Add codegen support for NEON vst3 intrinsics with <1 x i64> vectors.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83518 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Bob Wilson 2009-10-08 00:28:28 +00:00
parent 24e04c535f
commit 5adf60c03b
4 changed files with 15 additions and 0 deletions

View File

@ -1616,6 +1616,7 @@ SDNode *ARMDAGToDAGISel::Select(SDValue Op) {
case MVT::v4i16: Opc = ARM::VST3d16; break;
case MVT::v2f32:
case MVT::v2i32: Opc = ARM::VST3d32; break;
case MVT::v1i64: Opc = ARM::VST3d64; break;
}
SDValue Chain = N->getOperand(0);
const SDValue Ops[] = { MemAddr, MemUpdate, MemOpc,

View File

@ -374,6 +374,10 @@ class VST3WB<bits<4> op7_4, string OpcodeStr>
def VST3d8 : VST3D<0b0000, "vst3.8">;
def VST3d16 : VST3D<0b0100, "vst3.16">;
def VST3d32 : VST3D<0b1000, "vst3.32">;
def VST3d64 : NLdSt<0,0b00,0b0110,0b1100, (outs),
(ins addrmode6:$addr, DPR:$src1, DPR:$src2, DPR:$src3),
IIC_VST,
"vst1.64\t\\{$src1,$src2,$src3\\}, $addr", "", []>;
// vst3 to double-spaced even registers.
def VST3q8a : VST3WB<0b0000, "vst3.8">;

View File

@ -143,6 +143,7 @@ static bool isNEONMultiRegOp(int Opcode, unsigned &FirstOpnd, unsigned &NumRegs,
case ARM::VST3d8:
case ARM::VST3d16:
case ARM::VST3d32:
case ARM::VST3d64:
case ARM::VST3LNd8:
case ARM::VST3LNd16:
case ARM::VST3LNd32:

View File

@ -32,6 +32,14 @@ define void @vst3f(float* %A, <2 x float>* %B) nounwind {
ret void
}
define void @vst3i64(i64* %A, <1 x i64>* %B) nounwind {
;CHECK: vst3i64:
;CHECK: vst1.64
%tmp1 = load <1 x i64>* %B
call void @llvm.arm.neon.vst3.v1i64(i64* %A, <1 x i64> %tmp1, <1 x i64> %tmp1, <1 x i64> %tmp1)
ret void
}
define void @vst3Qi8(i8* %A, <16 x i8>* %B) nounwind {
;CHECK: vst3Qi8:
;CHECK: vst3.8
@ -72,6 +80,7 @@ declare void @llvm.arm.neon.vst3.v8i8(i8*, <8 x i8>, <8 x i8>, <8 x i8>) nounwin
declare void @llvm.arm.neon.vst3.v4i16(i8*, <4 x i16>, <4 x i16>, <4 x i16>) nounwind
declare void @llvm.arm.neon.vst3.v2i32(i8*, <2 x i32>, <2 x i32>, <2 x i32>) nounwind
declare void @llvm.arm.neon.vst3.v2f32(i8*, <2 x float>, <2 x float>, <2 x float>) nounwind
declare void @llvm.arm.neon.vst3.v1i64(i8*, <1 x i64>, <1 x i64>, <1 x i64>) nounwind
declare void @llvm.arm.neon.vst3.v16i8(i8*, <16 x i8>, <16 x i8>, <16 x i8>) nounwind
declare void @llvm.arm.neon.vst3.v8i16(i8*, <8 x i16>, <8 x i16>, <8 x i16>) nounwind