netpoll.c 18.6 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
/*
 * Common framework for low-level network console, dump, and debugger code
 *
 * Sep 8 2003  Matt Mackall <mpm@selenic.com>
 *
 * based on the netconsole code from:
 *
 * Copyright (C) 2001  Ingo Molnar <mingo@redhat.com>
 * Copyright (C) 2002  Red Hat, Inc.
 */

#include <linux/smp_lock.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/string.h>
16
#include <linux/if_arp.h>
L
Linus Torvalds 已提交
17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37
#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)

38
static struct sk_buff_head skb_pool;
L
Linus Torvalds 已提交
39 40 41

static atomic_t trapped;

S
Stephen Hemminger 已提交
42
#define USEC_PER_POLL	50
L
Linus Torvalds 已提交
43 44 45 46 47 48 49 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);
51
static void arp_reply(struct sk_buff *skb);
L
Linus Torvalds 已提交
52

D
David Howells 已提交
53
static void queue_process(struct work_struct *work)
L
Linus Torvalds 已提交
54
{
55 56
	struct netpoll_info *npinfo =
		container_of(work, struct netpoll_info, tx_work.work);
L
Linus Torvalds 已提交
57
	struct sk_buff *skb;
I
Ingo Molnar 已提交
58
	unsigned long flags;
L
Linus Torvalds 已提交
59

S
Stephen Hemminger 已提交
60 61
	while ((skb = skb_dequeue(&npinfo->txq))) {
		struct net_device *dev = skb->dev;
L
Linus Torvalds 已提交
62

S
Stephen Hemminger 已提交
63 64 65 66
		if (!netif_device_present(dev) || !netif_running(dev)) {
			__kfree_skb(skb);
			continue;
		}
L
Linus Torvalds 已提交
67

I
Ingo Molnar 已提交
68 69
		local_irq_save(flags);
		netif_tx_lock(dev);
S
Stephen Hemminger 已提交
70 71 72
		if (netif_queue_stopped(dev) ||
		    dev->hard_start_xmit(skb, dev) != NETDEV_TX_OK) {
			skb_queue_head(&npinfo->txq, skb);
I
Ingo Molnar 已提交
73 74
			netif_tx_unlock(dev);
			local_irq_restore(flags);
L
Linus Torvalds 已提交
75

S
Stephen Hemminger 已提交
76 77 78
			schedule_delayed_work(&npinfo->tx_work, HZ/10);
			return;
		}
I
Ingo Molnar 已提交
79 80
		netif_tx_unlock(dev);
		local_irq_restore(flags);
L
Linus Torvalds 已提交
81 82 83
	}
}

84 85
static __sum16 checksum_udp(struct sk_buff *skb, struct udphdr *uh,
			    unsigned short ulen, __be32 saddr, __be32 daddr)
L
Linus Torvalds 已提交
86
{
87
	__wsum psum;
88 89

	if (uh->check == 0 || skb->ip_summed == CHECKSUM_UNNECESSARY)
L
Linus Torvalds 已提交
90 91
		return 0;

92 93
	psum = csum_tcpudp_nofold(saddr, daddr, ulen, IPPROTO_UDP, 0);

94
	if (skb->ip_summed == CHECKSUM_COMPLETE &&
95
	    !csum_fold(csum_add(psum, skb->csum)))
96
		return 0;
L
Linus Torvalds 已提交
97

98
	skb->csum = psum;
L
Linus Torvalds 已提交
99

100
	return __skb_checksum_complete(skb);
L
Linus Torvalds 已提交
101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120
}

/*
 * 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 void poll_napi(struct netpoll *np)
{
121
	struct netpoll_info *npinfo = np->dev->npinfo;
L
Linus Torvalds 已提交
122 123 124
	int budget = 16;

	if (test_bit(__LINK_STATE_RX_SCHED, &np->dev->state) &&
125 126 127
	    npinfo->poll_owner != smp_processor_id() &&
	    spin_trylock(&npinfo->poll_lock)) {
		npinfo->rx_flags |= NETPOLL_RX_DROP;
L
Linus Torvalds 已提交
128 129 130 131 132
		atomic_inc(&trapped);

		np->dev->poll(np->dev, &budget);

		atomic_dec(&trapped);
133 134
		npinfo->rx_flags &= ~NETPOLL_RX_DROP;
		spin_unlock(&npinfo->poll_lock);
L
Linus Torvalds 已提交
135 136 137
	}
}

138 139 140 141 142 143 144 145 146 147 148 149 150 151 152
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);
	}
}

L
Linus Torvalds 已提交
153 154
void netpoll_poll(struct netpoll *np)
{
155
	if (!np->dev || !netif_running(np->dev) || !np->dev->poll_controller)
L
Linus Torvalds 已提交
156 157 158 159 160 161 162
		return;

	/* Process pending work on NIC */
	np->dev->poll_controller(np->dev);
	if (np->dev->poll)
		poll_napi(np);

163 164
	service_arp_queue(np->dev->npinfo);

L
Linus Torvalds 已提交
165 166 167 168 169 170 171 172
	zap_completion_queue();
}

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

173 174
	spin_lock_irqsave(&skb_pool.lock, flags);
	while (skb_pool.qlen < MAX_SKBS) {
L
Linus Torvalds 已提交
175 176 177 178
		skb = alloc_skb(MAX_SKB_SIZE, GFP_ATOMIC);
		if (!skb)
			break;

179
		__skb_queue_tail(&skb_pool, skb);
L
Linus Torvalds 已提交
180
	}
181
	spin_unlock_irqrestore(&skb_pool.lock, flags);
L
Linus Torvalds 已提交
182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199
}

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;
200
			if (skb->destructor)
L
Linus Torvalds 已提交
201 202 203 204 205 206 207 208 209
				dev_kfree_skb_any(skb); /* put this one back */
			else
				__kfree_skb(skb);
		}
	}

	put_cpu_var(softnet_data);
}

210
static struct sk_buff *find_skb(struct netpoll *np, int len, int reserve)
L
Linus Torvalds 已提交
211
{
212 213
	int count = 0;
	struct sk_buff *skb;
L
Linus Torvalds 已提交
214 215

	zap_completion_queue();
216
	refill_skbs();
L
Linus Torvalds 已提交
217 218 219
repeat:

	skb = alloc_skb(len, GFP_ATOMIC);
220 221
	if (!skb)
		skb = skb_dequeue(&skb_pool);
L
Linus Torvalds 已提交
222 223

	if (!skb) {
224 225 226
		if (++count < 10) {
			netpoll_poll(np);
			goto repeat;
L
Linus Torvalds 已提交
227
		}
228
		return NULL;
L
Linus Torvalds 已提交
229 230 231 232 233 234 235 236 237
	}

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

static void netpoll_send_skb(struct netpoll *np, struct sk_buff *skb)
{
S
Stephen Hemminger 已提交
238 239
	int status = NETDEV_TX_BUSY;
	unsigned long tries;
240 241
	struct net_device *dev = np->dev;
	struct netpoll_info *npinfo = np->dev->npinfo;
S
Stephen Hemminger 已提交
242

243 244 245 246
	if (!npinfo || !netif_running(dev) || !netif_device_present(dev)) {
		__kfree_skb(skb);
		return;
	}
S
Stephen Hemminger 已提交
247 248

	/* don't get messages out of order, and no recursion */
249
	if (skb_queue_len(&npinfo->txq) == 0 &&
250
		    npinfo->poll_owner != smp_processor_id()) {
251 252 253
		unsigned long flags;

		local_irq_save(flags);
254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269
		if (netif_tx_trylock(dev)) {
			/* try until next clock tick */
			for (tries = jiffies_to_usecs(1)/USEC_PER_POLL;
					tries > 0; --tries) {
				if (!netif_queue_stopped(dev))
					status = dev->hard_start_xmit(skb, dev);

				if (status == NETDEV_TX_OK)
					break;

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

				udelay(USEC_PER_POLL);
			}
			netif_tx_unlock(dev);
M
Matt Mackall 已提交
270
		}
271
		local_irq_restore(flags);
L
Linus Torvalds 已提交
272 273
	}

S
Stephen Hemminger 已提交
274
	if (status != NETDEV_TX_OK) {
S
Stephen Hemminger 已提交
275
		skb_queue_tail(&npinfo->txq, skb);
276
		schedule_delayed_work(&npinfo->tx_work,0);
L
Linus Torvalds 已提交
277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298
	}
}

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;

	memcpy(skb->data, msg, len);
	skb->len += len;

299
	skb->h.uh = udph = (struct udphdr *) skb_push(skb, sizeof(*udph));
L
Linus Torvalds 已提交
300 301 302 303
	udph->source = htons(np->local_port);
	udph->dest = htons(np->remote_port);
	udph->len = htons(udp_len);
	udph->check = 0;
304 305 306 307 308
	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)
309
		udph->check = CSUM_MANGLED_0;
L
Linus Torvalds 已提交
310

311
	skb->nh.iph = iph = (struct iphdr *)skb_push(skb, sizeof(*iph));
L
Linus Torvalds 已提交
312 313 314 315 316 317 318 319 320 321 322 323 324 325 326

	/* 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);
327
	skb_reset_mac_header(skb);
328
	skb->protocol = eth->h_proto = htons(ETH_P_IP);
L
Linus Torvalds 已提交
329 330 331 332 333 334 335 336 337 338
	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)
{
339
	struct netpoll_info *npinfo = skb->dev->npinfo;
L
Linus Torvalds 已提交
340 341 342
	struct arphdr *arp;
	unsigned char *arp_ptr;
	int size, type = ARPOP_REPLY, ptype = ETH_P_ARP;
A
Al Viro 已提交
343
	__be32 sip, tip;
344
	unsigned char *sha;
L
Linus Torvalds 已提交
345
	struct sk_buff *send_skb;
346
	struct netpoll *np = NULL;
L
Linus Torvalds 已提交
347

348 349
	if (npinfo->rx_np && npinfo->rx_np->dev == skb->dev)
		np = npinfo->rx_np;
350 351
	if (!np)
		return;
L
Linus Torvalds 已提交
352 353 354 355 356 357 358 359 360 361

	/* 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;

362 363
	skb_reset_network_header(skb);
	skb->h.raw = skb->data;
L
Linus Torvalds 已提交
364 365 366 367 368 369 370 371
	arp = skb->nh.arph;

	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;

372 373 374 375
	arp_ptr = (unsigned char *)(arp+1);
	/* save the location of the src hw addr */
	sha = arp_ptr;
	arp_ptr += skb->dev->addr_len;
L
Linus Torvalds 已提交
376
	memcpy(&sip, arp_ptr, 4);
377 378 379
	arp_ptr += 4;
	/* if we actually cared about dst hw addr, it would get copied here */
	arp_ptr += skb->dev->addr_len;
L
Linus Torvalds 已提交
380 381 382 383 384 385 386 387 388 389 390 391 392
	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)
		return;

393
	skb_reset_network_header(send_skb);
L
Linus Torvalds 已提交
394 395 396 397 398 399 400 401
	arp = (struct arphdr *) skb_put(send_skb, size);
	send_skb->dev = skb->dev;
	send_skb->protocol = htons(ETH_P_ARP);

	/* Fill the device header for the ARP frame */

	if (np->dev->hard_header &&
	    np->dev->hard_header(send_skb, skb->dev, ptype,
402
				 sha, np->local_mac,
403
				 send_skb->len) < 0) {
L
Linus Torvalds 已提交
404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425
		kfree_skb(send_skb);
		return;
	}

	/*
	 * Fill out the arp protocol part.
	 *
	 * we only support ethernet device type,
	 * which (according to RFC 1390) should always equal 1 (Ethernet).
	 */

	arp->ar_hrd = htons(np->dev->type);
	arp->ar_pro = htons(ETH_P_IP);
	arp->ar_hln = np->dev->addr_len;
	arp->ar_pln = 4;
	arp->ar_op = htons(type);

	arp_ptr=(unsigned char *)(arp + 1);
	memcpy(arp_ptr, np->dev->dev_addr, np->dev->addr_len);
	arp_ptr += np->dev->addr_len;
	memcpy(arp_ptr, &tip, 4);
	arp_ptr += 4;
426
	memcpy(arp_ptr, sha, np->dev->addr_len);
L
Linus Torvalds 已提交
427 428 429 430 431 432 433 434 435 436 437
	arp_ptr += np->dev->addr_len;
	memcpy(arp_ptr, &sip, 4);

	netpoll_send_skb(np, send_skb);
}

int __netpoll_rx(struct sk_buff *skb)
{
	int proto, len, ulen;
	struct iphdr *iph;
	struct udphdr *uh;
438 439 440
	struct netpoll_info *npi = skb->dev->npinfo;
	struct netpoll *np = npi->rx_np;

441
	if (!np)
L
Linus Torvalds 已提交
442 443 444 445 446
		goto out;
	if (skb->dev->type != ARPHRD_ETHER)
		goto out;

	/* check if netpoll clients need ARP */
447
	if (skb->protocol == htons(ETH_P_ARP) &&
L
Linus Torvalds 已提交
448
	    atomic_read(&trapped)) {
449
		skb_queue_tail(&npi->arp_tx, skb);
L
Linus Torvalds 已提交
450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474
		return 1;
	}

	proto = ntohs(eth_hdr(skb)->h_proto);
	if (proto != ETH_P_IP)
		goto out;
	if (skb->pkt_type == PACKET_OTHERHOST)
		goto out;
	if (skb_shared(skb))
		goto out;

	iph = (struct iphdr *)skb->data;
	if (!pskb_may_pull(skb, sizeof(struct iphdr)))
		goto out;
	if (iph->ihl < 5 || iph->version != 4)
		goto out;
	if (!pskb_may_pull(skb, iph->ihl*4))
		goto out;
	if (ip_fast_csum((u8 *)iph, iph->ihl) != 0)
		goto out;

	len = ntohs(iph->tot_len);
	if (skb->len < len || len < iph->ihl*4)
		goto out;

475 476 477 478 479 480 481
	/*
	 * Our transport medium may have padded the buffer out.
	 * Now We trim to the true length of the frame.
	 */
	if (pskb_trim_rcsum(skb, len))
		goto out;

L
Linus Torvalds 已提交
482 483 484 485 486 487 488 489 490
	if (iph->protocol != IPPROTO_UDP)
		goto out;

	len -= iph->ihl*4;
	uh = (struct udphdr *)(((char *)iph) + iph->ihl*4);
	ulen = ntohs(uh->len);

	if (ulen != len)
		goto out;
491
	if (checksum_udp(skb, uh, ulen, iph->saddr, iph->daddr))
L
Linus Torvalds 已提交
492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519
		goto out;
	if (np->local_ip && np->local_ip != ntohl(iph->daddr))
		goto out;
	if (np->remote_ip && np->remote_ip != ntohl(iph->saddr))
		goto out;
	if (np->local_port && np->local_port != ntohs(uh->dest))
		goto out;

	np->rx_hook(np, ntohs(uh->source),
		    (char *)(uh+1),
		    ulen - sizeof(struct udphdr));

	kfree_skb(skb);
	return 1;

out:
	if (atomic_read(&trapped)) {
		kfree_skb(skb);
		return 1;
	}

	return 0;
}

int netpoll_parse_options(struct netpoll *np, char *opt)
{
	char *cur=opt, *delim;

520
	if (*cur != '@') {
L
Linus Torvalds 已提交
521 522
		if ((delim = strchr(cur, '@')) == NULL)
			goto parse_failed;
523 524 525
		*delim = 0;
		np->local_port = simple_strtol(cur, NULL, 10);
		cur = delim;
L
Linus Torvalds 已提交
526 527 528 529
	}
	cur++;
	printk(KERN_INFO "%s: local port %d\n", np->name, np->local_port);

530
	if (*cur != '/') {
L
Linus Torvalds 已提交
531 532
		if ((delim = strchr(cur, '/')) == NULL)
			goto parse_failed;
533 534 535
		*delim = 0;
		np->local_ip = ntohl(in_aton(cur));
		cur = delim;
L
Linus Torvalds 已提交
536 537 538 539 540 541

		printk(KERN_INFO "%s: local IP %d.%d.%d.%d\n",
		       np->name, HIPQUAD(np->local_ip));
	}
	cur++;

542
	if (*cur != ',') {
L
Linus Torvalds 已提交
543 544 545
		/* parse out dev name */
		if ((delim = strchr(cur, ',')) == NULL)
			goto parse_failed;
546
		*delim = 0;
L
Linus Torvalds 已提交
547
		strlcpy(np->dev_name, cur, sizeof(np->dev_name));
548
		cur = delim;
L
Linus Torvalds 已提交
549 550 551 552 553
	}
	cur++;

	printk(KERN_INFO "%s: interface %s\n", np->name, np->dev_name);

554
	if (*cur != '@') {
L
Linus Torvalds 已提交
555 556 557
		/* dst port */
		if ((delim = strchr(cur, '@')) == NULL)
			goto parse_failed;
558 559 560
		*delim = 0;
		np->remote_port = simple_strtol(cur, NULL, 10);
		cur = delim;
L
Linus Torvalds 已提交
561 562 563 564 565 566 567
	}
	cur++;
	printk(KERN_INFO "%s: remote port %d\n", np->name, np->remote_port);

	/* dst ip */
	if ((delim = strchr(cur, '/')) == NULL)
		goto parse_failed;
568 569 570
	*delim = 0;
	np->remote_ip = ntohl(in_aton(cur));
	cur = delim + 1;
L
Linus Torvalds 已提交
571 572

	printk(KERN_INFO "%s: remote IP %d.%d.%d.%d\n",
573
	       np->name, HIPQUAD(np->remote_ip));
L
Linus Torvalds 已提交
574

575
	if (*cur != 0) {
L
Linus Torvalds 已提交
576 577 578
		/* MAC address */
		if ((delim = strchr(cur, ':')) == NULL)
			goto parse_failed;
579 580 581
		*delim = 0;
		np->remote_mac[0] = simple_strtol(cur, NULL, 16);
		cur = delim + 1;
L
Linus Torvalds 已提交
582 583
		if ((delim = strchr(cur, ':')) == NULL)
			goto parse_failed;
584 585 586
		*delim = 0;
		np->remote_mac[1] = simple_strtol(cur, NULL, 16);
		cur = delim + 1;
L
Linus Torvalds 已提交
587 588
		if ((delim = strchr(cur, ':')) == NULL)
			goto parse_failed;
589 590 591
		*delim = 0;
		np->remote_mac[2] = simple_strtol(cur, NULL, 16);
		cur = delim + 1;
L
Linus Torvalds 已提交
592 593
		if ((delim = strchr(cur, ':')) == NULL)
			goto parse_failed;
594 595 596
		*delim = 0;
		np->remote_mac[3] = simple_strtol(cur, NULL, 16);
		cur = delim + 1;
L
Linus Torvalds 已提交
597 598
		if ((delim = strchr(cur, ':')) == NULL)
			goto parse_failed;
599 600 601 602
		*delim = 0;
		np->remote_mac[4] = simple_strtol(cur, NULL, 16);
		cur = delim + 1;
		np->remote_mac[5] = simple_strtol(cur, NULL, 16);
L
Linus Torvalds 已提交
603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626
	}

	printk(KERN_INFO "%s: remote ethernet address "
	       "%02x:%02x:%02x:%02x:%02x:%02x\n",
	       np->name,
	       np->remote_mac[0],
	       np->remote_mac[1],
	       np->remote_mac[2],
	       np->remote_mac[3],
	       np->remote_mac[4],
	       np->remote_mac[5]);

	return 0;

 parse_failed:
	printk(KERN_INFO "%s: couldn't parse config at %s!\n",
	       np->name, cur);
	return -1;
}

int netpoll_setup(struct netpoll *np)
{
	struct net_device *ndev = NULL;
	struct in_device *in_dev;
627
	struct netpoll_info *npinfo;
628
	unsigned long flags;
S
Stephen Hemminger 已提交
629
	int err;
L
Linus Torvalds 已提交
630 631 632 633 634 635

	if (np->dev_name)
		ndev = dev_get_by_name(np->dev_name);
	if (!ndev) {
		printk(KERN_ERR "%s: %s doesn't exist, aborting.\n",
		       np->name, np->dev_name);
S
Stephen Hemminger 已提交
636
		return -ENODEV;
L
Linus Torvalds 已提交
637 638 639
	}

	np->dev = ndev;
640 641
	if (!ndev->npinfo) {
		npinfo = kmalloc(sizeof(*npinfo), GFP_KERNEL);
S
Stephen Hemminger 已提交
642 643
		if (!npinfo) {
			err = -ENOMEM;
644
			goto release;
S
Stephen Hemminger 已提交
645
		}
646

J
Jeff Moyer 已提交
647
		npinfo->rx_flags = 0;
648
		npinfo->rx_np = NULL;
649
		spin_lock_init(&npinfo->poll_lock);
650
		npinfo->poll_owner = -1;
S
Stephen Hemminger 已提交
651

652
		spin_lock_init(&npinfo->rx_lock);
653
		skb_queue_head_init(&npinfo->arp_tx);
S
Stephen Hemminger 已提交
654
		skb_queue_head_init(&npinfo->txq);
655
		INIT_DELAYED_WORK(&npinfo->tx_work, queue_process);
S
Stephen Hemminger 已提交
656

S
Stephen Hemminger 已提交
657 658
		atomic_set(&npinfo->refcnt, 1);
	} else {
659
		npinfo = ndev->npinfo;
S
Stephen Hemminger 已提交
660 661
		atomic_inc(&npinfo->refcnt);
	}
L
Linus Torvalds 已提交
662 663 664 665

	if (!ndev->poll_controller) {
		printk(KERN_ERR "%s: %s doesn't support polling, aborting.\n",
		       np->name, np->dev_name);
S
Stephen Hemminger 已提交
666
		err = -ENOTSUPP;
L
Linus Torvalds 已提交
667 668 669 670 671 672 673 674 675
		goto release;
	}

	if (!netif_running(ndev)) {
		unsigned long atmost, atleast;

		printk(KERN_INFO "%s: device %s not up yet, forcing it\n",
		       np->name, np->dev_name);

676
		rtnl_lock();
S
Stephen Hemminger 已提交
677 678 679 680
		err = dev_open(ndev);
		rtnl_unlock();

		if (err) {
L
Linus Torvalds 已提交
681
			printk(KERN_ERR "%s: failed to open %s\n",
S
Stephen Hemminger 已提交
682
			       np->name, ndev->name);
L
Linus Torvalds 已提交
683 684 685 686
			goto release;
		}

		atleast = jiffies + HZ/10;
687
		atmost = jiffies + 4*HZ;
L
Linus Torvalds 已提交
688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710
		while (!netif_carrier_ok(ndev)) {
			if (time_after(jiffies, atmost)) {
				printk(KERN_NOTICE
				       "%s: timeout waiting for carrier\n",
				       np->name);
				break;
			}
			cond_resched();
		}

		/* If carrier appears to come up instantly, we don't
		 * trust it and pause so that we don't pump all our
		 * queued console messages into the bitbucket.
		 */

		if (time_before(jiffies, atleast)) {
			printk(KERN_NOTICE "%s: carrier detect appears"
			       " untrustworthy, waiting 4 seconds\n",
			       np->name);
			msleep(4000);
		}
	}

711
	if (is_zero_ether_addr(np->local_mac) && ndev->dev_addr)
L
Linus Torvalds 已提交
712 713 714 715
		memcpy(np->local_mac, ndev->dev_addr, 6);

	if (!np->local_ip) {
		rcu_read_lock();
716
		in_dev = __in_dev_get_rcu(ndev);
L
Linus Torvalds 已提交
717 718 719 720 721

		if (!in_dev || !in_dev->ifa_list) {
			rcu_read_unlock();
			printk(KERN_ERR "%s: no IP address for %s, aborting\n",
			       np->name, np->dev_name);
S
Stephen Hemminger 已提交
722
			err = -EDESTADDRREQ;
L
Linus Torvalds 已提交
723 724 725 726 727 728 729 730 731
			goto release;
		}

		np->local_ip = ntohl(in_dev->ifa_list->ifa_local);
		rcu_read_unlock();
		printk(KERN_INFO "%s: local IP %d.%d.%d.%d\n",
		       np->name, HIPQUAD(np->local_ip));
	}

732 733 734 735 736 737
	if (np->rx_hook) {
		spin_lock_irqsave(&npinfo->rx_lock, flags);
		npinfo->rx_flags |= NETPOLL_RX_ENABLED;
		npinfo->rx_np = np;
		spin_unlock_irqrestore(&npinfo->rx_lock, flags);
	}
I
Ingo Molnar 已提交
738 739 740 741

	/* fill up the skb queue */
	refill_skbs();

742
	/* last thing to do is link it to the net device structure */
743
	ndev->npinfo = npinfo;
L
Linus Torvalds 已提交
744

745 746 747
	/* avoid racing with NAPI reading npinfo */
	synchronize_rcu();

L
Linus Torvalds 已提交
748 749 750
	return 0;

 release:
751 752
	if (!ndev->npinfo)
		kfree(npinfo);
L
Linus Torvalds 已提交
753 754
	np->dev = NULL;
	dev_put(ndev);
S
Stephen Hemminger 已提交
755
	return err;
L
Linus Torvalds 已提交
756 757
}

758 759
static int __init netpoll_init(void)
{
760 761 762 763 764
	skb_queue_head_init(&skb_pool);
	return 0;
}
core_initcall(netpoll_init);

L
Linus Torvalds 已提交
765 766
void netpoll_cleanup(struct netpoll *np)
{
767 768 769
	struct netpoll_info *npinfo;
	unsigned long flags;

770
	if (np->dev) {
771
		npinfo = np->dev->npinfo;
S
Stephen Hemminger 已提交
772 773 774 775 776 777 778 779 780 781 782
		if (npinfo) {
			if (npinfo->rx_np == np) {
				spin_lock_irqsave(&npinfo->rx_lock, flags);
				npinfo->rx_np = NULL;
				npinfo->rx_flags &= ~NETPOLL_RX_ENABLED;
				spin_unlock_irqrestore(&npinfo->rx_lock, flags);
			}

			np->dev->npinfo = NULL;
			if (atomic_dec_and_test(&npinfo->refcnt)) {
				skb_queue_purge(&npinfo->arp_tx);
783
				skb_queue_purge(&npinfo->txq);
S
Stephen Hemminger 已提交
784
				cancel_rearming_delayed_work(&npinfo->tx_work);
785
				flush_scheduled_work();
S
Stephen Hemminger 已提交
786 787 788

				kfree(npinfo);
			}
789
		}
S
Stephen Hemminger 已提交
790

791 792
		dev_put(np->dev);
	}
793

L
Linus Torvalds 已提交
794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816
	np->dev = NULL;
}

int netpoll_trap(void)
{
	return atomic_read(&trapped);
}

void netpoll_set_trap(int trap)
{
	if (trap)
		atomic_inc(&trapped);
	else
		atomic_dec(&trapped);
}

EXPORT_SYMBOL(netpoll_set_trap);
EXPORT_SYMBOL(netpoll_trap);
EXPORT_SYMBOL(netpoll_parse_options);
EXPORT_SYMBOL(netpoll_setup);
EXPORT_SYMBOL(netpoll_cleanup);
EXPORT_SYMBOL(netpoll_send_udp);
EXPORT_SYMBOL(netpoll_poll);