On Wed, May 24, 2023 at 8:31 PM Ilya Maximets wrote:
> On 5/19/23 09:36, Ales Musil wrote:
> > Use the backtrace functions that is provided by libc,
> > this allows us to get backtrace that is independent of
> > the current memory map of the process. Which in turn can
> > be used for debugging/tr
Use the backtrace functions that is provided by libc,
this allows us to get backtrace that is independent of
the current memory map of the process. Which in turn can
be used for debugging/tracing purpose. The backtrace
is not 100% accurate due to various optimizations, most
notably "-fomit-frame-po
On Wed, May 24, 2023 at 1:11 PM Ilya Maximets wrote:
> On 5/23/23 10:09, Ales Musil wrote:
> > The incremental processing is broken for addresses
> > that have mask which could "erase" portion of the address
> > itself e.g. 10.16.0.47/4, after applying the mask with token
> > parser the address b
The incremental processing is broken for addresses
that have mask which could "erase" portion of the address
itself e.g. 10.16.0.47/4, after applying the mask with token
parser the address becomes 0.0.0.0/4, which is fine for
logical flows. However, for the deletion/addition to database
we need the
ons. 24. mai 2023, 21:31 skrev Ilya Maximets :
> On 5/23/23 12:39, Frode Nordahl wrote:
>
>
>
> > int
> > tc_transact(struct ofpbuf *request, struct ofpbuf **replyp)
> > {
> > int error = nl_transact(NETLINK_ROUTE, request, replyp);
> > ofpbuf_uninit(request);
> > +
> > +if (!err
On Wed, May 24, 2023 at 7:11 AM Ilya Maximets wrote:
>
> On 5/23/23 10:09, Ales Musil wrote:
> > The incremental processing is broken for addresses
> > that have mask which could "erase" portion of the address
> > itself e.g. 10.16.0.47/4, after applying the mask with token
> > parser the address
Hi Lorenzo,
I tried running the system test from this patch against ovn main and it
passes even without the changes in northd. I think this is because the
alice1 and foo1 ports need to be bound on different hvs to properly test
the changes in northd.
See below for other comments.
On 5/24/23
On 5/23/23 12:39, Frode Nordahl wrote:
> int
> tc_transact(struct ofpbuf *request, struct ofpbuf **replyp)
> {
> int error = nl_transact(NETLINK_ROUTE, request, replyp);
> ofpbuf_uninit(request);
> +
> +if (!error && replyp && !(*replyp)->size) {
Not a full review, but shouldn'
On 5/24/23 15:39, Frode Nordahl wrote:
> The bareudp tests depend on specific kernel configuration to
> succeed. Skip the test if the feature is not enabled in the
> running kernel.
>
> Signed-off-by: Frode Nordahl
> ---
> tests/system-kmod-macros.at | 10 ++
> tests/system-layer3-
On 24 May 2023, at 21:05, Ilya Maximets wrote:
> On 5/23/23 15:11, Balazs Nemeth wrote:
>> The only way that stats->{n_packets,n_bytes} would decrease is due to an
>> overflow, or if there are bugs in how statistics are handled. In the
>> past, there were multiple issues that caused a jump back
On 5/23/23 15:11, Balazs Nemeth wrote:
> The only way that stats->{n_packets,n_bytes} would decrease is due to an
> overflow, or if there are bugs in how statistics are handled. In the
> past, there were multiple issues that caused a jump backward. A
> workaround was in place to set the statistics
On 5/19/23 09:36, Ales Musil wrote:
> Use the backtrace functions that is provided by libc,
> this allows us to get backtrace that is independent of
> the current memory map of the process. Which in turn can
> be used for debugging/tracing purpose. The backtrace
> is not 100% accurate due to variou
On 5/16/23 13:59, Kevin Traynor wrote:
> By default OVS configures 2048 descriptors for tx and rx queues
> on DPDK devices. It also allows the user to configure those values.
>
> If the values used are not acceptable to the device then queue setup
> would fail.
>
> The device exposes it's max/min
In the current codebase for distributed gw router port use-case,
it is not possible to add a load balancer that redirects the traffic
to a back-end if it is used as internal IP of a FIP NAT rule since
the reply traffic is never centralized. Fix the issue centralizing the
traffic if it is the reply
On Fri, May 19, 2023 at 05:13:23PM +0200, 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-active-ovsdb-server-probe-interval` c
Thanks Lorenzo,
a few suggestions from my side.
On Thu, May 18, 2023 at 10:20:34PM +0200, Lorenzo Bianconi wrote:
> In the current codebase for distributed gw router port use-case,
> it is not possible to add a load balancer that redirects the traffic
> to a backed if it is even the internal IP o
Kevin Traynor, May 24, 2023 at 16:54:
> > The "rss" item should be mandatory anyway. There should be no way to
> > disable it. I assume that it is what Ilya meant with these "+" symbols.
> > That would leave us with these examples:
> >
> > # lacp+bfd on separate queue, rss on other queues
> >
> >
On 5/24/23 16:54, Kevin Traynor wrote:
> On 24/05/2023 15:32, Robin Jarry wrote:
>> Kevin Traynor, May 24, 2023 at 16:06:
>>> Hmm, not sure on this one. I have a feeling that having a 'hash' mode
>>> for tx-steering that only applies to vhost devices, and 'hash' mode
>>> for rx-steering that only a
On Thu, May 18, 2023 at 04:08:34PM -0400, Mike Pattrick wrote:
> 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()
Hey Kevin,
Kevin Traynor, May 24, 2023 at 16:13:
> Hi Robin,
>
> I tested combinations of enabling/disabling cp-proto and enabling
> hwol. It is working as expected and hwol feature always has a clear
> priority, regardless of the order they are enabled in.
>
> I didn't test lacp traffic, but I wa
On 24/05/2023 15:32, Robin Jarry wrote:
Kevin Traynor, May 24, 2023 at 16:06:
Hmm, not sure on this one. I have a feeling that having a 'hash' mode
for tx-steering that only applies to vhost devices, and 'hash' mode
for rx-steering that only applies to NICs means people will miss the
subtlety an
On Wed, May 24, 2023 at 03:39:53PM +0200, Frode Nordahl wrote:
> The bareudp tests depend on specific kernel configuration to
> succeed. Skip the test if the feature is not enabled in the
> running kernel.
>
> Signed-off-by: Frode Nordahl
Tested-by: Simon Horman
Reviewed-by: Simon Horman
___
Kevin Traynor, May 24, 2023 at 16:06:
> Hmm, not sure on this one. I have a feeling that having a 'hash' mode
> for tx-steering that only applies to vhost devices, and 'hash' mode
> for rx-steering that only applies to NICs means people will miss the
> subtlety and try to enable the wrong hash mode
On 17/04/2023 13:37, 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 of
On 24/05/2023 11:41, Ilya Maximets wrote:
On 5/24/23 09:18, Robin Jarry wrote:
Ilya Maximets, May 23, 2023 at 22:04:
'rxq-isolate' will be confused with 'other_config:pmd-rxq-isolate'.
Same likely goes for the 'isolated-rxq'.
'rxq-steernig' may be confused with 'other_config:tx-steering'.
But
The bareudp tests depend on specific kernel configuration to
succeed. Skip the test if the feature is not enabled in the
running kernel.
Signed-off-by: Frode Nordahl
---
tests/system-kmod-macros.at | 10 ++
tests/system-layer3-tunnels.at | 2 ++
tests/system-userspace-macros.at
On 5/23/23 22:26, Ilya Maximets wrote:
On 5/23/23 20:51, Sayali Naval (sanaval) via dev wrote:
Does anyone have any insights on the below?
From: Sayali Naval (sanaval)
Sent: Friday, May 12, 2023 11:10 AM
To: d...@openvswitch.org
Subject: [ovs-dev] ovs-vsctl Br
On 5/17/23 05:11, Mike Pattrick wrote:
> From: Flavio Leitner
>
> The netdev receiving packets is supposed to provide the flags
> indicating if the IP checksum was verified and it is GOOD or BAD,
> otherwise the stack will check when appropriate by software.
>
> If the packet comes with good che
On 5/17/23 05:11, Mike Pattrick wrote:
> From: Flavio Leitner
>
> The netdev receiving packets is supposed to provide the flags
> indicating if the L4 checksum was verified and it is OK or BAD,
> otherwise the stack will check when appropriate by software.
>
> If the packet comes with good check
On 5/17/23 05:11, Mike Pattrick wrote:
> From: Flavio Leitner
>
> This patch modifies netdev_get_status to include information about
> checksum offload status by port, allowing the user to gain insight into
> where checksum offloading is active.
>
> Signed-off-by: Flavio Leitner
> Co-authored-b
On 5/17/23 05:11, Mike Pattrick wrote:
> From: Flavio Leitner
>
> Document the implementation of netdev hardware offloading
> in userspace datapath.
>
> Signed-off-by: Flavio Leitner
> Co-authored-by: Mike Pattrick
> Signed-off-by: Mike Pattrick
> ---
> Since v9:
> - Renamed documentation
On Tue, May 16, 2023 at 12:59:58PM +0100, Kevin Traynor wrote:
> By default OVS configures 2048 descriptors for tx and rx queues
> on DPDK devices. It also allows the user to configure those values.
>
> If the values used are not acceptable to the device then queue setup
> would fail.
>
> The dev
On Tue, May 16, 2023 at 12:59:57PM +0100, Kevin Traynor wrote:
> There is no need to display 'requested_rx/tx_descriptors' and
> 'configured_rx/tx_descriptors' as they will be the same.
>
> It is simpler to just have a single 'n_rxq/txq_desc' value.
>
> Suggested-by: Ilya Maximets
> Reviewed-by:
On Tue, May 16, 2023 at 2:00 PM Kevin Traynor wrote:
>
> By default OVS configures 2048 descriptors for tx and rx queues
> on DPDK devices. It also allows the user to configure those values.
>
> If the values used are not acceptable to the device then queue setup
> would fail.
>
> The device expos
On 5/23/23 10:09, Ales Musil wrote:
> The incremental processing is broken for addresses
> that have mask which could "erase" portion of the address
> itself e.g. 10.16.0.47/4, after applying the mask with token
> parser the address becomes 0.0.0.0/4, which is fine for
> logical flows. However, for
On Mon, May 22, 2023 at 11:18 PM Vladislav Odintsov
wrote:
> Depending on the udp service, it can reply with some udp data.
> In that case ovn-controller will warn with next message:
>
> pinctrl(ovn_pinctrl0)|WARN|handle service check: Unsupported protocol -
> [11]
>
> With this patch ovn-control
On 5/24/23 09:18, Robin Jarry wrote:
> Ilya Maximets, May 23, 2023 at 22:04:
>> 'rxq-isolate' will be confused with 'other_config:pmd-rxq-isolate'.
>> Same likely goes for the 'isolated-rxq'.
>>
>> 'rxq-steernig' may be confused with 'other_config:tx-steering'.
>> But this can be argued that it's e
On 23 May 2023, at 15:11, Balazs Nemeth wrote:
> The only way that stats->{n_packets,n_bytes} would decrease is due to an
> overflow, or if there are bugs in how statistics are handled. In the
> past, there were multiple issues that caused a jump backward. A
> workaround was in place to set the
Using cache improves performance of recomputation of lflows(by
about 30%)
Exising lflow cache for `matches` and `expressions` is adopted
to include `actions` as well.
Co-authored-by: Felix Huettner
Signed-off-by: Felix Huettner
Signed-off-by: Ihtisham ul Haq
---
controller/lflow-cache.c
Bleep bloop. Greetings Ihtisham ul Haq, 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:
ERROR: Co-author Felix Huettner needs to sign off.
ERROR: Improper whitespace around contro
Using cache improves performance of recomputation of lflows. We see
about 30% improvments during our tests
Exising lflow cache for `matches` and `expressions` is adopted
to include `actions` as well.
Co-authored-by: Felix Huettner
Signed-off-by: Ihtisham ul Haq
---
controller/lflow-cache.c
On 23/05/2023 19:06, Simon Horman wrote:
> On Mon, May 15, 2023 at 11:23:56AM +0300, Roi Dayan via dev wrote:
>> From: Gavin Li
>>
>> Add TC offload support for vxlan encap with gbp option
>>
>> Signed-off-by: Gavin Li
>> Reviewed-by: Gavi Teitz
>> Reviewed-by: Roi Dayan
>
> Reviewed-by: Si
Hi everyone,
Ilya mentioned to me that you will want to bring openstack examples to
ovn-heater.
I wanted to ask how to best join this effort. It would be great for us
to have a tool to scale test ovn. Also we have an ovn deployment with
currently 450 compute nodes where we can try to extract the
Ilya Maximets, May 23, 2023 at 22:04:
> 'rxq-isolate' will be confused with 'other_config:pmd-rxq-isolate'.
> Same likely goes for the 'isolated-rxq'.
>
> 'rxq-steernig' may be confused with 'other_config:tx-steering'.
> But this can be argued that it's essentially similar functionality,
> so shoul
44 matches
Mail list logo
