Update for Zvkb extension.

https://github.com/riscv/riscv-crypto/blob/c8ddeb7e64a3444dda0438316af1238aeed72041/doc/vector/riscv-crypto-vector-zvkb.adoc
Create `RISCV_HAS_ZVKB()` macro.
Use zvkb for SM4 instead of zvbb.
Use zvkb for ghash instead of zvbb.
We could just use the zvbb's subset `zvkb` for flexibility.

Signed-off-by: Jerry Shih <jerry.shih@sifive.com>
Signed-off-by: Phoebe Chen <phoebe.chen@sifive.com>

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from #21923)

crypto/modes/asm/ghash-riscv64-zvbb-zvbc.pl → crypto/modes/asm/ghash-riscv64-zvkb-zvbc.pl

@@ -35,9 +35,9 @@
 # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
 # - RV64I
-# - RISC-V vector ('V') with VLEN >= 128
-# - Vector Bit-manipulation used in Cryptography ('Zvbb')
-# - Vector Carryless Multiplication ('Zvbc')
+# - RISC-V Vector ('V') with VLEN >= 128
+# - RISC-V Vector Cryptography Bit-manipulation extension ('Zvkb')
+# - RISC-V Vector Carryless Multiplication extension ('Zvbc')
 
 use strict;
 use warnings;
@@ -59,20 +59,20 @@
 ___
 
 ################################################################################
-# void gcm_init_rv64i_zvbb_zvbc(u128 Htable[16], const u64 H[2]);
+# void gcm_init_rv64i_zvkb_zvbc(u128 Htable[16], const u64 H[2]);
 #
 # input:	H: 128-bit H - secret parameter E(K, 0^128)
-# output:	Htable: Preprocessed key data for gcm_gmult_rv64i_zvbb_zvbc and
-#                       gcm_ghash_rv64i_zvbb_zvbc
+# output:	Htable: Preprocessed key data for gcm_gmult_rv64i_zvkb_zvbc and
+#                       gcm_ghash_rv64i_zvkb_zvbc
 {
 my ($Htable,$H,$TMP0,$TMP1,$TMP2) = ("a0","a1","t0","t1","t2");
 my ($V0,$V1,$V2,$V3,$V4,$V5,$V6) = ("v0","v1","v2","v3","v4","v5","v6");
 
 $code .= <<___;
 .p2align 3
-.globl gcm_init_rv64i_zvbb_zvbc
-.type gcm_init_rv64i_zvbb_zvbc,\@function
-gcm_init_rv64i_zvbb_zvbc:
+.globl gcm_init_rv64i_zvkb_zvbc
+.type gcm_init_rv64i_zvkb_zvbc,\@function
+gcm_init_rv64i_zvkb_zvbc:
     # Load/store data in reverse order.
     # This is needed as a part of endianness swap.
     add $H, $H, 8
@@ -110,12 +110,12 @@
 
     @{[vse64_v $V1, $Htable]}        # vse64.v v1, (a0)
     ret
-.size gcm_init_rv64i_zvbb_zvbc,.-gcm_init_rv64i_zvbb_zvbc
+.size gcm_init_rv64i_zvkb_zvbc,.-gcm_init_rv64i_zvkb_zvbc
 ___
 }
 
 ################################################################################
-# void gcm_gmult_rv64i_zvbb_zvbc(u64 Xi[2], const u128 Htable[16]);
+# void gcm_gmult_rv64i_zvkb_zvbc(u64 Xi[2], const u128 Htable[16]);
 #
 # input:	Xi: current hash value
 #		Htable: preprocessed H
@@ -127,9 +127,9 @@
 $code .= <<___;
 .text
 .p2align 3
-.globl gcm_gmult_rv64i_zvbb_zvbc
-.type gcm_gmult_rv64i_zvbb_zvbc,\@function
-gcm_gmult_rv64i_zvbb_zvbc:
+.globl gcm_gmult_rv64i_zvkb_zvbc
+.type gcm_gmult_rv64i_zvkb_zvbc,\@function
+gcm_gmult_rv64i_zvkb_zvbc:
     ld $TMP0, ($Htable)
     ld $TMP1, 8($Htable)
     li $TMP2, 63
@@ -228,12 +228,12 @@
     @{[vrev8_v $V2, $V2]}            # vrev8.v v2, v2
     @{[vsse64_v $V2, $Xi, $TMP4]}    # vsse64.v v2, (a0), t4
     ret
-.size gcm_gmult_rv64i_zvbb_zvbc,.-gcm_gmult_rv64i_zvbb_zvbc
+.size gcm_gmult_rv64i_zvkb_zvbc,.-gcm_gmult_rv64i_zvkb_zvbc
 ___
 }
 
 ################################################################################
-# void gcm_ghash_rv64i_zvbb_zvbc(u64 Xi[2], const u128 Htable[16],
+# void gcm_ghash_rv64i_zvkb_zvbc(u64 Xi[2], const u128 Htable[16],
 #                                const u8 *inp, size_t len);
 #
 # input:	Xi: current hash value
@@ -247,9 +247,9 @@
 
 $code .= <<___;
 .p2align 3
-.globl gcm_ghash_rv64i_zvbb_zvbc
-.type gcm_ghash_rv64i_zvbb_zvbc,\@function
-gcm_ghash_rv64i_zvbb_zvbc:
+.globl gcm_ghash_rv64i_zvkb_zvbc
+.type gcm_ghash_rv64i_zvkb_zvbc,\@function
+gcm_ghash_rv64i_zvkb_zvbc:
     ld $TMP0, ($Htable)
     ld $TMP1, 8($Htable)
     li $TMP2, 63
@@ -361,7 +361,7 @@
 
     @{[vsse64_v $V5, $Xi, $M8]}    # vsse64.v v2, (a0), t4
     ret
-.size gcm_ghash_rv64i_zvbb_zvbc,.-gcm_ghash_rv64i_zvbb_zvbc
+.size gcm_ghash_rv64i_zvkb_zvbc,.-gcm_ghash_rv64i_zvkb_zvbc
 ___
 }
 

crypto/modes/asm/ghash-riscv64-zvkg.pl

@@ -35,8 +35,11 @@
 # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
 # - RV64I
-# - RISC-V vector ('V') with VLEN >= 128
-# - RISC-V vector crypto GHASH extension ('Zvkg')
+# - RISC-V Vector ('V') with VLEN >= 128
+# - RISC-V Vector GCM/GMAC extension ('Zvkg')
+#
+# Optional:
+# - RISC-V Vector Cryptography Bit-manipulation extension ('Zvkb')
 
 use strict;
 use warnings;
@@ -59,7 +62,7 @@
 
 ################################################################################
 # void gcm_init_rv64i_zvkg(u128 Htable[16], const u64 H[2]);
-# void gcm_init_rv64i_zvkg_zvbb(u128 Htable[16], const u64 H[2]);
+# void gcm_init_rv64i_zvkg_zvkb(u128 Htable[16], const u64 H[2]);
 #
 # input: H: 128-bit H - secret parameter E(K, 0^128)
 # output: Htable: Copy of secret parameter (in normalized byte order)
@@ -88,15 +91,15 @@
 
 $code .= <<___;
 .p2align 3
-.globl gcm_init_rv64i_zvkg_zvbb
-.type gcm_init_rv64i_zvkg_zvbb,\@function
-gcm_init_rv64i_zvkg_zvbb:
-    @{[vsetivli__x0_2_e64_m1_tu_mu]} # vsetivli x0, 2, e64, m1, tu, mu
+.globl gcm_init_rv64i_zvkg_zvkb
+.type gcm_init_rv64i_zvkg_zvkb,\@function
+gcm_init_rv64i_zvkg_zvkb:
+    @{[vsetivli__x0_2_e64_m1_tu_mu]} # vsetivli x0, 2, e64, m1, ta, ma
     @{[vle64_v $V0, $H]}             # vle64.v v0, (a1)
     @{[vrev8_v $V0, $V0]}            # vrev8.v v0, v0
     @{[vse64_v $V0, $Htable]}        # vse64.v v0, (a0)
     ret
-.size gcm_init_rv64i_zvkg_zvbb,.-gcm_init_rv64i_zvkg_zvbb
+.size gcm_init_rv64i_zvkg_zvkb,.-gcm_init_rv64i_zvkg_zvkb
 ___
 }
 

crypto/modes/build.info

@@ -43,7 +43,7 @@ IF[{- !$disabled{asm} -}]
   $MODESASM_c64xplus=ghash-c64xplus.s
   $MODESDEF_c64xplus=GHASH_ASM
 
-  $MODESASM_riscv64=ghash-riscv64.s ghash-riscv64-zvbb-zvbc.s ghash-riscv64-zvkg.s
+  $MODESASM_riscv64=ghash-riscv64.s ghash-riscv64-zvkb-zvbc.s ghash-riscv64-zvkg.s
   $MODESDEF_riscv64=GHASH_ASM
 
   # Now that we have defined all the arch specific variables, use the
@@ -91,5 +91,5 @@ GENERATE[ghash-s390x.S]=asm/ghash-s390x.pl
 INCLUDE[ghash-s390x.o]=..
 GENERATE[ghash-c64xplus.S]=asm/ghash-c64xplus.pl
 GENERATE[ghash-riscv64.s]=asm/ghash-riscv64.pl
-GENERATE[ghash-riscv64-zvbb-zvbc.s]=asm/ghash-riscv64-zvbb-zvbc.pl
+GENERATE[ghash-riscv64-zvkb-zvbc.s]=asm/ghash-riscv64-zvkb-zvbc.pl
 GENERATE[ghash-riscv64-zvkg.s]=asm/ghash-riscv64-zvkg.pl

crypto/modes/gcm128.c

@@ -413,14 +413,14 @@ void gcm_ghash_rv64i_zbc(u64 Xi[2], const u128 Htable[16],
                          const u8 *inp, size_t len);
 void gcm_ghash_rv64i_zbc__zbkb(u64 Xi[2], const u128 Htable[16],
                                const u8 *inp, size_t len);
-/* Zvbb/Zvbc (vector crypto with vclmul) based routines. */
-void gcm_init_rv64i_zvbb_zvbc(u128 Htable[16], const u64 Xi[2]);
-void gcm_gmult_rv64i_zvbb_zvbc(u64 Xi[2], const u128 Htable[16]);
-void gcm_ghash_rv64i_zvbb_zvbc(u64 Xi[2], const u128 Htable[16],
+/* zvkb/Zvbc (vector crypto with vclmul) based routines. */
+void gcm_init_rv64i_zvkb_zvbc(u128 Htable[16], const u64 Xi[2]);
+void gcm_gmult_rv64i_zvkb_zvbc(u64 Xi[2], const u128 Htable[16]);
+void gcm_ghash_rv64i_zvkb_zvbc(u64 Xi[2], const u128 Htable[16],
                                const u8 *inp, size_t len);
 /* Zvkg (vector crypto with vgmul.vv and vghsh.vv). */
 void gcm_init_rv64i_zvkg(u128 Htable[16], const u64 Xi[2]);
-void gcm_init_rv64i_zvkg_zvbb(u128 Htable[16], const u64 Xi[2]);
+void gcm_init_rv64i_zvkg_zvkb(u128 Htable[16], const u64 Xi[2]);
 void gcm_gmult_rv64i_zvkg(u64 Xi[2], const u128 Htable[16]);
 void gcm_ghash_rv64i_zvkg(u64 Xi[2], const u128 Htable[16],
                           const u8 *inp, size_t len);
@@ -524,16 +524,16 @@ static void gcm_get_funcs(struct gcm_funcs_st *ctx)
     ctx->ghash = gcm_ghash_4bit;
 
     if (RISCV_HAS_ZVKG() && riscv_vlen() >= 128) {
-        if (RISCV_HAS_ZVBB())
-            ctx->ginit = gcm_init_rv64i_zvkg_zvbb;
+        if (RISCV_HAS_ZVKB())
+            ctx->ginit = gcm_init_rv64i_zvkg_zvkb;
         else
             ctx->ginit = gcm_init_rv64i_zvkg;
         ctx->gmult = gcm_gmult_rv64i_zvkg;
         ctx->ghash = gcm_ghash_rv64i_zvkg;
-    } else if (RISCV_HAS_ZVBB() && RISCV_HAS_ZVBC() && riscv_vlen() >= 128) {
-        ctx->ginit = gcm_init_rv64i_zvbb_zvbc;
-        ctx->gmult = gcm_gmult_rv64i_zvbb_zvbc;
-        ctx->ghash = gcm_ghash_rv64i_zvbb_zvbc;
+    } else if (RISCV_HAS_ZVKB() && RISCV_HAS_ZVBC() && riscv_vlen() >= 128) {
+        ctx->ginit = gcm_init_rv64i_zvkb_zvbc;
+        ctx->gmult = gcm_gmult_rv64i_zvkb_zvbc;
+        ctx->ghash = gcm_ghash_rv64i_zvkb_zvbc;
     } else if (RISCV_HAS_ZBC()) {
         if (RISCV_HAS_ZBKB()) {
             ctx->ginit = gcm_init_rv64i_zbc__zbkb;

crypto/perlasm/riscv.pm

@@ -746,7 +746,18 @@ sub vxor_vv {
 
 # Vector crypto instructions
 
-## Zvbb instructions
+## Zvbb and Zvkb instructions
+##
+## vandn (also in zvkb)
+## vbrev
+## vbrev8 (also in zvkb)
+## vrev8 (also in zvkb)
+## vclz
+## vctz
+## vcpop
+## vrol (also in zvkb)
+## vror (also in zvkb)
+## vwsll
 
 sub vrev8_v {
     # vrev8.v vd, vs2, vm

crypto/sm4/asm/sm4-riscv64-zvksed.pl

@@ -36,9 +36,9 @@
 
 # The generated code of this file depends on the following RISC-V extensions:
 # - RV64I
-# - RISC-V vector ('V') with VLEN >= 128
-# - Vector Bit-manipulation used in Cryptography ('Zvbb')
-# - Vector ShangMi Suite: SM4 Block Cipher ('Zvksed')
+# - RISC-V Vector ('V') with VLEN >= 128
+# - RISC-V Vector Cryptography Bit-manipulation extension ('Zvkb')
+# - RISC-V Vector SM4 Block Cipher extension ('Zvksed')
 
 use strict;
 use warnings;

include/crypto/riscv_arch.def

@@ -35,12 +35,13 @@ RISCV_DEFINE_CAP(ZKT, 0, 13)
 RISCV_DEFINE_CAP(V, 0, 14)
 RISCV_DEFINE_CAP(ZVBB, 0, 15)
 RISCV_DEFINE_CAP(ZVBC, 0, 16)
-RISCV_DEFINE_CAP(ZVKG, 0, 17)
-RISCV_DEFINE_CAP(ZVKNED, 0, 18)
-RISCV_DEFINE_CAP(ZVKNHA, 0, 19)
-RISCV_DEFINE_CAP(ZVKNHB, 0, 20)
-RISCV_DEFINE_CAP(ZVKSED, 0, 21)
-RISCV_DEFINE_CAP(ZVKSH, 0, 22)
+RISCV_DEFINE_CAP(ZVKB, 0, 17)
+RISCV_DEFINE_CAP(ZVKG, 0, 18)
+RISCV_DEFINE_CAP(ZVKNED, 0, 19)
+RISCV_DEFINE_CAP(ZVKNHA, 0, 20)
+RISCV_DEFINE_CAP(ZVKNHB, 0, 21)
+RISCV_DEFINE_CAP(ZVKSED, 0, 22)
+RISCV_DEFINE_CAP(ZVKSH, 0, 23)
 
 /*
  * In the future ...

include/crypto/riscv_arch.h

@@ -60,10 +60,15 @@ static const size_t kRISCVNumCaps =
 #define RISCV_HAS_ZBB_AND_ZBC() (RISCV_HAS_ZBB() && RISCV_HAS_ZBC())
 #define RISCV_HAS_ZBKB_AND_ZKND_AND_ZKNE() (RISCV_HAS_ZBKB() && RISCV_HAS_ZKND() && RISCV_HAS_ZKNE())
 #define RISCV_HAS_ZKND_AND_ZKNE() (RISCV_HAS_ZKND() && RISCV_HAS_ZKNE())
-#define RISCV_HAS_ZVBB_AND_ZVKNHA() (RISCV_HAS_ZVBB() && RISCV_HAS_ZVKNHA())
-#define RISCV_HAS_ZVBB_AND_ZVKNHB() (RISCV_HAS_ZVBB() && RISCV_HAS_ZVKNHB())
-#define RISCV_HAS_ZVBB_AND_ZVKSED() (RISCV_HAS_ZVBB() && RISCV_HAS_ZVKSED())
-#define RISCV_HAS_ZVBB_AND_ZVKSH() (RISCV_HAS_ZVBB() && RISCV_HAS_ZVKSH())
+/*
+ * The ZVBB is the superset of ZVKB extension. We use macro here to replace the
+ * `RISCV_HAS_ZVKB()` with `RISCV_HAS_ZVBB() || RISCV_HAS_ZVKB()`.
+ */
+#define RISCV_HAS_ZVKB() (RISCV_HAS_ZVBB() || RISCV_HAS_ZVKB())
+#define RISCV_HAS_ZVKB_AND_ZVKNHA() (RISCV_HAS_ZVKB() && RISCV_HAS_ZVKNHA())
+#define RISCV_HAS_ZVKB_AND_ZVKNHB() (RISCV_HAS_ZVKB() && RISCV_HAS_ZVKNHB())
+#define RISCV_HAS_ZVKB_AND_ZVKSED() (RISCV_HAS_ZVKB() && RISCV_HAS_ZVKSED())
+#define RISCV_HAS_ZVKB_AND_ZVKSH() (RISCV_HAS_ZVKB() && RISCV_HAS_ZVKSH())
 
 /*
  * Get the size of a vector register in bits (VLEN).

providers/implementations/ciphers/cipher_sm4_ccm_hw_rv64i.inc

@@ -34,7 +34,7 @@ static const PROV_CCM_HW rv64i_zvksed_sm4_ccm = {
 
 const PROV_CCM_HW *ossl_prov_sm4_hw_ccm(size_t keybits)
 {
-    if (RISCV_HAS_ZVBB_AND_ZVKSED() && riscv_vlen() >= 128)
+    if (RISCV_HAS_ZVKB_AND_ZVKSED() && riscv_vlen() >= 128)
         return &rv64i_zvksed_sm4_ccm;
     else
         return &ccm_sm4;

providers/implementations/ciphers/cipher_sm4_gcm_hw_rv64i.inc

@@ -35,7 +35,7 @@ static const PROV_GCM_HW rv64i_zvksed_sm4_gcm = {
 
 const PROV_GCM_HW *ossl_prov_sm4_hw_gcm(size_t keybits)
 {
-    if (RISCV_HAS_ZVBB_AND_ZVKSED() && riscv_vlen() >= 128)
+    if (RISCV_HAS_ZVKB_AND_ZVKSED() && riscv_vlen() >= 128)
         return &rv64i_zvksed_sm4_gcm;
     else
         return &sm4_gcm;

providers/implementations/ciphers/cipher_sm4_hw_rv64i.inc

@@ -48,5 +48,5 @@ static const PROV_CIPHER_HW rv64i_zvksed_sm4_##mode = {                \
     cipher_hw_sm4_copyctx                                              \
 };
 #define PROV_CIPHER_HW_select(mode)                                    \
-if (RISCV_HAS_ZVBB_AND_ZVKSED() && riscv_vlen() >= 128)                \
+if (RISCV_HAS_ZVKB_AND_ZVKSED() && riscv_vlen() >= 128)                \
     return &rv64i_zvksed_sm4_##mode;

providers/implementations/ciphers/cipher_sm4_xts_hw_rv64i.inc

@@ -36,7 +36,7 @@ static const PROV_CIPHER_HW rv64i_zvksed_sm4_xts = {
 
 const PROV_CIPHER_HW *ossl_prov_cipher_hw_sm4_xts(size_t keybits)
 {
-    if (RISCV_HAS_ZVBB_AND_ZVKSED() && riscv_vlen() >= 128)
+    if (RISCV_HAS_ZVKB_AND_ZVKSED() && riscv_vlen() >= 128)
         return &rv64i_zvksed_sm4_xts;
     else
         return &sm4_generic_xts;