ISA_EXT_DATA_ENTRY(zvkned, PRIV_VERSION_1_12_0, ext_zvkned),
ISA_EXT_DATA_ENTRY(zvknha, PRIV_VERSION_1_12_0, ext_zvknha),
ISA_EXT_DATA_ENTRY(zvknhb, PRIV_VERSION_1_12_0, ext_zvknhb),
+ ISA_EXT_DATA_ENTRY(zvksh, PRIV_VERSION_1_12_0, ext_zvksh),
ISA_EXT_DATA_ENTRY(zhinx, PRIV_VERSION_1_12_0, ext_zhinx),
ISA_EXT_DATA_ENTRY(zhinxmin, PRIV_VERSION_1_12_0, ext_zhinxmin),
ISA_EXT_DATA_ENTRY(smaia, PRIV_VERSION_1_12_0, ext_smaia),
* In principle Zve*x would also suffice here, were they supported
* in qemu
*/
- if ((cpu->cfg.ext_zvbb || cpu->cfg.ext_zvkned || cpu->cfg.ext_zvknha) &&
- !cpu->cfg.ext_zve32f) {
+ if ((cpu->cfg.ext_zvbb || cpu->cfg.ext_zvkned || cpu->cfg.ext_zvknha ||
+ cpu->cfg.ext_zvksh) && !cpu->cfg.ext_zve32f) {
error_setg(errp,
"Vector crypto extensions require V or Zve* extensions");
return;
DEFINE_PROP_BOOL("x-zvkned", RISCVCPU, cfg.ext_zvkned, false),
DEFINE_PROP_BOOL("x-zvknha", RISCVCPU, cfg.ext_zvknha, false),
DEFINE_PROP_BOOL("x-zvknhb", RISCVCPU, cfg.ext_zvknhb, false),
+ DEFINE_PROP_BOOL("x-zvksh", RISCVCPU, cfg.ext_zvksh, false),
DEFINE_PROP_END_OF_LIST(),
};
vext_set_elems_1s(vd, vta, env->vl * esz, total_elems * esz);
env->vstart = 0;
}
+
+static inline uint32_t p1(uint32_t x)
+{
+ return x ^ rol32(x, 15) ^ rol32(x, 23);
+}
+
+static inline uint32_t zvksh_w(uint32_t m16, uint32_t m9, uint32_t m3,
+ uint32_t m13, uint32_t m6)
+{
+ return p1(m16 ^ m9 ^ rol32(m3, 15)) ^ rol32(m13, 7) ^ m6;
+}
+
+void HELPER(vsm3me_vv)(void *vd_vptr, void *vs1_vptr, void *vs2_vptr,
+ CPURISCVState *env, uint32_t desc)
+{
+ uint32_t esz = memop_size(FIELD_EX64(env->vtype, VTYPE, VSEW));
+ uint32_t total_elems = vext_get_total_elems(env, desc, esz);
+ uint32_t vta = vext_vta(desc);
+ uint32_t *vd = vd_vptr;
+ uint32_t *vs1 = vs1_vptr;
+ uint32_t *vs2 = vs2_vptr;
+
+ for (int i = env->vstart / 8; i < env->vl / 8; i++) {
+ uint32_t w[24];
+ for (int j = 0; j < 8; j++) {
+ w[j] = bswap32(vs1[H4((i * 8) + j)]);
+ w[j + 8] = bswap32(vs2[H4((i * 8) + j)]);
+ }
+ for (int j = 0; j < 8; j++) {
+ w[j + 16] =
+ zvksh_w(w[j], w[j + 7], w[j + 13], w[j + 3], w[j + 10]);
+ }
+ for (int j = 0; j < 8; j++) {
+ vd[(i * 8) + j] = bswap32(w[H4(j + 16)]);
+ }
+ }
+ vext_set_elems_1s(vd_vptr, vta, env->vl * esz, total_elems * esz);
+ env->vstart = 0;
+}
+
+static inline uint32_t ff1(uint32_t x, uint32_t y, uint32_t z)
+{
+ return x ^ y ^ z;
+}
+
+static inline uint32_t ff2(uint32_t x, uint32_t y, uint32_t z)
+{
+ return (x & y) | (x & z) | (y & z);
+}
+
+static inline uint32_t ff_j(uint32_t x, uint32_t y, uint32_t z, uint32_t j)
+{
+ return (j <= 15) ? ff1(x, y, z) : ff2(x, y, z);
+}
+
+static inline uint32_t gg1(uint32_t x, uint32_t y, uint32_t z)
+{
+ return x ^ y ^ z;
+}
+
+static inline uint32_t gg2(uint32_t x, uint32_t y, uint32_t z)
+{
+ return (x & y) | (~x & z);
+}
+
+static inline uint32_t gg_j(uint32_t x, uint32_t y, uint32_t z, uint32_t j)
+{
+ return (j <= 15) ? gg1(x, y, z) : gg2(x, y, z);
+}
+
+static inline uint32_t t_j(uint32_t j)
+{
+ return (j <= 15) ? 0x79cc4519 : 0x7a879d8a;
+}
+
+static inline uint32_t p_0(uint32_t x)
+{
+ return x ^ rol32(x, 9) ^ rol32(x, 17);
+}
+
+static void sm3c(uint32_t *vd, uint32_t *vs1, uint32_t *vs2, uint32_t uimm)
+{
+ uint32_t x0, x1;
+ uint32_t j;
+ uint32_t ss1, ss2, tt1, tt2;
+ x0 = vs2[0] ^ vs2[4];
+ x1 = vs2[1] ^ vs2[5];
+ j = 2 * uimm;
+ ss1 = rol32(rol32(vs1[0], 12) + vs1[4] + rol32(t_j(j), j % 32), 7);
+ ss2 = ss1 ^ rol32(vs1[0], 12);
+ tt1 = ff_j(vs1[0], vs1[1], vs1[2], j) + vs1[3] + ss2 + x0;
+ tt2 = gg_j(vs1[4], vs1[5], vs1[6], j) + vs1[7] + ss1 + vs2[0];
+ vs1[3] = vs1[2];
+ vd[3] = rol32(vs1[1], 9);
+ vs1[1] = vs1[0];
+ vd[1] = tt1;
+ vs1[7] = vs1[6];
+ vd[7] = rol32(vs1[5], 19);
+ vs1[5] = vs1[4];
+ vd[5] = p_0(tt2);
+ j = 2 * uimm + 1;
+ ss1 = rol32(rol32(vd[1], 12) + vd[5] + rol32(t_j(j), j % 32), 7);
+ ss2 = ss1 ^ rol32(vd[1], 12);
+ tt1 = ff_j(vd[1], vs1[1], vd[3], j) + vs1[3] + ss2 + x1;
+ tt2 = gg_j(vd[5], vs1[5], vd[7], j) + vs1[7] + ss1 + vs2[1];
+ vd[2] = rol32(vs1[1], 9);
+ vd[0] = tt1;
+ vd[6] = rol32(vs1[5], 19);
+ vd[4] = p_0(tt2);
+}
+
+void HELPER(vsm3c_vi)(void *vd_vptr, void *vs2_vptr, uint32_t uimm,
+ CPURISCVState *env, uint32_t desc)
+{
+ uint32_t esz = memop_size(FIELD_EX64(env->vtype, VTYPE, VSEW));
+ uint32_t total_elems = vext_get_total_elems(env, desc, esz);
+ uint32_t vta = vext_vta(desc);
+ uint32_t *vd = vd_vptr;
+ uint32_t *vs2 = vs2_vptr;
+ uint32_t v1[8], v2[8], v3[8];
+
+ for (int i = env->vstart / 8; i < env->vl / 8; i++) {
+ for (int k = 0; k < 8; k++) {
+ v2[k] = bswap32(vd[H4(i * 8 + k)]);
+ v3[k] = bswap32(vs2[H4(i * 8 + k)]);
+ }
+ sm3c(v1, v2, v3, uimm);
+ for (int k = 0; k < 8; k++) {
+ vd[i * 8 + k] = bswap32(v1[H4(k)]);
+ }
+ }
+ vext_set_elems_1s(vd_vptr, vta, env->vl * esz, total_elems * esz);
+ env->vstart = 0;
+}