learning:vpp/classify(2)

dpdk-vpp源码解读

发布于 2023-01-04 04:40:46

2.1K03

代码可运行

文章被收录于专栏：DPDK VPP源码分析DPDK VPP源码分析

运行总次数：3

代码可运行

classify table cli讲解

classify table [miss-next|l2-miss_next|acl-miss-next <next_index>]
 mask <mask-value> buckets <nn> [skip <n>] [match <n>]
 [current-data-flag <n>] [current-data-offset <n>] [table <n>]
 [memory-size <nn>[M][G]] [next-table <n>]
 [del] [del-chain]

作用如下图所示从table内存池上申请table内存，在main_heap上申请桶占用空间，从系统内存映射classify session中匹配规则表配置空间。这里操作和bihash一致。

下面分别来讲解： 1、classify miss处理，三者只能选择一个。当classify 匹配miss时，送到下一个node在当前classify node的slot id。个人理解时支持三种类型处理逻辑，如下： l2: 二层转发流程处理逻辑，对应node节点 l2-input-classify. l3: 三层转发流程处理逻辑，区别ipv4、ipv6节点，分别对应node节点ip4-classify、ip6-classify. acl: acl功能扽别支持二层、三层转发处理。 l2转发对应node节点 l2-input-acl,l2-output-acl、 L3转发 ip4-inacl，ip4-outacl，ip6-inacl，ip6-outacl）.

miss-next，主要是三层处理流程中配置，

#三层处理流程，
miss-next  [ip4-node <node name>] | [ip6-node <node name>]
| drop | local | rewrite | <slot id>

l2-miss_next 主要处理二层转发流程。

l2-miss_next drop | ethernet-input | ip4-input | <slot id> |
[input-node <node name> | output-node <node name>

acl-miss-next 配置s使能acl时使用

acl-miss-next deny | perimit | <slot id> |
[ip4-node <node name>] | [ip6-node <node name>]

2、mask匹配那些bit，或过滤那些bit。设置该表是用数据包中的哪些字段用于过滤；mask是以16个字节为一组的数据。 mask支持支持配置16进制字符串方式，或者指定报文内容。下面分别说明。

#设置匹配报文ip version及protocol字段。
mask hex 0000000000000000000000000000FF0000000000000000FF
                                     |                 |
 ip                                 version       protocol

mask l2 根据vlan帧格式及QinQ帧格式，来确定mask长度

mask l2 src | dst  | proto   | tag1 | tag2 | ignore-tag1 | 
 #    源mac |目的mac|eth type | vlan        | 忽略vlan
ignore-tag2 | cos1 | cos2 | dot1q   | dot1ad |
#           |Priority数值  | 一层vlan |二层vlan |

ignore-tag1（18字节）,ignore-tag2（22字节）用来确定二层头长度。不关心信息。

mask l3 支持匹配内容：

#ip4报文支持匹配内容
mask l3 ip4 version [hdr_length] [src/<prefix_len>] [dst/<prefix_len]
[src|src_address] [dst|dst_address] [tos] [length] [fragment_id] [ttl]
 [proto|protocol] [checksum]
#ip6报文支持匹配内容
mask l3 ip6 [version] [traffic-class] [flow-label]
[src|src_address] [dst|dst_address] [payload_length]
[hop_limit] [proto|protocol]

mask l4支持匹配内容

mask l4 [ [tcp |udp] [src_port | dst_port]  ] | src_port  | dst_port ]

这里需要注意，不支持单独匹配l4，必须需要指定l3内容，个人认为是因为无法确定mask长度，ip4和ip6报文头长度不一致。

3、其他参数说明： buckets：buckets的最大数目,默认是2，需要用户自己根据支持匹配规则的大小，来设置有效的桶大小。 [memory-size[M][G]]：classify table对应结构中mheap的大小。默认是2M的大小，需要用户跟组匹配规则的大小，来设置有效的mheap大小。

这里的hash处理等同于bihash，因为classify entry的不固定性，所以没办法直接shiyongbihash。

skip：mask中跳过的全0数据的组数（16字节为一组） match：mask中有效组数（16字节为一组） current-data-flag：标识数据包过滤的头信息获取是从vlib_buff的curretn_data处加current-data-offset 获取. current-data-offset：与current-data-flag配合使用。如下代码所示：

#判断current_data_flag 是否设置，获取classify匹配头起始位置。
if (t[0]->current_data_flag == CLASSIFY_FLAG_USE_CURR_DATA)
{ #标识数据包过滤的头信息获取是从vlib_buff的curretn_data处加current-data-offset 获取.
    h[0] = (void *) vlib_buffer_get_current (b[0]) + t[0]->current_data_offset;
}
else
{   #从原始报文data处获取classify 过滤头信息。
    h[0] = b[0]->data;
}
#计算hash
hash[0] = vnet_classify_hash_packet_inline (t[0], (u8 *) h[0]);
#查询报文是否命中classify 表内容。
e[0] = vnet_classify_find_entry_inline (t[0], (u8 *) h[0], hash[0], now);

table：表索引；新增就不填，只有更新表信息才指定索引，更新仅支持多table链时，设置next table index。

l2 classify acl测试

1、vpp基本配置

vpp2 接口GigabitEthernetb/0/0 接口与内核接口ens33直连。在vpp2上创建一个二层域bridge-domain 13及loopback接口，loop0接口及GigabitEthernetb/0/0加入二层域。从而实现一个二层转发流程。

#vpp2配置
#创建一个二层域BD 13
create bridge-domain 13 learn 1 forward 1 uu-flood 1 flood 1 arp-term 1
#
set interface state GigabitEthernetb/0/0 up
set interface l2 bridge GigabitEthernetb/0/0 13
#创建loopback接口
loopback create mac 11:22:33:44:55:66
#loopback接口加入二层域名，并设置为bvi接口
set interface l2 bridge loop0 13 bvi
#配置loopback接口ip地址
set interface ip table loop0 0
set interface state loop0 up
set interface ip addr loop0 192.168.3.1/24
#vpp2 接口配置查询情况：
learning_vpp2# show interface addr                 
GigabitEthernetb/0/0 (up):
  L2 bridge bd-id 13 idx 1 shg 0  
local0 (dn):
loop0 (up):
  L2 bridge bd-id 13 idx 1 shg 0 bvi
  L3 192.168.3.1/24
#内核ens33接口与GigabitEthernetb/0/0直连
[root@learning_vpp2 vpp]# ifconfig ens33
ens33: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet 192.168.3.2  netmask 255.255.255.0  broadcast

2、配置classify table 、session

 #创建clssify 表匹配IP version及IP Protocol。
classify table acl-miss-next deny mask hex 0000000000000000000000000000FF0000000000000000FF
#匹配规则：ipv4及tcp报文
classify session acl-hit-next permit table-index 0 match hex 000000000000000000000000000045000000000000000006
#匹配规则：ipv4及icmp报文
classify session acl-hit-next permit table-index 0 match hex 000000000000000000000000000045000000000000000001
##匹配规则：ipv4及icmp报文
classify session acl-hit-next permit table-index 0 match hex 000000000000000000000000000045000000000000000011

#查询配置表：
learning_vpp2# show classify tables index 0 verbose
  TableIdx  Sessions   NextTbl  NextNode
         0         3        -1         0
  Heap: base 0x7fffc920d000, size 2m, locked, unmap-on-destroy, name 'classify'
          page stats: page-size 4K, total 512, mapped 2, not-mapped 0, unknown 510
            numa 0: 2 pages, 8k bytes
          total: 1.99M, used: 1.92K, free: 1.99M, trimmable: 1.99M
  nbuckets 2, skip 0 match 2 flag 0 offset 0
  mask 0000000000000000000000000000ff0000000000000000ff0000000000000000
  linear-search buckets 0

[0]: heap offset 1856, elts 2, normal
    0: [1856]: next_index -1 advance 0 opaque -1 action 0 metadata 0
        k: 0000000000000000000000000000450000000000000000110000000000000000
        hits 0, last_heard 0.00

[1]: heap offset 1280, elts 2, normal
    0: [1280]: next_index -1 advance 0 opaque -1 action 0 metadata 0
        k: 0000000000000000000000000000450000000000000000060000000000000000
        hits 0, last_heard 0.00

    1: [1344]: next_index -1 advance 0 opaque -1 action 0 metadata 0
        k: 0000000000000000000000000000450000000000000000010000000000000000
        hits 8, last_heard 2062.06

    3 active elements
    1 free lists
    0 linear-search buckets

3、接口绑定clssify table

set interface input acl intfc  GigabitEthernetb/0/0 l2-table 0
set interface output acl intfc  GigabitEthernetb/0/0 l2-table 0

从内核发起ping报文，trace流程如下：

00:34:10:049788: dpdk-input
  GigabitEthernetb/0/0 rx queue 0
  buffer 0x9bad1: current data 0, length 98, buffer-pool 0, ref-count 1, totlen-nifb 0, trace handle 0x1000000
                  ext-hdr-valid 
                  l4-cksum-computed l4-cksum-correct 
  PKT MBUF: port 0, nb_segs 1, pkt_len 98
    buf_len 2176, data_len 98, ol_flags 0x80, data_off 128, phys_addr 0x53aeb4c0
    packet_type 0x91 l2_len 0 l3_len 0 outer_l2_len 0 outer_l3_len 0
    rss 0x0 fdir.hi 0x0 fdir.lo 0x0
    Packet Offload Flags
      PKT_RX_IP_CKSUM_GOOD (0x0080) IP cksum of RX pkt. is valid
    Packet Types
      RTE_PTYPE_L2_ETHER (0x0001) Ethernet packet
      RTE_PTYPE_L3_IPV4_EXT_UNKNOWN (0x0090) IPv4 packet with or without extension headers
  IP4: 00:0c:29:17:0a:58 -> 11:22:33:44:55:66
  ICMP: 192.168.3.2 -> 192.168.3.1
    tos 0x00, ttl 64, length 84, checksum 0x8734 dscp CS0 ecn NON_ECN
    fragment id 0x2c21, flags DONT_FRAGMENT
  ICMP echo_request checksum 0x8dcb id 2020
00:34:10:049819: ethernet-input
  frame: flags 0x1, hw-if-index 1, sw-if-index 1
  IP4: 00:0c:29:17:0a:58 -> 11:22:33:44:55:66
00:34:10:049838: l2-input
  l2-input: sw_if_index 1 dst 11:22:33:44:55:66 src 00:0c:29:17:0a:58 [l2-input-acl l2-learn l2-flood ]
00:34:10:049844: l2-input-acl
  INACL: sw_if_index 1, next_index 9, table 0, offset 1344
00:34:10:067214: l2-learn
  l2-learn: sw_if_index 1 dst 11:22:33:44:55:66 src 00:0c:29:17:0a:58 bd_index 1
00:34:10:067221: l2-flood
  l2-flood: sw_if_index 1 dst 11:22:33:44:55:66 src 00:0c:29:17:0a:58 bd_index 1
00:34:10:067228: ip4-input
  ICMP: 192.168.3.2 -> 192.168.3.1
    tos 0x00, ttl 64, length 84, checksum 0x8734 dscp CS0 ecn NON_ECN
    fragment id 0x2c21, flags DONT_FRAGMENT
  ICMP echo_request checksum 0x8dcb id 2020
00:34:10:067236: ip4-lookup
  fib 0 dpo-idx 7 flow hash: 0x00000000
  ICMP: 192.168.3.2 -> 192.168.3.1
    tos 0x00, ttl 64, length 84, checksum 0x8734 dscp CS0 ecn NON_ECN
    fragment id 0x2c21, flags DONT_FRAGMENT
  ICMP echo_request checksum 0x8dcb id 2020
00:34:10:067248: ip4-local
    ICMP: 192.168.3.2 -> 192.168.3.1
      tos 0x00, ttl 64, length 84, checksum 0x8734 dscp CS0 ecn NON_ECN
      fragment id 0x2c21, flags DONT_FRAGMENT
    ICMP echo_request checksum 0x8dcb id 2020
00:34:10:067457: ip4-icmp-input
  ICMP: 192.168.3.2 -> 192.168.3.1
    tos 0x00, ttl 64, length 84, checksum 0x8734 dscp CS0 ecn NON_ECN
    fragment id 0x2c21, flags DONT_FRAGMENT
  ICMP echo_request checksum 0x8dcb id 2020
00:34:10:067462: ip4-icmp-echo-request
  ICMP: 192.168.3.2 -> 192.168.3.1
    tos 0x00, ttl 64, length 84, checksum 0x8734 dscp CS0 ecn NON_ECN
    fragment id 0x2c21, flags DONT_FRAGMENT
  ICMP echo_request checksum 0x8dcb id 2020
00:34:10:067468: ip4-load-balance
  fib 0 dpo-idx 2 flow hash: 0x00000000
  ICMP: 192.168.3.1 -> 192.168.3.2
    tos 0x00, ttl 64, length 84, checksum 0x0266 dscp CS0 ecn NON_ECN
    fragment id 0xb0ef, flags DONT_FRAGMENT
  ICMP echo_reply checksum 0x95cb id 2020
00:34:10:067471: ip4-rewrite
  tx_sw_if_index 4 dpo-idx 2 : ipv4 via 192.168.3.2 loop0: mtu:9000 next:3 flags:[] 000c29170a581122334455660800 flow hash: 0x00000000
  00000000: 000c29170a58112233445566080045000054b0ef400040010266c0a80301c0a8
  00000020: 0302000095cb07e400017124c760000000005ef70c00000000001011
00:34:10:067474: loop0-output
  loop0 
  IP4: 11:22:33:44:55:66 -> 00:0c:29:17:0a:58
  ICMP: 192.168.3.1 -> 192.168.3.2
    tos 0x00, ttl 64, length 84, checksum 0x0266 dscp CS0 ecn NON_ECN
    fragment id 0xb0ef, flags DONT_FRAGMENT
  ICMP echo_reply checksum 0x95cb id 2020
00:34:10:067488: l2-input
  l2-input: sw_if_index 4 dst 00:0c:29:17:0a:58 src 11:22:33:44:55:66 [l2-fwd l2-flood arp-term-l2bd l2-flood ]
00:34:10:067490: l2-fwd
  l2-fwd:   sw_if_index 4 dst 00:0c:29:17:0a:58 src 11:22:33:44:55:66 bd_index 1 result [0x1160000000001, 1] none
00:34:10:067494: l2-output
  l2-output: sw_if_index 1 dst 00:0c:29:17:0a:58 src 11:22:33:44:55:66 data 08 00 45 00 00 54 b0 ef 40 00 40 01
00:34:10:067498: l2-output-acl
  OUTACL: sw_if_index 1, next_index 1, table 0, offset 1344
00:34:10:067505: GigabitEthernetb/0/0-output
  GigabitEthernetb/0/0 
  IP4: 11:22:33:44:55:66 -> 00:0c:29:17:0a:58
  ICMP: 192.168.3.1 -> 192.168.3.2
    tos 0x00, ttl 64, length 84, checksum 0x0266 dscp CS0 ecn NON_ECN
    fragment id 0xb0ef, flags DONT_FRAGMENT
  ICMP echo_reply checksum 0x95cb id 2020
00:34:10:067506: GigabitEthernetb/0/0-tx
  GigabitEthernetb/0/0 tx queue 1
  buffer 0x9bad1: current data 0, length 98, buffer-pool 0, ref-count 1, totlen-nifb 0, trace handle 0x1000000
                  ext-hdr-valid 
                  l4-cksum-computed l4-cksum-correct local l2-hdr-offset 0 l3-hdr-offset 14 
  PKT MBUF: port 0, nb_segs 1, pkt_len 98
    buf_len 2176, data_len 98, ol_flags 0x80, data_off 128, phys_addr 0x53aeb4c0
    packet_type 0x91 l2_len 0 l3_len 0 outer_l2_len 0 outer_l3_len 0
    rss 0x0 fdir.hi 0x0 fdir.lo 0x0
    Packet Offload Flags
      PKT_RX_IP_CKSUM_GOOD (0x0080) IP cksum of RX pkt. is valid
    Packet Types
      RTE_PTYPE_L2_ETHER (0x0001) Ethernet packet
      RTE_PTYPE_L3_IPV4_EXT_UNKNOWN (0x0090) IPv4 packet with or without extension headers
  IP4: 11:22:33:44:55:66 -> 00:0c:29:17:0a:58
  ICMP: 192.168.3.1 -> 192.168.3.2
    tos 0x00, ttl 64, length 84, checksum 0x0266 dscp CS0 ecn NON_ECN
    fragment id 0xb0ef, flags DONT_FRAGMENT
  ICMP echo_reply checksum 0x95cb id 2020