On 31/01/17(Tue) 12:16, Jeremie Courreges-Anglas wrote:
> Martin Pieuchot <[email protected]> writes:
> [...] 
> Yes, but I don't know which threshold warrants introducing a new
> function.  So far those are the first cases for counters_dec(); that
> includes wip rip6stat, icmp6stat, icmpstat and tcpstat.  And I'm not
> even sure those decrements make sense.  I guess we can add counters_dec()
> and then later remove it if it is deemed useless.

Let's convert all the counters first.  If some people think it doesn't
make sense they can change the code afterward.

> >> @@ -222,7 +225,9 @@ rip6_input(struct mbuf **mp, int *offp, 
> >>                        ICMP6_PARAMPROB_NEXTHEADER,
> >>                        prvnxtp - mtod(m, u_int8_t *));
> >>            }
> >> -          ip6stat.ip6s_delivered--;
> >> +          counters = counters_enter(&ref, ip6counters);
> >> +          counters[ip6s_delivered]--;
> >> +          counters_leave(&ref, ip6counters);
> >
> > Should we also use your counters_dec() here or do you think it's better
> > to merge icmp6_rip6_input() and rip6_input()?
> 
> I'm fine with using counters_dec(), I can see that those functions are
> almost twins but I don't want to open a can of worms...

Why not?  Fewer code is always good!  Anyway, this is something for
later.

Reply via email to