e1000_main.c 154.2 KB
Newer Older
L
Linus Torvalds 已提交
1 2
/*******************************************************************************

3 4 5 6 7 8 9 10 11 12
  Intel PRO/1000 Linux driver
  Copyright(c) 1999 - 2006 Intel Corporation.

  This program is free software; you can redistribute it and/or modify it
  under the terms and conditions of the GNU General Public License,
  version 2, as published by the Free Software Foundation.

  This program is distributed in the hope it will be useful, but WITHOUT
  ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
  FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
L
Linus Torvalds 已提交
13
  more details.
14

L
Linus Torvalds 已提交
15
  You should have received a copy of the GNU General Public License along with
16 17 18 19 20 21
  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".

L
Linus Torvalds 已提交
22 23
  Contact Information:
  Linux NICS <linux.nics@intel.com>
24
  e1000-devel Mailing List <e1000-devel@lists.sourceforge.net>
L
Linus Torvalds 已提交
25 26 27 28 29
  Intel Corporation, 5200 N.E. Elam Young Parkway, Hillsboro, OR 97124-6497

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

#include "e1000.h"
A
Andrew Morton 已提交
30
#include <net/ip6_checksum.h>
L
Linus Torvalds 已提交
31 32

char e1000_driver_name[] = "e1000";
33
static char e1000_driver_string[] = "Intel(R) PRO/1000 Network Driver";
L
Linus Torvalds 已提交
34 35 36 37 38
#ifndef CONFIG_E1000_NAPI
#define DRIVERNAPI
#else
#define DRIVERNAPI "-NAPI"
#endif
A
Auke Kok 已提交
39
#define DRV_VERSION "7.3.20-k2"DRIVERNAPI
S
Stephen Hemminger 已提交
40 41
const char e1000_driver_version[] = DRV_VERSION;
static const char e1000_copyright[] = "Copyright (c) 1999-2006 Intel Corporation.";
L
Linus Torvalds 已提交
42 43 44 45 46 47 48 49

/* e1000_pci_tbl - PCI Device ID Table
 *
 * Last entry must be all 0s
 *
 * Macro expands to...
 *   {PCI_DEVICE(PCI_VENDOR_ID_INTEL, device_id)}
 */
50 51 52 53 54 55
#ifdef CONFIG_E1000E_ENABLED
  #define PCIE(x) 
#else
  #define PCIE(x) x,
#endif

L
Linus Torvalds 已提交
56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75
static struct pci_device_id e1000_pci_tbl[] = {
	INTEL_E1000_ETHERNET_DEVICE(0x1000),
	INTEL_E1000_ETHERNET_DEVICE(0x1001),
	INTEL_E1000_ETHERNET_DEVICE(0x1004),
	INTEL_E1000_ETHERNET_DEVICE(0x1008),
	INTEL_E1000_ETHERNET_DEVICE(0x1009),
	INTEL_E1000_ETHERNET_DEVICE(0x100C),
	INTEL_E1000_ETHERNET_DEVICE(0x100D),
	INTEL_E1000_ETHERNET_DEVICE(0x100E),
	INTEL_E1000_ETHERNET_DEVICE(0x100F),
	INTEL_E1000_ETHERNET_DEVICE(0x1010),
	INTEL_E1000_ETHERNET_DEVICE(0x1011),
	INTEL_E1000_ETHERNET_DEVICE(0x1012),
	INTEL_E1000_ETHERNET_DEVICE(0x1013),
	INTEL_E1000_ETHERNET_DEVICE(0x1014),
	INTEL_E1000_ETHERNET_DEVICE(0x1015),
	INTEL_E1000_ETHERNET_DEVICE(0x1016),
	INTEL_E1000_ETHERNET_DEVICE(0x1017),
	INTEL_E1000_ETHERNET_DEVICE(0x1018),
	INTEL_E1000_ETHERNET_DEVICE(0x1019),
76
	INTEL_E1000_ETHERNET_DEVICE(0x101A),
L
Linus Torvalds 已提交
77 78 79 80 81
	INTEL_E1000_ETHERNET_DEVICE(0x101D),
	INTEL_E1000_ETHERNET_DEVICE(0x101E),
	INTEL_E1000_ETHERNET_DEVICE(0x1026),
	INTEL_E1000_ETHERNET_DEVICE(0x1027),
	INTEL_E1000_ETHERNET_DEVICE(0x1028),
82 83 84 85 86 87 88 89
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x1049))
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x104A))
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x104B))
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x104C))
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x104D))
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x105E))
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x105F))
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x1060))
L
Linus Torvalds 已提交
90 91 92 93 94 95 96 97
	INTEL_E1000_ETHERNET_DEVICE(0x1075),
	INTEL_E1000_ETHERNET_DEVICE(0x1076),
	INTEL_E1000_ETHERNET_DEVICE(0x1077),
	INTEL_E1000_ETHERNET_DEVICE(0x1078),
	INTEL_E1000_ETHERNET_DEVICE(0x1079),
	INTEL_E1000_ETHERNET_DEVICE(0x107A),
	INTEL_E1000_ETHERNET_DEVICE(0x107B),
	INTEL_E1000_ETHERNET_DEVICE(0x107C),
98 99 100
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x107D))
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x107E))
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x107F))
L
Linus Torvalds 已提交
101
	INTEL_E1000_ETHERNET_DEVICE(0x108A),
102 103 104 105
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x108B))
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x108C))
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x1096))
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x1098))
106
	INTEL_E1000_ETHERNET_DEVICE(0x1099),
107 108 109
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x109A))
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x10A4))
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x10A5))
110
	INTEL_E1000_ETHERNET_DEVICE(0x10B5),
111 112 113 114 115 116 117 118 119
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x10B9))
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x10BA))
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x10BB))
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x10BC))
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x10C4))
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x10C5))
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x10D5))
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x10D9))
PCIE(	INTEL_E1000_ETHERNET_DEVICE(0x10DA))
L
Linus Torvalds 已提交
120 121 122 123 124 125
	/* required last entry */
	{0,}
};

MODULE_DEVICE_TABLE(pci, e1000_pci_tbl);

126 127 128 129 130 131 132 133 134
int e1000_up(struct e1000_adapter *adapter);
void e1000_down(struct e1000_adapter *adapter);
void e1000_reinit_locked(struct e1000_adapter *adapter);
void e1000_reset(struct e1000_adapter *adapter);
int e1000_set_spd_dplx(struct e1000_adapter *adapter, uint16_t spddplx);
int e1000_setup_all_tx_resources(struct e1000_adapter *adapter);
int e1000_setup_all_rx_resources(struct e1000_adapter *adapter);
void e1000_free_all_tx_resources(struct e1000_adapter *adapter);
void e1000_free_all_rx_resources(struct e1000_adapter *adapter);
135
static int e1000_setup_tx_resources(struct e1000_adapter *adapter,
136
                             struct e1000_tx_ring *txdr);
137
static int e1000_setup_rx_resources(struct e1000_adapter *adapter,
138
                             struct e1000_rx_ring *rxdr);
139
static void e1000_free_tx_resources(struct e1000_adapter *adapter,
140
                             struct e1000_tx_ring *tx_ring);
141
static void e1000_free_rx_resources(struct e1000_adapter *adapter,
142 143
                             struct e1000_rx_ring *rx_ring);
void e1000_update_stats(struct e1000_adapter *adapter);
L
Linus Torvalds 已提交
144 145 146 147 148

static int e1000_init_module(void);
static void e1000_exit_module(void);
static int e1000_probe(struct pci_dev *pdev, const struct pci_device_id *ent);
static void __devexit e1000_remove(struct pci_dev *pdev);
149
static int e1000_alloc_queues(struct e1000_adapter *adapter);
L
Linus Torvalds 已提交
150 151 152 153 154 155
static int e1000_sw_init(struct e1000_adapter *adapter);
static int e1000_open(struct net_device *netdev);
static int e1000_close(struct net_device *netdev);
static void e1000_configure_tx(struct e1000_adapter *adapter);
static void e1000_configure_rx(struct e1000_adapter *adapter);
static void e1000_setup_rctl(struct e1000_adapter *adapter);
156 157 158 159 160 161
static void e1000_clean_all_tx_rings(struct e1000_adapter *adapter);
static void e1000_clean_all_rx_rings(struct e1000_adapter *adapter);
static void e1000_clean_tx_ring(struct e1000_adapter *adapter,
                                struct e1000_tx_ring *tx_ring);
static void e1000_clean_rx_ring(struct e1000_adapter *adapter,
                                struct e1000_rx_ring *rx_ring);
162
static void e1000_set_rx_mode(struct net_device *netdev);
L
Linus Torvalds 已提交
163 164 165 166 167 168 169
static void e1000_update_phy_info(unsigned long data);
static void e1000_watchdog(unsigned long data);
static void e1000_82547_tx_fifo_stall(unsigned long data);
static int e1000_xmit_frame(struct sk_buff *skb, struct net_device *netdev);
static struct net_device_stats * e1000_get_stats(struct net_device *netdev);
static int e1000_change_mtu(struct net_device *netdev, int new_mtu);
static int e1000_set_mac(struct net_device *netdev, void *p);
170
static irqreturn_t e1000_intr(int irq, void *data);
171
static irqreturn_t e1000_intr_msi(int irq, void *data);
172 173
static boolean_t e1000_clean_tx_irq(struct e1000_adapter *adapter,
                                    struct e1000_tx_ring *tx_ring);
L
Linus Torvalds 已提交
174
#ifdef CONFIG_E1000_NAPI
175
static int e1000_clean(struct napi_struct *napi, int budget);
L
Linus Torvalds 已提交
176
static boolean_t e1000_clean_rx_irq(struct e1000_adapter *adapter,
177
                                    struct e1000_rx_ring *rx_ring,
L
Linus Torvalds 已提交
178
                                    int *work_done, int work_to_do);
179
static boolean_t e1000_clean_rx_irq_ps(struct e1000_adapter *adapter,
180
                                       struct e1000_rx_ring *rx_ring,
181
                                       int *work_done, int work_to_do);
L
Linus Torvalds 已提交
182
#else
183 184 185 186
static boolean_t e1000_clean_rx_irq(struct e1000_adapter *adapter,
                                    struct e1000_rx_ring *rx_ring);
static boolean_t e1000_clean_rx_irq_ps(struct e1000_adapter *adapter,
                                       struct e1000_rx_ring *rx_ring);
L
Linus Torvalds 已提交
187
#endif
188
static void e1000_alloc_rx_buffers(struct e1000_adapter *adapter,
189 190
                                   struct e1000_rx_ring *rx_ring,
				   int cleaned_count);
191
static void e1000_alloc_rx_buffers_ps(struct e1000_adapter *adapter,
192 193
                                      struct e1000_rx_ring *rx_ring,
				      int cleaned_count);
L
Linus Torvalds 已提交
194 195 196 197 198 199
static int e1000_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd);
static int e1000_mii_ioctl(struct net_device *netdev, struct ifreq *ifr,
			   int cmd);
static void e1000_enter_82542_rst(struct e1000_adapter *adapter);
static void e1000_leave_82542_rst(struct e1000_adapter *adapter);
static void e1000_tx_timeout(struct net_device *dev);
200
static void e1000_reset_task(struct work_struct *work);
L
Linus Torvalds 已提交
201
static void e1000_smartspeed(struct e1000_adapter *adapter);
202 203
static int e1000_82547_fifo_workaround(struct e1000_adapter *adapter,
                                       struct sk_buff *skb);
L
Linus Torvalds 已提交
204 205 206 207 208 209

static void e1000_vlan_rx_register(struct net_device *netdev, struct vlan_group *grp);
static void e1000_vlan_rx_add_vid(struct net_device *netdev, uint16_t vid);
static void e1000_vlan_rx_kill_vid(struct net_device *netdev, uint16_t vid);
static void e1000_restore_vlan(struct e1000_adapter *adapter);

210
static int e1000_suspend(struct pci_dev *pdev, pm_message_t state);
A
Auke Kok 已提交
211
#ifdef CONFIG_PM
L
Linus Torvalds 已提交
212 213
static int e1000_resume(struct pci_dev *pdev);
#endif
214
static void e1000_shutdown(struct pci_dev *pdev);
L
Linus Torvalds 已提交
215 216 217 218 219 220

#ifdef CONFIG_NET_POLL_CONTROLLER
/* for netdump / net console */
static void e1000_netpoll (struct net_device *netdev);
#endif

221 222 223 224 225 226
#define COPYBREAK_DEFAULT 256
static unsigned int copybreak __read_mostly = COPYBREAK_DEFAULT;
module_param(copybreak, uint, 0644);
MODULE_PARM_DESC(copybreak,
	"Maximum size of packet that is copied to a new buffer on receive");

A
Auke Kok 已提交
227 228 229 230 231 232 233 234 235 236
static pci_ers_result_t e1000_io_error_detected(struct pci_dev *pdev,
                     pci_channel_state_t state);
static pci_ers_result_t e1000_io_slot_reset(struct pci_dev *pdev);
static void e1000_io_resume(struct pci_dev *pdev);

static struct pci_error_handlers e1000_err_handler = {
	.error_detected = e1000_io_error_detected,
	.slot_reset = e1000_io_slot_reset,
	.resume = e1000_io_resume,
};
237

L
Linus Torvalds 已提交
238 239 240 241 242
static struct pci_driver e1000_driver = {
	.name     = e1000_driver_name,
	.id_table = e1000_pci_tbl,
	.probe    = e1000_probe,
	.remove   = __devexit_p(e1000_remove),
243
#ifdef CONFIG_PM
L
Linus Torvalds 已提交
244 245
	/* Power Managment Hooks */
	.suspend  = e1000_suspend,
246
	.resume   = e1000_resume,
L
Linus Torvalds 已提交
247
#endif
A
Auke Kok 已提交
248 249
	.shutdown = e1000_shutdown,
	.err_handler = &e1000_err_handler
L
Linus Torvalds 已提交
250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276
};

MODULE_AUTHOR("Intel Corporation, <linux.nics@intel.com>");
MODULE_DESCRIPTION("Intel(R) PRO/1000 Network Driver");
MODULE_LICENSE("GPL");
MODULE_VERSION(DRV_VERSION);

static int debug = NETIF_MSG_DRV | NETIF_MSG_PROBE;
module_param(debug, int, 0);
MODULE_PARM_DESC(debug, "Debug level (0=none,...,16=all)");

/**
 * e1000_init_module - Driver Registration Routine
 *
 * e1000_init_module is the first routine called when the driver is
 * loaded. All it does is register with the PCI subsystem.
 **/

static int __init
e1000_init_module(void)
{
	int ret;
	printk(KERN_INFO "%s - version %s\n",
	       e1000_driver_string, e1000_driver_version);

	printk(KERN_INFO "%s\n", e1000_copyright);

277
	ret = pci_register_driver(&e1000_driver);
278 279 280 281 282 283 284
	if (copybreak != COPYBREAK_DEFAULT) {
		if (copybreak == 0)
			printk(KERN_INFO "e1000: copybreak disabled\n");
		else
			printk(KERN_INFO "e1000: copybreak enabled for "
			       "packets <= %u bytes\n", copybreak);
	}
L
Linus Torvalds 已提交
285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304
	return ret;
}

module_init(e1000_init_module);

/**
 * e1000_exit_module - Driver Exit Cleanup Routine
 *
 * e1000_exit_module is called just before the driver is removed
 * from memory.
 **/

static void __exit
e1000_exit_module(void)
{
	pci_unregister_driver(&e1000_driver);
}

module_exit(e1000_exit_module);

305 306 307
static int e1000_request_irq(struct e1000_adapter *adapter)
{
	struct net_device *netdev = adapter->netdev;
A
Al Viro 已提交
308
	irq_handler_t handler = e1000_intr;
309 310
	int irq_flags = IRQF_SHARED;
	int err;
311

312
	if (adapter->hw.mac_type >= e1000_82571) {
313 314
		adapter->have_msi = !pci_enable_msi(adapter->pdev);
		if (adapter->have_msi) {
A
Al Viro 已提交
315
			handler = e1000_intr_msi;
316
			irq_flags = 0;
317 318
		}
	}
319 320 321 322 323 324

	err = request_irq(adapter->pdev->irq, handler, irq_flags, netdev->name,
	                  netdev);
	if (err) {
		if (adapter->have_msi)
			pci_disable_msi(adapter->pdev);
325 326
		DPRINTK(PROBE, ERR,
		        "Unable to allocate interrupt Error: %d\n", err);
327
	}
328 329 330 331 332 333 334 335 336 337 338 339 340 341

	return err;
}

static void e1000_free_irq(struct e1000_adapter *adapter)
{
	struct net_device *netdev = adapter->netdev;

	free_irq(adapter->pdev->irq, netdev);

	if (adapter->have_msi)
		pci_disable_msi(adapter->pdev);
}

L
Linus Torvalds 已提交
342 343 344 345 346
/**
 * e1000_irq_disable - Mask off interrupt generation on the NIC
 * @adapter: board private structure
 **/

347
static void
L
Linus Torvalds 已提交
348 349 350 351 352 353 354 355 356 357 358 359 360
e1000_irq_disable(struct e1000_adapter *adapter)
{
	atomic_inc(&adapter->irq_sem);
	E1000_WRITE_REG(&adapter->hw, IMC, ~0);
	E1000_WRITE_FLUSH(&adapter->hw);
	synchronize_irq(adapter->pdev->irq);
}

/**
 * e1000_irq_enable - Enable default interrupt generation settings
 * @adapter: board private structure
 **/

361
static void
L
Linus Torvalds 已提交
362 363
e1000_irq_enable(struct e1000_adapter *adapter)
{
J
Jesse Brandeburg 已提交
364
	if (likely(atomic_dec_and_test(&adapter->irq_sem))) {
L
Linus Torvalds 已提交
365 366 367 368
		E1000_WRITE_REG(&adapter->hw, IMS, IMS_ENABLE_MASK);
		E1000_WRITE_FLUSH(&adapter->hw);
	}
}
369 370

static void
371 372 373 374 375
e1000_update_mng_vlan(struct e1000_adapter *adapter)
{
	struct net_device *netdev = adapter->netdev;
	uint16_t vid = adapter->hw.mng_cookie.vlan_id;
	uint16_t old_vid = adapter->mng_vlan_id;
J
Jesse Brandeburg 已提交
376
	if (adapter->vlgrp) {
D
Dan Aloni 已提交
377
		if (!vlan_group_get_device(adapter->vlgrp, vid)) {
J
Jesse Brandeburg 已提交
378
			if (adapter->hw.mng_cookie.status &
379 380 381 382 383
				E1000_MNG_DHCP_COOKIE_STATUS_VLAN_SUPPORT) {
				e1000_vlan_rx_add_vid(netdev, vid);
				adapter->mng_vlan_id = vid;
			} else
				adapter->mng_vlan_id = E1000_MNG_VLAN_NONE;
J
Jesse Brandeburg 已提交
384 385 386

			if ((old_vid != (uint16_t)E1000_MNG_VLAN_NONE) &&
					(vid != old_vid) &&
D
Dan Aloni 已提交
387
			    !vlan_group_get_device(adapter->vlgrp, old_vid))
388
				e1000_vlan_rx_kill_vid(netdev, old_vid);
389 390
		} else
			adapter->mng_vlan_id = vid;
391 392
	}
}
393 394 395 396 397 398 399 400

/**
 * e1000_release_hw_control - release control of the h/w to f/w
 * @adapter: address of board private structure
 *
 * e1000_release_hw_control resets {CTRL_EXT|FWSM}:DRV_LOAD bit.
 * For ASF and Pass Through versions of f/w this means that the
 * driver is no longer loaded. For AMT version (only with 82573) i
401
 * of the f/w this means that the network i/f is closed.
402
 *
403 404
 **/

405
static void
406 407 408 409 410 411 412 413 414 415 416
e1000_release_hw_control(struct e1000_adapter *adapter)
{
	uint32_t ctrl_ext;
	uint32_t swsm;

	/* Let firmware taken over control of h/w */
	switch (adapter->hw.mac_type) {
	case e1000_82573:
		swsm = E1000_READ_REG(&adapter->hw, SWSM);
		E1000_WRITE_REG(&adapter->hw, SWSM,
				swsm & ~E1000_SWSM_DRV_LOAD);
B
Bruce Allan 已提交
417 418 419 420
		break;
	case e1000_82571:
	case e1000_82572:
	case e1000_80003es2lan:
421
	case e1000_ich8lan:
B
Bruce Allan 已提交
422
		ctrl_ext = E1000_READ_REG(&adapter->hw, CTRL_EXT);
423
		E1000_WRITE_REG(&adapter->hw, CTRL_EXT,
B
Bruce Allan 已提交
424
				ctrl_ext & ~E1000_CTRL_EXT_DRV_LOAD);
425
		break;
426 427 428 429 430 431 432 433 434 435
	default:
		break;
	}
}

/**
 * e1000_get_hw_control - get control of the h/w from f/w
 * @adapter: address of board private structure
 *
 * e1000_get_hw_control sets {CTRL_EXT|FWSM}:DRV_LOAD bit.
436 437
 * For ASF and Pass Through versions of f/w this means that
 * the driver is loaded. For AMT version (only with 82573)
438
 * of the f/w this means that the network i/f is open.
439
 *
440 441
 **/

442
static void
443 444 445 446
e1000_get_hw_control(struct e1000_adapter *adapter)
{
	uint32_t ctrl_ext;
	uint32_t swsm;
447

448 449 450 451 452 453 454
	/* Let firmware know the driver has taken over */
	switch (adapter->hw.mac_type) {
	case e1000_82573:
		swsm = E1000_READ_REG(&adapter->hw, SWSM);
		E1000_WRITE_REG(&adapter->hw, SWSM,
				swsm | E1000_SWSM_DRV_LOAD);
		break;
B
Bruce Allan 已提交
455 456 457
	case e1000_82571:
	case e1000_82572:
	case e1000_80003es2lan:
458
	case e1000_ich8lan:
B
Bruce Allan 已提交
459 460 461
		ctrl_ext = E1000_READ_REG(&adapter->hw, CTRL_EXT);
		E1000_WRITE_REG(&adapter->hw, CTRL_EXT,
				ctrl_ext | E1000_CTRL_EXT_DRV_LOAD);
462
		break;
463 464 465 466 467
	default:
		break;
	}
}

468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513
static void
e1000_init_manageability(struct e1000_adapter *adapter)
{
	if (adapter->en_mng_pt) {
		uint32_t manc = E1000_READ_REG(&adapter->hw, MANC);

		/* disable hardware interception of ARP */
		manc &= ~(E1000_MANC_ARP_EN);

		/* enable receiving management packets to the host */
		/* this will probably generate destination unreachable messages
		 * from the host OS, but the packets will be handled on SMBUS */
		if (adapter->hw.has_manc2h) {
			uint32_t manc2h = E1000_READ_REG(&adapter->hw, MANC2H);

			manc |= E1000_MANC_EN_MNG2HOST;
#define E1000_MNG2HOST_PORT_623 (1 << 5)
#define E1000_MNG2HOST_PORT_664 (1 << 6)
			manc2h |= E1000_MNG2HOST_PORT_623;
			manc2h |= E1000_MNG2HOST_PORT_664;
			E1000_WRITE_REG(&adapter->hw, MANC2H, manc2h);
		}

		E1000_WRITE_REG(&adapter->hw, MANC, manc);
	}
}

static void
e1000_release_manageability(struct e1000_adapter *adapter)
{
	if (adapter->en_mng_pt) {
		uint32_t manc = E1000_READ_REG(&adapter->hw, MANC);

		/* re-enable hardware interception of ARP */
		manc |= E1000_MANC_ARP_EN;

		if (adapter->hw.has_manc2h)
			manc &= ~E1000_MANC_EN_MNG2HOST;

		/* don't explicitly have to mess with MANC2H since
		 * MANC has an enable disable that gates MANC2H */

		E1000_WRITE_REG(&adapter->hw, MANC, manc);
	}
}

514 515 516 517 518
/**
 * e1000_configure - configure the hardware for RX and TX
 * @adapter = private board structure
 **/
static void e1000_configure(struct e1000_adapter *adapter)
L
Linus Torvalds 已提交
519 520
{
	struct net_device *netdev = adapter->netdev;
521
	int i;
L
Linus Torvalds 已提交
522

523
	e1000_set_rx_mode(netdev);
L
Linus Torvalds 已提交
524 525

	e1000_restore_vlan(adapter);
526
	e1000_init_manageability(adapter);
L
Linus Torvalds 已提交
527 528 529 530

	e1000_configure_tx(adapter);
	e1000_setup_rctl(adapter);
	e1000_configure_rx(adapter);
531 532 533
	/* call E1000_DESC_UNUSED which always leaves
	 * at least 1 descriptor unused to make sure
	 * next_to_use != next_to_clean */
534
	for (i = 0; i < adapter->num_rx_queues; i++) {
535
		struct e1000_rx_ring *ring = &adapter->rx_ring[i];
536 537
		adapter->alloc_rx_buf(adapter, ring,
		                      E1000_DESC_UNUSED(ring));
538
	}
L
Linus Torvalds 已提交
539

540
	adapter->tx_queue_len = netdev->tx_queue_len;
541 542 543 544 545 546 547 548
}

int e1000_up(struct e1000_adapter *adapter)
{
	/* hardware has been reset, we need to reload some things */
	e1000_configure(adapter);

	clear_bit(__E1000_DOWN, &adapter->flags);
549

L
Linus Torvalds 已提交
550
#ifdef CONFIG_E1000_NAPI
551
	napi_enable(&adapter->napi);
L
Linus Torvalds 已提交
552
#endif
553 554
	e1000_irq_enable(adapter);

555 556
	/* fire a link change interrupt to start the watchdog */
	E1000_WRITE_REG(&adapter->hw, ICS, E1000_ICS_LSC);
L
Linus Torvalds 已提交
557 558 559
	return 0;
}

560 561 562 563 564 565 566 567 568 569
/**
 * e1000_power_up_phy - restore link in case the phy was powered down
 * @adapter: address of board private structure
 *
 * The phy may be powered down to save power and turn off link when the
 * driver is unloaded and wake on lan is not enabled (among others)
 * *** this routine MUST be followed by a call to e1000_reset ***
 *
 **/

570
void e1000_power_up_phy(struct e1000_adapter *adapter)
571 572 573 574 575 576 577 578 579 580 581 582 583 584 585
{
	uint16_t mii_reg = 0;

	/* Just clear the power down bit to wake the phy back up */
	if (adapter->hw.media_type == e1000_media_type_copper) {
		/* according to the manual, the phy will retain its
		 * settings across a power-down/up cycle */
		e1000_read_phy_reg(&adapter->hw, PHY_CTRL, &mii_reg);
		mii_reg &= ~MII_CR_POWER_DOWN;
		e1000_write_phy_reg(&adapter->hw, PHY_CTRL, mii_reg);
	}
}

static void e1000_power_down_phy(struct e1000_adapter *adapter)
{
586 587
	/* Power down the PHY so no link is implied when interface is down *
	 * The PHY cannot be powered down if any of the following is TRUE *
588 589 590 591
	 * (a) WoL is enabled
	 * (b) AMT is active
	 * (c) SoL/IDER session is active */
	if (!adapter->wol && adapter->hw.mac_type >= e1000_82540 &&
592
	   adapter->hw.media_type == e1000_media_type_copper) {
593
		uint16_t mii_reg = 0;
594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620

		switch (adapter->hw.mac_type) {
		case e1000_82540:
		case e1000_82545:
		case e1000_82545_rev_3:
		case e1000_82546:
		case e1000_82546_rev_3:
		case e1000_82541:
		case e1000_82541_rev_2:
		case e1000_82547:
		case e1000_82547_rev_2:
			if (E1000_READ_REG(&adapter->hw, MANC) &
			    E1000_MANC_SMBUS_EN)
				goto out;
			break;
		case e1000_82571:
		case e1000_82572:
		case e1000_82573:
		case e1000_80003es2lan:
		case e1000_ich8lan:
			if (e1000_check_mng_mode(&adapter->hw) ||
			    e1000_check_phy_reset_block(&adapter->hw))
				goto out;
			break;
		default:
			goto out;
		}
621 622 623 624 625
		e1000_read_phy_reg(&adapter->hw, PHY_CTRL, &mii_reg);
		mii_reg |= MII_CR_POWER_DOWN;
		e1000_write_phy_reg(&adapter->hw, PHY_CTRL, mii_reg);
		mdelay(1);
	}
626 627
out:
	return;
628 629
}

L
Linus Torvalds 已提交
630 631 632 633 634
void
e1000_down(struct e1000_adapter *adapter)
{
	struct net_device *netdev = adapter->netdev;

A
Auke Kok 已提交
635 636 637 638
	/* signal that we're down so the interrupt handler does not
	 * reschedule our watchdog timer */
	set_bit(__E1000_DOWN, &adapter->flags);

639
#ifdef CONFIG_E1000_NAPI
640
	napi_disable(&adapter->napi);
641
	atomic_set(&adapter->irq_sem, 0);
642
#endif
L
Linus Torvalds 已提交
643
	e1000_irq_disable(adapter);
644

L
Linus Torvalds 已提交
645 646 647 648
	del_timer_sync(&adapter->tx_fifo_stall_timer);
	del_timer_sync(&adapter->watchdog_timer);
	del_timer_sync(&adapter->phy_info_timer);

649
	netdev->tx_queue_len = adapter->tx_queue_len;
L
Linus Torvalds 已提交
650 651 652 653 654 655
	adapter->link_speed = 0;
	adapter->link_duplex = 0;
	netif_carrier_off(netdev);
	netif_stop_queue(netdev);

	e1000_reset(adapter);
656 657
	e1000_clean_all_tx_rings(adapter);
	e1000_clean_all_rx_rings(adapter);
L
Linus Torvalds 已提交
658 659
}

660 661 662 663 664 665 666 667 668
void
e1000_reinit_locked(struct e1000_adapter *adapter)
{
	WARN_ON(in_interrupt());
	while (test_and_set_bit(__E1000_RESETTING, &adapter->flags))
		msleep(1);
	e1000_down(adapter);
	e1000_up(adapter);
	clear_bit(__E1000_RESETTING, &adapter->flags);
L
Linus Torvalds 已提交
669 670 671 672 673
}

void
e1000_reset(struct e1000_adapter *adapter)
{
674
	uint32_t pba = 0, tx_space, min_tx_space, min_rx_space;
675
	uint16_t fc_high_water_mark = E1000_FC_HIGH_DIFF;
676
	boolean_t legacy_pba_adjust = FALSE;
L
Linus Torvalds 已提交
677 678 679 680 681

	/* Repartition Pba for greater than 9k mtu
	 * To take effect CTRL.RST is required.
	 */

682
	switch (adapter->hw.mac_type) {
683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698
	case e1000_82542_rev2_0:
	case e1000_82542_rev2_1:
	case e1000_82543:
	case e1000_82544:
	case e1000_82540:
	case e1000_82541:
	case e1000_82541_rev_2:
		legacy_pba_adjust = TRUE;
		pba = E1000_PBA_48K;
		break;
	case e1000_82545:
	case e1000_82545_rev_3:
	case e1000_82546:
	case e1000_82546_rev_3:
		pba = E1000_PBA_48K;
		break;
699
	case e1000_82547:
700
	case e1000_82547_rev_2:
701
		legacy_pba_adjust = TRUE;
702 703
		pba = E1000_PBA_30K;
		break;
704 705
	case e1000_82571:
	case e1000_82572:
706
	case e1000_80003es2lan:
707 708
		pba = E1000_PBA_38K;
		break;
709
	case e1000_82573:
710
		pba = E1000_PBA_20K;
711
		break;
712 713
	case e1000_ich8lan:
		pba = E1000_PBA_8K;
714 715
	case e1000_undefined:
	case e1000_num_macs:
716 717 718
		break;
	}

719 720 721
	if (legacy_pba_adjust == TRUE) {
		if (adapter->netdev->mtu > E1000_RXBUFFER_8192)
			pba -= 8; /* allocate more FIFO for Tx */
722

723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749
		if (adapter->hw.mac_type == e1000_82547) {
			adapter->tx_fifo_head = 0;
			adapter->tx_head_addr = pba << E1000_TX_HEAD_ADDR_SHIFT;
			adapter->tx_fifo_size =
				(E1000_PBA_40K - pba) << E1000_PBA_BYTES_SHIFT;
			atomic_set(&adapter->tx_fifo_stall, 0);
		}
	} else if (adapter->hw.max_frame_size > MAXIMUM_ETHERNET_FRAME_SIZE) {
		/* adjust PBA for jumbo frames */
		E1000_WRITE_REG(&adapter->hw, PBA, pba);

		/* To maintain wire speed transmits, the Tx FIFO should be
		 * large enough to accomodate two full transmit packets,
		 * rounded up to the next 1KB and expressed in KB.  Likewise,
		 * the Rx FIFO should be large enough to accomodate at least
		 * one full receive packet and is similarly rounded up and
		 * expressed in KB. */
		pba = E1000_READ_REG(&adapter->hw, PBA);
		/* upper 16 bits has Tx packet buffer allocation size in KB */
		tx_space = pba >> 16;
		/* lower 16 bits has Rx packet buffer allocation size in KB */
		pba &= 0xffff;
		/* don't include ethernet FCS because hardware appends/strips */
		min_rx_space = adapter->netdev->mtu + ENET_HEADER_SIZE +
		               VLAN_TAG_SIZE;
		min_tx_space = min_rx_space;
		min_tx_space *= 2;
750
		min_tx_space = ALIGN(min_tx_space, 1024);
751
		min_tx_space >>= 10;
752
		min_rx_space = ALIGN(min_rx_space, 1024);
753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783
		min_rx_space >>= 10;

		/* If current Tx allocation is less than the min Tx FIFO size,
		 * and the min Tx FIFO size is less than the current Rx FIFO
		 * allocation, take space away from current Rx allocation */
		if (tx_space < min_tx_space &&
		    ((min_tx_space - tx_space) < pba)) {
			pba = pba - (min_tx_space - tx_space);

			/* PCI/PCIx hardware has PBA alignment constraints */
			switch (adapter->hw.mac_type) {
			case e1000_82545 ... e1000_82546_rev_3:
				pba &= ~(E1000_PBA_8K - 1);
				break;
			default:
				break;
			}

			/* if short on rx space, rx wins and must trump tx
			 * adjustment or use Early Receive if available */
			if (pba < min_rx_space) {
				switch (adapter->hw.mac_type) {
				case e1000_82573:
					/* ERT enabled in e1000_configure_rx */
					break;
				default:
					pba = min_rx_space;
					break;
				}
			}
		}
L
Linus Torvalds 已提交
784
	}
785

L
Linus Torvalds 已提交
786 787 788
	E1000_WRITE_REG(&adapter->hw, PBA, pba);

	/* flow control settings */
789 790 791
	/* Set the FC high water mark to 90% of the FIFO size.
	 * Required to clear last 3 LSB */
	fc_high_water_mark = ((pba * 9216)/10) & 0xFFF8;
792 793 794 795 796 797
	/* We can't use 90% on small FIFOs because the remainder
	 * would be less than 1 full frame.  In this case, we size
	 * it to allow at least a full frame above the high water
	 *  mark. */
	if (pba < E1000_PBA_16K)
		fc_high_water_mark = (pba * 1024) - 1600;
798 799 800

	adapter->hw.fc_high_water = fc_high_water_mark;
	adapter->hw.fc_low_water = fc_high_water_mark - 8;
801 802 803 804
	if (adapter->hw.mac_type == e1000_80003es2lan)
		adapter->hw.fc_pause_time = 0xFFFF;
	else
		adapter->hw.fc_pause_time = E1000_FC_PAUSE_TIME;
L
Linus Torvalds 已提交
805 806 807
	adapter->hw.fc_send_xon = 1;
	adapter->hw.fc = adapter->hw.original_fc;

808
	/* Allow time for pending master requests to run */
L
Linus Torvalds 已提交
809
	e1000_reset_hw(&adapter->hw);
J
Jesse Brandeburg 已提交
810
	if (adapter->hw.mac_type >= e1000_82544)
L
Linus Torvalds 已提交
811
		E1000_WRITE_REG(&adapter->hw, WUC, 0);
812

J
Jesse Brandeburg 已提交
813
	if (e1000_init_hw(&adapter->hw))
L
Linus Torvalds 已提交
814
		DPRINTK(PROBE, ERR, "Hardware Error\n");
815
	e1000_update_mng_vlan(adapter);
816 817 818 819 820 821 822 823 824 825 826 827 828 829

	/* if (adapter->hwflags & HWFLAGS_PHY_PWR_BIT) { */
	if (adapter->hw.mac_type >= e1000_82544 &&
	    adapter->hw.mac_type <= e1000_82547_rev_2 &&
	    adapter->hw.autoneg == 1 &&
	    adapter->hw.autoneg_advertised == ADVERTISE_1000_FULL) {
		uint32_t ctrl = E1000_READ_REG(&adapter->hw, CTRL);
		/* clear phy power management bit if we are in gig only mode,
		 * which if enabled will attempt negotiation to 100Mb, which
		 * can cause a loss of link at power off or driver unload */
		ctrl &= ~E1000_CTRL_SWDPIN3;
		E1000_WRITE_REG(&adapter->hw, CTRL, ctrl);
	}

L
Linus Torvalds 已提交
830 831 832 833 834
	/* Enable h/w to recognize an 802.1Q VLAN Ethernet packet */
	E1000_WRITE_REG(&adapter->hw, VET, ETHERNET_IEEE_VLAN_TYPE);

	e1000_reset_adaptive(&adapter->hw);
	e1000_phy_get_info(&adapter->hw, &adapter->phy_info);
A
Auke Kok 已提交
835 836 837 838 839 840 841 842 843 844 845 846 847 848 849

	if (!adapter->smart_power_down &&
	    (adapter->hw.mac_type == e1000_82571 ||
	     adapter->hw.mac_type == e1000_82572)) {
		uint16_t phy_data = 0;
		/* speed up time to link by disabling smart power down, ignore
		 * the return value of this function because there is nothing
		 * different we would do if it failed */
		e1000_read_phy_reg(&adapter->hw, IGP02E1000_PHY_POWER_MGMT,
		                   &phy_data);
		phy_data &= ~IGP02E1000_PM_SPD;
		e1000_write_phy_reg(&adapter->hw, IGP02E1000_PHY_POWER_MGMT,
		                    phy_data);
	}

850
	e1000_release_manageability(adapter);
L
Linus Torvalds 已提交
851 852
}

853 854 855
/**
 *  Dump the eeprom for users having checksum issues
 **/
856
static void e1000_dump_eeprom(struct e1000_adapter *adapter)
857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910
{
	struct net_device *netdev = adapter->netdev;
	struct ethtool_eeprom eeprom;
	const struct ethtool_ops *ops = netdev->ethtool_ops;
	u8 *data;
	int i;
	u16 csum_old, csum_new = 0;

	eeprom.len = ops->get_eeprom_len(netdev);
	eeprom.offset = 0;

	data = kmalloc(eeprom.len, GFP_KERNEL);
	if (!data) {
		printk(KERN_ERR "Unable to allocate memory to dump EEPROM"
		       " data\n");
		return;
	}

	ops->get_eeprom(netdev, &eeprom, data);

	csum_old = (data[EEPROM_CHECKSUM_REG * 2]) +
		   (data[EEPROM_CHECKSUM_REG * 2 + 1] << 8);
	for (i = 0; i < EEPROM_CHECKSUM_REG * 2; i += 2)
		csum_new += data[i] + (data[i + 1] << 8);
	csum_new = EEPROM_SUM - csum_new;

	printk(KERN_ERR "/*********************/\n");
	printk(KERN_ERR "Current EEPROM Checksum : 0x%04x\n", csum_old);
	printk(KERN_ERR "Calculated              : 0x%04x\n", csum_new);

	printk(KERN_ERR "Offset    Values\n");
	printk(KERN_ERR "========  ======\n");
	print_hex_dump(KERN_ERR, "", DUMP_PREFIX_OFFSET, 16, 1, data, 128, 0);

	printk(KERN_ERR "Include this output when contacting your support "
	       "provider.\n");
	printk(KERN_ERR "This is not a software error! Something bad "
	       "happened to your hardware or\n");
	printk(KERN_ERR "EEPROM image. Ignoring this "
	       "problem could result in further problems,\n");
	printk(KERN_ERR "possibly loss of data, corruption or system hangs!\n");
	printk(KERN_ERR "The MAC Address will be reset to 00:00:00:00:00:00, "
	       "which is invalid\n");
	printk(KERN_ERR "and requires you to set the proper MAC "
	       "address manually before continuing\n");
	printk(KERN_ERR "to enable this network device.\n");
	printk(KERN_ERR "Please inspect the EEPROM dump and report the issue "
	       "to your hardware vendor\n");
	printk(KERN_ERR "or Intel Customer Support: linux-nics@intel.com\n");
	printk(KERN_ERR "/*********************/\n");

	kfree(data);
}

L
Linus Torvalds 已提交
911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928
/**
 * e1000_probe - Device Initialization Routine
 * @pdev: PCI device information struct
 * @ent: entry in e1000_pci_tbl
 *
 * Returns 0 on success, negative on failure
 *
 * e1000_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
e1000_probe(struct pci_dev *pdev,
            const struct pci_device_id *ent)
{
	struct net_device *netdev;
	struct e1000_adapter *adapter;
929

L
Linus Torvalds 已提交
930
	static int cards_found = 0;
931
	static int global_quad_port_a = 0; /* global ksp3 port a indication */
932
	int i, err, pci_using_dac;
933
	uint16_t eeprom_data = 0;
L
Linus Torvalds 已提交
934
	uint16_t eeprom_apme_mask = E1000_EEPROM_APME;
935 936
	DECLARE_MAC_BUF(mac);

J
Jesse Brandeburg 已提交
937
	if ((err = pci_enable_device(pdev)))
L
Linus Torvalds 已提交
938 939
		return err;

940 941
	if (!(err = pci_set_dma_mask(pdev, DMA_64BIT_MASK)) &&
	    !(err = pci_set_consistent_dma_mask(pdev, DMA_64BIT_MASK))) {
L
Linus Torvalds 已提交
942 943
		pci_using_dac = 1;
	} else {
944 945
		if ((err = pci_set_dma_mask(pdev, DMA_32BIT_MASK)) &&
		    (err = pci_set_consistent_dma_mask(pdev, DMA_32BIT_MASK))) {
L
Linus Torvalds 已提交
946
			E1000_ERR("No usable DMA configuration, aborting\n");
947
			goto err_dma;
L
Linus Torvalds 已提交
948 949 950 951
		}
		pci_using_dac = 0;
	}

J
Jesse Brandeburg 已提交
952
	if ((err = pci_request_regions(pdev, e1000_driver_name)))
953
		goto err_pci_reg;
L
Linus Torvalds 已提交
954 955 956

	pci_set_master(pdev);

957
	err = -ENOMEM;
L
Linus Torvalds 已提交
958
	netdev = alloc_etherdev(sizeof(struct e1000_adapter));
959
	if (!netdev)
L
Linus Torvalds 已提交
960 961 962 963 964
		goto err_alloc_etherdev;

	SET_NETDEV_DEV(netdev, &pdev->dev);

	pci_set_drvdata(pdev, netdev);
965
	adapter = netdev_priv(netdev);
L
Linus Torvalds 已提交
966 967 968 969 970
	adapter->netdev = netdev;
	adapter->pdev = pdev;
	adapter->hw.back = adapter;
	adapter->msg_enable = (1 << debug) - 1;

971
	err = -EIO;
972 973
	adapter->hw.hw_addr = ioremap(pci_resource_start(pdev, BAR_0),
				      pci_resource_len(pdev, BAR_0));
974
	if (!adapter->hw.hw_addr)
L
Linus Torvalds 已提交
975 976
		goto err_ioremap;

J
Jesse Brandeburg 已提交
977 978
	for (i = BAR_1; i <= BAR_5; i++) {
		if (pci_resource_len(pdev, i) == 0)
L
Linus Torvalds 已提交
979
			continue;
J
Jesse Brandeburg 已提交
980
		if (pci_resource_flags(pdev, i) & IORESOURCE_IO) {
L
Linus Torvalds 已提交
981 982 983 984 985 986 987 988 989
			adapter->hw.io_base = pci_resource_start(pdev, i);
			break;
		}
	}

	netdev->open = &e1000_open;
	netdev->stop = &e1000_close;
	netdev->hard_start_xmit = &e1000_xmit_frame;
	netdev->get_stats = &e1000_get_stats;
990
	netdev->set_rx_mode = &e1000_set_rx_mode;
L
Linus Torvalds 已提交
991 992 993 994 995 996 997
	netdev->set_mac_address = &e1000_set_mac;
	netdev->change_mtu = &e1000_change_mtu;
	netdev->do_ioctl = &e1000_ioctl;
	e1000_set_ethtool_ops(netdev);
	netdev->tx_timeout = &e1000_tx_timeout;
	netdev->watchdog_timeo = 5 * HZ;
#ifdef CONFIG_E1000_NAPI
998
	netif_napi_add(netdev, &adapter->napi, e1000_clean, 64);
L
Linus Torvalds 已提交
999 1000 1001 1002 1003 1004 1005
#endif
	netdev->vlan_rx_register = e1000_vlan_rx_register;
	netdev->vlan_rx_add_vid = e1000_vlan_rx_add_vid;
	netdev->vlan_rx_kill_vid = e1000_vlan_rx_kill_vid;
#ifdef CONFIG_NET_POLL_CONTROLLER
	netdev->poll_controller = e1000_netpoll;
#endif
1006
	strncpy(netdev->name, pci_name(pdev), sizeof(netdev->name) - 1);
L
Linus Torvalds 已提交
1007 1008 1009 1010 1011

	adapter->bd_number = cards_found;

	/* setup the private structure */

J
Jesse Brandeburg 已提交
1012
	if ((err = e1000_sw_init(adapter)))
L
Linus Torvalds 已提交
1013 1014
		goto err_sw_init;

1015
	err = -EIO;
1016 1017 1018 1019
	/* Flash BAR mapping must happen after e1000_sw_init
	 * because it depends on mac_type */
	if ((adapter->hw.mac_type == e1000_ich8lan) &&
	   (pci_resource_flags(pdev, 1) & IORESOURCE_MEM)) {
1020 1021 1022
		adapter->hw.flash_address =
			ioremap(pci_resource_start(pdev, 1),
				pci_resource_len(pdev, 1));
1023
		if (!adapter->hw.flash_address)
1024 1025 1026
			goto err_flashmap;
	}

1027
	if (e1000_check_phy_reset_block(&adapter->hw))
1028 1029
		DPRINTK(PROBE, INFO, "PHY reset is blocked due to SOL/IDER session.\n");

J
Jesse Brandeburg 已提交
1030
	if (adapter->hw.mac_type >= e1000_82543) {
L
Linus Torvalds 已提交
1031 1032 1033 1034 1035
		netdev->features = NETIF_F_SG |
				   NETIF_F_HW_CSUM |
				   NETIF_F_HW_VLAN_TX |
				   NETIF_F_HW_VLAN_RX |
				   NETIF_F_HW_VLAN_FILTER;
1036 1037
		if (adapter->hw.mac_type == e1000_ich8lan)
			netdev->features &= ~NETIF_F_HW_VLAN_FILTER;
L
Linus Torvalds 已提交
1038 1039
	}

J
Jesse Brandeburg 已提交
1040
	if ((adapter->hw.mac_type >= e1000_82544) &&
L
Linus Torvalds 已提交
1041 1042
	   (adapter->hw.mac_type != e1000_82547))
		netdev->features |= NETIF_F_TSO;
1043

J
Jesse Brandeburg 已提交
1044
	if (adapter->hw.mac_type > e1000_82547_rev_2)
A
Auke Kok 已提交
1045
		netdev->features |= NETIF_F_TSO6;
J
Jesse Brandeburg 已提交
1046
	if (pci_using_dac)
L
Linus Torvalds 已提交
1047 1048
		netdev->features |= NETIF_F_HIGHDMA;

1049 1050
	netdev->features |= NETIF_F_LLTX;

1051 1052
	adapter->en_mng_pt = e1000_enable_mng_pass_thru(&adapter->hw);

1053 1054 1055
	/* initialize eeprom parameters */
	if (e1000_init_eeprom_params(&adapter->hw)) {
		E1000_ERR("EEPROM initialization failed\n");
1056
		goto err_eeprom;
1057 1058
	}

J
Jesse Brandeburg 已提交
1059
	/* before reading the EEPROM, reset the controller to
L
Linus Torvalds 已提交
1060
	 * put the device in a known good starting state */
J
Jesse Brandeburg 已提交
1061

L
Linus Torvalds 已提交
1062 1063 1064
	e1000_reset_hw(&adapter->hw);

	/* make sure the EEPROM is good */
J
Jesse Brandeburg 已提交
1065
	if (e1000_validate_eeprom_checksum(&adapter->hw) < 0) {
L
Linus Torvalds 已提交
1066
		DPRINTK(PROBE, ERR, "The EEPROM Checksum Is Not Valid\n");
1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080
		e1000_dump_eeprom(adapter);
		/*
		 * set MAC address to all zeroes to invalidate and temporary
		 * disable this device for the user. This blocks regular
		 * traffic while still permitting ethtool ioctls from reaching
		 * the hardware as well as allowing the user to run the
		 * interface after manually setting a hw addr using
		 * `ip set address`
		 */
		memset(adapter->hw.mac_addr, 0, netdev->addr_len);
	} else {
		/* copy the MAC address out of the EEPROM */
		if (e1000_read_mac_addr(&adapter->hw))
			DPRINTK(PROBE, ERR, "EEPROM Read Error\n");
L
Linus Torvalds 已提交
1081
	}
1082
	/* don't block initalization here due to bad MAC address */
L
Linus Torvalds 已提交
1083
	memcpy(netdev->dev_addr, adapter->hw.mac_addr, netdev->addr_len);
1084
	memcpy(netdev->perm_addr, adapter->hw.mac_addr, netdev->addr_len);
L
Linus Torvalds 已提交
1085

1086
	if (!is_valid_ether_addr(netdev->perm_addr))
L
Linus Torvalds 已提交
1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102
		DPRINTK(PROBE, ERR, "Invalid MAC Address\n");

	e1000_get_bus_info(&adapter->hw);

	init_timer(&adapter->tx_fifo_stall_timer);
	adapter->tx_fifo_stall_timer.function = &e1000_82547_tx_fifo_stall;
	adapter->tx_fifo_stall_timer.data = (unsigned long) adapter;

	init_timer(&adapter->watchdog_timer);
	adapter->watchdog_timer.function = &e1000_watchdog;
	adapter->watchdog_timer.data = (unsigned long) adapter;

	init_timer(&adapter->phy_info_timer);
	adapter->phy_info_timer.function = &e1000_update_phy_info;
	adapter->phy_info_timer.data = (unsigned long) adapter;

1103
	INIT_WORK(&adapter->reset_task, e1000_reset_task);
L
Linus Torvalds 已提交
1104 1105 1106 1107 1108 1109 1110 1111

	e1000_check_options(adapter);

	/* Initial Wake on LAN setting
	 * If APM wake is enabled in the EEPROM,
	 * enable the ACPI Magic Packet filter
	 */

J
Jesse Brandeburg 已提交
1112
	switch (adapter->hw.mac_type) {
L
Linus Torvalds 已提交
1113 1114 1115 1116 1117 1118 1119 1120 1121
	case e1000_82542_rev2_0:
	case e1000_82542_rev2_1:
	case e1000_82543:
		break;
	case e1000_82544:
		e1000_read_eeprom(&adapter->hw,
			EEPROM_INIT_CONTROL2_REG, 1, &eeprom_data);
		eeprom_apme_mask = E1000_EEPROM_82544_APM;
		break;
1122 1123 1124 1125 1126
	case e1000_ich8lan:
		e1000_read_eeprom(&adapter->hw,
			EEPROM_INIT_CONTROL1_REG, 1, &eeprom_data);
		eeprom_apme_mask = E1000_EEPROM_ICH8_APME;
		break;
L
Linus Torvalds 已提交
1127 1128
	case e1000_82546:
	case e1000_82546_rev_3:
J
Jeff Kirsher 已提交
1129
	case e1000_82571:
1130
	case e1000_80003es2lan:
J
Jesse Brandeburg 已提交
1131
		if (E1000_READ_REG(&adapter->hw, STATUS) & E1000_STATUS_FUNC_1){
L
Linus Torvalds 已提交
1132 1133 1134 1135 1136 1137 1138 1139 1140 1141
			e1000_read_eeprom(&adapter->hw,
				EEPROM_INIT_CONTROL3_PORT_B, 1, &eeprom_data);
			break;
		}
		/* Fall Through */
	default:
		e1000_read_eeprom(&adapter->hw,
			EEPROM_INIT_CONTROL3_PORT_A, 1, &eeprom_data);
		break;
	}
J
Jesse Brandeburg 已提交
1142
	if (eeprom_data & eeprom_apme_mask)
1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160
		adapter->eeprom_wol |= E1000_WUFC_MAG;

	/* now that we have the eeprom settings, apply the special cases
	 * where the eeprom may be wrong or the board simply won't support
	 * wake on lan on a particular port */
	switch (pdev->device) {
	case E1000_DEV_ID_82546GB_PCIE:
		adapter->eeprom_wol = 0;
		break;
	case E1000_DEV_ID_82546EB_FIBER:
	case E1000_DEV_ID_82546GB_FIBER:
	case E1000_DEV_ID_82571EB_FIBER:
		/* Wake events only supported on port A for dual fiber
		 * regardless of eeprom setting */
		if (E1000_READ_REG(&adapter->hw, STATUS) & E1000_STATUS_FUNC_1)
			adapter->eeprom_wol = 0;
		break;
	case E1000_DEV_ID_82546GB_QUAD_COPPER_KSP3:
1161
	case E1000_DEV_ID_82571EB_QUAD_COPPER:
1162
	case E1000_DEV_ID_82571EB_QUAD_FIBER:
A
Auke Kok 已提交
1163
	case E1000_DEV_ID_82571EB_QUAD_COPPER_LOWPROFILE:
1164
	case E1000_DEV_ID_82571PT_QUAD_COPPER:
1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177
		/* if quad port adapter, disable WoL on all but port A */
		if (global_quad_port_a != 0)
			adapter->eeprom_wol = 0;
		else
			adapter->quad_port_a = 1;
		/* Reset for multiple quad port adapters */
		if (++global_quad_port_a == 4)
			global_quad_port_a = 0;
		break;
	}

	/* initialize the wol settings based on the eeprom settings */
	adapter->wol = adapter->eeprom_wol;
L
Linus Torvalds 已提交
1178

1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195
	/* print bus type/speed/width info */
	{
	struct e1000_hw *hw = &adapter->hw;
	DPRINTK(PROBE, INFO, "(PCI%s:%s:%s) ",
		((hw->bus_type == e1000_bus_type_pcix) ? "-X" :
		 (hw->bus_type == e1000_bus_type_pci_express ? " Express":"")),
		((hw->bus_speed == e1000_bus_speed_2500) ? "2.5Gb/s" :
		 (hw->bus_speed == e1000_bus_speed_133) ? "133MHz" :
		 (hw->bus_speed == e1000_bus_speed_120) ? "120MHz" :
		 (hw->bus_speed == e1000_bus_speed_100) ? "100MHz" :
		 (hw->bus_speed == e1000_bus_speed_66) ? "66MHz" : "33MHz"),
		((hw->bus_width == e1000_bus_width_64) ? "64-bit" :
		 (hw->bus_width == e1000_bus_width_pciex_4) ? "Width x4" :
		 (hw->bus_width == e1000_bus_width_pciex_1) ? "Width x1" :
		 "32-bit"));
	}

1196
	printk("%s\n", print_mac(mac, netdev->dev_addr));
1197

L
Linus Torvalds 已提交
1198 1199 1200
	/* reset the hardware with the new settings */
	e1000_reset(adapter);

1201 1202 1203 1204 1205 1206 1207
	/* If the controller is 82573 and f/w is AMT, do not set
	 * DRV_LOAD until the interface is up.  For all other cases,
	 * let the f/w know that the h/w is now under the control
	 * of the driver. */
	if (adapter->hw.mac_type != e1000_82573 ||
	    !e1000_check_mng_mode(&adapter->hw))
		e1000_get_hw_control(adapter);
1208

A
Auke Kok 已提交
1209 1210 1211
	/* tell the stack to leave us alone until e1000_open() is called */
	netif_carrier_off(netdev);
	netif_stop_queue(netdev);
1212 1213 1214 1215

	strcpy(netdev->name, "eth%d");
	if ((err = register_netdev(netdev)))
		goto err_register;
A
Auke Kok 已提交
1216

L
Linus Torvalds 已提交
1217 1218 1219 1220 1221 1222
	DPRINTK(PROBE, INFO, "Intel(R) PRO/1000 Network Connection\n");

	cards_found++;
	return 0;

err_register:
1223 1224 1225 1226 1227
	e1000_release_hw_control(adapter);
err_eeprom:
	if (!e1000_check_phy_reset_block(&adapter->hw))
		e1000_phy_hw_reset(&adapter->hw);

1228 1229 1230
	if (adapter->hw.flash_address)
		iounmap(adapter->hw.flash_address);
err_flashmap:
1231 1232 1233 1234 1235 1236 1237 1238 1239 1240
#ifdef CONFIG_E1000_NAPI
	for (i = 0; i < adapter->num_rx_queues; i++)
		dev_put(&adapter->polling_netdev[i]);
#endif

	kfree(adapter->tx_ring);
	kfree(adapter->rx_ring);
#ifdef CONFIG_E1000_NAPI
	kfree(adapter->polling_netdev);
#endif
L
Linus Torvalds 已提交
1241 1242 1243 1244 1245 1246
err_sw_init:
	iounmap(adapter->hw.hw_addr);
err_ioremap:
	free_netdev(netdev);
err_alloc_etherdev:
	pci_release_regions(pdev);
1247 1248 1249
err_pci_reg:
err_dma:
	pci_disable_device(pdev);
L
Linus Torvalds 已提交
1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266
	return err;
}

/**
 * e1000_remove - Device Removal Routine
 * @pdev: PCI device information struct
 *
 * e1000_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
e1000_remove(struct pci_dev *pdev)
{
	struct net_device *netdev = pci_get_drvdata(pdev);
1267
	struct e1000_adapter *adapter = netdev_priv(netdev);
1268 1269 1270
#ifdef CONFIG_E1000_NAPI
	int i;
#endif
L
Linus Torvalds 已提交
1271

1272
	cancel_work_sync(&adapter->reset_task);
J
Jeff Garzik 已提交
1273

1274
	e1000_release_manageability(adapter);
L
Linus Torvalds 已提交
1275

1276 1277 1278
	/* Release control of h/w to f/w.  If f/w is AMT enabled, this
	 * would have already happened in close and is redundant. */
	e1000_release_hw_control(adapter);
1279

1280
#ifdef CONFIG_E1000_NAPI
1281
	for (i = 0; i < adapter->num_rx_queues; i++)
1282
		dev_put(&adapter->polling_netdev[i]);
1283
#endif
L
Linus Torvalds 已提交
1284

1285 1286
	unregister_netdev(netdev);

J
Jesse Brandeburg 已提交
1287
	if (!e1000_check_phy_reset_block(&adapter->hw))
1288
		e1000_phy_hw_reset(&adapter->hw);
L
Linus Torvalds 已提交
1289

1290 1291 1292 1293 1294 1295
	kfree(adapter->tx_ring);
	kfree(adapter->rx_ring);
#ifdef CONFIG_E1000_NAPI
	kfree(adapter->polling_netdev);
#endif

L
Linus Torvalds 已提交
1296
	iounmap(adapter->hw.hw_addr);
1297 1298
	if (adapter->hw.flash_address)
		iounmap(adapter->hw.flash_address);
L
Linus Torvalds 已提交
1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320
	pci_release_regions(pdev);

	free_netdev(netdev);

	pci_disable_device(pdev);
}

/**
 * e1000_sw_init - Initialize general software structures (struct e1000_adapter)
 * @adapter: board private structure to initialize
 *
 * e1000_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
e1000_sw_init(struct e1000_adapter *adapter)
{
	struct e1000_hw *hw = &adapter->hw;
	struct net_device *netdev = adapter->netdev;
	struct pci_dev *pdev = adapter->pdev;
1321 1322 1323
#ifdef CONFIG_E1000_NAPI
	int i;
#endif
L
Linus Torvalds 已提交
1324 1325 1326 1327 1328 1329 1330

	/* PCI config space info */

	hw->vendor_id = pdev->vendor;
	hw->device_id = pdev->device;
	hw->subsystem_vendor_id = pdev->subsystem_vendor;
	hw->subsystem_id = pdev->subsystem_device;
1331
	hw->revision_id = pdev->revision;
L
Linus Torvalds 已提交
1332 1333 1334

	pci_read_config_word(pdev, PCI_COMMAND, &hw->pci_cmd_word);

1335
	adapter->rx_buffer_len = MAXIMUM_ETHERNET_VLAN_SIZE;
1336
	adapter->rx_ps_bsize0 = E1000_RXBUFFER_128;
L
Linus Torvalds 已提交
1337 1338 1339 1340 1341 1342
	hw->max_frame_size = netdev->mtu +
			     ENET_HEADER_SIZE + ETHERNET_FCS_SIZE;
	hw->min_frame_size = MINIMUM_ETHERNET_FRAME_SIZE;

	/* identify the MAC */

J
Jesse Brandeburg 已提交
1343
	if (e1000_set_mac_type(hw)) {
L
Linus Torvalds 已提交
1344 1345 1346 1347
		DPRINTK(PROBE, ERR, "Unknown MAC Type\n");
		return -EIO;
	}

J
Jesse Brandeburg 已提交
1348
	switch (hw->mac_type) {
L
Linus Torvalds 已提交
1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366
	default:
		break;
	case e1000_82541:
	case e1000_82547:
	case e1000_82541_rev_2:
	case e1000_82547_rev_2:
		hw->phy_init_script = 1;
		break;
	}

	e1000_set_media_type(hw);

	hw->wait_autoneg_complete = FALSE;
	hw->tbi_compatibility_en = TRUE;
	hw->adaptive_ifs = TRUE;

	/* Copper options */

J
Jesse Brandeburg 已提交
1367
	if (hw->media_type == e1000_media_type_copper) {
L
Linus Torvalds 已提交
1368 1369 1370 1371 1372
		hw->mdix = AUTO_ALL_MODES;
		hw->disable_polarity_correction = FALSE;
		hw->master_slave = E1000_MASTER_SLAVE;
	}

1373 1374
	adapter->num_tx_queues = 1;
	adapter->num_rx_queues = 1;
1375 1376 1377 1378 1379 1380 1381

	if (e1000_alloc_queues(adapter)) {
		DPRINTK(PROBE, ERR, "Unable to allocate memory for queues\n");
		return -ENOMEM;
	}

#ifdef CONFIG_E1000_NAPI
1382
	for (i = 0; i < adapter->num_rx_queues; i++) {
1383 1384 1385 1386
		adapter->polling_netdev[i].priv = adapter;
		dev_hold(&adapter->polling_netdev[i]);
		set_bit(__LINK_STATE_START, &adapter->polling_netdev[i].state);
	}
1387
	spin_lock_init(&adapter->tx_queue_lock);
1388 1389
#endif

1390 1391 1392 1393
	/* Explicitly disable IRQ since the NIC can be in any state. */
	atomic_set(&adapter->irq_sem, 0);
	e1000_irq_disable(adapter);

L
Linus Torvalds 已提交
1394 1395
	spin_lock_init(&adapter->stats_lock);

A
Auke Kok 已提交
1396 1397
	set_bit(__E1000_DOWN, &adapter->flags);

L
Linus Torvalds 已提交
1398 1399 1400
	return 0;
}

1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412
/**
 * e1000_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
 * number of queues at compile-time.  The polling_netdev array is
 * intended for Multiqueue, but should work fine with a single queue.
 **/

static int __devinit
e1000_alloc_queues(struct e1000_adapter *adapter)
{
Y
Yan Burman 已提交
1413 1414
	adapter->tx_ring = kcalloc(adapter->num_tx_queues,
	                           sizeof(struct e1000_tx_ring), GFP_KERNEL);
1415 1416 1417
	if (!adapter->tx_ring)
		return -ENOMEM;

Y
Yan Burman 已提交
1418 1419
	adapter->rx_ring = kcalloc(adapter->num_rx_queues,
	                           sizeof(struct e1000_rx_ring), GFP_KERNEL);
1420 1421 1422 1423 1424 1425
	if (!adapter->rx_ring) {
		kfree(adapter->tx_ring);
		return -ENOMEM;
	}

#ifdef CONFIG_E1000_NAPI
Y
Yan Burman 已提交
1426 1427 1428
	adapter->polling_netdev = kcalloc(adapter->num_rx_queues,
	                                  sizeof(struct net_device),
	                                  GFP_KERNEL);
1429 1430 1431 1432 1433 1434 1435 1436 1437 1438
	if (!adapter->polling_netdev) {
		kfree(adapter->tx_ring);
		kfree(adapter->rx_ring);
		return -ENOMEM;
	}
#endif

	return E1000_SUCCESS;
}

L
Linus Torvalds 已提交
1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454
/**
 * e1000_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
e1000_open(struct net_device *netdev)
{
1455
	struct e1000_adapter *adapter = netdev_priv(netdev);
L
Linus Torvalds 已提交
1456 1457
	int err;

1458
	/* disallow open during test */
A
Auke Kok 已提交
1459
	if (test_bit(__E1000_TESTING, &adapter->flags))
1460 1461
		return -EBUSY;

L
Linus Torvalds 已提交
1462
	/* allocate transmit descriptors */
1463 1464
	err = e1000_setup_all_tx_resources(adapter);
	if (err)
L
Linus Torvalds 已提交
1465 1466 1467
		goto err_setup_tx;

	/* allocate receive descriptors */
1468
	err = e1000_setup_all_rx_resources(adapter);
1469
	if (err)
1470
		goto err_setup_rx;
1471

1472 1473
	e1000_power_up_phy(adapter);

1474
	adapter->mng_vlan_id = E1000_MNG_VLAN_NONE;
J
Jesse Brandeburg 已提交
1475
	if ((adapter->hw.mng_cookie.status &
1476 1477 1478
			  E1000_MNG_DHCP_COOKIE_STATUS_VLAN_SUPPORT)) {
		e1000_update_mng_vlan(adapter);
	}
L
Linus Torvalds 已提交
1479

1480 1481 1482 1483 1484 1485
	/* If AMT is enabled, let the firmware know that the network
	 * interface is now open */
	if (adapter->hw.mac_type == e1000_82573 &&
	    e1000_check_mng_mode(&adapter->hw))
		e1000_get_hw_control(adapter);

1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498
	/* before we allocate an interrupt, we must be ready to handle it.
	 * Setting DEBUG_SHIRQ in the kernel makes it fire an interrupt
	 * as soon as we call pci_request_irq, so we have to setup our
	 * clean_rx handler before we do so.  */
	e1000_configure(adapter);

	err = e1000_request_irq(adapter);
	if (err)
		goto err_req_irq;

	/* From here on the code is the same as e1000_up() */
	clear_bit(__E1000_DOWN, &adapter->flags);

1499
#ifdef CONFIG_E1000_NAPI
1500
	napi_enable(&adapter->napi);
1501 1502
#endif

1503 1504 1505 1506 1507
	e1000_irq_enable(adapter);

	/* fire a link status change interrupt to start the watchdog */
	E1000_WRITE_REG(&adapter->hw, ICS, E1000_ICS_LSC);

L
Linus Torvalds 已提交
1508 1509
	return E1000_SUCCESS;

1510
err_req_irq:
1511 1512
	e1000_release_hw_control(adapter);
	e1000_power_down_phy(adapter);
1513
	e1000_free_all_rx_resources(adapter);
L
Linus Torvalds 已提交
1514
err_setup_rx:
1515
	e1000_free_all_tx_resources(adapter);
L
Linus Torvalds 已提交
1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536
err_setup_tx:
	e1000_reset(adapter);

	return err;
}

/**
 * e1000_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
e1000_close(struct net_device *netdev)
{
1537
	struct e1000_adapter *adapter = netdev_priv(netdev);
L
Linus Torvalds 已提交
1538

1539
	WARN_ON(test_bit(__E1000_RESETTING, &adapter->flags));
L
Linus Torvalds 已提交
1540
	e1000_down(adapter);
1541
	e1000_power_down_phy(adapter);
1542
	e1000_free_irq(adapter);
L
Linus Torvalds 已提交
1543

1544 1545
	e1000_free_all_tx_resources(adapter);
	e1000_free_all_rx_resources(adapter);
L
Linus Torvalds 已提交
1546

1547 1548
	/* kill manageability vlan ID if supported, but not if a vlan with
	 * the same ID is registered on the host OS (let 8021q kill it) */
J
Jesse Brandeburg 已提交
1549
	if ((adapter->hw.mng_cookie.status &
1550 1551
			  E1000_MNG_DHCP_COOKIE_STATUS_VLAN_SUPPORT) &&
	     !(adapter->vlgrp &&
D
Dan Aloni 已提交
1552
	       vlan_group_get_device(adapter->vlgrp, adapter->mng_vlan_id))) {
1553 1554
		e1000_vlan_rx_kill_vid(netdev, adapter->mng_vlan_id);
	}
1555 1556 1557 1558 1559 1560 1561

	/* If AMT is enabled, let the firmware know that the network
	 * interface is now closed */
	if (adapter->hw.mac_type == e1000_82573 &&
	    e1000_check_mng_mode(&adapter->hw))
		e1000_release_hw_control(adapter);

L
Linus Torvalds 已提交
1562 1563 1564 1565 1566 1567
	return 0;
}

/**
 * e1000_check_64k_bound - check that memory doesn't cross 64kB boundary
 * @adapter: address of board private structure
1568 1569
 * @start: address of beginning of memory
 * @len: length of memory
L
Linus Torvalds 已提交
1570
 **/
1571
static boolean_t
L
Linus Torvalds 已提交
1572 1573 1574 1575 1576 1577
e1000_check_64k_bound(struct e1000_adapter *adapter,
		      void *start, unsigned long len)
{
	unsigned long begin = (unsigned long) start;
	unsigned long end = begin + len;

1578 1579
	/* First rev 82545 and 82546 need to not allow any memory
	 * write location to cross 64k boundary due to errata 23 */
L
Linus Torvalds 已提交
1580
	if (adapter->hw.mac_type == e1000_82545 ||
1581
	    adapter->hw.mac_type == e1000_82546) {
L
Linus Torvalds 已提交
1582 1583 1584 1585 1586 1587 1588 1589 1590
		return ((begin ^ (end - 1)) >> 16) != 0 ? FALSE : TRUE;
	}

	return TRUE;
}

/**
 * e1000_setup_tx_resources - allocate Tx resources (Descriptors)
 * @adapter: board private structure
1591
 * @txdr:    tx descriptor ring (for a specific queue) to setup
L
Linus Torvalds 已提交
1592 1593 1594 1595
 *
 * Return 0 on success, negative on failure
 **/

1596
static int
1597 1598
e1000_setup_tx_resources(struct e1000_adapter *adapter,
                         struct e1000_tx_ring *txdr)
L
Linus Torvalds 已提交
1599 1600 1601 1602 1603
{
	struct pci_dev *pdev = adapter->pdev;
	int size;

	size = sizeof(struct e1000_buffer) * txdr->count;
1604
	txdr->buffer_info = vmalloc(size);
J
Jesse Brandeburg 已提交
1605
	if (!txdr->buffer_info) {
1606 1607
		DPRINTK(PROBE, ERR,
		"Unable to allocate memory for the transmit descriptor ring\n");
L
Linus Torvalds 已提交
1608 1609 1610 1611 1612 1613 1614
		return -ENOMEM;
	}
	memset(txdr->buffer_info, 0, size);

	/* round up to nearest 4K */

	txdr->size = txdr->count * sizeof(struct e1000_tx_desc);
1615
	txdr->size = ALIGN(txdr->size, 4096);
L
Linus Torvalds 已提交
1616 1617

	txdr->desc = pci_alloc_consistent(pdev, txdr->size, &txdr->dma);
J
Jesse Brandeburg 已提交
1618
	if (!txdr->desc) {
L
Linus Torvalds 已提交
1619 1620
setup_tx_desc_die:
		vfree(txdr->buffer_info);
1621 1622
		DPRINTK(PROBE, ERR,
		"Unable to allocate memory for the transmit descriptor ring\n");
L
Linus Torvalds 已提交
1623 1624 1625
		return -ENOMEM;
	}

1626
	/* Fix for errata 23, can't cross 64kB boundary */
L
Linus Torvalds 已提交
1627 1628 1629
	if (!e1000_check_64k_bound(adapter, txdr->desc, txdr->size)) {
		void *olddesc = txdr->desc;
		dma_addr_t olddma = txdr->dma;
1630 1631 1632
		DPRINTK(TX_ERR, ERR, "txdr align check failed: %u bytes "
				     "at %p\n", txdr->size, txdr->desc);
		/* Try again, without freeing the previous */
L
Linus Torvalds 已提交
1633
		txdr->desc = pci_alloc_consistent(pdev, txdr->size, &txdr->dma);
1634
		/* Failed allocation, critical failure */
J
Jesse Brandeburg 已提交
1635
		if (!txdr->desc) {
L
Linus Torvalds 已提交
1636 1637 1638 1639 1640 1641
			pci_free_consistent(pdev, txdr->size, olddesc, olddma);
			goto setup_tx_desc_die;
		}

		if (!e1000_check_64k_bound(adapter, txdr->desc, txdr->size)) {
			/* give up */
1642 1643
			pci_free_consistent(pdev, txdr->size, txdr->desc,
					    txdr->dma);
L
Linus Torvalds 已提交
1644 1645
			pci_free_consistent(pdev, txdr->size, olddesc, olddma);
			DPRINTK(PROBE, ERR,
1646 1647
				"Unable to allocate aligned memory "
				"for the transmit descriptor ring\n");
L
Linus Torvalds 已提交
1648 1649 1650
			vfree(txdr->buffer_info);
			return -ENOMEM;
		} else {
1651
			/* Free old allocation, new allocation was successful */
L
Linus Torvalds 已提交
1652 1653 1654 1655 1656 1657 1658
			pci_free_consistent(pdev, txdr->size, olddesc, olddma);
		}
	}
	memset(txdr->desc, 0, txdr->size);

	txdr->next_to_use = 0;
	txdr->next_to_clean = 0;
1659
	spin_lock_init(&txdr->tx_lock);
L
Linus Torvalds 已提交
1660 1661 1662 1663

	return 0;
}

1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676
/**
 * e1000_setup_all_tx_resources - wrapper to allocate Tx resources
 * 				  (Descriptors) for all queues
 * @adapter: board private structure
 *
 * Return 0 on success, negative on failure
 **/

int
e1000_setup_all_tx_resources(struct e1000_adapter *adapter)
{
	int i, err = 0;

1677
	for (i = 0; i < adapter->num_tx_queues; i++) {
1678 1679 1680 1681
		err = e1000_setup_tx_resources(adapter, &adapter->tx_ring[i]);
		if (err) {
			DPRINTK(PROBE, ERR,
				"Allocation for Tx Queue %u failed\n", i);
1682 1683 1684
			for (i-- ; i >= 0; i--)
				e1000_free_tx_resources(adapter,
							&adapter->tx_ring[i]);
1685 1686 1687 1688 1689 1690 1691
			break;
		}
	}

	return err;
}

L
Linus Torvalds 已提交
1692 1693 1694 1695 1696 1697 1698 1699 1700 1701
/**
 * e1000_configure_tx - Configure 8254x Transmit Unit after Reset
 * @adapter: board private structure
 *
 * Configure the Tx unit of the MAC after a reset.
 **/

static void
e1000_configure_tx(struct e1000_adapter *adapter)
{
1702 1703 1704
	uint64_t tdba;
	struct e1000_hw *hw = &adapter->hw;
	uint32_t tdlen, tctl, tipg, tarc;
1705
	uint32_t ipgr1, ipgr2;
L
Linus Torvalds 已提交
1706 1707 1708

	/* Setup the HW Tx Head and Tail descriptor pointers */

1709
	switch (adapter->num_tx_queues) {
1710 1711
	case 1:
	default:
1712 1713 1714 1715
		tdba = adapter->tx_ring[0].dma;
		tdlen = adapter->tx_ring[0].count *
			sizeof(struct e1000_tx_desc);
		E1000_WRITE_REG(hw, TDLEN, tdlen);
1716 1717
		E1000_WRITE_REG(hw, TDBAH, (tdba >> 32));
		E1000_WRITE_REG(hw, TDBAL, (tdba & 0x00000000ffffffffULL));
1718
		E1000_WRITE_REG(hw, TDT, 0);
1719
		E1000_WRITE_REG(hw, TDH, 0);
1720 1721
		adapter->tx_ring[0].tdh = ((hw->mac_type >= e1000_82543) ? E1000_TDH : E1000_82542_TDH);
		adapter->tx_ring[0].tdt = ((hw->mac_type >= e1000_82543) ? E1000_TDT : E1000_82542_TDT);
1722 1723
		break;
	}
L
Linus Torvalds 已提交
1724 1725

	/* Set the default values for the Tx Inter Packet Gap timer */
1726 1727 1728
	if (adapter->hw.mac_type <= e1000_82547_rev_2 &&
	    (hw->media_type == e1000_media_type_fiber ||
	     hw->media_type == e1000_media_type_internal_serdes))
1729 1730 1731 1732
		tipg = DEFAULT_82543_TIPG_IPGT_FIBER;
	else
		tipg = DEFAULT_82543_TIPG_IPGT_COPPER;

1733
	switch (hw->mac_type) {
L
Linus Torvalds 已提交
1734 1735 1736
	case e1000_82542_rev2_0:
	case e1000_82542_rev2_1:
		tipg = DEFAULT_82542_TIPG_IPGT;
1737 1738
		ipgr1 = DEFAULT_82542_TIPG_IPGR1;
		ipgr2 = DEFAULT_82542_TIPG_IPGR2;
L
Linus Torvalds 已提交
1739
		break;
1740 1741 1742 1743
	case e1000_80003es2lan:
		ipgr1 = DEFAULT_82543_TIPG_IPGR1;
		ipgr2 = DEFAULT_80003ES2LAN_TIPG_IPGR2;
		break;
L
Linus Torvalds 已提交
1744
	default:
1745 1746 1747
		ipgr1 = DEFAULT_82543_TIPG_IPGR1;
		ipgr2 = DEFAULT_82543_TIPG_IPGR2;
		break;
L
Linus Torvalds 已提交
1748
	}
1749 1750
	tipg |= ipgr1 << E1000_TIPG_IPGR1_SHIFT;
	tipg |= ipgr2 << E1000_TIPG_IPGR2_SHIFT;
1751
	E1000_WRITE_REG(hw, TIPG, tipg);
L
Linus Torvalds 已提交
1752 1753 1754

	/* Set the Tx Interrupt Delay register */

1755 1756 1757
	E1000_WRITE_REG(hw, TIDV, adapter->tx_int_delay);
	if (hw->mac_type >= e1000_82540)
		E1000_WRITE_REG(hw, TADV, adapter->tx_abs_int_delay);
L
Linus Torvalds 已提交
1758 1759 1760

	/* Program the Transmit Control Register */

1761
	tctl = E1000_READ_REG(hw, TCTL);
L
Linus Torvalds 已提交
1762
	tctl &= ~E1000_TCTL_CT;
1763
	tctl |= E1000_TCTL_PSP | E1000_TCTL_RTLC |
L
Linus Torvalds 已提交
1764 1765
		(E1000_COLLISION_THRESHOLD << E1000_CT_SHIFT);

1766 1767
	if (hw->mac_type == e1000_82571 || hw->mac_type == e1000_82572) {
		tarc = E1000_READ_REG(hw, TARC0);
1768 1769
		/* set the speed mode bit, we'll clear it if we're not at
		 * gigabit link later */
1770
		tarc |= (1 << 21);
1771
		E1000_WRITE_REG(hw, TARC0, tarc);
1772 1773 1774 1775 1776 1777 1778
	} else if (hw->mac_type == e1000_80003es2lan) {
		tarc = E1000_READ_REG(hw, TARC0);
		tarc |= 1;
		E1000_WRITE_REG(hw, TARC0, tarc);
		tarc = E1000_READ_REG(hw, TARC1);
		tarc |= 1;
		E1000_WRITE_REG(hw, TARC1, tarc);
1779 1780
	}

1781
	e1000_config_collision_dist(hw);
L
Linus Torvalds 已提交
1782 1783

	/* Setup Transmit Descriptor Settings for eop descriptor */
1784 1785 1786 1787 1788
	adapter->txd_cmd = E1000_TXD_CMD_EOP | E1000_TXD_CMD_IFCS;

	/* only set IDE if we are delaying interrupts using the timers */
	if (adapter->tx_int_delay)
		adapter->txd_cmd |= E1000_TXD_CMD_IDE;
L
Linus Torvalds 已提交
1789

1790
	if (hw->mac_type < e1000_82543)
L
Linus Torvalds 已提交
1791 1792 1793 1794 1795 1796
		adapter->txd_cmd |= E1000_TXD_CMD_RPS;
	else
		adapter->txd_cmd |= E1000_TXD_CMD_RS;

	/* Cache if we're 82544 running in PCI-X because we'll
	 * need this to apply a workaround later in the send path. */
1797 1798
	if (hw->mac_type == e1000_82544 &&
	    hw->bus_type == e1000_bus_type_pcix)
L
Linus Torvalds 已提交
1799
		adapter->pcix_82544 = 1;
1800 1801 1802

	E1000_WRITE_REG(hw, TCTL, tctl);

L
Linus Torvalds 已提交
1803 1804 1805 1806 1807
}

/**
 * e1000_setup_rx_resources - allocate Rx resources (Descriptors)
 * @adapter: board private structure
1808
 * @rxdr:    rx descriptor ring (for a specific queue) to setup
L
Linus Torvalds 已提交
1809 1810 1811 1812
 *
 * Returns 0 on success, negative on failure
 **/

1813
static int
1814 1815
e1000_setup_rx_resources(struct e1000_adapter *adapter,
                         struct e1000_rx_ring *rxdr)
L
Linus Torvalds 已提交
1816 1817
{
	struct pci_dev *pdev = adapter->pdev;
1818
	int size, desc_len;
L
Linus Torvalds 已提交
1819 1820

	size = sizeof(struct e1000_buffer) * rxdr->count;
1821
	rxdr->buffer_info = vmalloc(size);
1822
	if (!rxdr->buffer_info) {
1823 1824
		DPRINTK(PROBE, ERR,
		"Unable to allocate memory for the receive descriptor ring\n");
L
Linus Torvalds 已提交
1825 1826 1827 1828
		return -ENOMEM;
	}
	memset(rxdr->buffer_info, 0, size);

Y
Yan Burman 已提交
1829 1830
	rxdr->ps_page = kcalloc(rxdr->count, sizeof(struct e1000_ps_page),
	                        GFP_KERNEL);
J
Jesse Brandeburg 已提交
1831
	if (!rxdr->ps_page) {
1832 1833 1834 1835 1836 1837
		vfree(rxdr->buffer_info);
		DPRINTK(PROBE, ERR,
		"Unable to allocate memory for the receive descriptor ring\n");
		return -ENOMEM;
	}

Y
Yan Burman 已提交
1838 1839 1840
	rxdr->ps_page_dma = kcalloc(rxdr->count,
	                            sizeof(struct e1000_ps_page_dma),
	                            GFP_KERNEL);
J
Jesse Brandeburg 已提交
1841
	if (!rxdr->ps_page_dma) {
1842 1843 1844 1845 1846 1847 1848
		vfree(rxdr->buffer_info);
		kfree(rxdr->ps_page);
		DPRINTK(PROBE, ERR,
		"Unable to allocate memory for the receive descriptor ring\n");
		return -ENOMEM;
	}

J
Jesse Brandeburg 已提交
1849
	if (adapter->hw.mac_type <= e1000_82547_rev_2)
1850 1851 1852 1853
		desc_len = sizeof(struct e1000_rx_desc);
	else
		desc_len = sizeof(union e1000_rx_desc_packet_split);

L
Linus Torvalds 已提交
1854 1855
	/* Round up to nearest 4K */

1856
	rxdr->size = rxdr->count * desc_len;
1857
	rxdr->size = ALIGN(rxdr->size, 4096);
L
Linus Torvalds 已提交
1858 1859 1860

	rxdr->desc = pci_alloc_consistent(pdev, rxdr->size, &rxdr->dma);

1861 1862 1863
	if (!rxdr->desc) {
		DPRINTK(PROBE, ERR,
		"Unable to allocate memory for the receive descriptor ring\n");
L
Linus Torvalds 已提交
1864 1865
setup_rx_desc_die:
		vfree(rxdr->buffer_info);
1866 1867
		kfree(rxdr->ps_page);
		kfree(rxdr->ps_page_dma);
L
Linus Torvalds 已提交
1868 1869 1870
		return -ENOMEM;
	}

1871
	/* Fix for errata 23, can't cross 64kB boundary */
L
Linus Torvalds 已提交
1872 1873 1874
	if (!e1000_check_64k_bound(adapter, rxdr->desc, rxdr->size)) {
		void *olddesc = rxdr->desc;
		dma_addr_t olddma = rxdr->dma;
1875 1876 1877
		DPRINTK(RX_ERR, ERR, "rxdr align check failed: %u bytes "
				     "at %p\n", rxdr->size, rxdr->desc);
		/* Try again, without freeing the previous */
L
Linus Torvalds 已提交
1878
		rxdr->desc = pci_alloc_consistent(pdev, rxdr->size, &rxdr->dma);
1879
		/* Failed allocation, critical failure */
1880
		if (!rxdr->desc) {
L
Linus Torvalds 已提交
1881
			pci_free_consistent(pdev, rxdr->size, olddesc, olddma);
1882 1883 1884
			DPRINTK(PROBE, ERR,
				"Unable to allocate memory "
				"for the receive descriptor ring\n");
L
Linus Torvalds 已提交
1885 1886 1887 1888 1889
			goto setup_rx_desc_die;
		}

		if (!e1000_check_64k_bound(adapter, rxdr->desc, rxdr->size)) {
			/* give up */
1890 1891
			pci_free_consistent(pdev, rxdr->size, rxdr->desc,
					    rxdr->dma);
L
Linus Torvalds 已提交
1892
			pci_free_consistent(pdev, rxdr->size, olddesc, olddma);
1893 1894 1895
			DPRINTK(PROBE, ERR,
				"Unable to allocate aligned memory "
				"for the receive descriptor ring\n");
1896
			goto setup_rx_desc_die;
L
Linus Torvalds 已提交
1897
		} else {
1898
			/* Free old allocation, new allocation was successful */
L
Linus Torvalds 已提交
1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909
			pci_free_consistent(pdev, rxdr->size, olddesc, olddma);
		}
	}
	memset(rxdr->desc, 0, rxdr->size);

	rxdr->next_to_clean = 0;
	rxdr->next_to_use = 0;

	return 0;
}

1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922
/**
 * e1000_setup_all_rx_resources - wrapper to allocate Rx resources
 * 				  (Descriptors) for all queues
 * @adapter: board private structure
 *
 * Return 0 on success, negative on failure
 **/

int
e1000_setup_all_rx_resources(struct e1000_adapter *adapter)
{
	int i, err = 0;

1923
	for (i = 0; i < adapter->num_rx_queues; i++) {
1924 1925 1926 1927
		err = e1000_setup_rx_resources(adapter, &adapter->rx_ring[i]);
		if (err) {
			DPRINTK(PROBE, ERR,
				"Allocation for Rx Queue %u failed\n", i);
1928 1929 1930
			for (i-- ; i >= 0; i--)
				e1000_free_rx_resources(adapter,
							&adapter->rx_ring[i]);
1931 1932 1933 1934 1935 1936 1937
			break;
		}
	}

	return err;
}

L
Linus Torvalds 已提交
1938
/**
1939
 * e1000_setup_rctl - configure the receive control registers
L
Linus Torvalds 已提交
1940 1941
 * @adapter: Board private structure
 **/
1942 1943
#define PAGE_USE_COUNT(S) (((S) >> PAGE_SHIFT) + \
			(((S) & (PAGE_SIZE - 1)) ? 1 : 0))
L
Linus Torvalds 已提交
1944 1945 1946
static void
e1000_setup_rctl(struct e1000_adapter *adapter)
{
1947 1948
	uint32_t rctl, rfctl;
	uint32_t psrctl = 0;
1949
#ifndef CONFIG_E1000_DISABLE_PACKET_SPLIT
1950 1951
	uint32_t pages = 0;
#endif
L
Linus Torvalds 已提交
1952 1953 1954 1955 1956 1957 1958 1959 1960

	rctl = E1000_READ_REG(&adapter->hw, RCTL);

	rctl &= ~(3 << E1000_RCTL_MO_SHIFT);

	rctl |= E1000_RCTL_EN | E1000_RCTL_BAM |
		E1000_RCTL_LBM_NO | E1000_RCTL_RDMTS_HALF |
		(adapter->hw.mc_filter_type << E1000_RCTL_MO_SHIFT);

1961
	if (adapter->hw.tbi_compatibility_on == 1)
L
Linus Torvalds 已提交
1962 1963 1964 1965
		rctl |= E1000_RCTL_SBP;
	else
		rctl &= ~E1000_RCTL_SBP;

1966 1967 1968 1969 1970
	if (adapter->netdev->mtu <= ETH_DATA_LEN)
		rctl &= ~E1000_RCTL_LPE;
	else
		rctl |= E1000_RCTL_LPE;

L
Linus Torvalds 已提交
1971
	/* Setup buffer sizes */
1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986
	rctl &= ~E1000_RCTL_SZ_4096;
	rctl |= E1000_RCTL_BSEX;
	switch (adapter->rx_buffer_len) {
		case E1000_RXBUFFER_256:
			rctl |= E1000_RCTL_SZ_256;
			rctl &= ~E1000_RCTL_BSEX;
			break;
		case E1000_RXBUFFER_512:
			rctl |= E1000_RCTL_SZ_512;
			rctl &= ~E1000_RCTL_BSEX;
			break;
		case E1000_RXBUFFER_1024:
			rctl |= E1000_RCTL_SZ_1024;
			rctl &= ~E1000_RCTL_BSEX;
			break;
1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000
		case E1000_RXBUFFER_2048:
		default:
			rctl |= E1000_RCTL_SZ_2048;
			rctl &= ~E1000_RCTL_BSEX;
			break;
		case E1000_RXBUFFER_4096:
			rctl |= E1000_RCTL_SZ_4096;
			break;
		case E1000_RXBUFFER_8192:
			rctl |= E1000_RCTL_SZ_8192;
			break;
		case E1000_RXBUFFER_16384:
			rctl |= E1000_RCTL_SZ_16384;
			break;
2001 2002
	}

2003
#ifndef CONFIG_E1000_DISABLE_PACKET_SPLIT
2004 2005 2006 2007 2008 2009 2010
	/* 82571 and greater support packet-split where the protocol
	 * header is placed in skb->data and the packet data is
	 * placed in pages hanging off of skb_shinfo(skb)->nr_frags.
	 * In the case of a non-split, skb->data is linearly filled,
	 * followed by the page buffers.  Therefore, skb->data is
	 * sized to hold the largest protocol header.
	 */
2011 2012
	/* allocations using alloc_page take too long for regular MTU
	 * so only enable packet split for jumbo frames */
2013
	pages = PAGE_USE_COUNT(adapter->netdev->mtu);
2014 2015
	if ((adapter->hw.mac_type >= e1000_82571) && (pages <= 3) &&
	    PAGE_SIZE <= 16384 && (rctl & E1000_RCTL_LPE))
2016 2017 2018
		adapter->rx_ps_pages = pages;
	else
		adapter->rx_ps_pages = 0;
2019
#endif
2020
	if (adapter->rx_ps_pages) {
2021 2022 2023
		/* Configure extra packet-split registers */
		rfctl = E1000_READ_REG(&adapter->hw, RFCTL);
		rfctl |= E1000_RFCTL_EXTEN;
A
Auke Kok 已提交
2024 2025 2026 2027 2028
		/* disable packet split support for IPv6 extension headers,
		 * because some malformed IPv6 headers can hang the RX */
		rfctl |= (E1000_RFCTL_IPV6_EX_DIS |
		          E1000_RFCTL_NEW_IPV6_EXT_DIS);

2029 2030
		E1000_WRITE_REG(&adapter->hw, RFCTL, rfctl);

2031
		rctl |= E1000_RCTL_DTYP_PS;
J
Jesse Brandeburg 已提交
2032

2033 2034
		psrctl |= adapter->rx_ps_bsize0 >>
			E1000_PSRCTL_BSIZE0_SHIFT;
2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047

		switch (adapter->rx_ps_pages) {
		case 3:
			psrctl |= PAGE_SIZE <<
				E1000_PSRCTL_BSIZE3_SHIFT;
		case 2:
			psrctl |= PAGE_SIZE <<
				E1000_PSRCTL_BSIZE2_SHIFT;
		case 1:
			psrctl |= PAGE_SIZE >>
				E1000_PSRCTL_BSIZE1_SHIFT;
			break;
		}
2048 2049

		E1000_WRITE_REG(&adapter->hw, PSRCTL, psrctl);
L
Linus Torvalds 已提交
2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064
	}

	E1000_WRITE_REG(&adapter->hw, RCTL, rctl);
}

/**
 * e1000_configure_rx - Configure 8254x Receive Unit after Reset
 * @adapter: board private structure
 *
 * Configure the Rx unit of the MAC after a reset.
 **/

static void
e1000_configure_rx(struct e1000_adapter *adapter)
{
2065 2066 2067
	uint64_t rdba;
	struct e1000_hw *hw = &adapter->hw;
	uint32_t rdlen, rctl, rxcsum, ctrl_ext;
2068

2069
	if (adapter->rx_ps_pages) {
2070
		/* this is a 32 byte descriptor */
2071
		rdlen = adapter->rx_ring[0].count *
2072 2073 2074 2075
			sizeof(union e1000_rx_desc_packet_split);
		adapter->clean_rx = e1000_clean_rx_irq_ps;
		adapter->alloc_rx_buf = e1000_alloc_rx_buffers_ps;
	} else {
2076 2077
		rdlen = adapter->rx_ring[0].count *
			sizeof(struct e1000_rx_desc);
2078 2079 2080
		adapter->clean_rx = e1000_clean_rx_irq;
		adapter->alloc_rx_buf = e1000_alloc_rx_buffers;
	}
L
Linus Torvalds 已提交
2081 2082

	/* disable receives while setting up the descriptors */
2083 2084
	rctl = E1000_READ_REG(hw, RCTL);
	E1000_WRITE_REG(hw, RCTL, rctl & ~E1000_RCTL_EN);
L
Linus Torvalds 已提交
2085 2086

	/* set the Receive Delay Timer Register */
2087
	E1000_WRITE_REG(hw, RDTR, adapter->rx_int_delay);
L
Linus Torvalds 已提交
2088

2089 2090
	if (hw->mac_type >= e1000_82540) {
		E1000_WRITE_REG(hw, RADV, adapter->rx_abs_int_delay);
J
Jesse Brandeburg 已提交
2091
		if (adapter->itr_setting != 0)
2092
			E1000_WRITE_REG(hw, ITR,
L
Linus Torvalds 已提交
2093 2094 2095
				1000000000 / (adapter->itr * 256));
	}

2096 2097
	if (hw->mac_type >= e1000_82571) {
		ctrl_ext = E1000_READ_REG(hw, CTRL_EXT);
2098
		/* Reset delay timers after every interrupt */
A
Auke Kok 已提交
2099
		ctrl_ext |= E1000_CTRL_EXT_INT_TIMER_CLR;
2100
#ifdef CONFIG_E1000_NAPI
J
Jesse Brandeburg 已提交
2101
		/* Auto-Mask interrupts upon ICR access */
2102
		ctrl_ext |= E1000_CTRL_EXT_IAME;
J
Jesse Brandeburg 已提交
2103
		E1000_WRITE_REG(hw, IAM, 0xffffffff);
2104
#endif
2105 2106 2107 2108
		E1000_WRITE_REG(hw, CTRL_EXT, ctrl_ext);
		E1000_WRITE_FLUSH(hw);
	}

2109 2110
	/* Setup the HW Rx Head and Tail Descriptor Pointers and
	 * the Base and Length of the Rx Descriptor Ring */
2111
	switch (adapter->num_rx_queues) {
2112 2113
	case 1:
	default:
2114 2115
		rdba = adapter->rx_ring[0].dma;
		E1000_WRITE_REG(hw, RDLEN, rdlen);
2116 2117
		E1000_WRITE_REG(hw, RDBAH, (rdba >> 32));
		E1000_WRITE_REG(hw, RDBAL, (rdba & 0x00000000ffffffffULL));
2118
		E1000_WRITE_REG(hw, RDT, 0);
2119
		E1000_WRITE_REG(hw, RDH, 0);
2120 2121
		adapter->rx_ring[0].rdh = ((hw->mac_type >= e1000_82543) ? E1000_RDH : E1000_82542_RDH);
		adapter->rx_ring[0].rdt = ((hw->mac_type >= e1000_82543) ? E1000_RDT : E1000_82542_RDT);
2122
		break;
2123 2124
	}

L
Linus Torvalds 已提交
2125
	/* Enable 82543 Receive Checksum Offload for TCP and UDP */
2126 2127
	if (hw->mac_type >= e1000_82543) {
		rxcsum = E1000_READ_REG(hw, RXCSUM);
J
Jesse Brandeburg 已提交
2128
		if (adapter->rx_csum == TRUE) {
2129 2130
			rxcsum |= E1000_RXCSUM_TUOFL;

2131
			/* Enable 82571 IPv4 payload checksum for UDP fragments
2132
			 * Must be used in conjunction with packet-split. */
J
Jesse Brandeburg 已提交
2133 2134
			if ((hw->mac_type >= e1000_82571) &&
			    (adapter->rx_ps_pages)) {
2135 2136 2137 2138 2139 2140
				rxcsum |= E1000_RXCSUM_IPPCSE;
			}
		} else {
			rxcsum &= ~E1000_RXCSUM_TUOFL;
			/* don't need to clear IPPCSE as it defaults to 0 */
		}
2141
		E1000_WRITE_REG(hw, RXCSUM, rxcsum);
L
Linus Torvalds 已提交
2142 2143
	}

2144 2145 2146 2147 2148 2149
	/* enable early receives on 82573, only takes effect if using > 2048
	 * byte total frame size.  for example only for jumbo frames */
#define E1000_ERT_2048 0x100
	if (hw->mac_type == e1000_82573)
		E1000_WRITE_REG(hw, ERT, E1000_ERT_2048);

L
Linus Torvalds 已提交
2150
	/* Enable Receives */
2151
	E1000_WRITE_REG(hw, RCTL, rctl);
L
Linus Torvalds 已提交
2152 2153 2154
}

/**
2155
 * e1000_free_tx_resources - Free Tx Resources per Queue
L
Linus Torvalds 已提交
2156
 * @adapter: board private structure
2157
 * @tx_ring: Tx descriptor ring for a specific queue
L
Linus Torvalds 已提交
2158 2159 2160 2161
 *
 * Free all transmit software resources
 **/

2162
static void
2163 2164
e1000_free_tx_resources(struct e1000_adapter *adapter,
                        struct e1000_tx_ring *tx_ring)
L
Linus Torvalds 已提交
2165 2166 2167
{
	struct pci_dev *pdev = adapter->pdev;

2168
	e1000_clean_tx_ring(adapter, tx_ring);
L
Linus Torvalds 已提交
2169

2170 2171
	vfree(tx_ring->buffer_info);
	tx_ring->buffer_info = NULL;
L
Linus Torvalds 已提交
2172

2173
	pci_free_consistent(pdev, tx_ring->size, tx_ring->desc, tx_ring->dma);
L
Linus Torvalds 已提交
2174

2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189
	tx_ring->desc = NULL;
}

/**
 * e1000_free_all_tx_resources - Free Tx Resources for All Queues
 * @adapter: board private structure
 *
 * Free all transmit software resources
 **/

void
e1000_free_all_tx_resources(struct e1000_adapter *adapter)
{
	int i;

2190
	for (i = 0; i < adapter->num_tx_queues; i++)
2191
		e1000_free_tx_resources(adapter, &adapter->tx_ring[i]);
L
Linus Torvalds 已提交
2192 2193
}

2194
static void
L
Linus Torvalds 已提交
2195 2196 2197
e1000_unmap_and_free_tx_resource(struct e1000_adapter *adapter,
			struct e1000_buffer *buffer_info)
{
J
Jesse Brandeburg 已提交
2198
	if (buffer_info->dma) {
2199 2200 2201 2202
		pci_unmap_page(adapter->pdev,
				buffer_info->dma,
				buffer_info->length,
				PCI_DMA_TODEVICE);
2203
		buffer_info->dma = 0;
L
Linus Torvalds 已提交
2204
	}
2205
	if (buffer_info->skb) {
L
Linus Torvalds 已提交
2206
		dev_kfree_skb_any(buffer_info->skb);
2207 2208 2209
		buffer_info->skb = NULL;
	}
	/* buffer_info must be completely set up in the transmit path */
L
Linus Torvalds 已提交
2210 2211 2212 2213 2214
}

/**
 * e1000_clean_tx_ring - Free Tx Buffers
 * @adapter: board private structure
2215
 * @tx_ring: ring to be cleaned
L
Linus Torvalds 已提交
2216 2217 2218
 **/

static void
2219 2220
e1000_clean_tx_ring(struct e1000_adapter *adapter,
                    struct e1000_tx_ring *tx_ring)
L
Linus Torvalds 已提交
2221 2222 2223 2224 2225 2226 2227
{
	struct e1000_buffer *buffer_info;
	unsigned long size;
	unsigned int i;

	/* Free all the Tx ring sk_buffs */

J
Jesse Brandeburg 已提交
2228
	for (i = 0; i < tx_ring->count; i++) {
L
Linus Torvalds 已提交
2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241
		buffer_info = &tx_ring->buffer_info[i];
		e1000_unmap_and_free_tx_resource(adapter, buffer_info);
	}

	size = sizeof(struct e1000_buffer) * tx_ring->count;
	memset(tx_ring->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;
J
Jeff Kirsher 已提交
2242
	tx_ring->last_tx_tso = 0;
L
Linus Torvalds 已提交
2243

2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257
	writel(0, adapter->hw.hw_addr + tx_ring->tdh);
	writel(0, adapter->hw.hw_addr + tx_ring->tdt);
}

/**
 * e1000_clean_all_tx_rings - Free Tx Buffers for all queues
 * @adapter: board private structure
 **/

static void
e1000_clean_all_tx_rings(struct e1000_adapter *adapter)
{
	int i;

2258
	for (i = 0; i < adapter->num_tx_queues; i++)
2259
		e1000_clean_tx_ring(adapter, &adapter->tx_ring[i]);
L
Linus Torvalds 已提交
2260 2261 2262 2263 2264
}

/**
 * e1000_free_rx_resources - Free Rx Resources
 * @adapter: board private structure
2265
 * @rx_ring: ring to clean the resources from
L
Linus Torvalds 已提交
2266 2267 2268 2269
 *
 * Free all receive software resources
 **/

2270
static void
2271 2272
e1000_free_rx_resources(struct e1000_adapter *adapter,
                        struct e1000_rx_ring *rx_ring)
L
Linus Torvalds 已提交
2273 2274 2275
{
	struct pci_dev *pdev = adapter->pdev;

2276
	e1000_clean_rx_ring(adapter, rx_ring);
L
Linus Torvalds 已提交
2277 2278 2279

	vfree(rx_ring->buffer_info);
	rx_ring->buffer_info = NULL;
2280 2281 2282 2283
	kfree(rx_ring->ps_page);
	rx_ring->ps_page = NULL;
	kfree(rx_ring->ps_page_dma);
	rx_ring->ps_page_dma = NULL;
L
Linus Torvalds 已提交
2284 2285 2286 2287 2288 2289 2290

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

	rx_ring->desc = NULL;
}

/**
2291
 * e1000_free_all_rx_resources - Free Rx Resources for All Queues
L
Linus Torvalds 已提交
2292
 * @adapter: board private structure
2293 2294 2295 2296 2297 2298 2299 2300 2301
 *
 * Free all receive software resources
 **/

void
e1000_free_all_rx_resources(struct e1000_adapter *adapter)
{
	int i;

2302
	for (i = 0; i < adapter->num_rx_queues; i++)
2303 2304 2305 2306 2307 2308 2309
		e1000_free_rx_resources(adapter, &adapter->rx_ring[i]);
}

/**
 * e1000_clean_rx_ring - Free Rx Buffers per Queue
 * @adapter: board private structure
 * @rx_ring: ring to free buffers from
L
Linus Torvalds 已提交
2310 2311 2312
 **/

static void
2313 2314
e1000_clean_rx_ring(struct e1000_adapter *adapter,
                    struct e1000_rx_ring *rx_ring)
L
Linus Torvalds 已提交
2315 2316
{
	struct e1000_buffer *buffer_info;
2317 2318
	struct e1000_ps_page *ps_page;
	struct e1000_ps_page_dma *ps_page_dma;
L
Linus Torvalds 已提交
2319 2320
	struct pci_dev *pdev = adapter->pdev;
	unsigned long size;
2321
	unsigned int i, j;
L
Linus Torvalds 已提交
2322 2323

	/* Free all the Rx ring sk_buffs */
J
Jesse Brandeburg 已提交
2324
	for (i = 0; i < rx_ring->count; i++) {
L
Linus Torvalds 已提交
2325
		buffer_info = &rx_ring->buffer_info[i];
J
Jesse Brandeburg 已提交
2326
		if (buffer_info->skb) {
L
Linus Torvalds 已提交
2327 2328 2329 2330 2331 2332 2333
			pci_unmap_single(pdev,
					 buffer_info->dma,
					 buffer_info->length,
					 PCI_DMA_FROMDEVICE);

			dev_kfree_skb(buffer_info->skb);
			buffer_info->skb = NULL;
2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344
		}
		ps_page = &rx_ring->ps_page[i];
		ps_page_dma = &rx_ring->ps_page_dma[i];
		for (j = 0; j < adapter->rx_ps_pages; j++) {
			if (!ps_page->ps_page[j]) break;
			pci_unmap_page(pdev,
				       ps_page_dma->ps_page_dma[j],
				       PAGE_SIZE, PCI_DMA_FROMDEVICE);
			ps_page_dma->ps_page_dma[j] = 0;
			put_page(ps_page->ps_page[j]);
			ps_page->ps_page[j] = NULL;
L
Linus Torvalds 已提交
2345 2346 2347 2348 2349
		}
	}

	size = sizeof(struct e1000_buffer) * rx_ring->count;
	memset(rx_ring->buffer_info, 0, size);
2350 2351 2352 2353
	size = sizeof(struct e1000_ps_page) * rx_ring->count;
	memset(rx_ring->ps_page, 0, size);
	size = sizeof(struct e1000_ps_page_dma) * rx_ring->count;
	memset(rx_ring->ps_page_dma, 0, size);
L
Linus Torvalds 已提交
2354 2355 2356 2357 2358 2359 2360 2361

	/* 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;

2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375
	writel(0, adapter->hw.hw_addr + rx_ring->rdh);
	writel(0, adapter->hw.hw_addr + rx_ring->rdt);
}

/**
 * e1000_clean_all_rx_rings - Free Rx Buffers for all queues
 * @adapter: board private structure
 **/

static void
e1000_clean_all_rx_rings(struct e1000_adapter *adapter)
{
	int i;

2376
	for (i = 0; i < adapter->num_rx_queues; i++)
2377
		e1000_clean_rx_ring(adapter, &adapter->rx_ring[i]);
L
Linus Torvalds 已提交
2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396
}

/* The 82542 2.0 (revision 2) needs to have the receive unit in reset
 * and memory write and invalidate disabled for certain operations
 */
static void
e1000_enter_82542_rst(struct e1000_adapter *adapter)
{
	struct net_device *netdev = adapter->netdev;
	uint32_t rctl;

	e1000_pci_clear_mwi(&adapter->hw);

	rctl = E1000_READ_REG(&adapter->hw, RCTL);
	rctl |= E1000_RCTL_RST;
	E1000_WRITE_REG(&adapter->hw, RCTL, rctl);
	E1000_WRITE_FLUSH(&adapter->hw);
	mdelay(5);

J
Jesse Brandeburg 已提交
2397
	if (netif_running(netdev))
2398
		e1000_clean_all_rx_rings(adapter);
L
Linus Torvalds 已提交
2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412
}

static void
e1000_leave_82542_rst(struct e1000_adapter *adapter)
{
	struct net_device *netdev = adapter->netdev;
	uint32_t rctl;

	rctl = E1000_READ_REG(&adapter->hw, RCTL);
	rctl &= ~E1000_RCTL_RST;
	E1000_WRITE_REG(&adapter->hw, RCTL, rctl);
	E1000_WRITE_FLUSH(&adapter->hw);
	mdelay(5);

J
Jesse Brandeburg 已提交
2413
	if (adapter->hw.pci_cmd_word & PCI_COMMAND_INVALIDATE)
L
Linus Torvalds 已提交
2414 2415
		e1000_pci_set_mwi(&adapter->hw);

J
Jesse Brandeburg 已提交
2416
	if (netif_running(netdev)) {
2417 2418
		/* No need to loop, because 82542 supports only 1 queue */
		struct e1000_rx_ring *ring = &adapter->rx_ring[0];
2419
		e1000_configure_rx(adapter);
2420
		adapter->alloc_rx_buf(adapter, ring, E1000_DESC_UNUSED(ring));
L
Linus Torvalds 已提交
2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434
	}
}

/**
 * e1000_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
e1000_set_mac(struct net_device *netdev, void *p)
{
2435
	struct e1000_adapter *adapter = netdev_priv(netdev);
L
Linus Torvalds 已提交
2436 2437
	struct sockaddr *addr = p;

J
Jesse Brandeburg 已提交
2438
	if (!is_valid_ether_addr(addr->sa_data))
L
Linus Torvalds 已提交
2439 2440 2441 2442
		return -EADDRNOTAVAIL;

	/* 82542 2.0 needs to be in reset to write receive address registers */

J
Jesse Brandeburg 已提交
2443
	if (adapter->hw.mac_type == e1000_82542_rev2_0)
L
Linus Torvalds 已提交
2444 2445 2446 2447 2448 2449 2450
		e1000_enter_82542_rst(adapter);

	memcpy(netdev->dev_addr, addr->sa_data, netdev->addr_len);
	memcpy(adapter->hw.mac_addr, addr->sa_data, netdev->addr_len);

	e1000_rar_set(&adapter->hw, adapter->hw.mac_addr, 0);

2451 2452 2453 2454 2455 2456
	/* With 82571 controllers, LAA may be overwritten (with the default)
	 * due to controller reset from the other port. */
	if (adapter->hw.mac_type == e1000_82571) {
		/* activate the work around */
		adapter->hw.laa_is_present = 1;

J
Jesse Brandeburg 已提交
2457 2458 2459
		/* Hold a copy of the LAA in RAR[14] This is done so that
		 * between the time RAR[0] gets clobbered  and the time it
		 * gets fixed (in e1000_watchdog), the actual LAA is in one
2460
		 * of the RARs and no incoming packets directed to this port
J
Jesse Brandeburg 已提交
2461
		 * are dropped. Eventaully the LAA will be in RAR[0] and
2462
		 * RAR[14] */
J
Jesse Brandeburg 已提交
2463
		e1000_rar_set(&adapter->hw, adapter->hw.mac_addr,
2464 2465 2466
					E1000_RAR_ENTRIES - 1);
	}

J
Jesse Brandeburg 已提交
2467
	if (adapter->hw.mac_type == e1000_82542_rev2_0)
L
Linus Torvalds 已提交
2468 2469 2470 2471 2472 2473
		e1000_leave_82542_rst(adapter);

	return 0;
}

/**
2474
 * e1000_set_rx_mode - Secondary Unicast, Multicast and Promiscuous mode set
L
Linus Torvalds 已提交
2475 2476
 * @netdev: network interface device structure
 *
2477 2478 2479
 * The set_rx_mode entry point is called whenever the unicast or multicast
 * address lists or the network interface flags are updated. This routine is
 * responsible for configuring the hardware for proper unicast, multicast,
L
Linus Torvalds 已提交
2480 2481 2482 2483
 * promiscuous mode, and all-multi behavior.
 **/

static void
2484
e1000_set_rx_mode(struct net_device *netdev)
L
Linus Torvalds 已提交
2485
{
2486
	struct e1000_adapter *adapter = netdev_priv(netdev);
L
Linus Torvalds 已提交
2487
	struct e1000_hw *hw = &adapter->hw;
2488 2489
	struct dev_addr_list *uc_ptr;
	struct dev_addr_list *mc_ptr;
L
Linus Torvalds 已提交
2490 2491
	uint32_t rctl;
	uint32_t hash_value;
2492
	int i, rar_entries = E1000_RAR_ENTRIES;
2493 2494 2495 2496 2497 2498
	int mta_reg_count = (hw->mac_type == e1000_ich8lan) ?
				E1000_NUM_MTA_REGISTERS_ICH8LAN :
				E1000_NUM_MTA_REGISTERS;

	if (adapter->hw.mac_type == e1000_ich8lan)
		rar_entries = E1000_RAR_ENTRIES_ICH8LAN;
L
Linus Torvalds 已提交
2499

2500 2501 2502
	/* reserve RAR[14] for LAA over-write work-around */
	if (adapter->hw.mac_type == e1000_82571)
		rar_entries--;
L
Linus Torvalds 已提交
2503

2504 2505
	/* Check for Promiscuous and All Multicast modes */

L
Linus Torvalds 已提交
2506 2507
	rctl = E1000_READ_REG(hw, RCTL);

J
Jesse Brandeburg 已提交
2508
	if (netdev->flags & IFF_PROMISC) {
L
Linus Torvalds 已提交
2509
		rctl |= (E1000_RCTL_UPE | E1000_RCTL_MPE);
J
Jesse Brandeburg 已提交
2510
	} else if (netdev->flags & IFF_ALLMULTI) {
L
Linus Torvalds 已提交
2511 2512
		rctl |= E1000_RCTL_MPE;
	} else {
2513 2514 2515 2516 2517 2518 2519 2520 2521
		rctl &= ~E1000_RCTL_MPE;
	}

	uc_ptr = NULL;
	if (netdev->uc_count > rar_entries - 1) {
		rctl |= E1000_RCTL_UPE;
	} else if (!(netdev->flags & IFF_PROMISC)) {
		rctl &= ~E1000_RCTL_UPE;
		uc_ptr = netdev->uc_list;
L
Linus Torvalds 已提交
2522 2523 2524 2525 2526 2527
	}

	E1000_WRITE_REG(hw, RCTL, rctl);

	/* 82542 2.0 needs to be in reset to write receive address registers */

J
Jesse Brandeburg 已提交
2528
	if (hw->mac_type == e1000_82542_rev2_0)
L
Linus Torvalds 已提交
2529 2530
		e1000_enter_82542_rst(adapter);

2531 2532 2533 2534
	/* load the first 14 addresses into the exact filters 1-14. Unicast
	 * addresses take precedence to avoid disabling unicast filtering
	 * when possible.
	 *
L
Linus Torvalds 已提交
2535 2536
	 * RAR 0 is used for the station MAC adddress
	 * if there are not 14 addresses, go ahead and clear the filters
2537
	 * -- with 82571 controllers only 0-13 entries are filled here
L
Linus Torvalds 已提交
2538 2539 2540
	 */
	mc_ptr = netdev->mc_list;

J
Jesse Brandeburg 已提交
2541
	for (i = 1; i < rar_entries; i++) {
2542 2543 2544 2545 2546
		if (uc_ptr) {
			e1000_rar_set(hw, uc_ptr->da_addr, i);
			uc_ptr = uc_ptr->next;
		} else if (mc_ptr) {
			e1000_rar_set(hw, mc_ptr->da_addr, i);
L
Linus Torvalds 已提交
2547 2548 2549
			mc_ptr = mc_ptr->next;
		} else {
			E1000_WRITE_REG_ARRAY(hw, RA, i << 1, 0);
2550
			E1000_WRITE_FLUSH(hw);
L
Linus Torvalds 已提交
2551
			E1000_WRITE_REG_ARRAY(hw, RA, (i << 1) + 1, 0);
2552
			E1000_WRITE_FLUSH(hw);
L
Linus Torvalds 已提交
2553 2554
		}
	}
2555
	WARN_ON(uc_ptr != NULL);
L
Linus Torvalds 已提交
2556 2557 2558

	/* clear the old settings from the multicast hash table */

2559
	for (i = 0; i < mta_reg_count; i++) {
L
Linus Torvalds 已提交
2560
		E1000_WRITE_REG_ARRAY(hw, MTA, i, 0);
2561 2562
		E1000_WRITE_FLUSH(hw);
	}
L
Linus Torvalds 已提交
2563 2564 2565

	/* load any remaining addresses into the hash table */

J
Jesse Brandeburg 已提交
2566
	for (; mc_ptr; mc_ptr = mc_ptr->next) {
2567
		hash_value = e1000_hash_mc_addr(hw, mc_ptr->da_addr);
L
Linus Torvalds 已提交
2568 2569 2570
		e1000_mta_set(hw, hash_value);
	}

J
Jesse Brandeburg 已提交
2571
	if (hw->mac_type == e1000_82542_rev2_0)
L
Linus Torvalds 已提交
2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596
		e1000_leave_82542_rst(adapter);
}

/* Need to wait a few seconds after link up to get diagnostic information from
 * the phy */

static void
e1000_update_phy_info(unsigned long data)
{
	struct e1000_adapter *adapter = (struct e1000_adapter *) data;
	e1000_phy_get_info(&adapter->hw, &adapter->phy_info);
}

/**
 * e1000_82547_tx_fifo_stall - Timer Call-back
 * @data: pointer to adapter cast into an unsigned long
 **/

static void
e1000_82547_tx_fifo_stall(unsigned long data)
{
	struct e1000_adapter *adapter = (struct e1000_adapter *) data;
	struct net_device *netdev = adapter->netdev;
	uint32_t tctl;

J
Jesse Brandeburg 已提交
2597 2598
	if (atomic_read(&adapter->tx_fifo_stall)) {
		if ((E1000_READ_REG(&adapter->hw, TDT) ==
L
Linus Torvalds 已提交
2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635
		    E1000_READ_REG(&adapter->hw, TDH)) &&
		   (E1000_READ_REG(&adapter->hw, TDFT) ==
		    E1000_READ_REG(&adapter->hw, TDFH)) &&
		   (E1000_READ_REG(&adapter->hw, TDFTS) ==
		    E1000_READ_REG(&adapter->hw, TDFHS))) {
			tctl = E1000_READ_REG(&adapter->hw, TCTL);
			E1000_WRITE_REG(&adapter->hw, TCTL,
					tctl & ~E1000_TCTL_EN);
			E1000_WRITE_REG(&adapter->hw, TDFT,
					adapter->tx_head_addr);
			E1000_WRITE_REG(&adapter->hw, TDFH,
					adapter->tx_head_addr);
			E1000_WRITE_REG(&adapter->hw, TDFTS,
					adapter->tx_head_addr);
			E1000_WRITE_REG(&adapter->hw, TDFHS,
					adapter->tx_head_addr);
			E1000_WRITE_REG(&adapter->hw, TCTL, tctl);
			E1000_WRITE_FLUSH(&adapter->hw);

			adapter->tx_fifo_head = 0;
			atomic_set(&adapter->tx_fifo_stall, 0);
			netif_wake_queue(netdev);
		} else {
			mod_timer(&adapter->tx_fifo_stall_timer, jiffies + 1);
		}
	}
}

/**
 * e1000_watchdog - Timer Call-back
 * @data: pointer to adapter cast into an unsigned long
 **/
static void
e1000_watchdog(unsigned long data)
{
	struct e1000_adapter *adapter = (struct e1000_adapter *) data;
	struct net_device *netdev = adapter->netdev;
J
Jeff Kirsher 已提交
2636
	struct e1000_tx_ring *txdr = adapter->tx_ring;
2637
	uint32_t link, tctl;
2638 2639 2640 2641 2642 2643 2644 2645 2646 2647
	int32_t ret_val;

	ret_val = e1000_check_for_link(&adapter->hw);
	if ((ret_val == E1000_ERR_PHY) &&
	    (adapter->hw.phy_type == e1000_phy_igp_3) &&
	    (E1000_READ_REG(&adapter->hw, CTRL) & E1000_PHY_CTRL_GBE_DISABLE)) {
		/* See e1000_kumeran_lock_loss_workaround() */
		DPRINTK(LINK, INFO,
			"Gigabit has been disabled, downgrading speed\n");
	}
2648

2649 2650
	if (adapter->hw.mac_type == e1000_82573) {
		e1000_enable_tx_pkt_filtering(&adapter->hw);
J
Jesse Brandeburg 已提交
2651
		if (adapter->mng_vlan_id != adapter->hw.mng_cookie.vlan_id)
2652
			e1000_update_mng_vlan(adapter);
J
Jesse Brandeburg 已提交
2653
	}
L
Linus Torvalds 已提交
2654

J
Jesse Brandeburg 已提交
2655
	if ((adapter->hw.media_type == e1000_media_type_internal_serdes) &&
L
Linus Torvalds 已提交
2656 2657 2658 2659 2660
	   !(E1000_READ_REG(&adapter->hw, TXCW) & E1000_TXCW_ANE))
		link = !adapter->hw.serdes_link_down;
	else
		link = E1000_READ_REG(&adapter->hw, STATUS) & E1000_STATUS_LU;

J
Jesse Brandeburg 已提交
2661 2662
	if (link) {
		if (!netif_carrier_ok(netdev)) {
2663
			uint32_t ctrl;
2664
			boolean_t txb2b = 1;
L
Linus Torvalds 已提交
2665 2666 2667 2668
			e1000_get_speed_and_duplex(&adapter->hw,
			                           &adapter->link_speed,
			                           &adapter->link_duplex);

2669 2670 2671 2672 2673 2674 2675 2676 2677 2678
			ctrl = E1000_READ_REG(&adapter->hw, CTRL);
			DPRINTK(LINK, INFO, "NIC Link is Up %d Mbps %s, "
			        "Flow Control: %s\n",
			        adapter->link_speed,
			        adapter->link_duplex == FULL_DUPLEX ?
			        "Full Duplex" : "Half Duplex",
			        ((ctrl & E1000_CTRL_TFCE) && (ctrl &
			        E1000_CTRL_RFCE)) ? "RX/TX" : ((ctrl &
			        E1000_CTRL_RFCE) ? "RX" : ((ctrl &
			        E1000_CTRL_TFCE) ? "TX" : "None" )));
L
Linus Torvalds 已提交
2679

2680 2681
			/* tweak tx_queue_len according to speed/duplex
			 * and adjust the timeout factor */
2682 2683
			netdev->tx_queue_len = adapter->tx_queue_len;
			adapter->tx_timeout_factor = 1;
2684 2685
			switch (adapter->link_speed) {
			case SPEED_10:
2686
				txb2b = 0;
2687 2688 2689 2690
				netdev->tx_queue_len = 10;
				adapter->tx_timeout_factor = 8;
				break;
			case SPEED_100:
2691
				txb2b = 0;
2692 2693 2694 2695 2696
				netdev->tx_queue_len = 100;
				/* maybe add some timeout factor ? */
				break;
			}

2697
			if ((adapter->hw.mac_type == e1000_82571 ||
2698
			     adapter->hw.mac_type == e1000_82572) &&
2699
			    txb2b == 0) {
2700 2701
				uint32_t tarc0;
				tarc0 = E1000_READ_REG(&adapter->hw, TARC0);
2702
				tarc0 &= ~(1 << 21);
2703 2704
				E1000_WRITE_REG(&adapter->hw, TARC0, tarc0);
			}
2705

2706 2707 2708 2709
			/* disable TSO for pcie and 10/100 speeds, to avoid
			 * some hardware issues */
			if (!adapter->tso_force &&
			    adapter->hw.bus_type == e1000_bus_type_pci_express){
2710 2711 2712
				switch (adapter->link_speed) {
				case SPEED_10:
				case SPEED_100:
2713 2714 2715
					DPRINTK(PROBE,INFO,
				        "10/100 speed: disabling TSO\n");
					netdev->features &= ~NETIF_F_TSO;
A
Auke Kok 已提交
2716
					netdev->features &= ~NETIF_F_TSO6;
2717 2718 2719
					break;
				case SPEED_1000:
					netdev->features |= NETIF_F_TSO;
A
Auke Kok 已提交
2720
					netdev->features |= NETIF_F_TSO6;
2721 2722 2723
					break;
				default:
					/* oops */
2724 2725 2726
					break;
				}
			}
2727 2728 2729 2730 2731 2732

			/* enable transmits in the hardware, need to do this
			 * after setting TARC0 */
			tctl = E1000_READ_REG(&adapter->hw, TCTL);
			tctl |= E1000_TCTL_EN;
			E1000_WRITE_REG(&adapter->hw, TCTL, tctl);
2733

L
Linus Torvalds 已提交
2734 2735
			netif_carrier_on(netdev);
			netif_wake_queue(netdev);
2736
			mod_timer(&adapter->phy_info_timer, round_jiffies(jiffies + 2 * HZ));
L
Linus Torvalds 已提交
2737
			adapter->smartspeed = 0;
2738 2739 2740 2741 2742 2743 2744
		} else {
			/* make sure the receive unit is started */
			if (adapter->hw.rx_needs_kicking) {
				struct e1000_hw *hw = &adapter->hw;
				uint32_t rctl = E1000_READ_REG(hw, RCTL);
				E1000_WRITE_REG(hw, RCTL, rctl | E1000_RCTL_EN);
			}
L
Linus Torvalds 已提交
2745 2746
		}
	} else {
J
Jesse Brandeburg 已提交
2747
		if (netif_carrier_ok(netdev)) {
L
Linus Torvalds 已提交
2748 2749 2750 2751 2752
			adapter->link_speed = 0;
			adapter->link_duplex = 0;
			DPRINTK(LINK, INFO, "NIC Link is Down\n");
			netif_carrier_off(netdev);
			netif_stop_queue(netdev);
2753
			mod_timer(&adapter->phy_info_timer, round_jiffies(jiffies + 2 * HZ));
2754 2755 2756 2757 2758 2759

			/* 80003ES2LAN workaround--
			 * For packet buffer work-around on link down event;
			 * disable receives in the ISR and
			 * reset device here in the watchdog
			 */
2760
			if (adapter->hw.mac_type == e1000_80003es2lan)
2761 2762
				/* reset device */
				schedule_work(&adapter->reset_task);
L
Linus Torvalds 已提交
2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781
		}

		e1000_smartspeed(adapter);
	}

	e1000_update_stats(adapter);

	adapter->hw.tx_packet_delta = adapter->stats.tpt - adapter->tpt_old;
	adapter->tpt_old = adapter->stats.tpt;
	adapter->hw.collision_delta = adapter->stats.colc - adapter->colc_old;
	adapter->colc_old = adapter->stats.colc;

	adapter->gorcl = adapter->stats.gorcl - adapter->gorcl_old;
	adapter->gorcl_old = adapter->stats.gorcl;
	adapter->gotcl = adapter->stats.gotcl - adapter->gotcl_old;
	adapter->gotcl_old = adapter->stats.gotcl;

	e1000_update_adaptive(&adapter->hw);

2782
	if (!netif_carrier_ok(netdev)) {
2783
		if (E1000_DESC_UNUSED(txdr) + 1 < txdr->count) {
L
Linus Torvalds 已提交
2784 2785 2786 2787
			/* 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). */
2788 2789
			adapter->tx_timeout_count++;
			schedule_work(&adapter->reset_task);
L
Linus Torvalds 已提交
2790 2791 2792 2793 2794 2795
		}
	}

	/* Cause software interrupt to ensure rx ring is cleaned */
	E1000_WRITE_REG(&adapter->hw, ICS, E1000_ICS_RXDMT0);

2796
	/* Force detection of hung controller every watchdog period */
L
Linus Torvalds 已提交
2797 2798
	adapter->detect_tx_hung = TRUE;

J
Jesse Brandeburg 已提交
2799
	/* With 82571 controllers, LAA may be overwritten due to controller
2800 2801 2802 2803
	 * reset from the other port. Set the appropriate LAA in RAR[0] */
	if (adapter->hw.mac_type == e1000_82571 && adapter->hw.laa_is_present)
		e1000_rar_set(&adapter->hw, adapter->hw.mac_addr, 0);

L
Linus Torvalds 已提交
2804
	/* Reset the timer */
2805
	mod_timer(&adapter->watchdog_timer, round_jiffies(jiffies + 2 * HZ));
L
Linus Torvalds 已提交
2806 2807
}

J
Jesse Brandeburg 已提交
2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846
enum latency_range {
	lowest_latency = 0,
	low_latency = 1,
	bulk_latency = 2,
	latency_invalid = 255
};

/**
 * e1000_update_itr - update the dynamic ITR value based on statistics
 *      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 e1000_param.c)
 * @adapter: pointer to adapter
 * @itr_setting: current adapter->itr
 * @packets: the number of packets during this measurement interval
 * @bytes: the number of bytes during this measurement interval
 **/
static unsigned int e1000_update_itr(struct e1000_adapter *adapter,
                                   uint16_t itr_setting,
                                   int packets,
                                   int bytes)
{
	unsigned int retval = itr_setting;
	struct e1000_hw *hw = &adapter->hw;

	if (unlikely(hw->mac_type < e1000_82540))
		goto update_itr_done;

	if (packets == 0)
		goto update_itr_done;

	switch (itr_setting) {
	case lowest_latency:
2847 2848 2849 2850
		/* jumbo frames get bulk treatment*/
		if (bytes/packets > 8000)
			retval = bulk_latency;
		else if ((packets < 5) && (bytes > 512))
J
Jesse Brandeburg 已提交
2851 2852 2853 2854
			retval = low_latency;
		break;
	case low_latency:  /* 50 usec aka 20000 ints/s */
		if (bytes > 10000) {
2855 2856 2857 2858
			/* jumbo frames need bulk latency setting */
			if (bytes/packets > 8000)
				retval = bulk_latency;
			else if ((packets < 10) || ((bytes/packets) > 1200))
J
Jesse Brandeburg 已提交
2859 2860 2861
				retval = bulk_latency;
			else if ((packets > 35))
				retval = lowest_latency;
2862 2863 2864
		} else if (bytes/packets > 2000)
			retval = bulk_latency;
		else if (packets <= 2 && bytes < 512)
J
Jesse Brandeburg 已提交
2865 2866 2867 2868 2869 2870
			retval = lowest_latency;
		break;
	case bulk_latency: /* 250 usec aka 4000 ints/s */
		if (bytes > 25000) {
			if (packets > 35)
				retval = low_latency;
2871 2872
		} else if (bytes < 6000) {
			retval = low_latency;
J
Jesse Brandeburg 已提交
2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900
		}
		break;
	}

update_itr_done:
	return retval;
}

static void e1000_set_itr(struct e1000_adapter *adapter)
{
	struct e1000_hw *hw = &adapter->hw;
	uint16_t current_itr;
	uint32_t new_itr = adapter->itr;

	if (unlikely(hw->mac_type < e1000_82540))
		return;

	/* for non-gigabit speeds, just fix the interrupt rate at 4000 */
	if (unlikely(adapter->link_speed != SPEED_1000)) {
		current_itr = 0;
		new_itr = 4000;
		goto set_itr_now;
	}

	adapter->tx_itr = e1000_update_itr(adapter,
	                            adapter->tx_itr,
	                            adapter->total_tx_packets,
	                            adapter->total_tx_bytes);
2901 2902 2903 2904
	/* conservative mode (itr 3) eliminates the lowest_latency setting */
	if (adapter->itr_setting == 3 && adapter->tx_itr == lowest_latency)
		adapter->tx_itr = low_latency;

J
Jesse Brandeburg 已提交
2905 2906 2907 2908
	adapter->rx_itr = e1000_update_itr(adapter,
	                            adapter->rx_itr,
	                            adapter->total_rx_packets,
	                            adapter->total_rx_bytes);
2909 2910 2911
	/* conservative mode (itr 3) eliminates the lowest_latency setting */
	if (adapter->itr_setting == 3 && adapter->rx_itr == lowest_latency)
		adapter->rx_itr = low_latency;
J
Jesse Brandeburg 已提交
2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944

	current_itr = max(adapter->rx_itr, adapter->tx_itr);

	switch (current_itr) {
	/* counts and packets in update_itr are dependent on these numbers */
	case lowest_latency:
		new_itr = 70000;
		break;
	case low_latency:
		new_itr = 20000; /* aka hwitr = ~200 */
		break;
	case bulk_latency:
		new_itr = 4000;
		break;
	default:
		break;
	}

set_itr_now:
	if (new_itr != adapter->itr) {
		/* this attempts to bias the interrupt rate towards Bulk
		 * by adding intermediate steps when interrupt rate is
		 * increasing */
		new_itr = new_itr > adapter->itr ?
		             min(adapter->itr + (new_itr >> 2), new_itr) :
		             new_itr;
		adapter->itr = new_itr;
		E1000_WRITE_REG(hw, ITR, 1000000000 / (new_itr * 256));
	}

	return;
}

L
Linus Torvalds 已提交
2945 2946 2947
#define E1000_TX_FLAGS_CSUM		0x00000001
#define E1000_TX_FLAGS_VLAN		0x00000002
#define E1000_TX_FLAGS_TSO		0x00000004
2948
#define E1000_TX_FLAGS_IPV4		0x00000008
L
Linus Torvalds 已提交
2949 2950 2951
#define E1000_TX_FLAGS_VLAN_MASK	0xffff0000
#define E1000_TX_FLAGS_VLAN_SHIFT	16

2952
static int
2953 2954
e1000_tso(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring,
          struct sk_buff *skb)
L
Linus Torvalds 已提交
2955 2956
{
	struct e1000_context_desc *context_desc;
J
Jeff Kirsher 已提交
2957
	struct e1000_buffer *buffer_info;
L
Linus Torvalds 已提交
2958 2959
	unsigned int i;
	uint32_t cmd_length = 0;
2960
	uint16_t ipcse = 0, tucse, mss;
L
Linus Torvalds 已提交
2961 2962 2963
	uint8_t ipcss, ipcso, tucss, tucso, hdr_len;
	int err;

H
Herbert Xu 已提交
2964
	if (skb_is_gso(skb)) {
L
Linus Torvalds 已提交
2965 2966 2967 2968 2969 2970
		if (skb_header_cloned(skb)) {
			err = pskb_expand_head(skb, 0, 0, GFP_ATOMIC);
			if (err)
				return err;
		}

2971
		hdr_len = skb_transport_offset(skb) + tcp_hdrlen(skb);
2972
		mss = skb_shinfo(skb)->gso_size;
A
Alexey Dobriyan 已提交
2973
		if (skb->protocol == htons(ETH_P_IP)) {
2974 2975 2976
			struct iphdr *iph = ip_hdr(skb);
			iph->tot_len = 0;
			iph->check = 0;
2977 2978 2979 2980
			tcp_hdr(skb)->check = ~csum_tcpudp_magic(iph->saddr,
								 iph->daddr, 0,
								 IPPROTO_TCP,
								 0);
2981
			cmd_length = E1000_TXD_CMD_IP;
2982
			ipcse = skb_transport_offset(skb) - 1;
2983
		} else if (skb->protocol == htons(ETH_P_IPV6)) {
2984
			ipv6_hdr(skb)->payload_len = 0;
2985
			tcp_hdr(skb)->check =
2986 2987 2988
				~csum_ipv6_magic(&ipv6_hdr(skb)->saddr,
						 &ipv6_hdr(skb)->daddr,
						 0, IPPROTO_TCP, 0);
2989 2990
			ipcse = 0;
		}
2991
		ipcss = skb_network_offset(skb);
2992
		ipcso = (void *)&(ip_hdr(skb)->check) - (void *)skb->data;
2993
		tucss = skb_transport_offset(skb);
2994
		tucso = (void *)&(tcp_hdr(skb)->check) - (void *)skb->data;
L
Linus Torvalds 已提交
2995 2996 2997
		tucse = 0;

		cmd_length |= (E1000_TXD_CMD_DEXT | E1000_TXD_CMD_TSE |
2998
			       E1000_TXD_CMD_TCP | (skb->len - (hdr_len)));
L
Linus Torvalds 已提交
2999

3000 3001
		i = tx_ring->next_to_use;
		context_desc = E1000_CONTEXT_DESC(*tx_ring, i);
J
Jeff Kirsher 已提交
3002
		buffer_info = &tx_ring->buffer_info[i];
L
Linus Torvalds 已提交
3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013

		context_desc->lower_setup.ip_fields.ipcss  = ipcss;
		context_desc->lower_setup.ip_fields.ipcso  = ipcso;
		context_desc->lower_setup.ip_fields.ipcse  = cpu_to_le16(ipcse);
		context_desc->upper_setup.tcp_fields.tucss = tucss;
		context_desc->upper_setup.tcp_fields.tucso = tucso;
		context_desc->upper_setup.tcp_fields.tucse = cpu_to_le16(tucse);
		context_desc->tcp_seg_setup.fields.mss     = cpu_to_le16(mss);
		context_desc->tcp_seg_setup.fields.hdr_len = hdr_len;
		context_desc->cmd_and_length = cpu_to_le32(cmd_length);

J
Jeff Kirsher 已提交
3014
		buffer_info->time_stamp = jiffies;
3015
		buffer_info->next_to_watch = i;
J
Jeff Kirsher 已提交
3016

3017 3018
		if (++i == tx_ring->count) i = 0;
		tx_ring->next_to_use = i;
L
Linus Torvalds 已提交
3019

3020
		return TRUE;
L
Linus Torvalds 已提交
3021
	}
3022
	return FALSE;
L
Linus Torvalds 已提交
3023 3024
}

3025
static boolean_t
3026 3027
e1000_tx_csum(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring,
              struct sk_buff *skb)
L
Linus Torvalds 已提交
3028 3029
{
	struct e1000_context_desc *context_desc;
J
Jeff Kirsher 已提交
3030
	struct e1000_buffer *buffer_info;
L
Linus Torvalds 已提交
3031 3032 3033
	unsigned int i;
	uint8_t css;

3034
	if (likely(skb->ip_summed == CHECKSUM_PARTIAL)) {
3035
		css = skb_transport_offset(skb);
L
Linus Torvalds 已提交
3036

3037
		i = tx_ring->next_to_use;
J
Jeff Kirsher 已提交
3038
		buffer_info = &tx_ring->buffer_info[i];
3039
		context_desc = E1000_CONTEXT_DESC(*tx_ring, i);
L
Linus Torvalds 已提交
3040

3041
		context_desc->lower_setup.ip_config = 0;
L
Linus Torvalds 已提交
3042
		context_desc->upper_setup.tcp_fields.tucss = css;
3043 3044
		context_desc->upper_setup.tcp_fields.tucso =
			css + skb->csum_offset;
L
Linus Torvalds 已提交
3045 3046 3047 3048
		context_desc->upper_setup.tcp_fields.tucse = 0;
		context_desc->tcp_seg_setup.data = 0;
		context_desc->cmd_and_length = cpu_to_le32(E1000_TXD_CMD_DEXT);

J
Jeff Kirsher 已提交
3049
		buffer_info->time_stamp = jiffies;
3050
		buffer_info->next_to_watch = i;
J
Jeff Kirsher 已提交
3051

3052 3053
		if (unlikely(++i == tx_ring->count)) i = 0;
		tx_ring->next_to_use = i;
L
Linus Torvalds 已提交
3054 3055 3056 3057 3058 3059 3060 3061 3062 3063

		return TRUE;
	}

	return FALSE;
}

#define E1000_MAX_TXD_PWR	12
#define E1000_MAX_DATA_PER_TXD	(1<<E1000_MAX_TXD_PWR)

3064
static int
3065 3066 3067
e1000_tx_map(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring,
             struct sk_buff *skb, unsigned int first, unsigned int max_per_txd,
             unsigned int nr_frags, unsigned int mss)
L
Linus Torvalds 已提交
3068 3069 3070 3071 3072 3073 3074 3075 3076
{
	struct e1000_buffer *buffer_info;
	unsigned int len = skb->len;
	unsigned int offset = 0, size, count = 0, i;
	unsigned int f;
	len -= skb->data_len;

	i = tx_ring->next_to_use;

J
Jesse Brandeburg 已提交
3077
	while (len) {
L
Linus Torvalds 已提交
3078 3079
		buffer_info = &tx_ring->buffer_info[i];
		size = min(len, max_per_txd);
J
Jeff Kirsher 已提交
3080 3081 3082
		/* Workaround for Controller erratum --
		 * descriptor for non-tso packet in a linear SKB that follows a
		 * tso gets written back prematurely before the data is fully
3083
		 * DMA'd to the controller */
J
Jeff Kirsher 已提交
3084
		if (!skb->data_len && tx_ring->last_tx_tso &&
H
Herbert Xu 已提交
3085
		    !skb_is_gso(skb)) {
J
Jeff Kirsher 已提交
3086 3087 3088 3089
			tx_ring->last_tx_tso = 0;
			size -= 4;
		}

L
Linus Torvalds 已提交
3090 3091
		/* Workaround for premature desc write-backs
		 * in TSO mode.  Append 4-byte sentinel desc */
J
Jesse Brandeburg 已提交
3092
		if (unlikely(mss && !nr_frags && size == len && size > 8))
L
Linus Torvalds 已提交
3093
			size -= 4;
3094 3095 3096 3097 3098
		/* work-around for errata 10 and it applies
		 * to all controllers in PCI-X mode
		 * The fix is to make sure that the first descriptor of a
		 * packet is smaller than 2048 - 16 - 16 (or 2016) bytes
		 */
J
Jesse Brandeburg 已提交
3099
		if (unlikely((adapter->hw.bus_type == e1000_bus_type_pcix) &&
3100 3101
		                (size > 2015) && count == 0))
		        size = 2015;
J
Jesse Brandeburg 已提交
3102

L
Linus Torvalds 已提交
3103 3104
		/* Workaround for potential 82544 hang in PCI-X.  Avoid
		 * terminating buffers within evenly-aligned dwords. */
J
Jesse Brandeburg 已提交
3105
		if (unlikely(adapter->pcix_82544 &&
L
Linus Torvalds 已提交
3106 3107 3108 3109 3110 3111 3112 3113 3114 3115 3116
		   !((unsigned long)(skb->data + offset + size - 1) & 4) &&
		   size > 4))
			size -= 4;

		buffer_info->length = size;
		buffer_info->dma =
			pci_map_single(adapter->pdev,
				skb->data + offset,
				size,
				PCI_DMA_TODEVICE);
		buffer_info->time_stamp = jiffies;
3117
		buffer_info->next_to_watch = i;
L
Linus Torvalds 已提交
3118 3119 3120 3121

		len -= size;
		offset += size;
		count++;
J
Jesse Brandeburg 已提交
3122
		if (unlikely(++i == tx_ring->count)) i = 0;
L
Linus Torvalds 已提交
3123 3124
	}

J
Jesse Brandeburg 已提交
3125
	for (f = 0; f < nr_frags; f++) {
L
Linus Torvalds 已提交
3126 3127 3128 3129 3130 3131
		struct skb_frag_struct *frag;

		frag = &skb_shinfo(skb)->frags[f];
		len = frag->size;
		offset = frag->page_offset;

J
Jesse Brandeburg 已提交
3132
		while (len) {
L
Linus Torvalds 已提交
3133 3134 3135 3136
			buffer_info = &tx_ring->buffer_info[i];
			size = min(len, max_per_txd);
			/* Workaround for premature desc write-backs
			 * in TSO mode.  Append 4-byte sentinel desc */
J
Jesse Brandeburg 已提交
3137
			if (unlikely(mss && f == (nr_frags-1) && size == len && size > 8))
L
Linus Torvalds 已提交
3138 3139 3140 3141
				size -= 4;
			/* Workaround for potential 82544 hang in PCI-X.
			 * Avoid terminating buffers within evenly-aligned
			 * dwords. */
J
Jesse Brandeburg 已提交
3142
			if (unlikely(adapter->pcix_82544 &&
L
Linus Torvalds 已提交
3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154
			   !((unsigned long)(frag->page+offset+size-1) & 4) &&
			   size > 4))
				size -= 4;

			buffer_info->length = size;
			buffer_info->dma =
				pci_map_page(adapter->pdev,
					frag->page,
					offset,
					size,
					PCI_DMA_TODEVICE);
			buffer_info->time_stamp = jiffies;
3155
			buffer_info->next_to_watch = i;
L
Linus Torvalds 已提交
3156 3157 3158 3159

			len -= size;
			offset += size;
			count++;
J
Jesse Brandeburg 已提交
3160
			if (unlikely(++i == tx_ring->count)) i = 0;
L
Linus Torvalds 已提交
3161 3162 3163 3164 3165 3166 3167 3168 3169 3170
		}
	}

	i = (i == 0) ? tx_ring->count - 1 : i - 1;
	tx_ring->buffer_info[i].skb = skb;
	tx_ring->buffer_info[first].next_to_watch = i;

	return count;
}

3171
static void
3172 3173
e1000_tx_queue(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring,
               int tx_flags, int count)
L
Linus Torvalds 已提交
3174 3175 3176 3177 3178 3179
{
	struct e1000_tx_desc *tx_desc = NULL;
	struct e1000_buffer *buffer_info;
	uint32_t txd_upper = 0, txd_lower = E1000_TXD_CMD_IFCS;
	unsigned int i;

J
Jesse Brandeburg 已提交
3180
	if (likely(tx_flags & E1000_TX_FLAGS_TSO)) {
L
Linus Torvalds 已提交
3181 3182
		txd_lower |= E1000_TXD_CMD_DEXT | E1000_TXD_DTYP_D |
		             E1000_TXD_CMD_TSE;
3183 3184
		txd_upper |= E1000_TXD_POPTS_TXSM << 8;

J
Jesse Brandeburg 已提交
3185
		if (likely(tx_flags & E1000_TX_FLAGS_IPV4))
3186
			txd_upper |= E1000_TXD_POPTS_IXSM << 8;
L
Linus Torvalds 已提交
3187 3188
	}

J
Jesse Brandeburg 已提交
3189
	if (likely(tx_flags & E1000_TX_FLAGS_CSUM)) {
L
Linus Torvalds 已提交
3190 3191 3192 3193
		txd_lower |= E1000_TXD_CMD_DEXT | E1000_TXD_DTYP_D;
		txd_upper |= E1000_TXD_POPTS_TXSM << 8;
	}

J
Jesse Brandeburg 已提交
3194
	if (unlikely(tx_flags & E1000_TX_FLAGS_VLAN)) {
L
Linus Torvalds 已提交
3195 3196 3197 3198 3199 3200
		txd_lower |= E1000_TXD_CMD_VLE;
		txd_upper |= (tx_flags & E1000_TX_FLAGS_VLAN_MASK);
	}

	i = tx_ring->next_to_use;

J
Jesse Brandeburg 已提交
3201
	while (count--) {
L
Linus Torvalds 已提交
3202 3203 3204 3205 3206 3207
		buffer_info = &tx_ring->buffer_info[i];
		tx_desc = E1000_TX_DESC(*tx_ring, i);
		tx_desc->buffer_addr = cpu_to_le64(buffer_info->dma);
		tx_desc->lower.data =
			cpu_to_le32(txd_lower | buffer_info->length);
		tx_desc->upper.data = cpu_to_le32(txd_upper);
J
Jesse Brandeburg 已提交
3208
		if (unlikely(++i == tx_ring->count)) i = 0;
L
Linus Torvalds 已提交
3209 3210 3211 3212 3213 3214 3215 3216 3217 3218 3219
	}

	tx_desc->lower.data |= cpu_to_le32(adapter->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;
3220
	writel(i, adapter->hw.hw_addr + tx_ring->tdt);
3221 3222 3223
	/* we need this if more than one processor can write to our tail
	 * at a time, it syncronizes IO on IA64/Altix systems */
	mmiowb();
L
Linus Torvalds 已提交
3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234 3235 3236 3237
}

/**
 * 82547 workaround to avoid controller hang in half-duplex environment.
 * The workaround is to avoid queuing a large packet that would span
 * the internal Tx FIFO ring boundary by notifying the stack to resend
 * the packet at a later time.  This gives the Tx FIFO an opportunity to
 * flush all packets.  When that occurs, we reset the Tx FIFO pointers
 * to the beginning of the Tx FIFO.
 **/

#define E1000_FIFO_HDR			0x10
#define E1000_82547_PAD_LEN		0x3E0

3238
static int
L
Linus Torvalds 已提交
3239 3240 3241 3242 3243
e1000_82547_fifo_workaround(struct e1000_adapter *adapter, struct sk_buff *skb)
{
	uint32_t fifo_space = adapter->tx_fifo_size - adapter->tx_fifo_head;
	uint32_t skb_fifo_len = skb->len + E1000_FIFO_HDR;

3244
	skb_fifo_len = ALIGN(skb_fifo_len, E1000_FIFO_HDR);
L
Linus Torvalds 已提交
3245

J
Jesse Brandeburg 已提交
3246
	if (adapter->link_duplex != HALF_DUPLEX)
L
Linus Torvalds 已提交
3247 3248
		goto no_fifo_stall_required;

J
Jesse Brandeburg 已提交
3249
	if (atomic_read(&adapter->tx_fifo_stall))
L
Linus Torvalds 已提交
3250 3251
		return 1;

J
Jesse Brandeburg 已提交
3252
	if (skb_fifo_len >= (E1000_82547_PAD_LEN + fifo_space)) {
L
Linus Torvalds 已提交
3253 3254 3255 3256 3257 3258
		atomic_set(&adapter->tx_fifo_stall, 1);
		return 1;
	}

no_fifo_stall_required:
	adapter->tx_fifo_head += skb_fifo_len;
J
Jesse Brandeburg 已提交
3259
	if (adapter->tx_fifo_head >= adapter->tx_fifo_size)
L
Linus Torvalds 已提交
3260 3261 3262 3263
		adapter->tx_fifo_head -= adapter->tx_fifo_size;
	return 0;
}

3264
#define MINIMUM_DHCP_PACKET_SIZE 282
3265
static int
3266 3267 3268 3269
e1000_transfer_dhcp_info(struct e1000_adapter *adapter, struct sk_buff *skb)
{
	struct e1000_hw *hw =  &adapter->hw;
	uint16_t length, offset;
J
Jesse Brandeburg 已提交
3270 3271
	if (vlan_tx_tag_present(skb)) {
		if (!((vlan_tx_tag_get(skb) == adapter->hw.mng_cookie.vlan_id) &&
3272 3273 3274 3275
			( adapter->hw.mng_cookie.status &
			  E1000_MNG_DHCP_COOKIE_STATUS_VLAN_SUPPORT)) )
			return 0;
	}
3276
	if (skb->len > MINIMUM_DHCP_PACKET_SIZE) {
3277
		struct ethhdr *eth = (struct ethhdr *) skb->data;
J
Jesse Brandeburg 已提交
3278 3279
		if ((htons(ETH_P_IP) == eth->h_proto)) {
			const struct iphdr *ip =
3280
				(struct iphdr *)((uint8_t *)skb->data+14);
J
Jesse Brandeburg 已提交
3281 3282 3283
			if (IPPROTO_UDP == ip->protocol) {
				struct udphdr *udp =
					(struct udphdr *)((uint8_t *)ip +
3284
						(ip->ihl << 2));
J
Jesse Brandeburg 已提交
3285
				if (ntohs(udp->dest) == 67) {
3286 3287 3288 3289
					offset = (uint8_t *)udp + 8 - skb->data;
					length = skb->len - offset;

					return e1000_mng_write_dhcp_info(hw,
J
Jesse Brandeburg 已提交
3290
							(uint8_t *)udp + 8,
3291 3292 3293 3294 3295 3296 3297 3298
							length);
				}
			}
		}
	}
	return 0;
}

3299 3300 3301 3302 3303 3304 3305 3306 3307 3308 3309 3310 3311 3312 3313 3314 3315 3316
static int __e1000_maybe_stop_tx(struct net_device *netdev, int size)
{
	struct e1000_adapter *adapter = netdev_priv(netdev);
	struct e1000_tx_ring *tx_ring = adapter->tx_ring;

	netif_stop_queue(netdev);
	/* 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(E1000_DESC_UNUSED(tx_ring) < size))
		return -EBUSY;

	/* A reprieve! */
	netif_start_queue(netdev);
3317
	++adapter->restart_queue;
3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328
	return 0;
}

static int e1000_maybe_stop_tx(struct net_device *netdev,
                               struct e1000_tx_ring *tx_ring, int size)
{
	if (likely(E1000_DESC_UNUSED(tx_ring) >= size))
		return 0;
	return __e1000_maybe_stop_tx(netdev, size);
}

L
Linus Torvalds 已提交
3329 3330 3331 3332
#define TXD_USE_COUNT(S, X) (((S) >> (X)) + 1 )
static int
e1000_xmit_frame(struct sk_buff *skb, struct net_device *netdev)
{
3333
	struct e1000_adapter *adapter = netdev_priv(netdev);
3334
	struct e1000_tx_ring *tx_ring;
L
Linus Torvalds 已提交
3335 3336 3337
	unsigned int first, max_per_txd = E1000_MAX_DATA_PER_TXD;
	unsigned int max_txd_pwr = E1000_MAX_TXD_PWR;
	unsigned int tx_flags = 0;
3338
	unsigned int len = skb->len - skb->data_len;
L
Linus Torvalds 已提交
3339
	unsigned long flags;
3340 3341
	unsigned int nr_frags;
	unsigned int mss;
L
Linus Torvalds 已提交
3342
	int count = 0;
3343
	int tso;
L
Linus Torvalds 已提交
3344 3345
	unsigned int f;

3346 3347 3348 3349
	/* This goes back to the question of how to logically map a tx queue
	 * to a flow.  Right now, performance is impacted slightly negatively
	 * if using multiple tx queues.  If the stack breaks away from a
	 * single qdisc implementation, we can look at this again. */
3350
	tx_ring = adapter->tx_ring;
3351

3352
	if (unlikely(skb->len <= 0)) {
L
Linus Torvalds 已提交
3353 3354 3355 3356
		dev_kfree_skb_any(skb);
		return NETDEV_TX_OK;
	}

3357 3358 3359 3360 3361
	/* 82571 and newer doesn't need the workaround that limited descriptor
	 * length to 4kB */
	if (adapter->hw.mac_type >= e1000_82571)
		max_per_txd = 8192;

3362
	mss = skb_shinfo(skb)->gso_size;
3363
	/* The controller does a simple calculation to
L
Linus Torvalds 已提交
3364 3365 3366 3367 3368
	 * make sure there is enough room in the FIFO before
	 * initiating the DMA for each buffer.  The calc is:
	 * 4 = ceil(buffer len/mss).  To make sure we don't
	 * overrun the FIFO, adjust the max buffer len if mss
	 * drops. */
J
Jesse Brandeburg 已提交
3369
	if (mss) {
J
Jeff Kirsher 已提交
3370
		uint8_t hdr_len;
L
Linus Torvalds 已提交
3371 3372
		max_per_txd = min(mss << 2, max_per_txd);
		max_txd_pwr = fls(max_per_txd) - 1;
J
Jeff Kirsher 已提交
3373

3374 3375 3376
		/* TSO Workaround for 82571/2/3 Controllers -- if skb->data
		* points to just header, pull a few bytes of payload from
		* frags into skb->data */
3377
		hdr_len = skb_transport_offset(skb) + tcp_hdrlen(skb);
3378
		if (skb->data_len && hdr_len == len) {
3379 3380
			switch (adapter->hw.mac_type) {
				unsigned int pull_size;
3381 3382 3383 3384 3385 3386 3387
			case e1000_82544:
				/* Make sure we have room to chop off 4 bytes,
				 * and that the end alignment will work out to
				 * this hardware's requirements
				 * NOTE: this is a TSO only workaround
				 * if end byte alignment not correct move us
				 * into the next dword */
3388
				if ((unsigned long)(skb_tail_pointer(skb) - 1) & 4)
3389 3390
					break;
				/* fall through */
3391 3392 3393
			case e1000_82571:
			case e1000_82572:
			case e1000_82573:
3394
			case e1000_ich8lan:
3395 3396
				pull_size = min((unsigned int)4, skb->data_len);
				if (!__pskb_pull_tail(skb, pull_size)) {
A
Auke Kok 已提交
3397
					DPRINTK(DRV, ERR,
3398 3399
						"__pskb_pull_tail failed.\n");
					dev_kfree_skb_any(skb);
J
Jeff Garzik 已提交
3400
					return NETDEV_TX_OK;
3401 3402 3403 3404 3405 3406
				}
				len = skb->len - skb->data_len;
				break;
			default:
				/* do nothing */
				break;
3407
			}
J
Jeff Kirsher 已提交
3408
		}
L
Linus Torvalds 已提交
3409 3410
	}

J
Jeff Kirsher 已提交
3411
	/* reserve a descriptor for the offload context */
3412
	if ((mss) || (skb->ip_summed == CHECKSUM_PARTIAL))
L
Linus Torvalds 已提交
3413
		count++;
3414
	count++;
J
Jeff Kirsher 已提交
3415 3416

	/* Controller Erratum workaround */
H
Herbert Xu 已提交
3417
	if (!skb->data_len && tx_ring->last_tx_tso && !skb_is_gso(skb))
J
Jeff Kirsher 已提交
3418 3419
		count++;

L
Linus Torvalds 已提交
3420 3421
	count += TXD_USE_COUNT(len, max_txd_pwr);

J
Jesse Brandeburg 已提交
3422
	if (adapter->pcix_82544)
L
Linus Torvalds 已提交
3423 3424
		count++;

J
Jesse Brandeburg 已提交
3425
	/* work-around for errata 10 and it applies to all controllers
3426 3427
	 * in PCI-X mode, so add one more descriptor to the count
	 */
J
Jesse Brandeburg 已提交
3428
	if (unlikely((adapter->hw.bus_type == e1000_bus_type_pcix) &&
3429 3430 3431
			(len > 2015)))
		count++;

L
Linus Torvalds 已提交
3432
	nr_frags = skb_shinfo(skb)->nr_frags;
J
Jesse Brandeburg 已提交
3433
	for (f = 0; f < nr_frags; f++)
L
Linus Torvalds 已提交
3434 3435
		count += TXD_USE_COUNT(skb_shinfo(skb)->frags[f].size,
				       max_txd_pwr);
J
Jesse Brandeburg 已提交
3436
	if (adapter->pcix_82544)
L
Linus Torvalds 已提交
3437 3438
		count += nr_frags;

3439 3440 3441

	if (adapter->hw.tx_pkt_filtering &&
	    (adapter->hw.mac_type == e1000_82573))
3442 3443
		e1000_transfer_dhcp_info(adapter, skb);

3444
	if (!spin_trylock_irqsave(&tx_ring->tx_lock, flags))
3445 3446
		/* Collision - tell upper layer to requeue */
		return NETDEV_TX_LOCKED;
L
Linus Torvalds 已提交
3447 3448 3449

	/* need: count + 2 desc gap to keep tail from touching
	 * head, otherwise try next time */
3450
	if (unlikely(e1000_maybe_stop_tx(netdev, tx_ring, count + 2))) {
3451
		spin_unlock_irqrestore(&tx_ring->tx_lock, flags);
L
Linus Torvalds 已提交
3452 3453 3454
		return NETDEV_TX_BUSY;
	}

J
Jesse Brandeburg 已提交
3455 3456
	if (unlikely(adapter->hw.mac_type == e1000_82547)) {
		if (unlikely(e1000_82547_fifo_workaround(adapter, skb))) {
L
Linus Torvalds 已提交
3457
			netif_stop_queue(netdev);
A
Auke Kok 已提交
3458
			mod_timer(&adapter->tx_fifo_stall_timer, jiffies + 1);
3459
			spin_unlock_irqrestore(&tx_ring->tx_lock, flags);
L
Linus Torvalds 已提交
3460 3461 3462 3463
			return NETDEV_TX_BUSY;
		}
	}

J
Jesse Brandeburg 已提交
3464
	if (unlikely(adapter->vlgrp && vlan_tx_tag_present(skb))) {
L
Linus Torvalds 已提交
3465 3466 3467 3468
		tx_flags |= E1000_TX_FLAGS_VLAN;
		tx_flags |= (vlan_tx_tag_get(skb) << E1000_TX_FLAGS_VLAN_SHIFT);
	}

3469
	first = tx_ring->next_to_use;
J
Jesse Brandeburg 已提交
3470

3471
	tso = e1000_tso(adapter, tx_ring, skb);
L
Linus Torvalds 已提交
3472 3473
	if (tso < 0) {
		dev_kfree_skb_any(skb);
3474
		spin_unlock_irqrestore(&tx_ring->tx_lock, flags);
L
Linus Torvalds 已提交
3475 3476 3477
		return NETDEV_TX_OK;
	}

J
Jeff Kirsher 已提交
3478 3479
	if (likely(tso)) {
		tx_ring->last_tx_tso = 1;
L
Linus Torvalds 已提交
3480
		tx_flags |= E1000_TX_FLAGS_TSO;
J
Jeff Kirsher 已提交
3481
	} else if (likely(e1000_tx_csum(adapter, tx_ring, skb)))
L
Linus Torvalds 已提交
3482 3483
		tx_flags |= E1000_TX_FLAGS_CSUM;

3484
	/* Old method was to assume IPv4 packet by default if TSO was enabled.
3485
	 * 82571 hardware supports TSO capabilities for IPv6 as well...
3486
	 * no longer assume, we must. */
A
Alexey Dobriyan 已提交
3487
	if (likely(skb->protocol == htons(ETH_P_IP)))
3488 3489
		tx_flags |= E1000_TX_FLAGS_IPV4;

3490 3491 3492
	e1000_tx_queue(adapter, tx_ring, tx_flags,
	               e1000_tx_map(adapter, tx_ring, skb, first,
	                            max_per_txd, nr_frags, mss));
L
Linus Torvalds 已提交
3493 3494 3495 3496

	netdev->trans_start = jiffies;

	/* Make sure there is space in the ring for the next send. */
3497
	e1000_maybe_stop_tx(netdev, tx_ring, MAX_SKB_FRAGS + 2);
L
Linus Torvalds 已提交
3498

3499
	spin_unlock_irqrestore(&tx_ring->tx_lock, flags);
L
Linus Torvalds 已提交
3500 3501 3502 3503 3504 3505 3506 3507 3508 3509 3510
	return NETDEV_TX_OK;
}

/**
 * e1000_tx_timeout - Respond to a Tx Hang
 * @netdev: network interface device structure
 **/

static void
e1000_tx_timeout(struct net_device *netdev)
{
3511
	struct e1000_adapter *adapter = netdev_priv(netdev);
L
Linus Torvalds 已提交
3512 3513

	/* Do the reset outside of interrupt context */
3514 3515
	adapter->tx_timeout_count++;
	schedule_work(&adapter->reset_task);
L
Linus Torvalds 已提交
3516 3517 3518
}

static void
3519
e1000_reset_task(struct work_struct *work)
L
Linus Torvalds 已提交
3520
{
3521 3522
	struct e1000_adapter *adapter =
		container_of(work, struct e1000_adapter, reset_task);
L
Linus Torvalds 已提交
3523

3524
	e1000_reinit_locked(adapter);
L
Linus Torvalds 已提交
3525 3526 3527 3528 3529 3530 3531 3532 3533 3534 3535 3536 3537
}

/**
 * e1000_get_stats - Get System Network Statistics
 * @netdev: network interface device structure
 *
 * Returns the address of the device statistics structure.
 * The statistics are actually updated from the timer callback.
 **/

static struct net_device_stats *
e1000_get_stats(struct net_device *netdev)
{
3538
	struct e1000_adapter *adapter = netdev_priv(netdev);
L
Linus Torvalds 已提交
3539

J
Jeff Kirsher 已提交
3540
	/* only return the current stats */
L
Linus Torvalds 已提交
3541 3542 3543 3544 3545 3546 3547 3548 3549 3550 3551 3552 3553 3554
	return &adapter->net_stats;
}

/**
 * e1000_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
e1000_change_mtu(struct net_device *netdev, int new_mtu)
{
3555
	struct e1000_adapter *adapter = netdev_priv(netdev);
L
Linus Torvalds 已提交
3556
	int max_frame = new_mtu + ENET_HEADER_SIZE + ETHERNET_FCS_SIZE;
3557
	uint16_t eeprom_data = 0;
L
Linus Torvalds 已提交
3558

J
Jesse Brandeburg 已提交
3559 3560 3561
	if ((max_frame < MINIMUM_ETHERNET_FRAME_SIZE) ||
	    (max_frame > MAX_JUMBO_FRAME_SIZE)) {
		DPRINTK(PROBE, ERR, "Invalid MTU setting\n");
L
Linus Torvalds 已提交
3562
		return -EINVAL;
3563
	}
L
Linus Torvalds 已提交
3564

3565 3566
	/* Adapter-specific max frame size limits. */
	switch (adapter->hw.mac_type) {
3567
	case e1000_undefined ... e1000_82542_rev2_1:
3568
	case e1000_ich8lan:
3569 3570
		if (max_frame > MAXIMUM_ETHERNET_FRAME_SIZE) {
			DPRINTK(PROBE, ERR, "Jumbo Frames not supported.\n");
3571 3572
			return -EINVAL;
		}
3573
		break;
3574
	case e1000_82573:
3575 3576 3577
		/* Jumbo Frames not supported if:
		 * - this is not an 82573L device
		 * - ASPM is enabled in any way (0x1A bits 3:2) */
3578 3579
		e1000_read_eeprom(&adapter->hw, EEPROM_INIT_3GIO_3, 1,
		                  &eeprom_data);
3580 3581
		if ((adapter->hw.device_id != E1000_DEV_ID_82573L) ||
		    (eeprom_data & EEPROM_WORD1A_ASPM_MASK)) {
3582 3583 3584 3585 3586 3587 3588
			if (max_frame > MAXIMUM_ETHERNET_FRAME_SIZE) {
				DPRINTK(PROBE, ERR,
			            	"Jumbo Frames not supported.\n");
				return -EINVAL;
			}
			break;
		}
3589 3590
		/* ERT will be enabled later to enable wire speed receives */

3591
		/* fall through to get support */
3592 3593
	case e1000_82571:
	case e1000_82572:
3594
	case e1000_80003es2lan:
3595 3596 3597 3598 3599 3600 3601 3602 3603
#define MAX_STD_JUMBO_FRAME_SIZE 9234
		if (max_frame > MAX_STD_JUMBO_FRAME_SIZE) {
			DPRINTK(PROBE, ERR, "MTU > 9216 not supported.\n");
			return -EINVAL;
		}
		break;
	default:
		/* Capable of supporting up to MAX_JUMBO_FRAME_SIZE limit. */
		break;
L
Linus Torvalds 已提交
3604 3605
	}

3606
	/* NOTE: netdev_alloc_skb reserves 16 bytes, and typically NET_IP_ALIGN
3607 3608 3609 3610 3611 3612 3613 3614 3615 3616 3617 3618 3619 3620 3621 3622 3623 3624 3625 3626 3627 3628 3629 3630
	 * means we reserve 2 more, this pushes us to allocate from the next
	 * larger slab size
	 * i.e. RXBUFFER_2048 --> size-4096 slab */

	if (max_frame <= E1000_RXBUFFER_256)
		adapter->rx_buffer_len = E1000_RXBUFFER_256;
	else if (max_frame <= E1000_RXBUFFER_512)
		adapter->rx_buffer_len = E1000_RXBUFFER_512;
	else if (max_frame <= E1000_RXBUFFER_1024)
		adapter->rx_buffer_len = E1000_RXBUFFER_1024;
	else if (max_frame <= E1000_RXBUFFER_2048)
		adapter->rx_buffer_len = E1000_RXBUFFER_2048;
	else if (max_frame <= E1000_RXBUFFER_4096)
		adapter->rx_buffer_len = E1000_RXBUFFER_4096;
	else if (max_frame <= E1000_RXBUFFER_8192)
		adapter->rx_buffer_len = E1000_RXBUFFER_8192;
	else if (max_frame <= E1000_RXBUFFER_16384)
		adapter->rx_buffer_len = E1000_RXBUFFER_16384;

	/* adjust allocation if LPE protects us, and we aren't using SBP */
	if (!adapter->hw.tbi_compatibility_on &&
	    ((max_frame == MAXIMUM_ETHERNET_FRAME_SIZE) ||
	     (max_frame == MAXIMUM_ETHERNET_VLAN_SIZE)))
		adapter->rx_buffer_len = MAXIMUM_ETHERNET_VLAN_SIZE;
3631

3632
	netdev->mtu = new_mtu;
3633
	adapter->hw.max_frame_size = max_frame;
3634

3635 3636
	if (netif_running(netdev))
		e1000_reinit_locked(adapter);
L
Linus Torvalds 已提交
3637 3638 3639 3640 3641 3642 3643 3644 3645 3646 3647 3648 3649

	return 0;
}

/**
 * e1000_update_stats - Update the board statistics counters
 * @adapter: board private structure
 **/

void
e1000_update_stats(struct e1000_adapter *adapter)
{
	struct e1000_hw *hw = &adapter->hw;
3650
	struct pci_dev *pdev = adapter->pdev;
L
Linus Torvalds 已提交
3651 3652 3653 3654 3655
	unsigned long flags;
	uint16_t phy_tmp;

#define PHY_IDLE_ERROR_COUNT_MASK 0x00FF

3656 3657 3658 3659
	/*
	 * Prevent stats update while adapter is being reset, or if the pci
	 * connection is down.
	 */
A
Auke Kok 已提交
3660
	if (adapter->link_speed == 0)
3661
		return;
3662
	if (pci_channel_offline(pdev))
A
Auke Kok 已提交
3663 3664
		return;

L
Linus Torvalds 已提交
3665 3666
	spin_lock_irqsave(&adapter->stats_lock, flags);

3667
	/* these counters are modified from e1000_tbi_adjust_stats,
L
Linus Torvalds 已提交
3668 3669 3670 3671 3672 3673 3674 3675 3676 3677 3678
	 * called from the interrupt context, so they must only
	 * be written while holding adapter->stats_lock
	 */

	adapter->stats.crcerrs += E1000_READ_REG(hw, CRCERRS);
	adapter->stats.gprc += E1000_READ_REG(hw, GPRC);
	adapter->stats.gorcl += E1000_READ_REG(hw, GORCL);
	adapter->stats.gorch += E1000_READ_REG(hw, GORCH);
	adapter->stats.bprc += E1000_READ_REG(hw, BPRC);
	adapter->stats.mprc += E1000_READ_REG(hw, MPRC);
	adapter->stats.roc += E1000_READ_REG(hw, ROC);
3679 3680

	if (adapter->hw.mac_type != e1000_ich8lan) {
3681 3682 3683 3684 3685 3686
		adapter->stats.prc64 += E1000_READ_REG(hw, PRC64);
		adapter->stats.prc127 += E1000_READ_REG(hw, PRC127);
		adapter->stats.prc255 += E1000_READ_REG(hw, PRC255);
		adapter->stats.prc511 += E1000_READ_REG(hw, PRC511);
		adapter->stats.prc1023 += E1000_READ_REG(hw, PRC1023);
		adapter->stats.prc1522 += E1000_READ_REG(hw, PRC1522);
3687
	}
L
Linus Torvalds 已提交
3688 3689 3690 3691 3692 3693 3694 3695 3696 3697 3698 3699 3700 3701 3702 3703 3704 3705 3706 3707 3708 3709 3710 3711 3712 3713 3714

	adapter->stats.symerrs += E1000_READ_REG(hw, SYMERRS);
	adapter->stats.mpc += E1000_READ_REG(hw, MPC);
	adapter->stats.scc += E1000_READ_REG(hw, SCC);
	adapter->stats.ecol += E1000_READ_REG(hw, ECOL);
	adapter->stats.mcc += E1000_READ_REG(hw, MCC);
	adapter->stats.latecol += E1000_READ_REG(hw, LATECOL);
	adapter->stats.dc += E1000_READ_REG(hw, DC);
	adapter->stats.sec += E1000_READ_REG(hw, SEC);
	adapter->stats.rlec += E1000_READ_REG(hw, RLEC);
	adapter->stats.xonrxc += E1000_READ_REG(hw, XONRXC);
	adapter->stats.xontxc += E1000_READ_REG(hw, XONTXC);
	adapter->stats.xoffrxc += E1000_READ_REG(hw, XOFFRXC);
	adapter->stats.xofftxc += E1000_READ_REG(hw, XOFFTXC);
	adapter->stats.fcruc += E1000_READ_REG(hw, FCRUC);
	adapter->stats.gptc += E1000_READ_REG(hw, GPTC);
	adapter->stats.gotcl += E1000_READ_REG(hw, GOTCL);
	adapter->stats.gotch += E1000_READ_REG(hw, GOTCH);
	adapter->stats.rnbc += E1000_READ_REG(hw, RNBC);
	adapter->stats.ruc += E1000_READ_REG(hw, RUC);
	adapter->stats.rfc += E1000_READ_REG(hw, RFC);
	adapter->stats.rjc += E1000_READ_REG(hw, RJC);
	adapter->stats.torl += E1000_READ_REG(hw, TORL);
	adapter->stats.torh += E1000_READ_REG(hw, TORH);
	adapter->stats.totl += E1000_READ_REG(hw, TOTL);
	adapter->stats.toth += E1000_READ_REG(hw, TOTH);
	adapter->stats.tpr += E1000_READ_REG(hw, TPR);
3715 3716

	if (adapter->hw.mac_type != e1000_ich8lan) {
3717 3718 3719 3720 3721 3722
		adapter->stats.ptc64 += E1000_READ_REG(hw, PTC64);
		adapter->stats.ptc127 += E1000_READ_REG(hw, PTC127);
		adapter->stats.ptc255 += E1000_READ_REG(hw, PTC255);
		adapter->stats.ptc511 += E1000_READ_REG(hw, PTC511);
		adapter->stats.ptc1023 += E1000_READ_REG(hw, PTC1023);
		adapter->stats.ptc1522 += E1000_READ_REG(hw, PTC1522);
3723 3724
	}

L
Linus Torvalds 已提交
3725 3726 3727 3728 3729 3730 3731 3732 3733 3734
	adapter->stats.mptc += E1000_READ_REG(hw, MPTC);
	adapter->stats.bptc += E1000_READ_REG(hw, BPTC);

	/* used for adaptive IFS */

	hw->tx_packet_delta = E1000_READ_REG(hw, TPT);
	adapter->stats.tpt += hw->tx_packet_delta;
	hw->collision_delta = E1000_READ_REG(hw, COLC);
	adapter->stats.colc += hw->collision_delta;

J
Jesse Brandeburg 已提交
3735
	if (hw->mac_type >= e1000_82543) {
L
Linus Torvalds 已提交
3736 3737 3738 3739 3740 3741 3742
		adapter->stats.algnerrc += E1000_READ_REG(hw, ALGNERRC);
		adapter->stats.rxerrc += E1000_READ_REG(hw, RXERRC);
		adapter->stats.tncrs += E1000_READ_REG(hw, TNCRS);
		adapter->stats.cexterr += E1000_READ_REG(hw, CEXTERR);
		adapter->stats.tsctc += E1000_READ_REG(hw, TSCTC);
		adapter->stats.tsctfc += E1000_READ_REG(hw, TSCTFC);
	}
J
Jesse Brandeburg 已提交
3743
	if (hw->mac_type > e1000_82547_rev_2) {
3744 3745
		adapter->stats.iac += E1000_READ_REG(hw, IAC);
		adapter->stats.icrxoc += E1000_READ_REG(hw, ICRXOC);
3746 3747

		if (adapter->hw.mac_type != e1000_ich8lan) {
3748 3749 3750 3751 3752 3753 3754
			adapter->stats.icrxptc += E1000_READ_REG(hw, ICRXPTC);
			adapter->stats.icrxatc += E1000_READ_REG(hw, ICRXATC);
			adapter->stats.ictxptc += E1000_READ_REG(hw, ICTXPTC);
			adapter->stats.ictxatc += E1000_READ_REG(hw, ICTXATC);
			adapter->stats.ictxqec += E1000_READ_REG(hw, ICTXQEC);
			adapter->stats.ictxqmtc += E1000_READ_REG(hw, ICTXQMTC);
			adapter->stats.icrxdmtc += E1000_READ_REG(hw, ICRXDMTC);
3755
		}
3756
	}
L
Linus Torvalds 已提交
3757 3758 3759 3760 3761 3762 3763

	/* Fill out the OS statistics structure */
	adapter->net_stats.multicast = adapter->stats.mprc;
	adapter->net_stats.collisions = adapter->stats.colc;

	/* Rx Errors */

3764 3765
	/* RLEC on some newer hardware can be incorrect so build
	* our own version based on RUC and ROC */
L
Linus Torvalds 已提交
3766 3767
	adapter->net_stats.rx_errors = adapter->stats.rxerrc +
		adapter->stats.crcerrs + adapter->stats.algnerrc +
3768 3769
		adapter->stats.ruc + adapter->stats.roc +
		adapter->stats.cexterr;
3770 3771
	adapter->stats.rlerrc = adapter->stats.ruc + adapter->stats.roc;
	adapter->net_stats.rx_length_errors = adapter->stats.rlerrc;
L
Linus Torvalds 已提交
3772 3773 3774 3775 3776
	adapter->net_stats.rx_crc_errors = adapter->stats.crcerrs;
	adapter->net_stats.rx_frame_errors = adapter->stats.algnerrc;
	adapter->net_stats.rx_missed_errors = adapter->stats.mpc;

	/* Tx Errors */
3777 3778
	adapter->stats.txerrc = adapter->stats.ecol + adapter->stats.latecol;
	adapter->net_stats.tx_errors = adapter->stats.txerrc;
L
Linus Torvalds 已提交
3779 3780 3781
	adapter->net_stats.tx_aborted_errors = adapter->stats.ecol;
	adapter->net_stats.tx_window_errors = adapter->stats.latecol;
	adapter->net_stats.tx_carrier_errors = adapter->stats.tncrs;
3782 3783 3784 3785 3786
	if (adapter->hw.bad_tx_carr_stats_fd &&
	    adapter->link_duplex == FULL_DUPLEX) {
		adapter->net_stats.tx_carrier_errors = 0;
		adapter->stats.tncrs = 0;
	}
L
Linus Torvalds 已提交
3787 3788 3789 3790

	/* Tx Dropped needs to be maintained elsewhere */

	/* Phy Stats */
J
Jesse Brandeburg 已提交
3791 3792
	if (hw->media_type == e1000_media_type_copper) {
		if ((adapter->link_speed == SPEED_1000) &&
L
Linus Torvalds 已提交
3793 3794 3795 3796 3797
		   (!e1000_read_phy_reg(hw, PHY_1000T_STATUS, &phy_tmp))) {
			phy_tmp &= PHY_IDLE_ERROR_COUNT_MASK;
			adapter->phy_stats.idle_errors += phy_tmp;
		}

J
Jesse Brandeburg 已提交
3798
		if ((hw->mac_type <= e1000_82546) &&
L
Linus Torvalds 已提交
3799 3800 3801 3802 3803
		   (hw->phy_type == e1000_phy_m88) &&
		   !e1000_read_phy_reg(hw, M88E1000_RX_ERR_CNTR, &phy_tmp))
			adapter->phy_stats.receive_errors += phy_tmp;
	}

3804 3805 3806 3807 3808 3809 3810
	/* Management Stats */
	if (adapter->hw.has_smbus) {
		adapter->stats.mgptc += E1000_READ_REG(hw, MGTPTC);
		adapter->stats.mgprc += E1000_READ_REG(hw, MGTPRC);
		adapter->stats.mgpdc += E1000_READ_REG(hw, MGTPDC);
	}

L
Linus Torvalds 已提交
3811 3812
	spin_unlock_irqrestore(&adapter->stats_lock, flags);
}
3813 3814 3815 3816 3817 3818 3819

/**
 * e1000_intr_msi - Interrupt Handler
 * @irq: interrupt number
 * @data: pointer to a network interface device structure
 **/

3820 3821
static irqreturn_t
e1000_intr_msi(int irq, void *data)
3822 3823 3824 3825 3826 3827 3828
{
	struct net_device *netdev = data;
	struct e1000_adapter *adapter = netdev_priv(netdev);
	struct e1000_hw *hw = &adapter->hw;
#ifndef CONFIG_E1000_NAPI
	int i;
#endif
3829
	uint32_t icr = E1000_READ_REG(hw, ICR);
3830 3831

#ifdef CONFIG_E1000_NAPI
3832 3833 3834
	/* read ICR disables interrupts using IAM, so keep up with our
	 * enable/disable accounting */
	atomic_inc(&adapter->irq_sem);
3835
#endif
3836 3837 3838 3839 3840 3841 3842 3843 3844 3845
	if (icr & (E1000_ICR_RXSEQ | E1000_ICR_LSC)) {
		hw->get_link_status = 1;
		/* 80003ES2LAN workaround-- For packet buffer work-around on
		 * link down event; disable receives here in the ISR and reset
		 * adapter in watchdog */
		if (netif_carrier_ok(netdev) &&
		    (adapter->hw.mac_type == e1000_80003es2lan)) {
			/* disable receives */
			uint32_t rctl = E1000_READ_REG(hw, RCTL);
			E1000_WRITE_REG(hw, RCTL, rctl & ~E1000_RCTL_EN);
3846
		}
3847 3848 3849
		/* guard against interrupt when we're going down */
		if (!test_bit(__E1000_DOWN, &adapter->flags))
			mod_timer(&adapter->watchdog_timer, jiffies + 1);
3850 3851 3852
	}

#ifdef CONFIG_E1000_NAPI
3853
	if (likely(netif_rx_schedule_prep(netdev, &adapter->napi))) {
J
Jesse Brandeburg 已提交
3854 3855 3856 3857
		adapter->total_tx_bytes = 0;
		adapter->total_tx_packets = 0;
		adapter->total_rx_bytes = 0;
		adapter->total_rx_packets = 0;
3858
		__netif_rx_schedule(netdev, &adapter->napi);
J
Jesse Brandeburg 已提交
3859
	} else
3860 3861
		e1000_irq_enable(adapter);
#else
J
Jesse Brandeburg 已提交
3862 3863 3864 3865 3866
	adapter->total_tx_bytes = 0;
	adapter->total_rx_bytes = 0;
	adapter->total_tx_packets = 0;
	adapter->total_rx_packets = 0;

3867 3868
	for (i = 0; i < E1000_MAX_INTR; i++)
		if (unlikely(!adapter->clean_rx(adapter, adapter->rx_ring) &
3869
		   !e1000_clean_tx_irq(adapter, adapter->tx_ring)))
3870
			break;
J
Jesse Brandeburg 已提交
3871 3872 3873

	if (likely(adapter->itr_setting & 3))
		e1000_set_itr(adapter);
3874 3875 3876 3877
#endif

	return IRQ_HANDLED;
}
L
Linus Torvalds 已提交
3878 3879 3880 3881 3882 3883 3884 3885

/**
 * e1000_intr - Interrupt Handler
 * @irq: interrupt number
 * @data: pointer to a network interface device structure
 **/

static irqreturn_t
3886
e1000_intr(int irq, void *data)
L
Linus Torvalds 已提交
3887 3888
{
	struct net_device *netdev = data;
3889
	struct e1000_adapter *adapter = netdev_priv(netdev);
L
Linus Torvalds 已提交
3890
	struct e1000_hw *hw = &adapter->hw;
3891
	uint32_t rctl, icr = E1000_READ_REG(hw, ICR);
3892
#ifndef CONFIG_E1000_NAPI
3893
	int i;
J
Jesse Brandeburg 已提交
3894 3895 3896 3897 3898 3899 3900 3901 3902 3903 3904
#endif
	if (unlikely(!icr))
		return IRQ_NONE;  /* Not our interrupt */

#ifdef CONFIG_E1000_NAPI
	/* IMS will not auto-mask if INT_ASSERTED is not set, and if it is
	 * not set, then the adapter didn't send an interrupt */
	if (unlikely(hw->mac_type >= e1000_82571 &&
	             !(icr & E1000_ICR_INT_ASSERTED)))
		return IRQ_NONE;

3905 3906 3907 3908 3909 3910
	/* Interrupt Auto-Mask...upon reading ICR,
	 * interrupts are masked.  No need for the
	 * IMC write, but it does mean we should
	 * account for it ASAP. */
	if (likely(hw->mac_type >= e1000_82571))
		atomic_inc(&adapter->irq_sem);
J
Jeff Garzik 已提交
3911
#endif
L
Linus Torvalds 已提交
3912

J
Jesse Brandeburg 已提交
3913
	if (unlikely(icr & (E1000_ICR_RXSEQ | E1000_ICR_LSC))) {
L
Linus Torvalds 已提交
3914
		hw->get_link_status = 1;
3915 3916 3917 3918 3919 3920 3921 3922 3923 3924 3925
		/* 80003ES2LAN workaround--
		 * For packet buffer work-around on link down event;
		 * disable receives here in the ISR and
		 * reset adapter in watchdog
		 */
		if (netif_carrier_ok(netdev) &&
		    (adapter->hw.mac_type == e1000_80003es2lan)) {
			/* disable receives */
			rctl = E1000_READ_REG(hw, RCTL);
			E1000_WRITE_REG(hw, RCTL, rctl & ~E1000_RCTL_EN);
		}
A
Auke Kok 已提交
3926 3927 3928
		/* guard against interrupt when we're going down */
		if (!test_bit(__E1000_DOWN, &adapter->flags))
			mod_timer(&adapter->watchdog_timer, jiffies + 1);
L
Linus Torvalds 已提交
3929 3930 3931
	}

#ifdef CONFIG_E1000_NAPI
3932
	if (unlikely(hw->mac_type < e1000_82571)) {
J
Jesse Brandeburg 已提交
3933
		/* disable interrupts, without the synchronize_irq bit */
3934 3935 3936 3937
		atomic_inc(&adapter->irq_sem);
		E1000_WRITE_REG(hw, IMC, ~0);
		E1000_WRITE_FLUSH(hw);
	}
3938
	if (likely(netif_rx_schedule_prep(netdev, &adapter->napi))) {
J
Jesse Brandeburg 已提交
3939 3940 3941 3942
		adapter->total_tx_bytes = 0;
		adapter->total_tx_packets = 0;
		adapter->total_rx_bytes = 0;
		adapter->total_rx_packets = 0;
3943
		__netif_rx_schedule(netdev, &adapter->napi);
J
Jesse Brandeburg 已提交
3944
	} else
3945 3946
		/* this really should not happen! if it does it is basically a
		 * bug, but not a hard error, so enable ints and continue */
3947
		e1000_irq_enable(adapter);
3948
#else
L
Linus Torvalds 已提交
3949
	/* Writing IMC and IMS is needed for 82547.
J
Jesse Brandeburg 已提交
3950 3951 3952 3953 3954 3955 3956 3957 3958 3959
	 * Due to Hub Link bus being occupied, an interrupt
	 * de-assertion message is not able to be sent.
	 * When an interrupt assertion message is generated later,
	 * two messages are re-ordered and sent out.
	 * That causes APIC to think 82547 is in de-assertion
	 * state, while 82547 is in assertion state, resulting
	 * in dead lock. Writing IMC forces 82547 into
	 * de-assertion state.
	 */
	if (hw->mac_type == e1000_82547 || hw->mac_type == e1000_82547_rev_2) {
L
Linus Torvalds 已提交
3960
		atomic_inc(&adapter->irq_sem);
3961
		E1000_WRITE_REG(hw, IMC, ~0);
L
Linus Torvalds 已提交
3962 3963
	}

J
Jesse Brandeburg 已提交
3964 3965 3966 3967 3968
	adapter->total_tx_bytes = 0;
	adapter->total_rx_bytes = 0;
	adapter->total_tx_packets = 0;
	adapter->total_rx_packets = 0;

J
Jesse Brandeburg 已提交
3969 3970
	for (i = 0; i < E1000_MAX_INTR; i++)
		if (unlikely(!adapter->clean_rx(adapter, adapter->rx_ring) &
3971
		   !e1000_clean_tx_irq(adapter, adapter->tx_ring)))
L
Linus Torvalds 已提交
3972 3973
			break;

J
Jesse Brandeburg 已提交
3974 3975 3976
	if (likely(adapter->itr_setting & 3))
		e1000_set_itr(adapter);

J
Jesse Brandeburg 已提交
3977
	if (hw->mac_type == e1000_82547 || hw->mac_type == e1000_82547_rev_2)
L
Linus Torvalds 已提交
3978
		e1000_irq_enable(adapter);
3979

3980
#endif
L
Linus Torvalds 已提交
3981 3982 3983 3984 3985 3986 3987 3988 3989 3990
	return IRQ_HANDLED;
}

#ifdef CONFIG_E1000_NAPI
/**
 * e1000_clean - NAPI Rx polling callback
 * @adapter: board private structure
 **/

static int
3991
e1000_clean(struct napi_struct *napi, int budget)
L
Linus Torvalds 已提交
3992
{
3993 3994
	struct e1000_adapter *adapter = container_of(napi, struct e1000_adapter, napi);
	struct net_device *poll_dev = adapter->netdev;
3995
	int tx_cleaned = 0, work_done = 0;
3996 3997 3998 3999

	/* Must NOT use netdev_priv macro here. */
	adapter = poll_dev->priv;

4000 4001 4002 4003 4004
	/* e1000_clean is called per-cpu.  This lock protects
	 * tx_ring[0] from being cleaned by multiple cpus
	 * simultaneously.  A failure obtaining the lock means
	 * tx_ring[0] is currently being cleaned anyway. */
	if (spin_trylock(&adapter->tx_queue_lock)) {
4005 4006
		tx_cleaned = e1000_clean_tx_irq(adapter,
						&adapter->tx_ring[0]);
4007
		spin_unlock(&adapter->tx_queue_lock);
4008 4009
	}

4010
	adapter->clean_rx(adapter, &adapter->rx_ring[0],
4011
	                  &work_done, budget);
J
Jesse Brandeburg 已提交
4012

4013 4014 4015
	if (tx_cleaned)
		work_done = budget;

4016 4017
	/* If budget not fully consumed, exit the polling mode */
	if (work_done < budget) {
J
Jesse Brandeburg 已提交
4018 4019
		if (likely(adapter->itr_setting & 3))
			e1000_set_itr(adapter);
4020
		netif_rx_complete(poll_dev, napi);
L
Linus Torvalds 已提交
4021 4022 4023
		e1000_irq_enable(adapter);
	}

4024
	return work_done;
L
Linus Torvalds 已提交
4025 4026 4027 4028 4029 4030 4031 4032 4033
}

#endif
/**
 * e1000_clean_tx_irq - Reclaim resources after transmit completes
 * @adapter: board private structure
 **/

static boolean_t
4034 4035
e1000_clean_tx_irq(struct e1000_adapter *adapter,
                   struct e1000_tx_ring *tx_ring)
L
Linus Torvalds 已提交
4036 4037 4038 4039 4040
{
	struct net_device *netdev = adapter->netdev;
	struct e1000_tx_desc *tx_desc, *eop_desc;
	struct e1000_buffer *buffer_info;
	unsigned int i, eop;
4041 4042 4043
#ifdef CONFIG_E1000_NAPI
	unsigned int count = 0;
#endif
4044
	boolean_t cleaned = FALSE;
J
Jesse Brandeburg 已提交
4045
	unsigned int total_tx_bytes=0, total_tx_packets=0;
L
Linus Torvalds 已提交
4046 4047 4048 4049 4050

	i = tx_ring->next_to_clean;
	eop = tx_ring->buffer_info[i].next_to_watch;
	eop_desc = E1000_TX_DESC(*tx_ring, eop);

4051
	while (eop_desc->upper.data & cpu_to_le32(E1000_TXD_STAT_DD)) {
J
Jesse Brandeburg 已提交
4052
		for (cleaned = FALSE; !cleaned; ) {
L
Linus Torvalds 已提交
4053 4054 4055 4056
			tx_desc = E1000_TX_DESC(*tx_ring, i);
			buffer_info = &tx_ring->buffer_info[i];
			cleaned = (i == eop);

J
Jesse Brandeburg 已提交
4057
			if (cleaned) {
4058
				struct sk_buff *skb = buffer_info->skb;
4059 4060 4061 4062 4063
				unsigned int segs, bytecount;
				segs = skb_shinfo(skb)->gso_segs ?: 1;
				/* multiply data chunks by size of headers */
				bytecount = ((segs - 1) * skb_headlen(skb)) +
				            skb->len;
4064
				total_tx_packets += segs;
4065
				total_tx_bytes += bytecount;
J
Jesse Brandeburg 已提交
4066
			}
J
Jeff Kirsher 已提交
4067
			e1000_unmap_and_free_tx_resource(adapter, buffer_info);
4068
			tx_desc->upper.data = 0;
L
Linus Torvalds 已提交
4069

J
Jesse Brandeburg 已提交
4070
			if (unlikely(++i == tx_ring->count)) i = 0;
L
Linus Torvalds 已提交
4071
		}
4072

L
Linus Torvalds 已提交
4073 4074
		eop = tx_ring->buffer_info[i].next_to_watch;
		eop_desc = E1000_TX_DESC(*tx_ring, eop);
4075 4076 4077
#ifdef CONFIG_E1000_NAPI
#define E1000_TX_WEIGHT 64
		/* weight of a sort for tx, to avoid endless transmit cleanup */
4078
		if (count++ == E1000_TX_WEIGHT) break;
4079
#endif
L
Linus Torvalds 已提交
4080 4081 4082 4083
	}

	tx_ring->next_to_clean = i;

4084
#define TX_WAKE_THRESHOLD 32
4085 4086 4087 4088 4089 4090
	if (unlikely(cleaned && netif_carrier_ok(netdev) &&
		     E1000_DESC_UNUSED(tx_ring) >= TX_WAKE_THRESHOLD)) {
		/* Make sure that anybody stopping the queue after this
		 * sees the new next_to_clean.
		 */
		smp_mb();
4091
		if (netif_queue_stopped(netdev)) {
4092
			netif_wake_queue(netdev);
4093 4094
			++adapter->restart_queue;
		}
4095
	}
4096

4097
	if (adapter->detect_tx_hung) {
4098
		/* Detect a transmit hang in hardware, this serializes the
L
Linus Torvalds 已提交
4099 4100
		 * check with the clearing of time_stamp and movement of i */
		adapter->detect_tx_hung = FALSE;
4101 4102
		if (tx_ring->buffer_info[eop].dma &&
		    time_after(jiffies, tx_ring->buffer_info[eop].time_stamp +
4103
		               (adapter->tx_timeout_factor * HZ))
4104
		    && !(E1000_READ_REG(&adapter->hw, STATUS) &
4105
		         E1000_STATUS_TXOFF)) {
4106 4107

			/* detected Tx unit hang */
4108
			DPRINTK(DRV, ERR, "Detected Tx Unit Hang\n"
4109
					"  Tx Queue             <%lu>\n"
4110 4111 4112 4113 4114 4115 4116 4117 4118
					"  TDH                  <%x>\n"
					"  TDT                  <%x>\n"
					"  next_to_use          <%x>\n"
					"  next_to_clean        <%x>\n"
					"buffer_info[next_to_clean]\n"
					"  time_stamp           <%lx>\n"
					"  next_to_watch        <%x>\n"
					"  jiffies              <%lx>\n"
					"  next_to_watch.status <%x>\n",
4119 4120
				(unsigned long)((tx_ring - adapter->tx_ring) /
					sizeof(struct e1000_tx_ring)),
4121 4122
				readl(adapter->hw.hw_addr + tx_ring->tdh),
				readl(adapter->hw.hw_addr + tx_ring->tdt),
4123
				tx_ring->next_to_use,
4124 4125
				tx_ring->next_to_clean,
				tx_ring->buffer_info[eop].time_stamp,
4126 4127 4128
				eop,
				jiffies,
				eop_desc->upper.fields.status);
L
Linus Torvalds 已提交
4129
			netif_stop_queue(netdev);
4130
		}
L
Linus Torvalds 已提交
4131
	}
J
Jesse Brandeburg 已提交
4132 4133
	adapter->total_tx_bytes += total_tx_bytes;
	adapter->total_tx_packets += total_tx_packets;
4134 4135
	adapter->net_stats.tx_bytes += total_tx_bytes;
	adapter->net_stats.tx_packets += total_tx_packets;
L
Linus Torvalds 已提交
4136 4137 4138 4139 4140
	return cleaned;
}

/**
 * e1000_rx_checksum - Receive Checksum Offload for 82543
4141 4142 4143 4144
 * @adapter:     board private structure
 * @status_err:  receive descriptor status and error fields
 * @csum:        receive descriptor csum field
 * @sk_buff:     socket buffer with received data
L
Linus Torvalds 已提交
4145 4146
 **/

4147
static void
L
Linus Torvalds 已提交
4148
e1000_rx_checksum(struct e1000_adapter *adapter,
4149 4150
		  uint32_t status_err, uint32_t csum,
		  struct sk_buff *skb)
L
Linus Torvalds 已提交
4151
{
4152 4153 4154 4155
	uint16_t status = (uint16_t)status_err;
	uint8_t errors = (uint8_t)(status_err >> 24);
	skb->ip_summed = CHECKSUM_NONE;

L
Linus Torvalds 已提交
4156
	/* 82543 or newer only */
J
Jesse Brandeburg 已提交
4157
	if (unlikely(adapter->hw.mac_type < e1000_82543)) return;
L
Linus Torvalds 已提交
4158
	/* Ignore Checksum bit is set */
J
Jesse Brandeburg 已提交
4159
	if (unlikely(status & E1000_RXD_STAT_IXSM)) return;
4160
	/* TCP/UDP checksum error bit is set */
J
Jesse Brandeburg 已提交
4161
	if (unlikely(errors & E1000_RXD_ERR_TCPE)) {
L
Linus Torvalds 已提交
4162 4163
		/* let the stack verify checksum errors */
		adapter->hw_csum_err++;
4164 4165 4166
		return;
	}
	/* TCP/UDP Checksum has not been calculated */
J
Jesse Brandeburg 已提交
4167 4168
	if (adapter->hw.mac_type <= e1000_82547_rev_2) {
		if (!(status & E1000_RXD_STAT_TCPCS))
4169
			return;
L
Linus Torvalds 已提交
4170
	} else {
J
Jesse Brandeburg 已提交
4171
		if (!(status & (E1000_RXD_STAT_TCPCS | E1000_RXD_STAT_UDPCS)))
4172 4173 4174 4175
			return;
	}
	/* It must be a TCP or UDP packet with a valid checksum */
	if (likely(status & E1000_RXD_STAT_TCPCS)) {
L
Linus Torvalds 已提交
4176 4177
		/* TCP checksum is good */
		skb->ip_summed = CHECKSUM_UNNECESSARY;
4178 4179 4180 4181 4182
	} else if (adapter->hw.mac_type > e1000_82547_rev_2) {
		/* IP fragment with UDP payload */
		/* Hardware complements the payload checksum, so we undo it
		 * and then put the value in host order for further stack use.
		 */
A
Al Viro 已提交
4183 4184
		__sum16 sum = (__force __sum16)htons(csum);
		skb->csum = csum_unfold(~sum);
4185
		skb->ip_summed = CHECKSUM_COMPLETE;
L
Linus Torvalds 已提交
4186
	}
4187
	adapter->hw_csum_good++;
L
Linus Torvalds 已提交
4188 4189 4190
}

/**
4191
 * e1000_clean_rx_irq - Send received data up the network stack; legacy
L
Linus Torvalds 已提交
4192 4193 4194 4195 4196
 * @adapter: board private structure
 **/

static boolean_t
#ifdef CONFIG_E1000_NAPI
4197 4198 4199
e1000_clean_rx_irq(struct e1000_adapter *adapter,
                   struct e1000_rx_ring *rx_ring,
                   int *work_done, int work_to_do)
L
Linus Torvalds 已提交
4200
#else
4201 4202
e1000_clean_rx_irq(struct e1000_adapter *adapter,
                   struct e1000_rx_ring *rx_ring)
L
Linus Torvalds 已提交
4203 4204 4205 4206
#endif
{
	struct net_device *netdev = adapter->netdev;
	struct pci_dev *pdev = adapter->pdev;
4207 4208
	struct e1000_rx_desc *rx_desc, *next_rxd;
	struct e1000_buffer *buffer_info, *next_buffer;
L
Linus Torvalds 已提交
4209 4210 4211 4212
	unsigned long flags;
	uint32_t length;
	uint8_t last_byte;
	unsigned int i;
4213
	int cleaned_count = 0;
4214
	boolean_t cleaned = FALSE;
J
Jesse Brandeburg 已提交
4215
	unsigned int total_rx_bytes=0, total_rx_packets=0;
L
Linus Torvalds 已提交
4216 4217 4218

	i = rx_ring->next_to_clean;
	rx_desc = E1000_RX_DESC(*rx_ring, i);
4219
	buffer_info = &rx_ring->buffer_info[i];
L
Linus Torvalds 已提交
4220

4221
	while (rx_desc->status & E1000_RXD_STAT_DD) {
4222
		struct sk_buff *skb;
4223
		u8 status;
4224

L
Linus Torvalds 已提交
4225
#ifdef CONFIG_E1000_NAPI
J
Jesse Brandeburg 已提交
4226
		if (*work_done >= work_to_do)
L
Linus Torvalds 已提交
4227 4228 4229
			break;
		(*work_done)++;
#endif
4230
		status = rx_desc->status;
4231
		skb = buffer_info->skb;
4232 4233
		buffer_info->skb = NULL;

4234 4235
		prefetch(skb->data - NET_IP_ALIGN);

4236 4237
		if (++i == rx_ring->count) i = 0;
		next_rxd = E1000_RX_DESC(*rx_ring, i);
4238 4239
		prefetch(next_rxd);

4240 4241
		next_buffer = &rx_ring->buffer_info[i];

4242 4243
		cleaned = TRUE;
		cleaned_count++;
4244 4245 4246
		pci_unmap_single(pdev,
		                 buffer_info->dma,
		                 buffer_info->length,
L
Linus Torvalds 已提交
4247 4248 4249 4250
		                 PCI_DMA_FROMDEVICE);

		length = le16_to_cpu(rx_desc->length);

4251 4252 4253 4254
		if (unlikely(!(status & E1000_RXD_STAT_EOP))) {
			/* All receives must fit into a single buffer */
			E1000_DBG("%s: Receive packet consumed multiple"
				  " buffers\n", netdev->name);
A
Auke Kok 已提交
4255
			/* recycle */
4256
			buffer_info->skb = skb;
L
Linus Torvalds 已提交
4257 4258 4259
			goto next_desc;
		}

J
Jesse Brandeburg 已提交
4260
		if (unlikely(rx_desc->errors & E1000_RXD_ERR_FRAME_ERR_MASK)) {
L
Linus Torvalds 已提交
4261
			last_byte = *(skb->data + length - 1);
4262
			if (TBI_ACCEPT(&adapter->hw, status,
L
Linus Torvalds 已提交
4263 4264
			              rx_desc->errors, length, last_byte)) {
				spin_lock_irqsave(&adapter->stats_lock, flags);
4265 4266
				e1000_tbi_adjust_stats(&adapter->hw,
				                       &adapter->stats,
L
Linus Torvalds 已提交
4267 4268 4269 4270 4271
				                       length, skb->data);
				spin_unlock_irqrestore(&adapter->stats_lock,
				                       flags);
				length--;
			} else {
4272 4273
				/* recycle */
				buffer_info->skb = skb;
L
Linus Torvalds 已提交
4274 4275
				goto next_desc;
			}
A
Auke Kok 已提交
4276
		}
L
Linus Torvalds 已提交
4277

4278 4279 4280 4281
		/* adjust length to remove Ethernet CRC, this must be
		 * done after the TBI_ACCEPT workaround above */
		length -= 4;

J
Jesse Brandeburg 已提交
4282 4283 4284 4285
		/* probably a little skewed due to removing CRC */
		total_rx_bytes += length;
		total_rx_packets++;

4286 4287 4288
		/* code added for copybreak, this should improve
		 * performance for small packets with large amounts
		 * of reassembly being done in the stack */
4289
		if (length < copybreak) {
4290
			struct sk_buff *new_skb =
4291
			    netdev_alloc_skb(netdev, length + NET_IP_ALIGN);
4292 4293
			if (new_skb) {
				skb_reserve(new_skb, NET_IP_ALIGN);
4294 4295 4296 4297 4298 4299
				skb_copy_to_linear_data_offset(new_skb,
							       -NET_IP_ALIGN,
							       (skb->data -
							        NET_IP_ALIGN),
							       (length +
							        NET_IP_ALIGN));
4300 4301 4302 4303
				/* save the skb in buffer_info as good */
				buffer_info->skb = skb;
				skb = new_skb;
			}
A
Auke Kok 已提交
4304 4305
			/* else just continue with the old one */
		}
4306
		/* end copybreak code */
A
Auke Kok 已提交
4307
		skb_put(skb, length);
L
Linus Torvalds 已提交
4308 4309

		/* Receive Checksum Offload */
4310 4311
		e1000_rx_checksum(adapter,
				  (uint32_t)(status) |
4312
				  ((uint32_t)(rx_desc->errors) << 24),
D
David S. Miller 已提交
4313
				  le16_to_cpu(rx_desc->csum), skb);
J
Jesse Brandeburg 已提交
4314

L
Linus Torvalds 已提交
4315 4316
		skb->protocol = eth_type_trans(skb, netdev);
#ifdef CONFIG_E1000_NAPI
J
Jesse Brandeburg 已提交
4317
		if (unlikely(adapter->vlgrp &&
4318
			    (status & E1000_RXD_STAT_VP))) {
L
Linus Torvalds 已提交
4319
			vlan_hwaccel_receive_skb(skb, adapter->vlgrp,
4320 4321
						 le16_to_cpu(rx_desc->special) &
						 E1000_RXD_SPC_VLAN_MASK);
L
Linus Torvalds 已提交
4322 4323 4324 4325
		} else {
			netif_receive_skb(skb);
		}
#else /* CONFIG_E1000_NAPI */
J
Jesse Brandeburg 已提交
4326
		if (unlikely(adapter->vlgrp &&
4327
			    (status & E1000_RXD_STAT_VP))) {
L
Linus Torvalds 已提交
4328 4329 4330 4331 4332 4333 4334 4335 4336 4337 4338 4339
			vlan_hwaccel_rx(skb, adapter->vlgrp,
					le16_to_cpu(rx_desc->special) &
					E1000_RXD_SPC_VLAN_MASK);
		} else {
			netif_rx(skb);
		}
#endif /* CONFIG_E1000_NAPI */
		netdev->last_rx = jiffies;

next_desc:
		rx_desc->status = 0;

4340 4341 4342 4343 4344 4345
		/* return some buffers to hardware, one at a time is too slow */
		if (unlikely(cleaned_count >= E1000_RX_BUFFER_WRITE)) {
			adapter->alloc_rx_buf(adapter, rx_ring, cleaned_count);
			cleaned_count = 0;
		}

4346
		/* use prefetched values */
4347 4348
		rx_desc = next_rxd;
		buffer_info = next_buffer;
L
Linus Torvalds 已提交
4349 4350
	}
	rx_ring->next_to_clean = i;
4351 4352 4353 4354

	cleaned_count = E1000_DESC_UNUSED(rx_ring);
	if (cleaned_count)
		adapter->alloc_rx_buf(adapter, rx_ring, cleaned_count);
4355

J
Jesse Brandeburg 已提交
4356 4357
	adapter->total_rx_packets += total_rx_packets;
	adapter->total_rx_bytes += total_rx_bytes;
4358 4359
	adapter->net_stats.rx_bytes += total_rx_bytes;
	adapter->net_stats.rx_packets += total_rx_packets;
4360 4361 4362 4363 4364 4365 4366 4367 4368 4369
	return cleaned;
}

/**
 * e1000_clean_rx_irq_ps - Send received data up the network stack; packet split
 * @adapter: board private structure
 **/

static boolean_t
#ifdef CONFIG_E1000_NAPI
4370 4371 4372
e1000_clean_rx_irq_ps(struct e1000_adapter *adapter,
                      struct e1000_rx_ring *rx_ring,
                      int *work_done, int work_to_do)
4373
#else
4374 4375
e1000_clean_rx_irq_ps(struct e1000_adapter *adapter,
                      struct e1000_rx_ring *rx_ring)
4376 4377
#endif
{
4378
	union e1000_rx_desc_packet_split *rx_desc, *next_rxd;
4379 4380
	struct net_device *netdev = adapter->netdev;
	struct pci_dev *pdev = adapter->pdev;
4381
	struct e1000_buffer *buffer_info, *next_buffer;
4382 4383
	struct e1000_ps_page *ps_page;
	struct e1000_ps_page_dma *ps_page_dma;
4384
	struct sk_buff *skb;
4385 4386
	unsigned int i, j;
	uint32_t length, staterr;
4387
	int cleaned_count = 0;
4388
	boolean_t cleaned = FALSE;
J
Jesse Brandeburg 已提交
4389
	unsigned int total_rx_bytes=0, total_rx_packets=0;
4390 4391 4392

	i = rx_ring->next_to_clean;
	rx_desc = E1000_RX_DESC_PS(*rx_ring, i);
4393
	staterr = le32_to_cpu(rx_desc->wb.middle.status_error);
4394
	buffer_info = &rx_ring->buffer_info[i];
4395

J
Jesse Brandeburg 已提交
4396
	while (staterr & E1000_RXD_STAT_DD) {
4397 4398 4399
		ps_page = &rx_ring->ps_page[i];
		ps_page_dma = &rx_ring->ps_page_dma[i];
#ifdef CONFIG_E1000_NAPI
J
Jesse Brandeburg 已提交
4400
		if (unlikely(*work_done >= work_to_do))
4401 4402 4403
			break;
		(*work_done)++;
#endif
4404 4405
		skb = buffer_info->skb;

4406 4407 4408
		/* in the packet split case this is header only */
		prefetch(skb->data - NET_IP_ALIGN);

4409 4410
		if (++i == rx_ring->count) i = 0;
		next_rxd = E1000_RX_DESC_PS(*rx_ring, i);
4411 4412
		prefetch(next_rxd);

4413 4414
		next_buffer = &rx_ring->buffer_info[i];

4415
		cleaned = TRUE;
4416
		cleaned_count++;
4417 4418 4419 4420
		pci_unmap_single(pdev, buffer_info->dma,
				 buffer_info->length,
				 PCI_DMA_FROMDEVICE);

J
Jesse Brandeburg 已提交
4421
		if (unlikely(!(staterr & E1000_RXD_STAT_EOP))) {
4422 4423 4424 4425 4426
			E1000_DBG("%s: Packet Split buffers didn't pick up"
				  " the full packet\n", netdev->name);
			dev_kfree_skb_irq(skb);
			goto next_desc;
		}
L
Linus Torvalds 已提交
4427

J
Jesse Brandeburg 已提交
4428
		if (unlikely(staterr & E1000_RXDEXT_ERR_FRAME_ERR_MASK)) {
4429 4430 4431 4432 4433 4434
			dev_kfree_skb_irq(skb);
			goto next_desc;
		}

		length = le16_to_cpu(rx_desc->wb.middle.length0);

J
Jesse Brandeburg 已提交
4435
		if (unlikely(!length)) {
4436 4437 4438 4439 4440 4441 4442 4443 4444
			E1000_DBG("%s: Last part of the packet spanning"
				  " multiple descriptors\n", netdev->name);
			dev_kfree_skb_irq(skb);
			goto next_desc;
		}

		/* Good Receive */
		skb_put(skb, length);

4445 4446 4447 4448 4449 4450 4451
		{
		/* this looks ugly, but it seems compiler issues make it
		   more efficient than reusing j */
		int l1 = le16_to_cpu(rx_desc->wb.upper.length[0]);

		/* page alloc/put takes too long and effects small packet
		 * throughput, so unsplit small packets and save the alloc/put*/
4452
		if (l1 && (l1 <= copybreak) && ((length + l1) <= adapter->rx_ps_bsize0)) {
4453
			u8 *vaddr;
4454
			/* there is no documentation about how to call
4455 4456 4457 4458 4459 4460 4461 4462
			 * kmap_atomic, so we can't hold the mapping
			 * very long */
			pci_dma_sync_single_for_cpu(pdev,
				ps_page_dma->ps_page_dma[0],
				PAGE_SIZE,
				PCI_DMA_FROMDEVICE);
			vaddr = kmap_atomic(ps_page->ps_page[0],
			                    KM_SKB_DATA_SOFTIRQ);
4463
			memcpy(skb_tail_pointer(skb), vaddr, l1);
4464 4465 4466 4467
			kunmap_atomic(vaddr, KM_SKB_DATA_SOFTIRQ);
			pci_dma_sync_single_for_device(pdev,
				ps_page_dma->ps_page_dma[0],
				PAGE_SIZE, PCI_DMA_FROMDEVICE);
4468 4469
			/* remove the CRC */
			l1 -= 4;
4470 4471 4472 4473
			skb_put(skb, l1);
			goto copydone;
		} /* if */
		}
4474

J
Jesse Brandeburg 已提交
4475
		for (j = 0; j < adapter->rx_ps_pages; j++) {
4476
			if (!(length= le16_to_cpu(rx_desc->wb.upper.length[j])))
4477 4478 4479 4480
				break;
			pci_unmap_page(pdev, ps_page_dma->ps_page_dma[j],
					PAGE_SIZE, PCI_DMA_FROMDEVICE);
			ps_page_dma->ps_page_dma[j] = 0;
4481 4482
			skb_fill_page_desc(skb, j, ps_page->ps_page[j], 0,
			                   length);
4483 4484 4485
			ps_page->ps_page[j] = NULL;
			skb->len += length;
			skb->data_len += length;
4486
			skb->truesize += length;
4487 4488
		}

4489 4490 4491 4492
		/* strip the ethernet crc, problem is we're using pages now so
		 * this whole operation can get a little cpu intensive */
		pskb_trim(skb, skb->len - 4);

4493
copydone:
J
Jesse Brandeburg 已提交
4494 4495 4496
		total_rx_bytes += skb->len;
		total_rx_packets++;

4497
		e1000_rx_checksum(adapter, staterr,
D
David S. Miller 已提交
4498
				  le16_to_cpu(rx_desc->wb.lower.hi_dword.csum_ip.csum), skb);
4499 4500
		skb->protocol = eth_type_trans(skb, netdev);

J
Jesse Brandeburg 已提交
4501
		if (likely(rx_desc->wb.upper.header_status &
D
David S. Miller 已提交
4502
			   cpu_to_le16(E1000_RXDPS_HDRSTAT_HDRSP)))
4503
			adapter->rx_hdr_split++;
4504
#ifdef CONFIG_E1000_NAPI
J
Jesse Brandeburg 已提交
4505
		if (unlikely(adapter->vlgrp && (staterr & E1000_RXD_STAT_VP))) {
4506
			vlan_hwaccel_receive_skb(skb, adapter->vlgrp,
4507 4508
				le16_to_cpu(rx_desc->wb.middle.vlan) &
				E1000_RXD_SPC_VLAN_MASK);
4509 4510 4511 4512
		} else {
			netif_receive_skb(skb);
		}
#else /* CONFIG_E1000_NAPI */
J
Jesse Brandeburg 已提交
4513
		if (unlikely(adapter->vlgrp && (staterr & E1000_RXD_STAT_VP))) {
4514
			vlan_hwaccel_rx(skb, adapter->vlgrp,
4515 4516
				le16_to_cpu(rx_desc->wb.middle.vlan) &
				E1000_RXD_SPC_VLAN_MASK);
4517 4518 4519 4520 4521 4522 4523
		} else {
			netif_rx(skb);
		}
#endif /* CONFIG_E1000_NAPI */
		netdev->last_rx = jiffies;

next_desc:
D
David S. Miller 已提交
4524
		rx_desc->wb.middle.status_error &= cpu_to_le32(~0xFF);
4525 4526
		buffer_info->skb = NULL;

4527 4528 4529 4530 4531 4532
		/* return some buffers to hardware, one at a time is too slow */
		if (unlikely(cleaned_count >= E1000_RX_BUFFER_WRITE)) {
			adapter->alloc_rx_buf(adapter, rx_ring, cleaned_count);
			cleaned_count = 0;
		}

4533
		/* use prefetched values */
4534 4535 4536
		rx_desc = next_rxd;
		buffer_info = next_buffer;

4537
		staterr = le32_to_cpu(rx_desc->wb.middle.status_error);
4538 4539
	}
	rx_ring->next_to_clean = i;
4540 4541 4542 4543

	cleaned_count = E1000_DESC_UNUSED(rx_ring);
	if (cleaned_count)
		adapter->alloc_rx_buf(adapter, rx_ring, cleaned_count);
L
Linus Torvalds 已提交
4544

J
Jesse Brandeburg 已提交
4545 4546
	adapter->total_rx_packets += total_rx_packets;
	adapter->total_rx_bytes += total_rx_bytes;
4547 4548
	adapter->net_stats.rx_bytes += total_rx_bytes;
	adapter->net_stats.rx_packets += total_rx_packets;
L
Linus Torvalds 已提交
4549 4550 4551 4552
	return cleaned;
}

/**
4553
 * e1000_alloc_rx_buffers - Replace used receive buffers; legacy & extended
L
Linus Torvalds 已提交
4554 4555 4556 4557
 * @adapter: address of board private structure
 **/

static void
4558
e1000_alloc_rx_buffers(struct e1000_adapter *adapter,
4559
                       struct e1000_rx_ring *rx_ring,
4560
		       int cleaned_count)
L
Linus Torvalds 已提交
4561 4562 4563 4564 4565 4566
{
	struct net_device *netdev = adapter->netdev;
	struct pci_dev *pdev = adapter->pdev;
	struct e1000_rx_desc *rx_desc;
	struct e1000_buffer *buffer_info;
	struct sk_buff *skb;
4567 4568
	unsigned int i;
	unsigned int bufsz = adapter->rx_buffer_len + NET_IP_ALIGN;
L
Linus Torvalds 已提交
4569 4570 4571 4572

	i = rx_ring->next_to_use;
	buffer_info = &rx_ring->buffer_info[i];

4573
	while (cleaned_count--) {
4574 4575
		skb = buffer_info->skb;
		if (skb) {
4576 4577 4578 4579
			skb_trim(skb, 0);
			goto map_skb;
		}

4580
		skb = netdev_alloc_skb(netdev, bufsz);
J
Jesse Brandeburg 已提交
4581
		if (unlikely(!skb)) {
L
Linus Torvalds 已提交
4582
			/* Better luck next round */
4583
			adapter->alloc_rx_buff_failed++;
L
Linus Torvalds 已提交
4584 4585 4586
			break;
		}

4587
		/* Fix for errata 23, can't cross 64kB boundary */
L
Linus Torvalds 已提交
4588 4589
		if (!e1000_check_64k_bound(adapter, skb->data, bufsz)) {
			struct sk_buff *oldskb = skb;
4590 4591 4592
			DPRINTK(RX_ERR, ERR, "skb align check failed: %u bytes "
					     "at %p\n", bufsz, skb->data);
			/* Try again, without freeing the previous */
4593
			skb = netdev_alloc_skb(netdev, bufsz);
4594
			/* Failed allocation, critical failure */
L
Linus Torvalds 已提交
4595 4596 4597 4598
			if (!skb) {
				dev_kfree_skb(oldskb);
				break;
			}
4599

L
Linus Torvalds 已提交
4600 4601 4602 4603 4604 4605
			if (!e1000_check_64k_bound(adapter, skb->data, bufsz)) {
				/* give up */
				dev_kfree_skb(skb);
				dev_kfree_skb(oldskb);
				break; /* while !buffer_info->skb */
			}
4606 4607 4608

			/* Use new allocation */
			dev_kfree_skb(oldskb);
L
Linus Torvalds 已提交
4609 4610 4611 4612 4613 4614 4615 4616 4617
		}
		/* Make buffer alignment 2 beyond a 16 byte boundary
		 * this will result in a 16 byte aligned IP header after
		 * the 14 byte MAC header is removed
		 */
		skb_reserve(skb, NET_IP_ALIGN);

		buffer_info->skb = skb;
		buffer_info->length = adapter->rx_buffer_len;
4618
map_skb:
L
Linus Torvalds 已提交
4619 4620 4621 4622 4623
		buffer_info->dma = pci_map_single(pdev,
						  skb->data,
						  adapter->rx_buffer_len,
						  PCI_DMA_FROMDEVICE);

4624 4625 4626 4627 4628 4629 4630 4631
		/* Fix for errata 23, can't cross 64kB boundary */
		if (!e1000_check_64k_bound(adapter,
					(void *)(unsigned long)buffer_info->dma,
					adapter->rx_buffer_len)) {
			DPRINTK(RX_ERR, ERR,
				"dma align check failed: %u bytes at %p\n",
				adapter->rx_buffer_len,
				(void *)(unsigned long)buffer_info->dma);
L
Linus Torvalds 已提交
4632 4633 4634
			dev_kfree_skb(skb);
			buffer_info->skb = NULL;

4635
			pci_unmap_single(pdev, buffer_info->dma,
L
Linus Torvalds 已提交
4636 4637 4638 4639 4640 4641 4642 4643
					 adapter->rx_buffer_len,
					 PCI_DMA_FROMDEVICE);

			break; /* while !buffer_info->skb */
		}
		rx_desc = E1000_RX_DESC(*rx_ring, i);
		rx_desc->buffer_addr = cpu_to_le64(buffer_info->dma);

J
Jesse Brandeburg 已提交
4644 4645
		if (unlikely(++i == rx_ring->count))
			i = 0;
L
Linus Torvalds 已提交
4646 4647 4648
		buffer_info = &rx_ring->buffer_info[i];
	}

4649 4650 4651 4652 4653 4654 4655 4656 4657 4658 4659 4660
	if (likely(rx_ring->next_to_use != i)) {
		rx_ring->next_to_use = i;
		if (unlikely(i-- == 0))
			i = (rx_ring->count - 1);

		/* 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();
		writel(i, adapter->hw.hw_addr + rx_ring->rdt);
	}
L
Linus Torvalds 已提交
4661 4662
}

4663 4664 4665 4666 4667 4668
/**
 * e1000_alloc_rx_buffers_ps - Replace used receive buffers; packet split
 * @adapter: address of board private structure
 **/

static void
4669
e1000_alloc_rx_buffers_ps(struct e1000_adapter *adapter,
4670 4671
                          struct e1000_rx_ring *rx_ring,
			  int cleaned_count)
4672 4673 4674 4675 4676 4677 4678 4679 4680 4681 4682 4683 4684 4685 4686
{
	struct net_device *netdev = adapter->netdev;
	struct pci_dev *pdev = adapter->pdev;
	union e1000_rx_desc_packet_split *rx_desc;
	struct e1000_buffer *buffer_info;
	struct e1000_ps_page *ps_page;
	struct e1000_ps_page_dma *ps_page_dma;
	struct sk_buff *skb;
	unsigned int i, j;

	i = rx_ring->next_to_use;
	buffer_info = &rx_ring->buffer_info[i];
	ps_page = &rx_ring->ps_page[i];
	ps_page_dma = &rx_ring->ps_page_dma[i];

4687
	while (cleaned_count--) {
4688 4689
		rx_desc = E1000_RX_DESC_PS(*rx_ring, i);

J
Jesse Brandeburg 已提交
4690
		for (j = 0; j < PS_PAGE_BUFFERS; j++) {
4691 4692 4693 4694
			if (j < adapter->rx_ps_pages) {
				if (likely(!ps_page->ps_page[j])) {
					ps_page->ps_page[j] =
						alloc_page(GFP_ATOMIC);
4695 4696
					if (unlikely(!ps_page->ps_page[j])) {
						adapter->alloc_rx_buff_failed++;
4697
						goto no_buffers;
4698
					}
4699 4700 4701 4702 4703 4704 4705
					ps_page_dma->ps_page_dma[j] =
						pci_map_page(pdev,
							    ps_page->ps_page[j],
							    0, PAGE_SIZE,
							    PCI_DMA_FROMDEVICE);
				}
				/* Refresh the desc even if buffer_addrs didn't
J
Jesse Brandeburg 已提交
4706
				 * change because each write-back erases
4707 4708 4709 4710 4711
				 * this info.
				 */
				rx_desc->read.buffer_addr[j+1] =
				     cpu_to_le64(ps_page_dma->ps_page_dma[j]);
			} else
A
Al Viro 已提交
4712
				rx_desc->read.buffer_addr[j+1] = ~cpu_to_le64(0);
4713 4714
		}

4715
		skb = netdev_alloc_skb(netdev,
4716
		                       adapter->rx_ps_bsize0 + NET_IP_ALIGN);
4717

4718 4719
		if (unlikely(!skb)) {
			adapter->alloc_rx_buff_failed++;
4720
			break;
4721
		}
4722 4723 4724 4725 4726 4727 4728 4729 4730 4731 4732 4733 4734 4735 4736

		/* Make buffer alignment 2 beyond a 16 byte boundary
		 * this will result in a 16 byte aligned IP header after
		 * the 14 byte MAC header is removed
		 */
		skb_reserve(skb, NET_IP_ALIGN);

		buffer_info->skb = skb;
		buffer_info->length = adapter->rx_ps_bsize0;
		buffer_info->dma = pci_map_single(pdev, skb->data,
						  adapter->rx_ps_bsize0,
						  PCI_DMA_FROMDEVICE);

		rx_desc->read.buffer_addr[0] = cpu_to_le64(buffer_info->dma);

J
Jesse Brandeburg 已提交
4737
		if (unlikely(++i == rx_ring->count)) i = 0;
4738 4739 4740 4741 4742 4743
		buffer_info = &rx_ring->buffer_info[i];
		ps_page = &rx_ring->ps_page[i];
		ps_page_dma = &rx_ring->ps_page_dma[i];
	}

no_buffers:
4744 4745 4746 4747 4748 4749 4750 4751 4752 4753 4754 4755 4756 4757 4758
	if (likely(rx_ring->next_to_use != i)) {
		rx_ring->next_to_use = i;
		if (unlikely(i-- == 0)) i = (rx_ring->count - 1);

		/* 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();
		/* Hardware increments by 16 bytes, but packet split
		 * descriptors are 32 bytes...so we increment tail
		 * twice as much.
		 */
		writel(i<<1, adapter->hw.hw_addr + rx_ring->rdt);
	}
4759 4760
}

L
Linus Torvalds 已提交
4761 4762 4763 4764 4765 4766 4767 4768 4769 4770 4771
/**
 * e1000_smartspeed - Workaround for SmartSpeed on 82541 and 82547 controllers.
 * @adapter:
 **/

static void
e1000_smartspeed(struct e1000_adapter *adapter)
{
	uint16_t phy_status;
	uint16_t phy_ctrl;

J
Jesse Brandeburg 已提交
4772
	if ((adapter->hw.phy_type != e1000_phy_igp) || !adapter->hw.autoneg ||
L
Linus Torvalds 已提交
4773 4774 4775
	   !(adapter->hw.autoneg_advertised & ADVERTISE_1000_FULL))
		return;

J
Jesse Brandeburg 已提交
4776
	if (adapter->smartspeed == 0) {
L
Linus Torvalds 已提交
4777 4778 4779
		/* If Master/Slave config fault is asserted twice,
		 * we assume back-to-back */
		e1000_read_phy_reg(&adapter->hw, PHY_1000T_STATUS, &phy_status);
J
Jesse Brandeburg 已提交
4780
		if (!(phy_status & SR_1000T_MS_CONFIG_FAULT)) return;
L
Linus Torvalds 已提交
4781
		e1000_read_phy_reg(&adapter->hw, PHY_1000T_STATUS, &phy_status);
J
Jesse Brandeburg 已提交
4782
		if (!(phy_status & SR_1000T_MS_CONFIG_FAULT)) return;
L
Linus Torvalds 已提交
4783
		e1000_read_phy_reg(&adapter->hw, PHY_1000T_CTRL, &phy_ctrl);
J
Jesse Brandeburg 已提交
4784
		if (phy_ctrl & CR_1000T_MS_ENABLE) {
L
Linus Torvalds 已提交
4785 4786 4787 4788
			phy_ctrl &= ~CR_1000T_MS_ENABLE;
			e1000_write_phy_reg(&adapter->hw, PHY_1000T_CTRL,
					    phy_ctrl);
			adapter->smartspeed++;
J
Jesse Brandeburg 已提交
4789
			if (!e1000_phy_setup_autoneg(&adapter->hw) &&
L
Linus Torvalds 已提交
4790 4791 4792 4793 4794 4795 4796 4797 4798
			   !e1000_read_phy_reg(&adapter->hw, PHY_CTRL,
				   	       &phy_ctrl)) {
				phy_ctrl |= (MII_CR_AUTO_NEG_EN |
					     MII_CR_RESTART_AUTO_NEG);
				e1000_write_phy_reg(&adapter->hw, PHY_CTRL,
						    phy_ctrl);
			}
		}
		return;
J
Jesse Brandeburg 已提交
4799
	} else if (adapter->smartspeed == E1000_SMARTSPEED_DOWNSHIFT) {
L
Linus Torvalds 已提交
4800 4801 4802 4803
		/* If still no link, perhaps using 2/3 pair cable */
		e1000_read_phy_reg(&adapter->hw, PHY_1000T_CTRL, &phy_ctrl);
		phy_ctrl |= CR_1000T_MS_ENABLE;
		e1000_write_phy_reg(&adapter->hw, PHY_1000T_CTRL, phy_ctrl);
J
Jesse Brandeburg 已提交
4804
		if (!e1000_phy_setup_autoneg(&adapter->hw) &&
L
Linus Torvalds 已提交
4805 4806 4807 4808 4809 4810 4811
		   !e1000_read_phy_reg(&adapter->hw, PHY_CTRL, &phy_ctrl)) {
			phy_ctrl |= (MII_CR_AUTO_NEG_EN |
				     MII_CR_RESTART_AUTO_NEG);
			e1000_write_phy_reg(&adapter->hw, PHY_CTRL, phy_ctrl);
		}
	}
	/* Restart process after E1000_SMARTSPEED_MAX iterations */
J
Jesse Brandeburg 已提交
4812
	if (adapter->smartspeed++ == E1000_SMARTSPEED_MAX)
L
Linus Torvalds 已提交
4813 4814 4815 4816 4817 4818 4819 4820 4821 4822 4823 4824 4825 4826 4827 4828 4829 4830 4831 4832 4833 4834 4835 4836 4837 4838 4839 4840 4841 4842 4843 4844 4845
		adapter->smartspeed = 0;
}

/**
 * e1000_ioctl -
 * @netdev:
 * @ifreq:
 * @cmd:
 **/

static int
e1000_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd)
{
	switch (cmd) {
	case SIOCGMIIPHY:
	case SIOCGMIIREG:
	case SIOCSMIIREG:
		return e1000_mii_ioctl(netdev, ifr, cmd);
	default:
		return -EOPNOTSUPP;
	}
}

/**
 * e1000_mii_ioctl -
 * @netdev:
 * @ifreq:
 * @cmd:
 **/

static int
e1000_mii_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd)
{
4846
	struct e1000_adapter *adapter = netdev_priv(netdev);
L
Linus Torvalds 已提交
4847 4848 4849 4850
	struct mii_ioctl_data *data = if_mii(ifr);
	int retval;
	uint16_t mii_reg;
	uint16_t spddplx;
4851
	unsigned long flags;
L
Linus Torvalds 已提交
4852

J
Jesse Brandeburg 已提交
4853
	if (adapter->hw.media_type != e1000_media_type_copper)
L
Linus Torvalds 已提交
4854 4855 4856 4857 4858 4859 4860
		return -EOPNOTSUPP;

	switch (cmd) {
	case SIOCGMIIPHY:
		data->phy_id = adapter->hw.phy_addr;
		break;
	case SIOCGMIIREG:
J
Jesse Brandeburg 已提交
4861
		if (!capable(CAP_NET_ADMIN))
L
Linus Torvalds 已提交
4862
			return -EPERM;
4863
		spin_lock_irqsave(&adapter->stats_lock, flags);
J
Jesse Brandeburg 已提交
4864
		if (e1000_read_phy_reg(&adapter->hw, data->reg_num & 0x1F,
4865 4866
				   &data->val_out)) {
			spin_unlock_irqrestore(&adapter->stats_lock, flags);
L
Linus Torvalds 已提交
4867
			return -EIO;
4868 4869
		}
		spin_unlock_irqrestore(&adapter->stats_lock, flags);
L
Linus Torvalds 已提交
4870 4871
		break;
	case SIOCSMIIREG:
J
Jesse Brandeburg 已提交
4872
		if (!capable(CAP_NET_ADMIN))
L
Linus Torvalds 已提交
4873
			return -EPERM;
J
Jesse Brandeburg 已提交
4874
		if (data->reg_num & ~(0x1F))
L
Linus Torvalds 已提交
4875 4876
			return -EFAULT;
		mii_reg = data->val_in;
4877
		spin_lock_irqsave(&adapter->stats_lock, flags);
J
Jesse Brandeburg 已提交
4878
		if (e1000_write_phy_reg(&adapter->hw, data->reg_num,
4879 4880
					mii_reg)) {
			spin_unlock_irqrestore(&adapter->stats_lock, flags);
L
Linus Torvalds 已提交
4881
			return -EIO;
4882
		}
4883
		spin_unlock_irqrestore(&adapter->stats_lock, flags);
4884
		if (adapter->hw.media_type == e1000_media_type_copper) {
L
Linus Torvalds 已提交
4885 4886
			switch (data->reg_num) {
			case PHY_CTRL:
J
Jesse Brandeburg 已提交
4887
				if (mii_reg & MII_CR_POWER_DOWN)
L
Linus Torvalds 已提交
4888
					break;
J
Jesse Brandeburg 已提交
4889
				if (mii_reg & MII_CR_AUTO_NEG_EN) {
L
Linus Torvalds 已提交
4890 4891 4892 4893 4894 4895 4896 4897 4898 4899
					adapter->hw.autoneg = 1;
					adapter->hw.autoneg_advertised = 0x2F;
				} else {
					if (mii_reg & 0x40)
						spddplx = SPEED_1000;
					else if (mii_reg & 0x2000)
						spddplx = SPEED_100;
					else
						spddplx = SPEED_10;
					spddplx += (mii_reg & 0x100)
4900 4901
						   ? DUPLEX_FULL :
						   DUPLEX_HALF;
L
Linus Torvalds 已提交
4902 4903
					retval = e1000_set_spd_dplx(adapter,
								    spddplx);
4904
					if (retval)
L
Linus Torvalds 已提交
4905 4906
						return retval;
				}
4907 4908 4909
				if (netif_running(adapter->netdev))
					e1000_reinit_locked(adapter);
				else
L
Linus Torvalds 已提交
4910 4911 4912 4913
					e1000_reset(adapter);
				break;
			case M88E1000_PHY_SPEC_CTRL:
			case M88E1000_EXT_PHY_SPEC_CTRL:
4914
				if (e1000_phy_reset(&adapter->hw))
L
Linus Torvalds 已提交
4915 4916 4917 4918 4919 4920
					return -EIO;
				break;
			}
		} else {
			switch (data->reg_num) {
			case PHY_CTRL:
J
Jesse Brandeburg 已提交
4921
				if (mii_reg & MII_CR_POWER_DOWN)
L
Linus Torvalds 已提交
4922
					break;
4923 4924 4925
				if (netif_running(adapter->netdev))
					e1000_reinit_locked(adapter);
				else
L
Linus Torvalds 已提交
4926 4927 4928 4929 4930 4931 4932 4933 4934 4935 4936 4937 4938 4939 4940
					e1000_reset(adapter);
				break;
			}
		}
		break;
	default:
		return -EOPNOTSUPP;
	}
	return E1000_SUCCESS;
}

void
e1000_pci_set_mwi(struct e1000_hw *hw)
{
	struct e1000_adapter *adapter = hw->back;
4941
	int ret_val = pci_set_mwi(adapter->pdev);
L
Linus Torvalds 已提交
4942

J
Jesse Brandeburg 已提交
4943
	if (ret_val)
4944
		DPRINTK(PROBE, ERR, "Error in setting MWI\n");
L
Linus Torvalds 已提交
4945 4946 4947 4948 4949 4950 4951 4952 4953 4954
}

void
e1000_pci_clear_mwi(struct e1000_hw *hw)
{
	struct e1000_adapter *adapter = hw->back;

	pci_clear_mwi(adapter->pdev);
}

4955 4956 4957 4958 4959 4960 4961 4962 4963 4964 4965 4966 4967 4968
int
e1000_pcix_get_mmrbc(struct e1000_hw *hw)
{
	struct e1000_adapter *adapter = hw->back;
	return pcix_get_mmrbc(adapter->pdev);
}

void
e1000_pcix_set_mmrbc(struct e1000_hw *hw, int mmrbc)
{
	struct e1000_adapter *adapter = hw->back;
	pcix_set_mmrbc(adapter->pdev, mmrbc);
}

4969 4970 4971 4972 4973 4974 4975 4976 4977 4978 4979 4980 4981 4982 4983
int32_t
e1000_read_pcie_cap_reg(struct e1000_hw *hw, uint32_t reg, uint16_t *value)
{
    struct e1000_adapter *adapter = hw->back;
    uint16_t cap_offset;

    cap_offset = pci_find_capability(adapter->pdev, PCI_CAP_ID_EXP);
    if (!cap_offset)
        return -E1000_ERR_CONFIG;

    pci_read_config_word(adapter->pdev, cap_offset + reg, value);

    return E1000_SUCCESS;
}

L
Linus Torvalds 已提交
4984 4985 4986 4987 4988 4989 4990 4991 4992
void
e1000_io_write(struct e1000_hw *hw, unsigned long port, uint32_t value)
{
	outl(value, port);
}

static void
e1000_vlan_rx_register(struct net_device *netdev, struct vlan_group *grp)
{
4993
	struct e1000_adapter *adapter = netdev_priv(netdev);
L
Linus Torvalds 已提交
4994 4995 4996 4997 4998
	uint32_t ctrl, rctl;

	e1000_irq_disable(adapter);
	adapter->vlgrp = grp;

J
Jesse Brandeburg 已提交
4999
	if (grp) {
L
Linus Torvalds 已提交
5000 5001 5002 5003 5004
		/* enable VLAN tag insert/strip */
		ctrl = E1000_READ_REG(&adapter->hw, CTRL);
		ctrl |= E1000_CTRL_VME;
		E1000_WRITE_REG(&adapter->hw, CTRL, ctrl);

5005
		if (adapter->hw.mac_type != e1000_ich8lan) {
5006 5007 5008 5009 5010 5011
			/* enable VLAN receive filtering */
			rctl = E1000_READ_REG(&adapter->hw, RCTL);
			rctl |= E1000_RCTL_VFE;
			rctl &= ~E1000_RCTL_CFIEN;
			E1000_WRITE_REG(&adapter->hw, RCTL, rctl);
			e1000_update_mng_vlan(adapter);
5012
		}
L
Linus Torvalds 已提交
5013 5014 5015 5016 5017 5018
	} else {
		/* disable VLAN tag insert/strip */
		ctrl = E1000_READ_REG(&adapter->hw, CTRL);
		ctrl &= ~E1000_CTRL_VME;
		E1000_WRITE_REG(&adapter->hw, CTRL, ctrl);

5019
		if (adapter->hw.mac_type != e1000_ich8lan) {
5020 5021 5022 5023 5024 5025 5026 5027 5028 5029
			/* disable VLAN filtering */
			rctl = E1000_READ_REG(&adapter->hw, RCTL);
			rctl &= ~E1000_RCTL_VFE;
			E1000_WRITE_REG(&adapter->hw, RCTL, rctl);
			if (adapter->mng_vlan_id !=
			    (uint16_t)E1000_MNG_VLAN_NONE) {
				e1000_vlan_rx_kill_vid(netdev,
				                       adapter->mng_vlan_id);
				adapter->mng_vlan_id = E1000_MNG_VLAN_NONE;
			}
5030
		}
L
Linus Torvalds 已提交
5031 5032 5033 5034 5035 5036 5037 5038
	}

	e1000_irq_enable(adapter);
}

static void
e1000_vlan_rx_add_vid(struct net_device *netdev, uint16_t vid)
{
5039
	struct e1000_adapter *adapter = netdev_priv(netdev);
L
Linus Torvalds 已提交
5040
	uint32_t vfta, index;
J
Jesse Brandeburg 已提交
5041 5042 5043 5044

	if ((adapter->hw.mng_cookie.status &
	     E1000_MNG_DHCP_COOKIE_STATUS_VLAN_SUPPORT) &&
	    (vid == adapter->mng_vlan_id))
5045
		return;
L
Linus Torvalds 已提交
5046 5047 5048 5049 5050 5051 5052 5053 5054 5055
	/* add VID to filter table */
	index = (vid >> 5) & 0x7F;
	vfta = E1000_READ_REG_ARRAY(&adapter->hw, VFTA, index);
	vfta |= (1 << (vid & 0x1F));
	e1000_write_vfta(&adapter->hw, index, vfta);
}

static void
e1000_vlan_rx_kill_vid(struct net_device *netdev, uint16_t vid)
{
5056
	struct e1000_adapter *adapter = netdev_priv(netdev);
L
Linus Torvalds 已提交
5057 5058 5059
	uint32_t vfta, index;

	e1000_irq_disable(adapter);
D
Dan Aloni 已提交
5060
	vlan_group_set_device(adapter->vlgrp, vid, NULL);
L
Linus Torvalds 已提交
5061 5062
	e1000_irq_enable(adapter);

J
Jesse Brandeburg 已提交
5063 5064
	if ((adapter->hw.mng_cookie.status &
	     E1000_MNG_DHCP_COOKIE_STATUS_VLAN_SUPPORT) &&
J
Jeff Kirsher 已提交
5065 5066 5067
	    (vid == adapter->mng_vlan_id)) {
		/* release control to f/w */
		e1000_release_hw_control(adapter);
5068
		return;
J
Jeff Kirsher 已提交
5069 5070
	}

L
Linus Torvalds 已提交
5071 5072 5073 5074 5075 5076 5077 5078 5079 5080 5081 5082
	/* remove VID from filter table */
	index = (vid >> 5) & 0x7F;
	vfta = E1000_READ_REG_ARRAY(&adapter->hw, VFTA, index);
	vfta &= ~(1 << (vid & 0x1F));
	e1000_write_vfta(&adapter->hw, index, vfta);
}

static void
e1000_restore_vlan(struct e1000_adapter *adapter)
{
	e1000_vlan_rx_register(adapter->netdev, adapter->vlgrp);

J
Jesse Brandeburg 已提交
5083
	if (adapter->vlgrp) {
L
Linus Torvalds 已提交
5084
		uint16_t vid;
J
Jesse Brandeburg 已提交
5085
		for (vid = 0; vid < VLAN_GROUP_ARRAY_LEN; vid++) {
D
Dan Aloni 已提交
5086
			if (!vlan_group_get_device(adapter->vlgrp, vid))
L
Linus Torvalds 已提交
5087 5088 5089 5090 5091 5092 5093 5094 5095 5096 5097
				continue;
			e1000_vlan_rx_add_vid(adapter->netdev, vid);
		}
	}
}

int
e1000_set_spd_dplx(struct e1000_adapter *adapter, uint16_t spddplx)
{
	adapter->hw.autoneg = 0;

5098
	/* Fiber NICs only allow 1000 gbps Full duplex */
J
Jesse Brandeburg 已提交
5099
	if ((adapter->hw.media_type == e1000_media_type_fiber) &&
5100 5101 5102 5103 5104
		spddplx != (SPEED_1000 + DUPLEX_FULL)) {
		DPRINTK(PROBE, ERR, "Unsupported Speed/Duplex configuration\n");
		return -EINVAL;
	}

J
Jesse Brandeburg 已提交
5105
	switch (spddplx) {
L
Linus Torvalds 已提交
5106 5107 5108 5109 5110 5111 5112 5113 5114 5115 5116 5117 5118 5119 5120 5121 5122 5123
	case SPEED_10 + DUPLEX_HALF:
		adapter->hw.forced_speed_duplex = e1000_10_half;
		break;
	case SPEED_10 + DUPLEX_FULL:
		adapter->hw.forced_speed_duplex = e1000_10_full;
		break;
	case SPEED_100 + DUPLEX_HALF:
		adapter->hw.forced_speed_duplex = e1000_100_half;
		break;
	case SPEED_100 + DUPLEX_FULL:
		adapter->hw.forced_speed_duplex = e1000_100_full;
		break;
	case SPEED_1000 + DUPLEX_FULL:
		adapter->hw.autoneg = 1;
		adapter->hw.autoneg_advertised = ADVERTISE_1000_FULL;
		break;
	case SPEED_1000 + DUPLEX_HALF: /* not supported */
	default:
5124
		DPRINTK(PROBE, ERR, "Unsupported Speed/Duplex configuration\n");
L
Linus Torvalds 已提交
5125 5126 5127 5128 5129 5130
		return -EINVAL;
	}
	return 0;
}

static int
5131
e1000_suspend(struct pci_dev *pdev, pm_message_t state)
L
Linus Torvalds 已提交
5132 5133
{
	struct net_device *netdev = pci_get_drvdata(pdev);
5134
	struct e1000_adapter *adapter = netdev_priv(netdev);
5135
	uint32_t ctrl, ctrl_ext, rctl, status;
L
Linus Torvalds 已提交
5136
	uint32_t wufc = adapter->wol;
A
Auke Kok 已提交
5137
#ifdef CONFIG_PM
5138
	int retval = 0;
A
Auke Kok 已提交
5139
#endif
L
Linus Torvalds 已提交
5140 5141 5142

	netif_device_detach(netdev);

5143 5144
	if (netif_running(netdev)) {
		WARN_ON(test_bit(__E1000_RESETTING, &adapter->flags));
L
Linus Torvalds 已提交
5145
		e1000_down(adapter);
5146
	}
L
Linus Torvalds 已提交
5147

5148
#ifdef CONFIG_PM
5149
	retval = pci_save_state(pdev);
5150 5151 5152 5153
	if (retval)
		return retval;
#endif

L
Linus Torvalds 已提交
5154
	status = E1000_READ_REG(&adapter->hw, STATUS);
J
Jesse Brandeburg 已提交
5155
	if (status & E1000_STATUS_LU)
L
Linus Torvalds 已提交
5156 5157
		wufc &= ~E1000_WUFC_LNKC;

J
Jesse Brandeburg 已提交
5158
	if (wufc) {
L
Linus Torvalds 已提交
5159
		e1000_setup_rctl(adapter);
5160
		e1000_set_rx_mode(netdev);
L
Linus Torvalds 已提交
5161 5162

		/* turn on all-multi mode if wake on multicast is enabled */
5163
		if (wufc & E1000_WUFC_MC) {
L
Linus Torvalds 已提交
5164 5165 5166 5167 5168
			rctl = E1000_READ_REG(&adapter->hw, RCTL);
			rctl |= E1000_RCTL_MPE;
			E1000_WRITE_REG(&adapter->hw, RCTL, rctl);
		}

J
Jesse Brandeburg 已提交
5169
		if (adapter->hw.mac_type >= e1000_82540) {
L
Linus Torvalds 已提交
5170 5171 5172 5173 5174 5175 5176 5177 5178 5179
			ctrl = E1000_READ_REG(&adapter->hw, CTRL);
			/* advertise wake from D3Cold */
			#define E1000_CTRL_ADVD3WUC 0x00100000
			/* phy power management enable */
			#define E1000_CTRL_EN_PHY_PWR_MGMT 0x00200000
			ctrl |= E1000_CTRL_ADVD3WUC |
				E1000_CTRL_EN_PHY_PWR_MGMT;
			E1000_WRITE_REG(&adapter->hw, CTRL, ctrl);
		}

J
Jesse Brandeburg 已提交
5180
		if (adapter->hw.media_type == e1000_media_type_fiber ||
L
Linus Torvalds 已提交
5181 5182 5183 5184 5185 5186 5187
		   adapter->hw.media_type == e1000_media_type_internal_serdes) {
			/* keep the laser running in D3 */
			ctrl_ext = E1000_READ_REG(&adapter->hw, CTRL_EXT);
			ctrl_ext |= E1000_CTRL_EXT_SDP7_DATA;
			E1000_WRITE_REG(&adapter->hw, CTRL_EXT, ctrl_ext);
		}

5188 5189 5190
		/* Allow time for pending master requests to run */
		e1000_disable_pciex_master(&adapter->hw);

L
Linus Torvalds 已提交
5191 5192
		E1000_WRITE_REG(&adapter->hw, WUC, E1000_WUC_PME_EN);
		E1000_WRITE_REG(&adapter->hw, WUFC, wufc);
5193 5194
		pci_enable_wake(pdev, PCI_D3hot, 1);
		pci_enable_wake(pdev, PCI_D3cold, 1);
L
Linus Torvalds 已提交
5195 5196 5197
	} else {
		E1000_WRITE_REG(&adapter->hw, WUC, 0);
		E1000_WRITE_REG(&adapter->hw, WUFC, 0);
5198 5199
		pci_enable_wake(pdev, PCI_D3hot, 0);
		pci_enable_wake(pdev, PCI_D3cold, 0);
L
Linus Torvalds 已提交
5200 5201
	}

5202 5203 5204 5205 5206 5207
	e1000_release_manageability(adapter);

	/* make sure adapter isn't asleep if manageability is enabled */
	if (adapter->en_mng_pt) {
		pci_enable_wake(pdev, PCI_D3hot, 1);
		pci_enable_wake(pdev, PCI_D3cold, 1);
L
Linus Torvalds 已提交
5208 5209
	}

5210 5211 5212
	if (adapter->hw.phy_type == e1000_phy_igp_3)
		e1000_phy_powerdown_workaround(&adapter->hw);

5213 5214 5215
	if (netif_running(netdev))
		e1000_free_irq(adapter);

5216 5217 5218
	/* Release control of h/w to f/w.  If f/w is AMT enabled, this
	 * would have already happened in close and is redundant. */
	e1000_release_hw_control(adapter);
5219

L
Linus Torvalds 已提交
5220
	pci_disable_device(pdev);
5221

5222
	pci_set_power_state(pdev, pci_choose_state(pdev, state));
L
Linus Torvalds 已提交
5223 5224 5225 5226

	return 0;
}

5227
#ifdef CONFIG_PM
L
Linus Torvalds 已提交
5228 5229 5230 5231
static int
e1000_resume(struct pci_dev *pdev)
{
	struct net_device *netdev = pci_get_drvdata(pdev);
5232
	struct e1000_adapter *adapter = netdev_priv(netdev);
5233
	uint32_t err;
L
Linus Torvalds 已提交
5234

5235
	pci_set_power_state(pdev, PCI_D0);
5236
	pci_restore_state(pdev);
5237 5238 5239 5240
	if ((err = pci_enable_device(pdev))) {
		printk(KERN_ERR "e1000: Cannot enable PCI device from suspend\n");
		return err;
	}
5241
	pci_set_master(pdev);
L
Linus Torvalds 已提交
5242

5243 5244
	pci_enable_wake(pdev, PCI_D3hot, 0);
	pci_enable_wake(pdev, PCI_D3cold, 0);
L
Linus Torvalds 已提交
5245

5246 5247 5248 5249
	if (netif_running(netdev) && (err = e1000_request_irq(adapter)))
		return err;

	e1000_power_up_phy(adapter);
L
Linus Torvalds 已提交
5250 5251 5252
	e1000_reset(adapter);
	E1000_WRITE_REG(&adapter->hw, WUS, ~0);

5253 5254
	e1000_init_manageability(adapter);

J
Jesse Brandeburg 已提交
5255
	if (netif_running(netdev))
L
Linus Torvalds 已提交
5256 5257 5258 5259
		e1000_up(adapter);

	netif_device_attach(netdev);

5260 5261 5262 5263 5264 5265 5266
	/* If the controller is 82573 and f/w is AMT, do not set
	 * DRV_LOAD until the interface is up.  For all other cases,
	 * let the f/w know that the h/w is now under the control
	 * of the driver. */
	if (adapter->hw.mac_type != e1000_82573 ||
	    !e1000_check_mng_mode(&adapter->hw))
		e1000_get_hw_control(adapter);
5267

L
Linus Torvalds 已提交
5268 5269 5270
	return 0;
}
#endif
5271 5272 5273 5274 5275 5276

static void e1000_shutdown(struct pci_dev *pdev)
{
	e1000_suspend(pdev, PMSG_SUSPEND);
}

L
Linus Torvalds 已提交
5277 5278 5279 5280 5281 5282 5283
#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
5284
e1000_netpoll(struct net_device *netdev)
L
Linus Torvalds 已提交
5285
{
5286
	struct e1000_adapter *adapter = netdev_priv(netdev);
5287

L
Linus Torvalds 已提交
5288
	disable_irq(adapter->pdev->irq);
5289
	e1000_intr(adapter->pdev->irq, netdev);
A
Andrew Morton 已提交
5290
	e1000_clean_tx_irq(adapter, adapter->tx_ring);
J
Jeff Kirsher 已提交
5291 5292 5293
#ifndef CONFIG_E1000_NAPI
	adapter->clean_rx(adapter, adapter->rx_ring);
#endif
L
Linus Torvalds 已提交
5294 5295 5296 5297
	enable_irq(adapter->pdev->irq);
}
#endif

A
Auke Kok 已提交
5298 5299 5300 5301 5302 5303 5304 5305 5306 5307 5308 5309 5310 5311 5312 5313 5314
/**
 * e1000_io_error_detected - called when PCI error is detected
 * @pdev: Pointer to PCI device
 * @state: The current pci conneection state
 *
 * This function is called after a PCI bus error affecting
 * this device has been detected.
 */
static pci_ers_result_t e1000_io_error_detected(struct pci_dev *pdev, pci_channel_state_t state)
{
	struct net_device *netdev = pci_get_drvdata(pdev);
	struct e1000_adapter *adapter = netdev->priv;

	netif_device_detach(netdev);

	if (netif_running(netdev))
		e1000_down(adapter);
5315
	pci_disable_device(pdev);
A
Auke Kok 已提交
5316 5317 5318 5319 5320 5321 5322 5323 5324 5325 5326 5327 5328 5329 5330 5331 5332 5333 5334 5335 5336 5337 5338

	/* Request a slot slot reset. */
	return PCI_ERS_RESULT_NEED_RESET;
}

/**
 * e1000_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. Implementation
 * resembles the first-half of the e1000_resume routine.
 */
static pci_ers_result_t e1000_io_slot_reset(struct pci_dev *pdev)
{
	struct net_device *netdev = pci_get_drvdata(pdev);
	struct e1000_adapter *adapter = netdev->priv;

	if (pci_enable_device(pdev)) {
		printk(KERN_ERR "e1000: Cannot re-enable PCI device after reset.\n");
		return PCI_ERS_RESULT_DISCONNECT;
	}
	pci_set_master(pdev);

5339 5340
	pci_enable_wake(pdev, PCI_D3hot, 0);
	pci_enable_wake(pdev, PCI_D3cold, 0);
A
Auke Kok 已提交
5341 5342 5343 5344 5345 5346 5347 5348 5349 5350 5351 5352 5353 5354 5355 5356 5357 5358 5359

	e1000_reset(adapter);
	E1000_WRITE_REG(&adapter->hw, WUS, ~0);

	return PCI_ERS_RESULT_RECOVERED;
}

/**
 * e1000_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. Implementation resembles the
 * second-half of the e1000_resume routine.
 */
static void e1000_io_resume(struct pci_dev *pdev)
{
	struct net_device *netdev = pci_get_drvdata(pdev);
	struct e1000_adapter *adapter = netdev->priv;
5360 5361

	e1000_init_manageability(adapter);
A
Auke Kok 已提交
5362 5363 5364 5365 5366 5367 5368 5369 5370 5371

	if (netif_running(netdev)) {
		if (e1000_up(adapter)) {
			printk("e1000: can't bring device back up after reset\n");
			return;
		}
	}

	netif_device_attach(netdev);

5372 5373 5374 5375 5376 5377 5378
	/* If the controller is 82573 and f/w is AMT, do not set
	 * DRV_LOAD until the interface is up.  For all other cases,
	 * let the f/w know that the h/w is now under the control
	 * of the driver. */
	if (adapter->hw.mac_type != e1000_82573 ||
	    !e1000_check_mng_mode(&adapter->hw))
		e1000_get_hw_control(adapter);
A
Auke Kok 已提交
5379 5380 5381

}

L
Linus Torvalds 已提交
5382
/* e1000_main.c */