Re: [ovs-dev] [PATCH v6 2/2] handlers: Fix handlers mapping

On 7/19/22 08:59, Michael Santana wrote: > The handler and CPU mapping in upcalls are incorrect, and this is > specially noticeable systems with cpu isolation enabled. > > Say we have a 12 core system where only every even number CPU is enabled > C0, C2, C4, C6, C8, C10 > > This means we will

Re: [ovs-dev] [PATCH ovn v3] multicast: Properly flood IGMP queries and reports.

On Tue, Aug 2, 2022 at 10:46 AM Dumitru Ceara wrote: > > RFC 4541 "Considerations for Internet Group Management Protocol (IGMP) > and Multicast Listener Discovery (MLD) Snooping Switches" [0] states: > > For IGMP packets: > 2.1.1. IGMP Forwarding Rules >1) A snooping switch should

Re: [ovs-dev] [PATCH v6 1/2] handlers: Create additional handler threads when using CPU isolation

On 7/19/22 08:59, Michael Santana wrote: > Additional threads are required to service upcalls when we have CPU > isolation (in per-cpu dispatch mode). The reason additional threads > are required is because it creates a more fair distribution. With more > threads we decrease the load of each

Re: [ovs-dev] [PATCH ovn v2] northd: handle virtual lport type update

On Tue, Aug 2, 2022 at 10:22 AM Mohammad Heib wrote: > > ovn-northd re-create sbrec row for lport of type > virtual when the type explicitly updated in NBDB. > > This approach was applied to handle container lport > type updates and avoid handling issues in the ovn-controller, > this patch

Re: [ovs-dev] [PATCH ovn v2 0/2] Fix IPv6 prefix delegation and the test flakiness

On Mon, Jul 25, 2022 at 6:50 AM Ales Musil wrote: > > > > Ales Musil (2): > system-tests: Reduce flakiness of IPv6 prefix delegation > controller: Fix IPv6 prefix delegation > Thanks. Applied both the patches to main and backported upto branch-22.03. Numan > controller/binding.c|

Re: [ovs-dev] [PATCH v2] dpif-netdev: Simplify AVX512 build time checks to enhance readability.

On 8/2/22 22:32, Ilya Maximets wrote: > On 8/2/22 09:47, Pai G, Sunil wrote: -Original Message- From: Pai G, Sunil Sent: Monday 4 July 2022 13:27 To: d...@openvswitch.org Cc: echau...@redhat.com; Finn, Emma ; Van Haaren, Harry ; Ferriter, Cian

Re: [ovs-dev] [PATCH v2] dpif-netdev: Simplify AVX512 build time checks to enhance readability.

On 8/2/22 09:47, Pai G, Sunil wrote: >>> -Original Message- >>> From: Pai G, Sunil >>> Sent: Monday 4 July 2022 13:27 >>> To: d...@openvswitch.org >>> Cc: echau...@redhat.com; Finn, Emma ; Van Haaren, >>> Harry ; Ferriter, Cian >>> >>> Subject: [PATCH v2] dpif-netdev: Simplify AVX512

[ovs-dev] [PATCHv2] bfd: Add IPv6 address support to BFD packets

Added config options for source and destination IPv6 addresses to BFD packets. IPv6 address presence overrides IPv4 address and packets are crafted as IPv6. Added IP address values to status to ease setting up IP addresses. Signed-off-by: Toms Atteka --- lib/bfd.c| 119

[ovs-dev] [PATCH] bfd: Add IPv6 address support to BFD packets

Added config options for source and destination IPv6 addresses to BFD packets. IPv6 address presence overrides IPv4 address and packets are crafted as IPv6. Added IP address values to status to ease setting up IP addresses. Signed-off-by: Toms Atteka --- lib/bfd.c| 119

[ovs-dev] [OVN PATCH] northd: add unsnat/undnat lflow for established connections

snat/dnat rules of logical router, no logical flows for established connection, all natted packets deliver to kernel conntrack module by ct commit, this is low performance and difficult to offload. add another logical flows without ct commit forestablished on pipeline stage of unsnat/undnat for

Re: [ovs-dev] [PATCH ovn] ofctrl.c: mff_ovn_geneve should be available at state WAIT_BEFORE_CLEAR.

On Wed, Jul 27, 2022 at 9:06 PM Han Zhou wrote: > > When introducing the feature ovn-ofctrl-wait-before-clear, a new state > WAIT_BEFORE_CLEAR is added before CLEAR_FLOWS. In the WAIT_BEFORE_CLEAR > state the mff_ovn_geneve is already available. However, the function > ofctrl_get_mf_field_id()

[ovs-dev] [PATCH] netdev-dpdk: fix tx_dropped counters value

Packets that could not be transmitted because the TXQ are full should be taken into account in the global ovs_tx_failure_drops as it was the case before commit 29b94e12d57d ("netdev-dpdk: Refactor the DPDK transmit path."). Also, the global tx_dropped should take all ovs_*_tx_drops counters into

[ovs-dev] [PATCH ovn v3] multicast: Properly flood IGMP queries and reports.

RFC 4541 "Considerations for Internet Group Management Protocol (IGMP) and Multicast Listener Discovery (MLD) Snooping Switches" [0] states: For IGMP packets: 2.1.1. IGMP Forwarding Rules 1) A snooping switch should forward IGMP Membership Reports only to those ports where

Re: [ovs-dev] [PATCH ovn v2 2/2] multicast: Properly flood IGMP queries and reports.

On Tue, Aug 2, 2022 at 4:19 PM Dumitru Ceara wrote: > On 7/20/22 10:13, Ales Musil wrote: > > Hi Dumitru, > > > > Hi Ales, > > > the code looks good to me, with one small nit see below. > > > > Thanks for the review! > > > On Tue, Jul 19, 2022 at 10:43 PM Dumitru Ceara > wrote: > > > >> RFC

[ovs-dev] [PATCH ovn v2] northd: handle virtual lport type update

ovn-northd re-create sbrec row for lport of type virtual when the type explicitly updated in NBDB. This approach was applied to handle container lport type updates and avoid handling issues in the ovn-controller, this patch expanded that approach to handle lport of type virtual in the same way.

Re: [ovs-dev] [PATCH ovn] northd: handle virtual lport type update

On Tue, Aug 2, 2022 at 5:00 PM Dumitru Ceara wrote: > On 8/2/22 15:56, Mohammad Heib wrote: > > HI Dumitru, > > thank you for reviewing the patch, > > > > > > On Tue, Aug 2, 2022 at 1:27 PM Dumitru Ceara wrote: > > > >> On 7/19/22 15:33, Mohammad Heib wrote: > >>> ovn-northd re-create sbrec row

Re: [ovs-dev] [PATCH ovn v2 2/2] multicast: Properly flood IGMP queries and reports.

On 7/20/22 10:13, Ales Musil wrote: > Hi Dumitru, > Hi Ales, > the code looks good to me, with one small nit see below. > Thanks for the review! > On Tue, Jul 19, 2022 at 10:43 PM Dumitru Ceara wrote: > >> RFC 4541 "Considerations for Internet Group Management Protocol (IGMP) >> and

Re: [ovs-dev] [PATCH ovn] northd: handle virtual lport type update

On 8/2/22 15:56, Mohammad Heib wrote: > HI Dumitru, > thank you for reviewing the patch, > > > On Tue, Aug 2, 2022 at 1:27 PM Dumitru Ceara wrote: > >> On 7/19/22 15:33, Mohammad Heib wrote: >>> ovn-northd re-create sbrec row for lport of type >>> virtual when the type explicitly updated in

Re: [ovs-dev] [PATCH ovn] northd: handle virtual lport type update

HI Dumitru, thank you for reviewing the patch, On Tue, Aug 2, 2022 at 1:27 PM Dumitru Ceara wrote: > On 7/19/22 15:33, Mohammad Heib wrote: > > ovn-northd re-create sbrec row for lport of type > > virtual when the type explicitly updated in NBDB. > > > > This approach was applied to handle

Re: [ovs-dev] [PATCH] ovsdb: Fix copying weak references into transaction history.

On 8/2/22 15:23, Ilya Maximets wrote: > Transaction history is used only to construct row data updates for > clients, it's not used for checking data integrity, hence it doesn't > need a copy of weak references. > > Not copying this data saves a lot of CPU cycles and memory in some > cases. For

[ovs-dev] [PATCH] ovsdb: Fix copying weak references into transaction history.

Transaction history is used only to construct row data updates for clients, it's not used for checking data integrity, hence it doesn't need a copy of weak references. Not copying this data saves a lot of CPU cycles and memory in some cases. For example, in 250-node density-heavy scenario in

Re: [ovs-dev] [PATCH] rhel: make the version, displayed to the user, customizable

Timothy Redaelli writes: > Since on CentOS/RHEL the builds are based on stable branches and not on > tags for debugging purpose it's better to have the downstream version as > version so it's easier to know which commits are included in a build. > > This commit adds DISPLAY_VERSION as

Re: [ovs-dev] [PATCH v2] netdev-linux: set correct action for packets that passed policer

On Tue, Aug 02, 2022 at 10:38:03AM +0300, Vlad Buslov wrote: > On Wed 27 Jul 2022 at 11:49, Vlad Buslov wrote: > > On Wed 27 Jul 2022 at 10:26, Simon Horman wrote: > >> On Tue, Jul 26, 2022 at 07:52:31AM -0500, Marcelo Ricardo Leitner wrote: > >>> On Tue, Jul 26, 2022 at 09:24:12AM +0200, Vlad

Re: [ovs-dev] [PATCH] netdev-offload-tc: Disable offload of IPv6 fragments.

On 2022-07-29 11:38 PM, Ilya Maximets wrote: OVS kernel datapath and TC are parsing IPv6 fragments differently. For IPv6 later fragments, according to the original design [1], OVS always sets nw_proto=44 (IPPROTO_FRAMENT), regardless of the type of the L4 protocol. typo IPPROTO_FRAGMENT

Re: [ovs-dev] [PATCH] test-ovsdb: Fix false-positive leaks from LeakSanitizer.

On 8/2/22 12:31, Dumitru Ceara wrote: > On 7/29/22 19:52, Ilya Maximets wrote: >> On 7/29/22 08:05, Ales Musil wrote: >>> >>> On Thu, Jul 28, 2022 at 5:43 PM Ilya Maximets >> > wrote: >>> >>> LeakSanitizer for some reason reports these json objects as leaked, >>>

Re: [ovs-dev] [PATCH] test-ovsdb: Fix false-positive leaks from LeakSanitizer.

On 7/29/22 19:52, Ilya Maximets wrote: > On 7/29/22 08:05, Ales Musil wrote: >> >> On Thu, Jul 28, 2022 at 5:43 PM Ilya Maximets > > wrote: >> >> LeakSanitizer for some reason reports these json objects as leaked, >> even though we do have references to them at

Re: [ovs-dev] [PATCH ovn] northd: handle virtual lport type update

On 7/19/22 15:33, Mohammad Heib wrote: > ovn-northd re-create sbrec row for lport of type > virtual when the type explicitly updated in NBDB. > > This approach was applied to handle container lport > type updates and avoid handling issues in the ovn-controller, > this patch expanded that approach

Re: [ovs-dev] [PATCH v2] netdev-linux: set correct action for packets that passed policer

On Wed 27 Jul 2022 at 11:49, Vlad Buslov wrote: > On Wed 27 Jul 2022 at 10:26, Simon Horman wrote: >> On Tue, Jul 26, 2022 at 07:52:31AM -0500, Marcelo Ricardo Leitner wrote: >>> On Tue, Jul 26, 2022 at 09:24:12AM +0200, Vlad Buslov wrote: >>> > Referenced commit changed policer action type from

Re: [ovs-dev] [PATCH v2] dpif-netdev: Simplify AVX512 build time checks to enhance readability.

> > -Original Message- > > From: Pai G, Sunil > > Sent: Monday 4 July 2022 13:27 > > To: d...@openvswitch.org > > Cc: echau...@redhat.com; Finn, Emma ; Van Haaren, > > Harry ; Ferriter, Cian > > > > Subject: [PATCH v2] dpif-netdev: Simplify AVX512 build time checks to > enhance

[ovs-dev] [PATCH ovn] ovn-ctl: Ensure that log/run directory have correct permission

When someone creates or changes permission of the var/log directory it stays that way. This can cause issues for logrotate as it needs to have correct permission on directory. Attempt to fix the permission every time we run start_ovn_daemon. Reported-at: https://bugzilla.redhat.com/2113855

Re: [ovs-dev] [PATCH 0/2] m4: Replace obsolete AC_HELP_STRING and AC_TRY_RUN.

Hi Ilya, > >> > >> It also complains about AC_PROG_CC_C99, but I'm not sure if we want > >> to or how to correctly replace it. > >> Suggested AC_PROG_CC enables C11 by default and that might not be a > >> desired behavior. > > > > Not an expert in this area, but to ensure C99 support, the > >

Re: [ovs-dev] [PATCH] checkpatch: Add check for a Fixes tag.

> -Original Message- > From: dev On Behalf Of Ilya Maximets > Sent: Tuesday, July 26, 2022 7:00 PM > To: ovs-dev@openvswitch.org > Cc: Ilya Maximets > Subject: [ovs-dev] [PATCH] checkpatch: Add check for a Fixes tag. > > A new check for common mistakes while formatting a 'Fixes:'

Re: [ovs-dev] [PATCH v2 4/5] netdev-offload-tc: Fix ignoring unknown tunnel keys.

On 2022-08-02 9:53 AM, Roi Dayan wrote: On 2022-08-01 9:31 AM, Roi Dayan wrote: On 2022-07-29 5:53 PM, Ilya Maximets wrote: Current offloading code supports only limited number of tunnel keys and silently ignores everything it doesn't understand.  This is causing, for example, offloaded

Re: [ovs-dev] [PATCH v2 4/5] netdev-offload-tc: Fix ignoring unknown tunnel keys.

On 2022-08-01 9:31 AM, Roi Dayan wrote: On 2022-07-29 5:53 PM, Ilya Maximets wrote: Current offloading code supports only limited number of tunnel keys and silently ignores everything it doesn't understand.  This is causing, for example, offloaded ERSPAN tunnels to not work, because flow is