Revisiting SHA-512/t HMAC_GENERAL

[Dina Kurktchi <dina.kurktchi@oracle.com>]

While following the example text from SHA1 and SHA512 in the
previous version specification, my proposal for SHA-512/t
variants included these 3 mechanisms:
        CKM_SHA512_224_HMAC_GENERAL
        CKM_SHA512_256_HMAC_GENERAL
        CKM_SHA512_T_HMAC_GENERAL

The question raised in our group now is whether the *_GENERAL
variants are needed at all, in the case of SHA-512/t.  If one
needs a shorter hash, selecting a smaller "t" should accomplish
the same thing.  Should these HMAC_GENERALs stay or be removed
(via errata perhaps)?


An ancillary question is about these existing mechanisms:
        CKM_SHA1_HMAC_GENERAL
        CKM_SHA224_HMAC_GENERAL
        CKM_SHA256_HMAC_GENERAL
        CKM_SHA384_HMAC_GENERAL
        CKM_SHA512_HMAC_GENERAL
Does anyone have background on why were these were originally
added to the specification?  Are they still necessary to keep?


It seems like I remember someone on-list brought up questions
in this area, but it wasn't clear if he was talking about
SHA-512/t itself, or the HMAC_GENERAL variants of all digest
mechanisms.


Any comments?


Thanks,
D.