It was missing from the VLD1/VST1 handling logic, even though the
corresponding instructions exist (same form as v2i64).
In preparation for a future patch.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@223832
91177308-0d34-0410-b5e6-
96231b3b80d8
case MVT::v8i16: OpcodeIndex = 1; break;
case MVT::v4f32:
case MVT::v4i32: OpcodeIndex = 2; break;
+ case MVT::v2f64:
case MVT::v2i64: OpcodeIndex = 3;
assert(NumVecs == 1 && "v2i64 type only supported for VLD1");
break;
case MVT::v8i16: OpcodeIndex = 1; break;
case MVT::v4f32:
case MVT::v4i32: OpcodeIndex = 2; break;
+ case MVT::v2f64:
case MVT::v2i64: OpcodeIndex = 3;
assert(NumVecs == 1 && "v2i64 type only supported for VST1");
break;
ret <2 x i64> %tmp1
}
+define <2 x double> @vld1Qf64(double* %A) nounwind {
+;CHECK-LABEL: vld1Qf64:
+;CHECK: vld1.64
+ %tmp0 = bitcast double* %A to i8*
+ %tmp1 = call <2 x double> @llvm.arm.neon.vld1.v2f64(i8* %tmp0, i32 1)
+ ret <2 x double> %tmp1
+}
+
declare <8 x i8> @llvm.arm.neon.vld1.v8i8(i8*, i32) nounwind readonly
declare <4 x i16> @llvm.arm.neon.vld1.v4i16(i8*, i32) nounwind readonly
declare <2 x i32> @llvm.arm.neon.vld1.v2i32(i8*, i32) nounwind readonly
declare <4 x i32> @llvm.arm.neon.vld1.v4i32(i8*, i32) nounwind readonly
declare <4 x float> @llvm.arm.neon.vld1.v4f32(i8*, i32) nounwind readonly
declare <2 x i64> @llvm.arm.neon.vld1.v2i64(i8*, i32) nounwind readonly
+declare <2 x double> @llvm.arm.neon.vld1.v2f64(i8*, i32) nounwind readonly
; Radar 8355607
; Do not crash if the vld1 result is not used.
ret void
}
+define void @vst1Qf64(double* %A, <2 x double>* %B) nounwind {
+;CHECK-LABEL: vst1Qf64:
+;CHECK: vst1.64
+ %tmp0 = bitcast double* %A to i8*
+ %tmp1 = load <2 x double>* %B
+ call void @llvm.arm.neon.vst1.v2f64(i8* %tmp0, <2 x double> %tmp1, i32 1)
+ ret void
+}
+
declare void @llvm.arm.neon.vst1.v8i8(i8*, <8 x i8>, i32) nounwind
declare void @llvm.arm.neon.vst1.v4i16(i8*, <4 x i16>, i32) nounwind
declare void @llvm.arm.neon.vst1.v2i32(i8*, <2 x i32>, i32) nounwind
declare void @llvm.arm.neon.vst1.v4i32(i8*, <4 x i32>, i32) nounwind
declare void @llvm.arm.neon.vst1.v4f32(i8*, <4 x float>, i32) nounwind
declare void @llvm.arm.neon.vst1.v2i64(i8*, <2 x i64>, i32) nounwind
+declare void @llvm.arm.neon.vst1.v2f64(i8*, <2 x double>, i32) nounwind