31 Commits

Author SHA1 Message Date
d0fecbc03c Add sync_perf_output BCC example 2025-10-20 15:43:57 +05:30
174095973b Fix userspace calling in sys_sync 2025-10-20 15:10:12 +05:30
3273620447 Fix userspace calling in sync_timing 2025-10-20 15:08:27 +05:30
610cbe82a8 Fix userspace calling in sync_count 2025-10-20 15:04:40 +05:30
54c97e648b Fix userspace calling in hello_fields 2025-10-20 15:02:31 +05:30
dd9411b7b9 Fix userspace calling in hello_world 2025-10-20 14:59:13 +05:30
aa85d0e0ef Remove unnecessary attached var in hello_perf_output 2025-10-20 14:58:50 +05:30
5bba8dce12 Complete hello_perf_output BCC example 2025-10-20 04:02:34 +05:30
8c976e46ae Fix loglevel and pylibbpf import in codegen 2025-10-20 04:00:30 +05:30
60737d9894 Improve error handling in compile, pass structs_sym_tab and maps_sym_tab to BpfProgram 2025-10-17 03:25:15 +05:30
fc55b7ecaa Add passing ptr_to_char_array test for strings 2025-10-16 23:42:03 +05:30
c143739a04 Add passing test struct_field_to_var_str for strings 2025-10-16 23:21:55 +05:30
009b11aca6 Implement bpf_probe_read_kernel_str helper, Allow i8* to i8 ArrayType conversion 2025-10-15 23:52:15 +05:30
9fc3c85b75 Add struct-field to named-var assignment of char arrays 2025-10-15 20:02:18 +05:30
fd630293f7 Remove duplicate alignment logic from allocation_pass 2025-10-15 18:14:13 +05:30
81f72a7698 Support var-to-var and var-to-struct-fld allocations 2025-10-15 18:10:04 +05:30
fb480639a5 Make char array struct fields work 2025-10-15 17:27:43 +05:30
13a804f7ac Implement bpf_get_current_comm_emitter 2025-10-15 14:03:09 +05:30
a0d954b20b Register bpf_get_current_comm_emitter for comm 2025-10-15 12:47:30 +05:30
b105c70b38 Add hello_perf_output BCC-Example skeleton 2025-10-15 12:35:59 +05:30
0a1557e318 Add sync_count BCC example to use tuple-like assignment 2025-10-14 23:33:54 +05:30
c56928bc8a Add create_targets_and_rvals, use it in handle_assign to enable tuple assignment 2025-10-14 23:30:59 +05:30
dd3fc74d09 Add support for tuple style multiiassignment in allocation_pass 2025-10-14 23:06:43 +05:30
4a79f9b9b2 Add sync_count BCC-Example 2025-10-14 18:50:46 +05:30
b676a5ebb4 Fix return in BCC-Examples 2025-10-14 17:05:05 +05:30
d7329ad3d7 Add BCC sync_timing example 2025-10-14 16:07:55 +05:30
903654daff Add hello_fields BCC Example 2025-10-14 04:42:12 +05:30
263402d137 Add trace_fields 2025-10-14 04:22:17 +05:30
37d1e1b143 Add sys_sync BCC example 2025-10-14 03:54:02 +05:30
edc33733d9 Add trace_pipe utility 2025-10-14 03:51:43 +05:30
18d62d605a Add hello_world BCC example 2025-10-13 20:11:27 +05:30
20 changed files with 1143 additions and 130 deletions

View File

@ -0,0 +1,34 @@
from pythonbpf import bpf, section, bpfglobal, BPF, trace_fields
from ctypes import c_void_p, c_int64
@bpf
@section("tracepoint/syscalls/sys_enter_clone")
def hello_world(ctx: c_void_p) -> c_int64:
print("Hello, World!")
return 0 # type: ignore [return-value]
@bpf
@bpfglobal
def LICENSE() -> str:
return "GPL"
# Compile and load
b = BPF()
b.load()
b.attach_all()
# header
print(f"{'TIME(s)':<18} {'COMM':<16} {'PID':<6} {'MESSAGE'}")
# format output
while True:
try:
(task, pid, cpu, flags, ts, msg) = trace_fields()
except ValueError:
continue
except KeyboardInterrupt:
exit()
print(f"{ts:<18} {task:<16} {pid:<6} {msg}")

View File

@ -0,0 +1,61 @@
from pythonbpf import bpf, map, struct, section, bpfglobal, BPF
from pythonbpf.helper import ktime, pid, comm
from pythonbpf.maps import PerfEventArray
from ctypes import c_void_p, c_int64, c_uint64
@bpf
@struct
class data_t:
pid: c_uint64
ts: c_uint64
comm: str(16) # type: ignore [valid-type]
@bpf
@map
def events() -> PerfEventArray:
return PerfEventArray(key_size=c_int64, value_size=c_int64)
@bpf
@section("tracepoint/syscalls/sys_enter_clone")
def hello(ctx: c_void_p) -> c_int64:
dataobj = data_t()
dataobj.pid, dataobj.ts = pid(), ktime()
comm(dataobj.comm)
events.output(dataobj)
return 0 # type: ignore [return-value]
@bpf
@bpfglobal
def LICENSE() -> str:
return "GPL"
# Compile and load
b = BPF()
b.load()
b.attach_all()
start = 0
def callback(cpu, event):
global start
if start == 0:
start = event.ts
ts = (event.ts - start) / 1e9
print(f"[CPU {cpu}] PID: {event.pid}, TS: {ts}, COMM: {event.comm.decode()}")
perf = b["events"].open_perf_buffer(callback, struct_name="data_t")
print("Starting to poll... (Ctrl+C to stop)")
print("Try running: fork() or clone() system calls to trigger events")
try:
while True:
b["events"].poll(1000)
except KeyboardInterrupt:
print("Stopping...")

View File

@ -0,0 +1,23 @@
from pythonbpf import bpf, section, bpfglobal, BPF, trace_pipe
from ctypes import c_void_p, c_int64
@bpf
@section("tracepoint/syscalls/sys_enter_clone")
def hello_world(ctx: c_void_p) -> c_int64:
print("Hello, World!")
return 0 # type: ignore [return-value]
@bpf
@bpfglobal
def LICENSE() -> str:
return "GPL"
# Compile and load
b = BPF()
b.load()
b.attach_all()
trace_pipe()

View File

@ -0,0 +1,58 @@
from pythonbpf import bpf, map, section, bpfglobal, BPF, trace_fields
from pythonbpf.helper import ktime
from pythonbpf.maps import HashMap
from ctypes import c_void_p, c_int64
@bpf
@map
def last() -> HashMap:
return HashMap(key=c_int64, value=c_int64, max_entries=2)
@bpf
@section("tracepoint/syscalls/sys_enter_sync")
def do_trace(ctx: c_void_p) -> c_int64:
ts_key, cnt_key = 0, 1
tsp, cntp = last.lookup(ts_key), last.lookup(cnt_key)
if not cntp:
last.update(cnt_key, 0)
cntp = last.lookup(cnt_key)
if tsp:
delta = ktime() - tsp
if delta < 1000000000:
time_ms = delta // 1000000
print(f"{time_ms} {cntp}")
last.delete(ts_key)
else:
last.update(ts_key, ktime())
last.update(cnt_key, cntp + 1)
return 0 # type: ignore [return-value]
@bpf
@bpfglobal
def LICENSE() -> str:
return "GPL"
# Compile and load
b = BPF()
b.load()
b.attach_all()
print("Tracing for quick sync's... Ctrl-C to end")
# format output
start = 0
while True:
try:
task, pid, cpu, flags, ts, msg = trace_fields()
if start == 0:
start = ts
ts -= start
ms, cnt = msg.split()
print(f"At time {ts} s: Multiple syncs detected, last {ms} ms ago. Count {cnt}")
except KeyboardInterrupt:
exit()

View File

@ -0,0 +1,78 @@
from pythonbpf import bpf, map, struct, section, bpfglobal, BPF
from pythonbpf.helper import ktime
from pythonbpf.maps import HashMap
from pythonbpf.maps import PerfEventArray
from ctypes import c_void_p, c_int64
@bpf
@struct
class data_t:
ts: c_int64
ms: c_int64
@bpf
@map
def events() -> PerfEventArray:
return PerfEventArray(key_size=c_int64, value_size=c_int64)
@bpf
@map
def last() -> HashMap:
return HashMap(key=c_int64, value=c_int64, max_entries=1)
@bpf
@section("tracepoint/syscalls/sys_enter_sync")
def do_trace(ctx: c_void_p) -> c_int64:
dat, dat.ts, key = data_t(), ktime(), 0
tsp = last.lookup(key)
if tsp:
delta = ktime() - tsp
if delta < 1000000000:
dat.ms = delta // 1000000
events.output(dat)
last.delete(key)
else:
last.update(key, ktime())
return 0 # type: ignore [return-value]
@bpf
@bpfglobal
def LICENSE() -> str:
return "GPL"
# Compile and load
b = BPF()
b.load()
b.attach_all()
print("Tracing for quick sync's... Ctrl-C to end")
# format output
start = 0
def callback(cpu, event):
global start
if start == 0:
start = event.ts
event.ts -= start
print(
f"At time {event.ts / 1e9} s: Multiple sync detected, Last sync: {event.ms} ms ago"
)
perf = b["events"].open_perf_buffer(callback, struct_name="data_t")
print("Starting to poll... (Ctrl+C to stop)")
print("Try running: fork() or clone() system calls to trigger events")
try:
while True:
b["events"].poll(1000)
except KeyboardInterrupt:
print("Stopping...")

View File

@ -0,0 +1,53 @@
from pythonbpf import bpf, map, section, bpfglobal, BPF, trace_fields
from pythonbpf.helper import ktime
from pythonbpf.maps import HashMap
from ctypes import c_void_p, c_int64
@bpf
@map
def last() -> HashMap:
return HashMap(key=c_int64, value=c_int64, max_entries=1)
@bpf
@section("tracepoint/syscalls/sys_enter_sync")
def do_trace(ctx: c_void_p) -> c_int64:
key = 0
tsp = last.lookup(key)
if tsp:
delta = ktime() - tsp
if delta < 1000000000:
time_ms = delta // 1000000
print(f"{time_ms}")
last.delete(key)
else:
last.update(key, ktime())
return 0 # type: ignore [return-value]
@bpf
@bpfglobal
def LICENSE() -> str:
return "GPL"
# Compile and load
b = BPF()
b.load()
b.attach_all()
print("Tracing for quick sync's... Ctrl-C to end")
# format output
start = 0
while True:
try:
task, pid, cpu, flags, ts, ms = trace_fields()
if start == 0:
start = ts
ts -= start
print(f"At time {ts} s: Multiple syncs detected, last {ms} ms ago")
except KeyboardInterrupt:
exit()

23
BCC-Examples/sys_sync.py Normal file
View File

@ -0,0 +1,23 @@
from pythonbpf import bpf, section, bpfglobal, BPF, trace_pipe
from ctypes import c_void_p, c_int64
@bpf
@section("tracepoint/syscalls/sys_enter_sync")
def hello_world(ctx: c_void_p) -> c_int64:
print("sys_sync() called")
return c_int64(0)
@bpf
@bpfglobal
def LICENSE() -> str:
return "GPL"
# Compile and load
b = BPF()
b.load()
b.attach_all()
print("Tracing sys_sync()... Ctrl-C to end.")
trace_pipe()

View File

@ -1,5 +1,6 @@
from .decorators import bpf, map, section, bpfglobal, struct
from .codegen import compile_to_ir, compile, BPF
from .utils import trace_pipe, trace_fields
__all__ = [
"bpf",
@ -10,4 +11,6 @@ __all__ = [
"compile_to_ir",
"compile",
"BPF",
"trace_pipe",
"trace_fields",
]

View File

@ -22,44 +22,68 @@ class LocalSymbol:
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):
if not isinstance(stmt.value, ast.Tuple):
logger.warning("Mismatched multi-target assignment, skipping allocation")
return
targets, rvals = stmt.targets[0].elts, stmt.value.elts
if len(targets) != len(rvals):
logger.warning("length of LHS != length of RHS, skipping allocation")
return
return targets, rvals
return stmt.targets, [stmt.value]
def handle_assign_allocation(builder, stmt, local_sym_tab, structs_sym_tab):
"""Handle memory allocation for assignment statements."""
# Validate assignment
if len(stmt.targets) != 1:
logger.warning("Multi-target assignment not supported, skipping allocation")
return
logger.info(f"Handling assignment for allocation: {ast.dump(stmt)}")
target = stmt.targets[0]
# NOTE: Support multi-target assignments (e.g.: a, b = 1, 2)
targets, rvals = create_targets_and_rvals(stmt)
# Skip non-name targets (e.g., struct field assignments)
if isinstance(target, ast.Attribute):
logger.debug(f"Struct field assignment to {target.attr}, no allocation needed")
return
for target, rval in zip(targets, rvals):
# Skip non-name targets (e.g., struct field assignments)
if isinstance(target, ast.Attribute):
logger.debug(
f"Struct field assignment to {target.attr}, no allocation needed"
)
continue
if not isinstance(target, ast.Name):
logger.warning(f"Unsupported assignment target type: {type(target).__name__}")
return
if not isinstance(target, ast.Name):
logger.warning(
f"Unsupported assignment target type: {type(target).__name__}"
)
continue
var_name = target.id
rval = stmt.value
var_name = target.id
# Skip if already allocated
if var_name in local_sym_tab:
logger.debug(f"Variable {var_name} already allocated, skipping")
return
# Skip if already allocated
if var_name in local_sym_tab:
logger.debug(f"Variable {var_name} already allocated, skipping")
continue
# 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)
elif isinstance(rval, ast.Constant):
_allocate_for_constant(builder, var_name, rval, local_sym_tab)
elif isinstance(rval, ast.BinOp):
_allocate_for_binop(builder, var_name, local_sym_tab)
else:
logger.warning(
f"Unsupported assignment value type for {var_name}: {type(rval).__name__}"
)
# 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)
elif isinstance(rval, ast.Constant):
_allocate_for_constant(builder, var_name, rval, local_sym_tab)
elif isinstance(rval, ast.BinOp):
_allocate_for_binop(builder, var_name, local_sym_tab)
elif isinstance(rval, ast.Name):
# Variable-to-variable assignment (b = a)
_allocate_for_name(builder, var_name, rval, local_sym_tab)
elif isinstance(rval, ast.Attribute):
# Struct field-to-variable assignment (a = dat.fld)
_allocate_for_attribute(
builder, var_name, rval, local_sym_tab, structs_sym_tab
)
else:
logger.warning(
f"Unsupported assignment value type for {var_name}: {type(rval).__name__}"
)
def _allocate_for_call(builder, var_name, rval, local_sym_tab, structs_sym_tab):
@ -176,3 +200,88 @@ def allocate_temp_pool(builder, max_temps, local_sym_tab):
temp_var = builder.alloca(ir.IntType(64), name=temp_name)
temp_var.align = 8
local_sym_tab[temp_name] = LocalSymbol(temp_var, ir.IntType(64))
def _allocate_for_name(builder, var_name, rval, local_sym_tab):
"""Allocate memory for variable-to-variable assignment (b = a)."""
source_var = rval.id
if source_var not in local_sym_tab:
logger.error(f"Source variable '{source_var}' not found in symbol table")
return
# Get type and metadata from source variable
source_symbol = local_sym_tab[source_var]
# Allocate with same type and alignment
var = _allocate_with_type(builder, var_name, source_symbol.ir_type)
local_sym_tab[var_name] = LocalSymbol(
var, source_symbol.ir_type, source_symbol.metadata
)
logger.info(
f"Pre-allocated {var_name} from {source_var} with type {source_symbol.ir_type}"
)
def _allocate_for_attribute(builder, var_name, rval, local_sym_tab, structs_sym_tab):
"""Allocate memory for struct field-to-variable assignment (a = dat.fld)."""
if not isinstance(rval.value, ast.Name):
logger.warning(f"Complex attribute access not supported for {var_name}")
return
struct_var = rval.value.id
field_name = rval.attr
# Validate struct and field
if struct_var not in local_sym_tab:
logger.error(f"Struct variable '{struct_var}' not found")
return
struct_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")
return
struct_info = structs_sym_tab[struct_type]
if field_name not in struct_info.fields:
logger.error(f"Field '{field_name}' not found in struct '{struct_type}'")
return
# Get field type
field_type = struct_info.field_type(field_name)
# Special case: char array -> allocate as i8* pointer instead
if (
isinstance(field_type, ir.ArrayType)
and isinstance(field_type.element, ir.IntType)
and field_type.element.width == 8
):
alloc_type = ir.PointerType(ir.IntType(8))
logger.info(f"Allocating {var_name} as i8* (pointer to char array)")
else:
alloc_type = field_type
var = _allocate_with_type(builder, var_name, alloc_type)
local_sym_tab[var_name] = LocalSymbol(var, alloc_type)
logger.info(
f"Pre-allocated {var_name} from {struct_var}.{field_name} with type {alloc_type}"
)
def _allocate_with_type(builder, var_name, ir_type):
"""Allocate variable with appropriate alignment for type."""
var = builder.alloca(ir_type, name=var_name)
var.align = _get_alignment(ir_type)
return var
def _get_alignment(ir_type):
"""Get appropriate alignment for IR type."""
if isinstance(ir_type, ir.IntType):
return ir_type.width // 8
elif isinstance(ir_type, ir.ArrayType) and isinstance(ir_type.element, ir.IntType):
return ir_type.element.width // 8
else:
return 8 # Default: pointer size

View File

@ -2,6 +2,7 @@ import ast
import logging
from llvmlite import ir
from pythonbpf.expr import eval_expr
from pythonbpf.helper import emit_probe_read_kernel_str_call
logger = logging.getLogger(__name__)
@ -27,27 +28,82 @@ def handle_struct_field_assignment(
# Get field pointer and evaluate value
field_ptr = struct_info.gep(builder, local_sym_tab[var_name].var, field_name)
val = eval_expr(
field_type = struct_info.field_type(field_name)
val_result = eval_expr(
func, module, builder, rval, local_sym_tab, map_sym_tab, structs_sym_tab
)
if val is None:
if val_result is None:
logger.error(f"Failed to evaluate value for {var_name}.{field_name}")
return
# TODO: Handle string assignment to char array (not a priority)
field_type = struct_info.field_type(field_name)
if isinstance(field_type, ir.ArrayType) and val[1] == ir.PointerType(ir.IntType(8)):
logger.warning(
f"String to char array assignment not implemented for {var_name}.{field_name}"
val, val_type = val_result
# Special case: i8* string to [N x i8] char array
if _is_char_array(field_type) and _is_i8_ptr(val_type):
_copy_string_to_char_array(
func,
module,
builder,
val,
field_ptr,
field_type,
local_sym_tab,
map_sym_tab,
structs_sym_tab,
)
logger.info(f"Copied string to char array {var_name}.{field_name}")
return
# Store the value
builder.store(val[0], field_ptr)
# Regular assignment
builder.store(val, field_ptr)
logger.info(f"Assigned to struct field {var_name}.{field_name}")
def _copy_string_to_char_array(
func,
module,
builder,
src_ptr,
dst_ptr,
array_type,
local_sym_tab,
map_sym_tab,
struct_sym_tab,
):
"""Copy string (i8*) to char array ([N x i8]) using bpf_probe_read_kernel_str"""
array_size = array_type.count
# Get pointer to first element: [N x i8]* -> i8*
dst_i8_ptr = builder.gep(
dst_ptr,
[ir.Constant(ir.IntType(32), 0), ir.Constant(ir.IntType(32), 0)],
inbounds=True,
)
# Use the shared emitter function
emit_probe_read_kernel_str_call(builder, dst_i8_ptr, array_size, src_ptr)
def _is_char_array(ir_type):
"""Check if type is [N x i8]."""
return (
isinstance(ir_type, ir.ArrayType)
and isinstance(ir_type.element, ir.IntType)
and ir_type.element.width == 8
)
def _is_i8_ptr(ir_type):
"""Check if type is i8*."""
return (
isinstance(ir_type, ir.PointerType)
and isinstance(ir_type.pointee, ir.IntType)
and ir_type.pointee.width == 8
)
def handle_variable_assignment(
func, module, builder, var_name, rval, local_sym_tab, map_sym_tab, structs_sym_tab
):
@ -71,6 +127,17 @@ def handle_variable_assignment(
logger.info(f"Initialized struct {struct_name} for variable {var_name}")
return True
# Special case: struct field char array -> pointer
# Handle this before eval_expr to get the pointer, not the value
if isinstance(rval, ast.Attribute) and isinstance(rval.value, ast.Name):
converted_val = _try_convert_char_array_to_ptr(
rval, var_type, builder, local_sym_tab, structs_sym_tab
)
if converted_val is not None:
builder.store(converted_val, var_ptr)
logger.info(f"Assigned char array pointer to {var_name}")
return True
val_result = eval_expr(
func, module, builder, rval, local_sym_tab, map_sym_tab, structs_sym_tab
)
@ -106,3 +173,52 @@ def handle_variable_assignment(
builder.store(val, var_ptr)
logger.info(f"Assigned value to variable {var_name}")
return True
def _try_convert_char_array_to_ptr(
rval, var_type, builder, local_sym_tab, structs_sym_tab
):
"""Try to convert char array field to i8* pointer"""
# Only convert if target is i8*
if not (
isinstance(var_type, ir.PointerType)
and isinstance(var_type.pointee, ir.IntType)
and var_type.pointee.width == 8
):
return None
struct_var = rval.value.id
field_name = rval.attr
# Validate struct
if struct_var not in local_sym_tab:
return None
struct_type = local_sym_tab[struct_var].metadata
if not struct_type or struct_type not in structs_sym_tab:
return None
struct_info = structs_sym_tab[struct_type]
if field_name not in struct_info.fields:
return None
field_type = struct_info.field_type(field_name)
# Check if it's a char array
if not (
isinstance(field_type, ir.ArrayType)
and isinstance(field_type.element, ir.IntType)
and field_type.element.width == 8
):
return None
# Get pointer to struct field
struct_ptr = local_sym_tab[struct_var].var
field_ptr = struct_info.gep(builder, struct_ptr, field_name)
# GEP to first element: [N x i8]* -> i8*
return builder.gep(
field_ptr,
[ir.Constant(ir.IntType(32), 0), ir.Constant(ir.IntType(32), 0)],
inbounds=True,
)

View File

@ -15,7 +15,7 @@ import os
import subprocess
import inspect
from pathlib import Path
from pylibbpf import BpfProgram
from pylibbpf import BpfObject
import tempfile
from logging import Logger
import logging
@ -55,6 +55,7 @@ def processor(source_code, filename, module):
func_proc(tree, module, bpf_chunks, map_sym_tab, structs_sym_tab)
globals_list_creation(tree, module)
return structs_sym_tab, map_sym_tab
def compile_to_ir(filename: str, output: str, loglevel=logging.INFO):
@ -80,7 +81,7 @@ def compile_to_ir(filename: str, output: str, loglevel=logging.INFO):
True,
)
processor(source, filename, module)
structs_sym_tab, maps_sym_tab = processor(source, filename, module)
wchar_size = module.add_metadata(
[
@ -127,7 +128,7 @@ def compile_to_ir(filename: str, output: str, loglevel=logging.INFO):
f.write(str(module))
f.write("\n")
return output
return output, structs_sym_tab, maps_sym_tab
def _run_llc(ll_file, obj_file):
@ -157,7 +158,7 @@ def _run_llc(ll_file, obj_file):
return False
def compile(loglevel=logging.INFO) -> bool:
def compile(loglevel=logging.WARNING) -> bool:
# Look one level up the stack to the caller of this function
caller_frame = inspect.stack()[1]
caller_file = Path(caller_frame.filename).resolve()
@ -165,18 +166,19 @@ def compile(loglevel=logging.INFO) -> bool:
ll_file = Path("/tmp") / caller_file.with_suffix(".ll").name
o_file = caller_file.with_suffix(".o")
success = True
success = (
compile_to_ir(str(caller_file), str(ll_file), loglevel=loglevel) and success
_, structs_sym_tab, maps_sym_tab = compile_to_ir(
str(caller_file), str(ll_file), loglevel=loglevel
)
success = _run_llc(ll_file, o_file) and success
if not _run_llc(ll_file, o_file):
logger.error("Compilation to object file failed.")
return False
logger.info(f"Object written to {o_file}")
return success
return True
def BPF(loglevel=logging.INFO) -> BpfProgram:
def BPF(loglevel=logging.WARNING) -> BpfObject:
caller_frame = inspect.stack()[1]
src = inspect.getsource(caller_frame.frame)
with tempfile.NamedTemporaryFile(
@ -189,7 +191,9 @@ def BPF(loglevel=logging.INFO) -> BpfProgram:
f.write(src)
f.flush()
source = f.name
compile_to_ir(source, str(inter.name), loglevel=loglevel)
_, structs_sym_tab, maps_sym_tab = compile_to_ir(
source, str(inter.name), loglevel=loglevel
)
_run_llc(str(inter.name), str(obj_file.name))
return BpfProgram(str(obj_file.name))
return BpfObject(str(obj_file.name), structs=structs_sym_tab)

View File

@ -12,7 +12,11 @@ from pythonbpf.assign_pass import (
handle_variable_assignment,
handle_struct_field_assignment,
)
from pythonbpf.allocation_pass import handle_assign_allocation, allocate_temp_pool
from pythonbpf.allocation_pass import (
handle_assign_allocation,
allocate_temp_pool,
create_targets_and_rvals,
)
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
@ -140,48 +144,43 @@ def handle_assign(
):
"""Handle assignment statements in the function body."""
# TODO: Support this later
# GH #37
if len(stmt.targets) != 1:
logger.error("Multi-target assignment is not supported for now")
return
# NOTE: Support multi-target assignments (e.g.: a, b = 1, 2)
targets, rvals = create_targets_and_rvals(stmt)
target = stmt.targets[0]
rval = stmt.value
for target, rval in zip(targets, rvals):
if isinstance(target, ast.Name):
# NOTE: Simple variable assignment case: x = 5
var_name = target.id
result = handle_variable_assignment(
func,
module,
builder,
var_name,
rval,
local_sym_tab,
map_sym_tab,
structs_sym_tab,
)
if not result:
logger.error(f"Failed to handle assignment to {var_name}")
continue
if isinstance(target, ast.Name):
# NOTE: Simple variable assignment case: x = 5
var_name = target.id
result = handle_variable_assignment(
func,
module,
builder,
var_name,
rval,
local_sym_tab,
map_sym_tab,
structs_sym_tab,
)
if not result:
logger.error(f"Failed to handle assignment to {var_name}")
return
if isinstance(target, ast.Attribute):
# NOTE: Struct field assignment case: pkt.field = value
handle_struct_field_assignment(
func,
module,
builder,
target,
rval,
local_sym_tab,
map_sym_tab,
structs_sym_tab,
)
continue
if isinstance(target, ast.Attribute):
# NOTE: Struct field assignment case: pkt.field = value
handle_struct_field_assignment(
func,
module,
builder,
target,
rval,
local_sym_tab,
map_sym_tab,
structs_sym_tab,
)
return
# Unsupported target type
logger.error(f"Unsupported assignment target: {ast.dump(target)}")
# Unsupported target type
logger.error(f"Unsupported assignment target: {ast.dump(target)}")
def handle_cond(

View File

@ -1,7 +1,7 @@
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
from .bpf_helper_handler import handle_helper_call, emit_probe_read_kernel_str_call
from .helpers import ktime, pid, deref, comm, probe_read_str, XDP_DROP, XDP_PASS
# Register the helper handler with expr module
@ -59,9 +59,12 @@ __all__ = [
"HelperHandlerRegistry",
"reset_scratch_pool",
"handle_helper_call",
"emit_probe_read_kernel_str_call",
"ktime",
"pid",
"deref",
"comm",
"probe_read_str",
"XDP_DROP",
"XDP_PASS",
]

View File

@ -7,6 +7,9 @@ from .helper_utils import (
get_or_create_ptr_from_arg,
get_flags_val,
get_data_ptr_and_size,
get_buffer_ptr_and_size,
get_char_array_ptr_and_size,
get_ptr_from_arg,
)
from .printk_formatter import simple_string_print, handle_fstring_print
@ -23,7 +26,9 @@ class BPFHelperID(Enum):
BPF_KTIME_GET_NS = 5
BPF_PRINTK = 6
BPF_GET_CURRENT_PID_TGID = 14
BPF_GET_CURRENT_COMM = 16
BPF_PERF_EVENT_OUTPUT = 25
BPF_PROBE_READ_KERNEL_STR = 115
@HelperHandlerRegistry.register("ktime")
@ -234,6 +239,63 @@ def bpf_map_delete_elem_emitter(
return result, None
@HelperHandlerRegistry.register("comm")
def bpf_get_current_comm_emitter(
call,
map_ptr,
module,
builder,
func,
local_sym_tab=None,
struct_sym_tab=None,
map_sym_tab=None,
):
"""
Emit LLVM IR for bpf_get_current_comm helper function call.
Accepts: comm(dataobj.field) or comm(my_buffer)
"""
if not call.args or len(call.args) != 1:
raise ValueError(
f"comm expects exactly one argument (buffer), got {len(call.args)}"
)
buf_arg = call.args[0]
# Extract buffer pointer and size
buf_ptr, buf_size = get_buffer_ptr_and_size(
buf_arg, builder, local_sym_tab, struct_sym_tab
)
# Validate it's a char array
if not isinstance(
buf_ptr.type.pointee, ir.ArrayType
) or buf_ptr.type.pointee.element != ir.IntType(8):
raise ValueError(
f"comm expects a char array buffer, got {buf_ptr.type.pointee}"
)
# Cast to void* and call helper
buf_void_ptr = builder.bitcast(buf_ptr, ir.PointerType())
fn_type = ir.FunctionType(
ir.IntType(64),
[ir.PointerType(), ir.IntType(32)],
var_arg=False,
)
fn_ptr = builder.inttoptr(
ir.Constant(ir.IntType(64), BPFHelperID.BPF_GET_CURRENT_COMM.value),
ir.PointerType(fn_type),
)
result = builder.call(
fn_ptr, [buf_void_ptr, ir.Constant(ir.IntType(32), buf_size)], tail=False
)
logger.info(f"Emitted bpf_get_current_comm with {buf_size} byte buffer")
return result, None
@HelperHandlerRegistry.register("pid")
def bpf_get_current_pid_tgid_emitter(
call,
@ -309,6 +371,68 @@ def bpf_perf_event_output_handler(
return result, None
def emit_probe_read_kernel_str_call(builder, dst_ptr, dst_size, src_ptr):
"""Emit LLVM IR call to bpf_probe_read_kernel_str"""
fn_type = ir.FunctionType(
ir.IntType(64),
[ir.PointerType(), ir.IntType(32), ir.PointerType()],
var_arg=False,
)
fn_ptr = builder.inttoptr(
ir.Constant(ir.IntType(64), BPFHelperID.BPF_PROBE_READ_KERNEL_STR.value),
ir.PointerType(fn_type),
)
result = builder.call(
fn_ptr,
[
builder.bitcast(dst_ptr, ir.PointerType()),
ir.Constant(ir.IntType(32), dst_size),
builder.bitcast(src_ptr, ir.PointerType()),
],
tail=False,
)
logger.info(f"Emitted bpf_probe_read_kernel_str (size={dst_size})")
return result
@HelperHandlerRegistry.register("probe_read_str")
def bpf_probe_read_kernel_str_emitter(
call,
map_ptr,
module,
builder,
func,
local_sym_tab=None,
struct_sym_tab=None,
map_sym_tab=None,
):
"""Emit LLVM IR for bpf_probe_read_kernel_str helper."""
if len(call.args) != 2:
raise ValueError(
f"probe_read_str expects 2 args (dst, src), got {len(call.args)}"
)
# Get destination buffer (char array -> i8*)
dst_ptr, dst_size = get_char_array_ptr_and_size(
call.args[0], builder, local_sym_tab, struct_sym_tab
)
# Get source pointer (evaluate expression)
src_ptr, src_type = get_ptr_from_arg(
call.args[1], func, module, builder, local_sym_tab, map_sym_tab, struct_sym_tab
)
# Emit the helper call
result = emit_probe_read_kernel_str_call(builder, dst_ptr, dst_size, src_ptr)
logger.info(f"Emitted bpf_probe_read_kernel_str (size={dst_size})")
return result, ir.IntType(64)
def handle_helper_call(
call,
module,

View File

@ -4,6 +4,7 @@ import logging
from llvmlite import ir
from pythonbpf.expr import (
get_operand_value,
eval_expr,
)
logger = logging.getLogger(__name__)
@ -136,3 +137,140 @@ def get_data_ptr_and_size(data_arg, local_sym_tab, struct_sym_tab):
raise NotImplementedError(
"Only simple object names are supported as data in perf event output."
)
def get_buffer_ptr_and_size(buf_arg, builder, local_sym_tab, struct_sym_tab):
"""Extract buffer pointer and size from either a struct field or variable."""
# Case 1: Struct field (obj.field)
if isinstance(buf_arg, ast.Attribute):
if not isinstance(buf_arg.value, ast.Name):
raise ValueError(
"Only simple struct field access supported (e.g., obj.field)"
)
struct_name = buf_arg.value.id
field_name = buf_arg.attr
# Lookup struct
if not local_sym_tab or struct_name not in local_sym_tab:
raise ValueError(f"Struct '{struct_name}' not found")
struct_type = local_sym_tab[struct_name].metadata
if not struct_sym_tab or struct_type not in struct_sym_tab:
raise ValueError(f"Struct type '{struct_type}' not found")
struct_info = struct_sym_tab[struct_type]
# Get field pointer and type
struct_ptr = local_sym_tab[struct_name].var
field_ptr = struct_info.gep(builder, struct_ptr, field_name)
field_type = struct_info.field_type(field_name)
if not isinstance(field_type, ir.ArrayType):
raise ValueError(f"Field '{field_name}' must be an array type")
return field_ptr, field_type.count
# Case 2: Variable name
elif isinstance(buf_arg, ast.Name):
var_name = buf_arg.id
if not local_sym_tab or var_name not in local_sym_tab:
raise ValueError(f"Variable '{var_name}' not found")
var_ptr = local_sym_tab[var_name].var
var_type = local_sym_tab[var_name].ir_type
if not isinstance(var_type, ir.ArrayType):
raise ValueError(f"Variable '{var_name}' must be an array type")
return var_ptr, var_type.count
else:
raise ValueError(
"comm expects either a struct field (obj.field) or variable name"
)
def get_char_array_ptr_and_size(buf_arg, builder, local_sym_tab, struct_sym_tab):
"""Get pointer to char array and its size."""
# Struct field: obj.field
if isinstance(buf_arg, ast.Attribute) and isinstance(buf_arg.value, ast.Name):
var_name = buf_arg.value.id
field_name = buf_arg.attr
if not (local_sym_tab and var_name in local_sym_tab):
raise ValueError(f"Variable '{var_name}' not found")
struct_type = local_sym_tab[var_name].metadata
if not (struct_sym_tab and struct_type in struct_sym_tab):
raise ValueError(f"Struct type '{struct_type}' not found")
struct_info = struct_sym_tab[struct_type]
if field_name not in struct_info.fields:
raise ValueError(f"Field '{field_name}' not found")
field_type = struct_info.field_type(field_name)
if not _is_char_array(field_type):
raise ValueError("Expected char array field")
struct_ptr = local_sym_tab[var_name].var
field_ptr = struct_info.gep(builder, struct_ptr, field_name)
# GEP to first element: [N x i8]* -> i8*
buf_ptr = builder.gep(
field_ptr,
[ir.Constant(ir.IntType(32), 0), ir.Constant(ir.IntType(32), 0)],
inbounds=True,
)
return buf_ptr, field_type.count
elif isinstance(buf_arg, ast.Name):
# NOTE: We shouldn't be doing this as we can't get size info
var_name = buf_arg.id
if not (local_sym_tab and var_name in local_sym_tab):
raise ValueError(f"Variable '{var_name}' not found")
var_ptr = local_sym_tab[var_name].var
var_type = local_sym_tab[var_name].ir_type
if not isinstance(var_type, ir.PointerType) and not isinstance(
var_type.pointee, ir.IntType(8)
):
raise ValueError("Expected str ptr variable")
return var_ptr, 256 # Size unknown for str ptr, using 256 as default
else:
raise ValueError("Expected struct field or variable name")
def _is_char_array(ir_type):
"""Check if IR type is [N x i8]."""
return (
isinstance(ir_type, ir.ArrayType)
and isinstance(ir_type.element, ir.IntType)
and ir_type.element.width == 8
)
def get_ptr_from_arg(
arg, func, module, builder, local_sym_tab, map_sym_tab, struct_sym_tab
):
"""Evaluate argument and return pointer value"""
result = eval_expr(
func, module, builder, arg, local_sym_tab, map_sym_tab, struct_sym_tab
)
if not result:
raise ValueError("Failed to evaluate argument")
val, val_type = result
if not isinstance(val_type, ir.PointerType):
raise ValueError(f"Expected pointer type, got {val_type}")
return val, val_type

View File

@ -2,19 +2,31 @@ import ctypes
def ktime():
"""get current ktime"""
return ctypes.c_int64(0)
def pid():
"""get current process id"""
return ctypes.c_int32(0)
def deref(ptr):
"dereference a pointer"
"""dereference a pointer"""
result = ctypes.cast(ptr, ctypes.POINTER(ctypes.c_void_p)).contents.value
return result if result is not None else 0
def comm(buf):
"""get current process command name"""
return ctypes.c_int64(0)
def probe_read_str(dst, src):
"""Safely read a null-terminated string from kernel memory"""
return ctypes.c_int64(0)
XDP_ABORTED = ctypes.c_int64(0)
XDP_DROP = ctypes.c_int64(1)
XDP_PASS = ctypes.c_int64(2)

View File

@ -173,6 +173,15 @@ def _populate_fval(ftype, node, fmt_parts, exprs):
raise NotImplementedError(
f"Unsupported pointer target type in f-string: {target}"
)
elif isinstance(ftype, ir.ArrayType):
if isinstance(ftype.element, ir.IntType) and ftype.element.width == 8:
# Char array
fmt_parts.append("%s")
exprs.append(node)
else:
raise NotImplementedError(
f"Unsupported array element type in f-string: {ftype.element}"
)
else:
raise NotImplementedError(f"Unsupported field type in f-string: {ftype}")
@ -197,44 +206,100 @@ def _create_format_string_global(fmt_str, func, module, builder):
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,
# Special case: struct field char array needs pointer to first element
char_array_ptr = _get_struct_char_array_ptr(
expr, builder, local_sym_tab, struct_sym_tab
)
if char_array_ptr:
return char_array_ptr
# Regular expression evaluation
val, _ = eval_expr(func, module, builder, expr, local_sym_tab, None, struct_sym_tab)
if not val:
logger.warning("Failed to evaluate expression for bpf_printk, defaulting to 0")
return ir.Constant(ir.IntType(64), 0)
# Convert value to bpf_printk compatible type
if isinstance(val.type, ir.PointerType):
return _handle_pointer_arg(val, func, builder)
elif isinstance(val.type, ir.IntType):
return _handle_int_arg(val, builder)
else:
logger.warning(f"Unsupported type {val.type} in bpf_printk, defaulting to 0")
return ir.Constant(ir.IntType(64), 0)
def _get_struct_char_array_ptr(expr, builder, local_sym_tab, struct_sym_tab):
"""Get pointer to first element of char array in struct field, or None."""
if not (isinstance(expr, ast.Attribute) and isinstance(expr.value, ast.Name)):
return None
var_name = expr.value.id
field_name = expr.attr
# Check if it's a valid struct field
if not (
local_sym_tab
and var_name in local_sym_tab
and struct_sym_tab
and local_sym_tab[var_name].metadata in struct_sym_tab
):
return None
struct_type = local_sym_tab[var_name].metadata
struct_info = struct_sym_tab[struct_type]
if field_name not in struct_info.fields:
return None
field_type = struct_info.field_type(field_name)
# Check if it's a char array
is_char_array = (
isinstance(field_type, ir.ArrayType)
and isinstance(field_type.element, ir.IntType)
and field_type.element.width == 8
)
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
if not is_char_array:
return None
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."
)
# Get field pointer and GEP to first element: [N x i8]* -> i8*
struct_ptr = local_sym_tab[var_name].var
field_ptr = struct_info.gep(builder, struct_ptr, field_name)
return builder.gep(
field_ptr,
[ir.Constant(ir.IntType(32), 0), ir.Constant(ir.IntType(32), 0)],
inbounds=True,
)
def _handle_pointer_arg(val, func, builder):
"""Convert pointer type for bpf_printk."""
target, depth = get_base_type_and_depth(val.type)
if not isinstance(target, ir.IntType):
logger.warning("Only int pointers supported in bpf_printk, defaulting to 0")
return ir.Constant(ir.IntType(64), 0)
# i8* is string - use as-is
if target.width == 8 and depth == 1:
return val
# Integer pointers: dereference and sign-extend to i64
if target.width >= 32:
val = deref_to_depth(func, builder, val, depth)
return builder.sext(val, ir.IntType(64))
logger.warning("Unsupported pointer width in bpf_printk, defaulting to 0")
return ir.Constant(ir.IntType(64), 0)
def _handle_int_arg(val, builder):
"""Convert integer type for bpf_printk (sign-extend to i64)."""
if val.type.width < 64:
return builder.sext(val, ir.IntType(64))
return val

56
pythonbpf/utils.py Normal file
View File

@ -0,0 +1,56 @@
import subprocess
def trace_pipe():
"""Util to read from the trace pipe."""
try:
subprocess.run(["cat", "/sys/kernel/tracing/trace_pipe"])
except KeyboardInterrupt:
print("Tracing stopped.")
def trace_fields():
"""Parse one line from trace_pipe into fields."""
with open("/sys/kernel/tracing/trace_pipe", "rb", buffering=0) as f:
while True:
line = f.readline().rstrip()
if not line:
continue
# Skip lost event lines
if line.startswith(b"CPU:"):
continue
# Parse BCC-style: first 16 bytes = task
task = line[:16].lstrip().decode("utf-8")
line = line[17:] # Skip past task field and space
# Find the colon that ends "pid cpu flags timestamp"
ts_end = line.find(b":")
if ts_end == -1:
raise ValueError("Cannot parse trace line")
# Split "pid [cpu] flags timestamp"
try:
parts = line[:ts_end].split()
if len(parts) < 4:
raise ValueError("Not enough fields")
pid = int(parts[0])
cpu = parts[1][1:-1] # Remove brackets from [cpu]
cpu = int(cpu)
flags = parts[2]
ts = float(parts[3])
except (ValueError, IndexError):
raise ValueError("Cannot parse trace line")
# Get message: skip ": symbol:" part
line = line[ts_end + 1 :] # Skip first ":"
sym_end = line.find(b":")
if sym_end != -1:
msg = line[sym_end + 2 :].decode("utf-8") # Skip ": " after symbol
else:
msg = line.lstrip().decode("utf-8")
return (task, pid, cpu, flags, ts, msg)

View File

@ -0,0 +1,28 @@
from pythonbpf import bpf, struct, section, bpfglobal
from pythonbpf.helper import comm
from ctypes import c_void_p, c_int64
@bpf
@struct
class data_t:
comm: str(16) # type: ignore [valid-type]
copp: str(16) # type: ignore [valid-type]
@bpf
@section("tracepoint/syscalls/sys_enter_clone")
def hello(ctx: c_void_p) -> c_int64:
dataobj = data_t()
comm(dataobj.comm)
strobj = dataobj.comm
dataobj.copp = strobj
print(f"clone called by comm {dataobj.copp}")
return 0
@bpf
@bpfglobal
def LICENSE() -> str:
return "GPL"

View File

@ -0,0 +1,26 @@
from pythonbpf import bpf, struct, section, bpfglobal
from pythonbpf.helper import comm
from ctypes import c_void_p, c_int64
@bpf
@struct
class data_t:
comm: str(16) # type: ignore [valid-type]
@bpf
@section("tracepoint/syscalls/sys_enter_clone")
def hello(ctx: c_void_p) -> c_int64:
dataobj = data_t()
comm(dataobj.comm)
strobj = dataobj.comm
print(f"clone called by comm {strobj}")
return 0
@bpf
@bpfglobal
def LICENSE() -> str:
return "GPL"