Allow MachineCSE to coalesce trivial subregister copies the same way that it coalesces normal copies.

Without this, MachineCSE is powerless to handle redundant operations with truncated source operands.

This required fixing the 2-addr pass to handle tied subregisters. It isn't clear what combinations of subregisters can legally be tied, but the simple case of truncated source operands is now safely handled:

     %vreg11<def> = COPY %vreg1:sub_32bit; GR32:%vreg11 GR64:%vreg1
     %vreg12<def> = COPY %vreg2:sub_32bit; GR32:%vreg12 GR64:%vreg2
     %vreg13<def,tied1> = ADD32rr %vreg11<tied0>, %vreg12<kill>, %EFLAGS<imp-def>

Test case: cse-add-with-overflow.ll.

This exposed an existing bug in
PPCInstrInfo::commuteInstruction. Thanks to Rafael for the test case:
PowerPC/crash.ll.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@197465 91177308-0d34-0410-b5e6-96231b3b80d8

2 files changed, 40 insertions, 10 deletions

diff --git a/lib/CodeGen/MachineCSE.cpp b/lib/CodeGen/MachineCSE.cpp
index 2e90f7472e..80982bca8c 100644
--- a/lib/CodeGen/MachineCSE.cpp
+++ b/lib/CodeGen/MachineCSE.cpp
@@ -131,13 +131,18 @@ bool MachineCSE::PerformTrivialCoalescing(MachineInstr *MI,
     unsigned SrcReg = DefMI->getOperand(1).getReg();
     if (!TargetRegisterInfo::isVirtualRegister(SrcReg))
       continue;
-    if (DefMI->getOperand(0).getSubReg() || DefMI->getOperand(1).getSubReg())
+    if (DefMI->getOperand(0).getSubReg())
       continue;
-    if (!MRI->constrainRegClass(SrcReg, MRI->getRegClass(Reg)))
+    unsigned SrcSubReg = DefMI->getOperand(1).getSubReg();
+    const TargetRegisterClass *RC = MRI->getRegClass(Reg);
+    if (SrcSubReg)
+      RC = TRI->getMatchingSuperRegClass(MRI->getRegClass(SrcReg), RC,
+                                         SrcSubReg);
+    if (!MRI->constrainRegClass(SrcReg, RC))
       continue;
     DEBUG(dbgs() << "Coalescing: " << *DefMI);
     DEBUG(dbgs() << "***     to: " << *MI);
-    MO.setReg(SrcReg);
+    MO.substVirtReg(SrcReg, SrcSubReg, *TRI);
     MRI->clearKillFlags(SrcReg);
     DefMI->eraseFromParent();
     ++NumCoalesces;
diff --git a/lib/CodeGen/TwoAddressInstructionPass.cpp b/lib/CodeGen/TwoAddressInstructionPass.cpp
index b9a6b479c3..15105d4b3f 100644
--- a/lib/CodeGen/TwoAddressInstructionPass.cpp
+++ b/lib/CodeGen/TwoAddressInstructionPass.cpp
@@ -1317,13 +1317,14 @@ collectTiedOperands(MachineInstr *MI, TiedOperandMap &TiedOperands) {
     assert(SrcReg && SrcMO.isUse() && "two address instruction invalid");
 
     // Deal with <undef> uses immediately - simply rewrite the src operand.
-    if (SrcMO.isUndef()) {
+    if (SrcMO.isUndef() && !DstMO.getSubReg()) {
       // Constrain the DstReg register class if required.
       if (TargetRegisterInfo::isVirtualRegister(DstReg))
         if (const TargetRegisterClass *RC = TII->getRegClass(MCID, SrcIdx,
                                                              TRI, *MF))
           MRI->constrainRegClass(DstReg, RC);
       SrcMO.setReg(DstReg);
+      SrcMO.setSubReg(0);
       DEBUG(dbgs() << "\t\trewrite undef:\t" << *MI);
       continue;
     }
@@ -1349,6 +1350,7 @@ TwoAddressInstructionPass::processTiedPairs(MachineInstr *MI,
   unsigned LastCopiedReg = 0;
   SlotIndex LastCopyIdx;
   unsigned RegB = 0;
+  unsigned SubRegB = 0;
   for (unsigned tpi = 0, tpe = TiedPairs.size(); tpi != tpe; ++tpi) {
     unsigned SrcIdx = TiedPairs[tpi].first;
     unsigned DstIdx = TiedPairs[tpi].second;
@@ -1359,6 +1361,7 @@ TwoAddressInstructionPass::processTiedPairs(MachineInstr *MI,
     // Grab RegB from the instruction because it may have changed if the
     // instruction was commuted.
     RegB = MI->getOperand(SrcIdx).getReg();
+    SubRegB = MI->getOperand(SrcIdx).getSubReg();
 
     if (RegA == RegB) {
       // The register is tied to multiple destinations (or else we would
@@ -1383,8 +1386,25 @@ TwoAddressInstructionPass::processTiedPairs(MachineInstr *MI,
 #endif
 
     // Emit a copy.
-    BuildMI(*MI->getParent(), MI, MI->getDebugLoc(),
-            TII->get(TargetOpcode::COPY), RegA).addReg(RegB);
+    MachineInstrBuilder MIB = BuildMI(*MI->getParent(), MI, MI->getDebugLoc(),
+                                      TII->get(TargetOpcode::COPY), RegA);
+    // If this operand is folding a truncation, the truncation now moves to the
+    // copy so that the register classes remain valid for the operands.
+    MIB.addReg(RegB, 0, SubRegB);
+    const TargetRegisterClass *RC = MRI->getRegClass(RegB);
+    if (SubRegB) {
+      if (TargetRegisterInfo::isVirtualRegister(RegA)) {
+        assert(TRI->getMatchingSuperRegClass(RC, MRI->getRegClass(RegA),
+                                             SubRegB) &&
+               "tied subregister must be a truncation");
+        // The superreg class will not be used to constrain the subreg class.
+        RC = 0;
+      }
+      else {
+        assert(TRI->getMatchingSuperReg(RegA, SubRegB, MRI->getRegClass(RegB))
+               && "tied subregister must be a truncation");
+      }
+    }
 
     // Update DistanceMap.
     MachineBasicBlock::iterator PrevMI = MI;
@@ -1404,7 +1424,7 @@ TwoAddressInstructionPass::processTiedPairs(MachineInstr *MI,
       }
     }
 
-    DEBUG(dbgs() << "\t\tprepend:\t" << *PrevMI);
+    DEBUG(dbgs() << "\t\tprepend:\t" << *MIB);
 
     MachineOperand &MO = MI->getOperand(SrcIdx);
     assert(MO.isReg() && MO.getReg() == RegB && MO.isUse() &&
@@ -1417,9 +1437,12 @@ TwoAddressInstructionPass::processTiedPairs(MachineInstr *MI,
     // Make sure regA is a legal regclass for the SrcIdx operand.
     if (TargetRegisterInfo::isVirtualRegister(RegA) &&
         TargetRegisterInfo::isVirtualRegister(RegB))
-      MRI->constrainRegClass(RegA, MRI->getRegClass(RegB));
-
+      MRI->constrainRegClass(RegA, RC);
     MO.setReg(RegA);
+    // The getMatchingSuper asserts guarantee that the register class projected
+    // by SubRegB is compatible with RegA with no subregister. So regardless of
+    // whether the dest oper writes a subreg, the source oper should not.
+    MO.setSubReg(0);
 
     // Propagate SrcRegMap.
     SrcRegMap[RegA] = RegB;
@@ -1431,12 +1454,14 @@ TwoAddressInstructionPass::processTiedPairs(MachineInstr *MI,
       // Replace other (un-tied) uses of regB with LastCopiedReg.
       for (unsigned i = 0, e = MI->getNumOperands(); i != e; ++i) {
         MachineOperand &MO = MI->getOperand(i);
-        if (MO.isReg() && MO.getReg() == RegB && MO.isUse()) {
+        if (MO.isReg() && MO.getReg() == RegB && MO.getSubReg() == SubRegB &&
+            MO.isUse()) {
           if (MO.isKill()) {
             MO.setIsKill(false);
             RemovedKillFlag = true;
           }
           MO.setReg(LastCopiedReg);
+          MO.setSubReg(0);
         }
       }
     }