ext4: allow directory holes

commit 4e19d6b65fb4fc42e352ce9883649e049da14743 upstream.

The largedir feature was intended to allow ext4 directories to have
unmapped directory blocks (e.g., directory holes).  And so the
released e2fsprogs no longer enforces this for largedir file systems;
however, the corresponding change to the kernel-side code was not made.

This commit fixes this oversight.

Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Cc: stable@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

2 files changed, 46 insertions, 18 deletions

diff --git a/fs/ext4/dir.c b/fs/ext4/dir.c
index dc676714454a..446b6c375b6f 100644
--- a/fs/ext4/dir.c
+++ b/fs/ext4/dir.c
@@ -107,7 +107,6 @@ static int ext4_readdir(struct file *file, struct dir_context *ctx)
 	struct inode *inode = file_inode(file);
 	struct super_block *sb = inode->i_sb;
 	struct buffer_head *bh = NULL;
-	int dir_has_error = 0;
 	struct fscrypt_str fstr = FSTR_INIT(NULL, 0);
 
 	if (ext4_encrypted_inode(inode)) {
@@ -143,8 +142,6 @@ static int ext4_readdir(struct file *file, struct dir_context *ctx)
 			return err;
 	}
 
-	offset = ctx->pos & (sb->s_blocksize - 1);
-
 	while (ctx->pos < inode->i_size) {
 		struct ext4_map_blocks map;
 
@@ -153,9 +150,18 @@ static int ext4_readdir(struct file *file, struct dir_context *ctx)
 			goto errout;
 		}
 		cond_resched();
+		offset = ctx->pos & (sb->s_blocksize - 1);
 		map.m_lblk = ctx->pos >> EXT4_BLOCK_SIZE_BITS(sb);
 		map.m_len = 1;
 		err = ext4_map_blocks(NULL, inode, &map, 0);
+		if (err == 0) {
+			/* m_len should never be zero but let's avoid
+			 * an infinite loop if it somehow is */
+			if (map.m_len == 0)
+				map.m_len = 1;
+			ctx->pos += map.m_len * sb->s_blocksize;
+			continue;
+		}
 		if (err > 0) {
 			pgoff_t index = map.m_pblk >>
 					(PAGE_SHIFT - inode->i_blkbits);
@@ -174,13 +180,6 @@ static int ext4_readdir(struct file *file, struct dir_context *ctx)
 		}
 
 		if (!bh) {
-			if (!dir_has_error) {
-				EXT4_ERROR_FILE(file, 0,
-						"directory contains a "
-						"hole at offset %llu",
-					   (unsigned long long) ctx->pos);
-				dir_has_error = 1;
-			}
 			/* corrupt size?  Maybe no more blocks to read */
 			if (ctx->pos > inode->i_blocks << 9)
 				break;
diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
index 070660cb5b91..162e853dc5d6 100644
--- a/fs/ext4/namei.c
+++ b/fs/ext4/namei.c
@@ -80,8 +80,18 @@ static struct buffer_head *ext4_append(handle_t *handle,
 static int ext4_dx_csum_verify(struct inode *inode,
 			       struct ext4_dir_entry *dirent);
 
+/*
+ * Hints to ext4_read_dirblock regarding whether we expect a directory
+ * block being read to be an index block, or a block containing
+ * directory entries (and if the latter, whether it was found via a
+ * logical block in an htree index block).  This is used to control
+ * what sort of sanity checkinig ext4_read_dirblock() will do on the
+ * directory block read from the storage device.  EITHER will means
+ * the caller doesn't know what kind of directory block will be read,
+ * so no specific verification will be done.
+ */
 typedef enum {
-	EITHER, INDEX, DIRENT
+	EITHER, INDEX, DIRENT, DIRENT_HTREE
 } dirblock_type_t;
 
 #define ext4_read_dirblock(inode, block, type) \
@@ -107,11 +117,14 @@ static struct buffer_head *__ext4_read_dirblock(struct inode *inode,
 
 		return bh;
 	}
-	if (!bh) {
+	if (!bh && (type == INDEX || type == DIRENT_HTREE)) {
 		ext4_error_inode(inode, func, line, block,
-				 "Directory hole found");
+				 "Directory hole found for htree %s block",
+				 (type == INDEX) ? "index" : "leaf");
 		return ERR_PTR(-EFSCORRUPTED);
 	}
+	if (!bh)
+		return NULL;
 	dirent = (struct ext4_dir_entry *) bh->b_data;
 	/* Determine whether or not we have an index block */
 	if (is_dx(inode)) {
@@ -978,7 +991,7 @@ static int htree_dirblock_to_tree(struct file *dir_file,
 
 	dxtrace(printk(KERN_INFO "In htree dirblock_to_tree: block %lu\n",
 							(unsigned long)block));
-	bh = ext4_read_dirblock(dir, block, DIRENT);
+	bh = ext4_read_dirblock(dir, block, DIRENT_HTREE);
 	if (IS_ERR(bh))
 		return PTR_ERR(bh);
 
@@ -1508,7 +1521,7 @@ static struct buffer_head * ext4_dx_find_entry(struct inode *dir,
 		return (struct buffer_head *) frame;
 	do {
 		block = dx_get_block(frame->at);
-		bh = ext4_read_dirblock(dir, block, DIRENT);
+		bh = ext4_read_dirblock(dir, block, DIRENT_HTREE);
 		if (IS_ERR(bh))
 			goto errout;
 
@@ -2088,6 +2101,11 @@ static int ext4_add_entry(handle_t *handle, struct dentry *dentry,
 	blocks = dir->i_size >> sb->s_blocksize_bits;
 	for (block = 0; block < blocks; block++) {
 		bh = ext4_read_dirblock(dir, block, DIRENT);
+		if (bh == NULL) {
+			bh = ext4_bread(handle, dir, block,
+					EXT4_GET_BLOCKS_CREATE);
+			goto add_to_new_block;
+		}
 		if (IS_ERR(bh)) {
 			retval = PTR_ERR(bh);
 			bh = NULL;
@@ -2108,6 +2126,7 @@ static int ext4_add_entry(handle_t *handle, struct dentry *dentry,
 		brelse(bh);
 	}
 	bh = ext4_append(handle, dir, &block);
+add_to_new_block:
 	if (IS_ERR(bh)) {
 		retval = PTR_ERR(bh);
 		bh = NULL;
@@ -2152,7 +2171,7 @@ again:
 		return PTR_ERR(frame);
 	entries = frame->entries;
 	at = frame->at;
-	bh = ext4_read_dirblock(dir, dx_get_block(frame->at), DIRENT);
+	bh = ext4_read_dirblock(dir, dx_get_block(frame->at), DIRENT_HTREE);
 	if (IS_ERR(bh)) {
 		err = PTR_ERR(bh);
 		bh = NULL;
@@ -2700,7 +2719,10 @@ bool ext4_empty_dir(struct inode *inode)
 		EXT4_ERROR_INODE(inode, "invalid size");
 		return true;
 	}
-	bh = ext4_read_dirblock(inode, 0, EITHER);
+	/* The first directory block must not be a hole,
+	 * so treat it as DIRENT_HTREE
+	 */
+	bh = ext4_read_dirblock(inode, 0, DIRENT_HTREE);
 	if (IS_ERR(bh))
 		return true;
 
@@ -2722,6 +2744,10 @@ bool ext4_empty_dir(struct inode *inode)
 			brelse(bh);
 			lblock = offset >> EXT4_BLOCK_SIZE_BITS(sb);
 			bh = ext4_read_dirblock(inode, lblock, EITHER);
+			if (bh == NULL) {
+				offset += sb->s_blocksize;
+				continue;
+			}
 			if (IS_ERR(bh))
 				return true;
 			de = (struct ext4_dir_entry_2 *) bh->b_data;
@@ -3292,7 +3318,10 @@ static struct buffer_head *ext4_get_first_dir_block(handle_t *handle,
 	struct buffer_head *bh;
 
 	if (!ext4_has_inline_data(inode)) {
-		bh = ext4_read_dirblock(inode, 0, EITHER);
+		/* The first directory block must not be a hole, so
+		 * treat it as DIRENT_HTREE
+		 */
+		bh = ext4_read_dirblock(inode, 0, DIRENT_HTREE);
 		if (IS_ERR(bh)) {
 			*retval = PTR_ERR(bh);
 			return NULL;