diff options
author | Johannes Doerfert <johannes@jdoerfert.de> | 2022-08-10 22:26:31 -0500 |
---|---|---|
committer | Johannes Doerfert <johannes@jdoerfert.de> | 2022-08-11 09:55:56 -0500 |
commit | a8cda3290944687b4fd0138e63cd980ea497a438 (patch) | |
tree | 182d18dac7104242e78bde030183c6aba48e577a /openmp | |
parent | 6e19e6ce36e44554ac9fbf8b2780de05e922c849 (diff) | |
download | llvm-a8cda3290944687b4fd0138e63cd980ea497a438.zip llvm-a8cda3290944687b4fd0138e63cd980ea497a438.tar.gz llvm-a8cda3290944687b4fd0138e63cd980ea497a438.tar.bz2 |
[OpenMP][FIX] Ensure __kmpc_kernel_parallel is reachable
The problem is we create the call to __kmpc_kernel_parallel in the
openmp-opt pass but while we optimize the code, the call is not there
yet. Thus, we assume we never reach it from __kmpc_target_deinit. That
allows us to remove the store in there (`ParallelRegionFn = nullptr`),
which leads to bad results later on.
This is a shortstop solution until we come up with something better.
Fixes https://github.com/llvm/llvm-project/issues/57064
Diffstat (limited to 'openmp')
-rw-r--r-- | openmp/libomptarget/DeviceRTL/src/Kernel.cpp | 16 |
1 files changed, 14 insertions, 2 deletions
diff --git a/openmp/libomptarget/DeviceRTL/src/Kernel.cpp b/openmp/libomptarget/DeviceRTL/src/Kernel.cpp index 74c22a6..d682652 100644 --- a/openmp/libomptarget/DeviceRTL/src/Kernel.cpp +++ b/openmp/libomptarget/DeviceRTL/src/Kernel.cpp @@ -35,7 +35,7 @@ static void genericStateMachine(IdentTy *Ident) { uint32_t TId = mapping::getThreadIdInBlock(); do { - ParallelRegionFnTy WorkFn = 0; + ParallelRegionFnTy WorkFn = nullptr; // Wait for the signal that we have a new work function. synchronize::threads(); @@ -100,8 +100,20 @@ int32_t __kmpc_target_init(IdentTy *Ident, int8_t Mode, // doing any work. mapping::getBlockSize() does not include any of the main // thread's warp, so none of its threads can ever be active worker threads. if (UseGenericStateMachine && - mapping::getThreadIdInBlock() < mapping::getBlockSize(IsSPMD)) + mapping::getThreadIdInBlock() < mapping::getBlockSize(IsSPMD)) { genericStateMachine(Ident); + } else { + // Retrieve the work function just to ensure we always call + // __kmpc_kernel_parallel even if a custom state machine is used. + // TODO: this is not super pretty. The problem is we create the call to + // __kmpc_kernel_parallel in the openmp-opt pass but while we optimize it is + // not there yet. Thus, we assume we never reach it from + // __kmpc_target_deinit. That allows us to remove the store in there to + // ParallelRegionFn, which leads to bad results later on. + ParallelRegionFnTy WorkFn = nullptr; + __kmpc_kernel_parallel(&WorkFn); + ASSERT(WorkFn == nullptr); + } return mapping::getThreadIdInBlock(); } |