Merge remote-tracking branch 'intel/sycl' into steffen/func_ptr_case

Author: steffenlarsen

.github/CODEOWNERS

@@ -223,3 +223,6 @@ llvm/test/Instrumentation/ThreadSanitizer/ @intel/dpcpp-sanitizers-review
 sycl/test-e2e/AddressSanitizer/ @intel/dpcpp-sanitizers-review
 sycl/test-e2e/MemorySanitizer/ @intel/dpcpp-sanitizers-review
 sycl/test-e2e/ThreadSanitizer/ @intel/dpcpp-sanitizers-review
+
+# ABI compatibility
+devops/compat_ci_exclude.sycl-rel-** @gmlueck @xtian-github

.github/workflows/sycl-linux-build.yml

@@ -296,6 +296,9 @@ jobs:
         cmake --build $GITHUB_WORKSPACE/build --target install-llvm-cov
         cmake --build $GITHUB_WORKSPACE/build --target install-llvm-profdata
         cmake --build $GITHUB_WORKSPACE/build --target install-compiler-rt
+        # This is required to perform the DeviceConfigFile consistency test, see
+        # sycl/test-e2e/Basic/device_config_file_consistency.cpp.
+        cmake --install $GITHUB_WORKSPACE/build --component DeviceConfigFile
     - name: Additional Install for "--shared-libs" build
       if: ${{ always() && !cancelled() && steps.build.conclusion == 'success' && contains(inputs.build_configure_extra_args, '--shared-libs') }}
       run: |

.github/workflows/sycl-ur-perf-benchmarking.yml

@@ -28,8 +28,18 @@ on:
           deployment branch will be used.
         required: false
         default: ''
+      save_name:
+        type: string
+        description: |
+          Specify a custom name to use for the benchmark result: If uploading
+          results, this will be the name used to refer results from the current
+          run.
+        required: false
+        default: ''
       upload_results:
         type: string # true/false: workflow_dispatch does not support booleans
+        description: |
+          Upload results to https://intel.github.io/llvm/benchmarks/.
         required: true
       runner:
         type: string
@@ -67,8 +77,14 @@ on:
           Leave both pr_no and commit_hash empty to use latest commit.
         required: false
         default: ''
+      save_name:
+        type: string
+        description: |
+          Name to use for the benchmark result:
+        required: false
+        default: ''
       upload_results:
-        description: 'Save and upload results'
+        description: 'Save and upload results (to https://intel.github.io/llvm/benchmarks)'
         type: choice
         options:
           - false
@@ -90,16 +106,61 @@ on:
 permissions: read-all
 
 jobs:
+  sanitize_inputs:
+    name: Sanitize inputs
+    runs-on: ubuntu-latest
+    env:
+      COMMIT_HASH: ${{ inputs.commit_hash }}
+      PR_NO: ${{ inputs.pr_no }}
+      SAVE_NAME: ${{ inputs.save_name }}
+    outputs:
+      benchmark_save_name: ${{ steps.sanitize.outputs.benchmark_save_name }}
+      build_ref: ${{ steps.sanitize.outputs.build_ref }}
+    steps:
+      - id: sanitize
+        run: |
+          # Validate user inputs:
+          # usage: check_if_nonempty <var> <regex to check var against> <err message>
+          check_nonempty() {
+            [ -z "$1" ] && return
+            if [ -z "$(echo "$1" | grep -P "$2")" ]; then
+              echo "$3"
+              exit 1
+            fi
+          }
+          check_nonempty "$COMMIT_HASH" '^[0-9a-f]{7,}$' "Bad commit hash (or hash short)."
+          check_nonempty "$PR_NO" '^[0-9]+$' "Bad PR number."
+          check_nonempty "$SAVE_NAME" '^[A-Za-z][A-Za-z0-9_-]+$' "Bad save name."
+
+          BENCHMARK_SAVE_NAME=""
+          BUILD_REF="${{ github.ref }}"
+          if [ -n "$SAVE_NAME" ]; then
+            BENCHMARK_SAVE_NAME="$(echo "$SAVE_NAME" | tr -cd 'A-Za-z0-9_-')"
+          fi;
+          if [ -n "$COMMIT_HASH" ]; then
+            echo "Using commit hash $COMMIT_HASH for build..."
+            BUILD_REF="$COMMIT_HASH"
+            shortened_commit="$(echo "$COMMIT_HASH" | cut -c 1-7)"
+            [ -z "$BENCHMARK_SAVE_NAME" ] && BENCHMARK_SAVE_NAME="Commit_${shortened_commit}"
+          elif [ -n "$PR_NO" ]; then
+            echo "Using PR no. $PR_NO for build..."
+            BUILD_REF="refs/pull/$PR_NO/head"
+            [ -z "$BENCHMARK_SAVE_NAME" ] && BENCHMARK_SAVE_NAME="PR_${PR_NO}"
+          fi
+          [ -z "$BENCHMARK_SAVE_NAME" ] && BENCHMARK_SAVE_NAME="Baseline"
+
+          echo "benchmark_save_name=$BENCHMARK_SAVE_NAME" >> $GITHUB_OUTPUT
+          echo "build_ref=$BUILD_REF" >> $GITHUB_OUTPUT
+
+          echo "Final sanitized values:"
+          cat $GITHUB_OUTPUT
+
   build_sycl:
     name: Build SYCL
+    needs: [ sanitize_inputs ]
     uses: ./.github/workflows/sycl-linux-build.yml
     with:
-      build_ref: |
-        ${{
-          inputs.commit_hash != '' && inputs.commit_hash ||
-          inputs.pr_no != '' && format('refs/pull/{0}/head', inputs.pr_no) ||
-          github.ref
-        }}
+      build_ref: ${{ needs.sanitize_inputs.outputs.build_ref }}
       build_cache_root: "/__w/"
       build_cache_suffix: "prod_noassert"
       build_configure_extra_args: "--no-assertions"
@@ -112,14 +173,12 @@ jobs:
 
   run_benchmarks_build:
     name: Run Benchmarks on Build
-    needs: [ build_sycl ]
+    needs: [ build_sycl, sanitize_inputs ]
     strategy:
       matrix:
         include:
-          - ref: ${{ inputs.commit_hash != '' && inputs.commit_hash || inputs.pr_no != '' && format('refs/pull/{0}/head', inputs.pr_no) || github.ref }}
-            save_name: ${{ inputs.commit_hash != '' && format('Commit{0}', inputs.commit_hash) || inputs.pr_no != '' && format('PR{0}', inputs.pr_no) || 'Baseline' }}
-            # Set default values if not specified:
-            runner: ${{ inputs.runner || '["PVC_PERF"]' }}
+          # Set default values if not specified:
+          - runner: ${{ inputs.runner || '["PVC_PERF"]' }}
             backend: ${{ inputs.backend || 'level_zero:gpu' }}
     uses: ./.github/workflows/sycl-linux-run-tests.yml
     secrets: inherit
@@ -131,9 +190,9 @@ jobs:
       target_devices: ${{ matrix.backend }}
       tests_selector: benchmarks
       benchmark_upload_results: ${{ inputs.upload_results }}
-      benchmark_save_name: ${{ matrix.save_name }}
+      benchmark_save_name: ${{ needs.sanitize_inputs.outputs.benchmark_save_name }}
       benchmark_preset: ${{ inputs.preset }}
-      repo_ref: ${{ matrix.ref }}
+      repo_ref: ${{ needs.sanitize_inputs.outputs.build_ref }}
       toolchain_artifact: ${{ needs.build_sycl.outputs.toolchain_artifact }}
       toolchain_artifact_filename: ${{ needs.build_sycl.outputs.toolchain_artifact_filename }}
       toolchain_decompress_command: ${{ needs.build_sycl.outputs.toolchain_decompress_command }}

buildbot/configure.py

@@ -66,7 +66,6 @@ def do_configure(args, passthrough_args):
     xpti_enable_werror = "OFF"
     llvm_enable_zstd = "OFF"
     spirv_enable_dis = "OFF"
-    sycl_install_device_config_file = "OFF"
 
     if sys.platform != "darwin":
         # For more info on the enablement of level_zero_v2 refer to this document:
@@ -162,7 +161,6 @@ def do_configure(args, passthrough_args):
                 libclc_targets_to_build += libclc_nvidia_target_names
             libclc_gen_remangled_variants = "ON"
             spirv_enable_dis = "ON"
-            sycl_install_device_config_file = "ON"
 
     if args.enable_backends:
         sycl_enabled_backends += args.enable_backends
@@ -211,7 +209,6 @@ def do_configure(args, passthrough_args):
         "-DSYCL_ENABLE_EXTENSION_JIT={}".format(sycl_enable_jit),
         "-DSYCL_ENABLE_MAJOR_RELEASE_PREVIEW_LIB={}".format(sycl_preview_lib),
         "-DBUG_REPORT_URL=https://github.com/intel/llvm/issues",
-        "-DSYCL_INSTALL_DEVICE_CONFIG_FILE={}".format(sycl_install_device_config_file),
     ]
 
     if libclc_enabled:

clang/include/clang/Basic/DiagnosticSemaKinds.td

@@ -13038,6 +13038,8 @@ def err_free_function_first_occurrence_missing_attr: Error<
   "the first occurrence of SYCL kernel free function should be declared with 'sycl-nd-range-kernel' or 'sycl-single-task-kernel' compile time properties">;
 def err_free_function_class_method : Error<
   "%select{static |}0class method cannot be used to define a SYCL kernel free function kernel">;
+def err_sycl_kernel_virtual_arg : Error<
+  "argument type '%0' virtually inherited from base class `%1` is not supported as a SYCL kernel argument">;
 
 
 // SYCL kernel entry point diagnostics

clang/include/clang/Driver/CudaInstallationDetector.h

@@ -0,0 +1,76 @@
+//===-- CudaInstallationDetector.h - Cuda Instalation Detector --*- 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
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_CLANG_DRIVER_CUDAINSTALLATIONDETECTOR_H
+#define LLVM_CLANG_DRIVER_CUDAINSTALLATIONDETECTOR_H
+
+#include "clang/Basic/Cuda.h"
+#include "clang/Driver/Driver.h"
+#include <bitset>
+
+namespace clang {
+namespace driver {
+
+/// A class to find a viable CUDA installation
+class CudaInstallationDetector {
+private:
+  const Driver &D;
+  bool IsValid = false;
+  CudaVersion Version = CudaVersion::UNKNOWN;
+  std::string InstallPath;
+  std::string BinPath;
+  std::string LibDevicePath;
+  std::string IncludePath;
+  llvm::StringMap<std::string> LibDeviceMap;
+
+  // CUDA architectures for which we have raised an error in
+  // CheckCudaVersionSupportsArch.
+  mutable std::bitset<(int)OffloadArch::LAST> ArchsWithBadVersion;
+
+public:
+  CudaInstallationDetector(const Driver &D, const llvm::Triple &HostTriple,
+                           const llvm::opt::ArgList &Args);
+
+  void AddCudaIncludeArgs(const llvm::opt::ArgList &DriverArgs,
+                          llvm::opt::ArgStringList &CC1Args) const;
+
+  /// Emit an error if Version does not support the given Arch.
+  ///
+  /// If either Version or Arch is unknown, does not emit an error.  Emits at
+  /// most one error per Arch.
+  void CheckCudaVersionSupportsArch(OffloadArch Arch) const;
+
+  /// Check whether we detected a valid Cuda install.
+  bool isValid() const { return IsValid; }
+  /// Print information about the detected CUDA installation.
+  void print(raw_ostream &OS) const;
+
+  /// Get the detected Cuda install's version.
+  CudaVersion version() const {
+    return Version == CudaVersion::NEW ? CudaVersion::PARTIALLY_SUPPORTED
+                                       : Version;
+  }
+  /// Get the detected Cuda installation path.
+  StringRef getInstallPath() const { return InstallPath; }
+  /// Get the detected path to Cuda's bin directory.
+  StringRef getBinPath() const { return BinPath; }
+  /// Get the detected Cuda Include path.
+  StringRef getIncludePath() const { return IncludePath; }
+  /// Get the detected Cuda device library path.
+  StringRef getLibDevicePath() const { return LibDevicePath; }
+  /// Get libdevice file for given architecture
+  std::string getLibDeviceFile(StringRef Gpu) const {
+    return LibDeviceMap.lookup(Gpu);
+  }
+  void WarnIfUnsupportedVersion() const;
+};
+
+} // namespace driver
+} // namespace clang
+
+#endif // LLVM_CLANG_DRIVER_CUDAINSTALLATIONDETECTOR_H

clang/lib/Driver/ToolChains/LazyDetector.h renamed to clang/include/clang/Driver/LazyDetector.h

@@ -6,8 +6,8 @@
 //
 //===----------------------------------------------------------------------===//
 
-#ifndef LLVM_CLANG_LIB_DRIVER_TOOLCHAINS_LAZYDETECTOR_H
-#define LLVM_CLANG_LIB_DRIVER_TOOLCHAINS_LAZYDETECTOR_H
+#ifndef LLVM_CLANG_DRIVER_LAZYDETECTOR_H
+#define LLVM_CLANG_DRIVER_LAZYDETECTOR_H
 
 #include "clang/Driver/Tool.h"
 #include "clang/Driver/ToolChain.h"
@@ -42,4 +42,4 @@ template <class T> class LazyDetector {
 
 } // end namespace clang
 
-#endif // LLVM_CLANG_LIB_DRIVER_TOOLCHAINS_LAZYDETECTOR_H
+#endif // LLVM_CLANG_DRIVER_LAZYDETECTOR_H

clang/include/clang/Driver/Options.td

@@ -9265,14 +9265,11 @@ class CLRemainingArgsJoined<string name,
 // (We don't put any of these in cl_compile_Group as the options they alias are
 // already in the right group.)
 
-// INTEL_CUSTOMIZATION
 def _SLASH_Qfp_accuracy_EQ : CLJoined<"Qfp-accuracy=">,
   Alias<ffp_accuracy_EQ>;
 def _SLASH_Qfp_accuracy_COL : CLJoined<"Qfp-accuracy:">,
   Alias<ffp_accuracy_EQ>,HelpText<"Specifies the required accuracy for "
   "floating-point operations and library calls.">;
-// END INTEL_CUSTOMIZATION
-
 def _SLASH_Brepro : CLFlag<"Brepro">,
   HelpText<"Do not write current time into COFF output (breaks link.exe /incremental)">,
   Alias<mno_incremental_linker_compatible>;

clang/lib/Driver/ToolChains/ROCm.h renamed to clang/include/clang/Driver/RocmInstallationDetector.h

@@ -1,24 +1,15 @@
-//===--- ROCm.h - ROCm installation detector --------------------*- C++ -*-===//
+//===-- RocmInstallationDetector.h - ROCm Instalation Detector --*- 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
 //
 //===----------------------------------------------------------------------===//
 
-#ifndef LLVM_CLANG_LIB_DRIVER_TOOLCHAINS_ROCM_H
-#define LLVM_CLANG_LIB_DRIVER_TOOLCHAINS_ROCM_H
+#ifndef LLVM_CLANG_DRIVER_ROCMINSTALLATIONDETECTOR_H
+#define LLVM_CLANG_DRIVER_ROCMINSTALLATIONDETECTOR_H
 
-#include "clang/Basic/Cuda.h"
-#include "clang/Basic/LLVM.h"
 #include "clang/Driver/Driver.h"
-#include "clang/Driver/Options.h"
-#include "clang/Driver/SanitizerArgs.h"
-#include "llvm/ADT/SmallString.h"
-#include "llvm/ADT/StringMap.h"
-#include "llvm/Option/ArgList.h"
-#include "llvm/Support/VersionTuple.h"
-#include "llvm/TargetParser/Triple.h"
 
 namespace clang {
 namespace driver {
@@ -77,6 +68,24 @@ class RocmInstallationDetector {
           SPACKReleaseStr(SPACKReleaseStr.str()) {}
   };
 
+  struct CommonBitcodeLibsPreferences {
+    CommonBitcodeLibsPreferences(const Driver &D,
+                                 const llvm::opt::ArgList &DriverArgs,
+                                 StringRef GPUArch,
+                                 const Action::OffloadKind DeviceOffloadingKind,
+                                 const bool NeedsASanRT);
+
+    DeviceLibABIVersion ABIVer;
+    bool IsOpenMP;
+    bool Wave64;
+    bool DAZ;
+    bool FiniteOnly;
+    bool UnsafeMathOpt;
+    bool FastRelaxedMath;
+    bool CorrectSqrt;
+    bool GPUSan;
+  };
+
   const Driver &D;
   bool HasHIPRuntime = false;
   bool HasDeviceLibrary = false;
@@ -175,11 +184,11 @@ class RocmInstallationDetector {
 
   /// Get file paths of default bitcode libraries common to AMDGPU based
   /// toolchains.
-  llvm::SmallVector<ToolChain::BitCodeLibraryInfo, 12> getCommonBitcodeLibs(
-      const llvm::opt::ArgList &DriverArgs, StringRef LibDeviceFile,
-      bool Wave64, bool DAZ, bool FiniteOnly, bool UnsafeMathOpt,
-      bool FastRelaxedMath, bool CorrectSqrt, DeviceLibABIVersion ABIVer,
-      bool GPUSan, bool isOpenMP) const;
+  llvm::SmallVector<ToolChain::BitCodeLibraryInfo, 12>
+  getCommonBitcodeLibs(const llvm::opt::ArgList &DriverArgs,
+                       StringRef LibDeviceFile, StringRef GPUArch,
+                       const Action::OffloadKind DeviceOffloadingKind,
+                       const bool NeedsASanRT) const;
   /// Check file paths of default bitcode libraries common to AMDGPU based
   /// toolchains. \returns false if there are invalid or missing files.
   bool checkCommonBitcodeLibs(StringRef GPUArch, StringRef LibDeviceFile,
@@ -288,7 +297,7 @@ class RocmInstallationDetector {
   StringRef getHIPVersion() const { return DetectedVersion; }
 };
 
-} // end namespace driver
-} // end namespace clang
+} // namespace driver
+} // namespace clang
 
-#endif // LLVM_CLANG_LIB_DRIVER_TOOLCHAINS_ROCM_H
+#endif // LLVM_CLANG_DRIVER_ROCMINSTALLATIONDETECTOR_H