On Feb 13, 2013, at 16:44 , ext Lorand Jakab wrote:
> +static int lisp_tnl_send(struct vport *vport, struct sk_buff *skb)
> +{
> + int network_offset = skb_network_offset(skb);
> +
> + /* We only encapsulate IPv4 and IPv6 packets */
> + switch (ntohs(skb->protocol)) {
> + case ETH_
On Feb 14, 2013, at 12:59 PM, Rajahalme, Jarno (NSN - FI/Espoo) wrote:
>
> On Feb 13, 2013, at 16:44 , ext Lorand Jakab wrote:
>> +static int lisp_tnl_send(struct vport *vport, struct sk_buff *skb)
>> +{
>> +int network_offset = skb_network_offset(skb);
>> +
>> +/* We only encapsulate IP
With the recent flow based tunneling changes, VXLAN support was broken. These
patches address this by enabling VXLAN to work as it did before flow based
tunneling was pushed upstream.
Kyle Mestery (6):
Modify netdev_vport_get_dpif_port() to return a name for VXLAN ports
which includes th
Add utility functions to the simap structure. These are
used by future patches in this seris. The functions added are.
Signed-off-by: Kyle Mestery
Acked-by: Ethan Jackson
---
lib/simap.c | 19 +++
lib/simap.h | 2 ++
2 files changed, 21 insertions(+)
diff --git a/lib/simap.c b
Move dpif_backer->tnl_backers from a "struct sset" to a
"struct simap". Store odp_port in the new map. This will make it easier to
access the odp_port for future patches.
Signed-off-by: Kyle Mestery
Acked-by: Ethan Jackson
---
ofproto/ofproto-dpif.c | 19 +++
1 file changed, 11
Modify netdev_vport_get_dpif_port() to return a name for
VXLAN ports which includes the destination UDP port number as a part of the
name.
Signed-off-by: Kyle Mestery
Acked-by: Ethan Jackson
---
lib/netdev-vport.c | 36 +---
1 file changed, 33 insertions(+), 3 de
In tnl_set_config(), when determining if a tunnel port
already exists, make sure to also check the destination port. For VXLAN, this
can be different and allows multiple VXLAN ports in the datapath to be
created.
Signed-off-by: Kyle Mestery
---
datapath/tunnel.c | 18 --
1 file c
Garbage collect tnl_backers during type_run(). Add new
tnl_backers if a VXLAN ports UDP port changes.
Signed-off-by: Kyle Mestery
---
ofproto/ofproto-dpif.c | 118 +
1 file changed, 71 insertions(+), 47 deletions(-)
diff --git a/ofproto/ofproto-dp
Signed-off-by: Kyle Mestery
Acked-by: Ethan Jackson
---
lib/dpif-linux.c | 11 +++
1 file changed, 11 insertions(+)
diff --git a/lib/dpif-linux.c b/lib/dpif-linux.c
index b6eba39..34b9532 100644
--- a/lib/dpif-linux.c
+++ b/lib/dpif-linux.c
@@ -482,12 +482,14 @@ dpif_linux_port_add(stru
On Feb 14, 2013, at 8:37 AM, Kyle Mestery
wrote:
> With the recent flow based tunneling changes, VXLAN support was broken. These
> patches address this by enabling VXLAN to work as it did before flow based
> tunneling was pushed upstream.
>
> Kyle Mestery (6):
> Modify netdev_vport_get_dpif_por
On Thu, Feb 14, 2013 at 09:37:27AM -0500, Kyle Mestery wrote:
> Add utility functions to the simap structure. These are
> used by future patches in this seris. The functions added are.
>
> Signed-off-by: Kyle Mestery
> Acked-by: Ethan Jackson
Applied to master, thanks.
The standards fascist in
On Feb 14, 2013, at 11:37 AM, Ben Pfaff wrote:
> On Thu, Feb 14, 2013 at 09:37:27AM -0500, Kyle Mestery wrote:
>> Add utility functions to the simap structure. These are
>> used by future patches in this seris. The functions added are.
>>
>> Signed-off-by: Kyle Mestery
>> Acked-by: Ethan Jackson
On Thu, Feb 14, 2013 at 09:37:25AM -0500, Kyle Mestery wrote:
> Modify netdev_vport_get_dpif_port() to return a name for
> VXLAN ports which includes the destination UDP port number as a part of the
> name.
>
> Signed-off-by: Kyle Mestery
> Acked-by: Ethan Jackson
I applied the following increm
On Feb 14, 2013, at 11:55 AM, Ben Pfaff wrote:
> On Thu, Feb 14, 2013 at 09:37:25AM -0500, Kyle Mestery wrote:
>> Modify netdev_vport_get_dpif_port() to return a name for
>> VXLAN ports which includes the destination UDP port number as a part of the
>> name.
>>
>> Signed-off-by: Kyle Mestery
>>
On Thu, Feb 14, 2013 at 09:37:26AM -0500, Kyle Mestery wrote:
> Signed-off-by: Kyle Mestery
> Acked-by: Ethan Jackson
I'm pretty sure this introduces a memory leak because ofpbuf_init()
allocates data and I don't see it getting freed anywhere. So I applied
the following incremental:
diff --git
On Feb 14, 2013, at 12:00 PM, Ben Pfaff
wrote:
> On Thu, Feb 14, 2013 at 09:37:26AM -0500, Kyle Mestery wrote:
>> Signed-off-by: Kyle Mestery
>> Acked-by: Ethan Jackson
>
> I'm pretty sure this introduces a memory leak because ofpbuf_init()
> allocates data and I don't see it getting freed any
On Thu, Feb 14, 2013 at 09:37:28AM -0500, Kyle Mestery wrote:
> Move dpif_backer->tnl_backers from a "struct sset" to a
> "struct simap". Store odp_port in the new map. This will make it easier to
> access the odp_port for future patches.
>
> Signed-off-by: Kyle Mestery
> Acked-by: Ethan Jackson
On Thu, Feb 14, 2013 at 09:37:29AM -0500, Kyle Mestery wrote:
> In tnl_set_config(), when determining if a tunnel port
> already exists, make sure to also check the destination port. For VXLAN, this
> can be different and allows multiple VXLAN ports in the datapath to be
> created.
>
> Signed-off-
On Feb 14, 2013, at 12:07 PM, Ben Pfaff wrote:
> On Thu, Feb 14, 2013 at 09:37:28AM -0500, Kyle Mestery wrote:
>> Move dpif_backer->tnl_backers from a "struct sset" to a
>> "struct simap". Store odp_port in the new map. This will make it easier to
>> access the odp_port for future patches.
>>
>>
On Thu, Feb 14, 2013 at 09:37:30AM -0500, Kyle Mestery wrote:
> Garbage collect tnl_backers during type_run(). Add new
> tnl_backers if a VXLAN ports UDP port changes.
>
> Signed-off-by: Kyle Mestery
The error handling here is bad. If it fails to add or remove a port,
then ovs_assert() aborts t
On Feb 14, 2013, at 12:33 PM, Ben Pfaff wrote:
> On Thu, Feb 14, 2013 at 09:37:30AM -0500, Kyle Mestery wrote:
>> Garbage collect tnl_backers during type_run(). Add new
>> tnl_backers if a VXLAN ports UDP port changes.
>>
>> Signed-off-by: Kyle Mestery
>
> The error handling here is bad. If it
On Thu, Feb 14, 2013 at 06:47:56PM +, Kyle Mestery (kmestery) wrote:
> On Feb 14, 2013, at 12:33 PM, Ben Pfaff wrote:
> > On Thu, Feb 14, 2013 at 09:37:30AM -0500, Kyle Mestery wrote:
> >> Garbage collect tnl_backers during type_run(). Add new
> >> tnl_backers if a VXLAN ports UDP port changes
On Wed, Feb 13, 2013 at 03:31:42PM -0800, Justin Pettit wrote:
> For VLAN splinters, an "initial_tci" value was introduced that is passed
> around during flow processing to be used later for action translation.
> This commit switches to passing around a struct so that additional
> values beyond TCI
This patch removes unused TNL_F_IPSEC flag.
Issue: 14870
Signed-off-by: Ansis Atteka
---
datapath/tunnel.h|2 +-
include/openvswitch/tunnel.h |1 -
2 files changed, 1 insertion(+), 2 deletions(-)
diff --git a/datapath/tunnel.h b/datapath/tunnel.h
index 7e4d1a6..93ce243 10064
If IPsec policy uses skb mark, then we have to do route look up
with skb mark as well.
Issue: 14870
Signed-off-by: Ansis Atteka
---
datapath/tunnel.c |7 ---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/datapath/tunnel.c b/datapath/tunnel.c
index 6193891..5564d32 100644
The new ovs-monitor-ipsec implementation will use skb marks in
IPsec policies. This patch will configure datapath to use these
skb marks for IPsec tunnel packets.
Issue: 14870
Signed-off-by: Ansis Atteka
---
lib/odp-util.c | 12 +---
lib/odp-util.h |4 ++--
ofproto/
All these 3 patches should be applied on branch-1.10.
On Thu, Feb 14, 2013 at 11:50 AM, Ansis Atteka wrote:
> This patch removes unused TNL_F_IPSEC flag.
>
> Issue: 14870
> Signed-off-by: Ansis Atteka
> ---
> datapath/tunnel.h|2 +-
> include/openvswitch/tunnel.h |1 -
> 2 f
On Wed, Feb 13, 2013 at 03:31:43PM -0800, Justin Pettit wrote:
> When a packet arrives on an IP tunnel, store the TOS value for later
> use. This value will be used in a couple of future commits.
>
> Signed-off-by: Justin Pettit
...
> @@ -4922,6 +4928,7 @@ flow_push_stats(struct rule_dpif *rul
On Wed, Feb 13, 2013 at 03:31:44PM -0800, Justin Pettit wrote:
> Move the check for whether tunneled packets should be dropped due to
> congestion encountered (CE) when the encapsulated packet is not ECN
> capable (non-ECT). This also adds some additional tests for ECN
> handling on tunnel decapsu
On Wed, Feb 13, 2013 at 03:31:45PM -0800, Justin Pettit wrote:
> In the kernel tunnel implementation, if a packet was marked as ECN CE on
> the outer packet then we would carry this over to the inner packet on
> decapsulation. With the switch to flow based tunneling, this stopped
> happening. Thi
v2 fixes error handling noticed by Ben.
---
Garbage collect tnl_backers during type_run(). Add new
tnl_backers if a VXLAN ports UDP port changes.
Signed-off-by: Kyle Mestery
---
ofproto/ofproto-dpif.c | 120 ++---
1 file changed, 73 insertions(+), 47 d
On Feb 14, 2013, at 12:49 PM, Ben Pfaff wrote:
> On Thu, Feb 14, 2013 at 06:47:56PM +, Kyle Mestery (kmestery) wrote:
>> On Feb 14, 2013, at 12:33 PM, Ben Pfaff wrote:
>>> On Thu, Feb 14, 2013 at 09:37:30AM -0500, Kyle Mestery wrote:
Garbage collect tnl_backers during type_run(). Add new
On Thu, Feb 14, 2013 at 08:19:29PM +, Kyle Mestery (kmestery) wrote:
> On Feb 14, 2013, at 12:49 PM, Ben Pfaff wrote:
> > On Thu, Feb 14, 2013 at 06:47:56PM +, Kyle Mestery (kmestery) wrote:
> >> On Feb 14, 2013, at 12:33 PM, Ben Pfaff wrote:
> >>> On Thu, Feb 14, 2013 at 09:37:30AM -0500
On Thu, Feb 14, 2013 at 11:50 AM, Ansis Atteka wrote:
> This patch removes unused TNL_F_IPSEC flag.
>
> Issue: 14870
> Signed-off-by: Ansis Atteka
> ---
> datapath/tunnel.h|2 +-
> include/openvswitch/tunnel.h |1 -
> 2 files changed, 1 insertion(+), 2 deletions(-)
>
> diff -
On Thu, Feb 14, 2013 at 11:50 AM, Ansis Atteka wrote:
> If IPsec policy uses skb mark, then we have to do route look up
> with skb mark as well.
>
> Issue: 14870
> Signed-off-by: Ansis Atteka
> ---
> datapath/tunnel.c |7 ---
> 1 file changed, 4 insertions(+), 3 deletions(-)
>
> diff --g
On Thu, Feb 14, 2013 at 11:50:53AM -0800, Ansis Atteka wrote:
> The new ovs-monitor-ipsec implementation will use skb marks in
> IPsec policies. This patch will configure datapath to use these
> skb marks for IPsec tunnel packets.
>
> Issue: 14870
> Signed-off-by: Ansis Atteka
This provides a wa
On Thu, Feb 14, 2013 at 11:50 AM, Ansis Atteka wrote:
> The new ovs-monitor-ipsec implementation will use skb marks in
> IPsec policies. This patch will configure datapath to use these
> skb marks for IPsec tunnel packets.
>
> Issue: 14870
> Signed-off-by: Ansis Atteka
> ---
> lib/odp-util.c
On Thu, Feb 14, 2013 at 01:13:29PM -0800, Pravin Shelar wrote:
> On Thu, Feb 14, 2013 at 11:50 AM, Ansis Atteka wrote:
> > The new ovs-monitor-ipsec implementation will use skb marks in
> > IPsec policies. This patch will configure datapath to use these
> > skb marks for IPsec tunnel packets.
> >
The default values can be different and usually comes from /etc/logrotate.conf.
Signed-off-by: Gurucharan Shetty
---
rhel/etc_logrotate.d_openvswitch |2 ++
xenserver/etc_logrotate.d_openvswitch |2 ++
2 files changed, 4 insertions(+)
diff --git a/rhel/etc_logrotate.d_openvswitch b
In my tests, I see that creating a tar.gz takes a lot less
time than creating a tar.bz2. The difference in compressed
size of the debug bundle is not much different when either
of the above is used. So, use tar.gz as the default debug
bundle type.
Test results in my setup:
For an uncompressed debu
There is some code that is unused and some commands whose
output we probably don't need to debug openvswitch issues.
Signed-off-by: Gurucharan Shetty
---
utilities/bugtool/ovs-bugtool.in | 95 --
1 file changed, 95 deletions(-)
diff --git a/utilities/bugtoo
Currently we have a 50 MB size limitation for all logs. This looks
quite less because a single uncompressed log can be 50 MB which
will result in ovs-bugtool picking a single log.
While debugging issues related to openvswitch, it is important that
we have all logs related to openvswitch atleast. T
There can be a few hundred virtual interfaces in a hypervisor.
Some of the ethtool commands that we currently run on these devices
probably does not provide any extra information. So remove them
for tap and vif interfaces.
Also bump up the size limitation for CAP_NETWORK_STATUS. The
current value
A big reason for a large debug bundle size is the size of log
files. By default we collect 20 rotated logs for each logfile.
Most of the times we collect the debug bundle as soon as we
hit a bug. In such cases, we know that we need only one day's
worth of logs.
This patch adds an option, '--log-da
On Thu, Feb 14, 2013 at 01:28:16PM -0800, Gurucharan Shetty wrote:
> The default values can be different and usually comes from
> /etc/logrotate.conf.
>
> Signed-off-by: Gurucharan Shetty
Do you happen to know what the defaults are on RHEL and XenServer? That
is, are we making a real change he
On Thu, Feb 14, 2013 at 01:28:18PM -0800, Gurucharan Shetty wrote:
> A big reason for a large debug bundle size is the size of log
> files. By default we collect 20 rotated logs for each logfile.
> Most of the times we collect the debug bundle as soon as we
> hit a bug. In such cases, we know that
>
> Do you happen to know what the defaults are on RHEL and XenServer? That
> is, are we making a real change here? (It looks like the Debian
> /etc/logrotate.conf default is weekly rotation and no compression.)
For xenserver it is daily and compress. Fro rhel it is weekly and no-compress.
I gues
On Thu, Feb 14, 2013 at 01:28:21PM -0800, Gurucharan Shetty wrote:
> There is some code that is unused and some commands whose
> output we probably don't need to debug openvswitch issues.
>
> Signed-off-by: Gurucharan Shetty
This patch and the others that I didn't already comment on (I think
tha
On Thu, Feb 14, 2013 at 02:14:20PM -0800, Gurucharan Shetty wrote:
> >
> > Do you happen to know what the defaults are on RHEL and XenServer? That
> > is, are we making a real change here? (It looks like the Debian
> > /etc/logrotate.conf default is weekly rotation and no compression.)
> For xens
On Thu, Feb 14, 2013 at 2:13 PM, Ben Pfaff wrote:
> On Thu, Feb 14, 2013 at 01:28:18PM -0800, Gurucharan Shetty wrote:
>> A big reason for a large debug bundle size is the size of log
>> files. By default we collect 20 rotated logs for each logfile.
>> Most of the times we collect the debug bundle
On Thu, Feb 14, 2013 at 2:15 PM, Ben Pfaff wrote:
> On Thu, Feb 14, 2013 at 02:14:20PM -0800, Gurucharan Shetty wrote:
>> >
>> > Do you happen to know what the defaults are on RHEL and XenServer? That
>> > is, are we making a real change here? (It looks like the Debian
>> > /etc/logrotate.conf d
On Thu, Feb 14, 2013 at 1:16 PM, Ben Pfaff wrote:
> On Thu, Feb 14, 2013 at 01:13:29PM -0800, Pravin Shelar wrote:
>> On Thu, Feb 14, 2013 at 11:50 AM, Ansis Atteka wrote:
>> > The new ovs-monitor-ipsec implementation will use skb marks in
>> > IPsec policies. This patch will configure datapath t
On Wed, Feb 13, 2013 at 6:44 AM, Lorand Jakab wrote:
> LISP is an experimental layer 3 tunneling protocol, described in RFC
> 6830. This patch adds support for LISP tunneling. Since LISP
> encapsulated packets do not carry an Ethernet header, it is removed
> before encapsulation, and added with
On Thu, Feb 14, 2013 at 3:29 PM, Ansis Atteka wrote:
> On Thu, Feb 14, 2013 at 1:16 PM, Ben Pfaff wrote:
>> On Thu, Feb 14, 2013 at 01:13:29PM -0800, Pravin Shelar wrote:
>>> On Thu, Feb 14, 2013 at 11:50 AM, Ansis Atteka wrote:
>>> > The new ovs-monitor-ipsec implementation will use skb marks i
Also I verified that an item in OPENFLOW-1.1+ was already implemented, so
this commit removes it.
Signed-off-by: Ben Pfaff
---
DESIGN| 120 +++--
OPENFLOW-1.1+ |4 --
2 files changed, 116 insertions(+), 8 deletions(-)
diff --git a
This patch removes unused TNL_F_IPSEC flag.
Issue: 14870
Signed-off-by: Ansis Atteka
---
datapath/tunnel.h|2 +-
include/openvswitch/tunnel.h |2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/datapath/tunnel.h b/datapath/tunnel.h
index 7e4d1a6..93ce243 100
If IPsec policy uses skb mark, then we have to do route look up
with skb mark as well.
Issue: 14870
Signed-off-by: Ansis Atteka
---
datapath/tunnel.c | 16 +---
1 file changed, 13 insertions(+), 3 deletions(-)
diff --git a/datapath/tunnel.c b/datapath/tunnel.c
index 6193891..b2d7b
The new ovs-monitor-ipsec implementation will use skb marks in
IPsec policies. This patch will configure datapath to use these
skb marks for IPsec tunnel packets.
Issue: 14870
Signed-off-by: Ansis Atteka
---
lib/odp-util.c |9 +
lib/odp-util.h |2 ++
ofproto/ofpro
On Thu, Feb 14, 2013 at 6:17 PM, Ansis Atteka wrote:
> This patch removes unused TNL_F_IPSEC flag.
>
> Issue: 14870
> Signed-off-by: Ansis Atteka
Acked-by: Jesse Gross
___
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/de
On Feb 14, 2013, at 11:07 AM, Ben Pfaff wrote:
> On Wed, Feb 13, 2013 at 03:31:42PM -0800, Justin Pettit wrote:
>> For VLAN splinters, an "initial_tci" value was introduced that is passed
>> around during flow processing to be used later for action translation.
>> This commit switches to passing
On Thu, Feb 14, 2013 at 6:20 PM, Ansis Atteka wrote:
> If IPsec policy uses skb mark, then we have to do route look up
> with skb mark as well.
>
> Issue: 14870
> Signed-off-by: Ansis Atteka
Acked-by: Jesse Gross
___
dev mailing list
dev@openvswitch.o
On Thu, Feb 14, 2013 at 6:20 PM, Ansis Atteka wrote:
> diff --git a/lib/odp-util.c b/lib/odp-util.c
> index 7e48981..08195b7 100644
> --- a/lib/odp-util.c
> +++ b/lib/odp-util.c
> +void
> +odp_put_skb_mark_action(const uint32_t skb_mark,
> +struct ofpbuf *odp_actions)
> +{
Hi All,
We are using the VXLAN changes done to the open vswitch. We have taken
these changes from https://github.com/mestery/ovs-vxlan. I think this is
not added to main line openvswitch source code.
We are observing a Kernel crash issue with this code. Just wanted to check
if this issue is alrea
The VXLAN changes were upstreamed a little while ago. Can you try either the
tip of master or the branch "branch-1.10"?
--Justin
On Feb 14, 2013, at 11:28 PM, Jagadish Nadimpalli
wrote:
> Hi All,
>
> We are using the VXLAN changes done to the open vswitch. We have taken these
> changes fr
64 matches
Mail list logo
