Re: [Cfrg] I-D Action:draft-kiyomoto-kcipher2-02.txt

["Jim Schaad" <[e-mail]>]

Personal opinion.  It will be easier and clearer if you go back to using circular shift operations.  

Jim


> -----Original Message-----
> From: Wook Shin [mailto:[e-mail]]
> Sent: Wednesday, April 13, 2011 1:31 AM
> To: [e-mail]; Jim Schaad
> Subject: Re: I-D Action:draft-kiyomoto-kcipher2-02.txt
> 
> Jim,
> 
> That's my mistake. I think I made that mistake while converting circular shift
> operations to the combination of XOR and non-circular shit operations.
> 
> The notation in Section 2.1 should be corrected to
> 
>    <<n t     t-bit left shift in an n-bit register
>    >>n t     t-bit right shift in an n-bit register
> 
> and then, the code
> 
>    SubK2((IK[3] <<32 8) XOR (IK[3] >>32 24))
>      XOR (0x01, 0x00, 0x00, 0x00);
> 
> should mean that
> 
>   (1) IK[3] gets left-shifted by 8 bits
>   (2) IK[3] gets right-shifted by 24 bits
>   (3) call SubK2() by giving "(1) XOR (2) XOR 0x01000000"
>       as the input parameter
> 
> Best regards,
> 
> Wook
> 
> 
> 
> 
> On Wed, 13 Apr 2011 17:23:10 +0900, Jim Schaad <[e-mail]>
> wrote:
> 
> > You have a notation that is leaving me puzzled.
> >
> > In section 2.1 you say
> >
> >      <<n   Left circular rotation by n bits
> >
> >       >>n   Right circular rotation by n bits
> >
> > Which makes sense.
> >
> > However in section 2.3.2 you use the notation
> >
> >            SubK2((IK[3] <<32 8) XOR (IK[3] >>32 24)) XOR
> >             (0x01, 0x00, 0x00, 0x00);
> >
> > Which is not covered by the above.
> >
> > Additionally I wonder if you mean to use IK[3] in both locations here
> > as (I think)
> >
> > (IK[3] << 8) XOR (IK[3] >> 24) ---> (IK[3] << 8) XOR (IK[3] <<
> > (32-24))
> > --> (IK[3] << 8) XOR (IK[3] << 8)
> >
> > jim
> >
> >> -----Original Message-----
> >> From: Wook Shin [mailto:[e-mail]]
> >> Sent: Tuesday, April 12, 2011 6:39 PM
> >> To: [e-mail]
> >> Subject: Re: I-D Action:draft-kiyomoto-kcipher2-02.txt
> >>
> >> Hello all,
> >>
> >> I've revised our previous draft that explains the KCipher-2 stream
> >> cipher.
> >> The current version includes some code snippets written in the C
> >> language syntax, so that it could help implementers get clearer idea
> >> of how it works and help them check their outputs with the test
> >> vectors mentioned in the draft.
> >> The new version of the draft can be found at
> >>    http://www.ietf.org/internet-drafts/draft-kiyomoto-kcipher2-02.txt
> >>
> >> I have attached a naive implementation of KCipher-2 written in C.
> >> Two files, KCipher2.c and KCipher2.h, are in the attached .zip file,
> >> and no special option is required to compile and run the program.
> >> The code snippets and the test vector running examples in the draft
> >> are from the the attached code.
> >>
> >> Please look at the new version of the draft and the code, and let me
> >> have your comments.
> >>
> >> Thank you,
> >>
> >> Wook
> 
> 
> --
> Wook Shin
> Information Security Laboratory,
> KDDI R&D Laboratories, Inc.