ext4: WARN if a full dir leaf block has only one dentry

[PATCH] ext4: WARN if a full dir leaf block has only one dentry

Posted by libaokun@huaweicloud.com 1 month, 2 weeks ago

From: Baokun Li <libaokun1@huawei.com>

The maximum length of a filename is 255 and the minimum block size is 1024,
so it is always guaranteed that the number of entries is greater than or
equal to 2 when do_split() is called. So unless ext4_dx_add_entry() and
make_indexed_dir() or some other functions are buggy, 'split == 0' will
not occur.

Setting 'continued' to 0 in this case masks the problem that the file
system has become corrupted, even though it prevents possible out-of-bounds
access. Hence WARN_ON_ONCE() is used to check if 'split' is 0, and if it is
then warns and returns an error to abort split.

Suggested-by: Theodore Ts'o <tytso@mit.edu>
Link: https://lore.kernel.org/r/20240823160518.GA424729@mit.edu
Signed-off-by: Baokun Li <libaokun1@huawei.com>
---
 fs/ext4/namei.c | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)

diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
index 790db7eac6c2..08d15cd2b594 100644
--- a/fs/ext4/namei.c
+++ b/fs/ext4/namei.c
@@ -2000,8 +2000,17 @@ static struct ext4_dir_entry_2 *do_split(handle_t *handle, struct inode *dir,
 	else
 		split = count/2;
 
+	if (WARN_ON_ONCE(split == 0)) {
+		/* Should never happen, but avoid out-of-bounds access below */
+		ext4_error_inode_block(dir, (*bh)->b_blocknr, 0,
+			"bad indexed directory? hash=%08x:%08x count=%d move=%u",
+			hinfo->hash, hinfo->minor_hash, count, move);
+		err = -EFSCORRUPTED;
+		goto out;
+	}
+
 	hash2 = map[split].hash;
-	continued = split > 0 ? hash2 == map[split - 1].hash : 0;
+	continued = hash2 == map[split - 1].hash;
 	dxtrace(printk(KERN_INFO "Split block %lu at %x, %i/%i\n",
 			(unsigned long)dx_get_block(frame->at),
 					hash2, split, count-split));
@@ -2043,10 +2052,11 @@ static struct ext4_dir_entry_2 *do_split(handle_t *handle, struct inode *dir,
 	return de;
 
 journal_error:
+	ext4_std_error(dir->i_sb, err);
+out:
 	brelse(*bh);
 	brelse(bh2);
 	*bh = NULL;
-	ext4_std_error(dir->i_sb, err);
 	return ERR_PTR(err);
 }
 
-- 
2.46.1

Re: [PATCH] ext4: WARN if a full dir leaf block has only one dentry

Posted by Theodore Ts'o 3 weeks, 6 days ago

On Tue, 08 Oct 2024 20:11:52 +0800, libaokun@huaweicloud.com wrote:
> The maximum length of a filename is 255 and the minimum block size is 1024,
> so it is always guaranteed that the number of entries is greater than or
> equal to 2 when do_split() is called. So unless ext4_dx_add_entry() and
> make_indexed_dir() or some other functions are buggy, 'split == 0' will
> not occur.
> 
> Setting 'continued' to 0 in this case masks the problem that the file
> system has become corrupted, even though it prevents possible out-of-bounds
> access. Hence WARN_ON_ONCE() is used to check if 'split' is 0, and if it is
> then warns and returns an error to abort split.
> 
> [...]

Applied, thanks!

[1/1] ext4: WARN if a full dir leaf block has only one dentry
      commit: 4748df069284704b166b98737352677c525076bd

Best regards,
-- 
Theodore Ts'o <tytso@mit.edu>

Re: [PATCH] ext4: WARN if a full dir leaf block has only one dentry

Posted by Jan Kara 1 month, 2 weeks ago

On Tue 08-10-24 20:11:52, libaokun@huaweicloud.com wrote:
> From: Baokun Li <libaokun1@huawei.com>
> 
> The maximum length of a filename is 255 and the minimum block size is 1024,
> so it is always guaranteed that the number of entries is greater than or
> equal to 2 when do_split() is called. So unless ext4_dx_add_entry() and
> make_indexed_dir() or some other functions are buggy, 'split == 0' will
> not occur.
> 
> Setting 'continued' to 0 in this case masks the problem that the file
> system has become corrupted, even though it prevents possible out-of-bounds
> access. Hence WARN_ON_ONCE() is used to check if 'split' is 0, and if it is
> then warns and returns an error to abort split.
> 
> Suggested-by: Theodore Ts'o <tytso@mit.edu>
> Link: https://lore.kernel.org/r/20240823160518.GA424729@mit.edu
> Signed-off-by: Baokun Li <libaokun1@huawei.com>

Looks good. Feel free to add:

Reviewed-by: Jan Kara <jack@suse.cz>

								Honza

> ---
>  fs/ext4/namei.c | 14 ++++++++++++--
>  1 file changed, 12 insertions(+), 2 deletions(-)
> 
> diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
> index 790db7eac6c2..08d15cd2b594 100644
> --- a/fs/ext4/namei.c
> +++ b/fs/ext4/namei.c
> @@ -2000,8 +2000,17 @@ static struct ext4_dir_entry_2 *do_split(handle_t *handle, struct inode *dir,
>  	else
>  		split = count/2;
>  
> +	if (WARN_ON_ONCE(split == 0)) {
> +		/* Should never happen, but avoid out-of-bounds access below */
> +		ext4_error_inode_block(dir, (*bh)->b_blocknr, 0,
> +			"bad indexed directory? hash=%08x:%08x count=%d move=%u",
> +			hinfo->hash, hinfo->minor_hash, count, move);
> +		err = -EFSCORRUPTED;
> +		goto out;
> +	}
> +
>  	hash2 = map[split].hash;
> -	continued = split > 0 ? hash2 == map[split - 1].hash : 0;
> +	continued = hash2 == map[split - 1].hash;
>  	dxtrace(printk(KERN_INFO "Split block %lu at %x, %i/%i\n",
>  			(unsigned long)dx_get_block(frame->at),
>  					hash2, split, count-split));
> @@ -2043,10 +2052,11 @@ static struct ext4_dir_entry_2 *do_split(handle_t *handle, struct inode *dir,
>  	return de;
>  
>  journal_error:
> +	ext4_std_error(dir->i_sb, err);
> +out:
>  	brelse(*bh);
>  	brelse(bh2);
>  	*bh = NULL;
> -	ext4_std_error(dir->i_sb, err);
>  	return ERR_PTR(err);
>  }
>  
> -- 
> 2.46.1
> 
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR