[SparcV9]: Implement lowering of long double (fp128) arguments in Sparc64 ABI.

Also, pass fp128 arguments to varargs through integer registers if necessary.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@198145 91177308-0d34-0410-b5e6-96231b3b80d8

1 files changed, 60 insertions, 8 deletions

diff --git a/lib/Target/Sparc/SparcISelLowering.cpp b/lib/Target/Sparc/SparcISelLowering.cpp
index 64625f7018..b794f8702b 100644
--- a/lib/Target/Sparc/SparcISelLowering.cpp
+++ b/lib/Target/Sparc/SparcISelLowering.cpp
@@ -80,11 +80,14 @@ static bool CC_Sparc_Assign_f64(unsigned &ValNo, MVT &ValVT,
 static bool CC_Sparc64_Full(unsigned &ValNo, MVT &ValVT,
                             MVT &LocVT, CCValAssign::LocInfo &LocInfo,
                             ISD::ArgFlagsTy &ArgFlags, CCState &State) {
-  assert((LocVT == MVT::f32 || LocVT.getSizeInBits() == 64) &&
+  assert((LocVT == MVT::f32 || LocVT == MVT::f128
+          || LocVT.getSizeInBits() == 64) &&
          "Can't handle non-64 bits locations");
 
   // Stack space is allocated for all arguments starting from [%fp+BIAS+128].
-  unsigned Offset = State.AllocateStack(8, 8);
+  unsigned size      = (LocVT == MVT::f128) ? 16 : 8;
+  unsigned alignment = (LocVT == MVT::f128) ? 16 : 8;
+  unsigned Offset = State.AllocateStack(size, alignment);
   unsigned Reg = 0;
 
   if (LocVT == MVT::i64 && Offset < 6*8)
@@ -96,6 +99,9 @@ static bool CC_Sparc64_Full(unsigned &ValNo, MVT &ValVT,
   else if (LocVT == MVT::f32 && Offset < 16*8)
     // Promote floats to %f1, %f3, ...
     Reg = SP::F1 + Offset/4;
+  else if (LocVT == MVT::f128 && Offset < 16*8)
+    // Promote long doubles to %q0-%q28. (Which LLVM calls Q0-Q7).
+    Reg = SP::Q0 + Offset/16;
 
   // Promote to register when possible, otherwise use the stack slot.
   if (Reg) {
@@ -998,9 +1004,10 @@ static void fixupVariableFloatArgs(SmallVectorImpl<CCValAssign> &ArgLocs,
                                    ArrayRef<ISD::OutputArg> Outs) {
   for (unsigned i = 0, e = ArgLocs.size(); i != e; ++i) {
     const CCValAssign &VA = ArgLocs[i];
+    MVT ValTy = VA.getLocVT();
     // FIXME: What about f32 arguments? C promotes them to f64 when calling
     // varargs functions.
-    if (!VA.isRegLoc() || VA.getLocVT() != MVT::f64)
+    if (!VA.isRegLoc() || (ValTy != MVT::f64 && ValTy != MVT::f128))
       continue;
     // The fixed arguments to a varargs function still go in FP registers.
     if (Outs[VA.getValNo()].IsFixed)
@@ -1010,15 +1017,25 @@ static void fixupVariableFloatArgs(SmallVectorImpl<CCValAssign> &ArgLocs,
     CCValAssign NewVA;
 
     // Determine the offset into the argument array.
-    unsigned Offset = 8 * (VA.getLocReg() - SP::D0);
+    unsigned firstReg = (ValTy == MVT::f64) ? SP::D0 : SP::Q0;
+    unsigned argSize  = (ValTy == MVT::f64) ? 8 : 16;
+    unsigned Offset = argSize * (VA.getLocReg() - firstReg);
     assert(Offset < 16*8 && "Offset out of range, bad register enum?");
 
     if (Offset < 6*8) {
       // This argument should go in %i0-%i5.
       unsigned IReg = SP::I0 + Offset/8;
-      // Full register, just bitconvert into i64.
-      NewVA = CCValAssign::getReg(VA.getValNo(), VA.getValVT(),
-                                  IReg, MVT::i64, CCValAssign::BCvt);
+      if (ValTy == MVT::f64)
+        // Full register, just bitconvert into i64.
+        NewVA = CCValAssign::getReg(VA.getValNo(), VA.getValVT(),
+                                    IReg, MVT::i64, CCValAssign::BCvt);
+      else {
+        assert(ValTy == MVT::f128 && "Unexpected type!");
+        // Full register, just bitconvert into i128 -- We will lower this into
+        // two i64s in LowerCall_64.
+        NewVA = CCValAssign::getCustomReg(VA.getValNo(), VA.getValVT(),
+                                          IReg, MVT::i128, CCValAssign::BCvt);
+      }
     } else {
       // This needs to go to memory, we're out of integer registers.
       NewVA = CCValAssign::getMem(VA.getValNo(), VA.getValVT(),
@@ -1094,11 +1111,46 @@ SparcTargetLowering::LowerCall_64(TargetLowering::CallLoweringInfo &CLI,
       Arg = DAG.getNode(ISD::ANY_EXTEND, DL, VA.getLocVT(), Arg);
       break;
     case CCValAssign::BCvt:
-      Arg = DAG.getNode(ISD::BITCAST, DL, VA.getLocVT(), Arg);
+      // fixupVariableFloatArgs() may create bitcasts from f128 to i128. But
+      // SPARC does not support i128 natively. Lower it into two i64, see below.
+      if (!VA.needsCustom() || VA.getValVT() != MVT::f128
+          || VA.getLocVT() != MVT::i128)
+        Arg = DAG.getNode(ISD::BITCAST, DL, VA.getLocVT(), Arg);
       break;
     }
 
     if (VA.isRegLoc()) {
+      if (VA.needsCustom() && VA.getValVT() == MVT::f128
+          && VA.getLocVT() == MVT::i128) {
+        // Store and reload into the interger register reg and reg+1.
+        unsigned Offset = 8 * (VA.getLocReg() - SP::I0);
+        unsigned StackOffset = Offset + Subtarget->getStackPointerBias() + 128;
+        SDValue StackPtr = DAG.getRegister(SP::O6, getPointerTy());
+        SDValue HiPtrOff = DAG.getIntPtrConstant(StackOffset);
+        HiPtrOff         = DAG.getNode(ISD::ADD, DL, getPointerTy(), StackPtr,
+                                       HiPtrOff);
+        SDValue LoPtrOff = DAG.getIntPtrConstant(StackOffset + 8);
+        LoPtrOff         = DAG.getNode(ISD::ADD, DL, getPointerTy(), StackPtr,
+                                       LoPtrOff);
+
+        // Store to %sp+BIAS+128+Offset
+        SDValue Store = DAG.getStore(Chain, DL, Arg, HiPtrOff,
+                                     MachinePointerInfo(),
+                                     false, false, 0);
+        // Load into Reg and Reg+1
+        SDValue Hi64 = DAG.getLoad(MVT::i64, DL, Store, HiPtrOff,
+                                   MachinePointerInfo(),
+                                   false, false, false, 0);
+        SDValue Lo64 = DAG.getLoad(MVT::i64, DL, Store, LoPtrOff,
+                                   MachinePointerInfo(),
+                                   false, false, false, 0);
+        RegsToPass.push_back(std::make_pair(toCallerWindow(VA.getLocReg()),
+                                            Hi64));
+        RegsToPass.push_back(std::make_pair(toCallerWindow(VA.getLocReg()+1),
+                                            Lo64));
+        continue;
+      }
+
       // The custom bit on an i32 return value indicates that it should be
       // passed in the high bits of the register.
       if (VA.getValVT() == MVT::i32 && VA.needsCustom()) {