LLVM and SPIRV-LLVM-Translator pulldown (WW13 2024)

.github/CODEOWNERS

@@ -47,6 +47,12 @@ sycl/test-e2e/Plugin/dll-detach-order.cpp @intel/llvm-reviewers-runtime
 sycl/plugins/**/cuda/ @intel/llvm-reviewers-cuda
 sycl/plugins/**/hip/ @intel/llvm-reviewers-cuda
 
+# Transform Dialect in MLIR.
+/mlir/include/mlir/Dialect/Transform/* @ftynse @nicolasvasilache
+/mlir/lib/Dialect/Transform/* @ftynse @nicolasvasilache
+/mlir/**/*TransformOps* @ftynse @nicolasvasilache
+
+
 # CUDA specific runtime implementations
 sycl/include/sycl/ext/oneapi/experimental/cuda/ @intel/llvm-reviewers-cuda
 

.github/workflows/issue-write.yml

@@ -0,0 +1,128 @@
+name: Comment on an issue
+
+on:
+  workflow_run:
+    workflows: ["Check code formatting"]
+    types:
+      - completed
+
+permissions:
+  contents: read
+
+jobs:
+  pr-comment:
+    runs-on: ubuntu-latest
+    permissions:
+      pull-requests: write
+    if: >
+      github.event.workflow_run.event == 'pull_request'
+    steps:
+      - name: 'Download artifact'
+        uses: actions/download-artifact@6b208ae046db98c579e8a3aa621ab581ff575935 # v4.1.1
+        with:
+          github-token: ${{ secrets.ISSUE_WRITE_DOWNLOAD_ARTIFACT }}
+          run-id: ${{ github.event.workflow_run.id }}
+          name: workflow-args
+
+      - name: 'Comment on PR'
+        uses: actions/github-script@v3
+        with:
+          github-token: ${{ secrets.GITHUB_TOKEN }}
+          script: |
+            var fs = require('fs');
+            const comments = JSON.parse(fs.readFileSync('./comments'));
+            if (!comments) {
+              return;
+            }
+
+            let runInfo = await github.actions.getWorkflowRun({
+              owner: context.repo.owner,
+              repo: context.repo.repo,
+              run_id: context.payload.workflow_run.id
+            });
+
+            console.log(runInfo);
+
+
+            // Query to find the number of the pull request that triggered this job.
+            // The associated pull requests are based off of the branch name, so if
+            // you create a pull request for a branch, close it, and then create
+            // another pull request with the same branch, then this query will return
+            // two associated pull requests.  This is why we have to fetch all the
+            // associated pull requests and then iterate through them to find the
+            // one that is open.
+            const gql_query = `
+              query($repo_owner : String!, $repo_name : String!, $branch: String!) {
+                repository(owner: $repo_owner, name: $repo_name) {
+                  ref (qualifiedName: $branch) {
+                    associatedPullRequests(first: 100) {
+                      nodes {
+                        baseRepository {
+                          owner {
+                            login
+                          }
+                        }
+                        number
+                        state
+                      }
+                    }
+                  }
+                }
+              }
+            `
+            const gql_variables = {
+              repo_owner: runInfo.data.head_repository.owner.login,
+              repo_name: runInfo.data.head_repository.name,
+              branch: runInfo.data.head_branch
+            }
+            const gql_result = await github.graphql(gql_query, gql_variables);
+            console.log(gql_result);
+            console.log(gql_result.repository.ref.associatedPullRequests.nodes);
+
+            var pr_number = 0;
+            gql_result.repository.ref.associatedPullRequests.nodes.forEach((pr) => {
+              if (pr.baseRepository.owner.login = context.repo.owner && pr.state == 'OPEN') {
+                pr_number = pr.number;
+              }
+            });
+            if (pr_number == 0) {
+              console.log("Error retrieving pull request number");
+              return;
+            }
+
+            await comments.forEach(function (comment) {
+              if (comment.id) {
+                // Security check: Ensure that this comment was created by
+                // the github-actions bot, so a malicious input won't overwrite
+                // a user's comment.
+                github.issues.getComment({
+                  owner: context.repo.owner,
+                  repo: context.repo.repo,
+                  comment_id: comment.id
+                }).then((old_comment) => {
+                  console.log(old_comment);
+                  if (old_comment.data.user.login != "github-actions[bot]") {
+                    console.log("Invalid comment id: " + comment.id);
+                    return;
+                  }
+                  github.issues.updateComment({
+                    owner: context.repo.owner,
+                    repo: context.repo.repo,
+                    issue_number: pr_number,
+                    comment_id: comment.id,
+                    body: comment.body
+                  });
+                });
+              } else {
+                github.issues.createComment({
+                  owner: context.repo.owner,
+                  repo: context.repo.repo,
+                  issue_number: pr_number,
+                  body: comment.body
+                });
+              }
+            });
+
+      - name: Dump comments file
+        if: always()
+        run: cat comments

.github/workflows/pr-code-format.yml

@@ -1,15 +1,12 @@
 name: "Check code formatting"
 on:
-  pull_request_target:
+  pull_request:
     branches:
       - main
       - sycl
       - sycl-devops-pr/**
       - sycl-rel-**
 
-permissions:
-  pull-requests: write
-
 jobs:
   code_formatter:
     runs-on: ubuntu-latest
@@ -34,12 +31,13 @@ jobs:
           separator: ","
           skip_initial_fetch: true
 
-      # We need to make sure that we aren't executing/using any code from the
-      # PR for security reasons as we're using pull_request_target. Checkout
-      # the target branch with the necessary files.
+      # We need to pull the script from the main branch, so that we ensure
+      # we get the latest version of this script.
       - name: Fetch code formatting utils
         uses: actions/checkout@v4
         with:
+          reository: ${{ github.repository }}
+          ref: ${{ github.base_ref }}
           sparse-checkout: |
             llvm/utils/git/requirements_formatting.txt
             llvm/utils/git/code-format-helper.py
@@ -78,10 +76,20 @@ jobs:
         # to take advantage of the new --diff_from_common_commit option
         # explicitly in code-format-helper.py and not have to diff starting at
         # the merge base.
+        # Create an empty comments file so the pr-write job doesn't fail.
         run: |
+          echo "[]" > comments &&
           python ./code-format-tools/llvm/utils/git/code-format-helper.py \
+            --write-comment-to-file \
             --token ${{ secrets.GITHUB_TOKEN }} \
             --issue-number $GITHUB_PR_NUMBER \
             --start-rev $(git merge-base $START_REV $END_REV) \
             --end-rev $END_REV \
             --changed-files "$CHANGED_FILES"
+
+      - uses: actions/upload-artifact@26f96dfa697d77e81fd5907df203aa23a56210a8 #v4.3.0
+        if: always()
+        with:
+          name: workflow-args
+          path: |
+            comments

bolt/CMakeLists.txt

@@ -45,9 +45,9 @@ if (BOLT_ENABLE_RUNTIME)
   execute_process(COMMAND ls /proc/self/map_files
     RESULT_VARIABLE LS OUTPUT_QUIET ERROR_QUIET)
   if (LS)
-    set(BOLT_ENABLE_RUNTIME OFF)
     message(WARNING
-      "BOLT runtime is disabled as /proc/self/map_files is unreadable.")
+      "BOLT runtime may not be able to read /proc/self/map_files. Please use
+      `--instrumentation-binpath <path-to-instrumented-binary>` option.")
   endif()
 endif()
 

bolt/docs/BAT.md

@@ -14,9 +14,8 @@ binary onto the original binary.
 # Usage
 `--enable-bat` flag controls the generation of BAT section. Sampled profile
 needs to be passed along with the optimized binary containing BAT section to
-`perf2bolt` which reads BAT section and produces fdata profile for the original
-binary. Note that YAML profile generation is not supported since BAT doesn't
-contain the metadata for input functions.
+`perf2bolt` which reads BAT section and produces profile for the original
+binary.
 
 # Internals
 ## Section contents
@@ -43,21 +42,21 @@ and [BoltAddressTranslation.cpp](/bolt/lib/Profile/BoltAddressTranslation.cpp).
 ### Layout
 The general layout is as follows:
 ```
-Hot functions table header
-|------------------|
-|  Function entry  |
-| |--------------| |
-| | OutOff InOff | |
-| |--------------| |
-~~~~~~~~~~~~~~~~~~~~
+Hot functions table
+Cold functions table
 
-Cold functions table header
+Functions table:
 |------------------|
 |  Function entry  |
-| |--------------| |
-| | OutOff InOff | |
-| |--------------| |
-~~~~~~~~~~~~~~~~~~~~
+|                  |
+|     Address      |
+|   translation    |
+|      table       |
+|                  |
+| Secondary entry  |
+|      points      |
+|------------------|
+
 ```
 
 ### Functions table
@@ -75,28 +74,43 @@ internal offsets, and between hot and cold fragments, to better spread deltas
 and save space.
 
 Hot indices are delta encoded, implicitly starting at zero.
-| Entry  | Encoding | Description |
-| ------ | ------| ----------- |
-| `Address` | Continuous, Delta, ULEB128 | Function address in the output binary |
-| `HotIndex` | Delta, ULEB128 | Cold functions only: index of corresponding hot function in hot functions table |
-| `NumEntries` | ULEB128 | Number of address translation entries for a function |
-| `EqualElems` | ULEB128 | Hot functions only: number of equal offsets in the beginning of a function |
-| `BranchEntries` | Bitmask, `alignTo(EqualElems, 8)` bits | Hot functions only: if `EqualElems` is non-zero, bitmask denoting entries with `BRANCHENTRY` bit |
+| Entry  | Encoding | Description | Hot/Cold |
+| ------ | ------| ----------- | ------ |
+| `Address` | Continuous, Delta, ULEB128 | Function address in the output binary | Both |
+| `HotIndex` | Delta, ULEB128 | Index of corresponding hot function in hot functions table | Cold |
+| `FuncHash` | 8b | Function hash for input function | Hot |
+| `NumBlocks` | ULEB128 | Number of basic blocks in the original function | Hot |
+| `NumSecEntryPoints` | ULEB128 | Number of secondary entry points in the original function | Hot |
+| `NumEntries` | ULEB128 | Number of address translation entries for a function | Both |
+| `EqualElems` | ULEB128 | Number of equal offsets in the beginning of a function | Hot |
+| `BranchEntries` | Bitmask, `alignTo(EqualElems, 8)` bits | If `EqualElems` is non-zero, bitmask denoting entries with `BRANCHENTRY` bit | Hot |
 
-Function header is followed by `EqualElems` offsets (hot functions only) and
-`NumEntries-EqualElems` (`NumEntries` for cold functions) pairs of offsets for
-current function.
+Function header is followed by *Address Translation Table* with `NumEntries`
+total entries, and *Secondary Entry Points* table with `NumSecEntryPoints`
+entries (hot functions only).
 
 ### Address translation table
 Delta encoding means that only the difference with the previous corresponding
 entry is encoded. Input offsets implicitly start at zero.
-| Entry  | Encoding | Description |
-| ------ | ------| ----------- |
-| `OutputOffset` | Continuous, Delta, ULEB128 | Function offset in output binary |
-| `InputOffset` | Optional, Delta, SLEB128 | Function offset in input binary with `BRANCHENTRY` LSB bit |
+| Entry  | Encoding | Description | Branch/BB |
+| ------ | ------| ----------- | ------ |
+| `OutputOffset` | Continuous, Delta, ULEB128 | Function offset in output binary | Both |
+| `InputOffset` | Optional, Delta, SLEB128 | Function offset in input binary with `BRANCHENTRY` LSB bit | Both |
+| `BBHash` | Optional, 8b | Basic block hash in input binary | BB |
+| `BBIdx`  | Optional, Delta, ULEB128 | Basic block index in input binary | BB |
+
+For hot fragments, the table omits the first `EqualElems` input offsets
+where the input offset equals output offset.
 
 `BRANCHENTRY` bit denotes whether a given offset pair is a control flow source
 (branch or call instruction). If not set, it signifies a control flow target
 (basic block offset).
 `InputAddr` is omitted for equal offsets in input and output function. In this
 case, `BRANCHENTRY` bits are encoded separately in a `BranchEntries` bitvector.
+
+### Secondary Entry Points table
+The table is emitted for hot fragments only. It contains `NumSecEntryPoints`
+offsets denoting secondary entry points, delta encoded, implicitly starting at zero.
+| Entry | Encoding | Description |
+| ----- | -------- | ----------- |
+| `SecEntryPoint` | Delta, ULEB128 | Secondary entry point offset |

bolt/include/bolt/Core/DebugNames.h

@@ -68,6 +68,16 @@ class DWARF5AcceleratorTable {
   std::unique_ptr<DebugBufferVector> releaseBuffer() {
     return std::move(FullTableBuffer);
   }
+  /// Adds a DIE that is referenced across CUs.
+  void addCrossCUDie(const DIE *Die) {
+    CrossCUDies.insert({Die->getOffset(), Die});
+  }
+  /// Returns true if the DIE can generate an entry for a cross cu reference.
+  /// This only checks TAGs of a DIE because when this is invoked DIE might not
+  /// be fully constructed.
+  bool canGenerateEntryWithCrossCUReference(
+      const DWARFUnit &Unit, const DIE &Die,
+      const DWARFAbbreviationDeclaration::AttributeSpec &AttrSpec);
 
 private:
   BinaryContext &BC;
@@ -128,6 +138,7 @@ class DWARF5AcceleratorTable {
   llvm::DenseMap<uint64_t, uint32_t> CUOffsetsToPatch;
   // Contains a map of Entry ID to Entry relative offset.
   llvm::DenseMap<uint64_t, uint32_t> EntryRelativeOffsets;
+  llvm::DenseMap<uint64_t, const DIE *> CrossCUDies;
   /// Adds Unit to either CUList, LocalTUList or ForeignTUList.
   /// Input Unit being processed, and DWO ID if Unit is being processed comes
   /// from a DWO section.

bolt/include/bolt/Core/MCPlus.h

@@ -73,6 +73,7 @@ class MCAnnotation {
     kOffset,              /// Offset in the function.
     kLabel,               /// MCSymbol pointing to this instruction.
     kSize,                /// Size of the instruction.
+    kDynamicBranch,       /// Jit instruction patched at runtime.
     kGeneric              /// First generic annotation.
   };
 

bolt/include/bolt/Core/MCPlusBuilder.h

@@ -1199,6 +1199,16 @@ class MCPlusBuilder {
   /// Set instruction size.
   void setSize(MCInst &Inst, uint32_t Size) const;
 
+  /// Check if the branch instruction could be modified at runtime.
+  bool isDynamicBranch(const MCInst &Inst) const;
+
+  /// Return ID for runtime-modifiable instruction.
+  std::optional<uint32_t> getDynamicBranchID(const MCInst &Inst) const;
+
+  /// Mark instruction as a dynamic branch, i.e. a branch that can be
+  /// overwritten at runtime.
+  void setDynamicBranch(MCInst &Inst, uint32_t ID) const;
+
   /// Return MCSymbol that represents a target of this instruction at a given
   /// operand number \p OpNum. If there's no symbol associated with
   /// the operand - return nullptr.
@@ -1688,6 +1698,13 @@ class MCPlusBuilder {
     llvm_unreachable("not implemented");
   }
 
+  /// Create long conditional branch with a target-specific conditional code
+  /// \p CC.
+  virtual void createLongCondBranch(MCInst &Inst, const MCSymbol *Target,
+                                    unsigned CC, MCContext *Ctx) const {
+    llvm_unreachable("not implemented");
+  }
+
   /// Reverses the branch condition in Inst and update its taken target to TBB.
   ///
   /// Returns true on success.