Hi Guys,
Can you please review the patch? It has been in mailing list for long time.
Regards,
Nitin
> -Original Message-
> From: Nitin Katiyar
> Sent: Friday, October 04, 2019 11:06 AM
> To: 'ovs-dev@openvswitch.org'
> Cc: Anju Thomas ; 'Ilya Maximets'
> ; Ben Pfaff
> Subject: RE:
Thanks Ilya and kevin.
Regards,
Sriram.
-Original Message-
From: Ilya Maximets
Sent: 12 November 2019 00:08
To: Ilya Maximets ; Kevin Traynor ;
Sriram Vatala ; ovs-dev@openvswitch.org
Subject: Re: [PATCH v11 2/3] netdev-dpdk : Detailed packet drop statistics
On 11.11.2019 17:11, Ilya
> On Nov 12, 2019, at 12:35 PM, Pravin Shelar wrote:
>
>> On Sun, Nov 10, 2019 at 3:44 AM wrote:
>>
>> From: Tonghao Zhang
>>
>> When using the kernel datapath, the upcall don't
>> add skb hash info relatived. That will introduce
>> some problem, because the hash of skb is very
>>
On Sun, Nov 10, 2019 at 3:44 AM wrote:
>
> From: Tonghao Zhang
>
> When using the kernel datapath, the upcall don't
> add skb hash info relatived. That will introduce
> some problem, because the hash of skb is very
> important (e.g. vxlan module uses it for udp src port,
> tx queue selection on
On Mon, Nov 11, 2019 at 11:03 AM Dumitru Ceara wrote:
>
> On Mon, Nov 11, 2019 at 6:11 PM Han Zhou wrote:
> >
> > 1. Would there be problem for VLAN backed logical router use case,
since a chassis MAC is used as src MAC to send packets from router ports?
> > 2. How about checking if tpa == spa
21 de Noviembre | Horario de 10:00 a 17:00 hrs. | (hora del centro de México)
- Mapeo de Procesos - Webinar en Vivo
Nuestro webinar, te ofrece aplicar una metodología práctica y ordenada para el
mapeo de
procesos que puede ser aplicable a cualquier tipo de organización. Este taller
está
From: xiangxia.m@gmail.com
Date: Sun, 10 Nov 2019 19:54:04 +0800
> From: Tonghao Zhang
>
> When using the kernel datapath, the upcall don't
> add skb hash info relatived. That will introduce
> some problem, because the hash of skb is very
> important (e.g. vxlan module uses it for udp src
On 10/1/2019 3:10 PM, Eelco Chaudron wrote:
This patch adds support for multi-queue QoS to the DPDK datapath. Most of
the code is based on an earlier patch from a patchset sent out by
zhaozhanxu. The patch was titled "[ovs-dev, v2, 1/4] netdev-dpdk.c: Support
the multi-queue QoS configuration
On Mon, Nov 11, 2019 at 6:11 PM Han Zhou wrote:
>
> 1. Would there be problem for VLAN backed logical router use case, since a
> chassis MAC is used as src MAC to send packets from router ports?
> 2. How about checking if tpa == spa to make sure GARP is always flooded? (not
> directly supported
Hi Dumitru. Everything you've done looks good to me.
However, when reviewing the patch, I educated myself on how ovn-detrace
is implemented and I found a couple of problems. These problems aren't
introduced by you. However, I think that since you have broadened the
scope of the southbound
The conventional way for packet dumping in OVS is to use ovs-tcpdump
that works via traffic mirroring. DPDK pdump could probably be used
for some lower level debugging, but it is not commonly used for
various reasons.
There are lots of limitations for using this functionality in practice.
Most
On 11.11.2019 17:11, Ilya Maximets wrote:
>>> I'm not sure if clang annotations will work with rte_spinlock.
>>> DPDK doesn't have proper annotations for locking functions.
>>>
>>
>> Ah, good point, I didn't check the lock type. In that case nevermind,
>> patch+incremental LGTM as is.
>>
>>
On Thu, Nov 07, 2019 at 01:39:28PM +0100, Eelco Chaudron wrote:
>
>
> On 7 Nov 2019, at 12:36, Ilya Maximets wrote:
>
> >Until now there was only two options for XDP mode in OVS: SKB or DRV.
> >i.e. 'generic XDP' or 'native XDP with zero-copy enabled'.
> >
> >Devices like 'veth' interfaces in
Sorry, I missed the 'dpdk-latest' tag.
On 11.11.2019 16:01, Ilya Maximets wrote:
> DPDK commit 660098d61f57 ("pdump: use generic multi-process channel")
> switched pdump to use generic DPDK IPC instead of sockets.
> Old API was deprecated and removed. Updating OVS code accordingly.
>
>
1. Would there be problem for VLAN backed logical router use case, since a
chassis MAC is used as src MAC to send packets from router ports?
2. How about checking if tpa == spa to make sure GARP is always flooded?
(not directly supported by OF)
On Mon, Nov 11, 2019 at 5:32 AM Dumitru Ceara
On 2019-11-11 4:08 PM, Roi Dayan wrote:
>
>
> On 2019-11-11 3:39 PM, Simon Horman wrote:
>> On Mon, Nov 04, 2019 at 11:49:44AM +0200, Roi Dayan wrote:
>>> In kernels older than 4.8, struct tcf_t didn't have the firstuse.
>>> If openvswitch is compiled with the compat pkt_cls.h then there is
Hi Zhike
Thanks for clarifying
There is presently no support for multi-segment mbufs in OVS, so the patch
would
not be needed and in general patches are only proposed at the time they are
relevant.
Also note that in the hypothetical case of multisegment mbufs, the packet
would be linearized at
On 11.11.2019 17:06, Kevin Traynor wrote:
> On 11/11/2019 15:59, Ilya Maximets wrote:
>> On 11.11.2019 16:55, Kevin Traynor wrote:
>>> On 10/11/2019 23:20, Ilya Maximets wrote:
On 29.10.2019 15:50, Sriram Vatala wrote:> @@ -2388,12 +2412,16 @@
__netdev_dpdk_vhost_send(struct netdev
On 11/11/2019 15:59, Ilya Maximets wrote:
> On 11.11.2019 16:55, Kevin Traynor wrote:
>> On 10/11/2019 23:20, Ilya Maximets wrote:
>>> On 29.10.2019 15:50, Sriram Vatala wrote:> @@ -2388,12 +2412,16 @@
>>> __netdev_dpdk_vhost_send(struct netdev *netdev, int qid,
}
}
On Mon, Nov 04, 2019 at 06:34:49PM +0200, Roi Dayan wrote:
> From: Vlad Buslov
>
> Recent changes in Linux kernel TC action subsystem introduced new
> TCA_ACT_FLAGS_NO_PERCPU_STATS flag. The purpose of the flag is to request
> action implementation to skip allocating action stats with expensive
On 11.11.2019 16:55, Kevin Traynor wrote:
> On 10/11/2019 23:20, Ilya Maximets wrote:
>> On 29.10.2019 15:50, Sriram Vatala wrote:> @@ -2388,12 +2412,16 @@
>> __netdev_dpdk_vhost_send(struct netdev *netdev, int qid,
>>> }
>>> } while (cnt && (retries++ < max_retries));
>>>
>>>
On 10/11/2019 23:20, Ilya Maximets wrote:
> On 29.10.2019 15:50, Sriram Vatala wrote:> @@ -2388,12 +2412,16 @@
> __netdev_dpdk_vhost_send(struct netdev *netdev, int qid,
>> }
>> } while (cnt && (retries++ < max_retries));
>>
>> +tx_failure = cnt;
>>
On 11 Nov 2019, at 16:15, Ilya Maximets wrote:
This patch doesn't apply to master, but it applies to dpdk-latest.
dpdk-latest needs some rebase and this patch should be rebased
after that.
Yes I know, any idea when this will be? Or how we will handle these DPDK
related features?
Will
This patch doesn't apply to master, but it applies to dpdk-latest.
dpdk-latest needs some rebase and this patch should be rebased
after that.
Comments inline.
Best regards, Ilya Maximets.
On 25.10.2019 15:56, Eelco Chaudron wrote:
> When the dpdk vhost library executes an eventfd_write() call,
DPDK commit 660098d61f57 ("pdump: use generic multi-process channel")
switched pdump to use generic DPDK IPC instead of sockets.
Old API was deprecated and removed. Updating OVS code accordingly.
Signed-off-by: Ilya Maximets
---
Version 3:
* Added note about XDG_RUNTIME_DIR.
Version 2:
*
On 8 Nov 2019, at 6:02, Anju Thomas wrote:
Currently OVS maintains explicit packet drop/error counters only on
port level. Packets that are dropped as part of normal OpenFlow
processing are counted in flow stats of “drop” flows or as table
misses in table stats. These can only be
On 2019-11-11 3:39 PM, Simon Horman wrote:
> On Mon, Nov 04, 2019 at 11:49:44AM +0200, Roi Dayan wrote:
>> In kernels older than 4.8, struct tcf_t didn't have the firstuse.
>> If openvswitch is compiled with the compat pkt_cls.h then there is
>> a struct size mismatch between openvswitch and
Currently, OVS does not register and therefore not handle the
interface reset event from the DPDK framework. This would cause a
problem in cases where a VF is used as an interface, and its
configuration changes.
As an example in the following scenario the MAC change is not
detected/acted upon
On 7 Nov 2019, at 12:50, Ilya Maximets wrote:
On 06.11.2019 14:43, Eelco Chaudron wrote:
Currently, OVS does not register and therefore not handle the
interface reset event from the DPDK framework. This would cause a
problem in cases where a VF is used as an interface, and its
configuration
On Mon, Nov 04, 2019 at 11:49:44AM +0200, Roi Dayan wrote:
> In kernels older than 4.8, struct tcf_t didn't have the firstuse.
> If openvswitch is compiled with the compat pkt_cls.h then there is
> a struct size mismatch between openvswitch and the kernel which cause
> parsing netlink actions to
On Sat, Nov 9, 2019 at 8:35 AM Han Zhou wrote:
>
>
>
> On Fri, Nov 8, 2019 at 6:38 AM Dumitru Ceara wrote:
> >
> > ARP request and ND NS packets for router owned IPs were being
> > flooded in the complete L2 domain (using the MC_FLOOD multicast group).
> > However this creates a scaling issue in
On Mon, Nov 11, 2019 at 9:07 PM Simon Horman wrote:
>
> On Sun, Nov 10, 2019 at 07:44:18PM +0800, xiangxia.m@gmail.com wrote:
> > From: Tonghao Zhang
> >
> > When using the kernel datapath, the upcall don't
> > add skb hash info relatived. That will introduce
> > some problem, because the
On 25 Oct 2019, at 15:56, Eelco Chaudron wrote:
When the dpdk vhost library executes an eventfd_write() call,
i.e. waking up the guest, a new callback will be called.
This patch adds the callback to count the number of
interrupts sent to the VM to track the number of times
interrupts where
On Sun, Nov 10, 2019 at 07:44:18PM +0800, xiangxia.m@gmail.com wrote:
> From: Tonghao Zhang
>
> When using the kernel datapath, the upcall don't
> add skb hash info relatived. That will introduce
> some problem, because the hash of skb is very
> important (e.g. vxlan module uses it for udp
> Hi Darrell,
>
> In TSO case, the packet may use multi-segments mbuf, and I do not> think we
> need to make it linearal. In this case, we can NOT use
> pointer to calculate the tcp length.
Hi. Just wanted to mention that current main course for TSO enabling in
OVS seems to be using of extbuf
Hi Ilya,
Thanks for the review. I agree with your proposal to move the stats update
code to existing special functions. Thanks for the incremental patch, it looks
good to me. Will wait for Kevin's taught on this.
Thanks & Regards,
Sriram.
-Original Message-
From: Ilya Maximets
Sent:
36 matches
Mail list logo