diff options
author | Chandler Carruth <chandlerc@gmail.com> | 2017-02-12 05:34:04 +0000 |
---|---|---|
committer | Chandler Carruth <chandlerc@gmail.com> | 2017-02-12 05:34:04 +0000 |
commit | 9c2410924c0f9bc09c9f79a342aaa531d9be6eb4 (patch) | |
tree | 41bc217965a43162d30191ab067e43280dddb663 /lib/Passes | |
parent | 820dc664db7fe84669b76c50962b278e5ed19992 (diff) |
[PM] Enable GlobalsAA in the new PM's pipeline by default.
All the invalidation issues and bugs in this seem to be fixed, it has
survived a full build of the test suite plus SPEC with asserts and ASan
enabled on the Clang binary used.
Differential Revision: https://reviews.llvm.org/D29815
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@294887 91177308-0d34-0410-b5e6-96231b3b80d8
Diffstat (limited to 'lib/Passes')
-rw-r--r-- | lib/Passes/PassBuilder.cpp | 20 |
1 files changed, 6 insertions, 14 deletions
diff --git a/lib/Passes/PassBuilder.cpp b/lib/Passes/PassBuilder.cpp index 06190682f32c..9264cf11e504 100644 --- a/lib/Passes/PassBuilder.cpp +++ b/lib/Passes/PassBuilder.cpp @@ -431,10 +431,9 @@ PassBuilder::buildPerModuleDefaultPipeline(OptimizationLevel Level, GlobalCleanupPM.addPass(SimplifyCFGPass()); MPM.addPass(createModuleToFunctionPassAdaptor(std::move(GlobalCleanupPM))); - // FIXME: Enable this when cross-IR-unit analysis invalidation is working. -#if 0 - MPM.addPass(RequireAnalysisPass<GlobalsAA>()); -#endif + // Require the GlobalsAA analysis for the module so we can query it within + // the CGSCC pipeline. + MPM.addPass(RequireAnalysisPass<GlobalsAA, Module>()); // Now begin the main postorder CGSCC pipeline. // FIXME: The current CGSCC pipeline has its origins in the legacy pass @@ -482,17 +481,14 @@ PassBuilder::buildPerModuleDefaultPipeline(OptimizationLevel Level, // FIXME: Is this really an optimization rather than a canonicalization? MPM.addPass(ReversePostOrderFunctionAttrsPass()); - // Recompute GloblasAA here prior to function passes. This is particularly + // Re-require GloblasAA here prior to function passes. This is particularly // useful as the above will have inlined, DCE'ed, and function-attr // propagated everything. We should at this point have a reasonably minimal // and richly annotated call graph. By computing aliasing and mod/ref // information for all local globals here, the late loop passes and notably // the vectorizer will be able to use them to help recognize vectorizable // memory operations. - // FIXME: Enable this once analysis invalidation is fully supported. -#if 0 - MPM.addPass(Require<GlobalsAA>()); -#endif + MPM.addPass(RequireAnalysisPass<GlobalsAA, Module>()); FunctionPassManager OptimizePM(DebugLogging); OptimizePM.addPass(Float2IntPass()); @@ -766,12 +762,8 @@ AAManager PassBuilder::buildDefaultAAPipeline() { // Add support for querying global aliasing information when available. // Because the `AAManager` is a function analysis and `GlobalsAA` is a module // analysis, all that the `AAManager` can do is query for any *cached* - // results from `GlobalsAA` through a readonly proxy.. -#if 0 - // FIXME: Enable once the invalidation logic supports this. Currently, the - // `AAManager` will hold stale references to the module analyses. + // results from `GlobalsAA` through a readonly proxy. AA.registerModuleAnalysis<GlobalsAA>(); -#endif return AA; } |