2
0
mirror of https://github.com/xcat2/confluent.git synced 2026-05-01 04:47:45 +00:00
Files
confluent/confluent_server/aiohmi/ipmi/console.py
Jarrod Johnson bfc27595dc Fold aiohmi into confluent
If someone asks for it independently, we can break it out again.  But for now,
assume it's only for confluent.
2026-04-30 08:48:24 -04:00

552 lines
23 KiB
Python

# Copyright 2014 IBM Corporation
# Copyright 2015-2019 Lenovo
#
# 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.
#
"""This represents the low layer message framing portion of IPMI"""
import struct
import threading
import aiohmi.exceptions as exc
from aiohmi.ipmi.private import constants
from aiohmi.ipmi.private import session
from aiohmi.ipmi.private.util import _monotonic_time
class Console(object):
"""IPMI SOL class.
This object represents an SOL channel, multiplexing SOL data with
commands issued by ipmi.command.
:param bmc: hostname or ip address of BMC
:param userid: username to use to connect
:param password: password to connect to the BMC
:param iohandler: Either a function to call with bytes, a filehandle to
use for input and output, or a tuple of (input, output)
handles
:param force: Set to True to force on or False to force off
:param kg: optional parameter for BMCs configured to require it
"""
# TODO(jbjohnso): still need an exit and a data callin function
def __init__(self, bmc, userid, password,
iohandler, port=623,
force=False, kg=None):
self.outputlock = threading.RLock()
self.keepaliveid = None
self.connected = False
self.broken = False
self.out_handler = iohandler
self.remseq = 0
self.myseq = 0
self.lastsize = 0
self.retriedpayload = 0
self.pendingoutput = []
self.awaitingack = False
self.activated = False
self.force_session = force
self.port = port
self.ipmi_session = None
self.callgotsession = None
self.bmc = bmc
self.userid = userid
self.password = password
self.port = port
self.kg = kg
self.broken = False
async def connect(self):
bmc = self.bmc
userid = self.userid
password = self.password
port = self.port
kg = self.kg
self.ipmi_session = await session.Session(
bmc=bmc, userid=userid, password=password, port=port, kg=kg)
# induce one iteration of the loop, now that we would be
# prepared for it in theory
await self._got_session({})
async def _got_session(self, response):
"""Private function to navigate SOL payload activation"""
if 'error' in response:
await self._print_error(response['error'])
return
if not self.ipmi_session:
self.callgotsession = response
return
# Send activate sol payload directive
# netfn= 6 (application)
# command = 0x48 (activate payload)
# data = (1, sol payload type
# 1, first instance
# 0b11000000, -encrypt, authenticate,
# disable serial/modem alerts, CTS fine
# 0, 0, 0 reserved
response = await self.ipmi_session.raw_command(netfn=0x6, command=0x48,
data=(1, 1, 192, 0, 0, 0))
# given that these are specific to the command,
# it's probably best if one can grep the error
# here instead of in constants
sol_activate_codes = {
0x81: 'SOL is disabled',
0x82: 'Maximum SOL session count reached',
0x83: 'Cannot activate payload with encryption',
0x84: 'Cannot activate payload without encryption',
}
if 'code' in response and response['code']:
if response['code'] in constants.ipmi_completion_codes:
await self._print_error(
constants.ipmi_completion_codes[response['code']])
return
elif response['code'] == 0x80:
if self.force_session and not self.retriedpayload:
self.retriedpayload = 1
sessrsp = await self.ipmi_session.raw_command(
netfn=0x6,
command=0x49,
data=(1, 1, 0, 0, 0, 0))
await self._got_session(sessrsp)
return
else:
await self._print_error('SOL Session active for another client')
return
elif response['code'] in sol_activate_codes:
await self._print_error(sol_activate_codes[response['code']])
return
else:
await self._print_error(
'SOL encountered Unrecognized error code %d' %
response['code'])
return
if 'error' in response:
self._print_error(response['error'])
return
self.activated = True
# data[0:3] is reserved except for the test mode, which we don't use
data = response['data']
self.maxoutcount = (data[5] << 8) + data[4]
# BMC tells us this is the maximum allowed size
# data[6:7] is the promise of how small packets are going to be, but we
# don't have any reason to worry about it
# some BMCs disagree on the endianness, so do both
valid_ports = (self.port, struct.unpack(
'<H', struct.pack('>H', self.port))[0])
if (data[8] + (data[9] << 8)) not in valid_ports:
# TODO(jbjohnso): support atypical SOL port number
raise NotImplementedError("Non-standard SOL Port Number")
# ignore data[10:11] for now, the vlan detail, shouldn't matter to this
# code anyway...
# NOTE(jbjohnso):
# We will use a special purpose keepalive
if self.ipmi_session.sol_handler is not None:
# If there is erroneously another SOL handler already, notify
# it of newly established session
await self.ipmi_session.sol_handler({'error': 'Session Disconnected'})
self.keepaliveid = self.ipmi_session.register_keepalive(
cmd={'netfn': 6, 'command': 0x4b, 'data': (1, 1)},
callback=self._got_payload_instance_info)
self.ipmi_session.sol_handler = self._got_sol_payload
self.connected = True
# self._sendpendingoutput() checks len(self._sendpendingoutput)
await self._sendpendingoutput()
async def _got_payload_instance_info(self, response):
if 'error' in response:
self.activated = False
await self._print_error(response['error'])
return
currowner = struct.unpack(
"<I", struct.pack('4B', *response['data'][:4]))
if currowner[0] != self.ipmi_session.sessionid:
# the session is deactivated or active for something else
self.activated = False
await self._print_error('SOL deactivated')
return
# ok, still here, that means session is alive, but another
# common issue is firmware messing with mux on reboot
# this would be a nice thing to check, but the serial channel
# number is needed and there isn't an obvious means to reliably
# discern which channel or even *if* the serial port in question
# correlates at all to an ipmi channel to check mux
def _addpendingdata(self, data):
with self.outputlock:
if isinstance(data, dict):
self.pendingoutput.append(data)
else: # it is a text situation
if (len(self.pendingoutput) == 0
or isinstance(self.pendingoutput[-1], dict)):
self.pendingoutput.append(data)
else:
self.pendingoutput[-1] += data
def _got_cons_input(self, handle):
"""Callback for handle events detected by ipmi session"""
self._addpendingdata(handle.read())
if not self.awaitingack:
self._sendpendingoutput()
async def close(self):
"""Shut down an SOL session"""
if self.ipmi_session:
self.ipmi_session.unregister_keepalive(self.keepaliveid)
if self.activated and self.ipmi_session is not None:
try:
await self.ipmi_session.raw_command(netfn=6, command=0x49,
data=(1, 1, 0, 0, 0, 0))
except exc.IpmiException:
# if underlying ipmi session is not working, then
# run with the implicit success
pass
async def send_data(self, data):
if self.broken:
return
self._addpendingdata(data)
if not self.connected:
return
if not self.awaitingack:
await self._sendpendingoutput()
async def send_break(self):
self._addpendingdata({'break': 1})
if not self.connected:
return
if not self.awaitingack:
await self._sendpendingoutput()
@classmethod
def wait_for_rsp(cls, timeout):
"""Delay for no longer than timeout for next response.
This acts like a sleep that exits on activity.
:param timeout: Maximum number of seconds before returning
"""
return session.Session.wait_for_rsp(timeout=timeout)
async def _sendpendingoutput(self):
with self.outputlock:
dobreak = False
chunk = ''
if len(self.pendingoutput) == 0:
return
if isinstance(self.pendingoutput[0], dict):
if 'break' in self.pendingoutput[0]:
dobreak = True
else:
del self.pendingoutput[0]
raise ValueError
del self.pendingoutput[0]
elif len(self.pendingoutput[0]) > self.maxoutcount:
chunk = self.pendingoutput[0][:self.maxoutcount]
self.pendingoutput[0] = self.pendingoutput[0][
self.maxoutcount:]
else:
chunk = self.pendingoutput[0]
del self.pendingoutput[0]
await self._sendoutput(chunk, sendbreak=dobreak)
async def _sendoutput(self, output, sendbreak=False):
self.myseq += 1
self.myseq &= 0xf
if self.myseq == 0:
self.myseq = 1
# currently we don't try to combine ack with outgoing data
# so we use 0 for ack sequence number and accepted character
# count
breakbyte = 0
if sendbreak:
breakbyte = 0b10000
try:
payload = bytearray((self.myseq, 0, 0, breakbyte)) + output
except TypeError: # bytearray hits unicode...
payload = bytearray((self.myseq, 0, 0, breakbyte
)) + output.encode('utf8')
self.lasttextsize = len(output)
needskeepalive = False
if self.lasttextsize == 0:
needskeepalive = True
self.awaitingack = True
self.lastpayload = payload
await self.send_payload(payload, retry=False, needskeepalive=needskeepalive)
retries = 5
while retries and self.awaitingack:
expiry = _monotonic_time() + 5.5 - retries
while self.awaitingack and _monotonic_time() < expiry:
await self.wait_for_rsp(0.5)
if self.awaitingack:
await self.send_payload(payload, retry=False,
needskeepalive=needskeepalive)
retries -= 1
if not retries:
await self._print_error('Connection lost')
async def send_payload(self, payload, payload_type=1, retry=True,
needskeepalive=False):
while not (self.connected or self.broken):
session.Session.wait_for_rsp(timeout=10)
if self.ipmi_session is None or not self.ipmi_session.logged:
await self._print_error('Session no longer connected')
raise exc.IpmiException('Session no longer connected')
await self.ipmi_session.send_payload(payload,
payload_type=payload_type,
retry=retry,
needskeepalive=needskeepalive)
async def _print_info(self, info):
await self._print_data({'info': info})
async def _print_error(self, error):
self.broken = True
if self.ipmi_session:
self.ipmi_session.unregister_keepalive(self.keepaliveid)
if (self.ipmi_session.sol_handler
and self.ipmi_session.sol_handler.__self__ is self):
self.ipmi_session.sol_handler = None
self.ipmi_session = None
if type(error) == dict:
await self._print_data(error)
else:
await self._print_data({'error': error})
async def _print_data(self, data):
"""Convey received data back to caller in the format of their choice.
Caller may elect to provide this class filehandle(s) or else give a
callback function that this class will use to convey data back to
caller.
"""
await self.out_handler(data)
async def _got_sol_payload(self, payload):
"""SOL payload callback"""
# TODO(jbjohnso) test cases to throw some likely scenarios at functions
# for example, retry with new data, retry with no new data
# retry with unexpected sequence number
if type(payload) == dict: # we received an error condition
self.activated = False
await self._print_error(payload)
return
newseq = payload[0] & 0b1111
ackseq = payload[1] & 0b1111
ackcount = payload[2]
nacked = payload[3] & 0b1000000
poweredoff = payload[3] & 0b100000
deactivated = payload[3] & 0b10000
breakdetected = payload[3] & 0b100
# for now, ignore overrun. I assume partial NACK for this reason or
# for no reason would be treated the same, new payload with partial
# data.
remdata = ""
remdatalen = 0
if newseq != 0: # this packet at least has some data to send to us..
if len(payload) > 4:
remdatalen = len(payload[4:]) # store remote len before dupe
# retry logic, we must ack *this* many even if it is
# a retry packet with new partial data
remdata = bytes(payload[4:])
if newseq == self.remseq: # it is a retry, but could have new data
if remdatalen > self.lastsize:
remdata = bytes(remdata[4 + self.lastsize:])
else: # no new data...
remdata = ""
else: # TODO(jbjohnso) what if remote sequence number is wrong??
self.remseq = newseq
self.lastsize = remdatalen
if remdata: # Do not subject callers to empty data
await self._print_data(remdata)
ackpayload = bytearray((0, self.remseq, remdatalen, 0))
# Why not put pending data into the ack? because it's rare
# and might be hard to decide what to do in the context of
# retry situation
try:
await self.send_payload(ackpayload, retry=False)
except exc.IpmiException:
# if the session is broken, then close the SOL session
self.close()
if self.myseq != 0 and ackseq == self.myseq: # the bmc has something
# to say about last xmit
self.awaitingack = False
if nacked and not breakdetected: # the BMC was in some way unhappy
if poweredoff:
await self._print_info("Remote system is powered down")
if deactivated:
self.activated = False
await self._print_error("Remote IPMI console disconnected")
else: # retry all or part of packet, but in a new form
# also add pending output for efficiency and ease
newtext = self.lastpayload[4 + ackcount:]
with self.outputlock:
if (self.pendingoutput
and not isinstance(self.pendingoutput[0],
dict)):
self.pendingoutput[0] = \
newtext + self.pendingoutput[0]
else:
self.pendingoutput = [newtext] + self.pendingoutput
# self._sendpendingoutput() checks len(self._sendpendingoutput)
await self._sendpendingoutput()
elif ackseq != 0 and self.awaitingack:
# if an ack packet came in, but did not match what we
# expected, retry our payload now.
# the situation that was triggered was a senseless retry
# when data came in while we xmitted. In theory, a BMC
# should handle a retry correctly, but some do not, so
# try to mitigate by avoiding overeager retries
# occasional retry of a packet
# sooner than timeout suggests is evidently a big deal
await self.send_payload(payload=self.lastpayload, retry=False)
def main_loop(self):
"""Process all events until no more sessions exist.
If a caller is a simple little utility, provide a function to
eternally run the event loop. More complicated usage would be expected
to provide their own event loop behavior, though this could be used
within the greenthread implementation of caller's choice if desired.
"""
# wait_for_rsp promises to return a false value when no sessions are
# alive anymore
# TODO(jbjohnso): wait_for_rsp is not returning a true value for our
# own session
while (1):
session.Session.wait_for_rsp(timeout=600)
class ServerConsole(Console):
"""IPMI SOL class.
This object represents an SOL channel, multiplexing SOL data with
commands issued by ipmi.command.
:param session: IPMI session
:param iohandler: I/O handler
"""
def __init__(self, _session, iohandler, force=False):
self.outputlock = threading.RLock()
self.keepaliveid = None
self.connected = True
self.broken = False
self.out_handler = iohandler
self.remseq = 0
self.myseq = 0
self.lastsize = 0
self.retriedpayload = 0
self.pendingoutput = []
self.awaitingack = False
self.activated = True
self.force_session = force
self.ipmi_session = _session
self.ipmi_session.sol_handler = self._got_sol_payload
self.maxoutcount = 256
self.poweredon = True
session.Session.wait_for_rsp(0)
async def _got_sol_payload(self, payload):
"""SOL payload callback"""
# TODO(jbjohnso) test cases to throw some likely scenarios at functions
# for example, retry with new data, retry with no new data
# retry with unexpected sequence number
if type(payload) == dict: # we received an error condition
self.activated = False
await self._print_error(payload)
return
newseq = payload[0] & 0b1111
ackseq = payload[1] & 0b1111
ackcount = payload[2]
nacked = payload[3] & 0b1000000
breakdetected = payload[3] & 0b10000
# for now, ignore overrun. I assume partial NACK for this reason or
# for no reason would be treated the same, new payload with partial
# data.
remdata = ""
remdatalen = 0
flag = 0
if not self.poweredon:
flag |= 0b1100000
if not self.activated:
flag |= 0b1010000
if newseq != 0: # this packet at least has some data to send to us..
if len(payload) > 4:
remdatalen = len(payload[4:]) # store remote len before dupe
# retry logic, we must ack *this* many even if it is
# a retry packet with new partial data
remdata = bytes(payload[4:])
if newseq == self.remseq: # it is a retry, but could have new data
if remdatalen > self.lastsize:
remdata = bytes(remdata[4 + self.lastsize:])
else: # no new data...
remdata = ""
else: # TODO(jbjohnso) what if remote sequence number is wrong??
self.remseq = newseq
self.lastsize = remdatalen
ackpayload = bytearray((0, self.remseq, remdatalen, flag))
# Why not put pending data into the ack? because it's rare
# and might be hard to decide what to do in the context of
# retry situation
try:
self.send_payload(ackpayload, retry=False)
except exc.IpmiException:
# if the session is broken, then close the SOL session
self.close()
if remdata: # Do not subject callers to empty data
await self._print_data(remdata)
if self.myseq != 0 and ackseq == self.myseq: # the bmc has something
# to say about last xmit
self.awaitingack = False
if nacked and not breakdetected: # the BMC was in some way unhappy
newtext = self.lastpayload[4 + ackcount:]
with self.outputlock:
if (self.pendingoutput
and not isinstance(self.pendingoutput[0], dict)):
self.pendingoutput[0] = newtext + self.pendingoutput[0]
else:
self.pendingoutput = [newtext] + self.pendingoutput
# self._sendpendingoutput() checks len(self._sendpendingoutput)
self._sendpendingoutput()
elif ackseq != 0 and self.awaitingack:
# if an ack packet came in, but did not match what we
# expected, retry our payload now.
# the situation that was triggered was a senseless retry
# when data came in while we xmitted. In theory, a BMC
# should handle a retry correctly, but some do not, so
# try to mitigate by avoiding overeager retries
# occasional retry of a packet
# sooner than timeout suggests is evidently a big deal
self.send_payload(payload=self.lastpayload)
def send_payload(self, payload, payload_type=1, retry=True,
needskeepalive=False):
while not (self.connected or self.broken):
session.Session.wait_for_rsp(timeout=10)
self.ipmi_session.send_payload(payload,
payload_type=payload_type,
retry=retry,
needskeepalive=needskeepalive)
def close(self):
"""Shut down an SOL session"""
self.activated = False