Add read and write interceptors

.github/workflows/issue-write.yml

@@ -24,14 +24,21 @@ jobs:
         github.event.workflow_run.conclusion == 'failure'
       )
     steps:
+      - name: Fetch Sources
+        uses: actions/checkout@v4
+        with:
+          sparse-checkout: |
+            .github/workflows/unprivileged-download-artifact/action.yml
+          sparse-checkout-cone-mode: false
       - name: 'Download artifact'
-        uses: actions/download-artifact@6b208ae046db98c579e8a3aa621ab581ff575935 # v4.1.1
+        uses: ./.github/workflows/unprivileged-download-artifact
+        id: download-artifact
         with:
-          github-token: ${{ secrets.ISSUE_WRITE_DOWNLOAD_ARTIFACT }}
           run-id: ${{ github.event.workflow_run.id }}
-          name: workflow-args
+          artifact-name: workflow-args
 
       - name: 'Comment on PR'
+        if: steps.download-artifact.outputs.artifact-id != ''
         uses: actions/github-script@v3
         with:
           github-token: ${{ secrets.GITHUB_TOKEN }}
@@ -144,5 +151,7 @@ jobs:
             });
 
       - name: Dump comments file
-        if: always()
+        if: >-
+          always() &&
+          steps.download-artifact.outputs.artifact-id != ''
         run: cat comments

.github/workflows/unprivileged-download-artifact/action.yml

@@ -0,0 +1,81 @@
+name: Unprivileged Download Artifact
+description: >-
+  Download artifacts from another workflow run without using an access token.
+inputs:
+  run-id:
+    description: >-
+      The run-id for the workflow run that you want to download the artifact
+      from.  If ommitted it will download the most recently created artifact
+      from the repo with the artifact-name.
+    required: false
+  artifact-name:
+    desciption: The name of the artifact to download.
+    required: true
+
+
+outputs:
+  filename:
+    description: >-
+      The filename of the downloaded artifact or the empty string if the
+      artifact was not found.
+    value: ${{ steps.download-artifact.outputs.filename }}
+  artifact-id:
+    description: "The id of the artifact being downloaded."
+    value: ${{ steps.artifact-url.outputs.id }}
+
+
+runs:
+  using: "composite"
+  steps:
+    - uses: actions/github-script@60a0d83039c74a4aee543508d2ffcb1c3799cdea #v7.0.1
+      id: artifact-url
+      with:
+        script: |
+          var response;
+          if (!"${{ inputs.run-id }}") {
+            response = await github.rest.actions.listArtifactsForRepo({
+              owner: context.repo.owner,
+              repo: context.repo.repo,
+              name: "${{ inputs.artifact-name }}"
+            })
+          } else {
+            response = await github.rest.actions.listWorkflowRunArtifacts({
+              owner: context.repo.owner,
+              repo: context.repo.repo,
+              run_id: "${{ inputs.run-id }}",
+              name: "${{ inputs.artifact-name }}"
+            })
+          }
+
+          console.log(response)
+
+          for (artifact of response.data.artifacts) {
+            console.log(artifact);
+          }
+
+          if (response.data.artifacts.length == 0) {
+            console.log("Could not find artifact ${{ inputs.artifact-name }} for workflow run ${{ inputs.run-id }}")
+            return;
+          }
+
+          const url_response = await github.rest.actions.downloadArtifact({
+            owner: context.repo.owner,
+            repo: context.repo.repo,
+            artifact_id: response.data.artifacts[0].id,
+            archive_format: "zip"
+          })
+
+          core.setOutput("url", url_response.url);
+          core.setOutput("id", response.data.artifacts[0].id);
+
+    - shell: bash
+      if: steps.artifact-url.outputs.url != ''
+      id: download-artifact
+      run: |
+        curl -L -o ${{ inputs.artifact-name }}.zip "${{ steps.artifact-url.outputs.url }}"
+        echo "filename=${{ inputs.artifact-name }}.zip" >> $GITHUB_OUTPUT
+
+    - shell: bash
+      if: steps.download-artifact.outputs.filename != ''
+      run: |
+        unzip ${{ steps.download-artifact.outputs.filename }}

bolt/docs/CommandLineArgumentReference.md

@@ -113,11 +113,6 @@
 
   Prints out offsets for abbrev and debug_info of Skeleton CUs that get patched.
 
-- `--deterministic-debuginfo`
-
-  Disables parallel execution of tasks that may produce nondeterministic debug
-  info
-
 - `--dot-tooltip-code`
 
   Add basic block instructions as tool tips on nodes
@@ -686,6 +681,10 @@
   threshold means fewer functions to process. E.g threshold of 90 means only top
   10 percent of functions with profile will be processed.
 
+- `--match-with-call-graph`
+
+  Match functions with call graph
+
 - `--memcpy1-spec=<func1,func2:cs1:cs2,func3:cs1,...>`
 
   List of functions with call sites for which to specialize memcpy() for size 1

bolt/docs/OptimizingClang.md

@@ -49,6 +49,7 @@ $ cd ${TOPLEV}/stage3
 $ CPATH=${TOPLEV}/stage2-prof-use-lto/install/bin/
 $ cmake -G Ninja ${TOPLEV}/llvm -DLLVM_TARGETS_TO_BUILD=X86 -DCMAKE_BUILD_TYPE=Release \
     -DCMAKE_C_COMPILER=$CPATH/clang -DCMAKE_CXX_COMPILER=$CPATH/clang++ \
+    -DLLVM_ENABLE_PROJECTS="clang" \
     -DLLVM_USE_LINKER=lld -DCMAKE_INSTALL_PREFIX=${TOPLEV}/stage3/install
 $ perf record -e cycles:u -j any,u -- ninja clang
 ```

bolt/docs/generate_doc.py

@@ -45,7 +45,7 @@ def parse_bolt_options(output):
         cleaned_line = line.strip()
 
         if cleaned_line.casefold() in map(str.casefold, section_headers):
-            if prev_section != None:  # Save last option from prev section
+            if prev_section is not None:  # Save last option from prev section
                 add_info(sections, current_section, option, description)
                 option, description = None, []
 
@@ -76,7 +76,7 @@ def parse_bolt_options(output):
                 description = [descr]
                 if option.startswith("--print") or option.startswith("--time"):
                     current_section = "BOLT printing options:"
-                elif prev_section != None:
+                elif prev_section is not None:
                     current_section = prev_section
             continue
 

bolt/include/bolt/Core/BinaryContext.h

@@ -32,6 +32,7 @@
 #include "llvm/MC/MCContext.h"
 #include "llvm/MC/MCObjectFileInfo.h"
 #include "llvm/MC/MCObjectWriter.h"
+#include "llvm/MC/MCPseudoProbe.h"
 #include "llvm/MC/MCSectionELF.h"
 #include "llvm/MC/MCSectionMachO.h"
 #include "llvm/MC/MCStreamer.h"
@@ -246,6 +247,9 @@ class BinaryContext {
   /// DWP Context.
   std::shared_ptr<DWARFContext> DWPContext;
 
+  /// Decoded pseudo probes.
+  std::shared_ptr<MCPseudoProbeDecoder> PseudoProbeDecoder;
+
   /// A map of DWO Ids to CUs.
   using DWOIdToCUMapType = std::unordered_map<uint64_t, DWARFUnit *>;
   DWOIdToCUMapType DWOCUs;
@@ -377,6 +381,15 @@ class BinaryContext {
     RtLibrary = std::move(Lib);
   }
 
+  const MCPseudoProbeDecoder *getPseudoProbeDecoder() const {
+    return PseudoProbeDecoder.get();
+  }
+
+  void setPseudoProbeDecoder(std::shared_ptr<MCPseudoProbeDecoder> Decoder) {
+    assert(!PseudoProbeDecoder && "Cannot set pseudo probe decoder twice.");
+    PseudoProbeDecoder = Decoder;
+  }
+
   /// Return BinaryFunction containing a given \p Address or nullptr if
   /// no registered function contains the \p Address.
   ///
@@ -431,6 +444,9 @@ class BinaryContext {
     return nullptr;
   }
 
+  /// Deregister JumpTable registered at a given \p Address and delete it.
+  void deleteJumpTable(uint64_t Address);
+
   unsigned getDWARFEncodingSize(unsigned Encoding) {
     if (Encoding == dwarf::DW_EH_PE_omit)
       return 0;
@@ -1436,10 +1452,7 @@ class BinaryContext {
     std::unique_ptr<MCObjectWriter> OW = MAB->createObjectWriter(OS);
     std::unique_ptr<MCStreamer> Streamer(TheTarget->createMCObjectStreamer(
         *TheTriple, *Ctx, std::unique_ptr<MCAsmBackend>(MAB), std::move(OW),
-        std::unique_ptr<MCCodeEmitter>(MCE), *STI,
-        /* RelaxAll */ false,
-        /* IncrementalLinkerCompatible */ false,
-        /* DWARFMustBeAtTheEnd */ false));
+        std::unique_ptr<MCCodeEmitter>(MCE), *STI));
     return Streamer;
   }
 

bolt/include/bolt/Core/BinaryFunction.h

@@ -416,6 +416,9 @@ class BinaryFunction {
   /// different parameters by every pass.
   mutable uint64_t Hash{0};
 
+  /// Function GUID assigned externally.
+  uint64_t GUID{0};
+
   /// For PLT functions it contains a symbol associated with a function
   /// reference. It is nullptr for non-PLT functions.
   const MCSymbol *PLTSymbol{nullptr};
@@ -2256,6 +2259,11 @@ class BinaryFunction {
   /// Returns the last computed hash value of the function.
   size_t getHash() const { return Hash; }
 
+  /// Returns the function GUID.
+  uint64_t getGUID() const { return GUID; }
+
+  void setGUID(uint64_t Id) { GUID = Id; }
+
   using OperandHashFuncTy =
       function_ref<typename std::string(const MCOperand &)>;
 

bolt/include/bolt/Core/MCPlusBuilder.h

@@ -58,6 +58,8 @@ enum class IndirectBranchType : char {
   POSSIBLE_PIC_JUMP_TABLE, /// Possibly a jump table for PIC.
   POSSIBLE_GOTO,           /// Possibly a gcc's computed goto.
   POSSIBLE_FIXED_BRANCH,   /// Possibly an indirect branch to a fixed location.
+  POSSIBLE_PIC_FIXED_BRANCH, /// Possibly an indirect jump to a fixed entry in a
+                             /// PIC jump table.
 };
 
 class MCPlusBuilder {
@@ -1474,12 +1476,11 @@ class MCPlusBuilder {
   /// will be set to the different components of the branch.  \p MemLocInstr
   /// is the instruction that loads up the indirect function pointer.  It may
   /// or may not be same as \p Instruction.
-  virtual IndirectBranchType
-  analyzeIndirectBranch(MCInst &Instruction, InstructionIterator Begin,
-                        InstructionIterator End, const unsigned PtrSize,
-                        MCInst *&MemLocInstr, unsigned &BaseRegNum,
-                        unsigned &IndexRegNum, int64_t &DispValue,
-                        const MCExpr *&DispExpr, MCInst *&PCRelBaseOut) const {
+  virtual IndirectBranchType analyzeIndirectBranch(
+      MCInst &Instruction, InstructionIterator Begin, InstructionIterator End,
+      const unsigned PtrSize, MCInst *&MemLocInstr, unsigned &BaseRegNum,
+      unsigned &IndexRegNum, int64_t &DispValue, const MCExpr *&DispExpr,
+      MCInst *&PCRelBaseOut, MCInst *&FixedEntryLoadInst) const {
     llvm_unreachable("not implemented");
     return IndirectBranchType::UNKNOWN;
   }

bolt/include/bolt/Profile/ProfileYAMLMapping.h

@@ -93,11 +93,36 @@ template <> struct MappingTraits<bolt::SuccessorInfo> {
   static const bool flow = true;
 };
 
+namespace bolt {
+struct PseudoProbeInfo {
+  llvm::yaml::Hex64 GUID;
+  uint64_t Index;
+  uint8_t Type;
+
+  bool operator==(const PseudoProbeInfo &Other) const {
+    return GUID == Other.GUID && Index == Other.Index;
+  }
+  bool operator!=(const PseudoProbeInfo &Other) const {
+    return !(*this == Other);
+  }
+};
+} // end namespace bolt
+
+template <> struct MappingTraits<bolt::PseudoProbeInfo> {
+  static void mapping(IO &YamlIO, bolt::PseudoProbeInfo &PI) {
+    YamlIO.mapRequired("guid", PI.GUID);
+    YamlIO.mapRequired("id", PI.Index);
+    YamlIO.mapRequired("type", PI.Type);
+  }
+
+  static const bool flow = true;
+};
 } // end namespace yaml
 } // end namespace llvm
 
 LLVM_YAML_IS_FLOW_SEQUENCE_VECTOR(llvm::yaml::bolt::CallSiteInfo)
 LLVM_YAML_IS_FLOW_SEQUENCE_VECTOR(llvm::yaml::bolt::SuccessorInfo)
+LLVM_YAML_IS_FLOW_SEQUENCE_VECTOR(llvm::yaml::bolt::PseudoProbeInfo)
 
 namespace llvm {
 namespace yaml {
@@ -111,6 +136,7 @@ struct BinaryBasicBlockProfile {
   uint64_t EventCount{0};
   std::vector<CallSiteInfo> CallSites;
   std::vector<SuccessorInfo> Successors;
+  std::vector<PseudoProbeInfo> PseudoProbes;
 
   bool operator==(const BinaryBasicBlockProfile &Other) const {
     return Index == Other.Index;
@@ -132,6 +158,8 @@ template <> struct MappingTraits<bolt::BinaryBasicBlockProfile> {
                        std::vector<bolt::CallSiteInfo>());
     YamlIO.mapOptional("succ", BBP.Successors,
                        std::vector<bolt::SuccessorInfo>());
+    YamlIO.mapOptional("pseudo_probes", BBP.PseudoProbes,
+                       std::vector<bolt::PseudoProbeInfo>());
   }
 };
 
@@ -151,6 +179,8 @@ struct BinaryFunctionProfile {
   llvm::yaml::Hex64 Hash{0};
   uint64_t ExecCount{0};
   std::vector<BinaryBasicBlockProfile> Blocks;
+  llvm::yaml::Hex64 GUID{0};
+  llvm::yaml::Hex64 PseudoProbeDescHash{0};
   bool Used{false};
 };
 } // end namespace bolt
@@ -164,6 +194,9 @@ template <> struct MappingTraits<bolt::BinaryFunctionProfile> {
     YamlIO.mapRequired("nblocks", BFP.NumBasicBlocks);
     YamlIO.mapOptional("blocks", BFP.Blocks,
                        std::vector<bolt::BinaryBasicBlockProfile>());
+    YamlIO.mapOptional("guid", BFP.GUID, (uint64_t)0);
+    YamlIO.mapOptional("pseudo_probe_desc_hash", BFP.PseudoProbeDescHash,
+                       (uint64_t)0);
   }
 };