Hi Yi,
[auto build test ERROR on net-next/master]
url:
https://github.com/0day-ci/linux/commits/Yi-Yang/openvswitch-enable-NSH-support/20170909-124643
config: i386-randconfig-a1-201736 (attached as .config)
compiler: gcc-5 (Debian 5.4.1-2) 5.4.1 20160904
reproduce:
# save the attached
If destination port only receive mirrored traffic, this function do not
need add port with new type of taas. In this situation, the mirror flag is
needed.
But, I think, destination port receive both mirrored traffic and regular
traffic may be more flexible.
Thanks
Takashi YAMAMOTO
2017/09
Thanks Guru! I think this is a serious problem. I verified it is a problem
even with a single address set that is empty. It would impact some basic
use cases such as OpenStack Neutron security group. For example:
sec-group A:
rule1: ingress, remote group == sec-group B, ipv4, tcp 22 // allows acce
Hi Bhanu
What version of clang are you using and in what environment ?
Thanks Darrell
On 9/8/17, 10:59 AM, "ovs-dev-boun...@openvswitch.org on behalf of Bhanuprakash
Bodireddy" wrote:
Clang reports that value stored to ftp, seq_skew_dir never read inside
the function.
Signed
Hi Yuanhan
In the dpdk public meeting, we requested Simon to also review this series to
check
that it is sync with the HWOL for kernel with respect to one or two high level
aspects.
I had some comments related to coding standards for V2, but I’ll wait for you
to respond
to Simon’s comments befo
On 9/8/17, 9:47 AM, "ovs-dev-boun...@openvswitch.org on behalf of Simon Horman"
wrote:
On Tue, Sep 05, 2017 at 05:22:57PM +0800, Yuanhan Liu wrote:
> From: Finn Christensen
>
> The basic yet the major part of this patch is to translate the "match"
> to rte flow patterns.
The command error message is misleading, e.g.:
$ ovn-nbctl acl-del ls1 100 'outport=="lsp1" && ip4'
ovn-nbctl: cannot specify priority without match
$ ovn-nbctl acl-del ls1 to-lport 'outport=="lsp1" && ip4'
ovn-nbctl: cannot specify priority without match
This patch fixes the problem.
$
$ ovn-nb
Clang reports that value stored in to ac_offset is never read in the
function.
Signed-off-by: Bhanuprakash Bodireddy
---
ofproto/ofproto-dpif-xlate.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/ofproto/ofproto-dpif-xlate.c b/ofproto/ofproto-dpif-xlate.c
index 9e1f837..5d83666 100644
--- a
Clang reports that value stored to ftp, seq_skew_dir never read inside
the function.
Signed-off-by: Bhanuprakash Bodireddy
---
lib/conntrack.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/lib/conntrack.c b/lib/conntrack.c
index 419cb1d..a0838ee 100644
--- a/lib/conntr
By reordering elements in ofproto_bundle_settings structure, sum holes
and pad bytes can be reduced.
Before: structure size: 96, sum holes: 13, pad bytes: 7, cachelines:2
After : structure size: 80, sum holes: 4, pad bytes: 0, cachelines:2
Signed-off-by: Bhanuprakash Bodireddy
---
ofproto/ofpr
By reordering elements in ofproto_ipfix_flow_exporter_options structure,
sum holes can be reduced significantly.
Before: structure size: 64, sum holes: 11, cachelines:1
After : structure size: 56, sum holes: 3, cachelines:1
Signed-off-by: Bhanuprakash Bodireddy
---
ofproto/ofproto.h | 2 +-
1
By reordering elements in dp_netdev_rxq structure, pad bytes and a hole
can be removed.
Before: structure size: 104, sum holes: 1, sum padbytes:4, cachelines:2
After : structure size: 96, sum holes: 0, sum padbytes:0, cachelines:2
Signed-off-by: Bhanuprakash Bodireddy
---
lib/dpif-netdev.c | 2
By reordering elements in dpif_flow_put structure, holes can be removed.
Before: structure size: 80, sum holes: 8, cachelines:2
After : structure size: 72, sum holes: 0, cachelines:2
Signed-off-by: Bhanuprakash Bodireddy
---
lib/dpif.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff
By reordering elements in netdev structure, holes can be removed.
Before: structure size: 88, sum holes: 10, cachelines:2
After : structure size: 80, sum holes: 2, cachelines:2
Signed-off-by: Bhanuprakash Bodireddy
---
lib/netdev-provider.h | 11 ++-
1 file changed, 6 insertions(+), 5
By reordering bool in netdev_flow_dump structure, pad bytes can be
reduced.
Before: structure size: 32, sum holes: 4, sum padbytes:7, cachelines:1
After : structure size: 24, sum holes: 3, sum padbytes:0, cachelines:1
Signed-off-by: Bhanuprakash Bodireddy
---
lib/netdev-provider.h | 2 +-
1 fil
By reordering elements in netdev_tunnel_config structure, sum holes and
pad bytes can be reduced.
Before: structure size: 96, sum holes: 17, pad bytes: 4, cachelines:2
After : structure size: 80, sum holes: 5, pad bytes: 0, cachelines:2
Signed-off-by: Bhanuprakash Bodireddy
---
lib/netdev.h |
By reordering elements in ct_dpif_entry structure, sum holes and pad
bytes can be reduced.
Before: structure size: 232, sum holes: 5, pad bytes: 4, cachelines:4
After : structure size: 224, sum holes: 1, pad bytes: 0, cachelines:4
Signed-off-by: Bhanuprakash Bodireddy
---
lib/ct-dpif.h | 5 +++-
By reordering elements in tun_meta_entry structure, sum holes and pad
bytes can be reduced there by reducing the tun_table size.
Before: structure size: 56, sum holes: 4, pad bytes: 7 cachelines:1
After : structure size: 48, sum holes: 0, pad bytes: 3, cachelines:1
Signed-off-by: Bhanuprakash Bo
By reordering elements in dummy_packet_stream structure, sum holes
can be reduced, thus saving a cache line.
Before: structure size: 784, sum holes: 56, cachelines:13
After : structure size: 768, sum holes: 40, cachelines:12
Signed-off-by: Bhanuprakash Bodireddy
---
lib/netdev-dummy.c | 2 +-
1
By reordering elements in bond_slave structure, holes can be removed and
saves a cache line.
Before: structure size: 136, sum holes: 10, cachelines:3
After : structure size: 128, sum holes: 2, cachelines:2
Signed-off-by: Bhanuprakash Bodireddy
---
ofproto/bond.c | 4 ++--
1 file changed, 2 ins
This is a simple patch series that reorders members in some structures
to reduce/remove holes and pad bytes, thus saving memory footprint.
This series doesn't change any functionality and care has been taken not
to disturb the grouping of related members variables in structures.
Please note that n
Central Bank of Nigeria (C.B.N),
CBN CORPORATE HEADQUARTER,
NNPC Towers, Central Business District,
P.M.B. 190, Garki, Abuja-Nigeria.
Attn: Sir,
After our meetings few days ago,your contract will be paid in our
security firm united states of American in your country since you are
there in USA you
On 9/8/17, 9:46 AM, "ovs-dev-boun...@openvswitch.org on behalf of Simon Horman"
wrote:
On Tue, Sep 05, 2017 at 05:22:56PM +0800, Yuanhan Liu wrote:
> Flows offloaded to DPDK are identified by rte_flow pointer while OVS
> flows are identified by ufid. This patches adds a hmap to con
On 9/8/17, 9:44 AM, "ovs-dev-boun...@openvswitch.org on behalf of Simon Horman"
wrote:
On Tue, Sep 05, 2017 at 05:22:55PM +0800, Yuanhan Liu wrote:
> So that we could skip the heavy emc processing, notably, the
> miniflow_extract function. A simple PHY-PHY forwarding testing
>
"Bodireddy, Bhanuprakash" writes:
>>Bhanuprakash Bodireddy writes:
>>
>>> This commit registers the packet processing PMD cores to keepalive
>>> framework. Only PMDs that have rxqs mapped will be registered and
>>> actively monitored by KA framework.
>>>
>>> This commit spawns a keepalive thread
On Tue, Sep 05, 2017 at 05:22:59PM +0800, Yuanhan Liu wrote:
> From: Finn Christensen
>
> AFAIK, most (if not all) NICs (including Mellanox and Intel) do not
> support a pure MARK action. It's required to be used together with
> some other actions, like QUEUE.
>
> To workaround it, retry with a
On Tue, Sep 05, 2017 at 05:22:58PM +0800, Yuanhan Liu wrote:
> From: Shachar Beiser
>
> For the DPDK flow offload, which basically just binds a MARK action to
> a flow, the MARK is required to be used together with a QUEUE action for
> the most NICs I'm aware of. The QUEUE action then needs a que
On Tue, Sep 05, 2017 at 05:23:00PM +0800, Yuanhan Liu wrote:
> Inovke netdev class '->flow_del' method on flow deletion. The dpdk netdev
> implementation will then remove the rte flow associated with the ufid.
>
> Co-authored-by: Finn Christensen
> Signed-off-by: Yuanhan Liu
> Signed-off-by: Fin
On Tue, Sep 05, 2017 at 05:23:01PM +0800, Yuanhan Liu wrote:
> From: Finn Christensen
>
> The Intel i40e PMD driver requires the fdir mode set to
> RTE_FDIR_MODE_PERFECT, otherwise, the flow creation would
> be failed.
This appears to update a setting that is used more broadly than
with the hard
On Tue, Sep 05, 2017 at 05:22:57PM +0800, Yuanhan Liu wrote:
> From: Finn Christensen
>
> The basic yet the major part of this patch is to translate the "match"
> to rte flow patterns. And then, we create a rte flow with a MARK action.
> Afterwards, all pkts matches the flow will have the mark id
On Tue, Sep 05, 2017 at 05:22:56PM +0800, Yuanhan Liu wrote:
> Flows offloaded to DPDK are identified by rte_flow pointer while OVS
> flows are identified by ufid. This patches adds a hmap to convert ufid
> to dpdk flow (rte_flow).
>
> Most of the code are stolen from netdev-tc-offloads.c, with so
On Tue, Sep 05, 2017 at 05:22:55PM +0800, Yuanhan Liu wrote:
> So that we could skip the heavy emc processing, notably, the
> miniflow_extract function. A simple PHY-PHY forwarding testing
> shows 53% performance improvement.
>
> Note that though the heavy miniflow_extract is skipped, we
> still h
On Tue, Sep 05, 2017 at 05:22:54PM +0800, Yuanhan Liu wrote:
> This patch associate a flow with a mark id (a uint32_t number) by CMAP.
>
> It re-uses the flow API (more precisely, the ->flow_put method) to setup
> the hw flow. The flow_put implementation then is supposed to create a
> flow with MA
Hi Zoltan
The packet_type initialization that I was referring to as becoming redundant
now was the one moved from to
void
dp_packet_init_dpdk(struct dp_packet *b, size_t allocated)
{
dp_packet_set_allocated(b, allocated);
b->source = DPBUF_DPDK;
b->packet_type = htonl(PT_ETH); <<
Hi Wang,
https://mail.openvswitch.org/pipermail/ovs-dev/2017-August/337309.html
I see it's been acked and is due to be pushed to master with other changes on
the dpdk merge branch so you'll have to apply it manually for now.
/Billy.
> -Original Message-
> From: 王志克 [mailto:wangzh...@j
On 09/08/2017 06:35 AM, Weglicki, MichalX wrote:
Greg,
Patch is rebased and sent to mailing list as V3 (Last patch was supposed to be
V2 - Przemek by accident sent it again as V1).
Br,
Michal.
Thanks, I'll have a look at it!
- Greg
> -Original Message-
> From: Greg Rose [mailto:g
I'm sorry, this one became corrupt because of my e-mail client.
Sent a 4th one:
https://patchwork.ozlabs.org/patch/811617/
https://mail.openvswitch.org/pipermail/ovs-dev/2017-September/338486.html
Zoltan
> -Original Message-
> From: ovs-dev-boun...@openvswitch.org
> [mailto:ovs-dev-boun.
DPDK uses dp-packet pool for storing received packets. The pool is
reused by rxq_recv funcions of the DPDK netdevs. The datapath is
capable to modify the packet_type property of packets. For instance
when encapsulated L3 packets are received on a ptap gre port.
In this case the packet_type property
Greg,
Patch is rebased and sent to mailing list as V3 (Last patch was supposed to be
V2 - Przemek by accident sent it again as V1).
Br,
Michal.
> -Original Message-
> From: Greg Rose [mailto:gvrose8...@gmail.com]
> Sent: Thursday, September 7, 2017 12:02 AM
> To: Weglicki, MichalX
Extend flow key part of data record to include following Information Elements:
- ingressInterface
- ingressInterfaceType
- egressInterface
- egressInterfaceType
- interfaceName
- interfaceDescription
In case of input sampling we don't have information about egress port.
Define templates depending
This commit extends netdev_dpdk_get_status API to include additional
driver-related information: if_type and if_descr.
Signed-off-by: Przemyslaw Szczerbik
Signed-off-by: Michal Weglicki
---
v2-v3: Code rebase.
---
lib/netdev-dpdk.c | 11 +++
1 file changed, 11 insertions(+)
diff --git
On Wed, Sep 6, 2017 at 3:57 AM, Russell Bryant wrote:
> What if a mirror port *only* receives mirrored packets? If the only
> packets it ever receives are mirrored packets, a new flag would not be
> necessary.
>
> Do you intend for the port to operate as both a regular port *and* to
> receive a
If possible, use OpenFlow 1.4 atomic bundle transaction to restore flows.
The patch uses also the highest enabled OpenFlow version to do the queries.
With the actual implementation if you have the default openflow version
disabled ovs-save fails, this patch also fixes that.
Replaced "echo -n" wit
Hi Billy,
I used ovs2.7.0. I searched the git log, and not sure which commit it is. Do
you happen to know?
Yes, I cleared the stats after traffic run.
Br,
Wang Zhike
From: "O Mahony, Billy"
To: "wangzh...@jd.com" , Jan Scheurich
, Darrell Ball ,
"ovs-disc...@openvswitch.org"
Hi Wang,
Thanks for the figures. Unexpected results as you say. Two things come to mind:
I’m not sure what code you are using but the cycles per packet statistic was
broken for a while recently. Ilya posted a patch to fix it so make sure you
have that patch included.
Also remember to reset the
Hi,
Please ignore this patch. It's faulty. We cannot remove initialization
of packet_type from dp_packet_init__(). Beacause of non-dpdk netdevs
still need it. Here is a link to the corrected patch on patchwork:
https://patchwork.ozlabs.org/patch/811445/
Best regards,
Zoltan
> -Original Mes
DPDK uses dp-packet pool for storing received packets. The pool is
reused by rxq_recv funcions of the DPDK netdevs. The datapath is
capable to modify the packet_type property of packets. For instance
when encapsulated L3 packets are received on a ptap gre port.
In this case the packet_type property
> Hi,
>
> I have compiled and built ovs-dpdk using DPDK v17.08 and OVS v2.8.0. The
> NIC that I am using is Mellanox ConnectX-3 Pro, which is a dual port 10G
> NIC. The problem with this NIC is that it provides only one PCI address for
> both the 10G ports.
>
> So when I am trying to add the two
Hi Darrell,
Thank you for your comments! I created a second version and sent it to the dev
list:
https://patchwork.ozlabs.org/patch/811387/
Best regards,
Zoltan
> -Original Message-
> From: Darrell Ball [mailto:db...@vmware.com]
> Sent: Thursday, September 07, 2017 12:45 AM
> To: Zoltán
DPDK uses dp-packet pool for storing received packets. The pool is
reused by rxq_recv funcions of the DPDK netdevs. The datapath is
capable to modify the packet_type property of packets. For instance
when encapsulated L3 packets are received on a ptap gre port.
In this case the packet_type property
Hi All,
I tested below cases, and get some performance data. The data shows there is
little impact for cross NUMA communication, which is different from my
expectation. (Previously I mentioned that cross NUMA would add 60% cycles, but
I can NOT reproduce it any more).
@Jan,
You mentioned
51 matches
Mail list logo