kernel.control() raises error for multi-qubit control argument

[fedonman]

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

I am trying to create controlled versions of X gate. For the CX (1 control qubit), the code works correctly, but for CCX (2 control qubits) or more, I get an error.

Steps to reproduce the bug

The following works correcty:

nqubits = 3

kernel = cudaq.make_kernel()
qubits = kernel.qalloc(nqubits)

target_kernel, qubit = cudaq.make_kernel(cudaq.qubit)
target_kernel.x(qubit)

kernel.control(target_kernel, qubits[0], qubits[2])

When I try to add more control qubits, I get an error, no matter the syntax:

nqubits = 3

kernel = cudaq.make_kernel()
qubits = kernel.qalloc(nqubits)

target_kernel, qubit = cudaq.make_kernel(cudaq.qubit)
target_kernel.x(qubit)

kernel.control(target_kernel, [qubits[0], qubits[1]], qubits[2])  # This raises an error
kernel.control(target_kernel, qubits[0], qubits[1], qubits[2])  # This raises an error as well

In the first case the error is:

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
Cell In[7], line 10
      7 target_kernel.x(qubit)
      9 kernel.h(qubits[0])
---> 10 kernel.control(target_kernel, [qubits[0], qubits[1]], qubits[nqubits - 1])  # This raises an error
     11 # kernel.control(target_kernel, qubits[0], qubits[1], qubits[nqubits - 1])  # This raises an error as well

File ~/projects/qilisdk/.venv/lib/python3.10/site-packages/cudaq/kernel/kernel_builder.py:1239, in PyKernel.control(self, target, control, *target_arguments)
   1205 def control(self, target, control, *target_arguments):
   1206     """
   1207     Apply the `target` kernel as a controlled operation in-place to 
   1208     `self`.Uses the provided `control` as control qubit/s for the operation.
   (...)
   1237     ```
   1238     """
-> 1239     self.__applyControlOrAdjoint(target, False, [control.mlirValue],
   1240                                  *target_arguments)
   1241     return

AttributeError: 'list' object has no attribute 'mlirValue'

In the second case the error is:

---------------------------------------------------------------------------
IndexError                                Traceback (most recent call last)
Cell In[6], line 11
      9 kernel.h(qubits[0])
     10 # kernel.control(target_kernel, [qubits[0], qubits[1]], qubits[nqubits - 1])  # This raises an error
---> 11 kernel.control(target_kernel, qubits[0], qubits[1], qubits[nqubits - 1])  # This raises an error as well

File ~/projects/qilisdk/.venv/lib/python3.10/site-packages/cudaq/kernel/kernel_builder.py:1239, in PyKernel.control(self, target, control, *target_arguments)
   1205 def control(self, target, control, *target_arguments):
   1206     """
   1207     Apply the `target` kernel as a controlled operation in-place to 
   1208     `self`.Uses the provided `control` as control qubit/s for the operation.
   (...)
   1237     ```
   1238     """
-> 1239     self.__applyControlOrAdjoint(target, False, [control.mlirValue],
   1240                                  *target_arguments)
   1241     return

File ~/projects/qilisdk/.venv/lib/python3.10/site-packages/cudaq/kernel/kernel_builder.py:666, in PyKernel.__applyControlOrAdjoint(self, target, isAdjoint, controls, *args)
    664 mlirValues = []
    665 for i, v in enumerate(args):
--> 666     argTy = otherFTy[i].type
    667     if not isinstance(v, QuakeValue):
    668         # here we have to map constant Python data
    669         # to an MLIR Value
    670         value = self.__getMLIRValueFromPythonArg(v, argTy)

IndexError: index out of range

Expected behavior

I would expect the correct CCX gate to be created as is the case with the decorator syntax:

@cudaq.kernel
def kernel():
    qubits = cudaq.qvector(3)
    x.ctrl([qubits[0], qubits[1]], qubits[2])

print(cudaq.draw(kernel))

# Prints:
q0 : ──●──
       │  
q1 : ──●──
     ╭─┴─╮
q2 : ┤ x ├
     ╰───╯

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

I haven't tested it with older versions.

Environment

• CUDA-Q version: 0.9.1
• Python version: 3.10+
• C++ compiler: Irrelevant
• Operating system: Ubuntu 24.04

Suggestions

The error probably lies within how kernel_builder.__applyControlOrAdjoint() handles the QuakeValue of control argument.