[X86][SSE] Added support for MOVHPD/MOVLPD + MOVHPS/MOVLPS shuffle decoding.

llvm-svn: 260034

3 files changed, 48 insertions, 0 deletions

diff --git a/llvm/lib/Target/X86/InstPrinter/X86InstComments.cpp b/llvm/lib/Target/X86/InstPrinter/X86InstComments.cpp
index 04a937807ea..5e0dd4f7bfc 100644
--- a/llvm/lib/Target/X86/InstPrinter/X86InstComments.cpp
+++ b/llvm/lib/Target/X86/InstPrinter/X86InstComments.cpp
@@ -257,6 +257,38 @@ bool llvm::EmitAnyX86InstComments(const MCInst *MI, raw_ostream &OS,
     DecodeMOVHLPSMask(2, ShuffleMask);
     break;
 
+  case X86::MOVHPDrm:
+  case X86::VMOVHPDrm:
+  case X86::VMOVHPDZ128rm:
+    Src1Name = getRegName(MI->getOperand(1).getReg());
+    DestName = getRegName(MI->getOperand(0).getReg());
+    DecodeInsertElementMask(MVT::v2f64, 1, 1, ShuffleMask);
+    break;
+
+  case X86::MOVHPSrm:
+  case X86::VMOVHPSrm:
+  case X86::VMOVHPSZ128rm:
+    Src1Name = getRegName(MI->getOperand(1).getReg());
+    DestName = getRegName(MI->getOperand(0).getReg());
+    DecodeInsertElementMask(MVT::v4f32, 2, 2, ShuffleMask);
+    break;
+
+  case X86::MOVLPDrm:
+  case X86::VMOVLPDrm:
+  case X86::VMOVLPDZ128rm:
+    Src1Name = getRegName(MI->getOperand(1).getReg());
+    DestName = getRegName(MI->getOperand(0).getReg());
+    DecodeInsertElementMask(MVT::v2f64, 0, 1, ShuffleMask);
+    break;
+
+  case X86::MOVLPSrm:
+  case X86::VMOVLPSrm:
+  case X86::VMOVLPSZ128rm:
+    Src1Name = getRegName(MI->getOperand(1).getReg());
+    DestName = getRegName(MI->getOperand(0).getReg());
+    DecodeInsertElementMask(MVT::v4f32, 0, 2, ShuffleMask);
+    break;
+
   CASE_MOVDUP(MOVSLDUP, r)
     Src1Name = getRegName(MI->getOperand(MI->getNumOperands() - 1).getReg());
     // FALL THROUGH.
diff --git a/llvm/lib/Target/X86/Utils/X86ShuffleDecode.cpp b/llvm/lib/Target/X86/Utils/X86ShuffleDecode.cpp
index 08c039f3dec..e2e3f8cf0ea 100644
--- a/llvm/lib/Target/X86/Utils/X86ShuffleDecode.cpp
+++ b/llvm/lib/Target/X86/Utils/X86ShuffleDecode.cpp
@@ -44,6 +44,17 @@ void DecodeINSERTPSMask(unsigned Imm, SmallVectorImpl<int> &ShuffleMask) {
   if (ZMask & 8) ShuffleMask[3] = SM_SentinelZero;
 }
 
+void DecodeInsertElementMask(MVT VT, unsigned Idx, unsigned Len,
+                             SmallVectorImpl<int> &ShuffleMask) {
+  unsigned NumElts = VT.getVectorNumElements();
+  assert((Idx + Len) <= NumElts && "Insertion out of range");
+
+  for (unsigned i = 0; i != NumElts; ++i)
+    ShuffleMask.push_back(i);
+  for (unsigned i = 0; i != Len; ++i)
+    ShuffleMask[Idx + i] = NumElts + i;
+}
+
 // <3,1> or <6,7,2,3>
 void DecodeMOVHLPSMask(unsigned NElts, SmallVectorImpl<int> &ShuffleMask) {
   for (unsigned i = NElts / 2; i != NElts; ++i)
diff --git a/llvm/lib/Target/X86/Utils/X86ShuffleDecode.h b/llvm/lib/Target/X86/Utils/X86ShuffleDecode.h
index 185256fd1be..f0fd04ac8b2 100644
--- a/llvm/lib/Target/X86/Utils/X86ShuffleDecode.h
+++ b/llvm/lib/Target/X86/Utils/X86ShuffleDecode.h
@@ -29,6 +29,11 @@ enum { SM_SentinelUndef = -1, SM_SentinelZero = -2 };
 
 void DecodeINSERTPSMask(unsigned Imm, SmallVectorImpl<int> &ShuffleMask);
 
+// Insert the bottom Len elements from a second source into a vector starting at
+// element Idx.
+void DecodeInsertElementMask(MVT VT, unsigned Idx, unsigned Len,
+                             SmallVectorImpl<int> &ShuffleMask);
+
 // <3,1> or <6,7,2,3>
 void DecodeMOVHLPSMask(unsigned NElts, SmallVectorImpl<int> &ShuffleMask);