Merge pull request #60 from pythonbpf/vmlinux-handler

vmlinux handler with struct support for only int64 and unsigned uint64 type struct fields.

Author: varun-r-mallya

pythonbpf/allocation_pass.py

@@ -2,27 +2,15 @@
 import logging
 
 from llvmlite import ir
-from dataclasses import dataclass
-from typing import Any
+from .local_symbol import LocalSymbol
 from pythonbpf.helper import HelperHandlerRegistry
+from pythonbpf.vmlinux_parser.dependency_node import Field
 from .expr import VmlinuxHandlerRegistry
 from pythonbpf.type_deducer import ctypes_to_ir
 
 logger = logging.getLogger(__name__)
 
 
-@dataclass
-class LocalSymbol:
-    var: ir.AllocaInstr
-    ir_type: ir.Type
-    metadata: Any = None
-
-    def __iter__(self):
-        yield self.var
-        yield self.ir_type
-        yield self.metadata
-
-
 def create_targets_and_rvals(stmt):
     """Create lists of targets and right-hand values from an assignment statement."""
     if isinstance(stmt.targets[0], ast.Tuple):
@@ -60,21 +48,11 @@ def handle_assign_allocation(builder, stmt, local_sym_tab, structs_sym_tab):
             continue
 
         var_name = target.id
-
         # Skip if already allocated
         if var_name in local_sym_tab:
             logger.debug(f"Variable {var_name} already allocated, skipping")
             continue
 
-        # When allocating a variable, check if it's a vmlinux struct type
-        if isinstance(
-            stmt.value, ast.Name
-        ) and VmlinuxHandlerRegistry.is_vmlinux_struct(stmt.value.id):
-            # Handle vmlinux struct allocation
-            # This requires more implementation
-            print(stmt.value)
-            pass
-
         # Determine type and allocate based on rval
         if isinstance(rval, ast.Call):
             _allocate_for_call(builder, var_name, rval, local_sym_tab, structs_sym_tab)
@@ -248,9 +226,41 @@ def _allocate_for_attribute(builder, var_name, rval, local_sym_tab, structs_sym_
         logger.error(f"Struct variable '{struct_var}' not found")
         return
 
-    struct_type = local_sym_tab[struct_var].metadata
+    struct_type: type = local_sym_tab[struct_var].metadata
     if not struct_type or struct_type not in structs_sym_tab:
-        logger.error(f"Struct type '{struct_type}' not found")
+        if VmlinuxHandlerRegistry.is_vmlinux_struct(struct_type.__name__):
+            # Handle vmlinux struct field access
+            vmlinux_struct_name = struct_type.__name__
+            if not VmlinuxHandlerRegistry.has_field(vmlinux_struct_name, field_name):
+                logger.error(
+                    f"Field '{field_name}' not found in vmlinux struct '{vmlinux_struct_name}'"
+                )
+                return
+
+            field_type: tuple[ir.GlobalVariable, Field] = (
+                VmlinuxHandlerRegistry.get_field_type(vmlinux_struct_name, field_name)
+            )
+            field_ir, field = field_type
+            # TODO: For now, we only support integer type allocations.
+            # This always assumes first argument of function to be the context struct
+            base_ptr = builder.function.args[0]
+            local_sym_tab[
+                struct_var
+            ].var = base_ptr  # This is repurposing of var to store the pointer of the base type
+            local_sym_tab[struct_var].ir_type = field_ir
+
+            actual_ir_type = ir.IntType(64)
+
+            # Allocate with the actual IR type, not the GlobalVariable
+            var = _allocate_with_type(builder, var_name, actual_ir_type)
+            local_sym_tab[var_name] = LocalSymbol(var, actual_ir_type, field)
+
+            logger.info(
+                f"Pre-allocated {var_name} from vmlinux struct {vmlinux_struct_name}.{field_name}"
+            )
+            return
+        else:
+            logger.error(f"Struct type '{struct_type}' not found")
         return
 
     struct_info = structs_sym_tab[struct_type]

pythonbpf/assign_pass.py

@@ -3,6 +3,8 @@
 from llvmlite import ir
 from pythonbpf.expr import eval_expr
 from pythonbpf.helper import emit_probe_read_kernel_str_call
+from pythonbpf.type_deducer import ctypes_to_ir
+from pythonbpf.vmlinux_parser.dependency_node import Field
 
 logger = logging.getLogger(__name__)
 
@@ -148,7 +150,18 @@ def handle_variable_assignment(
     val, val_type = val_result
     logger.info(f"Evaluated value for {var_name}: {val} of type {val_type}, {var_type}")
     if val_type != var_type:
-        if isinstance(val_type, ir.IntType) and isinstance(var_type, ir.IntType):
+        if isinstance(val_type, Field):
+            logger.info("Handling assignment to struct field")
+            # TODO: handling only ctype struct fields for now. Handle other stuff too later.
+            if var_type == ctypes_to_ir(val_type.type.__name__):
+                builder.store(val, var_ptr)
+                logger.info(f"Assigned ctype struct field to {var_name}")
+                return True
+            logger.error(
+                f"Failed to assign ctype struct field to {var_name}: {val_type} != {var_type}"
+            )
+            return False
+        elif isinstance(val_type, ir.IntType) and isinstance(var_type, ir.IntType):
             # Allow implicit int widening
             if val_type.width < var_type.width:
                 val = builder.sext(val, var_type)

pythonbpf/codegen.py

@@ -37,6 +37,24 @@ def finalize_module(original_str):
     return re.sub(pattern, replacement, original_str)
 
 
+def bpf_passthrough_gen(module):
+    i32_ty = ir.IntType(32)
+    ptr_ty = ir.PointerType(ir.IntType(8))
+    fnty = ir.FunctionType(ptr_ty, [i32_ty, ptr_ty])
+
+    # Declare the intrinsic
+    passthrough = ir.Function(module, fnty, "llvm.bpf.passthrough.p0.p0")
+
+    # Set function attributes
+    # TODO: the ones commented are supposed to be there but cannot be added due to llvmlite limitations at the moment
+    # passthrough.attributes.add("nofree")
+    # passthrough.attributes.add("nosync")
+    passthrough.attributes.add("nounwind")
+    # passthrough.attributes.add("memory(none)")
+
+    return passthrough
+
+
 def find_bpf_chunks(tree):
     """Find all functions decorated with @bpf in the AST."""
     bpf_functions = []
@@ -57,6 +75,8 @@ def processor(source_code, filename, module):
     for func_node in bpf_chunks:
         logger.info(f"Found BPF function/struct: {func_node.name}")
 
+    bpf_passthrough_gen(module)
+
     vmlinux_symtab = vmlinux_proc(tree, module)
     if vmlinux_symtab:
         handler = VmlinuxHandler.initialize(vmlinux_symtab)
@@ -137,7 +157,7 @@ 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))
+    module_string: str = finalize_module(str(module))
 
     logger.info(f"IR written to {output}")
     with open(output, "w") as f:

pythonbpf/debuginfo/debug_info_generator.py

@@ -184,3 +184,83 @@ def create_global_var_debug_info(
             "DIGlobalVariableExpression",
             {"var": global_var, "expr": self.module.add_debug_info("DIExpression", {})},
         )
+
+    def get_int64_type(self):
+        return self.get_basic_type("long", 64, dc.DW_ATE_signed)
+
+    def create_subroutine_type(self, return_type, param_types):
+        """
+        Create a DISubroutineType given return type and list of parameter types.
+        Equivalent to: !DISubroutineType(types: !{ret, args...})
+        """
+        type_array = [return_type]
+        if isinstance(param_types, (list, tuple)):
+            type_array.extend(param_types)
+        else:
+            type_array.append(param_types)
+        return self.module.add_debug_info("DISubroutineType", {"types": type_array})
+
+    def create_local_variable_debug_info(
+        self, name: str, arg: int, var_type: Any
+    ) -> Any:
+        """
+        Create debug info for a local variable (DILocalVariable) without scope.
+        Example:
+        !DILocalVariable(name: "ctx", arg: 1, file: !3, line: 20, type: !7)
+        """
+        return self.module.add_debug_info(
+            "DILocalVariable",
+            {
+                "name": name,
+                "arg": arg,
+                "file": self.module._file_metadata,
+                "type": var_type,
+            },
+        )
+
+    def add_scope_to_local_variable(self, local_variable_debug_info, scope_value):
+        """
+        Add scope information to an existing local variable debug info object.
+        """
+        # TODO: this is a workaround a flaw in the debug info generation. Fix this if possible in the future.
+        # We should not be touching llvmlite's internals like this.
+        if hasattr(local_variable_debug_info, "operands"):
+            # LLVM metadata operands is a tuple, so we need to rebuild it
+            existing_operands = local_variable_debug_info.operands
+
+            # Convert tuple to list, add scope, convert back to tuple
+            operands_list = list(existing_operands)
+            operands_list.append(("scope", scope_value))
+
+            # Reassign the new tuple
+            local_variable_debug_info.operands = tuple(operands_list)
+
+    def create_subprogram(
+        self, name: str, subroutine_type: Any, retained_nodes: List[Any]
+    ) -> Any:
+        """
+        Create a DISubprogram for a function.
+
+        Args:
+            name: Function name
+            subroutine_type: DISubroutineType for the function signature
+            retained_nodes: List of DILocalVariable nodes for function parameters/variables
+
+        Returns:
+            DISubprogram metadata
+        """
+        return self.module.add_debug_info(
+            "DISubprogram",
+            {
+                "name": name,
+                "scope": self.module._file_metadata,
+                "file": self.module._file_metadata,
+                "type": subroutine_type,
+                # TODO: the following flags do not exist at the moment in our dwarf constants file. We need to add them.
+                # "flags": dc.DW_FLAG_Prototyped | dc.DW_FLAG_AllCallsDescribed,
+                # "spFlags": dc.DW_SPFLAG_Definition | dc.DW_SPFLAG_Optimized,
+                "unit": self.module._debug_compile_unit,
+                "retainedNodes": retained_nodes,
+            },
+            is_distinct=True,
+        )

pythonbpf/expr/expr_pass.py

@@ -72,20 +72,28 @@ def _handle_attribute_expr(
         if var_name in local_sym_tab:
             var_ptr, var_type, var_metadata = local_sym_tab[var_name]
             logger.info(f"Loading attribute {attr_name} from variable {var_name}")
-            logger.info(f"Variable type: {var_type}, Variable ptr: {var_ptr}")
+            logger.info(
+                f"Variable type: {var_type}, Variable ptr: {var_ptr}, Variable Metadata: {var_metadata}"
+            )
+            if (
+                hasattr(var_metadata, "__module__")
+                and var_metadata.__module__ == "vmlinux"
+            ):
+                # Try vmlinux handler when var_metadata is not a string, but has a module attribute.
+                # This has been done to keep everything separate in vmlinux struct handling.
+                vmlinux_result = VmlinuxHandlerRegistry.handle_attribute(
+                    expr, local_sym_tab, None, builder
+                )
+                if vmlinux_result is not None:
+                    return vmlinux_result
+                else:
+                    raise RuntimeError("Vmlinux struct did not process successfully")
             metadata = structs_sym_tab[var_metadata]
             if attr_name in metadata.fields:
                 gep = metadata.gep(builder, var_ptr, attr_name)
                 val = builder.load(gep)
                 field_type = metadata.field_type(attr_name)
                 return val, field_type
-
-        # Try vmlinux handler as fallback
-        vmlinux_result = VmlinuxHandlerRegistry.handle_attribute(
-            expr, local_sym_tab, None, builder
-        )
-        if vmlinux_result is not None:
-            return vmlinux_result
     return None
 
 

pythonbpf/expr/vmlinux_registry.py

@@ -1,13 +1,15 @@
 import ast
 
+from pythonbpf.vmlinux_parser.vmlinux_exports_handler import VmlinuxHandler
+
 
 class VmlinuxHandlerRegistry:
     """Registry for vmlinux handler operations"""
 
     _handler = None
 
     @classmethod
-    def set_handler(cls, handler):
+    def set_handler(cls, handler: VmlinuxHandler):
         """Set the vmlinux handler"""
         cls._handler = handler
 
@@ -37,9 +39,37 @@ def handle_attribute(cls, expr, local_sym_tab, module, builder):
             )
         return None
 
+    @classmethod
+    def get_struct_debug_info(cls, name):
+        if cls._handler is None:
+            return False
+        return cls._handler.get_struct_debug_info(name)
+
     @classmethod
     def is_vmlinux_struct(cls, name):
         """Check if a name refers to a vmlinux struct"""
         if cls._handler is None:
             return False
         return cls._handler.is_vmlinux_struct(name)
+
+    @classmethod
+    def get_struct_type(cls, name):
+        """Try to handle a struct name as vmlinux struct"""
+        if cls._handler is None:
+            return None
+        return cls._handler.get_vmlinux_struct_type(name)
+
+    @classmethod
+    def has_field(cls, vmlinux_struct_name, field_name):
+        """Check if a vmlinux struct has a specific field"""
+        if cls._handler is None:
+            return False
+        return cls._handler.has_field(vmlinux_struct_name, field_name)
+
+    @classmethod
+    def get_field_type(cls, vmlinux_struct_name, field_name):
+        """Get the type of a field in a vmlinux struct"""
+        if cls._handler is None:
+            return None
+        assert isinstance(cls._handler, VmlinuxHandler)
+        return cls._handler.get_field_type(vmlinux_struct_name, field_name)