[MLIR][ExecutionEngine] Enable PIC option #170995
Conversation
This PR enables the MLIR execution engine to dump object file as PIC code which is needed when the object file is bundled into a dynamic shared library.
tqchen
requested review from
ftynse,
makslevental,
rolfmorel and
stellaraccident
as code owners
|
Thank you for submitting a Pull Request (PR) to the LLVM Project! This PR will be automatically labeled and the relevant teams will be notified. If you wish to, you can add reviewers by using the "Reviewers" section on this page. If this is not working for you, it is probably because you do not have write permissions for the repository. In which case you can instead tag reviewers by name in a comment by using If you have received no comments on your PR for a week, you can request a review by "ping"ing the PR by adding a comment “Ping”. The common courtesy "ping" rate is once a week. Please remember that you are asking for valuable time from other developers. If you have further questions, they may be answered by the LLVM GitHub User Guide. You can also ask questions in a comment on this PR, on the LLVM Discord or on the forums. |
|
@llvm/pr-subscribers-mlir-execution-engine @llvm/pr-subscribers-mlir Author: Tianqi Chen (tqchen) ChangesThis PR enables the MLIR execution engine to dump object file as PIC code which is needed when the object file is bundled into a dynamic shared library. Full diff: https://github.com/llvm/llvm-project/pull/170995.diff 6 Files Affected:
diff --git a/mlir/include/mlir-c/ExecutionEngine.h b/mlir/include/mlir-c/ExecutionEngine.h
index 1a58d68533f24..348a1642ee6a0 100644
--- a/mlir/include/mlir-c/ExecutionEngine.h
+++ b/mlir/include/mlir-c/ExecutionEngine.h
@@ -44,7 +44,7 @@ DEFINE_C_API_STRUCT(MlirExecutionEngine, void);
/// TODO: figure out other options.
MLIR_CAPI_EXPORTED MlirExecutionEngine mlirExecutionEngineCreate(
MlirModule op, int optLevel, int numPaths,
- const MlirStringRef *sharedLibPaths, bool enableObjectDump);
+ const MlirStringRef *sharedLibPaths, bool enableObjectDump, bool enablePIC);
/// Initialize the ExecutionEngine. Global constructors specified by
/// `llvm.mlir.global_ctors` will be run. One common scenario is that kernel
diff --git a/mlir/lib/Bindings/Python/ExecutionEngineModule.cpp b/mlir/lib/Bindings/Python/ExecutionEngineModule.cpp
index 8bb493ed7240c..be0785b126eaa 100644
--- a/mlir/lib/Bindings/Python/ExecutionEngineModule.cpp
+++ b/mlir/lib/Bindings/Python/ExecutionEngineModule.cpp
@@ -75,13 +75,13 @@ NB_MODULE(_mlirExecutionEngine, m) {
"__init__",
[](PyExecutionEngine &self, MlirModule module, int optLevel,
const std::vector<std::string> &sharedLibPaths,
- bool enableObjectDump) {
+ bool enableObjectDump, bool enablePIC) {
llvm::SmallVector<MlirStringRef, 4> libPaths;
for (const std::string &path : sharedLibPaths)
libPaths.push_back({path.c_str(), path.length()});
- MlirExecutionEngine executionEngine =
- mlirExecutionEngineCreate(module, optLevel, libPaths.size(),
- libPaths.data(), enableObjectDump);
+ MlirExecutionEngine executionEngine = mlirExecutionEngineCreate(
+ module, optLevel, libPaths.size(), libPaths.data(),
+ enableObjectDump, enablePIC);
if (mlirExecutionEngineIsNull(executionEngine))
throw std::runtime_error(
"Failure while creating the ExecutionEngine.");
@@ -89,7 +89,7 @@ NB_MODULE(_mlirExecutionEngine, m) {
},
nb::arg("module"), nb::arg("opt_level") = 2,
nb::arg("shared_libs") = nb::list(),
- nb::arg("enable_object_dump") = true,
+ nb::arg("enable_object_dump") = true, nb::arg("enable_pic") = false,
"Create a new ExecutionEngine instance for the given Module. The "
"module must contain only dialects that can be translated to LLVM. "
"Perform transformations and code generation at the optimization "
diff --git a/mlir/lib/CAPI/ExecutionEngine/ExecutionEngine.cpp b/mlir/lib/CAPI/ExecutionEngine/ExecutionEngine.cpp
index 6c97499b28fd9..c3cb5e14f6160 100644
--- a/mlir/lib/CAPI/ExecutionEngine/ExecutionEngine.cpp
+++ b/mlir/lib/CAPI/ExecutionEngine/ExecutionEngine.cpp
@@ -22,7 +22,7 @@ using namespace mlir;
extern "C" MlirExecutionEngine
mlirExecutionEngineCreate(MlirModule op, int optLevel, int numPaths,
const MlirStringRef *sharedLibPaths,
- bool enableObjectDump) {
+ bool enableObjectDump, bool enablePIC) {
static bool initOnce = [] {
llvm::InitializeNativeTarget();
llvm::InitializeNativeTargetAsmParser(); // needed for inline_asm
@@ -43,6 +43,9 @@ mlirExecutionEngineCreate(MlirModule op, int optLevel, int numPaths,
consumeError(tmBuilderOrError.takeError());
return MlirExecutionEngine{nullptr};
}
+ if (enablePIC) {
+ tmBuilderOrError->setRelocationModel(llvm::Reloc::PIC_);
+ }
auto tmOrError = tmBuilderOrError->createTargetMachine();
if (!tmOrError) {
llvm::errs() << "Failed to create a TargetMachine for the host because: \n";
@@ -63,7 +66,8 @@ mlirExecutionEngineCreate(MlirModule op, int optLevel, int numPaths,
jitOptions.jitCodeGenOptLevel = static_cast<llvm::CodeGenOptLevel>(optLevel);
jitOptions.sharedLibPaths = libPaths;
jitOptions.enableObjectDump = enableObjectDump;
- auto jitOrError = ExecutionEngine::create(unwrap(op), jitOptions);
+ auto jitOrError = ExecutionEngine::create(unwrap(op), jitOptions,
+ std::move(tmOrError.get()));
if (!jitOrError) {
llvm::errs() << "Failed to create an ExecutionEngine because: \n";
consumeError(jitOrError.takeError());
diff --git a/mlir/test/CAPI/execution_engine.c b/mlir/test/CAPI/execution_engine.c
index 4751288c3ee4b..4df232f3caab3 100644
--- a/mlir/test/CAPI/execution_engine.c
+++ b/mlir/test/CAPI/execution_engine.c
@@ -69,7 +69,7 @@ void testSimpleExecution(void) {
mlirRegisterAllLLVMTranslations(ctx);
MlirExecutionEngine jit = mlirExecutionEngineCreate(
module, /*optLevel=*/2, /*numPaths=*/0, /*sharedLibPaths=*/NULL,
- /*enableObjectDump=*/false);
+ /*enableObjectDump=*/false, /*enablePIC=*/false);
if (mlirExecutionEngineIsNull(jit)) {
fprintf(stderr, "Execution engine creation failed");
exit(2);
@@ -125,7 +125,7 @@ void testOmpCreation(void) {
// against the OpenMP library.
MlirExecutionEngine jit = mlirExecutionEngineCreate(
module, /*optLevel=*/2, /*numPaths=*/0, /*sharedLibPaths=*/NULL,
- /*enableObjectDump=*/false);
+ /*enableObjectDump=*/false, /*enablePIC=*/false);
if (mlirExecutionEngineIsNull(jit)) {
fprintf(stderr, "Engine creation failed with OpenMP");
exit(2);
diff --git a/mlir/test/CAPI/global_constructors.c b/mlir/test/CAPI/global_constructors.c
index bd2fe1416f0df..9aacaf2c513f3 100644
--- a/mlir/test/CAPI/global_constructors.c
+++ b/mlir/test/CAPI/global_constructors.c
@@ -79,7 +79,7 @@ void testGlobalCtorJitCallback(void) {
// Create execution engine with initialization disabled
MlirExecutionEngine jit = mlirExecutionEngineCreate(
module, /*optLevel=*/2, /*numPaths=*/0, /*sharedLibPaths=*/NULL,
- /*enableObjectDump=*/false);
+ /*enableObjectDump=*/false, /*enablePIC=*/false);
if (mlirExecutionEngineIsNull(jit)) {
fprintf(stderr, "Execution engine creation failed");
diff --git a/mlir/test/python/execution_engine.py b/mlir/test/python/execution_engine.py
index 005813d1788f5..b11340f2c19ce 100644
--- a/mlir/test/python/execution_engine.py
+++ b/mlir/test/python/execution_engine.py
@@ -807,6 +807,7 @@ def testDumpToObjectFile():
# because RTDyldObjectLinkingLayer::emit will try to resolve symbols before dumping
# (see the jitLinkForORC call at the bottom there).
shared_libs=[MLIR_C_RUNNER_UTILS],
+ enable_pic=True,
)
# CHECK: Object file exists: True
|
just double checking my understanding: you mean when the object file is intended to be bundled into a shlib right? I.e. instead of into a static archive. |
makslevental
left a comment
makslevental
left a comment
There was a problem hiding this comment.
LGTM (just some qqs)
Co-authored-by: Mehdi Amini <[email protected]>
|
@tqchen let me know when you're ready - I'll merge. |
|
✅ With the latest revision this PR passed the C/C++ code formatter. |
tqchen
force-pushed
the
exec-engine-pic
branch
from
a252973 to
af0b61d
Compare
|
@tqchen Congratulations on having your first Pull Request (PR) merged into the LLVM Project! Your changes will be combined with recent changes from other authors, then tested by our build bots. If there is a problem with a build, you may receive a report in an email or a comment on this PR. Please check whether problems have been caused by your change specifically, as the builds can include changes from many authors. It is not uncommon for your change to be included in a build that fails due to someone else's changes, or infrastructure issues. How to do this, and the rest of the post-merge process, is covered in detail here. If your change does cause a problem, it may be reverted, or you can revert it yourself. This is a normal part of LLVM development. You can fix your changes and open a new PR to merge them again. If you don't get any reports, no action is required from you. Your changes are working as expected, well done! |
4 participants
This PR enables the MLIR execution engine to dump object file as PIC code which is needed when the object file is bundled into a dynamic shared library.