On Thu, 2009-05-28 at 12:48 +0400, Dmitry Eremin-Solenikov wrote:
> On Thu, May 28, 2009 at 04:08:51AM +0100, Ben Hutchings wrote:
> > On Tue, 2009-05-26 at 15:23 +0400, Dmitry Eremin-Solenikov wrote:
> > [...]
> > > diff --git a/include/net/ieee802154/af_ieee802154.h b/include/net/ieee802154/af_ieee802154.h
> > > new file mode 100644
> > > index 0000000..6eb7f51
> > > --- /dev/null
> > > +++ b/include/net/ieee802154/af_ieee802154.h
> > [...]
> > > +#ifdef __KERNEL__
> > > +#include <linux/skbuff.h>
> > > +#include <linux/netdevice.h>
> >
> > The struct declarations would be sufficient.
>
> Do you mean just:
> struct sk_buff;
> struct net_device;
Right.
[...]
> >
> > [...]
> > > +#define IEEE802154_FETCH_U64(skb, var) \
> > > + do { \
> > > + if (skb->len < IEEE802154_ADDR_LEN) \
> >
> > I see what you did there...
>
> ???
[...]
This macro works with u64 values, not specifically addresses, which
happen to be 8 bytes long too. I infer that someone search-and-replaced
instances of the literal 8 with IEEE802154_ADDR_LEN and wrongly replaced
this instance.
Ben.
--
Ben Hutchings
Logic doesn't apply to the real world. - Marvin Minsky
Attachment:
signature.asc
Description: This is a digitally signed message part
