Sam, With this patch we still need to use Yi Yang's "out-of-tree" OVS. Im just adding Ethernet+NSH between the SFF and SF so now we wont have to use that VXLAN workaround mentioned in section 5.8 here [0]
Regards, Brady [0] http://artifacts.opnfv.org/sfc/colorado/docs/design/index.html#document-architecture -----Original Message----- From: "Yang, Yi Y" <yi.y.y...@intel.com<mailto:%22Yang,%20yi%20y%22%20%3cyi.y.y...@intel.com%3e>> To: Sam Hague <sha...@redhat.com<mailto:sam%20hague%20%3csha...@redhat.com%3e>>, Brady Allen Johnson <brady.allen.john...@ericsson.com<mailto:brady%20allen%20johnson%20%3cbrady.allen.john...@ericsson.com%3e>> Cc: sfc-dev@lists.opendaylight.org <sfc-dev@lists.opendaylight.org<mailto:%22sfc-...@lists.opendaylight.org%22%20%3csfc-...@lists.opendaylight.org%3e>> Subject: RE: [sfc-dev] Ethernet+NSH encapsulation now implemented Date: Tue, 21 Feb 2017 00:25:58 +0000 Sam, previous opnfv sfc releases have used our out-of-tree ovs version, opnfv has its own package repo for our out-of-tree ovs version. From: sfc-dev-boun...@lists.opendaylight.org [mailto:sfc-dev-boun...@lists.opendaylight.org] On Behalf Of Sam Hague Sent: Tuesday, February 21, 2017 1:40 AM To: Brady Allen Johnson <brady.allen.john...@ericsson.com> Cc: sfc-dev@lists.opendaylight.org Subject: Re: [sfc-dev] Ethernet+NSH encapsulation now implemented Brady, does this still require an out-of-tree ovs version? Any idea how the work is going to get that upstreamed? Can we use the out-of-tree distro in the opnfv official tests or does that need to be an official ovs? Thanks, Sam On Mon, Feb 20, 2017 at 10:11 AM, Brady Allen Johnson <brady.allen.john...@ericsson.com<mailto:brady.allen.john...@ericsson.com>> wrote: Hello, I just submit a patch [0] that implements Ethernet+NSH transport/encapsulation between the SFF and SFs. The transport/encapsulation between the classifier-SFF and SFF-SFF will still need to be Vxgpe+NSH. I'll update the "resources/configuration-examples" with JSON configurations in a separate patch. This will be very nice for OPNFV SFC since it will allow us to remove the ugly work-around we are currently using to make the Vxgpe tunnel terminate in the SF so it can see the NSH header. I tested this locally and got the following flows: Formatted flows for easier reading: (complete flows below) cookie=0x14, table=0, priority=5 actions=goto_table:1 cookie=0x14, table=1, priority=250,nsp=6 actions=goto_table:4 cookie=0x14, table=1, priority=250,nsp=8388614 actions=goto_table:4 cookie=0x14, table=1, priority=5 actions=drop cookie=0x14, table=2, priority=5 actions=goto_table:3 cookie=0x14, table=3, priority=5 actions=goto_table:4 cookie=0x14, table=4, priority=5 actions=goto_table:10 cookie=0x14, table=4, priority=550,nsi=255,nsp=6 actions=load:0x3->NXM_NX_ENCAP_ETH_SRC[], load:0x2->NXM_NX_ENCAP_ETH_DST[], goto_table:10 cookie=0x14, table=4, priority=550,nsi=255,nsp=8388614 actions=load:0x3->NXM_NX_ENCAP_ETH_SRC[], load:0x2->NXM_NX_ENCAP_ETH_DST[], goto_table:10 cookie=0x14, table=10, priority=5 actions=drop cookie=0xba5eba1100000201, table=10, priority=650,nsi=255,nsp=6 actions=move:NXM_NX_NSH_C1[]->NXM_NX_NSH_C1[], move:NXM_NX_NSH_C2[]->NXM_NX_NSH_C2[], move:NXM_NX_NSI[]->NXM_NX_NSI[], move:NXM_NX_NSP[0..23]->NXM_NX_NSP[0..23], move:NXM_NX_NSH_MDTYPE[]->NXM_NX_NSH_MDTYPE[], load:0x3->NXM_NX_NSH_NP[], IN_PORT cookie=0xba5eba1100000102, table=10, priority=660,nsi=254,nsp=6,nshc1=0 actions=load:0x4->NXM_NX_TUN_GPE_NP[],IN_PORT cookie=0xba5eba1100000103, table=10, priority=650,nsi=254,nsp=6 actions=move:NXM_NX_NSH_MDTYPE[]->NXM_NX_NSH_MDTYPE[], move:NXM_NX_NSH_NP[]->NXM_NX_NSH_NP[], 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=0xba5eba1100000201, table=10, priority=650,nsi=255,nsp=8388614 actions=move:NXM_NX_NSH_C1[]->NXM_NX_NSH_C1[], move:NXM_NX_NSH_C2[]->NXM_NX_NSH_C2[], move:NXM_NX_NSI[]->NXM_NX_NSI[], move:NXM_NX_NSP[0..23]->NXM_NX_NSP[0..23], move:NXM_NX_NSH_MDTYPE[]->NXM_NX_NSH_MDTYPE[], load:0x3->NXM_NX_NSH_NP[], IN_PORT cookie=0xba5eba1100000102, table=10, priority=660,nsi=254,nsp=8388614,nshc1=0 actions=load:0x4->NXM_NX_TUN_GPE_NP[],IN_PORT cookie=0xba5eba1100000103, table=10, priority=650,nsi=254,nsp=8388614 actions=move:NXM_NX_NSH_MDTYPE[]->NXM_NX_NSH_MDTYPE[], move:NXM_NX_NSH_NP[]->NXM_NX_NSH_NP[], 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 Complete flow dump: cookie=0x14, duration=17.291s, table=0, n_packets=0, n_bytes=0, priority=5 actions=goto_table:1 cookie=0x14, duration=17.289s, table=1, n_packets=0, n_bytes=0, priority=250,nsp=6 actions=goto_table:4 cookie=0x14, duration=17.273s, table=1, n_packets=0, n_bytes=0, priority=250,nsp=8388614 actions=goto_table:4 cookie=0x14, duration=17.288s, table=1, n_packets=0, n_bytes=0, priority=5 actions=drop cookie=0x14, duration=17.341s, table=2, n_packets=0, n_bytes=0, priority=5 actions=goto_table:3 cookie=0x14, duration=17.296s, table=3, n_packets=0, n_bytes=0, priority=5 actions=goto_table:4 cookie=0x14, duration=17.293s, table=4, n_packets=0, n_bytes=0, priority=550,nsi=255,nsp=6 actions=load:0x3->NXM_NX_ENCAP_ETH_SRC[],load:0x2->NXM_NX_ENCAP_ETH_DST[],goto_table:10 cookie=0x14, duration=17.284s, table=4, n_packets=0, n_bytes=0, priority=550,nsi=255,nsp=8388614 actions=load:0x3->NXM_NX_ENCAP_ETH_SRC[],load:0x2->NXM_NX_ENCAP_ETH_DST[],goto_table:10 cookie=0x14, duration=17.291s, table=4, n_packets=0, n_bytes=0, priority=5 actions=goto_table:10 cookie=0xba5eba1100000102, duration=17.302s, table=10, n_packets=0, n_bytes=0, priority=660,nsi=254,nsp=6,nshc1=0 actions=load:0x4->NXM_NX_TUN_GPE_NP[],IN_PORT cookie=0xba5eba1100000102, duration=17.281s, table=10, n_packets=0, n_bytes=0, priority=660,nsi=254,nsp=8388614,nshc1=0 actions=load:0x4->NXM_NX_TUN_GPE_NP[],IN_PORT cookie=0xba5eba1100000103, duration=17.309s, table=10, n_packets=0, n_bytes=0, priority=650,nsi=254,nsp=6 actions=move:NXM_NX_NSH_MDTYPE[]->NXM_NX_NSH_MDTYPE[],move:NXM_NX_NSH_NP[]->NXM_NX_NSH_NP[],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=0xba5eba1100000201, duration=17.296s, table=10, n_packets=0, n_bytes=0, priority=650,nsi=255,nsp=6 actions=move:NXM_NX_NSH_C1[]->NXM_NX_NSH_C1[],move:NXM_NX_NSH_C2[]->NXM_NX_NSH_C2[],move:NXM_NX_NSI[]->NXM_NX_NSI[],move:NXM_NX_NSP[0..23]->NXM_NX_NSP[0..23],move:NXM_NX_NSH_MDTYPE[]->NXM_NX_NSH_MDTYPE[],load:0x3->NXM_NX_NSH_NP[],IN_PORT cookie=0xba5eba1100000201, duration=17.279s, table=10, n_packets=0, n_bytes=0, priority=650,nsi=255,nsp=8388614 actions=move:NXM_NX_NSH_C1[]->NXM_NX_NSH_C1[],move:NXM_NX_NSH_C2[]->NXM_NX_NSH_C2[],move:NXM_NX_NSI[]->NXM_NX_NSI[],move:NXM_NX_NSP[0..23]->NXM_NX_NSP[0..23],move:NXM_NX_NSH_MDTYPE[]->NXM_NX_NSH_MDTYPE[],load:0x3->NXM_NX_NSH_NP[],IN_PORT cookie=0xba5eba1100000103, duration=17.275s, table=10, n_packets=0, n_bytes=0, priority=650,nsi=254,nsp=8388614 actions=move:NXM_NX_NSH_MDTYPE[]->NXM_NX_NSH_MDTYPE[],move:NXM_NX_NSH_NP[]->NXM_NX_NSH_NP[],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=0x14, duration=17.336s, table=10, n_packets=0, n_bytes=0, priority=5 actions=drop Regards, Brady [0] https://git.opendaylight.org/gerrit/#/c/42313/ _______________________________________________ sfc-dev mailing list sfc-dev@lists.opendaylight.org<mailto:sfc-dev@lists.opendaylight.org> https://lists.opendaylight.org/mailman/listinfo/sfc-dev
_______________________________________________ sfc-dev mailing list sfc-dev@lists.opendaylight.org https://lists.opendaylight.org/mailman/listinfo/sfc-dev
