Hi Ben,
I confirmed it works well!
Thanks so much!
...hiroshi miyata
(13/04/17 9:04), miyahiro.d...@gmail.com wrote:
Hi Ben,
Thanks, I will try the patch!
I will sent you the result shortly.
Best Regards,
...hiroshi miyata
(13/04/16 6:02), Ben Pfaff wrote:
NXM puts the DSCP value in
On Thu, 2013-04-18 at 12:50 +0100, Ian Campbell wrote:
Based on Waldi's RFC at
http://lists.xen.org/archives/html/xen-devel/2012-09/msg00943.html
To use it set vif.default.script=vif-openvswitch in /etc/xen/xl.conf or use
script=vif-openvswitch in the vif configuration.
Appears to do the
(adding ovs list to this subthread too)
On Thu, 2013-04-18 at 13:19 +0100, Bastian Blank wrote:
On Thu, Apr 18, 2013 at 12:50:40PM +0100, Ian Campbell wrote:
Based on Waldi's RFC at
http://lists.xen.org/archives/html/xen-devel/2012-09/msg00943.html
BRIDGE_NAME[.VLAN][:TRUNK,TRUNK]
Change the base flow only if a corresponding kernel action is generated
in commit_odp_tunnel_action().
Signed-off-by: Jarno Rajahalme jarno.rajaha...@nsn.com
---
lib/odp-util.c |9 -
1 file changed, 4 insertions(+), 5 deletions(-)
diff --git a/lib/odp-util.c b/lib/odp-util.c
index
Adds tun_src and tun_dst match and set capabilities via new NXM fields
NXM_NX_TUN_IPV4_SRC and NXM_NX_TUN_IPV4_DST. This allows management of
large number of tunnels via flow tables, without requiring the tunnels
to be pre-configured.
Flow-based tunnels can be configured with options
As tunnel metadata is no longer cleared on input, and the input
values are retained in 'ctx-flow' accross tunnel output actions,
there is no need to store the tunnel.ip_tos to initial_vals.
When needed, tunnel.ip_tos is gotten from ctx-flow instead of
ctx-base_flow, since ctx-flow retains the
Do not clear tunnel metadata on tunnel input. This is used by a later patch.
Signed-off-by: Jarno Rajahalme jarno.rajaha...@nsn.com
---
lib/match.c|4
ofproto/ofproto-dpif.c | 26 --
ofproto/tunnel.c |3 +--
3 files changed, 13
This makes sure that output actions leave no changes to any flow fields,
while all explicit set_field actions are retained actross output actions.
Signed-off-by: Jarno Rajahalme jarno.rajaha...@nsn.com
---
ofproto/ofproto-dpif.c | 20
1 file changed, 16 insertions(+), 4
Adds tun_src and tun_dst match and set capabilities via new NXM fields
NXM_NX_TUN_IPV4_SRC and NXM_NX_TUN_IPV4_DST. This allows management of
large number of tunnels via flow tables, without requiring the tunnels
to be pre-configured.
NXM_NX_TUN_IPV4_SRC and NXM_NX_TUN_IPV4_DST are also included
As the flow is no longer modified (apart from the in_port) on tunnel
input, the perfect fitness can be retained.
Signed-off-by: Jarno Rajahalme jarno.rajaha...@nsn.com
---
ofproto/ofproto-dpif.c |3 ---
1 file changed, 3 deletions(-)
diff --git a/ofproto/ofproto-dpif.c
On Tue, Apr 02, 2013 at 08:38:17AM -0700, Justin Pettit wrote:
Hidden flows, such as in-band control, are not visible when using
OpenFlow to view stats. However, the OpenFlow Table Stats message would
report them back, which is confusing. This commit changes the code to
count only visible
All genl callbacks are serialized by genl-mutex. This can become
bottleneck in multi threaded case.
Following patch adds an parameter to genl_family so that a
particular family can get concurrent netlink callback without
genl_lock held.
New rw-sem is used to protect genl callback from genl family
Following patch uses of lockless genl callbacks for ovs.
Signed-off-by: Pravin B Shelar pshe...@nicira.com
---
net/openvswitch/datapath.c | 12
1 files changed, 8 insertions(+), 4 deletions(-)
diff --git a/net/openvswitch/datapath.c b/net/openvswitch/datapath.c
index
On Thu, Apr 18, 2013 at 09:46:30AM -0700, Justin Pettit wrote:
On Apr 18, 2013, at 9:28 AM, Ben Pfaff b...@nicira.com wrote:
But I'm not sure about the test. It tries to connect to the default
OpenFlow port on localhost. I guess that there will ordinarily be no
OpenFlow controller
On Wed, Apr 17, 2013 at 05:35:04PM -0700, Alex Wang wrote:
The commit 89d7ffa9 (python: Workaround UNIX socket path
length limits), fixes most failed tests. But it has a
typo and the typo causes the failure of test unixctl
server errors - Python when the path length is very
long (e.g. more
I forgot that I needed to push this myself.
I pushed this to master.
On Mon, Apr 15, 2013 at 06:19:56PM -0700, Ben Pfaff wrote:
I misunderstood. Thanks, I'm happy with it as is then.
On Apr 15, 2013 5:17 PM, Andy Zhou az...@nicira.com wrote:
This patch does the way you recommended already.
I snipped most of your responses, which look fine to me. Thanks.
I see one place worth a more detailed response.
On Wed, Apr 17, 2013 at 07:00:17PM -0700, Romain Lenglet wrote:
I wasn't sure about the policy regarding asserts, since there aren't
many in the code.
I have mixed feelings about
On 4/17/13 8:03 PM, Ben Pfaff wrote:
+ovs-ofctl add-flows br0 - 'EOF'
+table=4 reg0=0 priority=99 dl_vlan=20 actions=1,strip_vlan,2
+table=4 reg0=0 priority=99 dl_vlan=30 actions=1,strip_vlan,3,4
+table=4 reg0=0 priority=50actions=1
+EOF
Wouldn't the last flow (table=4
On Thu, Apr 18, 2013 at 02:42:10PM -0300, Luiz Ozaki wrote:
On 4/17/13 8:03 PM, Ben Pfaff wrote:
+ovs-ofctl add-flows br0 - 'EOF'
+table=4 reg0=0 priority=99 dl_vlan=20 actions=1,strip_vlan,2
+table=4 reg0=0 priority=99 dl_vlan=30 actions=1,strip_vlan,3,4
+table=4 reg0=0
Thanks, I applied this to master.
On Thu, Apr 18, 2013 at 07:26:35PM +0900, Hiroshi Miyata wrote:
Hi Ben,
I confirmed it works well!
Thanks so much!
...hiroshi miyata
(13/04/17 9:04), miyahiro.d...@gmail.com wrote:
Hi Ben,
Thanks, I will try the patch!
I will sent you the
On Thu, Apr 18, 2013 at 01:08:40PM +0100, Ian Campbell wrote:
On Thu, 2013-04-18 at 12:50 +0100, Ian Campbell wrote:
Based on Waldi's RFC at
http://lists.xen.org/archives/html/xen-devel/2012-09/msg00943.html
To use it set vif.default.script=vif-openvswitch in /etc/xen/xl.conf or
use
On Apr 18, 2013, at 9:56 AM, Ben Pfaff b...@nicira.com wrote:
On Thu, Apr 18, 2013 at 09:46:30AM -0700, Justin Pettit wrote:
On Apr 18, 2013, at 9:28 AM, Ben Pfaff b...@nicira.com wrote:
But I'm not sure about the test. It tries to connect to the default
OpenFlow port on localhost. I
Looks good.
--Justin
On Apr 3, 2013, at 12:02 PM, Ben Pfaff b...@nicira.com wrote:
An upcoming patch will start using the --log-file option with these test
programs, so they need to support it.
Signed-off-by: Ben Pfaff b...@nicira.com
---
tests/test-netflow.c | 12 +---
On Apr 3, 2013, at 12:02 PM, Ben Pfaff b...@nicira.com wrote:
port chosen by the user (he Open vSwitch automated tests currently do this)
s/he/the/
Otherwise, looks good.
--Justin
___
dev mailing list
dev@openvswitch.org
Thanks.
On Thu, Apr 18, 2013 at 10:09 AM, Ben Pfaff b...@nicira.com wrote:
I forgot that I needed to push this myself.
I pushed this to master.
On Mon, Apr 15, 2013 at 06:19:56PM -0700, Ben Pfaff wrote:
I misunderstood. Thanks, I'm happy with it as is then.
On Apr 15, 2013 5:17 PM,
On Thu, Apr 18, 2013 at 9:52 AM, Pravin B Shelar pshe...@nicira.com wrote:
Following patch uses of lockless genl callbacks for ovs.
Signed-off-by: Pravin B Shelar pshe...@nicira.com
I think this is clearly the right thing to do (assuming that the genl
part passes review from other people).
On Apr 3, 2013, at 12:02 PM, Ben Pfaff b...@nicira.com wrote:
diff --git a/tests/ofproto-dpif.at b/tests/ofproto-dpif.at
index 06ebf23..857e6ab 100644
--- a/tests/ofproto-dpif.at
+++ b/tests/ofproto-dpif.at
@@ -1208,10 +1208,12 @@ AT_CLEANUP
dnl Test that sFlow samples packets correctly.
On Thu, Apr 18, 2013 at 11:31 AM, Jesse Gross je...@nicira.com wrote:
On Thu, Apr 18, 2013 at 9:52 AM, Pravin B Shelar pshe...@nicira.com wrote:
Following patch uses of lockless genl callbacks for ovs.
Signed-off-by: Pravin B Shelar pshe...@nicira.com
I think this is clearly the right thing
OVS locking was recently changed to have private OVS lock which
simplified overall locking. Therefore there is no need to have
another global genl lock to protect OVS data structures. Following
patch uses of lockless genl family for OVS. This also allows
more granual OVS locking using ovs_mutex
From: Pravin B Shelar pshe...@nicira.com
Date: Thu, 18 Apr 2013 14:16:52 -0700
OVS locking was recently changed to have private OVS lock which
simplified overall locking. Therefore there is no need to have
another global genl lock to protect OVS data structures. Following
patch uses of
All genl callbacks are serialized by genl-mutex. This can become
bottleneck in multi threaded case.
Following patch adds an parameter to genl_family so that a
particular family can get concurrent netlink callback without
genl_lock held.
New rw-sem is used to protect genl callback from genl family
On Thu, Apr 18, 2013 at 2:18 PM, David Miller da...@davemloft.net wrote:
From: Pravin B Shelar pshe...@nicira.com
Date: Thu, 18 Apr 2013 14:16:52 -0700
OVS locking was recently changed to have private OVS lock which
simplified overall locking. Therefore there is no need to have
another
All genl callbacks are serialized by genl-mutex. This can become
bottleneck in multi threaded case.
Following patch adds an parameter to genl_family so that a
particular family can get concurrent netlink callback without
genl_lock held.
New rw-sem is used to protect genl callback from genl family
Thanks,
Pushed to master.
On Mon, Apr 15, 2013 at 2:36 PM, Jesse Gross je...@nicira.com wrote:
On Mon, Apr 15, 2013 at 11:32 AM, Pravin B Shelar pshe...@nicira.com wrote:
Signed-off-by: Pravin B Shelar pshe...@nicira.com
Acked-by: Jesse Gross je...@nicira.com
OVS locking was recently changed to have private OVS lock which
simplified overall locking. Therefore there is no need to have
another global genl lock to protect OVS data structures. Following
patch uses of lockless genl family for OVS. This also allows
more granual OVS locking using ovs_mutex
There is a inconsistent ordering in function ovs_vport_cmd_set()
between upstream and out of tree ovs module. Following patch
fixes it by releasing lock before calling ovs_notify.
Signed-off-by: Pravin B Shelar pshe...@nicira.com
---
datapath/datapath.c |3 +--
1 files changed, 1
On Thu, Apr 18, 2013 at 4:02 PM, Pravin B Shelar pshe...@nicira.com wrote:
There is a inconsistent ordering in function ovs_vport_cmd_set()
between upstream and out of tree ovs module. Following patch
fixes it by releasing lock before calling ovs_notify.
Signed-off-by: Pravin B Shelar
On Thu, Apr 18, 2013 at 4:07 PM, Pravin B Shelar pshe...@nicira.com wrote:
Removes remaining part of comment about `linkname` vport member.
Signed-off-by: Pravin B Shelar pshe...@nicira.com
Acked-by: Jesse Gross je...@nicira.com
___
dev mailing list
On Thu, Apr 18, 2013 at 11:53:12AM -0700, Justin Pettit wrote:
On Apr 3, 2013, at 12:02 PM, Ben Pfaff b...@nicira.com wrote:
diff --git a/tests/ofproto-dpif.at b/tests/ofproto-dpif.at
index 06ebf23..857e6ab 100644
--- a/tests/ofproto-dpif.at
+++ b/tests/ofproto-dpif.at
@@ -1208,10
Following patch series is factored out from tunnel upstream
series. This simplifies upstreaming patch series.
Pravin B Shelar (3):
datapath: Use generic struct pcpu_tstats.
tunneling: Remove struct tnl_vport and tnl_ops.
datapath: Move vport init to First port create.
After flow based tunneling, kernel tunneling is greatly simplified.
There is no need to have extra tunning layer between vport and
perticular protocol.
Following patch removes tunneling struct which make code easy to read.
Signed-off-by: Pravin B Shelar pshe...@nicira.com
---
datapath/tunnel.c
Rather than defining ovs specific stats struct (vport_percpu_stats),
we can use existing pcpu_tstats to achieve exactly same functionality.
Signed-off-by: Pravin B Shelar pshe...@nicira.com
---
datapath/linux/Modules.mk |1 +
vport-init and exit() functios are defined by gre and netdev vport
only and both can be moved to first port create.
Following patch does same, it moves vport init to respectve vport
create and get rid of vport-init() and vport-exit() fnctions.
Signed-off-by: Pravin B Shelar pshe...@nicira.com
On Thu, Apr 18, 2013 at 2:30 PM, Pravin B Shelar pshe...@nicira.com wrote:
OVS locking was recently changed to have private OVS lock which
simplified overall locking. Therefore there is no need to have
another global genl lock to protect OVS data structures. Following
patch uses of lockless
44 matches
Mail list logo