Some partial offloading test cases are failing inconsistently. The root
cause is that dummy netdev is assigned with "linux_tc" offloading API.
dpif-netdev - partial hw offload - dummy
dpif-netdev - partial hw offload - dummy-pmd
dpif-netdev - partial hw offload with packet modifications - dummy
dpi
In the same environment, but I used tap but not veth, retr number is 0 for
the case without this patch (of course, I applied Flavio's tap enable patch)
vagrant@ubuntu1804:~$ sudo ./run-iperf3.sh
Connecting to host 10.15.1.3, port 5201
[ 4] local 10.15.1.2 port 54572 connected to 10.15.1.3 port 52
This is the result in my VM for the case without this patch. Retr number is
really very very high, this is super abnormal. Physical machine has same
phenomenon.
vagrant@ubuntu1804:~$ sudo ./run-iperf3.sh
Connecting to host 10.15.1.3, port 5201
[ 4] local 10.15.1.2 port 54566 connected to 10.15.1
William, I tested it in VM because I doesn't have physical machine with new
kernel, the number is similar, mine is about 2 Gbps, can you show me your
iperf output for the case without this patch, I saw a very strange
phenomenon, with this patch, retransmit is almost 0, but without this patch,
retra
Read/write concurrency of pvector library is implemented by a temp vector
and RCU protection. Considering performance reason, insertion does not
follow this scheme.
In insertion function, a thread fence ensures size incrementation is done
after new entry is stored. But there is no barrier in the it
On Mon, Feb 24, 2020 at 02:47:54PM +, Roni Bar Yanai wrote:
>
>
> >-Original Message-
> >From: Yi-Hung Wei
> >Sent: Saturday, February 22, 2020 12:51 AM
> >To: d...@openvswitch.org; Roni Bar Yanai
> >Cc: Yi-Hung Wei
> >Subject: [PATCH] docs: Update conntrack established state descr
On Mon, Feb 24, 2020 at 5:01 AM wrote:
>
> From: Yi Yang
>
> We can avoid high system call overhead by using TPACKET_V3
> and using DPDK-like poll to receive and send packets (Note: send
> still needs to call sendto to trigger final packet transmission).
>
> From Linux kernel 3.10 on, TPACKET_V3
Currently ovs-bugtool does not work due to Python2 deprecated.
When moving to Python3, a couple of things are broken in ovs-bugtool,
ex: import libraries issues such as StringIO and commands.
Also there are some bytes to string convertion in this patch
because in Python3,strings are now always Unic
From: Numan Siddique
This patch reverts the below patches which added lflow expr caching
supported and follow up patches which fixed few issues.
With the present lflow expr caching we still have issues with logical
flows referencing port groups/address sets. If a port group/address set
change ha
Madame, Monsieur,
Vous devez vous connecter à votre compte de crédit mutuel
dès que possible et accepter les
nouvelles conditions pour confirmer et / ou mettre à jour vos
données personnelles.
Avec cette étape, nous pouvons nous assurer que votre adresse
électronique et votre numéro de
tél
> -Original Message-
> From: Ilya Maximets
> Sent: Tuesday, February 25, 2020 6:31 PM
> To: Yanqin Wei ; Ilya Maximets
> ; d...@openvswitch.org
> Cc: nd ; Lijian Zhang ; Gavin Hu
>
> Subject: Re: [ovs-dev] [PATCH v1] netdev: fix partial offloading test cases
> failure
>
> On 2/25/20 10
Hi Ilya,
> -Original Message-
> From: Ilya Maximets
> Sent: Tuesday, February 25, 2020 5:54 PM
> To: Yanqin Wei ; d...@openvswitch.org
> Cc: Lijian Zhang ; Gavin Hu ; nd
> ; i.maxim...@ovn.org; Ben Pfaff
> Subject: Re: [ovs-dev] [PATCH v1] lib: use acquire-release semantics for
> pvecto
On 2/25/20 10:59 AM, Yanqin Wei wrote:
> Hi Ilya,
>
>
>> -Original Message-
>> From: Ilya Maximets
>> Sent: Tuesday, February 25, 2020 5:25 PM
>> To: Yanqin Wei ; d...@openvswitch.org
>> Cc: nd ; Lijian Zhang ; Gavin Hu
>> ; i.maxim...@ovn.org
>> Subject: Re: [ovs-dev] [PATCH v1] netdev:
Découvrez notre formation pour changer de vie, devenez
Praticien, Installez vous à votre compte ! et donnez du sens
à votre vie ...
Médicale ou Thérapeutique, cette nouvelle façon de
soulager et rétablir l'harmonie des personnes en passant par
l'inconscient, intéresse les professionn
Hi Ilya,
> -Original Message-
> From: Ilya Maximets
> Sent: Tuesday, February 25, 2020 5:25 PM
> To: Yanqin Wei ; d...@openvswitch.org
> Cc: nd ; Lijian Zhang ; Gavin Hu
> ; i.maxim...@ovn.org
> Subject: Re: [ovs-dev] [PATCH v1] netdev: fix partial offloading test cases
> failure
>
> On
On 2/25/20 2:52 AM, Yanqin Wei wrote:
> Read/write concurrency of pvector library is implemented by a temp vector
> and RCU protection. Considering performance reason, insertion does not
> follow this scheme.
> In insertion function, a thread fence ensures size incrementation is done
> after new en
On Tue, Feb 25, 2020 at 5:09 AM Ben Pfaff wrote:
>
> The series that introduced redirect-type first used "vlan" as the
> name for the feature, then changed to "bridged" without updating all
> the comments and the documentation. This updates those.
>
> CC: Ankur Sharma
> Fixes: 37cd27b8a85e ("OVN
On Tue, Feb 25, 2020 at 2:43 PM Dumitru Ceara wrote:
>
> On 2/25/20 12:39 AM, Ben Pfaff wrote:
> > The call to packet_set_ipv6() calls into packet_rh_present(), which in
> > turn iterates through the L3 content. Without this commit, the l4_ofs
> > in the packet has its default value of UINT16_MAX
On 2/25/20 2:46 AM, Yanqin Wei wrote:
> Some partial offloading test cases are failing inconsistently. The root
> cause is that dummy netdev is assigned with incorrect offloading flow API.
> dpif-netdev - partial hw offload - dummy
> dpif-netdev - partial hw offload - dummy-pmd
> dpif-netdev - part
On 2/25/20 12:39 AM, Ben Pfaff wrote:
> The call to packet_set_ipv6() calls into packet_rh_present(), which in
> turn iterates through the L3 content. Without this commit, the l4_ofs
> in the packet has its default value of UINT16_MAX, which means that
> packet_rh_present() reads well beyond the r
I have worked around it. Looking at the source I can guarantee that it is a
work-around and not a 100% fix, but it has survived an overnight loop run on a
6 core/12 thread 3.5 GHz Ryzen.
The root cause is that instead of using a proper packet oriented interface (f.e
SOCK_SEQPACKET or SOCK_DATA
21 matches
Mail list logo