ixgbe_main.c 189.2 KB
Newer Older
1 2 3
/*******************************************************************************

  Intel 10 Gigabit PCI Express Linux driver
S
Shannon Nelson 已提交
4
  Copyright(c) 1999 - 2010 Intel Corporation.
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

  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.

  The full GNU General Public License is included in this distribution in
  the file called "COPYING".

  Contact Information:
  e1000-devel Mailing List <e1000-devel@lists.sourceforge.net>
  Intel Corporation, 5200 N.E. Elam Young Parkway, Hillsboro, OR 97124-6497

*******************************************************************************/

#include <linux/types.h>
#include <linux/module.h>
#include <linux/pci.h>
#include <linux/netdevice.h>
#include <linux/vmalloc.h>
#include <linux/string.h>
#include <linux/in.h>
#include <linux/ip.h>
#include <linux/tcp.h>
37
#include <linux/pkt_sched.h>
38 39 40 41 42
#include <linux/ipv6.h>
#include <net/checksum.h>
#include <net/ip6_checksum.h>
#include <linux/ethtool.h>
#include <linux/if_vlan.h>
43
#include <scsi/fc/fc_fcoe.h>
44 45 46

#include "ixgbe.h"
#include "ixgbe_common.h"
47
#include "ixgbe_dcb_82599.h"
48
#include "ixgbe_sriov.h"
49 50

char ixgbe_driver_name[] = "ixgbe";
S
Stephen Hemminger 已提交
51
static const char ixgbe_driver_string[] =
52
                              "Intel(R) 10 Gigabit PCI Express Network Driver";
53

P
Peter Waskiewicz 已提交
54
#define DRV_VERSION "2.0.62-k2"
S
Stephen Hemminger 已提交
55
const char ixgbe_driver_version[] = DRV_VERSION;
S
Shannon Nelson 已提交
56
static char ixgbe_copyright[] = "Copyright (c) 1999-2010 Intel Corporation.";
57 58

static const struct ixgbe_info *ixgbe_info_tbl[] = {
59
	[board_82598] = &ixgbe_82598_info,
60
	[board_82599] = &ixgbe_82599_info,
61 62 63 64 65 66 67 68 69 70
};

/* ixgbe_pci_tbl - PCI Device ID Table
 *
 * Wildcard entries (PCI_ANY_ID) should come last
 * Last entry must be all 0s
 *
 * { Vendor ID, Device ID, SubVendor ID, SubDevice ID,
 *   Class, Class Mask, private data (not used) }
 */
71
static DEFINE_PCI_DEVICE_TABLE(ixgbe_pci_tbl) = {
D
Don Skidmore 已提交
72 73
	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82598),
	 board_82598 },
74
	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82598AF_DUAL_PORT),
75
	 board_82598 },
76
	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82598AF_SINGLE_PORT),
77
	 board_82598 },
78 79
	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82598AT),
	 board_82598 },
80 81
	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82598AT2),
	 board_82598 },
82
	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82598EB_CX4),
83
	 board_82598 },
J
Jesse Brandeburg 已提交
84 85
	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82598_CX4_DUAL_PORT),
	 board_82598 },
D
Donald Skidmore 已提交
86 87 88 89
	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82598_DA_DUAL_PORT),
	 board_82598 },
	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82598_SR_DUAL_PORT_EM),
	 board_82598 },
90 91
	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82598EB_XF_LR),
	 board_82598 },
D
Donald Skidmore 已提交
92 93
	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82598EB_SFP_LOM),
	 board_82598 },
94 95
	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82598_BX),
	 board_82598 },
96 97
	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82599_KX4),
	 board_82599 },
98 99
	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82599_XAUI_LOM),
	 board_82599 },
100 101
	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82599_KR),
	 board_82599 },
102 103
	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82599_SFP),
	 board_82599 },
104 105
	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82599_SFP_EM),
	 board_82599 },
106 107
	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82599_KX4_MEZZ),
	 board_82599 },
108 109
	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82599_CX4),
	 board_82599 },
110 111
	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82599_COMBO_BACKPLANE),
	 board_82599 },
112 113 114 115 116 117

	/* required last entry */
	{0, }
};
MODULE_DEVICE_TABLE(pci, ixgbe_pci_tbl);

118
#ifdef CONFIG_IXGBE_DCA
119
static int ixgbe_notify_dca(struct notifier_block *, unsigned long event,
120
                            void *p);
121 122 123 124 125 126 127
static struct notifier_block dca_notifier = {
	.notifier_call = ixgbe_notify_dca,
	.next          = NULL,
	.priority      = 0
};
#endif

128 129 130 131 132 133 134
#ifdef CONFIG_PCI_IOV
static unsigned int max_vfs;
module_param(max_vfs, uint, 0);
MODULE_PARM_DESC(max_vfs, "Maximum number of virtual functions to allocate "
                 "per physical function");
#endif /* CONFIG_PCI_IOV */

135 136 137 138 139 140 141
MODULE_AUTHOR("Intel Corporation, <linux.nics@intel.com>");
MODULE_DESCRIPTION("Intel(R) 10 Gigabit PCI Express Network Driver");
MODULE_LICENSE("GPL");
MODULE_VERSION(DRV_VERSION);

#define DEFAULT_DEBUG_LEVEL_SHIFT 3

142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176
static inline void ixgbe_disable_sriov(struct ixgbe_adapter *adapter)
{
	struct ixgbe_hw *hw = &adapter->hw;
	u32 gcr;
	u32 gpie;
	u32 vmdctl;

#ifdef CONFIG_PCI_IOV
	/* disable iov and allow time for transactions to clear */
	pci_disable_sriov(adapter->pdev);
#endif

	/* turn off device IOV mode */
	gcr = IXGBE_READ_REG(hw, IXGBE_GCR_EXT);
	gcr &= ~(IXGBE_GCR_EXT_SRIOV);
	IXGBE_WRITE_REG(hw, IXGBE_GCR_EXT, gcr);
	gpie = IXGBE_READ_REG(hw, IXGBE_GPIE);
	gpie &= ~IXGBE_GPIE_VTMODE_MASK;
	IXGBE_WRITE_REG(hw, IXGBE_GPIE, gpie);

	/* set default pool back to 0 */
	vmdctl = IXGBE_READ_REG(hw, IXGBE_VT_CTL);
	vmdctl &= ~IXGBE_VT_CTL_POOL_MASK;
	IXGBE_WRITE_REG(hw, IXGBE_VT_CTL, vmdctl);

	/* take a breather then clean up driver data */
	msleep(100);
	if (adapter->vfinfo)
		kfree(adapter->vfinfo);
	adapter->vfinfo = NULL;

	adapter->num_vfs = 0;
	adapter->flags &= ~IXGBE_FLAG_SRIOV_ENABLED;
}

177 178 179 180 181 182 183
static void ixgbe_release_hw_control(struct ixgbe_adapter *adapter)
{
	u32 ctrl_ext;

	/* Let firmware take over control of h/w */
	ctrl_ext = IXGBE_READ_REG(&adapter->hw, IXGBE_CTRL_EXT);
	IXGBE_WRITE_REG(&adapter->hw, IXGBE_CTRL_EXT,
184
	                ctrl_ext & ~IXGBE_CTRL_EXT_DRV_LOAD);
185 186 187 188 189 190 191 192 193
}

static void ixgbe_get_hw_control(struct ixgbe_adapter *adapter)
{
	u32 ctrl_ext;

	/* Let firmware know the driver has taken over */
	ctrl_ext = IXGBE_READ_REG(&adapter->hw, IXGBE_CTRL_EXT);
	IXGBE_WRITE_REG(&adapter->hw, IXGBE_CTRL_EXT,
194
	                ctrl_ext | IXGBE_CTRL_EXT_DRV_LOAD);
195
}
196

197 198 199 200 201 202 203 204 205 206
/*
 * ixgbe_set_ivar - set the IVAR registers, mapping interrupt causes to vectors
 * @adapter: pointer to adapter struct
 * @direction: 0 for Rx, 1 for Tx, -1 for other causes
 * @queue: queue to map the corresponding interrupt to
 * @msix_vector: the vector to map to the corresponding queue
 *
 */
static void ixgbe_set_ivar(struct ixgbe_adapter *adapter, s8 direction,
	                   u8 queue, u8 msix_vector)
207 208
{
	u32 ivar, index;
209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243
	struct ixgbe_hw *hw = &adapter->hw;
	switch (hw->mac.type) {
	case ixgbe_mac_82598EB:
		msix_vector |= IXGBE_IVAR_ALLOC_VAL;
		if (direction == -1)
			direction = 0;
		index = (((direction * 64) + queue) >> 2) & 0x1F;
		ivar = IXGBE_READ_REG(hw, IXGBE_IVAR(index));
		ivar &= ~(0xFF << (8 * (queue & 0x3)));
		ivar |= (msix_vector << (8 * (queue & 0x3)));
		IXGBE_WRITE_REG(hw, IXGBE_IVAR(index), ivar);
		break;
	case ixgbe_mac_82599EB:
		if (direction == -1) {
			/* other causes */
			msix_vector |= IXGBE_IVAR_ALLOC_VAL;
			index = ((queue & 1) * 8);
			ivar = IXGBE_READ_REG(&adapter->hw, IXGBE_IVAR_MISC);
			ivar &= ~(0xFF << index);
			ivar |= (msix_vector << index);
			IXGBE_WRITE_REG(&adapter->hw, IXGBE_IVAR_MISC, ivar);
			break;
		} else {
			/* tx or rx causes */
			msix_vector |= IXGBE_IVAR_ALLOC_VAL;
			index = ((16 * (queue & 1)) + (8 * direction));
			ivar = IXGBE_READ_REG(hw, IXGBE_IVAR(queue >> 1));
			ivar &= ~(0xFF << index);
			ivar |= (msix_vector << index);
			IXGBE_WRITE_REG(hw, IXGBE_IVAR(queue >> 1), ivar);
			break;
		}
	default:
		break;
	}
244 245
}

246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261
static inline void ixgbe_irq_rearm_queues(struct ixgbe_adapter *adapter,
                                          u64 qmask)
{
	u32 mask;

	if (adapter->hw.mac.type == ixgbe_mac_82598EB) {
		mask = (IXGBE_EIMS_RTX_QUEUE & qmask);
		IXGBE_WRITE_REG(&adapter->hw, IXGBE_EICS, mask);
	} else {
		mask = (qmask & 0xFFFFFFFF);
		IXGBE_WRITE_REG(&adapter->hw, IXGBE_EICS_EX(0), mask);
		mask = (qmask >> 32);
		IXGBE_WRITE_REG(&adapter->hw, IXGBE_EICS_EX(1), mask);
	}
}

262
static void ixgbe_unmap_and_free_tx_resource(struct ixgbe_adapter *adapter,
263 264
                                             struct ixgbe_tx_buffer
                                             *tx_buffer_info)
265
{
266 267 268 269 270 271 272 273 274 275 276 277 278
	if (tx_buffer_info->dma) {
		if (tx_buffer_info->mapped_as_page)
			pci_unmap_page(adapter->pdev,
				       tx_buffer_info->dma,
				       tx_buffer_info->length,
				       PCI_DMA_TODEVICE);
		else
			pci_unmap_single(adapter->pdev,
					 tx_buffer_info->dma,
					 tx_buffer_info->length,
					 PCI_DMA_TODEVICE);
		tx_buffer_info->dma = 0;
	}
279 280 281 282
	if (tx_buffer_info->skb) {
		dev_kfree_skb_any(tx_buffer_info->skb);
		tx_buffer_info->skb = NULL;
	}
283
	tx_buffer_info->time_stamp = 0;
284 285 286
	/* tx_buffer_info must be completely set up in the transmit path */
}

287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303
/**
 * ixgbe_tx_is_paused - check if the tx ring is paused
 * @adapter: the ixgbe adapter
 * @tx_ring: the corresponding tx_ring
 *
 * If not in DCB mode, checks TFCS.TXOFF, otherwise, find out the
 * corresponding TC of this tx_ring when checking TFCS.
 *
 * Returns : true if paused
 */
static inline bool ixgbe_tx_is_paused(struct ixgbe_adapter *adapter,
                                      struct ixgbe_ring *tx_ring)
{
	u32 txoff = IXGBE_TFCS_TXOFF;

#ifdef CONFIG_IXGBE_DCB
	if (adapter->flags & IXGBE_FLAG_DCB_ENABLED) {
304
		int tc;
305 306 307
		int reg_idx = tx_ring->reg_idx;
		int dcb_i = adapter->ring_feature[RING_F_DCB].indices;

308 309
		switch (adapter->hw.mac.type) {
		case ixgbe_mac_82598EB:
310 311
			tc = reg_idx >> 2;
			txoff = IXGBE_TFCS_TXOFF0;
312 313
			break;
		case ixgbe_mac_82599EB:
314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331
			tc = 0;
			txoff = IXGBE_TFCS_TXOFF;
			if (dcb_i == 8) {
				/* TC0, TC1 */
				tc = reg_idx >> 5;
				if (tc == 2) /* TC2, TC3 */
					tc += (reg_idx - 64) >> 4;
				else if (tc == 3) /* TC4, TC5, TC6, TC7 */
					tc += 1 + ((reg_idx - 96) >> 3);
			} else if (dcb_i == 4) {
				/* TC0, TC1 */
				tc = reg_idx >> 6;
				if (tc == 1) {
					tc += (reg_idx - 64) >> 5;
					if (tc == 2) /* TC2, TC3 */
						tc += (reg_idx - 96) >> 4;
				}
			}
332 333 334
			break;
		default:
			tc = 0;
335 336 337 338 339 340 341
		}
		txoff <<= tc;
	}
#endif
	return IXGBE_READ_REG(&adapter->hw, IXGBE_TFCS) & txoff;
}

342
static inline bool ixgbe_check_tx_hang(struct ixgbe_adapter *adapter,
343 344
                                       struct ixgbe_ring *tx_ring,
                                       unsigned int eop)
345
{
346 347
	struct ixgbe_hw *hw = &adapter->hw;

348
	/* Detect a transmit hang in hardware, this serializes the
349
	 * check with the clearing of time_stamp and movement of eop */
350
	adapter->detect_tx_hung = false;
351
	if (tx_ring->tx_buffer_info[eop].time_stamp &&
352
	    time_after(jiffies, tx_ring->tx_buffer_info[eop].time_stamp + HZ) &&
353
	    !ixgbe_tx_is_paused(adapter, tx_ring)) {
354
		/* detected Tx unit hang */
355 356
		union ixgbe_adv_tx_desc *tx_desc;
		tx_desc = IXGBE_TX_DESC_ADV(*tx_ring, eop);
357
		DPRINTK(DRV, ERR, "Detected Tx Unit Hang\n"
358 359
			"  Tx Queue             <%d>\n"
			"  TDH, TDT             <%x>, <%x>\n"
360 361 362 363
			"  next_to_use          <%x>\n"
			"  next_to_clean        <%x>\n"
			"tx_buffer_info[next_to_clean]\n"
			"  time_stamp           <%lx>\n"
364 365
			"  jiffies              <%lx>\n",
			tx_ring->queue_index,
366 367
			IXGBE_READ_REG(hw, tx_ring->head),
			IXGBE_READ_REG(hw, tx_ring->tail),
368 369
			tx_ring->next_to_use, eop,
			tx_ring->tx_buffer_info[eop].time_stamp, jiffies);
370 371 372 373 374 375
		return true;
	}

	return false;
}

376 377
#define IXGBE_MAX_TXD_PWR       14
#define IXGBE_MAX_DATA_PER_TXD  (1 << IXGBE_MAX_TXD_PWR)
378 379 380 381 382

/* Tx Descriptors needed, worst case */
#define TXD_USE_COUNT(S) (((S) >> IXGBE_MAX_TXD_PWR) + \
			 (((S) & (IXGBE_MAX_DATA_PER_TXD - 1)) ? 1 : 0))
#define DESC_NEEDED (TXD_USE_COUNT(IXGBE_MAX_DATA_PER_TXD) /* skb->data */ + \
383
	MAX_SKB_FRAGS * TXD_USE_COUNT(PAGE_SIZE) + 1) /* for context */
384

385 386
static void ixgbe_tx_timeout(struct net_device *netdev);

387 388
/**
 * ixgbe_clean_tx_irq - Reclaim resources after transmit completes
389
 * @q_vector: structure containing interrupt and ring information
390
 * @tx_ring: tx ring to clean
391
 **/
392
static bool ixgbe_clean_tx_irq(struct ixgbe_q_vector *q_vector,
393
                               struct ixgbe_ring *tx_ring)
394
{
395
	struct ixgbe_adapter *adapter = q_vector->adapter;
396
	struct net_device *netdev = adapter->netdev;
397 398 399
	union ixgbe_adv_tx_desc *tx_desc, *eop_desc;
	struct ixgbe_tx_buffer *tx_buffer_info;
	unsigned int i, eop, count = 0;
400
	unsigned int total_bytes = 0, total_packets = 0;
401 402

	i = tx_ring->next_to_clean;
403 404 405 406
	eop = tx_ring->tx_buffer_info[i].next_to_watch;
	eop_desc = IXGBE_TX_DESC_ADV(*tx_ring, eop);

	while ((eop_desc->wb.status & cpu_to_le32(IXGBE_TXD_STAT_DD)) &&
407
	       (count < tx_ring->work_limit)) {
408 409 410
		bool cleaned = false;
		for ( ; !cleaned; count++) {
			struct sk_buff *skb;
411 412
			tx_desc = IXGBE_TX_DESC_ADV(*tx_ring, i);
			tx_buffer_info = &tx_ring->tx_buffer_info[i];
413
			cleaned = (i == eop);
414
			skb = tx_buffer_info->skb;
415

416
			if (cleaned && skb) {
417
				unsigned int segs, bytecount;
418
				unsigned int hlen = skb_headlen(skb);
419 420

				/* gso_segs is currently only valid for tcp */
421
				segs = skb_shinfo(skb)->gso_segs ?: 1;
422 423 424 425 426 427 428 429 430 431 432 433
#ifdef IXGBE_FCOE
				/* adjust for FCoE Sequence Offload */
				if ((adapter->flags & IXGBE_FLAG_FCOE_ENABLED)
				    && (skb->protocol == htons(ETH_P_FCOE)) &&
				    skb_is_gso(skb)) {
					hlen = skb_transport_offset(skb) +
						sizeof(struct fc_frame_header) +
						sizeof(struct fcoe_crc_eof);
					segs = DIV_ROUND_UP(skb->len - hlen,
						skb_shinfo(skb)->gso_size);
				}
#endif /* IXGBE_FCOE */
434
				/* multiply data chunks by size of headers */
435
				bytecount = ((segs - 1) * hlen) + skb->len;
436 437
				total_packets += segs;
				total_bytes += bytecount;
438
			}
439

440
			ixgbe_unmap_and_free_tx_resource(adapter,
441
			                                 tx_buffer_info);
442

443 444
			tx_desc->wb.status = 0;

445 446 447
			i++;
			if (i == tx_ring->count)
				i = 0;
448
		}
449 450 451 452 453

		eop = tx_ring->tx_buffer_info[i].next_to_watch;
		eop_desc = IXGBE_TX_DESC_ADV(*tx_ring, eop);
	}

454 455
	tx_ring->next_to_clean = i;

456
#define TX_WAKE_THRESHOLD (DESC_NEEDED * 2)
457 458
	if (unlikely(count && netif_carrier_ok(netdev) &&
	             (IXGBE_DESC_UNUSED(tx_ring) >= TX_WAKE_THRESHOLD))) {
459 460 461 462
		/* Make sure that anybody stopping the queue after this
		 * sees the new next_to_clean.
		 */
		smp_mb();
463 464 465
		if (__netif_subqueue_stopped(netdev, tx_ring->queue_index) &&
		    !test_bit(__IXGBE_DOWN, &adapter->state)) {
			netif_wake_subqueue(netdev, tx_ring->queue_index);
J
Jesse Brandeburg 已提交
466
			++tx_ring->restart_queue;
467
		}
468
	}
469

470 471 472 473 474 475 476 477 478
	if (adapter->detect_tx_hung) {
		if (ixgbe_check_tx_hang(adapter, tx_ring, i)) {
			/* schedule immediate reset if we believe we hung */
			DPRINTK(PROBE, INFO,
			        "tx hang %d detected, resetting adapter\n",
			        adapter->tx_timeout_count + 1);
			ixgbe_tx_timeout(adapter->netdev);
		}
	}
479

480
	/* re-arm the interrupt */
481 482
	if (count >= tx_ring->work_limit)
		ixgbe_irq_rearm_queues(adapter, ((u64)1 << q_vector->v_idx));
483

484 485 486
	tx_ring->total_bytes += total_bytes;
	tx_ring->total_packets += total_packets;
	tx_ring->stats.packets += total_packets;
487
	tx_ring->stats.bytes += total_bytes;
488
	return (count < tx_ring->work_limit);
489 490
}

491
#ifdef CONFIG_IXGBE_DCA
492
static void ixgbe_update_rx_dca(struct ixgbe_adapter *adapter,
493
                                struct ixgbe_ring *rx_ring)
494 495 496
{
	u32 rxctrl;
	int cpu = get_cpu();
497
	int q = rx_ring->reg_idx;
498

499
	if (rx_ring->cpu != cpu) {
500
		rxctrl = IXGBE_READ_REG(&adapter->hw, IXGBE_DCA_RXCTRL(q));
501 502 503 504 505 506 507 508
		if (adapter->hw.mac.type == ixgbe_mac_82598EB) {
			rxctrl &= ~IXGBE_DCA_RXCTRL_CPUID_MASK;
			rxctrl |= dca3_get_tag(&adapter->pdev->dev, cpu);
		} else if (adapter->hw.mac.type == ixgbe_mac_82599EB) {
			rxctrl &= ~IXGBE_DCA_RXCTRL_CPUID_MASK_82599;
			rxctrl |= (dca3_get_tag(&adapter->pdev->dev, cpu) <<
			           IXGBE_DCA_RXCTRL_CPUID_SHIFT_82599);
		}
509 510
		rxctrl |= IXGBE_DCA_RXCTRL_DESC_DCA_EN;
		rxctrl |= IXGBE_DCA_RXCTRL_HEAD_DCA_EN;
511 512
		rxctrl &= ~(IXGBE_DCA_RXCTRL_DESC_RRO_EN);
		rxctrl &= ~(IXGBE_DCA_RXCTRL_DESC_WRO_EN |
513
		            IXGBE_DCA_RXCTRL_DESC_HSRO_EN);
514
		IXGBE_WRITE_REG(&adapter->hw, IXGBE_DCA_RXCTRL(q), rxctrl);
515
		rx_ring->cpu = cpu;
516 517 518 519 520
	}
	put_cpu();
}

static void ixgbe_update_tx_dca(struct ixgbe_adapter *adapter,
521
                                struct ixgbe_ring *tx_ring)
522 523 524
{
	u32 txctrl;
	int cpu = get_cpu();
525
	int q = tx_ring->reg_idx;
526
	struct ixgbe_hw *hw = &adapter->hw;
527

528
	if (tx_ring->cpu != cpu) {
529
		if (adapter->hw.mac.type == ixgbe_mac_82598EB) {
530
			txctrl = IXGBE_READ_REG(hw, IXGBE_DCA_TXCTRL(q));
531 532
			txctrl &= ~IXGBE_DCA_TXCTRL_CPUID_MASK;
			txctrl |= dca3_get_tag(&adapter->pdev->dev, cpu);
533 534
			txctrl |= IXGBE_DCA_TXCTRL_DESC_DCA_EN;
			IXGBE_WRITE_REG(hw, IXGBE_DCA_TXCTRL(q), txctrl);
535
		} else if (adapter->hw.mac.type == ixgbe_mac_82599EB) {
536
			txctrl = IXGBE_READ_REG(hw, IXGBE_DCA_TXCTRL_82599(q));
537 538
			txctrl &= ~IXGBE_DCA_TXCTRL_CPUID_MASK_82599;
			txctrl |= (dca3_get_tag(&adapter->pdev->dev, cpu) <<
539 540 541
			          IXGBE_DCA_TXCTRL_CPUID_SHIFT_82599);
			txctrl |= IXGBE_DCA_TXCTRL_DESC_DCA_EN;
			IXGBE_WRITE_REG(hw, IXGBE_DCA_TXCTRL_82599(q), txctrl);
542
		}
543
		tx_ring->cpu = cpu;
544 545 546 547 548 549 550 551 552 553 554
	}
	put_cpu();
}

static void ixgbe_setup_dca(struct ixgbe_adapter *adapter)
{
	int i;

	if (!(adapter->flags & IXGBE_FLAG_DCA_ENABLED))
		return;

555 556 557
	/* always use CB2 mode, difference is masked in the CB driver */
	IXGBE_WRITE_REG(&adapter->hw, IXGBE_DCA_CTRL, 2);

558
	for (i = 0; i < adapter->num_tx_queues; i++) {
559 560
		adapter->tx_ring[i]->cpu = -1;
		ixgbe_update_tx_dca(adapter, adapter->tx_ring[i]);
561 562
	}
	for (i = 0; i < adapter->num_rx_queues; i++) {
563 564
		adapter->rx_ring[i]->cpu = -1;
		ixgbe_update_rx_dca(adapter, adapter->rx_ring[i]);
565 566 567 568 569 570 571 572 573 574 575
	}
}

static int __ixgbe_notify_dca(struct device *dev, void *data)
{
	struct net_device *netdev = dev_get_drvdata(dev);
	struct ixgbe_adapter *adapter = netdev_priv(netdev);
	unsigned long event = *(unsigned long *)data;

	switch (event) {
	case DCA_PROVIDER_ADD:
576 577 578
		/* if we're already enabled, don't do it again */
		if (adapter->flags & IXGBE_FLAG_DCA_ENABLED)
			break;
579
		if (dca_add_requester(dev) == 0) {
580
			adapter->flags |= IXGBE_FLAG_DCA_ENABLED;
581 582 583 584 585 586 587 588 589 590 591 592 593
			ixgbe_setup_dca(adapter);
			break;
		}
		/* Fall Through since DCA is disabled. */
	case DCA_PROVIDER_REMOVE:
		if (adapter->flags & IXGBE_FLAG_DCA_ENABLED) {
			dca_remove_requester(dev);
			adapter->flags &= ~IXGBE_FLAG_DCA_ENABLED;
			IXGBE_WRITE_REG(&adapter->hw, IXGBE_DCA_CTRL, 1);
		}
		break;
	}

594
	return 0;
595 596
}

597
#endif /* CONFIG_IXGBE_DCA */
598 599 600 601
/**
 * ixgbe_receive_skb - Send a completed packet up the stack
 * @adapter: board private structure
 * @skb: packet to send up
602 603 604
 * @status: hardware indication of status of receive
 * @rx_ring: rx descriptor ring (for a specific queue) to setup
 * @rx_desc: rx descriptor
605
 **/
H
Herbert Xu 已提交
606
static void ixgbe_receive_skb(struct ixgbe_q_vector *q_vector,
607
                              struct sk_buff *skb, u8 status,
608
                              struct ixgbe_ring *ring,
609
                              union ixgbe_adv_rx_desc *rx_desc)
610
{
H
Herbert Xu 已提交
611 612
	struct ixgbe_adapter *adapter = q_vector->adapter;
	struct napi_struct *napi = &q_vector->napi;
613 614
	bool is_vlan = (status & IXGBE_RXD_STAT_VP);
	u16 tag = le16_to_cpu(rx_desc->wb.upper.vlan);
615

616
	skb_record_rx_queue(skb, ring->queue_index);
617
	if (!(adapter->flags & IXGBE_FLAG_IN_NETPOLL)) {
618
		if (adapter->vlgrp && is_vlan && (tag & VLAN_VID_MASK))
H
Herbert Xu 已提交
619
			vlan_gro_receive(napi, adapter->vlgrp, tag, skb);
620
		else
H
Herbert Xu 已提交
621
			napi_gro_receive(napi, skb);
622
	} else {
623
		if (adapter->vlgrp && is_vlan && (tag & VLAN_VID_MASK))
624 625 626
			vlan_hwaccel_rx(skb, adapter->vlgrp, tag);
		else
			netif_rx(skb);
627 628 629
	}
}

630 631 632 633 634 635
/**
 * ixgbe_rx_checksum - indicate in skb if hw indicated a good cksum
 * @adapter: address of board private structure
 * @status_err: hardware indication of status of receive
 * @skb: skb currently being received and modified
 **/
636
static inline void ixgbe_rx_checksum(struct ixgbe_adapter *adapter,
637 638
				     union ixgbe_adv_rx_desc *rx_desc,
				     struct sk_buff *skb)
639
{
640 641
	u32 status_err = le32_to_cpu(rx_desc->wb.upper.status_error);

642 643
	skb->ip_summed = CHECKSUM_NONE;

644 645
	/* Rx csum disabled */
	if (!(adapter->flags & IXGBE_FLAG_RX_CSUM_ENABLED))
646
		return;
647 648 649 650

	/* if IP and error */
	if ((status_err & IXGBE_RXD_STAT_IPCS) &&
	    (status_err & IXGBE_RXDADV_ERR_IPE)) {
651 652 653
		adapter->hw_csum_rx_error++;
		return;
	}
654 655 656 657 658

	if (!(status_err & IXGBE_RXD_STAT_L4CS))
		return;

	if (status_err & IXGBE_RXDADV_ERR_TCPE) {
659 660 661 662 663 664 665 666 667 668
		u16 pkt_info = rx_desc->wb.lower.lo_dword.hs_rss.pkt_info;

		/*
		 * 82599 errata, UDP frames with a 0 checksum can be marked as
		 * checksum errors.
		 */
		if ((pkt_info & IXGBE_RXDADV_PKTTYPE_UDP) &&
		    (adapter->hw.mac.type == ixgbe_mac_82599EB))
			return;

669 670 671 672
		adapter->hw_csum_rx_error++;
		return;
	}

673
	/* It must be a TCP or UDP packet with a valid checksum */
674
	skb->ip_summed = CHECKSUM_UNNECESSARY;
675 676
}

677 678 679 680 681 682 683 684 685 686 687 688 689
static inline void ixgbe_release_rx_desc(struct ixgbe_hw *hw,
                                         struct ixgbe_ring *rx_ring, u32 val)
{
	/*
	 * Force memory writes to complete before letting h/w
	 * know there are new descriptors to fetch.  (Only
	 * applicable for weak-ordered memory model archs,
	 * such as IA-64).
	 */
	wmb();
	IXGBE_WRITE_REG(hw, IXGBE_RDT(rx_ring->reg_idx), val);
}

690 691 692 693 694
/**
 * ixgbe_alloc_rx_buffers - Replace used receive buffers; packet split
 * @adapter: address of board private structure
 **/
static void ixgbe_alloc_rx_buffers(struct ixgbe_adapter *adapter,
695 696
                                   struct ixgbe_ring *rx_ring,
                                   int cleaned_count)
697 698 699
{
	struct pci_dev *pdev = adapter->pdev;
	union ixgbe_adv_rx_desc *rx_desc;
700
	struct ixgbe_rx_buffer *bi;
701 702 703
	unsigned int i;

	i = rx_ring->next_to_use;
704
	bi = &rx_ring->rx_buffer_info[i];
705 706 707 708

	while (cleaned_count--) {
		rx_desc = IXGBE_RX_DESC_ADV(*rx_ring, i);

709
		if (!bi->page_dma &&
710
		    (rx_ring->flags & IXGBE_RING_RX_PS_ENABLED)) {
711
			if (!bi->page) {
712 713 714 715 716 717 718 719 720
				bi->page = alloc_page(GFP_ATOMIC);
				if (!bi->page) {
					adapter->alloc_rx_page_failed++;
					goto no_buffers;
				}
				bi->page_offset = 0;
			} else {
				/* use a half page if we're re-using */
				bi->page_offset ^= (PAGE_SIZE / 2);
721
			}
722 723 724 725 726

			bi->page_dma = pci_map_page(pdev, bi->page,
			                            bi->page_offset,
			                            (PAGE_SIZE / 2),
			                            PCI_DMA_FROMDEVICE);
727 728
		}

729
		if (!bi->skb) {
730
			struct sk_buff *skb;
J
Jesse Brandeburg 已提交
731 732 733
			/* netdev_alloc_skb reserves 32 bytes up front!! */
			uint bufsz = rx_ring->rx_buf_len + SMP_CACHE_BYTES;
			skb = netdev_alloc_skb(adapter->netdev, bufsz);
734 735 736 737 738 739

			if (!skb) {
				adapter->alloc_rx_buff_failed++;
				goto no_buffers;
			}

J
Jesse Brandeburg 已提交
740 741 742 743
			/* advance the data pointer to the next cache line */
			skb_reserve(skb, (PTR_ALIGN(skb->data, SMP_CACHE_BYTES)
			                  - skb->data));

744
			bi->skb = skb;
J
Jesse Brandeburg 已提交
745 746
			bi->dma = pci_map_single(pdev, skb->data,
			                         rx_ring->rx_buf_len,
747
			                         PCI_DMA_FROMDEVICE);
748 749 750
		}
		/* Refresh the desc even if buffer_addrs didn't change because
		 * each write-back erases this info. */
751
		if (rx_ring->flags & IXGBE_RING_RX_PS_ENABLED) {
752 753
			rx_desc->read.pkt_addr = cpu_to_le64(bi->page_dma);
			rx_desc->read.hdr_addr = cpu_to_le64(bi->dma);
754
		} else {
755
			rx_desc->read.pkt_addr = cpu_to_le64(bi->dma);
756 757 758 759 760
		}

		i++;
		if (i == rx_ring->count)
			i = 0;
761
		bi = &rx_ring->rx_buffer_info[i];
762
	}
763

764 765 766 767 768 769
no_buffers:
	if (rx_ring->next_to_use != i) {
		rx_ring->next_to_use = i;
		if (i-- == 0)
			i = (rx_ring->count - 1);

770
		ixgbe_release_rx_desc(&adapter->hw, rx_ring, i);
771 772 773
	}
}

774 775 776 777 778 779 780 781 782 783
static inline u16 ixgbe_get_hdr_info(union ixgbe_adv_rx_desc *rx_desc)
{
	return rx_desc->wb.lower.lo_dword.hs_rss.hdr_info;
}

static inline u16 ixgbe_get_pkt_info(union ixgbe_adv_rx_desc *rx_desc)
{
	return rx_desc->wb.lower.lo_dword.hs_rss.pkt_info;
}

A
Alexander Duyck 已提交
784 785 786 787 788 789 790 791 792 793
static inline u32 ixgbe_get_rsc_count(union ixgbe_adv_rx_desc *rx_desc)
{
	return (le32_to_cpu(rx_desc->wb.lower.lo_dword.data) &
	        IXGBE_RXDADV_RSCCNT_MASK) >>
	        IXGBE_RXDADV_RSCCNT_SHIFT;
}

/**
 * ixgbe_transform_rsc_queue - change rsc queue into a full packet
 * @skb: pointer to the last skb in the rsc queue
794
 * @count: pointer to number of packets coalesced in this context
A
Alexander Duyck 已提交
795 796 797 798 799
 *
 * This function changes a queue full of hw rsc buffers into a completed
 * packet.  It uses the ->prev pointers to find the first packet and then
 * turns it into the frag list owner.
 **/
800 801
static inline struct sk_buff *ixgbe_transform_rsc_queue(struct sk_buff *skb,
                                                        u64 *count)
A
Alexander Duyck 已提交
802 803 804 805 806 807 808 809
{
	unsigned int frag_list_size = 0;

	while (skb->prev) {
		struct sk_buff *prev = skb->prev;
		frag_list_size += skb->len;
		skb->prev = NULL;
		skb = prev;
810
		*count += 1;
A
Alexander Duyck 已提交
811 812 813 814 815 816 817 818 819 820
	}

	skb_shinfo(skb)->frag_list = skb->next;
	skb->next = NULL;
	skb->len += frag_list_size;
	skb->data_len += frag_list_size;
	skb->truesize += frag_list_size;
	return skb;
}

821 822 823 824 825 826
struct ixgbe_rsc_cb {
	dma_addr_t dma;
};

#define IXGBE_RSC_CB(skb) ((struct ixgbe_rsc_cb *)(skb)->cb)

H
Herbert Xu 已提交
827
static bool ixgbe_clean_rx_irq(struct ixgbe_q_vector *q_vector,
828 829
                               struct ixgbe_ring *rx_ring,
                               int *work_done, int work_to_do)
830
{
H
Herbert Xu 已提交
831
	struct ixgbe_adapter *adapter = q_vector->adapter;
832
	struct net_device *netdev = adapter->netdev;
833 834 835 836
	struct pci_dev *pdev = adapter->pdev;
	union ixgbe_adv_rx_desc *rx_desc, *next_rxd;
	struct ixgbe_rx_buffer *rx_buffer_info, *next_buffer;
	struct sk_buff *skb;
A
Alexander Duyck 已提交
837
	unsigned int i, rsc_count = 0;
838
	u32 len, staterr;
839 840
	u16 hdr_info;
	bool cleaned = false;
841
	int cleaned_count = 0;
842
	unsigned int total_rx_bytes = 0, total_rx_packets = 0;
843 844 845
#ifdef IXGBE_FCOE
	int ddp_bytes = 0;
#endif /* IXGBE_FCOE */
846 847 848 849 850 851 852

	i = rx_ring->next_to_clean;
	rx_desc = IXGBE_RX_DESC_ADV(*rx_ring, i);
	staterr = le32_to_cpu(rx_desc->wb.upper.status_error);
	rx_buffer_info = &rx_ring->rx_buffer_info[i];

	while (staterr & IXGBE_RXD_STAT_DD) {
853
		u32 upper_len = 0;
854 855 856 857
		if (*work_done >= work_to_do)
			break;
		(*work_done)++;

858
		rmb(); /* read descriptor and rx_buffer_info after status DD */
859
		if (rx_ring->flags & IXGBE_RING_RX_PS_ENABLED) {
860 861
			hdr_info = le16_to_cpu(ixgbe_get_hdr_info(rx_desc));
			len = (hdr_info & IXGBE_RXDADV_HDRBUFLEN_MASK) >>
862
			       IXGBE_RXDADV_HDRBUFLEN_SHIFT;
863 864 865
			if (len > IXGBE_RX_HDR_SIZE)
				len = IXGBE_RX_HDR_SIZE;
			upper_len = le16_to_cpu(rx_desc->wb.upper.length);
866
		} else {
867
			len = le16_to_cpu(rx_desc->wb.upper.length);
868
		}
869 870 871

		cleaned = true;
		skb = rx_buffer_info->skb;
J
Jesse Brandeburg 已提交
872
		prefetch(skb->data);
873 874
		rx_buffer_info->skb = NULL;

875
		if (rx_buffer_info->dma) {
876 877 878 879 880 881 882 883 884 885 886 887 888 889 890
			if ((adapter->flags2 & IXGBE_FLAG2_RSC_ENABLED) &&
			    (!(staterr & IXGBE_RXD_STAT_EOP)) &&
				 (!(skb->prev)))
				/*
				 * When HWRSC is enabled, delay unmapping
				 * of the first packet. It carries the
				 * header information, HW may still
				 * access the header after the writeback.
				 * Only unmap it when EOP is reached
				 */
				IXGBE_RSC_CB(skb)->dma = rx_buffer_info->dma;
			else
				pci_unmap_single(pdev, rx_buffer_info->dma,
				                 rx_ring->rx_buf_len,
				                 PCI_DMA_FROMDEVICE);
J
Jesse Brandeburg 已提交
891
			rx_buffer_info->dma = 0;
892 893 894 895 896
			skb_put(skb, len);
		}

		if (upper_len) {
			pci_unmap_page(pdev, rx_buffer_info->page_dma,
897
			               PAGE_SIZE / 2, PCI_DMA_FROMDEVICE);
898 899
			rx_buffer_info->page_dma = 0;
			skb_fill_page_desc(skb, skb_shinfo(skb)->nr_frags,
900 901 902 903 904 905 906 907 908
			                   rx_buffer_info->page,
			                   rx_buffer_info->page_offset,
			                   upper_len);

			if ((rx_ring->rx_buf_len > (PAGE_SIZE / 2)) ||
			    (page_count(rx_buffer_info->page) != 1))
				rx_buffer_info->page = NULL;
			else
				get_page(rx_buffer_info->page);
909 910 911 912 913 914 915 916 917 918 919 920 921

			skb->len += upper_len;
			skb->data_len += upper_len;
			skb->truesize += upper_len;
		}

		i++;
		if (i == rx_ring->count)
			i = 0;

		next_rxd = IXGBE_RX_DESC_ADV(*rx_ring, i);
		prefetch(next_rxd);
		cleaned_count++;
A
Alexander Duyck 已提交
922

923
		if (adapter->flags2 & IXGBE_FLAG2_RSC_CAPABLE)
A
Alexander Duyck 已提交
924 925 926 927 928 929 930 931 932 933
			rsc_count = ixgbe_get_rsc_count(rx_desc);

		if (rsc_count) {
			u32 nextp = (staterr & IXGBE_RXDADV_NEXTP_MASK) >>
				     IXGBE_RXDADV_NEXTP_SHIFT;
			next_buffer = &rx_ring->rx_buffer_info[nextp];
		} else {
			next_buffer = &rx_ring->rx_buffer_info[i];
		}

934
		if (staterr & IXGBE_RXD_STAT_EOP) {
A
Alexander Duyck 已提交
935
			if (skb->prev)
936 937
				skb = ixgbe_transform_rsc_queue(skb, &(rx_ring->rsc_count));
			if (adapter->flags2 & IXGBE_FLAG2_RSC_ENABLED) {
938 939 940 941
				if (IXGBE_RSC_CB(skb)->dma)
					pci_unmap_single(pdev, IXGBE_RSC_CB(skb)->dma,
					                 rx_ring->rx_buf_len,
					                 PCI_DMA_FROMDEVICE);
942 943 944 945 946 947
				if (rx_ring->flags & IXGBE_RING_RX_PS_ENABLED)
					rx_ring->rsc_count += skb_shinfo(skb)->nr_frags;
				else
					rx_ring->rsc_count++;
				rx_ring->rsc_flush++;
			}
948 949 950
			rx_ring->stats.packets++;
			rx_ring->stats.bytes += skb->len;
		} else {
951
			if (rx_ring->flags & IXGBE_RING_RX_PS_ENABLED) {
A
Alexander Duyck 已提交
952 953 954 955 956 957 958 959
				rx_buffer_info->skb = next_buffer->skb;
				rx_buffer_info->dma = next_buffer->dma;
				next_buffer->skb = skb;
				next_buffer->dma = 0;
			} else {
				skb->next = next_buffer->skb;
				skb->next->prev = skb;
			}
J
Jesse Brandeburg 已提交
960
			rx_ring->non_eop_descs++;
961 962 963 964 965 966 967 968
			goto next_desc;
		}

		if (staterr & IXGBE_RXDADV_ERR_FRAME_ERR_MASK) {
			dev_kfree_skb_irq(skb);
			goto next_desc;
		}

969
		ixgbe_rx_checksum(adapter, rx_desc, skb);
970 971 972 973 974

		/* probably a little skewed due to removing CRC */
		total_rx_bytes += skb->len;
		total_rx_packets++;

975
		skb->protocol = eth_type_trans(skb, adapter->netdev);
976 977
#ifdef IXGBE_FCOE
		/* if ddp, not passing to ULD unless for FCP_RSP or error */
978 979 980
		if (adapter->flags & IXGBE_FLAG_FCOE_ENABLED) {
			ddp_bytes = ixgbe_fcoe_ddp(adapter, rx_desc, skb);
			if (!ddp_bytes)
981
				goto next_desc;
982
		}
983
#endif /* IXGBE_FCOE */
984
		ixgbe_receive_skb(q_vector, skb, staterr, rx_ring, rx_desc);
985 986 987 988 989 990 991 992 993 994 995 996

next_desc:
		rx_desc->wb.upper.status_error = 0;

		/* return some buffers to hardware, one at a time is too slow */
		if (cleaned_count >= IXGBE_RX_BUFFER_WRITE) {
			ixgbe_alloc_rx_buffers(adapter, rx_ring, cleaned_count);
			cleaned_count = 0;
		}

		/* use prefetched values */
		rx_desc = next_rxd;
A
Alexander Duyck 已提交
997
		rx_buffer_info = &rx_ring->rx_buffer_info[i];
998 999

		staterr = le32_to_cpu(rx_desc->wb.upper.status_error);
1000 1001
	}

1002 1003 1004 1005 1006 1007
	rx_ring->next_to_clean = i;
	cleaned_count = IXGBE_DESC_UNUSED(rx_ring);

	if (cleaned_count)
		ixgbe_alloc_rx_buffers(adapter, rx_ring, cleaned_count);

1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022
#ifdef IXGBE_FCOE
	/* include DDPed FCoE data */
	if (ddp_bytes > 0) {
		unsigned int mss;

		mss = adapter->netdev->mtu - sizeof(struct fcoe_hdr) -
			sizeof(struct fc_frame_header) -
			sizeof(struct fcoe_crc_eof);
		if (mss > 512)
			mss &= ~511;
		total_rx_bytes += ddp_bytes;
		total_rx_packets += DIV_ROUND_UP(ddp_bytes, mss);
	}
#endif /* IXGBE_FCOE */

1023 1024
	rx_ring->total_packets += total_rx_packets;
	rx_ring->total_bytes += total_rx_bytes;
1025 1026
	netdev->stats.rx_bytes += total_rx_bytes;
	netdev->stats.rx_packets += total_rx_packets;
1027

1028 1029 1030
	return cleaned;
}

1031
static int ixgbe_clean_rxonly(struct napi_struct *, int);
1032 1033 1034 1035 1036 1037 1038 1039 1040
/**
 * ixgbe_configure_msix - Configure MSI-X hardware
 * @adapter: board private structure
 *
 * ixgbe_configure_msix sets up the hardware to properly generate MSI-X
 * interrupts.
 **/
static void ixgbe_configure_msix(struct ixgbe_adapter *adapter)
{
1041 1042 1043
	struct ixgbe_q_vector *q_vector;
	int i, j, q_vectors, v_idx, r_idx;
	u32 mask;
1044

1045
	q_vectors = adapter->num_msix_vectors - NON_Q_VECTORS;
1046

1047 1048
	/*
	 * Populate the IVAR table and set the ITR values to the
1049 1050 1051
	 * corresponding register.
	 */
	for (v_idx = 0; v_idx < q_vectors; v_idx++) {
1052
		q_vector = adapter->q_vector[v_idx];
1053
		/* XXX for_each_set_bit(...) */
1054
		r_idx = find_first_bit(q_vector->rxr_idx,
1055
		                       adapter->num_rx_queues);
1056 1057

		for (i = 0; i < q_vector->rxr_count; i++) {
1058
			j = adapter->rx_ring[r_idx]->reg_idx;
1059
			ixgbe_set_ivar(adapter, 0, j, v_idx);
1060
			r_idx = find_next_bit(q_vector->rxr_idx,
1061 1062
			                      adapter->num_rx_queues,
			                      r_idx + 1);
1063 1064
		}
		r_idx = find_first_bit(q_vector->txr_idx,
1065
		                       adapter->num_tx_queues);
1066 1067

		for (i = 0; i < q_vector->txr_count; i++) {
1068
			j = adapter->tx_ring[r_idx]->reg_idx;
1069
			ixgbe_set_ivar(adapter, 1, j, v_idx);
1070
			r_idx = find_next_bit(q_vector->txr_idx,
1071 1072
			                      adapter->num_tx_queues,
			                      r_idx + 1);
1073 1074 1075
		}

		if (q_vector->txr_count && !q_vector->rxr_count)
1076 1077
			/* tx only */
			q_vector->eitr = adapter->tx_eitr_param;
1078
		else if (q_vector->rxr_count)
1079 1080
			/* rx or mixed */
			q_vector->eitr = adapter->rx_eitr_param;
1081

1082
		ixgbe_write_eitr(q_vector);
1083 1084
	}

1085 1086 1087 1088 1089
	if (adapter->hw.mac.type == ixgbe_mac_82598EB)
		ixgbe_set_ivar(adapter, -1, IXGBE_IVAR_OTHER_CAUSES_INDEX,
		               v_idx);
	else if (adapter->hw.mac.type == ixgbe_mac_82599EB)
		ixgbe_set_ivar(adapter, -1, 1, v_idx);
1090 1091
	IXGBE_WRITE_REG(&adapter->hw, IXGBE_EITR(v_idx), 1950);

1092
	/* set up to autoclear timer, and the vectors */
1093
	mask = IXGBE_EIMS_ENABLE_MASK;
1094 1095 1096 1097 1098 1099
	if (adapter->num_vfs)
		mask &= ~(IXGBE_EIMS_OTHER |
			  IXGBE_EIMS_MAILBOX |
			  IXGBE_EIMS_LSC);
	else
		mask &= ~(IXGBE_EIMS_OTHER | IXGBE_EIMS_LSC);
1100
	IXGBE_WRITE_REG(&adapter->hw, IXGBE_EIAC, mask);
1101 1102
}

1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128
enum latency_range {
	lowest_latency = 0,
	low_latency = 1,
	bulk_latency = 2,
	latency_invalid = 255
};

/**
 * ixgbe_update_itr - update the dynamic ITR value based on statistics
 * @adapter: pointer to adapter
 * @eitr: eitr setting (ints per sec) to give last timeslice
 * @itr_setting: current throttle rate in ints/second
 * @packets: the number of packets during this measurement interval
 * @bytes: the number of bytes during this measurement interval
 *
 *      Stores a new ITR value based on packets and byte
 *      counts during the last interrupt.  The advantage of per interrupt
 *      computation is faster updates and more accurate ITR for the current
 *      traffic pattern.  Constants in this function were computed
 *      based on theoretical maximum wire speed and thresholds were set based
 *      on testing data as well as attempting to minimize response time
 *      while increasing bulk throughput.
 *      this functionality is controlled by the InterruptThrottleRate module
 *      parameter (see ixgbe_param.c)
 **/
static u8 ixgbe_update_itr(struct ixgbe_adapter *adapter,
1129 1130
                           u32 eitr, u8 itr_setting,
                           int packets, int bytes)
1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169
{
	unsigned int retval = itr_setting;
	u32 timepassed_us;
	u64 bytes_perint;

	if (packets == 0)
		goto update_itr_done;


	/* simple throttlerate management
	 *    0-20MB/s lowest (100000 ints/s)
	 *   20-100MB/s low   (20000 ints/s)
	 *  100-1249MB/s bulk (8000 ints/s)
	 */
	/* what was last interrupt timeslice? */
	timepassed_us = 1000000/eitr;
	bytes_perint = bytes / timepassed_us; /* bytes/usec */

	switch (itr_setting) {
	case lowest_latency:
		if (bytes_perint > adapter->eitr_low)
			retval = low_latency;
		break;
	case low_latency:
		if (bytes_perint > adapter->eitr_high)
			retval = bulk_latency;
		else if (bytes_perint <= adapter->eitr_low)
			retval = lowest_latency;
		break;
	case bulk_latency:
		if (bytes_perint <= adapter->eitr_high)
			retval = low_latency;
		break;
	}

update_itr_done:
	return retval;
}

1170 1171
/**
 * ixgbe_write_eitr - write EITR register in hardware specific way
1172
 * @q_vector: structure containing interrupt and ring information
1173 1174 1175 1176 1177
 *
 * This function is made to be called by ethtool and by the driver
 * when it needs to update EITR registers at runtime.  Hardware
 * specific quirks/differences are taken care of here.
 */
1178
void ixgbe_write_eitr(struct ixgbe_q_vector *q_vector)
1179
{
1180
	struct ixgbe_adapter *adapter = q_vector->adapter;
1181
	struct ixgbe_hw *hw = &adapter->hw;
1182 1183 1184
	int v_idx = q_vector->v_idx;
	u32 itr_reg = EITR_INTS_PER_SEC_TO_REG(q_vector->eitr);

1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197
	if (adapter->hw.mac.type == ixgbe_mac_82598EB) {
		/* must write high and low 16 bits to reset counter */
		itr_reg |= (itr_reg << 16);
	} else if (adapter->hw.mac.type == ixgbe_mac_82599EB) {
		/*
		 * set the WDIS bit to not clear the timer bits and cause an
		 * immediate assertion of the interrupt
		 */
		itr_reg |= IXGBE_EITR_CNT_WDIS;
	}
	IXGBE_WRITE_REG(hw, IXGBE_EITR(v_idx), itr_reg);
}

1198 1199 1200 1201 1202
static void ixgbe_set_itr_msix(struct ixgbe_q_vector *q_vector)
{
	struct ixgbe_adapter *adapter = q_vector->adapter;
	u32 new_itr;
	u8 current_itr, ret_itr;
1203
	int i, r_idx;
1204 1205 1206 1207
	struct ixgbe_ring *rx_ring, *tx_ring;

	r_idx = find_first_bit(q_vector->txr_idx, adapter->num_tx_queues);
	for (i = 0; i < q_vector->txr_count; i++) {
1208
		tx_ring = adapter->tx_ring[r_idx];
1209
		ret_itr = ixgbe_update_itr(adapter, q_vector->eitr,
1210 1211 1212
		                           q_vector->tx_itr,
		                           tx_ring->total_packets,
		                           tx_ring->total_bytes);
1213 1214
		/* if the result for this queue would decrease interrupt
		 * rate for this vector then use that result */
1215
		q_vector->tx_itr = ((q_vector->tx_itr > ret_itr) ?
1216
		                    q_vector->tx_itr - 1 : ret_itr);
1217
		r_idx = find_next_bit(q_vector->txr_idx, adapter->num_tx_queues,
1218
		                      r_idx + 1);
1219 1220 1221 1222
	}

	r_idx = find_first_bit(q_vector->rxr_idx, adapter->num_rx_queues);
	for (i = 0; i < q_vector->rxr_count; i++) {
1223
		rx_ring = adapter->rx_ring[r_idx];
1224
		ret_itr = ixgbe_update_itr(adapter, q_vector->eitr,
1225 1226 1227
		                           q_vector->rx_itr,
		                           rx_ring->total_packets,
		                           rx_ring->total_bytes);
1228 1229
		/* if the result for this queue would decrease interrupt
		 * rate for this vector then use that result */
1230
		q_vector->rx_itr = ((q_vector->rx_itr > ret_itr) ?
1231
		                    q_vector->rx_itr - 1 : ret_itr);
1232
		r_idx = find_next_bit(q_vector->rxr_idx, adapter->num_rx_queues,
1233
		                      r_idx + 1);
1234 1235
	}

1236
	current_itr = max(q_vector->rx_itr, q_vector->tx_itr);
1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252

	switch (current_itr) {
	/* counts and packets in update_itr are dependent on these numbers */
	case lowest_latency:
		new_itr = 100000;
		break;
	case low_latency:
		new_itr = 20000; /* aka hwitr = ~200 */
		break;
	case bulk_latency:
	default:
		new_itr = 8000;
		break;
	}

	if (new_itr != q_vector->eitr) {
1253 1254
		/* do an exponential smoothing */
		new_itr = ((q_vector->eitr * 90)/100) + ((new_itr * 10)/100);
1255 1256 1257

		/* save the algorithm value here, not the smoothed one */
		q_vector->eitr = new_itr;
1258 1259

		ixgbe_write_eitr(q_vector);
1260 1261 1262 1263 1264
	}

	return;
}

1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275
static void ixgbe_check_fan_failure(struct ixgbe_adapter *adapter, u32 eicr)
{
	struct ixgbe_hw *hw = &adapter->hw;

	if ((adapter->flags & IXGBE_FLAG_FAN_FAIL_CAPABLE) &&
	    (eicr & IXGBE_EICR_GPI_SDP1)) {
		DPRINTK(PROBE, CRIT, "Fan has stopped, replace the adapter\n");
		/* write to clear the interrupt */
		IXGBE_WRITE_REG(hw, IXGBE_EICR, IXGBE_EICR_GPI_SDP1);
	}
}
1276

1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294
static void ixgbe_check_sfp_event(struct ixgbe_adapter *adapter, u32 eicr)
{
	struct ixgbe_hw *hw = &adapter->hw;

	if (eicr & IXGBE_EICR_GPI_SDP1) {
		/* Clear the interrupt */
		IXGBE_WRITE_REG(hw, IXGBE_EICR, IXGBE_EICR_GPI_SDP1);
		schedule_work(&adapter->multispeed_fiber_task);
	} else if (eicr & IXGBE_EICR_GPI_SDP2) {
		/* Clear the interrupt */
		IXGBE_WRITE_REG(hw, IXGBE_EICR, IXGBE_EICR_GPI_SDP2);
		schedule_work(&adapter->sfp_config_module_task);
	} else {
		/* Interrupt isn't for us... */
		return;
	}
}

1295 1296 1297 1298 1299 1300 1301 1302 1303
static void ixgbe_check_lsc(struct ixgbe_adapter *adapter)
{
	struct ixgbe_hw *hw = &adapter->hw;

	adapter->lsc_int++;
	adapter->flags |= IXGBE_FLAG_NEED_LINK_UPDATE;
	adapter->link_check_timeout = jiffies;
	if (!test_bit(__IXGBE_DOWN, &adapter->state)) {
		IXGBE_WRITE_REG(hw, IXGBE_EIMC, IXGBE_EIMC_LSC);
1304
		IXGBE_WRITE_FLUSH(hw);
1305 1306 1307 1308
		schedule_work(&adapter->watchdog_task);
	}
}

1309 1310 1311 1312 1313
static irqreturn_t ixgbe_msix_lsc(int irq, void *data)
{
	struct net_device *netdev = data;
	struct ixgbe_adapter *adapter = netdev_priv(netdev);
	struct ixgbe_hw *hw = &adapter->hw;
1314 1315 1316 1317 1318 1319 1320 1321 1322 1323
	u32 eicr;

	/*
	 * Workaround for Silicon errata.  Use clear-by-write instead
	 * of clear-by-read.  Reading with EICS will return the
	 * interrupt causes without clearing, which later be done
	 * with the write to EICR.
	 */
	eicr = IXGBE_READ_REG(hw, IXGBE_EICS);
	IXGBE_WRITE_REG(hw, IXGBE_EICR, eicr);
1324

1325 1326
	if (eicr & IXGBE_EICR_LSC)
		ixgbe_check_lsc(adapter);
1327

1328 1329 1330
	if (eicr & IXGBE_EICR_MAILBOX)
		ixgbe_msg_task(adapter);

1331 1332
	if (hw->mac.type == ixgbe_mac_82598EB)
		ixgbe_check_fan_failure(adapter, eicr);
1333

1334
	if (hw->mac.type == ixgbe_mac_82599EB) {
1335
		ixgbe_check_sfp_event(adapter, eicr);
1336 1337 1338 1339 1340 1341 1342 1343 1344

		/* Handle Flow Director Full threshold interrupt */
		if (eicr & IXGBE_EICR_FLOW_DIR) {
			int i;
			IXGBE_WRITE_REG(hw, IXGBE_EICR, IXGBE_EICR_FLOW_DIR);
			/* Disable transmits before FDIR Re-initialization */
			netif_tx_stop_all_queues(netdev);
			for (i = 0; i < adapter->num_tx_queues; i++) {
				struct ixgbe_ring *tx_ring =
1345
				                            adapter->tx_ring[i];
1346 1347 1348 1349 1350 1351
				if (test_and_clear_bit(__IXGBE_FDIR_INIT_DONE,
				                       &tx_ring->reinit_state))
					schedule_work(&adapter->fdir_reinit_task);
			}
		}
	}
1352 1353
	if (!test_bit(__IXGBE_DOWN, &adapter->state))
		IXGBE_WRITE_REG(hw, IXGBE_EIMS, IXGBE_EIMS_OTHER);
1354 1355 1356 1357

	return IRQ_HANDLED;
}

1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391
static inline void ixgbe_irq_enable_queues(struct ixgbe_adapter *adapter,
					   u64 qmask)
{
	u32 mask;

	if (adapter->hw.mac.type == ixgbe_mac_82598EB) {
		mask = (IXGBE_EIMS_RTX_QUEUE & qmask);
		IXGBE_WRITE_REG(&adapter->hw, IXGBE_EIMS, mask);
	} else {
		mask = (qmask & 0xFFFFFFFF);
		IXGBE_WRITE_REG(&adapter->hw, IXGBE_EIMS_EX(0), mask);
		mask = (qmask >> 32);
		IXGBE_WRITE_REG(&adapter->hw, IXGBE_EIMS_EX(1), mask);
	}
	/* skip the flush */
}

static inline void ixgbe_irq_disable_queues(struct ixgbe_adapter *adapter,
                                            u64 qmask)
{
	u32 mask;

	if (adapter->hw.mac.type == ixgbe_mac_82598EB) {
		mask = (IXGBE_EIMS_RTX_QUEUE & qmask);
		IXGBE_WRITE_REG(&adapter->hw, IXGBE_EIMC, mask);
	} else {
		mask = (qmask & 0xFFFFFFFF);
		IXGBE_WRITE_REG(&adapter->hw, IXGBE_EIMC_EX(0), mask);
		mask = (qmask >> 32);
		IXGBE_WRITE_REG(&adapter->hw, IXGBE_EIMC_EX(1), mask);
	}
	/* skip the flush */
}

1392 1393
static irqreturn_t ixgbe_msix_clean_tx(int irq, void *data)
{
1394 1395
	struct ixgbe_q_vector *q_vector = data;
	struct ixgbe_adapter  *adapter = q_vector->adapter;
1396
	struct ixgbe_ring     *tx_ring;
1397 1398 1399 1400 1401 1402 1403
	int i, r_idx;

	if (!q_vector->txr_count)
		return IRQ_HANDLED;

	r_idx = find_first_bit(q_vector->txr_idx, adapter->num_tx_queues);
	for (i = 0; i < q_vector->txr_count; i++) {
1404
		tx_ring = adapter->tx_ring[r_idx];
1405 1406
		tx_ring->total_bytes = 0;
		tx_ring->total_packets = 0;
1407
		r_idx = find_next_bit(q_vector->txr_idx, adapter->num_tx_queues,
1408
		                      r_idx + 1);
1409
	}
1410

1411
	/* EIAM disabled interrupts (on this vector) for us */
1412 1413
	napi_schedule(&q_vector->napi);

1414 1415 1416
	return IRQ_HANDLED;
}

1417 1418 1419 1420 1421
/**
 * ixgbe_msix_clean_rx - single unshared vector rx clean (all queues)
 * @irq: unused
 * @data: pointer to our q_vector struct for this interrupt vector
 **/
1422 1423
static irqreturn_t ixgbe_msix_clean_rx(int irq, void *data)
{
1424 1425
	struct ixgbe_q_vector *q_vector = data;
	struct ixgbe_adapter  *adapter = q_vector->adapter;
1426
	struct ixgbe_ring  *rx_ring;
1427
	int r_idx;
1428
	int i;
1429 1430

	r_idx = find_first_bit(q_vector->rxr_idx, adapter->num_rx_queues);
1431
	for (i = 0;  i < q_vector->rxr_count; i++) {
1432
		rx_ring = adapter->rx_ring[r_idx];
1433 1434 1435 1436 1437 1438
		rx_ring->total_bytes = 0;
		rx_ring->total_packets = 0;
		r_idx = find_next_bit(q_vector->rxr_idx, adapter->num_rx_queues,
		                      r_idx + 1);
	}

1439 1440 1441 1442
	if (!q_vector->rxr_count)
		return IRQ_HANDLED;

	/* disable interrupts on this vector only */
1443
	/* EIAM disabled interrupts (on this vector) for us */
1444
	napi_schedule(&q_vector->napi);
1445 1446 1447 1448 1449 1450

	return IRQ_HANDLED;
}

static irqreturn_t ixgbe_msix_clean_many(int irq, void *data)
{
1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461
	struct ixgbe_q_vector *q_vector = data;
	struct ixgbe_adapter  *adapter = q_vector->adapter;
	struct ixgbe_ring  *ring;
	int r_idx;
	int i;

	if (!q_vector->txr_count && !q_vector->rxr_count)
		return IRQ_HANDLED;

	r_idx = find_first_bit(q_vector->txr_idx, adapter->num_tx_queues);
	for (i = 0; i < q_vector->txr_count; i++) {
1462
		ring = adapter->tx_ring[r_idx];
1463 1464 1465 1466 1467 1468 1469 1470
		ring->total_bytes = 0;
		ring->total_packets = 0;
		r_idx = find_next_bit(q_vector->txr_idx, adapter->num_tx_queues,
		                      r_idx + 1);
	}

	r_idx = find_first_bit(q_vector->rxr_idx, adapter->num_rx_queues);
	for (i = 0; i < q_vector->rxr_count; i++) {
1471
		ring = adapter->rx_ring[r_idx];
1472 1473 1474 1475 1476 1477
		ring->total_bytes = 0;
		ring->total_packets = 0;
		r_idx = find_next_bit(q_vector->rxr_idx, adapter->num_rx_queues,
		                      r_idx + 1);
	}

1478
	/* EIAM disabled interrupts (on this vector) for us */
1479
	napi_schedule(&q_vector->napi);
1480 1481 1482 1483

	return IRQ_HANDLED;
}

1484 1485 1486 1487 1488
/**
 * ixgbe_clean_rxonly - msix (aka one shot) rx clean routine
 * @napi: napi struct with our devices info in it
 * @budget: amount of work driver is allowed to do this pass, in packets
 *
1489 1490
 * This function is optimized for cleaning one queue only on a single
 * q_vector!!!
1491
 **/
1492 1493
static int ixgbe_clean_rxonly(struct napi_struct *napi, int budget)
{
1494
	struct ixgbe_q_vector *q_vector =
1495
	                       container_of(napi, struct ixgbe_q_vector, napi);
1496
	struct ixgbe_adapter *adapter = q_vector->adapter;
1497
	struct ixgbe_ring *rx_ring = NULL;
1498
	int work_done = 0;
1499
	long r_idx;
1500

1501
	r_idx = find_first_bit(q_vector->rxr_idx, adapter->num_rx_queues);
1502
	rx_ring = adapter->rx_ring[r_idx];
1503
#ifdef CONFIG_IXGBE_DCA
1504
	if (adapter->flags & IXGBE_FLAG_DCA_ENABLED)
1505
		ixgbe_update_rx_dca(adapter, rx_ring);
1506
#endif
1507

H
Herbert Xu 已提交
1508
	ixgbe_clean_rx_irq(q_vector, rx_ring, &work_done, budget);
1509

1510 1511
	/* If all Rx work done, exit the polling mode */
	if (work_done < budget) {
1512
		napi_complete(napi);
1513
		if (adapter->rx_itr_setting & 1)
1514
			ixgbe_set_itr_msix(q_vector);
1515
		if (!test_bit(__IXGBE_DOWN, &adapter->state))
1516 1517
			ixgbe_irq_enable_queues(adapter,
			                        ((u64)1 << q_vector->v_idx));
1518 1519 1520 1521 1522
	}

	return work_done;
}

1523
/**
1524
 * ixgbe_clean_rxtx_many - msix (aka one shot) rx clean routine
1525 1526 1527 1528 1529 1530
 * @napi: napi struct with our devices info in it
 * @budget: amount of work driver is allowed to do this pass, in packets
 *
 * This function will clean more than one rx queue associated with a
 * q_vector.
 **/
1531
static int ixgbe_clean_rxtx_many(struct napi_struct *napi, int budget)
1532 1533 1534 1535
{
	struct ixgbe_q_vector *q_vector =
	                       container_of(napi, struct ixgbe_q_vector, napi);
	struct ixgbe_adapter *adapter = q_vector->adapter;
1536
	struct ixgbe_ring *ring = NULL;
1537 1538
	int work_done = 0, i;
	long r_idx;
1539 1540 1541 1542
	bool tx_clean_complete = true;

	r_idx = find_first_bit(q_vector->txr_idx, adapter->num_tx_queues);
	for (i = 0; i < q_vector->txr_count; i++) {
1543
		ring = adapter->tx_ring[r_idx];
1544 1545 1546 1547 1548 1549 1550 1551
#ifdef CONFIG_IXGBE_DCA
		if (adapter->flags & IXGBE_FLAG_DCA_ENABLED)
			ixgbe_update_tx_dca(adapter, ring);
#endif
		tx_clean_complete &= ixgbe_clean_tx_irq(q_vector, ring);
		r_idx = find_next_bit(q_vector->txr_idx, adapter->num_tx_queues,
		                      r_idx + 1);
	}
1552 1553 1554 1555 1556 1557 1558

	/* attempt to distribute budget to each queue fairly, but don't allow
	 * the budget to go below 1 because we'll exit polling */
	budget /= (q_vector->rxr_count ?: 1);
	budget = max(budget, 1);
	r_idx = find_first_bit(q_vector->rxr_idx, adapter->num_rx_queues);
	for (i = 0; i < q_vector->rxr_count; i++) {
1559
		ring = adapter->rx_ring[r_idx];
1560
#ifdef CONFIG_IXGBE_DCA
1561
		if (adapter->flags & IXGBE_FLAG_DCA_ENABLED)
1562
			ixgbe_update_rx_dca(adapter, ring);
1563
#endif
1564
		ixgbe_clean_rx_irq(q_vector, ring, &work_done, budget);
1565 1566 1567 1568 1569
		r_idx = find_next_bit(q_vector->rxr_idx, adapter->num_rx_queues,
		                      r_idx + 1);
	}

	r_idx = find_first_bit(q_vector->rxr_idx, adapter->num_rx_queues);
1570
	ring = adapter->rx_ring[r_idx];
1571
	/* If all Rx work done, exit the polling mode */
1572
	if (work_done < budget) {
1573
		napi_complete(napi);
1574
		if (adapter->rx_itr_setting & 1)
1575 1576
			ixgbe_set_itr_msix(q_vector);
		if (!test_bit(__IXGBE_DOWN, &adapter->state))
1577 1578
			ixgbe_irq_enable_queues(adapter,
			                        ((u64)1 << q_vector->v_idx));
1579 1580 1581 1582 1583
		return 0;
	}

	return work_done;
}
1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602

/**
 * ixgbe_clean_txonly - msix (aka one shot) tx clean routine
 * @napi: napi struct with our devices info in it
 * @budget: amount of work driver is allowed to do this pass, in packets
 *
 * This function is optimized for cleaning one queue only on a single
 * q_vector!!!
 **/
static int ixgbe_clean_txonly(struct napi_struct *napi, int budget)
{
	struct ixgbe_q_vector *q_vector =
	                       container_of(napi, struct ixgbe_q_vector, napi);
	struct ixgbe_adapter *adapter = q_vector->adapter;
	struct ixgbe_ring *tx_ring = NULL;
	int work_done = 0;
	long r_idx;

	r_idx = find_first_bit(q_vector->txr_idx, adapter->num_tx_queues);
1603
	tx_ring = adapter->tx_ring[r_idx];
1604 1605 1606 1607 1608 1609 1610 1611
#ifdef CONFIG_IXGBE_DCA
	if (adapter->flags & IXGBE_FLAG_DCA_ENABLED)
		ixgbe_update_tx_dca(adapter, tx_ring);
#endif

	if (!ixgbe_clean_tx_irq(q_vector, tx_ring))
		work_done = budget;

1612
	/* If all Tx work done, exit the polling mode */
1613 1614
	if (work_done < budget) {
		napi_complete(napi);
1615
		if (adapter->tx_itr_setting & 1)
1616 1617 1618 1619 1620 1621 1622 1623
			ixgbe_set_itr_msix(q_vector);
		if (!test_bit(__IXGBE_DOWN, &adapter->state))
			ixgbe_irq_enable_queues(adapter, ((u64)1 << q_vector->v_idx));
	}

	return work_done;
}

1624
static inline void map_vector_to_rxq(struct ixgbe_adapter *a, int v_idx,
1625
                                     int r_idx)
1626
{
1627 1628 1629 1630
	struct ixgbe_q_vector *q_vector = a->q_vector[v_idx];

	set_bit(r_idx, q_vector->rxr_idx);
	q_vector->rxr_count++;
1631 1632 1633
}

static inline void map_vector_to_txq(struct ixgbe_adapter *a, int v_idx,
1634
                                     int t_idx)
1635
{
1636 1637 1638 1639
	struct ixgbe_q_vector *q_vector = a->q_vector[v_idx];

	set_bit(t_idx, q_vector->txr_idx);
	q_vector->txr_count++;
1640 1641
}

1642
/**
1643 1644 1645
 * ixgbe_map_rings_to_vectors - Maps descriptor rings to vectors
 * @adapter: board private structure to initialize
 * @vectors: allotted vector count for descriptor rings
1646
 *
1647 1648 1649 1650 1651
 * This function maps descriptor rings to the queue-specific vectors
 * we were allotted through the MSI-X enabling code.  Ideally, we'd have
 * one vector per ring/queue, but on a constrained vector budget, we
 * group the rings as "efficiently" as possible.  You would add new
 * mapping configurations in here.
1652
 **/
1653
static int ixgbe_map_rings_to_vectors(struct ixgbe_adapter *adapter,
1654
                                      int vectors)
1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666
{
	int v_start = 0;
	int rxr_idx = 0, txr_idx = 0;
	int rxr_remaining = adapter->num_rx_queues;
	int txr_remaining = adapter->num_tx_queues;
	int i, j;
	int rqpv, tqpv;
	int err = 0;

	/* No mapping required if MSI-X is disabled. */
	if (!(adapter->flags & IXGBE_FLAG_MSIX_ENABLED))
		goto out;
1667

1668 1669 1670 1671 1672 1673 1674
	/*
	 * The ideal configuration...
	 * We have enough vectors to map one per queue.
	 */
	if (vectors == adapter->num_rx_queues + adapter->num_tx_queues) {
		for (; rxr_idx < rxr_remaining; v_start++, rxr_idx++)
			map_vector_to_rxq(adapter, v_start, rxr_idx);
1675

1676 1677
		for (; txr_idx < txr_remaining; v_start++, txr_idx++)
			map_vector_to_txq(adapter, v_start, txr_idx);
1678 1679

		goto out;
1680
	}
1681

1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701
	/*
	 * If we don't have enough vectors for a 1-to-1
	 * mapping, we'll have to group them so there are
	 * multiple queues per vector.
	 */
	/* Re-adjusting *qpv takes care of the remainder. */
	for (i = v_start; i < vectors; i++) {
		rqpv = DIV_ROUND_UP(rxr_remaining, vectors - i);
		for (j = 0; j < rqpv; j++) {
			map_vector_to_rxq(adapter, i, rxr_idx);
			rxr_idx++;
			rxr_remaining--;
		}
	}
	for (i = v_start; i < vectors; i++) {
		tqpv = DIV_ROUND_UP(txr_remaining, vectors - i);
		for (j = 0; j < tqpv; j++) {
			map_vector_to_txq(adapter, i, txr_idx);
			txr_idx++;
			txr_remaining--;
1702 1703 1704
		}
	}

1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720
out:
	return err;
}

/**
 * ixgbe_request_msix_irqs - Initialize MSI-X interrupts
 * @adapter: board private structure
 *
 * ixgbe_request_msix_irqs allocates MSI-X vectors and requests
 * interrupts from the kernel.
 **/
static int ixgbe_request_msix_irqs(struct ixgbe_adapter *adapter)
{
	struct net_device *netdev = adapter->netdev;
	irqreturn_t (*handler)(int, void *);
	int i, vector, q_vectors, err;
R
Robert Olsson 已提交
1721
	int ri=0, ti=0;
1722 1723 1724 1725 1726 1727 1728 1729 1730 1731

	/* Decrement for Other and TCP Timer vectors */
	q_vectors = adapter->num_msix_vectors - NON_Q_VECTORS;

	/* Map the Tx/Rx rings to the vectors we were allotted. */
	err = ixgbe_map_rings_to_vectors(adapter, q_vectors);
	if (err)
		goto out;

#define SET_HANDLER(_v) ((!(_v)->rxr_count) ? &ixgbe_msix_clean_tx : \
1732 1733
                         (!(_v)->txr_count) ? &ixgbe_msix_clean_rx : \
                         &ixgbe_msix_clean_many)
1734
	for (vector = 0; vector < q_vectors; vector++) {
1735
		handler = SET_HANDLER(adapter->q_vector[vector]);
R
Robert Olsson 已提交
1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748

		if(handler == &ixgbe_msix_clean_rx) {
			sprintf(adapter->name[vector], "%s-%s-%d",
				netdev->name, "rx", ri++);
		}
		else if(handler == &ixgbe_msix_clean_tx) {
			sprintf(adapter->name[vector], "%s-%s-%d",
				netdev->name, "tx", ti++);
		}
		else
			sprintf(adapter->name[vector], "%s-%s-%d",
				netdev->name, "TxRx", vector);

1749
		err = request_irq(adapter->msix_entries[vector].vector,
1750
		                  handler, 0, adapter->name[vector],
1751
		                  adapter->q_vector[vector]);
1752 1753
		if (err) {
			DPRINTK(PROBE, ERR,
1754 1755
			        "request_irq failed for MSIX interrupt "
			        "Error: %d\n", err);
1756
			goto free_queue_irqs;
1757 1758 1759
		}
	}

1760 1761
	sprintf(adapter->name[vector], "%s:lsc", netdev->name);
	err = request_irq(adapter->msix_entries[vector].vector,
1762
	                  ixgbe_msix_lsc, 0, adapter->name[vector], netdev);
1763 1764 1765
	if (err) {
		DPRINTK(PROBE, ERR,
			"request_irq for msix_lsc failed: %d\n", err);
1766
		goto free_queue_irqs;
1767 1768 1769 1770
	}

	return 0;

1771 1772 1773
free_queue_irqs:
	for (i = vector - 1; i >= 0; i--)
		free_irq(adapter->msix_entries[--vector].vector,
1774
		         adapter->q_vector[i]);
1775 1776
	adapter->flags &= ~IXGBE_FLAG_MSIX_ENABLED;
	pci_disable_msix(adapter->pdev);
1777 1778
	kfree(adapter->msix_entries);
	adapter->msix_entries = NULL;
1779
out:
1780 1781 1782
	return err;
}

1783 1784
static void ixgbe_set_itr(struct ixgbe_adapter *adapter)
{
1785
	struct ixgbe_q_vector *q_vector = adapter->q_vector[0];
1786 1787
	u8 current_itr;
	u32 new_itr = q_vector->eitr;
1788 1789
	struct ixgbe_ring *rx_ring = adapter->rx_ring[0];
	struct ixgbe_ring *tx_ring = adapter->tx_ring[0];
1790

1791
	q_vector->tx_itr = ixgbe_update_itr(adapter, new_itr,
1792 1793 1794
	                                    q_vector->tx_itr,
	                                    tx_ring->total_packets,
	                                    tx_ring->total_bytes);
1795
	q_vector->rx_itr = ixgbe_update_itr(adapter, new_itr,
1796 1797 1798
	                                    q_vector->rx_itr,
	                                    rx_ring->total_packets,
	                                    rx_ring->total_bytes);
1799

1800
	current_itr = max(q_vector->rx_itr, q_vector->tx_itr);
1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817

	switch (current_itr) {
	/* counts and packets in update_itr are dependent on these numbers */
	case lowest_latency:
		new_itr = 100000;
		break;
	case low_latency:
		new_itr = 20000; /* aka hwitr = ~200 */
		break;
	case bulk_latency:
		new_itr = 8000;
		break;
	default:
		break;
	}

	if (new_itr != q_vector->eitr) {
1818 1819
		/* do an exponential smoothing */
		new_itr = ((q_vector->eitr * 90)/100) + ((new_itr * 10)/100);
1820 1821 1822

		/* save the algorithm value here, not the smoothed one */
		q_vector->eitr = new_itr;
1823 1824

		ixgbe_write_eitr(q_vector);
1825 1826 1827 1828 1829
	}

	return;
}

1830 1831 1832 1833 1834 1835 1836
/**
 * ixgbe_irq_enable - Enable default interrupt generation settings
 * @adapter: board private structure
 **/
static inline void ixgbe_irq_enable(struct ixgbe_adapter *adapter)
{
	u32 mask;
1837 1838

	mask = (IXGBE_EIMS_ENABLE_MASK & ~IXGBE_EIMS_RTX_QUEUE);
1839 1840
	if (adapter->flags & IXGBE_FLAG_FAN_FAIL_CAPABLE)
		mask |= IXGBE_EIMS_GPI_SDP1;
1841
	if (adapter->hw.mac.type == ixgbe_mac_82599EB) {
1842
		mask |= IXGBE_EIMS_ECC;
1843 1844
		mask |= IXGBE_EIMS_GPI_SDP1;
		mask |= IXGBE_EIMS_GPI_SDP2;
1845 1846
		if (adapter->num_vfs)
			mask |= IXGBE_EIMS_MAILBOX;
1847
	}
1848 1849 1850
	if (adapter->flags & IXGBE_FLAG_FDIR_HASH_CAPABLE ||
	    adapter->flags & IXGBE_FLAG_FDIR_PERFECT_CAPABLE)
		mask |= IXGBE_EIMS_FLOW_DIR;
1851

1852
	IXGBE_WRITE_REG(&adapter->hw, IXGBE_EIMS, mask);
1853
	ixgbe_irq_enable_queues(adapter, ~0);
1854
	IXGBE_WRITE_FLUSH(&adapter->hw);
1855 1856 1857 1858 1859

	if (adapter->num_vfs > 32) {
		u32 eitrsel = (1 << (adapter->num_vfs - 32)) - 1;
		IXGBE_WRITE_REG(&adapter->hw, IXGBE_EITRSEL, eitrsel);
	}
1860
}
1861

1862
/**
1863
 * ixgbe_intr - legacy mode Interrupt Handler
1864 1865 1866 1867 1868 1869 1870 1871
 * @irq: interrupt number
 * @data: pointer to a network interface device structure
 **/
static irqreturn_t ixgbe_intr(int irq, void *data)
{
	struct net_device *netdev = data;
	struct ixgbe_adapter *adapter = netdev_priv(netdev);
	struct ixgbe_hw *hw = &adapter->hw;
1872
	struct ixgbe_q_vector *q_vector = adapter->q_vector[0];
1873 1874
	u32 eicr;

1875 1876 1877 1878 1879 1880
	/*
	 * Workaround for silicon errata.  Mask the interrupts
	 * before the read of EICR.
	 */
	IXGBE_WRITE_REG(hw, IXGBE_EIMC, IXGBE_IRQ_CLEAR_MASK);

1881 1882 1883
	/* for NAPI, using EIAM to auto-mask tx/rx interrupt bits on read
	 * therefore no explict interrupt disable is necessary */
	eicr = IXGBE_READ_REG(hw, IXGBE_EICR);
1884 1885 1886 1887 1888
	if (!eicr) {
		/* shared interrupt alert!
		 * make sure interrupts are enabled because the read will
		 * have disabled interrupts due to EIAM */
		ixgbe_irq_enable(adapter);
1889
		return IRQ_NONE;	/* Not our interrupt */
1890
	}
1891

1892 1893
	if (eicr & IXGBE_EICR_LSC)
		ixgbe_check_lsc(adapter);
1894

1895 1896 1897
	if (hw->mac.type == ixgbe_mac_82599EB)
		ixgbe_check_sfp_event(adapter, eicr);

1898 1899
	ixgbe_check_fan_failure(adapter, eicr);

1900
	if (napi_schedule_prep(&(q_vector->napi))) {
1901 1902 1903 1904
		adapter->tx_ring[0]->total_packets = 0;
		adapter->tx_ring[0]->total_bytes = 0;
		adapter->rx_ring[0]->total_packets = 0;
		adapter->rx_ring[0]->total_bytes = 0;
1905
		/* would disable interrupts here but EIAM disabled it */
1906
		__napi_schedule(&(q_vector->napi));
1907 1908 1909 1910 1911
	}

	return IRQ_HANDLED;
}

1912 1913 1914 1915 1916
static inline void ixgbe_reset_q_vectors(struct ixgbe_adapter *adapter)
{
	int i, q_vectors = adapter->num_msix_vectors - NON_Q_VECTORS;

	for (i = 0; i < q_vectors; i++) {
1917
		struct ixgbe_q_vector *q_vector = adapter->q_vector[i];
1918 1919 1920 1921 1922 1923 1924
		bitmap_zero(q_vector->rxr_idx, MAX_RX_QUEUES);
		bitmap_zero(q_vector->txr_idx, MAX_TX_QUEUES);
		q_vector->rxr_count = 0;
		q_vector->txr_count = 0;
	}
}

1925 1926 1927 1928 1929 1930 1931
/**
 * ixgbe_request_irq - initialize interrupts
 * @adapter: board private structure
 *
 * Attempts to configure interrupts using the best available
 * capabilities of the hardware and kernel.
 **/
1932
static int ixgbe_request_irq(struct ixgbe_adapter *adapter)
1933 1934
{
	struct net_device *netdev = adapter->netdev;
1935
	int err;
1936

1937 1938 1939
	if (adapter->flags & IXGBE_FLAG_MSIX_ENABLED) {
		err = ixgbe_request_msix_irqs(adapter);
	} else if (adapter->flags & IXGBE_FLAG_MSI_ENABLED) {
1940
		err = request_irq(adapter->pdev->irq, ixgbe_intr, 0,
1941
		                  netdev->name, netdev);
1942
	} else {
1943
		err = request_irq(adapter->pdev->irq, ixgbe_intr, IRQF_SHARED,
1944
		                  netdev->name, netdev);
1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957
	}

	if (err)
		DPRINTK(PROBE, ERR, "request_irq failed, Error %d\n", err);

	return err;
}

static void ixgbe_free_irq(struct ixgbe_adapter *adapter)
{
	struct net_device *netdev = adapter->netdev;

	if (adapter->flags & IXGBE_FLAG_MSIX_ENABLED) {
1958
		int i, q_vectors;
1959

1960 1961 1962
		q_vectors = adapter->num_msix_vectors;

		i = q_vectors - 1;
1963 1964
		free_irq(adapter->msix_entries[i].vector, netdev);

1965 1966 1967
		i--;
		for (; i >= 0; i--) {
			free_irq(adapter->msix_entries[i].vector,
1968
			         adapter->q_vector[i]);
1969 1970 1971 1972 1973
		}

		ixgbe_reset_q_vectors(adapter);
	} else {
		free_irq(adapter->pdev->irq, netdev);
1974 1975 1976
	}
}

1977 1978 1979 1980 1981 1982
/**
 * ixgbe_irq_disable - Mask off interrupt generation on the NIC
 * @adapter: board private structure
 **/
static inline void ixgbe_irq_disable(struct ixgbe_adapter *adapter)
{
1983 1984 1985 1986 1987
	if (adapter->hw.mac.type == ixgbe_mac_82598EB) {
		IXGBE_WRITE_REG(&adapter->hw, IXGBE_EIMC, ~0);
	} else {
		IXGBE_WRITE_REG(&adapter->hw, IXGBE_EIMC, 0xFFFF0000);
		IXGBE_WRITE_REG(&adapter->hw, IXGBE_EIMC_EX(0), ~0);
1988
		IXGBE_WRITE_REG(&adapter->hw, IXGBE_EIMC_EX(1), ~0);
1989 1990
		if (adapter->num_vfs > 32)
			IXGBE_WRITE_REG(&adapter->hw, IXGBE_EITRSEL, 0);
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
	}
	IXGBE_WRITE_FLUSH(&adapter->hw);
	if (adapter->flags & IXGBE_FLAG_MSIX_ENABLED) {
		int i;
		for (i = 0; i < adapter->num_msix_vectors; i++)
			synchronize_irq(adapter->msix_entries[i].vector);
	} else {
		synchronize_irq(adapter->pdev->irq);
	}
}

2002 2003 2004 2005 2006 2007 2008 2009
/**
 * ixgbe_configure_msi_and_legacy - Initialize PIN (INTA...) and MSI interrupts
 *
 **/
static void ixgbe_configure_msi_and_legacy(struct ixgbe_adapter *adapter)
{
	struct ixgbe_hw *hw = &adapter->hw;

2010
	IXGBE_WRITE_REG(hw, IXGBE_EITR(0),
2011
	                EITR_INTS_PER_SEC_TO_REG(adapter->rx_eitr_param));
2012

2013 2014
	ixgbe_set_ivar(adapter, 0, 0, 0);
	ixgbe_set_ivar(adapter, 1, 0, 0);
2015 2016 2017 2018 2019

	map_vector_to_rxq(adapter, 0, 0);
	map_vector_to_txq(adapter, 0, 0);

	DPRINTK(HW, INFO, "Legacy interrupt IVAR setup done\n");
2020 2021 2022
}

/**
2023
 * ixgbe_configure_tx - Configure 8259x Transmit Unit after Reset
2024 2025 2026 2027 2028 2029
 * @adapter: board private structure
 *
 * Configure the Tx unit of the MAC after a reset.
 **/
static void ixgbe_configure_tx(struct ixgbe_adapter *adapter)
{
2030
	u64 tdba;
2031
	struct ixgbe_hw *hw = &adapter->hw;
2032
	u32 i, j, tdlen, txctrl;
2033 2034 2035

	/* Setup the HW Tx Head and Tail descriptor pointers */
	for (i = 0; i < adapter->num_tx_queues; i++) {
2036
		struct ixgbe_ring *ring = adapter->tx_ring[i];
2037 2038 2039
		j = ring->reg_idx;
		tdba = ring->dma;
		tdlen = ring->count * sizeof(union ixgbe_adv_tx_desc);
2040
		IXGBE_WRITE_REG(hw, IXGBE_TDBAL(j),
2041
		                (tdba & DMA_BIT_MASK(32)));
2042 2043 2044 2045
		IXGBE_WRITE_REG(hw, IXGBE_TDBAH(j), (tdba >> 32));
		IXGBE_WRITE_REG(hw, IXGBE_TDLEN(j), tdlen);
		IXGBE_WRITE_REG(hw, IXGBE_TDH(j), 0);
		IXGBE_WRITE_REG(hw, IXGBE_TDT(j), 0);
2046 2047
		adapter->tx_ring[i]->head = IXGBE_TDH(j);
		adapter->tx_ring[i]->tail = IXGBE_TDT(j);
2048 2049
		/*
		 * Disable Tx Head Writeback RO bit, since this hoses
2050 2051
		 * bookkeeping if things aren't delivered in order.
		 */
2052 2053 2054 2055 2056 2057 2058 2059 2060
		switch (hw->mac.type) {
		case ixgbe_mac_82598EB:
			txctrl = IXGBE_READ_REG(hw, IXGBE_DCA_TXCTRL(j));
			break;
		case ixgbe_mac_82599EB:
		default:
			txctrl = IXGBE_READ_REG(hw, IXGBE_DCA_TXCTRL_82599(j));
			break;
		}
2061
		txctrl &= ~IXGBE_DCA_TXCTRL_TX_WB_RO_EN;
2062 2063 2064 2065 2066 2067 2068 2069 2070
		switch (hw->mac.type) {
		case ixgbe_mac_82598EB:
			IXGBE_WRITE_REG(hw, IXGBE_DCA_TXCTRL(j), txctrl);
			break;
		case ixgbe_mac_82599EB:
		default:
			IXGBE_WRITE_REG(hw, IXGBE_DCA_TXCTRL_82599(j), txctrl);
			break;
		}
2071
	}
2072

2073
	if (hw->mac.type == ixgbe_mac_82599EB) {
2074
		u32 rttdcs;
2075
		u32 mask;
2076 2077 2078 2079 2080 2081

		/* disable the arbiter while setting MTQC */
		rttdcs = IXGBE_READ_REG(hw, IXGBE_RTTDCS);
		rttdcs |= IXGBE_RTTDCS_ARBDIS;
		IXGBE_WRITE_REG(hw, IXGBE_RTTDCS, rttdcs);

2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097
		/* set transmit pool layout */
		mask = (IXGBE_FLAG_SRIOV_ENABLED | IXGBE_FLAG_DCB_ENABLED);
		switch (adapter->flags & mask) {

		case (IXGBE_FLAG_SRIOV_ENABLED):
			IXGBE_WRITE_REG(hw, IXGBE_MTQC,
					(IXGBE_MTQC_VT_ENA | IXGBE_MTQC_64VF));
			break;

		case (IXGBE_FLAG_DCB_ENABLED):
			/* We enable 8 traffic classes, DCB only */
			IXGBE_WRITE_REG(hw, IXGBE_MTQC,
				      (IXGBE_MTQC_RT_ENA | IXGBE_MTQC_8TC_8TQ));
			break;

		default:
2098
			IXGBE_WRITE_REG(hw, IXGBE_MTQC, IXGBE_MTQC_64Q_1PB);
2099 2100
			break;
		}
2101 2102 2103 2104

		/* re-eable the arbiter */
		rttdcs &= ~IXGBE_RTTDCS_ARBDIS;
		IXGBE_WRITE_REG(hw, IXGBE_RTTDCS, rttdcs);
2105
	}
2106 2107
}

2108
#define IXGBE_SRRCTL_BSIZEHDRSIZE_SHIFT 2
2109

2110 2111
static void ixgbe_configure_srrctl(struct ixgbe_adapter *adapter,
                                   struct ixgbe_ring *rx_ring)
2112 2113
{
	u32 srrctl;
2114
	int index;
2115
	struct ixgbe_ring_feature *feature = adapter->ring_feature;
2116

2117 2118 2119
	index = rx_ring->reg_idx;
	if (adapter->hw.mac.type == ixgbe_mac_82598EB) {
		unsigned long mask;
2120
		mask = (unsigned long) feature[RING_F_RSS].mask;
2121
		index = index & mask;
2122 2123 2124 2125 2126 2127
	}
	srrctl = IXGBE_READ_REG(&adapter->hw, IXGBE_SRRCTL(index));

	srrctl &= ~IXGBE_SRRCTL_BSIZEHDR_MASK;
	srrctl &= ~IXGBE_SRRCTL_BSIZEPKT_MASK;

2128 2129 2130
	srrctl |= (IXGBE_RX_HDR_SIZE << IXGBE_SRRCTL_BSIZEHDRSIZE_SHIFT) &
		  IXGBE_SRRCTL_BSIZEHDR_MASK;

2131
	if (rx_ring->flags & IXGBE_RING_RX_PS_ENABLED) {
2132 2133 2134 2135 2136
#if (PAGE_SIZE / 2) > IXGBE_MAX_RXBUFFER
		srrctl |= IXGBE_MAX_RXBUFFER >> IXGBE_SRRCTL_BSIZEPKT_SHIFT;
#else
		srrctl |= (PAGE_SIZE / 2) >> IXGBE_SRRCTL_BSIZEPKT_SHIFT;
#endif
2137 2138
		srrctl |= IXGBE_SRRCTL_DESCTYPE_HDR_SPLIT_ALWAYS;
	} else {
2139 2140
		srrctl |= ALIGN(rx_ring->rx_buf_len, 1024) >>
			  IXGBE_SRRCTL_BSIZEPKT_SHIFT;
2141 2142
		srrctl |= IXGBE_SRRCTL_DESCTYPE_ADV_ONEBUF;
	}
2143

2144 2145
	IXGBE_WRITE_REG(&adapter->hw, IXGBE_SRRCTL(index), srrctl);
}
2146

2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158
static u32 ixgbe_setup_mrqc(struct ixgbe_adapter *adapter)
{
	u32 mrqc = 0;
	int mask;

	if (!(adapter->hw.mac.type == ixgbe_mac_82599EB))
		return mrqc;

	mask = adapter->flags & (IXGBE_FLAG_RSS_ENABLED
#ifdef CONFIG_IXGBE_DCB
				 | IXGBE_FLAG_DCB_ENABLED
#endif
2159
				 | IXGBE_FLAG_SRIOV_ENABLED
2160 2161 2162 2163 2164 2165
				);

	switch (mask) {
	case (IXGBE_FLAG_RSS_ENABLED):
		mrqc = IXGBE_MRQC_RSSEN;
		break;
2166 2167 2168
	case (IXGBE_FLAG_SRIOV_ENABLED):
		mrqc = IXGBE_MRQC_VMDQEN;
		break;
2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180
#ifdef CONFIG_IXGBE_DCB
	case (IXGBE_FLAG_DCB_ENABLED):
		mrqc = IXGBE_MRQC_RT8TCEN;
		break;
#endif /* CONFIG_IXGBE_DCB */
	default:
		break;
	}

	return mrqc;
}

2181 2182 2183 2184 2185
/**
 * ixgbe_configure_rscctl - enable RSC for the indicated ring
 * @adapter:    address of board private structure
 * @index:      index of ring to set
 **/
2186
static void ixgbe_configure_rscctl(struct ixgbe_adapter *adapter, int index)
2187 2188 2189 2190 2191
{
	struct ixgbe_ring *rx_ring;
	struct ixgbe_hw *hw = &adapter->hw;
	int j;
	u32 rscctrl;
2192
	int rx_buf_len;
2193

2194
	rx_ring = adapter->rx_ring[index];
2195
	j = rx_ring->reg_idx;
2196
	rx_buf_len = rx_ring->rx_buf_len;
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
	rscctrl = IXGBE_READ_REG(hw, IXGBE_RSCCTL(j));
	rscctrl |= IXGBE_RSCCTL_RSCEN;
	/*
	 * we must limit the number of descriptors so that the
	 * total size of max desc * buf_len is not greater
	 * than 65535
	 */
	if (rx_ring->flags & IXGBE_RING_RX_PS_ENABLED) {
#if (MAX_SKB_FRAGS > 16)
		rscctrl |= IXGBE_RSCCTL_MAXDESC_16;
#elif (MAX_SKB_FRAGS > 8)
		rscctrl |= IXGBE_RSCCTL_MAXDESC_8;
#elif (MAX_SKB_FRAGS > 4)
		rscctrl |= IXGBE_RSCCTL_MAXDESC_4;
#else
		rscctrl |= IXGBE_RSCCTL_MAXDESC_1;
#endif
	} else {
		if (rx_buf_len < IXGBE_RXBUFFER_4096)
			rscctrl |= IXGBE_RSCCTL_MAXDESC_16;
		else if (rx_buf_len < IXGBE_RXBUFFER_8192)
			rscctrl |= IXGBE_RSCCTL_MAXDESC_8;
		else
			rscctrl |= IXGBE_RSCCTL_MAXDESC_4;
	}
	IXGBE_WRITE_REG(hw, IXGBE_RSCCTL(j), rscctrl);
}

2225
/**
2226
 * ixgbe_configure_rx - Configure 8259x Receive Unit after Reset
2227 2228 2229 2230 2231 2232 2233 2234
 * @adapter: board private structure
 *
 * Configure the Rx unit of the MAC after a reset.
 **/
static void ixgbe_configure_rx(struct ixgbe_adapter *adapter)
{
	u64 rdba;
	struct ixgbe_hw *hw = &adapter->hw;
2235
	struct ixgbe_ring *rx_ring;
2236 2237
	struct net_device *netdev = adapter->netdev;
	int max_frame = netdev->mtu + ETH_HLEN + ETH_FCS_LEN;
2238
	int i, j;
2239
	u32 rdlen, rxctrl, rxcsum;
2240 2241 2242
	static const u32 seed[10] = { 0xE291D73D, 0x1805EC6C, 0x2A94B30D,
	                  0xA54F2BEC, 0xEA49AF7C, 0xE214AD3D, 0xB855AABE,
	                  0x6A3E67EA, 0x14364D17, 0x3BED200D};
2243
	u32 fctrl, hlreg0;
2244
	u32 reta = 0, mrqc = 0;
2245
	u32 rdrxctl;
2246
	int rx_buf_len;
2247 2248

	/* Decide whether to use packet split mode or not */
2249 2250 2251
	/* Do not use packet split if we're in SR-IOV Mode */
	if (!adapter->num_vfs)
		adapter->flags |= IXGBE_FLAG_RX_PS_ENABLED;
2252 2253 2254

	/* Set the RX buffer length according to the mode */
	if (adapter->flags & IXGBE_FLAG_RX_PS_ENABLED) {
2255
		rx_buf_len = IXGBE_RX_HDR_SIZE;
2256 2257 2258 2259 2260
		if (hw->mac.type == ixgbe_mac_82599EB) {
			/* PSRTYPE must be initialized in 82599 */
			u32 psrtype = IXGBE_PSRTYPE_TCPHDR |
			              IXGBE_PSRTYPE_UDPHDR |
			              IXGBE_PSRTYPE_IPV4HDR |
Y
Yi Zou 已提交
2261 2262
			              IXGBE_PSRTYPE_IPV6HDR |
			              IXGBE_PSRTYPE_L2HDR;
2263 2264 2265
			IXGBE_WRITE_REG(hw,
					IXGBE_PSRTYPE(adapter->num_vfs),
					psrtype);
2266
		}
2267
	} else {
2268
		if (!(adapter->flags2 & IXGBE_FLAG2_RSC_ENABLED) &&
A
Alexander Duyck 已提交
2269
		    (netdev->mtu <= ETH_DATA_LEN))
2270
			rx_buf_len = MAXIMUM_ETHERNET_VLAN_SIZE;
2271
		else
2272
			rx_buf_len = ALIGN(max_frame, 1024);
2273 2274 2275 2276
	}

	fctrl = IXGBE_READ_REG(&adapter->hw, IXGBE_FCTRL);
	fctrl |= IXGBE_FCTRL_BAM;
2277
	fctrl |= IXGBE_FCTRL_DPF; /* discard pause frames when FC enabled */
2278
	fctrl |= IXGBE_FCTRL_PMCF;
2279 2280 2281 2282 2283 2284 2285
	IXGBE_WRITE_REG(&adapter->hw, IXGBE_FCTRL, fctrl);

	hlreg0 = IXGBE_READ_REG(hw, IXGBE_HLREG0);
	if (adapter->netdev->mtu <= ETH_DATA_LEN)
		hlreg0 &= ~IXGBE_HLREG0_JUMBOEN;
	else
		hlreg0 |= IXGBE_HLREG0_JUMBOEN;
2286
#ifdef IXGBE_FCOE
2287
	if (netdev->features & NETIF_F_FCOE_MTU)
2288 2289
		hlreg0 |= IXGBE_HLREG0_JUMBOEN;
#endif
2290 2291
	IXGBE_WRITE_REG(hw, IXGBE_HLREG0, hlreg0);

2292
	rdlen = adapter->rx_ring[0]->count * sizeof(union ixgbe_adv_rx_desc);
2293 2294 2295 2296
	/* disable receives while setting up the descriptors */
	rxctrl = IXGBE_READ_REG(hw, IXGBE_RXCTRL);
	IXGBE_WRITE_REG(hw, IXGBE_RXCTRL, rxctrl & ~IXGBE_RXCTRL_RXEN);

2297 2298 2299 2300
	/*
	 * Setup the HW Rx Head and Tail Descriptor Pointers and
	 * the Base and Length of the Rx Descriptor Ring
	 */
2301
	for (i = 0; i < adapter->num_rx_queues; i++) {
2302
		rx_ring = adapter->rx_ring[i];
2303 2304
		rdba = rx_ring->dma;
		j = rx_ring->reg_idx;
2305
		IXGBE_WRITE_REG(hw, IXGBE_RDBAL(j), (rdba & DMA_BIT_MASK(32)));
2306 2307 2308 2309
		IXGBE_WRITE_REG(hw, IXGBE_RDBAH(j), (rdba >> 32));
		IXGBE_WRITE_REG(hw, IXGBE_RDLEN(j), rdlen);
		IXGBE_WRITE_REG(hw, IXGBE_RDH(j), 0);
		IXGBE_WRITE_REG(hw, IXGBE_RDT(j), 0);
2310 2311 2312
		rx_ring->head = IXGBE_RDH(j);
		rx_ring->tail = IXGBE_RDT(j);
		rx_ring->rx_buf_len = rx_buf_len;
2313

2314 2315
		if (adapter->flags & IXGBE_FLAG_RX_PS_ENABLED)
			rx_ring->flags |= IXGBE_RING_RX_PS_ENABLED;
2316 2317
		else
			rx_ring->flags &= ~IXGBE_RING_RX_PS_ENABLED;
2318

2319
#ifdef IXGBE_FCOE
2320
		if (netdev->features & NETIF_F_FCOE_MTU) {
2321 2322
			struct ixgbe_ring_feature *f;
			f = &adapter->ring_feature[RING_F_FCOE];
2323 2324 2325 2326 2327 2328
			if ((i >= f->mask) && (i < f->mask + f->indices)) {
				rx_ring->flags &= ~IXGBE_RING_RX_PS_ENABLED;
				if (rx_buf_len < IXGBE_FCOE_JUMBO_FRAME_SIZE)
					rx_ring->rx_buf_len =
					        IXGBE_FCOE_JUMBO_FRAME_SIZE;
			}
2329 2330 2331
		}

#endif /* IXGBE_FCOE */
2332
		ixgbe_configure_srrctl(adapter, rx_ring);
2333 2334
	}

2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345
	if (hw->mac.type == ixgbe_mac_82598EB) {
		/*
		 * For VMDq support of different descriptor types or
		 * buffer sizes through the use of multiple SRRCTL
		 * registers, RDRXCTL.MVMEN must be set to 1
		 *
		 * also, the manual doesn't mention it clearly but DCA hints
		 * will only use queue 0's tags unless this bit is set.  Side
		 * effects of setting this bit are only that SRRCTL must be
		 * fully programmed [0..15]
		 */
2346 2347 2348
		rdrxctl = IXGBE_READ_REG(hw, IXGBE_RDRXCTL);
		rdrxctl |= IXGBE_RDRXCTL_MVMEN;
		IXGBE_WRITE_REG(hw, IXGBE_RDRXCTL, rdrxctl);
2349
	}
2350

2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374
	if (adapter->flags & IXGBE_FLAG_SRIOV_ENABLED) {
		u32 vt_reg_bits;
		u32 reg_offset, vf_shift;
		u32 vmdctl = IXGBE_READ_REG(hw, IXGBE_VT_CTL);
		vt_reg_bits = IXGBE_VMD_CTL_VMDQ_EN
			| IXGBE_VT_CTL_REPLEN;
		vt_reg_bits |= (adapter->num_vfs <<
				IXGBE_VT_CTL_POOL_SHIFT);
		IXGBE_WRITE_REG(hw, IXGBE_VT_CTL, vmdctl | vt_reg_bits);
		IXGBE_WRITE_REG(hw, IXGBE_MRQC, 0);

		vf_shift = adapter->num_vfs % 32;
		reg_offset = adapter->num_vfs / 32;
		IXGBE_WRITE_REG(hw, IXGBE_VFRE(0), 0);
		IXGBE_WRITE_REG(hw, IXGBE_VFRE(1), 0);
		IXGBE_WRITE_REG(hw, IXGBE_VFTE(0), 0);
		IXGBE_WRITE_REG(hw, IXGBE_VFTE(1), 0);
		/* Enable only the PF's pool for Tx/Rx */
		IXGBE_WRITE_REG(hw, IXGBE_VFRE(reg_offset), (1 << vf_shift));
		IXGBE_WRITE_REG(hw, IXGBE_VFTE(reg_offset), (1 << vf_shift));
		IXGBE_WRITE_REG(hw, IXGBE_PFDTXGSWC, IXGBE_PFDTXGSWC_VT_LBEN);
		ixgbe_set_vmolr(hw, adapter->num_vfs);
	}

2375
	/* Program MRQC for the distribution of queues */
2376
	mrqc = ixgbe_setup_mrqc(adapter);
2377

2378
	if (adapter->flags & IXGBE_FLAG_RSS_ENABLED) {
2379
		/* Fill out redirection table */
2380 2381 2382 2383 2384 2385 2386 2387
		for (i = 0, j = 0; i < 128; i++, j++) {
			if (j == adapter->ring_feature[RING_F_RSS].indices)
				j = 0;
			/* reta = 4-byte sliding window of
			 * 0x00..(indices-1)(indices-1)00..etc. */
			reta = (reta << 8) | (j * 0x11);
			if ((i & 3) == 3)
				IXGBE_WRITE_REG(hw, IXGBE_RETA(i >> 2), reta);
2388 2389 2390 2391
		}

		/* Fill out hash function seeds */
		for (i = 0; i < 10; i++)
2392
			IXGBE_WRITE_REG(hw, IXGBE_RSSRK(i), seed[i]);
2393

2394 2395
		if (hw->mac.type == ixgbe_mac_82598EB)
			mrqc |= IXGBE_MRQC_RSSEN;
2396
		    /* Perform hash on these packet types */
2397 2398 2399 2400 2401 2402
		mrqc |= IXGBE_MRQC_RSS_FIELD_IPV4
		      | IXGBE_MRQC_RSS_FIELD_IPV4_TCP
		      | IXGBE_MRQC_RSS_FIELD_IPV4_UDP
		      | IXGBE_MRQC_RSS_FIELD_IPV6
		      | IXGBE_MRQC_RSS_FIELD_IPV6_TCP
		      | IXGBE_MRQC_RSS_FIELD_IPV6_UDP;
2403
	}
2404
	IXGBE_WRITE_REG(hw, IXGBE_MRQC, mrqc);
2405

2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419
	if (adapter->num_vfs) {
		u32 reg;

		/* Map PF MAC address in RAR Entry 0 to first pool
		 * following VFs */
		hw->mac.ops.set_vmdq(hw, 0, adapter->num_vfs);

		/* Set up VF register offsets for selected VT Mode, i.e.
		 * 64 VFs for SR-IOV */
		reg = IXGBE_READ_REG(hw, IXGBE_GCR_EXT);
		reg |= IXGBE_GCR_EXT_SRIOV;
		IXGBE_WRITE_REG(hw, IXGBE_GCR_EXT, reg);
	}

2420 2421 2422 2423 2424 2425
	rxcsum = IXGBE_READ_REG(hw, IXGBE_RXCSUM);

	if (adapter->flags & IXGBE_FLAG_RSS_ENABLED ||
	    adapter->flags & IXGBE_FLAG_RX_CSUM_ENABLED) {
		/* Disable indicating checksum in descriptor, enables
		 * RSS hash */
2426 2427
		rxcsum |= IXGBE_RXCSUM_PCSD;
	}
2428 2429 2430 2431 2432 2433 2434
	if (!(rxcsum & IXGBE_RXCSUM_PCSD)) {
		/* Enable IPv4 payload checksum for UDP fragments
		 * if PCSD is not set */
		rxcsum |= IXGBE_RXCSUM_IPPCSE;
	}

	IXGBE_WRITE_REG(hw, IXGBE_RXCSUM, rxcsum);
2435 2436 2437 2438

	if (hw->mac.type == ixgbe_mac_82599EB) {
		rdrxctl = IXGBE_READ_REG(hw, IXGBE_RDRXCTL);
		rdrxctl |= IXGBE_RDRXCTL_CRCSTRIP;
A
Alexander Duyck 已提交
2439
		rdrxctl &= ~IXGBE_RDRXCTL_RSCFRSTSIZE;
2440 2441
		IXGBE_WRITE_REG(hw, IXGBE_RDRXCTL, rdrxctl);
	}
A
Alexander Duyck 已提交
2442

2443
	if (adapter->flags2 & IXGBE_FLAG2_RSC_ENABLED) {
A
Alexander Duyck 已提交
2444
		/* Enable 82599 HW-RSC */
2445
		for (i = 0; i < adapter->num_rx_queues; i++)
2446
			ixgbe_configure_rscctl(adapter, i);
2447

A
Alexander Duyck 已提交
2448 2449 2450 2451
		/* Disable RSC for ACK packets */
		IXGBE_WRITE_REG(hw, IXGBE_RSCDBU,
		   (IXGBE_RSCDBU_RSCACKDIS | IXGBE_READ_REG(hw, IXGBE_RSCDBU)));
	}
2452 2453
}

2454 2455 2456 2457
static void ixgbe_vlan_rx_add_vid(struct net_device *netdev, u16 vid)
{
	struct ixgbe_adapter *adapter = netdev_priv(netdev);
	struct ixgbe_hw *hw = &adapter->hw;
2458
	int pool_ndx = adapter->num_vfs;
2459 2460

	/* add VID to filter table */
2461
	hw->mac.ops.set_vfta(&adapter->hw, vid, pool_ndx, true);
2462 2463 2464 2465 2466 2467
}

static void ixgbe_vlan_rx_kill_vid(struct net_device *netdev, u16 vid)
{
	struct ixgbe_adapter *adapter = netdev_priv(netdev);
	struct ixgbe_hw *hw = &adapter->hw;
2468
	int pool_ndx = adapter->num_vfs;
2469 2470 2471 2472 2473 2474 2475 2476 2477 2478

	if (!test_bit(__IXGBE_DOWN, &adapter->state))
		ixgbe_irq_disable(adapter);

	vlan_group_set_device(adapter->vlgrp, vid, NULL);

	if (!test_bit(__IXGBE_DOWN, &adapter->state))
		ixgbe_irq_enable(adapter);

	/* remove VID from filter table */
2479
	hw->mac.ops.set_vfta(&adapter->hw, vid, pool_ndx, false);
2480 2481
}

2482
static void ixgbe_vlan_rx_register(struct net_device *netdev,
2483
                                   struct vlan_group *grp)
2484 2485 2486
{
	struct ixgbe_adapter *adapter = netdev_priv(netdev);
	u32 ctrl;
2487
	int i, j;
2488

2489 2490
	if (!test_bit(__IXGBE_DOWN, &adapter->state))
		ixgbe_irq_disable(adapter);
2491 2492
	adapter->vlgrp = grp;

2493 2494 2495 2496 2497 2498
	/*
	 * For a DCB driver, always enable VLAN tag stripping so we can
	 * still receive traffic from a DCB-enabled host even if we're
	 * not in DCB mode.
	 */
	ctrl = IXGBE_READ_REG(&adapter->hw, IXGBE_VLNCTRL);
2499 2500 2501 2502 2503

	/* Disable CFI check */
	ctrl &= ~IXGBE_VLNCTRL_CFIEN;

	/* enable VLAN tag stripping */
2504
	if (adapter->hw.mac.type == ixgbe_mac_82598EB) {
2505
		ctrl |= IXGBE_VLNCTRL_VME;
2506 2507
	} else if (adapter->hw.mac.type == ixgbe_mac_82599EB) {
		for (i = 0; i < adapter->num_rx_queues; i++) {
2508
			u32 ctrl;
2509
			j = adapter->rx_ring[i]->reg_idx;
2510 2511 2512 2513
			ctrl = IXGBE_READ_REG(&adapter->hw, IXGBE_RXDCTL(j));
			ctrl |= IXGBE_RXDCTL_VME;
			IXGBE_WRITE_REG(&adapter->hw, IXGBE_RXDCTL(j), ctrl);
		}
2514
	}
2515 2516 2517

	IXGBE_WRITE_REG(&adapter->hw, IXGBE_VLNCTRL, ctrl);

2518
	ixgbe_vlan_rx_add_vid(netdev, 0);
2519

2520 2521
	if (!test_bit(__IXGBE_DOWN, &adapter->state))
		ixgbe_irq_enable(adapter);
2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537
}

static void ixgbe_restore_vlan(struct ixgbe_adapter *adapter)
{
	ixgbe_vlan_rx_register(adapter->netdev, adapter->vlgrp);

	if (adapter->vlgrp) {
		u16 vid;
		for (vid = 0; vid < VLAN_GROUP_ARRAY_LEN; vid++) {
			if (!vlan_group_get_device(adapter->vlgrp, vid))
				continue;
			ixgbe_vlan_rx_add_vid(adapter->netdev, vid);
		}
	}
}

2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552
static u8 *ixgbe_addr_list_itr(struct ixgbe_hw *hw, u8 **mc_addr_ptr, u32 *vmdq)
{
	struct dev_mc_list *mc_ptr;
	u8 *addr = *mc_addr_ptr;
	*vmdq = 0;

	mc_ptr = container_of(addr, struct dev_mc_list, dmi_addr[0]);
	if (mc_ptr->next)
		*mc_addr_ptr = mc_ptr->next->dmi_addr;
	else
		*mc_addr_ptr = NULL;

	return addr;
}

2553
/**
2554
 * ixgbe_set_rx_mode - Unicast, Multicast and Promiscuous mode set
2555 2556
 * @netdev: network interface device structure
 *
2557 2558 2559 2560
 * The set_rx_method entry point is called whenever the unicast/multicast
 * address list or the network interface flags are updated.  This routine is
 * responsible for configuring the hardware for proper unicast, multicast and
 * promiscuous mode.
2561
 **/
2562
void ixgbe_set_rx_mode(struct net_device *netdev)
2563 2564 2565
{
	struct ixgbe_adapter *adapter = netdev_priv(netdev);
	struct ixgbe_hw *hw = &adapter->hw;
A
Alexander Duyck 已提交
2566
	u32 fctrl, vlnctrl;
2567 2568
	u8 *addr_list = NULL;
	int addr_count = 0;
2569 2570 2571 2572

	/* Check for Promiscuous and All Multicast modes */

	fctrl = IXGBE_READ_REG(hw, IXGBE_FCTRL);
A
Alexander Duyck 已提交
2573
	vlnctrl = IXGBE_READ_REG(hw, IXGBE_VLNCTRL);
2574 2575

	if (netdev->flags & IFF_PROMISC) {
2576
		hw->addr_ctrl.user_set_promisc = 1;
2577
		fctrl |= (IXGBE_FCTRL_UPE | IXGBE_FCTRL_MPE);
A
Alexander Duyck 已提交
2578
		vlnctrl &= ~IXGBE_VLNCTRL_VFE;
2579
	} else {
2580 2581 2582 2583 2584 2585
		if (netdev->flags & IFF_ALLMULTI) {
			fctrl |= IXGBE_FCTRL_MPE;
			fctrl &= ~IXGBE_FCTRL_UPE;
		} else {
			fctrl &= ~(IXGBE_FCTRL_UPE | IXGBE_FCTRL_MPE);
		}
A
Alexander Duyck 已提交
2586
		vlnctrl |= IXGBE_VLNCTRL_VFE;
2587
		hw->addr_ctrl.user_set_promisc = 0;
2588 2589 2590
	}

	IXGBE_WRITE_REG(hw, IXGBE_FCTRL, fctrl);
A
Alexander Duyck 已提交
2591
	IXGBE_WRITE_REG(hw, IXGBE_VLNCTRL, vlnctrl);
2592

2593
	/* reprogram secondary unicast list */
2594
	hw->mac.ops.update_uc_addr_list(hw, netdev);
2595

2596
	/* reprogram multicast list */
2597
	addr_count = netdev_mc_count(netdev);
2598 2599
	if (addr_count)
		addr_list = netdev->mc_list->dmi_addr;
2600 2601
	hw->mac.ops.update_mc_addr_list(hw, addr_list, addr_count,
	                                ixgbe_addr_list_itr);
2602 2603
	if (adapter->num_vfs)
		ixgbe_restore_vf_multicasts(adapter);
2604 2605
}

2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616
static void ixgbe_napi_enable_all(struct ixgbe_adapter *adapter)
{
	int q_idx;
	struct ixgbe_q_vector *q_vector;
	int q_vectors = adapter->num_msix_vectors - NON_Q_VECTORS;

	/* legacy and MSI only use one vector */
	if (!(adapter->flags & IXGBE_FLAG_MSIX_ENABLED))
		q_vectors = 1;

	for (q_idx = 0; q_idx < q_vectors; q_idx++) {
2617
		struct napi_struct *napi;
2618
		q_vector = adapter->q_vector[q_idx];
2619
		napi = &q_vector->napi;
2620 2621 2622 2623 2624 2625 2626 2627
		if (adapter->flags & IXGBE_FLAG_MSIX_ENABLED) {
			if (!q_vector->rxr_count || !q_vector->txr_count) {
				if (q_vector->txr_count == 1)
					napi->poll = &ixgbe_clean_txonly;
				else if (q_vector->rxr_count == 1)
					napi->poll = &ixgbe_clean_rxonly;
			}
		}
2628 2629

		napi_enable(napi);
2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643
	}
}

static void ixgbe_napi_disable_all(struct ixgbe_adapter *adapter)
{
	int q_idx;
	struct ixgbe_q_vector *q_vector;
	int q_vectors = adapter->num_msix_vectors - NON_Q_VECTORS;

	/* legacy and MSI only use one vector */
	if (!(adapter->flags & IXGBE_FLAG_MSIX_ENABLED))
		q_vectors = 1;

	for (q_idx = 0; q_idx < q_vectors; q_idx++) {
2644
		q_vector = adapter->q_vector[q_idx];
2645 2646 2647 2648
		napi_disable(&q_vector->napi);
	}
}

J
Jeff Kirsher 已提交
2649
#ifdef CONFIG_IXGBE_DCB
2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671
/*
 * ixgbe_configure_dcb - Configure DCB hardware
 * @adapter: ixgbe adapter struct
 *
 * This is called by the driver on open to configure the DCB hardware.
 * This is also called by the gennetlink interface when reconfiguring
 * the DCB state.
 */
static void ixgbe_configure_dcb(struct ixgbe_adapter *adapter)
{
	struct ixgbe_hw *hw = &adapter->hw;
	u32 txdctl, vlnctrl;
	int i, j;

	ixgbe_dcb_check_config(&adapter->dcb_cfg);
	ixgbe_dcb_calculate_tc_credits(&adapter->dcb_cfg, DCB_TX_CONFIG);
	ixgbe_dcb_calculate_tc_credits(&adapter->dcb_cfg, DCB_RX_CONFIG);

	/* reconfigure the hardware */
	ixgbe_dcb_hw_config(&adapter->hw, &adapter->dcb_cfg);

	for (i = 0; i < adapter->num_tx_queues; i++) {
2672
		j = adapter->tx_ring[i]->reg_idx;
2673 2674 2675 2676 2677 2678 2679
		txdctl = IXGBE_READ_REG(hw, IXGBE_TXDCTL(j));
		/* PThresh workaround for Tx hang with DFP enabled. */
		txdctl |= 32;
		IXGBE_WRITE_REG(hw, IXGBE_TXDCTL(j), txdctl);
	}
	/* Enable VLAN tag insert/strip */
	vlnctrl = IXGBE_READ_REG(hw, IXGBE_VLNCTRL);
2680 2681 2682 2683 2684 2685 2686 2687 2688
	if (hw->mac.type == ixgbe_mac_82598EB) {
		vlnctrl |= IXGBE_VLNCTRL_VME | IXGBE_VLNCTRL_VFE;
		vlnctrl &= ~IXGBE_VLNCTRL_CFIEN;
		IXGBE_WRITE_REG(hw, IXGBE_VLNCTRL, vlnctrl);
	} else if (hw->mac.type == ixgbe_mac_82599EB) {
		vlnctrl |= IXGBE_VLNCTRL_VFE;
		vlnctrl &= ~IXGBE_VLNCTRL_CFIEN;
		IXGBE_WRITE_REG(hw, IXGBE_VLNCTRL, vlnctrl);
		for (i = 0; i < adapter->num_rx_queues; i++) {
2689
			j = adapter->rx_ring[i]->reg_idx;
2690 2691 2692 2693 2694
			vlnctrl = IXGBE_READ_REG(hw, IXGBE_RXDCTL(j));
			vlnctrl |= IXGBE_RXDCTL_VME;
			IXGBE_WRITE_REG(hw, IXGBE_RXDCTL(j), vlnctrl);
		}
	}
2695 2696 2697 2698
	hw->mac.ops.set_vfta(&adapter->hw, 0, 0, true);
}

#endif
2699 2700 2701
static void ixgbe_configure(struct ixgbe_adapter *adapter)
{
	struct net_device *netdev = adapter->netdev;
2702
	struct ixgbe_hw *hw = &adapter->hw;
2703 2704
	int i;

2705
	ixgbe_set_rx_mode(netdev);
2706 2707

	ixgbe_restore_vlan(adapter);
J
Jeff Kirsher 已提交
2708
#ifdef CONFIG_IXGBE_DCB
2709
	if (adapter->flags & IXGBE_FLAG_DCB_ENABLED) {
2710 2711 2712 2713
		if (hw->mac.type == ixgbe_mac_82598EB)
			netif_set_gso_max_size(netdev, 32768);
		else
			netif_set_gso_max_size(netdev, 65536);
2714 2715 2716 2717 2718 2719 2720
		ixgbe_configure_dcb(adapter);
	} else {
		netif_set_gso_max_size(netdev, 65536);
	}
#else
	netif_set_gso_max_size(netdev, 65536);
#endif
2721

2722 2723 2724 2725 2726
#ifdef IXGBE_FCOE
	if (adapter->flags & IXGBE_FLAG_FCOE_ENABLED)
		ixgbe_configure_fcoe(adapter);

#endif /* IXGBE_FCOE */
2727 2728
	if (adapter->flags & IXGBE_FLAG_FDIR_HASH_CAPABLE) {
		for (i = 0; i < adapter->num_tx_queues; i++)
2729
			adapter->tx_ring[i]->atr_sample_rate =
2730 2731 2732 2733 2734 2735
			                               adapter->atr_sample_rate;
		ixgbe_init_fdir_signature_82599(hw, adapter->fdir_pballoc);
	} else if (adapter->flags & IXGBE_FLAG_FDIR_PERFECT_CAPABLE) {
		ixgbe_init_fdir_perfect_82599(hw, adapter->fdir_pballoc);
	}

2736 2737 2738
	ixgbe_configure_tx(adapter);
	ixgbe_configure_rx(adapter);
	for (i = 0; i < adapter->num_rx_queues; i++)
2739 2740
		ixgbe_alloc_rx_buffers(adapter, adapter->rx_ring[i],
		                       (adapter->rx_ring[i]->count - 1));
2741 2742
}

2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757
static inline bool ixgbe_is_sfp(struct ixgbe_hw *hw)
{
	switch (hw->phy.type) {
	case ixgbe_phy_sfp_avago:
	case ixgbe_phy_sfp_ftl:
	case ixgbe_phy_sfp_intel:
	case ixgbe_phy_sfp_unknown:
	case ixgbe_phy_tw_tyco:
	case ixgbe_phy_tw_unknown:
		return true;
	default:
		return false;
	}
}

2758
/**
2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791
 * ixgbe_sfp_link_config - set up SFP+ link
 * @adapter: pointer to private adapter struct
 **/
static void ixgbe_sfp_link_config(struct ixgbe_adapter *adapter)
{
	struct ixgbe_hw *hw = &adapter->hw;

		if (hw->phy.multispeed_fiber) {
			/*
			 * In multispeed fiber setups, the device may not have
			 * had a physical connection when the driver loaded.
			 * If that's the case, the initial link configuration
			 * couldn't get the MAC into 10G or 1G mode, so we'll
			 * never have a link status change interrupt fire.
			 * We need to try and force an autonegotiation
			 * session, then bring up link.
			 */
			hw->mac.ops.setup_sfp(hw);
			if (!(adapter->flags & IXGBE_FLAG_IN_SFP_LINK_TASK))
				schedule_work(&adapter->multispeed_fiber_task);
		} else {
			/*
			 * Direct Attach Cu and non-multispeed fiber modules
			 * still need to be configured properly prior to
			 * attempting link.
			 */
			if (!(adapter->flags & IXGBE_FLAG_IN_SFP_MOD_TASK))
				schedule_work(&adapter->sfp_config_module_task);
		}
}

/**
 * ixgbe_non_sfp_link_config - set up non-SFP+ link
2792 2793 2794 2795
 * @hw: pointer to private hardware struct
 *
 * Returns 0 on success, negative on failure
 **/
2796
static int ixgbe_non_sfp_link_config(struct ixgbe_hw *hw)
2797 2798
{
	u32 autoneg;
2799
	bool negotiation, link_up = false;
2800 2801 2802 2803 2804 2805 2806 2807 2808
	u32 ret = IXGBE_ERR_LINK_SETUP;

	if (hw->mac.ops.check_link)
		ret = hw->mac.ops.check_link(hw, &autoneg, &link_up, false);

	if (ret)
		goto link_cfg_out;

	if (hw->mac.ops.get_link_capabilities)
2809
		ret = hw->mac.ops.get_link_capabilities(hw, &autoneg, &negotiation);
2810 2811 2812
	if (ret)
		goto link_cfg_out;

2813 2814
	if (hw->mac.ops.setup_link)
		ret = hw->mac.ops.setup_link(hw, autoneg, negotiation, link_up);
2815 2816 2817 2818
link_cfg_out:
	return ret;
}

2819 2820 2821 2822
#define IXGBE_MAX_RX_DESC_POLL 10
static inline void ixgbe_rx_desc_queue_enable(struct ixgbe_adapter *adapter,
	                                      int rxr)
{
2823
	int j = adapter->rx_ring[rxr]->reg_idx;
2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836
	int k;

	for (k = 0; k < IXGBE_MAX_RX_DESC_POLL; k++) {
		if (IXGBE_READ_REG(&adapter->hw,
		                   IXGBE_RXDCTL(j)) & IXGBE_RXDCTL_ENABLE)
			break;
		else
			msleep(1);
	}
	if (k >= IXGBE_MAX_RX_DESC_POLL) {
		DPRINTK(DRV, ERR, "RXDCTL.ENABLE on Rx queue %d "
		        "not set within the polling period\n", rxr);
	}
2837 2838
	ixgbe_release_rx_desc(&adapter->hw, adapter->rx_ring[rxr],
	                      (adapter->rx_ring[rxr]->count - 1));
2839 2840
}

2841 2842 2843 2844
static int ixgbe_up_complete(struct ixgbe_adapter *adapter)
{
	struct net_device *netdev = adapter->netdev;
	struct ixgbe_hw *hw = &adapter->hw;
2845
	int i, j = 0;
2846
	int num_rx_rings = adapter->num_rx_queues;
2847
	int err;
2848
	int max_frame = netdev->mtu + ETH_HLEN + ETH_FCS_LEN;
2849
	u32 txdctl, rxdctl, mhadd;
2850
	u32 dmatxctl;
2851
	u32 gpie;
2852
	u32 ctrl_ext;
2853

2854 2855
	ixgbe_get_hw_control(adapter);

2856 2857
	if ((adapter->flags & IXGBE_FLAG_MSIX_ENABLED) ||
	    (adapter->flags & IXGBE_FLAG_MSI_ENABLED)) {
2858 2859
		if (adapter->flags & IXGBE_FLAG_MSIX_ENABLED) {
			gpie = (IXGBE_GPIE_MSIX_MODE | IXGBE_GPIE_EIAME |
2860
			        IXGBE_GPIE_PBA_SUPPORT | IXGBE_GPIE_OCD);
2861 2862
		} else {
			/* MSI only */
2863
			gpie = 0;
2864
		}
2865 2866 2867 2868
		if (adapter->flags & IXGBE_FLAG_SRIOV_ENABLED) {
			gpie &= ~IXGBE_GPIE_VTMODE_MASK;
			gpie |= IXGBE_GPIE_VTMODE_64;
		}
2869 2870 2871
		/* XXX: to interrupt immediately for EICS writes, enable this */
		/* gpie |= IXGBE_GPIE_EIMEN; */
		IXGBE_WRITE_REG(hw, IXGBE_GPIE, gpie);
2872 2873
	}

2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889
	if (adapter->flags & IXGBE_FLAG_MSIX_ENABLED) {
		/*
		 * use EIAM to auto-mask when MSI-X interrupt is asserted
		 * this saves a register write for every interrupt
		 */
		switch (hw->mac.type) {
		case ixgbe_mac_82598EB:
			IXGBE_WRITE_REG(hw, IXGBE_EIAM, IXGBE_EICS_RTX_QUEUE);
			break;
		default:
		case ixgbe_mac_82599EB:
			IXGBE_WRITE_REG(hw, IXGBE_EIAM_EX(0), 0xFFFFFFFF);
			IXGBE_WRITE_REG(hw, IXGBE_EIAM_EX(1), 0xFFFFFFFF);
			break;
		}
	} else {
2890 2891 2892 2893
		/* legacy interrupts, use EIAM to auto-mask when reading EICR,
		 * specifically only auto mask tx and rx interrupts */
		IXGBE_WRITE_REG(hw, IXGBE_EIAM, IXGBE_EICS_RTX_QUEUE);
	}
2894

2895 2896 2897 2898 2899 2900 2901
	/* Enable fan failure interrupt if media type is copper */
	if (adapter->flags & IXGBE_FLAG_FAN_FAIL_CAPABLE) {
		gpie = IXGBE_READ_REG(hw, IXGBE_GPIE);
		gpie |= IXGBE_SDP1_GPIEN;
		IXGBE_WRITE_REG(hw, IXGBE_GPIE, gpie);
	}

2902 2903 2904 2905 2906 2907 2908
	if (hw->mac.type == ixgbe_mac_82599EB) {
		gpie = IXGBE_READ_REG(hw, IXGBE_GPIE);
		gpie |= IXGBE_SDP1_GPIEN;
		gpie |= IXGBE_SDP2_GPIEN;
		IXGBE_WRITE_REG(hw, IXGBE_GPIE, gpie);
	}

2909 2910
#ifdef IXGBE_FCOE
	/* adjust max frame to be able to do baby jumbo for FCoE */
2911
	if ((netdev->features & NETIF_F_FCOE_MTU) &&
2912 2913 2914 2915
	    (max_frame < IXGBE_FCOE_JUMBO_FRAME_SIZE))
		max_frame = IXGBE_FCOE_JUMBO_FRAME_SIZE;

#endif /* IXGBE_FCOE */
2916
	mhadd = IXGBE_READ_REG(hw, IXGBE_MHADD);
2917 2918 2919 2920 2921 2922 2923 2924
	if (max_frame != (mhadd >> IXGBE_MHADD_MFS_SHIFT)) {
		mhadd &= ~IXGBE_MHADD_MFS_MASK;
		mhadd |= max_frame << IXGBE_MHADD_MFS_SHIFT;

		IXGBE_WRITE_REG(hw, IXGBE_MHADD, mhadd);
	}

	for (i = 0; i < adapter->num_tx_queues; i++) {
2925
		j = adapter->tx_ring[i]->reg_idx;
2926
		txdctl = IXGBE_READ_REG(hw, IXGBE_TXDCTL(j));
2927 2928
		/* enable WTHRESH=8 descriptors, to encourage burst writeback */
		txdctl |= (8 << 16);
2929 2930 2931 2932 2933 2934 2935 2936 2937 2938
		IXGBE_WRITE_REG(hw, IXGBE_TXDCTL(j), txdctl);
	}

	if (hw->mac.type == ixgbe_mac_82599EB) {
		/* DMATXCTL.EN must be set after all Tx queue config is done */
		dmatxctl = IXGBE_READ_REG(hw, IXGBE_DMATXCTL);
		dmatxctl |= IXGBE_DMATXCTL_TE;
		IXGBE_WRITE_REG(hw, IXGBE_DMATXCTL, dmatxctl);
	}
	for (i = 0; i < adapter->num_tx_queues; i++) {
2939
		j = adapter->tx_ring[i]->reg_idx;
2940
		txdctl = IXGBE_READ_REG(hw, IXGBE_TXDCTL(j));
2941
		txdctl |= IXGBE_TXDCTL_ENABLE;
2942
		IXGBE_WRITE_REG(hw, IXGBE_TXDCTL(j), txdctl);
2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954
		if (hw->mac.type == ixgbe_mac_82599EB) {
			int wait_loop = 10;
			/* poll for Tx Enable ready */
			do {
				msleep(1);
				txdctl = IXGBE_READ_REG(hw, IXGBE_TXDCTL(j));
			} while (--wait_loop &&
			         !(txdctl & IXGBE_TXDCTL_ENABLE));
			if (!wait_loop)
				DPRINTK(DRV, ERR, "Could not enable "
				        "Tx Queue %d\n", j);
		}
2955 2956
	}

2957
	for (i = 0; i < num_rx_rings; i++) {
2958
		j = adapter->rx_ring[i]->reg_idx;
2959 2960 2961 2962 2963
		rxdctl = IXGBE_READ_REG(hw, IXGBE_RXDCTL(j));
		/* enable PTHRESH=32 descriptors (half the internal cache)
		 * and HTHRESH=0 descriptors (to minimize latency on fetch),
		 * this also removes a pesky rx_no_buffer_count increment */
		rxdctl |= 0x0020;
2964
		rxdctl |= IXGBE_RXDCTL_ENABLE;
2965
		IXGBE_WRITE_REG(hw, IXGBE_RXDCTL(j), rxdctl);
2966 2967
		if (hw->mac.type == ixgbe_mac_82599EB)
			ixgbe_rx_desc_queue_enable(adapter, i);
2968 2969 2970
	}
	/* enable all receives */
	rxdctl = IXGBE_READ_REG(hw, IXGBE_RXCTRL);
2971 2972 2973 2974 2975
	if (hw->mac.type == ixgbe_mac_82598EB)
		rxdctl |= (IXGBE_RXCTRL_DMBYPS | IXGBE_RXCTRL_RXEN);
	else
		rxdctl |= IXGBE_RXCTRL_RXEN;
	hw->mac.ops.enable_rx_dma(hw, rxdctl);
2976 2977 2978 2979 2980 2981 2982

	if (adapter->flags & IXGBE_FLAG_MSIX_ENABLED)
		ixgbe_configure_msix(adapter);
	else
		ixgbe_configure_msi_and_legacy(adapter);

	clear_bit(__IXGBE_DOWN, &adapter->state);
2983 2984 2985 2986 2987
	ixgbe_napi_enable_all(adapter);

	/* clear any pending interrupts, may auto mask */
	IXGBE_READ_REG(hw, IXGBE_EICR);

2988 2989
	ixgbe_irq_enable(adapter);

2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000
	/*
	 * If this adapter has a fan, check to see if we had a failure
	 * before we enabled the interrupt.
	 */
	if (adapter->flags & IXGBE_FLAG_FAN_FAIL_CAPABLE) {
		u32 esdp = IXGBE_READ_REG(hw, IXGBE_ESDP);
		if (esdp & IXGBE_ESDP_SDP1)
			DPRINTK(DRV, CRIT,
				"Fan has stopped, replace the adapter\n");
	}

3001 3002
	/*
	 * For hot-pluggable SFP+ devices, a new SFP+ module may have
3003 3004 3005
	 * arrived before interrupts were enabled but after probe.  Such
	 * devices wouldn't have their type identified yet. We need to
	 * kick off the SFP+ module setup first, then try to bring up link.
3006 3007 3008
	 * If we're not hot-pluggable SFP+, we just need to configure link
	 * and bring it up.
	 */
3009 3010 3011
	if (hw->phy.type == ixgbe_phy_unknown) {
		err = hw->phy.ops.identify(hw);
		if (err == IXGBE_ERR_SFP_NOT_SUPPORTED) {
3012 3013 3014 3015
			/*
			 * Take the device down and schedule the sfp tasklet
			 * which will unregister_netdev and log it.
			 */
3016
			ixgbe_down(adapter);
3017
			schedule_work(&adapter->sfp_config_module_task);
3018 3019
			return err;
		}
3020 3021 3022 3023 3024 3025 3026 3027 3028
	}

	if (ixgbe_is_sfp(hw)) {
		ixgbe_sfp_link_config(adapter);
	} else {
		err = ixgbe_non_sfp_link_config(hw);
		if (err)
			DPRINTK(PROBE, ERR, "link_config FAILED %d\n", err);
	}
3029

3030 3031
	for (i = 0; i < adapter->num_tx_queues; i++)
		set_bit(__IXGBE_FDIR_INIT_DONE,
3032
		        &(adapter->tx_ring[i]->reinit_state));
3033

3034 3035 3036
	/* enable transmits */
	netif_tx_start_all_queues(netdev);

3037 3038
	/* bring the link up in the watchdog, this could race with our first
	 * link up interrupt but shouldn't be a problem */
3039 3040
	adapter->flags |= IXGBE_FLAG_NEED_LINK_UPDATE;
	adapter->link_check_timeout = jiffies;
3041
	mod_timer(&adapter->watchdog_timer, jiffies);
3042 3043 3044 3045 3046 3047

	/* Set PF Reset Done bit so PF/VF Mail Ops can work */
	ctrl_ext = IXGBE_READ_REG(hw, IXGBE_CTRL_EXT);
	ctrl_ext |= IXGBE_CTRL_EXT_PFRSTD;
	IXGBE_WRITE_REG(hw, IXGBE_CTRL_EXT, ctrl_ext);

3048 3049 3050
	return 0;
}

3051 3052 3053 3054 3055 3056 3057 3058 3059 3060
void ixgbe_reinit_locked(struct ixgbe_adapter *adapter)
{
	WARN_ON(in_interrupt());
	while (test_and_set_bit(__IXGBE_RESETTING, &adapter->state))
		msleep(1);
	ixgbe_down(adapter);
	ixgbe_up(adapter);
	clear_bit(__IXGBE_RESETTING, &adapter->state);
}

3061 3062 3063 3064 3065 3066 3067 3068 3069 3070
int ixgbe_up(struct ixgbe_adapter *adapter)
{
	/* hardware has been reset, we need to reload some things */
	ixgbe_configure(adapter);

	return ixgbe_up_complete(adapter);
}

void ixgbe_reset(struct ixgbe_adapter *adapter)
{
3071
	struct ixgbe_hw *hw = &adapter->hw;
3072 3073 3074
	int err;

	err = hw->mac.ops.init_hw(hw);
3075 3076 3077 3078 3079 3080 3081
	switch (err) {
	case 0:
	case IXGBE_ERR_SFP_NOT_PRESENT:
		break;
	case IXGBE_ERR_MASTER_REQUESTS_PENDING:
		dev_err(&adapter->pdev->dev, "master disable timed out\n");
		break;
3082 3083 3084 3085 3086 3087 3088 3089 3090
	case IXGBE_ERR_EEPROM_VERSION:
		/* We are running on a pre-production device, log a warning */
		dev_warn(&adapter->pdev->dev, "This device is a pre-production "
		         "adapter/LOM.  Please be aware there may be issues "
		         "associated with your hardware.  If you are "
		         "experiencing problems please contact your Intel or "
		         "hardware representative who provided you with this "
		         "hardware.\n");
		break;
3091 3092 3093
	default:
		dev_err(&adapter->pdev->dev, "Hardware Error: %d\n", err);
	}
3094 3095

	/* reprogram the RAR[0] in case user changed it. */
3096 3097
	hw->mac.ops.set_rar(hw, 0, hw->mac.addr, adapter->num_vfs,
			    IXGBE_RAH_AV);
3098 3099 3100 3101 3102 3103 3104 3105
}

/**
 * ixgbe_clean_rx_ring - Free Rx Buffers per Queue
 * @adapter: board private structure
 * @rx_ring: ring to free buffers from
 **/
static void ixgbe_clean_rx_ring(struct ixgbe_adapter *adapter,
3106
                                struct ixgbe_ring *rx_ring)
3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 3118 3119
{
	struct pci_dev *pdev = adapter->pdev;
	unsigned long size;
	unsigned int i;

	/* Free all the Rx ring sk_buffs */

	for (i = 0; i < rx_ring->count; i++) {
		struct ixgbe_rx_buffer *rx_buffer_info;

		rx_buffer_info = &rx_ring->rx_buffer_info[i];
		if (rx_buffer_info->dma) {
			pci_unmap_single(pdev, rx_buffer_info->dma,
3120 3121
			                 rx_ring->rx_buf_len,
			                 PCI_DMA_FROMDEVICE);
3122 3123 3124
			rx_buffer_info->dma = 0;
		}
		if (rx_buffer_info->skb) {
A
Alexander Duyck 已提交
3125
			struct sk_buff *skb = rx_buffer_info->skb;
3126
			rx_buffer_info->skb = NULL;
A
Alexander Duyck 已提交
3127 3128
			do {
				struct sk_buff *this = skb;
3129 3130 3131 3132
				if (IXGBE_RSC_CB(this)->dma)
					pci_unmap_single(pdev, IXGBE_RSC_CB(this)->dma,
					                 rx_ring->rx_buf_len,
					                 PCI_DMA_FROMDEVICE);
A
Alexander Duyck 已提交
3133 3134 3135
				skb = skb->prev;
				dev_kfree_skb(this);
			} while (skb);
3136 3137 3138
		}
		if (!rx_buffer_info->page)
			continue;
J
Jesse Brandeburg 已提交
3139 3140 3141 3142 3143
		if (rx_buffer_info->page_dma) {
			pci_unmap_page(pdev, rx_buffer_info->page_dma,
			               PAGE_SIZE / 2, PCI_DMA_FROMDEVICE);
			rx_buffer_info->page_dma = 0;
		}
3144 3145
		put_page(rx_buffer_info->page);
		rx_buffer_info->page = NULL;
3146
		rx_buffer_info->page_offset = 0;
3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157
	}

	size = sizeof(struct ixgbe_rx_buffer) * rx_ring->count;
	memset(rx_ring->rx_buffer_info, 0, size);

	/* Zero out the descriptor ring */
	memset(rx_ring->desc, 0, rx_ring->size);

	rx_ring->next_to_clean = 0;
	rx_ring->next_to_use = 0;

3158 3159 3160 3161
	if (rx_ring->head)
		writel(0, adapter->hw.hw_addr + rx_ring->head);
	if (rx_ring->tail)
		writel(0, adapter->hw.hw_addr + rx_ring->tail);
3162 3163 3164 3165 3166 3167 3168 3169
}

/**
 * ixgbe_clean_tx_ring - Free Tx Buffers
 * @adapter: board private structure
 * @tx_ring: ring to be cleaned
 **/
static void ixgbe_clean_tx_ring(struct ixgbe_adapter *adapter,
3170
                                struct ixgbe_ring *tx_ring)
3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191
{
	struct ixgbe_tx_buffer *tx_buffer_info;
	unsigned long size;
	unsigned int i;

	/* Free all the Tx ring sk_buffs */

	for (i = 0; i < tx_ring->count; i++) {
		tx_buffer_info = &tx_ring->tx_buffer_info[i];
		ixgbe_unmap_and_free_tx_resource(adapter, tx_buffer_info);
	}

	size = sizeof(struct ixgbe_tx_buffer) * tx_ring->count;
	memset(tx_ring->tx_buffer_info, 0, size);

	/* Zero out the descriptor ring */
	memset(tx_ring->desc, 0, tx_ring->size);

	tx_ring->next_to_use = 0;
	tx_ring->next_to_clean = 0;

3192 3193 3194 3195
	if (tx_ring->head)
		writel(0, adapter->hw.hw_addr + tx_ring->head);
	if (tx_ring->tail)
		writel(0, adapter->hw.hw_addr + tx_ring->tail);
3196 3197 3198
}

/**
3199
 * ixgbe_clean_all_rx_rings - Free Rx Buffers for all queues
3200 3201
 * @adapter: board private structure
 **/
3202
static void ixgbe_clean_all_rx_rings(struct ixgbe_adapter *adapter)
3203 3204 3205
{
	int i;

3206
	for (i = 0; i < adapter->num_rx_queues; i++)
3207
		ixgbe_clean_rx_ring(adapter, adapter->rx_ring[i]);
3208 3209 3210
}

/**
3211
 * ixgbe_clean_all_tx_rings - Free Tx Buffers for all queues
3212 3213
 * @adapter: board private structure
 **/
3214
static void ixgbe_clean_all_tx_rings(struct ixgbe_adapter *adapter)
3215 3216 3217
{
	int i;

3218
	for (i = 0; i < adapter->num_tx_queues; i++)
3219
		ixgbe_clean_tx_ring(adapter, adapter->tx_ring[i]);
3220 3221 3222 3223 3224
}

void ixgbe_down(struct ixgbe_adapter *adapter)
{
	struct net_device *netdev = adapter->netdev;
3225
	struct ixgbe_hw *hw = &adapter->hw;
3226
	u32 rxctrl;
3227 3228
	u32 txdctl;
	int i, j;
3229 3230 3231 3232

	/* signal that we are down to the interrupt handler */
	set_bit(__IXGBE_DOWN, &adapter->state);

3233 3234 3235 3236 3237 3238 3239 3240 3241 3242 3243
	/* disable receive for all VFs and wait one second */
	if (adapter->num_vfs) {
		for (i = 0 ; i < adapter->num_vfs; i++)
			adapter->vfinfo[i].clear_to_send = 0;

		/* ping all the active vfs to let them know we are going down */
		ixgbe_ping_all_vfs(adapter);
		/* Disable all VFTE/VFRE TX/RX */
		ixgbe_disable_tx_rx(adapter);
	}

3244
	/* disable receives */
3245 3246
	rxctrl = IXGBE_READ_REG(hw, IXGBE_RXCTRL);
	IXGBE_WRITE_REG(hw, IXGBE_RXCTRL, rxctrl & ~IXGBE_RXCTRL_RXEN);
3247 3248 3249

	netif_tx_disable(netdev);

3250
	IXGBE_WRITE_FLUSH(hw);
3251 3252
	msleep(10);

3253 3254
	netif_tx_stop_all_queues(netdev);

3255 3256
	ixgbe_irq_disable(adapter);

3257
	ixgbe_napi_disable_all(adapter);
3258

3259 3260
	clear_bit(__IXGBE_SFP_MODULE_NOT_FOUND, &adapter->state);
	del_timer_sync(&adapter->sfp_timer);
3261
	del_timer_sync(&adapter->watchdog_timer);
3262
	cancel_work_sync(&adapter->watchdog_task);
3263

3264 3265 3266 3267
	if (adapter->flags & IXGBE_FLAG_FDIR_HASH_CAPABLE ||
	    adapter->flags & IXGBE_FLAG_FDIR_PERFECT_CAPABLE)
		cancel_work_sync(&adapter->fdir_reinit_task);

3268 3269
	/* disable transmits in the hardware now that interrupts are off */
	for (i = 0; i < adapter->num_tx_queues; i++) {
3270
		j = adapter->tx_ring[i]->reg_idx;
3271 3272 3273 3274
		txdctl = IXGBE_READ_REG(hw, IXGBE_TXDCTL(j));
		IXGBE_WRITE_REG(hw, IXGBE_TXDCTL(j),
		                (txdctl & ~IXGBE_TXDCTL_ENABLE));
	}
3275 3276 3277 3278 3279
	/* Disable the Tx DMA engine on 82599 */
	if (hw->mac.type == ixgbe_mac_82599EB)
		IXGBE_WRITE_REG(hw, IXGBE_DMATXCTL,
		                (IXGBE_READ_REG(hw, IXGBE_DMATXCTL) &
		                 ~IXGBE_DMATXCTL_TE));
3280

3281 3282
	netif_carrier_off(netdev);

3283 3284 3285
	/* clear n-tuple filters that are cached */
	ethtool_ntuple_flush(netdev);

3286 3287
	if (!pci_channel_offline(adapter->pdev))
		ixgbe_reset(adapter);
3288 3289 3290
	ixgbe_clean_all_tx_rings(adapter);
	ixgbe_clean_all_rx_rings(adapter);

3291
#ifdef CONFIG_IXGBE_DCA
3292
	/* since we reset the hardware DCA settings were cleared */
3293
	ixgbe_setup_dca(adapter);
3294
#endif
3295 3296 3297
}

/**
3298 3299 3300 3301 3302
 * ixgbe_poll - NAPI Rx polling callback
 * @napi: structure for representing this polling device
 * @budget: how many packets driver is allowed to clean
 *
 * This function is used for legacy and MSI, NAPI mode
3303
 **/
3304
static int ixgbe_poll(struct napi_struct *napi, int budget)
3305
{
3306 3307
	struct ixgbe_q_vector *q_vector =
	                        container_of(napi, struct ixgbe_q_vector, napi);
3308
	struct ixgbe_adapter *adapter = q_vector->adapter;
3309
	int tx_clean_complete, work_done = 0;
3310

3311
#ifdef CONFIG_IXGBE_DCA
3312
	if (adapter->flags & IXGBE_FLAG_DCA_ENABLED) {
3313 3314
		ixgbe_update_tx_dca(adapter, adapter->tx_ring[0]);
		ixgbe_update_rx_dca(adapter, adapter->rx_ring[0]);
3315 3316 3317
	}
#endif

3318 3319
	tx_clean_complete = ixgbe_clean_tx_irq(q_vector, adapter->tx_ring[0]);
	ixgbe_clean_rx_irq(q_vector, adapter->rx_ring[0], &work_done, budget);
3320

3321
	if (!tx_clean_complete)
3322 3323
		work_done = budget;

3324 3325
	/* If budget not fully consumed, exit the polling mode */
	if (work_done < budget) {
3326
		napi_complete(napi);
3327
		if (adapter->rx_itr_setting & 1)
3328
			ixgbe_set_itr(adapter);
3329
		if (!test_bit(__IXGBE_DOWN, &adapter->state))
3330
			ixgbe_irq_enable_queues(adapter, IXGBE_EIMS_RTX_QUEUE);
3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351
	}
	return work_done;
}

/**
 * ixgbe_tx_timeout - Respond to a Tx Hang
 * @netdev: network interface device structure
 **/
static void ixgbe_tx_timeout(struct net_device *netdev)
{
	struct ixgbe_adapter *adapter = netdev_priv(netdev);

	/* Do the reset outside of interrupt context */
	schedule_work(&adapter->reset_task);
}

static void ixgbe_reset_task(struct work_struct *work)
{
	struct ixgbe_adapter *adapter;
	adapter = container_of(work, struct ixgbe_adapter, reset_task);

3352 3353 3354 3355 3356
	/* If we're already down or resetting, just bail */
	if (test_bit(__IXGBE_DOWN, &adapter->state) ||
	    test_bit(__IXGBE_RESETTING, &adapter->state))
		return;

3357 3358
	adapter->tx_timeout_count++;

3359
	ixgbe_reinit_locked(adapter);
3360 3361
}

3362 3363
#ifdef CONFIG_IXGBE_DCB
static inline bool ixgbe_set_dcb_queues(struct ixgbe_adapter *adapter)
3364
{
3365
	bool ret = false;
3366
	struct ixgbe_ring_feature *f = &adapter->ring_feature[RING_F_DCB];
3367

3368 3369 3370 3371 3372 3373 3374
	if (!(adapter->flags & IXGBE_FLAG_DCB_ENABLED))
		return ret;

	f->mask = 0x7 << 3;
	adapter->num_rx_queues = f->indices;
	adapter->num_tx_queues = f->indices;
	ret = true;
3375

3376 3377 3378 3379
	return ret;
}
#endif

3380 3381 3382 3383 3384 3385 3386 3387
/**
 * ixgbe_set_rss_queues: Allocate queues for RSS
 * @adapter: board private structure to initialize
 *
 * This is our "base" multiqueue mode.  RSS (Receive Side Scaling) will try
 * to allocate one Rx queue per CPU, and if available, one Tx queue per CPU.
 *
 **/
3388 3389 3390
static inline bool ixgbe_set_rss_queues(struct ixgbe_adapter *adapter)
{
	bool ret = false;
3391
	struct ixgbe_ring_feature *f = &adapter->ring_feature[RING_F_RSS];
3392 3393

	if (adapter->flags & IXGBE_FLAG_RSS_ENABLED) {
3394 3395 3396
		f->mask = 0xF;
		adapter->num_rx_queues = f->indices;
		adapter->num_tx_queues = f->indices;
3397 3398 3399
		ret = true;
	} else {
		ret = false;
3400 3401
	}

3402 3403 3404
	return ret;
}

3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436
/**
 * ixgbe_set_fdir_queues: Allocate queues for Flow Director
 * @adapter: board private structure to initialize
 *
 * Flow Director is an advanced Rx filter, attempting to get Rx flows back
 * to the original CPU that initiated the Tx session.  This runs in addition
 * to RSS, so if a packet doesn't match an FDIR filter, we can still spread the
 * Rx load across CPUs using RSS.
 *
 **/
static bool inline ixgbe_set_fdir_queues(struct ixgbe_adapter *adapter)
{
	bool ret = false;
	struct ixgbe_ring_feature *f_fdir = &adapter->ring_feature[RING_F_FDIR];

	f_fdir->indices = min((int)num_online_cpus(), f_fdir->indices);
	f_fdir->mask = 0;

	/* Flow Director must have RSS enabled */
	if (adapter->flags & IXGBE_FLAG_RSS_ENABLED &&
	    ((adapter->flags & IXGBE_FLAG_FDIR_HASH_CAPABLE ||
	     (adapter->flags & IXGBE_FLAG_FDIR_PERFECT_CAPABLE)))) {
		adapter->num_tx_queues = f_fdir->indices;
		adapter->num_rx_queues = f_fdir->indices;
		ret = true;
	} else {
		adapter->flags &= ~IXGBE_FLAG_FDIR_HASH_CAPABLE;
		adapter->flags &= ~IXGBE_FLAG_FDIR_PERFECT_CAPABLE;
	}
	return ret;
}

3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448 3449 3450 3451 3452 3453 3454
#ifdef IXGBE_FCOE
/**
 * ixgbe_set_fcoe_queues: Allocate queues for Fiber Channel over Ethernet (FCoE)
 * @adapter: board private structure to initialize
 *
 * FCoE RX FCRETA can use up to 8 rx queues for up to 8 different exchanges.
 * The ring feature mask is not used as a mask for FCoE, as it can take any 8
 * rx queues out of the max number of rx queues, instead, it is used as the
 * index of the first rx queue used by FCoE.
 *
 **/
static inline bool ixgbe_set_fcoe_queues(struct ixgbe_adapter *adapter)
{
	bool ret = false;
	struct ixgbe_ring_feature *f = &adapter->ring_feature[RING_F_FCOE];

	f->indices = min((int)num_online_cpus(), f->indices);
	if (adapter->flags & IXGBE_FLAG_FCOE_ENABLED) {
3455 3456
		adapter->num_rx_queues = 1;
		adapter->num_tx_queues = 1;
3457 3458
#ifdef CONFIG_IXGBE_DCB
		if (adapter->flags & IXGBE_FLAG_DCB_ENABLED) {
3459
			DPRINTK(PROBE, INFO, "FCoE enabled with DCB \n");
3460 3461 3462 3463
			ixgbe_set_dcb_queues(adapter);
		}
#endif
		if (adapter->flags & IXGBE_FLAG_RSS_ENABLED) {
3464
			DPRINTK(PROBE, INFO, "FCoE enabled with RSS \n");
3465 3466 3467 3468 3469
			if ((adapter->flags & IXGBE_FLAG_FDIR_HASH_CAPABLE) ||
			    (adapter->flags & IXGBE_FLAG_FDIR_PERFECT_CAPABLE))
				ixgbe_set_fdir_queues(adapter);
			else
				ixgbe_set_rss_queues(adapter);
3470 3471 3472 3473
		}
		/* adding FCoE rx rings to the end */
		f->mask = adapter->num_rx_queues;
		adapter->num_rx_queues += f->indices;
3474
		adapter->num_tx_queues += f->indices;
3475 3476 3477 3478 3479 3480 3481 3482

		ret = true;
	}

	return ret;
}

#endif /* IXGBE_FCOE */
3483 3484 3485 3486 3487 3488 3489 3490 3491 3492 3493 3494 3495
/**
 * ixgbe_set_sriov_queues: Allocate queues for IOV use
 * @adapter: board private structure to initialize
 *
 * IOV doesn't actually use anything, so just NAK the
 * request for now and let the other queue routines
 * figure out what to do.
 */
static inline bool ixgbe_set_sriov_queues(struct ixgbe_adapter *adapter)
{
	return false;
}

3496 3497 3498 3499 3500 3501 3502 3503 3504 3505 3506
/*
 * ixgbe_set_num_queues: Allocate queues for device, feature dependant
 * @adapter: board private structure to initialize
 *
 * This is the top level queue allocation routine.  The order here is very
 * important, starting with the "most" number of features turned on at once,
 * and ending with the smallest set of features.  This way large combinations
 * can be allocated if they're turned on, and smaller combinations are the
 * fallthrough conditions.
 *
 **/
3507 3508
static void ixgbe_set_num_queues(struct ixgbe_adapter *adapter)
{
3509 3510 3511 3512 3513 3514 3515 3516 3517
	/* Start with base case */
	adapter->num_rx_queues = 1;
	adapter->num_tx_queues = 1;
	adapter->num_rx_pools = adapter->num_rx_queues;
	adapter->num_rx_queues_per_pool = 1;

	if (ixgbe_set_sriov_queues(adapter))
		return;

3518 3519 3520 3521 3522
#ifdef IXGBE_FCOE
	if (ixgbe_set_fcoe_queues(adapter))
		goto done;

#endif /* IXGBE_FCOE */
3523 3524
#ifdef CONFIG_IXGBE_DCB
	if (ixgbe_set_dcb_queues(adapter))
3525
		goto done;
3526 3527

#endif
3528 3529 3530
	if (ixgbe_set_fdir_queues(adapter))
		goto done;

3531
	if (ixgbe_set_rss_queues(adapter))
3532 3533 3534 3535 3536 3537 3538 3539 3540
		goto done;

	/* fallback to base case */
	adapter->num_rx_queues = 1;
	adapter->num_tx_queues = 1;

done:
	/* Notify the stack of the (possibly) reduced Tx Queue count. */
	adapter->netdev->real_num_tx_queues = adapter->num_tx_queues;
3541 3542
}

3543
static void ixgbe_acquire_msix_vectors(struct ixgbe_adapter *adapter,
3544
                                       int vectors)
3545 3546 3547 3548 3549 3550 3551 3552 3553 3554 3555 3556 3557 3558 3559 3560 3561 3562
{
	int err, vector_threshold;

	/* We'll want at least 3 (vector_threshold):
	 * 1) TxQ[0] Cleanup
	 * 2) RxQ[0] Cleanup
	 * 3) Other (Link Status Change, etc.)
	 * 4) TCP Timer (optional)
	 */
	vector_threshold = MIN_MSIX_COUNT;

	/* The more we get, the more we will assign to Tx/Rx Cleanup
	 * for the separate queues...where Rx Cleanup >= Tx Cleanup.
	 * Right now, we simply care about how many we'll get; we'll
	 * set them up later while requesting irq's.
	 */
	while (vectors >= vector_threshold) {
		err = pci_enable_msix(adapter->pdev, adapter->msix_entries,
3563
		                      vectors);
3564 3565 3566 3567 3568 3569 3570 3571 3572 3573 3574 3575 3576 3577 3578 3579 3580 3581 3582
		if (!err) /* Success in acquiring all requested vectors. */
			break;
		else if (err < 0)
			vectors = 0; /* Nasty failure, quit now */
		else /* err == number of vectors we should try again with */
			vectors = err;
	}

	if (vectors < vector_threshold) {
		/* Can't allocate enough MSI-X interrupts?  Oh well.
		 * This just means we'll go with either a single MSI
		 * vector or fall back to legacy interrupts.
		 */
		DPRINTK(HW, DEBUG, "Unable to allocate MSI-X interrupts\n");
		adapter->flags &= ~IXGBE_FLAG_MSIX_ENABLED;
		kfree(adapter->msix_entries);
		adapter->msix_entries = NULL;
	} else {
		adapter->flags |= IXGBE_FLAG_MSIX_ENABLED; /* Woot! */
3583 3584 3585 3586 3587 3588 3589
		/*
		 * Adjust for only the vectors we'll use, which is minimum
		 * of max_msix_q_vectors + NON_Q_VECTORS, or the number of
		 * vectors we were allocated.
		 */
		adapter->num_msix_vectors = min(vectors,
		                   adapter->max_msix_q_vectors + NON_Q_VECTORS);
3590 3591 3592 3593
	}
}

/**
3594
 * ixgbe_cache_ring_rss - Descriptor ring to register mapping for RSS
3595 3596
 * @adapter: board private structure to initialize
 *
3597 3598
 * Cache the descriptor ring offsets for RSS to the assigned rings.
 *
3599
 **/
3600
static inline bool ixgbe_cache_ring_rss(struct ixgbe_adapter *adapter)
3601
{
3602 3603 3604 3605 3606
	int i;
	bool ret = false;

	if (adapter->flags & IXGBE_FLAG_RSS_ENABLED) {
		for (i = 0; i < adapter->num_rx_queues; i++)
3607
			adapter->rx_ring[i]->reg_idx = i;
3608
		for (i = 0; i < adapter->num_tx_queues; i++)
3609
			adapter->tx_ring[i]->reg_idx = i;
3610 3611 3612 3613 3614 3615 3616 3617 3618 3619 3620 3621 3622 3623 3624 3625 3626 3627 3628 3629 3630 3631 3632 3633
		ret = true;
	} else {
		ret = false;
	}

	return ret;
}

#ifdef CONFIG_IXGBE_DCB
/**
 * ixgbe_cache_ring_dcb - Descriptor ring to register mapping for DCB
 * @adapter: board private structure to initialize
 *
 * Cache the descriptor ring offsets for DCB to the assigned rings.
 *
 **/
static inline bool ixgbe_cache_ring_dcb(struct ixgbe_adapter *adapter)
{
	int i;
	bool ret = false;
	int dcb_i = adapter->ring_feature[RING_F_DCB].indices;

	if (adapter->flags & IXGBE_FLAG_DCB_ENABLED) {
		if (adapter->hw.mac.type == ixgbe_mac_82598EB) {
3634 3635
			/* the number of queues is assumed to be symmetric */
			for (i = 0; i < dcb_i; i++) {
3636 3637
				adapter->rx_ring[i]->reg_idx = i << 3;
				adapter->tx_ring[i]->reg_idx = i << 2;
3638
			}
3639
			ret = true;
3640
		} else if (adapter->hw.mac.type == ixgbe_mac_82599EB) {
3641 3642 3643 3644 3645 3646 3647 3648 3649 3650 3651 3652 3653 3654
			if (dcb_i == 8) {
				/*
				 * Tx TC0 starts at: descriptor queue 0
				 * Tx TC1 starts at: descriptor queue 32
				 * Tx TC2 starts at: descriptor queue 64
				 * Tx TC3 starts at: descriptor queue 80
				 * Tx TC4 starts at: descriptor queue 96
				 * Tx TC5 starts at: descriptor queue 104
				 * Tx TC6 starts at: descriptor queue 112
				 * Tx TC7 starts at: descriptor queue 120
				 *
				 * Rx TC0-TC7 are offset by 16 queues each
				 */
				for (i = 0; i < 3; i++) {
3655 3656
					adapter->tx_ring[i]->reg_idx = i << 5;
					adapter->rx_ring[i]->reg_idx = i << 4;
3657 3658
				}
				for ( ; i < 5; i++) {
3659
					adapter->tx_ring[i]->reg_idx =
3660
					                         ((i + 2) << 4);
3661
					adapter->rx_ring[i]->reg_idx = i << 4;
3662 3663
				}
				for ( ; i < dcb_i; i++) {
3664
					adapter->tx_ring[i]->reg_idx =
3665
					                         ((i + 8) << 3);
3666
					adapter->rx_ring[i]->reg_idx = i << 4;
3667 3668 3669 3670 3671 3672 3673 3674 3675 3676 3677 3678
				}

				ret = true;
			} else if (dcb_i == 4) {
				/*
				 * Tx TC0 starts at: descriptor queue 0
				 * Tx TC1 starts at: descriptor queue 64
				 * Tx TC2 starts at: descriptor queue 96
				 * Tx TC3 starts at: descriptor queue 112
				 *
				 * Rx TC0-TC3 are offset by 32 queues each
				 */
3679 3680 3681 3682
				adapter->tx_ring[0]->reg_idx = 0;
				adapter->tx_ring[1]->reg_idx = 64;
				adapter->tx_ring[2]->reg_idx = 96;
				adapter->tx_ring[3]->reg_idx = 112;
3683
				for (i = 0 ; i < dcb_i; i++)
3684
					adapter->rx_ring[i]->reg_idx = i << 5;
3685 3686 3687 3688

				ret = true;
			} else {
				ret = false;
3689
			}
3690 3691
		} else {
			ret = false;
3692
		}
3693 3694
	} else {
		ret = false;
3695
	}
3696 3697 3698 3699 3700

	return ret;
}
#endif

3701 3702 3703 3704 3705 3706 3707 3708 3709 3710 3711 3712 3713 3714 3715 3716
/**
 * ixgbe_cache_ring_fdir - Descriptor ring to register mapping for Flow Director
 * @adapter: board private structure to initialize
 *
 * Cache the descriptor ring offsets for Flow Director to the assigned rings.
 *
 **/
static bool inline ixgbe_cache_ring_fdir(struct ixgbe_adapter *adapter)
{
	int i;
	bool ret = false;

	if (adapter->flags & IXGBE_FLAG_RSS_ENABLED &&
	    ((adapter->flags & IXGBE_FLAG_FDIR_HASH_CAPABLE) ||
	     (adapter->flags & IXGBE_FLAG_FDIR_PERFECT_CAPABLE))) {
		for (i = 0; i < adapter->num_rx_queues; i++)
3717
			adapter->rx_ring[i]->reg_idx = i;
3718
		for (i = 0; i < adapter->num_tx_queues; i++)
3719
			adapter->tx_ring[i]->reg_idx = i;
3720 3721 3722 3723 3724 3725
		ret = true;
	}

	return ret;
}

3726 3727 3728 3729 3730 3731 3732 3733 3734 3735
#ifdef IXGBE_FCOE
/**
 * ixgbe_cache_ring_fcoe - Descriptor ring to register mapping for the FCoE
 * @adapter: board private structure to initialize
 *
 * Cache the descriptor ring offsets for FCoE mode to the assigned rings.
 *
 */
static inline bool ixgbe_cache_ring_fcoe(struct ixgbe_adapter *adapter)
{
3736
	int i, fcoe_rx_i = 0, fcoe_tx_i = 0;
3737 3738 3739 3740 3741 3742
	bool ret = false;
	struct ixgbe_ring_feature *f = &adapter->ring_feature[RING_F_FCOE];

	if (adapter->flags & IXGBE_FLAG_FCOE_ENABLED) {
#ifdef CONFIG_IXGBE_DCB
		if (adapter->flags & IXGBE_FLAG_DCB_ENABLED) {
3743 3744
			struct ixgbe_fcoe *fcoe = &adapter->fcoe;

3745
			ixgbe_cache_ring_dcb(adapter);
3746
			/* find out queues in TC for FCoE */
3747 3748
			fcoe_rx_i = adapter->rx_ring[fcoe->tc]->reg_idx + 1;
			fcoe_tx_i = adapter->tx_ring[fcoe->tc]->reg_idx + 1;
3749 3750 3751 3752 3753 3754 3755 3756 3757 3758 3759 3760 3761 3762 3763 3764 3765
			/*
			 * In 82599, the number of Tx queues for each traffic
			 * class for both 8-TC and 4-TC modes are:
			 * TCs  : TC0 TC1 TC2 TC3 TC4 TC5 TC6 TC7
			 * 8 TCs:  32  32  16  16   8   8   8   8
			 * 4 TCs:  64  64  32  32
			 * We have max 8 queues for FCoE, where 8 the is
			 * FCoE redirection table size. If TC for FCoE is
			 * less than or equal to TC3, we have enough queues
			 * to add max of 8 queues for FCoE, so we start FCoE
			 * tx descriptor from the next one, i.e., reg_idx + 1.
			 * If TC for FCoE is above TC3, implying 8 TC mode,
			 * and we need 8 for FCoE, we have to take all queues
			 * in that traffic class for FCoE.
			 */
			if ((f->indices == IXGBE_FCRETA_SIZE) && (fcoe->tc > 3))
				fcoe_tx_i--;
3766 3767 3768
		}
#endif /* CONFIG_IXGBE_DCB */
		if (adapter->flags & IXGBE_FLAG_RSS_ENABLED) {
3769 3770 3771 3772 3773 3774
			if ((adapter->flags & IXGBE_FLAG_FDIR_HASH_CAPABLE) ||
			    (adapter->flags & IXGBE_FLAG_FDIR_PERFECT_CAPABLE))
				ixgbe_cache_ring_fdir(adapter);
			else
				ixgbe_cache_ring_rss(adapter);

3775 3776 3777 3778
			fcoe_rx_i = f->mask;
			fcoe_tx_i = f->mask;
		}
		for (i = 0; i < f->indices; i++, fcoe_rx_i++, fcoe_tx_i++) {
3779 3780
			adapter->rx_ring[f->mask + i]->reg_idx = fcoe_rx_i;
			adapter->tx_ring[f->mask + i]->reg_idx = fcoe_tx_i;
3781 3782 3783 3784 3785 3786 3787
		}
		ret = true;
	}
	return ret;
}

#endif /* IXGBE_FCOE */
3788 3789 3790 3791 3792 3793 3794 3795 3796 3797
/**
 * ixgbe_cache_ring_sriov - Descriptor ring to register mapping for sriov
 * @adapter: board private structure to initialize
 *
 * SR-IOV doesn't use any descriptor rings but changes the default if
 * no other mapping is used.
 *
 */
static inline bool ixgbe_cache_ring_sriov(struct ixgbe_adapter *adapter)
{
3798 3799
	adapter->rx_ring[0]->reg_idx = adapter->num_vfs * 2;
	adapter->tx_ring[0]->reg_idx = adapter->num_vfs * 2;
3800 3801 3802 3803 3804 3805
	if (adapter->num_vfs)
		return true;
	else
		return false;
}

3806 3807 3808 3809 3810 3811 3812 3813 3814 3815 3816 3817 3818 3819
/**
 * ixgbe_cache_ring_register - Descriptor ring to register mapping
 * @adapter: board private structure to initialize
 *
 * Once we know the feature-set enabled for the device, we'll cache
 * the register offset the descriptor ring is assigned to.
 *
 * Note, the order the various feature calls is important.  It must start with
 * the "most" features enabled at the same time, then trickle down to the
 * least amount of features turned on at once.
 **/
static void ixgbe_cache_ring_register(struct ixgbe_adapter *adapter)
{
	/* start with default case */
3820 3821
	adapter->rx_ring[0]->reg_idx = 0;
	adapter->tx_ring[0]->reg_idx = 0;
3822

3823 3824 3825
	if (ixgbe_cache_ring_sriov(adapter))
		return;

3826 3827 3828 3829 3830
#ifdef IXGBE_FCOE
	if (ixgbe_cache_ring_fcoe(adapter))
		return;

#endif /* IXGBE_FCOE */
3831 3832 3833 3834 3835
#ifdef CONFIG_IXGBE_DCB
	if (ixgbe_cache_ring_dcb(adapter))
		return;

#endif
3836 3837 3838
	if (ixgbe_cache_ring_fdir(adapter))
		return;

3839 3840
	if (ixgbe_cache_ring_rss(adapter))
		return;
3841 3842
}

3843 3844 3845 3846 3847
/**
 * ixgbe_alloc_queues - Allocate memory for all rings
 * @adapter: board private structure to initialize
 *
 * We allocate one ring per queue at run-time since we don't know the
3848 3849
 * number of queues at compile-time.  The polling_netdev array is
 * intended for Multiqueue, but should work fine with a single queue.
3850
 **/
3851
static int ixgbe_alloc_queues(struct ixgbe_adapter *adapter)
3852 3853
{
	int i;
3854
	int orig_node = adapter->node;
3855

3856
	for (i = 0; i < adapter->num_tx_queues; i++) {
3857 3858 3859 3860 3861 3862 3863 3864 3865 3866 3867 3868 3869 3870 3871 3872 3873 3874
		struct ixgbe_ring *ring = adapter->tx_ring[i];
		if (orig_node == -1) {
			int cur_node = next_online_node(adapter->node);
			if (cur_node == MAX_NUMNODES)
				cur_node = first_online_node;
			adapter->node = cur_node;
		}
		ring = kzalloc_node(sizeof(struct ixgbe_ring), GFP_KERNEL,
		                    adapter->node);
		if (!ring)
			ring = kzalloc(sizeof(struct ixgbe_ring), GFP_KERNEL);
		if (!ring)
			goto err_tx_ring_allocation;
		ring->count = adapter->tx_ring_count;
		ring->queue_index = i;
		ring->numa_node = adapter->node;

		adapter->tx_ring[i] = ring;
3875
	}
3876

3877 3878 3879
	/* Restore the adapter's original node */
	adapter->node = orig_node;

3880
	for (i = 0; i < adapter->num_rx_queues; i++) {
3881 3882 3883 3884 3885 3886 3887 3888 3889 3890 3891 3892 3893 3894 3895 3896 3897 3898
		struct ixgbe_ring *ring = adapter->rx_ring[i];
		if (orig_node == -1) {
			int cur_node = next_online_node(adapter->node);
			if (cur_node == MAX_NUMNODES)
				cur_node = first_online_node;
			adapter->node = cur_node;
		}
		ring = kzalloc_node(sizeof(struct ixgbe_ring), GFP_KERNEL,
		                    adapter->node);
		if (!ring)
			ring = kzalloc(sizeof(struct ixgbe_ring), GFP_KERNEL);
		if (!ring)
			goto err_rx_ring_allocation;
		ring->count = adapter->rx_ring_count;
		ring->queue_index = i;
		ring->numa_node = adapter->node;

		adapter->rx_ring[i] = ring;
3899 3900
	}

3901 3902 3903
	/* Restore the adapter's original node */
	adapter->node = orig_node;

3904 3905 3906 3907 3908
	ixgbe_cache_ring_register(adapter);

	return 0;

err_rx_ring_allocation:
3909 3910
	for (i = 0; i < adapter->num_tx_queues; i++)
		kfree(adapter->tx_ring[i]);
3911 3912 3913 3914 3915 3916 3917 3918 3919 3920 3921
err_tx_ring_allocation:
	return -ENOMEM;
}

/**
 * ixgbe_set_interrupt_capability - set MSI-X or MSI if supported
 * @adapter: board private structure to initialize
 *
 * Attempt to configure the interrupts using the best available
 * capabilities of the hardware and the kernel.
 **/
A
Al Viro 已提交
3922
static int ixgbe_set_interrupt_capability(struct ixgbe_adapter *adapter)
3923
{
3924
	struct ixgbe_hw *hw = &adapter->hw;
3925 3926 3927 3928 3929 3930 3931
	int err = 0;
	int vector, v_budget;

	/*
	 * It's easy to be greedy for MSI-X vectors, but it really
	 * doesn't do us much good if we have a lot more vectors
	 * than CPU's.  So let's be conservative and only ask for
3932
	 * (roughly) the same number of vectors as there are CPU's.
3933 3934
	 */
	v_budget = min(adapter->num_rx_queues + adapter->num_tx_queues,
3935
	               (int)num_online_cpus()) + NON_Q_VECTORS;
3936 3937 3938

	/*
	 * At the same time, hardware can only support a maximum of
3939 3940 3941 3942
	 * hw.mac->max_msix_vectors vectors.  With features
	 * such as RSS and VMDq, we can easily surpass the number of Rx and Tx
	 * descriptor queues supported by our device.  Thus, we cap it off in
	 * those rare cases where the cpu count also exceeds our vector limit.
3943
	 */
3944
	v_budget = min(v_budget, (int)hw->mac.max_msix_vectors);
3945 3946 3947 3948

	/* A failure in MSI-X entry allocation isn't fatal, but it does
	 * mean we disable MSI-X capabilities of the adapter. */
	adapter->msix_entries = kcalloc(v_budget,
3949
	                                sizeof(struct msix_entry), GFP_KERNEL);
3950 3951 3952
	if (adapter->msix_entries) {
		for (vector = 0; vector < v_budget; vector++)
			adapter->msix_entries[vector].entry = vector;
3953

3954
		ixgbe_acquire_msix_vectors(adapter, v_budget);
3955

3956 3957 3958
		if (adapter->flags & IXGBE_FLAG_MSIX_ENABLED)
			goto out;
	}
3959

3960 3961
	adapter->flags &= ~IXGBE_FLAG_DCB_ENABLED;
	adapter->flags &= ~IXGBE_FLAG_RSS_ENABLED;
3962 3963 3964
	adapter->flags &= ~IXGBE_FLAG_FDIR_HASH_CAPABLE;
	adapter->flags &= ~IXGBE_FLAG_FDIR_PERFECT_CAPABLE;
	adapter->atr_sample_rate = 0;
3965 3966 3967
	if (adapter->flags & IXGBE_FLAG_SRIOV_ENABLED)
		ixgbe_disable_sriov(adapter);

3968
	ixgbe_set_num_queues(adapter);
3969 3970 3971 3972 3973 3974

	err = pci_enable_msi(adapter->pdev);
	if (!err) {
		adapter->flags |= IXGBE_FLAG_MSI_ENABLED;
	} else {
		DPRINTK(HW, DEBUG, "Unable to allocate MSI interrupt, "
3975
		        "falling back to legacy.  Error: %d\n", err);
3976 3977 3978 3979 3980 3981 3982 3983
		/* reset err */
		err = 0;
	}

out:
	return err;
}

3984 3985 3986 3987 3988 3989 3990 3991 3992 3993 3994 3995 3996 3997 3998 3999 4000
/**
 * ixgbe_alloc_q_vectors - Allocate memory for interrupt vectors
 * @adapter: board private structure to initialize
 *
 * We allocate one q_vector per queue interrupt.  If allocation fails we
 * return -ENOMEM.
 **/
static int ixgbe_alloc_q_vectors(struct ixgbe_adapter *adapter)
{
	int q_idx, num_q_vectors;
	struct ixgbe_q_vector *q_vector;
	int napi_vectors;
	int (*poll)(struct napi_struct *, int);

	if (adapter->flags & IXGBE_FLAG_MSIX_ENABLED) {
		num_q_vectors = adapter->num_msix_vectors - NON_Q_VECTORS;
		napi_vectors = adapter->num_rx_queues;
4001
		poll = &ixgbe_clean_rxtx_many;
4002 4003 4004 4005 4006 4007 4008
	} else {
		num_q_vectors = 1;
		napi_vectors = 1;
		poll = &ixgbe_poll;
	}

	for (q_idx = 0; q_idx < num_q_vectors; q_idx++) {
4009 4010 4011 4012 4013
		q_vector = kzalloc_node(sizeof(struct ixgbe_q_vector),
		                        GFP_KERNEL, adapter->node);
		if (!q_vector)
			q_vector = kzalloc(sizeof(struct ixgbe_q_vector),
			                   GFP_KERNEL);
4014 4015 4016
		if (!q_vector)
			goto err_out;
		q_vector->adapter = adapter;
4017 4018 4019 4020
		if (q_vector->txr_count && !q_vector->rxr_count)
			q_vector->eitr = adapter->tx_eitr_param;
		else
			q_vector->eitr = adapter->rx_eitr_param;
4021
		q_vector->v_idx = q_idx;
4022
		netif_napi_add(adapter->netdev, &q_vector->napi, (*poll), 64);
4023 4024 4025 4026 4027 4028 4029 4030 4031 4032 4033 4034 4035 4036 4037 4038 4039 4040 4041 4042 4043 4044 4045 4046 4047 4048 4049 4050
		adapter->q_vector[q_idx] = q_vector;
	}

	return 0;

err_out:
	while (q_idx) {
		q_idx--;
		q_vector = adapter->q_vector[q_idx];
		netif_napi_del(&q_vector->napi);
		kfree(q_vector);
		adapter->q_vector[q_idx] = NULL;
	}
	return -ENOMEM;
}

/**
 * ixgbe_free_q_vectors - Free memory allocated for interrupt vectors
 * @adapter: board private structure to initialize
 *
 * This function frees the memory allocated to the q_vectors.  In addition if
 * NAPI is enabled it will delete any references to the NAPI struct prior
 * to freeing the q_vector.
 **/
static void ixgbe_free_q_vectors(struct ixgbe_adapter *adapter)
{
	int q_idx, num_q_vectors;

4051
	if (adapter->flags & IXGBE_FLAG_MSIX_ENABLED)
4052
		num_q_vectors = adapter->num_msix_vectors - NON_Q_VECTORS;
4053
	else
4054 4055 4056 4057 4058
		num_q_vectors = 1;

	for (q_idx = 0; q_idx < num_q_vectors; q_idx++) {
		struct ixgbe_q_vector *q_vector = adapter->q_vector[q_idx];
		adapter->q_vector[q_idx] = NULL;
4059
		netif_napi_del(&q_vector->napi);
4060 4061 4062 4063
		kfree(q_vector);
	}
}

4064
static void ixgbe_reset_interrupt_capability(struct ixgbe_adapter *adapter)
4065 4066 4067 4068 4069 4070 4071 4072 4073 4074 4075 4076 4077 4078 4079 4080 4081 4082 4083 4084 4085 4086 4087
{
	if (adapter->flags & IXGBE_FLAG_MSIX_ENABLED) {
		adapter->flags &= ~IXGBE_FLAG_MSIX_ENABLED;
		pci_disable_msix(adapter->pdev);
		kfree(adapter->msix_entries);
		adapter->msix_entries = NULL;
	} else if (adapter->flags & IXGBE_FLAG_MSI_ENABLED) {
		adapter->flags &= ~IXGBE_FLAG_MSI_ENABLED;
		pci_disable_msi(adapter->pdev);
	}
	return;
}

/**
 * ixgbe_init_interrupt_scheme - Determine proper interrupt scheme
 * @adapter: board private structure to initialize
 *
 * We determine which interrupt scheme to use based on...
 * - Kernel support (MSI, MSI-X)
 *   - which can be user-defined (via MODULE_PARAM)
 * - Hardware queue count (num_*_queues)
 *   - defined by miscellaneous hardware support/features (RSS, etc.)
 **/
4088
int ixgbe_init_interrupt_scheme(struct ixgbe_adapter *adapter)
4089 4090 4091 4092 4093 4094 4095 4096 4097 4098
{
	int err;

	/* Number of supported queues */
	ixgbe_set_num_queues(adapter);

	err = ixgbe_set_interrupt_capability(adapter);
	if (err) {
		DPRINTK(PROBE, ERR, "Unable to setup interrupt capabilities\n");
		goto err_set_interrupt;
4099 4100
	}

4101 4102 4103 4104 4105 4106 4107 4108 4109 4110 4111 4112 4113
	err = ixgbe_alloc_q_vectors(adapter);
	if (err) {
		DPRINTK(PROBE, ERR, "Unable to allocate memory for queue "
		        "vectors\n");
		goto err_alloc_q_vectors;
	}

	err = ixgbe_alloc_queues(adapter);
	if (err) {
		DPRINTK(PROBE, ERR, "Unable to allocate memory for queues\n");
		goto err_alloc_queues;
	}

4114
	DPRINTK(DRV, INFO, "Multiqueue %s: Rx Queue count = %u, "
4115 4116 4117
	        "Tx Queue count = %u\n",
	        (adapter->num_rx_queues > 1) ? "Enabled" :
	        "Disabled", adapter->num_rx_queues, adapter->num_tx_queues);
4118 4119 4120

	set_bit(__IXGBE_DOWN, &adapter->state);

4121
	return 0;
4122

4123 4124 4125 4126
err_alloc_queues:
	ixgbe_free_q_vectors(adapter);
err_alloc_q_vectors:
	ixgbe_reset_interrupt_capability(adapter);
4127
err_set_interrupt:
4128 4129 4130 4131 4132 4133 4134 4135 4136 4137 4138 4139
	return err;
}

/**
 * ixgbe_clear_interrupt_scheme - Clear the current interrupt scheme settings
 * @adapter: board private structure to clear interrupt scheme on
 *
 * We go through and clear interrupt specific resources and reset the structure
 * to pre-load conditions
 **/
void ixgbe_clear_interrupt_scheme(struct ixgbe_adapter *adapter)
{
4140 4141 4142 4143 4144 4145 4146 4147 4148 4149
	int i;

	for (i = 0; i < adapter->num_tx_queues; i++) {
		kfree(adapter->tx_ring[i]);
		adapter->tx_ring[i] = NULL;
	}
	for (i = 0; i < adapter->num_rx_queues; i++) {
		kfree(adapter->rx_ring[i]);
		adapter->rx_ring[i] = NULL;
	}
4150 4151 4152

	ixgbe_free_q_vectors(adapter);
	ixgbe_reset_interrupt_capability(adapter);
4153 4154
}

D
Donald Skidmore 已提交
4155 4156 4157 4158 4159 4160 4161 4162
/**
 * ixgbe_sfp_timer - worker thread to find a missing module
 * @data: pointer to our adapter struct
 **/
static void ixgbe_sfp_timer(unsigned long data)
{
	struct ixgbe_adapter *adapter = (struct ixgbe_adapter *)data;

4163 4164
	/*
	 * Do the sfp_timer outside of interrupt context due to the
D
Donald Skidmore 已提交
4165 4166 4167 4168 4169 4170 4171 4172 4173 4174 4175 4176 4177 4178 4179 4180 4181 4182 4183
	 * delays that sfp+ detection requires
	 */
	schedule_work(&adapter->sfp_task);
}

/**
 * ixgbe_sfp_task - worker thread to find a missing module
 * @work: pointer to work_struct containing our data
 **/
static void ixgbe_sfp_task(struct work_struct *work)
{
	struct ixgbe_adapter *adapter = container_of(work,
	                                             struct ixgbe_adapter,
	                                             sfp_task);
	struct ixgbe_hw *hw = &adapter->hw;

	if ((hw->phy.type == ixgbe_phy_nl) &&
	    (hw->phy.sfp_type == ixgbe_sfp_type_not_present)) {
		s32 ret = hw->phy.ops.identify_sfp(hw);
4184
		if (ret == IXGBE_ERR_SFP_NOT_PRESENT)
D
Donald Skidmore 已提交
4185 4186 4187
			goto reschedule;
		ret = hw->phy.ops.reset(hw);
		if (ret == IXGBE_ERR_SFP_NOT_SUPPORTED) {
D
Don Skidmore 已提交
4188 4189 4190 4191 4192
			dev_err(&adapter->pdev->dev, "failed to initialize "
				"because an unsupported SFP+ module type "
				"was detected.\n"
				"Reload the driver after installing a "
				"supported module.\n");
D
Donald Skidmore 已提交
4193 4194 4195 4196 4197 4198 4199 4200 4201 4202 4203 4204 4205 4206 4207
			unregister_netdev(adapter->netdev);
		} else {
			DPRINTK(PROBE, INFO, "detected SFP+: %d\n",
			        hw->phy.sfp_type);
		}
		/* don't need this routine any more */
		clear_bit(__IXGBE_SFP_MODULE_NOT_FOUND, &adapter->state);
	}
	return;
reschedule:
	if (test_bit(__IXGBE_SFP_MODULE_NOT_FOUND, &adapter->state))
		mod_timer(&adapter->sfp_timer,
		          round_jiffies(jiffies + (2 * HZ)));
}

4208 4209 4210 4211 4212 4213 4214 4215 4216 4217 4218 4219
/**
 * ixgbe_sw_init - Initialize general software structures (struct ixgbe_adapter)
 * @adapter: board private structure to initialize
 *
 * ixgbe_sw_init initializes the Adapter private data structure.
 * Fields are initialized based on PCI device information and
 * OS network device settings (MTU size).
 **/
static int __devinit ixgbe_sw_init(struct ixgbe_adapter *adapter)
{
	struct ixgbe_hw *hw = &adapter->hw;
	struct pci_dev *pdev = adapter->pdev;
4220
	struct net_device *dev = adapter->netdev;
4221
	unsigned int rss;
J
Jeff Kirsher 已提交
4222
#ifdef CONFIG_IXGBE_DCB
4223 4224 4225
	int j;
	struct tc_configuration *tc;
#endif
4226

4227 4228 4229 4230 4231 4232 4233 4234
	/* PCI config space info */

	hw->vendor_id = pdev->vendor;
	hw->device_id = pdev->device;
	hw->revision_id = pdev->revision;
	hw->subsystem_vendor_id = pdev->subsystem_vendor;
	hw->subsystem_device_id = pdev->subsystem_device;

4235 4236 4237 4238
	/* Set capability flags */
	rss = min(IXGBE_MAX_RSS_INDICES, (int)num_online_cpus());
	adapter->ring_feature[RING_F_RSS].indices = rss;
	adapter->flags |= IXGBE_FLAG_RSS_ENABLED;
4239
	adapter->ring_feature[RING_F_DCB].indices = IXGBE_MAX_DCB_INDICES;
4240 4241 4242
	if (hw->mac.type == ixgbe_mac_82598EB) {
		if (hw->device_id == IXGBE_DEV_ID_82598AT)
			adapter->flags |= IXGBE_FLAG_FAN_FAIL_CAPABLE;
4243
		adapter->max_msix_q_vectors = MAX_MSIX_Q_VECTORS_82598;
4244
	} else if (hw->mac.type == ixgbe_mac_82599EB) {
4245
		adapter->max_msix_q_vectors = MAX_MSIX_Q_VECTORS_82599;
4246 4247
		adapter->flags2 |= IXGBE_FLAG2_RSC_CAPABLE;
		adapter->flags2 |= IXGBE_FLAG2_RSC_ENABLED;
4248 4249 4250 4251 4252 4253 4254 4255 4256 4257
		if (dev->features & NETIF_F_NTUPLE) {
			/* Flow Director perfect filter enabled */
			adapter->flags |= IXGBE_FLAG_FDIR_PERFECT_CAPABLE;
			adapter->atr_sample_rate = 0;
			spin_lock_init(&adapter->fdir_perfect_lock);
		} else {
			/* Flow Director hash filters enabled */
			adapter->flags |= IXGBE_FLAG_FDIR_HASH_CAPABLE;
			adapter->atr_sample_rate = 20;
		}
4258 4259 4260
		adapter->ring_feature[RING_F_FDIR].indices =
		                                         IXGBE_MAX_FDIR_INDICES;
		adapter->fdir_pballoc = 0;
4261
#ifdef IXGBE_FCOE
4262 4263 4264
		adapter->flags |= IXGBE_FLAG_FCOE_CAPABLE;
		adapter->flags &= ~IXGBE_FLAG_FCOE_ENABLED;
		adapter->ring_feature[RING_F_FCOE].indices = 0;
4265
#ifdef CONFIG_IXGBE_DCB
4266 4267
		/* Default traffic class to use for FCoE */
		adapter->fcoe.tc = IXGBE_FCOE_DEFTC;
4268
#endif
4269
#endif /* IXGBE_FCOE */
A
Alexander Duyck 已提交
4270
	}
4271

J
Jeff Kirsher 已提交
4272
#ifdef CONFIG_IXGBE_DCB
4273 4274 4275 4276 4277 4278 4279 4280 4281 4282 4283 4284
	/* Configure DCB traffic classes */
	for (j = 0; j < MAX_TRAFFIC_CLASS; j++) {
		tc = &adapter->dcb_cfg.tc_config[j];
		tc->path[DCB_TX_CONFIG].bwg_id = 0;
		tc->path[DCB_TX_CONFIG].bwg_percent = 12 + (j & 1);
		tc->path[DCB_RX_CONFIG].bwg_id = 0;
		tc->path[DCB_RX_CONFIG].bwg_percent = 12 + (j & 1);
		tc->dcb_pfc = pfc_disabled;
	}
	adapter->dcb_cfg.bw_percentage[DCB_TX_CONFIG][0] = 100;
	adapter->dcb_cfg.bw_percentage[DCB_RX_CONFIG][0] = 100;
	adapter->dcb_cfg.rx_pba_cfg = pba_equal;
4285
	adapter->dcb_cfg.pfc_mode_enable = false;
4286 4287 4288 4289 4290 4291
	adapter->dcb_cfg.round_robin_enable = false;
	adapter->dcb_set_bitmap = 0x00;
	ixgbe_copy_dcb_cfg(&adapter->dcb_cfg, &adapter->temp_dcb_cfg,
	                   adapter->ring_feature[RING_F_DCB].indices);

#endif
4292 4293

	/* default flow control settings */
4294
	hw->fc.requested_mode = ixgbe_fc_full;
D
Don Skidmore 已提交
4295
	hw->fc.current_mode = ixgbe_fc_full;	/* init for ethtool output */
4296 4297 4298
#ifdef CONFIG_DCB
	adapter->last_lfc_mode = hw->fc.current_mode;
#endif
4299 4300 4301 4302
	hw->fc.high_water = IXGBE_DEFAULT_FCRTH;
	hw->fc.low_water = IXGBE_DEFAULT_FCRTL;
	hw->fc.pause_time = IXGBE_DEFAULT_FCPAUSE;
	hw->fc.send_xon = true;
D
Don Skidmore 已提交
4303
	hw->fc.disable_fc_autoneg = false;
4304

4305
	/* enable itr by default in dynamic mode */
4306 4307 4308 4309
	adapter->rx_itr_setting = 1;
	adapter->rx_eitr_param = 20000;
	adapter->tx_itr_setting = 1;
	adapter->tx_eitr_param = 10000;
4310 4311 4312 4313 4314 4315 4316 4317 4318

	/* set defaults for eitr in MegaBytes */
	adapter->eitr_low = 10;
	adapter->eitr_high = 20;

	/* set default ring sizes */
	adapter->tx_ring_count = IXGBE_DEFAULT_TXD;
	adapter->rx_ring_count = IXGBE_DEFAULT_RXD;

4319
	/* initialize eeprom parameters */
4320
	if (ixgbe_init_eeprom_params_generic(hw)) {
4321 4322 4323 4324
		dev_err(&pdev->dev, "EEPROM initialization failed\n");
		return -EIO;
	}

4325
	/* enable rx csum by default */
4326 4327
	adapter->flags |= IXGBE_FLAG_RX_CSUM_ENABLED;

4328 4329 4330
	/* get assigned NUMA node */
	adapter->node = dev_to_node(&pdev->dev);

4331 4332 4333 4334 4335 4336 4337 4338
	set_bit(__IXGBE_DOWN, &adapter->state);

	return 0;
}

/**
 * ixgbe_setup_tx_resources - allocate Tx resources (Descriptors)
 * @adapter: board private structure
4339
 * @tx_ring:    tx descriptor ring (for a specific queue) to setup
4340 4341 4342 4343
 *
 * Return 0 on success, negative on failure
 **/
int ixgbe_setup_tx_resources(struct ixgbe_adapter *adapter,
4344
                             struct ixgbe_ring *tx_ring)
4345 4346 4347 4348
{
	struct pci_dev *pdev = adapter->pdev;
	int size;

4349
	size = sizeof(struct ixgbe_tx_buffer) * tx_ring->count;
4350
	tx_ring->tx_buffer_info = vmalloc_node(size, tx_ring->numa_node);
4351 4352
	if (!tx_ring->tx_buffer_info)
		tx_ring->tx_buffer_info = vmalloc(size);
4353 4354
	if (!tx_ring->tx_buffer_info)
		goto err;
4355
	memset(tx_ring->tx_buffer_info, 0, size);
4356 4357

	/* round up to nearest 4K */
4358
	tx_ring->size = tx_ring->count * sizeof(union ixgbe_adv_tx_desc);
4359
	tx_ring->size = ALIGN(tx_ring->size, 4096);
4360

4361 4362
	tx_ring->desc = pci_alloc_consistent(pdev, tx_ring->size,
	                                     &tx_ring->dma);
4363 4364
	if (!tx_ring->desc)
		goto err;
4365

4366 4367 4368
	tx_ring->next_to_use = 0;
	tx_ring->next_to_clean = 0;
	tx_ring->work_limit = tx_ring->count;
4369
	return 0;
4370 4371 4372 4373 4374 4375 4376

err:
	vfree(tx_ring->tx_buffer_info);
	tx_ring->tx_buffer_info = NULL;
	DPRINTK(PROBE, ERR, "Unable to allocate memory for the transmit "
	                    "descriptor ring\n");
	return -ENOMEM;
4377 4378
}

4379 4380 4381 4382 4383 4384 4385 4386 4387 4388 4389 4390 4391 4392 4393
/**
 * ixgbe_setup_all_tx_resources - allocate all queues Tx resources
 * @adapter: board private structure
 *
 * If this function returns with an error, then it's possible one or
 * more of the rings is populated (while the rest are not).  It is the
 * callers duty to clean those orphaned rings.
 *
 * Return 0 on success, negative on failure
 **/
static int ixgbe_setup_all_tx_resources(struct ixgbe_adapter *adapter)
{
	int i, err = 0;

	for (i = 0; i < adapter->num_tx_queues; i++) {
4394
		err = ixgbe_setup_tx_resources(adapter, adapter->tx_ring[i]);
4395 4396 4397 4398 4399 4400 4401 4402 4403
		if (!err)
			continue;
		DPRINTK(PROBE, ERR, "Allocation for Tx Queue %u failed\n", i);
		break;
	}

	return err;
}

4404 4405 4406
/**
 * ixgbe_setup_rx_resources - allocate Rx resources (Descriptors)
 * @adapter: board private structure
4407
 * @rx_ring:    rx descriptor ring (for a specific queue) to setup
4408 4409 4410 4411
 *
 * Returns 0 on success, negative on failure
 **/
int ixgbe_setup_rx_resources(struct ixgbe_adapter *adapter,
4412
                             struct ixgbe_ring *rx_ring)
4413 4414
{
	struct pci_dev *pdev = adapter->pdev;
4415
	int size;
4416

4417
	size = sizeof(struct ixgbe_rx_buffer) * rx_ring->count;
4418 4419 4420
	rx_ring->rx_buffer_info = vmalloc_node(size, adapter->node);
	if (!rx_ring->rx_buffer_info)
		rx_ring->rx_buffer_info = vmalloc(size);
4421
	if (!rx_ring->rx_buffer_info) {
4422
		DPRINTK(PROBE, ERR,
4423
		        "vmalloc allocation failed for the rx desc ring\n");
4424
		goto alloc_failed;
4425
	}
4426
	memset(rx_ring->rx_buffer_info, 0, size);
4427 4428

	/* Round up to nearest 4K */
4429 4430
	rx_ring->size = rx_ring->count * sizeof(union ixgbe_adv_rx_desc);
	rx_ring->size = ALIGN(rx_ring->size, 4096);
4431

4432
	rx_ring->desc = pci_alloc_consistent(pdev, rx_ring->size, &rx_ring->dma);
4433

4434
	if (!rx_ring->desc) {
4435
		DPRINTK(PROBE, ERR,
4436
		        "Memory allocation failed for the rx desc ring\n");
4437
		vfree(rx_ring->rx_buffer_info);
4438
		goto alloc_failed;
4439 4440
	}

4441 4442
	rx_ring->next_to_clean = 0;
	rx_ring->next_to_use = 0;
4443 4444

	return 0;
4445 4446 4447

alloc_failed:
	return -ENOMEM;
4448 4449
}

4450 4451 4452 4453 4454 4455 4456 4457 4458 4459 4460 4461 4462 4463 4464 4465
/**
 * ixgbe_setup_all_rx_resources - allocate all queues Rx resources
 * @adapter: board private structure
 *
 * If this function returns with an error, then it's possible one or
 * more of the rings is populated (while the rest are not).  It is the
 * callers duty to clean those orphaned rings.
 *
 * Return 0 on success, negative on failure
 **/

static int ixgbe_setup_all_rx_resources(struct ixgbe_adapter *adapter)
{
	int i, err = 0;

	for (i = 0; i < adapter->num_rx_queues; i++) {
4466
		err = ixgbe_setup_rx_resources(adapter, adapter->rx_ring[i]);
4467 4468 4469 4470 4471 4472 4473 4474 4475
		if (!err)
			continue;
		DPRINTK(PROBE, ERR, "Allocation for Rx Queue %u failed\n", i);
		break;
	}

	return err;
}

4476 4477 4478 4479 4480 4481 4482
/**
 * ixgbe_free_tx_resources - Free Tx Resources per Queue
 * @adapter: board private structure
 * @tx_ring: Tx descriptor ring for a specific queue
 *
 * Free all transmit software resources
 **/
4483 4484
void ixgbe_free_tx_resources(struct ixgbe_adapter *adapter,
                             struct ixgbe_ring *tx_ring)
4485 4486 4487 4488 4489 4490 4491 4492 4493 4494 4495 4496 4497 4498 4499 4500 4501 4502 4503 4504 4505 4506 4507 4508
{
	struct pci_dev *pdev = adapter->pdev;

	ixgbe_clean_tx_ring(adapter, tx_ring);

	vfree(tx_ring->tx_buffer_info);
	tx_ring->tx_buffer_info = NULL;

	pci_free_consistent(pdev, tx_ring->size, tx_ring->desc, tx_ring->dma);

	tx_ring->desc = NULL;
}

/**
 * ixgbe_free_all_tx_resources - Free Tx Resources for All Queues
 * @adapter: board private structure
 *
 * Free all transmit software resources
 **/
static void ixgbe_free_all_tx_resources(struct ixgbe_adapter *adapter)
{
	int i;

	for (i = 0; i < adapter->num_tx_queues; i++)
4509 4510
		if (adapter->tx_ring[i]->desc)
			ixgbe_free_tx_resources(adapter, adapter->tx_ring[i]);
4511 4512 4513
}

/**
4514
 * ixgbe_free_rx_resources - Free Rx Resources
4515 4516 4517 4518 4519
 * @adapter: board private structure
 * @rx_ring: ring to clean the resources from
 *
 * Free all receive software resources
 **/
4520 4521
void ixgbe_free_rx_resources(struct ixgbe_adapter *adapter,
                             struct ixgbe_ring *rx_ring)
4522 4523 4524 4525 4526 4527 4528 4529 4530 4531 4532 4533 4534 4535 4536 4537 4538 4539 4540 4541 4542 4543 4544 4545
{
	struct pci_dev *pdev = adapter->pdev;

	ixgbe_clean_rx_ring(adapter, rx_ring);

	vfree(rx_ring->rx_buffer_info);
	rx_ring->rx_buffer_info = NULL;

	pci_free_consistent(pdev, rx_ring->size, rx_ring->desc, rx_ring->dma);

	rx_ring->desc = NULL;
}

/**
 * ixgbe_free_all_rx_resources - Free Rx Resources for All Queues
 * @adapter: board private structure
 *
 * Free all receive software resources
 **/
static void ixgbe_free_all_rx_resources(struct ixgbe_adapter *adapter)
{
	int i;

	for (i = 0; i < adapter->num_rx_queues; i++)
4546 4547
		if (adapter->rx_ring[i]->desc)
			ixgbe_free_rx_resources(adapter, adapter->rx_ring[i]);
4548 4549 4550 4551 4552 4553 4554 4555 4556 4557 4558 4559 4560 4561
}

/**
 * ixgbe_change_mtu - Change the Maximum Transfer Unit
 * @netdev: network interface device structure
 * @new_mtu: new value for maximum frame size
 *
 * Returns 0 on success, negative on failure
 **/
static int ixgbe_change_mtu(struct net_device *netdev, int new_mtu)
{
	struct ixgbe_adapter *adapter = netdev_priv(netdev);
	int max_frame = new_mtu + ETH_HLEN + ETH_FCS_LEN;

4562 4563
	/* MTU < 68 is an error and causes problems on some kernels */
	if ((new_mtu < 68) || (max_frame > IXGBE_MAX_JUMBO_FRAME_SIZE))
4564 4565
		return -EINVAL;

4566
	DPRINTK(PROBE, INFO, "changing MTU from %d to %d\n",
4567
	        netdev->mtu, new_mtu);
4568
	/* must set new MTU before calling down or up */
4569 4570
	netdev->mtu = new_mtu;

4571 4572
	if (netif_running(netdev))
		ixgbe_reinit_locked(adapter);
4573 4574 4575 4576 4577 4578 4579 4580 4581 4582 4583 4584 4585 4586 4587 4588 4589 4590 4591 4592

	return 0;
}

/**
 * ixgbe_open - Called when a network interface is made active
 * @netdev: network interface device structure
 *
 * Returns 0 on success, negative value on failure
 *
 * The open entry point is called when a network interface is made
 * active by the system (IFF_UP).  At this point all resources needed
 * for transmit and receive operations are allocated, the interrupt
 * handler is registered with the OS, the watchdog timer is started,
 * and the stack is notified that the interface is ready.
 **/
static int ixgbe_open(struct net_device *netdev)
{
	struct ixgbe_adapter *adapter = netdev_priv(netdev);
	int err;
4593 4594 4595 4596

	/* disallow open during test */
	if (test_bit(__IXGBE_TESTING, &adapter->state))
		return -EBUSY;
4597

4598 4599
	netif_carrier_off(netdev);

4600 4601 4602 4603 4604 4605 4606 4607 4608 4609 4610 4611
	/* allocate transmit descriptors */
	err = ixgbe_setup_all_tx_resources(adapter);
	if (err)
		goto err_setup_tx;

	/* allocate receive descriptors */
	err = ixgbe_setup_all_rx_resources(adapter);
	if (err)
		goto err_setup_rx;

	ixgbe_configure(adapter);

4612
	err = ixgbe_request_irq(adapter);
4613 4614 4615 4616 4617 4618 4619
	if (err)
		goto err_req_irq;

	err = ixgbe_up_complete(adapter);
	if (err)
		goto err_up;

4620 4621
	netif_tx_start_all_queues(netdev);

4622 4623 4624
	return 0;

err_up:
4625
	ixgbe_release_hw_control(adapter);
4626 4627 4628
	ixgbe_free_irq(adapter);
err_req_irq:
err_setup_rx:
4629
	ixgbe_free_all_rx_resources(adapter);
4630
err_setup_tx:
4631
	ixgbe_free_all_tx_resources(adapter);
4632 4633 4634 4635 4636 4637 4638 4639 4640 4641 4642 4643 4644 4645 4646 4647 4648 4649 4650 4651 4652 4653 4654 4655 4656 4657
	ixgbe_reset(adapter);

	return err;
}

/**
 * ixgbe_close - Disables a network interface
 * @netdev: network interface device structure
 *
 * Returns 0, this is not allowed to fail
 *
 * The close entry point is called when an interface is de-activated
 * by the OS.  The hardware is still under the drivers control, but
 * needs to be disabled.  A global MAC reset is issued to stop the
 * hardware, and all transmit and receive resources are freed.
 **/
static int ixgbe_close(struct net_device *netdev)
{
	struct ixgbe_adapter *adapter = netdev_priv(netdev);

	ixgbe_down(adapter);
	ixgbe_free_irq(adapter);

	ixgbe_free_all_tx_resources(adapter);
	ixgbe_free_all_rx_resources(adapter);

4658
	ixgbe_release_hw_control(adapter);
4659 4660 4661 4662

	return 0;
}

4663 4664 4665 4666 4667 4668 4669 4670 4671
#ifdef CONFIG_PM
static int ixgbe_resume(struct pci_dev *pdev)
{
	struct net_device *netdev = pci_get_drvdata(pdev);
	struct ixgbe_adapter *adapter = netdev_priv(netdev);
	u32 err;

	pci_set_power_state(pdev, PCI_D0);
	pci_restore_state(pdev);
4672 4673 4674 4675 4676
	/*
	 * pci_restore_state clears dev->state_saved so call
	 * pci_save_state to restore it.
	 */
	pci_save_state(pdev);
4677 4678

	err = pci_enable_device_mem(pdev);
4679
	if (err) {
4680
		printk(KERN_ERR "ixgbe: Cannot enable PCI device from "
4681 4682 4683 4684 4685
				"suspend\n");
		return err;
	}
	pci_set_master(pdev);

4686
	pci_wake_from_d3(pdev, false);
4687 4688 4689 4690 4691 4692 4693 4694 4695 4696

	err = ixgbe_init_interrupt_scheme(adapter);
	if (err) {
		printk(KERN_ERR "ixgbe: Cannot initialize interrupts for "
		                "device\n");
		return err;
	}

	ixgbe_reset(adapter);

4697 4698
	IXGBE_WRITE_REG(&adapter->hw, IXGBE_WUS, ~0);

4699 4700 4701 4702 4703 4704 4705 4706 4707 4708 4709
	if (netif_running(netdev)) {
		err = ixgbe_open(adapter->netdev);
		if (err)
			return err;
	}

	netif_device_attach(netdev);

	return 0;
}
#endif /* CONFIG_PM */
4710 4711

static int __ixgbe_shutdown(struct pci_dev *pdev, bool *enable_wake)
4712 4713 4714
{
	struct net_device *netdev = pci_get_drvdata(pdev);
	struct ixgbe_adapter *adapter = netdev_priv(netdev);
4715 4716 4717
	struct ixgbe_hw *hw = &adapter->hw;
	u32 ctrl, fctrl;
	u32 wufc = adapter->wol;
4718 4719 4720 4721 4722 4723 4724 4725 4726 4727 4728 4729
#ifdef CONFIG_PM
	int retval = 0;
#endif

	netif_device_detach(netdev);

	if (netif_running(netdev)) {
		ixgbe_down(adapter);
		ixgbe_free_irq(adapter);
		ixgbe_free_all_tx_resources(adapter);
		ixgbe_free_all_rx_resources(adapter);
	}
4730
	ixgbe_clear_interrupt_scheme(adapter);
4731 4732 4733 4734 4735

#ifdef CONFIG_PM
	retval = pci_save_state(pdev);
	if (retval)
		return retval;
4736

4737
#endif
4738 4739
	if (wufc) {
		ixgbe_set_rx_mode(netdev);
4740

4741 4742 4743 4744 4745 4746 4747 4748 4749 4750 4751 4752 4753 4754 4755 4756 4757
		/* turn on all-multi mode if wake on multicast is enabled */
		if (wufc & IXGBE_WUFC_MC) {
			fctrl = IXGBE_READ_REG(hw, IXGBE_FCTRL);
			fctrl |= IXGBE_FCTRL_MPE;
			IXGBE_WRITE_REG(hw, IXGBE_FCTRL, fctrl);
		}

		ctrl = IXGBE_READ_REG(hw, IXGBE_CTRL);
		ctrl |= IXGBE_CTRL_GIO_DIS;
		IXGBE_WRITE_REG(hw, IXGBE_CTRL, ctrl);

		IXGBE_WRITE_REG(hw, IXGBE_WUFC, wufc);
	} else {
		IXGBE_WRITE_REG(hw, IXGBE_WUC, 0);
		IXGBE_WRITE_REG(hw, IXGBE_WUFC, 0);
	}

4758 4759 4760 4761
	if (wufc && hw->mac.type == ixgbe_mac_82599EB)
		pci_wake_from_d3(pdev, true);
	else
		pci_wake_from_d3(pdev, false);
4762

4763 4764
	*enable_wake = !!wufc;

4765 4766 4767 4768
	ixgbe_release_hw_control(adapter);

	pci_disable_device(pdev);

4769 4770 4771 4772 4773 4774 4775 4776 4777 4778 4779 4780 4781 4782 4783 4784 4785 4786 4787
	return 0;
}

#ifdef CONFIG_PM
static int ixgbe_suspend(struct pci_dev *pdev, pm_message_t state)
{
	int retval;
	bool wake;

	retval = __ixgbe_shutdown(pdev, &wake);
	if (retval)
		return retval;

	if (wake) {
		pci_prepare_to_sleep(pdev);
	} else {
		pci_wake_from_d3(pdev, false);
		pci_set_power_state(pdev, PCI_D3hot);
	}
4788 4789 4790

	return 0;
}
4791
#endif /* CONFIG_PM */
4792 4793 4794

static void ixgbe_shutdown(struct pci_dev *pdev)
{
4795 4796 4797 4798 4799 4800 4801 4802
	bool wake;

	__ixgbe_shutdown(pdev, &wake);

	if (system_state == SYSTEM_POWER_OFF) {
		pci_wake_from_d3(pdev, wake);
		pci_set_power_state(pdev, PCI_D3hot);
	}
4803 4804
}

4805 4806 4807 4808 4809 4810
/**
 * ixgbe_update_stats - Update the board statistics counters.
 * @adapter: board private structure
 **/
void ixgbe_update_stats(struct ixgbe_adapter *adapter)
{
4811
	struct net_device *netdev = adapter->netdev;
4812
	struct ixgbe_hw *hw = &adapter->hw;
4813 4814
	u64 total_mpc = 0;
	u32 i, missed_rx = 0, mpc, bprc, lxon, lxoff, xon_off_tot;
4815
	u64 non_eop_descs = 0, restart_queue = 0;
4816

4817
	if (adapter->flags2 & IXGBE_FLAG2_RSC_ENABLED) {
A
Alexander Duyck 已提交
4818
		u64 rsc_count = 0;
4819
		u64 rsc_flush = 0;
4820 4821 4822
		for (i = 0; i < 16; i++)
			adapter->hw_rx_no_dma_resources +=
			                     IXGBE_READ_REG(hw, IXGBE_QPRDC(i));
4823
		for (i = 0; i < adapter->num_rx_queues; i++) {
4824 4825
			rsc_count += adapter->rx_ring[i]->rsc_count;
			rsc_flush += adapter->rx_ring[i]->rsc_flush;
4826 4827 4828
		}
		adapter->rsc_total_count = rsc_count;
		adapter->rsc_total_flush = rsc_flush;
4829 4830
	}

J
Jesse Brandeburg 已提交
4831 4832
	/* gather some stats to the adapter struct that are per queue */
	for (i = 0; i < adapter->num_tx_queues; i++)
4833
		restart_queue += adapter->tx_ring[i]->restart_queue;
4834
	adapter->restart_queue = restart_queue;
J
Jesse Brandeburg 已提交
4835 4836

	for (i = 0; i < adapter->num_rx_queues; i++)
4837
		non_eop_descs += adapter->rx_ring[i]->non_eop_descs;
4838
	adapter->non_eop_descs = non_eop_descs;
J
Jesse Brandeburg 已提交
4839

4840
	adapter->stats.crcerrs += IXGBE_READ_REG(hw, IXGBE_CRCERRS);
4841 4842 4843 4844 4845 4846
	for (i = 0; i < 8; i++) {
		/* for packet buffers not used, the register should read 0 */
		mpc = IXGBE_READ_REG(hw, IXGBE_MPC(i));
		missed_rx += mpc;
		adapter->stats.mpc[i] += mpc;
		total_mpc += adapter->stats.mpc[i];
4847 4848
		if (hw->mac.type == ixgbe_mac_82598EB)
			adapter->stats.rnbc[i] += IXGBE_READ_REG(hw, IXGBE_RNBC(i));
4849 4850 4851 4852
		adapter->stats.qptc[i] += IXGBE_READ_REG(hw, IXGBE_QPTC(i));
		adapter->stats.qbtc[i] += IXGBE_READ_REG(hw, IXGBE_QBTC(i));
		adapter->stats.qprc[i] += IXGBE_READ_REG(hw, IXGBE_QPRC(i));
		adapter->stats.qbrc[i] += IXGBE_READ_REG(hw, IXGBE_QBRC(i));
4853 4854 4855 4856 4857 4858 4859 4860 4861 4862 4863 4864
		if (hw->mac.type == ixgbe_mac_82599EB) {
			adapter->stats.pxonrxc[i] += IXGBE_READ_REG(hw,
			                                    IXGBE_PXONRXCNT(i));
			adapter->stats.pxoffrxc[i] += IXGBE_READ_REG(hw,
			                                   IXGBE_PXOFFRXCNT(i));
			adapter->stats.qprdc[i] += IXGBE_READ_REG(hw, IXGBE_QPRDC(i));
		} else {
			adapter->stats.pxonrxc[i] += IXGBE_READ_REG(hw,
			                                      IXGBE_PXONRXC(i));
			adapter->stats.pxoffrxc[i] += IXGBE_READ_REG(hw,
			                                     IXGBE_PXOFFRXC(i));
		}
4865 4866 4867
		adapter->stats.pxontxc[i] += IXGBE_READ_REG(hw,
		                                            IXGBE_PXONTXC(i));
		adapter->stats.pxofftxc[i] += IXGBE_READ_REG(hw,
4868
		                                             IXGBE_PXOFFTXC(i));
4869 4870 4871 4872 4873 4874
	}
	adapter->stats.gprc += IXGBE_READ_REG(hw, IXGBE_GPRC);
	/* work around hardware counting issue */
	adapter->stats.gprc -= missed_rx;

	/* 82598 hardware only has a 32 bit counter in the high register */
4875
	if (hw->mac.type == ixgbe_mac_82599EB) {
4876
		u64 tmp;
4877
		adapter->stats.gorc += IXGBE_READ_REG(hw, IXGBE_GORCL);
4878 4879
		tmp = IXGBE_READ_REG(hw, IXGBE_GORCH) & 0xF; /* 4 high bits of GORC */
		adapter->stats.gorc += (tmp << 32);
4880
		adapter->stats.gotc += IXGBE_READ_REG(hw, IXGBE_GOTCL);
4881 4882
		tmp = IXGBE_READ_REG(hw, IXGBE_GOTCH) & 0xF; /* 4 high bits of GOTC */
		adapter->stats.gotc += (tmp << 32);
4883 4884 4885 4886
		adapter->stats.tor += IXGBE_READ_REG(hw, IXGBE_TORL);
		IXGBE_READ_REG(hw, IXGBE_TORH); /* to clear */
		adapter->stats.lxonrxc += IXGBE_READ_REG(hw, IXGBE_LXONRXCNT);
		adapter->stats.lxoffrxc += IXGBE_READ_REG(hw, IXGBE_LXOFFRXCNT);
4887 4888
		adapter->stats.fdirmatch += IXGBE_READ_REG(hw, IXGBE_FDIRMATCH);
		adapter->stats.fdirmiss += IXGBE_READ_REG(hw, IXGBE_FDIRMISS);
4889 4890 4891 4892 4893 4894 4895 4896
#ifdef IXGBE_FCOE
		adapter->stats.fccrc += IXGBE_READ_REG(hw, IXGBE_FCCRC);
		adapter->stats.fcoerpdc += IXGBE_READ_REG(hw, IXGBE_FCOERPDC);
		adapter->stats.fcoeprc += IXGBE_READ_REG(hw, IXGBE_FCOEPRC);
		adapter->stats.fcoeptc += IXGBE_READ_REG(hw, IXGBE_FCOEPTC);
		adapter->stats.fcoedwrc += IXGBE_READ_REG(hw, IXGBE_FCOEDWRC);
		adapter->stats.fcoedwtc += IXGBE_READ_REG(hw, IXGBE_FCOEDWTC);
#endif /* IXGBE_FCOE */
4897 4898 4899 4900 4901 4902 4903
	} else {
		adapter->stats.lxonrxc += IXGBE_READ_REG(hw, IXGBE_LXONRXC);
		adapter->stats.lxoffrxc += IXGBE_READ_REG(hw, IXGBE_LXOFFRXC);
		adapter->stats.gorc += IXGBE_READ_REG(hw, IXGBE_GORCH);
		adapter->stats.gotc += IXGBE_READ_REG(hw, IXGBE_GOTCH);
		adapter->stats.tor += IXGBE_READ_REG(hw, IXGBE_TORH);
	}
4904 4905 4906
	bprc = IXGBE_READ_REG(hw, IXGBE_BPRC);
	adapter->stats.bprc += bprc;
	adapter->stats.mprc += IXGBE_READ_REG(hw, IXGBE_MPRC);
4907 4908
	if (hw->mac.type == ixgbe_mac_82598EB)
		adapter->stats.mprc -= bprc;
4909 4910 4911 4912 4913 4914 4915 4916
	adapter->stats.roc += IXGBE_READ_REG(hw, IXGBE_ROC);
	adapter->stats.prc64 += IXGBE_READ_REG(hw, IXGBE_PRC64);
	adapter->stats.prc127 += IXGBE_READ_REG(hw, IXGBE_PRC127);
	adapter->stats.prc255 += IXGBE_READ_REG(hw, IXGBE_PRC255);
	adapter->stats.prc511 += IXGBE_READ_REG(hw, IXGBE_PRC511);
	adapter->stats.prc1023 += IXGBE_READ_REG(hw, IXGBE_PRC1023);
	adapter->stats.prc1522 += IXGBE_READ_REG(hw, IXGBE_PRC1522);
	adapter->stats.rlec += IXGBE_READ_REG(hw, IXGBE_RLEC);
4917 4918 4919 4920
	lxon = IXGBE_READ_REG(hw, IXGBE_LXONTXC);
	adapter->stats.lxontxc += lxon;
	lxoff = IXGBE_READ_REG(hw, IXGBE_LXOFFTXC);
	adapter->stats.lxofftxc += lxoff;
4921 4922
	adapter->stats.ruc += IXGBE_READ_REG(hw, IXGBE_RUC);
	adapter->stats.gptc += IXGBE_READ_REG(hw, IXGBE_GPTC);
4923 4924 4925 4926 4927 4928 4929 4930
	adapter->stats.mptc += IXGBE_READ_REG(hw, IXGBE_MPTC);
	/*
	 * 82598 errata - tx of flow control packets is included in tx counters
	 */
	xon_off_tot = lxon + lxoff;
	adapter->stats.gptc -= xon_off_tot;
	adapter->stats.mptc -= xon_off_tot;
	adapter->stats.gotc -= (xon_off_tot * (ETH_ZLEN + ETH_FCS_LEN));
4931 4932 4933 4934 4935
	adapter->stats.ruc += IXGBE_READ_REG(hw, IXGBE_RUC);
	adapter->stats.rfc += IXGBE_READ_REG(hw, IXGBE_RFC);
	adapter->stats.rjc += IXGBE_READ_REG(hw, IXGBE_RJC);
	adapter->stats.tpr += IXGBE_READ_REG(hw, IXGBE_TPR);
	adapter->stats.ptc64 += IXGBE_READ_REG(hw, IXGBE_PTC64);
4936
	adapter->stats.ptc64 -= xon_off_tot;
4937 4938 4939 4940 4941 4942 4943 4944
	adapter->stats.ptc127 += IXGBE_READ_REG(hw, IXGBE_PTC127);
	adapter->stats.ptc255 += IXGBE_READ_REG(hw, IXGBE_PTC255);
	adapter->stats.ptc511 += IXGBE_READ_REG(hw, IXGBE_PTC511);
	adapter->stats.ptc1023 += IXGBE_READ_REG(hw, IXGBE_PTC1023);
	adapter->stats.ptc1522 += IXGBE_READ_REG(hw, IXGBE_PTC1522);
	adapter->stats.bptc += IXGBE_READ_REG(hw, IXGBE_BPTC);

	/* Fill out the OS statistics structure */
4945
	netdev->stats.multicast = adapter->stats.mprc;
4946 4947

	/* Rx Errors */
4948
	netdev->stats.rx_errors = adapter->stats.crcerrs +
4949
	                               adapter->stats.rlec;
4950 4951 4952 4953
	netdev->stats.rx_dropped = 0;
	netdev->stats.rx_length_errors = adapter->stats.rlec;
	netdev->stats.rx_crc_errors = adapter->stats.crcerrs;
	netdev->stats.rx_missed_errors = total_mpc;
4954 4955 4956 4957 4958 4959 4960 4961 4962
}

/**
 * ixgbe_watchdog - Timer Call-back
 * @data: pointer to adapter cast into an unsigned long
 **/
static void ixgbe_watchdog(unsigned long data)
{
	struct ixgbe_adapter *adapter = (struct ixgbe_adapter *)data;
4963
	struct ixgbe_hw *hw = &adapter->hw;
4964 4965
	u64 eics = 0;
	int i;
4966

4967 4968 4969 4970
	/*
	 *  Do the watchdog outside of interrupt context due to the lovely
	 * delays that some of the newer hardware requires
	 */
4971

4972 4973
	if (test_bit(__IXGBE_DOWN, &adapter->state))
		goto watchdog_short_circuit;
4974

4975 4976 4977 4978 4979 4980 4981 4982 4983 4984 4985 4986 4987 4988 4989 4990
	if (!(adapter->flags & IXGBE_FLAG_MSIX_ENABLED)) {
		/*
		 * for legacy and MSI interrupts don't set any bits
		 * that are enabled for EIAM, because this operation
		 * would set *both* EIMS and EICS for any bit in EIAM
		 */
		IXGBE_WRITE_REG(hw, IXGBE_EICS,
			(IXGBE_EICS_TCP_TIMER | IXGBE_EICS_OTHER));
		goto watchdog_reschedule;
	}

	/* get one bit for every active tx/rx interrupt vector */
	for (i = 0; i < adapter->num_msix_vectors - NON_Q_VECTORS; i++) {
		struct ixgbe_q_vector *qv = adapter->q_vector[i];
		if (qv->rxr_count || qv->txr_count)
			eics |= ((u64)1 << i);
4991
	}
4992

4993 4994 4995 4996 4997 4998 4999 5000
	/* Cause software interrupt to ensure rx rings are cleaned */
	ixgbe_irq_rearm_queues(adapter, eics);

watchdog_reschedule:
	/* Reset the timer */
	mod_timer(&adapter->watchdog_timer, round_jiffies(jiffies + 2 * HZ));

watchdog_short_circuit:
5001 5002 5003
	schedule_work(&adapter->watchdog_task);
}

5004 5005 5006 5007 5008 5009 5010 5011 5012 5013 5014
/**
 * ixgbe_multispeed_fiber_task - worker thread to configure multispeed fiber
 * @work: pointer to work_struct containing our data
 **/
static void ixgbe_multispeed_fiber_task(struct work_struct *work)
{
	struct ixgbe_adapter *adapter = container_of(work,
	                                             struct ixgbe_adapter,
	                                             multispeed_fiber_task);
	struct ixgbe_hw *hw = &adapter->hw;
	u32 autoneg;
5015
	bool negotiation;
5016 5017

	adapter->flags |= IXGBE_FLAG_IN_SFP_LINK_TASK;
5018 5019
	autoneg = hw->phy.autoneg_advertised;
	if ((!autoneg) && (hw->mac.ops.get_link_capabilities))
5020
		hw->mac.ops.get_link_capabilities(hw, &autoneg, &negotiation);
5021
	hw->mac.autotry_restart = false;
5022 5023
	if (hw->mac.ops.setup_link)
		hw->mac.ops.setup_link(hw, autoneg, negotiation, true);
5024 5025 5026 5027 5028 5029 5030 5031 5032 5033 5034 5035 5036 5037 5038 5039 5040
	adapter->flags |= IXGBE_FLAG_NEED_LINK_UPDATE;
	adapter->flags &= ~IXGBE_FLAG_IN_SFP_LINK_TASK;
}

/**
 * ixgbe_sfp_config_module_task - worker thread to configure a new SFP+ module
 * @work: pointer to work_struct containing our data
 **/
static void ixgbe_sfp_config_module_task(struct work_struct *work)
{
	struct ixgbe_adapter *adapter = container_of(work,
	                                             struct ixgbe_adapter,
	                                             sfp_config_module_task);
	struct ixgbe_hw *hw = &adapter->hw;
	u32 err;

	adapter->flags |= IXGBE_FLAG_IN_SFP_MOD_TASK;
5041 5042 5043

	/* Time for electrical oscillations to settle down */
	msleep(100);
5044
	err = hw->phy.ops.identify_sfp(hw);
5045

5046
	if (err == IXGBE_ERR_SFP_NOT_SUPPORTED) {
D
Don Skidmore 已提交
5047 5048 5049 5050
		dev_err(&adapter->pdev->dev, "failed to initialize because "
			"an unsupported SFP+ module type was detected.\n"
			"Reload the driver after installing a supported "
			"module.\n");
5051
		unregister_netdev(adapter->netdev);
5052 5053 5054 5055
		return;
	}
	hw->mac.ops.setup_sfp(hw);

5056
	if (!(adapter->flags & IXGBE_FLAG_IN_SFP_LINK_TASK))
5057 5058 5059 5060 5061
		/* This will also work for DA Twinax connections */
		schedule_work(&adapter->multispeed_fiber_task);
	adapter->flags &= ~IXGBE_FLAG_IN_SFP_MOD_TASK;
}

5062 5063 5064 5065 5066 5067 5068 5069 5070 5071 5072 5073 5074 5075 5076
/**
 * ixgbe_fdir_reinit_task - worker thread to reinit FDIR filter table
 * @work: pointer to work_struct containing our data
 **/
static void ixgbe_fdir_reinit_task(struct work_struct *work)
{
	struct ixgbe_adapter *adapter = container_of(work,
	                                             struct ixgbe_adapter,
	                                             fdir_reinit_task);
	struct ixgbe_hw *hw = &adapter->hw;
	int i;

	if (ixgbe_reinit_fdir_tables_82599(hw) == 0) {
		for (i = 0; i < adapter->num_tx_queues; i++)
			set_bit(__IXGBE_FDIR_INIT_DONE,
5077
			        &(adapter->tx_ring[i]->reinit_state));
5078 5079 5080 5081 5082 5083 5084 5085
	} else {
		DPRINTK(PROBE, ERR, "failed to finish FDIR re-initialization, "
		        "ignored adding FDIR ATR filters \n");
	}
	/* Done FDIR Re-initialization, enable transmits */
	netif_tx_start_all_queues(adapter->netdev);
}

5086 5087
static DEFINE_MUTEX(ixgbe_watchdog_lock);

5088
/**
5089 5090
 * ixgbe_watchdog_task - worker thread to bring link up
 * @work: pointer to work_struct containing our data
5091 5092 5093 5094 5095 5096 5097 5098
 **/
static void ixgbe_watchdog_task(struct work_struct *work)
{
	struct ixgbe_adapter *adapter = container_of(work,
	                                             struct ixgbe_adapter,
	                                             watchdog_task);
	struct net_device *netdev = adapter->netdev;
	struct ixgbe_hw *hw = &adapter->hw;
5099 5100
	u32 link_speed;
	bool link_up;
5101 5102 5103
	int i;
	struct ixgbe_ring *tx_ring;
	int some_tx_pending = 0;
5104

5105 5106 5107 5108
	mutex_lock(&ixgbe_watchdog_lock);

	link_up = adapter->link_up;
	link_speed = adapter->link_speed;
5109 5110 5111

	if (adapter->flags & IXGBE_FLAG_NEED_LINK_UPDATE) {
		hw->mac.ops.check_link(hw, &link_speed, &link_up, false);
5112 5113 5114 5115
		if (link_up) {
#ifdef CONFIG_DCB
			if (adapter->flags & IXGBE_FLAG_DCB_ENABLED) {
				for (i = 0; i < MAX_TRAFFIC_CLASS; i++)
5116
					hw->mac.ops.fc_enable(hw, i);
5117
			} else {
5118
				hw->mac.ops.fc_enable(hw, 0);
5119 5120
			}
#else
5121
			hw->mac.ops.fc_enable(hw, 0);
5122 5123 5124
#endif
		}

5125 5126 5127 5128
		if (link_up ||
		    time_after(jiffies, (adapter->link_check_timeout +
		                         IXGBE_TRY_LINK_TIMEOUT))) {
			adapter->flags &= ~IXGBE_FLAG_NEED_LINK_UPDATE;
5129
			IXGBE_WRITE_REG(hw, IXGBE_EIMS, IXGBE_EIMC_LSC);
5130 5131 5132 5133
		}
		adapter->link_up = link_up;
		adapter->link_speed = link_speed;
	}
5134 5135 5136

	if (link_up) {
		if (!netif_carrier_ok(netdev)) {
5137 5138 5139 5140 5141
			bool flow_rx, flow_tx;

			if (hw->mac.type == ixgbe_mac_82599EB) {
				u32 mflcn = IXGBE_READ_REG(hw, IXGBE_MFLCN);
				u32 fccfg = IXGBE_READ_REG(hw, IXGBE_FCCFG);
5142 5143
				flow_rx = !!(mflcn & IXGBE_MFLCN_RFCE);
				flow_tx = !!(fccfg & IXGBE_FCCFG_TFCE_802_3X);
5144 5145 5146
			} else {
				u32 frctl = IXGBE_READ_REG(hw, IXGBE_FCTRL);
				u32 rmcs = IXGBE_READ_REG(hw, IXGBE_RMCS);
5147 5148
				flow_rx = !!(frctl & IXGBE_FCTRL_RFCE);
				flow_tx = !!(rmcs & IXGBE_RMCS_TFCE_802_3X);
5149 5150
			}

5151 5152 5153 5154 5155 5156 5157
			printk(KERN_INFO "ixgbe: %s NIC Link is Up %s, "
			       "Flow Control: %s\n",
			       netdev->name,
			       (link_speed == IXGBE_LINK_SPEED_10GB_FULL ?
			        "10 Gbps" :
			        (link_speed == IXGBE_LINK_SPEED_1GB_FULL ?
			         "1 Gbps" : "unknown speed")),
5158 5159 5160
			       ((flow_rx && flow_tx) ? "RX/TX" :
			        (flow_rx ? "RX" :
			        (flow_tx ? "TX" : "None"))));
5161 5162 5163 5164 5165 5166 5167

			netif_carrier_on(netdev);
		} else {
			/* Force detection of hung controller */
			adapter->detect_tx_hung = true;
		}
	} else {
5168 5169
		adapter->link_up = false;
		adapter->link_speed = 0;
5170
		if (netif_carrier_ok(netdev)) {
5171 5172
			printk(KERN_INFO "ixgbe: %s NIC Link is Down\n",
			       netdev->name);
5173 5174 5175 5176
			netif_carrier_off(netdev);
		}
	}

5177 5178
	if (!netif_carrier_ok(netdev)) {
		for (i = 0; i < adapter->num_tx_queues; i++) {
5179
			tx_ring = adapter->tx_ring[i];
5180 5181 5182 5183 5184 5185 5186 5187 5188 5189 5190 5191 5192 5193 5194 5195
			if (tx_ring->next_to_use != tx_ring->next_to_clean) {
				some_tx_pending = 1;
				break;
			}
		}

		if (some_tx_pending) {
			/* We've lost link, so the controller stops DMA,
			 * but we've got queued Tx work that's never going
			 * to get done, so reset controller to flush Tx.
			 * (Do the reset outside of interrupt context).
			 */
			 schedule_work(&adapter->reset_task);
		}
	}

5196
	ixgbe_update_stats(adapter);
5197
	mutex_unlock(&ixgbe_watchdog_lock);
5198 5199 5200
}

static int ixgbe_tso(struct ixgbe_adapter *adapter,
5201 5202
                     struct ixgbe_ring *tx_ring, struct sk_buff *skb,
                     u32 tx_flags, u8 *hdr_len)
5203 5204 5205 5206 5207
{
	struct ixgbe_adv_tx_context_desc *context_desc;
	unsigned int i;
	int err;
	struct ixgbe_tx_buffer *tx_buffer_info;
J
Jesse Brandeburg 已提交
5208 5209
	u32 vlan_macip_lens = 0, type_tucmd_mlhl;
	u32 mss_l4len_idx, l4len;
5210 5211 5212 5213 5214 5215 5216 5217 5218 5219

	if (skb_is_gso(skb)) {
		if (skb_header_cloned(skb)) {
			err = pskb_expand_head(skb, 0, 0, GFP_ATOMIC);
			if (err)
				return err;
		}
		l4len = tcp_hdrlen(skb);
		*hdr_len += l4len;

A
Al Viro 已提交
5220
		if (skb->protocol == htons(ETH_P_IP)) {
5221 5222 5223 5224
			struct iphdr *iph = ip_hdr(skb);
			iph->tot_len = 0;
			iph->check = 0;
			tcp_hdr(skb)->check = ~csum_tcpudp_magic(iph->saddr,
5225 5226 5227
			                                         iph->daddr, 0,
			                                         IPPROTO_TCP,
			                                         0);
5228
		} else if (skb_is_gso_v6(skb)) {
5229 5230 5231
			ipv6_hdr(skb)->payload_len = 0;
			tcp_hdr(skb)->check =
			    ~csum_ipv6_magic(&ipv6_hdr(skb)->saddr,
5232 5233
			                     &ipv6_hdr(skb)->daddr,
			                     0, IPPROTO_TCP, 0);
5234 5235 5236 5237 5238 5239 5240 5241 5242 5243 5244 5245
		}

		i = tx_ring->next_to_use;

		tx_buffer_info = &tx_ring->tx_buffer_info[i];
		context_desc = IXGBE_TX_CTXTDESC_ADV(*tx_ring, i);

		/* VLAN MACLEN IPLEN */
		if (tx_flags & IXGBE_TX_FLAGS_VLAN)
			vlan_macip_lens |=
			    (tx_flags & IXGBE_TX_FLAGS_VLAN_MASK);
		vlan_macip_lens |= ((skb_network_offset(skb)) <<
5246
		                    IXGBE_ADVTXD_MACLEN_SHIFT);
5247 5248 5249 5250 5251 5252 5253 5254 5255
		*hdr_len += skb_network_offset(skb);
		vlan_macip_lens |=
		    (skb_transport_header(skb) - skb_network_header(skb));
		*hdr_len +=
		    (skb_transport_header(skb) - skb_network_header(skb));
		context_desc->vlan_macip_lens = cpu_to_le32(vlan_macip_lens);
		context_desc->seqnum_seed = 0;

		/* ADV DTYP TUCMD MKRLOC/ISCSIHEDLEN */
J
Jesse Brandeburg 已提交
5256
		type_tucmd_mlhl = (IXGBE_TXD_CMD_DEXT |
5257
		                   IXGBE_ADVTXD_DTYP_CTXT);
5258

A
Al Viro 已提交
5259
		if (skb->protocol == htons(ETH_P_IP))
5260 5261 5262 5263 5264
			type_tucmd_mlhl |= IXGBE_ADVTXD_TUCMD_IPV4;
		type_tucmd_mlhl |= IXGBE_ADVTXD_TUCMD_L4T_TCP;
		context_desc->type_tucmd_mlhl = cpu_to_le32(type_tucmd_mlhl);

		/* MSS L4LEN IDX */
J
Jesse Brandeburg 已提交
5265
		mss_l4len_idx =
5266 5267
		    (skb_shinfo(skb)->gso_size << IXGBE_ADVTXD_MSS_SHIFT);
		mss_l4len_idx |= (l4len << IXGBE_ADVTXD_L4LEN_SHIFT);
5268 5269
		/* use index 1 for TSO */
		mss_l4len_idx |= (1 << IXGBE_ADVTXD_IDX_SHIFT);
5270 5271 5272 5273 5274 5275 5276 5277 5278 5279 5280 5281 5282 5283 5284 5285
		context_desc->mss_l4len_idx = cpu_to_le32(mss_l4len_idx);

		tx_buffer_info->time_stamp = jiffies;
		tx_buffer_info->next_to_watch = i;

		i++;
		if (i == tx_ring->count)
			i = 0;
		tx_ring->next_to_use = i;

		return true;
	}
	return false;
}

static bool ixgbe_tx_csum(struct ixgbe_adapter *adapter,
5286 5287
                          struct ixgbe_ring *tx_ring,
                          struct sk_buff *skb, u32 tx_flags)
5288 5289 5290 5291 5292 5293 5294 5295 5296 5297 5298 5299 5300 5301 5302 5303
{
	struct ixgbe_adv_tx_context_desc *context_desc;
	unsigned int i;
	struct ixgbe_tx_buffer *tx_buffer_info;
	u32 vlan_macip_lens = 0, type_tucmd_mlhl = 0;

	if (skb->ip_summed == CHECKSUM_PARTIAL ||
	    (tx_flags & IXGBE_TX_FLAGS_VLAN)) {
		i = tx_ring->next_to_use;
		tx_buffer_info = &tx_ring->tx_buffer_info[i];
		context_desc = IXGBE_TX_CTXTDESC_ADV(*tx_ring, i);

		if (tx_flags & IXGBE_TX_FLAGS_VLAN)
			vlan_macip_lens |=
			    (tx_flags & IXGBE_TX_FLAGS_VLAN_MASK);
		vlan_macip_lens |= (skb_network_offset(skb) <<
5304
		                    IXGBE_ADVTXD_MACLEN_SHIFT);
5305 5306
		if (skb->ip_summed == CHECKSUM_PARTIAL)
			vlan_macip_lens |= (skb_transport_header(skb) -
5307
			                    skb_network_header(skb));
5308 5309 5310 5311 5312

		context_desc->vlan_macip_lens = cpu_to_le32(vlan_macip_lens);
		context_desc->seqnum_seed = 0;

		type_tucmd_mlhl |= (IXGBE_TXD_CMD_DEXT |
5313
		                    IXGBE_ADVTXD_DTYP_CTXT);
5314 5315

		if (skb->ip_summed == CHECKSUM_PARTIAL) {
5316 5317 5318 5319 5320 5321 5322 5323 5324 5325 5326 5327
			__be16 protocol;

			if (skb->protocol == cpu_to_be16(ETH_P_8021Q)) {
				const struct vlan_ethhdr *vhdr =
					(const struct vlan_ethhdr *)skb->data;

				protocol = vhdr->h_vlan_encapsulated_proto;
			} else {
				protocol = skb->protocol;
			}

			switch (protocol) {
5328
			case cpu_to_be16(ETH_P_IP):
5329
				type_tucmd_mlhl |= IXGBE_ADVTXD_TUCMD_IPV4;
5330 5331
				if (ip_hdr(skb)->protocol == IPPROTO_TCP)
					type_tucmd_mlhl |=
5332
					        IXGBE_ADVTXD_TUCMD_L4T_TCP;
5333 5334 5335
				else if (ip_hdr(skb)->protocol == IPPROTO_SCTP)
					type_tucmd_mlhl |=
					        IXGBE_ADVTXD_TUCMD_L4T_SCTP;
5336
				break;
5337
			case cpu_to_be16(ETH_P_IPV6):
5338 5339 5340
				/* XXX what about other V6 headers?? */
				if (ipv6_hdr(skb)->nexthdr == IPPROTO_TCP)
					type_tucmd_mlhl |=
5341
					        IXGBE_ADVTXD_TUCMD_L4T_TCP;
5342 5343 5344
				else if (ipv6_hdr(skb)->nexthdr == IPPROTO_SCTP)
					type_tucmd_mlhl |=
					        IXGBE_ADVTXD_TUCMD_L4T_SCTP;
5345 5346 5347 5348 5349 5350 5351 5352 5353
				break;
			default:
				if (unlikely(net_ratelimit())) {
					DPRINTK(PROBE, WARNING,
					 "partial checksum but proto=%x!\n",
					 skb->protocol);
				}
				break;
			}
5354 5355 5356
		}

		context_desc->type_tucmd_mlhl = cpu_to_le32(type_tucmd_mlhl);
5357
		/* use index zero for tx checksum offload */
5358 5359 5360 5361
		context_desc->mss_l4len_idx = 0;

		tx_buffer_info->time_stamp = jiffies;
		tx_buffer_info->next_to_watch = i;
J
Jesse Brandeburg 已提交
5362

5363 5364 5365 5366 5367 5368 5369
		i++;
		if (i == tx_ring->count)
			i = 0;
		tx_ring->next_to_use = i;

		return true;
	}
J
Jesse Brandeburg 已提交
5370

5371 5372 5373 5374
	return false;
}

static int ixgbe_tx_map(struct ixgbe_adapter *adapter,
5375
                        struct ixgbe_ring *tx_ring,
5376 5377
                        struct sk_buff *skb, u32 tx_flags,
                        unsigned int first)
5378
{
5379
	struct pci_dev *pdev = adapter->pdev;
5380
	struct ixgbe_tx_buffer *tx_buffer_info;
5381 5382
	unsigned int len;
	unsigned int total = skb->len;
5383 5384 5385 5386 5387 5388
	unsigned int offset = 0, size, count = 0, i;
	unsigned int nr_frags = skb_shinfo(skb)->nr_frags;
	unsigned int f;

	i = tx_ring->next_to_use;

5389 5390 5391 5392 5393
	if (tx_flags & IXGBE_TX_FLAGS_FCOE)
		/* excluding fcoe_crc_eof for FCoE */
		total -= sizeof(struct fcoe_crc_eof);

	len = min(skb_headlen(skb), total);
5394 5395 5396 5397 5398
	while (len) {
		tx_buffer_info = &tx_ring->tx_buffer_info[i];
		size = min(len, (uint)IXGBE_MAX_DATA_PER_TXD);

		tx_buffer_info->length = size;
5399 5400 5401 5402 5403 5404
		tx_buffer_info->mapped_as_page = false;
		tx_buffer_info->dma = pci_map_single(pdev,
						     skb->data + offset,
						     size, PCI_DMA_TODEVICE);
		if (pci_dma_mapping_error(pdev, tx_buffer_info->dma))
			goto dma_error;
5405 5406 5407 5408
		tx_buffer_info->time_stamp = jiffies;
		tx_buffer_info->next_to_watch = i;

		len -= size;
5409
		total -= size;
5410 5411
		offset += size;
		count++;
5412 5413 5414 5415 5416 5417

		if (len) {
			i++;
			if (i == tx_ring->count)
				i = 0;
		}
5418 5419 5420 5421 5422 5423
	}

	for (f = 0; f < nr_frags; f++) {
		struct skb_frag_struct *frag;

		frag = &skb_shinfo(skb)->frags[f];
5424
		len = min((unsigned int)frag->size, total);
5425
		offset = frag->page_offset;
5426 5427

		while (len) {
5428 5429 5430 5431
			i++;
			if (i == tx_ring->count)
				i = 0;

5432 5433 5434 5435
			tx_buffer_info = &tx_ring->tx_buffer_info[i];
			size = min(len, (uint)IXGBE_MAX_DATA_PER_TXD);

			tx_buffer_info->length = size;
5436 5437 5438 5439 5440 5441 5442
			tx_buffer_info->dma = pci_map_page(adapter->pdev,
							   frag->page,
							   offset, size,
							   PCI_DMA_TODEVICE);
			tx_buffer_info->mapped_as_page = true;
			if (pci_dma_mapping_error(pdev, tx_buffer_info->dma))
				goto dma_error;
5443 5444 5445 5446
			tx_buffer_info->time_stamp = jiffies;
			tx_buffer_info->next_to_watch = i;

			len -= size;
5447
			total -= size;
5448 5449 5450
			offset += size;
			count++;
		}
5451 5452
		if (total == 0)
			break;
5453
	}
5454

5455 5456 5457
	tx_ring->tx_buffer_info[i].skb = skb;
	tx_ring->tx_buffer_info[first].next_to_watch = i;

5458 5459 5460 5461 5462 5463 5464 5465 5466
	return count;

dma_error:
	dev_err(&pdev->dev, "TX DMA map failed\n");

	/* clear timestamp and dma mappings for failed tx_buffer_info map */
	tx_buffer_info->dma = 0;
	tx_buffer_info->time_stamp = 0;
	tx_buffer_info->next_to_watch = 0;
5467 5468
	if (count)
		count--;
5469 5470

	/* clear timestamp and dma mappings for remaining portion of packet */
5471 5472
	while (count--) {
		if (i==0)
5473
			i += tx_ring->count;
5474
		i--;
5475 5476 5477 5478
		tx_buffer_info = &tx_ring->tx_buffer_info[i];
		ixgbe_unmap_and_free_tx_resource(adapter, tx_buffer_info);
	}

5479
	return 0;
5480 5481 5482
}

static void ixgbe_tx_queue(struct ixgbe_adapter *adapter,
5483 5484
                           struct ixgbe_ring *tx_ring,
                           int tx_flags, int count, u32 paylen, u8 hdr_len)
5485 5486 5487 5488 5489 5490 5491 5492 5493 5494 5495 5496 5497 5498 5499 5500 5501 5502
{
	union ixgbe_adv_tx_desc *tx_desc = NULL;
	struct ixgbe_tx_buffer *tx_buffer_info;
	u32 olinfo_status = 0, cmd_type_len = 0;
	unsigned int i;
	u32 txd_cmd = IXGBE_TXD_CMD_EOP | IXGBE_TXD_CMD_RS | IXGBE_TXD_CMD_IFCS;

	cmd_type_len |= IXGBE_ADVTXD_DTYP_DATA;

	cmd_type_len |= IXGBE_ADVTXD_DCMD_IFCS | IXGBE_ADVTXD_DCMD_DEXT;

	if (tx_flags & IXGBE_TX_FLAGS_VLAN)
		cmd_type_len |= IXGBE_ADVTXD_DCMD_VLE;

	if (tx_flags & IXGBE_TX_FLAGS_TSO) {
		cmd_type_len |= IXGBE_ADVTXD_DCMD_TSE;

		olinfo_status |= IXGBE_TXD_POPTS_TXSM <<
5503
		                 IXGBE_ADVTXD_POPTS_SHIFT;
5504

5505 5506
		/* use index 1 context for tso */
		olinfo_status |= (1 << IXGBE_ADVTXD_IDX_SHIFT);
5507 5508
		if (tx_flags & IXGBE_TX_FLAGS_IPV4)
			olinfo_status |= IXGBE_TXD_POPTS_IXSM <<
5509
			                 IXGBE_ADVTXD_POPTS_SHIFT;
5510 5511 5512

	} else if (tx_flags & IXGBE_TX_FLAGS_CSUM)
		olinfo_status |= IXGBE_TXD_POPTS_TXSM <<
5513
		                 IXGBE_ADVTXD_POPTS_SHIFT;
5514

5515 5516 5517 5518 5519 5520 5521
	if (tx_flags & IXGBE_TX_FLAGS_FCOE) {
		olinfo_status |= IXGBE_ADVTXD_CC;
		olinfo_status |= (1 << IXGBE_ADVTXD_IDX_SHIFT);
		if (tx_flags & IXGBE_TX_FLAGS_FSO)
			cmd_type_len |= IXGBE_ADVTXD_DCMD_TSE;
	}

5522 5523 5524 5525 5526 5527 5528 5529
	olinfo_status |= ((paylen - hdr_len) << IXGBE_ADVTXD_PAYLEN_SHIFT);

	i = tx_ring->next_to_use;
	while (count--) {
		tx_buffer_info = &tx_ring->tx_buffer_info[i];
		tx_desc = IXGBE_TX_DESC_ADV(*tx_ring, i);
		tx_desc->read.buffer_addr = cpu_to_le64(tx_buffer_info->dma);
		tx_desc->read.cmd_type_len =
5530
		        cpu_to_le32(cmd_type_len | tx_buffer_info->length);
5531 5532 5533 5534 5535 5536 5537 5538 5539 5540 5541 5542 5543 5544 5545 5546 5547 5548 5549 5550
		tx_desc->read.olinfo_status = cpu_to_le32(olinfo_status);
		i++;
		if (i == tx_ring->count)
			i = 0;
	}

	tx_desc->read.cmd_type_len |= cpu_to_le32(txd_cmd);

	/*
	 * Force memory writes to complete before letting h/w
	 * know there are new descriptors to fetch.  (Only
	 * applicable for weak-ordered memory model archs,
	 * such as IA-64).
	 */
	wmb();

	tx_ring->next_to_use = i;
	writel(i, adapter->hw.hw_addr + tx_ring->tail);
}

5551 5552 5553 5554 5555 5556 5557 5558 5559 5560 5561 5562 5563 5564 5565 5566 5567 5568 5569 5570 5571 5572 5573 5574 5575 5576 5577 5578 5579 5580 5581 5582 5583 5584 5585 5586 5587 5588 5589 5590 5591 5592 5593 5594 5595
static void ixgbe_atr(struct ixgbe_adapter *adapter, struct sk_buff *skb,
	              int queue, u32 tx_flags)
{
	/* Right now, we support IPv4 only */
	struct ixgbe_atr_input atr_input;
	struct tcphdr *th;
	struct iphdr *iph = ip_hdr(skb);
	struct ethhdr *eth = (struct ethhdr *)skb->data;
	u16 vlan_id, src_port, dst_port, flex_bytes;
	u32 src_ipv4_addr, dst_ipv4_addr;
	u8 l4type = 0;

	/* check if we're UDP or TCP */
	if (iph->protocol == IPPROTO_TCP) {
		th = tcp_hdr(skb);
		src_port = th->source;
		dst_port = th->dest;
		l4type |= IXGBE_ATR_L4TYPE_TCP;
		/* l4type IPv4 type is 0, no need to assign */
	} else {
		/* Unsupported L4 header, just bail here */
		return;
	}

	memset(&atr_input, 0, sizeof(struct ixgbe_atr_input));

	vlan_id = (tx_flags & IXGBE_TX_FLAGS_VLAN_MASK) >>
	           IXGBE_TX_FLAGS_VLAN_SHIFT;
	src_ipv4_addr = iph->saddr;
	dst_ipv4_addr = iph->daddr;
	flex_bytes = eth->h_proto;

	ixgbe_atr_set_vlan_id_82599(&atr_input, vlan_id);
	ixgbe_atr_set_src_port_82599(&atr_input, dst_port);
	ixgbe_atr_set_dst_port_82599(&atr_input, src_port);
	ixgbe_atr_set_flex_byte_82599(&atr_input, flex_bytes);
	ixgbe_atr_set_l4type_82599(&atr_input, l4type);
	/* src and dst are inverted, think how the receiver sees them */
	ixgbe_atr_set_src_ipv4_82599(&atr_input, dst_ipv4_addr);
	ixgbe_atr_set_dst_ipv4_82599(&atr_input, src_ipv4_addr);

	/* This assumes the Rx queue and Tx queue are bound to the same CPU */
	ixgbe_fdir_add_signature_filter_82599(&adapter->hw, &atr_input, queue);
}

5596
static int __ixgbe_maybe_stop_tx(struct net_device *netdev,
5597
                                 struct ixgbe_ring *tx_ring, int size)
5598
{
5599
	netif_stop_subqueue(netdev, tx_ring->queue_index);
5600 5601 5602 5603 5604 5605 5606 5607 5608 5609 5610
	/* Herbert's original patch had:
	 *  smp_mb__after_netif_stop_queue();
	 * but since that doesn't exist yet, just open code it. */
	smp_mb();

	/* We need to check again in a case another CPU has just
	 * made room available. */
	if (likely(IXGBE_DESC_UNUSED(tx_ring) < size))
		return -EBUSY;

	/* A reprieve! - use start_queue because it doesn't call schedule */
5611
	netif_start_subqueue(netdev, tx_ring->queue_index);
J
Jesse Brandeburg 已提交
5612
	++tx_ring->restart_queue;
5613 5614 5615 5616
	return 0;
}

static int ixgbe_maybe_stop_tx(struct net_device *netdev,
5617
                              struct ixgbe_ring *tx_ring, int size)
5618 5619 5620 5621 5622 5623
{
	if (likely(IXGBE_DESC_UNUSED(tx_ring) >= size))
		return 0;
	return __ixgbe_maybe_stop_tx(netdev, tx_ring, size);
}

5624 5625 5626
static u16 ixgbe_select_queue(struct net_device *dev, struct sk_buff *skb)
{
	struct ixgbe_adapter *adapter = netdev_priv(dev);
5627
	int txq = smp_processor_id();
5628

K
Krishna Kumar 已提交
5629 5630 5631
	if (adapter->flags & IXGBE_FLAG_FDIR_HASH_CAPABLE) {
		while (unlikely(txq >= dev->real_num_tx_queues))
			txq -= dev->real_num_tx_queues;
5632
		return txq;
K
Krishna Kumar 已提交
5633
	}
5634

5635 5636 5637 5638 5639 5640 5641 5642
#ifdef IXGBE_FCOE
	if ((adapter->flags & IXGBE_FLAG_FCOE_ENABLED) &&
	    (skb->protocol == htons(ETH_P_FCOE))) {
		txq &= (adapter->ring_feature[RING_F_FCOE].indices - 1);
		txq += adapter->ring_feature[RING_F_FCOE].mask;
		return txq;
	}
#endif
5643 5644 5645 5646 5647 5648 5649 5650
	if (adapter->flags & IXGBE_FLAG_DCB_ENABLED) {
		if (skb->priority == TC_PRIO_CONTROL)
			txq = adapter->ring_feature[RING_F_DCB].indices-1;
		else
			txq = (skb->vlan_tci & IXGBE_TX_FLAGS_VLAN_PRIO_MASK)
			       >> 13;
		return txq;
	}
5651 5652 5653 5654

	return skb_tx_hash(dev, skb);
}

5655 5656
static netdev_tx_t ixgbe_xmit_frame(struct sk_buff *skb,
				    struct net_device *netdev)
5657 5658 5659
{
	struct ixgbe_adapter *adapter = netdev_priv(netdev);
	struct ixgbe_ring *tx_ring;
E
Eric Dumazet 已提交
5660
	struct netdev_queue *txq;
5661 5662
	unsigned int first;
	unsigned int tx_flags = 0;
5663
	u8 hdr_len = 0;
5664
	int tso;
5665 5666
	int count = 0;
	unsigned int f;
J
Jesse Brandeburg 已提交
5667 5668 5669

	if (adapter->vlgrp && vlan_tx_tag_present(skb)) {
		tx_flags |= vlan_tx_tag_get(skb);
5670 5671
		if (adapter->flags & IXGBE_FLAG_DCB_ENABLED) {
			tx_flags &= ~IXGBE_TX_FLAGS_VLAN_PRIO_MASK;
5672
			tx_flags |= ((skb->queue_mapping & 0x7) << 13);
5673 5674 5675 5676
		}
		tx_flags <<= IXGBE_TX_FLAGS_VLAN_SHIFT;
		tx_flags |= IXGBE_TX_FLAGS_VLAN;
	} else if (adapter->flags & IXGBE_FLAG_DCB_ENABLED) {
5677 5678 5679
		tx_flags |= ((skb->queue_mapping & 0x7) << 13);
		tx_flags <<= IXGBE_TX_FLAGS_VLAN_SHIFT;
		tx_flags |= IXGBE_TX_FLAGS_VLAN;
5680
	}
5681

5682
	tx_ring = adapter->tx_ring[skb->queue_mapping];
5683

5684
	if ((adapter->flags & IXGBE_FLAG_FCOE_ENABLED) &&
5685
	    (skb->protocol == htons(ETH_P_FCOE))) {
5686
		tx_flags |= IXGBE_TX_FLAGS_FCOE;
5687
#ifdef IXGBE_FCOE
5688 5689 5690 5691 5692 5693
#ifdef CONFIG_IXGBE_DCB
		tx_flags &= ~(IXGBE_TX_FLAGS_VLAN_PRIO_MASK
			      << IXGBE_TX_FLAGS_VLAN_SHIFT);
		tx_flags |= ((adapter->fcoe.up << 13)
			      << IXGBE_TX_FLAGS_VLAN_SHIFT);
#endif
5694 5695
#endif
	}
5696
	/* four things can cause us to need a context descriptor */
J
Jesse Brandeburg 已提交
5697 5698
	if (skb_is_gso(skb) ||
	    (skb->ip_summed == CHECKSUM_PARTIAL) ||
5699 5700
	    (tx_flags & IXGBE_TX_FLAGS_VLAN) ||
	    (tx_flags & IXGBE_TX_FLAGS_FCOE))
5701 5702
		count++;

J
Jesse Brandeburg 已提交
5703 5704
	count += TXD_USE_COUNT(skb_headlen(skb));
	for (f = 0; f < skb_shinfo(skb)->nr_frags; f++)
5705 5706
		count += TXD_USE_COUNT(skb_shinfo(skb)->frags[f].size);

5707
	if (ixgbe_maybe_stop_tx(netdev, tx_ring, count)) {
5708 5709 5710 5711 5712
		adapter->tx_busy++;
		return NETDEV_TX_BUSY;
	}

	first = tx_ring->next_to_use;
5713 5714 5715 5716 5717 5718 5719 5720 5721 5722 5723 5724 5725 5726 5727 5728 5729 5730 5731
	if (tx_flags & IXGBE_TX_FLAGS_FCOE) {
#ifdef IXGBE_FCOE
		/* setup tx offload for FCoE */
		tso = ixgbe_fso(adapter, tx_ring, skb, tx_flags, &hdr_len);
		if (tso < 0) {
			dev_kfree_skb_any(skb);
			return NETDEV_TX_OK;
		}
		if (tso)
			tx_flags |= IXGBE_TX_FLAGS_FSO;
#endif /* IXGBE_FCOE */
	} else {
		if (skb->protocol == htons(ETH_P_IP))
			tx_flags |= IXGBE_TX_FLAGS_IPV4;
		tso = ixgbe_tso(adapter, tx_ring, skb, tx_flags, &hdr_len);
		if (tso < 0) {
			dev_kfree_skb_any(skb);
			return NETDEV_TX_OK;
		}
5732

5733 5734 5735 5736 5737 5738
		if (tso)
			tx_flags |= IXGBE_TX_FLAGS_TSO;
		else if (ixgbe_tx_csum(adapter, tx_ring, skb, tx_flags) &&
			 (skb->ip_summed == CHECKSUM_PARTIAL))
			tx_flags |= IXGBE_TX_FLAGS_CSUM;
	}
5739

5740
	count = ixgbe_tx_map(adapter, tx_ring, skb, tx_flags, first);
5741
	if (count) {
5742 5743 5744 5745 5746 5747 5748 5749 5750 5751 5752
		/* add the ATR filter if ATR is on */
		if (tx_ring->atr_sample_rate) {
			++tx_ring->atr_count;
			if ((tx_ring->atr_count >= tx_ring->atr_sample_rate) &&
		             test_bit(__IXGBE_FDIR_INIT_DONE,
                                      &tx_ring->reinit_state)) {
				ixgbe_atr(adapter, skb, tx_ring->queue_index,
				          tx_flags);
				tx_ring->atr_count = 0;
			}
		}
E
Eric Dumazet 已提交
5753 5754 5755
		txq = netdev_get_tx_queue(netdev, tx_ring->queue_index);
		txq->tx_bytes += skb->len;
		txq->tx_packets++;
5756 5757 5758
		ixgbe_tx_queue(adapter, tx_ring, tx_flags, count, skb->len,
		               hdr_len);
		ixgbe_maybe_stop_tx(netdev, tx_ring, DESC_NEEDED);
5759

5760 5761 5762 5763 5764
	} else {
		dev_kfree_skb_any(skb);
		tx_ring->tx_buffer_info[first].time_stamp = 0;
		tx_ring->next_to_use = first;
	}
5765 5766 5767 5768 5769 5770 5771 5772 5773 5774 5775 5776 5777 5778

	return NETDEV_TX_OK;
}

/**
 * ixgbe_set_mac - Change the Ethernet Address of the NIC
 * @netdev: network interface device structure
 * @p: pointer to an address structure
 *
 * Returns 0 on success, negative on failure
 **/
static int ixgbe_set_mac(struct net_device *netdev, void *p)
{
	struct ixgbe_adapter *adapter = netdev_priv(netdev);
5779
	struct ixgbe_hw *hw = &adapter->hw;
5780 5781 5782 5783 5784 5785
	struct sockaddr *addr = p;

	if (!is_valid_ether_addr(addr->sa_data))
		return -EADDRNOTAVAIL;

	memcpy(netdev->dev_addr, addr->sa_data, netdev->addr_len);
5786
	memcpy(hw->mac.addr, addr->sa_data, netdev->addr_len);
5787

5788 5789
	hw->mac.ops.set_rar(hw, 0, hw->mac.addr, adapter->num_vfs,
			    IXGBE_RAH_AV);
5790 5791 5792 5793

	return 0;
}

5794 5795 5796 5797 5798 5799 5800 5801 5802 5803 5804 5805 5806 5807 5808 5809 5810 5811 5812 5813 5814 5815 5816 5817 5818 5819 5820 5821 5822 5823 5824 5825 5826 5827
static int
ixgbe_mdio_read(struct net_device *netdev, int prtad, int devad, u16 addr)
{
	struct ixgbe_adapter *adapter = netdev_priv(netdev);
	struct ixgbe_hw *hw = &adapter->hw;
	u16 value;
	int rc;

	if (prtad != hw->phy.mdio.prtad)
		return -EINVAL;
	rc = hw->phy.ops.read_reg(hw, addr, devad, &value);
	if (!rc)
		rc = value;
	return rc;
}

static int ixgbe_mdio_write(struct net_device *netdev, int prtad, int devad,
			    u16 addr, u16 value)
{
	struct ixgbe_adapter *adapter = netdev_priv(netdev);
	struct ixgbe_hw *hw = &adapter->hw;

	if (prtad != hw->phy.mdio.prtad)
		return -EINVAL;
	return hw->phy.ops.write_reg(hw, addr, devad, value);
}

static int ixgbe_ioctl(struct net_device *netdev, struct ifreq *req, int cmd)
{
	struct ixgbe_adapter *adapter = netdev_priv(netdev);

	return mdio_mii_ioctl(&adapter->hw.phy.mdio, if_mii(req), cmd);
}

5828 5829
/**
 * ixgbe_add_sanmac_netdev - Add the SAN MAC address to the corresponding
5830
 * netdev->dev_addrs
5831 5832 5833 5834 5835 5836 5837 5838 5839 5840 5841 5842 5843 5844 5845 5846 5847 5848 5849 5850
 * @netdev: network interface device structure
 *
 * Returns non-zero on failure
 **/
static int ixgbe_add_sanmac_netdev(struct net_device *dev)
{
	int err = 0;
	struct ixgbe_adapter *adapter = netdev_priv(dev);
	struct ixgbe_mac_info *mac = &adapter->hw.mac;

	if (is_valid_ether_addr(mac->san_addr)) {
		rtnl_lock();
		err = dev_addr_add(dev, mac->san_addr, NETDEV_HW_ADDR_T_SAN);
		rtnl_unlock();
	}
	return err;
}

/**
 * ixgbe_del_sanmac_netdev - Removes the SAN MAC address to the corresponding
5851
 * netdev->dev_addrs
5852 5853 5854 5855 5856 5857 5858 5859 5860 5861 5862 5863 5864 5865 5866 5867 5868 5869
 * @netdev: network interface device structure
 *
 * Returns non-zero on failure
 **/
static int ixgbe_del_sanmac_netdev(struct net_device *dev)
{
	int err = 0;
	struct ixgbe_adapter *adapter = netdev_priv(dev);
	struct ixgbe_mac_info *mac = &adapter->hw.mac;

	if (is_valid_ether_addr(mac->san_addr)) {
		rtnl_lock();
		err = dev_addr_del(dev, mac->san_addr, NETDEV_HW_ADDR_T_SAN);
		rtnl_unlock();
	}
	return err;
}

5870 5871 5872 5873 5874 5875 5876 5877 5878
#ifdef CONFIG_NET_POLL_CONTROLLER
/*
 * Polling 'interrupt' - used by things like netconsole to send skbs
 * without having to re-enable interrupts. It's not called while
 * the interrupt routine is executing.
 */
static void ixgbe_netpoll(struct net_device *netdev)
{
	struct ixgbe_adapter *adapter = netdev_priv(netdev);
5879
	int i;
5880

5881 5882 5883 5884
	/* if interface is down do nothing */
	if (test_bit(__IXGBE_DOWN, &adapter->state))
		return;

5885
	adapter->flags |= IXGBE_FLAG_IN_NETPOLL;
5886 5887 5888 5889 5890 5891 5892 5893 5894
	if (adapter->flags & IXGBE_FLAG_MSIX_ENABLED) {
		int num_q_vectors = adapter->num_msix_vectors - NON_Q_VECTORS;
		for (i = 0; i < num_q_vectors; i++) {
			struct ixgbe_q_vector *q_vector = adapter->q_vector[i];
			ixgbe_msix_clean_many(0, q_vector);
		}
	} else {
		ixgbe_intr(adapter->pdev->irq, netdev);
	}
5895 5896 5897 5898
	adapter->flags &= ~IXGBE_FLAG_IN_NETPOLL;
}
#endif

5899 5900 5901
static const struct net_device_ops ixgbe_netdev_ops = {
	.ndo_open 		= ixgbe_open,
	.ndo_stop		= ixgbe_close,
5902
	.ndo_start_xmit		= ixgbe_xmit_frame,
5903
	.ndo_select_queue	= ixgbe_select_queue,
5904
	.ndo_set_rx_mode        = ixgbe_set_rx_mode,
5905 5906 5907 5908 5909 5910 5911 5912
	.ndo_set_multicast_list	= ixgbe_set_rx_mode,
	.ndo_validate_addr	= eth_validate_addr,
	.ndo_set_mac_address	= ixgbe_set_mac,
	.ndo_change_mtu		= ixgbe_change_mtu,
	.ndo_tx_timeout		= ixgbe_tx_timeout,
	.ndo_vlan_rx_register	= ixgbe_vlan_rx_register,
	.ndo_vlan_rx_add_vid	= ixgbe_vlan_rx_add_vid,
	.ndo_vlan_rx_kill_vid	= ixgbe_vlan_rx_kill_vid,
5913
	.ndo_do_ioctl		= ixgbe_ioctl,
5914 5915 5916
#ifdef CONFIG_NET_POLL_CONTROLLER
	.ndo_poll_controller	= ixgbe_netpoll,
#endif
5917 5918 5919
#ifdef IXGBE_FCOE
	.ndo_fcoe_ddp_setup = ixgbe_fcoe_ddp_get,
	.ndo_fcoe_ddp_done = ixgbe_fcoe_ddp_put,
5920 5921
	.ndo_fcoe_enable = ixgbe_fcoe_enable,
	.ndo_fcoe_disable = ixgbe_fcoe_disable,
5922
	.ndo_fcoe_get_wwn = ixgbe_fcoe_get_wwn,
5923
#endif /* IXGBE_FCOE */
5924 5925
};

5926 5927 5928 5929 5930 5931 5932 5933 5934 5935 5936 5937 5938 5939 5940 5941 5942 5943 5944 5945 5946 5947 5948 5949 5950 5951 5952 5953 5954 5955 5956 5957 5958 5959 5960 5961 5962 5963 5964 5965 5966 5967 5968 5969 5970 5971 5972 5973 5974 5975 5976 5977 5978 5979 5980
static void __devinit ixgbe_probe_vf(struct ixgbe_adapter *adapter,
			   const struct ixgbe_info *ii)
{
#ifdef CONFIG_PCI_IOV
	struct ixgbe_hw *hw = &adapter->hw;
	int err;

	if (hw->mac.type != ixgbe_mac_82599EB || !max_vfs)
		return;

	/* The 82599 supports up to 64 VFs per physical function
	 * but this implementation limits allocation to 63 so that
	 * basic networking resources are still available to the
	 * physical function
	 */
	adapter->num_vfs = (max_vfs > 63) ? 63 : max_vfs;
	adapter->flags |= IXGBE_FLAG_SRIOV_ENABLED;
	err = pci_enable_sriov(adapter->pdev, adapter->num_vfs);
	if (err) {
		DPRINTK(PROBE, ERR,
			"Failed to enable PCI sriov: %d\n", err);
		goto err_novfs;
	}
	/* If call to enable VFs succeeded then allocate memory
	 * for per VF control structures.
	 */
	adapter->vfinfo =
		kcalloc(adapter->num_vfs,
			sizeof(struct vf_data_storage), GFP_KERNEL);
	if (adapter->vfinfo) {
		/* Now that we're sure SR-IOV is enabled
		 * and memory allocated set up the mailbox parameters
		 */
		ixgbe_init_mbx_params_pf(hw);
		memcpy(&hw->mbx.ops, ii->mbx_ops,
		       sizeof(hw->mbx.ops));

		/* Disable RSC when in SR-IOV mode */
		adapter->flags2 &= ~(IXGBE_FLAG2_RSC_CAPABLE |
				     IXGBE_FLAG2_RSC_ENABLED);
		return;
	}

	/* Oh oh */
	DPRINTK(PROBE, ERR,
		"Unable to allocate memory for VF "
		"Data Storage - SRIOV disabled\n");
	pci_disable_sriov(adapter->pdev);

err_novfs:
	adapter->flags &= ~IXGBE_FLAG_SRIOV_ENABLED;
	adapter->num_vfs = 0;
#endif /* CONFIG_PCI_IOV */
}

5981 5982 5983 5984 5985 5986 5987 5988 5989 5990 5991 5992
/**
 * ixgbe_probe - Device Initialization Routine
 * @pdev: PCI device information struct
 * @ent: entry in ixgbe_pci_tbl
 *
 * Returns 0 on success, negative on failure
 *
 * ixgbe_probe initializes an adapter identified by a pci_dev structure.
 * The OS initialization, configuring of the adapter private structure,
 * and a hardware reset occur.
 **/
static int __devinit ixgbe_probe(struct pci_dev *pdev,
5993
                                 const struct pci_device_id *ent)
5994 5995 5996 5997 5998 5999 6000
{
	struct net_device *netdev;
	struct ixgbe_adapter *adapter = NULL;
	struct ixgbe_hw *hw;
	const struct ixgbe_info *ii = ixgbe_info_tbl[ent->driver_data];
	static int cards_found;
	int i, err, pci_using_dac;
6001
	unsigned int indices = num_possible_cpus();
6002 6003 6004
#ifdef IXGBE_FCOE
	u16 device_caps;
#endif
6005
	u32 part_num, eec;
6006

6007
	err = pci_enable_device_mem(pdev);
6008 6009 6010
	if (err)
		return err;

6011 6012
	if (!pci_set_dma_mask(pdev, DMA_BIT_MASK(64)) &&
	    !pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(64))) {
6013 6014
		pci_using_dac = 1;
	} else {
6015
		err = pci_set_dma_mask(pdev, DMA_BIT_MASK(32));
6016
		if (err) {
6017
			err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(32));
6018
			if (err) {
6019 6020
				dev_err(&pdev->dev, "No usable DMA "
				        "configuration, aborting\n");
6021 6022 6023 6024 6025 6026
				goto err_dma;
			}
		}
		pci_using_dac = 0;
	}

6027 6028
	err = pci_request_selected_regions(pdev, pci_select_bars(pdev,
	                                   IORESOURCE_MEM), ixgbe_driver_name);
6029
	if (err) {
6030 6031
		dev_err(&pdev->dev,
		        "pci_request_selected_regions failed 0x%x\n", err);
6032 6033 6034
		goto err_pci_reg;
	}

6035
	pci_enable_pcie_error_reporting(pdev);
6036

6037
	pci_set_master(pdev);
6038
	pci_save_state(pdev);
6039

6040 6041 6042 6043 6044 6045 6046 6047 6048 6049 6050 6051
	if (ii->mac == ixgbe_mac_82598EB)
		indices = min_t(unsigned int, indices, IXGBE_MAX_RSS_INDICES);
	else
		indices = min_t(unsigned int, indices, IXGBE_MAX_FDIR_INDICES);

	indices = max_t(unsigned int, indices, IXGBE_MAX_DCB_INDICES);
#ifdef IXGBE_FCOE
	indices += min_t(unsigned int, num_possible_cpus(),
			 IXGBE_MAX_FCOE_INDICES);
#endif
	indices = min_t(unsigned int, indices, MAX_TX_QUEUES);
	netdev = alloc_etherdev_mq(sizeof(struct ixgbe_adapter), indices);
6052 6053 6054 6055 6056 6057 6058 6059 6060 6061 6062 6063 6064 6065 6066 6067
	if (!netdev) {
		err = -ENOMEM;
		goto err_alloc_etherdev;
	}

	SET_NETDEV_DEV(netdev, &pdev->dev);

	pci_set_drvdata(pdev, netdev);
	adapter = netdev_priv(netdev);

	adapter->netdev = netdev;
	adapter->pdev = pdev;
	hw = &adapter->hw;
	hw->back = adapter;
	adapter->msg_enable = (1 << DEFAULT_DEBUG_LEVEL_SHIFT) - 1;

6068 6069
	hw->hw_addr = ioremap(pci_resource_start(pdev, 0),
	                      pci_resource_len(pdev, 0));
6070 6071 6072 6073 6074 6075 6076 6077 6078 6079
	if (!hw->hw_addr) {
		err = -EIO;
		goto err_ioremap;
	}

	for (i = 1; i <= 5; i++) {
		if (pci_resource_len(pdev, i) == 0)
			continue;
	}

6080
	netdev->netdev_ops = &ixgbe_netdev_ops;
6081 6082 6083 6084 6085 6086 6087 6088
	ixgbe_set_ethtool_ops(netdev);
	netdev->watchdog_timeo = 5 * HZ;
	strcpy(netdev->name, pci_name(pdev));

	adapter->bd_number = cards_found;

	/* Setup hw api */
	memcpy(&hw->mac.ops, ii->mac_ops, sizeof(hw->mac.ops));
6089
	hw->mac.type  = ii->mac;
6090

6091 6092 6093 6094 6095 6096 6097 6098 6099
	/* EEPROM */
	memcpy(&hw->eeprom.ops, ii->eeprom_ops, sizeof(hw->eeprom.ops));
	eec = IXGBE_READ_REG(hw, IXGBE_EEC);
	/* If EEPROM is valid (bit 8 = 1), use default otherwise use bit bang */
	if (!(eec & (1 << 8)))
		hw->eeprom.ops.read = &ixgbe_read_eeprom_bit_bang_generic;

	/* PHY */
	memcpy(&hw->phy.ops, ii->phy_ops, sizeof(hw->phy.ops));
D
Donald Skidmore 已提交
6100
	hw->phy.sfp_type = ixgbe_sfp_type_unknown;
6101 6102 6103 6104 6105 6106 6107
	/* ixgbe_identify_phy_generic will set prtad and mmds properly */
	hw->phy.mdio.prtad = MDIO_PRTAD_NONE;
	hw->phy.mdio.mmds = 0;
	hw->phy.mdio.mode_support = MDIO_SUPPORTS_C45 | MDIO_EMULATE_C22;
	hw->phy.mdio.dev = netdev;
	hw->phy.mdio.mdio_read = ixgbe_mdio_read;
	hw->phy.mdio.mdio_write = ixgbe_mdio_write;
D
Donald Skidmore 已提交
6108 6109 6110 6111 6112 6113 6114 6115 6116

	/* set up this timer and work struct before calling get_invariants
	 * which might start the timer
	 */
	init_timer(&adapter->sfp_timer);
	adapter->sfp_timer.function = &ixgbe_sfp_timer;
	adapter->sfp_timer.data = (unsigned long) adapter;

	INIT_WORK(&adapter->sfp_task, ixgbe_sfp_task);
6117

6118 6119 6120 6121 6122 6123 6124
	/* multispeed fiber has its own tasklet, called from GPI SDP1 context */
	INIT_WORK(&adapter->multispeed_fiber_task, ixgbe_multispeed_fiber_task);

	/* a new SFP+ module arrival, called from GPI SDP2 context */
	INIT_WORK(&adapter->sfp_config_module_task,
	          ixgbe_sfp_config_module_task);

6125
	ii->get_invariants(hw);
6126 6127 6128 6129 6130 6131

	/* setup the private structure */
	err = ixgbe_sw_init(adapter);
	if (err)
		goto err_sw_init;

6132 6133 6134 6135
	/* Make it possible the adapter to be woken up via WOL */
	if (adapter->hw.mac.type == ixgbe_mac_82599EB)
		IXGBE_WRITE_REG(&adapter->hw, IXGBE_WUS, ~0);

6136 6137 6138 6139 6140 6141 6142 6143 6144 6145 6146
	/*
	 * If there is a fan on this device and it has failed log the
	 * failure.
	 */
	if (adapter->flags & IXGBE_FLAG_FAN_FAIL_CAPABLE) {
		u32 esdp = IXGBE_READ_REG(hw, IXGBE_ESDP);
		if (esdp & IXGBE_ESDP_SDP1)
			DPRINTK(PROBE, CRIT,
				"Fan has stopped, replace the adapter\n");
	}

6147 6148
	/* reset_hw fills in the perm_addr as well */
	err = hw->mac.ops.reset_hw(hw);
6149 6150 6151 6152 6153 6154 6155 6156 6157 6158 6159 6160
	if (err == IXGBE_ERR_SFP_NOT_PRESENT &&
	    hw->mac.type == ixgbe_mac_82598EB) {
		/*
		 * Start a kernel thread to watch for a module to arrive.
		 * Only do this for 82598, since 82599 will generate
		 * interrupts on module arrival.
		 */
		set_bit(__IXGBE_SFP_MODULE_NOT_FOUND, &adapter->state);
		mod_timer(&adapter->sfp_timer,
			  round_jiffies(jiffies + (2 * HZ)));
		err = 0;
	} else if (err == IXGBE_ERR_SFP_NOT_SUPPORTED) {
D
Don Skidmore 已提交
6161 6162 6163 6164
		dev_err(&adapter->pdev->dev, "failed to initialize because "
			"an unsupported SFP+ module type was detected.\n"
			"Reload the driver after installing a supported "
			"module.\n");
6165 6166
		goto err_sw_init;
	} else if (err) {
6167 6168 6169 6170
		dev_err(&adapter->pdev->dev, "HW Init failed: %d\n", err);
		goto err_sw_init;
	}

6171 6172
	ixgbe_probe_vf(adapter, ii);

6173
	netdev->features = NETIF_F_SG |
6174 6175 6176 6177
	                   NETIF_F_IP_CSUM |
	                   NETIF_F_HW_VLAN_TX |
	                   NETIF_F_HW_VLAN_RX |
	                   NETIF_F_HW_VLAN_FILTER;
6178

6179
	netdev->features |= NETIF_F_IPV6_CSUM;
6180 6181
	netdev->features |= NETIF_F_TSO;
	netdev->features |= NETIF_F_TSO6;
H
Herbert Xu 已提交
6182
	netdev->features |= NETIF_F_GRO;
6183

6184 6185 6186
	if (adapter->hw.mac.type == ixgbe_mac_82599EB)
		netdev->features |= NETIF_F_SCTP_CSUM;

6187 6188
	netdev->vlan_features |= NETIF_F_TSO;
	netdev->vlan_features |= NETIF_F_TSO6;
6189
	netdev->vlan_features |= NETIF_F_IP_CSUM;
6190
	netdev->vlan_features |= NETIF_F_IPV6_CSUM;
6191 6192
	netdev->vlan_features |= NETIF_F_SG;

6193 6194 6195
	if (adapter->flags & IXGBE_FLAG_SRIOV_ENABLED)
		adapter->flags &= ~(IXGBE_FLAG_RSS_ENABLED |
				    IXGBE_FLAG_DCB_ENABLED);
6196 6197 6198
	if (adapter->flags & IXGBE_FLAG_DCB_ENABLED)
		adapter->flags &= ~IXGBE_FLAG_RSS_ENABLED;

J
Jeff Kirsher 已提交
6199
#ifdef CONFIG_IXGBE_DCB
6200 6201 6202
	netdev->dcbnl_ops = &dcbnl_ops;
#endif

6203
#ifdef IXGBE_FCOE
6204
	if (adapter->flags & IXGBE_FLAG_FCOE_CAPABLE) {
6205 6206
		if (hw->mac.ops.get_device_caps) {
			hw->mac.ops.get_device_caps(hw, &device_caps);
6207 6208
			if (device_caps & IXGBE_DEVICE_CAPS_FCOE_OFFLOADS)
				adapter->flags &= ~IXGBE_FLAG_FCOE_CAPABLE;
6209 6210 6211
		}
	}
#endif /* IXGBE_FCOE */
6212 6213 6214
	if (pci_using_dac)
		netdev->features |= NETIF_F_HIGHDMA;

6215
	if (adapter->flags2 & IXGBE_FLAG2_RSC_ENABLED)
A
Alexander Duyck 已提交
6216 6217
		netdev->features |= NETIF_F_LRO;

6218
	/* make sure the EEPROM is good */
6219
	if (hw->eeprom.ops.validate_checksum(hw, NULL) < 0) {
6220 6221 6222 6223 6224 6225 6226 6227
		dev_err(&pdev->dev, "The EEPROM Checksum Is Not Valid\n");
		err = -EIO;
		goto err_eeprom;
	}

	memcpy(netdev->dev_addr, hw->mac.perm_addr, netdev->addr_len);
	memcpy(netdev->perm_addr, hw->mac.perm_addr, netdev->addr_len);

6228 6229
	if (ixgbe_validate_mac_addr(netdev->perm_addr)) {
		dev_err(&pdev->dev, "invalid MAC address\n");
6230 6231 6232 6233 6234 6235 6236 6237 6238
		err = -EIO;
		goto err_eeprom;
	}

	init_timer(&adapter->watchdog_timer);
	adapter->watchdog_timer.function = &ixgbe_watchdog;
	adapter->watchdog_timer.data = (unsigned long)adapter;

	INIT_WORK(&adapter->reset_task, ixgbe_reset_task);
6239
	INIT_WORK(&adapter->watchdog_task, ixgbe_watchdog_task);
6240

6241 6242 6243
	err = ixgbe_init_interrupt_scheme(adapter);
	if (err)
		goto err_sw_init;
6244

6245 6246
	switch (pdev->device) {
	case IXGBE_DEV_ID_82599_KX4:
6247 6248
		adapter->wol = (IXGBE_WUFC_MAG | IXGBE_WUFC_EX |
		                IXGBE_WUFC_MC | IXGBE_WUFC_BC);
6249 6250 6251 6252 6253 6254 6255
		break;
	default:
		adapter->wol = 0;
		break;
	}
	device_set_wakeup_enable(&adapter->pdev->dev, adapter->wol);

6256 6257 6258
	/* pick up the PCI bus settings for reporting later */
	hw->mac.ops.get_bus_info(hw);

6259
	/* print bus type/speed/width info */
J
Johannes Berg 已提交
6260
	dev_info(&pdev->dev, "(PCI Express:%s:%s) %pM\n",
6261 6262 6263 6264 6265
	        ((hw->bus.speed == ixgbe_bus_speed_5000) ? "5.0Gb/s":
	         (hw->bus.speed == ixgbe_bus_speed_2500) ? "2.5Gb/s":"Unknown"),
	        ((hw->bus.width == ixgbe_bus_width_pcie_x8) ? "Width x8" :
	         (hw->bus.width == ixgbe_bus_width_pcie_x4) ? "Width x4" :
	         (hw->bus.width == ixgbe_bus_width_pcie_x1) ? "Width x1" :
6266
	         "Unknown"),
J
Johannes Berg 已提交
6267
	        netdev->dev_addr);
6268
	ixgbe_read_pba_num_generic(hw, &part_num);
6269 6270 6271 6272 6273 6274 6275 6276
	if (ixgbe_is_sfp(hw) && hw->phy.sfp_type != ixgbe_sfp_type_not_present)
		dev_info(&pdev->dev, "MAC: %d, PHY: %d, SFP+: %d, PBA No: %06x-%03x\n",
		         hw->mac.type, hw->phy.type, hw->phy.sfp_type,
		         (part_num >> 8), (part_num & 0xff));
	else
		dev_info(&pdev->dev, "MAC: %d, PHY: %d, PBA No: %06x-%03x\n",
		         hw->mac.type, hw->phy.type,
		         (part_num >> 8), (part_num & 0xff));
6277

6278
	if (hw->bus.width <= ixgbe_bus_width_pcie_x4) {
6279
		dev_warn(&pdev->dev, "PCI-Express bandwidth available for "
6280 6281
		         "this card is not sufficient for optimal "
		         "performance.\n");
6282
		dev_warn(&pdev->dev, "For optimal performance a x8 "
6283
		         "PCI-Express slot is required.\n");
6284 6285
	}

6286 6287 6288
	/* save off EEPROM version number */
	hw->eeprom.ops.read(hw, 0x29, &adapter->eeprom_version);

6289
	/* reset the hardware with the new settings */
6290
	err = hw->mac.ops.start_hw(hw);
6291

6292 6293 6294 6295 6296 6297 6298 6299 6300
	if (err == IXGBE_ERR_EEPROM_VERSION) {
		/* We are running on a pre-production device, log a warning */
		dev_warn(&pdev->dev, "This device is a pre-production "
		         "adapter/LOM.  Please be aware there may be issues "
		         "associated with your hardware.  If you are "
		         "experiencing problems please contact your Intel or "
		         "hardware representative who provided you with this "
		         "hardware.\n");
	}
6301 6302 6303 6304 6305
	strcpy(netdev->name, "eth%d");
	err = register_netdev(netdev);
	if (err)
		goto err_register;

6306 6307 6308
	/* carrier off reporting is important to ethtool even BEFORE open */
	netif_carrier_off(netdev);

6309 6310 6311 6312
	if (adapter->flags & IXGBE_FLAG_FDIR_HASH_CAPABLE ||
	    adapter->flags & IXGBE_FLAG_FDIR_PERFECT_CAPABLE)
		INIT_WORK(&adapter->fdir_reinit_task, ixgbe_fdir_reinit_task);

6313
#ifdef CONFIG_IXGBE_DCA
6314
	if (dca_add_requester(&pdev->dev) == 0) {
6315 6316 6317 6318
		adapter->flags |= IXGBE_FLAG_DCA_ENABLED;
		ixgbe_setup_dca(adapter);
	}
#endif
6319 6320 6321 6322 6323 6324 6325
	if (adapter->flags & IXGBE_FLAG_SRIOV_ENABLED) {
		DPRINTK(PROBE, INFO, "IOV is enabled with %d VFs\n",
			adapter->num_vfs);
		for (i = 0; i < adapter->num_vfs; i++)
			ixgbe_vf_configuration(pdev, (i | 0x10000000));
	}

6326 6327
	/* add san mac addr to netdev */
	ixgbe_add_sanmac_netdev(netdev);
6328 6329 6330 6331 6332 6333

	dev_info(&pdev->dev, "Intel(R) 10 Gigabit Network Connection\n");
	cards_found++;
	return 0;

err_register:
6334
	ixgbe_release_hw_control(adapter);
6335
	ixgbe_clear_interrupt_scheme(adapter);
6336 6337
err_sw_init:
err_eeprom:
6338 6339
	if (adapter->flags & IXGBE_FLAG_SRIOV_ENABLED)
		ixgbe_disable_sriov(adapter);
D
Donald Skidmore 已提交
6340 6341 6342
	clear_bit(__IXGBE_SFP_MODULE_NOT_FOUND, &adapter->state);
	del_timer_sync(&adapter->sfp_timer);
	cancel_work_sync(&adapter->sfp_task);
6343 6344
	cancel_work_sync(&adapter->multispeed_fiber_task);
	cancel_work_sync(&adapter->sfp_config_module_task);
6345 6346 6347 6348
	iounmap(hw->hw_addr);
err_ioremap:
	free_netdev(netdev);
err_alloc_etherdev:
6349 6350
	pci_release_selected_regions(pdev, pci_select_bars(pdev,
	                             IORESOURCE_MEM));
6351 6352 6353 6354 6355 6356 6357 6358 6359 6360 6361 6362 6363 6364 6365 6366 6367 6368 6369 6370 6371
err_pci_reg:
err_dma:
	pci_disable_device(pdev);
	return err;
}

/**
 * ixgbe_remove - Device Removal Routine
 * @pdev: PCI device information struct
 *
 * ixgbe_remove is called by the PCI subsystem to alert the driver
 * that it should release a PCI device.  The could be caused by a
 * Hot-Plug event, or because the driver is going to be removed from
 * memory.
 **/
static void __devexit ixgbe_remove(struct pci_dev *pdev)
{
	struct net_device *netdev = pci_get_drvdata(pdev);
	struct ixgbe_adapter *adapter = netdev_priv(netdev);

	set_bit(__IXGBE_DOWN, &adapter->state);
D
Donald Skidmore 已提交
6372 6373 6374 6375
	/* clear the module not found bit to make sure the worker won't
	 * reschedule
	 */
	clear_bit(__IXGBE_SFP_MODULE_NOT_FOUND, &adapter->state);
6376 6377
	del_timer_sync(&adapter->watchdog_timer);

D
Donald Skidmore 已提交
6378 6379 6380
	del_timer_sync(&adapter->sfp_timer);
	cancel_work_sync(&adapter->watchdog_task);
	cancel_work_sync(&adapter->sfp_task);
6381 6382 6383 6384 6385 6386 6387 6388 6389 6390
	if (adapter->hw.phy.multispeed_fiber) {
		struct ixgbe_hw *hw = &adapter->hw;
		/*
		 * Restart clause 37 autoneg, disable and re-enable
		 * the tx laser, to clear & alert the link partner
		 * that it needs to restart autotry
		 */
		hw->mac.autotry_restart = true;
		hw->mac.ops.flap_tx_laser(hw);
	}
6391 6392
	cancel_work_sync(&adapter->multispeed_fiber_task);
	cancel_work_sync(&adapter->sfp_config_module_task);
6393 6394 6395
	if (adapter->flags & IXGBE_FLAG_FDIR_HASH_CAPABLE ||
	    adapter->flags & IXGBE_FLAG_FDIR_PERFECT_CAPABLE)
		cancel_work_sync(&adapter->fdir_reinit_task);
6396 6397
	flush_scheduled_work();

6398
#ifdef CONFIG_IXGBE_DCA
6399 6400 6401 6402 6403 6404 6405
	if (adapter->flags & IXGBE_FLAG_DCA_ENABLED) {
		adapter->flags &= ~IXGBE_FLAG_DCA_ENABLED;
		dca_remove_requester(&pdev->dev);
		IXGBE_WRITE_REG(&adapter->hw, IXGBE_DCA_CTRL, 1);
	}

#endif
6406 6407 6408 6409 6410
#ifdef IXGBE_FCOE
	if (adapter->flags & IXGBE_FLAG_FCOE_ENABLED)
		ixgbe_cleanup_fcoe(adapter);

#endif /* IXGBE_FCOE */
6411 6412 6413 6414

	/* remove the added san mac */
	ixgbe_del_sanmac_netdev(netdev);

D
Donald Skidmore 已提交
6415 6416
	if (netdev->reg_state == NETREG_REGISTERED)
		unregister_netdev(netdev);
6417

6418 6419 6420
	if (adapter->flags & IXGBE_FLAG_SRIOV_ENABLED)
		ixgbe_disable_sriov(adapter);

6421
	ixgbe_clear_interrupt_scheme(adapter);
6422

6423
	ixgbe_release_hw_control(adapter);
6424 6425

	iounmap(adapter->hw.hw_addr);
6426 6427
	pci_release_selected_regions(pdev, pci_select_bars(pdev,
	                             IORESOURCE_MEM));
6428

6429 6430
	DPRINTK(PROBE, INFO, "complete\n");

6431 6432
	free_netdev(netdev);

6433
	pci_disable_pcie_error_reporting(pdev);
6434

6435 6436 6437 6438 6439 6440 6441 6442 6443 6444 6445 6446
	pci_disable_device(pdev);
}

/**
 * ixgbe_io_error_detected - called when PCI error is detected
 * @pdev: Pointer to PCI device
 * @state: The current pci connection state
 *
 * This function is called after a PCI bus error affecting
 * this device has been detected.
 */
static pci_ers_result_t ixgbe_io_error_detected(struct pci_dev *pdev,
6447
                                                pci_channel_state_t state)
6448 6449
{
	struct net_device *netdev = pci_get_drvdata(pdev);
6450
	struct ixgbe_adapter *adapter = netdev_priv(netdev);
6451 6452 6453

	netif_device_detach(netdev);

6454 6455 6456
	if (state == pci_channel_io_perm_failure)
		return PCI_ERS_RESULT_DISCONNECT;

6457 6458 6459 6460
	if (netif_running(netdev))
		ixgbe_down(adapter);
	pci_disable_device(pdev);

6461
	/* Request a slot reset. */
6462 6463 6464 6465 6466 6467 6468 6469 6470 6471 6472 6473
	return PCI_ERS_RESULT_NEED_RESET;
}

/**
 * ixgbe_io_slot_reset - called after the pci bus has been reset.
 * @pdev: Pointer to PCI device
 *
 * Restart the card from scratch, as if from a cold-boot.
 */
static pci_ers_result_t ixgbe_io_slot_reset(struct pci_dev *pdev)
{
	struct net_device *netdev = pci_get_drvdata(pdev);
6474
	struct ixgbe_adapter *adapter = netdev_priv(netdev);
6475 6476
	pci_ers_result_t result;
	int err;
6477

6478
	if (pci_enable_device_mem(pdev)) {
6479
		DPRINTK(PROBE, ERR,
6480
		        "Cannot re-enable PCI device after reset.\n");
6481 6482 6483 6484
		result = PCI_ERS_RESULT_DISCONNECT;
	} else {
		pci_set_master(pdev);
		pci_restore_state(pdev);
6485
		pci_save_state(pdev);
6486

6487
		pci_wake_from_d3(pdev, false);
6488

6489
		ixgbe_reset(adapter);
6490
		IXGBE_WRITE_REG(&adapter->hw, IXGBE_WUS, ~0);
6491 6492 6493 6494 6495 6496 6497 6498 6499
		result = PCI_ERS_RESULT_RECOVERED;
	}

	err = pci_cleanup_aer_uncorrect_error_status(pdev);
	if (err) {
		dev_err(&pdev->dev,
		  "pci_cleanup_aer_uncorrect_error_status failed 0x%0x\n", err);
		/* non-fatal, continue */
	}
6500

6501
	return result;
6502 6503 6504 6505 6506 6507 6508 6509 6510 6511 6512 6513
}

/**
 * ixgbe_io_resume - called when traffic can start flowing again.
 * @pdev: Pointer to PCI device
 *
 * This callback is called when the error recovery driver tells us that
 * its OK to resume normal operation.
 */
static void ixgbe_io_resume(struct pci_dev *pdev)
{
	struct net_device *netdev = pci_get_drvdata(pdev);
6514
	struct ixgbe_adapter *adapter = netdev_priv(netdev);
6515 6516 6517 6518 6519 6520 6521 6522 6523 6524 6525 6526 6527 6528 6529 6530 6531 6532 6533 6534 6535 6536 6537 6538 6539 6540 6541 6542 6543 6544 6545 6546 6547 6548 6549 6550 6551 6552 6553 6554 6555 6556 6557 6558

	if (netif_running(netdev)) {
		if (ixgbe_up(adapter)) {
			DPRINTK(PROBE, INFO, "ixgbe_up failed after reset\n");
			return;
		}
	}

	netif_device_attach(netdev);
}

static struct pci_error_handlers ixgbe_err_handler = {
	.error_detected = ixgbe_io_error_detected,
	.slot_reset = ixgbe_io_slot_reset,
	.resume = ixgbe_io_resume,
};

static struct pci_driver ixgbe_driver = {
	.name     = ixgbe_driver_name,
	.id_table = ixgbe_pci_tbl,
	.probe    = ixgbe_probe,
	.remove   = __devexit_p(ixgbe_remove),
#ifdef CONFIG_PM
	.suspend  = ixgbe_suspend,
	.resume   = ixgbe_resume,
#endif
	.shutdown = ixgbe_shutdown,
	.err_handler = &ixgbe_err_handler
};

/**
 * ixgbe_init_module - Driver Registration Routine
 *
 * ixgbe_init_module is the first routine called when the driver is
 * loaded. All it does is register with the PCI subsystem.
 **/
static int __init ixgbe_init_module(void)
{
	int ret;
	printk(KERN_INFO "%s: %s - version %s\n", ixgbe_driver_name,
	       ixgbe_driver_string, ixgbe_driver_version);

	printk(KERN_INFO "%s: %s\n", ixgbe_driver_name, ixgbe_copyright);

6559
#ifdef CONFIG_IXGBE_DCA
6560 6561
	dca_register_notify(&dca_notifier);
#endif
6562

6563 6564 6565
	ret = pci_register_driver(&ixgbe_driver);
	return ret;
}
6566

6567 6568 6569 6570 6571 6572 6573 6574 6575 6576
module_init(ixgbe_init_module);

/**
 * ixgbe_exit_module - Driver Exit Cleanup Routine
 *
 * ixgbe_exit_module is called just before the driver is removed
 * from memory.
 **/
static void __exit ixgbe_exit_module(void)
{
6577
#ifdef CONFIG_IXGBE_DCA
6578 6579
	dca_unregister_notify(&dca_notifier);
#endif
6580 6581
	pci_unregister_driver(&ixgbe_driver);
}
6582

6583
#ifdef CONFIG_IXGBE_DCA
6584
static int ixgbe_notify_dca(struct notifier_block *nb, unsigned long event,
6585
                            void *p)
6586 6587 6588 6589
{
	int ret_val;

	ret_val = driver_for_each_device(&ixgbe_driver.driver, NULL, &event,
6590
	                                 __ixgbe_notify_dca);
6591 6592 6593

	return ret_val ? NOTIFY_BAD : NOTIFY_DONE;
}
6594

6595
#endif /* CONFIG_IXGBE_DCA */
6596 6597 6598 6599 6600 6601 6602 6603 6604 6605
#ifdef DEBUG
/**
 * ixgbe_get_hw_dev_name - return device name string
 * used by hardware layer to print debugging information
 **/
char *ixgbe_get_hw_dev_name(struct ixgbe_hw *hw)
{
	struct ixgbe_adapter *adapter = hw->back;
	return adapter->netdev->name;
}
6606

6607
#endif
6608 6609 6610
module_exit(ixgbe_exit_module);

/* ixgbe_main.c */