On 12/07/2021 05:53, Han Zhou wrote:
> That's weird. These two tests just pass when I revert this commit. With
> this commit it fails at:
>
> ../../tests/system-ovn.at:5752: ovs-appctl dpctl/dump-flows | grep
> 'ct_state(-new+est+rpl+trk).*ct_label(0x.*04010204/.*)' -c
> --- - 2021-07-
Allow adding label to an ACL to identify which ACL allowed a particular
flow in the connection tracking table.
The ACL label covers 32 bits at the end of ct_label. Since only allowed
connections are committed, only "allow" and "allow-related" ACLs can
have the label.
If the ACL allowing the conne
References: <20210712073436.24906-1-priyankar.j...@nutanix.com>
Bleep bloop. Greetings Priyankar Jain, 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: Line is 412 char
Whenever a Load_Balancer is updated, e.g., a VIP is added, the following
sequence of events happens:
1. The Southbound Load_Balancer record is updated.
2. The Southbound Datapath_Binding records on which the Load_Balancer is
applied are updated.
3. Southbound ovsdb-server sends updates about th
On Thu, Jul 8, 2021 at 3:53 PM Kevin Traynor wrote:
>
> This reworks the current rxq scheduling code to break it into more
> generic and reusable pieces.
>
> The behaviour does not change from a user perspective, except the logs
> are updated to be more consistent.
>
> From an implementation view,
On 7/9/21 6:11 PM, Han Zhou wrote:
>> To avoid this potentially expensive table walk, we use the load_balancer
>> uuids stored in the datapath record itself (it's probably best to see
>> those as hints I guess).
>>
> Thanks for the explain. What you described is indeed a dependency between
> lflow
On Thu, Jul 8, 2021 at 3:53 PM Kevin Traynor wrote:
>
> PMD auto load balance had its own separate implementation of the
> rxq scheduling that it used for dry runs. This was done because
> previously the rxq scheduling was not made reusable for a dry run.
>
> Apart from the code duplication (which
On Thu, Jul 8, 2021 at 3:54 PM Kevin Traynor wrote:
>
> The list of PMDs is round robined through for the selection
> when assigning an rxq to a PMD. The list is based on a
> hash map, so there is no defined order.
>
> It means the same set of PMDs may get assigned different rxqs
> on different ru
On Thu, Jul 8, 2021 at 3:54 PM Kevin Traynor wrote:
>
> Previously, if pmd-rxq-affinity was used to pin an rxq to
> a core that was not in pmd-cpu-mask the rxq was not polled
> for and the user received a warning. This meant that no traffic
> would be received from that rxq.
>
> Now that pinned an
On Thu, Jul 8, 2021 at 3:54 PM Kevin Traynor wrote:
>
> Add an rxq scheduling option that allows rxqs to be grouped
> on a pmd based purely on their load.
>
> The current default 'cycles' assignment sorts rxqs by measured
> processing load and then assigns them to a list of round robin PMDs.
> Thi
On Thu, Jul 8, 2021 at 3:54 PM Kevin Traynor wrote:
>
> Pinning an rxq to a PMD with pmd-rxq-affinity may be done for
> various reasons such as reserving a full PMD for an rxq, or to
> ensure that multiple rxqs from a port are handled on different PMDs.
>
> Previously pmd-rxq-affinity always isola
On Thu, Jul 8, 2021 at 3:54 PM Kevin Traynor wrote:
>
> Add some tests to ensure that if there are numa local
> PMDs they are used for polling an rxq.
>
> Also check that if there are only numa non-local PMDs they
> will be used ito poll the rxq and but the user will be warned.
>
> Signed-off-by:
On Thu, Jul 8, 2021 at 3:53 PM Kevin Traynor wrote:
>
> The first two patches do not provide new functionality for the user
> (except the logs are reworked). They are reworking to make the
> rxq scheduling and PMD auto load balance code more modular for cleanup
> and to be used by subsequent patch
On 9 Jul 2021, at 20:23, Ilya Maximets wrote:
> On 7/9/21 10:35 AM, Eelco Chaudron wrote:
>>
>>
>> On 8 Jul 2021, at 22:18, Ilya Maximets wrote:
>>
>>> On 5/17/21 3:20 PM, Eelco Chaudron wrote:
When OVs installs the flower rule, it only checks for the OK from the
kernel. It does not c
> -Original Message-
> From: Eli Britstein
> Sent: Sunday, July 11, 2021 6:15 AM
> To: d...@openvswitch.org; Ilya Maximets ; Van Haaren,
> Harry
> Cc: Gaetan Rivet ; Majd Dibbiny ; Eli
> Britstein
> Subject: [PATCH 2/2] netdev-offload-dpdk: Fix vxlan vni cast-align warnings
>
> Compilin
See some comments below…
For this patch series, I’m only looking at the diff from v6..v9, not a full
review.
I will do basic compilation and some tests at the end.
Cheers,
Eelco
On 12 Jul 2021, at 7:51, kumar Amber wrote:
> From: Kumar Amber
>
> This patch introduces the MFEX function point
On Mon, Jun 28, 2021 at 10:07 AM Tonghao Zhang
wrote:
> On Thu, Apr 1, 2021 at 9:34 PM Tonghao Zhang
> wrote:
> >
> > On Mon, Dec 14, 2020 at 11:11 AM wrote:
> > >
> > > From: Tonghao Zhang
> > >
> > > The root cause is that the old version of openvswitch doesn't
> > > remove the padding from
Introduce stopwatch instrumentation to the following ovn-controller
routines:
- commit_ct_zones
- bfd_run
- patch_run
- pinctrl_run
- if_status_mgr_update
- if_status_mgr_run
- ofctrl_seqno_run
Signed-off-by: Lorenzo Bianconi
---
This patch is based on controller:
"Add stopwatch to measure OF upd
On 12 Jul 2021, at 7:51, kumar Amber wrote:
> From: Kumar Amber
>
> The study function runs all the available implementations
> of miniflow_extract and makes a choice whose hitmask has
> maximum hits and sets the mfex to that function.
>
> Study can be run at runtime using the following comman
On Mon, Jul 12, 2021 at 10:28:15AM +0200, Eelco Chaudron wrote:
>
>
> On 9 Jul 2021, at 20:23, Ilya Maximets wrote:
>
> > On 7/9/21 10:35 AM, Eelco Chaudron wrote:
> >>
> >>
> >> On 8 Jul 2021, at 22:18, Ilya Maximets wrote:
> >>
> >>> On 5/17/21 3:20 PM, Eelco Chaudron wrote:
> When OVs insta
Bleep bloop. Greetings Lorenzo Bianconi, 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: Failed to merge in the changes.
hint: Use 'git am --show-current-patch' to see the faile
On 12 Jul 2021, at 7:51, kumar Amber wrote:
> From: Kumar Amber
>
> The study function runs all the available implementations
> of miniflow_extract and makes a choice whose hitmask has
> maximum hits and sets the mfex to that function.
>
> Study can be run at runtime using the following command
On 12 Jul 2021, at 7:51, kumar Amber wrote:
> From: Kumar Amber
>
> This commit adds a section to the dpdk/bridge.rst netdev documentation,
> detailing the added miniflow functionality. The newly added commands are
> documented, and sample output is provided.
>
> The use of auto-validator and s
On Mon, Jul 12, 2021 at 02:22:46PM +0200, Eelco Chaudron wrote:
> See some comments below…
>
> For this patch series, I’m only looking at the diff from v6..v9, not a full
> review.
> I will do basic compilation and some tests at the end.
>
> Cheers,
>
> Eelco
>
>
> On 12 Jul 2021, at 7:51, ku
On 12 Jul 2021, at 7:51, kumar Amber wrote:
> From: Kumar Amber
>
> This commit adds a new command to allow the user to enable
> autovalidatior by default at build time thus allowing for
> runnig unit test by default.
>
> $ ./configure --enable-mfex-default-autovalidator
>
> Signed-off-by: Ku
On 7/12/21 7:51 AM, kumar Amber wrote:
> From: Kumar Amber
>
> This patch introduces the MFEX function pointers which allows
> the user to switch between different miniflow extract implementations
> which are provided by the OVS based on optimized ISA CPU.
>
> The user can query for the availabl
> -Original Message-
> From: Ilya Maximets
> Sent: Monday, July 12, 2021 2:25 PM
> To: Amber, Kumar ; ovs-dev@openvswitch.org
> Cc: f...@sysclose.org; echau...@redhat.com; i.maxim...@ovn.org; Van Haaren,
> Harry ; Ferriter, Cian ;
> Stokes, Ian
> Subject: Re: [v9 01/12] dpif-netdev: Add c
Whenever a Load_Balancer is updated, e.g., a VIP is added, the following
sequence of events happens:
1. The Southbound Load_Balancer record is updated.
2. The Southbound Datapath_Binding records on which the Load_Balancer is
applied are updated.
3. Southbound ovsdb-server sends updates about th
On 7/12/21 10:11 AM, Dumitru Ceara wrote:
> On 7/9/21 6:11 PM, Han Zhou wrote:
>>> To avoid this potentially expensive table walk, we use the load_balancer
>>> uuids stored in the datapath record itself (it's probably best to see
>>> those as hints I guess).
>>>
>> Thanks for the explain. What you
On 7/12/21 4:02 PM, Van Haaren, Harry wrote:
>> -Original Message-
>> From: Ilya Maximets
>> Sent: Monday, July 12, 2021 2:25 PM
>> To: Amber, Kumar ; ovs-dev@openvswitch.org
>> Cc: f...@sysclose.org; echau...@redhat.com; i.maxim...@ovn.org; Van Haaren,
>> Harry ; Ferriter, Cian ;
>> Stoke
> -Original Message-
> From: Ilya Maximets
> Sent: Monday, July 12, 2021 3:43 PM
> To: Van Haaren, Harry ; Ilya Maximets
> ; Amber, Kumar ; ovs-
> d...@openvswitch.org
> Cc: f...@sysclose.org; echau...@redhat.com; Ferriter, Cian
> ; Stokes, Ian
> Subject: Re: [v9 01/12] dpif-netdev: Add c
Log flow modifications to help debugging.
Signed-off-by: Eli Britstein
Reviewed-by: Gaetan Rivet
---
lib/dpif-netdev.c | 101 +-
1 file changed, 55 insertions(+), 46 deletions(-)
diff --git a/lib/dpif-netdev.c b/lib/dpif-netdev.c
index 9b2b8d6d9..cae
Before flushing offloads of a removed port was supported by [1], it was
necessary to flush the 'marks'. In doing so, all offloads of the PMD are
removed, include the ones that are not related to the removed port and
that are not modified following this removal. As a result such flows are
evicted fr
This patch-set improves offloads transitions behavior.
Patch #1 avoids flushing PMD offloads unnecessarily.
Patch #2 fixes a race condition with flow modifications.
Patch #3 improves debuggability of flow modifications.
v2-v1: Rebase.
Travis:
v1: https://travis-ci.org/github/elibritstein/OVS/bui
Association of a mark to a flow is done as part of its offload handling,
in the offloading thread. However, the PMD thread specifies whether an
offload request is an "add" or "modify" by the association of a mark to
the flow.
This is exposed to a race condition. A flow might be created with
actions
On Mon, Jul 12, 2021 at 4:43 PM Ilya Maximets wrote:
> >> ovsrobot has issues with reporting the status right now, but this
> >> patch fails the build in GHA:
> >> https://github.com/ovsrobot/ovs/actions/runs/1021787643
> >
> > Thanks for linking on results.
> >
> > I've spot-checked a bunch of
Since this addressed Han's findings in v1 and I had already ACKed it, I
pushed this change to the main branch. Thanks, Dumitru and Han!
On 7/12/21 10:14 AM, Dumitru Ceara wrote:
Whenever a Load_Balancer is updated, e.g., a VIP is added, the following
sequence of events happens:
1. The Southbou
On 7/12/21 5:10 PM, David Marchand wrote:
> On Mon, Jul 12, 2021 at 4:43 PM Ilya Maximets wrote:
ovsrobot has issues with reporting the status right now, but this
patch fails the build in GHA:
https://github.com/ovsrobot/ovs/actions/runs/1021787643
>>>
>>> Thanks for linking on re
On 7/12/21 4:57 PM, Van Haaren, Harry wrote:
>> -Original Message-
>> From: Ilya Maximets
>> Sent: Monday, July 12, 2021 3:43 PM
>> To: Van Haaren, Harry ; Ilya Maximets
>> ; Amber, Kumar ; ovs-
>> d...@openvswitch.org
>> Cc: f...@sysclose.org; echau...@redhat.com; Ferriter, Cian
>> ; Stok
For the approach,
Acked-by: Mark Michelson
I have one final suggestion down below.
On 7/10/21 6:13 AM, Lorenzo Bianconi wrote:
Incrementally manage local_active_ports_ipv6_pd map for interfaces
where IPv6 prefix-delegation has been enabled. This patch allows to
avoid looping over all local in
Acked-by: Mark Michelson
On 7/10/21 6:13 AM, Lorenzo Bianconi wrote:
Incrementally manage local_active_ports_ras map for interfaces
where periodic router advertisement has been enabled. This patch
allows to avoid looping over all local interfaces to check if
periodic RA is running on the curren
On 12 Jul 2021, at 7:51, kumar Amber wrote:
> From: Kumar Amber
>
> This commit introduces additional command line paramter
> for mfex study function. If user provides additional packet out
> it is used in study to compare minimum packets which must be processed
> else a default value is choose
I pushed this to master and branch-21.06.
On 7/8/21 4:19 PM, Mark Michelson wrote:
Acked-by: Mark Michelson
On 7/7/21 10:09 AM, Dumitru Ceara wrote:
This was causing some IPv6 system tests to fail when run with
ovn-northd-ddlog.
Also fix cleanup of the northd process in system-ovn.at. A few
I pushed this to master.
On 7/8/21 4:27 PM, Mark Michelson wrote:
Acked-by: Mark Michelson
On 7/6/21 10:41 AM, Dumitru Ceara wrote:
Also, shorten the CONTROLLER_LOOP_STOPWATCH_NAME name as there is a bug
in lib/stopwatch.c which fails to report an error when the stopwatch
name is longer than
Ilya Maximets writes:
> On 7/12/21 5:10 PM, David Marchand wrote:
>> On Mon, Jul 12, 2021 at 4:43 PM Ilya Maximets wrote:
> ovsrobot has issues with reporting the status right now, but this
> patch fails the build in GHA:
> https://github.com/ovsrobot/ovs/actions/runs/1021787643
>>
Aaron Conole writes:
> Ilya Maximets writes:
>
>> On 7/12/21 5:10 PM, David Marchand wrote:
>>> On Mon, Jul 12, 2021 at 4:43 PM Ilya Maximets wrote:
>> ovsrobot has issues with reporting the status right now, but this
>> patch fails the build in GHA:
>> https://github.com/ovsrobot
From: Anton Ivanov
time_poll() makes an excessive number of time_msec() calls
which incur a performance penalty.
1. Avoid time_msec() call for timeout calculation when time_poll()
is asked to skip poll()
2. Reuse the time_msec() result from deadline calculation for
last_wakeup and timeout calcu
From: Anton Ivanov
If we are not obtaining any useful information out of the poll(),
such as is a fd busy or not, we do not need to do a poll() if
an immediate_wake() has been requested.
This cuts out all the pollfd hash additions, forming the poll
arguments and the actual poll() after a call to
Statically add IPv6 neighbor MAC addresses to avoid NS messages
evicting datapath flows causing occasional test failures.
We also configure all interfaces to send only one IPv6 router
solicitation message. These messages can cause datapath flows
to be unexpectedly evicted causing test failures.
F
Fix 2 test flakes that have been observed in the OVS CI
Mark Gray (2):
system-test: Fix flake in ECMP IPv6 symmetric reply test
ipsec.at: Fix ipsec test flake
tests/ovn-ipsec.at | 8 ---
tests/system-ovn.at | 51 +
2 files changed, 33 inserti
Change order of command execution and add `ovn-nbctl --wait=hv sync`.
This ensures the vswitchd ovsdb instance is updated by the time
it is checked.
Fixes: ff2b6ff69740 ("ovn-controller: Add 'local_ip' option to tunnel ports for
IPsec case")
Signed-off-by: Mark Gray
---
tests/ovn-ipsec.at | 8 +
Bleep bloop. Greetings Anton Ivanov, 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: Line has trailing whitespace
#171 FILE: lib/timeval.c:327:
* shortcut. Otherw
Hi Joe,
Maybe you can take a look...
Thanks,
fbl
On Thu, Jul 08, 2021 at 11:40:12AM -0300, Flavio Leitner wrote:
>
> Hi Pravin,
>
> Any thoughts on this patch? We are closing OVS 2.16, so it would
> be nice to know if it looks okay or needs changes, specially
> changes related to the userspac
https://bugzilla.redhat.com/show_bug.cgi?id=1944220
Changes since v2:
- use smap_get_bool instead of smap_get in update_active_pb_ras_pd routine
Changes since v1:
- use shash instead of hamp
- always remove the entry from shash if the user removed the ipv6_pd/ras info
in port_binding option col
Incrementally manage local_active_ports_ras map for interfaces
where periodic router advertisement has been enabled. This patch
allows to avoid looping over all local interfaces to check if
periodic RA is running on the current port binding.
Acked-by: Mark Michelson
Signed-off-by: Lorenzo Biancon
Incrementally manage local_active_ports_ipv6_pd map for interfaces
where IPv6 prefix-delegation has been enabled. This patch allows to
avoid looping over all local interfaces to check if prefix-delegation
is running on the current port binding.
Acked-by: Mark Michelson
Signed-off-by: Lorenzo Bian
Hi Kumar,
There is an issue with the signed-offs reported by 0-day Robot.
For additional info, please check the link below and look for the
tag Co-authored-by:
https://github.com/openvswitch/ovs/blob/master/Documentation/internals/contributing/submitting-patches.rst#tags
Otherwise the patch loo
On Thu, Jun 17, 2021 at 3:36 AM Dumitru Ceara wrote:
>
> Also store the timestamp when ovn-controller started up. This helps
> implementing alerts on the CMS side to detect whether ovn-controller is
> still alive and functioning well.
>
> Reported-at: https://bugzilla.redhat.com/1924751
> Reporte
On Fri, Jul 9, 2021 at 11:54 AM Dumitru Ceara wrote:
>
> The backport looks good to me, thanks!
Thanks. I applied this patch to branch-20.09.
Numan
>
> On 7/9/21 4:00 PM, Vladislav Odintsov wrote:
> > Signed-off-by: Vladislav Odintsov
> >
> > Regards,
> > Vladislav Odintsov
> >
> >> On 9 Jul
On Fri, Jul 9, 2021 at 9:01 AM Dumitru Ceara wrote:
>
> On 7/8/21 6:34 PM, Vladislav Odintsov wrote:
> > Hi,
>
> Hi Vladislav,
>
> >
> > I see constantly failing test while OVN branch-20.09 against OVS higher
> > than 2.14.0 (2.14.1, 2.14.2, branch-2.14):
> > ovn -- ensure one gw controller resta
On Fri, Jul 9, 2021 at 2:49 PM Han Zhou wrote:
>
> lb_force_snat_ip is a flag set in logical flow pipeline, while
> lb_force_snat_ip is the option configured in NB DB. In NB document we
> should mention the actual option configured in NB instead of the flow
> details.
>
> Signed-off-by: Han Zhou
On Wed, May 12, 2021 at 05:08:08PM +0200, Timothy Redaelli wrote:
> Systemd unit file generates warnings about PID file path since /var/run
> is a legacy path so just use /run instead of /var/run.
>
> /var/run is a symlink of /run starting from RHEL7 (and any other distribution
> that uses systemd
On 6/29/21 1:20 PM, Dumitru Ceara wrote:
> Until now clients that needed to reconnect immediately could only use
> reconnect_force_reconnect(). However, reconnect_force_reconnect()
> doesn't reset the backoff for connections that were alive long enough
> (more than backoff seconds).
>
> Moreover,
On 6/19/21 8:01 PM, Mark Gray wrote:
> On 12/06/2021 03:00, Ilya Maximets wrote:
>> This will be used to apply update3 type updates to ovsdb tables
>> while processing updates for future ovsdb 'relay' service model.
>>
>> 'ovsdb_datum_apply_diff' is allowed to fail, so adding support
>> to return t
Sorry Lorenzo but I found one more issue. Sorry for not noticing it
during an earlier review.
On 7/12/21 2:18 PM, Lorenzo Bianconi wrote:
Incrementally manage local_active_ports_ipv6_pd map for interfaces
where IPv6 prefix-delegation has been enabled. This patch allows to
avoid looping over all
On 7/12/21 2:18 PM, Lorenzo Bianconi wrote:
Incrementally manage local_active_ports_ras map for interfaces
where periodic router advertisement has been enabled. This patch
allows to avoid looping over all local interfaces to check if
periodic RA is running on the current port binding.
Acked-by:
On 6/19/21 8:30 PM, Mark Gray wrote:
> On 12/06/2021 03:00, Ilya Maximets wrote:
>> New database service model 'relay' that is needed to scale out
>> read-mostly database access, e.g. ovn-controller connections to
>> OVN_Southbound.
>>
>> In this service model ovsdb-server connects to existing OVSD
On 6/25/21 3:34 PM, Dumitru Ceara wrote:
> On 6/12/21 4:00 AM, Ilya Maximets wrote:
>> New database service model 'relay' that is needed to scale out
>> read-mostly database access, e.g. ovn-controller connections to
>> OVN_Southbound.
>>
>> In this service model ovsdb-server connects to existing O
On 7/2/21 12:24 PM, Mark Gray wrote:
> On 12/06/2021 03:00, Ilya Maximets wrote:
>> Current version of ovsdb relay allows to scale out read-only
>> access to the primary database. However, many clients are not
>> read-only but read-mostly. For example, ovn-controller.
>>
>> In order to scale out
On 6/25/21 3:34 PM, Dumitru Ceara wrote:
> On 6/12/21 4:00 AM, Ilya Maximets wrote:
>> Current version of ovsdb relay allows to scale out read-only
>> access to the primary database. However, many clients are not
>> read-only but read-mostly. For example, ovn-controller.
>>
>> In order to scale o
On 6/25/21 3:34 PM, Dumitru Ceara wrote:
> On 6/12/21 4:00 AM, Ilya Maximets wrote:
>> It might be important for clients to know that relay lost connection
>> with the relay remote, so they could re-connect to other relay.
>>
>> Signed-off-by: Ilya Maximets
>> ---
>
> [...]
>
>>
>> +#define RE
On 7/2/21 12:48 PM, Mark Gray wrote:
> On 12/06/2021 03:00, Ilya Maximets wrote:
>> It might be important for clients to know that relay lost connection
>> with the relay remote, so they could re-connect to other relay.
>
> Yeah this makes sense. I guess there are some deployment scenarios we
> sh
On Thu, Jul 08, 2021 at 08:59:24PM +0200, Dumitru Ceara wrote:
> Hi Ben,
>
> As discussed earlier, during the OVN meeting, I've noticed a new
> performance issue with ovn-northd-ddlog when running it against a
> database from one of our more recent scale tests:
>
> http://people.redhat.com/~dcear
Hi Dumitru. Can you please rebase this? There's a conflict due to
384a7c6237da8f88ab68a9abd0982f92d7d8c2d2 (northd: Refactor Logical Flows
for routers with DNAT/Load Balancers).
On 7/6/21 6:45 AM, Lorenzo Bianconi wrote:
This allows creating the match strings for each LB VIP exactly once,
inst
On 6/25/21 3:35 PM, Dumitru Ceara wrote:
> On 6/12/21 4:00 AM, Ilya Maximets wrote:
>> Main documentation for the service model and tutorial with the use case
>> and configuration examples.
>>
>> Signed-off-by: Ilya Maximets
>> ---
>
> I left a few minor comments below. With them addressed:
>
>
On 7/2/21 1:05 PM, Mark Gray wrote:
> On 12/06/2021 03:00, Ilya Maximets wrote:
>> Main documentation for the service model and tutorial with the use case
>> and configuration examples.
>>
>> Signed-off-by: Ilya Maximets
>> ---
>> Documentation/automake.mk| 1 +
>> Documentation/ref
On 6/25/21 3:33 PM, Dumitru Ceara wrote:
> On 6/12/21 3:59 AM, Ilya Maximets wrote:
>> Replication can be used to scale out read-only access to the database.
>> But there are clients that are not read-only, but read-mostly.
>> One of the main examples is ovn-controller that mostly monitors
>> updat
On Mon, Jul 12, 2021 at 10:28 AM Mark Gray wrote:
>
> Statically add IPv6 neighbor MAC addresses to avoid NS messages
> evicting datapath flows causing occasional test failures.
>
> We also configure all interfaces to send only one IPv6 router
> solicitation message. These messages can cause datap
On Mon, Jul 12, 2021 at 12:42 PM Numan Siddique wrote:
>
> On Fri, Jul 9, 2021 at 2:49 PM Han Zhou wrote:
> >
> > lb_force_snat_ip is a flag set in logical flow pipeline, while
> > lb_force_snat_ip is the option configured in NB DB. In NB document we
> > should mention the actual option configur
From: Harry van Haaren
This commit avoids many instances of "using subtable X for miniflow (x,y)"
in the ovs-vswitchd log when using the DPCLS Autovalidator. This occurs
when no specialized subtable is found, and the generic "_any" version of
the avx512 subtable search implementation was used. Th
Hi Eelco ,
Thanks for Ack. Will also fix the minor changes in next patch 😊
> > +for (int i = MFEX_IMPL_START_IDX; i < MFEX_IMPL_MAX; i++) {
> > +VLOG_DBG("MFEX study results for implementation %s:"
> > + " (hits %d/%d pkts)",
> > +
Hi Eelco
fixed all typos in v10
> -Original Message-
> From: Eelco Chaudron
> Sent: Monday, July 12, 2021 6:37 PM
> To: Amber, Kumar
> Cc: ovs-dev@openvswitch.org; f...@sysclose.org; i.maxim...@ovn.org; Van
> Haaren, Harry ; Ferriter, Cian
> ; Stokes, Ian
> Subject: Re: [v9 04/12] doc
Hi Eelco, Flavio ,
Pls find my comments Inline.
> > +/* For the first call, this will be choosen based on the
> > + * compile time flag and if nor flag is set it is set to
> > + * default scalar.
> > + */
> > +if (OVS_UNLIKELY(!default_mfex_func_set)) {
> > +VLOG_INF
Hi Flavio,
All fixed ready to merge
> -Original Message-
> From: Flavio Leitner
> Sent: Monday, July 12, 2021 11:53 PM
> To: Amber, Kumar
> Cc: ovs-dev@openvswitch.org
> Subject: Re: [ovs-dev] [v4] dpif/dpcls: limit count subtable search info logs
>
>
> Hi Kumar,
>
> There is an iss
Hi Eelco,
Fixed all and reworked default.
> -Original Message-
> From: Eelco Chaudron
> Sent: Monday, July 12, 2021 6:50 PM
> To: Amber, Kumar
> Cc: ovs-dev@openvswitch.org; f...@sysclose.org; i.maxim...@ovn.org; Van
> Haaren, Harry ; Ferriter, Cian
> ; Stokes, Ian
> Subject: Re: [v9 0
Hi Eelco,
Pls find my comments inline.
> -Original Message-
> From: Eelco Chaudron
> Sent: Monday, July 12, 2021 9:17 PM
> To: Amber, Kumar
> Cc: ovs-dev@openvswitch.org; f...@sysclose.org; i.maxim...@ovn.org; Van
> Haaren, Harry ; Ferriter, Cian
> ; Stokes, Ian
> Subject: Re: [v9 06/1
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")
Signed-off-by: Eli Britstein
Reviewed-by: Salem Sol
---
lib/netdev-offload-dpdk.c | 16
1 fi
For VLANs, the match of ethernet type should be specified in inner_type
field of the vlan match, and not type field in ethernet match.
Fix it.
Fixes: e8a2b5bf92bb ("netdev-dpdk: implement flow offload with rte flow")
Signed-off-by: Eli Britstein
Reviewed-by: Salem Sol
---
lib/netdev-offload-dpd
v10 update:
- re-worked the default implementation
- fix comments from Flavio and Eelco
- Include Acks from Eelco in study
v9 update:
- Include review comments from Flavio
- Rebase onto Master
- Include Acks from Flavio
v8 updates:
- Include documentation on AVX512 MFEX as per Eelco's suggestion on
From: Kumar Amber
This patch introduced the auto-validation function which
allows users to compare the batch of packets obtained from
different miniflow implementations against the linear
miniflow extract and return a hitmask.
The autovaidator function can be triggered at runtime using the
follo
From: Kumar Amber
This patch introduces the MFEX function pointers which allows
the user to switch between different miniflow extract implementations
which are provided by the OVS based on optimized ISA CPU.
The user can query for the available minflow extract variants available
for that CPU by
From: Kumar Amber
The study function runs all the available implementations
of miniflow_extract and makes a choice whose hitmask has
maximum hits and sets the mfex to that function.
Study can be run at runtime using the following command:
$ ovs-appctl dpif-netdev/miniflow-parser-set study
Sign
From: Kumar Amber
This commit adds a section to the dpdk/bridge.rst netdev documentation,
detailing the added miniflow functionality. The newly added commands are
documented, and sample output is provided.
The use of auto-validator and special study function is also described
in detail as well a
From: Kumar Amber
This commit adds a new command to allow the user to enable
autovalidatior by default at build time thus allowing for
runnig unit test by default.
$ ./configure --enable-mfex-default-autovalidator
Signed-off-by: Kumar Amber
Co-authored-by: Harry van Haaren
Signed-off-by: Har
From: Kumar Amber
This commit introduces additional command line paramter
for mfex study function. If user provides additional packet out
it is used in study to compare minimum packets which must be processed
else a default value is choosen.
Also introduces a third paramter for choosing a particu
From: Kumar Amber
Tests:
6: OVS-DPDK - MFEX Autovalidator
7: OVS-DPDK - MFEX Autovalidator Fuzzy
Added a new directory to store the PCAP file used
in the tests and a script to generate the fuzzy traffic
type pcap to be used in fuzzy unit test.
Signed-off-by: Kumar Amber
Acked-by: Flavio Le
From: Harry van Haaren
This commit adds a new counter to be displayed to the user when
requesting datapath packet statistics. It counts the number of
packets that are parsed and a miniflow built up from it by the
optimized miniflow extract parsers.
The ovs-appctl command "dpif-netdev/pmd-perf-sh
From: Harry van Haaren
This commit enables OVS to at runtime check for more detailed
AVX512 capabilities, specifically Byte and Word (BW) extensions,
and Vector Bit Manipulation Instructions (VBMI).
These instructions will be used in the CPU ISA optimized
implementations of traffic profile aware
From: Harry van Haaren
This commit adds AVX512 implementations of miniflow extract.
By using the 64 bytes available in an AVX512 register, it is
possible to convert a packet to a miniflow data-structure in
a small quantity instructions.
The implementation here probes for Ether()/IP()/UDP() traff
From: Harry van Haaren
This commit adds 3 new traffic profile implementations to the
existing avx512 miniflow extract infrastructure. The profiles added are:
- Ether()/IP()/TCP()
- Ether()/Dot1Q()/IP()/UDP()
- Ether()/Dot1Q()/IP()/TCP()
The design of the avx512 code here is for scalability to ad
1 - 100 of 116 matches
Mail list logo
