On Wed, Jun 27, 2012 at 02:47:47PM +0900, Isaku Yamahata wrote:
> On Tue, Jun 26, 2012 at 09:39:23PM -0700, Ben Pfaff wrote:
> > On Wed, Jun 27, 2012 at 01:27:00PM +0900, Isaku Yamahata wrote:
> > > Signed-off-by: Isaku Yamahata <yamah...@valinux.co.jp>
> > > ---
> > > lib/meta-flow.h | 1 +
> > > 1 files changed, 1 insertions(+), 0 deletions(-)
> > >
> > > diff --git a/lib/meta-flow.h b/lib/meta-flow.h
> > > index ffde5cc..b4a15e5 100644
> > > --- a/lib/meta-flow.h
> > > +++ b/lib/meta-flow.h
> > > @@ -58,6 +58,7 @@ enum mf_field_id {
> > > #if FLOW_N_REGS > 7
> > > MFF_REG7, /* be32 */
> > > #endif
> > > +#define MFF_REG_END (MFF_REG0 + FLOW_N_REGS - 1)
> >
> > Do you plan to use it for something? Oh, I see, in patch 7. But
> > case A ... B:
> > is not valid C syntax (not even in C11 as far as I know) and it's not
> > among the GCC extensions that we commonly use in Open vSwitch.
> >
> > I'd take a patch that defined a macro to
> > case MFF_REG0: case MFF_REG1: case MFF_REG2: ...
> > etc. based on the number of registers, so that we could just use that
> > macro instead of writing out all the cases every time.
>
> I see. I'll create such patch.
Thanks. We can actually use such a macro in a number of places in
meta-flow.c already (e.g. mf_is_all_wild(), mf_get_mask()), so it'd be a
nice cleanup.
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
