Hi Matt,
thanks for generating and sending out these test cases! GMAC test
vectors has been on my to-do list for a while - it was a conspicuous
omission from the specification. Anyway, I fired up my GCM
implementation and confirmed your results. I'd also like to generate
some GMAC test cases for other AAD sizes (including ones with no byte
alignment) and larger AES keys. Would you mind verifying those as well?
Best regards,
David
On Dec 16, 2005, at 12:00 PM, Matt Ball wrote:
Thanks for the pointer to the source code. I went ahead and
compiled it and generated two new test vectors for a 128-bit and
256-bit AAD (with no plaintext). Here they are for reference:
KEY feffe9928665731c6d6a8f9467308308
IV cafebabefacedbaddecaf888
AAD feedfacedeadbeeffeedfacedeadbeef
TAG 54df474f4e71a9ef8a09bf30da7b1a92
KEY feffe9928665731c6d6a8f9467308308
IV cafebabefacedbaddecaf888
AAD feedfacedeadbeeffeedfacedeadbeefabaddad242831ec2217774244b7221b7
TAG 1cbe3936e553b08f25c08d7b8dc39fdb
-Matt
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of
Colin
Sinclair
Sent: Friday, December 16, 2005 9:59 AM
To: [EMAIL PROTECTED]
Subject: RE: P1619.1 (Tape): GMAC Test vectors
Well I never, that's the first time I've come across a URL which is
case sensitive. Try capitalising
the AES then it works. I was browsing on a different machine to my
email; I just typed it in without
thinking.
http://fp.gladman.plus.com/AES/index.htm
-----Original Message-----
From: Matt Ball [mailto:[EMAIL PROTECTED]
Sent: 16 December 2005 16:39
To: [EMAIL PROTECTED]; [EMAIL PROTECTED]
Subject: RE: P1619.1 (Tape): GMAC Test vectors
Hi Colin,
Would you mind double-checking the link below? I tried it out,
but got a 'Not Found'
error on my web browser. It looks like Brian has deleted the /
aes/ directory.
Thanks,
-Matt
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of
Colin
Sinclair
Sent: Friday, December 16, 2005 9:16 AM
To: [EMAIL PROTECTED]
Subject: RE: P1619.1 (Tape): GMAC Test vectors
There's been a lot of traffic on this list recently
I'm guilty as charged! Apologies to those not interested in finer
points of LRW!
I had a quick question about the GMAC proposal. Could we get
some GMAC-only test vectors
added to the GCM spec? In particular, I'm interested in a test
vector that has a
non-zero AAD length and a zero plaintext length. Preferably, the
AAD would be at least
256-bits long. David McGrew, would you be able to provide
something like this? This
would be really helpful for testing GMAC.
The best freely available, and fully-featured, GCM software
implementation is from
Brian Gladman, http://fp.gladman.plus.com/aes/index.htm,
scroll down and look for link to zip file called modes.zip.
Just write a simple C program around this to print out some test
vectors.
Colin.
