[OpenMP][libomptarget] Add utility header for AMDGPU plugins

[kevinsala]

Summary

This patch prepares the PluginInterface for the new AMDGPU NextGen plugin. The original and the NextGen plugin will share some structures and functionalities. We use this header for defining them and avoiding code duplication.

[tianshilei1992]

openmp/libomptarget/plugins-nextgen/amdgpu/utils/UtilitiesRTL.h

29

C++ style

openmp/libomptarget/plugins/amdgpu/CMakeLists.txt

100

Not related to this patch, but I think we should really set a CMake variable for the header and use the variable instead of relative path.

[tschuett]

openmp/libomptarget/plugins-nextgen/amdgpu/utils/UtilitiesRTL.h

22

nested namespaces?

[kevinsala]

openmp/libomptarget/plugins-nextgen/amdgpu/utils/UtilitiesRTL.h

22

Could you add more context to the comment please? Are nested namespaces a problem? Thanks

[kevinsala]

Using C++ style for struct.

[jdoerfert]

LG, we can modify it in-tree. We need to merge the plugin part to modify it w/ more ppl.

openmp/libomptarget/plugins-nextgen/amdgpu/utils/UtilitiesRTL.h

+//===----RTLs/amdgpu/utils/UtilitiesRTL.h ------------------------- 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
+//
+//===----------------------------------------------------------------------===//
+//
+// RTL Utilities for AMDGPU plugins
+//
+//===----------------------------------------------------------------------===//
+
+#include <cstdint>
+
+#include "Debug.h"
+#include "omptarget.h"
+
+#include "llvm/ADT/StringMap.h"
+#include "llvm/ADT/StringRef.h"
+
+namespace llvm {
+namespace omp {
+namespace target {
+namespace plugin {
+namespace utils {
+
+// The implicit arguments of AMDGPU kernels.
+struct AMDGPUImplicitArgsTy {
+  uint64_t OffsetX;
+  uint64_t OffsetY;
+  uint64_t OffsetZ;
+  uint64_t HostcallPtr;
+  uint64_t Unused0;
+  uint64_t Unused1;
+  uint64_t Unused2;
+};
+
+static_assert(sizeof(AMDGPUImplicitArgsTy) == 56,
+              "Unexpected size of implicit arguments");
+
+/// Parse a TargetID to get processor arch and feature map.
+/// Returns processor subarch.
+/// Returns TargetID features in \p FeatureMap argument.
+/// If the \p TargetID contains feature+, FeatureMap it to true.
+/// If the \p TargetID contains feature-, FeatureMap it to false.
+/// If the \p TargetID does not contain a feature (default), do not map it.
+StringRef parseTargetID(StringRef TargetID, StringMap<bool> &FeatureMap) {
+  if (TargetID.empty())
+    return llvm::StringRef();
+
+  auto ArchFeature = TargetID.split(":");
+  auto Arch = ArchFeature.first;
+  auto Features = ArchFeature.second;
+  if (Features.empty())
+    return Arch;
+
+  if (Features.contains("sramecc+")) {
+    FeatureMap.insert(std::pair<StringRef, bool>("sramecc", true));
+  } else if (Features.contains("sramecc-")) {
+    FeatureMap.insert(std::pair<StringRef, bool>("sramecc", false));
+  }
+  if (Features.contains("xnack+")) {
+    FeatureMap.insert(std::pair<StringRef, bool>("xnack", true));
+  } else if (Features.contains("xnack-")) {
+    FeatureMap.insert(std::pair<StringRef, bool>("xnack", false));
+  }
+
+  return Arch;
+}
+
+/// Check if an image is compatible with current system's environment.
+bool isImageCompatibleWithEnv(const __tgt_image_info *Info,
+                              StringRef EnvTargetID) {
+  llvm::StringRef ImageTargetID(Info->Arch);
+
+  // Compatible in case of exact match.
+  if (ImageTargetID == EnvTargetID) {
+    DP("Compatible: Exact match \t[Image: %s]\t:\t[Env: %s]\n",
+       ImageTargetID.data(), EnvTargetID.data());
+    return true;
+  }
+
+  // Incompatible if Archs mismatch.
+  StringMap<bool> ImgMap, EnvMap;
+  StringRef ImgArch = utils::parseTargetID(ImageTargetID, ImgMap);
+  StringRef EnvArch = utils::parseTargetID(EnvTargetID, EnvMap);
+
+  // Both EnvArch and ImgArch can't be empty here.
+  if (EnvArch.empty() || ImgArch.empty() || !ImgArch.contains(EnvArch)) {
+    DP("Incompatible: Processor mismatch \t[Image: %s]\t:\t[Env: %s]\n",
+       ImageTargetID.data(), EnvTargetID.data());
+    return false;
+  }
+
+  // Incompatible if image has more features than the environment,
+  // irrespective of type or sign of features.
+  if (ImgMap.size() > EnvMap.size()) {
+    DP("Incompatible: Image has more features than the Environment \t[Image: "
+       "%s]\t:\t[Env: %s]\n",
+       ImageTargetID.data(), EnvTargetID.data());
+    return false;
+  }
+
+  // Compatible if each target feature specified by the environment is
+  // compatible with target feature of the image. The target feature is
+  // compatible if the iamge does not specify it (meaning Any), or if it
+  // specifies it with the same value (meaning On or Off).
+  for (const auto &ImgFeature : ImgMap) {
+    auto EnvFeature = EnvMap.find(ImgFeature.first());
+    if (EnvFeature == EnvMap.end() ||
+        (EnvFeature->first() == ImgFeature.first() &&
+         EnvFeature->second != ImgFeature.second)) {
+      DP("Incompatible: Value of Image's non-ANY feature is not matching with "
+         "the Environment's non-ANY feature \t[Image: %s]\t:\t[Env: %s]\n",
+         ImageTargetID.data(), EnvTargetID.data());
+      return false;
+    }
+  }
+
+  // Image is compatible if all features of Environment are:
+  //   - either, present in the Image's features map with the same sign,
+  //   - or, the feature is missing from Image's features map i.e. it is
+  //   set to ANY
+  DP("Compatible: Target IDs are compatible \t[Image: %s]\t:\t[Env: %s]\n",
+     ImageTargetID.data(), EnvTargetID.data());
+
+  return true;
+}
+
+} // namespace utils
+} // namespace plugin
+} // namespace target
+} // namespace omp
+} // namespace llvm

openmp/libomptarget/plugins/amdgpu/CMakeLists.txt

add_llvm_library(omptarget.rtl.amdgpu SHARED
   impl/system.cpp
   impl/msgpack.cpp
   src/rtl.cpp
   ${LIBOMPTARGET_EXTRA_SOURCE}
 
   ADDITIONAL_HEADER_DIRS
   ${LIBOMPTARGET_INCLUDE_DIR}
   ${CMAKE_CURRENT_SOURCE_DIR}/impl
+  ${CMAKE_CURRENT_SOURCE_DIR}/../../plugins-nextgen/amdgpu/utils
 
   LINK_COMPONENTS
   Support
   Object
 
   LINK_LIBS 
   PRIVATE
   elf_common
   ${LIBOMPTARGET_DEP_LIBRARIES}
   ${OPENMP_PTHREAD_LIB}
   "-Wl,--version-script=${CMAKE_CURRENT_SOURCE_DIR}/../exports"
   ${LDFLAGS_UNDEFINED}
 
   NO_INSTALL_RPATH
 )
 add_dependencies(omptarget.rtl.amdgpu omptarget.devicertl.amdgpu)
 
 target_include_directories(
   omptarget.rtl.amdgpu
   PRIVATE
   ${LIBOMPTARGET_INCLUDE_DIR}
   ${CMAKE_CURRENT_SOURCE_DIR}/impl
+  ${CMAKE_CURRENT_SOURCE_DIR}/../../plugins-nextgen/amdgpu/utils
 )
 
 
 # Install plugin under the lib destination folder.
 install(TARGETS omptarget.rtl.amdgpu LIBRARY DESTINATION "${OPENMP_INSTALL_LIBDIR}")
 set_target_properties(omptarget.rtl.amdgpu PROPERTIES 
   INSTALL_RPATH "$ORIGIN" BUILD_RPATH "$ORIGIN:${CMAKE_CURRENT_BINARY_DIR}/.."
   CXX_VISIBILITY_PRESET protected)

openmp/libomptarget/plugins/amdgpu/impl/internal.h

 #ifndef TARGET_NAME
 #error "Missing TARGET_NAME macro"
 #endif
 #define DEBUG_PREFIX "Target " GETNAME(TARGET_NAME) " RTL"
 #include "Debug.h"
 
 #define MAX_NUM_KERNELS (1024 * 16)
 
-typedef struct impl_implicit_args_s {
-  uint64_t offset_x;
-  uint64_t offset_y;
-  uint64_t offset_z;
-  uint64_t hostcall_ptr;
-  uint64_t unused0;
-  uint64_t unused1;
-  uint64_t unused2;
-} impl_implicit_args_t;
-static_assert(sizeof(impl_implicit_args_t) == 56, "");
-
 // ---------------------- Kernel Start -------------
 typedef struct atl_kernel_info_s {
   uint64_t kernel_object;
   uint32_t group_segment_size;
   uint32_t private_segment_size;
   uint32_t sgpr_count;
   uint32_t vgpr_count;
   uint32_t sgpr_spill_count;

openmp/libomptarget/plugins/amdgpu/src/rtl.cpp

 //===--- amdgpu/src/rtl.cpp --------------------------------------- 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
 //
 //===----------------------------------------------------------------------===//
 //
 // RTL for AMD hsa machine
 //
 //===----------------------------------------------------------------------===//
 
-#include "llvm/ADT/StringMap.h"
 #include "llvm/ADT/StringRef.h"
 #include "llvm/Frontend/OpenMP/OMPConstants.h"
 #include "llvm/Frontend/OpenMP/OMPGridValues.h"
 #include "llvm/Object/ELF.h"
 #include "llvm/Object/ELFObjectFile.h"
 
 #include <algorithm>
 #include <assert.h>
 #include <cstdio>
 #include <cstdlib>
 #include <cstring>
 #include <functional>
 #include <list>
 #include <memory>
 #include <mutex>
 #include <shared_mutex>
 #include <unordered_map>
 #include <vector>
 
 #include "ELFSymbols.h"
 #include "impl_runtime.h"
 #include "interop_hsa.h"
 
+#include "UtilitiesRTL.h"
 #include "internal.h"
 #include "rt.h"
 
 #include "DeviceEnvironment.h"
 #include "get_elf_mach_gfx_name.h"
 #include "omptargetplugin.h"
 #include "print_tracing.h"
 
 using namespace llvm;
 using namespace llvm::object;
 using namespace llvm::ELF;
+using namespace llvm::omp::target::plugin::utils;
 
 // hostrpc interface, FIXME: consider moving to its own include these are
 // statically linked into amdgpu/plugin if present from hostrpc_services.a,
 // linked as --whole-archive to override the weak symbols that are used to
 // implement a fallback for toolchains that do not yet have a hostrpc library.
 extern "C" {
 uint64_t hostrpc_assign_buffer(hsa_agent_t Agent, hsa_queue_t *ThisQ,
                                uint32_t DeviceId);
private:
   static pthread_mutex_t Mutex;
 
 public:
   uint32_t KernargSegmentSize;
   void *KernargRegion = nullptr;
   std::queue<int> FreeKernargSegments;
 
   uint32_t kernargSizeIncludingImplicit() {
-    return KernargSegmentSize + sizeof(impl_implicit_args_t);
+    return KernargSegmentSize + sizeof(AMDGPUImplicitArgsTy);
   }
 
   ~KernelArgPool() {
     if (KernargRegion) {
       auto R = hsa_amd_memory_pool_free(KernargRegion);
       if (R != HSA_STATUS_SUCCESS) {
         DP("hsa_amd_memory_pool_free failed: %s\n", get_error_string(R));
       }
if (!ArgPool) {
       }
 
       // Copy explicit arguments
       for (int I = 0; I < ArgNum; I++) {
         memcpy((char *)KernArg + sizeof(void *) * I, Args[I], sizeof(void *));
       }
 
       // Initialize implicit arguments. TODO: Which of these can be dropped
-      impl_implicit_args_t *ImplArgs = reinterpret_cast<impl_implicit_args_t *>(
+      AMDGPUImplicitArgsTy *ImplArgs = reinterpret_cast<AMDGPUImplicitArgsTy *>(
           static_cast<char *>(KernArg) + ArgPool->KernargSegmentSize);
       memset(ImplArgs, 0,
-             sizeof(impl_implicit_args_t)); // may not be necessary
-      ImplArgs->offset_x = 0;
-      ImplArgs->offset_y = 0;
-      ImplArgs->offset_z = 0;
+             sizeof(AMDGPUImplicitArgsTy)); // may not be necessary
+      ImplArgs->OffsetX = 0;
+      ImplArgs->OffsetY = 0;
+      ImplArgs->OffsetZ = 0;
 
       // assign a hostcall buffer for the selected Q
       if (__atomic_load_n(&DeviceInfo().HostcallRequired, __ATOMIC_ACQUIRE)) {
         // hostrpc_assign_buffer is not thread safe, and this function is
         // under a multiple reader lock, not a writer lock.
         static pthread_mutex_t HostcallInitLock = PTHREAD_MUTEX_INITIALIZER;
         pthread_mutex_lock(&HostcallInitLock);
         uint64_t Buffer = hostrpc_assign_buffer(
if (!ArgPool) {
                "args: %d\n",
                Offset + 8, ArgPool->kernargSizeIncludingImplicit());
           } else {
             memcpy(static_cast<char *>(KernArg) + Offset, &Buffer, 8);
           }
         }
 
         // initialise pointer for implicit_argument_count == 0 ABI
-        ImplArgs->hostcall_ptr = Buffer;
+        ImplArgs->HostcallPtr = Buffer;
       }
 
       Packet->kernarg_address = KernArg;
     }
 
     hsa_signal_t S = DeviceInfo().FreeSignalPool.pop();
     if (S.handle == 0) {
       DP("Failed to get signal instance\n");
static hsa_status_t GetIsaInfo(hsa_isa_t isa, void *data) {
 
   auto TripleTargetID = llvm::StringRef(TargetID);
   if (TripleTargetID.consume_front("amdgcn-amd-amdhsa")) {
     DeviceInfo().TargetID.push_back(TripleTargetID.ltrim('-').str());
   }
   return HSA_STATUS_SUCCESS;
 }
 
-/// Parse a TargetID to get processor arch and feature map.
-/// Returns processor subarch.
-/// Returns TargetID features in \p FeatureMap argument.
-/// If the \p TargetID contains feature+, FeatureMap it to true.
-/// If the \p TargetID contains feature-, FeatureMap it to false.
-/// If the \p TargetID does not contain a feature (default), do not map it.
-StringRef parseTargetID(StringRef TargetID, StringMap<bool> &FeatureMap) {
-  if (TargetID.empty())
-    return llvm::StringRef();
-
-  auto ArchFeature = TargetID.split(":");
-  auto Arch = ArchFeature.first;
-  auto Features = ArchFeature.second;
-  if (Features.empty())
-    return Arch;
-
-  if (Features.contains("sramecc+")) {
-    FeatureMap.insert(std::pair<std::string, bool>("sramecc", true));
-  } else if (Features.contains("sramecc-")) {
-    FeatureMap.insert(std::pair<std::string, bool>("sramecc", false));
-  }
-  if (Features.contains("xnack+")) {
-    FeatureMap.insert(std::pair<std::string, bool>("xnack", true));
-  } else if (Features.contains("xnack-")) {
-    FeatureMap.insert(std::pair<std::string, bool>("xnack", false));
-  }
-
-  return Arch;
-}
-
-/// Checks if an image \p ImgInfo is compatible with current
-/// system's environment \p EnvInfo
-bool IsImageCompatibleWithEnv(const char *ImgInfo, std::string EnvInfo) {
-  llvm::StringRef ImgTID(ImgInfo), EnvTID(EnvInfo);
-
-  // Compatible in case of exact match
-  if (ImgTID == EnvTID) {
-    DP("Compatible: Exact match \t[Image: %s]\t:\t[Environment: %s]\n",
-       ImgTID.data(), EnvTID.data());
-    return true;
-  }
-
-  // Incompatible if Archs mismatch.
-  StringMap<bool> ImgMap, EnvMap;
-  StringRef ImgArch = parseTargetID(ImgTID, ImgMap);
-  StringRef EnvArch = parseTargetID(EnvTID, EnvMap);
-
-  // Both EnvArch and ImgArch can't be empty here.
-  if (EnvArch.empty() || ImgArch.empty() || !ImgArch.contains(EnvArch)) {
-    DP("Incompatible: Processor mismatch \t[Image: %s]\t:\t[Environment: %s]\n",
-       ImgTID.data(), EnvTID.data());
-    return false;
-  }
-
-  // Incompatible if image has more features than the environment, irrespective
-  // of type or sign of features.
-  if (ImgMap.size() > EnvMap.size()) {
-    DP("Incompatible: Image has more features than the environment \t[Image: "
-       "%s]\t:\t[Environment: %s]\n",
-       ImgTID.data(), EnvTID.data());
-    return false;
-  }
-
-  // Compatible if each target feature specified by the environment is
-  // compatible with target feature of the image. The target feature is
-  // compatible if the iamge does not specify it (meaning Any), or if it
-  // specifies it with the same value (meaning On or Off).
-  for (const auto &ImgFeature : ImgMap) {
-    auto EnvFeature = EnvMap.find(ImgFeature.first());
-    if (EnvFeature == EnvMap.end()) {
-      DP("Incompatible: Value of Image's non-ANY feature is not matching with "
-         "the Environment feature's ANY value \t[Image: %s]\t:\t[Environment: "
-         "%s]\n",
-         ImgTID.data(), EnvTID.data());
-      return false;
-    } else if (EnvFeature->first() == ImgFeature.first() &&
-               EnvFeature->second != ImgFeature.second) {
-      DP("Incompatible: Value of Image's non-ANY feature is not matching with "
-         "the Environment feature's non-ANY value \t[Image: "
-         "%s]\t:\t[Environment: %s]\n",
-         ImgTID.data(), EnvTID.data());
-      return false;
-    }
-  }
-
-  // Image is compatible if all features of Environment are:
-  //   - either, present in the Image's features map with the same sign,
-  //   - or, the feature is missing from Image's features map i.e. it is
-  //   set to ANY
-  DP("Compatible: Target IDs are compatible \t[Image: %s]\t:\t[Environment: "
-     "%s]\n",
-     ImgTID.data(), EnvTID.data());
-  return true;
-}
-
 extern "C" {
 int32_t __tgt_rtl_is_valid_binary(__tgt_device_image *Image) {
   return elfMachineIdIsAmdgcn(Image);
 }
 
 int32_t __tgt_rtl_is_valid_binary_info(__tgt_device_image *image,
                                        __tgt_image_info *info) {
   if (!__tgt_rtl_is_valid_binary(image))
     return false;
 
   // A subarchitecture was not specified. Assume it is compatible.
   if (!info->Arch)
     return true;
 
   int32_t NumberOfDevices = __tgt_rtl_number_of_devices();
 
   for (int32_t DeviceId = 0; DeviceId < NumberOfDevices; ++DeviceId) {
     __tgt_rtl_init_device(DeviceId);
     hsa_agent_t agent = DeviceInfo().HSAAgents[DeviceId];
     hsa_status_t err = hsa_agent_iterate_isas(agent, GetIsaInfo, &DeviceId);
     if (err != HSA_STATUS_SUCCESS) {
       DP("Error iterating ISAs\n");
       return false;
     }
-    if (!IsImageCompatibleWithEnv(info->Arch, DeviceInfo().TargetID[DeviceId]))
+    if (!isImageCompatibleWithEnv(info, DeviceInfo().TargetID[DeviceId]))
       return false;
   }
   DP("Image has Target ID compatible with the current environment: %s\n",
      info->Arch);
   return true;
 }
 
 int32_t __tgt_rtl_init_plugin() { return OFFLOAD_SUCCESS; }