1
0
mirror of https://github.com/trezor/trezor-firmware.git synced 2024-11-23 16:08:17 +00:00
trezor-firmware/trezorlib/protocol_v1.py

94 lines
3.1 KiB
Python

# This file is part of the TREZOR project.
#
# Copyright (C) 2012-2016 Marek Palatinus <slush@satoshilabs.com>
# Copyright (C) 2012-2016 Pavol Rusnak <stick@satoshilabs.com>
#
# This library is free software: you can redistribute it and/or modify
# it under the terms of the GNU Lesser General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This library is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU Lesser General Public License for more details.
#
# You should have received a copy of the GNU Lesser General Public License
# along with this library. If not, see <http://www.gnu.org/licenses/>.
from __future__ import absolute_import
from io import BytesIO
import logging
import struct
from typing import Tuple, Type
from . import mapping
from . import protobuf
from .transport import Transport
REPLEN = 64
LOG = logging.getLogger(__name__)
class ProtocolV1:
def session_begin(self, transport: Transport) -> None:
pass
def session_end(self, transport: Transport) -> None:
pass
def write(self, transport: Transport, msg: protobuf.MessageType) -> None:
LOG.debug("sending message: {}".format(msg.__class__.__name__),
extra={'protobuf': msg})
data = BytesIO()
protobuf.dump_message(data, msg)
ser = data.getvalue()
header = struct.pack(">HL", mapping.get_type(msg), len(ser))
data = bytearray(b"##" + header + ser)
while data:
# Report ID, data padded to 63 bytes
chunk = b'?' + data[:REPLEN - 1]
chunk = chunk.ljust(REPLEN, b'\x00')
transport.write_chunk(chunk)
data = data[63:]
def read(self, transport: Transport) -> protobuf.MessageType:
# Read header with first part of message data
chunk = transport.read_chunk()
msg_type, datalen, data = self.parse_first(chunk)
# Read the rest of the message
while len(data) < datalen:
chunk = transport.read_chunk()
data.extend(self.parse_next(chunk))
# Strip padding
data = BytesIO(data[:datalen])
# Parse to protobuf
msg = protobuf.load_message(data, mapping.get_class(msg_type))
LOG.debug("received message: {}".format(msg.__class__.__name__),
extra={'protobuf': msg})
return msg
def parse_first(self, chunk: bytes) -> Tuple[int, int, bytes]:
if chunk[:3] != b'?##':
raise RuntimeError('Unexpected magic characters')
try:
headerlen = struct.calcsize('>HL')
msg_type, datalen = struct.unpack('>HL', chunk[3:3 + headerlen])
except:
raise RuntimeError('Cannot parse header')
data = chunk[3 + headerlen:]
return msg_type, datalen, data
def parse_next(self, chunk: bytes) -> bytes:
if chunk[:1] != b'?':
raise RuntimeError('Unexpected magic characters')
return chunk[1:]