summaryrefslogtreecommitdiffstats
path: root/llvm/lib/ExecutionEngine/Orc/IRCompileLayer.cpp
diff options
context:
space:
mode:
authorLang Hames <lhames@gmail.com>2019-08-02 15:21:37 +0000
committerLang Hames <lhames@gmail.com>2019-08-02 15:21:37 +0000
commit809e9d1efa2f58b6333b1a0445e8574beedffc22 (patch)
tree2e9b52fc726ab240ba28823ba6d5f3ccacaac248 /llvm/lib/ExecutionEngine/Orc/IRCompileLayer.cpp
parent7eacefedab6fcdaab760ce28665fff57458a4e2f (diff)
downloadbcm5719-llvm-809e9d1efa2f58b6333b1a0445e8574beedffc22.tar.gz
bcm5719-llvm-809e9d1efa2f58b6333b1a0445e8574beedffc22.zip
[ORC] Change the locking scheme for ThreadSafeModule.
ThreadSafeModule/ThreadSafeContext are used to manage lifetimes and locking for LLVMContexts in ORCv2. Prior to this patch contexts were locked as soon as an associated Module was emitted (to be compiled and linked), and were not unlocked until the emit call returned. This could lead to deadlocks if interdependent modules that shared contexts were compiled on different threads: when, during emission of the first module, the dependence was discovered the second module (which would provide the required symbol) could not be emitted as the thread emitting the first module still held the lock. This patch eliminates this possibility by moving to a finer-grained locking scheme. Each client holds the module lock only while they are actively operating on it. To make this finer grained locking simpler/safer to implement this patch removes the explicit lock method, 'getContextLock', from ThreadSafeModule and replaces it with a new method, 'withModuleDo', that implicitly locks the context, calls a user-supplied function object to operate on the Module, then implicitly unlocks the context before returning the result. ThreadSafeModule TSM = getModule(...); size_t NumFunctions = TSM.withModuleDo( [](Module &M) { // <- context locked before entry to lambda. return M.size(); }); Existing ORCv2 layers that operate on ThreadSafeModules are updated to use the new method. This method is used to introduce Module locking into each of the existing layers. llvm-svn: 367686
Diffstat (limited to 'llvm/lib/ExecutionEngine/Orc/IRCompileLayer.cpp')
-rw-r--r--llvm/lib/ExecutionEngine/Orc/IRCompileLayer.cpp4
1 files changed, 2 insertions, 2 deletions
diff --git a/llvm/lib/ExecutionEngine/Orc/IRCompileLayer.cpp b/llvm/lib/ExecutionEngine/Orc/IRCompileLayer.cpp
index 81dfc02f55b..d311f34179c 100644
--- a/llvm/lib/ExecutionEngine/Orc/IRCompileLayer.cpp
+++ b/llvm/lib/ExecutionEngine/Orc/IRCompileLayer.cpp
@@ -22,9 +22,9 @@ void IRCompileLayer::setNotifyCompiled(NotifyCompiledFunction NotifyCompiled) {
void IRCompileLayer::emit(MaterializationResponsibility R,
ThreadSafeModule TSM) {
- assert(TSM.getModule() && "Module must not be null");
+ assert(TSM && "Module must not be null");
- if (auto Obj = Compile(*TSM.getModule())) {
+ if (auto Obj = TSM.withModuleDo(Compile)) {
{
std::lock_guard<std::mutex> Lock(IRLayerMutex);
if (NotifyCompiled)
OpenPOWER on IntegriCloud