[Qemu-devel] [PATCH v4 07/69] target/arm: Simplify UMAAL

[Richard Henderson]

Since all of the inputs and outputs are i32, dispense with
the intermediate promotion to i64 and use tcg_gen_mulu2_i32
and tcg_gen_add2_i32.

Reviewed-by: Peter Maydell <address@hidden>
Signed-off-by: Richard Henderson <address@hidden>
---
 target/arm/translate.c | 34 ++++++++++++----------------------
 1 file changed, 12 insertions(+), 22 deletions(-)

diff --git a/target/arm/translate.c b/target/arm/translate.c
index 9bcf1e0964..8d70f15a4f 100644
--- a/target/arm/translate.c
+++ b/target/arm/translate.c
@@ -7346,21 +7346,6 @@ static void gen_storeq_reg(DisasContext *s, int rlow, 
int rhigh, TCGv_i64 val)
     store_reg(s, rhigh, tmp);
 }
 
-/* load a 32-bit value from a register and perform a 64-bit accumulate.  */
-static void gen_addq_lo(DisasContext *s, TCGv_i64 val, int rlow)
-{
-    TCGv_i64 tmp;
-    TCGv_i32 tmp2;
-
-    /* Load value and extend to 64 bits.  */
-    tmp = tcg_temp_new_i64();
-    tmp2 = load_reg(s, rlow);
-    tcg_gen_extu_i32_i64(tmp, tmp2);
-    tcg_temp_free_i32(tmp2);
-    tcg_gen_add_i64(val, val, tmp);
-    tcg_temp_free_i64(tmp);
-}
-
 /* load and add a 64-bit value from a register pair.  */
 static void gen_addq(DisasContext *s, TCGv_i64 val, int rlow, int rhigh)
 {
@@ -8119,8 +8104,7 @@ static bool trans_SMLAL(DisasContext *s, arg_SMLAL *a)
 
 static bool trans_UMAAL(DisasContext *s, arg_UMAAL *a)
 {
-    TCGv_i32 t0, t1;
-    TCGv_i64 t64;
+    TCGv_i32 t0, t1, t2, zero;
 
     if (s->thumb
         ? !arm_dc_feature(s, ARM_FEATURE_THUMB_DSP)
@@ -8130,11 +8114,17 @@ static bool trans_UMAAL(DisasContext *s, arg_UMAAL *a)
 
     t0 = load_reg(s, a->rm);
     t1 = load_reg(s, a->rn);
-    t64 = gen_mulu_i64_i32(t0, t1);
-    gen_addq_lo(s, t64, a->ra);
-    gen_addq_lo(s, t64, a->rd);
-    gen_storeq_reg(s, a->ra, a->rd, t64);
-    tcg_temp_free_i64(t64);
+    tcg_gen_mulu2_i32(t0, t1, t0, t1);
+    zero = tcg_const_i32(0);
+    t2 = load_reg(s, a->ra);
+    tcg_gen_add2_i32(t0, t1, t0, t1, t2, zero);
+    tcg_temp_free_i32(t2);
+    t2 = load_reg(s, a->rd);
+    tcg_gen_add2_i32(t0, t1, t0, t1, t2, zero);
+    tcg_temp_free_i32(t2);
+    tcg_temp_free_i32(zero);
+    store_reg(s, a->ra, t0);
+    store_reg(s, a->rd, t1);
     return true;
 }
 
-- 
2.17.1