PostgreSQL RLS 显式策略

Question

从安全的角度来看，以下策略的实施是否等效？

隐式：

CREATE POLICY test_access_policy ON test 
  TO PUBLIC 
  USING (id = (current_setting('rls.id'::TEXT))) 
  WITH CHECK (TRUE);

显式：

CREATE POLICY test_insert_policy ON test 
  FOR INSERT TO PUBLIC
  WITH CHECK (TRUE);   

CREATE POLICY test_select_policy ON test
  FOR SELECT TO PUBLIC 
  USING (id = (current_setting('rls.id'::TEXT)));  

CREATE POLICY test_update_policy ON test 
  FOR UPDATE TO PUBLIC
  USING (id = (current_setting('rls.id'::TEXT)));   

CREATE POLICY test_delete_policy ON test 
  FOR DELETE TO PUBLIC 
USING (id = (current_setting('rls.id'::TEXT)));

我担心的是更新政策，如 docs 中所述：

Any rows whose updated values do not pass the WITH CHECK expression will cause an error, and the entire command will be aborted. If only a USING clause is specified, then that clause will be used for both USING and WITH CHECK cases.

根据我的理解，隐式版本（oneliner）的等效更新策略如下：

CREATE POLICY test_update_policy ON test 
  FOR UPDATE TO PUBLIC 
  USING (id = (current_setting('rls.id'::TEXT))) WITH CHECK (TRUE);

而显式版本是：

CREATE POLICY test_update_policy ON test 
  FOR UPDATE TO PUBLIC 
  USING (id = (current_setting('rls.id'::TEXT))) WITH CHECK (id =
(current_setting('rls.id'::TEXT)));

在测试这两种情况后，我没有找到任何安全桥，我是否遗漏了什么？

Answer 1

您的解释是正确的，您必须将 WITH CHECK (TRUE) 添加到 FOR UPDATE 政策才能获得等效定义。

不同之处在于 WITH CHECK (TRUE) 允许您将值更改为任何值，如果没有它，如果新行版本与条件不匹配，您将收到错误消息。

PostgreSQL RLS 显式策略

PostgreSQL RLS explicit Policies

postgresql

security-policy

rls