path: root/net/test/bpf.py

#!/usr/bin/python3
#
# Copyright 2016 The Android Open Source Project
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

"""kernel net test library for bpf testing."""

import ctypes
import os
import resource
import socket
import sys

import csocket
import cstruct
import net_test

# __NR_bpf syscall numbers for various architectures.
# NOTE: If python inherited COMPAT_UTS_MACHINE, uname's 'machine' field will
# return the 32-bit architecture name, even if python itself is 64-bit. To work
# around this problem and pick the right syscall nr, we can additionally check
# the bitness of the python interpreter. Assume that the 64-bit architectures
# are not running with COMPAT_UTS_MACHINE and must be 64-bit at all times.
__NR_bpf = {  # pylint: disable=invalid-name
    "aarch64-32bit": 386,
    "aarch64-64bit": 280,
    "armv7l-32bit": 386,
    "armv8l-32bit": 386,
    "armv8l-64bit": 280,
    "i686-32bit": 357,
    "i686-64bit": 321,
    "x86_64-32bit": 357,
    "x86_64-64bit": 321,
    "riscv64-64bit": 280,
}[os.uname()[4] + "-" + ("64" if sys.maxsize > 0x7FFFFFFF else "32") + "bit"]

# After ACK merge of 5.10.168 is when support for this was backported from
# upstream Linux 5.14 and was merged into ACK android{12,13}-5.10 branches.
#   ACK android12-5.10 was >= 5.10.168 without this support only for ~4.5 hours
#   ACK android13-4.10 was >= 5.10.168 without this support only for ~25 hours
# as such we can >= 5.10.168 instead of > 5.10.168
HAVE_SO_NETNS_COOKIE = net_test.LINUX_VERSION >= (5, 10, 168)

# Note: This is *not* correct for parisc & sparc architectures
SO_NETNS_COOKIE = 71

LOG_LEVEL = 1
LOG_SIZE = 65536

# BPF syscall commands constants.
BPF_MAP_CREATE = 0
BPF_MAP_LOOKUP_ELEM = 1
BPF_MAP_UPDATE_ELEM = 2
BPF_MAP_DELETE_ELEM = 3
BPF_MAP_GET_NEXT_KEY = 4
BPF_PROG_LOAD = 5
BPF_OBJ_PIN = 6
BPF_OBJ_GET = 7
BPF_PROG_ATTACH = 8
BPF_PROG_DETACH = 9
SO_ATTACH_BPF = 50

# BPF map type constant.
BPF_MAP_TYPE_UNSPEC = 0
BPF_MAP_TYPE_HASH = 1
BPF_MAP_TYPE_ARRAY = 2
BPF_MAP_TYPE_PROG_ARRAY = 3
BPF_MAP_TYPE_PERF_EVENT_ARRAY = 4

# BPF program type constant.
BPF_PROG_TYPE_UNSPEC = 0
BPF_PROG_TYPE_SOCKET_FILTER = 1
BPF_PROG_TYPE_KPROBE = 2
BPF_PROG_TYPE_SCHED_CLS = 3
BPF_PROG_TYPE_SCHED_ACT = 4
BPF_PROG_TYPE_TRACEPOINT = 5
BPF_PROG_TYPE_XDP = 6
BPF_PROG_TYPE_PERF_EVENT = 7
BPF_PROG_TYPE_CGROUP_SKB = 8
BPF_PROG_TYPE_CGROUP_SOCK = 9

# BPF program attach type.
BPF_CGROUP_INET_INGRESS = 0
BPF_CGROUP_INET_EGRESS = 1
BPF_CGROUP_INET_SOCK_CREATE = 2

# BPF register constant
BPF_REG_0 = 0
BPF_REG_1 = 1
BPF_REG_2 = 2
BPF_REG_3 = 3
BPF_REG_4 = 4
BPF_REG_5 = 5
BPF_REG_6 = 6
BPF_REG_7 = 7
BPF_REG_8 = 8
BPF_REG_9 = 9
BPF_REG_10 = 10

# BPF instruction constants
BPF_PSEUDO_MAP_FD = 1
BPF_LD = 0x00
BPF_LDX = 0x01
BPF_ST = 0x02
BPF_STX = 0x03
BPF_ALU = 0x04
BPF_JMP = 0x05
BPF_RET = 0x06
BPF_MISC = 0x07
BPF_W = 0x00
BPF_H = 0x08
BPF_B = 0x10
BPF_IMM = 0x00
BPF_ABS = 0x20
BPF_IND = 0x40
BPF_MEM = 0x60
BPF_LEN = 0x80
BPF_MSH = 0xa0
BPF_ADD = 0x00
BPF_SUB = 0x10
BPF_MUL = 0x20
BPF_DIV = 0x30
BPF_OR = 0x40
BPF_AND = 0x50
BPF_LSH = 0x60
BPF_RSH = 0x70
BPF_NEG = 0x80
BPF_MOD = 0x90
BPF_XOR = 0xa0
BPF_JA = 0x00
BPF_JEQ = 0x10
BPF_JGT = 0x20
BPF_JGE = 0x30
BPF_JSET = 0x40
BPF_K = 0x00
BPF_X = 0x08
BPF_ALU64 = 0x07
BPF_DW = 0x18
BPF_XADD = 0xc0
BPF_MOV = 0xb0

BPF_ARSH = 0xc0
BPF_END = 0xd0
BPF_TO_LE = 0x00
BPF_TO_BE = 0x08

BPF_JNE = 0x50
BPF_JSGT = 0x60

BPF_JSGE = 0x70
BPF_CALL = 0x80
BPF_EXIT = 0x90

# BPF helper function constants
# pylint: disable=invalid-name
BPF_FUNC_unspec = 0
BPF_FUNC_map_lookup_elem = 1
BPF_FUNC_map_update_elem = 2
BPF_FUNC_map_delete_elem = 3
BPF_FUNC_ktime_get_ns = 5
BPF_FUNC_get_current_uid_gid = 15
BPF_FUNC_skb_change_head = 43
BPF_FUNC_get_socket_cookie = 46
BPF_FUNC_get_socket_uid = 47
BPF_FUNC_ktime_get_boot_ns = 125
# pylint: enable=invalid-name

BPF_F_RDONLY = 1 << 3
BPF_F_WRONLY = 1 << 4

#  These object below belongs to the same kernel union and the types below
#  (e.g., bpf_attr_create) aren't kernel struct names but just different
#  variants of the union.
# pylint: disable=invalid-name
BpfAttrCreate = cstruct.Struct(
    "bpf_attr_create", "=IIIII",
    "map_type key_size value_size max_entries, map_flags")
BpfAttrOps = cstruct.Struct(
    "bpf_attr_ops", "=QQQQ",
    "map_fd key_ptr value_ptr flags")
BpfAttrProgLoad = cstruct.Struct(
    "bpf_attr_prog_load", "=IIQQIIQI", "prog_type insn_cnt insns"
    " license log_level log_size log_buf kern_version")
BpfAttrProgAttach = cstruct.Struct(
    "bpf_attr_prog_attach", "=III", "target_fd attach_bpf_fd attach_type")
BpfInsn = cstruct.Struct("bpf_insn", "=BBhi", "code dst_src_reg off imm")
# pylint: enable=invalid-name

libc = ctypes.CDLL(ctypes.util.find_library("c"), use_errno=True)
HAVE_EBPF_4_19 = net_test.LINUX_VERSION >= (4, 19, 0)
HAVE_EBPF_5_4 = net_test.LINUX_VERSION >= (5, 4, 0)

# set memlock resource 1 GiB
resource.setrlimit(resource.RLIMIT_MEMLOCK, (1073741824, 1073741824))


# BPF program syscalls
def BpfSyscall(op, attr):
  ret = libc.syscall(__NR_bpf, op, csocket.VoidPointer(attr), len(attr))
  csocket.MaybeRaiseSocketError(ret)
  return ret


def CreateMap(map_type, key_size, value_size, max_entries, map_flags=0):
  attr = BpfAttrCreate((map_type, key_size, value_size, max_entries, map_flags))
  return BpfSyscall(BPF_MAP_CREATE, attr)


def UpdateMap(map_fd, key, value, flags=0):
  c_value = ctypes.c_uint32(value)
  c_key = ctypes.c_uint32(key)
  value_ptr = ctypes.addressof(c_value)
  key_ptr = ctypes.addressof(c_key)
  attr = BpfAttrOps((map_fd, key_ptr, value_ptr, flags))
  BpfSyscall(BPF_MAP_UPDATE_ELEM, attr)


def LookupMap(map_fd, key):
  c_value = ctypes.c_uint32(0)
  c_key = ctypes.c_uint32(key)
  attr = BpfAttrOps(
      (map_fd, ctypes.addressof(c_key), ctypes.addressof(c_value), 0))
  BpfSyscall(BPF_MAP_LOOKUP_ELEM, attr)
  return c_value


def GetNextKey(map_fd, key):
  """Get the next key in the map after the specified key."""
  if key is not None:
    c_key = ctypes.c_uint32(key)
    c_next_key = ctypes.c_uint32(0)
    key_ptr = ctypes.addressof(c_key)
  else:
    key_ptr = 0
  c_next_key = ctypes.c_uint32(0)
  attr = BpfAttrOps(
      (map_fd, key_ptr, ctypes.addressof(c_next_key), 0))
  BpfSyscall(BPF_MAP_GET_NEXT_KEY, attr)
  return c_next_key


def GetFirstKey(map_fd):
  return GetNextKey(map_fd, None)


def DeleteMap(map_fd, key):
  c_key = ctypes.c_uint32(key)
  attr = BpfAttrOps((map_fd, ctypes.addressof(c_key), 0, 0))
  BpfSyscall(BPF_MAP_DELETE_ELEM, attr)


def BpfProgLoad(prog_type, instructions, prog_license=b"GPL"):
  bpf_prog = b"".join(instructions)
  insn_buff = ctypes.create_string_buffer(bpf_prog)
  gpl_license = ctypes.create_string_buffer(prog_license)
  log_buf = ctypes.create_string_buffer(b"", LOG_SIZE)
  attr = BpfAttrProgLoad((prog_type, len(insn_buff) // len(BpfInsn),
                          ctypes.addressof(insn_buff),
                          ctypes.addressof(gpl_license), LOG_LEVEL,
                          LOG_SIZE, ctypes.addressof(log_buf), 0))
  return BpfSyscall(BPF_PROG_LOAD, attr)


# Attach a socket eBPF filter to a target socket
def BpfProgAttachSocket(sock_fd, prog_fd):
  uint_fd = ctypes.c_uint32(prog_fd)
  ret = libc.setsockopt(sock_fd, socket.SOL_SOCKET, SO_ATTACH_BPF,
                        ctypes.pointer(uint_fd), ctypes.sizeof(uint_fd))
  csocket.MaybeRaiseSocketError(ret)


# Attach a eBPF filter to a cgroup
def BpfProgAttach(prog_fd, target_fd, prog_type):
  attr = BpfAttrProgAttach((target_fd, prog_fd, prog_type))
  return BpfSyscall(BPF_PROG_ATTACH, attr)


# Detach a eBPF filter from a cgroup
def BpfProgDetach(target_fd, prog_type):
  attr = BpfAttrProgAttach((target_fd, 0, prog_type))
  return BpfSyscall(BPF_PROG_DETACH, attr)


# BPF program command constructors
def BpfMov64Reg(dst, src):
  code = BPF_ALU64 | BPF_MOV | BPF_X
  dst_src = src << 4 | dst
  ret = BpfInsn((code, dst_src, 0, 0))
  return ret.Pack()


def BpfLdxMem(size, dst, src, off):
  code = BPF_LDX | (size & 0x18) | BPF_MEM
  dst_src = src << 4 | dst
  ret = BpfInsn((code, dst_src, off, 0))
  return ret.Pack()


def BpfStxMem(size, dst, src, off):
  code = BPF_STX | (size & 0x18) | BPF_MEM
  dst_src = src << 4 | dst
  ret = BpfInsn((code, dst_src, off, 0))
  return ret.Pack()


def BpfStMem(size, dst, off, imm):
  code = BPF_ST | (size & 0x18) | BPF_MEM
  dst_src = dst
  ret = BpfInsn((code, dst_src, off, imm))
  return ret.Pack()


def BpfAlu64Imm(op, dst, imm):
  code = BPF_ALU64 | (op & 0xf0) | BPF_K
  dst_src = dst
  ret = BpfInsn((code, dst_src, 0, imm))
  return ret.Pack()


def BpfJumpImm(op, dst, imm, off):
  code = BPF_JMP | (op & 0xf0) | BPF_K
  dst_src = dst
  ret = BpfInsn((code, dst_src, off, imm))
  return ret.Pack()


def BpfRawInsn(code, dst, src, off, imm):
  ret = BpfInsn((code, (src << 4 | dst), off, imm))
  return ret.Pack()


def BpfMov64Imm(dst, imm):
  code = BPF_ALU64 | BPF_MOV | BPF_K
  dst_src = dst
  ret = BpfInsn((code, dst_src, 0, imm))
  return ret.Pack()


def BpfExitInsn():
  code = BPF_JMP | BPF_EXIT
  ret = BpfInsn((code, 0, 0, 0))
  return ret.Pack()


def BpfLoadMapFd(map_fd, dst):
  code = BPF_LD | BPF_DW | BPF_IMM
  dst_src = BPF_PSEUDO_MAP_FD << 4 | dst
  insn1 = BpfInsn((code, dst_src, 0, map_fd))
  insn2 = BpfInsn((0, 0, 0, map_fd >> 32))
  return insn1.Pack() + insn2.Pack()


def BpfFuncCall(func):
  code = BPF_JMP | BPF_CALL
  dst_src = 0
  ret = BpfInsn((code, dst_src, 0, func))
  return ret.Pack()