target/mips: msa: Split helpers for MULV.<B|H|W|D>

Achieves clearer code and slightly better performance.

Reviewed-by: Aleksandar Rikalo <aleksandar.rikalo@syrmia.com>
Signed-off-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
Message-Id: <20200613152133.8964-15-aleksandar.qemu.devel@gmail.com>

target/mips/helper.h

@@ -960,6 +960,11 @@ DEF_HELPER_4(msa_msubv_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_msubv_w, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_msubv_d, void, env, i32, i32, i32)
 
+DEF_HELPER_4(msa_mulv_b, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_mulv_h, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_mulv_w, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_mulv_d, void, env, i32, i32, i32)
+
 DEF_HELPER_4(msa_asub_s_b, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_asub_s_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_asub_s_w, void, env, i32, i32, i32)
@@ -1098,7 +1103,6 @@ DEF_HELPER_5(msa_srlri_df, void, env, i32, i32, i32, i32)
 
 DEF_HELPER_5(msa_binsl_df, void, env, i32, i32, i32, i32)
 DEF_HELPER_5(msa_binsr_df, void, env, i32, i32, i32, i32)
-DEF_HELPER_5(msa_mulv_df, void, env, i32, i32, i32, i32)
 
 DEF_HELPER_4(msa_dotp_s_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_dotp_s_w, void, env, i32, i32, i32)

target/mips/msa_helper.c

@@ -3360,6 +3360,78 @@ void helper_msa_msubv_d(CPUMIPSState *env,
 }
 
 
+static inline int64_t msa_mulv_df(uint32_t df, int64_t arg1, int64_t arg2)
+{
+    return arg1 * arg2;
+}
+
+void helper_msa_mulv_b(CPUMIPSState *env,
+                       uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->b[0]  = msa_mulv_df(DF_BYTE, pws->b[0],  pwt->b[0]);
+    pwd->b[1]  = msa_mulv_df(DF_BYTE, pws->b[1],  pwt->b[1]);
+    pwd->b[2]  = msa_mulv_df(DF_BYTE, pws->b[2],  pwt->b[2]);
+    pwd->b[3]  = msa_mulv_df(DF_BYTE, pws->b[3],  pwt->b[3]);
+    pwd->b[4]  = msa_mulv_df(DF_BYTE, pws->b[4],  pwt->b[4]);
+    pwd->b[5]  = msa_mulv_df(DF_BYTE, pws->b[5],  pwt->b[5]);
+    pwd->b[6]  = msa_mulv_df(DF_BYTE, pws->b[6],  pwt->b[6]);
+    pwd->b[7]  = msa_mulv_df(DF_BYTE, pws->b[7],  pwt->b[7]);
+    pwd->b[8]  = msa_mulv_df(DF_BYTE, pws->b[8],  pwt->b[8]);
+    pwd->b[9]  = msa_mulv_df(DF_BYTE, pws->b[9],  pwt->b[9]);
+    pwd->b[10] = msa_mulv_df(DF_BYTE, pws->b[10], pwt->b[10]);
+    pwd->b[11] = msa_mulv_df(DF_BYTE, pws->b[11], pwt->b[11]);
+    pwd->b[12] = msa_mulv_df(DF_BYTE, pws->b[12], pwt->b[12]);
+    pwd->b[13] = msa_mulv_df(DF_BYTE, pws->b[13], pwt->b[13]);
+    pwd->b[14] = msa_mulv_df(DF_BYTE, pws->b[14], pwt->b[14]);
+    pwd->b[15] = msa_mulv_df(DF_BYTE, pws->b[15], pwt->b[15]);
+}
+
+void helper_msa_mulv_h(CPUMIPSState *env,
+                       uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->h[0]  = msa_mulv_df(DF_HALF, pws->h[0],  pwt->h[0]);
+    pwd->h[1]  = msa_mulv_df(DF_HALF, pws->h[1],  pwt->h[1]);
+    pwd->h[2]  = msa_mulv_df(DF_HALF, pws->h[2],  pwt->h[2]);
+    pwd->h[3]  = msa_mulv_df(DF_HALF, pws->h[3],  pwt->h[3]);
+    pwd->h[4]  = msa_mulv_df(DF_HALF, pws->h[4],  pwt->h[4]);
+    pwd->h[5]  = msa_mulv_df(DF_HALF, pws->h[5],  pwt->h[5]);
+    pwd->h[6]  = msa_mulv_df(DF_HALF, pws->h[6],  pwt->h[6]);
+    pwd->h[7]  = msa_mulv_df(DF_HALF, pws->h[7],  pwt->h[7]);
+}
+
+void helper_msa_mulv_w(CPUMIPSState *env,
+                       uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->w[0]  = msa_mulv_df(DF_WORD, pws->w[0],  pwt->w[0]);
+    pwd->w[1]  = msa_mulv_df(DF_WORD, pws->w[1],  pwt->w[1]);
+    pwd->w[2]  = msa_mulv_df(DF_WORD, pws->w[2],  pwt->w[2]);
+    pwd->w[3]  = msa_mulv_df(DF_WORD, pws->w[3],  pwt->w[3]);
+}
+
+void helper_msa_mulv_d(CPUMIPSState *env,
+                       uint32_t wd, uint32_t ws, uint32_t wt)
+{
+    wr_t *pwd = &(env->active_fpu.fpr[wd].wr);
+    wr_t *pws = &(env->active_fpu.fpr[ws].wr);
+    wr_t *pwt = &(env->active_fpu.fpr[wt].wr);
+
+    pwd->d[0]  = msa_mulv_df(DF_DOUBLE, pws->d[0],  pwt->d[0]);
+    pwd->d[1]  = msa_mulv_df(DF_DOUBLE, pws->d[1],  pwt->d[1]);
+}
+
+
 /*
  * Int Subtract
  * ------------
@@ -5444,11 +5516,6 @@ MSA_TEROP_IMMU_DF(binsli, binsl)
 MSA_TEROP_IMMU_DF(binsri, binsr)
 #undef MSA_TEROP_IMMU_DF
 
-static inline int64_t msa_mulv_df(uint32_t df, int64_t arg1, int64_t arg2)
-{
-    return arg1 * arg2;
-}
-
 #define CONCATENATE_AND_SLIDE(s, k)             \
     do {                                        \
         for (i = 0; i < s; i++) {               \
@@ -5566,8 +5633,6 @@ void helper_msa_ ## func ## _df(CPUMIPSState *env, uint32_t df,         \
     }                                                                   \
 }
 
-MSA_BINOP_DF(mulv)
-
 MSA_BINOP_DF(mul_q)
 MSA_BINOP_DF(mulr_q)
 #undef MSA_BINOP_DF

target/mips/translate.c

@@ -29315,7 +29315,20 @@ static void gen_msa_3r(CPUMIPSState *env, DisasContext *ctx)
         }
         break;
     case OPC_MULV_df:
-        gen_helper_msa_mulv_df(cpu_env, tdf, twd, tws, twt);
+        switch (df) {
+        case DF_BYTE:
+            gen_helper_msa_mulv_b(cpu_env, twd, tws, twt);
+            break;
+        case DF_HALF:
+            gen_helper_msa_mulv_h(cpu_env, twd, tws, twt);
+            break;
+        case DF_WORD:
+            gen_helper_msa_mulv_w(cpu_env, twd, tws, twt);
+            break;
+        case DF_DOUBLE:
+            gen_helper_msa_mulv_d(cpu_env, twd, tws, twt);
+            break;
+        }
         break;
     case OPC_SLD_df:
         gen_helper_msa_sld_df(cpu_env, tdf, twd, tws, twt);