Re: [PATCH] drm/vmwgfx_kms: Fix potential NULL pointer dereference

["Ville Syrjälä" <ville.syrjala@linux.intel.com>]

On Mon, Dec 04, 2017 at 03:54:18PM -0600, Gustavo A. R. Silva wrote:
> crtc_state is being null checked in a previous code block, which implies
> that such pointer might be null.
> 
> crtc_state is dereferenced in drm_atomic_helper_check_plane_state, hence
> there is a potential null pointer dereference.

This is a false positive. drm_atomic_helper_check_plane_state() will not
dereference crtc_state when plane_state->crtc is NULL.

> 
> Fix this by warning-on and returning -EINVAL in case crtc_state is null.
> 
> Addresses-Coverity-ID: 1462412 ("Dereference after null check")
> Fixes: a01cb8ba3f62 ("drm: Move drm_plane_helper_check_state() into drm_atomic_helper.c")
> Signed-off-by: Gustavo A. R. Silva <garsilva@embeddedor.com>
> ---
>  drivers/gpu/drm/vmwgfx/vmwgfx_kms.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
> index a2a93d7..72c3b290 100644
> --- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
> +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
> @@ -454,6 +454,9 @@ int vmw_du_primary_plane_atomic_check(struct drm_plane *plane,
>  		clip.y2 = crtc_state->adjusted_mode.vdisplay;
>  	}
>  
> +	if (WARN_ON(!crtc_state))
> +		return -EINVAL;

This would in fact break the driver because it would flag an error
whenever the plane is disabled.

> +
>  	ret = drm_atomic_helper_check_plane_state(state, crtc_state, &clip,
>  						  DRM_PLANE_HELPER_NO_SCALING,
>  						  DRM_PLANE_HELPER_NO_SCALING,
> -- 
> 2.7.4
> 
> _______________________________________________
> dri-devel mailing list
> dri-devel@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/dri-devel

-- 
Ville Syrjälä
Intel OTC
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel