MIME-Version: 1.0
References: <YaZP2HzTQw1QJxOK@intel.com> <20211201152904.182293-1-zhou1615@umn.edu>
In-Reply-To: <20211201152904.182293-1-zhou1615@umn.edu>
From:   Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Date:   Wed, 1 Dec 2021 16:57:24 +0100
Message-ID: <CAMeQTsY0RHJsJzNW-77gK_=WCetvhrUVbF4KA+Qu3uaV1CCR_w@mail.gmail.com>
Subject: Re: [PATCH v2] drm/gma500/cdv: Fix a wild pointer dereference in cdv_intel_dp_get_modes()
To:     Zhou Qingyang <zhou1615@umn.edu>
Cc:     kjlu@umn.edu, David Airlie <airlied@linux.ie>,
        Daniel Vetter <daniel@ffwll.ch>,
        Dave Airlie <airlied@redhat.com>,
        Alan Cox <alan@linux.intel.com>,
        Zhao Yakui <yakui.zhao@intel.com>,
        dri-devel <dri-devel@lists.freedesktop.org>,
        linux-kernel <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Wed, Dec 1, 2021 at 4:29 PM Zhou Qingyang <zhou1615@umn.edu> wrote:
>
> In cdv_intel_dp_get_modes(), the third return value of
> drm_mode_duplicate() is assigned to mode and used in
> drm_mode_probed_add(). drm_mode_probed_add() passes mode->head to
> list_add_tail(). list_add_tail() will further call __list_add() and
> there is a dereference of mode->head in __list_add(), which could lead
> to a wild pointer dereference on failure of drm_mode_duplicate().
>
> Fix this bug by adding a check of mode.
>
> This bug was found by a static analyzer. The analysis employs
> differential checking to identify inconsistent security operations
> (e.g., checks or kfrees) between two code paths and confirms that the
> inconsistent operations are not recovered in the current function or
> the callers, so they constitute bugs.
>
> Note that, as a bug found by static analysis, it can be a false
> positive or hard to trigger. Multiple researchers have cross-reviewed
> the bug.
>

Is it really necessary to explain what the static analyzer does and
that it can be faulty in every single patch?
"This bug was found by a static analyzer" is enough for me.

> Builds with CONFIG_DRM_GMA500=m show no new warnings,
> and our static analyzer no longer warns about this code.

I assume all patches to be at least compile tested before submitted,
so if you didn't actually run this code on hardware it's better to
replace the above with:
"Only compile tested".

-Patrik

>
> Fixes: d112a8163f83 ("gma500/cdv: Add eDP support")
> Signed-off-by: Zhou Qingyang <zhou1615@umn.edu>
> ---
> Changes in V2:
>   -  Instead of returning -ENOMEM, this patch returns 0
>   -  Use DRM_DEBUG_KMS to report the failure of drm_mode_duplicate()
>
>  drivers/gpu/drm/gma500/cdv_intel_dp.c | 5 +++++
>  1 file changed, 5 insertions(+)
>
> diff --git a/drivers/gpu/drm/gma500/cdv_intel_dp.c b/drivers/gpu/drm/gma500/cdv_intel_dp.c
> index ba6ad1466374..bf47db488b7b 100644
> --- a/drivers/gpu/drm/gma500/cdv_intel_dp.c
> +++ b/drivers/gpu/drm/gma500/cdv_intel_dp.c
> @@ -1773,6 +1773,11 @@ static int cdv_intel_dp_get_modes(struct drm_connector *connector)
>                 if (intel_dp->panel_fixed_mode != NULL) {
>                         struct drm_display_mode *mode;
>                         mode = drm_mode_duplicate(dev, intel_dp->panel_fixed_mode);
> +                       if (!mode) {
> +                               DRM_DEBUG_KMS("Failure in drm_mode_duplicate()\n");
> +                               return 0;
> +                       }
> +
>                         drm_mode_probed_add(connector, mode);
>                         return 1;
>                 }
> --
> 2.25.1
>