On Fri, Jun 19, 2015 at 03:44:44PM +0200, Patrick McHardy wrote:
> On 19.06, Pablo Neira Ayuso wrote:
> > On Fri, Jun 19, 2015 at 03:13:36PM +0200, Patrick McHardy wrote:
> > > > static int do_add_setelems(struct netlink_ctx *ctx, const struct handle *h,
> > > > - const struct expr *expr)
> > > > + const struct location *loc, struct expr *expr)
> > > > {
> > > > - if (netlink_add_setelems(ctx, h, expr) < 0)
> > > > + struct set *set;
> > > > +
> > > > + if (netlink_get_set(ctx, h, loc) < 0)
> > >
> > > I think we should get it from the internal list and not from the
> > > kernel.
> >
> > There is no such internal list at this moment, we retrieve the list
> > only for do_command_list(). Are you suggesting to add the code to
> > retrieve it inconditionally initially?
>
> We do have the table->sets list. Yeah, but we don't add it to that
> list in the creation part, I see. Actually we should be doing that,
> since otherwise we also won't support creating a set and adding
> new elements in seperate commands but a single transaction.
Right.
> > > We can't add intervals to existing sets so far anyways, and this
> > > would allow it, but it wouldn't work.
> >
> > Not sure what you mean with this.
>
> Intervals need to know the entire set content to be created correctly.
> We don't handle incremental updates with intervals correctly ATM.
What's the problem?
--
To unsubscribe from this list: send the line "unsubscribe netfilter-devel" in
