ARM: implement @llvm.readcyclecounter intrinsic

This implements the @llvm.readcyclecounter intrinsic as the specific
MRC instruction specified in the ARM manuals for CPUs with the Power
Management extensions.

Older CPUs had slightly different methods which may also have to be
implemented eventually, but this should cover all v7 cases.

rdar://problem/13939186

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@182603 91177308-0d34-0410-b5e6-96231b3b80d8

1 files changed, 43 insertions, 1 deletions

diff --git a/lib/Target/ARM/ARMISelLowering.cpp b/lib/Target/ARM/ARMISelLowering.cpp
index a1443d1b02..4238c72c13 100644
--- a/lib/Target/ARM/ARMISelLowering.cpp
+++ b/lib/Target/ARM/ARMISelLowering.cpp
@@ -681,6 +681,8 @@ ARMTargetLowering::ARMTargetLowering(TargetMachine &TM)
   setOperationAction(ISD::CTTZ_ZERO_UNDEF  , MVT::i32  , Expand);
   setOperationAction(ISD::CTLZ_ZERO_UNDEF  , MVT::i32  , Expand);
 
+  setOperationAction(ISD::READCYCLECOUNTER, MVT::i64, Custom);
+
   // Only ARMv6 has BSWAP.
   if (!Subtarget->hasV6Ops())
     setOperationAction(ISD::BSWAP, MVT::i32, Expand);
@@ -5702,7 +5704,6 @@ static SDValue LowerAtomicLoadStore(SDValue Op, SelectionDAG &DAG) {
   return SDValue();
 }
 
-
 static void
 ReplaceATOMIC_OP_64(SDNode *Node, SmallVectorImpl<SDValue>& Results,
                     SelectionDAG &DAG, unsigned NewOp) {
@@ -5736,6 +5737,44 @@ ReplaceATOMIC_OP_64(SDNode *Node, SmallVectorImpl<SDValue>& Results,
   Results.push_back(Result.getValue(2));
 }
 
+static void ReplaceREADCYCLECOUNTER(SDNode *N,
+                                    SmallVectorImpl<SDValue> &Results,
+                                    SelectionDAG &DAG,
+                                    const ARMSubtarget *Subtarget) {
+  DebugLoc DL = N->getDebugLoc();
+  SDValue Cycles32, OutChain;
+
+  if (Subtarget->hasPerfMon()) {
+    // Under Power Management extensions, the cycle-count is:
+    //    mrc p15, #0, <Rt>, c9, c13, #0
+    SDValue Ops[] = { N->getOperand(0), // Chain
+                      DAG.getConstant(Intrinsic::arm_mrc, MVT::i32),
+                      DAG.getConstant(15, MVT::i32),
+                      DAG.getConstant(0, MVT::i32),
+                      DAG.getConstant(9, MVT::i32),
+                      DAG.getConstant(13, MVT::i32),
+                      DAG.getConstant(0, MVT::i32)
+    };
+
+    Cycles32 = DAG.getNode(ISD::INTRINSIC_W_CHAIN, DL,
+                           DAG.getVTList(MVT::i32, MVT::Other), &Ops[0],
+                           array_lengthof(Ops));
+    OutChain = Cycles32.getValue(1);
+  } else {
+    // Intrinsic is defined to return 0 on unsupported platforms. Technically
+    // there are older ARM CPUs that have implementation-specific ways of
+    // obtaining this information (FIXME!).
+    Cycles32 = DAG.getConstant(0, MVT::i32);
+    OutChain = DAG.getEntryNode();
+  }
+
+
+  SDValue Cycles64 = DAG.getNode(ISD::BUILD_PAIR, DL, MVT::i64,
+                                 Cycles32, DAG.getConstant(0, MVT::i32));
+  Results.push_back(Cycles64);
+  Results.push_back(OutChain);
+}
+
 SDValue ARMTargetLowering::LowerOperation(SDValue Op, SelectionDAG &DAG) const {
   switch (Op.getOpcode()) {
   default: llvm_unreachable("Don't know how to custom lower this!");
@@ -5813,6 +5852,9 @@ void ARMTargetLowering::ReplaceNodeResults(SDNode *N,
   case ISD::SRA:
     Res = Expand64BitShift(N, DAG, Subtarget);
     break;
+  case ISD::READCYCLECOUNTER:
+    ReplaceREADCYCLECOUNTER(N, Results, DAG, Subtarget);
+    return;
   case ISD::ATOMIC_LOAD_ADD:
     ReplaceATOMIC_OP_64(N, Results, DAG, ARMISD::ATOMADD64_DAG);
     return;