RE: [RFC PATCH v3 1/1] virtio crypto device specification: asymmetric crypto service

["Zeng, Xin" <xin.zeng@intel.com>]

On Wednesday, January 18, 2017 10:04 AM, Gonglei (Arei) wrote:
> I update the virtio crypto spec to support non-session based crypto
> Operations and your asym service also needs change struct
> virtio_crypto_op_data_req which maybe cause the compatibility complaint
> with the pre-existing code.
> 
> So I strongly suggest you rebase my v16 (will be submitted today), and put
> asym service to struct virtio_crypto_op_data_req_mixed.
> 

Yes, sure.

> struct virtio_crypto_op_data_req_mixed {
>     struct virtio_crypto_op_header header;
> 
>     union {
>         struct virtio_crypto_sym_data_req   sym_req;
>         struct virtio_crypto_hash_data_req  hash_req;
>         struct virtio_crypto_mac_data_req   mac_req;
>         struct virtio_crypto_aead_data_req  aead_req;
>         struct virtio_crypto_sym_data_req_non_sess   sym_non_sess_req;
>         struct virtio_crypto_hash_data_req_non_sess  hash_non_sess_req;
>         struct virtio_crypto_mac_data_req_non_sess   mac_non_sess_req;
>         struct virtio_crypto_aead_data_req_non_sess  aead_non_sess_req;
> 
> 		/* asymmetric crypto requests here */
> 		...
>     } u;
> };
> 
> > We also plan to add asymmetric crypto service part support in LKCF
> > based virtio crypto frontend driver, and upstream DPDK vhost library
> > based vhost user for cryptodev which also focuses on asymmetric crypto
> service , any comments?
> >
> Sounds great. It's very helpful to accelerate the progress of reviewing.
> After all they are kinds of special knowledge.
> 
> Meanwhile, I can submit the cryptodev-vhost-user/cryptodev-vhost-kernel
> backends in QEMU if needs.
> 

That's great. We planned to upstream vhost-user for cryptodev to DPDK in Q3 this year. 
Let's have another collaboration.

> Thanks,
> -Gonglei
>