ovs dpdk debug

[root@localhost ~]# ovs-ofctl dump-flows br0
 cookie=0x0, duration=706691.312s, table=0, n_packets=84180530, n_bytes=7456796909, priority=0 actions=NORMAL
 cookie=0x0, duration=706444.352s, table=1, n_packets=0, n_bytes=0, ip,nw_dst=192.168.2.210 actions=mod_dl_dst:2e:a9:be:9e:4d:07
[root@localhost ~]# ovs-ofctl del-flows br0
[root@localhost ~]# ovs-ofctl del-flows br0
[root@localhost ~]# ovs-ofctl dump-flows br0
[root@localhost ~]# 

 虚拟机

qemu-system-aarch64  -name vm2 -nographic  \
  -enable-kvm -M virt,usb=off -cpu host -smp 2 -m 4096 \
  -global virtio-blk-device.scsi=off \
  -device virtio-scsi-device,id=scsi \
  -kernel vmlinuz-4.18 --append "console=ttyAMA0  root=UUID=6a09973e-e8fd-4a6d-a8c0-1deb9556f477" \
  -initrd initramfs-4.18 \
  -drive file=vhuser-test1.qcow2  \
  -m 2048M -numa node,memdev=mem -mem-prealloc \
  -object memory-backend-file,id=mem,size=2048M,mem-path=/dev/hugepages,share=on \
  -chardev socket,id=char1,path=$VHOST_SOCK_DIR/vhost-user1 \
  -netdev type=vhost-user,id=mynet1,chardev=char1,vhostforce \
  -device virtio-net-pci,netdev=mynet1,mac=00:00:00:00:00:01,mrg_rxbuf=off \

[root@localhost ~]# ip a
1: lo:  mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: eth0:  mtu 1500 qdisc fq_codel state UP group default qlen 1000
    link/ether 00:00:00:00:00:01 brd ff:ff:ff:ff:ff:ff
    inet6 fe80::200:ff:fe00:1/64 scope link 
       valid_lft forever preferred_lft forever
[root@localhost ~]# ip a add   10.10.103.229/24 dev eth0
[root@localhost ~]# ping 10.10.103.81
PING 10.10.103.81 (10.10.103.81) 56(84) bytes of data.
64 bytes from 10.10.103.81: icmp_seq=1 ttl=64 time=0.374 ms
64 bytes from 10.10.103.81: icmp_seq=2 ttl=64 time=0.144 ms

--- 10.10.103.81 ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1001ms
rtt min/avg/max/mdev = 0.144/0.259/0.374/0.115 ms

(gdb) b dp_netdev_upcall
Breakpoint 1 at 0x96c638: file lib/dpif-netdev.c, line 6434.
(gdb) c
Continuing.
[Switching to Thread 0xfffd53ffd510 (LWP 117402)]

Breakpoint 1, dp_netdev_upcall (pmd=pmd@entry=0x18f60810, packet_=packet_@entry=0x16e0f4b80, flow=flow@entry=0xfffd53ff48d8, wc=wc@entry=0xfffd53ff4b78, ufid=ufid@entry=0xfffd53ff4328, type=type@entry=DPIF_UC_MISS, userdata=userdata@entry=0x0, 
    actions=actions@entry=0xfffd53ff4348, put_actions=0xfffd53ff4388) at lib/dpif-netdev.c:6434
6434    {
(gdb) bt
#0  dp_netdev_upcall (pmd=pmd@entry=0x18f60810, packet_=packet_@entry=0x16e0f4b80, flow=flow@entry=0xfffd53ff48d8, wc=wc@entry=0xfffd53ff4b78, ufid=ufid@entry=0xfffd53ff4328, type=type@entry=DPIF_UC_MISS, userdata=userdata@entry=0x0, 
    actions=actions@entry=0xfffd53ff4348, put_actions=0xfffd53ff4388) at lib/dpif-netdev.c:6434
#1  0x0000000000976558 in handle_packet_upcall (put_actions=0xfffd53ff4388, actions=0xfffd53ff4348, key=0xfffd53ff9350, packet=0x16e0f4b80, pmd=0x18f60810) at lib/dpif-netdev.c:6823
#2  fast_path_processing (pmd=pmd@entry=0x18f60810, packets_=packets_@entry=0xfffd53ffca80, keys=keys@entry=0xfffd53ff7d60, flow_map=flow_map@entry=0xfffd53ff5640, index_map=index_map@entry=0xfffd53ff5620 "", in_port=)
    at lib/dpif-netdev.c:6942
#3  0x0000000000977f58 in dp_netdev_input__ (pmd=pmd@entry=0x18f60810, packets=packets@entry=0xfffd53ffca80, md_is_valid=md_is_valid@entry=false, port_no=port_no@entry=2) at lib/dpif-netdev.c:7031
#4  0x0000000000978680 in dp_netdev_input (port_no=2, packets=0xfffd53ffca80, pmd=0x18f60810) at lib/dpif-netdev.c:7069
#5  dp_netdev_process_rxq_port (pmd=pmd@entry=0x18f60810, rxq=0x1721f440, port_no=2) at lib/dpif-netdev.c:4480
#6  0x0000000000978a24 in pmd_thread_main (f_=0x18f60810) at lib/dpif-netdev.c:5731
#7  0x00000000009fc5dc in ovsthread_wrapper (aux_=) at lib/ovs-thread.c:383
#8  0x0000ffffb19f7d38 in start_thread (arg=0xfffd53ffd510) at pthread_create.c:309
#9  0x0000ffffb16df690 in thread_start () from /lib64/libc.so.6
(gdb)

ovs

现在Open vSwitch主要由三个部分组成：

1. ovsdb-server：OpenFlow本身被设计成网络数据包的一种处理流程，它没有考虑软件交换机的配 置，例如配置QoS，关联SDN控制器等。ovsdb-server是Open vSwitch对于OpenFlow实现的补 充，它作为OpenvSwitch的configuration database，保存Open vSwitch的持久化数据。
2. ovs-vswitchd：运行在用户空间的转发程序，接收SDN控制器下发的OpenFlow规则。并且通知 OVS内核模块该如何处理网络数据包。
3. OVS内核模块：运行在内核空间的转发程序，根据ovs-vswitchd的指示，处理网络数据包

 Open vSwitch中有快速路径（fast path）和慢速路径（slow path）。其中ovs-vswitchd代表了slow path，OVS内核模块代表了fast path。现在OpenFlow存储在slow path中，但是为了快速转发，网络包 应该尽可能的在fast path中转发。因此，OpenVSwitch按照下面的逻辑完成转发。

当一个网络连接的第一个网络数据包（首包）被发出时，OVS内核模块会先收到这个数据包。但是内核 模块现在还不知道如何处理这个包，因为所有的OpenFlow都存在ovs-vswitchd，因此它的默认行为是 将这个包上送到ovs-vswitchd。ovs-vswitchd通过OpenFlow pipeline，处理完网络数据包送回给OVS内 核模块，同时，ovs-vswitchd还会生成一串类似于OpenFlow Action，但是更简单的datapath action。 这串datapath action会一起送到OVS内核模块。因为同一个网络连接的所有网络数据包特征（IP， MAC，端口号）都一样，当OVS内核模块收到其他网络包的时候，可以直接应用datapath action。

这样，成功的解决了之前的问题。首先，内核模块不用关心OpenFlow的变化，不用考虑OpenVSwitch的代码更新，这些都在ovs-vswitchd完成。其次，整个网络连接，只有首包需要经过OpenFlow pipeline的处理，其余的包在内核模块匹配网络包特征，直接应用datapath action，就可以完成转发。OVS内核模块通过缓存来保持datapath action的记录。稍微早期的内核模块实现了名为microflow的缓存，这个缓存就是一个hash map，key是所有OpenFlow可能匹配的值对应的hash值，包括了网络2-4层header数据和一些其他的metadata例如in_port，value就是datapath action。hash map可以实现O(1)的查找时间，这样可以在OVS内核模块中实现高效的查找转发。

kernel收包

 通过vport注册的回调函数netdev_frame_hook()->netdev_frame_hook()-> netdev_port_receive()->ovs_vport_receive()处理接收报文，ovs_flow_key_extract()函数生成flow的key内容用以接下来进行流表匹配，最后调用ovs_dp_process_packet()函数进入真正的ovs数据包处理,代码流程如下：

action处理

ovs的action类型如下，使用nla_type()函数获取nl_type的值，入口处理函数为do_execute_actions()。

enum ovs_action_attr {
        OVS_ACTION_ATTR_UNSPEC,
        OVS_ACTION_ATTR_OUTPUT,       /* u32 port number. */
        OVS_ACTION_ATTR_USERSPACE,    /* Nested OVS_USERSPACE_ATTR_*. */
        OVS_ACTION_ATTR_SET,          /* One nested OVS_KEY_ATTR_*. */
        OVS_ACTION_ATTR_PUSH_VLAN,    /* struct ovs_action_push_vlan. */
        OVS_ACTION_ATTR_POP_VLAN,     /* No argument. */
        OVS_ACTION_ATTR_SAMPLE,       /* Nested OVS_SAMPLE_ATTR_*. */
        OVS_ACTION_ATTR_RECIRC,       /* u32 recirc_id. */
        OVS_ACTION_ATTR_HASH,         /* struct ovs_action_hash. */
        OVS_ACTION_ATTR_PUSH_MPLS,    /* struct ovs_action_push_mpls. */
        OVS_ACTION_ATTR_POP_MPLS,     /* __be16 ethertype. */
        OVS_ACTION_ATTR_SET_MASKED,   /* One nested OVS_KEY_ATTR_* including
                                       * data immediately followed by a mask.
                                       * The data must be zero for the unmasked
                                       * bits. */
        OVS_ACTION_ATTR_CT,           /* Nested OVS_CT_ATTR_* . */
        OVS_ACTION_ATTR_TRUNC,        /* u32 struct ovs_action_trunc. */
        OVS_ACTION_ATTR_PUSH_ETH,     /* struct ovs_action_push_eth. */
        OVS_ACTION_ATTR_POP_ETH,      /* No argument. */
        OVS_ACTION_ATTR_CT_CLEAR,     /* No argument. */
        OVS_ACTION_ATTR_PUSH_NSH,     /* Nested OVS_NSH_KEY_ATTR_*. */
        OVS_ACTION_ATTR_POP_NSH,      /* No argument. */
        OVS_ACTION_ATTR_METER,        /* u32 meter ID. */
        OVS_ACTION_ATTR_CLONE,        /* Nested OVS_CLONE_ATTR_*.  */
        OVS_ACTION_ATTR_CHECK_PKT_LEN, /* Nested OVS_CHECK_PKT_LEN_ATTR_*. */

        __OVS_ACTION_ATTR_MAX,        /* Nothing past this will be accepted
                                       * from userspace. */

#ifdef __KERNEL__
        OVS_ACTION_ATTR_SET_TO_MASKED, /* Kernel module internal masked
                                        * set action converted from
                                        * OVS_ACTION_ATTR_SET. */
#endif
};

OVS_ACTION_ATTR_OUTPUT：获取port号，调用do_output()发送报文到该port；
OVS_ACTION_ATTR_USERSPACE：调用output_userspace()发送到用户态；
OVS_ACTION_ATTR_HASH：调用execute_hash()获取skb的hash赋值到ovs_flow_hash
OVS_ACTION_ATTR_PUSH_VLAN：调用push_vlan()增加vlan头部

upcall处理

当没有找到匹配的流表时，内核通过netlink发送报文到用户层处理，入口函数ovs_dp_upcall()，该函数调用queue_userspace_packet()构造发往用户层的skb，通过netlink通信机制发送到用户层，其中形成的主要数据格式如下：

output_userspace
      ovs_dp_upcall 
         queue_userspace_packet
             skb_zerocopy(user_skb, skb, skb->len, hlen);    //upcall信息对象添加报文

void ovs_dp_process_packet(struct sk_buff *skb, struct sw_flow_key *key)
{
        const struct vport *p = OVS_CB(skb)->input_vport;
        struct datapath *dp = p->dp;
        struct sw_flow *flow;
        struct sw_flow_actions *sf_acts;
        struct dp_stats_percpu *stats;
        u64 *stats_counter;
        u32 n_mask_hit;

        stats = this_cpu_ptr(dp->stats_percpu);

        /* Look up flow. */
        flow = ovs_flow_tbl_lookup_stats(&dp->table, key, &n_mask_hit);
        if (unlikely(!flow)) {
                struct dp_upcall_info upcall;
                int error;

                memset(&upcall, 0, sizeof(upcall));
                upcall.cmd = OVS_PACKET_CMD_MISS;
                upcall.portid = ovs_vport_find_upcall_portid(p, skb);
                upcall.mru = OVS_CB(skb)->mru;
                error = ovs_dp_upcall(dp, skb, key, &upcall, 0);
                if (unlikely(error))
                        kfree_skb(skb);
                else
                        consume_skb(skb);
                stats_counter = &stats->n_missed;
                goto out;
        }

        ovs_flow_stats_update(flow, key->tp.flags, skb);
        sf_acts = rcu_dereference(flow->sf_acts);
        ovs_execute_actions(dp, skb, sf_acts, key);

        stats_counter = &stats->n_hit;

out:
        /* Update datapath statistics. */
        u64_stats_update_begin(&stats->syncp);
        (*stats_counter)++;
        stats->n_mask_hit += n_mask_hit;
        u64_stats_update_end(&stats->syncp);
}

学习地址: Dpdk/网络协议栈/vpp/OvS/DDos/NFV/虚拟化/高性能专家-学习视频教程-腾讯课堂  【免费订阅】
更多DPDK相关学习资料有需要的可以自行报名学习,免费订阅,久学习,或点击这里加qun免费
领取,关注我持续更新哦! !  

action

遍历所有的action，支持的有OFPACT_OUTPUT、OFPACT_GROUP、OFPACT_CONTROLLER、OFPACT_ENQUEUE、OFPACT_SET_VLAN_VID、OFPACT_SET_VLAN_PCP、OFPACT_STRIP_VLAN、OFPACT_PUSH_VLAN、OFPACT_SET_ETH_SRC、OFPACT_SET_ETH_DST、OFPACT_SET_IPV4_SRC、
OFPACT_SET_IPV4_DST、OFPACT_SET_IP_DSCP、OFPACT_SET_IP_ECN、OFPACT_SET_IP_TTL、OFPACT_SET_L4_SRC_PORT、OFPACT_SET_L4_DST_PORT、OFPACT_RESUBMIT、OFPACT_SET_TUNNEL、OFPACT_SET_QUEUE、OFPACT_POP_QUEUE、OFPACT_REG_MOVE、
OFPACT_SET_FIELD、OFPACT_STACK_PUSH、OFPACT_STACK_POP、OFPACT_PUSH_MPLS、OFPACT_POP_MPLS、OFPACT_SET_MPLS_LABEL、OFPACT_SET_MPLS_TC、OFPACT_SET_MPLS_TTL、OFPACT_DEC_MPLS_TTL、OFPACT_DEC_TTL、OFPACT_NOTE、OFPACT_MULTIPATH、
OFPACT_BUNDLE、OFPACT_OUTPUT_REG、OFPACT_OUTPUT_TRUNC、OFPACT_LEARN、OFPACT_CONJUNCTION、OFPACT_EXIT、OFPACT_UNROLL_XLATE、OFPACT_FIN_TIMEOUT、OFPACT_CLEAR_ACTIONS、OFPACT_WRITE_ACTIONS、OFPACT_WRITE_METADATA、OFPACT_METER、
OFPACT_GOTO_TABLE、OFPACT_SAMPLE、OFPACT_CLONE、OFPACT_CT、OFPACT_CT_CLEAR、OFPACT_NAT、OFPACT_DEBUG_RECIRC，
action太多，我们先介绍几个常用的
OFPACT_OUTPUT，xlate_output_action会根据端口情况进行一些操作，这块不细看了
OFPACT_CONTROLLER，execute_controller_action生成一个packet_in报文，然后发送
OFPACT_SET_ETH_SRC、OFPACT_SET_ETH_DST、OFPACT_SET_IPV4_SRC、OFPACT_SET_IPV4_DST、OFPACT_SET_IP_DSCP、OFPACT_SET_IP_ECN、OFPACT_SET_IP_TTL、OFPACT_SET_L4_SRC_PORT、OFPACT_SET_L4_DST_PORT，修改源目的mac、IP以及DSCP、ECN、TTL和L4的源目的端口
OFPACT_RESUBMIT，xlate_ofpact_resubmit会继续查找指定的table的流表
OFPACT_SET_TUNNEL，设置tunnel id
OFPACT_CT，compose_conntrack_action执行完ct的设置之后回调do_xlate_actions执行其他的action

packetin数据包

ovs-ofctl add-flow br0 in_port=1,actions=controller

OFPP_CONTROLLER
xlate_actions
   ┣━do_xlate_actions(ofpacts, ofpacts_len, &ctx, true, false);
        ┣━xlate_output_action
do_xlate_actions
{
    switch (a->type) {
        case OFPACT_OUTPUT:
            xlate_output_action(ctx, ofpact_get_OUTPUT(a)->port,
                                ofpact_get_OUTPUT(a)->max_len, true, last,
                                false, group_bucket_action);
            break;
}
 
 xlate_output_action(struct xlate_ctx *ctx, ofp_port_t port,
                    uint16_t controller_len, bool may_packet_in,
                    bool is_last_action, bool truncate,
                    bool group_bucket_action)
{
        case OFPP_CONTROLLER:
        xlate_controller_action(ctx, controller_len,
                                (ctx->in_packet_out ? OFPR_PACKET_OUT
                                 : group_bucket_action ? OFPR_GROUP
                                 : ctx->in_action_set ? OFPR_ACTION_SET
                                 : OFPR_ACTION),
                                0, UINT32_MAX, NULL, 0);
}