2010-12-25 13:13:50

by Jarek Poplawski

Subject: [2.6.37-rc7] drivers/gpu/drm/drm_edid.c:1303: warning: array subscript is above array bounds

Hi,
There is this compile warning in Linus' tree (gcc version 4.3.3):

drivers/gpu/drm/drm_edid.c: In function 'drm_detect_hdmi_monitor':
drivers/gpu/drm/drm_edid.c:1303: warning: array subscript is above array bounds
drivers/gpu/drm/drm_edid.c: In function 'drm_detect_monitor_audio':
drivers/gpu/drm/drm_edid.c:1303: warning: array subscript is above array bounds

Cheers,
Jarek P.

2010-12-25 13:40:46

by Chris Wilson

Subject: Re: [2.6.37-rc7] drivers/gpu/drm/drm_edid.c:1303: warning: array subscript is above array bounds

On Sat, 25 Dec 2010 14:13:43 +0100, Jarek Poplawski <[email protected]> wrote:
> Hi,
> There is this compile warning in Linus' tree (gcc version 4.3.3):
>
> drivers/gpu/drm/drm_edid.c: In function 'drm_detect_hdmi_monitor':
> drivers/gpu/drm/drm_edid.c:1303: warning: array subscript is above array bounds
> drivers/gpu/drm/drm_edid.c: In function 'drm_detect_monitor_audio':
> drivers/gpu/drm/drm_edid.c:1303: warning: array subscript is above array bounds

What array? It's declared as "u8 *edid_ext;". At first glance looks like a
compiler bug, applying subscript testing to an object with unknown bounds.
-Chris

--
Chris Wilson, Intel Open Source Technology Centre

2010-12-25 19:20:18

by Jarek Poplawski

Subject: Re: [2.6.37-rc7] drivers/gpu/drm/drm_edid.c:1303: warning: array subscript is above array bounds

On Sat, Dec 25, 2010 at 01:38:56PM +0000, Chris Wilson wrote:
> On Sat, 25 Dec 2010 14:13:43 +0100, Jarek Poplawski <[email protected]> wrote:
> > Hi,
> > There is this compile warning in Linus' tree (gcc version 4.3.3):
> >
> > drivers/gpu/drm/drm_edid.c: In function 'drm_detect_hdmi_monitor':
> > drivers/gpu/drm/drm_edid.c:1303: warning: array subscript is above array bounds
> > drivers/gpu/drm/drm_edid.c: In function 'drm_detect_monitor_audio':
> > drivers/gpu/drm/drm_edid.c:1303: warning: array subscript is above array bounds
>
> What array? It's declared as "u8 *edid_ext;". At first glance looks like a
> compiler bug, applying subscript testing to an object with unknown bounds.

Maybe a bug but then also something very unique in this place wrt the
rest of the kernel (at least in my config). Btw, a patch below removes
this warning.

Jarek P.

--- drm_edid.c- 2010-12-25 19:24:55.000000000 +0100
+++ drm_edid.c 2010-12-25 19:51:45.000000000 +0100
@@ -1290,7 +1290,7 @@ add_detailed_modes(struct drm_connector
*/
static u8 *drm_find_cea_extension(struct edid *edid)
{
- u8 *edid_ext = NULL;
+ char *edid_ext = NULL;
int i;

/* No EDID or EDID extensions */
@@ -1299,7 +1299,7 @@ static u8 *drm_find_cea_extension(struct

/* Find CEA extension */
for (i = 0; i < edid->extensions; i++) {
- edid_ext = (u8 *)edid + EDID_LENGTH * (i + 1);
+ edid_ext = (char *)edid + EDID_LENGTH * (i + 1);
if (edid_ext[0] == CEA_EXT)
break;
}