On 2023/6/28 5:14 PM, Richard Henderson wrote:
On 6/27/23 19:45, Max Chou wrote:
+void HELPER(vsha2cl_vv)(void *vd, void *vs1, void *vs2,
CPURISCVState *env,
+ uint32_t desc)
+{
+ uint32_t sew = FIELD_EX64(env->vtype, VTYPE, VSEW);
+ uint32_t esz = sew == MO_64 ? 8 : 4;
+ uint32_t total_elems;
+ uint32_t vta = vext_vta(desc);
+
+ for (uint32_t i = env->vstart / 4; i < env->vl / 4; i++) {
+ if (sew == MO_64) {
+ vsha2c_64(((uint64_t *)vs2) + 4 * i, ((uint64_t *)vd) +
4 * i,
+ (((uint64_t *)vs1) + 4 * i));
+ } else {
+ vsha2c_32(((uint32_t *)vs2) + 4 * i, ((uint32_t *)vd) +
4 * i,
+ (((uint32_t *)vs1) + 4 * i));
+ }
+ }
You should have two separate functions for two separate VSEW, so that
you do not need to pass the value nor decode the value here. The
selection of SEW should happen during translation using DisasContext.sew.
r~
Thanks for the suggestion. I'll provide the v7 patch set for this.
Thanks,
Max
