1 files changed, 179 insertions, 469 deletions

diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index 176b432035ae..91b00eb2440e 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -30,329 +30,13 @@
 #include "delalloc-space.h"
 #include "reflink.h"
 #include "subpage.h"
-
-static struct kmem_cache *btrfs_inode_defrag_cachep;
-/*
- * when auto defrag is enabled we
- * queue up these defrag structs to remember which
- * inodes need defragging passes
- */
-struct inode_defrag {
-	struct rb_node rb_node;
-	/* objectid */
-	u64 ino;
-	/*
-	 * transid where the defrag was added, we search for
-	 * extents newer than this
-	 */
-	u64 transid;
-
-	/* root objectid */
-	u64 root;
-
-	/*
-	 * The extent size threshold for autodefrag.
-	 *
-	 * This value is different for compressed/non-compressed extents,
-	 * thus needs to be passed from higher layer.
-	 * (aka, inode_should_defrag())
-	 */
-	u32 extent_thresh;
-};
-
-static int __compare_inode_defrag(struct inode_defrag *defrag1,
-				  struct inode_defrag *defrag2)
-{
-	if (defrag1->root > defrag2->root)
-		return 1;
-	else if (defrag1->root < defrag2->root)
-		return -1;
-	else if (defrag1->ino > defrag2->ino)
-		return 1;
-	else if (defrag1->ino < defrag2->ino)
-		return -1;
-	else
-		return 0;
-}
-
-/* pop a record for an inode into the defrag tree.  The lock
- * must be held already
- *
- * If you're inserting a record for an older transid than an
- * existing record, the transid already in the tree is lowered
- *
- * If an existing record is found the defrag item you
- * pass in is freed
- */
-static int __btrfs_add_inode_defrag(struct btrfs_inode *inode,
-				    struct inode_defrag *defrag)
-{
-	struct btrfs_fs_info *fs_info = inode->root->fs_info;
-	struct inode_defrag *entry;
-	struct rb_node **p;
-	struct rb_node *parent = NULL;
-	int ret;
-
-	p = &fs_info->defrag_inodes.rb_node;
-	while (*p) {
-		parent = *p;
-		entry = rb_entry(parent, struct inode_defrag, rb_node);
-
-		ret = __compare_inode_defrag(defrag, entry);
-		if (ret < 0)
-			p = &parent->rb_left;
-		else if (ret > 0)
-			p = &parent->rb_right;
-		else {
-			/* if we're reinserting an entry for
-			 * an old defrag run, make sure to
-			 * lower the transid of our existing record
-			 */
-			if (defrag->transid < entry->transid)
-				entry->transid = defrag->transid;
-			entry->extent_thresh = min(defrag->extent_thresh,
-						   entry->extent_thresh);
-			return -EEXIST;
-		}
-	}
-	set_bit(BTRFS_INODE_IN_DEFRAG, &inode->runtime_flags);
-	rb_link_node(&defrag->rb_node, parent, p);
-	rb_insert_color(&defrag->rb_node, &fs_info->defrag_inodes);
-	return 0;
-}
-
-static inline int __need_auto_defrag(struct btrfs_fs_info *fs_info)
-{
-	if (!btrfs_test_opt(fs_info, AUTO_DEFRAG))
-		return 0;
-
-	if (btrfs_fs_closing(fs_info))
-		return 0;
-
-	return 1;
-}
-
-/*
- * insert a defrag record for this inode if auto defrag is
- * enabled
- */
-int btrfs_add_inode_defrag(struct btrfs_trans_handle *trans,
-			   struct btrfs_inode *inode, u32 extent_thresh)
-{
-	struct btrfs_root *root = inode->root;
-	struct btrfs_fs_info *fs_info = root->fs_info;
-	struct inode_defrag *defrag;
-	u64 transid;
-	int ret;
-
-	if (!__need_auto_defrag(fs_info))
-		return 0;
-
-	if (test_bit(BTRFS_INODE_IN_DEFRAG, &inode->runtime_flags))
-		return 0;
-
-	if (trans)
-		transid = trans->transid;
-	else
-		transid = inode->root->last_trans;
-
-	defrag = kmem_cache_zalloc(btrfs_inode_defrag_cachep, GFP_NOFS);
-	if (!defrag)
-		return -ENOMEM;
-
-	defrag->ino = btrfs_ino(inode);
-	defrag->transid = transid;
-	defrag->root = root->root_key.objectid;
-	defrag->extent_thresh = extent_thresh;
-
-	spin_lock(&fs_info->defrag_inodes_lock);
-	if (!test_bit(BTRFS_INODE_IN_DEFRAG, &inode->runtime_flags)) {
-		/*
-		 * If we set IN_DEFRAG flag and evict the inode from memory,
-		 * and then re-read this inode, this new inode doesn't have
-		 * IN_DEFRAG flag. At the case, we may find the existed defrag.
-		 */
-		ret = __btrfs_add_inode_defrag(inode, defrag);
-		if (ret)
-			kmem_cache_free(btrfs_inode_defrag_cachep, defrag);
-	} else {
-		kmem_cache_free(btrfs_inode_defrag_cachep, defrag);
-	}
-	spin_unlock(&fs_info->defrag_inodes_lock);
-	return 0;
-}
-
-/*
- * pick the defragable inode that we want, if it doesn't exist, we will get
- * the next one.
- */
-static struct inode_defrag *
-btrfs_pick_defrag_inode(struct btrfs_fs_info *fs_info, u64 root, u64 ino)
-{
-	struct inode_defrag *entry = NULL;
-	struct inode_defrag tmp;
-	struct rb_node *p;
-	struct rb_node *parent = NULL;
-	int ret;
-
-	tmp.ino = ino;
-	tmp.root = root;
-
-	spin_lock(&fs_info->defrag_inodes_lock);
-	p = fs_info->defrag_inodes.rb_node;
-	while (p) {
-		parent = p;
-		entry = rb_entry(parent, struct inode_defrag, rb_node);
-
-		ret = __compare_inode_defrag(&tmp, entry);
-		if (ret < 0)
-			p = parent->rb_left;
-		else if (ret > 0)
-			p = parent->rb_right;
-		else
-			goto out;
-	}
-
-	if (parent && __compare_inode_defrag(&tmp, entry) > 0) {
-		parent = rb_next(parent);
-		if (parent)
-			entry = rb_entry(parent, struct inode_defrag, rb_node);
-		else
-			entry = NULL;
-	}
-out:
-	if (entry)
-		rb_erase(parent, &fs_info->defrag_inodes);
-	spin_unlock(&fs_info->defrag_inodes_lock);
-	return entry;
-}
-
-void btrfs_cleanup_defrag_inodes(struct btrfs_fs_info *fs_info)
-{
-	struct inode_defrag *defrag;
-	struct rb_node *node;
-
-	spin_lock(&fs_info->defrag_inodes_lock);
-	node = rb_first(&fs_info->defrag_inodes);
-	while (node) {
-		rb_erase(node, &fs_info->defrag_inodes);
-		defrag = rb_entry(node, struct inode_defrag, rb_node);
-		kmem_cache_free(btrfs_inode_defrag_cachep, defrag);
-
-		cond_resched_lock(&fs_info->defrag_inodes_lock);
-
-		node = rb_first(&fs_info->defrag_inodes);
-	}
-	spin_unlock(&fs_info->defrag_inodes_lock);
-}
-
-#define BTRFS_DEFRAG_BATCH	1024
-
-static int __btrfs_run_defrag_inode(struct btrfs_fs_info *fs_info,
-				    struct inode_defrag *defrag)
-{
-	struct btrfs_root *inode_root;
-	struct inode *inode;
-	struct btrfs_ioctl_defrag_range_args range;
-	int ret = 0;
-	u64 cur = 0;
-
-again:
-	if (test_bit(BTRFS_FS_STATE_REMOUNTING, &fs_info->fs_state))
-		goto cleanup;
-	if (!__need_auto_defrag(fs_info))
-		goto cleanup;
-
-	/* get the inode */
-	inode_root = btrfs_get_fs_root(fs_info, defrag->root, true);
-	if (IS_ERR(inode_root)) {
-		ret = PTR_ERR(inode_root);
-		goto cleanup;
-	}
-
-	inode = btrfs_iget(fs_info->sb, defrag->ino, inode_root);
-	btrfs_put_root(inode_root);
-	if (IS_ERR(inode)) {
-		ret = PTR_ERR(inode);
-		goto cleanup;
-	}
-
-	if (cur >= i_size_read(inode)) {
-		iput(inode);
-		goto cleanup;
-	}
-
-	/* do a chunk of defrag */
-	clear_bit(BTRFS_INODE_IN_DEFRAG, &BTRFS_I(inode)->runtime_flags);
-	memset(&range, 0, sizeof(range));
-	range.len = (u64)-1;
-	range.start = cur;
-	range.extent_thresh = defrag->extent_thresh;
-
-	sb_start_write(fs_info->sb);
-	ret = btrfs_defrag_file(inode, NULL, &range, defrag->transid,
-				       BTRFS_DEFRAG_BATCH);
-	sb_end_write(fs_info->sb);
-	iput(inode);
-
-	if (ret < 0)
-		goto cleanup;
-
-	cur = max(cur + fs_info->sectorsize, range.start);
-	goto again;
-
-cleanup:
-	kmem_cache_free(btrfs_inode_defrag_cachep, defrag);
-	return ret;
-}
-
-/*
- * run through the list of inodes in the FS that need
- * defragging
- */
-int btrfs_run_defrag_inodes(struct btrfs_fs_info *fs_info)
-{
-	struct inode_defrag *defrag;
-	u64 first_ino = 0;
-	u64 root_objectid = 0;
-
-	atomic_inc(&fs_info->defrag_running);
-	while (1) {
-		/* Pause the auto defragger. */
-		if (test_bit(BTRFS_FS_STATE_REMOUNTING,
-			     &fs_info->fs_state))
-			break;
-
-		if (!__need_auto_defrag(fs_info))
-			break;
-
-		/* find an inode to defrag */
-		defrag = btrfs_pick_defrag_inode(fs_info, root_objectid,
-						 first_ino);
-		if (!defrag) {
-			if (root_objectid || first_ino) {
-				root_objectid = 0;
-				first_ino = 0;
-				continue;
-			} else {
-				break;
-			}
-		}
-
-		first_ino = defrag->ino + 1;
-		root_objectid = defrag->root;
-
-		__btrfs_run_defrag_inode(fs_info, defrag);
-	}
-	atomic_dec(&fs_info->defrag_running);
-
-	/*
-	 * during unmount, we use the transaction_wait queue to
-	 * wait for the defragger to stop
-	 */
-	wake_up(&fs_info->transaction_wait);
-	return 0;
-}
+#include "fs.h"
+#include "accessors.h"
+#include "extent-tree.h"
+#include "file-item.h"
+#include "ioctl.h"
+#include "file.h"
+#include "super.h"
 
 /* simple helper to fault in pages and copy.  This should go away
  * and be replaced with calls into generic code.
@@ -696,7 +380,10 @@ next_slot:
 						args->start - extent_offset,
 						0, false);
 				ret = btrfs_inc_extent_ref(trans, &ref);
-				BUG_ON(ret); /* -ENOMEM */
+				if (ret) {
+					btrfs_abort_transaction(trans, ret);
+					break;
+				}
 			}
 			key.offset = args->start;
 		}
@@ -783,7 +470,10 @@ delete_extent_item:
 						key.offset - extent_offset, 0,
 						false);
 				ret = btrfs_free_extent(trans, &ref);
-				BUG_ON(ret); /* -ENOMEM */
+				if (ret) {
+					btrfs_abort_transaction(trans, ret);
+					break;
+				}
 				args->bytes_found += extent_end - key.offset;
 			}
 
@@ -1302,7 +992,8 @@ lock_and_cleanup_extent_if_need(struct btrfs_inode *inode, struct page **pages,
 		struct btrfs_ordered_extent *ordered;
 
 		if (nowait) {
-			if (!try_lock_extent(&inode->io_tree, start_pos, last_pos)) {
+			if (!try_lock_extent(&inode->io_tree, start_pos, last_pos,
+					     cached_state)) {
 				for (i = 0; i < num_pages; i++) {
 					unlock_page(pages[i]);
 					put_page(pages[i]);
@@ -1372,6 +1063,7 @@ int btrfs_check_nocow_lock(struct btrfs_inode *inode, loff_t pos,
 {
 	struct btrfs_fs_info *fs_info = inode->root->fs_info;
 	struct btrfs_root *root = inode->root;
+	struct extent_state *cached_state = NULL;
 	u64 lockstart, lockend;
 	u64 num_bytes;
 	int ret;
@@ -1388,12 +1080,14 @@ int btrfs_check_nocow_lock(struct btrfs_inode *inode, loff_t pos,
 	num_bytes = lockend - lockstart + 1;
 
 	if (nowait) {
-		if (!btrfs_try_lock_ordered_range(inode, lockstart, lockend)) {
+		if (!btrfs_try_lock_ordered_range(inode, lockstart, lockend,
+						  &cached_state)) {
 			btrfs_drew_write_unlock(&root->snapshot_lock);
 			return -EAGAIN;
 		}
 	} else {
-		btrfs_lock_and_flush_ordered_range(inode, lockstart, lockend, NULL);
+		btrfs_lock_and_flush_ordered_range(inode, lockstart, lockend,
+						   &cached_state);
 	}
 	ret = can_nocow_extent(&inode->vfs_inode, lockstart, &num_bytes,
 			NULL, NULL, NULL, nowait, false);
@@ -1402,7 +1096,7 @@ int btrfs_check_nocow_lock(struct btrfs_inode *inode, loff_t pos,
 	else
 		*write_bytes = min_t(size_t, *write_bytes ,
 				     num_bytes - pos + lockstart);
-	unlock_extent(&inode->io_tree, lockstart, lockend, NULL);
+	unlock_extent(&inode->io_tree, lockstart, lockend, &cached_state);
 
 	return ret;
 }
@@ -1505,7 +1199,7 @@ static noinline ssize_t btrfs_buffered_write(struct kiocb *iocb,
 	if (nowait)
 		ilock_flags |= BTRFS_ILOCK_TRY;
 
-	ret = btrfs_inode_lock(inode, ilock_flags);
+	ret = btrfs_inode_lock(BTRFS_I(inode), ilock_flags);
 	if (ret < 0)
 		return ret;
 
@@ -1598,14 +1292,19 @@ static noinline ssize_t btrfs_buffered_write(struct kiocb *iocb,
 						write_bytes);
 			else
 				btrfs_check_nocow_unlock(BTRFS_I(inode));
+
+			if (nowait && ret == -ENOSPC)
+				ret = -EAGAIN;
 			break;
 		}
 
 		release_bytes = reserve_bytes;
 again:
 		ret = balance_dirty_pages_ratelimited_flags(inode->i_mapping, bdp_flags);
-		if (ret)
+		if (ret) {
+			btrfs_delalloc_release_extents(BTRFS_I(inode), reserve_bytes);
 			break;
+		}
 
 		/*
 		 * This is going to setup the pages array with the number of
@@ -1735,7 +1434,7 @@ again:
 		iocb->ki_pos += num_written;
 	}
 out:
-	btrfs_inode_unlock(inode, ilock_flags);
+	btrfs_inode_unlock(BTRFS_I(inode), ilock_flags);
 	return num_written ? num_written : ret;
 }
 
@@ -1765,6 +1464,7 @@ static ssize_t btrfs_direct_write(struct kiocb *iocb, struct iov_iter *from)
 	loff_t endbyte;
 	ssize_t err;
 	unsigned int ilock_flags = 0;
+	struct iomap_dio *dio;
 
 	if (iocb->ki_flags & IOCB_NOWAIT)
 		ilock_flags |= BTRFS_ILOCK_TRY;
@@ -1774,19 +1474,19 @@ static ssize_t btrfs_direct_write(struct kiocb *iocb, struct iov_iter *from)
 		ilock_flags |= BTRFS_ILOCK_SHARED;
 
 relock:
-	err = btrfs_inode_lock(inode, ilock_flags);
+	err = btrfs_inode_lock(BTRFS_I(inode), ilock_flags);
 	if (err < 0)
 		return err;
 
 	err = generic_write_checks(iocb, from);
 	if (err <= 0) {
-		btrfs_inode_unlock(inode, ilock_flags);
+		btrfs_inode_unlock(BTRFS_I(inode), ilock_flags);
 		return err;
 	}
 
 	err = btrfs_write_check(iocb, from, err);
 	if (err < 0) {
-		btrfs_inode_unlock(inode, ilock_flags);
+		btrfs_inode_unlock(BTRFS_I(inode), ilock_flags);
 		goto out;
 	}
 
@@ -1797,13 +1497,13 @@ relock:
 	 */
 	if ((ilock_flags & BTRFS_ILOCK_SHARED) &&
 	    pos + iov_iter_count(from) > i_size_read(inode)) {
-		btrfs_inode_unlock(inode, ilock_flags);
+		btrfs_inode_unlock(BTRFS_I(inode), ilock_flags);
 		ilock_flags &= ~BTRFS_ILOCK_SHARED;
 		goto relock;
 	}
 
 	if (check_direct_IO(fs_info, from, pos)) {
-		btrfs_inode_unlock(inode, ilock_flags);
+		btrfs_inode_unlock(BTRFS_I(inode), ilock_flags);
 		goto buffered;
 	}
 
@@ -1825,11 +1525,22 @@ relock:
 	 * So here we disable page faults in the iov_iter and then retry if we
 	 * got -EFAULT, faulting in the pages before the retry.
 	 */
-again:
 	from->nofault = true;
-	err = btrfs_dio_rw(iocb, from, written);
+	dio = btrfs_dio_write(iocb, from, written);
 	from->nofault = false;
 
+	/*
+	 * iomap_dio_complete() will call btrfs_sync_file() if we have a dsync
+	 * iocb, and that needs to lock the inode. So unlock it before calling
+	 * iomap_dio_complete() to avoid a deadlock.
+	 */
+	btrfs_inode_unlock(BTRFS_I(inode), ilock_flags);
+
+	if (IS_ERR_OR_NULL(dio))
+		err = PTR_ERR_OR_ZERO(dio);
+	else
+		err = iomap_dio_complete(dio);
+
 	/* No increment (+=) because iomap returns a cumulative value. */
 	if (err > 0)
 		written = err;
@@ -1855,12 +1566,10 @@ again:
 		} else {
 			fault_in_iov_iter_readable(from, left);
 			prev_left = left;
-			goto again;
+			goto relock;
 		}
 	}
 
-	btrfs_inode_unlock(inode, ilock_flags);
-
 	/*
 	 * If 'err' is -ENOTBLK or we have not written all data, then it means
 	 * we must fallback to buffered IO.
@@ -1872,8 +1581,8 @@ buffered:
 	/*
 	 * If we are in a NOWAIT context, then return -EAGAIN to signal the caller
 	 * it must retry the operation in a context where blocking is acceptable,
-	 * since we currently don't have NOWAIT semantics support for buffered IO
-	 * and may block there for many reasons (reserving space for example).
+	 * because even if we end up not blocking during the buffered IO attempt
+	 * below, we will block when flushing and waiting for the IO.
 	 */
 	if (iocb->ki_flags & IOCB_NOWAIT) {
 		err = -EAGAIN;
@@ -1913,7 +1622,7 @@ static ssize_t btrfs_encoded_write(struct kiocb *iocb, struct iov_iter *from,
 	loff_t count;
 	ssize_t ret;
 
-	btrfs_inode_lock(inode, 0);
+	btrfs_inode_lock(BTRFS_I(inode), 0);
 	count = encoded->len;
 	ret = generic_write_checks_count(iocb, &count);
 	if (ret == 0 && count != encoded->len) {
@@ -1932,7 +1641,7 @@ static ssize_t btrfs_encoded_write(struct kiocb *iocb, struct iov_iter *from,
 
 	ret = btrfs_do_encoded_write(iocb, from, encoded);
 out:
-	btrfs_inode_unlock(inode, 0);
+	btrfs_inode_unlock(BTRFS_I(inode), 0);
 	return ret;
 }
 
@@ -1993,10 +1702,12 @@ int btrfs_release_file(struct inode *inode, struct file *filp)
 {
 	struct btrfs_file_private *private = filp->private_data;
 
-	if (private && private->filldir_buf)
+	if (private) {
 		kfree(private->filldir_buf);
-	kfree(private);
-	filp->private_data = NULL;
+		free_extent_state(private->llseek_cached_state);
+		kfree(private);
+		filp->private_data = NULL;
+	}
 
 	/*
 	 * Set by setattr when we are about to truncate a file from a non-zero
@@ -2103,7 +1814,7 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 	if (ret)
 		goto out;
 
-	btrfs_inode_lock(inode, BTRFS_ILOCK_MMAP);
+	btrfs_inode_lock(BTRFS_I(inode), BTRFS_ILOCK_MMAP);
 
 	atomic_inc(&root->log_batch);
 
@@ -2127,7 +1838,7 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 	 */
 	ret = start_ordered_ops(inode, start, end);
 	if (ret) {
-		btrfs_inode_unlock(inode, BTRFS_ILOCK_MMAP);
+		btrfs_inode_unlock(BTRFS_I(inode), BTRFS_ILOCK_MMAP);
 		goto out;
 	}
 
@@ -2230,7 +1941,7 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 	 * file again, but that will end up using the synchronization
 	 * inside btrfs_sync_log to keep things safe.
 	 */
-	btrfs_inode_unlock(inode, BTRFS_ILOCK_MMAP);
+	btrfs_inode_unlock(BTRFS_I(inode), BTRFS_ILOCK_MMAP);
 
 	if (ret == BTRFS_NO_LOG_SYNC) {
 		ret = btrfs_end_transaction(trans);
@@ -2298,7 +2009,7 @@ out:
 
 out_release_extents:
 	btrfs_release_log_ctx_extents(&ctx);
-	btrfs_inode_unlock(inode, BTRFS_ILOCK_MMAP);
+	btrfs_inode_unlock(BTRFS_I(inode), BTRFS_ILOCK_MMAP);
 	goto out;
 }
 
@@ -2893,7 +2604,7 @@ static int btrfs_punch_hole(struct file *file, loff_t offset, loff_t len)
 	bool truncated_block = false;
 	bool updated_inode = false;
 
-	btrfs_inode_lock(inode, BTRFS_ILOCK_MMAP);
+	btrfs_inode_lock(BTRFS_I(inode), BTRFS_ILOCK_MMAP);
 
 	ret = btrfs_wait_ordered_range(inode, offset, len);
 	if (ret)
@@ -2941,7 +2652,7 @@ static int btrfs_punch_hole(struct file *file, loff_t offset, loff_t len)
 		truncated_block = true;
 		ret = btrfs_truncate_block(BTRFS_I(inode), offset, 0, 0);
 		if (ret) {
-			btrfs_inode_unlock(inode, BTRFS_ILOCK_MMAP);
+			btrfs_inode_unlock(BTRFS_I(inode), BTRFS_ILOCK_MMAP);
 			return ret;
 		}
 	}
@@ -3040,7 +2751,7 @@ out_only_mutex:
 				ret = ret2;
 		}
 	}
-	btrfs_inode_unlock(inode, BTRFS_ILOCK_MMAP);
+	btrfs_inode_unlock(BTRFS_I(inode), BTRFS_ILOCK_MMAP);
 	return ret;
 }
 
@@ -3351,7 +3062,7 @@ static long btrfs_fallocate(struct file *file, int mode,
 	if (mode & FALLOC_FL_PUNCH_HOLE)
 		return btrfs_punch_hole(file, offset, len);
 
-	btrfs_inode_lock(inode, BTRFS_ILOCK_MMAP);
+	btrfs_inode_lock(BTRFS_I(inode), BTRFS_ILOCK_MMAP);
 
 	if (!(mode & FALLOC_FL_KEEP_SIZE) && offset + len > inode->i_size) {
 		ret = inode_newsize_ok(inode, offset + len);
@@ -3401,7 +3112,7 @@ static long btrfs_fallocate(struct file *file, int mode,
 
 	if (mode & FALLOC_FL_ZERO_RANGE) {
 		ret = btrfs_zero_range(inode, offset, len, mode);
-		btrfs_inode_unlock(inode, BTRFS_ILOCK_MMAP);
+		btrfs_inode_unlock(BTRFS_I(inode), BTRFS_ILOCK_MMAP);
 		return ret;
 	}
 
@@ -3499,7 +3210,7 @@ out_unlock:
 	unlock_extent(&BTRFS_I(inode)->io_tree, alloc_start, locked_end,
 		      &cached_state);
 out:
-	btrfs_inode_unlock(inode, BTRFS_ILOCK_MMAP);
+	btrfs_inode_unlock(BTRFS_I(inode), BTRFS_ILOCK_MMAP);
 	extent_changeset_free(data_reserved);
 	return ret;
 }
@@ -3511,117 +3222,106 @@ out:
  * looping while it gets adjacent subranges, and merging them together.
  */
 static bool find_delalloc_subrange(struct btrfs_inode *inode, u64 start, u64 end,
+				   struct extent_state **cached_state,
+				   bool *search_io_tree,
 				   u64 *delalloc_start_ret, u64 *delalloc_end_ret)
 {
-	const u64 len = end + 1 - start;
-	struct extent_map_tree *em_tree = &inode->extent_tree;
-	struct extent_map *em;
-	u64 em_end;
-	u64 delalloc_len;
+	u64 len = end + 1 - start;
+	u64 delalloc_len = 0;
+	struct btrfs_ordered_extent *oe;
+	u64 oe_start;
+	u64 oe_end;
 
 	/*
 	 * Search the io tree first for EXTENT_DELALLOC. If we find any, it
 	 * means we have delalloc (dirty pages) for which writeback has not
 	 * started yet.
 	 */
-	*delalloc_start_ret = start;
-	delalloc_len = count_range_bits(&inode->io_tree, delalloc_start_ret, end,
-					len, EXTENT_DELALLOC, 1);
-	/*
-	 * If delalloc was found then *delalloc_start_ret has a sector size
-	 * aligned value (rounded down).
-	 */
-	if (delalloc_len > 0)
+	if (*search_io_tree) {
+		spin_lock(&inode->lock);
+		if (inode->delalloc_bytes > 0) {
+			spin_unlock(&inode->lock);
+			*delalloc_start_ret = start;
+			delalloc_len = count_range_bits(&inode->io_tree,
+							delalloc_start_ret, end,
+							len, EXTENT_DELALLOC, 1,
+							cached_state);
+		} else {
+			spin_unlock(&inode->lock);
+		}
+	}
+
+	if (delalloc_len > 0) {
+		/*
+		 * If delalloc was found then *delalloc_start_ret has a sector size
+		 * aligned value (rounded down).
+		 */
 		*delalloc_end_ret = *delalloc_start_ret + delalloc_len - 1;
 
+		if (*delalloc_start_ret == start) {
+			/* Delalloc for the whole range, nothing more to do. */
+			if (*delalloc_end_ret == end)
+				return true;
+			/* Else trim our search range for ordered extents. */
+			start = *delalloc_end_ret + 1;
+			len = end + 1 - start;
+		}
+	} else {
+		/* No delalloc, future calls don't need to search again. */
+		*search_io_tree = false;
+	}
+
 	/*
-	 * Now also check if there's any extent map in the range that does not
-	 * map to a hole or prealloc extent. We do this because:
+	 * Now also check if there's any ordered extent in the range.
+	 * We do this because:
 	 *
 	 * 1) When delalloc is flushed, the file range is locked, we clear the
-	 *    EXTENT_DELALLOC bit from the io tree and create an extent map for
-	 *    an allocated extent. So we might just have been called after
-	 *    delalloc is flushed and before the ordered extent completes and
-	 *    inserts the new file extent item in the subvolume's btree;
+	 *    EXTENT_DELALLOC bit from the io tree and create an extent map and
+	 *    an ordered extent for the write. So we might just have been called
+	 *    after delalloc is flushed and before the ordered extent completes
+	 *    and inserts the new file extent item in the subvolume's btree;
 	 *
-	 * 2) We may have an extent map created by flushing delalloc for a
+	 * 2) We may have an ordered extent created by flushing delalloc for a
 	 *    subrange that starts before the subrange we found marked with
 	 *    EXTENT_DELALLOC in the io tree.
+	 *
+	 * We could also use the extent map tree to find such delalloc that is
+	 * being flushed, but using the ordered extents tree is more efficient
+	 * because it's usually much smaller as ordered extents are removed from
+	 * the tree once they complete. With the extent maps, we mau have them
+	 * in the extent map tree for a very long time, and they were either
+	 * created by previous writes or loaded by read operations.
 	 */
-	read_lock(&em_tree->lock);
-	em = lookup_extent_mapping(em_tree, start, len);
-	read_unlock(&em_tree->lock);
-
-	/* extent_map_end() returns a non-inclusive end offset. */
-	em_end = em ? extent_map_end(em) : 0;
-
-	/*
-	 * If we have a hole/prealloc extent map, check the next one if this one
-	 * ends before our range's end.
-	 */
-	if (em && (em->block_start == EXTENT_MAP_HOLE ||
-		   test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) && em_end < end) {
-		struct extent_map *next_em;
+	oe = btrfs_lookup_first_ordered_range(inode, start, len);
+	if (!oe)
+		return (delalloc_len > 0);
 
-		read_lock(&em_tree->lock);
-		next_em = lookup_extent_mapping(em_tree, em_end, len - em_end);
-		read_unlock(&em_tree->lock);
+	/* The ordered extent may span beyond our search range. */
+	oe_start = max(oe->file_offset, start);
+	oe_end = min(oe->file_offset + oe->num_bytes - 1, end);
 
-		free_extent_map(em);
-		em_end = next_em ? extent_map_end(next_em) : 0;
-		em = next_em;
-	}
-
-	if (em && (em->block_start == EXTENT_MAP_HOLE ||
-		   test_bit(EXTENT_FLAG_PREALLOC, &em->flags))) {
-		free_extent_map(em);
-		em = NULL;
-	}
+	btrfs_put_ordered_extent(oe);
 
-	/*
-	 * No extent map or one for a hole or prealloc extent. Use the delalloc
-	 * range we found in the io tree if we have one.
-	 */
-	if (!em)
-		return (delalloc_len > 0);
-
-	/*
-	 * We don't have any range as EXTENT_DELALLOC in the io tree, so the
-	 * extent map is the only subrange representing delalloc.
-	 */
+	/* Don't have unflushed delalloc, return the ordered extent range. */
 	if (delalloc_len == 0) {
-		*delalloc_start_ret = em->start;
-		*delalloc_end_ret = min(end, em_end - 1);
-		free_extent_map(em);
+		*delalloc_start_ret = oe_start;
+		*delalloc_end_ret = oe_end;
 		return true;
 	}
 
 	/*
-	 * The extent map represents a delalloc range that starts before the
-	 * delalloc range we found in the io tree.
+	 * We have both unflushed delalloc (io_tree) and an ordered extent.
+	 * If the ranges are adjacent returned a combined range, otherwise
+	 * return the leftmost range.
 	 */
-	if (em->start < *delalloc_start_ret) {
-		*delalloc_start_ret = em->start;
-		/*
-		 * If the ranges are adjacent, return a combined range.
-		 * Otherwise return the extent map's range.
-		 */
-		if (em_end < *delalloc_start_ret)
-			*delalloc_end_ret = min(end, em_end - 1);
-
-		free_extent_map(em);
-		return true;
+	if (oe_start < *delalloc_start_ret) {
+		if (oe_end < *delalloc_start_ret)
+			*delalloc_end_ret = oe_end;
+		*delalloc_start_ret = oe_start;
+	} else if (*delalloc_end_ret + 1 == oe_start) {
+		*delalloc_end_ret = oe_end;
 	}
 
-	/*
-	 * The extent map starts after the delalloc range we found in the io
-	 * tree. If it's adjacent, return a combined range, otherwise return
-	 * the range found in the io tree.
-	 */
-	if (*delalloc_end_ret + 1 == em->start)
-		*delalloc_end_ret = min(end, em_end - 1);
-
-	free_extent_map(em);
 	return true;
 }
 
@@ -3633,6 +3333,8 @@ static bool find_delalloc_subrange(struct btrfs_inode *inode, u64 start, u64 end
  *                       sector size aligned.
  * @end:                 The end offset (inclusive value) of the search range.
  *                       It does not need to be sector size aligned.
+ * @cached_state:        Extent state record used for speeding up delalloc
+ *                       searches in the inode's io_tree. Can be NULL.
  * @delalloc_start_ret:  Output argument, set to the start offset of the
  *                       subrange found with delalloc (may not be sector size
  *                       aligned).
@@ -3644,10 +3346,12 @@ static bool find_delalloc_subrange(struct btrfs_inode *inode, u64 start, u64 end
  * end offsets of the subrange.
  */
 bool btrfs_find_delalloc_in_range(struct btrfs_inode *inode, u64 start, u64 end,
+				  struct extent_state **cached_state,
 				  u64 *delalloc_start_ret, u64 *delalloc_end_ret)
 {
 	u64 cur_offset = round_down(start, inode->root->fs_info->sectorsize);
 	u64 prev_delalloc_end = 0;
+	bool search_io_tree = true;
 	bool ret = false;
 
 	while (cur_offset < end) {
@@ -3656,6 +3360,7 @@ bool btrfs_find_delalloc_in_range(struct btrfs_inode *inode, u64 start, u64 end,
 		bool delalloc;
 
 		delalloc = find_delalloc_subrange(inode, cur_offset, end,
+						  cached_state, &search_io_tree,
 						  &delalloc_start,
 						  &delalloc_end);
 		if (!delalloc)
@@ -3701,13 +3406,14 @@ bool btrfs_find_delalloc_in_range(struct btrfs_inode *inode, u64 start, u64 end,
  * is found, it updates @start_ret with the start of the subrange.
  */
 static bool find_desired_extent_in_hole(struct btrfs_inode *inode, int whence,
+					struct extent_state **cached_state,
 					u64 start, u64 end, u64 *start_ret)
 {
 	u64 delalloc_start;
 	u64 delalloc_end;
 	bool delalloc;
 
-	delalloc = btrfs_find_delalloc_in_range(inode, start, end,
+	delalloc = btrfs_find_delalloc_in_range(inode, start, end, cached_state,
 						&delalloc_start, &delalloc_end);
 	if (delalloc && whence == SEEK_DATA) {
 		*start_ret = delalloc_start;
@@ -3750,11 +3456,13 @@ static bool find_desired_extent_in_hole(struct btrfs_inode *inode, int whence,
 	return false;
 }
 
-static loff_t find_desired_extent(struct btrfs_inode *inode, loff_t offset,
-				  int whence)
+static loff_t find_desired_extent(struct file *file, loff_t offset, int whence)
 {
+	struct btrfs_inode *inode = BTRFS_I(file->f_mapping->host);
+	struct btrfs_file_private *private = file->private_data;
 	struct btrfs_fs_info *fs_info = inode->root->fs_info;
 	struct extent_state *cached_state = NULL;
+	struct extent_state **delalloc_cached_state;
 	const loff_t i_size = i_size_read(&inode->vfs_inode);
 	const u64 ino = btrfs_ino(inode);
 	struct btrfs_root *root = inode->root;
@@ -3779,6 +3487,22 @@ static loff_t find_desired_extent(struct btrfs_inode *inode, loff_t offset,
 	    inode_get_bytes(&inode->vfs_inode) == i_size)
 		return i_size;
 
+	if (!private) {
+		private = kzalloc(sizeof(*private), GFP_KERNEL);
+		/*
+		 * No worries if memory allocation failed.
+		 * The private structure is used only for speeding up multiple
+		 * lseek SEEK_HOLE/DATA calls to a file when there's delalloc,
+		 * so everything will still be correct.
+		 */
+		file->private_data = private;
+	}
+
+	if (private)
+		delalloc_cached_state = &private->llseek_cached_state;
+	else
+		delalloc_cached_state = NULL;
+
 	/*
 	 * offset can be negative, in this case we start finding DATA/HOLE from
 	 * the very start of the file.
@@ -3856,6 +3580,7 @@ static loff_t find_desired_extent(struct btrfs_inode *inode, loff_t offset,
 				search_start = offset;
 
 			found = find_desired_extent_in_hole(inode, whence,
+							    delalloc_cached_state,
 							    search_start,
 							    key.offset - 1,
 							    &found_start);
@@ -3890,6 +3615,7 @@ static loff_t find_desired_extent(struct btrfs_inode *inode, loff_t offset,
 				search_start = offset;
 
 			found = find_desired_extent_in_hole(inode, whence,
+							    delalloc_cached_state,
 							    search_start,
 							    extent_end - 1,
 							    &found_start);
@@ -3931,7 +3657,8 @@ static loff_t find_desired_extent(struct btrfs_inode *inode, loff_t offset,
 
 	/* We have an implicit hole from the last extent found up to i_size. */
 	if (!found && start < i_size) {
-		found = find_desired_extent_in_hole(inode, whence, start,
+		found = find_desired_extent_in_hole(inode, whence,
+						    delalloc_cached_state, start,
 						    i_size - 1, &start);
 		if (!found)
 			start = i_size;
@@ -3959,9 +3686,9 @@ static loff_t btrfs_file_llseek(struct file *file, loff_t offset, int whence)
 		return generic_file_llseek(file, offset, whence);
 	case SEEK_DATA:
 	case SEEK_HOLE:
-		btrfs_inode_lock(inode, BTRFS_ILOCK_SHARED);
-		offset = find_desired_extent(BTRFS_I(inode), offset, whence);
-		btrfs_inode_unlock(inode, BTRFS_ILOCK_SHARED);
+		btrfs_inode_lock(BTRFS_I(inode), BTRFS_ILOCK_SHARED);
+		offset = find_desired_extent(file, offset, whence);
+		btrfs_inode_unlock(BTRFS_I(inode), BTRFS_ILOCK_SHARED);
 		break;
 	}
 
@@ -4016,7 +3743,7 @@ static ssize_t btrfs_direct_read(struct kiocb *iocb, struct iov_iter *to)
 	if (check_direct_read(btrfs_sb(inode->i_sb), to, iocb->ki_pos))
 		return 0;
 
-	btrfs_inode_lock(inode, BTRFS_ILOCK_SHARED);
+	btrfs_inode_lock(BTRFS_I(inode), BTRFS_ILOCK_SHARED);
 again:
 	/*
 	 * This is similar to what we do for direct IO writes, see the comment
@@ -4035,7 +3762,7 @@ again:
 	 */
 	pagefault_disable();
 	to->nofault = true;
-	ret = btrfs_dio_rw(iocb, to, read);
+	ret = btrfs_dio_read(iocb, to, read);
 	to->nofault = false;
 	pagefault_enable();
 
@@ -4065,7 +3792,7 @@ again:
 			goto again;
 		}
 	}
-	btrfs_inode_unlock(inode, BTRFS_ILOCK_SHARED);
+	btrfs_inode_unlock(BTRFS_I(inode), BTRFS_ILOCK_SHARED);
 	return ret < 0 ? ret : read;
 }
 
@@ -4102,23 +3829,6 @@ const struct file_operations btrfs_file_operations = {
 	.remap_file_range = btrfs_remap_file_range,
 };
 
-void __cold btrfs_auto_defrag_exit(void)
-{
-	kmem_cache_destroy(btrfs_inode_defrag_cachep);
-}
-
-int __init btrfs_auto_defrag_init(void)
-{
-	btrfs_inode_defrag_cachep = kmem_cache_create("btrfs_inode_defrag",
-					sizeof(struct inode_defrag), 0,
-					SLAB_MEM_SPREAD,
-					NULL);
-	if (!btrfs_inode_defrag_cachep)
-		return -ENOMEM;
-
-	return 0;
-}
-
 int btrfs_fdatawrite_range(struct inode *inode, loff_t start, loff_t end)
 {
 	int ret;