1 files changed, 41 insertions, 12 deletions

diff --git a/lib/Target/R600/SIISelLowering.cpp b/lib/Target/R600/SIISelLowering.cpp
index 7d794b8..bd0c3c2 100644
--- a/lib/Target/R600/SIISelLowering.cpp
+++ b/lib/Target/R600/SIISelLowering.cpp
@@ -172,16 +172,12 @@ SITargetLowering::SITargetLowering(TargetMachine &TM,
   setOperationAction(ISD::UDIV, MVT::i64, Expand);
   setOperationAction(ISD::UREM, MVT::i64, Expand);
 
-  // We only support LOAD/STORE and vector manipulation ops for vectors
-  // with > 4 elements.
-  MVT VecTypes[] = {
-    MVT::v8i32, MVT::v8f32, MVT::v16i32, MVT::v16f32
-  };
-
   setOperationAction(ISD::SELECT_CC, MVT::i1, Expand);
   setOperationAction(ISD::SELECT, MVT::i1, Promote);
 
-  for (MVT VT : VecTypes) {
+  // We only support LOAD/STORE and vector manipulation ops for vectors
+  // with > 4 elements.
+  for (MVT VT : {MVT::v8i32, MVT::v8f32, MVT::v16i32, MVT::v16f32}) {
     for (unsigned Op = 0; Op < ISD::BUILTIN_OP_END; ++Op) {
       switch(Op) {
       case ISD::LOAD:
@@ -206,10 +202,10 @@ SITargetLowering::SITargetLowering(TargetMachine &TM,
   if (Subtarget->getGeneration() >= AMDGPUSubtarget::SEA_ISLANDS) {
     setOperationAction(ISD::FTRUNC, MVT::f64, Legal);
     setOperationAction(ISD::FCEIL, MVT::f64, Legal);
-    setOperationAction(ISD::FFLOOR, MVT::f64, Legal);
     setOperationAction(ISD::FRINT, MVT::f64, Legal);
   }
 
+  setOperationAction(ISD::FFLOOR, MVT::f64, Legal);
   setOperationAction(ISD::FDIV, MVT::f32, Custom);
   setOperationAction(ISD::FDIV, MVT::f64, Custom);
 
@@ -932,6 +928,12 @@ SDValue SITargetLowering::LowerINTRINSIC_WO_CHAIN(SDValue Op,
                        Op.getOperand(1),
                        Op.getOperand(2),
                        Op.getOperand(3));
+
+  case AMDGPUIntrinsic::AMDGPU_fract:
+  case AMDGPUIntrinsic::AMDIL_fraction: // Legacy name.
+    return DAG.getNode(ISD::FSUB, DL, VT, Op.getOperand(1),
+                       DAG.getNode(ISD::FFLOOR, DL, VT, Op.getOperand(1)));
+
   default:
     return AMDGPUTargetLowering::LowerOperation(Op, DAG);
   }
@@ -1346,6 +1348,35 @@ SDValue SITargetLowering::performUCharToFloatCombine(SDNode *N,
   return SDValue();
 }
 
+/// \brief Return true if the given offset Size in bytes can be folded into
+/// the immediate offsets of a memory instruction for the given address space.
+static bool canFoldOffset(unsigned OffsetSize, unsigned AS,
+                          const AMDGPUSubtarget &STI) {
+  switch (AS) {
+  case AMDGPUAS::GLOBAL_ADDRESS: {
+    // MUBUF instructions a 12-bit offset in bytes.
+    return isUInt<12>(OffsetSize);
+  }
+  case AMDGPUAS::CONSTANT_ADDRESS: {
+    // SMRD instructions have an 8-bit offset in dwords on SI and
+    // a 20-bit offset in bytes on VI.
+    if (STI.getGeneration() >= AMDGPUSubtarget::VOLCANIC_ISLANDS)
+      return isUInt<20>(OffsetSize);
+    else
+      return (OffsetSize % 4 == 0) && isUInt<8>(OffsetSize / 4);
+  }
+  case AMDGPUAS::LOCAL_ADDRESS:
+  case AMDGPUAS::REGION_ADDRESS: {
+    // The single offset versions have a 16-bit offset in bytes.
+    return isUInt<16>(OffsetSize);
+  }
+  case AMDGPUAS::PRIVATE_ADDRESS:
+  // Indirect register addressing does not use any offsets.
+  default:
+    return 0;
+  }
+}
+
 // (shl (add x, c1), c2) -> add (shl x, c2), (shl c1, c2)
 
 // This is a variant of
@@ -1377,13 +1408,10 @@ SDValue SITargetLowering::performSHLPtrCombine(SDNode *N,
   if (!CAdd)
     return SDValue();
 
-  const SIInstrInfo *TII =
-      static_cast<const SIInstrInfo *>(Subtarget->getInstrInfo());
-
   // If the resulting offset is too large, we can't fold it into the addressing
   // mode offset.
   APInt Offset = CAdd->getAPIntValue() << CN1->getAPIntValue();
-  if (!TII->canFoldOffset(Offset.getZExtValue(), AddrSpace))
+  if (!canFoldOffset(Offset.getZExtValue(), AddrSpace, *Subtarget))
     return SDValue();
 
   SelectionDAG &DAG = DCI.DAG;
@@ -1595,6 +1623,7 @@ SDValue SITargetLowering::PerformDAGCombine(SDNode *N,
   case AMDGPUISD::UMAX:
   case AMDGPUISD::UMIN: {
     if (DCI.getDAGCombineLevel() >= AfterLegalizeDAG &&
+        N->getValueType(0) != MVT::f64 &&
         getTargetMachine().getOptLevel() > CodeGenOpt::None)
       return performMin3Max3Combine(N, DCI);
     break;