zulip/zerver/views/user_groups.py

443 lines
17 KiB
Python
Raw Normal View History

from typing import List, Optional, Sequence
from django.conf import settings
user_groups: Make locks required for updating user group memberships. **Background** User groups are expected to comply with the DAG constraint for the many-to-many inter-group membership. The check for this constraint has to be performed recursively so that we can find all direct and indirect subgroups of the user group to be added. This kind of check is vulnerable to phantom reads which is possible at the default read committed isolation level because we cannot guarantee that the check is still valid when we are adding the subgroups to the user group. **Solution** To avoid having another transaction concurrently update one of the to-be-subgroup after the recursive check is done, and before the subgroup is added, we use SELECT FOR UPDATE to lock the user group rows. The lock needs to be acquired before a group membership change is about to occur before any check has been conducted. Suppose that we are adding subgroup B to supergroup A, the locking protocol is specified as follows: 1. Acquire a lock for B and all its direct and indirect subgroups. 2. Acquire a lock for A. For the removal of user groups, we acquire a lock for the user group to be removed with all its direct and indirect subgroups. This is the special case A=B, which is still complaint with the protocol. **Error handling** We currently rely on Postgres' deadlock detection to abort transactions and show an error for the users. In the future, we might need some recovery mechanism or at least better error handling. **Notes** An important note is that we need to reuse the recursive CTE query that finds the direct and indirect subgroups when applying the lock on the rows. And the lock needs to be acquired the same way for the addition and removal of direct subgroups. User membership change (as opposed to user group membership) is not affected. Read-only queries aren't either. The locks only protect critical regions where the user group dependency graph might violate the DAG constraint, where users are not participating. **Testing** We implement a transaction test case targeting some typical scenarios when an internal server error is expected to happen (this means that the user group view makes the correct decision to abort the transaction when something goes wrong with locks). To achieve this, we add a development view intended only for unit tests. It has a global BARRIER that can be shared across threads, so that we can synchronize them to consistently reproduce certain potential race conditions prevented by the database locks. The transaction test case lanuches pairs of threads initiating possibly conflicting requests at the same time. The tests are set up such that exactly N of them are expected to succeed with a certain error message (while we don't know each one). **Security notes** get_recursive_subgroups_for_groups will no longer fetch user groups from other realms. As a result, trying to add/remove a subgroup from another realm results in a UserGroup not found error response. We also implement subgroup-specific checks in has_user_group_access to keep permission managing in a single place. Do note that the API currently don't have a way to violate that check because we are only checking the realm ID now.
2023-06-17 04:39:52 +02:00
from django.db import transaction
from django.http import HttpRequest, HttpResponse
from django.utils.translation import gettext as _
from django.utils.translation import override as override_language
from zerver.actions.message_send import do_send_messages, internal_prep_private_message
from zerver.actions.user_groups import (
add_subgroups_to_user_group,
bulk_add_members_to_user_group,
check_add_user_group,
check_delete_user_group,
do_change_user_group_permission_setting,
do_update_user_group_description,
do_update_user_group_name,
remove_members_from_user_group,
remove_subgroups_from_user_group,
)
from zerver.decorator import require_member_or_admin, require_user_group_edit_permission
from zerver.lib.exceptions import JsonableError
from zerver.lib.mention import MentionBackend, silent_mention_syntax_for_user
from zerver.lib.request import REQ, has_request_variables
from zerver.lib.response import json_success
from zerver.lib.user_groups import (
access_user_group_by_id,
access_user_group_for_setting,
check_user_group_name,
get_direct_memberships_of_users,
get_subgroup_ids,
get_user_group_direct_member_ids,
get_user_group_member_ids,
is_user_in_group,
user_groups: Make locks required for updating user group memberships. **Background** User groups are expected to comply with the DAG constraint for the many-to-many inter-group membership. The check for this constraint has to be performed recursively so that we can find all direct and indirect subgroups of the user group to be added. This kind of check is vulnerable to phantom reads which is possible at the default read committed isolation level because we cannot guarantee that the check is still valid when we are adding the subgroups to the user group. **Solution** To avoid having another transaction concurrently update one of the to-be-subgroup after the recursive check is done, and before the subgroup is added, we use SELECT FOR UPDATE to lock the user group rows. The lock needs to be acquired before a group membership change is about to occur before any check has been conducted. Suppose that we are adding subgroup B to supergroup A, the locking protocol is specified as follows: 1. Acquire a lock for B and all its direct and indirect subgroups. 2. Acquire a lock for A. For the removal of user groups, we acquire a lock for the user group to be removed with all its direct and indirect subgroups. This is the special case A=B, which is still complaint with the protocol. **Error handling** We currently rely on Postgres' deadlock detection to abort transactions and show an error for the users. In the future, we might need some recovery mechanism or at least better error handling. **Notes** An important note is that we need to reuse the recursive CTE query that finds the direct and indirect subgroups when applying the lock on the rows. And the lock needs to be acquired the same way for the addition and removal of direct subgroups. User membership change (as opposed to user group membership) is not affected. Read-only queries aren't either. The locks only protect critical regions where the user group dependency graph might violate the DAG constraint, where users are not participating. **Testing** We implement a transaction test case targeting some typical scenarios when an internal server error is expected to happen (this means that the user group view makes the correct decision to abort the transaction when something goes wrong with locks). To achieve this, we add a development view intended only for unit tests. It has a global BARRIER that can be shared across threads, so that we can synchronize them to consistently reproduce certain potential race conditions prevented by the database locks. The transaction test case lanuches pairs of threads initiating possibly conflicting requests at the same time. The tests are set up such that exactly N of them are expected to succeed with a certain error message (while we don't know each one). **Security notes** get_recursive_subgroups_for_groups will no longer fetch user groups from other realms. As a result, trying to add/remove a subgroup from another realm results in a UserGroup not found error response. We also implement subgroup-specific checks in has_user_group_access to keep permission managing in a single place. Do note that the API currently don't have a way to violate that check because we are only checking the realm ID now.
2023-06-17 04:39:52 +02:00
lock_subgroups_with_respect_to_supergroup,
user_groups_in_realm_serialized,
)
from zerver.lib.users import access_user_by_id, user_ids_to_users
from zerver.lib.validator import check_bool, check_int, check_list
from zerver.models import UserGroup, UserProfile, get_system_bot
from zerver.views.streams import compose_views
@require_user_group_edit_permission
@has_request_variables
def add_user_group(
request: HttpRequest,
user_profile: UserProfile,
name: str = REQ(),
members: Sequence[int] = REQ(json_validator=check_list(check_int), default=[]),
description: str = REQ(),
can_mention_group_id: Optional[int] = REQ(
"can_mention_group", json_validator=check_int, default=None
),
) -> HttpResponse:
user_profiles = user_ids_to_users(members, user_profile.realm)
name = check_user_group_name(name)
group_settings_map = {}
request_settings_dict = locals()
for setting_name, permission_config in UserGroup.GROUP_PERMISSION_SETTINGS.items():
setting_group_id_name = permission_config.id_field_name
if setting_group_id_name not in request_settings_dict: # nocoverage
continue
if request_settings_dict[setting_group_id_name] is not None:
setting_value_group_id = request_settings_dict[setting_group_id_name]
setting_value_group = access_user_group_for_setting(
setting_value_group_id,
user_profile,
setting_name=setting_name,
require_system_group=permission_config.require_system_group,
allow_internet_group=permission_config.allow_internet_group,
allow_owners_group=permission_config.allow_owners_group,
allow_nobody_group=permission_config.allow_nobody_group,
allow_everyone_group=permission_config.allow_everyone_group,
)
group_settings_map[setting_name] = setting_value_group
check_add_user_group(
user_profile.realm,
name,
user_profiles,
description,
group_settings_map=group_settings_map,
acting_user=user_profile,
)
return json_success(request)
@require_member_or_admin
@has_request_variables
def get_user_group(request: HttpRequest, user_profile: UserProfile) -> HttpResponse:
user_groups = user_groups_in_realm_serialized(user_profile.realm)
return json_success(request, data={"user_groups": user_groups})
user_groups: Make locks required for updating user group memberships. **Background** User groups are expected to comply with the DAG constraint for the many-to-many inter-group membership. The check for this constraint has to be performed recursively so that we can find all direct and indirect subgroups of the user group to be added. This kind of check is vulnerable to phantom reads which is possible at the default read committed isolation level because we cannot guarantee that the check is still valid when we are adding the subgroups to the user group. **Solution** To avoid having another transaction concurrently update one of the to-be-subgroup after the recursive check is done, and before the subgroup is added, we use SELECT FOR UPDATE to lock the user group rows. The lock needs to be acquired before a group membership change is about to occur before any check has been conducted. Suppose that we are adding subgroup B to supergroup A, the locking protocol is specified as follows: 1. Acquire a lock for B and all its direct and indirect subgroups. 2. Acquire a lock for A. For the removal of user groups, we acquire a lock for the user group to be removed with all its direct and indirect subgroups. This is the special case A=B, which is still complaint with the protocol. **Error handling** We currently rely on Postgres' deadlock detection to abort transactions and show an error for the users. In the future, we might need some recovery mechanism or at least better error handling. **Notes** An important note is that we need to reuse the recursive CTE query that finds the direct and indirect subgroups when applying the lock on the rows. And the lock needs to be acquired the same way for the addition and removal of direct subgroups. User membership change (as opposed to user group membership) is not affected. Read-only queries aren't either. The locks only protect critical regions where the user group dependency graph might violate the DAG constraint, where users are not participating. **Testing** We implement a transaction test case targeting some typical scenarios when an internal server error is expected to happen (this means that the user group view makes the correct decision to abort the transaction when something goes wrong with locks). To achieve this, we add a development view intended only for unit tests. It has a global BARRIER that can be shared across threads, so that we can synchronize them to consistently reproduce certain potential race conditions prevented by the database locks. The transaction test case lanuches pairs of threads initiating possibly conflicting requests at the same time. The tests are set up such that exactly N of them are expected to succeed with a certain error message (while we don't know each one). **Security notes** get_recursive_subgroups_for_groups will no longer fetch user groups from other realms. As a result, trying to add/remove a subgroup from another realm results in a UserGroup not found error response. We also implement subgroup-specific checks in has_user_group_access to keep permission managing in a single place. Do note that the API currently don't have a way to violate that check because we are only checking the realm ID now.
2023-06-17 04:39:52 +02:00
@transaction.atomic
@require_user_group_edit_permission
@has_request_variables
def edit_user_group(
request: HttpRequest,
user_profile: UserProfile,
user_group_id: int = REQ(json_validator=check_int, path_only=True),
name: Optional[str] = REQ(default=None),
description: Optional[str] = REQ(default=None),
can_mention_group_id: Optional[int] = REQ(
"can_mention_group", json_validator=check_int, default=None
),
) -> HttpResponse:
if name is None and description is None and can_mention_group_id is None:
raise JsonableError(_("No new data supplied"))
user_group = access_user_group_by_id(user_group_id, user_profile, for_read=False)
if name is not None and name != user_group.name:
name = check_user_group_name(name)
do_update_user_group_name(user_group, name, acting_user=user_profile)
if description is not None and description != user_group.description:
do_update_user_group_description(user_group, description, acting_user=user_profile)
request_settings_dict = locals()
for setting_name, permission_config in UserGroup.GROUP_PERMISSION_SETTINGS.items():
setting_group_id_name = permission_config.id_field_name
if setting_group_id_name not in request_settings_dict: # nocoverage
continue
if request_settings_dict[setting_group_id_name] is not None and request_settings_dict[
setting_group_id_name
] != getattr(user_group, setting_group_id_name):
setting_value_group_id = request_settings_dict[setting_group_id_name]
setting_value_group = access_user_group_for_setting(
setting_value_group_id,
user_profile,
setting_name=setting_name,
require_system_group=permission_config.require_system_group,
allow_internet_group=permission_config.allow_internet_group,
allow_owners_group=permission_config.allow_owners_group,
allow_nobody_group=permission_config.allow_nobody_group,
allow_everyone_group=permission_config.allow_everyone_group,
)
do_change_user_group_permission_setting(
user_group, setting_name, setting_value_group, acting_user=user_profile
)
return json_success(request)
@require_user_group_edit_permission
@has_request_variables
def delete_user_group(
request: HttpRequest,
user_profile: UserProfile,
user_group_id: int = REQ(json_validator=check_int, path_only=True),
) -> HttpResponse:
user_groups: Make locks required for updating user group memberships. **Background** User groups are expected to comply with the DAG constraint for the many-to-many inter-group membership. The check for this constraint has to be performed recursively so that we can find all direct and indirect subgroups of the user group to be added. This kind of check is vulnerable to phantom reads which is possible at the default read committed isolation level because we cannot guarantee that the check is still valid when we are adding the subgroups to the user group. **Solution** To avoid having another transaction concurrently update one of the to-be-subgroup after the recursive check is done, and before the subgroup is added, we use SELECT FOR UPDATE to lock the user group rows. The lock needs to be acquired before a group membership change is about to occur before any check has been conducted. Suppose that we are adding subgroup B to supergroup A, the locking protocol is specified as follows: 1. Acquire a lock for B and all its direct and indirect subgroups. 2. Acquire a lock for A. For the removal of user groups, we acquire a lock for the user group to be removed with all its direct and indirect subgroups. This is the special case A=B, which is still complaint with the protocol. **Error handling** We currently rely on Postgres' deadlock detection to abort transactions and show an error for the users. In the future, we might need some recovery mechanism or at least better error handling. **Notes** An important note is that we need to reuse the recursive CTE query that finds the direct and indirect subgroups when applying the lock on the rows. And the lock needs to be acquired the same way for the addition and removal of direct subgroups. User membership change (as opposed to user group membership) is not affected. Read-only queries aren't either. The locks only protect critical regions where the user group dependency graph might violate the DAG constraint, where users are not participating. **Testing** We implement a transaction test case targeting some typical scenarios when an internal server error is expected to happen (this means that the user group view makes the correct decision to abort the transaction when something goes wrong with locks). To achieve this, we add a development view intended only for unit tests. It has a global BARRIER that can be shared across threads, so that we can synchronize them to consistently reproduce certain potential race conditions prevented by the database locks. The transaction test case lanuches pairs of threads initiating possibly conflicting requests at the same time. The tests are set up such that exactly N of them are expected to succeed with a certain error message (while we don't know each one). **Security notes** get_recursive_subgroups_for_groups will no longer fetch user groups from other realms. As a result, trying to add/remove a subgroup from another realm results in a UserGroup not found error response. We also implement subgroup-specific checks in has_user_group_access to keep permission managing in a single place. Do note that the API currently don't have a way to violate that check because we are only checking the realm ID now.
2023-06-17 04:39:52 +02:00
# For deletion, the user group's recursive subgroups and the user group itself are locked.
with lock_subgroups_with_respect_to_supergroup(
[user_group_id], user_group_id, acting_user=user_profile
) as context:
check_delete_user_group(context.supergroup, acting_user=user_profile)
return json_success(request)
@require_user_group_edit_permission
@has_request_variables
def update_user_group_backend(
request: HttpRequest,
user_profile: UserProfile,
user_group_id: int = REQ(json_validator=check_int, path_only=True),
delete: Sequence[int] = REQ(json_validator=check_list(check_int), default=[]),
add: Sequence[int] = REQ(json_validator=check_list(check_int), default=[]),
) -> HttpResponse:
if not add and not delete:
raise JsonableError(_('Nothing to do. Specify at least one of "add" or "delete".'))
thunks = [
lambda: add_members_to_group_backend(
request, user_profile, user_group_id=user_group_id, members=add
),
lambda: remove_members_from_group_backend(
request, user_profile, user_group_id=user_group_id, members=delete
),
python: Convert assignment type annotations to Python 3.6 style. This commit was split by tabbott; this piece covers the vast majority of files in Zulip, but excludes scripts/, tools/, and puppet/ to help ensure we at least show the right error messages for Xenial systems. We can likely further refine the remaining pieces with some testing. Generated by com2ann, with whitespace fixes and various manual fixes for runtime issues: - invoiced_through: Optional[LicenseLedger] = models.ForeignKey( + invoiced_through: Optional["LicenseLedger"] = models.ForeignKey( -_apns_client: Optional[APNsClient] = None +_apns_client: Optional["APNsClient"] = None - notifications_stream: Optional[Stream] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE) - signup_notifications_stream: Optional[Stream] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE) + notifications_stream: Optional["Stream"] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE) + signup_notifications_stream: Optional["Stream"] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE) - author: Optional[UserProfile] = models.ForeignKey('UserProfile', blank=True, null=True, on_delete=CASCADE) + author: Optional["UserProfile"] = models.ForeignKey('UserProfile', blank=True, null=True, on_delete=CASCADE) - bot_owner: Optional[UserProfile] = models.ForeignKey('self', null=True, on_delete=models.SET_NULL) + bot_owner: Optional["UserProfile"] = models.ForeignKey('self', null=True, on_delete=models.SET_NULL) - default_sending_stream: Optional[Stream] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE) - default_events_register_stream: Optional[Stream] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE) + default_sending_stream: Optional["Stream"] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE) + default_events_register_stream: Optional["Stream"] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE) -descriptors_by_handler_id: Dict[int, ClientDescriptor] = {} +descriptors_by_handler_id: Dict[int, "ClientDescriptor"] = {} -worker_classes: Dict[str, Type[QueueProcessingWorker]] = {} -queues: Dict[str, Dict[str, Type[QueueProcessingWorker]]] = {} +worker_classes: Dict[str, Type["QueueProcessingWorker"]] = {} +queues: Dict[str, Dict[str, Type["QueueProcessingWorker"]]] = {} -AUTH_LDAP_REVERSE_EMAIL_SEARCH: Optional[LDAPSearch] = None +AUTH_LDAP_REVERSE_EMAIL_SEARCH: Optional["LDAPSearch"] = None Signed-off-by: Anders Kaseorg <anders@zulipchat.com>
2020-04-22 01:09:50 +02:00
]
data = compose_views(thunks)
return json_success(request, data)
def notify_for_user_group_subscription_changes(
acting_user: UserProfile,
recipient_users: List[UserProfile],
user_group: UserGroup,
*,
send_subscription_message: bool = False,
send_unsubscription_message: bool = False,
) -> None:
realm = acting_user.realm
mention_backend = MentionBackend(realm.id)
notifications = []
notification_bot = get_system_bot(settings.NOTIFICATION_BOT, realm.id)
for recipient_user in recipient_users:
if recipient_user.id == acting_user.id:
# Don't send notification message if you subscribed/unsubscribed yourself.
continue
if recipient_user.is_bot:
# Don't send notification message to bots.
continue
if not recipient_user.is_active:
# Don't send notification message to deactivated users.
continue
with override_language(recipient_user.default_language):
if send_subscription_message:
message = _("{user_full_name} added you to the group {group_name}.").format(
user_full_name=silent_mention_syntax_for_user(acting_user),
group_name=f"@_*{user_group.name}*",
)
if send_unsubscription_message:
message = _("{user_full_name} removed you from the group {group_name}.").format(
user_full_name=silent_mention_syntax_for_user(acting_user),
group_name=f"@_*{user_group.name}*",
)
notifications.append(
internal_prep_private_message(
sender=notification_bot,
recipient_user=recipient_user,
content=message,
mention_backend=mention_backend,
)
)
if len(notifications) > 0:
do_send_messages(notifications)
user_groups: Make locks required for updating user group memberships. **Background** User groups are expected to comply with the DAG constraint for the many-to-many inter-group membership. The check for this constraint has to be performed recursively so that we can find all direct and indirect subgroups of the user group to be added. This kind of check is vulnerable to phantom reads which is possible at the default read committed isolation level because we cannot guarantee that the check is still valid when we are adding the subgroups to the user group. **Solution** To avoid having another transaction concurrently update one of the to-be-subgroup after the recursive check is done, and before the subgroup is added, we use SELECT FOR UPDATE to lock the user group rows. The lock needs to be acquired before a group membership change is about to occur before any check has been conducted. Suppose that we are adding subgroup B to supergroup A, the locking protocol is specified as follows: 1. Acquire a lock for B and all its direct and indirect subgroups. 2. Acquire a lock for A. For the removal of user groups, we acquire a lock for the user group to be removed with all its direct and indirect subgroups. This is the special case A=B, which is still complaint with the protocol. **Error handling** We currently rely on Postgres' deadlock detection to abort transactions and show an error for the users. In the future, we might need some recovery mechanism or at least better error handling. **Notes** An important note is that we need to reuse the recursive CTE query that finds the direct and indirect subgroups when applying the lock on the rows. And the lock needs to be acquired the same way for the addition and removal of direct subgroups. User membership change (as opposed to user group membership) is not affected. Read-only queries aren't either. The locks only protect critical regions where the user group dependency graph might violate the DAG constraint, where users are not participating. **Testing** We implement a transaction test case targeting some typical scenarios when an internal server error is expected to happen (this means that the user group view makes the correct decision to abort the transaction when something goes wrong with locks). To achieve this, we add a development view intended only for unit tests. It has a global BARRIER that can be shared across threads, so that we can synchronize them to consistently reproduce certain potential race conditions prevented by the database locks. The transaction test case lanuches pairs of threads initiating possibly conflicting requests at the same time. The tests are set up such that exactly N of them are expected to succeed with a certain error message (while we don't know each one). **Security notes** get_recursive_subgroups_for_groups will no longer fetch user groups from other realms. As a result, trying to add/remove a subgroup from another realm results in a UserGroup not found error response. We also implement subgroup-specific checks in has_user_group_access to keep permission managing in a single place. Do note that the API currently don't have a way to violate that check because we are only checking the realm ID now.
2023-06-17 04:39:52 +02:00
@transaction.atomic
def add_members_to_group_backend(
request: HttpRequest, user_profile: UserProfile, user_group_id: int, members: Sequence[int]
) -> HttpResponse:
if not members:
return json_success(request)
user_group = access_user_group_by_id(user_group_id, user_profile, for_read=False)
member_users = user_ids_to_users(members, user_profile.realm)
existing_member_ids = set(get_direct_memberships_of_users(user_group, member_users))
for member_user in member_users:
if member_user.id in existing_member_ids:
raise JsonableError(
_("User {user_id} is already a member of this group").format(
user_id=member_user.id,
)
)
member_user_ids = [member_user.id for member_user in member_users]
bulk_add_members_to_user_group(user_group, member_user_ids, acting_user=user_profile)
notify_for_user_group_subscription_changes(
acting_user=user_profile,
recipient_users=member_users,
user_group=user_group,
send_subscription_message=True,
)
return json_success(request)
user_groups: Make locks required for updating user group memberships. **Background** User groups are expected to comply with the DAG constraint for the many-to-many inter-group membership. The check for this constraint has to be performed recursively so that we can find all direct and indirect subgroups of the user group to be added. This kind of check is vulnerable to phantom reads which is possible at the default read committed isolation level because we cannot guarantee that the check is still valid when we are adding the subgroups to the user group. **Solution** To avoid having another transaction concurrently update one of the to-be-subgroup after the recursive check is done, and before the subgroup is added, we use SELECT FOR UPDATE to lock the user group rows. The lock needs to be acquired before a group membership change is about to occur before any check has been conducted. Suppose that we are adding subgroup B to supergroup A, the locking protocol is specified as follows: 1. Acquire a lock for B and all its direct and indirect subgroups. 2. Acquire a lock for A. For the removal of user groups, we acquire a lock for the user group to be removed with all its direct and indirect subgroups. This is the special case A=B, which is still complaint with the protocol. **Error handling** We currently rely on Postgres' deadlock detection to abort transactions and show an error for the users. In the future, we might need some recovery mechanism or at least better error handling. **Notes** An important note is that we need to reuse the recursive CTE query that finds the direct and indirect subgroups when applying the lock on the rows. And the lock needs to be acquired the same way for the addition and removal of direct subgroups. User membership change (as opposed to user group membership) is not affected. Read-only queries aren't either. The locks only protect critical regions where the user group dependency graph might violate the DAG constraint, where users are not participating. **Testing** We implement a transaction test case targeting some typical scenarios when an internal server error is expected to happen (this means that the user group view makes the correct decision to abort the transaction when something goes wrong with locks). To achieve this, we add a development view intended only for unit tests. It has a global BARRIER that can be shared across threads, so that we can synchronize them to consistently reproduce certain potential race conditions prevented by the database locks. The transaction test case lanuches pairs of threads initiating possibly conflicting requests at the same time. The tests are set up such that exactly N of them are expected to succeed with a certain error message (while we don't know each one). **Security notes** get_recursive_subgroups_for_groups will no longer fetch user groups from other realms. As a result, trying to add/remove a subgroup from another realm results in a UserGroup not found error response. We also implement subgroup-specific checks in has_user_group_access to keep permission managing in a single place. Do note that the API currently don't have a way to violate that check because we are only checking the realm ID now.
2023-06-17 04:39:52 +02:00
@transaction.atomic
def remove_members_from_group_backend(
request: HttpRequest, user_profile: UserProfile, user_group_id: int, members: Sequence[int]
) -> HttpResponse:
if not members:
return json_success(request)
user_profiles = user_ids_to_users(members, user_profile.realm)
user_group = access_user_group_by_id(user_group_id, user_profile, for_read=False)
group_member_ids = get_user_group_direct_member_ids(user_group)
for member in members:
if member not in group_member_ids:
raise JsonableError(
_("There is no member '{user_id}' in this user group").format(user_id=member)
)
user_profile_ids = [user.id for user in user_profiles]
remove_members_from_user_group(user_group, user_profile_ids, acting_user=user_profile)
notify_for_user_group_subscription_changes(
acting_user=user_profile,
recipient_users=user_profiles,
user_group=user_group,
send_unsubscription_message=True,
)
return json_success(request)
def add_subgroups_to_group_backend(
request: HttpRequest, user_profile: UserProfile, user_group_id: int, subgroup_ids: Sequence[int]
) -> HttpResponse:
if not subgroup_ids:
return json_success(request)
user_groups: Make locks required for updating user group memberships. **Background** User groups are expected to comply with the DAG constraint for the many-to-many inter-group membership. The check for this constraint has to be performed recursively so that we can find all direct and indirect subgroups of the user group to be added. This kind of check is vulnerable to phantom reads which is possible at the default read committed isolation level because we cannot guarantee that the check is still valid when we are adding the subgroups to the user group. **Solution** To avoid having another transaction concurrently update one of the to-be-subgroup after the recursive check is done, and before the subgroup is added, we use SELECT FOR UPDATE to lock the user group rows. The lock needs to be acquired before a group membership change is about to occur before any check has been conducted. Suppose that we are adding subgroup B to supergroup A, the locking protocol is specified as follows: 1. Acquire a lock for B and all its direct and indirect subgroups. 2. Acquire a lock for A. For the removal of user groups, we acquire a lock for the user group to be removed with all its direct and indirect subgroups. This is the special case A=B, which is still complaint with the protocol. **Error handling** We currently rely on Postgres' deadlock detection to abort transactions and show an error for the users. In the future, we might need some recovery mechanism or at least better error handling. **Notes** An important note is that we need to reuse the recursive CTE query that finds the direct and indirect subgroups when applying the lock on the rows. And the lock needs to be acquired the same way for the addition and removal of direct subgroups. User membership change (as opposed to user group membership) is not affected. Read-only queries aren't either. The locks only protect critical regions where the user group dependency graph might violate the DAG constraint, where users are not participating. **Testing** We implement a transaction test case targeting some typical scenarios when an internal server error is expected to happen (this means that the user group view makes the correct decision to abort the transaction when something goes wrong with locks). To achieve this, we add a development view intended only for unit tests. It has a global BARRIER that can be shared across threads, so that we can synchronize them to consistently reproduce certain potential race conditions prevented by the database locks. The transaction test case lanuches pairs of threads initiating possibly conflicting requests at the same time. The tests are set up such that exactly N of them are expected to succeed with a certain error message (while we don't know each one). **Security notes** get_recursive_subgroups_for_groups will no longer fetch user groups from other realms. As a result, trying to add/remove a subgroup from another realm results in a UserGroup not found error response. We also implement subgroup-specific checks in has_user_group_access to keep permission managing in a single place. Do note that the API currently don't have a way to violate that check because we are only checking the realm ID now.
2023-06-17 04:39:52 +02:00
with lock_subgroups_with_respect_to_supergroup(
subgroup_ids, user_group_id, user_profile
) as context:
existing_direct_subgroup_ids = context.supergroup.direct_subgroups.all().values_list(
"id", flat=True
)
for group in context.direct_subgroups:
if group.id in existing_direct_subgroup_ids:
raise JsonableError(
_("User group {group_id} is already a subgroup of this group.").format(
group_id=group.id
)
)
user_groups: Make locks required for updating user group memberships. **Background** User groups are expected to comply with the DAG constraint for the many-to-many inter-group membership. The check for this constraint has to be performed recursively so that we can find all direct and indirect subgroups of the user group to be added. This kind of check is vulnerable to phantom reads which is possible at the default read committed isolation level because we cannot guarantee that the check is still valid when we are adding the subgroups to the user group. **Solution** To avoid having another transaction concurrently update one of the to-be-subgroup after the recursive check is done, and before the subgroup is added, we use SELECT FOR UPDATE to lock the user group rows. The lock needs to be acquired before a group membership change is about to occur before any check has been conducted. Suppose that we are adding subgroup B to supergroup A, the locking protocol is specified as follows: 1. Acquire a lock for B and all its direct and indirect subgroups. 2. Acquire a lock for A. For the removal of user groups, we acquire a lock for the user group to be removed with all its direct and indirect subgroups. This is the special case A=B, which is still complaint with the protocol. **Error handling** We currently rely on Postgres' deadlock detection to abort transactions and show an error for the users. In the future, we might need some recovery mechanism or at least better error handling. **Notes** An important note is that we need to reuse the recursive CTE query that finds the direct and indirect subgroups when applying the lock on the rows. And the lock needs to be acquired the same way for the addition and removal of direct subgroups. User membership change (as opposed to user group membership) is not affected. Read-only queries aren't either. The locks only protect critical regions where the user group dependency graph might violate the DAG constraint, where users are not participating. **Testing** We implement a transaction test case targeting some typical scenarios when an internal server error is expected to happen (this means that the user group view makes the correct decision to abort the transaction when something goes wrong with locks). To achieve this, we add a development view intended only for unit tests. It has a global BARRIER that can be shared across threads, so that we can synchronize them to consistently reproduce certain potential race conditions prevented by the database locks. The transaction test case lanuches pairs of threads initiating possibly conflicting requests at the same time. The tests are set up such that exactly N of them are expected to succeed with a certain error message (while we don't know each one). **Security notes** get_recursive_subgroups_for_groups will no longer fetch user groups from other realms. As a result, trying to add/remove a subgroup from another realm results in a UserGroup not found error response. We also implement subgroup-specific checks in has_user_group_access to keep permission managing in a single place. Do note that the API currently don't have a way to violate that check because we are only checking the realm ID now.
2023-06-17 04:39:52 +02:00
recursive_subgroup_ids = {
recursive_subgroup.id for recursive_subgroup in context.recursive_subgroups
}
if user_group_id in recursive_subgroup_ids:
raise JsonableError(
_(
"User group {user_group_id} is already a subgroup of one of the passed subgroups."
).format(user_group_id=user_group_id)
)
user_groups: Make locks required for updating user group memberships. **Background** User groups are expected to comply with the DAG constraint for the many-to-many inter-group membership. The check for this constraint has to be performed recursively so that we can find all direct and indirect subgroups of the user group to be added. This kind of check is vulnerable to phantom reads which is possible at the default read committed isolation level because we cannot guarantee that the check is still valid when we are adding the subgroups to the user group. **Solution** To avoid having another transaction concurrently update one of the to-be-subgroup after the recursive check is done, and before the subgroup is added, we use SELECT FOR UPDATE to lock the user group rows. The lock needs to be acquired before a group membership change is about to occur before any check has been conducted. Suppose that we are adding subgroup B to supergroup A, the locking protocol is specified as follows: 1. Acquire a lock for B and all its direct and indirect subgroups. 2. Acquire a lock for A. For the removal of user groups, we acquire a lock for the user group to be removed with all its direct and indirect subgroups. This is the special case A=B, which is still complaint with the protocol. **Error handling** We currently rely on Postgres' deadlock detection to abort transactions and show an error for the users. In the future, we might need some recovery mechanism or at least better error handling. **Notes** An important note is that we need to reuse the recursive CTE query that finds the direct and indirect subgroups when applying the lock on the rows. And the lock needs to be acquired the same way for the addition and removal of direct subgroups. User membership change (as opposed to user group membership) is not affected. Read-only queries aren't either. The locks only protect critical regions where the user group dependency graph might violate the DAG constraint, where users are not participating. **Testing** We implement a transaction test case targeting some typical scenarios when an internal server error is expected to happen (this means that the user group view makes the correct decision to abort the transaction when something goes wrong with locks). To achieve this, we add a development view intended only for unit tests. It has a global BARRIER that can be shared across threads, so that we can synchronize them to consistently reproduce certain potential race conditions prevented by the database locks. The transaction test case lanuches pairs of threads initiating possibly conflicting requests at the same time. The tests are set up such that exactly N of them are expected to succeed with a certain error message (while we don't know each one). **Security notes** get_recursive_subgroups_for_groups will no longer fetch user groups from other realms. As a result, trying to add/remove a subgroup from another realm results in a UserGroup not found error response. We also implement subgroup-specific checks in has_user_group_access to keep permission managing in a single place. Do note that the API currently don't have a way to violate that check because we are only checking the realm ID now.
2023-06-17 04:39:52 +02:00
add_subgroups_to_user_group(
context.supergroup, context.direct_subgroups, acting_user=user_profile
)
return json_success(request)
def remove_subgroups_from_group_backend(
request: HttpRequest, user_profile: UserProfile, user_group_id: int, subgroup_ids: Sequence[int]
) -> HttpResponse:
if not subgroup_ids:
return json_success(request)
user_groups: Make locks required for updating user group memberships. **Background** User groups are expected to comply with the DAG constraint for the many-to-many inter-group membership. The check for this constraint has to be performed recursively so that we can find all direct and indirect subgroups of the user group to be added. This kind of check is vulnerable to phantom reads which is possible at the default read committed isolation level because we cannot guarantee that the check is still valid when we are adding the subgroups to the user group. **Solution** To avoid having another transaction concurrently update one of the to-be-subgroup after the recursive check is done, and before the subgroup is added, we use SELECT FOR UPDATE to lock the user group rows. The lock needs to be acquired before a group membership change is about to occur before any check has been conducted. Suppose that we are adding subgroup B to supergroup A, the locking protocol is specified as follows: 1. Acquire a lock for B and all its direct and indirect subgroups. 2. Acquire a lock for A. For the removal of user groups, we acquire a lock for the user group to be removed with all its direct and indirect subgroups. This is the special case A=B, which is still complaint with the protocol. **Error handling** We currently rely on Postgres' deadlock detection to abort transactions and show an error for the users. In the future, we might need some recovery mechanism or at least better error handling. **Notes** An important note is that we need to reuse the recursive CTE query that finds the direct and indirect subgroups when applying the lock on the rows. And the lock needs to be acquired the same way for the addition and removal of direct subgroups. User membership change (as opposed to user group membership) is not affected. Read-only queries aren't either. The locks only protect critical regions where the user group dependency graph might violate the DAG constraint, where users are not participating. **Testing** We implement a transaction test case targeting some typical scenarios when an internal server error is expected to happen (this means that the user group view makes the correct decision to abort the transaction when something goes wrong with locks). To achieve this, we add a development view intended only for unit tests. It has a global BARRIER that can be shared across threads, so that we can synchronize them to consistently reproduce certain potential race conditions prevented by the database locks. The transaction test case lanuches pairs of threads initiating possibly conflicting requests at the same time. The tests are set up such that exactly N of them are expected to succeed with a certain error message (while we don't know each one). **Security notes** get_recursive_subgroups_for_groups will no longer fetch user groups from other realms. As a result, trying to add/remove a subgroup from another realm results in a UserGroup not found error response. We also implement subgroup-specific checks in has_user_group_access to keep permission managing in a single place. Do note that the API currently don't have a way to violate that check because we are only checking the realm ID now.
2023-06-17 04:39:52 +02:00
with lock_subgroups_with_respect_to_supergroup(
subgroup_ids, user_group_id, user_profile
) as context:
# While the recursive subgroups in the context are not used, it is important that
# we acquire a lock for these rows while updating the subgroups to acquire the locks
# in a consistent order for subgroup membership changes.
existing_direct_subgroup_ids = context.supergroup.direct_subgroups.all().values_list(
"id", flat=True
)
for group in context.direct_subgroups:
if group.id not in existing_direct_subgroup_ids:
raise JsonableError(
_("User group {group_id} is not a subgroup of this group.").format(
group_id=group.id
)
)
user_groups: Make locks required for updating user group memberships. **Background** User groups are expected to comply with the DAG constraint for the many-to-many inter-group membership. The check for this constraint has to be performed recursively so that we can find all direct and indirect subgroups of the user group to be added. This kind of check is vulnerable to phantom reads which is possible at the default read committed isolation level because we cannot guarantee that the check is still valid when we are adding the subgroups to the user group. **Solution** To avoid having another transaction concurrently update one of the to-be-subgroup after the recursive check is done, and before the subgroup is added, we use SELECT FOR UPDATE to lock the user group rows. The lock needs to be acquired before a group membership change is about to occur before any check has been conducted. Suppose that we are adding subgroup B to supergroup A, the locking protocol is specified as follows: 1. Acquire a lock for B and all its direct and indirect subgroups. 2. Acquire a lock for A. For the removal of user groups, we acquire a lock for the user group to be removed with all its direct and indirect subgroups. This is the special case A=B, which is still complaint with the protocol. **Error handling** We currently rely on Postgres' deadlock detection to abort transactions and show an error for the users. In the future, we might need some recovery mechanism or at least better error handling. **Notes** An important note is that we need to reuse the recursive CTE query that finds the direct and indirect subgroups when applying the lock on the rows. And the lock needs to be acquired the same way for the addition and removal of direct subgroups. User membership change (as opposed to user group membership) is not affected. Read-only queries aren't either. The locks only protect critical regions where the user group dependency graph might violate the DAG constraint, where users are not participating. **Testing** We implement a transaction test case targeting some typical scenarios when an internal server error is expected to happen (this means that the user group view makes the correct decision to abort the transaction when something goes wrong with locks). To achieve this, we add a development view intended only for unit tests. It has a global BARRIER that can be shared across threads, so that we can synchronize them to consistently reproduce certain potential race conditions prevented by the database locks. The transaction test case lanuches pairs of threads initiating possibly conflicting requests at the same time. The tests are set up such that exactly N of them are expected to succeed with a certain error message (while we don't know each one). **Security notes** get_recursive_subgroups_for_groups will no longer fetch user groups from other realms. As a result, trying to add/remove a subgroup from another realm results in a UserGroup not found error response. We also implement subgroup-specific checks in has_user_group_access to keep permission managing in a single place. Do note that the API currently don't have a way to violate that check because we are only checking the realm ID now.
2023-06-17 04:39:52 +02:00
remove_subgroups_from_user_group(
context.supergroup, context.direct_subgroups, acting_user=user_profile
)
return json_success(request)
@require_user_group_edit_permission
@has_request_variables
def update_subgroups_of_user_group(
request: HttpRequest,
user_profile: UserProfile,
user_group_id: int = REQ(json_validator=check_int, path_only=True),
delete: Sequence[int] = REQ(json_validator=check_list(check_int), default=[]),
add: Sequence[int] = REQ(json_validator=check_list(check_int), default=[]),
) -> HttpResponse:
if not add and not delete:
raise JsonableError(_('Nothing to do. Specify at least one of "add" or "delete".'))
thunks = [
lambda: add_subgroups_to_group_backend(
request, user_profile, user_group_id=user_group_id, subgroup_ids=add
),
lambda: remove_subgroups_from_group_backend(
request, user_profile, user_group_id=user_group_id, subgroup_ids=delete
),
]
data = compose_views(thunks)
return json_success(request, data)
@require_member_or_admin
@has_request_variables
def get_is_user_group_member(
request: HttpRequest,
user_profile: UserProfile,
user_group_id: int = REQ(json_validator=check_int, path_only=True),
user_id: int = REQ(json_validator=check_int, path_only=True),
direct_member_only: bool = REQ(json_validator=check_bool, default=False),
) -> HttpResponse:
user_group = access_user_group_by_id(user_group_id, user_profile, for_read=True)
target_user = access_user_by_id(user_profile, user_id, for_admin=False)
return json_success(
request,
data={
"is_user_group_member": is_user_in_group(
user_group, target_user, direct_member_only=direct_member_only
)
},
)
@require_member_or_admin
@has_request_variables
def get_user_group_members(
request: HttpRequest,
user_profile: UserProfile,
user_group_id: int = REQ(json_validator=check_int, path_only=True),
direct_member_only: bool = REQ(json_validator=check_bool, default=False),
) -> HttpResponse:
user_group = access_user_group_by_id(user_group_id, user_profile, for_read=True)
return json_success(
request,
data={
"members": get_user_group_member_ids(user_group, direct_member_only=direct_member_only)
},
)
@require_member_or_admin
@has_request_variables
def get_subgroups_of_user_group(
request: HttpRequest,
user_profile: UserProfile,
user_group_id: int = REQ(json_validator=check_int, path_only=True),
direct_subgroup_only: bool = REQ(json_validator=check_bool, default=False),
) -> HttpResponse:
user_group = access_user_group_by_id(user_group_id, user_profile, for_read=True)
return json_success(
request,
data={"subgroups": get_subgroup_ids(user_group, direct_subgroup_only=direct_subgroup_only)},
)