原文出处：
本文只是一个内核网络协议的实践的例子，先说明添加的目的。
内核版本：2.6.34；在支持802.1主机上，报文的一般格式：
现在需要支持一种新的协议[二层] – BRCM协议，与IP等协议不同，它位于2层，拥有6字节的头部和4字节的尾部，添加的层次决定了比起添加其它协议要复杂一些，新的报文格式如下，而我们的目的就是要网络协议栈能正常处理这样的报文：
实际上BRCM是一种交换机的内部协议，用处是让交换机管理端口能通过BRCM获取报文来自于交换机的哪个端口，或者指定报文从交换机哪个端口出去；当然，这不是我们关心的内容，我们只需要为它挑选一个协议号0x8744，其余内容置0就可以了。因此，brcm头部会填写成 88 74 00 00 00 00，brcm尾部会填写成 00 00。一个新协议的报文内容用wireshark等捕包工具查看的形式如下：
[源/目的mac]02 03 04 05 06 07 10 11 12 13 14 15
[BRCM报头]88 74 00 00 00 00
[Vlan报头]81 00 00 01
[报文内容]……..
先从设备的概念来看下添加BRCM协议后的层次图：
eth1代表实际的B4401物理网卡；eth1.X/brcm0.x代表VLAN创建的虚拟网卡，后面的数字X是vlan号；brcm0代表BRCM创建的虚拟网卡，数字0表示测试用；从图中可以看到，brcm协议的添加是通过添加brcmX虚拟网卡接口实现的。
如果BRCM协议添加正确，那么最终的结果应该是：
在接收报文时，通过虚拟设备BRCM会脱去brcm的6字节的头部与4字节尾部，当然，协议做的远远不只这些，但这是核心。
在发送报文时，通过虚拟设备BRCM会添加6字节的头部与4字节的尾部，当然，协议做的远远不只这些，但这是核心。
要做的工作大致是以下几项：
1. 确定brcm_dev的存储数据结构
2. 编写struct ptype_base brcm_packet_type
3. 添加ioctl调用供用户空间调用，至少包括brcm_dev的创建和删除
4. 添加notifier机制、netlink机制、proc机制
5. 添加编译用的Kconfig, Makefile等，并修改Menuconfig
数据结构是核心，每一步也是一个网络协议模块的核心。
内核版本：2.6.34
实现思路：
报文在网络协议栈中的流动，对于接收来讲，是对报文的脱壳的过程，由于报文是已知的输入，只要逐个解析协议号；对于发送来讲，是各层发送函数的嵌套调用，由于没有已知的输入，只能按事先设计好的协议进行层层构造。但无论报文怎样的流动，核心是报文所在设备(skb->dev)的变化，相当于各层之间传递的交接棒。
按照上述思路，brcm协议接收的处理作为模块brcm_packet_type加入到ptype_base中就可以了；brcm协议发送的处理则复杂一点，发送的嵌套调用完全是依赖于设备来推动的，因此要有一种新创建的设备X，插入到vlan设备和网卡设备之间。
因此，至少要有brcm_packet_type来加入ptype_base和register_brcm_dev()来向系统注册设备X。进一步考虑，设备X在全局量init_net中有存储，但我们还需要知道设备X与vlan设备以及网卡设备是何种组织关系，所以在这里设计了brcm_group_hash来存储这种关系。为了对设备感兴趣的事件作出响应，添加自己的notifier到netdev_chain中。另外，为了用户空间具有一定控制能力(如创建、删除)，还需要添加brcm相关的ioctl调用。为了让它看起来更完整，一种新的设备在proc中也应有对应项，用来调试和查看设备。
从最简单开始要让网络协议栈能够接收一种新协议是很简单的，由于已经有报文作为输入，我们要做的仅仅是编写好brcm_packet_type，然后在注册模块时只用做一件事：dev_add_pack。
staticint__init brcm_proto_init(void)
{
dev_add_pack(&brcm_packet_type);
}
staticstructpacket_type brcm_packet_type __read_mostly = {
.type = cpu_to_be16(ETH_P_BRCM),
.func = brcm_skb_recv,/* BRCM receive method */
};
intbrcm_skb_recv(structsk_buff *skb,structnet_device *dev,
structpacket_type *ptype,structnet_device *orig_dev)
{
structbrcm_hdr *bhdr;
structbrcm_rx_stats *rx_stats;
skb = skb_share_check(skb, GFP_ATOMIC);
if(!skb)
gotoerr_free;
bhdr = (structbrcm_hdr *)skb->data;
rcu_read_lock();
skb_pull_rcsum(skb, BRCM_HLEN);
// set protocol
skb->protocol = bhdr->brcm_encapsulated_proto;
// reorder skb
skb = brcm_check_reorder_header(skb);
if(!skb)
gotoerr_unlock;
netif_rx(skb);
rcu_read_unlock();
returnNET_RX_SUCCESS;
err_unlock:
rcu_read_unlock();
err_free:
kfree_skb(skb);
returnNET_RX_DROP;
}static int __init brcm_proto_init(void){ dev_add_pack(&brcm_packet_type);}static struct packet_type brcm_packet_type __read_mostly = { .type = cpu_to_be16(ETH_P_BRCM), .func = brcm_skb_recv, /* BRCM receive method */};int brcm_skb_recv(struct sk_buff *skb, struct net_device *dev, struct packet_type *ptype, struct net_device *orig_dev){ struct brcm_hdr *bhdr; struct brcm_rx_stats *rx_stats; skb = skb_share_check(skb, GFP_ATOMIC); if(!skb) goto err_free; bhdr = (struct brcm_hdr *)skb->data; rcu_read_lock(); skb_pull_rcsum(skb, BRCM_HLEN); // set protocol skb->protocol = bhdr->brcm_encapsulated_proto; // reorder skb skb = brcm_check_reorder_header(skb); if (!skb) goto err_unlock; netif_rx(skb); rcu_read_unlock(); return NET_RX_SUCCESS;err_unlock: rcu_read_unlock();err_free: kfree_skb(skb); return NET_RX_DROP;}
注册这个模块后，协议栈就能正常接收带brcm报头的报文的，代码中ETH_P_BRCM是brcm的协议号，BRCM_HLEN是brcm的报头长度。正是由于有报文作为输入，接收变得十分简单。
但这仅仅是能接收而已，发送的报文还是不带brcm报头的，而且接收的这段代码也很粗略，没有变更skb的设备，没有记录流量，没有对brcm报头作有意义的处理，下面逐一进行添加。
设备的相关定义一种设备就是net_device类型，而每种设备都有自己的私有变量，它存储在net_device末尾，定义如下，其中real_dev指向下层设备，这是最基本属性，其余可以视需要自己设定，brcm_rx_stats则是该设备接收流量统计：
structbrcm_dev_info{
structnet_device *real_dev;
u16 brcm_port;
unsignedcharreal_dev_addr[ETH_ALEN];
structproc_dir_entry *dent;
structbrcm_rx_stats __percpu *brcm_rx_stats;
};
structbrcm_rx_stats {
unsignedlongrx_packets;
unsignedlongrx_bytes;
unsignedlongmulticast;
unsignedlongrx_errors;
};struct brcm_dev_info{ struct net_device *real_dev; u16 brcm_port; unsigned char real_dev_addr[ETH_ALEN]; struct proc_dir_entry *dent; struct brcm_rx_stats __percpu *brcm_rx_stats;};struct brcm_rx_stats { unsigned long rx_packets; unsigned long rx_bytes; unsigned long multicast; unsigned long rx_errors;};
设备间的关系问题如果brcm仅仅是只有一个设备，则无需数据结构来存储这种关系，一个全局全变的brcm_dev就可以了。这里的设计考虑的是复杂的情况，可以存在多个下层设备，多个brcm设备，之间没有固定的关系。所以需要一种数据结构来存储这种关系- brcm_group_hash。下面是一个简单的图示：
各个数据结构定义如下：
staticstructhlist_head brcm_group_hash[BRCM_GRP_HASH_SIZE];
structbrcm_group {
structhlist_node hlist;
structnet_device *real_dev;
intnr_ports;
intkillall;
structnet_device *brcm_devices_array[BRCM_GROUP_ARRAY_LEN];
structrcu_head rcu;
};static struct hlist_head brcm_group_hash[BRCM_GRP_HASH_SIZE];struct brcm_group { struct hlist_node hlist; struct net_device *real_dev; int nr_ports; int killall; struct net_device *brcm_devices_array[BRCM_GROUP_ARRAY_LEN]; struct rcu_head rcu;};
brcm_group_hash作为全局变量存在，以hash表形式组织，brcm_group被插入到brcm_group_hash中，brcm_group存储了它与下层设备的关系(eth与brcm)，real_dev指向e下层设备，而brcm设备则存储在brcm_devices_array数组中。
下面完成由下层设备转换成brcm设备的函数，brcm_port是报头中的值，可以自己设定它的含义，这里设定它表示报文来自于哪个端口。
structnet_device *find_brcm_dev(structnet_device *real_dev, u16 brcm_port)
{
structbrcm_group *grp = brcm_find_group(real_dev);
if(grp)
brcm_dev = grp->brcm_devices_array[brcm_port];
returnNULL;
}struct net_device *find_brcm_dev(struct net_device *real_dev, u16 brcm_port){ struct brcm_group *grp = brcm_find_group(real_dev); if (grp) brcm_dev = grp->brcm_devices_array[brcm_port]; return NULL;}
因为在接收报文时，报文到达brcm层开始处理时，skb->dev指向的仍是下层设备，这时通过skb->dev查到brcm_group->real_dev相匹配的hash项，然后通过报文brcm报头的信息，确定brcm_group->brcm_devices_array中哪个brcm设备作为skb的新设备；
而在发送报文时，报文到达brcm层开始处理时，skb->dev指向的是brcm设备，为了继续向下传递，需要变更为它的下层设备，在设备数据net_device的私有数据部分，一般会存储一个指针，指向它的下层设备，因此skb->dev只要变更为brcm_dev_info(dev)->real_dev。
流量统计在数据结构中，brcm设备的私有数据brcm_dev_info中brcm_rx_stats记录接收的流量信息；而dev->_tx[index]则会记录发送的流量信息。
在接收函数brcm_skb_rcv()中对于成功接收的报文会增加流量统计：
rx_stats = per_cpu_ptr(brcm_dev_info(skb->dev)->brcm_rx_stats,
smp_processor_id());
rx_stats->rx_packets++;
rx_stats->rx_bytes += skb->len;rx_stats = per_cpu_ptr(brcm_dev_info(skb->dev)->brcm_rx_stats, smp_processor_id());rx_stats->rx_packets++;rx_stats->rx_bytes += skb->len;
在发送函数brcm_dev_hard_start_xmit()中对于发送的报文会增加相应流量统计：
if(likely(ret == NET_XMIT_SUCCESS)) {
txq->tx_packets++;
txq->tx_bytes += len;
}else
txq->tx_dropped++;if (likely(ret == NET_XMIT_SUCCESS)) {txq->tx_packets++;txq->tx_bytes += len;} else txq->tx_dropped++;
而brcm_netdev_ops->ndo_get_stats()即brcm_dev_get_stats()函数，则会将brcm网卡设备中记录的发送和接收流量信息汇总成通用的格式net_device_stats，像ifconfig等命令使用的就是net_device_stats转换后的结果。
完整收发函数有了这些后接收函数brcm_skb_recv()就可以完整了，其中关于报头brcm_hdr的处理可以略过，由于是空想的协议，含义是可以自己设定的：
intbrcm_skb_recv(structsk_buff *skb,structnet_device *dev,
structpacket_type *ptype,structnet_device *orig_dev)
{
structbrcm_hdr *bhdr;
structbrcm_rx_stats *rx_stats;
intop, brcm_port;
skb = skb_share_check(skb, GFP_ATOMIC);
if(!skb)
gotoerr_free;
bhdr = (structbrcm_hdr *)skb->data;
op = bhdr->brcm_tag.brcm_53242_op;
brcm_port = bhdr->brcm_tag.brcm_53242_src_portid- 23;
rcu_read_lock();
// drop wrong brcm tag packet
if(op != BRCM_RCV_OP || brcm_port
|| brcm_port > 27)
gotoerr_unlock;
skb->dev = find_brcm_dev(dev, brcm_port);
if(!skb->dev) {
gotoerr_unlock;
}
rx_stats = per_cpu_ptr(brcm_dev_info(skb->dev)->brcm_rx_stats,
smp_processor_id());
rx_stats->rx_packets++;
rx_stats->rx_bytes += skb->len;
skb_pull_rcsum(skb, BRCM_HLEN);
switch(skb->pkt_type) {
casePACKET_BROADCAST:/* Yeah, stats collect these together.. */
/* stats->broadcast ++; // no such counter :-( */
break;
casePACKET_MULTICAST:
rx_stats->multicast++;
break;
casePACKET_OTHERHOST:
/* Our lower layer thinks this is not local, let's make sure.
* This allows the VLAN to have a different MAC than the
* underlying device, and still route correctly.
*/
if(!compare_ether_addr(eth_hdr(skb)->h_dest,
skb->dev->dev_addr))
skb->pkt_type = PACKET_HOST;
break;
default:
break;
}
// set protocol
skb->protocol = bhdr->brcm_encapsulated_proto;
// reorder skb
skb = brcm_check_reorder_header(skb);
if(!skb) {
rx_stats->rx_errors++;
gotoerr_unlock;
}
netif_rx(skb);
rcu_read_unlock();
returnNET_RX_SUCCESS;
err_unlock:
rcu_read_unlock();
err_free:
kfree_skb(skb);
returnNET_RX_DROP;
}int brcm_skb_recv(struct sk_buff *skb, struct net_device *dev, struct packet_type *ptype, struct net_device *orig_dev){ struct brcm_hdr *bhdr; struct brcm_rx_stats *rx_stats; int op, brcm_port; skb = skb_share_check(skb, GFP_ATOMIC); if(!skb) goto err_free; bhdr = (struct brcm_hdr *)skb->data; op = bhdr->brcm_tag.brcm_53242_op; brcm_port = bhdr->brcm_tag.brcm_53242_src_portid- 23; rcu_read_lock(); // drop wrong brcm tag packet if (op != BRCM_RCV_OP || brcm_port 27) goto err_unlock; skb->dev = find_brcm_dev(dev, brcm_port); if (!skb->dev) { goto err_unlock; } rx_stats = per_cpu_ptr(brcm_dev_info(skb->dev)->brcm_rx_stats, smp_processor_id()); rx_stats->rx_packets++; rx_stats->rx_bytes += skb->len; skb_pull_rcsum(skb, BRCM_HLEN); switch (skb->pkt_type) { case PACKET_BROADCAST: /* Yeah, stats collect these together.. */ /* stats->broadcast ++; // no such counter :-( */ break; case PACKET_MULTICAST: rx_stats->multicast++; break; case PACKET_OTHERHOST: /* Our lower layer thinks this is not local, let's make sure. * This allows the VLAN to have a different MAC than the * underlying device, and still route correctly. */ if (!compare_ether_addr(eth_hdr(skb)->h_dest, skb->dev->dev_addr)) skb->pkt_type = PACKET_HOST; break; default: break; } // set protocol skb->protocol = bhdr->brcm_encapsulated_proto; // reorder skb skb = brcm_check_reorder_header(skb); if (!skb) { rx_stats->rx_errors++; goto err_unlock; } netif_rx(skb); rcu_read_unlock(); return NET_RX_SUCCESS;err_unlock: rcu_read_unlock();err_free: kfree_skb(skb); return NET_RX_DROP;}
同时，发送函数brcm_dev_hard_start_xmit()可以完整了，同样，其中关于brcm_hdr的处理可以略过：
staticnetdev_tx_t brcm_dev_hard_start_xmit(structsk_buff *skb,
structnet_device *dev)
{
inti = skb_get_queue_mapping(skb);
structnetdev_queue *txq = netdev_get_tx_queue(dev, i);
structbrcm_ethhdr *beth = (structbrcm_ethhdr *)(skb->data);
unsignedintlen;
u16 brcm_port;
intret;
/* Handle non-VLAN frames if they are sent to us, for example by DHCP.
*
* NOTE: THIS ASSUMES DIX ETHERNET, SPECIFICALLY NOT SUPPORTING
* OTHER THINGS LIKE FDDI/TokenRing/802.3 SNAPs...
*/
if(beth->h_brcm_proto != htons(ETH_P_BRCM)){
//unsigned int orig_headroom = skb_headroom(skb);
brcm_t brcm_tag;
brcm_port = brcm_dev_info(dev)->brcm_port;
if(brcm_port == BRCM_ANY_PORT) {
brcm_tag.brcm_op_53242 = 0;
brcm_tag.brcm_tq_53242 = 0;
brcm_tag.brcm_te_53242 = 0;
brcm_tag.brcm_dst_53242 = 0;
}else{
brcm_tag.brcm_op_53242 = BRCM_SND_OP;
brcm_tag.brcm_tq_53242 = 0;
brcm_tag.brcm_te_53242 = 0;
brcm_tag.brcm_dst_53242 = brcm_port + 23;
}
skb = brcm_put_tag(skb, *(u32 *)(&brcm_tag));
if(!skb) {
txq->tx_dropped++;
returnNETDEV_TX_OK;
}
}
skb_set_dev(skb, brcm_dev_info(dev)->real_dev);
len = skb->len;
ret = dev_queue_xmit(skb);
if(likely(ret == NET_XMIT_SUCCESS)) {
txq->tx_packets++;
txq->tx_bytes += len;
}else
txq->tx_dropped++;
returnret;
}static netdev_tx_t brcm_dev_hard_start_xmit(struct sk_buff *skb, struct net_device *dev){ int i = skb_get_queue_mapping(skb); struct netdev_queue *txq = netdev_get_tx_queue(dev, i); struct brcm_ethhdr *beth = (struct brcm_ethhdr *)(skb->data); unsigned int len; u16 brcm_port; int ret; /* Handle non-VLAN frames if they are sent to us, for example by DHCP. * * NOTE: THIS ASSUMES DIX ETHERNET, SPECIFICALLY NOT SUPPORTING * OTHER THINGS LIKE FDDI/TokenRing/802.3 SNAPs... */ if (beth->h_brcm_proto != htons(ETH_P_BRCM)){ //unsigned int orig_headroom = skb_headroom(skb); brcm_t brcm_tag; brcm_port = brcm_dev_info(dev)->brcm_port; if (brcm_port == BRCM_ANY_PORT) { brcm_tag.brcm_op_53242 = 0; brcm_tag.brcm_tq_53242 = 0; brcm_tag.brcm_te_53242 = 0; brcm_tag.brcm_dst_53242 = 0; }else { brcm_tag.brcm_op_53242 = BRCM_SND_OP; brcm_tag.brcm_tq_53242 = 0; brcm_tag.brcm_te_53242 = 0; brcm_tag.brcm_dst_53242 = brcm_port + 23; } skb = brcm_put_tag(skb, *(u32 *)(&brcm_tag)); if (!skb) { txq->tx_dropped++; return NETDEV_TX_OK; } } skb_set_dev(skb, brcm_dev_info(dev)->real_dev); len = skb->len; ret = dev_queue_xmit(skb); if (likely(ret == NET_XMIT_SUCCESS)) { txq->tx_packets++; txq->tx_bytes += len; } else txq->tx_dropped++; return ret;}
注册设备接收通过dev_add_pack()，就可以融入协议栈了，前面几篇的分析已经讲过通过ptype_base对报文进行脱壳。现在要融入的发送，函数已经完成了，既然发送是一种嵌套的调用，并且是由dev来推过的，那么发送函数的融入一定在设备进行注册时，作为设备的一种发送方法。
创建一种设备时，一定会有设备的XXX_setup()初始化，大部分设备都会用ether_setup()来作初始化，再进行适当更改。下面是brcm_setup()：
voidbrcm_setup(structnet_device *dev)
{
ether_setup(dev);
dev->priv_flags |= IFF_BRCM_TAG;
dev->priv_flags &= ~IFF_XMIT_DST_RELEASE;
dev->tx_queue_len = 0;
dev->netdev_ops = &brcm_netdev_ops;
dev->destructor = free_netdev;
dev->ethtool_ops = &brcm_ethtool_ops;
memset(dev->broadcast, 0, ETH_ALEN);
}void brcm_setup(struct net_device *dev){ ether_setup(dev); dev->priv_flags |= IFF_BRCM_TAG; dev->priv_flags &= ~IFF_XMIT_DST_RELEASE; dev->tx_queue_len = 0; dev->netdev_ops = &brcm_netdev_ops; dev->destructor = free_netdev; dev->ethtool_ops = &brcm_ethtool_ops; memset(dev->broadcast, 0, ETH_ALEN);}
其中发送函数就在brcm_netdev_ops中，每层设备都会这样调用：dev->netdev_ops->ndo_start_xmit()。
staticconststructnet_device_ops brcm_netdev_ops = {
.ndo_change_mtu = brcm_dev_change_mtu,
.ndo_init = brcm_dev_init,
.ndo_uninit = brcm_dev_uninit,
.ndo_open = brcm_dev_open,
.ndo_stop = brcm_dev_stop,
.ndo_start_xmit = brcm_dev_hard_start_xmit,
.ndo_validate_addr = eth_validate_addr,
.ndo_set_mac_address = brcm_dev_set_mac_address,
.ndo_set_rx_mode = brcm_dev_set_rx_mode,
.ndo_set_multicast_list = brcm_dev_set_rx_mode,
.ndo_change_rx_flags = brcm_dev_change_rx_flags,
//.ndo_do_ioctl = brcm_dev_ioctl,
.ndo_neigh_setup = brcm_dev_neigh_setup,
.ndo_get_stats = brcm_dev_get_stats,
};static const struct net_device_ops brcm_netdev_ops = {.ndo_change_mtu= brcm_dev_change_mtu,.ndo_init= brcm_dev_init,.ndo_uninit= brcm_dev_uninit,.ndo_open= brcm_dev_open,.ndo_stop= brcm_dev_stop,.ndo_start_xmit = brcm_dev_hard_start_xmit,.ndo_validate_addr= eth_validate_addr,.ndo_set_mac_address= brcm_dev_set_mac_address,.ndo_set_rx_mode= brcm_dev_set_rx_mode,.ndo_set_multicast_list= brcm_dev_set_rx_mode,.ndo_change_rx_flags= brcm_dev_change_rx_flags,//.ndo_do_ioctl= brcm_dev_ioctl,.ndo_neigh_setup= brcm_dev_neigh_setup,.ndo_get_stats= brcm_dev_get_stats,};
而设备的初始化应该发生在创建设备时，也就是向网络注册它时，也就是register_brcm_dev()，注册一个新设备，需要知道它的下层设备real_dev以及唯一标识brcm设备的brcm_port。首先确定该设备没有被创建，然后用alloc_netdev_mq创建新设备new_dev，然后设置相关属性，特别是它的私有属性brcm_dev_info(new_dev)，然后添加它到brcm_group_hash中，最后发生真正的注册register_netdevice()。
staticintregister_brcm_dev(structnet_device *real_dev, u16 brcm_port)
{
structnet_device *new_dev;
structnet *net = dev_net(real_dev);
structbrcm_group *grp;
charname[IFNAMSIZ];
interr;
if(brcm_port >= BRCM_PORT_MASK)
return-ERANGE;
// exist yet
if(find_brcm_dev(real_dev, brcm_port) != NULL)
return-EEXIST;
snprintf(name, IFNAMSIZ,"brcm%i", brcm_port);
new_dev = alloc_netdev_mq(sizeof(structbrcm_dev_info), name,
brcm_setup, 1);
if(new_dev == NULL)
return-ENOBUFS;
new_dev->real_num_tx_queues = real_dev->real_num_tx_queues;
dev_net_set(new_dev, net);
new_dev->mtu = real_dev->mtu;
brcm_dev_info(new_dev)->brcm_port = brcm_port;
brcm_dev_info(new_dev)->real_dev = real_dev;
brcm_dev_info(new_dev)->dent = NULL;
//new_dev->rtnl_link_ops = &brcm_link_ops;
grp = brcm_find_group(real_dev);
if(!grp)
grp = brcm_group_alloc(real_dev);
err = register_netdevice(new_dev);
if(err
gotoout_free_newdev;
/* Account for reference in struct vlan_dev_info */
dev_hold(real_dev);
brcm_group_set_device(grp, brcm_port, new_dev);
return0;
out_free_newdev:
free_netdev(new_dev);
returnerr;
}static int register_brcm_dev(struct net_device *real_dev, u16 brcm_port){ struct net_device *new_dev; struct net *net = dev_net(real_dev); struct brcm_group *grp; char name[IFNAMSIZ]; int err; if(brcm_port >= BRCM_PORT_MASK) return -ERANGE; // exist yet if (find_brcm_dev(real_dev, brcm_port) != NULL) return -EEXIST; snprintf(name, IFNAMSIZ, "brcm%i", brcm_port); new_dev = alloc_netdev_mq(sizeof(struct brcm_dev_info), name, brcm_setup, 1); if (new_dev == NULL) return -ENOBUFS; new_dev->real_num_tx_queues = real_dev->real_num_tx_queues; dev_net_set(new_dev, net); new_dev->mtu = real_dev->mtu; brcm_dev_info(new_dev)->brcm_port = brcm_port; brcm_dev_info(new_dev)->real_dev = real_dev; brcm_dev_info(new_dev)->dent = NULL; //new_dev->rtnl_link_ops = &brcm_link_ops; grp = brcm_find_group(real_dev); if (!grp) grp = brcm_group_alloc(real_dev); err = register_netdevice(new_dev); if (err
ioctl由于brcm设备可以存在多个，并且和下层设备不是固定的对应关系，因此它的创建应该可以人为控制，因此通过ioctl由用户进行创建。这里只为brcm提供了两种操作-添加与删除。一种设备添加一定是与下层设备成关系的，因此添加时需要手动指明这种下层设备，然后通过__dev_get_by_name()从网络空间中找到这种设备，就可以调用register_brcm_dev()来完成注册了。而设备的删除则是直接删除，直接删除unregister_brcm_dev()。
staticintbrcm_ioctl_handler(structnet *net,void__user *arg)
{
interr;
structbrcm_ioctl_args args;
structnet_device *dev = NULL;
if(copy_from_user(&args, arg,sizeof(structbrcm_ioctl_args)))
return-EFAULT;
/* Null terminate this sucker, just in case. */
args.device1[23] = 0;
args.u.device2[23] = 0;
rtnl_lock();
switch(args.cmd) {
caseADD_BRCM_CMD:
caseDEL_BRCM_CMD:
err = -ENODEV;
dev = __dev_get_by_name(net, args.device1);
if(!dev)
gotoout;
err = -EINVAL;
if(args.cmd != ADD_BRCM_CMD && !is_brcm_dev(dev))
gotoout;
}
switch(args.cmd) {
caseADD_BRCM_CMD:
err = -EPERM;
if(!capable(CAP_NET_ADMIN))
break;
err = register_brcm_dev(dev, args.u.port);
break;
caseDEL_BRCM_CMD:
err = -EPERM;
if(!capable(CAP_NET_ADMIN))
break;
unregister_brcm_dev(dev, NULL);
err = 0;
break;
default:
err = -EOPNOTSUPP;
break;
}
out:
rtnl_unlock();
returnerr;
}static int brcm_ioctl_handler(struct net *net, void __user *arg){int err;struct brcm_ioctl_args args;struct net_device *dev = NULL;if (copy_from_user(&args, arg, sizeof(struct brcm_ioctl_args)))return -EFAULT;/* Null terminate this sucker, just in case. */args.device1[23] = 0;args.u.device2[23] = 0;rtnl_lock();switch (args.cmd) {case ADD_BRCM_CMD:case DEL_BRCM_CMD:err = -ENODEV;dev = __dev_get_by_name(net, args.device1);if (!dev)goto out;err = -EINVAL;if (args.cmd != ADD_BRCM_CMD && !is_brcm_dev(dev))goto out;}switch (args.cmd) {case ADD_BRCM_CMD:err = -EPERM;if (!capable(CAP_NET_ADMIN))break;err = register_brcm_dev(dev, args.u.port);break;case DEL_BRCM_CMD:err = -EPERM;if (!capable(CAP_NET_ADMIN))break;unregister_brcm_dev(dev, NULL);err = 0;break;default:err = -EOPNOTSUPP;break;}out:rtnl_unlock();return err;}
这些是brcm协议模块的主体部分了，当然它还不完整，在下篇中继续完成brcm协议的添加，为它完善一些细节：proc文件系统， notifier机制等等，以及内核Makefile的编写，当然还有协议的测试。
为了用户方便查看brcm设备的工作状态，使用proc文件系统是很好的方式。一个网络协议模块可以注册到网络空间中register_pernet_subsys()，这个函数会为子空间分配一个id号，通过id可以在网络空间中找到分配给该子空间的内存：init_net->gen->ptr[id - 1]。而我们正是利用这块内存去存储proc中的相关信息：struct brcm_net，它记录了brcm设备在proc文件系统中的位置。
structbrcm_net {
/* /proc/net/brcm */
structproc_dir_entry *proc_brcm_dir;
/* /proc/net/brcm/config */
structproc_dir_entry *proc_brcm_conf;
};struct brcm_net {/* /proc/net/brcm */struct proc_dir_entry *proc_brcm_dir;/* /proc/net/brcm/config */struct proc_dir_entry *proc_brcm_conf;};
在加载brcm模块时会注册子空间，brcm_init_net创建在proc中的相关项，并记录路径在brcm_net中；brcm_exit_net删除在proc中的相关项；brcm_net_id记录分配给子空间的id，这样通过init_net->gen->ptr[brcm_net_id - 1]就可以操作brcm_net了。
err = register_pernet_subsys(&brcm_net_ops);
staticstructpernet_operations brcm_net_ops = {
.init = brcm_init_net,
.exit = brcm_exit_net,
.id = &brcm_net_id,
.size =sizeof(structbrcm_net),
};err = register_pernet_subsys(&brcm_net_ops);static struct pernet_operations brcm_net_ops = {.init = brcm_init_net,.exit = brcm_exit_net,.id = &brcm_net_id,.size = sizeof(struct brcm_net),};
注意到在brcm_init_net和brcm_exit_net中添加和删除的仅仅是/proc/net/brcm目录和config文件，而在使用中brcm设备是可以动态创建的，因此这部分代码应该发生在添加和删除brcm设备时，而不是在brcm模块注册和删除时。最简单的是直接添加在register方法或unregister方法中，但内核提供了更好的机制：事件，将对proc的操作分离出来，因为proc的操作实际上属于附加的操作而不是必须的操作。下面就来看event机制。
前面几篇已经有描述过event机制，这里的事件都是关于设备的事件，使用的是register_netdevice_notifier()来，注册notifier到netdev_chain链表上。在加载brcm模块时注册notifier，brcm_notifier_block包含了brcm设备对所关心的事件作出的反应。
err = register_netdevice_notifier(&brcm_notifier_block);
staticstructnotifier_block brcm_notifier_block __read_mostly = {
.notifier_call = brcm_device_event,
};err = register_netdevice_notifier(&brcm_notifier_block);static struct notifier_block brcm_notifier_block __read_mostly = {.notifier_call = brcm_device_event,};
设备对哪些事件会感兴趣，首先，brcm设备对注册和注销是感兴趣的，要操作proc文件系统；其次，对于发往brcm下层设备的事件，要考虑这些事件造成的连带影响(比如eth1被down掉，则其上的brcm设备也应该被down掉)，一般是下层设备的事件对其上的所有brcm设备都进行相应操作。
所以在brcm_device_event有两类进入的设备dev会进行操作，一类是brcm设备，它仅仅是操作proc文件系统。判断是否为brcm设备，是的话则由__brcm_device_event() 处理。
if(is_brcm_dev(dev))
__brcm_device_event(dev, event);
staticvoid__brcm_device_event(structnet_device *dev, unsignedlongevent)
{
switch(event) {
caseNETDEV_CHANGENAME:
brcm_proc_rem_dev(dev);
if(brcm_proc_add_dev(dev)
pr_warning("BRCM: failed to change proc name for %s\n",
dev->name);
break;
caseNETDEV_REGISTER:
if(brcm_proc_add_dev(dev)
pr_warning("BRCM: failed to add proc entry for %s\n",
dev->name);
break;
caseNETDEV_UNREGISTER:
brcm_proc_rem_dev(dev);
break;
}
}if (is_brcm_dev(dev))__brcm_device_event(dev, event);static void __brcm_device_event(struct net_device *dev, unsigned long event){switch (event) {case NETDEV_CHANGENAME:brcm_proc_rem_dev(dev);if (brcm_proc_add_dev(dev) name);break;case NETDEV_REGISTER:if (brcm_proc_add_dev(dev) name);break;case NETDEV_UNREGISTER:brcm_proc_rem_dev(dev);break;}}
如何是brcm的下层设备，如根据brcm_group_hash中的映射关系，对下层设备相关的所有brcm设备进行操作：
switch(event) {
caseNETDEV_CHANGE:
/* Propagate real device state to vlan devices */
for(i = 0; i
brcmdev = brcm_group_get_device(grp, i);
if(!brcmdev)
continue;
netif_stacked_transfer_operstate(dev, brcmdev);
}
break;
caseNETDEV_CHANGEADDR:
/* Adjust unicast filters on underlying device */
for(i = 0; i
brcmdev = brcm_group_get_device(grp, i);
if(!brcmdev)
continue;
flgs = brcmdev->flags;
if(!(flgs & IFF_UP))
continue;
brcm_sync_address(dev, brcmdev);
}
break;
caseNETDEV_CHANGEMTU:
for(i = 0; i
brcmdev = brcm_group_get_device(grp, i);
if(!brcmdev)
continue;
if(brcmdev->mtu <= dev->mtu)
continue;
dev_set_mtu(brcmdev, dev->mtu);
}
break;
caseNETDEV_DOWN:
/* Put all VLANs for this dev in the down state too. */
for(i = 0; i
brcmdev = brcm_group_get_device(grp, i);
if(!brcmdev)
continue;
flgs = brcmdev->flags;
if(!(flgs & IFF_UP))
continue;
brcm = brcm_dev_info(brcmdev);
dev_change_flags(brcmdev, flgs & ~IFF_UP);
netif_stacked_transfer_operstate(dev, brcmdev);
}
break;
caseNETDEV_UP:
/* Put all VLANs for this dev in the up state too. */
for(i = 0; i
brcmdev = brcm_group_get_device(grp, i);
if(!brcmdev)
continue;
flgs = brcmdev->flags;
if(flgs & IFF_UP)
continue;
brcm = brcm_dev_info(brcmdev);
dev_change_flags(brcmdev, flgs | IFF_UP);
netif_stacked_transfer_operstate(dev, brcmdev);
}
break;
caseNETDEV_UNREGISTER:
/* Delete all BRCMs for this dev. */
grp->killall = 1;
for(i = 0; i
brcmdev = brcm_group_get_device(grp, i);
if(!brcmdev)
continue;
/* unregistration of last brcm destroys group, abort
* afterwards */
if(grp->nr_ports == 1)
i = BRCM_GROUP_ARRAY_LEN;
unregister_brcm_dev(brcmdev, &list);
}
unregister_netdevice_many(&list);
break;
}switch (event) {case NETDEV_CHANGE:/* Propagate real device state to vlan devices */for (i = 0; i flags;if (!(flgs & IFF_UP))continue;brcm_sync_address(dev, brcmdev);}break;case NETDEV_CHANGEMTU:for (i = 0; i mtu <= dev->mtu)continue;dev_set_mtu(brcmdev, dev->mtu);}break;case NETDEV_DOWN:/* Put all VLANs for this dev in the down state too. */for (i = 0; i flags;if (!(flgs & IFF_UP))continue;brcm = brcm_dev_info(brcmdev);dev_change_flags(brcmdev, flgs & ~IFF_UP);netif_stacked_transfer_operstate(dev, brcmdev);}break;case NETDEV_UP:/* Put all VLANs for this dev in the up state too. */for (i = 0; i flags;if (flgs & IFF_UP)continue;brcm = brcm_dev_info(brcmdev);dev_change_flags(brcmdev, flgs | IFF_UP);netif_stacked_transfer_operstate(dev, brcmdev);}break;case NETDEV_UNREGISTER:/* Delete all BRCMs for this dev. */grp->killall = 1;for (i = 0; i nr_ports == 1)i = BRCM_GROUP_ARRAY_LEN;unregister_brcm_dev(brcmdev, &list);}unregister_netdevice_many(&list);break;}
到这里，协议的添加就大致完成了，当然还包括一些头文件的修改，宏变量的添加等就不一一详述，具体可见最后的附件。
为了编译进内核，还需要修改以下文件：
$(linux)/net/Kconfig $(linux)/net/Makefile
最后，在make menuconfig选择添加brcm协议
Networking Support -> Networking options
同时，需要一个简单的用户空间工具来配置我们的brcm设备，就像vconfig用来配置vlan设备一样；编写的简单的bconfig工具，命令格式：
"Usage: add [interface-name] [brcm_port]\n"
" rem [dev-name]";
内核编译完成后就该进行测试了，如果开启了内核调试信息，启动内核就看到以下信息：
然后启用网卡，可以查看到添加了brcm设备后的状态：
可以使用原生套接字自己打上brcm头后发送报文让协议栈接收，或者用wireshark等捕获协议栈发出的报文，下图即是捕获到的报文：
这是主机发出的arp报文，可以看到，在源mac后接的不是vlan报头，而是我们添加的brcm报文，协议号是8744。
查看proc中信息：
附：patch补丁 && 重要的源文件 && bconfig工具源码
以上是原作者上传的code，但CSDN下载需要1积分，CSDN的积分实在没有了，我托人下来的，并在51CTO上分享了一下，这里的下载豆很好弄到。
1823
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
