ResponseConsumer: tolerate daft exception PotentialDataLoss
[hippotat] / 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
30 import io
31
32 class GeneralResponseConsumer(twisted.internet.protocol.Protocol):
33 def __init__(self, cl, req, resp, desc):
34 self._cl = cl
35 self._req = req
36 self._resp = resp
37 self._desc = desc
38
39 def _log(self, dflag, msg, **kwargs):
40 self._cl.log(dflag, '%s: %s' % (self._desc, msg), idof=self._req, **kwargs)
41
42 def connectionMade(self):
43 self._log(DBG.HTTP_CTRL, 'connectionMade')
44
45 def connectionLostOK(self, reason):
46 return (reason.check(twisted.web.client.ResponseDone) or
47 reason.check(twisted.web.client.PotentialDataLoss))
48 # twisted.web.client.PotentialDataLoss is an entirely daft
49 # exception. It will occur every time if the origin server does
50 # not provide a Content-Length. (hippotatd does, of course, but
51 # the HTTP transaction might be proxied.)
52
53 class ResponseConsumer(GeneralResponseConsumer):
54 def __init__(self, cl, req, resp):
55 super().__init__(cl, req, resp, 'RC')
56 ssddesc = '[%s] %s' % (id(req), self._desc)
57 self._ssd = SlipStreamDecoder(ssddesc, partial(queue_inbound, cl.ipif))
58 self._log(DBG.HTTP_CTRL, '__init__')
59
60 def dataReceived(self, data):
61 self._log(DBG.HTTP, 'dataReceived', d=data)
62 try:
63 self._ssd.inputdata(data)
64 except Exception as e:
65 self._handleexception()
66
67 def connectionLost(self, reason):
68 reason_msg = 'connectionLost ' + str(reason)
69 self._log(DBG.HTTP_CTRL, reason_msg)
70 if not self.connectionLostOK(reason):
71 self._latefailure(reason_msg)
72 return
73 try:
74 self._log(DBG.HTTP, 'ResponseDone')
75 self._ssd.flush()
76 self._cl.req_fin(self._req)
77 except Exception as e:
78 self._handleexception()
79 self._cl.report_running()
80
81 def _handleexception(self):
82 self._latefailure(traceback.format_exc())
83
84 def _latefailure(self, reason):
85 self._log(DBG.HTTP_CTRL, '_latefailure ' + str(reason))
86 self._cl.req_err(self._req, reason)
87
88 class ErrorResponseConsumer(GeneralResponseConsumer):
89 def __init__(self, cl, req, resp):
90 super().__init__(cl, req, resp, 'ERROR-RC')
91 self._m = b''
92 try:
93 self._phrase = resp.phrase.decode('utf-8')
94 except Exception:
95 self._phrase = repr(resp.phrase)
96 self._log(DBG.HTTP_CTRL, '__init__ %d %s' % (resp.code, self._phrase))
97
98 def dataReceived(self, data):
99 self._log(DBG.HTTP_CTRL, 'dataReceived ' + repr(data))
100 self._m += data
101
102 def connectionLost(self, reason):
103 try:
104 mbody = self._m.decode('utf-8')
105 except Exception:
106 mbody = repr(self._m)
107 if not self.connectionLostOK(reason):
108 mbody += ' || ' + str(reason)
109 self._cl.req_err(self._req,
110 "FAILED %d %s | %s"
111 % (self._resp.code, self._phrase, mbody))
112
113 class Client():
114 def __init__(cl, c,ss,cs):
115 cl.c = c
116 cl.outstanding = { }
117 cl.desc = '[%s %s] ' % (ss,cs)
118 cl.running_reported = False
119 cl.log_info('setting up')
120
121 def log_info(cl, msg):
122 log.info(cl.desc + msg, dflag=False)
123
124 def report_running(cl):
125 if not cl.running_reported:
126 cl.log_info('running OK')
127 cl.running_reported = True
128
129 def log(cl, dflag, msg, **kwargs):
130 log_debug(dflag, cl.desc + msg, **kwargs)
131
132 def log_outstanding(cl):
133 cl.log(DBG.CTRL_DUMP, 'OS %s' % cl.outstanding)
134
135 def start(cl):
136 cl.queue = PacketQueue('up', cl.c.max_queue_time)
137 cl.agent = twisted.web.client.Agent(
138 reactor, connectTimeout = cl.c.http_timeout)
139
140 def outbound(cl, packet, saddr, daddr):
141 #print('OUT ', saddr, daddr, repr(packet))
142 cl.queue.append(packet)
143 cl.check_outbound()
144
145 def req_ok(cl, req, resp):
146 cl.log(DBG.HTTP_CTRL,
147 'req_ok %d %s %s' % (resp.code, repr(resp.phrase), str(resp)),
148 idof=req)
149 if resp.code == 200:
150 rc = ResponseConsumer(cl, req, resp)
151 else:
152 rc = ErrorResponseConsumer(cl, req, resp)
153
154 resp.deliverBody(rc)
155 # now rc is responsible for calling req_fin
156
157 def req_err(cl, req, err):
158 # called when the Deferred fails, or (if it completes),
159 # later, by ResponsConsumer or ErrorResponsConsumer
160 try:
161 cl.log(DBG.HTTP_CTRL, 'req_err ' + str(err), idof=req)
162 cl.running_reported = False
163 if isinstance(err, twisted.python.failure.Failure):
164 err = err.getTraceback()
165 print('%s[%#x] %s' % (cl.desc, id(req), err.strip('\n').replace('\n',' / ')),
166 file=sys.stderr)
167 if not isinstance(cl.outstanding[req], int):
168 raise RuntimeError('[%#x] previously %s' %
169 (id(req), cl.outstanding[req]))
170 cl.outstanding[req] = err
171 cl.log_outstanding()
172 reactor.callLater(cl.c.http_retry, partial(cl.req_fin, req))
173 except Exception as e:
174 crash(traceback.format_exc() + '\n----- handling -----\n' + err)
175
176 def req_fin(cl, req):
177 del cl.outstanding[req]
178 cl.log(DBG.HTTP_CTRL, 'req_fin OS=%d' % len(cl.outstanding), idof=req)
179 cl.check_outbound()
180
181 def check_outbound(cl):
182 while True:
183 if len(cl.outstanding) >= cl.c.max_outstanding:
184 break
185
186 if (not cl.queue.nonempty() and
187 len(cl.outstanding) >= cl.c.target_requests_outstanding):
188 break
189
190 d = b''
191 def moredata(s): nonlocal d; d += s
192 cl.queue.process((lambda: len(d)),
193 moredata,
194 cl.c.max_batch_up)
195
196 d = mime_translate(d)
197
198 token = authtoken_make(cl.c.secret)
199
200 crlf = b'\r\n'
201 lf = b'\n'
202 mime = (b'--b' + crlf +
203 b'Content-Type: text/plain; charset="utf-8"' + crlf +
204 b'Content-Disposition: form-data; name="m"' + crlf + crlf +
205 str(cl.c.client) .encode('ascii') + crlf +
206 token + crlf +
207 str(cl.c.target_requests_outstanding)
208 .encode('ascii') + crlf +
209 str(cl.c.http_timeout) .encode('ascii') + crlf +
210 ((
211 b'--b' + crlf +
212 b'Content-Type: application/octet-stream' + crlf +
213 b'Content-Disposition: form-data; name="d"' + crlf + crlf +
214 d + crlf
215 ) if len(d) else b'') +
216 b'--b--' + crlf)
217
218 #df = open('data.dump.dbg', mode='wb')
219 #df.write(mime)
220 #df.close()
221 # POST -use -c 'multipart/form-data; boundary="b"' http://localhost:8099/ <data.dump.dbg
222
223 cl.log(DBG.HTTP_FULL, 'requesting: ' + str(mime))
224
225 hh = { 'User-Agent': ['hippotat'],
226 'Content-Type': ['multipart/form-data; boundary="b"'],
227 'Content-Length': [str(len(mime))] }
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 process_cfg(_opts, putative_servers, putative_clients):
250 global clients
251
252 for ss in putative_servers.values():
253 for (ci,cs) in putative_clients.items():
254 c = ConfigResults()
255
256 sections = cfg_process_client_common(c,ss,cs,ci)
257 if not sections: continue
258
259 log_debug_config('processing client [%s %s]' % (ss, cs))
260
261 def srch(getter,key): return cfg_search(getter,key,sections)
262
263 c.http_timeout += srch(cfg.getint, 'http_timeout_grace')
264 c.max_outstanding = srch(cfg.getint, 'max_requests_outstanding')
265 c.max_batch_up = srch(cfg.getint, 'max_batch_up')
266 c.http_retry = srch(cfg.getint, 'http_retry')
267 c.max_queue_time = srch(cfg.getint, 'max_queue_time')
268 c.vroutes = srch(cfg.get, 'vroutes')
269
270 try: c.ifname = srch(cfg_get_raw, 'ifname_client')
271 except NoOptionError: pass
272
273 try: c.url = srch(cfg.get,'url')
274 except NoOptionError:
275 cfg_process_saddrs(c, ss)
276 c.url = c.saddrs[0].url()
277
278 c.client = ci
279
280 cfg_process_vaddr(c,ss)
281
282 cfg_process_ipif(c,
283 sections,
284 (('local','client'),
285 ('peer', 'vaddr'),
286 ('rnets','vroutes')))
287
288 clients.append(Client(c,ss,cs))
289
290 common_startup(process_cfg)
291
292 for cl in clients:
293 cl.start()
294 cl.ipif = start_ipif(cl.c.ipif_command, cl.outbound)
295 cl.check_outbound()
296
297 common_run()