LLVM asserts "Resolving symbol with incorrect flags" - c++

Question

I'm following the example of Kaleidoscope to write a minimum IR file interpreter. It takes one command line argument, which is a path to .ll file, and executes the main function in the file. But when I tested it on an IR file, it failed with:
Assertion failed: (KV.second.getFlags() & ~WeakFlags) == (I->second & ~WeakFlags) && "Resolving symbol with incorrect flags", file <path>\llvm\lib\ExecutionEngine\Orc\Core.cpp, line 2775
Considering the simplicity of my code (which only has 63 lines), I can't figure out what's wrong in it. Please help 😭!!!
Full Source Code
#include <iostream>
#include <llvm/IR/DataLayout.h>
#include <llvm/IR/LLVMContext.h>
#include <llvm/IR/Module.h>
#include <llvm/IRReader/IRReader.h>
#include <llvm/Support/SourceMgr.h>
#include <llvm/Support/TargetSelect.h>
#include <llvm/ExecutionEngine/Orc/CompileUtils.h>
#include <llvm/ExecutionEngine/Orc/Core.h>
#include <llvm/ExecutionEngine/Orc/ExecutionUtils.h>
#include <llvm/ExecutionEngine/Orc/ExecutorProcessControl.h>
#include <llvm/ExecutionEngine/Orc/IRCompileLayer.h>
#include <llvm/ExecutionEngine/Orc/JITTargetMachineBuilder.h>
#include <llvm/ExecutionEngine/Orc/RTDyldObjectLinkingLayer.h>
#include <llvm/ExecutionEngine/SectionMemoryManager.h>
using namespace llvm;
int
main(int argc, char* argv[])
{
InitializeNativeTarget();
InitializeNativeTargetAsmPrinter();
InitializeNativeTargetAsmParser();
orc::ThreadSafeContext tsctx(std::make_unique<LLVMContext>());
SMDiagnostic error;
auto mod = parseIRFile(argv[1], error, *tsctx.getContext());
auto epc = orc::SelfExecutorProcessControl::Create();
cantFail(epc.takeError());
orc::ExecutionSession es(std::move(*epc));
auto triple = es.getExecutorProcessControl().getTargetTriple();
orc::JITTargetMachineBuilder jtmb(triple);
auto dl = jtmb.getDefaultDataLayoutForTarget();
cantFail(dl.takeError());
orc::RTDyldObjectLinkingLayer ol(
es, []() { return std::make_unique<SectionMemoryManager>(); });
orc::IRCompileLayer cl(
es, ol, std::make_unique<orc::ConcurrentIRCompiler>(std::move(jtmb)));
auto& jd = es.createBareJITDylib("jd");
jd.addGenerator(
cantFail(orc::DynamicLibrarySearchGenerator::GetForCurrentProcess(
dl->getGlobalPrefix())));
cantFail(cl.add(jd, orc::ThreadSafeModule(std::move(mod), tsctx)));
orc::MangleAndInterner mangle(es, *dl);
auto f = es.lookup({ &jd }, mangle("main"));
cantFail(f.takeError());
return reinterpret_cast<int (*)()>(f->getAddress())();
}
Test .ll file
; ModuleID = 'sum.ll'
source_filename = "sum.c"
target datalayout = "e-m:w-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
target triple = "x86_64-pc-windows-msvc19.29.30133"
%struct._iobuf = type { i8* }
%struct.__crt_locale_pointers = type { %struct.__crt_locale_data*, %struct.__crt_multibyte_data* }
%struct.__crt_locale_data = type opaque
%struct.__crt_multibyte_data = type opaque
$scanf = comdat any
$__local_stdio_scanf_options = comdat any
$"??_C#_02DPKJAMEF#?$CFd?$AA#" = comdat any
#"??_C#_02DPKJAMEF#?$CFd?$AA#" = linkonce_odr dso_local unnamed_addr constant [3 x i8] c"%d\00", comdat, align 1
#__local_stdio_scanf_options._OptionsStorage = internal global i64 0, align 8
; Function Attrs: nounwind uwtable
define dso_local i32 #main() local_unnamed_addr #0 {
%1 = alloca i32, align 4
%2 = bitcast i32* %1 to i8*
call void #llvm.lifetime.start.p0i8(i64 4, i8* nonnull %2) #6
%3 = call i32 (i8*, ...) #scanf(i8* getelementptr inbounds ([3 x i8], [3 x i8]* #"??_C#_02DPKJAMEF#?$CFd?$AA#", i64 0, i64 0), i32* nonnull %1)
%4 = load i32, i32* %1, align 4, !tbaa !4
%5 = call i32 (i8*, ...) #scanf(i8* getelementptr inbounds ([3 x i8], [3 x i8]* #"??_C#_02DPKJAMEF#?$CFd?$AA#", i64 0, i64 0), i32* nonnull %1)
%6 = load i32, i32* %1, align 4, !tbaa !4
%7 = add nsw i32 %6, %4
%8 = call i32 (i8*, ...) #scanf(i8* getelementptr inbounds ([3 x i8], [3 x i8]* #"??_C#_02DPKJAMEF#?$CFd?$AA#", i64 0, i64 0), i32* nonnull %1)
%9 = load i32, i32* %1, align 4, !tbaa !4
%10 = add nsw i32 %7, %9
%11 = call i32 (i8*, ...) #scanf(i8* getelementptr inbounds ([3 x i8], [3 x i8]* #"??_C#_02DPKJAMEF#?$CFd?$AA#", i64 0, i64 0), i32* nonnull %1)
%12 = load i32, i32* %1, align 4, !tbaa !4
%13 = add nsw i32 %10, %12
%14 = call i32 (i8*, ...) #scanf(i8* getelementptr inbounds ([3 x i8], [3 x i8]* #"??_C#_02DPKJAMEF#?$CFd?$AA#", i64 0, i64 0), i32* nonnull %1)
%15 = load i32, i32* %1, align 4, !tbaa !4
%16 = add nsw i32 %13, %15
%17 = call i32 (i8*, ...) #scanf(i8* getelementptr inbounds ([3 x i8], [3 x i8]* #"??_C#_02DPKJAMEF#?$CFd?$AA#", i64 0, i64 0), i32* nonnull %1)
%18 = load i32, i32* %1, align 4, !tbaa !4
%19 = add nsw i32 %16, %18
%20 = call i32 (i8*, ...) #scanf(i8* getelementptr inbounds ([3 x i8], [3 x i8]* #"??_C#_02DPKJAMEF#?$CFd?$AA#", i64 0, i64 0), i32* nonnull %1)
%21 = load i32, i32* %1, align 4, !tbaa !4
%22 = add nsw i32 %19, %21
%23 = call i32 (i8*, ...) #scanf(i8* getelementptr inbounds ([3 x i8], [3 x i8]* #"??_C#_02DPKJAMEF#?$CFd?$AA#", i64 0, i64 0), i32* nonnull %1)
%24 = load i32, i32* %1, align 4, !tbaa !4
%25 = add nsw i32 %22, %24
%26 = call i32 (i8*, ...) #scanf(i8* getelementptr inbounds ([3 x i8], [3 x i8]* #"??_C#_02DPKJAMEF#?$CFd?$AA#", i64 0, i64 0), i32* nonnull %1)
%27 = load i32, i32* %1, align 4, !tbaa !4
%28 = add nsw i32 %25, %27
%29 = call i32 (i8*, ...) #scanf(i8* getelementptr inbounds ([3 x i8], [3 x i8]* #"??_C#_02DPKJAMEF#?$CFd?$AA#", i64 0, i64 0), i32* nonnull %1)
%30 = load i32, i32* %1, align 4, !tbaa !4
%31 = add nsw i32 %28, %30
call void #llvm.lifetime.end.p0i8(i64 4, i8* nonnull %2) #6
ret i32 %31
}
; Function Attrs: argmemonly mustprogress nofree nosync nounwind willreturn
declare void #llvm.lifetime.start.p0i8(i64 immarg, i8* nocapture) #1
; Function Attrs: inlinehint nobuiltin nounwind uwtable
define linkonce_odr dso_local i32 #scanf(i8* %0, ...) local_unnamed_addr #2 comdat {
%2 = alloca i8*, align 8
%3 = bitcast i8** %2 to i8*
call void #llvm.lifetime.start.p0i8(i64 8, i8* nonnull %3) #6
call void #llvm.va_start(i8* nonnull %3)
%4 = load i8*, i8** %2, align 8, !tbaa !8
%5 = call %struct._iobuf* #__acrt_iob_func(i32 0) #6
%6 = call i64* #__local_stdio_scanf_options() #6
%7 = load i64, i64* %6, align 8, !tbaa !10
%8 = call i32 #__stdio_common_vfscanf(i64 %7, %struct._iobuf* %5, i8* %0, %struct.__crt_locale_pointers* null, i8* %4) #6
call void #llvm.va_end(i8* nonnull %3)
call void #llvm.lifetime.end.p0i8(i64 8, i8* nonnull %3) #6
ret i32 %8
}
; Function Attrs: argmemonly mustprogress nofree nosync nounwind willreturn
declare void #llvm.lifetime.end.p0i8(i64 immarg, i8* nocapture) #1
; Function Attrs: mustprogress nofree nosync nounwind willreturn
declare void #llvm.va_start(i8*) #3
; Function Attrs: mustprogress nofree nosync nounwind willreturn
declare void #llvm.va_end(i8*) #3
declare dso_local %struct._iobuf* #__acrt_iob_func(i32) local_unnamed_addr #4
declare dso_local i32 #__stdio_common_vfscanf(i64, %struct._iobuf*, i8*, %struct.__crt_locale_pointers*, i8*) local_unnamed_addr #4
; Function Attrs: noinline nounwind uwtable
define linkonce_odr dso_local i64* #__local_stdio_scanf_options() local_unnamed_addr #5 comdat {
ret i64* #__local_stdio_scanf_options._OptionsStorage
}
attributes #0 = { nounwind uwtable "frame-pointer"="none" "min-legal-vector-width"="0" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-cpu"="x86-64" "target-features"="+cx8,+fxsr,+mmx,+sse,+sse2,+x87" "tune-cpu"="generic" }
attributes #1 = { argmemonly mustprogress nofree nosync nounwind willreturn }
attributes #2 = { inlinehint nobuiltin nounwind uwtable "frame-pointer"="none" "min-legal-vector-width"="0" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-cpu"="x86-64" "target-features"="+cx8,+fxsr,+mmx,+sse,+sse2,+x87" "tune-cpu"="generic" }
attributes #3 = { mustprogress nofree nosync nounwind willreturn }
attributes #4 = { "frame-pointer"="none" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-cpu"="x86-64" "target-features"="+cx8,+fxsr,+mmx,+sse,+sse2,+x87" "tune-cpu"="generic" }
attributes #5 = { noinline nounwind uwtable "frame-pointer"="none" "min-legal-vector-width"="0" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-cpu"="x86-64" "target-features"="+cx8,+fxsr,+mmx,+sse,+sse2,+x87" "tune-cpu"="generic" }
attributes #6 = { nounwind }
!llvm.module.flags = !{!0, !1, !2}
!llvm.ident = !{!3}
!0 = !{i32 1, !"wchar_size", i32 2}
!1 = !{i32 7, !"PIC Level", i32 2}
!2 = !{i32 7, !"uwtable", i32 1}
!3 = !{!"clang version 13.0.1"}
!4 = !{!5, !5, i64 0}
!5 = !{!"int", !6, i64 0}
!6 = !{!"omnipotent char", !7, i64 0}
!7 = !{!"Simple C/C++ TBAA"}
!8 = !{!9, !9, i64 0}
!9 = !{!"any pointer", !6, i64 0}
!10 = !{!11, !11, i64 0}
!11 = !{!"long long", !6, i64 0}
Which is compiled from:
#include <stdio.h>
int
main()
{
int n, sum = 0;
for (int i = 0; i < 10; ++i) {
scanf("%d", &n);
sum += n;
}
return sum;
}

Answer 1

Well, it turns out that THERE IS NOTHING WRONG WITH MY CODE!
I compile and test the same code in Linux environment (WSL2), and everything works fine. I'm pretty sure that this is somewhat compatibility problem between Linux and Windows.
Maybe this is a bug of LLVM?