sys/dev/netif/em/if_em.c

   1 /*
   2  * Copyright (c) 2004 Joerg Sonnenberger <joerg@bec.de>.  All rights reserved.
   3  *
   4  * Copyright (c) 2001-2008, Intel Corporation
   5  * All rights reserved.
   6  *
   7  * Redistribution and use in source and binary forms, with or without
   8  * modification, are permitted provided that the following conditions are met:
   9  *
  10  *  1. Redistributions of source code must retain the above copyright notice,
  11  *     this list of conditions and the following disclaimer.
  12  *
  13  *  2. Redistributions in binary form must reproduce the above copyright
  14  *     notice, this list of conditions and the following disclaimer in the
  15  *     documentation and/or other materials provided with the distribution.
  16  *
  17  *  3. Neither the name of the Intel Corporation nor the names of its
  18  *     contributors may be used to endorse or promote products derived from
  19  *     this software without specific prior written permission.
  20  *
  21  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
  22  * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  23  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  24  * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
  25  * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
  26  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
  27  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
  28  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
  29  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
  30  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
  31  * POSSIBILITY OF SUCH DAMAGE.
  32  *
  33  *
  34  * Copyright (c) 2005 The DragonFly Project.  All rights reserved.
  35  *
  36  * This code is derived from software contributed to The DragonFly Project
  37  * by Matthew Dillon <dillon@backplane.com>
  38  *
  39  * Redistribution and use in source and binary forms, with or without
  40  * modification, are permitted provided that the following conditions
  41  * are met:
  42  *
  43  * 1. Redistributions of source code must retain the above copyright
  44  *    notice, this list of conditions and the following disclaimer.
  45  * 2. Redistributions in binary form must reproduce the above copyright
  46  *    notice, this list of conditions and the following disclaimer in
  47  *    the documentation and/or other materials provided with the
  48  *    distribution.
  49  * 3. Neither the name of The DragonFly Project nor the names of its
  50  *    contributors may be used to endorse or promote products derived
  51  *    from this software without specific, prior written permission.
  52  *
  53  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
  54  * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
  55  * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
  56  * FOR A PARTICULAR PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE
  57  * COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
  58  * INCIDENTAL, SPECIAL, EXEMPLARY OR CONSEQUENTIAL DAMAGES (INCLUDING,
  59  * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
  60  * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
  61  * AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
  62  * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
  63  * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  64  * SUCH DAMAGE.
  65  *
  66  */
  67 /*
  68  * SERIALIZATION API RULES:
  69  *
  70  * - If the driver uses the same serializer for the interrupt as for the
  71  *   ifnet, most of the serialization will be done automatically for the
  72  *   driver.
  73  *
  74  * - ifmedia entry points will be serialized by the ifmedia code using the
  75  *   ifnet serializer.
  76  *
  77  * - if_* entry points except for if_input will be serialized by the IF
  78  *   and protocol layers.
  79  *
  80  * - The device driver must be sure to serialize access from timeout code
  81  *   installed by the device driver.
  82  *
  83  * - The device driver typically holds the serializer at the time it wishes
  84  *   to call if_input.
  85  *
  86  * - We must call lwkt_serialize_handler_enable() prior to enabling the
  87  *   hardware interrupt and lwkt_serialize_handler_disable() after disabling
  88  *   the hardware interrupt in order to avoid handler execution races from
  89  *   scheduled interrupt threads.
  90  *
  91  *   NOTE!  Since callers into the device driver hold the ifnet serializer,
  92  *   the device driver may be holding a serializer at the time it calls
  93  *   if_input even if it is not serializer-aware.
  94  */
  95
  96 #include "opt_polling.h"
  97
  98 #include <sys/param.h>
  99 #include <sys/bus.h>
 100 #include <sys/endian.h>
 101 #include <sys/interrupt.h>
 102 #include <sys/kernel.h>
 103 #include <sys/ktr.h>
 104 #include <sys/malloc.h>
 105 #include <sys/mbuf.h>
 106 #include <sys/proc.h>
 107 #include <sys/rman.h>
 108 #include <sys/serialize.h>
 109 #include <sys/socket.h>
 110 #include <sys/sockio.h>
 111 #include <sys/sysctl.h>
 112 #include <sys/systm.h>
 113
 114 #include <net/bpf.h>
 115 #include <net/ethernet.h>
 116 #include <net/if.h>
 117 #include <net/if_arp.h>
 118 #include <net/if_dl.h>
 119 #include <net/if_media.h>
 120 #include <net/ifq_var.h>
 121 #include <net/vlan/if_vlan_var.h>
 122 #include <net/vlan/if_vlan_ether.h>
 123
 124 #include <netinet/ip.h>
 125 #include <netinet/tcp.h>
 126 #include <netinet/udp.h>
 127
 128 #include <bus/pci/pcivar.h>
 129 #include <bus/pci/pcireg.h>
 130
 131 #include <dev/netif/ig_hal/e1000_api.h>
 132 #include <dev/netif/ig_hal/e1000_82571.h>
 133 #include <dev/netif/em/if_em.h>
 134
 135 #define EM_NAME "Intel(R) PRO/1000 Network Connection "
 136 #define EM_VER  " 7.2.4"
 137
 138 #define _EM_DEVICE(id, ret)     \
 139         { EM_VENDOR_ID, E1000_DEV_ID_##id, ret, EM_NAME #id EM_VER }
 140 #define EM_EMX_DEVICE(id)       _EM_DEVICE(id, -100)
 141 #define EM_DEVICE(id)           _EM_DEVICE(id, 0)
 142 #define EM_DEVICE_NULL  { 0, 0, 0, NULL }
 143
 144 static const struct em_vendor_info em_vendor_info_array[] = {
 145         EM_DEVICE(82540EM),
 146         EM_DEVICE(82540EM_LOM),
 147         EM_DEVICE(82540EP),
 148         EM_DEVICE(82540EP_LOM),
 149         EM_DEVICE(82540EP_LP),
 150
 151         EM_DEVICE(82541EI),
 152         EM_DEVICE(82541ER),
 153         EM_DEVICE(82541ER_LOM),
 154         EM_DEVICE(82541EI_MOBILE),
 155         EM_DEVICE(82541GI),
 156         EM_DEVICE(82541GI_LF),
 157         EM_DEVICE(82541GI_MOBILE),
 158
 159         EM_DEVICE(82542),
 160
 161         EM_DEVICE(82543GC_FIBER),
 162         EM_DEVICE(82543GC_COPPER),
 163
 164         EM_DEVICE(82544EI_COPPER),
 165         EM_DEVICE(82544EI_FIBER),
 166         EM_DEVICE(82544GC_COPPER),
 167         EM_DEVICE(82544GC_LOM),
 168
 169         EM_DEVICE(82545EM_COPPER),
 170         EM_DEVICE(82545EM_FIBER),
 171         EM_DEVICE(82545GM_COPPER),
 172         EM_DEVICE(82545GM_FIBER),
 173         EM_DEVICE(82545GM_SERDES),
 174
 175         EM_DEVICE(82546EB_COPPER),
 176         EM_DEVICE(82546EB_FIBER),
 177         EM_DEVICE(82546EB_QUAD_COPPER),
 178         EM_DEVICE(82546GB_COPPER),
 179         EM_DEVICE(82546GB_FIBER),
 180         EM_DEVICE(82546GB_SERDES),
 181         EM_DEVICE(82546GB_PCIE),
 182         EM_DEVICE(82546GB_QUAD_COPPER),
 183         EM_DEVICE(82546GB_QUAD_COPPER_KSP3),
 184
 185         EM_DEVICE(82547EI),
 186         EM_DEVICE(82547EI_MOBILE),
 187         EM_DEVICE(82547GI),
 188
 189         EM_EMX_DEVICE(82571EB_COPPER),
 190         EM_EMX_DEVICE(82571EB_FIBER),
 191         EM_EMX_DEVICE(82571EB_SERDES),
 192         EM_EMX_DEVICE(82571EB_SERDES_DUAL),
 193         EM_EMX_DEVICE(82571EB_SERDES_QUAD),
 194         EM_EMX_DEVICE(82571EB_QUAD_COPPER),
 195         EM_EMX_DEVICE(82571EB_QUAD_COPPER_BP),
 196         EM_EMX_DEVICE(82571EB_QUAD_COPPER_LP),
 197         EM_EMX_DEVICE(82571EB_QUAD_FIBER),
 198         EM_EMX_DEVICE(82571PT_QUAD_COPPER),
 199
 200         EM_EMX_DEVICE(82572EI_COPPER),
 201         EM_EMX_DEVICE(82572EI_FIBER),
 202         EM_EMX_DEVICE(82572EI_SERDES),
 203         EM_EMX_DEVICE(82572EI),
 204
 205         EM_EMX_DEVICE(82573E),
 206         EM_EMX_DEVICE(82573E_IAMT),
 207         EM_EMX_DEVICE(82573L),
 208
 209         EM_DEVICE(82583V),
 210
 211         EM_EMX_DEVICE(80003ES2LAN_COPPER_SPT),
 212         EM_EMX_DEVICE(80003ES2LAN_SERDES_SPT),
 213         EM_EMX_DEVICE(80003ES2LAN_COPPER_DPT),
 214         EM_EMX_DEVICE(80003ES2LAN_SERDES_DPT),
 215
 216         EM_DEVICE(ICH8_IGP_M_AMT),
 217         EM_DEVICE(ICH8_IGP_AMT),
 218         EM_DEVICE(ICH8_IGP_C),
 219         EM_DEVICE(ICH8_IFE),
 220         EM_DEVICE(ICH8_IFE_GT),
 221         EM_DEVICE(ICH8_IFE_G),
 222         EM_DEVICE(ICH8_IGP_M),
 223         EM_DEVICE(ICH8_82567V_3),
 224
 225         EM_DEVICE(ICH9_IGP_M_AMT),
 226         EM_DEVICE(ICH9_IGP_AMT),
 227         EM_DEVICE(ICH9_IGP_C),
 228         EM_DEVICE(ICH9_IGP_M),
 229         EM_DEVICE(ICH9_IGP_M_V),
 230         EM_DEVICE(ICH9_IFE),
 231         EM_DEVICE(ICH9_IFE_GT),
 232         EM_DEVICE(ICH9_IFE_G),
 233         EM_DEVICE(ICH9_BM),
 234
 235         EM_EMX_DEVICE(82574L),
 236         EM_EMX_DEVICE(82574LA),
 237
 238         EM_DEVICE(ICH10_R_BM_LM),
 239         EM_DEVICE(ICH10_R_BM_LF),
 240         EM_DEVICE(ICH10_R_BM_V),
 241         EM_DEVICE(ICH10_D_BM_LM),
 242         EM_DEVICE(ICH10_D_BM_LF),
 243         EM_DEVICE(ICH10_D_BM_V),
 244
 245         EM_DEVICE(PCH_M_HV_LM),
 246         EM_DEVICE(PCH_M_HV_LC),
 247         EM_DEVICE(PCH_D_HV_DM),
 248         EM_DEVICE(PCH_D_HV_DC),
 249
 250         EM_DEVICE(PCH2_LV_LM),
 251         EM_DEVICE(PCH2_LV_V),
 252
 253         /* required last entry */
 254         EM_DEVICE_NULL
 255 };
 256
 257 static int      em_probe(device_t);
 258 static int      em_attach(device_t);
 259 static int      em_detach(device_t);
 260 static int      em_shutdown(device_t);
 261 static int      em_suspend(device_t);
 262 static int      em_resume(device_t);
 263
 264 static void     em_init(void *);
 265 static void     em_stop(struct adapter *);
 266 static int      em_ioctl(struct ifnet *, u_long, caddr_t, struct ucred *);
 267 static void     em_start(struct ifnet *);
 268 #ifdef DEVICE_POLLING
 269 static void     em_poll(struct ifnet *, enum poll_cmd, int);
 270 #endif
 271 static void     em_watchdog(struct ifnet *);
 272 static void     em_media_status(struct ifnet *, struct ifmediareq *);
 273 static int      em_media_change(struct ifnet *);
 274 static void     em_timer(void *);
 275
 276 static void     em_intr(void *);
 277 static void     em_intr_mask(void *);
 278 static void     em_intr_body(struct adapter *, boolean_t);
 279 static void     em_rxeof(struct adapter *, int);
 280 static void     em_txeof(struct adapter *);
 281 static void     em_tx_collect(struct adapter *);
 282 static void     em_tx_purge(struct adapter *);
 283 static void     em_enable_intr(struct adapter *);
 284 static void     em_disable_intr(struct adapter *);
 285
 286 static int      em_dma_malloc(struct adapter *, bus_size_t,
 287                     struct em_dma_alloc *);
 288 static void     em_dma_free(struct adapter *, struct em_dma_alloc *);
 289 static void     em_init_tx_ring(struct adapter *);
 290 static int      em_init_rx_ring(struct adapter *);
 291 static int      em_create_tx_ring(struct adapter *);
 292 static int      em_create_rx_ring(struct adapter *);
 293 static void     em_destroy_tx_ring(struct adapter *, int);
 294 static void     em_destroy_rx_ring(struct adapter *, int);
 295 static int      em_newbuf(struct adapter *, int, int);
 296 static int      em_encap(struct adapter *, struct mbuf **);
 297 static void     em_rxcsum(struct adapter *, struct e1000_rx_desc *,
 298                     struct mbuf *);
 299 static int      em_txcsum(struct adapter *, struct mbuf *,
 300                     uint32_t *, uint32_t *);
 301 static int      em_tso_pullup(struct adapter *, struct mbuf **);
 302 static int      em_tso_setup(struct adapter *, struct mbuf *,
 303                     uint32_t *, uint32_t *);
 304
 305 static int      em_get_hw_info(struct adapter *);
 306 static int      em_is_valid_eaddr(const uint8_t *);
 307 static int      em_alloc_pci_res(struct adapter *);
 308 static void     em_free_pci_res(struct adapter *);
 309 static int      em_reset(struct adapter *);
 310 static void     em_setup_ifp(struct adapter *);
 311 static void     em_init_tx_unit(struct adapter *);
 312 static void     em_init_rx_unit(struct adapter *);
 313 static void     em_update_stats(struct adapter *);
 314 static void     em_set_promisc(struct adapter *);
 315 static void     em_disable_promisc(struct adapter *);
 316 static void     em_set_multi(struct adapter *);
 317 static void     em_update_link_status(struct adapter *);
 318 static void     em_smartspeed(struct adapter *);
 319 static void     em_set_itr(struct adapter *, uint32_t);
 320 static void     em_disable_aspm(struct adapter *);
 321
 322 /* Hardware workarounds */
 323 static int      em_82547_fifo_workaround(struct adapter *, int);
 324 static void     em_82547_update_fifo_head(struct adapter *, int);
 325 static int      em_82547_tx_fifo_reset(struct adapter *);
 326 static void     em_82547_move_tail(void *);
 327 static void     em_82547_move_tail_serialized(struct adapter *);
 328 static uint32_t em_82544_fill_desc(bus_addr_t, uint32_t, PDESC_ARRAY);
 329
 330 static void     em_print_debug_info(struct adapter *);
 331 static void     em_print_nvm_info(struct adapter *);
 332 static void     em_print_hw_stats(struct adapter *);
 333
 334 static int      em_sysctl_stats(SYSCTL_HANDLER_ARGS);
 335 static int      em_sysctl_debug_info(SYSCTL_HANDLER_ARGS);
 336 static int      em_sysctl_int_throttle(SYSCTL_HANDLER_ARGS);
 337 static int      em_sysctl_int_tx_nsegs(SYSCTL_HANDLER_ARGS);
 338 static void     em_add_sysctl(struct adapter *adapter);
 339
 340 /* Management and WOL Support */
 341 static void     em_get_mgmt(struct adapter *);
 342 static void     em_rel_mgmt(struct adapter *);
 343 static void     em_get_hw_control(struct adapter *);
 344 static void     em_rel_hw_control(struct adapter *);
 345 static void     em_enable_wol(device_t);
 346
 347 static device_method_t em_methods[] = {
 348         /* Device interface */
 349         DEVMETHOD(device_probe,         em_probe),
 350         DEVMETHOD(device_attach,        em_attach),
 351         DEVMETHOD(device_detach,        em_detach),
 352         DEVMETHOD(device_shutdown,      em_shutdown),
 353         DEVMETHOD(device_suspend,       em_suspend),
 354         DEVMETHOD(device_resume,        em_resume),
 355         { 0, 0 }
 356 };
 357
 358 static driver_t em_driver = {
 359         "em",
 360         em_methods,
 361         sizeof(struct adapter),
 362 };
 363
 364 static devclass_t em_devclass;
 365
 366 DECLARE_DUMMY_MODULE(if_em);
 367 MODULE_DEPEND(em, ig_hal, 1, 1, 1);
 368 DRIVER_MODULE(if_em, pci, em_driver, em_devclass, NULL, NULL);
 369
 370 /*
 371  * Tunables
 372  */
 373 static int      em_int_throttle_ceil = EM_DEFAULT_ITR;
 374 static int      em_rxd = EM_DEFAULT_RXD;
 375 static int      em_txd = EM_DEFAULT_TXD;
 376 static int      em_smart_pwr_down = 0;
 377
 378 /* Controls whether promiscuous also shows bad packets */
 379 static int      em_debug_sbp = FALSE;
 380
 381 static int      em_82573_workaround = 1;
 382 static int      em_msi_enable = 1;
 383
 384 TUNABLE_INT("hw.em.int_throttle_ceil", &em_int_throttle_ceil);
 385 TUNABLE_INT("hw.em.rxd", &em_rxd);
 386 TUNABLE_INT("hw.em.txd", &em_txd);
 387 TUNABLE_INT("hw.em.smart_pwr_down", &em_smart_pwr_down);
 388 TUNABLE_INT("hw.em.sbp", &em_debug_sbp);
 389 TUNABLE_INT("hw.em.82573_workaround", &em_82573_workaround);
 390 TUNABLE_INT("hw.em.msi.enable", &em_msi_enable);
 391
 392 /* Global used in WOL setup with multiport cards */
 393 static int      em_global_quad_port_a = 0;
 394
 395 /* Set this to one to display debug statistics */
 396 static int      em_display_debug_stats = 0;
 397
 398 #if !defined(KTR_IF_EM)
 399 #define KTR_IF_EM       KTR_ALL
 400 #endif
 401 KTR_INFO_MASTER(if_em);
 402 KTR_INFO(KTR_IF_EM, if_em, intr_beg, 0, "intr begin");
 403 KTR_INFO(KTR_IF_EM, if_em, intr_end, 1, "intr end");
 404 KTR_INFO(KTR_IF_EM, if_em, pkt_receive, 4, "rx packet");
 405 KTR_INFO(KTR_IF_EM, if_em, pkt_txqueue, 5, "tx packet");
 406 KTR_INFO(KTR_IF_EM, if_em, pkt_txclean, 6, "tx clean");
 407 #define logif(name)     KTR_LOG(if_em_ ## name)
 408
 409 static int
 410 em_probe(device_t dev)
 411 {
 412         const struct em_vendor_info *ent;
 413         uint16_t vid, did;
 414
 415         vid = pci_get_vendor(dev);
 416         did = pci_get_device(dev);
 417
 418         for (ent = em_vendor_info_array; ent->desc != NULL; ++ent) {
 419                 if (vid == ent->vendor_id && did == ent->device_id) {
 420                         device_set_desc(dev, ent->desc);
 421                         device_set_async_attach(dev, TRUE);
 422                         return (ent->ret);
 423                 }
 424         }
 425         return (ENXIO);
 426 }
 427
 428 static int
 429 em_attach(device_t dev)
 430 {
 431         struct adapter *adapter = device_get_softc(dev);
 432         struct ifnet *ifp = &adapter->arpcom.ac_if;
 433         int tsize, rsize;
 434         int error = 0;
 435         uint16_t eeprom_data, device_id, apme_mask;
 436         driver_intr_t *intr_func;
 437
 438         adapter->dev = adapter->osdep.dev = dev;
 439
 440         callout_init_mp(&adapter->timer);
 441         callout_init_mp(&adapter->tx_fifo_timer);
 442
 443         /* Determine hardware and mac info */
 444         error = em_get_hw_info(adapter);
 445         if (error) {
 446                 device_printf(dev, "Identify hardware failed\n");
 447                 goto fail;
 448         }
 449
 450         /* Setup PCI resources */
 451         error = em_alloc_pci_res(adapter);
 452         if (error) {
 453                 device_printf(dev, "Allocation of PCI resources failed\n");
 454                 goto fail;
 455         }
 456
 457         /*
 458          * For ICH8 and family we need to map the flash memory,
 459          * and this must happen after the MAC is identified.
 460          */
 461         if (adapter->hw.mac.type == e1000_ich8lan ||
 462             adapter->hw.mac.type == e1000_ich9lan ||
 463             adapter->hw.mac.type == e1000_ich10lan ||
 464             adapter->hw.mac.type == e1000_pchlan ||
 465             adapter->hw.mac.type == e1000_pch2lan) {
 466                 adapter->flash_rid = EM_BAR_FLASH;
 467
 468                 adapter->flash = bus_alloc_resource_any(dev, SYS_RES_MEMORY,
 469                                         &adapter->flash_rid, RF_ACTIVE);
 470                 if (adapter->flash == NULL) {
 471                         device_printf(dev, "Mapping of Flash failed\n");
 472                         error = ENXIO;
 473                         goto fail;
 474                 }
 475                 adapter->osdep.flash_bus_space_tag =
 476                     rman_get_bustag(adapter->flash);
 477                 adapter->osdep.flash_bus_space_handle =
 478                     rman_get_bushandle(adapter->flash);
 479
 480                 /*
 481                  * This is used in the shared code
 482                  * XXX this goof is actually not used.
 483                  */
 484                 adapter->hw.flash_address = (uint8_t *)adapter->flash;
 485         }
 486
 487         switch (adapter->hw.mac.type) {
 488         case e1000_82571:
 489         case e1000_82572:
 490                 /*
 491                  * Pullup extra 4bytes into the first data segment, see:
 492                  * 82571/82572 specification update errata #7
 493                  *
 494                  * NOTE:
 495                  * 4bytes instead of 2bytes, which are mentioned in the
 496                  * errata, are pulled; mainly to keep rest of the data
 497                  * properly aligned.
 498                  */
 499                 adapter->flags |= EM_FLAG_TSO_PULLEX;
 500                 /* FALL THROUGH */
 501
 502         case e1000_82573:
 503         case e1000_82574:
 504         case e1000_80003es2lan:
 505                 adapter->flags |= EM_FLAG_TSO;
 506                 break;
 507
 508         default:
 509                 break;
 510         }
 511
 512         /* Do Shared Code initialization */
 513         if (e1000_setup_init_funcs(&adapter->hw, TRUE)) {
 514                 device_printf(dev, "Setup of Shared code failed\n");
 515                 error = ENXIO;
 516                 goto fail;
 517         }
 518
 519         e1000_get_bus_info(&adapter->hw);
 520
 521         /*
 522          * Validate number of transmit and receive descriptors.  It
 523          * must not exceed hardware maximum, and must be multiple
 524          * of E1000_DBA_ALIGN.
 525          */
 526         if ((em_txd * sizeof(struct e1000_tx_desc)) % EM_DBA_ALIGN != 0 ||
 527             (adapter->hw.mac.type >= e1000_82544 && em_txd > EM_MAX_TXD) ||
 528             (adapter->hw.mac.type < e1000_82544 && em_txd > EM_MAX_TXD_82543) ||
 529             em_txd < EM_MIN_TXD) {
 530                 device_printf(dev, "Using %d TX descriptors instead of %d!\n",
 531                     EM_DEFAULT_TXD, em_txd);
 532                 adapter->num_tx_desc = EM_DEFAULT_TXD;
 533         } else {
 534                 adapter->num_tx_desc = em_txd;
 535         }
 536         if ((em_rxd * sizeof(struct e1000_rx_desc)) % EM_DBA_ALIGN != 0 ||
 537             (adapter->hw.mac.type >= e1000_82544 && em_rxd > EM_MAX_RXD) ||
 538             (adapter->hw.mac.type < e1000_82544 && em_rxd > EM_MAX_RXD_82543) ||
 539             em_rxd < EM_MIN_RXD) {
 540                 device_printf(dev, "Using %d RX descriptors instead of %d!\n",
 541                     EM_DEFAULT_RXD, em_rxd);
 542                 adapter->num_rx_desc = EM_DEFAULT_RXD;
 543         } else {
 544                 adapter->num_rx_desc = em_rxd;
 545         }
 546
 547         adapter->hw.mac.autoneg = DO_AUTO_NEG;
 548         adapter->hw.phy.autoneg_wait_to_complete = FALSE;
 549         adapter->hw.phy.autoneg_advertised = AUTONEG_ADV_DEFAULT;
 550         adapter->rx_buffer_len = MCLBYTES;
 551
 552         /*
 553          * Interrupt throttle rate
 554          */
 555         if (em_int_throttle_ceil == 0) {
 556                 adapter->int_throttle_ceil = 0;
 557         } else {
 558                 int throttle = em_int_throttle_ceil;
 559
 560                 if (throttle < 0)
 561                         throttle = EM_DEFAULT_ITR;
 562
 563                 /* Recalculate the tunable value to get the exact frequency. */
 564                 throttle = 1000000000 / 256 / throttle;
 565
 566                 /* Upper 16bits of ITR is reserved and should be zero */
 567                 if (throttle & 0xffff0000)
 568                         throttle = 1000000000 / 256 / EM_DEFAULT_ITR;
 569
 570                 adapter->int_throttle_ceil = 1000000000 / 256 / throttle;
 571         }
 572
 573         e1000_init_script_state_82541(&adapter->hw, TRUE);
 574         e1000_set_tbi_compatibility_82543(&adapter->hw, TRUE);
 575
 576         /* Copper options */
 577         if (adapter->hw.phy.media_type == e1000_media_type_copper) {
 578                 adapter->hw.phy.mdix = AUTO_ALL_MODES;
 579                 adapter->hw.phy.disable_polarity_correction = FALSE;
 580                 adapter->hw.phy.ms_type = EM_MASTER_SLAVE;
 581         }
 582
 583         /* Set the frame limits assuming standard ethernet sized frames. */
 584         adapter->max_frame_size = ETHERMTU + ETHER_HDR_LEN + ETHER_CRC_LEN;
 585         adapter->min_frame_size = ETH_ZLEN + ETHER_CRC_LEN;
 586
 587         /* This controls when hardware reports transmit completion status. */
 588         adapter->hw.mac.report_tx_early = 1;
 589
 590         /*
 591          * Create top level busdma tag
 592          */
 593         error = bus_dma_tag_create(NULL, 1, 0,
 594                         BUS_SPACE_MAXADDR, BUS_SPACE_MAXADDR,
 595                         NULL, NULL,
 596                         BUS_SPACE_MAXSIZE_32BIT, 0, BUS_SPACE_MAXSIZE_32BIT,
 597                         0, &adapter->parent_dtag);
 598         if (error) {
 599                 device_printf(dev, "could not create top level DMA tag\n");
 600                 goto fail;
 601         }
 602
 603         /*
 604          * Allocate Transmit Descriptor ring
 605          */
 606         tsize = roundup2(adapter->num_tx_desc * sizeof(struct e1000_tx_desc),
 607                          EM_DBA_ALIGN);
 608         error = em_dma_malloc(adapter, tsize, &adapter->txdma);
 609         if (error) {
 610                 device_printf(dev, "Unable to allocate tx_desc memory\n");
 611                 goto fail;
 612         }
 613         adapter->tx_desc_base = adapter->txdma.dma_vaddr;
 614
 615         /*
 616          * Allocate Receive Descriptor ring
 617          */
 618         rsize = roundup2(adapter->num_rx_desc * sizeof(struct e1000_rx_desc),
 619                          EM_DBA_ALIGN);
 620         error = em_dma_malloc(adapter, rsize, &adapter->rxdma);
 621         if (error) {
 622                 device_printf(dev, "Unable to allocate rx_desc memory\n");
 623                 goto fail;
 624         }
 625         adapter->rx_desc_base = adapter->rxdma.dma_vaddr;
 626
 627         /* Allocate multicast array memory. */
 628         adapter->mta = kmalloc(ETH_ADDR_LEN * MAX_NUM_MULTICAST_ADDRESSES,
 629             M_DEVBUF, M_WAITOK);
 630
 631         /* Indicate SOL/IDER usage */
 632         if (e1000_check_reset_block(&adapter->hw)) {
 633                 device_printf(dev,
 634                     "PHY reset is blocked due to SOL/IDER session.\n");
 635         }
 636
 637         /*
 638          * Start from a known state, this is important in reading the
 639          * nvm and mac from that.
 640          */
 641         e1000_reset_hw(&adapter->hw);
 642
 643         /* Make sure we have a good EEPROM before we read from it */
 644         if (e1000_validate_nvm_checksum(&adapter->hw) < 0) {
 645                 /*
 646                  * Some PCI-E parts fail the first check due to
 647                  * the link being in sleep state, call it again,
 648                  * if it fails a second time its a real issue.
 649                  */
 650                 if (e1000_validate_nvm_checksum(&adapter->hw) < 0) {
 651                         device_printf(dev,
 652                             "The EEPROM Checksum Is Not Valid\n");
 653                         error = EIO;
 654                         goto fail;
 655                 }
 656         }
 657
 658         /* Copy the permanent MAC address out of the EEPROM */
 659         if (e1000_read_mac_addr(&adapter->hw) < 0) {
 660                 device_printf(dev, "EEPROM read error while reading MAC"
 661                     " address\n");
 662                 error = EIO;
 663                 goto fail;
 664         }
 665         if (!em_is_valid_eaddr(adapter->hw.mac.addr)) {
 666                 device_printf(dev, "Invalid MAC address\n");
 667                 error = EIO;
 668                 goto fail;
 669         }
 670
 671         /* Allocate transmit descriptors and buffers */
 672         error = em_create_tx_ring(adapter);
 673         if (error) {
 674                 device_printf(dev, "Could not setup transmit structures\n");
 675                 goto fail;
 676         }
 677
 678         /* Allocate receive descriptors and buffers */
 679         error = em_create_rx_ring(adapter);
 680         if (error) {
 681                 device_printf(dev, "Could not setup receive structures\n");
 682                 goto fail;
 683         }
 684
 685         /* Manually turn off all interrupts */
 686         E1000_WRITE_REG(&adapter->hw, E1000_IMC, 0xffffffff);
 687
 688         /* Determine if we have to control management hardware */
 689         if (e1000_enable_mng_pass_thru(&adapter->hw))
 690                 adapter->flags |= EM_FLAG_HAS_MGMT;
 691
 692         /*
 693          * Setup Wake-on-Lan
 694          */
 695         apme_mask = EM_EEPROM_APME;
 696         eeprom_data = 0;
 697         switch (adapter->hw.mac.type) {
 698         case e1000_82542:
 699         case e1000_82543:
 700                 break;
 701
 702         case e1000_82573:
 703         case e1000_82583:
 704                 adapter->flags |= EM_FLAG_HAS_AMT;
 705                 /* FALL THROUGH */
 706
 707         case e1000_82546:
 708         case e1000_82546_rev_3:
 709         case e1000_82571:
 710         case e1000_82572:
 711         case e1000_80003es2lan:
 712                 if (adapter->hw.bus.func == 1) {
 713                         e1000_read_nvm(&adapter->hw,
 714                             NVM_INIT_CONTROL3_PORT_B, 1, &eeprom_data);
 715                 } else {
 716                         e1000_read_nvm(&adapter->hw,
 717                             NVM_INIT_CONTROL3_PORT_A, 1, &eeprom_data);
 718                 }
 719                 break;
 720
 721         case e1000_ich8lan:
 722         case e1000_ich9lan:
 723         case e1000_ich10lan:
 724         case e1000_pchlan:
 725         case e1000_pch2lan:
 726                 apme_mask = E1000_WUC_APME;
 727                 adapter->flags |= EM_FLAG_HAS_AMT;
 728                 eeprom_data = E1000_READ_REG(&adapter->hw, E1000_WUC);
 729                 break;
 730
 731         default:
 732                 e1000_read_nvm(&adapter->hw,
 733                     NVM_INIT_CONTROL3_PORT_A, 1, &eeprom_data);
 734                 break;
 735         }
 736         if (eeprom_data & apme_mask)
 737                 adapter->wol = E1000_WUFC_MAG | E1000_WUFC_MC;
 738
 739         /*
 740          * We have the eeprom settings, now apply the special cases
 741          * where the eeprom may be wrong or the board won't support
 742          * wake on lan on a particular port
 743          */
 744         device_id = pci_get_device(dev);
 745         switch (device_id) {
 746         case E1000_DEV_ID_82546GB_PCIE:
 747                 adapter->wol = 0;
 748                 break;
 749
 750         case E1000_DEV_ID_82546EB_FIBER:
 751         case E1000_DEV_ID_82546GB_FIBER:
 752         case E1000_DEV_ID_82571EB_FIBER:
 753                 /*
 754                  * Wake events only supported on port A for dual fiber
 755                  * regardless of eeprom setting
 756                  */
 757                 if (E1000_READ_REG(&adapter->hw, E1000_STATUS) &
 758                     E1000_STATUS_FUNC_1)
 759                         adapter->wol = 0;
 760                 break;
 761
 762         case E1000_DEV_ID_82546GB_QUAD_COPPER_KSP3:
 763         case E1000_DEV_ID_82571EB_QUAD_COPPER:
 764         case E1000_DEV_ID_82571EB_QUAD_FIBER:
 765         case E1000_DEV_ID_82571EB_QUAD_COPPER_LP:
 766                 /* if quad port adapter, disable WoL on all but port A */
 767                 if (em_global_quad_port_a != 0)
 768                         adapter->wol = 0;
 769                 /* Reset for multiple quad port adapters */
 770                 if (++em_global_quad_port_a == 4)
 771                         em_global_quad_port_a = 0;
 772                 break;
 773         }
 774
 775         /* XXX disable wol */
 776         adapter->wol = 0;
 777
 778         /* Setup OS specific network interface */
 779         em_setup_ifp(adapter);
 780
 781         /* Add sysctl tree, must after em_setup_ifp() */
 782         em_add_sysctl(adapter);
 783
 784         /* Reset the hardware */
 785         error = em_reset(adapter);
 786         if (error) {
 787                 device_printf(dev, "Unable to reset the hardware\n");
 788                 goto fail;
 789         }
 790
 791         /* Initialize statistics */
 792         em_update_stats(adapter);
 793
 794         adapter->hw.mac.get_link_status = 1;
 795         em_update_link_status(adapter);
 796
 797         /* Do we need workaround for 82544 PCI-X adapter? */
 798         if (adapter->hw.bus.type == e1000_bus_type_pcix &&
 799             adapter->hw.mac.type == e1000_82544)
 800                 adapter->pcix_82544 = TRUE;
 801         else
 802                 adapter->pcix_82544 = FALSE;
 803
 804         if (adapter->pcix_82544) {
 805                 /*
 806                  * 82544 on PCI-X may split one TX segment
 807                  * into two TX descs, so we double its number
 808                  * of spare TX desc here.
 809                  */
 810                 adapter->spare_tx_desc = 2 * EM_TX_SPARE;
 811         } else {
 812                 adapter->spare_tx_desc = EM_TX_SPARE;
 813         }
 814         if (adapter->flags & EM_FLAG_TSO)
 815                 adapter->spare_tx_desc = EM_TX_SPARE_TSO;
 816
 817         /*
 818          * Keep following relationship between spare_tx_desc, oact_tx_desc
 819          * and tx_int_nsegs:
 820          * (spare_tx_desc + EM_TX_RESERVED) <=
 821          * oact_tx_desc <= EM_TX_OACTIVE_MAX <= tx_int_nsegs
 822          */
 823         adapter->oact_tx_desc = adapter->num_tx_desc / 8;
 824         if (adapter->oact_tx_desc > EM_TX_OACTIVE_MAX)
 825                 adapter->oact_tx_desc = EM_TX_OACTIVE_MAX;
 826         if (adapter->oact_tx_desc < adapter->spare_tx_desc + EM_TX_RESERVED)
 827                 adapter->oact_tx_desc = adapter->spare_tx_desc + EM_TX_RESERVED;
 828
 829         adapter->tx_int_nsegs = adapter->num_tx_desc / 16;
 830         if (adapter->tx_int_nsegs < adapter->oact_tx_desc)
 831                 adapter->tx_int_nsegs = adapter->oact_tx_desc;
 832
 833         /* Non-AMT based hardware can now take control from firmware */
 834         if ((adapter->flags & (EM_FLAG_HAS_MGMT | EM_FLAG_HAS_AMT)) ==
 835             EM_FLAG_HAS_MGMT && adapter->hw.mac.type >= e1000_82571)
 836                 em_get_hw_control(adapter);
 837
 838         /*
 839          * Missing Interrupt Following ICR read:
 840          *
 841          * 82571/82572 specification update errata #76
 842          * 82573 specification update errata #31
 843          * 82574 specification update errata #12
 844          * 82583 specification update errata #4
 845          */
 846         intr_func = em_intr;
 847         if ((adapter->flags & EM_FLAG_SHARED_INTR) &&
 848             (adapter->hw.mac.type == e1000_82571 ||
 849              adapter->hw.mac.type == e1000_82572 ||
 850              adapter->hw.mac.type == e1000_82573 ||
 851              adapter->hw.mac.type == e1000_82574 ||
 852              adapter->hw.mac.type == e1000_82583))
 853                 intr_func = em_intr_mask;
 854
 855         error = bus_setup_intr(dev, adapter->intr_res, INTR_MPSAFE,
 856                                intr_func, adapter, &adapter->intr_tag,
 857                                ifp->if_serializer);
 858         if (error) {
 859                 device_printf(dev, "Failed to register interrupt handler");
 860                 ether_ifdetach(&adapter->arpcom.ac_if);
 861                 goto fail;
 862         }
 863
 864         ifp->if_cpuid = rman_get_cpuid(adapter->intr_res);
 865         KKASSERT(ifp->if_cpuid >= 0 && ifp->if_cpuid < ncpus);
 866         return (0);
 867 fail:
 868         em_detach(dev);
 869         return (error);
 870 }
 871
 872 static int
 873 em_detach(device_t dev)
 874 {
 875         struct adapter *adapter = device_get_softc(dev);
 876
 877         if (device_is_attached(dev)) {
 878                 struct ifnet *ifp = &adapter->arpcom.ac_if;
 879
 880                 lwkt_serialize_enter(ifp->if_serializer);
 881
 882                 em_stop(adapter);
 883
 884                 e1000_phy_hw_reset(&adapter->hw);
 885
 886                 em_rel_mgmt(adapter);
 887                 em_rel_hw_control(adapter);
 888
 889                 if (adapter->wol) {
 890                         E1000_WRITE_REG(&adapter->hw, E1000_WUC,
 891                                         E1000_WUC_PME_EN);
 892                         E1000_WRITE_REG(&adapter->hw, E1000_WUFC, adapter->wol);
 893                         em_enable_wol(dev);
 894                 }
 895
 896                 bus_teardown_intr(dev, adapter->intr_res, adapter->intr_tag);
 897
 898                 lwkt_serialize_exit(ifp->if_serializer);
 899
 900                 ether_ifdetach(ifp);
 901         } else if (adapter->memory != NULL) {
 902                 em_rel_hw_control(adapter);
 903         }
 904         bus_generic_detach(dev);
 905
 906         em_free_pci_res(adapter);
 907
 908         em_destroy_tx_ring(adapter, adapter->num_tx_desc);
 909         em_destroy_rx_ring(adapter, adapter->num_rx_desc);
 910
 911         /* Free Transmit Descriptor ring */
 912         if (adapter->tx_desc_base)
 913                 em_dma_free(adapter, &adapter->txdma);
 914
 915         /* Free Receive Descriptor ring */
 916         if (adapter->rx_desc_base)
 917                 em_dma_free(adapter, &adapter->rxdma);
 918
 919         /* Free top level busdma tag */
 920         if (adapter->parent_dtag != NULL)
 921                 bus_dma_tag_destroy(adapter->parent_dtag);
 922
 923         /* Free sysctl tree */
 924         if (adapter->sysctl_tree != NULL)
 925                 sysctl_ctx_free(&adapter->sysctl_ctx);
 926
 927         if (adapter->mta != NULL)
 928                 kfree(adapter->mta, M_DEVBUF);
 929
 930         return (0);
 931 }
 932
 933 static int
 934 em_shutdown(device_t dev)
 935 {
 936         return em_suspend(dev);
 937 }
 938
 939 static int
 940 em_suspend(device_t dev)
 941 {
 942         struct adapter *adapter = device_get_softc(dev);
 943         struct ifnet *ifp = &adapter->arpcom.ac_if;
 944
 945         lwkt_serialize_enter(ifp->if_serializer);
 946
 947         em_stop(adapter);
 948
 949         em_rel_mgmt(adapter);
 950         em_rel_hw_control(adapter);
 951
 952         if (adapter->wol) {
 953                 E1000_WRITE_REG(&adapter->hw, E1000_WUC, E1000_WUC_PME_EN);
 954                 E1000_WRITE_REG(&adapter->hw, E1000_WUFC, adapter->wol);
 955                 em_enable_wol(dev);
 956         }
 957
 958         lwkt_serialize_exit(ifp->if_serializer);
 959
 960         return bus_generic_suspend(dev);
 961 }
 962
 963 static int
 964 em_resume(device_t dev)
 965 {
 966         struct adapter *adapter = device_get_softc(dev);
 967         struct ifnet *ifp = &adapter->arpcom.ac_if;
 968
 969         lwkt_serialize_enter(ifp->if_serializer);
 970
 971         em_init(adapter);
 972         em_get_mgmt(adapter);
 973         if_devstart(ifp);
 974
 975         lwkt_serialize_exit(ifp->if_serializer);
 976
 977         return bus_generic_resume(dev);
 978 }
 979
 980 static void
 981 em_start(struct ifnet *ifp)
 982 {
 983         struct adapter *adapter = ifp->if_softc;
 984         struct mbuf *m_head;
 985
 986         ASSERT_SERIALIZED(ifp->if_serializer);
 987
 988         if ((ifp->if_flags & (IFF_RUNNING | IFF_OACTIVE)) != IFF_RUNNING)
 989                 return;
 990
 991         if (!adapter->link_active) {
 992                 ifq_purge(&ifp->if_snd);
 993                 return;
 994         }
 995
 996         while (!ifq_is_empty(&ifp->if_snd)) {
 997                 /* Now do we at least have a minimal? */
 998                 if (EM_IS_OACTIVE(adapter)) {
 999                         em_tx_collect(adapter);
1000                         if (EM_IS_OACTIVE(adapter)) {
1001                                 ifp->if_flags |= IFF_OACTIVE;
1002                                 adapter->no_tx_desc_avail1++;
1003                                 break;
1004                         }
1005                 }
1006
1007                 logif(pkt_txqueue);
1008                 m_head = ifq_dequeue(&ifp->if_snd, NULL);
1009                 if (m_head == NULL)
1010                         break;
1011
1012                 if (em_encap(adapter, &m_head)) {
1013                         ifp->if_oerrors++;
1014                         em_tx_collect(adapter);
1015                         continue;
1016                 }
1017
1018                 /* Send a copy of the frame to the BPF listener */
1019                 ETHER_BPF_MTAP(ifp, m_head);
1020
1021                 /* Set timeout in case hardware has problems transmitting. */
1022                 ifp->if_timer = EM_TX_TIMEOUT;
1023         }
1024 }
1025
1026 static int
1027 em_ioctl(struct ifnet *ifp, u_long command, caddr_t data, struct ucred *cr)
1028 {
1029         struct adapter *adapter = ifp->if_softc;
1030         struct ifreq *ifr = (struct ifreq *)data;
1031         uint16_t eeprom_data = 0;
1032         int max_frame_size, mask, reinit;
1033         int error = 0;
1034
1035         ASSERT_SERIALIZED(ifp->if_serializer);
1036
1037         switch (command) {
1038         case SIOCSIFMTU:
1039                 switch (adapter->hw.mac.type) {
1040                 case e1000_82573:
1041                         /*
1042                          * 82573 only supports jumbo frames
1043                          * if ASPM is disabled.
1044                          */
1045                         e1000_read_nvm(&adapter->hw,
1046                             NVM_INIT_3GIO_3, 1, &eeprom_data);
1047                         if (eeprom_data & NVM_WORD1A_ASPM_MASK) {
1048                                 max_frame_size = ETHER_MAX_LEN;
1049                                 break;
1050                         }
1051                         /* FALL THROUGH */
1052
1053                 /* Limit Jumbo Frame size */
1054                 case e1000_82571:
1055                 case e1000_82572:
1056                 case e1000_ich9lan:
1057                 case e1000_ich10lan:
1058                 case e1000_pch2lan:
1059                 case e1000_82574:
1060                 case e1000_82583:
1061                 case e1000_80003es2lan:
1062                         max_frame_size = 9234;
1063                         break;
1064
1065                 case e1000_pchlan:
1066                         max_frame_size = 4096;
1067                         break;
1068
1069                 /* Adapters that do not support jumbo frames */
1070                 case e1000_82542:
1071                 case e1000_ich8lan:
1072                         max_frame_size = ETHER_MAX_LEN;
1073                         break;
1074
1075                 default:
1076                         max_frame_size = MAX_JUMBO_FRAME_SIZE;
1077                         break;
1078                 }
1079                 if (ifr->ifr_mtu > max_frame_size - ETHER_HDR_LEN -
1080                     ETHER_CRC_LEN) {
1081                         error = EINVAL;
1082                         break;
1083                 }
1084
1085                 ifp->if_mtu = ifr->ifr_mtu;
1086                 adapter->max_frame_size =
1087                     ifp->if_mtu + ETHER_HDR_LEN + ETHER_CRC_LEN;
1088
1089                 if (ifp->if_flags & IFF_RUNNING)
1090                         em_init(adapter);
1091                 break;
1092
1093         case SIOCSIFFLAGS:
1094                 if (ifp->if_flags & IFF_UP) {
1095                         if ((ifp->if_flags & IFF_RUNNING)) {
1096                                 if ((ifp->if_flags ^ adapter->if_flags) &
1097                                     (IFF_PROMISC | IFF_ALLMULTI)) {
1098                                         em_disable_promisc(adapter);
1099                                         em_set_promisc(adapter);
1100                                 }
1101                         } else {
1102                                 em_init(adapter);
1103                         }
1104                 } else if (ifp->if_flags & IFF_RUNNING) {
1105                         em_stop(adapter);
1106                 }
1107                 adapter->if_flags = ifp->if_flags;
1108                 break;
1109
1110         case SIOCADDMULTI:
1111         case SIOCDELMULTI:
1112                 if (ifp->if_flags & IFF_RUNNING) {
1113                         em_disable_intr(adapter);
1114                         em_set_multi(adapter);
1115                         if (adapter->hw.mac.type == e1000_82542 &&
1116                             adapter->hw.revision_id == E1000_REVISION_2)
1117                                 em_init_rx_unit(adapter);
1118 #ifdef DEVICE_POLLING
1119                         if (!(ifp->if_flags & IFF_POLLING))
1120 #endif
1121                                 em_enable_intr(adapter);
1122                 }
1123                 break;
1124
1125         case SIOCSIFMEDIA:
1126                 /* Check SOL/IDER usage */
1127                 if (e1000_check_reset_block(&adapter->hw)) {
1128                         device_printf(adapter->dev, "Media change is"
1129                             " blocked due to SOL/IDER session.\n");
1130                         break;
1131                 }
1132                 /* FALL THROUGH */
1133
1134         case SIOCGIFMEDIA:
1135                 error = ifmedia_ioctl(ifp, ifr, &adapter->media, command);
1136                 break;
1137
1138         case SIOCSIFCAP:
1139                 reinit = 0;
1140                 mask = ifr->ifr_reqcap ^ ifp->if_capenable;
1141                 if (mask & IFCAP_RXCSUM) {
1142                         ifp->if_capenable ^= IFCAP_RXCSUM;
1143                         reinit = 1;
1144                 }
1145                 if (mask & IFCAP_TXCSUM) {
1146                         ifp->if_capenable ^= IFCAP_TXCSUM;
1147                         if (ifp->if_capenable & IFCAP_TXCSUM)
1148                                 ifp->if_hwassist |= EM_CSUM_FEATURES;
1149                         else
1150                                 ifp->if_hwassist &= ~EM_CSUM_FEATURES;
1151                 }
1152                 if (mask & IFCAP_TSO) {
1153                         ifp->if_capenable ^= IFCAP_TSO;
1154                         if (ifp->if_capenable & IFCAP_TSO)
1155                                 ifp->if_hwassist |= CSUM_TSO;
1156                         else
1157                                 ifp->if_hwassist &= ~CSUM_TSO;
1158                 }
1159                 if (mask & IFCAP_VLAN_HWTAGGING) {
1160                         ifp->if_capenable ^= IFCAP_VLAN_HWTAGGING;
1161                         reinit = 1;
1162                 }
1163                 if (reinit && (ifp->if_flags & IFF_RUNNING))
1164                         em_init(adapter);
1165                 break;
1166
1167         default:
1168                 error = ether_ioctl(ifp, command, data);
1169                 break;
1170         }
1171         return (error);
1172 }
1173
1174 static void
1175 em_watchdog(struct ifnet *ifp)
1176 {
1177         struct adapter *adapter = ifp->if_softc;
1178
1179         ASSERT_SERIALIZED(ifp->if_serializer);
1180
1181         /*
1182          * The timer is set to 5 every time start queues a packet.
1183          * Then txeof keeps resetting it as long as it cleans at
1184          * least one descriptor.
1185          * Finally, anytime all descriptors are clean the timer is
1186          * set to 0.
1187          */
1188
1189         if (E1000_READ_REG(&adapter->hw, E1000_TDT(0)) ==
1190             E1000_READ_REG(&adapter->hw, E1000_TDH(0))) {
1191                 /*
1192                  * If we reach here, all TX jobs are completed and
1193                  * the TX engine should have been idled for some time.
1194                  * We don't need to call if_devstart() here.
1195                  */
1196                 ifp->if_flags &= ~IFF_OACTIVE;
1197                 ifp->if_timer = 0;
1198                 return;
1199         }
1200
1201         /*
1202          * If we are in this routine because of pause frames, then
1203          * don't reset the hardware.
1204          */
1205         if (E1000_READ_REG(&adapter->hw, E1000_STATUS) &
1206             E1000_STATUS_TXOFF) {
1207                 ifp->if_timer = EM_TX_TIMEOUT;
1208                 return;
1209         }
1210
1211         if (e1000_check_for_link(&adapter->hw) == 0)
1212                 if_printf(ifp, "watchdog timeout -- resetting\n");
1213
1214         ifp->if_oerrors++;
1215         adapter->watchdog_events++;
1216
1217         em_init(adapter);
1218
1219         if (!ifq_is_empty(&ifp->if_snd))
1220                 if_devstart(ifp);
1221 }
1222
1223 static void
1224 em_init(void *xsc)
1225 {
1226         struct adapter *adapter = xsc;
1227         struct ifnet *ifp = &adapter->arpcom.ac_if;
1228         device_t dev = adapter->dev;
1229         uint32_t pba;
1230
1231         ASSERT_SERIALIZED(ifp->if_serializer);
1232
1233         em_stop(adapter);
1234
1235         /*
1236          * Packet Buffer Allocation (PBA)
1237          * Writing PBA sets the receive portion of the buffer
1238          * the remainder is used for the transmit buffer.
1239          *
1240          * Devices before the 82547 had a Packet Buffer of 64K.
1241          *   Default allocation: PBA=48K for Rx, leaving 16K for Tx.
1242          * After the 82547 the buffer was reduced to 40K.
1243          *   Default allocation: PBA=30K for Rx, leaving 10K for Tx.
1244          *   Note: default does not leave enough room for Jumbo Frame >10k.
1245          */
1246         switch (adapter->hw.mac.type) {
1247         case e1000_82547:
1248         case e1000_82547_rev_2: /* 82547: Total Packet Buffer is 40K */
1249                 if (adapter->max_frame_size > 8192)
1250                         pba = E1000_PBA_22K; /* 22K for Rx, 18K for Tx */
1251                 else
1252                         pba = E1000_PBA_30K; /* 30K for Rx, 10K for Tx */
1253                 adapter->tx_fifo_head = 0;
1254                 adapter->tx_head_addr = pba << EM_TX_HEAD_ADDR_SHIFT;
1255                 adapter->tx_fifo_size =
1256                     (E1000_PBA_40K - pba) << EM_PBA_BYTES_SHIFT;
1257                 break;
1258
1259         /* Total Packet Buffer on these is 48K */
1260         case e1000_82571:
1261         case e1000_82572:
1262         case e1000_80003es2lan:
1263                 pba = E1000_PBA_32K; /* 32K for Rx, 16K for Tx */
1264                 break;
1265
1266         case e1000_82573: /* 82573: Total Packet Buffer is 32K */
1267                 pba = E1000_PBA_12K; /* 12K for Rx, 20K for Tx */
1268                 break;
1269
1270         case e1000_82574:
1271         case e1000_82583:
1272                 pba = E1000_PBA_20K; /* 20K for Rx, 20K for Tx */
1273                 break;
1274
1275         case e1000_ich8lan:
1276                 pba = E1000_PBA_8K;
1277                 break;
1278
1279         case e1000_ich9lan:
1280         case e1000_ich10lan:
1281 #define E1000_PBA_10K   0x000A
1282                 pba = E1000_PBA_10K;
1283                 break;
1284
1285         case e1000_pchlan:
1286         case e1000_pch2lan:
1287                 pba = E1000_PBA_26K;
1288                 break;
1289
1290         default:
1291                 /* Devices before 82547 had a Packet Buffer of 64K.   */
1292                 if (adapter->max_frame_size > 8192)
1293                         pba = E1000_PBA_40K; /* 40K for Rx, 24K for Tx */
1294                 else
1295                         pba = E1000_PBA_48K; /* 48K for Rx, 16K for Tx */
1296         }
1297         E1000_WRITE_REG(&adapter->hw, E1000_PBA, pba);
1298
1299         /* Get the latest mac address, User can use a LAA */
1300         bcopy(IF_LLADDR(ifp), adapter->hw.mac.addr, ETHER_ADDR_LEN);
1301
1302         /* Put the address into the Receive Address Array */
1303         e1000_rar_set(&adapter->hw, adapter->hw.mac.addr, 0);
1304
1305         /*
1306          * With the 82571 adapter, RAR[0] may be overwritten
1307          * when the other port is reset, we make a duplicate
1308          * in RAR[14] for that eventuality, this assures
1309          * the interface continues to function.
1310          */
1311         if (adapter->hw.mac.type == e1000_82571) {
1312                 e1000_set_laa_state_82571(&adapter->hw, TRUE);
1313                 e1000_rar_set(&adapter->hw, adapter->hw.mac.addr,
1314                     E1000_RAR_ENTRIES - 1);
1315         }
1316
1317         /* Reset the hardware */
1318         if (em_reset(adapter)) {
1319                 device_printf(dev, "Unable to reset the hardware\n");
1320                 /* XXX em_stop()? */
1321                 return;
1322         }
1323         em_update_link_status(adapter);
1324
1325         /* Setup VLAN support, basic and offload if available */
1326         E1000_WRITE_REG(&adapter->hw, E1000_VET, ETHERTYPE_VLAN);
1327
1328         if (ifp->if_capenable & IFCAP_VLAN_HWTAGGING) {
1329                 uint32_t ctrl;
1330
1331                 ctrl = E1000_READ_REG(&adapter->hw, E1000_CTRL);
1332                 ctrl |= E1000_CTRL_VME;
1333                 E1000_WRITE_REG(&adapter->hw, E1000_CTRL, ctrl);
1334         }
1335
1336         /* Configure for OS presence */
1337         em_get_mgmt(adapter);
1338
1339         /* Prepare transmit descriptors and buffers */
1340         em_init_tx_ring(adapter);
1341         em_init_tx_unit(adapter);
1342
1343         /* Setup Multicast table */
1344         em_set_multi(adapter);
1345
1346         /* Prepare receive descriptors and buffers */
1347         if (em_init_rx_ring(adapter)) {
1348                 device_printf(dev, "Could not setup receive structures\n");
1349                 em_stop(adapter);
1350                 return;
1351         }
1352         em_init_rx_unit(adapter);
1353
1354         /* Don't lose promiscuous settings */
1355         em_set_promisc(adapter);
1356
1357         ifp->if_flags |= IFF_RUNNING;
1358         ifp->if_flags &= ~IFF_OACTIVE;
1359
1360         callout_reset(&adapter->timer, hz, em_timer, adapter);
1361         e1000_clear_hw_cntrs_base_generic(&adapter->hw);
1362
1363         /* MSI/X configuration for 82574 */
1364         if (adapter->hw.mac.type == e1000_82574) {
1365                 int tmp;
1366
1367                 tmp = E1000_READ_REG(&adapter->hw, E1000_CTRL_EXT);
1368                 tmp |= E1000_CTRL_EXT_PBA_CLR;
1369                 E1000_WRITE_REG(&adapter->hw, E1000_CTRL_EXT, tmp);
1370                 /*
1371                  * XXX MSIX
1372                  * Set the IVAR - interrupt vector routing.
1373                  * Each nibble represents a vector, high bit
1374                  * is enable, other 3 bits are the MSIX table
1375                  * entry, we map RXQ0 to 0, TXQ0 to 1, and
1376                  * Link (other) to 2, hence the magic number.
1377                  */
1378                 E1000_WRITE_REG(&adapter->hw, E1000_IVAR, 0x800A0908);
1379         }
1380
1381 #ifdef DEVICE_POLLING
1382         /*
1383          * Only enable interrupts if we are not polling, make sure
1384          * they are off otherwise.
1385          */
1386         if (ifp->if_flags & IFF_POLLING)
1387                 em_disable_intr(adapter);
1388         else
1389 #endif /* DEVICE_POLLING */
1390                 em_enable_intr(adapter);
1391
1392         /* AMT based hardware can now take control from firmware */
1393         if ((adapter->flags & (EM_FLAG_HAS_MGMT | EM_FLAG_HAS_AMT)) ==
1394             (EM_FLAG_HAS_MGMT | EM_FLAG_HAS_AMT) &&
1395             adapter->hw.mac.type >= e1000_82571)
1396                 em_get_hw_control(adapter);
1397
1398         /* Don't reset the phy next time init gets called */
1399         adapter->hw.phy.reset_disable = TRUE;
1400 }
1401
1402 #ifdef DEVICE_POLLING
1403
1404 static void
1405 em_poll(struct ifnet *ifp, enum poll_cmd cmd, int count)
1406 {
1407         struct adapter *adapter = ifp->if_softc;
1408         uint32_t reg_icr;
1409
1410         ASSERT_SERIALIZED(ifp->if_serializer);
1411
1412         switch (cmd) {
1413         case POLL_REGISTER:
1414                 em_disable_intr(adapter);
1415                 break;
1416
1417         case POLL_DEREGISTER:
1418                 em_enable_intr(adapter);
1419                 break;
1420
1421         case POLL_AND_CHECK_STATUS:
1422                 reg_icr = E1000_READ_REG(&adapter->hw, E1000_ICR);
1423                 if (reg_icr & (E1000_ICR_RXSEQ | E1000_ICR_LSC)) {
1424                         callout_stop(&adapter->timer);
1425                         adapter->hw.mac.get_link_status = 1;
1426                         em_update_link_status(adapter);
1427                         callout_reset(&adapter->timer, hz, em_timer, adapter);
1428                 }
1429                 /* FALL THROUGH */
1430         case POLL_ONLY:
1431                 if (ifp->if_flags & IFF_RUNNING) {
1432                         em_rxeof(adapter, count);
1433                         em_txeof(adapter);
1434
1435                         if (!ifq_is_empty(&ifp->if_snd))
1436                                 if_devstart(ifp);
1437                 }
1438                 break;
1439         }
1440 }
1441
1442 #endif /* DEVICE_POLLING */
1443
1444 static void
1445 em_intr(void *xsc)
1446 {
1447         em_intr_body(xsc, TRUE);
1448 }
1449
1450 static void
1451 em_intr_body(struct adapter *adapter, boolean_t chk_asserted)
1452 {
1453         struct ifnet *ifp = &adapter->arpcom.ac_if;
1454         uint32_t reg_icr;
1455
1456         logif(intr_beg);
1457         ASSERT_SERIALIZED(ifp->if_serializer);
1458
1459         reg_icr = E1000_READ_REG(&adapter->hw, E1000_ICR);
1460
1461         if (chk_asserted &&
1462             ((adapter->hw.mac.type >= e1000_82571 &&
1463               (reg_icr & E1000_ICR_INT_ASSERTED) == 0) ||
1464              reg_icr == 0)) {
1465                 logif(intr_end);
1466                 return;
1467         }
1468
1469         /*
1470          * XXX: some laptops trigger several spurious interrupts
1471          * on em(4) when in the resume cycle. The ICR register
1472          * reports all-ones value in this case. Processing such
1473          * interrupts would lead to a freeze. I don't know why.
1474          */
1475         if (reg_icr == 0xffffffff) {
1476                 logif(intr_end);
1477                 return;
1478         }
1479
1480         if (ifp->if_flags & IFF_RUNNING) {
1481                 if (reg_icr &
1482                     (E1000_ICR_RXT0 | E1000_ICR_RXDMT0 | E1000_ICR_RXO))
1483                         em_rxeof(adapter, -1);
1484                 if (reg_icr & E1000_ICR_TXDW) {
1485                         em_txeof(adapter);
1486                         if (!ifq_is_empty(&ifp->if_snd))
1487                                 if_devstart(ifp);
1488                 }
1489         }
1490
1491         /* Link status change */
1492         if (reg_icr & (E1000_ICR_RXSEQ | E1000_ICR_LSC)) {
1493                 callout_stop(&adapter->timer);
1494                 adapter->hw.mac.get_link_status = 1;
1495                 em_update_link_status(adapter);
1496
1497                 /* Deal with TX cruft when link lost */
1498                 em_tx_purge(adapter);
1499
1500                 callout_reset(&adapter->timer, hz, em_timer, adapter);
1501         }
1502
1503         if (reg_icr & E1000_ICR_RXO)
1504                 adapter->rx_overruns++;
1505
1506         logif(intr_end);
1507 }
1508
1509 static void
1510 em_intr_mask(void *xsc)
1511 {
1512         struct adapter *adapter = xsc;
1513
1514         E1000_WRITE_REG(&adapter->hw, E1000_IMC, 0xffffffff);
1515         /*
1516          * NOTE:
1517          * ICR.INT_ASSERTED bit will never be set if IMS is 0,
1518          * so don't check it.
1519          */
1520         em_intr_body(adapter, FALSE);
1521         E1000_WRITE_REG(&adapter->hw, E1000_IMS, IMS_ENABLE_MASK);
1522 }
1523
1524 static void
1525 em_media_status(struct ifnet *ifp, struct ifmediareq *ifmr)
1526 {
1527         struct adapter *adapter = ifp->if_softc;
1528         u_char fiber_type = IFM_1000_SX;
1529
1530         ASSERT_SERIALIZED(ifp->if_serializer);
1531
1532         em_update_link_status(adapter);
1533
1534         ifmr->ifm_status = IFM_AVALID;
1535         ifmr->ifm_active = IFM_ETHER;
1536
1537         if (!adapter->link_active)
1538                 return;
1539
1540         ifmr->ifm_status |= IFM_ACTIVE;
1541
1542         if (adapter->hw.phy.media_type == e1000_media_type_fiber ||
1543             adapter->hw.phy.media_type == e1000_media_type_internal_serdes) {
1544                 if (adapter->hw.mac.type == e1000_82545)
1545                         fiber_type = IFM_1000_LX;
1546                 ifmr->ifm_active |= fiber_type | IFM_FDX;
1547         } else {
1548                 switch (adapter->link_speed) {
1549                 case 10:
1550                         ifmr->ifm_active |= IFM_10_T;
1551                         break;
1552                 case 100:
1553                         ifmr->ifm_active |= IFM_100_TX;
1554                         break;
1555
1556                 case 1000:
1557                         ifmr->ifm_active |= IFM_1000_T;
1558                         break;
1559                 }
1560                 if (adapter->link_duplex == FULL_DUPLEX)
1561                         ifmr->ifm_active |= IFM_FDX;
1562                 else
1563                         ifmr->ifm_active |= IFM_HDX;
1564         }
1565 }
1566
1567 static int
1568 em_media_change(struct ifnet *ifp)
1569 {
1570         struct adapter *adapter = ifp->if_softc;
1571         struct ifmedia *ifm = &adapter->media;
1572
1573         ASSERT_SERIALIZED(ifp->if_serializer);
1574
1575         if (IFM_TYPE(ifm->ifm_media) != IFM_ETHER)
1576                 return (EINVAL);
1577
1578         switch (IFM_SUBTYPE(ifm->ifm_media)) {
1579         case IFM_AUTO:
1580                 adapter->hw.mac.autoneg = DO_AUTO_NEG;
1581                 adapter->hw.phy.autoneg_advertised = AUTONEG_ADV_DEFAULT;
1582                 break;
1583
1584         case IFM_1000_LX:
1585         case IFM_1000_SX:
1586         case IFM_1000_T:
1587                 adapter->hw.mac.autoneg = DO_AUTO_NEG;
1588                 adapter->hw.phy.autoneg_advertised = ADVERTISE_1000_FULL;
1589                 break;
1590
1591         case IFM_100_TX:
1592                 adapter->hw.mac.autoneg = FALSE;
1593                 adapter->hw.phy.autoneg_advertised = 0;
1594                 if ((ifm->ifm_media & IFM_GMASK) == IFM_FDX)
1595                         adapter->hw.mac.forced_speed_duplex = ADVERTISE_100_FULL;
1596                 else
1597                         adapter->hw.mac.forced_speed_duplex = ADVERTISE_100_HALF;
1598                 break;
1599
1600         case IFM_10_T:
1601                 adapter->hw.mac.autoneg = FALSE;
1602                 adapter->hw.phy.autoneg_advertised = 0;
1603                 if ((ifm->ifm_media & IFM_GMASK) == IFM_FDX)
1604                         adapter->hw.mac.forced_speed_duplex = ADVERTISE_10_FULL;
1605                 else
1606                         adapter->hw.mac.forced_speed_duplex = ADVERTISE_10_HALF;
1607                 break;
1608
1609         default:
1610                 if_printf(ifp, "Unsupported media type\n");
1611                 break;
1612         }
1613
1614         /*
1615          * As the speed/duplex settings my have changed we need to
1616          * reset the PHY.
1617          */
1618         adapter->hw.phy.reset_disable = FALSE;
1619
1620         em_init(adapter);
1621
1622         return (0);
1623 }
1624
1625 static int
1626 em_encap(struct adapter *adapter, struct mbuf **m_headp)
1627 {
1628         bus_dma_segment_t segs[EM_MAX_SCATTER];
1629         bus_dmamap_t map;
1630         struct em_buffer *tx_buffer, *tx_buffer_mapped;
1631         struct e1000_tx_desc *ctxd = NULL;
1632         struct mbuf *m_head = *m_headp;
1633         uint32_t txd_upper, txd_lower, txd_used, cmd = 0;
1634         int maxsegs, nsegs, i, j, first, last = 0, error;
1635
1636         if (m_head->m_pkthdr.csum_flags & CSUM_TSO) {
1637                 error = em_tso_pullup(adapter, m_headp);
1638                 if (error)
1639                         return error;
1640                 m_head = *m_headp;
1641         }
1642
1643         txd_upper = txd_lower = 0;
1644         txd_used = 0;
1645
1646         /*
1647          * Capture the first descriptor index, this descriptor
1648          * will have the index of the EOP which is the only one
1649          * that now gets a DONE bit writeback.
1650          */
1651         first = adapter->next_avail_tx_desc;
1652         tx_buffer = &adapter->tx_buffer_area[first];
1653         tx_buffer_mapped = tx_buffer;
1654         map = tx_buffer->map;
1655
1656         maxsegs = adapter->num_tx_desc_avail - EM_TX_RESERVED;
1657         KASSERT(maxsegs >= adapter->spare_tx_desc,
1658                 ("not enough spare TX desc"));
1659         if (adapter->pcix_82544) {
1660                 /* Half it; see the comment in em_attach() */
1661                 maxsegs >>= 1;
1662         }
1663         if (maxsegs > EM_MAX_SCATTER)
1664                 maxsegs = EM_MAX_SCATTER;
1665
1666         error = bus_dmamap_load_mbuf_defrag(adapter->txtag, map, m_headp,
1667                         segs, maxsegs, &nsegs, BUS_DMA_NOWAIT);
1668         if (error) {
1669                 if (error == ENOBUFS)
1670                         adapter->mbuf_alloc_failed++;
1671                 else
1672                         adapter->no_tx_dma_setup++;
1673
1674                 m_freem(*m_headp);
1675                 *m_headp = NULL;
1676                 return error;
1677         }
1678         bus_dmamap_sync(adapter->txtag, map, BUS_DMASYNC_PREWRITE);
1679
1680         m_head = *m_headp;
1681         adapter->tx_nsegs += nsegs;
1682
1683         if (m_head->m_pkthdr.csum_flags & CSUM_TSO) {
1684                 /* TSO will consume one TX desc */
1685                 adapter->tx_nsegs += em_tso_setup(adapter, m_head,
1686                     &txd_upper, &txd_lower);
1687         } else if (m_head->m_pkthdr.csum_flags & EM_CSUM_FEATURES) {
1688                 /* TX csum offloading will consume one TX desc */
1689                 adapter->tx_nsegs += em_txcsum(adapter, m_head,
1690                                                &txd_upper, &txd_lower);
1691         }
1692         i = adapter->next_avail_tx_desc;
1693
1694         /* Set up our transmit descriptors */
1695         for (j = 0; j < nsegs; j++) {
1696                 /* If adapter is 82544 and on PCIX bus */
1697                 if(adapter->pcix_82544) {
1698                         DESC_ARRAY desc_array;
1699                         uint32_t array_elements, counter;
1700
1701                         /*
1702                          * Check the Address and Length combination and
1703                          * split the data accordingly
1704                          */
1705                         array_elements = em_82544_fill_desc(segs[j].ds_addr,
1706                                                 segs[j].ds_len, &desc_array);
1707                         for (counter = 0; counter < array_elements; counter++) {
1708                                 KKASSERT(txd_used < adapter->num_tx_desc_avail);
1709
1710                                 tx_buffer = &adapter->tx_buffer_area[i];
1711                                 ctxd = &adapter->tx_desc_base[i];
1712
1713                                 ctxd->buffer_addr = htole64(
1714                                     desc_array.descriptor[counter].address);
1715                                 ctxd->lower.data = htole32(
1716                                     E1000_TXD_CMD_IFCS | txd_lower |
1717                                     desc_array.descriptor[counter].length);
1718                                 ctxd->upper.data = htole32(txd_upper);
1719
1720                                 last = i;
1721                                 if (++i == adapter->num_tx_desc)
1722                                         i = 0;
1723
1724                                 txd_used++;
1725                         }
1726                 } else {
1727                         tx_buffer = &adapter->tx_buffer_area[i];
1728                         ctxd = &adapter->tx_desc_base[i];
1729
1730                         ctxd->buffer_addr = htole64(segs[j].ds_addr);
1731                         ctxd->lower.data = htole32(E1000_TXD_CMD_IFCS |
1732                                                    txd_lower | segs[j].ds_len);
1733                         ctxd->upper.data = htole32(txd_upper);
1734
1735                         last = i;
1736                         if (++i == adapter->num_tx_desc)
1737                                 i = 0;
1738                 }
1739         }
1740
1741         adapter->next_avail_tx_desc = i;
1742         if (adapter->pcix_82544) {
1743                 KKASSERT(adapter->num_tx_desc_avail > txd_used);
1744                 adapter->num_tx_desc_avail -= txd_used;
1745         } else {
1746                 KKASSERT(adapter->num_tx_desc_avail > nsegs);
1747                 adapter->num_tx_desc_avail -= nsegs;
1748         }
1749
1750         /* Handle VLAN tag */
1751         if (m_head->m_flags & M_VLANTAG) {
1752                 /* Set the vlan id. */
1753                 ctxd->upper.fields.special =
1754                     htole16(m_head->m_pkthdr.ether_vlantag);
1755
1756                 /* Tell hardware to add tag */
1757                 ctxd->lower.data |= htole32(E1000_TXD_CMD_VLE);
1758         }
1759
1760         tx_buffer->m_head = m_head;
1761         tx_buffer_mapped->map = tx_buffer->map;
1762         tx_buffer->map = map;
1763
1764         if (adapter->tx_nsegs >= adapter->tx_int_nsegs) {
1765                 adapter->tx_nsegs = 0;
1766
1767                 /*
1768                  * Report Status (RS) is turned on
1769                  * every tx_int_nsegs descriptors.
1770                  */
1771                 cmd = E1000_TXD_CMD_RS;
1772
1773                 /*
1774                  * Keep track of the descriptor, which will
1775                  * be written back by hardware.
1776                  */
1777                 adapter->tx_dd[adapter->tx_dd_tail] = last;
1778                 EM_INC_TXDD_IDX(adapter->tx_dd_tail);
1779                 KKASSERT(adapter->tx_dd_tail != adapter->tx_dd_head);
1780         }
1781
1782         /*
1783          * Last Descriptor of Packet needs End Of Packet (EOP)
1784          */
1785         ctxd->lower.data |= htole32(E1000_TXD_CMD_EOP | cmd);
1786
1787         /*
1788          * Advance the Transmit Descriptor Tail (TDT), this tells the E1000
1789          * that this frame is available to transmit.
1790          */
1791         if (adapter->hw.mac.type == e1000_82547 &&
1792             adapter->link_duplex == HALF_DUPLEX) {
1793                 em_82547_move_tail_serialized(adapter);
1794         } else {
1795                 E1000_WRITE_REG(&adapter->hw, E1000_TDT(0), i);
1796                 if (adapter->hw.mac.type == e1000_82547) {
1797                         em_82547_update_fifo_head(adapter,
1798                             m_head->m_pkthdr.len);
1799                 }
1800         }
1801         return (0);
1802 }
1803
1804 /*
1805  * 82547 workaround to avoid controller hang in half-duplex environment.
1806  * The workaround is to avoid queuing a large packet that would span
1807  * the internal Tx FIFO ring boundary.  We need to reset the FIFO pointers
1808  * in this case.  We do that only when FIFO is quiescent.
1809  */
1810 static void
1811 em_82547_move_tail_serialized(struct adapter *adapter)
1812 {
1813         struct e1000_tx_desc *tx_desc;
1814         uint16_t hw_tdt, sw_tdt, length = 0;
1815         bool eop = 0;
1816
1817         ASSERT_SERIALIZED(adapter->arpcom.ac_if.if_serializer);
1818
1819         hw_tdt = E1000_READ_REG(&adapter->hw, E1000_TDT(0));
1820         sw_tdt = adapter->next_avail_tx_desc;
1821
1822         while (hw_tdt != sw_tdt) {
1823                 tx_desc = &adapter->tx_desc_base[hw_tdt];
1824                 length += tx_desc->lower.flags.length;
1825                 eop = tx_desc->lower.data & E1000_TXD_CMD_EOP;
1826                 if (++hw_tdt == adapter->num_tx_desc)
1827                         hw_tdt = 0;
1828
1829                 if (eop) {
1830                         if (em_82547_fifo_workaround(adapter, length)) {
1831                                 adapter->tx_fifo_wrk_cnt++;
1832                                 callout_reset(&adapter->tx_fifo_timer, 1,
1833                                         em_82547_move_tail, adapter);
1834                                 break;
1835                         }
1836                         E1000_WRITE_REG(&adapter->hw, E1000_TDT(0), hw_tdt);
1837                         em_82547_update_fifo_head(adapter, length);
1838                         length = 0;
1839                 }
1840         }
1841 }
1842
1843 static void
1844 em_82547_move_tail(void *xsc)
1845 {
1846         struct adapter *adapter = xsc;
1847         struct ifnet *ifp = &adapter->arpcom.ac_if;
1848
1849         lwkt_serialize_enter(ifp->if_serializer);
1850         em_82547_move_tail_serialized(adapter);
1851         lwkt_serialize_exit(ifp->if_serializer);
1852 }
1853
1854 static int
1855 em_82547_fifo_workaround(struct adapter *adapter, int len)
1856 {
1857         int fifo_space, fifo_pkt_len;
1858
1859         fifo_pkt_len = roundup2(len + EM_FIFO_HDR, EM_FIFO_HDR);
1860
1861         if (adapter->link_duplex == HALF_DUPLEX) {
1862                 fifo_space = adapter->tx_fifo_size - adapter->tx_fifo_head;
1863
1864                 if (fifo_pkt_len >= (EM_82547_PKT_THRESH + fifo_space)) {
1865                         if (em_82547_tx_fifo_reset(adapter))
1866                                 return (0);
1867                         else
1868                                 return (1);
1869                 }
1870         }
1871         return (0);
1872 }
1873
1874 static void
1875 em_82547_update_fifo_head(struct adapter *adapter, int len)
1876 {
1877         int fifo_pkt_len = roundup2(len + EM_FIFO_HDR, EM_FIFO_HDR);
1878
1879         /* tx_fifo_head is always 16 byte aligned */
1880         adapter->tx_fifo_head += fifo_pkt_len;
1881         if (adapter->tx_fifo_head >= adapter->tx_fifo_size)
1882                 adapter->tx_fifo_head -= adapter->tx_fifo_size;
1883 }
1884
1885 static int
1886 em_82547_tx_fifo_reset(struct adapter *adapter)
1887 {
1888         uint32_t tctl;
1889
1890         if ((E1000_READ_REG(&adapter->hw, E1000_TDT(0)) ==
1891              E1000_READ_REG(&adapter->hw, E1000_TDH(0))) &&
1892             (E1000_READ_REG(&adapter->hw, E1000_TDFT) ==
1893              E1000_READ_REG(&adapter->hw, E1000_TDFH)) &&
1894             (E1000_READ_REG(&adapter->hw, E1000_TDFTS) ==
1895              E1000_READ_REG(&adapter->hw, E1000_TDFHS)) &&
1896             (E1000_READ_REG(&adapter->hw, E1000_TDFPC) == 0)) {
1897                 /* Disable TX unit */
1898                 tctl = E1000_READ_REG(&adapter->hw, E1000_TCTL);
1899                 E1000_WRITE_REG(&adapter->hw, E1000_TCTL,
1900                     tctl & ~E1000_TCTL_EN);
1901
1902                 /* Reset FIFO pointers */
1903                 E1000_WRITE_REG(&adapter->hw, E1000_TDFT,
1904                     adapter->tx_head_addr);
1905                 E1000_WRITE_REG(&adapter->hw, E1000_TDFH,
1906                     adapter->tx_head_addr);
1907                 E1000_WRITE_REG(&adapter->hw, E1000_TDFTS,
1908                     adapter->tx_head_addr);
1909                 E1000_WRITE_REG(&adapter->hw, E1000_TDFHS,
1910                     adapter->tx_head_addr);
1911
1912                 /* Re-enable TX unit */
1913                 E1000_WRITE_REG(&adapter->hw, E1000_TCTL, tctl);
1914                 E1000_WRITE_FLUSH(&adapter->hw);
1915
1916                 adapter->tx_fifo_head = 0;
1917                 adapter->tx_fifo_reset_cnt++;
1918
1919                 return (TRUE);
1920         } else {
1921                 return (FALSE);
1922         }
1923 }
1924
1925 static void
1926 em_set_promisc(struct adapter *adapter)
1927 {
1928         struct ifnet *ifp = &adapter->arpcom.ac_if;
1929         uint32_t reg_rctl;
1930
1931         reg_rctl = E1000_READ_REG(&adapter->hw, E1000_RCTL);
1932
1933         if (ifp->if_flags & IFF_PROMISC) {
1934                 reg_rctl |= (E1000_RCTL_UPE | E1000_RCTL_MPE);
1935                 /* Turn this on if you want to see bad packets */
1936                 if (em_debug_sbp)
1937                         reg_rctl |= E1000_RCTL_SBP;
1938                 E1000_WRITE_REG(&adapter->hw, E1000_RCTL, reg_rctl);
1939         } else if (ifp->if_flags & IFF_ALLMULTI) {
1940                 reg_rctl |= E1000_RCTL_MPE;
1941                 reg_rctl &= ~E1000_RCTL_UPE;
1942                 E1000_WRITE_REG(&adapter->hw, E1000_RCTL, reg_rctl);
1943         }
1944 }
1945
1946 static void
1947 em_disable_promisc(struct adapter *adapter)
1948 {
1949         uint32_t reg_rctl;
1950
1951         reg_rctl = E1000_READ_REG(&adapter->hw, E1000_RCTL);
1952
1953         reg_rctl &= ~E1000_RCTL_UPE;
1954         reg_rctl &= ~E1000_RCTL_MPE;
1955         reg_rctl &= ~E1000_RCTL_SBP;
1956         E1000_WRITE_REG(&adapter->hw, E1000_RCTL, reg_rctl);
1957 }
1958
1959 static void
1960 em_set_multi(struct adapter *adapter)
1961 {
1962         struct ifnet *ifp = &adapter->arpcom.ac_if;
1963         struct ifmultiaddr *ifma;
1964         uint32_t reg_rctl = 0;
1965         uint8_t *mta;
1966         int mcnt = 0;
1967
1968         mta = adapter->mta;
1969         bzero(mta, ETH_ADDR_LEN * MAX_NUM_MULTICAST_ADDRESSES);
1970
1971         if (adapter->hw.mac.type == e1000_82542 &&
1972             adapter->hw.revision_id == E1000_REVISION_2) {
1973                 reg_rctl = E1000_READ_REG(&adapter->hw, E1000_RCTL);
1974                 if (adapter->hw.bus.pci_cmd_word & CMD_MEM_WRT_INVALIDATE)
1975                         e1000_pci_clear_mwi(&adapter->hw);
1976                 reg_rctl |= E1000_RCTL_RST;
1977                 E1000_WRITE_REG(&adapter->hw, E1000_RCTL, reg_rctl);
1978                 msec_delay(5);
1979         }
1980
1981         TAILQ_FOREACH(ifma, &ifp->if_multiaddrs, ifma_link) {
1982                 if (ifma->ifma_addr->sa_family != AF_LINK)
1983                         continue;
1984
1985                 if (mcnt == MAX_NUM_MULTICAST_ADDRESSES)
1986                         break;
1987
1988                 bcopy(LLADDR((struct sockaddr_dl *)ifma->ifma_addr),
1989                     &mta[mcnt * ETHER_ADDR_LEN], ETHER_ADDR_LEN);
1990                 mcnt++;
1991         }
1992
1993         if (mcnt >= MAX_NUM_MULTICAST_ADDRESSES) {
1994                 reg_rctl = E1000_READ_REG(&adapter->hw, E1000_RCTL);
1995                 reg_rctl |= E1000_RCTL_MPE;
1996                 E1000_WRITE_REG(&adapter->hw, E1000_RCTL, reg_rctl);
1997         } else {
1998                 e1000_update_mc_addr_list(&adapter->hw, mta, mcnt);
1999         }
2000
2001         if (adapter->hw.mac.type == e1000_82542 &&
2002             adapter->hw.revision_id == E1000_REVISION_2) {
2003                 reg_rctl = E1000_READ_REG(&adapter->hw, E1000_RCTL);
2004                 reg_rctl &= ~E1000_RCTL_RST;
2005                 E1000_WRITE_REG(&adapter->hw, E1000_RCTL, reg_rctl);
2006                 msec_delay(5);
2007                 if (adapter->hw.bus.pci_cmd_word & CMD_MEM_WRT_INVALIDATE)
2008                         e1000_pci_set_mwi(&adapter->hw);
2009         }
2010 }
2011
2012 /*
2013  * This routine checks for link status and updates statistics.
2014  */
2015 static void
2016 em_timer(void *xsc)
2017 {
2018         struct adapter *adapter = xsc;
2019         struct ifnet *ifp = &adapter->arpcom.ac_if;
2020
2021         lwkt_serialize_enter(ifp->if_serializer);
2022
2023         em_update_link_status(adapter);
2024         em_update_stats(adapter);
2025
2026         /* Reset LAA into RAR[0] on 82571 */
2027         if (e1000_get_laa_state_82571(&adapter->hw) == TRUE)
2028                 e1000_rar_set(&adapter->hw, adapter->hw.mac.addr, 0);
2029
2030         if (em_display_debug_stats && (ifp->if_flags & IFF_RUNNING))
2031                 em_print_hw_stats(adapter);
2032
2033         em_smartspeed(adapter);
2034
2035         callout_reset(&adapter->timer, hz, em_timer, adapter);
2036
2037         lwkt_serialize_exit(ifp->if_serializer);
2038 }
2039
2040 static void
2041 em_update_link_status(struct adapter *adapter)
2042 {
2043         struct e1000_hw *hw = &adapter->hw;
2044         struct ifnet *ifp = &adapter->arpcom.ac_if;
2045         device_t dev = adapter->dev;
2046         uint32_t link_check = 0;
2047
2048         /* Get the cached link value or read phy for real */
2049         switch (hw->phy.media_type) {
2050         case e1000_media_type_copper:
2051                 if (hw->mac.get_link_status) {
2052                         /* Do the work to read phy */
2053                         e1000_check_for_link(hw);
2054                         link_check = !hw->mac.get_link_status;
2055                         if (link_check) /* ESB2 fix */
2056                                 e1000_cfg_on_link_up(hw);
2057                 } else {
2058                         link_check = TRUE;
2059                 }
2060                 break;
2061
2062         case e1000_media_type_fiber:
2063                 e1000_check_for_link(hw);
2064                 link_check =
2065                         E1000_READ_REG(hw, E1000_STATUS) & E1000_STATUS_LU;
2066                 break;
2067
2068         case e1000_media_type_internal_serdes:
2069                 e1000_check_for_link(hw);
2070                 link_check = adapter->hw.mac.serdes_has_link;
2071                 break;
2072
2073         case e1000_media_type_unknown:
2074         default:
2075                 break;
2076         }
2077
2078         /* Now check for a transition */
2079         if (link_check && adapter->link_active == 0) {
2080                 e1000_get_speed_and_duplex(hw, &adapter->link_speed,
2081                     &adapter->link_duplex);
2082
2083                 /*
2084                  * Check if we should enable/disable SPEED_MODE bit on
2085                  * 82571/82572
2086                  */
2087                 if (adapter->link_speed != SPEED_1000 &&
2088                     (hw->mac.type == e1000_82571 ||
2089                      hw->mac.type == e1000_82572)) {
2090                         int tarc0;
2091
2092                         tarc0 = E1000_READ_REG(hw, E1000_TARC(0));
2093                         tarc0 &= ~SPEED_MODE_BIT;
2094                         E1000_WRITE_REG(hw, E1000_TARC(0), tarc0);
2095                 }
2096                 if (bootverbose) {
2097                         device_printf(dev, "Link is up %d Mbps %s\n",
2098                             adapter->link_speed,
2099                             ((adapter->link_duplex == FULL_DUPLEX) ?
2100                             "Full Duplex" : "Half Duplex"));
2101                 }
2102                 adapter->link_active = 1;
2103                 adapter->smartspeed = 0;
2104                 ifp->if_baudrate = adapter->link_speed * 1000000;
2105                 ifp->if_link_state = LINK_STATE_UP;
2106                 if_link_state_change(ifp);
2107         } else if (!link_check && adapter->link_active == 1) {
2108                 ifp->if_baudrate = adapter->link_speed = 0;
2109                 adapter->link_duplex = 0;
2110                 if (bootverbose)
2111                         device_printf(dev, "Link is Down\n");
2112                 adapter->link_active = 0;
2113 #if 0
2114                 /* Link down, disable watchdog */
2115                 if->if_timer = 0;
2116 #endif
2117                 ifp->if_link_state = LINK_STATE_DOWN;
2118                 if_link_state_change(ifp);
2119         }
2120 }
2121
2122 static void
2123 em_stop(struct adapter *adapter)
2124 {
2125         struct ifnet *ifp = &adapter->arpcom.ac_if;
2126         int i;
2127
2128         ASSERT_SERIALIZED(ifp->if_serializer);
2129
2130         em_disable_intr(adapter);
2131
2132         callout_stop(&adapter->timer);
2133         callout_stop(&adapter->tx_fifo_timer);
2134
2135         ifp->if_flags &= ~(IFF_RUNNING | IFF_OACTIVE);
2136         ifp->if_timer = 0;
2137
2138         e1000_reset_hw(&adapter->hw);
2139         if (adapter->hw.mac.type >= e1000_82544)
2140                 E1000_WRITE_REG(&adapter->hw, E1000_WUC, 0);
2141
2142         for (i = 0; i < adapter->num_tx_desc; i++) {
2143                 struct em_buffer *tx_buffer = &adapter->tx_buffer_area[i];
2144
2145                 if (tx_buffer->m_head != NULL) {
2146                         bus_dmamap_unload(adapter->txtag, tx_buffer->map);
2147                         m_freem(tx_buffer->m_head);
2148                         tx_buffer->m_head = NULL;
2149                 }
2150         }
2151
2152         for (i = 0; i < adapter->num_rx_desc; i++) {
2153                 struct em_buffer *rx_buffer = &adapter->rx_buffer_area[i];
2154
2155                 if (rx_buffer->m_head != NULL) {
2156                         bus_dmamap_unload(adapter->rxtag, rx_buffer->map);
2157                         m_freem(rx_buffer->m_head);
2158                         rx_buffer->m_head = NULL;
2159                 }
2160         }
2161
2162         if (adapter->fmp != NULL)
2163                 m_freem(adapter->fmp);
2164         adapter->fmp = NULL;
2165         adapter->lmp = NULL;
2166
2167         adapter->csum_flags = 0;
2168         adapter->csum_lhlen = 0;
2169         adapter->csum_iphlen = 0;
2170         adapter->csum_thlen = 0;
2171         adapter->csum_mss = 0;
2172         adapter->csum_pktlen = 0;
2173
2174         adapter->tx_dd_head = 0;
2175         adapter->tx_dd_tail = 0;
2176         adapter->tx_nsegs = 0;
2177 }
2178
2179 static int
2180 em_get_hw_info(struct adapter *adapter)
2181 {
2182         device_t dev = adapter->dev;
2183
2184         /* Save off the information about this board */
2185         adapter->hw.vendor_id = pci_get_vendor(dev);
2186         adapter->hw.device_id = pci_get_device(dev);
2187         adapter->hw.revision_id = pci_get_revid(dev);
2188         adapter->hw.subsystem_vendor_id = pci_get_subvendor(dev);
2189         adapter->hw.subsystem_device_id = pci_get_subdevice(dev);
2190
2191         /* Do Shared Code Init and Setup */
2192         if (e1000_set_mac_type(&adapter->hw))
2193                 return ENXIO;
2194         return 0;
2195 }
2196
2197 static int
2198 em_alloc_pci_res(struct adapter *adapter)
2199 {
2200         device_t dev = adapter->dev;
2201         u_int intr_flags;
2202         int val, rid, msi_enable;
2203
2204         /* Enable bus mastering */
2205         pci_enable_busmaster(dev);
2206
2207         adapter->memory_rid = EM_BAR_MEM;
2208         adapter->memory = bus_alloc_resource_any(dev, SYS_RES_MEMORY,
2209                                 &adapter->memory_rid, RF_ACTIVE);
2210         if (adapter->memory == NULL) {
2211                 device_printf(dev, "Unable to allocate bus resource: memory\n");
2212                 return (ENXIO);
2213         }
2214         adapter->osdep.mem_bus_space_tag =
2215             rman_get_bustag(adapter->memory);
2216         adapter->osdep.mem_bus_space_handle =
2217             rman_get_bushandle(adapter->memory);
2218
2219         /* XXX This is quite goofy, it is not actually used */
2220         adapter->hw.hw_addr = (uint8_t *)&adapter->osdep.mem_bus_space_handle;
2221
2222         /* Only older adapters use IO mapping */
2223         if (adapter->hw.mac.type > e1000_82543 &&
2224             adapter->hw.mac.type < e1000_82571) {
2225                 /* Figure our where our IO BAR is ? */
2226                 for (rid = PCIR_BAR(0); rid < PCIR_CARDBUSCIS;) {
2227                         val = pci_read_config(dev, rid, 4);
2228                         if (EM_BAR_TYPE(val) == EM_BAR_TYPE_IO) {
2229                                 adapter->io_rid = rid;
2230                                 break;
2231                         }
2232                         rid += 4;
2233                         /* check for 64bit BAR */
2234                         if (EM_BAR_MEM_TYPE(val) == EM_BAR_MEM_TYPE_64BIT)
2235                                 rid += 4;
2236                 }
2237                 if (rid >= PCIR_CARDBUSCIS) {
2238                         device_printf(dev, "Unable to locate IO BAR\n");
2239                         return (ENXIO);
2240                 }
2241                 adapter->ioport = bus_alloc_resource_any(dev, SYS_RES_IOPORT,
2242                                         &adapter->io_rid, RF_ACTIVE);
2243                 if (adapter->ioport == NULL) {
2244                         device_printf(dev, "Unable to allocate bus resource: "
2245                             "ioport\n");
2246                         return (ENXIO);
2247                 }
2248                 adapter->hw.io_base = 0;
2249                 adapter->osdep.io_bus_space_tag =
2250                     rman_get_bustag(adapter->ioport);
2251                 adapter->osdep.io_bus_space_handle =
2252                     rman_get_bushandle(adapter->ioport);
2253         }
2254
2255         /*
2256          * Don't enable MSI-X on 82574, see:
2257          * 82574 specification update errata #15
2258          *
2259          * Don't enable MSI on PCI/PCI-X chips, see:
2260          * 82540 specification update errata #6
2261          * 82545 specification update errata #4
2262          *
2263          * Don't enable MSI on 82571/82572, see:
2264          * 82571/82572 specification update errata #63
2265          */
2266         msi_enable = em_msi_enable;
2267         if (msi_enable &&
2268             (!pci_is_pcie(dev) ||
2269              adapter->hw.mac.type == e1000_82571 ||
2270              adapter->hw.mac.type == e1000_82572))
2271                 msi_enable = 0;
2272
2273         adapter->intr_type = pci_alloc_1intr(dev, msi_enable,
2274             &adapter->intr_rid, &intr_flags);
2275
2276         if (adapter->intr_type == PCI_INTR_TYPE_LEGACY) {
2277                 int unshared;
2278
2279                 unshared = device_getenv_int(dev, "irq.unshared", 0);
2280                 if (!unshared) {
2281                         adapter->flags |= EM_FLAG_SHARED_INTR;
2282                         if (bootverbose)
2283                                 device_printf(dev, "IRQ shared\n");
2284                 } else {
2285                         intr_flags &= ~RF_SHAREABLE;
2286                         if (bootverbose)
2287                                 device_printf(dev, "IRQ unshared\n");
2288                 }
2289         }
2290
2291         adapter->intr_res = bus_alloc_resource_any(dev, SYS_RES_IRQ,
2292             &adapter->intr_rid, intr_flags);
2293         if (adapter->intr_res == NULL) {
2294                 device_printf(dev, "Unable to allocate bus resource: "
2295                     "interrupt\n");
2296                 return (ENXIO);
2297         }
2298
2299         adapter->hw.bus.pci_cmd_word = pci_read_config(dev, PCIR_COMMAND, 2);
2300         adapter->hw.back = &adapter->osdep;
2301         return (0);
2302 }
2303
2304 static void
2305 em_free_pci_res(struct adapter *adapter)
2306 {
2307         device_t dev = adapter->dev;
2308
2309         if (adapter->intr_res != NULL) {
2310                 bus_release_resource(dev, SYS_RES_IRQ,
2311                     adapter->intr_rid, adapter->intr_res);
2312         }
2313
2314         if (adapter->intr_type == PCI_INTR_TYPE_MSI)
2315                 pci_release_msi(dev);
2316
2317         if (adapter->memory != NULL) {
2318                 bus_release_resource(dev, SYS_RES_MEMORY,
2319                     adapter->memory_rid, adapter->memory);
2320         }
2321
2322         if (adapter->flash != NULL) {
2323                 bus_release_resource(dev, SYS_RES_MEMORY,
2324                     adapter->flash_rid, adapter->flash);
2325         }
2326
2327         if (adapter->ioport != NULL) {
2328                 bus_release_resource(dev, SYS_RES_IOPORT,
2329                     adapter->io_rid, adapter->ioport);
2330         }
2331 }
2332
2333 static int
2334 em_reset(struct adapter *adapter)
2335 {
2336         device_t dev = adapter->dev;
2337         uint16_t rx_buffer_size;
2338
2339         /* When hardware is reset, fifo_head is also reset */
2340         adapter->tx_fifo_head = 0;
2341
2342         /* Set up smart power down as default off on newer adapters. */
2343         if (!em_smart_pwr_down &&
2344             (adapter->hw.mac.type == e1000_82571 ||
2345              adapter->hw.mac.type == e1000_82572)) {
2346                 uint16_t phy_tmp = 0;
2347
2348                 /* Speed up time to link by disabling smart power down. */
2349                 e1000_read_phy_reg(&adapter->hw,
2350                     IGP02E1000_PHY_POWER_MGMT, &phy_tmp);
2351                 phy_tmp &= ~IGP02E1000_PM_SPD;
2352                 e1000_write_phy_reg(&adapter->hw,
2353                     IGP02E1000_PHY_POWER_MGMT, phy_tmp);
2354         }
2355
2356         /*
2357          * These parameters control the automatic generation (Tx) and
2358          * response (Rx) to Ethernet PAUSE frames.
2359          * - High water mark should allow for at least two frames to be
2360          *   received after sending an XOFF.
2361          * - Low water mark works best when it is very near the high water mark.
2362          *   This allows the receiver to restart by sending XON when it has
2363          *   drained a bit. Here we use an arbitary value of 1500 which will
2364          *   restart after one full frame is pulled from the buffer. There
2365          *   could be several smaller frames in the buffer and if so they will
2366          *   not trigger the XON until their total number reduces the buffer
2367          *   by 1500.
2368          * - The pause time is fairly large at 1000 x 512ns = 512 usec.
2369          */
2370         rx_buffer_size =
2371                 (E1000_READ_REG(&adapter->hw, E1000_PBA) & 0xffff) << 10;
2372
2373         adapter->hw.fc.high_water = rx_buffer_size -
2374                                     roundup2(adapter->max_frame_size, 1024);
2375         adapter->hw.fc.low_water = adapter->hw.fc.high_water - 1500;
2376
2377         if (adapter->hw.mac.type == e1000_80003es2lan)
2378                 adapter->hw.fc.pause_time = 0xFFFF;
2379         else
2380                 adapter->hw.fc.pause_time = EM_FC_PAUSE_TIME;
2381
2382         adapter->hw.fc.send_xon = TRUE;
2383
2384         adapter->hw.fc.requested_mode = e1000_fc_full;
2385
2386         /* Workaround: no TX flow ctrl for PCH */
2387         if (adapter->hw.mac.type == e1000_pchlan)
2388                 adapter->hw.fc.requested_mode = e1000_fc_rx_pause;
2389
2390         /* Override - settings for PCH2LAN, ya its magic :) */
2391         if (adapter->hw.mac.type == e1000_pch2lan) {
2392                 adapter->hw.fc.high_water = 0x5C20;
2393                 adapter->hw.fc.low_water = 0x5048;
2394                 adapter->hw.fc.pause_time = 0x0650;
2395                 adapter->hw.fc.refresh_time = 0x0400;
2396
2397                 /* Jumbos need adjusted PBA */
2398                 if (adapter->arpcom.ac_if.if_mtu > ETHERMTU)
2399                         E1000_WRITE_REG(&adapter->hw, E1000_PBA, 12);
2400                 else
2401                         E1000_WRITE_REG(&adapter->hw, E1000_PBA, 26);
2402         }
2403
2404         /* Issue a global reset */
2405         e1000_reset_hw(&adapter->hw);
2406         if (adapter->hw.mac.type >= e1000_82544)
2407                 E1000_WRITE_REG(&adapter->hw, E1000_WUC, 0);
2408         em_disable_aspm(adapter);
2409
2410         if (e1000_init_hw(&adapter->hw) < 0) {
2411                 device_printf(dev, "Hardware Initialization Failed\n");
2412                 return (EIO);
2413         }
2414
2415         E1000_WRITE_REG(&adapter->hw, E1000_VET, ETHERTYPE_VLAN);
2416         e1000_get_phy_info(&adapter->hw);
2417         e1000_check_for_link(&adapter->hw);
2418
2419         return (0);
2420 }
2421
2422 static void
2423 em_setup_ifp(struct adapter *adapter)
2424 {
2425         struct ifnet *ifp = &adapter->arpcom.ac_if;
2426
2427         if_initname(ifp, device_get_name(adapter->dev),
2428                     device_get_unit(adapter->dev));
2429         ifp->if_softc = adapter;
2430         ifp->if_flags = IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST;
2431         ifp->if_init =  em_init;
2432         ifp->if_ioctl = em_ioctl;
2433         ifp->if_start = em_start;
2434 #ifdef DEVICE_POLLING
2435         ifp->if_poll = em_poll;
2436 #endif
2437         ifp->if_watchdog = em_watchdog;
2438         ifq_set_maxlen(&ifp->if_snd, adapter->num_tx_desc - 1);
2439         ifq_set_ready(&ifp->if_snd);
2440
2441         ether_ifattach(ifp, adapter->hw.mac.addr, NULL);
2442
2443         ifp->if_capabilities = IFCAP_VLAN_HWTAGGING | IFCAP_VLAN_MTU;
2444         if (adapter->hw.mac.type >= e1000_82543)
2445                 ifp->if_capabilities |= IFCAP_HWCSUM;
2446         if (adapter->flags & EM_FLAG_TSO)
2447                 ifp->if_capabilities |= IFCAP_TSO;
2448         ifp->if_capenable = ifp->if_capabilities;
2449
2450         if (ifp->if_capenable & IFCAP_TXCSUM)
2451                 ifp->if_hwassist |= EM_CSUM_FEATURES;
2452         if (ifp->if_capenable & IFCAP_TSO)
2453                 ifp->if_hwassist |= CSUM_TSO;
2454
2455         /*
2456          * Tell the upper layer(s) we support long frames.
2457          */
2458         ifp->if_data.ifi_hdrlen = sizeof(struct ether_vlan_header);
2459
2460         /*
2461          * Specify the media types supported by this adapter and register
2462          * callbacks to update media and link information
2463          */
2464         ifmedia_init(&adapter->media, IFM_IMASK,
2465                      em_media_change, em_media_status);
2466         if (adapter->hw.phy.media_type == e1000_media_type_fiber ||
2467             adapter->hw.phy.media_type == e1000_media_type_internal_serdes) {
2468                 u_char fiber_type = IFM_1000_SX; /* default type */
2469
2470                 if (adapter->hw.mac.type == e1000_82545)
2471                         fiber_type = IFM_1000_LX;
2472                 ifmedia_add(&adapter->media, IFM_ETHER | fiber_type | IFM_FDX,
2473                             0, NULL);
2474                 ifmedia_add(&adapter->media, IFM_ETHER | fiber_type, 0, NULL);
2475         } else {
2476                 ifmedia_add(&adapter->media, IFM_ETHER | IFM_10_T, 0, NULL);
2477                 ifmedia_add(&adapter->media, IFM_ETHER | IFM_10_T | IFM_FDX,
2478                             0, NULL);
2479                 ifmedia_add(&adapter->media, IFM_ETHER | IFM_100_TX,
2480                             0, NULL);
2481                 ifmedia_add(&adapter->media, IFM_ETHER | IFM_100_TX | IFM_FDX,
2482                             0, NULL);
2483                 if (adapter->hw.phy.type != e1000_phy_ife) {
2484                         ifmedia_add(&adapter->media,
2485                                 IFM_ETHER | IFM_1000_T | IFM_FDX, 0, NULL);
2486                         ifmedia_add(&adapter->media,
2487                                 IFM_ETHER | IFM_1000_T, 0, NULL);
2488                 }
2489         }
2490         ifmedia_add(&adapter->media, IFM_ETHER | IFM_AUTO, 0, NULL);
2491         ifmedia_set(&adapter->media, IFM_ETHER | IFM_AUTO);
2492 }
2493
2494
2495 /*
2496  * Workaround for SmartSpeed on 82541 and 82547 controllers
2497  */
2498 static void
2499 em_smartspeed(struct adapter *adapter)
2500 {
2501         uint16_t phy_tmp;
2502
2503         if (adapter->link_active || adapter->hw.phy.type != e1000_phy_igp ||
2504             adapter->hw.mac.autoneg == 0 ||
2505             (adapter->hw.phy.autoneg_advertised & ADVERTISE_1000_FULL) == 0)
2506                 return;
2507
2508         if (adapter->smartspeed == 0) {
2509                 /*
2510                  * If Master/Slave config fault is asserted twice,
2511                  * we assume back-to-back
2512                  */
2513                 e1000_read_phy_reg(&adapter->hw, PHY_1000T_STATUS, &phy_tmp);
2514                 if (!(phy_tmp & SR_1000T_MS_CONFIG_FAULT))
2515                         return;
2516                 e1000_read_phy_reg(&adapter->hw, PHY_1000T_STATUS, &phy_tmp);
2517                 if (phy_tmp & SR_1000T_MS_CONFIG_FAULT) {
2518                         e1000_read_phy_reg(&adapter->hw,
2519                             PHY_1000T_CTRL, &phy_tmp);
2520                         if (phy_tmp & CR_1000T_MS_ENABLE) {
2521                                 phy_tmp &= ~CR_1000T_MS_ENABLE;
2522                                 e1000_write_phy_reg(&adapter->hw,
2523                                     PHY_1000T_CTRL, phy_tmp);
2524                                 adapter->smartspeed++;
2525                                 if (adapter->hw.mac.autoneg &&
2526                                     !e1000_phy_setup_autoneg(&adapter->hw) &&
2527                                     !e1000_read_phy_reg(&adapter->hw,
2528                                      PHY_CONTROL, &phy_tmp)) {
2529                                         phy_tmp |= MII_CR_AUTO_NEG_EN |
2530                                                    MII_CR_RESTART_AUTO_NEG;
2531                                         e1000_write_phy_reg(&adapter->hw,
2532                                             PHY_CONTROL, phy_tmp);
2533                                 }
2534                         }
2535                 }
2536                 return;
2537         } else if (adapter->smartspeed == EM_SMARTSPEED_DOWNSHIFT) {
2538                 /* If still no link, perhaps using 2/3 pair cable */
2539                 e1000_read_phy_reg(&adapter->hw, PHY_1000T_CTRL, &phy_tmp);
2540                 phy_tmp |= CR_1000T_MS_ENABLE;
2541                 e1000_write_phy_reg(&adapter->hw, PHY_1000T_CTRL, phy_tmp);
2542                 if (adapter->hw.mac.autoneg &&
2543                     !e1000_phy_setup_autoneg(&adapter->hw) &&
2544                     !e1000_read_phy_reg(&adapter->hw, PHY_CONTROL, &phy_tmp)) {
2545                         phy_tmp |= MII_CR_AUTO_NEG_EN | MII_CR_RESTART_AUTO_NEG;
2546                         e1000_write_phy_reg(&adapter->hw, PHY_CONTROL, phy_tmp);
2547                 }
2548         }
2549
2550         /* Restart process after EM_SMARTSPEED_MAX iterations */
2551         if (adapter->smartspeed++ == EM_SMARTSPEED_MAX)
2552                 adapter->smartspeed = 0;
2553 }
2554
2555 static int
2556 em_dma_malloc(struct adapter *adapter, bus_size_t size,
2557               struct em_dma_alloc *dma)
2558 {
2559         dma->dma_vaddr = bus_dmamem_coherent_any(adapter->parent_dtag,
2560                                 EM_DBA_ALIGN, size, BUS_DMA_WAITOK,
2561                                 &dma->dma_tag, &dma->dma_map,
2562                                 &dma->dma_paddr);
2563         if (dma->dma_vaddr == NULL)
2564                 return ENOMEM;
2565         else
2566                 return 0;
2567 }
2568
2569 static void
2570 em_dma_free(struct adapter *adapter, struct em_dma_alloc *dma)
2571 {
2572         if (dma->dma_tag == NULL)
2573                 return;
2574         bus_dmamap_unload(dma->dma_tag, dma->dma_map);
2575         bus_dmamem_free(dma->dma_tag, dma->dma_vaddr, dma->dma_map);
2576         bus_dma_tag_destroy(dma->dma_tag);
2577 }
2578
2579 static int
2580 em_create_tx_ring(struct adapter *adapter)
2581 {
2582         device_t dev = adapter->dev;
2583         struct em_buffer *tx_buffer;
2584         int error, i;
2585
2586         adapter->tx_buffer_area =
2587                 kmalloc(sizeof(struct em_buffer) * adapter->num_tx_desc,
2588                         M_DEVBUF, M_WAITOK | M_ZERO);
2589
2590         /*
2591          * Create DMA tags for tx buffers
2592          */
2593         error = bus_dma_tag_create(adapter->parent_dtag, /* parent */
2594                         1, 0,                   /* alignment, bounds */
2595                         BUS_SPACE_MAXADDR,      /* lowaddr */
2596                         BUS_SPACE_MAXADDR,      /* highaddr */
2597                         NULL, NULL,             /* filter, filterarg */
2598                         EM_TSO_SIZE,            /* maxsize */
2599                         EM_MAX_SCATTER,         /* nsegments */
2600                         PAGE_SIZE,              /* maxsegsize */
2601                         BUS_DMA_WAITOK | BUS_DMA_ALLOCNOW |
2602                         BUS_DMA_ONEBPAGE,       /* flags */
2603                         &adapter->txtag);
2604         if (error) {
2605                 device_printf(dev, "Unable to allocate TX DMA tag\n");
2606                 kfree(adapter->tx_buffer_area, M_DEVBUF);
2607                 adapter->tx_buffer_area = NULL;
2608                 return error;
2609         }
2610
2611         /*
2612          * Create DMA maps for tx buffers
2613          */
2614         for (i = 0; i < adapter->num_tx_desc; i++) {
2615                 tx_buffer = &adapter->tx_buffer_area[i];
2616
2617                 error = bus_dmamap_create(adapter->txtag,
2618                                           BUS_DMA_WAITOK | BUS_DMA_ONEBPAGE,
2619                                           &tx_buffer->map);
2620                 if (error) {
2621                         device_printf(dev, "Unable to create TX DMA map\n");
2622                         em_destroy_tx_ring(adapter, i);
2623                         return error;
2624                 }
2625         }
2626         return (0);
2627 }
2628
2629 static void
2630 em_init_tx_ring(struct adapter *adapter)
2631 {
2632         /* Clear the old ring contents */
2633         bzero(adapter->tx_desc_base,
2634             (sizeof(struct e1000_tx_desc)) * adapter->num_tx_desc);
2635
2636         /* Reset state */
2637         adapter->next_avail_tx_desc = 0;
2638         adapter->next_tx_to_clean = 0;
2639         adapter->num_tx_desc_avail = adapter->num_tx_desc;
2640 }
2641
2642 static void
2643 em_init_tx_unit(struct adapter *adapter)
2644 {
2645         uint32_t tctl, tarc, tipg = 0;
2646         uint64_t bus_addr;
2647
2648         /* Setup the Base and Length of the Tx Descriptor Ring */
2649         bus_addr = adapter->txdma.dma_paddr;
2650         E1000_WRITE_REG(&adapter->hw, E1000_TDLEN(0),
2651             adapter->num_tx_desc * sizeof(struct e1000_tx_desc));
2652         E1000_WRITE_REG(&adapter->hw, E1000_TDBAH(0),
2653             (uint32_t)(bus_addr >> 32));
2654         E1000_WRITE_REG(&adapter->hw, E1000_TDBAL(0),
2655             (uint32_t)bus_addr);
2656         /* Setup the HW Tx Head and Tail descriptor pointers */
2657         E1000_WRITE_REG(&adapter->hw, E1000_TDT(0), 0);
2658         E1000_WRITE_REG(&adapter->hw, E1000_TDH(0), 0);
2659
2660         /* Set the default values for the Tx Inter Packet Gap timer */
2661         switch (adapter->hw.mac.type) {
2662         case e1000_82542:
2663                 tipg = DEFAULT_82542_TIPG_IPGT;
2664                 tipg |= DEFAULT_82542_TIPG_IPGR1 << E1000_TIPG_IPGR1_SHIFT;
2665                 tipg |= DEFAULT_82542_TIPG_IPGR2 << E1000_TIPG_IPGR2_SHIFT;
2666                 break;
2667
2668         case e1000_80003es2lan:
2669                 tipg = DEFAULT_82543_TIPG_IPGR1;
2670                 tipg |= DEFAULT_80003ES2LAN_TIPG_IPGR2 <<
2671                     E1000_TIPG_IPGR2_SHIFT;
2672                 break;
2673
2674         default:
2675                 if (adapter->hw.phy.media_type == e1000_media_type_fiber ||
2676                     adapter->hw.phy.media_type ==
2677                     e1000_media_type_internal_serdes)
2678                         tipg = DEFAULT_82543_TIPG_IPGT_FIBER;
2679                 else
2680                         tipg = DEFAULT_82543_TIPG_IPGT_COPPER;
2681                 tipg |= DEFAULT_82543_TIPG_IPGR1 << E1000_TIPG_IPGR1_SHIFT;
2682                 tipg |= DEFAULT_82543_TIPG_IPGR2 << E1000_TIPG_IPGR2_SHIFT;
2683                 break;
2684         }
2685
2686         E1000_WRITE_REG(&adapter->hw, E1000_TIPG, tipg);
2687
2688         /* NOTE: 0 is not allowed for TIDV */
2689         E1000_WRITE_REG(&adapter->hw, E1000_TIDV, 1);
2690         if(adapter->hw.mac.type >= e1000_82540)
2691                 E1000_WRITE_REG(&adapter->hw, E1000_TADV, 0);
2692
2693         if (adapter->hw.mac.type == e1000_82571 ||
2694             adapter->hw.mac.type == e1000_82572) {
2695                 tarc = E1000_READ_REG(&adapter->hw, E1000_TARC(0));
2696                 tarc |= SPEED_MODE_BIT;
2697                 E1000_WRITE_REG(&adapter->hw, E1000_TARC(0), tarc);
2698         } else if (adapter->hw.mac.type == e1000_80003es2lan) {
2699                 tarc = E1000_READ_REG(&adapter->hw, E1000_TARC(0));
2700                 tarc |= 1;
2701                 E1000_WRITE_REG(&adapter->hw, E1000_TARC(0), tarc);
2702                 tarc = E1000_READ_REG(&adapter->hw, E1000_TARC(1));
2703                 tarc |= 1;
2704                 E1000_WRITE_REG(&adapter->hw, E1000_TARC(1), tarc);
2705         }
2706
2707         /* Program the Transmit Control Register */
2708         tctl = E1000_READ_REG(&adapter->hw, E1000_TCTL);
2709         tctl &= ~E1000_TCTL_CT;
2710         tctl |= E1000_TCTL_PSP | E1000_TCTL_RTLC | E1000_TCTL_EN |
2711                 (E1000_COLLISION_THRESHOLD << E1000_CT_SHIFT);
2712
2713         if (adapter->hw.mac.type >= e1000_82571)
2714                 tctl |= E1000_TCTL_MULR;
2715
2716         /* This write will effectively turn on the transmit unit. */
2717         E1000_WRITE_REG(&adapter->hw, E1000_TCTL, tctl);
2718 }
2719
2720 static void
2721 em_destroy_tx_ring(struct adapter *adapter, int ndesc)
2722 {
2723         struct em_buffer *tx_buffer;
2724         int i;
2725
2726         if (adapter->tx_buffer_area == NULL)
2727                 return;
2728
2729         for (i = 0; i < ndesc; i++) {
2730                 tx_buffer = &adapter->tx_buffer_area[i];
2731
2732                 KKASSERT(tx_buffer->m_head == NULL);
2733                 bus_dmamap_destroy(adapter->txtag, tx_buffer->map);
2734         }
2735         bus_dma_tag_destroy(adapter->txtag);
2736
2737         kfree(adapter->tx_buffer_area, M_DEVBUF);
2738         adapter->tx_buffer_area = NULL;
2739 }
2740
2741 /*
2742  * The offload context needs to be set when we transfer the first
2743  * packet of a particular protocol (TCP/UDP).  This routine has been
2744  * enhanced to deal with inserted VLAN headers.
2745  *
2746  * If the new packet's ether header length, ip header length and
2747  * csum offloading type are same as the previous packet, we should
2748  * avoid allocating a new csum context descriptor; mainly to take
2749  * advantage of the pipeline effect of the TX data read request.
2750  *
2751  * This function returns number of TX descrptors allocated for
2752  * csum context.
2753  */
2754 static int
2755 em_txcsum(struct adapter *adapter, struct mbuf *mp,
2756           uint32_t *txd_upper, uint32_t *txd_lower)
2757 {
2758         struct e1000_context_desc *TXD;
2759         int curr_txd, ehdrlen, csum_flags;
2760         uint32_t cmd, hdr_len, ip_hlen;
2761
2762         csum_flags = mp->m_pkthdr.csum_flags & EM_CSUM_FEATURES;
2763         ip_hlen = mp->m_pkthdr.csum_iphlen;
2764         ehdrlen = mp->m_pkthdr.csum_lhlen;
2765
2766         if (adapter->csum_lhlen == ehdrlen &&
2767             adapter->csum_iphlen == ip_hlen &&
2768             adapter->csum_flags == csum_flags) {
2769                 /*
2770                  * Same csum offload context as the previous packets;
2771                  * just return.
2772                  */
2773                 *txd_upper = adapter->csum_txd_upper;
2774                 *txd_lower = adapter->csum_txd_lower;
2775                 return 0;
2776         }
2777
2778         /*
2779          * Setup a new csum offload context.
2780          */
2781
2782         curr_txd = adapter->next_avail_tx_desc;
2783         TXD = (struct e1000_context_desc *)&adapter->tx_desc_base[curr_txd];
2784
2785         cmd = 0;
2786
2787         /* Setup of IP header checksum. */
2788         if (csum_flags & CSUM_IP) {
2789                 /*
2790                  * Start offset for header checksum calculation.
2791                  * End offset for header checksum calculation.
2792                  * Offset of place to put the checksum.
2793                  */
2794                 TXD->lower_setup.ip_fields.ipcss = ehdrlen;
2795                 TXD->lower_setup.ip_fields.ipcse =
2796                     htole16(ehdrlen + ip_hlen - 1);
2797                 TXD->lower_setup.ip_fields.ipcso =
2798                     ehdrlen + offsetof(struct ip, ip_sum);
2799                 cmd |= E1000_TXD_CMD_IP;
2800                 *txd_upper |= E1000_TXD_POPTS_IXSM << 8;
2801         }
2802         hdr_len = ehdrlen + ip_hlen;
2803
2804         if (csum_flags & CSUM_TCP) {
2805                 /*
2806                  * Start offset for payload checksum calculation.
2807                  * End offset for payload checksum calculation.
2808                  * Offset of place to put the checksum.
2809                  */
2810                 TXD->upper_setup.tcp_fields.tucss = hdr_len;
2811                 TXD->upper_setup.tcp_fields.tucse = htole16(0);
2812                 TXD->upper_setup.tcp_fields.tucso =
2813                     hdr_len + offsetof(struct tcphdr, th_sum);
2814                 cmd |= E1000_TXD_CMD_TCP;
2815                 *txd_upper |= E1000_TXD_POPTS_TXSM << 8;
2816         } else if (csum_flags & CSUM_UDP) {
2817                 /*
2818                  * Start offset for header checksum calculation.
2819                  * End offset for header checksum calculation.
2820                  * Offset of place to put the checksum.
2821                  */
2822                 TXD->upper_setup.tcp_fields.tucss = hdr_len;
2823                 TXD->upper_setup.tcp_fields.tucse = htole16(0);
2824                 TXD->upper_setup.tcp_fields.tucso =
2825                     hdr_len + offsetof(struct udphdr, uh_sum);
2826                 *txd_upper |= E1000_TXD_POPTS_TXSM << 8;
2827         }
2828
2829         *txd_lower = E1000_TXD_CMD_DEXT |       /* Extended descr type */
2830                      E1000_TXD_DTYP_D;          /* Data descr */
2831
2832         /* Save the information for this csum offloading context */
2833         adapter->csum_lhlen = ehdrlen;
2834         adapter->csum_iphlen = ip_hlen;
2835         adapter->csum_flags = csum_flags;
2836         adapter->csum_txd_upper = *txd_upper;
2837         adapter->csum_txd_lower = *txd_lower;
2838
2839         TXD->tcp_seg_setup.data = htole32(0);
2840         TXD->cmd_and_length =
2841             htole32(E1000_TXD_CMD_IFCS | E1000_TXD_CMD_DEXT | cmd);
2842
2843         if (++curr_txd == adapter->num_tx_desc)
2844                 curr_txd = 0;
2845
2846         KKASSERT(adapter->num_tx_desc_avail > 0);
2847         adapter->num_tx_desc_avail--;
2848
2849         adapter->next_avail_tx_desc = curr_txd;
2850         return 1;
2851 }
2852
2853 static void
2854 em_txeof(struct adapter *adapter)
2855 {
2856         struct ifnet *ifp = &adapter->arpcom.ac_if;
2857         struct em_buffer *tx_buffer;
2858         int first, num_avail;
2859
2860         if (adapter->tx_dd_head == adapter->tx_dd_tail)
2861                 return;
2862
2863         if (adapter->num_tx_desc_avail == adapter->num_tx_desc)
2864                 return;
2865
2866         num_avail = adapter->num_tx_desc_avail;
2867         first = adapter->next_tx_to_clean;
2868
2869         while (adapter->tx_dd_head != adapter->tx_dd_tail) {
2870                 struct e1000_tx_desc *tx_desc;
2871                 int dd_idx = adapter->tx_dd[adapter->tx_dd_head];
2872
2873                 tx_desc = &adapter->tx_desc_base[dd_idx];
2874                 if (tx_desc->upper.fields.status & E1000_TXD_STAT_DD) {
2875                         EM_INC_TXDD_IDX(adapter->tx_dd_head);
2876
2877                         if (++dd_idx == adapter->num_tx_desc)
2878                                 dd_idx = 0;
2879
2880                         while (first != dd_idx) {
2881                                 logif(pkt_txclean);
2882
2883                                 num_avail++;
2884
2885                                 tx_buffer = &adapter->tx_buffer_area[first];
2886                                 if (tx_buffer->m_head) {
2887                                         ifp->if_opackets++;
2888                                         bus_dmamap_unload(adapter->txtag,
2889                                                           tx_buffer->map);
2890                                         m_freem(tx_buffer->m_head);
2891                                         tx_buffer->m_head = NULL;
2892                                 }
2893
2894                                 if (++first == adapter->num_tx_desc)
2895                                         first = 0;
2896                         }
2897                 } else {
2898                         break;
2899                 }
2900         }
2901         adapter->next_tx_to_clean = first;
2902         adapter->num_tx_desc_avail = num_avail;
2903
2904         if (adapter->tx_dd_head == adapter->tx_dd_tail) {
2905                 adapter->tx_dd_head = 0;
2906                 adapter->tx_dd_tail = 0;
2907         }
2908
2909         if (!EM_IS_OACTIVE(adapter)) {
2910                 ifp->if_flags &= ~IFF_OACTIVE;
2911
2912                 /* All clean, turn off the timer */
2913                 if (adapter->num_tx_desc_avail == adapter->num_tx_desc)
2914                         ifp->if_timer = 0;
2915         }
2916 }
2917
2918 static void
2919 em_tx_collect(struct adapter *adapter)
2920 {
2921         struct ifnet *ifp = &adapter->arpcom.ac_if;
2922         struct em_buffer *tx_buffer;
2923         int tdh, first, num_avail, dd_idx = -1;
2924
2925         if (adapter->num_tx_desc_avail == adapter->num_tx_desc)
2926                 return;
2927
2928         tdh = E1000_READ_REG(&adapter->hw, E1000_TDH(0));
2929         if (tdh == adapter->next_tx_to_clean)
2930                 return;
2931
2932         if (adapter->tx_dd_head != adapter->tx_dd_tail)
2933                 dd_idx = adapter->tx_dd[adapter->tx_dd_head];
2934
2935         num_avail = adapter->num_tx_desc_avail;
2936         first = adapter->next_tx_to_clean;
2937
2938         while (first != tdh) {
2939                 logif(pkt_txclean);
2940
2941                 num_avail++;
2942
2943                 tx_buffer = &adapter->tx_buffer_area[first];
2944                 if (tx_buffer->m_head) {
2945                         ifp->if_opackets++;
2946                         bus_dmamap_unload(adapter->txtag,
2947                                           tx_buffer->map);
2948                         m_freem(tx_buffer->m_head);
2949                         tx_buffer->m_head = NULL;
2950                 }
2951
2952                 if (first == dd_idx) {
2953                         EM_INC_TXDD_IDX(adapter->tx_dd_head);
2954                         if (adapter->tx_dd_head == adapter->tx_dd_tail) {
2955                                 adapter->tx_dd_head = 0;
2956                                 adapter->tx_dd_tail = 0;
2957                                 dd_idx = -1;
2958                         } else {
2959                                 dd_idx = adapter->tx_dd[adapter->tx_dd_head];
2960                         }
2961                 }
2962
2963                 if (++first == adapter->num_tx_desc)
2964                         first = 0;
2965         }
2966         adapter->next_tx_to_clean = first;
2967         adapter->num_tx_desc_avail = num_avail;
2968
2969         if (!EM_IS_OACTIVE(adapter)) {
2970                 ifp->if_flags &= ~IFF_OACTIVE;
2971
2972                 /* All clean, turn off the timer */
2973                 if (adapter->num_tx_desc_avail == adapter->num_tx_desc)
2974                         ifp->if_timer = 0;
2975         }
2976 }
2977
2978 /*
2979  * When Link is lost sometimes there is work still in the TX ring
2980  * which will result in a watchdog, rather than allow that do an
2981  * attempted cleanup and then reinit here.  Note that this has been
2982  * seens mostly with fiber adapters.
2983  */
2984 static void
2985 em_tx_purge(struct adapter *adapter)
2986 {
2987         struct ifnet *ifp = &adapter->arpcom.ac_if;
2988
2989         if (!adapter->link_active && ifp->if_timer) {
2990                 em_tx_collect(adapter);
2991                 if (ifp->if_timer) {
2992                         if_printf(ifp, "Link lost, TX pending, reinit\n");
2993                         ifp->if_timer = 0;
2994                         em_init(adapter);
2995                 }
2996         }
2997 }
2998
2999 static int
3000 em_newbuf(struct adapter *adapter, int i, int init)
3001 {
3002         struct mbuf *m;
3003         bus_dma_segment_t seg;
3004         bus_dmamap_t map;
3005         struct em_buffer *rx_buffer;
3006         int error, nseg;
3007
3008         m = m_getcl(init ? MB_WAIT : MB_DONTWAIT, MT_DATA, M_PKTHDR);
3009         if (m == NULL) {
3010                 adapter->mbuf_cluster_failed++;
3011                 if (init) {
3012                         if_printf(&adapter->arpcom.ac_if,
3013                                   "Unable to allocate RX mbuf\n");
3014                 }
3015                 return (ENOBUFS);
3016         }
3017         m->m_len = m->m_pkthdr.len = MCLBYTES;
3018
3019         if (adapter->max_frame_size <= MCLBYTES - ETHER_ALIGN)
3020                 m_adj(m, ETHER_ALIGN);
3021
3022         error = bus_dmamap_load_mbuf_segment(adapter->rxtag,
3023                         adapter->rx_sparemap, m,
3024                         &seg, 1, &nseg, BUS_DMA_NOWAIT);
3025         if (error) {
3026                 m_freem(m);
3027                 if (init) {
3028                         if_printf(&adapter->arpcom.ac_if,
3029                                   "Unable to load RX mbuf\n");
3030                 }
3031                 return (error);
3032         }
3033
3034         rx_buffer = &adapter->rx_buffer_area[i];
3035         if (rx_buffer->m_head != NULL)
3036                 bus_dmamap_unload(adapter->rxtag, rx_buffer->map);
3037
3038         map = rx_buffer->map;
3039         rx_buffer->map = adapter->rx_sparemap;
3040         adapter->rx_sparemap = map;
3041
3042         rx_buffer->m_head = m;
3043
3044         adapter->rx_desc_base[i].buffer_addr = htole64(seg.ds_addr);
3045         return (0);
3046 }
3047
3048 static int
3049 em_create_rx_ring(struct adapter *adapter)
3050 {
3051         device_t dev = adapter->dev;
3052         struct em_buffer *rx_buffer;
3053         int i, error;
3054
3055         adapter->rx_buffer_area =
3056                 kmalloc(sizeof(struct em_buffer) * adapter->num_rx_desc,
3057                         M_DEVBUF, M_WAITOK | M_ZERO);
3058
3059         /*
3060          * Create DMA tag for rx buffers
3061          */
3062         error = bus_dma_tag_create(adapter->parent_dtag, /* parent */
3063                         1, 0,                   /* alignment, bounds */
3064                         BUS_SPACE_MAXADDR,      /* lowaddr */
3065                         BUS_SPACE_MAXADDR,      /* highaddr */
3066                         NULL, NULL,             /* filter, filterarg */
3067                         MCLBYTES,               /* maxsize */
3068                         1,                      /* nsegments */
3069                         MCLBYTES,               /* maxsegsize */
3070                         BUS_DMA_WAITOK | BUS_DMA_ALLOCNOW, /* flags */
3071                         &adapter->rxtag);
3072         if (error) {
3073                 device_printf(dev, "Unable to allocate RX DMA tag\n");
3074                 kfree(adapter->rx_buffer_area, M_DEVBUF);
3075                 adapter->rx_buffer_area = NULL;
3076                 return error;
3077         }
3078
3079         /*
3080          * Create spare DMA map for rx buffers
3081          */
3082         error = bus_dmamap_create(adapter->rxtag, BUS_DMA_WAITOK,
3083                                   &adapter->rx_sparemap);
3084         if (error) {
3085                 device_printf(dev, "Unable to create spare RX DMA map\n");
3086                 bus_dma_tag_destroy(adapter->rxtag);
3087                 kfree(adapter->rx_buffer_area, M_DEVBUF);
3088                 adapter->rx_buffer_area = NULL;
3089                 return error;
3090         }
3091
3092         /*
3093          * Create DMA maps for rx buffers
3094          */
3095         for (i = 0; i < adapter->num_rx_desc; i++) {
3096                 rx_buffer = &adapter->rx_buffer_area[i];
3097
3098                 error = bus_dmamap_create(adapter->rxtag, BUS_DMA_WAITOK,
3099                                           &rx_buffer->map);
3100                 if (error) {
3101                         device_printf(dev, "Unable to create RX DMA map\n");
3102                         em_destroy_rx_ring(adapter, i);
3103                         return error;
3104                 }
3105         }
3106         return (0);
3107 }
3108
3109 static int
3110 em_init_rx_ring(struct adapter *adapter)
3111 {
3112         int i, error;
3113
3114         /* Reset descriptor ring */
3115         bzero(adapter->rx_desc_base,
3116             (sizeof(struct e1000_rx_desc)) * adapter->num_rx_desc);
3117
3118         /* Allocate new ones. */
3119         for (i = 0; i < adapter->num_rx_desc; i++) {
3120                 error = em_newbuf(adapter, i, 1);
3121                 if (error)
3122                         return (error);
3123         }
3124
3125         /* Setup our descriptor pointers */
3126         adapter->next_rx_desc_to_check = 0;
3127
3128         return (0);
3129 }
3130
3131 static void
3132 em_init_rx_unit(struct adapter *adapter)
3133 {
3134         struct ifnet *ifp = &adapter->arpcom.ac_if;
3135         uint64_t bus_addr;
3136         uint32_t rctl;
3137
3138         /*
3139          * Make sure receives are disabled while setting
3140          * up the descriptor ring
3141          */
3142         rctl = E1000_READ_REG(&adapter->hw, E1000_RCTL);
3143         E1000_WRITE_REG(&adapter->hw, E1000_RCTL, rctl & ~E1000_RCTL_EN);
3144
3145         if (adapter->hw.mac.type >= e1000_82540) {
3146                 uint32_t itr;
3147
3148                 /*
3149                  * Set the interrupt throttling rate. Value is calculated
3150                  * as ITR = 1 / (INT_THROTTLE_CEIL * 256ns)
3151                  */
3152                 if (adapter->int_throttle_ceil)
3153                         itr = 1000000000 / 256 / adapter->int_throttle_ceil;
3154                 else
3155                         itr = 0;
3156                 em_set_itr(adapter, itr);
3157         }
3158
3159         /* Disable accelerated ackknowledge */
3160         if (adapter->hw.mac.type == e1000_82574) {
3161                 E1000_WRITE_REG(&adapter->hw,
3162                     E1000_RFCTL, E1000_RFCTL_ACK_DIS);
3163         }
3164
3165         /* Receive Checksum Offload for TCP and UDP */
3166         if (ifp->if_capenable & IFCAP_RXCSUM) {
3167                 uint32_t rxcsum;
3168
3169                 rxcsum = E1000_READ_REG(&adapter->hw, E1000_RXCSUM);
3170                 rxcsum |= (E1000_RXCSUM_IPOFL | E1000_RXCSUM_TUOFL);
3171                 E1000_WRITE_REG(&adapter->hw, E1000_RXCSUM, rxcsum);
3172         }
3173
3174         /*
3175          * XXX TEMPORARY WORKAROUND: on some systems with 82573
3176          * long latencies are observed, like Lenovo X60. This
3177          * change eliminates the problem, but since having positive
3178          * values in RDTR is a known source of problems on other
3179          * platforms another solution is being sought.
3180          */
3181         if (em_82573_workaround && adapter->hw.mac.type == e1000_82573) {
3182                 E1000_WRITE_REG(&adapter->hw, E1000_RADV, EM_RADV_82573);
3183                 E1000_WRITE_REG(&adapter->hw, E1000_RDTR, EM_RDTR_82573);
3184         }
3185
3186         /*
3187          * Setup the Base and Length of the Rx Descriptor Ring
3188          */
3189         bus_addr = adapter->rxdma.dma_paddr;
3190         E1000_WRITE_REG(&adapter->hw, E1000_RDLEN(0),
3191             adapter->num_rx_desc * sizeof(struct e1000_rx_desc));
3192         E1000_WRITE_REG(&adapter->hw, E1000_RDBAH(0),
3193             (uint32_t)(bus_addr >> 32));
3194         E1000_WRITE_REG(&adapter->hw, E1000_RDBAL(0),
3195             (uint32_t)bus_addr);
3196
3197         /*
3198          * Setup the HW Rx Head and Tail Descriptor Pointers
3199          */
3200         E1000_WRITE_REG(&adapter->hw, E1000_RDH(0), 0);
3201         E1000_WRITE_REG(&adapter->hw, E1000_RDT(0), adapter->num_rx_desc - 1);
3202
3203         /* Set early receive threshold on appropriate hw */
3204         if (((adapter->hw.mac.type == e1000_ich9lan) ||
3205             (adapter->hw.mac.type == e1000_pch2lan) ||
3206             (adapter->hw.mac.type == e1000_ich10lan)) &&
3207             (ifp->if_mtu > ETHERMTU)) {
3208                 uint32_t rxdctl;
3209
3210                 rxdctl = E1000_READ_REG(&adapter->hw, E1000_RXDCTL(0));
3211                 E1000_WRITE_REG(&adapter->hw, E1000_RXDCTL(0), rxdctl | 3);
3212                 E1000_WRITE_REG(&adapter->hw, E1000_ERT, 0x100 | (1 << 13));
3213         }
3214
3215         if (adapter->hw.mac.type == e1000_pch2lan) {
3216                 if (ifp->if_mtu > ETHERMTU)
3217                         e1000_lv_jumbo_workaround_ich8lan(&adapter->hw, TRUE);
3218                 else
3219                         e1000_lv_jumbo_workaround_ich8lan(&adapter->hw, FALSE);
3220         }
3221
3222         /* Setup the Receive Control Register */
3223         rctl &= ~(3 << E1000_RCTL_MO_SHIFT);
3224         rctl |= E1000_RCTL_EN | E1000_RCTL_BAM | E1000_RCTL_LBM_NO |
3225                 E1000_RCTL_RDMTS_HALF |
3226                 (adapter->hw.mac.mc_filter_type << E1000_RCTL_MO_SHIFT);
3227
3228         /* Make sure VLAN Filters are off */
3229         rctl &= ~E1000_RCTL_VFE;
3230
3231         if (e1000_tbi_sbp_enabled_82543(&adapter->hw))
3232                 rctl |= E1000_RCTL_SBP;
3233         else
3234                 rctl &= ~E1000_RCTL_SBP;
3235
3236         switch (adapter->rx_buffer_len) {
3237         default:
3238         case 2048:
3239                 rctl |= E1000_RCTL_SZ_2048;
3240                 break;
3241
3242         case 4096:
3243                 rctl |= E1000_RCTL_SZ_4096 |
3244                     E1000_RCTL_BSEX | E1000_RCTL_LPE;
3245                 break;
3246
3247         case 8192:
3248                 rctl |= E1000_RCTL_SZ_8192 |
3249                     E1000_RCTL_BSEX | E1000_RCTL_LPE;
3250                 break;
3251
3252         case 16384:
3253                 rctl |= E1000_RCTL_SZ_16384 |
3254                     E1000_RCTL_BSEX | E1000_RCTL_LPE;
3255                 break;
3256         }
3257
3258         if (ifp->if_mtu > ETHERMTU)
3259                 rctl |= E1000_RCTL_LPE;
3260         else
3261                 rctl &= ~E1000_RCTL_LPE;
3262
3263         /* Enable Receives */
3264         E1000_WRITE_REG(&adapter->hw, E1000_RCTL, rctl);
3265 }
3266
3267 static void
3268 em_destroy_rx_ring(struct adapter *adapter, int ndesc)
3269 {
3270         struct em_buffer *rx_buffer;
3271         int i;
3272
3273         if (adapter->rx_buffer_area == NULL)
3274                 return;
3275
3276         for (i = 0; i < ndesc; i++) {
3277                 rx_buffer = &adapter->rx_buffer_area[i];
3278
3279                 KKASSERT(rx_buffer->m_head == NULL);
3280                 bus_dmamap_destroy(adapter->rxtag, rx_buffer->map);
3281         }
3282         bus_dmamap_destroy(adapter->rxtag, adapter->rx_sparemap);
3283         bus_dma_tag_destroy(adapter->rxtag);
3284
3285         kfree(adapter->rx_buffer_area, M_DEVBUF);
3286         adapter->rx_buffer_area = NULL;
3287 }
3288
3289 static void
3290 em_rxeof(struct adapter *adapter, int count)
3291 {
3292         struct ifnet *ifp = &adapter->arpcom.ac_if;
3293         uint8_t status, accept_frame = 0, eop = 0;
3294         uint16_t len, desc_len, prev_len_adj;
3295         struct e1000_rx_desc *current_desc;
3296         struct mbuf *mp;
3297         int i;
3298
3299         i = adapter->next_rx_desc_to_check;
3300         current_desc = &adapter->rx_desc_base[i];
3301
3302         if (!(current_desc->status & E1000_RXD_STAT_DD))
3303                 return;
3304
3305         while ((current_desc->status & E1000_RXD_STAT_DD) && count != 0) {
3306                 struct mbuf *m = NULL;
3307
3308                 logif(pkt_receive);
3309
3310                 mp = adapter->rx_buffer_area[i].m_head;
3311
3312                 /*
3313                  * Can't defer bus_dmamap_sync(9) because TBI_ACCEPT
3314                  * needs to access the last received byte in the mbuf.
3315                  */
3316                 bus_dmamap_sync(adapter->rxtag, adapter->rx_buffer_area[i].map,
3317                                 BUS_DMASYNC_POSTREAD);
3318
3319                 accept_frame = 1;
3320                 prev_len_adj = 0;
3321                 desc_len = le16toh(current_desc->length);
3322                 status = current_desc->status;
3323                 if (status & E1000_RXD_STAT_EOP) {
3324                         count--;
3325                         eop = 1;
3326                         if (desc_len < ETHER_CRC_LEN) {
3327                                 len = 0;
3328                                 prev_len_adj = ETHER_CRC_LEN - desc_len;
3329                         } else {
3330                                 len = desc_len - ETHER_CRC_LEN;
3331                         }
3332                 } else {
3333                         eop = 0;
3334                         len = desc_len;
3335                 }
3336
3337                 if (current_desc->errors & E1000_RXD_ERR_FRAME_ERR_MASK) {
3338                         uint8_t last_byte;
3339                         uint32_t pkt_len = desc_len;
3340
3341                         if (adapter->fmp != NULL)
3342                                 pkt_len += adapter->fmp->m_pkthdr.len;
3343
3344                         last_byte = *(mtod(mp, caddr_t) + desc_len - 1);
3345                         if (TBI_ACCEPT(&adapter->hw, status,
3346                             current_desc->errors, pkt_len, last_byte,
3347                             adapter->min_frame_size, adapter->max_frame_size)) {
3348                                 e1000_tbi_adjust_stats_82543(&adapter->hw,
3349                                     &adapter->stats, pkt_len,
3350                                     adapter->hw.mac.addr,
3351                                     adapter->max_frame_size);
3352                                 if (len > 0)
3353                                         len--;
3354                         } else {
3355                                 accept_frame = 0;
3356                         }
3357                 }
3358
3359                 if (accept_frame) {
3360                         if (em_newbuf(adapter, i, 0) != 0) {
3361                                 ifp->if_iqdrops++;
3362                                 goto discard;
3363                         }
3364
3365                         /* Assign correct length to the current fragment */
3366                         mp->m_len = len;
3367
3368                         if (adapter->fmp == NULL) {
3369                                 mp->m_pkthdr.len = len;
3370                                 adapter->fmp = mp; /* Store the first mbuf */
3371                                 adapter->lmp = mp;
3372                         } else {
3373                                 /*
3374                                  * Chain mbuf's together
3375                                  */
3376
3377                                 /*
3378                                  * Adjust length of previous mbuf in chain if
3379                                  * we received less than 4 bytes in the last
3380                                  * descriptor.
3381                                  */
3382                                 if (prev_len_adj > 0) {
3383                                         adapter->lmp->m_len -= prev_len_adj;
3384                                         adapter->fmp->m_pkthdr.len -=
3385                                             prev_len_adj;
3386                                 }
3387                                 adapter->lmp->m_next = mp;
3388                                 adapter->lmp = adapter->lmp->m_next;
3389                                 adapter->fmp->m_pkthdr.len += len;
3390                         }
3391
3392                         if (eop) {
3393                                 adapter->fmp->m_pkthdr.rcvif = ifp;
3394                                 ifp->if_ipackets++;
3395
3396                                 if (ifp->if_capenable & IFCAP_RXCSUM) {
3397                                         em_rxcsum(adapter, current_desc,
3398                                                   adapter->fmp);
3399                                 }
3400
3401                                 if (status & E1000_RXD_STAT_VP) {
3402                                         adapter->fmp->m_pkthdr.ether_vlantag =
3403                                             (le16toh(current_desc->special) &
3404                                             E1000_RXD_SPC_VLAN_MASK);
3405                                         adapter->fmp->m_flags |= M_VLANTAG;
3406                                 }
3407                                 m = adapter->fmp;
3408                                 adapter->fmp = NULL;
3409                                 adapter->lmp = NULL;
3410                         }
3411                 } else {
3412                         ifp->if_ierrors++;
3413 discard:
3414 #ifdef foo
3415                         /* Reuse loaded DMA map and just update mbuf chain */
3416                         mp = adapter->rx_buffer_area[i].m_head;
3417                         mp->m_len = mp->m_pkthdr.len = MCLBYTES;
3418                         mp->m_data = mp->m_ext.ext_buf;
3419                         mp->m_next = NULL;
3420                         if (adapter->max_frame_size <= (MCLBYTES - ETHER_ALIGN))
3421                                 m_adj(mp, ETHER_ALIGN);
3422 #endif
3423                         if (adapter->fmp != NULL) {
3424                                 m_freem(adapter->fmp);
3425                                 adapter->fmp = NULL;
3426                                 adapter->lmp = NULL;
3427                         }
3428                         m = NULL;
3429                 }
3430
3431                 /* Zero out the receive descriptors status. */
3432                 current_desc->status = 0;
3433
3434                 if (m != NULL)
3435                         ifp->if_input(ifp, m);
3436
3437                 /* Advance our pointers to the next descriptor. */
3438                 if (++i == adapter->num_rx_desc)
3439                         i = 0;
3440                 current_desc = &adapter->rx_desc_base[i];
3441         }
3442         adapter->next_rx_desc_to_check = i;
3443
3444         /* Advance the E1000's Receive Queue #0  "Tail Pointer". */
3445         if (--i < 0)
3446                 i = adapter->num_rx_desc - 1;
3447         E1000_WRITE_REG(&adapter->hw, E1000_RDT(0), i);
3448 }
3449
3450 static void
3451 em_rxcsum(struct adapter *adapter, struct e1000_rx_desc *rx_desc,
3452           struct mbuf *mp)
3453 {
3454         /* 82543 or newer only */
3455         if (adapter->hw.mac.type < e1000_82543 ||
3456             /* Ignore Checksum bit is set */
3457             (rx_desc->status & E1000_RXD_STAT_IXSM))
3458                 return;
3459
3460         if ((rx_desc->status & E1000_RXD_STAT_IPCS) &&
3461             !(rx_desc->errors & E1000_RXD_ERR_IPE)) {
3462                 /* IP Checksum Good */
3463                 mp->m_pkthdr.csum_flags |= CSUM_IP_CHECKED | CSUM_IP_VALID;
3464         }
3465
3466         if ((rx_desc->status & E1000_RXD_STAT_TCPCS) &&
3467             !(rx_desc->errors & E1000_RXD_ERR_TCPE)) {
3468                 mp->m_pkthdr.csum_flags |= CSUM_DATA_VALID |
3469                                            CSUM_PSEUDO_HDR |
3470                                            CSUM_FRAG_NOT_CHECKED;
3471                 mp->m_pkthdr.csum_data = htons(0xffff);
3472         }
3473 }
3474
3475 static void
3476 em_enable_intr(struct adapter *adapter)
3477 {
3478         uint32_t ims_mask = IMS_ENABLE_MASK;
3479
3480         lwkt_serialize_handler_enable(adapter->arpcom.ac_if.if_serializer);
3481
3482 #if 0
3483         /* XXX MSIX */
3484         if (adapter->hw.mac.type == e1000_82574) {
3485                 E1000_WRITE_REG(&adapter->hw, EM_EIAC, EM_MSIX_MASK);
3486                 ims_mask |= EM_MSIX_MASK;
3487         }
3488 #endif
3489         E1000_WRITE_REG(&adapter->hw, E1000_IMS, ims_mask);
3490 }
3491
3492 static void
3493 em_disable_intr(struct adapter *adapter)
3494 {
3495         uint32_t clear = 0xffffffff;
3496
3497         /*
3498          * The first version of 82542 had an errata where when link was forced
3499          * it would stay up even up even if the cable was disconnected.
3500          * Sequence errors were used to detect the disconnect and then the
3501          * driver would unforce the link.  This code in the in the ISR.  For
3502          * this to work correctly the Sequence error interrupt had to be
3503          * enabled all the time.
3504          */
3505         if (adapter->hw.mac.type == e1000_82542 &&
3506             adapter->hw.revision_id == E1000_REVISION_2)
3507                 clear &= ~E1000_ICR_RXSEQ;
3508         else if (adapter->hw.mac.type == e1000_82574)
3509                 E1000_WRITE_REG(&adapter->hw, EM_EIAC, 0);
3510
3511         E1000_WRITE_REG(&adapter->hw, E1000_IMC, clear);
3512
3513         lwkt_serialize_handler_disable(adapter->arpcom.ac_if.if_serializer);
3514 }
3515
3516 /*
3517  * Bit of a misnomer, what this really means is
3518  * to enable OS management of the system... aka
3519  * to disable special hardware management features
3520  */
3521 static void
3522 em_get_mgmt(struct adapter *adapter)
3523 {
3524         /* A shared code workaround */
3525 #define E1000_82542_MANC2H E1000_MANC2H
3526         if (adapter->flags & EM_FLAG_HAS_MGMT) {
3527                 int manc2h = E1000_READ_REG(&adapter->hw, E1000_MANC2H);
3528                 int manc = E1000_READ_REG(&adapter->hw, E1000_MANC);
3529
3530                 /* disable hardware interception of ARP */
3531                 manc &= ~(E1000_MANC_ARP_EN);
3532
3533                 /* enable receiving management packets to the host */
3534                 if (adapter->hw.mac.type >= e1000_82571) {
3535                         manc |= E1000_MANC_EN_MNG2HOST;
3536 #define E1000_MNG2HOST_PORT_623 (1 << 5)
3537 #define E1000_MNG2HOST_PORT_664 (1 << 6)
3538                         manc2h |= E1000_MNG2HOST_PORT_623;
3539                         manc2h |= E1000_MNG2HOST_PORT_664;
3540                         E1000_WRITE_REG(&adapter->hw, E1000_MANC2H, manc2h);
3541                 }
3542
3543                 E1000_WRITE_REG(&adapter->hw, E1000_MANC, manc);
3544         }
3545 }
3546
3547 /*
3548  * Give control back to hardware management
3549  * controller if there is one.
3550  */
3551 static void
3552 em_rel_mgmt(struct adapter *adapter)
3553 {
3554         if (adapter->flags & EM_FLAG_HAS_MGMT) {
3555                 int manc = E1000_READ_REG(&adapter->hw, E1000_MANC);
3556
3557                 /* re-enable hardware interception of ARP */
3558                 manc |= E1000_MANC_ARP_EN;
3559
3560                 if (adapter->hw.mac.type >= e1000_82571)
3561                         manc &= ~E1000_MANC_EN_MNG2HOST;
3562
3563                 E1000_WRITE_REG(&adapter->hw, E1000_MANC, manc);
3564         }
3565 }
3566
3567 /*
3568  * em_get_hw_control() sets {CTRL_EXT|FWSM}:DRV_LOAD bit.
3569  * For ASF and Pass Through versions of f/w this means that
3570  * the driver is loaded.  For AMT version (only with 82573)
3571  * of the f/w this means that the network i/f is open.
3572  */
3573 static void
3574 em_get_hw_control(struct adapter *adapter)
3575 {
3576         /* Let firmware know the driver has taken over */
3577         if (adapter->hw.mac.type == e1000_82573) {
3578                 uint32_t swsm;
3579
3580                 swsm = E1000_READ_REG(&adapter->hw, E1000_SWSM);
3581                 E1000_WRITE_REG(&adapter->hw, E1000_SWSM,
3582                     swsm | E1000_SWSM_DRV_LOAD);
3583         } else {
3584                 uint32_t ctrl_ext;
3585
3586                 ctrl_ext = E1000_READ_REG(&adapter->hw, E1000_CTRL_EXT);
3587                 E1000_WRITE_REG(&adapter->hw, E1000_CTRL_EXT,
3588                     ctrl_ext | E1000_CTRL_EXT_DRV_LOAD);
3589         }
3590         adapter->flags |= EM_FLAG_HW_CTRL;
3591 }
3592
3593 /*
3594  * em_rel_hw_control() resets {CTRL_EXT|FWSM}:DRV_LOAD bit.
3595  * For ASF and Pass Through versions of f/w this means that the
3596  * driver is no longer loaded.  For AMT version (only with 82573)
3597  * of the f/w this means that the network i/f is closed.
3598  */
3599 static void
3600 em_rel_hw_control(struct adapter *adapter)
3601 {
3602         if ((adapter->flags & EM_FLAG_HW_CTRL) == 0)
3603                 return;
3604         adapter->flags &= ~EM_FLAG_HW_CTRL;
3605
3606         /* Let firmware taken over control of h/w */
3607         if (adapter->hw.mac.type == e1000_82573) {
3608                 uint32_t swsm;
3609
3610                 swsm = E1000_READ_REG(&adapter->hw, E1000_SWSM);
3611                 E1000_WRITE_REG(&adapter->hw, E1000_SWSM,
3612                     swsm & ~E1000_SWSM_DRV_LOAD);
3613         } else {
3614                 uint32_t ctrl_ext;
3615
3616                 ctrl_ext = E1000_READ_REG(&adapter->hw, E1000_CTRL_EXT);
3617                 E1000_WRITE_REG(&adapter->hw, E1000_CTRL_EXT,
3618                     ctrl_ext & ~E1000_CTRL_EXT_DRV_LOAD);
3619         }
3620 }
3621
3622 static int
3623 em_is_valid_eaddr(const uint8_t *addr)
3624 {
3625         char zero_addr[ETHER_ADDR_LEN] = { 0, 0, 0, 0, 0, 0 };
3626
3627         if ((addr[0] & 1) || !bcmp(addr, zero_addr, ETHER_ADDR_LEN))
3628                 return (FALSE);
3629
3630         return (TRUE);
3631 }
3632
3633 /*
3634  * Enable PCI Wake On Lan capability
3635  */
3636 void
3637 em_enable_wol(device_t dev)
3638 {
3639         uint16_t cap, status;
3640         uint8_t id;
3641
3642         /* First find the capabilities pointer*/
3643         cap = pci_read_config(dev, PCIR_CAP_PTR, 2);
3644
3645         /* Read the PM Capabilities */
3646         id = pci_read_config(dev, cap, 1);
3647         if (id != PCIY_PMG)     /* Something wrong */
3648                 return;
3649
3650         /*
3651          * OK, we have the power capabilities,
3652          * so now get the status register
3653          */
3654         cap += PCIR_POWER_STATUS;
3655         status = pci_read_config(dev, cap, 2);
3656         status |= PCIM_PSTAT_PME | PCIM_PSTAT_PMEENABLE;
3657         pci_write_config(dev, cap, status, 2);
3658 }
3659
3660
3661 /*
3662  * 82544 Coexistence issue workaround.
3663  *    There are 2 issues.
3664  *       1. Transmit Hang issue.
3665  *    To detect this issue, following equation can be used...
3666  *        SIZE[3:0] + ADDR[2:0] = SUM[3:0].
3667  *        If SUM[3:0] is in between 1 to 4, we will have this issue.
3668  *
3669  *       2. DAC issue.
3670  *    To detect this issue, following equation can be used...
3671  *        SIZE[3:0] + ADDR[2:0] = SUM[3:0].
3672  *        If SUM[3:0] is in between 9 to c, we will have this issue.
3673  *
3674  *    WORKAROUND:
3675  *        Make sure we do not have ending address
3676  *        as 1,2,3,4(Hang) or 9,a,b,c (DAC)
3677  */
3678 static uint32_t
3679 em_82544_fill_desc(bus_addr_t address, uint32_t length, PDESC_ARRAY desc_array)
3680 {
3681         uint32_t safe_terminator;
3682
3683         /*
3684          * Since issue is sensitive to length and address.
3685          * Let us first check the address...
3686          */
3687         if (length <= 4) {
3688                 desc_array->descriptor[0].address = address;
3689                 desc_array->descriptor[0].length = length;
3690                 desc_array->elements = 1;
3691                 return (desc_array->elements);
3692         }
3693
3694         safe_terminator =
3695         (uint32_t)((((uint32_t)address & 0x7) + (length & 0xF)) & 0xF);
3696
3697         /* If it does not fall between 0x1 to 0x4 and 0x9 to 0xC then return */
3698         if (safe_terminator == 0 ||
3699             (safe_terminator > 4 && safe_terminator < 9) ||
3700             (safe_terminator > 0xC && safe_terminator <= 0xF)) {
3701                 desc_array->descriptor[0].address = address;
3702                 desc_array->descriptor[0].length = length;
3703                 desc_array->elements = 1;
3704                 return (desc_array->elements);
3705         }
3706
3707         desc_array->descriptor[0].address = address;
3708         desc_array->descriptor[0].length = length - 4;
3709         desc_array->descriptor[1].address = address + (length - 4);
3710         desc_array->descriptor[1].length = 4;
3711         desc_array->elements = 2;
3712         return (desc_array->elements);
3713 }
3714
3715 static void
3716 em_update_stats(struct adapter *adapter)
3717 {
3718         struct ifnet *ifp = &adapter->arpcom.ac_if;
3719
3720         if (adapter->hw.phy.media_type == e1000_media_type_copper ||
3721             (E1000_READ_REG(&adapter->hw, E1000_STATUS) & E1000_STATUS_LU)) {
3722                 adapter->stats.symerrs +=
3723                         E1000_READ_REG(&adapter->hw, E1000_SYMERRS);
3724                 adapter->stats.sec += E1000_READ_REG(&adapter->hw, E1000_SEC);
3725         }
3726         adapter->stats.crcerrs += E1000_READ_REG(&adapter->hw, E1000_CRCERRS);
3727         adapter->stats.mpc += E1000_READ_REG(&adapter->hw, E1000_MPC);
3728         adapter->stats.scc += E1000_READ_REG(&adapter->hw, E1000_SCC);
3729         adapter->stats.ecol += E1000_READ_REG(&adapter->hw, E1000_ECOL);
3730
3731         adapter->stats.mcc += E1000_READ_REG(&adapter->hw, E1000_MCC);
3732         adapter->stats.latecol += E1000_READ_REG(&adapter->hw, E1000_LATECOL);
3733         adapter->stats.colc += E1000_READ_REG(&adapter->hw, E1000_COLC);
3734         adapter->stats.dc += E1000_READ_REG(&adapter->hw, E1000_DC);
3735         adapter->stats.rlec += E1000_READ_REG(&adapter->hw, E1000_RLEC);
3736         adapter->stats.xonrxc += E1000_READ_REG(&adapter->hw, E1000_XONRXC);
3737         adapter->stats.xontxc += E1000_READ_REG(&adapter->hw, E1000_XONTXC);
3738         adapter->stats.xoffrxc += E1000_READ_REG(&adapter->hw, E1000_XOFFRXC);
3739         adapter->stats.xofftxc += E1000_READ_REG(&adapter->hw, E1000_XOFFTXC);
3740         adapter->stats.fcruc += E1000_READ_REG(&adapter->hw, E1000_FCRUC);
3741         adapter->stats.prc64 += E1000_READ_REG(&adapter->hw, E1000_PRC64);
3742         adapter->stats.prc127 += E1000_READ_REG(&adapter->hw, E1000_PRC127);
3743         adapter->stats.prc255 += E1000_READ_REG(&adapter->hw, E1000_PRC255);
3744         adapter->stats.prc511 += E1000_READ_REG(&adapter->hw, E1000_PRC511);
3745         adapter->stats.prc1023 += E1000_READ_REG(&adapter->hw, E1000_PRC1023);
3746         adapter->stats.prc1522 += E1000_READ_REG(&adapter->hw, E1000_PRC1522);
3747         adapter->stats.gprc += E1000_READ_REG(&adapter->hw, E1000_GPRC);
3748         adapter->stats.bprc += E1000_READ_REG(&adapter->hw, E1000_BPRC);
3749         adapter->stats.mprc += E1000_READ_REG(&adapter->hw, E1000_MPRC);
3750         adapter->stats.gptc += E1000_READ_REG(&adapter->hw, E1000_GPTC);
3751
3752         /* For the 64-bit byte counters the low dword must be read first. */
3753         /* Both registers clear on the read of the high dword */
3754
3755         adapter->stats.gorc += E1000_READ_REG(&adapter->hw, E1000_GORCH);
3756         adapter->stats.gotc += E1000_READ_REG(&adapter->hw, E1000_GOTCH);
3757
3758         adapter->stats.rnbc += E1000_READ_REG(&adapter->hw, E1000_RNBC);
3759         adapter->stats.ruc += E1000_READ_REG(&adapter->hw, E1000_RUC);
3760         adapter->stats.rfc += E1000_READ_REG(&adapter->hw, E1000_RFC);
3761         adapter->stats.roc += E1000_READ_REG(&adapter->hw, E1000_ROC);
3762         adapter->stats.rjc += E1000_READ_REG(&adapter->hw, E1000_RJC);
3763
3764         adapter->stats.tor += E1000_READ_REG(&adapter->hw, E1000_TORH);
3765         adapter->stats.tot += E1000_READ_REG(&adapter->hw, E1000_TOTH);
3766
3767         adapter->stats.tpr += E1000_READ_REG(&adapter->hw, E1000_TPR);
3768         adapter->stats.tpt += E1000_READ_REG(&adapter->hw, E1000_TPT);
3769         adapter->stats.ptc64 += E1000_READ_REG(&adapter->hw, E1000_PTC64);
3770         adapter->stats.ptc127 += E1000_READ_REG(&adapter->hw, E1000_PTC127);
3771         adapter->stats.ptc255 += E1000_READ_REG(&adapter->hw, E1000_PTC255);
3772         adapter->stats.ptc511 += E1000_READ_REG(&adapter->hw, E1000_PTC511);
3773         adapter->stats.ptc1023 += E1000_READ_REG(&adapter->hw, E1000_PTC1023);
3774         adapter->stats.ptc1522 += E1000_READ_REG(&adapter->hw, E1000_PTC1522);
3775         adapter->stats.mptc += E1000_READ_REG(&adapter->hw, E1000_MPTC);
3776         adapter->stats.bptc += E1000_READ_REG(&adapter->hw, E1000_BPTC);
3777
3778         if (adapter->hw.mac.type >= e1000_82543) {
3779                 adapter->stats.algnerrc +=
3780                 E1000_READ_REG(&adapter->hw, E1000_ALGNERRC);
3781                 adapter->stats.rxerrc +=
3782                 E1000_READ_REG(&adapter->hw, E1000_RXERRC);
3783                 adapter->stats.tncrs +=
3784                 E1000_READ_REG(&adapter->hw, E1000_TNCRS);
3785                 adapter->stats.cexterr +=
3786                 E1000_READ_REG(&adapter->hw, E1000_CEXTERR);
3787                 adapter->stats.tsctc +=
3788                 E1000_READ_REG(&adapter->hw, E1000_TSCTC);
3789                 adapter->stats.tsctfc +=
3790                 E1000_READ_REG(&adapter->hw, E1000_TSCTFC);
3791         }
3792
3793         ifp->if_collisions = adapter->stats.colc;
3794
3795         /* Rx Errors */
3796         ifp->if_ierrors =
3797             adapter->dropped_pkts + adapter->stats.rxerrc +
3798             adapter->stats.crcerrs + adapter->stats.algnerrc +
3799             adapter->stats.ruc + adapter->stats.roc +
3800             adapter->stats.mpc + adapter->stats.cexterr;
3801
3802         /* Tx Errors */
3803         ifp->if_oerrors =
3804             adapter->stats.ecol + adapter->stats.latecol +
3805             adapter->watchdog_events;
3806 }
3807
3808 static void
3809 em_print_debug_info(struct adapter *adapter)
3810 {
3811         device_t dev = adapter->dev;
3812         uint8_t *hw_addr = adapter->hw.hw_addr;
3813
3814         device_printf(dev, "Adapter hardware address = %p \n", hw_addr);
3815         device_printf(dev, "CTRL = 0x%x RCTL = 0x%x \n",
3816             E1000_READ_REG(&adapter->hw, E1000_CTRL),
3817             E1000_READ_REG(&adapter->hw, E1000_RCTL));
3818         device_printf(dev, "Packet buffer = Tx=%dk Rx=%dk \n",
3819             ((E1000_READ_REG(&adapter->hw, E1000_PBA) & 0xffff0000) >> 16),\
3820             (E1000_READ_REG(&adapter->hw, E1000_PBA) & 0xffff) );
3821         device_printf(dev, "Flow control watermarks high = %d low = %d\n",
3822             adapter->hw.fc.high_water,
3823             adapter->hw.fc.low_water);
3824         device_printf(dev, "tx_int_delay = %d, tx_abs_int_delay = %d\n",
3825             E1000_READ_REG(&adapter->hw, E1000_TIDV),
3826             E1000_READ_REG(&adapter->hw, E1000_TADV));
3827         device_printf(dev, "rx_int_delay = %d, rx_abs_int_delay = %d\n",
3828             E1000_READ_REG(&adapter->hw, E1000_RDTR),
3829             E1000_READ_REG(&adapter->hw, E1000_RADV));
3830         device_printf(dev, "fifo workaround = %lld, fifo_reset_count = %lld\n",
3831             (long long)adapter->tx_fifo_wrk_cnt,
3832             (long long)adapter->tx_fifo_reset_cnt);
3833         device_printf(dev, "hw tdh = %d, hw tdt = %d\n",
3834             E1000_READ_REG(&adapter->hw, E1000_TDH(0)),
3835             E1000_READ_REG(&adapter->hw, E1000_TDT(0)));
3836         device_printf(dev, "hw rdh = %d, hw rdt = %d\n",
3837             E1000_READ_REG(&adapter->hw, E1000_RDH(0)),
3838             E1000_READ_REG(&adapter->hw, E1000_RDT(0)));
3839         device_printf(dev, "Num Tx descriptors avail = %d\n",
3840             adapter->num_tx_desc_avail);
3841         device_printf(dev, "Tx Descriptors not avail1 = %ld\n",
3842             adapter->no_tx_desc_avail1);
3843         device_printf(dev, "Tx Descriptors not avail2 = %ld\n",
3844             adapter->no_tx_desc_avail2);
3845         device_printf(dev, "Std mbuf failed = %ld\n",
3846             adapter->mbuf_alloc_failed);
3847         device_printf(dev, "Std mbuf cluster failed = %ld\n",
3848             adapter->mbuf_cluster_failed);
3849         device_printf(dev, "Driver dropped packets = %ld\n",
3850             adapter->dropped_pkts);
3851         device_printf(dev, "Driver tx dma failure in encap = %ld\n",
3852             adapter->no_tx_dma_setup);
3853 }
3854
3855 static void
3856 em_print_hw_stats(struct adapter *adapter)
3857 {
3858         device_t dev = adapter->dev;
3859
3860         device_printf(dev, "Excessive collisions = %lld\n",
3861             (long long)adapter->stats.ecol);
3862 #if (DEBUG_HW > 0)  /* Dont output these errors normally */
3863         device_printf(dev, "Symbol errors = %lld\n",
3864             (long long)adapter->stats.symerrs);
3865 #endif
3866         device_printf(dev, "Sequence errors = %lld\n",
3867             (long long)adapter->stats.sec);
3868         device_printf(dev, "Defer count = %lld\n",
3869             (long long)adapter->stats.dc);
3870         device_printf(dev, "Missed Packets = %lld\n",
3871             (long long)adapter->stats.mpc);
3872         device_printf(dev, "Receive No Buffers = %lld\n",
3873             (long long)adapter->stats.rnbc);
3874         /* RLEC is inaccurate on some hardware, calculate our own. */
3875         device_printf(dev, "Receive Length Errors = %lld\n",
3876             ((long long)adapter->stats.roc + (long long)adapter->stats.ruc));
3877         device_printf(dev, "Receive errors = %lld\n",
3878             (long long)adapter->stats.rxerrc);
3879         device_printf(dev, "Crc errors = %lld\n",
3880             (long long)adapter->stats.crcerrs);
3881         device_printf(dev, "Alignment errors = %lld\n",
3882             (long long)adapter->stats.algnerrc);
3883         device_printf(dev, "Collision/Carrier extension errors = %lld\n",
3884             (long long)adapter->stats.cexterr);
3885         device_printf(dev, "RX overruns = %ld\n", adapter->rx_overruns);
3886         device_printf(dev, "watchdog timeouts = %ld\n",
3887             adapter->watchdog_events);
3888         device_printf(dev, "XON Rcvd = %lld\n",
3889             (long long)adapter->stats.xonrxc);
3890         device_printf(dev, "XON Xmtd = %lld\n",
3891             (long long)adapter->stats.xontxc);
3892         device_printf(dev, "XOFF Rcvd = %lld\n",
3893             (long long)adapter->stats.xoffrxc);
3894         device_printf(dev, "XOFF Xmtd = %lld\n",
3895             (long long)adapter->stats.xofftxc);
3896         device_printf(dev, "Good Packets Rcvd = %lld\n",
3897             (long long)adapter->stats.gprc);
3898         device_printf(dev, "Good Packets Xmtd = %lld\n",
3899             (long long)adapter->stats.gptc);
3900 }
3901
3902 static void
3903 em_print_nvm_info(struct adapter *adapter)
3904 {
3905         uint16_t eeprom_data;
3906         int i, j, row = 0;
3907
3908         /* Its a bit crude, but it gets the job done */
3909         kprintf("\nInterface EEPROM Dump:\n");
3910         kprintf("Offset\n0x0000  ");
3911         for (i = 0, j = 0; i < 32; i++, j++) {
3912                 if (j == 8) { /* Make the offset block */
3913                         j = 0; ++row;
3914                         kprintf("\n0x00%x0  ",row);
3915                 }
3916                 e1000_read_nvm(&adapter->hw, i, 1, &eeprom_data);
3917                 kprintf("%04x ", eeprom_data);
3918         }
3919         kprintf("\n");
3920 }
3921
3922 static int
3923 em_sysctl_debug_info(SYSCTL_HANDLER_ARGS)
3924 {
3925         struct adapter *adapter;
3926         struct ifnet *ifp;
3927         int error, result;
3928
3929         result = -1;
3930         error = sysctl_handle_int(oidp, &result, 0, req);
3931         if (error || !req->newptr)
3932                 return (error);
3933
3934         adapter = (struct adapter *)arg1;
3935         ifp = &adapter->arpcom.ac_if;
3936
3937         lwkt_serialize_enter(ifp->if_serializer);
3938
3939         if (result == 1)
3940                 em_print_debug_info(adapter);
3941
3942         /*
3943          * This value will cause a hex dump of the
3944          * first 32 16-bit words of the EEPROM to
3945          * the screen.
3946          */
3947         if (result == 2)
3948                 em_print_nvm_info(adapter);
3949
3950         lwkt_serialize_exit(ifp->if_serializer);
3951
3952         return (error);
3953 }
3954
3955 static int
3956 em_sysctl_stats(SYSCTL_HANDLER_ARGS)
3957 {
3958         int error, result;
3959
3960         result = -1;
3961         error = sysctl_handle_int(oidp, &result, 0, req);
3962         if (error || !req->newptr)
3963                 return (error);
3964
3965         if (result == 1) {
3966                 struct adapter *adapter = (struct adapter *)arg1;
3967                 struct ifnet *ifp = &adapter->arpcom.ac_if;
3968
3969                 lwkt_serialize_enter(ifp->if_serializer);
3970                 em_print_hw_stats(adapter);
3971                 lwkt_serialize_exit(ifp->if_serializer);
3972         }
3973         return (error);
3974 }
3975
3976 static void
3977 em_add_sysctl(struct adapter *adapter)
3978 {
3979         sysctl_ctx_init(&adapter->sysctl_ctx);
3980         adapter->sysctl_tree = SYSCTL_ADD_NODE(&adapter->sysctl_ctx,
3981                                         SYSCTL_STATIC_CHILDREN(_hw), OID_AUTO,
3982                                         device_get_nameunit(adapter->dev),
3983                                         CTLFLAG_RD, 0, "");
3984         if (adapter->sysctl_tree == NULL) {
3985                 device_printf(adapter->dev, "can't add sysctl node\n");
3986         } else {
3987                 SYSCTL_ADD_PROC(&adapter->sysctl_ctx,
3988                     SYSCTL_CHILDREN(adapter->sysctl_tree),
3989                     OID_AUTO, "debug", CTLTYPE_INT|CTLFLAG_RW, adapter, 0,
3990                     em_sysctl_debug_info, "I", "Debug Information");
3991
3992                 SYSCTL_ADD_PROC(&adapter->sysctl_ctx,
3993                     SYSCTL_CHILDREN(adapter->sysctl_tree),
3994                     OID_AUTO, "stats", CTLTYPE_INT|CTLFLAG_RW, adapter, 0,
3995                     em_sysctl_stats, "I", "Statistics");
3996
3997                 SYSCTL_ADD_INT(&adapter->sysctl_ctx,
3998                     SYSCTL_CHILDREN(adapter->sysctl_tree),
3999                     OID_AUTO, "rxd", CTLFLAG_RD,
4000                     &adapter->num_rx_desc, 0, NULL);
4001                 SYSCTL_ADD_INT(&adapter->sysctl_ctx,
4002                     SYSCTL_CHILDREN(adapter->sysctl_tree),
4003                     OID_AUTO, "txd", CTLFLAG_RD,
4004                     &adapter->num_tx_desc, 0, NULL);
4005
4006                 if (adapter->hw.mac.type >= e1000_82540) {
4007                         SYSCTL_ADD_PROC(&adapter->sysctl_ctx,
4008                             SYSCTL_CHILDREN(adapter->sysctl_tree),
4009                             OID_AUTO, "int_throttle_ceil",
4010                             CTLTYPE_INT|CTLFLAG_RW, adapter, 0,
4011                             em_sysctl_int_throttle, "I",
4012                             "interrupt throttling rate");
4013                 }
4014                 SYSCTL_ADD_PROC(&adapter->sysctl_ctx,
4015                     SYSCTL_CHILDREN(adapter->sysctl_tree),
4016                     OID_AUTO, "int_tx_nsegs",
4017                     CTLTYPE_INT|CTLFLAG_RW, adapter, 0,
4018                     em_sysctl_int_tx_nsegs, "I",
4019                     "# segments per TX interrupt");
4020         }
4021 }
4022
4023 static int
4024 em_sysctl_int_throttle(SYSCTL_HANDLER_ARGS)
4025 {
4026         struct adapter *adapter = (void *)arg1;
4027         struct ifnet *ifp = &adapter->arpcom.ac_if;
4028         int error, throttle;
4029
4030         throttle = adapter->int_throttle_ceil;
4031         error = sysctl_handle_int(oidp, &throttle, 0, req);
4032         if (error || req->newptr == NULL)
4033                 return error;
4034         if (throttle < 0 || throttle > 1000000000 / 256)
4035                 return EINVAL;
4036
4037         if (throttle) {
4038                 /*
4039                  * Set the interrupt throttling rate in 256ns increments,
4040                  * recalculate sysctl value assignment to get exact frequency.
4041                  */
4042                 throttle = 1000000000 / 256 / throttle;
4043
4044                 /* Upper 16bits of ITR is reserved and should be zero */
4045                 if (throttle & 0xffff0000)
4046                         return EINVAL;
4047         }
4048
4049         lwkt_serialize_enter(ifp->if_serializer);
4050
4051         if (throttle)
4052                 adapter->int_throttle_ceil = 1000000000 / 256 / throttle;
4053         else
4054                 adapter->int_throttle_ceil = 0;
4055
4056         if (ifp->if_flags & IFF_RUNNING)
4057                 em_set_itr(adapter, throttle);
4058
4059         lwkt_serialize_exit(ifp->if_serializer);
4060
4061         if (bootverbose) {
4062                 if_printf(ifp, "Interrupt moderation set to %d/sec\n",
4063                           adapter->int_throttle_ceil);
4064         }
4065         return 0;
4066 }
4067
4068 static int
4069 em_sysctl_int_tx_nsegs(SYSCTL_HANDLER_ARGS)
4070 {
4071         struct adapter *adapter = (void *)arg1;
4072         struct ifnet *ifp = &adapter->arpcom.ac_if;
4073         int error, segs;
4074
4075         segs = adapter->tx_int_nsegs;
4076         error = sysctl_handle_int(oidp, &segs, 0, req);
4077         if (error || req->newptr == NULL)
4078                 return error;
4079         if (segs <= 0)
4080                 return EINVAL;
4081
4082         lwkt_serialize_enter(ifp->if_serializer);
4083
4084         /*
4085          * Don't allow int_tx_nsegs to become:
4086          * o  Less the oact_tx_desc
4087          * o  Too large that no TX desc will cause TX interrupt to
4088          *    be generated (OACTIVE will never recover)
4089          * o  Too small that will cause tx_dd[] overflow
4090          */
4091         if (segs < adapter->oact_tx_desc ||
4092             segs >= adapter->num_tx_desc - adapter->oact_tx_desc ||
4093             segs < adapter->num_tx_desc / EM_TXDD_SAFE) {
4094                 error = EINVAL;
4095         } else {
4096                 error = 0;
4097                 adapter->tx_int_nsegs = segs;
4098         }
4099
4100         lwkt_serialize_exit(ifp->if_serializer);
4101
4102         return error;
4103 }
4104
4105 static void
4106 em_set_itr(struct adapter *adapter, uint32_t itr)
4107 {
4108         E1000_WRITE_REG(&adapter->hw, E1000_ITR, itr);
4109         if (adapter->hw.mac.type == e1000_82574) {
4110                 int i;
4111
4112                 /*
4113                  * When using MSIX interrupts we need to
4114                  * throttle using the EITR register
4115                  */
4116                 for (i = 0; i < 4; ++i) {
4117                         E1000_WRITE_REG(&adapter->hw,
4118                             E1000_EITR_82574(i), itr);
4119                 }
4120         }
4121 }
4122
4123 static void
4124 em_disable_aspm(struct adapter *adapter)
4125 {
4126         uint16_t link_cap, link_ctrl, disable;
4127         uint8_t pcie_ptr, reg;
4128         device_t dev = adapter->dev;
4129
4130         switch (adapter->hw.mac.type) {
4131         case e1000_82571:
4132         case e1000_82572:
4133         case e1000_82573:
4134                 /*
4135                  * 82573 specification update
4136                  * errata #8 disable L0s
4137                  * errata #41 disable L1
4138                  *
4139                  * 82571/82572 specification update
4140                  # errata #13 disable L1
4141                  * errata #68 disable L0s
4142                  */
4143                 disable = PCIEM_LNKCTL_ASPM_L0S | PCIEM_LNKCTL_ASPM_L1;
4144                 break;
4145
4146         case e1000_82574:
4147         case e1000_82583:
4148                 /*
4149                  * 82574 specification update errata #20
4150                  * 82583 specification update errata #9
4151                  *
4152                  * There is no need to disable L1
4153                  */
4154                 disable = PCIEM_LNKCTL_ASPM_L0S;
4155                 break;
4156
4157         default:
4158                 return;
4159         }
4160
4161         pcie_ptr = pci_get_pciecap_ptr(dev);
4162         if (pcie_ptr == 0)
4163                 return;
4164
4165         link_cap = pci_read_config(dev, pcie_ptr + PCIER_LINKCAP, 2);
4166         if ((link_cap & PCIEM_LNKCAP_ASPM_MASK) == 0)
4167                 return;
4168
4169         if (bootverbose) {
4170                 if_printf(&adapter->arpcom.ac_if,
4171                     "disable ASPM %#02x\n", disable);
4172         }
4173
4174         reg = pcie_ptr + PCIER_LINKCTRL;
4175         link_ctrl = pci_read_config(dev, reg, 2);
4176         link_ctrl &= ~disable;
4177         pci_write_config(dev, reg, link_ctrl, 2);
4178 }
4179
4180 static int
4181 em_tso_pullup(struct adapter *adapter, struct mbuf **mp)
4182 {
4183         int iphlen, hoff, thoff, ex = 0;
4184         struct mbuf *m;
4185         struct ip *ip;
4186
4187         m = *mp;
4188         KASSERT(M_WRITABLE(m), ("TSO mbuf not writable"));
4189
4190         iphlen = m->m_pkthdr.csum_iphlen;
4191         thoff = m->m_pkthdr.csum_thlen;
4192         hoff = m->m_pkthdr.csum_lhlen;
4193
4194         KASSERT(iphlen > 0, ("invalid ip hlen"));
4195         KASSERT(thoff > 0, ("invalid tcp hlen"));
4196         KASSERT(hoff > 0, ("invalid ether hlen"));
4197
4198         if (adapter->flags & EM_FLAG_TSO_PULLEX)
4199                 ex = 4;
4200
4201         if (m->m_len < hoff + iphlen + thoff + ex) {
4202                 m = m_pullup(m, hoff + iphlen + thoff + ex);
4203                 if (m == NULL) {
4204                         *mp = NULL;
4205                         return ENOBUFS;
4206                 }
4207                 *mp = m;
4208         }
4209         ip = mtodoff(m, struct ip *, hoff);
4210         ip->ip_len = 0;
4211
4212         return 0;
4213 }
4214
4215 static int
4216 em_tso_setup(struct adapter *adapter, struct mbuf *mp,
4217     uint32_t *txd_upper, uint32_t *txd_lower)
4218 {
4219         struct e1000_context_desc *TXD;
4220         int hoff, iphlen, thoff, hlen;
4221         int mss, pktlen, curr_txd;
4222
4223         iphlen = mp->m_pkthdr.csum_iphlen;
4224         thoff = mp->m_pkthdr.csum_thlen;
4225         hoff = mp->m_pkthdr.csum_lhlen;
4226         mss = mp->m_pkthdr.tso_segsz;
4227         pktlen = mp->m_pkthdr.len;
4228
4229         if (adapter->csum_flags == CSUM_TSO &&
4230             adapter->csum_iphlen == iphlen &&
4231             adapter->csum_lhlen == hoff &&
4232             adapter->csum_thlen == thoff &&
4233             adapter->csum_mss == mss &&
4234             adapter->csum_pktlen == pktlen) {
4235                 *txd_upper = adapter->csum_txd_upper;
4236                 *txd_lower = adapter->csum_txd_lower;
4237                 return 0;
4238         }
4239         hlen = hoff + iphlen + thoff;
4240
4241         /*
4242          * Setup a new TSO context.
4243          */
4244
4245         curr_txd = adapter->next_avail_tx_desc;
4246         TXD = (struct e1000_context_desc *)&adapter->tx_desc_base[curr_txd];
4247
4248         *txd_lower = E1000_TXD_CMD_DEXT |       /* Extended descr type */
4249                      E1000_TXD_DTYP_D |         /* Data descr type */
4250                      E1000_TXD_CMD_TSE;         /* Do TSE on this packet */
4251
4252         /* IP and/or TCP header checksum calculation and insertion. */
4253         *txd_upper = (E1000_TXD_POPTS_IXSM | E1000_TXD_POPTS_TXSM) << 8;
4254
4255         /*
4256          * Start offset for header checksum calculation.
4257          * End offset for header checksum calculation.
4258          * Offset of place put the checksum.
4259          */
4260         TXD->lower_setup.ip_fields.ipcss = hoff;
4261         TXD->lower_setup.ip_fields.ipcse = htole16(hoff + iphlen - 1);
4262         TXD->lower_setup.ip_fields.ipcso = hoff + offsetof(struct ip, ip_sum);
4263
4264         /*
4265          * Start offset for payload checksum calculation.
4266          * End offset for payload checksum calculation.
4267          * Offset of place to put the checksum.
4268          */
4269         TXD->upper_setup.tcp_fields.tucss = hoff + iphlen;
4270         TXD->upper_setup.tcp_fields.tucse = 0;
4271         TXD->upper_setup.tcp_fields.tucso =
4272             hoff + iphlen + offsetof(struct tcphdr, th_sum);
4273
4274         /*
4275          * Payload size per packet w/o any headers.
4276          * Length of all headers up to payload.
4277          */
4278         TXD->tcp_seg_setup.fields.mss = htole16(mss);
4279         TXD->tcp_seg_setup.fields.hdr_len = hlen;
4280         TXD->cmd_and_length = htole32(E1000_TXD_CMD_IFCS |
4281                                 E1000_TXD_CMD_DEXT |    /* Extended descr */
4282                                 E1000_TXD_CMD_TSE |     /* TSE context */
4283                                 E1000_TXD_CMD_IP |      /* Do IP csum */
4284                                 E1000_TXD_CMD_TCP |     /* Do TCP checksum */
4285                                 (pktlen - hlen));       /* Total len */
4286
4287         /* Save the information for this TSO context */
4288         adapter->csum_flags = CSUM_TSO;
4289         adapter->csum_lhlen = hoff;
4290         adapter->csum_iphlen = iphlen;
4291         adapter->csum_thlen = thoff;
4292         adapter->csum_mss = mss;
4293         adapter->csum_pktlen = pktlen;
4294         adapter->csum_txd_upper = *txd_upper;
4295         adapter->csum_txd_lower = *txd_lower;
4296
4297         if (++curr_txd == adapter->num_tx_desc)
4298                 curr_txd = 0;
4299
4300         KKASSERT(adapter->num_tx_desc_avail > 0);
4301         adapter->num_tx_desc_avail--;
4302
4303         adapter->next_avail_tx_desc = curr_txd;
4304         return 1;
4305 }