Hi,
On Wed, Mar 16, 2022 at 4:32 PM Tonghao Zhang wrote:
>
> On Wed, Mar 16, 2022 at 4:09 PM Harold Huang wrote:
> >
> > When we send parallel flows such as VXLAN to a PF[1] port in OVS-DPDK with
> > multiple PMDs. OVS will create a RTE flow with Mark and RSS actions to send
> > flows to the sof
Bleep bloop. Greetings Chris Mi, I am a robot and I have tried out your patch.
Thanks for your contribution.
I encountered some error that I wasn't expecting. See the details below.
git-am:
error: sha1 information is lacking or useless (lib/netdev-offload-tc.c).
error: could not build fake anc
Bleep bloop. Greetings Chris Mi, I am a robot and I have tried out your patch.
Thanks for your contribution.
I encountered some error that I wasn't expecting. See the details below.
Patch skipped due to previous failure.
Please check this out. If you feel there has been an error, please emai
Create a unique group ID to map the sFlow info when offloading sFlow
action to TC. When showing the offloaded datapath flows, translate the
group ID from TC sample action to sFlow info using the mapping.
Signed-off-by: Chris Mi
Reviewed-by: Eli Britstein
Acked-by: Eelco Chaudron
---
NEWS
Add two sFlow offload test caes:
3: sflow offloads with sampling=1 - ping between two ports - offloads enabled
ok
4: sflow offloads with sampling=2 - ping between two ports - offloads enabled
ok
Signed-off-by: Chris Mi
Acked-by: Eelco Chaudron
---
tests/system-offloads-traffic.at | 101 +
Process sFlow offload packet in handler thread if handler id is 0.
Signed-off-by: Chris Mi
Reviewed-by: Eli Britstein
Acked-by: Eelco Chaudron
---
ofproto/ofproto-dpif-upcall.c | 73 +++
1 file changed, 73 insertions(+)
diff --git a/ofproto/ofproto-dpif-upcall.
Implement dpif-offload API for netlink datapath.
Signed-off-by: Chris Mi
Reviewed-by: Eli Britstein
---
lib/automake.mk | 1 +
lib/dpif-netdev.c | 3 +-
lib/dpif-netlink.c | 20 +++-
lib/dpif-offload-netlink.c | 221
lib/
When offloading sample action to TC, userspace creates a unique ID
to map sFlow action and tunnel info and passes this ID to kernel instead
of the sFlow info. psample will send this ID and sampled packet to
userspace. Using the ID, userspace can recover the sFlow info and send
sampled packet to the
Some offload actions require functionality that is not netdev
based, but dpif. For example, sFlow action requires to create
a psample netlink socket to receive the sampled packets from
TC or kernel driver.
Create dpif-offload-provider layer to support such actions.
Signed-off-by: Chris Mi
Review
Load kernel module psample to receive sampled packets from TC.
Before removing kernel module psample, remove act_sample first.
Signed-off-by: Chris Mi
Reviewed-by: Eli Britstein
Acked-by: Eelco Chaudron
---
utilities/ovs-kmod-ctl.in | 14 ++
1 file changed, 14 insertions(+)
diff -
Update kernel UAPI to support psample and the tc sample action.
Signed-off-by: Chris Mi
Reviewed-by: Eli Britstein
Acked-by: Eelco Chaudron
---
include/linux/automake.mk| 4 ++-
include/linux/psample.h | 62
include/linux/tc_act/tc_sample.h |
This patch set adds offload support for sFlow.
Psample is a genetlink channel for packet sampling. TC action act_sample
uses psample to send sampled packets to userspace.
When offloading sample action to TC, userspace creates a unique ID to
map sFlow action and tunnel info and passes this ID to k
On 2022-03-11 8:53 PM, Eelco Chaudron wrote:
On 11 Mar 2022, at 3:14, Chris Mi wrote:
On 2022-03-04 6:50 PM, Eelco Chaudron wrote:
I’m re-adding some of my v18 comments, which I know are depending on Ilya’s
feedback, although I still think we should have provider classes, i.e., not
based on
Currently we allocate and format a string in dp_netdev_flow_add in case
miniflow_bits needs to be printed by dpctl/dump-flows/get-flow. However,
this adds unneeded calls to realloc in the pmd hot path, while
the resulting string may never be viewed.
Instead of keeping a pointer to an allocated str
>
>
> >
> > Hi Han,
> >
> > >
> > > This can be m_existing->name, and so the parameter m_desired is not needed
> > > in this function.
> >
> > ack, I will fix it in v11.
> >
> Hi Lorenzo,
>
> You are replying to an old email. No need for v11 :)
Sorry Han, this mail for some reason this email was q
>
> Hi Han,
>
> >
> > This can be m_existing->name, and so the parameter m_desired is not
needed
> > in this function.
>
> ack, I will fix it in v11.
>
Hi Lorenzo,
You are replying to an old email. No need for v11 :)
Thanks,
Han
___
dev mailing list
d..
> On Tue, Mar 8, 2022 at 4:14 AM Lorenzo Bianconi
> wrote:
> >
> > At the moment ovs meters are reconfigured by ovn just when a
> > meter is allocated or removed while updates for an already
> > allocated meter are ignored. This issue can be easily verified
> > with the following reproducer:
> >
>
This commit doesn't change any functionality, it just performs the
following refactor work:
a. consolidates the old build_ovn_lbs(), build_lrouter_lbs() and
build_ovn_lr_lbs() functions into a single function, build_lbs(),
that builds all the load balancer port-independent information.
Pa
Bleep bloop. Greetings Anurag Agawral, I am a robot and I have tried out your
patch.
Thanks for your contribution.
I encountered some error that I wasn't expecting. See the details below.
checkpatch:
WARNING: Unexpected sign-offs from developers who are not authors or co-authors
or committer
From: Jan Scheurich
Today dpif-netdev considers PMD threads on a non-local NUMA node for
automatic assignment of the rxqs of a port only if there are no local,
non-isolated PMDs.
On typical servers with both physical ports on one NUMA node, this
often
leaves the PMDs on the other NUMA node
On 14 Jul 2021, at 14:33, Eelco Chaudron wrote:
> On 8 Jul 2021, at 21:21, Ilya Maximets wrote:
>
>> On 5/24/21 2:39 PM, Timothy Redaelli wrote:
>>> On Tue, 18 May 2021 06:17:48 -0400
>>> Eelco Chaudron wrote:
>>>
Due to flow lookup optimizations, especially in the resubmit/clone cases,
>
> -Original Message-
> From: Ilya Maximets
> Sent: Wednesday, March 16, 2022 2:43 PM
> To: Eli Britstein ; d...@openvswitch.org; Emma Finn
>
> Cc: i.maxim...@ovn.org; Ian Stokes
> Subject: Re: [ovs-dev] [PATCH V4 1/2] netdev-offload-dpdk: Use has_vlan
> match attribute
>
> External e
On 2/7/22 17:56, Eli Britstein via dev wrote:
> DPDK 20.11 introduced an ability to specify existance/non-existance of
> VLAN tag by [1].
> Use this attribute.
>
> [1]: 09315fc83861 ("ethdev: add VLAN attributes to ethernet and VLAN items")
Hi, Eli. I'm afraid we still can't use the 'has_vlan' i
Hello Kevin,
Thanks for your inputs.
In this scenario we have one VM each on NUMA0 and NUMA1 (VM1 is on NUMA0, VM2
is on NUMA1), dpdk port is on NUMA1.
Without cross-numa-polling, VM/VHU queue traffic is evenly distributed based on
load on their respective NUMA sockets.
However, DPDK
On Wed, Mar 16, 2022 at 4:09 PM Harold Huang wrote:
>
> When we send parallel flows such as VXLAN to a PF[1] port in OVS-DPDK with
> multiple PMDs. OVS will create a RTE flow with Mark and RSS actions to send
> flows to the software data path. But the RSS action does not work well and
> all the fl
When we send parallel flows such as VXLAN to a PF[1] port in OVS-DPDK with
multiple PMDs. OVS will create a RTE flow with Mark and RSS actions to send
flows to the software data path. But the RSS action does not work well and
all the flows are forwarded to a single PMD. This is because RSS hash typ
26 matches
Mail list logo