Add a config option for validating 'next_link' parameters against a domain whitelist (#8275)

This is a config option ported over from DINUM's Sydent: https://github.com/matrix-org/sydent/pull/285

They've switched to validating 3PIDs via Synapse rather than Sydent, and would like to retain this functionality.

This original purpose for this change is phishing prevention. This solution could also potentially be replaced by a similar one to https://github.com/matrix-org/synapse/pull/8004, but across all `*/submit_token` endpoint.

This option may still be useful to enterprise even with that safeguard in place though, if they want to be absolutely sure that their employees don't follow links to other domains.

tests/rest/client/v2_alpha/test_account.py

@@ -14,11 +14,11 @@
 # 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 json
 import os
 import re
 from email.parser import Parser
+from typing import Optional
 
 import pkg_resources
 
@@ -29,6 +29,7 @@ from synapse.rest.client.v1 import login, room
 from synapse.rest.client.v2_alpha import account, register
 
 from tests import unittest
+from tests.unittest import override_config
 
 
 class PasswordResetTestCase(unittest.HomeserverTestCase):
@@ -668,16 +669,104 @@ class ThreepidEmailRestTestCase(unittest.HomeserverTestCase):
         self.assertEqual(200, int(channel.result["code"]), msg=channel.result["body"])
         self.assertFalse(channel.json_body["threepids"])
 
-    def _request_token(self, email, client_secret):
+    @override_config({"next_link_domain_whitelist": None})
+    def test_next_link(self):
+        """Tests a valid next_link parameter value with no whitelist (good case)"""
+        self._request_token(
+            "something@example.com",
+            "some_secret",
+            next_link="https://example.com/a/good/site",
+            expect_code=200,
+        )
+
+    @override_config({"next_link_domain_whitelist": None})
+    def test_next_link_exotic_protocol(self):
+        """Tests using a esoteric protocol as a next_link parameter value.
+        Someone may be hosting a client on IPFS etc.
+        """
+        self._request_token(
+            "something@example.com",
+            "some_secret",
+            next_link="some-protocol://abcdefghijklmopqrstuvwxyz",
+            expect_code=200,
+        )
+
+    @override_config({"next_link_domain_whitelist": None})
+    def test_next_link_file_uri(self):
+        """Tests next_link parameters cannot be file URI"""
+        # Attempt to use a next_link value that points to the local disk
+        self._request_token(
+            "something@example.com",
+            "some_secret",
+            next_link="file:///host/path",
+            expect_code=400,
+        )
+
+    @override_config({"next_link_domain_whitelist": ["example.com", "example.org"]})
+    def test_next_link_domain_whitelist(self):
+        """Tests next_link parameters must fit the whitelist if provided"""
+        self._request_token(
+            "something@example.com",
+            "some_secret",
+            next_link="https://example.com/some/good/page",
+            expect_code=200,
+        )
+
+        self._request_token(
+            "something@example.com",
+            "some_secret",
+            next_link="https://example.org/some/also/good/page",
+            expect_code=200,
+        )
+
+        self._request_token(
+            "something@example.com",
+            "some_secret",
+            next_link="https://bad.example.org/some/bad/page",
+            expect_code=400,
+        )
+
+    @override_config({"next_link_domain_whitelist": []})
+    def test_empty_next_link_domain_whitelist(self):
+        """Tests an empty next_lint_domain_whitelist value, meaning next_link is essentially
+        disallowed
+        """
+        self._request_token(
+            "something@example.com",
+            "some_secret",
+            next_link="https://example.com/a/page",
+            expect_code=400,
+        )
+
+    def _request_token(
+        self,
+        email: str,
+        client_secret: str,
+        next_link: Optional[str] = None,
+        expect_code: int = 200,
+    ) -> str:
+        """Request a validation token to add an email address to a user's account
+
+        Args:
+            email: The email address to validate
+            client_secret: A secret string
+            next_link: A link to redirect the user to after validation
+            expect_code: Expected return code of the call
+
+        Returns:
+            The ID of the new threepid validation session
+        """
+        body = {"client_secret": client_secret, "email": email, "send_attempt": 1}
+        if next_link:
+            body["next_link"] = next_link
+
         request, channel = self.make_request(
-            "POST",
-            b"account/3pid/email/requestToken",
-            {"client_secret": client_secret, "email": email, "send_attempt": 1},
+            "POST", b"account/3pid/email/requestToken", body,
         )
         self.render(request)
-        self.assertEquals(200, channel.code, channel.result)
+        self.assertEquals(expect_code, channel.code, channel.result)
 
-        return channel.json_body["sid"]
+        return channel.json_body.get("sid")
 
     def _request_token_invalid_email(
         self, email, expected_errcode, expected_error, client_secret="foobar",