linux网卡驱动注册与接受数据处理

Linux网络接受

本文基于2.6.15的内核版本来简单学习一下有关Linux内核如何进行网络数据的处理流程。

网卡驱动注册

以ixgb驱动为例，在网卡注册的时候会进入如下流程。

static struct pci_driver ixgb_driver = {
	.name     = ixgb_driver_name,
	.id_table = ixgb_pci_tbl,
	.probe    = ixgb_probe,
	.remove   = __devexit_p(ixgb_remove),
};

...
/**
 * ixgb_init_module - Driver Registration Routine
 *
 * ixgb_init_module is the first routine called when the driver is
 * loaded. All it does is register with the PCI subsystem.
 **/

static int __init
ixgb_init_module(void)
{
	printk(KERN_INFO "%s - version %s\n",
	       ixgb_driver_string, ixgb_driver_version);

	printk(KERN_INFO "%s\n", ixgb_copyright);

	return pci_module_init(&ixgb_driver);
}

module_init(ixgb_init_module);
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

PCI注册的流程会将调用probe方法，即ixgb_probe。

/**
 * ixgb_probe - Device Initialization Routine
 * @pdev: PCI device information struct
 * @ent: entry in ixgb_pci_tbl
 *
 * Returns 0 on success, negative on failure
 *
 * ixgb_probe initializes an adapter identified by a pci_dev structure.
 * The OS initialization, configuring of the adapter private structure,
 * and a hardware reset occur.
 **/

static int __devinit
ixgb_probe(struct pci_dev *pdev,
		const struct pci_device_id *ent)
{
	struct net_device *netdev = NULL;
	struct ixgb_adapter *adapter;
	static int cards_found = 0;
	unsigned long mmio_start;
	int mmio_len;
	int pci_using_dac;
	int i;
	int err;

	if((err = pci_enable_device(pdev)))  // 是否可以注册该设备
		return err;

	if(!(err = pci_set_dma_mask(pdev, DMA_64BIT_MASK))) { //设置DMA的掩码
		pci_using_dac = 1;
	} else {
		if((err = pci_set_dma_mask(pdev, DMA_32BIT_MASK))) {
			IXGB_ERR("No usable DMA configuration, aborting\n");
			return err;
		}
		pci_using_dac = 0;
	}

	if((err = pci_request_regions(pdev, ixgb_driver_name)))
		return err;

	pci_set_master(pdev);

	netdev = alloc_etherdev(sizeof(struct ixgb_adapter)); // 申请设备结构体
	if(!netdev) {
		err = -ENOMEM;
		goto err_alloc_etherdev;
	}

	SET_MODULE_OWNER(netdev);
	SET_NETDEV_DEV(netdev, &pdev->dev);

	pci_set_drvdata(pdev, netdev);    // 设置DMA对应的管理的内存空间
	adapter = netdev_priv(netdev);
	adapter->netdev = netdev;
	adapter->pdev = pdev;
	adapter->hw.back = adapter;

	mmio_start = pci_resource_start(pdev, BAR_0);
	mmio_len = pci_resource_len(pdev, BAR_0);

	adapter->hw.hw_addr = ioremap(mmio_start, mmio_len);
	if(!adapter->hw.hw_addr) {
		err = -EIO;
		goto err_ioremap;
	}

	for(i = BAR_1; i <= BAR_5; i++) {
		if(pci_resource_len(pdev, i) == 0)
			continue;
		if(pci_resource_flags(pdev, i) & IORESOURCE_IO) {
			adapter->hw.io_base = pci_resource_start(pdev, i);
			break;
		}
	}

	netdev->open = &ixgb_open;  // 设置ethtool对应的操作的方法，即设备的打开的方法
	netdev->stop = &ixgb_close;   // 设备的关闭的方法
	netdev->hard_start_xmit = &ixgb_xmit_frame;
	netdev->get_stats = &ixgb_get_stats;
	netdev->set_multicast_list = &ixgb_set_multi;
	netdev->set_mac_address = &ixgb_set_mac;
	netdev->change_mtu = &ixgb_change_mtu;  // 设置mtu长度
	ixgb_set_ethtool_ops(netdev);
	netdev->tx_timeout = &ixgb_tx_timeout;  // 设置超时时间
	netdev->watchdog_timeo = HZ;
#ifdef CONFIG_IXGB_NAPI
	netdev->poll = &ixgb_clean;    // 设置设备的poll方法 一般用在NAPI模式中
	netdev->weight = 64;
#endif
	netdev->vlan_rx_register = ixgb_vlan_rx_register;  // 设置vlan的方法
	netdev->vlan_rx_add_vid = ixgb_vlan_rx_add_vid;
	netdev->vlan_rx_kill_vid = ixgb_vlan_rx_kill_vid;
#ifdef CONFIG_NET_POLL_CONTROLLER
	netdev->poll_controller = ixgb_netpoll;  //是否设置了netpoll的方式
#endif

	netdev->mem_start = mmio_start;
	netdev->mem_end = mmio_start + mmio_len;
	netdev->base_addr = adapter->hw.io_base;

	adapter->bd_number = cards_found;
	adapter->link_speed = 0;
	adapter->link_duplex = 0;

	/* setup the private structure */

	if((err = ixgb_sw_init(adapter)))
		goto err_sw_init;

	netdev->features = NETIF_F_SG |
			   NETIF_F_HW_CSUM |
			   NETIF_F_HW_VLAN_TX |
			   NETIF_F_HW_VLAN_RX |
			   NETIF_F_HW_VLAN_FILTER;
#ifdef NETIF_F_TSO
	netdev->features |= NETIF_F_TSO;
#endif

	if(pci_using_dac)
		netdev->features |= NETIF_F_HIGHDMA;

	/* make sure the EEPROM is good */

	if(!ixgb_validate_eeprom_checksum(&adapter->hw)) {
		printk(KERN_ERR "The EEPROM Checksum Is Not Valid\n");
		err = -EIO;
		goto err_eeprom;
	}

	ixgb_get_ee_mac_addr(&adapter->hw, netdev->dev_addr);
	memcpy(netdev->perm_addr, netdev->dev_addr, netdev->addr_len);

	if(!is_valid_ether_addr(netdev->perm_addr)) {
		err = -EIO;
		goto err_eeprom;
	}

	adapter->part_num = ixgb_get_ee_pba_number(&adapter->hw);

	init_timer(&adapter->watchdog_timer);   // 设置定时器
	adapter->watchdog_timer.function = &ixgb_watchdog;
	adapter->watchdog_timer.data = (unsigned long)adapter;

	INIT_WORK(&adapter->tx_timeout_task,
		  (void (*)(void *))ixgb_tx_timeout_task, netdev);

	if((err = register_netdev(netdev)))  // 注册网络设备
		goto err_register;

	/* we're going to reset, so assume we have no link for now */

	netif_carrier_off(netdev);     // 重置网络设备的参数
	netif_stop_queue(netdev);

	printk(KERN_INFO "%s: Intel(R) PRO/10GbE Network Connection\n",
		   netdev->name);
	ixgb_check_options(adapter);
	/* reset the hardware with the new settings */

	ixgb_reset(adapter);

	cards_found++;
	return 0;

err_register:
err_sw_init:
err_eeprom:
	iounmap(adapter->hw.hw_addr);
err_ioremap:
	free_netdev(netdev);
err_alloc_etherdev:
	pci_release_regions(pdev);
	return err;
}
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

注册的方式主要就是通过配置一些网络的参数。

打开网卡接受数据

如果我们使用ethtool工具是该网卡打开，此时就会执行如下流程。

static int
ixgb_open(struct net_device *netdev)
{
	struct ixgb_adapter *adapter = netdev_priv(netdev);
	int err;

	/* allocate transmit descriptors */

	if((err = ixgb_setup_tx_resources(adapter)))
		goto err_setup_tx;

	/* allocate receive descriptors */

	if((err = ixgb_setup_rx_resources(adapter)))
		goto err_setup_rx;

	if((err = ixgb_up(adapter)))   // 注册中断函数，响应网卡数据
		goto err_up;

	return 0;

err_up:
	ixgb_free_rx_resources(adapter);
err_setup_rx:
	ixgb_free_tx_resources(adapter);
err_setup_tx:
	ixgb_reset(adapter);

	return err;
}
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

此时就会调用ixgb_open的函数，进行发送和接受的资源设置，并通过ixgb_up注册网卡的中断回调函数。

int
ixgb_up(struct ixgb_adapter *adapter)
{
	struct net_device *netdev = adapter->netdev;
	int err;
	int max_frame = netdev->mtu + ENET_HEADER_SIZE + ENET_FCS_LENGTH;  // 获取帧大小
	struct ixgb_hw *hw = &adapter->hw;

	/* hardware has been reset, we need to reload some things */

	ixgb_set_multi(netdev);

	ixgb_restore_vlan(adapter);

	ixgb_configure_tx(adapter);  // 配置发送
	ixgb_setup_rctl(adapter);
	ixgb_configure_rx(adapter);  // 配置接受
	ixgb_alloc_rx_buffers(adapter);  // 设置接受ring缓冲区

#ifdef CONFIG_PCI_MSI
	{
	boolean_t pcix = (IXGB_READ_REG(&adapter->hw, STATUS) & 
						  IXGB_STATUS_PCIX_MODE) ? TRUE : FALSE;
	adapter->have_msi = TRUE;

	if (!pcix)
	   adapter->have_msi = FALSE;
	else if((err = pci_enable_msi(adapter->pdev))) {
		printk (KERN_ERR
		 "Unable to allocate MSI interrupt Error: %d\n", err);
		adapter->have_msi = FALSE;
		/* proceed to try to request regular interrupt */
	}
	}

#endif
	if((err = request_irq(adapter->pdev->irq, &ixgb_intr,
				  SA_SHIRQ | SA_SAMPLE_RANDOM,
				  netdev->name, netdev)))  // 注册网卡的中断处理函数，并注册网卡的中断号
		return err;

	/* disable interrupts and get the hardware into a known state */
	IXGB_WRITE_REG(&adapter->hw, IMC, 0xffffffff);

	if((hw->max_frame_size != max_frame) ||
		(hw->max_frame_size !=
		(IXGB_READ_REG(hw, MFS) >> IXGB_MFS_SHIFT))) {

		hw->max_frame_size = max_frame;

		IXGB_WRITE_REG(hw, MFS, hw->max_frame_size << IXGB_MFS_SHIFT);

		if(hw->max_frame_size >
		   IXGB_MAX_ENET_FRAME_SIZE_WITHOUT_FCS + ENET_FCS_LENGTH) {
			uint32_t ctrl0 = IXGB_READ_REG(hw, CTRL0);

			if(!(ctrl0 & IXGB_CTRL0_JFE)) {
				ctrl0 |= IXGB_CTRL0_JFE;
				IXGB_WRITE_REG(hw, CTRL0, ctrl0);
			}
		}
	}

	mod_timer(&adapter->watchdog_timer, jiffies);
	ixgb_irq_enable(adapter);   // 开启中断处理

#ifdef CONFIG_IXGB_NAPI
	netif_poll_enable(netdev);
#endif
	return 0;
}
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

主要是设置了接受的数据缓冲区，设置了中断的处理函数ixgb_intr，当数据来了之后就会调用ixgb_intr函数进行处理。

static irqreturn_t
ixgb_intr(int irq, void *data, struct pt_regs *regs)
{
	struct net_device *netdev = data;
	struct ixgb_adapter *adapter = netdev_priv(netdev);
	struct ixgb_hw *hw = &adapter->hw;
	uint32_t icr = IXGB_READ_REG(hw, ICR);
#ifndef CONFIG_IXGB_NAPI
	unsigned int i;
#endif

	if(unlikely(!icr))
		return IRQ_NONE;  /* Not our interrupt */

	if(unlikely(icr & (IXGB_INT_RXSEQ | IXGB_INT_LSC))) {
		mod_timer(&adapter->watchdog_timer, jiffies);
	}

#ifdef CONFIG_IXGB_NAPI
	if(netif_rx_schedule_prep(netdev)) {  // 是否是NAPI模式运行

		/* Disable interrupts and register for poll. The flush 
		  of the posted write is intentionally left out.
		*/

		atomic_inc(&adapter->irq_sem);
		IXGB_WRITE_REG(&adapter->hw, IMC, ~0);
		__netif_rx_schedule(netdev);
	}
#else
	/* yes, that is actually a & and it is meant to make sure that
	 * every pass through this for loop checks both receive and
	 * transmit queues for completed descriptors, intended to
	 * avoid starvation issues and assist tx/rx fairness. */
	for(i = 0; i < IXGB_MAX_INTR; i++)  // 通过ixgb_clean_rx_irq进行数据处理
		if(!ixgb_clean_rx_irq(adapter) &
		   !ixgb_clean_tx_irq(adapter))
			break;
#endif 
	return IRQ_HANDLED;
}
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

此时就看配置的是网卡的哪种数据处理类型，一般有如下两种网卡数据处理流程。

NAPI模式

通过softirq内核线程来进行网卡数据的进行poll进行，从而避免过多的中断来进行处理提升效率。

static inline void __netif_rx_schedule(struct net_device *dev)
{
	unsigned long flags;

	local_irq_save(flags);
	dev_hold(dev);
	list_add_tail(&dev->poll_list, &__get_cpu_var(softnet_data).poll_list); // 加入当前设备的队列从而使在软中断中内容持续接受网卡的数据
	if (dev->quota < 0)
		dev->quota += dev->weight;
	else
		dev->quota = dev->weight;
	__raise_softirq_irqoff(NET_RX_SOFTIRQ);  // 发出软中断，发送数据接受软中断
	local_irq_restore(flags);
}
1
2
3
4
5
6
7
8
9
10
11
12
13
14

此时就会调用在NET_RX_SOFTIRQ注册的信号方法。

open_softirq(NET_RX_SOFTIRQ, net_rx_action, NULL);
1

此时调用的就是网络处理。

static void net_rx_action(struct softirq_action *h)
{
	struct softnet_data *queue = &__get_cpu_var(softnet_data);
	unsigned long start_time = jiffies;
	int budget = netdev_budget;
	void *have;

	local_irq_disable();

	while (!list_empty(&queue->poll_list)) {  // 检查链表是否为空
		struct net_device *dev;

		if (budget <= 0 || jiffies - start_time > 1)
			goto softnet_break;

		local_irq_enable();  //使能中断

		dev = list_entry(queue->poll_list.next,
				 struct net_device, poll_list);  // 获取设备
		have = netpoll_poll_lock(dev);    

		if (dev->quota <= 0 || dev->poll(dev, &budget)) {  // 获取链表上面的数据
			netpoll_poll_unlock(have);
			local_irq_disable();
			list_del(&dev->poll_list);  
			list_add_tail(&dev->poll_list, &queue->poll_list);
			if (dev->quota < 0)
				dev->quota += dev->weight;
			else
				dev->quota = dev->weight;
		} else {
			netpoll_poll_unlock(have);
			dev_put(dev);  // 将获取数据放回
			local_irq_disable();
		}
	}
out:
	local_irq_enable();  //使能中断
	return;

softnet_break:
	__get_cpu_var(netdev_rx_stat).time_squeeze++;
	__raise_softirq_irqoff(NET_RX_SOFTIRQ);
	goto out;
}
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

通过调用dev的poll函数来进行数据的处理。ixgb对应的poll函数为ixgb_clean。

#ifdef CONFIG_IXGB_NAPI
/**
 * ixgb_clean - NAPI Rx polling callback
 * @adapter: board private structure
 **/

static int
ixgb_clean(struct net_device *netdev, int *budget)
{
	struct ixgb_adapter *adapter = netdev_priv(netdev);
	int work_to_do = min(*budget, netdev->quota);
	int tx_cleaned;
	int work_done = 0;

	tx_cleaned = ixgb_clean_tx_irq(adapter);
	ixgb_clean_rx_irq(adapter, &work_done, work_to_do); // 处理单个网络接受数据

	*budget -= work_done;
	netdev->quota -= work_done;  // 减去王超的数量

	/* if no Tx and not enough Rx work done, exit the polling mode */
	if((!tx_cleaned && (work_done == 0)) || !netif_running(netdev)) {
		netif_rx_complete(netdev);  // 如果没了就离开该模式
		ixgb_irq_enable(adapter);
		return 0;
	}

	return 1;
}
#endif
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

其中ixgb_clean_rx_irq函数，就是我们不经过NAPI模式处理的流程。

数据直接传递

数据直接传递就是直接通过ixgb_clean_rx_irq函数来处理网络包。

static boolean_t
#ifdef CONFIG_IXGB_NAPI
ixgb_clean_rx_irq(struct ixgb_adapter *adapter, int *work_done, int work_to_do)
#else
ixgb_clean_rx_irq(struct ixgb_adapter *adapter)
#endif
{
	struct ixgb_desc_ring *rx_ring = &adapter->rx_ring;
	struct net_device *netdev = adapter->netdev;
	struct pci_dev *pdev = adapter->pdev;
	struct ixgb_rx_desc *rx_desc, *next_rxd;
	struct ixgb_buffer *buffer_info, *next_buffer, *next2_buffer;
	uint32_t length;
	unsigned int i, j;
	boolean_t cleaned = FALSE;

	i = rx_ring->next_to_clean;
	rx_desc = IXGB_RX_DESC(*rx_ring, i);
	buffer_info = &rx_ring->buffer_info[i];

	while(rx_desc->status & IXGB_RX_DESC_STATUS_DD) {
		struct sk_buff *skb, *next_skb;  // skb数据包
		u8 status;

#ifdef CONFIG_IXGB_NAPI
		if(*work_done >= work_to_do)
			break;

		(*work_done)++;
#endif
		status = rx_desc->status;
		skb = buffer_info->skb;  //获取网卡处理好的skb数据

		prefetch(skb->data);

		if(++i == rx_ring->count) i = 0;
		next_rxd = IXGB_RX_DESC(*rx_ring, i);
		prefetch(next_rxd);

		if((j = i + 1) == rx_ring->count) j = 0;
		next2_buffer = &rx_ring->buffer_info[j];  // 获取ring的数据
		prefetch(next2_buffer);

		next_buffer = &rx_ring->buffer_info[i];
		next_skb = next_buffer->skb;
		prefetch(next_skb);

		cleaned = TRUE;

		pci_unmap_single(pdev,
				 buffer_info->dma,
				 buffer_info->length,
				 PCI_DMA_FROMDEVICE);

		length = le16_to_cpu(rx_desc->length);

		if(unlikely(!(status & IXGB_RX_DESC_STATUS_EOP))) {

			/* All receives must fit into a single buffer */

			IXGB_DBG("Receive packet consumed multiple buffers "
					 "length<%x>\n", length);

			dev_kfree_skb_irq(skb);
			goto rxdesc_done;
		}

		if (unlikely(rx_desc->errors
			     & (IXGB_RX_DESC_ERRORS_CE | IXGB_RX_DESC_ERRORS_SE
				| IXGB_RX_DESC_ERRORS_P |
				IXGB_RX_DESC_ERRORS_RXE))) {

			dev_kfree_skb_irq(skb);
			goto rxdesc_done;
		}

		/* Good Receive */
		skb_put(skb, length);   //放入skb中保存

		/* Receive Checksum Offload */
		ixgb_rx_checksum(adapter, rx_desc, skb); //检查checksum

		skb->protocol = eth_type_trans(skb, netdev);  //获取协议
#ifdef CONFIG_IXGB_NAPI
		if(adapter->vlgrp && (status & IXGB_RX_DESC_STATUS_VP)) {
			vlan_hwaccel_receive_skb(skb, adapter->vlgrp,
				le16_to_cpu(rx_desc->special) &
					IXGB_RX_DESC_SPECIAL_VLAN_MASK);
		} else {
			netif_receive_skb(skb);  // 如果是NAPI模式 直接调用netif_receive_skb处理
		}
#else /* CONFIG_IXGB_NAPI */
		if(adapter->vlgrp && (status & IXGB_RX_DESC_STATUS_VP)) {
			vlan_hwaccel_rx(skb, adapter->vlgrp,
				le16_to_cpu(rx_desc->special) &
					IXGB_RX_DESC_SPECIAL_VLAN_MASK);
		} else {
			netif_rx(skb);   // 走单个软中段的处理流程
		}
#endif /* CONFIG_IXGB_NAPI */
		netdev->last_rx = jiffies;

rxdesc_done:
		/* clean up descriptor, might be written over by hw */
		rx_desc->status = 0;
		buffer_info->skb = NULL;

		/* use prefetched values */
		rx_desc = next_rxd;
		buffer_info = next_buffer;
	}

	rx_ring->next_to_clean = i;   

	ixgb_alloc_rx_buffers(adapter);

	return cleaned;
}
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

在非NAPI模式下，会调用netif_rx进入到软中段的网络数据处理中，只不过该步就没有调用poll模式处理。

int netif_rx(struct sk_buff *skb)
{
	struct softnet_data *queue;
	unsigned long flags;

	/* if netpoll wants it, pretend we never saw it */
	if (netpoll_rx(skb))
		return NET_RX_DROP;

	if (!skb->tstamp.off_sec)
		net_timestamp(skb);

	/*
	 * The code is rearranged so that the path is the most
	 * short when CPU is congested, but is still operating.
	 */
	local_irq_save(flags);
	queue = &__get_cpu_var(softnet_data);

	__get_cpu_var(netdev_rx_stat).total++;
	if (queue->input_pkt_queue.qlen <= netdev_max_backlog) {
		if (queue->input_pkt_queue.qlen) {
enqueue:
			dev_hold(skb->dev);
			__skb_queue_tail(&queue->input_pkt_queue, skb);
			local_irq_restore(flags);
			return NET_RX_SUCCESS;
		}

		netif_rx_schedule(&queue->backlog_dev);  // 调用队列进行接受数据的处理
		goto enqueue;
	}

	__get_cpu_var(netdev_rx_stat).dropped++;
	local_irq_restore(flags);

	kfree_skb(skb);
	return NET_RX_DROP;
}
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

基本的两种模式下面的网络数据处理就基本分析完成。

总结

本文简单根据Linux-2.6系列的源码，基于网卡ixgb进行了一个网卡数据的简单的接受流程分析，网卡通过ethtool完成对应的接口，并注册对应的驱动，在通过ethtool工具使网卡开始运行的时候，就注册中断处理函数，并通过是否是NAPI处理流程来进行不同的处理，使用了NAPI会使用poll函数在softirq内核线程中，进行网卡数据的处理，效率上会比非NAPI模式要高，在非NAPI模式下面，最后也是通过通过softirq内核线程将获取的网络包进行处理，从而完成网卡数据的接受与处理。由于本人才疏学浅，如有错误请批评指正。