mdw@git.distorted.org.uk Git - hippotat/blob - hippotat

   1 #!/usr/bin/python3
   2 #
   3 # Hippotat - Asinine IP Over HTTP program
   4 # ./hippotat - client main program
   5 #
   6 # Copyright 2017 Ian Jackson
   7 #
   8 # GPLv3+
   9 #
  10 #    This program is free software: you can redistribute it and/or modify
  11 #    it under the terms of the GNU General Public License as published by
  12 #    the Free Software Foundation, either version 3 of the License, or
  13 #    (at your option) any later version.
  14 #
  15 #    This program is distributed in the hope that it will be useful,
  16 #    but WITHOUT ANY WARRANTY; without even the implied warranty of
  17 #    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  18 #    GNU General Public License for more details.
  19 #
  20 #    You should have received a copy of the GNU General Public License
  21 #    along with this program, in the file GPLv3.  If not,
  22 #    see <http://www.gnu.org/licenses/>.
  23
  24 #@ import sys; sys.path.append('@PYBUILD_INSTALL_DIR@')
  25 from hippotatlib import *
  26
  27 import twisted.web
  28 import twisted.web.client
  29 import urllib.parse
  30
  31 import io
  32
  33 class GeneralResponseConsumer(twisted.internet.protocol.Protocol):
  34   def __init__(self, cl, req, resp, desc):
  35     self._cl = cl
  36     self._req = req
  37     self._resp = resp
  38     self._desc = desc
  39
  40   def _log(self, dflag, msg, **kwargs):
  41     self._cl.log(dflag, '%s: %s' % (self._desc, msg), idof=self._req, **kwargs)
  42
  43   def connectionMade(self):
  44     self._log(DBG.HTTP_CTRL, 'connectionMade')
  45
  46   def connectionLostOK(self, reason):
  47     return (reason.check(twisted.web.client.ResponseDone) or
  48             reason.check(twisted.web.client.PotentialDataLoss))
  49     # twisted.web.client.PotentialDataLoss is an entirely daft
  50     # exception.  It will occur every time if the origin server does
  51     # not provide a Content-Length.  (hippotatd does, of course, but
  52     # the HTTP transaction might be proxied.)
  53
  54 class ResponseConsumer(GeneralResponseConsumer):
  55   def __init__(self, cl, req, resp):
  56     super().__init__(cl, req, resp, 'RC')
  57     ssddesc = '[%s] %s' % (id(req), self._desc)
  58     self._ssd = SlipStreamDecoder(ssddesc, partial(queue_inbound, cl.ipif))
  59     self._log(DBG.HTTP_CTRL, '__init__')
  60
  61   def dataReceived(self, data):
  62     self._log(DBG.HTTP, 'dataReceived', d=data)
  63     try:
  64       self._ssd.inputdata(data)
  65     except Exception as e:
  66       self._handleexception()
  67
  68   def connectionLost(self, reason):
  69     reason_msg = 'connectionLost ' + str(reason)
  70     self._log(DBG.HTTP_CTRL, reason_msg)
  71     if not self.connectionLostOK(reason):
  72       self._latefailure(reason_msg)
  73       return
  74     try:
  75       self._log(DBG.HTTP, 'ResponseDone')
  76       self._ssd.flush()
  77       self._cl.req_fin(self._req)
  78     except Exception as e:
  79       self._handleexception()
  80     self._cl.report_running()
  81
  82   def _handleexception(self):
  83     self._latefailure(traceback.format_exc())
  84
  85   def _latefailure(self, reason):
  86     self._log(DBG.HTTP_CTRL, '_latefailure ' + str(reason))
  87     self._cl.req_err(self._req, reason)
  88
  89 class ErrorResponseConsumer(GeneralResponseConsumer):
  90   def __init__(self, cl, req, resp):
  91     super().__init__(cl, req, resp, 'ERROR-RC')
  92     self._m = b''
  93     try:
  94       self._phrase = resp.phrase.decode('utf-8')
  95     except Exception:
  96       self._phrase = repr(resp.phrase)
  97     self._log(DBG.HTTP_CTRL, '__init__ %d %s' % (resp.code, self._phrase))
  98
  99   def dataReceived(self, data):
 100     self._log(DBG.HTTP_CTRL, 'dataReceived ' + repr(data))
 101     self._m += data
 102
 103   def connectionLost(self, reason):
 104     try:
 105       mbody = self._m.decode('utf-8')
 106     except Exception:
 107       mbody = repr(self._m)
 108     if not self.connectionLostOK(reason):
 109       mbody += ' || ' + str(reason)
 110     self._cl.req_err(self._req,
 111             "FAILED %d %s | %s"
 112             % (self._resp.code, self._phrase, mbody))
 113
 114 class Client():
 115   def __init__(cl, c,ss,cs):
 116     cl.c = c
 117     cl.outstanding = { }
 118     cl.desc = '[%s %s] ' % (ss,cs)
 119     cl.running_reported = False
 120     cl.log_info('setting up')
 121
 122   def log_info(cl, msg):
 123     log.info(cl.desc + msg, dflag=False)
 124
 125   def report_running(cl):
 126     if not cl.running_reported:
 127       cl.log_info('running OK')
 128       cl.running_reported = True
 129
 130   def log(cl, dflag, msg, **kwargs):
 131     log_debug(dflag, cl.desc + msg, **kwargs)
 132
 133   def log_outstanding(cl):
 134     cl.log(DBG.CTRL_DUMP, 'OS %s' % cl.outstanding)
 135
 136   def start(cl):
 137     cl.queue = PacketQueue('up', cl.c.max_queue_time)
 138     cl.agent = twisted.web.client.Agent(
 139       reactor, connectTimeout = cl.c.http_timeout)
 140
 141   def outbound(cl, packet, saddr, daddr):
 142     #print('OUT ', saddr, daddr, repr(packet))
 143     cl.queue.append(packet)
 144     cl.check_outbound()
 145
 146   def req_ok(cl, req, resp):
 147     cl.log(DBG.HTTP_CTRL,
 148             'req_ok %d %s %s' % (resp.code, repr(resp.phrase), str(resp)),
 149             idof=req)
 150     if resp.code == 200:
 151       rc = ResponseConsumer(cl, req, resp)
 152     else:
 153       rc = ErrorResponseConsumer(cl, req, resp)
 154
 155     resp.deliverBody(rc)
 156     # now rc is responsible for calling req_fin
 157
 158   def req_err(cl, req, err):
 159     # called when the Deferred fails, or (if it completes),
 160     # later, by ResponsConsumer or ErrorResponsConsumer
 161     try:
 162       cl.log(DBG.HTTP_CTRL, 'req_err ' + str(err), idof=req)
 163       cl.running_reported = False
 164       if isinstance(err, twisted.python.failure.Failure):
 165         err = err.getTraceback()
 166       print('%s[%#x] %s' % (cl.desc, id(req), err.strip('\n').replace('\n',' / ')),
 167             file=sys.stderr)
 168       if not isinstance(cl.outstanding[req], int):
 169         raise RuntimeError('[%#x] previously %s' %
 170                            (id(req), cl.outstanding[req]))
 171       cl.outstanding[req] = err
 172       cl.log_outstanding()
 173       reactor.callLater(cl.c.http_retry, partial(cl.req_fin, req))
 174     except Exception as e:
 175       crash(traceback.format_exc() + '\n----- handling -----\n' + err)
 176
 177   def req_fin(cl, req):
 178     del cl.outstanding[req]
 179     cl.log(DBG.HTTP_CTRL, 'req_fin OS=%d' % len(cl.outstanding), idof=req)
 180     cl.check_outbound()
 181
 182   def check_outbound(cl):
 183     while True:
 184       if len(cl.outstanding) >= cl.c.max_outstanding:
 185         break
 186
 187       if (not cl.queue.nonempty() and
 188           len(cl.outstanding) >= cl.c.target_requests_outstanding):
 189         break
 190
 191       d = b''
 192       def moredata(s): nonlocal d; d += s
 193       cl.queue.process((lambda: len(d)),
 194                     moredata,
 195                     cl.c.max_batch_up)
 196
 197       d = mime_translate(d)
 198
 199       token = authtoken_make(cl.c.secret)
 200
 201       crlf = b'\r\n'
 202       lf   =   b'\n'
 203       mime = (b'--b'                                        + crlf +
 204               b'Content-Type: text/plain; charset="utf-8"'  + crlf +
 205               b'Content-Disposition: form-data; name="m"'   + crlf + crlf +
 206               str(cl.c.client)            .encode('ascii')  + crlf +
 207               token                                         + crlf +
 208               str(cl.c.target_requests_outstanding)
 209                                           .encode('ascii')  + crlf +
 210               str(cl.c.http_timeout)      .encode('ascii')  + crlf +
 211             ((
 212               b'--b'                                        + crlf +
 213               b'Content-Type: application/octet-stream'     + crlf +
 214               b'Content-Disposition: form-data; name="d"'   + crlf + crlf +
 215               d                                             + crlf
 216              ) if len(d) else b'')                               +
 217               b'--b--'                                      + crlf)
 218
 219       #df = open('data.dump.dbg', mode='wb')
 220       #df.write(mime)
 221       #df.close()
 222       # POST -use -c 'multipart/form-data; boundary="b"' http://localhost:8099/ <data.dump.dbg
 223
 224       cl.log(DBG.HTTP_FULL, 'requesting: ' + str(mime))
 225
 226       hh = { 'User-Agent': ['hippotat'],
 227              'Content-Type': ['multipart/form-data; boundary="b"'] }
 228
 229       bytesreader = io.BytesIO(mime)
 230       producer = twisted.web.client.FileBodyProducer(bytesreader)
 231
 232       req = cl.agent.request(b'POST',
 233                           cl.c.url,
 234                           twisted.web.client.Headers(hh),
 235                           producer)
 236
 237       cl.outstanding[req] = len(d)
 238       cl.log(DBG.HTTP_CTRL,
 239              'request OS=%d' % len(cl.outstanding),
 240              idof=req, d=d)
 241       req.addTimeout(cl.c.http_timeout, reactor)
 242       req.addCallback(partial(cl.req_ok, req))
 243       req.addErrback(partial(cl.req_err, req))
 244
 245     cl.log_outstanding()
 246
 247 clients = [ ]
 248
 249 def encode_url(urlstr):
 250   # Oh, this is a disaster.  We're given a URL as a `str', but the underlying
 251   # machinery insists on having `bytes'.  Assume we've been given a sensible
 252   # URL, with escaping in all of the necessary places, except that it may
 253   # contain non-ASCII characters: then encode as UTF-8 and squash the top-
 254   # bit-set bytes down to percent escapes.
 255   #
 256   # This conses like it's going out of fashion, but it gets the job done.
 257   return b''.join(bytes([b]) if b < 128 else '%%%02X' % b
 258                   for b in urlstr.encode('utf-8'))
 259
 260 def process_cfg(_opts, putative_servers, putative_clients):
 261   global clients
 262
 263   for ss in putative_servers.values():
 264     for (ci,cs) in putative_clients.items():
 265       c = ConfigResults()
 266
 267       sections = cfg_process_client_common(c,ss,cs,ci)
 268       if not sections: continue
 269
 270       log_debug_config('processing client [%s %s]' % (ss, cs))
 271
 272       def srch(getter,key): return cfg_search(getter,key,sections)
 273
 274       c.http_timeout   += srch(cfg.getint, 'http_timeout_grace')
 275       c.max_outstanding = srch(cfg.getint, 'max_requests_outstanding')
 276       c.max_batch_up    = srch(cfg.getint, 'max_batch_up')
 277       c.http_retry      = srch(cfg.getint, 'http_retry')
 278       c.max_queue_time  = srch(cfg.getint, 'max_queue_time')
 279       c.vroutes         = srch(cfg.get,    'vroutes')
 280
 281       try: c.ifname     = srch(cfg_get_raw, 'ifname_client')
 282       except NoOptionError: pass
 283
 284       try: c.url = encode_url(srch(cfg.get,'url'))
 285       except NoOptionError:
 286         cfg_process_saddrs(c, ss)
 287         c.url = c.saddrs[0].url()
 288
 289       c.client = ci
 290
 291       cfg_process_vaddr(c,ss)
 292
 293       cfg_process_ipif(c,
 294                        sections,
 295                        (('local','client'),
 296                         ('peer', 'vaddr'),
 297                         ('rnets','vroutes')))
 298
 299       clients.append(Client(c,ss,cs))
 300
 301 common_startup(process_cfg)
 302
 303 for cl in clients:
 304   cl.start()
 305   cl.ipif = start_ipif(cl.c.ipif_command, cl.outbound)
 306   cl.check_outbound()
 307
 308 common_run()