Thanks William. I am trying out your suggestions. Will keep you posted how
it goes.
Meanwhile, I found another issue without any of my changes related to
flooding and multiple tunnels in the same bridge. So, I initiated separate
discussion. Would you mind comment on my flooding issue. I am
NO, it is different issue.
The current deadlock mentioned in this patch is caused by some blocking
function (like ovsrcu_synchronize) in application (like OVS).
In this patch, the application is needed to break the logical deadlock.
At 2020-04-27 16:09:31, "Maxime
From: Numan Siddique
This patch add a new column 'selection_fields' in Load_Balancer
table in NB DB. CMS can define a set of packet headers to use
while selecting a backend. If this column is set, OVN will add the
flow in group table with selection method as 'hash' with the set fields.
Otherwise
On Sat, May 2, 2020 at 1:48 AM Mark Michelson wrote:
> I think you should follow this up with some tests.
>
> On 4/30/20 1:20 PM, num...@ovn.org wrote:
> > From: Numan Siddique
> >
> > This patch add a new column 'selection_fields' in Load_Balancer
> > table in NB DB. CMS can define a set of
On Tue, May 5, 2020 at 5:42 AM Mark Michelson wrote:
>
> Acked-by: Mark Michelson
>
> On 5/5/20 7:08 AM, Ilya Maximets wrote:
> > If route count reaches UINT16_MAX, ecmp_groups_add_route() will leak the
> > allocated route structure. Also, since group->route_count incremented
> >
On Mon, May 4, 2020 at 9:22 AM Han Zhou wrote:
> On Thu, Apr 30, 2020 at 10:20 AM wrote:
> >
> > From: Numan Siddique
> >
> > This patch add a new column 'selection_fields' in Load_Balancer
> > table in NB DB. CMS can define a set of packet headers to use
> > while selecting a backend. If this
On Mon, May 4, 2020 at 9:22 AM Han Zhou wrote:
> On Thu, Apr 30, 2020 at 10:20 AM wrote:
> >
> > From: Numan Siddique
> >
> > When adding flows to the group table, we need to use OFP15_VERSION to
> > set the selection_method. Right now ovn-controller is setting
> > select_method=dp_hash for
On 2020-04-21 12:22 PM, Roi Dayan wrote:
>
>
> On 2020-04-21 10:59 AM, Roi Dayan wrote:
>>
>>
>> On 2020-04-12 11:31 AM, Roi Dayan wrote:
>>> Hi,
>>>
>>> The two patches fix debian packaging.
>>> The first one to fix required package names.
>>> The second fix man page locations that were
A very gentle ping.
Thanks
Numan
On Thu, Apr 30, 2020 at 10:28 PM wrote:
> From: Numan Siddique
>
> This patch series handles port binding, datapath binding, ovs interface
> changes,
> runtime data changes, sb chassis changes incrementally.
>
> Below are the results of some testing I did
On Tue, Apr 21, 2020 at 11:47:03PM +0900, Toshiaki Makita wrote:
> This provider offloads classifier to software XDP.
>
> It works only when a custom XDP object is loaded by afxdp netdev.
> The BPF program needs to support classifier with array-of-maps for
> subtable hashmaps and arraymap for
On Tue, May 05, 2020 at 11:43:58AM +0900, Toshiaki Makita wrote:
> On 2020/05/04 1:22, William Tu wrote:
> >On Tue, Apr 21, 2020 at 11:47:04PM +0900, Toshiaki Makita wrote:
> >>This adds a reference program, flowtable_afxdp.o, which can be used to
> >>offload flows to XDP through
On Tue, May 5, 2020 at 3:26 PM Ilya Maximets wrote:
>
> On 5/5/20 3:12 PM, Stokes, Ian wrote:
> >
> >
> > On 4/28/2020 1:08 PM, David Marchand wrote:
> >> Add Travis jobs to check compilation with DPDK experimental API enabled.
> >> This will help us catch issues for the day we need one of them.
On Tue, May 5, 2020 at 1:48 AM Ilya Maximets wrote:
>
> On 5/5/20 3:08 AM, William Tu wrote:
> > On Mon, May 04, 2020 at 04:55:24PM -0700, Han Zhou wrote:
> >> On Mon, May 4, 2020 at 4:10 PM William Tu wrote:
> >>>
> >>> On Mon, May 04, 2020 at 09:55:41PM +0200, Ilya Maximets wrote:
>
On Tue, May 05, 2020 at 11:50:38AM +0900, Toshiaki Makita wrote:
> On 2020/05/05 1:24, William Tu wrote:
> >On Tue, Apr 21, 2020 at 11:47:00PM +0900, Toshiaki Makita wrote:
> >>This patch adds an XDP-based flow cache using the OVS netdev-offload
> >>flow API provider. When an OVS device with XDP
On Tue, Apr 21, 2020 at 11:47:00PM +0900, Toshiaki Makita wrote:
> This patch adds an XDP-based flow cache using the OVS netdev-offload
> flow API provider. When an OVS device with XDP offload enabled,
> packets first are processed in the XDP flow cache (with parse, and
> table lookup implemented
Find Attached
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
On 5/5/20 3:12 PM, Stokes, Ian wrote:
>
>
> On 4/28/2020 1:08 PM, David Marchand wrote:
>> Add Travis jobs to check compilation with DPDK experimental API enabled.
>> This will help us catch issues for the day we need one of them.
>>
>
> Am I right in thinking this is specific to dpdk-latest
On 5/5/20 2:44 PM, Stokes, Ian wrote:
>
>
> On 4/28/2020 1:03 PM, David Marchand wrote:
>> This header is duplicated from the DPDK generic header.
>> Fix typo identified in DPDK [1].
>>
>> While at it, RTE_EXEC_ENV_BSDAPP has been replaced with
>> RTE_EXEC_ENV_FREEBSD in 19.05 [2].
>>
>> 1:
On 4/28/2020 1:08 PM, David Marchand wrote:
Add Travis jobs to check compilation with DPDK experimental API enabled.
This will help us catch issues for the day we need one of them.
Am I right in thinking this is specific to dpdk-latest and wont be
applied to master?
If that's the case
On 4/28/2020 1:08 PM, David Marchand wrote:
rfc4115 meter APIs are now part of 20.02 stable ABI so we can remove the
forced ALLOW_EXPERIMENTAL_API define.
Signed-off-by: David Marchand
---
Changelog since v1:
- split travis changes in a separate patch
LGTM, just awaiting travis and will
>
> From: Numan Siddique
>
> The icmp packet generated by ovn-controller for reject ACL action
> for non TCP packets is not getting delivered to the sender of
> the original packet. This is because the icmp packets are skipped
> from out_pre_lb/out_pre_acl logical switch egress pipeline and this
On 4/28/2020 1:08 PM, David Marchand wrote:
When building against 20.05-rc1 which introduced a tracing framework,
mempool header now triggers an error with sparse.
.../dpdk-dir/build/include/rte_mempool_trace_fp.h:96:1:
error: undefined identifier '__atomic_load_n'
On 4/28/2020 1:08 PM, David Marchand wrote:
DPDK disabled kmods compilation in 20.02 [1].
Nothing left to be done on OVS side.
1: https://git.dpdk.org/dpdk/commit/?id=91a861e54164
Signed-off-by: David Marchand
LGTM, just waiting on the travis result (although I expect it to be
fine).
On 4/28/2020 1:03 PM, David Marchand wrote:
This header is duplicated from the DPDK generic header.
Fix typo identified in DPDK [1].
While at it, RTE_EXEC_ENV_BSDAPP has been replaced with
RTE_EXEC_ENV_FREEBSD in 19.05 [2].
1: https://git.dpdk.org/dpdk/commit/?id=a3e283ed904c
2:
Acked-by: Mark Michelson
On 5/5/20 7:08 AM, Ilya Maximets wrote:
If route count reaches UINT16_MAX, ecmp_groups_add_route() will leak the
allocated route structure. Also, since group->route_count incremented
unconditionally, next attempt to add new route will succeed, because the
value of
*My Dear friend,I am Mrs. Grace KaderI have emailed you earlier without any
response. I expected an urgent response to anticipate my previous
message.Regards;Mrs. Grace Kader.*
___
dev mailing list
d...@openvswitch.org
If route count reaches UINT16_MAX, ecmp_groups_add_route() will leak the
allocated route structure. Also, since group->route_count incremented
unconditionally, next attempt to add new route will succeed, because the
value of 'route_count' is zero now and out of sync with the real number
of
On Wed, Apr 29, 2020 at 10:24 PM wrote:
> From: Numan Siddique
>
> When adding flows to the group table, we need to use OFP15_VERSION to
> set the selection_method. Right now ovn-controller is setting
> select_method=dp_hash for OVN load balancers, but when encoding the
> group mod, it is
On 5/5/20 3:08 AM, William Tu wrote:
> On Mon, May 04, 2020 at 04:55:24PM -0700, Han Zhou wrote:
>> On Mon, May 4, 2020 at 4:10 PM William Tu wrote:
>>>
>>> On Mon, May 04, 2020 at 09:55:41PM +0200, Ilya Maximets wrote:
Function raft_command_initiate() returns correctly referenced command
On Mon, May 4, 2020 at 9:52 AM Ihar Hrachyshka wrote:
>
> On Fri, May 1, 2020 at 3:28 AM Han Zhou wrote:
> >
> >
> >
> > On Tue, Apr 28, 2020 at 1:38 PM Ihar Hrachyshka
wrote:
> > >
> > > Assuming only a single localnet port is actually plugged mapped on
> > > each chassis, this allows to
30 matches
Mail list logo