IR blocks for calling function pointers

Project Infrastructure LLVM Dev List Archives
1

Hello,

I am trying to create IR block for making a call to function pointer.

For creating the IR for a function call to “foo”, with “foo” being defined as “void foo(int)”, I can use the “getOrInsertFunction” call from Module class as follows:

std::vector<Type*> FooArgs;
FooArgs.push_back(IRB.getInt64Ty());
Value FooFunction = M.getOrInsertFunction(std::string(“foo”),
FunctionType::get(IRB.getVoidTy(), ArrayRef<Type>(FooArgs), false));
IRB.CreateCall(FooFunction, IRB.CreateLoad(LenAlloca));

I want to create a similar call, but to the pointer of foo (“foo_ptr” is defined as “(void) *foo_ptr (int)”). In C code, I would just call “foo_ptr(var)”, and it would just work. But, replacing the function name with the name of the function pointer, in the IRBlock does not work:

std::vector<Type*> FooArgs;
FooArgs.push_back(IRB.getInt64Ty());
Value FooFunctionPtr = M.getOrInsertFunction(std::string(“foo_ptr”),
FunctionType::get(IRB.getVoidTy(), ArrayRef<Type>(FooArgs), false));
IRB.CreateCall(FooFunctionPtr, IRB.CreateLoad(LenAlloca));

IRBlocks generated from the above crashes.
I see the LLVM blocks generated for a sample call to function pointer. It looks as follows:

%4 = load void (i64)** @foo_ptr, align 8, !dbg !26
%call = call i32 %4(i32 %3), !dbg !26

What LLVM function calls should be used to generate the above IR blocks?

Thanks,

Avi.

2

Hi,

  %4 = load void (i64)** @foo_ptr, align 8, !dbg !26
  %call = call i32 %4(i32 %3), !dbg !26

What LLVM function calls should be used to generate the above IR blocks?

I think the key thing you should notice in the "correct" code is that
@foo_ptr isn't actually a function. It's a generic global with type
void(i64)* (which means, as with all other LLVM globals that the
symbol @foo_ptr has type void(i64)** and you must load it to get the
actual value of the function pointer). If you look at the module where
it's defined you'll see something like

    @foo_ptr = global void(i64)* null

instead of a function definition. So you'd get a handle to it via a call like

    auto FooFunctionPtr = M.getOrInsertGlobal("foo_ptr",
PointerType::get(FunctionType::get(IRB.getVoidTy(), FooArgs, false)));

after that (and given that FooFunctionPtr is a Value with type
"void(i64)**") there's really only one sane set of operations you can
perform on it:

    auto FooFunction = IRB.getLoad(FooFunctionPtr)
    IRB.CreateCall(FooFunction, IRB.CreateLoad(LenAlloca));

Does that help?

Cheers.

Tim.

3

Yes. That is exactly what I was looking for. I tried it, and it works.

Thanks, Tim!