LLVM and SPIRV-LLVM-Translator pulldown (WW45 2024)

.ci/generate-buildkite-pipeline-premerge

@@ -191,14 +191,17 @@ function keep-modified-projects() {
 }
 
 function check-targets() {
+  # Do not use "check-all" here because if there is "check-all" plus a
+  # project specific target like "check-clang", that project's tests
+  # will be run twice.
   projects=${@}
   for project in ${projects}; do
     case ${project} in
     clang-tools-extra)
       echo "check-clang-tools"
     ;;
     compiler-rt)
-      echo "check-all"
+      echo "check-compiler-rt"
     ;;
     cross-project-tests)
       echo "check-cross-project"
@@ -216,10 +219,10 @@ function check-targets() {
       echo "check-lldb"
     ;;
     pstl)
-      echo "check-all"
+      # Currently we do not run pstl tests in CI.
     ;;
     libclc)
-      echo "check-all"
+      # Currently there is no testing for libclc.
     ;;
     *)
       echo "check-${project}"

.gitattributes

@@ -1,4 +1,5 @@
 .github/CODEOWNERS merge=ours
+
 libcxx/src/**/*.cpp     merge=libcxx-reformat
 libcxx/include/**/*.h   merge=libcxx-reformat
 

.github/new-prs-labeler.yml

@@ -69,7 +69,7 @@ PGO:
   - llvm/**/llvm-profdata/**/*
   - llvm/**/llvm-profgen/**/*
 
-vectorization:
+vectorizers:
   - llvm/lib/Transforms/Vectorize/**/*
   - llvm/include/llvm/Transforms/Vectorize/**/*
 

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

@@ -193,17 +193,24 @@ jobs:
             **/crash_diagnostics/*
 
   macos:
-    runs-on: macos-14
     needs: [ stage1 ]
     strategy:
-      fail-fast: true
+      fail-fast: false
       matrix:
-        config: [
-          generic-cxx03,
-          generic-cxx23,
-          generic-modules,
-          apple-configuration
-        ]
+        include:
+        - config: generic-cxx03
+          os: macos-latest
+        - config: generic-cxx23
+          os: macos-latest
+        - config: generic-modules
+          os: macos-latest
+        - config: apple-configuration
+          os: macos-latest
+        - config: apple-system
+          os: macos-13
+        - config: apple-system-hardened
+          os: macos-13
+    runs-on: ${{ matrix.os }}
     steps:
       - uses: actions/checkout@v4
       - uses: maxim-lobanov/setup-xcode@v1

.github/workflows/release-binaries-all.yml

@@ -43,6 +43,7 @@ on:
       - '.github/workflows/release-binaries.yml'
       - '.github/workflows/release-binaries-setup-stage/*'
       - '.github/workflows/release-binaries-save-stage/*'
+      - 'clang/cmake/caches/Release.cmake'
 
 concurrency:
   group: ${{ github.workflow }}-${{ github.event.pull_request.number || 'dispatch' }}

.github/workflows/release-binaries.yml

@@ -328,7 +328,7 @@ jobs:
       run: |
         # Build some of the mlir tools that take a long time to link
         if [ "${{ needs.prepare.outputs.build-flang }}" = "true" ]; then
-          ninja -C ${{ steps.setup-stage.outputs.build-prefix }}/build/tools/clang/stage2-bins/ -j2 flang-new bbc
+          ninja -C ${{ steps.setup-stage.outputs.build-prefix }}/build/tools/clang/stage2-bins/ -j2 flang bbc
         fi
         ninja -C ${{ steps.setup-stage.outputs.build-prefix }}/build/tools/clang/stage2-bins/ \
             mlir-bytecode-parser-fuzzer \
@@ -420,6 +420,14 @@ jobs:
       attestations: write # For artifact attestations
 
     steps:
+    - name: Checkout Release Scripts
+      uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11 # v4.1.1
+      with:
+        sparse-checkout: |
+          llvm/utils/release/github-upload-release.py
+          llvm/utils/git/requirements.txt
+        sparse-checkout-cone-mode: false
+
     - name: 'Download artifact'
       uses: actions/download-artifact@6b208ae046db98c579e8a3aa621ab581ff575935 # v4.1.1
       with:
@@ -442,14 +450,6 @@ jobs:
         name: ${{ needs.prepare.outputs.release-binary-filename }}-attestation
         path: ${{ needs.prepare.outputs.release-binary-filename }}.jsonl
 
-    - name: Checkout Release Scripts
-      uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11 # v4.1.1
-      with:
-        sparse-checkout: |
-          llvm/utils/release/github-upload-release.py
-          llvm/utils/git/requirements.txt
-        sparse-checkout-cone-mode: false
-
     - name: Install Python Requirements
       run: |
         pip install --require-hashes -r ./llvm/utils/git/requirements.txt

.github/workflows/release-documentation.yml

@@ -73,17 +73,20 @@ jobs:
           ref: main
           fetch-depth: 0
           path: www-releases
+          persist-credentials: false
 
       - name: Upload Release Notes
         if: env.upload
         env:
-          WWW_RELEASES_TOKEN: ${{ secrets.WWW_RELEASES_TOKEN }}
+          GH_TOKEN: ${{ secrets.WWW_RELEASES_TOKEN }}
         run: |
-          mkdir -p ../www-releases/${{ inputs.release-version }}
-          mv ./docs-build/html-export/* ../www-releases/${{ inputs.release-version }}
-          cd ../www-releases
+          mkdir -p www-releases/${{ inputs.release-version }}
+          mv ./docs-build/html-export/* www-releases/${{ inputs.release-version }}
+          cd www-releases
+          git checkout -b ${{ inputs.release-version }}
           git add ${{ inputs.release-version }}
           git config user.email "[email protected]"
           git config user.name "llvmbot"
           git commit -a -m "Add ${{ inputs.release-version }} documentation"
-          git push "https://[email protected]/${{ github.repository_owner }}/www-releases" main:main
+          git push --force  "https://[email protected]/llvmbot/www-releases.git" HEAD:refs/heads/${{ inputs.release-version }}
+          gh pr create -f -B main -H ${{ inputs.release-version }} -R llvmbot/www-releases

.gitignore

@@ -55,6 +55,7 @@ autoconf/autom4te.cache
 /CMakeSettings.json
 # CLion project configuration
 /.idea
+/cmake-build*
 
 #==============================================================================#
 # Directories to ignore (do not add trailing '/'s, they skip symlinks).

bolt/include/bolt/Core/BinaryBasicBlock.h

@@ -19,6 +19,7 @@
 #include "bolt/Core/MCPlus.h"
 #include "llvm/ADT/GraphTraits.h"
 #include "llvm/ADT/StringRef.h"
+#include "llvm/Config/llvm-config.h" // for LLVM_ON_UNIX
 #include "llvm/MC/MCInst.h"
 #include "llvm/MC/MCSymbol.h"
 #include "llvm/Support/ErrorOr.h"

bolt/include/bolt/Core/BinaryContext.h

@@ -71,14 +71,15 @@ struct SegmentInfo {
   uint64_t FileOffset;        /// Offset in the file.
   uint64_t FileSize;          /// Size in file.
   uint64_t Alignment;         /// Alignment of the segment.
+  bool IsExecutable;          /// Is the executable bit set on the Segment?
 
   void print(raw_ostream &OS) const {
-    OS << "SegmentInfo { Address: 0x"
-       << Twine::utohexstr(Address) << ", Size: 0x"
-       << Twine::utohexstr(Size) << ", FileOffset: 0x"
+    OS << "SegmentInfo { Address: 0x" << Twine::utohexstr(Address)
+       << ", Size: 0x" << Twine::utohexstr(Size) << ", FileOffset: 0x"
        << Twine::utohexstr(FileOffset) << ", FileSize: 0x"
        << Twine::utohexstr(FileSize) << ", Alignment: 0x"
-       << Twine::utohexstr(Alignment) << "}";
+       << Twine::utohexstr(Alignment) << ", " << (IsExecutable ? "x" : " ")
+       << "}";
   };
 };
 

bolt/include/bolt/Core/BinaryData.h

@@ -226,7 +226,6 @@ inline raw_ostream &operator<<(raw_ostream &OS,
     Sep = ",\n        ";
     TotalCount += AccessInfo.Count;
   }
-  SS.flush();
 
   OS << TotalCount << " total counts : " << TempString;
   return OS;

bolt/include/bolt/Core/BinaryFunction.h

@@ -117,7 +117,6 @@ inline raw_ostream &operator<<(raw_ostream &OS,
     TotalCount += CSP.Count;
     TotalMispreds += CSP.Mispreds;
   }
-  SS.flush();
 
   OS << TotalCount << " (" << TotalMispreds << " misses) :" << TempString;
   return OS;

bolt/include/bolt/Core/DIEBuilder.h

@@ -314,7 +314,7 @@ class DIEBuilder {
 
     BC.errs()
         << "BOLT-ERROR: unable to find TypeUnit for Type Unit at offset 0x"
-        << DU.getOffset() << "\n";
+        << Twine::utohexstr(DU.getOffset()) << "\n";
     return nullptr;
   }
 

bolt/include/bolt/Passes/ContinuityStats.h

@@ -0,0 +1,61 @@
+//===- bolt/Passes/ContinuityStats.h ----------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// This pass checks how well the BOLT input profile satisfies the following
+// "CFG continuity" property of a perfect profile:
+//
+//        Each positive-execution-count block in the function’s CFG
+//        should be *reachable* from a positive-execution-count function
+//        entry block through a positive-execution-count path.
+//
+// More specifically, for each of the hottest 1000 functions, the pass
+// calculates the function’s fraction of basic block execution counts
+// that is *unreachable*. It then reports the 95th percentile of the
+// distribution of the 1000 unreachable fractions in a single BOLT-INFO line.
+// The smaller the reported value is, the better the BOLT profile
+// satisfies the CFG continuity property.
+
+// The default value of 1000 above can be changed via the hidden BOLT option
+// `-num-functions-for-continuity-check=[N]`.
+// If more detailed stats are needed, `-v=1` can be used: the hottest N
+// functions will be grouped into 5 equally-sized buckets, from the hottest
+// to the coldest; for each bucket, various summary statistics of the
+// distribution of the unreachable fractions and the raw unreachable execution
+// counts will be reported.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef BOLT_PASSES_CONTINUITYSTATS_H
+#define BOLT_PASSES_CONTINUITYSTATS_H
+
+#include "bolt/Passes/BinaryPasses.h"
+#include <vector>
+
+namespace llvm {
+
+class raw_ostream;
+
+namespace bolt {
+class BinaryContext;
+
+/// Compute and report to the user the function CFG continuity quality
+class PrintContinuityStats : public BinaryFunctionPass {
+public:
+  explicit PrintContinuityStats(const cl::opt<bool> &PrintPass)
+      : BinaryFunctionPass(PrintPass) {}
+
+  bool shouldOptimize(const BinaryFunction &BF) const override;
+  const char *getName() const override { return "continuity-stats"; }
+  bool shouldPrint(const BinaryFunction &) const override { return false; }
+  Error runOnFunctions(BinaryContext &BC) override;
+};
+
+} // namespace bolt
+} // namespace llvm
+
+#endif // BOLT_PASSES_CONTINUITYSTATS_H

bolt/include/bolt/Profile/BoltAddressTranslation.h

@@ -141,15 +141,13 @@ class BoltAddressTranslation {
                          uint64_t FuncOutputAddress) const;
 
   /// Write the serialized address translation table for a function.
-  template <bool Cold>
-  void writeMaps(std::map<uint64_t, MapTy> &Maps, uint64_t &PrevAddress,
-                 raw_ostream &OS);
+  template <bool Cold> void writeMaps(uint64_t &PrevAddress, raw_ostream &OS);
 
   /// Read the serialized address translation table for a function.
   /// Return a parse error if failed.
   template <bool Cold>
-  void parseMaps(std::vector<uint64_t> &HotFuncs, uint64_t &PrevAddress,
-                 DataExtractor &DE, uint64_t &Offset, Error &Err);
+  void parseMaps(uint64_t &PrevAddress, DataExtractor &DE, uint64_t &Offset,
+                 Error &Err);
 
   /// Returns the bitmask with set bits corresponding to indices of BRANCHENTRY
   /// entries in function address translation map.
@@ -161,6 +159,9 @@ class BoltAddressTranslation {
 
   std::map<uint64_t, MapTy> Maps;
 
+  /// Ordered vector with addresses of hot functions.
+  std::vector<uint64_t> HotFuncs;
+
   /// Map a function to its basic blocks count
   std::unordered_map<uint64_t, size_t> NumBasicBlocksMap;
 

bolt/include/bolt/Rewrite/RewriteInstance.h

@@ -510,12 +510,11 @@ class RewriteInstance {
   };
 
   /// Different types of X86-64 PLT sections.
-  const PLTSectionInfo X86_64_PLTSections[4] = {
-      { ".plt", 16 },
-      { ".plt.got", 8 },
-      { ".plt.sec", 8 },
-      { nullptr, 0 }
-  };
+  const PLTSectionInfo X86_64_PLTSections[5] = {{".plt", 16},
+                                                {".plt.got", 8},
+                                                {".plt.sec", 8},
+                                                {".iplt", 16},
+                                                {nullptr, 0}};
 
   /// AArch64 PLT sections.
   const PLTSectionInfo AArch64_PLTSections[4] = {

bolt/lib/Core/BinaryContext.cpp

@@ -1294,8 +1294,8 @@ bool BinaryContext::handleAArch64Veneer(uint64_t Address, bool MatchOnly) {
     Veneer->getOrCreateLocalLabel(Address);
     Veneer->setMaxSize(TotalSize);
     Veneer->updateState(BinaryFunction::State::Disassembled);
-    LLVM_DEBUG(dbgs() << "BOLT-DEBUG: handling veneer function at 0x" << Address
-                      << "\n");
+    LLVM_DEBUG(dbgs() << "BOLT-DEBUG: handling veneer function at 0x"
+                      << Twine::utohexstr(Address) << "\n");
     return true;
   };
 
@@ -2021,6 +2021,9 @@ BinaryContext::getBaseAddressForMapping(uint64_t MMapAddress,
   // Find a segment with a matching file offset.
   for (auto &KV : SegmentMapInfo) {
     const SegmentInfo &SegInfo = KV.second;
+    // Only consider executable segments.
+    if (!SegInfo.IsExecutable)
+      continue;
     // FileOffset is got from perf event,
     // and it is equal to alignDown(SegInfo.FileOffset, pagesize).
     // If the pagesize is not equal to SegInfo.Alignment.

bolt/lib/Core/BinaryFunction.cpp

@@ -165,6 +165,12 @@ bool shouldPrint(const BinaryFunction &Function) {
     }
   }
 
+  std::optional<StringRef> Origin = Function.getOriginSectionName();
+  if (Origin && llvm::any_of(opts::PrintOnly, [&](const std::string &Name) {
+        return Name == *Origin;
+      }))
+    return true;
+
   return false;
 }
 
@@ -3678,9 +3684,8 @@ BinaryFunction::BasicBlockListType BinaryFunction::dfs() const {
     BinaryBasicBlock *BB = Stack.top();
     Stack.pop();
 
-    if (Visited.find(BB) != Visited.end())
+    if (!Visited.insert(BB).second)
       continue;
-    Visited.insert(BB);
     DFS.push_back(BB);
 
     for (BinaryBasicBlock *SuccBB : BB->landing_pads()) {
@@ -3873,11 +3878,8 @@ void BinaryFunction::disambiguateJumpTables(
       JumpTable *JT = getJumpTable(Inst);
       if (!JT)
         continue;
-      auto Iter = JumpTables.find(JT);
-      if (Iter == JumpTables.end()) {
-        JumpTables.insert(JT);
+      if (JumpTables.insert(JT).second)
         continue;
-      }
       // This instruction is an indirect jump using a jump table, but it is
       // using the same jump table of another jump. Try all our tricks to
       // extract the jump table symbol and make it point to a new, duplicated JT

bolt/lib/Core/DIEBuilder.cpp

@@ -57,11 +57,9 @@ getDWOName(llvm::DWARFUnit &CU,
          "DW_AT_dwo_name/DW_AT_GNU_dwo_name does not exist.");
   if (DwarfOutputPath) {
     DWOName = std::string(sys::path::filename(DWOName));
-    auto Iter = NameToIndexMap.find(DWOName);
-    if (Iter == NameToIndexMap.end())
-      Iter = NameToIndexMap.insert({DWOName, 0}).first;
-    DWOName.append(std::to_string(Iter->second));
-    ++Iter->second;
+    uint32_t &Index = NameToIndexMap[DWOName];
+    DWOName.append(std::to_string(Index));
+    ++Index;
   }
   DWOName.append(".dwo");
   return DWOName;