On Fri, Feb 14, 2020 at 8:10 PM Yi Yang (杨燚)-云服务集团 wrote:
>
> William, I don't know why I can't receive your comments in my outlook,
> https://mail.openvswitch.org/pipermail/ovs-dev/2020-February/367860.html
>
Are you able to use github? if OK, then travis can directly links
to you github repo. So
On Fri, Feb 14, 2020 at 01:14:10PM -0800, William Tu wrote:
> On Tue, Feb 11, 2020 at 06:21:37PM +0800, yang_y...@126.com wrote:
> > From: Yi Yang
> >
> > We can avoid high system call overhead by using TPACKET_V3
> > and using DPDK-like poll to receive and send pa
s about 1.98
> Gbps, but it was 1.47 Gbps before.
>
> Note: Linux kernel TPACKET_V3 can't support TSO, so the performance
> is very poor, please turn off tso for veth interfces in case
> userspace-tso-enable is set to true.
>
> Signed-off-by: Yi Yang
> Co-authored-b
d frame size to accommodate the
TSO 64K-size packet?
William
> -邮件原件-----
> 发件人: William Tu [mailto:u9012...@gmail.com]
> 发送时间: 2020年2月14日 8:38
> 收件人: Ilya Maximets
> 抄送: yang_y...@126.com; ovs-dev ; yang_y_yi
> ; Ben Pfaff ; Yi Yang (杨燚)-云服务集团
>
> 主题: Re: [ovs-dev]
On Fri, Feb 7, 2020 at 6:43 AM Ilya Maximets wrote:
>
> On 2/7/20 12:50 PM, yang_y...@126.com 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
On Thu, Feb 13, 2020 at 3:50 PM Ben Pfaff wrote:
>
> On Tue, Feb 11, 2020 at 11:13:35AM -0800, William Tu wrote:
> > Currently we use MSVC to compile OVS on Windows. The patch tries
> > to cross-compile OVS for windows using gcc from mingw-w64.
> > The patch still shows lo
mingw32 \
CXX=x86_64-w64-mingw32-g++ \
CXXCPP=x86_64-w64-mingw32-cpp-win32 \
--includedir=/usr/x86_64-w64-mingw32/include \
--prefix="C:/openvswitch/usr" \
--localstatedir="C:/openvswitch/var" \
--sysconfdir="C:/openvswitch/etc"
Signed-off-by: William
On Mon, Feb 10, 2020 at 9:44 AM Ben Pfaff wrote:
>
> Hi everyone! The 2.13 release should, by our calendar, happen around
> Feb. 15. I would like to do it Feb. 14, if possible, because I'm
> planning to take next week off.
>
> Does anyone have anything they'd like to get into 2.13 or know a reas
On Sun, Feb 2, 2020 at 8:06 PM Yi Yang (杨燚)-云服务集团 wrote:
>
> Hi, William
>
> Sorry for last reply, I don't know why I always can't get your comments
> email from my outlook, Ben's comments are ok, I also can't see your comments
> in outlook junk box.
>
> About your comments in
> https://mail.openv
Hi Ilya,
Thanks for your quick response.
On Thu, Jan 30, 2020 at 1:14 AM Ilya Maximets wrote:
>
> Hi.
> I didn't test or carefully review this. Just a couple of comments
> inline.
>
> Best regards, Ilya Maximets.
>
> On 30.01.2020 00:23, William Tu wrote:
>
around 6Gbps, similar to that with DPDK.
Tested-at: https://travis-ci.org/williamtu/ovs-travis/builds/643599592
Signed-off-by: William Tu
---
lib/dp-packet.h | 95 -
lib/userspace-tso.c | 9 ++---
2 files changed, 62 insertions(+), 42 deletions(-)
On Wed, Jan 29, 2020 at 3:25 AM Flavio Leitner wrote:
>
> On Tue, Jan 28, 2020 at 03:23:02PM -0800, Yifeng Sun wrote:
> > Sure.
> >
> > Firstly, make sure userspace-tso-enable is true
> > # ovs-vsctl get Open_vSwitch . other_config
> > {dpdk-init="true", enable-statistics="true", pmd-cpu-mask="0xf
On Thu, Jan 23, 2020 at 09:03:11AM -0800, William Tu wrote:
> The 'XDP_RING_NEED_WAKEUP' and related flags are defined if_xdp.h, so if
> users are building their own kernel, users have to update the kernel's
> header files, by doing:
>
> $ make headers
On Fri, Dec 20, 2019 at 01:16:39PM -0800, Ben Pfaff wrote:
> On Fri, Dec 20, 2019 at 09:51:08AM -0800, Yi-Hung Wei wrote:
> > In connection tracking system, a connection is established if we
> > see packets from both directions. However, in userspace datapath's
> > conntrack, if we send a connecti
Hi Eelco and Ilya,
Do you think this patch is ok?
Thanks
William
On Thu, Jan 16, 2020 at 1:49 PM William Tu wrote:
>
> Now netdev-afxdp always forwards all packets to userspace because
> it is using libbpf's default XDP program, see 'xsk_load_xdp_prog'.
> There are so
On Fri, Jan 24, 2020 at 1:38 PM Flavio Leitner wrote:
>
> On Fri, Jan 24, 2020 at 10:17:10AM -0800, William Tu wrote:
> > On Fri, Jan 24, 2020 at 6:40 AM Flavio Leitner wrote:
> > >
> > > On Wed, Jan 22, 2020 at 10:33:59AM -0800, William Tu wrote:
> > > &g
On Fri, Jan 24, 2020 at 6:40 AM Flavio Leitner wrote:
>
> On Wed, Jan 22, 2020 at 10:33:59AM -0800, William Tu wrote:
> > On Wed, Jan 22, 2020 at 12:54 AM Flavio Leitner wrote:
> > >
> > >
> > > Hi Ben,
> > >
> > > Thanks for reviewing
rial/ovs-afxdp-installation/
Signed-off-by: William Tu
Acked-by: Ben Pfaff
---
Documentation/intro/install/afxdp.rst | 8
1 file changed, 8 insertions(+)
diff --git a/Documentation/intro/install/afxdp.rst
b/Documentation/intro/install/afxdp.rst
index c4685fa7ebac..99003e4dbdb2 100644
-
irst use in this function)
> > return *r->flags & XDP_RING_NEED_WAKEUP;
>
> > Reported-by: Tomek Osinski
> > Reported-at: https://osinstom.github.io/en/tutorial/ovs-afxdp-installation/
> > Signed-off-by: William Tu
> > ---
> > Documentation/intro/i
Hi Roni,
Thanks for your feedbacks. See my reply below.
On Tue, Jan 21, 2020 at 11:20 AM Roni Bar Yanai wrote:
>
> Hi William,
>
> Two comments, please inline.
>
> > struct dp_packet *
> >+netdev_gtpu_pop_header(struct dp_packet *packet)
> >+{
> >+struct pkt_metadata *md = &packet->md;
>
t; unixctl_server_wait(unixctl);
> > netdev_wait();
> > poll_block();
> >
> > In the next step, it will be better if let pmd_thread to handle
> > tap and veth interface.
> >
> > Signed-off-by: Yi Yang
> > Co-authored-by: Willi
On Wed, Jan 22, 2020 at 12:54 AM Flavio Leitner wrote:
>
>
> Hi Ben,
>
> Thanks for reviewing it!
>
> On Tue, Jan 21, 2020 at 01:35:39PM -0800, Ben Pfaff wrote:
> > On Sat, Jan 18, 2020 at 12:08:06AM +0100, Ilya Maximets wrote:
> > > On 18.01.2020 00:03, Stokes, Ian wrote:
> > > > Thanks all for r
ait();
> unixctl_server_wait(unixctl);
> netdev_wait();
> poll_block();
>
> In the next step, it will be better if let pmd_thread to handle
> tap and veth interface.
>
> Signed-off-by: Yi Yang
> Co-authored-by: William Tu
> Signed-off-by: William T
On Wed, Jan 22, 2020 at 12:26 AM Flavio Leitner wrote:
>
> On Tue, Jan 21, 2020 at 10:39:14AM -0800, William Tu wrote:
> > On Thu, Jan 16, 2020 at 9:01 AM Flavio Leitner wrote:
> > >
> > > Abbreviated as TSO, TCP Segmentation Offload is a feature which enables
>
On Thu, Jan 16, 2020 at 9:01 AM Flavio Leitner wrote:
>
> Abbreviated as TSO, TCP Segmentation Offload is a feature which enables
> the network stack to delegate the TCP segmentation to the NIC reducing
> the per packet CPU overhead.
>
> A guest using vhost-user interface with TSO enabled can send
.h: In function 'xsk_ring_prod__needs_wakeup':
/usr/local/include/bpf/xsk.h:82:21: error: 'XDP_RING_NEED_WAKEUP' undeclared \
(first use in this function)
return *r->flags & XDP_RING_NEED_WAKEUP;
Reported-by: Tomek Osinski
Reported-at: https://osinstom.github.io/en/tutorial/ovs-
:
https://tools.ietf.org/html/draft-hmm-dmm-5g-uplane-analysis-00
Signed-off-by: Feng Yang
Co-authored-by: Feng Yang
Signed-off-by: Yi Yang
Co-authored-by: Yi Yang
Signed-off-by: William Tu
Acked-by: Ben Pfaff
---
v7 -> v8:
- Add Feng Yang as co-authored
- fix checkpatch error
- ht
On Fri, Jan 17, 2020 at 7:00 PM 0-day Robot wrote:
>
> Bleep bloop. Greetings William Tu, 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.
>
>
> checkp
:
https://tools.ietf.org/html/draft-hmm-dmm-5g-uplane-analysis-00
Signed-off-by: Yi Yang
Co-authored-by: Yi Yang
Signed-off-by: William Tu
Acked-by: Ben Pfaff
---
v6 -> v7:
- address Ben's feedback
- function name: use netdev_tnl_calc_udp_csum
- remove unnecessary be32_to_be
On Fri, Jan 17, 2020 at 02:14:27PM -0800, Ben Pfaff wrote:
> On Thu, Jan 16, 2020 at 11:45:20AM -0800, William Tu wrote:
> > GTP, GPRS Tunneling Protocol, is a group of IP-based communications
> > protocols used to carry general packet radio service (GPRS) within
> > GSM,
On Fri, Jan 17, 2020 at 2:58 PM Ilya Maximets wrote:
>
> On 04.01.2020 02:13, Yi-Hung Wei wrote:
> > Currently, the AF_XDP socket (XSK) related memory are allocated by main
> > thread in the main thread's NUMA domain. With the patch that detects
> > netdev-linux's NUMA node id, the PMD thread of
On Fri, Jan 17, 2020 at 1:41 PM Stokes, Ian wrote:
>
>
>
> On 1/17/2020 9:37 PM, Flavio Leitner wrote:
> > On Fri, Jan 17, 2020 at 12:37:56PM -0800, William Tu wrote:
> >> On Fri, Jan 17, 2020 at 04:58:57PM -0300, Flavio Leitner wrote:
> >>> On Fri, Jan 17,
On Fri, Jan 17, 2020 at 04:58:57PM -0300, Flavio Leitner wrote:
> On Fri, Jan 17, 2020 at 06:58:56PM +0100, Ilya Maximets wrote:
> > On 16.01.2020 18:00, Flavio Leitner wrote:
> > > Abbreviated as TSO, TCP Segmentation Offload is a feature which enables
> > > the network stack to delegate the TCP s
be processed in kernel.
The patch enables loading the user-provided XDP program by
$ovs-vsctl -- set int afxdp-p0 options:xdp-obj=
So users can implement their filtering logic or traffic steering idea
in their XDP program, and rest of the traffic passes to AF_XDP socket
handled by OVS.
Signed-off
On Tue, Jan 7, 2020 at 2:53 PM Ben Pfaff wrote:
>
> On Wed, Dec 11, 2019 at 04:24:13PM -0800, William Tu wrote:
> > GTP, GPRS Tunneling Protocol, is a group of IP-based communications
> > protocols used to carry general packet radio service (GPRS) within
> > GSM, UMT
:
https://tools.ietf.org/html/draft-hmm-dmm-5g-uplane-analysis-00
Signed-off-by: Yi Yang
Co-authored-by: Yi Yang
Signed-off-by: William Tu
---
v5 -> v6:
- rebase to master
- travis: https://travis-ci.org/williamtu/ovs-travis/builds/638083655
v4 -> v5:
- address Ben and Aaron comments
Thanks for the feedback.
The original problem we faced is that when using AF_XDP on physical device,
users often forgets to setup combined channel (ethool -L eth0 combined N).
Assume the device's default combined channel is 8, and ovs creates only 1 n_rxq,
(ovs-vsctl -- set int eth0 n_rxq=1 ...).
ocating
> the memory in the net device's NUMA domain.
>
> Signed-off-by: Yi-Hung Wei
Thanks for working on this!
Acked-by: William Tu
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
from that port. Thus, abort the port setup when
> the case is detected.
>
> Signed-off-by: Yi-Hung Wei
Looks good to me, thanks!
CC Ilya to see if he has more insight/comments.
Acked-by: William Tu
> ---
> Travis CI: https://travis-ci.org/YiHungWei/ovs/builds/627972465
>
sh = nl_attr_get_u64(dupcall->hash);
> > }
> >
> > error = upcall_receive(upcall, udpif->backer, &dupcall->packet,
> > @@ -830,6 +831,7 @@ recv_upcalls(struct handler *handler)
> > upcall->key = dupcall->key;
> > u
t; Fix that by fully initializing dpif_execute structure. Using
> designated initializers to avoid such issues in the future.
>
> Fixes: 0442bfb11d6c ("ofproto-dpif-upcall: Echo HASH attribute back to
> datapath.")
> Signed-off-by: Ilya Maximets
> ---
Looks good to me.
Acked-
On Mon, Jan 06, 2020 at 11:52:43AM -0800, William Tu wrote:
> On Sun, Jan 05, 2020 at 01:51:19AM +0100, Ilya Maximets wrote:
> > Kernel uses 'xsk_generic_xmit()' for all modes where zerocopy is
> > not enabled:
> >
> >net/xdp/xsk.c
> >43
return xs->zc ? xsk_zc_xmit(xs) : xsk_generic_xmit(sk);
>443 }
>
> 'xsk_generic_xmit ()' sends packets synchronously and no more than 16
> packets at a time. This means that we have to kick Tx with sendmsg()
> for every 16 packets in simple native mode too, otherw
Hi Nicolas,
I was wrong. Actually with existing OVS internal device, the xdp generic mode
works fine. Please see my test on kernel 5.3 below
#!/bin/bash
modprobe openvswitch
ovs-dpctl add-dp br0
ip addr add dev br0 10.1.1.2/24
ip link set dev br0 up
ip netns add at_ns0
ip link add p0 type veth
On Fri, Jan 03, 2020 at 02:28:03PM +0100, Ilya Maximets wrote:
> On 03.01.2020 00:54, William Tu wrote:
> > On Mon, Dec 23, 2019 at 5:22 PM Yi Yang (杨燚)-云服务集团
> > wrote:
> >>
> >> William, maybe you don't know that kind of tap interface you're saying
On Fri, Jan 3, 2020 at 5:28 AM Ilya Maximets wrote:
>
> On 03.01.2020 00:54, William Tu wrote:
> > On Mon, Dec 23, 2019 at 5:22 PM Yi Yang (杨燚)-云服务集团
> > wrote:
> >>
> >> William, maybe you don't know that kind of tap interface you're say
On Mon, Dec 23, 2019 at 5:22 PM Yi Yang (杨燚)-云服务集团 wrote:
>
> William, maybe you don't know that kind of tap interface you're saying only
> can be used for VM, that is why openvswitch has to introduce internal type
> for the case I'm saying.
>
> In OVS DPDK case, if you create the below interfac
"queue_size= "
"queues= "
"iface= "
"server=<0|1> "
"mrg_rxbuf=<0|1> "
"in_order=<0|1> "
"packed_vq=<0|1>");
Thanks
William
> -邮件原件-
> 发件人: dev [mailto:dev-boun
Hi,
I'm trying to find a faster way to communicate from userspace OVS to
kernel. So I create a virtio_user port at OVS-DPDK, and send packets
to kernel's tap device.
packets in OVS userspace -> virtio-user port -> vhost-net (kernel) ->
tap device (kernel)
As described in paper[1], figure 1 for le
- set interface tapX type=afxdp
Regards,
William
>
> Anybody can tell me how I can create such a tap interface without using "
> ovs-vsctl add-port tapX"
>
> By the way, I tried af_packet for veth, the performance is very good, it is
> about 4Gbps on my machine, but it used TPAC
On Mon, Dec 23, 2019 at 12:29:25AM +, Yi Yang (杨燚)-云服务集团 wrote:
> Thanks William,
> https://www.kernel.org/doc/Documentation/networking/packet_mmap.txt is very
> good document for TPACKET_V*, I completely agree TPCKET_V3 is the best way to
> improve tap and veth performance. Can you share us
>
> At 2019-12-17 02:55:50, "William Tu" wrote:
> >On Fri, Dec 06, 2019 at 02:09:24AM -0500, yang_y...@163.com wrote:
> >> From: Yi Yang
> >>
> >> Current netdev_linux_rxq_recv_tap and netdev_linux_rxq_recv_sock
> >> just receive single pa
On Fri, Dec 20, 2019 at 01:14:37PM -0800, Ben Pfaff wrote:
> On Fri, Dec 20, 2019 at 09:49:44AM -0800, William Tu wrote:
> > On Thu, Dec 19, 2019 at 08:44:30PM -0800, Ben Pfaff wrote:
> > > On Thu, Dec 19, 2019 at 04:41:25PM -0800, William Tu wrote:
> > > > Curren
ue by allocating
> the memory in the net device's NUMA domain.
>
> Signed-off-by: Yi-Hung Wei
LGTM, Thanks for working on this!
(I wasn't able to test NIC on different NUMA id, becuase I don't
have physical access to hardware. So only make sure numa id=0 works)
Teste
On Thu, Dec 19, 2019 at 08:44:30PM -0800, Ben Pfaff wrote:
> On Thu, Dec 19, 2019 at 04:41:25PM -0800, William Tu wrote:
> > Currently the performance of sending packets from userspace
> > ovs to kernel veth device is pretty bad as reported from YiYang[1].
> > The patch adds
4. Added user-configurable knobs:
4.1 block::timeout
4.2 tpkt_hdr::sk_rxhash
- RX Hash data available in user space
- TX_RING semantics are conceptually similar to TPACKET_V2;
Thanks
William
>
> -邮件原件
Currently the performance of sending packets from userspace
ovs to kernel veth device is pretty bad as reported from YiYang[1].
The patch adds AF_PACKET v3, tpacket v3, as another way to
tx/rx packet to linux device, hopefully showing better performance.
AF_PACKET v3 should get closed to 1Mpps, as
On Tue, Dec 17, 2019 at 4:34 PM Ilya Maximets wrote:
>
> On 09.12.2019 21:42, William Tu wrote:
> > On Sat, Dec 07, 2019 at 03:46:18PM +0100, Ilya Maximets wrote:
> >> New file created for AF_XDP specific tests.
> >>
> >> Signed-off-by: Ilya Maximets
On Tue, Dec 17, 2019 at 07:59:42AM -0800, William Tu wrote:
> On Tue, Dec 17, 2019 at 09:14:00AM -0500, Nicolas Bouliane wrote:
> > >
> > >
> > > type of this port?
> > >
> > Internal
> >
> > We need to have an IP address set on
On Tue, Dec 17, 2019 at 09:14:00AM -0500, Nicolas Bouliane wrote:
> >
> >
> > type of this port?
> >
> Internal
>
> We need to have an IP address set on the interface, which is why we use the
> internal type.
>
>
> > Can you share your "ovs-vsctl show"
> > If meta0 is "type: internal", then it d
Hi,
We are measuring the performance of OVS using AF_XDP, and
have the following use case for XDP checksum:
High level Topology, two physical machines (PM1,2):
VM1 -> OVS encap with Geneve (PM1) -> physical NIC
-> network ->
physical NIC -> OVS decap (PM2) -> VM2
When running an iperf TCP
On Fri, Dec 06, 2019 at 02:09:24AM -0500, yang_y...@163.com wrote:
> From: Yi Yang
>
> Current netdev_linux_rxq_recv_tap and netdev_linux_rxq_recv_sock
> just receive single packet, that is very inefficient, per my test
> case which adds two tap ports or veth ports into OVS bridge
> (datapath_typ
n all other cases
> alignment only produces inefficient memory usage.
yes, now I realize umem->lock does not have the false
sharing problem.
>
> Also, CACHE_LINE_SIZE macro should be used instead of 64 as different
> platforms may have different cache line sizes.
>
> Using
On Mon, Dec 16, 2019 at 9:47 AM Ilya Maximets wrote:
>
> On 16.12.2019 18:23, William Tu wrote:
> > On Mon, Dec 16, 2019 at 7:27 AM Ilya Maximets wrote:
> >>
> >> On 16.12.2019 15:27, William Tu wrote:
> >>> On Mon, Dec 16, 2019 at 5:57 AM Ilya Maximets
On Mon, Dec 16, 2019 at 7:27 AM Ilya Maximets wrote:
>
> On 16.12.2019 15:27, William Tu wrote:
> > On Mon, Dec 16, 2019 at 5:57 AM Ilya Maximets wrote:
> >>
> >> On 16.12.2019 14:43, William Tu wrote:
> >>> On Mon, Dec 16, 2019 at 1:41 AM Ilya Maximets
On Mon, Dec 16, 2019 at 5:57 AM Ilya Maximets wrote:
>
> On 16.12.2019 14:43, William Tu wrote:
> > On Mon, Dec 16, 2019 at 1:41 AM Ilya Maximets wrote:
> >>
> >> On 13.12.2019 21:59, William Tu wrote:
> >>> Debugging netdev-afxdp is hard because tc
Signed-off-by: William Tu
---
Documentation/ref/ovs-tcpdump.8.rst | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/Documentation/ref/ovs-tcpdump.8.rst
b/Documentation/ref/ovs-tcpdump.8.rst
index 048b70f23daf..b9f8cdf6f786 100644
--- a/Documentation/ref/ovs-tcpdump.8.rst
+++ b
On Mon, Dec 16, 2019 at 1:41 AM Ilya Maximets wrote:
>
> On 13.12.2019 21:59, William Tu wrote:
> > Debugging netdev-afxdp is hard because tcpdump does not work
> > at all, even for generic mode. ovs-tcpdump which uses port
> > mirroring also does not work for netdev-
On Fri, Dec 13, 2019 at 01:31:04PM -0500, Nicolas Bouliane via dev wrote:
> Hi,
>
> I have a very simple ebpf program that just print something when a packet
> goes through. When I attach the ebpf program to the TAP interface (tun
> driver) of a virtual machine, I see the expected output as expect
On Fri, Dec 06, 2019 at 03:15:01PM -0800, Yifeng Sun wrote:
> Looks good to me, thanks YiHung.
>
> Reviewed-by: Yifeng Sun
>
Thank you.
Applied to master and branch-2.12
William
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/
Debugging netdev-afxdp is hard because tcpdump does not work
at all, even for generic mode. ovs-tcpdump which uses port
mirroring also does not work for netdev-afxdp. This patch
adds options:pcap= to enable capture the packets
from afxdp device.
Signed-off-by: William Tu
Tested-At: https
On Wed, Dec 04, 2019 at 03:06:07PM -0800, Ben Pfaff wrote:
> This makes the output both shorter and easier to read.
>
> Signed-off-by: Ben Pfaff
LGTM, I applied the series to master.
Thanks!
William
___
dev mailing list
d...@openvswitch.org
https://mai
core (per direction, according to RSS symmetric cfg).
I see. This looks like a performance issue.
Maybe we can improve it later.
>
> See inline.
>
> >-Original Message-
> >From: William Tu
> >Sent: Wednesday, December 11, 2019 8:20 PM
> >To: Roni B
:
https://tools.ietf.org/html/draft-hmm-dmm-5g-uplane-analysis-00
Signed-off-by: Yi Yang
Co-authored-by: Yi Yang
Signed-off-by: William Tu
---
v4 -> v5:
- address Ben and Aaron comments
1) flow_get_metadata, format_flow_tunnel
2) use of ?: in MSVS
3) tun_key_to_attr
4) handl
On Sun, Dec 08, 2019 at 08:11:19AM +, Roni Bar Yanai wrote:
> Hi William,
>
> GTP-U header size is not constant, you *must* take into account the flags,
> mainly
> the sequence. The use of sequence in GTP-U is optional but some devices do
> use
> it. see from 3GPP definition:
> "For PGW,
to prevent packet re-order as handover might have packets
> on the fly. So I would expected that END MARKER will do the exact same
> forwarding as the GTP-U data.
OK, thanks.
>
> see inline
>
> >-Original Message-
> >From: dev On Behalf Of William Tu
> &
mber/363038.html
> Signed-off-by: Ilya Maximets
Not sure about the Mellanox's bifurcated driver use case, but
I tested the afxdp port config fails cases, and I think the current
solution looks pretty ok to me.
Acked-by: William Tu
___
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
On Sat, Dec 07, 2019 at 03:46:18PM +0100, Ilya Maximets wrote:
> New file created for AF_XDP specific tests.
>
> Signed-off-by: Ilya Maximets
tested and LGTM, one minor comment.
Acked-by: William Tu
> ---
> tests/automake.mk | 3 ++-
> tests/system-afxdp
about infinite loop when enable loading XDP program at
https://patchwork.ozlabs.org/patch/1199734/
Acked-by: William Tu
> ---
> lib/netdev-afxdp.c | 13 +
> 1 file changed, 13 insertions(+)
>
> diff --git a/lib/netdev-afxdp.c b/lib/netdev-afxdp.c
> index ca2dfd005..bdabf97f4 10
On Thu, Dec 05, 2019 at 03:32:19PM +0100, Eelco Chaudron wrote:
>
>
> On 23 Nov 2019, at 1:26, William Tu wrote:
>
> >Now netdev-afxdp always forwards all packets to userspace because
> >it is using libbpf's default XDP program, see 'xsk_load_xdp_prog'.
On Sun, Dec 08, 2019 at 08:42:34PM +0100, Ilya Maximets wrote:
> This seems useful as I'm usually making a lot of typing mistakes.
> Also, few commonly used words added to the extended dictionary.
>
> Signed-off-by: Ilya Maximets
> ---
Acked-by: William Tu
> utili
t; of ones that starts with numbers.
>
> Signed-off-by: Ilya Maximets
Acked-by: William Tu
> ---
> utilities/checkpatch.py | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/utilities/checkpatch.py b/utilities/checkpatch.py
> index 6d95b0e2f..1b14c
id this
> issue we could pass non-stripped version of the word to the dictionary
> checker too.
>
> Signed-off-by: Ilya Maximets
LGTM,
Acked-by: William Tu
> ---
> utilities/checkpatch.py | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/u
On Thu, Dec 5, 2019 at 11:09 PM wrote:
>
> From: Yi Yang
>
> Current netdev_linux_rxq_recv_tap and netdev_linux_rxq_recv_sock
> just receive single packet, that is very inefficient, per my test
> case which adds two tap ports or veth ports into OVS bridge
> (datapath_type=netdev) and use iperf3 t
On Thu, Dec 05, 2019 at 08:53:35PM -0800, Ben Pfaff wrote:
> On Thu, Dec 05, 2019 at 12:43:31PM -0800, William Tu wrote:
> > GTP, GPRS Tunneling Protocol, is a group of IP-based communications
> > protocols used to carry general packet radio service (GPRS) within
> > GSM,
On Thu, Dec 5, 2019 at 5:43 AM William Tu wrote:
>
> On Wed, Dec 4, 2019 at 10:06 PM William Tu wrote:
> >
> > The patch works around an error when reconfigure the netdev-afxdp
> > device into different modes. Currently, when OVS destroy xsk, the
> > xsk_de
:
https://tools.ietf.org/html/draft-hmm-dmm-5g-uplane-analysis-00
Signed-off-by: Yi Yang
Co-authored-by: Yi Yang
Signed-off-by: William Tu
---
v3 -> v4:
- applied Ben's doc revise
- increment FLOW_WC_SEQ to 42
- minor fixes
- travis: https://travis-ci.org/williamtu/ovs-travi
On Thu, Dec 05, 2019 at 09:26:06AM -0800, Ben Pfaff wrote:
> On Thu, Dec 05, 2019 at 05:41:45AM -0800, William Tu wrote:
> > One question I have in this patch is whether to increase FLOW_WC_SEQ.
> > And what's the purpose of this number?
> >
> > /* This seque
On Wed, Dec 4, 2019 at 10:06 PM William Tu wrote:
>
> The patch works around an error when reconfigure the netdev-afxdp
> device into different modes. Currently, when OVS destroy xsk, the
> xsk_destruct() in linux kernel calls xdp_put_umem() and defers
> calling xdp_umem_r
On Wed, Dec 4, 2019 at 9:10 PM Ben Pfaff wrote:
>
> On Wed, Dec 04, 2019 at 06:37:38PM -0800, William Tu wrote:
> > On Wed, Dec 4, 2019 at 3:49 PM Ben Pfaff wrote:
> > >
> > > On Mon, Dec 02, 2019 at 02:19:29PM -0800, William Tu wrote:
> > > > GTP, GPRS
afxdp" options:xdp-mode=best-effort
ovs-vsctl -- set interface afxdp-p0 \
options:n_rxq=1 type="afxdp" options:xdp-mode=generic
The patch fixes it by adding a delay, hopefully the umem has
been properly cleanup.
Signed-off-by: William Tu
---
lib/netdev-afxdp.c | 3 +++
1 file c
On Wed, Dec 4, 2019 at 3:49 PM Ben Pfaff wrote:
>
> On Mon, Dec 02, 2019 at 02:19:29PM -0800, William Tu wrote:
> > GTP, GPRS Tunneling Protocol, is a group of IP-based communications
> > protocols used to carry general packet radio service (GPRS) within
> > GSM, UMT
On Tue, Dec 3, 2019 at 2:42 PM Ben Pfaff wrote:
>
> On Tue, Dec 03, 2019 at 01:44:08PM -0800, William Tu wrote:
> > Fix indentation and whitespace.
> >
> > Signed-off-by: William Tu
>
> Acked-by: Ben Pfaff
Thanks, applied to master.
_
On Tue, Dec 3, 2019 at 2:41 PM Ben Pfaff wrote:
>
> On Tue, Dec 03, 2019 at 01:44:07PM -0800, William Tu wrote:
> > Signed-off-by: William Tu
> > ---
> > AUTHORS.rst | 1 +
> > 1 file changed, 1 insertion(+)
> >
> > diff --git a/AUTHORS.rst b/AUTHOR
Fix indentation and whitespace.
Signed-off-by: William Tu
---
lib/match.c | 2 +-
lib/netdev-native-tnl.c | 4 ++--
lib/odp-util.c | 4 ++--
3 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/lib/match.c b/lib/match.c
index 052daee146f2..ae568280a2bb 100644
Signed-off-by: William Tu
---
AUTHORS.rst | 1 +
1 file changed, 1 insertion(+)
diff --git a/AUTHORS.rst b/AUTHORS.rst
index 4823badfbd4b..5043aeb27126 100644
--- a/AUTHORS.rst
+++ b/AUTHORS.rst
@@ -421,6 +421,7 @@ wenxu we...@ucloud.cn
wisd0me
On Fri, Nov 22, 2019 at 09:28:14AM -0800, Ben Pfaff wrote:
> On Thu, Nov 21, 2019 at 11:09:02AM -0800, William Tu wrote:
> > Make the code simpler by removing the use of
> > xasprintf and free, and use smap_add_format.
> >
> > Cc: Ben Pfaff
> > Signed-off-by:
:
https://tools.ietf.org/html/draft-hmm-dmm-5g-uplane-analysis-00
Signed-off-by: Yi Yang
Co-authored-by: Yi Yang
Signed-off-by: William Tu
---
v2 -> v3:
- pick up the code from v2, rebase to master
- many fixes in code, docs, and add more tests
- travis: https://travis-ci.org/williamtu/
On Tue, Nov 26, 2019 at 12:25 PM Ilya Maximets wrote:
>
> 'get_datapath_cap' callback was mistakenly placed in
> 'Connection tracking' section of the 'struct dpif_class'
> while belongs to the 'Datapath information'.
>
> CC: William
On Fri, Nov 22, 2019 at 10:50 AM Ilya Maximets wrote:
>
> On 22.11.2019 18:51, William Tu wrote:
> > Hi Ilya and Eelco,
> >
> > Yiyang reports very poor TCP performance on his setup and I can
> > also reproduce it on my machine. Somehow I think this might be a
>
501 - 600 of 1254 matches
Mail list logo
