[PATCH] KEYS: asymmetric: Fix ECDSA use via keyctl uapi

[PATCH] KEYS: asymmetric: Fix ECDSA use via keyctl uapi

[Denis Kenzior]

When support for ECDSA keys was added, constraints for data & signature
sizes were never updated.  This makes it impossible to use such keys via
keyctl API from userspace; fix that.

Fixes: 299f561a6693 ("x509: Add support for parsing x509 certs with ECDSA keys")
Signed-off-by: Denis Kenzior <denkenz@gmail.com>
---
 crypto/asymmetric_keys/public_key.c | 24 ++++++++++++++++++++++--
 1 file changed, 22 insertions(+), 2 deletions(-)

diff --git a/crypto/asymmetric_keys/public_key.c b/crypto/asymmetric_keys/public_key.c
index 2f8352e88860..eca5671ad3f2 100644
--- a/crypto/asymmetric_keys/public_key.c
+++ b/crypto/asymmetric_keys/public_key.c
@@ -186,8 +186,28 @@ static int software_key_query(const struct kernel_pkey_params *params,
 
 	len = crypto_akcipher_maxsize(tfm);
 	info->key_size = len * 8;
-	info->max_data_size = len;
-	info->max_sig_size = len;
+
+	if (strncmp(pkey->pkey_algo, "ecdsa", 5) == 0) {
+		/*
+		 * ECDSA key sizes are much smaller than RSA, and thus could
+		 * operate on (hashed) inputs that are larger than key size.
+		 * For example SHA384-hashed input used with secp256r1
+		 * based keys.  Set max_data_size to be at least as large as
+		 * the largest supported hash size (SHA512)
+		 */
+		info->max_data_size = 64;
+
+		/*
+		 * Verify takes ECDSA-Sig (described in RFC 5480) as input,
+		 * which is actually 2 'key_size'-bit integers encoded in
+		 * ASN.1.  Account for the ASN.1 encoding overhead here.
+		 */
+		info->max_sig_size = 2 * (len + 3) + 2;
+	} else {
+		info->max_data_size = len;
+		info->max_sig_size = len;
+	}
+
 	info->max_enc_size = len;
 	info->max_dec_size = len;
 	info->supported_ops = (KEYCTL_SUPPORTS_ENCRYPT |
-- 
2.35.1

Re: [PATCH] KEYS: asymmetric: Fix ECDSA use via keyctl uapi

[Stefan Berger]

On 7/15/22 14:28, Denis Kenzior wrote:
> When support for ECDSA keys was added, constraints for data & signature
> sizes were never updated.  This makes it impossible to use such keys via
> keyctl API from userspace; fix that.
> 
> Fixes: 299f561a6693 ("x509: Add support for parsing x509 certs with ECDSA keys")
> Signed-off-by: Denis Kenzior <denkenz@gmail.com>
> ---
>   crypto/asymmetric_keys/public_key.c | 24 ++++++++++++++++++++++--
>   1 file changed, 22 insertions(+), 2 deletions(-)
> 
> diff --git a/crypto/asymmetric_keys/public_key.c b/crypto/asymmetric_keys/public_key.c
> index 2f8352e88860..eca5671ad3f2 100644
> --- a/crypto/asymmetric_keys/public_key.c
> +++ b/crypto/asymmetric_keys/public_key.c
> @@ -186,8 +186,28 @@ static int software_key_query(const struct kernel_pkey_params *params,
>   
>   	len = crypto_akcipher_maxsize(tfm);
>   	info->key_size = len * 8;
> -	info->max_data_size = len;
> -	info->max_sig_size = len;
> +
> +	if (strncmp(pkey->pkey_algo, "ecdsa", 5) == 0) {
> +		/*
> +		 * ECDSA key sizes are much smaller than RSA, and thus could
> +		 * operate on (hashed) inputs that are larger than key size.
> +		 * For example SHA384-hashed input used with secp256r1
> +		 * based keys.  Set max_data_size to be at least as large as
> +		 * the largest supported hash size (SHA512)
> +		 */
> +		info->max_data_size = 64;
> +
> +		/*
> +		 * Verify takes ECDSA-Sig (described in RFC 5480) as input,
> +		 * which is actually 2 'key_size'-bit integers encoded in
> +		 * ASN.1.  Account for the ASN.1 encoding overhead here.
> +		 */
> +		info->max_sig_size = 2 * (len + 3) + 2;


This part matches what I wrote about here: 
https://lore.kernel.org/linux-integrity/127811ae-ef71-d225-0450-710a6ee6afef@linux.ibm.com/T/#m7aea8ca697e7c35773c3f8f110329553f4f8d91e

Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>

> +	} else {
> +		info->max_data_size = len;
> +		info->max_sig_size = len;
> +	}
> +
>   	info->max_enc_size = len;
>   	info->max_dec_size = len;
>   	info->supported_ops = (KEYCTL_SUPPORTS_ENCRYPT |