mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-11-26 04:32:44 +01:00
979e07bdb3
FAILED: unittests/Target/AMDGPU/AMDGPUTests ... ld.lld: error: undefined symbol: llvm::MCRegisterInfo::getLLVMRegNum(unsigned int, bool) const >>> referenced by DwarfRegMappings.cpp:60 (/usr/local/google/home/maskray/llvm/llvm/unittests/Target/AMDGPU/DwarfRegMappings.cpp:60) >>> unittests/Target/AMDGPU/CMakeFiles/AMDGPUTests.dir/DwarfRegMappings.cpp.o:(AMDGPUDwarfRegMappingTests_TestWave64DwarfRegMapping_Test::TestBody()) >>> referenced by DwarfRegMappings.cpp:82 (/usr/local/google/home/maskray/llvm/llvm/unittests/Target/AMDGPU/DwarfRegMappings.cpp:82) >>> unittests/Target/AMDGPU/CMakeFiles/AMDGPUTests.dir/DwarfRegMappings.cpp.o:(AMDGPUDwarfRegMappingTests_TestWave32DwarfRegMapping_Test::TestBody()) A -DBUILD_SHARED_LIBS=off build is good because AMDGPUCodeGen pulls in MC. A -DBUILD_SHARED_LIBS=on build requires all direct dependencies (MC) to be listed becuase llvm/cmake/modules/HandleLLVMOptions.cmake uses -Wl,-z,defs
12 lines
149 B
CMake
12 lines
149 B
CMake
set(LLVM_LINK_COMPONENTS
|
|
AMDGPUCodeGen
|
|
AMDGPUDesc
|
|
AMDGPUInfo
|
|
MC
|
|
Support
|
|
)
|
|
|
|
add_llvm_unittest(AMDGPUDwarfTests
|
|
DwarfRegMappings.cpp
|
|
)
|