK);
-	struct xfs_bmbt_irec		got, prev, del;
+	struct xfs_bmbt_irec		got, del;
 	xfs_extnum_t			idx;
 	xfs_fsblock_t			firstfsb;
 	struct xfs_defer_ops		dfops;
-	int				error = 0, eof = 0;
+	int				error = 0;
 
 	if (!xfs_is_reflink_inode(ip))
 		return 0;
-
-	xfs_bmap_search_extents(ip, offset_fsb, XFS_COW_FORK, &eof, &idx,
-			&got, &prev);
-	if (eof)
+	if (!xfs_iext_lookup_extent(ip, ifp, offset_fsb, &idx, &got))
 		return 0;
 
 	while (got.br_startoff < end_fsb) {
@@ -566,9 +520,8 @@ xfs_reflink_cancel_cow_blocks(
 			xfs_bmap_del_extent_cow(ip, &idx, &got, &del);
 		}
 
-		if (++idx >= ifp->if_bytes / sizeof(struct xfs_bmbt_rec))
+		if (!xfs_iext_get_extent(ifp, ++idx, &got))
 			break;
-		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx), &got);
 	}
 
 	/* clear tag if cow fork is emptied */
@@ -638,13 +591,13 @@ xfs_reflink_end_cow(
 	xfs_off_t			count)
 {
 	struct xfs_ifork		*ifp = XFS_IFORK_PTR(ip, XFS_COW_FORK);
-	struct xfs_bmbt_irec		got, prev, del;
+	struct xfs_bmbt_irec		got, del;
 	struct xfs_trans		*tp;
 	xfs_fileoff_t			offset_fsb;
 	xfs_fileoff_t			end_fsb;
 	xfs_fsblock_t			firstfsb;
 	struct xfs_defer_ops		dfops;
-	int				error, eof = 0;
+	int				error;
 	unsigned int			resblks;
 	xfs_filblks_t			rlen;
 	xfs_extnum_t			idx;
@@ -668,13 +621,11 @@ xfs_reflink_end_cow(
 	xfs_ilock(ip, XFS_ILOCK_EXCL);
 	xfs_trans_ijoin(tp, ip, 0);
 
-	xfs_bmap_search_extents(ip, end_fsb - 1, XFS_COW_FORK, &eof, &idx,
-			&got, &prev);
-
 	/* If there is a hole at end_fsb - 1 go to the previous extent */
-	if (eof || got.br_startoff > end_fsb) {
+	if (!xfs_iext_lookup_extent(ip, ifp, end_fsb - 1, &idx, &got) ||
+	    got.br_startoff > end_fsb) {
 		ASSERT(idx > 0);
-		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, --idx), &got);
+		xfs_iext_get_extent(ifp, --idx, &got);
 	}
 
 	/* Walk backwards until we're out of the I/O range... */
@@ -722,11 +673,9 @@ xfs_reflink_end_cow(
 		error = xfs_defer_finish(&tp, &dfops, ip);
 		if (error)
 			goto out_defer;
-
 next_extent:
-		if (idx < 0)
+		if (!xfs_iext_get_extent(ifp, idx, &got))
 			break;
-		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx), &got);
 	}
 
 	error = xfs_trans_commit(tp);
@@ -1302,13 +1251,11 @@ xfs_reflink_remap_range(
 		return -EIO;
 
 	/* Lock both files against IO */
-	if (same_inode) {
-		xfs_ilock(src, XFS_IOLOCK_EXCL);
+	lock_two_nondirectories(inode_in, inode_out);
+	if (same_inode)
 		xfs_ilock(src, XFS_MMAPLOCK_EXCL);
-	} else {
-		xfs_lock_two_inodes(src, dest, XFS_IOLOCK_EXCL);
+	else
 		xfs_lock_two_inodes(src, dest, XFS_MMAPLOCK_EXCL);
-	}
 
 	/* Don't touch certain kinds of inodes */
 	ret = -EPERM;
@@ -1345,8 +1292,14 @@ xfs_reflink_remap_range(
 		goto out_unlock;
 	}
 
-	if (len == 0)
+	/* Zero length dedupe exits immediately; reflink goes to EOF. */
+	if (len == 0) {
+		if (is_dedupe) {
+			ret = 0;
+			goto out_unlock;
+		}
 		len = isize - pos_in;
+	}
 
 	/* Ensure offsets don't wrap and the input is inside i_size */
 	if (pos_in + len < pos_in || pos_out + len < pos_out ||
@@ -1447,11 +1400,9 @@ xfs_reflink_remap_range(
 
 out_unlock:
 	xfs_iunlock(src, XFS_MMAPLOCK_EXCL);
-	xfs_iunlock(src, XFS_IOLOCK_EXCL);
-	if (src->i_ino != dest->i_ino) {
+	if (!same_inode)
 		xfs_iunlock(dest, XFS_MMAPLOCK_EXCL);
-		xfs_iunlock(dest, XFS_IOLOCK_EXCL);
-	}
+	unlock_two_nondirectories(inode_in, inode_out);
 	if (ret)
 		trace_xfs_reflink_remap_range_error(dest, ret, _RET_IP_);
 	return ret;
@@ -1697,37 +1648,3 @@ out:
 	trace_xfs_reflink_unshare_error(ip, error, _RET_IP_);
 	return error;
 }
-
-/*
- * Does this inode have any real CoW reservations?
- */
-bool
-xfs_reflink_has_real_cow_blocks(
-	struct xfs_inode		*ip)
-{
-	struct xfs_bmbt_irec		irec;
-	struct xfs_ifork		*ifp;
-	struct xfs_bmbt_rec_host	*gotp;
-	xfs_extnum_t			idx;
-
-	if (!xfs_is_reflink_inode(ip))
-		return false;
-
-	/* Go find the old extent in the CoW fork. */
-	ifp = XFS_IFORK_PTR(ip, XFS_COW_FORK);
-	gotp = xfs_iext_bno_to_ext(ifp, 0, &idx);
-	while (gotp) {
-		xfs_bmbt_get_all(gotp, &irec);
-
-		if (!isnullstartblock(irec.br_startblock))
-			return true;
-
-		/* Roll on... */
-		idx++;
-		if (idx >= ifp->if_bytes / sizeof(xfs_bmbt_rec_t))
-			break;
-		gotp = xfs_iext_get_ext(ifp, idx);
-	}
-
-	return false;
-}

fs/xfs/xfs_reflink.h

@@ -31,8 +31,8 @@ extern int xfs_reflink_reserve_cow(struct xfs_inode *ip,
 extern int xfs_reflink_allocate_cow_range(struct xfs_inode *ip,
 		xfs_off_t offset, xfs_off_t count);
 extern bool xfs_reflink_find_cow_mapping(struct xfs_inode *ip, xfs_off_t offset,
-		struct xfs_bmbt_irec *imap, bool *need_alloc);
-extern int xfs_reflink_trim_irec_to_next_cow(struct xfs_inode *ip,
+		struct xfs_bmbt_irec *imap);
+extern void xfs_reflink_trim_irec_to_next_cow(struct xfs_inode *ip,
 		xfs_fileoff_t offset_fsb, struct xfs_bmbt_irec *imap);
 
 extern int xfs_reflink_cancel_cow_blocks(struct xfs_inode *ip,
@@ -50,6 +50,4 @@ extern int xfs_reflink_clear_inode_flag(struct xfs_inode *ip,
 extern int xfs_reflink_unshare(struct xfs_inode *ip, xfs_off_t offset,
 		xfs_off_t len);
 
-extern bool xfs_reflink_has_real_cow_blocks(struct xfs_inode *ip);
-
 #endif /* __XFS_REFLINK_H */

fs/xfs/xfs_stats.c

@@ -80,9 +80,9 @@ int xfs_stats_format(struct xfsstats __percpu *stats, char *buf)
 	}
 	/* extra precision counters */
 	for_each_possible_cpu(i) {
-		xs_xstrat_bytes += per_cpu_ptr(stats, i)->xs_xstrat_bytes;
-		xs_write_bytes += per_cpu_ptr(stats, i)->xs_write_bytes;
-		xs_read_bytes += per_cpu_ptr(stats, i)->xs_read_bytes;
+		xs_xstrat_bytes += per_cpu_ptr(stats, i)->s.xs_xstrat_bytes;
+		xs_write_bytes += per_cpu_ptr(stats, i)->s.xs_write_bytes;
+		xs_read_bytes += per_cpu_ptr(stats, i)->s.xs_read_bytes;
 	}
 
 	len += snprintf(buf + len, PATH_MAX-len, "xpc %Lu %Lu %Lu\n",
@@ -106,9 +106,9 @@ void xfs_stats_clearall(struct xfsstats __percpu *stats)
 	for_each_possible_cpu(c) {
 		preempt_disable();
 		/* save vn_active, it's a universal truth! */
-		vn_active = per_cpu_ptr(stats, c)->vn_active;
+		vn_active = per_cpu_ptr(stats, c)->s.vn_active;
 		memset(per_cpu_ptr(stats, c), 0, sizeof(*stats));
-		per_cpu_ptr(stats, c)->vn_active = vn_active;
+		per_cpu_ptr(stats, c)->s.vn_active = vn_active;
 		preempt_enable();
 	}
 }

fs/xfs/xfs_stats.h

@@ -21,10 +21,38 @@
 
 #include <linux/percpu.h>
 
+/*
+ * The btree stats arrays have fixed offsets for the different stats. We
+ * store the base index in the btree cursor via XFS_STATS_CALC_INDEX() and
+ * that allows us to use fixed offsets into the stats array for each btree
+ * stat. These index offsets are defined in the order they will be emitted
+ * in the stats files, so it is possible to add new btree stat types by
+ * appending to the enum list below.
+ */
+enum {
+	__XBTS_lookup = 0,
+	__XBTS_compare = 1,
+	__XBTS_insrec = 2,
+	__XBTS_delrec = 3,
+	__XBTS_newroot = 4,
+	__XBTS_killroot = 5,
+	__XBTS_increment = 6,
+	__XBTS_decrement = 7,
+	__XBTS_lshift = 8,
+	__XBTS_rshift = 9,
+	__XBTS_split = 10,
+	__XBTS_join = 11,
+	__XBTS_alloc = 12,
+	__XBTS_free = 13,
+	__XBTS_moves = 14,
+
+	__XBTS_MAX = 15,
+};
+
 /*
  * XFS global statistics
  */
-struct xfsstats {
+struct __xfsstats {
 # define XFSSTAT_END_EXTENT_ALLOC	4
 	__uint32_t		xs_allocx;
 	__uint32_t		xs_allocb;
@@ -117,118 +145,20 @@ struct xfsstats {
 	__uint32_t		xb_page_found;
 	__uint32_t		xb_get_read;
 /* Version 2 btree counters */
-#define XFSSTAT_END_ABTB_V2		(XFSSTAT_END_BUF+15)
-	__uint32_t		xs_abtb_2_lookup;
-	__uint32_t		xs_abtb_2_compare;
-	__uint32_t		xs_abtb_2_insrec;
-	__uint32_t		xs_abtb_2_delrec;
-	__uint32_t		xs_abtb_2_newroot;
-	__uint32_t		xs_abtb_2_killroot;
-	__uint32_t		xs_abtb_2_increment;
-	__uint32_t		xs_abtb_2_decrement;
-	__uint32_t		xs_abtb_2_lshift;
-	__uint32_t		xs_abtb_2_rshift;
-	__uint32_t		xs_abtb_2_split;
-	__uint32_t		xs_abtb_2_join;
-	__uint32_t		xs_abtb_2_alloc;
-	__uint32_t		xs_abtb_2_free;
-	__uint32_t		xs_abtb_2_moves;
-#define XFSSTAT_END_ABTC_V2		(XFSSTAT_END_ABTB_V2+15)
-	__uint32_t		xs_abtc_2_lookup;
-	__uint32_t		xs_abtc_2_compare;
-	__uint32_t		xs_abtc_2_insrec;
-	__uint32_t		xs_abtc_2_delrec;
-	__uint32_t		xs_abtc_2_newroot;
-	__uint32_t		xs_abtc_2_killroot;
-	__uint32_t		xs_abtc_2_increment;
-	__uint32_t		xs_abtc_2_decrement;
-	__uint32_t		xs_abtc_2_lshift;
-	__uint32_t		xs_abtc_2_rshift;
-	__uint32_t		xs_abtc_2_split;
-	__uint32_t		xs_abtc_2_join;
-	__uint32_t		xs_abtc_2_alloc;
-	__uint32_t		xs_abtc_2_free;
-	__uint32_t		xs_abtc_2_moves;
-#define XFSSTAT_END_BMBT_V2		(XFSSTAT_END_ABTC_V2+15)
-	__uint32_t		xs_bmbt_2_lookup;
-	__uint32_t		xs_bmbt_2_compare;
-	__uint32_t		xs_bmbt_2_insrec;
-	__uint32_t		xs_bmbt_2_delrec;
-	__uint32_t		xs_bmbt_2_newroot;
-	__uint32_t		xs_bmbt_2_killroot;
-	__uint32_t		xs_bmbt_2_increment;
-	__uint32_t		xs_bmbt_2_decrement;
-	__uint32_t		xs_bmbt_2_lshift;
-	__uint32_t		xs_bmbt_2_rshift;
-	__uint32_t		xs_bmbt_2_split;
-	__uint32_t		xs_bmbt_2_join;
-	__uint32_t		xs_bmbt_2_alloc;
-	__uint32_t		xs_bmbt_2_free;
-	__uint32_t		xs_bmbt_2_moves;
-#define XFSSTAT_END_IBT_V2		(XFSSTAT_END_BMBT_V2+15)
-	__uint32_t		xs_ibt_2_lookup;
-	__uint32_t		xs_ibt_2_compare;
-	__uint32_t		xs_ibt_2_insrec;
-	__uint32_t		xs_ibt_2_delrec;
-	__uint32_t		xs_ibt_2_newroot;
-	__uint32_t		xs_ibt_2_killroot;
-	__uint32_t		xs_ibt_2_increment;
-	__uint32_t		xs_ibt_2_decrement;
-	__uint32_t		xs_ibt_2_lshift;
-	__uint32_t		xs_ibt_2_rshift;
-	__uint32_t		xs_ibt_2_split;
-	__uint32_t		xs_ibt_2_join;
-	__uint32_t		xs_ibt_2_alloc;
-	__uint32_t		xs_ibt_2_free;
-	__uint32_t		xs_ibt_2_moves;
-#define XFSSTAT_END_FIBT_V2		(XFSSTAT_END_IBT_V2+15)
-	__uint32_t		xs_fibt_2_lookup;
-	__uint32_t		xs_fibt_2_compare;
-	__uint32_t		xs_fibt_2_insrec;
-	__uint32_t		xs_fibt_2_delrec;
-	__uint32_t		xs_fibt_2_newroot;
-	__uint32_t		xs_fibt_2_killroot;
-	__uint32_t		xs_fibt_2_increment;
-	__uint32_t		xs_fibt_2_decrement;
-	__uint32_t		xs_fibt_2_lshift;
-	__uint32_t		xs_fibt_2_rshift;
-	__uint32_t		xs_fibt_2_split;
-	__uint32_t		xs_fibt_2_join;
-	__uint32_t		xs_fibt_2_alloc;
-	__uint32_t		xs_fibt_2_free;
-	__uint32_t		xs_fibt_2_moves;
-#define XFSSTAT_END_RMAP_V2		(XFSSTAT_END_FIBT_V2+15)
-	__uint32_t		xs_rmap_2_lookup;
-	__uint32_t		xs_rmap_2_compare;
-	__uint32_t		xs_rmap_2_insrec;
-	__uint32_t		xs_rmap_2_delrec;
-	__uint32_t		xs_rmap_2_newroot;
-	__uint32_t		xs_rmap_2_killroot;
-	__uint32_t		xs_rmap_2_increment;
-	__uint32_t		xs_rmap_2_decrement;
-	__uint32_t		xs_rmap_2_lshift;
-	__uint32_t		xs_rmap_2_rshift;
-	__uint32_t		xs_rmap_2_split;
-	__uint32_t		xs_rmap_2_join;
-	__uint32_t		xs_rmap_2_alloc;
-	__uint32_t		xs_rmap_2_free;
-	__uint32_t		xs_rmap_2_moves;
-#define XFSSTAT_END_REFCOUNT		(XFSSTAT_END_RMAP_V2 + 15)
-	__uint32_t		xs_refcbt_2_lookup;
-	__uint32_t		xs_refcbt_2_compare;
-	__uint32_t		xs_refcbt_2_insrec;
-	__uint32_t		xs_refcbt_2_delrec;
-	__uint32_t		xs_refcbt_2_newroot;
-	__uint32_t		xs_refcbt_2_killroot;
-	__uint32_t		xs_refcbt_2_increment;
-	__uint32_t		xs_refcbt_2_decrement;
-	__uint32_t		xs_refcbt_2_lshift;
-	__uint32_t		xs_refcbt_2_rshift;
-	__uint32_t		xs_refcbt_2_split;
-	__uint32_t		xs_refcbt_2_join;
-	__uint32_t		xs_refcbt_2_alloc;
-	__uint32_t		xs_refcbt_2_free;
-	__uint32_t		xs_refcbt_2_moves;
+#define XFSSTAT_END_ABTB_V2		(XFSSTAT_END_BUF + __XBTS_MAX)
+	__uint32_t		xs_abtb_2[__XBTS_MAX];
+#define XFSSTAT_END_ABTC_V2		(XFSSTAT_END_ABTB_V2 + __XBTS_MAX)
+	__uint32_t		xs_abtc_2[__XBTS_MAX];
+#define XFSSTAT_END_BMBT_V2		(XFSSTAT_END_ABTC_V2 + __XBTS_MAX)
+	__uint32_t		xs_bmbt_2[__XBTS_MAX];
+#define XFSSTAT_END_IBT_V2		(XFSSTAT_END_BMBT_V2 + __XBTS_MAX)
+	__uint32_t		xs_ibt_2[__XBTS_MAX];
+#define XFSSTAT_END_FIBT_V2		(XFSSTAT_END_IBT_V2 + __XBTS_MAX)
+	__uint32_t		xs_fibt_2[__XBTS_MAX];
+#define XFSSTAT_END_RMAP_V2		(XFSSTAT_END_FIBT_V2 + __XBTS_MAX)
+	__uint32_t		xs_rmap_2[__XBTS_MAX];
+#define XFSSTAT_END_REFCOUNT		(XFSSTAT_END_RMAP_V2 + __XBTS_MAX)
+	__uint32_t		xs_refcbt_2[__XBTS_MAX];
 #define XFSSTAT_END_XQMSTAT		(XFSSTAT_END_REFCOUNT + 6)
 	__uint32_t		xs_qm_dqreclaims;
 	__uint32_t		xs_qm_dqreclaim_misses;
@@ -245,26 +175,58 @@ struct xfsstats {
 	__uint64_t		xs_read_bytes;
 };
 
+struct xfsstats {
+	union {
+		struct __xfsstats	s;
+		uint32_t		a[XFSSTAT_END_XQMSTAT];
+	};
+};
+
+/*
+ * simple wrapper for getting the array index of s struct member offset
+ */
+#define XFS_STATS_CALC_INDEX(member)	\
+	(offsetof(struct __xfsstats, member) / (int)sizeof(__uint32_t))
+
+
 int xfs_stats_format(struct xfsstats __percpu *stats, char *buf);
 void xfs_stats_clearall(struct xfsstats __percpu *stats);
 extern struct xstats xfsstats;
 
 #define XFS_STATS_INC(mp, v)					\
 do {								\
-	per_cpu_ptr(xfsstats.xs_stats, current_cpu())->v++;	\
-	per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->v++;	\
+	per_cpu_ptr(xfsstats.xs_stats, current_cpu())->s.v++;	\
+	per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->s.v++;	\
 } while (0)
 
 #define XFS_STATS_DEC(mp, v)					\
 do {								\
-	per_cpu_ptr(xfsstats.xs_stats, current_cpu())->v--;	\
-	per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->v--;	\
+	per_cpu_ptr(xfsstats.xs_stats, current_cpu())->s.v--;	\
+	per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->s.v--;	\
 } while (0)
 
 #define XFS_STATS_ADD(mp, v, inc)					\
 do {									\
-	per_cpu_ptr(xfsstats.xs_stats, current_cpu())->v += (inc);	\
-	per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->v += (inc);	\
+	per_cpu_ptr(xfsstats.xs_stats, current_cpu())->s.v += (inc);	\
+	per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->s.v += (inc);	\
+} while (0)
+
+#define XFS_STATS_INC_OFF(mp, off)				\
+do {								\
+	per_cpu_ptr(xfsstats.xs_stats, current_cpu())->a[off]++;	\
+	per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->a[off]++;	\
+} while (0)
+
+#define XFS_STATS_DEC_OFF(mp, off)					\
+do {								\
+	per_cpu_ptr(xfsstats.xs_stats, current_cpu())->a[off];	\
+	per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->a[off];	\
+} while (0)
+
+#define XFS_STATS_ADD_OFF(mp, off, inc)					\
+do {									\
+	per_cpu_ptr(xfsstats.xs_stats, current_cpu())->a[off] += (inc);	\
+	per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->a[off] += (inc);	\
 } while (0)
 
 #if defined(CONFIG_PROC_FS)

fs/xfs/xfs_super.c

@@ -104,9 +104,6 @@ static const match_table_t tokens = {
 	{Opt_sysvgroups,"sysvgroups"},	/* group-ID from current process */
 	{Opt_allocsize,	"allocsize=%s"},/* preferred allocation size */
 	{Opt_norecovery,"norecovery"},	/* don't run XFS recovery */
-	{Opt_barrier,	"barrier"},	/* use writer barriers for log write and
-					 * unwritten extent conversion */
-	{Opt_nobarrier,	"nobarrier"},	/* .. disable */
 	{Opt_inode64,	"inode64"},	/* inodes can be allocated anywhere */
 	{Opt_inode32,   "inode32"},	/* inode allocation limited to
 					 * XFS_MAXINUMBER_32 */
@@ -134,6 +131,12 @@ static const match_table_t tokens = {
 	{Opt_nodiscard,	"nodiscard"},	/* Do not discard unused blocks */
 
 	{Opt_dax,	"dax"},		/* Enable direct access to bdev pages */
+
+	/* Deprecated mount options scheduled for removal */
+	{Opt_barrier,	"barrier"},	/* use writer barriers for log write and
+					 * unwritten extent conversion */
+	{Opt_nobarrier,	"nobarrier"},	/* .. disable */
+
 	{Opt_err,	NULL},
 };
 
@@ -301,12 +304,6 @@ xfs_parseargs(
 		case Opt_nouuid:
 			mp->m_flags |= XFS_MOUNT_NOUUID;
 			break;
-		case Opt_barrier:
-			mp->m_flags |= XFS_MOUNT_BARRIER;
-			break;
-		case Opt_nobarrier:
-			mp->m_flags &= ~XFS_MOUNT_BARRIER;
-			break;
 		case Opt_ikeep:
 			mp->m_flags |= XFS_MOUNT_IKEEP;
 			break;
@@ -374,6 +371,14 @@ xfs_parseargs(
 			mp->m_flags |= XFS_MOUNT_DAX;
 			break;
 #endif
+		case Opt_barrier:
+			xfs_warn(mp, "%s option is deprecated, ignoring.", p);
+			mp->m_flags |= XFS_MOUNT_BARRIER;
+			break;
+		case Opt_nobarrier:
+			xfs_warn(mp, "%s option is deprecated, ignoring.", p);
+			mp->m_flags &= ~XFS_MOUNT_BARRIER;
+			break;
 		default:
 			xfs_warn(mp, "unknown mount option [%s].", p);
 			return -EINVAL;
@@ -943,7 +948,7 @@ xfs_fs_destroy_inode(
 
 	trace_xfs_destroy_inode(ip);
 
-	ASSERT(!rwsem_is_locked(&ip->i_iolock.mr_lock));
+	ASSERT(!rwsem_is_locked(&inode->i_rwsem));
 	XFS_STATS_INC(ip->i_mount, vn_rele);
 	XFS_STATS_INC(ip->i_mount, vn_remove);
 
@@ -1238,9 +1243,11 @@ xfs_fs_remount(
 		token = match_token(p, tokens, args);
 		switch (token) {
 		case Opt_barrier:
+			xfs_warn(mp, "%s option is deprecated, ignoring.", p);
 			mp->m_flags |= XFS_MOUNT_BARRIER;
 			break;
 		case Opt_nobarrier:
+			xfs_warn(mp, "%s option is deprecated, ignoring.", p);
 			mp->m_flags &= ~XFS_MOUNT_BARRIER;
 			break;
 		case Opt_inode64:

fs/xfs/xfs_symlink.c

@@ -238,8 +238,7 @@ xfs_symlink(
 	if (error)
 		goto out_release_inode;
 
-	xfs_ilock(dp, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL |
-		      XFS_IOLOCK_PARENT | XFS_ILOCK_PARENT);
+	xfs_ilock(dp, XFS_ILOCK_EXCL | XFS_ILOCK_PARENT);
 	unlock_dp_on_error = true;
 
 	/*
@@ -287,7 +286,7 @@ xfs_symlink(
 	 * the transaction cancel unlocking dp so don't do it explicitly in the
 	 * error path.
 	 */
-	xfs_trans_ijoin(tp, dp, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL);
+	xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL);
 	unlock_dp_on_error = false;
 
 	/*
@@ -412,7 +411,7 @@ out_release_inode:
 	xfs_qm_dqrele(pdqp);
 
 	if (unlock_dp_on_error)
-		xfs_iunlock(dp, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL);
+		xfs_iunlock(dp, XFS_ILOCK_EXCL);
 	return error;
 }
 

fs/xfs/xfs_trace.h

@@ -355,7 +355,6 @@ DEFINE_BUF_EVENT(xfs_buf_rele);
 DEFINE_BUF_EVENT(xfs_buf_iodone);
 DEFINE_BUF_EVENT(xfs_buf_submit);
 DEFINE_BUF_EVENT(xfs_buf_submit_wait);
-DEFINE_BUF_EVENT(xfs_buf_bawrite);
 DEFINE_BUF_EVENT(xfs_buf_lock);
 DEFINE_BUF_EVENT(xfs_buf_lock_done);
 DEFINE_BUF_EVENT(xfs_buf_trylock_fail);
@@ -367,19 +366,15 @@ DEFINE_BUF_EVENT(xfs_buf_delwri_queue);
 DEFINE_BUF_EVENT(xfs_buf_delwri_queued);
 DEFINE_BUF_EVENT(xfs_buf_delwri_split);
 DEFINE_BUF_EVENT(xfs_buf_get_uncached);
-DEFINE_BUF_EVENT(xfs_bdstrat_shut);
 DEFINE_BUF_EVENT(xfs_buf_item_relse);
 DEFINE_BUF_EVENT(xfs_buf_item_iodone_async);
 DEFINE_BUF_EVENT(xfs_buf_error_relse);
 DEFINE_BUF_EVENT(xfs_buf_wait_buftarg);
-DEFINE_BUF_EVENT(xfs_trans_read_buf_io);
 DEFINE_BUF_EVENT(xfs_trans_read_buf_shut);
 
 /* not really buffer traces, but the buf provides useful information */
 DEFINE_BUF_EVENT(xfs_btree_corrupt);
-DEFINE_BUF_EVENT(xfs_da_btree_corrupt);
 DEFINE_BUF_EVENT(xfs_reset_dqcounts);
-DEFINE_BUF_EVENT(xfs_inode_item_push);
 
 /* pass flags explicitly */
 DECLARE_EVENT_CLASS(xfs_buf_flags_class,
@@ -541,7 +536,6 @@ DEFINE_BUF_ITEM_EVENT(xfs_trans_bjoin);
 DEFINE_BUF_ITEM_EVENT(xfs_trans_bhold);
 DEFINE_BUF_ITEM_EVENT(xfs_trans_bhold_release);
 DEFINE_BUF_ITEM_EVENT(xfs_trans_binval);
-DEFINE_BUF_ITEM_EVENT(xfs_trans_buf_ordered);
 
 DECLARE_EVENT_CLASS(xfs_filestream_class,
 	TP_PROTO(struct xfs_inode *ip, xfs_agnumber_t agno),
@@ -680,7 +674,6 @@ DEFINE_INODE_EVENT(xfs_ioctl_setattr);
 DEFINE_INODE_EVENT(xfs_dir_fsync);
 DEFINE_INODE_EVENT(xfs_file_fsync);
 DEFINE_INODE_EVENT(xfs_destroy_inode);
-DEFINE_INODE_EVENT(xfs_evict_inode);
 DEFINE_INODE_EVENT(xfs_update_time);
 
 DEFINE_INODE_EVENT(xfs_dquot_dqalloc);
@@ -798,7 +791,6 @@ TRACE_EVENT(xfs_irec_merge_post,
 DEFINE_EVENT(xfs_iref_class, name, \
 	TP_PROTO(struct xfs_inode *ip, unsigned long caller_ip), \
 	TP_ARGS(ip, caller_ip))
-DEFINE_IREF_EVENT(xfs_ihold);
 DEFINE_IREF_EVENT(xfs_irele);
 DEFINE_IREF_EVENT(xfs_inode_pin);
 DEFINE_IREF_EVENT(xfs_inode_unpin);
@@ -939,7 +931,6 @@ DEFINE_DQUOT_EVENT(xfs_dqget_miss);
 DEFINE_DQUOT_EVENT(xfs_dqget_freeing);
 DEFINE_DQUOT_EVENT(xfs_dqget_dup);
 DEFINE_DQUOT_EVENT(xfs_dqput);
-DEFINE_DQUOT_EVENT(xfs_dqput_wait);
 DEFINE_DQUOT_EVENT(xfs_dqput_free);
 DEFINE_DQUOT_EVENT(xfs_dqrele);
 DEFINE_DQUOT_EVENT(xfs_dqflush);
@@ -1815,7 +1806,6 @@ DEFINE_ATTR_EVENT(xfs_attr_sf_addname);
 DEFINE_ATTR_EVENT(xfs_attr_sf_create);
 DEFINE_ATTR_EVENT(xfs_attr_sf_lookup);
 DEFINE_ATTR_EVENT(xfs_attr_sf_remove);
-DEFINE_ATTR_EVENT(xfs_attr_sf_removename);
 DEFINE_ATTR_EVENT(xfs_attr_sf_to_leaf);
 
 DEFINE_ATTR_EVENT(xfs_attr_leaf_add);
@@ -1844,7 +1834,6 @@ DEFINE_ATTR_EVENT(xfs_attr_leaf_toosmall);
 
 DEFINE_ATTR_EVENT(xfs_attr_node_addname);
 DEFINE_ATTR_EVENT(xfs_attr_node_get);
-DEFINE_ATTR_EVENT(xfs_attr_node_lookup);
 DEFINE_ATTR_EVENT(xfs_attr_node_replace);
 DEFINE_ATTR_EVENT(xfs_attr_node_removename);
 
@@ -2440,11 +2429,9 @@ DEFINE_DEFER_EVENT(xfs_defer_finish_done);
 
 DEFINE_DEFER_ERROR_EVENT(xfs_defer_trans_roll_error);
 DEFINE_DEFER_ERROR_EVENT(xfs_defer_finish_error);
-DEFINE_DEFER_ERROR_EVENT(xfs_defer_op_finish_error);
 
 DEFINE_DEFER_PENDING_EVENT(xfs_defer_intake_work);
 DEFINE_DEFER_PENDING_EVENT(xfs_defer_intake_cancel);
-DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_commit);
 DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_cancel);
 DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_finish);
 DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_abort);
@@ -3092,87 +3079,6 @@ DEFINE_EVENT(xfs_double_io_class, name,	\
 		 struct xfs_inode *dest, xfs_off_t doffset), \
 	TP_ARGS(src, soffset, len, dest, doffset))
 
-/* two-file vfs io tracepoint class */
-DECLARE_EVENT_CLASS(xfs_double_vfs_io_class,
-	TP_PROTO(struct inode *src, u64 soffset, u64 len,
-		 struct inode *dest, u64 doffset),
-	TP_ARGS(src, soffset, len, dest, doffset),
-	TP_STRUCT__entry(
-		__field(dev_t, dev)
-		__field(unsigned long, src_ino)
-		__field(loff_t, src_isize)
-		__field(loff_t, src_offset)
-		__field(size_t, len)
-		__field(unsigned long, dest_ino)
-		__field(loff_t, dest_isize)
-		__field(loff_t, dest_offset)
-	),
-	TP_fast_assign(
-		__entry->dev = src->i_sb->s_dev;
-		__entry->src_ino = src->i_ino;
-		__entry->src_isize = i_size_read(src);
-		__entry->src_offset = soffset;
-		__entry->len = len;
-		__entry->dest_ino = dest->i_ino;
-		__entry->dest_isize = i_size_read(dest);
-		__entry->dest_offset = doffset;
-	),
-	TP_printk("dev %d:%d count %zd "
-		  "ino 0x%lx isize 0x%llx offset 0x%llx -> "
-		  "ino 0x%lx isize 0x%llx offset 0x%llx",
-		  MAJOR(__entry->dev), MINOR(__entry->dev),
-		  __entry->len,
-		  __entry->src_ino,
-		  __entry->src_isize,
-		  __entry->src_offset,
-		  __entry->dest_ino,
-		  __entry->dest_isize,
-		  __entry->dest_offset)
-)
-
-#define DEFINE_DOUBLE_VFS_IO_EVENT(name)	\
-DEFINE_EVENT(xfs_double_vfs_io_class, name,	\
-	TP_PROTO(struct inode *src, u64 soffset, u64 len, \
-		 struct inode *dest, u64 doffset), \
-	TP_ARGS(src, soffset, len, dest, doffset))
-
-/* CoW write tracepoint */
-DECLARE_EVENT_CLASS(xfs_copy_on_write_class,
-	TP_PROTO(struct xfs_inode *ip, xfs_fileoff_t lblk, xfs_fsblock_t pblk,
-		 xfs_extlen_t len, xfs_fsblock_t new_pblk),
-	TP_ARGS(ip, lblk, pblk, len, new_pblk),
-	TP_STRUCT__entry(
-		__field(dev_t, dev)
-		__field(xfs_ino_t, ino)
-		__field(xfs_fileoff_t, lblk)
-		__field(xfs_fsblock_t, pblk)
-		__field(xfs_extlen_t, len)
-		__field(xfs_fsblock_t, new_pblk)
-	),
-	TP_fast_assign(
-		__entry->dev = VFS_I(ip)->i_sb->s_dev;
-		__entry->ino = ip->i_ino;
-		__entry->lblk = lblk;
-		__entry->pblk = pblk;
-		__entry->len = len;
-		__entry->new_pblk = new_pblk;
-	),
-	TP_printk("dev %d:%d ino 0x%llx lblk 0x%llx pblk 0x%llx "
-		  "len 0x%x new_pblk %llu",
-		  MAJOR(__entry->dev), MINOR(__entry->dev),
-		  __entry->ino,
-		  __entry->lblk,
-		  __entry->pblk,
-		  __entry->len,
-		  __entry->new_pblk)
-)
-
-#define DEFINE_COW_EVENT(name)	\
-DEFINE_EVENT(xfs_copy_on_write_class, name,	\
-	TP_PROTO(struct xfs_inode *ip, xfs_fileoff_t lblk, xfs_fsblock_t pblk, \
-		 xfs_extlen_t len, xfs_fsblock_t new_pblk), \
-	TP_ARGS(ip, lblk, pblk, len, new_pblk))
-
 /* inode/irec events */
 DECLARE_EVENT_CLASS(xfs_inode_irec_class,
 	TP_PROTO(struct xfs_inode *ip, struct xfs_bmbt_irec *irec),
@@ -3292,8 +3198,6 @@ DEFINE_DOUBLE_IO_EVENT(xfs_reflink_remap_range);
 DEFINE_INODE_ERROR_EVENT(xfs_reflink_remap_range_error);
 DEFINE_INODE_ERROR_EVENT(xfs_reflink_set_inode_flag_error);
 DEFINE_INODE_ERROR_EVENT(xfs_reflink_update_inode_size_error);
-DEFINE_INODE_ERROR_EVENT(xfs_reflink_reflink_main_loop_error);
-DEFINE_INODE_ERROR_EVENT(xfs_reflink_read_iomap_error);
 DEFINE_INODE_ERROR_EVENT(xfs_reflink_remap_blocks_error);
 DEFINE_INODE_ERROR_EVENT(xfs_reflink_remap_extent_error);
 
@@ -3302,9 +3206,6 @@ DEFINE_DOUBLE_IO_EVENT(xfs_reflink_compare_extents);
 DEFINE_INODE_ERROR_EVENT(xfs_reflink_compare_extents_error);
 
 /* ioctl tracepoints */
-DEFINE_DOUBLE_VFS_IO_EVENT(xfs_ioctl_reflink);
-DEFINE_DOUBLE_VFS_IO_EVENT(xfs_ioctl_clone_range);
-DEFINE_DOUBLE_VFS_IO_EVENT(xfs_ioctl_file_extent_same);
 TRACE_EVENT(xfs_ioctl_clone,
 	TP_PROTO(struct inode *src, struct inode *dest),
 	TP_ARGS(src, dest),
@@ -3334,11 +3235,7 @@ TRACE_EVENT(xfs_ioctl_clone,
 
 /* unshare tracepoints */
 DEFINE_SIMPLE_IO_EVENT(xfs_reflink_unshare);
-DEFINE_SIMPLE_IO_EVENT(xfs_reflink_cow_eof_block);
-DEFINE_PAGE_EVENT(xfs_reflink_unshare_page);
 DEFINE_INODE_ERROR_EVENT(xfs_reflink_unshare_error);
-DEFINE_INODE_ERROR_EVENT(xfs_reflink_cow_eof_block_error);
-DEFINE_INODE_ERROR_EVENT(xfs_reflink_dirty_page_error);
 
 /* copy on write */
 DEFINE_INODE_IREC_EVENT(xfs_reflink_trim_around_shared);
@@ -3361,14 +3258,8 @@ DEFINE_INODE_ERROR_EVENT(xfs_reflink_allocate_cow_range_error);
 DEFINE_INODE_ERROR_EVENT(xfs_reflink_cancel_cow_range_error);
 DEFINE_INODE_ERROR_EVENT(xfs_reflink_end_cow_error);
 
-DEFINE_COW_EVENT(xfs_reflink_fork_buf);
-DEFINE_COW_EVENT(xfs_reflink_finish_fork_buf);
-DEFINE_INODE_ERROR_EVENT(xfs_reflink_fork_buf_error);
-DEFINE_INODE_ERROR_EVENT(xfs_reflink_finish_fork_buf_error);
 
-DEFINE_INODE_EVENT(xfs_reflink_cancel_pending_cow);
 DEFINE_INODE_IREC_EVENT(xfs_reflink_cancel_cow);
-DEFINE_INODE_ERROR_EVENT(xfs_reflink_cancel_pending_cow_error);
 
 /* rmap swapext tracepoints */
 DEFINE_INODE_IREC_EVENT(xfs_swap_extent_rmap_remap);

fs/xfs/xfs_xattr.c

@@ -130,7 +130,7 @@ const struct xattr_handler *xfs_xattr_handlers[] = {
 	NULL
 };
 
-static int
+static void
 __xfs_xattr_put_listent(
 	struct xfs_attr_list_context *context,
 	char *prefix,
@@ -148,7 +148,7 @@ __xfs_xattr_put_listent(
 	if (arraytop > context->firstu) {
 		context->count = -1;	/* insufficient space */
 		context->seen_enough = 1;
-		return 0;
+		return;
 	}
 	offset = (char *)context->alist + context->count;
 	strncpy(offset, prefix, prefix_len);
@@ -159,10 +159,10 @@ __xfs_xattr_put_listent(
 
 compute_size:
 	context->count += prefix_len + namelen + 1;
-	return 0;
+	return;
 }
 
-static int
+static void
 xfs_xattr_put_listent(
 	struct xfs_attr_list_context *context,
 	int		flags,
@@ -180,23 +180,19 @@ xfs_xattr_put_listent(
 		if (namelen == SGI_ACL_FILE_SIZE &&
 		    strncmp(name, SGI_ACL_FILE,
 			    SGI_ACL_FILE_SIZE) == 0) {
-			int ret = __xfs_xattr_put_listent(
+			__xfs_xattr_put_listent(
 					context, XATTR_SYSTEM_PREFIX,
 					XATTR_SYSTEM_PREFIX_LEN,
 					XATTR_POSIX_ACL_ACCESS,
 					strlen(XATTR_POSIX_ACL_ACCESS));
-			if (ret)
-				return ret;
 		} else if (namelen == SGI_ACL_DEFAULT_SIZE &&
 			 strncmp(name, SGI_ACL_DEFAULT,
 				 SGI_ACL_DEFAULT_SIZE) == 0) {
-			int ret = __xfs_xattr_put_listent(
+			__xfs_xattr_put_listent(
 					context, XATTR_SYSTEM_PREFIX,
 					XATTR_SYSTEM_PREFIX_LEN,
 					XATTR_POSIX_ACL_DEFAULT,
 					strlen(XATTR_POSIX_ACL_DEFAULT));
-			if (ret)
-				return ret;
 		}
 #endif
 
@@ -205,7 +201,7 @@ xfs_xattr_put_listent(
 		 * see them.
 		 */
 		if (!capable(CAP_SYS_ADMIN))
-			return 0;
+			return;
 
 		prefix = XATTR_TRUSTED_PREFIX;
 		prefix_len = XATTR_TRUSTED_PREFIX_LEN;
@@ -217,8 +213,9 @@ xfs_xattr_put_listent(
 		prefix_len = XATTR_USER_PREFIX_LEN;
 	}
 
-	return __xfs_xattr_put_listent(context, prefix, prefix_len, name,
-				       namelen);
+	__xfs_xattr_put_listent(context, prefix, prefix_len, name,
+				namelen);
+	return;
 }
 
 ssize_t

include/linux/iomap.h

@@ -50,6 +50,7 @@ struct iomap {
 #define IOMAP_ZERO		(1 << 1) /* zeroing operation, may skip holes */
 #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 */
 
 struct iomap_ops {
 	/*
@@ -83,4 +84,14 @@ int iomap_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf,
 int iomap_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 		loff_t start, loff_t len, struct iomap_ops *ops);
 
+/*
+ * Flags for direct I/O ->end_io:
+ */
+#define IOMAP_DIO_UNWRITTEN	(1 << 0)	/* covers unwritten extent(s) */
+#define IOMAP_DIO_COW		(1 << 1)	/* covers COW extent(s) */
+typedef int (iomap_dio_end_io_t)(struct kiocb *iocb, ssize_t ret,
+		unsigned flags);
+ssize_t iomap_dio_rw(struct kiocb *iocb, struct iov_iter *iter,
+		struct iomap_ops *ops, iomap_dio_end_io_t end_io);
+
 #endif /* LINUX_IOMAP_H */

include/linux/lockdep.h

@@ -338,9 +338,18 @@ extern void lock_acquire(struct lockdep_map *lock, unsigned int subclass,
 extern void lock_release(struct lockdep_map *lock, int nested,
 			 unsigned long ip);
 
-#define lockdep_is_held(lock)	lock_is_held(&(lock)->dep_map)
+/*
+ * Same "read" as for lock_acquire(), except -1 means any.
+ */
+extern int lock_is_held_type(struct lockdep_map *lock, int read);
+
+static inline int lock_is_held(struct lockdep_map *lock)
+{
+	return lock_is_held_type(lock, -1);
+}
 
-extern int lock_is_held(struct lockdep_map *lock);
+#define lockdep_is_held(lock)		lock_is_held(&(lock)->dep_map)
+#define lockdep_is_held_type(lock, r)	lock_is_held_type(&(lock)->dep_map, (r))
 
 extern void lock_set_class(struct lockdep_map *lock, const char *name,
 			   struct lock_class_key *key, unsigned int subclass,
@@ -372,6 +381,14 @@ extern void lock_unpin_lock(struct lockdep_map *lock, struct pin_cookie);
 		WARN_ON(debug_locks && !lockdep_is_held(l));	\
 	} while (0)
 
+#define lockdep_assert_held_exclusive(l)	do {			\
+		WARN_ON(debug_locks && !lockdep_is_held_type(l, 0));	\
+	} while (0)
+
+#define lockdep_assert_held_read(l)	do {				\
+		WARN_ON(debug_locks && !lockdep_is_held_type(l, 1));	\
+	} while (0)
+
 #define lockdep_assert_held_once(l)	do {				\
 		WARN_ON_ONCE(debug_locks && !lockdep_is_held(l));	\
 	} while (0)
@@ -428,7 +445,11 @@ struct lock_class_key { };
 
 #define lockdep_depth(tsk)	(0)
 
+#define lockdep_is_held_type(l, r)		(1)
+
 #define lockdep_assert_held(l)			do { (void)(l); } while (0)
+#define lockdep_assert_held_exclusive(l)	do { (void)(l); } while (0)
+#define lockdep_assert_held_read(l)		do { (void)(l); } while (0)
 #define lockdep_assert_held_once(l)		do { (void)(l); } while (0)
 
 #define lockdep_recursing(tsk)			(0)

kernel/locking/lockdep.c

@@ -3191,7 +3191,7 @@ print_lock_nested_lock_not_held(struct task_struct *curr,
 	return 0;
 }
 
-static int __lock_is_held(struct lockdep_map *lock);
+static int __lock_is_held(struct lockdep_map *lock, int read);
 
 /*
  * This gets called for every mutex_lock*()/spin_lock*() operation.
@@ -3332,7 +3332,7 @@ static int __lock_acquire(struct lockdep_map *lock, unsigned int subclass,
 	}
 	chain_key = iterate_chain_key(chain_key, class_idx);
 
-	if (nest_lock && !__lock_is_held(nest_lock))
+	if (nest_lock && !__lock_is_held(nest_lock, -1))
 		return print_lock_nested_lock_not_held(curr, hlock, ip);
 
 	if (!validate_chain(curr, lock, hlock, chain_head, chain_key))
@@ -3579,7 +3579,7 @@ found_it:
 	return 1;
 }
 
-static int __lock_is_held(struct lockdep_map *lock)
+static int __lock_is_held(struct lockdep_map *lock, int read)
 {
 	struct task_struct *curr = current;
 	int i;
@@ -3587,8 +3587,12 @@ static int __lock_is_held(struct lockdep_map *lock)
 	for (i = 0; i < curr->lockdep_depth; i++) {
 		struct held_lock *hlock = curr->held_locks + i;
 
-		if (match_held_lock(hlock, lock))
-			return 1;
+		if (match_held_lock(hlock, lock)) {
+			if (read == -1 || hlock->read == read)
+				return 1;
+
+			return 0;
+		}
 	}
 
 	return 0;
@@ -3772,7 +3776,7 @@ void lock_release(struct lockdep_map *lock, int nested,
 }
 EXPORT_SYMBOL_GPL(lock_release);
 
-int lock_is_held(struct lockdep_map *lock)
+int lock_is_held_type(struct lockdep_map *lock, int read)
 {
 	unsigned long flags;
 	int ret = 0;
@@ -3784,13 +3788,13 @@ int lock_is_held(struct lockdep_map *lock)
 	check_flags(flags);
 
 	current->lockdep_recursion = 1;
-	ret = __lock_is_held(lock);
+	ret = __lock_is_held(lock, read);
 	current->lockdep_recursion = 0;
 	raw_local_irq_restore(flags);
 
 	return ret;
 }
-EXPORT_SYMBOL_GPL(lock_is_held);
+EXPORT_SYMBOL_GPL(lock_is_held_type);
 
 struct pin_cookie lock_pin_lock(struct lockdep_map *lock)
 {