Re: [dpdk-dev] [PATCH] cryptodev: formalize key wrap method in API

[Matan Azrad <matan@nvidia.com>]

From: Akhil Goyal
> > Hi Akhil
> >
> > From: Akhil Goyal
> > > Hi Matan,
> > >
> > > > The Key Wrap approach is used by applications in order to protect
> > > > keys located in untrusted storage or transmitted over untrusted
> > > > communications networks. The constructions are typically built
> > > > from standard primitives such as block ciphers and cryptographic
> > > > hash functions.
> > > >
> > > > The Key Wrap method and its parameters are a secret between the
> > > > keys provider and the device, means that the device is
> > > > preconfigured for this method using very secured way.
> > > >
> > > > The key wrap method may change the key length and layout.
> > > >
> > > > Add a description for the cipher transformation key to allow
> > > > wrapped key to be forwarded by the same API.
> > > >
> > > > Signed-off-by: Matan Azrad <matan@nvidia.com>
> > > > ---
> > >
> > > How will the driver gets notified whether the key is wrapped or not?
> > > The driver would expect the keys are as per the capabilities exposed.
> > > If it does not check as per the capabilities, how will it identify a
> > > bad Key len
> > and
> > > the wrapped key length?
> >
> > As I wrote, the device is preconfigured to the key wrapping mode in
> > very secured way.
> > No one knows this secret except the device and the key provider
> > (outside the crypto lib and dpdk application).
> > The driver may know something but not necessarily.
> > If the device validation to the key is failed, the driver will return
> > an error to the app.
> >
> Shouldn't this be a capability/feature flag whether it can be supported by the
> underlying Driver or not?

Makes sense to me.

> I am not saying that the driver should know the details of the wrapping mode.
> But it can be asked, whether it support this wrapping mode or not?

It can say that it support key wrapping - no more than that.

What do you think?


> 
> >
> > > >  lib/librte_cryptodev/rte_crypto_sym.h | 7 +++++++
> > > >  1 file changed, 7 insertions(+)
> > > >
> > > > diff --git a/lib/librte_cryptodev/rte_crypto_sym.h
> > > > b/lib/librte_cryptodev/rte_crypto_sym.h
> > > > index 5973e31..6aca2c7 100644
> > > > --- a/lib/librte_cryptodev/rte_crypto_sym.h
> > > > +++ b/lib/librte_cryptodev/rte_crypto_sym.h
> > > > @@ -200,6 +200,13 @@ struct rte_crypto_cipher_xform {
> > > >               uint16_t length;        /**< key length in bytes */
> > > >       } key;
> > > >       /**< Cipher key
> > > > +      * The original key data may be provided wrapped (encrypted)
> > > > + using
> > > > a key
> > > > +      * wrap algorithm such as AES key wrap (from rfc3394) or other.
> > > > + In
> > > > such
> > > > +      * case, the wrapping details is a secret between the key provider and
> > > > +      * the device. Such key wrapping may increase the length of
> > > > + the
> > > > provided
> > > > +      * key beyond the advertised supported key size. Hence it is the
> > > > +      * responsibility of the driver/device to validate the length of the
> > > > +      * provided key.
> > > >        *
> > > >        * For the RTE_CRYPTO_CIPHER_AES_F8 mode of operation,
> > > > key.data will
> > > >        * point to a concatenation of the AES encryption key
> > > > followed by a
> > > > --
> > > > 1.8.3.1