target/mips: msa: Split helpers for SUBV.<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: <20200613152133.8964-14-aleksandar.qemu.devel@gmail.com>

3 files changed, 91 insertions, 11 deletions

diff --git a/target/mips/helper.h b/target/mips/helper.h
index 4795c97f47..5d7ba6a847 100644
--- a/target/mips/helper.h
+++ b/target/mips/helper.h
@@ -998,6 +998,11 @@ DEF_HELPER_4(msa_subsuu_s_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_subsuu_s_w, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_subsuu_s_d, void, env, i32, i32, i32)
 
+DEF_HELPER_4(msa_subv_b, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_subv_h, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_subv_w, void, env, i32, i32, i32)
+DEF_HELPER_4(msa_subv_d, void, env, i32, i32, i32)
+
 DEF_HELPER_4(msa_ilvev_b, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_ilvev_h, void, env, i32, i32, i32)
 DEF_HELPER_4(msa_ilvev_w, void, env, i32, i32, i32)
@@ -1093,7 +1098,6 @@ 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_subv_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)
diff --git a/target/mips/msa_helper.c b/target/mips/msa_helper.c
index 27a9c36a89..d099e00b40 100644
--- a/target/mips/msa_helper.c
+++ b/target/mips/msa_helper.c
@@ -3553,9 +3553,6 @@ void helper_msa_asub_u_d(CPUMIPSState *env,
 }
 
 
-/* TODO: insert the rest of Int Subtract group helpers here */
-
-
 static inline int64_t msa_hsub_s_df(uint32_t df, int64_t arg1, int64_t arg2)
 {
     return SIGNED_ODD(arg1, df) - SIGNED_EVEN(arg2, df);
@@ -3970,6 +3967,78 @@ void helper_msa_subsuu_s_d(CPUMIPSState *env,
 }
 
 
+static inline int64_t msa_subv_df(uint32_t df, int64_t arg1, int64_t arg2)
+{
+    return arg1 - arg2;
+}
+
+void helper_msa_subv_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_subv_df(DF_BYTE, pws->b[0],  pwt->b[0]);
+    pwd->b[1]  = msa_subv_df(DF_BYTE, pws->b[1],  pwt->b[1]);
+    pwd->b[2]  = msa_subv_df(DF_BYTE, pws->b[2],  pwt->b[2]);
+    pwd->b[3]  = msa_subv_df(DF_BYTE, pws->b[3],  pwt->b[3]);
+    pwd->b[4]  = msa_subv_df(DF_BYTE, pws->b[4],  pwt->b[4]);
+    pwd->b[5]  = msa_subv_df(DF_BYTE, pws->b[5],  pwt->b[5]);
+    pwd->b[6]  = msa_subv_df(DF_BYTE, pws->b[6],  pwt->b[6]);
+    pwd->b[7]  = msa_subv_df(DF_BYTE, pws->b[7],  pwt->b[7]);
+    pwd->b[8]  = msa_subv_df(DF_BYTE, pws->b[8],  pwt->b[8]);
+    pwd->b[9]  = msa_subv_df(DF_BYTE, pws->b[9],  pwt->b[9]);
+    pwd->b[10] = msa_subv_df(DF_BYTE, pws->b[10], pwt->b[10]);
+    pwd->b[11] = msa_subv_df(DF_BYTE, pws->b[11], pwt->b[11]);
+    pwd->b[12] = msa_subv_df(DF_BYTE, pws->b[12], pwt->b[12]);
+    pwd->b[13] = msa_subv_df(DF_BYTE, pws->b[13], pwt->b[13]);
+    pwd->b[14] = msa_subv_df(DF_BYTE, pws->b[14], pwt->b[14]);
+    pwd->b[15] = msa_subv_df(DF_BYTE, pws->b[15], pwt->b[15]);
+}
+
+void helper_msa_subv_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_subv_df(DF_HALF, pws->h[0],  pwt->h[0]);
+    pwd->h[1]  = msa_subv_df(DF_HALF, pws->h[1],  pwt->h[1]);
+    pwd->h[2]  = msa_subv_df(DF_HALF, pws->h[2],  pwt->h[2]);
+    pwd->h[3]  = msa_subv_df(DF_HALF, pws->h[3],  pwt->h[3]);
+    pwd->h[4]  = msa_subv_df(DF_HALF, pws->h[4],  pwt->h[4]);
+    pwd->h[5]  = msa_subv_df(DF_HALF, pws->h[5],  pwt->h[5]);
+    pwd->h[6]  = msa_subv_df(DF_HALF, pws->h[6],  pwt->h[6]);
+    pwd->h[7]  = msa_subv_df(DF_HALF, pws->h[7],  pwt->h[7]);
+}
+
+void helper_msa_subv_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_subv_df(DF_WORD, pws->w[0],  pwt->w[0]);
+    pwd->w[1]  = msa_subv_df(DF_WORD, pws->w[1],  pwt->w[1]);
+    pwd->w[2]  = msa_subv_df(DF_WORD, pws->w[2],  pwt->w[2]);
+    pwd->w[3]  = msa_subv_df(DF_WORD, pws->w[3],  pwt->w[3]);
+}
+
+void helper_msa_subv_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_subv_df(DF_DOUBLE, pws->d[0],  pwt->d[0]);
+    pwd->d[1]  = msa_subv_df(DF_DOUBLE, pws->d[1],  pwt->d[1]);
+}
+
+
 /*
  * Interleave
  * ----------
@@ -5194,11 +5263,6 @@ void helper_msa_shf_df(CPUMIPSState *env, uint32_t df, uint32_t wd,
     msa_move_v(pwd, pwx);
 }
 
-static inline int64_t msa_subv_df(uint32_t df, int64_t arg1, int64_t arg2)
-{
-    return arg1 - arg2;
-}
-
 #define MSA_BINOP_IMM_DF(helper, func)                                  \
 void helper_msa_ ## helper ## _df(CPUMIPSState *env, uint32_t df,       \
                         uint32_t wd, uint32_t ws, int32_t u5)           \
@@ -5502,7 +5566,6 @@ void helper_msa_ ## func ## _df(CPUMIPSState *env, uint32_t df,         \
     }                                                                   \
 }
 
-MSA_BINOP_DF(subv)
 MSA_BINOP_DF(mulv)
 
 MSA_BINOP_DF(mul_q)
diff --git a/target/mips/translate.c b/target/mips/translate.c
index 6d9ae1de4b..9ca17ed02c 100644
--- a/target/mips/translate.c
+++ b/target/mips/translate.c
@@ -29324,7 +29324,20 @@ static void gen_msa_3r(CPUMIPSState *env, DisasContext *ctx)
         gen_helper_msa_vshf_df(cpu_env, tdf, twd, tws, twt);
         break;
     case OPC_SUBV_df:
-        gen_helper_msa_subv_df(cpu_env, tdf, twd, tws, twt);
+        switch (df) {
+        case DF_BYTE:
+            gen_helper_msa_subv_b(cpu_env, twd, tws, twt);
+            break;
+        case DF_HALF:
+            gen_helper_msa_subv_h(cpu_env, twd, tws, twt);
+            break;
+        case DF_WORD:
+            gen_helper_msa_subv_w(cpu_env, twd, tws, twt);
+            break;
+        case DF_DOUBLE:
+            gen_helper_msa_subv_d(cpu_env, twd, tws, twt);
+            break;
+        }
         break;
     case OPC_SUBS_U_df:
         switch (df) {