1 files changed, 59 insertions, 66 deletions

diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index be9e3900cce8..864c08d08a35 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -2014,14 +2014,8 @@ static ssize_t btrfs_file_write_iter(struct kiocb *iocb,
 	else
 		num_written = btrfs_buffered_write(iocb, from);
 
-	/*
-	 * We also have to set last_sub_trans to the current log transid,
-	 * otherwise subsequent syncs to a file that's been synced in this
-	 * transaction will appear to have already occurred.
-	 */
-	spin_lock(&inode->lock);
-	inode->last_sub_trans = inode->root->log_transid;
-	spin_unlock(&inode->lock);
+	btrfs_set_inode_last_sub_trans(inode);
+
 	if (num_written > 0)
 		num_written = generic_write_sync(iocb, num_written);
 
@@ -2122,7 +2116,7 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 	if (ret)
 		goto out;
 
-	inode_lock(inode);
+	btrfs_inode_lock(inode, BTRFS_ILOCK_MMAP);
 
 	atomic_inc(&root->log_batch);
 
@@ -2135,11 +2129,11 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 			     &BTRFS_I(inode)->runtime_flags);
 
 	/*
-	 * Before we acquired the inode's lock, someone may have dirtied more
-	 * pages in the target range. We need to make sure that writeback for
-	 * any such pages does not start while we are logging the inode, because
-	 * if it does, any of the following might happen when we are not doing a
-	 * full inode sync:
+	 * Before we acquired the inode's lock and the mmap lock, someone may
+	 * have dirtied more pages in the target range. We need to make sure
+	 * that writeback for any such pages does not start while we are logging
+	 * the inode, because if it does, any of the following might happen when
+	 * we are not doing a full inode sync:
 	 *
 	 * 1) We log an extent after its writeback finishes but before its
 	 *    checksums are added to the csum tree, leading to -EIO errors
@@ -2154,7 +2148,7 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 	 */
 	ret = start_ordered_ops(inode, start, end);
 	if (ret) {
-		inode_unlock(inode);
+		btrfs_inode_unlock(inode, BTRFS_ILOCK_MMAP);
 		goto out;
 	}
 
@@ -2255,7 +2249,7 @@ 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.
 	 */
-	inode_unlock(inode);
+	btrfs_inode_unlock(inode, BTRFS_ILOCK_MMAP);
 
 	if (ret != BTRFS_NO_LOG_SYNC) {
 		if (!ret) {
@@ -2285,7 +2279,7 @@ out:
 
 out_release_extents:
 	btrfs_release_log_ctx_extents(&ctx);
-	inode_unlock(inode);
+	btrfs_inode_unlock(inode, BTRFS_ILOCK_MMAP);
 	goto out;
 }
 
@@ -2605,16 +2599,17 @@ static int btrfs_insert_replace_extent(struct btrfs_trans_handle *trans,
  * extents without inserting a new one, so we must abort the transaction to avoid
  * a corruption.
  */
-int btrfs_replace_file_extents(struct inode *inode, struct btrfs_path *path,
-			   const u64 start, const u64 end,
-			   struct btrfs_replace_extent_info *extent_info,
-			   struct btrfs_trans_handle **trans_out)
+int btrfs_replace_file_extents(struct btrfs_inode *inode,
+			       struct btrfs_path *path, const u64 start,
+			       const u64 end,
+			       struct btrfs_replace_extent_info *extent_info,
+			       struct btrfs_trans_handle **trans_out)
 {
 	struct btrfs_drop_extents_args drop_args = { 0 };
-	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
+	struct btrfs_root *root = inode->root;
+	struct btrfs_fs_info *fs_info = root->fs_info;
 	u64 min_size = btrfs_calc_insert_metadata_size(fs_info, 1);
-	u64 ino_size = round_up(inode->i_size, fs_info->sectorsize);
-	struct btrfs_root *root = BTRFS_I(inode)->root;
+	u64 ino_size = round_up(inode->vfs_inode.i_size, fs_info->sectorsize);
 	struct btrfs_trans_handle *trans = NULL;
 	struct btrfs_block_rsv *rsv;
 	unsigned int rsv_count;
@@ -2662,10 +2657,10 @@ int btrfs_replace_file_extents(struct inode *inode, struct btrfs_path *path,
 	drop_args.drop_cache = true;
 	while (cur_offset < end) {
 		drop_args.start = cur_offset;
-		ret = btrfs_drop_extents(trans, root, BTRFS_I(inode), &drop_args);
+		ret = btrfs_drop_extents(trans, root, inode, &drop_args);
 		/* If we are punching a hole decrement the inode's byte count */
 		if (!extent_info)
-			btrfs_update_inode_bytes(BTRFS_I(inode), 0,
+			btrfs_update_inode_bytes(inode, 0,
 						 drop_args.bytes_found);
 		if (ret != -ENOSPC) {
 			/*
@@ -2685,8 +2680,8 @@ int btrfs_replace_file_extents(struct inode *inode, struct btrfs_path *path,
 
 		if (!extent_info && cur_offset < drop_args.drop_end &&
 		    cur_offset < ino_size) {
-			ret = fill_holes(trans, BTRFS_I(inode), path,
-					 cur_offset, drop_args.drop_end);
+			ret = fill_holes(trans, inode, path, cur_offset,
+					 drop_args.drop_end);
 			if (ret) {
 				/*
 				 * If we failed then we didn't insert our hole
@@ -2704,7 +2699,7 @@ int btrfs_replace_file_extents(struct inode *inode, struct btrfs_path *path,
 			 * know to not set disk_i_size in this area until a new
 			 * file extent is inserted here.
 			 */
-			ret = btrfs_inode_clear_file_extent_range(BTRFS_I(inode),
+			ret = btrfs_inode_clear_file_extent_range(inode,
 					cur_offset,
 					drop_args.drop_end - cur_offset);
 			if (ret) {
@@ -2723,8 +2718,8 @@ int btrfs_replace_file_extents(struct inode *inode, struct btrfs_path *path,
 			u64 replace_len = drop_args.drop_end -
 					  extent_info->file_offset;
 
-			ret = btrfs_insert_replace_extent(trans, BTRFS_I(inode),
-					path, extent_info, replace_len,
+			ret = btrfs_insert_replace_extent(trans, inode,	path,
+					extent_info, replace_len,
 					drop_args.bytes_found);
 			if (ret) {
 				btrfs_abort_transaction(trans, ret);
@@ -2735,9 +2730,7 @@ int btrfs_replace_file_extents(struct inode *inode, struct btrfs_path *path,
 			extent_info->file_offset += replace_len;
 		}
 
-		cur_offset = drop_args.drop_end;
-
-		ret = btrfs_update_inode(trans, root, BTRFS_I(inode));
+		ret = btrfs_update_inode(trans, root, inode);
 		if (ret)
 			break;
 
@@ -2756,9 +2749,10 @@ int btrfs_replace_file_extents(struct inode *inode, struct btrfs_path *path,
 		BUG_ON(ret);	/* shouldn't happen */
 		trans->block_rsv = rsv;
 
-		if (!extent_info) {
-			ret = find_first_non_hole(BTRFS_I(inode), &cur_offset,
-						  &len);
+		cur_offset = drop_args.drop_end;
+		len = end - cur_offset;
+		if (!extent_info && len) {
+			ret = find_first_non_hole(inode, &cur_offset, &len);
 			if (unlikely(ret < 0))
 				break;
 			if (ret && !len) {
@@ -2771,14 +2765,11 @@ int btrfs_replace_file_extents(struct inode *inode, struct btrfs_path *path,
 	/*
 	 * If we were cloning, force the next fsync to be a full one since we
 	 * we replaced (or just dropped in the case of cloning holes when
-	 * NO_HOLES is enabled) extents and extent maps.
-	 * This is for the sake of simplicity, and cloning into files larger
-	 * than 16Mb would force the full fsync any way (when
-	 * try_release_extent_mapping() is invoked during page cache truncation.
+	 * NO_HOLES is enabled) file extent items and did not setup new extent
+	 * maps for the replacement extents (or holes).
 	 */
 	if (extent_info && !extent_info->is_new_extent)
-		set_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
-			&BTRFS_I(inode)->runtime_flags);
+		set_bit(BTRFS_INODE_NEEDS_FULL_SYNC, &inode->runtime_flags);
 
 	if (ret)
 		goto out_trans;
@@ -2804,8 +2795,8 @@ int btrfs_replace_file_extents(struct inode *inode, struct btrfs_path *path,
 	 */
 	if (!extent_info && cur_offset < ino_size &&
 	    cur_offset < drop_args.drop_end) {
-		ret = fill_holes(trans, BTRFS_I(inode), path,
-				 cur_offset, drop_args.drop_end);
+		ret = fill_holes(trans, inode, path, cur_offset,
+				 drop_args.drop_end);
 		if (ret) {
 			/* Same comment as above. */
 			btrfs_abort_transaction(trans, ret);
@@ -2813,8 +2804,8 @@ int btrfs_replace_file_extents(struct inode *inode, struct btrfs_path *path,
 		}
 	} else if (!extent_info && cur_offset < drop_args.drop_end) {
 		/* See the comment in the loop above for the reasoning here. */
-		ret = btrfs_inode_clear_file_extent_range(BTRFS_I(inode),
-				cur_offset, drop_args.drop_end - cur_offset);
+		ret = btrfs_inode_clear_file_extent_range(inode, cur_offset,
+					drop_args.drop_end - cur_offset);
 		if (ret) {
 			btrfs_abort_transaction(trans, ret);
 			goto out_trans;
@@ -2822,7 +2813,7 @@ int btrfs_replace_file_extents(struct inode *inode, struct btrfs_path *path,
 
 	}
 	if (extent_info) {
-		ret = btrfs_insert_replace_extent(trans, BTRFS_I(inode), path,
+		ret = btrfs_insert_replace_extent(trans, inode, path,
 				extent_info, extent_info->data_len,
 				drop_args.bytes_found);
 		if (ret) {
@@ -2868,7 +2859,7 @@ static int btrfs_punch_hole(struct inode *inode, loff_t offset, loff_t len)
 	if (ret)
 		return ret;
 
-	inode_lock(inode);
+	btrfs_inode_lock(inode, BTRFS_ILOCK_MMAP);
 	ino_size = round_up(inode->i_size, fs_info->sectorsize);
 	ret = find_first_non_hole(BTRFS_I(inode), &offset, &len);
 	if (ret < 0)
@@ -2908,7 +2899,7 @@ static int btrfs_punch_hole(struct inode *inode, loff_t offset, loff_t len)
 		truncated_block = true;
 		ret = btrfs_truncate_block(BTRFS_I(inode), offset, 0, 0);
 		if (ret) {
-			inode_unlock(inode);
+			btrfs_inode_unlock(inode, BTRFS_ILOCK_MMAP);
 			return ret;
 		}
 	}
@@ -2967,8 +2958,8 @@ static int btrfs_punch_hole(struct inode *inode, loff_t offset, loff_t len)
 		goto out;
 	}
 
-	ret = btrfs_replace_file_extents(inode, path, lockstart, lockend, NULL,
-				     &trans);
+	ret = btrfs_replace_file_extents(BTRFS_I(inode), path, lockstart,
+					 lockend, NULL, &trans);
 	btrfs_free_path(path);
 	if (ret)
 		goto out;
@@ -3009,7 +3000,7 @@ out_only_mutex:
 				ret = ret2;
 		}
 	}
-	inode_unlock(inode);
+	btrfs_inode_unlock(inode, BTRFS_ILOCK_MMAP);
 	return ret;
 }
 
@@ -3260,8 +3251,11 @@ reserve_space:
 			goto out;
 		ret = btrfs_qgroup_reserve_data(BTRFS_I(inode), &data_reserved,
 						alloc_start, bytes_to_reserve);
-		if (ret)
+		if (ret) {
+			unlock_extent_cached(&BTRFS_I(inode)->io_tree, lockstart,
+					     lockend, &cached_state);
 			goto out;
+		}
 		ret = btrfs_prealloc_file_range(inode, mode, alloc_start,
 						alloc_end - alloc_start,
 						i_blocksize(inode),
@@ -3332,7 +3326,7 @@ static long btrfs_fallocate(struct file *file, int mode,
 			return ret;
 	}
 
-	btrfs_inode_lock(inode, 0);
+	btrfs_inode_lock(inode, BTRFS_ILOCK_MMAP);
 
 	if (!(mode & FALLOC_FL_KEEP_SIZE) && offset + len > inode->i_size) {
 		ret = inode_newsize_ok(inode, offset + len);
@@ -3374,7 +3368,7 @@ static long btrfs_fallocate(struct file *file, int mode,
 
 	if (mode & FALLOC_FL_ZERO_RANGE) {
 		ret = btrfs_zero_range(inode, offset, len, mode);
-		inode_unlock(inode);
+		btrfs_inode_unlock(inode, BTRFS_ILOCK_MMAP);
 		return ret;
 	}
 
@@ -3484,7 +3478,7 @@ out_unlock:
 	unlock_extent_cached(&BTRFS_I(inode)->io_tree, alloc_start, locked_end,
 			     &cached_state);
 out:
-	inode_unlock(inode);
+	btrfs_inode_unlock(inode, BTRFS_ILOCK_MMAP);
 	/* Let go of our reservation. */
 	if (ret != 0 && !(mode & FALLOC_FL_ZERO_RANGE))
 		btrfs_free_reserved_data_space(BTRFS_I(inode), data_reserved,
@@ -3493,13 +3487,13 @@ out:
 	return ret;
 }
 
-static loff_t find_desired_extent(struct inode *inode, loff_t offset,
+static loff_t find_desired_extent(struct btrfs_inode *inode, loff_t offset,
 				  int whence)
 {
-	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
+	struct btrfs_fs_info *fs_info = inode->root->fs_info;
 	struct extent_map *em = NULL;
 	struct extent_state *cached_state = NULL;
-	loff_t i_size = inode->i_size;
+	loff_t i_size = inode->vfs_inode.i_size;
 	u64 lockstart;
 	u64 lockend;
 	u64 start;
@@ -3522,11 +3516,10 @@ static loff_t find_desired_extent(struct inode *inode, loff_t offset,
 	lockend--;
 	len = lockend - lockstart + 1;
 
-	lock_extent_bits(&BTRFS_I(inode)->io_tree, lockstart, lockend,
-			 &cached_state);
+	lock_extent_bits(&inode->io_tree, lockstart, lockend, &cached_state);
 
 	while (start < i_size) {
-		em = btrfs_get_extent_fiemap(BTRFS_I(inode), start, len);
+		em = btrfs_get_extent_fiemap(inode, start, len);
 		if (IS_ERR(em)) {
 			ret = PTR_ERR(em);
 			em = NULL;
@@ -3548,7 +3541,7 @@ static loff_t find_desired_extent(struct inode *inode, loff_t offset,
 		cond_resched();
 	}
 	free_extent_map(em);
-	unlock_extent_cached(&BTRFS_I(inode)->io_tree, lockstart, lockend,
+	unlock_extent_cached(&inode->io_tree, lockstart, lockend,
 			     &cached_state);
 	if (ret) {
 		offset = ret;
@@ -3572,7 +3565,7 @@ static loff_t btrfs_file_llseek(struct file *file, loff_t offset, int whence)
 	case SEEK_DATA:
 	case SEEK_HOLE:
 		btrfs_inode_lock(inode, BTRFS_ILOCK_SHARED);
-		offset = find_desired_extent(inode, offset, whence);
+		offset = find_desired_extent(BTRFS_I(inode), offset, whence);
 		btrfs_inode_unlock(inode, BTRFS_ILOCK_SHARED);
 		break;
 	}
@@ -3634,7 +3627,7 @@ static ssize_t btrfs_file_read_iter(struct kiocb *iocb, struct iov_iter *to)
 			return ret;
 	}
 
-	return generic_file_buffered_read(iocb, to, ret);
+	return filemap_read(iocb, to, ret);
 }
 
 const struct file_operations btrfs_file_operations = {