On Thu, 17 Nov 2016 14:14:21 +0800
Peter Pan <peterpans...@gmail.com> wrote:
> Hi Boris,
>
> On Sun, Oct 16, 2016 at 10:35 PM, Boris Brezillon
> <boris.brezil...@free-electrons.com> wrote:
> > Now that raw NAND header has been moved to rawnand.h, we can add a new
> > nand.h file and define the common nand stuff in there.
> >
> > Signed-off-by: Boris Brezillon <boris.brezil...@free-electrons.com>
> > ---
> > include/linux/mtd/nand.h | 480
> > +++++++++++++++++++++++++++++++++++++++++++++++
> > 1 file changed, 480 insertions(+)
> > create mode 100644 include/linux/mtd/nand.h
> >
> > diff --git a/include/linux/mtd/nand.h b/include/linux/mtd/nand.h
> > new file mode 100644
> > index 000000000000..b9b5d3c1bdf9
> > --- /dev/null
> > +++ b/include/linux/mtd/nand.h
> [...]
> > +
> > +/**
> > + * nand_offs_to_die - Convert an absolute offset to a die number
> > + * @nand: NAND device
> > + * @offs: absolute offset
> > + *
> > + * Returns the die number containing @offs.
> > + */
> > +static inline int nand_offs_to_die(struct nand_device *nand, loff_t offs)
> > +{
> > + u64 die = offs;
> > +
> > + do_div(die, nand->memorg.diesize);
> > +
> > + return die;
>
> Is it proper to use do_div here?. If I'm right, the divisor of do_div is
> 32bit.
> It may cause division 0 exception when die size is larger than 4GB.
> div64_u64 may be better.
Nice catch! Indeed, div64_u64() should be used here. I'll fix that.
Thanks,
Boris
