feat: Integrate databus in the private kernels

[sirasistant]

Integrates the databus in the private kernels. We do this by annotating with call_data(0) for previous kernel public inputs and call_data(1) for app public inputs. Kernels and apps expose their own public inputs as return_data except the tail kernel who does it via the traditional public inputs mechanism for the tube.

[github-actions]

Changes to circuit sizes

Generated at commit: aca9fec1e1e1fdb10876b3e0deeb05fcd004b654, compared to commit: 97e4b9b2e0d37575b6b5e4c7a22f85b60d1f418b

🧾 Summary (100% most significant diffs)

Program	ACIR opcodes (+/-)	%	Circuit size (+/-)	%
private_kernel_init	-8 ✅	-0.03%	+1,628 ❌	+5.39%
private_kernel_reset_4_4_4_4_4_4_4_4_1	-4 ✅	-0.01%	+843 ❌	+1.15%
private_kernel_reset	-4 ✅	-0.00%	-2,422 ✅	-0.52%
private_kernel_inner	-19 ✅	-0.04%	-2,982 ✅	-5.43%
private_kernel_tail_to_public	-5 ✅	-0.02%	-2,332 ✅	-5.58%
private_kernel_tail	-3 ✅	-0.06%	-1,904 ✅	-17.42%

---

Full diff report 👇

Program	ACIR opcodes (+/-)	%	Circuit size (+/-)	%
private_kernel_init	24,893 (-8)	-0.03%	31,835 (+1,628)	+5.39%
private_kernel_reset_4_4_4_4_4_4_4_4_1	34,895 (-4)	-0.01%	74,459 (+843)	+1.15%
private_kernel_reset	91,931 (-4)	-0.00%	467,711 (-2,422)	-0.52%
private_kernel_inner	43,932 (-19)	-0.04%	51,905 (-2,982)	-5.43%
private_kernel_tail_to_public	29,826 (-5)	-0.02%	39,492 (-2,332)	-5.58%
private_kernel_tail	4,742 (-3)	-0.06%	9,023 (-1,904)	-17.42%

[LeilaWang]

Nice!! 🎉