There is no such animal as an inverse key (if there was one, and it was easy to find, it wouldn't be a very good encryption algorithm). Decryption with AES requires using the same key and a different algorithm which runs the steps in the reverse order. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.opencores.org/forums.cgi/cores/attachments/20061011/030dc78c/attachment.htm

While using the AES128 core (aes_crypto_core) I have come across a
problem; after encrypting data with a key and receiving an output, how
do I know what the inverse key is in order to decrypt the output? I had
an example from the fips document about AES so I was able to test
given their key and output key, and was successful in doing so.
However, making up a key wont work for me as I have no idea what the
inverse key is. Clearing this issue up would really help me out. Thanks.

Hi Jonathan,
I am sorry for not being able to reply earlier, I have been travelling a
lot and hardly have checked my mails in between. Hmm.. that was a long
time back and dont remember exactly but if my memory is correct the
inverse key is the key that is output after an encryption is completed. If
you observe, both data and key undergo a series of transformation and at
the end there is both a garbled key and encrypted data coming out. This
key is taken as the inverse key for decryption. Infact I have used one set
of key pairs as default in my test bench. This information is also
explained in the FIPS document.

Hope this helps,
Hemanth

----------------------------------------- Stay ahead of the information curve. Receive EDA news and jobs on your desktop daily. Subscribe today to the EDA CafeNews newsletter. [ http://www10.edacafe.com/nl/newsletter_subscribe.php ] It's informative and essential. This message was sent to you from a machine at 128.88.255.123

Decryption is the reverse(inverse) of encryption. In AES you perform a series of operations, say you have 16 rounds. Then 16 times data is serially encrypted with a different key. It is only the key that is different between the rounds. Nothing else. In the end, however some more mixing is done. This process is repeated for each of the blocks of input data. The blocks also interact with one another depending on the ciphering mode, and the mixing is very good if you operate in CBC mode. Of couse the IV also comes into play.

Encryption is nothing but confusion and diffusion; substitution and transposition.

Decryption is nothing but doing the same thing in reverse. So in this case, you would take the key , feed it to encrypted data, apply the last stage mixing first and then do the series of rounds. You get back the data because encryption at a fundamental level is an XOR operation and you XOR twice to get back the data.

Sorry, I have oversimplified things and also have been wrong at a few places. But this may help clear some confusion for u. Who knows? :-)

Best,
Girish

Yes I understand Hemanth but the inverse key is not an output of the AES core, I dont know where to access it. Everything you said makes perfect sense so if I could figure out how to access the inverse key I'll be fine. Thanks! ----- Original Message ----- From: Hemanthhemanths at d...> To: Date: Wed Oct 18 13:37:36 CEST 2006 Subject: [oc] AES128 Question

Jonathan/Marcio,
I am sorry for the confusion but I have clearly stated in the tb_readme
doc that the mode for encryption or decryption has to be set in the TB
module aes_tester for generic inputs. I have for example used the default
key/inverse-key pair of:
000102030405060708090A0B0C0D0E0F
13111D7FE3944A17F307A78B4D2B30C5
in the above test bench. The enryption/decryption is selected by setting
the "mode_tb" variable "manually" in this TB module. I have NOT devised
the TB to to run encryption-decryption back to back. You also need to copy
the encrypted output data to the file "encoded_text.txt" to run the
decryption on the same ouput to get back the original message.
To be specific, I would suggest you to run the test with the "mode_tb"
signal set to "1"(it is currently set to "0" in the project) which will
enable encryption. After running the test, copy the ouput from the file
"coded_text.txt" to the input file for decryption "encoded_text.txt" and
run the test again with the "mode_tb" set to "0" which will enable
decryption and you can see your original message in file
"decoded_text.txt".
I am assuming that you faced the problems when running the generic input
tests run the TB "aes_tester".

Thanks,
Hemanth

----------------------------------------- Stay ahead of the information curve. Receive EDA news and jobs on your desktop daily. Subscribe today to the EDA CafeNews newsletter. [ http://www10.edacafe.com/nl/newsletter_subscribe.php ] It's informative and essential. This message was sent to you from a machine at 128.88.255.35

Hmm..,
I got where you are pointing at, I think I missed answering this in my
previous post. You are right the RTL does'nt ouput the inverse key. I
think my assumption was that the RTL need'nt take the burden of it. So
looks like you have to record the inverse key by looking at the waveform
or better still adding a piece of code to the TB to spy on the interior
signals of the key expander module at the end of its encryption/decryption
phase or even better you could have an excel spread sheet or some program
do that.

Hope this answers your doubts,
Hemanth

----------------------------------------- Stay ahead of the information curve. Receive EDA news and jobs on your desktop daily. Subscribe today to the EDA CafeNews newsletter. [ http://www10.edacafe.com/nl/newsletter_subscribe.php ] It's informative and essential. This message was sent to you from a machine at 128.88.255.35

Thanks Hemanth. What is the easiest way to "spy" on the key during the key's transformation? Much of the AES code is rather complicated for me. ----- Original Message ----- From: Hemanthhemanths at d...> To: Date: Mon Oct 23 08:15:29 CEST 2006 Subject: [oc] AES128 Question

Jonathan,
The signal spy option is provided by many simulators and in Modelsim it
goes like this:
signal_spy:process
begin
init_signal_spy("/aes_tester/aes_i/expand_key/new_key0",
"/aes_tester/key_op0", 0);
init_signal_spy("/aes_tester/aes_i/expand_key/new_key1",
"/aes_tester/key_op1", 0);
init_signal_spy("/aes_tester/aes_i/expand_key/new_key2",
"/aes_tester/key_op2", 0);
init_signal_spy("/aes_tester/aes_i/expand_key/new_key3",
"/aes_tester/key_op3", 0);
wait;
end process signal_spy;

In the above code new_key0,new_key1,new_key2,new_key3 are the signals
declared in the TB module aes_tester. The syntax being
init_signal_spy(src,dest,0). You should refer to Modelsim user guide for
further explanation. Now with the key values available on new_key* you can
use them depending on your requirement.

Best wishes,
Hemanth

Ah I see....Thanks a lot Hemanth, sorry for the late response. ----- Original Message ----- From: Hemanthhemanths at d...> To: Date: Fri Oct 27 10:31:22 CEST 2006 Subject: [oc] AES128 Question

Any luck on getting the decryption to work? Thanks, ----- Original Message ----- From: Hemanthhemanths at d...> To: Date: Fri Oct 27 10:31:22 CEST 2006 Subject: [oc] AES128 Question