nsh_sfc-dev@lists.fd.io | VPP-43 NSH removal progress

Home Messages Hashtags Subgroups

Date Date 1 - 4 of 4

VPP-43 NSH removal progress

Keith Burns <alagalah@...>

Got VXLAN-GPE separated from NSH

Just have to find a way to test GRE and it should be pretty straight forward to make generic GRE talk to new NSH node. Then I'll publish patch for review.

New commands:

vxlan-gpe tunnel:

vxlan-gpe tunnel local 192.168.150.90 remote 192.168.150.91 vni 16777215 next-nsh

(note next-nsh is not necessary... this is just there for classifier case... its not really needed at all)

nsh entry: (defines an NSH header)

nsh entry nsp 1 nsi 255 c1 9901 c2 9902 c3 9903 c4 9904 md-type 1

nsh map: maps two nsh headers together. Note that if you do "encap-none" it will attempt to decap it ... this is not currently supported and can be done in NSH_SFC

nsh map nsp 1 nsi 255 mapped-nsp 1 mapped-nsi 255 encap-vxlan-gpe-intf 6

Packet 6

00:00:14:216827: dpdk-input

GigabitEthernet0/5/0 rx queue 0

buffer 0x10d33: current data 0, length 106, free-list 0, totlen-nifb 0, trace 0x5

PKT MBUF: port 0, nb_segs 1, pkt_len 106

buf_len 2176, data_len 106, ol_flags 0x0,

packet_type 0x0

IP4: 52:54:00:24:19:a5 -> 52:54:00:27:0b:2f

UDP: 192.168.150.91 -> 192.168.150.90

tos 0x00, ttl 64, length 92, checksum 0x4d08

fragment id 0x3f82, flags DONT_FRAGMENT

UDP: 4790 -> 4790

length 72, checksum 0x9947

00:00:14:216861: ethernet-input

IP4: 52:54:00:24:19:a5 -> 52:54:00:27:0b:2f

00:00:14:216875: ip4-input

UDP: 192.168.150.91 -> 192.168.150.90

tos 0x00, ttl 64, length 92, checksum 0x4d08

fragment id 0x3f82, flags DONT_FRAGMENT

UDP: 4790 -> 4790

length 72, checksum 0x9947

00:00:14:216888: ip4-local

fib 0 adj-idx 4 : local 192.168.150.90/24 flow hash: 0x00000000

00:00:14:216898: ip4-udp-lookup

UDP: src-port 4790 dst-port 4790

00:00:14:216908: vxlan-gpe-input

VXLAN-GPE: tunnel 0 next 4 error 0

00:00:14:216917: nsh-input

nsh ver 0 len 0 (0 bytes) md_type 1 next_protocol 1

service path 1 service index 255

c1 9901 c2 9902 c3 9903 c4 9904

00:00:14:216925: vxlan-gpe-encap

VXLAN-GPE-ENCAP: tunnel 0

00:00:14:216932: ip4-rewrite-transit

fib 0 adj-idx 5 : GigabitEthernet0/5/0

IP4: 52:54:00:27:0b:2f -> 52:54:00:24:19:a5 flow hash: 0x00000000

IP4: 52:54:00:27:0b:2f -> 52:54:00:24:19:a5

UDP: 192.168.150.90 -> 192.168.150.91

tos 0x00, ttl 253, length 92, checksum 0x0f8a

fragment id 0x0000

UDP: 4790 -> 4790

length 72, checksum 0x0000

00:00:14:216944: GigabitEthernet0/5/0-output

GigabitEthernet0/5/0

IP4: 52:54:00:27:0b:2f -> 52:54:00:24:19:a5

UDP: 192.168.150.90 -> 192.168.150.91

tos 0x00, ttl 253, length 92, checksum 0x0f8a

fragment id 0x0000

UDP: 4790 -> 4790

length 72, checksum 0x0000

00:00:14:216952: GigabitEthernet0/5/0-tx

GigabitEthernet0/5/0 tx queue 0

buffer 0x10d33: current data 0, length 106, free-list 0, totlen-nifb 0, trace 0x5

IP4: 52:54:00:27:0b:2f -> 52:54:00:24:19:a5

UDP: 192.168.150.90 -> 192.168.150.91

tos 0x00, ttl 253, length 92, checksum 0x0f8a

fragment id 0x0000

UDP: 4790 -> 4790

length 72, checksum 0x0000

Ni, Hongjun

#65

Hi Keith,

Just a question about the nsh map:

For both nsh headers that have the same SPI and SI, are the context data of them the same?

If they are not, how can distinguish these two nsh headers?

Thanks,

Hongjun

Show quoted text

From: nsh_sfc-dev-bounces@... [mailto:nsh_sfc-dev-bounces@...] On Behalf Of Keith Burns
Sent: Monday, May 9, 2016 6:37 AM
To: nsh_sfc-dev@...; Dave Barach <dbarach@...>; Ed Warnicke (eaw) <eaw@...>
Subject: [nsh_sfc-dev] VPP-43 NSH removal progress

Got VXLAN-GPE separated from NSH

Just have to find a way to test GRE and it should be pretty straight forward to make generic GRE talk to new NSH node. Then I'll publish patch for review.

New commands:

vxlan-gpe tunnel:

vxlan-gpe tunnel local 192.168.150.90 remote 192.168.150.91 vni 16777215 next-nsh

(note next-nsh is not necessary... this is just there for classifier case... its not really needed at all)

nsh entry: (defines an NSH header)

nsh entry nsp 1 nsi 255 c1 9901 c2 9902 c3 9903 c4 9904 md-type 1

nsh map: maps two nsh headers together. Note that if you do "encap-none" it will attempt to decap it ... this is not currently supported and can be done in NSH_SFC

nsh map nsp 1 nsi 255 mapped-nsp 1 mapped-nsi 255 encap-vxlan-gpe-intf 6

Packet 6

00:00:14:216827: dpdk-input

GigabitEthernet0/5/0 rx queue 0

buffer 0x10d33: current data 0, length 106, free-list 0, totlen-nifb 0, trace 0x5

PKT MBUF: port 0, nb_segs 1, pkt_len 106

buf_len 2176, data_len 106, ol_flags 0x0,

packet_type 0x0

IP4: 52:54:00:24:19:a5 -> 52:54:00:27:0b:2f

UDP: 192.168.150.91 -> 192.168.150.90

tos 0x00, ttl 64, length 92, checksum 0x4d08

fragment id 0x3f82, flags DONT_FRAGMENT

UDP: 4790 -> 4790

length 72, checksum 0x9947

00:00:14:216861: ethernet-input

IP4: 52:54:00:24:19:a5 -> 52:54:00:27:0b:2f

00:00:14:216875: ip4-input

UDP: 192.168.150.91 -> 192.168.150.90

tos 0x00, ttl 64, length 92, checksum 0x4d08

fragment id 0x3f82, flags DONT_FRAGMENT

UDP: 4790 -> 4790

length 72, checksum 0x9947

00:00:14:216888: ip4-local

fib 0 adj-idx 4 : local 192.168.150.90/24 flow hash: 0x00000000

00:00:14:216898: ip4-udp-lookup

UDP: src-port 4790 dst-port 4790

00:00:14:216908: vxlan-gpe-input

VXLAN-GPE: tunnel 0 next 4 error 0

00:00:14:216917: nsh-input

nsh ver 0 len 0 (0 bytes) md_type 1 next_protocol 1

service path 1 service index 255

c1 9901 c2 9902 c3 9903 c4 9904

00:00:14:216925: vxlan-gpe-encap

VXLAN-GPE-ENCAP: tunnel 0

00:00:14:216932: ip4-rewrite-transit

fib 0 adj-idx 5 : GigabitEthernet0/5/0

IP4: 52:54:00:27:0b:2f -> 52:54:00:24:19:a5 flow hash: 0x00000000

IP4: 52:54:00:27:0b:2f -> 52:54:00:24:19:a5

UDP: 192.168.150.90 -> 192.168.150.91

tos 0x00, ttl 253, length 92, checksum 0x0f8a

fragment id 0x0000

UDP: 4790 -> 4790

length 72, checksum 0x0000

00:00:14:216944: GigabitEthernet0/5/0-output

GigabitEthernet0/5/0

IP4: 52:54:00:27:0b:2f -> 52:54:00:24:19:a5

UDP: 192.168.150.90 -> 192.168.150.91

tos 0x00, ttl 253, length 92, checksum 0x0f8a

fragment id 0x0000

UDP: 4790 -> 4790

length 72, checksum 0x0000

00:00:14:216952: GigabitEthernet0/5/0-tx

GigabitEthernet0/5/0 tx queue 0

buffer 0x10d33: current data 0, length 106, free-list 0, totlen-nifb 0, trace 0x5

IP4: 52:54:00:27:0b:2f -> 52:54:00:24:19:a5

UDP: 192.168.150.90 -> 192.168.150.91

tos 0x00, ttl 253, length 92, checksum 0x0f8a

fragment id 0x0000

UDP: 4790 -> 4790

length 72, checksum 0x0000

Keith Burns <alagalah@...>

#66

Hongjun,

Implementation detail we can sort out when we move to nshsfc repo but for now key is NSP NSI.

So you can't have the situation you describe.

Show quoted text

On Sun, May 8, 2016, 7:26 PM Ni, Hongjun <hongjun.ni@...> wrote:

Hi Keith,

Just a question about the nsh map:

For both nsh headers that have the same SPI and SI, are the context data of them the same?

If they are not, how can distinguish these two nsh headers?

Thanks,

Hongjun

From: nsh_sfc-dev-bounces@... [mailto:nsh_sfc-dev-bounces@...] On Behalf Of Keith Burns
Sent: Monday, May 9, 2016 6:37 AM
To: nsh_sfc-dev@...; Dave Barach <dbarach@...>; Ed Warnicke (eaw) <eaw@...>
Subject: [nsh_sfc-dev] VPP-43 NSH removal progress

Got VXLAN-GPE separated from NSH

Just have to find a way to test GRE and it should be pretty straight forward to make generic GRE talk to new NSH node. Then I'll publish patch for review.

New commands:

vxlan-gpe tunnel:

vxlan-gpe tunnel local 192.168.150.90 remote 192.168.150.91 vni 16777215 next-nsh

(note next-nsh is not necessary... this is just there for classifier case... its not really needed at all)

nsh entry: (defines an NSH header)

nsh entry nsp 1 nsi 255 c1 9901 c2 9902 c3 9903 c4 9904 md-type 1

nsh map: maps two nsh headers together. Note that if you do "encap-none" it will attempt to decap it ... this is not currently supported and can be done in NSH_SFC

nsh map nsp 1 nsi 255 mapped-nsp 1 mapped-nsi 255 encap-vxlan-gpe-intf 6

Packet 6

00:00:14:216827: dpdk-input

GigabitEthernet0/5/0 rx queue 0

buffer 0x10d33: current data 0, length 106, free-list 0, totlen-nifb 0, trace 0x5

PKT MBUF: port 0, nb_segs 1, pkt_len 106

buf_len 2176, data_len 106, ol_flags 0x0,

packet_type 0x0

IP4: 52:54:00:24:19:a5 -> 52:54:00:27:0b:2f

UDP: 192.168.150.91 -> 192.168.150.90

tos 0x00, ttl 64, length 92, checksum 0x4d08

fragment id 0x3f82, flags DONT_FRAGMENT

UDP: 4790 -> 4790

length 72, checksum 0x9947

00:00:14:216861: ethernet-input

IP4: 52:54:00:24:19:a5 -> 52:54:00:27:0b:2f

00:00:14:216875: ip4-input

UDP: 192.168.150.91 -> 192.168.150.90

tos 0x00, ttl 64, length 92, checksum 0x4d08

fragment id 0x3f82, flags DONT_FRAGMENT

UDP: 4790 -> 4790

length 72, checksum 0x9947

00:00:14:216888: ip4-local

fib 0 adj-idx 4 : local 192.168.150.90/24 flow hash: 0x00000000

00:00:14:216898: ip4-udp-lookup

UDP: src-port 4790 dst-port 4790

00:00:14:216908: vxlan-gpe-input

VXLAN-GPE: tunnel 0 next 4 error 0

00:00:14:216917: nsh-input

nsh ver 0 len 0 (0 bytes) md_type 1 next_protocol 1

service path 1 service index 255

c1 9901 c2 9902 c3 9903 c4 9904

00:00:14:216925: vxlan-gpe-encap

VXLAN-GPE-ENCAP: tunnel 0

00:00:14:216932: ip4-rewrite-transit

fib 0 adj-idx 5 : GigabitEthernet0/5/0

IP4: 52:54:00:27:0b:2f -> 52:54:00:24:19:a5 flow hash: 0x00000000

IP4: 52:54:00:27:0b:2f -> 52:54:00:24:19:a5

UDP: 192.168.150.90 -> 192.168.150.91

tos 0x00, ttl 253, length 92, checksum 0x0f8a

fragment id 0x0000

UDP: 4790 -> 4790

length 72, checksum 0x0000

00:00:14:216944: GigabitEthernet0/5/0-output

GigabitEthernet0/5/0

IP4: 52:54:00:27:0b:2f -> 52:54:00:24:19:a5

UDP: 192.168.150.90 -> 192.168.150.91

tos 0x00, ttl 253, length 92, checksum 0x0f8a

fragment id 0x0000

UDP: 4790 -> 4790

length 72, checksum 0x0000

00:00:14:216952: GigabitEthernet0/5/0-tx

GigabitEthernet0/5/0 tx queue 0

buffer 0x10d33: current data 0, length 106, free-list 0, totlen-nifb 0, trace 0x5

IP4: 52:54:00:27:0b:2f -> 52:54:00:24:19:a5

UDP: 192.168.150.90 -> 192.168.150.91

tos 0x00, ttl 253, length 92, checksum 0x0f8a

fragment id 0x0000

UDP: 4790 -> 4790

length 72, checksum 0x0000

Ni, Hongjun

#67

Hi Keith,

Got it.

Thanks,

Hongjun

Show quoted text

From: Keith Burns [mailto:alagalah@...]
Sent: Monday, May 9, 2016 10:33 AM
To: Ni, Hongjun <hongjun.ni@...>; nsh_sfc-dev@...; Dave Barach <dbarach@...>; Ed Warnicke (eaw) <eaw@...>
Subject: Re: [nsh_sfc-dev] VPP-43 NSH removal progress

Hongjun,

Implementation detail we can sort out when we move to nshsfc repo but for now key is NSP NSI.

So you can't have the situation you describe.

On Sun, May 8, 2016, 7:26 PM Ni, Hongjun <hongjun.ni@...> wrote:

Hi Keith,

Just a question about the nsh map:

For both nsh headers that have the same SPI and SI, are the context data of them the same?

If they are not, how can distinguish these two nsh headers?

Thanks,

Hongjun

From: nsh_sfc-dev-bounces@... [mailto:nsh_sfc-dev-bounces@...] On Behalf Of Keith Burns
Sent: Monday, May 9, 2016 6:37 AM
To: nsh_sfc-dev@...; Dave Barach <dbarach@...>; Ed Warnicke (eaw) <eaw@...>
Subject: [nsh_sfc-dev] VPP-43 NSH removal progress

Got VXLAN-GPE separated from NSH

Just have to find a way to test GRE and it should be pretty straight forward to make generic GRE talk to new NSH node. Then I'll publish patch for review.

New commands:

vxlan-gpe tunnel:

vxlan-gpe tunnel local 192.168.150.90 remote 192.168.150.91 vni 16777215 next-nsh

(note next-nsh is not necessary... this is just there for classifier case... its not really needed at all)

nsh entry: (defines an NSH header)

nsh entry nsp 1 nsi 255 c1 9901 c2 9902 c3 9903 c4 9904 md-type 1

nsh map: maps two nsh headers together. Note that if you do "encap-none" it will attempt to decap it ... this is not currently supported and can be done in NSH_SFC

nsh map nsp 1 nsi 255 mapped-nsp 1 mapped-nsi 255 encap-vxlan-gpe-intf 6

Packet 6

00:00:14:216827: dpdk-input

GigabitEthernet0/5/0 rx queue 0

buffer 0x10d33: current data 0, length 106, free-list 0, totlen-nifb 0, trace 0x5

PKT MBUF: port 0, nb_segs 1, pkt_len 106

buf_len 2176, data_len 106, ol_flags 0x0,

packet_type 0x0

IP4: 52:54:00:24:19:a5 -> 52:54:00:27:0b:2f

UDP: 192.168.150.91 -> 192.168.150.90

tos 0x00, ttl 64, length 92, checksum 0x4d08

fragment id 0x3f82, flags DONT_FRAGMENT

UDP: 4790 -> 4790

length 72, checksum 0x9947

00:00:14:216861: ethernet-input

IP4: 52:54:00:24:19:a5 -> 52:54:00:27:0b:2f

00:00:14:216875: ip4-input

UDP: 192.168.150.91 -> 192.168.150.90

tos 0x00, ttl 64, length 92, checksum 0x4d08

fragment id 0x3f82, flags DONT_FRAGMENT

UDP: 4790 -> 4790

length 72, checksum 0x9947

00:00:14:216888: ip4-local

fib 0 adj-idx 4 : local 192.168.150.90/24 flow hash: 0x00000000

00:00:14:216898: ip4-udp-lookup

UDP: src-port 4790 dst-port 4790

00:00:14:216908: vxlan-gpe-input

VXLAN-GPE: tunnel 0 next 4 error 0

00:00:14:216917: nsh-input

nsh ver 0 len 0 (0 bytes) md_type 1 next_protocol 1

service path 1 service index 255

c1 9901 c2 9902 c3 9903 c4 9904

00:00:14:216925: vxlan-gpe-encap

VXLAN-GPE-ENCAP: tunnel 0

00:00:14:216932: ip4-rewrite-transit

fib 0 adj-idx 5 : GigabitEthernet0/5/0

IP4: 52:54:00:27:0b:2f -> 52:54:00:24:19:a5 flow hash: 0x00000000

IP4: 52:54:00:27:0b:2f -> 52:54:00:24:19:a5

UDP: 192.168.150.90 -> 192.168.150.91

tos 0x00, ttl 253, length 92, checksum 0x0f8a

fragment id 0x0000

UDP: 4790 -> 4790

length 72, checksum 0x0000

00:00:14:216944: GigabitEthernet0/5/0-output

GigabitEthernet0/5/0

IP4: 52:54:00:27:0b:2f -> 52:54:00:24:19:a5

UDP: 192.168.150.90 -> 192.168.150.91

tos 0x00, ttl 253, length 92, checksum 0x0f8a

fragment id 0x0000

UDP: 4790 -> 4790

length 72, checksum 0x0000

00:00:14:216952: GigabitEthernet0/5/0-tx

GigabitEthernet0/5/0 tx queue 0

buffer 0x10d33: current data 0, length 106, free-list 0, totlen-nifb 0, trace 0x5

IP4: 52:54:00:27:0b:2f -> 52:54:00:24:19:a5

UDP: 192.168.150.90 -> 192.168.150.91

tos 0x00, ttl 253, length 92, checksum 0x0f8a

fragment id 0x0000

UDP: 4790 -> 4790

length 72, checksum 0x0000

1 - 4 of 4

Previous Topic Next Topic

Home Hashtags Subgroups Terms