Re: [PATCH] PKCS#7: fix a possible memory leak when calculating the digest

[Eric Biggers <ebiggers@kernel.org>]

On Thu, Feb 24, 2022 at 07:09:12PM +0000, Denis Glazkov wrote:
> In function `pkcs7_digest`, if there is an error allocating memory
> for the `shash_desc` structure, the public key signature digest
> remains unfreed.
> 
> Signed-off-by: Denis Glazkov <d.glazkov@omp.ru>
> ---
>  crypto/asymmetric_keys/pkcs7_verify.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/crypto/asymmetric_keys/pkcs7_verify.c b/crypto/asymmetric_keys/pkcs7_verify.c
> index 0b4d07aa8811..e6f648dcc02a 100644
> --- a/crypto/asymmetric_keys/pkcs7_verify.c
> +++ b/crypto/asymmetric_keys/pkcs7_verify.c
> @@ -50,7 +50,7 @@ static int pkcs7_digest(struct pkcs7_message *pkcs7,
>  	ret = -ENOMEM;
>  	sig->digest = kmalloc(sig->digest_size, GFP_KERNEL);
>  	if (!sig->digest)
> -		goto error_no_desc;
> +		goto error_no_digest;
>  
>  	desc = kzalloc(desc_size, GFP_KERNEL);
>  	if (!desc)
> @@ -117,6 +117,8 @@ static int pkcs7_digest(struct pkcs7_message *pkcs7,
>  error:
>  	kfree(desc);
>  error_no_desc:
> +	kfree(sig->digest);
> +error_no_digest:
>  	crypto_free_shash(tfm);
>  	kleave(" = %d", ret);
>  	return ret;
> -- 
> 2.25.1

Doesn't this introduce a double free?  public_key_signature_free() frees this
too.

- Eric