On Fri, Jul 7, 2023 at 1:56 PM wrote:
>
> From: Numan Siddique
>
> When a logical router gets updated due to load balancer
> or load balancer groups changes, it is now incrementally
> handled in the 'northd' engine node. Other logical router
> updates result in the full recompute of 'northd'
On Fri, Jul 7, 2023 at 1:55 PM wrote:
>
> From: Numan Siddique
>
> For every a given load balancer group 'A', northd engine data maintains
> a bitmap of datapaths associated to this lb group. So when lb group 'A'
> gets associated to a logical switch 's1', the bitmap index of 's1' is set
> in
On Fri, Jul 14, 2023 at 9:12 PM Numan Siddique wrote:
>
> On Thu, Jul 13, 2023 at 7:40 AM Han Zhou wrote:
> >
> > On Wed, Jul 12, 2023 at 11:54 PM Han Zhou wrote:
> > >
> > >
> > >
> > > On Sat, Jul 8, 2023 at 3:57 AM Mark Michelson
wrote:
> > > >
> > > > Hi Numan,
> > > >
> > > > I have one
Sorry, there was a typo.
regards,
Vladislav Odintsov
> On 17 Jul 2023, at 23:59, Vladislav Odintsov wrote:
>
> Hi Ilya,
>
> My usecase is to run ovsdb-server to serve hardware_vtep ovsdb schema in OVN
> controller vtep scenario. Container is used to run service on top of
> linux-based
Hi Ilya,
My usecase is to run ovsdb-server to serve hardware_vtep ovsdb schema in OVN
controller vtep scenario. Container is used to run service on top of
linux-based network switch.
For now, I just use this patchset internally and it seems sufficient to provide
a solution for that. The only
Hi, everyone.
branch-3.2 was just created. Please, test it and report issues!
Official release date, according to our release process, should be
on Thursday, August 17th (1 month from now).
Best regards, Ilya Maximets.
___
dev mailing list
On 7/17/23 16:03, Eelco Chaudron wrote:
>
>
> On 17 Jul 2023, at 15:22, Ilya Maximets wrote:
>
>> Signed-off-by: Ilya Maximets
>
>
> Changes look good to me!
>
> Acked-by: Eelco Chaudron
>
Applied to master. Thanks!
Best regards, Ilya Maximets.
On 7/17/23 16:03, Eelco Chaudron wrote:
>
>
> On 17 Jul 2023, at 15:22, Ilya Maximets wrote:
>
>> Signed-off-by: Ilya Maximets
>
> Changes look good to me!
>
> Acked-by: Eelco Chaudron
>
Thanks! Applied to master and a new branch-3.2.
Best regards, Ilya Maximets.
Hi Xavier,
For the series,
Acked-by: Mark Michelson
On 7/17/23 11:06, Xavier Simonart wrote:
Xavier Simonart (3):
tests: fixed typos in macro logs
binding: fixed ovn-installed not properly removed (migration)
binding: fixed ovn-installed not properly removed (recomputes)
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
On 7/17/23 10:08, Adrian Moreno wrote:
> There are currently two issues that limit our ability to configure QoS
> on certain cards in linux:
>
> 1) Firstly, the maximum link speed (which is used as maximum rate in
> some tc classes), is obtained from netdev feature bits.
> This is quite
On 7/17/23 11:06, Felix Huettner via dev wrote:
> 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-relay-source-probe-interval` command.
>
> Reviewed-by:
On 7/17/23 06:06, Mike Pattrick wrote:
> On Fri, Jul 14, 2023 at 7:31 AM Ilya Maximets wrote:
>>
>> On 7/11/23 14:46, Eelco Chaudron wrote:
>>>
>>>
>>> On 10 Jul 2023, at 17:34, Mike Pattrick wrote:
>>>
Currently OVS keeps track of which mirrors that each packet has been
sent to for the
If, while a port was claimed, a recompute is triggered at the same time
as the iface-id is removed, this was resulting in ovn-installed not
being properly removed.
Signed-off-by: Xavier Simonart
---
controller/binding.c | 8 ++
tests/ovn-macros.at | 9 +++
tests/ovn.at | 64
Fixed another potential race condition causing ovn-controller to
not properly remove ovn-install.
Before this patch, ovn-installed was not removed if an hypervisor
was receiving the port_binding->chassis change (to another hv)
within the same idl as the requested-chassis change. This caused
random
Fixes: 0794a6edf40b ("qos: fix potential double deletion of ovs idl row")
Signed-off-by: Xavier Simonart
---
tests/ovn-macros.at | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/tests/ovn-macros.at b/tests/ovn-macros.at
index 7c23db514..4898b0bc8 100644
---
Xavier Simonart (3):
tests: fixed typos in macro logs
binding: fixed ovn-installed not properly removed (migration)
binding: fixed ovn-installed not properly removed (recomputes)
controller/binding.c | 21 +-
controller/if-status.c | 19 +++--
controller/if-status.h | 2 +
On 17 Jul 2023, at 15:22, Ilya Maximets wrote:
> Signed-off-by: Ilya Maximets
Changes look good to me!
Acked-by: Eelco Chaudron
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
On 17 Jul 2023, at 15:22, Ilya Maximets wrote:
> Signed-off-by: Ilya Maximets
Changes look good to me!
Acked-by: Eelco Chaudron
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
On 6/22/23 14:51, Dumitru Ceara wrote:
> On 6/22/23 14:45, Dumitru Ceara wrote:
>> On 6/22/23 13:56, Aaron Conole wrote:
>>> Dumitru Ceara writes:
>>>
This avoids manual intervention when upstream ovn-kubernetes changes
its dependency versions.
Signed-off-by: Patryk Diak
On 17 Jul 2023, at 10:08, 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 bytes/s (~ 34.3Gbps), but using the
On 7/17/23 11:13, Ales Musil wrote:
> On Tue, Jun 20, 2023 at 3:59 PM Xavier Simonart wrote:
>
>> Before this patch, if iface did not exist when qos I+P was run the first
>> time
>> a qos queue was still created and qos I+P was not applying qos to OVS.
>> qos was potentially applied later,
On 7/17/23 11:10, Ales Musil wrote:
> On Thu, Jul 13, 2023 at 9:42 PM Dumitru Ceara wrote:
>
>> On 7/7/23 10:59, Ales Musil wrote:
>>> On Tue, Jun 20, 2023 at 4:26 PM Xavier Simonart
>> wrote:
>>>
If an interface with an qos option is deleted at the same
time as an ofport notification
On 7/14/23 08:41, Ales Musil wrote:
> On Thu, Jul 13, 2023 at 4:39 PM Dumitru Ceara wrote:
>
>> Regular dp-hash is not a canonical L4 hash (at least with the netlink
>> datapath). If the datapath supports l4 symmetrical dp-hash use that one
>> instead.
>>
>> Reported-at:
On 7/13/23 10:23, Ales Musil wrote:
> The requested and acked seqno values are allowed
> to be uint64_t, however the values that were added
> to the hmap were truncated to uint32_t. This would
> lead to loss of information when the value is bigger.
>
> Use uin64_t for the function signatures and
On 17 Jul 2023, at 10:08, Adrian Moreno wrote:
> Currently, the netdev's speed is being calculated by taking the link's
> feature bits (using netdev_get_features()) and transforming them into
> bps.
>
> This mechanism can be both inaccurate and difficult to maintain, mainly
> because we
On 7/14/23 19:11, Ilya Maximets wrote:
On 7/13/23 11:26, Viacheslav Galaktionov via dev wrote:
Currently, if the user wants to track related connections, they have to
specify a helper in all CT actions, which contradicts the behaviour
described in the documentation.
Fix this by using the
Signed-off-by: Ilya Maximets
---
NEWS | 4
configure.ac | 2 +-
debian/changelog | 6 ++
3 files changed, 11 insertions(+), 1 deletion(-)
diff --git a/NEWS b/NEWS
index b998fcf5c..077a34225 100644
--- a/NEWS
+++ b/NEWS
@@ -1,3 +1,7 @@
+Post-v3.2.0
+
Signed-off-by: Ilya Maximets
---
Documentation/faq/releases.rst | 1 +
NEWS | 2 +-
configure.ac | 2 +-
debian/changelog | 4 ++--
debian/rules | 4 ++--
5 files changed, 7 insertions(+), 6 deletions(-)
diff --git
The plan is to branch today by the EOD. There is still a couple of
patches on the list that can be applied before that.
Ilya Maximets (2):
Prepare for 3.2.0.
Prepare for post-3.2.0 (3.2.90).
Documentation/faq/releases.rst | 1 +
NEWS | 6 +-
configure.ac
On 7/14/23 20:06, Kevin Traynor wrote:
> Patches are about adding some visiblity and preparing for
> new functionality in a later series.
>
> Patch 1 renames 'pmd-maxsleep' to 'pmd-sleep-max'. To soften
> the change over, 'pmd-maxsleep' is deprecated and may still be used.
>
> v4
> - Drop
On 7/17/23 11:14, Eelco Chaudron wrote:
>
>
> On 15 Jul 2023, at 4:49, miter wrote:
>
>> Thank you for your reply.
>>
>> On 7/14/2023 11:01 PM, Ilya Maximets wrote:
>>> On 7/14/23 14:20, Eelco Chaudron wrote:
On 14 Jul 2023, at 13:52, Eelco Chaudron wrote:
> On 1 Jul 2023, at
On 15 Jul 2023, at 4:49, miter wrote:
> Thank you for your reply.
>
> On 7/14/2023 11:01 PM, Ilya Maximets wrote:
>> On 7/14/23 14:20, Eelco Chaudron wrote:
>>>
>>> On 14 Jul 2023, at 13:52, Eelco Chaudron wrote:
>>>
On 1 Jul 2023, at 16:40, mit...@outlook.com wrote:
> From: Lin
On Tue, Jun 20, 2023 at 3:59 PM Xavier Simonart wrote:
> Before this patch, if iface did not exist when qos I+P was run the first
> time
> a qos queue was still created and qos I+P was not applying qos to OVS.
> qos was potentially applied later, if/when (for any reason) runtime_data
> had to be
On Thu, Jul 13, 2023 at 9:42 PM Dumitru Ceara wrote:
> On 7/7/23 10:59, Ales Musil wrote:
> > On Tue, Jun 20, 2023 at 4:26 PM Xavier Simonart
> wrote:
> >
> >> If an interface with an qos option is deleted at the same
> >> time as an ofport notification from ovs (causing runtime_data recompute)
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-relay-source-probe-interval` command.
Reviewed-by: Simon Horman
Signed-off-by: Felix Huettner
---
v3:
- use a
On Wed, Jul 12, 2023 at 01:15:57PM +0200, Ilya Maximets wrote:
> On 7/10/23 08:19, Felix Huettner via dev wrote:
> > 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
> >
In preparation for supporting 64-bit rates in tc policies, move the
allocation and initialization of struct tc_police object inside
nl_msg_put_act_police(). That way, the function is now called with the
actual rates.
Acked-by: Eelco Chaudron
Reviewed-by: Simon Horman
Signed-off-by: Adrian
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 rates using 32-bit integers.
Acked-by: Eelco Chaudron
Reported-at: https://bugzilla.redhat.com/show_bug.cgi?id=2137643
It is equivalent to tc_policer_init() so remove the duplicated function.
Reviewed-by: Simon Horman
Acked-by: Eelco Chaudron
Signed-off-by: Adrian Moreno
---
lib/netdev-linux.c | 27 +++
1 file changed, 3 insertions(+), 24 deletions(-)
diff --git a/lib/netdev-linux.c
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 bytes/s (~ 34.3Gbps), but using the actual link speed
instead of the feature bits, we can at
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 test this, create a dummy tuntap device and set it's
tc uses these "rtab" tables to estimate the time (ticks) that it takes
to send a packet of different sizes. In preparation for the introduction
of 64-bit rates, add an argument to tc_put_rtab() to allow an external
64-bit rate.
Also use 64bits for other burst buffer calculation functions.
Currently, the netdev's speed is being calculated by taking the link's
feature bits (using netdev_get_features()) and transforming them into
bps.
This mechanism can be both inaccurate and difficult to maintain, mainly
because we currently use the feature bits supported by OpenFlow which
would
There are currently two issues that limit our ability to configure QoS
on certain cards in linux:
1) Firstly, the maximum link speed (which is used as maximum rate in
some tc classes), is obtained from netdev feature bits.
This is quite problematic because netdev feature bits do not keep up
with
On 7/12/23 17:28, Eelco Chaudron wrote:
On 12 Jul 2023, at 9:02, Adrian Moreno wrote:
Currently, the netdev's speed is being calculated by taking the link's
feature bits (using netdev_get_features()) and transforming them into
bps.
This mechanism can be both inaccurate and difficult to
46 matches
Mail list logo