On Mon, Dec 10, 2012 at 10:21:52PM +0000, Przemo Firszt wrote:
> Dnia 2012-12-10, pon o godzinie 12:41 +0300, Dan Carpenter pisze:
> > On Sun, Dec 09, 2012 at 10:15:09AM +0000, Przemo Firszt wrote:
> > > A clean up change suggested by checkpatch.pl
> > >
> > > Signed-off-by: Przemo Firszt <przemo@xxxxxxxxx>
> > > ---
> > > drivers/staging/rtl8712/rtl871x_ioctl_linux.c | 6 ++----
> > > 1 file changed, 2 insertions(+), 4 deletions(-)
> > >
> > > diff --git a/drivers/staging/rtl8712/rtl871x_ioctl_linux.c b/drivers/staging/rtl8712/rtl871x_ioctl_linux.c
> > > index cdb51d7..b131b61 100644
> > > --- a/drivers/staging/rtl8712/rtl871x_ioctl_linux.c
> > > +++ b/drivers/staging/rtl8712/rtl871x_ioctl_linux.c
> > > @@ -188,8 +188,7 @@ static inline char *translate_scan(struct _adapter *padapter,
> > > /* Add the ESSID */
> > > iwe.cmd = SIOCGIWESSID;
> > > iwe.u.data.flags = 1;
> > > - iwe.u.data.length = (u16)min((u16)pnetwork->network.Ssid.SsidLength,
> > > - (u16)32);
> > > + iwe.u.data.length = min_t(u16, pnetwork->network.Ssid.SsidLength, 32);
> >
> > pnetwork->network.Ssid.SsidLength is a u32 so it would be better to
> > not truncate the upper bits away. It's not going to cause a
> > problem, but its slightly messy.
> >
> > This is a common problem where people take the type of
> > iwe.u.data.length and cast to that instead of considering the types
> > for the data they are comparing.
> >
> Dan,
> Thanks for the comment!
>
> iew.u.data.length is __u16
>
> min_t is defined as:
> #define min_t(type, x, y) ({ \
> type __min1 = (x); \
> type __min2 = (y); \
> __min1 < __min2 ? __min1: __min2; })
>
> If I understand you correctly I should use this:
>
> iwe.u.data.length = min_t(u32, pnetwork->network.Ssid.SsidLength, 32);
>
> the result of min_t is u32, but iwe.u.data.lenght is __u16
>
At the end iwe.u.data.length is going to be a number between 0 and
32. That can easily fit in 16 bits no problem.
The difference is that, imagine ->network.Ssid.SsidLength is larger
than u16 like 0x10001. The cast to u16 changes it to 0x0001 which
is less than 32 so we would say the minimum is 1 when actually we
want to say 32 is the min.
regards,
dan carpenter
_______________________________________________
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxx
http://driverdev.linuxdriverproject.org/mailman/listinfo/devel
