R600: Use KCache for kernel arguments

Reviewed-by: Vincent Lejeune <vljn at ovi.com>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@186918 91177308-0d34-0410-b5e6-96231b3b80d8

4 files changed, 22 insertions, 49 deletions

diff --git a/lib/Target/R600/AMDGPU.h b/lib/Target/R600/AMDGPU.h
index 7621422a1f..fbf1fcee58 100644
--- a/lib/Target/R600/AMDGPU.h
+++ b/lib/Target/R600/AMDGPU.h
@@ -74,6 +74,12 @@ enum AddressSpaces {
   ADDRESS_NONE     = 5, ///< Address space for unknown memory.
   PARAM_D_ADDRESS  = 6, ///< Address space for direct addressible parameter memory (CONST0)
   PARAM_I_ADDRESS  = 7, ///< Address space for indirect addressible parameter memory (VTX1)
+
+  // Do not re-order the CONSTANT_BUFFER_* enums.  Several places depend on this
+  // order to be able to dynamically index a constant buffer, for example:
+  //
+  // ConstantBufferAS = CONSTANT_BUFFER_0 + CBIdx
+
   CONSTANT_BUFFER_0 = 8,
   CONSTANT_BUFFER_1 = 9,
   CONSTANT_BUFFER_2 = 10,
diff --git a/lib/Target/R600/AMDILISelDAGToDAG.cpp b/lib/Target/R600/AMDILISelDAGToDAG.cpp
index e79ab3c0db..78036a8a14 100644
--- a/lib/Target/R600/AMDILISelDAGToDAG.cpp
+++ b/lib/Target/R600/AMDILISelDAGToDAG.cpp
@@ -565,24 +565,11 @@ bool AMDGPUDAGToDAGISel::isRegionStore(const StoreSDNode *N) {
   return checkType(N->getSrcValue(), AMDGPUAS::REGION_ADDRESS);
 }
 
-bool AMDGPUDAGToDAGISel::isConstantLoad(const LoadSDNode *N, int cbID) const {
-  if (checkType(N->getSrcValue(), AMDGPUAS::CONSTANT_ADDRESS)) {
-    return true;
-  }
-
-  const DataLayout *DL = TM.getDataLayout();
-  MachineMemOperand *MMO = N->getMemOperand();
-  const Value *V = MMO->getValue();
-  const Value *BV = GetUnderlyingObject(V, DL, 0);
-  if (MMO
-      && MMO->getValue()
-      && ((V && dyn_cast<GlobalValue>(V))
-          || (BV && dyn_cast<GlobalValue>(
-                GetUnderlyingObject(MMO->getValue(), DL, 0))))) {
-    return checkType(N->getSrcValue(), AMDGPUAS::PRIVATE_ADDRESS);
-  } else {
-    return false;
+bool AMDGPUDAGToDAGISel::isConstantLoad(const LoadSDNode *N, int CbId) const {
+  if (CbId == -1) {
+    return checkType(N->getSrcValue(), AMDGPUAS::CONSTANT_ADDRESS);
   }
+  return checkType(N->getSrcValue(), AMDGPUAS::CONSTANT_BUFFER_0 + CbId);
 }
 
 bool AMDGPUDAGToDAGISel::isGlobalLoad(const LoadSDNode *N) const {
diff --git a/lib/Target/R600/R600ISelLowering.cpp b/lib/Target/R600/R600ISelLowering.cpp
index 7f93f23f28..dd613d56a6 100644
--- a/lib/Target/R600/R600ISelLowering.cpp
+++ b/lib/Target/R600/R600ISelLowering.cpp
@@ -72,10 +72,10 @@ R600TargetLowering::R600TargetLowering(TargetMachine &TM) :
   setOperationAction(ISD::LOAD, MVT::i32, Custom);
   setOperationAction(ISD::LOAD, MVT::v2i32, Expand);
   setOperationAction(ISD::LOAD, MVT::v4i32, Custom);
-  setLoadExtAction(ISD::SEXTLOAD, MVT::i8, Expand);
-  setLoadExtAction(ISD::SEXTLOAD, MVT::i16, Expand);
-  setLoadExtAction(ISD::ZEXTLOAD, MVT::i8, Expand);
-  setLoadExtAction(ISD::ZEXTLOAD, MVT::i16, Expand);
+  setLoadExtAction(ISD::SEXTLOAD, MVT::i8, Custom);
+  setLoadExtAction(ISD::SEXTLOAD, MVT::i16, Custom);
+  setLoadExtAction(ISD::ZEXTLOAD, MVT::i8, Custom);
+  setLoadExtAction(ISD::ZEXTLOAD, MVT::i16, Custom);
   setOperationAction(ISD::STORE, MVT::i8, Custom);
   setOperationAction(ISD::STORE, MVT::i32, Custom);
   setOperationAction(ISD::STORE, MVT::v2i32, Expand);
@@ -775,7 +775,7 @@ SDValue R600TargetLowering::LowerImplicitParameter(SelectionDAG &DAG, EVT VT,
                                                    unsigned DwordOffset) const {
   unsigned ByteOffset = DwordOffset * 4;
   PointerType * PtrType = PointerType::get(VT.getTypeForEVT(*DAG.getContext()),
-                                      AMDGPUAS::PARAM_I_ADDRESS);
+                                      AMDGPUAS::CONSTANT_BUFFER_0);
 
   // We shouldn't be using an offset wider than 16-bits for implicit parameters.
   assert(isInt<16>(ByteOffset));
@@ -1219,40 +1219,20 @@ SDValue R600TargetLowering::LowerFormalArguments(
 
   AnalyzeFormalArguments(CCInfo, Ins);
 
-  Function::const_arg_iterator FuncArg =
-                            DAG.getMachineFunction().getFunction()->arg_begin();
-  for (unsigned i = 0, e = Ins.size(); i < e; ++i, ++FuncArg) {
+  for (unsigned i = 0, e = Ins.size(); i < e; ++i) {
     CCValAssign &VA = ArgLocs[i];
     EVT VT = VA.getLocVT();
-    Type *ArgType = FuncArg->getType();
-    unsigned ArgSizeInBits = ArgType->isPointerTy() ?
-                             32 : ArgType->getPrimitiveSizeInBits();
-    unsigned ArgBytes = ArgSizeInBits >> 3;
-    EVT ArgVT;
-    if (ArgSizeInBits < VT.getSizeInBits()) {
-      assert(!ArgType->isFloatTy() &&
-             "Extending floating point arguments not supported yet");
-      ArgVT = MVT::getIntegerVT(ArgSizeInBits);
-    } else {
-      ArgVT = VT;
-    }
-
-    ISD::LoadExtType LoadType = ISD::EXTLOAD;
-    if (Ins[i].Flags.isZExt()) {
-      LoadType = ISD::ZEXTLOAD;
-    } else if (Ins[i].Flags.isSExt()) {
-      LoadType = ISD::SEXTLOAD;
-    }
 
     PointerType *PtrTy = PointerType::get(VT.getTypeForEVT(*DAG.getContext()),
-                                                    AMDGPUAS::PARAM_I_ADDRESS);
+                                                   AMDGPUAS::CONSTANT_BUFFER_0);
 
     // The first 36 bytes of the input buffer contains information about
     // thread group and global sizes.
-    SDValue Arg = DAG.getExtLoad(LoadType, DL, VT, DAG.getRoot(),
+    SDValue Arg = DAG.getLoad(VT, DL, Chain,
                            DAG.getConstant(36 + VA.getLocMemOffset(), MVT::i32),
-                           MachinePointerInfo(UndefValue::get(PtrTy)),
-                           ArgVT, false, false, ArgBytes);
+                           MachinePointerInfo(UndefValue::get(PtrTy)), false,
+                           false, false, 4); // 4 is the prefered alignment for
+                                             // the CONSTANT memory space.
     InVals.push_back(Arg);
   }
   return Chain;
diff --git a/lib/Target/R600/R600Instructions.td b/lib/Target/R600/R600Instructions.td
index 8a49a8df1d..632cbcf48a 100644
--- a/lib/Target/R600/R600Instructions.td
+++ b/lib/Target/R600/R600Instructions.td
@@ -313,7 +313,7 @@ class VTX_READ <string name, bits<8> buffer_id, dag outs, list<dag> pattern>
 
 class LoadParamFrag <PatFrag load_type> : PatFrag <
   (ops node:$ptr), (load_type node:$ptr),
-  [{ return isParamLoad(dyn_cast<LoadSDNode>(N)); }]
+  [{ return isConstantLoad(dyn_cast<LoadSDNode>(N), 0); }]
 >;
 
 def load_param : LoadParamFrag<load>;