Merge branch 'sycl' of https://github.com/intel/llvm into sycl

Author: bso-intel

README.md

@@ -6,7 +6,7 @@ Intel staging area for llvm.org contribution.
 Home for Intel LLVM-based projects:
  - oneAPI Data Parallel C++ compiler - see **sycl** branch. More information on
    oneAPI and DPC++ is available at
-([https://www.oneapi.com/](https://www.oneapi.com/))
+[https://www.oneapi.com/](https://www.oneapi.com/)
    - [![Linux Post Commit Checks](https://github.com/intel/llvm/workflows/Linux%20Post%20Commit%20Checks/badge.svg)](https://github.com/intel/llvm/actions?query=workflow%3A%22Linux+Post+Commit+Checks%22)
    [![Generate Doxygen documentation](https://github.com/intel/llvm/workflows/Generate%20Doxygen%20documentation/badge.svg)](https://github.com/intel/llvm/actions?query=workflow%3A%22Generate+Doxygen+documentation%22)
 

libdevice/device.h

@@ -9,10 +9,6 @@
 #ifndef __LIBDEVICE_DEVICE_H__
 #define __LIBDEVICE_DEVICE_H__
 
-// We need the following header to ensure the definition of all spirv variables
-// required by the wrapper libraries.
-#include "spirv_vars.hpp"
-
 #ifdef __cplusplus
 #define EXTERN_C extern "C"
 #else // __cplusplus
@@ -37,4 +33,8 @@
 
 #define DEVICE_EXTERN_C DEVICE_EXTERNAL EXTERN_C
 
+// We need the following header to ensure the definition of all spirv variables
+// required by the wrapper libraries.
+#include "spirv_vars.hpp"
+
 #endif // __LIBDEVICE_DEVICE_H__

libdevice/spirv_vars.hpp

@@ -15,37 +15,37 @@
 
 #ifdef __SYCL_NVPTX__
 
-SYCL_EXTERNAL size_t __spirv_GlobalInvocationId_x();
-SYCL_EXTERNAL size_t __spirv_GlobalInvocationId_y();
-SYCL_EXTERNAL size_t __spirv_GlobalInvocationId_z();
+DEVICE_EXTERNAL size_t __spirv_GlobalInvocationId_x();
+DEVICE_EXTERNAL size_t __spirv_GlobalInvocationId_y();
+DEVICE_EXTERNAL size_t __spirv_GlobalInvocationId_z();
 
-SYCL_EXTERNAL size_t __spirv_LocalInvocationId_x();
-SYCL_EXTERNAL size_t __spirv_LocalInvocationId_y();
-SYCL_EXTERNAL size_t __spirv_LocalInvocationId_z();
+DEVICE_EXTERNAL size_t __spirv_LocalInvocationId_x();
+DEVICE_EXTERNAL size_t __spirv_LocalInvocationId_y();
+DEVICE_EXTERNAL size_t __spirv_LocalInvocationId_z();
 
 #else // __SYCL_NVPTX__
 
 typedef size_t size_t_vec __attribute__((ext_vector_type(3)));
 extern "C" const size_t_vec __spirv_BuiltInGlobalInvocationId;
 extern "C" const size_t_vec __spirv_BuiltInLocalInvocationId;
 
-SYCL_EXTERNAL inline size_t __spirv_GlobalInvocationId_x() {
+DEVICE_EXTERNAL inline size_t __spirv_GlobalInvocationId_x() {
   return __spirv_BuiltInGlobalInvocationId.x;
 }
-SYCL_EXTERNAL inline size_t __spirv_GlobalInvocationId_y() {
+DEVICE_EXTERNAL inline size_t __spirv_GlobalInvocationId_y() {
   return __spirv_BuiltInGlobalInvocationId.y;
 }
-SYCL_EXTERNAL inline size_t __spirv_GlobalInvocationId_z() {
+DEVICE_EXTERNAL inline size_t __spirv_GlobalInvocationId_z() {
   return __spirv_BuiltInGlobalInvocationId.z;
 }
 
-SYCL_EXTERNAL inline size_t __spirv_LocalInvocationId_x() {
+DEVICE_EXTERNAL inline size_t __spirv_LocalInvocationId_x() {
   return __spirv_BuiltInLocalInvocationId.x;
 }
-SYCL_EXTERNAL inline size_t __spirv_LocalInvocationId_y() {
+DEVICE_EXTERNAL inline size_t __spirv_LocalInvocationId_y() {
   return __spirv_BuiltInLocalInvocationId.y;
 }
-SYCL_EXTERNAL inline size_t __spirv_LocalInvocationId_z() {
+DEVICE_EXTERNAL inline size_t __spirv_LocalInvocationId_z() {
   return __spirv_BuiltInLocalInvocationId.z;
 }
 

sycl/source/detail/config.cpp

@@ -27,7 +27,7 @@ namespace detail {
 
 #define CONFIG(Name, MaxSize, CompileTimeDef)                                  \
   const char *SYCLConfigBase<Name>::MValueFromFile = nullptr;                  \
-  char SYCLConfigBase<Name>::MStorage[MaxSize];                                \
+  char SYCLConfigBase<Name>::MStorage[MaxSize + 1];                            \
   const char *const SYCLConfigBase<Name>::MCompileTimeDef =                    \
       getStrOrNullptr(STRINGIFY_LINE(CompileTimeDef));                         \
   const char *const SYCLConfigBase<Name>::MConfigName = STRINGIFY_LINE(Name);
@@ -40,6 +40,7 @@ static void initValue(const char *Key, const char *Value) {
 #define CONFIG(Name, MaxSize, CompileTimeDef)                                  \
   if (0 == strncmp(Key, SYCLConfigBase<Name>::MConfigName, MAX_CONFIG_NAME)) { \
     strncpy(SYCLConfigBase<Name>::MStorage, Value, MaxSize);                   \
+    SYCLConfigBase<Name>::MStorage[MaxSize] = '\0';                            \
     SYCLConfigBase<Name>::MValueFromFile = SYCLConfigBase<Name>::MStorage;     \
     return;                                                                    \
   }

sycl/source/detail/config.hpp

@@ -63,7 +63,7 @@ template <ConfigID Config> class SYCLConfigBase;
   public:                                                                      \
     /*Preallocated storage for config value which is extracted from a config   \
      * file*/                                                                  \
-    static char MStorage[MaxSize];                                             \
+    static char MStorage[MaxSize + 1];                                         \
     /*Points to the storage if config is set in the file, nullptr otherwise*/  \
     static const char *MValueFromFile;                                         \
     /*The name of the config*/                                                 \

sycl/test/abi/sycl_symbols_linux.dump

@@ -1,4 +1,4 @@
-# RUN: env LLVM_BIN_PATH=%llvm_build_bin_dir python %sycl_tools_src_dir/abi_check.py --mode check_symbols --reference %s %llvm_build_lib_dir/libsycl.so
+# RUN: env LLVM_BIN_PATH=%llvm_build_bin_dir python %sycl_tools_src_dir/abi_check.py --mode check_symbols --reference %s %sycl_libs_dir/libsycl.so
 # REQUIRES: linux
 
 _Z20__spirv_ocl_prefetchPKcm

sycl/test/basic_tests/image.cpp

@@ -1,12 +1,11 @@
+// UNSUPPORTED: cuda
+// CUDA cannot support SYCL 1.2.1 images.
+//
 // RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
 // RUN: env SYCL_DEVICE_TYPE=HOST %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 // RUN: %GPU_RUN_PLACEHOLDER %t.out
 
-// TODO: No CUDA image support
-// TODO: ptxas fatal   : Unresolved extern function '_Z17__spirv_ImageReadIDv4_f14ocl_image2d_roDv2_iET_T0_T1_'
-// XFAIL: cuda
-
 //==------------------- image.cpp - SYCL image basic test -----------------==//
 //
 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.

sycl/test/basic_tests/image_accessor_readsampler.cpp

@@ -1,8 +1,10 @@
-// RUN: %clangxx -fsycl %s -o %t.out
+// UNSUPPORTED: cuda
+// CUDA cannot support SYCL 1.2.1 images.
+//
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
 // RUN: env SYCL_DEVICE_TYPE=HOST %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 // RUN: %GPU_RUN_PLACEHOLDER %t.out
-// REQUIRES: opencl
 //==------------------- image_accessor_readsampler.cpp ---------------------==//
 //==-----------------image_accessor read API test with sampler--------------==//
 //

sycl/test/basic_tests/image_accessor_readwrite.cpp

@@ -1,9 +1,11 @@
-// RUN: %clangxx -fsycl %s -o %t.out
+// UNSUPPORTED: cuda
+// CUDA cannot support SYCL 1.2.1 images.
+//
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
 // RUN: env SYCL_DEVICE_TYPE=HOST %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 // RUN: %GPU_RUN_PLACEHOLDER %t.out
 
-// UNSUPPORTED: cuda
 //==--------------------image_accessor_readwrite.cpp ----------------------==//
 //==----------image_accessor read without sampler & write API test---------==//
 //

sycl/test/basic_tests/image_accessor_readwrite_half.cpp

@@ -1,9 +1,11 @@
-// RUN: %clangxx -fsycl %s -o %t.out
+// UNSUPPORTED: cuda
+// CUDA cannot support SYCL 1.2.1 images.
+//
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
 // RUN: env SYCL_DEVICE_TYPE=HOST %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 // RUN: %GPU_RUN_PLACEHOLDER %t.out
 
-// UNSUPPORTED: cuda
 //==--------------------image_accessor_readwrite_half.cpp -------------------==//
 //==-image_accessor read (without sampler)& write API test for half datatype-==//
 //

sycl/test/basic_tests/image_array.cpp

@@ -1,12 +1,11 @@
+// UNSUPPORTED: cuda
+// CUDA cannot support SYCL 1.2.1 images.
+//
 // RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
 // RUNx: env SYCL_DEVICE_TYPE=HOST %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 // RUNx: %GPU_RUN_PLACEHOLDER %t.out
 
-// TODO: No CUDA image support
-// TODO: ptxas fatal   : Unresolved extern function '_Z17__spirv_ImageReadIDv4_f14ocl_image2d_roDv2_iET_T0_T1_'
-// XFAIL: cuda
-
 //==------------------- image.cpp - SYCL image basic test -----------------==//
 //
 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.

sycl/test/basic_tests/image_constructors.cpp

@@ -1,3 +1,6 @@
+// UNSUPPORTED: cuda
+// CUDA cannot support SYCL 1.2.1 images.
+//
 // RUN: %clangxx %s -o %t1.out -lsycl -I %sycl_include
 // RUN: env SYCL_DEVICE_TYPE=HOST %t1.out
 // RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t2.out

sycl/test/basic_tests/stream/auto_flush.cpp

@@ -1,10 +1,8 @@
-// RUN: %clangxx -fsycl %s -o %t.out
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
 // TODO: Enable on host when commands cleanup will be implemented in scheduler
 // RUN: %CPU_RUN_PLACEHOLDER %t.out %CPU_CHECK_PLACEHOLDER
 // RUN: %GPU_RUN_ON_LINUX_PLACEHOLDER %t.out %GPU_CHECK_ON_LINUX_PLACEHOLDER
 // RUN: %ACC_RUN_PLACEHOLDER %t.out %ACC_CHECK_PLACEHOLDER
-// XFAIL: cuda
-// cuda fail due to unimplemented param_name 4131 in cuda_piDeviceGetInfo
 //==-------------- copy.cpp - SYCL stream obect auto flushing test ---------==//
 //
 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.

sycl/test/built-ins/printf.cpp

@@ -1,11 +1,12 @@
-// RUN: %clangxx -fsycl %s -o %t.out
+// UNSUPPORTED: cuda
+// CUDA does not support printf.
+//
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
 // RUN: env SYCL_DEVICE_TYPE=HOST %t.out | FileCheck %s
 // RUN: %CPU_RUN_PLACEHOLDER %t.out %CPU_CHECK_PLACEHOLDER
 // RUN: %GPU_RUN_PLACEHOLDER %t.out %GPU_CHECK_PLACEHOLDER
 // RUN: %ACC_RUN_PLACEHOLDER %t.out %ACC_CHECK_PLACEHOLDER
 
-// XFAIL: cuda
-
 #include <CL/sycl.hpp>
 
 #include <cstdint>

sycl/test/device-code-split/aot-gpu.cpp

@@ -1,6 +1,11 @@
 // REQUIRES: ocloc, gpu
-
-// RUN: %clangxx -fsycl -fsycl-device-code-split=per_source -fsycl-targets=spir64_gen-unknown-unknown-sycldevice -Xsycl-target-backend=spir64_gen-unknown-unknown-sycldevice "-device skl" -I %S/Inputs -o %t.out %S/split-per-source-main.cpp %S/Inputs/split-per-source-second-file.cpp
+// UNSUPPORTED: cuda
+// CUDA does neither support device code splitting nor SPIR.
+//
+// RUN: %clangxx -fsycl -fsycl-device-code-split=per_source \
+// RUN:   -fsycl-targets=spir64_gen-unknown-unknown-sycldevice \
+// RUN:   -Xsycl-target-backend=spir64_gen-unknown-unknown-sycldevice \
+// RUN:   "-device skl" -I %S/Inputs -o %t.out \
+// RUN:   %S/split-per-source-main.cpp \
+// RUN:   %S/Inputs/split-per-source-second-file.cpp
 // RUN: %GPU_RUN_PLACEHOLDER %t.out
-
-// XFAIL: cuda

sycl/test/device-code-split/split-per-kernel.cpp

@@ -1,10 +1,11 @@
-// RUN: %clangxx -fsycl -fsycl-device-code-split=per_kernel -o %t.out %s
+// UNSUPPORTED: cuda
+// CUDA does not support device code splitting.
+//
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple -fsycl-device-code-split=per_kernel -o %t.out %s
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 // RUN: %GPU_RUN_PLACEHOLDER %t.out
 // RUN: %ACC_RUN_PLACEHOLDER %t.out
 
-// XFAIL: cuda
-
 #include <CL/sycl.hpp>
 
 class Kern1;

sycl/test/device-code-split/split-per-source-main.cpp

@@ -1,10 +1,11 @@
-// RUN: %clangxx -fsycl -fsycl-device-code-split=per_source -I %S/Inputs -o %t.out %s %S/Inputs/split-per-source-second-file.cpp
+// UNSUPPORTED: cuda
+// CUDA does not support device code splitting.
+//
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple -fsycl-device-code-split=per_source -I %S/Inputs -o %t.out %s %S/Inputs/split-per-source-second-file.cpp
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 // RUN: %GPU_RUN_PLACEHOLDER %t.out
 // RUN: %ACC_RUN_PLACEHOLDER %t.out
 
-// XFAIL: cuda
-
 #include "Inputs/split-per-source.h"
 
 int main () {

sycl/test/ordered_queue/oq_kernels.cpp

@@ -1,9 +1,11 @@
-// RUN: %clangxx -fsycl -fsycl-unnamed-lambda %s -o %t.out
+// UNSUPPORTED: cuda
+// CUDA does not support unnamed lambdas.
+//
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple -fsycl-unnamed-lambda %s -o %t.out
 // RUN: env SYCL_DEVICE_TYPE=HOST %t.out
 // RUN: %ACC_RUN_PLACEHOLDER %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 // RUN: %GPU_RUN_PLACEHOLDER %t.out
-// XFAIL: cuda
 
 //==------ oq_kernels.cpp - SYCL ordered queue kernel shortcut test --------==//
 //

sycl/test/ordered_queue/ordered_buffs.cpp

@@ -1,10 +1,9 @@
 // REQUIRES: opencl
 
-// RUN: %clangxx -fsycl %s -o %t.out -L %opencl_libs_dir -lOpenCL
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out -L %opencl_libs_dir -lOpenCL
 // RUN: %ACC_RUN_PLACEHOLDER %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 // RUN: %GPU_RUN_PLACEHOLDER %t.out
-// XFAIL: cuda
 //==-------- ordered_buffs.cpp - SYCL buffers in ordered queues test--------==//
 //
 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.

sycl/test/ordered_queue/ordered_dmemll.cpp

@@ -1,6 +1,6 @@
 // REQUIRES: opencl
 
-// RUN: %clangxx -fsycl %s -o %t1.out -L %opencl_libs_dir -lOpenCL
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t1.out -L %opencl_libs_dir -lOpenCL
 // RUN: %CPU_RUN_PLACEHOLDER %t1.out
 // RUN: %GPU_RUN_PLACEHOLDER %t1.out
 //==----------- ordered_dmemll.cpp - Device Memory Linked List test --------==//

sycl/test/ordered_queue/ordered_queue.cpp

@@ -1,6 +1,6 @@
 // REQUIRES: opencl
 //
-// RUN: %clangxx -fsycl %s -o %t.out -L %opencl_libs_dir -lOpenCL
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out -L %opencl_libs_dir -lOpenCL
 // RUN: env SYCL_DEVICE_TYPE=HOST %t.out
 //==---------- ordered_queue.cpp - SYCL ordered queue test -----------------==//
 //

sycl/test/ordered_queue/prop.cpp

@@ -1,6 +1,6 @@
 // REQUIRES: opencl
 
-// RUN: %clangxx -fsycl %s -o %t1.out -L %opencl_libs_dir -lOpenCL
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t1.out -L %opencl_libs_dir -lOpenCL
 // RUN: %CPU_RUN_PLACEHOLDER %t1.out
 // RUN: %GPU_RUN_PLACEHOLDER %t1.out
 

sycl/test/separate-compile/sycl-external.cpp

@@ -1,8 +1,8 @@
 // Test1 - check that kernel can call a SYCL_EXTERNAL function defined in a
 // different object file.
-// RUN: %clangxx -fsycl -DSOURCE1 -c %s -o %t1.o
-// RUN: %clangxx -fsycl -DSOURCE2 -c %s -o %t2.o
-// RUN: %clangxx -fsycl %t1.o %t2.o -o %t.exe
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple -DSOURCE1 -c %s -o %t1.o
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple -DSOURCE2 -c %s -o %t2.o
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %t1.o %t2.o -o %t.exe
 // RUN: %CPU_RUN_PLACEHOLDER %t.exe
 // RUN: %GPU_RUN_PLACEHOLDER %t.exe
 // RUN: %ACC_RUN_PLACEHOLDER %t.exe
@@ -11,11 +11,10 @@
 // static library.
 // RUN: rm -f %t.a
 // RUN: llvm-ar crv %t.a %t1.o
-// RUN: %clangxx -fsycl %t2.o -foffload-static-lib=%t.a -o %t.exe
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %t2.o -foffload-static-lib=%t.a -o %t.exe
 // RUN: %CPU_RUN_PLACEHOLDER %t.exe
 // RUN: %GPU_RUN_PLACEHOLDER %t.exe
 // RUN: %ACC_RUN_PLACEHOLDER %t.exe
-// XFAIL: cuda
 
 #include <CL/sycl.hpp>
 #include <iostream>

sycl/test/tools/abi_check_negative.dump

@@ -1,4 +1,4 @@
-# RUN: not env LLVM_BIN_PATH=%llvm_build_bin_dir python %sycl_tools_src_dir/abi_check.py --mode check_symbols --reference %s %llvm_build_lib_dir/libsycl.so | FileCheck %s
+# RUN: not env LLVM_BIN_PATH=%llvm_build_bin_dir python %sycl_tools_src_dir/abi_check.py --mode check_symbols --reference %s %sycl_libs_dir/libsycl.so | FileCheck %s
 # REQUIRES: linux
 
 # CHECK: The following symbols are missing from the new object file: