2009-12-20 07:10:38

by Marin Mitov

Subject: [PATCH]Silancing a false positive: "may be used uninitialized"

Hello all,

Silancing a false positive:
warning: 'width' may be used uninitialized in this function
drivers/gpu/drm/drm_edid.c

Signed-off-by: Marin Mitov <[email protected]>
==========================================================================
--- a/drivers/gpu/drm/drm_edid.c 2009-12-20 07:43:57.000000000 +0200
+++ b/drivers/gpu/drm/drm_edid.c 2009-12-20 07:45:49.000000000 +0200
@@ -913,7 +913,7 @@
const int rates[] = { 60, 85, 75, 60, 50 };

for (i = 0; i < 4; i++) {
- int width, height;
+ int uninitialized_var(width), height;
cvt = &(timing->data.other_data.data.cvt[i]);

height = (cvt->code[0] + ((cvt->code[1] & 0xf0) << 8) + 1) * 2;

2009-12-20 09:56:48

by Simon Horman

Subject: Re: [PATCH]Silancing a false positive: "may be used uninitialized"

On Sun, Dec 20, 2009 at 09:03:27AM +0200, Marin Mitov wrote:
> Hello all,
>
> Silancing a false positive:
> warning: 'width' may be used uninitialized in this function
> drivers/gpu/drm/drm_edid.c

Is it guaranteed that the switch will always see a value covered by the
four cases it has? If not width would be used without having been initialised.

Perhaps adding

default:
BUG();

would be worthwhile?

>
> Signed-off-by: Marin Mitov <[email protected]>
> ==========================================================================
> --- a/drivers/gpu/drm/drm_edid.c 2009-12-20 07:43:57.000000000 +0200
> +++ b/drivers/gpu/drm/drm_edid.c 2009-12-20 07:45:49.000000000 +0200
> @@ -913,7 +913,7 @@
> const int rates[] = { 60, 85, 75, 60, 50 };
>
> for (i = 0; i < 4; i++) {
> - int width, height;
> + int uninitialized_var(width), height;
> cvt = &(timing->data.other_data.data.cvt[i]);
>
> height = (cvt->code[0] + ((cvt->code[1] & 0xf0) << 8) + 1) * 2;
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/

2009-12-20 10:44:53

by Marin Mitov

Subject: Re: [PATCH]Silancing a false positive: "may be used uninitialized"

On Sunday 20 December 2009 11:56:44 am Simon Horman wrote:
> On Sun, Dec 20, 2009 at 09:03:27AM +0200, Marin Mitov wrote:
> > Hello all,
> >
> > Silancing a false positive:
> > warning: 'width' may be used uninitialized in this function
> > drivers/gpu/drm/drm_edid.c
>
> Is it guaranteed that the switch will always see a value covered by the
> four cases it has?

Yes it is. The value is masked (& 0xc0) so all the possible values are
(0x00, 0x40, 0x80, 0xc0) covered by the switch. That's why it is false positive.

> If not width would be used without having been initialised.
>
> Perhaps adding
>
> default:
> BUG();
>
> would be worthwhile?
>
> >
> > Signed-off-by: Marin Mitov <[email protected]>
> > ==========================================================================
> > --- a/drivers/gpu/drm/drm_edid.c 2009-12-20 07:43:57.000000000 +0200
> > +++ b/drivers/gpu/drm/drm_edid.c 2009-12-20 07:45:49.000000000 +0200
> > @@ -913,7 +913,7 @@
> > const int rates[] = { 60, 85, 75, 60, 50 };
> >
> > for (i = 0; i < 4; i++) {
> > - int width, height;
> > + int uninitialized_var(width), height;
> > cvt = &(timing->data.other_data.data.cvt[i]);
> >
> > height = (cvt->code[0] + ((cvt->code[1] & 0xf0) << 8) + 1) * 2;
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> > the body of a message to [email protected]
> > More majordomo info at http://vger.kernel.org/majordomo-info.html
> > Please read the FAQ at http://www.tux.org/lkml/
>

2009-12-20 21:14:39

by Simon Horman

Subject: Re: [PATCH]Silancing a false positive: "may be used uninitialized"

On Sun, Dec 20, 2009 at 12:43:19PM +0200, Marin Mitov wrote:
> On Sunday 20 December 2009 11:56:44 am Simon Horman wrote:
> > On Sun, Dec 20, 2009 at 09:03:27AM +0200, Marin Mitov wrote:
> > > Hello all,
> > >
> > > Silancing a false positive:
> > > warning: 'width' may be used uninitialized in this function
> > > drivers/gpu/drm/drm_edid.c
> >
> > Is it guaranteed that the switch will always see a value covered by the
> > four cases it has?
>
> Yes it is. The value is masked (& 0xc0) so all the possible values are
> (0x00, 0x40, 0x80, 0xc0) covered by the switch. That's why it is false positive.

Of course, silly me.