netdev - Re: [PATCH v2 net-next 9/9] l2tp: flush workqueue before draining it

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <6f0e2c8e-205c-71ae-2b93-02538122097a@katalix.com>
Date: Mon, 12 Aug 2024 10:24:11 +0100
From: James Chapman <jchapman@...alix.com>
To: Cong Wang <xiyou.wangcong@...il.com>
Cc: netdev@...r.kernel.org, davem@...emloft.net, edumazet@...gle.com,
 kuba@...nel.org, pabeni@...hat.com, dsahern@...nel.org, tparkin@...alix.com,
 horms@...nel.org, syzbot+0e85b10481d2f5478053@...kaller.appspotmail.com
Subject: Re: [PATCH v2 net-next 9/9] l2tp: flush workqueue before draining it

On 11/08/2024 18:54, Cong Wang wrote:
> On Wed, Aug 07, 2024 at 07:54:52AM +0100, James Chapman wrote:
>> syzbot exposes a race where a net used by l2tp is removed while an
>> existing pppol2tp socket is closed. In l2tp_pre_exit_net, l2tp queues
>> TUNNEL_DELETE work items to close each tunnel in the net. When these
>> are run, new SESSION_DELETE work items are queued to delete each
>> session in the tunnel. This all happens in drain_workqueue. However,
>> drain_workqueue allows only new work items if they are queued by other
>> work items which are already in the queue. If pppol2tp_release runs
>> after drain_workqueue has started, it may queue a SESSION_DELETE work
>> item, which results in the warning below in drain_workqueue.
>>
>> Address this by flushing the workqueue before drain_workqueue such
>> that all queued TUNNEL_DELETE work items run before drain_workqueue is
>> started. This will queue SESSION_DELETE work items for each session in
>> the tunnel, hence pppol2tp_release or other API requests won't queue
>> SESSION_DELETE requests once drain_workqueue is started.
>>
> 
> I am not convinced here.
> 
> 1) There is a __flush_workqueue() inside drain_workqueue() too. Why
> calling it again?

Once drain_workqueue starts, it considers any new work item queued 
outside of its processing as an unexpected condition. By doing 
__flush_workqueue first, we ensure that all TUNNEL_DELETE items are 
processed, which means that SESSION_DELETE work items are queued for all 
existing sessions before drain_workqueue starts. Now if an API request 
is processed to delete a session, l2tp_session_delete will do nothing 
because session->dead has already been set.

> 2) What prevents new work items be queued right between your
> __flush_workqueue() and drain_workqueue()?

Hmm, to do so, a new tunnel would need to be created and then deleted by 
API request once drain_workqueue starts. Is it possible to create new 
sockets in a net once a net's pre_exit starts? I didn't think so, but 
I'll recheck. Thanks!