On Wed, Aug 09, 2017 at 03:38:33PM +0100, Daniel Stone wrote:
> On 9 August 2017 at 15:36, Sean Paul <seanp...@chromium.org> wrote:
> > On Wed, Aug 09, 2017 at 02:19:06PM +0300, Dan Carpenter wrote:
> >> "plane->format_count" can go up to 64. (It's capped in
> >> drm_universal_plane_init().) So we should be using ULL type instead of
> >> int here to prevent shift wrapping.
> >> Fixes: db1689aa61bd ("drm: Create a format/modifier blob")
> >> Signed-off-by: Dan Carpenter <dan.carpen...@oracle.com>
> > Thank you for the fix, Dan.
> > I've applied it to drm-misc-next.
> Yes, thanks Dan!
> Out of interest, how was this found? With sparse?
These are Smatch checks that I haven't totally cleaned up enough to
publish yet. I have a couple versions of this check. This one is doing
cross function analysis so it knows that ->format_count can go up to 64
dri-devel mailing list