MIME-Version: 1.0
References: <20230312145305.1908607-1-zyytlz.wz@163.com> <CANDhNCr=hdhKS4c+U=+W1ONHDWv6BrwL5TovGjs0G2G+Reqc9g@mail.gmail.com>
In-Reply-To: <CANDhNCr=hdhKS4c+U=+W1ONHDWv6BrwL5TovGjs0G2G+Reqc9g@mail.gmail.com>
From:   Zheng Hacker <hackerzheng666@gmail.com>
Date:   Tue, 14 Mar 2023 09:01:46 +0800
Message-ID: <CAJedcCyJnV+KnFF5h+2-0W1R4uaUxUxXFUH3Q9HGYh-5F5LmBQ@mail.gmail.com>
Subject: Re: [PATCH] misc: hisi_hikey_usb: Fix use after free bug in
 hisi_hikey_usb_remove due to race condition
To:     John Stultz <jstultz@google.com>
Cc:     Zheng Wang <zyytlz.wz@163.com>,
        Yongqin Liu <yongqin.liu@linaro.org>,
        Sumit Semwal <sumit.semwal@linaro.org>, arnd@arndb.de,
        gregkh@linuxfoundation.org, linux-kernel@vger.kernel.org,
        1395428693sheep@gmail.com, alex000young@gmail.com,
        Mauro Carvalho Chehab <mchehab@kernel.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

John Stultz <jstultz@google.com> =E4=BA=8E2023=E5=B9=B43=E6=9C=8814=E6=97=
=A5=E5=91=A8=E4=BA=8C 03:57=E5=86=99=E9=81=93=EF=BC=9A
>
> On Sun, Mar 12, 2023 at 7:53=E2=80=AFAM Zheng Wang <zyytlz.wz@163.com> wr=
ote:
> >
> > In hisi_hikey_usb_probe, it called hisi_hikey_usb_of_role_switch
> > and bound &hisi_hikey_usb->work with relay_set_role_switch.
> > When it calls hub_usb_role_switch_set, it will finally call
> > schedule_work to start the work.
> >
> > When we call hisi_hikey_usb_remove to remove the driver, there
> > may be a sequence as follows:
> >
> > Fix it by finishing the work before cleanup in hisi_hikey_usb_remove.
> >
> > CPU0                  CPU1
> >
> >                     |relay_set_role_switch
> > hisi_hikey_usb_remove|
> >   usb_role_switch_put|
> >     usb_role_switch_release  |
> >      kfree(sw)     |
> >                     | usb_role_switch_set_role
> >                     |   //use
> >
> > Fixes: 7a6ff4c4cbc3 ("misc: hisi_hikey_usb: Driver to support onboard U=
SB gpio hub on Hikey960")
> > Signed-off-by: Zheng Wang <zyytlz.wz@163.com>
> > ---
> >  drivers/misc/hisi_hikey_usb.c | 1 +
> >  1 file changed, 1 insertion(+)
> >
> > diff --git a/drivers/misc/hisi_hikey_usb.c b/drivers/misc/hisi_hikey_us=
b.c
> > index 2165ec35a343..26fc895c4418 100644
> > --- a/drivers/misc/hisi_hikey_usb.c
> > +++ b/drivers/misc/hisi_hikey_usb.c
> > @@ -242,6 +242,7 @@ static int hisi_hikey_usb_probe(struct platform_dev=
ice *pdev)
> >  static int  hisi_hikey_usb_remove(struct platform_device *pdev)
> >  {
> >         struct hisi_hikey_usb *hisi_hikey_usb =3D platform_get_drvdata(=
pdev);
> > +       cancel_work_sync(&hisi_hikey_usb->work);
> >
> >         if (hisi_hikey_usb->hub_role_sw) {
> >                 usb_role_switch_unregister(hisi_hikey_usb->hub_role_sw)=
;
>
> Looks sane to me.
> Pulling in Sumit and YongQin as they have hardware and can test with it.
>
Hi John,

Thanks for your reply. Thank Sumit and YongQin for being willing to
test the solution with their hardware.

Best regards,
Zheng