aboutsummaryrefslogtreecommitdiffstats
path: root/RELEASE.md
blob: 80ad4227092799f582efe44c8d4cae5d6bcfce3c (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
# Release Notes    {#release_notes}

* @subpage release_notes_1707
* @subpage release_notes_1704
* @subpage release_notes_17011
* @subpage release_notes_1701
* @subpage release_notes_1609
* @subpage release_notes_1606

@page release_notes_1707 Release notes for VPP 17.07

More than 400 commits since the 1704 release.

## Features
- Infrastructure
  - make test; improved debuggability.
  - TAB auto-completion on the CLI
  - DPDK 17.05
  - python 3 support in test infra

- Host stack
  - Improved Linux TCP stack compatibility using IWL test suite (https://jira.fd.io/browse/VPP-720)
  - Improved loss recovery (RFC5681, RFC6582, RF6675)
  - Basic implementation of Eifel detection algorithm (RFC3522)
  - Basic support for buffer chains
  - Refactored session layer API
  - Overall performance, scale and hardening

- Interfaces
  - memif: IP mode, jumbo frames, multi queue
  - virtio-user support
  - vhost-usr; adaptive (poll/interupt) support.

- Network features
  - MPLS Multicast FIB

  - BFD FIB integration

  - NAT64 support

  - GRE over IPv6

  - Segement routing MPLS

  - IOAM configuration for SRv6 localsid

  - LISP
    - NSH support
    - native forward static routes
    - L2 ARP

  - ACL multi-core suuport

  - Flowprobe:
    - Add flowstartns, flowendns and tcpcontrolbits
    - Stateful flows and IPv6, L4 recording

  - GTP-U support

  - VXLAN GPE support for FIB2.0 and bypass.


## Known issues

For the full list of issues please reffer to fd.io [JIRA](https://jira.fd.io).

## Issues fixed

For the full list of fixed issues please reffer to:
- fd.io [JIRA](https://jira.fd.io)
- git [commit log](https://git.fd.io/vpp/log/?h=stable/1707)

@page release_notes_1704 Release notes for VPP 17.04

More than 500 commits since the 1701 release.

## Features
- Infrastructure
  - make test improvements
  - vnet: add device-input threadplacement infra
  - 64 bit per-thread counters
  - process restart cli
  - High performance timer wheels
  - Plugin infrastructure improvements
    - Support for .default_disabled, .version_required
  - Added MAINTAINERS file

- Host stack
  - TCP stack (experimental)
  - DHCPv4 / DHCPv6 relay multi-destination
  - DHCPv4 option 82
  - ND proxy
  - Attached hosts
  - Consolidated DHCPv4 and DHCPv6 implementation

- Interfaces
  - DPDK 17.02 (retire support for DPDK 16.07)
  - Add memif - packet memory interface for intra-host communication
  - vhost: support interrupt mode
  - DPDK as plugin (retired vpp_lite)
  - DPDPK input optimizations
  - Loopback interface allocation scheme

- Network features
  - IP Multicast FIB

  - Bridging
    - Learning on local interfaces
    - Flushing of MACs from the L2 FIB

  - SNAT
    - CGN (Deterministic and dynamic)
    - CGN configurable port allocation algorithm
    - ICMP support
    - Tentant VRF id for SNAT outside addresses
    - Session dump / User dump
    - Port allocation per protocol

  - Security groups
    - Routed interface support
    - L2+L3 unified processing node
    - Improve fragment handling

  - Segement routing v6
    - SR policies with weighted SID lists
    - Binding SID
    - SR steering policies
    - SR Local SIDs
    - Framework to expand local SIDs w/plugins
    - Documentation

  - IOAM
    - UDP Pinger w/path fault isolation
    - IOAM as type 2 metadata in NSH
    - IAOM raw IPFIX collector and analyzer
    - Anycast active server selection
    - Documentation
    - SRv6 Local SID
    - IP6 HBH header and SR header co-existence
    - Active probe

  - LISP
    - Statistics collection
    - Generalize encap for overlay transport (vxlan-gpe support)
    - Improve data plane speed

  - GPE
    - CLI
    - NSH added to encap/decap path
    - Renamed LISP GPE API to GPE

  - MPLS
    - Performance improvements (quad loop)

  - BFD
    - Command line interface
    - Echo function
    - Remote demand mode
    - SHA1 authentication

  - IPsec
    - IKEv2 initiator features

  - VXLAN
    - unify IP4/IP6 control plane handling

## API changes

- Python API: To avoid conflicts between VPP API messages names and
  the Python API binding function names, VPP API methods are put in a
  separate proxy object.
  https://gerrit.fd.io/r/#/c/5570/
  The api methods are now referenced as:
    vpp_handle = VPP(jsonfiles)
    vpp_handle.connect(...)
    vpp = vpp_handle.api
    vpp.show_version()
    vpp_handle.disconnect()

  For backwards compatibility VPP API methods are left in the main
  name space (VPP), but will be removed from 17.07.

  - Python API: Change from cPython to CFFI.

- create_loopback message to be replaced with create_loopback_instance
  create_loopback will be removed from 17.07.
  https://gerrit.fd.io/r/#/c/5572/

## Known issues

For the full list of issues please reffer to fd.io [JIRA](https://jira.fd.io).

## Issues fixed

For the full list of fixed issues please reffer to:
- fd.io [JIRA](https://jira.fd.io)
- git [commit log](https://git.fd.io/vpp/log/?h=stable/1704)

@page release_notes_17011 Release notes for VPP 17.01.1

This is bug fix release.

For the full list of fixed issues please reffer to:
- fd.io [JIRA](https://jira.fd.io)
- git [commit log](https://git.fd.io/vpp/log/?h=stable/1701)


@page release_notes_17011 Release notes for VPP 17.01.1

This is bug fix release.

For the full list of fixed issues please reffer to:
- fd.io [JIRA](https://jira.fd.io)
- git [commit log](https://git.fd.io/vpp/log/?h=stable/1701)

@page release_notes_1701 Release notes for VPP 17.01

@note This release was for a while known as 16.12.

## Features

- [Integrated November 2016 DPDK release](http://www.dpdk.org/doc/guides/rel_notes/release_16_11.html)

- Complete rework of Forwarding Information Base (FIB)

- Performance Improvements
  - Improvements in DPDK input and output nodes
  - Improvements in L2 path
  - Improvmeents in IPv4 lookup node

- Feature Arcs Improvements
  - Consolidation of the code
  - New feature arcs
    - device-input
    - interface-output

- DPDK Cryptodev Support
  - Software and Hardware Crypto Support

- DPDK HQoS support

- Simple Port Analyzer (SPAN)

- Bidirectional Forwarding Detection
  - Basic implementation

- IPFIX Improvements

- L2 GRE over IPSec tunnels

- Link Layer Discovery Protocol (LLDP)

- Vhost-user Improvements
  - Performance Improvements
  - Multiqueue
  - Reconnect

- LISP Enhancements
  - Source/Dest control plane support
  - L2 over LISP and GRE
  - Map-Register/Map-Notify/RLOC-probing support
  - L2 API improvements, overall code hardening

- Plugins:
  - New: ACL
  - New: Flow per Packet
  - Improved: SNAT
    - Mutlithreading
    - Flow export

- Doxygen Enhancements

- Luajit API bindings

- API Refactoring
  - file split
  - message signatures

- Python and Scapy based unit testing infrastructure
  - Infrastructure
  - Various tests

- Packet Generator improvements

- TUN/TAP jumbo frames support

- Other various bug fixes and improvements

## Known issues

For the full list of issues please reffer to fd.io [JIRA](https://jira.fd.io).

## Issues fixed

For the full list of fixed issues please reffer to:
- fd.io [JIRA](https://jira.fd.io)
- git [commit log](https://git.fd.io/vpp/log/?h=stable/1701)

@page release_notes_1609 Release notes for VPP 16.09

## Features

- [Integrated July 2016 DPDK release](http://www.dpdk.org/doc/guides/rel_notes/release_16_07.html)
  - DPDK-vhost is depreciated pending a complete rework of the original integration and
    addressing of rx performance deltas.
  - Patches required for DPDK 16.07:
    - Correctly setting the Packet Type in the IGB, IXGBE and i40e drivers.
    - Correctly setting checksum in the i40e driver.
    - NXP DPAA2 PMD Driver.
    - rte_delay (yield) functionality.

- Add “in tree” plugins:
  - IPv6 ILA.
  - iOAM.
  - Load Balancer.
  - SNAT.

- High-performance (line-rate) “neutron like” L4 port-filtering.

- API refactoring - addressing some of the issues around JVPP bindings.
  - Accommodating plugins [(e.g. NSH_SFC)](https://wiki.fd.io/view/NSH_SFC)
  - Binding for [python](https://wiki.fd.io/view/VPP/Python_API)

- LISP
  - L2 LISP overlays
  -  Multitenancy
  - Multihoming
  - RTR mode
  - Map-resolver failover algorithm

- Support 64-bit vector lengths, huge shared-memory segments.

- Dynamic IP Feature ordering
  - IP Features can now specify features they appear before and after

- 16.09 Builds
  - Ubuntu 14.04 LTS - Trusty Tahr
  - Ubuntu 16.04 LTS - Xenial Xerus
  - CentOS 7
  - More information on [VPP wiki](https://wiki.fd.io/view/VPP/Installing_VPP_binaries_from_packages)

- Performance, characterize and document performance for this release
  [(more information on CSIT page)](https://wiki.fd.io/view/CSIT)

   - IPv4 and IPv6 Scale - performance tests.
     - Bidirectional 10k/100k/1M flows.
     - 64B,570B, 1518B,9000B packet sizes.
   - IPv6 iACL - performance
     - DUT1 and DUT2 are configured with IPv6 routing, two static IPv6 /64 routes and IPv6 iAcl
       security whitelist ingress /64 filter entries applied on links.
     - TG traffic profile contains two L3 flow-groups (flow-group per direction, 253 flows per
       flow-group) with all packets containing Ethernet header, IPv6 header and generated payload.
       MAC addresses are matching MAC addresses of the TG node interfaces.

   - L2XC VXLANoIPv4 - performance
     - DUT1 and DUT2 are configured with L2 cross-connect. VXLAN tunnels are configured between
       L2XCs on DUT1 and DUT2.
     - TG traffic profile contains two L3 flow-groups (flow-group per direction, 253 flows per
       flow-group) with all packets containing Ethernet header, IPv4 header with IP protocol=61
       and generated payload. MAC addresses are matching MAC addresses of the TG node interfaces.

- Documentation
  - Autogenerated CLI documentation.
  - Using doxygen to automate API/Node documentation.
  - [(available online)](https://docs.fd.io/vpp/16.09/)

- Resolved all static analysis issues found by Coverity
  - Beginning of 16.09 cycle: 505 issues.
  - Release: 0 outstanding issues.


## Known issues

Issues in fd.io are tracked in [JIRA](https://jira.fd.io).

Issue | Description
--- | ---
VPP-391 |   vpp debug version assert appeared in the process of start
VPP-380 |   Mapping algorithm compute wrong ea-bits when IPv4 prefix 0.0.0.0/0
VPP-371 |   load_one_plugin:63: Loaded plugin: message from vppctl
VPP-367 |   vpp packages need to depend on specific versions of each other
VPP-312 |   IP6 FIB gets in indeterminate state by duplicating commands
VPP-224 |   Lookup-in-vrf can not be set correctly
VPP-206 |   Fix classify table delete
VPP-203 |   Fix binary API for reading vpp node graph
VPP-147 |   Inconsistent behaviour when adding L2 FIB filter entry
VPP-99  |  VPP doesn't discard DHCPOFFER message with wrong XID


## Issues fixed

Issues in fd.io are tracked in [JIRA](https://jira.fd.io).

Issue | Description
--- | ---
VPP-396 |   Ubuntu systems Graphviz bug
VPP-390 |   vpp-lib rpm fails to include *.so symlinks, causing linking problems with out of tree builds
VPP-388 |   IPSec output feature assumes packets have been ethernet rewritten
VPP-385 |   ARP for indirect adjacencies not working correctly
VPP-361 |   Memory leak on delete of VXLAN over IPv6 tunnel
VPP-357 |   VNI not set correctly when removing LISP fwd entries
VPP-349 |   sw_interface_vhost_user_dump not working
VPP-345 |   net/enic: bad L4 checksum ptype set on ICMP packets
VPP-340 |   MAP-T wrong destination address
VPP-330 |   Use fifo to store LISP pending map-requests
VPP-326 |   map_add_domain VAT command: unable to configure domain with mtu parameter
VPP-318 |   The map_add_domain VAT command accepts invalid arguments
VPP-315 |   Fix "show vxlan-gpe" issue
VPP-310 |   Mapping algorithm compute wrong ea-bits
VPP-239 |   LISP IP forwarding does not tag packets that hit negative mapping entries
VPP-235 |   Invalid help in VAT for sw_interface_set_l2_bridge
VPP-228 |   Mapping algorithm sends packet to wrong IPv6 address
VPP-214 |   vpp-api-test: api_ipsec_sad_add_del_entry: vector "ck" not initialized
VPP-200 |   VPP - TAP port create problem
VPP-189 |   Coverity Issues for 16.09
VPP-184 |   u16 translating to char ,not short
VPP-179 |   Adjacency share-count botch
VPP-163 |   "show ip6 interface" ignores non-global addresses
VPP-155 |   Netmap: Inconsistency in interface state between "show hardware" and "show interface"
VPP-145 |   Dynamically compute IP feature ordering based on constraints
VPP-137 |   VPP sends ARP with wrong requested IP
VPP-118 |   JVpp: 0 length arrays not handled properly in VPP responses
VPP-112 |   linux kernel info missing from build log
VPP-110 |   vxlan encap node should never touch a deleted tunnel
VPP-107 |   RPM build broken in master
VPP-92  |   segment routing is not properly filling out the segment list
VPP-91  |   segment routing add/del tunnel lookup doesn't work
VPP-84  |   af_packet throws a fatal error on EAGAIN
VPP-74  |   Clang compile fails due to warning in vlib/unix/cli.c
VPP-64  |   Top level "make pkg-deb" fails if CDPATH is set in user env.
VPP-48  |   Traceroute does not terminate when VPP is the target
VPP-23  |   CLI pager does not gracefully handle lines longer than the terminal width


@page release_notes_1606 Release notes for VPP 16.06


The FD.io Project, relentlessly focused on data IO speed and efficiency
supporting the creation of high performance, flexible, and scalable software
defined infrastructures, announces the availability of the community’s first
software release (16.06).

In the four months since launching, FD.io has brought together more than 75
developers from 11 different companies including network operators, solution
providers chip vendors, and network equipment vendors who are collaborating to
enhance and innovate around the Vector Packet Processing (VPP) technology. The
FD.io community has quickly formed to grow the number of projects from the
initial VPP project to an additional 6 projects addressing a diverse set of
requirements and usability across a variety of deployment environments.

The 16.06 release brings unprecedented performance: 480Gbps/200mpps with 8
million routes and 2k whitelist entries on standard high volume x86 servers.


## Features

In addition to the existing full suite of vswitch/vrouter features, the new
16.06 release adds:

* Enhanced Switching and Routing:
  * IPv6 Segment Routing multicast support.
  * LISP xTR support.
  * VXLAN over IPv6 underlay.
  * Per interface whitelists.
  * Shared adjacencies in FIB.

* New and improved interface support:
  * Jumbo frame support for vhost-user.
  * Netmap interface support.
  * AF_Packet interface support.

* Expanded and improved programmability:
  * Python API bindings.
  * Enhanced JVPP Java API bindings.
  * Debugging CLI.

* Expanded Hardware and Software Support:
  * Support for ARM 32 targets including Rasberry Pi single-board computer.
  * Support for DPDK 16.04.
lor: #dd7700 } /* Name.Variable.Global */ .highlight .vi { color: #3333bb } /* Name.Variable.Instance */ .highlight .vm { color: #336699 } /* Name.Variable.Magic */ .highlight .il { color: #0000DD; font-weight: bold } /* Literal.Number.Integer.Long */
/*
 * node.c - ipfix probe graph node
 *
 * Copyright (c) 2017 Cisco and/or its affiliates.
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at:
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */
#include <vlib/vlib.h>
#include <vnet/vnet.h>
#include <vppinfra/crc32.h>
#include <vppinfra/xxhash.h>
#include <vppinfra/error.h>
#include <flowprobe/flowprobe.h>
#include <vnet/ip/ip6_packet.h>
#include <vnet/udp/udp_local.h>
#include <vlibmemory/api.h>

static void flowprobe_export_entry (vlib_main_t * vm, flowprobe_entry_t * e);

/**
 * @file node.c
 * flow record generator graph node
 */

typedef struct
{
  /** interface handle */
  u32 rx_sw_if_index;
  u32 tx_sw_if_index;
  /** packet timestamp */
  u64 timestamp;
  /** size of the buffer */
  u16 buffer_size;

  /** L2 information */
  u8 src_mac[6];
  u8 dst_mac[6];
  /** Ethertype */
  u16 ethertype;

  /** L3 information */
  ip46_address_t src_address;
  ip46_address_t dst_address;
  u8 protocol;
  u8 tos;

  /** L4 information */
  u16 src_port;
  u16 dst_port;

  flowprobe_variant_t which;
} flowprobe_trace_t;

static char *flowprobe_variant_strings[] = {
  [FLOW_VARIANT_IP4] = "IP4",
  [FLOW_VARIANT_IP6] = "IP6",
  [FLOW_VARIANT_L2] = "L2",
  [FLOW_VARIANT_L2_IP4] = "L2-IP4",
  [FLOW_VARIANT_L2_IP6] = "L2-IP6",
};

/* packet trace format function */
static u8 *
format_flowprobe_trace (u8 * s, va_list * args)
{
  CLIB_UNUSED (vlib_main_t * vm) = va_arg (*args, vlib_main_t *);
  CLIB_UNUSED (vlib_node_t * node) = va_arg (*args, vlib_node_t *);
  flowprobe_trace_t *t = va_arg (*args, flowprobe_trace_t *);
  u32 indent = format_get_indent (s);

  s = format (s,
	      "FLOWPROBE[%s]: rx_sw_if_index %d, tx_sw_if_index %d, "
	      "timestamp %lld, size %d", flowprobe_variant_strings[t->which],
	      t->rx_sw_if_index, t->tx_sw_if_index,
	      t->timestamp, t->buffer_size);

  if (t->which == FLOW_VARIANT_L2)
    s = format (s, "\n%U -> %U", format_white_space, indent,
		format_ethernet_address, &t->src_mac,
		format_ethernet_address, &t->dst_mac);

  if (t->protocol > 0
      && (t->which == FLOW_VARIANT_L2_IP4 || t->which == FLOW_VARIANT_IP4
	  || t->which == FLOW_VARIANT_L2_IP6 || t->which == FLOW_VARIANT_IP6))
    s =
      format (s, "\n%U%U: %U -> %U", format_white_space, indent,
	      format_ip_protocol, t->protocol, format_ip46_address,
	      &t->src_address, IP46_TYPE_ANY, format_ip46_address,
	      &t->dst_address, IP46_TYPE_ANY);
  return s;
}

vlib_node_registration_t flowprobe_input_ip4_node;
vlib_node_registration_t flowprobe_input_ip6_node;
vlib_node_registration_t flowprobe_input_l2_node;
vlib_node_registration_t flowprobe_output_ip4_node;
vlib_node_registration_t flowprobe_output_ip6_node;
vlib_node_registration_t flowprobe_output_l2_node;

/* No counters at the moment */
#define foreach_flowprobe_error			\
_(COLLISION, "Hash table collisions")		\
_(BUFFER, "Buffer allocation error")		\
_(EXPORTED_PACKETS, "Exported packets")		\
_(INPATH, "Exported packets in path")

typedef enum
{
#define _(sym,str) FLOWPROBE_ERROR_##sym,
  foreach_flowprobe_error
#undef _
    FLOWPROBE_N_ERROR,
} flowprobe_error_t;

static char *flowprobe_error_strings[] = {
#define _(sym,string) string,
  foreach_flowprobe_error
#undef _
};

typedef enum
{
  FLOWPROBE_NEXT_DROP,
  FLOWPROBE_NEXT_IP4_LOOKUP,
  FLOWPROBE_N_NEXT,
} flowprobe_next_t;

#define FLOWPROBE_NEXT_NODES {					\
    [FLOWPROBE_NEXT_DROP] = "error-drop",			\
    [FLOWPROBE_NEXT_IP4_LOOKUP] = "ip4-lookup",		\
}

static inline flowprobe_variant_t
flowprobe_get_variant (flowprobe_variant_t which,
		       flowprobe_record_t flags, u16 ethertype)
{
  if (which == FLOW_VARIANT_L2
      && (flags & FLOW_RECORD_L3 || flags & FLOW_RECORD_L4))
    return ethertype == ETHERNET_TYPE_IP6 ? FLOW_VARIANT_L2_IP6 : ethertype ==
      ETHERNET_TYPE_IP4 ? FLOW_VARIANT_L2_IP4 : FLOW_VARIANT_L2;
  return which;
}

/*
 * NTP rfc868 : 2 208 988 800 corresponds to 00:00  1 Jan 1970 GMT
 */
#define NTP_TIMESTAMP 2208988800LU

static inline u32
flowprobe_common_add (vlib_buffer_t * to_b, flowprobe_entry_t * e, u16 offset)
{
  u16 start = offset;

  /* Ingress interface */
  u32 rx_if = clib_host_to_net_u32 (e->key.rx_sw_if_index);
  clib_memcpy_fast (to_b->data + offset, &rx_if, sizeof (rx_if));
  offset += sizeof (rx_if);

  /* Egress interface */
  u32 tx_if = clib_host_to_net_u32 (e->key.tx_sw_if_index);
  clib_memcpy_fast (to_b->data + offset, &tx_if, sizeof (tx_if));
  offset += sizeof (tx_if);

  /* Flow direction
     0x00: ingress flow
     0x01: egress flow */
  to_b->data[offset++] = (e->key.direction == FLOW_DIRECTION_TX);

  /* packet delta count */
  u64 packetdelta = clib_host_to_net_u64 (e->packetcount);
  clib_memcpy_fast (to_b->data + offset, &packetdelta, sizeof (u64));
  offset += sizeof (u64);

  /* flowStartNanoseconds */
  u32 t = clib_host_to_net_u32 (e->flow_start.sec + NTP_TIMESTAMP);
  clib_memcpy_fast (to_b->data + offset, &t, sizeof (u32));
  offset += sizeof (u32);
  t = clib_host_to_net_u32 (e->flow_start.nsec);
  clib_memcpy_fast (to_b->data + offset, &t, sizeof (u32));
  offset += sizeof (u32);

  /* flowEndNanoseconds */
  t = clib_host_to_net_u32 (e->flow_end.sec + NTP_TIMESTAMP);
  clib_memcpy_fast (to_b->data + offset, &t, sizeof (u32));
  offset += sizeof (u32);
  t = clib_host_to_net_u32 (e->flow_end.nsec);
  clib_memcpy_fast (to_b->data + offset, &t, sizeof (u32));
  offset += sizeof (u32);

  return offset - start;
}

static inline u32
flowprobe_l2_add (vlib_buffer_t * to_b, flowprobe_entry_t * e, u16 offset)
{
  u16 start = offset;

  /* src mac address */
  clib_memcpy_fast (to_b->data + offset, &e->key.src_mac, 6);
  offset += 6;

  /* dst mac address */
  clib_memcpy_fast (to_b->data + offset, &e->key.dst_mac, 6);
  offset += 6;

  /* ethertype */
  clib_memcpy_fast (to_b->data + offset, &e->key.ethertype, 2);
  offset += 2;

  return offset - start;
}

static inline u32
flowprobe_l3_ip6_add (vlib_buffer_t * to_b, flowprobe_entry_t * e, u16 offset)
{
  u16 start = offset;

  /* ip6 src address */
  clib_memcpy_fast (to_b->data + offset, &e->key.src_address,
		    sizeof (ip6_address_t));
  offset += sizeof (ip6_address_t);

  /* ip6 dst address */
  clib_memcpy_fast (to_b->data + offset, &e->key.dst_address,
		    sizeof (ip6_address_t));
  offset += sizeof (ip6_address_t);

  /* Protocol */
  to_b->data[offset++] = e->key.protocol;

  /* octetDeltaCount */
  u64 octetdelta = clib_host_to_net_u64 (e->octetcount);
  clib_memcpy_fast (to_b->data + offset, &octetdelta, sizeof (u64));
  offset += sizeof (u64);

  return offset - start;
}

static inline u32
flowprobe_l3_ip4_add (vlib_buffer_t * to_b, flowprobe_entry_t * e, u16 offset)
{
  u16 start = offset;

  /* ip4 src address */
  clib_memcpy_fast (to_b->data + offset, &e->key.src_address.ip4,
		    sizeof (ip4_address_t));
  offset += sizeof (ip4_address_t);

  /* ip4 dst address */
  clib_memcpy_fast (to_b->data + offset, &e->key.dst_address.ip4,
		    sizeof (ip4_address_t));
  offset += sizeof (ip4_address_t);

  /* Protocol */
  to_b->data[offset++] = e->key.protocol;

  /* octetDeltaCount */
  u64 octetdelta = clib_host_to_net_u64 (e->octetcount);
  clib_memcpy_fast (to_b->data + offset, &octetdelta, sizeof (u64));
  offset += sizeof (u64);

  return offset - start;
}

static inline u32
flowprobe_l4_add (vlib_buffer_t * to_b, flowprobe_entry_t * e, u16 offset)
{
  u16 start = offset;

  /* src port */
  clib_memcpy_fast (to_b->data + offset, &e->key.src_port, 2);
  offset += 2;

  /* dst port */
  clib_memcpy_fast (to_b->data + offset, &e->key.dst_port, 2);
  offset += 2;

  /* tcp control bits */
  u16 control_bits = htons (e->prot.tcp.flags);
  clib_memcpy_fast (to_b->data + offset, &control_bits, 2);
  offset += 2;

  return offset - start;
}

static inline u32
flowprobe_hash (flowprobe_key_t * k)
{
  flowprobe_main_t *fm = &flowprobe_main;
  u32 h = 0;

#ifdef clib_crc32c_uses_intrinsics
  h = clib_crc32c ((u8 *) k, sizeof (*k));
#else
  int i;
  u64 tmp = 0;
  for (i = 0; i < sizeof (*k) / 8; i++)
    tmp ^= ((u64 *) k)[i];

  h = clib_xxhash (tmp);
#endif

  return h >> (32 - fm->ht_log2len);
}

flowprobe_entry_t *
flowprobe_lookup (u32 my_cpu_number, flowprobe_key_t * k, u32 * poolindex,
		  bool * collision)
{
  flowprobe_main_t *fm = &flowprobe_main;
  flowprobe_entry_t *e;
  u32 h;

  h = (fm->active_timer) ? flowprobe_hash (k) : 0;

  /* Lookup in the flow state pool */
  *poolindex = fm->hash_per_worker[my_cpu_number][h];
  if (*poolindex != ~0)
    {
      e = pool_elt_at_index (fm->pool_per_worker[my_cpu_number], *poolindex);
      if (e)
	{
	  /* Verify key or report collision */
	  if (memcmp (k, &e->key, sizeof (flowprobe_key_t)))
	    *collision = true;
	  return e;
	}
    }

  return 0;
}

flowprobe_entry_t *
flowprobe_create (u32 my_cpu_number, flowprobe_key_t * k, u32 * poolindex)
{
  flowprobe_main_t *fm = &flowprobe_main;
  u32 h;

  flowprobe_entry_t *e;

  /* Get my index */
  h = (fm->active_timer) ? flowprobe_hash (k) : 0;

  pool_get (fm->pool_per_worker[my_cpu_number], e);
  *poolindex = e - fm->pool_per_worker[my_cpu_number];
  fm->hash_per_worker[my_cpu_number][h] = *poolindex;

  e->key = *k;

  if (fm->passive_timer > 0)
    {
      e->passive_timer_handle = tw_timer_start_2t_1w_2048sl
	(fm->timers_per_worker[my_cpu_number], *poolindex, 0,
	 fm->passive_timer);
    }
  return e;
}

static inline void
add_to_flow_record_state (vlib_main_t *vm, vlib_node_runtime_t *node,
			  flowprobe_main_t *fm, vlib_buffer_t *b,
			  timestamp_nsec_t timestamp, u16 length,
			  flowprobe_variant_t which,
			  flowprobe_direction_t direction,
			  flowprobe_trace_t *t)
{
  if (fm->disabled)
    return;

  ASSERT (direction == FLOW_DIRECTION_RX || direction == FLOW_DIRECTION_TX);

  u32 my_cpu_number = vm->thread_index;
  u16 octets = 0;

  flowprobe_record_t flags = fm->context[which].flags;
  bool collect_ip4 = false, collect_ip6 = false;
  ASSERT (b);
  ethernet_header_t *eth = ethernet_buffer_get_header (b);
  u16 ethertype = clib_net_to_host_u16 (eth->type);
  u16 l2_hdr_sz = sizeof (ethernet_header_t);
  /* *INDENT-OFF* */
  flowprobe_key_t k = {};
  /* *INDENT-ON* */
  ip4_header_t *ip4 = 0;
  ip6_header_t *ip6 = 0;
  udp_header_t *udp = 0;
  tcp_header_t *tcp = 0;
  u8 tcp_flags = 0;

  if (flags & FLOW_RECORD_L3 || flags & FLOW_RECORD_L4)
    {
      collect_ip4 = which == FLOW_VARIANT_L2_IP4 || which == FLOW_VARIANT_IP4;
      collect_ip6 = which == FLOW_VARIANT_L2_IP6 || which == FLOW_VARIANT_IP6;
    }

  k.rx_sw_if_index = vnet_buffer (b)->sw_if_index[VLIB_RX];
  k.tx_sw_if_index = vnet_buffer (b)->sw_if_index[VLIB_TX];

  k.which = which;
  k.direction = direction;

  if (flags & FLOW_RECORD_L2)
    {
      clib_memcpy_fast (k.src_mac, eth->src_address, 6);
      clib_memcpy_fast (k.dst_mac, eth->dst_address, 6);
      k.ethertype = ethertype;
    }
  if (ethertype == ETHERNET_TYPE_VLAN)
    {
      /*VLAN TAG*/
      ethernet_vlan_header_tv_t *ethv =
	(ethernet_vlan_header_tv_t *) (&(eth->type));
      /*Q in Q possibility */
      while (clib_net_to_host_u16 (ethv->type) == ETHERNET_TYPE_VLAN)
	{
	  ethv++;
	  l2_hdr_sz += sizeof (ethernet_vlan_header_tv_t);
	}
      k.ethertype = ethertype = clib_net_to_host_u16 ((ethv)->type);
    }
  if (collect_ip6 && ethertype == ETHERNET_TYPE_IP6)
    {
      ip6 = (ip6_header_t *) (b->data + l2_hdr_sz);
      if (flags & FLOW_RECORD_L3)
	{
	  k.src_address.as_u64[0] = ip6->src_address.as_u64[0];
	  k.src_address.as_u64[1] = ip6->src_address.as_u64[1];
	  k.dst_address.as_u64[0] = ip6->dst_address.as_u64[0];
	  k.dst_address.as_u64[1] = ip6->dst_address.as_u64[1];
	}
      k.protocol = ip6->protocol;
      if (k.protocol == IP_PROTOCOL_UDP)
	udp = (udp_header_t *) (ip6 + 1);
      else if (k.protocol == IP_PROTOCOL_TCP)
	tcp = (tcp_header_t *) (ip6 + 1);

      octets = clib_net_to_host_u16 (ip6->payload_length)
	+ sizeof (ip6_header_t);
    }
  if (collect_ip4 && ethertype == ETHERNET_TYPE_IP4)
    {
      ip4 = (ip4_header_t *) (b->data + l2_hdr_sz);
      if (flags & FLOW_RECORD_L3)
	{
	  k.src_address.ip4.as_u32 = ip4->src_address.as_u32;
	  k.dst_address.ip4.as_u32 = ip4->dst_address.as_u32;
	}
      k.protocol = ip4->protocol;
      if ((flags & FLOW_RECORD_L4) && k.protocol == IP_PROTOCOL_UDP)
	udp = (udp_header_t *) (ip4 + 1);
      else if ((flags & FLOW_RECORD_L4) && k.protocol == IP_PROTOCOL_TCP)
	tcp = (tcp_header_t *) (ip4 + 1);

      octets = clib_net_to_host_u16 (ip4->length);
    }

  if (udp)
    {
      k.src_port = udp->src_port;
      k.dst_port = udp->dst_port;
    }
  else if (tcp)
    {
      k.src_port = tcp->src_port;
      k.dst_port = tcp->dst_port;
      tcp_flags = tcp->flags;
    }

  if (t)
    {
      t->rx_sw_if_index = k.rx_sw_if_index;
      t->tx_sw_if_index = k.tx_sw_if_index;
      clib_memcpy_fast (t->src_mac, k.src_mac, 6);
      clib_memcpy_fast (t->dst_mac, k.dst_mac, 6);
      t->ethertype = k.ethertype;
      t->src_address.ip4.as_u32 = k.src_address.ip4.as_u32;
      t->dst_address.ip4.as_u32 = k.dst_address.ip4.as_u32;
      t->protocol = k.protocol;
      t->src_port = k.src_port;
      t->dst_port = k.dst_port;
      t->which = k.which;
    }

  flowprobe_entry_t *e = 0;
  f64 now = vlib_time_now (vm);
  if (fm->active_timer > 0)
    {
      u32 poolindex = ~0;
      bool collision = false;

      e = flowprobe_lookup (my_cpu_number, &k, &poolindex, &collision);
      if (collision)
	{
	  /* Flush data and clean up entry for reuse. */
	  if (e->packetcount)
	    flowprobe_export_entry (vm, e);
	  e->key = k;
	  e->flow_start = timestamp;
	  vlib_node_increment_counter (vm, node->node_index,
				       FLOWPROBE_ERROR_COLLISION, 1);
	}
      if (!e)			/* Create new entry */
	{
	  e = flowprobe_create (my_cpu_number, &k, &poolindex);
	  e->last_exported = now;
	  e->flow_start = timestamp;
	}
    }
  else
    {
      e = &fm->stateless_entry[my_cpu_number];
      e->key = k;
    }

  if (e)
    {
      /* Updating entry */
      e->packetcount++;
      e->octetcount += octets;
      e->last_updated = now;
      e->flow_end = timestamp;
      e->prot.tcp.flags |= tcp_flags;
      if (fm->active_timer == 0
	  || (now > e->last_exported + fm->active_timer))
	flowprobe_export_entry (vm, e);
    }
}

static u16
flowprobe_get_headersize (void)
{
  return sizeof (ip4_header_t) + sizeof (udp_header_t) +
    sizeof (ipfix_message_header_t) + sizeof (ipfix_set_header_t);
}

static void
flowprobe_export_send (vlib_main_t * vm, vlib_buffer_t * b0,
		       flowprobe_variant_t which)
{
  flowprobe_main_t *fm = &flowprobe_main;
  flow_report_main_t *frm = &flow_report_main;
  ipfix_exporter_t *exp = pool_elt_at_index (frm->exporters, 0);
  vlib_frame_t *f;
  ip4_ipfix_template_packet_t *tp;
  ipfix_set_header_t *s;
  ipfix_message_header_t *h;
  ip4_header_t *ip;
  udp_header_t *udp;
  flowprobe_record_t flags = fm->context[which].flags;
  u32 my_cpu_number = vm->thread_index;

  /* Fill in header */
  flow_report_stream_t *stream;

  /* Nothing to send */
  if (fm->context[which].next_record_offset_per_worker[my_cpu_number] <=
      flowprobe_get_headersize ())
    return;

  u32 i, index = vec_len (exp->streams);
  for (i = 0; i < index; i++)
    if (exp->streams[i].domain_id == 1)
      {
	index = i;
	break;
      }
  if (i == vec_len (exp->streams))
    {
      vec_validate (exp->streams, index);
      exp->streams[index].domain_id = 1;
    }
  stream = &exp->streams[index];

  tp = vlib_buffer_get_current (b0);
  ip = (ip4_header_t *) & tp->ip4;
  udp = (udp_header_t *) (ip + 1);
  h = (ipfix_message_header_t *) (udp + 1);
  s = (ipfix_set_header_t *) (h + 1);

  ip->ip_version_and_header_length = 0x45;
  ip->ttl = 254;
  ip->protocol = IP_PROTOCOL_UDP;
  ip->flags_and_fragment_offset = 0;
  ip->src_address.as_u32 = exp->src_address.ip.ip4.as_u32;
  ip->dst_address.as_u32 = exp->ipfix_collector.ip.ip4.as_u32;
  udp->src_port = clib_host_to_net_u16 (stream->src_port);
  udp->dst_port = clib_host_to_net_u16 (exp->collector_port);
  udp->checksum = 0;

  /* FIXUP: message header export_time */
  h->export_time = (u32)
    (((f64) frm->unix_time_0) +
     (vlib_time_now (frm->vlib_main) - frm->vlib_time_0));
  h->export_time = clib_host_to_net_u32 (h->export_time);
  h->domain_id = clib_host_to_net_u32 (stream->domain_id);

  /* FIXUP: message header sequence_number */
  h->sequence_number = stream->sequence_number++;
  h->sequence_number = clib_host_to_net_u32 (h->sequence_number);

  s->set_id_length = ipfix_set_id_length (fm->template_reports[flags],
					  b0->current_length -
					  (sizeof (*ip) + sizeof (*udp) +
					   sizeof (*h)));
  h->version_length = version_length (b0->current_length -
				      (sizeof (*ip) + sizeof (*udp)));

  ip->length = clib_host_to_net_u16 (b0->current_length);

  ip->checksum = ip4_header_checksum (ip);
  udp->length = clib_host_to_net_u16 (b0->current_length - sizeof (*ip));

  if (exp->udp_checksum)
    {
      /* RFC 7011 section 10.3.2. */
      udp->checksum = ip4_tcp_udp_compute_checksum (vm, b0, ip);
      if (udp->checksum == 0)
	udp->checksum = 0xffff;
    }

  ASSERT (ip4_header_checksum_is_valid (ip));

  /* Find or allocate a frame */
  f = fm->context[which].frames_per_worker[my_cpu_number];
  if (PREDICT_FALSE (f == 0))
    {
      u32 *to_next;
      f = vlib_get_frame_to_node (vm, ip4_lookup_node.index);
      fm->context[which].frames_per_worker[my_cpu_number] = f;
      u32 bi0 = vlib_get_buffer_index (vm, b0);

      /* Enqueue the buffer */
      to_next = vlib_frame_vector_args (f);
      to_next[0] = bi0;
      f->n_vectors = 1;
    }

  vlib_put_frame_to_node (vm, ip4_lookup_node.index, f);
  vlib_node_increment_counter (vm, flowprobe_output_l2_node.index,
			       FLOWPROBE_ERROR_EXPORTED_PACKETS, 1);

  fm->context[which].frames_per_worker[my_cpu_number] = 0;
  fm->context[which].buffers_per_worker[my_cpu_number] = 0;
  fm->context[which].next_record_offset_per_worker[my_cpu_number] =
    flowprobe_get_headersize ();
}

static vlib_buffer_t *
flowprobe_get_buffer (vlib_main_t * vm, flowprobe_variant_t which)
{
  flowprobe_main_t *fm = &flowprobe_main;
  ipfix_exporter_t *exp = pool_elt_at_index (flow_report_main.exporters, 0);
  vlib_buffer_t *b0;
  u32 bi0;
  u32 my_cpu_number = vm->thread_index;

  /* Find or allocate a buffer */
  b0 = fm->context[which].buffers_per_worker[my_cpu_number];

  /* Need to allocate a buffer? */
  if (PREDICT_FALSE (b0 == 0))
    {
      if (vlib_buffer_alloc (vm, &bi0, 1) != 1)
	{
	  vlib_node_increment_counter (vm, flowprobe_output_l2_node.index,
				       FLOWPROBE_ERROR_BUFFER, 1);
	  return 0;
	}

      /* Initialize the buffer */
      b0 = fm->context[which].buffers_per_worker[my_cpu_number] =
	vlib_get_buffer (vm, bi0);

      b0->current_data = 0;
      b0->current_length = flowprobe_get_headersize ();
      b0->flags |=
	(VLIB_BUFFER_TOTAL_LENGTH_VALID | VNET_BUFFER_F_FLOW_REPORT);
      vnet_buffer (b0)->sw_if_index[VLIB_RX] = 0;
      vnet_buffer (b0)->sw_if_index[VLIB_TX] = exp->fib_index;
      fm->context[which].next_record_offset_per_worker[my_cpu_number] =
	b0->current_length;
    }

  return b0;
}

static void
flowprobe_export_entry (vlib_main_t * vm, flowprobe_entry_t * e)
{
  u32 my_cpu_number = vm->thread_index;
  flowprobe_main_t *fm = &flowprobe_main;
  ipfix_exporter_t *exp = pool_elt_at_index (flow_report_main.exporters, 0);
  vlib_buffer_t *b0;
  bool collect_ip4 = false, collect_ip6 = false;
  flowprobe_variant_t which = e->key.which;
  flowprobe_record_t flags = fm->context[which].flags;
  u16 offset =
    fm->context[which].next_record_offset_per_worker[my_cpu_number];

  if (offset < flowprobe_get_headersize ())
    offset = flowprobe_get_headersize ();

  b0 = flowprobe_get_buffer (vm, which);
  /* No available buffer, what to do... */
  if (b0 == 0)
    return;

  if (flags & FLOW_RECORD_L3)
    {
      collect_ip4 = which == FLOW_VARIANT_L2_IP4 || which == FLOW_VARIANT_IP4;
      collect_ip6 = which == FLOW_VARIANT_L2_IP6 || which == FLOW_VARIANT_IP6;
    }

  offset += flowprobe_common_add (b0, e, offset);

  if (flags & FLOW_RECORD_L2)
    offset += flowprobe_l2_add (b0, e, offset);
  if (collect_ip6)
    offset += flowprobe_l3_ip6_add (b0, e, offset);
  if (collect_ip4)
    offset += flowprobe_l3_ip4_add (b0, e, offset);
  if (flags & FLOW_RECORD_L4)
    offset += flowprobe_l4_add (b0, e, offset);

  /* Reset per flow-export counters */
  e->packetcount = 0;
  e->octetcount = 0;
  e->last_exported = vlib_time_now (vm);

  b0->current_length = offset;

  fm->context[which].next_record_offset_per_worker[my_cpu_number] = offset;
  /* Time to flush the buffer? */
  if (offset + fm->template_size[flags] > exp->path_mtu)
    flowprobe_export_send (vm, b0, which);
}

uword
flowprobe_node_fn (vlib_main_t *vm, vlib_node_runtime_t *node,
		   vlib_frame_t *frame, flowprobe_variant_t which,
		   flowprobe_direction_t direction)
{
  u32 n_left_from, *from, *to_next;
  flowprobe_next_t next_index;
  flowprobe_main_t *fm = &flowprobe_main;
  timestamp_nsec_t timestamp;

  unix_time_now_nsec_fraction (&timestamp.sec, &timestamp.nsec);

  from = vlib_frame_vector_args (frame);
  n_left_from = frame->n_vectors;
  next_index = node->cached_next_index;

  while (n_left_from > 0)
    {
      u32 n_left_to_next;

      vlib_get_next_frame (vm, node, next_index, to_next, n_left_to_next);

      while (n_left_from >= 4 && n_left_to_next >= 2)
	{
	  u32 next0 = FLOWPROBE_NEXT_DROP;
	  u32 next1 = FLOWPROBE_NEXT_DROP;
	  u16 len0, len1;
	  u32 bi0, bi1;
	  vlib_buffer_t *b0, *b1;

	  /* Prefetch next iteration. */
	  {
	    vlib_buffer_t *p2, *p3;

	    p2 = vlib_get_buffer (vm, from[2]);
	    p3 = vlib_get_buffer (vm, from[3]);

	    vlib_prefetch_buffer_header (p2, LOAD);
	    vlib_prefetch_buffer_header (p3, LOAD);

	    clib_prefetch_store (p2->data);
	    clib_prefetch_store (p3->data);
	  }

	  /* speculatively enqueue b0 and b1 to the current next frame */
	  to_next[0] = bi0 = from[0];
	  to_next[1] = bi1 = from[1];
	  from += 2;
	  to_next += 2;
	  n_left_from -= 2;
	  n_left_to_next -= 2;

	  b0 = vlib_get_buffer (vm, bi0);
	  b1 = vlib_get_buffer (vm, bi1);

	  vnet_feature_next (&next0, b0);
	  vnet_feature_next (&next1, b1);

	  len0 = vlib_buffer_length_in_chain (vm, b0);
	  ethernet_header_t *eh0 = vlib_buffer_get_current (b0);
	  u16 ethertype0 = clib_net_to_host_u16 (eh0->type);

	  if (PREDICT_TRUE ((b0->flags & VNET_BUFFER_F_FLOW_REPORT) == 0))
	    add_to_flow_record_state (
	      vm, node, fm, b0, timestamp, len0,
	      flowprobe_get_variant (which, fm->context[which].flags,
				     ethertype0),
	      direction, 0);

	  len1 = vlib_buffer_length_in_chain (vm, b1);
	  ethernet_header_t *eh1 = vlib_buffer_get_current (b1);
	  u16 ethertype1 = clib_net_to_host_u16 (eh1->type);

	  if (PREDICT_TRUE ((b1->flags & VNET_BUFFER_F_FLOW_REPORT) == 0))
	    add_to_flow_record_state (
	      vm, node, fm, b1, timestamp, len1,
	      flowprobe_get_variant (which, fm->context[which].flags,
				     ethertype1),
	      direction, 0);

	  /* verify speculative enqueues, maybe switch current next frame */
	  vlib_validate_buffer_enqueue_x2 (vm, node, next_index,
					   to_next, n_left_to_next,
					   bi0, bi1, next0, next1);
	}

      while (n_left_from > 0 && n_left_to_next > 0)
	{
	  u32 bi0;
	  vlib_buffer_t *b0;
	  u32 next0 = FLOWPROBE_NEXT_DROP;
	  u16 len0;

	  /* speculatively enqueue b0 to the current next frame */
	  bi0 = from[0];
	  to_next[0] = bi0;
	  from += 1;
	  to_next += 1;
	  n_left_from -= 1;
	  n_left_to_next -= 1;

	  b0 = vlib_get_buffer (vm, bi0);

	  vnet_feature_next (&next0, b0);

	  len0 = vlib_buffer_length_in_chain (vm, b0);
	  ethernet_header_t *eh0 = vlib_buffer_get_current (b0);
	  u16 ethertype0 = clib_net_to_host_u16 (eh0->type);

	  if (PREDICT_TRUE ((b0->flags & VNET_BUFFER_F_FLOW_REPORT) == 0))
	    {
	      flowprobe_trace_t *t = 0;
	      if (PREDICT_FALSE ((node->flags & VLIB_NODE_FLAG_TRACE)
				 && (b0->flags & VLIB_BUFFER_IS_TRACED)))
		t = vlib_add_trace (vm, node, b0, sizeof (*t));

	      add_to_flow_record_state (
		vm, node, fm, b0, timestamp, len0,
		flowprobe_get_variant (which, fm->context[which].flags,
				       ethertype0),
		direction, t);
	    }

	  /* verify speculative enqueue, maybe switch current next frame */
	  vlib_validate_buffer_enqueue_x1 (vm, node, next_index,
					   to_next, n_left_to_next,
					   bi0, next0);
	}

      vlib_put_next_frame (vm, node, next_index, n_left_to_next);
    }
  return frame->n_vectors;
}

static uword
flowprobe_input_ip4_node_fn (vlib_main_t *vm, vlib_node_runtime_t *node,
			     vlib_frame_t *frame)
{
  return flowprobe_node_fn (vm, node, frame, FLOW_VARIANT_IP4,
			    FLOW_DIRECTION_RX);
}

static uword
flowprobe_input_ip6_node_fn (vlib_main_t *vm, vlib_node_runtime_t *node,
			     vlib_frame_t *frame)
{
  return flowprobe_node_fn (vm, node, frame, FLOW_VARIANT_IP6,
			    FLOW_DIRECTION_RX);
}

static uword
flowprobe_input_l2_node_fn (vlib_main_t *vm, vlib_node_runtime_t *node,
			    vlib_frame_t *frame)
{
  return flowprobe_node_fn (vm, node, frame, FLOW_VARIANT_L2,
			    FLOW_DIRECTION_RX);
}

static uword
flowprobe_output_ip4_node_fn (vlib_main_t *vm, vlib_node_runtime_t *node,
			      vlib_frame_t *frame)
{
  return flowprobe_node_fn (vm, node, frame, FLOW_VARIANT_IP4,
			    FLOW_DIRECTION_TX);
}

static uword
flowprobe_output_ip6_node_fn (vlib_main_t *vm, vlib_node_runtime_t *node,
			      vlib_frame_t *frame)
{
  return flowprobe_node_fn (vm, node, frame, FLOW_VARIANT_IP6,
			    FLOW_DIRECTION_TX);
}

static uword
flowprobe_output_l2_node_fn (vlib_main_t *vm, vlib_node_runtime_t *node,
			     vlib_frame_t *frame)
{
  return flowprobe_node_fn (vm, node, frame, FLOW_VARIANT_L2,
			    FLOW_DIRECTION_TX);
}

static inline void
flush_record (flowprobe_variant_t which)
{
  vlib_main_t *vm = vlib_get_main ();
  vlib_buffer_t *b = flowprobe_get_buffer (vm, which);
  if (b)
    flowprobe_export_send (vm, b, which);
}

void
flowprobe_flush_callback_ip4 (void)
{
  flush_record (FLOW_VARIANT_IP4);
}

void
flowprobe_flush_callback_ip6 (void)
{
  flush_record (FLOW_VARIANT_IP6);
}

void
flowprobe_flush_callback_l2 (void)
{
  flush_record (FLOW_VARIANT_L2);
  flush_record (FLOW_VARIANT_L2_IP4);
  flush_record (FLOW_VARIANT_L2_IP6);
}


static void
flowprobe_delete_by_index (u32 my_cpu_number, u32 poolindex)
{
  flowprobe_main_t *fm = &flowprobe_main;
  flowprobe_entry_t *e;
  u32 h;

  e = pool_elt_at_index (fm->pool_per_worker[my_cpu_number], poolindex);

  /* Get my index */
  h = flowprobe_hash (&e->key);

  /* Reset hash */
  fm->hash_per_worker[my_cpu_number][h] = ~0;

  pool_put_index (fm->pool_per_worker[my_cpu_number], poolindex);
}


/* Per worker process processing the active/passive expired entries */
static uword
flowprobe_walker_process (vlib_main_t * vm,
			  vlib_node_runtime_t * rt, vlib_frame_t * f)
{
  flowprobe_main_t *fm = &flowprobe_main;
  flowprobe_entry_t *e;
  ipfix_exporter_t *exp = pool_elt_at_index (flow_report_main.exporters, 0);

  /*
   * $$$$ Remove this check from here and track FRM status and disable
   * this process if required.
   */
  if (ip_address_is_zero (&exp->ipfix_collector) ||
      ip_address_is_zero (&exp->src_address))
    {
      fm->disabled = true;
      return 0;
    }
  fm->disabled = false;

  u32 cpu_index = os_get_thread_index ();
  u32 *to_be_removed = 0, *i;

  /*
   * Tick the timer when required and process the vector of expired
   * timers
   */
  f64 start_time = vlib_time_now (vm);
  u32 count = 0;

  tw_timer_expire_timers_2t_1w_2048sl (fm->timers_per_worker[cpu_index],
				       start_time);

  vec_foreach (i, fm->expired_passive_per_worker[cpu_index])
  {
    u32 exported = 0;
    f64 now = vlib_time_now (vm);
    if (now > start_time + 100e-6
	|| exported > FLOW_MAXIMUM_EXPORT_ENTRIES - 1)
      break;

    if (pool_is_free_index (fm->pool_per_worker[cpu_index], *i))
      {
	clib_warning ("Element is %d is freed already\n", *i);
	continue;
      }
    else
      e = pool_elt_at_index (fm->pool_per_worker[cpu_index], *i);

    /* Check last update timestamp. If it is longer than passive time nuke
     * entry. Otherwise restart timer with what's left
     * Premature passive timer by more than 10%
     */
    if ((now - e->last_updated) < (u64) (fm->passive_timer * 0.9))
      {
	u64 delta = fm->passive_timer - (now - e->last_updated);
	e->passive_timer_handle = tw_timer_start_2t_1w_2048sl
	  (fm->timers_per_worker[cpu_index], *i, 0, delta);
      }
    else			/* Nuke entry */
      {
	vec_add1 (to_be_removed, *i);
      }
    /* If anything to report send it to the exporter */
    if (e->packetcount && now > e->last_exported + fm->active_timer)
      {
	exported++;
	flowprobe_export_entry (vm, e);
      }
    count++;
  }
  if (count)
    vec_delete (fm->expired_passive_per_worker[cpu_index], count, 0);

  vec_foreach (i, to_be_removed) flowprobe_delete_by_index (cpu_index, *i);
  vec_free (to_be_removed);

  return 0;
}

/* *INDENT-OFF* */
VLIB_REGISTER_NODE (flowprobe_input_ip4_node) = {
  .function = flowprobe_input_ip4_node_fn,
  .name = "flowprobe-input-ip4",
  .vector_size = sizeof (u32),
  .format_trace = format_flowprobe_trace,
  .type = VLIB_NODE_TYPE_INTERNAL,
  .n_errors = ARRAY_LEN (flowprobe_error_strings),
  .error_strings = flowprobe_error_strings,
  .n_next_nodes = FLOWPROBE_N_NEXT,
  .next_nodes = FLOWPROBE_NEXT_NODES,
};
VLIB_REGISTER_NODE (flowprobe_input_ip6_node) = {
  .function = flowprobe_input_ip6_node_fn,
  .name = "flowprobe-input-ip6",
  .vector_size = sizeof (u32),
  .format_trace = format_flowprobe_trace,
  .type = VLIB_NODE_TYPE_INTERNAL,
  .n_errors = ARRAY_LEN (flowprobe_error_strings),
  .error_strings = flowprobe_error_strings,
  .n_next_nodes = FLOWPROBE_N_NEXT,
  .next_nodes = FLOWPROBE_NEXT_NODES,
};
VLIB_REGISTER_NODE (flowprobe_input_l2_node) = {
  .function = flowprobe_input_l2_node_fn,
  .name = "flowprobe-input-l2",
  .vector_size = sizeof (u32),
  .format_trace = format_flowprobe_trace,
  .type = VLIB_NODE_TYPE_INTERNAL,
  .n_errors = ARRAY_LEN (flowprobe_error_strings),
  .error_strings = flowprobe_error_strings,
  .n_next_nodes = FLOWPROBE_N_NEXT,
  .next_nodes = FLOWPROBE_NEXT_NODES,
};
VLIB_REGISTER_NODE (flowprobe_output_ip4_node) = {
  .function = flowprobe_output_ip4_node_fn,
  .name = "flowprobe-output-ip4",
  .vector_size = sizeof (u32),
  .format_trace = format_flowprobe_trace,
  .type = VLIB_NODE_TYPE_INTERNAL,
  .n_errors = ARRAY_LEN (flowprobe_error_strings),
  .error_strings = flowprobe_error_strings,
  .n_next_nodes = FLOWPROBE_N_NEXT,
  .next_nodes = FLOWPROBE_NEXT_NODES,
};
VLIB_REGISTER_NODE (flowprobe_output_ip6_node) = {
  .function = flowprobe_output_ip6_node_fn,
  .name = "flowprobe-output-ip6",
  .vector_size = sizeof (u32),
  .format_trace = format_flowprobe_trace,
  .type = VLIB_NODE_TYPE_INTERNAL,
  .n_errors = ARRAY_LEN (flowprobe_error_strings),
  .error_strings = flowprobe_error_strings,
  .n_next_nodes = FLOWPROBE_N_NEXT,
  .next_nodes = FLOWPROBE_NEXT_NODES,
};
VLIB_REGISTER_NODE (flowprobe_output_l2_node) = {
  .function = flowprobe_output_l2_node_fn,
  .name = "flowprobe-output-l2",
  .vector_size = sizeof (u32),
  .format_trace = format_flowprobe_trace,
  .type = VLIB_NODE_TYPE_INTERNAL,
  .n_errors = ARRAY_LEN (flowprobe_error_strings),
  .error_strings = flowprobe_error_strings,
  .n_next_nodes = FLOWPROBE_N_NEXT,
  .next_nodes = FLOWPROBE_NEXT_NODES,
};
VLIB_REGISTER_NODE (flowprobe_walker_node) = {
  .function = flowprobe_walker_process,
  .name = "flowprobe-walker",
  .type = VLIB_NODE_TYPE_INPUT,
  .state = VLIB_NODE_STATE_INTERRUPT,
};
/* *INDENT-ON* */

/*
 * fd.io coding-style-patch-verification: ON
 *
 * Local Variables:
 * eval: (c-set-style "gnu")
 * End:
 */