Hi Jafar and Andrew,
I think that both development are complementary. The point that bother me is to
call this 'vrf-lite', it seems to be more policy based routing. As you said, we
can easily imagine a scenario with netns + multiple routing table. Calling it
vrf will confuse the user.
With only
Le 02/06/2015 08:54, David Lamparter a écrit :
On Fri, May 22, 2015 at 11:39:51AM +0200, Nicolas Dichtel wrote:
Here is the last update of this series. There is no major change with the v2 as
no technical issue were spotted.
Patches are also available at https://github.com/6WIND/quagga/commits
Le 02/06/2015 21:52, Donald Sharp a écrit :
Some minor comments inline
On Tue, Jun 2, 2015 at 11:57 AM, Nicolas Dichtel nicolas.dich...@6wind.com
wrote:
[snip]
diff --git a/lib/vrf.c b/lib/vrf.c
index 683026e5c069..c3e870ef8960 100644
--- a/lib/vrf.c
+++ b/lib/vrf.c
@@ -22,14 +22,48
Le 03/06/2015 11:27, Donald Sharp a écrit :
s/HAVE_NETNS/HAVE_SETNS in my post. I fat fingered the #define name :(. I
still think using the #define setup in the confgure.ac( and then config.h )
would be better than VRF_USE_NETNS as well as GNU_LINUX.
Ok, I will add a 'HAVE_NETNS'.
Regards,
lu.f...@6wind.com
Reviewed-by: Alain Ritoux alain.rit...@6wind.com
Signed-off-by: Nicolas Dichtel nicolas.dich...@6wind.com
Acked-by: Donald Sharp sha...@cumulusnetworks.com
---
lib/memtypes.c | 1 +
zebra/if_ioctl.c | 9 +-
zebra/if_ioctl_solaris.c | 9 +-
zebra/if_netlink.c
the client can do anything it wants in this
callback. For example, it may send requests for both default VRF
and some non-default VRFs.
Signed-off-by: Feng Lu lu.f...@6wind.com
Reviewed-by: Alain Ritoux alain.rit...@6wind.com
Signed-off-by: Nicolas Dichtel nicolas.dich...@6wind.com
Acked
Le 03/06/2015 19:40, Paul Jakma a écrit :
On Wed, 3 Jun 2015, Paul Jakma wrote:
At present, we are not clear on the design because several people seem to
think the ZServ API added by this patch-set adds still is compatible with
multiple-processes.
Oh, and this isn't about VRFs and zebra
to the specified VRF by
using the stored file descriptor, and then can allocate
a socket which is working in the associated netns.
Signed-off-by: Feng Lu lu.f...@6wind.com
Reviewed-by: Alain Ritoux alain.rit...@6wind.com
Signed-off-by: Nicolas Dichtel nicolas.dich...@6wind.com
---
configure.ac
Le 05/06/2015 09:01, Nicolas Dichtel a écrit :
[snip]
After the last three patches, I've one more ready in my queue to
rename vrf to domain because the term 'vrf' is too controversial.
Not sure: was there a consensus on the name?
domain? logical-table? vrf?
Regards,
Nicolas
Le 27/05/2015 03:26, Xiaorong (Andrew) Qu a écrit :
Hi Jafar,
Either way of merging should work.
As for terminology, using your example VRF 3 table 2 is not ideal to me.
I would like netns 3|blue table 2 as netns is the
keyword for namespace anyway.
That is right, what I am working on is
Le 28/05/2015 03:28, Xiaorong (Andrew) Qu a écrit :
I agree with that netns is linux specific.
Considering netns is really like the concept of VDC or Virtual router which
is
a physical admin domain, hence
how about something more neutual:
domain 1 table 2?
I really want to avoid the VRF key
lu.f...@6wind.com
Reviewed-by: Alain Ritoux alain.rit...@6wind.com
Signed-off-by: Nicolas Dichtel nicolas.dich...@6wind.com
---
lib/memtypes.c | 1 +
zebra/if_ioctl.c | 9 +-
zebra/if_ioctl_solaris.c | 9 +-
zebra/if_netlink.c | 4 +-
zebra/if_sysctl.c| 10
Le 23/05/2015 05:16, David Lamparter a écrit :
[snip]
Do you want to roll a patch or should I?
You wrote the right version, so I will let you submit it ;-)
Thank you,
Nicolas
___
Quagga-dev mailing list
Quagga-dev@lists.quagga.net
Le 23/05/2015 05:27, David Lamparter a écrit :
On Fri, May 22, 2015 at 11:39:52AM +0200, Nicolas Dichtel wrote:
babeld is gone in commit 336724d628be, let's remove all related code.
Note that this patch fixes vtysh, which cannot start after this commit:
$ vtysh
Command node 17 doesn't exist
the important thing is to move forward.
On 05/25/2015 07:12 PM, Paul Jakma wrote:
Hi Nicolas,
On Fri, 22 May 2015, Nicolas Dichtel wrote:
Here is the last update of this series. There is no major change with
the v2 as no technical issue were spotted. Patches are also available at
https
babeld is gone in commit 336724d628be, let's remove all related code.
Note that this patch fixes vtysh, which cannot start after this commit:
$ vtysh
Command node 17 doesn't exist, please check it
Fixes: 336724d628be (babeld: Remove babeld from Quagga)
Signed-off-by: Nicolas Dichtel nicolas.dich
...@6wind.com
Signed-off-by: Nicolas Dichtel nicolas.dich...@6wind.com
---
bgpd/bgp_network.c | 2 +-
bgpd/bgp_zebra.c | 6 +++---
ospf6d/ospf6_route.c | 15 +++
ospf6d/ospf6_zebra.c | 5 ++---
4 files changed, 13 insertions(+), 15 deletions(-)
diff --git a/bgpd/bgp_network.c
Here is the last update of this series. There is no major change with the v2 as
no technical issue were spotted.
Patches are also available at https://github.com/6WIND/quagga/commits/vrf
It would be nice to have some guide lines from maintainers about how to get
this series included.
If people
the client can do anything it wants in this
callback. For example, it may send requests for both default VRF
and some non-default VRFs.
Signed-off-by: Feng Lu lu.f...@6wind.com
Reviewed-by: Alain Ritoux alain.rit...@6wind.com
Signed-off-by: Nicolas Dichtel nicolas.dich...@6wind.com
---
bgpd
...@6wind.com
Reviewed-by: Alain Ritoux alain.rit...@6wind.com
Signed-off-by: Nicolas Dichtel nicolas.dich...@6wind.com
---
zebra/rib.h |6 +
zebra/zebra_rib.c | 14 +-
zebra/zebra_vty.c | 1091 ++---
3 files changed, 962 insertions(+), 149
lu.f...@6wind.com
Reviewed-by: Alain Ritoux alain.rit...@6wind.com
Signed-off-by: Nicolas Dichtel nicolas.dich...@6wind.com
---
lib/memtypes.c | 1 +
zebra/if_ioctl.c | 9 +-
zebra/if_ioctl_solaris.c | 9 +-
zebra/if_netlink.c | 4 +-
zebra/if_sysctl.c| 10
is formed.
In vrf_socket(), we first switch to the specified VRF by
using the stored file descriptor, and then can allocate
a socket which is working in the associated netns.
Signed-off-by: Feng Lu lu.f...@6wind.com
Reviewed-by: Alain Ritoux alain.rit...@6wind.com
Signed-off-by: Nicolas Dichtel
; or
- which have a parameter in type of struct interface.
All incoming routes are set to default VRF.
In fact, all routes in FIB are kept in default VRF. And the logic
is not changed.
Signed-off-by: Feng Lu lu.f...@6wind.com
Reviewed-by: Alain Ritoux alain.rit...@6wind.com
Signed-off-by: Nicolas
...@6wind.com
Signed-off-by: Nicolas Dichtel nicolas.dich...@6wind.com
---
zebra/rib.h| 8
zebra/zebra_rib.c | 3 ++-
zebra/zebra_routemap.c | 8 ++--
3 files changed, 16 insertions(+), 3 deletions(-)
diff --git a/zebra/rib.h b/zebra/rib.h
index 2d8805a88a83..84cd3da6c02c
.
* ripng_zebra_ipv6_add() and ripng_zebra_ipv6_delete() now can share
the common code. The common part is moved to ripng_zebra_ipv6_send().
Signed-off-by: Feng Lu lu.f...@6wind.com
Reviewed-by: Alain Ritoux alain.rit...@6wind.com
Signed-off-by: Nicolas Dichtel nicolas.dich...@6wind.com
---
ripngd/ripng_interface.c
show interface description vrf N
or all VRFs:
show interface vrf all
show interface IFNAME vrf all
show interface description vrf all
Signed-off-by: Feng Lu lu.f...@6wind.com
Reviewed-by: Alain Ritoux alain.rit...@6wind.com
Signed-off-by: Nicolas Dichtel nicolas.dich...@6wind.com
() was called or iflist was initialized, now call
vrf_init() instead;
- if if_terminate() was called or iflist was destroyed, now call
vrf_terminate() instead.
Signed-off-by: Feng Lu lu.f...@6wind.com
Reviewed-by: Alain Ritoux alain.rit...@6wind.com
Signed-off-by: Nicolas Dichtel nicolas.dich
to move the interace list into the
struct vrf. The main logic is not changed.
Signed-off-by: Feng Lu lu.f...@6wind.com
Reviewed-by: Alain Ritoux alain.rit...@6wind.com
Signed-off-by: Nicolas Dichtel nicolas.dich...@6wind.com
---
lib/if.c | 8
lib/if.h | 2 ++
zebra
-by: Feng Lu lu.f...@6wind.com
Reviewed-by: Alain Ritoux alain.rit...@6wind.com
Signed-off-by: Nicolas Dichtel nicolas.dich...@6wind.com
---
zebra/main.c | 40
zebra/rib.h | 1 +
zebra/test_main.c | 37 +
zebra
support multiple VRFs
(till now only zebra).
Signed-off-by: Feng Lu lu.f...@6wind.com
Reviewed-by: Alain Ritoux alain.rit...@6wind.com
Signed-off-by: Nicolas Dichtel nicolas.dich...@6wind.com
---
lib/if.c | 226 --
lib/if.h | 33
-by: Alain Ritoux alain.rit...@6wind.com
Signed-off-by: Nicolas Dichtel nicolas.dich...@6wind.com
---
lib/vrf.c | 102 +-
lib/vrf.h | 9 ++
2 files changed, 110 insertions(+), 1 deletion(-)
diff --git a/lib/vrf.c b/lib/vrf.c
index
Le 20/08/2015 00:40, Martin Winter a écrit :
Nicholas,
-h ;-)
Any update (or patch to test) for this?
Sorry no. I was busy (and off some weeks). I will look at this pb asap (but not
before the next week or the week after).
Regards,
Nicolas
___
Le 16/10/2015 14:24, Paul Jakma a écrit :
On Fri, 16 Oct 2015, Nicolas Dichtel wrote:
Hi Paul,
we are pausing with this project right now, due to some other priorities. We
still plan to upstream vrf support in ospf, but it is delayed. As you said, a
first milestone has been reached, I'm
Le 14/10/2015 22:39, Paul Jakma a écrit :
On Wed, 14 Oct 2015, Paul Jakma wrote:
This needs finishing off, to a state that it makes sense for a release - i.e.
it needs to at least be a waypoint along a path to a coherent end-state.
And all that may be needed is some kind of rough plan, and
Le 12/11/2015 18:43, Paul Jakma a écrit :
On Thu, 12 Nov 2015, Donald Sharp wrote:
So question:
What constitutes an ack? I posted up 60 odd patches over the last few days
and only have gotten some conversation around 2-3 of those patches. Are
the rest, by omission, assumed to be ok then?
Le 06/07/2015 09:50, Martin Winter a écrit :
It seems we have a bad commit in git master.
Commit c99f348 “*: add VRF ID in the API message header” is the bad commit.
It seems (at least) direct connected next hops in received BGP routes show up
as inaccessible with “ebgp-multihop” configured.
This patch adds a runtime check to determine if netns are available. Some
systems like OpenWRT have the system call setns() but don't have the kernel
option CONFIG_NET_NS enabled.
Reported-by: Christian Franke <ch...@opensourcerouting.org>
Signed-off-by: Nicolas Dichtel <nicolas.dich...@
)
Reported-by: Martin Winter <mwin...@opensourcerouting.org>
Signed-off-by: Nicolas Dichtel <nicolas.dich...@6wind.com>
Acked-by: Donald Sharp <sha...@cumulusnetworks.com>
Tested-by: Martin Winter <mwin...@opensourcerouting.org>
---
v4: add credits to Martin
v3: ensur
Le 16/09/2015 04:00, Martin Winter a écrit :
ACK
Tested-by: Martin Winter
Confirmed fixed. Thanks a lot!
Tested on top of latest quagga master (git 52c0bc7)
(Tested BGP with & without 4-byte extensions, each for IPv4 & IPv6. It’s now all
back to same state as
Signed-off-by: Nicolas Dichtel <nicolas.dich...@6wind.com>
Acked-by: Donald Sharp <sha...@cumulusnetworks.com>
---
Martin, could you test this new version?
v3: ensure to read the whole message before accessing it (call stream_read()
at the end of (zclient_read_header())
v2: a
Le 23/12/2015 20:26, Vipin Kumar a écrit :
Thanks Nicolas.
So, I assume you are fine with changing zebra commands (such as this) to..
'show ip route vrf A.B.C.D/M'
from its previous form..
'show ip route A.B.C.D/M vrf '
Yes ;-)
Regards,
Nicolas
Le 21/12/2015 22:21, Vipin Kumar a écrit :
As some of you might recall, we had converged on adding the vrf option in
BGP commands just they way it has for a view. I have made the necessary
changes and (painfully n boringly) adjusted the argument indexes. As for
the pending work, view option
Le 24/02/2016 01:10, Martin Winter a écrit :
Ok, false alert.
Lovely CI system picks random execution nodes during the git bisect
and one of the node had a stuck route - so by random it happened
to end up with this specific commit as “bad”
Ok, thank you for the explanation, the relationship
client.
(zebra_client_create) Call previous helper to send all VRFs to new clients
immediately.
* zserv.h: Export zsend_vrf_update helper.
* main.c: (zebra_vrf_{en,dis}able) Have the vrf enable/disable hooks
call zsend_vrf_update to update all clients.
---
Acked-by: Nicolas Dichtel <nic
Le 20/10/2016 à 12:15, Paul Jakma a écrit :
> Oh,
>
> I'm also thinking of deleting the various round-8 intermediate proposed
> branches, that were rendered unreferenced by any other head or tag after
> rebases
> (i.e. they're not in the history of current 'master'). I don't see a need for
>
Le 20/10/2016 à 12:12, Paul Jakma a écrit :
> On Thu, 20 Oct 2016, Nicolas Dichtel wrote:
>
>> For example, a clone of the repo does not fetch by default this kind of tag.
>> You need to do 'git fetch -t'.
>
> Ok, pushed a stable/1.0 branch, with the head comm
Le 18/10/2016 à 16:50, Paul Jakma a écrit :
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA256
>
> Quagga 1.1.0 has been released, available from the usual place:
>
> https://download.savannah.gnu.org/releases/quagga/
A naive question: is it expected that the tag of the release is not on a
Le 20/10/2016 à 11:20, Paul Jakma a écrit :
> On Thu, 20 Oct 2016, Nicolas Dichtel wrote:
>
>> A naive question: is it expected that the tag of the release is not on a
>> branch?
>
> A point-release of an old release maybe could just be referenced by the tag I
> g
48 matches
Mail list logo