Update llvm, clang and lldb to release_38 branch r258968.

8 files changed, 34 insertions, 3 deletions

diff --git a/contrib/llvm/lib/Analysis/DemandedBits.cpp b/contrib/llvm/lib/Analysis/DemandedBits.cpp
index 912c5ce..143d0b7 100644
--- a/contrib/llvm/lib/Analysis/DemandedBits.cpp
+++ b/contrib/llvm/lib/Analysis/DemandedBits.cpp
@@ -244,7 +244,7 @@ void DemandedBits::determineLiveOperandBits(
     break;
   case Instruction::ICmp:
     // Count the number of leading zeroes in each operand.
-    ComputeKnownBits(BitWidth, I, UserI->getOperand(1));
+    ComputeKnownBits(BitWidth, UserI->getOperand(0), UserI->getOperand(1));
     auto NumLeadingZeroes = std::min(KnownZero.countLeadingOnes(),
                                      KnownZero2.countLeadingOnes());
     AB = ~APInt::getHighBitsSet(BitWidth, NumLeadingZeroes);
diff --git a/contrib/llvm/lib/Analysis/ScalarEvolution.cpp b/contrib/llvm/lib/Analysis/ScalarEvolution.cpp
index 34074ef..ef1bb3a 100644
--- a/contrib/llvm/lib/Analysis/ScalarEvolution.cpp
+++ b/contrib/llvm/lib/Analysis/ScalarEvolution.cpp
@@ -5368,6 +5368,14 @@ ScalarEvolution::computeExitLimitFromCond(const Loop *L,
           BECount = EL0.Exact;
       }
 
+      // There are cases (e.g. PR26207) where computeExitLimitFromCond is able
+      // to be more aggressive when computing BECount than when computing
+      // MaxBECount.  In these cases it is possible for EL0.Exact and EL1.Exact
+      // to match, but for EL0.Max and EL1.Max to not.
+      if (isa<SCEVCouldNotCompute>(MaxBECount) &&
+          !isa<SCEVCouldNotCompute>(BECount))
+        MaxBECount = BECount;
+
       return ExitLimit(BECount, MaxBECount);
     }
     if (BO->getOpcode() == Instruction::Or) {
diff --git a/contrib/llvm/lib/Target/AMDGPU/AMDGPU.td b/contrib/llvm/lib/Target/AMDGPU/AMDGPU.td
index db869cf..79c6604 100644
--- a/contrib/llvm/lib/Target/AMDGPU/AMDGPU.td
+++ b/contrib/llvm/lib/Target/AMDGPU/AMDGPU.td
@@ -138,6 +138,11 @@ def FeatureEnableHugeScratchBuffer : SubtargetFeature<"huge-scratch-buffer",
         "true",
         "Enable scratch buffer sizes greater than 128 GB">;
 
+def FeatureEnableSIScheduler : SubtargetFeature<"si-scheduler",
+        "EnableSIScheduler",
+        "true",
+        "Enable SI Machine Scheduler">;
+
 class SubtargetFeatureFetchLimit <string Value> :
                           SubtargetFeature <"fetch"#Value,
         "TexVTXClauseSize",
diff --git a/contrib/llvm/lib/Target/AMDGPU/AMDGPUSubtarget.cpp b/contrib/llvm/lib/Target/AMDGPU/AMDGPUSubtarget.cpp
index c6af5b9..7d70fa7 100644
--- a/contrib/llvm/lib/Target/AMDGPU/AMDGPUSubtarget.cpp
+++ b/contrib/llvm/lib/Target/AMDGPU/AMDGPUSubtarget.cpp
@@ -78,7 +78,7 @@ AMDGPUSubtarget::AMDGPUSubtarget(const Triple &TT, StringRef GPU, StringRef FS,
       EnableVGPRSpilling(false), SGPRInitBug(false), IsGCN(false),
       GCN1Encoding(false), GCN3Encoding(false), CIInsts(false), LDSBankCount(0),
       IsaVersion(ISAVersion0_0_0), EnableHugeScratchBuffer(false),
-      FrameLowering(nullptr),
+      EnableSIScheduler(false), FrameLowering(nullptr),
       InstrItins(getInstrItineraryForCPU(GPU)), TargetTriple(TT) {
 
   initializeSubtargetDependencies(TT, GPU, FS);
diff --git a/contrib/llvm/lib/Target/AMDGPU/AMDGPUSubtarget.h b/contrib/llvm/lib/Target/AMDGPU/AMDGPUSubtarget.h
index d371227..4796e9e 100644
--- a/contrib/llvm/lib/Target/AMDGPU/AMDGPUSubtarget.h
+++ b/contrib/llvm/lib/Target/AMDGPU/AMDGPUSubtarget.h
@@ -90,6 +90,7 @@ private:
   int LDSBankCount;
   unsigned IsaVersion;
   bool EnableHugeScratchBuffer;
+  bool EnableSIScheduler;
 
   std::unique_ptr<AMDGPUFrameLowering> FrameLowering;
   std::unique_ptr<AMDGPUTargetLowering> TLInfo;
@@ -280,6 +281,10 @@ public:
     return EnableHugeScratchBuffer;
   }
 
+  bool enableSIScheduler() const {
+    return EnableSIScheduler;
+  }
+
   bool dumpCode() const {
     return DumpCode;
   }
diff --git a/contrib/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp b/contrib/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
index b1be619..519ae5c 100644
--- a/contrib/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
+++ b/contrib/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
@@ -147,6 +147,8 @@ public:
     const AMDGPUSubtarget &ST = *getAMDGPUTargetMachine().getSubtargetImpl();
     if (ST.getGeneration() <= AMDGPUSubtarget::NORTHERN_ISLANDS)
       return createR600MachineScheduler(C);
+    else if (ST.enableSIScheduler())
+      return createSIMachineScheduler(C);
     return nullptr;
   }
 
diff --git a/contrib/llvm/lib/Target/X86/X86ISelLowering.cpp b/contrib/llvm/lib/Target/X86/X86ISelLowering.cpp
index 6904714..34f3919 100644
--- a/contrib/llvm/lib/Target/X86/X86ISelLowering.cpp
+++ b/contrib/llvm/lib/Target/X86/X86ISelLowering.cpp
@@ -21880,7 +21880,8 @@ X86TargetLowering::EmitLoweredSelect(MachineInstr *MI,
   if (LastCMOV == MI &&
       NextMIIt != BB->end() && NextMIIt->getOpcode() == MI->getOpcode() &&
       NextMIIt->getOperand(2).getReg() == MI->getOperand(2).getReg() &&
-      NextMIIt->getOperand(1).getReg() == MI->getOperand(0).getReg()) {
+      NextMIIt->getOperand(1).getReg() == MI->getOperand(0).getReg() &&
+      NextMIIt->getOperand(1).isKill()) {
     CascadedCMOV = &*NextMIIt;
   }
 
diff --git a/contrib/llvm/lib/Transforms/Instrumentation/GCOVProfiling.cpp b/contrib/llvm/lib/Transforms/Instrumentation/GCOVProfiling.cpp
index fa939ae..ffde7f8 100644
--- a/contrib/llvm/lib/Transforms/Instrumentation/GCOVProfiling.cpp
+++ b/contrib/llvm/lib/Transforms/Instrumentation/GCOVProfiling.cpp
@@ -494,6 +494,11 @@ void GCOVProfiler::emitProfileNotes() {
     // LTO, we'll generate the same .gcno files.
 
     auto *CU = cast<DICompileUnit>(CU_Nodes->getOperand(i));
+
+    // Skip module skeleton (and module) CUs.
+    if (CU->getDWOId())
+      continue;
+
     std::error_code EC;
     raw_fd_ostream out(mangleName(CU, "gcno"), EC, sys::fs::F_None);
     std::string EdgeDestinations;
@@ -853,6 +858,11 @@ Function *GCOVProfiler::insertCounterWriteout(
   if (CU_Nodes) {
     for (unsigned i = 0, e = CU_Nodes->getNumOperands(); i != e; ++i) {
       auto *CU = cast<DICompileUnit>(CU_Nodes->getOperand(i));
+
+      // Skip module skeleton (and module) CUs.
+      if (CU->getDWOId())
+        continue;
+
       std::string FilenameGcda = mangleName(CU, "gcda");
       uint32_t CfgChecksum = FileChecksums.empty() ? 0 : FileChecksums[i];
       Builder.CreateCall(StartFile,