Re: [vpp-dev] Is VPP IPSec implementation thread safe?

2018-06-29 Thread Damjan Marion
Dear Vamsi, It was long long time ago when I wrote original ipsec code, so I don't remember details anymore. For IKEv2, i would suggest using external implementation, as ikev2 stuff is just PoC quality code. -- Damjan > On 29 Jun 2018, at 18:38, Vamsi Krishna wrote: > > Hi Damjan, Dave, >

Re: [vpp-dev] Is VPP IPSec implementation thread safe?

2018-06-29 Thread Vamsi Krishna
Hi Damjan, Dave, Can you please also answer the questions I had in the email just before Jim hijacked the thread. Thanks Vamsi On Fri, Jun 29, 2018 at 3:06 PM, Damjan Marion wrote: > > Hi Jim, > > Atomic add sounds like a reasonable solution to me... > > -- > Damjan > > On 28 Jun 2018, at 09:2

[vpp-dev] VPP ARM Build/Installation

2018-06-29 Thread Jit Mehta
Hello, Is there a way I can download and install aarch64 binaries? If not, is there a wiki that lists how to build aarch64 targets? Thanks, Jit -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#9746): https://lists.fd.io/g/vpp-dev/message/9746 Mute

Re: [vpp-dev] Connection issue between container (slave) and host vpp (master) with memif

2018-06-29 Thread Damjan Marion
Depends on what do you expect from packet generator. I personally map physical NIC VF into container and use them to feed packets in from remote T-Rex box. You can also use VPP packet-generator to play PCAP file or craft packets and send them to memif interface, but I wold not expect great perf

Re: [vpp-dev] ip header checksum calculation issue

2018-06-29 Thread Dave Barach via Lists.Fd.Io
Glad to hear it. Note that fully-unaligned checksum performance probably won’t be anything to write home about, but at least the code produces the right answer... D. From: vpp-dev@lists.fd.io On Behalf Of Yao, Chengqiang Sent: Friday, June 29, 2018 6:03 AM To: Damjan Marion Cc: vpp-dev@lists.

Re: [vpp-dev] ip header checksum calculation issue

2018-06-29 Thread Yao, Chengqiang
Hi Marion, Yes, I just tried latest code committed 10 days ago, it works. Thanks for your reply. I will also try your optimized code later. Best Regards, Chengqiang Yao From: Damjan Marion [mailto:dmar...@me.com] Sent: Friday, June 29, 2018 5:34 PM To: Yao, Chengqiang Cc: vpp-dev@lists.f

Re: [vpp-dev] CPU instructions

2018-06-29 Thread Damjan Marion
N2600 seems to be very old ATOM CPU. and it doesn't support SSE4.2 which is our baseline. To get VPP working it will likely be few hour exercise to get things right at several places. For DPDK, you likely need to change dpdk/Makefile -- Damjan > On 27 Jun 2018, at 06:16, xulang wrote: > >

Re: [vpp-dev] Is VPP IPSec implementation thread safe?

2018-06-29 Thread Damjan Marion
Hi Jim, Atomic add sounds like a reasonable solution to me... -- Damjan > On 28 Jun 2018, at 09:26, Jim Thompson wrote: > > All, > > I don't know if any of the previously-raised issues occur in real-life. > Goodness knows we've run billions of IPsec packets in the test harnesses > (harne

Re: [vpp-dev] ip header checksum calculation issue

2018-06-29 Thread Damjan Marion
Hi Yao, Dave recently fixed bug related to alignment, Are you using latest code? Also, few days ago i merged new significantly-faster SIMD based ip checksum code[1], you can try that one also... [1] https://git.fd.io/vpp/commit/?id=08bca80 Thanks, -- Damjan > On 29 Jun 2018, at 09:27, Yao,

Re: [vpp-dev] Connection issue between container (slave) and host vpp (master) with memif

2018-06-29 Thread Damjan Marion
Hi Chakri, How do you map socket file into the container filesystem? Tnx, -- Damjan > On 29 Jun 2018, at 07:40, chakravarthy.arise...@viasat.com wrote: > > Hi, > > How do we connect memif inside host to memif inside a container? Somehow, > the container is not able to communicate with host

[vpp-dev] ip header checksum calculation issue

2018-06-29 Thread Yao, Chengqiang
Hi, I have an issue when calculating the checksum of two IP headers with ip4_header_checksum function. The content of these two IP headers are exactly the same. The only difference between these 2 IP headers is that, one IP header's address is 64 bits aligned, while the other one is not (My tes