Re: [ovs-dev] [PATCH ovn v2] northd: Fix address set incremental processing

On Tue, Jun 6, 2023 at 10:17 PM Han Zhou wrote: > > > On Thu, Jun 1, 2023 at 2:55 PM Han Zhou wrote: > > > > > > > > On Wed, May 31, 2023 at 11:16 PM Ales Musil wrote: > > > > > > > > > > > > On Thu, Jun 1, 2023 at 7:44 AM Han Zhou wrote: > > >> > > >> > > >> > > >> On Wed, May 24, 2023 at 11:

[ovs-dev] [PATCH ovn v3] northd: Fix address set incremental processing

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

[ovs-dev] [PATCH 1/2] utilities: add "--detach" option to ovs-ctl

Default is yes (current behavior). This means that after start process will be run in background. When "no" is given, process is run in foreground with `exec` call, which replaces current process (ovs-ctl) with wanted ovs process (ovsdb-server or ovs-vswitchd). This option is useful when running

[ovs-dev] [PATCH 2/2] utilities: add --in-db-ssl option to ovs-ctl

It is possible to parametrize ovs-ctl script to start ovsdb-server with DB_SCHEME other than Open_vSwitch. This scheme may not have currently required table "SSL" with "key", "cert" and "cacert" columns. This patch adds configuration knob "--in-db-ssl", which has default behavior as it is now: ru

[ovs-dev] [PATCH ovn v3] northd: Add logical flow to skip GARP with LLA

Skip GARP packet with link-local address being advertised when "always_learn_from_arp_request=false", this should prevent huge grow of MAC Binding table. To keep the option consistent overwrite the previous MAC with LLA if it was already stored in DB. Reported-at: https://bugzilla.redhat.com/22112

Re: [ovs-dev] [PATCH ovn v2] northd: Add logical flow to skip GARP with LLA

On Tue, Jun 6, 2023 at 11:37 PM Ihar Hrachyshka wrote: > On Tue, May 30, 2023 at 6:00 AM Ales Musil wrote: > >> Skip GARP packet with link-local address being advertised >> when "always_learn_from_arp_request=false", this should >> prevent huge grow of MAC Binding table. To keep the option >> co

Re: [ovs-dev] [PATCH net v2] net: openvswitch: fix upcall counter access before allocation

Thanks for fix this, in common enviroment, it's a small probability event. > Eelco Chaudron writes: > Currently, the per cpu upcall counters are allocated after the vport > is created and inserted into the system. This could lead to the > datapath accessing the counters before they are alloca

[ovs-dev] [PATCH ovn] call ovsrcu_exit() before exit in ovn-northd and ovn-controller to make valgrind happy

From: Igor Zhukov You can check logs by running: make check-valgrind TESTSUITEFLAGS="246" (Actually almost every test affected but for example we need one test) Valgrind message looks like ==65437== 304 bytes in 1 blocks are possibly lost in loss record 265 of 289 ==65437==at 0x483DD99: ca

Re: [ovs-dev] [PATCH 6/6] ovsdb: monitor: Keep and maintain the initial change set.

On Tue, Jun 6, 2023 at 3:08 PM Ilya Maximets wrote: > > On 6/6/23 22:12, Han Zhou wrote: > > > > > > On Tue, Jun 6, 2023 at 12:10 PM Ilya Maximets mailto:i.maxim...@ovn.org>> wrote: > >> > >> On 3/27/23 21:43, Ilya Maximets wrote: > >> > Change sets in OVSDB monitor are storing all the changes tha

Re: [ovs-dev] [PATCH 6/6] ovsdb: monitor: Keep and maintain the initial change set.

On 6/6/23 22:12, Han Zhou wrote: > > > On Tue, Jun 6, 2023 at 12:10 PM Ilya Maximets > wrote: >> >> On 3/27/23 21:43, Ilya Maximets wrote: >> > Change sets in OVSDB monitor are storing all the changes that happened >> > between a particular transaction ID and now.  Ini

Re: [ovs-dev] [PATCH net v2] net: openvswitch: fix upcall counter access before allocation

Eelco Chaudron writes: > Currently, the per cpu upcall counters are allocated after the vport is > created and inserted into the system. This could lead to the datapath > accessing the counters before they are allocated resulting in a kernel > Oops. > > Here is an example: > > PID: 59693TAS

Re: [ovs-dev] [PATCH ovn v2] northd: Add logical flow to skip GARP with LLA

On Tue, May 30, 2023 at 6:00 AM Ales Musil wrote: > Skip GARP packet with link-local address being advertised > when "always_learn_from_arp_request=false", this should > prevent huge grow of MAC Binding table. To keep the option > consistent overwrite the previous MAC with LLA if it was > already

Re: [ovs-dev] [PATCH net] net: openvswitch: fix upcall counter access before allocation

Eelco Chaudron writes: > Currently, the per cpu upcall counters are allocated after the vport is > created and inserted into the system. This could lead to the datapath > accessing the counters before they are allocated resulting in a kernel > Oops. > > Here is an example: > > PID: 59693TAS

Re: [ovs-dev] [PATCH ovn v2] northd: Fix address set incremental processing

On Thu, Jun 1, 2023 at 2:55 PM Han Zhou wrote: > > > > On Wed, May 31, 2023 at 11:16 PM Ales Musil wrote: > > > > > > > > On Thu, Jun 1, 2023 at 7:44 AM Han Zhou wrote: > >> > >> > >> > >> On Wed, May 24, 2023 at 11:27 PM Ales Musil wrote: > >> > > >> > The incremental processing is broken for

Re: [ovs-dev] [PATCH 6/6] ovsdb: monitor: Keep and maintain the initial change set.

On Tue, Jun 6, 2023 at 12:10 PM Ilya Maximets wrote: > > On 3/27/23 21:43, Ilya Maximets wrote: > > Change sets in OVSDB monitor are storing all the changes that happened > > between a particular transaction ID and now. Initial change set > > basically contains all the data. > > > > On each monit

Re: [ovs-dev] [PATCH 6/6] ovsdb: monitor: Keep and maintain the initial change set.

On 6/6/23 21:10, Ilya Maximets wrote: > On 3/27/23 21:43, Ilya Maximets wrote: >> Change sets in OVSDB monitor are storing all the changes that happened >> between a particular transaction ID and now. Initial change set >> basically contains all the data. >> >> On each monitor request a new initia

Re: [ovs-dev] [PATCH 6/6] ovsdb: monitor: Keep and maintain the initial change set.

On 3/27/23 21:43, Ilya Maximets wrote: > Change sets in OVSDB monitor are storing all the changes that happened > between a particular transaction ID and now. Initial change set > basically contains all the data. > > On each monitor request a new initial change set is created by creating > an emp

Re: [ovs-dev] [PATCH ovn branch-22.09 1/5] Track ip version of tunnel in chassis_tunnel struct

For the series: Acked-by: Mark Michelson On 5/31/23 16:10, Ihar Hrachyshka wrote: This will be used in a later patch to calculate tunneling overhead for effective path MTU. Acked-by: Dumitru Ceara Signed-off-by: Ihar Hrachyshka Signed-off-by: Mark Michelson (cherry picked from commit 93503

Re: [ovs-dev] [PATCH ovn branch-22.12 1/5] Track ip version of tunnel in chassis_tunnel struct

For the entire series: Acked-by: Mark Michelson On 5/31/23 16:08, Ihar Hrachyshka wrote: This will be used in a later patch to calculate tunneling overhead for effective path MTU. Acked-by: Dumitru Ceara Signed-off-by: Ihar Hrachyshka Signed-off-by: Mark Michelson (cherry picked from commi

Re: [ovs-dev] [PATCH ovn branch-23.03 1/5] Track ip version of tunnel in chassis_tunnel struct

For the series: Acked-by: Mark Michelson On 5/31/23 16:04, Ihar Hrachyshka wrote: This will be used in a later patch to calculate tunneling overhead for effective path MTU. Acked-by: Dumitru Ceara Signed-off-by: Ihar Hrachyshka Signed-off-by: Mark Michelson (cherry picked from commit 93503

Re: [ovs-dev] |fail| pw1791271 [PATCH ovn] ovn-controller: Reduce size of the SB monitor condition.

On 6/6/23 20:00, 0-day Robot wrote: > From: ro...@bytheb.org > > Test-Label: github-robot: ovn-kubernetes > Test-Status: fail > http://patchwork.ozlabs.org/api/patches/1791271/ > > _github build: failed_ > Build URL: https://github.com/ovsrobot/ovn/actions/runs/5190342903 > Build Logs: The actua

[ovs-dev] [PATCH v7] tc: fix crash on malformed reply from kernel.

The tc module combines the use of the `tc_transact` helper function for communication with the in-kernel tc infrastructure with assertions on the reply data by `ofpbuf_at_assert` on the received data prior to further processing. With the presence of bugs on the kernel side, we need to treat the ke

Re: [ovs-dev] [PATCH v6] tc: fix crash on malformed reply from kernel.

On 6/6/23 18:08, Frode Nordahl wrote: > On Tue, Jun 6, 2023 at 5:46 PM Ilya Maximets wrote: >> >> On 6/6/23 11:38, Frode Nordahl wrote: >>> The tc module combines the use of the `tc_transact` helper >>> function for communication with the in-kernel tc infrastructure >>> with assertions on the repl

Re: [ovs-dev] [PATCH v6] tc: fix crash on malformed reply from kernel.

On Tue, Jun 6, 2023 at 5:46 PM Ilya Maximets wrote: > > On 6/6/23 11:38, Frode Nordahl wrote: > > The tc module combines the use of the `tc_transact` helper > > function for communication with the in-kernel tc infrastructure > > with assertions on the reply data by `ofpbuf_at_assert` on the > > re

Re: [ovs-dev] [PATCH v6] tc: fix crash on malformed reply from kernel.

On 6/6/23 11:38, Frode Nordahl wrote: > The tc module combines the use of the `tc_transact` helper > function for communication with the in-kernel tc infrastructure > with assertions on the reply data by `ofpbuf_at_assert` on the > received data prior to further processing. > > With the presence o

[ovs-dev] [PATCH ovn] ovn-controller: Reduce size of the SB monitor condition.

We don't need to explicitly add port bindings that were already bound locally. We implicitly get those because we monitor the datapaths they're attached to. When performing an ovn-heater 500-node density-heavy scale test [0], with conditional monitoring enabled, the unreasonably long poll interva

Re: [ovs-dev] [PATCH v4] ofproto-dpif-xlate: Remove repeated function for judge garp

On Mon, Feb 20, 2023 at 5:05 AM Han Ding wrote: > > > Function is_gratuitous_arp() and function is_garp() are all used to judge > whether the flow is gratuitous arp. It is not necessary to use two functions > to do the same thing and just keep one. > > Gratuitous ARP message is a generally link-le

Re: [ovs-dev] [PATCH net v2] net: openvswitch: fix upcall counter access before allocation

On Tue, Jun 06, 2023 at 01:56:35PM +0200, Eelco Chaudron wrote: > Currently, the per cpu upcall counters are allocated after the vport is > created and inserted into the system. This could lead to the datapath > accessing the counters before they are allocated resulting in a kernel > Oops. > > Her

Re: [ovs-dev] [PATCH ovn] Remove lport from related_lports if there is no binding

Hi Priyankar, Mark Thanks for the patch. I agree with Mark - the description is really great ! Based on your description, I tried creating a unit-test reproducing the issue, and checking that your patch fixes it. I came up with [0]. It reproduces some issues (flows not deleted) on origin/main, and

Re: [ovs-dev] [PATCH v6 2/2] netdev-offload-dpdk: replace action PORT_ID with REPRESENTED_PORT

Bleep bloop. Greetings Ivan Malov, 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: Inappropriate bracing around statement #28 FILE: lib/netdev-offload-dpdk.c:743: if

Re: [ovs-dev] [PATCH v6 1/2] netdev-dpdk: negotiate delivery of per-packet Rx metadata

Bleep bloop. Greetings Ivan Malov, 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: Inappropriate bracing around statement #37 FILE: lib/netdev-dpdk.c:1152: if (dev->rx_m

Re: [ovs-dev] [PATCH net] net: openvswitch: fix upcall counter access before allocation

On 6 Jun 2023, at 13:43, Paolo Abeni wrote: > On Mon, 2023-06-05 at 16:38 +0200, Simon Horman wrote: >> On Mon, Jun 05, 2023 at 03:53:59PM +0200, Eelco Chaudron wrote: Yeah, I see that. And I might have done the same thing. But, OTOH, this change is making the error path more complex

[ovs-dev] [PATCH net v2] net: openvswitch: fix upcall counter access before allocation

Currently, the per cpu upcall counters are allocated after the vport is created and inserted into the system. This could lead to the datapath accessing the counters before they are allocated resulting in a kernel Oops. Here is an example: PID: 59693TASK: 0005f4f51500 CPU: 0COMMAND:

Re: [ovs-dev] [PATCH net] net: openvswitch: fix upcall counter access before allocation

On Mon, Jun 05, 2023 at 03:53:59PM +0200, Eelco Chaudron wrote: > > > On 5 Jun 2023, at 15:07, Simon Horman wrote: > > > On Mon, Jun 05, 2023 at 02:54:35PM +0200, Eelco Chaudron wrote: > >> > >> > >> On 5 Jun 2023, at 14:41, Simon Horman wrote: > >> > >>> On Mon, Jun 05, 2023 at 10:59:50AM +0200

Re: [ovs-dev] [PATCH net] net: openvswitch: fix upcall counter access before allocation

On Mon, 2023-06-05 at 16:38 +0200, Simon Horman wrote: > On Mon, Jun 05, 2023 at 03:53:59PM +0200, Eelco Chaudron wrote: > > > Yeah, I see that. And I might have done the same thing. > > > But, OTOH, this change is making the error path more complex > > > (or at least more prone to error). > > > >

[ovs-dev] [PATCH v6 2/2] netdev-offload-dpdk: replace action PORT_ID with REPRESENTED_PORT

Action PORT_ID has been deprecated. Use REPRESENTED_PORT instead. Signed-off-by: Ivan Malov --- lib/netdev-offload-dpdk.c | 31 +-- 1 file changed, 17 insertions(+), 14 deletions(-) diff --git a/lib/netdev-offload-dpdk.c b/lib/netdev-offload-dpdk.c index 14bc87771..e

[ovs-dev] [PATCH v6 0/2] DPDK: align flow offloads with 22.11

Address two problems using two corresponding features in DPDK, which have been around for a year and are now stable: 1) The need to make sure that metadata generated by flow rule execution be delivered from NIC to application; 2) Replacing PORT_ID action with REPRESENTED_PORT; --- v2: amendme

[ovs-dev] [PATCH v6 1/2] netdev-dpdk: negotiate delivery of per-packet Rx metadata

This may be required by some PMDs in offload scenarios. Signed-off-by: Ivan Malov --- lib/netdev-dpdk.c | 50 +++ 1 file changed, 50 insertions(+) diff --git a/lib/netdev-dpdk.c b/lib/netdev-dpdk.c index 8cb1a7703..98765fe6e 100644 --- a/lib/netdev-dp

Re: [ovs-dev] [ovn] ha-chassis-group false positive failover

Hi Han, thanks for your inputs! Please see comments inline. > On 5 Jun 2023, at 21:01, Han Zhou wrote: > > On Mon, Jun 5, 2023 at 8:58 AM Vladislav Odintsov > wrote: >> >> Hi Dumitru, Ilya, Han, >> >> do you guys have any comments here? >> I’m gonna start working on

[ovs-dev] [PATCH v6] tc: fix crash on malformed reply from kernel.

The tc module combines the use of the `tc_transact` helper function for communication with the in-kernel tc infrastructure with assertions on the reply data by `ofpbuf_at_assert` on the received data prior to further processing. With the presence of bugs on the kernel side, we need to treat the ke

Re: [ovs-dev] [PATCH v5 2/2] netdev-offload-dpdk: replace action PORT_ID with REPRESENTED_PORT

Bleep bloop. Greetings Ivan Malov, 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: Inappropriate bracing around statement #28 FILE: lib/netdev-offload-dpdk.c:743: if

[ovs-dev] [PATCH v5 2/2] netdev-offload-dpdk: replace action PORT_ID with REPRESENTED_PORT

Action PORT_ID has been deprecated. Use REPRESENTED_PORT instead. Signed-off-by: Ivan Malov --- lib/netdev-offload-dpdk.c | 31 +-- 1 file changed, 17 insertions(+), 14 deletions(-) diff --git a/lib/netdev-offload-dpdk.c b/lib/netdev-offload-dpdk.c index 14bc87771..e

[ovs-dev] [PATCH v5 0/2] DPDK: align flow offloads with 22.11

Address two problems using two corresponding features in DPDK, which have been around for a year and are now stable: 1) The need to make sure that metadata generated by flow rule execution be delivered from NIC to application; 2) Replacing PORT_ID action with REPRESENTED_PORT. --- v2: amendme

[ovs-dev] [PATCH v5 1/2] netdev-dpdk: negotiate delivery of per-packet Rx metadata

This may be required by some PMDs in offload scenarios. Signed-off-by: Ivan Malov --- lib/netdev-dpdk.c | 43 +++ 1 file changed, 43 insertions(+) diff --git a/lib/netdev-dpdk.c b/lib/netdev-dpdk.c index 8cb1a7703..774ca8d51 100644 --- a/lib/netdev-dpdk.c

Re: [ovs-dev] [PATCH v4 3/3] netdev-offload-dpdk: use flow transfer proxy mechanism

Hi Eli, Thanks for responding. PSB. On Tue, 6 Jun 2023, Eli Britstein wrote: -Original Message- From: Ivan Malov Sent: Sunday, 4 June 2023 15:58 To: Eli Britstein Cc: ovs-dev@openvswitch.org; Ilya Maximets ; Ori Kam ; David Marchand Subject: RE: [PATCH v4 3/3] netdev-offload-dpdk