Support 128-bit constant store in 64-bit STVhjl/pieces/move

We can load a 128-bit constant from memory and store it.

gcc/

	* config/i386/i386.c (timode_scalar_to_vector_candidate_p): Allow
	store from 128-bit constant.
	(timode_scalar_chain::convert_insn): Handle store from 128-bit
	constant.

gcc/testsuite/

	* gcc.target/i386/pieces-strcpy-1.c: New test.
	* gcc.target/i386/pieces-strcpy-2.c: Likewise.

3 files changed, 57 insertions, 0 deletions

diff --git a/gcc/config/i386/i386.c b/gcc/config/i386/i386.c
index 60dc16052e0..00a15e028d8 100644
--- a/gcc/config/i386/i386.c
+++ b/gcc/config/i386/i386.c
@@ -2877,6 +2877,12 @@ timode_scalar_to_vector_candidate_p (rtx_insn *insn)
 	case REG:
 	  return true;
 
+	case CONST_WIDE_INT:
+	  /* For store from 128-bit constant, memory must be aligned
+	     or unaligned store is optimal.  */
+	  return (!misaligned_operand (dst, TImode)
+		  || TARGET_SSE_UNALIGNED_STORE_OPTIMAL);
+
 	case CONST_INT:
 	  return standard_sse_constant_p (src, TImode);
 	}
@@ -3868,6 +3874,19 @@ timode_scalar_chain::convert_insn (rtx_insn *insn)
       PUT_MODE (src, V1TImode);
       break;
 
+    case CONST_WIDE_INT:
+      if (NONDEBUG_INSN_P (insn))
+	{
+	  /* Since there are no instructions to store 128-bit constant,
+	     temporary register usage is required.  */
+	  rtx tmp = gen_reg_rtx (V1TImode);
+	  src = gen_rtx_CONST_VECTOR (V1TImode, gen_rtvec (1, src));
+	  src = validize_mem (force_const_mem (V1TImode, src));
+	  emit_conversion_insns (gen_rtx_SET (dst, tmp), insn);
+	  dst = tmp;
+	}
+      break;
+
     case CONST_INT:
       switch (standard_sse_constant_p (src, TImode))
 	{
diff --git a/gcc/testsuite/gcc.target/i386/pieces-strcpy-1.c b/gcc/testsuite/gcc.target/i386/pieces-strcpy-1.c
new file mode 100644
index 00000000000..dd77a2cccda
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/pieces-strcpy-1.c
@@ -0,0 +1,19 @@
+/* { dg-do compile { target { ! ia32 } } } */
+/* { dg-options "-O2 -mno-avx -msse2 -mtune=generic" } */
+
+extern char *strcpy (char *, const char *);
+
+void
+foo (char *s)
+{
+  strcpy (s,
+	  "1234567890abcdef123456abcdef5678123456abcdef567abcdef678"
+	  "1234567");
+}
+
+/* { dg-final { scan-assembler-times "movdqa\[ \\t\]+\[^\n\]*%xmm" 4 } } */
+/* { dg-final { scan-assembler-times "movups\[ \\t\]+\[^\n\]*%xmm" 4 } } */
+/* No need to dynamically realign the stack here.  */
+/* { dg-final { scan-assembler-not "and\[^\n\r]*%\[re\]sp" } } */
+/* Nor use a frame pointer.  */
+/* { dg-final { scan-assembler-not "%\[re\]bp" } } */
diff --git a/gcc/testsuite/gcc.target/i386/pieces-strcpy-2.c b/gcc/testsuite/gcc.target/i386/pieces-strcpy-2.c
new file mode 100644
index 00000000000..f597e5d7c95
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/pieces-strcpy-2.c
@@ -0,0 +1,19 @@
+/* { dg-do compile { target { ! ia32 } } } */
+/* { dg-options "-O2 -mno-avx2 -mavx -mtune=sandybridge" } */
+
+extern char *strcpy (char *, const char *);
+
+void
+foo (char *s)
+{
+  strcpy (s,
+	  "1234567890abcdef123456abcdef5678123456abcdef567abcdef678"
+	  "1234567");
+}
+
+/* { dg-final { scan-assembler-times "vmovdqa\[ \\t\]+\[^\n\]*%xmm" 4 } } */
+/* { dg-final { scan-assembler-times "vmovups\[ \\t\]+\[^\n\]*%xmm" 4 } } */
+/* No need to dynamically realign the stack here.  */
+/* { dg-final { scan-assembler-not "and\[^\n\r]*%\[re\]sp" } } */
+/* Nor use a frame pointer.  */
+/* { dg-final { scan-assembler-not "%\[re\]bp" } } */