Thanks, I pushed this to master.
On 3 July 2014 08:42, Ben Pfaff wrote:
> On Tue, Jul 01, 2014 at 01:54:18PM +1200, Joe Stringer wrote:
> > Commit acaa8dac49 (revalidator: Eliminate duplicate flow handling.)
> > ensured that a ukey will always exist for a given flow, even if it is
> > about to
On 3 July 2014 10:30, Ben Pfaff wrote:
> On Wed, Jul 02, 2014 at 08:14:15PM +1200, Joe Stringer wrote:
> > If the datapath doesn't dump a flow for some reason, and the current
> > dump is expected to revalidate all flows in the datapath, then perform
> > revalidation for those flows by fetching t
I appreciate the expedient review. This series doesn't actually depend on
the batched flow_get, but this was a starting point/RFC for discussion.
On 3 July 2014 09:24, Ben Pfaff wrote:
> On Wed, Jul 02, 2014 at 08:14:14PM +1200, Joe Stringer wrote:
> > This patch adds support for batching flow_g
The handle_missed_revalidation() call is inverse from what I had intended:
diff --git a/ofproto/ofproto-dpif-upcall.c b/ofproto/ofproto-dpif-upcall.c
index 9c2a8a4..70f9a43 100644
--- a/ofproto/ofproto-dpif-upcall.c
+++ b/ofproto/ofproto-dpif-upcall.c
@@ -1583,7 +1583,7 @@ revalidator_sweep__(stru
Woops, I forgot to indicate that this series is for branch-2.3.
On 3 July 2014 12:29, Joe Stringer wrote:
> Change the interface to allow implementations to pass back a buffer, and
> allow callers to specify which of actions, mask, and stats they wish to
> receive. This will be used in the next
On Wed, Jul 2, 2014 at 4:46 PM, Thomas Graf wrote:
> On 07/01/14 at 05:43pm, Jesse Gross wrote:
>> There are many possible ways that a flow can be invalid so we've
>> added logging for most of them. This adds logs for the remaining
>> possible cases so there isn't any ambiguity while debugging.
>>
If the datapath doesn't dump a flow for some reason, and the current
dump is expected to revalidate all flows in the datapath, then perform
revalidation for those flows by fetching them during the sweep phase.
If revalidation is not required, then leave the flow in the datapath and
don't revalidate
Change the interface to allow implementations to pass back a buffer, and
allow callers to specify which of actions, mask, and stats they wish to
receive. This will be used in the next commit.
Signed-off-by: Joe Stringer
---
lib/dpif-linux.c| 23 +--
lib/dpif-netdev.c
On Wed, Jul 2, 2014 at 4:24 PM, Pritesh Kothari (pritkoth)
wrote:
> Thanks for the review.
>
>>> + OVS_GREP_IFELSE([$KSRC/include/linux/if.h], [IFF_LIVE_ADDR_CHANGE])
>>> +
>>> OVS_GREP_IFELSE([$KSRC/include/linux/if_vlan.h], [vlan_set_encap_proto])
>>>
>> Why are you changing existing compat c
On 07/01/14 at 05:43pm, Jesse Gross wrote:
> There are many possible ways that a flow can be invalid so we've
> added logging for most of them. This adds logs for the remaining
> possible cases so there isn't any ambiguity while debugging.
>
> CC: Federico Iezzi
> Signed-off-by: Jesse Gross
LGT
On 07/01/14 at 05:43pm, Jesse Gross wrote:
> There are many possible ways that a flow can be invalid so we've
> added logging for most of them. This adds logs for the remaining
> possible cases so there isn't any ambiguity while debugging.
>
> CC: Federico Iezzi
> Signed-off-by: Jesse Gross
It'
On Wed, Jul 2, 2014 at 3:40 PM, Ben Pfaff wrote:
> On Wed, Jul 02, 2014 at 02:21:58PM -0700, Pravin B Shelar wrote:
>> Following patch enables all available tunnel GSO features for OVS
>> bridge device so that ovs can use hardware offloads available to
>> underling device.
>>
>> Signed-off-by: Pra
Thanks for the discussion. Given the time constraints, I will prepare a
fresh version for branch-2.3 first, then address the issues on master
separately.
On 3 July 2014 08:56, Ben Pfaff wrote:
> On Wed, Jul 02, 2014 at 08:14:13PM +1200, Joe Stringer wrote:
> > There is no guarantee that perform
Thanks for the review.
>> + OVS_GREP_IFELSE([$KSRC/include/linux/if.h], [IFF_LIVE_ADDR_CHANGE])
>> +
>> OVS_GREP_IFELSE([$KSRC/include/linux/if_vlan.h], [vlan_set_encap_proto])
>>
> Why are you changing existing compat code in this patch? Even if you
> have a reason for such change please send
On Tue, Jul 01, 2014 at 05:43:39PM -0700, Jesse Gross wrote:
> There are many possible ways that a flow can be invalid so we've
> added logging for most of them. This adds logs for the remaining
> possible cases so there isn't any ambiguity while debugging.
>
> CC: Federico Iezzi
> Signed-off-by:
On Wed, Jul 02, 2014 at 02:21:58PM -0700, Pravin B Shelar wrote:
> Following patch enables all available tunnel GSO features for OVS
> bridge device so that ovs can use hardware offloads available to
> underling device.
>
> Signed-off-by: Pravin B Shelar
> ---
> acinclude.m4
On Wed, Jul 2, 2014 at 11:08 AM, Pritesh Kothari
wrote:
> Signed-off-by: Pritesh Kothari
> ---
> v1: make check passes for 3.2/3.13/3.14/3.15
> ---
> FAQ | 1 +
> NEWS| 2 +-
> acinclude.m4
On Wed, Jul 02, 2014 at 08:14:15PM +1200, Joe Stringer wrote:
> If the datapath doesn't dump a flow for some reason, and the current
> dump is expected to revalidate all flows in the datapath, then perform
> revalidation for those flows by fetching them during the sweep phase.
>
> Signed-off-by: J
This fixes valgrind errors.
Signed-off-by: Ryan Wilson
---
ovsdb/ovsdb-server.c | 25 -
1 file changed, 16 insertions(+), 9 deletions(-)
diff --git a/ovsdb/ovsdb-server.c b/ovsdb/ovsdb-server.c
index a85a672..0a1be0f 100644
--- a/ovsdb/ovsdb-server.c
+++ b/ovsdb/ovsdb-
On Wed, Jul 02, 2014 at 08:14:14PM +1200, Joe Stringer wrote:
> This patch adds support for batching flow_get operations to the dpif. As
> part of this, it also now allows fetching the mask (this was previously
> not possible using a 'get' operation).
>
> Signed-off-by: Joe Stringer
This approac
Following patch enables all available tunnel GSO features for OVS
bridge device so that ovs can use hardware offloads available to
underling device.
Signed-off-by: Pravin B Shelar
---
acinclude.m4 | 2 ++
.../linux/compat/include/linux/netdev_features.h |
On Wed, Jul 02, 2014 at 08:14:13PM +1200, Joe Stringer wrote:
> There is no guarantee that performing a flow_dump will return a complete
> version of the datapath's flow table. Occasionally, a flow may be
> completely missed even if it is present in the datapath (particularly if
> the datapath deci
On Tue, Jul 01, 2014 at 01:54:18PM +1200, Joe Stringer wrote:
> Commit acaa8dac49 (revalidator: Eliminate duplicate flow handling.)
> ensured that a ukey will always exist for a given flow, even if it is
> about to be deleted. This means that push_dump_ops__() no longer needs
> to handle the case w
On Mon, Jun 30, 2014 at 02:59:45PM -0700, Ben Pfaff wrote:
> The upstream kernel net/netlink/af_netlink.c netlink_recvmsg() contains the
> following code to refill the Netlink socket buffer with more dump skbs
> while a dump is in progress:
>
> if (nlk->cb && atomic_read(&sk->sk_rmem_alloc)
On Wed, Jul 02, 2014 at 01:07:14PM -0700, Jesse Gross wrote:
> On Wed, Jul 2, 2014 at 11:53 AM, Ben Pfaff wrote:
> > On Tue, Jul 01, 2014 at 05:52:23PM -0700, Jesse Gross wrote:
> >> On Tue, Jul 1, 2014 at 5:10 PM, Ben Pfaff wrote:
> >> > On Tue, Jul 01, 2014 at 05:00:45PM -0700, Jesse Gross wrot
On Wed, Jul 2, 2014 at 11:53 AM, Ben Pfaff wrote:
> On Tue, Jul 01, 2014 at 05:52:23PM -0700, Jesse Gross wrote:
>> On Tue, Jul 1, 2014 at 5:10 PM, Ben Pfaff wrote:
>> > On Tue, Jul 01, 2014 at 05:00:45PM -0700, Jesse Gross wrote:
>> >> On Tue, Jul 1, 2014 at 3:49 PM, Ben Pfaff wrote:
>> >> > I
Applied, thanks!
On Wed, Jul 02, 2014 at 07:43:08PM +, Pritesh Kothari (pritkoth) wrote:
> Acked-by: Pritesh Kothari
>
> On Jul 2, 2014, at 11:47 AM, Flavio Leitner wrote:
>
> > Signed-off-by: Flavio Leitner
> > ---
> > INSTALL.DPDK | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
On Mon, Jun 30, 2014 at 06:03:57PM -0300, Flavio Leitner wrote:
> On Mon, Jun 30, 2014 at 11:59:31AM -0700, Ben Pfaff wrote:
> > If this seems like an improvement, I can do the same thing for the RHEL
> > version.
> >
> > CC: Flavio Leitner
> > Signed-off-by: Ben Pfaff
>
> I think it's more cle
Acked-by: Pritesh Kothari
On Jul 2, 2014, at 11:47 AM, Flavio Leitner wrote:
> Signed-off-by: Flavio Leitner
> ---
> INSTALL.DPDK | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/INSTALL.DPDK b/INSTALL.DPDK
> index 2a6d7ef..36bbe54 100644
> --- a/INSTALL.DPDK
> +++ b/IN
On Tue, Jul 01, 2014 at 05:52:23PM -0700, Jesse Gross wrote:
> On Tue, Jul 1, 2014 at 5:10 PM, Ben Pfaff wrote:
> > On Tue, Jul 01, 2014 at 05:00:45PM -0700, Jesse Gross wrote:
> >> On Tue, Jul 1, 2014 at 3:49 PM, Ben Pfaff wrote:
> >> > I found when reconfiguring my kernel that if I turned on NE
Signed-off-by: Flavio Leitner
---
INSTALL.DPDK | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/INSTALL.DPDK b/INSTALL.DPDK
index 2a6d7ef..36bbe54 100644
--- a/INSTALL.DPDK
+++ b/INSTALL.DPDK
@@ -91,7 +91,7 @@ Start ovsdb-server as discussed in INSTALL doc:
start ovsdb-ser
Hey Joe,
Awesome work for the two solutions.
Few comments:
1. For this patch, could we make it for branch-2.3 only?
On master, I'm still seeing dp flow deleted only with a single ping~!!!
(without flow table modification & hping3)
So, there must be something wrong else.
2. After brief discussio
Signed-off-by: Pritesh Kothari
---
v1: make check passes for 3.2/3.13/3.14/3.15
---
FAQ | 1 +
NEWS| 2 +-
acinclude.m4| 17 -
datapath/linux/Modules.mk
Trying to use debian wheezy openvswitch with xen.
I think the real problem is that openvswitch-switch is being started after
networking but it needs to start before networking.
I put this in /etc/network/interfaces:
# XEN Bridge
auto xenbr0
iface xenbr0 inet dhcp
pre-up service openvswitch-sw
Hi Zoli,
On 01/07/14 20:52, Zoltan Kiss wrote:
This patch modifies the LACP selection logic by prefering a slaves with up and
running partners when looking for a lead.
That fixes the following scenario:
- bond has 2 ports, A and B, their other ends are in separate chassis with
MC-LAG sync
- t
If the datapath doesn't dump a flow for some reason, and the current
dump is expected to revalidate all flows in the datapath, then perform
revalidation for those flows by fetching them during the sweep phase.
Signed-off-by: Joe Stringer
---
This relies on the mask dumping from the previous patch
There is no guarantee that performing a flow_dump will return a complete
version of the datapath's flow table. Occasionally, a flow may be
completely missed even if it is present in the datapath (particularly if
the datapath decides to rehash its flow tables). Previously we would
delete such "misse
This patch adds support for batching flow_get operations to the dpif. As
part of this, it also now allows fetching the mask (this was previously
not possible using a 'get' operation).
Signed-off-by: Joe Stringer
---
lib/dpif-linux.c| 70 +++-
lib/dpi
There are a few corner cases where datapaths may fail to dump flows to
userspace. If userspace had seen the flow before, then rather than attempt to
handle the flow, it would simply delete it. This is the simplest, easiest way
to tackle the problem, but has the downside that if the flow is importan
39 matches
Mail list logo
