LoongArch: Implement vec_init<M><N> where N is a LSX vector mode

This patch implements more vec_init optabs that can handle two LSX vectors producing a LASX
vector by concatenating them. When an lsx vector is concatenated with an LSX const_vector of
zeroes, the vec_concatz pattern can be used effectively. For example as below

typedef short v8hi __attribute__ ((vector_size (16)));
typedef short v16hi __attribute__ ((vector_size (32)));
v8hi a, b;

v16hi vec_initv16hiv8hi ()
{
 return __builtin_shufflevector (a, b, 0, 8, 1, 9, 2, 10, 3, 11, 4, 12, 5, 13, 6, 14, 7, 15);
}

Before this patch:

vec_initv16hiv8hi:
    addi.d  $r3,$r3,-64
    .cfi_def_cfa_offset 64
    xvrepli.h   $xr0,0
    la.local    $r12,.LANCHOR0
    xvst    $xr0,$r3,0
    xvst    $xr0,$r3,32
    vld $vr0,$r12,0
    vst $vr0,$r3,0
    vld $vr0,$r12,16
    vst $vr0,$r3,32
    xvld    $xr1,$r3,32
    xvld    $xr2,$r3,32
    xvld    $xr0,$r3,0
    xvilvh.h    $xr0,$xr1,$xr0
    xvld    $xr1,$r3,0
    xvilvl.h    $xr1,$xr2,$xr1
    addi.d  $r3,$r3,64
    .cfi_def_cfa_offset 0
    xvpermi.q   $xr0,$xr1,32
    jr  $r1

After this patch:

vec_initv16hiv8hi:
    la.local        $r12,.LANCHOR0
    vld     $vr0,$r12,32
    vld     $vr2,$r12,48
    xvilvh.h        $xr1,$xr2,$xr0
    xvilvl.h        $xr0,$xr2,$xr0
    xvpermi.q       $xr1,$xr0,32
    xvst    $xr1,$r4,0
    jr      $r1

gcc/ChangeLog:

	* config/loongarch/lasx.md (vec_initv32qiv16qi): Rename to ..
	(vec_init<mode><lasxhalf>): .. this, and extend to mode.
	(@vec_concatz<mode>): New insn pattern.
	* config/loongarch/loongarch.cc (loongarch_expand_vector_group_init):
	Handle VALS containing two vectors.

gcc/testsuite/ChangeLog:

	* gcc.target/loongarch/vector/lasx/lasx-vec-init-2.c: New test.

3 files changed, 128 insertions, 7 deletions

diff --git a/gcc/config/loongarch/lasx.md b/gcc/config/loongarch/lasx.md
index 6c7e373..c2bde4d 100644
--- a/gcc/config/loongarch/lasx.md
+++ b/gcc/config/loongarch/lasx.md
@@ -465,6 +465,11 @@
    (V16HI "w")
    (V32QI "w")])
 
+;; Half modes of all LASX vector modes, in lower-case.
+(define_mode_attr lasxhalf [(V32QI "v16qi")  (V16HI "v8hi")
+             (V8SI "v4si")  (V4DI  "v2di")
+             (V8SF  "v4sf") (V4DF  "v2df")])
+
 (define_expand "vec_init<mode><unitmode>"
   [(match_operand:LASX 0 "register_operand")
    (match_operand:LASX 1 "")]
@@ -474,9 +479,9 @@
   DONE;
 })
 
-(define_expand "vec_initv32qiv16qi"
- [(match_operand:V32QI 0 "register_operand")
-  (match_operand:V16QI 1 "")]
+(define_expand "vec_init<mode><lasxhalf>"
+ [(match_operand:LASX 0 "register_operand")
+  (match_operand:<VHMODE256_ALL> 1 "")]
   "ISA_HAS_LASX"
 {
   loongarch_expand_vector_group_init (operands[0], operands[1]);
@@ -577,6 +582,21 @@
   [(set_attr "type" "simd_insert")
    (set_attr "mode" "<MODE>")])
 
+(define_insn "@vec_concatz<mode>"
+  [(set (match_operand:LASX 0 "register_operand" "=f")
+    (vec_concat:LASX
+      (match_operand:<VHMODE256_ALL> 1 "nonimmediate_operand")
+      (match_operand:<VHMODE256_ALL> 2 "const_0_operand")))]
+  "ISA_HAS_LASX"
+{
+  if (MEM_P (operands[1]))
+    return "vld\t%w0,%1";
+  else
+    return "vori.b\t%w0,%w1,0";
+}
+  [(set_attr "type" "simd_splat")
+   (set_attr "mode" "<MODE>")])
+
 (define_insn "vec_concat<mode>"
   [(set (match_operand:LASX 0 "register_operand" "=f")
 	(vec_concat:LASX
diff --git a/gcc/config/loongarch/loongarch.cc b/gcc/config/loongarch/loongarch.cc
index ec376a7..1c90afc 100644
--- a/gcc/config/loongarch/loongarch.cc
+++ b/gcc/config/loongarch/loongarch.cc
@@ -9847,10 +9847,46 @@ loongarch_gen_const_int_vector_shuffle (machine_mode mode, int val)
 void
 loongarch_expand_vector_group_init (rtx target, rtx vals)
 {
-  rtx ops[2] = { force_reg (E_V16QImode, XVECEXP (vals, 0, 0)),
-      force_reg (E_V16QImode, XVECEXP (vals, 0, 1)) };
-  emit_insn (gen_rtx_SET (target, gen_rtx_VEC_CONCAT (E_V32QImode, ops[0],
-						      ops[1])));
+  machine_mode vmode = GET_MODE (target);
+  machine_mode half_mode = VOIDmode;
+  rtx low = XVECEXP (vals, 0, 0);
+  rtx high = XVECEXP (vals, 0, 1);
+
+  switch (vmode)
+    {
+    case E_V32QImode:
+      half_mode = V16QImode;
+      break;
+    case E_V16HImode:
+      half_mode = V8HImode;
+      break;
+    case E_V8SImode:
+      half_mode = V4SImode;
+      break;
+    case E_V4DImode:
+      half_mode = V2DImode;
+      break;
+    case E_V8SFmode:
+      half_mode = V4SFmode;
+      break;
+    case E_V4DFmode:
+      half_mode = V2DFmode;
+      break;
+    default:
+      gcc_unreachable ();
+    }
+
+  if (high == CONST0_RTX (half_mode))
+    emit_insn (gen_vec_concatz (vmode, target, low, high));
+  else
+    {
+      if (!register_operand (low, half_mode))
+	low = force_reg (half_mode, low);
+      if (!register_operand (high, half_mode))
+	high = force_reg (half_mode, high);
+      emit_insn (gen_rtx_SET (target,
+			      gen_rtx_VEC_CONCAT (vmode, low, high)));
+    }
 }
 
 /* Expand initialization of a vector which has all same elements.  */
diff --git a/gcc/testsuite/gcc.target/loongarch/vector/lasx/lasx-vec-init-2.c b/gcc/testsuite/gcc.target/loongarch/vector/lasx/lasx-vec-init-2.c
new file mode 100644
index 0000000..7592198
--- /dev/null
+++ b/gcc/testsuite/gcc.target/loongarch/vector/lasx/lasx-vec-init-2.c
@@ -0,0 +1,65 @@
+/* { dg-do compile } */
+/* { dg-options "-O3 -fno-vect-cost-model -mlasx" } */
+/* { dg-final { scan-assembler-times "vld" 12 } } */
+
+
+typedef char v16qi __attribute__ ((vector_size (16)));
+typedef char v32qi __attribute__ ((vector_size (32)));
+
+typedef short v8hi __attribute__ ((vector_size (16)));
+typedef short v16hi __attribute__ ((vector_size (32)));
+
+typedef int v4si __attribute__ ((vector_size (16)));
+typedef int v8si __attribute__ ((vector_size (32)));
+
+typedef long v2di __attribute__ ((vector_size (16)));
+typedef long v4di __attribute__ ((vector_size (32)));
+
+typedef float v4sf __attribute__ ((vector_size (16)));
+typedef float v8sf __attribute__ ((vector_size (32)));
+
+typedef double v2df __attribute__ ((vector_size (16)));
+typedef double v4df __attribute__ ((vector_size (32)));
+
+v16qi a_qi, b_qi;
+v8hi  a_hi, b_hi;
+v4si  a_si, b_si;
+v2di  a_di, b_di;
+v4sf  a_sf, b_sf;
+v2df  a_df, b_df;
+
+v32qi
+foo_v32qi ()
+{
+  return __builtin_shufflevector (a_qi, b_qi, 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23, 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31);
+}
+
+v16hi
+foo_v16qi ()
+{
+  return __builtin_shufflevector (a_hi, b_hi, 0, 8, 1, 9, 2, 10, 3, 11, 4, 12, 5, 13, 6, 14, 7, 15);
+}
+
+v8si
+foo_v8si ()
+{
+  return __builtin_shufflevector (a_si, b_si, 0, 4, 1, 5, 2, 6, 3, 7);
+}
+
+v4di
+foo_v4di ()
+{
+  return __builtin_shufflevector (a_di, b_di, 0, 2, 1, 3);
+}
+
+v8sf
+foo_v8sf ()
+{
+  return __builtin_shufflevector (a_sf, b_sf, 0, 4, 1, 5, 2, 6, 3, 7);
+}
+
+v4df
+foo_v4df ()
+{
+  return __builtin_shufflevector (a_df, b_df, 0, 2, 1, 3);
+}