[WPD] Update llvm.public.type.test after importing functions

[tejohnson]

Summary

I noticed that we are converting llvm.public.type.test to regular
llvm.type.test too early, and thus not updating those in imported
functions. This would result in losing out on WPD opportunities. Move
the update to after function importing, and improve test to cover this
case.

[aeubanks]

llvm/test/ThinLTO/X86/public-type-test.ll

35

should this be checked twice for both public type tests?

[tejohnson]

llvm/test/ThinLTO/X86/public-type-test.ll

35

Yes, fixed.

[tejohnson]

Address comment

clang/test/CodeGenCXX/thinlto_public_type_test_distributed.ll

 ; REQUIRES: x86-registered-target
 
 ; Check that we properly update @llvm.public.type.test with distributed ThinLTO.
 
 ; RUN: opt -thinlto-bc -o %t.o %s
 
 ; RUN: llvm-lto2 run -thinlto-distributed-indexes %t.o \
 ; RUN:   -o %t2.index \
 ; RUN:   -r=%t.o,f,px
 
 ; RUN: %clang_cc1 -triple x86_64-grtev4-linux-gnu \
 ; RUN:   -emit-obj -fthinlto-index=%t.o.thinlto.bc \
 ; RUN:   -o %t.native.o -x ir %t.o --save-temps=obj
-; RUN: llvm-dis %t.native.o.0.preopt.bc -o - | FileCheck %s --check-prefix=PUBLIC
+; RUN: llvm-dis %t.native.o.3.import.bc -o - | FileCheck %s --check-prefix=PUBLIC
 
 ; RUN: llvm-lto2 run -thinlto-distributed-indexes %t.o --whole-program-visibility \
 ; RUN:   -o %t2.index \
 ; RUN:   -r=%t.o,f,px
 
 ; RUN: %clang_cc1 -triple x86_64-grtev4-linux-gnu \
 ; RUN:   -emit-obj -fthinlto-index=%t.o.thinlto.bc \
 ; RUN:   -o %t.native.o -x ir %t.o --save-temps=obj
-; RUN: llvm-dis %t.native.o.0.preopt.bc -o - | FileCheck %s --check-prefix=HIDDEN
+; RUN: llvm-dis %t.native.o.3.import.bc -o - | FileCheck %s --check-prefix=HIDDEN
 
 ; PUBLIC-NOT: call {{.*}}@llvm.public.type.test
 ; PUBLIC-NOT: call {{.*}}@llvm.type.test
 ; PUBLIC: call void @llvm.assume(i1 true)
 
 ; HIDDEN-NOT: call {{.*}}@llvm.public.type.test
 ; HIDDEN: call {{.*}}@llvm.type.test
 

lld/test/ELF/lto/update_public_type_test.ll

 ; REQUIRES: x86
 
 ; Check that we RAUW llvm.public.type.test with either llvm.type.test when --lto-whole-program-visibility is specified, or with true otherwise.
 
 ; RUN: opt --thinlto-bc -o %t.o %s
 ; RUN: ld.lld %t.o -o %t2.o --save-temps
-; RUN: llvm-dis %t.o.0.preopt.bc -o - | FileCheck %s --check-prefix=PUB
+; RUN: llvm-dis %t.o.3.import.bc -o - | FileCheck %s --check-prefix=PUB
 ; RUN: ld.lld %t.o -o %t3.o --save-temps --lto-whole-program-visibility
-; RUN: llvm-dis %t.o.0.preopt.bc -o - | FileCheck %s --check-prefix=WPV
+; RUN: llvm-dis %t.o.3.import.bc -o - | FileCheck %s --check-prefix=WPV
 
 target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
 target triple = "x86_64-grtev4-linux-gnu"
 
 declare i1 @llvm.public.type.test(ptr, metadata)
 declare void @llvm.assume(i1)
 
 ; PUB-NOT: call {{.*}}@llvm.public.type.test

llvm/lib/LTO/LTOBackend.cpp

Error lto::thinBackend(const Config &Conf, unsigned Task, AddStreamFn AddStream,
   if (!DiagFileOrErr)
     return DiagFileOrErr.takeError();
   auto DiagnosticOutputFile = std::move(*DiagFileOrErr);
 
   // Set the partial sample profile ratio in the profile summary module flag of
   // the module, if applicable.
   Mod.setPartialSampleProfileRatio(CombinedIndex);
 
-  updatePublicTypeTestCalls(Mod, CombinedIndex.withWholeProgramVisibility());
-
   if (Conf.CodeGenOnly) {
     codegen(Conf, TM.get(), AddStream, Task, Mod, CombinedIndex);
     return finalizeOptimizationRemarks(std::move(DiagnosticOutputFile));
   }
 
   if (Conf.PreOptModuleHook && !Conf.PreOptModuleHook(Task, Mod))
     return finalizeOptimizationRemarks(std::move(DiagnosticOutputFile));
 
Error lto::thinBackend(const Config &Conf, unsigned Task, AddStreamFn AddStream,
 
   FunctionImporter Importer(CombinedIndex, ModuleLoader,
                             ClearDSOLocalOnDeclarations);
   if (Error Err = Importer.importFunctions(Mod, ImportList).takeError())
     return Err;
 
   // Do this after any importing so that imported code is updated.
   updateMemProfAttributes(Mod, CombinedIndex);
+  updatePublicTypeTestCalls(Mod, CombinedIndex.withWholeProgramVisibility());
 
   if (Conf.PostImportModuleHook && !Conf.PostImportModuleHook(Task, Mod))
     return finalizeOptimizationRemarks(std::move(DiagnosticOutputFile));
 
   return OptimizeAndCodegen(Mod, TM.get(), std::move(DiagnosticOutputFile));
 }
 
 BitcodeModule *lto::findThinLTOModule(MutableArrayRef<BitcodeModule> BMs) {

llvm/lib/LTO/ThinLTOCodeGenerator.cpp

ProcessThinLTOModule(Module &TheModule, ModuleSummaryIndex &Index,
                      const FunctionImporter::ImportMapTy &ImportList,
                      const FunctionImporter::ExportSetTy &ExportList,
                      const DenseSet<GlobalValue::GUID> &GUIDPreservedSymbols,
                      const GVSummaryMapTy &DefinedGlobals,
                      const ThinLTOCodeGenerator::CachingOptions &CacheOptions,
                      bool DisableCodeGen, StringRef SaveTempsDir,
                      bool Freestanding, unsigned OptLevel, unsigned count,
                      bool DebugPassManager) {
-  // See comment at call to updateVCallVisibilityInIndex() for why
-  // WholeProgramVisibilityEnabledInLTO is false.
-  updatePublicTypeTestCalls(TheModule,
-                            /* WholeProgramVisibilityEnabledInLTO */ false);
-
   // "Benchmark"-like optimization: single-source case
   bool SingleModule = (ModuleMap.size() == 1);
 
   // When linking an ELF shared object, dso_local should be dropped. We
   // conservatively do this for -fpic.
   bool ClearDSOLocalOnDeclarations =
       TM.getTargetTriple().isOSBinFormatELF() &&
       TM.getRelocationModel() != Reloc::Static &&
ProcessThinLTOModule(Module &TheModule, ModuleSummaryIndex &Index,
   if (!ExportList.empty() || !GUIDPreservedSymbols.empty()) {
     // Apply summary-based internalization decisions.
     thinLTOInternalizeModule(TheModule, DefinedGlobals);
   }
 
   // Save internalized bitcode
   saveTempBitcode(TheModule, SaveTempsDir, count, ".2.internalized.bc");
 
-  if (!SingleModule) {
+  if (!SingleModule)
     crossImportIntoModule(TheModule, Index, ModuleMap, ImportList,
                           ClearDSOLocalOnDeclarations);
 
-    // Save temps: after cross-module import.
-    saveTempBitcode(TheModule, SaveTempsDir, count, ".3.imported.bc");
-  }
+  // Do this after any importing so that imported code is updated.
+  // See comment at call to updateVCallVisibilityInIndex() for why
+  // WholeProgramVisibilityEnabledInLTO is false.
+  updatePublicTypeTestCalls(TheModule,
+                            /* WholeProgramVisibilityEnabledInLTO */ false);
+
 
   optimizeModule(TheModule, TM, OptLevel, Freestanding, DebugPassManager,
                  &Index);
 
   saveTempBitcode(TheModule, SaveTempsDir, count, ".4.opt.bc");
 
   if (DisableCodeGen) {
     // Configured to stop before CodeGen, serialize the bitcode and return.

llvm/test/ThinLTO/X86/devirt_vcall_vis_public.ll

 ; RUN:   -r=%t2.o,test_public,px \
 ; RUN:   -r=%t2.o,_ZN1A1nEi,p \
 ; RUN:   -r=%t2.o,_ZN1B1fEi,p \
 ; RUN:   -r=%t2.o,_ZN1C1fEi,p \
 ; RUN:   -r=%t2.o,_ZN1D1mEi,p \
 ; RUN:   -r=%t2.o,_ZTV1B,px \
 ; RUN:   -r=%t2.o,_ZTV1C,px \
 ; RUN:   -r=%t2.o,_ZTV1D,px 2>&1 | FileCheck %s --check-prefix=REMARK
-; RUN: llvm-dis %t3.1.0.preopt.bc -o - | FileCheck %s --check-prefix=CHECK-TT
+; RUN: llvm-dis %t3.1.3.import.bc -o - | FileCheck %s --check-prefix=CHECK-TT
 ; RUN: llvm-dis %t3.1.4.opt.bc -o - | FileCheck %s --check-prefix=CHECK-IR
 
 ; Hybrid WPD
 ; Generate split module with summary for hybrid Thin/Regular LTO WPD.
 ; RUN: opt -thinlto-bc -thinlto-split-lto-unit -o %t.o %s
 ; RUN: llvm-lto2 run %t.o -save-temps -pass-remarks=. \
 ; RUN:   -whole-program-visibility \
 ; RUN:   -o %t3 \
 ; RUN:   -r=%t.o,test,px \
 ; RUN:   -r=%t.o,test_public,px \
 ; RUN:   -r=%t.o,_ZN1A1nEi,p \
 ; RUN:   -r=%t.o,_ZN1B1fEi,p \
 ; RUN:   -r=%t.o,_ZN1C1fEi,p \
 ; RUN:   -r=%t.o,_ZN1D1mEi,p \
 ; RUN:   -r=%t.o,_ZTV1B, \
 ; RUN:   -r=%t.o,_ZTV1C, \
 ; RUN:   -r=%t.o,_ZTV1D, \
 ; RUN:   -r=%t.o,_ZN1A1nEi, \
 ; RUN:   -r=%t.o,_ZN1B1fEi, \
 ; RUN:   -r=%t.o,_ZN1C1fEi, \
 ; RUN:   -r=%t.o,_ZN1D1mEi, \
 ; RUN:   -r=%t.o,_ZTV1B,px \
 ; RUN:   -r=%t.o,_ZTV1C,px \
 ; RUN:   -r=%t.o,_ZTV1D,px 2>&1 | FileCheck %s --check-prefix=REMARK --dump-input=fail
-; RUN: llvm-dis %t3.1.0.preopt.bc -o - | FileCheck %s --check-prefix=CHECK-TT
+; RUN: llvm-dis %t3.1.3.import.bc -o - | FileCheck %s --check-prefix=CHECK-TT
 ; RUN: llvm-dis %t3.1.4.opt.bc -o - | FileCheck %s --check-prefix=CHECK-IR
 
 ; Regular LTO WPD
 ; RUN: opt -o %t4.o %s
 ; RUN: llvm-lto2 run %t4.o -save-temps -pass-remarks=. \
 ; RUN:   -whole-program-visibility \
 ; RUN:   -o %t5 \
 ; RUN:   -r=%t4.o,test,px \

 ; RUN:   -r=%t2.o,test_public,px \
 ; RUN:   -r=%t2.o,_ZN1A1nEi,p \
 ; RUN:   -r=%t2.o,_ZN1B1fEi,p \
 ; RUN:   -r=%t2.o,_ZN1C1fEi,p \
 ; RUN:   -r=%t2.o,_ZN1D1mEi,p \
 ; RUN:   -r=%t2.o,_ZTV1B,px \
 ; RUN:   -r=%t2.o,_ZTV1C,px \
 ; RUN:   -r=%t2.o,_ZTV1D,px 2>&1 | FileCheck %s --implicit-check-not single-impl --allow-empty
-; RUN: llvm-dis %t3.1.0.preopt.bc -o - | FileCheck %s --check-prefix=CHECK-TT
+; RUN: llvm-dis %t3.1.3.import.bc -o - | FileCheck %s --check-prefix=CHECK-TT
 ; RUN: llvm-dis %t3.1.4.opt.bc -o - | FileCheck %s --check-prefix=CHECK-NODEVIRT-IR
 
 ; Hybrid WPD
 ; RUN: llvm-lto2 run %t.o -save-temps -pass-remarks=. \
 ; RUN:   -o %t3 \
 ; RUN:   -r=%t.o,test,px \
 ; RUN:   -r=%t.o,test_public,px \
 ; RUN:   -r=%t.o,_ZN1A1nEi,p \
 ; RUN:   -r=%t.o,_ZN1B1fEi,p \
 ; RUN:   -r=%t.o,_ZN1C1fEi,p \
 ; RUN:   -r=%t.o,_ZN1D1mEi,p \
 ; RUN:   -r=%t.o,_ZTV1B, \
 ; RUN:   -r=%t.o,_ZTV1C, \
 ; RUN:   -r=%t.o,_ZTV1D, \
 ; RUN:   -r=%t.o,_ZN1A1nEi, \
 ; RUN:   -r=%t.o,_ZN1B1fEi, \
 ; RUN:   -r=%t.o,_ZN1C1fEi, \
 ; RUN:   -r=%t.o,_ZN1D1mEi, \
 ; RUN:   -r=%t.o,_ZTV1B,px \
 ; RUN:   -r=%t.o,_ZTV1C,px \
 ; RUN:   -r=%t.o,_ZTV1D,px 2>&1 | FileCheck %s --implicit-check-not single-impl --allow-empty
-; RUN: llvm-dis %t3.1.0.preopt.bc -o - | FileCheck %s --check-prefix=CHECK-TT
+; RUN: llvm-dis %t3.1.3.import.bc -o - | FileCheck %s --check-prefix=CHECK-TT
 ; RUN: llvm-dis %t3.1.4.opt.bc -o - | FileCheck %s --check-prefix=CHECK-NODEVIRT-IR
 
 ; Regular LTO WPD
 ; RUN: llvm-lto2 run %t4.o -save-temps -pass-remarks=. \
 ; RUN:   -o %t5 \
 ; RUN:   -r=%t4.o,test,px \
 ; RUN:   -r=%t4.o,test_public,px \
 ; RUN:   -r=%t4.o,_ZN1A1nEi,p \

 ; RUN:   -r=%t2.o,test_public,px \
 ; RUN:   -r=%t2.o,_ZN1A1nEi,p \
 ; RUN:   -r=%t2.o,_ZN1B1fEi,p \
 ; RUN:   -r=%t2.o,_ZN1C1fEi,p \
 ; RUN:   -r=%t2.o,_ZN1D1mEi,p \
 ; RUN:   -r=%t2.o,_ZTV1B,px \
 ; RUN:   -r=%t2.o,_ZTV1C,px \
 ; RUN:   -r=%t2.o,_ZTV1D,px 2>&1 | FileCheck %s --implicit-check-not single-impl --allow-empty
-; RUN: llvm-dis %t3.1.0.preopt.bc -o - | FileCheck %s --check-prefix=CHECK-TT
+; RUN: llvm-dis %t3.1.3.import.bc -o - | FileCheck %s --check-prefix=CHECK-TT
 ; RUN: llvm-dis %t3.1.4.opt.bc -o - | FileCheck %s --check-prefix=CHECK-NODEVIRT-IR
 
 ; CHECK-TT-NOT: call {{.*}}@llvm.public.type.test
 
 target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
 target triple = "x86_64-grtev4-linux-gnu"
 
 %struct.A = type { ptr }

llvm/test/ThinLTO/X86/public-type-test.ll

-; Test to ensure that the legacy LTO API lowers @llvm.public.type.test.
+; Test to ensure that the LTO API (legacy and new) lowers @llvm.public.type.test.
 
-; RUN: opt -module-summary %s -o %t.bc
-; RUN: llvm-lto --thinlto-action=run -exported-symbol=_main %t.bc --thinlto-save-temps=%t2
-; RUN: llvm-dis -o - %t20.2.internalized.bc | FileCheck %s --check-prefix=PUBLIC
-; RUN: llvm-lto --thinlto-action=run -exported-symbol=_main %t.bc --thinlto-save-temps=%t2 --whole-program-visibility
-; RUN: llvm-dis -o - %t20.2.internalized.bc | FileCheck %s --check-prefix=HIDDEN
+; RUN: split-file %s %t
+
+; RUN: opt -module-summary %t/main.ll -o %t/main.bc
+; RUN: opt -module-summary %t/foo.ll -o %t/foo.bc
+; RUN: llvm-lto --thinlto-action=run -exported-symbol=_main %t/main.bc %t/foo.bc --thinlto-save-temps=%t2.
+; RUN: llvm-dis -o - %t2.0.3.imported.bc | FileCheck %s --check-prefix=PUBLIC
+; RUN: llvm-lto --thinlto-action=run -exported-symbol=_main %t/main.bc %t/foo.bc --thinlto-save-temps=%t2. --whole-program-visibility
+; RUN: llvm-dis -o - %t2.0.3.imported.bc | FileCheck %s --check-prefix=HIDDEN
+
+; RUN: llvm-lto2 run %t/main.bc %t/foo.bc -save-temps -pass-remarks=. \
+; RUN:   -whole-program-visibility \
+; RUN:   -o %t3 \
+; RUN:   -r=%t/main.bc,_main,px \
+; RUN:   -r=%t/main.bc,_bar,px \
+; RUN:   -r=%t/main.bc,_foo, \
+; RUN:   -r=%t/foo.bc,_foo,px
+; RUN: llvm-dis %t3.1.3.import.bc -o - | FileCheck %s --check-prefix=HIDDEN
+; RUN: llvm-lto2 run %t/main.bc %t/foo.bc -save-temps -pass-remarks=. \
+; RUN:   -o %t3 \
+; RUN:   -r=%t/main.bc,_main,px \
+; RUN:   -r=%t/main.bc,_bar,px \
+; RUN:   -r=%t/main.bc,_foo, \
+; RUN:   -r=%t/foo.bc,_foo,px
+; RUN: llvm-dis %t3.1.3.import.bc -o - | FileCheck %s --check-prefix=PUBLIC
 
 ; PUBLIC-NOT: call {{.*}}@llvm.public.type.test
 ; PUBLIC-NOT: call {{.*}}@llvm.type.test
+;; We should have converted the type tests from both main and the imported
+;; copy of foo to non-public.
+; HIDDEN-NOT: call {{.*}}@llvm.public.type.test
+; HIDDEN: call {{.*}}@llvm.type.test
 ; HIDDEN-NOT: call {{.*}}@llvm.public.type.test
 ; HIDDEN: call {{.*}}@llvm.type.test
 
+;--- main.ll
 target datalayout = "e-m:o-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
 target triple = "x86_64-apple-macosx10.9"
 
 define i32 @main(ptr %vtable) {
 entry:
   %p = call i1 @llvm.public.type.test(ptr %vtable, metadata !"_ZTS1A")
   call void @llvm.assume(i1 %p)
+  call void @bar(ptr %vtable)
   ret i32 0
 }
 
+define void @bar(ptr %vtable) {
+entry:
+  call void @foo(ptr %vtable)
+  ret void
+}
+
+declare void @foo(ptr %vtable)
+
+declare void @llvm.assume(i1)
+declare i1 @llvm.public.type.test(ptr, metadata)
+
+;--- foo.ll
+target datalayout = "e-m:o-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-apple-macosx10.9"
+
+define void @foo(ptr %vtable) {
+entry:
+  %p = call i1 @llvm.public.type.test(ptr %vtable, metadata !"_ZTS1A")
+  call void @llvm.assume(i1 %p)
+  ret void
+}
+
 declare void @llvm.assume(i1)
 declare i1 @llvm.public.type.test(ptr, metadata)