Re: [RFC][PATCH] bitfields API

[Alexey Dobriyan <adobriyan@gmail.com>]

On Thu, Aug 28, 2008 at 08:46:43PM +0200, Vegard Nossum wrote:
> On Thu, Aug 28, 2008 at 8:40 PM, Alexey Dobriyan <adobriyan@gmail.com> wrote:
> > On Thu, Aug 28, 2008 at 08:32:23PM +0200, Vegard Nossum wrote:
> >> How do you feel about this patch? It's all about making kmemcheck more
> >> useful... and not much else. Does it have any chance of entering the
> >> kernel along with kmemcheck (when/if that happens)?
> >
> > DEFINE_BITFIELD is horrible.
> >
> >> @@ -285,11 +286,12 @@ struct sk_buff {
> >>               };
> >>       };
> >>       __u32                   priority;
> >> -     __u8                    local_df:1,
> >> +     DEFINE_BITFIELD(__u8,   flags1,
> >> +                             local_df:1,
> >>                               cloned:1,
> >>                               ip_summed:2,
> >>                               nohdr:1,
> >> -                             nfctinfo:3;
> >> +                             nfctinfo:3);
> >>       __u8                    pkt_type:3,
> >>                               fclone:2,
> >>                               ipvs_property:1,
> 
> Ok, that's constructive :-P
> 
> Can we skip the type and always assume that it should be __u8/uint8_t?

Of course, no.

> I read somewhere that bitfields should anyway always be 1 byte wide if
> the bitfield should be "portable".

It should be signed int or unsigned int for maximum portability.

> Would it help (to make this less
> horrible) to omit the type declaration and have just the bitfield
> members as arguments to the macro?

Or you can parse instruction stream a bit more.