Re: [PATCH v2 3/4] xfsprogs: mkfs: fix unintentional integer overflow

[Dave Chinner <david@xxxxxxxxxxxxx>]

On Sun, Dec 20, 2015 at 07:36:49AM -0800, Christoph Hellwig wrote:
> > +++ b/mkfs/xfs_mkfs.c
> > @@ -2022,7 +2022,7 @@ _("warning: sparse inodes not supported without CRC 
> > support, disabled.\n"));
> >              * and the underlying volume is striped, then set rtextblocks
> >              * to the stripe width.
> >              */
> > -           int             rswidth;
> > +           __uint64_t      rswidth;
> >             __uint64_t      rtextbytes;
> 
> This looks odd.  We initiallize assigned ft.rtswidth (which is an int)
> or 0 to it.  I think you want a separate variable for the result of
> the  DTOBT(rswidth) statement to make this clear.

I don't see any point in doing that. It really doesn't matter that
the variable it is initialised from is only a 32 bit int, and having
yet another variable in mkfs is not going to make the code easier to
understand...

Cheers,

Dave.
-- 
Dave Chinner
david@xxxxxxxxxxxxx