i386: Optimize code generation of __mm256_zextsi128_si256(__mm_set1_epi8(-1))

gcc/ChangeLog:

	PR target/94962
	* config/i386/constraints.md (BH): New define_constraint.
	* config/i386/i386.cc (standard_sse_constant_p): Add return
	3/4 when operand matches new predicate.
	(standard_sse_constant_opcode): Add new alternative branch to
	return "vpcmpeqd".
	* config/i386/predicates.md
	(vector_all_ones_zero_extend_half_operand): New define_predicate.
	(vector_all_ones_zero_extend_quarter_operand): Ditto.
	* config/i386/sse.md: Add constraint to insn "mov<mode>_internal".

gcc/testsuite/ChangeLog:

	PR target/94962
	* gcc.target/i386/avx256-unaligned-load-1.c: Modify test.
	* gcc.target/i386/avx256-unaligned-store-1.c: Ditto.
	* gcc.target/i386/avx256-unaligned-store-2.c: Ditto.
	* gcc.target/i386/avx256-unaligned-store-3.c: Ditto.
	* gcc.target/i386/pr94962-1.c: New test.
	* gcc.target/i386/pr94962-2.c: Ditto.
	* gcc.target/i386/pr94962-3.c: Ditto.
	* gcc.target/i386/pr94962-4.c: Ditto.

4 files changed, 86 insertions, 5 deletions

diff --git a/gcc/config/i386/constraints.md b/gcc/config/i386/constraints.md
index 7361687..95b2b14 100644
--- a/gcc/config/i386/constraints.md
+++ b/gcc/config/i386/constraints.md
@@ -168,6 +168,9 @@
 ;;  z  Constant call address operand.
 ;;  C  Integer SSE constant with all bits set operand.
 ;;  F  Floating-point SSE constant with all bits set operand.
+;;  H  Integer SSE constant that is 128/256bit all ones
+;;     and zero-extand to 256/512bit, or 128bit all ones
+;;     and zero-extend to 512bit.
 ;;  M  x86-64 memory operand.
 
 (define_constraint "Bf"
@@ -233,6 +236,11 @@
   (and (match_test "TARGET_SSE")
        (match_operand 0 "float_vector_all_ones_operand")))
 
+(define_constraint "BH"
+  "@internal integer constant with last half/quarter bits set operand."
+  (ior (match_operand 0 "vector_all_ones_zero_extend_half_operand")
+       (match_operand 0 "vector_all_ones_zero_extend_quarter_operand")))
+
 ;; NB: Similar to 'm', but don't use define_memory_constraint on x86-64
 ;; to prevent LRA from converting the operand to the form '(mem (reg X))'
 ;; where X is a base register.
diff --git a/gcc/config/i386/i386.cc b/gcc/config/i386/i386.cc
index dadf453..ca799da 100644
--- a/gcc/config/i386/i386.cc
+++ b/gcc/config/i386/i386.cc
@@ -5186,7 +5186,8 @@ standard_80387_constant_rtx (int idx)
 				       XFmode);
 }
 
-/* Return 1 if X is all bits 0 and 2 if X is all bits 1
+/* Return 1 if X is all bits 0, 2 if X is all bits 1
+   and 3 if X is all bits 1 with zero extend
    in supported SSE/AVX vector mode.  */
 
 int
@@ -5234,6 +5235,10 @@ standard_sse_constant_p (rtx x, machine_mode pred_mode)
 	}
     }
 
+  if (vector_all_ones_zero_extend_half_operand (x, mode)
+      || vector_all_ones_zero_extend_quarter_operand (x, mode))
+    return 3;
+
   return 0;
 }
 
@@ -5341,6 +5346,25 @@ standard_sse_constant_opcode (rtx_insn *insn, rtx *operands)
 	  gcc_unreachable ();
 	}
    }
+  else if (vector_all_ones_zero_extend_half_operand (x, mode))
+    {
+      if (GET_MODE_SIZE (mode) == 64)
+	{
+	  gcc_assert (TARGET_AVX512F);
+	  return "vpcmpeqd \t %t0, %t0, %t0";
+	}
+      else if (GET_MODE_SIZE (mode) == 32)
+	{
+	  gcc_assert (TARGET_AVX);
+	  return "vpcmpeqd \t %x0, %x0, %x0";
+	}
+      gcc_unreachable ();
+    }
+  else if (vector_all_ones_zero_extend_quarter_operand (x, mode))
+    {
+      gcc_assert (TARGET_AVX512F);
+      return "vpcmpeqd \t %x0, %x0, %x0";
+    }
 
   gcc_unreachable ();
 }
diff --git a/gcc/config/i386/predicates.md b/gcc/config/i386/predicates.md
index 4f16bb7..655eabf 100644
--- a/gcc/config/i386/predicates.md
+++ b/gcc/config/i386/predicates.md
@@ -1159,6 +1159,55 @@
        (match_test "INTEGRAL_MODE_P (GET_MODE (op))")
        (match_test "op == CONSTM1_RTX (GET_MODE (op))")))
 
+/* Return true if operand is an 128/256bit all ones vector
+   that zero-extends to 256/512bit.  */
+(define_predicate "vector_all_ones_zero_extend_half_operand"
+  (match_code "const_vector")
+{
+  mode = GET_MODE (op);
+  if (GET_MODE_CLASS (mode) != MODE_VECTOR_INT
+      || (GET_MODE_SIZE (mode) != 32
+	  && GET_MODE_SIZE (mode) != 64))
+    return false;
+
+  int nelts = CONST_VECTOR_NUNITS (op);
+  for (int i = 0; i != nelts; i++)
+    {
+      rtx elt = CONST_VECTOR_ELT (op, i);
+      if (i < nelts / 2
+	  && elt != CONSTM1_RTX (GET_MODE_INNER (mode)))
+	return false;
+      if (i >= nelts / 2
+	  && elt != CONST0_RTX (GET_MODE_INNER (mode)))
+	return false;
+    }
+  return true;
+})
+
+/* Return true if operand is an 128bit all ones vector
+   that zero extends to 512bit.  */
+(define_predicate "vector_all_ones_zero_extend_quarter_operand"
+  (match_code "const_vector")
+{
+  mode = GET_MODE (op);
+  if (GET_MODE_CLASS (mode) != MODE_VECTOR_INT
+      || GET_MODE_SIZE (mode) != 64)
+    return false;
+
+  int nelts = CONST_VECTOR_NUNITS (op);
+  for (int i = 0; i != nelts; i++)
+    {
+      rtx elt = CONST_VECTOR_ELT (op, i);
+      if (i < nelts / 4
+	  && elt != CONSTM1_RTX (GET_MODE_INNER (mode)))
+	return false;
+      if (i >= nelts / 4
+	  && elt != CONST0_RTX (GET_MODE_INNER (mode)))
+	return false;
+    }
+  return true;
+})
+
 ; Return true when OP is operand acceptable for vector memory operand.
 ; Only AVX can have misaligned memory operand.
 (define_predicate "vector_memory_operand"
diff --git a/gcc/config/i386/sse.md b/gcc/config/i386/sse.md
index b60c0d3..5c18963 100644
--- a/gcc/config/i386/sse.md
+++ b/gcc/config/i386/sse.md
@@ -1323,9 +1323,9 @@
 
 (define_insn "mov<mode>_internal"
   [(set (match_operand:VMOVE 0 "nonimmediate_operand"
-	 "=v,v ,v ,m")
+	 "=v,v ,v,v ,m")
 	(match_operand:VMOVE 1 "nonimmediate_or_sse_const_operand"
-	 " C,<sseconstm1>,vm,v"))]
+	 " C,<sseconstm1>,BH,vm,v"))]
   "TARGET_SSE
    && (register_operand (operands[0], <MODE>mode)
        || register_operand (operands[1], <MODE>mode))
@@ -1343,7 +1343,7 @@
       gcc_unreachable ();
     }
 }
-  [(set_attr "type" "sselog1,sselog1,ssemov,ssemov")
+  [(set_attr "type" "sselog1,sselog1,sselog1,ssemov,ssemov")
    (set_attr "prefix" "maybe_vex")
    (set (attr "mode")
 	(cond [(match_test "TARGET_AVX")
@@ -1354,7 +1354,7 @@
 	       (and (match_test "<MODE>mode == V2DFmode")
 		    (match_test "TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL"))
 		 (const_string "V4SF")
-	       (and (eq_attr "alternative" "3")
+	       (and (eq_attr "alternative" "4")
 		    (match_test "TARGET_SSE_TYPELESS_STORES"))
 		 (const_string "V4SF")
 	       (and (eq_attr "alternative" "0")