fcoe.c 60.6 KB
Newer Older
1
/*
2
 * Copyright(c) 2007 - 2009 Intel Corporation. All rights reserved.
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
 *
 * 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
 * this program; if not, write to the Free Software Foundation, Inc.,
 * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
 *
 * Maintained at www.Open-FCoE.org
 */

#include <linux/module.h>
#include <linux/version.h>
#include <linux/spinlock.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/ethtool.h>
#include <linux/if_ether.h>
#include <linux/if_vlan.h>
#include <linux/crc32.h>
#include <linux/cpu.h>
#include <linux/fs.h>
#include <linux/sysfs.h>
#include <linux/ctype.h>
#include <scsi/scsi_tcq.h>
#include <scsi/scsicam.h>
#include <scsi/scsi_transport.h>
#include <scsi/scsi_transport_fc.h>
#include <net/rtnetlink.h>

#include <scsi/fc/fc_encaps.h>
40
#include <scsi/fc/fc_fip.h>
41 42 43 44 45

#include <scsi/libfc.h>
#include <scsi/fc_frame.h>
#include <scsi/libfcoe.h>

46
#include "fcoe.h"
47

48 49
MODULE_AUTHOR("Open-FCoE.org");
MODULE_DESCRIPTION("FCoE");
50
MODULE_LICENSE("GPL v2");
51

52 53 54 55 56 57
/* Performance tuning parameters for fcoe */
static unsigned int fcoe_ddp_min;
module_param_named(ddp_min, fcoe_ddp_min, uint, S_IRUGO | S_IWUSR);
MODULE_PARM_DESC(ddp_min, "Minimum I/O size in bytes for "	\
		 "Direct Data Placement (DDP).");

58 59
DEFINE_MUTEX(fcoe_config_mutex);

60 61 62
/* fcoe_percpu_clean completion.  Waiter protected by fcoe_create_mutex */
static DECLARE_COMPLETION(fcoe_flush_completion);

63
/* fcoe host list */
64
/* must only by accessed under the RTNL mutex */
65
LIST_HEAD(fcoe_hostlist);
66
DEFINE_PER_CPU(struct fcoe_percpu_s, fcoe_percpu);
67

68
/* Function Prototypes */
69 70 71 72 73 74 75 76 77 78 79 80
static int fcoe_reset(struct Scsi_Host *shost);
static int fcoe_xmit(struct fc_lport *, struct fc_frame *);
static int fcoe_rcv(struct sk_buff *, struct net_device *,
		    struct packet_type *, struct net_device *);
static int fcoe_percpu_receive_thread(void *arg);
static void fcoe_clean_pending_queue(struct fc_lport *lp);
static void fcoe_percpu_clean(struct fc_lport *lp);
static int fcoe_link_ok(struct fc_lport *lp);

static struct fc_lport *fcoe_hostlist_lookup(const struct net_device *);
static int fcoe_hostlist_add(const struct fc_lport *);

81
static void fcoe_check_wait_queue(struct fc_lport *, struct sk_buff *);
82 83 84
static int fcoe_device_notification(struct notifier_block *, ulong, void *);
static void fcoe_dev_setup(void);
static void fcoe_dev_cleanup(void);
85 86
static struct fcoe_interface *
	fcoe_hostlist_lookup_port(const struct net_device *dev);
87 88 89 90 91 92

/* notification function from net device */
static struct notifier_block fcoe_notifier = {
	.notifier_call = fcoe_device_notification,
};

93 94
static struct scsi_transport_template *fcoe_transport_template;
static struct scsi_transport_template *fcoe_vport_transport_template;
95

96 97 98
static int fcoe_vport_destroy(struct fc_vport *vport);
static int fcoe_vport_create(struct fc_vport *vport, bool disabled);
static int fcoe_vport_disable(struct fc_vport *vport, bool disable);
99
static void fcoe_set_vport_symbolic_name(struct fc_vport *vport);
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
struct fc_function_template fcoe_transport_function = {
	.show_host_node_name = 1,
	.show_host_port_name = 1,
	.show_host_supported_classes = 1,
	.show_host_supported_fc4s = 1,
	.show_host_active_fc4s = 1,
	.show_host_maxframe_size = 1,

	.show_host_port_id = 1,
	.show_host_supported_speeds = 1,
	.get_host_speed = fc_get_host_speed,
	.show_host_speed = 1,
	.show_host_port_type = 1,
	.get_host_port_state = fc_get_host_port_state,
	.show_host_port_state = 1,
	.show_host_symbolic_name = 1,

	.dd_fcrport_size = sizeof(struct fc_rport_libfc_priv),
	.show_rport_maxframe_size = 1,
	.show_rport_supported_classes = 1,

	.show_host_fabric_name = 1,
	.show_starget_node_name = 1,
	.show_starget_port_name = 1,
	.show_starget_port_id = 1,
	.set_rport_dev_loss_tmo = fc_set_rport_loss_tmo,
	.show_rport_dev_loss_tmo = 1,
	.get_fc_host_stats = fc_get_host_stats,
	.issue_fc_host_lip = fcoe_reset,

	.terminate_rport_io = fc_rport_terminate_io,
132 133 134 135

	.vport_create = fcoe_vport_create,
	.vport_delete = fcoe_vport_destroy,
	.vport_disable = fcoe_vport_disable,
136
	.set_vport_symbolic_name = fcoe_set_vport_symbolic_name,
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
struct fc_function_template fcoe_vport_transport_function = {
	.show_host_node_name = 1,
	.show_host_port_name = 1,
	.show_host_supported_classes = 1,
	.show_host_supported_fc4s = 1,
	.show_host_active_fc4s = 1,
	.show_host_maxframe_size = 1,

	.show_host_port_id = 1,
	.show_host_supported_speeds = 1,
	.get_host_speed = fc_get_host_speed,
	.show_host_speed = 1,
	.show_host_port_type = 1,
	.get_host_port_state = fc_get_host_port_state,
	.show_host_port_state = 1,
	.show_host_symbolic_name = 1,

	.dd_fcrport_size = sizeof(struct fc_rport_libfc_priv),
	.show_rport_maxframe_size = 1,
	.show_rport_supported_classes = 1,

	.show_host_fabric_name = 1,
	.show_starget_node_name = 1,
	.show_starget_port_name = 1,
	.show_starget_port_id = 1,
	.set_rport_dev_loss_tmo = fc_set_rport_loss_tmo,
	.show_rport_dev_loss_tmo = 1,
	.get_fc_host_stats = fc_get_host_stats,
	.issue_fc_host_lip = fcoe_reset,

	.terminate_rport_io = fc_rport_terminate_io,
};

172 173 174 175 176 177 178 179 180 181 182 183
static struct scsi_host_template fcoe_shost_template = {
	.module = THIS_MODULE,
	.name = "FCoE Driver",
	.proc_name = FCOE_NAME,
	.queuecommand = fc_queuecommand,
	.eh_abort_handler = fc_eh_abort,
	.eh_device_reset_handler = fc_eh_device_reset,
	.eh_host_reset_handler = fc_eh_host_reset,
	.slave_alloc = fc_slave_alloc,
	.change_queue_depth = fc_change_queue_depth,
	.change_queue_type = fc_change_queue_type,
	.this_id = -1,
184
	.cmd_per_lun = 3,
185 186 187 188 189 190
	.can_queue = FCOE_MAX_OUTSTANDING_COMMANDS,
	.use_clustering = ENABLE_CLUSTERING,
	.sg_tablesize = SG_ALL,
	.max_sectors = 0xffff,
};

191 192 193 194 195 196 197 198 199
static int fcoe_fip_recv(struct sk_buff *skb, struct net_device *dev,
			 struct packet_type *ptype,
			 struct net_device *orig_dev);
/**
 * fcoe_interface_setup()
 * @fcoe: new fcoe_interface
 * @netdev : ptr to the associated netdevice struct
 *
 * Returns : 0 for success
200
 * Locking: must be called with the RTNL mutex held
201 202 203 204 205 206 207
 */
static int fcoe_interface_setup(struct fcoe_interface *fcoe,
				struct net_device *netdev)
{
	struct fcoe_ctlr *fip = &fcoe->ctlr;
	struct netdev_hw_addr *ha;
	u8 flogi_maddr[ETH_ALEN];
208
	const struct net_device_ops *ops;
209 210 211

	fcoe->netdev = netdev;

212 213 214 215 216 217 218 219
	/* Let LLD initialize for FCoE */
	ops = netdev->netdev_ops;
	if (ops->ndo_fcoe_enable) {
		if (ops->ndo_fcoe_enable(netdev))
			FCOE_NETDEV_DBG(netdev, "Failed to enable FCoE"
					" specific feature for LLD.\n");
	}

220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272
	/* Do not support for bonding device */
	if ((netdev->priv_flags & IFF_MASTER_ALB) ||
	    (netdev->priv_flags & IFF_SLAVE_INACTIVE) ||
	    (netdev->priv_flags & IFF_MASTER_8023AD)) {
		return -EOPNOTSUPP;
	}

	/* look for SAN MAC address, if multiple SAN MACs exist, only
	 * use the first one for SPMA */
	rcu_read_lock();
	for_each_dev_addr(netdev, ha) {
		if ((ha->type == NETDEV_HW_ADDR_T_SAN) &&
		    (is_valid_ether_addr(fip->ctl_src_addr))) {
			memcpy(fip->ctl_src_addr, ha->addr, ETH_ALEN);
			fip->spma = 1;
			break;
		}
	}
	rcu_read_unlock();

	/* setup Source Mac Address */
	if (!fip->spma)
		memcpy(fip->ctl_src_addr, netdev->dev_addr, netdev->addr_len);

	/*
	 * Add FCoE MAC address as second unicast MAC address
	 * or enter promiscuous mode if not capable of listening
	 * for multiple unicast MACs.
	 */
	memcpy(flogi_maddr, (u8[6]) FC_FCOE_FLOGI_MAC, ETH_ALEN);
	dev_unicast_add(netdev, flogi_maddr);
	if (fip->spma)
		dev_unicast_add(netdev, fip->ctl_src_addr);
	dev_mc_add(netdev, FIP_ALL_ENODE_MACS, ETH_ALEN, 0);

	/*
	 * setup the receive function from ethernet driver
	 * on the ethertype for the given device
	 */
	fcoe->fcoe_packet_type.func = fcoe_rcv;
	fcoe->fcoe_packet_type.type = __constant_htons(ETH_P_FCOE);
	fcoe->fcoe_packet_type.dev = netdev;
	dev_add_pack(&fcoe->fcoe_packet_type);

	fcoe->fip_packet_type.func = fcoe_fip_recv;
	fcoe->fip_packet_type.type = htons(ETH_P_FIP);
	fcoe->fip_packet_type.dev = netdev;
	dev_add_pack(&fcoe->fip_packet_type);

	return 0;
}

static void fcoe_fip_send(struct fcoe_ctlr *fip, struct sk_buff *skb);
273 274
static void fcoe_update_src_mac(struct fc_lport *lport, u8 *addr);
static u8 *fcoe_get_src_mac(struct fc_lport *lport);
275
static void fcoe_destroy_work(struct work_struct *work);
276

277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292
/**
 * fcoe_interface_create()
 * @netdev: network interface
 *
 * Returns: pointer to a struct fcoe_interface or NULL on error
 */
static struct fcoe_interface *fcoe_interface_create(struct net_device *netdev)
{
	struct fcoe_interface *fcoe;

	fcoe = kzalloc(sizeof(*fcoe), GFP_KERNEL);
	if (!fcoe) {
		FCOE_NETDEV_DBG(netdev, "Could not allocate fcoe structure\n");
		return NULL;
	}

293
	dev_hold(netdev);
294
	kref_init(&fcoe->kref);
295 296 297 298 299 300 301

	/*
	 * Initialize FIP.
	 */
	fcoe_ctlr_init(&fcoe->ctlr);
	fcoe->ctlr.send = fcoe_fip_send;
	fcoe->ctlr.update_mac = fcoe_update_src_mac;
302
	fcoe->ctlr.get_src_addr = fcoe_get_src_mac;
303 304

	fcoe_interface_setup(fcoe, netdev);
305 306 307 308

	return fcoe;
}

309 310 311
/**
 * fcoe_interface_cleanup() - clean up netdev configurations
 * @fcoe:
312 313
 *
 * Caller must be holding the RTNL mutex
314 315 316 317 318 319
 */
void fcoe_interface_cleanup(struct fcoe_interface *fcoe)
{
	struct net_device *netdev = fcoe->netdev;
	struct fcoe_ctlr *fip = &fcoe->ctlr;
	u8 flogi_maddr[ETH_ALEN];
320
	const struct net_device_ops *ops;
321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337

	/*
	 * Don't listen for Ethernet packets anymore.
	 * synchronize_net() ensures that the packet handlers are not running
	 * on another CPU. dev_remove_pack() would do that, this calls the
	 * unsyncronized version __dev_remove_pack() to avoid multiple delays.
	 */
	__dev_remove_pack(&fcoe->fcoe_packet_type);
	__dev_remove_pack(&fcoe->fip_packet_type);
	synchronize_net();

	/* Delete secondary MAC addresses */
	memcpy(flogi_maddr, (u8[6]) FC_FCOE_FLOGI_MAC, ETH_ALEN);
	dev_unicast_delete(netdev, flogi_maddr);
	if (fip->spma)
		dev_unicast_delete(netdev, fip->ctl_src_addr);
	dev_mc_delete(netdev, FIP_ALL_ENODE_MACS, ETH_ALEN, 0);
338 339 340 341 342 343 344 345

	/* Tell the LLD we are done w/ FCoE */
	ops = netdev->netdev_ops;
	if (ops->ndo_fcoe_disable) {
		if (ops->ndo_fcoe_disable(netdev))
			FCOE_NETDEV_DBG(netdev, "Failed to disable FCoE"
					" specific feature for LLD.\n");
	}
346 347
}

348 349 350 351 352 353 354
/**
 * fcoe_interface_release() - fcoe_port kref release function
 * @kref: embedded reference count in an fcoe_interface struct
 */
static void fcoe_interface_release(struct kref *kref)
{
	struct fcoe_interface *fcoe;
355
	struct net_device *netdev;
356 357

	fcoe = container_of(kref, struct fcoe_interface, kref);
358 359 360
	netdev = fcoe->netdev;
	/* tear-down the FCoE controller */
	fcoe_ctlr_destroy(&fcoe->ctlr);
361
	kfree(fcoe);
362
	dev_put(netdev);
363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382
}

/**
 * fcoe_interface_get()
 * @fcoe:
 */
static inline void fcoe_interface_get(struct fcoe_interface *fcoe)
{
	kref_get(&fcoe->kref);
}

/**
 * fcoe_interface_put()
 * @fcoe:
 */
static inline void fcoe_interface_put(struct fcoe_interface *fcoe)
{
	kref_put(&fcoe->kref, fcoe_interface_release);
}

383 384 385 386 387 388 389 390 391 392 393 394 395
/**
 * fcoe_fip_recv - handle a received FIP frame.
 * @skb: the receive skb
 * @dev: associated &net_device
 * @ptype: the &packet_type structure which was used to register this handler.
 * @orig_dev: original receive &net_device, in case @dev is a bond.
 *
 * Returns: 0 for success
 */
static int fcoe_fip_recv(struct sk_buff *skb, struct net_device *dev,
			 struct packet_type *ptype,
			 struct net_device *orig_dev)
{
396
	struct fcoe_interface *fcoe;
397

398
	fcoe = container_of(ptype, struct fcoe_interface, fip_packet_type);
399
	fcoe_ctlr_recv(&fcoe->ctlr, skb);
400 401 402 403 404 405 406 407 408 409
	return 0;
}

/**
 * fcoe_fip_send() - send an Ethernet-encapsulated FIP frame.
 * @fip: FCoE controller.
 * @skb: FIP Packet.
 */
static void fcoe_fip_send(struct fcoe_ctlr *fip, struct sk_buff *skb)
{
410
	skb->dev = fcoe_from_ctlr(fip)->netdev;
411 412 413 414 415
	dev_queue_xmit(skb);
}

/**
 * fcoe_update_src_mac() - Update Ethernet MAC filters.
416 417
 * @lport: libfc lport
 * @addr: Unicast MAC address to add.
418 419 420 421
 *
 * Remove any previously-set unicast MAC filter.
 * Add secondary FCoE MAC address filter for our OUI.
 */
422
static void fcoe_update_src_mac(struct fc_lport *lport, u8 *addr)
423
{
424 425
	struct fcoe_port *port = lport_priv(lport);
	struct fcoe_interface *fcoe = port->fcoe;
426

427
	rtnl_lock();
428 429 430 431 432
	if (!is_zero_ether_addr(port->data_src_addr))
		dev_unicast_delete(fcoe->netdev, port->data_src_addr);
	if (!is_zero_ether_addr(addr))
		dev_unicast_add(fcoe->netdev, addr);
	memcpy(port->data_src_addr, addr, ETH_ALEN);
433 434 435
	rtnl_unlock();
}

436 437 438 439 440 441 442 443 444 445 446
/**
 * fcoe_get_src_mac() - return the Ethernet source address for an lport
 * @lport: libfc lport
 */
static u8 *fcoe_get_src_mac(struct fc_lport *lport)
{
	struct fcoe_port *port = lport_priv(lport);

	return port->data_src_addr;
}

447 448 449 450 451 452 453 454 455 456 457
/**
 * fcoe_lport_config() - sets up the fc_lport
 * @lp: ptr to the fc_lport
 *
 * Returns: 0 for success
 */
static int fcoe_lport_config(struct fc_lport *lp)
{
	lp->link_up = 0;
	lp->qfull = 0;
	lp->max_retry_count = 3;
458
	lp->max_rport_retry_count = 3;
459 460 461 462
	lp->e_d_tov = 2 * 1000;	/* FC-FS default */
	lp->r_a_tov = 2 * 2 * 1000;
	lp->service_params = (FCP_SPPF_INIT_FCN | FCP_SPPF_RD_XRDY_DIS |
			      FCP_SPPF_RETRY | FCP_SPPF_CONF_COMPL);
463
	lp->does_npiv = 1;
464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479

	fc_lport_init_stats(lp);

	/* lport fc_lport related configuration */
	fc_lport_config(lp);

	/* offload related configuration */
	lp->crc_offload = 0;
	lp->seq_offload = 0;
	lp->lro_enabled = 0;
	lp->lro_xid = 0;
	lp->lso_max = 0;

	return 0;
}

V
Vasu Dev 已提交
480 481 482 483 484 485 486 487 488 489 490 491
/**
 * fcoe_queue_timer() - fcoe queue timer
 * @lp: the fc_lport pointer
 *
 * Calls fcoe_check_wait_queue on timeout
 *
 */
static void fcoe_queue_timer(ulong lp)
{
	fcoe_check_wait_queue((struct fc_lport *)lp, NULL);
}

492 493 494 495 496 497 498 499 500 501 502 503 504
/**
 * fcoe_netdev_config() - Set up netdev for SW FCoE
 * @lp : ptr to the fc_lport
 * @netdev : ptr to the associated netdevice struct
 *
 * Must be called after fcoe_lport_config() as it will use lport mutex
 *
 * Returns : 0 for success
 */
static int fcoe_netdev_config(struct fc_lport *lp, struct net_device *netdev)
{
	u32 mfs;
	u64 wwnn, wwpn;
505
	struct fcoe_interface *fcoe;
506
	struct fcoe_port *port;
507 508

	/* Setup lport private data to point to fcoe softc */
509
	port = lport_priv(lp);
510
	fcoe = port->fcoe;
511 512 513 514 515 516

	/*
	 * Determine max frame size based on underlying device and optional
	 * user-configured limit.  If the MFS is too low, fcoe_link_ok()
	 * will return 0, so do this first.
	 */
517 518 519 520 521 522
	mfs = netdev->mtu;
	if (netdev->features & NETIF_F_FCOE_MTU) {
		mfs = FCOE_MTU;
		FCOE_NETDEV_DBG(netdev, "Supports FCOE_MTU of %d bytes\n", mfs);
	}
	mfs -= (sizeof(struct fcoe_hdr) + sizeof(struct fcoe_crc_eof));
523 524 525 526
	if (fc_set_mfs(lp, mfs))
		return -EINVAL;

	/* offload features support */
527
	if (netdev->features & NETIF_F_SG)
528 529 530 531
		lp->sg_supp = 1;

	if (netdev->features & NETIF_F_FCOE_CRC) {
		lp->crc_offload = 1;
532
		FCOE_NETDEV_DBG(netdev, "Supports FCCRC offload\n");
533 534 535 536
	}
	if (netdev->features & NETIF_F_FSO) {
		lp->seq_offload = 1;
		lp->lso_max = netdev->gso_max_size;
537 538
		FCOE_NETDEV_DBG(netdev, "Supports LSO for max len 0x%x\n",
				lp->lso_max);
539 540 541 542
	}
	if (netdev->fcoe_ddp_xid) {
		lp->lro_enabled = 1;
		lp->lro_xid = netdev->fcoe_ddp_xid;
543 544
		FCOE_NETDEV_DBG(netdev, "Supports LRO for max xid 0x%x\n",
				lp->lro_xid);
545
	}
546 547 548
	skb_queue_head_init(&port->fcoe_pending_queue);
	port->fcoe_pending_queue_active = 0;
	setup_timer(&port->timer, fcoe_queue_timer, (unsigned long)lp);
549

550 551 552 553 554 555 556
	if (!lp->vport) {
		wwnn = fcoe_wwn_from_mac(netdev->dev_addr, 1, 0);
		fc_set_wwnn(lp, wwnn);
		/* XXX - 3rd arg needs to be vlan id */
		wwpn = fcoe_wwn_from_mac(netdev->dev_addr, 2, 0);
		fc_set_wwpn(lp, wwpn);
	}
557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579

	return 0;
}

/**
 * fcoe_shost_config() - Sets up fc_lport->host
 * @lp : ptr to the fc_lport
 * @shost : ptr to the associated scsi host
 * @dev : device associated to scsi host
 *
 * Must be called after fcoe_lport_config() and fcoe_netdev_config()
 *
 * Returns : 0 for success
 */
static int fcoe_shost_config(struct fc_lport *lp, struct Scsi_Host *shost,
				struct device *dev)
{
	int rc = 0;

	/* lport scsi host config */
	lp->host->max_lun = FCOE_MAX_LUN;
	lp->host->max_id = FCOE_MAX_FCP_TARGET;
	lp->host->max_channel = 0;
580 581 582 583
	if (lp->vport)
		lp->host->transportt = fcoe_vport_transport_template;
	else
		lp->host->transportt = fcoe_transport_template;
584 585 586 587

	/* add the new host to the SCSI-ml */
	rc = scsi_add_host(lp->host, dev);
	if (rc) {
588 589
		FCOE_NETDEV_DBG(fcoe_netdev(lp), "fcoe_shost_config: "
				"error on scsi_add_host\n");
590 591
		return rc;
	}
592 593 594 595

	if (!lp->vport)
		fc_host_max_npiv_vports(lp->host) = USHORT_MAX;

596 597 598
	snprintf(fc_host_symbolic_name(lp->host), FC_SYMBOLIC_NAME_SIZE,
		 "%s v%s over %s", FCOE_NAME, FCOE_VERSION,
		 fcoe_netdev(lp)->name);
599 600 601 602

	return 0;
}

603 604 605 606 607 608 609 610
/*
 * fcoe_oem_match() - match for read types IO
 * @fp: the fc_frame for new IO.
 *
 * Returns : true for read types IO, otherwise returns false.
 */
bool fcoe_oem_match(struct fc_frame *fp)
{
611 612
	return fc_fcp_is_read(fr_fsp(fp)) &&
		(fr_fsp(fp)->data_len > fcoe_ddp_min);
613 614
}

615 616
/**
 * fcoe_em_config() - allocates em for this lport
617
 * @lp: the fcoe that em is to allocated for
618 619 620 621 622
 *
 * Returns : 0 on success
 */
static inline int fcoe_em_config(struct fc_lport *lp)
{
623
	struct fcoe_port *port = lport_priv(lp);
624 625
	struct fcoe_interface *fcoe = port->fcoe;
	struct fcoe_interface *oldfcoe = NULL;
626
	struct net_device *old_real_dev, *cur_real_dev;
627 628 629 630 631 632 633 634 635 636 637 638 639 640
	u16 min_xid = FCOE_MIN_XID;
	u16 max_xid = FCOE_MAX_XID;

	/*
	 * Check if need to allocate an em instance for
	 * offload exchange ids to be shared across all VN_PORTs/lport.
	 */
	if (!lp->lro_enabled || !lp->lro_xid || (lp->lro_xid >= max_xid)) {
		lp->lro_xid = 0;
		goto skip_oem;
	}

	/*
	 * Reuse existing offload em instance in case
641
	 * it is already allocated on real eth device
642
	 */
643 644
	if (fcoe->netdev->priv_flags & IFF_802_1Q_VLAN)
		cur_real_dev = vlan_dev_real_dev(fcoe->netdev);
645
	else
646
		cur_real_dev = fcoe->netdev;
647

648 649 650
	list_for_each_entry(oldfcoe, &fcoe_hostlist, list) {
		if (oldfcoe->netdev->priv_flags & IFF_802_1Q_VLAN)
			old_real_dev = vlan_dev_real_dev(oldfcoe->netdev);
651
		else
652
			old_real_dev = oldfcoe->netdev;
653 654

		if (cur_real_dev == old_real_dev) {
655
			fcoe->oem = oldfcoe->oem;
656 657 658 659
			break;
		}
	}

660 661
	if (fcoe->oem) {
		if (!fc_exch_mgr_add(lp, fcoe->oem, fcoe_oem_match)) {
662 663
			printk(KERN_ERR "fcoe_em_config: failed to add "
			       "offload em:%p on interface:%s\n",
664
			       fcoe->oem, fcoe->netdev->name);
665 666 667
			return -ENOMEM;
		}
	} else {
668
		fcoe->oem = fc_exch_mgr_alloc(lp, FC_CLASS_3,
669 670
					    FCOE_MIN_XID, lp->lro_xid,
					    fcoe_oem_match);
671
		if (!fcoe->oem) {
672 673
			printk(KERN_ERR "fcoe_em_config: failed to allocate "
			       "em for offload exches on interface:%s\n",
674
			       fcoe->netdev->name);
675 676 677 678 679 680 681 682 683 684 685 686
			return -ENOMEM;
		}
	}

	/*
	 * Exclude offload EM xid range from next EM xid range.
	 */
	min_xid += lp->lro_xid + 1;

skip_oem:
	if (!fc_exch_mgr_alloc(lp, FC_CLASS_3, min_xid, max_xid, NULL)) {
		printk(KERN_ERR "fcoe_em_config: failed to "
687
		       "allocate em on interface %s\n", fcoe->netdev->name);
688
		return -ENOMEM;
689
	}
690 691 692 693 694 695

	return 0;
}

/**
 * fcoe_if_destroy() - FCoE software HBA tear-down function
696
 * @lport: fc_lport to destroy
697
 */
698
static void fcoe_if_destroy(struct fc_lport *lport)
699
{
700 701
	struct fcoe_port *port = lport_priv(lport);
	struct fcoe_interface *fcoe = port->fcoe;
702
	struct net_device *netdev = fcoe->netdev;
703

704
	FCOE_NETDEV_DBG(netdev, "Destroying interface\n");
705 706

	/* Logout of the fabric */
707
	fc_fabric_logoff(lport);
708

709 710 711 712 713 714
	/* Cleanup the fc_lport */
	fc_lport_destroy(lport);
	fc_fcp_destroy(lport);

	/* Stop the transmit retry timer */
	del_timer_sync(&port->timer);
715

716 717 718
	/* Free existing transmit skbs */
	fcoe_clean_pending_queue(lport);

719 720 721 722 723
	rtnl_lock();
	if (!is_zero_ether_addr(port->data_src_addr))
		dev_unicast_delete(netdev, port->data_src_addr);
	rtnl_unlock();

724 725
	/* receives may not be stopped until after this */
	fcoe_interface_put(fcoe);
726

727
	/* Free queued packets for the per-CPU receive threads */
728
	fcoe_percpu_clean(lport);
729

730
	/* Detach from the scsi-ml */
731 732
	fc_remove_host(lport->host);
	scsi_remove_host(lport->host);
733 734

	/* There are no more rports or I/O, free the EM */
735
	fc_exch_mgr_free(lport);
736 737

	/* Free memory used by statistical counters */
738
	fc_lport_free_stats(lport);
739

740
	/* Release the Scsi_Host */
741
	scsi_host_put(lport->host);
742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757
}

/*
 * fcoe_ddp_setup - calls LLD's ddp_setup through net_device
 * @lp:	the corresponding fc_lport
 * @xid: the exchange id for this ddp transfer
 * @sgl: the scatterlist describing this transfer
 * @sgc: number of sg items
 *
 * Returns : 0 no ddp
 */
static int fcoe_ddp_setup(struct fc_lport *lp, u16 xid,
			     struct scatterlist *sgl, unsigned int sgc)
{
	struct net_device *n = fcoe_netdev(lp);

758
	if (n->netdev_ops->ndo_fcoe_ddp_setup)
759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774
		return n->netdev_ops->ndo_fcoe_ddp_setup(n, xid, sgl, sgc);

	return 0;
}

/*
 * fcoe_ddp_done - calls LLD's ddp_done through net_device
 * @lp:	the corresponding fc_lport
 * @xid: the exchange id for this ddp transfer
 *
 * Returns : the length of data that have been completed by ddp
 */
static int fcoe_ddp_done(struct fc_lport *lp, u16 xid)
{
	struct net_device *n = fcoe_netdev(lp);

775
	if (n->netdev_ops->ndo_fcoe_ddp_done)
776 777 778 779
		return n->netdev_ops->ndo_fcoe_ddp_done(n, xid);
	return 0;
}

780 781 782 783 784
static struct fc_seq *fcoe_elsct_send(struct fc_lport *lport,
		u32 did, struct fc_frame *fp, unsigned int op,
		void (*resp)(struct fc_seq *, struct fc_frame *, void *),
		void *arg, u32 timeout);

785 786 787 788
static struct libfc_function_template fcoe_libfc_fcn_templ = {
	.frame_send = fcoe_xmit,
	.ddp_setup = fcoe_ddp_setup,
	.ddp_done = fcoe_ddp_done,
789
	.elsct_send = fcoe_elsct_send,
790 791 792
};

/**
793 794
 * fcoe_if_create() - this function creates the fcoe port
 * @fcoe: fcoe_interface structure to create an fc_lport instance on
795
 * @parent: device pointer to be the parent in sysfs for the SCSI host
796
 * @npiv: is this a vport?
797
 *
798
 * Creates fc_lport struct and scsi_host for lport, configures lport.
799
 *
800
 * Returns : The allocated fc_lport or an error pointer
801
 */
802
static struct fc_lport *fcoe_if_create(struct fcoe_interface *fcoe,
803
				       struct device *parent, int npiv)
804 805
{
	int rc;
806
	struct fc_lport *lport = NULL;
807
	struct fcoe_port *port;
808
	struct Scsi_Host *shost;
809
	struct net_device *netdev = fcoe->netdev;
810 811 812 813 814
	/*
	 * parent is only a vport if npiv is 1,
	 * but we'll only use vport in that case so go ahead and set it
	 */
	struct fc_vport *vport = dev_to_vport(parent);
815

816
	FCOE_NETDEV_DBG(netdev, "Create Interface\n");
817

818 819 820 821 822 823 824
	if (!npiv) {
		lport = libfc_host_alloc(&fcoe_shost_template,
					 sizeof(struct fcoe_port));
	} else	{
		lport = libfc_vport_create(vport,
					   sizeof(struct fcoe_port));
	}
825
	if (!lport) {
826
		FCOE_NETDEV_DBG(netdev, "Could not allocate host structure\n");
827
		rc = -ENOMEM;
828
		goto out;
829
	}
830
	shost = lport->host;
831
	port = lport_priv(lport);
832
	port->lport = lport;
833
	port->fcoe = fcoe;
834
	INIT_WORK(&port->destroy_work, fcoe_destroy_work);
835 836

	/* configure fc_lport, e.g., em */
837
	rc = fcoe_lport_config(lport);
838
	if (rc) {
839 840
		FCOE_NETDEV_DBG(netdev, "Could not configure lport for the "
				"interface\n");
841 842 843
		goto out_host_put;
	}

844 845 846 847 848 849 850
	if (npiv) {
		FCOE_NETDEV_DBG(netdev, "Setting vport names, 0x%llX 0x%llX\n",
			vport->node_name, vport->port_name);
		fc_set_wwnn(lport, vport->node_name);
		fc_set_wwpn(lport, vport->port_name);
	}

851
	/* configure lport network properties */
852
	rc = fcoe_netdev_config(lport, netdev);
853
	if (rc) {
854 855
		FCOE_NETDEV_DBG(netdev, "Could not configure netdev for the "
				"interface\n");
856
		goto out_lp_destroy;
857
	}
858

859
	/* configure lport scsi host properties */
860
	rc = fcoe_shost_config(lport, shost, parent);
861
	if (rc) {
862 863
		FCOE_NETDEV_DBG(netdev, "Could not configure shost for the "
				"interface\n");
864
		goto out_lp_destroy;
865 866
	}

867
	/* Initialize the library */
868
	rc = fcoe_libfc_config(lport, &fcoe_libfc_fcn_templ);
869
	if (rc) {
870
		FCOE_NETDEV_DBG(netdev, "Could not configure libfc for the "
871
				"interface\n");
872
		goto out_lp_destroy;
873 874
	}

875 876 877 878 879 880 881 882 883 884
	if (!npiv) {
		/*
		 * fcoe_em_alloc() and fcoe_hostlist_add() both
		 * need to be atomic with respect to other changes to the
		 * hostlist since fcoe_em_alloc() looks for an existing EM
		 * instance on host list updated by fcoe_hostlist_add().
		 *
		 * This is currently handled through the fcoe_config_mutex
		 * begin held.
		 */
885

886 887 888 889 890 891 892
		/* lport exch manager allocation */
		rc = fcoe_em_config(lport);
		if (rc) {
			FCOE_NETDEV_DBG(netdev, "Could not configure the EM "
						"for the interface\n");
			goto out_lp_destroy;
		}
893 894
	}

895
	fcoe_interface_get(fcoe);
896
	return lport;
897 898

out_lp_destroy:
899
	fc_exch_mgr_free(lport);
900
out_host_put:
901 902 903
	scsi_host_put(lport->host);
out:
	return ERR_PTR(rc);
904 905 906 907 908 909 910 911 912 913
}

/**
 * fcoe_if_init() - attach to scsi transport
 *
 * Returns : 0 on success
 */
static int __init fcoe_if_init(void)
{
	/* attach to scsi transport */
914 915 916
	fcoe_transport_template = fc_attach_transport(&fcoe_transport_function);
	fcoe_vport_transport_template =
		fc_attach_transport(&fcoe_vport_transport_function);
917

918
	if (!fcoe_transport_template) {
919
		printk(KERN_ERR "fcoe: Failed to attach to the FC transport\n");
920 921 922 923 924 925 926 927 928 929 930 931 932
		return -ENODEV;
	}

	return 0;
}

/**
 * fcoe_if_exit() - detach from scsi transport
 *
 * Returns : 0 on success
 */
int __exit fcoe_if_exit(void)
{
933 934 935 936
	fc_release_transport(fcoe_transport_template);
	fc_release_transport(fcoe_vport_transport_template);
	fcoe_transport_template = NULL;
	fcoe_vport_transport_template = NULL;
937 938 939
	return 0;
}

940 941 942 943 944 945 946 947 948 949 950 951 952 953
/**
 * fcoe_percpu_thread_create() - Create a receive thread for an online cpu
 * @cpu: cpu index for the online cpu
 */
static void fcoe_percpu_thread_create(unsigned int cpu)
{
	struct fcoe_percpu_s *p;
	struct task_struct *thread;

	p = &per_cpu(fcoe_percpu, cpu);

	thread = kthread_create(fcoe_percpu_receive_thread,
				(void *)p, "fcoethread/%d", cpu);

954
	if (likely(!IS_ERR(thread))) {
955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982
		kthread_bind(thread, cpu);
		wake_up_process(thread);

		spin_lock_bh(&p->fcoe_rx_list.lock);
		p->thread = thread;
		spin_unlock_bh(&p->fcoe_rx_list.lock);
	}
}

/**
 * fcoe_percpu_thread_destroy() - removes the rx thread for the given cpu
 * @cpu: cpu index the rx thread is to be removed
 *
 * Destroys a per-CPU Rx thread. Any pending skbs are moved to the
 * current CPU's Rx thread. If the thread being destroyed is bound to
 * the CPU processing this context the skbs will be freed.
 */
static void fcoe_percpu_thread_destroy(unsigned int cpu)
{
	struct fcoe_percpu_s *p;
	struct task_struct *thread;
	struct page *crc_eof;
	struct sk_buff *skb;
#ifdef CONFIG_SMP
	struct fcoe_percpu_s *p0;
	unsigned targ_cpu = smp_processor_id();
#endif /* CONFIG_SMP */

983
	FCOE_DBG("Destroying receive thread for CPU %d\n", cpu);
984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004

	/* Prevent any new skbs from being queued for this CPU. */
	p = &per_cpu(fcoe_percpu, cpu);
	spin_lock_bh(&p->fcoe_rx_list.lock);
	thread = p->thread;
	p->thread = NULL;
	crc_eof = p->crc_eof_page;
	p->crc_eof_page = NULL;
	p->crc_eof_offset = 0;
	spin_unlock_bh(&p->fcoe_rx_list.lock);

#ifdef CONFIG_SMP
	/*
	 * Don't bother moving the skb's if this context is running
	 * on the same CPU that is having its thread destroyed. This
	 * can easily happen when the module is removed.
	 */
	if (cpu != targ_cpu) {
		p0 = &per_cpu(fcoe_percpu, targ_cpu);
		spin_lock_bh(&p0->fcoe_rx_list.lock);
		if (p0->thread) {
1005 1006
			FCOE_DBG("Moving frames from CPU %d to CPU %d\n",
				 cpu, targ_cpu);
1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037

			while ((skb = __skb_dequeue(&p->fcoe_rx_list)) != NULL)
				__skb_queue_tail(&p0->fcoe_rx_list, skb);
			spin_unlock_bh(&p0->fcoe_rx_list.lock);
		} else {
			/*
			 * The targeted CPU is not initialized and cannot accept
			 * new  skbs. Unlock the targeted CPU and drop the skbs
			 * on the CPU that is going offline.
			 */
			while ((skb = __skb_dequeue(&p->fcoe_rx_list)) != NULL)
				kfree_skb(skb);
			spin_unlock_bh(&p0->fcoe_rx_list.lock);
		}
	} else {
		/*
		 * This scenario occurs when the module is being removed
		 * and all threads are being destroyed. skbs will continue
		 * to be shifted from the CPU thread that is being removed
		 * to the CPU thread associated with the CPU that is processing
		 * the module removal. Once there is only one CPU Rx thread it
		 * will reach this case and we will drop all skbs and later
		 * stop the thread.
		 */
		spin_lock_bh(&p->fcoe_rx_list.lock);
		while ((skb = __skb_dequeue(&p->fcoe_rx_list)) != NULL)
			kfree_skb(skb);
		spin_unlock_bh(&p->fcoe_rx_list.lock);
	}
#else
	/*
1038
	 * This a non-SMP scenario where the singular Rx thread is
1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071
	 * being removed. Free all skbs and stop the thread.
	 */
	spin_lock_bh(&p->fcoe_rx_list.lock);
	while ((skb = __skb_dequeue(&p->fcoe_rx_list)) != NULL)
		kfree_skb(skb);
	spin_unlock_bh(&p->fcoe_rx_list.lock);
#endif

	if (thread)
		kthread_stop(thread);

	if (crc_eof)
		put_page(crc_eof);
}

/**
 * fcoe_cpu_callback() - fcoe cpu hotplug event callback
 * @nfb: callback data block
 * @action: event triggering the callback
 * @hcpu: index for the cpu of this event
 *
 * This creates or destroys per cpu data for fcoe
 *
 * Returns NOTIFY_OK always.
 */
static int fcoe_cpu_callback(struct notifier_block *nfb,
			     unsigned long action, void *hcpu)
{
	unsigned cpu = (unsigned long)hcpu;

	switch (action) {
	case CPU_ONLINE:
	case CPU_ONLINE_FROZEN:
1072
		FCOE_DBG("CPU %x online: Create Rx thread\n", cpu);
1073 1074 1075 1076
		fcoe_percpu_thread_create(cpu);
		break;
	case CPU_DEAD:
	case CPU_DEAD_FROZEN:
1077
		FCOE_DBG("CPU %x offline: Remove Rx thread\n", cpu);
1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089
		fcoe_percpu_thread_destroy(cpu);
		break;
	default:
		break;
	}
	return NOTIFY_OK;
}

static struct notifier_block fcoe_cpu_notifier = {
	.notifier_call = fcoe_cpu_callback,
};

1090
/**
1091
 * fcoe_rcv() - this is the fcoe receive function called by NET_RX_SOFTIRQ
1092 1093 1094
 * @skb: the receive skb
 * @dev: associated net device
 * @ptype: context
1095
 * @olddev: last device
1096 1097 1098 1099
 *
 * this function will receive the packet and build fc frame and pass it up
 *
 * Returns: 0 for success
1100
 */
1101 1102 1103 1104 1105
int fcoe_rcv(struct sk_buff *skb, struct net_device *dev,
	     struct packet_type *ptype, struct net_device *olddev)
{
	struct fc_lport *lp;
	struct fcoe_rcv_info *fr;
1106
	struct fcoe_interface *fcoe;
1107 1108
	struct fc_frame_header *fh;
	struct fcoe_percpu_s *fps;
1109
	unsigned int cpu;
1110

1111
	fcoe = container_of(ptype, struct fcoe_interface, fcoe_packet_type);
1112
	lp = fcoe->ctlr.lp;
1113
	if (unlikely(lp == NULL)) {
1114
		FCOE_NETDEV_DBG(dev, "Cannot find hba structure");
1115 1116
		goto err2;
	}
1117 1118
	if (!lp->link_up)
		goto err2;
1119

1120 1121 1122 1123 1124
	FCOE_NETDEV_DBG(dev, "skb_info: len:%d data_len:%d head:%p "
			"data:%p tail:%p end:%p sum:%d dev:%s",
			skb->len, skb->data_len, skb->head, skb->data,
			skb_tail_pointer(skb), skb_end_pointer(skb),
			skb->csum, skb->dev ? skb->dev->name : "<NULL>");
1125 1126 1127

	/* check for FCOE packet type */
	if (unlikely(eth_hdr(skb)->h_proto != htons(ETH_P_FCOE))) {
1128
		FCOE_NETDEV_DBG(dev, "Wrong FC type frame");
1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145
		goto err;
	}

	/*
	 * Check for minimum frame length, and make sure required FCoE
	 * and FC headers are pulled into the linear data area.
	 */
	if (unlikely((skb->len < FCOE_MIN_FRAME) ||
	    !pskb_may_pull(skb, FCOE_HEADER_LEN)))
		goto err;

	skb_set_transport_header(skb, sizeof(struct fcoe_hdr));
	fh = (struct fc_frame_header *) skb_transport_header(skb);

	fr = fcoe_dev_from_skb(skb);
	fr->fr_dev = lp;
	fr->ptype = ptype;
1146

1147
	/*
1148 1149 1150 1151
	 * In case the incoming frame's exchange is originated from
	 * the initiator, then received frame's exchange id is ANDed
	 * with fc_cpu_mask bits to get the same cpu on which exchange
	 * was originated, otherwise just use the current cpu.
1152
	 */
1153 1154 1155 1156
	if (ntoh24(fh->fh_f_ctl) & FC_FC_EX_CTX)
		cpu = ntohs(fh->fh_ox_id) & fc_cpu_mask;
	else
		cpu = smp_processor_id();
1157

1158
	fps = &per_cpu(fcoe_percpu, cpu);
1159
	spin_lock_bh(&fps->fcoe_rx_list.lock);
1160 1161 1162 1163 1164 1165
	if (unlikely(!fps->thread)) {
		/*
		 * The targeted CPU is not ready, let's target
		 * the first CPU now. For non-SMP systems this
		 * will check the same CPU twice.
		 */
1166 1167 1168
		FCOE_NETDEV_DBG(dev, "CPU is online, but no receive thread "
				"ready for incoming skb- using first online "
				"CPU.\n");
1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184

		spin_unlock_bh(&fps->fcoe_rx_list.lock);
		cpu = first_cpu(cpu_online_map);
		fps = &per_cpu(fcoe_percpu, cpu);
		spin_lock_bh(&fps->fcoe_rx_list.lock);
		if (!fps->thread) {
			spin_unlock_bh(&fps->fcoe_rx_list.lock);
			goto err;
		}
	}

	/*
	 * We now have a valid CPU that we're targeting for
	 * this skb. We also have this receive thread locked,
	 * so we're free to queue skbs into it's queue.
	 */
1185 1186 1187 1188 1189 1190 1191 1192
	__skb_queue_tail(&fps->fcoe_rx_list, skb);
	if (fps->fcoe_rx_list.qlen == 1)
		wake_up_process(fps->thread);

	spin_unlock_bh(&fps->fcoe_rx_list.lock);

	return 0;
err:
1193
	fc_lport_get_stats(lp)->ErrorFrames++;
1194 1195 1196 1197 1198 1199 1200

err2:
	kfree_skb(skb);
	return -1;
}

/**
1201
 * fcoe_start_io() - pass to netdev to start xmit for fcoe
1202 1203 1204
 * @skb: the skb to be xmitted
 *
 * Returns: 0 for success
1205
 */
1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218
static inline int fcoe_start_io(struct sk_buff *skb)
{
	int rc;

	skb_get(skb);
	rc = dev_queue_xmit(skb);
	if (rc != 0)
		return rc;
	kfree_skb(skb);
	return 0;
}

/**
1219
 * fcoe_get_paged_crc_eof() - in case we need to alloc a page for crc_eof
1220 1221 1222 1223
 * @skb: the skb to be xmitted
 * @tlen: total len
 *
 * Returns: 0 for success
1224
 */
1225 1226 1227 1228 1229
static int fcoe_get_paged_crc_eof(struct sk_buff *skb, int tlen)
{
	struct fcoe_percpu_s *fps;
	struct page *page;

1230
	fps = &get_cpu_var(fcoe_percpu);
1231 1232 1233 1234
	page = fps->crc_eof_page;
	if (!page) {
		page = alloc_page(GFP_ATOMIC);
		if (!page) {
1235
			put_cpu_var(fcoe_percpu);
1236 1237 1238
			return -ENOMEM;
		}
		fps->crc_eof_page = page;
1239
		fps->crc_eof_offset = 0;
1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254
	}

	get_page(page);
	skb_fill_page_desc(skb, skb_shinfo(skb)->nr_frags, page,
			   fps->crc_eof_offset, tlen);
	skb->len += tlen;
	skb->data_len += tlen;
	skb->truesize += tlen;
	fps->crc_eof_offset += sizeof(struct fcoe_crc_eof);

	if (fps->crc_eof_offset >= PAGE_SIZE) {
		fps->crc_eof_page = NULL;
		fps->crc_eof_offset = 0;
		put_page(page);
	}
1255
	put_cpu_var(fcoe_percpu);
1256 1257 1258 1259
	return 0;
}

/**
1260
 * fcoe_fc_crc() - calculates FC CRC in this fcoe skb
1261
 * @fp: the fc_frame containing data to be checksummed
1262
 *
1263
 * This uses crc32() to calculate the crc for port frame
1264
 * Return   : 32 bit crc
1265
 */
1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294
u32 fcoe_fc_crc(struct fc_frame *fp)
{
	struct sk_buff *skb = fp_skb(fp);
	struct skb_frag_struct *frag;
	unsigned char *data;
	unsigned long off, len, clen;
	u32 crc;
	unsigned i;

	crc = crc32(~0, skb->data, skb_headlen(skb));

	for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
		frag = &skb_shinfo(skb)->frags[i];
		off = frag->page_offset;
		len = frag->size;
		while (len > 0) {
			clen = min(len, PAGE_SIZE - (off & ~PAGE_MASK));
			data = kmap_atomic(frag->page + (off >> PAGE_SHIFT),
					   KM_SKB_DATA_SOFTIRQ);
			crc = crc32(crc, data + (off & ~PAGE_MASK), clen);
			kunmap_atomic(data, KM_SKB_DATA_SOFTIRQ);
			off += clen;
			len -= clen;
		}
	}
	return crc;
}

/**
1295
 * fcoe_xmit() - FCoE frame transmit function
1296
 * @lp:	the associated local fcoe
1297 1298 1299
 * @fp: the fc_frame to be transmitted
 *
 * Return   : 0 for success
1300
 */
1301 1302
int fcoe_xmit(struct fc_lport *lp, struct fc_frame *fp)
{
1303
	int wlen;
1304 1305 1306 1307 1308 1309 1310 1311 1312
	u32 crc;
	struct ethhdr *eh;
	struct fcoe_crc_eof *cp;
	struct sk_buff *skb;
	struct fcoe_dev_stats *stats;
	struct fc_frame_header *fh;
	unsigned int hlen;		/* header length implies the version */
	unsigned int tlen;		/* trailer length */
	unsigned int elen;		/* eth header, may include vlan */
1313 1314
	struct fcoe_port *port = lport_priv(lp);
	struct fcoe_interface *fcoe = port->fcoe;
1315 1316 1317 1318 1319 1320
	u8 sof, eof;
	struct fcoe_hdr *hp;

	WARN_ON((fr_len(fp) % sizeof(u32)) != 0);

	fh = fc_frame_header_get(fp);
1321 1322 1323 1324
	skb = fp_skb(fp);
	wlen = skb->len / FCOE_WORD_TO_BYTE;

	if (!lp->link_up) {
1325
		kfree_skb(skb);
1326
		return 0;
1327 1328
	}

1329
	if (unlikely(fh->fh_r_ctl == FC_RCTL_ELS_REQ) &&
1330
	    fcoe_ctlr_els_send(&fcoe->ctlr, lp, skb))
1331 1332
		return 0;

1333 1334 1335
	sof = fr_sof(fp);
	eof = fr_eof(fp);

1336
	elen = sizeof(struct ethhdr);
1337 1338 1339 1340 1341 1342
	hlen = sizeof(struct fcoe_hdr);
	tlen = sizeof(struct fcoe_crc_eof);
	wlen = (skb->len - tlen + sizeof(crc)) / FCOE_WORD_TO_BYTE;

	/* crc offload */
	if (likely(lp->crc_offload)) {
1343
		skb->ip_summed = CHECKSUM_PARTIAL;
1344 1345 1346 1347 1348 1349 1350 1351
		skb->csum_start = skb_headroom(skb);
		skb->csum_offset = skb->len;
		crc = 0;
	} else {
		skb->ip_summed = CHECKSUM_NONE;
		crc = fcoe_fc_crc(fp);
	}

1352
	/* copy port crc and eof to the skb buff */
1353 1354 1355
	if (skb_is_nonlinear(skb)) {
		skb_frag_t *frag;
		if (fcoe_get_paged_crc_eof(skb, tlen)) {
R
Roel Kluin 已提交
1356
			kfree_skb(skb);
1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374
			return -ENOMEM;
		}
		frag = &skb_shinfo(skb)->frags[skb_shinfo(skb)->nr_frags - 1];
		cp = kmap_atomic(frag->page, KM_SKB_DATA_SOFTIRQ)
			+ frag->page_offset;
	} else {
		cp = (struct fcoe_crc_eof *)skb_put(skb, tlen);
	}

	memset(cp, 0, sizeof(*cp));
	cp->fcoe_eof = eof;
	cp->fcoe_crc32 = cpu_to_le32(~crc);

	if (skb_is_nonlinear(skb)) {
		kunmap_atomic(cp, KM_SKB_DATA_SOFTIRQ);
		cp = NULL;
	}

1375
	/* adjust skb network/transport offsets to match mac/fcoe/port */
1376 1377 1378 1379
	skb_push(skb, elen + hlen);
	skb_reset_mac_header(skb);
	skb_reset_network_header(skb);
	skb->mac_len = elen;
1380
	skb->protocol = htons(ETH_P_FCOE);
1381
	skb->dev = fcoe->netdev;
1382 1383 1384 1385

	/* fill up mac and fcoe headers */
	eh = eth_hdr(skb);
	eh->h_proto = htons(ETH_P_FCOE);
1386
	if (fcoe->ctlr.map_dest)
1387 1388 1389
		fc_fcoe_set_mac(eh->h_dest, fh->fh_d_id);
	else
		/* insert GW address */
1390
		memcpy(eh->h_dest, fcoe->ctlr.dest_addr, ETH_ALEN);
1391

1392 1393
	if (unlikely(fcoe->ctlr.flogi_oxid != FC_XID_UNKNOWN))
		memcpy(eh->h_source, fcoe->ctlr.ctl_src_addr, ETH_ALEN);
1394
	else
1395
		memcpy(eh->h_source, port->data_src_addr, ETH_ALEN);
1396 1397 1398 1399 1400 1401 1402

	hp = (struct fcoe_hdr *)(eh + 1);
	memset(hp, 0, sizeof(*hp));
	if (FC_FCOE_VER)
		FC_FCOE_ENCAPS_VER(hp, FC_FCOE_VER);
	hp->fcoe_sof = sof;

1403 1404 1405 1406 1407 1408 1409 1410
	/* fcoe lso, mss is in max_payload which is non-zero for FCP data */
	if (lp->seq_offload && fr_max_payload(fp)) {
		skb_shinfo(skb)->gso_type = SKB_GSO_FCOE;
		skb_shinfo(skb)->gso_size = fr_max_payload(fp);
	} else {
		skb_shinfo(skb)->gso_type = 0;
		skb_shinfo(skb)->gso_size = 0;
	}
1411
	/* update tx stats: regardless if LLD fails */
1412 1413 1414
	stats = fc_lport_get_stats(lp);
	stats->TxFrames++;
	stats->TxWords += wlen;
1415 1416 1417

	/* send down to lld */
	fr_dev(fp) = lp;
1418
	if (port->fcoe_pending_queue.qlen)
1419 1420 1421
		fcoe_check_wait_queue(lp, skb);
	else if (fcoe_start_io(skb))
		fcoe_check_wait_queue(lp, skb);
1422 1423 1424 1425

	return 0;
}

1426 1427 1428 1429 1430 1431 1432 1433 1434
/**
 * fcoe_percpu_flush_done() - Indicate percpu queue flush completion.
 * @skb: the skb being completed.
 */
static void fcoe_percpu_flush_done(struct sk_buff *skb)
{
	complete(&fcoe_flush_completion);
}

1435 1436
/**
 * fcoe_percpu_receive_thread() - recv thread per cpu
1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452
 * @arg: ptr to the fcoe per cpu struct
 *
 * Return: 0 for success
 */
int fcoe_percpu_receive_thread(void *arg)
{
	struct fcoe_percpu_s *p = arg;
	u32 fr_len;
	struct fc_lport *lp;
	struct fcoe_rcv_info *fr;
	struct fcoe_dev_stats *stats;
	struct fc_frame_header *fh;
	struct sk_buff *skb;
	struct fcoe_crc_eof crc_eof;
	struct fc_frame *fp;
	u8 *mac = NULL;
1453
	struct fcoe_port *port;
1454 1455
	struct fcoe_hdr *hp;

1456
	set_user_nice(current, -20);
1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473

	while (!kthread_should_stop()) {

		spin_lock_bh(&p->fcoe_rx_list.lock);
		while ((skb = __skb_dequeue(&p->fcoe_rx_list)) == NULL) {
			set_current_state(TASK_INTERRUPTIBLE);
			spin_unlock_bh(&p->fcoe_rx_list.lock);
			schedule();
			set_current_state(TASK_RUNNING);
			if (kthread_should_stop())
				return 0;
			spin_lock_bh(&p->fcoe_rx_list.lock);
		}
		spin_unlock_bh(&p->fcoe_rx_list.lock);
		fr = fcoe_dev_from_skb(skb);
		lp = fr->fr_dev;
		if (unlikely(lp == NULL)) {
1474 1475
			if (skb->destructor != fcoe_percpu_flush_done)
				FCOE_NETDEV_DBG(skb->dev, "NULL lport in skb");
1476 1477 1478 1479
			kfree_skb(skb);
			continue;
		}

1480 1481 1482 1483 1484 1485
		FCOE_NETDEV_DBG(skb->dev, "skb_info: len:%d data_len:%d "
				"head:%p data:%p tail:%p end:%p sum:%d dev:%s",
				skb->len, skb->data_len,
				skb->head, skb->data, skb_tail_pointer(skb),
				skb_end_pointer(skb), skb->csum,
				skb->dev ? skb->dev->name : "<NULL>");
1486 1487 1488 1489

		/*
		 * Save source MAC address before discarding header.
		 */
1490
		port = lport_priv(lp);
1491 1492
		if (skb_is_nonlinear(skb))
			skb_linearize(skb);	/* not ideal */
1493
		mac = eth_hdr(skb)->h_source;
1494 1495 1496 1497 1498 1499 1500 1501

		/*
		 * Frame length checks and setting up the header pointers
		 * was done in fcoe_rcv already.
		 */
		hp = (struct fcoe_hdr *) skb_network_header(skb);
		fh = (struct fc_frame_header *) skb_transport_header(skb);

1502
		stats = fc_lport_get_stats(lp);
1503
		if (unlikely(FC_FCOE_DECAPS_VER(hp) != FC_FCOE_VER)) {
1504
			if (stats->ErrorFrames < 5)
1505
				printk(KERN_WARNING "fcoe: FCoE version "
1506 1507 1508 1509 1510 1511
				       "mismatch: The frame has "
				       "version %x, but the "
				       "initiator supports version "
				       "%x\n", FC_FCOE_DECAPS_VER(hp),
				       FC_FCOE_VER);
			stats->ErrorFrames++;
1512 1513 1514 1515 1516 1517 1518
			kfree_skb(skb);
			continue;
		}

		skb_pull(skb, sizeof(struct fcoe_hdr));
		fr_len = skb->len - sizeof(struct fcoe_crc_eof);

1519 1520
		stats->RxFrames++;
		stats->RxWords += fr_len / FCOE_WORD_TO_BYTE;
1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543

		fp = (struct fc_frame *)skb;
		fc_frame_init(fp);
		fr_dev(fp) = lp;
		fr_sof(fp) = hp->fcoe_sof;

		/* Copy out the CRC and EOF trailer for access */
		if (skb_copy_bits(skb, fr_len, &crc_eof, sizeof(crc_eof))) {
			kfree_skb(skb);
			continue;
		}
		fr_eof(fp) = crc_eof.fcoe_eof;
		fr_crc(fp) = crc_eof.fcoe_crc32;
		if (pskb_trim(skb, fr_len)) {
			kfree_skb(skb);
			continue;
		}

		/*
		 * We only check CRC if no offload is available and if it is
		 * it's solicited data, in which case, the FCP layer would
		 * check it during the copy.
		 */
1544
		if (lp->crc_offload && skb->ip_summed == CHECKSUM_UNNECESSARY)
1545 1546 1547 1548 1549 1550 1551
			fr_flags(fp) &= ~FCPHF_CRC_UNCHECKED;
		else
			fr_flags(fp) |= FCPHF_CRC_UNCHECKED;

		fh = fc_frame_header_get(fp);
		if (fh->fh_r_ctl == FC_RCTL_DD_SOL_DATA &&
		    fh->fh_type == FC_TYPE_FCP) {
1552
			fc_exch_recv(lp, fp);
1553 1554 1555 1556 1557
			continue;
		}
		if (fr_flags(fp) & FCPHF_CRC_UNCHECKED) {
			if (le32_to_cpu(fr_crc(fp)) !=
			    ~crc32(~0, skb->data, fr_len)) {
1558
				if (stats->InvalidCRCCount < 5)
1559 1560 1561 1562 1563 1564 1565 1566 1567
					printk(KERN_WARNING "fcoe: dropping "
					       "frame with CRC error\n");
				stats->InvalidCRCCount++;
				stats->ErrorFrames++;
				fc_frame_free(fp);
				continue;
			}
			fr_flags(fp) &= ~FCPHF_CRC_UNCHECKED;
		}
1568
		fc_exch_recv(lp, fp);
1569 1570 1571 1572 1573
	}
	return 0;
}

/**
1574 1575
 * fcoe_check_wait_queue() - attempt to clear the transmit backlog
 * @lp: the fc_lport
1576 1577 1578
 *
 * This empties the wait_queue, dequeue the head of the wait_queue queue
 * and calls fcoe_start_io() for each packet, if all skb have been
1579
 * transmitted, return qlen or -1 if a error occurs, then restore
1580
 * wait_queue and try again later.
1581 1582
 *
 * The wait_queue is used when the skb transmit fails. skb will go
1583
 * in the wait_queue which will be emptied by the timer function or
1584
 * by the next skb transmit.
1585
 */
1586
static void fcoe_check_wait_queue(struct fc_lport *lp, struct sk_buff *skb)
1587
{
1588
	struct fcoe_port *port = lport_priv(lp);
1589
	int rc;
1590

1591
	spin_lock_bh(&port->fcoe_pending_queue.lock);
1592 1593

	if (skb)
1594
		__skb_queue_tail(&port->fcoe_pending_queue, skb);
1595

1596
	if (port->fcoe_pending_queue_active)
1597
		goto out;
1598
	port->fcoe_pending_queue_active = 1;
1599

1600
	while (port->fcoe_pending_queue.qlen) {
1601
		/* keep qlen > 0 until fcoe_start_io succeeds */
1602 1603
		port->fcoe_pending_queue.qlen++;
		skb = __skb_dequeue(&port->fcoe_pending_queue);
1604

1605
		spin_unlock_bh(&port->fcoe_pending_queue.lock);
1606
		rc = fcoe_start_io(skb);
1607
		spin_lock_bh(&port->fcoe_pending_queue.lock);
1608 1609

		if (rc) {
1610
			__skb_queue_head(&port->fcoe_pending_queue, skb);
1611
			/* undo temporary increment above */
1612
			port->fcoe_pending_queue.qlen--;
1613
			break;
1614
		}
1615
		/* undo temporary increment above */
1616
		port->fcoe_pending_queue.qlen--;
1617
	}
1618

1619
	if (port->fcoe_pending_queue.qlen < FCOE_LOW_QUEUE_DEPTH)
1620
		lp->qfull = 0;
1621 1622 1623
	if (port->fcoe_pending_queue.qlen && !timer_pending(&port->timer))
		mod_timer(&port->timer, jiffies + 2);
	port->fcoe_pending_queue_active = 0;
1624
out:
1625
	if (port->fcoe_pending_queue.qlen > FCOE_MAX_QUEUE_DEPTH)
1626
		lp->qfull = 1;
1627
	spin_unlock_bh(&port->fcoe_pending_queue.lock);
1628
	return;
1629 1630 1631
}

/**
1632 1633
 * fcoe_dev_setup() - setup link change notification interface
 */
1634
static void fcoe_dev_setup(void)
1635 1636 1637 1638 1639
{
	register_netdevice_notifier(&fcoe_notifier);
}

/**
1640
 * fcoe_dev_cleanup() - cleanup link change notification interface
1641
 */
1642 1643 1644 1645 1646 1647
static void fcoe_dev_cleanup(void)
{
	unregister_netdevice_notifier(&fcoe_notifier);
}

/**
1648
 * fcoe_device_notification() - netdev event notification callback
1649 1650 1651 1652 1653 1654 1655
 * @notifier: context of the notification
 * @event: type of event
 * @ptr: fixed array for output parsed ifname
 *
 * This function is called by the ethernet driver in case of link change event
 *
 * Returns: 0 for success
1656
 */
1657 1658 1659 1660
static int fcoe_device_notification(struct notifier_block *notifier,
				    ulong event, void *ptr)
{
	struct fc_lport *lp = NULL;
1661
	struct net_device *netdev = ptr;
1662
	struct fcoe_interface *fcoe;
1663
	struct fcoe_port *port;
1664
	struct fcoe_dev_stats *stats;
1665
	u32 link_possible = 1;
1666 1667 1668
	u32 mfs;
	int rc = NOTIFY_OK;

1669
	list_for_each_entry(fcoe, &fcoe_hostlist, list) {
1670
		if (fcoe->netdev == netdev) {
1671
			lp = fcoe->ctlr.lp;
1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682
			break;
		}
	}
	if (lp == NULL) {
		rc = NOTIFY_DONE;
		goto out;
	}

	switch (event) {
	case NETDEV_DOWN:
	case NETDEV_GOING_DOWN:
1683
		link_possible = 0;
1684 1685 1686 1687 1688
		break;
	case NETDEV_UP:
	case NETDEV_CHANGE:
		break;
	case NETDEV_CHANGEMTU:
1689 1690
		if (netdev->features & NETIF_F_FCOE_MTU)
			break;
1691 1692
		mfs = netdev->mtu - (sizeof(struct fcoe_hdr) +
				     sizeof(struct fcoe_crc_eof));
1693 1694 1695 1696 1697
		if (mfs >= FC_MIN_MAX_FRAME)
			fc_set_mfs(lp, mfs);
		break;
	case NETDEV_REGISTER:
		break;
1698 1699 1700 1701 1702 1703 1704
	case NETDEV_UNREGISTER:
		list_del(&fcoe->list);
		port = lport_priv(fcoe->ctlr.lp);
		fcoe_interface_cleanup(fcoe);
		schedule_work(&port->destroy_work);
		goto out;
		break;
1705
	default:
1706
		FCOE_NETDEV_DBG(netdev, "Unknown event %ld "
1707
				"from netdev netlink\n", event);
1708
	}
1709
	if (link_possible && !fcoe_link_ok(lp))
1710 1711
		fcoe_ctlr_link_up(&fcoe->ctlr);
	else if (fcoe_ctlr_link_down(&fcoe->ctlr)) {
1712 1713 1714
		stats = fc_lport_get_stats(lp);
		stats->LinkFailureCount++;
		fcoe_clean_pending_queue(lp);
1715 1716 1717 1718 1719 1720
	}
out:
	return rc;
}

/**
1721
 * fcoe_if_to_netdev() - parse a name buffer to get netdev
1722 1723
 * @buffer: incoming buffer to be copied
 *
1724
 * Returns: NULL or ptr to net_device
1725
 */
1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741
static struct net_device *fcoe_if_to_netdev(const char *buffer)
{
	char *cp;
	char ifname[IFNAMSIZ + 2];

	if (buffer) {
		strlcpy(ifname, buffer, IFNAMSIZ);
		cp = ifname + strlen(ifname);
		while (--cp >= ifname && *cp == '\n')
			*cp = '\0';
		return dev_get_by_name(&init_net, ifname);
	}
	return NULL;
}

/**
1742
 * fcoe_destroy() - handles the destroy from sysfs
1743
 * @buffer: expected to be an eth if name
1744 1745 1746
 * @kp: associated kernel param
 *
 * Returns: 0 for success
1747
 */
1748 1749
static int fcoe_destroy(const char *buffer, struct kernel_param *kp)
{
1750
	struct fcoe_interface *fcoe;
1751
	struct net_device *netdev;
1752
	int rc = 0;
1753

1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766
	mutex_lock(&fcoe_config_mutex);
#ifdef CONFIG_FCOE_MODULE
	/*
	 * Make sure the module has been initialized, and is not about to be
	 * removed.  Module paramter sysfs files are writable before the
	 * module_init function is called and after module_exit.
	 */
	if (THIS_MODULE->state != MODULE_STATE_LIVE) {
		rc = -ENODEV;
		goto out_nodev;
	}
#endif

1767 1768 1769 1770 1771
	netdev = fcoe_if_to_netdev(buffer);
	if (!netdev) {
		rc = -ENODEV;
		goto out_nodev;
	}
1772

1773
	rtnl_lock();
1774 1775
	fcoe = fcoe_hostlist_lookup_port(netdev);
	if (!fcoe) {
1776
		rtnl_unlock();
1777 1778 1779
		rc = -ENODEV;
		goto out_putdev;
	}
1780 1781 1782 1783
	list_del(&fcoe->list);
	fcoe_interface_cleanup(fcoe);
	rtnl_unlock();
	fcoe_if_destroy(fcoe->ctlr.lp);
1784 1785 1786
out_putdev:
	dev_put(netdev);
out_nodev:
1787
	mutex_unlock(&fcoe_config_mutex);
1788 1789 1790
	return rc;
}

1791 1792 1793 1794 1795 1796 1797 1798 1799 1800
static void fcoe_destroy_work(struct work_struct *work)
{
	struct fcoe_port *port;

	port = container_of(work, struct fcoe_port, destroy_work);
	mutex_lock(&fcoe_config_mutex);
	fcoe_if_destroy(port->lport);
	mutex_unlock(&fcoe_config_mutex);
}

1801
/**
1802
 * fcoe_create() - Handles the create call from sysfs
1803
 * @buffer: expected to be an eth if name
1804 1805 1806
 * @kp: associated kernel param
 *
 * Returns: 0 for success
1807
 */
1808 1809 1810
static int fcoe_create(const char *buffer, struct kernel_param *kp)
{
	int rc;
1811
	struct fcoe_interface *fcoe;
1812
	struct fc_lport *lport;
1813 1814
	struct net_device *netdev;

1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827
	mutex_lock(&fcoe_config_mutex);
#ifdef CONFIG_FCOE_MODULE
	/*
	 * Make sure the module has been initialized, and is not about to be
	 * removed.  Module paramter sysfs files are writable before the
	 * module_init function is called and after module_exit.
	 */
	if (THIS_MODULE->state != MODULE_STATE_LIVE) {
		rc = -ENODEV;
		goto out_nodev;
	}
#endif

1828
	rtnl_lock();
1829 1830 1831 1832 1833
	netdev = fcoe_if_to_netdev(buffer);
	if (!netdev) {
		rc = -ENODEV;
		goto out_nodev;
	}
1834

1835 1836 1837 1838 1839 1840
	/* look for existing lport */
	if (fcoe_hostlist_lookup(netdev)) {
		rc = -EEXIST;
		goto out_putdev;
	}

1841 1842 1843 1844 1845 1846
	fcoe = fcoe_interface_create(netdev);
	if (!fcoe) {
		rc = -ENOMEM;
		goto out_putdev;
	}

1847
	lport = fcoe_if_create(fcoe, &netdev->dev, 0);
1848
	if (IS_ERR(lport)) {
1849
		printk(KERN_ERR "fcoe: Failed to create interface (%s)\n",
1850 1851
		       netdev->name);
		rc = -EIO;
1852
		fcoe_interface_cleanup(fcoe);
1853
		goto out_free;
1854
	}
1855

1856 1857 1858
	/* Make this the "master" N_Port */
	fcoe->ctlr.lp = lport;

1859 1860 1861
	/* add to lports list */
	fcoe_hostlist_add(lport);

1862 1863 1864 1865 1866
	/* start FIP Discovery and FLOGI */
	lport->boot_time = jiffies;
	fc_fabric_login(lport);
	if (!fcoe_link_ok(lport))
		fcoe_ctlr_link_up(&fcoe->ctlr);
1867

1868
	rc = 0;
1869
out_free:
1870 1871 1872 1873
	/*
	 * Release from init in fcoe_interface_create(), on success lport
	 * should be holding a reference taken in fcoe_if_create().
	 */
1874
	fcoe_interface_put(fcoe);
1875 1876 1877
out_putdev:
	dev_put(netdev);
out_nodev:
1878
	rtnl_unlock();
1879
	mutex_unlock(&fcoe_config_mutex);
1880 1881 1882 1883 1884
	return rc;
}

module_param_call(create, fcoe_create, NULL, NULL, S_IWUSR);
__MODULE_PARM_TYPE(create, "string");
1885
MODULE_PARM_DESC(create, "Create fcoe fcoe using net device passed in.");
1886 1887
module_param_call(destroy, fcoe_destroy, NULL, NULL, S_IWUSR);
__MODULE_PARM_TYPE(destroy, "string");
1888
MODULE_PARM_DESC(destroy, "Destroy fcoe fcoe");
1889

1890 1891
/**
 * fcoe_link_ok() - Check if link is ok for the fc_lport
1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905
 * @lp: ptr to the fc_lport
 *
 * Any permanently-disqualifying conditions have been previously checked.
 * This also updates the speed setting, which may change with link for 100/1000.
 *
 * This function should probably be checking for PAUSE support at some point
 * in the future. Currently Per-priority-pause is not determinable using
 * ethtool, so we shouldn't be restrictive until that problem is resolved.
 *
 * Returns: 0 if link is OK for use by FCoE.
 *
 */
int fcoe_link_ok(struct fc_lport *lp)
{
1906
	struct fcoe_port *port = lport_priv(lp);
1907
	struct net_device *dev = port->fcoe->netdev;
1908 1909
	struct ethtool_cmd ecmd = { ETHTOOL_GSET };

1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923
	if ((dev->flags & IFF_UP) && netif_carrier_ok(dev) &&
	    (!dev_ethtool_get_settings(dev, &ecmd))) {
		lp->link_supported_speeds &=
			~(FC_PORTSPEED_1GBIT | FC_PORTSPEED_10GBIT);
		if (ecmd.supported & (SUPPORTED_1000baseT_Half |
				      SUPPORTED_1000baseT_Full))
			lp->link_supported_speeds |= FC_PORTSPEED_1GBIT;
		if (ecmd.supported & SUPPORTED_10000baseT_Full)
			lp->link_supported_speeds |=
				FC_PORTSPEED_10GBIT;
		if (ecmd.speed == SPEED_1000)
			lp->link_speed = FC_PORTSPEED_1GBIT;
		if (ecmd.speed == SPEED_10000)
			lp->link_speed = FC_PORTSPEED_10GBIT;
1924

1925 1926 1927
		return 0;
	}
	return -1;
1928 1929
}

1930 1931
/**
 * fcoe_percpu_clean() - Clear the pending skbs for an lport
1932
 * @lp: the fc_lport
1933 1934 1935 1936 1937 1938 1939
 *
 * Must be called with fcoe_create_mutex held to single-thread completion.
 *
 * This flushes the pending skbs by adding a new skb to each queue and
 * waiting until they are all freed.  This assures us that not only are
 * there no packets that will be handled by the lport, but also that any
 * threads already handling packet have returned.
1940 1941 1942 1943 1944 1945 1946 1947
 */
void fcoe_percpu_clean(struct fc_lport *lp)
{
	struct fcoe_percpu_s *pp;
	struct fcoe_rcv_info *fr;
	struct sk_buff_head *list;
	struct sk_buff *skb, *next;
	struct sk_buff *head;
1948
	unsigned int cpu;
1949

1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961
	for_each_possible_cpu(cpu) {
		pp = &per_cpu(fcoe_percpu, cpu);
		spin_lock_bh(&pp->fcoe_rx_list.lock);
		list = &pp->fcoe_rx_list;
		head = list->next;
		for (skb = head; skb != (struct sk_buff *)list;
		     skb = next) {
			next = skb->next;
			fr = fcoe_dev_from_skb(skb);
			if (fr->fr_dev == lp) {
				__skb_unlink(skb, list);
				kfree_skb(skb);
1962 1963
			}
		}
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979

		if (!pp->thread || !cpu_online(cpu)) {
			spin_unlock_bh(&pp->fcoe_rx_list.lock);
			continue;
		}

		skb = dev_alloc_skb(0);
		if (!skb) {
			spin_unlock_bh(&pp->fcoe_rx_list.lock);
			continue;
		}
		skb->destructor = fcoe_percpu_flush_done;

		__skb_queue_tail(&pp->fcoe_rx_list, skb);
		if (pp->fcoe_rx_list.qlen == 1)
			wake_up_process(pp->thread);
1980
		spin_unlock_bh(&pp->fcoe_rx_list.lock);
1981 1982

		wait_for_completion(&fcoe_flush_completion);
1983 1984 1985 1986
	}
}

/**
1987
 * fcoe_clean_pending_queue() - Dequeue a skb and free it
1988 1989 1990
 * @lp: the corresponding fc_lport
 *
 * Returns: none
1991
 */
1992 1993
void fcoe_clean_pending_queue(struct fc_lport *lp)
{
1994
	struct fcoe_port  *port = lport_priv(lp);
1995 1996
	struct sk_buff *skb;

1997 1998 1999
	spin_lock_bh(&port->fcoe_pending_queue.lock);
	while ((skb = __skb_dequeue(&port->fcoe_pending_queue)) != NULL) {
		spin_unlock_bh(&port->fcoe_pending_queue.lock);
2000
		kfree_skb(skb);
2001
		spin_lock_bh(&port->fcoe_pending_queue.lock);
2002
	}
2003
	spin_unlock_bh(&port->fcoe_pending_queue.lock);
2004 2005
}

2006 2007
/**
 * fcoe_reset() - Resets the fcoe
2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
 * @shost: shost the reset is from
 *
 * Returns: always 0
 */
int fcoe_reset(struct Scsi_Host *shost)
{
	struct fc_lport *lport = shost_priv(shost);
	fc_lport_reset(lport);
	return 0;
}

2019
/**
2020
 * fcoe_hostlist_lookup_port() - find the corresponding lport by a given device
2021
 * @dev: this is currently ptr to net_device
2022
 *
2023
 * Returns: NULL or the located fcoe_port
2024
 * Locking: must be called with the RNL mutex held
2025
 */
2026 2027
static struct fcoe_interface *
fcoe_hostlist_lookup_port(const struct net_device *dev)
2028
{
2029
	struct fcoe_interface *fcoe;
2030

2031
	list_for_each_entry(fcoe, &fcoe_hostlist, list) {
2032
		if (fcoe->netdev == dev)
2033
			return fcoe;
2034 2035 2036 2037
	}
	return NULL;
}

2038 2039
/**
 * fcoe_hostlist_lookup() - Find the corresponding lport by netdev
2040 2041 2042
 * @netdev: ptr to net_device
 *
 * Returns: 0 for success
2043
 * Locking: must be called with the RTNL mutex held
2044
 */
2045
static struct fc_lport *fcoe_hostlist_lookup(const struct net_device *netdev)
2046
{
2047
	struct fcoe_interface *fcoe;
2048

2049
	fcoe = fcoe_hostlist_lookup_port(netdev);
2050
	return (fcoe) ? fcoe->ctlr.lp : NULL;
2051 2052
}

2053 2054
/**
 * fcoe_hostlist_add() - Add a lport to lports list
2055
 * @lp: ptr to the fc_lport to be added
2056 2057
 *
 * Returns: 0 for success
2058
 * Locking: must be called with the RTNL mutex held
2059
 */
2060
static int fcoe_hostlist_add(const struct fc_lport *lport)
2061
{
2062 2063 2064 2065 2066 2067 2068 2069
	struct fcoe_interface *fcoe;
	struct fcoe_port *port;

	fcoe = fcoe_hostlist_lookup_port(fcoe_netdev(lport));
	if (!fcoe) {
		port = lport_priv(lport);
		fcoe = port->fcoe;
		list_add_tail(&fcoe->list, &fcoe_hostlist);
2070 2071 2072 2073 2074
	}
	return 0;
}

/**
2075
 * fcoe_init() - fcoe module loading initialization
2076 2077
 *
 * Returns 0 on success, negative on failure
2078
 */
2079 2080
static int __init fcoe_init(void)
{
2081
	unsigned int cpu;
2082
	int rc = 0;
2083 2084
	struct fcoe_percpu_s *p;

2085 2086
	mutex_lock(&fcoe_config_mutex);

2087
	for_each_possible_cpu(cpu) {
2088
		p = &per_cpu(fcoe_percpu, cpu);
2089 2090 2091
		skb_queue_head_init(&p->fcoe_rx_list);
	}

2092 2093
	for_each_online_cpu(cpu)
		fcoe_percpu_thread_create(cpu);
2094

2095 2096 2097 2098
	/* Initialize per CPU interrupt thread */
	rc = register_hotcpu_notifier(&fcoe_cpu_notifier);
	if (rc)
		goto out_free;
2099

2100
	/* Setup link change notification */
2101 2102
	fcoe_dev_setup();

2103 2104 2105
	rc = fcoe_if_init();
	if (rc)
		goto out_free;
2106

2107
	mutex_unlock(&fcoe_config_mutex);
2108
	return 0;
2109 2110 2111 2112 2113

out_free:
	for_each_online_cpu(cpu) {
		fcoe_percpu_thread_destroy(cpu);
	}
2114
	mutex_unlock(&fcoe_config_mutex);
2115
	return rc;
2116 2117 2118 2119
}
module_init(fcoe_init);

/**
2120
 * fcoe_exit() - fcoe module unloading cleanup
2121 2122
 *
 * Returns 0 on success, negative on failure
2123
 */
2124 2125
static void __exit fcoe_exit(void)
{
2126
	unsigned int cpu;
2127
	struct fcoe_interface *fcoe, *tmp;
2128
	struct fcoe_port *port;
2129

2130 2131
	mutex_lock(&fcoe_config_mutex);

2132 2133
	fcoe_dev_cleanup();

2134
	/* releases the associated fcoe hosts */
2135 2136
	rtnl_lock();
	list_for_each_entry_safe(fcoe, tmp, &fcoe_hostlist, list) {
2137
		list_del(&fcoe->list);
2138 2139 2140
		port = lport_priv(fcoe->ctlr.lp);
		fcoe_interface_cleanup(fcoe);
		schedule_work(&port->destroy_work);
2141
	}
2142
	rtnl_unlock();
2143

2144 2145
	unregister_hotcpu_notifier(&fcoe_cpu_notifier);

2146
	for_each_online_cpu(cpu)
2147
		fcoe_percpu_thread_destroy(cpu);
2148

2149
	mutex_unlock(&fcoe_config_mutex);
2150 2151 2152 2153

	/* flush any asyncronous interface destroys,
	 * this should happen after the netdev notifier is unregistered */
	flush_scheduled_work();
2154 2155 2156
	/* That will flush out all the N_Ports on the hostlist, but now we
	 * may have NPIV VN_Ports scheduled for destruction */
	flush_scheduled_work();
2157 2158 2159 2160

	/* detach from scsi transport
	 * must happen after all destroys are done, therefor after the flush */
	fcoe_if_exit();
2161 2162
}
module_exit(fcoe_exit);
2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253

/**
 * fcoe_flogi_resp() - FCoE specific FLOGI and FDISC response handler
 * @seq: active sequence in the FLOGI or FDISC exchange
 * @fp: response frame, or error encoded in a pointer (timeout)
 * @arg: pointer the the fcoe_ctlr structure
 *
 * This handles MAC address managment for FCoE, then passes control on to
 * the libfc FLOGI response handler.
 */
static void fcoe_flogi_resp(struct fc_seq *seq, struct fc_frame *fp, void *arg)
{
	struct fcoe_ctlr *fip = arg;
	struct fc_exch *exch = fc_seq_exch(seq);
	struct fc_lport *lport = exch->lp;
	u8 *mac;

	if (IS_ERR(fp))
		goto done;

	mac = fr_cb(fp)->granted_mac;
	if (is_zero_ether_addr(mac)) {
		/* pre-FIP */
		mac = eth_hdr(&fp->skb)->h_source;
		if (fcoe_ctlr_recv_flogi(fip, lport, fp, mac)) {
			fc_frame_free(fp);
			return;
		}
	} else {
		/* FIP, libfcoe has already seen it */
		fip->update_mac(lport, fr_cb(fp)->granted_mac);
	}
done:
	fc_lport_flogi_resp(seq, fp, lport);
}

/**
 * fcoe_logo_resp() - FCoE specific LOGO response handler
 * @seq: active sequence in the LOGO exchange
 * @fp: response frame, or error encoded in a pointer (timeout)
 * @arg: pointer the the fcoe_ctlr structure
 *
 * This handles MAC address managment for FCoE, then passes control on to
 * the libfc LOGO response handler.
 */
static void fcoe_logo_resp(struct fc_seq *seq, struct fc_frame *fp, void *arg)
{
	struct fcoe_ctlr *fip = arg;
	struct fc_exch *exch = fc_seq_exch(seq);
	struct fc_lport *lport = exch->lp;
	static u8 zero_mac[ETH_ALEN] = { 0 };

	if (!IS_ERR(fp))
		fip->update_mac(lport, zero_mac);
	fc_lport_logo_resp(seq, fp, lport);
}

/**
 * fcoe_elsct_send - FCoE specific ELS handler
 *
 * This does special case handling of FIP encapsualted ELS exchanges for FCoE,
 * using FCoE specific response handlers and passing the FIP controller as
 * the argument (the lport is still available from the exchange).
 *
 * Most of the work here is just handed off to the libfc routine.
 */
static struct fc_seq *fcoe_elsct_send(struct fc_lport *lport,
		u32 did, struct fc_frame *fp, unsigned int op,
		void (*resp)(struct fc_seq *, struct fc_frame *, void *),
		void *arg, u32 timeout)
{
	struct fcoe_port *port = lport_priv(lport);
	struct fcoe_interface *fcoe = port->fcoe;
	struct fcoe_ctlr *fip = &fcoe->ctlr;
	struct fc_frame_header *fh = fc_frame_header_get(fp);

	switch (op) {
	case ELS_FLOGI:
	case ELS_FDISC:
		return fc_elsct_send(lport, did, fp, op, fcoe_flogi_resp,
				     fip, timeout);
	case ELS_LOGO:
		/* only hook onto fabric logouts, not port logouts */
		if (ntoh24(fh->fh_d_id) != FC_FID_FLOGI)
			break;
		return fc_elsct_send(lport, did, fp, op, fcoe_logo_resp,
				     fip, timeout);
	}
	return fc_elsct_send(lport, did, fp, op, resp, arg, timeout);
}

2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330
/**
 * fcoe_vport_create() - create an fc_host/scsi_host for a vport
 * @vport: fc_vport object to create a new fc_host for
 * @disabled: start the new fc_host in a disabled state by default?
 *
 * Returns: 0 for success
 */
static int fcoe_vport_create(struct fc_vport *vport, bool disabled)
{
	struct Scsi_Host *shost = vport_to_shost(vport);
	struct fc_lport *n_port = shost_priv(shost);
	struct fcoe_port *port = lport_priv(n_port);
	struct fcoe_interface *fcoe = port->fcoe;
	struct net_device *netdev = fcoe->netdev;
	struct fc_lport *vn_port;

	mutex_lock(&fcoe_config_mutex);
	vn_port = fcoe_if_create(fcoe, &vport->dev, 1);
	mutex_unlock(&fcoe_config_mutex);

	if (IS_ERR(vn_port)) {
		printk(KERN_ERR "fcoe: fcoe_vport_create(%s) failed\n",
		       netdev->name);
		return -EIO;
	}

	if (disabled) {
		fc_vport_set_state(vport, FC_VPORT_DISABLED);
	} else {
		vn_port->boot_time = jiffies;
		fc_fabric_login(vn_port);
		fc_vport_setlink(vn_port);
	}
	return 0;
}

/**
 * fcoe_vport_destroy() - destroy the fc_host/scsi_host for a vport
 * @vport: fc_vport object that is being destroyed
 *
 * Returns: 0 for success
 */
static int fcoe_vport_destroy(struct fc_vport *vport)
{
	struct Scsi_Host *shost = vport_to_shost(vport);
	struct fc_lport *n_port = shost_priv(shost);
	struct fc_lport *vn_port = vport->dd_data;
	struct fcoe_port *port = lport_priv(vn_port);

	mutex_lock(&n_port->lp_mutex);
	list_del(&vn_port->list);
	mutex_unlock(&n_port->lp_mutex);
	schedule_work(&port->destroy_work);
	return 0;
}

/**
 * fcoe_vport_disable() - change vport state
 * @vport: vport to bring online/offline
 * @disable: should the vport be disabled?
 */
static int fcoe_vport_disable(struct fc_vport *vport, bool disable)
{
	struct fc_lport *lport = vport->dd_data;

	if (disable) {
		fc_vport_set_state(vport, FC_VPORT_DISABLED);
		fc_fabric_logoff(lport);
	} else {
		lport->boot_time = jiffies;
		fc_fabric_login(lport);
		fc_vport_setlink(lport);
	}

	return 0;
}

2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361
/**
 * fcoe_vport_set_symbolic_name() - append vport string to symbolic name
 * @vport: fc_vport with a new symbolic name string
 *
 * After generating a new symbolic name string, a new RSPN_ID request is
 * sent to the name server.  There is no response handler, so if it fails
 * for some reason it will not be retried.
 */
static void fcoe_set_vport_symbolic_name(struct fc_vport *vport)
{
	struct fc_lport *lport = vport->dd_data;
	struct fc_frame *fp;
	size_t len;

	snprintf(fc_host_symbolic_name(lport->host), FC_SYMBOLIC_NAME_SIZE,
		 "%s v%s over %s : %s", FCOE_NAME, FCOE_VERSION,
		 fcoe_netdev(lport)->name, vport->symbolic_name);

	if (lport->state != LPORT_ST_READY)
		return;

	len = strnlen(fc_host_symbolic_name(lport->host), 255);
	fp = fc_frame_alloc(lport,
			    sizeof(struct fc_ct_hdr) +
			    sizeof(struct fc_ns_rspn) + len);
	if (!fp)
		return;
	lport->tt.elsct_send(lport, FC_FID_DIR_SERV, fp, FC_NS_RSPN_ID,
			     NULL, NULL, lport->e_d_tov);
}