[REFRACTOR] Transform OpenOS into Modular Research Kernel Architecture (OpenOS-RK)

[orgito1015]

Refactor: Transform OpenOS into Modular Research-Oriented Kernel Architecture (OpenOS-RK)

Summary

This pull request performs a full architectural refactor of OpenOS,
transitioning it from an educational monolithic kernel prototype into a
structured, research-oriented modular kernel architecture (OpenOS-RK).

The purpose of this refactor is not feature expansion but architectural
formalization. The changes introduce strict layering, a formal memory
stack, kernel object abstraction, scheduler skeleton integration,
slab-based allocation, and research instrumentation hooks. The kernel
remains monolithic, but now follows clear dependency boundaries and
modular separation principles.

This refactor establishes a clean foundation for future experimentation
in scheduling, memory management, isolation models, tracing, and
kernel-level research.

---

Motivation

The previous architecture of OpenOS had structural limitations including
mixed layers, lack of object abstraction, informal memory
initialization, and no instrumentation capability. This refactor
resolves these issues to establish a research-grade kernel foundation.

---

Architectural Goals

1. Maintain monolithic kernel design.
2. Enforce strict modular separation.
3. Formalize memory layering.
4. Introduce kernel object abstraction.
5. Prepare scheduler abstraction layer.
6. Add research instrumentation capability.
7. Eliminate circular dependencies.
8. Centralize memory layout management.
9. Make the kernel experiment-ready.

---

New Directory Structure

/arch
    /x86
        cpu.c
        gdt.c
        idt.c
        isr.S
        pic.c
        cpuid.c

/core
    kernel.c
    panic.c
    logger.c
    trace.c
    config.h

/memory
    pmm.c
    vmm.c
    heap.c
    slab.c
    mem_layout.c

/process
    scheduler.c
    task.c
    context_switch.S

/object
    object.c
    object.h

/fs
    vfs.c
    ramfs.c
    file.c

/drivers
    keyboard.c
    timer.c
    console.c
    driver_manager.c

/ipc
    pipe.c
    message.c

/research
    instrumentation.c
    profiler.c

---

Dependency Rules

arch → memory → core → object → process → fs → drivers

• arch must not depend on higher layers.
• memory must not depend on drivers or process.
• core must not manipulate hardware directly.
• No circular includes are allowed.
• Architecture-specific code is isolated under /arch/x86.

---

Memory Stack Formalization

pmm_init();
vmm_init();
heap_init();
slab_init();

No dynamic allocation occurs before heap initialization. Slab becomes
the primary object allocation layer.

---

Slab Allocator Integration

New files:

• memory/slab.c
• memory/slab.h

Purpose:

• Reduce fragmentation
• Enable object caching
• Prepare for per-CPU allocation models
• Replace direct kmalloc usage in higher-level subsystems

---

Kernel Object Model

struct kobject {
    uint32_t type;
    uint32_t flags;
    uint32_t refcount;
};

Benefits:

• Reference counting
• Structured resource lifetime management
• Foundation for handle-based access
• Clean resource ownership model

---

Scheduler Skeleton Integration

Timer IRQ routes through:

scheduler_tick();

Establishes deterministic execution control and preemption foundation.

---

Instrumentation and Research Hooks

void trace_event(uint32_t id);
void profile_enter(uint32_t id);
void profile_exit(uint32_t id);

Used for latency measurement, profiling, and research experiments.

---

Backward Compatibility

• Boot process unchanged
• Interrupts functional
• Timer and keyboard operational
• No external behavior changes

---

Future Research Roadmap

• SMP support (APIC)
• Per-CPU slab caches
• Lock-free scheduling experiments
• Capability-based security model
• Isolation research
• Page cache experimentation
• IPC performance studies

---

Conclusion

This refactor establishes a clean architectural baseline and positions
OpenOS as a structured research kernel platform suitable for advanced
systems experimentation and long-term evolution.