[PATCH staging 3/3] wilc1000 : Use BIT() macro where possible

[Dan Carpenter]

On Fri, Sep 25, 2015 at 09:14:29AM -0700, Anish Bhatt wrote:
> On Fri, Sep 25, 2015 at 3:50 AM, Dan Carpenter <dan.carpenter at oracle.com> wrote:
> > On Fri, Sep 25, 2015 at 12:45:17AM -0700, Anish Bhatt wrote:
> >> @@ -1672,11 +1673,11 @@ static int wilc_wlan_stop(void)
> >>               }
> >>
> >>       } while (timeout);
> >> -     reg = ((1 << 0) | (1 << 1) | (1 << 2) | (1 << 3) | (1 << 8) |
> >> -            (1 << 9) | (1 << 26) | (1 << 29) | (1 << 30) | (1 << 31));
> >> +     reg = (BIT(0) | BIT(1) | BIT(2) | BIT(3) | BIT(8) | BIT(9) | BIT(26) |
> >> +            BIT(29) | BIT(30) | BIT(31));
> >>
> >>       p->hif_func.hif_write_reg(WILC_GLB_RESET_0, reg);
> >> -     reg = ~(1 << 10);
> >> +     reg = (u32)~BIT(10);
> >
> > Why did you add this cast?  There is no need.
> >
> > regards,
> > dan carpenter
> >
> 
> Since BIT(x) is defined as (1UL << x), ~BIT here results in a unsigned
> long being stored in a u32,
> giving a -Woverflow warning. The cast fixes this.
> 

Is that a default warning?

As static checker warnings go that is a stupid warning.  I have a
similar check in Smatch but I haven't published it because it's a stupid
to warning.

regards,
dan carpenter