Hi Yi, I can see three patches related to NSH extensions
https://git.opendaylight.org/gerrit/#/c/30481/ https://git.opendaylight.org/gerrit/#/c/37937/ https://git.opendaylight.org/gerrit/#/c/27643/ Can you please point me to correct patch, because 30481 and 37937 has similar code in the patch. I am reviewing 37937 because looks like it has eth_nsh support. Apart from this, looks like NSH extensions are not yet finalized (i am looking from OpenFlow message format perspective)? is my understanding correct? On Thu, Apr 28, 2016 at 5:15 PM, Yang, Yi Y <[email protected]> wrote: > Dave, thank you for your information, I will send this to ovs-dev mail > list for more feedbacks, current roadblock is Linux kernel, Jiri Benc has > pushed VxLAN-gpe implementation into net-next tree, but it only can support > L3 mode, that means VxLAN-gpe port only can handle VxLAN-gpe + NSH + > original L2 but current ovs requires VxLAN-gpe + Eth + NSH + Original L2, > Simon Horman (Netronome guy) has posted a very-large patch set to remove > this ovs requirement and appended two new openflow actions push_eth and > pop_eth, but ovs community haven't merged it so far because this change > broke ovs requirement, this is another roadblock. > > Actually ovs community can't accept ovs patches until kernel part of > implementation has been accepted by Linux kernel, this is the real > roadblock. > > -----Original Message----- > From: Dave Neary [mailto:[email protected]] > Sent: Friday, April 29, 2016 7:54 AM > To: Thomas F Herbert <[email protected]>; Brady Allen Johnson < > [email protected]>; Yang, Yi Y <[email protected]>; Sam > Hague <[email protected]> > Cc: opendaylight sfc <[email protected]>; > [email protected]; [email protected]; > [email protected] > Subject: Re: [opnfv-tech-discuss] [SFC] [sfc-dev] ovs nsh patches are > available for public use > > Hi, > > The feedback I got from the OVS maintainers this week is that they are > happy to consider the patches, but are concerned that it takes some months > to get a new revision when a patchset gets feedback. > > I would like to understand - what do people feel is the roadblock from our > end? > > Thanks, > Dave. > > > On 04/28/2016 11:18 AM, Thomas F Herbert wrote: > > > > > > On 4/27/16 8:31 AM, Brady Allen Johnson wrote: > >> Yi, > >> > >> I installed your OVS NSH patch, and also built ODL with your > >> OpenFlowPlugin and SFC patches and am able to get the below NSH flows > >> that look good to me. > >> > >> Once we get the patches merged in ODL, we can start working on using > >> this in the OPNFV SFC Colorado release. > > I don't think we can merge these patches yet into ovs-nfv project yet > > because they are not upstream. However, I have created temporarily a > > fork of ovs-nfv repo that builds these patches into an RPM if you want > > to try it. This RPM should support both Linux kernel data-plane and DPDK. > > https://github.com/tfherbert/buildovsnsh > > > > --TFH > >> > >> Nice work :) > >> > >> Regards, > >> > >> Brady > >> > >> OFPST_FLOW reply (OF1.3) (xid=0x2): > >> cookie=0x14, duration=0.001s, table=0, n_packets=0, n_bytes=0, > >> priority=5 actions=goto_table:1 > >> cookie=0x14, duration=0.001s, table=1, n_packets=0, n_bytes=0, > >> priority=250,nsp=19 actions=goto_table:4 > >> cookie=0x14, duration=0.001s, table=1, n_packets=0, n_bytes=0, > >> priority=250,nsp=8388627 actions=goto_table:4 > >> cookie=0x14, duration=0.001s, table=1, n_packets=0, n_bytes=0, > >> priority=5 actions=drop > >> cookie=0x14, duration=0.001s, table=2, n_packets=0, n_bytes=0, > >> priority=5 actions=goto_table:3 > >> cookie=0x14, duration=0.001s, table=3, n_packets=0, n_bytes=0, > >> priority=5 actions=goto_table:4 > >> cookie=0x14, duration=0.001s, table=4, n_packets=0, n_bytes=0, > >> priority=550,nsi=255,nsp=19 > >> actions=load:0xbfa8386e->NXM_NX_TUN_IPV4_DST[],goto_table:10 > >> cookie=0x14, duration=0.001s, table=4, n_packets=0, n_bytes=0, > >> priority=550,nsi=255,nsp=8388627 > >> actions=load:0xbfa8386e->NXM_NX_TUN_IPV4_DST[],goto_table:10 > >> cookie=0x14, duration=0.001s, table=4, n_packets=0, n_bytes=0, > >> priority=5 actions=goto_table:10 > >> cookie=0xba5eba11ba5eba11, duration=0.976s, table=10, > >> n_packets=0, n_bytes=0, priority=660,nsi=254,nsp=19,nshc1=0 > >> actions=load:0x4->NXM_NX_TUN_GPE_NP[],IN_PORT > >> cookie=0xba5eba11ba5eba11, duration=0.976s, table=10, > >> n_packets=0, n_bytes=0, priority=660,nsi=254,nsp=8388627,nshc1=0 > >> actions=load:0x4->NXM_NX_TUN_GPE_NP[],IN_PORT > >> cookie=0xba5eba11ba5eba11, duration=0.976s, table=10, > >> n_packets=0, n_bytes=0, priority=650,nsi=254,nsp=19 > >> > > actions=move:NXM_NX_NSI[]->NXM_NX_NSI[],move:NXM_NX_NSP[0..23]->NXM_NX_NSP[0..23],move:NXM_NX_NSH_C1[]->NXM_NX_TUN_IPV4_DST[],move:NXM_NX_NSH_C2[]->NXM_NX_TUN_ID[0..31],load:0x4->NXM_NX_TUN_GPE_NP[],IN_PORT > >> cookie=0xba5eba11ba5eba11, duration=0.001s, table=10, > >> n_packets=0, n_bytes=0, priority=650,nsi=255,nsp=19 > >> > > actions=move:NXM_NX_NSH_C1[]->NXM_NX_NSH_C1[],move:NXM_NX_NSH_C2[]->NXM_NX_NSH_C2[],move:NXM_NX_TUN_ID[0..31]->NXM_NX_TUN_ID[0..31],load:0x4->NXM_NX_TUN_GPE_NP[],IN_PORT > >> cookie=0xba5eba11ba5eba11, duration=0.001s, table=10, > >> n_packets=0, n_bytes=0, priority=650,nsi=254,nsp=8388627 > >> > > actions=move:NXM_NX_NSI[]->NXM_NX_NSI[],move:NXM_NX_NSP[0..23]->NXM_NX_NSP[0..23],move:NXM_NX_NSH_C1[]->NXM_NX_TUN_IPV4_DST[],move:NXM_NX_NSH_C2[]->NXM_NX_TUN_ID[0..31],load:0x4->NXM_NX_TUN_GPE_NP[],IN_PORT > >> cookie=0xba5eba11ba5eba11, duration=0.001s, table=10, > >> n_packets=0, n_bytes=0, priority=650,nsi=255,nsp=8388627 > >> > > actions=move:NXM_NX_NSH_C1[]->NXM_NX_NSH_C1[],move:NXM_NX_NSH_C2[]->NXM_NX_NSH_C2[],move:NXM_NX_TUN_ID[0..31]->NXM_NX_TUN_ID[0..31],load:0x4->NXM_NX_TUN_GPE_NP[],IN_PORT > >> cookie=0x14, duration=0.001s, table=10, n_packets=0, n_bytes=0, > >> priority=5 actions=drop > >> > >> > >> =========================== > >> == Easier to read output == > >> =========================== > >> > >> table=0, priority=5 actions=goto_table:1 > >> > >> table=1, priority=250,nsp=19 actions=goto_table:4 > >> table=1, priority=250,nsp=8388627 actions=goto_table:4 > >> table=1, priority=5 actions=drop > >> > >> table=2, priority=5 actions=goto_table:3 > >> table=3, priority=5 actions=goto_table:4 > >> > >> table=4, priority=5 actions=goto_table:10 > >> table=4, priority=550,nsi=255,nsp=19 > >> actions=load:0xbfa8386e->NXM_NX_TUN_IPV4_DST[],goto_table:10 > >> table=4, priority=550,nsi=255,nsp=8388627 > >> actions=load:0xbfa8386e->NXM_NX_TUN_IPV4_DST[],goto_table:10 > >> > >> > >> table=10, priority=5 actions=drop > >> table=10, priority=660,nsi=254,nsp=19,nshc1=0 > >> actions=load:0x4->NXM_NX_TUN_GPE_NP[],IN_PORT > >> table=10, priority=650,nsi=255,nsp=19 > >> actions=move:NXM_NX_NSH_C1[]->NXM_NX_NSH_C1[], > >> move:NXM_NX_NSH_C2[]->NXM_NX_NSH_C2[], > >> move:NXM_NX_TUN_ID[0..31]->NXM_NX_TUN_ID[0..31], > >> load:0x4->NXM_NX_TUN_GPE_NP[], > >> IN_PORT > >> table=10, priority=650,nsi=254,nsp=19 > >> actions=move:NXM_NX_NSI[]->NXM_NX_NSI[], > >> move:NXM_NX_NSP[0..23]->NXM_NX_NSP[0..23], > >> move:NXM_NX_NSH_C1[]->NXM_NX_TUN_IPV4_DST[], > >> move:NXM_NX_NSH_C2[]->NXM_NX_TUN_ID[0..31], > >> load:0x4->NXM_NX_TUN_GPE_NP[], > >> IN_PORT > >> > >> table=10, priority=660,nsi=254,nsp=8388627,nshc1=0 > >> actions=load:0x4->NXM_NX_TUN_GPE_NP[],IN_PORT > >> table=10, priority=650,nsi=255,nsp=8388627 > >> actions=move:NXM_NX_NSH_C1[]->NXM_NX_NSH_C1[], > >> move:NXM_NX_NSH_C2[]->NXM_NX_NSH_C2[], > >> move:NXM_NX_TUN_ID[0..31]->NXM_NX_TUN_ID[0..31], > >> load:0x4->NXM_NX_TUN_GPE_NP[], > >> IN_PORT > >> table=10, priority=650,nsi=254,nsp=8388627 > >> actions=move:NXM_NX_NSI[]->NXM_NX_NSI[], > >> move:NXM_NX_NSP[0..23]->NXM_NX_NSP[0..23], > >> move:NXM_NX_NSH_C1[]->NXM_NX_TUN_IPV4_DST[], > >> move:NXM_NX_NSH_C2[]->NXM_NX_TUN_ID[0..31], > >> load:0x4->NXM_NX_TUN_GPE_NP[], > >> IN_PORT > >> > >> > >> > >> > >> > >> On 18/04/16 03:20, Yang, Yi Y wrote: > >>> > >>> Hi, Sam > >>> > >>> > >>> > >>> Replies inline, please check. > >>> > >>> > >>> > >>> *From:*Sam Hague [mailto:[email protected]] > >>> *Sent:* Friday, April 15, 2016 7:41 PM > >>> *To:* Brady Allen Johnson <[email protected]> > >>> *Cc:* Yang, Yi Y <[email protected]>; > >>> [email protected]; opendaylight sfc > >>> <[email protected]>; [email protected]; > >>> [email protected] > >>> *Subject:* Re: [opnfv-tech-discuss] [sfc-dev] ovs nsh patches are > >>> available for public use > >>> > >>> > >>> > >>> - what minimum linux kernel is required? OVS version? > >>> > >>> This depends on ovs itself, current ovs git master > >>> > >>> > >>> > >>> - I thought there were changes needed to the kernel, but these look > >>> to be all ovs patches > >>> > >>> ovs includes kernel modules, but to upstream them, kernel modules > >>> must be accepted by Linux firstly, our team is working on this, but > >>> progress is very slow. > >>> > >>> > >>> > >>> - do we need the related kernel gpe patch? > >>> > >>> Jiri has pushed gpe patch, but that only support L3 mode, but > >>> current ovs requires it in L2 mode, they didn't make an agreement on > >>> this work, progress will be slow. > >>> > >>> > >>> > >>> - what versions of linux distributions have been used? > >>> > >>> It is distribution agnostic, just build it as before. > >>> > >>> > >>> > >>> - eth_nsh is not in the odl ovsdb - is a patch needed there also? > >>> > >>> - does openflowplugin require support? > >>> > >>> Yes, https://git.opendaylight.org/gerrit/30481 is for your > >>> reference, will update it to adapt to these patches. > >>> > >>> > >>> > >>> On Fri, Apr 15, 2016 at 5:07 AM, Brady Allen Johnson > >>> <[email protected]> wrote: > >>> > >>> Yang, > >>> > >>> Thanks for this! > >>> > >>> We'll start looking into using this now. > >>> > >>> Do you know what's the latest status of getting NSH into OVS > >>> officially? What's the difference between this patch in your > >>> github and what will be available officially? > >>> > >>> What are the system requirements for using your patch, like Linux > >>> kernel, OVS, etc. It doesnt look like it, but is DPDK needed with > >>> this patch? > >>> > >>> Thanks, > >>> > >>> Brady > >>> > >>> On 15/04/16 10:53, Yang, Yi Y wrote: > >>> > >>> Hi, folks > >>> > >>> > >>> > >>> We have pushed ovs nsh patches in > >>> https://github.com/yyang13/ovs_nsh_patches, you can follow > >>> instructions in README.md to apply these patches and build > >>> ovs to support nsh, please let me know if you have any > >>> problem during using. > >>> > >>> > >>> > >>> > >>> > >>> _______________________________________________ > >>> > >>> sfc-dev mailing list > >>> > >>> [email protected] > >>> <mailto:[email protected]> > >>> > >>> https://lists.opendaylight.org/mailman/listinfo/sfc-dev > >>> > >>> > >>> > >>> > >>> _______________________________________________ > >>> opnfv-tech-discuss mailing list > >>> [email protected] > >>> <mailto:[email protected]> > >>> https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss > >>> > >>> > >>> > >> > >> > >> > >> _______________________________________________ > >> opnfv-tech-discuss mailing list > >> [email protected] > >> https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss > > > > -- > > *Thomas F Herbert* > > SDN Group > > Office of Technology > > *Red Hat* > > > > > > _______________________________________________ > > opnfv-tech-discuss mailing list > > [email protected] > > https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss > > > > -- > Dave Neary - NFV/SDN Community Strategy > Open Source and Standards, Red Hat - http://community.redhat.com > Ph: +1-978-399-2182 / Cell: +1-978-799-3338 > _______________________________________________ > openflowplugin-dev mailing list > [email protected] > https://lists.opendaylight.org/mailman/listinfo/openflowplugin-dev > -- Thanks Anil
_______________________________________________ openflowplugin-dev mailing list [email protected] https://lists.opendaylight.org/mailman/listinfo/openflowplugin-dev
