Hello,
I'd like to show the interoperability between openssl 3.4.1 and Python
3.13/cryptography 44.0.1 library (which generally works). I signed with DSA a
short message using Python/Cryptography with:
message (in French): Je tente de signer un message un peu plus gros que le
precedent
def sign(message, sk):
chosen_hash = hashes.SHA256()
hasher = hashes.Hash(chosen_hash)
hasher.update(message)
digest = hasher.finalize()
signature = sk.sign(
digest, utils.Prehashed(chosen_hash)
)
return signature
with the DSA key
-----BEGIN PRIVATE KEY-----
MIIBSwIBADCCASsGByqGSM44BAEwggEeAoGBAJeeE6sU5RBpDfNHj9J9abda8TYk
BZNBtnWBfPvsqsuiMpqPWOrs5N9m5kKZuKSF639XlHVKeCCh+k1YfOw02qZuP7w8
Q8v87FrMW5b9o35DeqGDlHyjYCdYRnDjM7Gs8VfAl1+8PCb1AbOhscF/ojP5lrqj
n9m2XbAHe+afCRLXAhUA9krLBVX+ckL7mb+KrhD3JehLGHsCgYB3jMw4QDHsQDnR
VZoo+NeAe5LkJeZbPMOrxHQ6H9LX0kKuAo1lZarkql3ACWyVLo9ytb2KPDCw4r0I
taj1C742ty8FjlreC/E8vsArflV5joUyx7hlmV8Xuirpp2E0rC/feUXJ/zbnpGCc
vrA823h8UZ2ehQoGnj1tfBBihk7ivQQXAhUAmeKg5l0VXaVoBLsLH6c0vpC4uBs=
-----END PRIVATE KEY-----
The signature of the message is saved as a byte array in a file test.sig.
The verification with openssl fails with the message Verification failure after
the command
openssl dgst -verify dsa_public_key.pem -signature test.sig test.txt
I ASN.1 decoded the pica signature which prints
RSAPrivateKey SEQUENCE (2 items)
version Version INTEGER (160 bit)
modulus INTEGER (160 bit)
And the same decoding holds as well for a signature obtained with OpenSSL (with
different values) but the latter is successfully verified. Bellow are two DSA
signatures of the message in base64
pyca: MC4CFQDGylJyxlks53gkzb0wFv86MjlbowIVALAkqD4+JoqYQdb//f6YhSQZ/f4S
openssl : MC4CFQCq1x7axNVxFr7Y7fR/TZRBHHEMWQIVAKOkA3I6y1gEzBHBA/Ng99JIDeMs
I asked the question on crypto stackexchange with two comments: one replying
the openssl verification works on python3.10.6/pyca3.4.8 on Ubuntu22.04 the
other confirms that OpenSSL accepts the openssl signature but refuses the pyca
signature with Verification failure.
Thanks in advance for any answer/explanation of this failure and direct replies
since I am not a subscriber of the list.
Bruno Martin
_______________________________________________
Cryptography-dev mailing list
[email protected]
https://mail.python.org/mailman/listinfo/cryptography-dev
