members generated for simple ctypes

Author: varun-r-mallya

pythonbpf/debuginfo/debug_info_generator.py

@@ -101,6 +101,21 @@ def create_struct_member(self, name: str, base_type: Any, offset: int) -> Any:
             },
         )
 
+    def create_struct_member_vmlinux(self, name: str, base_type_with_size: Any, offset: int) -> Any:
+        """Create a struct member with the given name, type, and offset"""
+        base_type, type_size = base_type_with_size
+        return self.module.add_debug_info(
+            "DIDerivedType",
+            {
+                "tag": dc.DW_TAG_member,
+                "name": name,
+                "file": self.module._file_metadata,
+                "baseType": base_type,
+                "size": getattr(base_type, "size", type_size),
+                "offset": offset,
+            },
+        )
+
     def create_struct_type(
         self, members: List[Any], size: int, is_distinct: bool
     ) -> Any:
@@ -116,6 +131,22 @@ def create_struct_type(
             is_distinct=is_distinct,
         )
 
+    def create_struct_type_with_name(
+        self, name: str, members: List[Any], size: int, is_distinct: bool
+    ) -> Any:
+        """Create a struct type with the given members and size"""
+        return self.module.add_debug_info(
+            "DICompositeType",
+            {
+                "name": name,
+                "tag": dc.DW_TAG_structure_type,
+                "file": self.module._file_metadata,
+                "size": size,
+                "elements": members,
+            },
+            is_distinct=is_distinct,
+        )
+
     def create_global_var_debug_info(
         self, name: str, var_type: Any, is_local: bool = False
     ) -> Any:

pythonbpf/vmlinux_parser/ir_gen/debug_info_gen.py

@@ -1,20 +1,155 @@
-from pythonbpf.debuginfo import DebugInfoGenerator
+from pythonbpf.debuginfo import DebugInfoGenerator, dwarf_constants as dc
 from ..dependency_node import DependencyNode
+import ctypes
+import logging
+from typing import List, Any, Tuple, Optional
 
+logger = logging.getLogger(__name__)
 
 def debug_info_generation(
-    struct: DependencyNode, llvm_module, generated_debug_info: list
-):
+        struct: DependencyNode, llvm_module, generated_debug_info: List[Tuple[DependencyNode, Any]]
+) -> Any:
+    """
+    Generate DWARF debug information for a struct defined in a DependencyNode.
+
+    Args:
+        struct: The dependency node containing struct information
+        llvm_module: The LLVM module to add debug info to
+        generated_debug_info: List of tuples (struct, debug_info) to track generated debug info
+
+    Returns:
+        The generated global variable debug info
+    """
+    # Set up debug info generator
     generator = DebugInfoGenerator(llvm_module)
+
+    # Check if debug info for this struct has already been generated
+    for existing_struct, debug_info in generated_debug_info:
+        if existing_struct.name == struct.name:
+            return debug_info
+
+    # Process all fields and create members for the struct
     members = []
-    uint32type = generator.get_uint32_type()
     for field_name, field in struct.fields.items():
-        members.append(generator.create_struct_member(field_name, uint32type, field.offset))
-
-    struct_type = generator.create_struct_type(members, struct.__sizeof__(), is_distinct=True)
+        # Get appropriate debug type for this field
+        field_type = _get_field_debug_type(
+            field_name, field, generator, struct, generated_debug_info
+        )
+        # Create struct member with proper offset
+        member = generator.create_struct_member_vmlinux(
+            field_name, field_type, field.offset * 8
+        )
+        members.append(member)
 
-    global_var = generator.create_global_var_debug_info(
-        struct.name, struct_type, is_local=False
+    if struct.name.startswith("struct_"):
+        struct_name = struct.name.removeprefix("struct_")
+    else:
+        raise ValueError("Unions are not supported in the current version")
+    # Create struct type with all members
+    struct_type = generator.create_struct_type_with_name(
+        struct_name, members, struct.__sizeof__() * 8, is_distinct=True
     )
 
-    return global_var
+    return struct_type
+
+
+def _get_field_debug_type(
+        field_name: str,
+        field,
+        generator: DebugInfoGenerator,
+        parent_struct: DependencyNode,
+        generated_debug_info: List[Tuple[DependencyNode, Any]]
+) -> Any:
+    """
+    Determine the appropriate debug type for a field based on its Python/ctypes type.
+
+    Args:
+        field_name: Name of the field
+        field: Field object containing type information
+        generator: DebugInfoGenerator instance
+        parent_struct: The parent struct containing this field
+        generated_debug_info: List of already generated debug info
+
+    Returns:
+        The debug info type for this field
+    """
+    # Handle complex types (arrays, pointers)
+    if field.ctype_complex_type is not None:
+        if issubclass(field.ctype_complex_type, ctypes.Array):
+            # Handle array types
+            element_type = _get_basic_debug_type(field.containing_type, generator)
+            return generator.create_array_type(element_type, field.type_size)
+        elif issubclass(field.ctype_complex_type, ctypes._Pointer):
+            # Handle pointer types
+            pointee_type = _get_basic_debug_type(field.containing_type, generator)
+            return generator.create_pointer_type(pointee_type)
+
+    # Handle other vmlinux types (nested structs)
+    if field.type.__module__ == "vmlinux":
+        # If it's a struct from vmlinux, check if we've already generated debug info for it
+        struct_name = field.type.__name__
+
+        # Look for existing debug info in the list
+        for existing_struct, debug_info in generated_debug_info:
+            if existing_struct.name == struct_name:
+                # Use existing debug info
+                return debug_info
+
+        # If not found, create a forward declaration
+        # This will be completed when the actual struct is processed
+        logger.warning("Forward declaration in struct created")
+        forward_type = generator.create_struct_type([], 0, is_distinct=True)
+        return forward_type
+
+    # Handle basic C types
+    return _get_basic_debug_type(field.type, generator)
+
+
+def _get_basic_debug_type(ctype, generator: DebugInfoGenerator) -> Any:
+    """
+    Map a ctypes type to a DWARF debug type.
+
+    Args:
+        ctype: A ctypes type or Python type
+        generator: DebugInfoGenerator instance
+
+    Returns:
+        The corresponding debug type
+    """
+    # Map ctypes to debug info types
+    if ctype == ctypes.c_char or ctype == ctypes.c_byte:
+        return generator.get_basic_type("char", 8, dc.DW_ATE_signed_char), 8
+    elif ctype == ctypes.c_ubyte or ctype == ctypes.c_uint8:
+        return generator.get_basic_type("unsigned char", 8, dc.DW_ATE_unsigned_char), 8
+    elif ctype == ctypes.c_short or ctype == ctypes.c_int16:
+        return generator.get_basic_type("short", 16, dc.DW_ATE_signed), 16
+    elif ctype == ctypes.c_ushort or ctype == ctypes.c_uint16:
+        return generator.get_basic_type("unsigned short", 16, dc.DW_ATE_unsigned), 16
+    elif ctype == ctypes.c_int or ctype == ctypes.c_int32:
+        return generator.get_basic_type("int", 32, dc.DW_ATE_signed), 32
+    elif ctype == ctypes.c_uint or ctype == ctypes.c_uint32:
+        return generator.get_basic_type("unsigned int", 32, dc.DW_ATE_unsigned), 32
+    elif ctype == ctypes.c_long:
+        return generator.get_basic_type("long", 64, dc.DW_ATE_signed), 64
+    elif ctype == ctypes.c_ulong:
+        return generator.get_basic_type("unsigned long", 64, dc.DW_ATE_unsigned), 64
+    elif ctype == ctypes.c_longlong or ctype == ctypes.c_int64:
+        return generator.get_basic_type("long long", 64, dc.DW_ATE_signed), 64
+    elif ctype == ctypes.c_ulonglong or ctype == ctypes.c_uint64:
+        return generator.get_basic_type("unsigned long long", 64, dc.DW_ATE_unsigned), 64
+    elif ctype == ctypes.c_float:
+        return generator.get_basic_type("float", 32, dc.DW_ATE_float), 32
+    elif ctype == ctypes.c_double:
+        return generator.get_basic_type("double", 64, dc.DW_ATE_float), 64
+    elif ctype == ctypes.c_bool:
+        return generator.get_basic_type("bool", 8, dc.DW_ATE_boolean), 8
+    elif ctype == ctypes.c_char_p:
+        char_type = generator.get_basic_type("char", 8, dc.DW_ATE_signed_char), 8
+        return generator.create_pointer_type(char_type)
+    elif ctype == ctypes.c_void_p:
+        void_type = generator.module.add_debug_info(
+            "DIBasicType", {"name": "void"}
+        )
+        return generator.create_pointer_type(void_type), 64
+    else:
+        return generator.get_uint64_type(), 64

tests/passing_tests/vmlinux/xdp_pass.py

@@ -3,7 +3,7 @@
 from pythonbpf.helper import XDP_PASS
 from vmlinux import TASK_COMM_LEN  # noqa: F401
 from vmlinux import struct_xdp_md
-from vmlinux import struct_trace_event_raw_sys_enter  # noqa: F401
+# from vmlinux import struct_trace_event_raw_sys_enter  # noqa: F401
 from ctypes import c_int64
 
 # Instructions to how to run this program
@@ -13,27 +13,9 @@
 # 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