[Offload] Unify offloading entries into a single section (#125731)

Summary:
This patch unifies the existing offloading entires into a single section
called `llvm_offload_entires`. This lets us use a more unified
offloading infrastructure so that all targets share the same handling.
The effect is that people in the runtimes now need to check if the kind
is what they expect, but the expectation is that you can combine
multiple potential providers into a compile job. Doesn't fully work
yet because of other runtime issues, but some day. Mostly this helps the
future of liboffload where we want to handle different languages than
OpenMP.

Author: jhuber6

clang/lib/CodeGen/CGCUDANV.cpp

@@ -40,8 +40,6 @@ class CGNVCUDARuntime : public CGCUDARuntime {
 
   /// The prefix used for function calls and section names (CUDA, HIP, LLVM)
   StringRef Prefix;
-  /// TODO: We should transition the OpenMP section to LLVM/Offload
-  StringRef SectionPrefix;
 
 private:
   llvm::IntegerType *IntTy, *SizeTy;
@@ -234,13 +232,12 @@ CGNVCUDARuntime::CGNVCUDARuntime(CodeGenModule &CGM)
   VoidTy = CGM.VoidTy;
   PtrTy = CGM.UnqualPtrTy;
 
-  if (CGM.getLangOpts().OffloadViaLLVM) {
+  if (CGM.getLangOpts().OffloadViaLLVM)
     Prefix = "llvm";
-    SectionPrefix = "omp";
-  } else if (CGM.getLangOpts().HIP)
-    SectionPrefix = Prefix = "hip";
+  else if (CGM.getLangOpts().HIP)
+    Prefix = "hip";
   else
-    SectionPrefix = Prefix = "cuda";
+    Prefix = "cuda";
 }
 
 llvm::FunctionCallee CGNVCUDARuntime::getSetupArgumentFn() const {
@@ -1198,17 +1195,19 @@ void CGNVCUDARuntime::transformManagedVars() {
 // register the symbols with the linked device image.
 void CGNVCUDARuntime::createOffloadingEntries() {
   SmallVector<char, 32> Out;
-  StringRef Section = (SectionPrefix + "_offloading_entries").toStringRef(Out);
   llvm::object::OffloadKind Kind = CGM.getLangOpts().HIP
                                        ? llvm::object::OffloadKind::OFK_HIP
                                        : llvm::object::OffloadKind::OFK_Cuda;
+  // For now, just spoof this as OpenMP because that's the runtime it uses.
+  if (CGM.getLangOpts().OffloadViaLLVM)
+    Kind = llvm::object::OffloadKind::OFK_OpenMP;
 
   llvm::Module &M = CGM.getModule();
   for (KernelInfo &I : EmittedKernels)
     llvm::offloading::emitOffloadingEntry(
         M, Kind, KernelHandles[I.Kernel->getName()],
         getDeviceSideName(cast<NamedDecl>(I.D)), /*Flags=*/0, /*Data=*/0,
-        llvm::offloading::OffloadGlobalEntry, Section);
+        llvm::offloading::OffloadGlobalEntry);
 
   for (VarInfo &I : DeviceVars) {
     uint64_t VarSize =
@@ -1233,23 +1232,23 @@ void CGNVCUDARuntime::createOffloadingEntries() {
         llvm::offloading::emitOffloadingEntry(
             M, Kind, I.Var, getDeviceSideName(I.D), VarSize,
             llvm::offloading::OffloadGlobalManagedEntry | Flags,
-            /*Data=*/I.Var->getAlignment(), Section, ManagedVar);
+            /*Data=*/I.Var->getAlignment(), ManagedVar);
       } else {
         llvm::offloading::emitOffloadingEntry(
             M, Kind, I.Var, getDeviceSideName(I.D), VarSize,
             llvm::offloading::OffloadGlobalEntry | Flags,
-            /*Data=*/0, Section);
+            /*Data=*/0);
       }
     } else if (I.Flags.getKind() == DeviceVarFlags::Surface) {
       llvm::offloading::emitOffloadingEntry(
           M, Kind, I.Var, getDeviceSideName(I.D), VarSize,
           llvm::offloading::OffloadGlobalSurfaceEntry | Flags,
-          I.Flags.getSurfTexType(), Section);
+          I.Flags.getSurfTexType());
     } else if (I.Flags.getKind() == DeviceVarFlags::Texture) {
       llvm::offloading::emitOffloadingEntry(
           M, Kind, I.Var, getDeviceSideName(I.D), VarSize,
           llvm::offloading::OffloadGlobalTextureEntry | Flags,
-          I.Flags.getSurfTexType(), Section);
+          I.Flags.getSurfTexType());
     }
   }
 }

clang/test/CodeGenCUDA/offloading-entries.cu

@@ -14,18 +14,18 @@
 // RUN: clang-linker-wrapper --print-wrapped-module --dry-run --host-triple=x86_64-unknown-windows-gnu \
 // RUN:   --linker-path=/usr/bin/ld %t.o -o a.out 2>&1 | FileCheck %s --check-prefixes=OPENMP,OPENMP-COFF
 
-//      OPENMP-ELF: @__start_omp_offloading_entries = external hidden constant [0 x %struct.__tgt_offload_entry]
-// OPENMP-ELF-NEXT: @__stop_omp_offloading_entries = external hidden constant [0 x %struct.__tgt_offload_entry]
-// OPENMP-ELF-NEXT: @__dummy.omp_offloading_entries = internal constant [0 x %struct.__tgt_offload_entry] zeroinitializer, section "omp_offloading_entries"
+//      OPENMP-ELF: @__start_llvm_offload_entries = external hidden constant [0 x %struct.__tgt_offload_entry]
+// OPENMP-ELF-NEXT: @__stop_llvm_offload_entries = external hidden constant [0 x %struct.__tgt_offload_entry]
+// OPENMP-ELF-NEXT: @__dummy.llvm_offload_entries = internal constant [0 x %struct.__tgt_offload_entry] zeroinitializer, section "llvm_offload_entries"
 
-//      OPENMP-COFF: @__start_omp_offloading_entries = weak_odr hidden constant [0 x %struct.__tgt_offload_entry] zeroinitializer, section "omp_offloading_entries$OA"
-// OPENMP-COFF-NEXT: @__stop_omp_offloading_entries = weak_odr hidden constant [0 x %struct.__tgt_offload_entry] zeroinitializer, section "omp_offloading_entries$OZ"
+//      OPENMP-COFF: @__start_llvm_offload_entries = weak_odr hidden constant [0 x %struct.__tgt_offload_entry] zeroinitializer, section "llvm_offload_entries$OA"
+// OPENMP-COFF-NEXT: @__stop_llvm_offload_entries = weak_odr hidden constant [0 x %struct.__tgt_offload_entry] zeroinitializer, section "llvm_offload_entries$OZ"
 
 // OPENMP-REL: @.omp_offloading.device_image = internal unnamed_addr constant [[[SIZE:[0-9]+]] x i8] c"\10\FF\10\AD{{.*}}", section ".llvm.offloading.relocatable", align 8
 
 //      OPENMP: @.omp_offloading.device_image = internal unnamed_addr constant [[[SIZE:[0-9]+]] x i8] c"\10\FF\10\AD{{.*}}", section ".llvm.offloading", align 8
-// OPENMP-NEXT: @.omp_offloading.device_images = internal unnamed_addr constant [1 x %__tgt_device_image] [%__tgt_device_image { ptr getelementptr ([[[BEGIN:[0-9]+]] x i8], ptr @.omp_offloading.device_image, i64 0, i64 144), ptr getelementptr ([[[END:[0-9]+]] x i8], ptr @.omp_offloading.device_image, i64 0, i64 144), ptr @__start_omp_offloading_entries, ptr @__stop_omp_offloading_entries }]
-// OPENMP-NEXT: @.omp_offloading.descriptor = internal constant %__tgt_bin_desc { i32 1, ptr @.omp_offloading.device_images, ptr @__start_omp_offloading_entries, ptr @__stop_omp_offloading_entries }
+// OPENMP-NEXT: @.omp_offloading.device_images = internal unnamed_addr constant [1 x %__tgt_device_image] [%__tgt_device_image { ptr getelementptr ([[[BEGIN:[0-9]+]] x i8], ptr @.omp_offloading.device_image, i64 0, i64 144), ptr getelementptr ([[[END:[0-9]+]] x i8], ptr @.omp_offloading.device_image, i64 0, i64 144), ptr @__start_llvm_offload_entries, ptr @__stop_llvm_offload_entries }]
+// OPENMP-NEXT: @.omp_offloading.descriptor = internal constant %__tgt_bin_desc { i32 1, ptr @.omp_offloading.device_images, ptr @__start_llvm_offload_entries, ptr @__stop_llvm_offload_entries }
 // OPENMP-NEXT: @llvm.global_ctors = appending global [1 x { i32, ptr, ptr }] [{ i32, ptr, ptr } { i32 101, ptr @.omp_offloading.descriptor_reg, ptr null }]
 
 //      OPENMP: define internal void @.omp_offloading.descriptor_reg() section ".text.startup" {
@@ -51,12 +51,12 @@
 // RUN: clang-linker-wrapper --print-wrapped-module --dry-run --host-triple=x86_64-unknown-windows-gnu \
 // RUN:   --linker-path=/usr/bin/ld %t.o -o a.out 2>&1 | FileCheck %s --check-prefixes=CUDA,CUDA-COFF
 
-//      CUDA-ELF: @__start_cuda_offloading_entries = external hidden constant [0 x %struct.__tgt_offload_entry]
-// CUDA-ELF-NEXT: @__stop_cuda_offloading_entries = external hidden constant [0 x %struct.__tgt_offload_entry]
-// CUDA-ELF-NEXT: @__dummy.cuda_offloading_entries = internal constant [0 x %struct.__tgt_offload_entry] zeroinitializer, section "cuda_offloading_entries"
+//      CUDA-ELF: @__start_llvm_offload_entries = external hidden constant [0 x %struct.__tgt_offload_entry]
+// CUDA-ELF-NEXT: @__stop_llvm_offload_entries = external hidden constant [0 x %struct.__tgt_offload_entry]
+// CUDA-ELF-NEXT: @__dummy.llvm_offload_entries = internal constant [0 x %struct.__tgt_offload_entry] zeroinitializer, section "llvm_offload_entries"
 
-//      CUDA-COFF: @__start_cuda_offloading_entries = weak_odr hidden constant [0 x %struct.__tgt_offload_entry] zeroinitializer, section "cuda_offloading_entries$OA"
-// CUDA-COFF-NEXT: @__stop_cuda_offloading_entries = weak_odr hidden constant [0 x %struct.__tgt_offload_entry] zeroinitializer, section "cuda_offloading_entries$OZ"
+//      CUDA-COFF: @__start_llvm_offload_entries = weak_odr hidden constant [0 x %struct.__tgt_offload_entry] zeroinitializer, section "llvm_offload_entries$OA"
+// CUDA-COFF-NEXT: @__stop_llvm_offload_entries = weak_odr hidden constant [0 x %struct.__tgt_offload_entry] zeroinitializer, section "llvm_offload_entries$OZ"
 
 //      CUDA: @.fatbin_image = internal constant [0 x i8] zeroinitializer, section ".nv_fatbin"
 // CUDA-NEXT: @.fatbin_wrapper = internal constant %fatbin_wrapper { i32 1180844977, i32 1, ptr @.fatbin_image, ptr null }, section ".nvFatBinSegment", align 8
@@ -83,11 +83,11 @@
 //
 //      CUDA: define internal void @.cuda.globals_reg(ptr %0) section ".text.startup" {
 // CUDA-NEXT: entry:
-// CUDA-NEXT:   %1 = icmp ne ptr @__start_cuda_offloading_entries, @__stop_cuda_offloading_entries
+// CUDA-NEXT:   %1 = icmp ne ptr @__start_llvm_offload_entries, @__stop_llvm_offload_entries
 // CUDA-NEXT:   br i1 %1, label %while.entry, label %while.end
 //
 //      CUDA: while.entry:
-// CUDA-NEXT:   %entry1 = phi ptr [ @__start_cuda_offloading_entries, %entry ], [ %16, %if.end ]
+// CUDA-NEXT:   %entry1 = phi ptr [ @__start_llvm_offload_entries, %entry ], [ %16, %if.end ]
 // CUDA-NEXT:   %2 = getelementptr inbounds %struct.__tgt_offload_entry, ptr %entry1, i32 0, i32 4
 // CUDA-NEXT:   %addr = load ptr, ptr %2, align 8
 // CUDA-NEXT:   %3 = getelementptr inbounds %struct.__tgt_offload_entry, ptr %entry1, i32 0, i32 8
@@ -145,7 +145,7 @@
 //
 //      CUDA: if.end:
 // CUDA-NEXT:   %16 = getelementptr inbounds %struct.__tgt_offload_entry, ptr %entry1, i64 1
-// CUDA-NEXT:   %17 = icmp eq ptr %16, @__stop_cuda_offloading_entries
+// CUDA-NEXT:   %17 = icmp eq ptr %16, @__stop_llvm_offload_entries
 // CUDA-NEXT:   br i1 %17, label %while.end, label %while.entry
 //
 //      CUDA: while.end:
@@ -162,12 +162,12 @@
 // RUN: clang-linker-wrapper --print-wrapped-module --dry-run --host-triple=x86_64-unknown-windows-gnu \
 // RUN:   --linker-path=/usr/bin/ld %t.o -o a.out 2>&1 | FileCheck %s --check-prefixes=HIP,HIP-COFF
 
-//      HIP-ELF: @__start_hip_offloading_entries = external hidden constant [0 x %struct.__tgt_offload_entry]
-// HIP-ELF-NEXT: @__stop_hip_offloading_entries = external hidden constant [0 x %struct.__tgt_offload_entry]
-// HIP-ELF-NEXT: @__dummy.hip_offloading_entries = internal constant [0 x %struct.__tgt_offload_entry] zeroinitializer, section "hip_offloading_entries"
+//      HIP-ELF: @__start_llvm_offload_entries = external hidden constant [0 x %struct.__tgt_offload_entry]
+// HIP-ELF-NEXT: @__stop_llvm_offload_entries = external hidden constant [0 x %struct.__tgt_offload_entry]
+// HIP-ELF-NEXT: @__dummy.llvm_offload_entries = internal constant [0 x %struct.__tgt_offload_entry] zeroinitializer, section "llvm_offload_entries"
 
-//      HIP-COFF: @__start_hip_offloading_entries = weak_odr hidden constant [0 x %struct.__tgt_offload_entry] zeroinitializer, section "hip_offloading_entries$OA"
-// HIP-COFF-NEXT: @__stop_hip_offloading_entries = weak_odr hidden constant [0 x %struct.__tgt_offload_entry] zeroinitializer, section "hip_offloading_entries$OZ"
+//      HIP-COFF: @__start_llvm_offload_entries = weak_odr hidden constant [0 x %struct.__tgt_offload_entry] zeroinitializer, section "llvm_offload_entries$OA"
+// HIP-COFF-NEXT: @__stop_llvm_offload_entries = weak_odr hidden constant [0 x %struct.__tgt_offload_entry] zeroinitializer, section "llvm_offload_entries$OZ"
 
 //      HIP: @.fatbin_image = internal constant [0 x i8] zeroinitializer, section ".hip_fatbin"
 // HIP-NEXT: @.fatbin_wrapper = internal constant %fatbin_wrapper { i32 1212764230, i32 1, ptr @.fatbin_image, ptr null }, section ".hipFatBinSegment", align 8
@@ -193,11 +193,11 @@
 //
 //      HIP: define internal void @.hip.globals_reg(ptr %0) section ".text.startup" {
 // HIP-NEXT: entry:
-// HIP-NEXT:   %1 = icmp ne ptr @__start_hip_offloading_entries, @__stop_hip_offloading_entries
+// HIP-NEXT:   %1 = icmp ne ptr @__start_llvm_offload_entries, @__stop_llvm_offload_entries
 // HIP-NEXT:   br i1 %1, label %while.entry, label %while.end
 //
 //      HIP: while.entry:
-// HIP-NEXT:   %entry1 = phi ptr [ @__start_hip_offloading_entries, %entry ], [ %16, %if.end ]
+// HIP-NEXT:   %entry1 = phi ptr [ @__start_llvm_offload_entries, %entry ], [ %16, %if.end ]
 // HIP-NEXT:   %2 = getelementptr inbounds %struct.__tgt_offload_entry, ptr %entry1, i32 0, i32 4
 // HIP-NEXT:   %addr = load ptr, ptr %2, align 8
 // HIP-NEXT:   %3 = getelementptr inbounds %struct.__tgt_offload_entry, ptr %entry1, i32 0, i32 8
@@ -257,7 +257,7 @@
 //
 //      HIP: if.end:
 // HIP-NEXT:   %16 = getelementptr inbounds %struct.__tgt_offload_entry, ptr %entry1, i64 1
-// HIP-NEXT:   %17 = icmp eq ptr %16, @__stop_hip_offloading_entries
+// HIP-NEXT:   %17 = icmp eq ptr %16, @__stop_llvm_offload_entries
 // HIP-NEXT:   br i1 %17, label %while.end, label %while.entry
 //
 //      HIP: while.end:

clang/test/Driver/linker-wrapper-image.c

@@ -27,11 +27,11 @@
 // HOST: [[SIZES:@.+]] = private unnamed_addr constant [3 x i64] [i64 4, i64 4, i64 4]
 // HOST: [[MAPTYPES:@.+]] = private unnamed_addr constant [3 x i64] [i64 35, i64 531, i64 531]
 // HOST: @.offloading.entry_name{{.*}} = internal unnamed_addr constant [{{[0-9]+}} x i8] c"c_decl_tgt_ref_ptr\00"
-// HOST: @.offloading.entry.c_decl_tgt_ref_ptr = weak constant %struct.__tgt_offload_entry { i64 0, i16 1, i16 1, i32 1, ptr @c_decl_tgt_ref_ptr, ptr @.offloading.entry_name, i64 8, i64 0, ptr null }, section "omp_offloading_entries", align 1 
-// HOST-COFF: @.offloading.entry.{{.*}} = weak constant %struct.__tgt_offload_entry { {{.*}} }, section "omp_offloading_entries$OE", align 1 
+// HOST: @.offloading.entry.c_decl_tgt_ref_ptr = weak constant %struct.__tgt_offload_entry { i64 0, i16 1, i16 1, i32 1, ptr @c_decl_tgt_ref_ptr, ptr @.offloading.entry_name, i64 8, i64 0, ptr null }, section "llvm_offload_entries"
+// HOST-COFF: @.offloading.entry.{{.*}} = weak constant %struct.__tgt_offload_entry { {{.*}} }, section "llvm_offload_entries$OE"
 // DEVICE-NOT: internal unnamed_addr constant [{{[0-9]+}} x i8] c"c_{{.*}}_decl_tgt_ref_ptr\00"
 // HOST: @.offloading.entry_name{{.*}} = internal unnamed_addr constant [{{[0-9]+}} x i8] c"_{{.*}}d_{{.*}}_decl_tgt_ref_ptr\00"
-// HOST: @.offloading.entry.[[D_PTR]] = weak constant %struct.__tgt_offload_entry { i64 0, i16 1, i16 1, i32 1, ptr @[[D_PTR]], ptr @.offloading.entry_name.3, i64 8, i64 0, ptr null }, section "omp_offloading_entries", align 1 
+// HOST: @.offloading.entry.[[D_PTR]] = weak constant %struct.__tgt_offload_entry { i64 0, i16 1, i16 1, i32 1, ptr @[[D_PTR]], ptr @.offloading.entry_name.3, i64 8, i64 0, ptr null }, section "llvm_offload_entries"
 
 extern int c;
 #pragma omp declare target link(c)

clang/test/OpenMP/declare_target_link_codegen.cpp

@@ -47,10 +47,10 @@ int bar(int n){
 // CHECK-HOST: [[OFFLOAD_MAPTYPES:@.+]] = private unnamed_addr constant [2 x i64] [i64 800, i64 800]
 
 // CHECK-HOST: [[OMP_OFFLOAD_ENTRY_LINK_VAR_PTR_NAME:@.+]] = internal unnamed_addr constant [21 x i8]
-// CHECK-HOST: [[OMP_OFFLOAD_ENTRY_LINK_VAR_PTR:@.+]] = weak constant %struct.__tgt_offload_entry { i64 0, i16 1, i16 1, i32 1, ptr [[VAR_DECL_TGT_LINK_PTR]], ptr [[OMP_OFFLOAD_ENTRY_LINK_VAR_PTR_NAME]], i64 8, i64 0, ptr null }, section "omp_offl
+// CHECK-HOST: [[OMP_OFFLOAD_ENTRY_LINK_VAR_PTR:@.+]] = weak constant %struct.__tgt_offload_entry { i64 0, i16 1, i16 1, i32 1, ptr [[VAR_DECL_TGT_LINK_PTR]], ptr [[OMP_OFFLOAD_ENTRY_LINK_VAR_PTR_NAME]], i64 8, i64 0, ptr null }, section "llvm_offload_entries"
 
 // CHECK-HOST: [[OMP_OFFLOAD_ENTRY_TO_VAR_PTR_NAME:@.+]] = internal unnamed_addr constant [24 x i8]
-// CHECK-HOST: [[OMP_OFFLOAD_ENTRY_TO_VAR_PTR:@.+]] = weak constant %struct.__tgt_offload_entry { i64 0, i16 1, i16 1, i32 0, ptr [[VAR_DECL_TGT_TO_PTR]], ptr [[OMP_OFFLOAD_ENTRY_TO_VAR_PTR_NAME]], i64 8, i64 0, ptr null }, section "omp_offloading_entries", align 1
+// CHECK-HOST: [[OMP_OFFLOAD_ENTRY_TO_VAR_PTR:@.+]] = weak constant %struct.__tgt_offload_entry { i64 0, i16 1, i16 1, i32 0, ptr [[VAR_DECL_TGT_TO_PTR]], ptr @.offloading.entry_name.1, i64 8, i64 0, ptr null }, section "llvm_offload_entries"
 
 // CHECK-HOST: [[N_CASTED:%.+]] = alloca i64
 // CHECK-HOST: [[SUM_CASTED:%.+]] = alloca i64

clang/test/OpenMP/nvptx_target_requires_unified_shared_memory.cpp

@@ -691,22 +691,19 @@ wrapDeviceImages(ArrayRef<std::unique_ptr<MemoryBuffer>> Buffers,
   switch (Kind) {
   case OFK_OpenMP:
     if (Error Err = offloading::wrapOpenMPBinaries(
-            M, BuffersToWrap,
-            offloading::getOffloadEntryArray(M, "omp_offloading_entries"),
+            M, BuffersToWrap, offloading::getOffloadEntryArray(M),
             /*Suffix=*/"", /*Relocatable=*/Args.hasArg(OPT_relocatable)))
       return std::move(Err);
     break;
   case OFK_Cuda:
     if (Error Err = offloading::wrapCudaBinary(
-            M, BuffersToWrap.front(),
-            offloading::getOffloadEntryArray(M, "cuda_offloading_entries"),
+            M, BuffersToWrap.front(), offloading::getOffloadEntryArray(M),
             /*Suffix=*/"", /*EmitSurfacesAndTextures=*/false))
       return std::move(Err);
     break;
   case OFK_HIP:
     if (Error Err = offloading::wrapHIPBinary(
-            M, BuffersToWrap.front(),
-            offloading::getOffloadEntryArray(M, "hip_offloading_entries")))
+            M, BuffersToWrap.front(), offloading::getOffloadEntryArray(M)))
       return std::move(Err);
     break;
   default:

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

@@ -82,8 +82,8 @@ StructType *getEntryTy(Module &M);
 /// \param AuxAddr An extra pointer if needed.
 void emitOffloadingEntry(Module &M, object::OffloadKind Kind, Constant *Addr,
                          StringRef Name, uint64_t Size, uint32_t Flags,
-                         uint64_t Data, StringRef SectionName,
-                         Constant *AuxAddr = nullptr);
+                         uint64_t Data, Constant *AuxAddr = nullptr,
+                         StringRef SectionName = "llvm_offload_entries");
 
 /// Create a constant struct initializer used to register this global at
 /// runtime.
@@ -96,7 +96,7 @@ getOffloadingEntryInitializer(Module &M, object::OffloadKind Kind,
 /// Creates a pair of globals used to iterate the array of offloading entries by
 /// accessing the section variables provided by the linker.
 std::pair<GlobalVariable *, GlobalVariable *>
-getOffloadEntryArray(Module &M, StringRef SectionName);
+getOffloadEntryArray(Module &M, StringRef SectionName = "llvm_offload_entries");
 
 namespace amdgpu {
 /// Check if an image is compatible with current system's environment. The

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

@@ -83,8 +83,8 @@ offloading::getOffloadingEntryInitializer(Module &M, object::OffloadKind Kind,
 void offloading::emitOffloadingEntry(Module &M, object::OffloadKind Kind,
                                      Constant *Addr, StringRef Name,
                                      uint64_t Size, uint32_t Flags,
-                                     uint64_t Data, StringRef SectionName,
-                                     Constant *AuxAddr) {
+                                     uint64_t Data, Constant *AuxAddr,
+                                     StringRef SectionName) {
   llvm::Triple Triple(M.getTargetTriple());
 
   auto [EntryInitializer, NameGV] = getOffloadingEntryInitializer(
@@ -103,7 +103,7 @@ void offloading::emitOffloadingEntry(Module &M, object::OffloadKind Kind,
     Entry->setSection((SectionName + "$OE").str());
   else
     Entry->setSection(SectionName);
-  Entry->setAlignment(Align(1));
+  Entry->setAlignment(Align(object::OffloadBinary::getAlignment()));
 }
 
 std::pair<GlobalVariable *, GlobalVariable *>
@@ -135,6 +135,7 @@ offloading::getOffloadEntryArray(Module &M, StringRef SectionName) {
         M, ZeroInitilaizer->getType(), true, GlobalVariable::InternalLinkage,
         ZeroInitilaizer, "__dummy." + SectionName);
     DummyEntry->setSection(SectionName);
+    DummyEntry->setAlignment(Align(object::OffloadBinary::getAlignment()));
     appendToCompilerUsed(M, DummyEntry);
   } else {
     // The COFF linker will merge sections containing a '$' together into a