[PATCH] lib: ecdsa: fix prevent memory leak in ecdsa_add_verify_data

[Quentin Schulz]

Hi Anton,

On 2/25/25 2:49 PM, Anton Moryakov wrote:
> - Ensure `free_ctx` is called in both error and success paths.
> - Fix memory leak in `ctx.signature` when `do_add` fails."
> 
> Triggers found by static analyzer Svace.
> 
> Signed-off-by: Anton Moryakov <ant.v.moryakov at gmail.com>
> ---
>   lib/ecdsa/ecdsa-libcrypto.c | 6 ++++--
>   1 file changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/lib/ecdsa/ecdsa-libcrypto.c b/lib/ecdsa/ecdsa-libcrypto.c
> index 1c5dde6069..f0095e9dbc 100644
> --- a/lib/ecdsa/ecdsa-libcrypto.c
> +++ b/lib/ecdsa/ecdsa-libcrypto.c
> @@ -363,8 +363,10 @@ int ecdsa_add_verify_data(struct image_sign_info *info, void *fdt)
>   	ret = prepare_ctx(&ctx, info);
>   	if (ret >= 0) {
>   		ret = do_add(&ctx, fdt, fdt_key_name, info);
> -		if (ret < 0)
> -			ret = ret == -FDT_ERR_NOSPACE ? -ENOSPC : -EIO;
> +		if (ret < 0) {
> +			free_ctx(&ctx);
> +			return ret == -FDT_ERR_NOSPACE ? -ENOSPC : -EIO;
> +		}

If I read the code properly, this is changing nothing in terms of 
behavior, I believe this is a false positive from the static analyzer.

We don't return in the if block, so free_ctx() will be called.

ret will be set appropriately and the return value as well, we don't 
need to return earlier.

What am I missing here?

Cheers,
Quentin