Re: [PATCH] ext4: fix memory leak in ext4_fname_setup_filename() error path

From: Luís Henriques
Date: Wed Aug 02 2023 - 04:13:36 EST


Eric Biggers <ebiggers@xxxxxxxxxx> writes:

> On Tue, Aug 01, 2023 at 03:41:36PM +0100, Luís Henriques wrote:
>> If casefolding the filename fails, we'll be leaking fscrypt_buf name.
>> Make sure we free it in the error path.
>>
>> Signed-off-by: Luís Henriques <lhenriques@xxxxxxx>
>> ---
>> fs/ext4/crypto.c | 2 ++
>> 1 file changed, 2 insertions(+)
>>
>> diff --git a/fs/ext4/crypto.c b/fs/ext4/crypto.c
>> index e20ac0654b3f..9e4503b051c4 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;
>> }
>
> Doesn't ext4_fname_prepare_lookup() have the same bug?

Doh! I looked into that function too and then completely forgot about it.
It may also call fscrypt_setup_filename(), so it has to do the free in the
error path as well. I'll send out v2 to include that fix too.

> Also, please include a Fixes tag.

Ok, makes sense. Thanks!

Cheers,
--
Luís