[PATCH 5/5] f2fs: Handle casefolding with Encryption

[Eric Biggers]

On Wed, Sep 23, 2020 at 01:01:51AM +0000, Daniel Rosenberg wrote:
> Expand f2fs's casefolding support to include encrypted directories.  To
> index casefolded+encrypted directories, we use the SipHash of the
> casefolded name, keyed by a key derived from the directory's fscrypt
> master key.  This ensures that the dirhash doesn't leak information
> about the plaintext filenames.
> 
> Encryption keys are unavailable during roll-forward recovery, so we
> can't compute the dirhash when recovering a new dentry in an encrypted +
> casefolded directory.  To avoid having to force a checkpoint when a new
> file is fsync'ed, store the dirhash on-disk appended to i_name.
> 
> This patch incorporates work by Eric Biggers <ebiggers at google.com>
> and Jaegeuk Kim <jaegeuk at kernel.org>.
> 
> Co-developed-by: Eric Biggers <ebiggers at google.com>
> Signed-off-by: Eric Biggers <ebiggers at google.com>
> Signed-off-by: Daniel Rosenberg <drosen at google.com>
> ---

Generally looks good.  If it's needed, you can add:

Reviewed-by: Eric Biggers <ebiggers at google.com>

(Though, some may claim I can't give Reviewed-by since this patch already has my
Co-developed-by.)

One comment below, though:

> @@ -218,9 +219,28 @@ static bool f2fs_match_ci_name(const struct inode *dir, const struct qstr *name,
>  {
>  	const struct super_block *sb = dir->i_sb;
>  	const struct unicode_map *um = sb->s_encoding;
> +	struct fscrypt_str decrypted_name = FSTR_INIT(NULL, de_name_len);
>  	struct qstr entry = QSTR_INIT(de_name, de_name_len);
>  	int res;
>  
> +	if (IS_ENCRYPTED(dir)) {
> +		const struct fscrypt_str encrypted_name =
> +			FSTR_INIT((u8 *)de_name, de_name_len);
> +
> +		if (WARN_ON_ONCE(!fscrypt_has_encryption_key(dir)))
> +			return false;
> +
> +		decrypted_name.name = kmalloc(de_name_len, GFP_KERNEL);
> +		if (!decrypted_name.name)
> +			return false;
> +		res = fscrypt_fname_disk_to_usr(dir, 0, 0, &encrypted_name,
> +						&decrypted_name);
> +		if (res < 0)
> +			goto out;

We probably should be passing up errors from here to f2fs_match_name(), then to
f2fs_find_target_dentry(), then to f2fs_find_in_inline_dir() or find_in_block().

Ignoring the filename may be okay if fscrypt_fname_disk_to_usr() returns
-EUCLEAN, indicating that it's invalid.  However, if the error is -ENOMEM,
either from the kmalloc() or from fscrypt_fname_disk_to_usr(), then the caller
should receive an error rather than the filename being ignored.

- Eric