Re: [ovs-dev] [PATCH] tc: Fix stats dump when using same meter table

On Mon, Sep 12, 2022 at 03:02:06PM +0200, Eelco Chaudron wrote: > > > On 7 Sep 2022, at 11:36, Simon Horman wrote: > > > From: Tianyu Yuan > > > > When we apply meter police on both directions of TCP traffic, the > > dumped stats is shown same (as shown below). This issue is introduced > > by m

[ovs-dev] [PATCH v2] netdev-dpdk: Drop TSO in case of conflicting virtio features.

At some point in OVS history, some virtio features were announced as supported (ECN and UFO virtio features). The userspace TSO code, which has been added later, does not support those features and tries to disable them. This breaks OVS upgrades: if an existing VM already negotiated such features

Re: [ovs-dev] [PATCH] netdev-linux: skip some internal kernel stats gathering

On 9/12/22 17:58, Jon Kohler wrote: > > >> On Sep 12, 2022, at 4:17 PM, Ilya Maximets wrote: >> >> On 6/24/22 22:01, Ilya Maximets wrote: >>> On 6/16/22 16:07, Jon Kohler wrote: > On Jun 3, 2022, at 1:00 PM, Ilya Maximets wrote: > > On 6/3/22 16:47, Jon Kohler wrote: >

Re: [ovs-dev] [PATCH ovn v3 4/4] northd: Re-use IP sets created for load balancer groups.

On 9/12/22 22:53, Han Zhou wrote: > > > On Fri, Sep 9, 2022 at 2:33 PM Ilya Maximets > wrote: >> >> If the same load balancer group is attached to multiple routers, >> IP sets will be constructed for each of them by sequentially calling >> sset_add() for each IP for ea

Re: [ovs-dev] [PATCH] tc: Fix stats dump when using same meter table

On 9/13/22 09:51, Simon Horman wrote: > On Mon, Sep 12, 2022 at 03:02:06PM +0200, Eelco Chaudron wrote: >> >> >> On 7 Sep 2022, at 11:36, Simon Horman wrote: >> >>> From: Tianyu Yuan >>> >>> When we apply meter police on both directions of TCP traffic, the >>> dumped stats is shown same (as shown

[ovs-dev] [PATCH ovn] northd: Use bitmaps instead of exclusive hash map for datapath groups.

Every time a new logical flow is created, ovn-northd creates an exclusive hash map to collect all the datapaths this logical flow is applicable to. While adding a new datapath to existing logical flow hmapx will perform a lookup by hash and then add a new entry by allocating a piece of memory and

Re: [ovs-dev] [PATCH] tc: Fix stats dump when using same meter table

On Tue, Sep 13, 2022 at 12:41:33PM +0200, Ilya Maximets wrote: > On 9/13/22 09:51, Simon Horman wrote: > > On Mon, Sep 12, 2022 at 03:02:06PM +0200, Eelco Chaudron wrote: > >> > >> > >> On 7 Sep 2022, at 11:36, Simon Horman wrote: > >> > >>> From: Tianyu Yuan > >>> > >>> When we apply meter police

Re: [ovs-dev] [PATCH] system-dpdk: Improve user configured mempool test.

On 05/09/2022 12:50, Sunil Pai G wrote: Improve the test by adding and varying the MTU of a vhost port to check if relevant mempools are created/(re)used. Signed-off-by: Sunil Pai G Thanks Sunil. Checked on master and branch-3.0 and working fine. Acked-by: Kevin Traynor --- tests/system

[ovs-dev] [PATCH] odp-execute: Add ISA implementation of set_masked IPv6 action

This commit adds support for the AVX512 implementation of the ipv6_set_addrs action as well as an AVX512 implementation of updating the L4 checksums. Signed-off-by: Emma Finn --- lib/odp-execute-avx512.c | 163 +++ 1 file changed, 163 insertions(+) diff --git

Re: [ovs-dev] [PATCH] system-dpdk: Improve user configured mempool test.

On 9/13/22 14:27, Kevin Traynor wrote: > On 05/09/2022 12:50, Sunil Pai G wrote: >> Improve the test by adding and varying the MTU of a vhost port to check if >> relevant mempools are created/(re)used. >> >> Signed-off-by: Sunil Pai G > > Thanks Sunil. Checked on master and branch-3.0 and working

Re: [ovs-dev] [PATCH v4] dpif-netdev: ct maxconns config persistence

Ilya Maximets writes: > On 5/21/22 05:55, lic121 wrote: >> Max allowed userspace dp conntrack entries is configurable with >> 'ovs-appctl dpctl/ct-set-maxconns' command. In real scenarios, >> this configuration is expected to survive from host reboot, from >> ovs service restart. >> >> Signed-of

Re: [ovs-dev] 答复: [PATCH] daemon-unix: fix bug of pipe residuals when ovn-northd restart by monitor

"Fengqi Li (李丰启)" writes: > <#part type=application/pkcs7-mime name="smime.p7m" smime-type="signed-data" > filename="smime.p7m" disposition=attachment buffer=" *mml*"> > <#/part> NOTE: This didn't come through my email client properly. Seems it thinks there is smime content here.

Re: [ovs-dev] [PATCH] daemon-unix: fix bug of pipe residuals when ovn-northd restart by monitor

"Fengqi Li (李丰启)" writes: > My ovs version is 2.13.2 and I checked the ff8decf1a318 commit .Ovs 2.13.2 > contains this commit and this bug still exists. I also checked the matser > branch and it may also have the same problem. > By the way,is there any way to change the title? The request is to

Re: [ovs-dev] [PATCH v4] ovs-tcpdump: Cleanup mirror port on SIGHUP/SIGTERM

Daniel Ding writes: > On 2022/9/10 1:43, Ilya Maximets wrote: > > On 9/8/22 09:37, Daniel Ding wrote: > > If ovs-tcpdump received HUP or TERM signal, mirror and mirror > interface should be destroyed. This often happens, when > controlling terminal is closed, like ssh session closed, and > othe

Re: [ovs-dev] [PATCH v3] conntrack: narrow the scope of ct_lock in new conn setup

Ilya Maximets writes: > On 8/2/22 06:01, we...@chinatelecom.cn wrote: >> From: wenxu >> >> There is a big scope ct_lock for new conn setup. The >> ct_lock should be hold only for conns map insert and >> expire rculist insert. >> >> Signed-off-by: wenxu >> --- >> lib/conntrack.c | 32

[ovs-dev] [PATCH v5] ovs-tcpdump: Cleanup mirror port on SIGHUP/SIGTERM

If ovs-tcpdump received HUP or TERM signal, mirror and mirror interface should be destroyed. This often happens, when controlling terminal is closed, like ssh session closed, and other users use kill to terminate it. Acked-by: Mike Pattrick Signed-off-by: Daniel Ding --- utilities/ovs-tcpdump.i

[ovs-dev] [PATCH] tests: Fix tests with GNU grep 3.8

GNU grep 3.8 started to emit warnings when invoking egrep/fgrep. In some cases this breaks tests that check stderr. Replace the commands with their grep -E and grep -F counterparts throughout. Signed-off-by: Andreas Stieger Reported-at: https://bugzilla.opensuse.org/show_bug.cgi?id=1203239 Submi

Re: [ovs-dev] [PATCH ovn v3 4/4] northd: Re-use IP sets created for load balancer groups.

On Tue, Sep 13, 2022 at 3:18 AM Ilya Maximets wrote: > > On 9/12/22 22:53, Han Zhou wrote: > > > > > > On Fri, Sep 9, 2022 at 2:33 PM Ilya Maximets mailto:i.maxim...@ovn.org>> wrote: > >> > >> If the same load balancer group is attached to multiple routers, > >> IP sets will be constructed for eac

Re: [ovs-dev] [PATCH ovn v3 0/4] northd: Optimize preparation of load balancers.

On Mon, Sep 12, 2022 at 3:25 AM Dumitru Ceara wrote: > > On 9/9/22 23:32, Ilya Maximets wrote: > > Re-compute of 'northd' node in ovn-northd may take almost half of the > > total processing time in case there is a big number of load balancers > > applied to multiple switches/routers or if there ar

[ovs-dev] [PATCH v2 0/2] bond: Avoid deadlock while updating post recirculation rules.

Version 2: - Added unit test for the double locking issue. - Added a new patch with debug commands to pause/resume revalidators. These are required for the unit test to work properly. Ilya Maximets (2): ofproto-dpif-upcall: Add debug commands to pause/resume revalidators. bond: Av

[ovs-dev] [PATCH v2 2/2] bond: Avoid deadlock while updating post recirculation rules.

If the PACKET_OUT from controller ends up with sending packet to a bond interface, the main thread will take locks in the following order: handle_openflow --> take ofproto_mutex handle_packet_out packet_xlate output_normal bond_update_post_recirc_rules --> take rwlock in bond.c If at

[ovs-dev] [PATCH v2 1/2] ofproto-dpif-upcall: Add debug commands to pause/resume revalidators.

New commands 'revalidator/pause' and 'revalidator/resume'. Not documented, since these should not be used in production environments. Will be used for unit tests in the next commit. Signed-off-by: Ilya Maximets --- ofproto/ofproto-dpif-upcall.c | 33 + 1 file cha

Re: [ovs-dev] [PATCH] bond: Avoid deadlock while updating post recirculation rules.

On 8/22/22 17:21, Mike Pattrick wrote: > On Thu, Aug 11, 2022 at 10:55 AM Ilya Maximets wrote: >> >> If the PACKET_OUT from controller ends up with sending packet to >> a bond interface, the main thread will take locks in the following >> order: >> handle_openflow >> --> take ofproto_mutex >>

[ovs-dev] stateless acl on ovn-k8s interface

Hi, Han, Numan: While testing a use case in our ovn-k8s cluster we ran into an issue where wecouldn't  effectively use stateless ACL on the OVN interface. Turns out we will track all the packets here, since there will be a some VIPs configured on the logical switch and...     if (od->has_lb_

Re: [ovs-dev] stateless acl on ovn-k8s interface

Bleep bloop. Greetings venugopal iyer, 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. git-am: error: corrupt patch at line 6 error: could not build fake ancestor hint: Use 'git am --show-curr

[ovs-dev] [PATCH v2] daemon-unix: Fix file descriptor leak when monitor restarts child

When segmentation fault occured in ovn-northd, monitor will try to restart the ovn-northd daemon process every 10s. Assume the following scenarios: There is a segmentation fault and the ovn-northd daemon process doen not restart properly everytime. Nws fds are created each time the ovn-northd daemo