@@ -148,6 +148,7 @@ bool aarch64_legitimate_pic_operand_p (rtx);
bool aarch64_move_imm (HOST_WIDE_INT, enum machine_mode);
bool aarch64_mov_operand_p (rtx, enum aarch64_symbol_context,
enum machine_mode);
+char *aarch64_output_scalar_simd_mov_immediate (rtx, enum machine_mode);
char *aarch64_output_simd_mov_immediate (rtx, enum machine_mode, unsigned);
bool aarch64_pad_arg_upward (enum machine_mode, const_tree);
bool aarch64_pad_reg_upward (enum machine_mode, const_tree, bool);
@@ -5988,32 +5988,57 @@ aarch64_vector_mode_supported_p (enum machine_mode mode)
return false;
}
-/* Return quad mode as the preferred SIMD mode. */
+/* Return appropriate SIMD container
+ for MODE within a vector of WIDTH bits. */
static enum machine_mode
-aarch64_preferred_simd_mode (enum machine_mode mode)
+aarch64_simd_container_mode (enum machine_mode mode, unsigned width)
{
+ gcc_assert (width == 64 || width == 128);
if (TARGET_SIMD)
- switch (mode)
- {
- case DFmode:
- return V2DFmode;
- case SFmode:
- return V4SFmode;
- case SImode:
- return V4SImode;
- case HImode:
- return V8HImode;
- case QImode:
- return V16QImode;
- case DImode:
- return V2DImode;
- break;
-
- default:;
- }
+ {
+ if (width == 128)
+ switch (mode)
+ {
+ case DFmode:
+ return V2DFmode;
+ case SFmode:
+ return V4SFmode;
+ case SImode:
+ return V4SImode;
+ case HImode:
+ return V8HImode;
+ case QImode:
+ return V16QImode;
+ case DImode:
+ return V2DImode;
+ default:
+ break;
+ }
+ else
+ switch (mode)
+ {
+ case SFmode:
+ return V2SFmode;
+ case SImode:
+ return V2SImode;
+ case HImode:
+ return V4HImode;
+ case QImode:
+ return V8QImode;
+ default:
+ break;
+ }
+ }
return word_mode;
}
+/* Return 128-bit container as the preferred SIMD mode for MODE. */
+static enum machine_mode
+aarch64_preferred_simd_mode (enum machine_mode mode)
+{
+ return aarch64_simd_container_mode (mode, 128);
+}
+
/* Return the bitmask of possible vector sizes for the vectorizer
to iterate over. */
static unsigned int
@@ -7280,6 +7305,18 @@ aarch64_output_simd_mov_immediate (rtx const_vector,
return templ;
}
+char*
+aarch64_output_scalar_simd_mov_immediate (rtx immediate,
+ enum machine_mode mode)
+{
+ enum machine_mode vmode;
+
+ gcc_assert (!VECTOR_MODE_P (mode));
+ vmode = aarch64_simd_container_mode (mode, 64);
+ rtx v_op = aarch64_simd_gen_const_vector_dup (vmode, INTVAL (immediate));
+ return aarch64_output_simd_mov_immediate (v_op, vmode, 64);
+}
+
/* Split operands into moves from op[1] + op[2] into op[0]. */
void
@@ -774,17 +774,34 @@
(match_operand:SHORT 1 "general_operand" " r,M,D<hq>,m, m,rZ,*w,*w, r,*w"))]
"(register_operand (operands[0], <MODE>mode)
|| aarch64_reg_or_zero (operands[1], <MODE>mode))"
- "@
- mov\\t%w0, %w1
- mov\\t%w0, %1
- movi\\t%0.<Vallxd>, %1
- ldr<size>\\t%w0, %1
- ldr\\t%<size>0, %1
- str<size>\\t%w1, %0
- str\\t%<size>1, %0
- umov\\t%w0, %1.<v>[0]
- dup\\t%0.<Vallxd>, %w1
- dup\\t%0, %1.<v>[0]"
+{
+ switch (which_alternative)
+ {
+ case 0:
+ return "mov\t%w0, %w1";
+ case 1:
+ return "mov\t%w0, %1";
+ case 2:
+ return aarch64_output_scalar_simd_mov_immediate (operands[1],
+ <MODE>mode);
+ case 3:
+ return "ldr<size>\t%w0, %1";
+ case 4:
+ return "ldr\t%<size>0, %1";
+ case 5:
+ return "str<size>\t%w1, %0";
+ case 6:
+ return "str\t%<size>1, %0";
+ case 7:
+ return "umov\t%w0, %1.<v>[0]";
+ case 8:
+ return "dup\t%0.<Vallxd>, %w1";
+ case 9:
+ return "dup\t%0, %1.<v>[0]";
+ default:
+ gcc_unreachable ();
+ }
+}
[(set_attr "v8type" "move,alu,alu,load1,load1,store1,store1,*,*,*")
(set_attr "simd_type" "*,*,simd_move_imm,*,*,*,*,simd_movgp,simd_dupgp,simd_dup")
(set_attr "simd" "*,*,yes,*,*,*,*,yes,yes,yes")
new file mode 100644
@@ -0,0 +1,13 @@
+/* { dg-do compile } */
+/* { dg-options "-O2" } */
+
+void
+dummy (short* b)
+{
+ /* { dg-final { scan-assembler "movi\tv\[0-9\]+\.4h, 0x4, lsl 8" } } */
+ /* { dg-final { scan-assembler-not "movi\tv\[0-9\]+\.4h, 0x400" } } */
+ /* { dg-final { scan-assembler-not "movi\tv\[0-9\]+\.4h, 1024" } } */
+ register short x asm ("h8") = 1024;
+ asm volatile ("" : : "w" (x));
+ *b = x;
+}