linux-kernel - Re: [PATCH] fbdev: atyfb: fix array overflow

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Thu, 23 Jun 2016 20:26:48 +0300
From:	Ville Syrjälä <syrjala@....fi>
To:	Arnd Bergmann <arnd@...db.de>
Cc:	Tomi Valkeinen <tomi.valkeinen@...com>,
	Jean-Christophe Plagniol-Villard <plagnioj@...osoft.com>,
	Ingo Molnar <mingo@...nel.org>,
	"Luis R. Rodriguez" <mcgrof@...e.com>,
	Borislav Petkov <bp@...e.de>, linux-fbdev@...r.kernel.org,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH] fbdev: atyfb: fix array overflow

On Thu, Jun 23, 2016 at 11:06:11AM +0200, Arnd Bergmann wrote:
> On Thursday, June 23, 2016 3:28:25 AM CEST Ville Syrjälä wrote:
> > On Wed, Jun 22, 2016 at 02:37:11PM +0200, Arnd Bergmann wrote:
> > > When building with CONFIG_UBSAN_SANITIZE_ALL on ARM, I get this
> > > gcc warning for atyfb:
> > > 
> > > drivers/video/fbdev/aty/atyfb_base.c: In function 'aty_bl_update_status':
> > > drivers/video/fbdev/aty/atyfb_base.c:167:33: warning: array subscript is above array bounds [-Warray-bounds]
> > > drivers/video/fbdev/aty/atyfb_base.c:152:26: warning: array subscript is above array bounds [-Warray-bounds]
> > > 
> > > Apparently the warning is correct and there is indeed an overflow,
> > 
> > Nope. All the LCD register indexes on the Rage LT (the only relevant
> > chip for this code path) should stay below the table size. At least
> > I can't see any place where we'd walk past the end.
> 
> I don't understand what you mean: the warning is about LCD_MISC_CNTL,
> which is defined as 0x14, while the array size is 9 and that is smaller.
> 
> Is there something more subtle going on than what gcc sees?

The LCD_MISC_CNTL access is in the backlight code, and thanks to the
following piece of code

if (M64_HAS(MOBIL_BUS) && ...) {
	aty_bl_init(...);
}

we register the backlight only on Rage Mobility. Rage LT is not a Rage
Mobility, so everything is fine.

-- 
Ville Syrjälä
syrjala@....fi
http://www.sci.fi/~syrjala/