Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
References: <CAG48ez14Q0-F8LqsvcNbyR2o6gPW8SHXsm4u5jmD9MpsteM2Tw@mail.gmail.com>
 <20191008130159.10161-1-christian.brauner@ubuntu.com> <20191008180516.GB143258@google.com>
 <20191009104011.rzfdvq7otkkj533m@wittgenstein>
In-Reply-To: <20191009104011.rzfdvq7otkkj533m@wittgenstein>
From:   Todd Kjos <tkjos@google.com>
Date:   Wed, 9 Oct 2019 08:56:28 -0700
Message-ID: <CAHRSSExAL3fQMP9x9p34qF8dnFKLSp7EzhZ7Y5y-qbzExCRo1w@mail.gmail.com>
Subject: Re: [PATCH] binder: prevent UAF read in print_binder_transaction_log_entry()
To:     Christian Brauner <christian.brauner@ubuntu.com>
Cc:     Joel Fernandes <joel@joelfernandes.org>,
        Todd Kjos <tkjos@android.com>, Jann Horn <jannh@google.com>,
        =?UTF-8?B?QXJ2ZSBIasO4bm5ldsOlZw==?= <arve@android.com>,
        Christian Brauner <christian@brauner.io>,
        "open list:ANDROID DRIVERS" <devel@driverdev.osuosl.org>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        LKML <linux-kernel@vger.kernel.org>,
        Martijn Coenen <maco@android.com>,
        Hridya Valsaraju <hridya@google.com>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Wed, Oct 9, 2019 at 3:40 AM Christian Brauner
<christian.brauner@ubuntu.com> wrote:
>
> On Tue, Oct 08, 2019 at 02:05:16PM -0400, Joel Fernandes wrote:
> > On Tue, Oct 08, 2019 at 03:01:59PM +0200, Christian Brauner wrote:

[...]

> >
> > One more thought, this can be made dependent on CONFIG_BINDERFS since regular
> > binder devices cannot be unregistered AFAICS and as Jann said, the problem is
> > BINDERFS specific. That way we avoid the memcpy for _every_ transaction.
> > These can be thundering when Android starts up.
>
> Unless Todd sees this as a real performance problem I'm weary to
> introduce additional checking and record a pointer for non-binderfs and
> a memcpy() for binderfs devices. :)
>

I don't see this as a real problem. In practice, memcpy will be moving
< 10 bytes. Also, by the time this code is in an android device,
CONFIG_BINDERFS will always be enabled since this is how we are
removing binder's use of debugfs. So a micro-optimization of the
!BINDERFS case will not be meaningful.

[...]