[ovs-dev] [PATCH] netdev: Remove modulo operation in dpdk/vhost/afxdp tx path

After 324c8374852a ("dpif-netdev: XPS (Transmit Packet Steering) implementation."), it is impossible that qid is greater than or equal to n_txq of netdev. so remove this modulo operation which is in hot path Signed-off-by: Li RongQing --- lib/netdev-afxdp.c | 1 - lib/netdev-dpdk.c | 3 +-- 2

Re: [ovs-dev] [PATCH] rhel: increase timeout of ovsdb-server.service to 300s

Hi Timothy, Timothy Redaelli writes: > In some scenarios starting ovsdb-server takes more than 90 seconds and > so ovsdb-server.service is marked as failed. > > Signed-off-by: Timothy Redaelli > --- > NOTE: Currently systemd will not start openvswitch.service > and ovs-vswitchd.service if

Re: [ovs-dev] [PATCH v2] conntrack: Add a test for IPv4 UDP zero checksum

Hi Aaron, Aaron Conole writes: > Recently, during some conntrack testing a bug was uncovered in a DPDK > PMD, which doesn't support an IPv4 packet with a zero checksum value. > In order to show that the connection tracking code in userspace > supports IPv4 UDP with a zero checksum, add a test

Re: [ovs-dev] [PATCH ovn v2 6/9] tests: Amend release stale port binding test for RBAC

Hi Frode, For patches 6-9: Acked-by: Mark Michelson Can you specify which of the patches need to be backported and to which branches they need to be backported? I can take care of all the merges at the same time once I get this info. Thanks. On 3/5/21 7:16 AM, Frode Nordahl wrote: The

Re: [ovs-dev] [PATCH v3 2/3] ovsdb-idl: Preserve references for deleted rows.

On Mon, Mar 22, 2021 at 4:27 AM Dumitru Ceara wrote: > > On 3/19/21 6:55 PM, Han Zhou wrote: > > On Fri, Mar 19, 2021 at 1:42 AM Dumitru Ceara wrote: > >> > >> On 3/19/21 2:41 AM, Han Zhou wrote: > >>> > >>> > >>> On Thu, Mar 18, 2021 at 6:27 PM Han Zhou >>> > wrote: >

Re: [ovs-dev] [PATCH] conntrack: document NULL SNAT behavior and add a test case

Hi Eelco, Thanks for working on this, very useful indeed. not a full review, but I have a question about a minor thing. Eelco Chaudron writes: > Currently, conntrack in the kernel has an undocumented feature referred > to as NULL SNAT. Basically, when a source port collision is detected >

Re: [ovs-dev] [PATCH ovn v3] binding: Fix the crashes seen when port binding type changes.

On Fri, Mar 19, 2021 at 8:10 PM Dumitru Ceara wrote: > > On 3/19/21 1:05 PM, num...@ovn.org wrote: > > From: Numan Siddique > > > > When a port binding type changes from type 'A' to type 'B', then > > there are many code paths in the existing binding.c which results > > in crashes due to

Re: [ovs-dev] [PATCH ovn] ovn-nbctl: update BFD rows in nbctl_lr_route_del routine

Correct me if I'm wrong, but couldn't this be done automatically by just setting "isRoot: false" for the BFD table in ovn-nb.ovsschema? On 3/1/21 8:12 AM, Lorenzo Bianconi wrote: Remove BFD entry if its the last reference in static_routes table has been removed. Signed-off-by: Lorenzo

Re: [ovs-dev] [PATCH ovn] configure: Save OVSDIR as an absolute path.

On 3/22/21 2:02 PM, Mark Michelson wrote: Lately, `make check` runs have been failing. When "--with-ovs-source" is not specified at configure time, the configure script automatically sets the location to "$srcdir/ovs". This location is passed along the chain and eventually added to AUTOTEST_PATH

[ovs-dev] [PATCH ovn] configure: Save OVSDIR as an absolute path.

Lately, `make check` runs have been failing. When "--with-ovs-source" is not specified at configure time, the configure script automatically sets the location to "$srcdir/ovs". This location is passed along the chain and eventually added to AUTOTEST_PATH when running the tests. $srcdir is a

Re: [ovs-dev] [PATCH ovn] northd: Remove the usage of 'ct.inv' in logical flows.

On Fri, Mar 5, 2021 at 11:05 PM Numan Siddique wrote: > > On Fri, Mar 5, 2021 at 11:53 AM Han Zhou wrote: > > > > On Thu, Mar 4, 2021 at 5:25 PM Numan Siddique wrote: > > > > > > On Fri, Mar 5, 2021 at 4:22 AM Han Zhou wrote: > > > > > > > > On Mon, Mar 1, 2021 at 5:40 AM Numan Siddique

[ovs-dev] [RFC PATCH v2] conntrack: extract l4 information for SCTP

since a27d70a89 ("conntrack: add generic IP protocol support") all the unrecognized IP protocols get handled using ct_proto_other ops and are managed as L3 using 3 tuples. This patch stores L4 information for SCTP in the conn_key so that multiple conn instances, instead of one with ports zeroed,

Re: [ovs-dev] [PATCH 1/1] daemon-unix: Support OVS-DPDK HW offloads for non-root user

Hello Gaëtan, On Fri, Mar 19, 2021 at 5:59 PM Gaetan Rivet wrote: > Our rte_flow implementation uses ICM mappings to program our hardware, > which requires super privileged access. We are looking into ways to avoid it. Ok, thanks for looking into it. > > In the meantime, we failed to properly

Re: [ovs-dev] [PATCH] dpif-netdev-lookup: Allow AVX512 dpcls without building core OVS with SSE4.2.

On 3/22/21 12:45 PM, Van Haaren, Harry wrote: > Hi Ilya, > >> -Original Message- >> From: Ilya Maximets >> Sent: Monday, March 22, 2021 10:59 AM >> To: ovs-dev@openvswitch.org; Van Haaren, Harry >> >> Cc: Stokes, Ian ; Timothy Redaelli >> ; Flavio Leitner ; Ilya Maximets >> >> Subject:

Re: [ovs-dev] [PATCH ovn] expr: Combine multiple ipv4 with wildcard mask.

On 2021/3/19, 9:02 PM, "Dumitru Ceara" wrote: On 3/10/21 2:29 PM, gmingchen(陈供明) wrote: > From: Gongming Chen > > This patch merges ipv4 addresses with wildcard masks, and replaces this > ipv4 addresses with the combined ip/mask. This will greatly reduce the >

Re: [ovs-dev] [PATCH v3] dpif-netdev: Allow PMD auto load balance with cross-numa.

On 3/18/21 12:34 PM, Kevin Traynor wrote: > Previously auto load balance did not trigger a reassignment when > there was any cross-numa polling as an rxq could be polled from a > different numa after reassign and it could impact estimates. > > In the case where there is only one numa with pmds

Re: [ovs-dev] [PATCH V2] rhel: ovs-kmod-manage.sh: Disable unneeded warning

On 8/21/20 12:27 AM, Yi-Hung Wei wrote: > On Thu, Aug 20, 2020 at 2:56 PM Greg Rose wrote: >> >> The script itself says which versions the script is needed for but >> it is run on RHEL 8.x as well where it is not needed. Disable the >> warning and change the exit code to zero since it may

Re: [ovs-dev] [PATCH ovn 1/2] northd: Optimize ct nat for load balancer traffic.

Bleep bloop. Greetings Numan Siddique, 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 has trailing whitespace #1230 FILE: tests/ovn-northd.at:3068: AT_CLEANUP

Re: [ovs-dev] [PATCH] rhel: Do not update kmod RPM newer major revision kernels

On 2/9/21 8:47 PM, Greg Rose wrote: > The ovs-kmod-manage.sh script will run weak-updates even on newer > release kernels installing a non-compatible or un-runnable kernel > module. > > Update the script to never install weak-updates onto kernels with > newer major release versions. > >

Re: [ovs-dev] [PATCH] dpif-netdev-lookup: Allow AVX512 dpcls without building core OVS with SSE4.2.

Hi Ilya, > -Original Message- > From: Ilya Maximets > Sent: Monday, March 22, 2021 10:59 AM > To: ovs-dev@openvswitch.org; Van Haaren, Harry > > Cc: Stokes, Ian ; Timothy Redaelli > ; Flavio Leitner ; Ilya Maximets > > Subject: [PATCH] dpif-netdev-lookup: Allow AVX512 dpcls without

Re: [ovs-dev] [PATCH v3 2/3] ovsdb-idl: Preserve references for deleted rows.

On 3/19/21 6:55 PM, Han Zhou wrote: > On Fri, Mar 19, 2021 at 1:42 AM Dumitru Ceara wrote: >> >> On 3/19/21 2:41 AM, Han Zhou wrote: >>> >>> >>> On Thu, Mar 18, 2021 at 6:27 PM Han Zhou >> > wrote: On Thu, Mar 18, 2021 at 11:07 AM Ilya Maximets >>

[ovs-dev] [PATCH ovn 2/2] northd: Provide the option to not use ct.inv in lflows.

From: Numan Siddique Presently we add 65535 priority lflows in the stages - 'ls_in_acl' and 'ls_out_acl' to drop packets which match on 'ct.inv'. This patch provides an option to not use this field in the logical flow matches for the following reasons: • Some of the smart NICs which support

[ovs-dev] [PATCH ovn 1/2] northd: Optimize ct nat for load balancer traffic.

From: Numan Siddique For a load balancer traffic destined to VIP, we do the below actions related to conntrack in the ingress logical switch pipeline. 1. Send the packet to conntrack - ct() 2a. if ct.new then ct_lb(backends) 2b. if ct.est then ct(nat) The step 2b is unnecessary and can

[ovs-dev] [PATCH] dpif-netdev-lookup: Allow AVX512 dpcls without building core OVS with SSE4.2.

It's only required to build lib/dpif-netdev-lookup-avx512-gather.c with SSE4.2. All hash functions in lib/hash.h are defined as 'static inline', so they will use fast crc instructions while building this module. The minimal requirement for core OVS could be left intact. With this change support

[ovs-dev] [PATCH ovn] northd: Fix routing loop in LRs with one-to-many SNAT

If there are snat entries on the router, and some logical_ip are set to network instead of an IP address then given SNAT is masquerade. In such case ct_snat action is used in lr_in_unsnat table to ensure that the packet is matched against conntrack and destination IP is replaced with one from