Received-SPF: pass (google.com: domain of linux-kernel+bounces-206515-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45d1:ec00::1 as permitted sender) client-ip=2604:1380:45d1:ec00::1;
Sender: Tejun Heo <htejun@gmail.com>
Date: Fri, 7 Jun 2024 06:53:42 -1000
From: Tejun Heo <tj@kernel.org>
To: Tim Van Patten <timvp@chromium.org>
Cc: LKML <linux-kernel@vger.kernel.org>, druth@chromium.org,
	Tim Van Patten <timvp@google.com>,
	Lai Jiangshan <jiangshanlai@gmail.com>
Subject: Re: [PATCH] workqueue: Prevent delayed work UAF kernel panic
Message-ID: <ZmM7FovMyez71TcC@slm.duckdns.org>
References: <20240607104610.1.I6c23f4fdb786f7200f6f1255af57b4e9621bcc66@changeid>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20240607104610.1.I6c23f4fdb786f7200f6f1255af57b4e9621bcc66@changeid>

Hello,

On Fri, Jun 07, 2024 at 10:46:11AM -0600, Tim Van Patten wrote:
...
> Commit 33e3f0a3358b ("workqueue: Add a new flag to spot the potential
> UAF error")added the flag __WQ_DESTROYING to help avoid this. However,
> this solution still allows work to be queued if it's from the same
> workqueue, even if the workqueue has been fully destroyed, which is
> possible with queue_delayed_work().
> 
> 1. queue_delayed_work()
> 2. destroy_workqueue()
> 3. [delayed work timer expires]: delayed_work_timer_fn()
> 
> To prevent kernel panics, check if the pwq and pwq->pool pointers are
> valid before derefencing them, and discard the work if they're not.

Nothing guarantees that they'd stay NULL after wq destruction, right?

> Discarding all work once __WQ_DESTROYING has been set (including from
> the same workqueue) causes breakage, so we must check the pointers
> directly.

There's only so much protection we can offer for buggy code and I'd much
prefer an approach where the overhead is in the destruction path rather than
the queueing path.

Thanks.

-- 
tejun