LLVM ir how to import other modules?

Question

I am learning LLVM IR by this LangRef.

As stated in this ref:

LLVM programs are composed of Modules, each of which is a translation unit of the input programs. Each module consists of functions, global variables, and symbol table entries.

This suggests that a program can be made up of multiple modules, but I didn't find a way to combine multiple modules.

The most likely way I found is by DICompileUnit or DIImportedEntry, but I don't know how to do it exactly.

Answer 1

You can combine LLVM modules similarly to how you combine .o objects: by linking them.

For example take these 2 c files:

a.c

int f(int);

int main(void) {
    return f(1);
}

b.c

int f(int x){
    return x*x;
}

Convert them to LLVM IR:

clang -c -emit-llvm a.c  # Produces a.bc
clang -c -emit-llvm b.c  # Produces b.bc

Disassemble them:

llvm-dis a.bc  # Produces a.ll
llvm-dis b.bc  # Produces b.ll

And now link them:

llvm-link a.bc b.c -o c.bc
llvm-dis c.bc  # Produces c.ll

The original IR files are:

a.ll

; ModuleID = 'a.bc'
source_filename = "a.c"
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-pc-linux-gnu"

; Function Attrs: noinline nounwind optnone sspstrong uwtable
define dso_local i32 @main() #0 {
  %1 = alloca i32, align 4
  store i32 0, i32* %1, align 4
  %2 = call i32 @f(i32 1)
  ret i32 %2
}

declare i32 @f(i32) #1

attributes #0 = { noinline nounwind optnone sspstrong uwtable "correctly-rounded-divide-sqrt-fp-math"="false" "disable-tail-calls"="false" "frame-pointer"="all" "less-precise-fpmad"="false" "min-legal-vector-width"="0" "no-infs-fp-math"="false" "no-jump-tables"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "target-cpu"="x86-64" "target-features"="+cx8,+fxsr,+mmx,+sse,+sse2,+x87" "unsafe-fp-math"="false" "use-soft-float"="false" }
attributes #1 = { "correctly-rounded-divide-sqrt-fp-math"="false" "disable-tail-calls"="false" "frame-pointer"="all" "less-precise-fpmad"="false" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "target-cpu"="x86-64" "target-features"="+cx8,+fxsr,+mmx,+sse,+sse2,+x87" "unsafe-fp-math"="false" "use-soft-float"="false" }

!llvm.module.flags = !{!0, !1, !2}
!llvm.ident = !{!3}

!0 = !{i32 1, !"wchar_size", i32 4}
!1 = !{i32 7, !"PIC Level", i32 2}
!2 = !{i32 7, !"PIE Level", i32 2}
!3 = !{!"clang version 10.0.0 "}

b.ll

; ModuleID = 'b.bc'
source_filename = "b.c"
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-pc-linux-gnu"

; Function Attrs: noinline nounwind optnone sspstrong uwtable
define dso_local i32 @f(i32 %0) #0 {
  %2 = alloca i32, align 4
  store i32 %0, i32* %2, align 4
  %3 = load i32, i32* %2, align 4
  %4 = load i32, i32* %2, align 4
  %5 = mul nsw i32 %3, %4
  ret i32 %5
}

attributes #0 = { noinline nounwind optnone sspstrong uwtable "correctly-rounded-divide-sqrt-fp-math"="false" "disable-tail-calls"="false" "frame-pointer"="all" "less-precise-fpmad"="false" "min-legal-vector-width"="0" "no-infs-fp-math"="false" "no-jump-tables"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "target-cpu"="x86-64" "target-features"="+cx8,+fxsr,+mmx,+sse,+sse2,+x87" "unsafe-fp-math"="false" "use-soft-float"="false" }

!llvm.module.flags = !{!0, !1, !2}
!llvm.ident = !{!3}

!0 = !{i32 1, !"wchar_size", i32 4}
!1 = !{i32 7, !"PIC Level", i32 2}
!2 = !{i32 7, !"PIE Level", i32 2}
!3 = !{!"clang version 10.0.0 "}

And the resulting linked file is:

c.ll

; ModuleID = 'c.bc'
source_filename = "llvm-link"
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-pc-linux-gnu"

; Function Attrs: noinline nounwind optnone sspstrong uwtable
define dso_local i32 @main() #0 {
  %1 = alloca i32, align 4
  store i32 0, i32* %1, align 4
  %2 = call i32 @f(i32 1)
  ret i32 %2
}

; Function Attrs: noinline nounwind optnone sspstrong uwtable
define dso_local i32 @f(i32 %0) #0 {
  %2 = alloca i32, align 4
  store i32 %0, i32* %2, align 4
  %3 = load i32, i32* %2, align 4
  %4 = load i32, i32* %2, align 4
  %5 = mul nsw i32 %3, %4
  ret i32 %5
}

attributes #0 = { noinline nounwind optnone sspstrong uwtable "correctly-rounded-divide-sqrt-fp-math"="false" "disable-tail-calls"="false" "frame-pointer"="all" "less-precise-fpmad"="false" "min-legal-vector-width"="0" "no-infs-fp-math"="false" "no-jump-tables"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "target-cpu"="x86-64" "target-features"="+cx8,+fxsr,+mmx,+sse,+sse2,+x87" "unsafe-fp-math"="false" "use-soft-float"="false" }

!llvm.ident = !{!0, !0}
!llvm.module.flags = !{!1, !2, !3}

!0 = !{!"clang version 10.0.0 "}
!1 = !{i32 1, !"wchar_size", i32 4}
!2 = !{i32 7, !"PIC Level", i32 2}
!3 = !{i32 7, !"PIE Level", i32 2}

You can now compile c.bc and run the resulting executable:

clang c.bc
./a.out  # Exit code 1 because 1*1=1

A similar alternative is to compile each bitcode file to an object file and then link those.