On Mon, May 30, 2016 at 02:46:18PM +0200, Martin Pieuchot wrote:
> Here's a corrected version of the previous diff that got backed out.
>
> The idea is still to stop calling nd6_output() with a NULL ``rt''
> argument in order to always use the same code path when inserting
> routes.
>
> ok?
OK b
Here's a corrected version of the previous diff that got backed out.
The idea is still to stop calling nd6_output() with a NULL ``rt''
argument in order to always use the same code path when inserting
routes.
ok?
Index: net/pf.c
===
On 24/05/16(Tue) 00:19, Alexander Bluhm wrote:
> On Mon, May 23, 2016 at 04:55:06PM +0200, Martin Pieuchot wrote:
> > @@ -1532,42 +1527,17 @@ nd6_output(struct ifnet *ifp, struct mbu
> ...
> > - /*
> > -* Since nd6_is_addr_neighbor() internally calls nd6_lookup(),
> > -
On Mon, May 23, 2016 at 04:55:06PM +0200, Martin Pieuchot wrote:
> +++ net/pf.c 23 May 2016 14:43:19 -
> +++ net/pf_norm.c 23 May 2016 14:48:01 -
I have tested the pf part with regress/sys/net/pf_forward. This
is OK bluhm@.
> @@ -1532,42 +1527,17 @@ nd6_output(struct ifnet *ifp, str
On 23/05/16(Mon) 15:59, Alexander Bluhm wrote:
> On Mon, May 23, 2016 at 02:47:02PM +0200, Martin Pieuchot wrote:
> > @@ -750,7 +750,15 @@ pf_refragment6(struct mbuf **m0, struct
> > if (ifp == NULL) {
> > ip6_forward(m, 0);
> > }
On Mon, May 23, 2016 at 02:47:02PM +0200, Martin Pieuchot wrote:
> @@ -5506,6 +5506,7 @@ pf_route(struct mbuf **m, struct pf_rule
> int error = 0;
> unsigned int rtableid;
>
> +
> if (m == NULL || *m == NULL || r == NULL ||
> (dir != PF
This is the companion diff to the ether_output() one. Since bluhm@
mentioned it, here it is!
The idea is, once again, to stop inserting cloned route entries in L2
functions. The only place that insert cloned route entries should be
inside rtalloc(9) when RT_RESOLVE is passed.
Index: net/pf.c
==
