Thanks Numan for the review. Please see my response below.
On Tue, Apr 2, 2019 at 3:59 AM Numan Siddique wrote:
>
>
>
> On Mon, Apr 1, 2019 at 1:15 AM Han Zhou wrote:
>>
>> From: Han Zhou
>>
>> In the main loop, if the SB DB is disconnected when there is a pending
>> transaction, there can be
Hello OVS team,
This issue is observed in ovs-master with dpdk-18.11.
Please find the steps to reproduce.
root@ubuntu16:/home/sdn# ovs-vswitchd --version
ovs-vswitchd (Open vSwitch) 2.11.90
DPDK 18.11.0
root@ubuntu16:/home/sdn# apport-retrace -g
Cursos TOP - Webinar Interactivo – Martes 09 de Abril
¿Cómo vender por Instagram?
No es sólo una red social de compartir imágenes sino de generar una conexión
con tu comunidad y con el cliente, permitiendo
ganar más relevancia y confianza con tu público-objetivo. En este webinar
aprenderás
It took a while, but I finally got this reviewed. Looks good to me,
Acked-by: Mark Michelson
On 3/29/19 1:53 PM, nusid...@redhat.com wrote:
From: Numan Siddique
This is an RFC series to address the MTU issues for OVN reported
here [1].
To address this issue, a new OVS action -
From: Dan Carpenter
Date: Tue, 2 Apr 2019 09:53:14 +0300
> We free "ct_info->ct" and then use it on the next line when we pass it
> to nf_ct_destroy_timeout(). This patch swaps the order to avoid the use
> after free.
>
> Fixes: 06bd2bdf19d2 ("openvswitch: Add timeout support to ct action")
>
On Tue, Apr 2, 2019 at 4:05 PM Roi Dayan wrote:
>
>
>
> On 02/04/2019 12:27, John Hurley wrote:
> > The TC datapath only permits the offload of mirr actions if they are
> > egress. To offload TC actions that output to OvS internal ports, ingress
> > mirr actions are required. At the TC layer, an
On Tue, Apr 2, 2019 at 4:20 PM Roi Dayan wrote:
>
>
>
> On 02/04/2019 12:27, John Hurley wrote:
> > Offloading rules to a TC datapath only allows the creating of ingress hook
> > qdiscs and the application of filters to these. However, there may be
> > certain situations where an egress qdisc is
On Tue, Apr 2, 2019 at 3:53 PM Roi Dayan wrote:
>
>
>
> On 02/04/2019 12:27, John Hurley wrote:
> > OvS includes compat code for serval TC actions including vlan, mirr and
>
> typo serval/several
>
> in mirr you mean mirred ? i would prefer mirred.
>
good spot. Will change. Thanks
> > tunnel
On Mon, Apr 1, 2019 at 11:53 PM Dan Carpenter wrote:
>
> We free "ct_info->ct" and then use it on the next line when we pass it
> to nf_ct_destroy_timeout(). This patch swaps the order to avoid the use
> after free.
>
> Fixes: 06bd2bdf19d2 ("openvswitch: Add timeout support to ct action")
>
On 02/04/2019 12:27, John Hurley wrote:
> Offloading rules to a TC datapath only allows the creating of ingress hook
> qdiscs and the application of filters to these. However, there may be
> certain situations where an egress qdisc is more applicable (e.g. when
> offloading to TC rules applied
On 02/04/2019 12:27, John Hurley wrote:
> The TC datapath only permits the offload of mirr actions if they are
> egress. To offload TC actions that output to OvS internal ports, ingress
> mirr actions are required. At the TC layer, an ingress mirr action passes
> the packet back into the
Adding the elaboration on each patch. I'll add it in V2 cover-letter if needed.
On 4/2/2019 5:57 PM, Eli Britstein wrote:
Current HW offloading solutions use embedded switches to offload OVS DP
rules while using SRIOV pass-through interfaces to the guest VMs. This
architecture requires the VM
Introduce peer name property for dpdk ports, as a pre-step towards
accelarating peer port forwarding.
Signed-off-by: Eli Britstein
Reviewed-by: Oz Shlomo
---
lib/netdev-dpdk.c | 55 ---
1 file changed, 52 insertions(+), 3 deletions(-)
diff --git
Packet parsing and rule matching introduce a large processing overhead
for simple port-to-port forwarding scenarios. Accelerate port-to-port
forwarding by bypassing the datapath packet processing if a peer-port
property is defined.
Signed-off-by: Eli Britstein
Reviewed-by: Oz Shlomo
---
Use peer port as a generic netdev property instead of a specific
patch-port property.
Signed-off-by: Eli Britstein
Reviewed-by: Oz Shlomo
---
ofproto/ofproto-dpif.c | 6 +-
1 file changed, 1 insertion(+), 5 deletions(-)
diff --git a/ofproto/ofproto-dpif.c b/ofproto/ofproto-dpif.c
index
Replace specific patch port API for retreiving peer port name with the
generic netdev API.
Signed-off-by: Eli Britstein
Reviewed-by: Oz Shlomo
---
lib/netdev-vport.c | 40
lib/netdev-vport.h | 2 --
ofproto/ofproto-dpif.c | 6 +++---
3 files
Currently peer port is defined as a patch port property. However, peer
ports can be used as generic netdevice property to accelerate peer ports
forwarding. Introduce a generic method for retreiving peer port name.
Signed-off-by: Eli Britstein
Reviewed-by: Oz Shlomo
---
lib/netdev-provider.h |
Current HW offloading solutions use embedded switches to offload OVS DP
rules while using SRIOV pass-through interfaces to the guest VMs. This
architecture requires the VM guests to install vendor specific drivers
and is challenging for live-migration. Such caveats may force some users
to fall
On 02/04/2019 12:27, John Hurley wrote:
> OvS includes compat code for serval TC actions including vlan, mirr and
typo serval/several
in mirr you mean mirred ? i would prefer mirred.
> tunnel key. Add support for using skbedit actions when compiling
> user-space code against older kernel
Dear Customer,
Thank you for banking with Standard Chartered Bank
Attached is your Debit Advice as per your request. Kindly note
that your e-Advice is in PDF format.
If you are unable to open the attachment please download Adobe
Acrobat Reader by clicking at:
www.adobe.com
The PDF file is
+d...@openvswitch.org 'ovs-disc...@openvswitch.org'
Hi Ilya & Ben
Thanks for your reply and suggestion. The following is my patch to fix this
bug(refer to Ilya's double postponing method).
It is useful via our fault-injection test. Could you help to check if it
will cause other
On Mon, Apr 1, 2019 at 1:15 AM Han Zhou wrote:
> From: Han Zhou
>
> In the main loop, if the SB DB is disconnected when there is a pending
> transaction, there can be busy loop causing 100% CPU of ovn-controller,
> until SB DB is connected again.
>
> The root cause is that when a transaction is
Rules applied to OvS internal ports are not represented in TC datapaths.
However, it is possible to support rules matching on internal ports in TC.
The start_xmit ndo of OvS internal ports directs packets back into the OvS
kernel datapath where they are rematched with the ingress port now being
Offloading rules to a TC datapath only allows the creating of ingress hook
qdiscs and the application of filters to these. However, there may be
certain situations where an egress qdisc is more applicable (e.g. when
offloading to TC rules applied to OvS internal ports).
Extend the TC API in OvS
The TC datapath only permits the offload of mirr actions if they are
egress. To offload TC actions that output to OvS internal ports, ingress
mirr actions are required. At the TC layer, an ingress mirr action passes
the packet back into the network stack as if it came in the action port
rather
OvS includes compat code for serval TC actions including vlan, mirr and
tunnel key. Add support for using skbedit actions when compiling
user-space code against older kernel headers.
Signed-off-by: John Hurley
Reviewed-by: Simon Horman
---
acinclude.m4 | 7 +
Common use-cases in OvS can produce datapath rules that match on OvS
internal ports. For example, when the endpoint IP address of a VXLAN
tunnel is on the OvS bridge itself, datapath rules may take the form:
1. in_port(eth1),actions:set(tunnel(...)),vxlan_sys_4789
2.
vport offloaded functions should have a different implementation for
linux based OVS versus dpdk based OVS. Currently there is only support
for linux based offloaded API. The code in the file named netdev-vport.c
checks 'ifdef __linux__' to select the linux based offloaded functions,
without
This patch series look good to me.
Thanks Greg for the work.
Reviewed-by: Yifeng Sun
On Wed, Mar 27, 2019 at 10:45 AM Gregory Rose wrote:
>
> On 3/27/2019 9:16 AM, 0-day Robot wrote:
> > Bleep bloop. Greetings Gregory Rose, I am a robot and I have tried out
> > your patch.
> > Thanks for
We free "ct_info->ct" and then use it on the next line when we pass it
to nf_ct_destroy_timeout(). This patch swaps the order to avoid the use
after free.
Fixes: 06bd2bdf19d2 ("openvswitch: Add timeout support to ct action")
Signed-off-by: Dan Carpenter
---
net/openvswitch/conntrack.c | 2 +-
30 matches
Mail list logo
