2007-12-03 13:32:04

by Jan Glauber

Subject: HMAC broken on s390

Hi Herbert,

commit 788fefa33b0b50581585925c53c230a36af35d0e in cryptodev breaks hmac
on s390 du to the usage of sg_chain():

static inline void sg_chain(struct scatterlist *prv, unsigned int prv_nents,
struct scatterlist *sgl)
{
#ifndef ARCH_HAS_SG_CHAIN
BUG();
#endif

ARCH_HAS_SG_CHAIN is false for s390 (and also for some other arch's).

What should we do with this?

Cheers,
Jan

2007-12-03 22:09:12

by Herbert Xu

Subject: Re: HMAC broken on s390

On Mon, Dec 03, 2007 at 02:31:40PM +0100, Jan Glauber wrote:
> Hi Herbert,
>
> commit 788fefa33b0b50581585925c53c230a36af35d0e in cryptodev breaks hmac
> on s390 du to the usage of sg_chain():
>
> static inline void sg_chain(struct scatterlist *prv, unsigned int prv_nents,
> struct scatterlist *sgl)
> {
> #ifndef ARCH_HAS_SG_CHAIN
> BUG();
> #endif
>
> ARCH_HAS_SG_CHAIN is false for s390 (and also for some other arch's).
>
> What should we do with this?

Looks like we took a step backwards because the chaining I had
before worked on all architectures :)

I suppose either we'll have to do our own chaining again or implement
it for all architectures. I'll look into this.

Thanks,
--
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

2007-12-04 08:27:19

by Jens Axboe

Subject: Re: HMAC broken on s390

On Tue, Dec 04 2007, Herbert Xu wrote:
> On Mon, Dec 03, 2007 at 02:31:40PM +0100, Jan Glauber wrote:
> > Hi Herbert,
> >
> > commit 788fefa33b0b50581585925c53c230a36af35d0e in cryptodev breaks hmac
> > on s390 du to the usage of sg_chain():
> >
> > static inline void sg_chain(struct scatterlist *prv, unsigned int prv_nents,
> > struct scatterlist *sgl)
> > {
> > #ifndef ARCH_HAS_SG_CHAIN
> > BUG();
> > #endif
> >
> > ARCH_HAS_SG_CHAIN is false for s390 (and also for some other arch's).
> >
> > What should we do with this?
>
> Looks like we took a step backwards because the chaining I had
> before worked on all architectures :)
>
> I suppose either we'll have to do our own chaining again or implement
> it for all architectures. I'll look into this.

You should continue to do your own chaining if it's a requirement for
the driver/layer to function, until we get rid of ARCH_HAS_SG_CHAIN.
That'll happen when all archs are converted.

So don't rely on scatterlist helping you with chaining just yet, sorry.

--
Jens Axboe

2007-12-04 08:51:38

by Herbert Xu

Subject: Re: HMAC broken on s390

On Tue, Dec 04, 2007 at 09:25:11AM +0100, Jens Axboe wrote:
>
> You should continue to do your own chaining if it's a requirement for
> the driver/layer to function, until we get rid of ARCH_HAS_SG_CHAIN.
> That'll happen when all archs are converted.

Thanks for the confirmation. I'll back out my conversion patch.

Cheers,
--
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt