C_DeriveKeys

[Michael StJohns <msj@nthpermutation.com>]

Attached is a draft submission for C_DeriveKeys - this augments/replaces 
C_DeriveKey

This may or may not be a 2.40 submission, but I'd like to chat about that.

This also includes some related changes to fix the issues with TLS key 
derivation (e.g. key data can leak into public data).

The reason for CKM_GENERIC_DERIVE is to somewhat simplify the interface 
for key derivation.  There are way too many structures for way too many 
ways of dealing with key derivation.  This basically takes all of those 
and provides a packed encoding.  I admit that the rationale for this is 
to make life simpler for things like java.

I'm also thinking that another variation on this might deal with 
compound KDFs (e.g. basic mechanism is X, but needs a PRF of Y - see 
NIST SP800-108 for example).

Mike

Attachment: pkcs11-derivekeys.docx
Description: application/vnd.openxmlformats-officedocument.wordprocessingml.document