[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <YsVYQAQ8ylvMQtR2@google.com>
Date:   Wed, 6 Jul 2022 10:39:12 +0100
From:   Lee Jones <lee.jones@...aro.org>
To:     Eric Biggers <ebiggers@...nel.org>
Cc:     Haimin Zhang <tcs.kernel@...il.com>,
        Alexander Viro <viro@...iv.linux.org.uk>,
        David Howells <dhowells@...hat.com>,
        "Fabio M. De Francesco" <fmdefrancesco@...il.com>,
        linux-kernel@...r.kernel.org, linux-fsdevel@...r.kernel.org,
        Haimin Zhang <tcs_kernel@...cent.com>,
        TCS Robot <tcs_robot@...cent.com>
Subject: Re: [PATCH v2] fs/pipe: Deinitialize the watch_queue when pipe is
 freed
On Mon, 09 May 2022, Eric Biggers wrote:
> On Mon, May 09, 2022 at 09:17:26PM +0800, Haimin Zhang wrote:
> > From: Haimin Zhang <tcs_kernel@...cent.com>
> > 
> > Add a new function call to deinitialize the watch_queue of a freed pipe.
> > When a pipe node is freed, it doesn't make pipe->watch_queue->pipe null.
> > Later when function post_one_notification is called, it will use this
> > field, but it has been freed and watch_queue->pipe is a dangling pointer.
> > It makes a uaf issue.
> > Check wqueu->defunct before pipe check since pipe becomes invalid once all
> > watch queues were cleared.
> > 
> > Reported-by: TCS Robot <tcs_robot@...cent.com>
> > Signed-off-by: Haimin Zhang <tcs_kernel@...cent.com>
> 
> Is this fixing something?  If so it should have a "Fixes" tag.
It sure is.
Haimin, are you planning a v3?
-- 
Lee Jones [李琼斯]
Principal Technical Lead - Developer Services
Linaro.org │ Open source software for Arm SoCs
Follow Linaro: Facebook | Twitter | Blog
Powered by blists - more mailing lists
 
