Goodbye from our Newsletter, sorry to see you go.
You have been unsubscribed from our newsletters.
This is the last email you will receive from us. Our newsletter system,
phpList,
will refuse to send you any further messages, without manual intervention
by our administrator.
If there is an
On 6/1/2020 8:29 PM, Sriharsha Basavapatna wrote:
On Mon, Jun 1, 2020 at 9:18 PM Eli Britstein wrote:
On 6/1/2020 6:15 PM, Sriharsha Basavapatna wrote:
On Mon, Jun 1, 2020 at 7:58 PM Eli Britstein wrote:
On 5/28/2020 11:19 AM, Sriharsha Basavapatna wrote:
In this patch we check if
miniflow_extract checks the validation of tunnel metadata by comparing
tunnel destination address, including 16 bytes ipv6 address.
This patch introduces a 'tunnel_valid' flag. If it is false,
md->cacheline2 will not be touched. This improvement is beneficial to
miniflow_extract performance for
OVS userspace datapath is a program with heavy memory access. It needs to
load/store a large number of memory, including packet header, metadata,
EMC/SMC/DPCLS tables and so on. It causes a lot of cache line missing and
refilling, which has a great impact on flow scalability. And in some cases,
In case of 10k+ flows, emc lookup will usually miss. Flow hash value is
always calculated in this case no matter smc is enabled or not. This patch
moves it from smc_insert function into fast_path_processing and
handle_packet_upcall function to avoid unnecessary hash calculation and memory
Hi Yanqin,
Thank you for your suggestions. I will send a new patch.
-Original Message-
From: Yanqin Wei [mailto:yanqin@arm.com]
Sent: Tuesday, June 2, 2020 11:51 AM
To: Linhaifeng ; d...@openvswitch.org
Cc: nd
Subject: RE: [PATCH] ovs rcu: update rcu pointer first
Hi Haifeng,
It
Hi Haifeng,
One more comment. Since this is a bug fix, it is possible that the maintainer
will backport to the previous branch. Therefore, it is recommended to split
into several patches and add fixes tag.
From: Anton Ivanov
1. Pull out buffering and send/receive ops from json rpc
2. Make the SSL send zero copy (it was creating an ofpbuf
out of an existing ofpbuf data without necessity).
3. Add vector IO to stream-fd to make flushing more
efficient. Also makes queueing for stream-fd and
On Fri, May 29, 2020 at 5:06 PM Simon Horman wrote:
>
> On Mon, May 18, 2020 at 09:44:42AM +0800, xiangxia.m@gmail.com wrote:
> > From: Tonghao Zhang
> >
> > This patch allows users to offload the TC flower rules with tunnel
> > mask. In some case, mask is useful as wildcards.
> >
> > For
Good Tuesday,
Please check if you have any requirement for Disposable Medical/Protective
masks / N95 / Face Shields / Protective suits with CE and FDA certificate.
*Country of Origin: Vietnam *
We can ship from our US/Australia offices or directly from Vietnam.
Below is the price list in USD.
From: Malvika Gupta
dfc_processing function returns the number of packets left to be processed
in 'packets' array via dp_packet_batch_size() function. dfc_processing function
is called only from dp_netdev_input__ and its return value is not checked upon
function return. Moreover,
Before sending packets on netdev, feature compatibility is always checked
and incompatible traffic should be dropped. But, packet batch is refilled
even when no packet needs to be dropped. This patch improves it by keeping
the original batch if no packet should be dropped.
Reviewed-by: Lijian
In the emc lookup function, hash/flow/key are checked for matching entry.
Each entry comparison loads several cachelines into CPU. So in the
multifow case, processor will wait for the data to be fetched from lower
level cacheline or main memory because of cache miss.
This patch modifies emc table
Key length is only useful for emc table. This patch moves the key length
calculation into emc_change_entry for fast path performance.
Reviewed-by: Lijian Zhang
Reviewed-by: Malvika Gupta
Reviewed-by: Ruifeng Wang
Signed-off-by: Yanqin Wei
---
lib/dpif-netdev.c | 10 ++
1 file
From: Anton Ivanov
A ByteQ with unlocked head and tail is unsafe for simultaneous
consume/produce.
If simultaneous use is desired, these either need to be locked
or there needs to be a third atomic or lock guarded variable
"used".
An atomic "used" allows the producer to enqueue safely because
From: Anton Ivanov
1. Fix probe logic. The stream_or_pstream_needs_probes
function returning a mix of integer and boolean. As a
result probes were NOT turned off in a number of cases on
unix domain sockets and other transports where there
should be no probing. It now returns -1 (do not know),
0
From: Anton Ivanov
Signed-off-by: Anton Ivanov
---
lib/byteq.c | 14 +++---
lib/byteq.h | 12 ++--
2 files changed, 13 insertions(+), 13 deletions(-)
diff --git a/lib/byteq.c b/lib/byteq.c
index da40c2530..8a7b4f1dc 100644
--- a/lib/byteq.c
+++ b/lib/byteq.c
@@ -38,16 +38,16
From: Anton Ivanov
The merge on backlogged behavior is predicated on prohibiting
processing of incoming transactions while there is an outstanding
backlog.
This behavior does not make sense if the xmit/recv is async from
actual business logic.
Signed-off-by: Anton Ivanov
---
I have updated the async io patch series and added to them the patch
to switch from "jsonrpc echo" to kernel probing where supported by the
OS.
The patchset is tested with OVN at scale and results in better stability
of OVN at high scales comparet to master. F.e. - I get 1 out of 3 OVN scale
We should update rcu pointer first then use ovsrcu_postpone to free
otherwise maybe cause use-after-free.
e.g.,reader indicates momentary quiescent and access old pointer after
writer postpone free old pointer and before setting new pointer.
Signed-off-by: Linhaifeng
---
lib/classifier.c
On 2020-06-02 1:35 AM, 0-day Robot wrote:
> Bleep bloop. Greetings Roi Dayan, 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 80 characters
If 'stats->n_packets' is less than 'op->ukey->stats.n_packets',
the calculation result will be wrong.
'stats->n_bytes' is the same.
Signed-off-by: zhaozhanxu
---
ofproto/ofproto-dpif-upcall.c | 8 ++--
1 file changed, 6 insertions(+), 2 deletions(-)
diff --git
On Tue, Jun 02, 2020 at 04:17:21PM +0800, Tonghao Zhang wrote:
> On Fri, May 29, 2020 at 5:06 PM Simon Horman
> wrote:
> >
> > On Mon, May 18, 2020 at 09:44:42AM +0800, xiangxia.m@gmail.com wrote:
> > > From: Tonghao Zhang
> > >
> > > This patch allows users to offload the TC flower rules
On Tue, Jun 2, 2020 at 4:33 PM Simon Horman wrote:
>
> On Tue, Jun 02, 2020 at 04:17:21PM +0800, Tonghao Zhang wrote:
> > On Fri, May 29, 2020 at 5:06 PM Simon Horman
> > wrote:
> > >
> > > On Mon, May 18, 2020 at 09:44:42AM +0800, xiangxia.m@gmail.com wrote:
> > > > From: Tonghao Zhang
>
We should update rcu pointer first then use ovsrcu_postpone to free
otherwise maybe cause use-after-free.
e.g.,reader indicates momentary quiescent and access old pointer after
writer postpone free old pointer and before setting new pointer.
CC: Ben Pfaff
Fixes: 0f2ea84841e1 (\ovs-rcu: New
use ovsrcu_set first then use ovsrcu_postpone
CC: Jarno Rajahalme
Fixes: da9cfca6e2d7 (\Revert "pvector: Expose non-concurrent priority
vector."\)
Acked-by: Yanqin Wei
Signed-off-by: Linhaifeng
---
lib/pvector.c | 15 ---
1 file changed, 8 insertions(+), 7 deletions(-)
diff
use ovsrcu_set first then use ovsrcu_postpone
CC: Jarno Rajahalme
Fixes: 4f6780691653 (\mirror: Allow concurrent lookups.\)
Acked-by: Yanqin Wei
Signed-off-by: Linhaifeng
---
ofproto/ofproto-dpif-mirror.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git
Greetings.
Is there an estimate on when this patch can be merged?
Thanks!
> On May 22, 2020, at 5:12 PM, Jeff Squyres wrote:
>
> In AB bonding, if the current active slave becomes disabled, a
> replacement slave is arbitrarily picked from the remaining set of
> enabled slaves. This commit
On Tue, Jun 2, 2020 at 5:34 PM Simon Horman wrote:
>
> On Tue, Jun 02, 2020 at 04:58:32PM +0800, Tonghao Zhang wrote:
> > On Tue, Jun 2, 2020 at 4:33 PM Simon Horman
> > wrote:
> > >
> > > On Tue, Jun 02, 2020 at 04:17:21PM +0800, Tonghao Zhang wrote:
> > > > On Fri, May 29, 2020 at 5:06 PM
Thanks a lot!
I proposed the new approach at:
https://patchwork.ozlabs.org/project/openvswitch/patch/20200525152821.19838-1-dalva...@redhat.com/
On Sat, May 23, 2020 at 8:38 PM Han Zhou wrote:
>
>
>
> On Sat, May 23, 2020 at 12:06 AM Daniel Alvarez wrote:
> >
> >
> > Thanks a lot Terry!
> >
>
On Tue, Jun 02, 2020 at 01:21:16PM +0200, Simon Horman wrote:
> On Mon, Jun 01, 2020 at 03:54:54PM +0300, Roi Dayan wrote:
> > In order to improve revalidator performance, extend netdev-offload with
> > terse
> > dump support. In terse dump mode modify code that parses netlink to flower
> > and
From: Tonghao Zhang
This series patch allows user to match tunnel src/dst
address and port with masked values. User can use that
to match specified IP address or port and then drop packets
(DDOS network attack from a gateway or others) or other actions.
patch [1]: allow user install TC Flowers
From: Tonghao Zhang
Not bugfix, make the codes more readable.
Cc: Simon Horman
Cc: Paul Blakey
Cc: Roi Dayan
Cc: Ben Pfaff
Cc: William Tu
Cc: Ilya Maximets
Signed-off-by: Tonghao Zhang
Acked-by: Roi Dayan
---
lib/netdev-offload-tc.c | 6 ++
lib/tc.c| 6 ++
2
From: Tonghao Zhang
To support more use case, for example, DDOS, which
packets should be dropped in hardware, this patch
allows users to match only the tunnel source IPs with
masked value.
$ ovs-appctl dpctl/add-flow
"tunnel(src=2.2.2.0/255.255.255.0,tp_dst=4789,ttl=64),\
From: Tonghao Zhang
To support installing the TC flowers to HW, via "ovs-appctl dpctl/add-flow"
command, there should be an ufid. This patch will check whether ufid exists,
if not, generate an ufid. Should to know that when processing upcall packets,
ufid is generated in parse_odp_packet for
yes,check dpdk gso lib……
在 2020年3月10日星期二,txfh2007 via dev 写道:
> Hi Flavio and all:
>
> Is there a way to support software TSO for DPDK tunnel network ? I
> have tried userspace TSO function, and running on tunnel network, I have
> got the following error:
> "Tunneling packets with HW
From: Tonghao Zhang
This patch allows users to offload the TC flower rules with
tunnel mask. This patch allows masked match of the following,
where previously supported an exact match was supported:
* Remote (dst) tunnel endpoint address
* Local (src) tunnel endpoint address
* Remote (dst)
On 6/1/2020 7:37 PM, Martin Varghese wrote:
Hi Greg,
I apologise for the confusion.
All the bareudp patches got integrated into the net-next tree during
5.6 -rcX time frame.
I was not careful enough while announcing the linux kernel version
bareudp is supported.
I checked the linux 5.7
use ovsrcu_set first then use ovsrcu_postpone
CC: Eelco Chaudron
Fixes: f82b3b6a2f4d (\ofproto-dpif-upcall: Only call ovsrcu_postpone()
on active actions\)
Acked-by: Yanqin Wei
Signed-off-by: Linhaifeng
---
ofproto/ofproto-dpif-upcall.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
use ovsrcu_set first then use ovsrcu_postpone
CC: Ben Pfaff
Fixes: 18080541d276 (\classifier: Add support for conjunctive matches.\)
Acked-by: Yanqin Wei
Signed-off-by: Linhaifeng
---
lib/classifier.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/lib/classifier.c
On Tue, Jun 02, 2020 at 04:58:32PM +0800, Tonghao Zhang wrote:
> On Tue, Jun 2, 2020 at 4:33 PM Simon Horman
> wrote:
> >
> > On Tue, Jun 02, 2020 at 04:17:21PM +0800, Tonghao Zhang wrote:
> > > On Fri, May 29, 2020 at 5:06 PM Simon Horman
> > > wrote:
> > > >
> > > > On Mon, May 18, 2020 at
This patch injects timing some important functions as listed
below in datapath. This can be expanded but the list is short
for now.
Signed-off-by: Gowrishankar Muthukrishnan
---
lib/dpif-netdev.c | 5 -
lib/netdev-dpdk.c | 3 +++
2 files changed, 7 insertions(+), 1 deletion(-)
diff --git
This is sample tiny daemon which receives event notification
from the registered events and displays the message info in the
terminal.
Signed-off-by: Gowrishankar Muthukrishnan
---
utilities/automake.mk | 6 +-
utilities/ovs-testeventd.c | 156
Documentation notes on event library and its usage is included
in this patch.
Signed-off-by: Gowrishankar Muthukrishnan
---
Documentation/automake.mk| 1 +
Documentation/topics/index.rst | 1 +
Documentation/topics/user-defined-events.rst | 306
Add iperf3 performance testsuite for checking events.
Signed-off-by: Gowrishankar Muthukrishnan
---
This is not complete test suite as I would like to get first set of
feedbacks on implementation and write unit tests accordingly. This
test is to check if event API works expectedly while running
Roi Dayan writes:
> On 2020-06-02 1:35 AM, 0-day Robot wrote:
>> Bleep bloop. Greetings Roi Dayan, 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:
>>
On Mon, Jun 01, 2020 at 03:54:54PM +0300, Roi Dayan wrote:
> In order to improve revalidator performance, extend netdev-offload with terse
> dump support. In terse dump mode modify code that parses netlink to flower and
> flower to match to only provide the essential data for conversion instead of
This patch enables monitoring coverage counters through event API.
Signed-off-by: Gowrishankar Muthukrishnan
---
lib/coverage.c | 61 ++
lib/coverage.h | 4
lib/timeval.c | 3 +++
3 files changed, 68 insertions(+)
diff --git
Timer notifications in event library require handling stopwatch
across one or more events, so share few required functions.
Signed-off-by: Gowrishankar Muthukrishnan
---
lib/stopwatch.c | 20 +++-
lib/stopwatch.h | 9 +
2 files changed, 28 insertions(+), 1 deletion(-)
User defined notification is helpful in troubleshooting OVS for
a particular event to appear as user will define. Based on the
type of event, event is handled by a dedicated thread and user
is notified in registered socket (jsonrpc) if provided.
This patch comprises of below:
1. event API which
This patchset enables new feature "user defined events and notification".
In short, by this feature, one can enable asynchronous and less intrusive
monitoring in Openvswitch data plane and control plane resources, by
defining one or more events to populate inside Openvswitch. Based on the
type of
On Mon, Jun 01, 2020 at 07:27:09PM -0400, Shahaji Bhosle via dev wrote:
> Hi Ben/Ilya,
> Hope you guys are doing well and staying safe. I have been chasing a weird
> problem with small drops and I think that is causing lots of TCP
> retransmission.
>
> Setup details
> iPerf3(1k-5K
>
If there are no other comments, can this be applied to master ?
Thanks,
-Harsha
On Mon, Jun 1, 2020 at 7:54 PM Eli Britstein wrote:
>
> Acked-by: Eli Britstein
>
> On 5/29/2020 9:33 AM, Sriharsha Basavapatna wrote:
> > Parse VLAN PUSH/POP OVS datapath actions and add respective RTE actions.
>
On Tue, Jun 02, 2020 at 07:27:59AM +, Linhaifeng wrote:
> We should update rcu pointer first then use ovsrcu_postpone to free
> otherwise maybe cause use-after-free.
> e.g.,reader indicates momentary quiescent and access old pointer after
> writer postpone free old pointer and before setting
On 5/25/2020 8:31 PM, Martin Varghese wrote:
From: Martin Varghese
There are various L3 encapsulation standards using UDP being discussed to
leverage the UDP based load balancing capability of different networks.
MPLSoUDP (__ https://tools.ietf.org/html/rfc7510) is one among them.
The
Hi Flavio,
Thanks for your reply.
I have captured the suggested information but do not see anything that
could cause the packet drops.
Can you please take a look at the below data and see if you can find
something unusual ?
The PMDs are running on CPU 1,2,3,4 and CPU 1-7 are isolated cores.
-Original Message-
From: Yanqin Wei [mailto:yanqin@arm.com]
Sent: Wednesday, June 3, 2020 7:23 AM
To: Ben Pfaff ; Linhaifeng
Cc: d...@openvswitch.org; nd ; Lilijun (Jerry)
; chenchanghu ; Lichunhe
Subject: RE: [ovs-dev] [PATCH v2] ovs rcu: update rcu pointer first
Hi Ben,
If
-Original Message-
From: Ben Pfaff [mailto:b...@ovn.org]
Sent: Wednesday, June 3, 2020 1:28 AM
To: Linhaifeng
Cc: Yanqin Wei ; d...@openvswitch.org; nd ;
Lilijun (Jerry) ; chenchanghu
; Lichunhe
Subject: Re: [ovs-dev] [PATCH v2] ovs rcu: update rcu pointer first
On Tue, Jun 02,
Hi Ben,
If my understanding is correct, the writer could not be a rcu thread because it
does not need report holding or not holding pointers.
So old memory will be freed after all rcu thread report quiesce.
Best Regards,
Wei Yanqin
> -Original Message-
> From: Ben Pfaff
> Sent:
Buenos día
Quise aprovechar la oportunidad de hacerte una invitación para tomar nuestro
curso de 6 horas :
Nombre: Taller de Contabilidad Electrónica 2020
Fecha: Martes 23 de Junio
Horario: 10:00 am a 1:00 pm y 2:00 pm a 05:00 pm
Formato: En línea con interacción en vivo.
Lugar: En Vivo desde
This is not how RCU works in OVS. Every thread is by default considered
active. They rarely quiesce except implicitly inside poll_block().
Please read the large comment at the top of ovs-rcu.h.
Is your patch based on actual bugs that you have found, or is it just
some kind of precaution? If it
Hi Ben,
This patch is from Linhai, but I have the same concern about this. I will read
ovs-rcu comments and feedback.
Thanks for your time.
Best Regards,
Wei Yanqin
> -Original Message-
> From: Ben Pfaff
> Sent: Wednesday, June 3, 2020 8:35 AM
> To: Yanqin Wei
> Cc: Linhaifeng ;
Oh, I apologize that I made a mistake about the author.
I appreciate feedback from anyone.
On Wed, Jun 03, 2020 at 12:37:27AM +, Yanqin Wei wrote:
> Hi Ben,
>
> This patch is from Linhai, but I have the same concern about this. I will
> read ovs-rcu comments and feedback.
> Thanks for
Hi Lorenzo,
Please find my comments inline.
Regards,
Ankur
From: dev on behalf of Ankur Sharma
Sent: Monday, June 1, 2020 9:05 PM
To: Lorenzo Bianconi ; ovs-dev@openvswitch.org
Subject: Re: [ovs-dev] [PATCH ovn] test: add more tests to IP-buffering
On Thu, May 28, 2020 at 4:05 AM wrote:
>
> From: Numan Siddique
>
> This patch handles SB port binding changes and OVS interface changes
> incrementally in the runtime_data stage which otherwise would have
> resulted in calls to binding_run().
>
> Prior to this patch, port binding changes were
On Wed, Jun 03, 2020 at 01:22:52AM +, Linhaifeng wrote:
>
>
> -Original Message-
> From: Ben Pfaff [mailto:b...@ovn.org]
> Sent: Wednesday, June 3, 2020 1:28 AM
> To: Linhaifeng
> Cc: Yanqin Wei ; d...@openvswitch.org; nd ;
> Lilijun (Jerry) ; chenchanghu
> ; Lichunhe
> Subject:
Thanks for the update. LGTM, except some minor changes related to the
comment in v9 4/8.
On Thu, May 28, 2020 at 4:05 AM wrote:
>
> From: Numan Siddique
>
> This patch handles ct zone changes and OVS interface changes incrementally
> in the flow output stage.
>
> Any changes to ct zone can be
On Thu, May 28, 2020 at 4:35 PM wrote:
> From: Venkata Anil
>
> This patch processes the logical flows of tracked datapaths
> and tracked logical ports. To handle the tracked logical port
> changes, reference of logical flows to port bindings is maintained.
>
> Acked-by: Mark Michelson
>
> -Original Message-
> From: Ben Pfaff [mailto:b...@ovn.org]
> Sent: Wednesday, June 3, 2020 8:35 AM
> To: Yanqin Wei
> Cc: Linhaifeng ; d...@openvswitch.org; nd
> ; Lilijun (Jerry) ; chenchanghu
> ; Lichunhe
> Subject: Re: [ovs-dev] [PATCH v2] ovs rcu: update rcu pointer first
>
>
69 matches
Mail list logo