If casefolding the filename fails, we'll be leaking fscrypt_buf name.
Make sure we free it in the error paths of ext4_fname_setup_filename() and
ext4_fname_prepare_lookup() functions.

Fixes: 1ae98e295fa2 ("ext4: optimize match for casefolded encrypted dirs")
Signed-off-by: Luís Henriques <lhenriques@suse.de>
---
Changes since v1:
- Include fix to ext4_fname_prepare_lookup() as well
- Add 'Fixes:' tag

 fs/ext4/crypto.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/fs/ext4/crypto.c b/fs/ext4/crypto.c
index e20ac0654b3f..3c05c7f3415b 100644
--- a/fs/ext4/crypto.c
+++ b/fs/ext4/crypto.c
@@ -33,6 +33,8 @@ int ext4_fname_setup_filename(struct inode *dir, const struct qstr *iname,
 
 #if IS_ENABLED(CONFIG_UNICODE)
 	err = ext4_fname_setup_ci_filename(dir, iname, fname);
+	if (err)
+		fscrypt_free_filename(&name);
 #endif
 	return err;
 }
@@ -51,6 +53,8 @@ int ext4_fname_prepare_lookup(struct inode *dir, struct dentry *dentry,
 
 #if IS_ENABLED(CONFIG_UNICODE)
 	err = ext4_fname_setup_ci_filename(dir, &dentry->d_name, fname);
+	if (err)
+		fscrypt_free_filename(&name);
 #endif
 	return err;
 }

On Wed, Aug 02, 2023 at 10:49:31AM +0100, Luís Henriques wrote:
> If casefolding the filename fails, we'll be leaking fscrypt_buf name.

fscrypt_buf => fscrypt_name

> diff --git a/fs/ext4/crypto.c b/fs/ext4/crypto.c
> index e20ac0654b3f..3c05c7f3415b 100644
> --- a/fs/ext4/crypto.c
> +++ b/fs/ext4/crypto.c
> @@ -33,6 +33,8 @@ int ext4_fname_setup_filename(struct inode *dir, const struct qstr *iname,
>	struct fscrypt_name name;
>	int err;
>
>	err = fscrypt_setup_filename(dir, iname, lookup, &name);
>	if (err)
>		return err;
>
>	ext4_fname_from_fscrypt_name(fname, &name);
>
>  #if IS_ENABLED(CONFIG_UNICODE)
>  	err = ext4_fname_setup_ci_filename(dir, iname, fname);
> +	if (err)
> +		fscrypt_free_filename(&name);
>  #endif
>  	return err;
>  }
> @@ -51,6 +53,8 @@ int ext4_fname_prepare_lookup(struct inode *dir, struct dentry *dentry,
>	struct fscrypt_name name;
>	int err;
>
>	err = fscrypt_prepare_lookup(dir, dentry, &name);
>	if (err)
>		return err;
>
>	ext4_fname_from_fscrypt_name(fname, &name);
>
>  #if IS_ENABLED(CONFIG_UNICODE)
>  	err = ext4_fname_setup_ci_filename(dir, &dentry->d_name, fname);
> +	if (err)
> +		fscrypt_free_filename(&name);
>  #endif
>  	return err;
>  }

This works, but it's a bit weird that the freeing happens on the original struct
fscrypt_name after it has already been "moved" to the struct ext4_filename by
ext4_fname_from_fscrypt_name().  That leaves a dangling pointer in the struct
ext4_filename.  Maybe you should call ext4_fname_free_filename() instead, even
though it would do some unnecessary work?

- Eric

Eric Biggers <ebiggers@kernel.org> writes:

> On Wed, Aug 02, 2023 at 10:49:31AM +0100, Luís Henriques wrote:
>> If casefolding the filename fails, we'll be leaking fscrypt_buf name.
>
> fscrypt_buf => fscrypt_name
>
>> diff --git a/fs/ext4/crypto.c b/fs/ext4/crypto.c
>> index e20ac0654b3f..3c05c7f3415b 100644
>> --- a/fs/ext4/crypto.c
>> +++ b/fs/ext4/crypto.c
>> @@ -33,6 +33,8 @@ int ext4_fname_setup_filename(struct inode *dir, const struct qstr *iname,
>>	struct fscrypt_name name;
>>	int err;
>>
>>	err = fscrypt_setup_filename(dir, iname, lookup, &name);
>>	if (err)
>>		return err;
>>
>>	ext4_fname_from_fscrypt_name(fname, &name);
>>
>>  #if IS_ENABLED(CONFIG_UNICODE)
>>  	err = ext4_fname_setup_ci_filename(dir, iname, fname);
>> +	if (err)
>> +		fscrypt_free_filename(&name);
>>  #endif
>>  	return err;
>>  }
>> @@ -51,6 +53,8 @@ int ext4_fname_prepare_lookup(struct inode *dir, struct dentry *dentry,
>>	struct fscrypt_name name;
>>	int err;
>>
>>	err = fscrypt_prepare_lookup(dir, dentry, &name);
>>	if (err)
>>		return err;
>>
>>	ext4_fname_from_fscrypt_name(fname, &name);
>>
>>  #if IS_ENABLED(CONFIG_UNICODE)
>>  	err = ext4_fname_setup_ci_filename(dir, &dentry->d_name, fname);
>> +	if (err)
>> +		fscrypt_free_filename(&name);
>>  #endif
>>  	return err;
>>  }
>
> This works, but it's a bit weird that the freeing happens on the original struct
> fscrypt_name after it has already been "moved" to the struct ext4_filename by
> ext4_fname_from_fscrypt_name().  That leaves a dangling pointer in the struct
> ext4_filename.  Maybe you should call ext4_fname_free_filename() instead, even
> though it would do some unnecessary work?

That makes sense, specially because fname is a parameter and it's probably
a good idea to clean-up everything before returning an error.  Thanks.

Cheers,
-- 
Luís