Re: AW: problem with testing a CTR block cipher mode which is partially working

[Stephan Mueller <smueller@chronox.de>]

Am Montag, 30. März 2015, 18:08:28 schrieb Markus Stockhausen:

Hi Markus,

> > Von: linux-crypto-owner@vger.kernel.org
> > [linux-crypto-owner@vger.kernel.org]" im Auftrag von "Corentin
> > LABBE [clabbe.montjoie@gmail.com] Gesendet: Montag, 30. März 2015 19:59
> > An: linux-crypto@vger.kernel.org
> > Cc: linux-sunxi@googlegroups.com
> > Betreff: problem with testing a CTR block cipher mode which is partially
> > working
> > 
> > hello
> > 
> > I am trying to add the CTR (counter) block cipher mode for AES on my
> > Security System driver.
> > 
> > When testing with the tcrypt module I got the following result:
> > [ 1256.986989] alg: skcipher: Test 1 failed on encryption for
> > ctr-aes-sunxi-ss [ 1256.987004] 00000000: 87 4d 61 91 b6 20 e3 26 1b ef
> > 68 64 99 0d b6 ce [ 1256.987013] 00000010: 40 94 25 91 d7 b4 4f 49 ab c1
> > 9d 33 a4 4e f6 54 [ 1256.987023] 00000020: ce 58 d2 f0 01 8f 92 a2 5f 2c
> > bb 66 13 8b 9d 76 [ 1256.987032] 00000030: 30 fa 4a 40 b1 67 2e f3 46 b7
> > 9a 7c ba 91 0b a2
> > 
> > As you can see the first ciphered block is correct (according to
> > testmgr.h), the subsequent blocks are bad.
> > 
> > So Could I assume that the setting of key and IV are good (at least for
> > the first cipher pass.
> > 
> > The number of inputs(register) are limited and I have tested near all the
> > possibility. Any idea of what could be wrong.
> 
> had a similar challenge a few months ago. I had to take care about
> 
> - counter IV is big endian (implemented it little endian in first place)

Use crypto_inc for the counter which properly increments in big endian.

> - CTR allows to encrypt data that does not need to be amultiple of 16 bytes.
> 
> Markus


-- 
Ciao
Stephan