Add codegen and encoding support for the immediate form of vbic.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118291 91177308-0d34-0410-b5e6-96231b3b80d8

5 files changed, 102 insertions, 6 deletions

diff --git a/lib/Target/ARM/ARMISelLowering.cpp b/lib/Target/ARM/ARMISelLowering.cpp
index 5a3520e..31bde5f 100644
--- a/lib/Target/ARM/ARMISelLowering.cpp
+++ b/lib/Target/ARM/ARMISelLowering.cpp
@@ -673,8 +673,10 @@ ARMTargetLowering::ARMTargetLowering(TargetMachine &TM)
   setTargetDAGCombine(ISD::SUB);
   setTargetDAGCombine(ISD::MUL);
 
-  if (Subtarget->hasV6T2Ops())
+  if (Subtarget->hasV6T2Ops() || Subtarget->hasNEON())
     setTargetDAGCombine(ISD::OR);
+  if (Subtarget->hasNEON())
+    setTargetDAGCombine(ISD::AND);
 
   setStackPointerRegisterToSaveRestore(ARM::SP);
 
@@ -4443,6 +4445,36 @@ static SDValue PerformMULCombine(SDNode *N,
   return SDValue();
 }
 
+static SDValue PerformANDCombine(SDNode *N,
+                                TargetLowering::DAGCombinerInfo &DCI) {
+  // Attempt to use immediate-form VBIC
+  BuildVectorSDNode *BVN = dyn_cast<BuildVectorSDNode>(N->getOperand(1));
+  DebugLoc dl = N->getDebugLoc();
+  EVT VT = N->getValueType(0);
+  SelectionDAG &DAG = DCI.DAG;
+  
+  APInt SplatBits, SplatUndef;
+  unsigned SplatBitSize;
+  bool HasAnyUndefs;
+  if (BVN &&
+      BVN->isConstantSplat(SplatBits, SplatUndef, SplatBitSize, HasAnyUndefs)) {
+    if (SplatBitSize <= 64) {
+      EVT VbicVT;
+      SDValue Val = isNEONModifiedImm((~SplatBits).getZExtValue(),
+                                      SplatUndef.getZExtValue(), SplatBitSize,
+                                      DAG, VbicVT, VT.is128BitVector(), false);
+      if (Val.getNode()) {
+        SDValue Input =
+          DAG.getNode(ISD::BIT_CONVERT, dl, VbicVT, N->getOperand(0));
+        SDValue Vbic = DAG.getNode(ARMISD::VBICIMM, dl, VbicVT, Input, Val);
+        return DAG.getNode(ISD::BIT_CONVERT, dl, VT, Vbic);
+      }
+    }
+  }
+  
+  return SDValue();
+}
+
 /// PerformORCombine - Target-specific dag combine xforms for ISD::OR
 static SDValue PerformORCombine(SDNode *N,
                                 TargetLowering::DAGCombinerInfo &DCI,
@@ -5066,6 +5098,7 @@ SDValue ARMTargetLowering::PerformDAGCombine(SDNode *N,
   case ISD::SUB:        return PerformSUBCombine(N, DCI);
   case ISD::MUL:        return PerformMULCombine(N, DCI, Subtarget);
   case ISD::OR:         return PerformORCombine(N, DCI, Subtarget);
+  case ISD::AND:        return PerformANDCombine(N, DCI);
   case ARMISD::VMOVRRD: return PerformVMOVRRDCombine(N, DCI);
   case ARMISD::VMOVDRR: return PerformVMOVDRRCombine(N, DCI.DAG);
   case ISD::BUILD_VECTOR: return PerformBUILD_VECTORCombine(N, DCI.DAG);
diff --git a/lib/Target/ARM/ARMISelLowering.h b/lib/Target/ARM/ARMISelLowering.h
index 4a7dec2..4e74193 100644
--- a/lib/Target/ARM/ARMISelLowering.h
+++ b/lib/Target/ARM/ARMISelLowering.h
@@ -165,7 +165,9 @@ namespace llvm {
       BFI,
       
       // Vector OR with immediate
-      VORRIMM
+      VORRIMM,
+      // Vector AND with NOT of immediate
+      VBICIMM
     };
   }
 
diff --git a/lib/Target/ARM/ARMInstrNEON.td b/lib/Target/ARM/ARMInstrNEON.td
index 29a1f2c..e64fefc 100644
--- a/lib/Target/ARM/ARMInstrNEON.td
+++ b/lib/Target/ARM/ARMInstrNEON.td
@@ -72,6 +72,7 @@ def NEONvmvnImm   : SDNode<"ARMISD::VMVNIMM", SDTARMVMOVIMM>;
 def SDTARMVORRIMM : SDTypeProfile<1, 2, [SDTCisVec<0>, SDTCisSameAs<0, 1>,
                                            SDTCisVT<2, i32>]>;
 def NEONvorrImm   : SDNode<"ARMISD::VORRIMM", SDTARMVORRIMM>;
+def NEONvbicImm   : SDNode<"ARMISD::VBICIMM", SDTARMVORRIMM>;
 
 def NEONvdup      : SDNode<"ARMISD::VDUP", SDTypeProfile<1, 1, [SDTCisVec<0>]>>;
 
@@ -3308,13 +3309,13 @@ def VORRiv4i16 : N1ModImm<1, 0b000, {1,0,?,1}, 0, 0, 0, 1,
   let Inst{9} = SIMM{9};
 }
 
-def VORRiv2i32 : N1ModImm<1, 0b000, {?,?,?,1}, 0, 0, 0, 1,
+def VORRiv2i32 : N1ModImm<1, 0b000, {0,?,?,1}, 0, 0, 0, 1,
                           (outs DPR:$Vd), (ins nModImm:$SIMM, DPR:$src),
                           IIC_VMOVImm,
                           "vorr", "i32", "$Vd, $SIMM", "$src = $Vd",
                           [(set DPR:$Vd,
                             (v2i32 (NEONvorrImm DPR:$src, timm:$SIMM)))]> {
-  let Inst{11-9} = SIMM{11-9};
+  let Inst{10-9} = SIMM{10-9};
 }
 
 def VORRiv8i16 : N1ModImm<1, 0b000, {1,0,?,1}, 0, 1, 0, 1,
@@ -3326,13 +3327,13 @@ def VORRiv8i16 : N1ModImm<1, 0b000, {1,0,?,1}, 0, 1, 0, 1,
   let Inst{9} = SIMM{9};
 }
 
-def VORRiv4i32 : N1ModImm<1, 0b000, {?,?,?,1}, 0, 1, 0, 1,
+def VORRiv4i32 : N1ModImm<1, 0b000, {0,?,?,1}, 0, 1, 0, 1,
                           (outs QPR:$Vd), (ins nModImm:$SIMM, QPR:$src),
                           IIC_VMOVImm,
                           "vorr", "i32", "$Vd, $SIMM", "$src = $Vd",
                           [(set QPR:$Vd,
                             (v4i32 (NEONvorrImm QPR:$src, timm:$SIMM)))]> {
-  let Inst{11-9} = SIMM{11-9};
+  let Inst{10-9} = SIMM{10-9};
 }
 
 
@@ -3348,6 +3349,42 @@ def  VBICq    : N3VX<0, 0, 0b01, 0b0001, 1, 1, (outs QPR:$dst),
                      [(set QPR:$dst, (v4i32 (and QPR:$src1,
                                                  (vnotq QPR:$src2))))]>;
 
+def VBICiv4i16 : N1ModImm<1, 0b000, {1,0,?,1}, 0, 0, 1, 1,
+                          (outs DPR:$Vd), (ins nModImm:$SIMM, DPR:$src),
+                          IIC_VMOVImm,
+                          "vbic", "i16", "$Vd, $SIMM", "$src = $Vd",
+                          [(set DPR:$Vd,
+                            (v4i16 (NEONvbicImm DPR:$src, timm:$SIMM)))]> {
+  let Inst{9} = SIMM{9};
+}
+
+def VBICiv2i32 : N1ModImm<1, 0b000, {0,?,?,1}, 0, 0, 1, 1,
+                          (outs DPR:$Vd), (ins nModImm:$SIMM, DPR:$src),
+                          IIC_VMOVImm,
+                          "vbic", "i32", "$Vd, $SIMM", "$src = $Vd",
+                          [(set DPR:$Vd,
+                            (v2i32 (NEONvbicImm DPR:$src, timm:$SIMM)))]> {
+  let Inst{10-9} = SIMM{10-9};
+}
+
+def VBICiv8i16 : N1ModImm<1, 0b000, {1,0,?,1}, 0, 1, 1, 1,
+                          (outs QPR:$Vd), (ins nModImm:$SIMM, QPR:$src),
+                          IIC_VMOVImm,
+                          "vbic", "i16", "$Vd, $SIMM", "$src = $Vd",
+                          [(set QPR:$Vd,
+                            (v8i16 (NEONvbicImm QPR:$src, timm:$SIMM)))]> {
+  let Inst{9} = SIMM{9};
+}
+
+def VBICiv4i32 : N1ModImm<1, 0b000, {0,?,?,1}, 0, 1, 1, 1,
+                          (outs QPR:$Vd), (ins nModImm:$SIMM, QPR:$src),
+                          IIC_VMOVImm,
+                          "vbic", "i32", "$Vd, $SIMM", "$src = $Vd",
+                          [(set QPR:$Vd,
+                            (v4i32 (NEONvbicImm QPR:$src, timm:$SIMM)))]> {
+  let Inst{10-9} = SIMM{10-9};
+}
+
 //   VORN     : Vector Bitwise OR NOT
 def  VORNd    : N3VX<0, 0, 0b11, 0b0001, 0, 1, (outs DPR:$dst),
                      (ins DPR:$src1, DPR:$src2), N3RegFrm, IIC_VBINiD,
diff --git a/test/CodeGen/ARM/vbits.ll b/test/CodeGen/ARM/vbits.ll
index b6077c0..880495d 100644
--- a/test/CodeGen/ARM/vbits.ll
+++ b/test/CodeGen/ARM/vbits.ll
@@ -525,3 +525,23 @@ define <16 x i8> @v_orrimmQ(<16 x i8>* %A) nounwind {
 	%tmp3 = or <16 x i8> %tmp1, <i8 0, i8 0, i8 0, i8 1, i8 0, i8 0, i8 0, i8 1, i8 0, i8 0, i8 0, i8 1, i8 0, i8 0, i8 0, i8 1>
 	ret <16 x i8> %tmp3
 }
+
+define <8 x i8> @v_bicimm(<8 x i8>* %A) nounwind {
+; CHECK: v_bicimm:
+; CHECK-NOT: vmov
+; CHECK-NOT vmvn
+; CHECK: vbic
+	%tmp1 = load <8 x i8>* %A
+	%tmp3 = and <8 x i8> %tmp1, < i8 -1, i8 -1, i8 -1, i8 0, i8 -1, i8 -1, i8 -1, i8 0 >
+	ret <8 x i8> %tmp3
+}
+
+define <16 x i8> @v_bicimmQ(<16 x i8>* %A) nounwind {
+; CHECK: v_bicimmQ:
+; CHECK-NOT: vmov
+; CHECK-NOT: vmvn
+; CHECK: vbic
+	%tmp1 = load <16 x i8>* %A
+	%tmp3 = and <16 x i8> %tmp1, < i8 -1, i8 -1, i8 -1, i8 0, i8 -1, i8 -1, i8 -1, i8 0, i8 -1, i8 -1, i8 -1, i8 0, i8 -1, i8 -1, i8 -1, i8 0 >
+	ret <16 x i8> %tmp3
+}
diff --git a/test/MC/ARM/neon-bitwise-encoding.s b/test/MC/ARM/neon-bitwise-encoding.s
index 43a73ff..7a4c8fd 100644
--- a/test/MC/ARM/neon-bitwise-encoding.s
+++ b/test/MC/ARM/neon-bitwise-encoding.s
@@ -26,6 +26,10 @@
 	vbic	d16, d17, d16
 @ CHECK: vbic	q8, q8, q9              @ encoding: [0xf2,0x01,0x50,0xf2]
 	vbic	q8, q8, q9
+@ CHECK: vbic.i32	d16, #0xFF000000 @ encoding: [0x3f,0x07,0xc7,0xf3]
+  vbic.i32	d16, #0xFF000000
+@ CHECK: vbic.i32	q8, #0xFF000000 @ encoding: [0x7f,0x07,0xc7,0xf3]
+  vbic.i32	q8, #0xFF000000
 
 @ CHECK: vorn	d16, d17, d16           @ encoding: [0xb0,0x01,0x71,0xf2]
 	vorn	d16, d17, d16