netvsc_drv.c 14.0 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13
/*
 * Copyright (c) 2009, Microsoft Corporation.
 *
 * This program is free software; you can redistribute it and/or modify it
 * under the terms and conditions of the GNU General Public License,
 * version 2, as published by the Free Software Foundation.
 *
 * This program is distributed in the hope it will be useful, but WITHOUT
 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
 * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
 * more details.
 *
 * You should have received a copy of the GNU General Public License along with
14
 * this program; if not, see <http://www.gnu.org/licenses/>.
15 16
 *
 * Authors:
17
 *   Haiyang Zhang <haiyangz@microsoft.com>
18 19
 *   Hank Janssen  <hjanssen@microsoft.com>
 */
20 21
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

22
#include <linux/init.h>
23
#include <linux/atomic.h>
24 25 26 27 28 29 30 31 32
#include <linux/module.h>
#include <linux/highmem.h>
#include <linux/device.h>
#include <linux/io.h>
#include <linux/delay.h>
#include <linux/netdevice.h>
#include <linux/inetdevice.h>
#include <linux/etherdevice.h>
#include <linux/skbuff.h>
33
#include <linux/if_vlan.h>
34
#include <linux/in.h>
35
#include <linux/slab.h>
36 37 38 39
#include <net/arp.h>
#include <net/route.h>
#include <net/sock.h>
#include <net/pkt_sched.h>
40

41
#include "hyperv_net.h"
42 43

struct net_device_context {
44
	/* point back to our device context */
45
	struct hv_device *device_ctx;
46
	struct delayed_work dwork;
47
	struct work_struct work;
48 49
};

50
#define RING_SIZE_MIN 64
51
static int ring_size = 128;
S
Stephen Hemminger 已提交
52 53
module_param(ring_size, int, S_IRUGO);
MODULE_PARM_DESC(ring_size, "Ring buffer size (# of pages)");
54

55 56
static void do_set_multicast(struct work_struct *w)
{
57 58
	struct net_device_context *ndevctx =
		container_of(w, struct net_device_context, work);
59 60 61 62
	struct netvsc_device *nvdev;
	struct rndis_device *rdev;

	nvdev = hv_get_drvdata(ndevctx->device_ctx);
63 64
	if (nvdev == NULL || nvdev->ndev == NULL)
		return;
65 66 67

	rdev = nvdev->extension;
	if (rdev == NULL)
68
		return;
69

70
	if (nvdev->ndev->flags & IFF_PROMISC)
71 72 73 74 75 76 77 78 79
		rndis_filter_set_packet_filter(rdev,
			NDIS_PACKET_TYPE_PROMISCUOUS);
	else
		rndis_filter_set_packet_filter(rdev,
			NDIS_PACKET_TYPE_BROADCAST |
			NDIS_PACKET_TYPE_ALL_MULTICAST |
			NDIS_PACKET_TYPE_DIRECTED);
}

80
static void netvsc_set_multicast_list(struct net_device *net)
81
{
82
	struct net_device_context *net_device_ctx = netdev_priv(net);
83

84
	schedule_work(&net_device_ctx->work);
85 86 87 88 89
}

static int netvsc_open(struct net_device *net)
{
	struct net_device_context *net_device_ctx = netdev_priv(net);
90
	struct hv_device *device_obj = net_device_ctx->device_ctx;
91 92
	struct netvsc_device *nvdev;
	struct rndis_device *rdev;
93
	int ret = 0;
94

95 96
	netif_carrier_off(net);

97 98 99 100 101
	/* Open up the device */
	ret = rndis_filter_open(device_obj);
	if (ret != 0) {
		netdev_err(net, "unable to open device (ret %d).\n", ret);
		return ret;
102 103
	}

104 105
	netif_start_queue(net);

106 107 108 109 110
	nvdev = hv_get_drvdata(device_obj);
	rdev = nvdev->extension;
	if (!rdev->link_state)
		netif_carrier_on(net);

111 112 113 114 115 116
	return ret;
}

static int netvsc_close(struct net_device *net)
{
	struct net_device_context *net_device_ctx = netdev_priv(net);
117
	struct hv_device *device_obj = net_device_ctx->device_ctx;
118
	int ret;
119

120
	netif_tx_disable(net);
121

122 123
	/* Make sure netvsc_set_multicast_list doesn't re-enable filter! */
	cancel_work_sync(&net_device_ctx->work);
124
	ret = rndis_filter_close(device_obj);
125
	if (ret != 0)
126
		netdev_err(net, "unable to close device (ret %d).\n", ret);
127 128 129 130 131 132

	return ret;
}

static void netvsc_xmit_completion(void *context)
{
133
	struct hv_netvsc_packet *packet = (struct hv_netvsc_packet *)context;
134
	struct sk_buff *skb = (struct sk_buff *)
135
		(unsigned long)packet->completion.send.send_completion_tid;
136 137 138

	kfree(packet);

139
	if (skb)
140
		dev_kfree_skb_any(skb);
141 142
}

143
static int netvsc_start_xmit(struct sk_buff *skb, struct net_device *net)
144 145
{
	struct net_device_context *net_device_ctx = netdev_priv(net);
146
	struct hv_netvsc_packet *packet;
147
	int ret;
148
	unsigned int i, num_pages, npg_data;
149

150
	/* Add multipages for skb->data and additional 2 for RNDIS */
151 152
	npg_data = (((unsigned long)skb->data + skb_headlen(skb) - 1)
		>> PAGE_SHIFT) - ((unsigned long)skb->data >> PAGE_SHIFT) + 1;
153
	num_pages = skb_shinfo(skb)->nr_frags + npg_data + 2;
154

155
	/* Allocate a netvsc packet based on # of frags. */
156
	packet = kzalloc(sizeof(struct hv_netvsc_packet) +
157
			 (num_pages * sizeof(struct hv_page_buffer)) +
158 159
			 sizeof(struct rndis_filter_packet) +
			 NDIS_VLAN_PPI_SIZE, GFP_ATOMIC);
160
	if (!packet) {
161
		/* out of memory, drop packet */
162
		netdev_err(net, "unable to allocate hv_netvsc_packet\n");
163 164 165

		dev_kfree_skb(skb);
		net->stats.tx_dropped++;
166
		return NETDEV_TX_OK;
167 168
	}

169 170
	packet->vlan_tci = skb->vlan_tci;

171
	packet->extension = (void *)(unsigned long)packet +
172
				sizeof(struct hv_netvsc_packet) +
173
				    (num_pages * sizeof(struct hv_page_buffer));
174

175 176
	/* If the rndis msg goes beyond 1 page, we will add 1 later */
	packet->page_buf_cnt = num_pages - 1;
177

178
	/* Initialize it from the skb */
179
	packet->total_data_buflen = skb->len;
180

181
	/* Start filling in the page buffers starting after RNDIS buffer. */
182 183
	packet->page_buf[1].pfn = virt_to_phys(skb->data) >> PAGE_SHIFT;
	packet->page_buf[1].offset
184
		= (unsigned long)skb->data & (PAGE_SIZE - 1);
185 186 187 188 189 190 191 192 193 194 195 196 197 198 199
	if (npg_data == 1)
		packet->page_buf[1].len = skb_headlen(skb);
	else
		packet->page_buf[1].len = PAGE_SIZE
			- packet->page_buf[1].offset;

	for (i = 2; i <= npg_data; i++) {
		packet->page_buf[i].pfn = virt_to_phys(skb->data
			+ PAGE_SIZE * (i-1)) >> PAGE_SHIFT;
		packet->page_buf[i].offset = 0;
		packet->page_buf[i].len = PAGE_SIZE;
	}
	if (npg_data > 1)
		packet->page_buf[npg_data].len = (((unsigned long)skb->data
			+ skb_headlen(skb) - 1) & (PAGE_SIZE - 1)) + 1;
200 201 202

	/* Additional fragments are after SKB data */
	for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
E
Eric Dumazet 已提交
203
		const skb_frag_t *f = &skb_shinfo(skb)->frags[i];
204

205 206 207 208
		packet->page_buf[i+npg_data+1].pfn =
			page_to_pfn(skb_frag_page(f));
		packet->page_buf[i+npg_data+1].offset = f->page_offset;
		packet->page_buf[i+npg_data+1].len = skb_frag_size(f);
209 210
	}

211
	/* Set the completion routine */
212 213 214
	packet->completion.send.send_completion = netvsc_xmit_completion;
	packet->completion.send.send_completion_ctx = packet;
	packet->completion.send.send_completion_tid = (unsigned long)skb;
215

216
	ret = rndis_filter_send(net_device_ctx->device_ctx,
217 218
				  packet);
	if (ret == 0) {
219 220
		net->stats.tx_bytes += skb->len;
		net->stats.tx_packets++;
221
	} else {
222
		kfree(packet);
223 224 225 226
		if (ret != -EAGAIN) {
			dev_kfree_skb_any(skb);
			net->stats.tx_dropped++;
		}
227 228
	}

229
	return (ret == -EAGAIN) ? NETDEV_TX_BUSY : NETDEV_TX_OK;
230 231
}

232
/*
233 234
 * netvsc_linkstatus_callback - Link up/down notification
 */
235
void netvsc_linkstatus_callback(struct hv_device *device_obj,
236
				       unsigned int status)
237
{
238
	struct net_device *net;
239
	struct net_device_context *ndev_ctx;
240
	struct netvsc_device *net_device;
241
	struct rndis_device *rdev;
242 243

	net_device = hv_get_drvdata(device_obj);
244 245 246 247
	rdev = net_device->extension;

	rdev->link_state = status != 1;

248
	net = net_device->ndev;
249

250
	if (!net || net->reg_state != NETREG_REGISTERED)
251 252
		return;

253
	ndev_ctx = netdev_priv(net);
254
	if (status == 1) {
255
		schedule_delayed_work(&ndev_ctx->dwork, 0);
256
		schedule_delayed_work(&ndev_ctx->dwork, msecs_to_jiffies(20));
257
	} else {
258
		schedule_delayed_work(&ndev_ctx->dwork, 0);
259 260 261
	}
}

262 263 264
/*
 * netvsc_recv_callback -  Callback when we receive a packet from the
 * "wire" on the specified device.
265
 */
266
int netvsc_recv_callback(struct hv_device *device_obj,
267
				struct hv_netvsc_packet *packet)
268
{
269
	struct net_device *net;
270 271
	struct sk_buff *skb;

272
	net = ((struct netvsc_device *)hv_get_drvdata(device_obj))->ndev;
273
	if (!net || net->reg_state != NETREG_REGISTERED) {
274
		packet->status = NVSP_STAT_FAIL;
275 276 277
		return 0;
	}

278
	/* Allocate a skb - TODO direct I/O to pages? */
279
	skb = netdev_alloc_skb_ip_align(net, packet->total_data_buflen);
280 281
	if (unlikely(!skb)) {
		++net->stats.rx_dropped;
282
		packet->status = NVSP_STAT_FAIL;
283 284
		return 0;
	}
285

286 287 288 289
	/*
	 * Copy to skb. This copy is needed here since the memory pointed by
	 * hv_netvsc_packet cannot be deallocated
	 */
290 291
	memcpy(skb_put(skb, packet->total_data_buflen), packet->data,
		packet->total_data_buflen);
292 293 294

	skb->protocol = eth_type_trans(skb, net);
	skb->ip_summed = CHECKSUM_NONE;
295 296 297
	if (packet->vlan_tci & VLAN_TAG_PRESENT)
		__vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q),
				       packet->vlan_tci);
298

299
	net->stats.rx_packets++;
300
	net->stats.rx_bytes += packet->total_data_buflen;
301

302 303
	/*
	 * Pass the skb back up. Network stack will deallocate the skb when it
304 305
	 * is done.
	 * TODO - use NAPI?
306
	 */
307
	netif_rx(skb);
308 309 310 311

	return 0;
}

312 313 314
static void netvsc_get_drvinfo(struct net_device *net,
			       struct ethtool_drvinfo *info)
{
315 316
	strlcpy(info->driver, KBUILD_MODNAME, sizeof(info->driver));
	strlcpy(info->fw_version, "N/A", sizeof(info->fw_version));
317 318
}

319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336
static int netvsc_change_mtu(struct net_device *ndev, int mtu)
{
	struct net_device_context *ndevctx = netdev_priv(ndev);
	struct hv_device *hdev =  ndevctx->device_ctx;
	struct netvsc_device *nvdev = hv_get_drvdata(hdev);
	struct netvsc_device_info device_info;
	int limit = ETH_DATA_LEN;

	if (nvdev == NULL || nvdev->destroy)
		return -ENODEV;

	if (nvdev->nvsp_version == NVSP_PROTOCOL_VERSION_2)
		limit = NETVSC_MTU;

	if (mtu < 68 || mtu > limit)
		return -EINVAL;

	nvdev->start_remove = true;
337
	cancel_work_sync(&ndevctx->work);
338
	netif_tx_disable(ndev);
339 340 341 342 343 344 345 346 347 348 349 350 351
	rndis_filter_device_remove(hdev);

	ndev->mtu = mtu;

	ndevctx->device_ctx = hdev;
	hv_set_drvdata(hdev, ndev);
	device_info.ring_size = ring_size;
	rndis_filter_device_add(hdev, &device_info);
	netif_wake_queue(ndev);

	return 0;
}

352 353 354 355 356 357

static int netvsc_set_mac_addr(struct net_device *ndev, void *p)
{
	struct net_device_context *ndevctx = netdev_priv(ndev);
	struct hv_device *hdev =  ndevctx->device_ctx;
	struct sockaddr *addr = p;
358
	char save_adr[ETH_ALEN];
359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379
	unsigned char save_aatype;
	int err;

	memcpy(save_adr, ndev->dev_addr, ETH_ALEN);
	save_aatype = ndev->addr_assign_type;

	err = eth_mac_addr(ndev, p);
	if (err != 0)
		return err;

	err = rndis_filter_set_device_mac(hdev, addr->sa_data);
	if (err != 0) {
		/* roll back to saved MAC */
		memcpy(ndev->dev_addr, save_adr, ETH_ALEN);
		ndev->addr_assign_type = save_aatype;
	}

	return err;
}


380 381 382 383 384
static const struct ethtool_ops ethtool_ops = {
	.get_drvinfo	= netvsc_get_drvinfo,
	.get_link	= ethtool_op_get_link,
};

385 386 387 388
static const struct net_device_ops device_ops = {
	.ndo_open =			netvsc_open,
	.ndo_stop =			netvsc_close,
	.ndo_start_xmit =		netvsc_start_xmit,
389
	.ndo_set_rx_mode =		netvsc_set_multicast_list,
390
	.ndo_change_mtu =		netvsc_change_mtu,
391
	.ndo_validate_addr =		eth_validate_addr,
392
	.ndo_set_mac_address =		netvsc_set_mac_addr,
393 394
};

395 396 397 398
/*
 * Send GARP packet to network peers after migrations.
 * After Quick Migration, the network is not immediately operational in the
 * current context when receiving RNDIS_STATUS_MEDIA_CONNECT event. So, add
399
 * another netif_notify_peers() into a delayed work, otherwise GARP packet
400
 * will not be sent after quick migration, and cause network disconnection.
401
 * Also, we update the carrier status here.
402
 */
403
static void netvsc_link_change(struct work_struct *w)
404 405 406
{
	struct net_device_context *ndev_ctx;
	struct net_device *net;
407
	struct netvsc_device *net_device;
408 409 410 411
	struct rndis_device *rdev;
	bool notify;

	rtnl_lock();
412

413
	ndev_ctx = container_of(w, struct net_device_context, dwork.work);
414
	net_device = hv_get_drvdata(ndev_ctx->device_ctx);
415
	rdev = net_device->extension;
416
	net = net_device->ndev;
417 418 419 420 421 422 423 424 425 426 427 428 429

	if (rdev->link_state) {
		netif_carrier_off(net);
		notify = false;
	} else {
		netif_carrier_on(net);
		notify = true;
	}

	rtnl_unlock();

	if (notify)
		netdev_notify_peers(net);
430 431 432
}


433 434
static int netvsc_probe(struct hv_device *dev,
			const struct hv_vmbus_device_id *dev_id)
435 436 437 438 439 440
{
	struct net_device *net = NULL;
	struct net_device_context *net_device_ctx;
	struct netvsc_device_info device_info;
	int ret;

441
	net = alloc_etherdev(sizeof(struct net_device_context));
442
	if (!net)
443
		return -ENOMEM;
444

445 446
	netif_carrier_off(net);

447
	net_device_ctx = netdev_priv(net);
448
	net_device_ctx->device_ctx = dev;
449
	hv_set_drvdata(dev, net);
450
	INIT_DELAYED_WORK(&net_device_ctx->dwork, netvsc_link_change);
451
	INIT_WORK(&net_device_ctx->work, do_set_multicast);
452 453 454

	net->netdev_ops = &device_ops;

455
	/* TODO: Add GSO and Checksum offload */
456 457
	net->hw_features = 0;
	net->features = NETIF_F_HW_VLAN_CTAG_TX;
458

459
	SET_ETHTOOL_OPS(net, &ethtool_ops);
460
	SET_NETDEV_DEV(net, &dev->device);
461

462 463 464 465 466
	/* Notify the netvsc driver of the new device */
	device_info.ring_size = ring_size;
	ret = rndis_filter_device_add(dev, &device_info);
	if (ret != 0) {
		netdev_err(net, "unable to add netvsc device (ret %d)\n", ret);
467
		free_netdev(net);
468
		hv_set_drvdata(dev, NULL);
469
		return ret;
470
	}
471 472
	memcpy(net->dev_addr, device_info.mac_adr, ETH_ALEN);

473 474 475 476 477
	ret = register_netdev(net);
	if (ret != 0) {
		pr_err("Unable to register netdev.\n");
		rndis_filter_device_remove(dev);
		free_netdev(net);
478 479
	} else {
		schedule_delayed_work(&net_device_ctx->dwork, 0);
480 481
	}

482 483 484
	return ret;
}

485
static int netvsc_remove(struct hv_device *dev)
486
{
487
	struct net_device *net;
488
	struct net_device_context *ndev_ctx;
489 490 491 492
	struct netvsc_device *net_device;

	net_device = hv_get_drvdata(dev);
	net = net_device->ndev;
493 494

	if (net == NULL) {
495
		dev_err(&dev->device, "No net device to remove\n");
496 497 498
		return 0;
	}

499 500
	net_device->start_remove = true;

501 502
	ndev_ctx = netdev_priv(net);
	cancel_delayed_work_sync(&ndev_ctx->dwork);
503
	cancel_work_sync(&ndev_ctx->work);
504

505
	/* Stop outbound asap */
506
	netif_tx_disable(net);
507 508 509 510 511 512 513

	unregister_netdev(net);

	/*
	 * Call to the vsc driver to let it know that the device is being
	 * removed
	 */
514
	rndis_filter_device_remove(dev);
515 516

	free_netdev(net);
517
	return 0;
518 519
}

520
static const struct hv_vmbus_device_id id_table[] = {
521
	/* Network guid */
522
	{ HV_NIC_GUID, },
523
	{ },
524 525 526 527
};

MODULE_DEVICE_TABLE(vmbus, id_table);

528
/* The one and only one */
529
static struct  hv_driver netvsc_drv = {
530
	.name = KBUILD_MODNAME,
531
	.id_table = id_table,
532 533
	.probe = netvsc_probe,
	.remove = netvsc_remove,
534
};
535

536
static void __exit netvsc_drv_exit(void)
537
{
538
	vmbus_driver_unregister(&netvsc_drv);
539 540
}

541
static int __init netvsc_drv_init(void)
542
{
543 544 545 546 547
	if (ring_size < RING_SIZE_MIN) {
		ring_size = RING_SIZE_MIN;
		pr_info("Increased ring_size to %d (min allowed)\n",
			ring_size);
	}
548
	return vmbus_driver_register(&netvsc_drv);
549 550
}

551
MODULE_LICENSE("GPL");
552
MODULE_DESCRIPTION("Microsoft Hyper-V network driver");
553

554
module_init(netvsc_drv_init);
555
module_exit(netvsc_drv_exit);