On Mon, Jul 10, 2023 at 5:26 PM Dumitru Ceara wrote:
> If we want to catch new failures faster we have a better chance if CI
> doesn't auto-retry (once).
>
> There are some tests that are still "unstable" and fail every now and
> then. In order to reduce the number of false negatives keep the
>
From: simon
Fix bug of ovs-tcpdump, which will cause megaflow action wrong.
As use ovs-tcpdump will add mipxxx NIC, and this NIC has IPv6 address by
default.
For vxlan topology, mipxxx will be treated as tunnel port, and will got
error actions.
For detail discuss, refer:
Hi Dumitru,
thanks for digging into this! I highly appreciate your help!
Please, see my answers inline.
> On 10 Jul 2023, at 15:28, Dumitru Ceara wrote:
>
> On 7/10/23 12:57, Dumitru Ceara wrote:
>> On 7/6/23 13:00, Vladislav Odintsov wrote:
>>>
>>>
On 5 Jul 2023, at 20:07, Vladislav
On 7/10/23 15:51, Robin Jarry wrote:
> Ilya Maximets, Jul 10, 2023 at 15:48:
>> In order to speed up the process a bit, proposing a following incremental
>> change:
>>
>> diff --git a/lib/netdev-dpdk.c b/lib/netdev-dpdk.c
>> index 27b2fa5e0..aa87ee546 100644
>> --- a/lib/netdev-dpdk.c
>> +++
On Mon, 10 Jul 2023 20:39:11 +0200 Ilya Maximets wrote:
> > As far as I understand what you're proposing, yes :)
>
> OK. Just to spell it all out:
>
> Userspace will install a flow with an OVS_FLOW_CMD_NEW:
>
> match:ip,tcp,... actions:something,something,drop(0)
> match:ip,udp,...
On 7/10/23 19:01, Jakub Kicinski wrote:
> On Mon, 10 Jul 2023 18:51:19 +0200 Ilya Maximets wrote:
>> Makes sense. I wasn't sure that's a good solution from a kernel perspective
>> either. It's better than defining all these reasons, IMO, but it's not good
>> enough to be considered acceptable, I
On Mon, Jul 10, 2023 at 06:51:19PM +0200, Ilya Maximets wrote:
> On 7/8/23 00:06, Jakub Kicinski wrote:
> > On Fri, 7 Jul 2023 18:04:36 +0200 Ilya Maximets wrote:
> That already exists, right? Johannes added it in the last release for
> WiFi.
> >>>
> >>> I'm not sure. The
On Mon, 10 Jul 2023 18:51:19 +0200 Ilya Maximets wrote:
> Makes sense. I wasn't sure that's a good solution from a kernel perspective
> either. It's better than defining all these reasons, IMO, but it's not good
> enough to be considered acceptable, I agree.
>
> How about we define just 2
On 7/8/23 00:06, Jakub Kicinski wrote:
> On Fri, 7 Jul 2023 18:04:36 +0200 Ilya Maximets wrote:
That already exists, right? Johannes added it in the last release for
WiFi.
>>>
>>> I'm not sure. The SKB_DROP_REASON_SUBSYS_MAC80211_UNUSABLE behaves
>>> similarly
>>> to that on a
Adrian Moreno writes:
> On 6/28/23 18:27, Aaron Conole wrote:
>> Building on the previous work, add a very simplistic NAT case
>> using ipv4. This just tests dnat transformation
>> Signed-off-by: Aaron Conole
>
> Hi Aaron,
>
> I know that the goal is not to support the full syntax, and that
Adrian Moreno writes:
> On 6/28/23 18:27, Aaron Conole wrote:
>> Forwarding via ct() action is an important use case for openvswitch, but
>> generally would require using a full ovs-vswitchd to get working. Add a
>> ct action parser for basic ct test case.
>> Signed-off-by: Aaron Conole
>> ---
On 7/7/23 17:04, Ilya Maximets wrote:
On 7/6/23 18:22, Eelco Chaudron wrote:
On 2 Jun 2023, at 16:13, Adrian Moreno wrote:
Use TCA_POLICE_RATE64 if the rate cannot be expressed using 32bits.
This breaks the 32Gbps barrier. The new barrier is ~4Tbps caused by
netdev's API expressing kbps
On 7/10/23 17:36, Terry Wilson wrote:
> On Mon, Jul 10, 2023 at 10:32 AM Terry Wilson wrote:
>>
>> I accidentally forgot to click reply-to-all.
>>
>> On Fri, Jun 30, 2023 at 10:27 AM Ilya Maximets wrote:
>>>
>>> On 6/30/23 16:54, Adrian Moreno wrote:
On 6/30/23 14:35, Ilya
Bleep bloop. Greetings Dumitru Ceara, 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 85 characters long (recommended limit is 79)
#171 FILE:
On Mon, Jul 10, 2023 at 10:32 AM Terry Wilson wrote:
>
> I accidentally forgot to click reply-to-all.
>
> On Fri, Jun 30, 2023 at 10:27 AM Ilya Maximets wrote:
> >
> > On 6/30/23 16:54, Adrian Moreno wrote:
> > >
> > >
> > > On 6/30/23 14:35, Ilya Maximets wrote:
> > >> On 6/30/23 14:23, Adrian
Currently OVS keeps track of which mirrors that each packet has been
sent to for the purpose of deduplication. However, this doesn't consider
that openflow rules can make significant changes to packets after
ingress.
For example, OVN can create OpenFlow rules that turn an echo request
into an
I accidentally forgot to click reply-to-all.
On Fri, Jun 30, 2023 at 10:27 AM Ilya Maximets wrote:
>
> On 6/30/23 16:54, Adrian Moreno wrote:
> >
> >
> > On 6/30/23 14:35, Ilya Maximets wrote:
> >> On 6/30/23 14:23, Adrian Moreno wrote:
> >>>
> >>>
> >>> On 6/30/23 13:40, Ilya Maximets wrote:
>
Bleep bloop. Greetings Kevin Traynor, 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 80 characters long (recommended limit is 79)
#50 FILE:
On 7/7/23 14:27, Ales Musil wrote:
> On Wed, Jul 5, 2023 at 7:08 PM Dumitru Ceara wrote:
>
>> If we want to catch new failures faster we have a better chance if CI
>> doesn't auto-retry (once).
>>
>> There are some tests that are still "unstable" and fail every now and
>> then. In order to
If we want to catch new failures faster we have a better chance if CI
doesn't auto-retry (once).
There are some tests that are still "unstable" and fail every now and
then. In order to reduce the number of false negatives keep the
--recheck for them. To achieve that we use a new macro,
Several xlate actions used in recursive translation currently store a
large amount of information on the stack. This can result in handler
threads quickly running out of stack space despite before
xlate_resubmit_resource_check() is able to terminate translation. This
patch reduces stack usage by
Add unit tests for new per pmd options of pmd-sleep-max.
Signed-off-by: Kevin Traynor
Reviewed-by: David Marchand
---
tests/pmd.at | 152 +++
1 file changed, 152 insertions(+)
diff --git a/tests/pmd.at b/tests/pmd.at
index 6040d558c..31b832e3e
Extend 'pmd-sleep-max' so that individual PMD thread cores
may have a specified max sleep request value.
Any PMD thread core without a value will use the datapath default
(no sleep request) or datapath global value set by the user.
To set PMD thread cores 8 and 9 to never request a load based
Signed-off-by: Kevin Traynor
Reviewed-by: David Marchand
---
Documentation/topics/dpdk/pmd.rst | 4 ++--
NEWS | 1 +
2 files changed, 3 insertions(+), 2 deletions(-)
diff --git a/Documentation/topics/dpdk/pmd.rst
b/Documentation/topics/dpdk/pmd.rst
index
Max requested sleep time and status for a PMD thread
is logged at start up or when changed, but it can be
convenient to have a command to dump this information
explicitly.
It is envisaged that this will be expanded when future
additions are added.
Signed-off-by: Kevin Traynor
Reviewed-by: David
other_config:pmd-maxsleep is a config option to allow
PMD thread cores to sleep under low or no load conditions.
Rename it to 'pmd-sleep-max' to allow a more structured
name and so that additional options or command can follow
the 'pmd-sleep-xyz' pattern.
Use of other_config:pmd-maxsleep will
This is just cosmetic. There is no change to the tests.
Signed-off-by: Kevin Traynor
Reviewed-by: David Marchand
---
tests/pmd.at | 39 ---
1 file changed, 24 insertions(+), 15 deletions(-)
diff --git a/tests/pmd.at b/tests/pmd.at
index 374ad7217..4dd775bd3
Patches 1-4 are about polishing the existing functionality
and preparing for new functionality later in the series.
Patch 1 renames 'pmd-maxsleep' to 'pmd-sleep-max'. I think
this is allowed as it is still experimental. A warning will
be displayed if pmd-maxsleep is attempted to be used.
Patches
On 10 Jul 2023, at 16:26, Adrian Moreno wrote:
> On 7/10/23 16:19, Adrian Moreno wrote:
>>
>>
>> On 7/6/23 14:50, Eelco Chaudron wrote:
>>>
>>>
>>> On 2 Jun 2023, at 16:13, Adrian Moreno wrote:
>>>
Instead of relying on feature bits, use the speed value directly as
maximum rate for
On 6/2/23 16:13, Adrian Moreno wrote:
> Currently, htb rates are capped at ~34Gbps because they are internally
> expressed as 32-bit fields.
>
> Move min and max rates to 64-bit fields and use TCA_HTB_RATE64 and
> TCA_HTB_CEIL64 to configure HTC classes to break this barrier.
>
> In order to
On 7/10/23 16:19, Adrian Moreno wrote:
On 7/6/23 14:50, Eelco Chaudron wrote:
On 2 Jun 2023, at 16:13, Adrian Moreno wrote:
Instead of relying on feature bits, use the speed value directly as
maximum rate for htb and hfsc classes.
There is still a limitation with the maximum rate that
On 7/6/23 15:54, Eelco Chaudron wrote:
On 2 Jun 2023, at 16:13, Adrian Moreno wrote:
Currently, htb rates are capped at ~34Gbps because they are internally
expressed as 32-bit fields.
Move min and max rates to 64-bit fields and use TCA_HTB_RATE64 and
TCA_HTB_CEIL64 to configure HTC
On 7/6/23 14:50, Eelco Chaudron wrote:
On 2 Jun 2023, at 16:13, Adrian Moreno wrote:
Instead of relying on feature bits, use the speed value directly as
maximum rate for htb and hfsc classes.
There is still a limitation with the maximum rate that we can express
with a 32-bit number in
On 29/06/2023 15:30, Simon Horman wrote:
On Wed, Jun 14, 2023 at 02:36:39PM +0100, Kevin Traynor wrote:
other_config:pmd-maxsleep is a config option to allow
PMD thread cores to sleep under low or no load conditions.
Rename it to 'pmd-sleep-max' to allow a more structured
name and so that
Ilya Maximets, Jul 10, 2023 at 15:48:
> In order to speed up the process a bit, proposing a following incremental
> change:
>
> diff --git a/lib/netdev-dpdk.c b/lib/netdev-dpdk.c
> index 27b2fa5e0..aa87ee546 100644
> --- a/lib/netdev-dpdk.c
> +++ b/lib/netdev-dpdk.c
> @@ -2074,23 +2074,23 @@
On 7/4/23 21:59, Robin Jarry wrote:
> Some control protocols are used to maintain link status between
> forwarding engines (e.g. LACP). When the system is not sized properly,
> the PMD threads may not be able to process all incoming traffic from the
> configured Rx queues. When a signaling packet
On 07/07/2023 14:22, David Marchand wrote:
On Wed, Jun 21, 2023 at 11:26 AM Kevin Traynor wrote:
Extend 'pmd-sleep-max' so that individual PMD thread cores
may have a specified max sleep request value.
Any PMD thread core without a value will use the datapath default
(no sleep request) or
On 7/10/23 12:57, Dumitru Ceara wrote:
> On 7/6/23 13:00, Vladislav Odintsov wrote:
>>
>>
>>> On 5 Jul 2023, at 20:07, Vladislav Odintsov wrote:
>>>
>>> Hi Dumitru,
>>>
>>> thanks for the quick response!
>>>
On 5 Jul 2023, at 19:53, Dumitru Ceara wrote:
On 7/5/23 17:14, Vladislav
To achieve that add thread that will handle
statistics requests and delegate the processing
to defined functions. This allows the thread to
be flexible enough, so it could be extended in future
if needed.
At the same time connected the thread with the MAC
cache I-P node to have the timestamp
The node currently keeps track of MAC bindings
that have aging enabled. The purpose of this node
is to have structures that can be used by the
thread to update the timestamp when the MAC
binding is used. The I-P is generic enough
to be used by FDB timestamp refresh in very similar
way to the MAC
Synchrinoize the MAC binding age threashold to
SB datapath.
This is a preparation for the MAC binding refresh mechanism.
Signed-off-by: Ales Musil
---
northd/northd.c | 7 +++
1 file changed, 7 insertions(+)
diff --git a/northd/northd.c b/northd/northd.c
index 4e3c5d02a..7958304cf 100644
Populate and use the MAC cache table. The MAC cache has
the following entries:
ip,reg14=,metadata=,dl_src=,nw_src=
actions=drop
ipv6,reg14=,metadata=,dl_src=,ipv6_src=
actions=drop
The "mac_cache_use" action will resubmit packets from
"lr_in_learn_neighbor" table for MAC bindings that we
The mac_cache_use is just resubmit to MAC use cache table,
which will be later on used by the timestamp refresh
mechanism.
This is a preparation for the MAC binding refresh mechanism.
Signed-off-by: Ales Musil
---
include/ovn/actions.h | 3 +++
lib/actions.c | 17 +
The timestamp refresh mechanism is based on
separate table. The new thread will collect
flow statistics from this table and based on
"idle_age" data we can decide whether we should
bump the timestamp (MAC binding still in use)
or keep it as is.
The timestamp refresh mechanism works only if
the
On 7/6/23 13:00, Vladislav Odintsov wrote:
>
>
>> On 5 Jul 2023, at 20:07, Vladislav Odintsov wrote:
>>
>> Hi Dumitru,
>>
>> thanks for the quick response!
>>
>>> On 5 Jul 2023, at 19:53, Dumitru Ceara wrote:
>>>
>>> On 7/5/23 17:14, Vladislav Odintsov wrote:
Hi,
>>>
>>> Hi
If there is no route in table ip_routing, the route policy
item in table policy is useless.
Because route policy has a higher priority than ip routing,
so moddify the table id of IP_ROUTING and POLICY is a little
better.By this way, there is no need reroute any more, this
should be more
Have now sent out an invite to everyone participating in this thread.
For the benefit of anyone else wanting to attend I'm also sharing the
video link and other resources here:
Date/Time: Tuesday July 11th 13:30 UTC
Video link: https://meet.google.com/jno-ouvk-rfs
Meeting notes:
previously it was not possible to set the probe interval for the
connection from a relay to the backing ovsdb-server. With this change it
is now possible using the
`ovsdb-server/set-active-ovsdb-server-probe-interval` command.
The command `ovsdb-server/set-active-ovsdb-server-probe-interval` is
48 matches
Mail list logo