Merge branch 'users/meinersbur/flang_runtime_FLANG_INCLUDE_RUNTIME' into users/meinersbur/flang_runtime_flang_rt

Author: Meinersbur

clang/lib/StaticAnalyzer/Core/ExprEngine.cpp

@@ -74,6 +74,7 @@
 #include "llvm/Support/ErrorHandling.h"
 #include "llvm/Support/GraphWriter.h"
 #include "llvm/Support/SaveAndRestore.h"
+#include "llvm/Support/TimeProfiler.h"
 #include "llvm/Support/raw_ostream.h"
 #include <cassert>
 #include <cstdint>
@@ -1031,6 +1032,7 @@ void ExprEngine::removeDead(ExplodedNode *Pred, ExplodedNodeSet &Out,
                             const LocationContext *LC,
                             const Stmt *DiagnosticStmt,
                             ProgramPoint::Kind K) {
+  llvm::TimeTraceScope TimeScope("ExprEngine::removeDead");
   assert((K == ProgramPoint::PreStmtPurgeDeadSymbolsKind ||
           ReferenceStmt == nullptr || isa<ReturnStmt>(ReferenceStmt))
           && "PostStmt is not generally supported by the SymbolReaper yet");

clang/lib/StaticAnalyzer/Core/RegionStore.cpp

@@ -29,6 +29,7 @@
 #include "clang/StaticAnalyzer/Core/PathSensitive/ProgramStateTrait.h"
 #include "llvm/ADT/ImmutableMap.h"
 #include "llvm/ADT/STLExtras.h"
+#include "llvm/Support/TimeProfiler.h"
 #include "llvm/Support/raw_ostream.h"
 #include <optional>
 #include <utility>
@@ -112,6 +113,13 @@ class BindingKey {
 
   LLVM_DUMP_METHOD void dump() const;
 };
+
+std::string locDescr(Loc L) {
+  std::string S;
+  llvm::raw_string_ostream OS(S);
+  L.dumpToStream(OS);
+  return OS.str();
+}
 } // end anonymous namespace
 
 BindingKey BindingKey::Make(const MemRegion *R, Kind k) {
@@ -2408,6 +2416,8 @@ StoreRef RegionStoreManager::killBinding(Store ST, Loc L) {
 
 RegionBindingsRef
 RegionStoreManager::bind(RegionBindingsConstRef B, Loc L, SVal V) {
+  llvm::TimeTraceScope TimeScope("RegionStoreManager::bind",
+                                 [&L]() { return locDescr(L); });
   // We only care about region locations.
   auto MemRegVal = L.getAs<loc::MemRegionVal>();
   if (!MemRegVal)
@@ -2514,6 +2524,8 @@ RegionBindingsRef
 RegionStoreManager::bindArray(RegionBindingsConstRef B,
                               const TypedValueRegion* R,
                               SVal Init) {
+  llvm::TimeTraceScope TimeScope("RegionStoreManager::bindArray",
+                                 [R]() { return R->getDescriptiveName(); });
 
   const ArrayType *AT =cast<ArrayType>(Ctx.getCanonicalType(R->getValueType()));
   QualType ElementTy = AT->getElementType();
@@ -2578,6 +2590,8 @@ RegionStoreManager::bindArray(RegionBindingsConstRef B,
 RegionBindingsRef RegionStoreManager::bindVector(RegionBindingsConstRef B,
                                                  const TypedValueRegion* R,
                                                  SVal V) {
+  llvm::TimeTraceScope TimeScope("RegionStoreManager::bindVector",
+                                 [R]() { return R->getDescriptiveName(); });
   QualType T = R->getValueType();
   const VectorType *VT = T->castAs<VectorType>(); // Use castAs for typedefs.
 
@@ -2700,6 +2714,8 @@ std::optional<RegionBindingsRef> RegionStoreManager::tryBindSmallStruct(
 RegionBindingsRef RegionStoreManager::bindStruct(RegionBindingsConstRef B,
                                                  const TypedValueRegion *R,
                                                  SVal V) {
+  llvm::TimeTraceScope TimeScope("RegionStoreManager::bindStruct",
+                                 [R]() { return R->getDescriptiveName(); });
   QualType T = R->getValueType();
   assert(T->isStructureOrClassType());
 
@@ -2818,6 +2834,8 @@ RegionBindingsRef
 RegionStoreManager::bindAggregate(RegionBindingsConstRef B,
                                   const TypedRegion *R,
                                   SVal Val) {
+  llvm::TimeTraceScope TimeScope("RegionStoreManager::bindAggregate",
+                                 [R]() { return R->getDescriptiveName(); });
   // Remove the old bindings, using 'R' as the root of all regions
   // we will invalidate. Then add the new binding.
   return removeSubRegionBindings(B, R).addBinding(R, BindingKey::Default, Val);

clang/test/Analysis/ftime-trace-bind.cpp

@@ -0,0 +1,24 @@
+// RUN: %clang_analyze_cc1 -analyzer-checker=core %s -ftime-trace=%t.raw.json -ftime-trace-granularity=0 -verify
+// RUN: %python -c 'import json, sys; print(json.dumps(json.load(sys.stdin), indent=4))' < %t.raw.json > %t.formatted.json
+// RUN: FileCheck --input-file=%t.formatted.json --check-prefix=CHECK %s
+
+// CHECK:          "name": "RegionStoreManager::bindArray",
+// CHECK-NEXT:     "args": {
+//
+// The below does not necessarily follow immediately,
+// depending on what parts of the array are initialized first.
+//
+// CHECK:              "detail": "'arr[0][1]'"
+// CHECK-NEXT:     }
+//
+// CHECK:              "detail": "'arr[0]'"
+// CHECK-NEXT:     }
+//
+// CHECK:              "detail": "'arr'"
+// CHECK-NEXT:     }
+
+int f() {
+    int arr[2][2][2] = {{{1, 2}, {3, 4}}, {{5, 6}, {7, 8}}};
+    return arr[1][0][1];
+}
+// expected-no-diagnostics

clang/test/Analysis/ftime-trace-removeDead.cpp

@@ -0,0 +1,17 @@
+// RUN: %clang_analyze_cc1 -analyzer-checker=core %s -ftime-trace=%t.raw.json -ftime-trace-granularity=0 -verify
+// RUN: %python -c 'import json, sys; print(json.dumps(json.load(sys.stdin), indent=4))' < %t.raw.json > %t.formatted.json
+// RUN: FileCheck --input-file=%t.formatted.json --check-prefix=CHECK %s
+
+// The trace file is rather large, but it should contain at least one scope for removeDead:
+//
+// CHECK:          "name": "ExprEngine::removeDead"
+
+bool coin();
+int f() {
+    int x = 0;
+    int y = 0;
+    while (coin()) {
+        x = 1;
+    }
+    return x / y; // expected-warning{{Division by zero}}
+}

clang/test/Tooling/clang-linker-wrapper-spirv-elf.cpp

@@ -0,0 +1,18 @@
+// Verify the ELF packaging of OpenMP SPIR-V device images.
+// REQUIRES: system-linux
+// REQUIRES: spirv-tools
+// RUN: mkdir -p %t_tmp
+// RUN: cd %t_tmp
+// RUN: %clangxx -fopenmp -fopenmp-targets=spirv64-intel -nogpulib -c -o %t_clang-linker-wrapper-spirv-elf.o %s
+// RUN: not clang-linker-wrapper -o a.out %t_clang-linker-wrapper-spirv-elf.o --save-temps --linker-path=ld
+// RUN: clang-offload-packager --image=triple=spirv64-intel,kind=openmp,file=%t.elf  %t_tmp/a.out.openmp.image.wrapper.o
+// RUN: llvm-readelf -t %t.elf | FileCheck -check-prefix=CHECK-SECTION %s
+// RUN: llvm-readelf -n %t.elf | FileCheck -check-prefix=CHECK-NOTES %s
+
+// CHECK-SECTION: .note.inteloneompoffload
+// CHECK-SECTION: __openmp_offload_spirv_0
+
+// CHECK-NOTES-COUNT-3: INTELONEOMPOFFLOAD
+int main(int argc, char** argv) {
+  return 0;
+}

clang/test/Tooling/lit.local.cfg

@@ -1,2 +1,8 @@
 if not config.root.clang_staticanalyzer:
     config.unsupported = True
+
+if config.spirv_tools_tests:
+    config.available_features.add("spirv-tools")
+    config.substitutions.append(("spirv-dis", os.path.join(config.llvm_tools_dir, "spirv-dis")))
+    config.substitutions.append(("spirv-val", os.path.join(config.llvm_tools_dir, "spirv-val")))
+    config.substitutions.append(("spirv-as", os.path.join(config.llvm_tools_dir, "spirv-as")))

clang/test/lit.site.cfg.py.in

@@ -43,6 +43,7 @@ config.llvm_external_lit = path(r"@LLVM_EXTERNAL_LIT@")
 config.standalone_build = @CLANG_BUILT_STANDALONE@
 config.ppc_linux_default_ieeelongdouble = @PPC_LINUX_DEFAULT_IEEELONGDOUBLE@
 config.have_llvm_driver = @LLVM_TOOL_LLVM_DRIVER_BUILD@
+config.spirv_tools_tests = "@LLVM_INCLUDE_SPIRV_TOOLS_TESTS@"
 config.substitutions.append(("%llvm-version-major", "@LLVM_VERSION_MAJOR@"))
 
 import lit.llvm

clang/tools/clang-linker-wrapper/ClangLinkerWrapper.cpp

@@ -605,6 +605,15 @@ Expected<StringRef> linkDevice(ArrayRef<StringRef> InputFiles,
   }
 }
 
+Error containerizeRawImage(std::unique_ptr<MemoryBuffer> &Img, OffloadKind Kind,
+                           const ArgList &Args) {
+  llvm::Triple Triple(Args.getLastArgValue(OPT_triple_EQ));
+  if (Kind != OFK_OpenMP || !Triple.isSPIRV() ||
+      Triple.getVendor() != llvm::Triple::Intel)
+    return Error::success();
+  return offloading::intel::containerizeOpenMPSPIRVImage(Img);
+}
+
 Expected<StringRef> writeOffloadFile(const OffloadFile &File) {
   const OffloadBinary &Binary = *File.getBinary();
 
@@ -957,6 +966,10 @@ Expected<SmallVector<StringRef>> linkAndWrapDeviceFiles(
           return createFileError(*OutputOrErr, EC);
       }
 
+      // Manually containerize offloading images not in ELF format.
+      if (Error E = containerizeRawImage(*FileOrErr, Kind, LinkerArgs))
+        return E;
+
       std::scoped_lock<decltype(ImageMtx)> Guard(ImageMtx);
       OffloadingImage TheImage{};
       TheImage.TheImageKind =

flang/test/Driver/linker-flags.f90

@@ -41,7 +41,6 @@
 ! BSD-F128NONE-NOT: FortranFloat128Math
 ! BSD-F128LIBQUADMATH-SAME: "-lFortranFloat128Math" "--as-needed" "-lquadmath" "--no-as-needed"
 ! BSD-SAME: -lFortranRuntime
-! BSD-SAME: -lFortranDecimal
 ! BSD-SAME: -lexecinfo
 
 ! DARWIN-LABEL:  "{{.*}}ld{{(\.exe)?}}"

llvm/include/llvm/Frontend/Offloading/Utility.h

@@ -10,6 +10,7 @@
 #define LLVM_FRONTEND_OFFLOADING_UTILITY_H
 
 #include <cstdint>
+#include <memory>
 
 #include "llvm/ADT/StringMap.h"
 #include "llvm/ADT/StringRef.h"
@@ -152,6 +153,12 @@ Error getAMDGPUMetaDataFromImage(MemoryBufferRef MemBuffer,
                                  StringMap<AMDGPUKernelMetaData> &KernelInfoMap,
                                  uint16_t &ELFABIVersion);
 } // namespace amdgpu
+
+namespace intel {
+/// Containerizes an offloading binary into the ELF binary format expected by
+/// the Intel runtime offload plugin.
+Error containerizeOpenMPSPIRVImage(std::unique_ptr<MemoryBuffer> &Binary);
+} // namespace intel
 } // namespace offloading
 } // namespace llvm
 

llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp

@@ -23188,7 +23188,7 @@ SDValue DAGCombiner::visitEXTRACT_VECTOR_ELT(SDNode *N) {
     auto *VecLoad = dyn_cast<LoadSDNode>(VecOp);
     if (VecLoad && VecLoad->isSimple()) {
       if (SDValue Scalarized = TLI.scalarizeExtractedVectorLoad(
-              ExtVT, SDLoc(N), VecVT, Index, VecLoad, DAG)) {
+              ScalarVT, SDLoc(N), VecVT, Index, VecLoad, DAG)) {
         ++OpsNarrowed;
         return Scalarized;
       }

llvm/lib/Frontend/Offloading/CMakeLists.txt

@@ -12,6 +12,7 @@ add_llvm_component_library(LLVMFrontendOffloading
   Core
   BinaryFormat
   Object
+  ObjectYAML
   Support
   TransformUtils
   TargetParser

llvm/lib/Frontend/Offloading/Utility.cpp

@@ -15,6 +15,8 @@
 #include "llvm/IR/GlobalVariable.h"
 #include "llvm/IR/Value.h"
 #include "llvm/Object/ELFObjectFile.h"
+#include "llvm/ObjectYAML/ELFYAML.h"
+#include "llvm/ObjectYAML/yaml2obj.h"
 #include "llvm/Support/MemoryBufferRef.h"
 #include "llvm/Transforms/Utils/ModuleUtils.h"
 
@@ -374,3 +376,86 @@ Error llvm::offloading::amdgpu::getAMDGPUMetaDataFromImage(
   }
   return Error::success();
 }
+Error offloading::intel::containerizeOpenMPSPIRVImage(
+    std::unique_ptr<MemoryBuffer> &Img) {
+  constexpr char INTEL_ONEOMP_OFFLOAD_VERSION[] = "1.0";
+  constexpr int NT_INTEL_ONEOMP_OFFLOAD_VERSION = 1;
+  constexpr int NT_INTEL_ONEOMP_OFFLOAD_IMAGE_COUNT = 2;
+  constexpr int NT_INTEL_ONEOMP_OFFLOAD_IMAGE_AUX = 3;
+
+  // Start creating notes for the ELF container.
+  std::vector<ELFYAML::NoteEntry> Notes;
+  std::string Version = toHex(INTEL_ONEOMP_OFFLOAD_VERSION);
+  Notes.emplace_back(ELFYAML::NoteEntry{"INTELONEOMPOFFLOAD",
+                                        yaml::BinaryRef(Version),
+                                        NT_INTEL_ONEOMP_OFFLOAD_VERSION});
+
+  // The AuxInfo string will hold auxiliary information for the image.
+  // ELFYAML::NoteEntry structures will hold references to the
+  // string, so we have to make sure the string is valid.
+  std::string AuxInfo;
+
+  // TODO: Pass compile/link opts
+  StringRef CompileOpts = "";
+  StringRef LinkOpts = "";
+
+  unsigned ImageFmt = 1; // SPIR-V format
+
+  AuxInfo = toHex((Twine(0) + Twine('\0') + Twine(ImageFmt) + Twine('\0') +
+                   CompileOpts + Twine('\0') + LinkOpts)
+                      .str());
+  Notes.emplace_back(ELFYAML::NoteEntry{"INTELONEOMPOFFLOAD",
+                                        yaml::BinaryRef(AuxInfo),
+                                        NT_INTEL_ONEOMP_OFFLOAD_IMAGE_AUX});
+
+  std::string ImgCount = toHex(Twine(1).str()); // always one image per ELF
+  Notes.emplace_back(ELFYAML::NoteEntry{"INTELONEOMPOFFLOAD",
+                                        yaml::BinaryRef(ImgCount),
+                                        NT_INTEL_ONEOMP_OFFLOAD_IMAGE_COUNT});
+
+  std::string YamlFile;
+  llvm::raw_string_ostream YamlFileStream(YamlFile);
+
+  // Write the YAML template file.
+
+  // We use 64-bit little-endian ELF currently.
+  ELFYAML::FileHeader Header{};
+  Header.Class = ELF::ELFCLASS64;
+  Header.Data = ELF::ELFDATA2LSB;
+  Header.Type = ELF::ET_DYN;
+  // Use an existing Intel machine type as there is not one specifically for
+  // Intel GPUs.
+  Header.Machine = ELF::EM_IA_64;
+
+  // Create a section with notes.
+  ELFYAML::NoteSection Section{};
+  Section.Type = ELF::SHT_NOTE;
+  Section.AddressAlign = 0;
+  Section.Name = ".note.inteloneompoffload";
+  Section.Notes.emplace(std::move(Notes));
+
+  ELFYAML::Object Object{};
+  Object.Header = Header;
+  Object.Chunks.push_back(
+      std::make_unique<ELFYAML::NoteSection>(std::move(Section)));
+
+  // Create the section that will hold the image
+  ELFYAML::RawContentSection ImageSection{};
+  ImageSection.Type = ELF::SHT_PROGBITS;
+  ImageSection.AddressAlign = 0;
+  std::string Name = "__openmp_offload_spirv_0";
+  ImageSection.Name = Name;
+  ImageSection.Content =
+      llvm::yaml::BinaryRef(arrayRefFromStringRef(Img->getBuffer()));
+  Object.Chunks.push_back(
+      std::make_unique<ELFYAML::RawContentSection>(std::move(ImageSection)));
+  Error Err = Error::success();
+  llvm::yaml::yaml2elf(
+      Object, YamlFileStream,
+      [&Err](const Twine &Msg) { Err = createStringError(Msg); }, UINT64_MAX);
+  if (Err)
+    return Err;
+
+  Img = MemoryBuffer::getMemBufferCopy(YamlFile);
+  return Error::success();
+}