[vector-crypto] Clarifying relationship between Zvknhb and Zvknha

[nibrunieAtSi5]

The current vector crypto specification does not mention any specific link between the two extensions. Their respective description implies that Zvknhb is a subset of Zvknha. This patch attempts at clarifying this relationship as simply as possible.

I have not modified any of the definition (Zvknhb still defines support for both SEW=32 and 64, and Zvknha defines support for SEW=32, other values being reserved encodings).

[aswaterman]

Do normative tags for extension dependence rules really make sense? How could you write a test that tests this property?

[nibrunieAtSi5]

Do normative tags for extension dependence rules really make sense? How could you write a test that tests this property?

I will have to consult @james-ball-qualcomm and the authors of #2518, as I simply reproduced what was done for other extensions in vector-crypto:

riscv-isa-manual/normative_rule_defs/vector-crypto.yaml

Lines 40 to 48 in b9f1438

	- names: [Zvknhb_depZve64x, Zvbc_depZve64x, Zvkn_depZve64x, Zvknc_depZve64x, Zvkng_depZve64x]
	tags: ["norm:Zvknhb_Zvbc_Zvkn_Zvknc_Zvkng_depZve64x"]
	- names: [Zvbb_depZve32x, Zvkb_depZve32x, Zvkg_depZve32x, Zvkned_depZve32x, Zvknha_depZve32x, Zvksed_depZve32x, Zvksh_depZve32x, Zvks_depZve32x, Zvksc_depZve32x,
	Zvksg_depZve32x, Zvkt_depZve32x]
	tags: ["norm:Zvbb_Zvkb_Zvkg_Zvkned_Zvknha_Zvksed_Zvksh_Zvks_Zvksc_Zvksg_Zvkt_depZve32x"]
	- name: veccrypto_indepZkt
	tags: ["norm:veccrypto_indepZkt"]

[radimkrcmar]

I think this would be better as a non-normative clarification, because Zvknha is already defined as subset of Zvknhb.
(The subset of Zvknhb is behaviorally indistinguishable from Zvknha.)

The change wouldn't have any effect on the ISA string requirements either... Normative rules should have a meaning.