RuntimeError when using kernel.control() alongside other controlled operations

[bharat-thotakura]

Required prerequisites

• Consult the security policy. If reporting a security vulnerability, do not report the bug using this form. Use the process described in the policy to report the issue.
• Make sure you've read the documentation. Your issue may be addressed there.
• Search the issue tracker to verify that this hasn't already been reported. +1 or comment there if it has.
• If possible, make a PR with a failing test to give us a starting point to work on!

Describe the bug

Adding broader controlled operations (e.g., controlled on a 'sub kernel') can result in a RuntimeError if paired with gates such as cswap, ccz, or ccx.

Steps to reproduce the bug

This issue can be seen, for example, in a snippet like:

import cudaq

cudaq.set_target("nvidia")

kernel = cudaq.make_kernel()
qvector = kernel.qalloc(4)

kernel.h(qvector)

qubit_placeholders = [cudaq.qubit for _ in range(2)]
sub_kernel, *control_qubits = cudaq.make_kernel(*qubit_placeholders)
sub_kernel.cx(*control_qubits)
kernel.control(sub_kernel, qvector[0], qvector[3], qvector[2])

kernel.cswap(qvector[1], qvector[2], qvector[0])
kernel.mz(qvector)

cudaq.sample(kernel).dump()
# error: 'llvm.call' op operand type mismatch for operand 3: '!llvm.ptr<func<void (ptr<struct<"Array", opaque>>, ptr<struct<"Qubit", opaque>>, ptr<struct<"Qubit", opaque>>)>>' != '!llvm.ptr<func<void (ptr<struct<"Array", opaque>>, ptr<struct<"Qubit", opaque>>)>>'
# RuntimeError: cudaq::builder failed to JIT compile the Quake representation.

whereas, the following works:

import cudaq

cudaq.set_target("nvidia")

kernel = cudaq.make_kernel()
qvector = kernel.qalloc(4)

kernel.h(qvector)

qubit_placeholders = [cudaq.qubit for _ in range(2)]
sub_kernel, *control_qubits = cudaq.make_kernel(*qubit_placeholders)
sub_kernel.cx(*control_qubits)
kernel.control(sub_kernel, qvector[0], qvector[3], qvector[2])

kernel.mz(qvector)

cudaq.sample(kernel).dump()
# { 0000:60 0001:60 0010:64 0011:69 0100:57 0101:67 0110:64 0111:59 1000:66 1001:65 1010:71 1011:58 1100:50 1101:73 1110:61 1111:56 }

with print(cudaq.draw(kernel)) showing that the ccx was correctly interpreted:

Expected behavior

I would expect kernel to support all kernel.control(), cswap, ccx, ccz operations

Is this a regression? If it is, put the last known working version (or commit) here.

Not a regression

Environment

• CUDA-Q version: 0.9.1 (https://github.com/NVIDIA/cuda-quantum a3ab722)
• Python version: 3.11.11
• Operating system: Ubuntu