On 7/22/2015 7:19 AM, Tadeusz Struk wrote:

On 07/21/2015 06:32 PM, Herbert Xu wrote:

quoted

I think we should finish the conversion of the only in-kernel
user of RSA before we add the user-space interface.  Otherwise
this unnecessarily ties our hands to the current API.

For example, do we want an SG interface for the input and output?
As it stands even if a hardware device came with SG support we
would not be able to use it.

Tadeusz, can you confirm again that qat does not support SG with
respect to RSA input and output?

No, it doesn't. Even if we would add SG support on the API it will
need to be copied into contiguous buffers before it is sent to the HW.

OTOH, caam has SG support for all PK operations, including rsa-encrypt,
rsa-decrypt primitives.
We are working at upstreaming - aligning our internal caam-pkc with
akcipher.

Please consider designing a flexible AF_ALG interface, as Stephan suggested.

Thanks,
Horia