KEYS: asymmetric: Fix sign/verify on pkcs1pad without a hash

Commit Message

Herbert Xu Oct. 16, 2023, 8:35 a.m. UTC

On Thu, Oct 12, 2023 at 10:08:46AM -0500, Denis Kenzior wrote:
>
> Looks like something took out the ability to run sign/verify without a hash
> on asymmetric keys.

Indeed this is what it was.  Please try this patch.  Thanks!

---8<---
The new sign/verify code broke the case of pkcs1pad without a
hash algorithm.  Fix it by setting issig correctly for this case.

Fixes: 63ba4d67594a ("KEYS: asymmetric: Use new crypto interface without scatterlists")
Cc: [e-mail] # v6.5
Reported-by: Denis Kenzior <[e-mail]>
Signed-off-by: Herbert Xu <[e-mail]>

Comments

Denis Kenzior Oct. 16, 2023, 7:37 p.m. UTC | #1

Hi Herbert,

On 10/16/23 03:35, Herbert Xu wrote:
> On Thu, Oct 12, 2023 at 10:08:46AM -0500, Denis Kenzior wrote:
>>
>> Looks like something took out the ability to run sign/verify without a hash
>> on asymmetric keys.
> 
> Indeed this is what it was.  Please try this patch.  Thanks!
> 

I can confirm that this fix does make all unit tests pass again.  Feel free to add:

Tested-by: Denis Kenzior <[e-mail]>

Regards,
-Denis

Patch

diff --git a/crypto/asymmetric_keys/public_key.c b/crypto/asymmetric_keys/public_key.c
index abeecb8329b3..2f9181c4cd59 100644
--- a/crypto/asymmetric_keys/public_key.c
+++ b/crypto/asymmetric_keys/public_key.c
@@ -81,14 +81,13 @@  software_key_determine_akcipher(const struct public_key *pkey,
 		 * RSA signatures usually use EMSA-PKCS1-1_5 [RFC3447 sec 8.2].
 		 */
 		if (strcmp(encoding, "pkcs1") == 0) {
+			*sig = op == kernel_pkey_sign ||
+			       op == kernel_pkey_verify;
 			if (!hash_algo) {
-				*sig = false;
 				n = snprintf(alg_name, CRYPTO_MAX_ALG_NAME,
 					     "pkcs1pad(%s)",
 					     pkey->pkey_algo);
 			} else {
-				*sig = op == kernel_pkey_sign ||
-				       op == kernel_pkey_verify;
 				n = snprintf(alg_name, CRYPTO_MAX_ALG_NAME,
 					     "pkcs1pad(%s,%s)",
 					     pkey->pkey_algo, hash_algo);