In ofproto/ofproto-dpif.c function rule_dpif_lookup_from_table the first
fragment ports are _temporarily_ set to 0 for the purpose of flow lookup.
However, the ports are then restored, and multipath code sees non-0 ports for
the first fragment
We could skip the saving/restore of the original
Thanks Darrell.
So glad to make the requirement clear by this discussion, expecting to see the
latest patch.
发件人: Darrell Ball
发送时间: 2019年6月6日 1:12
收件人: 姜立东
抄送: d...@openvswitch.org; 王志克
主题: Re: [ovs-dev] [PATCH] conntrack: add tcp_in_liberal option in userspace
conntrack
On Wed, Jun 5,
On Wed, May 29, 2019 at 04:05:07PM +0200, Lorenzo Bianconi wrote:
> > On Thu, May 16, 2019 at 06:05:24PM +0200, Lorenzo Bianconi wrote:
> > > Add Controller_Event table to OVN SBDB in order to
> > > report CMS related event.
> > > Introduce event_table hashmap array and controller_event related
>
There is a switch config, called 'frag_mode'; the default is 'normal' where
tcp ports, udp ports
and ICMP type/code are set to 0 for all fragments.
http://www.openvswitch.org/support/dist-docs/ovs-ofctl.8.pdf
This is referenced during datapath flow translation right now.
It might be worth
It looks good to me,thank you.
Darrell Ball wrote:
> From: solomon
>
> ICMP/ICMPv6 fails, if the src/dst port is set in a common NAT rule.
> For example:
> actions=ct(nat(dst=172.16.1.100:5000),commit,table=40)
>
> Fixes: 4cd0481c9e8b ("conntrack: Fix wasted work for ICMP NAT.")
> CC:
Hello,
We offer loan at low Interest rate 3% and with not credit check, we offer
Personal loan, debt consolidation loan, venture capital, business loan,
Education Loan, home loan, Investment Loan OR "loan for any reason! Contact us
with your Information below;
1. Name Of Applicant:...
2.
From: "Enrico Weigelt, metux IT consult"
Date: Wed, 5 Jun 2019 23:06:40 +0200
> From: Enrico Weigelt
>
> IS_ERR() already calls unlikely(), so this extra likely() call
> around the !IS_ERR() is not needed.
>
> Signed-off-by: Enrico Weigelt
Applied.
Hello,
We offer loan at low Interest rate 3% and with not credit check, we offer
Personal loan, debt consolidation loan, venture capital, business loan,
Education Loan, home loan, Investment Loan OR "loan for any reason! Contact us
with your Information below;
1. Name Of Applicant:...
2.
Ben,
Looking into the current OVS behavior w.r.t. IP fragments: Based on
datapath/flow.c function key_extract, it looks like OVS would treat a
multi-fragment UDP packet as 2 different flows:
1) Unfragmented packets + first fragments for a given 5-tuple
2) Subsequent fragments ( for src/dst
For a series of IP fragments, only the first packet includes the transport
header (TCP/UDP/SCTP) and the src/dst ports. By including these port
numbers in the hash, it may happen that a first fragment hashes to a
different value than subsequent packets, causing different packets from
the same flow
From: solomon
ICMP/ICMPv6 fails, if the src/dst port is set in a common NAT rule.
For example:
actions=ct(nat(dst=172.16.1.100:5000),commit,table=40)
Fixes: 4cd0481c9e8b ("conntrack: Fix wasted work for ICMP NAT.")
CC: Darrell Ball
Signed-off-by: solomon
Signed-off-by: Darrell Ball
On 6/5/2019 1:18 PM, Ben Pfaff wrote:
On Fri, May 31, 2019 at 01:18:53PM -0700, Gregory Rose wrote:
On 5/30/2019 10:37 PM, Van Bemmel, Jeroen (Nokia - US) wrote:
Hello all,
Back in 2015 I submitted some code to improve ECMP hashing algorithms used in
OVS, see
will rebase and resend
On Wed, Jun 5, 2019 at 2:18 PM Darrell Ball wrote:
> From: solomon
>
> ICMP/ICMPv6 fails, if the src/dst port is set in a common NAT rule.
> For example:
> actions=ct(nat(dst=172.16.1.100:5000),commit,table=40)
>
> Fixes: 4cd0481c9e8b ("conntrack: Fix wasted work for
On Tue, May 28, 2019 at 11:14:44AM -0700, Darrell Ball wrote:
> The new api incorporates the hash calculation which can be a
> distraction.
>
> Signed-off-by: Darrell Ball
I applied this series to master. If any of it needs backporting, please
let me know.
On Wed, Jun 5, 2019 at 2:48 PM Ben Pfaff wrote:
> On Wed, Jun 05, 2019 at 02:30:43PM -0700, Darrell Ball wrote:
> > I resent the patch on your behalf as the test was corrupted somehow.
>
> I recommend adding a hyphen to this word form, to avoid confusion:
> "re-sent".
>
good idea :-)
On Wed, Jun 05, 2019 at 02:30:43PM -0700, Darrell Ball wrote:
> I resent the patch on your behalf as the test was corrupted somehow.
I recommend adding a hyphen to this word form, to avoid confusion:
"re-sent".
___
dev mailing list
d...@openvswitch.org
Hi Solomon
I resent the patch on your behalf as the test was corrupted somehow.
Thanks Darrell
On Sat, Jun 1, 2019 at 8:09 PM solomon wrote:
>
> ICMP/ICMPv6 fails, if the src/dst port is set in a common NAT rule.
> like this:
> actions=ct(nat(dst=172.16.1.100:5000),commit,table=40)
>
>
>
From: solomon
ICMP/ICMPv6 fails, if the src/dst port is set in a common NAT rule.
For example:
actions=ct(nat(dst=172.16.1.100:5000),commit,table=40)
Fixes: 4cd0481c9e8b ("conntrack: Fix wasted work for ICMP NAT.")
CC: Darrell Ball
Signed-off-by: solomon
Signed-off-by: Darrell Ball
From: Enrico Weigelt
IS_ERR() already calls unlikely(), so this extra likely() call
around the !IS_ERR() is not needed.
Signed-off-by: Enrico Weigelt
---
net/openvswitch/datapath.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/net/openvswitch/datapath.c
On Wed, Jun 05, 2019 at 08:34:50PM +, Van Bemmel, Jeroen (Nokia - US) wrote:
> Hi Greg, Ben,
>
> I doubt we would see a measurable difference in performance, with the
> additional conditional jump based on the packet flags. That does bring
> up an interesting question: Shouldn't fragmented
The patch introduces experimental AF_XDP support for OVS netdev.
AF_XDP, the Address Family of the eXpress Data Path, is a new Linux socket
type built upon the eBPF and XDP technology. It is aims to have comparable
performance to DPDK but cooperate better with existing kernel's networking
stack.
On Fri, May 31, 2019 at 03:57:45PM -0700, Yi-Hung Wei wrote:
> On Fri, May 31, 2019 at 2:59 PM Yifeng Sun wrote:
> >
> > RHEL 7.6's new kernel 3.10.0-957.12.1.el7 introduced some ABI changes.
> > This leads to an issue that OVS kernel module built against
> > kernels newer than 12.1 cannot be
Hi Greg, Ben,
I doubt we would see a measurable difference in performance, with the
additional conditional jump based on the packet flags. That does bring up an
interesting question: Shouldn't fragmented packets all hash to the same single
flow, and shouldn't the resulting multipath hash value
On Fri, May 31, 2019 at 01:18:53PM -0700, Gregory Rose wrote:
>
> On 5/30/2019 10:37 PM, Van Bemmel, Jeroen (Nokia - US) wrote:
> > Hello all,
> >
> > Back in 2015 I submitted some code to improve ECMP hashing algorithms used
> > in OVS, see
> >
On Fri, May 31, 2019 at 05:37:14AM +, Van Bemmel, Jeroen (Nokia - US) wrote:
> Hello all,
>
> Back in 2015 I submitted some code to improve ECMP hashing algorithms used in
> OVS, see https://mail.openvswitch.org/pipermail/ovs-dev/2015-July/300748.html
>
> We have now found an issue with the
Thanks for working on the problem.
I agree that this will solve this particular problem. At the same time,
it will break OVS in every situation where a flow's actions need to
change from "drop" to anything else. We can't accept that regression,
which would be a correctness issue, not just a
On Wed, Jun 05, 2019 at 09:27:38AM +, Rostyslav Fridman wrote:
> > This explains the semantics but not why they're useful. Why would one use
> > a passthrough nat_type?
> Basically this replicates the behavior of iptables ACCEPT rules in nat table.
>
> target prot opt source
On Wed, Jun 5, 2019 at 9:35 AM Darrell Ball wrote:
>
>
> On Tue, Jun 4, 2019 at 11:41 PM 姜立东 wrote:
>
>> *comments* *inline*.
>>
>>
>>
>> *发件人:* Darrell Ball
>> *发送时间:* 2019年6月5日 13:31
>> *收件人:* 姜立东
>> *抄送:* d...@openvswitch.org; 王志克
>> *主题:* Re: [ovs-dev] [PATCH] conntrack: add
On Tue, Jun 4, 2019 at 11:41 PM 姜立东 wrote:
> *comments* *inline*.
>
>
>
> *发件人:* Darrell Ball
> *发送时间:* 2019年6月5日 13:31
> *收件人:* 姜立东
> *抄送:* d...@openvswitch.org; 王志克
> *主题:* Re: [ovs-dev] [PATCH] conntrack: add tcp_in_liberal option in
> userspace conntrack
>
>
>
>
>
>
>
> On Tue, Jun 4,
On 05.06.2019 16:15, Cyprian Lech wrote:
>>From e0d636846fcb37e6d0eeb93aeb1ad89fb392e126 Mon Sep 17 00:00:00 2001
> From: Cyprian Lech
> Date: Wed, 5 Jun 2019 14:20:08 +0200
> Subject: [PATCH] netdev: Fix to disable netdev flow offloading
>
> Netdev flow functions were called although
>
On 5/8/2019 4:13 PM, Harry van Haaren wrote:
This commit splits the generic hash-lookup-verify
function to its own file. In doing so, we must move
some MACRO definitions to dpif-netdev.h
Signed-off-by: Harry van Haaren
---
v6:
- Fixup some checkpatch warnings on whitespace with MACROs (Ilya)
Bleep bloop. Greetings Cyprian Lech, 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:
fatal: patch fragment without header at line 6: @@ -2174,6 +2182,10 @@
>From e0d636846fcb37e6d0eeb93aeb1ad89fb392e126 Mon Sep 17 00:00:00 2001
From: Cyprian Lech
Date: Wed, 5 Jun 2019 14:20:08 +0200
Subject: [PATCH] netdev: Fix to disable netdev flow offloading
Netdev flow functions were called although
hw-offload was disabled, e.g. netdev_flow_dump_create.
It
> This explains the semantics but not why they're useful. Why would one use a
> passthrough nat_type?
Basically this replicates the behavior of iptables ACCEPT rules in nat table.
target prot opt source destination
ACCEPT all -- 10.0.0.0/8 192.168.0.0/16
comments inline.
发件人: Darrell Ball
发送时间: 2019年6月5日 13:31
收件人: 姜立东
抄送: d...@openvswitch.org; 王志克
主题: Re: [ovs-dev] [PATCH] conntrack: add tcp_in_liberal option in userspace
conntrack
On Tue, Jun 4, 2019 at 7:13 PM 姜立东
mailto:jianglido...@jd.com>> wrote:
Please check comments inline.
BR,
35 matches
Mail list logo