[RISCV] Removed the requirement of XLenVT for performSELECTCombine

[mgudim]

Not sure why we have this restriction. I've ran most of spec without it, all the tests passed and even got some small improvements.

[craig.topper]

Tests?

[mgudim]

Tests?

I'll update soon. I just wanted to first see if there are any principal objections.

[mgudim]

updated tests

[craig.topper]

LGTM

[asb]

@mgudim are you happy to go ahead and commit this? Thanks!

[mgudim]

@abs Yes, will try to commit this today.

llvm/lib/Target/RISCV/RISCVISelLowering.cpp

static SDValue tryFoldSelectIntoOp(SDNode *N, SelectionDAG &DAG,
   return DAG.getNode(TrueVal.getOpcode(), DL, VT, FalseVal, NewSel);
 }
 
 static SDValue performSELECTCombine(SDNode *N, SelectionDAG &DAG,
                                     const RISCVSubtarget &Subtarget) {
   if (Subtarget.hasShortForwardBranchOpt())
     return SDValue();
 
-  // Only support XLenVT.
-  if (N->getValueType(0) != Subtarget.getXLenVT())
-    return SDValue();
-
   SDValue TrueVal = N->getOperand(1);
   SDValue FalseVal = N->getOperand(2);
   if (SDValue V = tryFoldSelectIntoOp(N, DAG, TrueVal, FalseVal, /*Swapped*/false))
     return V;
   return tryFoldSelectIntoOp(N, DAG, FalseVal, TrueVal, /*Swapped*/true);
 }
 
 // If we're concatenating a series of vector loads like