Merged main:6ae657b08d624f9634fa6ebbf5d6fd7a22dc3b4d into amd-gfx:57a9efc3a4ac

Local branch amd-gfx 57a9efc Merged main:2f91e98120f168b7ded6cb34d546dba178515cc4 into amd-gfx:58ed5842fd7a
Remote branch main 6ae657b [lldb] Adapt Plugins/Process/Windows to new Status API

Author: SC llvm team

.github/workflows/libcxx-build-and-test.yaml

@@ -191,6 +191,42 @@ jobs:
             **/CMakeError.log
             **/CMakeOutput.log
             **/crash_diagnostics/*
+
+  macos:
+    runs-on: macos-14
+    needs: [ stage1 ]
+    strategy:
+      fail-fast: true
+      matrix:
+        config: [
+          generic-cxx03,
+          generic-cxx23,
+          generic-modules,
+          apple-configuration
+        ]
+    steps:
+      - uses: actions/checkout@v4
+      - uses: maxim-lobanov/setup-xcode@v1
+        with:
+          xcode-version: 'latest-stable'
+      - uses: seanmiddleditch/gha-setup-ninja@master
+      - name: Build and test
+        run: |
+          python3 -m venv .venv
+          source .venv/bin/activate
+          python -m pip install psutil
+          bash libcxx/utils/ci/run-buildbot ${{ matrix.config }}
+      - uses: actions/upload-artifact@26f96dfa697d77e81fd5907df203aa23a56210a8 # v4.3.0
+        if: always()  # Upload artifacts even if the build or test suite fails
+        with:
+          name: macos-${{ matrix.config }}-results
+          path: |
+            **/test-results.xml
+            **/*.abilist
+            **/CMakeError.log
+            **/CMakeOutput.log
+            **/crash_diagnostics/*
+
   windows:
     runs-on: windows-2022
     needs: [ stage1 ]

bolt/include/bolt/Core/BinaryFunction.h

@@ -1692,6 +1692,8 @@ class BinaryFunction {
 
   void setPseudo(bool Pseudo) { IsPseudo = Pseudo; }
 
+  void setPreserveNops(bool Value) { PreserveNops = Value; }
+
   BinaryFunction &setUsesGnuArgsSize(bool Uses = true) {
     UsesGnuArgsSize = Uses;
     return *this;

bolt/lib/Core/BinaryFunction.cpp

@@ -1339,22 +1339,10 @@ Error BinaryFunction::disassemble() {
                   BC.getBinaryFunctionContainingAddress(TargetAddress))
             TargetFunc->setIgnored();
 
-        if (IsCall && containsAddress(TargetAddress)) {
-          if (TargetAddress == getAddress()) {
-            // Recursive call.
-            TargetSymbol = getSymbol();
-          } else {
-            if (BC.isX86()) {
-              // Dangerous old-style x86 PIC code. We may need to freeze this
-              // function, so preserve the function as is for now.
-              PreserveNops = true;
-            } else {
-              BC.errs() << "BOLT-WARNING: internal call detected at 0x"
-                        << Twine::utohexstr(AbsoluteInstrAddr)
-                        << " in function " << *this << ". Skipping.\n";
-              IsSimple = false;
-            }
-          }
+        if (IsCall && TargetAddress == getAddress()) {
+          // A recursive call. Calls to internal blocks are handled by
+          // ValidateInternalCalls pass.
+          TargetSymbol = getSymbol();
         }
 
         if (!TargetSymbol) {

bolt/lib/Passes/ValidateInternalCalls.cpp

@@ -302,24 +302,27 @@ bool ValidateInternalCalls::analyzeFunction(BinaryFunction &Function) const {
 }
 
 Error ValidateInternalCalls::runOnFunctions(BinaryContext &BC) {
-  if (!BC.isX86())
-    return Error::success();
-
   // Look for functions that need validation. This should be pretty rare.
   std::set<BinaryFunction *> NeedsValidation;
   for (auto &BFI : BC.getBinaryFunctions()) {
     BinaryFunction &Function = BFI.second;
     for (BinaryBasicBlock &BB : Function) {
       for (MCInst &Inst : BB) {
         if (getInternalCallTarget(Function, Inst)) {
+          BC.errs() << "BOLT-WARNING: internal call detected in function "
+                    << Function << '\n';
           NeedsValidation.insert(&Function);
           Function.setSimple(false);
+          Function.setPreserveNops(true);
           break;
         }
       }
     }
   }
 
+  if (!BC.isX86())
+    return Error::success();
+
   // Skip validation for non-relocation mode
   if (!BC.HasRelocations)
     return Error::success();

bolt/lib/Profile/DataAggregator.cpp

@@ -2415,17 +2415,15 @@ std::error_code DataAggregator::writeBATYAML(BinaryContext &BC,
         Fragments.insert(BF);
         for (const BinaryFunction *F : Fragments) {
           const uint64_t FuncAddr = F->getAddress();
-          const auto &FragmentProbes =
-              llvm::make_range(ProbeMap.lower_bound(FuncAddr),
-                               ProbeMap.lower_bound(FuncAddr + F->getSize()));
-          for (const auto &[OutputAddress, Probes] : FragmentProbes) {
+          for (const MCDecodedPseudoProbe &Probe :
+               ProbeMap.find(FuncAddr, FuncAddr + F->getSize())) {
+            const uint32_t OutputAddress = Probe.getAddress();
             const uint32_t InputOffset = BAT->translate(
                 FuncAddr, OutputAddress - FuncAddr, /*IsBranchSrc=*/true);
             const unsigned BlockIndex = getBlock(InputOffset).second;
-            for (const MCDecodedPseudoProbe &Probe : Probes)
-              YamlBF.Blocks[BlockIndex].PseudoProbes.emplace_back(
-                  yaml::bolt::PseudoProbeInfo{Probe.getGuid(), Probe.getIndex(),
-                                              Probe.getType()});
+            YamlBF.Blocks[BlockIndex].PseudoProbes.emplace_back(
+                yaml::bolt::PseudoProbeInfo{Probe.getGuid(), Probe.getIndex(),
+                                            Probe.getType()});
           }
         }
       }

bolt/lib/Profile/YAMLProfileWriter.cpp

@@ -193,13 +193,10 @@ YAMLProfileWriter::convert(const BinaryFunction &BF, bool UseDFS,
       const uint64_t FuncAddr = BF.getAddress();
       const std::pair<uint64_t, uint64_t> &BlockRange =
           BB->getInputAddressRange();
-      const auto &BlockProbes =
-          llvm::make_range(ProbeMap.lower_bound(FuncAddr + BlockRange.first),
-                           ProbeMap.lower_bound(FuncAddr + BlockRange.second));
-      for (const auto &[_, Probes] : BlockProbes)
-        for (const MCDecodedPseudoProbe &Probe : Probes)
-          YamlBB.PseudoProbes.emplace_back(yaml::bolt::PseudoProbeInfo{
-              Probe.getGuid(), Probe.getIndex(), Probe.getType()});
+      for (const MCDecodedPseudoProbe &Probe : ProbeMap.find(
+               FuncAddr + BlockRange.first, FuncAddr + BlockRange.second))
+        YamlBB.PseudoProbes.emplace_back(yaml::bolt::PseudoProbeInfo{
+            Probe.getGuid(), Probe.getIndex(), Probe.getType()});
     }
 
     YamlBF.Blocks.emplace_back(YamlBB);

bolt/lib/Rewrite/PseudoProbeRewriter.cpp

@@ -143,7 +143,6 @@ void PseudoProbeRewriter::parsePseudoProbe() {
   if (!ProbeDecoder.buildAddress2ProbeMap(
           reinterpret_cast<const uint8_t *>(Contents.data()), Contents.size(),
           GuidFilter, FuncStartAddrs)) {
-    ProbeDecoder.getAddress2ProbesMap().clear();
     errs() << "BOLT-WARNING: fail in building Address2ProbeMap\n";
     return;
   }
@@ -156,7 +155,8 @@ void PseudoProbeRewriter::parsePseudoProbe() {
     ProbeDecoder.printProbesForAllAddresses(outs());
   }
 
-  for (const auto &[GUID, FuncDesc] : ProbeDecoder.getGUID2FuncDescMap()) {
+  for (const auto &FuncDesc : ProbeDecoder.getGUID2FuncDescMap()) {
+    uint64_t GUID = FuncDesc.FuncGUID;
     if (!FuncStartAddrs.contains(GUID))
       continue;
     BinaryFunction *BF = BC.getBinaryFunctionAtAddress(FuncStartAddrs[GUID]);
@@ -174,59 +174,50 @@ void PseudoProbeRewriter::updatePseudoProbes() {
   AddressProbesMap &Address2ProbesMap = ProbeDecoder.getAddress2ProbesMap();
   const GUIDProbeFunctionMap &GUID2Func = ProbeDecoder.getGUID2FuncDescMap();
 
-  for (auto &AP : Address2ProbesMap) {
-    BinaryFunction *F = BC.getBinaryFunctionContainingAddress(AP.first);
+  for (MCDecodedPseudoProbe &Probe : Address2ProbesMap) {
+    uint64_t Address = Probe.getAddress();
+    BinaryFunction *F = BC.getBinaryFunctionContainingAddress(Address);
     // If F is removed, eliminate all probes inside it from inline tree
     // Setting probes' addresses as INT64_MAX means elimination
     if (!F) {
-      for (MCDecodedPseudoProbe &Probe : AP.second)
-        Probe.setAddress(INT64_MAX);
+      Probe.setAddress(INT64_MAX);
       continue;
     }
     // If F is not emitted, the function will remain in the same address as its
     // input
     if (!F->isEmitted())
       continue;
 
-    uint64_t Offset = AP.first - F->getAddress();
+    uint64_t Offset = Address - F->getAddress();
     const BinaryBasicBlock *BB = F->getBasicBlockContainingOffset(Offset);
     uint64_t BlkOutputAddress = BB->getOutputAddressRange().first;
     // Check if block output address is defined.
     // If not, such block is removed from binary. Then remove the probes from
     // inline tree
     if (BlkOutputAddress == 0) {
-      for (MCDecodedPseudoProbe &Probe : AP.second)
-        Probe.setAddress(INT64_MAX);
+      Probe.setAddress(INT64_MAX);
       continue;
     }
 
-    unsigned ProbeTrack = AP.second.size();
-    std::list<MCDecodedPseudoProbe>::iterator Probe = AP.second.begin();
-    while (ProbeTrack != 0) {
-      if (Probe->isBlock()) {
-        Probe->setAddress(BlkOutputAddress);
-      } else if (Probe->isCall()) {
-        // A call probe may be duplicated due to ICP
-        // Go through output of InputOffsetToAddressMap to collect all related
-        // probes
-        auto CallOutputAddresses = BC.getIOAddressMap().lookupAll(AP.first);
-        auto CallOutputAddress = CallOutputAddresses.first;
-        if (CallOutputAddress == CallOutputAddresses.second) {
-          Probe->setAddress(INT64_MAX);
-        } else {
-          Probe->setAddress(CallOutputAddress->second);
-          CallOutputAddress = std::next(CallOutputAddress);
-        }
-
-        while (CallOutputAddress != CallOutputAddresses.second) {
-          AP.second.push_back(*Probe);
-          AP.second.back().setAddress(CallOutputAddress->second);
-          Probe->getInlineTreeNode()->addProbes(&(AP.second.back()));
-          CallOutputAddress = std::next(CallOutputAddress);
-        }
+    if (Probe.isBlock()) {
+      Probe.setAddress(BlkOutputAddress);
+    } else if (Probe.isCall()) {
+      // A call probe may be duplicated due to ICP
+      // Go through output of InputOffsetToAddressMap to collect all related
+      // probes
+      auto CallOutputAddresses = BC.getIOAddressMap().lookupAll(Address);
+      auto CallOutputAddress = CallOutputAddresses.first;
+      if (CallOutputAddress == CallOutputAddresses.second) {
+        Probe.setAddress(INT64_MAX);
+      } else {
+        Probe.setAddress(CallOutputAddress->second);
+        CallOutputAddress = std::next(CallOutputAddress);
+      }
+
+      while (CallOutputAddress != CallOutputAddresses.second) {
+        ProbeDecoder.addInjectedProbe(Probe, CallOutputAddress->second);
+        CallOutputAddress = std::next(CallOutputAddress);
       }
-      Probe = std::next(Probe);
-      ProbeTrack--;
     }
   }
 
@@ -242,22 +233,16 @@ void PseudoProbeRewriter::updatePseudoProbes() {
             BinaryBlock.getName();
 
     // scan all addresses -> correlate probe to block when print out
-    std::vector<uint64_t> Addresses;
-    for (auto &Entry : Address2ProbesMap)
-      Addresses.push_back(Entry.first);
-    llvm::sort(Addresses);
-    for (uint64_t Key : Addresses) {
-      for (MCDecodedPseudoProbe &Probe : Address2ProbesMap[Key]) {
-        if (Probe.getAddress() == INT64_MAX)
-          outs() << "Deleted Probe: ";
-        else
-          outs() << "Address: " << format_hex(Probe.getAddress(), 8) << " ";
-        Probe.print(outs(), GUID2Func, true);
-        // print block name only if the probe is block type and undeleted.
-        if (Probe.isBlock() && Probe.getAddress() != INT64_MAX)
-          outs() << format_hex(Probe.getAddress(), 8) << " Probe is in "
-                 << Addr2BlockNames[Probe.getAddress()] << "\n";
-      }
+    for (MCDecodedPseudoProbe &Probe : Address2ProbesMap) {
+      if (Probe.getAddress() == INT64_MAX)
+        outs() << "Deleted Probe: ";
+      else
+        outs() << "Address: " << format_hex(Probe.getAddress(), 8) << " ";
+      Probe.print(outs(), GUID2Func, true);
+      // print block name only if the probe is block type and undeleted.
+      if (Probe.isBlock() && Probe.getAddress() != INT64_MAX)
+        outs() << format_hex(Probe.getAddress(), 8) << " Probe is in "
+               << Addr2BlockNames[Probe.getAddress()] << "\n";
     }
     outs() << "=======================================\n";
   }
@@ -333,7 +318,7 @@ void PseudoProbeRewriter::encodePseudoProbes() {
       ProbeDecoder.getDummyInlineRoot();
   for (auto Child = Root.getChildren().begin();
        Child != Root.getChildren().end(); ++Child)
-    Inlinees[Child->first] = Child->second.get();
+    Inlinees[Child->getInlineSite()] = &*Child;
 
   for (auto Inlinee : Inlinees)
     // INT64_MAX is "placeholder" of unused callsite index field in the pair
@@ -359,25 +344,37 @@ void PseudoProbeRewriter::encodePseudoProbes() {
     EmitInt(Cur->Guid, 8);
     // Emit number of probes in this node
     uint64_t Deleted = 0;
-    for (MCDecodedPseudoProbe *&Probe : Cur->getProbes())
+    for (MCDecodedPseudoProbe *&Probe :
+         llvm::make_pointer_range(Cur->getProbes()))
       if (Probe->getAddress() == INT64_MAX)
         Deleted++;
     LLVM_DEBUG(dbgs() << "Deleted Probes:" << Deleted << "\n");
-    uint64_t ProbesSize = Cur->getProbes().size() - Deleted;
+    size_t InjectedProbes = ProbeDecoder.getNumInjectedProbes(Cur);
+    uint64_t ProbesSize = Cur->getProbes().size() - Deleted + InjectedProbes;
     EmitULEB128IntValue(ProbesSize);
     // Emit number of direct inlinees
     EmitULEB128IntValue(Cur->getChildren().size());
     // Emit probes in this group
-    for (MCDecodedPseudoProbe *&Probe : Cur->getProbes()) {
+    for (MCDecodedPseudoProbe *&Probe :
+         llvm::make_pointer_range(Cur->getProbes())) {
       if (Probe->getAddress() == INT64_MAX)
         continue;
       EmitDecodedPseudoProbe(Probe);
       LastProbe = Probe;
     }
+    if (InjectedProbes) {
+      for (MCDecodedPseudoProbe *&Probe :
+           llvm::make_pointer_range(ProbeDecoder.getInjectedProbes(Cur))) {
+        if (Probe->getAddress() == INT64_MAX)
+          continue;
+        EmitDecodedPseudoProbe(Probe);
+        LastProbe = Probe;
+      }
+    }
 
     for (auto Child = Cur->getChildren().begin();
          Child != Cur->getChildren().end(); ++Child)
-      Inlinees[Child->first] = Child->second.get();
+      Inlinees[Child->getInlineSite()] = &*Child;
     for (const auto &Inlinee : Inlinees) {
       assert(Cur->Guid != 0 && "non root tree node must have nonzero Guid");
       NextNodes.push_back({std::get<1>(Inlinee.first), Inlinee.second});