bump version to v0.1.5

Adds IR and debug info generation capabilities for vmlinux imported structs

.gitignore

@@ -7,3 +7,6 @@ __pycache__/
 *.ll
 *.o
 .ipynb_checkpoints/
+vmlinux.py
+~*
+vmlinux.h

.pre-commit-config.yaml

@@ -12,7 +12,7 @@
 #
 # See https://github.com/pre-commit/pre-commit
 
-exclude: 'vmlinux.*\.py$'
+exclude: 'vmlinux.py'
 
 ci:
   autoupdate_commit_msg: "chore: update pre-commit hooks"
@@ -41,7 +41,7 @@ repos:
     - id: ruff
       args: ["--fix", "--show-fixes"]
     - id: ruff-format
-      exclude: ^(docs)|^(tests)|^(examples)
+#      exclude: ^(docs)|^(tests)|^(examples)
 
 # Checking static types
 - repo: https://github.com/pre-commit/mirrors-mypy

README.md

@@ -83,14 +83,14 @@ def hist() -> HashMap:
 def hello(ctx: c_void_p) -> c_int64:
     process_id = pid()
     one = 1
-    prev = hist().lookup(process_id)
+    prev = hist.lookup(process_id)
     if prev:
         previous_value = prev + 1
         print(f"count: {previous_value} with {process_id}")
-        hist().update(process_id, previous_value)
+        hist.update(process_id, previous_value)
         return c_int64(0)
     else:
-        hist().update(process_id, one)
+        hist.update(process_id, one)
     return c_int64(0)
 
 

TODO.md

@@ -1,13 +0,0 @@
-## Short term
-
-- Implement enough functionality to port the BCC tutorial examples in PythonBPF
-- Add all maps
-- XDP support in pylibbpf
-- ringbuf support
-- Add oneline IfExpr conditionals (wishlist)
-
-## Long term
-
-- Refactor the codebase to be better than a hackathon project
-- Port to C++ and use actual LLVM?
-- Fix struct_kioctx issue in the vmlinux transpiler

examples/clone-matplotlib.ipynb

@@ -308,6 +308,7 @@
     "def hist() -> HashMap:\n",
     "    return HashMap(key=c_int32, value=c_uint64, max_entries=4096)\n",
     "\n",
+    "\n",
     "@bpf\n",
     "@section(\"tracepoint/syscalls/sys_enter_clone\")\n",
     "def hello(ctx: c_void_p) -> c_int64:\n",
@@ -329,6 +330,7 @@
     "def LICENSE() -> str:\n",
     "    return \"GPL\"\n",
     "\n",
+    "\n",
     "b = BPF()"
    ]
   },
@@ -357,7 +359,6 @@
     }
    ],
    "source": [
-    "\n",
     "b.load_and_attach()\n",
     "hist = BpfMap(b, hist)\n",
     "print(\"Recording\")\n",

examples/kprobes.py

@@ -8,12 +8,14 @@ def hello_world(ctx: c_void_p) -> c_int64:
     print("Hello, World!")
     return c_int64(0)
 
+
 @bpf
 @section("kprobe/do_unlinkat")
 def hello_world2(ctx: c_void_p) -> c_int64:
     print("Hello, World!")
     return c_int64(0)
 
+
 @bpf
 @bpfglobal
 def LICENSE() -> str:

examples/struct_and_perf.py

@@ -27,7 +27,7 @@ def hello(ctx: c_void_p) -> c_int32:
     dataobj.pid = pid()
     dataobj.ts = ktime()
     # dataobj.comm = strobj
-    print(f"clone called at {dataobj.ts} by pid" f"{dataobj.pid}, comm {strobj}")
+    print(f"clone called at {dataobj.ts} by pid{dataobj.pid}, comm {strobj}")
     events.output(dataobj)
     return c_int32(0)
 

examples/xdp_pass.py

@@ -1,8 +1,8 @@
-from pythonbpf import bpf, map, section, bpfglobal, compile
+from pythonbpf import bpf, map, section, bpfglobal, compile, compile_to_ir
 from pythonbpf.helper import XDP_PASS
 from pythonbpf.maps import HashMap
+from ctypes import c_int64, c_void_p
 
-from ctypes import c_void_p, c_int64
 
 # Instructions to how to run this program
 # 1. Install PythonBPF: pip install pythonbpf
@@ -41,4 +41,5 @@ def LICENSE() -> str:
     return "GPL"
 
 
+compile_to_ir("xdp_pass.py", "xdp_pass.ll")
 compile()

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "pythonbpf"
-version = "0.1.4"
+version = "0.1.5"
 description = "Reduced Python frontend for eBPF"
 authors = [
   { name = "r41k0u", email="pragyanshchaturvedi18@gmail.com" },

pythonbpf/allocation_pass.py

@@ -22,19 +22,6 @@ class LocalSymbol:
         yield self.metadata
 
 
-def _is_helper_call(call_node):
-    """Check if a call node is a BPF helper function call."""
-    if isinstance(call_node.func, ast.Name):
-        # Exclude print from requiring temps (handles f-strings differently)
-        func_name = call_node.func.id
-        return HelperHandlerRegistry.has_handler(func_name) and func_name != "print"
-
-    elif isinstance(call_node.func, ast.Attribute):
-        return HelperHandlerRegistry.has_handler(call_node.func.attr)
-
-    return False
-
-
 def handle_assign_allocation(builder, stmt, local_sym_tab, structs_sym_tab):
     """Handle memory allocation for assignment statements."""
 

pythonbpf/binary_ops.py

@@ -1,110 +0,0 @@
-import ast
-from llvmlite import ir
-from logging import Logger
-import logging
-
-from pythonbpf.expr import get_base_type_and_depth, deref_to_depth, eval_expr
-
-logger: Logger = logging.getLogger(__name__)
-
-
-def get_operand_value(
-    func, module, operand, builder, local_sym_tab, map_sym_tab, structs_sym_tab=None
-):
-    """Extract the value from an operand, handling variables and constants."""
-    logger.info(f"Getting operand value for: {ast.dump(operand)}")
-    if isinstance(operand, ast.Name):
-        if operand.id in local_sym_tab:
-            var = local_sym_tab[operand.id].var
-            var_type = var.type
-            base_type, depth = get_base_type_and_depth(var_type)
-            logger.info(f"var is {var}, base_type is {base_type}, depth is {depth}")
-            val = deref_to_depth(func, builder, var, depth)
-            return val
-        raise ValueError(f"Undefined variable: {operand.id}")
-    elif isinstance(operand, ast.Constant):
-        if isinstance(operand.value, int):
-            cst = ir.Constant(ir.IntType(64), int(operand.value))
-            return cst
-        raise TypeError(f"Unsupported constant type: {type(operand.value)}")
-    elif isinstance(operand, ast.BinOp):
-        res = handle_binary_op_impl(
-            func, module, operand, builder, local_sym_tab, map_sym_tab, structs_sym_tab
-        )
-        return res
-    else:
-        res = eval_expr(
-            func, module, builder, operand, local_sym_tab, map_sym_tab, structs_sym_tab
-        )
-        if res is None:
-            raise ValueError(f"Failed to evaluate call expression: {operand}")
-        val, _ = res
-        logger.info(f"Evaluated expr to {val} of type {val.type}")
-        base_type, depth = get_base_type_and_depth(val.type)
-        if depth > 0:
-            val = deref_to_depth(func, builder, val, depth)
-        return val
-    raise TypeError(f"Unsupported operand type: {type(operand)}")
-
-
-def handle_binary_op_impl(
-    func, module, rval, builder, local_sym_tab, map_sym_tab, structs_sym_tab=None
-):
-    op = rval.op
-    left = get_operand_value(
-        func, module, rval.left, builder, local_sym_tab, map_sym_tab, structs_sym_tab
-    )
-    right = get_operand_value(
-        func, module, rval.right, builder, local_sym_tab, map_sym_tab, structs_sym_tab
-    )
-    logger.info(f"left is {left}, right is {right}, op is {op}")
-
-    # NOTE: Before doing the operation, if the operands are integers
-    # we always extend them to i64. The assignment to LHS will take
-    # care of truncation if needed.
-    if isinstance(left.type, ir.IntType) and left.type.width < 64:
-        left = builder.sext(left, ir.IntType(64))
-    if isinstance(right.type, ir.IntType) and right.type.width < 64:
-        right = builder.sext(right, ir.IntType(64))
-
-    # Map AST operation nodes to LLVM IR builder methods
-    op_map = {
-        ast.Add: builder.add,
-        ast.Sub: builder.sub,
-        ast.Mult: builder.mul,
-        ast.Div: builder.sdiv,
-        ast.Mod: builder.srem,
-        ast.LShift: builder.shl,
-        ast.RShift: builder.lshr,
-        ast.BitOr: builder.or_,
-        ast.BitXor: builder.xor,
-        ast.BitAnd: builder.and_,
-        ast.FloorDiv: builder.udiv,
-    }
-
-    if type(op) in op_map:
-        result = op_map[type(op)](left, right)
-        return result
-    else:
-        raise SyntaxError("Unsupported binary operation")
-
-
-def handle_binary_op(
-    func,
-    module,
-    rval,
-    builder,
-    var_name,
-    local_sym_tab,
-    map_sym_tab,
-    structs_sym_tab=None,
-):
-    result = handle_binary_op_impl(
-        func, module, rval, builder, local_sym_tab, map_sym_tab, structs_sym_tab
-    )
-    if var_name and var_name in local_sym_tab:
-        logger.info(
-            f"Storing result {result} into variable {local_sym_tab[var_name].var}"
-        )
-        builder.store(result, local_sym_tab[var_name].var)
-    return result, result.type

pythonbpf/codegen.py

@@ -4,6 +4,7 @@ from .license_pass import license_processing
 from .functions import func_proc
 from .maps import maps_proc
 from .structs import structs_proc
+from .vmlinux_parser import vmlinux_proc
 from .globals_pass import (
     globals_list_creation,
     globals_processing,
@@ -18,10 +19,20 @@ from pylibbpf import BpfProgram
 import tempfile
 from logging import Logger
 import logging
+import re
 
 logger: Logger = logging.getLogger(__name__)
 
-VERSION = "v0.1.4"
+VERSION = "v0.1.5"
+
+
+def finalize_module(original_str):
+    """After all IR generation is complete, we monkey patch btf_ama attribute"""
+
+    # Create a string with applied transformation of btf_ama attribute addition to BTF struct field accesses.
+    pattern = r'(@"llvm\.[^"]+:[^"]*" = external global i64, !llvm\.preserve\.access\.index ![0-9]+)'
+    replacement = r'\1 "btf_ama"'
+    return re.sub(pattern, replacement, original_str)
 
 
 def find_bpf_chunks(tree):
@@ -44,6 +55,7 @@ def processor(source_code, filename, module):
     for func_node in bpf_chunks:
         logger.info(f"Found BPF function/struct: {func_node.name}")
 
+    vmlinux_proc(tree, module)
     populate_global_symbol_table(tree, module)
     license_processing(tree, module)
     globals_processing(tree, module)
@@ -119,15 +131,44 @@ def compile_to_ir(filename: str, output: str, loglevel=logging.INFO):
 
     module.add_named_metadata("llvm.ident", [f"PythonBPF {VERSION}"])
 
+    module_string = finalize_module(str(module))
+
     logger.info(f"IR written to {output}")
     with open(output, "w") as f:
         f.write(f'source_filename = "{filename}"\n')
-        f.write(str(module))
+        f.write(module_string)
         f.write("\n")
 
     return output
 
 
+def _run_llc(ll_file, obj_file):
+    """Compile LLVM IR to BPF object file using llc."""
+
+    logger.info(f"Compiling IR to object: {ll_file} -> {obj_file}")
+    result = subprocess.run(
+        [
+            "llc",
+            "-march=bpf",
+            "-filetype=obj",
+            "-O2",
+            str(ll_file),
+            "-o",
+            str(obj_file),
+        ],
+        check=True,
+        capture_output=True,
+        text=True,
+    )
+
+    if result.returncode == 0:
+        logger.info(f"Object file written to {obj_file}")
+        return True
+    else:
+        logger.error(f"llc compilation failed: {result.stderr}")
+        return False
+
+
 def compile(loglevel=logging.INFO) -> bool:
     # Look one level up the stack to the caller of this function
     caller_frame = inspect.stack()[1]
@@ -141,21 +182,7 @@ def compile(loglevel=logging.INFO) -> bool:
         compile_to_ir(str(caller_file), str(ll_file), loglevel=loglevel) and success
     )
 
-    success = bool(
-        subprocess.run(
-            [
-                "llc",
-                "-march=bpf",
-                "-filetype=obj",
-                "-O2",
-                str(ll_file),
-                "-o",
-                str(o_file),
-            ],
-            check=True,
-        )
-        and success
-    )
+    success = _run_llc(ll_file, o_file) and success
 
     logger.info(f"Object written to {o_file}")
     return success
@@ -175,17 +202,6 @@ def BPF(loglevel=logging.INFO) -> BpfProgram:
         f.flush()
         source = f.name
         compile_to_ir(source, str(inter.name), loglevel=loglevel)
-        subprocess.run(
-            [
-                "llc",
-                "-march=bpf",
-                "-filetype=obj",
-                "-O2",
-                str(inter.name),
-                "-o",
-                str(obj_file.name),
-            ],
-            check=True,
-        )
+        _run_llc(str(inter.name), str(obj_file.name))
 
         return BpfProgram(str(obj_file.name))

pythonbpf/expr/__init__.py

@@ -1,5 +1,7 @@
-from .expr_pass import eval_expr, handle_expr
-from .type_normalization import convert_to_bool, get_base_type_and_depth, deref_to_depth
+from .expr_pass import eval_expr, handle_expr, get_operand_value
+from .type_normalization import convert_to_bool, get_base_type_and_depth
+from .ir_ops import deref_to_depth
+from .call_registry import CallHandlerRegistry
 
 __all__ = [
     "eval_expr",
@@ -7,4 +9,6 @@ __all__ = [
     "convert_to_bool",
     "get_base_type_and_depth",
     "deref_to_depth",
+    "get_operand_value",
+    "CallHandlerRegistry",
 ]

pythonbpf/expr/call_registry.py

@@ -0,0 +1,20 @@
+class CallHandlerRegistry:
+    """Registry for handling different types of calls (helpers, etc.)"""
+
+    _handler = None
+
+    @classmethod
+    def set_handler(cls, handler):
+        """Set the handler for unknown calls"""
+        cls._handler = handler
+
+    @classmethod
+    def handle_call(
+        cls, call, module, builder, func, local_sym_tab, map_sym_tab, structs_sym_tab
+    ):
+        """Handle a call using the registered handler"""
+        if cls._handler is None:
+            return None
+        return cls._handler(
+            call, module, builder, func, local_sym_tab, map_sym_tab, structs_sym_tab
+        )

pythonbpf/expr/expr_pass.py

@@ -5,10 +5,20 @@ import logging
 from typing import Dict
 
 from pythonbpf.type_deducer import ctypes_to_ir, is_ctypes
-from .type_normalization import convert_to_bool, handle_comparator
+from .call_registry import CallHandlerRegistry
+from .type_normalization import (
+    convert_to_bool,
+    handle_comparator,
+    get_base_type_and_depth,
+    deref_to_depth,
+)
 
 logger: Logger = logging.getLogger(__name__)
 
+# ============================================================================
+# Leaf Handlers (No Recursive eval_expr calls)
+# ============================================================================
+
 
 def _handle_name_expr(expr: ast.Name, local_sym_tab: Dict, builder: ir.IRBuilder):
     """Handle ast.Name expressions."""
@@ -21,10 +31,24 @@ def _handle_name_expr(expr: ast.Name, local_sym_tab: Dict, builder: ir.IRBuilder
         return None
 
 
-def _handle_constant_expr(expr: ast.Constant):
+def _handle_constant_expr(module, builder, expr: ast.Constant):
     """Handle ast.Constant expressions."""
     if isinstance(expr.value, int) or isinstance(expr.value, bool):
         return ir.Constant(ir.IntType(64), int(expr.value)), ir.IntType(64)
+    elif isinstance(expr.value, str):
+        str_name = f".str.{id(expr)}"
+        str_bytes = expr.value.encode("utf-8") + b"\x00"
+        str_type = ir.ArrayType(ir.IntType(8), len(str_bytes))
+        str_constant = ir.Constant(str_type, bytearray(str_bytes))
+
+        # Create global variable
+        global_str = ir.GlobalVariable(module, str_type, name=str_name)
+        global_str.linkage = "internal"
+        global_str.global_constant = True
+        global_str.initializer = str_constant
+
+        str_ptr = builder.bitcast(global_str, ir.PointerType(ir.IntType(8)))
+        return str_ptr, ir.PointerType(ir.IntType(8))
     else:
         logger.error(f"Unsupported constant type {ast.dump(expr)}")
         return None
@@ -88,6 +112,118 @@ def _handle_deref_call(expr: ast.Call, local_sym_tab: Dict, builder: ir.IRBuilde
     return val, local_sym_tab[arg.id].ir_type
 
 
+# ============================================================================
+# Binary Operations
+# ============================================================================
+
+
+def get_operand_value(
+    func, module, operand, builder, local_sym_tab, map_sym_tab, structs_sym_tab=None
+):
+    """Extract the value from an operand, handling variables and constants."""
+    logger.info(f"Getting operand value for: {ast.dump(operand)}")
+    if isinstance(operand, ast.Name):
+        if operand.id in local_sym_tab:
+            var = local_sym_tab[operand.id].var
+            var_type = var.type
+            base_type, depth = get_base_type_and_depth(var_type)
+            logger.info(f"var is {var}, base_type is {base_type}, depth is {depth}")
+            val = deref_to_depth(func, builder, var, depth)
+            return val
+        raise ValueError(f"Undefined variable: {operand.id}")
+    elif isinstance(operand, ast.Constant):
+        if isinstance(operand.value, int):
+            cst = ir.Constant(ir.IntType(64), int(operand.value))
+            return cst
+        raise TypeError(f"Unsupported constant type: {type(operand.value)}")
+    elif isinstance(operand, ast.BinOp):
+        res = _handle_binary_op_impl(
+            func, module, operand, builder, local_sym_tab, map_sym_tab, structs_sym_tab
+        )
+        return res
+    else:
+        res = eval_expr(
+            func, module, builder, operand, local_sym_tab, map_sym_tab, structs_sym_tab
+        )
+        if res is None:
+            raise ValueError(f"Failed to evaluate call expression: {operand}")
+        val, _ = res
+        logger.info(f"Evaluated expr to {val} of type {val.type}")
+        base_type, depth = get_base_type_and_depth(val.type)
+        if depth > 0:
+            val = deref_to_depth(func, builder, val, depth)
+        return val
+    raise TypeError(f"Unsupported operand type: {type(operand)}")
+
+
+def _handle_binary_op_impl(
+    func, module, rval, builder, local_sym_tab, map_sym_tab, structs_sym_tab=None
+):
+    op = rval.op
+    left = get_operand_value(
+        func, module, rval.left, builder, local_sym_tab, map_sym_tab, structs_sym_tab
+    )
+    right = get_operand_value(
+        func, module, rval.right, builder, local_sym_tab, map_sym_tab, structs_sym_tab
+    )
+    logger.info(f"left is {left}, right is {right}, op is {op}")
+
+    # NOTE: Before doing the operation, if the operands are integers
+    # we always extend them to i64. The assignment to LHS will take
+    # care of truncation if needed.
+    if isinstance(left.type, ir.IntType) and left.type.width < 64:
+        left = builder.sext(left, ir.IntType(64))
+    if isinstance(right.type, ir.IntType) and right.type.width < 64:
+        right = builder.sext(right, ir.IntType(64))
+
+    # Map AST operation nodes to LLVM IR builder methods
+    op_map = {
+        ast.Add: builder.add,
+        ast.Sub: builder.sub,
+        ast.Mult: builder.mul,
+        ast.Div: builder.sdiv,
+        ast.Mod: builder.srem,
+        ast.LShift: builder.shl,
+        ast.RShift: builder.lshr,
+        ast.BitOr: builder.or_,
+        ast.BitXor: builder.xor,
+        ast.BitAnd: builder.and_,
+        ast.FloorDiv: builder.udiv,
+    }
+
+    if type(op) in op_map:
+        result = op_map[type(op)](left, right)
+        return result
+    else:
+        raise SyntaxError("Unsupported binary operation")
+
+
+def _handle_binary_op(
+    func,
+    module,
+    rval,
+    builder,
+    var_name,
+    local_sym_tab,
+    map_sym_tab,
+    structs_sym_tab=None,
+):
+    result = _handle_binary_op_impl(
+        func, module, rval, builder, local_sym_tab, map_sym_tab, structs_sym_tab
+    )
+    if var_name and var_name in local_sym_tab:
+        logger.info(
+            f"Storing result {result} into variable {local_sym_tab[var_name].var}"
+        )
+        builder.store(result, local_sym_tab[var_name].var)
+    return result, result.type
+
+
+# ============================================================================
+# Comparison and Unary Operations
+# ============================================================================
+
+
 def _handle_ctypes_call(
     func,
     module,
@@ -180,8 +316,6 @@ def _handle_unary_op(
         logger.error("Only 'not' and '-' unary operators are supported")
         return None
 
-    from pythonbpf.binary_ops import get_operand_value
-
     operand = get_operand_value(
         func, module, expr.operand, builder, local_sym_tab, map_sym_tab, structs_sym_tab
     )
@@ -200,6 +334,11 @@ def _handle_unary_op(
         return result, ir.IntType(64)
 
 
+# ============================================================================
+# Boolean Operations
+# ============================================================================
+
+
 def _handle_and_op(func, builder, expr, local_sym_tab, map_sym_tab, structs_sym_tab):
     """Handle `and` boolean operations."""
 
@@ -330,6 +469,11 @@ def _handle_boolean_op(
         return None
 
 
+# ============================================================================
+# Expression Dispatcher
+# ============================================================================
+
+
 def eval_expr(
     func,
     module,
@@ -343,7 +487,7 @@ def eval_expr(
     if isinstance(expr, ast.Name):
         return _handle_name_expr(expr, local_sym_tab, builder)
     elif isinstance(expr, ast.Constant):
-        return _handle_constant_expr(expr)
+        return _handle_constant_expr(module, builder, expr)
     elif isinstance(expr, ast.Call):
         if isinstance(expr.func, ast.Name) and expr.func.id == "deref":
             return _handle_deref_call(expr, local_sym_tab, builder)
@@ -359,57 +503,18 @@ def eval_expr(
                 structs_sym_tab,
             )
 
-        # delayed import to avoid circular dependency
-        from pythonbpf.helper import HelperHandlerRegistry, handle_helper_call
+        result = CallHandlerRegistry.handle_call(
+            expr, module, builder, func, local_sym_tab, map_sym_tab, structs_sym_tab
+        )
+        if result is not None:
+            return result
 
-        if isinstance(expr.func, ast.Name) and HelperHandlerRegistry.has_handler(
-            expr.func.id
-        ):
-            return handle_helper_call(
-                expr,
-                module,
-                builder,
-                func,
-                local_sym_tab,
-                map_sym_tab,
-                structs_sym_tab,
-            )
-        elif isinstance(expr.func, ast.Attribute):
-            logger.info(f"Handling method call: {ast.dump(expr.func)}")
-            if isinstance(expr.func.value, ast.Call) and isinstance(
-                expr.func.value.func, ast.Name
-            ):
-                method_name = expr.func.attr
-                if HelperHandlerRegistry.has_handler(method_name):
-                    return handle_helper_call(
-                        expr,
-                        module,
-                        builder,
-                        func,
-                        local_sym_tab,
-                        map_sym_tab,
-                        structs_sym_tab,
-                    )
-            elif isinstance(expr.func.value, ast.Name):
-                obj_name = expr.func.value.id
-                method_name = expr.func.attr
-                if obj_name in map_sym_tab:
-                    if HelperHandlerRegistry.has_handler(method_name):
-                        return handle_helper_call(
-                            expr,
-                            module,
-                            builder,
-                            func,
-                            local_sym_tab,
-                            map_sym_tab,
-                            structs_sym_tab,
-                        )
+        logger.warning(f"Unknown call: {ast.dump(expr)}")
+        return None
     elif isinstance(expr, ast.Attribute):
         return _handle_attribute_expr(expr, local_sym_tab, structs_sym_tab, builder)
     elif isinstance(expr, ast.BinOp):
-        from pythonbpf.binary_ops import handle_binary_op
-
-        return handle_binary_op(
+        return _handle_binary_op(
             func,
             module,
             expr,

pythonbpf/expr/ir_ops.py

@@ -0,0 +1,50 @@
+import logging
+from llvmlite import ir
+
+logger = logging.getLogger(__name__)
+
+
+def deref_to_depth(func, builder, val, target_depth):
+    """Dereference a pointer to a certain depth."""
+
+    cur_val = val
+    cur_type = val.type
+
+    for depth in range(target_depth):
+        if not isinstance(val.type, ir.PointerType):
+            logger.error("Cannot dereference further, non-pointer type")
+            return None
+
+        # dereference with null check
+        pointee_type = cur_type.pointee
+        null_check_block = builder.block
+        not_null_block = func.append_basic_block(name=f"deref_not_null_{depth}")
+        merge_block = func.append_basic_block(name=f"deref_merge_{depth}")
+
+        null_ptr = ir.Constant(cur_type, None)
+        is_not_null = builder.icmp_signed("!=", cur_val, null_ptr)
+        logger.debug(f"Inserted null check for pointer at depth {depth}")
+
+        builder.cbranch(is_not_null, not_null_block, merge_block)
+
+        builder.position_at_end(not_null_block)
+        dereferenced_val = builder.load(cur_val)
+        logger.debug(f"Dereferenced to depth {depth - 1}, type: {pointee_type}")
+        builder.branch(merge_block)
+
+        builder.position_at_end(merge_block)
+        phi = builder.phi(pointee_type, name=f"deref_result_{depth}")
+
+        zero_value = (
+            ir.Constant(pointee_type, 0)
+            if isinstance(pointee_type, ir.IntType)
+            else ir.Constant(pointee_type, None)
+        )
+        phi.add_incoming(zero_value, null_check_block)
+
+        phi.add_incoming(dereferenced_val, not_null_block)
+
+        # Continue with phi result
+        cur_val = phi
+        cur_type = pointee_type
+    return cur_val

pythonbpf/expr/type_normalization.py

@@ -1,6 +1,7 @@
-from llvmlite import ir
 import logging
 import ast
+from llvmlite import ir
+from .ir_ops import deref_to_depth
 
 logger = logging.getLogger(__name__)
 
@@ -26,52 +27,6 @@ def get_base_type_and_depth(ir_type):
     return cur_type, depth
 
 
-def deref_to_depth(func, builder, val, target_depth):
-    """Dereference a pointer to a certain depth."""
-
-    cur_val = val
-    cur_type = val.type
-
-    for depth in range(target_depth):
-        if not isinstance(val.type, ir.PointerType):
-            logger.error("Cannot dereference further, non-pointer type")
-            return None
-
-        # dereference with null check
-        pointee_type = cur_type.pointee
-        null_check_block = builder.block
-        not_null_block = func.append_basic_block(name=f"deref_not_null_{depth}")
-        merge_block = func.append_basic_block(name=f"deref_merge_{depth}")
-
-        null_ptr = ir.Constant(cur_type, None)
-        is_not_null = builder.icmp_signed("!=", cur_val, null_ptr)
-        logger.debug(f"Inserted null check for pointer at depth {depth}")
-
-        builder.cbranch(is_not_null, not_null_block, merge_block)
-
-        builder.position_at_end(not_null_block)
-        dereferenced_val = builder.load(cur_val)
-        logger.debug(f"Dereferenced to depth {depth - 1}, type: {pointee_type}")
-        builder.branch(merge_block)
-
-        builder.position_at_end(merge_block)
-        phi = builder.phi(pointee_type, name=f"deref_result_{depth}")
-
-        zero_value = (
-            ir.Constant(pointee_type, 0)
-            if isinstance(pointee_type, ir.IntType)
-            else ir.Constant(pointee_type, None)
-        )
-        phi.add_incoming(zero_value, null_check_block)
-
-        phi.add_incoming(dereferenced_val, not_null_block)
-
-        # Continue with phi result
-        cur_val = phi
-        cur_type = pointee_type
-    return cur_val
-
-
 def _normalize_types(func, builder, lhs, rhs):
     """Normalize types for comparison."""
 

pythonbpf/functions/func_registry_handlers.py

@@ -1,22 +0,0 @@
-from typing import Dict
-
-
-class StatementHandlerRegistry:
-    """Registry for statement handlers."""
-
-    _handlers: Dict = {}
-
-    @classmethod
-    def register(cls, stmt_type):
-        """Register a handler for a specific statement type."""
-
-        def decorator(handler):
-            cls._handlers[stmt_type] = handler
-            return handler
-
-        return decorator
-
-    @classmethod
-    def __getitem__(cls, stmt_type):
-        """Get the handler for a specific statement type."""
-        return cls._handlers.get(stmt_type, None)

pythonbpf/functions/function_metadata.py

@@ -0,0 +1,88 @@
+import ast
+
+
+def get_probe_string(func_node):
+    """Extract the probe string from the decorator of the function node"""
+    # TODO: right now we have the whole string in the section decorator
+    # But later we can implement typed tuples for tracepoints and kprobes
+    # For helper functions, we return "helper"
+
+    for decorator in func_node.decorator_list:
+        if isinstance(decorator, ast.Name) and decorator.id == "bpfglobal":
+            return None
+        if isinstance(decorator, ast.Call) and isinstance(decorator.func, ast.Name):
+            if decorator.func.id == "section" and len(decorator.args) == 1:
+                arg = decorator.args[0]
+                if isinstance(arg, ast.Constant) and isinstance(arg.value, str):
+                    return arg.value
+    return "helper"
+
+
+def is_global_function(func_node):
+    """Check if the function is a global"""
+    for decorator in func_node.decorator_list:
+        if isinstance(decorator, ast.Name) and decorator.id in (
+            "map",
+            "bpfglobal",
+            "struct",
+        ):
+            return True
+    return False
+
+
+def infer_return_type(func_node: ast.FunctionDef):
+    if not isinstance(func_node, (ast.FunctionDef, ast.AsyncFunctionDef)):
+        raise TypeError("Expected ast.FunctionDef")
+    if func_node.returns is not None:
+        try:
+            return ast.unparse(func_node.returns)
+        except Exception:
+            node = func_node.returns
+            if isinstance(node, ast.Name):
+                return node.id
+            if isinstance(node, ast.Attribute):
+                return getattr(node, "attr", type(node).__name__)
+            try:
+                return str(node)
+            except Exception:
+                return type(node).__name__
+    found_type = None
+
+    def _expr_type(e):
+        if e is None:
+            return "None"
+        if isinstance(e, ast.Constant):
+            return type(e.value).__name__
+        if isinstance(e, ast.Name):
+            return e.id
+        if isinstance(e, ast.Call):
+            f = e.func
+            if isinstance(f, ast.Name):
+                return f.id
+            if isinstance(f, ast.Attribute):
+                try:
+                    return ast.unparse(f)
+                except Exception:
+                    return getattr(f, "attr", type(f).__name__)
+            try:
+                return ast.unparse(f)
+            except Exception:
+                return type(f).__name__
+        if isinstance(e, ast.Attribute):
+            try:
+                return ast.unparse(e)
+            except Exception:
+                return getattr(e, "attr", type(e).__name__)
+        try:
+            return ast.unparse(e)
+        except Exception:
+            return type(e).__name__
+
+    for walked_node in ast.walk(func_node):
+        if isinstance(walked_node, ast.Return):
+            t = _expr_type(walked_node.value)
+            if found_type is None:
+                found_type = t
+            elif found_type != t:
+                raise ValueError(f"Conflicting return types: {found_type} vs {t}")
+    return found_type or "None"

pythonbpf/functions/functions_pass.py

@@ -14,27 +14,125 @@ from pythonbpf.assign_pass import (
 )
 from pythonbpf.allocation_pass import handle_assign_allocation, allocate_temp_pool
 
-from .return_utils import _handle_none_return, _handle_xdp_return, _is_xdp_name
+from .return_utils import handle_none_return, handle_xdp_return, is_xdp_name
+from .function_metadata import get_probe_string, is_global_function, infer_return_type
 
 
 logger = logging.getLogger(__name__)
 
 
-def get_probe_string(func_node):
-    """Extract the probe string from the decorator of the function node."""
-    # TODO: right now we have the whole string in the section decorator
-    # But later we can implement typed tuples for tracepoints and kprobes
-    # For helper functions, we return "helper"
+# ============================================================================
+# SECTION 1: Memory Allocation
+# ============================================================================
 
-    for decorator in func_node.decorator_list:
-        if isinstance(decorator, ast.Name) and decorator.id == "bpfglobal":
-            return None
-        if isinstance(decorator, ast.Call) and isinstance(decorator.func, ast.Name):
-            if decorator.func.id == "section" and len(decorator.args) == 1:
-                arg = decorator.args[0]
-                if isinstance(arg, ast.Constant) and isinstance(arg.value, str):
-                    return arg.value
-    return "helper"
+
+def count_temps_in_call(call_node, local_sym_tab):
+    """Count the number of temporary variables needed for a function call."""
+
+    count = 0
+    is_helper = False
+
+    # NOTE: We exclude print calls for now
+    if isinstance(call_node.func, ast.Name):
+        if (
+            HelperHandlerRegistry.has_handler(call_node.func.id)
+            and call_node.func.id != "print"
+        ):
+            is_helper = True
+    elif isinstance(call_node.func, ast.Attribute):
+        if HelperHandlerRegistry.has_handler(call_node.func.attr):
+            is_helper = True
+
+    if not is_helper:
+        return 0
+
+    for arg in call_node.args:
+        # NOTE: Count all non-name arguments
+        # For struct fields, if it is being passed as an argument,
+        # The struct object should already exist in the local_sym_tab
+        if not isinstance(arg, ast.Name) and not (
+            isinstance(arg, ast.Attribute) and arg.value.id in local_sym_tab
+        ):
+            count += 1
+
+    return count
+
+
+def handle_if_allocation(
+    module, builder, stmt, func, ret_type, map_sym_tab, local_sym_tab, structs_sym_tab
+):
+    """Recursively handle allocations in if/else branches."""
+    if stmt.body:
+        allocate_mem(
+            module,
+            builder,
+            stmt.body,
+            func,
+            ret_type,
+            map_sym_tab,
+            local_sym_tab,
+            structs_sym_tab,
+        )
+    if stmt.orelse:
+        allocate_mem(
+            module,
+            builder,
+            stmt.orelse,
+            func,
+            ret_type,
+            map_sym_tab,
+            local_sym_tab,
+            structs_sym_tab,
+        )
+
+
+def allocate_mem(
+    module, builder, body, func, ret_type, map_sym_tab, local_sym_tab, structs_sym_tab
+):
+    max_temps_needed = 0
+
+    def update_max_temps_for_stmt(stmt):
+        nonlocal max_temps_needed
+        temps_needed = 0
+
+        if isinstance(stmt, ast.If):
+            for s in stmt.body:
+                update_max_temps_for_stmt(s)
+            for s in stmt.orelse:
+                update_max_temps_for_stmt(s)
+            return
+
+        for node in ast.walk(stmt):
+            if isinstance(node, ast.Call):
+                temps_needed += count_temps_in_call(node, local_sym_tab)
+        max_temps_needed = max(max_temps_needed, temps_needed)
+
+    for stmt in body:
+        update_max_temps_for_stmt(stmt)
+
+        # Handle allocations
+        if isinstance(stmt, ast.If):
+            handle_if_allocation(
+                module,
+                builder,
+                stmt,
+                func,
+                ret_type,
+                map_sym_tab,
+                local_sym_tab,
+                structs_sym_tab,
+            )
+        elif isinstance(stmt, ast.Assign):
+            handle_assign_allocation(builder, stmt, local_sym_tab, structs_sym_tab)
+
+    allocate_temp_pool(builder, max_temps_needed, local_sym_tab)
+
+    return local_sym_tab
+
+
+# ============================================================================
+# SECTION 2: Statement Handlers
+# ============================================================================
 
 
 def handle_assign(
@@ -146,9 +244,9 @@ def handle_if(
 def handle_return(builder, stmt, local_sym_tab, ret_type):
     logger.info(f"Handling return statement: {ast.dump(stmt)}")
     if stmt.value is None:
-        return _handle_none_return(builder)
-    elif isinstance(stmt.value, ast.Name) and _is_xdp_name(stmt.value.id):
-        return _handle_xdp_return(stmt, builder, ret_type)
+        return handle_none_return(builder)
+    elif isinstance(stmt.value, ast.Name) and is_xdp_name(stmt.value.id):
+        return handle_xdp_return(stmt, builder, ret_type)
     else:
         val = eval_expr(
             func=None,
@@ -207,108 +305,9 @@ def process_stmt(
     return did_return
 
 
-def handle_if_allocation(
-    module, builder, stmt, func, ret_type, map_sym_tab, local_sym_tab, structs_sym_tab
-):
-    """Recursively handle allocations in if/else branches."""
-    if stmt.body:
-        allocate_mem(
-            module,
-            builder,
-            stmt.body,
-            func,
-            ret_type,
-            map_sym_tab,
-            local_sym_tab,
-            structs_sym_tab,
-        )
-    if stmt.orelse:
-        allocate_mem(
-            module,
-            builder,
-            stmt.orelse,
-            func,
-            ret_type,
-            map_sym_tab,
-            local_sym_tab,
-            structs_sym_tab,
-        )
-
-
-def count_temps_in_call(call_node, local_sym_tab):
-    """Count the number of temporary variables needed for a function call."""
-
-    count = 0
-    is_helper = False
-
-    # NOTE: We exclude print calls for now
-    if isinstance(call_node.func, ast.Name):
-        if (
-            HelperHandlerRegistry.has_handler(call_node.func.id)
-            and call_node.func.id != "print"
-        ):
-            is_helper = True
-    elif isinstance(call_node.func, ast.Attribute):
-        if HelperHandlerRegistry.has_handler(call_node.func.attr):
-            is_helper = True
-
-    if not is_helper:
-        return 0
-
-    for arg in call_node.args:
-        # NOTE: Count all non-name arguments
-        # For struct fields, if it is being passed as an argument,
-        # The struct object should already exist in the local_sym_tab
-        if not isinstance(arg, ast.Name) and not (
-            isinstance(arg, ast.Attribute) and arg.value.id in local_sym_tab
-        ):
-            count += 1
-
-    return count
-
-
-def allocate_mem(
-    module, builder, body, func, ret_type, map_sym_tab, local_sym_tab, structs_sym_tab
-):
-    max_temps_needed = 0
-
-    def update_max_temps_for_stmt(stmt):
-        nonlocal max_temps_needed
-        temps_needed = 0
-
-        if isinstance(stmt, ast.If):
-            for s in stmt.body:
-                update_max_temps_for_stmt(s)
-            for s in stmt.orelse:
-                update_max_temps_for_stmt(s)
-            return
-
-        for node in ast.walk(stmt):
-            if isinstance(node, ast.Call):
-                temps_needed += count_temps_in_call(node, local_sym_tab)
-        max_temps_needed = max(max_temps_needed, temps_needed)
-
-    for stmt in body:
-        update_max_temps_for_stmt(stmt)
-
-        # Handle allocations
-        if isinstance(stmt, ast.If):
-            handle_if_allocation(
-                module,
-                builder,
-                stmt,
-                func,
-                ret_type,
-                map_sym_tab,
-                local_sym_tab,
-                structs_sym_tab,
-            )
-        elif isinstance(stmt, ast.Assign):
-            handle_assign_allocation(builder, stmt, local_sym_tab, structs_sym_tab)
-
-    allocate_temp_pool(builder, max_temps_needed, local_sym_tab)
-
-    return local_sym_tab
+# ============================================================================
+# SECTION 3: Function Body Processing
+# ============================================================================
 
 
 def process_func_body(
@@ -390,18 +389,14 @@ def process_bpf_chunk(func_node, module, return_type, map_sym_tab, structs_sym_t
     return func
 
 
+# ============================================================================
+# SECTION 4: Top-Level Function Processor
+# ============================================================================
+
+
 def func_proc(tree, module, chunks, map_sym_tab, structs_sym_tab):
     for func_node in chunks:
-        is_global = False
-        for decorator in func_node.decorator_list:
-            if isinstance(decorator, ast.Name) and decorator.id in (
-                "map",
-                "bpfglobal",
-                "struct",
-            ):
-                is_global = True
-                break
-        if is_global:
+        if is_global_function(func_node):
             continue
         func_type = get_probe_string(func_node)
         logger.info(f"Found probe_string of {func_node.name}: {func_type}")
@@ -415,67 +410,7 @@ def func_proc(tree, module, chunks, map_sym_tab, structs_sym_tab):
         )
 
 
-def infer_return_type(func_node: ast.FunctionDef):
-    if not isinstance(func_node, (ast.FunctionDef, ast.AsyncFunctionDef)):
-        raise TypeError("Expected ast.FunctionDef")
-    if func_node.returns is not None:
-        try:
-            return ast.unparse(func_node.returns)
-        except Exception:
-            node = func_node.returns
-            if isinstance(node, ast.Name):
-                return node.id
-            if isinstance(node, ast.Attribute):
-                return getattr(node, "attr", type(node).__name__)
-            try:
-                return str(node)
-            except Exception:
-                return type(node).__name__
-    found_type = None
-
-    def _expr_type(e):
-        if e is None:
-            return "None"
-        if isinstance(e, ast.Constant):
-            return type(e.value).__name__
-        if isinstance(e, ast.Name):
-            return e.id
-        if isinstance(e, ast.Call):
-            f = e.func
-            if isinstance(f, ast.Name):
-                return f.id
-            if isinstance(f, ast.Attribute):
-                try:
-                    return ast.unparse(f)
-                except Exception:
-                    return getattr(f, "attr", type(f).__name__)
-            try:
-                return ast.unparse(f)
-            except Exception:
-                return type(f).__name__
-        if isinstance(e, ast.Attribute):
-            try:
-                return ast.unparse(e)
-            except Exception:
-                return getattr(e, "attr", type(e).__name__)
-        try:
-            return ast.unparse(e)
-        except Exception:
-            return type(e).__name__
-
-    for walked_node in ast.walk(func_node):
-        if isinstance(walked_node, ast.Return):
-            t = _expr_type(walked_node.value)
-            if found_type is None:
-                found_type = t
-            elif found_type != t:
-                raise ValueError(f"Conflicting return types: {found_type} vs {t}")
-    return found_type or "None"
-
-
-# For string assignment to fixed-size arrays
-
-
+# TODO: WIP, for string assignment to fixed-size arrays
 def assign_string_to_array(builder, target_array_ptr, source_string_ptr, array_length):
     """
     Copy a string (i8*) to a fixed-size array ([N x i8]*)

pythonbpf/functions/return_utils.py

@@ -14,19 +14,19 @@ XDP_ACTIONS = {
 }
 
 
-def _handle_none_return(builder) -> bool:
+def handle_none_return(builder) -> bool:
     """Handle return or return None -> returns 0."""
     builder.ret(ir.Constant(ir.IntType(64), 0))
     logger.debug("Generated default return: 0")
     return True
 
 
-def _is_xdp_name(name: str) -> bool:
+def is_xdp_name(name: str) -> bool:
     """Check if a name is an XDP action"""
     return name in XDP_ACTIONS
 
 
-def _handle_xdp_return(stmt: ast.Return, builder, ret_type) -> bool:
+def handle_xdp_return(stmt: ast.Return, builder, ret_type) -> bool:
     """Handle XDP returns"""
     if not isinstance(stmt.value, ast.Name):
         return False
@@ -37,7 +37,6 @@ def _handle_xdp_return(stmt: ast.Return, builder, ret_type) -> bool:
         raise ValueError(
             f"Unknown XDP action: {action_name}. Available: {XDP_ACTIONS.keys()}"
         )
-        return False
 
     value = XDP_ACTIONS[action_name]
     builder.ret(ir.Constant(ret_type, value))

pythonbpf/helper/__init__.py

@@ -1,7 +1,60 @@
-from .helper_utils import HelperHandlerRegistry, reset_scratch_pool
+from .helper_registry import HelperHandlerRegistry
+from .helper_utils import reset_scratch_pool
 from .bpf_helper_handler import handle_helper_call
 from .helpers import ktime, pid, deref, XDP_DROP, XDP_PASS
 
+
+# Register the helper handler with expr module
+def _register_helper_handler():
+    """Register helper call handler with the expression evaluator"""
+    from pythonbpf.expr.expr_pass import CallHandlerRegistry
+
+    def helper_call_handler(
+        call, module, builder, func, local_sym_tab, map_sym_tab, structs_sym_tab
+    ):
+        """Check if call is a helper and handle it"""
+        import ast
+
+        # Check for direct helper calls (e.g., ktime(), print())
+        if isinstance(call.func, ast.Name):
+            if HelperHandlerRegistry.has_handler(call.func.id):
+                return handle_helper_call(
+                    call,
+                    module,
+                    builder,
+                    func,
+                    local_sym_tab,
+                    map_sym_tab,
+                    structs_sym_tab,
+                )
+
+        # Check for method calls (e.g., map.lookup())
+        elif isinstance(call.func, ast.Attribute):
+            method_name = call.func.attr
+
+            # Handle: my_map.lookup(key)
+            if isinstance(call.func.value, ast.Name):
+                obj_name = call.func.value.id
+                if map_sym_tab and obj_name in map_sym_tab:
+                    if HelperHandlerRegistry.has_handler(method_name):
+                        return handle_helper_call(
+                            call,
+                            module,
+                            builder,
+                            func,
+                            local_sym_tab,
+                            map_sym_tab,
+                            structs_sym_tab,
+                        )
+
+        return None
+
+    CallHandlerRegistry.set_handler(helper_call_handler)
+
+
+# Register on module import
+_register_helper_handler()
+
 __all__ = [
     "HelperHandlerRegistry",
     "reset_scratch_pool",

pythonbpf/helper/bpf_helper_handler.py

@@ -1,14 +1,15 @@
 import ast
 from llvmlite import ir
 from enum import Enum
+
+from .helper_registry import HelperHandlerRegistry
 from .helper_utils import (
-    HelperHandlerRegistry,
     get_or_create_ptr_from_arg,
     get_flags_val,
-    handle_fstring_print,
-    simple_string_print,
     get_data_ptr_and_size,
 )
+from .printk_formatter import simple_string_print, handle_fstring_print
+
 from logging import Logger
 import logging
 
@@ -135,7 +136,7 @@ def bpf_printk_emitter(
     fn_ptr = builder.inttoptr(fn_addr, fn_ptr_type)
 
     builder.call(fn_ptr, args, tail=True)
-    return None
+    return True
 
 
 @HelperHandlerRegistry.register("update")

pythonbpf/helper/helper_registry.py

@@ -0,0 +1,27 @@
+from typing import Callable
+
+
+class HelperHandlerRegistry:
+    """Registry for BPF helpers"""
+
+    _handlers: dict[str, Callable] = {}
+
+    @classmethod
+    def register(cls, helper_name):
+        """Decorator to register a handler function for a helper"""
+
+        def decorator(func):
+            cls._handlers[helper_name] = func
+            return func
+
+        return decorator
+
+    @classmethod
+    def get_handler(cls, helper_name):
+        """Get the handler function for a helper"""
+        return cls._handlers.get(helper_name)
+
+    @classmethod
+    def has_handler(cls, helper_name):
+        """Check if a handler function is registered for a helper"""
+        return helper_name in cls._handlers

pythonbpf/helper/helper_utils.py

@@ -1,40 +1,14 @@
 import ast
 import logging
-from collections.abc import Callable
 
 from llvmlite import ir
-from pythonbpf.expr import eval_expr, get_base_type_and_depth, deref_to_depth
-from pythonbpf.binary_ops import get_operand_value
+from pythonbpf.expr import (
+    get_operand_value,
+)
 
 logger = logging.getLogger(__name__)
 
 
-class HelperHandlerRegistry:
-    """Registry for BPF helpers"""
-
-    _handlers: dict[str, Callable] = {}
-
-    @classmethod
-    def register(cls, helper_name):
-        """Decorator to register a handler function for a helper"""
-
-        def decorator(func):
-            cls._handlers[helper_name] = func
-            return func
-
-        return decorator
-
-    @classmethod
-    def get_handler(cls, helper_name):
-        """Get the handler function for a helper"""
-        return cls._handlers.get(helper_name)
-
-    @classmethod
-    def has_handler(cls, helper_name):
-        """Check if a handler function is registered for a helper"""
-        return helper_name in cls._handlers
-
-
 class ScratchPoolManager:
     """Manage the temporary helper variables in local_sym_tab"""
 
@@ -70,6 +44,11 @@ def reset_scratch_pool():
     _temp_pool_manager.reset()
 
 
+# ============================================================================
+# Argument Preparation
+# ============================================================================
+
+
 def get_var_ptr_from_name(var_name, local_sym_tab):
     """Get a pointer to a variable from the symbol table."""
     if local_sym_tab and var_name in local_sym_tab:
@@ -134,234 +113,6 @@ def get_flags_val(arg, builder, local_sym_tab):
     )
 
 
-def simple_string_print(string_value, module, builder, func):
-    """Prepare arguments for bpf_printk from a simple string value"""
-    fmt_str = string_value + "\n\0"
-    fmt_ptr = _create_format_string_global(fmt_str, func, module, builder)
-
-    args = [fmt_ptr, ir.Constant(ir.IntType(32), len(fmt_str))]
-    return args
-
-
-def handle_fstring_print(
-    joined_str,
-    module,
-    builder,
-    func,
-    local_sym_tab=None,
-    struct_sym_tab=None,
-):
-    """Handle f-string formatting for bpf_printk emitter."""
-    fmt_parts = []
-    exprs = []
-
-    for value in joined_str.values:
-        logger.debug(f"Processing f-string value: {ast.dump(value)}")
-
-        if isinstance(value, ast.Constant):
-            _process_constant_in_fstring(value, fmt_parts, exprs)
-        elif isinstance(value, ast.FormattedValue):
-            _process_fval(
-                value,
-                fmt_parts,
-                exprs,
-                local_sym_tab,
-                struct_sym_tab,
-            )
-        else:
-            raise NotImplementedError(f"Unsupported f-string value type: {type(value)}")
-
-    fmt_str = "".join(fmt_parts)
-    args = simple_string_print(fmt_str, module, builder, func)
-
-    # NOTE: Process expressions (limited to 3 due to BPF constraints)
-    if len(exprs) > 3:
-        logger.warning("bpf_printk supports up to 3 args, extra args will be ignored.")
-
-    for expr in exprs[:3]:
-        arg_value = _prepare_expr_args(
-            expr,
-            func,
-            module,
-            builder,
-            local_sym_tab,
-            struct_sym_tab,
-        )
-        args.append(arg_value)
-
-    return args
-
-
-def _process_constant_in_fstring(cst, fmt_parts, exprs):
-    """Process constant values in f-string."""
-    if isinstance(cst.value, str):
-        fmt_parts.append(cst.value)
-    elif isinstance(cst.value, int):
-        fmt_parts.append("%lld")
-        exprs.append(ir.Constant(ir.IntType(64), cst.value))
-    else:
-        raise NotImplementedError(
-            f"Unsupported constant type in f-string: {type(cst.value)}"
-        )
-
-
-def _process_fval(fval, fmt_parts, exprs, local_sym_tab, struct_sym_tab):
-    """Process formatted values in f-string."""
-    logger.debug(f"Processing formatted value: {ast.dump(fval)}")
-
-    if isinstance(fval.value, ast.Name):
-        _process_name_in_fval(fval.value, fmt_parts, exprs, local_sym_tab)
-    elif isinstance(fval.value, ast.Attribute):
-        _process_attr_in_fval(
-            fval.value,
-            fmt_parts,
-            exprs,
-            local_sym_tab,
-            struct_sym_tab,
-        )
-    else:
-        raise NotImplementedError(
-            f"Unsupported formatted value in f-string: {type(fval.value)}"
-        )
-
-
-def _process_name_in_fval(name_node, fmt_parts, exprs, local_sym_tab):
-    """Process name nodes in formatted values."""
-    if local_sym_tab and name_node.id in local_sym_tab:
-        _, var_type, tmp = local_sym_tab[name_node.id]
-        _populate_fval(var_type, name_node, fmt_parts, exprs)
-
-
-def _process_attr_in_fval(attr_node, fmt_parts, exprs, local_sym_tab, struct_sym_tab):
-    """Process attribute nodes in formatted values."""
-    if (
-        isinstance(attr_node.value, ast.Name)
-        and local_sym_tab
-        and attr_node.value.id in local_sym_tab
-    ):
-        var_name = attr_node.value.id
-        field_name = attr_node.attr
-
-        var_type = local_sym_tab[var_name].metadata
-        if var_type not in struct_sym_tab:
-            raise ValueError(
-                f"Struct '{var_type}' for '{var_name}' not in symbol table"
-            )
-
-        struct_info = struct_sym_tab[var_type]
-        if field_name not in struct_info.fields:
-            raise ValueError(f"Field '{field_name}' not found in struct '{var_type}'")
-
-        field_type = struct_info.field_type(field_name)
-        _populate_fval(field_type, attr_node, fmt_parts, exprs)
-    else:
-        raise NotImplementedError(
-            "Only simple attribute on local vars is supported in f-strings."
-        )
-
-
-def _populate_fval(ftype, node, fmt_parts, exprs):
-    """Populate format parts and expressions based on field type."""
-    if isinstance(ftype, ir.IntType):
-        # TODO: We print as signed integers only for now
-        if ftype.width == 64:
-            fmt_parts.append("%lld")
-            exprs.append(node)
-        elif ftype.width == 32:
-            fmt_parts.append("%d")
-            exprs.append(node)
-        else:
-            raise NotImplementedError(
-                f"Unsupported integer width in f-string: {ftype.width}"
-            )
-    elif isinstance(ftype, ir.PointerType):
-        target, depth = get_base_type_and_depth(ftype)
-        if isinstance(target, ir.IntType):
-            if target.width == 64:
-                fmt_parts.append("%lld")
-                exprs.append(node)
-            elif target.width == 32:
-                fmt_parts.append("%d")
-                exprs.append(node)
-            elif target.width == 8 and depth == 1:
-                # NOTE: Assume i8* is a string
-                fmt_parts.append("%s")
-                exprs.append(node)
-            else:
-                raise NotImplementedError(
-                    f"Unsupported pointer target type in f-string: {target}"
-                )
-        else:
-            raise NotImplementedError(
-                f"Unsupported pointer target type in f-string: {target}"
-            )
-    else:
-        raise NotImplementedError(f"Unsupported field type in f-string: {ftype}")
-
-
-def _create_format_string_global(fmt_str, func, module, builder):
-    """Create a global variable for the format string."""
-    fmt_name = f"{func.name}____fmt{func._fmt_counter}"
-    func._fmt_counter += 1
-
-    fmt_gvar = ir.GlobalVariable(
-        module, ir.ArrayType(ir.IntType(8), len(fmt_str)), name=fmt_name
-    )
-    fmt_gvar.global_constant = True
-    fmt_gvar.initializer = ir.Constant(
-        ir.ArrayType(ir.IntType(8), len(fmt_str)), bytearray(fmt_str.encode("utf8"))
-    )
-    fmt_gvar.linkage = "internal"
-    fmt_gvar.align = 1
-
-    return builder.bitcast(fmt_gvar, ir.PointerType())
-
-
-def _prepare_expr_args(expr, func, module, builder, local_sym_tab, struct_sym_tab):
-    """Evaluate and prepare an expression to use as an arg for bpf_printk."""
-    val, _ = eval_expr(
-        func,
-        module,
-        builder,
-        expr,
-        local_sym_tab,
-        None,
-        struct_sym_tab,
-    )
-
-    if val:
-        if isinstance(val.type, ir.PointerType):
-            target, depth = get_base_type_and_depth(val.type)
-            if isinstance(target, ir.IntType):
-                if target.width >= 32:
-                    val = deref_to_depth(func, builder, val, depth)
-                    val = builder.sext(val, ir.IntType(64))
-                elif target.width == 8 and depth == 1:
-                    # NOTE: i8* is string, no need to deref
-                    pass
-
-            else:
-                logger.warning(
-                    "Only int and ptr supported in bpf_printk args. Others default to 0."
-                )
-                val = ir.Constant(ir.IntType(64), 0)
-        elif isinstance(val.type, ir.IntType):
-            if val.type.width < 64:
-                val = builder.sext(val, ir.IntType(64))
-        else:
-            logger.warning(
-                "Only int and ptr supported in bpf_printk args. Others default to 0."
-            )
-            val = ir.Constant(ir.IntType(64), 0)
-        return val
-    else:
-        logger.warning(
-            "Failed to evaluate expression for bpf_printk argument. "
-            "It will be converted to 0."
-        )
-        return ir.Constant(ir.IntType(64), 0)
-
-
 def get_data_ptr_and_size(data_arg, local_sym_tab, struct_sym_tab):
     """Extract data pointer and size information for perf event output."""
     if isinstance(data_arg, ast.Name):

pythonbpf/helper/helpers.py

@@ -15,5 +15,8 @@ def deref(ptr):
     return result if result is not None else 0
 
 
+XDP_ABORTED = ctypes.c_int64(0)
 XDP_DROP = ctypes.c_int64(1)
 XDP_PASS = ctypes.c_int64(2)
+XDP_TX = ctypes.c_int64(3)
+XDP_REDIRECT = ctypes.c_int64(4)

pythonbpf/helper/printk_formatter.py

@@ -0,0 +1,240 @@
+import ast
+import logging
+
+from llvmlite import ir
+from pythonbpf.expr import eval_expr, get_base_type_and_depth, deref_to_depth
+
+logger = logging.getLogger(__name__)
+
+
+def simple_string_print(string_value, module, builder, func):
+    """Prepare arguments for bpf_printk from a simple string value"""
+    fmt_str = string_value + "\n\0"
+    fmt_ptr = _create_format_string_global(fmt_str, func, module, builder)
+
+    args = [fmt_ptr, ir.Constant(ir.IntType(32), len(fmt_str))]
+    return args
+
+
+def handle_fstring_print(
+    joined_str,
+    module,
+    builder,
+    func,
+    local_sym_tab=None,
+    struct_sym_tab=None,
+):
+    """Handle f-string formatting for bpf_printk emitter."""
+    fmt_parts = []
+    exprs = []
+
+    for value in joined_str.values:
+        logger.debug(f"Processing f-string value: {ast.dump(value)}")
+
+        if isinstance(value, ast.Constant):
+            _process_constant_in_fstring(value, fmt_parts, exprs)
+        elif isinstance(value, ast.FormattedValue):
+            _process_fval(
+                value,
+                fmt_parts,
+                exprs,
+                local_sym_tab,
+                struct_sym_tab,
+            )
+        else:
+            raise NotImplementedError(f"Unsupported f-string value type: {type(value)}")
+
+    fmt_str = "".join(fmt_parts)
+    args = simple_string_print(fmt_str, module, builder, func)
+
+    # NOTE: Process expressions (limited to 3 due to BPF constraints)
+    if len(exprs) > 3:
+        logger.warning("bpf_printk supports up to 3 args, extra args will be ignored.")
+
+    for expr in exprs[:3]:
+        arg_value = _prepare_expr_args(
+            expr,
+            func,
+            module,
+            builder,
+            local_sym_tab,
+            struct_sym_tab,
+        )
+        args.append(arg_value)
+
+    return args
+
+
+# ============================================================================
+# Internal Helpers
+# ============================================================================
+
+
+def _process_constant_in_fstring(cst, fmt_parts, exprs):
+    """Process constant values in f-string."""
+    if isinstance(cst.value, str):
+        fmt_parts.append(cst.value)
+    elif isinstance(cst.value, int):
+        fmt_parts.append("%lld")
+        exprs.append(ir.Constant(ir.IntType(64), cst.value))
+    else:
+        raise NotImplementedError(
+            f"Unsupported constant type in f-string: {type(cst.value)}"
+        )
+
+
+def _process_fval(fval, fmt_parts, exprs, local_sym_tab, struct_sym_tab):
+    """Process formatted values in f-string."""
+    logger.debug(f"Processing formatted value: {ast.dump(fval)}")
+
+    if isinstance(fval.value, ast.Name):
+        _process_name_in_fval(fval.value, fmt_parts, exprs, local_sym_tab)
+    elif isinstance(fval.value, ast.Attribute):
+        _process_attr_in_fval(
+            fval.value,
+            fmt_parts,
+            exprs,
+            local_sym_tab,
+            struct_sym_tab,
+        )
+    else:
+        raise NotImplementedError(
+            f"Unsupported formatted value in f-string: {type(fval.value)}"
+        )
+
+
+def _process_name_in_fval(name_node, fmt_parts, exprs, local_sym_tab):
+    """Process name nodes in formatted values."""
+    if local_sym_tab and name_node.id in local_sym_tab:
+        _, var_type, tmp = local_sym_tab[name_node.id]
+        _populate_fval(var_type, name_node, fmt_parts, exprs)
+
+
+def _process_attr_in_fval(attr_node, fmt_parts, exprs, local_sym_tab, struct_sym_tab):
+    """Process attribute nodes in formatted values."""
+    if (
+        isinstance(attr_node.value, ast.Name)
+        and local_sym_tab
+        and attr_node.value.id in local_sym_tab
+    ):
+        var_name = attr_node.value.id
+        field_name = attr_node.attr
+
+        var_type = local_sym_tab[var_name].metadata
+        if var_type not in struct_sym_tab:
+            raise ValueError(
+                f"Struct '{var_type}' for '{var_name}' not in symbol table"
+            )
+
+        struct_info = struct_sym_tab[var_type]
+        if field_name not in struct_info.fields:
+            raise ValueError(f"Field '{field_name}' not found in struct '{var_type}'")
+
+        field_type = struct_info.field_type(field_name)
+        _populate_fval(field_type, attr_node, fmt_parts, exprs)
+    else:
+        raise NotImplementedError(
+            "Only simple attribute on local vars is supported in f-strings."
+        )
+
+
+def _populate_fval(ftype, node, fmt_parts, exprs):
+    """Populate format parts and expressions based on field type."""
+    if isinstance(ftype, ir.IntType):
+        # TODO: We print as signed integers only for now
+        if ftype.width == 64:
+            fmt_parts.append("%lld")
+            exprs.append(node)
+        elif ftype.width == 32:
+            fmt_parts.append("%d")
+            exprs.append(node)
+        else:
+            raise NotImplementedError(
+                f"Unsupported integer width in f-string: {ftype.width}"
+            )
+    elif isinstance(ftype, ir.PointerType):
+        target, depth = get_base_type_and_depth(ftype)
+        if isinstance(target, ir.IntType):
+            if target.width == 64:
+                fmt_parts.append("%lld")
+                exprs.append(node)
+            elif target.width == 32:
+                fmt_parts.append("%d")
+                exprs.append(node)
+            elif target.width == 8 and depth == 1:
+                # NOTE: Assume i8* is a string
+                fmt_parts.append("%s")
+                exprs.append(node)
+            else:
+                raise NotImplementedError(
+                    f"Unsupported pointer target type in f-string: {target}"
+                )
+        else:
+            raise NotImplementedError(
+                f"Unsupported pointer target type in f-string: {target}"
+            )
+    else:
+        raise NotImplementedError(f"Unsupported field type in f-string: {ftype}")
+
+
+def _create_format_string_global(fmt_str, func, module, builder):
+    """Create a global variable for the format string."""
+    fmt_name = f"{func.name}____fmt{func._fmt_counter}"
+    func._fmt_counter += 1
+
+    fmt_gvar = ir.GlobalVariable(
+        module, ir.ArrayType(ir.IntType(8), len(fmt_str)), name=fmt_name
+    )
+    fmt_gvar.global_constant = True
+    fmt_gvar.initializer = ir.Constant(
+        ir.ArrayType(ir.IntType(8), len(fmt_str)), bytearray(fmt_str.encode("utf8"))
+    )
+    fmt_gvar.linkage = "internal"
+    fmt_gvar.align = 1
+
+    return builder.bitcast(fmt_gvar, ir.PointerType())
+
+
+def _prepare_expr_args(expr, func, module, builder, local_sym_tab, struct_sym_tab):
+    """Evaluate and prepare an expression to use as an arg for bpf_printk."""
+    val, _ = eval_expr(
+        func,
+        module,
+        builder,
+        expr,
+        local_sym_tab,
+        None,
+        struct_sym_tab,
+    )
+
+    if val:
+        if isinstance(val.type, ir.PointerType):
+            target, depth = get_base_type_and_depth(val.type)
+            if isinstance(target, ir.IntType):
+                if target.width >= 32:
+                    val = deref_to_depth(func, builder, val, depth)
+                    val = builder.sext(val, ir.IntType(64))
+                elif target.width == 8 and depth == 1:
+                    # NOTE: i8* is string, no need to deref
+                    pass
+
+            else:
+                logger.warning(
+                    "Only int and ptr supported in bpf_printk args. Others default to 0."
+                )
+                val = ir.Constant(ir.IntType(64), 0)
+        elif isinstance(val.type, ir.IntType):
+            if val.type.width < 64:
+                val = builder.sext(val, ir.IntType(64))
+        else:
+            logger.warning(
+                "Only int and ptr supported in bpf_printk args. Others default to 0."
+            )
+            val = ir.Constant(ir.IntType(64), 0)
+        return val
+    else:
+        logger.warning(
+            "Failed to evaluate expression for bpf_printk argument. "
+            "It will be converted to 0."
+        )
+        return ir.Constant(ir.IntType(64), 0)

pythonbpf/maps/map_debug_info.py

@@ -0,0 +1,93 @@
+from pythonbpf.debuginfo import DebugInfoGenerator
+from .map_types import BPFMapType
+
+
+def create_map_debug_info(module, map_global, map_name, map_params):
+    """Generate debug info metadata for BPF maps HASH and PERF_EVENT_ARRAY"""
+    generator = DebugInfoGenerator(module)
+
+    uint_type = generator.get_uint32_type()
+    ulong_type = generator.get_uint64_type()
+    array_type = generator.create_array_type(
+        uint_type, map_params.get("type", BPFMapType.UNSPEC).value
+    )
+    type_ptr = generator.create_pointer_type(array_type, 64)
+    key_ptr = generator.create_pointer_type(
+        array_type if "key_size" in map_params else ulong_type, 64
+    )
+    value_ptr = generator.create_pointer_type(
+        array_type if "value_size" in map_params else ulong_type, 64
+    )
+
+    elements_arr = []
+
+    # Create struct members
+    # scope field does not appear for some reason
+    cnt = 0
+    for elem in map_params:
+        if elem == "max_entries":
+            continue
+        if elem == "type":
+            ptr = type_ptr
+        elif "key" in elem:
+            ptr = key_ptr
+        else:
+            ptr = value_ptr
+        # TODO: the best way to do this is not 64, but get the size each time. this will not work for structs.
+        member = generator.create_struct_member(elem, ptr, cnt * 64)
+        elements_arr.append(member)
+        cnt += 1
+
+    if "max_entries" in map_params:
+        max_entries_array = generator.create_array_type(
+            uint_type, map_params["max_entries"]
+        )
+        max_entries_ptr = generator.create_pointer_type(max_entries_array, 64)
+        max_entries_member = generator.create_struct_member(
+            "max_entries", max_entries_ptr, cnt * 64
+        )
+        elements_arr.append(max_entries_member)
+
+    # Create the struct type
+    struct_type = generator.create_struct_type(
+        elements_arr, 64 * len(elements_arr), is_distinct=True
+    )
+
+    # Create global variable debug info
+    global_var = generator.create_global_var_debug_info(
+        map_name, struct_type, is_local=False
+    )
+
+    # Attach debug info to the global variable
+    map_global.set_metadata("dbg", global_var)
+
+    return global_var
+
+
+def create_ringbuf_debug_info(module, map_global, map_name, map_params):
+    """Generate debug information metadata for BPF RINGBUF map"""
+    generator = DebugInfoGenerator(module)
+
+    int_type = generator.get_int32_type()
+
+    type_array = generator.create_array_type(
+        int_type, map_params.get("type", BPFMapType.RINGBUF).value
+    )
+    type_ptr = generator.create_pointer_type(type_array, 64)
+    type_member = generator.create_struct_member("type", type_ptr, 0)
+
+    max_entries_array = generator.create_array_type(int_type, map_params["max_entries"])
+    max_entries_ptr = generator.create_pointer_type(max_entries_array, 64)
+    max_entries_member = generator.create_struct_member(
+        "max_entries", max_entries_ptr, 64
+    )
+
+    elements_arr = [type_member, max_entries_member]
+
+    struct_type = generator.create_struct_type(elements_arr, 128, is_distinct=True)
+
+    global_var = generator.create_global_var_debug_info(
+        map_name, struct_type, is_local=False
+    )
+    map_global.set_metadata("dbg", global_var)
+    return global_var

pythonbpf/maps/map_types.py

@@ -0,0 +1,39 @@
+from enum import Enum
+
+
+class BPFMapType(Enum):
+    UNSPEC = 0
+    HASH = 1
+    ARRAY = 2
+    PROG_ARRAY = 3
+    PERF_EVENT_ARRAY = 4
+    PERCPU_HASH = 5
+    PERCPU_ARRAY = 6
+    STACK_TRACE = 7
+    CGROUP_ARRAY = 8
+    LRU_HASH = 9
+    LRU_PERCPU_HASH = 10
+    LPM_TRIE = 11
+    ARRAY_OF_MAPS = 12
+    HASH_OF_MAPS = 13
+    DEVMAP = 14
+    SOCKMAP = 15
+    CPUMAP = 16
+    XSKMAP = 17
+    SOCKHASH = 18
+    CGROUP_STORAGE_DEPRECATED = 19
+    CGROUP_STORAGE = 19
+    REUSEPORT_SOCKARRAY = 20
+    PERCPU_CGROUP_STORAGE_DEPRECATED = 21
+    PERCPU_CGROUP_STORAGE = 21
+    QUEUE = 22
+    STACK = 23
+    SK_STORAGE = 24
+    DEVMAP_HASH = 25
+    STRUCT_OPS = 26
+    RINGBUF = 27
+    INODE_STORAGE = 28
+    TASK_STORAGE = 29
+    BLOOM_FILTER = 30
+    USER_RINGBUF = 31
+    CGRP_STORAGE = 32

pythonbpf/maps/maps_pass.py

@@ -1,10 +1,11 @@
 import ast
+import logging
 from logging import Logger
 from llvmlite import ir
-from enum import Enum
+
 from .maps_utils import MapProcessorRegistry
-from pythonbpf.debuginfo import DebugInfoGenerator
-import logging
+from .map_types import BPFMapType
+from .map_debug_info import create_map_debug_info, create_ringbuf_debug_info
 
 logger: Logger = logging.getLogger(__name__)
 
@@ -26,44 +27,6 @@ def is_map(func_node):
     )
 
 
-class BPFMapType(Enum):
-    UNSPEC = 0
-    HASH = 1
-    ARRAY = 2
-    PROG_ARRAY = 3
-    PERF_EVENT_ARRAY = 4
-    PERCPU_HASH = 5
-    PERCPU_ARRAY = 6
-    STACK_TRACE = 7
-    CGROUP_ARRAY = 8
-    LRU_HASH = 9
-    LRU_PERCPU_HASH = 10
-    LPM_TRIE = 11
-    ARRAY_OF_MAPS = 12
-    HASH_OF_MAPS = 13
-    DEVMAP = 14
-    SOCKMAP = 15
-    CPUMAP = 16
-    XSKMAP = 17
-    SOCKHASH = 18
-    CGROUP_STORAGE_DEPRECATED = 19
-    CGROUP_STORAGE = 19
-    REUSEPORT_SOCKARRAY = 20
-    PERCPU_CGROUP_STORAGE_DEPRECATED = 21
-    PERCPU_CGROUP_STORAGE = 21
-    QUEUE = 22
-    STACK = 23
-    SK_STORAGE = 24
-    DEVMAP_HASH = 25
-    STRUCT_OPS = 26
-    RINGBUF = 27
-    INODE_STORAGE = 28
-    TASK_STORAGE = 29
-    BLOOM_FILTER = 30
-    USER_RINGBUF = 31
-    CGRP_STORAGE = 32
-
-
 def create_bpf_map(module, map_name, map_params):
     """Create a BPF map in the module with given parameters and debug info"""
 
@@ -84,114 +47,37 @@ def create_bpf_map(module, map_name, map_params):
     return map_global
 
 
-def create_map_debug_info(module, map_global, map_name, map_params):
-    """Generate debug info metadata for BPF maps HASH and PERF_EVENT_ARRAY"""
-    generator = DebugInfoGenerator(module)
+def _parse_map_params(rval, expected_args=None):
+    """Parse map parameters from call arguments and keywords."""
 
-    uint_type = generator.get_uint32_type()
-    ulong_type = generator.get_uint64_type()
-    array_type = generator.create_array_type(
-        uint_type, map_params.get("type", BPFMapType.UNSPEC).value
-    )
-    type_ptr = generator.create_pointer_type(array_type, 64)
-    key_ptr = generator.create_pointer_type(
-        array_type if "key_size" in map_params else ulong_type, 64
-    )
-    value_ptr = generator.create_pointer_type(
-        array_type if "value_size" in map_params else ulong_type, 64
-    )
+    params = {}
 
-    elements_arr = []
+    # Parse positional arguments
+    if expected_args:
+        for i, arg_name in enumerate(expected_args):
+            if i < len(rval.args):
+                arg = rval.args[i]
+                if isinstance(arg, ast.Name):
+                    params[arg_name] = arg.id
+                elif isinstance(arg, ast.Constant):
+                    params[arg_name] = arg.value
 
-    # Create struct members
-    # scope field does not appear for some reason
-    cnt = 0
-    for elem in map_params:
-        if elem == "max_entries":
-            continue
-        if elem == "type":
-            ptr = type_ptr
-        elif "key" in elem:
-            ptr = key_ptr
-        else:
-            ptr = value_ptr
-        # TODO: the best way to do this is not 64, but get the size each time. this will not work for structs.
-        member = generator.create_struct_member(elem, ptr, cnt * 64)
-        elements_arr.append(member)
-        cnt += 1
+    # Parse keyword arguments (override positional)
+    for keyword in rval.keywords:
+        if isinstance(keyword.value, ast.Name):
+            params[keyword.arg] = keyword.value.id
+        elif isinstance(keyword.value, ast.Constant):
+            params[keyword.arg] = keyword.value.value
 
-    if "max_entries" in map_params:
-        max_entries_array = generator.create_array_type(
-            uint_type, map_params["max_entries"]
-        )
-        max_entries_ptr = generator.create_pointer_type(max_entries_array, 64)
-        max_entries_member = generator.create_struct_member(
-            "max_entries", max_entries_ptr, cnt * 64
-        )
-        elements_arr.append(max_entries_member)
-
-    # Create the struct type
-    struct_type = generator.create_struct_type(
-        elements_arr, 64 * len(elements_arr), is_distinct=True
-    )
-
-    # Create global variable debug info
-    global_var = generator.create_global_var_debug_info(
-        map_name, struct_type, is_local=False
-    )
-
-    # Attach debug info to the global variable
-    map_global.set_metadata("dbg", global_var)
-
-    return global_var
-
-
-def create_ringbuf_debug_info(module, map_global, map_name, map_params):
-    """Generate debug information metadata for BPF RINGBUF map"""
-    generator = DebugInfoGenerator(module)
-
-    int_type = generator.get_int32_type()
-
-    type_array = generator.create_array_type(
-        int_type, map_params.get("type", BPFMapType.RINGBUF).value
-    )
-    type_ptr = generator.create_pointer_type(type_array, 64)
-    type_member = generator.create_struct_member("type", type_ptr, 0)
-
-    max_entries_array = generator.create_array_type(int_type, map_params["max_entries"])
-    max_entries_ptr = generator.create_pointer_type(max_entries_array, 64)
-    max_entries_member = generator.create_struct_member(
-        "max_entries", max_entries_ptr, 64
-    )
-
-    elements_arr = [type_member, max_entries_member]
-
-    struct_type = generator.create_struct_type(elements_arr, 128, is_distinct=True)
-
-    global_var = generator.create_global_var_debug_info(
-        map_name, struct_type, is_local=False
-    )
-    map_global.set_metadata("dbg", global_var)
-    return global_var
+    return params
 
 
 @MapProcessorRegistry.register("RingBuf")
 def process_ringbuf_map(map_name, rval, module):
     """Process a BPF_RINGBUF map declaration"""
     logger.info(f"Processing Ringbuf: {map_name}")
-    map_params = {"type": BPFMapType.RINGBUF}
-
-    # Parse max_entries if present
-    if len(rval.args) >= 1 and isinstance(rval.args[0], ast.Constant):
-        const_val = rval.args[0].value
-        if isinstance(const_val, int):
-            map_params["max_entries"] = const_val
-
-    for keyword in rval.keywords:
-        if keyword.arg == "max_entries" and isinstance(keyword.value, ast.Constant):
-            const_val = keyword.value.value
-            if isinstance(const_val, int):
-                map_params["max_entries"] = const_val
+    map_params = _parse_map_params(rval, expected_args=["max_entries"])
+    map_params["type"] = BPFMapType.RINGBUF
 
     logger.info(f"Ringbuf map parameters: {map_params}")
 
@@ -204,27 +90,8 @@ def process_ringbuf_map(map_name, rval, module):
 def process_hash_map(map_name, rval, module):
     """Process a BPF_HASH map declaration"""
     logger.info(f"Processing HashMap: {map_name}")
-    map_params = {"type": BPFMapType.HASH}
-
-    # Assuming order: key_type, value_type, max_entries
-    if len(rval.args) >= 1 and isinstance(rval.args[0], ast.Name):
-        map_params["key"] = rval.args[0].id
-    if len(rval.args) >= 2 and isinstance(rval.args[1], ast.Name):
-        map_params["value"] = rval.args[1].id
-    if len(rval.args) >= 3 and isinstance(rval.args[2], ast.Constant):
-        const_val = rval.args[2].value
-        if isinstance(const_val, (int, str)):  # safe check
-            map_params["max_entries"] = const_val
-
-    for keyword in rval.keywords:
-        if keyword.arg == "key" and isinstance(keyword.value, ast.Name):
-            map_params["key"] = keyword.value.id
-        elif keyword.arg == "value" and isinstance(keyword.value, ast.Name):
-            map_params["value"] = keyword.value.id
-        elif keyword.arg == "max_entries" and isinstance(keyword.value, ast.Constant):
-            const_val = keyword.value.value
-            if isinstance(const_val, (int, str)):
-                map_params["max_entries"] = const_val
+    map_params = _parse_map_params(rval, expected_args=["key", "value", "max_entries"])
+    map_params["type"] = BPFMapType.HASH
 
     logger.info(f"Map parameters: {map_params}")
     map_global = create_bpf_map(module, map_name, map_params)
@@ -237,18 +104,8 @@ def process_hash_map(map_name, rval, module):
 def process_perf_event_map(map_name, rval, module):
     """Process a BPF_PERF_EVENT_ARRAY map declaration"""
     logger.info(f"Processing PerfEventArray: {map_name}")
-    map_params = {"type": BPFMapType.PERF_EVENT_ARRAY}
-
-    if len(rval.args) >= 1 and isinstance(rval.args[0], ast.Name):
-        map_params["key_size"] = rval.args[0].id
-    if len(rval.args) >= 2 and isinstance(rval.args[1], ast.Name):
-        map_params["value_size"] = rval.args[1].id
-
-    for keyword in rval.keywords:
-        if keyword.arg == "key_size" and isinstance(keyword.value, ast.Name):
-            map_params["key_size"] = keyword.value.id
-        elif keyword.arg == "value_size" and isinstance(keyword.value, ast.Name):
-            map_params["value_size"] = keyword.value.id
+    map_params = _parse_map_params(rval, expected_args=["key_size", "value_size"])
+    map_params["type"] = BPFMapType.PERF_EVENT_ARRAY
 
     logger.info(f"Map parameters: {map_params}")
     map_global = create_bpf_map(module, map_name, map_params)

pythonbpf/vmlinux_parser/__init__.py

@@ -0,0 +1,3 @@
+from .import_detector import vmlinux_proc
+
+__all__ = ["vmlinux_proc"]

pythonbpf/vmlinux_parser/class_handler.py

@@ -0,0 +1,205 @@
+import logging
+from functools import lru_cache
+import importlib
+from .dependency_handler import DependencyHandler
+from .dependency_node import DependencyNode
+import ctypes
+from typing import Optional, Any, Dict
+
+logger = logging.getLogger(__name__)
+
+
+@lru_cache(maxsize=1)
+def get_module_symbols(module_name: str):
+    imported_module = importlib.import_module(module_name)
+    return [name for name in dir(imported_module)], imported_module
+
+
+def process_vmlinux_class(node, llvm_module, handler: DependencyHandler):
+    symbols_in_module, imported_module = get_module_symbols("vmlinux")
+    if node.name in symbols_in_module:
+        vmlinux_type = getattr(imported_module, node.name)
+        process_vmlinux_post_ast(vmlinux_type, llvm_module, handler)
+    else:
+        raise ImportError(f"{node.name} not in vmlinux")
+
+
+def process_vmlinux_post_ast(
+    elem_type_class, llvm_handler, handler: DependencyHandler, processing_stack=None
+):
+    # Initialize processing stack on first call
+    if processing_stack is None:
+        processing_stack = set()
+    symbols_in_module, imported_module = get_module_symbols("vmlinux")
+
+    current_symbol_name = elem_type_class.__name__
+    logger.info(f"Begin {current_symbol_name} Processing")
+    field_table: Dict[str, list] = {}
+    is_complex_type = False
+    containing_type: Optional[Any] = None
+    ctype_complex_type: Optional[Any] = None
+    type_length: Optional[int] = None
+    module_name = getattr(elem_type_class, "__module__", None)
+
+    # Check if already processed
+    if handler.has_node(current_symbol_name):
+        logger.debug(f"Node {current_symbol_name} already processed and ready")
+        return True
+
+    # XXX:Check it's use. It's probably not being used.
+    if current_symbol_name in processing_stack:
+        logger.debug(
+            f"Dependency already in processing stack for {current_symbol_name}, skipping"
+        )
+        return True
+
+    processing_stack.add(current_symbol_name)
+
+    if module_name == "vmlinux":
+        if hasattr(elem_type_class, "_type_"):
+            pass
+        else:
+            new_dep_node = DependencyNode(name=current_symbol_name)
+
+            # elem_type_class is the actual vmlinux struct/class
+            new_dep_node.set_ctype_struct(elem_type_class)
+
+            handler.add_node(new_dep_node)
+            class_obj = getattr(imported_module, current_symbol_name)
+            # Inspect the class fields
+            if hasattr(class_obj, "_fields_"):
+                for field_elem in class_obj._fields_:
+                    field_name: str = ""
+                    field_type: Optional[Any] = None
+                    bitfield_size: Optional[int] = None
+                    if len(field_elem) == 2:
+                        field_name, field_type = field_elem
+                    elif len(field_elem) == 3:
+                        field_name, field_type, bitfield_size = field_elem
+                    field_table[field_name] = [field_type, bitfield_size]
+            elif hasattr(class_obj, "__annotations__"):
+                for field_elem in class_obj.__annotations__.items():
+                    if len(field_elem) == 2:
+                        field_name, field_type = field_elem
+                        bitfield_size = None
+                    elif len(field_elem) == 3:
+                        field_name, field_type, bitfield_size = field_elem
+                    else:
+                        raise ValueError(
+                            "Number of fields in items() of class object unexpected"
+                        )
+                    field_table[field_name] = [field_type, bitfield_size]
+            else:
+                raise TypeError("Could not get required class and definition")
+
+            logger.debug(f"Extracted fields for {current_symbol_name}: {field_table}")
+            for elem in field_table.items():
+                elem_name, elem_temp_list = elem
+                [elem_type, elem_bitfield_size] = elem_temp_list
+                local_module_name = getattr(elem_type, "__module__", None)
+                new_dep_node.add_field(elem_name, elem_type, ready=False)
+                if local_module_name == ctypes.__name__:
+                    new_dep_node.set_field_bitfield_size(elem_name, elem_bitfield_size)
+                    new_dep_node.set_field_ready(elem_name, is_ready=True)
+                    logger.debug(
+                        f"Field {elem_name} is direct ctypes type: {elem_type}"
+                    )
+                elif local_module_name == "vmlinux":
+                    new_dep_node.set_field_bitfield_size(elem_name, elem_bitfield_size)
+                    logger.debug(
+                        f"Processing vmlinux field: {elem_name}, type: {elem_type}"
+                    )
+                    if hasattr(elem_type, "_type_"):
+                        is_complex_type = True
+                        containing_type = elem_type._type_
+                        if hasattr(elem_type, "_length_") and is_complex_type:
+                            type_length = elem_type._length_
+
+                        if containing_type.__module__ == "vmlinux":
+                            new_dep_node.add_dependent(
+                                elem_type._type_.__name__
+                                if hasattr(elem_type._type_, "__name__")
+                                else str(elem_type._type_)
+                            )
+                        elif containing_type.__module__ == ctypes.__name__:
+                            if isinstance(elem_type, type):
+                                if issubclass(elem_type, ctypes.Array):
+                                    ctype_complex_type = ctypes.Array
+                                elif issubclass(elem_type, ctypes._Pointer):
+                                    ctype_complex_type = ctypes._Pointer
+                            else:
+                                raise TypeError("Unsupported ctypes subclass")
+                        else:
+                            raise ImportError(
+                                f"Unsupported module of {containing_type}"
+                            )
+                        logger.debug(
+                            f"{containing_type} containing type of parent {elem_name} with {elem_type} and ctype {ctype_complex_type} and length {type_length}"
+                        )
+                        new_dep_node.set_field_containing_type(
+                            elem_name, containing_type
+                        )
+                        new_dep_node.set_field_type_size(elem_name, type_length)
+                        new_dep_node.set_field_ctype_complex_type(
+                            elem_name, ctype_complex_type
+                        )
+                        new_dep_node.set_field_type(elem_name, elem_type)
+                        if containing_type.__module__ == "vmlinux":
+                            containing_type_name = (
+                                containing_type.__name__
+                                if hasattr(containing_type, "__name__")
+                                else str(containing_type)
+                            )
+
+                            # Check for self-reference or already processed
+                            if containing_type_name == current_symbol_name:
+                                # Self-referential pointer
+                                logger.debug(
+                                    f"Self-referential pointer in {current_symbol_name}.{elem_name}"
+                                )
+                                new_dep_node.set_field_ready(elem_name, True)
+                            elif handler.has_node(containing_type_name):
+                                # Already processed
+                                logger.debug(
+                                    f"Reusing already processed {containing_type_name}"
+                                )
+                                new_dep_node.set_field_ready(elem_name, True)
+                            else:
+                                # Process recursively - THIS WAS MISSING
+                                new_dep_node.add_dependent(containing_type_name)
+                                process_vmlinux_post_ast(
+                                    containing_type,
+                                    llvm_handler,
+                                    handler,
+                                    processing_stack,
+                                )
+                                new_dep_node.set_field_ready(elem_name, True)
+                        elif containing_type.__module__ == ctypes.__name__:
+                            logger.debug(f"Processing ctype internal{containing_type}")
+                            new_dep_node.set_field_ready(elem_name, True)
+                        else:
+                            raise TypeError(
+                                "Module not supported in recursive resolution"
+                            )
+                    else:
+                        new_dep_node.add_dependent(
+                            elem_type.__name__
+                            if hasattr(elem_type, "__name__")
+                            else str(elem_type)
+                        )
+                        process_vmlinux_post_ast(
+                            elem_type, llvm_handler, handler, processing_stack
+                        )
+                        new_dep_node.set_field_ready(elem_name, True)
+                else:
+                    raise ValueError(
+                        f"{elem_name} with type {elem_type} from module {module_name} not supported in recursive resolver"
+                    )
+
+    else:
+        raise ImportError("UNSUPPORTED Module")
+
+    logging.info(
+        f"{current_symbol_name} processed and handler readiness {handler.is_ready}"
+    )
+    return True

pythonbpf/vmlinux_parser/dependency_handler.py

@@ -0,0 +1,173 @@
+from typing import Optional, Dict, List, Iterator
+from .dependency_node import DependencyNode
+
+
+class DependencyHandler:
+    """
+    Manages a collection of DependencyNode objects with no duplicates.
+
+    Ensures that no two nodes with the same name can be added and provides
+    methods to check readiness and retrieve specific nodes.
+
+    Example usage:
+        # Create a handler
+        handler = DependencyHandler()
+
+        # Create some dependency nodes
+        node1 = DependencyNode(name="node1")
+        node1.add_field("field1", str)
+        node1.set_field_value("field1", "value1")
+
+        node2 = DependencyNode(name="node2")
+        node2.add_field("field1", int)
+
+        # Add nodes to the handler
+        handler.add_node(node1)
+        handler.add_node(node2)
+
+        # Check if a specific node exists
+        print(handler.has_node("node1"))  # True
+
+        # Get a reference to a node and modify it
+        node = handler.get_node("node2")
+        node.set_field_value("field1", 42)
+
+        # Check if all nodes are ready
+        print(handler.is_ready)  # False (node2 is ready, but node1 isn't)
+    """
+
+    def __init__(self):
+        # Using a dictionary with node names as keys ensures name uniqueness
+        # and provides efficient lookups
+        self._nodes: Dict[str, DependencyNode] = {}
+
+    def add_node(self, node: DependencyNode) -> bool:
+        """
+        Add a dependency node to the handler.
+
+        Args:
+            node: The DependencyNode to add
+
+        Returns:
+            bool: True if the node was added, False if a node with the same name already exists
+
+        Raises:
+            TypeError: If the provided object is not a DependencyNode
+        """
+        if not isinstance(node, DependencyNode):
+            raise TypeError(f"Expected DependencyNode, got {type(node).__name__}")
+
+        # Check if a node with this name already exists
+        if node.name in self._nodes:
+            return False
+
+        self._nodes[node.name] = node
+        return True
+
+    @property
+    def is_ready(self) -> bool:
+        """
+        Check if all nodes are ready.
+
+        Returns:
+            bool: True if all nodes are ready (or if there are no nodes), False otherwise
+        """
+        if not self._nodes:
+            return True
+
+        return all(node.is_ready for node in self._nodes.values())
+
+    def has_node(self, name: str) -> bool:
+        """
+        Check if a node with the given name exists.
+
+        Args:
+            name: The name to check
+
+        Returns:
+            bool: True if a node with the given name exists, False otherwise
+        """
+        return name in self._nodes
+
+    def get_node(self, name: str) -> Optional[DependencyNode]:
+        """
+        Get a node by name for manipulation.
+
+        Args:
+            name: The name of the node to retrieve
+
+        Returns:
+            Optional[DependencyNode]: The node with the given name, or None if not found
+        """
+        return self._nodes.get(name)
+
+    def remove_node(self, node_or_name) -> bool:
+        """
+        Remove a node by name or reference.
+
+        Args:
+            node_or_name: The node to remove or its name
+
+        Returns:
+            bool: True if the node was removed, False if not found
+        """
+        if isinstance(node_or_name, DependencyNode):
+            name = node_or_name.name
+        else:
+            name = node_or_name
+
+        if name in self._nodes:
+            del self._nodes[name]
+            return True
+        return False
+
+    def get_all_nodes(self) -> List[DependencyNode]:
+        """
+        Get all nodes stored in the handler.
+
+        Returns:
+            List[DependencyNode]: List of all nodes
+        """
+        return list(self._nodes.values())
+
+    def __iter__(self) -> Iterator[DependencyNode]:
+        """
+        Iterate over all nodes.
+
+        Returns:
+            Iterator[DependencyNode]: Iterator over all nodes
+        """
+        return iter(self._nodes.values())
+
+    def __len__(self) -> int:
+        """
+        Get the number of nodes in the handler.
+
+        Returns:
+            int: The number of nodes
+        """
+        return len(self._nodes)
+
+    def __getitem__(self, name: str) -> DependencyNode:
+        """
+        Get a node by name using dictionary-style access.
+
+        Args:
+            name: The name of the node to retrieve
+
+        Returns:
+            DependencyNode: The node with the given name
+
+        Raises:
+            KeyError: If no node with the given name exists
+
+        Example:
+            node = handler["some-dep_node_name"]
+        """
+        if name not in self._nodes:
+            raise KeyError(f"No node with name '{name}' found")
+        return self._nodes[name]
+
+    @property
+    def nodes(self):
+        return self._nodes

pythonbpf/vmlinux_parser/dependency_node.py

@@ -0,0 +1,363 @@
+from dataclasses import dataclass, field
+from typing import Dict, Any, Optional
+import ctypes
+
+
+# TODO: FIX THE FUCKING TYPE NAME CONVENTION.
+@dataclass
+class Field:
+    """Represents a field in a dependency node with its type and readiness state."""
+
+    name: str
+    type: type
+    ctype_complex_type: Optional[Any]
+    containing_type: Optional[Any]
+    type_size: Optional[int]
+    bitfield_size: Optional[int]
+    offset: int
+    value: Any = None
+    ready: bool = False
+
+    def set_ready(self, is_ready: bool = True) -> None:
+        """Set the readiness state of this field."""
+        self.ready = is_ready
+
+    def set_value(self, value: Any, mark_ready: bool = False) -> None:
+        """Set the value of this field and optionally mark it as ready."""
+        self.value = value
+        if mark_ready:
+            self.ready = True
+
+    def set_type(self, given_type, mark_ready: bool = False) -> None:
+        """Set value of the type field and mark as ready"""
+        self.type = given_type
+        if mark_ready:
+            self.ready = True
+
+    def set_containing_type(
+        self, containing_type: Optional[Any], mark_ready: bool = False
+    ) -> None:
+        """Set the containing_type of this field and optionally mark it as ready."""
+        self.containing_type = containing_type
+        if mark_ready:
+            self.ready = True
+
+    def set_type_size(self, type_size: Any, mark_ready: bool = False) -> None:
+        """Set the type_size of this field and optionally mark it as ready."""
+        self.type_size = type_size
+        if mark_ready:
+            self.ready = True
+
+    def set_ctype_complex_type(
+        self, ctype_complex_type: Any, mark_ready: bool = False
+    ) -> None:
+        """Set the ctype_complex_type of this field and optionally mark it as ready."""
+        self.ctype_complex_type = ctype_complex_type
+        if mark_ready:
+            self.ready = True
+
+    def set_bitfield_size(self, bitfield_size: Any, mark_ready: bool = False) -> None:
+        """Set the bitfield_size of this field and optionally mark it as ready."""
+        self.bitfield_size = bitfield_size
+        if mark_ready:
+            self.ready = True
+
+    def set_offset(self, offset: int) -> None:
+        """Set the offset of this field"""
+        self.offset = offset
+
+
+@dataclass
+class DependencyNode:
+    """
+    A node with typed fields and readiness tracking.
+
+    Example usage:
+        # Create a dependency node for a Person
+        somestruct = DependencyNode(name="struct_1")
+
+        # Add fields with their types
+        somestruct.add_field("field_1", str)
+        somestruct.add_field("field_2", int)
+        somestruct.add_field("field_3", str)
+
+        # Check if the node is ready (should be False initially)
+        print(f"Is node ready? {somestruct.is_ready}")  # False
+
+        # Set some field values
+        somestruct.set_field_value("field_1", "someproperty")
+        somestruct.set_field_value("field_2", 30)
+
+        # Check if the node is ready (still False because email is not ready)
+        print(f"Is node ready? {somestruct.is_ready}")  # False
+
+        # Set the last field and make the node ready
+        somestruct.set_field_value("field_3", "anotherproperty")
+
+        # Now the node should be ready
+        print(f"Is node ready? {somestruct.is_ready}")  # True
+
+        # You can also mark a field as not ready
+        somestruct.set_field_ready("field_3", False)
+
+        # Now the node is not ready again
+        print(f"Is node ready? {somestruct.is_ready}")  # False
+
+        # Get all field values
+        print(somestruct.get_field_values())  # {'field_1': 'someproperty', 'field_2': 30, 'field_3': 'anotherproperty'}
+
+        # Get only ready fields
+        ready_fields = somestruct.get_ready_fields()
+        print(f"Ready fields: {[field.name for field in ready_fields.values()]}")  # ['field_1', 'field_2']
+    """
+
+    name: str
+    depends_on: Optional[list[str]] = None
+    fields: Dict[str, Field] = field(default_factory=dict)
+    _ready_cache: Optional[bool] = field(default=None, repr=False)
+    current_offset: int = 0
+    ctype_struct: Optional[Any] = field(default=None, repr=False)
+
+    def add_field(
+        self,
+        name: str,
+        field_type: type,
+        initial_value: Any = None,
+        containing_type: Optional[Any] = None,
+        type_size: Optional[int] = None,
+        ctype_complex_type: Optional[int] = None,
+        bitfield_size: Optional[int] = None,
+        ready: bool = False,
+        offset: int = 0,
+    ) -> None:
+        """Add a field to the node with an optional initial value and readiness state."""
+        if self.depends_on is None:
+            self.depends_on = []
+        self.fields[name] = Field(
+            name=name,
+            type=field_type,
+            value=initial_value,
+            ready=ready,
+            containing_type=containing_type,
+            type_size=type_size,
+            ctype_complex_type=ctype_complex_type,
+            bitfield_size=bitfield_size,
+            offset=offset,
+        )
+        # Invalidate readiness cache
+        self._ready_cache = None
+
+    def set_ctype_struct(self, ctype_struct: Any) -> None:
+        """Set the ctypes structure for automatic offset calculation."""
+        self.ctype_struct = ctype_struct
+
+    def __sizeof__(self):
+        # If we have a ctype_struct, use its size
+        if self.ctype_struct is not None:
+            return ctypes.sizeof(self.ctype_struct)
+        return self.current_offset
+
+    def get_field(self, name: str) -> Field:
+        """Get a field by name."""
+        return self.fields[name]
+
+    def set_field_value(self, name: str, value: Any, mark_ready: bool = False) -> None:
+        """Set a field's value and optionally mark it as ready."""
+        if name not in self.fields:
+            raise KeyError(f"Field '{name}' does not exist in node '{self.name}'")
+
+        self.fields[name].set_value(value, mark_ready)
+        # Invalidate readiness cache
+        self._ready_cache = None
+
+    def set_field_type(self, name: str, type: Any, mark_ready: bool = False) -> None:
+        """Set a field's type and optionally mark it as ready."""
+        if name not in self.fields:
+            raise KeyError(f"Field '{name}' does not exist in node '{self.name}'")
+
+        self.fields[name].set_type(type, mark_ready)
+        # Invalidate readiness cache
+        self._ready_cache = None
+
+    def set_field_containing_type(
+        self, name: str, containing_type: Any, mark_ready: bool = False
+    ) -> None:
+        """Set a field's containing_type and optionally mark it as ready."""
+        if name not in self.fields:
+            raise KeyError(f"Field '{name}' does not exist in node '{self.name}'")
+
+        self.fields[name].set_containing_type(containing_type, mark_ready)
+        # Invalidate readiness cache
+        self._ready_cache = None
+
+    def set_field_type_size(
+        self, name: str, type_size: Any, mark_ready: bool = False
+    ) -> None:
+        """Set a field's type_size and optionally mark it as ready."""
+        if name not in self.fields:
+            raise KeyError(f"Field '{name}' does not exist in node '{self.name}'")
+
+        self.fields[name].set_type_size(type_size, mark_ready)
+        # Invalidate readiness cache
+        self._ready_cache = None
+
+    def set_field_ctype_complex_type(
+        self, name: str, ctype_complex_type: Any, mark_ready: bool = False
+    ) -> None:
+        """Set a field's ctype_complex_type and optionally mark it as ready."""
+        if name not in self.fields:
+            raise KeyError(f"Field '{name}' does not exist in node '{self.name}'")
+
+        self.fields[name].set_ctype_complex_type(ctype_complex_type, mark_ready)
+        # Invalidate readiness cache
+        self._ready_cache = None
+
+    def set_field_bitfield_size(
+        self, name: str, bitfield_size: Any, mark_ready: bool = False
+    ) -> None:
+        """Set a field's bitfield_size and optionally mark it as ready."""
+        if name not in self.fields:
+            raise KeyError(f"Field '{name}' does not exist in node '{self.name}'")
+
+        self.fields[name].set_bitfield_size(bitfield_size, mark_ready)
+        # Invalidate readiness cache
+        self._ready_cache = None
+
+    def set_field_ready(
+        self,
+        name: str,
+        is_ready: bool = False,
+        size_of_containing_type: Optional[int] = None,
+    ) -> None:
+        """Mark a field as ready or not ready."""
+        if name not in self.fields:
+            raise KeyError(f"Field '{name}' does not exist in node '{self.name}'")
+
+        self.fields[name].set_ready(is_ready)
+
+        # Use ctypes built-in offset if available
+        if self.ctype_struct is not None:
+            try:
+                self.fields[name].set_offset(getattr(self.ctype_struct, name).offset)
+            except AttributeError:
+                # Fallback to manual calculation if field not found in ctype_struct
+                self.fields[name].set_offset(self.current_offset)
+                self.current_offset += self._calculate_size(
+                    name, size_of_containing_type
+                )
+        else:
+            # Manual offset calculation when no ctype_struct is available
+            self.fields[name].set_offset(self.current_offset)
+            self.current_offset += self._calculate_size(name, size_of_containing_type)
+
+        # Invalidate readiness cache
+        self._ready_cache = None
+
+    def _calculate_size(
+        self, name: str, size_of_containing_type: Optional[int] = None
+    ) -> int:
+        processing_field = self.fields[name]
+        # size_of_field will be in bytes
+        if processing_field.type.__module__ == ctypes.__name__:
+            size_of_field = ctypes.sizeof(processing_field.type)
+            return size_of_field
+        elif processing_field.type.__module__ == "vmlinux":
+            if processing_field.ctype_complex_type is not None:
+                if issubclass(processing_field.ctype_complex_type, ctypes.Array):
+                    if processing_field.containing_type.__module__ == ctypes.__name__:
+                        if (
+                            processing_field.containing_type is not None
+                            and processing_field.type_size is not None
+                        ):
+                            size_of_field = (
+                                ctypes.sizeof(processing_field.containing_type)
+                                * processing_field.type_size
+                            )
+                        else:
+                            raise RuntimeError(
+                                f"{processing_field} has no containing_type or type_size"
+                            )
+                        return size_of_field
+                    elif processing_field.containing_type.__module__ == "vmlinux":
+                        if (
+                            size_of_containing_type is not None
+                            and processing_field.type_size is not None
+                        ):
+                            size_of_field = (
+                                size_of_containing_type * processing_field.type_size
+                            )
+                        else:
+                            raise RuntimeError(
+                                f"{processing_field} has no containing_type or type_size"
+                            )
+                        return size_of_field
+                elif issubclass(processing_field.ctype_complex_type, ctypes._Pointer):
+                    return ctypes.sizeof(ctypes.c_void_p)
+                else:
+                    raise NotImplementedError(
+                        "This subclass of ctype not supported yet"
+                    )
+            elif processing_field.type_size is not None:
+                # Handle vmlinux types with type_size but no ctype_complex_type
+                # This means it's a direct vmlinux struct field (not array/pointer wrapped)
+                # The type_size should already contain the full size of the struct
+                # But if there's a containing_type from vmlinux, we need that size
+                if processing_field.containing_type is not None:
+                    if processing_field.containing_type.__module__ == "vmlinux":
+                        # For vmlinux containing types, we need the pre-calculated size
+                        if size_of_containing_type is not None:
+                            return size_of_containing_type * processing_field.type_size
+                        else:
+                            raise RuntimeError(
+                                f"Field {name}: vmlinux containing_type requires size_of_containing_type"
+                            )
+                    else:
+                        raise ModuleNotFoundError(
+                            f"Containing type module {processing_field.containing_type.__module__} not supported"
+                        )
+                else:
+                    raise RuntimeError("Wrong type found with no containing type")
+            else:
+                # No ctype_complex_type and no type_size, must rely on size_of_containing_type
+                if size_of_containing_type is None:
+                    raise RuntimeError(
+                        f"Size of containing type {size_of_containing_type} is None"
+                    )
+                return size_of_containing_type
+
+        else:
+            raise ModuleNotFoundError("Module is not supported for the operation")
+        raise RuntimeError("control should not reach here")
+
+    @property
+    def is_ready(self) -> bool:
+        """Check if the node is ready (all fields are ready)."""
+        # Use cached value if available
+        if self._ready_cache is not None:
+            return self._ready_cache
+
+        # Calculate readiness only when needed
+        if not self.fields:
+            self._ready_cache = True
+            return True
+
+        self._ready_cache = all(elem.ready for elem in self.fields.values())
+        return self._ready_cache
+
+    def get_field_values(self) -> Dict[str, Any]:
+        """Get a dictionary of field names to their values."""
+        return {name: elem.value for name, elem in self.fields.items()}
+
+    def get_ready_fields(self) -> Dict[str, Field]:
+        """Get all fields that are marked as ready."""
+        return {name: elem for name, elem in self.fields.items() if elem.ready}
+
+    def get_not_ready_fields(self) -> Dict[str, Field]:
+        """Get all fields that are marked as not ready."""
+        return {name: elem for name, elem in self.fields.items() if not elem.ready}
+
+    def add_dependent(self, dep_type):
+        if dep_type in self.depends_on:
+            return
+        else:
+            self.depends_on.append(dep_type)

pythonbpf/vmlinux_parser/import_detector.py

@@ -0,0 +1,147 @@
+import ast
+import logging
+from typing import List, Tuple, Any
+import importlib
+import inspect
+
+from .dependency_handler import DependencyHandler
+from .ir_gen import IRGenerator
+from .class_handler import process_vmlinux_class
+
+logger = logging.getLogger(__name__)
+
+
+def detect_import_statement(tree: ast.AST) -> List[Tuple[str, ast.ImportFrom]]:
+    """
+    Parse AST and detect import statements from vmlinux.
+
+    Returns a list of tuples (module_name, imported_item) for vmlinux imports.
+    Raises SyntaxError for invalid import patterns.
+
+    Args:
+        tree: The AST to parse
+
+    Returns:
+        List of tuples containing (module_name, imported_item) for each vmlinux import
+
+    Raises:
+        SyntaxError: If multiple imports from vmlinux are attempted or import * is used
+    """
+    vmlinux_imports = []
+
+    for node in ast.walk(tree):
+        # Handle "from vmlinux import ..." statements
+        if isinstance(node, ast.ImportFrom):
+            if node.module == "vmlinux":
+                # Check for wildcard import: from vmlinux import *
+                if any(alias.name == "*" for alias in node.names):
+                    raise SyntaxError(
+                        "Wildcard imports from vmlinux are not supported. "
+                        "Please import specific types explicitly."
+                    )
+
+                # Check for multiple imports: from vmlinux import A, B, C
+                if len(node.names) > 1:
+                    imported_names = [alias.name for alias in node.names]
+                    raise SyntaxError(
+                        f"Multiple imports from vmlinux are not supported. "
+                        f"Found: {', '.join(imported_names)}. "
+                        f"Please use separate import statements for each type."
+                    )
+
+                # Check if no specific import is specified (should not happen with valid Python)
+                if len(node.names) == 0:
+                    raise SyntaxError(
+                        "Import from vmlinux must specify at least one type."
+                    )
+
+                # Valid single import
+                for alias in node.names:
+                    import_name = alias.name
+                    # Use alias if provided, otherwise use the original name (commented)
+                    # as_name = alias.asname if alias.asname else alias.name
+                    vmlinux_imports.append(("vmlinux", node))
+                    logger.info(f"Found vmlinux import: {import_name}")
+
+        # Handle "import vmlinux" statements (not typical but should be rejected)
+        elif isinstance(node, ast.Import):
+            for alias in node.names:
+                if alias.name == "vmlinux" or alias.name.startswith("vmlinux."):
+                    raise SyntaxError(
+                        "Direct import of vmlinux module is not supported. "
+                        "Use 'from vmlinux import <type>' instead."
+                    )
+
+    logger.info(f"Total vmlinux imports detected: {len(vmlinux_imports)}")
+    return vmlinux_imports
+
+
+def vmlinux_proc(tree: ast.AST, module):
+    import_statements = detect_import_statement(tree)
+
+    # initialise dependency handler
+    handler = DependencyHandler()
+    # initialise assignment dictionary of name to type
+    assignments: dict[str, tuple[type, Any]] = {}
+
+    if not import_statements:
+        logger.info("No vmlinux imports found")
+        return
+
+    # Import vmlinux module directly
+    try:
+        vmlinux_mod = importlib.import_module("vmlinux")
+    except ImportError:
+        logger.warning("Could not import vmlinux module")
+        return
+
+    source_file = inspect.getsourcefile(vmlinux_mod)
+    if source_file is None:
+        logger.warning("Cannot find source for vmlinux module")
+        return
+
+    with open(source_file, "r") as f:
+        mod_ast = ast.parse(f.read(), filename=source_file)
+
+    for import_mod, import_node in import_statements:
+        for alias in import_node.names:
+            imported_name = alias.name
+            found = False
+            for mod_node in mod_ast.body:
+                if (
+                    isinstance(mod_node, ast.ClassDef)
+                    and mod_node.name == imported_name
+                ):
+                    process_vmlinux_class(mod_node, module, handler)
+                    found = True
+                    break
+                if isinstance(mod_node, ast.Assign):
+                    for target in mod_node.targets:
+                        if isinstance(target, ast.Name) and target.id == imported_name:
+                            process_vmlinux_assign(mod_node, module, assignments)
+                            found = True
+                            break
+                if found:
+                    break
+            if not found:
+                logger.info(
+                    f"{imported_name} not found as ClassDef or Assign in vmlinux"
+                )
+
+    IRGenerator(module, handler)
+    return assignments
+
+
+def process_vmlinux_assign(node, module, assignments: dict[str, tuple[type, Any]]):
+    # Check if this is a simple assignment with a constant value
+    if len(node.targets) == 1 and isinstance(node.targets[0], ast.Name):
+        target_name = node.targets[0].id
+        if isinstance(node.value, ast.Constant):
+            assignments[target_name] = (type(node.value.value), node.value.value)
+            logger.info(
+                f"Added assignment: {target_name} = {node.value.value!r} of type {type(node.value.value)}"
+            )
+        else:
+            raise ValueError(f"Unsupported assignment type for {target_name}")
+    else:
+        raise ValueError("Not a simple assignment")

pythonbpf/vmlinux_parser/ir_gen/__init__.py

@@ -0,0 +1,3 @@
+from .ir_generation import IRGenerator
+
+__all__ = ["IRGenerator"]

pythonbpf/vmlinux_parser/ir_gen/debug_info_gen.py

@@ -0,0 +1,15 @@
+from pythonbpf.debuginfo import DebugInfoGenerator
+
+
+def debug_info_generation(struct, llvm_module):
+    generator = DebugInfoGenerator(llvm_module)
+    # this is sample debug info generation
+    # i64type = generator.get_uint64_type()
+
+    struct_type = generator.create_struct_type([], 64 * 4, is_distinct=True)
+
+    global_var = generator.create_global_var_debug_info(
+        struct.name, struct_type, is_local=False
+    )
+
+    return global_var

pythonbpf/vmlinux_parser/ir_gen/ir_generation.py

@@ -0,0 +1,161 @@
+import ctypes
+import logging
+from ..dependency_handler import DependencyHandler
+from .debug_info_gen import debug_info_generation
+from ..dependency_node import DependencyNode
+import llvmlite.ir as ir
+
+logger = logging.getLogger(__name__)
+
+
+class IRGenerator:
+    # get the assignments dict and add this stuff to it.
+    def __init__(self, llvm_module, handler: DependencyHandler, assignment=None):
+        self.llvm_module = llvm_module
+        self.handler: DependencyHandler = handler
+        self.generated: list[str] = []
+        if not handler.is_ready:
+            raise ImportError(
+                "Semantic analysis of vmlinux imports failed. Cannot generate IR"
+            )
+        for struct in handler:
+            self.struct_processor(struct)
+
+    def struct_processor(self, struct, processing_stack=None):
+        # Initialize processing stack on first call
+        if processing_stack is None:
+            processing_stack = set()
+
+        # If already generated, skip
+        if struct.name in self.generated:
+            return
+
+        # Detect circular dependency
+        if struct.name in processing_stack:
+            logger.info(
+                f"Circular dependency detected for {struct.name}, skipping recursive processing"
+            )
+            # For circular dependencies, we can either:
+            # 1. Use forward declarations (opaque pointers)
+            # 2. Mark as incomplete and process later
+            # 3. Generate a placeholder type
+            # Here we'll just skip and let it be processed in its own call
+            return
+
+        logger.info(f"IR generating for {struct.name}")
+
+        # Add to processing stack before processing dependencies
+        processing_stack.add(struct.name)
+
+        try:
+            # Process all dependencies first
+            if struct.depends_on is None:
+                pass
+            else:
+                for dependency in struct.depends_on:
+                    if dependency not in self.generated:
+                        # Check if dependency exists in handler
+                        if dependency in self.handler.nodes:
+                            dep_node_from_dependency = self.handler[dependency]
+                            # Pass the processing_stack down to track circular refs
+                            self.struct_processor(
+                                dep_node_from_dependency, processing_stack
+                            )
+                        else:
+                            raise RuntimeError(
+                                f"Warning: Dependency {dependency} not found in handler"
+                            )
+
+            # Actual processor logic here after dependencies are resolved
+            self.gen_ir(struct)
+            self.generated.append(struct.name)
+
+        finally:
+            # Remove from processing stack after we're done
+            processing_stack.discard(struct.name)
+
+    def gen_ir(self, struct):
+        # TODO: we add the btf_ama attribute by monkey patching in the end of compilation, but once llvmlite
+        #  accepts our issue, we will resort to normal accessed attribute based attribute addition
+        # currently we generate all possible field accesses for CO-RE and put into the assignment table
+        debug_info = debug_info_generation(struct, self.llvm_module)
+        field_index = 0
+        for field_name, field in struct.fields.items():
+            # does not take arrays and similar types into consideration yet.
+            if field.ctype_complex_type is not None and issubclass(
+                field.ctype_complex_type, ctypes.Array
+            ):
+                array_size = field.type_size
+                containing_type = field.containing_type
+                if containing_type.__module__ == ctypes.__name__:
+                    containing_type_size = ctypes.sizeof(containing_type)
+                    for i in range(0, array_size):
+                        field_co_re_name = self._struct_name_generator(
+                            struct, field, field_index, True, i, containing_type_size
+                        )
+                        globvar = ir.GlobalVariable(
+                            self.llvm_module, ir.IntType(64), name=field_co_re_name
+                        )
+                        globvar.linkage = "external"
+                        globvar.set_metadata("llvm.preserve.access.index", debug_info)
+                    field_index += 1
+            elif field.type_size is not None:
+                array_size = field.type_size
+                containing_type = field.containing_type
+                if containing_type.__module__ == "vmlinux":
+                    containing_type_size = self.handler[
+                        containing_type.__name__
+                    ].current_offset
+                    for i in range(0, array_size):
+                        field_co_re_name = self._struct_name_generator(
+                            struct, field, field_index, True, i, containing_type_size
+                        )
+                        globvar = ir.GlobalVariable(
+                            self.llvm_module, ir.IntType(64), name=field_co_re_name
+                        )
+                        globvar.linkage = "external"
+                        globvar.set_metadata("llvm.preserve.access.index", debug_info)
+                    field_index += 1
+            else:
+                field_co_re_name = self._struct_name_generator(
+                    struct, field, field_index
+                )
+                field_index += 1
+                globvar = ir.GlobalVariable(
+                    self.llvm_module, ir.IntType(64), name=field_co_re_name
+                )
+                globvar.linkage = "external"
+                globvar.set_metadata("llvm.preserve.access.index", debug_info)
+
+    def _struct_name_generator(
+        self,
+        struct: DependencyNode,
+        field,
+        field_index: int,
+        is_indexed: bool = False,
+        index: int = 0,
+        containing_type_size: int = 0,
+    ) -> str:
+        if is_indexed:
+            name = (
+                "llvm."
+                + struct.name.removeprefix("struct_")
+                + f":0:{field.offset + index * containing_type_size}"
+                + "$"
+                + f"0:{field_index}:{index}"
+            )
+            return name
+        elif struct.name.startswith("struct_"):
+            name = (
+                "llvm."
+                + struct.name.removeprefix("struct_")
+                + f":0:{field.offset}"
+                + "$"
+                + f"0:{field_index}"
+            )
+            return name
+        else:
+            print(self.handler[struct.name])
+            raise TypeError(
+                "Name generation cannot occur due to type name not starting with struct"
+            )

tests/c-form/ex2.bpf.c

@@ -1,11 +1,10 @@
-#include <linux/bpf.h>
+#include "vmlinux.h"
 #include <bpf/bpf_helpers.h>
-#define u64 unsigned long long
-#define u32 unsigned int
+#include <bpf/bpf_endian.h>
 
 SEC("xdp")
 int hello(struct xdp_md *ctx) {
-    bpf_printk("Hello, World!\n");
+    bpf_printk("Hello, World! %ud \n", ctx->data);
     return XDP_PASS;
 }
 

tests/c-form/ex7.bpf.c

@@ -1,23 +1,9 @@
 // SPDX-License-Identifier: GPL-2.0
 
-#include <linux/bpf.h>
+#include "vmlinux.h"
 #include <bpf/bpf_helpers.h>
 #include <bpf/bpf_tracing.h>
 
-struct trace_entry {
-  short unsigned int type;
-  unsigned char flags;
-  unsigned char preempt_count;
-  int pid;
-};
-
-struct trace_event_raw_sys_enter {
-  struct trace_entry ent;
-  long int id;
-  long unsigned int args[6];
-  char __data[0];
-};
-
 struct event {
   __u32 pid;
   __u32 uid;

tests/failing_tests/assign/retype.py

@@ -11,6 +11,7 @@ from pythonbpf.maps import HashMap
 # We can allow bitcasts in cases where the width of the types is the same in
 # the future. But for now, we do not allow any re-interpretation of variables.
 
+
 @bpf
 @map
 def last() -> HashMap:

tests/failing_tests/globals.py

@@ -3,16 +3,19 @@ import logging
 from pythonbpf import compile, bpf, section, bpfglobal, compile_to_ir
 from ctypes import c_void_p, c_int64, c_int32
 
+
 @bpf
 @bpfglobal
 def somevalue() -> c_int32:
     return c_int32(42)
 
+
 @bpf
 @bpfglobal
 def somevalue2() -> c_int64:
     return c_int64(69)
 
+
 @bpf
 @bpfglobal
 def somevalue1() -> c_int32:
@@ -21,12 +24,14 @@ def somevalue1() -> c_int32:
 
 # --- Passing examples ---
 
+
 # Simple constant return
 @bpf
 @bpfglobal
 def g1() -> c_int64:
     return c_int64(42)
 
+
 # Constructor with one constant argument
 @bpf
 @bpfglobal
@@ -62,15 +67,17 @@ def g2() -> c_int64:
 # def g6() -> c_int64:
 #     return c_int64(CONST)
 
+
 # Constructor with multiple args
-#TODO: this is not working. should it work ?
+# TODO: this is not working. should it work ?
 @bpf
 @bpfglobal
 def g7() -> c_int64:
     return c_int64(1)
 
+
 # Dataclass call
-#TODO: fails with dataclass
+# TODO: fails with dataclass
 # @dataclass
 # class Point:
 #     x: c_int64
@@ -91,6 +98,7 @@ def sometag(ctx: c_void_p) -> c_int64:
     print(f"{somevalue}")
     return c_int64(1)
 
+
 @bpf
 @bpfglobal
 def LICENSE() -> str:

tests/failing_tests/named_arg.py

@@ -11,6 +11,7 @@ from ctypes import c_void_p, c_int64
 # We cannot allocate space for the intermediate type now.
 # We probably need to track the ref/deref chain for each variable.
 
+
 @bpf
 @map
 def count() -> HashMap:

tests/failing_tests/undeclared_values.py

@@ -3,6 +3,7 @@ import logging
 from pythonbpf import compile, bpf, section, bpfglobal, compile_to_ir
 from ctypes import c_void_p, c_int64
 
+
 # This should not pass as somevalue is not declared at all.
 @bpf
 @section("tracepoint/syscalls/sys_enter_execve")
@@ -11,6 +12,7 @@ def sometag(ctx: c_void_p) -> c_int64:
     print(f"{somevalue}")  # noqa: F821
     return c_int64(1)
 
+
 @bpf
 @bpfglobal
 def LICENSE() -> str:

tests/failing_tests/xdp_pass.py

@@ -0,0 +1,54 @@
+from pythonbpf import bpf, map, section, bpfglobal, compile_to_ir
+from pythonbpf.maps import HashMap
+from pythonbpf.helper import XDP_PASS
+from vmlinux import TASK_COMM_LEN  # noqa: F401
+
+from vmlinux import struct_qspinlock  # noqa: F401
+
+# from vmlinux import struct_trace_event_raw_sys_enter  # noqa: F401
+# from vmlinux import struct_posix_cputimers  # noqa: F401
+from vmlinux import struct_xdp_md
+
+# from vmlinux import struct_trace_event_raw_sys_enter  # noqa: F401
+# from vmlinux import struct_ring_buffer_per_cpu  # noqa: F401
+# from vmlinux import struct_request  # noqa: F401
+from ctypes import c_int64
+
+# Instructions to how to run this program
+# 1. Install PythonBPF: pip install pythonbpf
+# 2. Run the program: python examples/xdp_pass.py
+# 3. Run the program with sudo: sudo tools/check.sh run examples/xdp_pass.o
+# 4. Attach object file to any network device with something like ./check.sh xdp examples/xdp_pass.o tailscale0
+# 5. send traffic through the device and observe effects
+
+
+@bpf
+@map
+def count() -> HashMap:
+    return HashMap(key=c_int64, value=c_int64, max_entries=1)
+
+
+@bpf
+@section("xdp")
+def hello_world(ctx: struct_xdp_md) -> c_int64:
+    key = 0
+    one = 1
+    prev = count().lookup(key)
+    if prev:
+        prevval = prev + 1
+        print(f"count: {prevval}")
+        count().update(key, prevval)
+        return XDP_PASS
+    else:
+        count().update(key, one)
+
+    return XDP_PASS
+
+
+@bpf
+@bpfglobal
+def LICENSE() -> str:
+    return "GPL"
+
+
+compile_to_ir("xdp_pass.py", "xdp_pass.ll")

tests/passing_tests/assign/comprehensive.py

@@ -9,6 +9,7 @@ from pythonbpf.helper import ktime
 # the compiler, it is recommended to use named variables to reduce the amount of
 # scratch space that needs to be allocated.
 
+
 @bpf
 @struct
 class data_t:
@@ -48,8 +49,12 @@ def hello_world(ctx: c_void_p) -> c_int64:
         a = last.lookup(0)
         print(f"a is {a}")
         last.update(9, 9)
-        last.update(0, last.lookup(last.lookup(0)) +
-                    last.lookup(last.lookup(0)) + last.lookup(last.lookup(0)))
+        last.update(
+            0,
+            last.lookup(last.lookup(0))
+            + last.lookup(last.lookup(0))
+            + last.lookup(last.lookup(0)),
+        )
         z = last.lookup(0)
         print(f"new map val at index 0 is {z}")
     else:

tests/passing_tests/return/typecast_var.py

@@ -6,8 +6,8 @@ from ctypes import c_void_p, c_int32
 @section("tracepoint/syscalls/sys_enter_execve")
 def hello_world(ctx: c_void_p) -> c_int32:
     print("Hello, World!")
-    a = 1 # int64
-    return c_int32(a) # typecast to int32
+    a = 1  # int64
+    return c_int32(a)  # typecast to int32
 
 
 @bpf

tools/vmlinux-gen.py

@@ -26,8 +26,13 @@ import tempfile
 
 
 class BTFConverter:
-    def __init__(self, btf_source="/sys/kernel/btf/vmlinux", output_file="vmlinux.py",
-                 keep_intermediate=False, verbose=False):
+    def __init__(
+        self,
+        btf_source="/sys/kernel/btf/vmlinux",
+        output_file="vmlinux.py",
+        keep_intermediate=False,
+        verbose=False,
+    ):
         self.btf_source = btf_source
         self.output_file = output_file
         self.keep_intermediate = keep_intermediate
@@ -44,11 +49,7 @@ class BTFConverter:
         self.log(f"{description}...")
         try:
             result = subprocess.run(
-                cmd,
-                shell=True,
-                check=True,
-                capture_output=True,
-                text=True
+                cmd, shell=True, check=True, capture_output=True, text=True
             )
             if self.verbose and result.stdout:
                 print(result.stdout)
@@ -69,51 +70,55 @@ class BTFConverter:
         """Step 1.5: Preprocess enum definitions."""
         self.log("Preprocessing enum definitions...")
 
-        with open(input_file, 'r') as f:
+        with open(input_file, "r") as f:
             original_code = f.read()
 
         # Extract anonymous enums
         enums = re.findall(
-            r'(?<!typedef\s)(enum\s*\{[^}]*\})\s*(\w+)\s*(?::\s*\d+)?\s*;',
-            original_code
+            r"(?<!typedef\s)(enum\s*\{[^}]*\})\s*(\w+)\s*(?::\s*\d+)?\s*;",
+            original_code,
         )
-        enum_defs = [enum_block + ';' for enum_block, _ in enums]
+        enum_defs = [enum_block + ";" for enum_block, _ in enums]
 
         # Replace anonymous enums with int declarations
         processed_code = re.sub(
-            r'(?<!typedef\s)enum\s*\{[^}]*\}\s*(\w+)\s*(?::\s*\d+)?\s*;',
-            r'int \1;',
-            original_code
+            r"(?<!typedef\s)enum\s*\{[^}]*\}\s*(\w+)\s*(?::\s*\d+)?\s*;",
+            r"int \1;",
+            original_code,
         )
 
         # Prepend enum definitions
         if enum_defs:
-            enum_text = '\n'.join(enum_defs) + '\n\n'
+            enum_text = "\n".join(enum_defs) + "\n\n"
             processed_code = enum_text + processed_code
 
         output_file = os.path.join(self.temp_dir, "vmlinux_processed.h")
-        with open(output_file, 'w') as f:
+        with open(output_file, "w") as f:
             f.write(processed_code)
 
         return output_file
 
     def step2_5_process_kioctx(self, input_file):
-        #TODO: this is a very bad bug and design decision. A single struct has an issue mostly.
+        # TODO: this is a very bad bug and design decision. A single struct has an issue mostly.
         """Step 2.5: Process struct kioctx to extract nested anonymous structs."""
         self.log("Processing struct kioctx nested structs...")
 
-        with open(input_file, 'r') as f:
+        with open(input_file, "r") as f:
             content = f.read()
 
         # Pattern to match struct kioctx with its full body (handles multiple nesting levels)
-        kioctx_pattern = r'struct\s+kioctx\s*\{(?:[^{}]|\{(?:[^{}]|\{[^{}]*\})*\})*\}\s*;'
+        kioctx_pattern = (
+            r"struct\s+kioctx\s*\{(?:[^{}]|\{(?:[^{}]|\{[^{}]*\})*\})*\}\s*;"
+        )
 
         def process_kioctx_replacement(match):
             full_struct = match.group(0)
             self.log(f"Found struct kioctx, length: {len(full_struct)} chars")
 
             # Extract the struct body (everything between outermost { and })
-            body_match = re.search(r'struct\s+kioctx\s*\{(.*)\}\s*;', full_struct, re.DOTALL)
+            body_match = re.search(
+                r"struct\s+kioctx\s*\{(.*)\}\s*;", full_struct, re.DOTALL
+            )
             if not body_match:
                 return full_struct
 
@@ -121,7 +126,7 @@ class BTFConverter:
 
             # Find all anonymous structs within the body
             # Pattern: struct { ... } followed by ; (not a member name)
-            anon_struct_pattern = r'struct\s*\{[^}]*\}'
+            # anon_struct_pattern = r"struct\s*\{[^}]*\}"
 
             anon_structs = []
             anon_counter = 4  # Start from 4, counting down to 1
@@ -131,7 +136,9 @@ class BTFConverter:
                 anon_struct_content = m.group(0)
 
                 # Extract the body of the anonymous struct
-                anon_body_match = re.search(r'struct\s*\{(.*)\}', anon_struct_content, re.DOTALL)
+                anon_body_match = re.search(
+                    r"struct\s*\{(.*)\}", anon_struct_content, re.DOTALL
+                )
                 if not anon_body_match:
                     return anon_struct_content
 
@@ -154,7 +161,7 @@ class BTFConverter:
             processed_body = body
 
             # Find all occurrences and process them
-            pattern_with_semicolon = r'struct\s*\{([^}]*)\}\s*;'
+            pattern_with_semicolon = r"struct\s*\{([^}]*)\}\s*;"
             matches = list(re.finditer(pattern_with_semicolon, body, re.DOTALL))
 
             if not matches:
@@ -178,14 +185,16 @@ class BTFConverter:
 
                 # Replace in the body
                 replacement = f"struct {anon_name} {member_name};"
-                processed_body = processed_body[:start_pos] + replacement + processed_body[end_pos:]
+                processed_body = (
+                    processed_body[:start_pos] + replacement + processed_body[end_pos:]
+                )
 
                 anon_counter -= 1
 
             # Rebuild the complete definition
             if anon_structs:
                 # Prepend the anonymous struct definitions
-                anon_definitions = '\n'.join(anon_structs) + '\n\n'
+                anon_definitions = "\n".join(anon_structs) + "\n\n"
                 new_struct = f"struct kioctx {{{processed_body}}};"
                 return anon_definitions + new_struct
             else:
@@ -193,14 +202,11 @@ class BTFConverter:
 
         # Apply the transformation
         processed_content = re.sub(
-            kioctx_pattern,
-            process_kioctx_replacement,
-            content,
-            flags=re.DOTALL
+            kioctx_pattern, process_kioctx_replacement, content, flags=re.DOTALL
         )
 
         output_file = os.path.join(self.temp_dir, "vmlinux_kioctx_processed.h")
-        with open(output_file, 'w') as f:
+        with open(output_file, "w") as f:
             f.write(processed_content)
 
         self.log(f"Saved kioctx-processed output to {output_file}")
@@ -218,7 +224,7 @@ class BTFConverter:
         output_file = os.path.join(self.temp_dir, "vmlinux_raw.py")
         cmd = (
             f"clang2py {input_file} -o {output_file} "
-            f"--clang-args=\"-fno-ms-extensions -I/usr/include -I/usr/include/linux\""
+            f'--clang-args="-fno-ms-extensions -I/usr/include -I/usr/include/linux"'
         )
         self.run_command(cmd, "Converting to Python ctypes")
         return output_file
@@ -234,14 +240,21 @@ class BTFConverter:
         data = re.sub(r"\('_[0-9]+',\s*ctypes\.[a-zA-Z0-9_]+,\s*0\),?\s*\n?", "", data)
 
         # Replace ('_20', ctypes.c_uint64, 64) → ('_20', ctypes.c_uint64)
-        data = re.sub(r"\('(_[0-9]+)',\s*(ctypes\.[a-zA-Z0-9_]+),\s*[0-9]+\)", r"('\1', \2)", data)
+        data = re.sub(
+            r"\('(_[0-9]+)',\s*(ctypes\.[a-zA-Z0-9_]+),\s*[0-9]+\)", r"('\1', \2)", data
+        )
 
         # Replace ('_20', ctypes.c_char, 8) with ('_20', ctypes.c_uint8, 8)
-        data = re.sub(
-            r"(ctypes\.c_char)(\s*,\s*\d+\))",
-            r"ctypes.c_uint8\2",
-            data
-        )
+        data = re.sub(r"(ctypes\.c_char)(\s*,\s*\d+\))", r"ctypes.c_uint8\2", data)
+
+        # below to replace those c_bool with bitfield greater than 8
+        def repl(m):
+            name, bits = m.groups()
+            return (
+                f"('{name}', ctypes.c_uint32, {bits})" if int(bits) > 8 else m.group(0)
+            )
+
+        data = re.sub(r"\('([^']+)',\s*ctypes\.c_bool,\s*(\d+)\)", repl, data)
 
         # Remove ctypes. prefix from invalid entries
         invalid_ctypes = ["bpf_iter_state", "_cache_type", "fs_context_purpose"]
@@ -258,6 +271,7 @@ class BTFConverter:
         if not self.keep_intermediate and self.temp_dir != ".":
             self.log(f"Cleaning up temporary directory: {self.temp_dir}")
             import shutil
+
             shutil.rmtree(self.temp_dir, ignore_errors=True)
 
     def convert(self):
@@ -281,6 +295,7 @@ class BTFConverter:
         except Exception as e:
             print(f"\n✗ Error during conversion: {e}", file=sys.stderr)
             import traceback
+
             traceback.print_exc()
             sys.exit(1)
         finally:
@@ -293,18 +308,13 @@ class BTFConverter:
         dependencies = {
             "bpftool": "bpftool --version",
             "clang": "clang --version",
-            "clang2py": "clang2py --version"
+            "clang2py": "clang2py --version",
         }
 
         missing = []
         for tool, cmd in dependencies.items():
             try:
-                subprocess.run(
-                    cmd,
-                    shell=True,
-                    check=True,
-                    capture_output=True
-                )
+                subprocess.run(cmd, shell=True, check=True, capture_output=True)
             except subprocess.CalledProcessError:
                 missing.append(tool)
 
@@ -326,31 +336,31 @@ Examples:
   %(prog)s
   %(prog)s -o kernel_types.py
   %(prog)s --btf-source /sys/kernel/btf/custom_module -k -v
-        """
+        """,
     )
 
     parser.add_argument(
         "--btf-source",
         default="/sys/kernel/btf/vmlinux",
-        help="Path to BTF source (default: /sys/kernel/btf/vmlinux)"
+        help="Path to BTF source (default: /sys/kernel/btf/vmlinux)",
     )
 
     parser.add_argument(
-        "-o", "--output",
+        "-o",
+        "--output",
         default="vmlinux.py",
-        help="Output Python file (default: vmlinux.py)"
+        help="Output Python file (default: vmlinux.py)",
     )
 
     parser.add_argument(
-        "-k", "--keep-intermediate",
+        "-k",
+        "--keep-intermediate",
         action="store_true",
-        help="Keep intermediate files (vmlinux.h, vmlinux_processed.h, etc.)"
+        help="Keep intermediate files (vmlinux.h, vmlinux_processed.h, etc.)",
     )
 
     parser.add_argument(
-        "-v", "--verbose",
-        action="store_true",
-        help="Enable verbose output"
+        "-v", "--verbose", action="store_true", help="Enable verbose output"
     )
 
     args = parser.parse_args()
@@ -359,7 +369,7 @@ Examples:
         btf_source=args.btf_source,
         output_file=args.output,
         keep_intermediate=args.keep_intermediate,
-        verbose=args.verbose
+        verbose=args.verbose,
     )
 
     converter.convert()