Re: [ovs-dev] [PATCH v2 ovn 0/3] Introduce lb affinity timeout support

2022-10-13 Thread Mark Michelson
Hi Lorenzo, Thanks for addressing my concerns so quickly. I see that you addressed my concerns about load balancers with no ports in patch 3. However there are still some lingering issues: * In patch 1, the commit_lb_aff() action still expects all parsed IP addresses to have a port. * In

Re: [ovs-dev] [PATCH ovn] controller: improve buffered packets management

2022-10-13 Thread Mark Michelson
Hi Lorenzo, I only have one comment down below: On 10/7/22 07:08, Lorenzo Bianconi wrote: Improve buffered packet management in ovn-controller avoid useless loop in run_buffered_binding routine and using datapath key and output port key as buffered_packets_map hashmap hash. Add new selftest for

[ovs-dev] [PATCH v2 ovn 1/3] actions: introduce commit_lb_aff action

2022-10-13 Thread Lorenzo Bianconi
commit_lb_aff action translates to an openflow "learn" action that inserts a new flow in the OFTABLE_CHK_LB_AFFINITY table. The new flow is used to match on the the 5-tuple and set REGBIT_KNOWN_LB_SESSION bit. Moreover the new flow stores backend IP and port in register REG4 and REG8[0..15]

[ovs-dev] [PATCH v2 ovn 3/3] northd: rely on new actions for lb affinity

2022-10-13 Thread Lorenzo Bianconi
Rely on the following new actions in order to introduce affinity timeout support to load-balancer sessions: - commit_lb_aff - chk_lb_aff Introduce the following tables in switch and router pipelines respectively: - S_SWITCH_IN_LB_AFF_CHECK - S_SWITCH_IN_LB_AFF_LEARN - S_ROUTER_IN_LB_AFF_CHECK -

[ovs-dev] [PATCH v2 ovn 2/3] actions: introduce chk_lb_aff action

2022-10-13 Thread Lorenzo Bianconi
chk_lb_aff action checks if the packet under consideration matches the new flow in table OFTABLE_CHK_LB_AFFINITY. If so, then the 1-bit destination register is set to 1. chk_lb_aff will be used to add affinity timeout support to load balancer connections. Acked-by: Mark Michelson Signed-off-by:

[ovs-dev] [PATCH v2 ovn 0/3] Introduce lb affinity timeout support

2022-10-13 Thread Lorenzo Bianconi
Introduce load-balancer affinity timeout in order to dnat connections received from the same client to a given load-balancer to the same backend if received in the affinity timeslot specified by the CMS. Reported-at: https://bugzilla.redhat.com/show_bug.cgi?id=2127959 Changes since v1: - rebase

[ovs-dev] [PATCH v1 1/1] netdev-windows: Add checking when creating netdev with system type on Windows

2022-10-13 Thread Wilson Peng
From: Wilson Peng In the recent Antrea project testing, some port could not be created on Windows. When doing debug, our team found there is one case happening when multiple ports are waiting for be created with correct port number. Some system type port will be created netdev successfully and

Re: [ovs-dev] [PATCH ovn 0/3] Introduce lb affinity timeout support

2022-10-13 Thread Mark Michelson
On 10/13/22 07:35, Lorenzo Bianconi wrote: Hi Lorenzo, Hi Mark, Is load balancer affinity intended to work with a) Load balancers that do not specify a protocol? If not specified tcp is used as default value. TCP is the default only if a port is specified but no protocol is specified.

Re: [ovs-dev] [PATCH v1 1/1] netdev-windows: Add checking when creating netdev with system type on Windows

2022-10-13 Thread 0-day Robot
Bleep bloop. Greetings Wilson Peng, 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 spacing in pointer declaration #46 FILE: lib/netdev-windows.c:173:

[ovs-dev] [PATCH v1 1/1] netdev-windows: Add checking when creating netdev with system type on Windows

2022-10-13 Thread Wilson Peng
From: Wilson Peng In the recent Antrea project testing, some port could not be created on Windows. When doing debug, our team found there is one case happening when multiple ports are waiting for be created with correct port number. Some system type port will be created netdev successfully and

Re: [ovs-dev] [PATCH ovn 0/3] Introduce lb affinity timeout support

2022-10-13 Thread Ales Musil
On Thu, Oct 13, 2022 at 1:35 PM Lorenzo Bianconi < lorenzo.bianc...@redhat.com> wrote: > > Hi Lorenzo, > > Hi Mark, > > > > > Is load balancer affinity intended to work with > > > > a) Load balancers that do not specify a protocol? > > If not specified tcp is used as default value. > But that's

Re: [ovs-dev] [PATCH ovn 0/3] Introduce lb affinity timeout support

2022-10-13 Thread Lorenzo Bianconi
> Hi Lorenzo, Hi Mark, > > Is load balancer affinity intended to work with > > a) Load balancers that do not specify a protocol? If not specified tcp is used as default value. > b) Load balancers that do not specify a VIP port? ack, I need to fix it. > c) Load balancers that do not specify

Re: [ovs-dev] [PATCH v1 1/1] netdev-windows: Add checking when creating netdev with system type on Windows

2022-10-13 Thread 0-day Robot
Bleep bloop. Greetings Wilson Peng, 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 is 94 characters long (recommended limit is 79) #45 FILE:

Re: [ovs-dev] [PATCH ovn 3/3] northd: rely on new actions for lb affinity

2022-10-13 Thread Lorenzo Bianconi
> Hi Lorenzo, Hi Mark, > > Please add documentation to ovn-nb.xml for the new affinity_timeout option. ack, I will fix it in v2. > > I have more comments below. > > On 10/4/22 09:02, Lorenzo Bianconi wrote: > > Rely on the following new actions in order to introduce affinity timeout > >

[ovs-dev] [PATCH v1 1/1] netdev-windows: Add checking when creating netdev with system type on Windows

2022-10-13 Thread Wilson Peng
From: Wilson Peng In the recent Antrea project testing, some port could not be created on Windows. When doing debug, our team found there is one case happening when multiple ports are waiting for be created with correct port number. Some system type port will be created netdev successfully and

Re: [ovs-dev] [PATCH ovn 1/3] actions: introduce commit_lb_aff action

2022-10-13 Thread Lorenzo Bianconi
> On 10/4/22 09:02, Lorenzo Bianconi wrote: > > commit_lb_aff action translates to an openflow "learn" action that > > inserts a new flow in the OFTABLE_CHK_LB_AFFINITY table. The new flow is > > used to match on the the 5-tuple and set REGBIT_KNOWN_LB_SESSION bit. > > Moreover the new flow stores

Re: [ovs-dev] [PATCH ovn 2/3] actions: introduce chk_lb_aff action

2022-10-13 Thread Lorenzo Bianconi
> I found just one tiny grammatical problem in this one. Otherwise, > > Acked-by: Mark Michelson > > On 10/4/22 09:02, Lorenzo Bianconi wrote: > > chk_lb_aff action checks if the packet under consideration matches the > > new flow in table OFTABLE_CHK_LB_AFFINITY. If so, then the 1-bit > >

Re: [ovs-dev] [syzbot] WARNING in ovs_dp_reset_user_features

2022-10-13 Thread Paolo Abeni
On Wed, 2022-10-12 at 10:43 -0700, syzbot wrote: > Hello, > > syzbot found the following issue on: > > HEAD commit:e8bc52cb8df8 Merge tag 'driver-core-6.1-rc1' of git://git... > git tree: upstream > console+strace: https://syzkaller.appspot.com/x/log.txt?x=134de04288 > kernel

Re: [ovs-dev] [syzbot] WARNING in ovs_dp_cmd_new (2)

2022-10-13 Thread Paolo Abeni
On Wed, 2022-10-12 at 10:33 -0700, syzbot wrote: > Hello, > > syzbot found the following issue on: > > HEAD commit:e8bc52cb8df8 Merge tag 'driver-core-6.1-rc1' of git://git... > git tree: upstream > console+strace: https://syzkaller.appspot.com/x/log.txt?x=1112f08a88 > kernel