mirror of
https://github.com/capstone-engine/llvm-capstone.git
synced 2024-11-23 22:00:10 +00:00
[OpenMP][libomptarget] Add utility header for AMDGPU plugins
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. Differential Revision: https://reviews.llvm.org/D139792
This commit is contained in:
parent
9a07780687
commit
39fe657b66
134
openmp/libomptarget/plugins-nextgen/amdgpu/utils/UtilitiesRTL.h
Normal file
134
openmp/libomptarget/plugins-nextgen/amdgpu/utils/UtilitiesRTL.h
Normal file
@ -0,0 +1,134 @@
|
||||
//===----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
|
@ -74,6 +74,7 @@ add_llvm_library(omptarget.rtl.amdgpu SHARED
|
||||
ADDITIONAL_HEADER_DIRS
|
||||
${LIBOMPTARGET_INCLUDE_DIR}
|
||||
${CMAKE_CURRENT_SOURCE_DIR}/impl
|
||||
${CMAKE_CURRENT_SOURCE_DIR}/../../plugins-nextgen/amdgpu/utils
|
||||
|
||||
LINK_COMPONENTS
|
||||
Support
|
||||
@ -96,6 +97,7 @@ target_include_directories(
|
||||
PRIVATE
|
||||
${LIBOMPTARGET_INCLUDE_DIR}
|
||||
${CMAKE_CURRENT_SOURCE_DIR}/impl
|
||||
${CMAKE_CURRENT_SOURCE_DIR}/../../plugins-nextgen/amdgpu/utils
|
||||
)
|
||||
|
||||
|
||||
|
@ -33,17 +33,6 @@
|
||||
|
||||
#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;
|
||||
|
@ -10,7 +10,6 @@
|
||||
//
|
||||
//===----------------------------------------------------------------------===//
|
||||
|
||||
#include "llvm/ADT/StringMap.h"
|
||||
#include "llvm/ADT/StringRef.h"
|
||||
#include "llvm/Frontend/OpenMP/OMPConstants.h"
|
||||
#include "llvm/Frontend/OpenMP/OMPGridValues.h"
|
||||
@ -34,6 +33,7 @@
|
||||
#include "impl_runtime.h"
|
||||
#include "interop_hsa.h"
|
||||
|
||||
#include "UtilitiesRTL.h"
|
||||
#include "internal.h"
|
||||
#include "rt.h"
|
||||
|
||||
@ -45,6 +45,7 @@
|
||||
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,
|
||||
@ -126,7 +127,7 @@ public:
|
||||
std::queue<int> FreeKernargSegments;
|
||||
|
||||
uint32_t kernargSizeIncludingImplicit() {
|
||||
return KernargSegmentSize + sizeof(impl_implicit_args_t);
|
||||
return KernargSegmentSize + sizeof(AMDGPUImplicitArgsTy);
|
||||
}
|
||||
|
||||
~KernelArgPool() {
|
||||
@ -1487,13 +1488,13 @@ int32_t runRegionLocked(int32_t DeviceId, void *TgtEntryPtr, void **TgtArgs,
|
||||
}
|
||||
|
||||
// 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)) {
|
||||
@ -1529,7 +1530,7 @@ int32_t runRegionLocked(int32_t DeviceId, void *TgtEntryPtr, void **TgtArgs,
|
||||
}
|
||||
|
||||
// initialise pointer for implicit_argument_count == 0 ABI
|
||||
ImplArgs->hostcall_ptr = Buffer;
|
||||
ImplArgs->HostcallPtr = Buffer;
|
||||
}
|
||||
|
||||
Packet->kernarg_address = KernArg;
|
||||
@ -1847,101 +1848,6 @@ static hsa_status_t GetIsaInfo(hsa_isa_t isa, void *data) {
|
||||
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);
|
||||
@ -1966,7 +1872,7 @@ int32_t __tgt_rtl_is_valid_binary_info(__tgt_device_image *image,
|
||||
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",
|
||||
|
Loading…
Reference in New Issue
Block a user