cbc(aes-aesni) a givcipher?

cbc(aes-aesni) a givcipher?

[Stephan Mueller]

Hi Herbert,

with the latest cryptodev-2.6 tree I find the following:

name         : cbc(aes)
driver       : cbc(aes-aesni)
module       : kernel
priority     : 300
refcnt       : 1
selftest     : passed
internal     : no
type         : givcipher
async        : no
blocksize    : 16
min keysize  : 16
max keysize  : 32
ivsize       : 16
geniv        : eseqiv

name         : cbc(aes)
driver       : cbc(aes-aesni)
module       : cbc
priority     : 300
refcnt       : 1
selftest     : passed
internal     : no
type         : blkcipher
blocksize    : 16
min keysize  : 16
max keysize  : 32
ivsize       : 16
geniv        : <default>


Note the type of the ciphers. With this, an allocation of cbc(aes-aesni) does 
not work (ENOENT).

Any ideas?

-- 
Ciao
Stephan

Re: cbc(aes-aesni) a givcipher?

[Stephan Mueller]

Am Mittwoch, 19. August 2015, 12:11:03 schrieb Stephan Mueller:

Hi,

> Hi Herbert,
> 
> with the latest cryptodev-2.6 tree I find the following:
> 
> name         : cbc(aes)
> driver       : cbc(aes-aesni)
> module       : kernel
> priority     : 300
> refcnt       : 1
> selftest     : passed
> internal     : no
> type         : givcipher
> async        : no
> blocksize    : 16
> min keysize  : 16
> max keysize  : 32
> ivsize       : 16
> geniv        : eseqiv
> 
> name         : cbc(aes)
> driver       : cbc(aes-aesni)
> module       : cbc
> priority     : 300
> refcnt       : 1
> selftest     : passed
> internal     : no
> type         : blkcipher
> blocksize    : 16
> min keysize  : 16
> max keysize  : 32
> ivsize       : 16
> geniv        : <default>
> 
> 
> Note the type of the ciphers. With this, an allocation of cbc(aes-aesni)
> does not work (ENOENT).
 
Just a followup: I think I see how I triggered it; I used cbc(aes-aesni) with 
crypto_alloc_aead due to an accident.

But that makes me think: would that be an issue for the AF_ALG interface? I.e. 
I try to use the algif_aead interface but provide cbc(aes-aesni) as a cipher 
name?
> 
> Any ideas?


-- 
Ciao
Stephan

Re: cbc(aes-aesni) a givcipher?

[Herbert Xu]

On Wed, Aug 19, 2015 at 12:31:13PM +0200, Stephan Mueller wrote:
> > 
> > with the latest cryptodev-2.6 tree I find the following:
> > 
> > name         : cbc(aes)
> > driver       : cbc(aes-aesni)
> > module       : kernel
> > priority     : 300
> > refcnt       : 1
> > selftest     : passed
> > internal     : no
> > type         : givcipher
> > async        : no
> > blocksize    : 16
> > min keysize  : 16
> > max keysize  : 32
> > ivsize       : 16
> > geniv        : eseqiv
> > 
> > name         : cbc(aes)
> > driver       : cbc(aes-aesni)
> > module       : cbc
> > priority     : 300
> > refcnt       : 1
> > selftest     : passed
> > internal     : no
> > type         : blkcipher
> > blocksize    : 16
> > min keysize  : 16
> > max keysize  : 32
> > ivsize       : 16
> > geniv        : <default>

This is how it's supposed to look like as ablkcipher also has
implicit IV generators like AEAD used to have.  I'm working on
a patch set to get rid of them just like what we did to AEAD.

> > Note the type of the ciphers. With this, an allocation of cbc(aes-aesni)
> > does not work (ENOENT).
>  
> Just a followup: I think I see how I triggered it; I used cbc(aes-aesni) with 
> crypto_alloc_aead due to an accident.

So are you saying that crypto_alloc_blkcipher is failing? If it's
only crypto_alloc_aead failing then I don't see the problem.

Cheers,
-- 
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt