Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <20200428110253.1.I926f6741079cafb04ecb592130aef75b24ad31ae@changeid>
 <c57b0892-d68d-f0f2-3cc1-b8549983227e@collabora.com> <CAONX=-dLVBpz+p1si4cWGHEmQ_toO8kW=fCPcdUuX2KLc-o=2A@mail.gmail.com>
 <CAONX=-dDWjMQDGncSgXNf86Bt3C5OUtFqh=Pe2HToGPSv5qa4A@mail.gmail.com>
In-Reply-To: <CAONX=-dDWjMQDGncSgXNf86Bt3C5OUtFqh=Pe2HToGPSv5qa4A@mail.gmail.com>
From:   Prashant Malani <pmalani@chromium.org>
Date:   Wed, 29 Apr 2020 17:43:24 -0700
Message-ID: <CACeCKacoEjQm0_VfZGYNEZVhKubPyqspNZupQk+7piJOsVzH+Q@mail.gmail.com>
Subject: Re: [PATCH] platform/chrome: cros_ec_typec: Handle NULL EC pointer
 during probe.
To:     Daniil Lunev <dlunev@chromium.org>
Cc:     Enric Balletbo i Serra <enric.balletbo@collabora.com>,
        LKML <linux-kernel@vger.kernel.org>,
        Benson Leung <bleung@chromium.org>,
        Guenter Roeck <groeck@chromium.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Wed, Apr 29, 2020 at 5:38 PM Daniil Lunev <dlunev@chromium.org> wrote:
>
> [to make it appear on the mailing list as I didn't realize I was in
> hypertext sending mode]
>
> On Thu, Apr 30, 2020 at 10:11 AM Daniil Lunev <dlunev@chromium.org> wrote=
:
> >
> > Hi Enric.
> > I encountered the issue on a Hatch device when trying running 5.4 kerne=
l on that. After talking to Prashant it seems that any device with coreboot=
 built before a certain point (a particular fix for device hierarchy in ACP=
I tables of Chrome devices which happened in mid-April) will not be able to=
 correctly initialize the driver and will get a kernel panic trying to do s=
o.

A clarifying detail here: This should not be seen in any current
*production* device. No prod device firmware will carry the erroneous
ACPI device entry.

> > Thanks,
> > Daniil
> >
> > On Thu, Apr 30, 2020 at 7:58 AM Enric Balletbo i Serra <enric.balletbo@=
collabora.com> wrote:
> >>
> >> Hi Daniil,
> >>
> >> Thank you for the patch.
> >>
> >> On 28/4/20 3:02, Daniil Lunev wrote:
> >> > Missing EC in device hierarchy causes NULL pointer to be returned to=
 the
> >> > probe function which leads to NULL pointer dereference when trying t=
o
> >> > send a command to the EC. This can be the case if the device is miss=
ing
> >> > or incorrectly configured in the firmware blob. Even if the situatio=
n
> >>
> >> There is any production device with a buggy firmware outside? Or this =
is just
> >> for defensive programming while developing the firmware? Which device =
is
> >> affected for this issue?
> >>
> >> Thanks,
> >>  Enric
> >>
> >> > occures, the driver shall not cause a kernel panic as the condition =
is
> >> > not critical for the system functions.
> >> >
> >> > Signed-off-by: Daniil Lunev <dlunev@chromium.org>
> >> > ---
> >> >
> >> >  drivers/platform/chrome/cros_ec_typec.c | 5 +++++
> >> >  1 file changed, 5 insertions(+)
> >> >
> >> > diff --git a/drivers/platform/chrome/cros_ec_typec.c b/drivers/platf=
orm/chrome/cros_ec_typec.c
> >> > index 874269c07073..30d99c930445 100644
> >> > --- a/drivers/platform/chrome/cros_ec_typec.c
> >> > +++ b/drivers/platform/chrome/cros_ec_typec.c
> >> > @@ -301,6 +301,11 @@ static int cros_typec_probe(struct platform_dev=
ice *pdev)
> >> >
> >> >       typec->dev =3D dev;
> >> >       typec->ec =3D dev_get_drvdata(pdev->dev.parent);
> >> > +     if (!typec->ec) {
> >> > +             dev_err(dev, "Failed to get Cros EC data\n");
> >> > +             return -EINVAL;
> >> > +     }
> >> > +
> >> >       platform_set_drvdata(pdev, typec);
> >> >
> >> >       ret =3D cros_typec_get_cmd_version(typec);
> >> >