> When the IP message enters the GRE、L2TP、MPLS tunnel， does the VPP support
> sharding?
I'm not quite sure what "sharding" would mean in this context.
Are you thinking of point to multipoint tunnels? Or ECMP to a set of anycast
tunnel destinations or something completely different?
Cheers,
Ole
I'm sorry I didn't say it clearly.
My questions:
Does the VPP support ip fragment in GRE、L2TP、MPLS tunnel?
Whether the tunnel supports configuring the MTU？
Whether the receiver supports ip reassembly？
Thanks，
xyxue
From: otroan
Date: 2017-05-08 15:16
To: 薛欣颖
CC: vpp-dev
Subject: Re: [vpp-dev
A detailed description of my problem：
I would like to test ip fragment reassembly,but it didn't work.Is there
anything wrong in my configuration：
set interface mtu 2000 host-eth2
Did I miss anything configuration?
If vpp support ip fragment reassembly on host interface . Does VPP support the
IP
Hi Xyhue,
As far as I know IPv4 fragmentation / reassembly is not supported on the tunnel
types you listed.
IP{4,6} fragmentation / reassembly is supported for IPv4 over IPv6 tunnels
(src/vnet/map), but I think that's it.
In general the recommendation is to avoid fragmentation in tunnels becaus
Hi
I have read the SecurityGroups page in your Wiki about acl plugin and
executed some of functionality test on this feature. So, my question is
whether we have input and output acl simultaneously? In other words, could
have we an acl that affect on packets that are recieved from interface A
and e
Hi,
I tested statefull ACL by writing an ACL rule with action of permit+reflect
and it didn't work. Is this feature completed in 17.04 and is expected to
work properly?
I define one deny ACL rule and set it as inbound ACL for interface B and
also define a permit+reflect ACL rule and set it as inb
>
> On 3 May 2017, at 13:28, Nagaprabhanjan Bellaru wrote:
>
> Hi,
>
> It looks like dpdk_device_input() - is not checking if there is a vlan header
> in the packet or not and always sets the buffer->current_data to 14
> (smac+dmac+ethtype). Because of that ip4_input is not able to recognize
On 3 May 2017, at 17:20, Jon Loeliger
mailto:j...@netgate.com>> wrote:
Hey VPP Builders,
Do you ever use "cd build-root; make distclean"?
Does it look sort of like this:
jdl $ cd build-root/
jdl $ make distclean
rm -rf /home/jdl/workspace/vpp/build-root/build-*/
rm -rf /home/jdl/workspace/vpp
On 5 May 2017, at 22:37, Mircea Orban
mailto:mior...@hotmail.com>> wrote:
I have a Fortville NIC (XL710-QDA1) with one QSFP+ port that supports two
modes: 1X40g and 4X10g.
While in 1X40g mode everything seems to be fine, when I run VPP in 4X10g mode
some issues seems to occur:
-
Here it is.
Thanks,
Mircea
vpp#
vpp# show pci
Address Sock VID:PID Link Speed Driver Product Name
Vital Product Data
:09:00.0 0 15b3:1007 8.0 GT/s x8 mlx4_core
:0b:00.0 0 8086:1584 8.0 GT/s x8 i40eXL710 40GbE
Controller
Thanks,
what about PF case? Can you also grab output for PF case?
On 8 May 2017, at 16:47, Mircea Orban
mailto:mior...@hotmail.com>> wrote:
Here it is.
Thanks,
Mircea
vpp#
vpp# show pci
Address Sock VID:PID Link Speed Driver Product Name
Vital Product
It would be the same output because it’s the same server:
- :0b:00.0-3 are the four 10G PFs
- :0b:02.0 to 4 - 5 VFs for :0b:00.0
- And :0b:06.0 to 4 - 5VFs for :0b:00.1.
Thanks,
Mircea
From: Damjan Marion (damarion) [mailto:damar...@cisco.com]
Hi,
I have installed FD.IO on our system and am currently receiving a video stream
that we will eventually be processing. We would like to do some benchmarking to
evaluate the performance of FD.IO in our situation.
Does anyone have a suggesting as to how this could be done.
Thanks,
Guy
__
Hi,
Reflect is not evaluated across multiple interfaces, both ACLs must be on the
same interface. Mentally you can imagine like a tiny firewall attached sitting
inline with the VPP interface.
You can look at the test/test_acl_plugin*.py files which are the unit tests for
more specifics on how
--a
> On 8 May 2017, at 13:25, mahmood gholipour
> wrote:
>
> Hi
> I have read the SecurityGroups page in your Wiki about acl plugin and
> executed some of functionality test on this feature. So, my question is
> whether we have input and output acl simultaneously?
Yes, an interface can h
Can you try this one:
https://gerrit.fd.io/r/#/c/6614/
It should fix PF case….
On 8 May 2017, at 17:01, Mircea Orban
mailto:mior...@hotmail.com>> wrote:
It would be the same output because it’s the same server:
- :0b:00.0-3 are the four 10G PFs
- :0b:02.0 to 4 - 5
16 matches
Mail list logo
