1 /*
2  * Network-device interface management.
3  *
4  * Copyright (c) 2004-2005, Keir Fraser
5  *
6  * This program is free software; you can redistribute it and/or
7  * modify it under the terms of the GNU General Public License version 2
8  * as published by the Free Software Foundation; or, when distributed
9  * separately from the Linux kernel or incorporated into other
10  * software packages, subject to the following license:
11  *
12  * Permission is hereby granted, free of charge, to any person obtaining a copy
13  * of this source file (the "Software"), to deal in the Software without
14  * restriction, including without limitation the rights to use, copy, modify,
15  * merge, publish, distribute, sublicense, and/or sell copies of the Software,
16  * and to permit persons to whom the Software is furnished to do so, subject to
17  * the following conditions:
18  *
19  * The above copyright notice and this permission notice shall be included in
20  * all copies or substantial portions of the Software.
21  *
22  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
23  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
24  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
25  * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
26  * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
27  * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
28  * IN THE SOFTWARE.
29  */
30 
31 #include "common.h"
32 
33 #include <linux/kthread.h>
34 #include <linux/sched/task.h>
35 #include <linux/ethtool.h>
36 #include <linux/rtnetlink.h>
37 #include <linux/if_vlan.h>
38 #include <linux/vmalloc.h>
39 
40 #include <xen/events.h>
41 #include <asm/xen/hypercall.h>
42 #include <xen/balloon.h>
43 
44 #define XENVIF_QUEUE_LENGTH 32
45 
46 /* Number of bytes allowed on the internal guest Rx queue. */
47 #define XENVIF_RX_QUEUE_BYTES (XEN_NETIF_RX_RING_SIZE/2 * PAGE_SIZE)
48 
49 /* This function is used to set SKBFL_ZEROCOPY_ENABLE as well as
50  * increasing the inflight counter. We need to increase the inflight
51  * counter because core driver calls into xenvif_zerocopy_callback
52  * which calls xenvif_skb_zerocopy_complete.
53  */
54 void xenvif_skb_zerocopy_prepare(struct xenvif_queue *queue,
55 				 struct sk_buff *skb)
56 {
57 	skb_shinfo(skb)->flags |= SKBFL_ZEROCOPY_ENABLE;
58 	atomic_inc(&queue->inflight_packets);
59 }
60 
61 void xenvif_skb_zerocopy_complete(struct xenvif_queue *queue)
62 {
63 	atomic_dec(&queue->inflight_packets);
64 
65 	/* Wake the dealloc thread _after_ decrementing inflight_packets so
66 	 * that if kthread_stop() has already been called, the dealloc thread
67 	 * does not wait forever with nothing to wake it.
68 	 */
69 	wake_up(&queue->dealloc_wq);
70 }
71 
72 static int xenvif_schedulable(struct xenvif *vif)
73 {
74 	return netif_running(vif->dev) &&
75 		test_bit(VIF_STATUS_CONNECTED, &vif->status) &&
76 		!vif->disabled;
77 }
78 
79 static bool xenvif_handle_tx_interrupt(struct xenvif_queue *queue)
80 {
81 	bool rc;
82 
83 	rc = RING_HAS_UNCONSUMED_REQUESTS(&queue->tx);
84 	if (rc)
85 		napi_schedule(&queue->napi);
86 	return rc;
87 }
88 
89 static irqreturn_t xenvif_tx_interrupt(int irq, void *dev_id)
90 {
91 	struct xenvif_queue *queue = dev_id;
92 	int old;
93 
94 	old = atomic_fetch_or(NETBK_TX_EOI, &queue->eoi_pending);
95 	WARN(old & NETBK_TX_EOI, "Interrupt while EOI pending\n");
96 
97 	if (!xenvif_handle_tx_interrupt(queue)) {
98 		atomic_andnot(NETBK_TX_EOI, &queue->eoi_pending);
99 		xen_irq_lateeoi(irq, XEN_EOI_FLAG_SPURIOUS);
100 	}
101 
102 	return IRQ_HANDLED;
103 }
104 
105 static int xenvif_poll(struct napi_struct *napi, int budget)
106 {
107 	struct xenvif_queue *queue =
108 		container_of(napi, struct xenvif_queue, napi);
109 	int work_done;
110 
111 	/* This vif is rogue, we pretend we've there is nothing to do
112 	 * for this vif to deschedule it from NAPI. But this interface
113 	 * will be turned off in thread context later.
114 	 */
115 	if (unlikely(queue->vif->disabled)) {
116 		napi_complete(napi);
117 		return 0;
118 	}
119 
120 	work_done = xenvif_tx_action(queue, budget);
121 
122 	if (work_done < budget) {
123 		napi_complete_done(napi, work_done);
124 		/* If the queue is rate-limited, it shall be
125 		 * rescheduled in the timer callback.
126 		 */
127 		if (likely(!queue->rate_limited))
128 			xenvif_napi_schedule_or_enable_events(queue);
129 	}
130 
131 	return work_done;
132 }
133 
134 static bool xenvif_handle_rx_interrupt(struct xenvif_queue *queue)
135 {
136 	bool rc;
137 
138 	rc = xenvif_have_rx_work(queue, false);
139 	if (rc)
140 		xenvif_kick_thread(queue);
141 	return rc;
142 }
143 
144 static irqreturn_t xenvif_rx_interrupt(int irq, void *dev_id)
145 {
146 	struct xenvif_queue *queue = dev_id;
147 	int old;
148 
149 	old = atomic_fetch_or(NETBK_RX_EOI, &queue->eoi_pending);
150 	WARN(old & NETBK_RX_EOI, "Interrupt while EOI pending\n");
151 
152 	if (!xenvif_handle_rx_interrupt(queue)) {
153 		atomic_andnot(NETBK_RX_EOI, &queue->eoi_pending);
154 		xen_irq_lateeoi(irq, XEN_EOI_FLAG_SPURIOUS);
155 	}
156 
157 	return IRQ_HANDLED;
158 }
159 
160 irqreturn_t xenvif_interrupt(int irq, void *dev_id)
161 {
162 	struct xenvif_queue *queue = dev_id;
163 	int old;
164 	bool has_rx, has_tx;
165 
166 	old = atomic_fetch_or(NETBK_COMMON_EOI, &queue->eoi_pending);
167 	WARN(old, "Interrupt while EOI pending\n");
168 
169 	has_tx = xenvif_handle_tx_interrupt(queue);
170 	has_rx = xenvif_handle_rx_interrupt(queue);
171 
172 	if (!has_rx && !has_tx) {
173 		atomic_andnot(NETBK_COMMON_EOI, &queue->eoi_pending);
174 		xen_irq_lateeoi(irq, XEN_EOI_FLAG_SPURIOUS);
175 	}
176 
177 	return IRQ_HANDLED;
178 }
179 
180 static u16 xenvif_select_queue(struct net_device *dev, struct sk_buff *skb,
181 			       struct net_device *sb_dev)
182 {
183 	struct xenvif *vif = netdev_priv(dev);
184 	unsigned int size = vif->hash.size;
185 	unsigned int num_queues;
186 
187 	/* If queues are not set up internally - always return 0
188 	 * as the packet going to be dropped anyway */
189 	num_queues = READ_ONCE(vif->num_queues);
190 	if (num_queues < 1)
191 		return 0;
192 
193 	if (vif->hash.alg == XEN_NETIF_CTRL_HASH_ALGORITHM_NONE)
194 		return netdev_pick_tx(dev, skb, NULL) %
195 		       dev->real_num_tx_queues;
196 
197 	xenvif_set_skb_hash(vif, skb);
198 
199 	if (size == 0)
200 		return skb_get_hash_raw(skb) % dev->real_num_tx_queues;
201 
202 	return vif->hash.mapping[vif->hash.mapping_sel]
203 				[skb_get_hash_raw(skb) % size];
204 }
205 
206 static netdev_tx_t
207 xenvif_start_xmit(struct sk_buff *skb, struct net_device *dev)
208 {
209 	struct xenvif *vif = netdev_priv(dev);
210 	struct xenvif_queue *queue = NULL;
211 	unsigned int num_queues;
212 	u16 index;
213 	struct xenvif_rx_cb *cb;
214 
215 	BUG_ON(skb->dev != dev);
216 
217 	/* Drop the packet if queues are not set up.
218 	 * This handler should be called inside an RCU read section
219 	 * so we don't need to enter it here explicitly.
220 	 */
221 	num_queues = READ_ONCE(vif->num_queues);
222 	if (num_queues < 1)
223 		goto drop;
224 
225 	/* Obtain the queue to be used to transmit this packet */
226 	index = skb_get_queue_mapping(skb);
227 	if (index >= num_queues) {
228 		pr_warn_ratelimited("Invalid queue %hu for packet on interface %s\n",
229 				    index, vif->dev->name);
230 		index %= num_queues;
231 	}
232 	queue = &vif->queues[index];
233 
234 	/* Drop the packet if queue is not ready */
235 	if (queue->task == NULL ||
236 	    queue->dealloc_task == NULL ||
237 	    !xenvif_schedulable(vif))
238 		goto drop;
239 
240 	if (vif->multicast_control && skb->pkt_type == PACKET_MULTICAST) {
241 		struct ethhdr *eth = (struct ethhdr *)skb->data;
242 
243 		if (!xenvif_mcast_match(vif, eth->h_dest))
244 			goto drop;
245 	}
246 
247 	cb = XENVIF_RX_CB(skb);
248 	cb->expires = jiffies + vif->drain_timeout;
249 
250 	/* If there is no hash algorithm configured then make sure there
251 	 * is no hash information in the socket buffer otherwise it
252 	 * would be incorrectly forwarded to the frontend.
253 	 */
254 	if (vif->hash.alg == XEN_NETIF_CTRL_HASH_ALGORITHM_NONE)
255 		skb_clear_hash(skb);
256 
257 	if (!xenvif_rx_queue_tail(queue, skb))
258 		goto drop;
259 
260 	xenvif_kick_thread(queue);
261 
262 	return NETDEV_TX_OK;
263 
264  drop:
265 	vif->dev->stats.tx_dropped++;
266 	dev_kfree_skb_any(skb);
267 	return NETDEV_TX_OK;
268 }
269 
270 static struct net_device_stats *xenvif_get_stats(struct net_device *dev)
271 {
272 	struct xenvif *vif = netdev_priv(dev);
273 	struct xenvif_queue *queue = NULL;
274 	unsigned int num_queues;
275 	u64 rx_bytes = 0;
276 	u64 rx_packets = 0;
277 	u64 tx_bytes = 0;
278 	u64 tx_packets = 0;
279 	unsigned int index;
280 
281 	rcu_read_lock();
282 	num_queues = READ_ONCE(vif->num_queues);
283 
284 	/* Aggregate tx and rx stats from each queue */
285 	for (index = 0; index < num_queues; ++index) {
286 		queue = &vif->queues[index];
287 		rx_bytes += queue->stats.rx_bytes;
288 		rx_packets += queue->stats.rx_packets;
289 		tx_bytes += queue->stats.tx_bytes;
290 		tx_packets += queue->stats.tx_packets;
291 	}
292 
293 	rcu_read_unlock();
294 
295 	vif->dev->stats.rx_bytes = rx_bytes;
296 	vif->dev->stats.rx_packets = rx_packets;
297 	vif->dev->stats.tx_bytes = tx_bytes;
298 	vif->dev->stats.tx_packets = tx_packets;
299 
300 	return &vif->dev->stats;
301 }
302 
303 static void xenvif_up(struct xenvif *vif)
304 {
305 	struct xenvif_queue *queue = NULL;
306 	unsigned int num_queues = vif->num_queues;
307 	unsigned int queue_index;
308 
309 	for (queue_index = 0; queue_index < num_queues; ++queue_index) {
310 		queue = &vif->queues[queue_index];
311 		napi_enable(&queue->napi);
312 		enable_irq(queue->tx_irq);
313 		if (queue->tx_irq != queue->rx_irq)
314 			enable_irq(queue->rx_irq);
315 		xenvif_napi_schedule_or_enable_events(queue);
316 	}
317 }
318 
319 static void xenvif_down(struct xenvif *vif)
320 {
321 	struct xenvif_queue *queue = NULL;
322 	unsigned int num_queues = vif->num_queues;
323 	unsigned int queue_index;
324 
325 	for (queue_index = 0; queue_index < num_queues; ++queue_index) {
326 		queue = &vif->queues[queue_index];
327 		disable_irq(queue->tx_irq);
328 		if (queue->tx_irq != queue->rx_irq)
329 			disable_irq(queue->rx_irq);
330 		napi_disable(&queue->napi);
331 		del_timer_sync(&queue->credit_timeout);
332 	}
333 }
334 
335 static int xenvif_open(struct net_device *dev)
336 {
337 	struct xenvif *vif = netdev_priv(dev);
338 	if (test_bit(VIF_STATUS_CONNECTED, &vif->status))
339 		xenvif_up(vif);
340 	netif_tx_start_all_queues(dev);
341 	return 0;
342 }
343 
344 static int xenvif_close(struct net_device *dev)
345 {
346 	struct xenvif *vif = netdev_priv(dev);
347 	if (test_bit(VIF_STATUS_CONNECTED, &vif->status))
348 		xenvif_down(vif);
349 	netif_tx_stop_all_queues(dev);
350 	return 0;
351 }
352 
353 static int xenvif_change_mtu(struct net_device *dev, int mtu)
354 {
355 	struct xenvif *vif = netdev_priv(dev);
356 	int max = vif->can_sg ? ETH_MAX_MTU - VLAN_ETH_HLEN : ETH_DATA_LEN;
357 
358 	if (mtu > max)
359 		return -EINVAL;
360 	dev->mtu = mtu;
361 	return 0;
362 }
363 
364 static netdev_features_t xenvif_fix_features(struct net_device *dev,
365 	netdev_features_t features)
366 {
367 	struct xenvif *vif = netdev_priv(dev);
368 
369 	if (!vif->can_sg)
370 		features &= ~NETIF_F_SG;
371 	if (~(vif->gso_mask) & GSO_BIT(TCPV4))
372 		features &= ~NETIF_F_TSO;
373 	if (~(vif->gso_mask) & GSO_BIT(TCPV6))
374 		features &= ~NETIF_F_TSO6;
375 	if (!vif->ip_csum)
376 		features &= ~NETIF_F_IP_CSUM;
377 	if (!vif->ipv6_csum)
378 		features &= ~NETIF_F_IPV6_CSUM;
379 
380 	return features;
381 }
382 
383 static const struct xenvif_stat {
384 	char name[ETH_GSTRING_LEN];
385 	u16 offset;
386 } xenvif_stats[] = {
387 	{
388 		"rx_gso_checksum_fixup",
389 		offsetof(struct xenvif_stats, rx_gso_checksum_fixup)
390 	},
391 	/* If (sent != success + fail), there are probably packets never
392 	 * freed up properly!
393 	 */
394 	{
395 		"tx_zerocopy_sent",
396 		offsetof(struct xenvif_stats, tx_zerocopy_sent),
397 	},
398 	{
399 		"tx_zerocopy_success",
400 		offsetof(struct xenvif_stats, tx_zerocopy_success),
401 	},
402 	{
403 		"tx_zerocopy_fail",
404 		offsetof(struct xenvif_stats, tx_zerocopy_fail)
405 	},
406 	/* Number of packets exceeding MAX_SKB_FRAG slots. You should use
407 	 * a guest with the same MAX_SKB_FRAG
408 	 */
409 	{
410 		"tx_frag_overflow",
411 		offsetof(struct xenvif_stats, tx_frag_overflow)
412 	},
413 };
414 
415 static int xenvif_get_sset_count(struct net_device *dev, int string_set)
416 {
417 	switch (string_set) {
418 	case ETH_SS_STATS:
419 		return ARRAY_SIZE(xenvif_stats);
420 	default:
421 		return -EINVAL;
422 	}
423 }
424 
425 static void xenvif_get_ethtool_stats(struct net_device *dev,
426 				     struct ethtool_stats *stats, u64 * data)
427 {
428 	struct xenvif *vif = netdev_priv(dev);
429 	unsigned int num_queues;
430 	int i;
431 	unsigned int queue_index;
432 
433 	rcu_read_lock();
434 	num_queues = READ_ONCE(vif->num_queues);
435 
436 	for (i = 0; i < ARRAY_SIZE(xenvif_stats); i++) {
437 		unsigned long accum = 0;
438 		for (queue_index = 0; queue_index < num_queues; ++queue_index) {
439 			void *vif_stats = &vif->queues[queue_index].stats;
440 			accum += *(unsigned long *)(vif_stats + xenvif_stats[i].offset);
441 		}
442 		data[i] = accum;
443 	}
444 
445 	rcu_read_unlock();
446 }
447 
448 static void xenvif_get_strings(struct net_device *dev, u32 stringset, u8 * data)
449 {
450 	int i;
451 
452 	switch (stringset) {
453 	case ETH_SS_STATS:
454 		for (i = 0; i < ARRAY_SIZE(xenvif_stats); i++)
455 			memcpy(data + i * ETH_GSTRING_LEN,
456 			       xenvif_stats[i].name, ETH_GSTRING_LEN);
457 		break;
458 	}
459 }
460 
461 static const struct ethtool_ops xenvif_ethtool_ops = {
462 	.get_link	= ethtool_op_get_link,
463 
464 	.get_sset_count = xenvif_get_sset_count,
465 	.get_ethtool_stats = xenvif_get_ethtool_stats,
466 	.get_strings = xenvif_get_strings,
467 };
468 
469 static const struct net_device_ops xenvif_netdev_ops = {
470 	.ndo_select_queue = xenvif_select_queue,
471 	.ndo_start_xmit	= xenvif_start_xmit,
472 	.ndo_get_stats	= xenvif_get_stats,
473 	.ndo_open	= xenvif_open,
474 	.ndo_stop	= xenvif_close,
475 	.ndo_change_mtu	= xenvif_change_mtu,
476 	.ndo_fix_features = xenvif_fix_features,
477 	.ndo_set_mac_address = eth_mac_addr,
478 	.ndo_validate_addr   = eth_validate_addr,
479 };
480 
481 struct xenvif *xenvif_alloc(struct device *parent, domid_t domid,
482 			    unsigned int handle)
483 {
484 	static const u8 dummy_addr[ETH_ALEN] = {
485 		0xfe, 0xff, 0xff, 0xff, 0xff, 0xff,
486 	};
487 	int err;
488 	struct net_device *dev;
489 	struct xenvif *vif;
490 	char name[IFNAMSIZ] = {};
491 
492 	snprintf(name, IFNAMSIZ - 1, "vif%u.%u", domid, handle);
493 	/* Allocate a netdev with the max. supported number of queues.
494 	 * When the guest selects the desired number, it will be updated
495 	 * via netif_set_real_num_*_queues().
496 	 */
497 	dev = alloc_netdev_mq(sizeof(struct xenvif), name, NET_NAME_UNKNOWN,
498 			      ether_setup, xenvif_max_queues);
499 	if (dev == NULL) {
500 		pr_warn("Could not allocate netdev for %s\n", name);
501 		return ERR_PTR(-ENOMEM);
502 	}
503 
504 	SET_NETDEV_DEV(dev, parent);
505 
506 	vif = netdev_priv(dev);
507 
508 	vif->domid  = domid;
509 	vif->handle = handle;
510 	vif->can_sg = 1;
511 	vif->ip_csum = 1;
512 	vif->dev = dev;
513 	vif->disabled = false;
514 	vif->drain_timeout = msecs_to_jiffies(rx_drain_timeout_msecs);
515 	vif->stall_timeout = msecs_to_jiffies(rx_stall_timeout_msecs);
516 
517 	/* Start out with no queues. */
518 	vif->queues = NULL;
519 	vif->num_queues = 0;
520 
521 	vif->xdp_headroom = 0;
522 
523 	spin_lock_init(&vif->lock);
524 	INIT_LIST_HEAD(&vif->fe_mcast_addr);
525 
526 	dev->netdev_ops	= &xenvif_netdev_ops;
527 	dev->hw_features = NETIF_F_SG |
528 		NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM |
529 		NETIF_F_TSO | NETIF_F_TSO6 | NETIF_F_FRAGLIST;
530 	dev->features = dev->hw_features | NETIF_F_RXCSUM;
531 	dev->ethtool_ops = &xenvif_ethtool_ops;
532 
533 	dev->tx_queue_len = XENVIF_QUEUE_LENGTH;
534 
535 	dev->min_mtu = ETH_MIN_MTU;
536 	dev->max_mtu = ETH_MAX_MTU - VLAN_ETH_HLEN;
537 
538 	/*
539 	 * Initialise a dummy MAC address. We choose the numerically
540 	 * largest non-broadcast address to prevent the address getting
541 	 * stolen by an Ethernet bridge for STP purposes.
542 	 * (FE:FF:FF:FF:FF:FF)
543 	 */
544 	eth_hw_addr_set(dev, dummy_addr);
545 
546 	netif_carrier_off(dev);
547 
548 	err = register_netdev(dev);
549 	if (err) {
550 		netdev_warn(dev, "Could not register device: err=%d\n", err);
551 		free_netdev(dev);
552 		return ERR_PTR(err);
553 	}
554 
555 	netdev_dbg(dev, "Successfully created xenvif\n");
556 
557 	__module_get(THIS_MODULE);
558 
559 	return vif;
560 }
561 
562 int xenvif_init_queue(struct xenvif_queue *queue)
563 {
564 	int err, i;
565 
566 	queue->credit_bytes = queue->remaining_credit = ~0UL;
567 	queue->credit_usec  = 0UL;
568 	timer_setup(&queue->credit_timeout, xenvif_tx_credit_callback, 0);
569 	queue->credit_window_start = get_jiffies_64();
570 
571 	queue->rx_queue_max = XENVIF_RX_QUEUE_BYTES;
572 
573 	skb_queue_head_init(&queue->rx_queue);
574 	skb_queue_head_init(&queue->tx_queue);
575 
576 	queue->pending_cons = 0;
577 	queue->pending_prod = MAX_PENDING_REQS;
578 	for (i = 0; i < MAX_PENDING_REQS; ++i)
579 		queue->pending_ring[i] = i;
580 
581 	spin_lock_init(&queue->callback_lock);
582 	spin_lock_init(&queue->response_lock);
583 
584 	/* If ballooning is disabled, this will consume real memory, so you
585 	 * better enable it. The long term solution would be to use just a
586 	 * bunch of valid page descriptors, without dependency on ballooning
587 	 */
588 	err = gnttab_alloc_pages(MAX_PENDING_REQS,
589 				 queue->mmap_pages);
590 	if (err) {
591 		netdev_err(queue->vif->dev, "Could not reserve mmap_pages\n");
592 		return -ENOMEM;
593 	}
594 
595 	for (i = 0; i < MAX_PENDING_REQS; i++) {
596 		queue->pending_tx_info[i].callback_struct = (struct ubuf_info_msgzc)
597 			{ { .callback = xenvif_zerocopy_callback },
598 			  { { .ctx = NULL,
599 			      .desc = i } } };
600 		queue->grant_tx_handle[i] = NETBACK_INVALID_HANDLE;
601 	}
602 
603 	return 0;
604 }
605 
606 void xenvif_carrier_on(struct xenvif *vif)
607 {
608 	rtnl_lock();
609 	if (!vif->can_sg && vif->dev->mtu > ETH_DATA_LEN)
610 		dev_set_mtu(vif->dev, ETH_DATA_LEN);
611 	netdev_update_features(vif->dev);
612 	set_bit(VIF_STATUS_CONNECTED, &vif->status);
613 	if (netif_running(vif->dev))
614 		xenvif_up(vif);
615 	rtnl_unlock();
616 }
617 
618 int xenvif_connect_ctrl(struct xenvif *vif, grant_ref_t ring_ref,
619 			unsigned int evtchn)
620 {
621 	struct net_device *dev = vif->dev;
622 	struct xenbus_device *xendev = xenvif_to_xenbus_device(vif);
623 	void *addr;
624 	struct xen_netif_ctrl_sring *shared;
625 	RING_IDX rsp_prod, req_prod;
626 	int err;
627 
628 	err = xenbus_map_ring_valloc(xendev, &ring_ref, 1, &addr);
629 	if (err)
630 		goto err;
631 
632 	shared = (struct xen_netif_ctrl_sring *)addr;
633 	rsp_prod = READ_ONCE(shared->rsp_prod);
634 	req_prod = READ_ONCE(shared->req_prod);
635 
636 	BACK_RING_ATTACH(&vif->ctrl, shared, rsp_prod, XEN_PAGE_SIZE);
637 
638 	err = -EIO;
639 	if (req_prod - rsp_prod > RING_SIZE(&vif->ctrl))
640 		goto err_unmap;
641 
642 	err = bind_interdomain_evtchn_to_irq_lateeoi(xendev, evtchn);
643 	if (err < 0)
644 		goto err_unmap;
645 
646 	vif->ctrl_irq = err;
647 
648 	xenvif_init_hash(vif);
649 
650 	err = request_threaded_irq(vif->ctrl_irq, NULL, xenvif_ctrl_irq_fn,
651 				   IRQF_ONESHOT, "xen-netback-ctrl", vif);
652 	if (err) {
653 		pr_warn("Could not setup irq handler for %s\n", dev->name);
654 		goto err_deinit;
655 	}
656 
657 	return 0;
658 
659 err_deinit:
660 	xenvif_deinit_hash(vif);
661 	unbind_from_irqhandler(vif->ctrl_irq, vif);
662 	vif->ctrl_irq = 0;
663 
664 err_unmap:
665 	xenbus_unmap_ring_vfree(xendev, vif->ctrl.sring);
666 	vif->ctrl.sring = NULL;
667 
668 err:
669 	return err;
670 }
671 
672 static void xenvif_disconnect_queue(struct xenvif_queue *queue)
673 {
674 	if (queue->task) {
675 		kthread_stop(queue->task);
676 		put_task_struct(queue->task);
677 		queue->task = NULL;
678 	}
679 
680 	if (queue->dealloc_task) {
681 		kthread_stop(queue->dealloc_task);
682 		queue->dealloc_task = NULL;
683 	}
684 
685 	if (queue->napi.poll) {
686 		netif_napi_del(&queue->napi);
687 		queue->napi.poll = NULL;
688 	}
689 
690 	if (queue->tx_irq) {
691 		unbind_from_irqhandler(queue->tx_irq, queue);
692 		if (queue->tx_irq == queue->rx_irq)
693 			queue->rx_irq = 0;
694 		queue->tx_irq = 0;
695 	}
696 
697 	if (queue->rx_irq) {
698 		unbind_from_irqhandler(queue->rx_irq, queue);
699 		queue->rx_irq = 0;
700 	}
701 
702 	xenvif_unmap_frontend_data_rings(queue);
703 }
704 
705 int xenvif_connect_data(struct xenvif_queue *queue,
706 			unsigned long tx_ring_ref,
707 			unsigned long rx_ring_ref,
708 			unsigned int tx_evtchn,
709 			unsigned int rx_evtchn)
710 {
711 	struct xenbus_device *dev = xenvif_to_xenbus_device(queue->vif);
712 	struct task_struct *task;
713 	int err;
714 
715 	BUG_ON(queue->tx_irq);
716 	BUG_ON(queue->task);
717 	BUG_ON(queue->dealloc_task);
718 
719 	err = xenvif_map_frontend_data_rings(queue, tx_ring_ref,
720 					     rx_ring_ref);
721 	if (err < 0)
722 		goto err;
723 
724 	init_waitqueue_head(&queue->wq);
725 	init_waitqueue_head(&queue->dealloc_wq);
726 	atomic_set(&queue->inflight_packets, 0);
727 
728 	netif_napi_add(queue->vif->dev, &queue->napi, xenvif_poll);
729 
730 	queue->stalled = true;
731 
732 	task = kthread_run(xenvif_kthread_guest_rx, queue,
733 			   "%s-guest-rx", queue->name);
734 	if (IS_ERR(task))
735 		goto kthread_err;
736 	queue->task = task;
737 	/*
738 	 * Take a reference to the task in order to prevent it from being freed
739 	 * if the thread function returns before kthread_stop is called.
740 	 */
741 	get_task_struct(task);
742 
743 	task = kthread_run(xenvif_dealloc_kthread, queue,
744 			   "%s-dealloc", queue->name);
745 	if (IS_ERR(task))
746 		goto kthread_err;
747 	queue->dealloc_task = task;
748 
749 	if (tx_evtchn == rx_evtchn) {
750 		/* feature-split-event-channels == 0 */
751 		err = bind_interdomain_evtchn_to_irqhandler_lateeoi(
752 			dev, tx_evtchn, xenvif_interrupt, 0,
753 			queue->name, queue);
754 		if (err < 0)
755 			goto err;
756 		queue->tx_irq = queue->rx_irq = err;
757 		disable_irq(queue->tx_irq);
758 	} else {
759 		/* feature-split-event-channels == 1 */
760 		snprintf(queue->tx_irq_name, sizeof(queue->tx_irq_name),
761 			 "%s-tx", queue->name);
762 		err = bind_interdomain_evtchn_to_irqhandler_lateeoi(
763 			dev, tx_evtchn, xenvif_tx_interrupt, 0,
764 			queue->tx_irq_name, queue);
765 		if (err < 0)
766 			goto err;
767 		queue->tx_irq = err;
768 		disable_irq(queue->tx_irq);
769 
770 		snprintf(queue->rx_irq_name, sizeof(queue->rx_irq_name),
771 			 "%s-rx", queue->name);
772 		err = bind_interdomain_evtchn_to_irqhandler_lateeoi(
773 			dev, rx_evtchn, xenvif_rx_interrupt, 0,
774 			queue->rx_irq_name, queue);
775 		if (err < 0)
776 			goto err;
777 		queue->rx_irq = err;
778 		disable_irq(queue->rx_irq);
779 	}
780 
781 	return 0;
782 
783 kthread_err:
784 	pr_warn("Could not allocate kthread for %s\n", queue->name);
785 	err = PTR_ERR(task);
786 err:
787 	xenvif_disconnect_queue(queue);
788 	return err;
789 }
790 
791 void xenvif_carrier_off(struct xenvif *vif)
792 {
793 	struct net_device *dev = vif->dev;
794 
795 	rtnl_lock();
796 	if (test_and_clear_bit(VIF_STATUS_CONNECTED, &vif->status)) {
797 		netif_carrier_off(dev); /* discard queued packets */
798 		if (netif_running(dev))
799 			xenvif_down(vif);
800 	}
801 	rtnl_unlock();
802 }
803 
804 void xenvif_disconnect_data(struct xenvif *vif)
805 {
806 	struct xenvif_queue *queue = NULL;
807 	unsigned int num_queues = vif->num_queues;
808 	unsigned int queue_index;
809 
810 	xenvif_carrier_off(vif);
811 
812 	for (queue_index = 0; queue_index < num_queues; ++queue_index) {
813 		queue = &vif->queues[queue_index];
814 
815 		xenvif_disconnect_queue(queue);
816 	}
817 
818 	xenvif_mcast_addr_list_free(vif);
819 }
820 
821 void xenvif_disconnect_ctrl(struct xenvif *vif)
822 {
823 	if (vif->ctrl_irq) {
824 		xenvif_deinit_hash(vif);
825 		unbind_from_irqhandler(vif->ctrl_irq, vif);
826 		vif->ctrl_irq = 0;
827 	}
828 
829 	if (vif->ctrl.sring) {
830 		xenbus_unmap_ring_vfree(xenvif_to_xenbus_device(vif),
831 					vif->ctrl.sring);
832 		vif->ctrl.sring = NULL;
833 	}
834 }
835 
836 /* Reverse the relevant parts of xenvif_init_queue().
837  * Used for queue teardown from xenvif_free(), and on the
838  * error handling paths in xenbus.c:connect().
839  */
840 void xenvif_deinit_queue(struct xenvif_queue *queue)
841 {
842 	gnttab_free_pages(MAX_PENDING_REQS, queue->mmap_pages);
843 }
844 
845 void xenvif_free(struct xenvif *vif)
846 {
847 	struct xenvif_queue *queues = vif->queues;
848 	unsigned int num_queues = vif->num_queues;
849 	unsigned int queue_index;
850 
851 	unregister_netdev(vif->dev);
852 	free_netdev(vif->dev);
853 
854 	for (queue_index = 0; queue_index < num_queues; ++queue_index)
855 		xenvif_deinit_queue(&queues[queue_index]);
856 	vfree(queues);
857 
858 	module_put(THIS_MODULE);
859 }
860