[PATCH] mfd: mt6370: add bounds checking to regmap_read/write functions

Lee Jones lee at kernel.org
Wed Sep 7 23:57:16 PDT 2022


On Mon, 22 Aug 2022, Dan Carpenter wrote:

> On Fri, Aug 19, 2022 at 09:27:13AM +0300, Andy Shevchenko wrote:
> > On Fri, Aug 19, 2022 at 8:25 AM Dan Carpenter <dan.carpenter at oracle.com> wrote:
> > >
> > > It looks like there are a potential out of bounds accesses in the
> > > read/write() functions.  Also can "len" be negative?  Let's check for
> > > that too.
> > 
> > ...
> > 
> > > Fixes: ab9905c5e38e ("mfd: mt6370: Add MediaTek MT6370 support")
> > 
> > > From static analysis.  This code is obviously harmless however it may
> > > not be required.  The regmap range checking is slightly complicated and
> > > I haven't remembered where all it's done.
> > 
> > Exactement! I do not think this Fixes anything, I believe you are
> > adding a dead code. So, can you do deeper analysis?
> 
> I spent a long time looking at this code before I sent it and I've
> spent a long time looking at it today.
> 
> Smatch said that these values come from the user, but now it seems
> less clear to me and I have rebuilt the DB so I don't have the same
> information I was looking at earlier.
> 
> So I can't see if these come from the user but neither can I find any
> bounds checking.

What's the consensus please?

-- 
Lee Jones [李琼斯]



More information about the Linux-mediatek mailing list