smueller at chronox
Jan 11, 2012, 8:47 AM
Post #4 of 4
On 11.01.2012 16:55:59, +0100, Simon Crowe <smncrowe [at] gmail> wrote:
> Stephan, thanks for the information with respect to the perl script.
> With respect to the MAC size, i need the MAC size for HMAC-SHA384
The CAVS test vectors are really problematic in this area. The Perl
script uses the following, which is the default -- there are other size
combinations (IIRC, CAVS may test a truncated SHA512 hash that is
supposed to have a size of 48). So, sync that with your CAVS vectors!
# XXX this is a hack - we need to decipher the HMAC REQ files in
# sane way
# This is a conversion table from the expected hash output size
# to the assumed hash type - we only define here the block size of
# the underlying hashes and do not allow any truncation
my %hashtype = (
20 => 1,
28 => 224,
32 => 256,
48 => 384,
64 => 512
> Stephan Mueller-5 wrote:
>> On 10.01.2012 21:59:19, +0100, Simon Crowe <smncrowe [at] gmail> wrote:
>> Hi Simon,
>>> I am trying to run the CAVP tests against a set of test vectors,
>>> specifically SigGen and SigVer, however I am getting an error message
>>> stating that "Error: X9.31 is not supported", this is specifically being
>>> caught by the perl script which is in the test directory.
>> X9.31 RSA is not implemented in the Perl script, i.e. there is no parser
>> for the CAVS test vectors.
>>> Can someone please tell me why the SigGen and SigVer vectors are not
>>> supported by the CAVP test suite.
>> The X9.31 RSA testing is not supported by the Perl script.
>>> Furthermore, can someone tell me what the minimum HMAC size is for AES384
>>> looked at the code for AES384, and it appears to be based upon AES512 and
>>> therefore have the same HMAC size ..
>> AES384? You mean SHA384?
>>> Apologies if I have gone on the wrong track, but I am a newbie.
>> Gcrypt-devel mailing list
>> Gcrypt-devel [at] gnupg
Gcrypt-devel mailing list
Gcrypt-devel [at] gnupg