netpoll.c

linux 内核源代码

/*
 * Common framework for low-level network console, dump, and debugger code
 *
 * Sep 8 2003  Matt Mackall <[e-mail]>
 *
 * based on the netconsole code from:
 *
 * Copyright (C) 2001  Ingo Molnar <[e-mail]>
 * Copyright (C) 2002  Red Hat, Inc.
 */

#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/string.h>
#include <linux/if_arp.h>
#include <linux/inetdevice.h>
#include <linux/inet.h>
#include <linux/interrupt.h>
#include <linux/netpoll.h>
#include <linux/sched.h>
#include <linux/delay.h>
#include <linux/rcupdate.h>
#include <linux/workqueue.h>
#include <net/tcp.h>
#include <net/udp.h>
#include <asm/unaligned.h>

/*
 * We maintain a small pool of fully-sized skbs, to make sure the
 * message gets out even in extreme OOM situations.
 */

#define MAX_UDP_CHUNK 1460
#define MAX_SKBS 32
#define MAX_QUEUE_DEPTH (MAX_SKBS / 2)

static struct sk_buff_head skb_pool;

static atomic_t trapped;

#define USEC_PER_POLL	50
#define NETPOLL_RX_ENABLED  1
#define NETPOLL_RX_DROP     2

#define MAX_SKB_SIZE \
		(MAX_UDP_CHUNK + sizeof(struct udphdr) + \
				sizeof(struct iphdr) + sizeof(struct ethhdr))

static void zap_completion_queue(void);
static void arp_reply(struct sk_buff *skb);

static void queue_process(struct work_struct *work)
{
	struct netpoll_info *npinfo =
		container_of(work, struct netpoll_info, tx_work.work);
	struct sk_buff *skb;
	unsigned long flags;

	while ((skb = skb_dequeue(&npinfo->txq))) {
		struct net_device *dev = skb->dev;

		if (!netif_device_present(dev) || !netif_running(dev)) {
			__kfree_skb(skb);
			continue;
		}

		local_irq_save(flags);
		netif_tx_lock(dev);
		if ((netif_queue_stopped(dev) ||
		     netif_subqueue_stopped(dev, skb)) ||
		     dev->hard_start_xmit(skb, dev) != NETDEV_TX_OK) {
			skb_queue_head(&npinfo->txq, skb);
			netif_tx_unlock(dev);
			local_irq_restore(flags);

			schedule_delayed_work(&npinfo->tx_work, HZ/10);
			return;
		}
		netif_tx_unlock(dev);
		local_irq_restore(flags);
	}
}

static __sum16 checksum_udp(struct sk_buff *skb, struct udphdr *uh,
			    unsigned short ulen, __be32 saddr, __be32 daddr)
{
	__wsum psum;

	if (uh->check == 0 || skb_csum_unnecessary(skb))
		return 0;

	psum = csum_tcpudp_nofold(saddr, daddr, ulen, IPPROTO_UDP, 0);

	if (skb->ip_summed == CHECKSUM_COMPLETE &&
	    !csum_fold(csum_add(psum, skb->csum)))
		return 0;

	skb->csum = psum;

	return __skb_checksum_complete(skb);
}

/*
 * Check whether delayed processing was scheduled for our NIC. If so,
 * we attempt to grab the poll lock and use ->poll() to pump the card.
 * If this fails, either we've recursed in ->poll() or it's already
 * running on another CPU.
 *
 * Note: we don't mask interrupts with this lock because we're using
 * trylock here and interrupts are already disabled in the softirq
 * case. Further, we test the poll_owner to avoid recursion on UP
 * systems where the lock doesn't exist.
 *
 * In cases where there is bi-directional communications, reading only
 * one message at a time can lead to packets being dropped by the
 * network adapter, forcing superfluous retries and possibly timeouts.
 * Thus, we set our budget to greater than 1.
 */
static int poll_one_napi(struct netpoll_info *npinfo,
			 struct napi_struct *napi, int budget)
{
	int work;

	/* net_rx_action's ->poll() invocations and our's are
	 * synchronized by this test which is only made while
	 * holding the napi->poll_lock.
	 */
	if (!test_bit(NAPI_STATE_SCHED, &napi->state))
		return budget;

	npinfo->rx_flags |= NETPOLL_RX_DROP;
	atomic_inc(&trapped);

	work = napi->poll(napi, budget);

	atomic_dec(&trapped);
	npinfo->rx_flags &= ~NETPOLL_RX_DROP;

	return budget - work;
}

static void poll_napi(struct netpoll *np)
{
	struct netpoll_info *npinfo = np->dev->npinfo;
	struct napi_struct *napi;
	int budget = 16;

	list_for_each_entry(napi, &np->dev->napi_list, dev_list) {
		if (napi->poll_owner != smp_processor_id() &&
		    spin_trylock(&napi->poll_lock)) {
			budget = poll_one_napi(npinfo, napi, budget);
			spin_unlock(&napi->poll_lock);

			if (!budget)
				break;
		}
	}
}

static void service_arp_queue(struct netpoll_info *npi)
{
	struct sk_buff *skb;

	if (unlikely(!npi))
		return;

	skb = skb_dequeue(&npi->arp_tx);

	while (skb != NULL) {
		arp_reply(skb);
		skb = skb_dequeue(&npi->arp_tx);
	}
}

void netpoll_poll(struct netpoll *np)
{
	if (!np->dev || !netif_running(np->dev) || !np->dev->poll_controller)
		return;

	/* Process pending work on NIC */
	np->dev->poll_controller(np->dev);
	if (!list_empty(&np->dev->napi_list))
		poll_napi(np);

	service_arp_queue(np->dev->npinfo);

	zap_completion_queue();
}

static void refill_skbs(void)
{
	struct sk_buff *skb;
	unsigned long flags;

	spin_lock_irqsave(&skb_pool.lock, flags);
	while (skb_pool.qlen < MAX_SKBS) {
		skb = alloc_skb(MAX_SKB_SIZE, GFP_ATOMIC);
		if (!skb)
			break;

		__skb_queue_tail(&skb_pool, skb);
	}
	spin_unlock_irqrestore(&skb_pool.lock, flags);
}

static void zap_completion_queue(void)
{
	unsigned long flags;
	struct softnet_data *sd = &get_cpu_var(softnet_data);

	if (sd->completion_queue) {
		struct sk_buff *clist;

		local_irq_save(flags);
		clist = sd->completion_queue;
		sd->completion_queue = NULL;
		local_irq_restore(flags);

		while (clist != NULL) {
			struct sk_buff *skb = clist;
			clist = clist->next;
			if (skb->destructor)
				dev_kfree_skb_any(skb); /* put this one back */
			else
				__kfree_skb(skb);
		}
	}

	put_cpu_var(softnet_data);
}

static struct sk_buff *find_skb(struct netpoll *np, int len, int reserve)
{
	int count = 0;
	struct sk_buff *skb;

	zap_completion_queue();
	refill_skbs();
repeat:

	skb = alloc_skb(len, GFP_ATOMIC);
	if (!skb)
		skb = skb_dequeue(&skb_pool);

	if (!skb) {
		if (++count < 10) {
			netpoll_poll(np);
			goto repeat;
		}
		return NULL;
	}

	atomic_set(&skb->users, 1);
	skb_reserve(skb, reserve);
	return skb;
}

static int netpoll_owner_active(struct net_device *dev)
{
	struct napi_struct *napi;

	list_for_each_entry(napi, &dev->napi_list, dev_list) {
		if (napi->poll_owner == smp_processor_id())
			return 1;
	}
	return 0;
}

static void netpoll_send_skb(struct netpoll *np, struct sk_buff *skb)
{
	int status = NETDEV_TX_BUSY;
	unsigned long tries;
	struct net_device *dev = np->dev;
	struct netpoll_info *npinfo = np->dev->npinfo;

	if (!npinfo || !netif_running(dev) || !netif_device_present(dev)) {
		__kfree_skb(skb);
		return;
	}

	/* don't get messages out of order, and no recursion */
	if (skb_queue_len(&npinfo->txq) == 0 && !netpoll_owner_active(dev)) {
		unsigned long flags;

		local_irq_save(flags);
		/* try until next clock tick */
		for (tries = jiffies_to_usecs(1)/USEC_PER_POLL;
		     tries > 0; --tries) {
			if (netif_tx_trylock(dev)) {
				if (!netif_queue_stopped(dev) &&
				    !netif_subqueue_stopped(dev, skb))
					status = dev->hard_start_xmit(skb, dev);
				netif_tx_unlock(dev);

				if (status == NETDEV_TX_OK)
					break;

			}

			/* tickle device maybe there is some cleanup */
			netpoll_poll(np);

			udelay(USEC_PER_POLL);
		}
		local_irq_restore(flags);
	}

	if (status != NETDEV_TX_OK) {
		skb_queue_tail(&npinfo->txq, skb);
		schedule_delayed_work(&npinfo->tx_work,0);
	}
}

void netpoll_send_udp(struct netpoll *np, const char *msg, int len)
{
	int total_len, eth_len, ip_len, udp_len;
	struct sk_buff *skb;
	struct udphdr *udph;
	struct iphdr *iph;
	struct ethhdr *eth;

	udp_len = len + sizeof(*udph);
	ip_len = eth_len = udp_len + sizeof(*iph);
	total_len = eth_len + ETH_HLEN + NET_IP_ALIGN;

	skb = find_skb(np, total_len, total_len - len);
	if (!skb)
		return;

	skb_copy_to_linear_data(skb, msg, len);
	skb->len += len;

	skb_push(skb, sizeof(*udph));
	skb_reset_transport_header(skb);
	udph = udp_hdr(skb);
	udph->source = htons(np->local_port);
	udph->dest = htons(np->remote_port);
	udph->len = htons(udp_len);
	udph->check = 0;
	udph->check = csum_tcpudp_magic(htonl(np->local_ip),
					htonl(np->remote_ip),
					udp_len, IPPROTO_UDP,
					csum_partial((unsigned char *)udph, udp_len, 0));
	if (udph->check == 0)
		udph->check = CSUM_MANGLED_0;

	skb_push(skb, sizeof(*iph));
	skb_reset_network_header(skb);
	iph = ip_hdr(skb);

	/* iph->version = 4; iph->ihl = 5; */
	put_unaligned(0x45, (unsigned char *)iph);
	iph->tos      = 0;
	put_unaligned(htons(ip_len), &(iph->tot_len));
	iph->id       = 0;
	iph->frag_off = 0;
	iph->ttl      = 64;
	iph->protocol = IPPROTO_UDP;
	iph->check    = 0;
	put_unaligned(htonl(np->local_ip), &(iph->saddr));
	put_unaligned(htonl(np->remote_ip), &(iph->daddr));
	iph->check    = ip_fast_csum((unsigned char *)iph, iph->ihl);

	eth = (struct ethhdr *) skb_push(skb, ETH_HLEN);
	skb_reset_mac_header(skb);
	skb->protocol = eth->h_proto = htons(ETH_P_IP);
	memcpy(eth->h_source, np->local_mac, 6);
	memcpy(eth->h_dest, np->remote_mac, 6);

	skb->dev = np->dev;

	netpoll_send_skb(np, skb);
}

static void arp_reply(struct sk_buff *skb)
{
	struct netpoll_info *npinfo = skb->dev->npinfo;
	struct arphdr *arp;
	unsigned char *arp_ptr;
	int size, type = ARPOP_REPLY, ptype = ETH_P_ARP;
	__be32 sip, tip;
	unsigned char *sha;
	struct sk_buff *send_skb;
	struct netpoll *np = NULL;

	if (npinfo->rx_np && npinfo->rx_np->dev == skb->dev)
		np = npinfo->rx_np;
	if (!np)
		return;

	/* No arp on this interface */
	if (skb->dev->flags & IFF_NOARP)
		return;

	if (!pskb_may_pull(skb, (sizeof(struct arphdr) +
				 (2 * skb->dev->addr_len) +
				 (2 * sizeof(u32)))))
		return;

	skb_reset_network_header(skb);
	skb_reset_transport_header(skb);
	arp = arp_hdr(skb);

	if ((arp->ar_hrd != htons(ARPHRD_ETHER) &&
	     arp->ar_hrd != htons(ARPHRD_IEEE802)) ||
	    arp->ar_pro != htons(ETH_P_IP) ||
	    arp->ar_op != htons(ARPOP_REQUEST))
		return;

	arp_ptr = (unsigned char *)(arp+1);
	/* save the location of the src hw addr */
	sha = arp_ptr;
	arp_ptr += skb->dev->addr_len;
	memcpy(&sip, arp_ptr, 4);
	arp_ptr += 4;
	/* if we actually cared about dst hw addr, it would get copied here */
	arp_ptr += skb->dev->addr_len;
	memcpy(&tip, arp_ptr, 4);

	/* Should we ignore arp? */
	if (tip != htonl(np->local_ip) || LOOPBACK(tip) || MULTICAST(tip))
		return;

	size = sizeof(struct arphdr) + 2 * (skb->dev->addr_len + 4);
	send_skb = find_skb(np, size + LL_RESERVED_SPACE(np->dev),
			    LL_RESERVED_SPACE(np->dev));

	if (!send_skb)