X86: add GATHER intrinsics (AVX2) in LLVM

[oota-llvm.git] / lib / Target / X86 / MCTargetDesc / X86MCCodeEmitter.cpp

diff --git a/lib/Target/X86/MCTargetDesc/X86MCCodeEmitter.cpp b/lib/Target/X86/MCTargetDesc/X86MCCodeEmitter.cpp
index f79073ff58848af5035ce574d44f4f62913dfd8c..52506fa185f65e0e620cb6c68f2989778d4edbbc 100644 (file)
--- a/lib/Target/X86/MCTargetDesc/X86MCCodeEmitter.cpp
+++ b/lib/Target/X86/MCTargetDesc/X86MCCodeEmitter.cpp
@@ -570,7 +570,11 @@ void X86MCCodeEmitter::EmitVEXOpcodePrefix(uint64_t TSFlags, unsigned &CurByte,
   }
 
   // Classify VEX_B, VEX_4V, VEX_R, VEX_X
+  unsigned NumOps = Desc.getNumOperands();
   unsigned CurOp = 0;
+  if (NumOps > 1 && Desc.getOperandConstraint(1, MCOI::TIED_TO) != -1)
+    ++CurOp;
+
   switch (TSFlags & X86II::FormMask) {
   case X86II::MRMInitReg: llvm_unreachable("FIXME: Remove this!");
   case X86II::MRMDestMem: {
@@ -603,11 +607,11 @@ void X86MCCodeEmitter::EmitVEXOpcodePrefix(uint64_t TSFlags, unsigned &CurByte,
     //  FMA4:
     //  dst(ModR/M.reg), src1(VEX_4V), src2(ModR/M), src3(VEX_I8IMM)
     //  dst(ModR/M.reg), src1(VEX_4V), src2(VEX_I8IMM), src3(ModR/M),
-    if (X86II::isX86_64ExtendedReg(MI.getOperand(0).getReg()))
+    if (X86II::isX86_64ExtendedReg(MI.getOperand(CurOp++).getReg()))
       VEX_R = 0x0;
 
     if (HasVEX_4V)
-      VEX_4V = getVEXRegisterEncoding(MI, 1);
+      VEX_4V = getVEXRegisterEncoding(MI, CurOp);
 
     if (X86II::isX86_64ExtendedReg(
                MI.getOperand(MemOperand+X86::AddrBaseReg).getReg()))
@@ -617,7 +621,12 @@ void X86MCCodeEmitter::EmitVEXOpcodePrefix(uint64_t TSFlags, unsigned &CurByte,
       VEX_X = 0x0;
 
     if (HasVEX_4VOp3)
-      VEX_4V = getVEXRegisterEncoding(MI, X86::AddrNumOperands+1);
+      // Instruction format for 4VOp3:
+      //   src1(ModR/M), MemAddr, src3(VEX_4V)
+      // CurOp points to start of the MemoryOperand,
+      //   it skips TIED_TO operands if exist, then increments past src1.
+      // CurOp + X86::AddrNumOperands will point to src3.
+      VEX_4V = getVEXRegisterEncoding(MI, CurOp+X86::AddrNumOperands);
     break;
   case X86II::MRM0m: case X86II::MRM1m:
   case X86II::MRM2m: case X86II::MRM3m: