This is to match the kernel implementation in the
patch "datapath: Add basic MPLS support to kernel"
by Leo Alterman.
Cc: Leo Alterman
Signed-off-by: Simon Horman
---
v2.2
* No change
v2.1
* Initial post
---
lib/dpif-netdev.c | 13 +
lib/odp-util.c| 43 +++
There are (or at least will be) cases where this check can produce false
positives. For example, a flow which matches a non-MPLS packet and then
applies an MPLS push action followed by an action to load the MPLS label.
I believe this is the same problem that was recently discussed in
relation to
Allow datapath to recognize and extract MPLS labels into flow keys
and execute actions which push, pop, and set labels on packets.
Based heavily on work by Leo Alterman.
Cc: Leo Alterman
Signed-off-by: Simon Horman
---
v2.2
* Call skb_reset_mac_header() in skb_cb_set_mpls_stack()
eth_hdr(sk
Hi,
This series implements basic MPLS actions and matches based on the kernel
datapath work by Leo Alterman.
Some limitations in the scope of this series are described in the changelog
entry for the last patch in the series.
The key differences between this series and the previous version (v2.1)
On 10/08/2012 10:24 PM, Jesse Gross wrote:
On Sat, Oct 6, 2012 at 10:21 AM, Gary Kotton wrote:
Hi,
When Quantum is using OVS and namespaces there are problems when namespace
is being deleted. More specifically this is when a default gateway is
assigned to an interface that is used by the OVS. A
On Tue, Oct 9, 2012 at 4:25 AM, Gary Kotton wrote:
> On 10/08/2012 10:24 PM, Jesse Gross wrote:
>>
>> On Sat, Oct 6, 2012 at 10:21 AM, Gary Kotton wrote:
>>>
>>> Hi,
>>> When Quantum is using OVS and namespaces there are problems when
>>> namespace
>>> is being deleted. More specifically this is
The change to execute_set_action(), it implies that OVS_KEY_ATTR_MPLS
got put in a place you don't want it in a previous patch:
parse_remaining_mpls() will read past the end of the packet.
I don't think that flow_extract() checks that the packet is long
enough for an IPv4 or IPv6 header when it s
On Mon, Oct 08, 2012 at 12:49:21PM -0700, Ethan Jackson wrote:
> The ESX userspace looks quite a bit like linux, but has some key
> differences which need to be specially handled in the build. To
> distinguish between ESX and systems use the linux datapath module,
> this patch adds two new macros
These patches build on the first two patches which Simon Horman sent out in
May to move Open vSwitch towards flow-based tunneling. The first patche adds
a tun_key, deprecating the tun_id member of the ovs_skb_cb struct. This patche
retain compatibilty with existing tunneling, but once the userspace
This is a first pass at providing a tun_key which can be
used as the basis for flow-based tunnelling. The
tun_key includes and replaces the tun_id in both struct
ovs_skb_cb and struct sw_tun_key.
This patch allows all existing tun_id behaviour to still work. Existing
users of tun_id are redirected
Move struct ovs_key_ipv4_tunnel to the top of sw_flow_key. Add a new
struct ovs_flow_hash, which contains both the key_len and the offset to
use when hashing. This allows the outer tunnel to be hashed and used when
looking up flows.
Signed-off-by: Kyle Mestery
---
datapath/actions.c | 1 +
dat
Commit 2de795adb96 (datapath: 64-bit GRE support) introduced a bug
for tunnels with no key. Following patch fixes it by setting tunnel
type to GRE type.
Bugs: 13511
Signed-off-by: Pravin B Shelar
---
datapath/vport-gre.c |5 -
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/
On Oct 9, 2012, at 2:55 PM, Pravin B Shelar wrote:
> Commit 2de795adb96 (datapath: 64-bit GRE support) introduced a bug
> for tunnels with no key. Following patch fixes it by setting tunnel
> type to GRE type.
>
> Bugs: 13511
> Signed-off-by: Pravin B Shelar
Acked-by: Kyle Mestery
___
> Does the 'argv' code in command-line.c actually malfunction on ESX?
> It'd be nice to just leave it in, if not. It's not really tied to
> having the Linux datapath, it's orthogonal. (It should actually work
> on many Unix and Unix-like systems, but we specialize it to Linux
> pending testing on
On Tue, Oct 09, 2012 at 01:54:28PM -0700, Ethan Jackson wrote:
> > Does the 'argv' code in command-line.c actually malfunction on ESX?
> > It'd be nice to just leave it in, if not. It's not really tied to
> > having the Linux datapath, it's orthogonal. (It should actually work
> > on many Unix an
> Maybe I'm mistaken, but IIRC ESX is the only case we know of where
> timer_create() doesn't work. We previously found this out only at
> runtime, but with your change we will instead disable timer caching at
> compile time so there's now (again) no need to disable it at runtime.
> No?
True, I a
On Tue, Oct 09, 2012 at 02:21:08PM -0700, Ethan Jackson wrote:
> > Maybe I'm mistaken, but IIRC ESX is the only case we know of where
> > timer_create() doesn't work. We previously found this out only at
> > runtime, but with your change we will instead disable timer caching at
> > compile time so
On 9 October 2012 17:18, Ben Pfaff wrote:
> On Tue, Oct 09, 2012 at 01:54:28PM -0700, Ethan Jackson wrote:
>> > Does the 'argv' code in command-line.c actually malfunction on ESX?
>> > It'd be nice to just leave it in, if not. It's not really tied to
>> > having the Linux datapath, it's orthogona
On Tue, Oct 09, 2012 at 06:03:32PM -0400, Ed Maste wrote:
> On 9 October 2012 17:18, Ben Pfaff wrote:
> > On Tue, Oct 09, 2012 at 01:54:28PM -0700, Ethan Jackson wrote:
> >> > Does the 'argv' code in command-line.c actually malfunction on ESX?
> >> > It'd be nice to just leave it in, if not. It's
Here's an incremental.
---
configure.ac| 4 ++--
lib/socket-util.c | 12 +---
lib/timeval.c | 42 +++---
lib/timeval.h | 16
tests/test-timeval.c| 4 ++--
vswitchd/system-stats.c | 21
On Tue, Oct 09, 2012 at 03:09:51PM -0700, Ethan Jackson wrote:
> Here's an incremental.
Thanks for humoring me, this seems fine.
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
Thanks for the review, I'll merge this shortly.
Ethan
On Tue, Oct 9, 2012 at 3:16 PM, Ben Pfaff wrote:
> On Tue, Oct 09, 2012 at 03:09:51PM -0700, Ethan Jackson wrote:
>> Here's an incremental.
>
> Thanks for humoring me, this seems fine.
___
dev maili
Cabestan. Layout Simple
Asegúrate de no perderte ninguna oferta, añade ofer...@globalbono.com a tu
lista de contactos.
Si no ves correctamente las imágenes, pulsa [
http://email.globalbono.com/E09102012154627.cfm?WL=1845&WS=208833_8707685&WA=965
] aquí
[ http://email.globalbono.com/Go/index.c
On Tue, Oct 9, 2012 at 12:57 PM, Kyle Mestery (kmestery)
wrote:
> On Oct 9, 2012, at 2:55 PM, Pravin B Shelar wrote:
>> Commit 2de795adb96 (datapath: 64-bit GRE support) introduced a bug
>> for tunnels with no key. Following patch fixes it by setting tunnel
>> type to GRE type.
>>
>> Bugs: 13511
Now GRE64 deals with tunnel with no key and tunnel with zero key
as same. This behaviour is different than standard GRE.
Signed-off-by: Pravin B Shelar
---
datapath/vport-gre.c | 12 +---
1 file changed, 9 insertions(+), 3 deletions(-)
diff --git a/datapath/vport-gre.c b/datapath/vpor
On Tue, Oct 9, 2012 at 4:31 PM, Pravin B Shelar wrote:
> Now GRE64 deals with tunnel with no key and tunnel with zero key
> as same. This behaviour is different than standard GRE.
>
> Signed-off-by: Pravin B Shelar
Acked-by: Jesse Gross
___
dev mailin
On Tue, Oct 9, 2012 at 5:10 PM, Jesse Gross wrote:
> On Tue, Oct 9, 2012 at 4:31 PM, Pravin B Shelar wrote:
>> Now GRE64 deals with tunnel with no key and tunnel with zero key
>> as same. This behaviour is different than standard GRE.
>>
>> Signed-off-by: Pravin B Shelar
>
> Acked-by: Jesse Gros
On Tue, Oct 09, 2012 at 04:08:32PM +0900, Simon Horman wrote:
> +static int push_mpls(struct sk_buff *skb, const struct ovs_action_push_mpls
> *mpls)
> +{
> + u32 l2_size;
> + __be32 *new_mpls_label;
> +
> + if (skb_cow_head(skb, MPLS_HLEN) < 0) {
> + kfree_skb(skb);
> +
On Tue, Oct 09, 2012 at 04:08:33PM +0900, Simon Horman wrote:
> This is to match the kernel implementation in the
> patch "datapath: Add basic MPLS support to kernel"
> by Leo Alterman.
>
> Cc: Leo Alterman
> Signed-off-by: Simon Horman
>
> ---
>
> v2.2
> * No change
>
> v2.1
> * Initial post
On Tue, Oct 09, 2012 at 04:08:35PM +0900, Simon Horman wrote:
> @@ -938,6 +990,15 @@ mf_is_value_valid(const struct mf_field *mf, const union
> mf_value *value)
> case MFF_IPV6_LABEL:
> return !(value->be32 & ~htonl(IPV6_LABEL_MASK));
>
> +case MFF_MPLS_LABEL:
> +return
30 matches
Mail list logo
