support federation queries through http connect proxy (#10475)

Signed-off-by: Marcus Hoffmann <bubu@bubu1.eu>
Signed-off-by: Dirk Klimpel dirk@klimpel.org

4 files changed, 152 insertions, 79 deletions

diff --git a/synapse/http/connectproxyclient.py b/synapse/http/connectproxyclient.py
index 17e1c5abb1..c577142268 100644
--- a/synapse/http/connectproxyclient.py
+++ b/synapse/http/connectproxyclient.py
@@ -12,8 +12,11 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import base64
 import logging
+from typing import Optional
 
+import attr
 from zope.interface import implementer
 
 from twisted.internet import defer, protocol
@@ -21,7 +24,6 @@ from twisted.internet.error import ConnectError
 from twisted.internet.interfaces import IReactorCore, IStreamClientEndpoint
 from twisted.internet.protocol import ClientFactory, Protocol, connectionDone
 from twisted.web import http
-from twisted.web.http_headers import Headers
 
 logger = logging.getLogger(__name__)
 
@@ -30,6 +32,22 @@ class ProxyConnectError(ConnectError):
     pass
 
 
+@attr.s
+class ProxyCredentials:
+    username_password = attr.ib(type=bytes)
+
+    def as_proxy_authorization_value(self) -> bytes:
+        """
+        Return the value for a Proxy-Authorization header (i.e. 'Basic abdef==').
+
+        Returns:
+            A transformation of the authentication string the encoded value for
+            a Proxy-Authorization header.
+        """
+        # Encode as base64 and prepend the authorization type
+        return b"Basic " + base64.encodebytes(self.username_password)
+
+
 @implementer(IStreamClientEndpoint)
 class HTTPConnectProxyEndpoint:
     """An Endpoint implementation which will send a CONNECT request to an http proxy
@@ -46,7 +64,7 @@ class HTTPConnectProxyEndpoint:
         proxy_endpoint: the endpoint to use to connect to the proxy
         host: hostname that we want to CONNECT to
         port: port that we want to connect to
-        headers: Extra HTTP headers to include in the CONNECT request
+        proxy_creds: credentials to authenticate at proxy
     """
 
     def __init__(
@@ -55,20 +73,20 @@ class HTTPConnectProxyEndpoint:
         proxy_endpoint: IStreamClientEndpoint,
         host: bytes,
         port: int,
-        headers: Headers,
+        proxy_creds: Optional[ProxyCredentials],
     ):
         self._reactor = reactor
         self._proxy_endpoint = proxy_endpoint
         self._host = host
         self._port = port
-        self._headers = headers
+        self._proxy_creds = proxy_creds
 
     def __repr__(self):
         return "<HTTPConnectProxyEndpoint %s>" % (self._proxy_endpoint,)
 
     def connect(self, protocolFactory: ClientFactory):
         f = HTTPProxiedClientFactory(
-            self._host, self._port, protocolFactory, self._headers
+            self._host, self._port, protocolFactory, self._proxy_creds
         )
         d = self._proxy_endpoint.connect(f)
         # once the tcp socket connects successfully, we need to wait for the
@@ -87,7 +105,7 @@ class HTTPProxiedClientFactory(protocol.ClientFactory):
         dst_host: hostname that we want to CONNECT to
         dst_port: port that we want to connect to
         wrapped_factory: The original Factory
-        headers: Extra HTTP headers to include in the CONNECT request
+        proxy_creds: credentials to authenticate at proxy
     """
 
     def __init__(
@@ -95,12 +113,12 @@ class HTTPProxiedClientFactory(protocol.ClientFactory):
         dst_host: bytes,
         dst_port: int,
         wrapped_factory: ClientFactory,
-        headers: Headers,
+        proxy_creds: Optional[ProxyCredentials],
     ):
         self.dst_host = dst_host
         self.dst_port = dst_port
         self.wrapped_factory = wrapped_factory
-        self.headers = headers
+        self.proxy_creds = proxy_creds
         self.on_connection = defer.Deferred()
 
     def startedConnecting(self, connector):
@@ -114,7 +132,7 @@ class HTTPProxiedClientFactory(protocol.ClientFactory):
             self.dst_port,
             wrapped_protocol,
             self.on_connection,
-            self.headers,
+            self.proxy_creds,
         )
 
     def clientConnectionFailed(self, connector, reason):
@@ -145,7 +163,7 @@ class HTTPConnectProtocol(protocol.Protocol):
         connected_deferred: a Deferred which will be callbacked with
             wrapped_protocol when the CONNECT completes
 
-        headers: Extra HTTP headers to include in the CONNECT request
+        proxy_creds: credentials to authenticate at proxy
     """
 
     def __init__(
@@ -154,16 +172,16 @@ class HTTPConnectProtocol(protocol.Protocol):
         port: int,
         wrapped_protocol: Protocol,
         connected_deferred: defer.Deferred,
-        headers: Headers,
+        proxy_creds: Optional[ProxyCredentials],
     ):
         self.host = host
         self.port = port
         self.wrapped_protocol = wrapped_protocol
         self.connected_deferred = connected_deferred
-        self.headers = headers
+        self.proxy_creds = proxy_creds
 
         self.http_setup_client = HTTPConnectSetupClient(
-            self.host, self.port, self.headers
+            self.host, self.port, self.proxy_creds
         )
         self.http_setup_client.on_connected.addCallback(self.proxyConnected)
 
@@ -205,30 +223,38 @@ class HTTPConnectSetupClient(http.HTTPClient):
     Args:
         host: The hostname to send in the CONNECT message
         port: The port to send in the CONNECT message
-        headers: Extra headers to send with the CONNECT message
+        proxy_creds: credentials to authenticate at proxy
     """
 
-    def __init__(self, host: bytes, port: int, headers: Headers):
+    def __init__(
+        self,
+        host: bytes,
+        port: int,
+        proxy_creds: Optional[ProxyCredentials],
+    ):
         self.host = host
         self.port = port
-        self.headers = headers
+        self.proxy_creds = proxy_creds
         self.on_connected = defer.Deferred()
 
     def connectionMade(self):
         logger.debug("Connected to proxy, sending CONNECT")
         self.sendCommand(b"CONNECT", b"%s:%d" % (self.host, self.port))
 
-        # Send any additional specified headers
-        for name, values in self.headers.getAllRawHeaders():
-            for value in values:
-                self.sendHeader(name, value)
+        # Determine whether we need to set Proxy-Authorization headers
+        if self.proxy_creds:
+            # Set a Proxy-Authorization header
+            self.sendHeader(
+                b"Proxy-Authorization",
+                self.proxy_creds.as_proxy_authorization_value(),
+            )
 
         self.endHeaders()
 
     def handleStatus(self, version: bytes, status: bytes, message: bytes):
         logger.debug("Got Status: %s %s %s", status, message, version)
         if status != b"200":
-            raise ProxyConnectError("Unexpected status on CONNECT: %s" % status)
+            raise ProxyConnectError(f"Unexpected status on CONNECT: {status!s}")
 
     def handleEndHeaders(self):
         logger.debug("End Headers")
diff --git a/synapse/http/federation/matrix_federation_agent.py b/synapse/http/federation/matrix_federation_agent.py
index c16b7f10e6..1238bfd287 100644
--- a/synapse/http/federation/matrix_federation_agent.py
+++ b/synapse/http/federation/matrix_federation_agent.py
@@ -14,6 +14,10 @@
 import logging
 import urllib.parse
 from typing import Any, Generator, List, Optional
+from urllib.request import (  # type: ignore[attr-defined]
+    getproxies_environment,
+    proxy_bypass_environment,
+)
 
 from netaddr import AddrFormatError, IPAddress, IPSet
 from zope.interface import implementer
@@ -30,9 +34,12 @@ from twisted.web.http_headers import Headers
 from twisted.web.iweb import IAgent, IAgentEndpointFactory, IBodyProducer, IResponse
 
 from synapse.crypto.context_factory import FederationPolicyForHTTPS
-from synapse.http.client import BlacklistingAgentWrapper
+from synapse.http import proxyagent
+from synapse.http.client import BlacklistingAgentWrapper, BlacklistingReactorWrapper
+from synapse.http.connectproxyclient import HTTPConnectProxyEndpoint
 from synapse.http.federation.srv_resolver import Server, SrvResolver
 from synapse.http.federation.well_known_resolver import WellKnownResolver
+from synapse.http.proxyagent import ProxyAgent
 from synapse.logging.context import make_deferred_yieldable, run_in_background
 from synapse.types import ISynapseReactor
 from synapse.util import Clock
@@ -57,6 +64,14 @@ class MatrixFederationAgent:
         user_agent:
             The user agent header to use for federation requests.
 
+        ip_whitelist: Allowed IP addresses.
+
+        ip_blacklist: Disallowed IP addresses.
+
+        proxy_reactor: twisted reactor to use for connections to the proxy server
+           reactor might have some blacklisting applied (i.e. for DNS queries),
+           but we need unblocked access to the proxy.
+
         _srv_resolver:
             SrvResolver implementation to use for looking up SRV records. None
             to use a default implementation.
@@ -71,11 +86,18 @@ class MatrixFederationAgent:
         reactor: ISynapseReactor,
         tls_client_options_factory: Optional[FederationPolicyForHTTPS],
         user_agent: bytes,
+        ip_whitelist: IPSet,
         ip_blacklist: IPSet,
         _srv_resolver: Optional[SrvResolver] = None,
         _well_known_resolver: Optional[WellKnownResolver] = None,
     ):
-        self._reactor = reactor
+        # proxy_reactor is not blacklisted
+        proxy_reactor = reactor
+
+        # We need to use a DNS resolver which filters out blacklisted IP
+        # addresses, to prevent DNS rebinding.
+        reactor = BlacklistingReactorWrapper(reactor, ip_whitelist, ip_blacklist)
+
         self._clock = Clock(reactor)
         self._pool = HTTPConnectionPool(reactor)
         self._pool.retryAutomatically = False
@@ -83,24 +105,27 @@ class MatrixFederationAgent:
         self._pool.cachedConnectionTimeout = 2 * 60
 
         self._agent = Agent.usingEndpointFactory(
-            self._reactor,
+            reactor,
             MatrixHostnameEndpointFactory(
-                reactor, tls_client_options_factory, _srv_resolver
+                reactor,
+                proxy_reactor,
+                tls_client_options_factory,
+                _srv_resolver,
             ),
             pool=self._pool,
         )
         self.user_agent = user_agent
 
         if _well_known_resolver is None:
-            # Note that the name resolver has already been wrapped in a
-            # IPBlacklistingResolver by MatrixFederationHttpClient.
             _well_known_resolver = WellKnownResolver(
-                self._reactor,
+                reactor,
                 agent=BlacklistingAgentWrapper(
-                    Agent(
-                        self._reactor,
+                    ProxyAgent(
+                        reactor,
+                        proxy_reactor,
                         pool=self._pool,
                         contextFactory=tls_client_options_factory,
+                        use_proxy=True,
                     ),
                     ip_blacklist=ip_blacklist,
                 ),
@@ -200,10 +225,12 @@ class MatrixHostnameEndpointFactory:
     def __init__(
         self,
         reactor: IReactorCore,
+        proxy_reactor: IReactorCore,
         tls_client_options_factory: Optional[FederationPolicyForHTTPS],
         srv_resolver: Optional[SrvResolver],
     ):
         self._reactor = reactor
+        self._proxy_reactor = proxy_reactor
         self._tls_client_options_factory = tls_client_options_factory
 
         if srv_resolver is None:
@@ -211,9 +238,10 @@ class MatrixHostnameEndpointFactory:
 
         self._srv_resolver = srv_resolver
 
-    def endpointForURI(self, parsed_uri):
+    def endpointForURI(self, parsed_uri: URI):
         return MatrixHostnameEndpoint(
             self._reactor,
+            self._proxy_reactor,
             self._tls_client_options_factory,
             self._srv_resolver,
             parsed_uri,
@@ -227,23 +255,45 @@ class MatrixHostnameEndpoint:
 
     Args:
         reactor: twisted reactor to use for underlying requests
+        proxy_reactor: twisted reactor to use for connections to the proxy server.
+           'reactor' might have some blacklisting applied (i.e. for DNS queries),
+           but we need unblocked access to the proxy.
         tls_client_options_factory:
             factory to use for fetching client tls options, or none to disable TLS.
         srv_resolver: The SRV resolver to use
         parsed_uri: The parsed URI that we're wanting to connect to.
+
+    Raises:
+        ValueError if the environment variables contain an invalid proxy specification.
+        RuntimeError if no tls_options_factory is given for a https connection
     """
 
     def __init__(
         self,
         reactor: IReactorCore,
+        proxy_reactor: IReactorCore,
         tls_client_options_factory: Optional[FederationPolicyForHTTPS],
         srv_resolver: SrvResolver,
         parsed_uri: URI,
     ):
         self._reactor = reactor
-
         self._parsed_uri = parsed_uri
 
+        # http_proxy is not needed because federation is always over TLS
+        proxies = getproxies_environment()
+        https_proxy = proxies["https"].encode() if "https" in proxies else None
+        self.no_proxy = proxies["no"] if "no" in proxies else None
+
+        # endpoint and credentials to use to connect to the outbound https proxy, if any.
+        (
+            self._https_proxy_endpoint,
+            self._https_proxy_creds,
+        ) = proxyagent.http_proxy_endpoint(
+            https_proxy,
+            proxy_reactor,
+            tls_client_options_factory,
+        )
+
         # set up the TLS connection params
         #
         # XXX disabling TLS is really only supported here for the benefit of the
@@ -273,9 +323,33 @@ class MatrixHostnameEndpoint:
             host = server.host
             port = server.port
 
+            should_skip_proxy = False
+            if self.no_proxy is not None:
+                should_skip_proxy = proxy_bypass_environment(
+                    host.decode(),
+                    proxies={"no": self.no_proxy},
+                )
+
+            endpoint: IStreamClientEndpoint
             try:
-                logger.debug("Connecting to %s:%i", host.decode("ascii"), port)
-                endpoint = HostnameEndpoint(self._reactor, host, port)
+                if self._https_proxy_endpoint and not should_skip_proxy:
+                    logger.debug(
+                        "Connecting to %s:%i via %s",
+                        host.decode("ascii"),
+                        port,
+                        self._https_proxy_endpoint,
+                    )
+                    endpoint = HTTPConnectProxyEndpoint(
+                        self._reactor,
+                        self._https_proxy_endpoint,
+                        host,
+                        port,
+                        proxy_creds=self._https_proxy_creds,
+                    )
+                else:
+                    logger.debug("Connecting to %s:%i", host.decode("ascii"), port)
+                    # not using a proxy
+                    endpoint = HostnameEndpoint(self._reactor, host, port)
                 if self._tls_options:
                     endpoint = wrapClientTLS(self._tls_options, endpoint)
                 result = await make_deferred_yieldable(
diff --git a/synapse/http/matrixfederationclient.py b/synapse/http/matrixfederationclient.py
index 2efa15bf04..2e9898997c 100644
--- a/synapse/http/matrixfederationclient.py
+++ b/synapse/http/matrixfederationclient.py
@@ -59,7 +59,6 @@ from synapse.api.errors import (
 from synapse.http import QuieterFileBodyProducer
 from synapse.http.client import (
     BlacklistingAgentWrapper,
-    BlacklistingReactorWrapper,
     BodyExceededMaxSize,
     ByteWriteable,
     encode_query_args,
@@ -69,7 +68,7 @@ from synapse.http.federation.matrix_federation_agent import MatrixFederationAgen
 from synapse.logging import opentracing
 from synapse.logging.context import make_deferred_yieldable
 from synapse.logging.opentracing import set_tag, start_active_span, tags
-from synapse.types import ISynapseReactor, JsonDict
+from synapse.types import JsonDict
 from synapse.util import json_decoder
 from synapse.util.async_helpers import timeout_deferred
 from synapse.util.metrics import Measure
@@ -325,13 +324,7 @@ class MatrixFederationHttpClient:
         self.signing_key = hs.signing_key
         self.server_name = hs.hostname
 
-        # We need to use a DNS resolver which filters out blacklisted IP
-        # addresses, to prevent DNS rebinding.
-        self.reactor: ISynapseReactor = BlacklistingReactorWrapper(
-            hs.get_reactor(),
-            hs.config.federation_ip_range_whitelist,
-            hs.config.federation_ip_range_blacklist,
-        )
+        self.reactor = hs.get_reactor()
 
         user_agent = hs.version_string
         if hs.config.user_agent_suffix:
@@ -342,6 +335,7 @@ class MatrixFederationHttpClient:
             self.reactor,
             tls_client_options_factory,
             user_agent,
+            hs.config.federation_ip_range_whitelist,
             hs.config.federation_ip_range_blacklist,
         )
 
diff --git a/synapse/http/proxyagent.py b/synapse/http/proxyagent.py
index 19e987f118..a3f31452d0 100644
--- a/synapse/http/proxyagent.py
+++ b/synapse/http/proxyagent.py
@@ -11,7 +11,6 @@
 # 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.
-import base64
 import logging
 import re
 from typing import Any, Dict, Optional, Tuple
@@ -21,7 +20,6 @@ from urllib.request import (  # type: ignore[attr-defined]
     proxy_bypass_environment,
 )
 
-import attr
 from zope.interface import implementer
 
 from twisted.internet import defer
@@ -38,7 +36,7 @@ from twisted.web.error import SchemeNotSupported
 from twisted.web.http_headers import Headers
 from twisted.web.iweb import IAgent, IBodyProducer, IPolicyForHTTPS
 
-from synapse.http.connectproxyclient import HTTPConnectProxyEndpoint
+from synapse.http.connectproxyclient import HTTPConnectProxyEndpoint, ProxyCredentials
 from synapse.types import ISynapseReactor
 
 logger = logging.getLogger(__name__)
@@ -46,22 +44,6 @@ logger = logging.getLogger(__name__)
 _VALID_URI = re.compile(br"\A[\x21-\x7e]+\Z")
 
 
-@attr.s
-class ProxyCredentials:
-    username_password = attr.ib(type=bytes)
-
-    def as_proxy_authorization_value(self) -> bytes:
-        """
-        Return the value for a Proxy-Authorization header (i.e. 'Basic abdef==').
-
-        Returns:
-            A transformation of the authentication string the encoded value for
-            a Proxy-Authorization header.
-        """
-        # Encode as base64 and prepend the authorization type
-        return b"Basic " + base64.encodebytes(self.username_password)
-
-
 @implementer(IAgent)
 class ProxyAgent(_AgentBase):
     """An Agent implementation which will use an HTTP proxy if one was requested
@@ -95,6 +77,7 @@ class ProxyAgent(_AgentBase):
     Raises:
         ValueError if use_proxy is set and the environment variables
             contain an invalid proxy specification.
+        RuntimeError if no tls_options_factory is given for a https connection
     """
 
     def __init__(
@@ -131,11 +114,11 @@ class ProxyAgent(_AgentBase):
             https_proxy = proxies["https"].encode() if "https" in proxies else None
             no_proxy = proxies["no"] if "no" in proxies else None
 
-        self.http_proxy_endpoint, self.http_proxy_creds = _http_proxy_endpoint(
+        self.http_proxy_endpoint, self.http_proxy_creds = http_proxy_endpoint(
             http_proxy, self.proxy_reactor, contextFactory, **self._endpoint_kwargs
         )
 
-        self.https_proxy_endpoint, self.https_proxy_creds = _http_proxy_endpoint(
+        self.https_proxy_endpoint, self.https_proxy_creds = http_proxy_endpoint(
             https_proxy, self.proxy_reactor, contextFactory, **self._endpoint_kwargs
         )
 
@@ -224,22 +207,12 @@ class ProxyAgent(_AgentBase):
             and self.https_proxy_endpoint
             and not should_skip_proxy
         ):
-            connect_headers = Headers()
-
-            # Determine whether we need to set Proxy-Authorization headers
-            if self.https_proxy_creds:
-                # Set a Proxy-Authorization header
-                connect_headers.addRawHeader(
-                    b"Proxy-Authorization",
-                    self.https_proxy_creds.as_proxy_authorization_value(),
-                )
-
             endpoint = HTTPConnectProxyEndpoint(
                 self.proxy_reactor,
                 self.https_proxy_endpoint,
                 parsed_uri.host,
                 parsed_uri.port,
-                headers=connect_headers,
+                self.https_proxy_creds,
             )
         else:
             # not using a proxy
@@ -268,10 +241,10 @@ class ProxyAgent(_AgentBase):
         )
 
 
-def _http_proxy_endpoint(
+def http_proxy_endpoint(
     proxy: Optional[bytes],
     reactor: IReactorCore,
-    tls_options_factory: IPolicyForHTTPS,
+    tls_options_factory: Optional[IPolicyForHTTPS],
     **kwargs,
 ) -> Tuple[Optional[IStreamClientEndpoint], Optional[ProxyCredentials]]:
     """Parses an http proxy setting and returns an endpoint for the proxy
@@ -294,6 +267,7 @@ def _http_proxy_endpoint(
 
     Raise:
         ValueError if proxy has no hostname or unsupported scheme.
+        RuntimeError if no tls_options_factory is given for a https connection
     """
     if proxy is None:
         return None, None
@@ -305,8 +279,13 @@ def _http_proxy_endpoint(
     proxy_endpoint = HostnameEndpoint(reactor, host, port, **kwargs)
 
     if scheme == b"https":
-        tls_options = tls_options_factory.creatorForNetloc(host, port)
-        proxy_endpoint = wrapClientTLS(tls_options, proxy_endpoint)
+        if tls_options_factory:
+            tls_options = tls_options_factory.creatorForNetloc(host, port)
+            proxy_endpoint = wrapClientTLS(tls_options, proxy_endpoint)
+        else:
+            raise RuntimeError(
+                f"No TLS options for a https connection via proxy {proxy!s}"
+            )
 
     return proxy_endpoint, credentials